Digitized by the Internet Archive

in 2016

https://archive.org/details/bothaliavolume1111unse

MARY GUNN LIBRARY

NATIONAL BOTANICAL INSTITUTE

private BAG X 101

PRETORIA

Republic of sc

MARY GUNN LIBRARY

0000017921

South African National

Biodiversity Inst i tut

BOTHALIA

Volume 11

Published by the

Botanical Research Institute, Pretoria

Department of Agricultural Technical Services

Gepubliseer deur die

Navorsingsinstituut vir Plantkunde, Pretoria

Departement van Landbou-tegniese Dienste

Edited by — Onder redaksie van

L. E. Codd (No. 1-3)

D. J. B. Killick (No. 4)

587

INDEX - INDEKS

Acacia Miller, 107, 103, 109, 110, 111, 112, 126, 231,247, 299,443,

445, 457, 463, 473, 476, 481, 498

abyssinica sensu Brenan, 131

accola Maiden & Betche, 469

adunca A. Cunn. ex G. Don, 469

albida Del., 109, 110, 111, 112, 113, 231, 476

cimboensis Schinz, 131

andongensis Welv. ex Hiern, 291

ankokib Chiov., 453, 461

arenaria, 476

armata R. Br., 466

asak (Forsk.) mild., 303, 453, 458

asak (Forsk.) Willd.

var. unispinosa, 303, 459

asak sensu auct. mult., 457

ataxacantha DC., 110, 112, 113, 127, 476

baileyana F. Muell., 466

blommaertii De Wild., 130

borleae, 476

brevispica Harms, 112, 476

bricchettiana Chiov., 299

burkei, 234, 476

bussei Harms, 1 1 3

caffra (Thunb.) Willd., 127, 337, 338, 476, 553, 570, 571, 572,

573, 575

var. tomentosa/3. E. Glover, 127

var. transvaalensis P. E. Glover, 127

callicoma Meisn., 231

caraniana Chiov., 453, 461

caven (Mol.) Mol., 1 1 1

cheilanthifolia Chiov., 299, 453, 458

var. cheilanthifolia, 458

var. hirtella Chiov., 458

circummarginata Chiov., 456

clavigora, 476

condyloclada Chiov., 453, 460

cufodontii Chiov., 300, 457

cultriformis A. Cunn. ex G. Don, 469

cyanophylla Lindl., 467

cyclops A. Cunn. ex G. Don, 463, 468

davyi, 127, 476

dealbata Link , 465

decurrerfs Willd., 466

decurrens sensu Bak. f., 465

dekindtiana A. Chev., 443

detinens, 109

dolichocephala Harms, 109, 113

drepanolobium Harms ex Sjostedt, 292

dudgeoni Craib ex Holl., 453, 454

edgeworthii T. Anders., 300, 445

ehrenbergiana Hayne, 445

elata A. Cunn. ex Benth., 469

elephantina Burch., 247, 249

elephantorrhiza DC., 247, 249

eriocarpa Brenan, 111, 112, 231

erioloba E. Mey., 443, 444, 445, 476, 550

erioloba E. Mey. x Acacia haematoxylon Willd., 444, 477

erubescens Welw. ex Oliv., 127, All

etbaica Schweinf.

var. hirta A. Chev., 291

exuvialis, 477

farnesiana (L.) Willd., Ill, 463, 465, 471

fimbriata A. Cunn. ex G. Don, 469

fischeri Harms, 291

fistula Schweinf., 445

fleckii, 477

galpinii Burtt Davy, 109, 110, 477

gerrardii Benth., All

var. gerrardii, 291

giraffae sensu auct. mult., 444

giraffae Willd., 1 10, 1 1 1, 1 12, 443, 444, All

giraffae Willd. x A. haematoxylon Willd., 444

glaucophylla Steud. ex A. Rich., 458

gloveri Gilliland, 299

goetzei Harms, 234, 291, 293

subsp. goetzei, 234

goetzei Harms

subsp. microphylla Brenan, 291, 292

gorinii Chiov., 300

gossweileri Bak. f., 291, 292

grandicornuta, All

haematoxylon Willd., 112,477

hamulosa Benth., 301, 447, 453, 457

hebeclada

subsp. hebeclada. All

subsp. tristis, 477

hebeclada sensu Bews, 131

hebeclada sensu Wood, 131

hereroensis Engl., 127, 477

hirtella E. Mey.

var. inermis Walp., 445

hockii De Wild., 294

horrida (L.) Willd., 113, 127

horrida sensu Gossw., 443

humifusa Chiov., 300

hunteri Oliv., 453, 454

karroo Hayne, 127, 167, 171, 212, 213, 345, 443, 445, 446, 447

katangensis De Wild., 131

kirkii. All

kraussiana, 477

laeta R. Br. ex Benth., 109

lahai Steud. & Hochst., 113

lathouwersii Staner, 292

lasiopetala sensu Burtt Davy, 1 3 1

longifolia (Andr.) Willd., 467, 470

var. longifolia, 468

var. sophorae (Labill.) F. Muell. ex Benth., 468

luederitzii Engl., 233, 234

var. luederitzii. All

var. retinens, 234, All

macalusoi Mattei, 292

maidenii F. Muell., 470

manubensis J. H. Ross, sp. nov., 292

mearnsii De Willd., 463, 465

melanoxylon R. Br., 468, 469

mellifera (Valil) Benth., 109, All

subsp. detinens (Burch.) Brenan, 109, 555

subsp. mellifera, 109

manga De Wild., 131

montis-usti. All

nebrownii, 478

nefasia (Hochst. ex A. Rich.) Schweinf., 131

nefasia sensu Lebrun, 130

nervosula Chiov., 301

nigrescens Oliv., 234, 293, 478, 557

nilotica (L.) Willd., 107, 1 10, 1 1 1, 449, 478

nubica Benth., 300

ogadensis Chiov., Ill, 453, 460

oliveri Vatke, 303, 453, 459

oxyosprion Chiov., 301

var. oxyosprion, 301, 457

paradoxa Chiov., 301, 457

pendula A. Cunn. ex G. Don, 470

pennata, 109, 110, 111

permixta, 478

persiciflora Pax, 109, 110

pervillei Benth., 127

petersiana Bolle, 127

pinetorum F. J. Hermann, 472

podalyriifolia A. Cunn. ex G. Don, 466

polyacantha, 478

pseudonigrescens Brenan & J. H. Ross, 293

puccioniana Chiov., 301

purpurascens Vatke, 130

pycnantha Benth., 467

redacta J. H. Ross, sp. nov., 231, 233

reficiens Wawra, 233, 234, 478

subsp. misera (Vatke) Brenan, 302

var. reficiens, 233, 234

rehmanniana sensu Exell, 253

rehmanniana sensu Hutch. & Dalz., 131, 478

retinens Sim, 234

retinodes Schlechtd., 469

robecchii Pirotta, 445

robusta, 478

robusta sensu Oliv., 443

robynsiana, 478

rostrata Sim, 457

rovumae Oliv., 127, 234, 292

rupestris Stocks ex Boiss., 456

saligna (Labill.) Wendl., 463, 467, 468

samoryana A. Chev., 454

schlechteri Harms, 234

schweinfurthii, 478

588

Senegal (L.) mild ., 1 10, 292, 300, 449, 453, 455

var. kerensis Schweinf, 456

var. leiorhachis Brenan, 456, 478

var. platyosprion Chev., 302, 456

var. pseudoglaucophylla, 302, 457

var. rostrata Brenan , 301, 302, 457, 478

var. Senegal, 302, 455

var. Senegal sensu Brenan, 456

seyal Del., 294

var. kassonionga De Wild., 294

var. lescrauwaetii De Wild., 294

var. seretii De Wild., 294

sieberana DC., 128, 130, 294

subsp. sieberana,

var. sieberana, 130, 294

subsp. vermoesenii Troupin,

var. vermoesenii (De Wild.) Keay & Brenan, 131

var. woodii (Burtt Davy) Keay & Brenan, 131

var. orientate Troupin, 128, 129, 131

var. rehmanniana (Schinz) Roberty, 131

var. sieberana, 128, 130

var. vermoesenii (De Wild.) Keay & Brenan, 128, 129, 131

var. villosa A. Chev., 128, 131

var. woodii (Burtt Davy) Keay & Brenan, 128, 129, 131, 478

sing Guill. & Perr., 130

somalensis Vatke, 300, 453, 459

spinosa E. Mey., 267

spinosa Marloth & Engl., nom. illegit., 457

spirocarpa Hochst. ex A. Rich., 127

stenocarpa Hochst. ex A. Rich., 294

stefanini Chiov., 302

stolonifera Burch., 131

stuhlmannii, 478

swazica, 478

tenuispina, 478

terminalis (Salisb.) Macbride, 469

thomasii Harms, 453, 460

tortilis ( Forsk .) Hayne, 127

subsp. heteracantha, 478

subsp. raddiana (Savi) Brenan,

var. pubescens A. Chev., 294

var. raddiana, 294

subsp. spirocarpa (Hochst. ex A. Rich.) Brenan, 127, 445, 478

var. lenticellosa Chiov., 294

var. pubescens Abner ex Burtt Davy, 294

triacantha Hochst. ex A. Rich., 458

trispinosa Marloth & Engl., 457

tristis Welw. ex Oliv., 131

subsp. tristis Schreib., 131

unispinosa (Fiori) Chiov., 259, 303

verek Guill. & Perr., 456

vermoesenii De Wild., 131

verrugera Schweinf., 130, 131

var. subinermis A. Chev., 294

virchowiana Vatke & Hildebr., 456

viscidula A. Cunn. ex Benth., 470

visite Griseb., 469

volkii Suesseng., 457

welwitschii Oliv., 234, 479

subsp. delagoensis (Harms) J. H. Ross & Brenan, 234

woodii Burtt Davy, 131

xanthophloea, 479

zanzibarica (S. Moore) Taub.

var. microphylla Brenan, 445

zizyphispina Chiov., 453, 461

Acalypha glabrata, 569

Achaetomium, 21 1

globosum, 21 1

Achaetomium

strumarium Rai, Tewari & Maker ji, 211

Acremomium,

curvulum Gams, 207

persicinum ( Nicot ) Gams, 207

Acremomium

strict um Gams, 207

Acroceras macrum Stapf, 286

Agapanthus, 27

Albizia,

harveyi Foam., 127

obbiadensis (Chiov.) Brenan, 301

ogadensis (Chiov.) Bak.f. ex Chiov., 460

Alloteropsis, 273

cimicina (L.) Stapf, 273, 296

semialata (R. Br.) Hitchc., 273, 274, 286, 287, 342, 343

Aloe, 119, 122

albida (Stapf ) Reyn., 1 19

bowiea Roem. & Schult.f, 119

buettneri Berg., 1 19

dewinteri Giess, sp. nov., 120, 122

dolomitica, 1 19

kniphofioides Bak., 119

modesta Reyn., 1 19

myriacantha (Haw.) Roem. & Schult.f, 119

namibensis Giess, 120

niebuhriana Lavranos, 1 22

sessiliflora, 119

Aloineae, 119

Alternaria, 21 1

consortiale (Thiim.) Groves & Hughes, 209

Amaryllis, 27, 28

belladonna, 42, 51

bulbisperma Burm.f., 50, 51

capensis L., 51

cape ns is Mill., 51

forbesii Lindl., 35, 45, 51

var. purpurea Lindl., 44

longi folia L., 51

var. riparia Ker-Gawl., 50

longifolia sensu Jacq. et al., 50

ornata Ait., 38

revoluta L’Herit., 41

var. gracilior, 41

revoluta sensu Ker-Gawl., 46

variabilis Jacq., 41

Amethystanthus Nakai, 117, 373, 436

Amethystoides, 117

Ammocharis, 28, 30

Ammocharis sensu Milne-Redh. & Schweick., 28

baumii (Harms) Milne-Redh. & Schweick., 28, 30

Amyris sensu L., 54

Anaphrenium concolor E. Mey., 277

Anthephora, 285

aequiglumis, 285

Anthericum,

pubescens Nees, 285

brevicaule Bak., 123

brevifolium Thunb., 123

contortum L.f., 123

dregeanum (Kunth) Bak., 123

scilliflorum Eckl. ex Bak., 123

zeyheri Bak., 123

Anthospermum rigidum, 343

conchata (Lloyd) Reid, comb, nov., 222

Antrodia Karst, sensu Donk, 222

serialis (Fr.) Donk, 222

Apodytes dimidiata, 156

Aristida, 273, 305

aequiglumis, 573

junciformis, 578

Arodes Kuntze, 8

aethiopicum (L.) Kuntze, 8, 9

albomaculatum (Hook.) Kuntze, 8, 18

angustilobum (Schott) Kuntze, 8, 18

hastatum (Hook.) Kuntze, 8, 18

melanoleucum (Hook.f.) Kuntze, 8, 18

Aroides Heist, ex Fabric., 7, 8

aethiopicum (L.) Heist, ex Fabric., 9

albomaculatum (Hook.) Kuntze, 18

angustilobum (Schott) Kuntze, 18

hastatum (Hook.) Kuntze, 18

Artemisia caffra, 337

Arum L., 5

aethiopicum Herm., 6

africanum Herm., 6

Arundinella nepalensis, 578

Ascocarydion G. Tayl., 374, 381

Asparagus krebsianus, 577

suaveolens, 342

Asperifolia Berger, 122

Asphades undulata (Thunb.) DC., 518

Astephanus frutescens E. Mey., 277

Astroloba, 119

Avena sativa L., 217

Babiana, 281

villosula, 281

Balessan Bruce, 54

Balsamea Gled., 54

angolensis (Engl.) Hiern, 90

capensis (Sond.) Engl., 96

harveyi Engl., 76

mollis (Oliv.) Engl., 73

pilosa Engl., 68

schimperi (O. Berg) Engl., 65

zanzibarica Baill., 85

Balsamodendrum Kunth, 53, 54

africanum (A. Rich.) Arn., 68

africanum sensu Oliv., 65

capense Sond., 96

/nolle Oliv., 73

schimperi O. Berg, 65

Balsamophloeos O. Berg, 54

Beilschmiedia, 118

natalensis J. H. Ross, sp. nov., 118

589

Bequaertiodendron magalismontanum, 565, 569, 577

natalense, 533, 534

Berula thunbergii, 345

Blumea alata, 571, 575

Boscia undulata Thunb., 518

Bowiea Harv. ex Hook, f., 119

africana Haw., 119

myriacantha, 119

volubilis Harv. ex Hook, f, 1 19

Brachiaria, 273

serrata, 571, 572

Brachyachne, 306

Bromus diandrus Roth., 296

Buddleia salviifolia, 345

Bulbine brevifolia (Thunb.) Roem. & Schult. f., 123

Btilbinella brevifolia (Thunb.) Kunth, 123

Bulbostylis bu rchellii, 340, 342

Burkea africana, 157, 554, 573, 574

Burnatastrum Briq., 282, 372, 374, 380

spicatum (E. Mey. ex Benth.) Briq., 374, 383

Caesia R. Br., 123

brevicaulis (Bak.) Dur. & Schinz, 123

brevifolia (Thunb.) Dur. & Schinz, 123

capensis (H. Bol.) Oberm., comb, nov., 123

contorta (L.f.) Dur. & Schinz, 123

dregeana Kunth, 123

eckloniana Roem. & Schult. f., 123

scilliflora (Eckl. ex Bak.) Dur. & Schinz, 123

thunbergii Roem. & Schult. f., 123

zeyheri (Bak.) Dur. & Schinz, 123

Calla L., 6, 7

aetliiopica L., 6, 7, 9, 10

elliottiana Watson, 14

ocidata Lindl., 8, 18, 21

pentlandii Whyte ex Watson, 15

Calodendron capense, 157

Canthium,

inerme (L.f.) Kimtze, 491, 493

mundianum Cham. & Schlechtd., 493

ventosum (L.) Kuntze, 491, 493

Caralluma,

caudata N.E. Br., 133, 137

subsp. caudata 133, 134, 135

subsp. rhodesiaca Leach, subsp. nov., 133, 134, 136

var. chibensis (LuckhofF) LuckhofF, 134

var. fusca LuckhofF, 1 34

var. milleri Nel, 134, 135

var. stevensonii, 134

caudata N.E. Br. X Stapelia kwebensis N.E. Br., 136, 137

chibensis LuckhofF, 134

lateritia N.E. Br., 136

longecornuta Croizat ex Gomes e Sousa, 134

lutea N.E. Br., 136

praegracilis Oberm., 134

Cassia siamea, 143, 145

Cassine, 277

eucleiformis (Eckl. & Zeyh.) Kuntze, 277

eucleiforme sensu Davison, 277

Cavacoa aurea, 533

Celastrus filiformis L.f., 277

pyracanthus L., 115

Celtis africana, 167, 171, 569

kraussiana, 227

rhamnifolia, 144, 221

Cerebella, 215

andropogonis Ces. apud Rabenhorst, 215

cynodontis Syd., 215

Ceriops Am., 106

Chamaealoe, 119

afr cana (Haw.) Berger, 119

Chloris radiata (L.) Swartz, 295

virgata Swartz, 295

Choiromyces echinulatus Trappe & Marasas, sp. nov., 139

Chortolirion, 119

angolense (Bak.) Berger, 119

stenophvllum (Bak.) Berger, 119

subspicatum (Bak.) Berger, 119

tenuifolium (Engl.) Berger, 119

Chrysocoma tenuifolia, 168, 171

Chrysopogon montanus, 338

Cladonia pyxidata, 228

Cladosporium Link ex Fr., 218

resinae (Lindau) de Vries, 218

Clavaria,

cladoniae Kalchbr., 221, 228

pyxidata Pers., 228.

Claviceps purpurea (Fr.) Tvl., 215

Codd, L. E. Author citation for Putterlickia pyracantha, 115

Codd’ L. E. The genus Rabdosia in South Africa, 117

Codd', L. E. The identity of Erythrina princeps, 269-271

Codd' L. E. New species of Plectranthus, 282-284

Codd, L. E. Plectranthus (Labiatae) and allied genera in Southern

Africa, 371-442

Codonocrinum Bak., 27, 38

Coetzee, B. J. Improvement of association-analysis classification by

Braun-Blanquet technique, 365-367

Coetzee, B. J. A phytosociological classification of the vegetation

of the Jack Scott Nature Reserve, 329-346

Coetzee, B. J. A phytosociological classification of the Rustenburg

Nature Reserve, 561-580

Coetzee, B. J. & Werger, M. J. A. On hierarchical syndrome analysis

and the Ziirich-Montpellier table method, 159-164

Coetzee, B. .1. & Werger, M. J. A. A west-east vegetation transect

through Africa south of the Tropic of Capricorn, 539-560

Cola natalensis, 533

Coleochloa, 306

pallidior, 306

setifera, 306, 577

Coleus Lour., 282, 373, 374, 387

sect. Aromaria Benth., 373, 387

sect. Calceolus Benth., 373, 387, 389

sect. Solenostemoides Briq., 373, 438

sect. Solenostemon Briq., 373, 387, 437

amboinicus Lour., 373, 374, 388

var. violaceus Guerke, 388

aromaticus Benth., 387, 388

barbatus (Andr.) Benth., 394

blumei Benth., 373, 438, 440

caninus (Roth) Vatke, 390

carnosus Dinter ms., 392

comosus Hochst. ex Guerke, 393

crassifolius Benth., 388

dazo A. Chev., 377

decumbens Guerke, 390

dysentericus Bak., 438

esculentus (N.E. Br.) G. Tayl., 377

flavovirens Guerke, 490

floribunius (N.E. Br.) Robyns & Lebrun, 377

floribundus (N.E. Br.) Robyns & Lebrun

var. longipes (N.E. Br.) Robyns & Lebrun, 377

forskolilii sensu Briq., 394

gazensis S. Moore, 382

latifolius Hochst. ex Benth., 439

leucophyllus Bak., 381

madagascariensis (Pers.) A. Chev., 403

matopensis S. Moore, 382

mirabilis Briq., 381

var. buchnerianus Briq., 381

var. hypisodontus Briq., 381

var. mechowianus Briq., 381

var. poggeanus Briq., 381

myrianthus (Briq.) Brenan, 382

neochilus (Schltr.) Codd, 392

omahekense Dinter, 390

pentheri Guerke, 392

polyanthus S. Moore, 382

rehmannii Briq., 439

rotundifolius (Poir.) A. Chev., 438

schinzii Guerke, 392

scutellarioides (L.) Benth., 387, 440

shirensis Bak., 440

spicatus Benth., 390

spicatus sensu A. Rich., 393

tetensis Bak., 390

vagatus E. A. Bruce, 390

zatarhendi (Forsk.) Benth., 398

Colocasia Link, 6, 7, 8

aetliiopica (L.) Link, 7, 9

Combretum apiculatuni, 555

zeyheri, 570, 571, 573, 574, 575

Commiphora Jacq., 53, 54, 55

africana (A. Rich.) Engl., 68

var. africana, 71

var. rubriflora (Engl.) Willd., 71

angolensis Engl., 90

berberidifolia Engl., 57

betschuanica Engl., 65

boehmii Engl., 73

calciicola Engl., 68

capensis (Sond.) Engl., 96

caryaefolia Oliv., 83

cervifolia Van der Walt, 99

chlorocarpa Engl., 81

cinerea Engl., 73

dekindtiana Engl., 73

edulis (Klotzsch) Engl., 81

glandulosa Schinz, 55, 57, 60

gossweileri Engl., 90

gracilifrondosa Dinter ex Van der Walt, 95

harveyi (Engl.) Engl., 76, 85

heterophylla Engl., 73

iringensis Engl., 73

590

krausei Engl., 73

loandensis Engl., 68

longebracteata Engl., 90

lugardae N.E. Br., 57

madagascariensis Jacq., 54, 55

marlothi Engl., 78

merkeri Engl. , 55, 63

mollis ( Oliv .) Engl., 73

montana Engl., 73

namaensis Schinz, 92

ndemfi Engl., 74

neglecta Verdoorn, 71

nigrescens Engl., 90

nkolola Engl., 68

oliveri Engl., 90

pilosa (Engl.) Engl., 68

pyracanthoides Engl., 55, 60

subsp. glandnlosa (Schinz) Wild, 57

rangeana Engl., 96

rehmannii Engl., 90

rotundifolia Dinter & Engl., 92

rubri flora Engl., 68

sambesiaca Engl., 68

schimperi ( O . Berg.) Engl., 65, 68

seineri Engl., 57

spondioides Engl., 85

stuhlmannii Engl., 73

tenuipetiolata Engl., 87

viminea Burtt Davy, 63

welwitschii Engl., 73

woodii Engl., 83, 85

zanzibarica (Bail!.) Engl., 85

Coniophora DC. ex Merat, 227

Coniothyrium glomeratum Corda, 2 1 1

Conium chaerophylloides (Thunb.) Eckl. & Zeyh., 521

maculatum L., 521

Conyza podocephala, 344

Coriolopsis occidentalis ( Klotzsch ) Murr., 227

Coriolus Quel., 223

durbinensis (Van der Byl) Reid , comb, nov., 223, 227

Crassula transvaalensis, 342

Crinum L., 27, 28

acaule Bak., 39, 40

africanum L., 27

algo ease Herb., 41

americanum L., 27, 28, 38

aquaticum Burch., 34

asiaticum L., 27

baumii Harms, 30

bulbispermum (Bunn, f.) Milne-Redh., 42, 49

buphanoides Welw. ex Bak., 31, 33

campanulatum Herb., 27, 34

capense (Mill.) Herb., 49, 51

capense sensu Herb., 50

carolo-schmidtii Dinter, 37

colenso Hort., 38

crassicaule Bak., 32

crassifolium Herb., 41

crispum Phillips, 45, 46

delagoense Verdoorn, 44, 45

euchrophyllum Verdoorn, 33

foetidum Verdoorn, 27, 28, 42

forbesianum Herb., 44, 45

var. punica Herb., 44

forbesii (Lindl.) Schultes f., 27, 44, 51

forbesii sensu Van der Walt, 35

giganteum Andr., 38

gouwsii Traub, 46, 49

graminicola Verdoorn, 43

imbrication Bak., 38

jagus (Thompson) Dandy, 37 38

kirkii Bak., 38

latifolium L., 27

leucophyllum Bak., 31

longifolium (L.) Thunb., 42, 50, 51

lineare L.f, 41, 42

lugardiae N.E. Br., 45

mackenii Hort., 38

macowanii Bak., 42, 45, 46, 51

subsp. confusion Verdoorn, 42, 46, 49

moorei Hook. /., 27, 28, 37, 49

minimum Milne-Redh., 40

natalense Hort., 38

nerinoides Bak., 30

occiduale R. A. Dyer, 37

ornatum Bury, 38, 39

paludosum Verdoorn, 35, 37

parvibulbosum Dinter ex Overkott, 40

polyphyllum Bak., 45, 46

rautanenianum Schinz, 35, 36, 37

revolution (L’Herit.) Herb., 41

ripariion Herb., 50

subcernuum Bak., 33

variabile (Jacq.) Herb., 41, 49, 51

zeylanicum L., 43

Croton gratissimus, 565, 577

Cryptocarya, 118

Cryptolepis oblongifolia, 575

Curtisia dentata, 156

Cussonia Thunb., 191, 193

sect. Capitata Strey, sect, nov., 193

sect. Cussonia, 193

subgen. Cussonia, 193

sect. Neocussonia Harms, 193

subgen. Paniculata Strey, subgen. nov., 193

subgen. Protocussonia Strey, subgen. nov., 193

angolensis Hiern, 193

arborea Hochst. ex A. Rich., 193

arenicola Strey, sp. nov., 193, 194, 197

chartacea Schinz 194, 201

gamtoosensis Strey, sp. nov., 515, 516

holstii Harms ex Engl., 193

kraussii Hochst., 194

natalensis Sond., 193

nicholsonii Strey, 193, 194, 197, 198

paniculata Eckl. & Zeyh., 193, 339, 340

sphaerocephala Strey, 191, 192, 193, 194, 198, 199, 201

spicata Thunb., 191, 192, 193, 194, 198, 201, 515

thyrsiflora Thunb., 193, 194, 195, 197

umbellifera (Sond.) Baill., 194

zuluensis Strey, 193, 194, 195, 197, 515

Cybistetes longifolia (L.) Milne-Redh. & Schweick., 51

Cymbopogon excavatus, 342

Cynodon dactylon, 344

Cyperus,

bellus, 306

obtusiflorus, 343

rupestris, 577, 578

sexangularis, 345

Daedalea Pers. ex Fr., 221

dregeana, 226

eatonii, 226

macowanii Kalchbr. ex Theum., 221

palisotii Fr., 221

Danthonia DC., 106

De Winter, B. The identity of some species of Hermannia represen-

ted in the Linnaean, Thunberg, Bergius and Cavanilles Herbaria,

263-264

De Winter, B. A new combination in Ozoroa, 277

De Winter, B. & Vorster, P. Comments on the treatment of the

Poaceae in the Prodromus einer Flora von Siidwestafrika (1970),

295-297

Dichrostachys, 1 10, 301

caffra Meisn. ex Benth., 267

cinerea (L.) Wight & Am., 265

subsp. nyassana (Taub.) Brenan, 249, 265

subsp. africana Brenan & Brummitt,

var. africana, 265

var. lugardiae (N.E. Br.) Brenan & Brummitt, 265

var. plurijuga Brenan & Brummitt, 265

var. setulosa (Welw. ex Oliv.) Brenan & Brummitt, 265

var. pubescens Brenan & Brummitt, 265

subsp. argillicola Brenan & Brummitt,

var. hirtipes Brenan & Brummitt, 265

subsp. forbesii (Benth.) Brenan & Brummitt, 265

nutans (Pers.) Benth., 267

Digitaria, 285

sect. Erianthae, 285

brazzae, 575, 576, 578

diagonalis, 343

monodactyla, 340

otaviensis, 285

paspalodes Michx., 244

smutsii, 570, 571

Dinebra retroflexa ( Vald ) Panz., 297

Diospyros whyteana, 569

Diplachne biflora Hack, ex Schinz, 296, 573

eleusine Nees, 296

Diporidium serrulatum Hochst., 517

Dombcya rotundifolia, 569

Drechslera, 513

multiformis Jooste, sp. nov., 511

Drypetes gerrardii, 534

Du Toit, P. C. V. The identity of Ochna atropurpurea, 517

Du Toit, P. C. V. A change in status and a new name for Ochna

atropurpurea var. angustifolia, 517

Du Toit, P. C. V. A change in status for Ochna oconnorii, 518

Dyer, R. A. A new species of Euphorbia, 278-280

Dyer, R. A. The status of Strelitzia juncea, 519-520

591

Echinochloa frumentacea (Roxb.) Link., 295, 297

crusgalli (L.) Beanv., 295

Ekebergia capensis, 533

Elephantorrhiza Benth., 247, 257

burchellii Benth., 247, 249

burkei Benth., 251

dinteri Phillips ined., 249

elephantina (Burch.) Skeels, 247, 249, 250, 252, 575, 576, 578

var. burkei (Benth.) Merr., 251

goetzei (Harms) Harms, 252, 253, 254, 256

subsp. goetzei, 253

subsp. lata Brenan & Brummitt , 253, 254

obliqua Burtt Davy, 248

var. glabra Phillips, 249

var. obliqua, 248, 249

petersiana Bolle, 253

praetermissa J. H. Ross, sp. nov., 252

rangei Harms, 254, 255

rangei sensu Phillips, 249

rubescens Gibbs, 253

schinziana Dinter, 255

suffruticosa Sch nz, 255, 256

transvaalensis Phillips ined., 249

wahlbergii (Harv.) Phillips, 251

woodii Phillips, 259, 251

var. pubescens Phillips, 251

var. woodii, 251

Ellis, R. P. Comparative leaf anatomy of Paspalum paspalodes and

P. vaginatum, 235-241

Ellis, R. P. Smodingium dermatitis: The intercellular secretory

canals of the aerial axis and their relationship to this toxicity,

259-262

Ellis. R. P. Anomalous vascular bundle sheath structure in A/lo-

teropsis semialata leaf blades, 273-275

Ellis, R. P. Non-Kranz (C3) leaf anatomy in the panicoid grasses:

additional records from South Africa, 286-287

Elsholtzia Willd., 117, 373, 436

sect. Rabdosia Bl., 117, 373, 436

javanica Bl., 1 17, 373

Encephalartos villosus, 534

Englerastrum Briq., 373

floribimdus (N.E. Br.) Th. Fries jun., 377

var. longipes (N.E. Br.) Th. Fries jun., 377

schweinfurthii Briq., 373

Ensete ventricosum, 556

Entada, 125

abyssinica Steud. ex A. Rich., 257

arenaria Schinz, 126

subsp. arenaria, 126

subsp. microcarpa (Brenan) J. H. Ross, comb, nov., 126

nana Harms, 126

subsp. microcarpa Brenan, 126

natalensis Benth., 125, 126

var. aculeata, 125

phaseoloides (L.) Merr., 125

pursaetha DC., 125

scandens (L.) Benth., 125

schlechteri (Harms) Harms, 126

spicata (E. Mey.) Druce, 125, 126

wahlbergii, 125

Entolasia imbricata Stapf, 297

Epicoccum, 215

andropogonis (Ces.) Schol-Schwarz, 215

purpurascens, 215

Eragrostis, 305

habrantha Rendle, 296

hispida, 306

lehmanniana Nees, 168, 295

nindensis, 306, 577, 578

racemosa, 573

trichophora Coss. & Dur., 295

Erica L., 106

conferta Andr., 1 15

krugeri E. G. H. Oliver, sp. nov., 115, 116

obtusata Klotzsch, 115, 116

oliveri H. A. Baker, 115, 116

polifolia Salisb., 115

Eriochloa, 273

Eriudaphus, 280

ecklonii, 280

mundii, 280

zeyheri, 280

Erythrina,

caffra, 156

caffra sensu Ker, 271

hastifolia Bertol. f., 271

humeana Spreng., 270 271

humei E. Mey., 271

var. hastifolia (Bertol. f.) Bak.f., 271

var. raja (Meisn.) Harv., 271

lysistemon Hutch., 269

princeps A. Dietr., 269, 271

raja Meisn., 271

Eucalyptus spp., 143, 144

diversicolor, 157

maculata, 145, 157

Euclea crispa, 171

Euphorbia,

berotica N.E. Br., 505, 508

bottae Boiss., 506

cameronii N.E. Br., 506

carunculifera Leach, 495

celata R. A. Dyer, sp. nov., 278, 280

chersina N.E. Br., 510

consobrina N.E. Br., 506

damarana Leach sp. nov., 495, 509

gariepina Boiss., 280, 495

gossypina Pax, 505, 596

gregaria Marloth, 495, 498, 548

gregaria sensu Dinter, 500

gummifera Boiss., 495, 497

hamata Sweet, 495

implexa Stapf, 506

juttae Dinter, 280

lateriflora Schum. & Thonn., 505

lathyris L., 495

mauritanica L., 505

var. foetens (Dinter) White, Dyer & Sloane, 510

var. lignosa White , Dyer & Sloane, 505, 510

var. namaquensis N.E. Br., 510

merkeri N.E. Br., 506

monteiri Hook, f, 495

nubica N.E. Br., 506

sessiliflora E. Mey., 497

schimperi Presl, 506

tirucalli L., 495

tirucalli sensu Hiern, 508

Eustachys mutica, 337, 570

Exidia glandulosa, 227

Faidherbia A. Chev., Ill, 112

Festuca L., 106

Ficus pretoriae, 570

Fomes, 226

Funalia protea

var. imbricata (Berk.) Reid, 223, 225

Gaff, D. F. & Ellis, R. P. Southern African grasses with foliage that

revives after dehydration, 305-308

Ganoderma Karst., 224

colossus (Fr.) Bres., 143, 145, 147

mollicarnosum (Lloyd) Sacc. & Trott., 224

nigrolucidum (Lloyd) Reid, comb, nov., 224

Germanea Lam., 372, 374

crassifolia (Vahl) Poir., 398

laxiflora (Benth.) Hiern, 434

maculosa Lam., 372, 374

rotundifolia Poir., 438

urticifolia Lam., 372, 374, 415

Gladiolus florentiae Marl., 281

Goldblatt, Peter. Montbretiopsis: reduced to synonymy in Tritonia

(Iridaceae), 281

Gunn, M. D. John Hutchinson, O.B.E., F.R.S., LL. D., V.M.H.,

F.L.S., 1-3

Haworthia, 119

angolensis Bak., 119

graminifolia G. G. Smith, 119

marginata, 119

margaritifera, 119

stenophylla Hook., 119

subspicata Bak., 119

tenuifolium Engl., 119

Heeria Meisn., 277

argentea ( Thunb .) Meisn., 277

concolor (C. Presl) Kuntze, 277

Helichrysum, 106

adscendens, 342, 343

Hemizygia pretoriae, 342

Hemprichia Ehrenb., 54

Hermannia, 263

alnifolia L., 519

aspera, 288

biserrata L.f., 263

cernua, 264

ciliaris L.f, 263, 264

confusa Salter, 263

diffusa L.f, 263

glabrata L.f., 263

grossularifolia L., 263

helicoidea Verdoorn, sp. nov., 288

linearis (Harv.) Hochr., 263

litoralis Verdoorn sp. nov., 518

multiflora Jacq., 288, 519

592

muricata, 288

myrrhifolia Thunb., 263, 264

pilosula (Harv.) Hochr., 263

pinnata L., 263, 264

pinnatisecta Salter, 263, 264

var. auriculata Salter, 264

procumbens Cav., 264

subsp. myrrhifolia (Tlumb.) De IVint., comb, nov., 264

pulchella L.f., 264

repetenda, 288

rigida, 288

rugosa Adamson, 519

tenuifolia Sims, 263

verticil lata Berg., 264

vesicaria Cav., 263

Herpolirion capense H. Bol., 122

Hesperantha, 281

hantamensis, 281

rosea, 281

Heteropogon contortus, 341

Heudelotia A. Rich., 54

africana A. Rich., 68, 71

Hitzeria Klotzsch, 54

edulis Klotzsch, 81

Homalocheilos J. K. Morton, 117, 373, 436

Hyparrhenia dregeana ( Nees ) Stapf ex Stent, 296

tamba, 345

Hypoestes verticillaris, 568

Hypoxis, 51

Ilex mitis, 569

Inga nefasia Hochst. ex A. Rich., 131

Inhambanella (Engl) Dubard, 106

Inonotus ochroporus ( Van der Byl) Pegler, 225

Isodon (Schrad. ex Benth.) Kudo, 117, 373, 436

Jooste, W. J. A new species of Dreclislera on Tribulus terrestris,

511-513

J uncus oxycarpus, 345

punctorius, 345

Kalanchoe paniculata, 570

Killick, D. J. B. A note on the author citation for Eriudaphus mtindii,

basionym of Scolopia mundii, 280

Killick, D. J. B. Scolopia stolzii, a new record for South Africa,

280-281

Killick, D. J. B. A change of status in Scolopia, 515

Kohautia amatymbica, 340

Lachnocladium Lev., 228

Landolphia capensis, 565, 569, 577

Lapeirousia, 281

montana, 281

plicata, 281

Lasiosiphon burchellii, 343

Laurembergia, 116, 117

repens Berg., 116, 117

subsp. brachypoda (Hiern) Oberm. comb, et stat. nov., 117

tetrandra (Schott) Kanitz, 117

subsp. brachypoda (Welw. ex Hiern) Raynal, 117

veronicifolia (Bory) SchindL, 117

Leach, L. C. Stapelieae (Asclepiadaceae) from South Tropical

Africa, VII, 133-137

Leach, L. C. Euphorbia gummifera, E. gregaria and a new species

from Damaraland, 495-503

Leach, L. C. Notes on Euphorbia mauritanica, E. gossypina and

some related species with an amplified description of E. berotica,

505-510

Lenzites palisotii ( Fr .) Fr., 221

Leptochloa uniflora Hochst. ex A. Rich, 297

Lessertia pauciflora, 168

Letty, C. The genus Zantedeschia, 5-26

Leucospermum saxosum S. Moore, 287

Lolium multiflorum Lam., 215

Loudetia, 287

simplex, 337, 340, 573

superba De Not., 287

Lotononis tenella, 340

Loudetiopsis glabrata ( K . Schum.) Conert, 297

Loxton, A. E. The taxonomy of Paspalum paspalodes and P. vagi-

natum as represented in South Africa, 243-245

Loxton, A. E. A note on a possible bigeneric hybrid between Digi-

taria and Anthephora, 285-286

Lumnitzera Willd., 106

Lycium,

barbatum Thunb., 493

inerme L.f., 491, 492, 493

Macrosporium abietis Tengwall, 209

consortiale Thum., 209, 210

Majana amboinica (Lour.) Kuntze, 388

Mahernia, 263

bipinnata L„

var. acutifolia Harv., 263

var. glandulosa Harv., 263

biserrata (L.f.) Cav., 263

diffusa (L.f.) Jacq., 263

glabrata (L.f.) Cav., 263

linearis Harv., 263

ovata E. Mey. ex Turcz., 263

pilosula Harv., 263

var. latifolia, 263

pulchella (L.f.) Cav., 264

vernicata Burch., 264

verticillata L., 264

Marasas, W.F.O. & Trappe, James M. Notes on Southern African

Tuberales, 139-141

Marasas, W. F. O. & Bredell, Ingrid H. A new species of Phaeora-

midaria (Fungi Imperfecti : Dematiaceae) from South Africa, 217-

219

Maytenus heterophylla, 171, 172

Medicago sativa L., 217

Melissoides, 117

Merulius,

gelatinosus Lloyd, 227

lacrymans, 227

Michrochloa, 306

caffra, 306, 340

Mimosa asak Forsk., 458

capensis Burnt. /., 445, 446

decurrens Donn, 466

farnesiana L., 465, 471, 472

longifolia Andr., 467

nilotica, 446, 449

reticulata L„ 446

saligna Labill., 467

scorpioides, 446

senegalensis Forsk., 447, 456, 457

senegalensis Houtt., 447, 456

Senegal L., 447, 449, 455, 456

Mimusops zeyheri, 568

Monocymbium ceresiiforme (Nees) Stapf , 296, 297, 342, 576, 578

Montbretia, 281

Montbretiopsis, 281

florentiae (Marl.) L. Boh, 281

Moraea, 281

falcifolia, 281

Morris, J. W. An automatic bibliography indexing programme,

325-328

Morris, .1. W. Progress in the computerization of herbarium proce-

dures, 349-353

Morris, J. W. & Guillerm, J. L. The ecological profiles technique

applied to data from Lichtenburg, South Africa, 355-364

Musil, C. F., Grunow, J. O. & Bornman, C. H. Classification and

ordination of aquatic macrophytes in the Pongola River pans.

Natal, 181-190

Mystroxylon eucleaeforme, 277

filiforme Eckl. & Zevh., 277

Nanolirion, 122

capense (H. Bol.) Benth. & Hook., 122

Nenax microphylla, 171

Neomullera Briq., 374

damarensis S. Moore, 382

welwitschii Briq., 374

Newtonia Baill., 106

Obermeyer, A. A. A note on Laurembergia repens, 116-117

Obermeyer, A. A. New taxa in the Turneraceae, 288

Obermeyer, A. A. & Giess, W. Aloe, Chamaealoe, Hawortliia,

Astroloba, Poellnitzia and Chortolirion, 119

Ochna arborea Burch, ex DC. var. oconnorii (P/till.) Du Toit, stat.

nov., 518

atropurpurea DC., 517

var. angustifolia Phill., 517

gamostigmata Du Toit, stat. et nom nov., 517

jabot apita var. P L., 517

mossambicensis Klotzsch, 517

oconnorii Phill., 518

pulchra, 573, 574

serrulata (Hochst.) Walp., 517

squarrosa var. PL., 517

Ocimum madagascariensis Pers., 403

racemosum Thunb., 407

scutellarioides L., 439

tomentosum Thunb., 403

verticillatum L.f., 407

zatarhendi Forsk., 398

Olea africana 171, 172

capensis subsp. macrocarpa 156, 157

Olinia Thunb., 491

cymosa (L.f.) Thunb., 491

ventosa (L.) Cufod., 491

Oliver, E. G. H. Notes on the Ericaceae, 115-116

Oplismenus, 273, 286

Oropetium, 306

Otosma Rafin., 8

aethiopica (L.) Rafin., 9

Oxalis obliquifolia, 341

Oxyporus (Board. & Galz.) Donk, 227

subflavus (Lloyd) Reid, 227

Ozoroa Del., 277

concolor (C. Presl ex Sond.) De Wint., comb, nov., 277

crassinervia (Engl.) R. & A. Fernandes, 111

dispar (C. Presl) R. & A. Fernandes, 277

Panicum, 207, 273, 286

aequinerve Nees, 286

ecklonii Nees, 286

hymeniochilum Nees, 286

Iaticomum Nees, 286

natalense Hochst., 286

parvifolium Lam., 287

Papendorf, M. C. & Jooste, W. J. The mycoflora of wheat field

debris, Part 1, 207-210

Papendorf, M. C. & Jooste, W. J. New and interesting records of

South African fungi, Part VIII, 211-215

Paspalidium platyrrhachis C. E. Hubb., 296

Paspalum,

subgen. Disticha, 243

digit aria Poir., 244

distichum L., 235, 243, 245, 295

paspalodes (Michx.) Scribn., 235, 236, 238, 239, 240, 243, 244,

295

vaginatum Swartz, 235, 236, 238, 239, 240, 243, 244, 245, 295

Pellaea calomelanos, 340, 341, 343

Pentanisia angustifolia, 340

Pentaschistis Stapf, 106

Perenniporia Murr., 222

Peyronellaea,

glomerata (Corda) Goidanich, 211

musae Joly, 212

nainensis Tandon & Bilgrami, 212

Phaeoramularia, 217, 218

kellermaniana Marasas & Bredell, sp. nov., 217, 218, 219

paradoxa Muntanola, 218

unamunoi (Castell.) Munatola, 218

Phaseolus Pat., 143, 145, 147

manihotis Heim, 143, 145, 146, 224

Phellinus gilvus (Schw.) Pat., 226

Phoma alternariaceum Brooks & Searle, 21 1

capitulum Pawar, Mathur & Thirumalachar, 211, 212

glomerata (Corda) Wollenw. & Hochapf., 211

glumarum Ell. & Tracy, 213

jolyana Pirozynski & Morgan- Jones, 212

musae (Joly) Boerema, Dorenb. & Kest., 212

ostiolata Pawer, Mathur & Thirumalachar, 211, 212

var. brunnea Pawar, Mathur & Thirumalachar, 211-212

Phragmitis australis, 345

mauritianus, 578

Phyllanthus incurvis, 343

Picea sitchensis, 146

Pinus mughus, 146

Piptadenia goetzei Harms, 253

Pittosporum viridiflorum, 569

Platyaster, 27, 38

Plectranthus LHerit., 117, 282, 372, 374

sect. Amethystoides Benth., 117, 372, 436

subgen. Burnatastrum (Briq.) Codd, 380

subgen. Calceolanthus Codd, 389

sect. Coleoides Benth., 282, 372, 395

subgen. Coleus (Lour.) Codd, 387

sect. Germanea Benth., 372, 395, 406

sect. Heterocylix Benth., 372

sect. Isodon Schrad. ex Benth., 372, 380, 436

subgen. Isodon (Schrad. ex Benth.) Briq., 436

sect. Melissoides Benth., 372

subgen. Nodiflorus Codd, 376

sect. Plectranthus, 406

subgen. Plectranthus, 395

sect. Pyramidium Benth., 372, 436

subgen. Xerophilus Codd, 378

aegyptiacus C. Chr., 398

ambiguus (Bol.) Codd, 282, 429

amboinicus (Lour.) Spreng, 388

aromaticus (Benth.) Roxb., 388

arthropodus Briq., 415, 421

aff. arthropodus sensu Compton, 420

aurifer Dinter ex Launert, 382

barbatus Andr., 394

behri Compton, 416

biflorus Bak., 376

blumei (Benth.) Launert, 439

bolusii T. Cooke, 435

calycinus Benth., 1 17, 436

var. pachystachyus (Briq.) T. Cooke, 436

candelabriformis Launert, 380

caninus Roth, 389, 390, 393

charianthus Briq., 415, 421

ciliatus E. Mey. ex Benth., 414

coloratus E. Mey. ex Benth., 429

comosus Sims, 393, 394

cooper i T. Cooke, 418, 424

crassifolius Vahl, 398

cylindraceus Hochst. ex Benth., 385

densiflorus T. Cooke, 385

densus N.E. Br., 378

dinteri Briq., 396

dolichopodus Briq., 423

draconis Briq., 401

dregei Codd, 429

ecklonii Benth., 431

elegantulus Briq., 412

esculentus N.E. Br., 377

fioribundus N.E. Br., 377

var. longipes N.E. Br., 377

forskohlaei sensu Ait.f., 394

fruticosus L'Herit., 372, 374, 412, 415

galpinii Schltr., 415

glome rat us R. A. Dyer, 385

grallatus Briq., 418, 421

grandidentatus Dinter ms., 382

grandidentatus Guerke, 396, 401

hereroensis Engl., 382

hilliardiae Codd, 282, 428

hirtus Benth., 403

hylophilus sensu Cooke, 434

javanicus (Bl.) Benth., 117

krookii Guerke ex Zahlbr., 418

var. grand) folia T. Cooke, 418

kuntzei Guerke, 431

kuntzeanus Domin, 409

lanceolatus Benth., "ill, 385

lavanduloides Bak., 372, 385

laxiflorus Benth., 434

madagascariensis (Pers.) Benth., 402, 409

var. aliciae Codd, 402, 404

var. madagascariensis, 403

var. ramosior Benth., 399, 402, 404

marrubioides Hochst. ex Benth., 385

matabelensis Bak., 382

mauritianus Boj., 403

melanocarpus Guerke, 376

mirabilis (Briq.) Launert, 381

moschosmoides Bak., 385

mutabilis Codd, 404

myrianthus Briq., 382

natalensis Guerke, 414

forma glandulosa Phillips, 418

neochilus Schltr., 392

nummularius Briq., 407

oertendahlii Th. Fries iun., 411

ornatus Codd, 393

otaviensis Dinter, 382

pachyphyllus Guerke ex T. Cooke, 398

pachystachyus Briq., 436

parviflorus Guerke, 409

peglerae T. Cooke, 416

petiolaris E. Mey. ex Benth., 431

praetervisus Briq., 418

primulinus Bak., 378

psammophilus Codd, 405

punctatus (L.f.) L'Herit., 372

purpuratus Harv., 410

pyramidatus Guerke, 436

ramosissimus Hook, f., 117

rehmannii Guerke, 421

rotundifolius (Poir.) Spreng., 438

rubropunctatus Codd, 418, 420

saccatus Benth., 426

var. longitubus Codd, 428

var. saccatus, 427

scutellarioides (L.) R. Br., 439

spicatus E. Mey. ex Benth., 372, 383

spiciformis R. A. Dyer, 385

strigosus Benth., 409

var. lucidus Benth., 409

subspicatus Hochst., 383

succulentus Dyer & Bruce, 435

swynnertonii S. Moore, 422

ternifolius D. Don, 117, 436

tetensis (Bak.) Agnew, 390

tetragonus Guerke, 376

thorncroftii S. Moore, 435

thunbergii Benth., 407

tomentosus Benth., 282, 399

594

transvaalensis Briq., 418

tysonii Guerke, 439

unguentarius Codd, 387

urticifolius (Lam.) Salisb., 415

verticillatus (L.f.) Dnice, 407

villosus T. Cooke, 385

woodii Guerke, 401

xerophilus Codd , sp. nov., 282, 378

zatarhendi ( Forsk .) E. A. Bruce, 396, 398, 402

var. tomentosus ( Benth .) Codd , 399

var. woodii ( Guerke ) Codd , 401

var. zatarhendi, 398

zeylanicus Benth., 399

zuluensis T. Cooke, 424

Plectronia L., 491

ventosa L., 491, 492

Poa annua L., 296

Podocarpus, 225

falcatus, 153, 156, 157

latifolius, 153, 534

Poellnitzia, 1 19

Polygala leptophylla, 171

Polyporus [(Mich.) Fr.} Fr„ 222, 224

arenosobasus Lloyd , 222

baudonii Pat., 143, 145, 146, 147, 224

colossus, 224

conchatus Lloyd, 222

durbanensis Van der Bvl, 222, 223

flexilis Van der By l, 223

goetzii, 222

griseus Boes., 222

lucidus, 224

mollicarnosus Lloyd, 224

nigrolucidus Lloyd, 224

ochroporus Van der Byl, 224

pectunculus Lloyd, 223

rugoso-porus, 222

rusticus Lloyd, 225

schweinitzii Fries., 143, 145, 147

trichiliae Van der Byl, 225

tuberaster Pers., 222

vibecinus, 221

var. antelopum Kalchbr., 221, 225

Polyscytalum Ries, 218

Populus, 222

Prosopis sensu E. Mey., 247

elephantorrhiza (DC.) Spreng., 249

vaginatum Swartz,

var. nanum Doell., 243

Prosopis, 110

Protea L., 106

caffra, 342, 343, 553, 572, 575, 576, 578

gaguedi, 576, 578

Protionopsis Blume, 54

Protium sensu Wight & Arn., 54

africanum Harv., 76

Protorhus longifolia, 534

Prunus africana, 156

Pseudechinolaena polystachya (H.B.K.) Stapf, 287

Pseudostemphyllum consortiale (Thiim.) C. V. Subramanian, 209

Ptaeroxylon obliquum, 153

Pteridium aquilinum, 345, 578

Putterlickia, 115

pyracantha (L.) Szyszyl., 115

Pyramidium, 117

Pyrrosia schimperana ( Mett .) Alston, 287

Rabdosia (Bl.) Hassk., 117, 373, 374, 437

subgen. Pyramidium (Benth.) Codd, 436

calycina (Benth.) Codd, 117, 436

javanica (Bl.) Hassk., 1 17, 436

ternifolia (D. Don) Hara, 117, 436

Ramaria Holmsk., 228

cladoniae (Kalchbr.) Reid, comb, nov., 228

Ramularia, 217

aspalia Brogn., 106

eid, Derek A. A reappraisal of type and authentic material

larger Basidiomycetes in the Pretoria Herbarium, 221-230

hinocladiella mansonii (Castell.) Schol-Schwarz, 214, 215

hus, 262

concolor C. Presl, 277

concolor E. Mey. ex C. Presl, nom nud., 277

diversiloba, 259, 261, 262

erosa, 168, 171, 172

laevigata, 226

lancea, 345

laurina, 259

ovata, 259

radicans, 259

undulata var. burchelln, 171

hynchelytrum repens, 172

setifolium, 340, 343

of the

Rhynchosia monophylla, 564, 575, 578

totta, 342

Richardia Kunth, 8

africana Kunth, 6, 8, 9

albomaculata Hook., 8, 18, 21

angustiloba Schott, 8, 15, 16, 18, 21, 22

elliottiana Watson, 14

hastata Hook. 8, 18, 21, 22

lehmannii, 1 1

macrocarpa (Engl.) Watson, 23

melanoleuca Hook, f., 8, 18, 22

var. tropicalis N.E. Br., 18, 22

pentlandii Watson, 1 5

rehmannii (Engl.) N.E. Br., 11

sprengeri Comes, 15, 17

Richardsonia Kunth, 8

Robbertse, P. J. The genus Acacia in South Africa. I. Stipules and

spines, 473-479

Robbertse, P. J. The genus Acacia in South Africa. IV. Morphology

of the mature pod, 481^189

Rogers, D. J. & Moll, E. J. A quantitative description of some coast

forests of Natal, 523-537

Ross, J. H. An analysis of the flora of Natal, 103-106

Ross, J. H. Towards a classification of the African Acacias, 107-1 13

Ross, J. H. A new generic record for South Africa and a new species,

118

Ross, J. H. Notes on Entada in Southern Africa, 125-126

Ross, J. H. Notes on Acacia species in Southern Africa, 127-131

Ross, J. H. Notes on Acacia species in Southern Africa, 231-234

Ross, J. H. The genus Elephantorrhiza, 247-257

Ross, J. H. A note on Dichrostachys cinerea in South Africa, 265-

268

Ross, J. H. A note on the identity of Celastrus filiformis, 277

Ross, J. H. Typification of Schotia capitata, 285

Ross, J. H. Notes on miscellaneous Acacia species from Tropical

Africa, 291-294

Ross, J. H. Notes on Acacia species from north-east Tropical Africa,

299-303

Ross, J. H. Notes on African Acacia species, 443—447

Ross, J. H. The typification of Mimosa Senegal, 449-451

Ross, J. H. The Acacia Senegal complex, 453-462

Ross, J. H. The naturalized and cultivated Acacia species, 463-470

Ross, J. H. The typification of Mimosa farnesiana, 471-472

Ross, J. H. The typification of Lycium inerme, 491-493

Ross, J. H. A note on Vepris lanceolata, 518

Ross, J. H. A note on the identity of Conium chaerophylloides, 521

Ruellia cordata, 338, 341

Sacciolepis, 286

Salix woodii, 345

Sarcocaulon spinosum, 546

Schefflera umbellifera (Sond.) Bail!., 193, 194, 201

Schizachyrium sanguineum, 340

Schizophora Velen., 225

subiculoides (Lloyd) Ryvarden, 225

Schlechterina Harms, 106

Schotia, 285

brachypetala Sond., 285

capitata, 285

Sclerocarya caflfra, 557

Scolopia, 280, 281

flanaganii (H. Bol.) Sim, 281

var. oreophila Sleum., 515

var. flanaganii, 515

mundii (Eckl. & Zeyh.) Warb., 156, 280

oreophila (Sleum ) Killick, stat. nov., 515

stolzii Gilg ex Sleum., 281

var. stolzii, 280

Scopolia lanceolata (Lam.) Spreng., 518

Scytalidium lignicolum Pesante, 213

Sebacina, 229

Secamone, 277

filiformis (L.f.) J. H. Ross, comb, nov., 277

frutescens (E. Mey.) Decne., 277

Seemannaralia gerrardii (Seemann) Vig., 193

Selaginella dregei, 577

Senecio L., 106

serra, 341

Septonema Corda, 218

Serissa capensis Thunb., 491, 493

Serpiculastrum, 117

Serpicula repens L., 116, 117

var. brachypoda Welw. ex Hiern, 117

Serpula Pers. ex S. F. Gray, 227

himantioides (Fr.) Bond, ex Parmasto, 221

lacrimans

var. himantioides, 227

Seseli chaerophylloides Thunb., 521

Setaria,

cana De Wit, 297

finita Launert, 295

lindenbergiana, 572, 573

59:

paliide-fusca ( Schumach .) Stapf & C. E. Hubb., 295

phragmitoides Stapf, 297

tenuiseta De Wit , 295

ustilata De Wit, 295

Silene burchellii, 574

Smodingium, 259, 261, 262

argutum E. Mey. ex Sond., 259, 261, 262

Solenostemon Schumach., 373, 374, 437

sect. Coleoidea J. K. Morton, 373, 438

subgen. Solenostemoides (Briq.) Codd, 438

latifolius ( Hochst . ex Benth.) J. K. Morton, 438

ocymoides Schumach. & Thonn., 373, 437

rotundifolius ( Poir .) J. K. Morton, 438

scutellarioides ( L .) Codd, 439

shirensis ( Guerke ) Codd, 440

zambesiacus Bak., 440

Sorosporium africanum Syd., 215

Sphenostylis angustifolius, 573

Sporobolus, 306

acinifolius Stapf, 295

africanus (Poir.) Robyns & Tournay, 296

baumianus Pilg. , 296

conrathii Chiov., 296

glaucus Mez, 295, 296

lampranthus, 306

macrothrix Pilg., 296

natalensis (St end.) Dur. & Schinz, 296

pyramidalis Beam., 296

salsus Mez, 295, 296

stapfianus Gand., 297

subtilis Kunth, 296

tenellus (Spreng.) Kunth, 295, 296

welwitschii Rendle, 296

Stachys burchelliana, 168, 171, 172

Stapelia,

gigantea N.E. Br., 136, 137

kwebensis N.E. Br., 136

tarantuloides R. A. Dyer, 136, 137

Stemphyllium congestion Newton, 209

consort iale (Thum.) Groves & Skolko, 209, 210

dendriticum De Sousa da Camara, 209

illicis (Tengwail) Neergaard, 209, 210

Stenolirion, 28

Stigmella graminicola Linder, 2 1 5

Stipagrostis, 547

amabilis, 550

Strelitzia Ait., 519, 520

alba (L.f.) Skeels, 519

augusta Thunb., 519

angustifolia Ait., 519

farinosa Ait., 519

juncea Link, 519, 520

ovata Ait., 519

parvifolia Ait., 519

var. juncea, 519

reginae Ait., 519

teretifolia Barrow ex Stead., 520

Strey, R. G. Notes on the genus Cussonia in South Africa, 191-201

Strey, R. G. A new species of Cussonia from the Southern Cape,

515-516

Sympodina, 213

coprophila Subramanian & Lodha, 213

Tarchonanthus camphoratus, 171, 551

Targidia, 285

aequiglumis, 285

Terfezia, 139, 140

austroafricana Marasas & Trappe, sp. nov., 140, 141

boudieri Chat., 139

claveryi Chat., 139, 140

decaryi Heim, 139, 141

leptoderma Tul. & Tul., 140

olbiensis Tul. & Tul., 140

pfeilii Hennings, 139, 140

pinoyi Maire, 140

spinosa Harkn., 141

terfezioides (Matt.) Trappe, 141

Terminalia sericea, 557

Thalassodendron Den Hartog, 106

Themeda triandra, 344

Thesium transvaalense, 343

Toddalia lanceolata Lam., 518

Toxicodendron, 261

radicans (L.) Kuntze, 259, 261, 262

Trametes, 222

albotexta Lloyd, 221, 225

capensis Lloyd apud Doidge, 225

devexa, 226

glabrescens, 227

griseo-lilacina Van der Byl, 226

keetii Van der Byl, 226

lilacino-gilva (Berk.) Lloyd, 226

meyenii ( Klotzsch ) Lloyd, 226

ochro-flava, 226

ochrolignea Lloyd, 226

subflava Lloyd, 227

tomentosa Van der Byl, 227

varians Van der Byl, 222, 227

zimmermanni, 226

Tremella Dill, ex Fr., 229

foliacea, 229

hemifoliacea Lloyd, 229

Tribulus terrestris L., 511

Tricalysia capensis, 534

Tricellula curvata Haskins , 208, 209

Trichilia emetica, 225

Tricholaena, 273

Tripogon, 305

Tristachya biserriata, 564, 573, 574, 575, 578

rehmannii, 340, 575, 576, 578

superba Schweinf. & Aschers., 287

Triticum aestivum L., 217

Tritonia, 281

flava, 281

florentiae (Marl.) Goldbl., comb, nov., 281

Turnera oculata Story,

var. paucipilosa Oberm. var. nov., 290

Ulocladium consortiale (Thum.) Simmons, 209, 210

Urelytrum squarrosum, 340

Urera tenax, 570

Vachellia, 1 1 1

Van der Walt, J. J. A. The South African species of Commiphora,

53-102

Van der Westhuizen, G. C. A. Polyporus baudoni Pat. on Eucalyp-

tus spp. in South Africa, 143

Vellozia retinervis, 340, 343

Vepris lanceolata (Lam) G. Don, 518

undulata (Thunb.) Verdoorn & C. A. Smith, 518

Verdoorn, I. C. The genus Crinum in Southern Africa, 27-52

Verdoorn, I. C. A new species of Hermannia, 288

Verdoorn, I. C. A new species of Hermannia, 518-519

Veronaea, 213

botryosa Cifferi & Montemartini, 213

simplex Papendorf, 213

Vorster, P. The correct author citation for Tristachya superba, 287

Vorster, P. Pyrrosia schimperana, a new fern record for South

Africa, 287

Wells, M. J. The effect of the wagon building industry on the Ama-

tola Forests, 153-157

Werger, M. J. A. An account of the plant communities of Tussen-

die-Riviere Game Farm, Orange Free State, 165-176

Werger, M. J. A. Notes on the phytogeographical affinities of the

Southern Kalahari, 177-180

Werger, M. J. A. On concepts and techniques applied in the Zurich-

Montpellier method of vegetation survey, 309-323

Wormskioldia,

juttae Dinter & Urb., 288

lacerata Oberm., sp. nov., 288, 289, 290

longipedunculata Mast., 288, 289

schinzii Urb., 288, 289

Xylodon P. Karst., 225

subiculoides (Lloyd) Reid, 225

Zantedeschia Spreng., 8

aethiopica (L.) Spreng., 5, 6, 8, 9

var. minor Engl., 9

var. umganiensis Leicht. & Engl., 26

albomaculata (Hook.) Bail!., 8, 15, 17, 18, 21

subsp. albomaculata, 9, 15, 18, 21

forma latifolia Engl., 18

subsp. macrocarpa (Engl.) Letty, stat. nov., 6, 9, 22, 23

subsp. valida Letty, subsp. nov., 6, 9, 25

angustiloba (Schott) Engl., 8, 16, 18, 22

aromatica (Roxb.) Spreng., 5

chloroleuca Engl. & Gilg, 18, 22

elliottiana (Watson) Engl., 5, 9, 14

hastata (Hook.) Engl., 18

jucunda Letty, 6, 9, 13

lutwychei N.E. Br., 25

macrocarpa Engl., 8, 15, 23, 25

melanoleuca (Hook, f.) Engl., 18, 22

var. concolor Burtt Davy, 23

var. tropicalis N.E. Br., 18, 22

occulta (Lour.) Spreng., 5

occulta (Lindl.) Engl., 18, 21

pentlandii (Watson) Wittm., 6, 9, 15, 22

rehmannii Engl., 8, 9, 11

tropicalis (N.E. Br.) Letty, 18, 22

Zeuxine africana Reichb.f, 106

Zygophyllum, 548

CONTENTS - INHOUD

Page/ Bladsy

Date/ Datum

No. 1 & 2 1-205 13 June/Junie 1973

No. 3 207-370 24 December/Desember 1974

No. 4 371-586 9 July/Julie 1975

Page

Bladsy

Codd, L. E. Author citation for Putterlickia pyracantha 115

The genus Rabdosia in South Africa 117

The identity of Erythrina princeps 269

New species of Plectranthus 282

Plectranthus (Labiatae) and allied genera in Southern Africa 371

Coetzee, B. J. A phytosociological classification of the Jack Scott Nature Reserve 329

Improvement of association-analysis classification by Braun- Blanquet technique 365

A phytosociological classification of the Rustenburg Nature Reserve 561

Coetzee, B. J. & Werger, M. J. A. On hierarchical syndrome analysis and the Zurich-Montpellier table

method 159

A west-east vegetation transect through Africa south of the tropic of Capricorn 539

De Winter, B. The identity of some species of Hermannia represented in the Linnaean, Thunberg, Bergius

and Cavanilles Herbaria 263

A new combination in Ozoroa 277

De Winter, B. & Vorster, P. Comments on the treatment of the Poaceae in the Prodromus einer Flora von

Sudwestafrika (1970) 295

Du Toit, P. C. The identity of Ochna atropurpurea 517

A change in status and a new name for Ochna atropurpurea var. angustifolia 517

A change in status for Ochna oconnorii 518

Dyer, R. A. A new species of Euphorbia 278

The status of Strelitzia juncea 519

Ellis, R. P. Comparative leaf anatomy of Paspalum paspalodes and P. vaginatum 235

Smodingium dermatitis: The intercellular secretory canals of the aerial axis and their relationship to this

toxicity 259

Anomalous vascular bundle sheath structure in Alloteropsis semialata leaf blades 273

Non-Kranz (C3) leaf anatomy in the panicoid grasses: additional records from South Africa 286

Gaff, D. F. & Ellis, R. P. Southern African grasses with foliage that revives after dehydration 305

Goldblatt, Peter. Montbretiopsis : reduced to synonymy in Tritonia (Iridaceae) 281

Gunn, M. D. John Hutchinson, O.B.E., F.R.S., LL. D., Y.M.H., F.L.S 1

Jooste, W. J. A new species of Drechslera on Tribulus terrestris 511

Killick, D. J. B. A note on the author citation for Eriudaphus mundii basionym of Scolopia mundii 280

Scolopia stolzii, a new record for South Africa 280

A change of status in Scolopia 515

Leach, L. C. Stapelieae (Asclepiadaceae) from South Tropical Africa, VII 133

Euphorbia gummifera, E. gregaria and a new species from Damaraland 495

Notes on Euphorbia mauritanica, E. gossypina and some related species with an amplified description of

E. berotica 505

Letty, C. The genus Zantedeschia 5

Loxton, A. E. The taxonomy of Paspalum paspalodes and P. vaginatum as represented in South Africa 243

A note on a possible bigeneric hybrid between Digitaria and Anthephora 285

Marasas, W. F. O. & Trappe, James M. Notes on Southern African Tuberales 139

Marasas, W. F. O. & Bredell, Ingrid H. A new species of Phaeoramularia (Fungi Imperfecti : Dematiaceae

from South Africa 217

Morris, J. W. An automatic bibliography indexing programme 325

Progress in the computerization of herbarium procedures 349

Morris, J. W. & Guillerm, J. L. The ecological profiles technique applied to data from Lichtenburg, South

Africa 355

Musil, C. F., Grunow, J. O. & Bornman, C. H. Classification and ordination of aquatic macrophytes in

the Pongola River pans, Natal 181

Obermeyer. A. A. A note on Laurembergia repens 116

New taxa in the Turneraceae 288

Obermeyer, A. A. & Giess, W. Aloe , Chamaealoe, Haworthia, Astroloba, Poellnitzia and Chortolirion 119

Oliver, E. G. H. Notes on the Ericaceae 115

Papendorf, M. C. & Jooste, W. J. The mycoflora of wheat field debris. Part 1 207

New and interesting records of South African fungi. Part VI 11 211

Reid, Derek A. A reappraisal of type and authentic material of the larger Basidiomycetes in the Pretoria

Herbarium 321

Robbertse, P. J. The genus Acacia in South Africa. I. Stipules and spines 4/3

The genus Acacia in South Africa. IV. Morphology of the mature pod 481

Rogers, D. J. & Moll, E. J. A quantitative description of some coast forests of Natal 523

Page

Bladsy

Ross, J. H. An analysis of the flora of Natal 103

Towards a classification of the African Acacias 107

A new generic record for South Africa and a new species 118

Notes on Entada in Southern Africa 125

Notes on Acacia species in Southern Africa 127

Notes on Acacia species in Southern Africa 231

The genus Elephantorrhiza 247

A note on Dichrostachys cinerea in South Africa 265

A note on the identity of Celastrus filiformis 277

Typification of Scholia capitata 285

Notes on miscellaneous Acacia species from Tropical Africa 291

Notes on Acacia species from north-east Tropical Africa 299

Notes on African Acacia species 443

The typification of Mimosa Senegal 449

The Acacia Senegal complex 453

The naturalized and cultivated Acacia species 463

The typification of Mimosa farnesiana 471

The typification of Lycium inerme 491

A note on Vepris lanceolata 518

A note on the identity of Conium chaerophylloides 521

Strey, R. G. Notes on the genus Cussonia in South Africa 191

A new species of Cussonia from the Southern Cape 515

Van der Walt, J. J. A. The South African species of Commiphora 53

Van der Westhuizen, G. C. A. Polyporus haudoni Pat. on Eucalyptus spp. in South Africa 143

Verdoorn, I. C. The genus Crinum in Southern Africa 27

A new species of Hermannia 288

A new species of Hermannia 518

Vorster, P. The correct author citation for Tristachya superba 287

Pyrrosia schimperana, a new fern record for South Africa 287

Wells, M. J. The effect of the wagon building industry on the Amatola Forests 153

Werger, M. J. A. An account of the plant communities of Tussen-die-Riviere Game Farm, Orange Free State 165

Notes on the phytogeographical affinities of the Southern Kalahari 177

On concepts and techniques applied in the Ziirich-Montpellier method of vegetation survey 309

Bothalia 11, 1 & 2: 1 — 3 (1973)

John Hutchinson, O.B.E., F.R.S., LL.D., F.L.S.

M. D. GUNN*

Dr. John Hutchinson, Nestor of British systematic

botanists, botanical author, explorer, collector and

authority on the floras of India and Africa, died at his

home at Kent Road, Kew, England, on 2nd September

1972. Born at Blindburn, Wark-on-Tyne, Northum-

berland, England, on 7th April 1884, he received his

schooling at Wark and Newcastle and his practical

training in horticulture in his native Northumberland

and the adjacent county of Durham. To gain further

experience he journeyed south to Kew and entered the

Royal Botanic Gardens as a student gardener in 1904.

His interest in the scientific side of horticulture and his

natural talent for drawing gained him early recognition

when he was selected to fill a temporary assistant post

in the Herbarium in 1905. In 1907 he was promoted

assistant in the Indian section and in 1909 was trans-

ferred as assistant for Tropical Africa. He returned a

second time to Indian botany from 1915-1919 and

from then onwards was in charge of the African sec-

tion until 1936 when he was appointed Keeper of the

Museums of Botany. He retired officially from Kew in

1948, but continued working there in the Herbarium

and in the writing of books. His 68 years association

with the Royal Botanic Gardens terminated with his

death.

Fig. 1. — Drs. R. A. Dyer, I. B. Pole Evans and J. Hutchinson. Photograph taken in 1953

♦Botanical Research Institute, Department of Agricultural Technical Services, Private Bag X 101, Pretoria.

JOHN HUTCHINSON, O.B.E., F.R.S., LL.D., V.M.H., F.L.S.

Although Hutchinson had gained a wide knowledge

of African plant material at Kew and in the great

herbaria of Berlin-Dahlem and Paris and other con-

tinental botanical institutions, it was not until 1928

that he was able to visit the southern part of the

African Continent and study the veld and its flora.

His visit was made possible through a grant made by

the then Empire Marketing Board. He arrived in Table

Bay on 1 3th August 1928 and was met by a small

group of prominent local botanists. He began botaniz-

ing in the vicinity of Cape Town on the day he arrived

and continued collecting and recording his observa-

tions across the length and breadth of the country

until he set sail for England on April 12th 1929, having

travelled some 1 1 000 kilometres and collected over

3 000 plant species.

During his 1928-29 tour Hutchinson had botanized

with and been the guest of General Smuts at his home

“Doornkloof” at Irene near Pretoria. Hutchinson re-

ceived an invitation from the statesman to be his guest

on a botanical expedition he was planning to Lake

Tanganyika. As a result he returned for the second time

to South Africa travelling in the company of Mrs.

Arthur Gillett and her two sons Jan and Anthony

(Mrs. Gillett and Jan Gillett were both keen collectors

and botanists) who had also been invited to join the

Smuts Expedition, arriving at Irene on 29th June 1930.

Mrs. Smuts and some members of her family accom-

panied the party as far as the Victoria Falls. Dr. I. B.

Pole Evans then Chief of the Division of Botany joined

the expedition at Beit Bridge, he and two assistants

travelling in a well equipped motor caravan. The ex-

pedition journeyed through Southern and Northern

Rhodesia (now Rhodesia and Zambia) on to Abercorn

on the shores of the Lake Tanganyika. On the return

journey Hutchinson left the party at Broken Hill and

went on into the Belgian Congo. He returned to Eng-

land in August 1930.

An account of his travels is given in his book A

Botanist in Southern Africa published in London in

1946. The book, now an item of Africana, is profusely

illustrated and the numerous black and white line

drawings of plants bear testimony to his talent as a

botanical artist.

On his departure from South Africa he sent a print-

ed message of farewell (Fig. 2) to the many old and

new friends who had given him assistance during his

travels. This gesture on the part of Hutchinson was

typical of the man. His friendship towards South

Africans continued throughout his life. Any South

African visiting the Kew Herbarium was soon taken

under his wing, including a long line of liaison officers

from the National Herbarium in Pretoria.

Fig. 2. -Farewell message sent by Dr. Hutchinson to his friends on leaving South Africa.

M. D. GUNN

3

Apart from his monumental A Botanist in Southern

Africa, Hutchinson contributed to South African bo-

tany in numerous papers to the Kew Bulletin and in

revisions of families for the Flora Capensis, namely

Myricaceae and Moraceae and contributions to Cy-

cadaceae, Euphorbiaceae and Proteaceae. In a wider

context he was known to botanists in South Africa

for his classic works Families of Flowering Plants

(1926, 1934 and 1958), Flora of West Tropical Africa

(with Dalziel, 1927-1936), the Genera of Flowering

Plants (1963, 1967) and Evolution and Phylogeny of

Flowering Plants (1969).

Hutchinson received many honours. In 1934 the

University of St. Andrews awarded him an honorary

LL.D. degree; in 1947 he was elected Fellow of the

Royal Society; he received the Victoria Medal of Ho-

nour from the Royal Horticultural Society in 1944,

the Darwin Wallace Centenary Medal in 1958 and the

Linnean Gold Medal in 1968. In 1972, shortly before

his death, he was awarded the O.B.E. A fitting tribute

was paid to Hutchinson by Dr. R. A. Dyer, his old

and valued friend, formerly Chief of the Division of

Botany and Editor of Flowering Plants of Africa,

when he dedicated Vol. 31 (1956) to him. The dedica-

tion reads: —

TO

JOHN HUTCHINSON

LL.D., F.R.S., V.M.H., F.L.S

Botanist, artist, for many years in charge of the African col-

lection in the Herbarium of the Royal Botanic Gardens, Kew,

later Keeper of the Museums, retired in 1948, author of nume-

rous botanical works during his official career including “The

Flora of West Tropical Africa” in joint authorship with J. M.

Dalziel, “The Families of Flowering Plants”, “A Botanist in

Southern Africa”, and since retirement a series of popular books

on the British Flora illustrated by his facile pen and for many

years employed on the mammoth task of a new genera cf

flowering plants of the world, this volume is dedicated as a

mark of appreciation for seeing through the press in England

most of the first 26 volumes of “Flowering Plants of Africa” and

for his invaluable guidance of and friendship towards many

South African botanists including the present grateful editor of

this work.

On learning of his death, Dr. Dyer wrote: “From

a humble beginning in the Gardens, he reached the

pinnacle of botanical fame”.

.

Bothalia 11, 1 & 2: 5-26 (1973)

The Genus Zantedeschia

CYTHNA LETTY*

ABSTRACT

A revision of the genus Zantedeschia , with special reference to Southern Africa, is presented and

a key is provided as a guide to the six species and two subspecies recognized. The new combinations

and subspecies are: Z. albomaculata (Hook.) subsp. macrocarpa (Engl.) Letty and subsp. valida Letty.

Each species and subspecies is illustrated in colour.

INTRODUCTION

Although the genus Zantedeschia Spreng. has been

revised as recently as 1915 by Engler in Das Pflan-

zenreich and in 1948 by Traub in Plant Life, these

authors appear to have cleared up little of the con-

fusion existing, chiefly through lack of a wide range

of properly documented material and the oppor-

tunity for field work.

Most of the previous taxonomical studies in the

genus deal with specimens from Africa with scanty or

no information, or from plants grown in European

gardens and hot-houses. The inevitable consequence

is that species were described without knowledge of

the variation found in nature while some, e.g. Z.

elliottiana , may be of hybrid origin.

For this reason a revision is tentatively offered,

and will, it is hoped, point out features for study of

the plants occurring naturally in Southern Africa

and further afield.

Species of Zantedeschia are commonly known as

Arum Lilies, Pig Lilies, Cape Arums, Black-eyed

Arums, Yellow Arums, Pink Arums and in Afrikaans

as Varkblomme, Varkore and Aronskelke.

The use of the term Arum results from a superficial

resemblance of the South African plants to members

of the genus Arum L. occurring in Europe and

western Asia.

The genus was named by Sprengel in 1826 in

honour of a Professor Zantedeschi; there is uncer-

tainty, however, whether Giovanni or Francesco

Zantedeschi was intended. Engler states that it was

the latter but Marloth and contemporary Italiar

botanists feel sure that it was the former. According

to a letter from Professor Rodolfo Pichi-Sermolli

of the University of Florence: “As regards the

question of Zantedeschia I believe that you are

wrong in thinking that the genus is named in honour

of Francesco Zantedeschi. According to Saccardo

(La Botanica in Italia, Mem. R. 1st. Veneto Sc. Lett.

Art. 25, 4: 176, 1895) the genus Zantedeschia is

dedicated to Giovanni Zantedeschi (1773-1846), a

botanist of Breschia. I think that Saccardo is right.

In fact in 1826 when Sprengel established the genus,

Giovanni Zantedeschi was aged enough to be known,

and had published already several botanical papers.

On the contrary Francesco Zantedeschi (1797-1872)

in 1826 was 29 years old only, and probably was

completely unknown to Sprengel, above all because

he was not a botanist.”

* Formerly of the Botanical Research Institute, Department

of Agricultural Technical Services, Private Bag x 101, Pretoria.

Sprengel based the genus on three species: Z.

aethiopica (L.) Spreng. from South Africa and two

Asiatic species, Z. aromatica (Roxb.) Spreng. and

Z. occulta (Lour.) Spreng. Although further Asiatic

species were added by C. Koch, Ind. Sem. Hort.

Berol. (1854), it is generally accepted that Zante-

deschia should be restricted to the African species,

while the Asiatic species are mainly accommodated

in the genus Homalomena Schott. Engler, Pflanzenr.

Fig. 1. — Zantedeschia aethiopica, illustrated by Commelin in

Horti medici Amstelaedamensis, Vol. 1. Plate 50 (1697).

6

THE GENUS ZANTEDESCHIA

4, 23 Da: 24 (1912), places Homalomena in the

tribe Philodendreae and Zantedeschia in the tribe

Zan tedeschiaeae .

Until Elsie Garret Rice published the book Wild

Flowers of the Cape of Good Hope in 1952, the

striking Cape Arum seems to have been too common

for inclusion in the numerous publications of Cape

Flowering Plants. It received earlier recognition in

Europe. Probably the first illustration of the plant

was made by Commelin in Horti medici Amstelae-

damensis Rariorum Plantarum Historia (1697), page

95, plate 50, in 1697, from a plant raised from tubers

which, to translate his words, he had the honour of

receiving from his Excellency, Simon van der Stel,

Governor of the Cape of Good Hope. He describes

it as a Calf’s-foot from the land of the Moors (More-

landse kalfsvoet) literally: Arum aethiopicum. This

is the Calla ctelhiopica of Linnaeus, Species Plantarum

p. 968 (1753), where he cites Hortus Cliffortianus

453 (1737) among other refrences.

In Link, Diss. Bot. (1795), the species was trans-

ferred to the genus Colocasia, while Kunth (1815)

made use of the phrase name of Hermann and renamed

it Richardia africana. Since the generic name Richar-

dia had been used by Linnaeus for a plant in Rubia-

ceae, Sprengel (1826) established the genus Zante-

deschia and reverted to the correct specific epithet

aetliiopica. After further vicissitudes in the genera

Arodes, Richardia and Calla the name Zantedeschia

aethiopica (L.) Spreng. was upheld by Engler in Das

Pflanzenreich in 1915. Eventually the name Zante-

deschia was conserved against earlier generic names.

As far as is known, the genus is confined to the

African continent and is concentrated mainly in

Southern Africa.

Six species, with two subspecies, are recognized.

In addition to the common Arum Lily, Z. aethiopica ,

which is widespread in the Cape Province and Natal

and ranges as far afield as Clanwilliam, Lesotho and

north-eastern Transvaal, there are the Black-eyed

Arums (Z. albomaculata with its two subspecies

macrocarpa and valida), the Pink Arum (Z. rehman-

nii ) and the Yellow Arums (Z. pentlandii , Z. jucunda

and Z. elliottiana). The last mentioned is tentatively

upheld at this stage though its status will remain

uncertain until material matching the type has been

located in the wild state.

The Black-eyed Arums or Albomaculata group

show great variation in size, shape and maculation

of the leaf, the size, shape and colouring of the

spathe, the size of the berries and number of ovules.

Their classification presents one of the greatest

problems in the genus and a broad view of the species

limits appears to be justified.

ACKNOWLEDGEMENTS

I wish to express my thanks and appreciation to

Dr. L. E. Codd, Director, Botanical Research Insti-

tute, for his guidance and encouragement in under-

taking this revision and for examining certain type

specimens in European herbaria.

I am also grateful to the staff of the National

Herbarium for their assistance at all times, especially

to Miss Mary Gunn, Librarian of the Institute, for

her help with historical information and innumerable

references.

The assistance of the Directors of the following

institutes who kindly lent specimens, including types,

for the luitherance of this revision, is gratefully

acknowledged: Royal Botanic Gardens, Kew; Jardin

Botanique de l’Etat, Brussells; National Botanic

Gardens, Kirstenbosch; Natal Herbarium, Durban;

State Herbarium, Stellenbosh.

PRE-LINNAEAN HISTORY

1687. The first reference made to our popular Cape

Arum, which is now known as Zantedeschia aethio-

pica, appears to have been by Paul Hermann in his

Horti Academici Lugduno-Batavi Catalogus, page

60 (1687) where he used the phrase: “Arum aethio-

picum flore albo odorato, moschum redolente.”

1697. Commelin produced the first known illus-

tration of the species, accompanied by a description

in Latin and Dutch in his Horti Medici Amstelo-

damensis Rariorum Plantarum Historia, p. 95, plate

50 (1797).

1705. Paul Hermann referred again to the species

in his Paradisus Batavus, p. 74 (1705), under the

appellation “Arum africanum, flore albo odorato.”

1720 (?). There is a good picture of Zantedeschia

aethiopica in a book of wild-flower facsimile paintings

(probable year of painting about 1720), annotated

in 1755 by Johannes Burman, in the Library of the

Botanical Research Institute, Pretoria.

1737. Linnaeus quotes earlier literature references

when he describes this plant, which he places in

the genus Calla, in his list of the Hortus Cliffortianus.

Fig. 2. — Zantedeschia aethiopica, from a folio of early eignteentn

century water-colour illustrations in the library of the

Botanical Research Institute, Pretoria. The artist is net

known but the illustrations were annotated by Johannes

Burman in 1 755.

CYTHNA LETTY

7

POST-LINNAEAN HISTORY

1753. Linnaeus described the species as Cal la

aethiopica in his Species Plantarum ed. 1 : 968 (1753).

He gives its habitat as Aethiopia, a name often

used for Africa. It may be noted that the genus Calla

is now restricted to Europe.

1763. The genus Aroides Heister ex Fabricius, in

Enum. Plant Hort. Helmst. ed. 2, 2 : 42 (1763), was

separated from Calla to accommodate the species

aethiopica.

1768. In his Dictionary, ed. 8: no. 1 (1768), Miller

adds: “Found at the Cape of Good Hope”, to his

quotation from former literature.

1790. The identity of Colocasia Neck., Elem. 3:

289 (1790), is uncertain and it is treated as a nomen

dubium.

1795. Link, Diss. Bot. Sverin 77 (1795), lists our

species as Colocasia aethiopica. The name Colo-

casia was used by pre-Linnaean authors, even as

early as 1601 (Clusius, Rar. pi. hist., XXV), for a

different plant.

The above publication by Link was the first time

that the name Colocasia was used in post-Linnaean

times. The fact that Link lists the combination

Colocasia aethiopica together with a reference to

Commelin, Hort. Med. Amstel. Rar. PI. 95, t 50

(1697), where a description is provided, would

probably validate Colocasia Link in the above publi-

cation as a monotypic genus under the International

Code of Botanical Nomenclature, Art. 42 (1961

Code), based on Calla aethiopica L.

1805. John Sims, in Curtis’s Botanical Magazine

t. 832 (1805), published a description and a life-like

plate by Syd. Edwards of this species and followed

Linnaeus by naming the plants Calla aethiopica.

Sims throws doubt on Miller’s identification of

the plant because the latter says “the pistils and

stamens are so intermixed as not to be easily dis-

tinguished without the aid of glasses and that a

few of the flowers situate near the extremity of the

spadix are succeeded by berries . . .”.

Fig. 3. — Zantedeschia aethiopica, from

a slide taken by Mr. E. G. H. Oliver of

a specimen in Hort. Cliff. Herb. (BM).

8

THE GENUS ZANTEDESCHIA

1818. Kunth, in Mem. Mus. Nat. Paris 4: 443

(1818), apparently also unaware of the genera created

by Fabricius (1763) and Link (1795), established the

genus Richardia to accommodate Calla aethiopica

L., renaming it Richardia africana Kunth, probably

alluding to the phrase in Hermann (Paradisus Batavus

74, 1705) “Arum africanum flore albo odorato.” The

epithet “africana” is superfluous and therefore illegi-

timate according to our present International Code

of Botanical Nomenclature, and Richardia Kunth is

also inadmissible, being a homonym of Richardia

L. (1753), (Rubiaceae), which he renamed Richard-

sonia Kunth.

1826. Sprengel, Syst. Yeg. 3: 765 (1826), created

the genus Zantedeschia, making the combination

Z. aethiopica (L.) Spreng. and including together

with it certain other Aroids which, according to our

present views, belong in other genera. The name

Zantedeschia was given in honour of a contemporary

scientist, either Giovanni or Francesco Zantedeschi

of Italy (see Introduction).

1829. Link, Handb. 1: 267 (1829), continues to

use the generic name Colocasia for our species.

1837. Rafinesque, New Flora Amer. 2: 90 (title

page 1836, published 1837 according to Merrill,

Index Rafin. 1949), states: “Co//<2 aethiopica is a

peculiar genus, which I call Otosma." In Flora Tel-

luriana 4: 8 (1836) he gives a Latin description of

“this well-known plant of Africa . .

1837. Endlicher reverted to the name Richardia in

Gen. Plant, p. 238 (1837).

1841. Kunth continued to use the name Richardia

in Enum. 3: 58 (1841) retaining the epithet africana

in place of aethiopica.

1856; 1858. Schott, Syn. 131 (1856) and Gen.

Aroid. t. 62 (1858), follows Kunth by using the name

Richardia africana.

1859. W. J. Hooker in Curtis's Botanical Magazine

t. 5140 (1859) described a plant with a deep purple

eye at the base of the spathe and with spotted leaves

as Richardia albo-maculata which was received from

“Messrs Backhouse of York Nurseries . . . from

Natal.” This is the second species known from

South Africa.

1859. Lindley described a species on the same day

as the foregoing as Calla oculata in the Gardeners

Chronicle 40: 788 (1859) stating that “Messrs Veitch

received it from Natal in 1857 . . . spathe yellowish

green with a deep purple eye.”

1860. Schott, Prodr. Aroid. 324, still upheld

Richardia africana for the Cape plant.

I860. W. J. Hooker published a plate and descrip-

tion in Curtis’s Botanical Magazine t. 5176 (1860)

of another plant with a dark eye and cream spathe

which he named Richardia hastata , though he thought

it might be a form of R. alboinaculata with immaculate

leaves.

1865. Schott described a plant ( Welwitsch 230) in

Journal of Botany 3: 35 (1865) which was “common

in deep ponds between the islands of Calemba and

Quisonde on the right banks of the river Cuanza in

Pungo Andongo district of Angola. He named it

Richardia angustiloba Schott.

. 1869. Richardia melanolenca was described by

J. D. Hooker, with a plate, in Curtis’s Bot. Mag. t.

5765 (1869). Its type locality is given as Natal, the

plants having been imported and grown by “Mr Bull

of Chelsea.”

1879. Engler in DC., Mon. Phan 2: 327, upholds

the genus Richardia and refers to the Cape plant

as Richardia africana Kunth.

1880. Baillon in Bull. Soc. Linn. Paris 1: 254

(1880) discusses the confusion in nomenclature of

the Aroid genus which wrongly received the name

Richardia Kunth (1815), a name previously given

to a plant in the Rubiaceae by Linnaeus (1753) but

changed to Richardsonia by Kunth, and comes to

the conclusion that the Calla aethiopica L. which

Kunth named Richardia africana should bear the

name Zantedeschia aethiopica as given by Sprengel

in 1826. He also published the combination Z. albo-

maculata (Hook.) Baill.

1883. Bentham and Hook, f., Gen. PI. 3: 982

(1883), uphold Richardia Kunth, under which Zante-

deschia Spreng. (partly) is placed in synonomy.

1883. Engler, when describing the species Zante-

deschia rehniannii Engl, and Z. macrocarpa Engl,

in Bot. Jahrb. 4: 63 (1883), followed Baillon, ex-

plaining that he had previously used Kunth’s generic

name Richardia to prevent confusion but, after

taking everything into consideration and carrying

out the rules of priority in nomenclature, he agreed

with the findings of Baillon.

1889. Engler continues to use the generic name

Zantedeschia in Engler and Prantl, Pflanzenfamilien

2, 3: 136 (1889).

1891. Otto Kuntze, Rev. Gen. PI. 2: 739-40 (1891),

seems to have been the first to realise the priority

of the generic name Aroides, but he changed it to

Arodes for grammatical reasons. He made the com-

binations A. aethiopicum (L.) Kuntze, A. albo-

maculatum (Hook.) Kuntze, A. angustilobum (Schott)

Kuntze, A. hastatum (Hook.) Kuntze and A. melano-

leucum (Hoolc.f.) Kuntze.

1915. Engler, in his revision of the genus Zante-

deschia in Pflanzenreich 4, 23 Dc: 61 (1915), recog-

nizes 8 species. He, however, wrongly places the

species Z. pentlandii (Whyte ex Watson) Wittm. in

synonomy under Z. angustiloba (Schott) Engl.

1948. Traub in Plant Life Vol. 4 (1948) revised

the genus on the same lines as Engler in Pflanzen-

reich (1915). He also upheld 8 species and, as in

the case of Engler, he suffered from lack of knowledge

of the plants in the field, with the result that his

revision perpetuates much of the earlier confusion.

ZANTEDESCHIA

Zantedeschia Spreng ., Syst. Veg. 3: 765 (1826);

Baill. in Bull. Soc. Linn. Paris 1: 254 (1880); Engl,

in Bot. Jahrb. 4: 64 (1883); Pflanzenr. 4, 23 Dc:

61 (1915); Marloth, FI. S. Afr. 4: 52 (1915); Bailey,

Cycl. Hort. p. 3534 (1963); Traub, Plant Life 4:

12 (1948); Adamson & Salter, FI. Cape Penins. 132

(1950); Phill., Gen. S. Afr. Flow. PI. ed. 2: 161 (1951);

nom. cons. Type species: Z. aethiopica (L.) Spreng.

Calla L., Sp. PI. 968 (1753), pro. parte.

Aroides Heist, ex Farbricius, Enum. PI. Hort.

Helmst. ed. 2, 2: 42 (1763); Kuntze, Rev. Gen. 2:

737 (1891) (as “Arodes”).

Colocasia Link, Diss. Bot. Sverin 77 (1795); Handb.

1 : 267 (1829).

Richardia Kunth in Mem. Mus. Hist. Nat. Paris

4: 433, t. 20 (1818); Engl, in DC., Mon. Phan. 2:

326 (1879); N.E. Br. in FI. Cap. 7: 36 (1897); FI.

Trop. Afr. 167: 8 (1901); Phill., Gen. S. Afr. Flow.

PI. ed. 1: 125 (1926).

CYTHNA LETTY

9

Perennial herbs with fleshy rhizomes. Leaves

deciduous or evergreen, petiolate, radical, contem-

porary with flowers; petioles long, sometimes stained

or mottled with purple, sometimes setulose, clasping

at the base to form a short neck; leaf blades varying

with age and species from lanceolate to orbicular,

maculate or immaculate, base cuneate, truncate,

hastate, sagittate or cordate, apex acute or obtuse

with a subulate tip. Peduncle as long as, or longer

than, the leaves, green, sometimes stained or mottled

with purple, sometimes setulose. Spalhes subcylind-

rical to funnel-shaped, convolute at the base up

to 17 cm long, ivory-white, cream, yellow, pink or

rosy-purple, often with a purple blotch at the base

inside, persistent; limb spreading, oblique, truncate

to acute, terminating in a subulate tip. Spadix monoe-

cious, free, sessile or sometimes more or less stipi-

tate, shorter than the spathe; staminate and pistillate

sections distinct, with male flowers in the upper

sections; staminodes sometimes interspersed among

the ovaries, other sterile organs absent. Perianth O.

Stamens free, crowded together haphazardly; anthers

sessile, oblong, laterally compressed, truncate at the

apex, opening by terminal pores from which pollen

is emitted in long beaded strings. Ovaries numerous,

spirally arranged, 3-locular; ovules 1-8 per locule,

attached dorsally to the central placenta; style short;

stigma discoid. Fruit a berry; berries clustered, few

to many seeded, green or orange-coloured when

ripe, up to 2 cm in diam., usually surrounded by

and contained in, the withering spathe. Seed leathery,

subglobose or ovoid, often compressed.

The genus is restricted to the African continent

extending from the Cape Province to the eastern

Orange Free State, Natal, Lesotho, Swaziland, Trans-

vaal, Rhodesia, Malawi, Zambia and Angola, with

a recent gathering from northern Nigeria which has

not been examined. Six species are recognized, one

of which, Z. e/liottiana, has not been recorded with

certainty in the wild state and is suspected of being

of hybrid origin. The plants grow in marshy places,

on grassy slopes and at forest margins.

The species fall into two distinct sections: (a) the

typical section containing one species, Z. aethiopica ,

in which the plants do not die down in winter, the

female flowers in the lower part of the spadix are

interspersed with staminodes, and the fruits turn

orange in colour and become soft and later muci-

laginous on maturity; and (b) a section consisting

of the remaining five species in which the plants

die down in winter, there are no staminodes among

the female flowers, and the fruits remain firm and

green until they finally wither and rupture or decom-

pose on the ground.

Key to Species

Plants evergreen; female flowers interspersed with staminodes 1. Z. aethiopica

Plants deciduous; female flowers not interspersed with staminodes:

Leaves narrowly lanceolate, cuneate at the base 2. Z. rehmannii

Leaves hastate to cordate at the base:

Spathes golden yellow to lemon yellow:

Leaves triangular-hastate, always maculate; spathe golden yellow 3. Z. jucunda

Leaves oblong-hastate to ovate-orbicular, cordate, maculate or immaculate:

Leaves ovate-orbicular, maculate; spathe without a purple blotch at base 4. Z. elliottiana

Leaves oblong-hastate, usually immaculate; spathe usually with a purple blotch at base

5. Z. pent land ii

Spathes white, cream, ivory, pale greenish-yellow or rarely coral-pink:

Leaves either immaculate or conspicuously maculate, usually oblong-hastate, rarely ovate-

hastate (and then white-spotted), lower lobes short to triangular-spreading or strap-shaped ;

spathes with limb more or less spreading, usually tapering gradually to the apex

6a. Z. albomaculata subsp. albomaculata

Leaves immaculate or slightly maculate, rarely conspicuously maculate, triangular-hastate to

ovate-orbicular-cordate, lower lobes bluntly triangular; spathes more or less truncate,

abruptly apiculate:

Leaves usually triangular-hastate, usually slightly maculate; spathes more or less truncate;

berries large (up to 2 cm diam.)

6b. Z. albomaculata subsp. macrocarpa

Leaves ovate-cordate to ovate-orbicular-cordate, immaculate; spathes large, limb somewhat

spreading; berries medium-sized (up to

a

1. Zantedeschia aethiopica (L.) Spreng., Syst.

Veg. 3: 765 (1826); Baill. in Bull. Soc. Linn. Paris

1 : "254 (1880); Engl, in Bot. Jahrb. 4: 64 (1883);

Pflanzenr. 4, 23 Dc: 62 (1915); Marloth, FI. S. Afr.

4: 52 (1915); Traub in Plant Life 4: 12 (1948);

Adamson & Salter, FI. Cape Penins. 132 (1950);

Letty in Flow. PI. Afr. 30: t. 1190 (1954-55). Fecto-

type: Specimen in Hort. Cliff. Herb. (BM; PRE,

photo.).

Calla aethiopica L., Sp. PI. 968 (1753); Mill., Diet. ed. 8

(1768); Ait., Hort. Kew. ed. 1,3:318 (1789): Sims in Bot. Mag.

t. 832 (1805).

Aroides aethiopicum (L.) Heist, ex Fabric., Enum. PI. ed.

2: 42 (1763); Kuntze, Rev. Gen. 2: 740 (1891) (as “Arodes").

Colocasia aethiopica (L.) Link, Diss. Bot. Sverin 77 (1795);

Handb. 1: 267 (1829).

Otosma aethiopica (L.) Rafin., New FI. Amer. 2: 90 (1836);

FI. Tellur. 4: 8 (1836).

1 cm in diam.)

6c. Z. albomaculata subsp. valida

Richardia africana Kunth, Mem. Mus. Hist Nat. Paris 4:

433, t. 20 (1818); Engl, in DC., Mon. Phan. 2: 326 (1879);

N.E. Br. in FI. Cap. 7: 36 (1897); nom. illegit. Type: same as

Z. aethiopica.

Zantedeschia aethiopica var. minor Engl, in Pflanzenr. 4,

23 Dc. : 63 (1915). Type: based on Gard. Chron. 27: 755,

fig. 153 (1890) of Z. aethiopica “Little Gem." — var. umganien-

sis Leicht. & Engl, in Pflanzenr. 4, 23 Dc: 64 (1915). Type:

Natal, Howick Falls, 1899, Nelson.

Plants glabrous up to 60 cm tall (sometimes up

to 2,5 m under trees). Leaves evergreen, up to 60

cm long; petiole green; blade usually immaculate,

characteristically more or less spreading and leathery,

varied in shape but usually broadly ovate-cordate

or hastate, 15-20 cm long, 10-15 cm broad at the

base, apex obtuse or acute, the length of the blade

above the basal lobes usually less than twice the

width. Peduncle 30-60 cm long, green, glabrous,

CYTHNA LETTY

Fig. 5. — Fruit of Zantedeschia

aethiopica from plants

grown in Rivonia, Trans-

vaal.

agree in all except the last characters, for example

De Winter 768 from Margate, Letty 263 from Beach

Terminus and Codd 8001, all from near Port Shep-

stone, and Bruce 164 from Sheffield Beach near

Durban, have the spadix creamy white and a more

deeply folded spathe which turns purplish on fading,

a character which suggests a tendency towards

the purple blotch which occurs in all the species

except Z. aethiopica, Z. rehmannii and the as yet

not located Z. elliottiana

Young and depauperate plants, such as Codd

8001 mentioned above, are widespread and differ

only in size. The spathes are often smaller in pro-

portion to the length of the peduncle, which can

reach up to 2,5 m long in marshy ground under

trees. Hybrids of Z. aethiopica are not often found

in the wild state, perhaps because the species is

almost completely confined to the eastern and

southern coastal belt, although occasional plants

of obvious hybrid origin do occur in the Transvaal

and Lesotho, where Z. aethiopica overlaps with

Z. albomaculata. Horticultural hybridization is general

and many varieties in size, colour and scent have

been produced. Abnormalities, such as a branched

spadix, multiple spathes and modification of leaves

into large bi-coloured additional spathes are,

occasionally, found in the field, but mostly in culti-

vated plants. There is a form with a green and white

leafy spathe which is very popular in gardens and

greenhouses. Plants with maculate leaves are found

in gardens but are seldom collected in the wild.

2. Zantedeschia rehmannii Engl, in Bot. Jahrb.

4: 63 (1883) (as “Rehmanni”); Pflanzenr. 4, 23

Dc: 65, fig. 30 A-D (1915); Marloth, FI. S. Afr. 4:

53, t. 13 (1915); Traub in Plant Life 4: 25 (1948);

Letty, Wild Flows. Tvaal. 8, t. 3 (1962). Type:

‘New Castle”, Rehmann 80 (Z, holo.) (See note

below).

Richardia rehmannii (Engl.) N.E.Br. ex Harrow in Garn.

Chron. 4: 570 (1888), as “Lehmanni”; Watson in Gard. Chrod.

12: 124 (1892); N.E.Br. in Gard. Chron. 13: 568 (1893), as

“Rhemannii”; Krelage in Gard. Chron. 14: 564, t. 94 (1893),

as “Lehmannii”; Watson in Gard. Chron. 16: 364 (1894);

Krelage in Gartenfl. 43: 12, fig. 7 (1894); Hook.f. in Bot. Mag.

t. 7436 (1895); N.E.Br. in FI. Cap. 7: 36 (1897); Spreng. in

Wien III. Garten Zeit. 415 (1901), as “Stehmannii"; Medley

Wood, Natal Plants 6: t. 512 (1912); Phill. in Flow. PI. S.Afr.

1: t. 15 (1921).

Plants glabrous from 24-80 cm tall but usually

between 40-60 cm. Leaves deciduous, petiolate,

up to 40 cm long; petioles 10-20 cm long, clasping

at the base; blade dark green, immaculate, lanceolate,

acuminate, subulate at the apex, cuneate at the base,

15-40 cm long, 2-7 cm broad. Peduncle up to 60

cm long, green, glabrous, channelled. Spathe 11-12

cm long, 1,5-2 cm broad, folded from the base

for two thirds of its length into a narrow funnel,

colour ranging from white through shades of pink

to dark maroon (almost black), immaculate at

base within, limb slightly spreading, recurved, with

a tapering tip. Spadix usually stipitate about half

as long as the spathe; male zone about 2,2 cm long,

anthers yellow, pollen white; female zone about

0,8 cm long. Fruit with 2-4 ovules; up to 6 seeds.

Zantedeschia rehmannii grows among rocks on

grassy hillsides at medium and quite high altitudes

from Harrismith in the Orange Free State, northern

Natal, through Swaziland to southern and eastern

Transvaal. It is often found in semi-shade at forest

margins and in sandy runnels.

The flowering period is, more or less, from Sep-

tember until February with a peak in November,

December and January; in cultivation it blooms

occasionally out of season.

12

THE GENUS ZANTEDESCHIA

Fig. 6. — Zantedeschia rehman-

nii, growing wild in Utrecht

District; flowers pink. ( Codd

& Dyer 6268). Photo by

Dr R. A. Dyer.

Transvaal. — 2430 (Pilgrim’s Rest): Pilgrim's Rest (-DD),

Strey 3252. 2530 (Lydenburg): Oshoek (-AB), Letty 331;

Belfast (-CA), Leendertz 9210; 9212; 45 km N. of Belfast

(-CA), Prosser 1936; Dalmanutha (-CA), Codd 8084; S.E. of

Machadodorp (-CA), Bruce 474; between Belfast and Machado-

dorp (-CA), Van der Schijff 5818; Machadodorp (-CA), Galpin

12962; Donkerhoek (-CA), Letty 392; Dullstroom (-CA),

Van der Merwe s.n. 2531 (Komatipoort): Barberton (-CC),

Thorncroft 4976. 2629 (Bethal): Joubert’s farm, Ermelo (-DB),

Henrici 1707; farm Nooitgedacht, Ermelo (-DB), Henrici 1171 ;

Mockford’s farm Blairmore (-DB), Gunn s.n.; Ermelo (-DB),

Leendertz 7840; 2630 (Carolina): Steynsdorp (-AA), Diepe-

rink s.n. 2730 (Vryheid): Wakkerstroom (-AC), Devenish 219;

farm Oshoek (-AC), Mauve 4487.

Swaziland. — 2631 (Mbabane): Mbabane (-AC), Rogers

11664; 9 km E. of Oshoek on road to Forbes Reef (-AC), Codd

6395; Little Usulu River, (-AC), Compton 28456.

Natal. — 2729 (Volksrust): farm “The Elms”, Droogdal

(-DD), Schweikerdt 134; Majuba Mt., near Charlestown (-BD),

Haygarth s.//.(NH). 2730 (Vryheid): near Paulpietersburg

(-BD), Galpin 9690; Utrecht (-BD), Wahl s.n.; 24 km N. of

Utrecht (-BD), Codd & Dyer 6257 ; 9 km S.E. of P.O. Groenvlei

(-BD), Codd & Dyer 6268. 2731 (Louwsburg): Ngome (-CD),

Gerstner 5188; Codd 9592. 2830 (Dundee): Glencoe (-AA),

Sheperd 97 (NH); near Nqutu (-BA), Codd 8917; Weenen

County (-CC), Wood 5204.

O.F.S. — 2829 (Harrismith): Koolhoek, Thode 2674 (STE).

Cultivated. — Cambridge Botanic Garden 1888, plant sent

by Mr Adlam from Natal (K); Kew Gardens 1894, type of Bot.

Mag. t. (7436 K).

The character that separates Z. rehmannii from

all other species is its lanceolate leaves which are

cuneate at the base, not lobed. The colour range

of the spathe, from ivory-white, through pink to

dark maroon, also seems to be unique. The plants

are not as robust as the majority of species; the

spathe is more tightly folded into a narrow tube

with the limb tapering into a recurved apiculus.

The maturing fruits enlarge and their weight bends

the withering peduncle until the spathe tip touches

the ground where the seeds germinate and sometimes

disappear underground.

Fig. 7. — Hybrid crowd of

Zantedeschia rehmannii x

Z. albomaculata, Utrecht

District ( Codd & Dyer

6286). Photo Dr R. A.

Dyer.

Plate 2. — Zantedeschia rehmannii Engl. From Wild Flowers of the Transvaal, PI. 3 (1962).

CYTHNA LETTY

13

In Berlin there is no Rehmann material of Z.

reftmannii. In Zurich there are two sheets of Rehmann

80: (a) one sheet with 4 leaves has just “Natal 80“

on it; and (b) the other (which may be regarded

as the holotype) is annotated “ Zantedeschia rehman-

nii Engl.” in Engler’s writing and has leaves and

fruit. The label on this sheet, in Rehmann’s writing,

reads: “80 Natal auf trockenen. steinigen, begrasten

Hiigeln 5 engl. M. westlich van (Standarton) New

Castle, bei einem Farmer dessen Namen ich nicht

herausfinden kan. Vielleicht ist der Name in meinen

Reisebiichern enthalten?” The locality Standarton

is crossed out and New Castle substituted by Rehmann

himself, so the latter locality is probably correct

(supported by the former sheet which has just “Natal”

on it).

Rehmann spells the name Standerton incorrectly:

It is called after a man named Stander.

Z. rehmannii is easily crossed with other species,

giving rise to hybrids with lobed leaves, and spathes

varied in colour and shape. Abnormalities of leaves

and spathes are often found such as: a double spathe,

Pretoria garden, Pollock s.n.; Belfast, Galpin 11921;

modified leaf with spadix only, Pretoria garden,

Jarmer s.n.; modified leaf with spathe, Pretoria

garden, Pollock s.n.; leaves with basal lobes, Belfast,

Leendertz 9210; Pretoria garden, Gouws s.n.

3. Zantedeschia jucunda Letty in Bothalia 7:

455 (1961); Wild Flows. Tvaal.'l2, t. 7 (1962).

Type: Transvaal, Lydenburg, Barnard & Mogg 991

(PRE, holo.).

Plants glabrous, up to 60 cm tall. Leaves deciduous,

petiolate, up to 60 cm long; petioles 10-30 cm long;

blade triangular-hastate 17-30 cm long, 5-15 cm

broad across the base, densely maculate with long

white translucent spots, semi-folded along the midrib

and upwardly spreading, apex acute with a subulate

tip, the length of the blade above the basal lobes

usually less than twice the width. Spathe about

10-16 cm long, golden-yellow with a purple blotch

at the base inside, very slightly rugose within, deep

yellow without, folded from below the insertion

of the spadix into a cup-shaped funnel; limb spreading,

slightly recurved backwards and tapering to a subu-

late tip. Spadix 3-4 cm long; male zone apical about

2-2,5 cm long; female zone basal more or less

2,5 cm long; ovaries subglobose, angled from

pressure, sessile, about 24 in number. Fruits about

1-3 cm in diam., 3 loculed with 5-8 ovules.

The distribution of this species seems to be restric-

ted to the Magnet Heights region of the Lulu moun-

tains, Sekhukhuniland, where it flowers from Novem-

ber to January.

Transvaal. — 2429 (Zebediela): Farm Het Fort, Lulu Mts.

(-DB), Barnard & Mogg 991; Geluk’s Location (-DB), cult.

Pretoria, Native Commissioner s.n.; Schoonoord (-DD), Bar-

nard 180; Du Plessis sub PRE 28835; Magnet Heights (-DD),

cult. Haenertsburg, Thompson sub PRE 28836.

Zantedeschia jucunda is most nearly related to

Z. pentlandii and Z. albomaculata. It differs from

the former in having triangular-hastate, copiously

spotted leaves with upwardly spreading basal lobes

and an acute apex, while the spathe is more cup-

shaped, a deeper yellow without and within, where

it is also slightly rugose, and the limb more spreading

and tapering to a subulate tip; whereas Z. pent-

landii has immaculate (or rarely maculate), oblong-

lanceolate leaves with short rounded basal lobes,

the spathe regularly funnel-shaped, limb erect,

Fig. 8. — Fruiting spathes of Zantedeschia rehmannii touching

the ground, with a cluster of young plants in foreground. In

a garden, Pretoria.

f[9- 9. — Fruit of Zantedeschia rehmannii ( Codd 6257), grown

in Pretoria.

5292-2

14

THE GENUS ZANTEDESCHIA

Fig. 10. — Zantedeschia jucun-

da, grown in Pretoria from

plants sent in from the

Native Commissioner, Se-

kukuniland.

lighter in colour, greenish outside, terminating

abruptly in a subulate tip. From the form of Z.

albomaculata subsp. albomaculata occurring in the

eastern Transvaal, Z. jucanda differs in its consis-

tently brilliant yellow spathes, whereas the spathes

of Z. albomaculata may be white, cream or pale-

yellow with an occasional coral-pink variation.

4. Zantedeschia elliottiana ( Watson ) Engl, in

Pflanzenr. 4, 23 Dc: 61 (1915); Burtt Davy in Kew

Bull. 234 (1924), partly, excluding specimen cited;

Traub in Plant Life 4: 20 (1948). Neotype: Hort.

Kew Gardens 27th April, 1897, type of Bot. Mag.

t. 7577 (K!).

Richardia elliottiana Watson in Gard. Chron. 12: 124 (1892);

Mottet in Rev. Hort. 67: 38 (1895); De Duren in Rev. Hort.

Belg. 23: 13, t. 13 (1897); Hook.f. in Bot. Mag. t. 7577 (1898);

N.E. Br. in FI. Trap. Afr. 8: 167 (1901).

Calla elliottiana Watson in Garden and Forest 5: 330 (1892);

Knight in J. Roy. Hort. Soc. 12: Proc. 58 (1890), nom. nud.

Plant robust, up to 60 cm tall. Leaves deciduous,

glabrous; blade up to 27 cm long and 27 cm broad,

sometimes slightly longer than broad, orbicular-

ovate, deeply cordate, apiculate, green with many

translucent white spots; petiole more or less as long

as the blade, occasionally much longer. Peduncle

about 40 cm long. Spatbe up to 13 cm long, funnel-

shaped above, campanulate below, bright “golden

yellow” throughout, immaculate and smooth within,

tip caudate. Spadix up to 7 cm long, lengths of male

and female zones variable, pollen orange yellow;

ovaries green. Fruit a cluster of berries up to 2,5

cm in diam.

Known only from cultivated specimens; not yet

found, with certainty, in the wild state.

Plate 3. — Zantedeschia jucunda Letty. From Wild Flowers of the Transvaal, PI. 7 (1962).

.

CYTHNA LETTY

15

Transvaal.- — Cultivated, “from Transvaal”, Marloth 7697.

Belfast: Machadodorp, Galpin 11820; Dullstroom (cultivated),

O'Connor s.n.

Cultivated — Natal, Durban Bot. Garden 15/12/1909 (NH).

Hort. Kew Gardens 27 April 1897, type of Bot. Mag. t. 7577 (K)

No specimen was preserved at the time of the

original description of this plant; it was grown

from seed without locality of origin in the garden

of Captain Elliott of Farnboro Park, Hampshire,

England, and very inadequately described in the

Journal of the Royal Horticultural Society, Vol.

12, Proc. 58 (1890), and later by W. Watson in

the Gard. Chron. 12: 124 (1892); de Duren, who

published the first illustration of it in Rev. Hort.

Belg. 23: t. 13 (1897), was not certain that it was

not a variety of Z. alboniacu/ata.

The first full description, by J. D. Hooker, ac-

companies Bot. Mag. t. 7577 (1898), which is sup-

ported by a specimen in Kew Herbarium. The above

two plates and this specimen must form the basis

of the concept for identifying this plant because,

up to the present, nothing to match it or to fit into

the description has, with certainty, been found

in the wild state in South Africa. That no standard

was used for the term “yellow” adds to the difficul-

ties.

The specimens cited from the Transvaal come

near to the description of the species but all differ

slightly. Marloth 7697 has an abnormal double

spathe with signs of a purple area at the base inside.

Galpin 11820 has the same characteristics and there

is no certainty that it is not a cultivated specimen.

O'Connor s.n. has a normal spathe which has a

definite purple area at the base inside and is a culti-

vated plant. A plant grown from seed sent as Z.

elliottiana from California, U.S.A., flowered at the

Botanical Research Institute, Pretoria, and proved

to be a hybrid, probably with Z. elliottiana as a

parent, as the shape of the leaves was altogether

different from the description, and the spathe was

a yellowish cream with a pink margin.

The fact that no plants matching the description

have been found under natural conditions in South

Africa suggests that the original seeds acquired

by Captain Elliott were of hybrid origin.

The species is upheld, provisionally, until more

is known about it. Although the spathe is somewhat

similar to those of Z. pentlandii and Z. jucunda in

shape and colour, the species can be distinguished

from these two by the orbicular-ovate leaves. Some

specimens of Z. albomaculata subsp. albomaculata

from the eastern Transvaal have leaves of this shape

but, in this species, the spathe is narrowly funnel-

shaped, normally cream-coloured (suffused with

salmon-orange or coral pink in some abnormal

specimens), with a distinct purple blotch at the

base.

5. Zantedeschia pentlandii (Watson) Wittm. in

Gartenfl. 17: t. 1456 (1898); Letty, Wild Flows.

Transv. 9, t. 6 (1962). Type: Bot. Mag. t. 7397

(1895).

Calla pentlandii Whyte ex Watson in Gard. Chron. 12: 124

(1892), nomen subnudum.

Richardia pentlandii Watson in Gard. Chron. Ser. 3, 15: 590

(1894); Hook. f. in Bot. Mag. t. 7397 (1895). R. angustiloba

sensu N.E. Br. in FI. Cap. 7: 36 (1897); FI. Trop. Afr. 8: 169

(1901), partly; sensu Phillips in Flow. PI. S. Afr. 1 : t. 10 (1921).

R. sprengeri Comes in Att. R. 1st. Incorazz. Napoli. Ser. V.

Vol. 3. Mem. 7 (1902). Type: Hort. Sprenger, Naples (K).

Zantedeschia macrocarpa sensu Engl, in Pflanzenr. IV, 23

Dc: 67 (1915). Z. angustiloba sensu Marloth, FI. S. Afr. 4:

54, t. 13 (1915); sensu Traub in Plant Life 4: 16 (1948), partly;

sensu Batten and Bokelmann, Wild Flows. E. Cape Province

9, t. 4, 2 (1966).

Plants glabrous, up to 60 cm, seldom taller. Leaves

deciduous, up to 60 cm long, usually shorter than

the peduncles, petiolate, petioles up to 30 cm long,

clasping at the base; blade yellowish grass-green,

usually immaculate (seldom maculate) oblong-elliptic

to oblong-lanceolate, hastate, obtuse to acute,

abruptly apiculate, up to 35 cm long and 15 cm wide.

Spathe up to 13 cm long, lemon-chrome yellow with

a dark purple blotch at the base inside, slightly

marked with green outside at the junction with the

peduncle, also at the apiculate apex, regularly funnel-

shaped; limb loosely folded, seldom recurved back-

wards, with a 0,5-2 cm long apiculus. Spadix stipitate,

usually less than half as long as the spathe; male

zone slightly more than two thirds of the length,

anthers yellow; female portion less than one third

of the length. Ovaries 3-locular, 5-8 ovules in each

locule. Fruit green, few in number, with up to 20

seeds.

This species occurs between rocks and beside

streamlets on grassy mountains and seems to be

restricted to the Mapoch region of the eastern Trans-

vaal, comprising the northern part of the Belfast

district and adjoining part of the Lydenburg dis-

trict. Its flowering time ranges through November

and December.

Fig 13.— Zantedeschia pentlandii. Photograph of Bot. Mag. t.

7397.

16

THE GENUS ZANTEDESCHIA

14. — Zantedeschia pentlandii,

growing beside a mountain

stream in the Steenkamps-

berg, Transvaal ( Letty 334).

Photo by Dr. L. E. Codd.

Fig. 15. — Zantedeschia pent-

landii, in natural habitat in

valley of the Steenkamps-

berg, Belfast District, Trans-

vaal ( Letty 333). Photo by

Dr. L. E. Codd.

Fig. 16. — Fruit of Zantedeschia pentlandii from a plant culti-

vated in Pretoria. Berries are usually few but seeds numerous.

Transvaal. — 2530 (Lydenburg): Willem Botha’s farm,

Steenkampsberg (-AA), Letty 333; northern slopes of mountains,

Saaihoek (-CA), Letty 334 , 335 (leaves spotted); 3 km W. of

Draaikraal (-CA), Codd 8227; near Draaikraal (-AB), Du

Plessis 52.

Cultivated. — Cape Town, “plant from Transvaal ex hort.”,

Marloth 7691; Botanical Research Institute, figured for Flow.

PI. S. Afr. 1 : 10 (1921); ex Louis Trichardt in garden of Mrs

Walters; Tsolo: leaves from plant figured in Wild Flows. E.

Cape Province 9, t. 4, 2, 1966 as Z. angustiloba. Batten s.n.;

Hort. Sprenger, Naples (K).

An illustration agreeing with Z. pentlandii appears

under the name Z. angustiloba in Wild Flowers of

the E. Cape Province by Batten and Bokelmann

p. 9, t. 4, 2 (1966). The original plant is reported

to occur at Tsolo, and similar plants are said to have

been seen at Mt. AylifF in the Eastern Province.

Up to the present no specimens of naturally oc-

curring plants from that region have been seen.

Z. pentlandii has immaculate (or very rarely spotted)

oblong-lanceolate leaves, slightly constricted above

the short, hastate, rounded lobes; the apex is obtuse

(or sometimes acute) with a subulate, seldom re-

curving tip. The spathe is clear lemon-chrome yellow,

is funnel-shaped, greenish outside below and rounded

at the apex with an abruptly subulate tip.

/ "

'

• - - . - • - ■

■

■d

_ . t*

I f

I

CYTHNA LETTY

17

The description by Watson in the Gardener’s

Chronicle Ser. 3, 15: 590 (1894) is the first that

qualifies as a valid description of the species and

is indirectly associated with the Bot. Mag. Plate

7397 (1895); it appears, therefore, permissible to

accept the illustration as typifying the species. As

far as is known, no specimen of the figured plant

was preserved. The illustrated plant was grown

from a tuber taken to Kew by the well-known col-

lector E. E. Galpin, who is reported to have stated

that the tubers of a yellow-flowered Richardia which

he had brought to Kew had been presented to him

by a gentleman in the Transvaal who had obtained

them from a “Staatsartilleryman” who got them

from a native chief while on active service. Later

E. E. Galpin verified the precise locality as the Mapoch

district, Lydenburg, Transvaal. This information

is borne out by the fact that there had been a military

engagement in that area a year or two previously.

Richardia sprengeri is reduced to synonymy. There

is a specimen at Kew, made by Sprenger in 1902,

probably from the same material as described by

Comes, which may be regarded as typifying the

species. This specimen agrees with Z. pentlandii

in all characters except the spotted leaves which are

occasionally found on naturally occurring plants

(our example Letty 335).

6. Zantedeschia albomaculata {Hook.) Bail!, in

Bull. Soc. Linn. Paris 1: 254 (1880). Type: Hort.

Kew, type of Bot. Mag. t. 5140 (1859) (K!, holo.).

Plants 40-75 cm tall. Leaves deciduous, up to

75 cm long; petiolate; petiole variable in length

and proportion to leaf-blade, green or purple or

striped and mottled with purple at the base, some-

times with soft hairs towards the base; blade maculate

or immaculate, oblong- or triangular- to orbicular

hastate, up to 40 cm long and 25 cm broad; lower

lobes variable from short and rounded to long

triangular with rounded apex, or strap-shaped with

a blunt or acute apex. Peduncle up to 75 cm long,

green or sometimes stained or mottled with purple,

sometimes with soft hairs towards the base. Spathe

2,5-17 cm long, seldom without a dark purple area

at the base inside, varying in colour from white

through ivory, cream, straw-coloured to pale yellow

and, seldom, coral pink, the colour outside greenish

at the base, intensifying upwards as the spathe

matures; tube sometimes open to the base, mostly

variably convolute, broadening towards the apex;

limb subtruncate to gradually tapering, apiculate,

somewhat recurved. Spadix sessile or stipitate;

ovaries not interspersed with staminodes, 3-locular

with 3-4 ovules per locule. Fruits few to many, up

to 2 cm in diam., green, crowded, 4-8 seeded, upper

surface flattened, persistent stigma short and

sunk.

The distribution of the aggregate species extends

from the eastern Cape Province through Natal,

northern Lesotho, north-eastern Orange Free State,

Swaziland, south-eastern and eastern Transvaal,

and continues northwards into Rhodesia, Malawi,

Zambia and Angola, with a recent record from West

Tropical Africa which requires further study.

Fig. 17. — Zantedeschia albomaculata subsp. albomaculata.

Photograph of Bot. Mag. t. 5140.

Z. albomaculata was the second species in the

genus to be described and the description was based

on a plant introduced from “Natal”, without further

information as to its origin. It differs from the first,

Z. aethiopica , in many important respects: it is

deciduous, it has no staminodes among the female

organs, its fruit remains a firm berry until germination

and does not turn orange-coloured and become

pulpy while still on the peduncle.

The present concept of Z. albomaculata is a relatively

broad one and includes several forms which have

in the past been upheld as distinct species. A study

of plants in their habitats, together with the more

extensive herbarium collections now available, has

indicated that the following entities should be in-

cluded within the broader view of Z. albomaculata :

Z. oculata (1859), Z. has tat a (1860), Z. melanoleuca

(1869), Z. angustiloba (1879), Z. macrocarpa (1883),

Z. melanoleuca var. tropicalis (1901), Z. chloro-

leuca (1903) and Z. melanoleuca var. concolor (1924).

In this complex there is considerable variation

in leaf shape and degree of maculation, as well as

in spathe shape and colour. To some extent this

variation is linked with geographical distribution

and therefore three subspecies are upheld.

18

THE GENUS ZANTEDESCH1A

For key to subspecies, see species key.

(a) subsp. albomaculata

Richardia albomaculata Hook, in Bot. Mag. t. 5140 (1859);

Schott, Prodr. 325 (1860); Van Houtte in FI. des Serr. 13: 97,

t. 1343 (1860); 21 : 165, t. 2258 (1875); Engl, in DC., Mon. Phan.

2; 327 (1879); N.E. Br. in FI. Cap. 7: 37 (1897). Type Hort.

Kew ex Messrs. Backhouse of York, originally from “Natal”,

type of Bot. Mag. t. 5140 (Kl, holo.). R. hastata Hook, in

Bot. Mag. t. 5176 (1860); Engl, in DC., Mon. Phan. 2: 328

(1879); Watson in Gard. Chron. ser. 3, 12: 123 (1892); N.E.

Br. in FI. Cap. 7: 38 (1897); FI. Trop. Afr. 8: 168 (1901). Type:

Hort. Kew ex Messrs. Veitch of Exeter, originally from “Natal”,

type of Bot. Mag. t. 5176 (K!, holo.). R. angustiloba Schott

in J. Bot. Lond. 3: 35 (1865); Engl, in DC., Mon. Phan. 2: 329

(1879); N.E. Br. in FI. Cap. 7: 37 (1897); FI. Trop. Afr. 8:

169 (1901). Type: Angola, Pungo Andongo, between Calemba

and Quisonde, on left bank of river Cuanza, Welwitsch 230

(BM!, holo; Kl). R. melanoleuca Hook. f. in Bot. Mag. t. 5765

(1869); Engl, in DC., Mon. Phan. 2: 328 (1879); Watson in

Gard. Chron. 12: 124 (1892); N.E. Br. in FI. Cap. 7: 38 (1897);

Phill. in Flow. PI. S. Afr. 4: t. 141 (1924). Type: Hort. Bull,

Chelsea, from “Africa”, type of Bot. Mag. t. 5765 (Kl, holo.).

— var. tropicalis N.E. Br. in FI. Trop. Afr. 8: 168 (1901).

Lectotype: Malawi, Namasi, Cameron s.n. (Kl).

Calla oculata Lindl. in Gard. Chron. 40: 788 (1859). Type:

no specimen extant: “Messrs. Veitch received it from Natal”.

Zantedeschia albomaculata (Hook.) Baill. in Bull. Soc. Linn.

Paris 1 : 254 (1880); Engl, in Bot. Jahrb. 4: 64 (1883); Pflanzenr.

4, 23 Dc: 67, t. 30 E-G (1915); Bailey, Cycl. Hort. 3534 t.

4028 (1919); Phill. in Flow. PI. S. Afr. 19: t. 735 (1939); Traub

in Plant Life 4: 14 (1948). — forma latifolia Engl, in Bot. Jahrb.

4: 64 (1883). Syntypes: Transvaal, Houtbosch, Rehmann 93,

94 (Z). Z. hastata (Hook.) Engl, in Bot. Jahrb. 4: 64 (1883);

Dur. & Schinz, Consp. FI. Afr. 5: All (1894); Traub, l.c. 4:

20 (1948). Z. angustiloba (Schott) Engl, in Bot. Jahrb. 4: 64

(1883); Dur. & Schinz, l.c. 5: 477 (1895); Engl, in Pflanzenr. 4,

23 Dc: 67 (1915); Traub, l.c. 4: 16 (1948). Z. melanoleuca

(Hook.f.) Engl, in Bot. Jahrb. 4: 64 (1883); Pflanzenr. 4, 23

Dc: 67 (1915); Bailey, Cycl. Hort. 3536 (1919); Traub, l.c. 4:

22 (1948). — var. tropicalis (N.E. Br.) N.E. Br. ex Engl, in

Pflanzenr. 4, 23 Dc: 68 (1915); Traub, l.c. 4: 24 (1948). Z.

chloroleuca Engl. & Gilg in Warburg, Kunene Exp. 180 (1903).

Type: Angola, Baum 398. Z. oculata (Lindl.) Engl, in

Pflanzenr. 4, 23 Dc: 68 (1915). Z. tropicalis (N.E. Br.)

Letty in Bothalia 7: 456 (1961).

Aroides albomaculatum (Hook.) Kuntze, Rev. Gen. PI. 2:

740 (1891) (as Arodes albomaculatum). A. hastatum (Hook.)

Kuntze, l.c. (1891) (as Arodes hastatum); Rendle, Cat. Afr.

PI. Welw. 2, 1 : 90 (1899). A. angustilobum (Schott) Kuntze,

l.c. 740 (1891) (as Arodes angustilobum); Rendle, l.c. 90 (1899).

Leaves conspicuously maculate or immaculate,

usually oblong-hastate rarely ovate-hastate (and

then white spotted), lower lobes short to triangular-

spreading or strap-shaped; spathes white, ivory or

cream coloured, rarely pale yellow or coral pink,

with limb more or less spreading, usually tapering

gradually to the apex; fruit a cluster of medium-

sized berries up to 1,8 cm in diam.

Z. albomaculata subsp. albomaculata grows in

marshy ground beside rivers or streams and in vleis

but is also found in pockets among rocks on moun-

tains and grassy hillsides, and in forest margins.

It is able to survive, usually in a depauperate state,

under fairly dry conditions. It flowers from November

to April with a peak in December.

Its distribution, according to the present concept,

ranges from the eastern Cape Province, through

Natal, northern Lesotho, north-eastern Orange Free

State, Swaziland, south-eastern and eastern Trans-

vaal and, northwards, into Rhodesia, Malawi, Zambia

and Angola,

Fig. 18. — Zantedeschia albomaculata subsp. albomaculata,

photograph of Bot. Mag. t. 5176, figured as Z. hastata.

Fig. 19. — Zantedeschia albomaculata subsp. albomaculata,

photograph of Bot. Mag. t. 5765, figured as Z. melanoleuca.

Plate 5. — Zantedeschia albomaculata (Hook.) Engl, subsp. albomaculata, showing two forms of leaf.

CYTHNA LETTY

19

Fig. 20. — Zantedeschia albomaculata subsp. albomaculata,

Cathedral Peak Forestry Station, Natal ( Codd & Dyer 6248).

Photo by Dr R. A. Dyer.

Fig. 21. — Fruit of Zantedeschia albomaculata subsp. albom-

aculata from Nottingham Road ( Codd 8527).

Fig. 22. — Zantedeschia albomaculata subsp. albomaculata

with immaculate leaves from Natal ( Codd 8527).

Transvaal. — 2228 (Maasstroom): Mohlakeng Plateau, Blou-

berg (-DB), Codd 9148. 2328 (Baltimore): Pyramid Estate,

near Potgietersrust (-DD), Galpin 9100. 2329 (Pietersburg):

6 km N. of Louis Trichardt (-AB), Admiraal s.n.; 8 km N.E.

of Louis Trichardt (-AB), Codd 8326; on road to Sibasa, 42

km E. of Louis Trichardt (-AB), Codd 8401; Entabene Forestry

Station (-AB), Codd 84)0; Dorpsrivier, Louis Trichardt (-AB),

Gerstner 5/05; Kloofsig near Daviesville (-BD), Markotter

s.n. (STE); Forest Station, Woodbush (-DD), Van Dam s.n.;

Woodbush (-DD), Swierstra s.n.; 5 km N. of Haenertsburg

(-DD), Codd 8415; 1,5 km E. of Haenertsburg (-DD),

Reynolds 5806; Mountain Home Farm, Woodbush (-DD),

Mogg s.n.; Paardevlei Farm (-DD), Mauve 4289. 2330 (Tza-

neen): Elim (-AA), Obermeyer 834; Duiwelskloof (-CA), Haar

s.n.; Scheepers 36; Magoebaskloof (-CC), Codd 8407; Prosser

1827; Tzaneen (-CC), Rogers 12501; New Agatha (-CC),

McCallum 137. 2430 (Pilgrim's Rest): Wolkberg (-AA),

Turner s.n.; The Downs (-AA), Codd 9465, Murray s.n.;

Mariepskop (-DB), FitzSimons s.n.; on Blyde-forestry road

(-DD), Van der Schijff 6515; Kowyns pass, Graskop (-DD),

Plowes 2378. 2431 (Acornhoek): Klaseriebrug, Marieps-

kop (-AC), Van der Schijff 4797. 2528: (Pretoria): Water-

kloof (-CC), Verdoorn62. 2530 (Lydenburg): Dullstroom

(-AC) Galpin 13372; Langverwag (-AC), Louw s.n.; Schagen

Hill, Rosehaugh (-BD), Codd 6210. 2531 (Komatipoort): Kaap-

muiden (-CB), Mogg s.n.; Louws Creek (-CB), Bradshaw 110;

Barberton (-CC), Thorncroft 30; 12 km S.E. of Barberton on

road to Havelock (-CC), Codd 8159. 2729 (Volksrust): Volks-

rust (-BD), Jenkins s.n. 2730 (Vryheid): North Hill, Honey-

moonkloof (-AC), Galpin 9882; Laingsnek road, (-AC), Galpin

11792; farm Nauwpoort (-AC), Adendorff 3; farm Oshoek

(-AC), Letty 501 ; farm Tafelberg (-BB), Letty 500.

O.F.S. — "Basutoland”, Cooper 3327 (K). 2828 (Bethlehem):

Golden Gate, Hoogland Park (-AB), Liebenberg 7442; Tradou

(-CA), Steyn s.n.; Dunnelin Farm (-CA), Potts 3088; Generaal’s

Nek (-CC), Steyn s.n. 2829 (Harrismith): Sanky 320 (K); 29

km S.E. of Harrismith (-AC), Marais 332.

Swaziland. — 2631 (Mbabane): Gobolo (-AC), Compton

30347; Little Usutu River (-AC), Compton 28455.

20

THE GENUS ZANTEDESCHIA

Eig. 23. — Zantedcschia albomaculata subsp. albomaculata

with robust habit from Haenertsberg, Transvaal ( Codd

8415).

Natal. — Without locality: Gerrard 1525 (K); Buchanan s.n.

2731 (Louwsburg): Ngome Forest (-CB), Gerstner 5187 2829

(Harrismith): 38 km N.W. of Ladysmith (-BC), Marais 333;

Oliviershoek Pass (-CA), Thode 4295 (STE); Cathedral Peak

Forestry Station (-CC), Killick & Marais 2120; Killick 1201;

2178; Codd & Dyer 6248; Cathedral Peak area (-CC), Ger-

mishuizen 17. 2830 (Dundee): Culvers (-CC), Rogers 28319;

Krantzkop (-DD) Thode 2675. 2831 (Nkandla): Eshowe-

Entumeni road (-CD), Gerstner s.n. 2929 (Underberg): Gaints

Castle Game Reserve (-AB), Trauseld 303; Oribi Ridge area,

Giants Castle (-AB), Trauseld 492; Thabamhlope Research

Station (-AB), West 635; Champagne Castle (-AB), Bayer

1209; Sweetwaters (-BA), Dyer 3130; west of Estcourt (-BB),

Munro s.n.; Harris Hill, Glenbella (-BB), Plowes 2398; Fair-

leigh (-BB), Plowes 2388; 2390; 16 km W. of Underberg (-CD),

Dyer 3210; 16 km S.W. of Underberg (-CD), Dyer 3294; 3295;

3296; 3297; Loteni Nature Reserve (-DB), Killick 3887; 3,5

km from Donnybrook (-DD), Killick & Marais 2091; 2093;

2094. 2930 (Pietermaritzburg): Balgowan (-AC), Mogg 3537;

Nottingham Road (-AC), Dyer & Verdoorn 2374; 2375; Codd

8527; Everdon farm, 5 km from Howick (-AC), Hilliard 4049;

Lions River district (-AC), Smuts 1410; 18 km from Greytown

on Rietvlei Road (-BA), Marais 352; 43 km S. of Nottingham

Road (-BA), Smuts 1106; 36 km S. of Nottingham Road (-BA),

Smuts 1408; 1409; Dargle (-CA), Hill 2; 3; Griffins Dargle

(-CA), Moll 2669; Table Mountain, Pietermaritzburg (-CB),

Ward 644; Huntley s.n.; Little Noodberg (-DB), Strey 6256;

Kloof (-DD), Letty 449. 3029 (Kokstad): 38 km from Otters-

burg to Harding (-DB), Admiraal 318; Weza, Ingeli slopes

(-DB), Strey 6288. 3030 (Port Shepstone) Ixopo (-AA), Mogg

s.n.; 1,5 km from Ixopo (-AA), Strey 6129; Drakensberg,

Upper Umkomaas (-AB) Oherdieck & Weidemann 1383;

Dumisa (-AD), Rudatis 1237 (K); near Friedcnau (-BC), Strey

6222; 7038; Port Shepstone-Bizana road (-CB), Admiraal 305.

Lesotho. — 2828 (Bethlehem): Leribe (-CC), Dieter lien 306;

Moreng, Butha-Buthe (-CC), Jacot-Guillarmod, 1829, 1830.

2927 (Maseru): Mamathes, Cannibal’s Cave, (-BB), Jacot-

Guillarmod 4645; Mamathes (-BB), Jacot-Guillarmod 1818;

Collett 470.

Fig. 24.— Zantedeschia albomaculata subsp. albomaculata,

fruit of a plant with robust habit from Magoebaskloof

(Codd 8407).

Fig. 25. — Tubers of Zantedeschia albomaculata subsp. albom-

aculata from Maclear, Cape Province ( C. Giddy), X j .

CYTHNA LETTY

21

Fig. 27. — Zantedeschia albomaculata subsp. albomaculata from

Angola ( Codd 7687), with cream coloured spathe.

As will be seen from the extensive synonymy, a

good deal of variation is included in subsp. albo-

maculata. Characters on which species or even varieties

were previously separated, for example, degree

of spotting of the leaf, the presence of bristles on

petioles and peduncles and, to some extent, leaf

shape, have been found to be unreliable when a

wide range of material is examined.

The type of Z. albomaculata , on which the illus-

tration in Bot. Mag. t. 5140 (1859) is based, is a

plant cultivated at Kew which was obtained from

Messrs Backhouse of York and a specimen is pre-

served in Kew Herbarium. It is said to have been intro-

duced from Natal. The leaf is oblong-hastate, macu-

late, and the spathe is ivory-white with the limb

tapering gradually to the apex. Corresponding plants

are known from the eastern Cape Province to the

Natal Midlands, with isolated specimens from the

Harrismith (Orange Free State) and Wakkerstroom

(Transvaal) areas.

Calla oculata Lindl. was described in the Gardener’s

Chronicle 40: 788 (1859) on the same day (1st October

1859) as Richardia albomaculata Hook, in Bot.

Mag. t. 5140 (1859). No type specimen has been

traced and, from its description, it cannot be separated

from Z. albomaculata subsp. albomaculata as inter-

preted in the present publication. Although it has

been equated with Richardia hastata Hook, by most

authors, it has not previously been linked with

Cape. — 3025 (Colesburg): Top of Katberg Pass (-CB), Dyer

2311. 3027 (Lady Grey): Witte Bergen (-CA), Drege 3527 (K).

3028 (Matatiele): Halcyons Drift, (-CD), Marais 867. 3029

(Kokstad): Kingscote (-AB), Marais 344; Mount Currie (-CD),

Comins 308; near Kokstad (-CD), Marais 346; Tyson 1590

(K); 24 km S.E. of Bizana (-DD), Codd 10694. 3126 (Queens-

town): Queenstown (-DD), Galpin 1946. 3128 (Umtata):

Bazeia (-CB), Baur 433 (K). 3129 (Port St. Johns): 3 km W.

of Holy Cross Mission (-BA), Codd 9327; 18 km S.W. of Lusi-

kisiki (-BC), Codd 10693; 21 km E. of Libode (-CA), Codd

10690; 16 km E. of Umtata (-CC), Codd 10688. 3227 (Stutter-

heim): Dohne (-CB), Sim 20119; near Komgha (-DB), Flana-

gan, 1091.

Rhodesia. — Abercorn: Nkali Dambo, Richards 342 (K).

Inyanga: Mare River; Wild 4928; Rheenan Wattle Estate,

Plowes 2289. Melsetter: Crook s.n.; farm Gungunyama, Mt.

Silinda, Obermeyer s.n.

Malawi. — Nivera Hill, Benson 81. Namasi, Cameron s.n.

(K).

Zambia. — Mwinilunga: R. Muzhila, Milne-Redhead 3823.

Angola, — Cubango: near St. Cruz Mission on Cuando

River, Codd 7687 (cult, in Pretoria). Huilla: near Lopollo,

Welwitsch 232; Tchivinguiro, Barbosa & Morena 10062. Kitu-

vico-Kela: Baixo de Cassange, near River Lui, Gossweiler

9570 (K). Libolo: on Longa River, Dawe 352 (K). Moxico:

R. Masanu, MUne-Redhead 4153 (K). Pungo Andongo: between

islands of Calemba and Quisonde, Welwitsch 230 (BM; K).

Cultivated. — “From Natal”, from Messrs. Backhouse of

York, type ofZ. albomaculata in Bot. Mag. t. 5140 (K). “From

Natal”, ex Messrs. Veitch, Exeter, type of Z. hastata in Bot.

Mag. t. 5176 (K). “From Africa”, ex Hort. Bull, Chelsea, type

ofZ. melanoleuca in Bot. Mag. t. 5765 (K). Cult. Hort. Breslau.

Engter 230 (K). “From F. Sander, April 17th 1899” (K).

Fig. 26. — Zantedeschia albomaculata subsp. albomaculata,

type specimen of Z. angustiloba showing leaf variation.

(Welwitsch 230 in BM).

THE GENUS ZANTEDESCHIA

22

Z. albomaculata. A clue to its identity is the state-

ment that it came from Messrs Veitch’s nursery, as

did the type of Z. has tat a.

When describing Richardia hastata in Bot. Mag.

t. 5176 (1860), Hooker mentioned that it was very

closely allied to R. albomaculata, differing mainly

in the absence of pellucid spots on the leaves. Field

studies have shown that this character has no taxo-

nomic significance within subsp. albomaculata al-

though it is interesting to note that in this subspecies

the leaves are either immaculate or conspicuously

spotted, even on adjacent plants, but rarely sparsely

spotted as in subsp. macrocarpa. The type of R.

hastata is a plant cultivated at Kew, said to be from

“Natal”, supplied by Messrs Veitch of Exeter. A

specimen of a single spathe is preserved in the Kew

herbarium. The leaves are depicted as immaculate.

The spathe, which is a little broader than typical

R. albomaculata, is described as “rather dirty yellow-

green”. According to the plate it is light yellow-

green (Ridgway 6. 31. d). Specimens matching the

plate and description of R. hastata grow together

with plants matching the plate of R. albomaculata

in Natal and have also been found as far afield as

Angola ( Welwitsch 232, Codd 7687).

Among the specimens on the sheets bearing the

type number of R. angustiloba Schott ( Welwitsch

230) at the British Museum, only one differs in

having an exceptionally long upper leaf lobe with

prolonged strap-shaped side-lobes, from which, no

doubt, the epithet angustiloba was derived. The

variation within one gathering, added to certain

marked similarities with Z. hastata, has prompted the

present inclusion of R. angustiloba in the albom-

aculata group.

In the original description Schott does not men-

tion the colour of the spathe of R. angustiloba, nor

is it indicated on the type specimen, Welwitsch 230.

However, on two sheets of Welwitsch 232 in BM

(included by subsequent authors in this species) and

annotated, respectively, “/?. angolensis" and “/?.

hastata ”, are the words: “spathe . . . intense sulp-

hurea” and “spatha sulphurea”. Although sulphur

is a pale greenish-yellow, authors have apparently

interpreted the colour as a bright yellow, with the

result that Z. pentlandii was placed as a synonym

of Z. angustiloba (see N.E. Brown in FI. Trop. Afr.

8: 169, 1902).

Later, a plant collected by Baum in Angola was

separated as Z. chloroleuca Engl. & Gilg; it was

described as being related to Z. hastata and was later

placed by Engler as a synonym of Z. angustiloba

(Pflanzenr. 4, 23 Dc: 65, 1915).

The variation in the leaf-shape in specimens from

Angola, often on the same sheet, is not fully under-

stood, and more extensive collection in Angola and

Zambia is necessary in order to reach a satisfactory

decision on the status of plants in these territories.

In this connection it is of considerable interest

to note that Z. angustiloba has recently been recorded

in northern Nigeria by Hepper in Kew Bull. 21:

493 (1968) and FI. W. Trop. Afr. ed 2, 3: 121 (1968).

The spathe is stated to be a striking yellow colour.

This specimen has not been seen and further material

is desired.

Richardia melanoleuca Hook.f. in Bot. Mag. t.

5765, (1869), called the “black-throated” Richardia

by its author, has not only the white pellucid spots

on the leaves in common with the albomaculata

group but also the distinctive dark blotch at the

base of the spathe inside. The type of the plate in

the Botanical Magazine is in the Kew Herbarium

with “Hort. Bull, Chelsea, from Africa” written

on the sheet. Obviously a small plant, it has ovate,

hastate leaves with width almost equal to the length

and copiously maculate. The spathe, described as

“straw-coloured” is also small, wide-open to the

base exposing the whole spadix and the blackish-

purple patch inside. It is felt that the type of R.

melanoleuca is an abnormal specimen because it

is not exactly matched by any specimens from the

wild state. Certain gatherings from Natal (Inanda:

Wood 431 in K and NH, and Wood 1370 in NH)

approach it in appearance but the spathes, although

small, are not open to the base. Individual plants

with spathes matching the type are sometimes found

but with immaculate leaves ( Comins 708 from Mount

Currie). Odd spathes of this type also occur among

normal spathes on adjacent plants. As far as leaf

shape is concerned, oblong-hastate to orbicular-

cordate leaves are occasionally found on the same

plant.

Two specimens also from Inanda, Strey 4874 and

4876 are difficult to allocate. No. 4874 has the small

truncate spathe of subsp. macrocarpa but with

leaf-shape and maculation nearer to subsp. albom-

aculata', No. 4876, with spathe- and leaf-shape more

nearly matching macrocarpa, is immaculate. In

addition, in Kew Herbarium, mounted on the same

sheet as Gerrard 1525 (July 1865), is a specimen of

Sanderson 209 from “Port Natal” which is inter-

mediate between subsp. albomaculata and subsp.

macrocarpa, having the leaves of the former and

spathe of the latter.

R. melanoleuca var. tropicalis N.E. Br., based on

two specimens, Cameron s.n. and Evelyn Cecil 149

from Malawi and Rhodesia respectively, comes nearer

to the type of Z. albomaculata than to that of Z.

melanoleuca. It is characterized by long-triangular-

hastate leaves which are copiously translucent-

spotted and up to 30 cm or more long; and by the

spathes being folded into a cylindrical tube and

varying in colour from cream to pale yellow (with

an aberrant form with handsome, coral pink spathes

from north-eastern Transvaal known as “Helen

O’Connor”).

When dealing with the genus, prior to the publi-

cation of The Wild Flowers of the Transvaal (Letty,

1962), var. tropicalis was given specific status (Bot-

halia 7: 456, 1961) but, after examining a wider

range of material, the differences have not proved

sufficient to exclude it from Z. albomaculata subsp.

albomaculata, though it is almost distinct enough

to be regarded as a variety. Its habitat is mostly

at forest margins or bush in partial shade and its

flowering reaches its peak in November. It is distri-

buted mainly in the Transvaal and northwards into

Rhodesia, but there are several specimens from the

Natal coastal areas which grade into this form.

Plate 7. — Zantedeschia albomaculata (Hook.) Engl, subsp. macrocarpa (Engl.) Letty. From Wild Flowers of Transvaal, PI. 4 (1962).

CYTHNA LETTY

23

Certain specimens collected in the Barberton

area ( Thorncroft 30 and Codd 8159), although agreeing

with Z. albomaculala subsp. albomaculata in habit

of plant and colour and shape of spathe, have leaves

which match those of Z. elliottiana, being broadly

orbicular-cordate and densely maculate. More

material and further study may justify separate

status for this form.

Fig. 28. — Zantedeschia albomaculata subsp. macrocarpa grow-

ing in Lydenburg District ( Codd 8201). Photo by Dr L. E.

Codd.

(b) subsp. macrocarpa {Engl.) Letty, stat. nov.

Z. macrocarpa Engl, in Bot. Jahrb. 4: 64 (1883); Dur. &

Schinz, Consp. FI. Afr. 5: 477 (1895); Engl, in Pflanzenr. 4,

23: 67 (1915). Type: Trigardsfontein, Rehmann 82 (B!, holo.;

Z!; PRE, photo). Z. melanoleuca var. concolor Burtt Davy in

Kew. Bull. 1924: 233 (1924). Type: Hort. Sander & Co. (K!).

Z. oculata sensu Burtt Davy, l.c. 234 (1924), partly; sensu

Letty, Wild Flows. Transv. 8, t. 4 (1962). Z. angustiloba sensu

Traub in Plant Life 4: 16 (1948), partly.

Richardia macrocarpa (Engl.) Watson in Gard. Chron. 12

124 (1892). R. angustiloba sensu N.E. Br. in FI. Cap. 7: 37

(1897), partly; FI. Trop. Afr. 8: 169 (1901), partly.

This subspecies is characterized by having the

leaves sparsely maculate or sometimes immaculate,

triangular hastate, lower lobes bluntly triangular;

spathes cream or straw-coloured with limb more or

less truncate and terminating abruptly in a subulate

tip; fruits usually large, up to 2 cm in diameter,

3 locular, 2-3 ovuled.

Occurs on marshy ground beside streams and in

grassy vleis. The flowering period is from October

to February with a peak in November. Its distri-

bution ranges over the higher regions of the Trans-

vaal, Swaziland, Natal, Lesotho and eastern Orange

Free State.

Transvaal. — 2430 (Pilgrim's Rest): Mac-Mac Falls (-DD),

Codd 6751. 2528 (Pretoria): Debbes Ravine (-CA), Mogg

14842. 2529 (Witbank): Middelburg (-CC), Jenkins s.n.; Van

der Schijff 4323. 2530 (Lydenburg): Steenkampsberg (-AA),

Letty 336; 51 km beyond Belfast on Machadodorp road, (-AA),

Letty 338; Draaikraal (-AA), Du Plessis s.n.; 27 km N. of

Lydenburg on road to Steenkampsberg (-AA), Codd 8201;

W. of Lydenburg (-AB), Letty 330; Horseshoe Cataract (-AB),

Letty 322; near The Bonnet (-AB), Letty 323; Boshoek, W.

of Steenkampsberg (-AA), Young A373; Wemmershoek (-AB),

Strey 3154; Lydenburg (-AB), Wilms 1574; Wilms s.n.; Dull-

stroom (-AC) Van der Merwe s.n.; Galpin 11922; 11923; 12

km N. of Dullstroom (-AC), Codd 6656; Western end of

Schoemanskloof (-AC), Young A3 17; near Belfast turn-off to

Lydenburg (-CA), Letty 321; Belfast (-CA), Leendertz s.n.;

Machadodorp (-CB), Galpin 13110; Letty 327; 337; 4 km

beyond Belfast to Machadodorp (-CB), Letty 328; between

Belfast and Machadodorp (-CB), Van der Schijff s.n.; Reynolds

2697 ; 2564; N.E. of Machadodorp on Lydenburg Road (-CB),

Codd and de Winter 3362; 1,5 km E. of Dalmanutha (-CC),

Codd 8088; Slaaihoek (-DA), Bruce 503; Nelshoogte (-DD),

Codd 8127. 2531 (Komatipoort): Umkomati Valley (-CC),

Galpin 1360. 2627 (Potchefstroom): Vereeniging (-DB),

Leendertz 3913. 2628 (Johannesburg): Linden-Emmerentia

road (-AA), Gilliland s.n.; Witpoortjie valley (-AA), Marloth

7770; Wilgepoort (-AD), Leendertz s.n. 2629 (Bethal): 6 km

beyond Oshoek (-DA), Letty 332; Trichardtsfontein (-DB),

Rehmann 82 (B); Adendorff 1; 9 km beyond Ermelo on Bethal

road (-DB), Letty 339; Ermelo (-DB), Letty 340; banks of

Vaal River on farm Leeuwenberg (-DB) Robertson 23. 2630

(Carolina): 32 km S.E. of Machadodorp on Slaaihoek road

(-AA), Letty 469; Carolina (-AA), Rademacher s.n.; Mock-

ford’s farm, Blairmore (-BC), Gunn s.n.; Mavieristad (-CA)

Pott 5148; 24 km S.E. of Ermelo (-DA) Codd 4771. 2729 (Volks-

rust): Maquabi (-BB), Burtt- Davy 4095; Rolfontein (-BB),

Burtt-Davy 4123; Amersfoort (-BB), Adendorff 2; Laingsnek

(-BD), Letty 494. 2730 (Vryheid): farm Oshoek (-AC), Letty

489; Devenish 1086; Piet Retief (-BB), Collins s.n.

O.F.S.— 2828 (Bethlehem): Golden Gate National Park

(-CA), Roberts 3146. 2829 (Harrismith): 4 km S.E. of Swin-

burne (-AD), Codd 10724; Van Reenens Pass (-AD), Letty

472.

Swaziland. — 2631 (Mbabane): Forbes Reef road (-AA),

Compton 29551; Hlatikulu (-CD), Collins s.n.

Natal.— ex Natal, Wood s.n. (left-half of sheet, K); 2729

(Volksrust); Majuba (-BD), Schweickerdt 974. 2730 (Vryheid):

Paulpietersburg (-BD), Galpin 9691; Utrecht (-CB), Codd

6259. 2829 (Harrismith) Mlambonjiva valley, Bergville (-CB),

Killick & Marais 2133; 2134. 2830 (Dundee): Glencoe (-AA),

Shepherd 63 (NH); Nqutu (-BA), Codd 9405. 2929 (Under-

berg): Major Stewart's farm, Underberg (-CD), Dyer 3228.

2930 (Pietermaritzburg): (-CB), Wilms 2335 (K).

Cultivated. — Hort. F. Stander & Co. (K); Hort. C. G. van

Tubergen (K); “from Natal” (K).

Zantedeschia albomaculata subsp. macrocarpa is

characterised by, and differs from subsp. alboma-

culata, in having triangular-hastate, sparsely speckled

leaves, and medium-sized spathes which are straw-

coloured and truncate at the apex. Its epithet is

derived from the, generally, few and large-sized

berries which develop in the withering spathe; these

weigh the flaccid peduncle down until it lies on the

ground and the seeds germinate.

24

THE GENUS ZANTEDESCHIA

Fig. 29. — Zantedeschia albo-

maculata subsp. macrocarpa

on the marshy banks of a

Transvaal stream, Lyden-

burg District (Codd 8201).

Photo by Dr L. E. Codd.

Plate 8. — Zantedeschia albomaculata (Hook.) Engl, subsp. valida Letty.

CYTHNA LETTY

25

For purposes of typification, the specimen of

Rehmann 82 in Berlin-Dahlem Herbarium may be

accepted as being the holotype. It consists of one

leaf and a capsule containing large berries and bears

labels with “ Zanledeschia macrocarpa ” in Engler’s

handwritting and “Transvaal, Trigardsfontein, Reh-

mann No. 82” in Rehmann’s handwriting.

In Zurich there are several Rehmann specimens

called Z. macrocarpa, including a specimen of Reh-

mann 82 annotated in Engler's handwriting. In

addition there is Rehmann 72, consisting of three

leaves, from “Natal, Drakensberg, Laingsnek” (in

Rehmann’s writing) with “Z. macrocarpa Engl.?”.

For notes on Rehmann 83 and 7096 see subsp. vali-

da (below).

(c) subsp. valida Letty, subsp. nov., subsp. macro-

carapae (Engl.) Letty affinis, planta robustiore

foliis crassioribus ovato- vel ovato-orbiculato-cor-

datis differt.

Planta robusta ca 70 cm alta. Folia petiolata;

lamina 30 cm longa, 20 cm lata, crassa, ovato- vel

ovato-orbiculato-cordata, immaculata; petiola 30

cm longa. Spatha variabilis grandis ca 15 cm longa,

12 cm lata in fundo purpurae. Baccae 1 cm diam.,

numerosae.

Type: Natal, farm Blanerne, 29 km N. of Lady-

smith, Letty 492 (PRE, holo.).

Plant robust up to 70 cm tall. Leaves petiolate,

petiole up to 30 cm long; blades thickish textured

up to 30 cm long, up to 20 cm broad, ovate-cordate

to ovate-orbicular-cordate with apiculate apex, imma-

culate. Spathes vary in size up to 15 cm long, 12

cm broad, ivory to cream coloured with purple area

at base inside. Berries medium sized about 1 cm

diam., numerous, 3 locular, 2 to 3 ovuled.

Found in rock-clefts on the mountains and foot-

hills as well as on the banks of streams and in vleis,

it flowers from October to March with the peak in

November. Its distribution seems to be restricted

to the region bounded by the Biggarsberg, Giants

Castle and Collings Pass, in Natal.

Fig. 32. — Tubers of Zantedeschia albomaculata subsp. vaiiaa

from Klip River District, Natal (Letty 492 and 493) xC

Natal. — 2829 (Harrismith): Biggarsberg (cult. Pretoria),

Marais 360; 29 km N. of Ladysmith on farm Blanerne (-BD),

Letty 492; 10 km from Biggarsberg store on road to Collings

Pass (-BD), Letty 493; on farm Blanerne, Elandslaagte, Big-

garsberg (-BD), Mitchell-Innes s.n. 2929 (Underberg): Giant's

Castle (-AD), Trauseld 302.

Without the flowering spathes, plants of Zante-

deschia albomaculata subsp. valida could easily be

mistaken for Z. aethiopica by the robust habit and

the shape and texture of the leaves. Apart from this

Fig. 33. — Zantedeschia albomaculata subsp. valida, habit of

plant figured on Plate 8, from Biggarsberg, Natal (Marais

360).

resemblance, they differ in all other respects except,

perhaps, the colour of the spathes. The nearest

affinity of subsp. valida is subsp. macrocarpa with

which it agrees in being deciduous, and also in the

shape and the colouring of the spathe, but differs

in the shape and texture of the leaves.

Among specimens examined by Dr Codd in Zurich,

one labelled “ Rehmann 83, Transvaal, Trigards-

fontein” and annotated as Z. macrocarpa by Engler

has leaves somewhat larger and broader than Reh-

mann 82 (the type of Z. macrocarpa), and is practi-

cally identical with another specimen which bears

a rubber stamp with the words “7096, Drakensberge,

Biggarsberge”. As these two specimens Rehmann 83

and 7096 are considered to represent Z. albo-

maculata subsp. valida, the locality given for No.

7096 appears to be correct.

Z. lutwychei N.E. Br., described in Gardener’s

Chronicle 13: 568 (1893), was grown from speci-

mens said to have been imported from the neigh-

bourhood of Lake Nyanza and called “Pride of

the Congo”. It was illustrated in Rev. Hort. Paris,

p. 60 (1896). Although the Zantedeschia specimens

in several European herbaria have been seen, no

specimens from the Congo have been encountered.

The existing material of Z. lutwychei , which has been

examined, consists of a single leaf in Kew Herbarium

of the plant described in the Gardener’s Chronicle

(1893), and the illustration in Rev. Horticole (1896).

From this evidence it seems that Z. lutwychei re-

sembles Z. albomaculata subsp. valida but, until

matching specimens from the supposed area of

origin of Z. lutwychei are forthcoming, its status is

doubtful. It is, therefore, treated as a nomen dubium.

26

THE GENUS ZANTEDESCHJA

Arodes Kuntze

aethiopicum (L.) Kuntze

albomaculatum (Hook.) Kuntze

angustilobum (Schott) Kuntze . . ,

hastatum (Hook.) Kuntze

melanoleucum (Hook.f.) Kuntze

Aroides Heist, ex Fabric

aethiopicum (L.) Heist ex Fabric.

albomaculatum (Hook.) Kuntze

angustilobum (Schott) Kuntze . . ,

hastatum (Hook.) Kuntze

Arum L

aethiopicum Herm

africanum Herm

Calla L

aethiopica L

elliottiana Watson

oculata Lindl

pentlandii Whyte ex Watson

Colocasia Link

aethiopica Link

Otosma Rafin

aethiopica (L.) Rafin

Richardia Kunth

africana Kunth

albomaculata Hook

angustiloba Schott

elliottiana Watson

hastata Hook

Lehmannii

macrocarpa (Engl.) Watson . . .

me/anoleuca Hook.f

INDEX

8

8, 9

8,18

8, 18

7, 8

9

18

5

6

6, 7

6, 7, 9, 10

14

8, 18, 21

15

6, 7, 8

7, 9

8

9

8

6, 8, 9

8, 18, 21

15, 16, 18,21,22

14

. . . 8, 18, 21, 22

11

23

8, 18, 22

me/anoleuca Hook.f. var. tropicalis N.E.Br.

pentlandii Watson

rehmannii (Engl.) N.E.Br

rhemannii Watson

sprengeri Comes

Stehmannii

Richardsonia Kunth

Zantedeschia Spreng

Z. aethiopica (L.) Spreng

var. minor Engl

var. umganiensis Leicht. & Engl

albomaculata (Hook.) Bail I

subsp. albomaculata

forma latifolia Engl

subsp. macrocarpa (Engl.) Letty

subsp. valida Letty

angustiloba (Schott) Engl

aromatica (Roxb.) Spreng

chloroleuca Engl. & Gilg

elliottiana (Watson) Eng!

hastata (Hook.) Engl

jucunda Letty

lutwychei N.E.Br

macrocarpa Engl

melanoleuca (Hook.f.) Engl

var. concolor Burtt Davy

var. tropicalis N.E.Br

occulta (Lour.) Spreng

oculata (Lindl.) Engl

pentlandii ( Watson) Wittm

rehmannii Engl

tropicalis (N.E.Br.) Letty

15

11

15, 17

11

5, 6, 8, 9

9

8, 15, 17, 18,21

. 9. 15, 18, 21

18

... 6, 9, 22, 23

6,9,25

. . . 8, 16, 18, 22

5

18, 22

5, 9, 14

18

6, 9, 13

25

. . . 8, 15, 23, 25

18,22

23

18, 22

5

18,21

. ... 6, 9, 15, 22

8,9,11

18,22

Bothalia 11, 1 &2: 27-52 (1973)

The Genus Crinum

in Southern Africa

I. C. VERDOORN*

ABSTRACT

The 21 species of Crinum recognized in southern Africa are reviewed and illustrated, mostly in

colour. Five of these were described by the author during the last two decades. A key to the species

and notes on the history, diagnostic features and distribution in southern Africa are provided.

INTRODUCTION

Linnaeus described four species of Crinum in his

Species Plantarum, 1753, and the following year the

genus was defined in the 5th edition of his Genera

Plantarum. The four species listed were C. latifolium ,

C. asiaticum, C. americanum and C. africanum. Be-

cause C. latifolium is the first in the list, Uphof (1942)

considers it to be the type species. But Hitchcock &

Green proposed C. americanum as the lectotype, in the

International Rules of Botanical Nomenclature, p.

140, 1935, and Traub (1963) follows them. Linnaeus

described Crinum americanum in his Hortus Cliffor-

tianus in 1737 and this supports the choice made by

Hitchcock and Green.

Of these four species described by Linnaeus, three

belong to the genus as defined today, but C. africanum

is now classified as Agapanthus.

For many years after the establishment of Crinum

we find members of this genus being described among

the assortment of Amaryllids under the name Amaryl-

lis. In 1837 appeared the interesting and useful publi-

cation “Amaryllidaceae” by W. Herbert. He defined

the genus Crinum satisfactorily and described 46 spec-

ies and a number of varieties. These included several

which had originally been described as species of

Amaryllis. In addition to his descriptive or taxonomic

work, The Honourable and Reverend William Herbert,

Dean of Manchester, kinsman of the Earl of Car-

narvon, cultivated Amaryllids in his famous garden at

Spofforth and experimented with hybridization.

Fifty-one years later, in 1888, Baker’s “Handbook

of the Amaryllidaceae” appeared. J. G. Baker recog-

nised 79 species and divided them into three subgenera.

It is interesting to note that the two Linnaean species

which have claim to being the type species, C. latifo-

lium and C. americanum, fall into different subgenera

as defined by Baker: C. latifolium in the subgenus

Codonocrinum, defined as “perianth funnel-shaped,

tube permanently curved; stamens and style conti-

guous, declinate”, and C. americanum in the subgenus

Platyaster with “perianth erect hypocrateriform, seg-

ments lanceolate; stamens spreading”. The third

subgenus is Stenaster “perianth erect hypocrateriform,

segments linear; stamens spreading”. If species are

grouped according to these divisions, one finds an

assortment of forms in each group, that is, while spe-

cies agree in the diagnostic features, they may differ

considerably in other respects. In addition, there are

species which do not exactly fit into the definition of

any of the three subgenera. For instance C. campanu-

latum and C. moorei are both listed by Baker and

‘Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101 Pretoria.

subsequent authors in Codonocrinum but the one, C

campanulatum, has the lobes forming a distinct cup

and the stamens not declinate, and C. moorei has sub-

spreading lobes and declinate stamens. Because of the

occasional odd forms found in this genus it seems

advisable not to attempt any subdivisions.

Baker is also responsible for the revision of the genus

Crinum in the Flora Capensis (1896) and Flora of

Tropical Africa (1898). In South Africa only 8 species

were recognised at the time and in tropical Africa 38.

In a most helpful article Uphof (1942) sums up the

major and minor works on the genus and lists all the

species described to that date, 130 in all, classifying

them, where possible, under Baker’s subgenera, with

15 unclassified species. Recent contributors to our

knowledge of the genus are H. P. Traub and L. S.

Hannibal, members of the Amaryllis Society of

of America. An example of the work of each of

them is listed under the references Traub (1963)

and Hannibal (1967). In South Africa a cytological

study published by Van der Walt, Geerthsen and Rob-

bertse (1970), is of particular value because the mater-

ial worked with was identified at this Institute (Botani-

cal Research Institute). There have been some name

changes since the paper was prepared but it is known

to which species the names refer. For instance “C.

forbesii” is the recently described C. paludosum and

the “new species” has been described as C. foetidum.

DIAGNOSTIC FEATURES

The most striking feature about all species of Cri-

num, except C. moorei, growing in nature or cultivated

in the open in South Africa is that leaves more than

one year old have lost their tips by a clean cut. This is

because the leaves die back to the base in the winter

with the result that in spring the previous seasons

leaves grow out again with truncate tips, while only a

few leaves in the centre are new and have their apices

intact. This is not the case in C. moorei where all the

leaves in a season are intact and the many leaf-scars

on the false stems show no sign of growing out for a

second year. In this connection it is illuminating to

read Hannibal’s observations (1967) which in effect

state that the leaves of Crinum species from semi-arid

areas are slow growing and take three years to reach

maturity whereas the leaves of C. moorei from “hea-

vier rainfall areas with dry seasons” mature in one

year. The leaves of all Crinums are rosulate or rarely

distichous (but not biflabellate), and definitely sheath-

ing at the base. They vary in colour, width, ciliation

and texture. The solid scapes are short or long and

naked except for two terminal spathe-valves. The um-

bels bear from 1 to 40 flowers which may be subsessile

28

THE GENUS CR1NUM IN SOUTHERN AFRICA

or pedicellate and the fruit in some species is beaked

by the persistent base of the perianth-tube. The length

of the tube, the form in which the segments are ar-

ranged and whether they are keeled with red or not,

are of specific significance and possibly to a lesser ex-

tent, the length of the peaked apices of the segments.

This peak is formed by the true apex being introrse and

forming a peak like a dunce’s cap. This false apex or

peak may be short and blunt in some species or long

and pointed in others. Another interesting feature is

the way the flowers fade. In some species they turn

brown and in others red. The tunicated bulbs are

usually large and they either narrow gradually toward

the apex or abruptly into a short or long neck. In

some cases the neck, formed by the sheathing bases of

the leaves, elongates into a short, up to 60 cm long,

leafy stem-like structure, and in one species, C. moorei ,

this structure grows quite tall and the leaf-bases harden

and thicken forming a false stem leafy only in the

upper portion. This species is also unique in South

Africa in having ensiform leaves that do not grow out

for a second year and they have an indication of a

midrib. In all Crinums the seed-coats of the large,

more or less globose seeds are impervious to water

and an analyses of the seed by Isaac and McGil-

livray (1965) has shown that the water content is

very high. This accounts for a peculiarity of the

genus, namely that the seeds germinate readily with-

out water and in the driest conditions. It also supports

the claim (Hannibal, 1967) that the seeds are dis-

persed by water. The seed coats are usually smooth

and only in one species, C. foetidum, have they been

found to be papillate. A feature observed by Herbert

(1837) is that the ovary, which is situated between the

pedicel and the perianth tube is “thickest in the mid-

dle”. In some species new bulbs form prolifically

round the mother bulb.

DISTRIBUTION

The genus is found in the tropics of Asia, Africa and

America and extends into the temperate regions of

both hemispheres: in the south to South Africa and

Australia and, in the north, to Japan in the Old World

and the southern regions of the U.S.A. in the New

World. Over 150 species have been described but this

number may be much reduced when the genus is stud-

ied as a whole. Over half the number are from Africa.

In South Africa plants are found sparsely on moun-

tain- or hill-slopes with a concentration in low lying

areas, on river banks and at the coast. Some species

flourish in marshes and pans.

AUTHOR'S NOTE

It is regretted that this account deals only with the

southern African representatives. Species of Crinum,

as in many other genera, are difficult to assess from

dried specimens or from descriptions, unless the diag-

nostic features are known. In order to establish such

features the living plants in their natural habitat have

to be studied. Throughout many years, working at

endless identifications in the National Herbarium, 1

have, from time to time, been faced with the problem

of identifying a Crinum. Each case required investiga-

tion, first to know the plant in nature and something

of its distribution and then to search literature, old

and new, for its correct name. So that, by final retiring

time, it seemed logical to publish the data so far col-

lected in order that future taxonomists may extend the

study to the rest of the genus. This may result in name

changes as far as specific epithets are concerned, for the

distribution of the species may be much wider than our

present records show. In the notes under some species a

possible relationship with an extra South African spe-

cies is mentioned. However it is my hope that a future

reviewer will resist the modern tendency to make dras-

tic hanges at the generic level. It is conceivable that

each reviewer will in turn redefine sections of this genus

and possibly include its three related African genera,

Amaryllis, Ammocharis and Stenolirion in the rearran-

gement. It has not been possible to place all our species

in the subgeneric divisions that exist because of inter-

mediates and anomalies. It seems wiser therefore to

draw attention to the distinctions but not to erect

watertight compartments for them.

REFERENCES

Baker, J. G., 1888. Bak. Harnib. Amaryll. London: George

Bell & Sons.

Baker, J. G., 1896. Amaryllidaceae. FI. Cap. 6: 171-246.

Baker, J. G., 1898. Amaryllidaceae. FI. Trop. Afr. 7: 376-413.

Hannibal, L. S., 1967. Evolution of Crinum. Paper presented

at the Rhodesian Scientific Congress May, 1967. Not yet

published?

Herbert, W., 1837. Herb. Amaryll. London: James Ridgeway

& Sons.

Isaac & McGillivray, 1965. A Note on the Water-storing

Seeds of Two Species of Crinum and of Some Other S.A.

Amaryllidaceae. Ann. Bot. n.s. 29:739.

Traub, H. P., 1963. The Genera of Amaryllidaceae. California:

The American Plant Life Society.

Uphof, J. C. T., 1942. A Review of the Species of Crinum.

Herbertia 9: 63-84.

Van der Walt, E. L., Geerthsen, J. M. P. & Robbertse, P. J.,

1970. ’n Sitogenetiese Studie van die genus Crinum Linn, in

Suid-Afrika. Agroplantae 2: 7-14.

CRINUM

Crinum L., Sp. PI. 291 (1753); Gen. PI. ed. 5: 141,

No. 366 (1754); Herb., Amaryll. 242 (1837); Benth. &

Hook.f., Gen. PI. 3, 2: 726 (1883); Bak., Handb.

Amaryll. 74 (1888); FI. Cap. 6: 198 (1896); FI. Trop.

Afr. 7: 373 (1898); Phill., Gen. ed. 2: 203 (1951); Up-

hof in Herbertia 9: 63 (1942); Traub, the Genera of

Amaryllidaceae 60 (1963). Type species: C. america-

num L.

Ammocharis sensu Miln-Redhead & Schweick. in

J. Linn. Soc. 52: 170 (1939), partly, as to A. baumii.

Herbaceous plants with large tunicated bulbs which

are sometimes produced into a neck or rarely into a

false stem made up of the indurated sheathing bases of

the old leaves; in some species new bulbs are readily

produced laterally. Leaves linear, lorate or ensiform,

sheathing at the base, rosulate or distichous (but not

biflabellately spreading) often dying back in the winter,

the previous season’s leaves growing out again in spring

with a few new leaves in the centre. Inflorescence aris-

ing laterally, umbellate, 1 -many-flowered; peduncle

solid ; spathe-valves 2 ; bracteoles at base of each flower

narrowly linear. Flowers subsessile or pedicelled;

perianth-tube long; segments linear to broadly lanceo-

late, the inner slightly broader than the outer, spread-

ing or connivent into a campanulate or trumpet shape.

Stamens inserted in the throat of the perianth-tube;

filaments arcuate, ascending or declinate; anthers

versatile. Ovary inferior, obvious as a swelling be-

tween the apex of the pedicel and the base of the pe-

rianth tube, turbinate. Fruit subglobose, bulging with

the large seeds, eventually bursting irregularly to re-

lease the subglobose seeds, sometimes beaked with the

persistent base of the perianth-tube. Seeds subglobose.

I. C. VERDOORN

29

A genus with over 130 species described. Distributed

throughout the tropics and warm temperate regions

of the Old and New Worlds. The largest number of

species are found in Africa. There are 21 species in

South Africa, occurring in all the provinces but absent

from the Penisula.

Key to species

Perianth-tube straight, cylindric; segments narrow, mostly under 8 mm wide, spreading abruptly and

recurving; stamens not declinate:

Dwarf species; inflorescence under 20 cm high:

Umbel 1-2-flowered; perianth-tube about 12 cm long 1. C. baumii

Umbel about 7-flowered; perianth-tube about 2 cm long 2. C. nerinoides

Inflorescence over 20 cm high; umbels many-flowered; buds cernuous:

Leaves distichous; plants grow in dry areas 3. C. buphanoides

Leaves rosulate; plants usually grow in marshy country 4. C. crassicaide

Perianth-tube usually curved outwards; segments broad, mostly over 1 cm wide, conniving into a cup,

funnel or trumpet shape:

Leaves distichous 5. C. euchrophyllum

Leaves rosulate :

Leaves with a longitudinal thickened line suggestive of a midrib:

Perianth segments subspreading or forming a wide (loose) funnel, white or pink (without a red keel);

plants with a stem-like structure, leafy in the upper portion; leaves broad, ensiform .... 10. C moore i

Perianth segments closed to form a trumpet shape, nodding, distinctly keeled with red; leaves long

acuminate 11. C. kirkii

Leaves without a midrib:

Perianth segments shallowly cup-shaped in the lower half, upper half subspreading, suffused with

rose; stamens not declinate; plants growing in marshes or pans 6. C. campanulatum

Perianth segments deeply cup-shaped or funnel- or trumpet-shaped; stamens declinate:

Open flowers suberect, segments predominantly white; plants growing in marshes or pans:

Median leaves up to 5 cm broad measured shortly above the base, flaccid, somewhat glaucous

green; flowers up to 7 in an umbel 7. C. palndosum

Median leaves under 2 cm broad above the base :

Leaves firm, suberect, U-shaped in cross section 8. C. rautanenianum

Leaves flaccid, spreading, shallowly canaliculate 9. C. carolo-schmidtii

Open flowers spreading horizontally and then becoming nodding:

Peduncle short, 4-15 cm long, rarely slightly longer; umbel 1- rarely 2-flowered:

Leaves stiff, erect, about 10 mm broad, canaliculate, finely toothed on the margin. ... 12. C. acaide

Leaves flaccid, narrow, 3 mm broad, smooth on the margins 13. C. minimum

Peduncle 15-90 cm long, umbels few to many flowered:

Perianth-tube short, about 3 cm long, shorter than the segments:

Leaves grass-green, rather stiff; perianth becoming pink as it fades:

Leaves up to 2 cm broad; pedicels 0-2,5 cm long 14. C. hneare

Leaves up to 5 cm broad; pedicels up to 5 cm long 15. C. variabde

Leaves glaucous-green, flaccid, usually undulate, spreading at ground level; perianth be-

coming light brown as it fades 20. C. macowanii

Perianth-tube long, 7 cm or more long, segments about as long or shorter than the tube:

Leaves comparatively few, about 1 1 , broad, up to 1 5 cm broad, spreading along the ground :

Leaves blue-green (not glaucous), rather thin; perianth predominantly white with long

acuminate segments 16. C.foetidum

Leaves dark or grass green, rather thick; perianth with a distinct purplish-pink keel or

suffused with purplish-pink; segments not acuminate:

Peduncle up to 30 cm long; pedicels short, up to 2 cm long; perianth-tube persisting

like a beak on the fruit 17. C. grammicola

Peduncle up to 60 cm long or more; pedicels up to 7 cm long; perianth-tube dying

back to a short crown on the fruit 18. C. delagoense

Leaves many, rarely few but then not spreading along the ground but suberect with the

upper portion arching downwards:

Leaves narrow, the majority under 3 cm broad at the middle:

Leaves strongly undulate at first and spreading, the inner suberect and not undulate,

margins finely ciliate-dentate; perianth-segments forming a rather wide trumpet

shape with revolute apices; old flowers drying brown; fruit beaked 19. C. lugardiae

Leaves never undulate, suberect, margins smooth; perianth-segments spreading at the

apex old flowers drying reddish; fruit not beaked; plants restricted to the eastern

Cape coastal region 14- C- Hneare

Leaves broad, the majority over 5 cm broad at the middle:

Perianth-segments forming a wide trumpet shape; anthers black showing up on the

white inside; old flowers drying brown; fruit beaked; new leaves few in the

centre 20- c • macowanii

Perianth-segments forming a narrow trumpet shape; anthers grey or brown, old

flowers drying pink or red; fruit not beaked; new leaves many 21. C. bubispenmum

5292—3

30

THE GENUS CRINUM IN SOUTHERN AFRICA

1. Crinum baumii Harms in Warb., Kunene —

Samb. Exped. 199 (1903); Verdoorn in Flow. PI. Afr.

36: 1. 1432 (1964); Solch in Prodr. FI. S.W.Afr. 150:

6 (1969). Type: Angola, Chirumbu, Baum 273.

Ammocharis baumii (Harms) Milne-Redhead & Schweick. in

J. Linn. Soc. Bot. 52: 187 (1939).

Bulb 2-5 cm long, abruptly narrowed into a neck

1-5 cm long, covered entirely with chartaceous tunics

produced at the apex to surround the new leaves at the

base. Leaves rosulate, about 6, erect, canaliculate, 9

cm long, or longer, 2 mm broad. Peduncle terete, green,

erect, 3-6 cm long. Spathe-valves 3-5 cm long, about 5

mm broad at the base. Umbel 1-2-flowered; pedicels

0-5 mm long. Perianth white turning pink; tube

slender, erect, about 12 cm long; segments white with

a deep rose keel, abruptly spreading, revolute, linear,

5 cm long, 5 mm broad. Stamens arcuate, ascending;

filaments white turning rose, about 4 cm long; anthers

about 7 mm long. Ovary green, about 5 mm long, 4

mm broad. Fruit subglobse, tinged with red, about 15

mm diam, with an apical crown (not beaked). Plate 1.

Grows in the flood plains of rivers. Recorded from

South West Africa and Angola.

S.W.A. — 1719 (Rundu): about 5 km S. of Rundu (-DD), De

Winter 3774.

This charming Crinum was collected by Dr. B. de

Winter in the flood plains of the Okavango near

Rundu. It was locally common and grew in deep white

sand. Besides the type collection which is from near

Chirumbu, Angola, “in damp soil” along the Kuban-

go, only one other record exists, recorded by Milne-

Redhead & Schweickerdt (1939) but has not been seen

by me. It is from Huilla which lies in the drainage

basin of the Kunene River, and was collected at the

Gambos Mission Station by F. Newton, the Portugu-

ese naturalist, who collected in Angola and on St.

Thomas Island.

When Milne-Redhead and Schweickerdt transferred

this species to Ammocharis they expressed some un-

certainty because they did not know whether the ar-

rangement of the leaves was biflabellate or not (bifla-

bellate arrangement is one of the diagnostic features

of Ammocharis). In our growing plants it was obvious

that the leaves were rosulate. Whether this alone is

sufficient to separate Crinum from Ammocharis has

not yet been determined. As long as one has the species

well defined it seems of little real consequence. Here

the original classification is followed.

2. Crinum nerinoides Bak. in Bull. Herb. Boiss.

ser. II, 3: 666 (1903); Milne-Redhead & Schweick. in

J. Finn. Soc. 52: 189 (1939), as a doubtful species of

Ammocharis ; Solch in Prodr.Fl. S.W. Afr. 150: 3

(1969) as a doubtful species of Ammocharis. Type:

Herero-land, Dove s.n. (Z).

Bulb 2,5 cm long, produced into a short or long

neck, 2-3,5 cm broad, tunics produced at the apex to

cover the base of the leaves and peduncle. Leaves about

5, narrowly linear, 5-15 cm long, about 0,2 cm broad,

apex of outer leaves truncate (leaves having grown

out after dying back). Peduncle 4-10 cm long, slender.

Spathe-valves 3-4 cm long. Umbel 3-7-flowered; pe-

dicels 1-2 cm long. Perianth “pink”, “deep rose”;

tube slender, straight, 1-2,5 cm long; segments even-

tually spreading and revolute in the upper quarter,

about 3, 5-4, 5 cm long, 0,4-0, 6 cm broad. Stamens

arcuate erect, filaments about 3-3,5 cm long. Ovary

about 5 mm long, 3 mm diam. Fruit not seen. Fig. 1.

Grows in “loamy soil on the edge of a pan” ^“flood-

ed limestone”. Recorded from the Grootfontein and

Gobabis district of South West Africa.

S.W.A. — Hereroland, without precise locality, Dove s.n. (Z).

1917(Tsumeb): Farm Heidelberg (-BB), W. & E. Walter 331 ,

sub. WIND 382. 2218 (Gobabis): Witvlei (-AD), Mason &

Boshoff 2516.

This species, like C. baumii, is known in herbaria

from very few gatherings. In addition to the type, only

three specimens have been collected and all within the

last two decades. From the notes on these specimens

one gathers that the plants are restricted to groups

around limestone pans.

Fig. 1. — Crinum nerinoides,

a pressed specimen in the

Windhoek Herbarium, Wal-

ters 331 from near Tsumeb,

South-West Africa.

Plate 1. — Crinum baumii Harms. From Flowering Plants of Africa, Vol. 36, PI. 1432 (1964).

I. C. VERDOORN

31

3. Crinum buphanoides Welw. ex Bak. in J. Bot.

Lond. 16: 195 (1878); Bak., Handb. Amaryllid. 80

(1888); FI. Trop. Afr. 7: 398 (1898); Yerdoorn in

Flow. PI. S.Afr. 23: t.887 (1943); Solch in Prodr. FI.

S.W. Afr. 150: 6 (1969). Type: Angola, Pungo Andon-

go, Welwitsch 4016.

C. leucophyllum Bak. in Bot. Mag. t.6783 (1884). Type:

Damaraland, without precise locality, Een s.n., cult. Kew.

Bulb large, about 15 cm diam., produced into a neck

of varying lengths. Leaves light green, distichous,

sheathing at the base, forming a leafy false stem up to

about 30 cm high, the lowest the broadest, up to 15

cm broad, each successive leaf above narrower, the

uppermost (the youngest), about 1 cm broad, margins

ciliate with minute cartilaginous teeth. Peduncle lateral

stout more or less 45 cm long, 2,5 cm broad and 1 ,5

cm thick. Spathe-valves about 10 cm long, 3,5 cm wide

at the base, reflexed; bracteoles white, narrowly linear,

about 5 cm long. Umbel 13-40-flowered. Pedicels

1 ,5-4,5 cm long. Perianth with a long slender tube,

7-10 cm long; segments white to pink, usually with a

deep red keel, narrow and spreading abruptly, recurv-

ed, about 6,5 cm long, usually under 1 cm wide, lower

portion with infolded margins covering the basal

portion of the filaments, peaked apex usually under

5 mm long, reflexed apex bearded. Stamens arcuate

erect, filaments rose in the upper portion. Style deep

rose in the upper part. Fruit not seen. Fig. 2, 3, 4.

Grows in deep sand or loam in dry areas. Recorded

from the northern regions of South West Africa, and

the northwestern, northern central and eastern Trans-

vaal. Also occurs in Ngamiland and Angola.

S.W. A. — 1715 (Ondangwa): 19 km E. of Ondangwa (-BC),

De Winter & Giess 6927. 1821 (Andara): Island opposite

Shitangadimbo Camp, De Winter & Marais 4801. 2115 (Kari-

bib): Omaruru (-BD), Barnard 59. 2215 (Trekkopje): Onanis

(-DD), Hardy & De Winter 1402.

Fig. 2. — Crinum buphanoides,

showing the distichous

leaves, grown in Pretoria,

originally from Elmeston,

N.W. Transvaal ( Codd

2124).

Fig. 3. — Crinum buphanoides,

near Francistown. Photo

by Herbert Lang.

32

THE GENUS CR1NUM IN SOUTHERN AFRICA

Transvaal. — 2229 (Waterpoort): Eyem, north of Blaauw-

berg (-CC), Obermeyer, Schweickerdt & Verdoorn 86. 2230

(Messina): Messina, Pole Evans sub PRE 13121 A. 2231 (Pafuri):

Punda Milia (-CC), Lang sub TRV 32276. 2327 (Ellisras):

Elmeston (-DC), Codd & Erens 2124. 2328 (Baltimore): 32 km

N. of Potgietersrust (-DD), Galpin 11580. 2330 (Tzaneen):

Baiandbai (-BB), Lang sub TRV 32255; 24 km from Duiwels-

kloof on the Louis Moore Mine road, (-CA), Thompson s.n.

2427 (Thabazimbi): 32 km N.W. of P.O. Rooiberg on road to

Krantzberg (-DA), Codd & Erens 2071. 2428 (Nylstroom):

banks of the Palala River near Naboomspruit (-DA), Galpin

14665 ; 24 km N.E. of Naboomspruit on road to Potgieters-

rust (-BD), Reynolds 4049. 2431 (Acornhoek); Tshokwane

(-DD), Van der Sehijff 1422; Sabie River, Skukuza (-CD),

Stevenson-Hamilton s.n. 2528 (Pretoria): Rust de Winter

(-BA), Codd 6229. 2531 (Komatipoort): 9 km N. of Malelane

(-BC), Codd 5509; 15 km W. of Skukuza (-AA), Codd 5700;

Lower Sabie (-BB), Lang in TRV 30409; Barberton (-CC),

Williams in TRV 6227; Thorncroft in TRV 13428; Ellers s.n.;

Klokwane (-AD), Van der Sehijff; between Klokwane and

Mhamban (-AD), Van der Sehijff 3252; 3946; Gomondwane

(-BD) Van der Sehijff 4029.

Eig. 4. — Bulb of Crinum buphanoides from near Krantzberg,

Transvaal.

The clearly distichous arrangement of the leaves is

unusual in South African species of Crinum. This

feature is not obvious from most herbarium specimens,

that is if there are no photographs of the growing plant

attached or if the collector has not made a note about

the leaf arrangement. In the original description there

is no mention of distichous leaves and our concept of

the species is based on a specimen named at Kew in

1935 ( Obermeyer , Schweickert & Verdoorn 1937). From

the Botanical Magazine Plate, No. 6783, and from the

origin of the specimen figured, Damaraland, there can

be little doubt that C. leucophyllum is synonymous.

For the features by which to distinguish C. buphanoides

from the following species, C. crassicaule, see the notes

under that species.

4. Crinum crassicaule Bak ., Handb. Amaryll. 85 ]

(1888); N. E. Brown in Kew Bull. 1909: 142 (1909) (

partly, excluding the description and Lugard 45; Ver-

doorn in Bothalia 10: 57 (1969), in notes under C. I

foetidum n. sp. Type: Ngamiland, Koobie, Chapman |

s.n., painted by Baines (K, holo. ; PRE, photo.).

Bulb large, more or less 24 cm long, 14 cm diam.,

narrowing at the apex into a rather thick neck about |

7 cm diam. and varying in length. Leaves rosulate,

the lower spreading from near the base, arching down-

wards, the new entire leaves in the centre few, suberect,

tapering towards the apex, median leaves about 7 cm

broad; margins with a thin cartilaginous border and

small, fairly widely spaced, cartilaginous hairs. Pedun-

cle lateral about 50 cm long, 2-3 cm broad, 1,3 cm

thick, often tinged purplish brown. Spathe-valves fad-

ing soon and reflexed, about 7 cm long 2 cm broad at j

the base; bracteoles narrowly linear, white, about 4

cm long. Umbel 8-16-flowered. Pedicels 1 ,3-4 cm long.

Perianth with a long slender tube 9-1 5 cm long, usually

tinged brownish purple, buds erect at first, then cer-

nuous and as the flower opens erect again; segments /

spreading abruptly and reflexed from the base; ,

white or pale pink on the surface with a dorsal purplish 1

or brownish red band down the centre, 5,5-8 cm long,

1-1,3 cm broad in the upper two thirds with the peak-

ed apex usually over 5 mm long and green. Stamens

arcuate erect (not declinate); filaments a brilliant

maroon red in upper two thirds. Style brilliant red.

Fruit not seen. Plate 2; Fig. 5.

Found in or near river swamps. Recorded from the

Caprivi Strip. Also from the Okavango swamps in

Ngamiland and along the Mashi River in Barotseland.

S.W.A. — 1724 (Katima Molilo): Mpola, 24 km from Katima

Molilo on road to Ngoma, Killick and Leistner 3296. 1821

(Andara): on the Okavango River N. of Mohembo, Renew

s.n., cult. PRE, G.N. 22574; 22575.

Fig. 5. — Crinum crassicaule, showing leaves not distichous and

fading flowers with segments hanging down. Originally

from the Caprivi Strip (Renew G.N. 22575).

Plate 2 .—Crinum crassicaule Bak. From the Caprivi Strip, grown in Pretoria ( Killick & Leistner 3296).

Plate 3. — Crinum euchrophyllum Verdoorn. From Flowering Plants of Africa, Vol. 42, PL 1642 (1972).

I. C. VERDOORN

33

Fig. 6. — Bulbs of Crinum crassicaule from Nangweshi, Barotse-

land. Photo by Dr. L. E. Codd.

The interesting features observed on the growing

plants, among which were specimens collected by L. E.

Codd in Barotseland, are not obvious on herbarium

specimens. The young buds are stiffly erect. One in-

florescence observed commenced with 4 such buds and

seemed to bear out Baines’s remarks: “the pedicels

are arranged not in a circle but four lozengewise in the

centre then a row of four on each side etc.”. The

second stage finds the much elongated tubes and larger

buds (but still closed) cernuous. Next, as the flowers

open they become erect again and the segments spread

and rapidly become reflexed from the base, not arcuate

recurved as in C. buphanoides. In the meantime, the

following series in the same umbel were going through

the same three stages. The segments of the flowers are

slightly broader than those of C. buphanoides and the

peaked apices somewhat longer.

Besides the differences in degree mentioned above

and the rosulate arrangement of the leaves this species

differs from C. buphanoides in distribution and ha-

bitat. C. buphanoides grows in the dry areas of the

Transvaal, Botswana, S.W. Africa and Angola whereas

our species grows in swamps. It has been recorded

from the drainage basin of the Okavango and the Zam-

besi. A future reviewer of the genus should investigate

C. subcernuum Bak. from the Zambesi delta. This

could be our species and since it was described in 1881

would be the correct name for it.

5. Crinum euchrophyllum V erdoorn in Flow. PI.

Afr. 42: t. 1642 (1972). Type: South West Africa,

eastern Caprivi, McFerren sub PRE 31940 (PRE,

holo.).

Bulb subglobose, about 7 cm long, 6 cm diam.,

narrowed into a neck. Leaves distichous, about 12,

glaucous, flaccid, arcuate, canaliculate, the median

about 70 cm long, 2,5 cm broad near the base, width

increasing in the lower leaves to about 4 cm broad at

the base, narrowing towards the apex, margin with

fairly sparse, minute cartilaginous teeth. Peduncle erect,

about 30 cm long, slightly laterally compressed, 1,5

cm broad. Spathe-valves erect, about 8 cm long, 2 cm

broad at base, narrowing towards the apex. Umbel

1-2-flowers opening one at a time, buds at first erect

then spreading, open flower suberect, white, as it

matures it becomes spreading to cernuous and turns

a deep rose colour. Pedicels 0-5 mm long. Perianth

with a tube about 11 cm long; lobes with a faint rosy

keel, conniving to form a rather wide funnel shape, 10

cm long, the outer about 2 cm wide in the centre,

inner about 3 cm wide. Capsule not beaked. Seeds

subreniform, pale green, smooth. Plate 3, Fig. 7, 8, 9.

Found in areas subject to annual flooding. Recorded

from the eastern Caprivi.

Fig. 7. — Crinum euchrophyllum showing the distichous leaves

and 1-flowered umbel.

Fig. 8. — Crinum euchrophyllum with a 2-flowered umbel

34

THE GENUS CRINUM IN SOUTHERN AFRICA

Fig. 9. — Bulb of Crinum euchrophyllum, grown in a pot.

S.W.A. — 1723 (Singalamwe): eastern Caprivi, McFerren.

cult. PRE, G.N. 22582; McFerren sub P RE 31940.

The most striking feature of this species is the dis-

tichous, gracefully arching leaves which are covered

with a very fine bloom giving them a pleasing silvery

gray colour. The flowers, opening one at a time, are

short-lived and strikingly beautiful. When open the

single flower is suberect and a glistening white. As it

closes it becomes spreading and turns a deep rose co-

lour or, under some conditions, only faintly pink.

6. Crinum campanulatum Herb, in Bot. Mag. sub.

t.2 1 2 1 (1820); Amaryll. 270 (1837); Bak., Handbook.

Amaryll. 92 (1888); FI. Cap. 6: 200 (1896); Uphof in

Herbertia 9:81 (1942); Verdoorn in Flow. PI. Afr. 37:

1. 1455 (1965). Type: Bathurst district, between Kaf-

firdrift and Blaaukrantz, Burchel! 3785.

C. aquaticum Burch, ms.; Herb, in Bot. Mag. t.2352 (1822),

nom. illeg.

Bulb fairly small, about 7 cm long, 4 cm diam., nar-

rowing into a neck 5 cm or more long, producing new

bulbs laterally. Leaves linear, from 50 cm to over 1 m

long, 14-25 mm wide near the base, canaliculate above,

margins with a narrow cartilaginous edge and sparsely

ciliate with cartilaginous hairs. Peduncle lateral, 30

cm or more long, about 1 cm broad near the base, 7

mm thick. Spathe-valves up to 6 cm long, overlapping

in the lower half to appear united and encasing the

bases of the flowers, (not early reflexing); bracteoles

narrowly linear. Umbel few-flowered, 4-7-flowered;

pedicels 1-4,5 cm long. Perianth with a cylindric, cer-

nuous tube, crateriform at the apex, and about 5 cm

long; segments 4-5 cm long, 1 ,5-2 cm broad in the

middle conniving to form a shallow cup in the lower

half with only the apices spreading to somewhat re-

curved, white, mottled or suffused with shades of

rose, becoming deep rose as they mature, usually with

a green eye in the centre within. Stamens equally dis-

posed, subarcuate erect, faintly suffused with rose.

Style eventually overtopping the stamens, stigma ca-

pitate. Fruit subglobose, about 2,5 cm diam., not

beaked. Plate 4; Fig. 10.

Found in periodic marshes, vleis or pans in the

eastern Cape. Recorded from the Alexandria, Albany,

Bathurst, and East London districts.

Cape. — 3227 (Stutterheim): near Fort Jackson (-DC), Dodd

in Herb. Galpin 7957. 3326 (Grahamstown): Bathurst district,

Sidey 3337; Albany district, Story s.n.; Prior; s.n. Grahamstown

Flats, Galpin 7958; Waaiheuwel turn off on Salem-Alexandria

road (-CB) Archibald 5370; 9 km S.E. of Salem (-DA), Acocks

23289; Jagersdrift (-DA), Zeyher s.n.; 5 km from Shaw Park on

Fish River road (-DB), Welts & Dyer 2888; Bathurst (-DB),

Comins 2850.

This species which, in some respects, is unique

among the South African Crinums and, as far as can

be judged, among the tropical species as well, is

restricted in its distribution to the eastern Cape. It is

found in fairly permanent marshes near the coast from

Alexandria to East London and further inland, for

Fig. 10. — Crinum campanula-

tum, in a vlei in the Ba-

thurst District, eastern

Cape Province. Photo by

Dr. R. A. Dyer.

Plate 5. — Crinum rautanenianum Schinz. From Flowering Plants of Africa, Vol. 42, PI. 1643, (1972).

I. C. VERDOORN

35

instance north of Grahamstown, in vleis or pans which

are flooded for short periods only. The bulbs survive

the long dry season in hard-baked soil.

The features in which this species is unique are the

equally disposed stamens, that is they are not declinate,

combined with broad perianth-segments that form a

cup not a funnel. This combination would exclude it

from the three subgenera described but, probably on

account of the curved perianth-tube, Baker, the author

of the subgenus places it in Codonocrinum where

together with the curved tube the stamens are declinate

and the perianth funnel-shaped.

7. Crinum paludosum Verdoorn in Flow. PI. Afr.

39, 1. 1 523 (1968). Type: Natal, Ingwavuma, Opon-

dwini stream, Codd & Verdoorn 10297 (PRE, holo.).

C. forbesii sensu Van der Walt in Agroplantae 2, 1 : 8 etc.

(1970).

Bulb subglobose, 5-20 cm diam., narrowing more

or less gradually into a long neck. Leaves lorate, light

green with a slight metallic sheen, flexuous, arcuate,

somewhat undulate, the median 3-5 cm broad, new

leaves in the centre few, margin with a very narrow

cartilaginous border and obscurely to clearly ciliate

with short cartilaginous hairs. Peduncle usually 20-55

cm long, 1-2 cm diam. Spathe-valves 6-8 cm long,

2-2,5 cm broad at the base, soon withering and re-

flexing. Umbel 5-1 1-flowered. Pedicels 0-10 mm some-

times longer in fruiting specimen. Perianth with the

cylindric tube 9-12 cm long; segments from pure

white to delicate pink, drying to a deeper pink, some-

times keeled or striped with pale, to deep pink, in

open flower suberect, slightly spreading, apical por-

tion not recurved or only slightly so, 8-10 cm long,

I, 3-3 cm broad in the middle, apical peak short,

about 3 mm long. Stamens declinate, white turning

pink in the upper half. Style crimson in the upper half,

stigma 3-lobed, about 2 mm diam. Fruit not beaked,

about 3,5 cm diam; seeds reticulate. Plate 5; Fig.

II, 12.

Grows in pans, vleis or periodic streams. Recorded

from South West Africa eastwards through Botswana

to the western and central Transvaal and Zululand.

(Possibly in Swaziland too but not yet confirmed, see

Compton 2974).

S.W.A. — 2417 (Mariental): Sandhof (-CA), Giess 9109.

Transvaal. — 2427 (Thabazimbi): Vaalpenskraal, about 65

km N.W. of Thabazimbi (-AC), Verdoorn 2494; 2496. 2428

(Nylstroom): almost 5 km N.E. of Naboomspruit (-BD), Ver-

doorn 2469; Dyer 5853; 5854; about 9 km S.E. of Naboomspruit

(-DB), Codd & Verdoorn 10380; between Crecy and Naboom-

spruit (-DB), Meeuse 9445. 2528 (Pretoria): 14 km W. of

Pienaarsriver Station on road to Makapanstad (-CA), Tolken

1275.

Natal. — 2632 (Bella Vista): Ingwavuma-Ndumu Road

(-CC), Strey 3766; Vaze Swamp (-DC), De Winter & Vahr-

meyer 8595. 2731 (Louwsburg): E. of Pongola (-BC), McNeil

100; 101 ; 4,5 km E. of Pongola Poort (-BC), Dyer & Verdoorn

5840.

This species, which makes a striking show in pans,

marshes and along watercourses, was recently inves-

tigated in Zululand. When the diagnostic features

were established it was found to have a much wider

distribution than at first thought, occurring at intervals

across the Continent from east to west, following the

main rivers. It proved to be conspecific with specimens

collected in Botswana over 40 years ago (see photo)

which I had named from description as “probably

C. rautanenianum Schinz”. An examination of the

type of the latter species showed it to be distinct. For

the features by which to distinguish it see under that

species. It seemed possible, too, that because the flowers

are often of a “delicate pink colour”, and because it is

recorded by Exell in Kirkia 1 : 333 (1941) that Forbes

dug up “Amaryllidaceae” bulbs along a branch of the

Tembe river which runs parallel with the Pongola,

that it could have been the plant which Lindley

described (see Trans. Hort. Soc. 6, 1: 87, 1825) and

called Amaryllis forbesii, but other considerations

ruled this out and the species was finally described

as new.

C. paludosum is characterised by the gracefully

arching leaves, described by a collector as “like young

mealies”, the suberect, predominantly white flowers

which are suffused to a greater or lesser degree by pale

pink, and the conspicuous crimson stigma.

Fig. 1 1. — Crinum paludosum, in

a pan near Lake Makari-

kari, Botswana. Photo by

Herbert Lang.

36

THE GENUS CRINUM IN SOUTHERN AFRICA

Fig. 12. — Bulbs of Crinum paludosum from Mosdene, near

Naboomspruit.

Mention is made of this species in an article by J.

Freidberg in Landbouweekblad, 4th May, 1971. Here

the value of the species as a cattle feed, in the dry S.W.

African country, is reported. The leaves are said to be

browsed for several months after flowering.

8. Crinum rautanenianum Schinz in Bull. Herb.

Boiss. 4, App. Ill: 48 (1896); Bak. in FI. Trop. Afr. 7:

402 (1898); Solch in Prodr. FI. S.W. Afr. 150: 8 (1969);

Verdoorn in Flow. PI. Afr. 42: 1. 1 643 (1972). Type:

Olukonda, Schinz 822 (Z, holo.; PRE, photo.).

C. rautanenianum Schinz in Dur. & Schinz, Consp. FI. Afr.

5: 250 (1895), nomen nudum; Uphof in Herbertia9: 82(1943).

Bulb about 7 cm long, 6 cm diam., narrowed into a

neck at the apex. Leaves suberect to erect, rather firm,

1 , 5-3 cm broad at the base, narrowed gradually but

perceptibly towards the apex, about 1,5 mm broad at

the middle, and becoming deeply channelled from just

above the base, the upper portion appearing terete,

strongly many nerved, margin rough with small, dis-

tant, hard teeth. Peduncle erect, fairly stout, 40-65 cm

long. Spathe valves about 6 cm long, 1 ,8 cm broad at

base, closely enveloping the base of the flowers, soon

becoming papery; bracteoles narrow linear. Umbel

1-5-flowered. Pedicels 0 or very short. Perianth white,

faintly pink keeled, buds erect becoming distinctly

nodding, open flower suberect; tube 10-13 cm long;

segments about 8-10 cm long, the inner about 3 cm

broad and the outer slightly narrower, conniving in

a broad funnel with only the apices reflexed; peak

obtuse on the inner segments up to 3 mm long on the

outer. Stamens declinate. Fruit crowned but not

beaked. Plate 6; Fig. 13, 14.

Growing in open pans and shallow water, on open

sandy flats among large trees. Recorded from areas

around Ondangwa and Olukonda which are subject

to annual flooding.

S.W. A. — 1715 (Ondangwa); near Ondangwa (-DD), De

Winter & Giess 6857; J. N. du Plessis per P. J. le Roux s. n.

cult, in Bot. Gards. No. 22591.4.70. 1716 (Enana): Olukonda

(-CC), Schinz 822 (Z).

Although this species is so common locally very few

specimens have been preserved in South African her-

baria. Besides the type material in Zurich there are

two specimens in the Windhoek Herbarium collected

by Prof. Walter Nos. 508 & 2713, and one in the herba-

rium of the Botanical Research Institute collected by

De Winter and Giess and cited above.

Fig. 13. — Crinum rautanenia-

num, in a pan at Ondan-

gua. South West Africa.

Photo by W. Giess.

Plate 6. — Crinum carolo-schmidtii Dinter. From Flowering Plants of Africa, Vol. 41, PI. 1629 (1972).

Plate l.—Crinum moorei Hook.f. From Flowering Plants of Africa, Vol. 34, PI. 1351 (1961).

I. C. VERDOORN

37

Fig. 14. — Bulb of Crinum rautanenianum originally from On-

dangwa, growing in Pretoria.

The general appearance of the flowers in the wild

is white like those of C. carolo-schimidtii, C. euchro-

phyllum and C. paludosum. The main difference is in the

characteristic leaves of C. rautanenianum which are

rather firm, suberect and so deeply channelled to ap-

pear terete towards the apex. Bulbs received from

Ondangwa and grown at this Institute flowered in

December 1971. One plant had a 3-flowered umbel

and the other was 2-flowered. The flowers opened one

at a time and were short lived. At first the small buds

were stiffly erect, then developing quickly, one by one,

they became cernuous and as the flower opened it

assumed the erect or suberect position.

9. Crinum carolo-schmidtii Dinter, Neue Pfl. S.W.

Afr. 26 (1914); Solch, Beitrage zu einer FI. S.W.Afr.,

95 (1960); Prodr. FI. S.W.Afr. 150.7 (1969) (under

doubtful species); Yerdoorn in Flow. PI. Afr. 41:

t.1629 (1973). Type: South West Africa, Guntas

Dinter 2307 (B, holo.).

C. occiduate R. A. Dyer in Herbertia 15: 31 (1948); Solch in

Prodr. FI. S.W. Afr. 150: 7 (1969). Type: South West Africa,

exact locality unknown, cultivated in Pretoria, sub PRE 28308

(PRE, holo.).

Bulb 3-6 cm long, 3-5 cm diam., usually narrowed

into a long neck at the apex. Leaves 4-12 in a rosette

or somewhat 3-ranked, narrowly linear, flaccid, 20-60

cm long, 3-15 mm broad when flattened, canaliculate,

clearly nerved, margin sometimes with minute, soft,

distant teeth. Peduncle erect, 15-40 cm tall. Spathe-

valves 4,5-8 cm long, 0,5-1 ,5 cm broad at the base,

becoming papery and reflexing at least in the upper

half; bracteoles narrowly linear. Umbel 1-3-flowered

sometimes 4-flowered. Pedicels 0,3-2 cm long. Perianth

with a tube about 10 cm long; segments white, ob-

scurely pink flushed or with a pale pink keel, usually

becoming deep pink as they mature, conniving in a

fairly narrow funnel-shape as they close, with only the

apical portion recurved, about 8 cm long, the outer

about 1,6 cm broad the inner somewhat broader.

Stamens declinate, filaments white or faintly pink

towards the apices. Style deep rose towards apex,

stigma deep rose. Fruit not seen. Plate 7.

Grows in shallow water or clay in or near pans,

usually lime pans, or in flood plains of rivers. Record-

ed from the Okavango southwards to the Gobabis

district and eastwards to the Caprivi Strip.

S.W. A. — 1719 (Rundu): 16 km E. of Rundu (-DB), De

Winter 3810. 1720 (Sambio): 6 km E. of Masari Camp Exp.

Farm or road to Nyangana (-CC), De Winter & Wiss 4106.

1724 (Katima Molilo): 32 km from Katima on road to Linyanti

Killick & Leistner 3112. 1920 (Tsumkwe): Simkue, 251 km E.

of Grootfontein (-DA), Story 6112. 21 18 (Steinhausen): Farm

Okatjikurie (-BC), Tolken sub PRE 30940.

This species is distinguished from two other species

in South West Africa that grow in shallow water and

have the open flowers predominantly white, namely

C. paludosum and C. rautanenianum , by the very nar-

row, flaccid leaves. The areas of distribution differ

too. C. rautanenianum and C. carolo-schmidtii have a

restricted distribution in the northern regions of South

West Africa whereas C. paludosum is found in the

centre of the province and eastwards to Zululand.

As noted by the author of the species, K. Dinter,

these pan Crinums can be cultivated in the usual way

and need not necessarily be partly submerged as in

nature, where, at least for part of the year, they grow

in masses in shallow still water.

From a third shallow-water species, C. euchrophyl-

lum, recently recorded from the eastern Caprivi, C.

carolo-schmidtii differs in the very narrow, green,

rosulate leaves, as opposed to the distichous and glau-

cous leaves of C. euclirophyllum.

10. Crinum moorei Hook, f in Bot. Mag. t.6 1 13

(1874); Bak. Handb. Amaryll. 93 (1888); FI. Cap. 6:

200(1896); Medley Wood, Natal Plants 6: t. 504 (1912);

Verdoorn in Flow. PI. Afr. 34, 1. 1 35 1 (1961). Type:

Cult. Glasnevin, Eire, originally from S. Africa, exact

locality unknown.

C. imbrication Bak. in Gard. Chron. 1881, 16: 760 (1881)!

Handb. Amaryll. 92 (1888); FI. Cap. 6: 200 (1896). Type’

Eastern Cape, Butterworth Forest, MacOwan 2027. C. maco-

wanii Bak. in Bot. Mag. t.638 1 (1878) partly, as to figure only*

Bulb subglobose, varying in size, up to 19 cm diam.,

narrowing abruptly into a long stem-like neck, 30-120

cm high, multiplying prolifically from the base. Leaves

from the top of the false stem, ensiform, narrowing

towards apex and base, about 65-150 cm long and

6-12 cm broad, parallel veins obvious and the midrib

thickened on the lower surface, margins slightly un-

dulate, smooth. Peduncle about 70 cm long or longer,

somewhat flattened 1,5-3 cm broad. Spathe-valves

5-12 cm long, 1,5-3 cm broad at base, narrowing

towards apex; bracteoles narrowly linear. Umbel 5-10-

flowered. Pedicels 1-8 cm long. Perianth with a long

greenish cylindric tube, widening slightly at the apex,

8-10 cm long, erect at first, becoming cernuous as the

flower matures; segments white, usually suffused with

pink (but without a deep red keel) subspreading or

forming a rather wide funnel, oblong-elliptic, about

8-10 cm long, 2,4-4 cm broad, the inner slightly

broader than the outer, peaked apex green, 3-6 mm

long. Filaments declinate, tinged pink or red in the

upper portion. Style red towards the apex. Fruit not

beaked. Plate 8.

38

THE GENUS CRINUM IN SOUTHERN AFRICA

Grows in damp or marshy places and along rivers

and rivulets. Recorded from the coastal and semi-

coastal areas of the eastern Cape from Butterworth

district northwards into Natal and as far as the Ngo-

me Forest in Zululand.

Natal. — 2731 (Louwsberg): Ngome Forest Reserve (-CD),

Justin in F.D. Herb. 6663; Ngome Forest (-CD), Marasas 682;

Strey 9383; Codd 9598. 2831 (Nkandla): I. Garland's farm

(-DD), Venter 2770. 2931 (Pietermaritzburg): Krantzkloof

(-DD), Moll 4032. 3030 (Port Shepstone): Umzimkulu, Levey

in Herb. Galpin 3200.

Cape. — 3129 (Port St. Johns): Between Ngqeleni and Umtata

River Mouth (-CC), Lewis 67508; near Umgaza River (-CB),

Codd 9745; 16 km N. of Umzimvubu River, Theron 1554. 3228

(Butterworth): Butterworth (-AC), Van der Wa! s.n.; Kentani

(-AD )Pegler 336; The Haven (-BB), Gordon-Gray 1319 (NH);

banks of Qora River (-BC), Van Breda 884.

C. moorei differs from all other South African Cri-

nums in several respects. For instance, the neck of the

bulb, made up of the thickened and hardened leaf-

bases, forms a false stem, leafy only at the apex;

the leaves mature in one year, that is they do not die

back and grow out again the following year with a

truncated apex; the segments of the flowers do not

form the usual narrow funnel but are more spreading

when open and they are not keeled dorsally with a

deeper coloured band.

C. moorei was one of the first species of Crinum to

be cultivated in Europe mostly under the horticultural

names of C. colenso , C. nata/ense and C. mackenii.

It is closely related to C. jagus (=C. giganteum Andr.)

of tropical Africa and in some ways to the type species

C. americanum as far as one can tell from the descrip-

tion and plate which do not include the habit of that

species. Baker places C. americanum in the Platyaster

group with perianth erect and stamens spreading, and

C. moorei in the Codonocrinum group which is charac-

terised by the perianth being funnel-shaped and sta-

mens declinate. The stamens are certainly declinate

but the perianth hardly funnel-shaped.

The type specimen of our species was grown by Dr.

Moore at Glasnevin, Dublin, from seed collected by

a Mr. Webb when serving in the army in South Africa

in the 1860’s. When Dr. Hooker described the species

in the Botanical Magazine he called it after Dr. Moore

who had communicated it to him.

The flowers are white or pink and so it was con-

cluded that C. imbricatum Bak. is merely a colour

form of this species.

11. Crinum kirkii Bak. in Bot. Mag. 36: t.65 1 2

(1880); Handb. Amaryll. 91 (1888); FI. Trop. Afr. 7:

402 (1898); Uphof in Herbertia 9: 79 (1942). Type:

Zanzibar, Kirk s.n.

C. ornatum Bury, Hexand. Plate 18 (1834), partly, as to

specimen figured, excluding syn. Amaryllis ornata Ait.

Bulb varying in size, about 15 cm diam., narrowing

into a neck up to about 1 5 cm long. Leaves in a rosette

at ground level, erect at first becoming flaccid and

somewhat undulate, reflexed from about the middle,

60 cm long or longer, 5-10 cm broad, long acuminate

with a longitudinal dorsal thickening resembling a

midrib; margins sparsely ciliate with cartilaginous

hairs. Peduncle lateral, erect, about 60 cm long, up to

3 cm broad, 1,5 cm thick, with a thin bloom and

reddish at the base. Spathe valves 6-10 cm long, 3,5-5

cm broad, acuminate only in the upper half, overlap-

ping at the base and tightly enveloping the bases of

the flowers usually with only the upper portion reflex-

ing; bracteoles narrowly linear. Umbel 10-1 5-flowered.

Pedicels 0 or very short. Perianth with the tube about

10 cm long, nodding; segments white with a distinct

maroon band on the keel, about 9 cm long, 3-4 cm

broad in the centre, connivent in a trumpet of bell

shape with only the upper portion spreading and

recurved, apical peak aristate, 5 mm or more long.

Stamens declinate, filaments red in upper portion,

anthers black. Style red towards the apex. Fruit not

seen. Fig. 15, 16.

Grows on banks of rivers or edges of marshes.

Recorded from the Caprivi Strip in South-West Africa

on Mpilila Island. Also in Zambia and northwards to

Kenya.

Fig. 15. — Crinum kirkii showing the spathe valves tightly en-

veloping the bases of the flowers, and the long acuminate

leaves. Originally from the Caprivi Strip (Vahrmeijer 2204).

S.W.A. — 1725 (Fivingstone): Caprivi Strip, Mpilila Island,

Killick & Leistner 3384: Schuckmansburg, Vahrmeijer 2204.

Zambia. — Barotseland, Nangweshi, Codd 7141.

My concept of this species stems from bulbs and

specimens collected by Dr. Codd on the banks of the

Zambesi at Nangweshi, Zambia. When the plants

flowered at Pretoria they were studied in detail and

certain characteristics noted. For example the spathe-

valves remained erect longer than in other related

species known, overlapped each other and tightly

enveloped the bases of the flowers; the flowers in this

instance did not open widely at maturity and the seg-

ments are clearly banded with maroon; the peaks to

the segments were long and awn-like and the leaves

had an indication of a midrib. It was found that these

plants matched the one figured in the Botanical Maga-

zine t.65 12, the type of C. kirkii, and also specimens

collected in Kenya by Colonel Bayliss under that name.

■

I. C. VERDOORN

39

Fig. 16. — A close-up of the

umbel of Crinum kirkii

showing the distinct band

along the keel of the seg-

ments ( Vahrmeijer 2204).

It was found too that they match the specimen figured

by Mrs. Bury (1836) named C. ornatum (Ait.) Bury.

Baker, in his Handbook of Amaryllidaceae (page 90)

and in the Flora of Tropical Africa, cited Mrs. Bury’s

plate under C. sanderianum Bak. This raises the ques-

tion as to whether C. kirkii and C. sanderianum are

conspecific. An authority on tropical African species

of Crinum will probably settle this question in time.

C. kirkii shares with C. moorei the feature of the

dorsal, longitudinal thickening of the leaf which sug-

gests a midrib but in most other respects it does not

approach that species. There is no false stem to C.

kirkii , the segments of the flower do not open as

widely as C. moorei and they have a decided band of

colour on the keel.

A specimen collected in the Caprivi Strip, Vahr-

meyer 2204, flowered in Pretoria on 21st February,

1972. In the morning the segments were fairly wide

open showing the black anthers, but at midday they

closed at the throat with upper portions only reflexed.

12. Crinum acaule Bak. in FI. Cap. 6: 532 (1897);

Uphof in Herbertia 9: 82 (1942). Type: Zululand,

Sambaans territory, Saunders s.n. (K, holo.).

Bulb from about 7 to 12 cm long, 4-8 cm diam.,

produced into a neck about 4-9 cm long. Leaves linear,

usually deeply troughed, firm, 20-50 cm long or longer,

4-14 mm broad, margin ciliate with minute teeth.

Peduncle mostly underground, about 5 cm long.

Spathe-valves 9-13 cm long, 0,8-1 ,2 cm broad at the

base, narrowing gradually towards the obtuse apex,

more or less erect, the basal part encasing the basal por-

tion of the flower; bracteoles narrowly linear, 5-8 cm

long. Umbel 1-2-flowered, rarely 3-flowered. Pedicels 0

or very short. Perianth funnel shaped, with the tube

about 10 cm long; segments white with a pink flush,

usually keeled with deep pink, 10-12 cm long, the inner

about 2,5 cm broad, the outer slightly narrower, con-

niving to form a funnel with only the upper third

spreading to slightly recurved; apical peak on the inner

segments very short, about 1 mm long, on outer up to

5 mm long. Fruit beaked. Fig. 17, 18.

Fig. 17. — Crinum acaule, on the Makatini Flats, Natal, taken at

dusk. Photo by Dr. R. A. Dyer.

40 THE GENUS CRINUM IN SOUTHERN AFRICA

Found in sandy grassland. Recorded from Zululand,

from Ndumu in the north to Hlabisa district in the

south.

Natal. — 2632 (Bela Vista): Ndumu Hill (-CC), Pooley 106

(NH). 2732 (Ubombo): 19 km E. of Pongola Poort (-AC),

Dyer & Verdoorn 5842. 2832 (Mtubatuba): Palm Ridge Farm

(-AC), Harrison 337; near Makakatana Forest, Lake St. Lucia

(-AD), Codd 10263. Without precise locality, Zululand, Roux

s.n.

The description of this species, with its unusually

large flowers in comparison with the width of the leaves,

has been drawn up from the few specimens cited.

Although bulbs were planted in the gardens of the

Institute over eight years ago, they have not yet

flowered. While it can be expected that this species, like

several others, has an east to west distribution, to date

it has not been found outside Zululand. It resembles

the more widely distributed species, C. minimum ,

which also has one- to two-flowered umbels and rather

large flowers compared with the narrow leaves, but

differs mainly in the leaves being flaccid, not stiff and

toothed on the margins as in C. acaule.

The specimen figured on Plate 1253 of Curtis’s Bo-

tanical Magazine under the name C. ornatum (several

distinct elements have been figured under that name),

looks remarkably like our species having a one-

flowered umbel, a large trumpet-shaped flower and

narrow, firm leaves, but the leaves are distichous and

the peduncle longer than in C. acaule. It is reputed to

have come from Seirra Leone.

Several collectors remarked on the pleasant scent

emitted by this species, one described it as “scent like

carnations”.

13. Crinum minimum Milne-Redh. in Kew Bull.

1947: 33 (1947); Solch in FI. S.W. Afr. 150: 5 (1969);

Verdoorn in FI. PI. Afr. 40: 1. 1 577 (1969). Type:

Zambia, Mwinilunga, Cha Mwana Plain, Milne- Red-

head 2761 (K, holo. ; PRE).

C. parvibulbosum Dinter ex Overkott in Mitt. Bot. Munchen

1 : 444 (1954). Type: Karibib, Dinter 6793 (M, holo.). C. walteri

Overkott, l.c. (1954). Type: Benhop, Schwerdtfeger in herb.

Walter 1/302 (M, holo.).

Bulbs from about 5 to 8 cm long, 3-5 cm diam.,

produced into a neck from about 3 to 10 cm long,

tunics at the apex comparatively thin and papery.

Leaves about 8, flaccid, narrowly linear, about 18 cm

long, deeply canaliculate, 1,5-4 mm broad when

flattened, glabrous, margins smooth, many nerved on

dried specimens, nerves obscure in fresh leaves.

Peduncle short arising from dense papery tunics at the

apex of the bulb, 3-11 cm long. Spathe-valves soon

becoming thin and papery, 7-10 cm long, 3-9 mm

broad at the base, narrowing towards the apex. Umbel

usually 1 -flowered. Pedicel 0 or very short. Perianth

with a tube 8-12 cm long, nodding; segments white

with a deep rose dorsal keel, sometimes pale pink

tinted, 9-12 cm long, 0,8-1 ,6 cm broad, attenuate to

an acute apex, the inner slightly broader than the

outer, conniving in a funnel shape, at a stage some-

what zygomorphic, the three lower segments spreading

and the three upper ascending and somewhat recurved,

apical peak 2-5 cm long. Stamens declinate, filaments

white or pink flushed. Style declinate, deep rose. Fruit

subglobose, (not known whether beaked or not), about

2 cm diam. ; seed smooth. Plate 9.

Found in hot dry country in pebbly, sandy soil on

gentle slopes or in open patches near rivers, in Mo-

pane-Combretum veld and near Baobab trees. Re-

corded from Karibib in South-West Africa eastwards

through Botswana and the northern Transvaal to the

Kruger National Park. Also recorded from Rhodesia,

Zambia, and Tanzania.

S.W.A.— 2115 (Karibib): Karibib (-DD) Dinter 6793. 2119

(Epukiro): Farm Anderson (-CC), Tolken 1300. 2214 (Swakop-

mund): 55 km E. of Henties Bay (-BB), Hardy & de Winter

1405. 2218 (Gobabis): Farm Renette (-CC), Merxmuller 1064.

Transvaal. — 2229 (Waterpoort): near Zoutpan (-CC),

Schlieben 478. 2230 (Messina): Messina (-AC), Pole Evans

1777. 2231 (Pafuri): Between Punda Milia and Pafuri (-CA),

Mockford s.n. 2327 (Ellisras): Spruyts Kloof, Theron 2013.

2328 (Baltimore): Glen Alpine Dam (-BA), Verdoorn 2500;

Erens s.n.; Hardy s.n. 2330 (Tzaneen): Hans Merensky Nature

Reserve, Oates 03. 2531 (Komatipoort): between Skukuza

and Malelane (-BC), Van Wyk 4747.

Botswana. — 2426 (Mochudi): 48 km N. of Sikwana (-AB),

Reyneke 435; 9 km N. of Sikwana (-CB), Reyneke 450.

This species is characterised by the usually 1 -flowered

umbel, the large funnel-shaped flowers on short

peduncles and the few rather short, very narrow,

flaccid leaves. It is distinguished from C. acaule of

Zululand mainly by the somewhat smaller flowers and

the narrower flaccid leaves with smooth margins. The

distribution extends from near the west coast in S.W.

Africa eastwards through hot, dry country to the east-

ern Transvaal. Our specimens agree well with herba-

rium specimens collected in Rhodesia, Zambia and

Tanzania. These include the type specimen of C.

minimum from Mwinilunga, Zambia, and consequently

bear that specific name.

Plate 9. — Crinum lineare L.f. From Flowering Plants of Africa, Vol. 37, PI. 1471 (1966).

Plate 10. — -Crinum variabile (Jacq.) Herb. From Flowering Plants of Africa, Vol. 36, PI. 1433 (1964).

I. C. VERDOORN

41

14. Crinum linear eL.f., Suppl. 195 (1781); Thunb.‘

FI. Cap. ed. Schult. 301 (1823); Bak. Handb. Amaryll.

92 (1888); FI. Cap. 6: 199 (1896); Uphof in Herbertia

9: 81 (1942); Verdoorn in Flow. PI. Afr. 37: 1. 1 47 1

(1966.) Type: Cape, between Van Stadens and Swart-

kops Rivers, Thunberg s.n. (UPS, holo.)

C. revolutiini (L'Herit.) Herb., in Bot. Mag. sub. t.2121 (1820)

in obs. ; Amaryll. 267 (1837).

Amaryllis revoluta L’Herit., Sert., Angl. 14 (1788); Gawl. in

Bot. Mag. t.915 (1806). Type: Originally from “Prom, bonae

Spei”. — var. gracilior Ker-Gawl. in Bot. Reg. t.623 (1822).

Bulbs 7-14 cm long, 6-9 cm broad, produced into a

neck sometimes up to 14 cm long, new bulbs form

readily from the base. Leaves grass green, linear,

canaliculate, rather firm, about 65 cm long or longer,

0,5-2, 5 cm broad near the base when flattened,

arching from the base; margins smooth or with minute

distant teeth. Peduncle 18-60 cm long, subcompressed

in the lower portion, 1-1,8 cm broad, 0,7 cm thick.

Spathe-valves 4-7 cm long, 1-2 cm broad at the base,

fairly substantial on preserved flowering specimens,

reflexed. Umbels 4-14-flowered. Pedicels from very

short to 2,5 cm long. Perianth with the tube shorter

than the segments in some plants and longer in others,

3-10 cm long, curving; segments white suffused with

pale pink and with a deep-rose dorsal keel, fading

pink, the outer 1,4-2 cm broad, the inner slightly

broader, conniving to form a funnel with the upper

third recurved; apical peaks 4-5 mm long on the outer

segments, shorter on the inner. Stamens declinate;

filaments white or the upper portion purplish red;

anthers black. Style purplish-red towards the apex.

Fruit with a short beak. Plate 10.

Found in sandy soil on the northern slopes of

coastal hills or in open grassveld a few kilometres from

the sea, often among Acacias. Recorded from the

vicinity of Port Elizabeth and eastwards from near

Port Alfred and also from near Waterloo Bay in the

Peddie district.

Cape. — 3325 (Port Elizabeth): Walmer (-DC) Urton s.n.;

Draaifontein (-DC), Archibald 4884 ; Witteklip (-CD), Long

1342. 3326 (Grahamstown): near Port Alfred (-DB), McNeil

19b; east of Port Alfred Archibald 5 05; near the Aerodrome,

McNeil s.n. 3327 (Peddie): N. of Waterloo Bay (-AC), Story

4862.

For many years after this species had been collected

by the Swedish Botanist, Carl Thunberg, in 1773, it

seems to have escaped the attention of plant collectors.

Only since the year 1936 have a few specimens reached

this Institute. The first few, like the type specimen,

had the unusual feature of a perianth-tube that is

shorter than the segments. But soon specimens came

in from slightly further east with long perianth-tubes.

They were obviously the same species, characterised

by the narrow, linear, rather dark green and some-

what firm leaves and the readily multiplying bulbs

forming groups of two or more plants. This is one of

several species with a limited distribution.

C. algoense Herb., described in 1837 (Amaryll. 272),

is probably a synonym of this species but the flowers

were not seen and so it must remain a doubtful species.

15. Crinum variabile ( Jacq .) Herb., Amaryll. 268,

t.44, fig. 23 (1837); Bak. in FI. Cap. 6: 199 (1896),

partly, excluding specimen from British Caffraria;

Uphof in Herbertia 9: 81 (1942), partly, excl. syn. var.

roseum; Verdoorn in Flow. PI. Afr. 36: 1. 1 433 (1964).

Type: Cult. Vienna, originally from the Cape, speci-

men figured in Jacq. Hort. Schoenbr. t.429.

Amaryllis variabilis Jacq., Hort. Schoenbr. 4:14 t.429 (1804).

Crinum crassifolium Herb., Append. 23 (1821 ), nomen nudum.

Bulb about 26 cm long, narrowing gradually to-

wards the apex (not abruptly narrowed into a neck),

5-9 cm diam. near the base, new plants sprouting

around the bulb and from within the tunics. Leaves

deep green, rather firm, varying in length and width

on the same plant, 8-35 cm long, 0,4-4 cm broad;

margins with a cartilaginous border and with minute,

distant teeth. Peduncle 30-40 cm long, slightly com-

pressed, 1-1,8 cm broad. Spathe-valves 6-7 cm long,

1,5-2 cm broad at the base, reflexing and becoming

somewhat papery. Umbel 6-12-flowered. Pedicels 2,5-

5 cm long. Perianth with the tube shorter than the

segments, 3-4 cm long, curved; segments white suf-

fused at the base with green and with rose dorsally,

turning deep pink as they fade, conniving into a funnel

shape, the upper portion spreading and recurving, 6-9

cm long, 1,5-2 cm broad; apical peak small about

2 mm long. Ovary swelling rather more obvious than

in other species, not tapering as much to base and

apex, fruit without a beak or shortly beaked; seed

smooth. Plate 1 1 ; Fig. 19, 20.

Found in beds of the rivers that flow to the Atlantic

and are often dry for long periods. Recorded from

Namaqualand from near Garies and near Nieuwoudt-

ville.

Cape. — 3017 (Hondeklip Bay): Groen River near Garies

(-DB), Hardy 513; 1709. 3119 (Calvinia): 11 km south of

Nieuwoudtville (-AC), Hall 1080.

Fig. 19. — Fruiting umbel of Crinum variabile, originally from

Garies, Namaqualand.

42

THE GENUS CRINUM IN SOUTHERN AFRICA

Fig. 20. — Lower portion of a Crinum variabile plant to show

new plants sprouting around the bulb and from within

tunics.

This species is another with a restricted distribution.

To date it has not been found outside Namaqualand.

For years it has been confused with another species

with a short perianth tube. Herbert was the first to

distinguish between the two, noting that in the one the

leaves are dark green and rather stiff and the flowers

fade dark pink whereas in the other the leaves are a

lighter green, rather flaccid and the flowers fade

brownish (Amaryll. 268). This was investigated and

confirmed and resulted in the description of Crinum

macowanii subsp. confusum Verdoorn. It has been

found since, see under C. macowanii, that it is merely

a local form of C. macowanii.

C. variabile resembles C. bulbispermum in the bulb

which is not abruptly narrowed and the flowers that

fade a deep pink or red, but it differs very markedly

in the shape of the flowers and the size, colour and

texture of the leaves. From its nearest relative C.

lineare it differs mainly in the broader leaves and the

distribution. It resembles C. lineare in the colour and

texture of the leaves and the way young plants develop

prolifically around the parent bulb. Milne-Redhead

& Schweickerdt (1939) state that one inflorescence on

the Thunberg specimen named by him C. longifolium

is C. variabile (the others are Amaryllis belladonna).

Since it is an inflorescence only, there is the possibility

that it may be one of the other two short-tubed Cri-

nums, C. lineare or C. macowanii. Thunberg could

have collected these, but probably not C. variabile.

16. Crinum foetidum Verdoorn in Bothalia 10:

56 (1969). Type: 112 km N. of Vaalwater, Louw 3460

(PRE, holo.).

Bulb more or less globose, 15-18 cm diam., abruptly

narrowed into a neck of varying lengths. Leaves

about 15, dull blue-green, about 70 cm long, 12 cm

broad, distinctly ciliate with cartilaginous hairs which

are mostly over 1 cm long. Peduncle suberect, short,

14-25 cm long, 2,5 cm broad, 1,5 cm thick. Spathe-

valves 8-11 cm long, 2-3 cm broad at the base, nar-

rowing towards the apex, soon reflexing. Umbel 7—1 1-

flowered. Pedicels 0-1 ,5 cm long. Perianth with a tube

about 10 cm long; segments conniving in a funnel-

shape with the apical third recurved, white with a pale-

pink dorsal keel, about 10 cm long, outer about 2 cm

broad, inner 2,5 cm broad, all acuminate in the upper

third, apical peak often deep rose, 2-8 mm long

sometimes slightly longer. Stamens declinate filaments

white or tinged with rose towards the apices. Style

red in the upper portion. Capsule subglobose, beaked

at first eventually shortly so; seeds sepia coloured

rough with raised papillose ridges (lamellate?). Fig.

21, 22.

Plate 11. — Crinum graminicola Verdoorn. From Flowering Plants of Africa, Vol. 29, PI. 1155 (1953).

Plate 12. — Crinum delagoense Verdoorn. From Flowering Plants of Africa, Vol. 35, PI. 1389 (1962).

I. C. VERDOORN

43

Fig. 22. — Fruiting specimen of

Crinum foetidum showing

the long beaks to the cap-

sules, but these shorten

later to some extent.

Found growing in deep sand in open woodland in

the drainage basin of the upper reaches of the Lim-

popo and Okavango Rivers. Also recorded from the

Victoria Falls, Ngamiland and Tanzania.

S.W.A. — 1719 (Rundu): about 12 km E. of Rundu (-DC),

De Winter 3777. 1720 (Sambio): about 64 km W. of Rundu

(-CD), Le Roux s.n. 2118 (Steinhausen): farm Sturmfeld (-DB),

Tolken 1003.

Transvaal. — 2328 (Baltimore): 132 km N. of Vaalwater,

Louw 3460; Steilloop (-BC), Verdoorn 24 8; 2499. 2427 (Thaba-

zimbi): near Vaalpenskraal (-AC), Verdoorn 2495; west of

Vaalwater, near Sterkstroom (-BD), Verdoorn 2497.

This species is characterised by the broad dull,

blue-green, distinctly ciliate leaves. It was found

when these plants were examined in the veld, that the

leaves tore readily and a watery sap dripped from

them. The flowers were strikingly white with a pale

rose keel to the segments. The funnel was narrow,

and the segments were perceptibly acuminate in the

upper third. The subglobose fruits, bulging with the

large rounded seeds, were beaked by the remains of the

long perianth tube. A feature not observed before

in any other species was the dark colour and rough-

ness of the seed coats. Parts of the plant that were

injured or torn turned blackish, and emitted a foetid

odour. This character may not be constant as it

was not observed in cultivated plants the following

season.

The specimens cited came from open woodland

near the tributaries of the Limpopo in the north-

western Transvaal and from near the Okavango River

in South-West Africa. Baines’s painting of a Crinum

growing in Namaqualand. now housed in the Kew

Herbarium, and mentioned on page 220 of his “Ex-

ploration of S.W. Africa,” appears to be this species.

It is obvious, too, that the plant of which a photograph

appears in “Guide to the Victoria Falls” by H. Wild,

under the name C. z eylanicum L., is our species.

Because not one of the published descriptions was

found to fit this species exactly, it was described as

new.

17. Crinum graminicola Verdoorn in Flow. PI. Afr.

29: 1. 1 1 55 (1953), spalm graminicolum; African Wild

Life 17: 153 (1963). Type: Moot Halt, about 57 km

west of Pretoria, Verdoorn 2408b (PRE, holo.).

Bulb more or less globose, 12-14 cm diam., abruptly

narrowed into a neck 4-10 cm long. Leaves few,

willow-green, broad, spreading at ground level, up to

14 cm broad, obviously ciliate with cartilaginous hairs.

Peduncle arising laterally, arcuate ascending, up to 40

cm long, about 2,5 cm broad and 1,5 cm thick.

Spathe-valves 7-1 1 cm long, up to 3 cm broad at the

base, narrowing gradually towards the apex, not early

withering and reflexing; bracteoles narrowly linear,

up to 10 cm long. Umbels several to many flowered,

8-30 recorded. Pedicels subsessile or 0,5-2, 5 cm long.

Perianth with a curved tube, 7-11 cm long; segments

broadly keeled with deep rose or almost entirely that

colour, fading towards the margins, conniving into a

funnel shape with only the apical portion recurved,

about 9 cm long, the outer 1 ,5 cm broad just above

the middle, inner 2,5 cm broad above the middle,

apical peak 3-7 mm long, the reflexed true apex often

papillate. Stamens declinate, filaments white or suf-

fused with deep rose towards the apex. Style deep rose

in the upturned apical portion. Capsule oblong to

oblong-globose, bulging with seeds, beaked with the

basal portion of the perianth tube. Seeds subglobose

of irregular shape, about 2 cm diam. Plate 12; Fig.

23.

Grows in open grass country usually in sandy soil.

Recorded from central and northern Transvaal and as

far south as the Dundee district in Natal.

Transvaal. — 2229 (Waterpoort): About 6 km along Blue-

gums Poort Road (-DD), Leach 11339. 2230 (Messina): Origi-

nally from Rusfontein, R. Schlieben's farm (-DD), Schlieben &

Strey 8360. 2329 (Pietersburg): Between Duivelskloof and

Louis Trichardt (-BD), Werdermann & Oberdieck 1935; 28 km

E. of Pietersburg (-DC), Van Vuuren 1369; Boyne (-DD),

Thompson s.n. 2526 (Zeerust): Zeerust (-CA), Thode A 1506.

2527 (Rustenburg): Rustenburg (-CB), Collins 57; Horn’s Nek

(-DB), Dyer & Erens 3513; Moot Halt, 57 km W. of Pretoria

(-DC), Verdoorn 2408a,b,c, & d. 2528 (Pretoria): Wonderboom

aerodrome (-CA), Robertson 9; 10; Villieria (-CA), Schwei-

ckerdt 1726; Rietondale (-CA), Pole Evans s.n.; Pretoria West

(-CA), Kresfelder s.n.; University farm, Pretoria (-CA), Ver-

doorn 2344; Codd 3699; 4184. 2529 (Witbank): about 3 km W.

of Middelburg (-CD), Reynolds 3349; Middelburg townlands

(-CD), Van der Merwe s.n.; about 6 km S. of Middelburg (-CD).

Verdoorn 2422; 17 km E. of Middelburg (-CD), Bruce 467.

2626 (Klerksdorp) : Grasfontein (-AA), Sutton 340; Uitgevon-

44

THE GENUS CR1NUM IN SOUTHERN AFRICA

Fig. 23. — Fruiting specimen of

Crinum graminicola, near

Pretoria, showing the

beaked fruits. Photo by

Dr. L. E. Codd.

den, Lichtenburg, Liebenberg 19. 2627 (Potchefstroom): Farm

Somerville, 64 km W. of Krugersdorp (-AB), Codd 2127; Cl

km W. of Tarlton (-AC), Phillips 21; Jackson’s Drift (-BD?)

Gilliland in Moss Herb 26268; Randfontein (-BA) Gunn 1;

Krugersdorp (-BA or BB) Webster s.n. 2628 (Johannesburg):

near Bapsfontein (-AB), Dyer & Verdoorn 5852; Kaalfontein

(-AB), Pole Evans in Govt. Herb. 16825: Mogg s.n. ; 5 km W. of

Kendal (-BB), Reynolds 3347; Delmas (-BD), Van der Plank

s.n. 2630 (Carolina): Piet Retief, Galpin 10894; near Idalia

(-DC), Sidey 2319.

Natal. — 2729 (Volksrust): 8 km N. of Newcastle (-DB),

Reynolds 4055. 2830 (Dundee): near Rorkes Drift (-AD),

Codd 2241.

In the broad, distinctly ciliate leaves which spread

at ground level, and the shape and colour of the flo-

wers this species resembles C. delagoense which is

described next, No. 18. It differs principally in the

shorter peduncle and pedicels and the fruits which are

oblong-globose and long beaked at the apex. The

fruits of C. delagoense are subglobose, without a beak,

and more brightly coloured at maturity. The areas of

distribution differ too. C. graminicola occurs in open

grassveld at medium to high altitudes of the Transvaal

and bordering areas of Natal while C. delagoense is

found in the lowveld and in the basins of the Limpopo

and Zambesi Rivers with a concentration of plants at

the coast near the river mouths.

18. Crinum delagoense Verdoorn in Flow. PI. Afr.

35: 1. 1 389 (1962). Type: Inhaca Island, A/oggsub PRE

29010 (PRE, holo.).

Amaryllis forbesi var. purpurea Lindl. in Trans. Hort. Soc.

Lond. 6, 1: 285 (1826). Type: No specimen preserved but presu-

mably this species.

Crinum forbesianum var. punica Herb., Amaryll. 267 (1837).

Type: no specimen preserved but presumably this species. C.

forbesianum sensu Bak. in Bot. Mag. t.6545 (1881); FI. Cap. 6:

201 (1896), pro majore parte, excl. syn.

Bulb large, 11-22 cm diam., abruptly narrowed in

a neck of varying lengths. Leaves grass green, spread-

ing at ground level, comparatively few, broad, 1 1-21

cm broad, distinctly ciliate. Peduncle arcuate ascending

40-50 cm long, up to 3 cm broad. Spathe-valves up to

about 8 cm long, about 3,5 cm broad at the base,

narrowing gradually towards the apex; bracteoles

narrowly linear about 8 cm long or longer. Umbel

10-30-flowered. Pedicels 2-5 cm long; longer in

fruiting specimens. Perianth with a long curved tube,

7-10 cm long; segments white with a distinct broad,

deep-rose or purplish red, dorsal keel, or the whole

flushed rose or purplish red with the keel a deeper

rose, conniving in a funnel shape with only the apical

third recurved, about 9 cm long, the inner about 2,5

cm broad just above the middle, the outer somewhat

narrower, apical peak short and obtuse on the inner

lobes, about 1,5 mm long, and longer on the outer

3-5 mm long, reflexed true apex papillate. Stamens

declinate, filaments white or suffused with rose. Style

deep rose in the upper upturned portion. Fruit sub-

globose, 5-7 cm diam., purplish red or scarlet when

mature, not beaked, the perianth dying back to a short

crown at the base. Plate 13; Fig. 24.

Found in deep sand in the Transvaal lowveld and in

the basins of the Limpopo and Zambesi rivers with a

concentration of plants at the coast from Zululand

northwards to Mozambique as far as the Zambesi

mouth and probably beyond. Recorded near tribu-

taries of the Limpopo in the western Transvaal and

eastwards to the lowveld of the Transvaal and Zulu-

land.

Transvaal.— 2327 (Ellisras): Rietspruit (-DC), Louw 4210.

2328 (Baltimore): Tambotiekloof (-CC), Theron s.n., cult. PRE ,

G.N. 22207. 2331 (Phalaborwa): Madeira Farm (-DC), Me

Neil s.n. 2430 (Pilgrim's Rest): 25 km, S. of Olifants River

(-AC), Reynolds 5800. 2431 (Acornhoek): Tshokwane (-DD)

Codd 5725. 2531 (Komatipoort): Skukuza (-AB), Van der

Scliijff 3462; 3461 ; Codd 5091; between Skukuza and Lower

Sabie (-BA or BB) Van der Scliijff 2879; 8 km S. of Skukuza

(-BC), Van der Scliijff 2145; Komatipoort (-BD), Rogers 22213;

Tonetti SidingCB), Leach 12673.

Natal. — 2632 (Bella Vista): Mabibi (-DC), Edwards 2592;

24 km from Ingwavuma/Ndumu (-DC), Moll 4127. 2732

(Ubombo): 17 km E. of Pongola Poort, S. of Pongola River

(-AC), Dyer & Verdoorn 5841. 2832 (Matubatuba): 19 km S.

of Lake Bengazi (-AB), Codd 10252; near St. Lucia (-AB),

Codd 10265; Dukuduku farm (-AC), Strey 6132.

This species is characterised by the few broad, ciliate

leaves spreading at ground level. One collector de-

scribed the plants as “Welwitschia-like, sprawling on

the sand.” The bulb is large and the umbel many-

flowered, each flower striped like a barber’s pole or

Plate 13. — Crinum lugardiae N.E.Br. From Flowering Plants of Africa, Vol. 14, PI. 532 (1934) under the name Crinum crispum.

I. C. VERDOORN

45

Fig. 24. — Bulb and fruiting umbel of Crinum delagoense. Note

the perianth dies back to a short crown at the apex of the capsule.

suffused with purplish red. The fruits too are brightly

coloured, and Mrs. Monteiro (Verdoorn, l.c., 1962) de-

scribes them as: “almost as showy as the flowers,

each one growing to the size of a large apple and when

ripe turning a bright crimson.”

In the past C. delagoense has been very generally

known as Crinum forbesianum Herb, and it is under

that name that it appears in the Botanical Magazine

(Baker, l.c., 1881) and in the Flora Capensis (Bak, l.c.,

1896). Crinum forbesianum is an illegitimate name

based on Amaryllis forbesii Lindl. As pointed out by

Verdoorn, l.c., Lindley’s original description does not

fit this species and the conclusion has been reached

that, since it is impossible to establish what it describes,

the name must be discarded as confused. To date no

type specimen has been traced and the diagnostic

characters taken together are not applicable to any

known species of Crinum in that area. The large bulbs

and many-flowered umbels (Lindley, l.c.) characterise

two species in the vicinity, C. delagoense and C.

macowanii , but neither of these has “flowers of a most

delicate pink colour” nor leaves that are narrow,

flaccid and glacous” (Lindley, l.c.). The species C.

paludosum Verdoorn has flowers white suffused with

pale pink and comparatively narrow leaves and Forbes

could have collected this species along the Tembe

River, but the bulbs are not large and the umbels are

not many flowered.

19. Crinum lugardiae N.E. Br. in Gard. Chron.

2: 49 (25th July, 1903); Kew Bull. 1909: 142 (1909);

Uphof in Herbertia 9: 82 (1942); Solch in Prodr. FI.

S.W. Afr. 150: 7 (1969). Type: Kwebe Hills, Ngami-

land, Lugard 43 (K, holo.).

C. polyphyllum Bak. in Bull. Herb. Boiss 2nd series, 3: 667

(31st July, 1903); Uphof in Herbertia 9: 81 (1942); Solch in

Prodr. FI. S.W. Afr. 150: 7 (1969). Type: Hereroland, east of

Windhoek, Dimer 826. C. crispum Phillips in FI. PI. Afr. 14:

t.532 (1934); Uphof in Herbertia 9: 82 (1942). Type: Roode-

plaat, near Pienaars River, Letty 17 sub PRE 15877 (PRE, holo.).

Bulb 6-9 cm long, 4-7 cm broad, narrowed into a

neck 4-12 cm long. Leaves many, the majority narrow,

innermost 0,2 cm broad, the outermost up to 2,5 cm

broad and in certain circumstances undulate and

spreading at ground level but usually suberect and

canaliculate, margin ciliate. Peduncle 5-30 cm long,

0,8-2, 5 cm broad and up to 9 mm thick. Umbel 4-10-

flowered, often 8-flowered. Spathe-valves 4,5-11 cm

long, 1-2,5 cm broad at the base, tapering to the apex,

some becoming papery; bracteoles narrowly linear.

Pedicels from very short up to 2 cm long. Perianth

with the tube 8-13 cm long, from slightly to distinctly

curved; segments white with a distinct deep rose keel,

8-9 cm long, 1 ,5 to 2 cm broad conniving in a funnel

with the upper third curved and revolute; apical peaks

about 1 mm long on the broadly rounded inner seg-

ments, about 2 mm long on the slightly narrower

outer segments. Stamens declinate, filaments white,

anthers black. Style rose coloured. Fruits subglobose

about 3 cm diam., distinctly beaked; seed globose

about 1 cm diam. Plate 14; Fig. 25.

5292-4

Fig, 25. — Crinum lugardiae,

near Gaberone, Botswana.

Photo by Dr. L. E. Codd.

46

THE GENUS CRINUM IN SOUTHERN AFRICA

Found in grassy depressions or black turf flats

usually in thorn veld. Recorded from South-West

Africa, Botswana, central Transvaal and Griqualand

West.

S.W.A. — 2217 (Windhoek): Otjiserwa (-BD), Kinges 790 >'

Otjiveru, 24 km W. of Omitara (-BD), Tolken s.n.; 11 km W-

of Omitara (-BD), Codd 5822; Gochaganas (-CC), Giess 8368;

Haigamas (-CC), Giess 9597.

Transvaal. — 2427 (Thabazimbi): Makoppa (-AC), Codd

8650. 2428 (Nylstroom): 17 km N.E. of Naboomspruit (-BD),

Reynolds 5810; Nylstroom (-CB), Lang sub TRV 32751; Du

Toils Kraal (-D), Galpin s.n.; Klippan (-C), Meeuse s.n.;

Mosdene (-DB), Galpin M 337 ; Grootvlei (-DB), Galpin 11678.

2429 (Zebediela): Potgietersrus (-AA), Leach s.n.; Commins

902; 14 km N.W. of Marble Hall (-CC), Codd & Verdoorn

10443. 2527 (Rustenburg): Welgevonden near Brits (-DB),

Fochema s.n. 2528 (Pretoria): 0,5 km from Pienaars River

(-AB), Kies & Bruce 35 ;1 km S. of Pienaars River (-AB), Codd

& de Winter 3441 ; Rooikop (-BA) Smuts & Gillett 3050; Rust de

Winter (-BA), Mogg 17318; Pretoria (-CA), Phillips s.n.;

Zeekoegat, Vogts s.n.

Cape. — 2625 (Delareyville): 64 km S. of Mafeking (-DA),

Godfrey & Gilliland V.H. 1825. 2824 (Kimberley): 25 km S.

of Kimberley (-DA), Leistner 2104. 2922 (Prieska): Spitzkop

Marloth 974.

In the past this species was known to us as C.

polyphyllum Bak. with C. crispum Phillips as a syno-

nym but in 1967 when, through the courtesy of the

Director of the Royal Botanic Gardens, Kew, photo-

graphs of paintings by Mrs. Lugard and Thomas

Baines of the species described as “C. lugardae ” were

seen, it became obvious that the last mentioned was

also conspecific. Being the first published name, al-

though only 6 days earlier than C. polyphyllum, it is

the correct name for the species. The auther, N. E.

Brown, published the specific epithet as “lugardae”

but according to later Rules of Nomenolature this

must now be changed to “lugardiae”.

C. lugardiae is characterised by its many narrow

leaves in a rather tight tuft. This feature, by the way,

presumably suggested the specific epithet “polyphyl-

lum”. Under some conditions, or at certain stages of

growth the outer leaves are markedly undulate (the

reason for the name “crispum”) and spread at ground

level, but usually all the leaves are suberect and

eventually arcuate; they are deeply channelled and the

outer only obscurely undulate. The flowers resemble

those of C. macowanii in that the deep rose colour is

confined to the strip down the keel while the rest of the

segment is white and the black anthers show up against

the white inner face of the segments. The fruits also

resemble those of C. macowanii being distinctly beaked

but they are smaller in size.

20. Crinum macowanii Bak. in Gard. Chron. 9:

298 (1878), as to MacOwan 2122 and part of the

description; Bot. Mag. sub. t.638 1 (1878), partly, excl.

t.638 1 ; Handb. Amaryll. 94 (1888), partly; FI. Cap.

6: 202 (1896), partly, inch MacOwan 2122 and 508;

Uphof in Herbertia 9: 80 (1942), partly, excl. Bot.

Mag. t.638 1 ; Verdoorn in J. S. Afr. Bot. 22: 79 (1956).

Lectotype: Eastern Cape, No Mans Land, MacOwan

2122 (K, lecto. ; PRE, photo).

C. gouwsii Traub in Herbertia, Plant Life: 40 (1954). Type:

Cult. California from seed from Duivelskloof, Transvaal, Traub

522 (TRA, holo.).

C. macowanii subsp. confusum Verdoorn in J. S. Afr. Bot.

32: 67 (1966). Type: Albany, Carlisle Bridge near Graham-

stown, Archibald 7513 (PRE, holo.).

Amaryllis revoluta sensu Ker-Gawl. in Bot. Mag. 1. 1 1 78

(1809).

Fig. 26. — Crinum macowanii. near Pietermaritzburg, the same

form as that which grows in the eastern Cape Province

from where the type specimen came. Photo by Dr. R. A.

Dyer.

Bulb varying in size, large in some localities, 6-25

cm diam., fairly abruptly to abruptly narrowed into

a short to long neck. Leaves variable, green or glau-

cous, spreading at ground level or arcuate from a

short stemlike base, deeply canaliculate, from slightly )

to very strongly undulate, 80 cm long, or longer, 2-16

cm broad, margins with a narrow cartilaginous border i

and sparsely to fairly densely ciliate with short to

about 1 mm long cartilaginous hairs. Peduncle arcuate-

erect or erect and arcuate at the base only, varying

in length, 1 8-90 cm long, 1 , 5-3 cm broad and 0 , 6-1 , 7

cm thick. Spathe-valves 6-10 cm long, 2-4 cm broad

just above the base, rather thick, becoming parchment-

like but not very thin, eventually partly reflexed;

bracteoles narrow linear. Umbels on the average 8-25-

flowered. Pedicels varying in length, 1-4 cm long.

Perianth with a long or short tube, the segments con-

niving in a wide trumpet shape with the apical portion

recurved; tube 3-11 cm long, cernuous; segments

about 7-1 1 cm long, white with a distinct rose coloured ,

keel, rarely, in some areas, the whole segment rosy

with a deep carmine keel; apical peak on outer seg-

ments about 5 mm long, on obtuse inner segments,

short and broad. Stamens declinate, filaments usually

white; anthers black. Style white with the apical por-

tion red. Fruit subglobose, 3-6 cm diam. beaked with

I. C. VERDOORN

47

Fig. 27. — Bulb of Crinum macowanii which can be up to, or

over, 20 cm in diameter. Originally from Greytown, Natal.

the basal remains of the perianth tube; seeds sub-

globose to depressed globose, about 2 cm diam. in

greatest width, very pale green or whitish, skin smooth

becoming reticulate. Fig. 26, 27, 28, 29, 30, 31.

Found in various types of habitat, from mountain

grassveld and stony slopes to grassveld in hard dry

shale, gravelly soil or sandy flats, along rivers and at

the coast. Recorded from all the provinces but absent

from the south-western Cape. Also occurring in neigh-

bouring tropical countries to the north of South

Africa.

S.W.A.— 1917 (Tsumeb): Nosib (-BD), Schoenfelder 932.

2116 (Okahandja): 32 km E. of Karibib on road to Okahandja

(-CD), Van Vuuren 994. 2117 (Otjosandu): Quickborn (-AA),

Bradfield 55; farm Okarupa (-CC), Giess 9594. 2416 Maltahohe :

Buellspoort (-AB), Strey (photo only).

Botswana. — 2426 (Mochude): 24 km N. of Sikwane (-AD),

Reyneke 445.

Transvaal. — 2230 (Messina): Lake Funduzi (-CD), Codd

4511 ; Rambuda Location (-DA), Van Warmelo 5121911 ; Mbay-

inbayi (as Boiandboi), 44 km S.W. of Punda Milia (-DD),

Lang in TRV 33275 A; Tshaulu, 38 km N.E. of Sibasa (-DD),

Codd 6914. 2231 (Pafuri): 9,6 km S. of Punda Milia (-CC),

Codd 6003; Klopperfontein (-DA), Van der Schijff 4091. 2328

(Baltimore): Blaauberg, (-BB), Codd & Dyer 9191 (photo only).

2329 (Pietersburg): Haenertsburg (-DD), Thompson s.n. 2330

(Tzaneen): Tzaneen, Vygeboompoort (-AB?), Van Dam sub

TRV 13624; Merensky Dam (-CA) Scheepers 842; 12,8 km

N.W. of Tzaneen (-CA), Reynolds 5803; 1,6 km S.E. of

Tzaneen (-CC), Reynolds 5804; Magoebaskloof (-CC), Mogg

10675; near Debegeni Falls, Magoebaskloof (-CD), Mauve

4318. 2427 (Thabazimbi): 8 km S. of Thabazimbi (-CB),

Fig. 28. — Crinum macowanii,

the colour form growing

near Estcourt Natal. The

segments of the flower are

suffused with rose and the

dorsal band is a deep

carmine. Photo by Dr. R.

A. Dyer.

48

THE GENUS CRINUM IN SOUTHERN AFRICA

Fig. 29. — Crinum macowanii, form with glaucous leaves growing

near Lydenburg. Note the beaked fruits on the right. Photo

by Dr. L. E. Codd.

Codd 8676. 2428 (Nylstroom): about 13 km W. of Roedtan

(-DB), Codd & Verdoorn 10379; about 11 km S. of Warmbaths

(-DD), Verdoorn 2404; 13 km S. of Warmbaths (- DO) Ver-

doorn 2405; 2406; 2406a; 2507. 2430 (Pilgrim’s Rest): Bosbou-

stasie, Mariepskop (-DB), Van der Schijff 5071; Van Dam in

TRV 26250; 22,4 km W. of Pilgrim’s Rest (-DC), Codd &

Verdoorn 7606; 7607; 28 km W. of Pilgrim's Rest (-DC), Codd

& Verdoorn 7608; 22,4 km N. of Lydenburg near Krugerspos

(-O C), Codd & Verdoorn 7609; 20,8 km N. of Lydenburg

(-DC), Codd & de Winter 3356. 2431 (Acornhoek): Tshokwane

(-DD), Van der Schijff 1358. 2524 (Zeerust): 24 km N. of

Groot-Marico (-AD), Louw s.n. 2527 (Rustenburg): Farm

Welgevonden Brits (-DB), Mogg 14504. 2528 (Pretoria): Pre-

toria, without precise locality, Reck 598; Gouws s.n.; 8 km N.

of Radium (-AB), Bruce & Kies 38; Onderstepoort (-BA),

Verdoorn 2395; 2396; s.n. sub G.N. 8990 (figured but

not published); Haakdoringboom, Onderstepoort (-BA)

Hanekom 1323, 0,8 km S. of Pyramid Station (-BA),

Codd & de Winter 3443 (photo only); N. of Onderstepoort

(-BA); Laughton s.n.; Rietondale (-CA), Verdoorn s.n.; N. of

Derdepoort (-CA), Makink s.n.; Kaalplaats, about 17 km N.W.

of Pretoria (-CA), Du Toit s.n.; Pretoria University Farm

(-CA), Verdoorn 2339; Wonderboom (-CA), Leendertz 964;

Pyramids (-CA), Mogg 16776; Doornkloof, Irene (-CD), Ver-

doorn 2391; 2392; 2394 ; 2398; 2399. 2530 (Lydenburg): 18

km S.W. of Lydenburg (-AB), Codd & Verdoorn 7594; 6 km

S. of Lydenburg (-AB), Codd & Verdoorn 7619; 7623; 17 km

S.W. of Lydenburg (-AB), Codd & de Winter 3242; 9,7 km S.W.

of Lydenburg (-AB), Codd 5643; 9 km E. of Lydenburg (-BA),

Cod'd & Verdoorn 7598; 13 km W. of Sabie (-BA), Codd & Ver-

doorn 7603; 7 km E. of Sabie (-BB), Codd 5652; 4 km S. of

Machadodorp (-CB), Codd 8067; Mariepskop, between sawmill

and waterfall (-DB), Van der Schijff 5855. 2531 (Komatipoort):

Sigaas (-AD) Van der Schijff 3219 ; across Crocodile River from

Malelane (-BC), Godfrey SV 1686; Barberton (-CC), Codd

8172; Umvoti Creek, Barberton (-CC), Galpin 667; 14 km S.E.

of Kobinga (-DD), Codd 7801. 2630 (Carolina): Carolina

(-AA), Cross s.n. sub G.N. 7683.

O.F.S.— 2826 (Brandfort): 8 km S.W. of Winburg (-DB),

Codd (photo only).

Swaziland. — 2631 (Mbabane): between Piggs Peak and

Bremersdorp (-AD), Wells 2034; below Stegi (-BD), Rodin

4570; Hlatikulu (-CD), Compton 28136.

Natal. — 2729 (Volksrust): Near Laings Nek (-DB), Mauve

4466; 17,8 km N.W. of Newcastle on Memel road (-DB),

Marais 926. 2731 (Louwsburg): Between Pongola Poort and

road bridge to Transvaal (-BC), Dyer & Verdoorn 5847. 2732

(Ubombo): Ingwavuma (-AA), Codd 7035; 17,9 km N. of

Otobotini Store (-AC), Codd & Verdoorn 10298. 2829 (Harri-

smith) : Van Reenen’s Pass, Mauve 4473; 9 , 6 km W. of Weenen

(-DD), Hardy 1910; 4 km out of Estcourt on road to Innesdale

(-DD), West 390. 2830 (Dundee): Mashunka (-CB), Edwards

939; 8,9 km from Tugela Ferry on road to Greytown (-CB),

Killick & Marais 2115; Dundee (-AA or AB), Dahlstrand 39;

Weenen (-CC), Codd 8614. 2831 (Nkandla): east of Middledrift,

bank of Tugela (-CD), Strey 4179; Hlabisa (-BB), Codd 10266;

Umlalazi Game Reserve (-DD), Ward 4324. 2929 (Underberg):

western outskirts of Estcourt (-BB), Dyer 4861 ; 4862; Willow-

grange (-BB), Plowes 2203; Griffins Hill (-BB), Mauve 4472.

2930 (Pietermaritzburg): 4 km from Greytown on road to

Maritzburg (-BA), Killick 2115A.B; between Scottsville and

Alexandra Road (-CB), Dyer 4869; Pietermaritzburg (-CD),

Goossens 134; Inanda Farm (-DB), Strey 4836. 2931 (Stanger):

Umhlali (-CA), Meeuse 9658; Isipingo (-CC), Ward 5897. 3030

(Port Shepstone): Umtwalumi-Umgeni junction (-BC), Strey

4446.

Fig. 30. — The robust, many-

flowered form of Crinum

macowanii found in Mo-

zambique and the Kruger

National Park. Photo by

Herbert Lang.

Plate 14. — Crinum bulbispermum (Burm.f.) Milne-Redh. & Schweick. From Flowering Plants of Africa, Vol. 29, PI. 1 150 (1939).

I. C. VERDOORN

49

Fig. 3 1 . — The short-tubed form

of Crinum macowanii for-

merly described as subsp.

confusion of this species.

Plant originally from near

East London.

Cape. — 3126 (Queenstown): 27,2 km N.W. of Queenstown

(-DC), Codd lc679; 25,6 km N. oi Queenstown (-DC), Mauve

s.n.; Queenstown (-DD), Galpin 2210; Theron 2146. 3128

(Umtata): 8,8 km from Umtata (-DA), Killick & Marais 2068.

3227 (Stutterheim): 8 km from Cathcart (-AC), McNeil s.n.;

15,5 km from King William’s Town to Berlin (-CD), Comins

1439; 1440; 9,6 km W. of Komgha (-DB), Codd 6349; Gonubie

Drift (-DB), Flanagan 1708; grassy hills near Komgha (-DB),

Flanagan 624. 3228 (Butterworth): Butterworth (-AC), Leach

12555; Pegler 1802; 4,8 km from Butterworth along railway

line (-AC), Dyer 4509; near Kei mouth (-CB), Flanagan 625.

3326 (Grahamstown): 6 km S. of Carlisle Bridge (-AB), Archi-

bald 7513; Dikkop Flats (-AB), Archibald 7566.

This is the most widely distributed species in south-

ern Africa reaching from the tropics southwards

through South West Africa, Botswana, O.F.S., Swa-

ziland and Natal to the north-western and eastern

Cape. Although it varies considerably having several

local forms yet it is recognisable, when living plants in

the wild have been studied, as a definable species,

characterised, among other things, by the rather open

mouth of the curved, trumpet-shaped flower with

black anthers showing up against the usually white

inner surface of the recurved segments. The leaves are

flaccid but vary in colour, width and degree of undula-

tion. Only a few new leaves develop each year and can

be seen as complete narrow leaves in the centre of the

leaf cluster. As noted by most collectors of this species,

the flowers are sweetly scented. This is not so in all

the species.

Baker, when intending to describe a dried specimen

sent in by Dr. P. MacOwan, No. 2122, originally from

“No Man’s Land” in the eastern Cape, mistakenly con-

cluded that it was the same as two plants flowering at

that time in the Kew Gardens. One of these plants

was figured on Plate 6381 of the Botanical Magazine,

and it is in no wise like MacOwan 2122. It is C. moorei

Hook.f., described four years earlier. Anyone who has

tried to identify a Crinum from pressed material will

appreciate that such a mistake could be made but

from the living plants it is obvious that these two be-

long to distinct species. Having studied the living

Crinums in the eastern Cape there can be no doubt

that MacOwan 2122 represents the species here de-

scribed. The herbarium specimen with which Baker

dealt, MacOwan 2122, is preserved in the Kew Herba-

rium and since, in Baker’s handwriting, the name

Crinum macowanii (written “Macowani”) appears on

the label, this specimen was selected as the lectotype

by Verdoorn (1956).

With regard to the subspecies confusum Verdoorn

(1966), a further study of the species in the eastern

Cape revealed a number of intermediates in the same

area and it is now concluded that the specimens with

the short perianth tube and long pedicels represent

one of the several localized variants of this species.

Two other such variants are the colour-form in the

Estcourt area and the large, luxuriant form around

Punda Milia and in Portuguese East Africa. The short-

tubed variant of the eastern Cape (i.e. “subspecies

confusum”) is the subject of notes by Herbert (1837)

where he points out under C. variabile that the Crinum

figured on Plate 1178 of the Botanical Magazine is not

C. variabile because the leaves are glaucous and the

flowers do not become “red" as they fade. In this he

is correct but his further conclusion that the figure

“agrees better with the plants that I have of C.

capense ” are not very helpful because, as far as one

can judge, his concept of C. capense (illegitimate name)

includes C. bulbispermum. The figured specimen is

undoubtedly the short-tubed C. macowanii.

A study of the description and photograph of the

type specimen of Crinum gouwsii Traub (1954) to-

gether with specimens collected by Gouws and housed

in the herbarium of the University of Pretoria, show

this to be C. macowanii. It will be interesting to learn

whether any investigator has again found, as Prof.

Gouws did, the unusual chromosome number of

2n = 72 in this or any other species of Crinum.

21. Crinum bulbispermum ( Burm.f ) Milne- Red-

head & Schweickerdt in J. Linn Soc. (Bot.) 53: 161

(1939); Verdoorn in Flow. PI. Afr. 29: t.l 1 50 (1953).

Type: not known to have been preserved.

50

THE GENUS CRINUM IN SOUTHERN AFRICA

Amaryllis bulbisperma Burm.f., Prodr. FI. Cap. 9(1768).

A. longifolia sensu Jacq., Icon. PI. Rar., t.362 (1786);sensu L.f.

in Ait. Hort Kew 1 : 419 (1789), partly, as to Masson’s specime.

cited; sensu Red., Liliac. t.347 (1802-1815); sensu Ker-Gawl

in Bot. Mag. t.661 (1803).

Crinum riparium Herb., App. 23 (1821); Amaryll. 269 (1837).

Type: Burchell’s specimen from the Nu-Gariep figured in Bot.

Reg. on t.546 as A. longifolia var. riparia Ker-Gawl. C. capense

sensu Herb., Amaryll. 269 (1837), non (Mill.) Herb. C. longi-

folium (L.) Thunb., Prodr. 59 (1794), partly, as to descr. and

loc. ; sensu Bak., Handb. Amaryll. 93 (1888); FI. Cap. 6: 201

(1896), partly, excl. name and specimens cited from the coastal

and eastern regions of the Cape.

Bulbs 7-13 cm diam. near the base and narrowing

gradually towards the apex (not abruptly narrowed

into a neck). Leaves glaucous green, sheathing at the

base to form a false stem up to about 30 cm high,

flaccid, arcuate, the outermost the broadest, up to 11

cm broad, all with a narrow cartilaginous border and

scattered minute cartilaginous teeth or hairs, several to

many new leaves complete with tip produced in the

centre. Peduncle from 50-90 cm long, up to 2,5 cm

broad and 1 , 5 cm thick. Spatlie-valves up to about

8 cm long, 3,5 cm broad at base, becoming charta-

ceous but not very thin and reflexed at least the upper

half; bracteoles linear. Umbels 6-16-flowered. Pedicels

of different lengths, 4-9 cm long. Perianth normally

with a long cylindrical tube up to 11 cm long, rarely

tube only about 5 cm in young flowers; segments

conniving in a narrow funnel with the apices sub-

spreading or slightly recurved, white with a dark red

keel or entirely suffused with red and the keel very

dark red, 6-10 cm long, the outer about 2 cm wide,

the inner about 3 cm wide about the middle or in upper

half, apical peak on inner segments broadly acuminate

short and blunt, on the acuminate outer segments

2-4 mm long with minute papillose hairs on inner

base. Stamens declinate white or partly suffused with

deep pink; anthers greyish or light brown; style deep

pink in the upper portion, stigma small. Fruit sub-

globose, green, partly suffused with red-purple, 3-7

cm in diam., crowned with a ring but not beaked.

Seeds large 1-2 cm diam., appearing smooth when

turgid, reticulate later. Plate 15; Fig. 32.

Found along rivers and streams or in damp depres-

sions, in black clay or sandy soil. Recorded from the

Transvaal, Orange Free State, Lesotho and the north-

western Cape in the drainage basins of the Orange and

Vaal rivers practically throughout their lengths; also

from the south-eastern Transvaal (Wakkerstroom and

Piet Retief) and Natal (Estcourt, Ladysmith and Dun-

dee) in the catchment areas of the Pongola and the

Tugela rivers with exceptional records from nearer

the coast, in the Pietermaritzburg region and near the

mouth of the Umhlatuzi River, that is near Richard’s

Bay.

Transvaal. — 2529 (Witbank): Loskop dam (-AD), Theron

1821 ; Middelburg (-CD), Van der Merwe s.n.; Wassenaar s.n.;

King A. 2530 (Lydenburg); 5,6 km W. of Machadodorp (-CA),

Codd & Verdoorn 7593. 2627 (Potchefstroom): Frederikstad

(-CA), Louw 452; Louw s.n.; 9,7 km N. of Vereeniging (-DB),

Verdoorn 2340; Vereeniging (-DB), Leendertz 3890. Henley-on-

Klip (-DB), Bruyn 174. 2628 (Johannesburg): Delmas (-BA),

Dyer & Codd 4765; 0,8 km N. of Delmas (-BA), Codd 2181;

Leslie (-BD), Van der Plank s.n.; 27 km S.F.. of Vereeniging on

north bank of Vaal River (-CC), Verdoorn 2341; between Balfour

and Greylingstad (-DA), Reynolds 4051; Greylingstad (-DB),

Dyer & Verdoorn 2378. 2629 (Bethal): 1 ,6 km W. of Trichardt

(-AD), Tolken 1101; 12,8 km W. of Bethal (-AD), Dyer & Codd

4764 14,4 km S.E. of Ermelo (-DB), Codd 4773; Nooitgedacht

(-DB), Henrici 1694; S. of Ermelo near Vaal River (-DB),

Dyer & Verdoorn 5826. 2630 (Carolina): Carolina (-AA),

Rademacher sub TRV 7482; Mavieriestad (-CA), Pott 5166;

Iswepe (-DC), Sidey s.n. 2725 (Bloemhof): between Bloemhof

and Christiana (-CB), Marloth 800a; 800b. 2729 (Volksrust):

12,8 km S.E. of Paardekop (-BA), Reynolds 4053; Sandspruit

(-BD), Burtt Davy sub TRV 10464; Volksrust (-BD), Weder-

mann & Oberdieck 1229; near Volksrust (-BD) Schweickerdt

633; 3 km N.W. of Sandspruit (-BD), Reynolds 4054. 2730

(Vryheid): St. Helena near Dirkiesdorp (-AC), Devenish 1059;

Piet Retief (-BB), Sidey 2046; Collins sub TRV 13402.

O.F.S. — 2627 (Potchefstroom): Venters Kroon (-CD), Pole

Evans 19477 H. 2727 (Kroonstad): near Vais River (-CA), Pont

304. 2728 (Frankfort): 14,4 km S. of Marsala (-BC), Acocks

13887 ; Wilge Rivier (-BC), Acocks 13896. 2828 (Bethlehem):

31,7 km from Bethlehem on Lindley road, near Vais River

(-AB), Marais 1091; Clarence (-CB), Van Hoepen sub TRV

18158. 2925 (Jagersfontein): Piet River near Fauresmith (-CB),

Henrici 3511.

Natal. — 2729 (Volksrust): Ballengeich, 16 km S. of New-

castle (-DD), Reynolds 4056. 2730 (Vryheid): Vryheid (-DD),

Gerstner (photo only). 2829 (Harrismith): Matiwane station,

19 km N. of Ladysmith (-BD), Reynolds 4057; 20,8 km N.E.

of Ladysmith (-BD), Codd & Dver 6255; 22,4 km W. of Berg-

ville, (-CB), Codd & Dyer 6237. 2830 (Dundee): Vantsdrift,

33,6 km E. of Dundee (-AA), Codd 5613. 2831 (Nkandla):

near Umhlatuzi River mouth (-DD), Dyer 4285; Umhlatuzi,

Fig. 32. — Crinum bulbispermum

on the banks of the Vaal

River near Vereeniging.

Photo by Dr. L. E. Codd.

I. C. VERDOORN

51

Sugar Estate (-DD), Dyer 4285. 2929 (Underberg): 8 km W. of

Estcourt on banks of Little Bushmans River (-BB), Dyer 4860.

2930 (Pietermaritzburg): Howick (-AC), Moll 853; 1097;

Dargle Road (-AC), Mogg 5743; 0,4 km from Albert Falls

(-AD), Bruce 458.

Lesotho. — 2828 (Bethlehem): Leribe (-CC), Dieterlen 304;

304a. 2927 (Maseru): Mamathes (-BB), Guillarmod s.n.

Cape. — 2819 (Ariamsvlei): Namaqualand, on banks of the

Orange River (-CC), Kraphol 4992. 2821 (Upington): Louisvale

(-AC), Van Wyk s.n. 2824 (Kimberley): Warrenton, on banks

of Vaal River (-BB), Hafstrom H964; Caerwinning, on banks

of Vaal (-DA), Esterhuysen 1193. 2923 (Douglas): Reads Drift,

40 km from the junction of the Orange and Vaal rivers (-BA),

Anderson 611.

As pointed out by Milne-Redhead and Schweick-

erdt (1939), this species, generally known as the Oran-

ge River Lily, was mistakenly identified, towards the

end of the 18th century, as Amaryllis longifolia L. (see

L.f., 1784), which these authors have shown to be

Cybistetes longifolia (L.) Milne-Redh. & Schweick.

They suggested C. bulbispermum (Burm.) Milne-Redh.

& Schweick. as the correct name for the Orange River

Lily and Burmann’s description of Amaryllis bulbis-

perma (Burmann 1768) exactly describes the species

as we know it today. Because of the mistaken appli-

cation of Linneaus’s name, it follows that Thunberg’s

combination under Crinum , C. longifolium (L.) Thunb.

was consistently and erroneously applied to this spe-

cies and Baker in Flora Capensis (1896) used it in this

sense, but not only is the name incorrect in that work

but Baker has included at least two foreign elements

in his concept. Having today established the distribu-

tion of C. bulbispermum it is obvious that the specimen

from near Cape Town must belong to another genus

(there is no Crinum native to the Peninsula) and those

from the Transkei are probably Crinum macowanii.

Herbert in his Amaryllidaceae (1837) places the

Orange River Lily under the name C. capense (Mill.)

Herb, which, as pointed out by Milne-Redhead and

Schweickerdt (1939), was an illegitimate name as well

as a wrong identification. From Herbert’s notes under

C. variabile (Herb. 1837) he too included C. macowanii,

as depicted on Bot. Mag. 1. 1 178, in his concept of

our species, the Orange River Lily.

C. bulbispermum is characterised by a combination

of the following features: Leaves flaccid consistently

glaucous green, their sheathing bases forming a stem-

like column up to about 30 cm long with several

narrow, flaccid new leaves developing in the centre;

the perianth segments conniving in a narrow funnel

shape with the apical portion spreading to slightly

recurved, becoming reddish as the flower fades; the

pedicels are long and of various lengths in the same

umbel; anthers not black and fruit not beaked.

EXCLUDED SPECIES

Crinum capense (Mill.) Herb., Amaryll. 269 (1837),

based on Amaryllis capensis Mill., Gard. Diet. ed. 8

(1768), nom. illeg., antedated by Amaryllis capensis

L. = Hypoxis. In addition Miller’s description does not

fit Herbert’s specimen but rather describes Amaryllis

belladonna. As far as is known, no specimen was

preserved of Miller’s species. Herbert was dealing

with the Crinum figured on Bot. Mag. t.661 (in error

631) which is now classified as C. bulbispermum.

Crinum forbesii (Lindl.) Schultes f. in Syst. Veg. 7 :

864 (1830), based on Amaryllis forbesii Lindl. in Trans.

Hort. Soc. Lond. 6, 1 : 87 (1826), confused name. No

type specimen has been traced to date and the descrip-

tion does not fit any known Crinum from the area

travelled by Forbes (see C. delagoense , p. 44,45)

Crinum longifolium (L.) Thunb., Prodr. 59 (1794),

based on Amaryllis longifolia L., Sp. PI. 421 (1753)=

Cybistetes longifolia ( L .) Milne-Redhead 8c. Schweick.

(1939).

52

THE GENUS CRINUM IN SOUTHERN AFRICA

Page

Agapanthus 27

Amaryllis 27,28

belladonna 42, 51

bulbisperma Burm.f 50, 51

cape ns is L 51

cape ns is Mill 51

forbesii Lindl 35, 45, 51

var. purpurea Lindl 44

longifolia L 51

longifolia sensu Jacq. et. al 50

var. riparia Ker-Gawl 50

ornata Ait 38

revoluta L’Herit 41

revoluta sensu Ker-Gawl 46

revoluta L’Herit. var. gracilior 41

variabilis Jacq 41

Ammocharis 28,30

Ammocharis sensu Milne-Redh. & Schweick 28

baumii (Harms) Milne-Redh. & Schweick 28, 30

Baker, J. G 27, 28

Codonocrinum Bak 27, 38

Crinum L 27, 28

acaule Bak 39, 40

africanum L 27

algoense Herb 41

americanum L 27, 28, 38

aquaticum Burch 34

asiaticum L 27

baumii Harms 30

bulbispermum (Burm.f.) Milne-Redh 42, 49

buphanoides Wetw. ex Bak 31, 33

campanulatum Herb 27, 34

capense (Mill.) Herb 49, 51

capense sensu Herb 50

carolo-schmidtii Dinter 37

colenso Hort ., 38

crassicaule Bak 32

crassifolium Herb 41

crispum Phillips 45, 46

delagoense Verdoorn 44, 45

euchrophyllum Verdoorn 33

foetidum Verdoorn 27, 28, 32, 34

forbesianum Herb 44, 45

var. punica Herb 44

forbesii (Lindl.) Schultes f 27, 44, 51

forbesii sensu van der Walt 35

giganteum Andr 38

gouwsii Traub 46, 49

graminicola Verdoorn 43

imbricatum Bak 38

jagus ( Thompson ) Dandy 37, 38

kirkii Bak 38

latifolium L 27

leucophyllum Bak 31

longifolium (L.) Thunb 42, 50, 51

lineare L.f. 41,42

lugardiae N.E. Br 45

mackenii Hort 38

macowanii Bak 42, 45, 46, 51

subsp. confusum Verdoorn 42, 46, 49

moorei Hook.f. 27, 28, 37, 49

minimum Milne-Redh 40

natalense Hort 38

nerinoides Bak 30

occiduale R.A. Dyer 37

ornatum Bury 38, 39

paludosum Verdoorn 35, 37

parvibulbosum Dinter ex Overkott 40

polyphyllum Bak 45, 46

rautanenianum Schinz 35, 36, 37

revolutum (L’Herit) Herb 41

riparium Herb 50

subcernuum Bak 33

variabile (Jacq.) Herb 41, 49. 51

zeylanicum L 43

Cybistetes longifolia ( L .) Milne-Redh. & Schweick. . 51

Excluded species 51

Friedberg, J 36

Hannibal, L. S 27, 28

Herbert, The Hon. & Rev. W 27

Hitchcock & Green 27

Hypoxis 51

Isaac & McGillivray 28

Key to Species 29

Linnaeus 27

Platyaster 27, 38

References 28

Stenolirion 28

Newton, F 30

Traub. H. P 27, 28

Uphof, J. C 27, 28

Van der Walt, Geerthsen & Robbfrtse 27, 28

Bothnia 11, 1 &2: 53—102(1973)

The South African Species of Commiphora *

J. J. A. VAN DER WALT |

ABSTRACT

A revision of the South African species of Commiphora (Burseraceae) is presented in which 2 keys are provided

to the 18 species recognized. A comprehensive morphological study, including an anatomical study of the

stems and leaves, was regarded as essential for an accurate delimitation of the different species. Maps, sketches

and photographs serve for illustration.

CONTENTS

Page

Introduction 53

Generic Citation 54

Key to the Species Based on Vegetative

Characteristics 55

Key to the Species Based on all charac-

teristics 56

1. Commiphora glandidosa Schinz 57

2. C. pyracanthoides Engl 60

3. C. merkeri Engl 63

4. C. schimperi (O. Berg) Engl 65

5. C. africana (A. Rich.) Engl 68

6. C. neglecta Verdoorn 71

7. C. mollis (Oliv.) Engl 73

8. C. harveyi (Engl.) Engl 76

9. C. marlothii Engl 78

Page

10. C. edulis (Klotzsch) Engl 81

11. C. woodii Engl 83

12. C. zanzibarica (Baill.) Engl 85

13. C. tenuipetiolata Engl 87

14. C. angolensis Engl 90

15. C. namaensis Schinz 92

16. C. gracilifrondosa Dinter ex Van der

Walt 95

17. C. capensis (Sond.) Engl 96

18. C. cervifolia Van der Walt 99

Discussion of the Relationship of the

Species 101

Opsomming 102

References 102

INTRODUCTION

Berg (1 862) was apparently the first author to publish

a classification of Commiphora under the name

Balsamodendrum Kunth. He divided the 13 species

into two sections, using the type of inflorescence and

the structure of the calyx as criteria. These two

sections were divided into various subsections,

mainly based on leaf characteristics. Berg did not

assign names to these sections or subsections.

Engler (1883) extended the classification of Berg,

recognizing 35 Commiphora species, which he divided

into 18 subsections. This classification of Engler was

primarily based on leaf characteristics and again no

names were assigned to the subsections.

In 1896 Engler revised his classification of 1883

and recognized 63 species. The classification of 1896

was once more extended by Engler in 1913. In this

much more elaborate classification, 129 species were

divided into 43 sections which he published validly

with names and diagnoses. The classifications of

Engler published in 1915 and 1931 were principally

repetitions of the 1913 classification with a few

modifications and additions. Although Engler based

the ultimate division of the 43 sections mainly on leaf

characteristics such as the type of leaf, hairiness of

the leaves, number, colour, shape and margins of the

leaflets, he already realized the taxonomic importance

of the pseudaril. The variable structure of the pseudaril

was used as a criterion for distinguishing between two

of the sections.

* Forms part of a D.Sc. thesis submitted to the University

of Pretoria.

f Botany Department, University of Stellenbosch.

Sprague (1927) and Chiovenda (1932) adopted

Engler’s classification, but Wild (1959a) regarded the

system as artificial providing no guide to the natural

relationships of the species. The classification of

Wild (1959a) was mainly based on characteristics of

the fruit, inflorescence and flower, although leaf

characteristics were used in the subdivision of the

sections and subsections. According to Wild the

structure of the pseudaril, shape and surface of the

putamen and the structure of the disk in the flowers

are of vital importance. Wild reduced the number of

sections considerably and he also showed that many

of Engler’s species were in fact synonyms. He dis-

tinguished 185 species although 266 names for

Commiphora species had been published in the Index

Kewensis. There are many reasons for the great

number of synonyms. Burtt (1935) mentioned that

specimens in European herbaria often consist of

leafless twigs without any flowers or fruit. The irregular

branching, the presence of thorns and the fact that

the leaves often fall off during the preparation of the

specimens contribute to the poor quality of the

specimens obtained. White (1962) noted that the

plants are leafless for a great part of the year even

when flowers and fruits are produced. In addition,

it should be mentioned that important taxonomic

features such as those of the pseudaril and flower

are lost during the drying process. Ripe fruits with

exposed pseudarils are very attractive to birds and

are seldom found on plants.

54

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Sonder (1860), Harvey (1862), Hiern (1896), Burtt

Davy (1932), Verdoorn (1951), Codd (1951), Von

Breitenbach (1965) and de Winter (1968) were the

main contributors to the knowledge of the South

African species of Commiphora. In the past, descrip-

tions of the species were mainly based on external

morphological features, but due to insufficient

available material, the knowledge of these features

was also incomplete. South African botanists, such

as Verdoorn and de Winter, who are particularly

interested in this genus, realize that it is essential to

obtain mature flowers and ripe fruit for diagnostic

descriptions and for the investigation of the relation-

ships of the different species.

This investigation was conducted on 18 species of

Commiphora , so far the only representatives of the

Burseraceae recorded in South Africa. The majority

of species is widely distributed in the central and

northern parts of Transvaal, but they are particularly

common in the dry bushveld of the northern and

north-eastern Transvaal. In the Transvaal, north of

the Tropic of Capricorn, 11 species occur and large

areas north of the Soutpansberg can be designated

as Commiphora-\e Id. Twelve species are recorded from

the Kruger National Park, while the genus is also well

represented in Zululand. A few mesophytic species

occur along the east and south coast, extending as

far south as East London. So far two species from

the northern Cape, and four from the north-western

Cape, have been recorded. The species occurring in

the north-western Cape represent the most xerophytic

species studied.

The aim of this investigation was primarily to

make a contribution to the knowledge of the South

African flora by an accurate delimitation of the

indigenous species of Commiphora. The morphological

investigation was conducted on fresh material

collected for each of the species. A comprehensive

organographic study of the stems, leaves, flowers

and ripe fruit, as well as an anatomical study of the

leaves and stems, was regarded as essential for the

accurate delimitation of the different species. For a

comparative anatomical study of the leaves, it was

decided to study the terminal leaflets of all the species,

and the transverse sections were made a third of the

distance from the base of the leaflets. The anatomy

of the petioles has also been studied from transverse

sections made through the distal part of the petioles.

The anatomical study of the stems included a study

of the young stems and stems with a diameter of

2,5 cm of each species.

The type specimens of all the species including

those of the synonyms, have been studied and, where

applicable, lectotypes have been indicated. All

gatherings cited are represented in the National

Herbarium, Pretoria (PRE), unless otherwise indicated

by the herbarium abbreviation shown after the

collector’s number.

ACKNOWLEDGEMENTS

I wish to express my sincere thanks to my pro-

moters, Prof. H. P. van der Schijff and Dr. B. de

Winter. The greater part of my research work was

done at the Department of Botany, University of

Stellenbosch. I deeply appreciate the facilities put

at my disposal by Prof. P. G. Jordaan, chairman of

the Department. Mr. E. G. H. Olivier of the Botanical

Research Unit, Stellenbosch, acted in liaison with

overseas herbaria in procuring most of the type

specimens. I wish to thank him for this important

contribution and for help in translating this manus-

cript.

I offer thanks to Dr. L. E. Codd and his staff of

the Botanical Research Institute, Pretoria, for their

willing colaboration. Only through the financial aid

given by the C.S.I.R. and the grants made available

by the Research Committee of the University of

Pretoria, could the costs incurred, of collecting trips,

preparation of research material and the actual

compilation of the manuscript, have been defrayed.

For this financial support I am deeply grateful.

COMMIPHORA

Commiphora Jacq., Hort. Schoenbr. 2: 66, t.

249 (1797); Engl, in A. DC., Monogr. Phan. 4: 7

(1883); Bot Jahrb. 15: 94 (1893); in Pflanzenfam.

3,4: 251 (1896); Bot. Jahrb. 26: 368 (1899); Bot.

Jahrb. 34: 303 (1905); Guillaumin in Ann. Sc. Nat.

9,10: 279 (1909); Engl, in Bot. Jahrb. 44: 144 (1910);

Bot. Jahrb. 46: 289 (1912); Bot. Jahrb. 48: 449

(1913); Pflanzenw. Afr. 3,1: 786 (1915); Bot. Jahrb.

54: 292 (1917); Hutch. & Dalz., FI. W. Trop. Afr.

1: 488 (1928); Engl, in Pflanzenfam. ed. 2,19a: 429

(1931); Chiov., FI. Somala 2: 53 (1932); Burtt in Kew

Bull. 1935: 101 (1935); Webber in Lilloa 6: 443 (1941);

Perr. Bathie in FI. Madag. 5: 5 (1946); Exell &

Mendonca in Consp. FI. Angol. 1 : 298 (1951); Miller

in J. S. Afr. Bot. 18: 38 (1952); Wild in Bol. Soc.

Brot. 2,33: 76 (1959); Dale & Greenway, Kenya

Trees: 76 (1961); Capuron in Adansonia 2: 270

(1962); White, For FI. N. Rhod.: 173 (1962); Wild

in FI. Zamb. 2,1: 263 (1963); Yon Breitenbach, Ind.

Trees S. Afr. 3,2: 429 (1965); De Wint. in Trees S.

Afr. 20,1: 3 (1968); Merxm., Prod. FI. S.W. Afr.

23: 1 (1968). Type species: C. madagascariensis

Jacq., Hort. Schoenbr. 2: 66, t. 249 (1797).

Amyris sensu Linn., Mant. : 65 (1767).

Balsamea Gled. in Berl. Ges. Naturf. Fr. Schr. 3:

127 (1782); Engl., Bot. Jahrb. 1: 41 (1881).

Balessan Bruce, Trav. 5: t. 25 (1790).

Balsamodendrum Kunth in Ann. Sc. Nat. 1,2: 348

(1824); DC., Prodr. 2: 76 (1825); Sond. in FI. Cap.

1: 526 (1860); O. Berg in Bot. Ztg. 21: 161 (1862);

Marchand in Adansonia 8: 34, 67 (1867); Oliv. in

FI. Trop. Afr. 1: 324 (1868).

Hemprichia Ehrenb. in Linnaea 4: 396 (1829);

Marchand in Adansonia 8: 69 (1867).

Heudelotia A. Rich, in Guill., Perr. & A. Rich.,

FI. Sen. 1: 150, t. 39 (1832).

Protium sensu Wight & Arn. in Prod. FI. Ind.:

176 (1834); Harv. in FI. Cap. 2: 592 (1862).

Protionopsin Blume in Mus. Bot. Lugd.-Bat. 1 :

229 (1850) nom. nud.

Hitzeria Klotzsch in Peters, Reise Mossamb. Bot.

1: 89 (1861).

Balsamophloeos O. Berg in Bot. Ztg. 20- 163

(1862).

Dioecious or polygamous but rarely monoecious

many-stemmed shrubs or shrubs with the trunk

branching repeatedly above soil level or trees with a

single main stem of variable height; bark often peeling

or flaking in papery pieces or strips; resin ducts

secreting an odoriferous resin occurring in the phloem;

wood relatively light and consisting mainly of septated

fibres; branchlets often spine-tipped, glabrous, pilose

or tomentose. Leaves petiolate but rarely sessile or

J. J. A. VAN DER WALT

55

subsessile, alternate, usually grouped at the ends of

the branches, simple, trifoliolate or impari-pinnate,

margins or leaflets usually crenate, serrate or lobed

but seldom entire, glabrous, pilose or tomentose,

leaflets dorsiventral or isobilateral; petioles of a few

species with medullary vascular bundles. Flowers

unisexual rarely bisexual, perigynous or hypogynous,

male flowers usually larger than female flowers,

appearing before or with the leaves and occasionally

after the leaves in axillary simple or compound

dichasial cymes, in paniculate cymes or singly in

clusters. Pedicels of variable length, glabrous or

pilose to tomentose. Calyx infundibuliform, cam-

panulate or broadly campanulate with 4 valvate

persistent lobes, usually yellowish-green or reddish-

green, glabrous, glandular or pilose to tomentose, in

perigynous flowers continuous with hypanthium, in

hypogynous flowers inserted on receptacle. Petals 4,

usually yellow to green, apex incurved, glabrous or

occasionally pilose on outside. Disk in perigynous

flowers adnate to hypanthium, cylindrical, rarely

fleshy, sometimes lobed; in hypogynous flowers not

adnate to calyx or corolla, intrastaminal, cylindrical,

usually with 4 large lobes but in some species with

4 large and 4 small lobes, lobes bifid or not bifid;

disk in male flowers usually more fleshy than in

female flowers, glabrous or occasionally pilose.

Stamens 8 or in a few species 4, obdiplostemonous,

4 antisepalous stamens longer than other 4; filaments

subterete but lower part usually flattened and broad-

ened, inserted on the outside or on top of disk;

anthers introrse and adnate; staminodes in female

flowers. Gynoecium rudimentary in male flowers;

half inferior in perigynous flowers and superior in

hypogynous flowers, usually glabrous but occasionally

glandular or pilose; ovary ovoid, 2-locular with 2

epitropous ovules per loculus; style of variable

length but usually relatively short; stigma capitate,

obscurely 2-4 lobed. Fruit an ovoid, ellipsoid or

subglobose drupe, usually asymmetrically flattened;

exocarp relatively thin, glabrous but occasionally

pilose; mesocarp usually fleshy, consisting of spongy

tissue with resin ducts; exocarp and mesocarp

splitting in ripe fruit into 2 longitudinal valves (4

valves in a few species outside our area); endocarp

forming a crustaceous or bony putamen and usually

also a pseudaril; putamen ellipsoid or subglobose,

irregularly flattened, smooth or rugose, usually

enclosing one fertile loculus and a much smaller

abortive loculus; seed with a straight embryo,

cotyledons much folded; pseudaril clasping putamen,

usually red or yellowish, usually fleshy but in a few

species thin or membranous or absent, cupular with

short lobes or arms or with 2-4 relatively long arms

or covering almost whole putamen without distinct

arms.

Commiphora is represented in Arabia and western

India by only eight species, all the other species

occurring on the continent of Africa or on islands

along the east coast of Africa. The genus is well

represented in Madagascar and the Mascarenes where

25 species occur.

According to Jacquin (1779) the type species, C.

madagascariensis Jacq., is a plant from Madagascar

and Mauritius although the specimen from which

the plant was described was a cultivated plant. C.

madagascariensis has apparently never been re-

collected in either Madagascar or Mauritius. There is

evidence that species of Commiphora have been much

in demand for their resin from the earliest times

and it may be that the type species was widely cul-

tivated in the past (Wild, 1959a). According to Wild

it could easily have been in cultivation in Madagascar

or Mauritius before 1797 when Jacquin described it.

Engler (1931) and Wild (1959a&b, 1963) regarded

the leaves of C. glandulosa, C. pyracanthoides and C.

merkeri as unifoliolate. According to them, repre-

sentatives of the genus with pinnate leaves are

primitive, and the unifoliolate leaves are developed

by way of reduction. Sinia (1938) and Leenhouts

(1959) rejected this theory and stated that the pinnate

condition is advanced. The phylogeny of Commiphora

species needs further investigation, but observations

made during this study, support the view of Sinia and

Leenhouts. Since no articulation exists in the petioles

of C. glandulosa, C. pyracanthoides and C. merkeri,

I prefer to designate the leaves as simple rather than

unifoliolate.

Key to the Species Based on Vegetative Characteristics

Branchlets spine-tipped :

Leaves simple or trifoliolate with 2 much smaller lateral leaflets:

Bark grey with large black lenticels and peeling off around the stems in yellowish papery strips,

branchlets smooth and purplish, leaves glaucous, small trees with a single stem 3. C. merkeri

Bark yellow to green and flaking in yellowish papery pieces, branchlets greyish, leaves green, trees

with a single main stem or many-stemmed shrubs:

Many-stemmed shrubs up to 3 m tall, terminal leaflet up to 4x2 cm, leaflet-margins finely crenate-

serrate or entire 2. C. pyracanthoides

Trees with a single main stem up to 8 m tall, terminal leaflet up to 6x3 cm, leaflet-margins finely

crenate-serrate, rarely entire L C. glandulosa

Leaves trifoliolate with the lateral leaflets at least half the size of terminal leaflet :

Branchlets and leaves pilose to tomentose 5. C. africana

Branchlets and leaves glabrous or with a few scattered short hairs:

Branchlets and leaves glabrous, leaflets elliptic to broadly elliptic, margins coarsely crenate-serrate

especially in upper half, terminal leaflet dorsiventral 4. C. schimperi

Branchlets and leaves with a few scattered short hairs, leaflets elliptic to ovate to broadly ovate,

margins entire or upper half finely crenate-serrate, terminal leaflet isobilateral 6. C. neglecta

Branchlets not spine-tipped:

Leaves simple 15. C. namaensis

Leaves trifoliolate or impari-pinnate:

Branchlets and leaves pilose to tomentose:

Leaflets often distinctly paler below, margins entire, bark peeling in thick discs, trunk otter irregu-

larly fluted 7. C. mollis

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Leaflets not distinctly paler below, margins not entire (or entire as usually the case with C. edulis ),

bark flaking or peeling in thin papery pieces, trunk not fluted:

Leaves trifoliolate or impari-pinnate, lateral leaflets up to 2x1,2 cm, branchlets not obtuse,

terminal leaflets isobilateral and with hespiridin crystals, petiole without medullary vascular

bundles 14. C. artgolensis

Leaves only impari-pinnate, lateral leaflets larger than 2x1,2 cm, branchlets obtuse, terminal

leaflet dorsiventral and without hesperidin crystals, petiole with medullary vascular bundles:

Bark peeling in large yellowish papery pieces, tree with a single main stem, stems not entwined,

leaves dark green, leaflets obovate to broadly elliptic, margins crenate-serrate to finely lobed

9. C. mar lot hi i

Bark flaking in small yellowish papery pieces, many-stemmed shrub or small tree, stems usually

entwined, leaves greyish-green, leaflets narrowly elliptic to narrowly ovate, margins usually

entire, rarely finely crenate-serrate 10. C. edulis

Branchlets and leaves glabrous (or with a few scattered short hairs in the case of C. harveyi):

Leaves impari-pinnate or trifoliolate with relatively large leaflets and the petiole usually much

longer than 2 cm, trees with a single long trunk or many-stemmed:

Branchlets and leaves with a few scattered short hairs, bark usually peeling in large brown

papery pieces 8. C. harveyi

Branchlets and leaves glabrous, bark not peeling or peeling in white papery pieces:

Leaves trifoliolate or impari-pinnate, petiole slender, terminal leaflet isobilateral and with

hesperidin crystals, bark usually peeling in large white papery pieces to expose a glaucous

underlayer 13. C. terwipetiolata

Leaves only impari-pinnate, petiole not slender, terminal leaflet dorsiventral and without

hesperidin crystals:

Leaflets oblanceolate to narrowly elliptic, margins entire to finely serrate, petiole with

medullary vascular bundles 12. C. zanzibarica

Leaflets narrowly elliptic to elliptic, margins crenate-serrate to coarsely crenate-serrate,

petiole without medullary vascular bundles 11. C. woodii

Leaves trifoliolate with relatively small leaflets and the petiole not longer than 2 cm, shrubs with

a short trunk branching repeatedly above soil level:

Leaflets cordate, orbicular or obovate but without irregular lobes, apex usually emarginate but

sometimes obtuse, margins finely lobed 17. C. capensis

Leaflets linear or cultrate and usually with irregular lobes, apex acute to obtuse, margins entire

or coarsely dentate-serrate:

Leaves up to 8 cm long, leaflets linear to cultrate, margins coarsely dentate-serrate, branchlets

slender 16. C. gracilifrondosa

Leaves up to 2 cm long, leaflets cultrate, margins entire irrespective of lobes, branchlets short

and stout 18. C. cervifolia

Key lo the Species Based on all Characteristics

Leaves simple or trifoliolate with 2 much smaller lateral leaflets:

Branchlets not spine-tipped, leaves simple and usually orbicular, flowers perigynous; pseudaril cupular,

covering the lower J of putamen, with 2 arms on seam of putamen ; putamen smooth 1 5. C. namaensis

Branchlets spine-tipped, leaves simple or trifoliolate with 2 much smaller leaflets, leaflets obovate to

elliptic; flowers hypogynous; pseudaril covering most of putamen, with or without 4 arms; putamen

rugose :

Calyx glandular, flowers usually bisexual or female but male flowers rare, fruit subglobose

1. C. glandutosa

Calyx glabrous, flowers predominantly unisexual, fruit distinctly apiculate:

Bark grey with large black lenticels and peeling off around the stem in yellowish papery strips,

small tree with a single main stem, disk in flowers not fluted, pseudaril yellow and without

distinct arms, stems without a sclerenchymatous pericycle 3. C. merkeri

Bark yellow to green and flaking in yellowish papery pieces, many-stemmed shrub, disk in flowers

fluted, pseudaril red and with 4 distinct arms, stems with a sclerenchymatous pericycle

2. C. pyracanthoides

Leaves impari-pinnate or trifoliolate with the 2 lateral leaflets at least half the size of terminal leaflet:

Branchlets spine-tipped, leaves trifoliolate, flowers hypogynous:

Branchlets and leaves pilose to tomentose, leaflets obovate or occasionally elliptic, flowers only

unisexual, fruit subglobose, putamen rugose 5. C. africana

Branchlets and leaves glabrous or with a few scattered short hairs, leaflets elliptic or ovate, flowers

bisexual or unisexual, fruit subglobose or ellipsoid and apiculate, putamen rugose or smooth:

Branchlets and leaves glabrous, leaflet-margins coarsely crenate-serrate, flowers only bisexual,

fruit ellipsoid and distinctly apiculate, pseudaril membranous and without distinct arms,

putamen rugose 4. C. schimperi

Branchlets and leaves with a few short scattered hairs, leaflet-margins entire or upper half finely

crenate-serrate, flowers bisexual or unisexual, fruit subglobose, pseudaril fleshy and with

4 arms, putamen smooth 6. C. neglecta

Branchlets not spine-tipped, leaves trifoliolate or impari-pinnate, flowers hypogynous or perigynous:

Branchlets and leaves pilose to tomentose:

Leaves trifoliolate or impari-pinnate, lateral leaflets not larger than 3,5 X 2 cm, veins on lower

surface of leaflets not conspicuously raised, petiole without medullary vascular bundles,

fruit less than 1,5 cm in diameter:

Leaflets often distinctly paler below and margins entire, flowers hypogynous, petals pilose to

pubescent on outside, disk fleshy and not adnate to calyx and corolla, pseudaril with 4

winged arms, tree with a single main stem 7. C. mollis

J. J. A. VAN DER WALT

57

Leaflets not distinctly paler below and margins crenate-serrate, flowers perigynous, petals

glabrous or outside sparsely pilose, disk reduced and adnate to hypanthium, pseudaril

cupular with 2 lobes, many-stemmed shrub 14. C. angolensis

Leaves impari-pinnate, lateral leaflets usually much larger than 3,5x2 cm, veins on lower surface

of leaflets conspicuously raised, petiole with medullary vascular bundles, fruit more than

1,5 cm in diameter:

Bark peeling in large yellowish papery pieces, leaves dark green, leaflets obovate to broad elliptic,

flowers hypogynous; disk fleshy, pilose, not adnate to calyx or corolla; pseudaril yellow,

with 2 long and 2 short arms 9. C. marlothii

Bark flaking in small yellowish papery pieces, leaves greyish-green, leaflets narrowly elliptic to

narrowly ovate, flowers perigynous; disk much reduced, glabrous, adnate to hypanthium;

pseudaril red, cupular with 4 short lobes 10. C. edulis

Branchlets and leaves glabrous (or with a few scattered short hairs in the case of C. harveyi):

Leaves impari-pinnate or trifoliolate with relatively large leaflets and the petiole usually much

longer than 2 cm, inflorescences relatively long paniculate dichasial cymes or dichasial cymes,

trees with a long trunk or many-stemmed trees:

Branchlets and leaves with a few scattered short hairs, bark usually peeling in large brown papery

pieces, inflorescences with large leaf-like bracts, flowers hypogynous, disk not adnate to

calyx or corolla, pseudaril with 4 arms 8. C. harveyi

Branchlets and leaves glabrous, bark not peeling or peeling in white papery pieces, inflorescences

with relatively small bracts, flowers perigynous, disk adnate to hypanthium, pseudaril

cupular with 0-2 short lobes:

Leaves trifoliolate or impari-pinnate, petiole slender, bark usually peeling in large white papery

pieces to expose a glaucous underlayer, inflorescences simple or compound dichasial

cymes, pseudaril cupular with 2 lobes of varying length 13. C. tenuipetiolata

Leaves only impari-pinnate, petiole not slender, bark not peeling, inflorescences paniculate

dichasial cymes, pseudaril cupular without lobes or with one short lobe:

Leaflets oblanceolate to narrowly elliptic, margins entire to finely serrate, petiole with

medullary vascular bundles, inflorescences very long, pedicels relatively long (4-6 mm);

pseudaril cupular, covering the lower \ of putamen, without lobes, margin coarsely

crenate 12. C. zanzibarica

Leaflets narrowly elliptic to elliptic, margins crenate-serrate, petiole without medullary

vascular bundles, inflorescences long, pedicels relatively short (less than 1 mm); pseudaril

cupular, covering the lower 3 of putamen, with 1 short lobe, margin finely crenate

11. C. woodii

Leaves trifoliolate with relatively small leaflets and the petiole not longer than 2 cm, flowers borne

singly or in short simple dichasial cymes, shrubs with a short trunk branching repeatedly

above soil level :

Leaves up to 8 cm long, flowers with only 4 stamens/staminodes, fruit subglobose to ellipsoid,

pseudaril fleshy, cupular with 2 arms 16. C. gracilifrondosa

Leaves up to 8 cm long, flowers with 8 stamens/staminodes, fruit ellipsoid and flattened, pseudaril

absent:

Leaflets cordate, orbicular or obovate, without irregular lobes, apex usually emarginate but

sometimes obtuse, margins finely lobed 17. C. capensis

Leaflets cultrate, irregularly lobed, margins entire irrespective of lobes 18. C. cervifolia

1. Commiphora glandulosa Schinz in Bull. Herb.

Boiss. 2,8: 633 (1908); Codd, Mem. Bot. Surv. S.

Afr. 26; 86 (1951); Exell & Mendonca in Consp. FI.

Angol. 1,2: 298 (1951); Miller in J. S. Afr. Bot. 18:

38 (1952); Brenan in Kew Bull. 1953: 106 (1953).

Syntypes: S.W.A., Ombandja, Schinz 767 (Z !) ;

Ondangau, Schinz s.n. (not seen). Lectotype: Schinz

767 (Z).

Commiphora lugardae N.E. Br. in Kew Bull. 1909: 99 (1909);

Miller in J. S. Afr. Bot. 18: 38 (1952). Type: Botswana, Kwebe

Hills, Lugard 23 (K, holo.!). C. seineri Engl, in Bot. Jahrb.

44: 145 (1910); Bot. Jahrb. 48: 480 (1913); Pflanzenfam.

2,19a: 437 (1931). Type: Zambia, Sesheko, Seiner 57 (B, holo.!;

K, photo!; BM, sketch!). C. berberidifolia Engl, in Bot.

Jahrb. 48: 480 (1913); Pflanzenfam. ed. 2,19a: 437 (1931).

Type: S.W.A., Okahandja, Waldau, Dinter 385 (B, holo.!;

K, fragment!). C. pyracanthoides subsp. glandulosa (Schinz)

Wild in Bol. Soc. Brot. 2,33: 44 (1959); FI. Zamb. 2,1: 268

(1963). Von Breitenbach, Ind. Trees S. Afr. 3,2: 433 (1965);

Merxm., Prod. FI. S.W.Afr. 23: 8 (1968).

Polygamous or dioecious tree up to 8 m tall with

a single main stem; bark purple-grey to green,

flaking in yellowish papery pieces to expose a green

underlayer; branchlets spine-tipped. Leaves simple

or trifoliolate, up to 6,3 cm long, with long glandular

hairs especially at base of laminae; petiole up to 3 mm

long, with long glandular hairs at distal end; petiolules

less than 1 mm long ; terminal leaflet up to 6 X 3 cm, nar-

rowly obovate to broadly obovate, rarely elliptic, apex

acute or obtuse, base cuneate, margins finely crenate-

serrate, rarely entire; lateral leaflets up to 1,5 x 0,8 cm,

elliptic. Flowers subsessile, bisexual or unisexual but

male flowers rare, hypogynous, appearing before the

leaves in axillary clusters on side shoots or spines, in

some cases reduced dichasial cymes up to 1,5 cm

long; bisexual and male flowers, 6-8 mm, larger than

female flowers, 4, 5-5, 5 mm. Bracteoles up to 1 mm

long, lanceolate, with numerous glandular hairs.

Pedicels usually less than 1 mm long, with numerous

glandular hairs. Calyx campanulate, green to red,

2-3,5 mm long, with numerous long glandular hairs,

lobes less than 1 mm long, apex acute. Petals yellowish

green to red, 4-6 mm long, without glandular hairs.

Disk fleshy, not adnate to calyx and corolla, cylin-

drical with 4 prominent lobes, indentations between

lobes shallow, lobes bifid. Stamens 8, 4 long stamens

up to 4,5 mm long, inserted high up on the outside

of disk lobes, 4 short stamens up to 3 mm long,

inserted on the outside of disk between lobes; filaments

subterete, lower part flattened and broadened;

staminodes in female flowers. Gynoecium: rudimentary

in male flowers; ovary superior; style variable in

length; stigma obscurely 4-lobed. Fruit 1,1x1 cm,

subglobose, slightly flattened, asymmetrical; exocarp

glabrous; mesocarp fleshy; putamen 7x5 mm,

ellipsoid, asymmetrical with one face more convex

than the other, rugose; pseudaril red, fleshy, with 4

58

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

arms of equal length reaching almost to apex of

putamen, margins of arms irregular, arm on more

convex face of putamen slightly broader than arm on

the other face. Fig 1-7.

Fig. 1. — Commiphora glandulosa near Waterpoort, northern

Transvaal (height ±4 m).

Fig. 2. — Close-up view of the trunk of Commiphora glandulosa

illustrating the bark flaking off in small papery pieces.

Fig. 3.— Commiphora glandulosa: A, branchlet with flowers and

young fruits; B— E, leaves; F, branchlet with leaves and

mature fruits.

Fig. 4. — Flowers of Commiphora glandulosa: A, bisexual

flower; B, bisexual flower with calyx and corolla partly

removed; C, diagrammatic representation of two disc lobes

illustrating the insertion of the filaments; D, female flower;

E, female flower with the calyx and corolla partly removed.

J. J. A. VAN DER WALT

59

Fig. 5. — Fruit of Commiphora glandulosa: A, sideview of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Fig. 7. — Commiphora glandu-

losa (in the background)

and C. pyracanthoides (in

the foreground) near Mes-

sina, northern Transvaal.

Distinctive anatomical features of the stems and leaves

Young stems with a few glandular hairs especially

near apex. Stems of 2,5 cm diameter: sclerenchyma-

tous pericycle-cylinder consisting of fibres and stone

cells, epithelium cells of resin ducts in xylem rays

surrounded by 1-2 layers of cells with slightly

thickened walls. Leaves with relatively long glandular

hairs; petiole semi-circular as seen in transverse

section, sclerenchymatous pericycle absent, vascular

bundles triangularly distributed as seen in transverse

section; terminal leaflet typically dorsiventral with

a single layer of palisade cells adaxially, remaining

mesophyll consisting of spongy parenchyma, bulli-

form cells confined to adaxial epidermis; stomata

mainly abaxial.

60

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Diagnostic features

Polygamous or dioecious tree with a single main

stem; bark purple-grey to green, flaking in yellowish

papery pieces to expose a green underlayer; branchlets

spine-tipped. Leaves simple or trifoliolate with 2

small lateral leaflets, with long glandular hairs at

base of laminae and distal end of petiole, terminal

leaflet typically dorsiventral. Flowers subsessile;

hypogynous; bisexual or unisexual but male flowers

rare; calyx with numerous long glandular hairs;

disc lobes 4, bifid. Fruit subglobose; putamen rugose;

pseudaril red with 4 arms of equal length reaching

almost to apex of putamen.

Widely distributed in northern Zululand, north-

western, northern, far-northern and north-eastern

Transvaal, but is particularly common north of the

Soutpansberg. Also collected in the northern Cape.

Usually grows in sandy, well-drained soil in areas

with a relatively low annual rainfall, and occurs in

savanna-woodland or in broken mopaniveld.

Also recorded from South West Africa, Botswana,

Rhodesia, Zambia, Mozambique and Angola.

Transvaal. — 2229(Waterpoort): Dongola (-BC), Pole-

Evans 4450; Verdoorn 2297 A ; 2325; Soutpan (-CD), Obermeyer,

Schweickerdt & Verdoorn 159; near Mopane (-DB), Strey

3501 ; 8 km E. of Waterpoort (-DC), Van der Walt 20; Wyllies-

poort (-DD), Story 1857. 2230( Messina): near Messina

(-AC), Hafstrom & Acocks 1882; Rogers 18475; Tshipise

(-CA), Van der Schijff 5238 (PRU); Van der Walt 3; 29 km E,

of Tshipise (-CB), Van der Schijff 5239. 2231(Pafuri) : 7 km

N.E. of Punda Milia(-CA), Codd & Dyer 4569. 2327(Ellisras):

16 km S.S.E. of Ellisras (-DD), Van der Walt 108. 2328(Balti-

more): 32 km N.W. of Melkrivier (-CD), Van der Walt 47;

near Sterkwater (-DD), Van der Walt 51. 2329(Pietersburg):

near Vivo (-AB), Strey 3518. 2331(Phalaborwa): Shingwidzi

Rest Camp (-AB), Codd 4652; Gorge Rest Camp (-DD),

Van der Schijff 839. 2427(Thabazimbi): 50 km S. of Ellisras,

(-BA), Van der Walt 54; 40 km N.N.W. of Vaalwater (-BB)

Smuts 352; 22 km S.E. of Bulge River (-BB), Van der Walt 52;

W. of Krantzberg (-CB), Codd 4432; Galpin 13377; 35 km N.W.

of Northam (-CC), Van der Walt 74. 2429(Zebediela): 13 km

S.S.E. of Roedtan (-CA), Meeuse 9497. 243I(Acornhoek):

near Olifants Rest Camp (-BA), Codd 4290; near Satara

Rest Camp (-BD), Van der Schijff 3497. 2527(Rustenburg):

S. of Pilanesberg (-AC), Codd 1107; at confluence of Crocodile

and Pienaars Rivers (-BA), Codd 9839. 2529(Witbank: Loskop-

dam Nature Reserve (-AD), Codd 10365; Van der Walt 14.

Natal. — 2632(Bela Vista): Ndumo Game Reserve (-CC),

Gerstner 3 148 ; Tinley 577 ; Van der Walt 99; 102. 2831(Eshowe):

Umfolozi Game Reserve (-BD), Feely 65; Ward 4061 (NH).

Cape. — 2724(Taung): River Valley at Tierkloof (-BA),

Brueckner.

Wild (1959b) considers this taxon as a subspecies

of C. pyracanthoides Engl. This taxonomic change

by Wild is based mainly on observations made by

Merxmliller in South West Africa where C. glandulosa

occurs in tree and shrub form. However,

the flower, and fruit structure of these two taxa

differ to such an extent that they should be considered

as different species.

This species is easily grown from pole cuttings

which are often planted as fencing poles.

Common names: Corkwood (“Kurkhout”) and

“Kanniedood”.

2. Commiphora pyracanthoides Engl, in Bot.

Jahrb. 26: 368 (1899); Bot. Jahrb. 48: 481 (1913);

Pflanzenfam. ed. 2,19a: 437 (1931); Burtt Davy, FI.

Transv. 2: 485 (1932); Miller in J.S. Afr. Bot. 18:

38 (1952); Brenan in Kew Bull. 1953: 104 (1953);

Wild in Bol. Soc. Brot. 2,33: 43, 82 (1959); White,

For. FI. N. Rhod.: 176, t. 34A (1962); Wild in FI.

Zamb. 2,1: 268 (1963); Yon Breitenbach, Ind. Trees

S. Afr. 3,2: 432 (1965); De Wint. in Trees S. Afr.

20,1: 16 (1968); Merxm., Prod. FI. S.W. Afr. 23:

8 (1968). Type: S.W. A., Otjimbingwe, Fischer 8

(holo.f; ?); Neotype: S.W. A., Little Karas Moun-

tains, Holoog, Pearson 9747 (K!).

Dioecious or polygamous many-stemmed shrub

up to 3 m tall; bark yellow to green, flaking in yel-

lowish papery pieces to expose a green underlayer;

branchlets spine-tipped. Leaves simple or trifoliolate

up to 4,5 cm long, with long glandular hairs especially

at base of laminae; petiole up to 3 mm long, with long

glandular hairs at distal end; petiolules less than 1 mm

long; terminal leaflet up to 4x2 cm, narrowly

obovate to broadly obovate, rarely elliptic, apex

acute to obtuse, base cuneate, margins finely crenate-

serrate, in some cases entire; lateral leaflets 2x 1 cm,

elliptic. Flowers subsessile, predominently unisexual,

rarely bisexual, hypogynous, appearing before the

leaves in axillary clusters on side shoots or spines, in

some cases in reduced dichasial cymes up to 1,5 cm

long; male and bisexual flowers, 8-10 mm, larger

than female flowers, 3,5-5 mm. Bracteoles up to

2 mm long, lanceolate, with a few glandular hairs.

Pedicels up to 1,5 mm long, without glandular hairs.

Calyx campanulate, green to red, 2-4 mm long,

without glandular hairs, lobes up to 1 mm long,

apex acute. Petals yellowish green to red, 3-7 mm

long, without glandular hairs. Disk fleshy, cylindrical,

not adnate to calyx or corolla, folded to form 4

large lobes towards the outside, indentations between

lobes shallow, lobes not bifid, inside of lobes deeply

grooved; disk of female and bisexual flowers smaller

and less fleshy than those of male flowers but the

lobes in some cases bifid. Stamens 8, 4 long stamens

up to 4 mm long, inserted high up on the outside

of lobes, 4 short stamens up to 3 mm long, filaments

subterete, lower part flattened and broadened;

staminodes in female flowers. Gynoecium: rudimentary

in male flowers; ovary superior; style relatively long;

stigma 4-lobed. Fruit 1,2 X 0,8 cm, ellipsoid, irregularly

flattened, asymmetrical, apiculate; exocarp glabrous;

mesocarp relatively thin ; putamen 8x6 mm, ellipsoid,

asymmetrical with one face more convex than the

other, rugose; pseudaril red, not very fleshy, with

4 arms and isolated fragments on putamen, arms of

equal length reaching almost to apex of putamen,

margins of arms irregular, arm on more convex face

of putamen broader than arm on the other face.

Fig. 8-13.

J. J. A. VAN DER WALT

61

Fig. 8. — Commiphora pyra-

canthoides near Waterpoort,

northern Transvaal (height

±1,5 m).

Fig. 1 0. — Commiphora pyracanthoides : A, branchlet with flowers ;

B, C, leaves; D, branchlet with leaves and mature fruits.

5292-5

62

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 12. — Flowers of Commiphora pyracanthoides: A, male

flower; B, longitudinal section of male flower; C, dia-

grammatic representation of two disc lobes illustrating the

insertion of the filaments; D, female flower; E, female

flower with the calyx and corolla partly removed.

Fig. 13. — Fruit of Commiphora pyracanthoides: A, side view

of the fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Distinctive anatomical features of the stems and leaves

Young stems with a few glandular hairs especially

near apex. Stems of 2,5 cm diameter: sclerenchyma-

tous pericycle-cylinder consisting of fibres and stone

cells, epithelium cells of resin ducts in xylem rays

surrounded by 1-2 layers of cells with slightly

thickened walls. Leaves with relatively long glandular

hairs; petiole semi-circular as seen in transverse

section, sclerenchymatous pericycle present or

absent, vascular bundles triangularly distributed as

seen in transverse section; terminal leaflet dorsiventral

with a single layer of palisade cells adaxially, remaining

mesophyll consisting of spongy or palisade-like cells,

bulliform cells confined to adaxial epidermis, stomata

mainly abaxial.

Diagnostic features

Dioecious or polygamous many stemmed shrub;

bark yellow to green, flaking in yellowish papery

pieces to expose a green underlayer; branchlets

spine-tipped. Leaves simple or trifoliolate with 2

small lateral leaflets, with long glandular hairs at

base of laminae and distal end of petiole, terminal

leaflet dorsiventral. Flowers subsessile; hypogynous;

predominantly unisexual, rarely bisexual; calyx gla- 1

brous; disk folded to form 4 large lobes towards the 1

outside, lobes in male flowers not bifid but in female

and bisexual flowers bifid. Fruit ellipsoid; apiculate; ^

putamen rugose; pseudaril red, with 4 arms of 1

equal length reaching almost to apex of putamen.

Widely distributed in northern Zululand, north-

western, northern, far-northern and north-eastern

Transvaal, but is particularly common north of the

Soutpansberg. Also collected in northern Cape.

Usually grows in sandy, well-drained soil in areas

with a relatively low annual rainfall, and occurs in

savanna-woodland, broken mopaniveld and shrub-

thornveld.

Also recorded from Swaziland, South West Africa,

Botswana, Rhodesia and Mozambique.

Transvaal. — 2229(Waterpoort): Dongola (-BC), Dyer 4314;

4300; Verdoorn 2336; 16 km N.E. of Alldays (-CA), Van der

Walt 65; Soutpan (-CD), Obermeyer, Schweickerdt & Verdoorn

48; 8 km E. of Waterpoort (-DC), Van der Walt 10. 2230

(Messina): near Messina (-AC), Dyer 4324; 48 km N.E. of

Tshipise (-BC), Van der Schijff 5241; Tshipise (-CA), Van

der Schijff 5217. 2231(Pafuri): near Klopperfontein (-AC),

Van der Schijff 1861; 21 km N.E. of Punda Milia Rest Camp

(-CA), Codd & Dyer 4578. 2327(Ellisras); 8 km E.N.E. of

Ellisras (-DB), Van der Walt 57; 10 km S.S.E. of Ellisras

(-DD), Van der Walt 119; 120. 2328(Baltimore): 16 km S.

of Marnitz (-AC), Van der Walt 58. 2329(Pietersburg): 21

km S. of Bandelierkop (-BD), Gerstner 5576; 3 km N.E. of

Kalkbank (-CB), Story 1562. 2331(Phalaborwa): 5 km S. of

Shingwidzi Rest Camp (-AB), Van der Walt 76. 2426(Mochudi):

5 km S.E. of Rooibokkraal (-BB), Leistner 3167. 2427(Thaba-

zimbi): 13 km E.N.E. of Rooibokkraal (-AA), Van der Walt 26.

2428(Nylstroom): 5 km N. of Tuinplaats (-DD), Strey 1373.

2429(Zebediela): near Immerpan (-CB), Meeuse 9488; 9 km

N.W. of Marble Hall (-CD), Codd 10367 . 2430 (Pilgrim's Rest):

2 km N. of Origstad (-DA). Codd 6753. 2531 (Komatipoort):

Numbi (-AA), Van der Schijff 731; near Skukuza Rest Camp on

Lower Sabie road (-BA), Van der Schijff 3418.

Natal. — 2632(Bela Vista): Ndumo Game Reserve (-CC),

Van der Walt 100; 101. 2731(Louwsburg): 11 km N.W. of

Candover (-BD), Acocks 13128; Ward 3697 (NH). 2831

(Eshowe): Umfolozi Game Reserve (-BD), Leibnitz, Fakttde

6 Hancox 10; Van der Walt 84.

Cape. — 2725(Bloemhof): near Schweizer-Reneke (-AB)

Van Wyk 14.

J. J. A. VAN DER WALT

63

Brenan (1953) stated that the type specimen

(Fischer 8) could not be traced. Pearson 9747 was

chosen by him as the neotype because this specimen

was sent to Berlin in 1929 where it was compared

with material which Engler himself designated as

C. pyracanthoides.

3. Commiphora merkeri Engl, in Bot. Jahrb.

44: 144 (1910); Bot. Jahrb. 48: 480 (1913); Pflanzen-

fam. ed. 2,19a: 437 (1931); Burtt in Kew Bull. 1935:

110 (1935); Wild in Bol. Soc. Brot. 2,33: 82 (1959);

Dale & Greenway, Kenya Trees 89 (1961); Wild in

FI. Zamb. 2,1: 269 (1963); Von Breitenbach, Ind.

Trees S. Afr. 3,2: 431 (1965); De Wint. in Trees S.

Afr. 20,1: 12 (1968). Type: Tanzania, Nguruka,

Merker 565 (B, holo.f; K, fragment!).

Commiphora viminea Burtt Davy, FI. Transv. 2: 485 (1932);

Codd, Mem. Bot. Surv. S. Afr. 26: 88 (1951); Brenan in Kew

Bull. 1953: 104 (1953). Type: Transvaal, Messina, Moss &

Rogers 184b (K, holo.!).

Dioecious small tree up to 5 m tall; bark grey

with large black lenticels, peeling off around the

stems in yellowish papery strips; branchlets spine-

tipped. Leaves simple or trifoliolate, glaucous, up to

5 cm long, with long glandular hairs especially at

base of laminae; petiole up to 5 mm long, with long

glandular hairs at distal end; petiolules less than 1 mm

long; terminal leaflet up to 4,5 x 2,5 cm, narrowly

obovate to obovate or elliptic, apex acute to obtuse,

base cuneate, margins crenate-serrate especially near

apex; lateral leaflets up to 7x5 mm, elliptic. Flowers

unisexual, hypogynous, appearing before the leaves

or with the young leaves in axillary clusters on side

shoots or spines; male flowers, 10-12 mm, usually

much larger than female flowers, 5-6 mm. Bracteoles

up to 1 mm long, ± triangular, with a few long

glandular hairs. Pedicels 2-7 mm long, without

glandular hairs. Calyx campanulate, yellowish green

to brown, 2-3 mm long, without glandular hairs,

lobes usually less than 1 mm long, apex acute. Petals

yellowish green 3-5,5 mm long, without glandular

hairs. Disk cylindrical, not adnatc to calyx or corolla;

in male flowers very fleshy with 4 prominent lobes,

indentation between lobes shallow, lobes not bifid,

outside of lobes deeply grooved and inside shallowly

grooved; disk in female flowers smaller. Stamens

8, 4 long stamens up to 5 mm long, inserted halfway

up on the outside of lobes; 4 short stamens up to

3,5 mm long, inserted on the outside of disk between

lobes; filaments subterete, lower part flattened and

broadened, lower part of 4 long filaments placed in

grooves of lobes; staminodes in female flowers.

Gynoecium: rudimentary in male flowers; ovary

superior; style long causing stigma to protrude above

petals; stigma obscurely 4-lobed. Fruit 1,3 x 0,7 cm,

ellipsoid, slightly flattened, asymmetrical, very apicu-

late; exocarp glabrous; mesocarp relatively thin;

putamen 8x5 mm, ellipsoid, asymmetrically and

irregularly flattened, rugose; pseudaril yellow, cover-

ing the whole putamen except the apex, forming a

prominent ridge on small face of putamen. Fig.

14-19.

Fig. 14. — Commiphora mer-

keri near Waterpoort,

northern Transvaal (height

±2,5 m).

64

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 15. — Close-up view of a branch of Commiphora merkeri

illustrating the large lenticels and bark peeling off around

the stem in papery strips.

Fig. 16. — Commiphora merkeri: A, branchlet with leaves and 1

flowers; B & C, leaves; D, branchlet with leaves and a

mature fruit.

Fig. 17. — Flowers of Commiphora merkeri: A, male flower; B,

longitudinal section of male flower; C, diagrammatic

representation of two disc lobes illustrating the insertion

of the filaments; D, female flower; E, female flower with

the calyx and corolla partly removed.

Fig. 18 — Fruit of Commiphora merkeri: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

and pseudaril as seen from above.

J. J. A. VAN DER WALT

65

Fig. 19. — Geographical distribution of Commiphora merkeri

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with a few glandular hairs especially

near apex. Stems of 2,5 cm diameter, sclerenchyma-

tous pericycle absent, epithelium cells of resin ducts

in xylem rays surrounded by 1-2 layers of sclereids.

Leaves with relatively long glandular hairs; petiole

semi-circular as seen in transverse section, scleren-

chymatous pericycle present or absent, vascular

bundles triangularly distributed as seen in transverse

section; terminal leaflet typically dorsiventral with

a single layer of palisade cells adaxially, remaining

mesophyll consisting of spongy parenchyma, bulliform

cells confined to adaxial epidermis, stomata mainly

abaxial.

Diagnostic features

Dioecious small tree; bark grey with large black

lenticels, peeling off around the stems in yellowish

papery strips; stems without sclerenchymatous peri-

cycle; branchlets spine-tipped. Leaves simple or

trifoliolate with 2 small lateral leaflets, glaucous, with

long glandular hairs at base of laminae and distal end

of petiole, terminal leaflet typically dorsiventral.

Flowers hypogynous; unisexual: calyx glabrous; disk

lobes not bifid. Fruit ellipsoid; very apiculate;

putamen rugose; pseudaril yellow without distinct

arms, covering the whole putamen except the apex.

Occurs in northern Transvaal from the border of

Botswana in the west to Mozambique in the east,

but is particularly common north of the Soutpansberg.

Grows in well-drained, sandy soil in warm areas with

a relatively low annual rainfall. Occurs in savanna-

woodland.

Also recorded from Rhodesia, Mozambique, Tan-

zania, Kenya and South West Africa.

Transvala. — 2228(Maasstroom): 16 km N.E. of Swartwater

(-CC), Van der Walt 60. 2229(Waterpoort): Dongola area

(-BC), Verdoorn 2301; Codd 4104; 4835; Mopane (-DB),

Strey 3505 ; 8 km E. of Waterpoort, Van der Walt 32. 2230

(Messina): Tshipise (-CA), Van der Walt 6; Van der Schijff

5199 (PRU). 2231(Pafuri): 3 km W. of Pafuri (-AC), Codd &

Dyer 4634; S.E. of Klopperfontein (-AC), Van der Schijff

2948. 2429(Zebediela): Chuniespoort (-AB), De Winter 2327.

The stems often exude large quantities of gum-resin.

Common name: Zebra Tree.

4. Commiphora schimperi (O. Berg) Engl, in

A. DC., Monogr. Phan. 4: 13 (1883); Schweinf. in

Bull. Herb. Boiss. 7,2: 288 (1899); Engl, in Bot.

Jahrb. 48: 477, t.2N (1913); Pflanzenfam. ed. 2,19a:

435, t.204 C-D (1931); Burtt in Kew Bull 1935: 110

(1935); Wild in Bol. Soc. Brot. 2,33: 88 (1959);

Dale & Greenway, Kenya Trees 80 (1961); Wild in

FI. Zamb. 2,1: 277 (1963); Von Breitenbach, Ind.

Trees S. Afr. 3,2: 437 (1965); De Wint. in Trees S.

Afr. 20,1: 16 (1968). Syntypes: Ethiopia, Takazze,

Schimper 1564 (Bf; W!; G!); Schimper 624 (Bf;

K!); Schoata, Schimper 1139 (Bf; W!; G!). Lecto-

type: Schimper 624 (K).

Balsamodendrum schimperi O. Berg in Bot. Ztg. 20: 162

(1862) B. africanum sensu Oliv., FI. Trop. Afr. 1: 325 (1868),

pro parte quoad specim. Schimper.

Balsamea schimperiiO. Berg) Engl, in Bot. Jahrb. 1: 41

(1881).

Commiphora betschuanica Engl, in Bot. Jahrb. 44: 149

(1910); Bot. Jahrb. 48: 478 (1913); Pflanzenfam. ed. 2,19a:

435 (1931); Burtt Davy, FI. Transv. 2: 484 (1932); Codd,

Mem. Bot. Surv. Afr. 26: 85 (1951); Miller in J.S. Afr. Bot.

18: 38 (1952). Type: Botswana, Mugnune, Seiner 64 (B, holo.f;

K, fragment!; BM, sketch!).

Shrub or small tree 2-6 m tall; bark peeling in

black discs or flaking in small yellowish papery

pieces to expose a green underlayer; branchlets spine-

tipped, glabrous. Leaves trifoliolate, glabrous; lamina

up to 7 , 5 cm long ; petiole up to 2 , 8 cm long ; petiolules

up to 2 mm long; leaflets elliptic to broad elliptic,

apex acute, base cuneate, margins coarsely crenate-

serrate especially in the upper half of leaflets; terminal

leaflet up to 5 X 3 , 5 cm ; lateral leaflets up to 2 , 5 x 2 , 2

cm. Flowers only bisexual, hypogynous, appearing

before the leaves in axillary clusters, often borne on

the spines, flowers 6-8 mm long. Bracteoles up to

1 mm long, triangular, with a few glandular hairs.

Pedicels 1-2 mm long, with a few glandular hairs.

Calyx campanulate, green to red, 3-4 mm long,

glabrous, lobes less than 0,5 mm long, apex acute or

obtuse. Petals yellow to pink, 6-7 mm long, glabrous.

Disk fleshy, not adnate to calyx or corolla, cylindrical

with 4 lobes, indentation between lobes shallow,

lobes in some cases bifid. Stamens 8, 4 long stamens

up to 6 mm long, inserted high up on the outside of

disk lobes, 4 short stamens up to 4,5 mm, inserted on

the outside of disk between the lobes; filaments

slightly flattened, lower part broadened. Gynoecium:

ovary superior; style relatively long; stigma 4-lobed.

Fruit 1,7x1 cm, ellipsoid, asymmetrically and irre-

gularly flattened, very apiculate, apex curved; exocarp

rugose and glabrous; mesocarp fleshy; putamen 1,1

X0,7 cm, ellipsoid, asymmetrically and irregularly

flattened, very rugose, slimy; pseudaril red, mem-

branous, covering almost the whole putamen.

Fig. 20-26.

66

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 21, Fig. 22. — Close-up view of different branches of

Commiphora schimperi illustrating the bark flaking in

papery pieces (Fig. 21) and peeling in thicker discs (Fig. 22).

J. J. A. VAN DER WALT

67

Fig. 23. — Commiphora schimperi: A, branchlet with flowers

and young fruit ; B, branchlet with leaves and mature fruits ;

C, branchlet with leaves.

Fig. 25. — Fruit of Commiphora schimperi: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril ; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Fig. 24. — Flowers of Commiphora schimperi: A, bisexual

flower; B, bisexual flower with calyx and corolla partly

removed; C, diagrammatic representation of two disc lobes

illustrating the insertion of the filaments; D, longitudinal

section of the bisexual flower with the gynoecium removed.

Fig. 26. — Geographical distribution of Commiphora schimperi

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with a few glandularhairs especially near

apex. Stems of 2,5 cm diameter: sclerenchymatous

pericycle consisting of separate fibre strands but stone

cells absent, epithelium cells of resin ducts in xylem

rays surrounded by thin-walled cells. Leaves with a

few glandular hairs; petiole heart-shaped as seen in

transverse section, sclerenchymatous pericycle pre-

sent, vascular bundles ± triangularly distributed as

seen in transverse section, vascular bundles on adaxial

side smaller than bundles on abaxial side; terminal

leaflet dorsiventral with a single layer of palisade

68

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

cells adaxially, remaining mesophyll consisting of

spongy or palisade-like cells, adaxial epidermis

consisting mainly of large bulliform cells but smaller

bulliform cells occur in abaxial epidermis, stomata

confined to abaxial epidermis.

Diagnostic features

Shrub or small tree; bark peeling in black discs or

flaking in small yellowish papery pieces to expose a

green underlayer; branchlets spine-tipped. Leaves

trifol iolate, glabrous, all leaflets elliptic to broad

elliptic, margins coarsely crenate-serrate, terminal

leaflet dorsiventral. Flowers hypogynous, only bisexual.

Fruit ellipsoid, very apiculate and apex curved;

putamen rugose, slimy; pseudaril red, membranous,

without distinct arms, covering almost the whole

putamen.

Widely distributed in central Transvaal, northern

Transvaal and Zululand, and occurs in savanna-

woodland. Grows in well-drained, sandy soil, usually

in warm areas with a relatively low annual rainfall.

Also recorded from Botswana, Rhodesia, Mozam-

bique, Tanzania, Kenya and Ethiopia.

Transvaal. — 2229(Waterpoort): Dongola (-BC), Hutchinson

2136; Verdoorn 2337; 8 km W. of Alldays (-CA), Van der

Watt 64; Masekwaspoort (-DD), Van der Walt 66. 2230

(Messina): near Messina (-AC), Gerstner 5716. 2231(Pafuri):

5 km W. of Klopperfontein (-AC), Codd 5437. 2328(Baltimore):

32 km W. of Blaauwberg (-BA), Mogg 24483; Blaauwberg

(-BB), Tscheuschner sub PRE 29494; 32 km N.W. of Melk-

rivier (-CD), Van der Wait 48. 2329(Pietersburg): near Vivo

(-AB), Van der Watt 11; 32 km N.N.E. of Kalkbank (-AD),

Codd 4048; near Kalkbank (-CB), Van der Walt 68. 2427

(Thabazimbi): 50 km S. of Ellisras (-BA), Van der Walt 55;

Kransberg (-CB), Prosser 1701. 2428 (Nylstroom): 5 km N.

of Nylstroom (-CB), Van der Walt 1 ; 34 km N.E. of Nylstroom

(-DA), Codd 4820. 2429(Zebediela): near Zebediela (-AD),

Pole-Evans 3112. 2430(Pilgrim's Rest): 6 km N. of Burgersfort

(-CB), Codd 6766. 2431(Acornhoek): 16 km S. of Satara Rest

Camp (-DB), Codd 4294. 2526(Zeerust): near Silkaatskop

(-AB), Van der Walt 78. 2527(Rustenburg): 21 km S. of

Northam (-AB), Van der Walt 73; near Pilanesberg (-AC),

Codd 1093; at confluence of Crocodile and Pienaars Rivers

(-BA), Codd 9338. 2528(Pretoria): Elandsberg near Rust de

Winter (-BA), Codd 6294. 2529(Witbank): Loskopdam Nature

Reserve (-AD), Codd 10363 ; Van der Walt 12; 71. 2531

(Komatipoort): 5 km N. of Malelane Rest Camp (-BC),

Codd 4369; near Louw’s Creek (-CB), Thorncroft 2265.

Natal. — 2731(Louwsburg): 8 km N.W. of Candover (-BD),

Van der Walt 105. 2732(Ubombo): near Ingwavuma (-AA),

Strey474l (NH); Jozini (-AC), Van der Walt 29. 2831(Eshowe):

Umfolozi Game Reserve (-BD), Van der Walt 85; N.E. of

Melmoth in Imfuli Valley (-CB), Acocks 12993.

C. schimperi and C. africana , two closely related

species, are often confused and this is mainly due to

the similarity of the leaves. Besides many other

differences such as in flower and fruit structure, the

two species can also be distinguished on the hairiness

of the leaves. The leaves of C. africana are pilose to

mentose, while those of C. schimperi possess only a

few glandular hairs.

A pungent resin odour is emitted when fresh leaves

are picked.

5. Commiphora africana (A. Rich.) Engl, in

A. DC., Monogr. Phan. 4: 14 (1883); Schweinf. in

Bull. Herb. Boiss. 7,2: 289 (1899); Sim, For. FI.

P.E. Afr.: 28 (1909); Engl, in Bot. Jahrb. 48: 484,

t.3N (1913); Hutch. & Dalz., FI. W. Trop. Afr. 1:

488 (1928); Engl, in Pflanzenfam. ed. 2,19a: 438

(1931); Eggeling, Trees Uganda 26 (1940); Exell

6 Mendonca in Consp. FI. Angol. 1,2: 3C0 (1951);

Wild in Bol. Soc. Brot. 2,33: 42 (1959); Dale &

Greenway, Kenya Trees 83, 1. 1 7 (1961); Irvine,

Woody Plants Ghana 510 (1961); White, For. FI.

N. Rhod. 176, t.34B (1962); Wild in FI. Zamb. 2,1:

276 (1963); Von Breitenbach, Ind. Trees S. Afr.

3,2: 437 (1965); De Wint. in Trees S. Afr. 20,1: 8

(1968); Merxm., Prod. FI. S.W. Afr. 23: 4 (1968);

Lisowski, Malaisse & Symoens in Bull. Jard. Bot.

Nat. Belg. 40: 357 (1970). Type: Senegal, Kayar,

Leprieitr s.n. (P, holo., only photo seen; isotypes!).

Heudelotia africana A. Rich, in Guill., Perr. & A. Rich.,

FI. Sen. Tent. 1: 150, t. 39 (1832).

Balsamodendrum africanum (A. Rich.) Arn. in Ann. Nat.

Hist. 3: 87 (1839); Oliv. in FI. Trop. Afr. 1: 325 (1868), pro

parte excl. syn. B. schimperi et vars. B. kotschyi O. Berg in Bot.

Ztg. 20: 162 (1862). Type: Sudan, Nubia, Kotschy 271 (B,

holo.f; SI; W!; MEL!; K!; BM!). Lectotype: Kotschy 111 (K).

Balsamea pilosa Engl, in Bot. Jahrb. 1: 41 (1881). Type:

Tanzania, Zanzibar, HUdebrandt 1184 (W, holo!). B. kotschyi

(O. Berg) Engl, in Bot. Jahrb. 1: 41 (1881).

Commiphora pilosa (Engl.) Engl, in A. DC., Monogr. Phan.

4: 12 (1883); Bot. Jahrb. 48: 488, t.3U (1913); Pflanzenfam.

ed. 2,19a: 440 (1931); Chiov., FI. Somala 2: 124, t.85 (1932);

Steedman, Trees S. Rhod. 31 (1933); Eggeling, Trees Uganda

27 (1940); Palgrave, Trees Cent. Afr. 58, t. & photo (1956).

C. loandensis Engl, in Bot. Jahrb. 26: 370 (1899). Syntypes:

Angola, Luanda, Welwitsch 4497 (BM!; LISU!); 4498 (BM!;

LISU!); 4498b (BM!; LISU!); 4500 (BM!; LISU!); 4501

(BM !; LISU!). Lectotype: Welwitsch 4497 (LISU). C. rubriflora

Engl, in Bot. Jahrb. 30: 336 (1902). Type: Tanzania, near

Rukwa Lake, Goetze 1406 (B, holo.t; K, fragment! and iso!).

C. nkolola Engl, in Bot. Jahrb. 34: 308 (1905); Bot. Jahrb. 48:

490(1913); Pflanzenfam. ed. 2,19a: 440(1931). Type: Tanzania,

Zanzibar coastal area, Basse 528 (B, holo.t; K, fragment!).

C. sambesiaca Engl, in Bot. Jahrb. 44: 146 (1910); Bot. Jahrb.

48: 490, t.3W (1913); Pflanzenfam. ed. 2,19a: 330 (1931);

Burtt Davy, FI. Transv. 2: 485 (1932). Type: Zambia,

Kazungula, Seiner 90 (B, holo.f; K, fragment!). C. calciicola

Engl, in Bot. Jahrb. 44: 147 (1910); Bot. Jahrb. 48: 490, t.3V

(1913); Pflanzenfam. ed. 2,19a: 440 (1931); Burtt Davy, FI.

Transv. 2: 485 (1932). Type: S.W. A., Grootfontein; Dinter

820 (B, holo.f; K, fragment!; BM, sketch!).

Dioecious shrub less than 1 m tall or small tree

up to 4 m tall, with a single main stem; dark grey to

green, flaking in small yellowish papery pieces to

expose a green underlayer; branchlets spine-tipped,

pilose to tomentose. Leaves trifoliolate, pilose to

tomentose, lamina up to 5 cm long; petiole up to

1,6 cm long; petiolules usually less than 2 mm long;

all leaflets obovate seldom elliptic, apex obtuse rarely

acute, base cuneate, margins coarsely crenate-serrate

or finely lobed; terminal leaflet up to 3,3x3 cm;

lateral leaflets up to 1, 4x1,1 cm. Flowers unisexual,

hypogynous, appearing before or with the leaves in

axillary clusters, often borne on the spines; male

flowers, 6-8 mm, in most cases larger than female

flowers, 5-7 mm. Bracteoles up to 1 mm long,

triangular, pilose to pubescent. Pedicels up to 2,5 mm

long. Calyx campanulate, green to red, 2, 5-4,5 mm

long, lobes less than 1 mm long, apex acute. Petals

green to red, 4, 5-5, 5 mm long, glabrous. Disk

cylindrical not adnate to calyx or corolla, with 4

prominent lobes, indentation between lobes very deep,

upper part of lobes grooved on outside, lobes in male

flowers fleshy but not bifid, in female flowers less

fleshy and bifid. Stamens 8, 4 long stamens up to

5,5 mm long, inserted high up on the outside of disk

lobes; 4 short stamens up to 3 mm, inserted on the

outside of disk between the lobes; filaments subterete,

lower part flattened and broadened and placed in

grooves on outside of disk lobes; staminodes in female

flowers. Gynoecium: rudimentary in male flowers;

ovary superior; style relative long; stigma obscurely

4-lobed. Fruit 1,5 x 1,2 cm, subglobose, slightly

flattened, asymmetrical; exocarp glabrous; mesocarp

J. J. A. VAN DER WALT

69

fleshy; putamen 1, Ox 0,8 cm, subglobose to ellipsoid,

asymmetrically and irregularly flattened, very rugose;

pseudaril red, fleshy, with 4 arms of variable size and

form and often also isolated fragments, 2 arms on

seam of putamen reaching almost to apex, 2 arms

on flattened faces of putamen of variable length, in

some cases arms not distinct and pseudaril covering

almost whole putamen. Fig. 27-32.

Fig. 27. — Commiphora afri-

cana near Dendron,

northern Transvaal (height

± 1 m).

Fig. 28. — Close-up view of a branch of Commiphora africana

illustrating the bark flaking in small papery pieces.

Fig. 29.— Commiphora africana: A, branchlet with flowers;

B, C, branch lets with leaves; D, branchlet with leaves and

mature fruits.

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 30. — Flowers of Commiphora africana: A, male flowers;

B, longitudinal section of male flower; C, diagrammatic

representation of two disc lobes illustrating the insertion

of the filaments; D, female flower; E, female flower with

the calyx and corolla partly removed.

Fig. 31. — Fruit of Commiphora africana: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Fig. 32. — Geographical distribution of Commiphora africana

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with a variable number of multi-

cellular non-glandular and glandular hairs. Stems

of 2,5 cm diameter: sclerenchymatous pericycle con-

sisting of separate fibre strands but stone cells absent,

epithelium cells of resin ducts in xylem rays sur-

rounded by thin-walled cells. Leaves with a variable

number of multicellular non-glandular and glandular

hairs; petiole ± triangular as seen in transverse

section, sclerenchymatous pericycle present, vascular

bundles ± circularly distributed as seen in transverse

section; terminal leaflet dorsiventral with a single

layer of palisade cells adaxially, remaining mesophyll

consisting of spongy or palisade-like cells, buliform

cells confined to adaxial epidermis, stomata mainly

abaxial.

Diagnostic features

Dioecious shrub or small tree; bark grey to green,

flaking in small yellowish papery pieces to expose

a green underlayer; branchlets spine-tipped, pilose

to tomentose. Leaves trifoliolate, pilose to tomentose;

all leaflets obovate, seldom elliptic, margins coarsely

crenate-serrate or finely lobed, terminal leaflet

dorsiventral. Flowers hypogynous, unisexual. Fruit

subglobose; putamen rugose; pseudaril red, fleshy,

with 4 arms of variable size and form and often also

isolated fragments.

Distributed in northern Zululand, north-western,

far-northern, north-eastern and eastern Transvaal,

but is particularly common north of the Soutpansberg.

Usually grows in sandy, well-drained soil in areas

with a relatively low annual rainfall. Occurs in

shrub-thornveld, savanna-woodland or in broken

mopaniveld.

Also recorded from Swaziland, the northern and

central parts of South West Africa, Botswana,

Rhodesia, Mozambique, Zambia, Malawi, Angola,

Tanzania, Kenya, Ethiopia, Uganda, Sudan, Gambia,

Senegal, Nigeria, Mauritania, Mali, Ghana, Togo,

Niger, Zaire and Rwanda.

Transvaal. — 2228(Maasstroom): 15 km S.W. of Swartwater

(-CC), Meeuse 10571. 2229(Waterpoort): Dongola (-BB),

Verdoorn 2329; N. of Soutpan (-CC), Ohermeyer , Schweickerdt

& Verdoorn 180. 2230( Messina): 30 km N.E. of Tshipise

(- AD), Van der Schijff 5237; Tshipise (-CA), Van der Schijff

J. J. A. VAN DER WALT

71

5201; Van der Walt 9. 2231(Pafuri): near Punda Milia Rest

Camp (-CA), Van der Schijff921. 2327(ElIisras): 20 km S. of

Ellisras (-DC), Van der Walt 56; 16 km S.S.E. of Ellisras

(-DD), Van der Walt 118. 2329(Pietersburg): near Dendron

(-AD), Codd 4436; Van der Walt 67; Verdoorn 2317b. 2330

(Tzaneen): Hans Merensky Nature Reserve (-CA), Gilliland

790. 2331(Phalaborwa) : 5 km S. of Shingwidzi Rest Camp

(-AB), Van der Walt 75; near Gorge Rest Camp (-DD),

Van der Schijff 3495. 2427(Thabazimbi): 50 km S. of Ellisras

(-BA), Van der Walt 23. 2430( Pilgrim's Rest): Abel Erasmus

Pass (-BD), Kil/ick & Strey 2527. 2431(Acornhoek): near

Skukuza Rest Camp (-DC), Van der Schijff 449; near Tshok-

wane (-DD), Van der Schijff 912. 2531(Komatipoort): near

Komatipoort (-DD), Van der Schijff 893.

Natal. — 2632(Bela Vista): Ndumo Game Reserve (-CC),

Van der Walt 30; 97. 2731(Louwsburg): 6 km N.W. of Candover

(-BD), Van der Walt 104; 106. 2831(Eshowe): Umfolozi

Game Reserve (-BD), Van der Walt 86.

Richard (1832) described Heudelotia africana as

the type species of this genus. In his description it is

mentioned that the flowers are bisexual. An illustration

of Dale & Greenway (1961) also indicates that the

flower of C. africana is bisexual. The staminodes of

the female flowers investigated, although relatively

long, are always sterile.

Wild (1963) distinghuishes the var. africana and

var. rubriflora (Engl.) Wild. The calyx and pedicels

of the var. rubriflora are hairy, while those of the var.

africana are glabrous. As far as could be determined,

only var. africana occurs in South Africa.

Wild (1963) mentions that the pseudaril of C.

africana is apparently absent. However, all the fruits

of this species studied possess a fleshy pseudaril.

According to Irvine (1961) the gumresin is used by

the natives for perfuming and fumigating huts. He

also mentions that it has several medicinal uses, and

it is also used as a varnish. The species is easily grown

from pole cuttings which are often planted as fencing

poles. Common name: African bdellium.

6. Commiphora neglecta Verdoorn in Bothalia

6,1: 214 (1951); Codd, Mem. Bot. Surv. S. Afr.

26: 88 (1951); Wild in Bol. Soc. Brot. 2,33: 86 (1959);

Wild in FI. Zamb. 2,1: 271 (1963); Von Breitenbach,

Ind. Trees S. Afr. 3,2: 438 (1965); Moll, For. Trees

Natal 80 (1967); De Wint. in Trees S. Afr. 20,1: 16

(1968). Type: Transvaal, Skukuza, Codd & Verdoorn

5498 (PRE, holo. !).

Polygamous or dioecious many-stemmed shrub or

small tree with a single main stem up to 8 m tall;

bark grey to green, smooth or flaking in small yel-

lowish papery pieces; branchlets spine-tipped and

with a few short hairs. Leaves trifoliolate, with a few

short hairs; lamina up to 7 cm long; petiole up to

4,5 cm long; petiolules usually less than 1 mm long;

leaflets elliptic or ovate to broadly ovate, apex acute,

base cuneate, margins entire or upper half finely

crenate-serrate; terminal leaflet up to 4,4 x 2,8 cm,

lateral leaflets up to 3 x 2,2 cm. Flowers bisexual or

unisexual but male flowers rare, hypogynous, ap-

pearing before or with the leaves in axillary dichasial

cymes up to 1,2 cm long or in clusters, usually on

spines; peduncles usually with a few short hairs;

bisexual and male flowers, 7-8 mm, usually slightly

larger than female flowers, 6-7 mm. Bracteoles up

to 3,5 mm long, lanceolate, with variable number of

short hairs. Pedicels 2-5 mm long, often with a few

short hairs. Calyx campanulate, yellow to green,

2-3 mm long, often with a few short hairs, lobes up to

1 mm long, apex acute to obtuse. Petals yellow to

green, 3-5 mm long, glabrous. Disk fleshy, not adnate

to calyx or corolla, cylindrical with 4 lobes, indentation

between lobes shallow, lobes of bisexual and male

flowers not bifid but in female flowers bifid. Stamens

8, 4 long stamens up to 4 mm long, inserted high

up on the outside of disk lobes, 4 short stamens up

to 2,8 mm long, inserted on the outside of disk

between lobes; filaments subterete, lower part flat-

tened and broadened; staminodes in female flowers.

Gynoecium: rudimentary in male flowers; ovary

superior; style relatively short; stigma obscurely

4-lobed. Fruit 1,5 X 1,4 cm, subglobose, slightly

Fig. 33. — Commiphora neg-

lecta in the Umfolozi Game

Reserve, Zululand (height

+ 3 m).

72

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

flattened and asymmetrical; exocarp glabrous; meso-

carp fleshy; putamen 8x7 mm, ellipsoid, asymmetri-

cally and irregularly flattened, smooth; pseudaril red,

very fleshy, with 4 arms. 2 arms on seam of putamen

reaching almost to apex. 2 arms on flattened faces

of putamen shorter and of equal length, but arm on

more convex face usually broader than arm on other

face. Fig. 33-38.

Fig. 34.— Close-up view of a branch of Commiphora neglecta

illustrating the bark flaking in papery pieces.

Fig. 36.— Flowers of Commiphora neglecta: A, bisexual flower

B, bisexual flower with the calyx and corolla partly

removed; C, diagrammatic representation of two disc

lobes illustrating the insertion of the filaments; D, female

flower; E, female flower with the calyx and corolla

partly removed.

Fig. 37. — Fruit of Commiphora neglecta: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

J. J. A. VAN DER WALT

73

Fig. 38. — Geographical distribution of Commiphora neglecta

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with a few short non-glandular hairs

(mostly unicellular) and glandular hairs. Stems of

2,5 cm diameter: sclerenchymatous pericycle-cylinder

consisting of fibres and stone cells, epithelium cells of

resin ducts in xylem rays surrounded by 1 layer of

cells with slightly thickened walls. Leaves with a few

short non-glandular hairs (mostly unicellular) and

glandular hairs; petiole ± triangular as seen in

transverse section, sclerenchymatous pericycle pre-

sent, vascular bundles ± circularly distributed as

seen in transverse section; terminal leaflet isobilateral

with a single layer of long palisade cells adaxially

and a single layer of shorter palisade cells abaxially,

remaining mesophyll consisting of spongy or palisade-

like cells, adaxial epidermis consisting mainly of

large bulliform cells but smaller bulliform cells

occur in adaxial epidermis, stomata mainly abaxial.

Diagnostic features

Polygamous or dioecious many-stemmed shrub or

small tree; bark grey to green, smooth or flaking in

small yellowish papery pieces; branchlets spine-tipped,

with a few short hairs. Leaves trifoliolate, with a few

short hairs, margins of leaflets entire or upper half

finely crenate-serrate, terminal leaflet isobilateral.

Flowers hypogynous, bisexual or unisexual but male

flowers rare. Fruit subglobose; exocarp glabrous;

putamen smooth; pseudaril red, very fleshy, with 4

arms, 2 arms on seam of putamen reaching almost

to apex, 2 arms on flattened faces of putamen

shorter.

Thisspecies occursin central and northernTransvaal’

is widely distributed in Natal and particularly common

in northern Zululand. It usually occurs on the slopes

of mountains or in sandy, well-drained soil in areas

with an annual rainfall of 500-700 m per year.

Also recorded from Mozambique.

Transvaal. — 2231(Pafuri): near Punda Milia Rest Camp

(-CA), Lang 31083. 2327(Ellisras): 16 km S. of Ellisras (-DD),

Codd 1013. 2428(Nylstroom): 5 km N.E. of Nylstroom (-CB),

Smuts & Gillett 3340; Van der Walt 2; 34 km N.E. of Nylstroom

(-DA), Codd 4821; Verdoorn 2362. 2431(Acornhoek): 24 km

E. of Skukuza (-DD), Codd & Verdoorn 5498 (PRE); Codd

4396; Codd & de Winter 5052. 2527(Rustenburg): 21 km S.

of Northam (-AB), Van der Walt 72. 2529(Witbank): Loskop-

dam Nature Reserve (-AD), Codd 10362; Van der Walt 17.

2531(Komatipoort): Klokwene (-AD), Van der Schijff 337;

14 km S.E. of Pretoriuskop Rest Camp (-AD), Codd & de

Winter 5166; 2 km W. of Skipberg (-AD), Van der Walt 69;

on Lower Sabie road (-BB), Van der Schijff 3268.

Natal. — 2632(Bela Vista): Ndumo Game Reserve (-CC),

Gerstner 3452 (NH); 45 km N.E. of Jozini, Van der Walt 98.

2723(Ubombo): Ingwavuma (-AA), Vahrmeijer 812; Man-

zengwena Forest Station near Ingwavuma (-AA), de Winter

& Vahrmeijer 8515; 40 km N.E. of Ubombo (-AB), Van der

Walt 103 ; Lalanek (-BA), Vahrmeijer 467; Mkuzi Game

Reserve (-CA), Ward 3570 (NH); near False Bay (-CD),

Gerstner 6767; Ward 1639 (NH); 2829(Harrismith): Tugela

Valley near Colenso ( DB), Edwards 2465. 2830(Dundee):

Tugela Ferry (-CD), Edwards 1256. 2831(Eshowe): near

Mahlabatini (-AB), Gerstner 2222; Umfolozi Game Reserve

(-BD), Van der Walt 87; 90; 38 km N.E. of Heatonville (-BD),

Van der Walt 82; near Empelengeni bridge (-CB), Lawn 1697

(NH). 2832(Mtubatuba): Hluhluwe Game Reserve (-AA),

Ward 1750. 2930( Pietermaritzburg): near Middledrift (-AB),

Edwards 2083; near Umgeni Dam (-DB), Bayer 729.

As Verdoorn (1951) mentions, this species differs

in the particular combination of characters rather

than in any outstanding characteristics. This is

probably the reason why it was described in 1951 for

the first time.

Although plants in the veld only develop trifolio-

late leaves, it has been observed that leaves of pole

cuttings can be impari-pinnate with two pairs of lateral

leaflets.

7. Commiphora mollis (Oliv.) Engl, in A. DC.,

Monogr. Phan. 4: 23 (1883); Bot Jahrb. 48: 472

(1913); Pflanzenfam. ed. 2,19a: 435 (1931); Burtt

Davy, FI. Transv. 2: 484 (1932); Burtt in Kew Bull.

1935: 111 (1935); Brenan in Kew Bull. 1950: 367

(1950); Exell & Mendonca in Consp. FI. Angol.

1,2: 298 (1951); Codd, Mem. Bot. Surv. S. Afr. 26:

87 (1951); Wild in Bol. Soc. Brot. 2,33: 85 (1959);

FI. Zamb. 2,1: 273 (1963); Von Breitenbach, Ind.

Trees S. Afr. 3,2: 435 (1965); De Wint., M. de Wint.

& Killick, Transv. Trees 90 (1966); De Wint. in

Trees S. Afr. 20,1: 16 (1968); Merxm., Prod. FI.

S.W. Afr. 23: 7 (1968). Type: Mozambique, Chi-

ramba, between Tete and coast, Kirk s.n. (K, holo.l).

Balsamodendntm mode Oliv. in FI. Trop. Afr. I : 326(1868).

Balsamea mollis (Oliv.) Engl, in Bot. Jahrb. I : 42 (1881).

Commiphora welwitschii Engl, in A. DC., Monogr. Phan. 4:

22 (1883); Bot. Jahrb. 48: 473 (1913); Pflanzenfam. ed. 2,19a:

435 (1931); Miller in J.S. Afr. Bot. 18: 39 (1952). Type: Angola,

Huila, Welwitsch 4493 (G, holo., only photo seen: LISCJ !).

C. cinerea Engl, in Bot. Jahrb. 19: 139 (1895); Bot. Jahrb. 48

473, t. IR (1913); Pflanzenfam. ed. 2,19a: 435 (1931). Type:

S.W. A., Otjitambi, Gurich 21 (B, holo.t; K, fragment!).

C. stuhlmanii Engl., Pflanzenw. O. Afr. C: 230 (1895); Bot.

Jahrb. 48: 474 (1913); Pflanzenfam. ed. 2,19a: 435 (1931);

Burtt in Kew Bull. 1935 : 1 1 1 (1935). Type: Tanzania, Bukombe,

Stuhlman 3450 (B, holo.t; K, fragment!; BM, sketch!). C.

dekindtiana Engl, in Bot. Jahrb. 34: 312 (1905); Bot. Jahrb.

48: 473, t. IQ (1913); Pflanzenfam. ed. 2,19a: 435 (1931).

Type: Angola, Eluila, Benguela, Dekindt 225 (B, holo.f; K,

fragment!) C. heterophylla Engl, in Bot. Jahrb. 34: 312 (1905);

Bot. Jahrb. 48: 475 (1913); Pflanzenfam. ed. 2,19a: 435 (1931).

Type: Tanzania, Kilimandjaro area between Taveta and Bura,

Engler 1906 (B, holo.t; K, fragment!). C. montana Engl, in

Bot. Jahrb. 34: 312 (1905); Bot. Jahrb. 48: 473 (1913); Pflan-

zenfam. 2,19a: 435 (1931). Type: Angola, Huila, Benguela,

Dekindt 46 (B, holo.f; K, fragment!). C. krausei Engl, in Bot.

Jahrb. 44: 152 (1910); Bot. Jahrb. 48: 472, t. IP (1913);

Pflanzenfam. ed. 2,19a: 435 (1931). Type: Tanzania, Tabora,

van Trotha 8a (B, holo.f; K, fragment!). C. iringensis Engl,

in Bot. Jahrb. 44: 150 (1910); Bot. Jahrb. 48: 472, t. 10 (1913);

Pflanzenfam. ed. 2, 19a: 435 (1931). Type: Tanzania, Uhehe,

Spiegel sub Amani Herbarium 2507 (B, holo.f; EA!). C.

boehmii Engl, in Bot. Jahrb. 48: 472 (1913); Pflanzenfam. ed.;

2,19a: 435 (1931). Syntypes: Tanzania, Gonda, Bohm 281

(Bf; K, fragment!). Salanda, Fischer 292 (B+; K, fragment!);

74

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Lectotype: Bohm 281 (K, fragment). C. ndemfi Engl, in Bot.

Jahrb. 54: 293 (1917); Pflanzenfam. ed. 2,19a: 435 (1931).

Type: Tanzania, Urambo, Stolz 1678 (B, holo.f; K!; Z ! ;

P. !). Lectotype: Stolz 1678 (K).

Dioecious tree with a rounded spreading crown up

to 8 m tall, trunk often irregularly fluted or with

large knobbly outgrowths; bark brown to greyish-

green, peeling in discs with an average diameter of

2,5 cm; branchlets pilose to pubescent. Leaves impari-

pinnate but in some cases trifoliolate: grey to green,

pilose to tomentose, often distinctly paler and

tomentose below; lamina up to 10 cm long; petiole

up to 2,5 cm; petiolules up to 3 mm long; leaflets

1-6-jugate, oblong-elliptic to elliptic or obovate, apex

acute to obtuse rarely emarginate, base cuneate to

broadly cuneate, margins entire; terminal leaflet up

to 5,2 x 3,6 cm; lateral leaflets up to 3, 2x1,7 cm.

Flowers unisexual, hypogynous, appearing before or

with the leaves, male inflorescences usually axillary

compound dichasial cymes up to 4 cm long, female

inflorescenes usually axillary simple dichasial cymes

up to 2 cm long, branches of inflorescences pilose to

pubescent; male flowers, 8-12 mm, usually larger than

female flowers, 4-7 mm. Bracteoles up to 5 mm long.

linear, pilose to pubescent. Pedicels 3-8 mm long,

pilose to pubescent. Calyx campanulate, green to

red, 2-3 mm long, pilose to pubescent, lobes 1-2 mm

long, apex acute. Petals green to red, 3-6 mm long,

pilose to pubescent on outside. Disk fleshy, not

adnate to calyx or corolla, cylindrical with 4 prominent

lobes, lobes bifid, indentation between lobes deep;

disk in female flowers smaller than in male flowers.

Stamens 8, 4 long stamens up to 5 mm long, inserted

high up on outside of disk lobes, 4 short stamens up

to 2,5 mm long, inserted on outside of disk

between lobes; filaments subterete, lower part

flattened and broadened; staminodes in female

flowers. Gynoecium rudimentary in male flowers;

ovary superior; style relatively short; stigma obscurely

4-lobed. Fruit 1,3 x 1,2 cm, subglobose, slightly flat-

tened, asymmetrical; exocarp pilose to pubescent;

mesocarp fleshy; putamen 1,0x0, 8 cm, ellipsoid,

asymmetrically and irregularly flattened, smooth;

pseudaril bright red, fleshy, with 4 winged arms, 2

arms on seam of putamen reaching almost to apex,

2 arms on flattened faces shorter than arms on seam,

arm on more convex face of putamen longer than

arm on other face. Fig. 39-44.

J. J. A. VAN DER WALT

75

Fig. 40. — Close-up view of the trunk of Commiphora mollis

illustrating the knobbly outgrowths and the bark peeling in

thick discs.

Fig. 41. — Commiphora mollis: A, branchlet with young leaves

and flowers; B — D, leaves; E, branchlet with leaves and

mature fruits.

Fig. 42. — Flowers of Commiphora mollis: A, male flower; B,

longitudinal section of male flower; C, diagrammatic

representation of two disc lobes illustrating the insertion

of the filaments; D, female flower; E, female flower with

the calyx and corolla partly removed.

Fig. 43. — Fruit of Commiphora mollis: A, side-view of the fruit;

B, view of the less convex face of putamen with pseudaril;

C, view of the more convex face of putamen with pseudaril;

D, putamen and pseudaril as seen from above.

76

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 44. — Geographical distribution of Commiphora mollis in

South Africa.

Distinctive features of the stems and leaves

Young stems with a variable number but usually

many multicellular non-glandular and glandular hairs,

hypodermis consisting mainly of secretory cells. Stems

of 2,5 cm diameter: sclerenchymatous pericycle-

cylinder consisting of fibres and stone cells, epithelium

cells of resin ducts in xylem rays surrounded by 1

layer of cells with slightly thickened walls. Leaves

with a variable number but usually many multicellular

non-glandular and glandular hairs; petiole db triangu-

lar as seen in transverse section, sclerenchymatous

pericycle present, vascular bundles ± circularly

distributed as seen in transverse section; terminal

leaflet dorsiventral with a single layer of palisade cells

adaxially, remaining mesophyll consisting of spongy

or palisade-like cells, bulliform cells confined to

adaxial epidermis, stomata mainly abaxial.

Diagnostic features

Dioecious tree, trunk often irregularly fluted; bark

brown to greyish-green, peeling in discs with an

average diameter of 2,5 cm; branchlets pubescent

to pilose. Leaves impari-pinnate but sometimes tri-

foliolate, pilose to tomentose, leaflets 1-6-jugate,

often distinctly paler and tomentose below, margins

entire, terminal leaflet dorsiventral. Flowers red,

hypogynous, unisexual; pedicels, calyx and corolla

pilose to pubescent. Fruit subglobose; exocarp pilose

to pubescent; putamen smooth; pseudaril bright

red, fleshy, with 4 winged arms, 2 arms on seam of

putamen reaching almost to apex, 2 arms on flattened

faces of putamen shorter.

Widely distributed in north-western, central-

northern, far-northern and north-eastern Transvaal.

Grows in savanna-woodland on stony hills or in

well-drained, sandy soil. Occurs in warm areas with

a relatively low annual rainfall.

Also recorded from South West Africa, Botswana,

Rhodesia, Zambia, Malawi, Tanzania, Angola and

Zaire.

Transvaal. — 2228(Maasstroom): 14 km N.E. of Koper-

spruit (-CA), Van der Walt 62\ 14 km S.W. of Swartwater

(-CC), Meettse 10573 ; 2229(Waterpoort): Dongola (-BA),

Code I 4114; Verdoorn 2269; Soutpan (-CD), Obermeyer,

Schweickerdt & Verdoorn 152; 8 km E. of Waterpoort (-DC),

Van der Walt 21 ; 5 km N. of Wylliespoort (-DD) Codd 4445;

Masekwaspoort (-DD), Hardy 2240. 2230( Messina): near

Messina (-AC), Gerstner 5713; Tshipise (-CA), Van der

Schijff 5202; Van der Walt 7. 223 1 (Paftiri) : Punda Milia Rest

Camp(-CA), Codd & Dyer 4612; Van der Walt 77; 19 km S.E.

of Punda Milia Rest Camp (-CC), Codd & Dyer 4589. 2328

(Baltimore): Marnitz (-AA), Van der Walt 59; IS. of Blaauwberg

(-BB), Obermeyer , Schweickerdt & Verdoorn 95; 32 km N.W.

of Melkrivier (-CD), Van der Walt 45; near Sterkwater (-DD),

Codd 657 1 ; Van der Walt 50. 2329(Pietersburg) : 13 km S.S.E.

of Dendron (-AB), Van der Walt 18; 5 km S. of Bandelierkop

(-BD), Gerstner 5575. 2330(Tzaneen): Modjadji Nature

Reserve (-DA), Krige 7. 2427(Thabazimbi: Sentrum (-AD),

Vahrineijer 1370 ; 50 km S. of Ellisras (-BA), Van der Walt 53;

32 km N.E. of Thabazimbi (-BC), Codd 3726. 2429(ZebedieIa):

Chuniespoort (-AD), De Winter 2328. 2430(Pilgrim's Rest):

27 km S.E. of Gravelotte (-BB), Codd & de Winter 3710;

Abel Erasmus Pass (-BD), Strey 3636; 8 km N. of Buffelsvlei

(-CD), Codd 6677. 2431(Acornhoek): Skukuza Rest Camp

(-DC), Letty 58; Van der Schijff 745; 6 km S. of Skukuza

Rest Camp, Van der Schijff 867. 2528(Pretoria): 1 km N. of

Pienaarsrivier (-AB), Codd 3709 A. 2513(Komatipoort):

Numbi (-AA), Van der Schijff 3439; Krododilpoort near

Baberton (-CB), Galpin 1077.

The variation in hairiness of the young stems and

leaves, as well as the variation in the form and size

of the leaflets, can account for the many synonyms.

This species is easily grown from pole cuttings

which are often planted as fencing poles. The young

branches are grazed by cattle and game.

Common name: “Ghor-Ghor”.

8. Commiphora harveyi {Engl.) Engl, in A. DC.,

Monogr. Phan. 4: 25 (1883); Bot. Jahrb. 48: 476,

t. 2J (1913); Pflanzenfam. ed. 2,19a: 435 (1931);

Burtt Davy, FI. Transv. 2: 484 (1932); Henkel,

Woody PI. Natal 213 (1934); Wild in Bol. Soc. Brot.

2,33: 85 (1959); Von Breitenbach, Ind. Trees S. Afr.

3,2: 442 (1965); Moll, For. Trees Natal 79 (1967); De

Wint. in Trees S. Afr. 20,1: 12 (1968). Type: Natal

Durban, Gerrard & McKen 689 (TCD, holo. !; K! ).

Protium africanum Harv. in FI. Cap. 2: 592 (1862); Swart,

Monog. Protium 393 (1942).

Balsamea harveyi Engl, in Bot. Jahrb. 1: 42 (1881).

Dioecious tree from 4 m up to 18 m tall; bark

peeling in large brown papery pieces or in thicker

discs; branchlets fluted and with a few short hairs.

Leaves impari-pinnate, or occasionally trifoliolate,

with a few short hairs; lamina up to 15 cm long;

petiole up to 6 cm long; petiolules up to 1,5 cm long;

leaflets 1-3-jugate, lanceolate to elliptic or ovate,

apex acute, base cuneate, margins crenate-serrate to

coarsely crenate-serrate; terminal leaflet up to

8x3 cm; lateral leaflets up to 6x2,5 cm. Flowers

unisexual, hypogynous, appearing after the leaves in

axillary paniculate cymes up to 10 cm long, male

inflorenscences usually longer than female inflore-

scences, peduncles with a few short hairs and

conspicious leaf-like bracts up to 6 mm;

male flowers, 5-7 mm, usually larger than female

flowers, 4-5,5 mm. Bracteoles linear, up to 3 mm

long, with a few short hairs. Pedicels 2-3 mm long,

usually with a few short hairs. Calyx campanulate,

yellowish green, 2-3 mm long, sometimes with a few

short hairs, lobes 1-1,8 mm long, apex acute. Petals

yellowish green, 2,5-4 mm long, without hairs. Disk

fleshy, not adnate to calyx and corolla, cylindrical,

with 4 lobes, indentation between lobes not very

deep, lobes in male flowers not bifid but in female

flowers bifid. Stamens 8, 4 long stamens up to 4 mm

long, inserted high up on outside of disk lobes, 4

short stamens up to 2,5 mm long, inserted on outside

of disk between lobes; filaments slightly flattened,

lower part broadened; staminodes in female flowers.

J. J. A. VAN DER WALT

77

Gynoecium : rudimentary in male flowers; ovary

superior; style relatively long; stigma 2-lobed. Fruit

1,4 x 1,2 cm, subglobose, slightly flattened, asymmetri-

cal; exocarp glabrous; mesocarp fleshy; putamen

0,9 X 0,7 cm, ellipsoid to obovate, asymmetrically and

irregularly flattened, smooth; pseudaril light red,

very fleshy, with 4 arms, 2 arms on seam of putamen

reaching almost to apex, 2 arms on flattened faces of

putamen variable in length and breadth but shorter

than arms on seam, arm on more convex face of

putamen usually shorter but broader than arm on

other face. Fig. 45-50.

Fig. 45. — Commiphora harveyi near Malelane in the Kruger

National Park (height ±6 m).

Fig. 46. — Close-up view of a stem of Commiphora harveyi

illustrating the bark peeling off in thick pieces.

5292-6

Fig. 47. — Commiphora harveyi: A, branchlet with a leaf and

flowers; B, branchlet with a leaf and mature fruits.

Fig. 48. — Flowers of Commiphora harveyi: A, male flowers;

B, longitudinal section of male flower; C, diagrammatic

representation of two disc lobes illustrating the insertion

of the filaments; D, female flower; E, female flower with

the calyx and corolla partly removed.

78

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 49. — Fruit of Commiphora harveyi: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

section, sclerenchymatous pericycle present, vascular

bundles ± circularly distributed as seen in transverse

section; terminal leaflet dorsiventral with a single

layer of palisade cells adaxially, remaining mesophyll

consisting of spongy parenchyma, bulliform cells

confined to adaxial epidermis, stomata mainly

abaxial.

Diagnostic features

Dioecious tree, bark peeling in large brown papery

pieces or in thicker discs; branchlets fluted, with a

few short hairs. Leaves impari-pinnate or occasionally

trifoliolate, with a few short hairs, leaflets 1-3-jugate,

margins crenate-serrate to coarsely crenate-serrate,

terminal leaflet typically dorsiventral. Flowers hypogy-

nous, unisexual, peduncles with large bracts. Fruit

subglobose; exocarp glabrous; putamen smooth;

pseudaril light red, very fleshy, with 4 arms, 2 arms

on seam of putamen reaching almost to apex, 2 arms

on flattened faces of putamen shorter.

Occurs in north-eastern Transvaal, eastern Trans-

vaal, the Transkei and eastern Cape as far south as

East London, but is widely distributed in Natal and

Zululand.

Usually grows on the slopes of mountains or in

kloofs as part of the coastal forests. Occurs in areas

with a rainfall up to 1 000 mm or more per annum.

Also recorded from Swaziland and Mozambique.

Transvaal. — 2430( Pilgrim's Rest): 6 km N. of Brand-

draai (-DA), Codd & de Winter 3259; Mariepskop (-DB),

Van der Schijff 6095 (PRU) 2531(Komatipoort): near Malelane

RestCamp (-AD), Codd 5263; Van der Walt 22; 14 km S.E. of

Pretoriuskop Rest Camp (-AD), Codd & De Winter 5156;

Van der Walt 70.

Natal. — 2829(Harrismith): Tugela Valley near Colenso

(-DB), Edwards 2467; 6 km N.E. of Estcourt (-DD), Cheadle

611. 2830(Dundee): near Muden (-CD), Edwards 2772. 2831

(Eshowe): Hlabisa (-BB), Gerstner 715; 49 km S.E. of Nongoma

(-BB), Van der Walt 107; 32 km E. of Melmoth (-DA), Van

der Walt 95; Ngoya forest at Ntimona (-DB), Venter 4228;

Umhlatuzana (-DD), Venter 3824; Umhlatuzi Valley at

Ntimona (-DD), Van der Walt 27. 2832(Mtubatuba): Hluhluwe

Game Reserve (-AA), Bayer 1455; Ward 2907 (NH). 2930

(Pietermaritzburg): Richmond (-CC), Moll & Morris 663;

Camperdown (-DA), Morris 593; Isipingo (-DD), Ward 848.

2931(Stanger): Mapumulo (-AA), Mott 1628; New Guelderland

(-AD), Stewart 121 (NH); Burman Bush in Durban (-CC),

Ross & Moll 1723 (NH). 3030(Port Shepstone): Ramsgate

( — CD), Nicholson 136 (NH).

Cape. — 3227(Stutterheim): Komga (-DB), Schlechter 1403.

3228(Butterworth): near Kentani (-CA), Pegler 2136. 3327

(Peddie): Keiskama Valley between Peddie and East London

(-AB), Dyer 4539; Elizabeth Island in East London (-BB),

Galpinl.

This species is easily grown from pole cuttings

which are often planted as fencing poles.

Common name: Paper Tree.

*

::

J

:

■f

'I

:

I'!

:

4

:

:I

■M

:;!i

S

it

■

i

a

Fig. 50. — Geographical distribution of Commiphora harveyi

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with a few short non-glandular hairs

(mostly unicellular) and glandular hairs; scleren-

chymatous pericycle-cylinder fluted. Stems of 2,5 cm

diameter: sclerenchymatous pericycle-cylinder con-

sisting of fibres and stone cells but in some stems

already cut off by the development of periderm;

epithelium cells of resin ducts in xylem rays sur-

rounded by 1-2 layers of sclereids. Leaves with a

few short non-glandular hairs (mostly unicellular) and

glandular hairs; petiole ± ovate as seen in transverse

9. Commiphora marlothii Engl, in Bot. Jahrb.

44: 155 (1910); Bot. Jahrb. 48: 485 (1913); Pflanzen- j

fam. ed. 2,19a: 438 (1931); Miller in J. S. Afr. Bot.

18: 38 (1952); Palgrave, Trees Cent. Afr. 55, t. &

photo (1956); Wild in FI. Zamb. 2,1: 281 (1963);!

Von Breitenbach, Ind. Trees S. Afr. 3,2: 440 (1965);;

De Wint. in Trees S. Afr. 20,1: 12 (1968); Lisowski,

Malaisse & Symoens in Bull. Jard. Bot. Nat. Belg.

40: 360 (1970). Syntypes: Rhodesia, Matopos,

Marloth 3397 (B|; K, fragment!; PRE!); 3402 (Bf).

Lectotype: Marloth 3397 (PRE).

Dioecious tree with a rounded crown up to 9 m

tall; bark peeling in large yellowish papery pieces to

expose a green underlayer; branchlets obtuse, densely

J. J. A. VAN DER WALT

79

irregularly flattened, slightly rugose; pseudaril yellow,

fleshy, with 4 arms, 2 arms on seam of putamen

reaching almost to apex, 2 arms on flattened faces

of putamen shorter than arms on seam, arm on more

convex face of putamen shorter than arm on other

face. Fig. 51-56.

Fig. 52. — Close-up view of a branch of Commiphora marlothii

illustrating the bark peeling off in large, papery pieces.

Fig. 53. — Commiphora marlothii: A, leaf; B, branchlet with

young leaves and flowers; C, branchlet with a leaf and

mature fruits.

pilose to pubescent. Leaves impari-pinnate, dark

green, pubescent to tomentose; lamina up to 26 cm

long; petiole up to 9,5 cm long; petiolules up to 2 mm

long; leaflets 3-5-jugate, obovate to broadly elliptic,

apex obtuse to acute, base cuneate or rounded,

margins crenate-serrate to finely lobed; terminal

leaflet up to 8x4 cm; lateral leaflets up to 5,8 X 3,2

cm. Flowers unisexual, hypogynous, appearing with

the leaves in axillary paniculate simple or compound

dichasial cymes up to 10 cm long, peduncles densely

pilose to pubescent, male inflorescences usually

longer than female inflorescences; male flowers, 6-7

mm, usually larger than female flowers, 4, 5-5, 5 mm.

Bracteoles up to 6 mm long, linear, pubescent. Pedicels

usually less than 1 mm, clustered, pubescent. Calyx

campanulate, yellowish green, up to 3 mm long,

pubescent, lobes up to 1 mm long, apex acute. Petals

yellowish green, 3-4 mm long, pilose outside. Disk

fleshy, pilose, not adnate to calyx and corolla, cylin-

drical, with 4 prominent lobes, indentation between

lobes rather shallow, lobes not bifid in male flowers

but bifid in female flowers. Stamens 8, 4 long stamens

! up to 3,5 mm long, inserted on top of disk lobes, 4

short stamens up to 2,5 mm long, inserted on top of

disk between lobes; filaments slightly flattened, lower

,part broadened; staminodes in female flowers.

Gynoecium: rudimentary in male flowers; ovary

superior, sparsely pilose; style relatively short,

sparsely pilose; stigma obscurely 4-lobed. Fruit

1,9 X 1,7 cm, subglobose, slightly flattened, asym-

metrical; exocarp pilose; mesocarp very fleshy;

putamen 1x0,8 cm, ellipsoid, asymmetrically and

IG. 51. — Commiphora marlothii near Sterkwater in the district

of Potgietersrus, northern Transvaal (height ±5 m).

80

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 54. — Flowers of Commiphora marlothii: A, male flower;

B, longitudinal section of male flower; C, diagrammatic

representation of two disc lobes illustrating the insertion

of the filaments; D, female flower; E, female flower with

the calyx and corolla partly removed.

Fig. 56. — Geographical distribution of Commiphora marlothii

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with numerous multicellular non-

glandular and glandular hairs; sclerenchymatous

pericycle-cylinder fluted. Stems of 2,5 cm diameter;

sclerenchymatous pericycle-cylinder consisting of

fibres and stone cells; epithelium cells of resin ducts

in xylem rays surrounded by 1-2 layers of cells with

slightly thickened walls. Leaves with numerous multi-

cellular non-glandular and glandular hairs; petiole

± triangular to ovate as seen in transverse section,

sclerenchymatous pericycle present, vascular bundles

± circularly distributed as seen in transverse section

but 3-8 medullary bundles also present; terminal

leaflet dorsiventral with a single layer of palisade cells

adaxially, remaining mesophyll consiting of spongy

or palisade-like cells, bulliform cells relatively small

and confined to adaxial epidermis, stomata mainly

abaxial.

Diagnostic features

Dioecious tree, bark peeling in large yellowish

papery pieces to expose a green underlayer; branchlets

obtuse, densely pilose to pubescent. Leaves impari-

pinnate, densely pilose to pubescent, leaflets 3-5-

jugate, margins crenate-serrate to finely lobed,

terminal leaflet dorsiventral, petiole with 3-8 medul-

lary vascular bundles. Flowers hypogynous, unisexual;

pedicels, calyx, corolla, disk and gynoecium pilose

to pubescent. Fruit subglobose; exocarp pilose;

putamen slightly rugose; pseudaril yellow, fleshy, with

4 arms, 2 arms on seam of putamen reaching almost

to apex, 2 arms on flattened faces of putamen shorter.

Widely distributed in central and northern Trans-

vaal. Usually grows on arid mountain slopes or on

granite kopjes.

Also recorded from Botswana, Rhodesia and

Zambia.

Fig. 55. — Fruit of Commiphora marlothii: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Transvaal. — 2229(Waterpoort): near Soutpan (-CD), Ober-

meyer , Schweikerdt & Verdoorn 121 ; 165 ; Wylliespoort (-DD),

Pole-Evans 3765; Masekwaspoort (-DD), Vahrmeijer 151.

2230( Messina): Tshipise (-CA), Van der Walt 8; Nshelele

(-CC), Van der Schijff 5275. 223 1 (Pafuri) : 50 km N.E. of

!

N

)

«•— -

J. J. A. VAN DER WAI.I

I

Punda Milia ( AC), Codd & de Winter 5534 , near Pafuri

( AC), Van der Schijff & Marais 3711. 2327(1 llisras): 74 km

N.W. of Vaalwatcr (-DC), Meeuse 10434. 2328( Baltimore):

near Stcrkwatcr ( DD), Van der Walt 40. 2319(Pictersburg):

Bandelierkop ( BD), Gerstner 5622. 2427(Thabazimbi): S I .

of Rooibcrg ( DC), Verdoorn 2370. 2430(Pilgrim’s Rest):

S.E. of Penge ( AD), Story 40HH; Abel Erasmus Pass (-BD),

Strey 3307; 6 km N. of Branddraai (-DA), Codd & de Winter

5260. 2528(Pretoria: near Rust de Winter (-BA), Codd 6200;

Pole-Evans 4240. 2529(Witbank): near Dcnnilton ( AC),

Codd 0000; Loskopdam Nature Reserve ( AD), Van der Walt

15; Mom 3 1000.

It is recorded that the papery bark is used by natives

as writing paper.

Common name: Paper Tree.

10. Commiphora edulis ( Klotzsclt ) Engl, in A. DC.,

Monogr. Phan. 4: 22 (1883); Bot. Jahrb. 48: 474,

t. IS (1913); Pflanzenfam. ed. 2,19a: 435 (1931);

Burtt Davy, FI. Transv. 2: 484 (1932); Burtt in Kew

Bull. 1935: 108 (1935); Codd, Mem. Bot. Surv. S.

Afr. 26: 86 (1951); Miller in J. S. Afr. Bot. 18: 38

(1952); Wild in Bol. Soc. Brot. 2,33: 91 (1959);

White, For. FI. N. Rhod. 176, t. 34E (1962); Wild

in FI. Zamb. 2,1 279 (1963); Von Breitenbach, Ind.

Trees S. Afr. 3,2 435 (1965); De Wint. in Trees S.

Afr. 20: 12 (1968). Type: Mozambique, Sena,

Peters s.n. (B, holo.f; K! ).

Hitzeria edidis Klotzsch in Peters, Reise Mossamb. Bot.

1: 89 (1861).

Commiphora elilorocarpa Engl, in Bot Jahrb. 28: 414, t. IN

(1901); Pflanzenfam. 2,19a: 435 (1931). Type: Tanzania,

Ruaha River, Goetze 452 (B, holo.t).

Dioecious many-stemmed shrub or small tree up

to 6 m tall, stems usually entwined; bark light grey,

flaking in small yellowish papery pieces, branehlcts

obtuse, densely pubescent. Leaves impari-pinnatc,

greyish green, pubescent; lamina up to 25 cm long;

petiole up to 8 cm long; pctiolules up to I cm long;

leaflets 2 6-jugati >wly ovate,

apex acute or rounded, base obtuse, margins usually

entire but in some cases finely crenate-serratc ; terminal

leaflet up to 6,5 3 cm; lateral leaflets up to 5 .6

cm. I low ers unisexual, perigynous, appearing with

the leaves in axillary paniculate simple or compound

dichasial cymes up to 15 cm long, branches of in-

floresences pubescent, male inflorescences usually

longer than female inflorescences; male flowers, 6 7

mm, usually larger than female flowers, 5 6 mm.

Bracteoles up to 3 mm long, linear, pubescent.

Pedicels I 1,5 mm long, clustered, pubescent Calyx

yellow to green, continuous with hypanthium,

pubescent, lobes 2-3 mm long, apex acute. Petals

yellowish green, 3-4,5 mm long, inserted on hypan-

thium, glabrous. Disk much reduced, cylindrical,

adnate to hypanthium. Stamens 8, inserted on top

of disk, 4 long stamens up to 3 mm long, 4 short

stamens up to 2 mm long; filaments subtercte. lower

part flattened and broadened: staminodes in female

flowers. Gynoecium; rudimentary in male flowers;

ovary half inferior, pilose; style relatively short,

pilose; stigma obscurely 4-lobed. Fruit 2,4 / 2,3 cm,

subglobose, slightly flattened asymmetrical; exocarp

pilose; mesocarp very fleshy; putamen 1,5 - 0.9 cm,

ellipsoid, much flattened, slightly asymmetrical

smooth; pseudaril red, very fleshy, cupular with 4

short lobes, covering the lower ^ of putamen. lobe on

less convex face of putamen longer than other 3

lobes. Fig. 57-62.

Fig. 57.- Commiphora edulis

near Messina, northern

Transvaal (height ’ 4 m).

82

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 58. — Close-up view of a branch of Commiphora edulis

illustrating the bark flaking off in small, papery pieces.

Fig. 60. — Flowers of Commiphora edulis: A, male flower;

B, longitudinal section of male flower; C, female flower;

D, female flower with the calyx and corolla partly removed

Fig. 59. — Commiphora edulis: A, branchlet with young leaves

and flowers; B, branchlet with leaves and mature fruits.

Fig. 61. — Fruit of Commiphora edulis: A, side-view of the

fruit; B, view of the less convex face of putamen and

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

J. J. A. VAN DER WALT

83

Fig. 62. — Geographical distribution of Commiphora edulis in

South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with numerous multicellular non-

glandular and glandular hairs; sclerenchymatous

pericycle-cylinder fluted. Stems of 2,5 cm diameter:

sclerenchymatous pericycle-cylinder consisting of

fibres and stone cells; epithelium cells of resin ducts

in xylem rays surrounded by thin-walled cells.

Leaves with numerous multicellular non-glandular

and glandular hairs; petiole ± triangular as seen in

transverse section, sclerenchymatous pericycle pre-

sent, vascular bundles ± circularly distributed as

seen in transverse section but 3-8 medullary bundles

also present; terminal leaflet typically dorsiventral

with a single layer of palisade cells adaxially, remaining

mesophyll consisting of spongy parenchyma, bulliform

cells confined to adaxial epidermis, stomata mainly

abaxial.

Diagnostic features

Dioecious many-stemmed shrub or small tree;

bark light grey, flaking in small yellowish papery

pieces, stems usually entwined; branchlets obtuse,

densely pubescent. Leaves impari-pinnate, pubescent,

leaflets 2-6-jugate, margins usually entire, terminal

leaflet typically dorsiventral, petiole with 3-8 medul-

lary vascular bundles. Flowers perigynous, unisexual,

pedicels and calyx pilose to pubescent. Fruit sub-

globose; exocarp pilose; putamen smooth; pseudaril

red, very fleshy, cupular with 4 short lobes.

This species is recorded from the far-northern and

north-eastern Transvaal. These areas are warm and

dry and the annual rainfall is less than 400 mm. It

usually occurs in the vicinity of the Limpopo River,

but is particularly common in the Messina area.

It grows in savanna-woodland or broken mopani-

veld in well-drained, sandy soil.

Also recorded from Botswana, Rhodesia, Zambia,

Tanzania, Mozambique and Malawi.

Transvaal. — 2229(Waterpoort): 24 km N.W. of Messina

(-BB), Van der Walt 24; 24 km W.N.W. of Messina (-BD),

Codd 4866. 2230(Messina): near Messina (-AC), Rogers 18473;

40 km N.E. of Tshipise (-AD), Van der Schijff 5222 (PRU).

2231(Pafuri): 45 km N.E. of Punda Milia Rest Camp (-AC),

Codd & Dyer 4622; near Pafuri (-AC), Van der Schijff 3566.

C. edulis is one of the first Commiphora species of

northern Transvaal to shed its leaves, the plants

being leafless as early as March.

The fruits are eaten by birds, rodents and baboons.

11. Commiphora woodii Engl, in Bot. Jahrb. 15:

97 (1893); Bot. Jahrb. 26: 371 (1899); Bot. Jahrb.

44: 154 (1910); Bot. Jahrb. 48: 476, t. 2E (1913);

Pflanzenfam. ed. 2,19a: 435 (1913); Wild in Bol. Soc.

Brot. 2,33: 91 (1959); Von Breitenbach, Ind. Trees S.

Afr. 3,2: 422 (1965); Moll, For. Trees Natal 80

(1967); De Wint. in Trees S. Afr. 20,1: 18 (1968).

Syntypes: Natal, Durban, Wood sub NH 861 (Bf ?;

BM!?); Pinetown, Rehmann s.n. (B|?); Inanda,

Rehmann s.n. (B|?). Lectotype: Wood s.n. (BM).

Commiphora caryaefolia Oliv. in Hook. Icon. PL 23: t,

2287 (1894); Engl, in Pflanzenfam. ed. 2,19a: 435 (1931).

Henkel, Woody PI. Natal 213 (1934). Syntypes: Natal, Durrban;

Wood 4095 (BOL!; NH!); Inanda, Wood 1046 (NH!); Wood

1409 (not seen); Flanagan 1107 (Z!).

Dioecious tree up to 12 m tall; bark grey, not

peeling; branchlets glabrous, shallowly fluted, obtuse.

Leaves impari-pinnate, glabrous; lamina up to 32 cm

long; petiole up to 9 cm long; petiolules up to 5 mm

long; leaflets 3-5-jugate, narrowly elliptic to elliptic,

Fig. 63. — Commiphora woodii

in the Ubisana valley near

Empangeni, Zululand

(height ±3,5 m).

84

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

apex acute, base cuneate to rounded, margins crenate-

serrate to coarsely crenate-serrate; terminal leaflet

up to 12x5 cm; lateral leaflets up to 13x5 cm.

Flowers unisexual, subsessile, perigynous, appearing

before the leaves in axillary paniculate simple or

compound dichasial cymes up to 10 cm long, male

inflorescences usually longer than female inflores-

cences; male and female flowers rt of equal size,

4-5 mm long, relatively broad. Bracteoles up to 2 mm

long, lanceolate. Pedicels usually less than 1 mm long.

Calyx yellowish green, glabrous, continuous with

hypanthium, lobes 2,5-3 mm long, apex acute.

Petals yellowish green, 3-4 mm long, inserted on

hypanthium. Disk fleshy or very fleshy in female

flowers, adnate to hypanthium, cylindrical, with 4

lobes. Stamens 8, inserted on top of disk, 4 long

stamens up to 2 mm long, 4 short stamens up to 1,5

mm long; filaments flattened and lower part broad-

ened; staminodes in female flowers. Gynoecium:

rudimentary in male flowers; ovary half inferior;

style relatively short; stigma obscurely 4-lobed.

Fruit 2x1,9 cm, subglobose, slightly flattened,

asymmetrical; exocarp glabrous; mesocarp very

fleshy; putamen 1x0,8 cm, ellipsoid, asymmetrically

and irregularly flattened, smooth; pseudaril red, very

fleshy, cupular, covering lower ^ of putamen, with 1

very short lobe on the less convex face of putamen,

margin finely crenate. Fig. 63-68.

Fig. 64. — Close-up view of a branch of Commiphora woodii.

Distinctive anatomical features of the stems and leaves

Young stems with a few glandular hairs especially

near apex; sclerenchymatous pericycle-cylinder fluted.

Stems of 2,5 cm diameter: sclerenchymatous peri-

cycle-cylinder consisting of fibres and stone cells;

epithelium cells of resin ducts in xylem rays sur-

rounded by thin-walled cells. Leaves with a few

glandular hairs; petiole ± oval to ovate as seen in

transverse section, sclerenchymatous pericycle

present, vascular bundles ± circularly distributed

as seen in transverse section; terminal leaflet typically

dorsiventral with a single layer of palisade cells

adaxially, remaining mesophyll consisting of spongy

parenchyma, bulliform cells confined to adaxial

epidermis, stomata exclusively abaxial.

Fig. 65. — Commiphora woodii: A, branchlet with a leaf and

flowers; B, branchlet with a leaf and mature fruits.

Fig. 66.— Flowers of Commiphora woodii: A, male flower;

B, longitudinal section of male flower; C, female flower:

D, female flower with the calyx and corolla partly removed.

J. J. A. VAN DER WALT

85

Fig. 67. — Fruit of Commiphora woodii: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Fig. 68. — Geographical distribution of Commiphora woodii

in South Africa.

Diagnostic features

Dioecious tree; bark grey, not peeling; branchlets

shallowly fluted, obtuse, glabrous. Leaves impari-

pinnate, glabrous, leaflets 3-5-jugate, margins crenate-

serrate to coarsely crenate-serrate, terminal leaflet

typically dorsiventral, petiole without medullary

vascular bundles. Flowers perigynous, unisexual, gla-

brous. Fruit subglobose; exocarp glabrous; putamen

smooth; pseudaril red, very fleshy, cupular with 1

very short lobe.

This species occurs near the coast, from Zululand

southwards to East London. It usually grows on the

slopes of mountains or in kloofs as part of the coastal

forests with a rainfall of 1 000 mm and more per

annum.

Also recorded from Mozambique.

Natal. — 2732(Ubombo): near Ingwavuma (-AA), Dutton

& Tinley 6; 29 km S. of Jozini (-CA), de Winter & Vahrmeijer

8481. 2830( Dundee): S.W. of Weenen (-CC), Pentz 545.

2831(Eshowe): Umhlatuzi Valley at Ntimona (-DD), Van der

Walt 28; Ubisana Valley near Kwa-Dlangezwa (-DD), Van

der Walt 83. 2832(Mtubatuba): near Hluhluwe (-AA), Wells

2130; Elluhluwe Game Reserve (-AA), Ward 2982 (NH).

29301 Pietermaritzburg): near Richmond (-CB), Moll & Morris

663; Camperdown at Shongweni Dam (-DC), Marais ?

30301 Port Shepstone): Port Shepstone (-CB), Nicholson

247 (NH).

Cape. — 3226(Fort Beaufort): Victoria East at Pefferskop

(-DD), Acocks 11900. 3227(Stutterheim): 18 km S.W. of

King William's Town (-CD), Acocks 11879 ; Kei Road (-DA)‘

Ranger ?. 3228(Butterworth): Kentani (-AD), Pegler 1137.

None of the three syntypes mentioned by Engler

(1893) could be traced and it is suspected that they

were destroyed in Berlin. A specimen from the

British Museum with a label of the Natal Herbarium,

collected by Wood in Berea, Durban, was seen.

Since no number appears on the label, it is uncertain

whether this specimen is an isotype.

C. woodii and C. zanzibarica, two closely related

species, can be distinguished on features of the

inflorescences, flowers and fruit. The inflorescences

and flowers of C. zanzibarica are relatively long.

Medullary vascular bundles occur in the petioles of

C. zanzibarica but are absent in those of C. woodii.

The leaves of C. woodii and C. harveyi are sometimes

also confused. Short hairs occur on the leaves of

C. harveyi while those of C. woodii are glabrous.

C. woodii grows easily from pole cuttings which

are often planted as fencing poles. Natives prepare

gum from the bark.

12. Commiphora zanzibarica [Bail!.) Engl, in

A. DC., Monogr. Phan. 4: 28 (1883); Bot. Jahrb.

48: 468, t. 1A (1913); Pflanzenfam. ed. 2,19a: 433

(1931); Burtt in Kew Bull. 1935: 111 (1935); Wild in

Bol. Soc. Brot. 2,33: 91 (1959); Dale & Greenway,

Kenya Trees 93 (1961); Wild in FI. Zamb. 2,1: 279

(1963). Type: Tanzania, Zanzibar, Jablonski s.n.

(P, holo! ).

Balsamea zanzibarica Baill. in Adonsonia 11: 180 (1874);

Engl, in Bot. Jahrb. 1: 42 (1881).

Commiphora spondioides Engl, in Bot. Jahrb. 26: 371 (1899);

Bot. Jahrb. 48: 468 (1913); Pflanzenfam. ed. 2,19a: 433 (1931).

Type: Mozambique, Lourengo Marques, Schlechter 11559

(B, holo.!; K!).

Dioecious tree, often many-stemmed, up to 7 m

tall; bark grey, not peeling; branchlets glabrous,

shallowly fluted. Leaves impari-pinnate, glabrous;

lamina up to 20 cm long; petiole up to 6 cm long;

petiolules up to 5 mm long; leaflets 3-5-jugate,

oblanceolate to narrowly elliptic, apex obtuse to

acute, base cuneate to rounded, margins entire to

finely serrate; terminal leaflet up to 7 x 3,3 cm ; lateral

leaflets up to 6,8 X 2,2 cm. Flowers unisexual, peri-

gynous, appearing after the leaves in axillary pani-

culate simple or compound dichasial cymes up to

30 cm long; male flowers, 1,1-1, 3 cm, larger than

female flowers, 7-8 mm. Bracteoles up to 1 mm long,

lanceolate. Pedicels 4-6 mm long. Calyx yellowish

green, glabrous, continuous with hypanthium, lobes

2-3 mm long, apex acute. Petals yellowish green.

2-5 mm long, inserted on hypanthium. Disk fleshy,

adnate to hypanthium, cylindrical, with 4 lobes.

Stamens 8, inserted on top of disk, 4 long stamens

up to 4 mm long, 4 short stamens up to 3 mm

long; filaments subterete, lower part flattened and

broadened; staminodes in female flowers. Gynoecium;

86

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

rudimentary in male flowers; ovary half inferior;

style relatively short; stigma obscurely 4-lobed.

Fruit 1,8 x 1,5 cm, subglobose, slightly flattened,

asymmetrical; exocarp glabrous; mesocarp very

fleshy; putamen 1,8 x 0,8 cm, ellipsoid asymmetrically

and irregularly flattened, smooth; pseudaril red, very

fleshy, cupular, covering the lower of putamen,

margin coarsely crenate. Fig. 69-74.

Fig. 69. — Commiphora zanzibarica near Makanies^Pont,

northern Zululand (height ±7 m).

Fig. 70. — Close-up view of a branch of Commiphora zanzibarica.

The white patches on the bark are lichens.

Fig. 71. — Commiphora zanzibarica: A, branchlet with a lea

and flowers; B, branchlet with a leaf and mature fruits.

Distinctive anatomical features of the stems and leaves

Young stems with a few peltate glandular hairs

especially near apex; sclerenchymatous pericycle-

cylinder fluted. Stems of 2,5 cm diameter: scleren-

chymatous pericycle-cylinder consisting of fibres and

stone cells, epithelium cells of resin ducts in xylem

rays surrounded by thin-walled cells. Leaves with a

few peltate glandular hairs; petiole ± triangular as

seen in transverse section, sclerenchymatous pericycle

present, vascular bundles i circularly distributed as

seen in transverse section but 4-14 medullary bundles

also present; terminal leaflet typically dorsiventral

with a single layer of palisade cells adaxially,

remaining mesophyll consisting of spongy parenchyma

bulliform cells confined to adaxial epidermis, stomata

mainly adaxial.

J. J. A. VAN DER WALT

87

Fig. 72. — Flowers of Commiphora zanzibarica: A, male flower;

B, longitudinal section of the male flower; C, female

flower; D, female flower with the calyx and corolla

partly removed.

Fig. 73. — Fruit of Commiphora zanzibarica: A, side-view of

the fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Diagnostic features

Dioecious tree, often many-stemmed; bark grey,

not peeling; branchlets shallowly fluted, glabrous.

Leaves impari-pinnate, glabrous, leaflets 3-5-jugate,

margins entire of finely serrate, terminal leaflet

typically dorsiventral, petiole with 4-14 medullary

vascular bundles. Flowers perigynous, unisexual,

glabrous, inflorescences relatively long (up to 30 cm).

Fruit subglobose; exocarp glabrous; putamen smooth;

pseudaril red, very fleshy, cupular without lobes.

So far this species has only been collected near the

Kumane Dam in the Kruger National Park and on

the Makatini Flats in northern Zululand where it

grows in deep sandy soil in savanna-woodland. The

annual rainfall in these areas lies between 400-550 mm.

Also recorded from Mozambique, Rhodesia, Zan-

zibar, Tanzania and Kenya.

Transvaal. — 243 l(Acornhoek): at Kumane Dam (-DB), Van

Wyk 4917.

Natal. — 2732(Ubombo): 8 km S.E. of Makanies Pont

(-AB), Van der Walt 89.

Dale & Greenway (1961) and also Wild (1963)

mention that the bark of this species peels off in

papery pieces. This is not the case with the plants from

which the material for this study was collected.

13. Commiphora tenuipetiolata Engl, in Bot.

Jahrb. 48: 483, 3L (1913); Pflanzenfam. ed. 2,19a:

438 (1931); Burtt Davy, FI. Transv. 2: 485, t. 52

(1932); Wild in Bol. Soc. Brot. 2,33: 93 (1959);

Wild in FI. Zamb. 2,1: 280 (1963); Von Breitenbach,

Ind. Trees S. Afr. 3,2: 434 (1965); De Wint. in Trees

S. Afr. 20,1: 16 (1968); Merxm., Prod. FI. S.W. Afr.

23: 9 (1968). Syntypes: S.W. A., Otjiwarongo, Ses-

fontein, Dinter 1721 (Bf; K, fragment!; BM, sketch!);

S.W. A., Bulspoort, Dialer 2109 (Bf; K, fragment!).

Dioecious tree with a single main stem up to 7 m

tall; bark grey to white, usually peeling in large

whitish papery pieces to expose a glaucous under-

layer; branchlets glabrous. Leaves trifoliolate or

impari-pinnate, glabrous; lamina up to 8 cm long;

petiole relatively long and slender, up to 5 cm long,

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

glabrous; petiolules up to 1 mm long; leaflets 1-3-

jugate, obovate to broadly elliptic or elliptic; apex

acute but more often obtuse, base cuneate, margins

entire or crenate-serrate in the upper half; terminal

leaflet up to 3x2 cm; lateral leaflet 3 x 1,8 cm.

Flowers unisexual, perigynous, appearing after the

leaves in axillary simple or compound dichasial cymes

up to 5,5 cm long; male flowers 1, 2-1,4 cm, usually

larger than female flowers, 1-2,2 cm. Bracteo/es

linear, up to 5 mm long. Pedicels usually relatively

long, 6-10 mm. Calyx yellowish green, glabrous,

continuous with hypanthium, lobes 1,2-1, 5 mm long,

apex acute. Petals yellowish green, 2-3 mm long,

inserted on hypanthium, glabrous. Disk reduced, not

fleshy, adnate to hypanthium, cylindrical, with 4

inconspicuous lobes. Stamens 8, inserted on disk, 4

long stamens up to 2 mm long, 4 short stamens up

to 1,5 mm long; filaments subterete, lower part

flattened and much broadened; staminodes in female

flowers. Gynoecium: rudimentary in male flowers,

ovary half inferior; style relatively short; stigma

2-lobed. Fruit 1,5 x 1,3 cm, subglobose, slightly

flattened and asymmetrical, very much flattened,

smooth; pseudaril red, fleshy, cupular with 2 lobes

of variable length and shape on flattened faces of

putamen, covering lower \ of putamen, lobe on

less convex face of putamen usually longer and more

acute than lobe on other face. Fig. 75-81.

Fig. 75. Commiphora tcnuipetiolata near Melkrivier in the

district of Vaalwater, northern Transvaal (height ^6 m).

Distinctive anatomical features of the stems and leaves

Young stems with a few glandular hairs especially

near apex; dendritic crystals (hesperidin or diosmin)

occurring in some epidermal and hypodermal cells.

Stems of 2,5 cm diameter: sclerenchymatous pericycle

consisting of fibres and stone cells, epithelium cells

of resin ducts in xylem rays surrounded by 1 layer of

sclereids. Leaves with a few glandular hairs, dendritic

crystals (hesperidin or diosmin) occurring in some

epidermal and hypodermal cells of the petioles and

leaflets; petiole ± circular as seen in transverse

section, sclerenchymatous pericycle present, vascular

bundles ± circularly distributed as seen in transverse

section; terminal leaflet isobilateral with a single layer

of long palisade cells adaxially and a single layer of

shorter palisade cells abaxially, remaining mesophyll

consisting of spongy or palisade-like cells, adaxial

epidermis consisting mainly of large bulliform cells

but smaller bulliform cells occur in abaxial epidermis,

stomata mainly abaxial.

Fig. 76, Fig. 77. — Close-up views of different branches of

Commiphora tenuipetiolata.

J. J. A. VAN DER WALT

89

Fig. 78 — Commiphora tenuipetiolata: A, branchlet with leaves

and flowers; B-E, leaves; F, branchlet with leaves and ma-

ture fruits.

Fig. 80. — Fruit of Commiphora tenuipetiolata: A, side-view

of the fruit; B, view of the less convex face of putamen

with pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Fig. 79. — Flowers of Commiphora tenuipetiolata: A, male

flower; B, longitudinal section of male flower; C, female

flower; D, female flower with the calyx and corolla partly

removed.

Fig. 81. — Geographical distribution of Commiphora tenuipetio-

lata in South Africa.

Diagnostic features

Dioecious tree with a single main stem; bark grey

to white, usually peeling in large whitish papery

pieces to expose a glaucous underlayer; branchlets

glabrous. Leaves trifoliolate or impari-pinnate, gla-

brous, leaflets 1-3-jugate, margins entire or crenate-

serrate in the upper half, terminal leaflet isobilateral,

dendritic crystals (hesperidin or diosmin) occurring

in some epidermal and hypodermal cells of petioles

and leaflets, petioles relatively long and slender (up

to 5 cm). Flowers perigynous, unisexual, glabrous.

Fruit subglobose; exocarp glabrous, putamen very

much flattened, smooth; pseudaril red, fleshy, cupular

with 2 lobes on flattened faces of putamen.

90

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

This species occurs in the far-northern and north-

eastern Transvaal, but is particularly common north

of the Soutpansberg. Grows in well-drained, sandy

soil in warm areas with a relatively low annual

rainfall.

Also recorded from South West Africa, Botswana

and Rhodesia.

Transvaal. — 2228(Maasstroom): near Koperspruit (-CB),

Van der Walt 61. 2229(Waterpoort): Dongola (-BC), Codd

4140; Gerstner 5460; 19 km W. of Messina (-BD), Gerstner

5460; 8 km E. of Waterpoort (-DC), Van der Walt 31. 2230

(Messina): near Messina (-AC), Rogers 20763; 40 km N.E.

of Tshipise in Nwanedzi River valley (-AD), Gerstner 6039;

Tshipise (-CA), Van der Schijff 5197 (PRU); Van der Walt 4;

15. 2231(Pafuri): Mzimbiti Kloof near Punda Milia Rest

Camp (-CA), Van Wyk & Pienaar 4696. 2328( Baltimore):

32 km N.E. of Melkrivier (-CD), Van der Walt 46.

On the basis of the size, form and margin of the

terminal leaflet, Burtt Davy (1932) distinghuishes the

var. tenuipetiolata and var. rogersii Burtt Davy. The

supposed existence of the two varieties is not supported

since these variations of the terminal leaflet occur on

the same plant.

It has been recorded that C. tenuipetiolata also

occurs in shrub form in South West Africa.

14. Commiphora angolensis Engl, in A. DC.,

Monogr. Phan 4: 24 (1883); Bot. Jahrb. 48: 486

(1913); Pflanzenfam. ed. 2,19a: 438 (1931); Exell &

Mendonca in Consp. FI. Angol. 1,2: 300 (1951);

Wild in Bol. Soc. Brot. 2,33: 39 (1959); White, For.

FI. N. Rhod.: 176, t. 34D (1962); Wild in FI. Zamb.

2,1: 281 (1963); Von Breitenbach, Ind. Trees S. Afr.

3,2: 440 (1965); De Wint. in Trees S. Afr. 20,1: 8

(1968); Merxm., Prod. FI. S.W. Afr. 23: 5 (1968).

Syntypes: Angola, Fuanda, Welwitsch 4485; sine

loc. 4488 (G, only photo seen; FISU!). Lectotype:

Welwitsch 4485 (L1SU).

Balsamea angolensis (Engl.) Hiern, Cat. Welw. PI 1,1: 24

(1896).

Commiphora oliveri Engl, in A. DC., Monogr. Phan 4: 24

(1883); Bot. Jahrb. 48: 483, t. 3K (1913); Pflanzenfam. ed.

2,19a: 438 (1931). Type: Botswana, Baines s.n. (K, holo. !).

C. rehmannii Engl, in A. DC., Monogr. Phan 4: 15 (1883);

Bot. Jahrb. 48: 483 (1913); Burtt Davy, FI. Transv. 2: 485

(1932). Type: Transvaal, Rehmann s.n. (B, holo.f; K, fragment

and photo of holo. !; BM, sketch of holo.!); Klippan, Rehmann

5324 (Z!). Lectotype: Rehmann 5324 (Z). C. longebracteata

Engl, in A.DC., Monogr. Phan. 4: 19 (1883); Bot. Jahrb. 48:

486 (1913); Pflanzenfam. ed. 2,19a: 438 (1931). Type: Angola,

Welwitsch 4494 (G, holo., only photo seen; LISU!). C.

kwebensis N.E. Br. in Kew Bull. 1909: 98 (1909); Miller in

J.S. Afr. Bot. 18: 38 (1952). Syntypes: Botswana, Kwebe

Hills, Lngard 34 (K !) ; Lugard 86 (K!). Lectotype: Lugard 86

(K). C. gossweileri Engl, in Bot. Jahrb. 44: 147 (1910). Type

Angola, Luanda, Gossweiler 442 (B, holo.f; K!; BM!). Lec-

totype: Gossweiler 442 (K). C. nigrescens Engl, in Bot. Jahrb.

44: 148 (1910); Bot. Jahrb. 48: 484 (1913); Pflanzenfam. ed.

2,19a: 438 (1931). Syntypes: S.W. A., Grootfontein, Dinter

727 (B,f; K, fragment!); Dinter 727a (B,f; BM, sketch!).

Lectotype: Dinter 727 (BM).

Dioecious many-stemmed shrub up to 3 m tall;

bark yellowish green to chestnut-brown, peeling in

yellowish pieces to expose a green underlayer;

branchlets pilose to densely pubescent. Leaves tri-

foliolate or impari-pinnate, sparsely pilose to densely

pubescent; lamina up to 9 cm long; petiole up to

2,2 cm long; petiolules up to 4 mm long; leaflets

1-3-jugate, terminal leaflet up to 3,5x2 cm, elliptic

to obovate; lateral leaflets up to 2x1,2 cm, elliptic;

apex of all leaflets acute to obtuse, base cuneate,

margins crenate-serrate. Flowers unisexual, peri-

gynous, appearing after the leaves in axillary dichasial

cymes, male inflorescences usually compound dichasial

cymes up to 5 cm long, female inflorescences usually

simple dichasial cymes up to 3 cm long; male flowers,

8-1 1 mm, usually larger than female flowers, 6-8 mm.

Bracteo/es linear, up to 6 mm long, sparsely pilose to

densely pubescent. Pedicels 4-5 mm long, sparsely

pilose to densely pubescent. Calyx yellow to green,

sparsely pilose to densely pubescent, continuous with

hypanthium, lobes 2-3 mm, apex acute. Petals yellow

to green, 2-3 mm long, usually glabrous but in some

cases sparsely pilose on outside. Disk reduced, not

fleshy, adnate to hypanthium, cylindrical, with 4

inconspicuous lobes. Stamens 8, inserted on disk, 4

long stamens up to 4 mm long, 4 short stamens up

to 3 mm long; filaments subterete, lower part much

flattened and broadened; staminodes in female

flowers. Gynoecium: rudimentary in male flowers;

ovary half inferior; style relatively short; stigma

obscurely 4-lobed. Fruit 1,1x0, 9 cm, subglobose to

ellipsoid, asymmetrically flattened; exocarp pilose;

mesocarp fleshy; putamen 9x7 mm ellipsoid, slightly

asymmetrical, much flattened, smooth; pseudaril red,

fleshy, cupular with 2 lobes of variable length and

shape on flattened faces of putamen, covering lower

i-y of putamen, lobe on less convex face of putamen

usually longer than lobe on other face. Fig. 82-87.

Fig. 82. — Commiphora angolensis near Rooibokkraal, north-

western Transvaal (height ±3 m).

J. J. A. VAN DER WALT

91

Fig. 83. — Close-up view of a branch of Commiphora angolensis

illustrating the bark peeling in papery pieces.

Fig. 85. — Flowers of Commiphora angolensis: A, male flower;

B, longitudinal section of male flower; C, female flower;

D, female flower with calyx and corolla partly removed.

Fig. 84. — Commiphora angolensis: A, branchlet with leaves

and flowers; B-D, leaves; E, branchlet with leaves and

mature fruits.

Fig. 86. — Fruit of Commiphora angolensis: A, side-view of

the fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D. putamen and pseudaril as seen from

above.

92

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 87. — Geographical distribution of Commiphora angolensis

in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with a variable number of multi-

cellular non-glandular and glandular hairs; dendritic

crystals (hesperidin or diosmin) occurring in some

epidermal and hypodermal cells. Stems of 2,5 cm

diameter: sclerenchymatous pericycle consisting of

fibres and stone cells, epithelium cells of resin ducts

in xylem rays surrounded by 2-3 layers of sclereids.

Leaves with a variable number of multicellular non-

glandular and glandular hairs; dendritic crystals

(hesperidin or diosmin) occurring in some epidermal

and hypodermal cells of the petioles and leaflets;

petiole ± triangular as seen in transverse section,

terminal leaflet isobilateral with a single layer of long

palisade cells adaxially and a single layer of shorter

palisade cells abaxially, remaining mesophyll consis-

ting of spongy or palisade-like cells, adaxial epidermis

consisting mainly of large bulliform cells but smaller

bulliform cells occur in abaxial epidermis, stomata

mainly abaxial.

Diagnostic features

Dioecious many-stemmed shrub; bark yellowish

green to chestnut-brown, peeling in yellowish pieces

to expose a green underlayer; branchlets pilose to

densely pubescent. Leaves trifoliolate or impari-

pinnate, sparsely pilose to densely pubescent, leaflets 1

1-3-j ugate, margins crenate-serrate, terminal leaflet

isobilateral, dendritic crystals (hesperidin or diosmin)

occurring in some epidermal and hypodermal cells

of petioles and leaflets, petioles up to 2,2 cm long.

Flowers perigynous, unisexual, pedicels and calyx !

pilose to pubescent, corolla in some cases sparsely

pilose on outside. Fruit subglobose to ellipsoid; j

exocarp pilose; putamen much flattened, smooth; j

pseudaril red, fleshy, cupular with 2 lobes on flattened

faces of putamen.

;

In South Africa this species is confined to a few

localities in the arid bushveld of north-western and

northern Transvaal north of the Soutpansberg. It

grows in deep sandy soil presumably derived from

the Kalahari.

Also recorded from Botswana, South West Africa,

Rhodesia, Zambia and Angola.

Transvaal. — 2229(Waterpoort): 10 km N.E. of Vivo (-CD), i

Van der Walt 19; near Mopane (-DB), Strey 3502. 2317

(Ellisras): 3 km W. of Monte Christo (-BC), Codd 6603;

Van der Watt 110; 121; 8 km N. of Steenbokpan (-CB), Van

der Walt 122. 2329(Pietersburg): 3 km N.E. of Vivo (-AB),

Strey 3516. 2427(Thabazimbi): 13 km E.N.E. of Rooibokkraal

(-AA), Theron & Marsh 258 (PRU); Van der Walt 25.

As far as it could be determined, this species

occurs only in shrub form in South Africa. In South

West Africa it usually develops into a tree with a

single main stem.

15. Commiphora nainaensis Schinz in Bull.

Herb. Boiss. 2,8: 633 (1908); Wild in Bol. Soc. Brot.

2,33: 92 (1959); Merxm., Prod. FI. S.W. Afr. 23: 7

(1968). Type: S.W. A., Inachab, Dinter 958 (Z, holo. !).

Commiphora rotundifolia Dinter & Engl, in Bot. Jahrb. 46:

289 (1912); Engl, in Bot. Jahrb. 48: 482, t. 3G (1913); Pflan-

zenfam. ed. 2,19a: 438 (1931). Type: S.W. A., Seeheim, Dinter

1203 (B, holo.f; K, fragment!).

Fig. 88. — Commiphora na-

maensis near Vioolsdrif,

north-western Cape (height

±1 m).

J. J. A. VAN DER WALT

93

Dioecious shrub less than 1 m up to 3 m tall;

trunk branching repeatedly above soil level, forming

many relatively thin side branches; bark light grey,

not peeling; branchlets glabrous. Leaves simple,

glabrous, lamina up to 1,2x1 cm, orbicular or

slightly oblong, apex obtuse, base cuneate, margin

dentate to coarsely dentate, petiole up to 7 mm long.

Flowers subsessile, unisexual, perigynous, ap-

pearing before the leaves in axillary clusters; male

flowers, 4-5 mm, larger than female flowers, 3-4 mm.

Bracteoles up to 0,2 mm long, ± triangular, glandular.

Pedicels less than 0,2 mm long. Calyx green to brown,

continuous with hypanthium, lobes 0,5-1 mm long,

apex acute. Petals yellow to brown, 2,5-4 mm long,

inserted on hypanthium. Disk adnate to hypanthium,

cylindrical with 4 fleshy lobes, lobes in male flowers

not bifid but in female flowers bifid. Stamens 8,

4 long stamens up to 4 mm long, inserted on top of

disk lobes, 4 short stamens up to 2,5 mm long,

inserted on top of disk between lobes; filaments

subterete, lower part flattened and broadened;

staminodes in female flowers. Gynoecium : rudimentary

in male flowers; ovary half inferior, glandular; style

relatively long, glandular; stigma obscurely lobed.

Fruit 1x0,8 cm, subglobose to ellipsoid, slightly

flattened, asymmetrical; exocarp glabrous; mesocarp

not very fleshy; putamen 8x6 mm, ellipsoid, asym-

metrically and irregularly flattened, slightly rugose;

pseudaril red, fleshy, cupular with 2 arms on seam

of putamen, covering the lower $ of the more convex

face of putamen and the lower } of the other face.

Fig. 88-93.

Fig. 89. — Close-up view of a branch of Commiphora namaensis.

Distinctive anatomical features of the stems and leaves

Young stems with numerous peltate glandular hairs

at apex. Stems of 2,5 cm diameter; sclerenchymatous

pericycle-cylinder consisting of fibres and stone

cells, epithelium cells of resin ducts in xylem rays

surrounded by thin-walled cells, resin ducts in

primary phloem very conspicuous (up to 2 mm in

diameter). Leaves with a few peltate glandular hairs;

petiole ± semi-circular as seen in transverse section,

sclerenchymatous pericycle present or absent, vas-

cular bundles mainly abaxial and distributed in the

form of an arc as seen in transverse section, in some

cases 1-2 much smaller bundles adaxially, with a

large number of stomata; terminal leaflet typically

isobilateral with 1-3 layers of palisade cells ad- and

abaxially, central mesophyll consisting of ± colourless

cells, ad- and abaxial epidermis consisting mainly

of large bulliform cells, with a large number of

evenly distributed stomata in the ad- and abaxial

epidermis.

Fig. 90. — Commiphora namaensis: A, branchlet with flowers;

B, branchlet with leaves and flowers; C-E, leaves; F,

branchlet with leaves and mature fruits.

Fig. 91. — Flowers of Commiphora namaensis: A, male flower;

B, longitudinal section of male flower; C, female flower;

D, female flower with the calyx and corolla partly removed.

5292—7

94

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 92. — Fruit of Commiphora namaensis: A, side-view of the

fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Diagnostic features

Dioecious shrub, trunk branching repeatedly above

soil level, forming many relatively thin side branches;

bark light grey, not peeling; stems with very large

resin ducts in primary phloem; branchlets with

Fig. 93. — Geographical distribution of Commiphora namaensis

in South Africa.

numerous peltate glandular hairs at apex but otherwise

glabrous. Leaves simple, glabrous, lamina orbicular

or slightly oblong, margin dentate to coarsely dentate, ,

typically isobilateral. Flowers subsessile, perigynous,

unisexual. Fruit subglobose to ellipsoid; exocarp

glabrous; putamen slightly rugose; pseudaril red,

fleshy, cupular with 2 arms on seam of putamen.

In South Africa this species is confined to the semi-

desert areas of the north-western Cape. It occurs in

the mountains near the Orange River from Goodhouse 1

westwards. These areas are extremely dry and hot

with a rainfall of less than 80 mm per annum.

Also recorded from South West Africa.

J. J. A. VAN DER WALT

95

Cape. — 2817(Vioolsdrif): 16 km W. of Vioolsdrif (-DC),

Vander Walt 113; 114. 2818(Warmbad): 10 km S. of Goodhouse

(-CC), Van der Walt 120.

All the plants seen in the veld have simple leaves.

Plants cultivated in a glass house at Stellenbosch

developed trifoliolate leaves in addition.

16. Commiphora gracilifrondosa Dinter ex Van

der Walt in J.S. Afr. Bot. 37,3: 190 (1971); Dinter

in Fedde, Rep. Beih. 53: 48 (1928), nom. subnud.

Type: Dinter 5124 (BOL, holo. !; S! ; Bf).

Dioecious shrub up to 3 m tall; trunk branching

repeatedly above soil level, stamens appearing succose ;

bark reddish brown with dark patches, not peeling;

branchlets slender, glabrous. Leaves trifoliolate but

terminal leaflet often 3-lobed, glabrous; lamina up

to 6 cm long; petiole up to 2 cm long; petiolules

up to 3 mm long; leaflets variable in size and form,

linear to cultrate, margins irregularly and rather

coarsely dentate-serrate, apex obtuse to acute, base

cuneate; terminal leaflet up to 4,3x0, 2 cm, lateral

leaflets up to 3,5x0, 2 cm. Flowers unisexual, peri-

gynous, appearing before or with the leaves in

axillary dichasial cymes or occasionally solitary, male

inflorescences up to 5 cm long, female inflorescences

up to 1 cm long; male flowers 6-7 mm, usually

larger than female flowers, 4-5 mm. Bracteoles

up to 4 mm long, linear, sparsely glandular. Calyx

yellow to green, continuous with hypanthium,

sparsely glandular, lobes up to 1 mm long, apex

acute. Petals yellow to green, 2, 5-3, 5 mm long,

inserted on hypanthium. Disk adnate to hypan-

thium, cylindrical with 4 fleshy lobes. Stamens only

4, up to 2,5 mm long, inserted on top of disk lobes;

filaments slender, subterete, lower part flattened and

broadened; staminodes in female flowers. Gynoecium:

rudimentary in male flowers; ovary half inferior,

sparsely glandular; style relatively long, sparsely

glandular; stigma obscurely lobed. Fruit 1x0,8 cm,

subglobose to ellipsoid, asymmetrical, slightly flat-

tened; exocarp glabrous; mesocarp not very fleshy;

putamen 8x5 mm, ellipsoid, asymmetrically flattened,

smooth; pseudaril red, not very fleshy, cupular with

2 arms on seam of putamen, covering the lower i

of more convex face of putamen and \ of the other

face. Fig. 94-99.

Fig. 95. — Close-up view of stems of Commiphora gracili-

frondosa.

Fig. 96. — Commiphora gracilifrondosa: A — D, leaves; E,

branchlet with leaves and flowers; F, branchlet with

leaves and mature fruits.

Fig. 97. — Flowers of Commiphora gracilifrondosa: A, male

flower; B, longitudinal section of male flower; C, female

flower; D, female flower with the calyx and corolla partly

removed.

96

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 98. — Fruit of Commiphora gracilifrondosa : A, side-view

of the fruit; B, view of the less convex face of putamen with

pseudaril; C, view of the more convex face of putamen

with pseudaril; D, putamen and pseudaril as seen from

above.

Fig. 99. — Geographical distribution of Commiphora gracili-

frondosa in South Africa.

Distinctive anatomical features of the stems and leaves

Young stems with numerous peltate glandular hairs

at apex. Stems of 2, 5 cm diameter : sclerenchymatous

pericycle-cylinder consisting of fibres and stone cells,

epithelium cells of resin ducts in xylem rays sur-

rounded by thin-walled cells. Leaves with a few

peltate glandular hairs; petiole ± circular as seen

in transverse section, sclerenchymatous pericycle

present or absent, vascular bundles mainly abaxial

and distributed in the form of an arc as seen in

transverse section, in some cases 1-2 much smaller

bundles adaxially, with a large number of stomata;

terminal leaflet typically isobilateral, mesophyll con-

sisting mainly of palisade cells with a few ± colourless

cells centrally, ad- and abaxial epidermis consisting

mainly of large bulliform cells, with a large number

of evenly distributed stomata in the ad- and abaxial

epidermis.

Diagnostic features

Dioecious shrub, trunk branching repeatedly above

soil level, stems appearing succose; bark reddish

brown with dark patches, not peeling; branchlets

slender, with numerous peltate glandular hairs at

apex but otherwise glabrous. Leaves trifoliolate but

terminal leaflet often 3-lobed, glabrous, leaflets

linear to cultrate, margins irregularly and rather

coarsely dentate-serrate, terminal leaflet typically

isobilateral. Flowers perigynous, unisexual, only 4

stamens/staminodes. Fruit subglobose to ellipsoid;

exocarp glabrous; putamen smooth; pseudaril red,

not very fleshy, cupular with 2 arms on seam of

putamen.

This species occurs in the north-western Cape from

Kenhardt in the east to Goodhouse in the west. It

grows on the arid mountains and kopjes in the

vicinity of the Orange River in areas with an

annual rainfall up to 160 mm.

Also recorded from the southern part of South

West Africa.

Cape. — 2818(Warmbad): 5 km E. of Goodhouse (-CD),

Van der Walt 124; 27 km W.N.W. of Pella (-DD), Van der

Walt 119. 2819 (Ariamvlei): 8 km N. of Pella (-CC), Van der

Walt 116. 2820(Kakamas): near Augrabies (-CB), Pearson

3567 (BOL); near Kakamas (-DC), Fuller 24 (BOL). 2919

(Pofadder): 6 km N.E. of Pofadder (-AB), Acocks 21795.

2921 (Kenhardt): S. of Kenhardt (-AC), Hutchinson 952 (BOL).

This species is closely related to C. oblanceo/ata

Schinz. The type specimen ( Dinter 1497) of the

latter species has been studied. Like C. gracilifrondosa,

C. oblanceolata has also only 4 stamens. De Winter,

who studied the material of these two taxa in Kew,

also concluded that they should be considered as

different species.

It was observed that goats and game graze on the

young branches. The local name of “Suikerkan” is

probably derived from the sweet taste of the wood.

17. Commiphora capensis ( Sond .) Engl, in A. DC.,

Monogr. Phan. 4; 18 (1883); Bot. Jahrb. 48: 470

(1913); Pflanzenfam. ed. 2,19a: 433 (1931); Wild in

Bol. Soc. Brot. 2,33: 89 (1959); Von Breitenbach,

Ind. Trees S. Afr. 3,2: 436 (1965); de Wint. in Trees

S. Afr. 20,1: 10 (1968); Merxm., Prod. FI. S.W. Afr.

23: 5 (1968). Type: North-western Cape, Natvoet,

Drege 6809 (ex parte) (S, holo.!; MEL, fragment!).

Balsamodendrum capense Sond. in FI. Cap. 1: 526 (1860).

Balsamea capensis (Sond.) Engl, in Bot. Jahrb. 1: 42 (1881).

Commiphora rangeana Engl, in Bot. Jahrb. 44: 149 (1910);

Bot. Jahrb. 48: 482, t. 3F (1913); Pflanzenfam. ed. 2,19a:

438 (1931). Type: S.W. A., Kovies Mountains, Range 172

(B, holo.t; BOL!).

Dioecious shrub up to 4 m tall; trunk branching

repeatedly above soil level, stems appearing succose;

bark brown to green with blackish patches, peeling

locally in small white papery pieces; branchlets

glabrous. Leaves trifoliolate, glabrous; lamina up to

2,3 cm long; petiole up to 1 cm long; petiolules up

to 1 ,5 mm long; leaflets usually cordate but in some

cases orbicular or obovate, apex usually erparginate

but in some cases obtuse, base cuneate to obtuse,

margins finely lobed; terminal leaflet up to 1,8x1, 4

i

■:

1 I

j

\

)

:

J. J. A. VAN DER WALT

97

cm; lateral leaflets up to 1,3x1 cm. Flowers unisexual,

perigynous, appearing with the leaves in axillary

simple dichasial cymes or solitary; male flowers,

5-6 mm, usually larger than female flowers, 4,5-5 mm.

Bracteoles up to 1 mm long, triangular, glandular.

Pedicels 0,5-1 mm long, glandular. Calyx yellow to

green, fleshy, continuous with fleshy hypanthium,

glandular, lobes 1,5-2 mm long, apex acute. Petals

yellow to green, 2-3 mm long, glandular, inserted

on hypanthium. Disk adnate to hypanthium, cylin-

drical with 4 fleshy lobes. Stamens 8, 4 long stamens

up to 2,5 mm, inserted on top of disk lobes, 4 short

stamens up to 2 mm long, inserted on top of disk

between lobes; filaments flattened and lower part

broadened; staminodes in female flowers. Gynoecium:

rudimentary in male flowers; ovary half inferior,

glandular; style relatively long, glandular; stigma

2-lobed. Fruit 1,2x1 cm, ellipsoid, asymmetrical,

very much flattened; exocarp glabrous; mesocarp

very thin; putamen 1,1/0, 9 cm, ellipsoid, asym-

metrical, very much flattened, smooth; pseudaril

lacking. Fig. 100-105.

Fig. 100. — Commiphora ca-

pensis in Klein Helskloof

near Vioolsdrif, north-

western Cape (height I m).

Fig 102 — Commiphora capensis: A. branchlet with flowers;

B— D, leaves; E, branchlet with leaves and mature Iruits.

%

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 103. — Flowers of Commiphora capensis: A, male flower;

B, longitudinal section of male flower; C, female flower;

D, female flower with the calyx and corolla partly removed.

Fig. 104. — Fruit of Commiphora capensis: A, side-view of the

fruit; B, view of the less convex face of putamen; C, view

of the more convex face of the putamen; D, putamen

as seen from above.

Distinctive anatomical features of the stems and leaves

Young stems with numerous glandular hairs at

apex. Stems of 2,5 cm diameter: sclerenchymatous

pericycle-cylinder consisting of fibres and stone cells,

Fig. 105. — Geographical distribution of Commiphora capensis

in South Africa.

epithelium cells of resin ducts in xylem rays sur-

rounded by thin-walled cells. Leaves with a few

glandular hairs; petiole ± heart-shaped as seen in

transverse section, sclerenchymatous pericycle

present or absent, vascular bundles mainly abaxial

and distributed in the form of an arc as seen in

transverse section, in some cases 1-2 much smaller

bundles (usually only phloem strands) adaxially,

with a large number of stomata; terminal leaflet

typically isobilateral with 1-3 layers of palisade cells

ad- and abaxially, central mesophyll consisting of

± colourless cells, ad- and abaxial epidermis con-

sisting mainly of bulliform cells, with a large number

of evenly distributed stomata in the ad- and abaxial

epidermis.

Diagnostic features

Dioecious shrub, trunk branching repeatedly above

soil level, stems appearing succose; bark brown to

green with blackish patches, peeling locally in small

white papery pieces; branchlets with numerous

glandular hairs at apex but otherwise glabrous.

Leaves trifoliolate, glabrous, leaflets usually cordate

but in some cases orbicular or obovate, margins

finely lobed, terminal leaflet typically isobilateral.

Flowers perigynous, unisexual, calyx fleshy. Fruit

ellipsoid, very much flattened; exocarp glabrous;

mesocarp very thin; putamen smooth; pseudaril

lacking.

This species is confined to the semi-desert areas

of the north-western Cape and south-western parts

of South West Africa. It grows in the mountains

and kopjes in the vicinity of the Orange River from

Goodhouse westwards to the Richtersveld. These

areas are extremely dry and hot with a rainfall of

less that 80 mm per annum.

Cape. — 2816(Oranjemund): S.E. of Sendlingsdrift (-BB),

Pillans 5001 (BOL). 2817(Vioolsdrif): 37 km W.S.W. of

Vioolsdrif in Klein Helskloof (-CD), Van der Walt 111 ; 112.

2818(Warmbad): 5 km E. of Goodhouse (-CD), Van der Walt

126; 127.

C. capensis and C. cervifolia are closely related

species and the two have many characteristics in

common, especially as far as growth form, external

features of the stems and fruits are concerned.

Although both species have trifoliolate leaves, the

form of the leaflets differs considerably.

J. J. A. VAN DER WALT

99

As in the case of C. cervifolia, but to a lesser

extent, the living shoots, on being touched, exude

an aromatic secretion in such quantities that the

stems become wet.

The fruits are eaten by animals.

18. Commiphora cervifolia Van der Walt in J.S.

Afr. Bot. 37,3: 189 (1971). Type: North-western

Cape, 8 km S. of Vioolsdrif, Van der Walt 128

(PRE, holo. ; PRU).

Dioecious shrub up to 2 m tall; trunk branching

repeatedly above soil level, stems appearing succose;

bark greyish green to yellowish brown with dark

patches, not peeling; branchlets short and stout,

glabrous. Leaves trifoliolate, glabrous; lamina up to

1,5 cm long; petiole up to 5 mm long; leaflets small,

cultrate, usually irregularly lobed, apex acute to

obtuse, base cuneate, margins entire irrespective of

lobes; terminal leaflet up to 1 x0,2 cm; lateral leaflets

up to 0,8 x 0,2 cm. Flowers unisexual, perigynous,

appearing before the leaves in axillary dichasial cymes

up to 2 cm long or solitary; male flowers, 6-7 mm,

usually larger than female flowers, 5-6 mm. Bracteoles

up to 0,5 mm long, lanceolate, sparsely glandular.

Pedicels 1-1,5 mm long, sparsely glandular. Calyx

yellowish green to brown, fleshy, continuous with

fleshy hypanthium, sparsely glandular, lobes up to

2 mm long, apex acute. Petals yellowish green to

brown, 2-3 mm long, inserted on hypanthium. Disk

adnate to hypanthium, cylindrical with 4 fleshy lobes.

Stamens 8, 4 long stamens up to 3 mm long, inserted

on top of disk lobes; 4 short stamens up to 2,2 mm

long, inserted on top of disk between lobes; filaments

slender, subterete, lower part flattened and broadened ;

staminodes in female flowers. Gynoecium : rudimentary

in male flowers ; ovary half inferior, sparsely glandular ;

style relatively short, sparsely glandular; stigma

obscurely 4-lobed. Fruit 1,1x1 cm, ellipsoid, asym-

metrically flattened; exocarp glabrous; mesocarp very

thin; putamen 9x8 mm, ellipsoid, asymmetrically

flattened; pseudaril lacking. Fig. 106-111.

Fig. 106. — Commiphora cervi-

folia on the farm Gesel-

skapbank in the district of

Goodhouse, north-western

Cape (height ±0,5 m).

Fig. 107. — Close-up view of a branch of Commiphora cervifolia.

100

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

Fig. 108. — Commiphora cervifolia: A, branchlet with young

leaves and flowers; B, branchlet with leaves and mature

fruits; C, leaves.

Distinctive anatomical features of the stems and leaves

Young stems with numerous glandular hairs at

apex. Stems of 2,5 cm diameter: sclerenchymatous

pericycle-cylinder consisting of fibres and stone cells,

epithelium cells of resin ducts in xylem rays sur-

rounded by thin-walled cells. Leaves with a few

glandular hairs; petiole ± heart-shaped as seen in

transverse section, sclerenchymatous pericycle usual-

ly absent, 1-3 but usually only 1 vascular bundle

abaxially and in some cases 1-2 phloem strands

adaxially, with a large number of stomata; terminal

leaflet typically isolateral, mesophyll consisting mainly

of palisade cells with a few ± colourless cells cen-

trally, ad- and abaxial epidermis consisting mainly

of large bulliform cells, with a large number of

evenly distributed stomata in the ad- and abaxial

epidermis.

Diagnostic features

Dioecious shrub, trunk branching repeatedly above

soil level, stems appearing succose; bark greyish

green to yellowish brown with dark patches, not

peeling; branchlets short and stout, with numerous

glandular hairs at apex but otherwise glabrous.

Leaves trifoliolate, glabrous, leaflets small, cultrate,

usually irregularly lobed, margins entire irrespective

of lobes, terminal leaflet typically isobilateral. Flowers

perigynous, unisexual, calyx fleshy. Fruit ellipsoid;

exocarp glabrous; mesocarp very thin, putamen

smooth; pseudaril lacking.

Fig. 109. — Flowers of Commiphora cervifolia: A, male flower;

B, longitudinal section of male flower; C, female flower;

D, female flower with the calyx and corolla partly removed.

Fig. 110. — Fruit of Commiphora cervifolia: A, side-view of the

fruit; B, view of the less convex face of putamen; C, view

of the more convex face of putamen; D, putamen as seen

from above,

J. J. A. VAN DER WALT

private BAG X 101

PRETORIA 0001

-REPUBLIC OF SOUTH AFRICA

101

Fig. 111. — Geographical distribution of Commiphora cervifolia

in South Africa.

This species is apparently confined to the semi-

desert areas of the north-western Cape from Good-

house in the east to Yioolsdrif in the west. It occurs

on the arid mountains or kopjes in the vicinity of

the Orange River in areas with an annual rainfall

of less than 80 mm.

Cape. — 2817(Vioolsdrif): 8 km S. of Vioolsdrif (-DC), Van

der Walt 128. 2818(Warmbad: 13 km S.E. of Goodhouse

(-CD), Van der Walt 115. 2918(Gamoep): 56 km N.E. of

Okiep at Geselskapbank (-AA), Van der Walt 123.

Living shoots, on being touched, exude an aromatic

secretion in such quantities that the stems become

wet. The form of the leaflets with the irregular lobes

resembles the antlers of a stag, hence the name of

the species.

DISCUSSION OF THE RELATIONSHIPS OF THE

SPECIES

The South African species of Commiphora can

be divided on the basis of characteristics of the

flower, inflorescence, fruit and leaf as follows:

1. C. glandulosa , C. pyracanthoides, C. merkeri, C.

schimperi, C. africana , C. neglecta, C. mollis , C.

harveyi and C. marlothii have hypogynous flowers.

The disk of the flowers is fleshy and not adnate

to the calyx or corolla. A great part of the putamen

is covered by the pseudaril.

1.1 C. glandulosa, C. pyracanthoides, D. merkeri,

C. schimperi and C. africana have the fol-

lowing characteristics in common: The pu-

tamen is rugose, the pseudaril with or without

four distinct arms and the flowers are borne

in clusters or in reduced cymes.

1.11 C. glandulosa, C. pyracanthoides and

C. merkeri have simple leaves or tri-

foliolate leaves with two much smaller

lateral leaflets.

1.12 The leaves of C. schimperi and C.

africana are exclusively trifoliolate.

1.2 The putamen of C. neglecta, C. mollis, C.

harveyi and C. marlothii is smooth and the

pseudaril forms four distinct arms.

1.21 The filaments of C. neglecta, C. mollis

and C. harveyi are inserted on the

outside of the disk and the leaves are

trifoliolate or impari-pinnate.

1 . 22 The filaments of C. marlothii are

inserted on top of the disk and the

leaves are exclusively impari-pinnate.

2. C. edulis, C. woodii, C. zanzibarica, C. tenuipetio-

lata, C. angolensis, C. namaensis , C. gracili-

frondosa, C. capensis and C. cervifolia have

perigynous flowers. The disk of the flowers is

adnate to the hypanthium and in most cases not

fleshy. The pseudaril covers only the lower part

of the putamen or is completely absent.

2.1 The pseudaril of C. edulis, C. woodii and C.

zanzibarica is cupular without long arms or

lobes, the leaves are impari-pinnate and the

flowers are borne in long paniculate cymes.

2.2 The pseudaril of C. angolensis and C. tenui-

petiolata is cupular with two lobes on the

flattened faces of the putamen, the leaves are

trifiolate or impari-pinnate and the flowers

are borne in simple or compound dichasial

cymes.

2 . 3 The pseudaril of C. namaensis and C. gracili-

frondosa is cupular with two arms on the

seam of the putamen. The flowers are borne

singly or in short dichasial cymes. The leaves

of C. namaensis are simple or occasionally

trifoliolate, while those of C. gracilifrondosa

are trifoliolate.

2.4 C. capensis and C. cervifolia have no pseudaril,

the flowers are borne singly or in short

dichasial cymes and the leaves are trifoliolate.

In the revised classification of Wild (1959a) the

genus is divided into the subgenera Commiphora and

Opobalsamum. According to this division all the

South African species belong to the subgenus Commi-

phora, characterized by the fruit splitting into two

valves at maturity, four stamens which are distinctly

shorter than the remaining four and the presence

of four or eight disk lobes. It should be noted,

however, that the flowers of C. oblanceolata Schinz

(not mentioned by Wild) and C. gracilifrondosa have

only four stamens.

In the division of the subgenus into sections, Wild

did not make use of the fact that some species possess

hypogynous and other perigynous flowers. However,

all the South African species of his sections Commi-

phora and Africanae have hypogynous flowers, while

those of the sections Coriaceae and Spondioideae

are perigynous.

The distinction between the sections Commiphora

and Africanae by Wild is mainly based on the

structure of the pseudaril. Representatives of the

section Commiphora have a pseudaril forming four

arms, while representatives of the section Africana

have no pseudaril. Wild, however, maintained that

the pseudaril is only apparently absent in the latter

section because it is probably united too intimately

with the putamen to be visible. A remarkable simi-

larity exists between the South African species

belonging to the section Africanae (C. africana and

C. schimperi) and the species of the subsection

Pyracanthoides (C. glandulosa, C. pyracanthoides and

C. merkeri ). The putamen of all these species is

rugose ; the fruit of C. pyracanthoides, C. merkeri and

C. schimperi is apiculate; the length and shape of

the four arms of the pseudaril of C. africana, C.

glandulosa and C. pyracanthoides are alike, while the

pseudaril of C. merkeri and C. schimperi (although

102

THE SOUTH AFRICAN SPECIES OF COMMIPHORA

thin and membranous and only visible in fresh fruits)

covers almost the whole putamen. Furthermore, C.

glandulosa, C. pyracanthoides and C. schimperi can

have bisexual flowers. In addition all these species

have spine-tipped branchlets, flaking or peeling bark

and petioles which show marked anatomical simi-

larities. A reasonable deduction can be made that a

closer affinity exists between these species than Wild

realized.

C. marlothii is the only South African representative

of the section Commiphora where the filaments are

not adnate to the outside of the disk, but inserted

on top of the disk. This feature, as well as the presence

of long, paniculate cymes, impari-pinnate leaves and

medullary vascular bundles in the petiole, suggests

and affinity with the subsection Cupulares of the

section Spondioideae.

In agreement with the classification of Wild, a

cupular pseudaril is found in all the indigenous repre-

sentatives of the section Spondioides. Species of

the subsection Cupulares (C. edulis, C. woodii and

C. zanzibarica) have large pinnate leaves, perigynous

flowers and a pseudaril without long lobes or arms.

These three species show marked similarities in stem

and leaf anatomy, although medullary vascular

bundles are absent from the petiole of C. woodii.

The short lobes of the pseudaril of C. edulis suggest

a probable affinity with the section Commiphora.

The two closely related species, C. angolensis and

C. tenuipetiolata of the subsection Glaucidulae, have

a pseudaril with two lobes on the flattened faces of

the putamen. The hypanthium of both species is

relatively long. The external morphological similarities

of the leaves of the two species are reflected ana-

tomically; it is also of particular interest that the

dendritic crystals of hesperidin or diosmin only occur

in these two species.

The new species C. gracilifrondosa should be

placed in Wild’s subsection Pruinosae. As in C.

namaensis (subsection Pruinosae), the pseudaril of

C. gracilifrondosa is cupular with two arms on the

seam of the putamen.

The absence of a pseudaril is the outstanding

feature of representatives of the section Coriaceae,

and the new species C. cervifolia should be placed

in this section. Wild’s division of the Coriaceae into

the subsections Rangeanae and Teretifoliolatae is

partly based on the structure of the calyx. He described

the calyx of Rangeanae (includes C. capensis ) as

campanulate and the calyx of Teretifoliolatae as

broadly campanulate. This criterion for the sub-

division is now unsatisfactory because the calyx of

C. cervifolia is broadly campanulate, but this species

is undoubtedly closely related to C. capensis.

OPSOMMING

Hierdie ondersoek behels 'n taksonomies-morfologiese

studie van die 18 Commiphora-spes/es wat tot dusver

in Suid-Afrika ver samel is. Die belangrikste oogmerk

met die studie was om die verskillende spesies duidelik

te omgrens. Om dit te kon bereik is 'n volledige ana-

tomiese studie van die stingels en blare , asook ’«

organografiese studie van die stingels, blare, blomme

en ryp vrugte gemaak. Die organografiese kenmerke

van hierdie organe is met behulp van sketse en foto's

geillustreer.

'n Volledige uiteensetting van die taksonomiese

literatuur en geografiese verspreiding in Suid-Afrika,

word van elke spesie gegee. Die tipe-eksemplare van

al die spesies en hulle sinonieme is bestudeer, en in

toepaslike gevalle is lektotipes aangewys. Die onder-

skeidende anatomiese kenmerke en diagnostiese mor-

fologiese kenmerke van elke spesie, word gegee.

Twee sleutels wat lei tot die identifikasie van die

verskillende spesies is opgestel. Die een sleutel is

slegs op vegetatiewe kenmerke van die stingels en blare

gebaseer, terwyl die ander sleutel op alle morfologiese

kenmerke gebaseer is. Die moontlike verwantskappe

van die spesies word bespreek.

REFERENCES

Berg, O., 1862. Die Balsamodendron- Arten der Berliner

Herbarien. Bot. Ztg. 21: 161-164.

Brenan, J. P. M., 1953. Tropical African plants: Burseraceae.

Kew Bull. 1953: 104-107.

Burtt, B. D., 1935. Observations on the genus Commiphora

and its distribution in Tanganyika Territory. Kew Bull.

1935: 101-117.

Burtt Davy, J., 1932. A manual of the flowering plants and

ferns of the Transvaal with Swaziland. London: Longmans

Green.

Chiovenda, E., 1932. Flora Somala vol. 2. Modena: R. Orto

Botanico.

Codd, L. E. W., 1951. Trees and shrubs of the Kruger National

Park. Mem. Bot. Surv. S. Afr. 26.

Dale, I. R. & Greenway, P. J., 1961. Kenya trees and shrubs.

Nairobi: Buchanan’s Kenya Estates.

De Winter, B., 1968. Some observations on the genus Commi-

phora in South and South West Africa. Trees S. Afr.

20,1: 2-18.

Engler, H. G. A., 1883. Burseraceae. A. DC., Monogr. Phan.

4: 1-29.

Engler, H. G. A., 1893. Beitrage zur Flora von Afrika. 2.

Burseraceae africanae. Bot. Jahrb. 15: 95-102.

Engler, H. G. A., 1896. Burseraceae. Engl. & Prantl, Pflan-

zenfam. 3,4: 231-257.

Engler, H. G. A., 1913. Die Verbreitung der afrikanischen

Burseraceen im Verhaltnis zu ihrer systemalischen Gliede-

rung und die Einteilung der Gattung Commiphora. Bot.

Jahrb. 48 : 443-490.

Engler, H. G. A., 1915. Die Pflanzenwelt Afrikas vol. 3.

Leipzig: Engelmann.

Engler, H. G. A., 1931. Burseraceae. Engl. & Prantl, Pflan-

zenfam. ed. 2,19a: 405^156.

Harvey, W. H., 1862. Protium Wight & Arn. FI. Cap. 2: 592.

Hiern, W. P., 1896. Catalogue of Welwitseh's African plants

vol. 1, 1. London: Longmans & Co.

Irvine, F. R., 1961. Woody plants of Ghana with special reference

to their uses. London: Oxford University Press.

Jacquin, N. J., 1797. Plantarum rariorum horti caesarei schoen-

brunnensis descriptiones et icones vol. 2. Vienna: C. F.

Wappler.

Leenhouts, P. W., 1959. Revision of the Burseraceae of the

Malaysian area in a wider sense. 10a. Canarium Stick.

Blumea 9: 275-475.

Richard, A., 1832. Heudelotia nob. Gull., Perr. & A. Rich.

FI. Sen. 1: 150, t. 39.

Sinia, H. R., 1938. Zur Phylogenie der Fiederblatter der

Burseraceen und verwandter Familien. Annls. Jard. bot.

Buitenz. 48: 69-101.

Sonder, O. W., 1860. Balsamodendron Kunth. FI. Cap. 1: 526.

Sprague, T. A., 1927. Commiphora spp. Hook. Icon PI. 2:

t. 3105-3112.

Verdoorn, I. C., 1951. Newly described species: Burseraceae

Bothalia 6,1: 214-215.

Von Breitenbach, F., 1965. The indigenous trees of Southern

Africa, vol. 3. Pretoria: Government Printer.

White, F., 1962. Forest flora of Northern Rhodesia. London:

Oxford University Press.

Wild, H., 1959a. A revised classification of the genus Com-

miphora Jacq. Bol. Soc. Brot. ser. 2,33: 69-95.

Wild, H., 1959b. New and little known species from the Flora

Zambesiaca area. Bol. Soc. Brot. s6r. 2,33: 35-45.

Wild, H., 1963. Burseraceae. FI. Zamb. 2,1: 263-285,

Bothalia 11, 1&2: 103-106 (1973)

An analysis of the Flora of Natal

J. H. ROSS*

ABSTRACT

The recently completed Flora of Natal (Ross, Mem. Bot. Surv. S.Afr. No. 39, 1973) is analysed

numerically and a brief comparison is made with the total number of families, genera and species

recorded by Bews in his Flora of Natal and Zululand (1921). A number of diagrams illustrating the

proportional representation of certain elements of the Flora are provided. The families whose species

constitute more than 0,5% of the total number of species in Natal, and those genera containing 15

or more species, are tabulated. The marked inverse relationship between numbers of genera and

families is illustrated. Brief mention is made of some of the elements contributing to the richness of

the flora.

Natal is the smallest of the four provinces, its

91 385 square kilometres amounting to only 8% of

the area of the Republic of South Africa.

However, despite its size, the province has a rich flora

as revealed by the recently completed Flora of Natal

(Ross, Mem. Bot. Surv. S. Afr. No. 39, 1973).

The Flora of Natal (Ross, l.c.) accounts for 179

families, 1238 genera and 4826 species (these figures

exclude Pteridophytes and Bryophytes). These figures

compare with the 148 families, 901 genera and 3786

species recorded by Bews in his Flora of Natal and

Zululand (1921). Therefore, in the fifty years that

have elapsed since Bews published his work, an

additional 32 families, 337 genera and 1 040 species

have been recorded from Natal. With the continued

tendency in taxonomy to divide families it would be

unwise to place too much emphasis on the increase

in the number of families when comparing the present

figures with those recorded by Bews. Of the 32 ad-

ditional families not recorded by Bews (Bews regarded

Hippocrateaceae as a distinct family but it is now

included in Celastraceae), 13 are new records for

Natal while the remaining 19 are the consequence

of existing families being fragmented into new smaller

families. For example, the plants included in Liliaceae

by Bews are now placed in Liliaceae, Agavaceae

and Smilacaceae. The 13 new families recorded

from Natal are: Ruppiaceae, Zannichelliaceae, Naja-

daceae, Alismataceae, Pontederiaceae (introduced, but

naturalized), Casuarinaceae (introduced, but natura-

lized) Balanophoraceae, Basellaceae, Vahliaceae,

Dichapetalaceae, Elatinaceae, Canellaceae and Tur-

neraceae.

Of the 179 families, 4 families (2,24%) are gym-

nosperms and 175 (97,76%) are angiosperms. Of

these angiosperm families, 34 (19,40%) are mono-

cotyledons and 141 (80,60%) are dicotyledons (see

Fig- 1).

*Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Fig. 1. — Proportional representation of monocotyledon families

and dicotyledon families in the Flora of Natal.

Of the 4 826 species recorded, 14 (0,29%) are

gymnosperms and 4 812 (99,71%) are angiosperms.

Of these 4 812 angiosperm species, 1 308 (27,16%)

are monocotyledons and 3 504 (72,84%) are dicotyle-

dons (see Fig. 2). The families whose species contri-

bute more than 0,5% of the total number of species

are listed in order of numerical importance in Table

1. The number of genera present in these families

is also reflected in Table 1 but as family position is

determined by the total number of species, the ar-

rangement of genera follows no strict sequence.

104

AN ANALYSIS OF THE FLORA OF NATAL

Table 1. — Synopsis of the Natal families whose species com-

prise more than 0,5% of the total number listed in order

of numerical importance, together with the number of

genera in each family.

The largest family is Compositae with 549 species

(11,37%) followed by Leguminosae with 420 species

(8,70%) and Poaceae with 414 species (8,57%).

These three largest families contribute 1 383 species

or 28,64% of the total number of species in Natal,

while the ten largest families contribute 2 605 species

or 53,98% of the total. The proportional represen-

tation of angiosperm species in families with over

175 species each in the Flora of Natal is shown in

Fig. 2.

Thirty-seven families (20,67%) in Natal are

represented by only one species, 21 families by two

species, 13 families by three species, three families

by four species, nine families by five species, 11

families by six species and five families by seven

species. Only 80 of the 179 families in Natal have

more than seven species. This proportional represen-

tation of the angiosperm families with seven or

fewer species each is shown in Fig. 3.

Fig. 2. — Proportional representation of angiosperm species in

families with more than 175 species each.

Fig. 3. — Proportional representation of families with seven or

fewer species each.

Of the 1 238 genera, 4 (0,32%) are gymnosperms

and 1 234 (99,68%) are angiosperms. Of these 1 234

angiosperm genera, 3 14 (25 , 45 %) are monocotyledons

and 920 (74,55%) are dicotyledons (see Fig. 4).

Although Compositae has by far the largest number

of species, Poaceae has the largest number of genera.

Poaceae with 137 genera (11 ,07%), Compositae with

112 genera (9,05%) and Leguminosae with 89

genera (7,19%) contribute 338 genera or 27,31%

of the total number of genera. The ten largest families

contribute 588 genera or 47,10% of the total number

of genera. The proportional representation of angio-

sperm genera in families with over 34 genera each is

shown in Fig. 4.

J. H. ROSS

105

Fig. 4. — Proportional representation of Angiosperm genera in

families with more than 34 genera each.

Sixty-eight families (37,99%) in Natal are re-

presented by only one genus, 29 families by two

genera, 13 families by three genera, 13 families by four

genera, 12 families by five genera and four families

by six genera. Only 44 of the 179 families in Natal

have more than six genera. The proportional represen-

tation of the angiosperm families with five or fewer

genera each is shown in Fig. 5.

This marked inverse relationship between the

number of genera and families is shown in Fig. 6.

It is quite apparent from this figure that most families

have few genera and only very few families have

Fig. 5. — Proportional representation of families with five or

fewer genera each.

many genera. A similar inverse relationship exists

between the number of species and genera.

The ratio of genera to species in Natal is 1: 3,89

which implies a high proportion of genera with

only one or two species. This figure for Natal is

intermediate between the ratio of 1: 3,52 for the

Flora of South West Africa (calculated from the

figures supplied by Merxmuller in Mitt. Bot. Munchen

10: 75, 1971) and the ratio of 1: 4,21 for the revised

edition of the Flora of West Tropical Africa (cal-

culated from the figures supplied by Hepper, l.c. :

25).

140

120

< 100

CL

LU

Z

lu 80

O

O 60

CL

UJ

CO 40

£

3

Z 20

20

40

60

80 100

FAMILIES

Fig. 6. — Histogram showing

the marked inverse rela-

tionship between number

of genera and families.

106

AN ANALYSIS OF THE FLORA OF NATAL

The genera in Natal with the largest number of

species are given in order of numerical importance

in Table 2. The two largest genera, Helichrysum

Mill, and Senecio L., both belong to Compositae,

but thereafter a number of other families, for example,

Crassulaceae, Leguminosae, Asclepiadaceae, Eup-

horbiaceae, Liliaceae and Cyperaceae feature.

Table 2. — Synopsis of the Natal genera with 15 or more

species listed in order of numerical importance.

1. Helichrysum

2. Senecio

3. Crassula

4. Indigofera

5. Schizoglossum

6. Euphorbia

7. Aloe

8. Cyperus, Tephrosia

10. Lotononis

11. Eragrostis

12. Erica, Wahlenbergia, Polygala

15. Asclepias, Berkheya, Kniphofia, Solanum. . .

19. Disa, Acacia

21. Eulophia, Rhus, Sutera

24. Hibiscus, Ipomoea, Rhynchosia

27. Argyrolobium, Delosperma, Pavetta, Plec-

tranthus, Stachys, Vernonia

33. Scirpus

34. Digitaria, Habenaria, Selago, Streptocarpus,

Thesium, Panicum

40. Disperis, Hypoxis

42. Pelargonium

43. Ceropegia, Hermannia, Lobelia

46. Gladiolus, Sebaea, Sporobolus

49. Asparagus, Crotalaria, Ficus, Mariscus

53. Brachystelma

54. Alepidea, Cyrtanthus

56. Clutia, Loranthus, Phyllanthus

109

92

55

49

48

41

39

37

33

32

31

30

29

28

26

25

24

23

22

21

20

19

18

17

16

15

Of the 4826 species recorded in Natal, 568 species

(11,77%) appear in the National Tree List (De

Winter and Vahrmeijer 1972). As a number of the

smaller woody shrubs are excluded from the National

Tree List, the number of woody species in Natal is

actually slightly higher than the 11,77% recorded

and probably approaches 15% of the total number

of species.

The 179 families, 1 238 genera and 4 826 species

in Natal compare with the 202 families, 1 669 genera

and 7 014 species (figures adjusted to exclude ferns)

recorded by Hepper (Mitt. Bot. Munchen 10: 24,

1971) from the area delimited for the Flora of West

Tropical Africa. Some indication of the richness

of the Natal flora can be gained when it is appreciated

that the area delimited for the Flora of West Tropical

Africa is almost fifty times as large as Natal.

The richness of the Natal flora is due in part to

the migration of subtropical elements southwards to

the Tongaland plain, the area east of the Lebombo

mountains. Four families (Ruppiaceae, Zannichel-

liaceae, Alismataceae and Vahliaceae), 77 genera

(6,22%) and 278 species (5,76%) out of the total

flora are restricted to this Tongaland plain. A number

of the genera, for example, Ceriops Arn., Inham-

banella (Engl.) Dubard, Lunmitzera Willd., Newtonia

Baill., Schlechterina Harms and Thalassodendron

Den Hartog are found nowhere else in the Republic.

Similarly, many of the 278 species restricted to this

plain are not recorded elsewhere in the Republic.

This Tongaland plain is an extremely interesting area

and one from which new records continue to come.

In addition to a subtropical element, the Natal

flora also contains a temperate element which extends

from the south-western Cape along the Drakensberg

to the mountains of tropical Africa and Europe. In

the Natal Drakensberg the flora becomes increasingly

temperate in character with increasing altitude. This

is well illustrated at high altitudes by the grasses where

temperate genera such as Danthonia DC., Festuca

L., and Pentaschistis Stapf are completely dominant.

Some Cape genera, for example, Erica L. and Protea

L., also enrich our flora. Other Cape elements enter

southern Natal along the coast, for example, Brunia-

ceae, which is represented in Natal by a solitary

species of Raspalia Brongn.

Natal is still relatively poorly collected botanically

and new records and new species continue to be

found. Indeed, since the Flora of Natal was com-

pleted in April 1971, five new generic records and

10 new species records (plus several undescribed

species and a new generic record for South Africa)

have come to light. A number of species are known

only from the type collection while many other records

for the province are based on a single gathering. The

current conservation status of many of our species,

particularly the herbaceous species, is unknown and

many are probably endangered. At least one species,

the orchid Zeuxine africana Reichb.f., known only

from the Durban Bay area, is thought to be extinct.

Clearly there is an urgent need for more extensive

and more intensive collecting in Natal.

Bothalia 11,1 & 2: 107-113 (1973)

Towards a Classification of the African Acacias

J. H. ROSS*

ABSTRACT

Some of the first attempts to subdivide the genus Acacia Mill, as a whole are discussed briefly.

Bentham’s work, in which the subdivisions of the genus were for the first time given names, is con-

sidered and his two series into which the African species fall are reproduced. The characters employed

by various workers to divide the African species into two main groups are mentioned and the advantages

of using each of these characters for the first dichotomy in a key are discussed. Recent work on pollen

morphology and seedling morphology is correlated with general morphology. A proposal put forward

by Guinet to divide Acacia into three large genera on the basis of pollen morphology is briefly dis-

cussed. Certain modifications to Bentham's series Vulgares and Gummiferae are suggested.

INTRODUCTION

Philip Miller (Gard. Diet, abridg. ed. 4, 1754) was

the first author to employ the name Acacia in a

generic sense subsequent to 1753 and is, therefore,

regarded as the author of Acacia. Miller’s generic

description, which is based on the “Egyptian Thorn”

[A. nilotica (L.) Willd. ex Del.], is as follows:

“It hath a tubulous flower consisting of one leaf, with

many stamina or threads, which are many of them collected

into a kind of sphere or globe: the pointal of the flower

afterwards becomes a pod in which are included several

seeds, each of which is separated by transverse diaphragms,

and are generally surrounded with a sweetish pulp.”

Acacia Mill, is a large tropical or subtropical genus

of about 850-900 species. The vast majority of species

(± 620) are found in Australia, while many (± 115)

occur in Africa, many in America, and fewer species

in Asia. Europe is the only large geographical area

which is devoid of the genus, while there are no

indigenous species in New Zealand despite its close

proximity to Australia.

EARLY GENERIC SUBDIVISIONS

Following his generic diagnosis, Miller (l.c.)

enumerated and discussed 24 species under Acacia

but made no attempt to divide the species into groups.

As the generic limits of Acacia were very broad it

is not surprising that a number of the species enume-

rated under Acacia are no longer referable to the

genus as it is at present defined.

Lamarck, Encycl. 1: 8 (1783), listed 58 species

under Acacia and divided the species into two groups

depending upon whether or not the stipules were

spinescent. The two groups were not prefaced by

any indication of rank.

Willdenow, Sp. PI. ed 4, 4: 1049 (1806), listed

102 species under Acacia and was among the first

to attempt to draw up a system of classification of

the Acacia species. Willdenow divided the species

into seven groups on the basis of vegetative characters

but, once again, the groups were not prefaced by any

indication of rank.

De Candolle, Prodr. 2: 448 (1825), listed 258 species

under Acacia and divided the species into four main

Sections. His Sections, which were not named, were

•Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X 101, Pretoria.

founded essentially on leaf characters with the nature

of the stipules and of the inflorescences being employ-

ed to distinguish groups of species within the Sections.

Sprengel, Syst. Veg. 3: 133 (1826), listed only 188

species under Acacia and divided the species into

three groups on the basis of leaf characters. His first

group was subdivided on the nature of the inflores-

cence, the second was subdivided on whether or not

the plants were armed, while the third group was

subdivided on whether or not the plants were armed,

and then on the nature of the inflorescence. Like the

preceding generic subdivisions, Sprengel’s groups were

not prefaced by any indication of rank.

In 1842 Bentham published his notes on Acacieae

in Hook. Lond. J. Bot. 1: 318-392. In his treatment

of Acacia , Bentham (l.c.: 319) wrote:

“A dry two-valved pod has been the character hitherto

chiefly relied upon for the distinction of this extensive genus;

but this has not only the great inconvenience that there are

but few cases where the ripe pod can be observed, but also

it is often even then very uncertain, and not at all consonant

with general habit and other characters. Many species, precisely

similar in almost every other respect, have very different

pods, and the same pod may be found in two Mimoseae

having scarcely any other point in common. I have, therefore,

thought it better to derive the principal character from the

flower, and by excluding all species with definite stamens, or

with the filaments connected in a cylindrical tube, it has

appeared to me that the genus Acacia becomes more natural

than it could be made by any other limitations hitherto

proposed, and certainly very much more clearly and easily

defined. Even in the subdivision of the genus, imperfectly

as a great number of the species are as yet known to us, it

becomes necessary to rely more on foliage and habit than

on the pod, however diversified may be the forms assumed

by that organ.”

Bentham divided Acacia into six Series, the Series

being delimited primarily on foliage, on whether

or not the plants were armed, and, if armed, upon

whether or not the stipules were spinescent. The

inflorescence played a far less important role in his

division of the genus than the vegetative characters.

For the first time the generic subdivisions were for-

mally given names. It is perhaps surprising that

Bentham only accorded his subdivisions the rank of

Series and did not, for example, recognize them as

subgenera. All of the African species fall into two

of Bentham’s Series, namely, Gummiferae and Vul-

gares.

108

CLASSIFICATION OF THE AFRICAN ACACIAS

Bentham’s paper was a most significant and im-

portant contribution because, for the first time, the

generic limits of Acacia were clearly defined and those

species which did not belong in it were excluded.

Prior to 1842 the generic limits of Acacia were so

vague and ill-defined that a rather heterogeneous

assemblage of plants was included under Acacia. The

early subdivisions of the genus, therefore, all suffered

from the same deficiency in that they had to make

provision for too many species which were actually

not referable to Acacia. Indeed, the generic limits

of Acacia have not been seriously in doubt since

Bentham’s work in 1842. His subsequent work in

Flora Australiensis 2: 301-421 (1864), in Genera

Plantarum 1: 594 (1865) and in his “Revision of

the Suborder Mimoseae” in Trans. Linn. Soc. Lond.

30: 336-664 (1875), served to clarify the genus further.

The generic subdivisions of Acacia in Bentham’s

revision in Trans. Linn. Soc. Lond. 30 (1875) were

based on his earlier treatment in 1842 but there are

a number of significant alterations. Bentham (l.c.

444, 1875) wrote:

“I have not either been able, in this my third careful

revision of the species, to divide it into sections founded

upon any character derived from the flowers or fruits; I

therefore here repeat the series, based upon habit, inflorescence

and geographical distribution, which are given in the Genera

Plantarum, subdividing them into subseries and minor

groups still less definitely limited, but of which the following

may be taken as the chief characters, neglecting minor

exceptions.”

Because of the relevance of Bentham’s Series

Gummiferae and Vulgares to the following discussion

they are reproduced below :

Series 4. — GUMMIFERAE. Arbores v. frutices non scan-

dentes nec acideati. Folia bipinnata. Stipulae nonnullae v. omnes

spinescentes. Capitula globosa v. spicae cylindraceae, ad axillas

v. in racemo terminali brevi pedunculata, rarius paniculala.

Tropicae v. subtropicae utrinsque orbis.

Subseries 1. — Summibracteatae . Involucellum annulare sub

capitulo ipso v. ab eo parum distans. Capitula globosa. Legumen

crassum, turgidum v. rarius planum, non v. vix dehiscens,

intus inter semina farctum. Americanae v. Africanae, una

cosmopolitana.

Subseries 2. — Medibracteatae. Involucellum in medio pedun-

culo v. paullo altius v. inferius situm, rarius O. Capitula globosa.

Legumen bivalve.

A. — Heteracanthae. Spinae minores recurvae, auctae rectae.

Legumen demum turgidum v. subteres. — Species gerontogeae.

B. — Moniliformes. Spinae omnes rectae, v. minores recurvae.

Legumen planum, saepe crassum, inter semina regulariter

constrictum v. depressum. Species gerontogeae.

C. — Thyrsiflorae. Spinae omnes rectae. Inflorescentia ter-

minalis, subaphylla, simplex. Legumen planum, continuum.

Species Africanae.

D. — Pubiflorae. Spinae rectae. Pedunculi axillares. Flores

pubescentes. Legumen planum, continuum. Species geronto-

geae.

E. — Normales. Spinae rectae. Pedunculi axillares. Flores

glabri v. parce puberuli. Legumen saepius planum, valvis

tenuibus. Pleraeque Africanae; paucae Indicae, Australicae v.

Mexicano-Texanae.

F. — Paniculatae. Spinae rectae. Panicula terminalis, sub-

aphylla. Species Asiaticae.

Subseries 3. — Basibracteatae . Involucellum nullum nisi ad

basin pedunculi. Spicae cylindraceae v. elongatae, v. in una

specie (A. sphaerocephala) globosa. Americanae, Africanae

v. Asiaticae.

Series 5. — VULGARES. Arbores v. frutices interdum scan-

dentes. Stipulae non spinescentes. Aculei infrastipulares sparsi

v. O. Folia bipinnata, petiolo saepissime glandulifero.

Subseries 1. — Gerontogeae spiciflorae.

A. — Triacanthae. Aculei terni, infrastipulares cum infra-

foliaceo.

B. — Diacanthae. Aculei gemini, infrastipulares.

C. - — Ataxacanthae . Aculei sparsi.

Subseries 2. — Americanae spiciflorae. Aculei sparsi v. O.

Subseries 3. — Americanae capitulatae.

Subseries 4. — Gerontogeae capitulatae.

There have subsequently been many criticisms of

Bentham’s classification. For example, Newman in

J. Linn. Soc. Bot. 49: 133-143 (1933), considered

Bentham’s classification to be too static in concept.

However, it must be borne in mind that almost a

century has elapsed since Bentham produced his final

classification in 1875, and that the number of species

now included in the genus is double the number

that he made provision for. Bentham was well aware

of many of the deficiencies, but many of his decisions

were, of necessity, based on specimens which, by

modern standards, would be considered quite inade-

quate. Despite the criticism there has been no com-

prehensive account of the genus as a whole, nor

any attempt to subdivide the entire genus since 1875.

Indeed, Bentham’s major subdivisions of the genus

have stood the test of time and evidence will be led

later in support of their retention.

Britton and Rose, N. Amer. FI. 23, 2: 84 (1928),

divided the American Acacieae into a number of

genera on the basis of pod characters, but posterity

has rejected most of these new genera. Newman (l.c.

137) drew up a phylogenetic classification to the

acacias, mostly the Australian species, based primarily

on the inflorescence. Each of the three groups thus

formed was subdivided on flower-group and then

on the foliar types. However, this classification has

not found favour either.

As the vast majority of the species occur in Aus-

tralia and are not of immediate concern to a study

of the African species, the general subdivisions

of the genus as a whole will not be considered further.

The methods employed to subdivide the African

species will now be considered.

SUBDIVISIONS OF THE AFRICAN SPECIES

A. Richard, Tent. FI. Abyss. 1: 237 (1847), divided

the Acacia species into two groups on the basis

of whether or not the stipules were spinescent.

Although the two groups were not prefaced by any

indication of rank, they correspond to Bentham’s

Gummiferae and Vulgares.

Harvey in FI. Cap. 2: 279 (1862) used Bentham’s

subdivisions of 1842, the species being placed either

in Gummiferae or in Vulgares. The nature of the

inflorescence was employed to subdivide the species

within each of these series. Engler in Bot. Jahrb.

10: 16 (1888), Taubert in Engl. Pflanzenfam. 3,

3: 108 (1891) and Glover in Ann. Bolus Herb. 1:

143 (1915) also followed Bentham’s classification.

Oliver in FI. Trop. Afr. 2: 337 (1871) did not

follow Bentham, but based his primary division

of the species on the nature of the inflorescence.

Two broad groups were recognized, namely, those

species with spicate inflorescences and those with

capitate inflorescences. The second dichotomy within

each group was based upon whether or not the

stipules were spinescent. This appears to have been

the first departure from Bentham’s classification for

the African species and it was in time to be followed

by the authors of nearly all of the major regional

African floras. Similar keys to the species using the

inflorescence for the first dichotomy were used by

Hutchinson & Dalziel in FI. W. Trop. Afr. 1 (2):

359 (1928); Burtt Davy, FI. Transv. 2: 333 (1932);

Torre in Consp. FI. Angol. 2: 269 (1956); Keay in

FI. W. Trop. Afr. 1 (2), ed. 2: 496 (1958); Brenan

in FI. Trop. E. Afr. Legum. — Mimos. : 49 (1959);

J. H. ROSS

109

in FI. Zamb. 3 (1): 53 (1970); F. White, For. FI.

N. Rhod.: 78 (1962) and Schreiber in FI. S.W. Afr.

58: 2 (1967). Gilbert & Boutique in FI. Congo Belg.

3: 146 (1952) alone in recent years have used stipular

spines versus prickles for the first dichotomy in a

key.

E. G. Baker in his Leguminosae of Tropical Africa

3: 815 (1930) also used the inflorescence for dividing

the species into two main groups. The species in

the group with spicate inflorescences were divided

according to whether or not the stipules were spine-

scent. Those species with prickles were then sub-

divided into four series depending upon whether

the prickles were in threes at the nodes, in pairs at

the nodes, solitary, or scattered along the inter-

nodes. The group with capitate inflorescences was

divided according to whether or not the stipules

were spinescent, and then on the type of pods.

Eighteen series in all were recognized by Baker.

The African species, therefore, have been divided

in the past by various authors into two main groups

on the basis of two different characters. Some authors

have given preference to whether or not the stipules

are spinescent as the primary character in separating

the two groups, while other authors have employed

the inflorescence. What then are the advantages

offered by each of these characters, and which

provides a more natural systematic arrangement of

the species?

1. The inflorescence

The inflorescence is a very convenient character

to employ for dividing the Acacia species into two

main groups. It is usually far easier to decide whether

the inflorescence is capitate or spicate than it is to

decide whether or not the stipules are spinescent

because in many species the stipules are very small

and rapidly deciduous. However, despite the con-

venience of capitate versus spicate inflorescences,

there are some difficulties as there is no absolute

distinction between the two groups. For example,

Brenan in F.T.E.A. Legum.-Mimos.: 80 (1959)

points out that A. dolichocephaJa Harms “seems to

bridge the gap between the capitate- and spicate-

flowered groups of Acacias”. In A. mellifera (Vahl)

Benth. subsp. mellifera from tropical Africa the

inflorescence is spicate while in subsp. delinens

(Burch.) Brenan from southern Africa the flowers

are in subglobose or ellipsoid heads and super-

ficially are easily mistaken for those of a capitate-

flowered species.

Several characters lend support to the contention

that the two groups, obtained by using the inflores-

cence for dividing the species, may be natural. These

are:

a. Almost all of the African species with spicate

inflorescences are armed with recurved prickles while

nearly all of the species with capitate inflorescences

are armed with stipular spines. The exceptional

species with stipular spines among the spicate-

flowered species and the species with recurved prickles

among the capitate-flowered species are readily and

conveniently separated from the bulk of the species

within each of these two main groups.

b. All of the species with spicate inflorescences

have pale yellowish-white flowers (except for A.

persiciflora Pax and A. galpinii Burtt Davy which

5292-8

have red or purplish calyces and corollas). Most

of the species with capitate inflorescences have deep-

or golden-yellow flowers, relatively few species

having pale yellowish-white, white or rarely pinkish

or purple flowers.

c. Further support derives from pollen-morphology

(Coetzee in S. Afr. J. Sci. 52: 23, 1955) although

admittedly only a small proportion of the species

have been investigated. Coetzee examined 25 southern

African species and found that the species with

capitate inflorescences had pollen grains provided

with furrows whereas the species with spicate in-

florescences had pollen grains without furrows. Two

exceptions were found, namely, A. detinens and A.

pennata which “have a capitate inflorescence and

have pollen in which no furrows occur”. A. detinens

(A. mellifera subsp. detinens ) is not really an ex-

ception because the inflorescence is ellipsoid and

the species belongs to the group with spicate in-

florescences. However, the anomalous pollen-mor-

phology of A. pennata cannot be explained on the

basis of capitate versus spicate inflorescences. It

must be mentioned here that all of the species with

spicate inflorescences examined by Coetzee are armed

with recurved prickles and that none has stipular

spines.

Van Zinderen Bakker and Coetzee (South African

Pollen Grains 3: 115, 1959) furnish the results of

an investigation into the pollen-morphology of

28 species. Once again all of the species with spicate

inflorescences (except for A. pennata) had pollen

grains without furrows while all of the species with

capitate inflorescences had pollen grains with fur-

rows.

d. Although the chromosomes of only a relatively

small number of species have been investigated, the

results do suggest certain tendencies. Darlington

and Wylie, Chromosome Atlas of Flowering Plants:

151 (1955), record that all species with spicate in-

florescences investigated have a diploid chromosome

number of 26 (except for A. laeta R. Br. ex Benth.

in which 2n = 52) whereas in the capitate-flowered

species 2n = 52. It has been suggested (Brenan l.c.

83, 1959) that A. laeta may be the outcome of hy-

bridization and this could perhaps account for this

chromosome number.

e. The seeds also provide support for using the

inflorescence for the first dichotomy in dividing the

species into two main groups. The seed of all Acacia

species show on each face an area, usually more or

less elliptic or oblong in shape, bounded by a fine

line which frequently appears as a fissure in the

testa. The size and shape of this area, termed the

areole (Brenan l.c. 1, 1959), are often important

taxonomically. The line is usually broken opposite

the micropyle although in some species the line

is almost continuous. Areole shape often provides

a useful means of distinguishing between the seed

of capitate- and spicate-flowered species. In spicate-

flowered species the areole is typically horse-shoe

shaped, fairly small, and occupies the central area

of the seed. In the capitate-flowered species the

areole is larger and conforms to the outline of the

seed. A. albida is anomalous among the spicate-

flowered species in having a large subcircular-lenti-

cular areole as in the capitate-flowered species.

110

CLASSIFICATION OF THE AFRICAN ACACIAS

2. The stipules

The presence of spinescent stipules or of non-

spinescent stipules provides a useful means of dividing

the African species into two main groups. As each

species is armed either with stipular spines, or with

non-stipular prickles, the distinction between the

two groups is absolute. A key in which the first

dichotomy is based on stipular spines versus non-

stipular prickles has the advantage that it can be

used for flowering, fruiting, or even sterile specimens.

Only very rarely, and in very few species, is an entire

plant unarmed.

Several characters lend support to the contention

that the two groups, obtained by using the nature

of the stipules for the first dichotomy, may be natural.

These are:

a. Nearly all of the species armed with prickles

have spicate inflorescences (except for a group of

climbers with recurved prickles scattered along the

internodes) while nearly all of the species armed

with stipular spines have capitate inflorescences.

b. All of the species armed with prickles have

pale yellowish-white flowers (except for A. galpinii

and A. persiciflora which have red or purplish

calyces and corollas). Most of the species armed with

stipular spines have deep- or golden-yellow flowers,

relatively few species having pale yellowish-white,

white or rarely pinkish or purple flowers. There is

a slightly better correlation between flower colour

and the nature of the stipules than there is between

flower colour and the type of inflorescence.

c. If Coetzee’s pollen-morphology studies are cor-

related with the nature of the stipules, it is found

that all the species armed with stipular spines have

pollen grains with furrows whereas the species armed

with prickles have pollen grains without furrows.

On this basis, the position of A. pennata is no longer

anomalous as its pollen structure is then the same

as that of all the other species armed with prickles.

However, Van Zinderen Bakker and Coetzee (l.c.)

studied a larger number of species including A.

albida. When their results are correlated with the

nature of the stipules it is found, once more, that

all of the species armed with prickles have pollen

grains without furrows while all of the species with

stipular spines (except for A. albida) have pollen

grains with furrows. The position of A. albida within

the genus, which is anomalous in several respects,

will be considered later in more detail.

Studies on pollen-morphology by Guinet (Inst.

Fr. Pondichery, Trav. Sec. Sci. Tech. 9, 1969) also

provide evidence in support of the division of the

African acacias on the basis of spinescent stipules

versus non-spinescent stipules. Guinet found a

definite correlation between the nature of the stipules

and the pollen-morphology in Acacia, Dichro-

stachys and Prosopis: in these genera it was found

that stipular spines are correlated with a high degree

of differentiation in the pollen grains (for example,

the presence of furrows in the exine). Conversely,

in species of Acacia without spinescent stipules

a low degree of differentiation in the pollen grains

was found. It should be emphasized here that there

is no similar correlation between pollen morphology

and the type of inflorescence.

d. When the chromosome numbers recorded by

Darlington and Wylie are correlated with the nature

of the stipules it is found that in all of the species

armed with stipular spines 2n = 52 (except for A.

albida in which 2n = 26) while in all of the species

armed with prickles 2n = 26 (except for A. laeta

in which 2n = 52).

i

Presumably the higher chromosome complement

of the species with spinescent stipules has given

these species greater genetic plasticity and the ability

to exploit new habitats. Could this be the reason

for the preponderance of species armed with stipular

spines in Africa? There is no real evidence that the

species with spinescent stipules are more widespread

in Africa than species with non-spinescent stipules.

For example, A. ataxacantha DC. and A. Senegal

(L.) Willd. are probably as widespread in Africa

as any of the species with stipular spines. Endemics

are found among species armed with spinescent

stipules and among species with non-spinescent

stipules. However, there is a suggestion that some 1

of the species with spinescent stipules are able to

occupy more adverse habitats than species with 1

non-spinescent stipules.

e. If areole shape is correlated with the nature

of the stipules, it is found that nearly all of the species

with non-spinescent stipules (except for those species

with prickles scattered irregularly along the inter-

nodes) have small horse-shoe shaped areoles, while

nearly all of the species armed with spinescent

stipules have large subcircular-lenticular or quadrate

areoles conforming to the shape of the seed.

/. There is a tendency in the species armed with (

prickles for the veins to run transversely across the

pods while in the species armed with stipular spines

the veins tend to run longitudinally. In many species,

for example in A. nilotica (L.) Willd. ex Del. and

A. giraffae Willd., no distinct venation is visible ,

on the pods, but in those species where the venation

is conspicuous there is a tendency for the above

distinctions to prevail. The species with prickles ,

scattered irregularly along the internodes have

transversely venose pods irrespective of whether the

inflorescences are spicate or capitate.

i

g. The development of secondary leaves (Ross, The

Acacia species of Natal: 6, 1971) seems to be almost

restricted to the species armed with stipular spines.

The secondary leaves are fascicular and arise from

dwarf lateral shoots at the nodes. These secondary

leaves are thought to be an adaptation which enables

a plant to produce new leaves, particularly in an

unfavourable season, without first having to draw

on its reserves to produce new branchlets to carry

primary leaves. Even most of the spicate-flowered

species armed with spines produce secondary leaves.

It. A study of the seedling morphology of some

of the African acacias by Vassal (Trav. Lab. For.

Toulouse, Tome 1, Vol. 8, 3, 1969) revealed differences

between those species armed with stipular spines and

those species armed with prickles. Once again, these

results support the division of the species on the

nature of the stipules rather than on the type of

inflorescence.

J. H. ROSS

111

DISCUSSION AND CONCLUSIONS

It is apparent from the above that the division

of the African species into two main groups on the

basis of the inflorescence has certain advantages,

while a division of the species into two groups on

the nature of the stipules likewise has certain

advantages. Some characters are better correlated

with the type of inflorescence than with the nature

of the stipules, and vice versa. The anomalous species

recorded when the species are divided on the basis

of the inflorescence are often explained away when

the species are divided on the basis of the stipules,

although such a division in turn usually creates a

further set of anomalous species. For example, when

Van Zinderen Bakker and Coetzee’s pollen-mor-

phology studies are correlated with the type of

inflorescence, it is found that all of the species with

pollen grains without furrows have spicate inflore-

scences (except for A. pennata) and all of the species

with pollen grains with furrows have capitate in-

florescences. However, when their studies are cor-

related with the nature of the stipules it is found

that all of the species with prickles (including A.

pennata) have pollen grains without furrows while

all of the species armed with stipular spines (except

for A. albida ) have pollen with furrows.

If the African acacias are divided according to

the type of inflorescence, within each of the groups

thus formed some of the species are armed with

stipular spines and some of the species are armed

with non-stipular prickles. Conversely, if the species

are divided on the nature of the stipules, within each

group some of the species have spicate inflorescences

and some of the species have capitate inflorescences.

In other words, if the division of the species on the

basis of the inflorescences is considered a natural

one, then spinescent and non-spinescent stipules

must have developed within each group, whereas

if the division on the nature of the stipules is con-

sidered a natural one, then capitate and spicate in-

florescences must have developed within each group.

Apparently, therefore, certain characters must have

evolved at least twice during the development of

the African acacias.

For use in regional African floras the inflorescence

is a very convenient character to employ for the

primary division of the species into two main arti-

ficial groups. However, when a more natural system-

atic division is sought, it is clear that it is the nature

of the stipules that must be employed for separating

the two main groups of species. This view is in

agreement with the work of Bentham and the two

groups thus obtained correspond to his series Gum-

miferae and Vulgares. Certain modifications within

each of Bentham’s series are, however, desirable.

At this stage it is not intended to formally propose

a system of classification of the African species,

but rather to put forward some tentative suggestions.

In any event the taxonomy of many of the species,

particularly in north-east tropical Africa, is still

confused so it seems desirable to wait until these

problems are resolved so that all of the species can

be accurately placed within the framework of a

classification. Furthermore, it seems pointless to

draw up a classification of the African species in

isolation without taking into account the species

in other areas of distribution, particularly those in

Australia which, after all, constitute the vast majority

of the genus. A classification of the African species

must fit within the framework of a classification

of the genus as a whole. The Australian species

are currently being investigated, but the American

and Asian species are badly in need of attention.

It is clear, therefore, that a considerable amount of

basic work is still required throughout the distri-

butional range of the genus.

The present attempt to arrive at a classification

of the African species must be seen in the light of

the important work on pollen-morphology by Guinet

(l.c.). On the basis of the number, position, complexity

and size of the apertures, the number of cells in

the polyads, the sculpturing of the exine and the

position of the furrows, Guinet has proposed that:

1. A. albida should be removed from the genus

Acacia and that the genus Faidherbia A. Chev. should

be resuscitated. Guinet and Vassal (l.c.) agree that

Faidherbia is a good genus.

2. A. farnesiana (L.) Willd., A. giraffae Willd.

and A. caven (Mol.) Mol. constitute a good group

within Bentham’s Series Gummiferae and should be

referred to the genus Vachellia.

3. The existing genus Acacia should be subdivided

into three large genera (see Guinet fig. 19). The

Gummiferae would constitute one genus. As the

type species of Acacia, A. nilotica, is a member of

Gummiferae, the species in Gummiferae would re-

main under Acacia. Bentham’s series Vulgares and

Filicinae would constitute a second genus, and the

Phyllodineae, Pulchellae and Botryocephalae the

third. New generic names would be required for

the last two groups.

In Africa, therefore, A. albida would be referred

to Faidherbia, A. giraffae to Vachellia, the remaining

species armed with stipular spines would remain in

Acacia, and the species with non-spinescent stipules

would be referred to yet another genus. All four of

these genera would occur in southern Africa.

As discussed elsewhere (Ross in Bol. Soc. Brot.,

ser 2, 40: 188, 1966), A. albida exhibits a number of

unusual characters, some of which are peculiar to

this species alone amongst the African acacias. It

differs in having leaves with eglandular petioles

but a gland on the rhachis at the point of attachment

of each pinnae pair, stamen filaments which are

shortly connate basally (also in A. eriocarpa Brenan

and in A. ogadensis Chiov.) and large anthers which

are eglandular even when in bud. Eglandular petioles

do occur in certain other African species although

not consistently while eglandular anthers are found

in most of the extra-African species of the genus.

The pollen of A. albida forms polyads of 30 cells

whereas in all of the other African species investi-

gated the polyads have only 16 cells, except for A.

giraffae with 26 to 48 cells (van Zinderen Bakker

and Coetzee, l.c.). Vassal (Bull. Soc. Hist. Nat.

Toulouse 103: 583, 1967) found from a study of

seedling development that the ontogeny of the leaf

in A. albida differed from all other members of the

Gummiferae studied in producing bipinnate leaves

from the outset.

Chevalier (Rev. Bot. Appl. 4: 876, 1934) con-

sidered A. albida to be sufficiently distinct from all

of the other species to transfer it to the monotypic

genus Faidherbia. A. albida is not closely related

to any of the other African species and there are

the above characters to suggest that the species would

be better placed in Faidherbia. However, although

A. albida differs from the other African acacias it

112

CLASSIFICATION OF THE AFRICAN ACACIAS

does nevertheless share many characters in common

with them. In deciding whether or not the species

should be excluded from Acacia , it depends upon

whether one is more influenced by the similarities

or by the differences. Despite the somewhat anomalous

position of A. albida, the species is being left in Acacia

for the account of Mimosoideae which is currently

being prepared for the Flora of Southern Africa,

although clearly it may ultimately be found better

to place the species in Faidherbia.

A. giraffae shares so many common characters

with other species in the Gummiferae that I would

be extremely reluctant to see it split off Acacia and

placed in another genus. It is unfortunate that

Guinet has not yet investigated the pollen-morpho-

logy of A. haematoxylon Willd. as A. giraffae and

A. haematoxylon hybridize. Consequently, it would

be interesting to know whether A. haematoxylon

has pollen similar to that of A. giraffae.

One important result of Guinet’s studies is that

they confirm Bentham’s broad subdivisions of the

genus for, as mentioned above, Guinet’s three pro-

posed genera may be arrived at by amalgamating

certain of Bentham’s series. I am opposed to the

idea of fragmenting Acacia into three large genera

because, despite the differences in pollen-morpho-

logy enumerated by Guinet, the species share so

many other characters in common. What would

really be achieved by giving the three groups generic

status? The three genera thus recognised would

still be more closely related to one another than to

any other genera. Surely a more satisfactory solution

would be to accord each of the three proposed

genera subgeneric rank within Acacia.

Irrespective of whether or not Acacia is fragmented

into three large genera, it would still be necessary

for each of the groups to be subdivided. A system

of subdividing the African Vulgares and Gummi-

ferae (this name would have to fall into disuse in

any new system of classification as the group contains

the type species of Acacia) would still be required.

Nearly all of the classifications to date have been

based on gross morphology alone but it is clear

that other characters, for example, pollen-morphol-

ogy and seedling-morphology yield valuable infor-

mation. It seems most desirable to correlate infor-

mation of this nature with gross morphology in

any new attempt to draw up a classification. Hope-

fully it will not be too long before this is possible.

Meanwhile, some suggestions are advanced for sub-

dividing the Vulgares and Gummiferae.

*T

]. Vulgares

Bentham recognized four subseries within the

Vulgares. In doing so he relied to a certain extent

on geographical distribution, dividing the series

into the capitate- and the spicate-flowered species of

the Old World, and the capitate- and the spicate-

flowered species of the New World. Despite the

convenience, it no longer seems desirable to segregate

the Old World and the New World species.

All of the species with prickles in pairs near the

nodes (Diacanthae) or prickles in threes (Triacanthae)

have spicate inflorescences. It is only among the

species armed with irregularly scattered prickles

that capitate- and spicate-flowered species occur.

Consequently, there is no need to draw upon the

nature of the inflorescence to provide the major

subdivisions within the Vulgares as the subdivisions

can be arrived at by using the arrangement of the

prickles. The envisaged subdivisions within Vulgares

would then be as follows (Bentham’s ranks and

names have been retained here purely for convenience

although clearly they are no longer appropriate):

Series Vulgares

Subseries 1. Triacanthae j

Subseries 2. Diacanthae

Subseries 3. Ataxacanthae

A. Spiciflorae

B. Capitulatae

The Ataxacanthae, when delimited as above,

would contain all of the species with prickles scat-!

tered irregularly along the internodes. These species

appear to form a far more convenient group than

is apparent from Bentham’s subdivision of the,

Vulgares. However, the division of Ataxacanthae

into those species with capitate inflorescences and!

those species with spicate inflorescences is desirable!

because, although the species share many characters

in common, there are also some notable differences.!

Guinet’s division of the Vulgares on the basis of,

pollen-morphology indicates that the pollen in the,

capitate- and in the spicate-flowered species with

irregularly scattered prickles is different. This is,

also supported by Vassal’s studies on seedling mor-

phology. Vassal maintains that the “T. pennata')

group and A. ataxacantha (unfortunately no othei

species with scattered prickles and spicate inflore-

scences were investigated), although sharing some

characters in common, should nevertheless be sepa-

rated because of differences in seedling morphology.1

Of the 14 African species in this group with scat-

tered prickles, only four have spicate inflorescences.

However, all of the species, irrespective of whethei

the inflorescence is capitate or spicate, have the

following characters in common:

1. Pale yellowish- white flowers.

2. Stipitate ovaries.

3. Pubescent ovaries (except for A. lujae De Wild.)

Pubescent ovaries are, of course, found in severa

other species of African acacias but it seems of some

significance that nearly all of the species in thi:

group have this character and one wonders whethei

there is possibly some evolutionary involvement

4. Almost all of the species are climbers or scanden

shrubs. A. ataxacantha and A. brevispica Harm:

sometimes grow as trees while A. eriocarpa apparently

always grows as a small tree.

5. They all lack secondary leaves. As mentionee

earlier, secondary leaves are thought of as an adap

tation for the production of new leaves withou

necessitating the formation of new branchlets. How

ever, climbers must continue to grow in order t<

compete for light otherwise they face the possibility

of being shaded out by the surrounding vegetation

The possession of secondary leaves may, therefore

not confer any significant benefit to a climber.

6. The pods of all species are essentially simila

in being ± umbonate over the seeds and in havin;

a fairly conspicuous transverse venation.

J. H. ROSS

113

Within this group of 14 species, the species

with spicate inflorescences differ from those with

capitate inflorescences in the following respects:

1. The size of the apertures of the pollen grains.

Guinet records that in the spicate-flowered species

the diameter of the apertures is 3-5/x whereas in the

capitate-flowered species the diameter is 1, 2-2,9 /x.

2. Seedling morphology.

3. Areole shape. In the species with spicate in-

florescences the areole is small and typically horse-

shoe shaped (in A. ataxacantha it is a small central

depression) while in the capitate-flowered species the

areoles tend to be larger and conform in shape to

the outline of the seeds.

II Gummiferae

The subdivision of the series Gummiferae presents

far greater difficulty. Bentham subdivided Gum-

miferae into three main groups primarily on the

position of the involucel on the peduncle, namely,

Summibracteatae, Medibracteatae and Basibractea-

tae. The position of the involucel in many species

varies within far wider limits than previously realized

so that this character is no longer suitable for deli-

miting major groups within the Gummiferae. The

character is, however, important in certain small

groups of species for distinguishing between closely

related species. Some new character or combination

of characters must, therefore, be employed to sub-

divide Gummiferae.

The obvious character for dividing Gummiferae

is the nature of the inflorescence but the result is

very disappointing because the vast majority of

species have capitate inflorescences. In Africa only

five species have a spicate inflorescence, namely, A.

albida, A. lahai Steud. &Hochst. ex Benth., A. horrida

(L.) Willd., A. bussei Harms ex Sjostedt and, doubt-

fully, A. dolichocep/iala. A. albida differs from the

other four species in several respects so it would

have to be split off from the rest, if not placed in

Faidherbia.

Flower colour is important taxonomically and

could be employed to distinguish two broad groups

within the capitate-flowered species. The two main

colour groups found are white to pale yellowish-

white (or occasionally pale pink or rarely purple

in A. xanthophloea Benth. in tropical east Africa)

and bright- or golden-yellow. A. xanthophloea has

bright-yellow flowers throughout most of its range

but, apart from the anomalous behaviour of this

species, all of the other species can be satisfactorily

placed in one or the other colour group.

The species within each of these groups could be

subdivided on the basis of the stipular spines, a

character employed by Bentham. The stipular spines

are usually either straight or almost so, or else they

are strongly recurved, while in certain species a

mixture of long straight and short hooked spines

occur together. The latter group of species corres-

ponds to Bentham’s “Heteracanthae”.

The pods are useful for breaking down the groups

further into those species with indehiscent pods

and those species with dehiscent pods. The species

with indehiscent pods could be divided according

to whether the valves are thin or whether they are

markedly thickened, woody or pulpy. Pod shape,

that is falcate as opposed to straight or almost so,

often varies within a single species and it is felt

that this is not a suitable character to employ for

further subdividing major groups of species with

dehiscent pods.

These above gross morphological characters, when

considered together with pollen and seedling mor-

phology, should enable the Gummiferae to be divided

satisfactorily.

Despite the need for a classification of the African

species, there is, however, an even greater need

to resolve numerous basic taxonomic problems,

particularly those in north-east tropical Africa. The

last fairly comprehensive account of the species was

by E. G. Baker in his Leguminosae of Tropical Africa

(1930), but unfortunately this treatment of the genus

is now quite out of date. The regional floras have

contributed greatly to our knowledge of the African

species, but many problems await elucidation.

■

Bothalia 11, 1 & 2: 115-123 (1973)

Notes on African Plants

VARIOUS AUTHORS

CELASTRACEAE

AUTHOR CITATION FOR PUTTERLICKIA PYRACANTHA

The combination Putterlickia pyracantha is generally

attributed to (L.) Endl., e.g. by Davison in Bothalia

2: 337 (1927) and Loesener in Pflanzenfam. 20b:

154 (1942). However, although Endlicher, Gen.

1086 (1840), described the genus Putterlickia , basing

it on Celastrus pyracanthus L., he did not effect the

combination. The earliest effective combination that

has been traced was made by Szyszylowicz in 1888

as follows:

Putterlickia pyracantha (L.) Szyszyl., Polypet.

Disc. Rehm. 36 (1888).

L. E. Codd

ERICACEAE

Erica krugeri E. G. H. Oliver, sp. nov., E. obtu-

satae Klotzsch ex Benth. et E. oliveri H. A. Baker

affinis sed bracteolis grandibus, sepalis pilis glan-

duliferis marginibus et supra interstitia corollarum

incurvatis, corolla valde quadrangulosa distinguitur.

Frutex diffusus repens, interdum nodis radicans.

Rami longi et graciles ad 1 m longi glabri longi-

tudine porcati. Folia ternata ad 20 mm longa et 2

mm lata, late disposita patentes vel interdum recur-

vata lineari-lanceolata longe acuminata minute sca-

brida ad apice aliter glabra. Flores ternati terminales

in extremis ramulorum lateralium pendentes parum

involucrati, pedicellis brevibus glabris, bracteolis 3

grandibus ovatis acuminatis ciliatomarginatis aliter

glabris vel sparsim pubescentibus ad apicem in

superficiebus ambabus albis rubrescentibus. Sepala

corollam aequantia supra interstitia corollarum

incurvata glabra margine pilis glanduliferis alba

interdum rubrescentia. Corolla ad 2,5 mm longa et

4mm lata initia oblato-globosa sed mox valde quadran-

gulosa et late cyathiformis, lobis crenatis tubum

aequantibus. Stamina 3 mm longa exserta, filamentis

planis glabris, antheris ad 1 mm longis terminali-

bus muticis poro \ cellula aequante. Ovarium ad

0,75 mm longum, 1 mm latum oblato-globosum

glabrum, stylo stamina longe superante, stigma sim-

plici.

Type: Cape. — 3418 (Simonstown): Somerset Sneeu-

kop, south-east slopes, 1 280 m, (-BB), Nov. 1969,

Oliver 3024 (STE, holo., PRE, NBG, BOL, K, S,

W, MO).

Diffuse shrublet creeping along the ground and

among surrounding vegetation, sometimes rooting

at the nodes. Branches long and slender, up to 1 m

long, glabrous, brown, longitudinally ridged. Leaves

3-nate, up to 20 mm long and 2 mm broad, widely

spaced, spreading or occasionally recurved, often

spreading in the plane of the branches, linear-

lanceolate, long acuminate with a brown seta up to

2 mm long, minutely scabrid towards the apex

otherwise glabrous, dark green; petiole up to 2 mm

long adpressed glabrous. Flowers 3-nate, terminal

on ends of lateral branches, slightly involucrate,

pendent; pedicels short, glabrous; bracteoles large,

up to 4 mm long and 2 mm wide, acuminate, edges

ciliate with short stout setose hairs intermingled

with fine hairs, otherwise glabrous, or sparsely

pubescent towards the apex on both surfaces, white,

becoming dark red, the edges remaining white.

Sepals as long as the corolla, curving inwards over

the corolla intersitices, elliptic-oblong, glabrous,

edged with conspicuous stout gland-tipped hairs,

white, sometimes turning reddish. Corolla 2,5 mm

long and 4 mm broad, at first oblate-globose, but

soon becoming distinctly 4-angled and broad cyathi-

form; lobes as long as the tube, crenate. Stamens

3 mm long, exserted; filaments very slightly nar-

rowed at the apex, glabrous; anthers 1 mm long,

terminal, muticous, dark brown, relatively smooth;

pore i the length of the cell. Ovary 0,75 mm long,

1 mm broad, oblate-globose, glabrous, pale green;

style 2,5 mm long, glabrous, exserted beyond the

stamens; stigma simple.

Cape.— 3418 (Simonstown): Somerset Sneeukop, south-east

slopes, 1 280 m (-BB), Nov. 1969, Oliver 3024 (STE, PRE,

BOL, NBG, K, S, W, MO); Dec. 1971, Vogelpoel sub Williams

1608 (STE, NBG, PRE, K); Landrostkop and Somerset Sneeu-

kop, Nov. 1934, Stokoe 4027 (BOL, NBG); Stokoe in SAM

51333 (SAM): Nov. 1938, Stokoe 8553 (BOL, NBG); Nov.

1939, Stokoe in SAM 55264 (SAM, NBG); Dec. 1939, Ester-

huysen 3618 (BOL, NBG); Moordenaars Kop (-BB), Nov.

1946, Stokoe in SAM 62447 (SAM). 3419 (Caledon): Lebanon

Forest Reserve, Groenlandberge, south slopes, 1 066 m (-AA),

Oct. 1967, Kruger 522 (BOL); Oct. 1969, Kruger in STE 30321

(STE); Oct. 1969, Oliver 3011 (STE, PRE, NBG, BOL, K,

S).

The section Desmia to which this species belongs

is characterised by shrublets with terminal flowers,

the corolla being urceolate and viscid and the stamens

subexserted, terminal and muticous. It contains

four species, E. obtusata Klotzsch ex Benth., E.

polifolia Salisb. ex Benth., E. conferta Andr. and

E. oliveri H. A. Baker, all occurring on mountains

from the Caledon district in the west to Uniondale

in the east.

Under the “incertae” of this section in the Bolus

Herbarium there has remained for many years some

material collected by Stokoe and by Esterhuysen from

the Hottentots-Holland Mountains. During an eco-

logical survey of an area on the Groenlandberg in

116

NOTES ON AFRICAN PLANTS

Erica krugeri E. G. H. Oli-

ver.— 1, flower, side view,

x 12; 2, flower, viewed

from above, x 12; 3,

anther, side and front view,

x25; 4, gynoecium, x 14

(from Oliver 3024).

the Lebanon Forest Reserve, Mr F. J. Kruger

recorded similar material which required identifi-

cation. This material shows affinities to E. obtusata

which occurs in the area and to E. oliveri which

occurs in the mountains near Bredasdorp. However,

it possesses certain characteristics which distinguish

it very clearly from these and all other species.

The first notable feature is the semi-involucrate

nature of the 3-nate terminal inflorescences. This is

produced by the asymmetrical development of the

bracteoles. As the flowers are pendent the “involucre”

hides the flowers and becomes distinctly dark red

in colour.

The characteristics of the flower are remarkable

in the genus. The sepals are distinctly arched inwards

over the interstices between the corolla lobes and

possess conspicuous gland-bearing hairs on their

margins. The corolla is unusually shaped being at

first oblate-globose changing to cyathiform and

4-angled and becoming almost 4-lobed. Basically,

however, the species has its affinities in the section

Desmia.

There is one slight difference between the two

populations of this species: in the Hottentots-Holland

plants the surface of the bracteoles is glabrous

whereas in the Lebanon plants both surfaces are

sparsely pubescent towards the apex.

In habit, E. krugeri is remarkable in being almost

invisible to the passer-by. It grows twining amongst

the thick short fynbos so characteristic of moist

south-facing slopes. Some of the branches measured

were growing as much as 1 m from the main stem

and were extremly difficult to extricate from the

surrounding vegetation. The pendulous flowers and

dark colour of the “involucres” added to the incon-

spicuousness of the species.

E. G. H. Oliver

HALORRHAGIDACEAE

A NOTE ON LA UREMBERGIA REPENS

Although plants of Laurembergia from southern

Africa are vegetatively similar in appearance and

agree in their choice of habitat, a distinction has

been made between the southern Cape species, L.

repens Berg, and plants occurring in the northern

parts, i.e. South-West Africa, Transvaal, etc., because

of differences in the flowers. Bergius in his Descrip-

tiones Plantarum ex Capite Bonae Spei p. 350, t.5,

f. 10 (August 1767) published the genus Laurembergia,

naming it after the Swedish botanist Peter Laurem-

berg and giving it the specific name repens because

of its prostrate habit. His book appeared two months

before the first Mantissa of Linnaeus (October

1767) in which the same genus received the name

Serpicula repens. Typical for this species from the

winter rainfall region are the male flowers which

are borne on long pedicels 10-15 mm long, and

clustered in the upper axils of the branches. The

VARIOUS AUTHORS

117

female flowers are sessile and form glomerules in

the lower axils. The plants may be monoecious as

was the case with the type, but they are predomi-

nantly dioecious in the Cape and were described

as such by Linnaeus.

Material collected further north in the summer

rainfall region lacks these long stalked male flowers

but instead bears 1-2 bisexual flowers on short

(c. 2 mm) pedicels in between clusters of females.

These bisexual flowers possess 4 petals similar to

those of the males. (In this genus the female flowers

are without petals or these may be rudimentary).

This taxon was placed under L. tetranda (Schott)

Kanitz, a South American species which, according

to Schindler, must have come from north Africa

originally.

In the eastern Cape and southern Natal the distinc-

tions between these two “species” break down.

Here we get plants in which the bisexual flower

turns into a short stalked male flower through sup-

pression of the ovary, (cf Strey 9322, a plant collected

at Ngome in the Nongoma district of Natal). In

other plants the pedicels of the male flowers have

become longer, c. 7 mm (e.g. Strey 6392 from Weza

in the Alfred district, Natal,) and so it links up with

the Cape collections. In several of the eastern Cape

plants a reduction in the number of female flowers

in the upper axils is also noticeable, leading to a

complete separation on the stems and ultimately

dioecism.

One could therefore interpret the evolution of

this Laurembergia taxon as follows in its migration

from north to south; 1, the monoecious plants bear

short, stalked, bisexual flowers in the midst of the

female, sessile, axillary glomerules; 2, the bisexual

flowers become male through the suppression of

the ovary; 3, the male flowers develop longer pedicels;

4, in the upper axils the female flowers tend to dis-

appear, likewise the males from the lower parts of

the stems; 5, separation into male and female plants.

L. repens from the southern Cape thus represents

a final stage in the evolution of its flowers.

Recently Raynal (1965) subdivided the African

species in 2 subgenera. L. repens Berg, and L. veronici-

folia (Bory) Schindl. from Madagascar were placed

in the subgenus Laurembergia while the other sub-

genus Serpiculastrum of Raynal contained only one

variable tropical African species, L. tetrandra (Schott)

Kanitz, which he divided into 2 subspecies and 3

varieties. He reached this conclusion after a survey

of all the tropical African material and thus reduced

a large number species, described by Schindler

(Pflanzenreich IV, 1905) and others, to synonyms

or to subspecific rank. Hiern in Cat. Welw. Afr.

Plants 1 : 332 ( 1 896), using the synonym Serpicula

repens L., placed the Angolan collection of Wel-

witsch under this Cape species as a variety, var.

brachypoda. Oliver in the Flora of Tropical Africa

2: 405 (1871) concurred. I too agree with this inter-

pretation but since there is a geographical separation

I follow Raynal in assigning to it the rank of sub-

species.

L. repens Berg, subsp. brachypoda {Hiern)

Oberm., comb, et stat. nov.

Serpicula repens L. var. brachypoda Welw. ex Hiern, Cat.

Welw. Afr. Plants 1: 332 (1896). Lectotype: Welwitsch 1612a

(BM, P).

L. tetrandra (Schott) Kanitz subsp. brachypoda (Welw. ex

Hiern) Raynal in Webbia 19: 694 (1965).

REFERENCES

Adamson, R. S., (1950). FI. Cape Peninsula. 608.

Bergius, P. J., (1767, August). Descr. Plant. Cap. 350, t.5

f. 10.

Harvey, W., (1862). FI. Cap. 2: 572.

Linnaeus, C., (1767, October). Mantissa 1:16.

Oliver, D., (1871). Flora of Tropical Africa 2: 405.

Raynal, A., (1965). Les Especes Africaines du Genre Laurem-

bergia. Webbia 19: 683.

Schindler, A. K., (1905). Halorrhagaceae in Pflanzenr. IV,

225: 61-74.

Van der Meijden, R. & Caspers, N., (1971). Flora Male-

siana ser. 1,71: 246.

A. A. Obermeyer

LABIATAE

THE GENUS RABDOSIA IN SOUTH AFRICA

In Mitt. Bot. Staatssamml. Miinchen 10: 250

(1971) reasons were advanced for keeping the genus

Isodon (Benth.) Kudo (1929) separate from Plec-

tranthus L’Herit. Support for this view comes from

various sources, the most recent being Blake in

Contr. Queensland Herb. 9: 4 (1971) who, however,

points out that there is an earlier name for the genus,

namely, Rabdosia (Bl.) Hassk. in Flora 25 Beibl.

2: 25 (1842), based on Elsholtzia Willd. sect. Rab-

dosia Bl., Bijdr. FI. Nederl. Indie 825 (1825), the

type of which is E. javanica Bl. = Rabdosia javanica

(Bl.) Hassk. = Plectranthus javanicus (Bl.) Benth.

Other generic names that have been put forward

are Amethyst anthus Nakai (1934), based mainly

on the Asiatic species included by Bentham in Plec-

tranthus sect. Amethystoides Benth., and Homa-

locheilos J. K. Morton (1962), based on a tropical

African member of the group, P. ramosissimus

Hook.f.

The genus Rabdosia is, therefore, accepted as

including the following sections described by Bent-

ham, Lab. Gen. Sp. 40-45 (1832): Isodon, Amethy-

stoides, Melissoides and Pyramidium. Section Pyra-

midium consits of only two closely related species:

P. ternifolius Don, which occurs in India, and P.

calycinus Benth., found in the eastern Cape Province,

Natal, Swaziland and eastern Transvaal. The Indian

species has recently been transferred as Rabdosia

ternifolia (D. Don) Hara in J. Jap. Bot. 47: 201

(1972).

These two species are remarkably similar in ap-

pearance, but the leaves of P. ternifolius are narrower

in relation to their length than those of P. calycinus,

while the calyx teeth of the former are shorter than

those of the latter species. Both differ from typical

Rabdosia in being stout, erect plants with usually

ternate leaves and dense, pyramidal inflorescences

of condensed scorpioid cymes. As in Rabdosia ,

the bracts are leaf-like, becoming progressively

smaller towards the apex of the inflorescence and

the calyx is equally 5-toothed. The South African

species is now transferred to Rabdosia.

Rabdosia calycina {Benth.) Codd, comb. nov.

Plectranthus calycinus Benth. in E.Mey., Comm.

PI. 230 (1837). Lectotype: Transkei, between Morley

and Umtata River, Drege (K).

L. E. Codd

118

NOTES ON AFRICAN PLANTS

LAURACEAE

A NEW GENERIC RECORD FOR SOUTH AFRICA AND A NEW SPECIES

Beilschmiedia natalensis J. H. Ross, sp. nov.,

affinis incerta.

Arbor sempervirens, 9-20 m alta; ramuli novelli

glabri. Folia opposita, 0,5-1 cm longe petiolata,

petiolo supra canaliculato, glabro vel glabrescente;

lamina elliptica, basi cuneata, apice acuta usque

breviter acuminata vel subobtusa, 3,5-10 cm longa

et 2-5 cm lata, coriacea, atro-viridis, glabra; costa

media subtus distincte prominens, glabra; nervi

laterales utrinsecus prominentes; reticulatio utrinque

prominula. Inflorescentiae ad apicem ramulorum

novellorum axillares; pedunculo subtiliter puberulae;

bracteae mox caducae. Flores parvi, albi, dr 1,5

mm longi et apice 2, 5-3 mm diam., glabri vel sub-

tiliter puberuli, pedicellis gracilibus 2-4 mm longis,

subtiliter puberulis. Tepala 6, ovato-triangulata,

subaequalia, apice acuta, erecto-incurvata, dr 1,25

mm longa et dr 1 rnm lata, glabra vel subtiliter

puberula. Stamina fertilia 9, quorum 6 exteriora

sessilia, dr 1 mm longa et 1 mm lata, basi glabra,

tepalis adnata, antheris introrsis; 3 interiora dr 1

mm longa et dr 0,9 mm lata, filamentis subtiliter

pubescentibus, antheris lateralibus usque subex-

trorsis, basi utrinsecus glandulis binis magnis, sub-

globulosis, dr 7 mm latis praedita; staminodia

subsessilia, dr 1 mm longa, ovato-triangulata,

puberula. Ovarium ovoideum, dr 1 mm longum et

0,8 mm diam., glabrum, in receptaculo immersum

et in stylum brevem 0,6 mm longum erectum at-

tenuatum. Drupa subglobosa, 1,5-2 cm diam.,

glabra.

Type: Natal, 3030 (Port Shepstone), Umdoni Park

(-BC), 21 Dec. 1971, H. B. Nicholson 1140 (PRE,

holo. ; BOL, E, FHO, K, L, NH, NU, iso.).

Evergreen tree 9-20 m high; bark brown, scaly;

young branchlets glabrous. Leaves opposite, petiole

0,5-1 cm long, canaliculate above, glabrous or

glabrescent; lamina elliptic, cuneate basally, apex

acute to shortly acuminate or subobtuse, 3,5-10 cm

long and 2-5 cm wide, coriaceous, dark-green, gla-

brous; midrib distinctly prominent below, glabrous;

lateral nerves in 5-8 pairs, prominent on both sur-

faces, erect patent, the basal ones often steeper,

reticulation dense, prominent on both surfaces.

Inflorescences axillary at the apices of the young

branchlets; peduncle minutely puberulous; floral

bracts small, soon deciduous. Flowers small, white,

dt 1 ,5 mm long and 2,5-3 mm in diameter apically,

glabrous or minutely puberulous, pedicel slender,

2-4 mm long, minutely puberulous. Tepals 6, ovate-

triangular, subequal, acute apically, erect-incurved,

± 1,25 mm long and ± • mm wide, glabrous or

minutely puberulous externally, pubescent internally

towards the base. Fertile stamens 9, outer 6 sessile,

± 1 mm long and 1 mm wide, glabrous, adnate to

the tepals basally, anthers introrse, obtuse; 3 interior

stamens db 1 mm long and ±0,9 mm wide, filaments

short, minutely pubescent, anthers lateral to sub-

extrorse, a large subglobular gland ± 7 mm wide

on either side basally; staminodia subsessile, ± 1

mm long, ovate-triangular, puberulous. Ovary ovoid,

± 1 mm long and 0,8 mm in diameter, glabrous,

immersed in the receptacle, gradually merging into

a short style 0,6 mm long; stigma ± conspicuous.

Drupe subglobose, 1 , 5-2 cm in diameter, glabrous ;

pedicel 0,8-1, 4 cm long, cylindrical, glabrous.

Natal. — 2831 (Nkandla): Ngoye forest (-DC), Welts &

Edwards 120 (K). 2930 (Pietermaritzburg): Nqutu Falls, just

west of Krantzkloof Nature Reserve (-DB), K. H. Cooper 49

(NH); Moll, Cooper & Zaloumis 5668 (NH). 3030 (Port Shep-

stone): Umdoni Park (-BC), Nicholson 1140; 1140a (NH);

Strey 10578 (NH); K. H. Cooper (NH). 3130 (Port Edward):

Foster’s Kloof, Craigadour Farm (-AA), Nicholson 1227 (NH);

Swanepoel’s Kloof, Beacon Hill East (-AA), Nicholson 1228

(NH).

It is now over ten years since the first specimen

of this species was collected in the Ngoye forest in

Zululand. However, as the specimen was sterile,

its identity remained unknown. Some five years ago

a fruiting specimen was collected near the Krantz-

kloof Nature Reserve but, once again, in the absence

of flowers, the plant remained unidentified. Following

the discovery of a large tree at Umdoni Park on

the Natal south coast, the plant was kept under

surveillance for any sign of flowers. The discovery

and collection of good flowering material in December

1971, which finally enabled the plant to be identified

as a Beilschmiedia, is due to the persistence and

patience of Mr H. B. Nicholson. The collection

of this flowering material by Mr Nicholson was

soon followed by the discovery of several more

plants much further south in the Port Edward

area.

Beilschmiedia is a pantropical genus of ± 300

species occurring mainly in Asia, with relatively few

species in Africa. The occurrence of a species of

Beilschmiedia in Natal is very interesting as the

nearest Beilschmiedia species occur in Angola and

in Zambia. There is, therefore, a large geographical

discontinuity between B. natalensis and the nearest

species of Beilschmiedia. B. natalensis does not ap-

pear to be very closely related to any of the African

species. It appears to be confined to kloof forests

along the Natal coast.

B. natalensis has 2-valved anthers as in the case

of Cryptocarya but, unlike all of our species of

Cryptocarya, B. natalensis has opposite leaves.

I am most grateful to Prof. Dr A. J. G. Kostermans,

Bogor, for advice on the generic position of this

plant and for reading this manuscript and to Mr

H. K. Airy Shaw for checking the Latin description.

J. H. Ross

VARIOUS AUTHORS

119

LILIACEAE

ALOE, CHAMAEALOE, HAWORTHIA, ASTROLOBA, POELLN1TZIA AND CHORTOLIRION

While engaged on drawing up a key for the Lilia-

ceae for inclusion in the third edition of the Genera

of South African Flowering Plants, the genera Aloe

and Haworthia were reviewed together with the small

related genera Chamaealoe, Astroloba , Poel/nitzia

and Chortolirion (tribe Aloineae).

In Bull. Afr. Succ. Plant Society, London 6: 145,

195, etc. (1971), a semi-popular magazine, un-

fortunately not readily available in South Africa,

C. A. Parr has sunk the genera Astroloba and Poell-

nitzia under Haworthia in a new section, the Quin-

quefariae. These two genera had been separated

from Haworthia because of their regular perianths,

those of Haworthia being two-lipped. A few Haworthia

species, however, e.g. H. marginata and H. margariti-

fera, also possess regular flowers and there being

no other distinctive characters correlated with this

regular perianth, it forms a weak basis for separation.

Regular or bilabiate perianths frequently occur in

the Iridaceae where species belonging to one genus,

or even varieties of one species, show both types.*

In the genus Aloe with its usually regular flowers

two species, A. albida (Stapf) Reyn, and A. myria-

cantha (Haw.) Roem. & Schult.f. also possess bila-

biate perianths.

It is interesting to observe the parallel develop-

ment of variations inherent in these two large genera,

Aloe and Haworthia. The tribe Aloineae to which

they belong is characterised by possessing a stem

or a rhizome but both have species producing a

bulbous base (the leaf-bases having become swollen),

e.g. Aloe kniphofioides Bak., A. buettneri Berg.,

A. modesta Reyn., etc., and Haworthia angolensis

( Chortolirion angolense). In both genera the roots

may be thin or swollen and the leaves variously

arranged in spiral or longitudinal rows along the

stems or in basal rosettes. The raceme may be simple

or branched, the perianth regular or two-lipped.

But in Aloe the flowers are usually regular and

showy and possess long, often exserted stamens

and style, whereas those of Haworthia are small

and inconspicuous, usually bilabiate with the stamens

and style short and included.

Haworthia angolensis Bak. in Trans. Linn. Soc.

ser. 2, Bot. 1: 263 (1878), J. Linn. Soc. 18: 210

(1880), FI. Trop. Afr. 7: 469 (1897). Type: Angola,

Huilla, Welwitsch 3756 (BM, holo., PRE, photo).

H. tenuifolium Engl, in Bot. Jahrb. 10: 2, t. 1 (1889). Type:

Cape, near Kuruman, Marloth 1049 (B, holo., PRE!). H.

stenophylla Hook., Icon. pi. t. 1974 (1891). Type: Transvaal,

Barberton, Galpin 858 (K, holo., PRE!). H. subspicata Bak.

in Bull. Herb. Boiss. 2, ser. 4: 998 (1904). Type: Transvaal,

Modderfontein near Johannesburg, Conrath 645 (Z, holo.).

*In the Iridaceae experiments showed that if gravity was

eliminated, the zygomorphic perianth would revert to an

actinomorphic pattern.

Chortolirion angolense (Bak.) Berger in Pflanzenr. 4, 38,

3, 2: 723 (1908). C. tenuifolium (Engl.) Berger, l.c. p. 73. C.

stenophyUam (Bak.) Berger, l.c. p. 72. C. subspicatum (Bak.)

Berger, l.c. p. 74.

Chortolirion, separated from Haworthia by Berger

because of its bulbous base, is closely related to

H. graminifolia G. G. Smith, which also possesses

thin, narrow denticulate leaves with a slight ten-

dency for these to swell at the base. Berger also

drew attention to the shape of the ovary which he

considered an important difference, for he found

it to be acuminate and not obtuse apically. He must

have examined young fruits for initially the ovary

is obtuse and well separated from the terete, short

style; apparently the upper ovules are sterile causing

the capsule to become attenuated. This may also

be seen in other species of Haworthia.

Although Berger recognized 4 species in Chorto-

lirion, a study of the ample material now available

shows the differences enumerated by him as mere

variations in this widespread summer rainfall species.

It is found in grassland where, owing to fires or

frost, its leaves may die back to ground-level.

Aloe bowiea Roem. & Schult.f. Syst. Veg. 7:

704 (1829); Kunth, Enum. 4: 515 (1843); Bak. in

J. Linn. Soc. 18: 158 (1880), FI. Cap. 6: 309 (1896).

Type: Cape, Bowie (K, holo.).

Bowiea africana Haw., Phil. Mag. 299 (1824), l.c. 123 (1827).

Chamaealoe africana (Haw.) Berger in Pflanzenr. 4, 38, 3,:

120(1908).

This small rare plant, first collected by Bowie

in 1822, most probably in the neighbourhood of

Uitenhage, where it is endemic, was described by

Haworth as Bowiea africana .* Another species,

Bowiea myriacantha , was placed next to it and the

genus described as possessing a bilabiate perianth

which is not correct for Bowiea africana, for here

it is regular. Roemer & Schultes f. removed them

both to Aloe (1829) but since the name Aloe africana

was not available, Bowiea africana Haw. became

Aloe bowiea R. & S. In 1908 Berger placed Aloe

bowiea in a separate genus, Chamaealoe africana.

Reynolds upheld Chamaealoe for it is not included

in his Aloes of South Africa. The genus was omitted

in Phillips’ Genera of South African Flowering

Plants. The flowers of Aloe bowiea, although smaller

and thinner, resemble those of Aloe dolomitica,

A. sessiliflora , etc., which also possess erect, short

perianth-segments and have the stamens and style

much exserted. In the genus Aloe the extent of the

fusion and/or connivence of the segments varies

and it cannot be used as a distinguishing character.

A. A. Obermeyer

*Bowiea Harv. ex Hook.f. (containing the single species

B. volubilis Harv. ex Hook.f.) is conserved against Bowiea

Haw.

120

NOTES ON AFRICAN PLANTS

A NEW SPECIES OF ALOE FROM SOUTH WEST AFRICA

Aloe dewinteri Giess, sp. nov., A. namibensi

Giess affinis, sed foliis majoribus patentibus utrinque

laevibus, florum bracteis plus minusve obtusis vel

leviter apiculatis, floribus deinde flavis fere albis

differt.

Plantae solitariae vel caulo brevi ad 10 cm longo.

Folia 14-22, rosulata, inferiora patentia, in quadrente

superiore recurvata, juniora erecta vel patentia,

lanceolato-attenuata, basin versus plana, apicem

versus aliquid canaliculata, ad 50 cm longa basi ad

15 cm lata, marginibus flavo-brunneis cartilagineis

dentibus deltoideis brunneis 1-2 mm longis 10-20

mm distantibus munitis, utrinque laevia, cano-viridia.

Inflorescentia 1-2 eadem rosula, simplex vel 2-3-

ramosa, aliquid arcuato-adscendens, ad 85 cm alta.

Pedunculus basin versus compressus, ad 3 cm latus;

bracteae ad 24 chartaceae amplectentes steriles

25-30 cm longae. Racemus dense multiflorus, anguste

cylindrico-acuminatus, 25-40 cm longus, ad 5 cm

diam., alabastris erectis pallide roseis; flores pen-

duli, primum corallini deinde flavi fere albi; brac-

teae oblongo-obovatae, obtusae vel aliquid api-

culatae. Pedicelli ad 4 mm longi, 1 mm diam. Perian-

thium cylindricum, trigonum, sensim in pedicellum

angustatum, 30-33 mm longum, ad 8 mm diam.;

segmenta exteriora 2/3— 3/4 longitudinis integra,

3-nervata, apicibus subobtusis, leviter patentia; seg-

menta interiora integra sed dorse ad exteriora adnata,

exteriora aequantia, alba, translucida, aliquid cari-

nata apicibus obtusis leviter patentibus. Filamenta

alba, complanata, interioribus angustiora 2-3 mm

longiora, deinde 6 mm exserta. Ovarium 5-6 mm

longum, ad 3 mm diam., viride, stylo ad 30 mm

longo. Fig. 2, 3 and 4.

Type: South West Africa, 1913 (Sesfontein):

Anibib, between Sesfontein and Warmquelle (-BA),

Bulir sub Giess 10990 (WIND, holo.; PRE, M).

Plants solitary, stemless or with a short stem up

to 10 cm long. Leaves 14-22, in a rosette, lower

ones spreading, recurved in upper quarter, younger

ones erect to spreading, lanceolate-attenuate, flat

towards the base, becoming somewhat canaliculate

towards the tip, up to 50 cm long and up to 15 cm

wide at the base, margins with yellowish-brown

horny ridge, armed with brown deltoid teeth 1-2

mm long and 10-20 mm apart, both surface smooth,

grey-green with slight powdery bloom. Inflorescence

1-2 per rosette, simple to 2-3-branched, somewhat

Fig. 2.- Aloe dewinteri, habit. Plant growing in an horizontal

position in a garden. In nature it hangs on perpendicular

clifT-faces ( Giess 10990).

VARIOUS AUTHORS

121

arcuate-ascending, up to 85 cm high. Peduncle

compressed towards the base, up to 3 cm broad

with up to 24 papery, clasping sterile bracts 15-30 mm

long. Raceme rather densely multi-flowered, nar-

rowly cylindrical-acuminate, 25-40 cm long and

up to 5 cm in diameter, buds erect, pale pink, flowers

pendulous, coral pink, turning yellowish to white

when open, bracts oblong-obovate, obtuse to some-

what apiculate. Pedicels up to 4 mm long and 1

mm in diameter. Perianth cylindric trigonous, gradual-

ly tapering into the pedicel, 30-33 mm long and up

to 8 mm in diameter, outer segments free for 2/3 or

3/4 of their length, 3-nerved, apices subobtuse, very

slightly spreading, inner segments free but dorsally

adnate to the outer, as long as the outer, white

translucent, somewhat carinate with obtuse, very

slightly spreading tips. Filaments white, flattened,

the outer narrower and about 2-3 mm longer than

the inner, eventually exserted for about 6 mm.

Ovary 5-6 mm long, up to 3 mm in diameter, green,

style up to 30 mm long.

South West Africa. — 1913 (Sesfontein): Anibib, between

Sesfontein and Warmquelle (-BA), Buhr sub Giess 10990

(WIND, holo.; PRE, M); 26 km W. of Sesfontein (-AA),

Haagner in PRE 32044; between Sesfontein and Warmbad

(-BA), De Winter & Leistner 5856.

Fig. 3. — Aloe dewinteri, inflorescence ( Giess 10990).

122

NOTES ON AFRICAN PLANTS

Fig. 4. — Aloe dewinteri, flowers and floral bracts showing the

development (left to right) from the bud to the withered stage,

xl ( Giess 10990).

The species has so far only been collected in the

vicinity of Warmbad and Sesfontein in the Kaoko-

veld, where it inhabits sheer cliff-faces.

Of the South African species, A. dewinteri shows

most affinity to those species of the series Asperi-

folia Berger. It differs from all of them, however,

by the leaves which are smooth on both surfaces

and are spreading, by the more or less obtuse floral

bracts and the cliff-hanging habit.

Bornman and Hardy in their book “Aloes of the

South African Veld”, Voortrekkerpers, Johannesburg,

1971, compare the species with A. niebuhriana Lav-

ranos, a species from Arabia.

The species was first collected in 1950 on the

farm Warmquelle in the Kaokoveld by Dr E. R.

Scherz and cultivated in a garden in Klein-Wind-

hoek. More plants from the same locality were

later received from Siegfried Triebner and again from

Scherz, but no flowering was recorded until a plant

from the same source, grown in the garden of Mrs

Ute Meyer of Klein-Windhoek, produced an in-

florescence in 1968. Material had also been collected

CAESIA AND

The genus Nanolirion was established by Bentham

in Benth. & Hook.f., Genera Plantarum 3, 2: 793

(1883) for a small Liliaceous species discovered by

Harry Bolus in the south-western Cape, in the Great

Winterhoek mountains near Worcester and described

by him as Herpolirion capense. Bentham removed

it from this genus and placed it in a (monospecific)

genus of its own, Nanolirion. With more complete

and better material now available, including cap-

sules and seeds, it was seen to be so similar to Caesia

by B. de Winter and O. A. Leistner from the same

area in 1957. The single surviving plant only pro-

duced its first inflorescence in the Pretoria National

Botanic Garden in December 1970. The honour of

painting this new species for the first time goes to I

Barbara Jeppe who, on the 30th November 1967,

figured a plant, collected by C. Haagner, which

was flowering in her garden in Johannesburg, and

which was published, with a short description, in 1

the third edition of her book on Aloes, 1970, pp.

132 A and 132B, as Aloe sp. nov. The present type

plant I received from Dr W. Buhr in 1966, and it

flowered in my garden in Windhoek in December

1969 and 1970.

The species is named after Dr B. de Winter, who

was one of the first persons to study, photograph |

and collect it in its natural habitat. The plate in

Bornman and Hardy, l.c., depicts a plant collected

by him.

W. Giess

NANOLIRION

that it was decided to transfer it to this genus. This

was already suggested by K. Krause in Pflanzenfam.

ed. 2, 15a: 289 (1930), who believed it to represent a

depauperate relative.

Caesia capensis ( H.Bol .) Oberm., comb. nov.

Herpolirion capense FI. Bolus in J. Linn. Soc. 18: 395 (1881).

Nanolirion capense (H. Bol.) Benth. & FFook., Gen. Plant.

3, 2: 793 (1883); Oliv. in Hook., Ic. t.1726 (1887); Bak. in

FI. Cap. 6: 402 (1897). Type: Cape, Worcester district, Great

Winterhoek mountains, H. Bolus 5170 (BOL, holo.).

VARIOUS AUTHORS

123

Small, caespitose rhizomatous plants up to 7

cm tall, with long spreading fibrous roots. Leaves

in compact, subdistichous tufts, subulate or sub-

terete, apiculate, sheathing and membranous at

the base, glabrous. Flowers solitary or few together,

pseudo-umbellate, situated amongst the leaves; pedi-

cel terete, stout, erect at anthesis recurved like a

shepherd’s crook when in fruit. Perianth-segments

forming a short tube below, 3-nerved, pale blue,

spirally twisted when faded, deciduous, apex conical

with an introrse minute, deflexed, papillate tail.

Stamens 6, the 3 outer shorter, filaments attached

to the perianth tube, smooth, somewhat dilated,

inserted in a basal anther pit, anthers opening by

longitudinal slits, dorsifix. Ovary 3-celled with 2

collateral ovules, the one projected upwards, the

other downwards; style filiform, stigma apical,

minute. Capsule loculicidal, tripartite, the loculi

globose, seeds bean- or kidney-shaped, black, shiny,

verrucose and with irregular rows of larger tubercles,

strophiole discoid, white, swollen.

The species is confined to the south-western Cape

growing on mountain ledges at high altitudes (c.

2 000 m.).

Cape. — 3319 Worcester: Great Winterhoek mountains

(-AA), H. Bolus 5170 (BOL, holo.); Marloth 1643; Ester-

huysen 19780; Little Winterhoek mountain (-AA), Marloth

2433; Milner Peak, Hex River mountains (-CB), Esterhuysen

9387; Matroosberg, (-BD), Marloth 2201. 3321 Ladismith:

Sevenweeks Poort (-AD), Primos 54.

The genus Caesia R.Br. is an Australian genus

with about 9 species recorded from that continent.

There is one endemic species in Madagascar. In

southern Africa the genus is confined to the Cape

and was believed to contain several species, but

the ample collections at present available show that

we are dealing with a variable species, C. contorta ,

some plants being better developed, others reduced

in size, the latter usually gathered on mountain

peaks. The previous species, C. capensis, can be

distinguished from C. contorta by its small size and

short stalked flowers placed amongst the leaves.

Caesia contorta ( L.F .) Dur. & Schinz, Consp.

FI. Afr. 5: 353 (1893); Leighton in FI. Cape Penins.

186 (1950). Type: Cape, Thunberg (LINN 432.21

holo., UPS).

Anthericum contortum L.f., Suppl. 202 (1781). A. brevi-

folium Thunb., Prodr. 62 (1794), Bak. in J. Bot. 1872: 139

(1872), FI. Cap. 6: 388 (1897). Type: Cape, Thunberg (UPS,

holo.). A. dregeanum (Kunth) Bak. in J. Bot. 1872: 139 (1872).

A. zeyheri Bak. in J. Bot. 1872: 140 (1872). Type: Cape, Zeyher

4234. A. brevicaule Bak. in J. Linn. Soc. 15: 298 (1876), FI.

Cap. 6: 391 (1897). Type: Cape, Thunberg (UPS). A. scilli-

florum Eckl. ex Bak. in J. Linn. Soc. 15: 298 (1876). A. scilli-

florum Eckl., nom. nud. Type: Cape, Ecklon, Herb. Cap. No.

35b.

Caesia thunbergii Roem. & Schult. f., Syst. 7: 1695 (1830);

Bak. in FI. Cap. 6: 401 (1897). Type: as for C. contorta. C. ?

eckloniana Roem. & Schult. f., Syst. Veg. 7: 1691 (1830);

Bak. in FI. Cap. 6: 401 (1897); Leighton in FI. Cape Penins.

186 (1950). Type: as for A. sciliifloruin. C. ? dregeana Kunth,

Enum. 4: 611 (1843); Bak. in FI. Cap. 6: 401 (1897); Leighton

in FI. Cape Penins. 186 (1950). Type: Cape, Drege 8767 , 8768

(L). C. brevicaulis (Bak.) Dur. & Schinz, Consp. FI. Afr. 5:

353 (1893). C. brevifolia (Thunb.) Dur. & Schinz, l.c. C.

sciiliflora (Eckl. ex Bak.) Dur. & Schinz, l.c. C. zeyheri (Bak.)

Dur. & Schinz, l.c.

Bulbine brevifolia (Thunb.) Roem. & Schult. f., Syst. Veg.

7:451(1830).

Bulbinella brevifolia (Thunb.) Kunth, Enum. 4: 573 (1843).

Rhizomatous, caespitose, glabrous grass-like plants

10-70 cm tall, with woody roots. Leaves distichous,

usually numerous, filiform to linear, 5-35 mm long,

1- 4 mm broad, closely ribbed. Inflorescence a lax,

much branched panicle well exserted above the

leaves, but often reduced to a long simple raceme;

the scape erect, terete, branches wiry, often with

2- 4 accessory side branches at the nodes; pedicels

thin, 3-10 mm long, lengthening in fruit and then

firmer and patent-recurved. Flowers 1-4 in each

axil, developing successively, bracts small. Perianth

deciduous, spirally twisted when faded, blue, the

segments free, subequal, 3-nerved, apex conical

with a short, introrse, minute tail. Stamens 6, the

filaments somewhat swollen, retrorsely scabridu-

lous, anthers opening by longitudinal slits, basifix

with a dorsal pit, curving backwards with age. Ovary

trilocular, with 2 ovules in each locule, the one

projected upwards, the other downwards, style

terete, minutely papillate, stigma apical, minute.

Capsule globose, about 3 mm in diam. black; seeds

beanshaped, black, verrucose and with larger tuber-

cles arranged in irregular rows, strophiole discoid,

white.

Widely distributed and common in the Cape

Province from Namaqualand to East London,

usually montane.

Cape. — 3017 (Hondeklipbay): Namaqualand, 8 km N.N.W.

of Kotzesrus (-DD), strandveld, rare, forming big clumps,

Acocks 23400. 3227 (Stutterheim): Keiskamma Hoek, 3

km N. of Cata Forest Station (-CA), Story 3834. 3318 (Cape

Town): Stellenbosch, Jonkershoek Forest Reserve (-DD),

Taylor 5215. 3319 (Worcester): Great Winterhoek moun-

tains (-AA), Marloth 2312; Ceres (-AD), Bolus 8343. 3321

(Ladismith): Sevenweeks Poort (-AD), Andreae 1211. 3326

(Grahamstown): east of Howiesons Poort (-AD), Galpin

3091. 3418 (Simonstown): Miller Point (-AB), Galpin 12278;

west of old manganese mine (-BD), Boucher 953. 3424 (Humans-

dorp): 16 km W. of Humansdorp on Storms River road (-BA),

Story 2845.

A. A. Obermeyer

Bothalia 11, 1 & 2: 125-126 (1973)

Notes on Entada in South Africa

J. H. ROSS*

ABSTRACT

Miscellaneous information concerning the Entada species which occur in the area delimited for the

Flora of Southern Africa is presented, and the distribution of each species within this area is indicated.

Of the four species of Entada which occur in the

area delimited for the Flora of Southern Africa, three

were known to Harvey when he wrote his account of

the genus in FI. Cap. 2: 276 (1862).

The first species dealt with by Harvey was E. pur-

saetha DC. [Harvey referred to this taxon as E.

scandens (L.) Benth., a species which is now known to

be a synonym of E. phaseoloides (L.) Merr.]. The only

specimen cited by Harvey was “Wahlberg and the

locality given was “South Africa”.

Harvey l.c,p. 277, described E. wahlbergii, basing his

description of the species on a specimen in the Stock-

holm herbarium collected by Wahlberg. Once again,

the locality recorded was “South Africa”. Examina-

tion of this type specimen revealed that the label

bears the locality “Goda Hopps-udden” (Cape of

Good Hope). The Wahlberg specimen cited by Harvey

under E. scandens carries the same locality. However,

neither E. wahlbergii nor E. pursaetha occurs in the

Cape Province. The only area within the territories

covered by Flora Capensis where each of these two

species occurs and could have been collected is in

Natal or, more probably, in Zululand. Fortunately

it is known that Wahlberg visited Natal and Zululand

in 1843 (ffolliott & Liversidge, Ludwig Krebs-Cape

Naturalist to the King of Prussia: 138, 1971). There

was a rather casual attitude towards localities from

which specimens were alleged to have come on the

part of some collectors during the eighteenth century

and this probably accounts for the locality “Cape of

Good Hope” on the labels.

E. pursaetha is found in low lying areas along the

Zululand coast in riverine fringing vegetation and in

swamp forest. Some years ago a solitary plant ( Ward

5315) was found at Isipingo just south of Durban, but

it is not known whether or not this plant was intro-

duced. In any event the plant has since disappeared

and the species is now not known to occur south of

the Mtunzini district (see Fig. 1). Seeds of E. pursaetha

are frequently washed up on Natal beaches and are

sometimes even found on the Transkei coast.

In our area E. wahlbergii is found almost entirely in

Zululand and at altitudes below 350 metres. There

is a single collection ( Edwards 1659) from just south of

the Tugela river in the lower Tugela valley (see Fig. 1).

As E. wahlbergii occurs north and south of the Tugela

river it is unfortunately not possible to establish whe-

ther Wahlberg collected the type specimen in Zululand

or in Natal.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

5292-9

Fig. 1.— The known distribution of Entada arenaria, E. pur-

saetha and E. wahlbergii in the area delimited for the Flora

of Southern Africa.

The third species dealt with by Harvey was E. na-

talensis Benth., now regarded as a synonym of E.

spicata (E. Mey.) Druce. Harvey recognized var. acu-

leata within E. natalensis, basing the variety on “a

garden plant raised at Cape Town from Natal seeds,

Commis. Genl. J. D. Watt (Herb. Hk., D)”. Recently

the material housed at Trinity College, Dublin, was

examined and among it were two specimens from J. D.

Watt’s garden; one in flower and the other in fruit.

Despite Harvey’s mention of the presence of a speci-

men in Hooker’s herbarium, there is apparently no

specimen in the Kew Herbarium now.

The young stems and leaf-rhachides of the two syn-

types of E. natalensis var. aculeata, particularly those

of the fruiting specimen, are covered with numerous

scattered recurved prickles, whence the varietal epithet.

Recurved prickles are very variable in their occurrence

in E. spicata, some specimens having only a few, some

none at all, while others are very heavily armed. Con-

sequently, the numerous prickles on the syntypes of

var. aculeata are considered as no more than part

of the range of variation of the species and not worthy

of varietal rank. Harvey also drew attention to the

broader and more glabrous leaflets in var. aculeata

but, once again, with the more abundant material now

available, it is apparent that these characters likewise

form part of the overall range of variation of the

species and do not serve to distinguish var. aculeata

from the remainder of the material.

126

NOTES ON ENTADA IN SOUTH AFRICA

E. spicata is endemic in South Africa (see Fig. 2).

The record of E. natalensis Benth. (i.e. E. spicata) by

Sim, For. FI. P.E. Afr. : 53 (1909), probably refers to

E. schlechteri (Harms) Harms although the description

indicates E. spicata. E. spicata has not yet been

collected in Mozambique but it may well occur in the

southern part of that territory. Similarly, E. schlech-

teri, which is known only from southern Mozambique,

may yet be found in our area. Unlike E. spicata which

has yellowish-white flowers, E. schlechteri has red

flowers.

Fig. 2. — The known distribution of Entada spicata.

The earliest available name for the fourth taxon oc-

curring in our area is E. arenaria Schinz. Schinz, in

Mem. Herb. Boiss. 1: 118 (1900), based his descrip-

tion of E. arenaria on a specimen he collected in

Amboland in South West Africa. The type specimen

in the Zurich herbarium is a very poor one consisting

of portion of a stem plus the remains of a few old

pods. With such poor type material it is hardly sur-

prising that the type description is very brief and

rather unsatisfactory. Brenan, in FI. Zamb. 3, 1 : 20

(1970), adopted the name E. nana Harms for this

taxon but, in doing so, stated: “It is possible that E.

arenaria Schinz (in Mem. Herb. Boiss. 1: 118, 1900),

based on a type from S.W. Africa, is conspecific with

E. nana, and if this were proven it would be the earliest

name for the species. The type of E. arenaria is, how-

ever, so inadequate, consisting only of a few ancient

fragmentary pods, that I consider that this name

should be rejected as being of uncertain application.”

I have examined Schinz 277, the type of E. arenaria

and, despite its rather fragmentary nature and the

above comment, am of the opinion that the specimen

can be positively identified. All of the specimens col-

lected in South West Africa, especially those from the

type locality of E. arenaria, agree with Schinz 277 in

having longitudinally striate and densely puberulous

stems, and pods which are of similar dimensions and

texture. The specimens also agree in habit. Conse-

quently, I am satisfied that E. arenaria is an earlier

name than E. nana and must therefore be adopted.

In some ways this is perhaps unfortunate because

Harms provided a comprehensive description of E.

nana in Warb., Kunene-Samb. Exped. : 244 (1903) and

there is an excellent isotype in the Edinburgh herba-

rium. Furthermore, Brenan in Kew Bull. 20: 373

(1966) recognised subsp. microcarpa within E. nana,

basing the subspecies on a specimen from Zambia. As

E. arenaria is an earlier name than E. nana it becomes

necessary, therefore, to transfer subsp. microcarpa to

E. arenaria.

Entada arenaria Schinz in Mem. Herb. Boiss. 1 :

118 (1900). Type: South West Africa, mittellauf des

Omuramba ua Matako, Schinz 277 (Z, holo.!).

subsp. arenaria.

E. nana Harms in Warb., Kunene-Samb. Exped. 244 (1903);

Harms in Engl., Pflanzenw. Afr. 3 (1) : 403 (1915); Bak.f.,

Leg. Trop. Afr. 3: 787 (1930); Torre in Consp. FI. Angol. 2:

258, t.51 (1956); F. White, For. FI. N. Rhod. 92 (1962); Brenan

in FI. Zamb. 3, 1: 19 (1970). Type: Angola, Habungu, Baum

471 (E, iso.!).

subsp. microcarpa ( Brenan ) J. H. Ross, comb. nov.

Entada nana subsp. microcarpa Brenan in Kew Bull. 20: 373

(1966); FI. Zamb. 3, 1: 20 (1970). Type: Zambia, Mwinilunga

Distr., Dobeka Bridge, Milne-Redhead 4496 (K, holo.!).

E. sp. 2, F. White, For. FI. N. Rhod. : 92 (1962).

The material of subsp. microcarpa is still rather

limited and by no means homogenous in its pods.

The correct taxonomic status of subsp. microcarpa is

still not absolutely clear at this stage but, as indicated

by Brenan, it seems more prudent to regard it as a

northern subspecies rather than to regard it as specifi-

cally distinct.

Bolhalia 11, I & 2: 127-131 (1973)

Notes on Acacia Species in Southern Africa: Ilf

J. H. ROSS*

ABSTRACT

Miscellaneous information relating to several of the Acacia species in Southern Africa is provided.

The identity of A. petersiana Bolle is confirmed and a review of A. sieberana is given, tog :thcr with

a map showing the distribution of the variants within A. sieberana.

ACACIA CAFFRA ( THUNB .) WILLD.

Glover, in Ann. Bolus Herb. I; 146 (1915), did

not nominate type specimens when describing var.

tomenlosa and var. transvaalensis within A. caffra

(Thunb.) Willd., so all of the specimens of each

variety cited must be regarded as syntypes. As

thirteen specimens of each variety were cited, it is

desirable to select a lectotype for each variety. This

is particularly so in the case of var. tomentosa as

one of the specimens cited, Galpin 564 (GRA), is

a mixed gathering of A. caffra and of A. ataxacantha

DC. and another specimen, Burtt Davy 1534 (BOL),

is A. hereroensis Engl.

The relevant syntypes have been assembled and

from these specimens I now select Rehmann 4603

(PRE) from Wonderboompoort near Pretoria as the

lectotype of var. transvaalensis Glover, and Flanagan

302 (BOL) from hillsides near Komgha, October

1891, as the lectotype of var. tomentosa Glover. It

must be mentioned that there are two sheets of

Flanagan 302 in the Bolus Herbarium; the one

selected above as lectotype and another from Pros-

pect farm, near Komgha collected in October 1889.

The latter specimen is not to be mistaken for the

lectotype.

ACACIA DAVYI N.E. BR,

Burtt Davy, FI. Transv. 2: 346 (1932), cited the

specimen Houseman 21 as “a co-type” of A. davyi

tN.E. Br. However, this specimen is not a co-type;

it was not cited by N. E. Brown in Kew Bull. 1908:

1161 (1908) when he described A. davyi and it was

collected subsequent to the publication of the de-

scription. The date 26 July 1908 appears on the

specimen while the part of the Kew Bulletin con-

taining the description of A. davyi in the Kew lib-

rary carries the date c. 12 May 1908. Houseman 21

is a very poor sterile specimen and it is perhaps

fortunate that it is not a co-type as it is actually a

specimen of Albizia harveyi Fourn.

ACACIA KARROO HA YNE

In a recent paper on A. karroo Hayne (Ross in

Bothalia 10 (2): 387, 1971), doubt was cast on the

identity of Luderitz 122 as this collection was cited

by Schinz in Mem. Herb. Boiss. 1: 113, 115, 116

(1900) under three different taxa. While examining

specimens in the Zurich herbarium last year, it

was found that there are at least two different speci-

mens collected by Luderitz and bearing the number

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

122. The specimen of Luderitz 122 cited by Schinz

under A. horrida is A. karroo. However, another

specimen of Luderitz 122, cited by Schinz l.c. : 116

under A. afT. trispinosae, is A. erubescens Welw. ex

Oliv. The latter is a flowering specimen without

leaves or fruits but, as the inflorescences are spicate,

it is unlikely to be confused with the A. karroo

gathering of Luderitz 122.

ACACIA PERVILLEI BENTH.

Among the Southern African Acacia material

received on loan from the Trinity College Dublin

recently was a very old unnamed specimen alle-

gedly from Delagoa Bay. This specimen proved

to be A. pervUlei Benth., a Madagascan species.

Unfortunately the label on the specimen in question

is illegible except for the locality “Delagoa Bay”

and the fact that it was collected in the early 1800's.

The most logical explanation seems to be that the

label does not belong to the specimen for there is

no other record of A. pervillei from the African

continent and, if it was collected in the vicinity

of Delagoa Bay, it would surely have been collected

subsequently. However, before dismissing the speci-

men outright it is perhaps as well to recall that A.

rovumae Oliv., a species found along the east coast

of Africa, also occurs fairly commonly on the west

coast of Madagascar. Pending the discovery of a

further specimen of A. pervillei from the African

continent, it seems unwise at this stage to record the

species from Mozambique.

ACACIA PETERSIANA BOLLE

The identity of A. petersiana Bolle in Peters,

Reise Mossamb. Bot. 1: 4 (1861) has long been in

some doubt. Bak. f.. Leg. Trop. Afr. 3: 842 (1930),

held that A. petersiana seemed “nearly allied” to

A. spirocarpa Hochst. ex A. Rich. As it has usually

been assumed that no authentic material of .J. peter-

siana had survived, it was of great interest to find

an isosyntype of A. petersiana in the Paris herbarium

last year. This specimen was collected by Peters at

Sena in 1846 and was subsequently distributed

from the Berlin herbarium. The specimen, although

armed with short recurved spines only, is undoubted l>

conspecific with A. tortilis (Forsk.) Hayne. It is

unfortunate that the specimen is a flowering one

and does not have any fruits, but the indumentum

on the young branchlets, petioles, leat-rachides and

peduncles suggest that the specimen would be best

placed in subsp. spirocarpa (Hochst. e\ A. Rich.)

Brenan. This confirms that A. petersiana is a syno-

nym of A. tortilis subsp. spirocarpa.

128

NOTES ON ACACIA SPECIES IN SOUTHERN AFRICA: III

ACACIA SIEBERANA DC.

Chevalier, in Bull. Soc. Bot. Fr. 74: 959 (1927),

was the first to recognize an infraspecific taxon

within A. sieberana DC. when he described var.

vi/losa from Oure in the French Sudan. Keay &

Brenan, in Kew Bull. 5: 364 (1951), in addition to

maintaining var. vil/osa, recognized two further

varieties, namely, var. vermoesenii (De Wild.) Keay

& Brenan and var. woodii (Burtt Davy) Keay &

Brenan. The four varieties thus created within A.

sieberana were distinguished as follows:

Young branchlets glabrous or nearly so; branches of crown usually ascending var. sieberana

Young branchlets ± hairy, usually densely so; branches of crown usually widely spreading:

Indumentum of shoots short and rather downy, usually neither markedly golden nor villous; fruits

glabrous or nearly so even when young :

Median leaflets of pinnae 2-2,5 mm long, very rarely to 3 mm (West African) var. villosa

Median leaflets of pinnae 3-6 mm long (East and Central Africa) var. vermoesenii

Indumentum of shoots much coarser, normally villous and markedly golden especially when young;

fruits densely pubescent when young and slightly so even when old (Southern Africa) var. woodii

Brenan in FI. Trop. E. Afr. Legum. — Mimos.

127 (1959) recorded varieties sieberana, vermoesenii

and woodii from the area delimited for this Flora.

Under var. woodii Brenan noted:

“The variations in East Africa of A. sieberana

seem to fall into two groups, var. sieberana in one

and vars. vermoesenii and woodii in the other. . . .

As far as can be ascertained the habit and ecology

of var. woodii are decidedly those of var. vermoesenii

and not var. sieberana. . . . Careful field-work

is greatly needed in various parts of the range of

A. sieberana, which may show that the two groups

mentioned in the first sentence are subspecies or

even species; ... At present it seems more prudent

to maintain the three recognized varieties, especially

because they are all connected by intermediates.

These are particularly frequent between vars. ver-

moesenii and woodii, and it is hard to refer them

either to one or the other.”

The same three varieties were recorded by Brenan

in FI. Zamb. 3, 1 : 109 (1970) from the area delimited

for Flora Zambesiaca. Once again, Brenan com-

mented: “Indeed the two last-named (varieties ver-

moesenii and woodii ) are linked by puzzling inter-

Troupin in Bull. Jard. Bot. Brux. 35: 449 (1965)

reviewed A. sieberana throughout its range in Africa

and made certain significant changes to the above

treatment of the species. Troupin described a fifth

variety, namely, var. orientale, and recognized two

subspecies within A. sieberana, namely, subsp.

sieberana and subsp. vermoesenii. Subspecies sie-

berana included var. sieberana and var. orientale

while subsp. vermoesenii included var. vermoesenii

and var. woodii. Although Troupin (l.c. : 451) men-

tions var. villosa, there is no indication to which

subspecies this variety would be referred. The two

subspecies were distinguished on differences in

growth form and on whether or not the stipular spines

are persistent. Subspecies sieberana was characterized

by the ascending branches which form a narrow

or rarely rounded crown and the spines being long

and persistent, and subsp. vermoesenii by the spreading

branches and the typically rounded or mushroom-

shaped crown and the spines being persistent or

not persistent. Within subsp. sieberana, var. orientale

differed from var. sieberana in having grey-tomentose

young branchlets, different coloured leaflets and

in the persistent spines being initially tomentose.

As Troupin’s treatment of the species affected

the status of the Southern African var. woodii, it

was necessary to examine the species once more in

preparation for the account of Acacia being prepared

for the Flora of Southern Africa.

It is soon apparent from an examination of material

from throughout the distributional range in Africa \

that A. sieberana is an extremely variable species j

and that the characters on which the infraspecific ,

categories are based are themselves variable and

often vary independently. As indicated by Brenan i

(1959), varieties sieberana, vermoesenii and woodii j

are all linked by intermediates and these intermediates

are particularly frequent between varieties ver-

moesenii and woodii. Likewise, var. villosa is also

linked to the other varieties by intermediates.

Var. vermoesenii is distinguished from var. woodii ] I

essentially on the degree of development and the l

colour of the indumentum on the young branchlets.

A further distinction recorded is that in var. ver-

moesenii the pods are glabrous or almost so, while

in var. woodii the pods are typically densely golden-

pubescent, particularly when young. In var. woodii, '

however, the pods are frequently quite glabrous and

are indistinguishable from those of var. vermoesenii.

The capitula in var. vermoesenii are often larger

than those in var. woodii, but there is a complete

overlap and no clear distinction can be drawn on

the basis of this character. It is frequently difficult

to refer a specimen to one variety or to the other

with certainty and it is often a matter of personal

opinion as to whether a specimen should be placed j

in var. vermoesenii or in var. woodii. Indeed, the |

same specimen has sometimes been referred to var.

vermoesenii by one worker and to var. woodii by

another, while duplicates of the same gathering have

been noted as having been referred to different

varieties in different herbaria.

In the southern part of the species range typical

var. woodii occurs and even here considerable varia-

tion is encountered. The indumentum on the young

branchlets, leaves, peduncles and pods in typical

var. woodii is villous and distinctly golden, especially

when young. However, there is considerable variation

in the degree of pubescence and in the colour oil

the indumentum. The young branchlets vary from

glabrous or subglabrous to sparingly or densely

pubescent while the indumentum varies in coloui

from golden to faintly golden or greyish-white

Often in amongst the greyish-white indumentum a |

faint tinge of gold is visible, especially at the base I

of the petioles, spines and peduncles. Occasionally

the indumentum on the old shoots is greyish-white

while on young branchlets from the same plant the

indumentum is golden. There is often a difference

in the degree of pubescence of different organs or

a plant. The golden indumentum tends to be bes

developed in the southern part of the species range

J. H. ROSS

129

but this is only an overall tendency and it is possible

to find densely golden pubescent specimens in

Uganda.

Growth form varies considerably in Southern Africa

and plants with spreading branches and a flattened

spreading crown or rounded crown and plants

with ascending branches and a narrow crown can

be observed in most populations.

The bark in var. woodii is typically papery and

peels off irregularly (see Fig. 1), whence the common

name “paper-bark Acacia”. Often, however, the

bark does not flake or peel off at all (see Fig. 2).

Both forms frequently occur in the same population.

There seems to be a tendency for specimens with

non-peeling bark to have glabrous or glabrescent

branchlets and in Natal these plants are often, but

by no means always, confined to the floors of river

[ valleys near the coast. These plants appear to occupy

i slightly different ecological conditions to those

occupied by the densely pubescent specimens with

papery peeling bark. These glabrous specimens seem

1 best regarded as ± glabrous forms of var.

woodii since they are otherwise indistinguishable

1 from specimens of var. woodii.

In view of the nature of the variation within A.

sieberana and the presence of so many intermediates

between each of the varieties, I do not consider the

rank of subspecies created by Troupin to be ap-

F‘g- 1- — The typical papery, peeling bark of Acacia sieberana

var. woodii.

propriate and, consequently, do not intend to uphold

the two subspecies. The typical forms of var. ver-

moesenii and var. woodii do appear different, but

as they are linked by so many and varied intermediates,

and as difficulty is experienced frequently in

attempting to refer specimens to one variety or to

the other with certainty, it is not considered

desirable to maintain both of these varieties. Con-

sequently, it is intended to maintain but one of the

varieties, namely, var. woodii, and to relegate var.

vermoesenii to synonymy under var. woodii.

Although one of the distinguishing characters of

subsp. sieberana (which included var. orient ale)

was the presence of long persistent spines, examination

of the holotype and two isotypes of var. orientate

revealed that the holotype and one isotype are

devoid of spines while on the other isotype the

longest spines are only 0,5 cm long. None of the

morphological characters held to typify var. orien-

tate is peculiar to this variety alone; all of the charac-

ters may be found in numerous specimens in East,

Central and Southern Africa. Variety orientate, which

has no real distinguishing characters, is best regarded

as a local variant or ecotype and, when viewed in

relation to the range of variation within the species

throughout its range, cannot be upheld. Var. orien-

tate is likewise reduced to synonymy under var.

woodii.

Fig. 2.— The variant of Acacia sieberana var. woodii with

non-peeling bark.

130

NOTES ON ACACIA SPECIES IN SOUTHERN AFRICA: III

As indicated above, the three varieties still recog- mediates. However, the following key, by relying

nized within A. sieberana, namely, varieties sieberana, to some extent on the geographical distributions,

vi/losa and woodii, are linked by numerous inter- should enable the three variants to be identified.

Young branchlets glabrous or nearly so; branches of crown usually ascending; occurs in West, Central

and in East Africa var. sieberana

Young branchlets ± hairy, usually densely so, seldom glabrescent; branches of crown typically widely

spreading:

Indument m on branchlets short and rather downy; median leaflets of pinnae small, 2-2,5 (very

rarely to 3) mm long; West Africa eastwards to the Sudan var. villosa

Indumentum on branchlets usually coarser and often villous, golden to greyish-white, seldom glabres-

cent; median leaflets of pinnae usually 3-6 mm long; occurs in East, Central and Southern Africa. . .

var. woodii

The distribution of the variants of A. sieberana in Africa is shown in Fig. 3.

Fig. 3. — The distribution of the variants of Acacia sieberana in Africa.

It seems opportune at this juncture to provide

a fairly detailed synonymy and literature citation

of A. sieberana.

Acacia sieberana DC., Prodr. 2: 463 (1825);

Oliv. in FI. Trop. Afr. 2: 347 (1871); Benth. in Trans.

Linn. Soc. 30: 503 (1875); Bak. f., Leg. Trop. Afr.

3: 836 (1930); Brenan, Checklist Tang. Terr. 335

(1949); Bogdan in Nature in E. Afr., ser. 2, No. 1:

14 (1949); Eggeling & Dale, Indig. Trees Uganda,

ed. 2: 214, fig. 48 i (1952); Gilbert & Boutique in

F.C.B. 3: 166 (1952); Brenan in FI. Trop. E. Afr.

Legum. Mimos.: 127, fig. 17/57 (1959); F. White,

For. FI. N. Rhod. 84, fig. 17K (1962); Troupin in

Bull. Jard. Bot. Brux. 35: 449 (1965); v. Breitenbach,

Indig. Trees S. Afr. 2: 290 (1965); Brenan in FI. Zamb.

3, 1 : 107 (1970). Type: Senegal, Sieber 43 (G, holo.,

K!).

var. sieberana. Keay in FI. W. Trop. Afr., ed.

2, 1 : 499 (1958); Brenan in FI. Trop. E. Afr. Legum. —

Mimos.: 127 (1959); in F.Z. 3, 1 : 107 (1970).

A. sieberana DC., Prodr. 2: 463 (1825) sensu stricto; Gilbert

& Boutique l.c. 166 (1952), pro parte; Mullenders, Veg. Kania-

ma: 462 (1954), pro parte quoad specim. Mullenders 1660.

A. sing Guill & Perr., Tent. FI. Seneg. 1: 251 (1832). Syn-

types from Senegal (P !).

A. verrugera Schweinf. in Linnaea 35: 340, t. 9, 10 (1867-8);

Oliv. in FI. Trop. Afr. 2: 354 (1871); Benth. in Trans. Linn.

Soc. 30: 510 (1875); Taub. in Engl., Pflanzenw. Ost-Afr. C.:

195 (1895); Harms in Mildbr., Deutsch. Zentr.-Afr. Exped.

2: 235 (1911); R.E. Fr„ Schwed. Rhod.-Kongo Exped. 1:

64 (1914); Crowfoot, Flow. PI. Northern & Central Sudan,

fig. 81 (1928). Type: the Sudan, Kassala, by the River Gasch,

Schweinfurth 1963 (BM!, EA, K!, P! iso.).

A. purpurascens Vatke in Oesterr. Bot. Zeitschr. 30: 277

(1880); Bogdan in Nature in E. Afr., ser. 2, No. 1 : 14 (1914);

Dale, Trees & Shrubs Kenya: 70 (1936); Brenan, Checklist

Tang. Terr.: 335 (1949). Type: Kenya, near Mombasa, Hilde-

brandt 1938 (BM!, K! iso.).

A. blonmiaertii De Wild., PI. Bequaet 3: 58 (1925); Type:

Zaire, Dungu, Blommaert 67 (BR, holo.!).

A. nefasia sensu Lebrun, Veg. Plaine Alluv. Sud Lac Edouard:

290 (1947), pro parte, non (Hochst. ex A. Rich.) Schweinf.

A. sieberana subsp. sieberana var. sieberana, Troupin in

Bull. Jard. Bot. Brux. 35: 453 (1965).

J. H. ROSS

131

var. villosa A. Chev. in Bull. Soc. Bot. France

74: 959 (1927); Keay & Brenan in Kew Bull. 5:

364 (1951); Keay in FI. W. Trop. Afr., ed 2, 1: 499

(1958). Type: French Soudan, Oure, Chevalier 700

(P, holo.!, K!).

A. rehmanniana sensu Hutch. & Dalz. in FI. W. Trop. Afr.

1: 361 (1927); A. Chev. in Rev. Bot. Appl. 8: 197 (1928);

Bak. f.. Leg. Trop. Afr. 3: 838 (1930), pro parte quoad syn.

A. sieberana var. villosa, non Schinz.

A. sieberana var. rehmanniana (Schinz) Roberty in Candollea

11 : 142 (1948), quoad syn. A. sieberana var. villosa.

var. woodii ( Burlt Davy) Keay & Brenan in

Kew Bull. 5: 364 (1951); Pardy in Rhod. Agric. J.

48: 406 (1951); Wild, S. Rhod. Bot. Diet. 49 (1953);

Palgrave, Trees Centr. Afr. 254 (1956); Torre in

Consp. FI. Angol. 2: 281 (1956); Brenan in FI.

Trop. E. Afr. Legum. — Mimos. 128 (1959); Palmer

& Pitman, Trees S. Afr. 163, t.x, 45 (1961); v. Breiten-

bach, Indig. Trees S. Afr. 2: 292 (1965); De Winter

et al, Sixty-six Transv. Trees 56 (1966); Brenan in

F.Z. 3, 1: 108, t. 16/18 (1970); Ross, Acacia spp.

Natal: 40, fig. 2/8 (1971); Flow. PI. Afr. 40: t.

1653 (1972); FI. Natal 193 (1973). Type: Natal,

between Estcourt and Colenso, Wood 3528 (K, holo.!,

NH!).

Inga nefasia Hochst. ex A. Rich., Tent. FI. Abyss. 1 : 237

(1847) sensu stricto. Type: Ethiopia, without locality, Schimper

940 (P, holo.!; BM!, FI!, K!, P!).

A. nefasia (Hochst. ex A. Rich.) Schweinf. in Bull. Herb.

Boiss. 4, app. 2: 209 (1896); Lebrun, Veg. Plaine Alluv. Sud

Lac Edouard: 290 (1947), pro parte; Lebrun et al, Contr. Et. FI.

Parc Nat. Kagera: 57 (1948). Type as for Inga nefasia.

A. amboensis Schinz in Mem. Herb. Boiss. 1: 105 (1900);

Dinter in Feddes Repert. 15: 78 (1917); Bak. f., Leg. Trop.

Afr. 3: 838 (1930); O. B. Miller, Checklist Bech. Prot. 16

(1948); J. S. Afr. Bot. 18: 18 (1952). Syntypes: Angola, Omu-

panda in Uukuanjama, Wulfhorst 2 (Z !) ; precise locality

unknown, “Kilevi am Kunene”, Schinz 763 (Z!).

A. lasiopetala sensu Burtt Davy in Kew Bull. 1908: 158

(1908); Glover in Ann. Bolus Herb. 1: 149, t. 19/23 (1915);

Bews, FI. Natal 115 (1921); Henkel, Woody PI. Natal 226

(1934); Stapleton, Common Transv. Trees 6 (1937); Suesseng.

& Merxm. in Proc. & Trans. Rhod. Sci. Ass. 43: 16 (1951),

non Oliv.

A. cf. hebeclada sensu Wood in Trans. S. Afr. Phil. Soc. 18:

152 (1908).

A. verrugera Schweinf. f. latisiliqua Harms in Zentr. Afr.

Exped. 2: 235 (1911). Syntypes, Zaire and Ruanda, Mild-

braedSSl, 1104, 2108 (all B, t); Tanzania, Mwanza and Bukoba,

Holtz 1551, 1630. A probable but doubtful synonym.

A. katangensis De Wild, in Fedde Repert. 11: 502 (1913);

Ann. Mus. Congo Beige Bot., ser. 4, 11: 42 (1913); Contr.

FI. Kat.: 68 (1921). Type: Zaire, Katanga, Hock s.n. (BR,

holo.!).!

A. hebeclada sensu Bews, FI. Natal: 114 (1921), non DC.

A. woodii Burtt Davy in Kew Bull. 1922: 332 (1922); FI.

Transv. 2: 344 (1932); Steedman, Trees etc. S. Rhod. 15 (1933);

Hutch., Botanist in S. Afr. 394 (1946); O. B. Miller, Check-

list Bech. Prot. 21 (1948); Brenan, Checklist Tang. Terr.:

335 (1949); West in Rhod. Agric. J. 47: 208 (1950); Codd,

Trees & Shrubs Kruger Nat. Park 51, figs. 44a, b, 45 (1951);

O. B. Miller in J. S. Afr. Bot. 18: 26 (1952). Type as for A.

sieberana var. woodii.

A. monga De Wild., PI. Bequaert 3: 62 (1925); Contr. FI.

Kat. Suppl.l: 17 (1927). Type: Zaire, between Albertville and

Kongolo, Delevoy 356 (BR, holo.!).

A. vermoesenii De Wild., l.c.: 68 (1925); Brenan, Checklist

Tang. Terr. 335 (1949). Type: Congo Brazzaville, Boma,

Vermoesen 1378 (BR, holo.!).

A. abyssinica sensu Brenan, Checklist Tang. Terr. 335 (1949),

non Hochst. ex Benth.

A. sieberana var. vermoesenii (De Wild.) Keay & Brenan

in Kew Bull. 5: 364 (1951); Pardy in Rhod. Agric. J. 48: 406

(1951); Wild, S. Rhod. Bot. Diet. 49 (1953); Brenan in FI.

Trop. E. Afr. Legum. — Mimos. 128 (1959); in FI. Zamb. 3,

1 : 108 (1970). Type as for A. vermoesenii.

A. cf. stolonifera Burch, sensu Torre in Consp. FI. Angol.

2: 282 (1956) quoad specim. Gossweiler 11035 (K), pro parte.

A. sieberana var. orientate Troupin in Bull Jard. Bot. Brux.

35: 454 (1965). Type: Rwanda, Mutara, env. de Mimuri,

Troupin 3264 (BR, holo.!, EA, K!, LWO, YB1).

A. sieberana subsp. vermoesenii Troupin var. vermoesenii

(De Wild.) Keay & Brenan, Troupin l.c. 455 (1965); Schreiber

in FI. S.W. Afr. 58: 12 (1967).

A. sieberana subsp. vermoesenii Troupin var. woodii (Burtt

Davy) Keay & Brenan, Troupin l.c. 457 (1965).

ACACIA TRISTIS WELW. EX OLIV.

Schreiber, in Mitt. Bot. Mtinchen 6: 251 (1966),

considered that A. tristis Welw. ex Oliv. was not

specifically distinct from A. hebeclada DC. and

reduced it to subspecific rank within A. hebeclada,

citing the new combination as subsp. tristis (Welw.

ex Oliv.) Schreiber. However, A. tristis Welw. ex

Oliv. in FI. Trop. Afr. 2: 349 (1871) is a later homo-

nym of A. tristis R. Graham in Bot. Mag.t. 3420

(1835) and is, therefore, an illegitimate name. Conse-

quently, subsp. tristis must be regarded as a nomen

novum even although based on the same type specimen

as A. tristis Welw. ex Oliv. and the correct author

citation for subsp. tristis is subsp. tristis Schreiber.

.

Bothalia 11, 1 & 2: 133-137 (1973)

Stapelieae (Asclepiadaceae) from South Tropical Africa: VII

L. C. LEACH*

ABSTRACT

The identity of typical Caralluma caudata N.E. Br. is discussed; two subspecies are recognized,

one of which, subsp. rhodesiaca Leach, is described here, and details of their synonymy, variation and

distribution are given. Two putative intergeneric hybrids, with C. caudata one parent, are assumed

and from the evidence submitted the identity of one of these with Stapelia tarantuloides R. A. Dyer

is adduced.

Caralluma caudata N.E.Br. in FI. Trop. Afr. 4:

485 (1904); Berger, Stap. u. Klein.: 88 (1910); White

& Sloane, Stap. ed. 2, 1:348,3: 1 155 (1937); Jacobsen,

Handb. Succ. PI. 1; 242 (1960) et Sukk. Lexikon 108

(1970). Type: Malawi (Nyasaland), Namasi (Namad-

zi), Cameron 25 (K!).

Chromosome number: 2n — 22.

Although the variations in the precise size and shape

of the corolla and corona lobes of this showy, albeit

most malodorous, species repeat the pattern found in

most members of the genus, two quite distinct sub-

species may be recognised. One, the typical, rare ex-

cept in the Mbala (Abercorn) region, appears to be

restricted mainly, if not entirely, to the northern parts

of the Zambesi basin ; the other seems to be distributed

solely to the south of that river and extending into

the Sabi and Limpopo catchment areas is possibly the

most widespread and common stapeliad in Rhodesia.

The two taxa are distinguished mainly by their vege-

tative characteristics and are easily recognized by the

colour of their stems, which in subsp. caudata are olive

to dark olive-green, sometimes appearing to be brown-

ish or purplish from the markings and crowded speck-

lings with which they are often covered. Those of

* 157 Rhodesville Avenue, Greendale, Salisbury, Rhodesia.

subsp. rhodesiaca are pale greenish grey to greyish

green, usually with purplish red markings, and differ

from those of the typical subspecies, perhaps even

more significantly, in the length of their tapering stem

teeth which are, on average, about twice as long as

those of its relative. The flowers of subsp. rhodesiaca

are also generally paler in colour and are often more

markedly pubescent, while the spots on the disc are

seldom so crowded as to appear as solid colour, as is

more usual in subsp. caudata. However, these corolla

characters are, like those of the corona, subject to

considerable variation and flowers as dark as those of

the typical subspecies are not unknown from Rhode-

sia. Subsp. rhodesiaca generally flowers earlier than

subsp. caudata and in fewer flowered umbels, with

flowers opening in a more widely spaced succession,

so that it is seldom that more than two are open at a

time in any one inflorescence. In the typical subspecies,

on the other hand, flower development occurs more

nearly simultaneously, with the result that umbels

with 5-6 open flowers are by no means uncommon.

Despite their discrete distributions and although so

readily distinguished from each other, it is considered

that subspecific status is most appropriate for these

two taxa in view of their almost identical floral

characteristics.

Fig. 1. — Caralluma caudata subsp. caudata. Plant from Mbala, Zambia, cult. Nelspruit (14883).

134

STAPELIEAE FROM SOUTH TROPICAL AFRICA : VII

subsp. caudata

Caralluma longecornuta Croizat ex Gomes e Sousa in Rev.

Mozambique 4: 44 (1935) et 6: 20 (1936); White & Sloane, op.

cit. 1 : 35 (1937), nom. nud. Type: Mozambique, Niassa Distr.,

Mandimba, R. da Torre 4 (COI! PRE!).

Caralluma praegracilis Oberm. in White & Sloane, op. cit.,

3: 1161 (1937); Luckhoff, Stap. S. Afr.: 56, 57 (1952); Jacobsen,

Handb. Succ. PIT: 253 et Sukk. Lexikon 1 14 (1970), p.p. excl.

distrib. Rhodesia. Type: Zululand, Nongoma, Gerstner 752.

Caralluma caudata N.E. Br. var. fusca Luckhoff in White &

Sloane, Stap. ed. 2, 3: 144 (1937); Luckhoff, Stap. S. Afr. 53

(1952); Jacobsen, Handb. Succ. PI. 1 : 243 (1960) et Sukk. Lexi-

kon 108 (1970). No type cited, locality unknown.

In the absence of material, this variety is assigned to this

subspecies on the evidence of the photograph in White & Sloane,

l.c.. Fig. 287, and is placed in synonymy as it is considered

probably to be merely a slight colour variant caused by the spots

becoming more or less confluent over the whole corolla rather

than this being restricted to the disc.

Chromosome number: 2n=22. Leach & Brunton

10078 (SRGH).

It is perhaps unfortunate that the type specimen

of C. caudata originated from an area where the spe-

cies seems to be of such rare occurrence: since Camer-

on’s original gathering from Namasi (modern spelling

Namadzi), which lies about half way between Blantyre

and Zomba, it appears from the records to have been

collected only once in Malawi ( Reynecke 30). Dr.

Reynecke’s plant appears to be a good match for the

type material and is certainly identical with those from

around Mbala and adjacent parts of Tanzania. It is

among these latter that examples occur which appear

to be identical in all respects with the plant described

as C. praegracilis , and to which White & Sloane l.c.

referred as being common in Rhodesian garden rocker-

ies. It seems quite possible that these cultivated plants

originated from Mbala; certainly they are equally

disease resistant and tolerant of heavy rainfall. It is

considered that the Reverend Gerstner’s specimen may

well have been an escape from similarly cultivated

garden plants, especially as the species appears never

to have been again collected in Natal.

The specimen from Mandimba in northern Mozam-

bique (provisionally named C. longecornuta by Croiz-

at) does not, unfortunately, include stems, but the

flowers appear to match those of the type, especially

in respect of the “T” -shaped outer coronal lobes; it

is therefore, and in conformity with the distributional

evidence, considered to belong here.

Tanzania. — T4: Ufipa Distr., near Zambian borderon Mbala-

-Sumbawanga road, granite slope in association with Aloe

mzimbana Christian, cult. Greendale & Nelspruit, Leach &

Brunton 10078 (BM; BOL; BR; K; LISC; PRE; SRGH; ZSS);

ibid. cult. PRE 15872, Morony s.n. (PRE).

Zambia. — N: Mbala Distr., Nchalanga Hill near Mbala, cult.

& fl. at Nelspruit, M. Richards s.n. sub Leach 13296 (K; NDO;

PRE; SRGH); ibid. Morony s.n. sub Plowes 3705 (K); near

Mbala, cult. Nelspruit, Whellan s.n. sub Leach 12165 (BOL;

LISC); ibid. cult. SRGH, fl.. 26.iv.1972, Whellan s.n. sub Leach

14883 (SRGH); ibid. cult. SRGH 4699 G. Williamson s.n. (M;

NBG).

Malawi. — C: Lilongwe, cult. PRE, fl. Apr. 1954, Reynecke 30

(PRE; SRGH); Namasi, fl. April 1899, Cameron 25 (K).

Mozambique. — N: Niassa Distr., Mandimba, fl. Jan. 1937, R.

da Torre 4 (COI; PRE).

Rhodesia. — Origin unknown, common in cult., fl. 27. ii. 1958,

Leach 5445 (SRGH), idem, fl. 7.iv.l960, Leach 9834 (SRGH).

subsp. rhodesiaca Leach, subsp. nov.

Caralluma chibensis Luckhoff in S. Afr. Gardening & Country

Life 25: 56 (1935). Type: Rhodesia, Chibi, Miss Jackson in

Herb. Luckhoff 182.

Caralluma caudata var. chibensis (Luckhoff) Luckhoff in

White & Sloane, Stap. ed. 2, 1 : 352 (1937); Luckhoff, Stap. S.

Afr. 53 (1952); Jacobsen, Handb. Succ. PI. 1 : 243 (1960) et Sukk.

Lexikon 108 (1970). Type: as above.

Caralluma caudata var. stevensonii Oberm. in White & Sloane,

op. cit. 3: 1156 (1937); Jacobsen, op. cit. 1 : 243 (I960) et tom.

cit. 108 (1970). Type: “near Salisbury”, No. 34947 in Herb.

Transv. Mus.

Caralluma caudata var. milleri Nel in White & Sloane, op. cit.

3: 1158 (1937); Luckhoff, op. cit. 54 (1952); Huber in Prodr.

Fl. S-W. Afr. 114: 13 (1967); Jacobsen, op. cit. 1:243 (1960) et

tom. cit. 108 (1970). Type: Angola, Cubango (Okavango)

River, A. H. Miller 7390 in Herb. Stellenbosch (STEf).

A subspecie typica caulibus viridi-griseis vel griseo-

viridibus; caulium dentibus plus minusve 2-plo longio-

ribus; floribus plerumque pallidioribus, plus pubes-

sentibus facile distinguenda.

Type: Rhodesia, S: Belingwe Distr., Leach, E. J.

& W. Bullock 13145 (BM; BOL; BR ;K; LISC; PRE;

SRGH, holo.; ZSS).

Chromosome number: 2n=22. Baker s.n. sub Leach

14123 (SRGH).

Fig. 2. — Caralluma caudata subsp. rhodesiaca. Plant from Shawanoe

River, Mrewa, Rhodesia, cult. Greendale, Salisbury (49396).

L. C. LEACH

135

As it has not been found possible to trace any of the

type specimens involved it is considered advisable to

adopt an entirely new epithet for this southerly sub-

species in order to ensure that no possibility of error

may arise regarding its typification. The several varie-

ties included in the synonymy were based by their

respective authors on minor variations apparently in

single individuals. All of these appear to fall within the

range of variability to be observed in single field popu-

lations, and sometimes, particularly in respect of the

shape of the inner coronal lobes, in a single individual,

or even in a single flower.

Angola. — Cuando-Cubango Distr., N. bank of Cubango

(Okavango) River between Runtu and Nyangana (?Inhangana),

A. H. Miller s.n. (photo, of a plant in cult. PRE); ? clonotype

cult. Umtali, A. H. Miller sub P/owes 2580 (SRGH). The type

locality of var. milleri was originally recorded as Okavango, S. W.

Africa, but it appears that this was an error since, in a letter to

Mr. Plowes, Mr. Miller is most explicit that plants were found on

the north bank of the river and did not occur on the south. I

have been unable to find the place name Nyangana (Mr. Miller's

spelling) in any maps or gazeteers available to me, but it seems

certain that this is the Inhangana which is shown on one map as

being ± 93 km E of Runtu.

S.W. Africa. — Caprivi Strip, near Katima Mulilo, cult. PRE,

fl. March 1953 and Feb. 1954, Codd 7595 (BM; PRE; SRGH).

Rhodesia. — N: Gokwe Distr., Msadzi, cult. Gokwe, fl. Feb.

1964, Bingham 1107a (SRGH). Mrewa Distr., Shawanoe Riv.,

cult. Greendale, fl. & fr. l.xii. 1 954, Leach in SRGH 49396

(SRGH); ibid. cult. NBG, fl. 14.iv.1958, H. Hall 1146 (NBG);

ibid. cult. Nelspruit, fl. Jan. 1965, Leach & Muller 12166 (COI;

K; M; NBG; SRGH; WIND). W: Matobo Distr., E. Matopos,

cult. Greendale, fl. 27.ii.1958, M. Paterson sub Leach 5719

(SRGH); ibid. fl. 12.i.l958, Leach 5726 (SRGH); “Besna Ko-

bila” rock outcrop, fl. Dec. 1956, O. B. Miller 3979 (SRGH); ibid,

termite mound, fr. Sept. 1961, O. B. Miller 8019 (SRGH);

Kezi, cult. Burnside, fl. 30.i. 1966, fr. 7.viii. 1966, W. Bullock

11314/5 (SRGH); Gladstone Farm, cult. Burnside, fl. 30.i. 1 966,

W. Bullock 47 (SRGH); Anglesea Farm, cult. Bulawayo Mus.,

fl. 30.xii.1956, M. Paterson 208 (BOL); Rowallan Park, Mato-

pos, cult. Burnside, fl. Feb. 1964, W. Bullock 12 (K); Matopos,

± 67 km S.S.E. of Bulawayo, cult. Umtali, fl. 12. i. 1972, Plowes

3836A (PRE); Gordon Park, Matopos, fl. Feb. 1948, drawing

only, Plowes (PRE). Bulawayo Distr., Burnside, fl. 15. i. 1966,

W. Bullock 77 (K). Bulalima Mangwe Distr., Plumtree, D. A.

Robinson in SRGH 41636 (SRGH); ibid. cult. SRGH, fl. Feb.

1968, Drummond 8484 (SRGH). Insiza Distr., Filabusi, fl.

March 1949, R. M. Davies in SRGH 22687 (SRGH); ibid.,

cult. NBG, fl. 13. i. 1954, E. A. Schelpe in NBG 44247 (NBG).

C: Gwelo Distr., Mlezu Agr. Sch. Farm, fl. 30.xii. 1 965, Simon

580 (SRGH); ibid., fl. 5.iv.l968, Biegel 2591 (SRGH). Makoni

Distr., Chiduku Res., Plowes 2465 (K). E: ? “Sabi River”,

without precise locality, fl. 22. i. 1934, Vereker sub Eyles 7636

(K). Chipinga Distr., Chibunje, fl. 10.iii.1965 (SRGH), idem

cult. Nelspruit, fl. Dec. 1965 (PRE), idem cult. Umtali, fl.

20.iii.l967/>/tnves 2476 (SRGH); Sabi Gorge (Sabi/Lundi junc-

tion) cult. Umtali, fl. 6.ii.l965, Plowes 2464 (BOL; K; LISC);

Melsetter Distr., ± 35 km W of Melsetter, fl. 14. i. 1972, Lancas-

ter 11 (SRGH). S: Belingwe Distr., south of Mnene Mission,

cult. Nelspruit, fl. Jan. -Feb. 1966, Leach & Bullock 13145 (BM;

BOL; BR; K; LISC; PRE; SRGH; ZSS). Chibi Distr., Tokwe

Riv., cult. Greendale, fl. 21.U957, Leach 5467 (SRGH); “Hippo

Pools", Lundi Riv., cult. Umtali, fl. 29. ii. 1968, Buckland sub

Plowes 2644 (SRGH); “Lundi”, without precise locality, fl.

9. i. 1934, Vereker in Herb. Eyles 7632 (K; PRE), idem cult.

Salisbury, fl. 3.ii. 1 936 (SRGH). Gwanda Distr., 16 km W of

Gwanda, M. Paterson 304; 309 (SRGH). Victoria Distr., Ma-

shaba, fl. 30.xii. 1955, Leach 5571 (SRGH); Mushandike Riv.,

cult. Nelspruit, fl. Jan. 1965, Leach 12059 (K; PRE). Shabani

Distr., Shabi Riv., cult. Greendale, fl. Apr. 1960, Leach 9948

(SRGH); Ngesi Riv., fl. 26.xii.1959, Leach 9707 (SRGH); idem

cult. Greendale, fl. 20.ii.1960, Leach 9766 (PRE; SRGH). Bikita

Distr., Moodie's Pass, cult. Greendale, fl. 1 5.ii. 1960, Leach

9761 (LISC).

There is greater variation in both colour and size of

flower in this subspecies than in the typical ; the corolla

ground colour varies from yellowish cream to greenish

yellow, and occasionally to dark yellow similar to that

of subsp. caadata. The spots also vary considerably

in size and density and in colour from light brownish

red to purple or reddish violet. Some exceptionally

small-flowered plants were found by Mr. Plowes near

Chibunje and in the gorge near the junction of the Sabi

and Lundi rivers; however, the size of the flowers in-

creased when plants were placed in cultivation in Um-

tali, while similarly small flowered plants have been

found at a number of localities, sometimes in associ-

ation with plants displaying larger, more normally di-

mensioned flowers e.g. 16 km W of Gwanda, M.

Paterson 304 and 309. Var. milleri was based apparently

on a similar rather small-flowered specimen with co-

rolla lobes 22 mm long. Cuttings, reputed to be from

the type plant of this variety, were obtained from Mr.

A. H. Miller (now living in Hermanus) by Mr. Plowes;

these, cultivated in Umtali, bore even smaller flowers

with lobes averaging only 18 mm in length. There ap-

pears therefore to be no justification for taxonomic

recognition of these small flowered plants as the

variation from smallest to largest flowers appears to be

more or less continuous and to have no distributional

correlation.

Fig. 3.— Putative hybrid: Caralluma caudata x Stapelia kweben-

sis, assumed Fx Plant from Chibunje, Sabi Valley, Rho-

desia, cult. Nelspruit (12789).

136

STAPELIEAE FROM SOUTH TROPICAL AFRICA : VII

Fig. 4. — Left: Stapelia kwebensis, Dumela, Mozambique

(12307); centre and right: Caralluma caudata x S. kwebensis,

Chibunje, Sabi Valley; centre, assumed Fx (12789); right,

assumed F2 (14073).

The chromosome numbers given for these two taxa

are those published by Professor G. Reese in Portu-

galiae Acta Biologia 12: 1-23 (1971).

ASSUMED INTERGENERIC HYBRIDS

Caralluma caudata N.E.Br. X Stapelia kwebensis

N.E.Br.

A plant having every indication of being of this

parentage was discovered by Mr. Plowes near Chi-

bunje in the lower Sabi valley area of Rhodesia; this

was growing amongst a number of plants of C. cau-

data subsp. rhodesiaca with, as far as could be ascertain-

ed, no S. kwebensis in the immediate vicinity, although

this species is known to be not uncommon in the

general area.

Cuttings from this plant, cultivated at Umtali and

Nelspruit, produced rather attractive dark-coloured

flowers which appeared to be nearer to those of S.

kwebensis than of C. caudata. Plants were also raised

from seed obtained from the original wild plant and

flowers from these proved to be still nearer to those of

S. kwebensis, although the stem characters displayed

by the original (assumedly Ft generation) were retained

to a greater or lesser extent in this F2 generation.

From the foregoing it seems that the floral charac-

teristics of C. caudata are recessive, at least when

hybridised with S. kwebensis, and that, in the present

instance, C. caudata was probably the mother plant,

with long stem teeth dominant.

Rhodesia. — E: Chipinga Distr., Chibunje, Sabi Valley, cult.

& fl. Umtali, Feb. 1965, D. C. H. Plowes 2477 ( PRE; SRGH);

idem cult. c6 fl. Nelspruit Sept. /Nov. 1965, Plowes, 2477 sub

Leach 12789 (K; KIEL; LISC; SRGH); ibidem, seedling ex

Plowes, assumed F2, cult. & fl. Salisbury, sub Leach 14073

(SRGH).

Stapelia tarantuloides R. A. Dyer in Flow. PI. Afr.

18: t.717 (1938). Type: Rhodesia, Gwanda Distr., S.

Thompson s.n. in PRE 24162 (PRE!).

It has long been suspected that the plant on which

this species was based might prove to be of hybrid

origin, with Stapelia gigantea N.E.Br. one parent. Ap-

parent confirmation of this has recently become evi-

dent in plants which have appeared spontaneously and

independently in Mr. & Mrs. Bullock’s Burnside

garden and in Mr. Cannell’s at Rangemore, both in

the Bulawayo District. No plants similar to these have

ever been collected by either Mr. & Mrs. Bullock or

Mr. Cannell, although the Bullocks, in particular, are

known to have made numerous trips over a number of

years in an effort to rediscover plants corresponding

to Sheilah Thompson’s. Other searches of the general

area west of Gwanda and especially along the Mwewe

River by Mr. D. C. H. Plowes, the author and others

have also all resulted in failure.

However, the plants which have appeared in the

two gardens appear to conform in all respects with

that originally described under the name S. tarantulo-

ides; that these are of hybrid origin seems to be beyond

dispute and since the only other species cultivated in

Mr. Cannell’s garden were S. gigantea and Caralluma

caudata N.E.Br., it seems almost equally certain that

these were the parents. The same two species were also

plentifully represented in the Burnside garden and

although C. lutea N.E.Br. (=C. lateritia N.E.Br.) was

also present, both naturally and in cultivation, this is

considered to be unlikely to have been involved.

Fig. 5. — Stapelia gigantea x Caralluma caudata, assumed Fi

hybrid (W. Bullock GH 2 sub Leach 14746B).

L. C. LEACH

137

Of the three species, S’, gigantea is plentiful in the

Mwewe River area and C. caudata is not uncommon;

C. lutea has not, on the other hand, so far been recor-

ded.

From the foregoing it seems reasonably safe to as-

sume that S. tarantuloides is a hybrid of 5. gigantea

and C. caudata. The assumption that C. caudata should

be one of the parents appears to receive support from

the similarity of the stems of the putative hybrid

Stapelia kwebensis N.E.Br. X C. caudata , discussed

elsewhere in this paper, and in which the floral charac-

teristics of C. caudata also appear to be recessive and

long stem teeth dominant.

Rhodesia. — W: ± 80 km W of Gwanda, near Mwewe Riv.,

in the Matopos Hills, under protection of low thorny Acacia

sp., Sheitah Thompson s.n., cult Messina, N. Transvaal, in

PRE 24162 (PRE!). Bulawayo Distr., hort. Bullock, spontane-

ous in cultivation, fl. 1970, Bullock GH2, sub Leach 14746

(PRE; SRGH), idem, with deeper tube and more attenuate

corolla lobes, sub Leach 14746A (PRE; SRGH), idem cult. Salis-

bury, fl. Nov. 1971, sub Leach 14746B (K; KIEL); hort. Can-

ned; spontaneous in cultivation, fl. May 1971, Canned 465

(PRE, SRGH).

ACKNOWLEDGEMENTS

I am much indebted to and wish to express my ap-

preciation and thanks to: —

Mr. & Mrs. E. J. Bullock and Mr. I. C. Cannell of

Bulawayo for many live plants of the Stapelieae and

for material preserved in liquid.

Dr. L. E. Codd, Director, Botanical Research Insti-

tute, Pretoria (PRE) and Mr. R. B. Drummond, Chief

Botanist, Government Herbarium, Salisbury, (SRGH)

for making the facilities of the respective herbaria and

libraries freely available to me as well as for much

technical assistance personally.

The Directors of the Department of Botany, British

Museum (BM); the Bolus Herbarium, Cape Town

(BOL); the Instituto Botanico Dr. Julio Henriques,

Coimbra, (COI); the Centro de Botanica da Junta de

Investigagdes do Ultramar, Lisbon (LISC); and the

National Botanic Gardens, Kirstenbosch (NBG), for

the loan of much valuable material.

The Director, Royal Botanic Gardens, Kew (K) for

the loan of type and other material and Mr. R. M.

Polhill of that institute for an extract from a letter

addressed by Mr. Cameron to a Mr. W. J. Thistleton,

which confirms the locality of “Namasi” and its

identity with Namadzi.

Mr. D. C. H. Plowes for a great many specimens of

Stapelieae in liquid as well as for live plants and

especially for much information, particularly that

relating to C. caudata var. milleri.

Bothalia 11, 1 & 2: 139-141 (1973)

Notes on Southern African Tuberales

W. F. O. MARASAS* and JAMES M. TRAPPEf

ABSTRACT

Three species of Tuberales have been found in Southern Africa. Terfezia pfeilii Henn. occurs in the

Kalahari Desert and adjacent areas of the Cape Province, Botswana and South-West Africa. The

other two, Terfezia austroafricana sp. nov. and Choiromyces echinulatus sp. nov., are known only

from the Cape. C. echinulatus is the first representative of that genus to be collected in Africa or the

Southern Hemisphere.

INTRODUCTION

The first record of Tuberales from Southern

Africa is Henning’s (1897) description of Terfezia

pfeilii Henn. from Damaraland, South-West Africa.

The occurrence of this species in Damaraland was

later noted by Pole Evans (1918); the cracks in the

Kalahari sand resulting from growth of its hypo-

geous ascocarps were illustrated by Leistner (1967).

Two other species, Terfezia boudieri Chat, and

T. claveryi Chat., were reported from the Kalahari

and Windhoek (Marloth, 1913; Pole Evans, 1918).

The “ Terfezia claveryi " that they examined as well

as all others so labelled in the Mycological Herbarium

of the National Herbarium, Pretoria (PRE), were

subsequently redetermined as T. pfeilii by Dr P. H. B.

Talbot (unpublished). Marloth did not describe

his “ Terfezia boudieri" and Pole Evans apparently

did not see any specimens. We have been unable

to locate any collections that were labelled “ Ter-

fezia boudieri" by either. Accordingly, until now

the only member of the Tuberales authenticated for

Southern Africa is T. pfeilii. The closest record of a

different species is that of Terfezia decaryi Heim

in Madagascar (Heim, 1934).

While noting that three species of Terfezia have

been reported from South Africa and South-West

Africa, Story (1958) stated that “some records are

vague and incomplete and cannot be checked.” In

view of the confusion about these species, we

restudied the Tuberales at PRE. Trappe’s earlier

studies of the types and supplementary collections

of the entire order of Tuberales, including the Ter-

feziaceae (Trappe, 1971) provided interpretive back-

ground. Collections of Southern African Tuberales

were also kindly provided by the Herbaria of the

Botanical Institute of the University of Torino

(TO), Oregon State University (OSC), and the U.S.

National Fungus Collections (BPI). This paper

should be regarded as a preliminary contribution,

since further collecting in Southern Africa will almost

certainly produce additional hypogeous fungi.

The collections examined were either dried or,

in the case of some at PRE, preserved in ethanol-

formalin-glycerine-water solutions. Tissues and spores

were examined in these mounting media: (1) 5%

KOH, (2) cotton blue-lactic acid, (3) Melzer’s reagent,

and (4) lactophenol, the mount being heated gently

over a gas flame. Spores were drawn from the cotton

blue mounts. Spores measured the same in all mount-

ing media.

Key to the Tuberales of Southern Africa

Asci mostly cylindric, borne in crowded hymenial layers; spores mostly uniseriate, prominently echinulate,

12-18//. broad (including spines) 1. Choiromyces echinulatus

Asci subglobose, ellipsoid, reniform, or asymmetric; spores mostly biseriate to irregularly arranged, mostly

broader than 1 8/x (including ornamentation):

Spores minutely echinulate, mostly 18-23// broad (including spines) 2. Terfezia pfeilii

Spores prominently spinose-reticulate, 25— 30/x broad (including spines) 3. Terfezia austroafricana

1. Choiromyces echinulatus Trappe & Marasas,

sp. nov.

Ascocarpa viva eburnea, in statu sicco nigrescens.

Gleba solida, marmorata, hymeniis inclusis. Asci

cylindrici, 140 x 12-17//, octospori, in hymeniis

mutue compressis. Sporae globosae, 12-18//. latae

(cum ornamentis), echinulatae bactulis et conis

1-2 x 0,5-1 (-1,5)//. Peridium cellulis inflatis mul-

tis.

* Plant Protection Research Institute, Department of Agri-

cultural Technical Services, Private Bag XI 34, Pretoria.

tU.S. Department of Agriculture, Forest Service, Pacific

North-west Forest and Range Experiment Station, Forestry

Sciences Laboratory, Corvallis, Oregon 97331, U.S. A.

Type: Cape, Gordonia near Upington, in red

sand dune, June 1, 1961, Leistner 2612 (PRE 42202,

holotype).

Ascocarp pale cream-colored and subglobose

when fresh, as dried with a black peridium and a

solid gleba marbled by dark brown veins embedding

hymenial palisades; opposing hymenial palisades

deformed from pressing against each other.

Hymenium of thin-walled, mostly collapsed para-

physes 4// broad growing among hyaline, thin-walled

asci. Asci mostly 8-spored; in youth cylindric to

ellipsoid or saccate and with spores irregularly

arranged; by maturity mostly cylindric, 140 X 12-17//,

140

NOTES ON SOUTHERN AFRICAN TUBERALES

hyaline, thin-walled, with stems long tapered below

the basal spores to narrow bases and spores uni-

seriate or occasionally biseriate.

Spores globose, 12-18/a broad with ornamen-

tation, 10-14/a excluding ornamentation, light yel-

low to light brownish yellow. Walls 1/a thick, light

blue in cotton blue. Ornamentation of round-tipped

rods and cones 1-2 X 0 , 5— 1 (— 1 ,5)/a, pale yellow

in KOH and light blue in cotton blue, ca 20-25

around the spore circumference, unconnected except

for some spores on which barely perceptible lines

on the spore surface join occasional ornaments.

Glebal veins of subparallel-interwoven, hyaline,

thin-walled hyphae 3-6/a broad at septa, the cells

frequently inflated. Peridium of hyphae with light

yellow, thin walls, 4-8/a broad at septa but with

many cells greatly inflated (15-50/a broad).

Cape. — Gordonia near Upington, in red sand dune, June

1, 1961, Leistner 2612 (PRE 42202, holotype).

This species has the solid gleba marbled with

veins and embedded hymenia with elongate asci

that typify the genus Choiromyces. It differs from

previously described Choiromyces spp. in having

echinulate spores. C. venosus (Fr.) Th. Fr., the most

common species of Europe, has larger spores promi-

nently ornamented with irregular tubes and rods.

The other Choiromyces species have either ridged

or pitted spores. C. echinulatus is the first member

of the genus found in Africa or, for that matter,

in the Southern Hemisphere.

2. Terfezia pfeilii Hennings, Engler Bot. Jahrb.

22:75 (1897).

Ascocarps 2, 5-6, 5 X 2, 5-5, 8 cm, subglobose to

obpyriform or turbinate, lacking a basal mycelial

tuft but with a basal attachment scar. Peridium

c. 1 mm thick, prominently wrinkled (particularly

on the upper surface), blackish brown with the

wrinkles yellowish. Gleba yellowish white, fleshy,

solid, marbled with white veins. Odour rather strongly

fungoid.

Asci randomly arranged in fertile pockets, (5-)

8- spored at maturity, typically subglobose but some-

times ellipsoid or obovoid, 70-100 x 50-80/a, hyaline,

thin-walled, sessile to substipitate, readily separable

from the glebal hyphae, spores loosely arranged

within.

Spores globose, ( 1 6— )1 8— 23(— 26)/a with ornamen-

tation, at first hyaline, by maturity pale brown.

Walls 1,5/a thick, 2-layered, blue in cotton blue.

Ornamentation appearing as a minutely papillose,

mucilaginous-granulose epispore in KOH or lacto-

phenol but clearly seen as densely crowded, minute,

deeply staining spines 1 — 1 , 5(— 2)/a tall in cotton

blue.

Glebal fertile pockets separated by sterile but

otherwise undifferentiated veins, the hyphae hyaline,

thin-walled, 5-1 2/a broad at septa but the cells generally

inflated (up to 20/a) to appear pseudoparenchymatous.

Peridium with an outer layer of large hyphae 4-12/a

broad at septa but with cells inflated to 15-30/a to

give a cellular appearance; inner peridial layer of

generally circumferentially aligned hyphae 4—1 0/u

broad at septa, the cells often slightly inflated.

Fruiting from April to June in sand dunes of the

Kalahari Desert and adjacent areas of the Cape,

South-West Africa and Botswana.

South-West Africa.— Damaraland: Hennings (PRE 15021,

lectotype; TO, possible isotype). Gibeon: Burger (PRE 42082,

PRE 42203). Gobabis: Verbiicheln (PRE 43905), Keetmans-

hoop: Kinges (PRE 36986). Kleinkaras: Hill (PRE 17799;

BPI, Lloyd 38083; OSL, Trappe 1309; TO). Windhoek: Gies I

(PRE 42076).

Botswana.— Ghanzi: Scholtz (PRE 41869).

Kalahari Desert. — Bottomley (PRE 44310); Nash (PRE

44254); Weintraub (PRE 32394). Askham: Strydom (PRE

44245). Kalahari Gemsbok National Park: Story 5616 (PRE

41602); Le Riche (PRE 41870).

Cape. — Kakamas: Oosthuizen (PRE 26335). Postmasburg:

Hunter (PRE 11293). Prieska: MacCleod (PRE 11619). Uping-

ton: Leistner 2610 (PRE 42201). Vryburg: Stephens 527 (PRE

36103).

Terfezia pfeilii has been suggested as a synonym

of several other Terfezia species by various European

authors. Our examinations of the types of all Ter-

fezia species have established beyond doubt that

T. pfeilii is distinct and confirm Mattirolo’s (1922)

illustrations that clearly show the differences in ,

spore ornamentation between T. pfeilii and the

other species. The earlier confusion about T. pfeilii .

stems in part from the paucity of good specimens

available for study by earlier authors and in part

from the nature of the spore ornamentation of T. pfeilii.

The spines are so crowded and minute that they

cannot readily be seen even with an oil immersion

objective unless stained.

Only two described species of Terfezia, T. olbiensis

Tul. & Tul. and T. leptoderma Tul. & Tul., resemble

T. pfeilii in having spores ornamented with very

small spines. The spores are larger and the spore

ornamentation much taller (2-3/a), coarser, and more

openly spaced in T. olbiensis and T. leptoderma

than in T. pfeilii. Although Mattirolo (1922) cor-

rectly illustrated the echinulate nature of spores

of T. pfeilii, he suspected it to be synonymous with

T. pinoyi Maire, which he also correctly illustrated

as lacking spines. The two are readily separable

by this difference in spore ornamentation as well

as by the amyloid reaction of asci of T. pinoyi with

Melzer’s reagent and the lack of that reaction by

T. pfeilii.

Dr Talbot’s redetermination of PRE collections

labeled “ Terfezia claveryi ” as T. pfeilii, noted earlier

in this paper, was confirmed by our studies. Ac-

cordingly, T. claveryi is not known to occur in

South Africa, notwithstanding the reports of Marloth

(1913), Pole Evans (1918), and Doidge (1950).

3. Terfezia austroafricana Marasas & Trappe,

sp. nov.

Ascocarpae glebis solidis, marmoratis. Asci el-

lipsoidei, obovoidei, subcylindrici, reniformes, vel

asymmetrici, 90-140 X 30-80/a, plerumque octospori.

Sporae globosae, 25-30/a latae (cum ornamentis),

spinosae-reticulatae, spinis (2-)3-5(-6)/a altis. Peri-

dia cellulis inflatis multis.

Type: Cape, Griqualand West near Barkly West,

E. L. Stephens (PRE 35577, holotype).

Dried ascocarps orange brown to brownish black,

smooth; dried gleba brown in young specimen,

ochraceous in older specimen, solid, marbled with

pallid veins.

Asci randomly arranged in fertile pockets, (4-)8-

spored at maturity, ellipsoid to obovoid, subcylindric,

reniform or asymmetric, 90-140 x 30-80/a, hyaline,

W. F. O. MARASAS AND JAMES M. TRAPPE

141

thin-walled, astipitate or with a short basal pro-

tuberance; spore arrangement occasionally uni-

seriate but mostly incompletely biseriate to irregular.

Spores globose, 25-30/* broad with ornamentation,

1 6— 22/u. excluding ornamentation, hyaline in youth

and pale yellow at maturity. Walls 1-2/* thick, light

blue in cotton blue. Ornamentation of truncate to

round-tipped spines (2-)3-5(-6) x 1-3/*, connected

by walls to form a partial to complete reticulum

of 4-6 sided, irregularly sized alveoli; alveoli 3-7

across the spore diameter; reticular walls 0,5/* thick,

variable in height from very low to as high as the

spines. Occasional spores ornamented with rounded

warts and no reticulum or with crowded spines

only erratically connected by low walls.

Glebal fertile pockets separated by sterile but

otherwise undifferentiated veins, the hyphae hyaline,

thin-walled, 5-12/* broad at the septa but the cells

generally inflated (up to 22/*). Peridium of hyphae

10-30/* broad at septa, the cells mostly inflated

(up to 60/*) to give a cellular appearance.

Cape. — Griqualand West: Barkly West, E. L. Stephens

(PRE 35577, holotype); Kimberley, April 10, 1918, Wilman

(PRE 11542, paratype).

Terfezia austroafricana belongs to subgenus Mat-

tirolomyces (Fischer) Trappe by virtue of its large,

elongate asci and uncrowded spores (Trappe, 1971).

The other species presently assignable to this sub-

genus— T. decaryi, T. terfezioides (Matt.) Trappe, and

T. spinosa Harkn. — also have prominently reticulate

spores. Of these, T. spinosa of North America most

closely resembles T. austroafricana in microscopic

characters (the fresh ascocarps have not been described

for either); T. austroafricana, however, has fewer,

larger, and more regular alveolae on the spore surface

and larger, more inflated cells in the peridium and

gleba. The spores of T. austroafricana are larger

than those of T. terfezioides, and the asci are smaller

than those of T. decaryi. All of these species are

closely related in anatomical characters, but only

T. terfezioides is known from a large number of

collections. As the others are collected again and

become better known, some may prove to merit

only varietal status.

ACKNOWLEDGEMENTS

Spores were drawn by Mrs Jackie Atzet. Trappe’s

participation in the studies was financed in part by

The American Philosophical Society, the Society

of the Sigma Xi, and the U.S. National Science

Foundation (GB-27378).

REFERENCES

Doidge, Ethel M., 1950. The South African fungi and lichens.

Bothalia 5: 1-1094.

Heim, R., 1934. Observations sur la flore mycologique mal-

gache. Ann. Cryptogam. Exot. 8: 1-10.

Hennings, P., 1897. Fungi camerunenses I. Engler Bot. Jahrb.

22: 72-111.

Leistner, O. A., 1967. The plant ecology of the southern

Kalahari. Mem. Bot. Surv. S. Afr. 38: 1-172.

Marloth, R., 1913. The flora of South Africa. Vol. 1, xi,

Eumycetes, p. 20-34. Cape Town.

Mattirolo, O., 1922. Osservazioni sopra due ipogei della

Cirenaica e considerazioni intorno ai generi Tirmania e

Terfezia. Mem. Reale Accad. Naz. de Lincei Roma Ser.

5, 13: 543-568.

Pole-Evans, I. B., 1918. Note on the genus Terfezia; A truffle

from the Kalahari. Trans. R. Soc. S. Afr. 7: 117-118.

Story, R., 1958. Some plants used by the Bushmen in ob-

taining food and water. Mem. Bot. Surv. S. Afr. 30:

1-115.

Trappe, J. M., 1971. A synopsis of the Carbomycetaceae and

Terfeziaceae (Tuberales). Trans. Brit. Mycol. Soc. 57:

85-92.

C'V'. i ■'

%■/ c.r

©

10 u

V ,

T

\

©

■

Fig. 1-3. — Ascospores. Fig.

1, Choiromyces echinulatus,

PRE 42202. Fig. 2, Ter-

fezia pfeilii, PRE 17799.

Fig. 3, Terfezia austroafri-

cana, PRE 15542.

5292-10

. ,

.

'

Bothalia 11, 1 & 2: 143-151 (1973)

Polyporus baudoni Pat. on Eucalyptus spp. in South Africa

G. C. A. VAN DER WESTHUIZEN*

ABSTRACT

The morphology and anatomy of the fruit-bodies and characteristics in pure culture, of Polyporus

baudoni Pat., which is associated with death of species of Eucalyptus and other trees in South Africa,

are described. The anatomical characters of these fruit-bodies agree with those of the type specimen,

as well as with those of the type specimen of Phaeolus manihotis Heim which is shown to be synony-

mous. The anatomical characters of the fruit-bodies and the cultural characteristics differ from those

of Polyporus schweinitzii Fr., the type of the genus Phaeolus Pat. The fruit-bodies and cultures display

combinations of characters not known to occur in any other species of polypore.

INTRODUCTION

A disease of Eucalyptus species in plantations in

northern Natal, which results in death of the trees

in increasing numbers, has been causing concern

during the past few years. Death of the trees is as-

sociated with the appearance of a fungus with large,

bright orange-yellow to yellow-brown stipitate, poroid

fruit-bodies at the bases of trunks and on the ground

among the trees. This disease was first reported by

Luckhoff (1955) who referred to this fungus as

Ganoderma colossum. Subsequent study by the present

author confirmed the association between the death

of the trees and presence of the fungus but raised

doubts about its identity.

The fungus has now been identified as Polyporus

baudoni Pat., first described from the Congo (Patouil-

lard, 1914). A fungus, Phaeolus manihotis Heim,

first described from Madagascar on manihoc (Heim,

1931), was reported by Brunck (1965) on Cassia

siamea from Ghana and Upper Volta, on Gmelina

arborea from the Ivory Coast and Upper Volta and

on Eucalyptus spp. from Brazza Congo and South

Africa. Browne (1968) recorded this fungus in as-

sociation with death of trees in East Africa and

suggested that Phaeolus manihotis Heim is identical

to Polyporus baudoni Pat. The fungus does not

appear to be well known and very few references

to it exist in the literature. It is the purpose, in this

paper, to furnish detailed descriptions of the mor-

phology and anatomical characters of the fruit-bodies

found in Natal and of the cultures made from them

and to report on the results of comparisons made

with the type specimens of Polyporus baudoni Pat.

and Phaeolus manihotis Heim and cultures and

fruit-bodies of Polyporus schweinitzii Fr., the type

species of the genus Phaeolus Pat. (Donk, 1960).

METHODS

Fresh and dried fruit-bodies were examined micro-

scopically according to the methods described by

Teixeira (1956) and Van der Westhuizen (1971).

Thin slices of tissue cut and removed in a radial-

longitudinal plane from various parts of the fruit-

bodies, were carefully teased apart in water or lacto-

phenol with the aid of fine, sharpened darning

*Plant Protection Research Institute, Department of Agri-

cultural Technical Services, Private Bag x 134, Pretoria.

needles under 25 X magnification of a stereo-micro-

scope. Hyphae dissected out in this way were mounted

in water or lactophenol and examined with a 100

X oil immersion lens.

Cultures were obtained by removing small pieces

of context tissue from freshly broken surfaces of

newly collected fruit-bodies with the aid of sterile,

fine-pointed forceps and placing these on agar plates

which contained 1,5% Oxoid malt extract soli-

dified by 1,5% agar in distilled water.

The cultural characters were studied and described

according to the methods of Nobles (1948) from

cultures growing on 1,5% Oxoid malt extract agar

plates as described above, inoculated at the side

of the dish, and incubated in the dark at 25 °C for

6 weeks. The cultures were tested for the presence

of extra-cellular oxidase enzymes by growing them

at 25°C for seven days on plates of 1,5% malt agar

to which 0,5% of gallic acid and 0,5% of tannic

acid had been added according to the method de-

scribed by Nobles (1948) and Van der Westhuizen

(1971).

All colours are named according to the notation

of Ridgway (1912).

DESCRIPTIONS

Carpophores (Figs. 1-7).

Fruit-bodies terricolous or lignicolous annual,

solitary, or grouped; pileus orbicular, applanate to

somewhat depressed or dimidiate and often depres-

sed behind, mostly broadly obconical, tapering

sharply to a reduced base and frequently attached

to a hypogeous, soft, pseudosclerotium or mycelial

sheath or pad attached to stem or root of host,

simple or with imbricate or connate lobes, spongy

when fresh drying to soft, brittle corky to soft woody,

light, 12-70 cm in diameter or up to 40 x 20 x 2-8

cm in dimidiate forms; upper surface finely pubescent

to glabrous and crustose over rounded lumps, un-

even, tuberculate, often unevenly concentrically sul-

cate, mat, in bright orange yellow colours, “Buff

yellow” to “Apricot yellow” “Warm buff” to “Zinc

orange” darkening to “Mars yellow” or “Sudan

brown” on drying. Margin obtuse, thick, rounded,

entire or undulate to somewhat lobate, concolourous,

sterile below. Pore surface bright yellow, “Lemon

yellow” to “Apricot yellow” when fresh, drying

darker, occasionally to “Antique brown” to “Mummy

144

POLYPORUS BAUDONI ON EUCALYPTUS SPP. IN SOUTH AFRICA

brown”, poroid; pores angular, to labyrinthiform

1-4/mm; dissepiments, even to somewhat lacerate,

thin; tubes concolourous with pore surface, 1-4

mm deep in one layer. Context “Pale orange yellow”

to “Orange buff” or “Mikado orange” when fresh

drying to “light orange yellow” “Warm buff” or

“Antimony yellow” . 5-7 cm thick, soft felty, azonate

or concentrically zoned, darkening in KOH.

Hyphal characters: (1) hyphae hyaline or faintly yel-

lowish, branching, thin-walled, nodose-septate, and sim-

ple-septate, 3, 0-5, 0/a in diameter (Fig. 8); (2) hyphae

as in (1) but somewhat swollen and filled with bright

yellow or yellowish brown refractive contents 7,0-

15,0ju in diameter (Fig. 10); (3) fibre hyphae straight

or flexuous, unbranched or occasionally branched

with walls thickened, yellowish, refractive and lumina

narrowed or occasionally occluded, aseptate, or

with occasional simple septa, 2,0-3,0/u in diameter

(Fig. 9), swelling and fracturing when mounted

in 5% KOH solution (Fig. 11).

Hymenium: basidia long-clavate, 14-24 x 5-8/m

with four slender sterigmata 2, 5-4,0 p long; basi-

diospores hyaline, long ellipsoid to almost cylin-

drical, thin-walled, smooth, 5-11 x 3,0-5,0/m, mostly

6, 5-7, 5 X 3, 5-4, 5 p, non-amyloid, (Fig. 13, 14).

Construction: At the margin the fruit-body con-

sists of branching, thin-walled, septate hyphae,

hyaline to pale yellowish, with dense, deeply staining

contents and agglutinated into strands. Immediately

behind the margin many narrow fibre hyphae are

present, intertwined with thin-walled, septate hyphae,

some of which have dark-coloured luminal contents.

At the upper surface, the thin-walled, septate hyphae

are branched and intertwined, many with dark-

coloured contents, and agglutinated, together with

fibre hyphae into bundles that project upwards, their

ends bent over in all directions and agglutinated

by a dark-brown, resinous substance into the thin,

hardened surface of the dry fruit-body (Fig. 12).

Dark-coloured zones, 1 0— 20/x thick, of similar con-

struction to the upper surface, may be present in the

context at different levels. The context consists of

hyaline or yellowish, flexuous, thick-walled, asep-

tate fibre hyphae and thin-walled, branching sep-

tate hyphae often with bright yellow or yellow-

brown refractive contents, intertwined into a fairly

loose, soft, homogeneous tissue. The dissepiments

consist mainly of yellowish aseptate fibre hyphae,

with walls slightly thickened, their lumina relatively

wide and often with yellowish contents, and tightly

intertwined in a vertical direction. Numerous thin-

walled septate hyphae with short, lateral branches

which bear the basidia at the hymenial surfaces

are present between the fibre hyphae. Intertwined

with the hyaline hyphae of the dissepiments, are

numerous, thin-walled, septate hyphae, somewhat

swollen and filled with a bright yellow, granular or

refractive substance, which apparently imparts the

bright yellow colour to the dissepiments.

Cultural characters (Fig. 15-22).

Growth is moderately rapid, the colonies reaching

radii of 12-35 mm in one week and covering the

plates in 3-4 weeks. The margin is even to slightly

bayed with the mycelium raised to the limit of growth.

The advancing mycelium is thin, silky, colourless or

faintly yellowish becoming gradually thicker, more

dense and woolly over the older part of the mat and

gradually forming woolly patches along the sides

of the dish. Over the inoculum the mycelium is quite

dense, woolly to felty and “Naphthalene yellow”

to “Buff yellow” in colour. This dense mycelium

gradually increases in size and darkens over the

younger parts over the mat to form dense felty areas

of “Mustard yellow” to “Buckthorn brown” whilst

patches of woolly mycelium on the sides of the dish

turn “Naphthalene yellow” or darker. The mat may

change colour slowly or the felty dark-coloured

aerial mycelium may spread and darken until the

entire mat may be felty and “Chamois”, “Naples

yellow” “Ochraceous tawny” or “Cinnamon brown”

in 5-6 weeks, or, the mat may become irregularly

appressed and lacunose in patches with low, ir-

regular ridges of dense felty “Warm buff” to “Anti-

mony yellow” mycelium, grown together in an ir-

regular pattern.

The reverse bleaches slowly and a strong, sweetish

but nauseating odour is given off. The agar is sof-

tened under the mat.

On gallic acid and tannic acid malt agar no diffusion

zones are formed but colonies of up to 35 mm in

diameter are formed in 7 days on both media.

Hyphal characters

Advancing mycelium: hyphae hyaline, thin-walled,

nodose-septate with simple clamps, branching near

the septa, 3, 0-5,0 p in diameter (Fig. 17).

Aerial mycelium: (a) hyphae as in the advancing

zone; (b) narrow hyaline hyphae with simple septa

and thin walls and deeply staining contents, branching

near the septa, the branches long, often forming

H-connections, 1,5-3,0/m in diameter (Fig. 18,21);

(c) hyphae as in (b) but with brownish, lacquer-like

contents and often distended in parts up to 10/m in

diameter (Fig. 20); (d) fibre hyphae hyaline or pale

yellowish, long, narrow, unbranched or occasionally

branched, the walls thickened, refractive and lumina

narrowed, aseptate with deeply staining contents

or with brownish contents, arising from septate,

thin-walled hyphae 2,0-4,0/a in diameter (Fig. 19,

22), swelling and fracturing readily in KOH mounts

(Fig. 11).

Submerged mycelium: hyphae hyaline, wide, thin-

walled, nodose-septate and simple septate, without

contents or with deeply staining contents.

Specimens examined.

in Herb PRE: 14204, Swaziland, 1918; 17052, on grass, Ma-

galiesberg, Pretoria Dist., Jan., 1923; 23764, on Celtis

rhamnifolia, Wonderboompoort, Pretoria, 1926; 30832,

on ground, Park Rynie, Natal, Jan., 1939; 32379, on

ground in thick grass, Park Rynie, Natal Jan., 1938;

35995, on dead wood, Marracuene, Mocambique, May,

1946; 40660, pine plantation, Dukuduku Forest Reserve,

Natal, Dec., 1952; 41021, on roots of leguminous liane,

Dukuduku Forest Reserve, Natal, Dec., 1954; *42148,

Dukuduku Forest Reserve, Natal, Dec., 1961; 43155,

base of tree in indigenous forest, George, C.P., Feb., 1966;

44621, base of living coppice of Eucalyptus, Dukuduku,

Dec., 1967; *44622, on ground among Eucalyptus, Duku-

duku, Dec., 1967; 44623, on ground near living Eucalyptus,

Dukuduku, Dec., 1967; 44624, at base of living Euca-

lyptus, Dukuduku, Dec., 1967 ; 44625, at base of Eucalyptus

stump with 2 living stems, Dukuduku, Dec., 1967; 44626,

on ground, Dukuduku, Dec., 1967; 44627, on ground

near decayed Pinus sp. stump, Dukuduku, Jan., 1969;

44628, on ground, attached to Eucalyptus, root, Dukuduku,

Jan., 1970; 44629, at base of dead Eucalyptus tree, Duku-

duku, Jan., 1970; *44630, at base of Eucalyptus coppice

shoot, Dukuduku, Jan., 1970; *44631, at base of dead

stem of Eucalyptus, Dukuduku, Jan., 1970; 44632, on

piece of dead wood on ground, Dukuduku, Jan., 1970;

lii

■ 'ip

%

\ \ •«

G. C. A. VAN DER WESTHUIZEN

145

*44633, on ground, on living root of Eucalyptus , Duku-

duku, Jan., 1970; *44634, on ground near dead Euca-

lyptus maculata , Dukuduku, Jan., 1970; *44636, with

pseudosclerotium on root of dead Eucalyptus , Dukuduku,

Jan., 1970; *44637, on root of Pinus taeda, Dukuduku,

Jan., 1970; *44638, on ground among Eucalyptus, Duku-

duku, Jan., 1970.

in Herb. Forest Products Research Laboratory, Princes Ris-

borough: as Ganoderma colossus (Fr.) Bres., on Cassia

siamea, Miralya near Mwanza, Tanganyika, July 1957.

(Culture No. 338); in Herb. Museum National d’His-

toire Naturelle, Paris; Phaeolus manihotis Heim sp. nov.

Aux environs du lac Alaotra (Madagascar) Avril 1930.

Leg. G. Bouriquet. (Type).

in Herb. Patouillard in Farlow Herbarium; Polyporus baudoni,

Mr. Baudon No. 1616, 29 May 1912, FH sheet No. 2456

(Type).

Observations in the field.

Fruit-bodies growing at the bases of trunks of

Eucalyptus trees, closely resemble those of large

bracket shaped polypores. Unlike these, the fruit-

bodies of P. baudoni are not formed by hyphae which

grow out of the trunk tissues. These fruit-bodies

develop instead, from pseudosclerotial tissue com-

posed of pale yellow mycelium and sand grains

that ensheath the roots and underground portions

of the stems of affected trees in a layer which may

vary from a thin weft of mycelium to up to 2 cm

in thickness (Fig. 6, 7). Consequently, these fruit-

bodies are very easily removed from the tree trunks

to which they are merely closely appressed. Removal

of the fruit-body hardly leaves a mark on the trunk

whilst the point of severance from the pseudosclero-

tium frequently escapes observation. The fruit-

bodies growing on the soil surface also come away

with little or no resistance when plucked and fre-

quently without showing scars where they have been

attached to the pseudosclerotia. Only by carefully

removing the soil underneath the fruit-body, does

its attachment to a root by a larger or smaller pseu-

dosclerotium become visible (Fig. 3).

The fresh, actively growing fruit-bodies are bright

in colour and quite watery. They appear on the

ground and against tree trunks and stumps from the

middle of December and may survive for six weeks

or longer, depending on weather conditions. In a

wet season they develop to maturity, then darken

with age until autolysis sets in when they deliquesce

into black, soggy masses which finally dry to a few,

black, friable, cinder-like flakes. In dry seasons

the fruit-bodies darken, dry up, and become rigid,

friable and light in weight. It not is known how these

disintegrate but they also disappear before the

following fruiting season.

Observations on type material

The type specimen of Phaeolus manihotis Heim

agrees with Heim’s (1931) description. The context

is felty and consists mainly of unbranched or oc-

casionally branching fibre hyphae with pale yellow

to brownish-yellow thickened walls, and narrow,

aseptate lumina, often flexuous to almost tortuous

and 2,0-6,0/x, mostly 3,0-3,5/x, in diameter, swelling

and fracturing when mounted in KOH. These are

intertwined with hyaline to pale yellow, thin-walled

hyphae, mostly collapsed, and thin-walled, nodose-

septate hyphae with brown, somewhat refractive

contents and often distended and up to 1 2/x in dia-

meter.

(indicates culture examined as well).

No septa or clamps were seen on the hyaline,

thin-walled hyphae but these are continuous with

the nodose-septate hyphae with brown contents.

The upper surface consists of fibre hyphae and

thin-walled hyphae agglutinated by a yellowish

brown substance into a dense hardened layer in

which no detail can be distinguished. No basidia or

spores were present (Fig. 23).

Heim (1931) described the basidia as clavate

measuring 11-14 x 6-8/i whilst basidiospores are

obovoid to sub-cylindrical with a small hilar appen-

dage and smooth thin, hyaline wall, measuring 5 , 5-7 x

3,2-4,3/x. He also emphasized the thin superficial

crust on some parts of the fruit-bodies, the charac-

teristic straw-yellow to rusty-fawn colour of the

felty context of the fruit-body which consists mostly

of thick-walled, unbranched hyphae, and thin-

walled excretive hyphae which originate from thin-

walled subhyaline hyphae with septa and clamps,

and the presence of a stipe which is a prolongation

of the pileus.

In phytopathological notes on P. manihotis, Heim

(1931) described the formation of a thick mycelial

muffler or sheath on the subterranean parts of

manioc and cajanus plants which results in death

of the plants and from which the large spongy fruit-

bodies are formed.

The type of Polyporus baudoni Pat. consits of a

number of slices cut from a large fruit-body. These

agree with the morphological characters of Patouil-

lard’s (1914) description. These fruit-bodies with

felty context consist mainly of fibre hyphae which

are hyaline or sub-hyaline, flexuous, unbranched

or occasionally branched, thick-walled, with narrow

lumina, aseptate and 2-5 p. in diameter. Thin-walled,

hyaline hyphae, mostly collapsed and shrunken are

present among them and thin-walled hyphae with

yellowish-brown contents and walls distended up

to 8/x in diameter are present in fairly large numbers.

Basidia are clavate with 4 slender sterigmata, mostly

20 X 6fi; basidiospores long ellipsoidal to short

cylindrical, with hilar apiculum, smooth, hyaline,

thin-walled, 5-7 X 3,0-3,5/t (Fig. 24).

In his description Patouillard (1914) stated that

the fungus resembles a triangle resting on its apex

and with the two lateral sides concave. The mycelium

forms greyish filaments which include soil and plant

debris but does not form a true sclerotium.

In the present investigation it was found that

the context hyphae of Patouillard's specimen swell

and fracture when mounted in 5% KOH so that

it was impossible to observe micro-morphological

detail in mounts for microscopic examination made

in this medium. Water or lactophenol could be

used for this purpose but no detail of the type of

septation of the thin-walled hyphae could be distin-

guished even in these media.

Fruit-bodies of Polyporus schweinitzii Fries, the

type of the genus Phaeolus Pat. (Donk, 1960) and

cultures made from them, were examined for com-

parison with those of the South African collections

of Polyporus baudoni. Their cultural characters agreed

closely with the descriptions by Nobles (1948). In

both cultures and carpophores only one type of

hypha, viz: brownish, branching hyphae with simple

septa and 3-12/n in diameter, made up to bulk of

the tissues whilst a brown, resin-like substance was

146

POLYPORUS BAUDONI ON EUCALYPTUS SPP. IN SOUTH AFRICA

present in segments of these hyphae in the trama

of the tubes of the fruit-bodies. These hyphae do

not swell and disintegrate when mounted in KOH.

Cultures and fruit-bodies examined: DAOM 1897,

under Pinus mughus. Central Experimental Farm,

Ottawa, (det. L.O. Overholts) July 1931; DAOM

F9420, on fallen Picea sitchensis, Oyster River, B.C.

August 1939; DAOM 31930, on ground. Fallow-

field Ont., August 1955; DAOM 72512, on ground.

Bells Corners, Ont. July 1957.

Basidiospore size.

It appears from the above descriptions that the

basidiospores of the South African collections are

both longer and wider than those of the type speci-

mens and of the original descriptions. For comparison,

dimensions of basidiospores of the type specimens

and from some South African collections selected

at random, are presented in Table 1.

Table 1. — Size of basidiospores of the type specimen of Poly-

porus baudoni, and the isotype of Phaeolus manihotis and

South African collections.

Specimen Size range

Polyporus baudoni. Type, in FH 5, 0-7,0 X 3 , 0-3 , 5 p.

*Phaeoius manihotis. Isotype, in K.. . . 5, 5-7, 4 X 3,2^4,0/x

P. manihotis, (description Heim, 1931) 5, 5-7,0 X 3,2-4,0/x

PRE 23764 5, 0-8,0 X 3,0-4, Op

PRE 44623 6, 0-8, 5 X 3,(M,5/x

PRE 44624 6,0-10,0 X 3, 5-4, 5p

PRE 44637 5,0-10,0 X 3, 0-4, 5/*

Although smaller, the spore sizes of the type

material and descriptions are within the limits of

and in agreement with the most common size, 7,0 x

4,0/a, of the basidiospores of the South African

collections.

DISCUSSION

From the description it is clear that the fruit-body

of Polyporus baudoni has a dimitic hyphal system

according to the concepts of Corner, (1932a, & b).

The hyphal characters of P. baudoni do not agree

in all respects with Corner’s (1932b) descriptions

as the thick-walled hyphae do not project beyond

the thin-walled hyphae at the margin of the carpo-

phore. The thick-walled hyphae are rather fragile and

develop behind the margin which is composed of thin-

walled, septate hyphae. The thick-walled hyphae are

however the terminal, thick-walled, aseptate, prolon-

gations of lateral branches of thin-walled, septate

hyphae and thus agree morphologically with Corner’s

(1932a, b) definition of skeletal hyphae.

The thin-walled, septate hyphae agree with Cor-

ner’s (1932a) definition of generative hyphae, since

all other structures in the carpophores arise from

them. They also give rise to the widened hyphae

with dark-coloured contents. These resemble con-

ducting hyphae but do not form a continuous net-

work or system, as described by Talbot (1954), in

the tissues. Their contents stain in phloxine but not

in Melzers’ solution. Although they also partly

resemble gloeocystidia, the absence of a staining

reaction with Melzers’ solution and the fact that

they may from part of generative hyphae, negate

this possibility. They therefore appear to be generative

hyphae in which metabolic products of unknown

nature have accumulated.

* (Size given by Dr D. N. Pegler at K. to Prof. J. L. Lowe,

private communication).

The fruit-bodies of Polyporus baudoni thus differ

in some aspects of their construction from fruit-

bodies with dimitic hyphal systems of other species

of polypores described before (Corner, 1932b; Van

der Westhuizen, 1971).

In cultural characters Polyporus baudoni is un-

usual in that the thin-walled advancing hyphae are

regularly and abundantly nodose-septate whilst

many young, thin-walled aerial hyphae are simple-

septate. This character together with the negative

oxidase reaction and the presence of thick-walled

fibre hyphae, refer cultures of Polyporus baudoni

to Nobles’ (1965) Key Code 1, 5, 8, a position which

it occupies by itself. This unique combination of

characters together with the characteristic bright

yellow colours of its cultures, allow the recognition

of cultures of this species without difficulty.

The structures formed in cultures of Polyporus

baudoni are also present in the carpophores from

which they were made. In the cultures, the thin-walled,

septate hyphae are prominent in the aerial mycelium

where they give rise to thick-walled fibre hyphae

whilst clamps develop gradually at some of their

septa. These hyphae differ from the thin-walled,

septate, branching hyphae in the context of the

carpophores only by being narrow and more or

less straight. They agree with the hyphae in the

carpophores in respect of their septation, the fact

that fibre hyphae arise from them and that their

contents may become dark-coloured in parts. These

narrow thin-walled, septate hyphae are thus homo-

logous structures to the thin-walled, septate hyphae

of the carpophores.

The thin-walled, nodose-septate hyphae of the

advancing zone of the cultures resemble the hyphae

from the margin of the fruit-bodies in all respects

whilst the fibre hyphae from the cultures are also

similar to those of the carpophores.

The South African specimens described here,

resemble the type specimen of Phaeolus manihotis

very closely in respect of morphology, hyphal charac-

ters and anatomy but basidiospores of the South

African collections are generally larger. This may

indicate a specific difference between the South

African and the type collections. In the South African

collections, however, considerable variation in spore

size is evident and the spore sizes of the type spe-

cimens occur within the range of sizes found in the

South African specimens. It appears that descrip-

tion of a new species for the South African material

is not justified and, furthermore, may be considered

to be conspecific with the type of Phaeolus mani-

hotis Heim.

The characters of the type specimen of Phaeolus \

manihotis Heim also agree closely with those of

the type specimen of Polyporus baudoni Pat. The

colour, construction and texture of the fruit-bodies

and the morphology of the thick-walled hyphae,

basidia and basidiospores of the two specimens >

agree very closely. However, although thin- walled \

hyphae are present in the fruit-body of Polyporus y

baudoni, it was not possible to determine the nature

of their septation with certainty because of the

poor condition of preservation and fragility of the

specimen. This raises a certain amount of doubt

about conspecificity of the two species but, because

they agree so well in other characters, it seems that

Browne’s (1968) suggestion that the two species

G. C. A. VAN DER WESTHUIZEN

147

are conspecific, is justified. For this reason, the

South African collections are referred to Polyporus

baudoni Pat.

Collections of Polyporus baudoni have been named

as Ganoderma colossum on previous occasions but,

although some superficial resemblance exist between

the two species, the absence of the characteristic

ganodermoid basidiospore from fruit-bodies of P.

baudoni, and their presence in those of G. colossum

(Furtado, 1965) negates any possible relationship

between these two species.

It is evident that types of hyphae present in cul-

tures and carpophores of Polyporus baudoni, as

described here, are not present in cultures and

carpophores of Polyporus schweinitzii Fr., the type

species of the genus Phaeolus Pat. (Donk, 1960).

Because the presence or absence of different types

of hyphae in carpophores are regarded as of phylo-

genetic importance by Nobles (1958), Teixeira (1962)

and Van der Westhuizen (1971), the two species

cannot be regarded as being phylogenetically related

and are thus not congeneric. Polyporus baudoni

thus cannot be classified in the genus Phaeolus Pat.

The species cannot be assigned to a more suitable

genus either, because no other genus in which the

complex of characters described for this species,

is known to exist at present. Since its cultural charac-

ters were shown to be unique, it is possible that a

new genus may have to be created to accommodate

Polyporus baudoni Pat.

ACKNOWLEDGEMENTS

Sincere thanks are due to Dr Derek A. Reid,

Royal Botanic Gardens, Kew, and Prof. Josiah L.

Lowe, State University College of Forestry, Syracuse

N.Y., for identification of this fungus and much

valued comments and information; to Mrs. Frances

F. Lombard of the U.S. Forest Disease Laboratory,

Madison, for valuable comments on cultures of this

fungus; to Dr Mildred K. Nobles and Dr J. H. Ginns,

Forest Mycologists of the Plant Research Insti-

tute, Ottawa, for cultures of Polyporus schweinitzii;

Dr J. A. Parmelee, Curator of the Mycological

Herbarium (DAOM) for the loan of fruit-bodies of

P. schweinitzii; and Dr R. Heim, Museum National

d’Histoire Naturelle, Paris and Dr C. Alippi, of

the Farlow Herbarium, Boston, for the loan of the

type specimens, and to my colleague, Dr W. F. O.

Marasas for his interest and valued comments and

to Mr C. S. Moses, formerly Pathologist to the

Department of Forestry, Pretoria, for bringing this

fungus to my attention, and valuable discussions.

REFERENCES

Browne, F. G., 1968. Pests and diseases of forest planta-

tion trees. An annotated list of the principal species occur-

ring in the British Commonwealth, p. 923. Clarendon Press,

Oxford.

Brunck, F., 1965. Parasites des Plantations Forestieres

d’Afrique tropicale et de Madagascar et mesures de pro-

tection. Bo is et Forets des Tropiques No. 103: 17-25.

Corner, E. J. H., 1932a. The fruit-body of Polystictus xan-

thopus, Fr. Ann. Bot. 46: 61-111.

Corner, E. J. H., 1932b. A Fomes with two systems of hyphae.

Trans. Brit. Mycol. Soc. 17: 51-81.

Corner, E. J. H., 1953. The construction of polypores. I.

Introduction: Polyporus sulphureus, P. squamosus, P.

betulinus and Polystictus microcyclus. Phytomorphology

3: 152-167.

Donk, M. A., 1960. The generic names proposed for Polypo-

raceae. Persoonia 1: 173-302.

Furtado, J. S., 1965. Ganoderma colossum and the status

of Tomophagus. Mycologia 57: 979-984.

Heim, R., 1931. Le Phaeolus manihotis sp. nov., parasite du

manioc a Madagascar, et considerations sur le genre

Phaeolus Pat. Annales de Cryptogamie Exotique , Tome 4:

175-189 + 3 plates.

Luckhoff, H. A., 1955. Two hitherto unrecorded fungal

diseases attacking pines and eucalypts in South Africa.

J. S. Afr. For. Ass. No. 26: 47-61.

Nobles, M. K., 1948. Studies in forest pathology VI. Identi-

fication of cultures of wood-rotting fungi. Can. J. Res.,

C, 26: 281-431.

Nobles, M. K., 1958. Cultural characters as a guide to

the taxonomy and phylogeny of the Polyporaceae. Can.

J. Bot. 36: 883-926.

Nobles, M. K., 1965. Identification of cultures of wood-

inhabiting Hymenomycetes. Can. J. Bot. 43: 1097-1139.

Patouillard, N., 1914. Quelques champignons du Congo.

Bull. Soc. Myc. Fr. 30: 337-338.

Ridgway, R., 1912. Color standards and color nomenclature.

Washington, D.C.

Talbot, P. H. B., 1954. Micromorphology of the lower

Hymenomycetes. Bothalia 6: 249-299.

Teixeira, A. R., 1956. Metodo para estudo das hifas do

carpoforo de fungos Poliporaceos. Sec. Agr. Est. Sao

Paulo, Inst. Bot. 1-23.

Teixeira, A. R., 1962. The taxonomy of the Polyporaceae.

Biological Reviews 37: 51-81.

Van der Westhuizen, G. C. A., 1958. Studies of wood-

rotting fungi: I. Cultural characteristics of some common

species. Bothalia 7: 83-111.

Van der Westhuizen, G. C. A., 1971. Cultural characters

and carpophore construction of some poroid Hymeno-

mycetes. Bothalia 10: 137-293.

148

POLYPORUS BAUDONI ON EUCALYPTUS SPP. IN SOUTH AFRICA

Fig. 1-7. — Polyporus baudoni. Fig. 1. Fruit-bodies on ground and at base of Eucalyptus trunk. Fig. 2.

Fruit-body arising from base of trunk of Eucalyptus sp. Fig. 3. Root of Pittus sp. with soil removed

to show attachment of fruit-body. Fig. 4. Pore surface of fruit-body. Fig. 5. Longitudinal section

of fruit-body showing inverted triangle shape. Fig. 6. Immature fruit-body with subterranean

pseudo-sclerotium. Fig. 7. Close-up view of Eucalyptus rootlet ensheathed by pseudo-sclerotium.

G. C. A. VAN DER WESTHUIZEN

149

Fig. 8-14. — Polyporus baudoni, micromorphological characters of carpophores. Fig. 8. Thin-walled

hyphae with clamps at septa from margin, xl 000 phase contrast. Fig. 9. Thick-walled fibre

hyphae from context, x500 phase contrast. Fig. 10. Thin-walled, distended hypha with dark-

coloured contents from context, x500 phase contrast. Fig. 11. Segments of fractured fibre hypha

from context, mounted in KOH, xl 000, phase contrast. Fig. 12. Radial longitudinal section from

context below upper surface, x400. Fig. 13. Basidia, xl 000 phase contrast. Fig. 14. Basidiospores,

xl 000 phase contrast.

150

POLYPORUS BAUDON1 ON EUCALYPTUS SPP. IN SOUTH AFRICA

Fig. 15-22. — Polyporus baudoni, cultural characters. Fig. 15. & Fig. 16. Cultures of different isolates

at 4 weeks. Fig. 17. Thin-walled hypha with clamped septum from advancing zone xl 000 phase

contrast. Fig. 18. Narrow, thin-walled deeply staining, branching, aerial hyphae showing simple

septum, xl 000. Fig. 19. Unbranched, aseptate, fibre hyphae from aerial mycelium x500, phase

contrast. Fig. 20. Thin-walled, aerial hypha with dark-coloured contents, xl 000. Fig. 21. H-con-

nection between thin-walled, deeply-staining, simple-septate aerial hyphae, xl 000. Fig. 22. Thick-

walled, fibre hypha (left) arising at clamped septum of thin-walled hypha (right) xl 000.

G. C. A. VAN DER WESTHUIZEN

151

Fig. 23-24. — Type specimens. Fig. 23. Phaeolus manihoti

Heim. Fig. 24. Polyporits baudoni Patouillard.

Bothalia 11, 1 & 2: 153-157 (1973)

The effect of the wagon building industry on the Amatola Forests

M. J. WELLS*

INTRODUCTION

Selective timber cropping has made the recognition

of primary plant communities in the Amatola Forests

extremely difficult. No complete record of timber

cropping in these forests is available. To reconstruct

the effects of this cropping a study has been made of

the reasons for and side effects of the extraction of

selected tree species from these forests.

King (1941) describes the intensive exploitation to

which many indigenous forests in the Eastern Cape

have been subjected. He estimates that the mountain

forests situated mainly on the Amatolas and subsidiary

ranges, and aggregating about 26 300 ha (65 000 acres)

have probably yielded upwards of 340 000 m3 (12 mil-

lion cubic ft) of timber during the 90 years prior to

1941. The percentage composition of the cut between

1892 and 1938-39 is given as: 67% yellowwood (pre-

sumably both Podocarpus falcatus and Podocarpus

latifolius), 25% wagon woods (unspecified), 3%

sneezewood ( Ptaeroxylon obliquum), and 5% other

species (unspecified).

Much of the woodcutter’s activities were concen-

trated in the lower, more accessible parts of the forests,

which often contained the best timber. King writes,

“To-day these parts resemble scrub and give one the

erroneous impression that they never contained large

sized timber trees”.

From Forestry Department and sawmill records it

is clear that much of the main timbers cut, Podocarpus

falcatus (common or outeniqua yellowwood) and Po-

docarpus latifolius (real yellowwood) was used for

building and mining timbers, for railway sleepers and

uses other than in wagon-building. But an assessment

of the other timber species selectively removed from

the forests can only be obtained by a study of the

wagon-building industry.

Sim (1906) states that, ever since South Africa has

had a European population, wagon-making has been

one of its largest permanent industries. He then quotes

an article based on census figures for 1891 giving the

chief localities where the wagon -building trade was

carried on as: — Paarl, Worcester, Oudtshoorn, Gra-

hamstown and King William’s Town — the total out-

put for the year being about 3 000 wagons, 4 000 carts

and 450 other vehicles.

The wagon-building industry in the Eastern Cape

is now extinct, with the closure of the last wagon-

building firm, Robert Ballantine and Co., of Keiskam-

mahoek. Except where otherwise indicated, the fol-

lowing facts have emerged from a study of this firm’s

operations.

‘Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X 101, Pretoria.

I am deeply indebted to the owner of the firm, Mr.

Wiltse Ballantine who has patiently answered my ques-

tions, and placed his firm’s books dating from 1886 at

my disposal. Without his interest and co-operation

this survey would not have been possible. My thanks

also go to his staff, especially to his wagon-maker, Mr.

P. L. Fourie, whose 42 years in the trade, a family one,

ensures the reliability of the observations.

‘Ballantines’ started sawing wagon timber from the

Amatola Forests in 1893, and in 1902 commenced

wagon-building using local timbers only. In 1913-14

they started buying wagon wood of indigenous species

from Knysna, and later exotic timbers replaced indi-

genous ones in some parts of the wagons. These and

other developments can only be discussed, however,

when the various wagon parts have been identified

and described (see Figs. 1, 2 and 3). The total output

of the Keiskammahoek industry will then be calculated

and an attempt will be made to link it with local forests

in terms of intake of sound timber and discard of

unsound timber. Side effects of timber extraction in-

cluding damage due to clearings about saw pits and

slip paths will also be referred to in a subsequent

paper.

AN INTRODUCTION TO WAGON PARTS AND TIMBERS

Wagon Parts

To aid in the identification of parts, as many names

as possible have been included in the captions to Figs.

1, 2 and 3. Authority for the use of the names, many

of which combine English and Afrikaans terms, is

indicated as follows: — (B) Ballantine and Co. 1928-

1964, Mr. W. Ballantine and staff 1965; (W) Walton,

1952; (S) Sturt, 1963, and (Sim) Sim, 1906.

There is some confusion in the use of the term 'bed'

as applied to wagons. In this matter Sturt (1963) has

been followed. He describes the term ‘axle-tree’ as

having given way to ‘axle-bed’ when all-wood axles

carved from a single tree, gave way to axles with metal,

wheel-bearing arms which were bedded into wood.

Now axles are all-metal but they are still bedded into

the lowest piece of wood in the undercarriage, which is

thus an ‘axle-bed’ and not an ‘axle-tree’. Nor is the

‘draai-board’, which is not adjacent to the axle, an

‘axle-bed’ (see Fig. 1).

To avoid confusion, the term ‘body’ is preferred to

the term ‘bed’ when referring to that part of the wagon

that receives the load.

Ballantines have produced a variety of vehicles, in-

cluding buck wagons, block trolleys, spring or horse

trolleys, scotch carts and skey carts, and special orders

such as coal carts and manure carts. This paper will

concentrate on the main production line, the buck

154

EFFECT OF WAGON BUILDING INDUSTRY ON THE AMATOLA FORESTS

Q JOl

Fig. 1. The wooden parts of a Ballantine Buck

Wagon with over-irons and a Cape brake

1. Fore or front wheels (B) — see Fig. 3.

2. Hind or back wheels (B) — see Fig. 3.

3. Onderstel (B,W), undercarriage (S).

3 A Front or fore stel (B) fore carriage (S).

3 A 1 Front axle-bed, bed (B,S), ex-bed (S),

front, axle-tree, voorasboom (W).

3 A 2 Draai board (B,Sim), upper axle-bed,

bed (W).

3 A 3 Front schamel, schamel (B), schamel

(W), pillar or pillow ? (S), bolster? (S).

3 A 4 Foretongue (B,W), voor-tang (W).

3 B Wood longwagon, longwagon (B),

houtlangwa (W), pole (S).

3 C Back or hind stel (B), hind carriage

(S).

3 C 1 Back axle-bed, bed (B,S), rear axle-

tree, agterasboom (W).

3 C 2 Back or hind schamel, schamel (B),

skamel (W).

3 C 3 Guide (B), after-tongue, agtertang

(W).

4. Towing pieces

4 A Disselboom (B, W), pole (W,S),

shaft, sharps (S).

4 B Yoke (B,W,Sim).

4 C Yoke skey (B).

5. Wagon — bed (B), Body (S).

5 A Buck-beam (B), side (S), onderleer-

boom (W).

5 B Bed planks (B), wagon-bottom boards

(S).

5 C Krom bar (B), top head (S).

5 D Buck rail, rail (B), outer rave, outrave,

rave (S).

5 E Rail plank (B), rave board (S).

5 F Crossbar (B).

5 G Side Box (B).

5 H Toe Piece (B).

6. Cape brake (B).

6 A Long brake bar, long brake,

brake bar (B).

6 B Short brake bar, short brake, brake

bar (B).

6 C Hard brake block, brake block (B).

6 D Soft brake block, brake block (B).

wagon (Fig. 1). This vehicle in sizes ranging from 3,7

to 5,5 m (12 ft to 18 ft) beam, has been produced to

an almost unaltered design by Ballantines since 1902.

It is almost identical, in the wooden parts of its under-

carriage and body, to the trek wagon with a foreton-

gue, described and illustrated by Walton (1952). Rails,

rail planks, krom bar and brake system are the main

additions found in the Ballantine Buck Wagon, which

also differs from the trek wagon in having a removable

tent (Fig. 2). In the buck wagon the buck beams do not

also serve as the bottom rail of the tent side, as in the

trek wagon illustrated by Walton (1952) and the tent

sides do not slope as in a ‘kakebeenwa’. Sometimes a

‘Natal’ forestel, with futchels instead of a foretongue,

is fitted to a buck wagon.

Fig. 2. The wooden parts of a Cape tent

1 . Tent (B), kaptent framework (W).

1 A Tent bow (B), arch (W).

1 B Reinforcing bow.

1 C Tent rib (B).

2. Tent sides (B).

2 A Framework (B), ladder*, leer (W).

2 A 1 Leermannetjie skey (B,W), oortjie or

little ear skey (W).

2 A 2 Skey (B), leerskei, rung (W), stand,

standard (S).

2 A 3 Top rail (B,W), bo-leerboom (W), top

rave (S).

2 A 4 Bottom rail (B,W), onderleerboom

(W),

2 A 5 Tent plank (B) side board (W,S).

♦Sturt (1963) refers to a ‘ladder' as a

framework for extending the loading capacity

of a cart or wagon at the back or front.

M. J. WELLS

155

The most popular tents manufactured by Ballan-

tines are the ‘Cape’ and the ‘Natal’. The wooden parts

of a Cape tent are shown in Fig. 2. Both Cape and

Natal tents stand higher and wider than the trek wagon

tent, the Cape tent being based on the sloping rail

planks and the Natal tent being based on the rails- the

outermost and uppermost part of the wagon body.

The Cape tent was often supplied with a cartel or

portable bed. Natal tents often had seats fitted over

the rail planks and the back of the tent was boarded

up at the sides, covering the ends of the seats. Half

tents of both Cape and Natal design were made.

Wagon front-boxes which also served as driver’s seats

were supplied with most tented and some untented

wagons. Other accessories included locally made water

barrels of imported oak.

The wooden parts of a wagon back wheel are shown

in Fig. 3. Wagon wheels differ mainly in the size and

number of their parts. Back wheels for a 4,8 m (16 ft)

wagon usually stand 1,3 m (4 ft 4 inches) high and

contain 16 spokes and 8 felloes, whilst front wheels

usually stand 1,0 m (3 ft 4 inches) high and contain

10 spokes with five felloes. Back wheels may contain

14, and front wheels 12, spokes.

All the wooden parts usually incorporated in a

Ballantine Buck Wagon with Cape tent and Cape

brake have been listed and some reference to variation

in size and design have been made. In order to calcu-

late the effect on the forests of building such a wagon

it is necessary to know the volume and kind of timber

used in making the parts.

Because there are many curved, rounded and inter-

locking parts in a wagon the volume of the finished

product is difficult to calculate accurately, and varies

from vehicle to vehicle. The volume of rough timber

from which a wagon is derived also varies, two of the

main variables being the amount of unsound timber

discarded between the sawpit and the wagon shop,

and the sizes of sound timber available. The only rela-

tively stable timber volume which can be calculated

accurately is that of the smallest rectangular pieces of

sound timber from which wagon parts of known di-

mensions can be cut. This volume henceforth referred

to as the minimum volume is shown in Table 1, and

is standard for 4,8 (16 ft) Buck Wagons. The ‘finished’

volumes shown in Table 1 refer to a particular vehicle,

and were calculated by deducting off-cut volumes from

the minimum volume.

Since the down-cutting of sound, rough timber to

minimum volume starts months or years before the

parts are finished and used, the sound, rough timber

volume used in a particular vehicle could not be calcu-

lated. The ‘rough’ volumes in Table 1 average the

examples of down-cutting of sound timber, encounter-

ed at Ballantines and in the Cata forest near Keiskam-

mahoek during 1965.

The percentage wastage or discard of sound timber

during the making of a wagon can be calculated if the

volume of the finished product is compared with that

of the sound, rough timber from which it is derived.

The figure obtained, 58% of the rough timber volume,

may seem unduly great, but several factors should be

taken into account. Firstly, the tall, straight-boled

Fig. 3. The wooden parts of a wagon wheel

1. Nave (B,W), stock, hub (S).

2. Nave wedge (B).

3. Spoke (B,W,S).

4. Spoke wedge (B).

5. Felloe (B.W.S), felling (B).

156

EFFECT OF WAGON BUILDING INDUSTRY ON THE AMATOLA FORESTS

Table 1. The species and volume of sound timber used by R. Ballantine and Co., in building a 4,8 m (16 ft) buck wagon with over-irons.

Cape brake and Cape tent

*The following initials have been assigned to species: Ba — Burkea africana (Wild Syringa); C — Carya sp. (Hickory); Cc —

Calodendrum capense (Cape Chestnut); Cd — Curtisia dentata (Assegai); Ec — Erythrina caffra (Kaffirboom); Ed — Eucalyptus diversicolor

(Karri); Em — Eucalyptus maculata (Australian Spotted Gum); OL — Olea capense subsp. macrocarpa (Black Ironwood); P — Pinus sp.

or spp. (Baltic deal and Oregon Pine); Pf — Podocarpus falcatus (Common or Outeniqua Yellowwood); S+A — Socolopia mundii (Red

Pear) and Apodytes dimidiata (White Pear). Initials of exotics are in italics, and initials of locally cut indigenous species are in brackets

fl Cubic foot =0,028 316 8 m.

forest trees available in South Africa, together with

changes in wagon design favouring the use of more

straight parts have eliminated from the art of wagon

building in South Africa the practice described by

Sturt (1963) of cutting curved wagon parts from simi-

larly curved tree trunks. Secondly, the decline of the

wagon trade has made it uneconomic to keep a large

variety of sizes of wagon timbers in stock.

Wagon parts are often cut from unnecessarily large

timbers, but the larger offcuts from e.g. foretongue

timbers, are sometimes used for making small parts

such as hard brake blocks and wedges.

Thirdly, the wagon maker aims to reduce the weight

of his vehicle as far as possible without reducing its

strength. Sturt (1963) estimates that in making a Sur-

rey farm wagon . . . “the eighth part at least of the

weight of squared wagon timber was taken away with

the draw-shave”. Timber discard of this sort is certain-

ly not as great as one eighth of the weight of a Ballan-

tine wagon, but several cubic feet of timber are

probably discarded in this way, contributing to a total

discard of about 24% of the minimum volume during

finishing.

Wagon Timbers

From the rough volumes shown in Table 1 it can be

calculated that in 1902 when Ballantines used only

locally grown indigenous timber the percentage com-

position of a wagon was likely to have been: 28,7%

Olea eapensis subsp. macrocarpa (black ironwood),

22,9% Podocarpus falcatus (common yellowwood),

10,3% Curtisia dentata (assegai), 32,1% Scolopia

mundii (red pear) and Apodytes dimidiata (white pear),

4,8 % Calodendron capense (Cape chestnut) and 1,2%

Erythrina caffra (kaffirboom).

These figures are subject to variation due to the

inclusion of parts made of other timber species. For

example, locally grown Prunus africana (red stink-

wood) is considered by Ballantines to be the best of

South African wagon woods and when available in the

right sizes was used in making any part of a wagon.

M. J. WELLS

157

Between 1913 and 1914 Ballantines found that the

longer wagon timbers were best made of timber from

Knysna, a fact which was attributed to the Knysna

wood being of longer and straighter grain than that

locally grown. From this date disselbooms, long-

wagons, buck beams and buck rails, representing

9,3% of the total rough volume of wood in a wagon,

were often made from Knysna timber, mainly Olea

capensis subsp. macrocarpa (black ironwood). More

recently, as timber cutting in local forests slowed

down, foretongues, long brake bars and other wagon

parts have also been made of black ironwood from

Knysna. Spokes and felloes of other indigenous woods,

the species not always known, have also been obtained

from Knysna.

Hardwood naves of Burkea africana (wild syringa),

from the Transvaal, were also sometimes used in

preference to naves of locally grown Podocarpus

falcatus (common yellowwood). In the 1920’s a wagon

made entirely of indigenous timbers might have had a

local timber content of as little as 42% of the rough

volume. But, as local felloes were preferred, the local

timber content was usually nearer 74% of the rough

volume.

Up until the 1920’s some wagons were made entirely

of indigenous timbers but thereafter exotic timbers,

especially Eucalyptus maculata (Australian spotted

gum) and Eucalyptus diversicolor (karri) replaced

Knysna timber in the disselboom longwagon, buck

beams and buck rails. After 1924 some bed planks and

rail planks were made of Baltic pine and of Oregon

pine in preferance to Podocarpus falcatus (common

yellowwood) which often warped, resulting in split and

wastage, when sawn into 3,75 cm (1^- inch) thick

planks. Only 2,5 cm (1 inch) thick planks of common

yellowwood are now used — in making the tent side,

and side box. The total contribution of indigenous

timber to the last Ballantine wagons made was usually

about 80% of their rough timber volume, about 66%

being locally grown.

REFERENCES

Walton, J., 1952. Homesteads and Villages of South Africa.

Pretoria: J. L. van Schaik. 71-78.

Ballantine, R. & Co., 1928, 1931, 1942 and 1964. Price lists

of wagon wood.

Sturt, G., 1963. The Wheelwright's Shop. University Press:

Cambridge.

Sim, T. R., 1906. The Forests and Forest Flora of the Colony of

the Cape of Good Hope. Aberdeen: Taylor & Henderson,

58, 60-62, 295.

King, N. L., 1941. The exploitation of the indigenous forests

of South Africa. J. S. Afr. For. Ass. 6: 26.

5292 — J1

■

.

■' • ' i

. :

Bothalia 11, 1 & 2: 159-164 (1973)

On Hierarchical Syndrome Analysis and the Zurich-Montpellier

Table Method

B. J. COETZEE* and M. J. A. WERGER*

ABSTRACT

Hierarchical Syndrome Analysis was applied to two sets of data drawn from Werger, et al. (1972).

The resulting classification was compared to that obtained by Werger et al., who used the Zurich-

Montpellier Table Method. It is concluded that Hierarchical Syndrome Analysis results in an advanced

stage of the process that leads to a meaningful classification of vegetation as expressed in a phyto-

sociological table obtained by Zurich-Montpellier techniques. However, the Hierarchical Syndrome

Analysis was found to be too rigid in some respects to finalize such a classification. The time and effort

involved in preparing the data for computation is another serious consideration. The main conclusion

is that the consistency of the results of the Hierarchical Sydrome Analysis with those obtained by the

Zurich-Montpellier Table Method, proves the objectivity of the latter.

INTRODUCTION

The once seemingly unbridgeable gap in phyto-

sociology between the schools using statistical

techniques and the Zurich-Montpellier School with

its so-called subjective methods, has been reduced

in recent years by various studies that compared

the results of both types of methods. Some of these

studies dealt with a comparison of efficiency of these

types of methods (e.g. Van Groenewoud, 1965;

Ivimey-Cook & Proctor, 1966; Moore, et al., 1970),

whereas in others the application of a method of

one type was initiated as an aid towards the inter-

pretation of results obtained with a method of the

other type (e.g. Van der Maarel, 1969; Coetzee,

1972; Werger, 1972). The results of a wide range

of statistical methods, such as association analysis,

inverse analysis, weighted pair group method of

cluster analysis using different coefficients of simi-

larity, principal component analysis, and others,

were compared in this way with those obtained by

the table method of the Ziirich-Montpellier School.

As stated recently by Goodall (1970), none of these

available statistical methods for clustering samples

of vegetation has yet established itself as consis-

tently preferable, and variety might be advantageous

in the present stage of development of this meth-

odology, since this variety might add new facets to

the description of the acquired clusters and their

interrelationships.

One of the methods of clustering that, as far as

we know, was never used on vegetation data, is

Hierarchical Syndrome Analysis as introduced by

McQuitty (1960). It is a relatively simple polythetic,

agglomerative method, that has proven its merit in

an analysis of climatological data (Welding, pers.

comm.).

In this study Hierarchical Syndrome Analysis

was applied to two sets of data taken from Werger,

et al. (1972), in an attempt to determine the value

of this method in the analysis of vegetation data.

The first set of data comprises eight stands of relatively

simple scrub and forest vegetation, whereas the second

*Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

set comprises 36 stands of floristically rich Fynbos

vegetation, both from Swartboschkloof near Stellen-

bosch in the South-western Cape Province.

These data were synthesized and interpreted by

Werger, et al. (1972) by means of the Zurich-Mont-

pellier Table Method. This facilitated the inter-

pretation of the results of the Hierarchical Syndrome

Analysis and allowed a comparison thereof with

the results of the Zurich-Montpellier Table Method.

METHODS

The method is used to group individuals (en-

tities) into classes of successively lower levels of

similarity, or into a “hierarchical system of syn-

dromes”, so that on the highest level every entity

in a class is more like every other entity in that class

than it is like any entity in any other class (McQuitty,

1960). At each following stage of synthesis an es-

tablished class is regarded as one synthetic individual

(entity). The similarity between a class and an indi-

vidual, or between two classes, is taken as the lowest

correlation between a member of that class and the

single individual or pair of individuals, one from

each class. This is because it is assumed that the

members of a class have as much in common as

the pair of members with the lowest similarity.

However, this assumption need not be absolutely

valid; the only requirement is that valid classes

have higher estimates for overall similarity than

competing ones (McQuitty, 1960). McQuitty actually

found that the assumption invariably over-estimated

the scores for overall similarity, but nevertheless

resulted in satisfactory classifications as compared

with those obtained by more elaborate methods.

Of the two versions described by McQuitty, the one

adopted here is the “replacement version”.

In this study correlations were calculated with

the product-moment correlation coefficient, reported

to give the best results in a cluster analysis carried

out by Moore, et al. (1970).

For each set of data two analyses were carried

out. one to group the stands and one to group the

species.

160

HIERARCHICAL SYNDROME ANALYSIS AND THE ZORICH-MONTPELLIER TABLE METHOD

For purpose of calculation cover-abundance figures

were transformed as follows:

r + 1 2 3 4 5

1 5 10 20 30 40 50

This was adopted after a number of trials which

included the transformation of cover-abundance

values proposed by Moore, et al. (1970) and which

was felt to underestimate especially the lower values,

+ and 1, of the scale.

The analysis was carried out using all eight stands

and all 74 species, i.e. all species occurring in Table

2 and Appendix 2 of Werger, et al. (1972), in the

case of the first set of data, whereas all 36 stands

and only the 147 species occurring in Table 1 of

Werger, et al. (1972) were used for the second set

of data. The Ziirich-Montpellier Table Method was

already discussed by Werger, et al. (1972).

RESULTS

The results of the analysis of the eight stands of

scrub and forest vegetation are represented in Fig.

la and lb. Fig. la shows three main legs of the

dendrogram, corresponding to the three communities

recognized by Werger, et al. (1972: Table 2). It

confirms that the two forest communities, the Rapanea

melanophloeos Community and the Heeria argentea

Community, are more similar to each other than

any of them to the Brabeium stellatifolium Riverine-

scrub Community.

In the classification of species of the first set of

data, as shown by Fig. lb, one main leg includes all

species characteristic of the Brabeium stellatifolium

Community, as well as the Fynbos species intruding

into this community. This leg also includes a number

of species occurring only once and in a stand of the

Brabeium stellatifolium Community. Two species of

general occurrence over this set of data, Restio

subverticillatus and Chironia baccifera , and three

species common to this and the Rapanea melano-

phloeos Community, Cunonia capensis, Brachylaena

neriifolia and Oplismenus hirtellus, all reaching re-

latively high cover-abundance values in the Bra-

beium stellatifolium Community, are also included

in this leg of the dendrogram.

A second main leg includes all species charac-

teristic of the Rapanea melanophloeos Community

and three species common to this community and

the Heeria argentea Community, Olea africana ,

Maytenus acuminatus and Halleria lucida. that reach

Fig. 1. — Classification of scrub and forest vegetation. Explanation in text.

a. Classification of stands. A Heeria argentea Community; ■ Rapanea melanophloeos Community; % Brabeium

stellatifolium Community.

b Classification of species. % Character species of Brabeium stellatifolium Community; O species intruding in

Brabeium stellatifolium Community; -|- species common to Brabeium stellatifolium, Rapanea melanophloeos

and Heeria argentea Communities; V species common to Brabeium stellatifolium and Rapanea melanophloeos

Communities; ■ character species of Rapanea melanophloeos Community; /\ species common to Rapanea

melanophloeos and Heeria argentea Communities; A character species of Heeria argentea Community. Species

occurring only once and in the same stand are correlated at level + 1 and represented in the dendrogram

by one ending. No symbol is assigned to these and other species occurring only once.

B. J. COETZEE AND M. J. A. WERGER

161

relatively high cover-abundance values in the Rapanea

melanophloeos Community. This leg further includes

two species common to the Rapanea melanophloeos

and the Brabeium stellatifolium Communities, As-

paragus scandens and Ilex mitis, three species of

general occurrence, Blechnum punctulatum , Myrsine

africana and Knowltonia capensis, and three species

of single occurrence.

The third main leg of the dendrogram includes

all species characteristic of the Heeria argentea Com-

munity, as well as the three remaining species com-

mon to this and the Rapanea melanophloeos Com-

munity. Three species of general occurrence, Maytenus

oleoides , Secamone alpini and Hartogia schinoides,

of which the latter two constantly reach higher

cover-abundance values within this community, are

also grouped in this main leg, together with four

species of single occurrence.

Fig. 2a shows the grouping of the 36 stands of

complex Fynbos vegetation into four main “syn-

dromes”, which correspond closely to the com-

munities recognized by Werger et al. (1972: Table 1).

One main leg comprises all four stands of the Ber-

zelia lanuginosa — Osmitopsis asteriscoides Community.

A second main leg combines all stands of the

Protea arborea — Rhus angustifolia community with

the two typical stands of the Restio perplexus com-

munity. The affinity between these two communities

was already emphasized by Werger, et ai (1972).

The originally unclassified stand 22, as well as stand

7 of the Brunia nodiflora — Psoralea rotundifolia

Community and stand 21 of the Thamnochortus

gracilis — Hypodiscus aristatis Community are also

grouped in this main leg. This classification of stands 7

and 21 may be due to some exceptional cover-abun-

dance values scored by species in these stands, for

example by Protea neriifolia, which is otherwise of

particular importance in the Protea arborea — Rhus

angustifolia Community.

A third main leg is composed of stands of the

Thamnochortus gracilis — Hypodiscus aristatus Com-

munity. Stand 34, which was originally described

as transitional between this and the Restio per-

plexus Community, is also included in this group.

Stands 27 and 29 of this Community, that were

originally pointed out as not fully representative,

are not included in this main leg, but in the fourth

one, that combines the stands of the Brunia nodiflora —

Psoralea rotundifolia communities, except for stand

7 as pointed out already. Regarding the species

classification of the Fynbos stands (Fig. 2b), it

should be noted that species labelled by Werger, et

al. (1972) as accompanying species, are generally

those that occur too infrequently to be included into

another group of the table. In the Hierarchical

Syndrome Analysis these species, were necessarily

classified on the basis of their distribution and cover-

abundance scores, despite their infrequent occur-

rences. Thus their occurrence in any group will be

regarded as insignificant, and will not be further

commented upon.

-I

Fig. 2a. — Classification of stands of

Fynbos vegetation. Explanation in

text.

A Berzelia lanuginosa-Osmitopsis as-

teriscoides Community; O Restio

perplexus Community; % Protea

arborea-Rhus angustifolia Com-

munity; ■ Thamnochortus gracilis-

Hypodiscus aristatus Community ;

+ Brunia nodiflora-Psoralea rotun-

difolia Community.

iitiOOMIM •••••• ■■■■■•■ O •♦♦♦♦♦♦•

162

HIERARCHICAL SYNDROME ANALYSIS AND THE ZORICH-MONTPELLIER TABLE METHOD

B. J. COETZEE AND M. J. A. WERGER

163

One leg of the dendrogram consists of the character

species Diospyros glabra , Rhus angustifolia, Cassvtha

ciliolata and Pteridium aquilinum of the Protea

arborea — Rhus angustifolia Community, that also

intrude into the Berzelia lanuginosa- — Osmitopsis

asteriscoides Community.

The remaining character species of this community,

which only score low cover-abundance values, are

spread over the dendrogram.

Two main legs of the dendrogram comprise the

character species of the Berzelia lanuginosa — Os-

mitopsis asteriscoides Community. The separation of

these species over two legs may be due to an uneven-

ness of cover-abundance scorings of these species

in the stands of this community.

Another main leg is composed of most of the

character species of the Thamnochortus gracilis —

Hypodiscus aristatus Community plus four species

of less restricted occurrence, Aristea capitata, Pen-

taschistis steudelii, Metalasia muricata and Erica

hispidula of which the latter two tend to have higher

frequencies and/or higher cover-abundance values

within this community.

Two of the character species of this community,

Blaeria dumosa and Pentameris macrocalycina, which

both reach relatively high cover-abundance values

in stand 14, appear together in a separate leg. Two

other character species, Restio sieberi and Coleonema

juniperinum are also classified away from the main

group of character species, and form a leg together

with Pentaschistus colorata and Clijfortia ruscifolia.

These four species score exceptionally high cover-

abundance values in stand 27, which was also sepa-

rated from the other stands of the Thamnochortus

gracilis — Hypodiscus aristatus Community in the

classification of stands (Fig. 2a).

Another major branch of dendrogram comprises

the character species of the Brunia nodiflora — Psoralea

rotundifolia Community, with the exception of Os-

teospermum tomentosum and Elytropappus glandu-

losus, that are grouped together with a number of

species of more general occurrence.

Restio filiformis and Pentaschistus curvifolia, which

respectively score high cover-abundance values and

are more frequent in the Thamnochortus gracilis-

Hypodiscus aristatus Community, are grouped together

with the character species of this community.

Yet another main leg combines the character

species of the Restio perplexus Community with

Watsonia pyramidata, Themeda triandra and Psoralea

obliqua, which are common to this and the Protea

arborea — Rhus angustifolia Community. Stoebe

plumosa, a species of general occurrence which scored

exceptionally high cover-abundance values in the

two typical stands of the Restio perplexus Community,

is also classified in this group.

The species common to two or more communities

are more scattered over the dendrogram, although

there are marked clusters of species common to the

Protea arborea — Rhus angustifolia and the Brunia

nodiflora — Psoralea rotundifolia Communities, and

of species common to these two and the Thamno-

chortus gracilis — Hypodiscus aristatus Communities.

DISCUSSION

The combined outcome of the Hierarchical Syn-

drome Analysis of stands and species of a set of

data lead to a result that is largely consistent with

a meaningfully interpreted classification of vegetation.

Possibly still better results might be obtained by

slightly changing the general transformation of

cover-abundance values and introducing a scale of

degree of absence similar to the one proposed by

Swan (1970). This means that Hierarchical Syndrome

Analysis can be a useful tool in the construction

of an advanced stage of a meaningful classification

of vegetation as is expressed in a phytosociological

table obtained by Zurich-Montpellier techniques.

However, in estimating the value of Hierarchical

Syndrome Analysis two things should be borne in

mind. Firstly, the result of a Hierarchical Syndrome

Analysis cannot be regarded as a completed analysis

of vegetation data, because the procedure rigidly

classifies stands and species. Some subtle decisions

based on more complex criteria than correlation

coefficients only have to be taken in order to arrive

at a correct, meaningful placing of certain stands

and species. For example, particular values for

a certain attribute should not always be given the

same weight, but the general pattern of occurrence

of a species in the data should in some cases be

permitted to affect the weighting of cover-abundance

values. Thus, the weight of the score of 2 of Restio

subverticillatus in stand 31 must have been the cause

for its placing in the group of character species

of the Brabeium stellatifolium Community (Fig. lb),

whereas its general distribution pattern (Werger,

et al. 1972: Table 2) shows that it is not at all typical

for that community. Secondly, the time and effort

involved in preparing the data for computation

should not be underestimated.

Perhaps the most important conclusion drawn

from this study is that the similarity of the results

of the Hierarchical Syndrome Anylysis and the

Zurich-Montpellier Table Method contributed to

prove the often doubted objectivity of the latter.

In this sense the present study showed that the

“amount of art” involved in constructing a Zurich-

Montpellier phytosociological table is largely “un-

reconstructed logic-in-use” and thus it helped to

clarify the theoretical significance of Zurich-Mont-

pellier phytosociological procedures (Hull, 1970).

Already Ellenberg (1956) had stressed the objectivity

of the table method and the success of Spatz &

Siegmund (1972), who used the same data as Ellen-

berg, in arriving at a virtually identical result by

computer processing, further illustrated this point.

The often heard arguments that the subjective

way of sampling or the assignment of estimated

importance values to species in the Zurich-Mont-

pellier School already introduces a subjective element,

that makes it impossible to obtain an objective result,

need not be important, since meaningful interpret-

able phytosociological tables can also be constructed

from systematically distributed samples (e.g. Taylor,

1969) and where measured importance values are

assigned to species.

ACKNOWLEDGEMENT

We are indebted to Mr M. C. Welding for his

kind assistance with the computations.

REFERENCES

Coetzee, B. J., 1972. 'n Plantsosiologiese studie van die Jack

Scott-Natuurreservaat. Unpubl. M.Sc. Thesis, Univ. Pretoria.

Ellenberg, H„ 1956. Aufgaben und Methoden der Vege-

tationskunde. in H. Walter: Einfiihrung in die Phytologie.

Bd. 4. T. 1. Stuttgart: Ulmer.

Goodall, D. W., 1970. Statistical plant ecology. Ann. Rev.

Ecol. Syst. 1: 99-124,

164

HIERARCHICAL SYNDROME ANALYSIS AND THE Z0RICH-MONTPELLIER TABLE METHOD

Hull, D. L., 1970. Contemporary systematic philosophies.

Ann. Rev. Ecol. Syst. 1 : 19-54.

Ivimey-Cook, R. B. & Proctor, M. C. F., 1966. The application

of association-analysis to phytosociology. J. Ecol. 54:

179-192.

McQuitty, L. L., 1960. Hierarchical Syndrome Analysis.

Ed uc. Psychol. Meas. 20: 293-304.

Moore, J. J., Fitzsimons, P., Lambe, E. & White, J., 1970.

A comparison and evaluation of some phytosociological

techniques. Vegetatio 20: 1-20.

Spatz, G. & Siegmund, J., 1972. Eine Methode zur tabel-

larischen Ordination, Klassifikation und okologischen

Auswertung von pflanzensoziologischen Bestandsaufnah-

men durch den Computer. Vegetatio, in the press.

Swan, J. M. A., 1970. An examination of some ordination

problems by use of simulated vegetation data. Ecology

51: 89-102.

Taylor, H. C., 1969. A vegetation survey of the Cape of Good

Hope Nature Reserve. Unpubl. M.Sc. thesis, Univ. Cape

Town.

Van der Maarel, E., 1969. On the use of ordination models

in phytosociology. Vegetatio 19: 21-46.

Van Groenewoud, H., 1965. Ordination and classification

of Swiss and Canadian coniferous forests by various

biometric and other methods. Ber. Geobot. Inst. ETH,

Stiftung Riibel 36: 2 -102.

Werger, M. J. A., 1972. On the use of association-analysis

and principal component analysis in interpreting a Braun-

Blanquet phytosociological table of a Dutch grassland.

Vegetatio: in the press.

Werger, M. J. A., Kruger, F. J. & Taylor, H. C., 1972. A

phytosociological study of the Cape Fynbos and other

vegetation at Jonkershoek, Stellenbosch. Bothalia 10:

599-614, also as: Pflanzensoziologische Studie der Fynbos

Vegetation am Kap der Guten Hoffnung. Vegetatio 24:

71-89.

Bothalia 11, 1 & 2: 165-176 (1973)

An account of the plant communities of Tussen die Riviere Game

Farm, Orange Free State

M. J. A. WERGER*

ABSTRACT

As part of the IBP survey of conservation sites, the vegetation of Tussen die Riviere Game Farm,

Orange Free State, was surveyed and analysed according to the Braun-Blanquet phytosociological

method. A classification of the plant communities occurring there is given. A way by which a hierarchical

classification of plant communities in South Africa could be constructed, is suggested. Each plant

community is physiognomically classified according to Fosberg's (1967) system. Two disadvantages

of this system are discussed briefly.

INTRODUCTION

The continental phytosociological method de-

scribed by Ellenberg( 1956) and Braun-Blanquet (1964)

has proved effective for the analysis of a wide variety

of vegetation types in different parts of the world.

For various reasons this method was not used in

southern Africa until recently, when it was also

shown to be a quick and useful method for analysing

the floristically rich South African and South West

African vegetation (Van Zinderen Bakker, 1971;

Volk & Leippert, 1971; Werger et al., 1972).

Using this technique, Tussen die Riviere Game

Farm was surveyed on a semi-detailed scale, as part

of the survey of conservation areas for the Interna-

tional Biological Programme. This could be a first

step towards a comprehensive classification of South

African plant communities. If such surveys were

undertaken in all conservation areas within South

Africa, this would result in classifications of the plant

communities at points distributed over a variety of

South African Veld Types (Acocks, 1953). Classifi-

cations of plant communities in areas that are floris-

tically not too different from one another, would

provide insight into the variation of these com-

munities. By comparing and combining vegetation

classifications of such floristically related areas, an

integrated, consistent and generally applicable clas-

sification could be worked out. Each of these would

represent key areas from where interpolation and

survey of the intermediate areas can be undertaken.

In such a way a comprehensive hierarchical classi-

fication of plant communities could rapidly be

developed.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Fig. 1.— Map of Tussen die Riviere Game Farm, showing the position of the releves.

166

PLANT COMMUNITIES OF TUSSEN DIE RIVIERE GAME FARM

THE STUDY AREA

Tussen die Riviere Game Farm is situated in the

southern Orange Free State at about 30° 30' S and

26° 15' E, on the peninsula at the confluence of the

Orange and Caledon Rivers (Fig. 1.).

The Game Farm was created in 1967, when the

area was withdrawn from farming in anticipation

of the inundation along the Caledon and Orange

Rivers by the filling of the Hendrik Verwoerd Dam

Lake.

Tussen die Riviere Game Farm is approximately

22 000 ha in extent, falling entirely in the False

Upper Karoo Veld Type (Acocks, 1953). Because of

its situation near the eastern border of this Veld

Type, a number of species from the wetter eastern

parts of the country occur on protected sites. According

to Acocks (1953), the plains scarcely differ from those

of the Central Upper Karoo, whereas the hills, that

are not severely overgrazed, are essentially grassy.

The principal shrub on these False Upper Karoo hills,

Rhus erosa, does not occur in the Central Upper

Karoo.

Since the withdrawal of the area covered by the

Game Farm from sheep and cattle farming, there

has been a marked recovery of the veld from its

former poor condition, which was typical of the

overgrazed and poorly managed veld of the whole

region.

More than a century ago, when the European

settlers still crossed the Orange River from the south

only temporarily, Andrew Smith travelled through

the area north of the Orange River and reported the

good condition of the veld there. Travelling in the

vicinity of the present-day Bethulie, he wrote in

his diary (Kirby, 1939): “. . . . on flats, but especially

on the hills an abundance of dry grass” (Vol. 1,

p. 81) and “The flats have scarcely a bush upon them,

and the grass grows in tufts closely set together.

There are two sorts of grass occurring, both sour

and sweet grass. The latter is the most abundant.

The grass on the hills is also said to be sour. Amongst

the rocks on the sides and tops of the hills a few

stunted trees and bushes.” (Vol. 1., p. 89). Soon

after the area was permanently settled, however, the

grassland started to change into karroid dwarf shrub

veld. Acocks (1953) described the conversion of this

former grassland area into eroded False Upper

Karoo vegetation, following the introduction of the

merino sheep as “a national disaster”.

According to the Koppen system, the climate of

this area can be defined as BSkw’: arid (steppe)

climate, cold and dry, with a mean annual temperature

lower than 18°C, although the mean annual tempera-

ture of the hottest month exceeds 18° C. The winter

is dry, and the rainy season is from November to

April with a maximum in March (Schulze, 1947;

Weather Bureau, 1954). During summer the climate

of the region is under the influence of moist equatorial

and maritime tropical air from which the precipi-

tation falls largely in thunderstorms, so that much

of it is lost to agriculture because of the rapid runoff.

In winter anticyclonic systems over the Orange Free

State bring fine, dry weather, with almost cloudless

skies and light winds (Kendrew, 1961). Since humidity

and temperature are of main importance to plant

life, the bioclimatic map of UNESCO-FAO (1963)

is based on ombrothermic diagrams (where precipi-

tation in mm and temperature in °C on twice the

precipitation scale are plotted against month of the

year) and the xerothermic index (hot weather drought

index). This map classifies the climate of the Tussen

die Riviere Game Farm area as intermediate temperate

tropical, which means that the dry season coincides

with the period of shortest daylight, the mean tempera-

ture of the coldest month is between 0° and 10° C,

and the xerothermic index is between 100 and 150.

The Weather Bureau (1954) published the statistics

shown in Table 1 for the weather stations at Bethulie

and Aliwal North, between which Tussen die Riviere

is located. A climate diagram for Bethulie is given

by Walter & Lieth (1960).

The mean number of days with frost at Aliwal

North is 65,4 (Weather Bureau, 1954).

General information on the geology of the area

from Du Toit (1954) and Truter & Rossouw (1955)

shows that the entire area consists of sandstones,

mudstones and shales of the Upper and Middle

Beaufort Series (Upper Permian and Lower Triassic)

of the Karoo System, interrupted by dolerite sills

and dykes. Recently the geology has been mapped

in detail by Harmse (1971).

Geomorphologically, Tussen die Riviere Game

Farm falls in the Highveld physiographic region

(Wellington, 1955; King, 1967), in which the almost

horizontal beds of the Karoo System are the main

landscape-determining features. Although the tabular

surface in the landscape often gives the appearance

of peneplanation, it is largely a structural phenomenon

TABLE 1. — Temperature, humidity and precipitation stasistics of Bethulie and Aliwal North

At Bethulie temperature and precipitation observations were taken over a period of 20 years, at Aliwal North over 18 years. Humi-

dity observations were taken at Bethulie over 13 years, at Aliwal North over 7 years.

M. J. A. WERGER

167

(Wellington, 1955). The relief in the landscape is

caused by dolerite dykes and sills, the latter often

capping the mesas and koppies. The slopes of softer

sandstone and mudstone underlying the dolerite are

usually steep.

Apart from the abundant lithosols in Tussen die

Riviere Game Farm, and some alluvium along the

rivers, the prevailing soil type is solonetzic (Van der

Merwe, 1962; Loxton, 1962; D’Hoore, 1963). These

solonetzic soils possess a fairly dense and structureless,

0,15 to 0,25 m thick A horizon of grey loam, that

breaks up into clods, lumps and finer material on

slight pressure. The pH of this horizon is slightly

acid. Separated from the A horizon by an abrupt

transition is the B horizon of dark blackish-brown

clay, which on drying, forms columns and is rather

impervious to water. Downwards the colour of this

layer gradually changes to a lighter brown. The pH

of this B horizon is alkaline. Soils from the dolerite

are colluvial. The A horizon consists of reddish-brown

structureless sandy loam and the B horizon of com-

pact, dark, reddish-brown columnar structured clay

(Van der Merwe, 1962).

METHODS

During 1970 and 1971, 64 sample plots were layed

out in the area (Fig. 1) and phytosociological tables

compiled according to the procedures described by

Braun-Blanquet (1964), Ellenberg (1956) and Knapp

(1971). Value 2 of the cover-abundance scale was

slightly modified following Barkman et al. (1964),

so that 2m means “very abundant, but covering less

than 5% of the plot area”, 2a “cover 5-12% indepen-

dent of number of individuals”, and 2b “cover

13-25% independent of number of individuals”. Plot

size was determined according to the method

described by Werger (1972), and fixed at the following

sizes: for riverine communities, 25 m2; for dwarf

shrub Karoo veld, 50 m2; and for shrubby hillside

vegetation, 100 m2. Only permanently recognizable

species were recorded, thus omitting most annuals

and geophytes. Two doubtful species. Cotyledon

decussata and C. orbiculata were not separately

distinguished, but treated as one complex taxon.

Site factors such as aspect, slope angle, soil depth

and pH of top soil were measured, and geological,

geomorphological and other environmental features

noted. The communities distinguished were classified

into physiognomic formations according to Fosberg’s

(1967) system, which is recommended for IBP surveys.

THE PLANT COMMUNITIES

The vegetation of Tussen die Riviere Game Farm

can be divided into three major groups of communities :

(1) The riverine communities;

(2) The communities of the flats and gently sloping

terrain; and

(3) The communities of the steep slopes.

(1) Riverine communities (Table 2).

The woody riverine vegetation on the levees of

fine alluvial sand with a pH between 6,5 and 7,5

mainly belongs to one community: the Acacia

karroo-Celtis africana Community (Fig. 3). It is

usually three-layered consisting of a tree layer from

6 to 10 m with an average cover of 75 %; an open

2 to 4 m tall shrub layer with an average cover of

40-50%; and a field layer up to about 0,60 m,

consisting of grasses, herbs and small shrubs, with

a greatly varied cover depending on the amount of

grazing.

Physiognomically, this community falls in Fosberg’s

(1967) class 1 A2 (5), deciduous thorn forest formation.

Differential species in the tree layer are Acacia

karroo and Celtis africana; in the shrub layer Rhus

pyroides , Diospyros lycioides subsp. lycioides , Lycium

hirsutum , Clematis brachiata and Melianthus comosus;

and in the field layer Atriplex semibaccata , Asparagus

TABLE 2. — Acacia karroo-Celtis africana Community

Species only occurring once in releves of Table 2: Eragrostis lehmanniana Nees (114: +), Enneapogon brachystachyus (Jaub.

et Spach) Stapf (101 : 1), Conyza podocephala DC. (101 : +), Eragrostis curvula (Schrad.) Nees (105: r), Chrysocoma tenuifolia Berg.

(105 : r), Pentarrhinum insipidum E. Mey. (141 : +), Phytolacca heptandra Retz. (118: +), Olea africana Mill. (112: i ), Pentzia globosa

Less. (112: r), Silene capensis Otth. (106: +).

168

PLANT COMMUNITIES OF TUSSEN DIE RIVIERE GAME FARM

suaveolens , A. setaceus , Bromus willdenowii, Chenopo-

dium mura/e, Melica decumbens , Achyranthes aspera

and Rabia cordifolia. This Acacia karroo-Celtis

africana Community is typical for the levees of this

section of the Orange River. A few kilometres up-

stream Acacia karroo disappears completely, whereas

some 50 km downstream Celtis africana does not

occur any more in the riverine vegetation. Detailed

documentation will be given in a forthcoming paper.

Releves 114 and 101 (Table 2) differ slightly from

others in that Rhus pyroides, A triplex semibaccata,

Lycium arenicolum and Asparagus setaceus have

higher cover-abundance values, whereas Melica decum-

bens, Achyranthes aspera , Rubia cordifolia and Acacia

karroo do not occur in these two plots. This is

because the releves 1 14 and 101 were taken in narrow

strips of riverine vegetation, where most of the

Acacia karroo-Celtis africana Community was pre-

viously cleared for cultivation leaving only a narrow,

rather disturbed strip on the bank of the Orange

River. Releves 114 and 101 can thus be regarded

as the rudimentary fragments of the community.

On the outer edge of the levee one sometimes

finds communities that are completely dominated

by Lycium arenicolum or by Salso/a rabieana, ac-

companied by pioneer species like Eragrostis lehman-

niana , Chrysocoma tenuifolia, Aster muricatus , Aspara-

gus suaveolens, Walafrida saxatilis. Tragus koelerioides,

seedlings of Acacia karroo, etc. Usually the areas

where these communities occur show traces of

former cultivation. The Lvc/wm-dominated community

falls into formation 1 B2 (4), deciduous thorn scrub;

and the Sa/so/a-dominated community into formation

1 B 1 (8c), Grey microphyllous evergreen scrub (Fosberg,

1967). These communities have not been sampled

on Tussen die Riviere Game Farm.

Where the levee drops away into the river channel

one often finds a community mixture of the Acacia

karroo-Celtis africana Community with Salix capensis,

Nicotiana glauca, Agrostis lachnantha and even

Salix babylonica. On dry sand banks in the river

channel a temporary weed community occurs, with

Agrostis lachnantha, Lactuca serriola, Erigeron flori-

bundus, Argemone subfusiformis. Polygonum lapathi-

folium, Xanthium spinosum, Chenopodium ambrosio-

ides, etc. No releves were made in these unstable

communities.

(2) Communities of the flats and gently sloping terrain

(Table 3)

All these communities are usually two-layered.

The grass and dwarf shrub layer from 0, 15 to 0,60 m

high is the most important, with a cover up to 50%,

but usually less. A very open layer of creeping and

rosette plants, up to 0,05 m high, with a cover less

than 5%, is common.

The communities of the flats are floristically poorer

than those of the sloping terrain. Physiognomically

these communities are classified under Fosberg’s

(1967) formation 2C1 (3), evergreen narrow sclero-

phyll dwarf steppe scrub. The Eragrostis lehmanniana-

Chrysocoma tenuifolia Community shows a tendency

towards 2F1 (2), evergreen narrow sclerophyll dwarf

shrub steppe savanna formation.

2 (a) Eragrostis lehmanniana-Chrysocoma tenuifolia

Community

This community differs from all the others in that

it is unstable, occurring on flat to slightly sloping old

lands with rather deep soils (0,50 m to 1 m). The top

soil is usually a sandy loam with a pH between 5,5

and 7,0. The community has no differential species,

but is typified by the high cover-abundance values

reached here by Eragrostis lehmanniana and Chryso-

coma tenuifolia. Pentzia globosa sometimes also plays

an important role. Due to the importance of Eragrostis

lehmanniana , this community has a higher cover,

about 60%, than other communities on the flats.

The community is floristically very poor in that

virtually only the earliest Karoo pioneer species

occur: apart from the above mentioned ones, Tragus

koelerioides, Walafrida saxatilis, Aristida congesta,

Indigofera alternans, Gnidia polycephala and Pterothrix

spinescens. Eventually this community will possibly

develop into one of the others, probably into the

Chrysocoma tenuifolia-Lessertia pauciflora Commu-

nity.

2 (b) Chrysocoma tenuifolia-Lessertia pauciflora

Community

This is one of the most common communities in

Tussen die Riviere Game Farm and neighbouring

Karoo. It is found on the flats with soils over 1 m

deep. The top soil is a fine and middle grained sandy

loam, that has a slightly acid pH from 5,5 to 7,0.

Sometimes one finds a few coarse gravel stones in

the top soil. Ant heaps occur frequently in this

community (Fig. 2).

Fig. 2. — Chrysocoma tenuifo-

lia-Lessertia pauciflora Com-

munity. Note the frequent

ant heaps and the domi-

nance of Chrysocoma tenui-

folia. In the right hand

background a slope with

mainly Rhus erosa-Stachys

burchelliana Community.

Table 3. — Communities of the flats and gently sloping terrain.

169-170

103

6

NE

3

Releve no

Number of species

Aspect

Slope (°) I (i

Geology (s= sand- & mudstone, d^dolerite) “

Soil depth

pH (top soil)

Total cover (%)

7,0

55

110

9

s

>1

6,0

70

113

15

E

1

>1

5,5

75

534

15

E

4

s

0,7

6,5

65

117

11

NE

4

>2

7,0

45

Differential species of the Chrysocoma tenuifolia-Lessertia pauciflora Community

Osteospermum scariosum DC

Lessertia pauciflora Harv

Cyperus usitatus Burch

Aptosimum depressum Burch

Salsola glabrescens Burtt Davy

Nestlera conferta DC

Species common to Chrysocoma tenuifolia-Nenax microphylla and Chrysocoma tenuifolia-Polygla leptophylla

Communities

Nenax microphylla (Sond.) Salter —

Aristida diffusa Trin

Hibiscus marlothianus K. Schum

Trichodiadema pomeridianum L. Bol.

Dianthus basuticus Burtt Davy

Euphorbia clavarioides Boiss

+

Differential species of the Chrysocoma tenuifolia-Polygala leptophylla Community

Convolvulus boedeckerianus Peter

Polygala leptophylla Burch

Schizoglossum capense (Schltr.) Huber

Helichrysum niveum (L.) Less •

Osteospermum leptolobum (Harv.) T. Norl

Thesium spartioides A. W. Hill

Nananthus sp

Eragrostis denudata Hack, ex Schinz

Species of the slopes

Gazania krebsiana Less

Lotononis laxa Eckl. & Zeyh

Heteropogon contortus Beauv

Cymbopogon plurinodus (Stapf) Stapf

Rhus ciliata Licht

Lightfootia albens Spreng. ex A. DC

Sutera halimifolia (Benth.) Kuntze

Specie, con,,, on to Chrysocoma (enuifolia-Lessertia pauciflora Community one, the communities of the Copes

Limeum aethiopicum Burm

Eragrostis curvula (Schrad.) Nees

Lycium salinicolum Verdoorn

Themeda triandra Forsk

Helichrysum dregeanum Sond & Harv.. .

Melolobium microphyllum Eckl. & Zeyh

+

Common Karoo species

Chrysocoma tenuifolia Berg

Pentzia globosa Less

Eragrostis lehmanniana Nees

Tragus koelerioides Aschers

Walafrida saxatilis Rolfe

Aristida congesta Roem. & Schult —

Eragrostis obtusa Munro

Indigofera alternans DC

Gnidia polycephala (C.A. Mey.) Gilg.

Pterothrix spinescens DC

2a

3

2a

2b

3

1

+ 2a 2a

+ r 2b

1 1 +

+

1 +

2a +

+ +

+

2a

1

2b

1

+

2a

2b

+

Other

accompanying species

Hermannia coccocarpa O. Ktze.. . . .

Sutera atropurpurea (Benth.) Hiern.

Solanum supinum Dun

Aster muricatus Lesi

Sporobolus fimbriatus Nees

Oropetium capense Stapf

Geigeria filifolia Mah.f.

Mariscus capensis Schrad

Enneapogon scoparius Stapf

Eriocephalus spinescens Burch

Asparagus suaveolens Burch

Hibiscus pusillus Thunb

Cynodon hirsutus Stent

Polygala uncinata E. Mey

Hermannia linearifolia Harv

Tragus berteronianus Schult

2a

2b

2a

+

2m

* rrh„nh Stanf (549 • +), Caralluma lutea

* villosa (Nees) C B. ££3*+.

Infrequent species in releves of Table 2: (549. + ), Aristida curvata (Nees) Trin & Rupr. (553. + ), , Atnp oma macrocephala DC. (53 . h HelichrySUm +> llu: -r, w--- “"mtAnra Verdoot

Anacampseros lanigera Burch. ,553: +). Argyrolobium lanceola.um ML * ; (^5; r) , Crasmla rransvaalensrs I >. S.«mi -lb —

£& SS- aaub. & Spac'h) Stapf (542: C.A.' & oS: 4-1 'kohautia amatymbrea ^ >Rhus erosa Thunb.j^R^P

lycioides Desf. (553

Jacq. var. glabresce

Nestlera humilis Less. (546: 1

ofrieana Mill. (9^1- Malg

Jacq. var. glabrescens (Harv.) Verdoorn (109: +, 549. +)

U .,113 . J \ 0^1,

M. J. A. WERGER

171

Fig. 3. — Showing sampling of

Chrysocoma tenuifolia-Poly-

gala leptophylla Community

in foreground. On the slope

in the centre left Rhus erosa-

Stachys burchelliana Com-

munity and in the centre

right Acacia karroo-Celtis

africana Riverine Commu-

nity.

The community has a number of good differential

species, which are not, however, abundant: Osteo-

spermum scariosum, Lessertia pauciflora , Cyperus

usitatus, Aptosimum depressum, Salsola glabrescens

and Nestlera conferta. Apart from the common

Karoo species, this community has also a number

of species in common with the communities of the

gently sloping terrain, although they are less constant

and abundant here: Limeum aethiopicum , Eragrostis

curvula, Lycium salinicolum, Themeda triandra , Heli-

chrysum dregeamim and Melolobium microphyllum.

Releves 537 and 546 represent heavily overgrazed

stands in this community.

2 (c) Chrysocoma tenuifolia-Nenax microphylla

Community

This community is found on the nearly flat terrain

and slopes up to 6° with shallow soils of 0,02 to

0,30 m, or with lithosols and outcrops of mudstone

and fine and medium-grained sandstone of the

Beaufort series. The pH of the soil is usually from

5.5 to 6,5, except in releve 140, where the pH was

7.5 and the soil gave a light positive calcrete reaction.

Although no absolute differential species has been

found, this community possesses a characteristic

species combination: species shared with the Chry-

socoma tenuifolia-Polygala leptophylla Community,

Nenax microphylla , Aristida diffusa. Hibiscus marloth-

ianus, Trichodiadema pomeridianum , Dianthus basuticus

subsp. basuticus and Euphorbia c/avarioides, occur

together with the common Karoo species. Both

Pentzia g/obosa and Chrysocoma tenuifo/ia are

usually abundant in this community. Although

Aristida diffusa is typical it does not reach such high

cover-abundance values in this community as in the

communities of the steep slopes (Table 4).

2 (d) Chrysocoma tenuifolia-Polygala leptophylla

Community

This community occurs on gentle slopes of 5

to 10° of mudstone and fine to medium-grained

sandstone of the Beaufort series. Outcrops occur

regularly. Sometimes there is a slight dolerite influence.

The soil is a shallow lithosol, 0,02 to 0,30 m thick

and slightly acid with pH 5,5 to 7,0 (Fig. 3). The

aspect is generally south to east. Differential species

appear to be Convolvulus boedeckerianus , Polygala

Fig. 4. — South-east facing

dolerite slope with Olea

africana-Maytenus hetero-

phylla Community, that is

also rich in Rhus undulata

var. burchellii, Euclea

crispa and Tarchonanthus

camphoratus.

172

PLANT COMMUNITIES OF TUSSEN DIE RIVIERE GAME FARM

leptophylla, Schizoglossum capense, Helichrysum ni-

veum, Osteospermum leptobo/um, Thesium spartioides,

Nananthus sp. and Eragrostis denudata. Further, this

community has species in common with the Chryso-

coma tenuifolia-Nenax microphylla and the Chrysocoma

tenuifolia-Lessertia pauciflora Community (see 2(b),

2(c) and Table 3). This community also has species

in common with the communities of the steeper

slopes and stonier areas, such as Gazania krebsiana,

Lotononis laxa, Heteropogon conforms, Cymbopogon

plurinodis, Rhus ciliata, Lightfootia albens and

Sutera halimifolia. The dominant species in this

community is usually Chrysocoma tenuifolia.

3. The communities of the steep slopes (Table 4)

Floristically and structurally the richest communi-

ties in Tussen die Riviere Game Farm occur on

broken veld.

The vegetation of all these communities consists

of three or four layers: a very sparse layer of creeping

and rosette plants up to 0,05 m high and usually

covering less than 5%, although sometimes con-

siderably higher; a grass, dwarf shrub and small

shrub layer up to about 0,80 m, covering usually

between 35% and 50%, but with exceptions of 15%

and 70%; a tall shrub layer, with an occasional low

tree, between 1 to 4 m high, of which the cover

changes between 10% and 30% in the Rhus erosa-

Stachys burchelliana Community and the Rhus

erosa-Rhynchelytrum repens group of communities

and between 25% and 55% in the O/ea africana-

Maytenus heterophyl/a group of communities. Some-

times there is also an open tree layer, up to 6 m in

height and with a cover of about 10% to 30%, except

in releve 556, where it covers 70%. Without further

refinement of Fosberg’s (1967) system, it is somewhat

difficult to classify these communities physiog-

nomically. The Rhus erosa-Rhynchelytrum repens

group of communities and the Rhus erosa-Stachys

burchelliana Community fit best under category

2B1 (2): evergreen broad sclerophyll steppe shrub,

whereas the O/ea africana-Maytenus heterophylla

group of communities fits best under formation 1B1

(4a): Mesophyllous evergreen broad sclerophyll scrub,

with transitions to formation 2B1 (2).

All these communities of the steep slopes have a

great number of species in common of which the

most important are Asparagus suaveolens, Aristida

diffusa, Diospyros lycioides subsp. lycioides, Rhus

erosa, Rhus undulata var. burchellii, Eustachys

paspaloides, Sporobolus fimbriatus, Olea africana,

Diospyros austro-africana , Enneapogon scoparius,

Sutera albiflora and Euclea coriacea. These com-

munities have also a number of species in common

with the Chrysocoma tenuifolia- Polygala leptophylla

Community of the gentle slopes [see 2 (d)] and

with the Chrysocoma tenuifolia-Lessertia pauciflora

Community [see 2 (b)]. Although the common Karoo

species are always present in these communities of

the steep slopes, their cover-abundance values

usually do not reach such high values as in the

communities of the flats and gently sloping terrain.

3 (a) Rhus erosa — Rhynchelytrum repens Communities

This is a group of three communities which

floristically differ only slightly from one another.

The three communities of this group are characterized

by the following differential species: Rhynchelytrum

repens, Pellaea calomelanos, Asclepias fruticosa, Phyl-

lanthus maderaspatensis, Aristida curvata, Haworthia

tesselata and Hyparrhenia hirta.

One of these communities, represented by releves

547, 545, 144 and 136, is relatively poor in species,

in that it lacks largely the above-listed differential

species, and also Olea africana, Euclea coriacea and

Themeda triandra. It occurs on steep sandstone and

mudstone slopes of 12° to 21° with a northern to

eastern aspect. The soil is always very shallow with

a pH between 6,0 and 7,0 and there are many rocky

outcrops. A second community in this group is rich

in species and is characterized by the occurrence of

Hyparrhenia hirta and the importance of Rhus

undulata var. burchellii. This community also

occurs on steep sandstone and mudstone slopes of

13° to 30 J with a northern aspect. The soil, which

has a pH varying between 5,5 and 7,0 is between

0,10 and 0,70 m deep, which means that it is not so

shallow here as in the first community of this group.

Although rocky outcrops occur, they do not play

such an important role in this community.

A third community of this group is characterized

by the absence of Hyparrhenia hirta, and the slightly

higher cover-abundance values reached here by

Euclea crispa and Enneapogon scoparius. It occurs

on steep (23° to 31°) dolerite slopes, also with a

northern to north-western aspect.

Releve 544 was situated on the nearly level summit

of a dolerite koppie, near the northern edge. Soil

only occurs in fissures and is slightly acid with a pH

from 5,5 to 7,0. These last two communities have a

few species in common with the group of Olea

africana-Maytenus heterophylla Communities: Chei-

lanthes eck/oniana, Lantana rugosa and Euclea crispa.

Releve 548 is different from the others in that is has

Argyro/obium variopile and Rhus lancea with high

cover-abundance values. Releve 541 is extraordinarily

rich in species.

3 (b) Rhus erosa — Stachys burchelliana Community

In this community Rhus erosa plays an important

role, although Olea africana, Tarchonanthus campho-

ratus, Rhus undulata var. burchellii, Rhus ciliata and

Diospyros austro-africana also can be important in

the shrub layer (Fig. 3). The community is charac-

terized by a number of differential species, of which

Cheilanthes hirta, Stachys burchelliana and Chamarea

capensis are the most constant. The community has

the following species in common with the group

of Olea africana-Maytenus heterophylla Communities:

Tarchonanthus camphoratus, Senecio hieracioides,

Celtis africana, Selago albida. Asparagus laricinus.

Cotyledon decussata-orbiculata complex, Elyonurus

argenteus and Pavonia patens.

The community occurs on steep dolerite slopes of

9 0 to 27 0 with a south-western aspect. The only

exceptions form releve 92, which lies on a sandstone

slope of 6° and releve 102 which has an eastern

aspect. The soil virtually occurs only in fissures and

is always very shallow with a pH from 5,5 to 6,0.

Releve 108 differs from the others in that is has

Acacia karroo with a high cover-abundance value.

3 (c) Olea africana-Maytenus heterophylla Com-

munities

On steep and protected south-facing sites in Tussen

die Riviere Game Farm a group of communities

occurs in which the tall shrub and tree component

t rr -> t

_ ^ ^

</i o «o

v"< O ^

> ^ VI X ''"

c/3 O

^ 3 JN *o o

v~i 7- ‘“O— -*-

2£ ->. — * wurun

x rn ;£ m - - *

72 ^0>C

<0*0 0

i - -r~~

1 O VC

*■> * rj- v> © ®

Z^S^n-T *oe

C/3 ^CC

^ * *n wom

X ' 72 ^ -r; x -r-~

^ O <o

2? •> r-~ «-> o <n

« -r-

C/3 °OvC

S; £ > :z "yn©

73 ^ O io

2*^2

C/5 w O «->

o§...S <o«oo

(N — > —

M ^oV°°

ON £5 f"0'«^0

M +©^°°

75 TO NO

+

4- 4-4- 4-

+

JJ1 £2* NO V> V"> in

^ “ O «n ^

+

S 25 i . vn o o

£ "I | X ©V'°

T T 2> on v-i O O

1 *V^ *■' ^ \j

^ o r-

T VO ^ <N ,

10 z

X CM > <N.

*° Z

oo m -^ —

T m ^ co .

<o rz

oo m r-~ vn in

£ Z <n -a - | «n

<NI oo ^ ro >o »o

- ~cn

I oo ^ I

> V

1 z

ro CN in VOON

Vf^lhy/V n nr— I

■O Z ^ 0 r-

25 ^ O T >n 'n

„, - »(N

<o in o o

-H UJ ^ «2

^nou(N •omjn

-• Z 00 cT vcT~

2 Si m t'' >n O O

U-l /TJ c nr-f-

10 Z M© r-^

£r 25 o -a «o o o

I * "T

v-> X ~to r~-

C/5

d o

C H

«8

x .' :

c • . ^

g-p

dflb

■ — 'f-j C/5 >

fli • ~ N p*"* Cl o, Q

'U _o tj w oft o °

> T q u Qt3 « — i

“ £ a Q.15 _ , w 3

q 3 « o <u 'S 35 o

ft Z <5 73 O 72 O.H

§

Q

+ +

+ + + + —

+

+ +

+

C

X

x

Qi

x

3

X

m

- X

U.S

T-J

£ 5

i/ C/3

C w

<U c/5

:h

a p X

■ a

: as

^3

a3

c

aj

>.

X

o

rt

Q, O

J«S

3 h

Z S'

& v> <U w 3

3 O T3 _* -72 T

“ o 3 i3 «.b

^ 1 — 1 ■

I5l &

<u cj

S o.2

2 £ o - r

H 0,c! « V;W.S

' ' -^ 3 3 cu e

<2 X 3 •— <U

P ca «o

c > « ^

; x ; :

• • .

. <U • .

:N.^ :

■ o^j 5 .

.' • 3 !

•S5 ;

iWO •

t: P’S : C

<U n S Jp ,«2

c -2 ° y <u

3 5w bv;

3 5 wx Z

j£rt®-

, C .x: C/5 —

3 £ g T ^

IgSog

H K >>"3 3

3"3 q X

&0

_cti

">5

X

a

o

+ +

+

C/5

- ^ rS

• X *-*•— a/ -a-/

• ^ Oi -

<-> CXa c

O

ssiliii

_^*=l o CnZ 0 Q.

Qi C/5 f— l_. cn >-»

ob a, < a, < X X

&

Q

U

- o 0,0. c

5 3 cn a> 2

a) 3 x:

^dc-S.S-

>. S (U 3 OT

x £ — 5? o

3 J 3

07- <- o n

7)U2a. J

§

Cj

X • « : 2

c • — ’ Q • -3

3 C C O

X

§ g

E-22

OJ T 3 x c

XffcOU

E > g IE E

3 03 St O —

^X rt 3 X

T 3 — 2 3

03 3 ■— o 3

^ to ^ .22 S

’qj ctf o aj

SoSiu:

N

s

o

m

_3 >

>5i2

x Z

x

O g

a3 ^

« 3

u

= J3 j

QJ 3

rt* 3

_ II-

•^C^X o3

d> 8i o®

s-m c^i x-

C/5 /t^ i • i- “ C/5

z-gz g g1!

■2!s c3 o

S«£ ° § x

TJ n x _C 2 „

OX g 3

o o. , eg 'C X

_j 3 >> c/5 O 3

.3 U. OJ Q u. 3

£ «x y tj e

O X

Nvl 4-/

x

>>

o

-N

U

“ on

U^3_

X T X

3 Q O £2

wuj 5 ^

_3 &$ M

=^5 :

• C/2

X g

- ° a.

uj

3

X

iS ' — '

3 3

'E .2 3

n O 35 "3 —

9-X fe c.2 3

22-gggS

3 h S o’C

Q v 3 c/5 ^ □

■§•^§■1,3

Cj

U( r

o U

x: Q

EC/5

0)

C

C 00 c^

ac Sg|3 0|

0 U .5

:2 :qc

■ — X

x" 3 Ju

3

CQ

1_X

?z

^ ^

— ^ CL> t- ^ Ctf

OOmujil

_z w).:

*— cj '

C/5 g • —

~~i c , ^ *—

X “ C I) O U

2-S 3 g_ 2513

03 q .5 ^ — o

UO-KU^

•c 5 3 o

5x3

5 ~ c .2

R 5 3 x

~U o .2

?Cu

7 u -

R

X

CeC

3

X

Oi ■

4- 4-X

T i nj

+

+ +

^ +

+

++

+

+ +

o

*- X ra X *-

oS T3 — X ccS

.2 X sq C CO X

S I

R 5

2 S

03’

2 §C

2^o o

. = q c ^3 3 X— ° >

rq u 2 r9 - -a

CJ on < CJ X Cl.

03 W 03 U.

— ; 3 <u Q

CJ C ”o ^

-X 3 C/5

O X X .a

— <N <N «N —

+

m ■mm —

(N — <N —

+ + +

N 1 - 1 N — 1

+ ++ +

„ +■

+ + +

-+ +

+£_3++ +_^++

+

+ + + +^

+

+ $+ +

"T

+ +

+

+ +

+ + + +

+

+ ~ +

■ + + +

+ + + +

+,

+ + + +

+

+ +

— — 1 CN

+ + + + + -

+ + +

+

+ +

+ + + +

+ ++++ +

+ +

+

+ + + + + +

+

+ +

+ «i +

+

+ <3+<3£++~

+

« + + +

+ “1“ +

+ +

+

+ +

+ + +

~+ ^+++ ^

<N +— +^«

+

+ + + +

+ - ++ +

~ +

+ +

+ + +

+

— — + + +

f3__+ + _++ +

+ + +

+

CD

o

C 3

O •“ u.

O X

X C

> rt O 3

Cj C/5 •— JZ

-or"

>^r-

^ 8

> 7

O h- -

o

CT 3

o c

o3

C/2

^ X

SO 03

CO g

p p

3 X u X

03 X T

03 g-E^

ctf 3X 3

—2 X r#1 3

3 2 ert _ O •«

x ”2.o g $px q-

.EoQ.S o5g2 1

a 3 3 03

c 3 X

03

>1 o

X

n c/2

<<

Q o.

3 i5 _

03 O

t/2 00 . o 52 O

.03 •- ”• w

X"2 3

y s .s = . ^ ^ ^ ^

O oS =

3 3 C ~ 2 3

I-- O 3 r- E 'R 3

3e^^«3X°-2

3EO.3-S3 0-C

g MX o‘E-2 “ 3 “

23rtog°EC.2

S-S 2q p S^2

x 3 g| g b e?-|^:a

(bttJMOQBMajKMOQ

Cj Cj O !

£&><:

r: ^

c^> C^

D

X) O C

c x ^

^ 3 -5-

^ x c rs c

^ H £ tuo o

>— J — * ' r" c-

0/5 ^ rj

C/5 3 ■ CJ

^ O.

«5 E - .■

X'C/S 02 X

-oa^ j 3

3 4 X < _g

3 < CnT

Cj

2 q/ 2

00 ^ X

72 E b o *1

-.- = w o 2

Cj x

03 . u

Zn S

X; o < E

3-02 -b _

3 X Z T

wti “x N

x g y c<^ g

o3a^ y

3 7- r n

R

o

"g

5

fs

UJ

so ^

,<u

5^ CJ £ 23 —

rr o

3 3-g^

.2 J g a o g o

Q o'

E75

3 ;—

E E

E 3

£ x o

3 gx

o/J »-- C“ /. N ^

•— rri CJ • —

X _0 Xj

7) •— o 3 X

cr c>2<<

"co X

O 0 03 _> >-,

3- 3 fc ^ S

g r- O 72 •“

U-£.S O C3

os a s- 71 • -

O 3 X *■*

f#r Eji

m _ P O

X OJ

.5 5^”^“

1— t/2 O 3

3 C C.2U 3

O O ~T 3

”o -R 3 O ~

■~ 3- o X £

oi

x -=

M

— 3

o O £ x ^

x x oa < cl

JCC[»r

3 O o _i 3

■3 50 W).2 3=

• 5 3 O -R 3

X X Q. 2

3 _g o£

^ OX o

- xg

, 2 -P

c ,

u

1J J-J

iS p

C/5

Ojc C

cd _

o

3 3 —

■= - —

?x o

C 3 O

3 3';

03 3 X

O R O

60. y y

q o. o

ft Oc

o O

“ O x

3 O

X < 2C

O

o

K 72 gQi

-1

C

.x a

3 -a x

g ’5 3

q._ ><

cd

^ 8

‘

_ cQ

3. S3 «

c c

a

E 3

= uQ

3 >

Rj

o UJ

03 C

^J3 o

330

C 3 X X

w U « C

03 03 r O

X — -= 72

. 73 .

O t- •- c

— =; <u c

- c

cd

£ gw

cd

3 cd 3

r cd ^

C/5 "O

O CJ C

■ “ 3

C/5 . —

CdrSi ^

— w cd

o y-a

"EE =r

X >,0.-3

Ob CJ Z X 72

N

5P tb q cd

JO

60 - • —

C6 « O

: — C >-

Blf-J

3 ^

quo

P5 £ u

XjI

3^ o

o -r-

O «(■-

7) 3 03

>> 60X

03 o3

5 Uh^

cd C/5

c/5 O -

O SXD-C —

JD c w cd

00 ^

— f\ c/5 cd

My,“ t

3 ts,«

.5 x o o

n x rv eo

T V2 2J- —

3 •— 3 "O

O ^3 1- c

X < UJ X

z

3X 3

— ^ s—- c.

rr cj .-

•S U X

O >> C/5

3 g 73

cj r _

o

vj n 'tS >’

O X c X

o = q x

Un&O

“ 3 s-

.3 p (j o

“ 72 p

3 > « :

“gi.

C/5 -> 1 <Ti -

— C/5.2

(>o Ox^

3 i|

C/5 / \ — ,

cd ^ cd ,

cd

X s

0 3-5

»2 2

3 c <u

wOx

C 3 2 =

C/5

C/5 •

3 .2 03

Ico2Z

3 ,

£ = o

c •— .^2

cd ”0 lc

SsS

U

x -3 S 3-'—! .S £

,- 3 “p^o E ^ x 3

j = ..'o = C

vi f ov .2?^ o -p .2

-3 m in "2 n . Or

=.s »“ I4* o g

sll+lsl1

n cinSo^S

J— cn 'xr *0 c/5 , ^ - — s CZ

cd •— £•*/-> ^ — CJ

X os py T3

^ 3 g J 3 Q _ 3

< o — CD .q wX TJ

X2H n • C r!2 c

1 U-3 tLj £ "-'X

I - tJ- « /S n N U

. /— n 0 T’ ^

!^+o 3H

On

02 — rs=

qouS.^w^

n(N Cdt^ +

^3 ..

5b X X tQ. p 00 _

3 c_ 5 « 3h

!X25 5 S-£x:

CS

a c 5

3 l“J U == '7J Q. o >,

«.sc ?«sSq

72_ggT = Eo3^

■3 C CJ ^ ’77, OQ

£OCJ n C. ^ ^ ^

3 3 g- SXX =

^ 3-§^ Er&x

g y a^3 5 52

«(•£ g r <u j: ^ x

3Up “>g-3-X e

xP^ a j- o -3

< .. 1 00 .2

jJ-2 g+iJM

jq ^UJ .=

'N — , V2 _ WR ”

JJ N p.23 Ov . . y W

3-C

X-g03 3

“ § V2 V2

J X O ,

3 u g I X ^

. ps ■£ c« '5 7!

W. W) C c >^oc

Z.— • • 3 cd r-* o rn

p*. 00 cd c ^ •zp

3.22^ u7 a 3

g-o^jg .«5.s

•nE-S^PES

03 u x ^T ca

> ^ X . x 3 (U

J ” y Jn ^ PT

oS^r+ES|

; w as— 3. .-g

172

To:'.—

vj 3 . . -o

w > T C

jJ 72 T 03

00 o3r ‘

UJ

.. • T • • ' • 3

>,>*oiw Ex

60 "5 3 T p -5 — ■

<Q «

m >n

_ 3-0.-

T £ \3 2 cu

mX _

: Sts 60 -j- o

• +-* m 1 • —

TD

Cl cd 13

< ~ 1 ^ CJ Q.

= $ >. n c

-•g g£x iiJSx

- sf s|

-X

O ”>> 3 _J_ Ob ..

• Cl-OST

X E 60 N o - -

C/5 ^ cd

-- 1b3 .. „

—

03 - — v,

Oj < • 72

'c x y -g

\|_y v O CJ '~l

PSg?II«S

5^-s=i°

^MiSON, . _

^ .. 03 w >- X

-'Tj q co • •

Tj- C c/5

• • vi 3 q ov

O hJ

I|x

£ 3 ;

>. eo

' C 60.

o O

< 2

o3

: O a>

: x x

ni ' ' “ ~ ^ ^ ^

C4 O £3 3 E'g

.S3 X X 3 'S 3 -ft

c 0 ^

>..S o^u5S.SP^

.j au ^

Q-5 0.8 E >.2

,q 3 x eb *3

£0- _«S

j^xg E-^r+l5

« yQ B^+.v g.

03

(U ■

« r- ..VO

>v C VO X

3 m V>

75

3U+^.^-

cj - • • 0 o r^

1 0 1 Z-— — [— —1—

X +T <D J -4 ~ 1

§^x^+s§§p:

O O • • S c >>X Ov

x-SivnT^ cuv-m

< C3C£t>0 3^o!q

. 3 (3 7) ^ ^ Jj w

n-v T V2 1 W

W _o pcw.2j

vg.22 •• p 03 £

X X Q <

— E o M ;

W C C 72 !

OS

» t! X u

. cd Cl o

■ C cdrr

_• 2 3 -6 . 0 o 0 x

<■ 3 *5 u £ § c 3

5x g-s « g ^ « E

J -^2 3 1 3 73 ca

§ 5^ l^c

o y~_i_£

o 3+E—ajjg‘N

.y oo 3 ^ p- v£ .

o£t u£?+« +

J3 ' . - .

<— C ^ ^ On

O lJ n< 3^ _LW R

<00 8 +3 « ^X

- .12 *0.^ 3 =

TobOipx:"^^-1-

.. 3 o^T 2 r 3 td

cN cd CX n . . w cd T N

JO s p -4-o J!: ^ = U

y cj -5 b cj w.

. <N

3

fc — PJ J •— X

• • b 3 . E N = OS 3

^ 7! d- 75 g ^>V'P £

wx gj3 J(V x2-

s-^sJ+-&S|

0 r -4-. 2 ^ ^

o< 1 s

• • O

-COv Jtr

cd

, nCtn .r: 5 rn x

S+-CI^ ft

cd

• »o ,

t\J -P ~ uu

§or|ii^i

r^“in + V I

-g JV.Q ..Ob 60

< §Jx°2??2

— ^ s + g 3

0,0 3ox.y

' — "X 3. r > C „•> H

•Ey ?ai -x

,,'U ft b? E +

X V2 C 72 • • ft <U F

IjS b 33 2> c

J-< £ftT 3 •$■ 4>

O . c o 7 a"^v2 ^

o— Uft 03 3 O'X

^-4” ' i: >. y os y q-

os ^uoiHeoos

Z72 • •

KSOvCC XO -

ui x0-

>,00

C/5 C

» SO

•n — ^ cj

cd

-ft <

X >-

' X <S! —

O C

>V O

' vO X 02

vO • •

m —

<n ov

o

M. J. A. WERGER

175

plays an important role. They are characterized by

the following differential species: Melica decumbens ,

Crassula harveyi, Mohria cajfrorum, Hibiscus aethio-

picus var. ovatus, Crassula filamentosa, Maytenus

heterophylla, Crassula setulosa and Grewia occidentals.

Olea africana is the most important species in this

group of communities. As has been mentioned before,

they have species in common with the Rhus erosa-

Stachys burchelliana Community and with two of the

Rhus erosa-Rhynchelytrum repens group of com-

munities.

On steep dolerite slopes of between 24 ° and 33 °,

protected by nearby ridges, a community occurs in

which apart from Olea africana , Rhus undulata var.

uurcnellii, Euclea crispa, Tarchonanthus camphoratus

and Themeda triandra can be important (Fig. 4).

Releves 115, 119, 138 and 550 are examples of this

community.

Releves 556, 536 and 1 1 1 are each examples of

different communities, belonging to this group of

communities.

Releve 556 represents the vegetation in a dolerite

kloof, near the ridge of a mountain. There is very

little soil, and species like Cussonia paniculata and

Koeleria cristata occur here.

Releve 536 represents the vegetation in a well

protected sandstone kloof. Here, also, soil occurs

only in fissures. Species like Carex spicato-paniculata

and Gerbera viridifolia are typical.

Both the communities, of which releves 556 and

536 are examples, occur frequently and more typically

further upstream in the Orange River Valley.

Releve 111 represents the vegetation on a very

steep (31 °) south-west facing sandstone slope. It is

floristically not so rich as the two above-mentioned

communities, and is a more open vegetation, in

which grasses like Aristida diffusa. Digit aria eriantha,

Cymbopogon plurinodis, Heteropogon contortus and

Eragrostis curvula play an important part.

DISCUSSION

As has been pointed out several times (Tiixen,

1970; Werger et ah, 1972), more releves from a wider

area are necessary for granting fixed status in a

hierarchy to communities recognized in a study such

as this. It can then also be determined with certainty

whether weedy species like Asclepias fruticosa and

Aristida curvata are indeed differential species of the

Rhus erosa-Rhynchelytrum repens group of commu-

nities, as they appear to be from the present data.

Only with more releves available can it also be decided

whether releves like 541, which are rather rich in

species, are fully representative examples for the

communities into which they have been incorporated

or whether they are merely exceptions.

It is an immportant advantage of the Braun-Blanquet

phytosociological method that without much extra

effort an unrestricted amount of new releve data can

be compared with and incorporated into the existing

classification and that gradually an hierarchical

classification of the plant communities of the South

African veld can be compiled. The method is also

of utmost use for planners, pasture and veld manage-

ment scientists and many others (Tiixen, 1968-69)

as, from the tables, grazing potential and stage of

deterioration for each vegetation spot (releve), as

well as for each community, can immediately be read,

because it shows cover-abundance values and total

species composition for each releve and variation

thereof within a community. Thus, as Tiixen (1970)

has pointed out, a phytosociological table, because

of its synthetic character, but without loosing detailed

information, contains more information than the sum

of information contained in all single releves.

The application of only a physiognomic system

for classification of the vegetation is of limited use,

since these systems are always highly artificial. Thus,

vegetation types which are floristically very different

may be classified in the same formation and vice

versa. In the present study we see that floristically

similar communities end up in different formation

classes, for example the Eragrostis lehmanniana-

Chrysocoma tenuifolia Community and the Chryso-

coma tenuifolia-Lessertia pauciflora Community.

The Fosberg (1967) system provides a reasonably

suitable physiognomic classification for these Karoo

communities, especially if some refinements could be

effected. However, the definitions of concepts are

very rigid. The names of the formations are intended

to have the double function of label and diagnosis.

These two features of the Fosberg (1967) system lead

to rather peculiar word combinations in formation

names, that offend slightly, because they would be

contradictory in more common language usage. A

less strictly defined system, and one where names have

not so much the function of diagnosis but more that

of labelling, like the Ellenberg & Mueller-Dombois

(1967) system, would overcome this problem. For

many workers this latter system has the disadvantage,

however, of using as criteria characters that are not

strictly vegetational.

ACKNOWLEDGEMENTS

I am indebted to Dr. D. Edwards for valuable

comments on the draft and for continued interest

in my work.

SAMEVATT1NG

As deel van IBP-opname van beskermde gebiede

is die plantegroei van die Tussen die Riviere-wild-

plaas, O.V.S., geanaliseer volgens die Braun-Blanquet

plantsosiologiese metode. ’n Klassifikasie van die ter

plaatse voorkomende plantgemeenskappe word gegee.

’n Manier word voorgestel waarop ’n hierargiese klas-

sifikasie van die plantgemeenskappe in Suid-Afrika op-

gestel kan word. Elke plantgemeenskap is volgens Fos-

berg ( 1 967) se sisteem van plantformasies geklassifiseer.

Twee nadele van hierdie sisteem word kortliks be-

spreek.

REFERENCES

Acocks, J. P. H„ 1953. Veld Types of South Africa. Mem.

Bot. Shiv. S. Afr. No. 28.

Barkman, J. J., Doing, H. & Segal, S., 1964. Kritische Be-

merkungen und Vorschlage zur quantitativen Vegetations-

analyse. Acta Bot. Neert. 13: 394-419.

Braun-Blanquet, J., 1964. Pflanzensoziologie. 3. ed. Wien

New York: Springer.

D'Hoore, J. L., 1963. Soils map of Africa. Bruxelles: C.C.T.A.

Inst. Geogr. Mil.

Du Toit, A. L., 1954. The geology of South Africa. Edinburgh:

Oliver & Boyd.

Ellenberg, H., 1956. Aufgaben und Methoden der Vege-

tationskunde. in: H. Walter: Eirtfiihrung in die Phytologie.

Bd. IV, T.l. Stuttgart: Ulmer.

Ellenberg, H. & Mueller-Dombois, D„ 1967. Tentative

physiognomic-ecological classification of plant formations

of the earth. Ber. Geobot. Inst. ETH. Stift. Riibel, 37:

21-55.

176

PLANT COMMUNITIES OF TUSSEN DIE RIVIERE GAME FARM

Fosberg, F. R., 1967. A classification of vegetation for general

purposes, in: G. F. Peterken: Guide to the check sheet

for IBP areas. IBP Handbook No. 4. Oxford: Blackwell.

Harmse, H. von M., 1971. Geological maps of the Upper

Orange River catchment area , I: 50 000. Potchefstroom:

Dept. Agric. Techn. Serv. Unpubl.

Kendrew, W. G., 1961. The climates of the continents. Oxford:

Clarendon Press.

King, L. C., 1967. South African scenery. Edinburgh: Oliver

& Boyd.

Kirby, P. R. (ed.), 1939. The dairy of Dr. Andrew Smith,

director of the “ Expedition for Exploring Central Africa",

1834-1836. 2 Vols. Cape Town: Van Riebeeck Soc.

Knapp, R., 1971. Einfuhrung in die Pflanzensoziologie. 3 ed.

Stuttgart: Ulmer.

Loxton, R. F., 1962. The soils of the Republic of South

Africa: a preliminary reclassification. 5. Afr. J. Sci. 58:

45-52.

Schulze, B. R., 1947. The climates of South Africa according

to the classifications of Koppen and Thornthwaite. S. Afr.

Geogr. J. 29: 32-42.

Truter, F. C. & Rossouw, P. J., 1955. Geological map of the

Union of South Africa. Pretoria: Govt. Printer.

Tuxen, R., 1968-69. Reflexions sur l'importance de la socio-

logie vegetale pour l’economie de l’herbage europeen.

Melhor ament o 21: 187-199.

Tuxen, R., 1970. Pflanzensoziologie als synthetische Wis-

senschaft. Meded. Bot. Tuinen Belmonte Arboretum Land-

bouwhogeschool Wageningen, Vol. 12: 141-159.

UNESCO-FAO, 1963. Bioclimatic map of the Mediterranean

Zone. Ecological study of the Mediterranean Zone.

Explanatory notes. Arid Zone Research 21. Paris:

UNESCO-FAO.

Van der Merwe, C. R., 1962. Soil groups and subgroups

of South Africa. Dept. Agr. Sci. Bulk No. 23. 2nd ed.

Pretoria: Govt. Printer.

Van Zinderen Barker, E. M., Jr., 1971. Ecological investi-

gation on ravine forest of the Eastern Orange Free State

( South Africa). Unpubl. M.Sc. thesis, Univ. O.F.S., Bloem-

fontein.

Volk, O. H. & Leippert, J., 1971. Vegetationsverhaltnisse

im Windhoeker Bergland, Siidwestafrika. J. S. W.A.

Wissensch. Ges. 25: 5-44.

Walter, H. & Leith, H., 1960. Klimadiagramm — Weltatlas.

Jena: VEB Fischer.

Weather Bureau, 1954. Klimaat van Suid-Afrika. I. Klimaat-

statistieke. Pretoria: Govt. Printer.

Wellington, J. FL, 1955. Southern Africa: a geographical

study. Vol. 1: Physical geography. Cambridge: University

Press.

Werger, M. J. A., 1972. Species-area relationship and plot

size: with some examples from South African Vegetation.

Bothalia 10: 583-594.

Werger, M. J. A., Kruger, F. J. & Taylor, H. C., 1972.

Pfianzensoziologische Studie der Fynbos- Vegetation am

Kap der Guten Hoffnung. Vegetatio 24: 71-89. also as:

A phytosociological study of the Cape Fynbos and other

vegetation at Jonkershoek, Stellenbosch. Bothalia 10:

599-614.

Bothalia 11, 1 & 2: 177-180 (1973)

Notes on the phytogeographical affinities of the southern Kalahari

M. J. A. WERGER*

ABSTRACT

The chorological division of southern Africa as proposed by White (1965, 1971) and Volk (1966) is

discussed. Reference is made to distribution patterns of some typical Karoo-Namib taxa as discussed

by Volk and to the remarkable distribution of disjunct taxa in the arid zones of Africa. On the basis

thereof and of the similarity in the floras of the southern Kalahari and the Karoo-Namib Region,

the conclusion is drawn that the chorological division as proposed by Volk (1966) seems justified.

In the early days of phytogeography, chorological

units were distinguished on extremely fragmentary

evidence. For instance, Grisebach (1872) based his

classification of the vegetation of the world largely

on the information extracted from the limited number

of diaries of early travellers and adventurers. Africa

was divided by him into the following five floristically

and physiognomically defined vegetation areas:

Mediterranean, Sahara, Sudan, Kalahari and Cape.

His Sudan area covers most of subsaharan Africa,

whereas the Kalahari comprises South-West Africa,

Botswana and most of the Transvaal and the Orange

Free State. He regarded the Orange River as the

boundary between the Kalahari and Cape areas, a

decision he based on Burchell's observations. Sub-

sequently extensive collections of herbarium material

have contributed decisively to alter and refine the

chorological classification of the world. Thus, pro-

gressing from the work of pioneers like Engler,

Lebrun and others, in recent years a fairly satisfac-

tory phytogeographical classification of the African

continent has been achieved. Monod (1957) dis-

tinguished three major chorological units or “groupes”

in Africa: the Mediterranean group, the tropical

African group and the Cape group. The tropical

African group he subdivided into four regions: 1

region soudano-angolane, II region guineo-congo-

laise. III region afro-alpine, and IV region Karroo-

Namib. He further subdivided the Sudano-Angolan

Region into three types, the Saharan Type covering

North Africa, the Sahelian Type covering Sub-

saharan North Africa, Ethiopia, East Africa, and

parts of Rhodesia, Botswana, Angola, South-West

Africa and South Africa, and the Sudanian Type,

covering the zone between the previous type and the

West and Central African rainforest zone. His

Sahelian Type is again subdivided into a northern

and a southern part, the latter of which comprises

the Kalahari, the Transvaal Lowveld and Bushveld,

and the Highveld. Within the Karoo-Namib region

Monod (1957) recognized three Domains: the Karroo

Domain, comprising the area that is also geographi-

cally indentified with that name except the coastal

zone; the Namaqualand Domain, comprising a

narrow inland zone in South West Africa along the

escarpment, tapering out on the southern part of

the Angolan coast; and the Namib Domain, covering

the entire arid winter rainfall area of South Africa

and the Namib desert.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

5292-12

This phytogeographical division of Africa was

largely accepted by White (1965, 1971) and Volk

(1966). White modified Monod's classification in so

far as he renamed the Sudano-Angolan Region to

the Sudano-Zambezian Region. He further regarded

the northern Sahelian zone as an impoverished

western extension of the Oriental Domain into the

Sudanian Domain, whereas he combined the southern

Sahelian zone with Monod’s Angolo-Zambezian

Domain into the Zambezian Domain of the Sudano-

Zambezian Region. Besides the Afro-Alpine Region,

White also distinguished an Afro-Montane Region

(Fig.l).

It is generally accepted that the Central Kalahari

has closer floristic affinities with the Zambezian

Domain and should be incorporated into the Sudano-

Zambezian Region, rather than into the Karoo-

Namib Region (Bremekamp, 1935; Monod, 1957;

Mendonca, 1961; White, 1965, 1971; Volk, 1966).

White (1965) discussed the distribution of the tree

species of the savannas and woodlands of the Zam-

bezian Domain. He found that in the Kalahari this

woody flora is poorly represented, and that the

species usually occur as shrubs, whereas they form

well-developed trees in the adjacent part of the

Zambezian Domain. He suggested the Kalahari is

transitional between the Sudano-Zambezian and

Karoo-Namib Regions, but included it on his map

in the former Region. Later, White (1971) accentuated

this point, and drew the boundary between the two

Regions within South Africa and South West Africa

just north of the Orange River and from there

northwards to southern Angola (Fig. 1). Thus, the

south-western boundry between the two Regions

coincided closely with the distribution boundary of

Acacia haematoxylon. Volk (1966) accepted White's

chorological division except for this south-western

boundary between the Karoo-Namib and the Sudano-

Zambezian Regions. He (Volk, 1964, 1966) discussed

the distribution of several species belonging to what

he called the Afro-Meridional-Occidental floral

element, and thus supplied evidence for his suggestion

to include the area of the Southern Kalahari as

delineated by Leistner (1967) into the Karoo-Namib

Region. Besides his discussion on the distribution ot

Rhigozum trichotomum , Stipagrostis ciliata, Parkin-

sonia africana , Aloe dichotoma, Stipagrostis hoch-

stetterana. Acacia haematoxylon and Sesamothamnus

guerichii, which he considered typical for the Afro-

Meridional-Occidental or Karoo-Namib Region, Volk

listed a considerable number of endemic, monotvpic

178

PHYTOGEOGRAPHICAL AFFINITIES OF THE SOUTHERN KALAHARI

or small genera that characterize this Region. Based

upon this evidence, he drew the boundary between

the Sudano-Zambezian and the Karoo-Namib Regions

north-east of the southern Kalahari, the western

Kalahari Thornveld and the Vryburg Shrub Bushveld

(Volk, 1966) (Fig. 2).

Apart from the distribution patterns discussed by

Volk, there is another important feature of the

flora of the southern Kalahari that supports its

inclusion in the Karoo-Namib Region. As pointed out

by Range (1932), De Winter (1966, 1971), Verdcourt

(1969) and particularly by Monod (1971), a large

number of taxa that occur in the Karoo-Namib

Region also occur in the northern Sahelian zone, the

Oriental Domain and the Saharo-Sindian Region. In

southern Africa these disjunct taxa occur in the

Karoo-Namib Region as delineated by White (1971)

as well as in the areas that Volk (1966) has additionally

included in this Region. However, these taxa are

absent or virtually so in the remainder part of the

Zambezian Domain as delineated by Volk (Fig. 2).

Recently Tolmachev (1971) pointed out that, in

various floras of the same floristic region, the sequence

of importance of those families that comprise most

of the species of a flora is remarkably constant.

Also the percentages of species contributed to a

flora by these families are fairly constant. This is

independent of the total number of species of a flora

or the extent of its area. These premises provide a

means of testing the floristic affinity of the southern

Kalahari with the two floristic Regions in question.

Since Acocks's Veld Types (1953) are floristically

well-circumscribed units (comp. Grunow & Morris,

1969), and rather complete species lists of several

Veld Types are available, they provide a suitable

basis for a comparison of their floristic composition

with that of the southern Kalahari. Therefore, the

composition of the flora of the southern Kalahari,

based on the check-list produced by Leistner (1967),

and a species list of the Vryburg Shrub Bushveld

with invading Karoo species [Acocks, Veld Type

17 (2)], was compared with the floras of the Karroid

Broken Veld (Acocks, Veld Type 26), the Central

Upper Karoo (Acocks, Veld Type 27), the Arid

Karoo (Acocks, Veld Type 29) and the Orange River (

Broken Veld (Acocks, Veld Type 32) as floras that

form part of the Karoo-Namib Region, and with

the flora of the Central variation of the Bankenveld

(Acocks, Veld Type 61,b), the flora of the Jack Scott

Nature Reserve, located in the Bankenveld (Coetzee, I

1972), and the flora of the Pretoria magisterial

district (B.R.I., n.d.), as floras that form part of the

Zambezian Domain of the Sudano-ZambezianRegion.

The numbers of species of the first nine families

that contribute the most species to their floras are

shown in percentages in Tables 1, 2 and 3.

M. J. A. WERGER

I Mediterranean

E Saharo - Sindian

IE 5udano - Zambezian.

S * Sudanian Domain

0 • Oriental Domain

3d* SaWalian extension of

Oriental Domain

Z » Zambezian Domain

If Guinea- Cor^olian

I iCarroo— Nam'ib

21 Cape

• Afro— Montane

O Afro- Alpine and Afro- Montane

Fig. 2. — Phytogeographical

Regions of Africa according

to Volk (from Volk, 1966).

TABLE 1. — Karoo-Namib Region (number of species per

family as percent)

From the tables it can firstly be noted that all the

floras discussed have a striking similarity. Amongst

the most important nine families, several are shared

by all the floras. Also, the total percentage of species

contributed to their floras by these first nine families

is rather similar, although there tends to be a decrease

in this figure from the more temperate floras in the

south-western part of South Africa towards the more

tropical north-eastern floras.

TABLE 2. — Southern Kalahari and Vryburg Shrub Bushveld

(Number of species per family as percent)

TABLE 3. — Zambezian Domain (Number of species per

family as percent)

180

PHYTOGEOGRAPHLCAL AFFINITIES OF THE SOUTHERN KALAHARI

Of the four floras of the Karoo-Namib Region

compared, the first four families listed are the same,

although their sequence differs slightly. There is

some variation in families that are sixth to ninth in

the sequence. The Central Upper and the Arid

Karoos have the Chenopodiaceae and the Stercu-

liaceae in common, whereas these families contribute

respectively 1,9% and 1,8% of the total number of

species in the flora of the Karroid Broken Veld, and

respectively 1,9% and 2,1% in the flora of the

Orange River Broken Veld. The family Asclepiadaceae

is absent from the first nine families of the flora of the

Central Upper Karoo, where it only scores 1,5%.

Crassulaceae is absent from the list of the Arid

Karoo, where it scores 1,3%, and Euphorbiaceae

scores 1,4% and 1,7% in the floras of the Central

Upper and Arid Karoos respectively.

Table 2 of the floras of the Southern Kalahari

and the Vryburg Shrub Bushveld shows a strong

similarity with Table 1 of the Karoo-Namib Region.

Poaceae, Asteraceae, Fabaceae, Liliaceae, Aizoaceae,

Scrophulariaceae, Sterculiaceae, Asclepiadaceae and

Euphorbiaceae reappear amongst the important

families, and the only major differences are the ap-

pearance of the Cucurbitaceae in the list of the

Southern Kalahari and the absence of Chenopodiaceae

and Crassulaceae, although it should be remarked

that the Chenopodiaceae scored the same value as

Euphorbiaceae in the Southern Kalahari.

Table 3 of the floras of the Zambezian Domain is

considerably different from Tables 1 and 2. In all

three lists represented in Table 3, the nine most

important families are the same and occur in virtually

the same sequence. Remarkable is the appearance of

the Cyperaceae and the Rubiaceae amongst these

nine most important families. In the floras of Table 1

these two families only scored between 0,6% and

1,8% and between 0,6% and 1,3% respectively,

whereas in the floras of Table 2 these values were

between 1,4% and 1,7% and between 1,0% and

1,6% respectively. Other remarkable differences

between Tables 1 and 2 on the one hand and Table 3

on the other are the lower values scored by the

Liliaceae and particularly the absence of Aizoaceae

in Table 3. Aizoaceae scored only between 0,2%

and 0,8% in the floras of Table 3. Chenopodiaceae,

Sterculiaceae and Crassulaceae also scored low

values here.

It is clear that comparison of the floras of the

three different areas leads to the conclusion, that the

floras of the southern Kalahari and the Vryburg

Shrub Bushveld have considerably greater affinities

with those of the Karoo-Namib Region than with

those of the Zambezian Domain of the Sudano-

Zambezian Region.

Thus, based upon the results of the comparison

of floristic similarity, the distribution of the disjunct

arid taxa and distribution patterns of several typical

taxa as discussed by Volk (1964, 1966), Volk’s

(1966) inclusion of the southern Kalahari and the

Vryburg Shrub Bushveld in the Karoo-Namib Region

seems fully justified.

ACKNOWLEDGEMENT

I am very much indebted to Mr J. P. H. Acocks

for his kindness in providing me with his unpublished

checklists of six Veld Types and to Mr B. J. Coetzee

for making available to me his unpublished checklist

of the Jack Scott Nature Reserve.

OPSOMMING

Die plantegeografiese indeling van suidelike Afrika

soos deur White (1965, 1971) en Volk (1966) voor-

gestel, word bespreek. Daar word verwys na die

verspreidingspatrone van sommige tipiese Karoo-

Namib taksa, soos hulle deur Volk bespreek word.

Ook word daar verwys na die opvallende ver-

spreidingspatrone van die disjunkte taksa oor die

droe sones van Afrika. Op grond hiervan en van die

ooreenkoms in die floras van die suidelike Kalahari

en die Karoo-Namib Streek word afgelei, dat die

plantegeografiese indeling soos voorgestel deur Volk

(1966), geregverdig lyk.

REFERENCES

Acocks, J. P. H., 1953. The Veld Types of South Africa-

Mem. Bot. Surv. S. Afr. No. 28.

Botanical Research Institute, n.d. Systematic list of the <

species of Angiosperms of the Pretoria Magisterial District.

mimeogr. pp. 19. Pretoria.

Bremekamp, C. E. B., 1935. The origin of the flora of the

Central Kalahari. Ann. Transv. Mas. 16: 443-^455.

Coetzee, B. J., 1972. 'n Plantsosiologie.se studie van die Jack |

Scott-Natinirreservaat. Unpubl. M.Sc. Thesis. Univ.

Pretoria.

De Winter, B., 1966. Remarks on the distribution of some \

desert plants in Africa in VAN ZINDEREN BAKKER

Sr., E. M. (ed.) : Palaeoecology of Africa 1: 188-189.

De Winter, B., 1971. Floristic relationships between the

northern and southern arid areas in Africa. Mitt. Bot.

Staatssamml. Miinchen 10: 424-437.

Grisebach, A., 1872. Die Vegetation der Erde nach ihrer

klimatischen Anordnung. 2 Bde. Leipzig: Engelmann.

Grunow, J. O. & Morris, J. W., 1969. Preliminary assess-

ment of ecological status of plant species in three South

African veld types. J. S. Afr. Bot. 35: 1-12.

Leistner, O. A., 1967. The plant ecology of the Southern

Kalahari. Mem. Bot. Surv. S. Afr. No. 38.

Mendonca, F. A., 1961. Indices fitocorologicos da vegetagao i

de Angola. Garcia de Orta 9: 479-483.

Monod, T., 1957. Les grandes divisions chorologiques de I'Afrique

Londres: C.C.T.A./C.S.A. Publ. No. 24.

Monod, T., 1971. Remarques sur les symmetries floristiques

des zones seches nord et sud en Afrique. Mitt. Bot.

Staatssamml. Miinchen. 10: 375^423.

Range, P., 1932. Die Flora des Namalandes. I. Feddes Repert.

( Repert . Spec. Nov. Regn. Veg.) 30: 129-158.

Tolmachev, A. I., 1971. t)ber einige quantitative Wech- U

selbeziehungen der Floren der Erde. Feddes Repert.

82: 343-356.

Verdcourt, B., 1969. The arid corridor between the north-

east and south-west areas of Africa in Van Zinderen i |

Barker Sr., E. M. (ed.): Palaeoecology of Africa 4:

140-144.

Volk, O. H., 1964. Die afro-meridional-occidentale Floren-

Region in Sudwestafrika. in KREEB, K. (ed.): Beitrdge

zur Phytologie p. 1-16.

Volk, O. H„ 1966. Die Florengebiete von Sudwestafrika.

J.S.W.A. Wissensch. Gesellsch. 20: 25-58.

White, F., 1965. The savanna woodlands of the Zambezian 1

and Sudanian Domains. Webbia 19: 651. .681.

White, F., 1971. The taxonomic and ecological basis of choro- It

logy. Mitt. Bot. Staatssamml. Miinchen 10: 91-112.

Bothalia 11, 1 & 2: 181-190 (1973)

Classification and ordination of aquatic macrophytes in the Pongolo

River Pans, Natal

C. F. MUSIL1 J. O. GRUNOW2 & C. H. BORNMAN3

ABSTRACT

The physiography, soils and climate are briefly described for the Pongolo River flood plain pans

situated in north-east Zululand (27° S, 32° E), Natal, South Africa. Quantitative vegetation data ob-

tained from 60 stands distributed over 9 pans were summarized using a Braun-Blanquet procedure

and a Principal Components Analysis (PCA). Both approaches indicated that the aquatic vegetation

could be subdivided into two main groups of growth forms: the submerged and the floating and/or

rooted-floating aquatic plants, each showing a distribution influenced primarily by physical factors

such as exposure and depth of water. In the PCA, each of the above groups was defined by two stand

noda, each nodum representing a plant community that could be observed in the field. The Braun-

Blanquet analysis further subdivided the floating and/or rooted-floating aquatic plants into groups

of species on the basis of chemical and substrate factors.

INTRODUCTION Pongolo and Usutu Rivers (Fig. 1). The entire flood

The Pongolo River flood plain pans are situated

in north-east Zululand near the co-ordinates 27° S

and 32° E.

In this study, only the (macrophytic) higher

aquatic plants were considered, that is, from the

order Charales in the phylum Alga upwards. Owing

to numerous often conflicting definitions of the term

“aquatic plant” (Inversen, 1936; Raunkiaer, 1934),

no strict definition was adhered to. For the purposes

of this study, only those plants found growing in

the water during sampling were considered aquatic.

The Braun-Blanquet system can be defined as a

phytosociological technique which aims at a des-

cription and classification of vegetation, while

Principal Components Analysis (PCA) can be regarded

as a technique of axis construction in order to achieve

an efficient ordination of stands (Seal, 1964). Ordina-

tion is defined as the uni- or multidimensional

arrangement of stands so that a statement of stand

position, relative to other stands or to the axis or

axes of the model, conveys the maximum amount

of information about its composition. Although it is

true that most ordinations are constructed on the

basis of compositional information, the arrangement

of stands should also follow environmental gradients

of some sort.

In this study, the aim of both approaches was the

same: to investigate the possible relationships of any

floristic differences to selected environmental factors.

PHYSIOGRAPHY AND SOILS

The Pongolo Pans lie on Cretaceous and Tertiary

sediments (Du Preez, 1967), and extend along the

flood plain from Otobotini Drift in the south, where

the Pongolo River emerges from the Lebombo

mountains, northwards to the confluence of the

1. Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X 101, Pretoria.

2. Department of Pasture Science, University of Pretoria,

Pretoria.

3. Department of Botany, University of Natal, Pietermaritz-

burg.

plain lies below the 75 m contour and is on the

average between 0,8 and 4,8 km wide. The system

of pans comprises some 25 major bodies of water

and a mass of smaller pans which total some 2 600 ha

under normal summer conditions. During peak

summer floods, 10 000 ha may be under water. The

largest pans are between 80 and 280 ha in extent

and average 1,5 m in depth, the depths varying

according to the extent of seasonal flooding or

drought. The pans previously relied for their water

supply on the annual summer flooding of the Pongolo

River but, since closure of the J. G. Strijdom Dam

in March 1970, depend on irregular excessive dis-

charges from the dam (Coke, 1970).

A general description of the soils was given by

Tinley (1958). Hensley (1968) studied the soils in

detail and recognizes two principal soil types, terrace

and sandy soils. The terrace soils are derived from

alluvium deposited by the Usutu, Pongolo and

Ingwavuma Rivers. These range from leached red

sandy clay loams and clays, — oxisols of the Msinga,

Shorrocks and Makatini Series on well drained

sites, to saline calcareous black clay vertisols, such

as those of the Rensburg Series, as drainage deterio-

rates. The sandy soils lie adjacent to the terrace

soils of the Pongolo River. The aeolian sandy soils

are of marine origin resulting from successive decreases

in sea level since the late Pleistocene.

CLIMATE

Two main seasons are experienced in the flood

plain area (Fig. 2 and 3).

From May to mid-September, the dry winter

season, temperatures vary from 12° C at night to

25° C in the day. Strong winds are experienced in

this season, especially during August.

The summer rainy season is from the middle or

latter part of September to April, with temperatures

rising to 32°C and higher. The annual rainfall is

ca. 580 mm.

J.G. Strijdom

182

CLASSIFICATION AND ORDINATION OF AQUATIC MACROPHYTES

Fig. 1. — Pongolo River flood

plain pans (partly after

Tinley, 1958).

C. F. MUSIL, J. O. GRUNOW AND C. H. BORNMAN

183

35

Max and Min. Temp. 'C Otobotini station

•s-

Fig. 2. — Monthly average max. and min. temperature in °C from Otobotini over 26 years

nor

90 _

80_

.OTOBOTINI 39 years

NDUMU 37years

5

a.

I

_l I l 1 I I IU. l l i i

JAN. FEB MAR APR MAY JUN JUL AUC. SEP OCT NCV CEC

Fig. 3. — Average monthly rainfall in mm from Otobotini and Ndumu stations.

METHODS

Vegetation

Relative percentage frequencies and cover values

of species were obtained in each of 60 stands dis-

tributed over 14 sampling sites on 9 pans. From

south to north these pans were, Msinyeni, Mniati,

Mhlabongwenya River, Tete, Omengu, Sivunguvungu,

Sokunte, Nhlole and Mandlankuzi.

Sample sites were located at each pan following a

subjective assessment of the overall species com-

position and habitat conditions. It was considered

most important that each of these sites be uniform

and this was subjectively assessed using all the

properties of the vegetation and site that could be

directly observed. Each sampling site was subdivided

into stands (stand samples) along the depth contours

5, 20, 40. 60, 80 and 100 cm (depth permitting).

Each stand, therefore, was a line following a chosen

depth contour parallel to a uniform shore and on a

uniform substrate (Fig. 4).

184

CLASSIFICATION AND ORDINATION OF AQUATIC MACROPHYTES

Fig. 4. — View southward of

eastern shore of Omengu

Pan with two sample sites

A and B. Sample site B

subdivided into stands at 5,

20 and 40 cm water-depths.

Relative percentage frequencies were calculated

from data obtained from forty 0,25 m2 quadrats

placed in a stratified random manner in each stand.

Species area curves (Fig. 5) were plotted to determine

whether 40 quadrats would provide a representative

sample. Curves were constructed on data obtained

from three different stands: (a) containing a dense

cover; (b) an intermediate cover; and (c) a low cover

of aquatic vegetation. Thirty quadrats per stand

proved to be the minimum satisfactory sample size.

Relative percentage covers were calculated from

point quadrats placed in a stratified random manner

in each stand (using a long pole of small diameter).

Submerged and floating aquatics were sampled

concurrently. Point quadrats were placed until 80

strikes on the dominant plant species present were

obtained. This ensured that the information content

of each stand would be constant.

Fig. 5. — Species area curve

from three localities to

determine a satisfactory

sample size.

(

185-186

18 Potamogeton crispus +4 +4

21 Najas pectinata

36 Nitella tenella

5 Cynodon dactylon

2 Cyperus sp ?

23 Rotala tenella

24 Limosella capensis

25 Phyla nodiflora

7 Ludwigia stolonifera

4 Echinochloa pyramidalis

9 Nymphaea lotus

10 Nymphaea caerulea

20 Limosella maior

1 1 Nymphaea capensis

8 Ipomoea aquatica

26 Scirpus litoralis

22 Ceratophyllum demersum

12 Trapa bispinosa

16 Utricularia gibba subsp. exoleta

14 Azolla pinnata var. africana

13 Pistia stratiotes

1 Scirpus cubensis

27 Neptunia natans

6 Paspalum vaginatum

17 Utricularia inflexa var. stellaris

15 Wulffia arrhiza

28 Spirodela polyrhiza

3 Cyperus nudicaulis

29 Cyperus fastigiatus

19 Potamogeton schweinfurthii

32 Phragmites mauritianus

33 Marsilea macrocarpa

30 Potamogeton pectinatus

31 Najas marina subsp. delilei

34 Typha latifolia subsp. capensis

35 Ludwigia octovalvis subsp. sessilifera. .

4-4 +1

4-4

4-1

+2 4-2

Y

-l

-l -l

-l

+i

4-2 4-3 — 3 — 1

-1

3 -1

•1 -3

4-1 -2 4-3

-2

4-4

+4

4-4

+ 4

4-4

-1

+4 4-4 4-4 -3

4-4 4-4 4-4 4-4

Y,

31

20

12

5

3

2

1

35

21

17

16

6

5

4

2

X,

4-4

W

4-4

-1

+4

4- 1

+ i i

+4 +3

1 +1 -hi

4-1

X|

4-2

4-4

-1

4-1

-3

+4

-1

4-1

-4

+4

4-4

-1

4-3 s

+4 +4

-1

—2

-2

12

13

8

5

Total

36

3 112

2 4 2 3

3 3 4 4 5

Water depth range cm. .

Water depth average cm

Exposure rating 4 4

Bottom substrate class A A

pH range

pH average

Specific conductance /umho’s range. . .

Specific conductance/umho’s average. .

Na range mg/1

» average „

K range mg/I

» average „

Ca range mg/1

» average „

Mg range mg/1

» average „

Na

C^+Mg range

Na

S+Mg average

Si°3 range mg/1

” average „

range mg/1

*> average „

20—60

60

4 4 4

B C C

8,4-9, 3

8 9

220-310-7 750

N/A

28,0-700,0

N/A

0,56-5,56-10,0

2,96

3,9-36,6

N/A

6,25-237,0

N/A

1,42-1,72

1,62

3,05-3,46

3,31

60,02-1 180,0

N/A

4 4

A A

4

B

4

B

5—100

71

444444444

bbbddddaa

7,7-9, 1

8,5

239-290-2 100

N/A

28,0-36,0-230

N/A

0,56-5,55

3,66

1,5-7,2-36,8

N/A

6,25-35,06

N/A

1 ,36-2,08

4

B

1,75

1,25-4,20

3,68

23,7-482,3

N/A

9

1

B

9

2

A

6 7 5 6 5 8

6 6 6 8 5

4 6 3

3 3 3

5—100

44

3 3 3 3 4 4 3

B B B B B B B

7.2- 7, 4

7,4

230-235

230

27,5

2. 2- 2, 5

2,45

1,5-3, 8

3,47

7,50-8,75

8,57

1,32-1,74

2 2

A A

5—100

41

2 2 2 1

A A A A

7, 3-8, 2

7.6

290-420

391

30.0- 45,0

42,0

3, 9-6, 4

4,4

5, 6-5, 7

5.6

9.0- 12,5

9.7

1,0-1,92

5—40

17

4 4 4 4

A B B B

7, 7-8, 2

7,9

235-272

261

27,5-30,0

29,2

2,20-5,55

4,7

1,90-7,2

3,6

6,25-7,25

7,0

1,36-2,0

1,38

1,73

1,84

3,4

22,0

2,35-7,4

3,36

22,69-36,17

33,47

1,25-3,65

3,05

29,8-31,2

30,8

5292

'

C. F. MUSIL, J. O. GRUNOW AND C. H. BORNMAN

187

To test for adequacy of sample size, appropriate

statistical procedures were applied to point quadrat

data obtained in a test stand with an intermediate

cover prior to the start of the investigation. This

revealed that the 80 strikes mentioned would ensure

that the cover data for all species had a standard error

of less that 10 per cent of the mean. To ensure that

stands included in the data would be homogeneous, a

Chi-square test for homogeneity (uniformity of distri-

bution) of the plant species present was also conducted

from point quadrat data obtained in each stand. For

the same reason and in the same way as Curtis and

McIntosh (1951), only the strikes contributed by

species with a relative percentage cover of 25 per cent

and above were tested for homogeneity at the 5 per

cent level of significance. Point quadrat data taken

sequentially were divided into four groups for the

test. The Chi-square test, therefore, had three degrees

of freedom corresponding to a tabular value of

7,815. If a calculated value for the observed data was

equal to or smaller than this, it indicated homogeneity.

Water and soils

To supplement the vegetation study, physico-

chemical analyses were conducted on surface waters

in each stand. The following factors were analysed:

pH; specific conductance; temperature; O., concen-

tration; turbidity; Na+, K+, Ca2+, Mg2+, and Fe2+

concentrations; total alkalinity as total C02 (HC03~

+ C03~); Si— Si03— , Cl- and P04-3 according to

the method of Golterman (1969); and finally N03_

concentration according to the method of Muller &

Widemann (1955).

In addition, each stand was assigned an exposure

rating which was a subjective assessment of the

degree to which the vegetation in a particular stand

was exposed to the effects of wind and wave action.

Four degrees of exposure were utilized, ranging from

1 in protected areas to 4 in exposed areas. Bottom

substrate features such as colour and textures were

also briefly assessed for each stand.

TREATMENT OF DATA

Braun- Blanquet Analysis

Frequency data obtained from some 2 480 quadrats,

comprising 40x0,25 m2 quadrats for 62 stands,

were utilized. In this study two additional stands

from the extreme east and western ends of Nyamithi

Pan were included.

Stands and species were arranged into a comparison

matrix. Braun-Blanquet indices of cover abundance

and sociability were not used in this study. Each

species in a stand was assigned a frequency index

value according to the number of quadrats in which

it occurred out of forty for each stand. An example

of the frequency index values used is as follows:

Frequency

Quadrat groups index

values

Oto 5 -1

6 to 10 +1

11 to 15 -2

16 to 20 +2

21 to 25 -3

26 to 30 +3

31 to 35 -4

36 to 40 +4

Entire columns and then rows were manipulated

in the unordered matrix to give the best possible

ordered matrix according to the method of Braun-

Blanquet (Kuchler, 1967). Selected habitat factors

measured in each stand were written in the table.

For greater clarity in presentation, the range of

variation and average values of each environmental

factor rather than the individual stands, were recorded

under each group of stands classified in the table.

Principal Components Analysis (PC A)

The basic technique of PCA has been described

by Kendall (1957).

Quantitative data on presence-and-absence and

cover obtained from 59 separate stands were utilized

in this analysis. Only those stands which passed

the Chi-square test for homogeneity were considered.

Species and stands were arranged into a comparison

matrix, each species in each stand being assigned an

importance value (IV) which was calculated as

IV=2x Relative % Cover + Relative % Frequency

3

Species density was not measured in the stands

as individuals could not be distinguished in some

species of aquatic plants, particularly those exhibiting

a stoloniferous habit. As the relative percentage

covers were regarded as more important they were

accordingly weighted appropriately.

Two separate PCA were conducted: one using

seven species occurring in 20, that is, 30 per cent or

more of the stands, and another using 12 species

occurring in 12, that is, 20 per cent or more of the

59 stands. In each case, an analysis was performed

utilizing untreated input data and again after nor-

malizing the input data to a standard deviation of

one and a mean of zero.

The behaviour of each species and selected environ-

mental factors were obtained by, respectively, plotting

their importance values and unmodified quantitative

data within the various 3-dimensional stand ordinations

constructed on information obtained in each PCA.

The most interpretable ordination of stands was

obtained in a PCA conducted on seven species

utilizing unnormalized input data.

RESULTS

Braun-Blanquet Analysis

Two basic communities labelled Cx and C2 were

delineated (Table 1).

Community C4 contained predominantly submerged

aquatic species, for example, Potamogeton crispus ,

Najas pectinata and Nitella tenella while, apart from

Ceratophyllum demersum, Community C2 contained

a predominance of floating and rooted-floating

aquatic species, for example Trapa bispinosa , Ludwigia

stolonifera and Echinochloa pyramidalis.

Species belonging to Community Q were usually

found in more exposed areas and in deeper and

more alkaline water than those of Community C2.

For example all stands in Community C4 were

assigned exposure ratings of 4 and their water pH

was generally above 8,4, with an average value of

8,6. Stands found in Community C2, on the other

hand, had exposure ratings ranging between 1 and 3

while their pH values were generally below 8,2 with

an average value of 7,5. The water depth ranged

from 5 to 100 cm in stands of both Communities;

however, the average water depth in Community C

188

CLASSIFICATION AND ORDINATION OF AQUATIC MACROPHYTES

was 54 cm while in C2 it was 31 cm. There was

considerable overlap in the other environmental

factors recorded from stands of both communities.

Community Cx was further subdivided into two

sub-communities: a negatively defined bare Sub-

community labelled X; and a Sub-community Y

containing a number of rooted submerged or partially

submerged marginal species, such as Rotala tenella ,

Phyla nodiflora, Limosella capensis and an unidentified

Cyperus sp. These species were generally found in

stands with shallower water whose depth ranged from

5 to 60 cm with an average value of 31 cm. In Sub-

community X, the stands were found on an average

in deeper water ranging from 5 to 100 cm, with an

average value of 71 cm.

Community C2, like Cx, was further subdivided,

but into three sub-communities: a bare Sub-com-

munity Zx; a Sub-community Yx characterized by

the species Trapa bispinosa and Ceratophyllum demer-

sum ; and a Sub-community Xx characterized by

Nymphaea capensis, Limosella maior , Ipomoea aquatica

and Scirpus litoralis.

The last mentioned Sub-community X1 was dis-

tinguished from the other two on the basis of the

considerably higher salinity of its water. The Na+

concentration was 10 x higher, that of Ca2+ and Mg2+

5x higher, and that of Cl- 15 to 20 x higher, in its

stands than for those composing sub-communities

Zx and Yx. In addition, the stands of Xl all had

substrates consisting of a firm black calcareous mud

whereas those of Yx and Zx were found only on

soft and firm brown muds.

Yx and Zx did not separate as two very distinct

sub-communities as there was a considerable amount

of overlap in the environmental data relating to

each of their component stands. On an average,

however, stands belonging to Sub-community Zx

occurred in shallower water and in more exposed

areas than those of Yx, these being found in deeper

water and in less exposed areas. For example, stands

belonging to Sub-community Zx were all assigned

exposure ratings of 4 and were found in water ranging

in depth from 5 to 40 cm with an average depth

of 17 cm. On the other hand, stands belonging to

Sub-community Yx had exposure ratings generally

ranging between 1 and 3 and were found in water

ranging in depth from 5 to 100 cm with an average

depth of 45 cm.

Sub-community Yx was further subdivided into

five variants labelled w, x, xx, y and z. Variant w

was bare and negatively defined; x contained the

species Pistia stratiotes, Azolla pinnata var. africana ,

Utricularia gibba subsp. exoleta and Scirpus cubensis;

while z contained Utricularia inflexa var. stellaris,

Wolffia arrhiza and Spirodela polyrhiza. Composi-

tionally, Variant y was considered intermediate

between x and z as it contained species from both,

that is, Utricularia gibba subsp. exoleta and Neptunia

natans from Variant x and Wolffla arhiza from Variant

z. Variant y was, however, distinguished from the

others by the fact that Echinochloa pyramidalis was

replaced by Paspalum vaginatum in its stands. Finally,

the fifth Variant xt was represented by a single stand

(No. 39). It was characterized by the additional

species Marsilea macrocarpa, Typha latifolia subsp.

capensis and Ludwigia octovalvis subsp. sessilifera.

All five variants were separated on the different

Mg2+ concentrations of their stands, which showed

no overlap. For example, the highest Mg2+ concen-

trations were found in Variant xx with 22,5 mg/1

(see x in Table 1), followed by z with a range of

12,5 to 9,0 mg/1, then Variant y with a range of

8,75 to 7,5 mg/1, Variant w with 7,5 to 6,25 mg/1

and, finally, the lowest Variant x with 6,3 to 5,86 mg/1.

In addition, stands composing Variant xx had the

highest specific conductance and Na+, K+, Ca2+ and

Ci~ concentrations (Table 1), followed by stands

composing variant z. The remaining three variants

had similar overlapping values. Moreover, stands

included in Variants xx and z were found on soft

brown muds, those in x and y occurred on firm

brown muds, while those of Variant w were found

on both soft and firm brown muds.

Principal Components Analysis

From the pattern of distribution of importance

values of each species over the stands, five distinct

groups of stands with similar floristic composition,

termed noda and named after the species with the

highest mean importance values in each group, were

delineated. Noda were delimited by delineating those

stands within the 3-dimensional ordination in which

each of the species used in the PCA showed very

little or no overlap with each other (Fig. 6). This

was found in those stands where each species occurred

with an importance value of 30,0 per cent or more.

The following noda were distinguished: the

Potamogeton crispus Nodum, Najas pectinata Nodum,

Ceratophyllum demersum Nodum, Trapa bispinosa

Nodum and Ludwigia stolonifera Nodum. These

noda were found to have meaning in the field where

they could be identified as definite plant communities.

The floristic composition of an average stand

within the areas occupied by each nodum are indicated

in Tables 2 to 6. Mean importance and constancy

values for each species occurring in more than two

of the various stands comprising each nodum are

included. Constancy is the frequency of occurrence

of the species in the stands of the nodum expressed

as a percentage of the total numbers of stands in

nodum.

TABLE 2. — Vegetational characterization of Potamogeton

crispus Nodum (14 stands)

TABLE 3.— Vegetational characterization of Najas pectinata

Nodum (6 stands)

C. F. MUSIL, J. O. GRUNOW AND C. H. BORNMAN

189

Fig. 6. — Delineation of noda within 3-dimensional stand ordination. Each nodum is named after the species with the

highest mean cover and frequency in the stands of the group.

TABLE 4. — Vegetational characterization of Trapa bispinosa

Nodum (6 stands)

TABLE 5. — Vegetational characterization of Ceratophyllum

demersum Nodum (4 stands)

TABLE 6. — Vegetational characterization of Ludwigia stoloni-

fera Nodum (21 stands)

The tables show that apart from certain species,

such as Cynodon dactylon in the Potamogeton crispum

Nodum; Paspalum vaginatum and Wolffia arrhiza in

190

CLASSIFICATION AND ORDINATION OF AQUATIC MACROPHYTES

the Trapa bispinosa Nodum; and Pi si i a stratiotes ,

Azo/la pinnata var. africana, Ipomoea aquatica, Scirpus

cubensis, Nymphaea capensis, Cyperus fastigiatus,

Limose/la maior , Neptunia natans and Polygonum

senegalense forma albotomentosum in the Ludwigia

stolonifera Nodum; no other species were restricted

entirely to the stands comprising each nodum.

Although there was some overlap in species between

the various noda, the relative importance values,

that is, cover and frequency of these species, differed

completely between the noda.

The stands comprising the Najas pectinata and

Potamogeton crispus Noda were characterized by

having high covers and frequencies of submerged

aquatic plants and low covers and frequencies of

floating and rooted-floating aquatic plants. On the

other hand, the stands comprising the Trapa bispinosa

and Ludwigia stolonifera Noda had high covers and

frequencies of floating and rooted-floating aquatic

plants but low covers and frequencies of submerged

aquatic plants. The stands comprising the Cerato-

phyllum demersum Nodum appeared to be somewhat

intermediate in compositon between the Najas

pectinata-Potamogeton crispus Noda and the Trapa

bispinosa-Ludwigia stolonifera Noda, in that they

contained more or less intermediate covers and

frequencies of both submerged and rooted-floating

and floating aquatic plants.

The relation of the various noda to site factors is

adequately summarized in Table 7. It is evident

that the submerged aquatic communities were

simple, with few species, and had a competitive

advantage in the exposed areas and alkaline waters.

The floating and rooted-floating aquatic communities,

on the other hand, were more complex and dominated

the protected areas and less alkaline or neutral

water.

TABLE 7.— Environmental and other factors characterizing

noda on the Pongola pans

CONCLUSIONS

The Braun-Blanquet and PCA approaches showed

that the aquatic vegetation in the pans of the Pongolo

River flood plain could be divided into two groups

of growth forms, the submerged and the floating

and/or rooted-floating aquatic plants, each showing

a distribution influenced primarily by physical habitat

factors. The submerged aquatics succeeded best in

the exposed areas and deeper and more alkaline

waters, whereas the rooted and/or floating aquatics

dominated the protected areas and the shallower and

less alkaline waters. This is in agreement with Segal

(1966), who considered the growth forms (that is,

morphological types adapted to their environment)

to be principally or completely determined by the

physical environment.

In the PCA each of the above groups was re-

presented by 2 stand noda with one somewhat

intermediate nodum. Each nodum was considered

a real group as it had meaning in the field and could

be identified with a definite aquatic plant community

observed there.

The Braun-Blanquet analysis showed that the

rooted and/or floating aquatic plant species could be

further subdivided into a number of different groups,

whose stands differed in the chemical composition of

their waters and other factors. This, also, is in

agreement with Segal (1966), who considered that

the occurrence of different taxa were determined by

other, for example, chemical factors.

ACKNOWLEDGEMENTS

It is a pleasure to acknowledge the advice and

assistance of various members of staff of the Depart-

ment of Botany of the University of Natal, the

Botanical Research Institute, and the Natal Parks,

Game and Fish Preservation Board. This paper

presents part of a thesis submitted for the degree

of M.Sc. at the University of Natal. This work was

supported in part by the Department of Agricultural

Technical Services and in part by a grant from the

C.S.I.R.

REFERENCES

Coke, M., 1970. The status of the Pongolo flood plain pans.

Natal Parks Board Report.

Coke, M., 1970. The status of the Pongolo flood plain pans

and the effects of a 1 000 cusec flood in December, 1970.

Natal Parks Board Report.

Curtis, J. T. & McIntosch, 1951. An upland forest con-

tinuum in the prairie forest border region of Wisconsin.

Ecology 32: 476-496.

Du Preez, J. W., 1967. The ecology and geohydrology of the

Makatini Flats with special reference to the problems

of salinization and water-logging. Dept. Mines Geol.

Survey Report.

Goltermann, H. L., 1969. Methods for chemical analysis

of fresh waters. I.B.P. Handbook No. 8 Oxford: Blackwell

Scientific Publications.

Inversen, J., 1936. Biologische Pflanzentypen als Hilfsmittel

in der Vegetationforschung. Kopenhagen.

Kendall, M. G., 1957. A course in multivariate analysis.

London: Griffin.

Kuchler, A. W., 1960. Vegetation mapping. New York:

Ronald Press.

Muller, R. & Wideman, O., 1957. Bestimmung des Nitrat-

ions in Wasser. Wasser 22: 247.

Raunkiaf.r, C., 1934. The life forms of plants and statistical

plant geography. Oxford : Clarendon Press.

Seal, H., 1964. Multivariate statistical analysis for biologists.

London: Methuen.

Segal, S.. 1966. Oecologie van hogere water planttn. Vakbl.

Biol. 46: 138-149.

Tinley, K. L., 1958. A preliminary report of the ecology of

the Pongolo and Mkuze flood plains with particular reference

to hippopotamus. Natal Parks Board Report.

Bolhalia II, 1 & 2: 191-201 (1973)

Notes on the genus Cussonia in South Africa

R. G. STREY*

ABSTRACT

Information gathered during studies of the genus Cussonia under field conditions is presented.

The development of the inflorescences, and branch and leaf initiation, is given special attention. The

subdivision of the genus Cussonia is discussed and the following new subgeneric taxa are created:

Subgenus Cussonia (type species: C. spicata Thunb.) with two sections: Sect. Cussonia and Sect.

Capitata Strey,- Subgenus Panicu/ata Strey (type species: C. paniculata Eckl. & Zeyh.); Subgenus

Protocussonia (type species: C. natalensis Sond.). It is pointed out that the names C. kraussii Hochst.

and C. chartacea Schinz have been misapplied to two undescribed species, and are synonyms of C.

spicata Thunb. and Schefflem umbellifera (Sond.) Baill. , respectively. Four new species are described

viz. C. arenicola Strey, C. zuluensis Strey, C. nicholsonii Strey and C. sphaerocephala Strey.

INTRODUCTION

In the course of studies of the genus Cussonia a

number of interesting facts emerged, especially from

the field studies that were undertaken. Several of

these features, particularly the development of the

inflorescence, the mode of flowering and the renewal

of vegetative growth after flowering, have apparently

not been recorded before and are therefore discussed

in the following pages. It was also discovered that

two names had been misapplied and that the species

masquerading under these names were new and

required to be described. In addition, two new species

were identified and are described here. I am indebted

to Dr D. J. B. Killick for the Latin descriptions.

The majority of specimens studied are deposited in

the National Herbarium, Pretoria, and Natal Her-

barium, Durban. Those seen in other herbaria are

specially indicated.

FIELD STUDIES

The development of the inflorescence

In all species of Cussonia the structures bearing

the flowers are borne in clusters at the apices of

branches or in some cases even at the apices of

fairly thick trunks. These clusters usually consist of

a few to many peduncled spikes, racemes or even

panicles originating in the axils of the apical bracts

which are numerous and in dormant branches com-

pletely cover the apices. At the beginning of a new

season when the trees start sprouting it is soon evident

that branch apices either produce a new flush of

leaves or somewhat more slowly develop inflorescen-

ces. Leaves and inflorescences are seldom, if ever,

simultaneously produced at the apex of the same

branch. Whereas branches producing inflorescences

may be leafy, these leaves are almost invariably

produced during the previous season and are usually

shed before the inflorescences are mature. Mature

and particularly fruiting inflorescences are therefore

borne at the apices of bare branches. In all the species

studied, the clusters of pedunculate flower-bearing

branches together form an apical compound in-

florescence, since no normal leaves are produced

‘Botanical Research Unit, Botanic Gardens Road, Durban.

by the bract-covered apices from which the in-

florescence-branches develop. The development of

the inflorescence of two of the species studied differs

fundamentally from the pattern displayed by the

others. In most species the inflorescence bud initials

are laid down in the bract-covered apices of some

of the branches towards the end of the growing

season and develop from these usually in the following

spring. The situation in Cussonia spicata and C.

sphaerocephala is quite different. Instead of producing

inflorescences, an “umber’ of fairly slender leafy

branches is produced at the apex of a trunk during

the first season. Each branch has a knob-like thick-

ening covered with bracts, at the apex. Only during

the following season are inflorescences produced at

the apices of these branches. Since the leafy branches

formed during the first season lose their year-old

leaves, as soon as the inflorescences start developing,

the final structure resembles a compound “umbel”

at the end of the second season. These pseudo-

inflorescences are therefore biennial structures in

contrast to the inflorescences produced by other

species which develop in one year.

The flowers and fruits

The fully-developed flowerbuds open up in irregular

sequence on almost any parts of the spikes or racemes,

at the base or the apex or at the middle, and in

addition one side of the spike may be in full flower

when no flowers have opened on the other side.

Flowers at various stages of development are there-

fore found irregularly grouped in the inflorescences.

The open flowers are star-shaped with the petals and

stamens stiffly spreading. The petals and stamens

are very easily detached and are never preserved

in situ in dried specimens.

Newly opened flowers are greenish-cream to

butter-yellow in colour, the top of the ovary green,

the stigmas pale and erect and the anthers yellow.

As the flower ages the petals and stamens drop, the

top of the ovary becomes yellow and exudes a shiny,

sticky fluid which attracts insects, the stigmas become

darker and the styles recurve. It is probably at this

stage that pollination is effected by the various types

of insects found on the flowers, including bees, flies

and beetles. Not all ovaries are fertilized. The ones

192

NOTES ON THE GENUS CUSSONIA IN SOUTH AFRICA

that are fertile increase in size and develop a fleshy

purple exocarp attractive to birds. Each fruit con-

tains one to two seeds. Unfertilized and parasitized

ovaries may develop but remain dry and hollow

structures. Mature inflorescences are attacked by

insects and the stem-apex in particular seems sus-

ceptible. When the stem-apex has decayed, the whole

inflorescence breaks off and drops to the ground.

The formation of new branches

The apices of the stems which have fruited and

have subsequently lost the inflorescences, are now

usually more or less truncate and dry out, or develop

a certain amount of callus-tissue. These truncated

branches and trunks eventually produce side-shoots

which continue the growth of the branch. It appears

that side-shoots are seldom formed on actively growing

vegetative branches and trunks in the sparingly

branched species such as C. spicata and C. sphaero-

cephala, but that usually one, rarely more, side-

branches are formed after the growth of a trunk or

branch has been temporarily halted by production

of an inflorescence or by injury. In species which

by nature have a more branched habit, such as

C. natalensis, several shoots originate from dormant

buds after the inflorescences have been discarded.

Leaf -form and development

The leaves, which in all species are clustered at

the ends of branches, are, in most species, produced

in flushes. The young leaves mature and usually

remain on the tree for a full year. Old leaves lower

on the stems are gradually cast off after a new flush

of leaves has appeared so that the clusters are always

situated apically. Usually only one flush of leaves per

year is produced. Renewal of leaves, as also the

production of inflorescences, is not only seasonally

determined and may vary from tree to tree.

In shape the leaves vary from simple and deeply

lobed to singly or doubly compound. The leaflets

of the compound leaf are arranged digitately (like the

spokes of a wheel) round the apex of the peduncle.

They spread in one plane and at an angle to the

peduncle forming a fan-like structure, sub-circular

in outline. The individual leaflets are vertebrate,

a term first used by Mirbel (1815, p. 655) for leaves

of Cussonia spicata. These leaflets resemble pinnae

with the rachillas very broadly winged. The leaf

therefore is a doubly compound structure with the

first division digitately compound the second verte-

brate or rarely doubly vertebrate. The leaflets are

arranged in pairs, lower pairs being the least de-

veloped and the single terminal leaflet showing the

Fig. I. -Semi-diagrammatic drawing of a leaf of Cussonia spicata to illustrate the terminology used in this paper.

R. G. STREY

193

most elaborate development. The vertebrate leaflets

have 1-5 articulations, characterized by turgid green

papillate scales on the lower suface. These scales

may represent reduced bracts. At these articulations

the pinnules are attached. The articulations may

be devoid of pinnules, or bear two, four or six pinnules

arranged in a whorl, the numbers increasing pro-

gressively upwards. The terminal pinnules of the

vertebrae usually repeat the form of the lowest

leaflet, if this is a simple structure. Because of the

size of the compound leaves, only single leaflets are

usually represented in herbaria and the leaf-structure

is usually not fully understood. In Fig. 1 a complete

leaf of C. spicata is semi-diagrammatically represented.

This figure also serves as a reference to the terminology

used in this paper.

Bark, wood and roots

The wood is soft, light, very coarse and fibrous

and the branches have a distinct pith. The bark of

most species is thick and corky and usually fissured,

but it is thinner and less corky and more smooth

in forest trees. The roots of all species investigated

are thick, fleshy and frequently tuberous. Depending

on the species the roots may form a single large

tuber, may be thick and fleshy or produce a series

of tuberous swellings along their length. Such roots

are often used as a source of water by animals and

also by man.

CLASSIFICATION AND SUBDIVISIONS OF THE GENUS

CUSSONIA

The family Araliaceae to which Cussonia belongs

is represented in South Africa by three genera.

Seemannaralia gerrardii (Seemann) Vig. and Schef-

flera umbellifera (Sond.) Baill. were formerly both

included in Cussonia but now are the sole represen-

tatives in South Africa of the genera to which they

have been transferred. Seemannaralia (a monotypic

genus) is not closely related to Cussonia and needs

further study to confirm its inclusion in the Araliaceae.

Schefflera umbellifera has a much closer relationship

to Cussonia. It was recently removed to Neocussonia

by Hutchinson (1967, p. 79). I prefer to follow

Bernardi (1969, p. 92) who confirmed that it and

other related species, should be placed in the genus

Schefflera as was done by Baillon as early as 1878.

When the South African species of Cussonia are

considered it is soon evident that they do not form

a homogeneous assemblage of species, but may be

divided into several groups of closely related species.

Hutchinson selected C. spicata Thunb. as the lecto-

type species of the genus and his typification is here

accepted. The species most closely related to C.

spicata is C. sphaerocephala Strey which is newly

described in this paper. Therefore these two species

represent typical Cussonia.

The development of the inflorescences in the latter

two species is described in detail earlier in this paper

and distinguishes them from the following species

which, in other respects are closely related to them,

namely, C. thyrsiflora Thunb., C. arenicola Strey,

C. nicholsonii Strey and C. zuluensis Strey. All the

previously mentioned species have the following

characteristics in common: they bear fairly large

pedunculate spikes or racemes in terminal compound

inflorescences on relatively sparsely branched plants;

their leaves are mostly doubly compound and the

leaflets of the vertebrate type, except for C. thyrsiflora

which usually has simple leaflets.

A second distinct group is represented by only

one species, namely, C. paniculata Eckl. & Zeyh.

which may be distinguished from all other species

by the paniculate branches of the inflorescence and

the elongate simple, shallowly to deeply lobed leaflets

which are never vertebrate.

The third group is composed of species which

have palmatifid, palmatisect or digitately com-

pound leaves. There seems to be a somewhat gradual

transition from palmatilobed— palmatifid— palmati-

sect— digitately compound leaves (see also Aitken,

1923, p. 58), and a further division of the species

into two groups, one possessing simple lobed leaves

(e.g. C. natalensis) and a group possessing compound

palmatisect or digitate leaves was contemplated

but abandoned.

The following subgenera and sections are described

to accommodate the South African species of the

genus. It seems likely that most of the tropical species

can, without much difficulty, be accommodated in

these subdivisions.

Cussonia Thunb. in Nova Acta R. Soc. Sci. 3:

210-213 (1780); Harms in Pflanzenfam. 3, 8: 1

(1894) pro parte, excl. Sect. Neocussonia Harms;

Phill. Gen. Ed 2: 545 (1951); Hutch. Gen. FI. PI.

2: 57 (1967). Type species: C. spicata Thunb.

Subgen. Cussonia

Trees or shrubs usually only sparsely branched,

roots swollen, fleshy, leaves palmatisect or twice

compound with the leaflets vertebrate in arrange-

ment; seeds plano-convex.

Sect. Cussonia

Trees, terminal pseudoinflorescence a double umbel

consisting of a number of pedunculate spikes arranged

at the apex of bare, somewhat flattened vegetative

branches formed the previous season.

Species included: C. spicata Thunb., C. sphaero-

cephala Strey.

Sect. Capitata Strey, sect. nov.

Arbores vel frutices; inflorescentia terminalis um-

bellata ex spicis vel racemis pluribus pedunculatis

ex apice bracteato caulium exorientibus constans.

Type species: C. thyrsiflora Thunb.

Other species included: C. arenicola Strey, C.

nicholsonii Strey, C. zuluensis Strey.

Subgen. Paniculata Strey, subgen. nov.

Arbores; folia digitato-composita marginibus folio-

lorum non profunde ad profunde sed non ad rhachim

lobatis; inflorescentia terminalis umbellata ex spicis

pluribus pedunculatis ex apice bracteato caulium

exorientibus constans.

Type species: C. paniculata Eckl. & Zeyh.

This subgenus is monotypic.

Subgen. Protocussonia Strey, subgen. nov.

Arbores; folia simplicia palmato-lobata (vel digi-

tato-composita) foliolis sessilibus vel petiolatis;

inflorescentia ex spicis vel racemis pluribus sessilibus

vel breviter pedunculatis in pseudo-umbella terminali

ad apicem ramulorum dispositis constans; semina

subglobosa.

Type species: C. natalensis Sond.

Other species included: C. natalensis Sond., C.

arborea Hochst. ex A. Rich., C. ango/ensis Hiern,

C. holstii Harms ex Engl, and others.

194

NOTES ON THE GENUS CUSSONIA IN SOUTH AFRICA

MISAPPLIED NAMES

During the investigation it was found that two

names of long standing were misapplied and should

be relegated to synonymy. The specimens to which

these two names were applied, represent three distinct

undescribed species.

Cussonia kraussii Hoc/ist. in Flora 27: 431

(1844).

The material on which the description of this

species was based was collected by Krauss at the

Geelbeck's River near George. The Geelbeck's River

is today called Geelbeck’s Vlei as it is no longer

the perennial stream of former times. It is a tributary

of the Klein Brak River situated between Mosselbay

and George. Only Cussonia spicata is found in this

region and an investigation of the type of C.

kraussii , which we had on loan from Zurich (photo,

PRE), confirmed its identity with this species. It is

therefore a synonym of C. spicata Thunb. In South

African herbaria the specimens placed under C.

kraussii were all from Natal, an error probably caused

by the erroneous citation of a single specimen col-

lected by Gueinzius at Port Natal in the Flora Capen-

sis (1862, p. 569) as C. kraussii. An examination of

these specimens revealed that two distinct species

were represented in the material. These two species

are here described as C. nicholsonii Strey and C.

zuluensis Strey.

Cussonia chartacea Schinz in Bull. Herb. Boiss.

2: 211 (1894).

An examination of the type, Rehmann 8096 from

Umbilo, Natal, showed that it represents a juvenile

leaf of Schefflera umbellifera (Sond.) Baill. of which

Cussonia umbellifera Sond. is the basionym. The

name was, however, probably due to the lack of

floral parts, misapplied by Harms (1921, p. 784)

and by Engler. Engler stated that he collected it

himself “urn Kearney bei Stander in Natal” an

obvious error for “Kearsney near Stanger”, and

applied the name to the tall, little branched species,

with globose apical clusters of leaves, fairly com-

monly found in Natal forests. Specimens of this

species are as a result frequently found under the

name C. chartacea in South African herbaria, but

in fact represents an underscribed species closely

related to C. spicata, here described as C. sphaero-

cephala Strey.

NEW SPECIES OF CUSSONIA

Cussonia arenicola Strey, sp. nov., C. thyrsiflorae

Thunb. affinis, sed caulibus gracilioribus unicis

erectis ex basi tuberosa unica turbinata ovoidea vel

globosa exorientibus, foliolis vertebratis, floribus

fructibusque parvioribus cupiformibus differt.

Fruticulus ligneus caule unico ex basi tuberosa

unica turbinata ovoidea vel globosa exorienti. Folia

digitato-composita ad apices ramorum exorientia;

stipulae c. 2 mm longae, interpetiolares basi junctae

ad petiolum adnatac; foliola (4)— 7— ( 1 2), vertebrata.

6-18 cm tonga; petioli brevissimi; rhachilla late

lobata, lobis obtriangularibus vel obhastatis, pinnulis

sessilibus ovatis oblanceolatis obovatis vel trullatis

marginibus serratis apice mucronato. Inflorescentia

ex racemis pedunculatis constans; racemi caulium

apice bracteato clavato umbellati. Flores breviter

pedicellati; bracteae unicae, subulatae. Calyx oram

lobis 5 minutis redactus. Petala 5, ovata, valvata.

Stamina 5, caduca; antherae ovoideae versatiles.

Ovarium 2-loculare; stylus brevissimus, stigmatibus

2 patentibus persistentibus. Fig. 2.

Type: Natal, 2632 CD (Bella Vista): at Makane,

Nov. 1970, Strey 10283 (PRE, holo.; NH).

Single-stemmed shrubs, 1-2 m high; stems 1-2

cm thick, arising singly from a globose, ovoid or

turbinate tuber; basal tubers up to 14 cm broad

and up to 25 cm long often several spaced along a

single root; bark smooth greenish or grey. Leaves

glabrous or with scattered scale-like papillae at the

articulations, twice compound, first division digitate,

bearing vertebrate leaflets; petiole ribbed, terete, up

to 25 cm long, 2-4 mm thick; stipules intrapetiolar,

joined at the base, adnate to the base of the petiole,

lobes 2-3 mm long; limb of the leaf sub-circular in

outline, about 20 cm in diam.; leaflets 4-7(12) per

leaf, 6-18 cm long, chartaceous, dark green above,

dull green beneath, margin revolute, serrate, 1-3

times vertebrate, petiolules about 1(2) cm long,

usually narrowly winged, rhachilla wings obtrian-

gular to obhastate; pinnules up to 9 cm long and

2-4 cm wide, sessile, 1-5, obovate-oblanceolate, or

trullate, base cuneate, rounded, acute or apiculate.

Inflorescence a terminal umbel consisting of (5)8-

15(23) pedunculate, dense, cylindrical racemes ar-

ranged on the often distinctly swollen, bracteate,

apex of the stems; peduncles of the racemes 3-18 cm

long, ( 1 )— 2— 4 mm thick; bracts subtending the

peduncles binate, up to 1 cm long; racemes 3-9 cm

long, 8-14 mm broad. Pedicels 1-2 mm long; bracts

subtending flowers subulate, 2 mm long. Flowers

loosely spirally arranged, white to greenish cream.

Calyx reduced to a 4-toothed rim. Petals broadly

ovate, caducous. Stamens about as long as the petals,

caducous; anthers ovoid, introrse, versatile. Ovary

inferior, 2-loculed; disc conical; style about 1 mm

long, stigmas 2, divergent. Fruit 2 (1) seeded, barrel-

shaped, about 4x4 mm, exocarp mauve and fleshy

when ripe, calyx rim and styles persistent. Seed globose

to subglobose, about 3 mm long; raphe with ovate

areole; endosperm ruminate.

A species mainly found on the coastal sand-dunes

of northern Natal but occasionally further inland in

sand forest (Fig. 3). The specific epithet arenicola

meaning “dwelling on sand” was applied to this

species, because it is the only species which is confined

to sandy habitats.

Mozambique. — 2532 (Lourenco Marques): (-DC), Borle

157; Lourenco Marques (-DC), Borle 204; Polana (-DC),

Borle 553; Santaca (-DD), Gomes e Souza; Inhaca Island

(-DD), Moss J. 28163; Breyer 20507; Mogg 28013; Mogg

26732; Mauve & Verdoorn 88; Moss J. 27535 & 28164.

Natal.— 2632 (Bella Vista): (-CC), Moll 4372A (NH, K);

Ndumu Game Reserve (-CC), De Moor 24; Makane Location

(-CD), Strey 9810 & 10283 (NH, K); do., Ross 2369 (NH);

do.. Mol! 5008 (NH, K); Kosi Bay (-DD), Strey & Moll 3845

& 3848 (NH); do., Ross & Moll 5100 (NH, K); do., Strey

10326, 10434, 10435 & 10436 (NH, K). 2732 (Ubornbo):

Manzengwenya (-BA), Ross & Moll 5103 & 5076 (NH, K);

do., Strey 10453 (NH, K); Sibaya (-BC), Sibaya Project 35/

(NH); do., ( BC), Strey 10300 (NH, K); Sordwana Bay (-DA),

Strey & Moll 3954 (NH, K); South of Sordwana Bay (-DA),

Strey 10465 & 10466 (NH, K.). 2832 (Mtubatuba): Bangazi

Pan, Cape Vidal (-BA), Edwards 4288 (NH, K).

For differences from C. zuluensis see notes under

that species.

R. G. STREY

195

Fig. 2.— 1, Cussonia arenicola , inflorescence X

i ( Strey 10283); 2, seed, front view x li;

2a, seed, back view x 1^; 2b, enlargement

of surface of seed x 5; 3, fruit X 1£; 4,

petal, front view x 3 ; 4a, petal, side view

X 3; 5, stamen x 3; 5a, anther x 4 \\

6, flower bud and bract x 3; 7, flower;

8, longitudinal section of flower at anthesis

x 3; 9, longitudinal section of mature

flower x 3; 10, cross section of raceme;

lOa-lOb longitudinal sections of inflore-

scence X 1£.

Cussonia zuluensis Strey, sp. nov., C. thyrsiflorae

Thunb. affinis, sed arboribus parvis, foliis bicom-

positis foliolis vertebratis 8-12 differt.

Arbores parvae multicaulescentes ad 4 m altae;

radices tumoribus fusiformibus. Folia in fasciculis

terminalibus disposita, bicomposita, divisione prima

digitata; foliola 1 vel 2-vertebrata. Inflorescentia ex

fasciculo terminali racemis pedunculatis constans;

racemi 8-26, caulium apice tumido bracteato sub-

umbellati; pedunculi 20-35 cm longi, partibus

floriferis teretis 20-30 cm longis 3-5 cm diam. Flores

pedicellati, spiraliter dispositi; pedicelli 1-3 cm longi,

bracteis subulatis subtentis. Calyx minute 5-dentatus.

Petala 5, valvata. Stamina 5. petalis alternantia.

Fructus poculiformes; exocarpium carnosum. Semen

1, albumine ruminato. Fig. 4.

Type: Natal, 283 IBB (Nkandhla): Mbabasi River,

Strey 9822 (PRE, holo.; NH, K).

Small trees with several stems, up to 4 m high,

trunks 2-5 cm thick, rarely branched; wood coarse,

long-fibrous, pithed; bark smooth, flaking, grey-

green; roots tuberous, with fleshy fusiform swellings.

Leaves twice (rarely thrice) compound, first division

digitate, bearing vertebrate leaflets ; petiole terete,

20-30 cm long, (2)3-3, 5(4) mm thick ribbed, glabrous.

7ig. 3. — Distribution of Cussonia arenicola.

5292-13

196

NOTES ON THE GENUS CUSSONIA IN SOUTH AFRICA

stipules intrapetiolar, adnate to the base of the

petiole and joined to each other at the base, up to

4 cm long, glabrous or hairy; leaflets spreading at

an angle to the petiole, the whole sub-circular in

outline, up to 30 cm in diam.; leaflets 8-12 per leaf,

vertebrate or twice or thrice vertebrate (rarely simple)

chartaceous, thickly coriaceous when mature, glossy

green above, dark green beneath; 15-25 cm long,

with 1-5 articulations, scaly-papillate at the arti-

culations, petioles 5-10 cm long, occasionally nar-

rowly winged, rhachilla wings obtriangular to ob-

hastate; pinnules 3-10 cm long, lanceolate, broadly

ovate, obovate or oblong, base cuneate, margins

sparsely to distinctly serrate, apex mucronate. In-

florescence a terminal umbel, consisting of 8-26

pedunculate fairly dense, cylindrical racemes arranged

on much abbreviated club-shaped, bracteate apices

of the stems; peduncles of racemes 20-35 cm long;

bracts subtending the peduncles joined at the base

and adnate to the peduncle, bilobed, up to 3 cm

long, often hairy, carinate, caducous; racemes 20-30

cm long, 3-5 cm in diameter. Pedicels slender, 0,7-3

cm long; bracts subulate, 5-8 mm long. Flowers

loosely spirally arranged fairly densely in the racemes,

but not compacted, buds sphaeroid 5-10 mm long;

petals in bud about equal or slightly longer than

the ovary. Calyx reduced to a 5-toothed rim. Petals

5, deltoid, greenish white, acute, spreading and

starlike when open, caducous. Stamens 5, spreading-

erect alternating with the petals; filaments as long

as the petals, white; anthers oblong, introrse, ver-

satile, caducous. Ovary 2-loculed, broadly oblong

3-4 mm long and 2,5-3 mm wide; disc conical,

raised, ribbed; style 2, spreading, eventually recurved.

Fruits l(2)-seeded, 5-8 mm long and 3-5 mm wide,

goblet-shaped, exocarp mauve, fleshy when ripe,

calyx and styles persistent. Seeds triangular in outline, l

planoconvex, rugose, crested in upper third; raphe

ending in an ovate areole, endosperm ruminate.

A species found in the coastal areas from Durban

to Mozambique (Fig. 5). It inhabits dry scrub and .

open dry forest, often found in the river valleys

and occurs in the thornveld as far inland as Weenen 1

in Natal. Specimens from riverine bush have larger

leaves which are less divided and greener.

Fig. 4. — 1, Cussonia zuluensis, inflorescence I

reduced to X 1/6; 2, seed front view X

2a, seed, back view X 2b seed surface

X 5; 3, petal, front view X 1£; 3a, petal,

side view x 1^; 4, stamen X 1|; 4a, anther

x 5; 5 flower bud and bract x 1£; 6,

flower at anthesis X li; 7, longitudinal i;;

section of flower at anthesis x 1£; 8,

longitudinal section of mature flower X

1£; 9, 10 longitudinal sections of the PS

racemes to show arrangement of flowers

x 1 I

9

R. G. STREY

197

Swaziland. — 2631 (Mbabane): Manzini, (-DA), Compton

32011.

Mozambique. — 2632 (Bella Vista): Catembe, (-BA), Borle

113 & 496; Ma/>uta, (-DC), Hornby 2669.

Natal. — 2731 (Louwsburg): A.T.S. Farm, Pongola (-BD),

Moll 5414 , 5478 & 5479 (NH, K); Rooirand (-DB), Strey

9821 (NH, K); Rooirand, Mkuzi-Nongoma road (-DB), Strey

lc467 (NH, K); Mkuzi-Nongoma main road 12 miles from

Mkuzi (-DB), Bower 4; 10 miles from Mkuzi Railway Station

on Mkuzi-Nongoma main road (-DB), Bower 1; 2732 (Ubom-

bo): False Bay (-CD), Gerstner (NH); Gwaliweni forest

(-AC), Moll 4486 (NH); 2830 (Dundee): Weenen (-CC), Pentz

2831 Nkandhla (-DB), Codd 1418; Inadi River and Tugela River

junction north of Greytown (-DC), Dyer 4387 I; Enseleni,

Richards Bay (-CA), Strey 9914 (NH, K) ; Inadi River and Tugela

River junction north of Greytown (-DC), Dyer 4387 II (NH);

2930 (Pietermaritzburg); Edge of Mooi River near Greytown

(-AB), Acocks 11587 (NH); 3 miles from Mandini off Tugela

Mouth road, Mtunzini district (-AB), Edwards 1626; Town

Hill (-CB), Bayer s.n. (NH); 2931 (Stanger): Tugela valley,

Mapumulo (-AD), West 1878 (NH); Lower Tugela (-AA),

Stewart 42 (NH); Lower Tugela (-AA), Edwards 3138 I & //

(NH); Nonoti (-AD), Wood 3867 (NH); 2831 (Nkandhla):

White Umfolosi bridge (-AA), Strev 10492 (NH, K); Mbabazi

River (-AB), Strey 9822 & 9824 (NH, K); Black Umfolosi

near Keza, (-AB), Strey 1944 (NH, K); Mhlabatini (-AB),

Gerstner s.n.; Hill between Mhlabatini and Black Umfolosi

River (-AB), Verdoorn 1726; Hlabisa (-BB), Strey 56166

(NH); Msuzi River banks (-CA), L. E. Codd 1418; I. Scott

Barnes’ farm (-DB), Moll 4962 & 4968 (NH, K); Umlalazi

Nature Reserve (-DB), Moll 4984 (NH, K); Umhlatuzi Valley,

(-DD), J. G. Lawn 524 (NH); 2832 (Mtubatuba): Hlabisa

(-AA), Ward 2171 (NH); Hluhluwe State Dam (-AA), Ward

5615 (NH); Hlabisa, Palm Ridge Farm (-AC), Harrison 449

(NH); Charter’s Creek on Lake St. Lucia (-AB), De Winter

8124; Hluhluwe (-AB), Strey 7355.

Fig. 5. — Distribution of Cussonia zuluensis.

C. zuluensis is usually easily distinguished from

all other South African species of Cussonia by the

length of the pedicels of the flowers and fruits which

are seldom much less than 1 cm in length and usually

much longer. From C. thyrsiflora it may be distin-

guished by the leaves which have (3)5— 7(8) leaflets

which are once or twice vertebrate in contrast to

the usually 4-5 simple (or rarely once vertebrate)

leaflets of C. thyrsiflora. Whereas this species is a

small tree with several erect stems, C. thyrsiflora

is a scandent scrambler with more or less patent

branches.

From C. nicholsonii it differs by the pedicellate

(not sessile) flowers, the less dense inflorescence

and the shape of the fruits which have a rounded,

not cuneate base.

Differs from C. arenicola by the cluster of tubers

producing a number of stems set closely together

and by the differently shaped fruits, those of C.

zuluensis being goblet-shaped and vertically grooved

between the locules.

The name is derived from the area to which it

is mainly confined, namely Zululand in northern

Natal.

Cussonia nicholsonii Strey, sp. nov., C. zuluensi

Strey affinis, sed floribus fructibusque sessilibus

cuneiformibus differt.

Arbores parvae multicaulescentes ad 3 m altae;

radices tumoribus tuberosis. Folia in fasciculis

terminalibus disposita, bis vel ter composita, divisione

prima digitata; foliola 1-3-vertebrata. Inflorescentia

ex fasciculo terminali spicis pedunculatis constans;

spicae 8, caulium apice tumido bracteato subum-

bellatae, 5-12 cm longae, 3-4 cm diam.; pedunculi

tereti, 6-22 cm longi. Flores sessiles vel fere sessiles,

spiraliter disposita, bracteis subulatis subtentis. Calyx

5-dentatus. Petala 5, valvata. Stamina 5, petalis

alternantia. Fructus cuneiformes; exocarpium car-

nosum. Semen 1, albumine ruminato. Fig. 6.

Type: Natal, 3030CB (Port Shepstone): Gibraltar,

Strey 10025 (PRE, holo.; NH, K, NU).

Small trees with several stems, up to 3 m tall,

trunks 3-6 cm thick, rarely branched; wood coarse,

long-fibrous, pithed; bark smooth, flaking, reddish-

grey to grey; roots tuberous with fleshy fusiform

swellings. Leaves twice compound, first division

digitate, bearing vertebrate leaflets; petiole terete,

20-30 cm long, 6 mm thick, ribbed, glabrous; stipules

intrapetiolar, adnate to the base of the petiole and

joined to each other at the base, up to 4 cm long,

hairy; leaflets spreading at an angle to the petiole,

the whole sub-circular in outline, up to 30 cm in

diam.; leaflets (5)— 7— 9— (1 1) per leaf, vertebrate to

thrice vertebrate, rarely simple, chartaceous, thickly

coriaceous when mature, glossy-green above, dark

green beneath, 10-20 cm long, with 1-3 articulations,

papillate-scaly at the articulations, petiolules 3-5

cm long, occasionally winged; rhachilla wings ob-

triangular to obhastate; pinnules up to 10 cm long,

2-4 cm broad, oblong or obovate, broadly ovate,

base cuneate, margins sparsely serrate to distinctly

serrate, rarely entire, mucronate. Inflorescence a

terminal umbel, consisting of 8-30 pedunculate

dense, cylindrical spikes arranged on the much

abbreviated club-shaped, bracteate apex of the stem;

peduncles of spikes 6-22 cm long; bracts subtending

the peduncles joined at the base and adnate to the

peduncle, bilobed, up to 3 cm long, caducous; spikes

5-12 cm long, 3-4 cm in diameter, bracts subulate,

up to 10 mm long. Flowers sessile or subsessile,

spirally arranged, densely compacted, buds slightly

domed; calyx reduced to a 5-toothed rim. Petals 5,

valvate, deltoid-acuminate, yellowish green; caducous.

Stamens 5, alternating with the petals; filaments as

long as the petals; anthers oblong, introrse, versatile,

caducous. Ovary 2-loculed, wedge-shaped, styles

198

NOTES ON THE GENUS CUSSONIA IN SOUTH AFRICA

4 mm long, spreading; stigma folded, introrse. Fruit

10-15 mm long, 8-12 mm wide, wedge-shaped,

angled by pressure, exocarp mauve and fleshy when

ripe, calyx rim and styles persistent, fruit 1 -seeded

by abortion of one ovule. Seeds 5-10 mm long,

obovoid, plano-convex, rugose, crested and beaked;

raphe ending in a round areole; endosperm ruminate.

A species found mainly in the coastal areas of

Natal (Fig. 7). It inhabits a fairly wide range of

ecological habitats from hillsides to Euphorbia scrub

and riverine bush.

Natal. — 2732 (Ubombo): Josini Dam (-CA), Strey &

Moll 3670 (NH). 2930 (Pietermaritzburg): Inanda, Umzimyati

Falls (-DB), Strey 7412 (NH); Inanda (-DB), Moll 2172

(NH); Paradise Valley Nature Reserve (-DD), Mol! 5007

(NH, K); Durban (-DD), Bevis (NH). 2931 (Stanger): Mt.

Edgecombe (-CA), Rault s.rt. (NH); Umgeni, Hutchinson

1840 (NH). 3030 (Port Shepstone): Amanzimtoti (-BB), Ward

5903 I, II, III, (NH); Wood 12306 (NH); Warner Beach (-BB),

Strey 8800 , 9490, 9925 & 10401 (NH, K); Gibraltar (-CB),

Strey 10025 (NH, K); Strey 9570 (NH, K); Mpunzi Bridge

(-CC), Strey 8710 A & B (NH); do., (-CC), Strey 9020a,

9020, 9274a 1C024 & 9710 (NH, K).

C. nicholsonii resembles C. spicata and C. sphaero-

cephala in the spike-like inflorescence and sessile

flowers, but the inflorescence is not a double umbel

as in the latter two species. The sessile flowers and

dense flowering spike distinguish it from all the other

species. The specific epithet was chosen to honour

Mr H. B. Nicholson of Skyline, St. Michaels-on-

Sea, Natal, and to acknowledge his companion-

ship and help on many botanical tours.

Fig. 6.— 1, Cussonia nicholsonii, flowering spike X

1/20; 2, 2a, seed, front view and back view X

2b, surface of seed X 5; 3, 3a, petal, front

view and side view x li; 4, stamen X IT 4a,

anther x 5; 5, flower bud and bract x IT; 6,

flower at anthesis X If; 7, longitudinal section

of flower at anthesis X IT; 8, longitudinal

section of fruit X IT 9, cross section of imma-

ture spike XT 10, 11, longitudinal sectionsof

fruiting and flowering spikes X IT

R. G. STREY

199

Cussonia sphaerocephala Strey, sp. nov., C.

spicatae Thunb. affinis, sed truncis plerumque unicis

relative tenuibus non nisi prope apicem ramosis,

ramo quoque fasciculo sphaerico foliorum terminanti

spicis plerumque brevioribus crassioribus, bracteis,

majoribus, bracteolis acute carinatis, calyx majore

accrescenti, seminibus longioribus planoconvexis

differt.

Arbor usque ad 25 m alta, truncis relative tenuibus

unicis vel parce ramosis. Rami fasciculo sphaerico

foliorum terminantes. Folia composita vel bis com-

posite, divisione prima digitata; foliola vertebrata

vel bis vertebrata; lamina circularis. Pseudoin-

florescentia in anno primo biennis, ex fasciculo ramis

sparse foliosis constans; rami ad apicem incrassatum

caulium portati quisque fasciculo bractearum imbri-

catarum 5-6 cm longarum terminans. Rami primarii

folia in extremo tempi exuti, spicas 4-8 pedunculatas

ex apice bracteato producentes. Spicae partibus

floriferis 8-14 cm longis, 4-6 latis. Flores dense

spiraliter dispositi, sessiles, flavovirides. Bracteae

lineares, chartaceae. Calyx ovarium inferiore adnatus

oram 5-dentatam 2-3 mm altam formans. Petala 5,

valvata, deltoidea. Stamina 5, petalis alternantia;

antherae oblongae. Fig. 8, 10 and 11.

Type: Natal, 2931 (Stanger): Mapumulo, mile-

stone, 3, Dec. 1969 (-AA) Strey 9470 (PRE, holo.,

NH, K).

Evergreen trees, up to 25 m high with a single

bole; trunk wider and fluted at base, sparingly

branched upwards, with a distinct pith; wood coarse,

fibres long; bark somewhat corky, smooth or slightly

fissured, reddish-grey, dark-grey with age; old leaves

well spaced due to elongation of boles, deciduous

shortly after flush of new closely spaced leaves,

which form the spherical terminal clusters, develops

Leaves alternate, twice compound, first division

digitate bearing vertebrate leaflets; petiole terete,

ribbed, glabrous, up to 90 cm long, up to 1 cm thick,

swollen at the base; stipules 2, intrapetiolar, more

or less joined at the base, adnate to the petiole,

approximate or fairly wide apart, broadly-subulate,

coriaceous, up to 3 cm long; outline of lamina

orbicular, up to 40 cm in diam., comprised of 6-12

leaflets; leaflets coriaceous to chartaceous depending

on age; shiny and darkgreen above, darkgreen

beneath, rarely simple, usually vertebrate with 1-5

articulations, 10-35 cm long; vertebrae petioluled

or sessile; petiolules with wings or up to 5 pinnules;

base of petiolule and articulations with scales, scales

rarely absent; pinnules broadly obovate, lanceolate

to lanceolate-oblong, cuneate; base narrowly to

broadly decurrent, apex often mucronate, margin

entire or sparsely to coarsely serrate, rarely sinuate-

serrate; rhachilla wings entire, obovate to obtriangu-

lar, base cuneate, main nerves raised on both surfaces,

secondary nerves immersed above, inconspicuous

beneath. Leaf-scars persistent on young trunks,

disappearing lower down. Pseudo-inflorescence bien-

nial; in the first year consisting of 10-40 spreading

sparsely leafy primary branches produced at the

thickened apex of the stem, each branch up to 60

cm long, 3-7 cm thick, warty, bearing a cluster of

5-6 cm long imbricate and stipular bracts at its

apex. Primary branches casting their leaves in the

second season and develop the true inflorescence at

its apex, consisting of 4-8 piano-terete peduncles

arranged in an umbel, each peduncle bearing a short

thick terete, 8-14 cm long, 4-6 cm thick flowering

spike; bracts at the base of each peduncle 3-5,

triangular in shape. Flowers densely spirally arranged,

sessile, yellowish green. Bracts linear, chartaceous,

ciliate to fimbriate, 6-12 mm long, upper 2-3 mm

caducous; bracteoles carinate, fimbriate, 2-5 mm

long. Calyx rim 2-3 mm high, margin 5-toothed,

fringed. Petals 5, valvate, deltoid, acuminate. Stamens

5, alternating with the petals; anthers oblong, in-

trorse, versatile, spreading, caducous. Ovary inferior,

bilocular; styles persistent-approximate, fleshy; stig-

ma folded, introrse, spreading. Disc flat, rugose.

Fruit up to 3 cm long, 1 cm broad, conical, mauve

and fleshy when fertile; calyx somewhat accrescent;

seed solitary by abortion, 1 cm long, plano-convex,

crested, rugose; raphe ventral; areole linear; testa

brown; endosperm ruminate.

This species occurs from Port St. Johns to Kosi

Bay in coastal dune forest, but is also common in

moist frost-free high forest kloofs up to an altitude

of 1 300 m preferring south-eastern aspects, crowns

often emerging from the surrounding canopy and

very conspicuous (Fig. 9). Not yet recorded outside

of the Transkei and Natal.

Natal. — 2731 (Louwsburg): Ngome (-DD), Strey 10472

(NH, K). 2732 (Ubombo): Mseleni River bridge on Mbaz-

wana-Mseleni Mssion road (-BC), De Winter & Vahrmeijer

8582. 2830 (Dundee): Qudeni Forest (-DB), Talbot; Qudeni

Forest (-DB), For. Dept. Herb. No. 7860. 2831 (Nkandhla):

South of Richards Bay (-DD), Strey 9918 (NH, K); Port

Durnford Native Reserve No. 10 (-DD), Strey 9920 (NH, K).

2832 (Mtubatuba): St. Lucia (-AD), Lansdell 54 (NH); 6 miles

from St. Lucia towards Mtubatuba (-AD), De Winter 8752.

2930 (Pietermaritzburg): Lions River (-AC), Moll 3536 (NH);

Ismont (-DC), Strey 9888 (NH, K); Glenwood, Durban (-DD),

Ward 5896 II, III (NH,); Pigeon Valley, Durban (-DD), Strey

9480 (NH, K); Strey 9681 (NH, K); Strey 9682 (NH, K);

Bulwer Park, Durban (-DD), Strey 9832a, b (NH). 2931

(Stanger): Mapumulo, milestone 3 (-AA), Strey 9470 (NH,

K). 3030 (Port Shepstone): Umzinto (-BC), Bourquin (NH);

Beacon Hill (-CB), Strey 6547 (NH); Mills Farm, Paddock

(-CB), Strey 6928 (NH, K). 3129 (Port St. Johns): Lusikisiki

Manteku (-BD), Strey 10200 (NH, K); Ndindini Forest (-BD),

Strey 10207 (NH, K); Lusikisiki (-BD), Marais 1184 (NH),

Ntafufu (-DA), Strey 8980 (NH, K). 3130 (Port Edward);

Umtamvuna Forest Reserve (-AA), Strey 10406 (NH, K)

200

NOTES ON THE GENUS CUSSONIA IN SOUTH AFRICA

Fig. 8.— 1, Cussonia sphaerocephala, habit;

2, pseudoinflorescence X 1/20; 3, 3a,

seed, front and back view; 3b, surface of

seed enlarged x 5; 4, 4a, petal front and

side view x 1£; 5, stamen x 1|; 6, anther

x 5; 7, flower bud x !£; 8, flower at

anthesis; 9, longitudinal section of flower

at anthesis X 1£; 10, longitudinal section

of mature flower x 1^; 11, buds showing

arrangement in spike x |; 12, flower

viewed from above x f; 13, arrangement

of bracts and bracteoles x If; 14-14b,

bracts, front view x 1|.

R. G. STREY

201

Fig. 9. — Distribution of Cussonia sphaerocephala.

Fig. 11. — Cluster of spikes at the apex of a first-year branch

which has cast its leaves. Note leaf-scar below inflorescence.

C. sphaerocephala has been erroneously referred

to as C. chartacea by several authors and this is

the name it generally bears in herbaria. As explained

earlier in this paper under “Misapplied Names”,

C. chartacea Schinz is a synonym of Schefflera um-

bellifera (Sond.) Baill. The habit and habitat of this

species are described by Aitken (1923, p. 59), who

relates it to C. spicata Thunb. In spite of mentioning

the possibility that it could be a distinct species,

he favoured the view that it was an ecological form

of C. spicata occurring on “slopes with southern

exposure covered with close bush”. He was apparently

not aware of the form of C. spicata which occurs

in forests, since he described the habitat of the latter

as “tree-veld on hillsides with northern exposure”.

The specific epithet was chosen to describe the

striking spherical, terminal leaf-clusters, which

characterize the species.

REFERENCES

Aitken, R. D., 1923. The distribution and ecology of the

genus Cussonia Thunb. with some remarks upon its

probable evolutionary history. Mem. Bot. Surv. S. Afr.

5: 56-70.

Bernardi, L., 1969. Revisio et taxa nova Schefflerarum.

Candollea 24/1 : 89-122.

Hutchinson, J., 1967. Geneta of Flowering Plants 2. Oxford:

Clarendon Press.

Fig. 10. — Cussonia sphaerocephala in clearing in the Dukuduku

Forest, Hlabisa District.

Mirbel, C. F. B. de, 1815. Elemens de Physiologie Vegetale

et Botanique. Paris.

■

'

203

Book Reviews

Excelsa. Journal of the Aloe, Cactus & Succulent Society

of Rhodesia. Ed. M. J. Kimberley. No. 1, December, 1971.

Pp. 77. Price ?

Excelsa is a new journal, being the official organ of the

Aloe, Cactus & Succulent Society of Rhodesia. The name

Excelsa is, of course, derived from Aloe excelsa , Rhodesia’s

tallest and probably best-known aloe. According to the Editor,

Michael Kimberley, “Excelsa” will appeal to all who are

interested in acquiring a wide knowledge and appreciation of

plants that are somewhat loosely referred to as succulents

it is intended to offer something of interest to everyone — not

only to the botanist, both professional and amateur, the

natural historian and the succulent collector, but also to the

ordinary gardener and man in the street, who find an attraction

in all or any of the several genera and many species of the

Liliaceae, the Asclepiadaceae, the Euphorbiaceae, the Cactaceae,

and a host of other botanical families.”

In this first issue the journal amply fulfils these objectives.

Following the introduction is a biographical note by the

editor on the late Dr. G. W. Reynolds, well-known aloe

specialist. Then there is a short article by R. W. S. Turner on

Linnaeus. An interesting article by D. C. H. Plowes on the

stapeliads of Rhodesia is beautifully illustrated by colour

photographs of 36 different species. The photographs are

amongst the best ever seen by the reviewer. M. W. Kamstra

has contributed an article on aloe seeds, while Robert Drum-

mond of the Salisbury Herbarium has provided a provisional

list of Rhodesian succulents excluding those mentioned elsewhere

in the journal. E. J. Bullock writes about the aloes of Matabele-

land, and Professor D. T. Cole about Lithops. An article by

Thomas Baines on the Great Tree of Damara Land, S.W. Africa,

first published in Nature & Art in 1866, is reproduced. The

Editor then discusses existing Rhodesian plant conservation

legislation and the impressive series of articles concludes with

two lists, one of the aloes occurring naturally in Rhodesia

by E. J. Bullock and the other of some of the cycads of the

world by E. A. Pienaar. Finally, there are book reviews, the

Society’s constitution and a list of members.

The standard of production, editing, content and illus-

tration of this first issue is exceptionally high and, if this

standard can be maintained in future issues, this journal will

be one of the best of its kind available to the ever-increasing

number of succulent enthusiasts in Southern Africa.

D. J. B. Killick

Eastern Cape Veld Flowers deur Eily Gledhill. Kaap-

stad: Die Departement van Natuurbewaring van die Kaap-

landse Provinsiale Administrasie . 1971 (?). Bl. 271, 4

kleurplate, 544 lyntekeninge. Prys R2,50.

Eastern Cape Veld Flowers is geskryf as ’n gids om

natuurliefhebbers in Oos-Kaapland vertroud te maak met die

plantegroei van die streek waarin daar volgens die skrywer

tussen 4 000 en 6 000 spesies van blomplante voorkom. Vyf-

honderd vier-en-veertig van hierdie spesies, behorende aan

125 families, word in hierdie boek behandel. Van elke plant

word die wetenskaplike naam en, waar bekend, ook volksname

in Engels, Afrikaans en Xhosa aangegee. ’n Beskrywing van

sowat 50-80 woorde en ’n aanduiding van die blomtyd, habitat

asook die verspreiding van elke plant word gegee. In sommige

gevalle word iets omtrent die gebruik en ander interessanthede

van die plante meegedeel. So word op bl. 92 gemeld dat

Dioscorea elephantipes (Elephant’s Foot) skaars begin word

as gevolg van die feit dat bokke dit graag eet en plantversame-

laars dit “plunder”.

Elke beskrywing is deur die outeur van ’n lyntekening

voorsien. Mev. Noel Urton is verantwoordelik vir vier kleur-

plate waarop elk drie of vier plantsoorte afgebeeld is. Die

544 lyntekeninge is in ’n sekere mate teleurstellend. Die

buitelyne van blare en blomme is dikwels vaag of verdwyn

heeltemal, terwyl skaduweestrepe ook nie altyd na wense

is nie. ’n Voorbeeld hiervan is Hibiscus ludwigii op bl. 165,

plaat 41, no. 6. ’n Aantal plate het te veel afbeeldinge daarop,

met die gevolg dat sommige tekeninge redelik klein is en die

kenmerke van die plante nie duidelik toon nie. Voorbeelde

hiervan is Passerina rigida, Phylica axillaris, Disparago ericoides

en Pentzia globosa.

Daar is opgemerk dat 'n aantal wetenskaplike plantname

gebruik word, wat reeds verouderd is. Sida longipes E. Mey.

word by voorbeeld op bl. 164 as die geldige naam vir S. dregei

Burtt Davy aangehaal. Ander sulke voorbeelde is ( =Sci/la )

Ledebouria, ( = Hartmannia) Oenothera, ( = Osyris abyssinica)

O. lanceolata en (= Lantana salvifolia ) L. rugosa. Outeurs-

afkortings word inkonsekwent gebruik. Die naam Bolus kom

as Bolus., Bol. en Bolus voor, terwyl dit as H. of L. Bol.

behoort aangedui te word. Gilg word soms aangedui as Gil.

of Gilg.; Th. sou beter deur Thunb. vervang word. Punte

word onnodiglik na outeursname soos Lewin, Druce, Vahl

en Don, wat nie afgekort word nie, gesit. ’n Aantal spelfoute

is opgemerk: op bl. 201 word Chascanum as Chaschanum

gespel, terwyl op bl. 168 Dovyalis as Dovialis en op bl. 198

Ipomoea as lpomea voorkom.

Op bl. 29 verskyn ’n sleutel tot die 17 families van die

monokotiele wat in die boek genoem word. Die sleutel is

gebaseer op die vraag-en-antwoordbeginsel. 'n Tweede sleutel

is in vorm van tabelle en behandel slegs dikotiele. Die plante

word volgens hulle groeivorm en grootte in 11 groepe verdeel.

Vir elk van die 11 groepe word ’n tabel gegee en op hierdie

tabelle kan spesies aan die hand van blomkleur en enkele

vegetatiewe kenmerke onderskei word.

Die skrywer gee aan die begin van die boek ’n kort

omskrywing van die deel van die land wat deur die werk

gedek word. Dit word geillustreer deur ’n kaart op bl. 241,

verkeerdelik aangehaal as bl. 235 op bl. 8, asook deur ’n

gedeelte van ’n topografiese kaart van Suidelike Afrika wat,

sowel voor as agter, aan die binnekant van die omslag gevind

word. Vervolgens word iets omtrent die klimaat gese. ’n

Vereenvoudigde, maar baie interessante oorsig, in ongeveer

4 bladsye, oor die geologie van die gebied, sluit die inleidende

deel af. Die volgende deel behandel die verskillende floristiese

streke wat in Oos-Kaapland aangetref word. Die skrywer

bespreek ook die rykdom en die noodsaaklikheid van die

bewaring van die flora.

’n Woordelys waarin ’n aantal plantkundige en ander

terme wat in die teks voorkom, verklaar word, verskyn op bl.

24-28. Dit word deur 11 diagramme op bl. 23 toegelig. Aan

die einde van die boek word ’n indeks tot die wetenskaplike

plantname gevind. Laastens word ’n indeks tot die volksname

gegee wat veral handig te pas sal kom vir diegene wat slegs

met sulke name vertroud is.

Eastern Cape Veld Flowers word in ’n handige formaat,

21 cm x 14 cm, uitgegee. ’n Kleurvoile, treffende en buigbare

omslag is vir die boek gekies. Die prys van die boek is besonder

laag en bied goeie waarde vir u geld. Dit kan beslis as handboek

aanbeveel word vir almal wat die plantegroei van Oos-Kaapland

wil leer ken.

Eizabeth Retief

204

GUIDE FOR AUTHORS

GENERAL

Bothalia is a medium for the publication of botanical

papers dealing with the flora and vegetation of Southern

Africa. Papers submitted for publication in Bothalia should

conform to the general style and layout of recent issues of

the journal (from Vol. 11 onwards) and may be written in

either English or Afrikaans.

TEXT

Manuscripts should be typed, double-spaced on one side

of uniformly-sized A4 paper having at least a margin of 3 cm

all round. Latin names of plants should be underlined to

indicate italics. All other marking of the copy should be left

to the editor. Metric units are to be used throughout. Manus-

cripts should be submitted in duplicate to the Editor, Botha-

lia, Private BagXIOl, Pretoria.

ABSTRACT

A short abstract of 100-200 words preferably in both

English and Afrikaans should be provided. In the abstract

the names of new species and new combinations should not

be underlined.

FIGURES

Black and white drawings, including graphs, should be in

jet-black Indian ink preferably on bristol board. Lines should

be bold enough to stand reduction. Indicate the desired lettering

lightly in pencil: the printer will insert the final lettering. If

authors prefer to do their own lettering, then use some printing

device such as stencilling, letraset, etc. It is recommended

that drawings should be twice the size of the final reduction.

Photographs submitted should be of good quality, glossy,

sharp and of moderate, but not excessive contrast. Photograph

mosaics should be composed by the authors themselves: the

component photographs should be mounted neatly on a

white card base leaving a narrow gap between each print;

number the prints using some printing device.

Figures should be planned to fit, after reduction, into

a width of 8 cm, 11 cm or 17 cm with a maximum vertical

length of 24 cm.

The number of each figure and the author’s name should

be written on the back of the figure using a soft pencil.

Captions for figures should be collected together and typed

on a separate page headed Captions for Figures. A copy of

each caption should be attached to the base of each figure.

Do not underline plant names in captions — only collectors’

names and numbers.

Authors should indicate in pencil in the text where they

would like their illustrations to appear.

TABLES

Tables should be set out on separate sheets and numbered

in Arabic numerals.

CITATION OF SPECIMENS

In citing specimens the grid reference system should be

used (Technical Note: Gen. 4, 4c). Provinces/countries should

be cited in the following order: S.W. Africa, Botswana,

Transvaal, Orange Free State, Swaziland, Natal, Lesotho and

the Cape. Grid references should be cited in numerical sequence.

Locality records for specimens should preferably be given to

within a quarter-degree square. Records from the same one-

degree square are given in alphabetical order i.e. (-AC)

precedes (-AD), etc. Records from the same quarter-degree

square are arranged alphabetically according to the collectors’

names; the quarter degree references must be repeated for

each specimen cited. The following example will explain the

procedure:

Natal. — 2731 (Louwsburg): 16 km E. of Nongoma (-DD),

Pelser 354; near Dwarsrand, Van der Merwe 4789. 2829

(Harrismith): near Groothoek (-AB), Smith 234; Koffiefontein

(-AB), Taylor 723; Cathedral Peak Forest Station (-CC),

Marriott 74; Wilgerfontein, Roux 426. Grid ref. unknown:

Sterkstroom, Strydom 12.

Records from outside Southern Africa should be cited

from north to south i.e. preceding those from Southern Africa.

The abbreviation “distr.” should be added to all district

names e.g.

Kenya. — Nairobi distr.: Nairobi plains beyond race course,

Napier 485.

GIDS VIR SKRYWERS

ALGEMEEN

Bothalia is ’n medium vir die publikasie van plantkundige

artikels wat handel oor die flora van Suidelike Afrika. Artikels

wat voorgele word vir publikasie in Bothalia behoort ooreen

te stem met die algemene styl en rangskikking van onlangse

uitgawes van die tydskrif (vanaf Vol. 11). Dit mag in Engels

of in Afrikaans geskryf word.

TEKS

Manuskripte moet getik wees in dubbelspasiering slegs

op een kant van ewegroot A4 papier, met reg rondom ’n

rand van minstens 3 cm breed. Latynse name van plante moet

onderstreep word om aan te dui dat dit kursief gedruk moet

word. Alle ander merke moet aan die redakteur oorgelaat

word. Metrieke eenhede moet deurgaans gebruik word.

Manuskripte moet in tweevoud ingedien word by die Redak-

teur, Bothalia, Privaatsak X101, Pretoria.

UITTREKSEL

’n Kort uittreksel van 100-200 woorde moet voorsien

word, verkieslik beide in Engels en Afrikaans. In die uittreksel

moet die name van nuwe soorte en nuwe kombinasies nie

onderstreep word nie.

AFBEELDINGS

Wit en swart tekeninge, insluitende grafieke, moet met

pikswart Indiese ink geteken word, verkieslik op “bristol

board”. Lyne moet dik genoeg wees om verklein te kan word.

Dui die verlangde byskrifte liggies in potlood aan: die drukker

sal die uiteindelike byskrifte invoeg. Indien skrywers verkies

om hulle eie byskrifte te maak, gebruik dan een of ander

hulpmiddel soos letraset of ’n sjabloon. Dit is wenslik dat

tekeninge tweemaal so groot as die uiteindelike verkleining

sal wees.

Fotos wat ingedien word, moet van hoe kwaliteit wees,

glansend, skerp en van matige maar nie oordrewe kontras.

Fotomosaieke moet deur die skrywer self saamgestel word:

die afsonderlike fotos moet netjies monteer word op ’n stuk

wit karton met ’n smal strokie tussen die fotos; nommer die

fotos met behulp van een of ander druk-hulpmiddel.

Afbeeldings moet so beplan word dat hulle na verkleining

sal pas in ’n breedte van 8 cm, 11 cm of 17 cm met ’n maksimum

vertikale lengte van 24 cm.

Die nommer van elke afbeelding sowel as die skrywer

se naam moet op die rugkant van die afbeelding geskryf word

met ’n sagte potlood.

Onderskrifte vir afbeeldings moet bymekaar getik word

op 'n afsonderlike bladsy met die opskrif Onderskrifte vir

Afbeeldings. ’n Afskrif van elke onderskrif moet aan die

onderkant van elke afbeelding vasgeheg word. Moenie plant-

name in onderskrifte onderstreep nie, slegs versamelaarsname

en -nommers.

Skrywers moet met potlood in die teks aandui waar hulle

graag hulle afbeeldings wil he.

TABELLE

Tabelle moet op afsonderlike velle papier kom en ge-

nommer word met Arabiese nommers.

SITERING VAN EKSEMPLARE

Wanneer eksemplare siteer word, moet die ruitverwysing-

stelsel gebruik word (Tegniese Nota: Gen. 4, 4c). Provinsies/

lande moet in die volgende volgorde siteer word: Suidwes-

Afrika, Botswana, Transvaal, Oranje-Vrystaat, Swaziland,

Natal, Lesotho en die Kaapprovinsie. Ruitverwysings moet

in numeriese volgorde siteer word. Lokaliteitsrekords vir

eksemplare moet verkieslik tot binne kwartgraadvierkante

gegee word. Rekords uit dieselfde eengraadvierkant word in

alfabetiese volgorde aangebied, nl. (-AC) kom voor (-AD),

ens. Rekords uit dieselfde kwartgraadvierkant word alfabeties

gerangskik volgens die versamelaars se name, en die kwart-

graadverwysings moet herhaal word vir elke eksemplaar wat

siteer word. Die volgende voorbeeld sal die metode verduidelik:

Natal. — 2731 (Louwsburg): 16 km O. van Nongoma

(-DD), Pelser 354; naby Dwarsrand, Van der Merwe 4789.

2829 (Harrismith): naby Groothoek (-AB), Smith 234; Koffie-

fontein (-AB), Taylor 720; Cathedral Peak Bosboustasie

(-CC), Marriott 74; Wilgerfontein, Roux 426; Ruitverwysing

onbekend: Sterkstroom, Strydom 12.

Rekords van buite Suidelike Afrika moet siteer word

van noord na suid, d.w.s. dit gaan die van Suidelike Afrika

vooraf. Die afkorting “distr.” behoort by alle distriksname

gevoeg te word, bv.

Kenya. — Nairobi-distr. : Nairobivlakte anderkant die ren-

baan, Napier 845.

205

REFERENCES

References in the text should be cited as follows: “Jones

(1955) stated. . or . (Smit, 1956)” when giving

a reference simply as authority for a statement. The list of

references at the end of the article should be arranged alphabe-

tically and the literature abbreviations used should conform

to the list of Literature Abreviations (Technical Note: Tax.

6/1) issued by the Botanical Research Institute, thus:

HUTCHINSON, J., 1946. A botanist in Southern Africa.

London: Gawthorn.

MORRIS, J. W., 1969. An ordination of the vegetaion of

Ntshongweni, Natal. Bothalia 10: 89-120.

If, as in many taxonomic papers, periodicals or books

are mentioned in the text, usually in the species synopsis, they

should be cited as in the following examples: Gilg & Ben. in

Bot. Jahrb. 53: 240 (1915) and Burtt Davy, FI. Transv. 1:

122 (1926).

REPRINTS

Authors receive 75 reprints gratis. If there is more than

one author, this number will have to be shared between or

among them.

VERWYSINGS

Verwysings in die teks moet as volg siteer word: “Jones

(1955) beweer . . .” of “. . . (Smit, 1956)” wanneer ’n

verwysing slegs as outoriteit vir ’n stelling gegee word. Die

verwysingslys aan die einde van die artikel moet alfabeties

gerangskik wees en die Iiteratuurafkortings wat gebruik word,

moet in ooreenstemming wees met die lys van Literatuur-

afkortings (Tegniese Nota: Tax. 6/1) wat uitgegee is deur die

Navorsingsinstituut vir Plantkunde, as volg:

HUTCHINSON, J., 1946. A botanist in Southern Africa.

London: Gawthorn.

MORRIS, J. W., 1969. An ordination of the vegetation of

Ntshongweni, Natal. Bothalia 10: 89-120.

Wanneer, soos in baie taksonomiese artikels die geval

is, tydskrifte of boeke in die teks genoem word, gewoonlik

in die soortsinopsis, behoort hulle siteer te word soos in die

volgende voorbeelde: Gilg & Ben. in Bot. Jahrb. 53: 240

(1915) en Burtt Davy, FI. Transv. 1: 122 (1926).

HERDRUKKE

Skrywers ontvang 75 herdrukke gratis. Wanneer daar

meer as een skrywer is, sal hierdie aantal tussen hulle verdeel

moet word.

' . •

.

—

Bothalia 11,3: 207-210 (1974)

The Mycoflora of Wheat Field Debris, Part I

M. C. PAPENDORF* and W. J. JOOSTE*

ABSTRACT

Descriptions are given of five fungus species isolated from stubble of a wheat field in the Heilbron

district: Acremonium strictum Gams, Acremonium persicinu.n (Nicot) Gams, Acremonium curvulum

Gams, Tricellula curvata Haskins and Ulocladium consortiale (Thum.) Simmons.

Outbreaks of photosensitivity amongst sheep

grazing Panicum grasses in cultivated lands in the

Orange Free State were described by Steyn (1928) and

Rimington & Quin (1937). The aetiology of this

disease, locally known as Dikoor, has not yet been

resolved. Steyn (1928) considered the possibility that

a smut fungus which infected the Panicum plants

might have been the cause. Kellerman & Marasas

(1971) advanced the hypothesis that Dikoor may be

a mycotoxicosis caused by the toxic spores of sapro-

phytic fungi growing on wheat straw in these reaped

lands. In an attempt to establish a causal relationship

between a saprophytic fungus and Dikoor, the

Deparment of Agricultural Technical Services initiated

a survey of the mycoflora of wheat debris and Panicum

leaves. The authors co-operated in this survey. In this

series of papers a number of fungi isolated from these

materials, obtained from a farm in the Heilbron

district, Orange Free State, will be described.

The material was cut into short lengths of 1-2 cm

each, and shaken up in sterile water. The resultant

suspensions were diluted serially and plated on malt

extract agar (Oxoid malt extract 1,5%, agar 2%)

to which penicillin and streptomycin were then added

to inhibit bacterial growth. For microscopic study,

the fungi were grown on malt extract agar or oatmeal

agar and incubated at 25 °C.

Five species isolated from this habitat are described

below. Live cultures of these have been deposited in

the Potchefstroom University Culture Collection and

dried down cultures in the Mycological Herbarium

(PREM) at 590 Vermeulen Street, Pretoria.

Acremonium strictum Gams in Cephalosporium-

artige Schimmelpilze, p. 42. Gustav Fischer (1971).

Figure 3.

Colonies on oatmeal agar moderately fast growing

reaching a diameter of 25 mm in 10 days at 25 °C,

appressed to the medium and delicately floccose in

the centre, white becoming rosaceous in light, reverse

buff-rosaceous. Hyphae thin-walled, hyaline, up to

1,5 /xm and usually less than 1 pm diam. Phialides

arising from submerged or prostrate hyphae and

mostly from hyphal ropes, mostly single and un-

branched, straight and gradually tapering, thin-walled,

without collarette, 20-52 pm long, basal and apical

diameter normally not exceeding 2,0 /xm and 1,0

pm respectively, basally septate and sometimes with

one or more additional septa. Conidia produced in

siimy heads, cylindrical or ellipsoidal, straight or

occasionally slightly curved, hyaline, smooth.

3, 0-6, Ox 1 ,0-2,0 pm. Chlamydospores absent.

Specimen examined : P.U. Culture Collection,

No. 1178, isolated from wheat debris, Heilbron,

O.F.S., Dec. 1971. PREM 44855 dried culture on

1,5% malt extract agar.

* Institute for Botanical Research. Potchefstroom University

for C.H.E., Potchefstroom.

This is the first record of this species in South

Africa. — M.C.P.

Acremonium persicinum (Nicot) Gams in Ceph-

alosporiumartige Schimmelpilze, p. 75. Gustav

Fischer (1971).

Figure 1, 2.

Colonies on oatmeal agar fast growing reaching

a diameter of 30-40 mm within 10 days at 25 °C,

chondroid, floccose in centre with funiculose aerial

mycelium about 3 mm high, rosaceous in light,

reverse cream to brownish or saffron (Rayner: Chart

1, 10), numerous elongated crystals in older colonies.

Hyphae septate, partly thick-walled, hyaline, 1 ,5-2,5

fim diam. Phialides arising from prostrate hyphae and

mainly from aerial hyphal ropes, mostly simple,

straight and gradually tapering, thin-walled, smooth,

without collarette, basally septate, 20-35 pm long,

basal and apical diameter normally not exceeding 2,5

and 1,0 pm respectively. Conidia in dry heads with

limited number of spores or in chains, obovoid,

basally somewhat acuminate or rostellate and apically

rounded, smooth and eventually very finely rough-

ened, 3, 5-7,0 X 1 ,5 — 3,0 jtxm. Chlamydospores absent.

Specimen examined : P.U. Culture Collection. No.

1179, isolated from wheat debris, Heilbron, O.F.S.,

Dec. 1971. PREM 44856 dried culture on 1,5% malt

extract agar.

This is the first record of this species in South

Africa. — M.C.P.

Acremonium curvulum Gams in Cephalo-

sporiumartige Schimmelpilze, p. 57. Gustav Fischer

(1971).

Figure 4.

Colonies moderately fast growing on oatmeal agar

reaching a diameter of 25 mm in 10 days at 25 °C

closely appressed to the medium, aerial mycelium

forming scattered sporiferous hyphal ropes, salmon

coloured in light. Hyphae branched, septate, thin-

walled, hyaline, 1,0-2, 5 pm diam. Phialides arising

from prostrate hyphae or aerial hyphal ropes, single

or in groups, simple or branched, straight or slightly

curved or with uneven outline, gradually tapering,

usually without collarette, 18-57 pm long, basal and

apical diameter normally not exceeding 3,0 and

1,5 pm respectively, thin-walled, basally septate.

Conidia produced in slimy heads, falcate, at times

basally slightly pointed, hyaline, smooth, 4,0-10,0 X

1 ,0-2,0 pm. Chlamydospores absent.

Specimen examined: P.U. Culture Collection, No.

1177, isolated from wheat debris, Heilbron. O.F.S.,

Dec. 1971. PREM 44857 dried culture on 1,5% malt

extract agar.

208

THE MYCOFLORA OF WHEAT FIELD lOEBRIS

This isolate agrees very closely with the original

description (Gams, loc. cit.) but has spores longer

than 6,7 /zm. In a personal communication Dr Gams

explained that his conidium measurements indicate

only the 95% interval within which the average

figure would fall and that the dimensions of individual

conidia may easily be outside this range.

This is the first record of this species in South

Africa.- M.C.P.

Tricellula curvata Haskins in Can. J. Microbiol.

4: 273-285 (1958); Petersen in Bull. Torr. Bot. Club

89: 287-293 (1962); Webster in Trans. Br. mycol.

Soc. 42: 416-420 (1959).

Figures 5, 6, 7.

Colony on malt agar slow growing, reaching a

diameter of 30 mm in 25 days at 25 °C, smooth and

appressed with occasional raised funiculose clusters,

salmon apricot ( Rayner, 1970), reverse similar, no

pigment diffusing into the agar. Hyphae smooth,

septate, branched, hyaline, 1, 0-2,0 /zm diain. Coni-

diophores macronematous, mononematous, short.

Fig. 1-7. — Fig. 1, Acremonium per-

sicinum, phialide and phialocom-

dia (phase contrast). Fig. 2, A.

persicinum, thick-walled hyphae of

chondroid mycelium (phase con-

trast). Fig. 3, Acremonium s'.ric-

tuni, phialides and phialoccnidia

(phase contrast). Fig. 4, Acre-

monium curvulum, phialide and

phialoconidia (phase contrast).

Fig. 5, Tricellula curvata young

conidiophore showing blastic de-

velopment of conidia (phase con-

trast). Fig. 6 and 7, T. curvata,

mature conidiophore and conidia

(phase contrast).

4, 0-7, Ox 2, 0-2, 5 /zm, mostly clavate and branching

in the apical region to form a somewhat penicillate

head, at times curved, non-septate, hyaline. Coni-

diogenous cell integrated, holoblastic, polyblastic,

determinate, denticulate. Conidia composed of three

separate unequal hyaline and smooth cells, viz a

basal cell with a lateral and apical cell connected to

its apex by means of narrow protoplasmic isthmuses

and eventually separated by deposition of cell wall

material, overall dimensions 8,0-11,0x7,0-9,0 /zm;

basal cell variable, mostly somewhat reniform or

clavate and pointed basally, 4, 0-6,0 x 1 , 5-2,0 /zm;

apical cell somewhat reniform, 4, 0-5, Ox 1 ,5-2,0/zm;

lateral cell reniform with the free end pointed, 3, 0-5,0

X 1 ,5-2,0 /zm.

Specimen examined: P.U. Culture Collection No.

1184, isolated from wheat debris, Heilbron, O.F.S.,

March 1972. PREM 44858 dried culture on 1 , 5 % malt

extract agar.

M. C. PAPENDORF AND W. J. JOOSTE

209

The morphological features and dimensions of this

isolate agree well with those given by Haskins (loc.

cit.). Its cultural characters, however, differ from

those supplied by Webster (loc. cit.) who described

the colony on malt agar as pale pink above and buff

below with radial furrows and sparse sporulation.

Cultures of the isolate described here did not have

any radial furrows and the sporulation was abundant.

The identity of this isolate was confirmed by Dr

J. A. von Arx, Baarn, Netherlands.

T. curvata and other members of the genus have

been previously recovered from air and water. It is,

therefore, the first record of this fungus from a

different habitat and the first record of the genus in

South Africa. — W.J.J.

Ulocladium consortiale (Thiim.) Simmons in Myco-

logia 59: 67-92 (1967).

Macrosporium consortiale Thiimen in Elerb. mycol. oeconom.

450 (1876).

Stemphyllium consortiale (Thiim.) Groves & Skolko in Can.

J. Res. C 22: 196, (1944).

Stemphyllium illicis (Tengwall) Neergaard in Danish species

of Alternaria and Stemphyllium. Copenhagen: Einar Munks-

gaard. (1945).

Macrosporium abietis Tengwall in Meded. Phytopath. Lab.

Willie Commelin Scholten 6: 34-35 (1924).

Stemphyllium congestion Newton in Phytopathology 18:

565-578 (1928).

Stemphyllium dendriticum De Sousa da Camara in Proposta

de divisao do genero Stemphyllium fungo da ordem das Hyp-

hales (Mart.) em Sacc. et Trav. p. 21, Lisbonne (1930).

Alternaria consortiale (Thiim.) Groves & Hughes in Hughes,

Can. J. Bot. 31 : 636 (1953).

Pseudostemphyllium consortiale (Thiim.) C. V. Subramanian

in Curr. Sci. 30: 423 (1961).

Figures 8, 9, 10, 11, 12.

Colony on malt agar growing rapidly at 25 °C

reaching a diameter of 50 mm in 5 days, mycelium

appressed, spore production prolific resulting in a

colony surface which is carbonaceous velvety, colour

olivaceous-black with the reverse of the colony

similar. Hyphae smooth to finely roughened, light

buff to honey, bright, septate, 3-5 pm diam. Coni-

diophores macronematous, smooth, light isabelline

(Rayner, 1970), bright, 50,0-150,0x4,0-5,0 /am,

geniculate in the conidiogenous area. Conidiogenous

cells integrated, polytretic, 1-2 pores per cell, scars

small and usually more than 10. Conidia dry, mostly

solitary, occasionally in short chains, often with a

geniculate false beak with up to 1 1 geniculations and

serving as a secondary conidiophore, colour variable,

light umber when young and dark umber at maturity,

shape variable, subglobose, obovoid or oblong-

elliptical, often basally tapering to a porate apiculus,

septate, usually with 1-4 transverse and 2-6 longi-

tudinal and oblique septa, slightly to moderately

constricted at the septa, 21-35x11-22 pm, ave.

Fig. 8-12. — Fig. 8 and 9, Ulocladium

consortiale, conidia, demonstrating

variability found in this species.

Fig. 10, U. consortiale, conidia and

conidium with branched geniculate

false beak X 1 000 (scanning elec-

tron micrograph). Fig. 11 and 12,

U. consortiale, conidia X 2 400

showing variations in spore surface

characteristics (scanning electron

micrograph).

210

THE MYCOFLORA OF WHEAT FIELD DEBRIS

27,6-18,2 fim, smooth, verruculose or verrucose;

scanning electron micrographs reveal that the sur-

face of the apparently smooth conidia are finely

verruculose while that of the roughened conidia are

irregularly cerebriform and often with roughened and

smooth areas on the same conidium. In culture sub-

merged globose to sub-globose spores develop in the

agar, mostly 1-2 septate with transverse and longi-

tudinal septa, slightly constricted at the septa, ver-

ruculose, umber, 11-17x11-15 pm, remaining at-

tached to the hyphae.

Specimen examined: P.U. Culture Collection No.

1214, isolated from wheat debris, Heilbron, O.F.S.,

March 1972. PREM 44859 dried culture on 1,5%

malt extract agar.

Groves & Skolko (loc. cit.) accepted Macrosporium

consortiale (Thiim.) as the type for Stemphyllium

consortiale in the section Pseudostemphyllium erected

by Wiltshire (Trans. Br. mycol. Soc. 21, 211-239,

1939). They described the conidia as “variable in

shape, quadrilateral to oblong-ovoid, at first nearly

smooth, becoming coarsely verrucose, muriform, more

or less constricted at the septa, very dark to brown or

black, ( 1 3)— 1 8— 25— (28) x (1 2)— 1 4—1 7— (23) ^m”. Sim-

mons (loc. cit.) made M. consortiale the type for U.

consortiale. However, he regarded the description

of S. consortiale by Grove & Skolko as incorrect and

thought it to be U. at rum. S. consortiale was never-

theless retained as a synonym of U. consortiale.

In his treatment of Stemphyllium illicis Tengwall in

the section Pseudostemphyllium, Neergaard (loc. cit.)

noted that the conidia are smooth, echinulate or more

or less verrucose, even warty. He furthermore stated

in the Appendix to his book that S. illicis is identical

with S. consortiale (Thiim.) Groves & Skolko and

the latter name should, therefore, have priority. It

follows that S. illicis should be regarded as a synonym

of U. consortiale as well.

The isolate described here corresponds well with

the description given by Simmons (loc. cit.) except for

the fact that it differs in the surface features of the

conidia. According to Simmons (loc. cit.) “t/. con-

sortiale is restricted to a species with smooth to

inconspicuously roughened conidia of medium colour

density.” The conidia of this isolate vary from

apparently smooth to definitely roughened and can

hardly be described as predominantly smooth. It is

also similar to S. illicis described by Neergaard

(loc. cit). Since there is no obvious basis for regarding

this isolate as a separate species it must be labelled

U. consortiale.

This is the first record of this species in South

Africa.— W.J.J.

ACKNOWLEDGEMENTS

The financial assistance of the South African

C.S.I.R. and the Dept, of Agricultureal Technical

Services is gratefully acknowledged. The Dept, of

Zoology of the Potchefstroom University for C.H.E.

is thanked for the use of the Cambridge Stereoscan

Electron Microscope, and Dr V. Hamilton-Attwell

for the scanning electron micrographs.

REFERENCES

Kellerman, T. S. & Marasas, W. F. O., 1971. The possible

role of fungi in the aetiology of geeldikkop. Unpublished

paper presented at the Congress of the S.Afr. vet. med. Ass..

East London, Sept. 1971

Rayner, R. W. 1970. A Mycological Colour Chart. Kew:

Commonwealth Mycological Institute.

Rimington, C. & Quin, J. I. 1937. Dikoor or geeldikkop

on grass-veld pastures. J. S. Afr. vet. med. /Us. 8: 141-145.

Steyn, D. G., 1928. Dikoor in sheep. J. S. Afr. vet. med. Ass.

1, 47-50.

Bothalia 11, 3: 211—215 (1974)

New and Interesting Records of South African Fungi, Part VIII

M. C. PAPENDORF* and W. J. JOOSTE*

ABSTRACT

Descriptions of the following eight fungi isolated from various habitats are given: Achaetomium

strumarium Rai, Tewari & Mukerji, from a leaf of Protea sp. ; Phoma glomerata (Corda) Wollenw.

& Hochapf. from wheat debris; Phoma glumarum Ell. & Tracy and Phoma capitulum Pawar,

Mathur & Thirumalachar from leaf litter of Acacia karroo Hayne; Phoma jolyana Pirozynski &

Morgan-Jones from contaminated agar plate; Veronaea botryosa Cifferi & Montemartini, Scytali-

dium ligmcolum Pesante and Rhmocladiella mansonii (Castell.) Schol-Schwarz from deteriorated

canvas; Cerebella andropogonis Ces. from Lolium nmltiflorum Lam. spikelets.

Eight fungal species isolated from various

habitats are described and discussed below. The

descriptions are based on agar cultures incubated at

25 °C. Live cultures of these have been deposited in

the Potchefstroom University Culture Collection and

dried down cultures in the Mycological Elerbarium

(PREM) at 590 Vermeulen Street, Pretoria.

Achaetomium strumarium Rai, Tewari & Mukerji

in Can. J. Bot. 42: 693-697 (1964).

Figure 4.

Colonies develop fairly rapidly on potato-dex-

trose agar reaching a diameter of 9 cm within 10 days,

closely appressed to the medium, buff to dull grey-

brown, reverse rosaceous with pigment diffusing into

the agar. Hyphae hyaline to light golden-brown,

branched, septate, 2, 0-5,0 ,um diam. Ascomata

mostly superficial, single or aggregated, subglobose

to ovoid, ostiolate, without distinct neck, dull golden-

brown to nearly black when mature, 30, 0-1 60, Ox

30,0-90,0 /zm; peridium delicate and consisting of

loosely woven hyphae, tawny, usually adorned with

delicate, hyaline to subhyaline radiating hairs up to

300 /zm long and 2, 0-4,0 pm diam. Asci evanescent,

fascicled, cylindrical, aporate, octosporous, 45,0-

85,0x6,0-8,0 /zm; ascospores uniseriate, olivaceous

to dark olive-brown, greenish-grey in mass at maturity,

ellipsoidal with slightly pointed ends, single basal

germ pore, 9,0-13,0x6,0-8,0 /zm.

Specimen examined: P.U. Culture Collection, No

1175, isolated from leaf of Protea sp., Kempton Park,

Transvaal, March 1972. PREM 44860 dried culture on

1,5% malt extract agar.

The genus Achaetonium was erected by Rai et al.

(loc. cit.) with A. globosum as type species. The genus

resembles members of the Chaetomiales and is,

according to the original description, characterized

by ostiolate ascomata with evanescent asci, dark one-

celled ascospores and a delicate, light coloured

peridium devoid of hairy outgrowths. With the

exception of the very obvious hairs on the perithecia

the features of this specimen agree closely with those

of A. strumarium and there are only minor differences

in the size of the ascomata, asci and ascospores.

Evaluating the relative importance of perithecial

hairs of certain Ascomycetes Dr von Arx commented,

in a personal communication, that this feature should

not be over-emphasized and often has only limited

taxonomic value. More important characters are the

nature of the germ pore and the colour of the perithe-

cium and ascospores. According to him Achaetonium

is characterized mainly by dark ascospores with a

single basal germ pore and a perithecium without a

terminal tuft of hairs surrounding the ostiole.

* Institute for Botanical Research, Potchefstroom University

for C.H.E., Potchefstroom.

This is the first record of the occurrence of this

genus in South Africa. — M.C.P.

Phoma glomerata {Corda) Wollenw. & Hochapf.

in Z. Parasit Kde. 8: 592 (1936).

Coniothyrium glomeratum Corda, Ic. Fung. 4, 39 (1940).

Peyronellaea glomerata (Corda) Goidanich in Rc. Accad.

Lincei 1, 445; 658 (1946).

Phoma alternariaceum Brooks & Searle in Trans. Br. mycol.

Soc. 7, 193 (1921).

Figure 2.

Colonies on malt agar attain a diameter of 8 cm

in 7 days, closely appressed to the medium with little

aerial mycelium, brown to olivaceous. Hyphae septate,

branched, brown-olivaceous, 3, 0-8,0 /zm diam.

Chlamydospores brown to dark brown; growing

from the aerial mycelium, prostrate hyphae and

mature pycnidia in simple or branched chains con-

sisting of 2-10 units; shape variable, but mostly

irregularly ovate, obpyriform or fusiform-ellipsoidal

with 3-8 transverse and occasional longitudinal or

oblique septa, 15,0-30,0x15,0-20,0 /zm. Pycnidia

superficial or immersed in medium, brown, subglo-

bose to obpyriform, up to 160 /zm diam., often

laterally compressed and oval from above, 20,0 x

150,0 /zm; ostiole single or occasionally 2-3; spore

mass opaque, rosaceous. Pycnidiospores hyaline to

subhyaline, olive when mature, guttulate; shape

variable, ellipsiodal, ovoid, oblong-ellipsoid or irregu-

lar; very rarely 1-septate, 4,0-10,0x2,0-6,0 /zm.

Specimen examined'. P.U. Culture Collection, No.

1187, isolated from wheat debris, Heilbron, O.F.S.,

December 1971. PREM 44861 dried culture on 1,5%

malt extract agar.

The identity of this specimen was verified by Dr

H. A. van der Aa, Baarn, Netherlands. The characters

of this isolate agree closely with the description given

by Boerema et al. (Persoonia 4: 53-54. 1965) and

Morgan-Jones (C.M.I. Descriptions of Pathogenic

Fungi and Bacteria, No. 134). Boerema (loc. cit.)

pointed out that this species is strongly influenced by

cultural conditions and displays wide variation in its

main features, but that it is easily recognised by the

characteristic chlamydospores produced in chains and

resembling the conihia of Alternaria. It should be

noted that typical chlamydospores are not readily

observed in very young cultures and that these should

not be used for identification purposes (Leudemann,

Mycologia 51, 772-780, 1959). The full synonymy of

this species is listed by Boerema (loc. cit).

This is the first record of the occurrence of this

species in South Africa. — M.C.P.

Phoma capitulum Pawar , Mathur & Thirumalachar

in Trans. Br. mycol. Soc. 50: 261 (1967).

Phoma ostiolata Pawer, Mathur & Thirumalachar, in Trans.

Br. mycol. Soc. 50: 262 (1967).

Phoma ostiolata var. bntnnea Pawar, Mathur & Thiruma-

lachar, in Trans. Br. mycol. Soc. 50: 263 (1967).

212

NEW AND INTERESTING RECORDS OF SOUTH AFRICAN FUNGI

Colonies on potato-dextrose agar slow growing and

reaching a diameter of 3,5 cm in 10 days, eventually

compact and spongy; aerial mycelium cottony, pure

white; reverse colourless or faintly rosaceous. Hyphae

hyaline, branched, up to 3,0 /m diam. Pycnidia

tardily produced in small numbers, superficial or

immersed in the medium, subglobose, at first light

brown becoming brown to nearly black, neck incon-

spicuous or absent, mostly with a single ostiole,

variable in size and up to 250 /z m diam. Pycnidiospores

hyaline, mainly ovoid, smooth, discharged in greyish-

white opaque mass, 2, 5-7, 0x2, 0-4,0 ^m.

Specimen examined : P.U. Culture Collection, No.

71, isolated from leaf litter of Acacia karroo , Potchef-

stroom, Transvaal, Jan. 1964. PREM 44862 dried

culture on 1,5% malt extract agar.

The identity of this specimen was confirmed by

Dr G. H. Boerema, Wageningen, Netherlands.

In its main features this isolate agrees closely with

the description of the type (Pawar et al., loc. ci t . ). It

differs in having darker pycnidia, slightly larger

spores and a grey-white discharged spore mass.

The fact that P. ostiolala and P. ostiolata var. hrunnea ,

described by Pawar et al. (loc. cit.) as distinct types

on the grounds of slight morphological variations,

were reduced to synonymy with P. capitulum (Boerema

& Dorenbosch, Trans. Br. mycol. Soc. 51 : 145-146,

Fig. 1-5. — Fig. 1, Plioma glumarum

intercalary chlamydospores. Fig. 2,

Phoma glomerata catenate chla-

mydospores. Fig. 3, Phoma joly-

ana chlamydospores. Fig. 4 and 5,

Achaetomium strumarium asci and

ascospores.

1968), indicates some variability in the characters of

this species.

This is the first record of the occurrence of this

species in South Africa. — M.C.P.

Phoma jolyana Pirozynski & Morgan- Jones in

Trans. Br. mycol. Soc. 51 : 200 (1968).

Phoma mitsae (Joly) Boerema, Dorenb. & Rest, in Persoonia

4: 47—68 (1965).

Peyronellaea musae Joly in Rev. Mycol. 26, 97 (1961).

PeyroneUaea nainensis Tandon & Bilgrami in Curr. Sci.

30, 344 (1961).

Figure 3.

Colonies grow fairly rapidly on malt agar attaining

a diameter of 7 cm in 7 days, appressed, with little

aerial mycelium, olive-brown becoming dark oliva-

ceous to nearly black. Hyphae olivaceous-buff,

branched, septate, 2,0-7,0/zm diam. Chlamydospores

brown, lateral or terminal on hyphae, sometimes

intercalary, simple or branched, shape variable but

mostly more or less clavate or obovoid, 1-20 septate

with occasional longitudinal or oblique septa, 12,0-

75,0x5,0-87,0 /im. Pycnidia abundant, scattered,

superficial or immersed in the medium, small pycnidia

often develop on aerial mycelium, subglobose to

obpyriform, normally with a single ostiole, brown to

nearly black, 50-210 fim diam. Pycnidiospores hyaline

to light olivaceous, shape variable, usually ovoid.

M. C. PAPENDORF AND W. J. JOOSTE

213

oblong or irregular, continuous, guttulate, 4, 0-8, 5 x

2.0- 4, 5 /xm.

Specimen examined : P.U. Culture Collection, No.

1059 isolated from contaminated agar plate, Potchef-

stroom, Transvaal, 1966. PREM 44863 dried culture

on 1,5% malt extract agar.

The identity of this specimen was verified by Dr

G. H. Boerema, Wageningen, Netherlands.

According to Boerema el al. (loc. cit.) this species is

easily recognised by the single, more or less clavate

dictyochlamydospores developing terminally or late-

rally on the hyphae. It is further pointed out that

the size and pigmentation of the pycnidia, pycnidio-

spores and chlamydospores are markedly affected by

the age of the cultures and by the C/N ratio of the

medium. These variations have caused considerable

confusion in the taxonomy of this group of fungi

and should therefore be taken into account in the

identification of species.

This is the first record of the occurrence of this

species in South Africa. — M.C.P.

Phoma glumarum Ell. & Tracy in J. Mycol.

4:123 (1888).

Figure 1.

Colonies on malt agar develop fairly rapidly

reaching a diameter of 6,5 cm in 5 days, appressed

to the medium with scanty aerial mycelium, distinct

red pigment diffusing into agar. Mycelium brownish

to olivaceous; hyphae septate, branched, 3,0-10,0

/xm diam. Chlamydospores usually intercalary, vary-

ing from unicellular to dictyochlamydospores, brown

to brown-olivaceous; unicellular chlamydospores

solitary or in short chains, 7-18 /xm diam; dictyo-

chlamydospores irregular in shape but mostly fusi-

form-ellipsoid or ovoid to globose, 15, 0-55, Ox

8.0- 32,0 /xm, individual cells up to 22,0 /xm diam.

Pycnidia superficial or partly immersed in agar,

brown to black, variably pyriform or subglobose,

diameter 40-200 /xm mostly with distinct neck and

ostiole, ostiole often lined with dark cells. Pycnidio-

spores hyaline to light olivaceous, usually guttulate,

ellipsoid to ovoid or globose, continous or rarely

1 -septate, 3, 0-7, Ox 2, 0-4,0 /xm.

Specimen examined : P.U. Culture Collection, No.

280, isolated from leaf litter of Acacia karroo , Potchef-

stroom, Transvaal, Jan. 1964. PREM 44864 dried

culture on 1,5% malt extract agar.

The identity of this specimen was confirmed by

Dr G. H. Boerema, Wageningen, Netherlands.

The full synonymy of this species is listed by

Boerema (Persoonia 6: 174-175, 1971).

According to Boerema (Persoonia 5: 203, 1968)

the characters of this species in vitro show wide

variation but the fungus is, nevertheless, easily recog-

nized by the dark, beaked pycnidia, the intercalary

dictyochlamydospores and the production of a dis-

tinct reddish (orange to red-purple) pigment. The

C/N ratio of the medium affects the production of

pycnidia and chlamydospores — a high ratio favouring

chlamydospore production and a low ratio the forma-

tion of pycnidia (Boerema, Persoonia 5: 203, 1968).

—M.C.P.

Veronaea botryosa Cifferi & Monte mart ini in

Atti. 1st. bot. Univ. Lab. crittogam Pavia Ser. 5.

15: 67-72 (1958); Ellis in Dematiaceous Hypho-

mycetes, Commonwealth Mycological Institute, Kew,

p. 245 (1971).

Sympodina coprophila Subramanian & Lodha in Antonie

van Leeuwenhoek 30: 317-330 (1964).

Figures 6, 7.

Colonies on malt agar slow growing, reaching a

diameter of 27 mm in 12 days, velvety-granular, dark

grey-olivaceous, reverse similar, no pigment diffusing

into the agar. Hyphae smooth, septate, branched,

with branches often at right angles and with a septum

in the main axis close to the point of branching,

slightly wavy with cell walls often irregular and

slightly bulging, rarely straight, 2-3 /xm diam.,

olivaceous-grey. Conidiophores macronematous,

mononematous, smooth-walled, septate, slightly wavy,

olivaceous-grey, branched or unbranched, arising at

right angles on trailing hyphae, often basally septate

and constricted at this septum, variable in length,

55,0-300,0x2,0-3,0 /xm. Conidiogenous cells inte-

grated, polyblastic, restrictedly sympodial, rarely

geniculate, scars small and flat, mostly determinate

but rarely resumes normal growth resulting in an

intercalary conidiogenous area. Conidia solitary, dry,

acropleurogenous, oblong-ellipsoid to ellipsoid,

rounded apically, rarely fusiform, slightly tapering

basally, mostly with a definite hilum, 0-1-septate

often constricted at the septa, smooth, thin-walled,

olivaceous-buff, 5,0-12,0x2,0-5,0 /xm.

Specimen examined'. P.U. Culture Collection, No.

1165. Isolated from deteriorated canvas, by M. C.

Papendorf, Potchefstroom, Sept. 1970. PREM 44865

dried culture on 1,5% malt extract agar.

At present only two species of Veronaea are known

viz. V. botryosa and V. simplex Papendorf (Trans. Br.

mycol. Soc. 52: 483-498, 1969). Papendorf (loc. cit.)

considered Sympodina congeneric with Veronaea and

Dr J. A. von Arx (personal communication) considers

S. coprophila , the only species of Sympodina , to be

identical with V. botryosa. The latter as described by

Cifferi & Montemartini (loc. cit.) differs from the

isolate described here as well as from 5. coprophila ,

in having club-shaped conidiophores with small

sterigmata and conidia with pointed ends and no

constrictions at the septa. Von Arx (personal com-

munication) found the present isolate to differ only

slightly from the type of V. botryosa. It appears that

Subramanian & Lodha’s (loc. cit.) diagnosis for

S. coprophila is actually more acceptable for V.

botryosa than the original description by Cifferi &

Montemartini (loc. cit.).

Although the conidia of V. botryosa are similar to

those of V. simplex , the latter species is distinguished

by its less elaborate, shorter, geniculate conidio-

phores with prominent protruding scars.

This is the first record of the occurrence of this

species in South Africa. — W.J.J.

Scytalidium lignicolum Pesante in Annali Sper.

agr. N.S. 11 : 249-266; Ellis in Dematiaceous Hypho-

mycetes. Commonwealth Mycological Institute, Kew,

p. 28 (1971).

Figure 8.

Colonies on malt agar growing rapidly reaching a

diameter of 90 mm in 5 days at 25 C, mycelium effuse

flocculose, initially rosy buff (Rayner, 1970) but

darkening rapidly to olivaceous-grey to dark oliva-

ceous (Rayner, 1970). Hyphae smooth, septate,

branched, hyaline and thin-walled, 1,5-12,0 /xm

diam., sections of the hyphae developing into inter-

calary, thick-walled, olivaceous cells (not arthro- or

chlamydospores), very narrow hyphae generally ap-

pearing after repeated branching of the wider hyphae.

Conidiophores micronematous, determinate, variable

in length but occasionally short, mainly when bearing

type II conidia, or very long and slender when bearing

both types of conidia. Conidiogenous cells thallic,

integrated, intercalary or terminal, smooth or ver-

ruculose, hyaline, fragmenting to form two types of

214

NEW AND INTERESTING RECORDS OF SOUTH AFRICAN FUNGI

Fig. 6-1 3. — Fig. 6, Veronaea botryosa

conidiophore and conidia (inter-

ference contrast). Fig. 7, V.

botryosa branched conidiophore

(phase contrast). Fig. 8, Scytali-

dium lignicolum Type I and Type

11 arthroconidia (phase contrast).

Fig. 9 and 10, Cerebella andropo-

gonis conidiophores and conidia.

Fig. 11, Epicoccum purpurascens

conidiophores and conidia. Fig.

12 and 13, Rhinocladiella man-

sonii conidia and multiple branched

conidiophores (phase contrast).

arthroconidia. Conidia type I terminal, often arising

from slender conidiogenous cells on slender conidio-

phores, catenate, hyaline, thin-walled, dry, cylindrical

to oblong-cylindrical and truncate at each end,

4, 0-8,0 x 1 ,5-3,0 ^.m; these conidia also occur on

roughened, terminal, phialide-like, conidiogenous cells

at the ends of chains of type IT conidia. Conidia type

II terminal or intercalary, variable in shape, mostly

sub-globose or oblong-elliptical to doliiform, mode-

rately thick walled, 0-1 -septate with septa prominent,

often constricted at the septa, cells equal or unequal,

smooth, olivaceous, 6,0-15,0x5,0-6,0 pm.

Specimen examined-. P.U. Culture Collection, No.

1161, isolated from deteriorated canvas, by M. C.

Papendorf, Potchefstroom, Sept. 1970. PR EM 44866

dried culture on 1,5% malt extract agar.

The isolate described here is essentially similar to

the type according to the description of Pesante

(loc. c i t . ) . Slight variations in the dimensions were

noted but no microsclerotia were observed. The

identity of this fungus was confirmed by Dr W. Gams,

Centraalbureau voor Schimmelcultures, Baarn.

This is the first record of the occurrence of this

genus in South Africa. — W.J.J.

Rhinocladiella mansonii (Cast ell.) Schol-Schwarz

in Antonie van Leeuwenhoek 34: 119-152 (1968).

Figures 12, 13.

Colonies on malt agar slow growing at 25 °C,

reaching a diameter of 25 mm in 12 days; mycelium

slightly raised, velvety, granular, grey-olivaceous,

reverse olivaceous-black, (Rayner, 1970). Hyphae

smooth with slight bulges, slightly wavy, septate,

honey-buff, 1,5-2, 5 pm diam., adjacent hyphae

frequently anastomosing. Conidiophores macrone-

matous, mononematous, septate, unbranched or

branched, length variable, mostly 20-60 /im, bright,

honey-buff, arising singly as side branches of the

hyphae or branched with each branch ending in a

conidiogenous cell, branches rarely of similar length.

Conidiogenous cells polyblastic, integrated, terminal,

elongating, denticulate with small flat scars, often

resuming growth resulting in intercalary conidio-

genous areas. Conidia hayline, smooth, thin-walled,

oblong-ellipsoidal to fusiform-ellipsoidal, tapering

M. C. PAPENDORF AND W. J. JOOSTE

215

slightly to a truncate base, continuous, 3, 0-8, Ox

1, 5-2,0 /a.m; when separated from the conidiophore

the conidia often become thick-walled and swollen

to produce large numbers of secondary blastoconidia

of various shapes and sizes, mainly obovoid and

basally truncate; older conidia frequently anastomose

to form irregular clusters.

Specimen examined : P.U. Culture Collection, No.

1160, isolated from deteriorated canvas by M. C.

Papendorf, Potchefstroom, Sept. 1970. PR EM 44867

dried culture on 1,5% malt extract agar.

This isolate differs in some aspects from the descrip-

tion given by Schol-Schwarz (loc. cit.). The colony is

not slimy, neither were the various spore types, viz

Phialophora annellate-type and Cladosporium- type

observed. The conidia are also larger than the 2, 0-3,0

(5 ,5) X 1 ,0-1 ,5 (2,0) pm reported by Schol-Schwarz

(loc. cit.). However, other characteristics such as

proliferation of secondary conidia and the anastomos-

ing of older conidia agree well with those of R.

mansonii.

The identity of this isolate was confirmed by Dr

S. de Hoog. Centraalbureau voor Schimmelcultures,

Baarn.

This is the first record of the occurrence of this

species in South Africa. — W.J.J.

Cerebella andropogonis Ces. apud Rabenhorst in

Bot. Ztg. 9: 669 (1851); Langdon in Mycol. Papers61

(1955); Ellis in Dematiaceous Hyphomycetes, Com-

monwealth Mycological Institute, Kew, p. 246 (1971).

Epicoccum andropogonis (Ces.) Schol-Schwarz in Trans. Br.

mycol. Soc. 42, 149-173 (1959).

Figure 9, 10.

Colonies on malt agar growing moderately fast and

reaching a diameter of 35 mm in 5 days at 25 °C.

Mycelium at first light cream, mainly submerged,

becoming delicately floccose, often funiculose and

producing a characteristic sienna-umber (Rayner,

1970) pigment in the agar. Hyphae smooth to finely

verruculose, septate, branched, hyaline. Conidio-

phores semi-macronematous, mononematous, short,

up to 10,0 pm long and 5,0 pm diam., light brown,

sometimes branched, smooth to finely verruculose,

septate, forming dense clusters resembling sporodo-

chia. Conidogenous cells integrated, holoblastic,

monoblastic and determinate. Conidia single, dry,

light to dark brown, shape variable but mainly sub-

globose; 2 to many-celled, mostly 5-6; relatively

thin-walled with longitudinal, transverse or oblique

septa, constricted at the septa, with protuberant

truncate basal stalk cell, smooth to verruculose,

variable in size, 13,0-20,0x10,0-20,0 pm. The

fungus forms a much convoluted cerebriform stroma

on the host.

Specimen examined'. P.U. Culture Collection, No.

1186, isolated from spikelets of Lolium multiflorum

Lam. infected with Claviceps purpurea (Fr.) Tul.,

collected at Underberg, Natal, by Dr W. F. O.

Marasas, Feb. 1972. PREM 44868 dried culture on

1,5% malt extract agar.

In an extensive study of the genus Cerebella ,

Langdon (loc. cit.) maintained it as a separate genus

with C. andropogonis as the only species and listed

32 synonyms. Schol-Schwarz (loc. cit.), however,

placed C. andropogonis in the genus Epicoccum as

E. andropogonis because of similarties in conidial

ontogeny. Ellis (loc. cit.) did not accept this and

maintained C. andropogonis on account of its distinc-

tive conidial morphology and the production of the

cerebriform stroma in vivo.

The isolate described here has conidial dimensions

which agree with the lower limits given by Langdon

(loc. cit.). Considering the variability of this genus as

indicated by Langdon (loc. cit.), there should be no

doubt about the identity of the isolate. It is readily

distinguished from E. purpurascens (fig. II) by the

marked constrictions at the septa, the relatively

smooth conidial surface and the smaller and lighter-

coloured conidia.

This species has been recorded in South Africa

before. According to the synonyms given by Langdon

(loc. cit.), Doidge in Bothalia 5, (1950) listed it as

Sorosporium africanum Syd. (p. 378), Stigmella

graminicola Linder (p. 704) and Cerebella cvnodonlis

Syd. (p. 730). —W.J.J.

ACKNOWLEDGEMENTS

We are indebted to the South African C.S.l.R. and

the Department of Agricultural Technical Services

for financial support received and to Dr G. C. A.

van der Westhuizen for valuable comments and for

reading the manuscript.

REFERENCES

Rayner, R.W., 1970. A Mycological Colour Chart. Kew:

Commonwealth Mycological Institute.

Bothalia 11, 3: 217—219 (1974)

A new species of Phaeorcimularia (Fungi Imperfecti: Dematiaceae)

from South Africa

W. F. O. MARASAS and INGRID H. BREDELL*

ABSTRACT

A dematiaceous Hyphomycete isolated from wheat and oat straw, as well as lucerne seed in South

Africa, is described as Phaeoramularia kellermaniana Marasas & Bredeli, sp. nov. The relationships of

P. kellermaniana to Cladosporium resinae (Lindau) de Vries and other species of Phaeoramularia are

discussed.

During an investigation of the litter mycoflora in

cultivated lands where outbreaks of photosensitivity in

sheep occurred, an undescribed dematiaceous Hypho-

mycete was isolated from wheat ( Triticum aestivum L.)

and oat (A vena sativa L.) straw. This fungus was also

isolated from South African lucerne ( Medicago

sativa L.) seed (Marasas & Bredeli, 1973). Isolation

and cultural methods were similar to those described

previously (Marasas & Schumann, 1972).

A culture of the undescribed Hyphomycete was

examined by Dr M. B. Ellis of the Commonwealth

Mycological Institute, Kew, England. He identified

it as a new species of Phaeoramularia Muntanola.

Phaeoramularia kellermaniana Marasas &

Bredeli, sp. nov.

Figures: 1-8.

Hyphae ramosae, septatae, hyalinae vel brunneae,

leves vel verrucosae, 2-7/u lat. Chlamydosporae

intercalares vel terminales, globosae, unicellulares,

laeves vel verrucosae, crassitunicatae, fuscae, 6— 1 8/x

in diametro. Conidiophora singula ex apicibus lateri-

busque hypharum oriunda, simplicia vel raro ramosa,

continua vel parce septata, glabro-tunicata, cylindrica

vel clavata, recta vel flexuosa, olivacea vel brunnea,

10-50 x2-7/x. Ce/lulae conidiogenae terminales, cylin-

dricae vel clavatae, ad et prope apicem 1-4 cicatrices

in denticulis gerentes. Conidia ex apicibus lateribusque

cellularum conidiogenarum in catenis longis simplici-

bus aut plerumque ramosis producta, olivacea vel

brunnea, laevia, ellipsoidea vel fusoidea, ad basim

rotundata vel truncata, ad apicem rotundata, obtusa

vel 1-4 cicatricibus praedita, continua vel 1-4 septata.

6-30 x 2-6/u.

Habitat in caulibus emortuis Tritici aestivi L.,

Kapgat, Calvinia, Cape Province, Africa australi

(PRE 44703, holotypus).

Colonies on potato-carrot agar at 25° C growing

rapidly (8-9 mm/day), covering the surface of a

85 mm diam. petri dish within 10 days and sporulating

well within 3 days, effuse, velvety, grayish-olive

(Ridgway, Plate XLVI). Colonies on 1,5% malt

extract agar at 25° C growing rapidly (6-7 mm/day),

covering the surface of a 85 mm diam. petri dish

within 10 days and sporulating well within 3-5 days,

woolly, olive green (Ridgway, Plate IV) to dark

ivy green (Ridgway, Plate XLVII), approaching

medal bronze (Ridgway Plate IV) in highly sporulating

sectors, reverse olive green becoming black in age.

Mycelium composed of hyaline to olivaceous or

brown branched, septate smooth or encrusted hyphae,

2-lp diam. Hyphae often forming strands and the

olivaceous or brown aerial hyphae composed of

cylindrical or ellipsoidal cells, 8-26x4-7/x, each of

which can give rise to a lateral conidiophore. In old

cultures these hyphal cells often become swollen.

* Plant Protection Research Institute, Department of

Agricultural Technical Services, Private Bag XI 34, Pretoria.

thick-walled, dark brown, globose, intercalary or

terminal chlamydospores, 6-18^. diam. Conidiophores

semi-macronematous or macronematous, mononema-

tous, acroauxic, arising at right angles to the aerial

hyphae as lateral outgrowths of the hyphal cells,

occasionally terminal on the hyphae, simple or rarely

sympodially branched, aseptate to multiseptate,

usually with a septum at the base, not constricted or

swollen at the septa, smooth-walled, olivaceous to

brown, straight or flexuous, cylindrical but frequently

swollen and truncate at the apex, 1 0-50 x 2-7 p.,

rarely 100^ or longer. Conidiogenous cells polyblastic,

giving rise to 1-4 conidial chains on terminal or

lateral protuberances, integrated, terminal, sympodial,

cylindrical to clavate or truncate at the apex, cica-

triced. Conidia blastospores that give rise to long

simple or repeatedly branched persistent chains of up

to 30 or more spores by successive budding in an

acropetal sequence, dry, at first hyaline becoming

olivaceous to brown, smooth-walled, ellipsoid or

fusoid with rounded or truncate ends, without a

protuberant scar at each end, mostly O-septate,

sometimes 1-4-septate and then slightly or not

constricted at the septa, 6-30x2-6/u.. Any single

conidium can also function as a polyblastic coni-

diogenous cell, giving rise by budding to 1-4 chains of

conidia. Conidia borne in positions where branching

of chains occur, are often clavate or "boot-shaped"

with 1-4 terminal, and sometimes also lateral, protu-

berant points of attachment for conidia. These conidia

resemble ramo-conidia, but since they also secede

easily, they are best considered conidia.

Specimens examined. — Cultures on 1,5% malt

extract agar and potato carrot agar: Cape- Kapgat,

Calvinia, isolated from wheat (Triticum aestivum L.)

straw, February, 1972, Marasas OP-76, PRE 44703

(PRE, holotype); Oudtshoorn, isolated from lucerne

( Medicago sativa L.) seed, July, 1970, Marasas 325,

PRE 44704 (PRE, paratype).

P. kellermaniana has also been isolated from oat

(Avena sativa L.) straw, Swinburne, Harrismith

distr.. Orange Free State, and from a number of

samples of lucerne seed produced in the Cape Province.

Type specimens in the form of dried-down cultures

on agar have been deposited in the National Her-

barium. PRE (Mycological Herbarium). Department

of Agricultural Technical Services, Private Bag XI 34,

Pretoria. Cultures of the type strain (PRE 44703)

have also been deposited in the Commonwealth

Mycological Institute, Kew, England (IMI 165252);

Centraalbureau voor Schimmelcultures. Baarn Nether-

lands; and American Type Culture Collection,

Rockville, Maryland.

The genus Phaeoramularia Muntanola is similar to

Ramidaria Unger except for the olivaceous or brown

colour of the spores and conidiophores (Muntanola,

I960). The generic relationships of Phaeoramularia

have been discussed by Muntanola (1960) and Sutton

(1970). The genus is rather difficult to distinguish from

218

A NEW SPECIES OF PHAEORAMULARIA

Fig. 1-8.- Phaeoramularia kellermaniana. Fig. 1, encrusted vegetative hypha, short conidio Viores,

polyblastic conidiogenous cell, and conidial chains (slide culture, X 1 000). Fig. 2, lateral

protuberances ( = short conidiophores) on vegetative hyphae. Note the terminal and lateral

budding by the two-septate conidium (slide culture, X 1 000). Fig. 3, cicatrised, two-septate

conidiophore (x I 000). Fig. 4, sympodially-branched coidiophore (x 1 000). Fig. 5, conidia

( x 2 000). Fig. 6, one-septate, thick-walled conidium with lateral protuberance ( x 2 000).

Fig. 7, polyblastic conidiogenous cell ( x 2 000). Fig. 8, two-septate, thick-walled, brown

conidium (x 2 000).

certain dematiaceous genera with catenate spores such

as Polyscytalum Ries, Septonema Corda and parti-

cularly C/adosporium Link ex Fr., as defined by De

Vries (1952). According to the keys of Ellis ( 1971 ), the

determinate or percurrent conidiogenous cells of

Polyscytalum and Septonema distinguish these two

genera from Phaeoramularia and Cladosporium with

sympodial conidiogenous cells. The latter two genera

are distinguished by the presence of a small, distinctly

protuberant scar at each end of the spore in C/ado-

sporium and the absence of these scars in Phaeora-

mularia (Ellis, 1971). It should be pointed out,

however, that in at least one species of Cladosporium ,

C. resinae (Lindau) de Vries, the conidial scars are not

prominent. The new species described here is placed in

Phaeoramularia because of the lack of distinctly

protuberant conidial scars. It is, however, closely

related to and somewhat difficult to distinguish from

C. resinae as described by De Vries (1955) and Par-

berry (1969).

In C. resinae the long conidiophores often bear

irregularly scattered warts and are almost always of

two kinds — some bearing long chains of conidia and

others solitary conidia (or very short chains) on

swollen brances of conidiogenous cells which secede

with difficulty (M. B. Ellis, 1972, personal commu-

nication). These characteristic conidiophores are not

found in P. kellermaniana which also has slightly

longer, more elongated-ellipsoidal spores that tend to

become septate.

Within the genus Phaeoramularia , P. kellermaniana

differs from the five described species (Muntanola,

1960; Sutton 1970) in the following respects:

1. All the other known species are leaf parashes of

higher plants. P. kellermaniana is apparently a sapro-

phyte which has been isolated from decomposing

wheat and oat straw and from the external surface of

lucerne seeds. According to Dr M. B. Ellis (personal

communication, 1972) this fungus also occurs on

Hordeum and stored grain in Turkey and Pakistan.

2. The conidial chains of P. kellermaniana are

much longer (up to 30 or more spores/chain) than

those described for the other species of Phaeoramularia.

The conidial chains of P. kellermaniana are identical

to those of Cladosporium in method of development,

tend to be persistent and are true chains as defined by

Subramanian (1972).

3. P. kellermaniana grows rapidly in culture, the

growth rate being 6-9 mm/day depending on the

medium. In contrast, very slow growth rates have been

reported for the two other species of Phaeoramularia

that have been grown in culture: — 2-3 mm diam.

after 16 days in P. paradoxa Muntanola (Muntanola,

I960) and 2,5-3mm diam. after 20 days to 20 mm

after 60 days in P. unamunoi (Castell.) Muntanola

(Muntanola, 1954; Kovachevsky, 1938). In this

respect, P. kellermaniana also appears to be more

closely related to the saprophytic Cladosporium spp.

than the phytopathogenic members of the genus

Phaeoramularia.

W. F. O MARASAS AND INGRID H. BREDELL

219

P. kellermaniana is named after Dr T. S. Kellerman

of the Veterinary Research Institute, Onderstepoort,

who collected the sample of wheat straw from which

the type strain was isolated.

ACKNOWLEDGEMENTS

We are indebted to Dr M. B. Ellis, Commonwealth Mycologi-

cal Institute, Kew, England, for suggesting the genus

Phaeoramalaria , valuable advice and critical reading of the

manuscript; Dr G. C. A. van der Westhuizen, Plant Protection

Research Institute, Pretoria, for unfailing interest and advice;

Mr A. M. du Bruyn, Veterinary Research Institute, Onderste-

poort, for assistance with the photography.

OPSOMMING

'n Hifomiseet wat geTsoleer is uit koring- en hawerstrooi

sowel as lusernsaad in Suid-Afrika, word beskryf as Phaeora-

mularia kellermaniana Marasas & Bredell sp. nov. Die verwant-

skappe van P. kellermaniana met Cladosporium resinae (Lindau)

de Vries en ander spesies van Phaeoramularia word bespreek.

REFERENCES

De Vries, G. A., 1952. Contribution to the knowledge of the

genus Cladosporium Link ex Fr. Baarn: Centraalbureau

voor Schimmelcultures.

De Vries, G. A., 1955. Cladosporium avellaneum de Vries a

synonym of Hormodendrum resinae Lindau. Antonie van

Leeuwenhoek 21:1 66- 168.

Ellis, M. B., 1971. Dematiaceous Hyphomycetes. Kew: Com-

monwealth Mycological Institute.

Kovachevsky, I. C., 1938. Die Braunfleckenkrankheit der

Paprikapflanze, Cladosporium capsici (March, et Stey.)

n. comb. Z. PflKrankh. 48 : 321-336.

Marasas, W. F. O. & Bredell, Ingrid H., 1973. Mycoflora of

South African lucerne (Medicago sativa L.) seed. Phyto-

phylactica (In Press).

Marasas, W. F. O. & Schumann, Ingrid EL, 1972. The genus

Pithomyces in South Africa. Bothalia 10: 509-516.

Muntanola, Maria, 1954. A study of a newly-identified

pepper disease in the Americas. Phytopathology 44: 233-239.

Muntanola, Maria, 1960. Algunos Hyphomycetes criticos.

Notas y descripciones. LiUoa 30: 165-232.

Parberry, D. G., 1969. Amorphotheca resinae gen. nov., sp.

nov.: The perfect state of Cladosporium resinae. Aust.

J. Bot. 17:331-357.

Ridgway, R., 1912. Color standards and color nomenclature.

Baltimore: Hoen.

Subram an i an, C. V., 1972. Conidial chains, their nature and

significance in the taxonomy of Hyphomycetes. Curr. Sci.

41 :43-49.

Sutton, B. C„ 1970. Forest microfungi IV. A leaf spot of

Populus caused by Cladosporium subsessile. Can. J. Bot.

48:471-477.

17749-2

Bothalia 11,3: 221-230 (1974)

A reappraisal of type and authentic material of the larger

Basidiomycetes in the Pretoria Herbarium

DEREK A. REID-

ABSTRACT

Accounts of the macroscopic appearance and microscopic structure of twenty type and authentic

specimens of the larger Basidiomycetes preserved in the Pretoria Herbarium are given. The species

concerned have been reassigned wherever possible to genera in accordance with modern taxonomic

concepts, and the following new combinations made: Antrodia conchata (Lloyd) Reid, Coriohis

durbanensis (Van der Byl) Reid, Ganoderma nigroluciduin (Lloyd) Reid, Ramaria cladoniae (Kalchbr.)

Reid.

This paper is an account of type and authentic

specimens of the larger Basidiomycetes in the

Pretoria Herbarium, exclusive of those species

described by, or reported on, by both Wakefield

and Talbot, and for which full and accurate diagnoses

already exist. It is a supplement to my paper (Reid,

1973) reporting the results of a study of type and

authentic material in the Van der Byl Herbarium,

Stellenbosch.

When that paper was written attention was drawn

to the many discrepancies in the numbering of Van

der Byl’s specimens in STE as compared with the

published data, and it was suggested that this may

have been due to a partial renumbering by Van der

Byl of his herbarium when he moved to Stellenbosch.

However, specimens bearing the published numbers

were subsequently found in PRE and. although these

collections are probably part of the gatherings

studied at STE, one cannot be sure without having

them side by side for comparison. For this reason L

have provided detailed accounts of these specimens

in PRE since they are either holotypes or isotypes.

In most instances there is virtually complete agreement

with my published descriptions of the STE material-

an exception involves Trametes albotexta in which

the spores of the isotype in PRE are appreciably

broader than those found in the STE material. This

is probably only a question of maturity.

In addition to the descriptions of these undoubted

holotypes or isotypes of species already studied at

STE, this paper includes accounts of similar material

of species described by both Van der Byl and Lloyd

which are not represented in STE. Modern descrip-

tions are also provided for three species published by

Kalchbrenner, the isotypes of which are preserved in

PRE, viz. Daedalea macowanii , Polyporus vibecinus

var. antelopum , and Clavaria cladoniae.

In most instances attempts have been made to

reassign the taxa to genera in accordance with current

taxonomic concepts.

The format follows that adopted in my previous

paper; the original description is reproduced followed

by my own account of the material. For those species

described by Lloyd, where the diagnosis is often brief,

I have combined the original description with that of

Van der Byl, who had often seen the specimens in

the fresh condition. In such cases Van der Byl’s

comments have been interspersed between brackets,

following on Lloyd's observations. For the species

described by Kalchbrenner the Latin diagnosis is

reproduced followed in turn by a rough translation

and then by my own observations.

* Royal Botanic Gardens, Kew, England.

ACKNOWLEDGEMENTS

It is with pleasure and gratitude that the author

wishes to thank the trustees for the award of the

Smuts Memorial Fellowship which made this work

possible. He also wishes to thank Prof. E. A.

Schelpe, University of Cape Town, who was respons-

ible for suggesting and helping to organize my visits

to South Africa to study the Basidiomycete flora.

Grateful thanks are also due to the Director of the

Botanical Research Institute, Pretoria, for allowing

me to work in his Institute and for the innumerable

kindnesses shown to me by him and his staff. It is

also a pleasure to acknowledge the help and friend-

ship so generously given to me by Dr and Mrs B. de

Winter, and Drs G. C. A. van der Westhuizen and

W. F. O. Marasas.

APHYLLOPHORALES

Daedalea macowanii Kalchbr. ex Thnem. in Flora

59: 362 (1876).

“Sessilis, suberosus, extus intusque concolor albidus et

demum alutaceus. Pileus difformis primum tuberis instar e

matrice egrediens, in optima evolutione semiorbicularis, con-

vexus, 3-4" longus latusve, ad basin -j-2" crassus, margine

attenuato, obtusiusculus, azonus, laevis, tomentoso-glabratus,

tectu mollis sed mox nudus. Pori majusculi, labyrinthici,

acie obtusi, marginem versus valde elongati-paralleli. Caro

intus zonata. — In cortice vivo Celtidis rhamnifoliae aliorumque

arborum ad ripas fluvii “Klyn Vischrivier”. Promont: bonae

spei. 1875. Leg. P. MacOwan (No. 1066).

“Sessile, suberose, concolorous outside and in, whitish be-

coming alutaceous. Pileus deformed, at first tuber-like as it

emerges from the matrix, when better developed semi-orbicular,

convex, 3-4 inches long and wide, at the base )— 2 inches thick,

with attenuated margin, obtuse, azonate, smooth, tomentose-

glabrous, surface soft but soon naked. Pores rather large,

labyrinthoid, with obtuse edge, toward the margin strongly

elongate-parallel. Flesh internally zonate. — On living bark

of Celtis rhanmifolia and other trees on the banks of the Klein

Visrivier, Cape of Good Hope, 1875. leg. P. MacOwan (No.

1066).”

This is typical of the very common tropical fungus

usually known as Lenzites palisotii (Fr.) Fr. and

determined as such by P. H. B. Talbot in a note on

the type sheet. It is, however, best left in the genus

Daedalea Pers. ex Fr. to which it was originally

referred as D. palisotii Fr.

An interesting feature concerning members of the

genus Daedalea sensu stricto is that they have the

interior of the tubes lined by a palisadic layer of

thick-walled hyphal endings which could be inter-

preted as very poorly differentiated cystidioles. The

basidia, when formed, grow up through this palisade,

222

LARGER BAS1DIOMYCETES IN THE PRETORIA HERBARIUM

Polyporus arenosobasus Lloyd in Mycological

Writings 6: 919, PI. 144, fig. 1643 (1920). (Fig. 1).

“(Plants annual, stipitate). Pileus about three inches in dia-

meter globose with a thin cuticle (pileus circular, 8 cm diam.;

surface with a thin pellicle). Color (dried) dark, fuliginous (fawn-

coloured changing to sepia, rugulose). Flesh firm, pale isabel-

line (context 0,1 to 1 cm, fibrous, corky, firm, light yellowish).

Stem short, prolonged underground into a sclerotium-like

body of agglutinated sand (stalk arising from a false sclerotium

of agglutinated sand; true stalk short, about 1,8 cm diam.;

false sclerotium 5 cm long by 3 to 4 cm diam.). Pores small,

round, trametoid (tubes 1 to 3 mm long, decurrent on stalk;

mouths subtrotund to irregular, 3 to 4 to the mm; edges thin,

entire. Spores not found; hyphae 6 g to 7,5// diam.).

This must be rare as I judge that Mr Van der Bijl found only

one specimen as he sent me only half a specimen. It should be

added to Section 38 of Ovinus. Another similar plant has been

collected in Africa, Polyporus goetzii (Cfr. Synopsis, Ovinus,

page 74, Fig. 496) but that has a true sclerotium and large

pores. The false sclerotium of this is only sand-agglutinated

and the only other known specimen with suc^ a body is Poly-

porus tuheraster. Fig. 509 in our Ovinus pamphlet.

(Distribution. — A single collection at Durban, Natal, by

J. B. Leslie. Type in Natal Herbarium, No. 710.

The plant differs from others recorded from South Africa in

its false sclerotium. Polyporus tuberaster Pers., a native of

Europe, has a similar false sclerotium. Here, as in P. arenoso-

basus, the false sclerotium is composed of sand particles

cemented together into a hard body by the mycelial threads

of the fungus.)”.

The specimen in PRE bears the following data:

Durban, 6 Aug. 1917 [Herb. PRE 14022]. There are

also a few' fragments in Herb K of a specimen bearing

the Van der Byl No. 710 and this appears to be

equated by Doidge (1950) with the Pretoria collection

No. 14022 and also with the specimen No. 56364 in

the Lloyd Catalogue which is the type number of

P. arenosobasus. This would tie up with Van der

Byl’s statement in 1922 that only a single collection

was known at that time. However, judged from

Stevenson & Cash (1936) there appear to be two

collections in the Lloyd herbarium and they cite the

specimen with the Lloyd Cat. No. 56364 as type and

equate the specimen bearing the Van der Byl number

710 with the other specimen in the Lloyd Herbarium

with the Cat. No. 56386. This does not agree with

the Van der Byl observation of a single collection

prior to 1922. A further reason for regarding the

PRE specimen as isotype is that Lloyd (1920)

mentions that only part of the fruitbody was sent

to him by Van der Byl and the specimen in PRE

consists of only ^ of a fruitbody in two slices. These

match the half specimen illustrated by Lloyd so

closely in dimensions that they must surely be part

of the same sporophore.

The pileus is smooth, fawn to sepia in colour,

with a poorly defined cuticle. The stipe, about 1 cm

long, expands below ground into a long false sclero-

tium, 5 cm in length and 3,5 cm wide, formed of

agglutinated sand grains. The pores, 3-4 per mm,

are ochraceous, angular with thin dissepiments; the

tubes, 5 mm long, are ochraceous fawn; and the

flesh, similarly coloured, is soft and cottony, up to

6 mm thick at the centre of the pileus, thinning to

1,5 mm at the incurved margin. Anatomically the

context is probably dimitic but generative hyphae

were not actually seen; the skeletal hyphae, 4,5-

8,0/x diam., are long, seldom branched, with dis-

tinctly thickened, strongly dextrinoid walls, but they

retain a very wide lumen which may be secondarily

septate on rare occasions. There are also a few narrow,

branching hyphae but these may be the rare arbori-

form endings of skeletal hyphae. Spores were not

seen.

In the Kew fragments the hyphal structure was

still somewhat uncertain but there appeared to be

occasional, very collapsed, short lengths of thin-

walled septate non-dextrinoid hyphae in the tubes

with clamp-connections, but this needs confirmation

on better material. In addition very collapsed and

fairly numerous thin-walled, hyaline, non-dextrinoid,

very broadly elliptic to ovate spores were observed

with a range of 6,2-7,2x4,75-5,75//..

This species is best left in Polyporus [(Mich.) Fr.]

Fr. The fact that the hyphae are strongly dextrinoid

is not seemingly of generic significance for currently

species are admitted to Perenniporia Murr. with

both dextrinoid and non-dextrinoid flesh. However,

with accumulation of additional data concerning

tropical polypores it may in the future be necessary

to refer this species to some other genus.

Polyporus conchatus Lloyd in Mycological Writ-

ings 5; Mycological Notes 49, p. 700, fig. 1048

(1917). (Fig. 2).

‘Largely resupinate, but with reflexed pileus, conchoid (Plants

effused reflexed to entirely resupinate; pileus conchate, 1 cm

to 6 cm by 0,5 cm to 1,5 cm by 0,1 cm to 0,5 cm). Color (also

of context) pinkish-buff. Surface dull mat. (Surface pale buff,

becoming darker, uneven; context firm, hard, buff with pinkish

spots, or pinky tint throughout, 1 mm to 4 mm). Pores medium,

round or elongated on portion growing vertical (tubes, 0,5

mm to 1,5 mm, concolorous; mouths subrotund to irregularly

angular, unequal, 2 to 3 to mm elongated on effused portion).

Cystidia none. Spores abundant, 4-5 8-10, hyaline, surface

uneven. (Spores hyaline 3,7// to 5,5// by 7,4// to 9,3//; hyphae

4// to 5// diam.; setae none).

This belongs in section 91 of the Apus Polyporus, and closely

related to Polyporus rugoso-porus. The color is entirely different.

( Distribution . — Found by the writer at Klapmuts, Cape

Province, on stump of Populus and known only from this

collection. The pinkish buff colour of context is peculiar. The

plant would also be looked for under Trametes. (Cotype in

Natal Herbarium, P.v.d.B., No. 358.)

The specimen in PRE bears the number 385 but

the information on the packet otherwise agrees with

the published collection data; evidently there has

been an error at some time in transcribing the figures.

The specimen was collected in Dec. 1916.

Sporophores vary from completely resupinate

patches to small effuso-reflexed brackets measuring

only 0,6 cm from the point of attachment to the

margin. There is little evidence of pink colour in

any of the fruitbodies now — the colour is pale

ochraceous bulT with the surface of the brackets

becoming almost blackish in the older parts. The

pores, 2-3 per mm, are round, angular or irregular

with thick dissepiments. Anatomically the context is

seemingly dimitic, comprising thick-walled, almost

solid, glassy skeletal hyphae, 3-5/t diam., and a few

remnants of thin-walled generative hyphae with

clamp-connections at the septa. Spores abundant,

hyaline, elliptical, 7,5-1 1 ,5 (-12,0) x 3, 2-4,0 (—4 , 2)/x.

This species would seem to belong in Antrodia

Karst, sensu Donk (1966), and shows remarkable

similarity to A. serialis (Fr.) Donk in regard to spore

size and the pink colouration of the flesh (see Donk,

1971). However, A. serialis is said to be restricted to

coniferous substrates so that one cannot feel justi-

fication for assigning the South African fungus to

synonymy under this species. Hence I propose the

following new combination : Antrodia conchata (Lloyd)

Reid, comb. nov.

Polyporus durbanensis Van der Byl in S. Afr. J.

Sci. 18: 261-262 (1922). (Fig. 16 a-c).

Polyporus griseus Boes. in Annls mycol. 10, 494 (1912), non

P. griseus Peck (1874).

Trametes various Van der Byl in S. Afr. J. Sci. 18: 281-282

(1922).

DEREK A. REID

223

Polyporus pectunculus Lloyd in Mycological Writings 7:

1315 (1924), non P. pectunculus Lev. (1846).

“Plants annual, sessile or effused reflexed; pileus dimidiate,

imbricate, applanate to conchate, laterally connate, 1 cm to

3 cm by 0,5 cm to 2,5 cm by 0,1 cm to 0,3 cm, coriaceous,

tough, becoming firm and rigid in drying; surface pearl to lead

or ash grey, finely tomentose pruinose, undulating, smooth

becoming scabrid and fuliginous in places; context 0,5 mm to

1,5 mm white to discoloured, firm, corky; tubes 0,5 mm to

1,5 mm long; mouths irregular, angular, 4 to 5 to the mm;

edges thin, entire, grey changing to yellowish; spores hyaline,

globose to oblong, 4p diam.; hyphae 4p to 5,5//.

Distribution. — Known only from around Durban where it

was collected on dead logs by the writer. (Type in Natal Elerb.

P.v.d.B., No. 896).”

In a previous paper I (Reid, 1973) ventured the

opinion that P. durbanensis belonged in the genus

Coriolus Quel, but this view was based on a study of

a specimen in STE collected by Van der Byl at

Durban and named by him. However, it was not type

material and I indicated that before any transfer

was made it would be necessary to study the type

collection. This has now been completed and the

results confirm that the species does in fact belong

in Coriolus. A description of this collection follows:

Specimen No. 896, Durban (PRE 14021).

Sporophores consisting of a series of imbricate

variously confluent brackets forming a compound

fructification about 5 cm in length and up to 3 cm

from the point of attachment to the margin. These

brackets are rigid and up to 2,5 mm in total thickness.

The surface varies from pale ochraceous fawn or

buff to grey-brown near the margin, and is not really

tomentose. Toward the grey-brown marginal zone

the pileus appears closely and densely radially streaky

due to small raised fibrils (s.l.), but further back has

a rough scrupose texture due seemingly to a pruinose

pubescence. There is no real cuticle. The flesh is pale

iignicolorous with darker, more or less horny lines;

it has a fibrous texture but is inclined to flake away

rather than to pull apart in a fluffy manner. The pores

are small, with thin papery dissepiments, angular,

and rather dark ochraceous brown. Tubes about 1 mm

deep. Anatomically the context at the base of the

fruitbody has a peculiar hyphal construction since it

is composed mainly of thick-walled hyphae (often

with a lumen occupying about | the width of the

hypha) which, although consisting of a very long

unbranched axis, develop numerous short side

branches of limited growth in some instances; fur-

thermore the main trunk bears clamp-connections.

In other similar hyphae, although clamps are also

present, side branches may be few or absent. The

side branches vary from very short and peg-like to

better-developed hyphae. Elsewhere the context

seems to be typically trimitic comprising more or

less solid skeletal hyphae. 4-6// wide; coralloid

binding hyphae, 2,5// wide forming indistinct tangled

complexes; and branched, hyaline, clamp-bearing

generative hyphae, 2,5// wide. Although a hymenium

was present, no spores were found.

This species is hereby transferred to the genus

Coriolus Quel, as Coriolus durbanensis ( Van der Byl)

Reid , comb. nov.

In PRE there is another collection of C. durba-

nensis'. Drummond, Natal, leg. W. G. Rump, June

1932 (PRE 26845). This consists of two portions.

One example (Fig. 16 a) comprises a well-developed

bracket together with several rudimentary imbricate

pilei; the entire compound fruitbody is 10 cm long

and 2 cm from the point of attachment to the margin.

The surface is buff or fawn becoming darker yellowish

brown or brown tow'ard the margin, which is quite

conspicuously radially ridged; the surface is minutely

pruinose-pubescent and under a lens appears to have

a structure consisting of a matted but very short

tomentum. The pores are small and grey-brown, but

there is a broad, sterile, obtuse margin. In texture the

Iignicolorous flesh is hard and fibrous, without horny

lines. Anatomically the context is seemingly trimitic

in the thick basal portion comprising solid, glassy

skeletal hyphae, up to 6// wide, which taper to the

apex; generative hyphae up to 2,5// wide, with

clamp-connections at the septa, and slightly thickened

walls; and what appear to be narrow binding hyphae

which are not coralloid but more elongated, solid,

2,5// wide. Clearly the trimitic structure is not so

obvious as in the type material but nevertheless the

specimen seems to be referable to C. durbanensis.

A few hyaline, elliptical spores were found in this

specimen, measuring 4, 5-6, 0x2,0//.

The second fruitbody (Fig. 16 b) consists of a series

of imbricate brackets with a vertical disposition, one

above the other. In this fruitbody the marginal area

of the brackets is dark blackish brown and very con-

spicuously ridged, but elsewhere the surface is inclined

to be rough and scrupose. In this fructification the

structure is more like that of the type. There are many

broad skeletals, 6// wide, but also other skeletal-like

hyphae with clamp-connections and, in addition,

branched thick-walled hyphae some of which appear

to be binding hyphae but others bear clamps and

are presumably thick-walled generatives.

The anatomy of these two fruitbodies would seem

to indicate that in this species there is a considerable

degree of plasticity of hyphal structure with almost

any of the hyphae able to develop clamp-connections,

at least under certain conditions.

As regards Polyporus pectunculus Lloyd (non P.

pectunculus Lev.), Doidge (1950) would seem to cite

the collection under P. durbanensis in such a way as

to indicate that the types of this species and of P.

pectunculus were based on the same gathering, “Stella

Bush, v.d. Byl 896, Type (Lloyd Myc. Coll. 23901)

14021”. Stevenson & Cash (1936) cite specimen No.

23901 as the type of P. pectunculus Lloyd but they do

not give any data to suggest that this is Van der

Byl’s specimen 896! Whether or not P. pectunculus

Lloyd is an obligate synonym of P. durbanensis, it is

interesting to note that Van der Byl (1928) regarded

the two species as identical. This has been confirmed

by a study of the type collection of P. pectunculus in

BPI, which has a trimitic hyphal structure but also

shows the presence of broad skeletal-type hyphae

with clamp-connections in the basal portion of the

fruitbody. The surface pruinose-pubescence is formed

of erect, thin-walled, clamped generative hyphae.

Spores are also present in this material although

mostly collapsed; they are hyaline, elliptic, 5,2-5,75 x

1 ,75-2,0/t. A more detailed account of P. pectunculus

Lloyd will be given in a later paper in this series.

Polyporus flexilis Van der Byl in S. Afr. J. Sci.

18: 271 (1922).

A study of the type collection from “Maritzburg,

on dead log, det. P. A. v. d. Byl, No. 810(PRE 14031)”

has shown that this species is probably the same as

Funalia protea var. imbricata (Berk.) Reid, as were

the specimens studied in STE, see Reid (1973) to

which description 1 have nothing to add.

224

LARGER BASIDIOMYCETES IN THE PRETORIA HERBARIUM

Polyporus mollicarnosus Lloyd in Mycological

Writings 4; Letter 60, p. 11 (1915). (Fig. 3).

Ganoderma mollicarnosum (Lloyd) Sacc. & Trott. in Syll.

Fung. 23: 401 (1925).

“Pileus sessile, a foot or more in diameter (Pileus annual or

persisting a second season, circular, 38 cm in diam., concave

above). Surface not laccate, pale buff color, smooth, soft to

touch. (Surface smooth, pale buff to bluish-green, smooth,

undulating). Context very soft and spongy, light color, varying

from buff to isabelline. (Context soft, spongy, floccose, buff

to yellowish-buff, I cm to 5 cm thick). Pores small, round, an

inch or more long, with concolorous mouths. (Tubes long,

0,5 cm to 3 cm, buff coloured; mouths reseda green, irregular,

angular to elongated). Spores 8 x 12-14, very pale color, smooth.

Spores (teste Lloyd) smooth, very pale colour, truncate, 8// by

12// to 14//; hyphae 4// to 11//.

Distribution. — A single specimen collected at Durban, Natal.

This plant impresses me as exceptional among the Gano-

dermus section in its very soft flesh and pale spores. I know of

no other with such flesh excepting Polyporus colossus , which

has a differently colored context, and larger spores. I would

class it in Section 102 of my recent Apus Polyporus pamphlet.

It is an evident annual and probably of rapid growth. Type 58

from I. B. Pole Evans, South Africa.

(In its soft spongy context the fungus resembles P. colossus ,

but differs from it in its habit, surface, long pores, spore

characters, and colour of context.

It is evidently normally an annual, though in part a second

pore layer had formed over the first)”.

Sporophore circular, 38 cm diam., occasionally

lobed at the obtusely rounded margin, with an un-

evenly undulating and radially folded surface and a

very thin papery cuticle. The fruitbody is centrally

stipitate; the stipe being 8 cm long and 12 cm diam.

The pores, 2-3 per mm, are angular with thin dis-

sepiments and the tubes up to 3,5 cm long; in part

there is a hint of indistinct stratification (but this is

probably an illusion). The flesh is about 3 cm thick

in the pileus but much thicker in the centre where it

continues into the stipe; it has a soft cottony texture

and is ochraceous fawn becoming yellowish brown in

the stipe. Anatomically the context is dimitic, con-

sisting of strongly dextrinoid skeletal hyphae, 6-1 l/i

wide, with slightly to distinctly thickened pale straw-

coloured walls. These hyphae are mostly unbranched

and taper to an elongated thin-walled hyaline apex;

however there may be very occasional branching.

No generative hyphae were seen, as these had pre-

sumably collapsed. Spores were very scanty although

eventually found at the mouths of the tubes near the

base of the fruitbody. They are broadly elliptical,

hyaline, 10,2-12,0x6,0//..

This species is liable to confusion with P. baudortii

Pat. (syn. Phaeolus manihotis Heim) but the latter

has appreciably smaller spores and the skeletal

hyphae of the context are both narrower and non-

dextrinoid. P. mollicarnosus is left in Polyporus

[(Mich.) Fr. ] Fr. for the time being; it was erroneously

referred to Ganoderma Karst, by Saccardo & Trotter.

Polyporus niTro’ncidus Lloyd in Mycological

Writings 6: 925, PL '46, Fig. 1664. (1920). (Fig.

4a-b).

In this instance Lloyd’s description is so vague it

is reproduced separately as follows:

“The European form of Polyporus lucidus was not even

suggested to us when we first saw this specimen but when we

came to “analyze” it we concluded that it is only a “species

form". Plant is mesopodial with the surface shiny and jet

black. Trama, pores and pore mouths isabelline. The interior

of the stem is so soft that it could be called pithy. Spores

(i 8 of the usual type. There is a black form of Polyporus

lucidus in Japan that corresponds to the European plant except

as to color. This does not exactly agree in any feature,”

Van der Byl’s (1922) description is more lucid:

“Plants terrestrial, annual, stalked; pileus depressed at the

centre, rarely reniform, 4 cm to 9 cm diam. by 5 cm, surface

black, laccate, rugulose; context white to brown, 1 mm to 3 mm

thick, soft; tubes 1 mm to 3 mm long, white within; mouths

subrotund to angular, 5 to 6 to the mm; edges thin, entire, white,

becoming discoloured; spores coloured, apiculate, truncate,

4// to 5,5// by 6// to 7,5//; hyphae 5,5// to 7,5//; stalk central,

rarely somewhat excentric or lateral, and deeply rooted. 12 cm

to 20 cm long by 5 mm to 7 mm diam., interior soft.

Distribution. — Found at Durban, Natal, by J. M. Wood,

and collected here also by the writer.

Distinguished by black laccate surface of pileus and stalk.

(Type in Natal Herbarium, No. 412). In young specimens surface

of the pileus is frequently dark chestnut brown in part and

in a few specimens the main stalk branched below the soil.”

There is some uncertainty concerning the status of

the collection in PRE (No. 13941, ex Van der Byl),

but the latter expressly stated that the type was in

the Natal Herbarium, and since the mycological col-

lections from that herbarium are now in PRE it

seems reasonable to conclude that it is indeed the

specimen referred to as type by Van der Byl. Reference

to Doidge’s (1950) list of South African Fungi and

Lichens is confusing since it is not clear whether she

is indicating that Van der Byl No. 900 is to be equated

with the specimen in the Lloyd collections, with the

Lloyd Catalogue No. 22888, and hence to be regarded

as type rather than specimen No. 13941 in PRE.

Certainly Stevenson & Cash (1936) merely quote

Lloyd’s collection No. 22888 from Van der Byl as

type without quoting any collector’s number.

The following notes are taken from collection No.

13941 in PRE: This consists of two fruitbodies with

circular, umbilicate, concentrically sulcate and radially

plicate pilei, the surface of which is laccate and of

a dark chestnut to almost black. The stipe likewise

has a black shining crust and, although broken, is up

to 6 cm long and up to 1 ,2 cm wide with pale ligni-

colorous flesh. The pores, 5 per mm, are minute,

angular, with thin dissepiments, and creamy fawn.

The flesh of the pileus is similarly coloured but more

ochraceous. Anatomically the surface of the pileus

consists of a cuticle formed of a palisade of very

short, clavate or capitate elements, but with very

slightly thickened walls. These elements measure

13-20x8-1!//,. In structure the flesh appears to be

dimitic, comprising skeletal hyphae, up to 5// wide,

with thin, pale straw-coloured dextrinoid w'alls but

retaining a wide, often abundantly secondarily septate

lumen. Some of these hyphae branch toward their

apex. Spores are frequent, 8,0-9,75x6,2-6,5//, and

typical of those found in the genus Ganoderma Karst,

to which genus this species is now transferred as

Ganoderma nigrolucidum (Lloyd) Reid , comb. nov.

The distinguishing features appear to be the

centrally-stalked fruitbody with dark laccate crust,

the pale flesh, the small size and thin-walled nature

of the cuticular elements, and the small pores.

Polyporus oehroporus Van der Byl in S. Afr. J.

Sci. 18: 269 (1922).

The type material from the Eastern Cape Forest

Conservancy, coll. Van der Byl (No. 1 15) in PRE has

almost been destroyed by insects. Both surface and

pores have been eaten and only two portions of

context remain. Examination of these shows the

hyphal anatomy to be as described for this species

in my previous paper (Reid, 1973).

DEREK A. REID

225

It should be noted that Doidge’s (1950) citation is

again confusing, for Van der Byl clearly stated that

the type No. 115 was in the Natal Herbarium and

came from the Eastern Cape Forest Conservancy,

collected by himself. There is nothing on this specimen

in PRE to suggest it came from Branders High

Forest and was collected by Keet. Neither is there

any suggestion of this data on the Kew specimen

which merely bears the following information “South

Africa, Van der Bijl; Don. C. J. Humphrey, July

1932. Lloyd Myc. Coll. Cat. No. 18843”. Just how

many collections are included in the one citation by

Doidge, and all apparently equated with the type

No. 115, is far from clear.

This fungus is now widely known as Inonotus

ochroporus (Van der Byl) Pegler.

Polyporus rusticus Lloyd in Mycological Writings

5, Mycological Notes 53, p. 751, Fig. 1124, (1918).

(Figs. 5 a-b).

The isotype collection in PRE: On dead pine,

Klapmuts, Cape, December 1916, No. 387 (PRE

13940), differs in no significant detail from the

specimen 285 in STE as described in my previous

paper (Reid, 1973). It is similar in macroscopic

appearance, hyphal structure, and in having hyaline,

elliptic spores, measuring 7, 5-9, Ox 3, 5-4,0/m. This

species is the same as Funalia protea var. imbricata

(Berk.) Reid.

Polyporus trichiliae Van der By/ in S. Afr. J. Sci.

18: 262 (1922). (Fig. 7 a-b).

Examination of the holotype in PRE: On Trichilia

emetica , Durban, No. 897 (PRE 14032), confirms that

this species is a synonym of Xylodon subiculoides

(Lloyd) Reid as suggested in my previous paper

(Reid, 1973). However, as the generic name Xylodon

P. Karst, has been superseded by Schizopora Velen.,

the fungus should be cited as: Schizopora subiculoides

{Lloyd) Ryvarden.

Polyporus vibecinus var. antelopum Kalchbr. in

Grevillea 10: 53 (1881).

“Pileo vix virgato, margine exciso-lobato, lobis passim

dichotomis; poris nudicoribus, reticulatis.

P. Natal (1. Wood, No. 99).”

“Pileus scarcely virgate, margin excised-lobate, lobes here

and there dichotomous; pores rough*, reticulate.

P. Natal (coll. Wood, No. 99).”

The isotype in PRE: Inanda, Natal, coll. J. M.

Wood, 1877/1888 No. 99, consists of six fruitbodies.

These are all dimidiate or spathulate and narrowed

behind into a short dorsiventrally flattened stipe-like

base. The upper surface is purplish brown, densely

radiately striate or virgate, and covered toward the

base with a sparse, discontinuous tomentum which

is visible only under a lens, and extends in progres-

sively more sparse form to about the centre of the

pileus. The margin is slightly laccerate but not lobed.

The pores are ashy-buff, angular and elongated

radially (favoloid) and 2, 5-3,0 per mm; the dis-

sepiments are either entire or laccerated and toothed.

Anatomically the structure agrees with that of the

var. vibecinus and is dimitic, comprising non-dextri-

noid, very elongated, thick-walled to almost solid,

glassy, unbranched skeletal hyphae, which at their

widest portion are 5-8/m diam., but taper at either

end to about the generative hyphae have collapsed

and disappeared.

* There appears to have been an error in the printing and it

seems likely that “rudioribus” was intended instead of “nudi-

coribus”. The former would presumably refer to the almost

toothed or laccerated dissepiments.

This taxon is in no way different from P. vibecinus

Fr., which is itself a synonym of Favolus spathu/atus

(Jungh.) Bres., and was listed as such by Doidge

(1950).

Trametes albotexta Lloyd in Mycological Writings

5; Mycological Notes 44: p. 614, fig. 868 (1916).

(Fig. 6).

“Pileus sessile, 2-3 x4—5 inches, an inch thick, surface reddish

brown, dull, matt, soft. (Pileus sessile, dimidiate, 8 cm to

10 cm by 5 cm by 1 ,4 cm to 4 cm; surface reddish brown, soft

to hard, rugulose, tomentose to subglabrous, smooth or scabrid

to touch). Context reddish brown, thin, soft, (context 0,4 cm to

0,7 cm, firm, tough, corky, reddish brown). Pores rigid, small,

round, an inch long. Pore tissue white, contrasting with the

brown hymenium so that a section is variegated, (tubes 0,5 mm

to 1,5 mm long, tissue white with reddish-brown hymenium;

mouths subrotund to angular, 2 to 3 to mm, edges entire or

produced into teeth, white changing to yellowish and dark

brown). Cystidia none. Spores pale brown, small, elliptical-

subglobose, 3-4x4— 5. (Spores pale brown in mass, subglobose

3,7// by 5//; hyphae 4// to 5//).

The coloration of the surface, pore surface and context is

reddish brown, the tissue of the pores white. I know of no

other Trametes or polyporoid with this peculiar color pore

contrast. The pale-colored spores might be the basis for a

“new genus”, but I think that would be “inutile”.

( Distribution . — On dead Podocarpus sp. in Hlatikulu Forest .

Type in Natal Herbarium, No. 181. Collected also by Miss

A. V. Duthie.

The surface and context colour are peculiar and would aid

in the recognition.)”

The isotype in PRE: On dead Podocarpus , Hlatikulu

Forest, Swaziland, 11 May 1915, No. 181 (PRE

15568), is very similar indeed to the specimen in

STE which I have already described (Reid, 1973),

but there is an appreciable difference in spore size

and shape. I can only conclude that the spores in the

STE collection are immature. Because of this dis-

crepancy I have reproduced my notes on the PRE

material in full :

Sporophores with a very dark red-brown to blackish

surface, consisting of alternating bands of tomentum

and glabrous zones. The tomentum appears rather

scrupose, especially under a lens. The pores, 2-3 per

mm, are rounded to irregular, pale brownish (due to

the colour of the hymenium), with thick buff-coloured

dissepiments. The tubes, up to 10 mm long, are lined

with the pale brown hymenium. Toward the base of

the fruitbody the flesh is up to 1 ,5 cm thick but is

only 3 mm thick elsewhere. The lignicolorous context

is rather fibrous and has a monomitic structure, com-

prising thick-walled, glassy generative hyphae, 5-8//.

diam., with clamp-connections at the septa. These

hyphae, which are non-dextrinoid, have a very broad

lumen. Toward the cuticle there are coloured con-

ducting elements with brown contents. There are

similar conducting elements in the tubes and some

of these appear to terminate in the hymenium as

gloeocystidia-like organs. The hyphae forming the

surface tomentum of the pileus are also brown. The

spores, 5, 0-5, 75x3, 75—4 , 2/x, vary in shape from

very broadly elliptic to ovate and have slightly

thickened hyaline walls, but some spores either

become brownish or become stained brownish. They

are non-dextrinoid.

If this account is compared with that of the STE

specimen it is difficult to believe that one is dealing

with two different species. I think it more probable

that the spores of the STE fruitbody are immature,

but obviously this is a point in need of clarification.

Further collecting will be necessary to resolve the

problem.

Trametes capensis Lloyd apud Doidge in Bothalia

5: 544 (1950), non rite publicatum.

226

LARGER BASIDIOMYCETES IN THE PRETORIA HERBARIUM

The description, without Latin diagnosis, is as

follows: “This fungus was named by Lloyd (l.c.) and

described in a letter dated 7 August 1920, but the

description was omitted from his published notes.

The following is Lloyd’s description: “Sessile, hard,

ligneous. Context and pores isabelline. Surface red-

dish-brown, minutely velutinate. Pores small, rigid,

irregular. Spores not found. This belongs in section

128 but is close to Trametes devexa (in 131). The yel-

lowish context and minutely velutinate surface alone

distinguish it. The spores have the same appearance.

Based on collection 8820, L. C. Turner.”

The specimen in PRE: “On Acacia horrida. Foun-

tains Valley, L. C. Turner, No. 8820” comprises

fruitbodies forming an imbricate series of woody

brackets, each with a somewhat uneven, pale tawny,

velutinate surface. The pores, 2-3 per mm, are rounded

to angular, with thick dissepiments and, over most of

the lower surface of the brackets, have become open

and plate-like or trametoid. The flesh, up to 6 mm

thick is tawny-brown, zoned, and there is a distinct

black, thin, horny cuticle. The context is trimitic,

comprising solid, glassy, slightly tinted, skeletal

hyphae, 5-6/x diam. ; solid binding hyphae, 2-3// diam.,

which are sometimes condensed into coralloid com-

plexes, but on occasion may be more elongated and

lax; generative hyphae were not seen and had pre-

sumably collapsed. Spores not seen.

This is typical Trametes meyenii (Klotzsch) Lloyd.

Trametes griseo-lilacina Van der Bvl in S. Afr.

J. Sci. 18: 283 (1922). (Fig. 8).

“Plants annual, sessile; pileus coriaceous to corky, applanate

imbricate, dimidiate to laterally extended, 3 cm to 10 cm by

2,5 cm to 4 cm by 4 cm to 0,8 cm; surface concentrically

sulcate, tomentose to glabrous, fasciculate setose, rugulose, grey

to greyish-fawn; margin acute, tomentose on upper surface;

context corky, firm, fibrous, lilac-mauve to mouse-coloured,

3 mm to 4 mm; tubes 0,5 mm to 2,5 mm long, lighter than

context; mouths unequal, irregular, round to elongated and

angular, 2 to 3 to the mm; edges thick, entire, concolorous;

lilac-mauve; spores hyaline, smooth, oblong, 3,7 // by 7,5//;

hyphae simple, 3,5// to 6//.

Distribution. — Single collection by Geo. Hobbs on railway

sleeper at Illovo River, Natal. (Type in Natal Herbarium, No.

921).

The context colour and colour of pore mouths is peculiar

and should aid in the identification of the fungus. The fungus

is evidently related to the Australian Trametes lilacino-gilva

(Berk.) Lloyd, but with rougher surface and irregular pore

mouths.”

The holotype in PRE: Jllovo River, Natal, leg.

G. Flobbs, No. 921 (PRE 13942), consists of sporo-

phores varying in size from 2-6 cm in length and

1, 5-4,0 cm from the point of attachment to the

margin. Some of the fruitbodies are imbricate, and

become laterally confluent. They have a conspicuously

radiately strigose surface consisting of adpressed but

raised strands of hyphae, although in some fructifica-

tions the surface of the basal portion has an almost

spiculose appearance (especially in young brackets).

The pores, mostly 3 per mm, are rather irregular with

somewhat thickened dissepiments and pinkish-grey-

brown in colour. The flesh, up to 5 mm thick, is

pinkish-fawn or pinkish-brown and has a felty texture.

Anatomically the context appears to be rudimentarily

trimitic and conforms closely to the structure described

for the specimen in STE (see Reid, 1973). Spores are

present, hyaline, elliptic, tapering to a snout, and

measuring 5,0-7,2x2,75-3,2//.

Trametes keetii Van der Byl in S. Afr. J. Sci.

18: 283 (1922). (Fig. 12).

“Plants sessile; pileus dimidiate, applanate, woody, firm and

rigid, slightly decurrent, 3 cm to 6 cm by 2 cm to 3,5 cm by

0,4 cm to 1 ,5 cm; surface grey to purplish black, finely tomen-

tose, smooth to scrupose; margin acute to rounded, grey;

context 0,2 cm to 1 cm, hard, rusty-brown: tubes firm, 0,2 mm

to 4 mm concolorous with context; mouths minute, subrotund,

6 to 9 to the mm; edges entire, thick, firm, concolorous; spores

not found, taken to be hyaline, setae present, slender, subulate,

21// to 28// long; hyphae 4// to 6// diam.

Distribution. — A single collection by J. D. Keet in Eastern

Cape Forest Conservancy on Rhus laevigata. (Type in Natal

Herbarium, No. 87).”.

The holotype in PRE: On Rhus laevigata. Eastern

Cape Forest Conservancy, Katberg Forest, 14 Aug.

1915, coll. J. D. Keet, No. 87 PRE (15636), consists

of sporophores up to 3,2 cm from point of attachment

to the margin, with a grey-black, blackish-brown or

purplish-black scrupose surface and a pinkish-fawn

marginal zone. In very young brackets the colour is

paler pinkish-beige. The pores, mostly 6 per mm and

scarcely visible to the naked eye, are very small,

rounded, with thick dissepiments, and either rusty or

rusty with a greyish bloom; the tubes are up to 4 mm

deep. The flesh is rather thick behind, reaching 1,2

cm, and is rusty-brown. Anatomically the context is

dimitic, comprising skeletal hyphae, 3, 5-4,0// wide,

with thick, brown walls although retaining a wide

lumen; and generative hyphae are paler, narrower,

branched, with thinner walls but lacking clamp-

connections at the septa. Setae 15-30/x long and 7-10//

wide, thick-walled, brown and subulate. Spores not

seen.

This is a thickish form of Phellinus gilvus (Schw.)

Pat.

Trametes ochrolignea Lloyd in Mycological Writ-

ings 5; Letter 63, p. 10 (1916). (Fig. 9).

“(Plant forms a hard flat, woody and unseparablemass, 2 mm

thick, on the wood of which it grows, the periphery of this mass

is continued into the pilei or the pilei develop separately from

it). Pibus woody, sessile, 4-6 inches in diameter, an inch thick

(pilei subcircular 6-10 cm diam., by 0,7 cm to 2,0 cm thick).

Surface uneven, not zoned, reddish-brown (Rood’s Brown,

Ridgway) minutely pubescent. (Surface uneven, tuberculate,

reddish brown, more yellow towards margin, tomentose, with

a few concentric furrows). Context hard, woody, yellow. (Con-

text 3 mm to 9 mm thick, corky to hard and woody, yellow,

shining). Pores round, or elongated, medium, somewhat ir-

regular. (Tubes about 1 mm long; mouths round to elongated,

irregular, 2 to 3 to the mm; edges thick or thin, somewhat

lacerate, yellow). Spores 3x6, cylindrical, hyaline, smooth.

(Spores hyaline, smooth, 3,7// by 7//; hyphae 3// to 7// diam.).

Trametes Zimmermanni , a species named in MSS at Berlin,

and Trametes ochro-flava of Brazil, which is really a Fomes,

are the only other species known to me with yellow context.

Although the hymenial configuration is so different, with same

texture, context color and spores, Daedalea Dregeana , Daedalea

Eatonii and Trametes ochrolignea are for me variations of the

same species. Zones of annual growth and indistinct pore

strata are seen in Trametes ochrolignea so that it is really a

Fomes.

( Distribution . — Found at Durban, Natal, on rotten log.

Distinguished from P. occidentalis by its habit of growth, more

tuberculate surface and context in older part and part above

substratum being extremely hard and woody. Cotype in Natal

Herbarium, No. 266).”

The isotype in PRE: Durban, Stella Bush, No. 226,

21 Feb. 1916. (PRE 11258), consists of imbricate

woody brackets, covered by an even, smooth, ochra-

ceous brown, felted tomentum, which is darker

brown behind. The brackets are about 5 cm long

and 3 cm from point of attachment to the margin.

The pores, which are irregular and angular to almost

sinuous behind, measure 2 per mm. The flesh, up to

1,2 cm at the base, is yellowish-brown to brown,

with a silky sheen where cut. Anatomically the con-

text is trimitic, comprising unbranched skeletal

hyphae, 4— 6(— 8)/x diam., with only slightly thickened,

pale brown walls and a very wide lumen (a very few

almost solid skeletals are also present); scanty, more

or less solid, hyaline, branched, binding hyphae,

2, 5-3,0 // diam., but not forming dense coralloid

complexes; and a few mostly collapsed, hyaline,

DEREK A. REID

227

generative hyphae, which appear to have very small

clamp-connections at the septa. Spores, often in

groups of 4, hyaline, elliptic, 6, 0-7, 2 X 2, 0-2, 2/*.

This is merely a thick form of Coriolopsis occident-

alis (Klotzsch) Murr. as suggested by Reid (1973).

Trametes subflava Lloyd in Mycological Writings

5; Letter 66, p. 8 (1917). (Fig. 10 a-b).

Examination of the isotype in PRE: On live Celtis

kraussiana, coll. J. D. Keet, in Eastern Cape Forest

Conservancy, No. 388 (PRE 31722), shows this to

accord in all details with the description which was

published (Reid, 1973) for the specimen in STE, and

which, despite the differing numbers, is probably

part of the same fruitbody. It certainly belongs in

the genus Oxyporus (Bourd. & Galz.) Donk to which

I referred it as O. subflavus (Lloyd) Reid.

Trametes tomentosa Van der Bvl in S. Afr. J.

Sci. 18: 285 (1922).

After examination of a specimen in STE (No. 707)

1 (Reid, 1973) concluded that this agreed with the

original diagnosis and was merely a large, thin,

flaccid form of Coriolopsis occidentalis (Klotzsch)

Murr. A study of the holotype in PRE: Durban,

No. 836 (PRE 14036), has confirmed that this species

is identical with C. occidentalis. The only unusual

feature of specimen 836 is that the surface tomentum

has become slightly more matted or felted and hence

smoother than that of material in pristine condition.

Trametes varians Van der Bvl in S. Afr. J. Sci.

18: 281-282 (1922). (Fig. 11).

“Plants effused-reflexed to largely resupinate; pileus corky

rigid, dimidiate, imbricate, at times conchate, 2 cm to 4 cm by

0,7 cm to 2 cm by 0,5 cm to I cm; surface azonate, minutely

pubescent, frequently somewhat tuberculate, at times rugulose

and scabrid, creamy white becoming ochraceous to somewhat

fuliginous; margin acute, cream coloured or fuliginous; context

1 mm to 2 mm thick, corky, creamy white; tubes 1 ,5 mm long,

discoloured within; mouths angular to elongated, irregular,

approximating 4 to the mm; edges entire, creamy white to

greyish and greenish drab; hyphae 7,2// diam.

Distribution. — Recorded from Eastern Cape Forest Con-

servancy. (Type in Natal Herbarium, No. 151).

Differs from Trametes glabrescens in habit, surface and pore-

characters.”

The holotype in PRE: Without field data on the

packet (but fide Van der Byl from the Eastern Cape

Forest Conservancy), det. Van der Byl and bearing

the number 151 (PRE 14027), consists of effuso-

reflexed, narrow brackets up to 2 cm from the point

of attachment to the margin of a grey-buff to ochra-

ceous-buff colour with a minutely roughened scrupose

surface, and an acute margin. These brackets are

rather thick behind, reaching 2 cm at the base. The

pores, 3-4 per mm, are angular with thin dissepiments

and the tubes up to 2 mm long. The flesh is con-

colorous and rudimentarily trimitic, comprising thick-

walled, glassy, skeletal hyphae, up to 6/* diam. and

also a few similar broad, thick-walled skeletal-type

hyphae, but with clamp-connections. The generative

hyphae are thin-walled, branched, with clamps at

the septa. There are no coralloid binding hyphae,

but there are long, narrow, glassy hyphae, 2,5/* diam.,

which are often kinked and bear abortive side-

branches at the angles. These hyphae wander through

the flesh and appear to be primitive binding hyphae.

Spores, 4,75-6,0x2,0-2,2/*, thin-walled, hyaline and

elliptic.

The above material seems to be conspecific with

the collection 527 in STE, but is apparently more

mature, which no doubt accounts for the very slightly

larger spores in the PRE fruitbodies. In my previous

paper (Reid, 1973) I suggested the STE material

represented young fructifications of a Coriolus. A

study of the more mature sporophores of the holotype

leads me to conclude that T. varians is a synonym of

Coriolus durbanensis (Van der Byl) Reid (see page

000).

Merulius gelatinosus Lloyd in Mycological Writ-

ings 7: 1 158, PI. 223, fig. 2293 (1922). (nec M. gelati-

nosus Petch, 1925). (Fig. 13).

“Resupinate, growing over leaves and debris with abundant

development of white subiculum. The hymenium is even, when

dry Dresden brown, and the quick change to bright russet when

moistened is remarkable. On soaking it swells with a gelatinous

appearance and the hymenium in folds turns dark, reminding

one of Exidia glandulosa. The basidia are hyaline but seated

in colored tissue. The abundant spores are elongated, 7x10,

deep colored, smooth. A section of soaked plant about 2-3 mm

thick shows 3 layers, the intermediate being white of compact

hyphae, the lower indefinite of coarse brown loosely woven

hyphae. The entire plant when soaked has a gelatinous ap-

pearance, not seen in others of this section. Its affinities other-

wise are very close to the notorious “dry rot” Merulius lacry-

mans.”

After a study of what is probably isotype material

in PRE: “On logs, debris, etc., Belvidere, Knysna,

C.P., coll. A. vr Duthie, July 1921 (PRE 31481),

there is little that can be added to the original diagno-

sis. As Lloyd stated, the hymenium of the dried

material is smooth, but on soaking in water it becomes

wrinkled in broad, obtuse undulations. The flesh is

up to 1 mm thick, and the white subiculum is formed

of hyaline, glassy hyphae with slightly thickened

walls and clamp-connections at the septa. Nearest

the substrate the hyphae have distinctly brown walls,

but are otherwise similar to those of the subiculum.

although less glassy. All the hyphae have a more or

less erect, entwined orientation. Below the hymenium

there is a dense subhymenial layer. The spores,

9,0-10,2x5,75-6,2/*, are elliptic, brown, with a

distinct wall.

Ginns (1969) noted that Talbot (1951) had con-

cluded that M. gelatinosus Lloyd differed from

Serpula himantioides (Fr.) Bond, ex Parmasto pri-

marily in having more frequent clamp-connections,

but Ginns argued that the specimen which Talbot

(and myself) studied, ie. Duthie 31481, may not

have been authentic, for he indicated that there was

no evidence from the specimen in the Lloyd herba-

rium to help establish its authenticity. (On the other

hand there seems to me absolutely no reason to

doubt it! There is an unusual feature shared by both

the type as figured by Lloyd and the PRE specimen,

in that both have numerous bird feathers in the sub-

strate). Ginns along with W. B. Cooke (1957) con-

sidered M. gelatinosus Lloyd to be a synonym of

Serpula lacrimans var himantioides (Fr.) W. B. Cooke.

However, many authorities would prefer to recognize

Serpula himantioides (Fr.) Karst, at specific level.

Despite the conclusions oc Ginns and W. B. Cooke

I am not entirely convinced that M. gelatinosus Lloyd

should be reduced to synonymy under S. himantioides.

The latter normally retains a distinctly reticulate-

poroid hymenial configuration in the dried state; it

does not swell conspicuously on soaking in water;

and 1 have never seen a specimen 1 mm in thickness

exclusive of pore tissue. I am uncertain whether this

species should be assigned to Serpula Pers. ex S. F.

Gray or to Coniophora DC. ex Merat. Further col-

lecting is needed to resolve this problem.

228

LARGER BASIDIOMYCETES IN THE PRETORIA HERBARIUM

CLAVARIACEAE

Clavaria cladoniae Kalchbr. in Grevillea 10: 105

(1882). (Fig. 14 a-d).

“Terrigena, simplex, ochraceo-ferruginea clavulis vix pol-

licem altis, sursum dilatatis et in ramulos breves, compressos,

subfoliaceos divisis.

Somerset East, I. McOw., No. 1431.

Basi mycelio radiculoso albo, supra folia serpente aucta.

Clavulae nonnunquam scyphum Cladoniae pyxidatae referunt.

Quam, Clavaria pyxidata, Pers. multo tenerior.”

"Growing on the ground, simple, ochraceous-ferrugineous,

clubs scarcely an inch high, dilated above and divided into

short, compressed, subfoliaceous branches.

Somerset East, coll. MacOwan, No. 1431.

With basal mycelium forming white root-like strands,

spreading and creeping over the leaves. Clubs sometimes

resembling the scyphus of Cladonia pyxidata ; much thinner

than Clavaria pyxidata. Pers."

The isotype in PRE comprises fruitbodies up to

I cm high, with a main trunk which is more or less

flattened and dilated above where there is the irregular

production of short prongs. The fructifications

Figs. 1-12. — Fig. 1, Polyporus areno-

sobasus, spores from fragments

in K of Van der Byl 710. Fig. 2,

Polyporus conchatus, spores from

PRE 385. Fig. 3, Polyporus molli-

carnosus, spores from presumed

isotype in PR.E. Figs. 4 a-b, Poly-

porus nigrolucidus, a, cuticular

elements; b, spores; from PRE

13941. Figs. 5 a-b, Polyporus

rusticus, a, basidia; b, spores;

from isotype in PRE. Fig. 6, Tra-

metes albotexia, spores from iso-

type in PRE. Figs. 7 a-b, Poly-

porus trichiiiae, a, cystidioles; b,

spores; from holotype in PRE.

Fig. 8, Trametes griseo-lilacina,

spores from holotype in PRE.

Fig. 9, Trametes ochrolignea,

spores from isotype in PRE. Figs.

10 a-b, Trametes subflava, spores,

from isotype in PRE. Fig. 11,

Trametes varians, spores, from

holotype in PRE. Fig. 12, Trametes

keetii, setae, from holotype in

PRE. (All x 866).

resemble miniature antlers, and are rusty-ochraceous

with a greyish tinge, but are attached to vegetable

debris by the copious development of cream, cottony

mycelial strands. The structure of these strands is

monomitic, comprising narrow, hyaline hyphae,

1,7-2,0/x wide, with slightly thickened, distinct walls

but retaining a wide lumen. These hyphae, which

have clamp-connections at the septa, are sometimes

encrusted with large amorphous granules. Anatom-

ically the flesh of the sporophore is formed of

similar clamp-bearing hyphae, 2-3m wide, although

some appear to be inflated up to 5/u. Basidia elongate-

clavate, with a basal clamp, and up to 4Ctyi in length.

Spores brown, echinulate, elliptic, 6,5-8,75x3,0-

3,75/t.

This fungus belongs in the genus Ramaria Flolmsk.

ex S. F. Gray and is accordingly so transferred as

Ramaria cladoniae ( Kalchbr .) Reid, comb. nov. It

should be noted that Corner (1950) suggested that

this species belonged in the genus Lachnocladium

Lev.

DEREK A. REID

229

TREMELLALES

Tremella hemifoliacea Llovd in Mycological

Writings 7: 1338, PI. 324, Figs'. 3104, 3105 (1925)-

(Fig. 1 5 a-b).

“When we first saw this it was something out of the usual,

for we never saw a Tremella that dried down to a thin brown

layer (Fig. 3104). When soaked it developed in perfect cere-

briform fruitbodies over portions of the surface but not all

(Fig. 3105). As to color, consistency and ‘structure’ it is the

same as the European species Tremella foliacea , but a com-

parison with our Fig. 1195, page 793, well demonstrates that

it cannot be the European plant. We will not further describe

it, for, excepting its different method of development, it is

same as Tremella foliacea. The brown color is exactly same, but

collector’s notes were a mass of dirty-white jelly.”

The isotype in PRE: On Gymnosporia peduncularis ,

The Glebe, Knysna, C.P., coll. E. M. Doidge, 13

May 1923 (PRE 17798), consists of sporop'nores which,

when dry, form a distinct greasy-looking (not varnish-

like) brown film with a metallic sheen, for the most

part smooth, but here and there with an occasional

wrinkle or fold. When soaked the fungus forms a

dark brown, firmly gelatinous, encrusting layer, rather

like a Sebacina, with a very closely-tuberculate or

wrinkled surface, which does not become gyrose,

foliose or cerebriform. The context is formed of very

narrow, gelatinized hyphae, difficult to distinguish

individually, but with brownish, granular contents

and minute clamp-connections at the septa. The

basidia are brown, more or less globose and 4-spored.

Spores, 6,2-8,2x4,75-6,0^, vary from ovate or

subglobose to elliptic, and germinate by repetition.

It should be noted that when Bandoni (1958)

examined the holotype in the Lloyd Herbarium, he

found it to be in poor condition and, beyond being

able to confirm that it was a tremellaceous fungus,

he could not even be sure of its generic disposition.

It seems best left in Tremella Dill, ex Fr.

Figs. 13-16. — Fig. 13, Merulius tre-

niellosus, spores, from probable

isotype in PRE. Figs. 14 a-d, Cla-

varia cladoniae, a, basidia; b,

spores; c, spore 1450: d, habit

sketch x T from holotype in

PRE. Figs. 15 a-b, Tremella hemi-

foliacea, a, basidium: b, spores;

from isotype in PRE. Figs. 16

a-c, Polyporus durbanensis, habit

sketches of sporophores, from col-

lection 26845 in PRE, ■ J ; c, habit

sketch, from holotype in PRE, xf

(All x 866 except where otherwise

indicated).

230

LARGER BASIDIOMYCETES IN THE PRETORIA HERBARIUM

REFERENCES

Bandoni, R. J., 1958. Some Tremellaceous Fungi in the C. G.

Lloyd Collection. Lloydia 21 : 145.

Cooke, W. B., 1957. The Genera Serpula and Meruliporia.

My colog ia 49: 208.

Corner, E. J. H., 1950. A Monograph of Clavaria and Allied

Genera. Oxford, p. 262.

Doidge, E. M., 1950. The South African Fungi and Lichens to

the end of 1945. Bothalia 5: pp. 1094.

Donk, M. A., 1966. Notes on European Polypores I. Persoonia

4: 339.

Donk, M. A., 1971. Notes on European Polypores VII. Proc.

K. ned. Akad. Wet. Section C 74: 26.

Ginns, J. H., 1969. The Genus Merulius II. Species of Merulius

and Pltlehia proposed by Lloyed. Mycologia 61: 368-369.

Lloyd, C. G., 1920. Mycological Notes 62. Mycological

Writings 6: 919.

Reid, D. A., 1973. A reappraisal of type and authentic speci-

mens of Basidiomvcetes in the Van der Byl Herbarium,

Stellenbosch. J. S. Afr. Bot. 39: 141-178.

Stevenson, J. A. & Cash, E. K., 1936. The new Fungus Names

proposed by C. G. Lloyd. Bull. Lloyd Libr. 35: 209.

Talbot, P. H. B., 1951. Studies of some South African Resu-

pinate Hymenomycetes. Bothalia 6: 32-33.

Van der Byl, P. A., 1928. Aantekeninge oor enige Suid-

Afrikaanse Swamme. S. Afr. J. Sci. 25: 183.

Bothalia 11, 3: 231-234 (1974)

Notes on Acacia species in Southern Africa: IV

J. H. ROSS*

ABSTRACT

Information concerning miscellaneous Acacia species is presented, and a new species, A. redacta

J. H. Ross, from the northern Cape Province is described.

ACACIA CALLICOMA MEISN.

Meisner, in Hook., Lond. J. Bot. 2: 104 (1843),

based his description of A. ca/licoma on a specimen

without fruits, and doubtfully from Natal, seen in

Krauss’ herbarium. No type specimen was cited and

1 have not succeeded in tracing a specimen bearing

this name or any further reference to the species in

literature. Neither Harvey in FI. Cap., vol. 2 (1862),

nor Bentham in his revision of Mimoseae in Trans.

Linn. Soc. Lond. vol. 30 (1875), mention A. callicoma.

The description suggests that the specimen Meisner

saw was not an Acacia but, in the absence of a

specimen, the precise identity of A. callicoma remains

unresolved.

ACACIA REDACTA /. H. ROSS

Acacia redacta J. H. Ross, sp. nov., ab omnibus

speciebus africanis distincta, affinitate incerta.

Frutex ramosissimus, 0,3-0, 6 m altus; ramuli

fusco-griseo-brunnei. Ramuli juveniles rubrobrunnei,

dense et persistenter appresse pubescentes, glandulis

numerosi minimi atropurpureis conspicuis inter pilos

immixtis. Spinae stipulares geminae, plus minusve

elongatae, 0,8-1 ,4 cm longae, rubrobrunneae, rectae

sed saepe deflexae. Folia: petiolus brevis, plerumque

2-6 mm longus, griseo-puberulus, apice eglanduiosus;

pinnae 1-jugatae; rhachillae 0,4-1, 8 cm longae,

subglabrae vel puberulae; foliola 2-4-juga, 2-5,5 mm

longa, 1,2-3, 5 mm lata, oblique oblonga vel elliptica

vel rb subrotundata, apice rotundata vel obtusa,

pagina appresse pubescenti, costa et nervis lateralibus

subtus haud perspicuis vel obscuris. Inflorescentiae ut

videtur capitatae, redactae; flores 2 -4 pro inflorescen-

tiam vel solitarii; pedunculi axillares, 2-6 mm longi, ut

ramuli conspicue glandulosi, dense appresse pubes-

centes. Flores ut videtur flavidi. Calyx cupularis,

0,8-1, 2 mm longus, quinquedenticulatus, dense pu-

bescens. Corolla 4-6 mm longa, praesertim in lobis

dense appresse pubescens vel tomentella. Stamina

numerosa, filamentis 14-17 mm longis, basi breviter

connatis et breviter tubulosis. Ovarium breviter stipi-

tatum, ±1 ,5 mm longum. Stylus ± 17-19 mm longus.

Legumina 2, 6-3, 2 cm longa, 0,9-1 , 1 cm lata, lineari-

oblonga, recta, 1-2-sperma, subroseo-brunnea, dense

appresse griseo-puberula, glandulis numerosis minimi

atropurpureis sessilibus conspicuis inter pilos immix-

tis, apice acuta vel acuminata, dehiscentia. Semina

non visa.

Type: Cape, 2817 (Vioolsdrif): 22,4 km north of

Stinkfontein on way to Jenkinskop ( — CB), on steep

west-facing ridge of schistoid granite, on border of

Namaqualand Broken Veld and Western Mountain

Karoo, M. J. A. Werger 1518 (PRE, holo.; K, iso.).

Much-branched shrub 0,3-0, 6 m high; branches

dark grey-brown, flaking minutely. Young branchlets

reddish-brown, densely and persistently appressed

pubescent, with numerous small, conspicuous, dark

purplish glands among the hairs. Stipules spinescent,

in pairs, 0,8-1, 4 cm long, reddish-brown, straight

or often deflexed, slender. Leaves: petiole short.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

mostly 2-6 mm long, grey-puberulous, eglandular

apically; pinnae 1 pair; rhachillae 0,4-1, 8 cm long,

subglabrous or puberulous; leaflets 2-4 pairs, 2-5,5

mm long, 1,2-3, 5 mm wide, oblique, oblong or

elliptic or ± subrotund, apex rounded or obtuse,

surfaces appressed pubescent, midrib and lateral

nerves not visible or inconspicuous beneath, with

minute reddish glands at the point of attachment of

the leaflets. Inflorescences apparently capitate, re-

duced; flowers 2-4 per inflorescence or sometimes

solitary; peduncles axillary, 2-6 mm long, conspicu-

ously glandular like the branchlets, densely appressed

pubescent. Flowers apparently pinkish. Calyx cupu-

lar, 0,8-1, 2 mm long, 5-toothed, densely pubescent.

Corolla 4-6 mm long, lobes densely appressed

pubescent or tomentellous. Stamens numerous, fila-

ments 14-17 mm long, shortly connate basally and

shortly tubular. Ovary shortly stipitate, ± 1,5 mm

long. Style d= 17-19 mm long. Pods 2, 6-3, 2 cm long,

0, 9-1,1 cm wide, linear-oblong, straight, 1-2 seeded,

pinkish-brown, densely appressed grey-puberulous,

numerous small, conspicuous, dark purplish sessile

glands among the hairs, apex acute or acuminate,

dehiscent. Seeds not seen.

The first specimens of this plant were collected by

Dr O. A. Leistner and Dr M. J. A. Werger in the

north-western Cape Province in 1969. In response to

a request for more material, the latter, who is under-

taking a survey of the Orange River, collected a

further set of specimens in December 1971.

While the capitate heads or elongate spikes in all

other African species of Acacia contain numerous

flowers, in this plant the inflorescences are greatly

reduced and the “heads” contain only 2-4 flowers or,

rarely, the flowers appear to be solitary. This reduction

in the number of flowers per “head” is coupled with

an increase in the size of the individual flowers, the

flowers being larger than in most other species of

African Acacia. The stamen-filaments are united

basally for ± 2 mm into a tube, which is a character

shared by only a few other species, for example,

A. albida Del. and A. eriocarpa Brenan. Unfortu-

nately no young anthers are available for examination

so it has not been possible to establish whether or

not the anthers have an apical caducous gland.

More flowering material is required and flower colour

needs confirmation. The pods, which are very small,

dehisce longitudinally from the apex downwards

and the two valves diverge. Unfortunately no seeds

are available to establish whether or not they have the

characteristic areole. However, even if absent, they

would not be without parallel for the seeds of A.

leucospira Brenan lack an areole.

Although it differs from all the other African species

of Acacia in lacking many-flowered capitula or

spikes, it would seem unwise to separate this plant

generically from Acacia. The presence of straight,

paired, stipular spines, small bipinnate leaves, flowers

with numerous stamens, ± straight, dehiscent pods,

and the overall facies, suggest that the plant is an

Acacia. The plants are of very small stature but this

232

NOTES ON ACACIA SPECIES IN SOUTHERN AFRICA

Fig. 1. — Specimen of Acacia redacta

growing on a steep, west-facing

ridge of schistoid granite, 22,4 km

north of Stinkfontein on the way

to Jenkinskop, northern Cape

Province. A dehisced pod with

diverging valves is visible towards

the right-hand side of the plant.

Fig. 2. — The habitat occupied by

Acacia redacta on the border of

Namaqualand Broken Veld. Many

Aizoaceous genera are present and

species of Crassula, Cotyledon and

Euphorbia are common.

J. H. ROSS

233

is thought to be the result of the extreme and inhospi-

table environment they occupy, an environment

which is conducive to shrubbiness and in which the

arborescent growth form is not well represented.

(See Figs. 1 & 2). The reduction in the number of

flowers per inflorescence is also a possible result of

the adverse environment.

A. redacta appears to have a very restricted distri-

bution in the north-western Cape Province. Appar-

ently it favours rocky ridges.

Cape. — 2817 (Vioolsdrif): 20,8 km north of Stink-

fontein (-CB), Leistner 3401 ; 22,4 km north of

Stinkfontein on way to Rosyntjieberg (-CB), (Verger

428; 22,4 km north of Stinkfontein on way to

Jenicinskop (-CB), (Verger 1518.

I am most grateful to Mr J. P. M. Brenan, Keeper

of the Herbarium, Royal Botanic Gardens, Kew, for

commenting on the above specimens, to Mr H. K.

Airy Shaw for checking the Latin description, and

to Dr Werger for supplying the photographs which

accompany this note and for the ecological infor-

mation.

ACACIA REFICIENS WA WRA — ACACIA LUEDERTIZII

ENGL.

The A. reficiens Wawra — A. luderitzii Engl, com-

plex was dealt with some years ago (Ross & Brenan

in Kew Bull. 21: 71, 1967). Although A. reficiens

and A. luederitzii are unquestionably closely related,

the two were maintained as distinct species. However,

Schreiber in FI. S.W. Afr. 58: 11 (1967) regarded

A. reficiens and A. luederitzii as conspecific, placing

the latter species as a synonym of A. reficiens.

Because of these conflicting views, the complex has

been re-investigated. The conclusions reached are

expressed below.

A. reficiens , which was based on Wawra 248 from

between Benguela and Catumbela in Angola, differs

from A. luederitzii mainly in its indumentum being

puberulous or pulverulent without spreading hairs;

by the consistently few pairs of leaflets, usually up

to 11, occasionally 13, pairs per pinna with their

margins glabrous or almost so, lacking the spreading

cilia; fewer pinnae pairs, 1-4; and by the tendency

for narrower pods. A. reficiens appears to be a fairly

uniform species and the two subspecies within it,

although separated by a large geographical discon-

tinuity (see Fig. 3), differ very slightly from each

other. A. luederitzii appears to be far more variable.

When working on this complex previously, we had

on loan the material from the National Herbarium,

Pretoria. It was found that the PRE specimen of

Marloth 1270, one of the syntypes of A. luederitzii,

was in fact a specimen of A. reficiens subsp. reficiens,

while the specimen of Marloth 1328, the other syn-

type, matched the illustration and the description

of A. luederitzii in Bot. Jahrb. 10: 23, t. 3 B (1888).

Consequently, the PRE specimen of Marloth 1328

was selected as the lectotype of A. luederitzii. Since

then specimens of Marloth 1328 from the Albany

Museum in Grahamstown, the Staatssammlung in

Munchen, and the Fielding Herbarium in Oxford

have been examined, and all three agree with the

PRE specimen. However, the specimen of Marloth

1328 from the Bolus Herbarium, University of Cape

Town, is unquestionably a specimen of A. reficiens

Fig. 3. — The known distributions of

Acacia reficiens and A. luederitzii.

234

NOTES ON ACACIA SPECIES IN SOUTHERN AFRICA

subsp. reficiens and not A. luederitzii. There is no

doubt about this. As the PRE sheet of Marloth 1270

is A. reficiens and the BOL sheet of Marloth 1328 is

also A. reficiens , it seems that some of Marloth's

specimens must have been muddled at some stage.

It is, of course, quite probable that the specimens

of Marloth 1270 and 1328 which Engler had before

him in Berlin when he described A. luederitzii, were

both A. luederitzii. However, in the absence of the

Berlin specimens, it will now unfortunately never

be known whether or not both specimens were

A. luederitzii.

A. retinens Sim, from Mozambique, the Transvaal,

Swaziland and Natal, is certainly not conspecific

with A. reficiens. It differs from A. reficiens mainly

in that some of the spines are greatly enlarged and

inflated into characteristic “ant-galls”, in having up

to 13 pinnae pairs, a spreading indumentum, and

leaflets with spreading marginal cilia. However,

A. retinens (A. luederitzii var. retinens) is not speci-

cally distinct from typical A. luederitzii. This is the

real problem within this complex: the fact that

A. reficiens and A. retinens , which are themselves

distinct, are almost linked through another species,

namely, A. luederitzii.

Although A. reficiens and typical A. luederitzii

are undoubtedly very closely related, and although

there are some problematical specimens in South

West Africa, it is still considered that the two species

should be maintained, particularly when material

throughout their ranges (and that of some of their

allies) is examined. If A. luederitzii were reduced,

for example, to subspecific or varietal rank within

A. reficiens, then this would necessitate according

A. retinens similar rank within A. reficiens. It is felt

that the inclusion of A. retinens under A. reficiens

would alter the concept of the latter species entirely.

As the taxonomy and diagnostic characters of A.

reficiens and A. luederitzii were enumerated in Kew

Bull, l.c., it is not necessary to repeat them here.

A. reficiens subsp. reficiens and A. luederitzii var.

luederitzii tend to occupy somewhat different geo-

graphical ranges in South West Africa (see Fig. 3);

the former occurring in the western areas of the terri-

tory and the latter in the eastern, although the ranges

of the two do show some overlap.

ACACIA SCHLECHTERl HARMS

Harms, in Bot. Jahrb. 51: 367 (1914), based his

description of A. schlechteri on the specimen Schlechter

1 1901 from Ressano Garcia in Mozambique. As it

had long been assumed that no type specimen sur-

vived, attempts had been made to identify the species

from Harms’ description. A. schlechteri was pro-

visionally referred to synonymy under A. welwitschii

Oiiv. subsp. delagoensis (Harms) Ross & Brenan

(Ross in J.S. Afr. Bot. 39: 265, 1968), while Brenan

in FI. Zamb. 3, 1: 76 (1970) suggested that it may

represent a plant similar to A. burkei Benth. It was

therefore of great interest to find an isotype of A.

schlechteri in the University of Zurich Herbarium.

The isotype consists of a single flowering twig.

The paired prickles are slightly recurved, the bark

is dark grey, and the young extremities are very

sparingly pubescent. The petioles, rhachides and

rhachillae are also sparingly pubescent. The leaves

have 2-4 pinnae pairs and the pinnae 3-4 leaflet

pairs. The leaflets, which vary considerably in size

and in shape, are up to 14x8 mm, discolorous,

asymmetric basally, glabrous or with few marginal

cilia, while some have a small tuft of hairs to one

side of the midrib on the lower surface. Some leaflets

are broadest approximately at the middle while

others are broader above the middle and are obovate

or obovate-oblong. The inflorescence axes are up to

8 cm long (excluding the peduncle) and sparingly

pubescent. Some of the calyces are glabrous, some

have occasional scattered hairs, while some are

sparingly pubescent.

A. schlechteri forms part of a taxonomically difficult

complex of species which includes A. burkei, A.

rovumae Oliv., A. goetzei Harms, A. nigrescens Oliv.

and A. welwitschii. Within this complex the degree

of pubescence of the calyx is the character of prime

importance in distinguishing between two groups:

A. burkei and A. rovumae alone have densely

pubescent calyces. Although some of the calyces

in the isotype of A. schlechteri are sparingly pubescent,

the specimen has a somewhat different facies to the

material of A. burkei and, in particular, the young

inflorescences lack the long, exserted bracts which

are present in A. burkei. Consequently, it is felt that

its affinities are not with A. burkei but with A. goetzei

subsp. goetzei or with A. welwitschii subsp. dela-

goensis, two very closely related and variable taxa.

Neither A. goetzei nor A. welwitschii typically has

sparingly pubescent calyces, but the sparingly pubes-

cent calyces exhibited by the type of A. schlechteri

could well represent a local variant of either species.

Sparingly pubescent inflorescence axes similar to

those of A. schlechteri do occur in several specimens

of A. goetzei in Mozambique, Tanzania and in Kenya,

although sparingly pubescent calyces have not been

definitely recorded in A. goetzei.

It is very unlikely that A. schlechteri represents a

“good” taxon, but from the available material it is

difficult to establish its precise affinities with certainty.

It is unfortunate that the type of A. schlechteri is

without pods because pods would enable the species

to be placed with a fair degree of certainty.

The asymmetric leaflet bases, leaflet shape, the

presence of a small basal tuft of hairs to one side

of the midrib on the lower surface of some leaflets

(a character not yet observed in A. welwitschii subsp.

delagoensis), the sparingly pubescent leaf-rhachides

and inflorescence axes, and the long inflorescence

axes, suggest that A. schlechteri is more closely

related to A. goetzei subsp. goetzei than it is to

A. welwitschii subsp. delagoensis. A. schlechteri is

most probably a local variant of A. goetzei subsp.

goetzei and not worthy of formal taxonomic recog-

nition. Clearly, more material from the type locality

of A. schlechteri is required in an attempt to evaluate

the incidence and significance of sparingly pubescent

calyces.

The difficulty encountered in attempting to place

this specimen of A. schlechteri emphasizes the need

for more field studies within this complex of species

in southern Mozambique. In southern Mozambique

A. burkei, A. goetzei, A. nigrescens and A. wel-

witschii subsp. delagoensis meet and the range of

variation within each species in this area requires

investigation. It is possible that some of the species

hybridize but, if they do, proof is required. Infor-

mation relating to the nature of any hybrids would

be extremely useful.

Bothalia 11, 3: 235-241 (1974)

Comparative Leaf Anatomy of Paspalum paspalodes and

P. vaginatum

R. P. ELLIS*

ABSTRACT

The leaf blade and epidermal anatomy of Paspalum paspalodes ( Michx.) Scribn. (syn. P. distichum

L.) and P. vaginatum Swartz is compared and discussed. Numerous anatomical differences are evident

and the species can be separated on the basis of the distribution of sclerenchyma and the shape and

thickness of the leaf margin. Another distinct diagnostic difference is the shape of the adaxial ribs and

furrows and their associated papillae. P. paspalodes has two distinct types of abaxial epidermis:

cuticular and papillate.

INTRODUCTION

The need for a detailed investigation of the relation-

ship between Paspalum paspalodes (Michx.) Scribn.

(syn. P. distichum L.) and P. vaginatum Swartz arose

because of the conflicting opinions in the literature

as to the nomenclature and status of these two related

species. These historical, morphological and taxono-

mic aspects are discussed by Loxton (1974). The

present anatomical investigation was conducted in

conjunction with the above taxonomic study and here

the leaf and epidermal anatomy are compared and

differences and similarities discussed.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

TABLE 1. — The differences between Paspalum paspalodes and P.

Leaf material was collected and fixed in FAA in the

field from a number of localities in South Africa.

Transverse sections of the leaf were prepared by use

of wax embedding techniques, and double-stained

in safranin and fast green. The method of Metcalfe

(1960) was used for preparing epidermal material for

microscopic study. To minimize the possibility of

intraspecific variation and to facilitate and standardize

comparison, tissue from a point midway between the

apex and the ligule of the leaf blade was examined.

vaginatum as seen in transverse sections of the leaf blade.

Character

Paspalum paspalodes

Paspalum vaginatum

1. Width of blade at a point halfway

between apex and ligule

2. Adaxial furrows

3. Second-order vascular bundles

4. First-order vascular bundle

5. Parenchyma sheath of third-order

vascular bundles

6. Parenchyma sheath of first-order

vascular bundles

7. Sclerenchyma of leaf margin in trans-

verse section

8. Colourless cells

9. Adaxial epidermal cells

10. Abaxial epidermal cells

1. More than 3,00 mm wide; usually wider

than 7,20 mm

2a. Shallow furrows only; usually with

distinct furrows less than i of the leaf

thickness deep, or leaf may be undulating

with no regular furrows associated with

the vascular bundles

2b. Furrows always wide and open; obtuse

angle formed by the furrow sides at the

base

3. Tali and narrow with vertical, parallel

sides; rarely some bundles may be ellip-

tical in shape

4a. Lysigenous cavity and protoxylem

vessel always present (Fig. 15, 16)

4b. Vessels wider than the cells ot the

parenchyma sheath and always angular

in section

5. Sheath rounded or circular in shape

6a. Commonly round or circular in shape;

some bundles may have elliptical, verti-

cally-elongated sheaths

6b. Parenchyma cells of sheath often

rounded in shape; elongated, elliptical

cells or cells with straight radial and in-

flated tangential walls uncommon

7. Relatively small to well-developed

sclerenchyma caps present; rounded in

shape, rarely narrow and pointed (Fig.

22, 23, 24)

8 No colourless cells present.

10 Two distinct types of epidermis occur:

(i) Epidermis may be composed of small,

inflated, bulliform cells with a very

thin cuticle and papillae

(ii) Epidermis may be composed of rela-

tively small cells with no papillae and

the outer tangential wall with a distinct,

continuous cuticle

1. Less than 3,00 mm wide

2a. Furrows of medium depth; depth

between \ and J) of the leaf thickness

(Fig. 10)

2b. Furrows in the form of a narrow cleft

3. Elliptical, vertically elongated in shape

4a. Lysigenous cavity and protoxylem

rarely present together; usually only a

protoxylem vessel is present but both may

be absent (Fig. 10)

4b. Vessels not as wide as the cells of the

parenchyma sheath and rounded or cur-

cular in section

5. Sheath usually elliptical, vertically elon-

gated; some bundles may possess round

sheaths

6a. Sheaths elliptical or egg-shaped with

broadest side adaxial

6b. Cells of sheath usually with straight

radial and inflated tangential walls;

rounded cells do occur infrequently

7. Very small sclerenchyma cap consisting

of a few fibres; pointed in shape (Fig.

20, 21)

8. Irregular abaxial groups of colourless

cells associated with the sclerenchyma

strands present (Fig. 10)

9. Sunken constricted bases of short, thick

macro-hairs relatively common

10 Epidermis composed of relatively large

inflated bulliform cells with a distinct

thick cuticle; no papillae

9 No macro-hair bases found

17749-3

236

LEAF ANATOMY OF PASPALUM PASPALODES AND P. VAGINATUM

ANATOMY OF THE LEAF

Leaf Blade

Sections of the leaf blade of P. paspalodes and

P. vaginatum, taken at a point midway between the

leaf apex and the ligule, exhibit the following common

anatomical features.

1. Adaxial ribs, wherever present, are rounded and

present only over the first- and second-order vascular

bundles. Adaxial furrows are situated over the third-

order vascular bundles.

2. Abaxial ribs are absent in both species. Specimens

of P. paspalodes with inflated, abaxial epidermal cells

may possess an irregular, undulating, abaxial surface,

but the undulations are not regularly associated with

any order of vascular bundle to form ribs (Fig. 1).

3. One, rarely two, third-order vascular bundle is

located between consecutive first- and second-order

vascular bundles.

4. The first- and second-order vascular bundles are

centrally positioned in the leaf section, whereas the

third-order bundles are abaxially situated.

5. The bundle sheath of all vascular bundles is

single, entire and without parenchymatous projec-

tions. The sheath cells contain specialized chloro-

plasts, which are larger than those of the chlorenchyma

and are often grouped near the outer tangential walls

of the sheath cells.

6. No adaxial or abaxial sclerenchyma strands or

girders are associated with the third-order vascular

bundles.

7. No sclerenchyma is developed between the

bundles.

8. Adaxial papillae are developed on all, or most,

epidermal cells in both species. They are always wider

than half the width of the epidermal cells and are

either thin- walled or have the distal wall thickened.

Notwithstanding the above structural similarities

between the two species, they can, nevertheless, be

readily separated and distinguished by the diagnostic

characters of the leaf blade set out in Table 1.

Fig. 1-9. — Leaf blade outline in transverse section. All x6,25 bright field, bigs. 1-6, Paspalum

paspalodes. Fig. I, Ellis 1 106. Fig. 2, Ellis 254. Fig. 3, Ellis 146. Fig. 4, Ellis 432. Fig. 5, Ellis 767.

Fig. 6, Ellis 726. Figs. 7-9, Paspalum vaginatum. Fig. 7, Ellis 1114. Fig. 8, Ellis 276. Fig. 9,

Werger 1622.

R. P. ELLIS

237

Although these diagnostic characters exhibit varia-

tion, there are distinct discontinuities between the

two species. Numerous other differences occur, but

the variation shown by these characters is relatively

great and may overlap with the extreme of the same

character in the other species. These cannot be used

alone as diagnostic characters but, when used in

combination with other differences, are nevertheless

important.

Fig. 10—19. — Transverse sections of leaf blade. All x40. Figs. 10-12, Paspalum vaginaium. Fig. 10,

Ellis 276 bright field. Fig. 11, Werger 1622 bright field. Fig. 12, Ellis 276 interference contrast.

Figs. 13-19, Paspalum paspalodes. Fig. 13, Ellis 1106 interference contrast. Fig. 14, Ellis 146

lateral region of leaf, bright field. Fig. 15, Ellis 146 central region of leaf, interference contrast.

Fig. 16, Ellis 254 bright field. Fig. 17, Ellis 432 interference contrast. Fig. 18, Ellis Ibl inter-

ference contrast. Fig. 19, Ellis 767 interference contrast.

238

LEAF ANATOMY OF PASPALUM PASPALODES AND P. VAGINATUM

The outline of the blade of P. paspalodes, as seen

in transverse section, is expanded and gently undu-

lating (Fig. 2) or broadly V-shaped with the arms of the

“V” outwardly bowed (Fig. 5, 6) or irregularly wavy

or bent (Fig. 2). Many leaves exhibit various degrees

of inrolling from both margins (Fig. 3, 4, 6) and this

loosely involute condition appears to be basic. Under

adverse moisture conditions probably all leaves

assume the involute shape of Fig. 4. Leaves of P.

vaginatum are either V- (Fig. 7) or U-shaped (Fig.

8, 9), with the lamina arms either straight or concave,

respectively. The angle formed by the arms of the

lamina at the median vascular bundle is usually small,

but may be 90° in certain cases.

In P. vaginatum the median vascular bundle is

distinguishable only by the central position in the

leaf section and is structurally identical to the other

first-order bundles of the leaf. Some specimens of

P. paspalodes possess only a median vascular bundle

(Fig. 2, 3, 4, 6), but in others a keel comprising 7-9

bundles is developed (Fig. 1, 5). Only a single first-

order bundle is present in the keel. Keel shape is

often not distinct from the leaf outline, the keel

merely being thicker than the rest of the lamina.

Well-developed keels may be semicircular or rounded

with a raised, flattened adaxial side.

The arrangement of the different orders of vascular

bundle along the leaf differs in the two species. In

P. paspalodes there are 3 or 4 second-order vascular

bundles between successive first-order bundles. There

is no regular arrangement in P. vaginatum and the

number of second-order bundles between successive

first-order bundles decreases rapidly from centre

to margin. There is only one second-order bundle

between the lateral pair of first-order bundles.

The bundle sheath of the first-order bundles of

both species is composed of smaller parenchyma

sheath cells than second- and third-order bundles.

These sheath cells are normally thin-walled with

large chloroplasts situated near the outer tangential

wall. In a few specimens of P. paspalodes, however,

the sheath cells are distinctly thickened on their

inner tangential and radial walls thus resembling a

mestome sheath. In these instances the chloroplasts

are concentrated near this thicker, inner tangential

wall (Fig. 14).

Sclerenchyma development associated with the

first- and second-order vascular bundles in P. vagina-

tum is scant and in the form of a shallow, sub-epider-

mal strip adaxially and a small strand abaxially

(Fig. 10, 11, 12). Girders were never found. In P.

paspalodes both the adaxial and abaxial sclerenchyma

is in the form of small strands (Fig. 18), but girders

are sometimes developed when these strands are in

contact with the bundle sheath (Fig. 15, 19).

The chlorenchyma of both species is irregular or

sometimes indistinctly radiate with the adaxial cells

tabular. It is continuous between the vascular bundles.

The indistinctly radiate condition is more commonly

found in P. vaginatum.

In both species the adaxial bulliform cells are

either extensive groups present throughout the epi-

dermis and only reduced opposite the first- and second-

order vascular bundles, or they are in fan-shaped

groups at the base of furrows.

The adaxial papillae of P. vaginatum are inflated,

thin-walled and as wide as the epidermal cell on which

the papillus is located (Fig. 12). They are longer than

those of P. paspalodes and almost every cell is papil-

late. Those of P. paspalodes are narrower and often

thickened distally (Fig. 13, 19).

Thus there are numerous anatomical differences

between P. paspalodes and P. vaginatum and they can

easily be separated on anatomical criteria of the leaf

blade as seen in transverse section. The most out-

standing and obvious difference is that of the scleren-

chyma distribution and shape, as seen in cross-

section, of the leaf margin. Figs. 20-24 illustrate this

clearly, together with the distinct difference in leaf

thickness at the margin. The shape of the adaxial

ribs and furrows and their associated papillae is

another distinct diagnostic difference between the two

species.

Fig. 20-24. — Transverse sections of the leaf margins. All x40. Figs. 20-21. Paspalum vaginatum.

Fig. 20, Ellis 276 bright field. Fig. 21, Wergev 1622 interference contrast. Figs. 22-24. Paspalum

paspalodes. Fig. 22, Ellis 254 bright field. Fig. 23, Ellis 1 106 interference contrast. Fig. 24, Ellis

726 interference contrast.

R. P. ELLIS

239

Epidermis

Epidermal scrapes of both the abaxial and adaxial

epidermides were prepared for microscopical examin-

ation. The description of the epidermis of these two

species is complicated by the fact that in P. paspalodes

two distinct types of abaxial epidermis are present:

those with a distinct, continuous cuticle and having no

papillae (Fig. 27, 28), and those with a thin, indistinct

cuticle and papillate epidermal cells (Fig. 29, 30).

Correlated with this difference are numerous others

in the abaxial as well as the adaxial epidermides.

For the sake of convenience these two forms of P.

paspalodes will be termed cuticular and papillate

respectively.

In P. vaginatum and cuticular P. paspalodes the

abaxial intercostal long cells are elongated, with

parallel side walls and vertical, anticlinal end walls.

These walls are moderately undulating (Fig. 26, 28).

In P. vaginatum cell shape and size are remarkably

constant, but in cuticular P. paspalodes, cell-length

is variable within a single file of long cells. Adjacent

long cells are frequently separated by silico-suberose

couples or solitary cork cells. In P. vaginatum and

cuticular P. paspalodes there are no bulliform cells

on the abaxial surface. In papillate P. paspalodes

(Fig. 29, 30), however, the intercostal long cells are

shortly rectangular or hexagonal in shape and

resemble bulliform cells. They are not interspersed

with intercostal short cells.

Two rows of stomata, separated by more than one

file of intercostal long cells, are found in each inter-

costal zone in both species. Their shape, as deter-

mined by the shape of both subsidiary cells, is low

triangular in P. vaginatum and cuticular P. paspalodes.

In papillate P. paspalodes the stomata are low dome-

shaped, with the subsidiary cells ovoid in shape.

The interstomatal long cells of both species are very

similar in shape and size to the intercostal long cells.

A single interstomatal cell is most commonly present

between successive stomata in P. vaginatum and

cuticular P. paspalodes, but in papillate P. paspalodes

three or more interstomatal cells are found.

No prickle hairs were observed on the epidermal

surface of either of the two species examined.

Macro-hairs were occasionally observed on the

abaxial surface of papillate. P. paspalodes. These are

typical cushion hairs with raised epidermal cells

surrounding the sunken base of the hair (Fig. 13).

Fig. 25-30. — Abaxial epidermis, interference contrast. Figs. 25-26, Paspalum vaginatum. Fig. 25,

Ellis 276 x40. Fig. 26, Ellis 276 xlOO. Figs. 27-30 Paspalum paspalodes. Fig. 27, Ellis 432

> 40. Fig. 28, Ellis 432 x 100. Fig. 29, Ellis 254 X40. Fig. 30, Ellis 146 X 100.

240

LEAF ANATOMY OF P ASP ALUM PASPALODES AND P. VAGINATUM

No micro-hairs were seen on P. vaginal um. On

P. paspalodes (both cuticular and papillate forms)

micro-hairs are infrequently present on isolated parts

of the leaf. These have thickened distal cells tapering

to a point. The basal cell is exceptionally short and

very difficult to see. This is in agreement with Met-

calfe (1960), but Tiirpe (1966) illustrates micro-hairs

with the basal cell up to one third the total length of

the micro-hair.

The silica bodies of papillate P. paspalodes are

horizontally elongated and nodular in outline on

both the abaxial (Fig. 30) and adaxial (Fig. 34)

surfaces. P. vaginatum has horizontally elongated,

but noticeably angular, silica bodies as well, but these

are only present on the adaxial surface. Abaxial

costal silica bodies are very rare (Fig. 25), but crescent-

shaped intercostal silica bodies are common in

P. vaginatum (Fig. 26). Cuticular P. paspalodes has

rows of shortly dumb-bell-shaped silica bodies with

wide central portions in the costal zones (Fig. 27, 28).

Specimens Examined

Paspalum paspalodes. Transvaal. — 2527 (Rusten-

burg): Schoemansville, Hartebeespoort Dam (-DD),

Ellis 432. 2528 (Pretoria): National Botanical Gar-

dens, Brummeria (-CA), Ellis 767. Constantia Park

Extension (-CB), Ellis 146. O.F.S.-2926 (Bloem-

fontein): Botanical Gardens, Bloemfontein (-AA),

van Heerden 95. Natal. — 3030 (Port Shepstone):

Eureka Farm, Izothsa (-CB), Ellis 726. Cape. —

3028 (Matatiele): 16 Km S.W. of Matatiele (-BD),

Ellis 254. 3228 (Butterworth): Fort Warden farm,

Komga (-CC), Ellis 1106.

Paspalum vaginatum. Transvaal. — 2528 (Preto-

ria): Glasshouse in National Botanical Gardens,

Brummeria (-CA), Ellis 1235. Natal. — 2930 (Pieter-

maritzburg): Isipingo beach mangrove fringes (-DD),

IVerger 1622. Cape. — 3228 (Butterworth): Mazeppa

Bay (-BC), Ellis 276. Haga Haga (-CC), Ellis 1114.

3422 (Mossel Bay): Wilderness lagoon (-BB), Ellis

1299.

P. vaginatum and cuticular P. paspalodes bear no

abaxial papillae. In papillate P. paspalodes the abaxial

papillae are inflated, unthickened and dome-shaped,

with a diameter greater than half the vertical width of

the long cell on which it is located (Fig. 30). There are

no papillae on the costal zones, but all intercostal

and interstomatal long cells bear a single, centrally-

positioned papillus. Adaxial papillae are found on

P. vaginatum and both forms of P. paspalodes. They

are extremely large, elongated, oblique and inflated

on P. vaginatum (Fig. 31, 32). These papillae are the

same width as the epidermal cell, which is thus

protruded to form the papillus. Most epidermal cells

on the adaxial ribs and furrows bear papillae. The

adaxial bulliform cells of cuticular P. paspalodes

bear papillae which taper to narrow, thickened distal

ends (Fig. 33), whereas those of papillate P. paspalodes

are more inflated and oblique, but are distally thick-

ened as well (Fig. 34).

DISCUSSION

The material examined can be separated satisfac-

torily by a number of apparently unrelated diagnostic

characters especially of the leaf blade as seen in

transverse section. The most easily studied of these

are the obvious differences of the leaf margin and of

the shape and size of the adaxial ribs and furrows and

their associated papillae. Distinct differences are not

so readily evident in epidermal preparations, because

of the complications caused by P. paspalodes having

two distinct types of epidermis: papillate and cuti-

cular.

Undoubtedly, when the abaxial epidermis is studied

alone (Fig. 25-30), cuticular P. paspalodes resembles

P. vaginatum more closely than it does papillate

P. paspalodes. It can also be reasonably assumed

that specimens of P. paspalodes will be found with

shorter and relatively thicker and narrower leaves

than typical P. paspalodes. Thus differences as seen in

Fie;. 31-34.- Adaxial epidermis, interference contrast. Figs. 31-32, Paspalum vaginatum. Fig. 31,

Ellis 276 x40. Fig. 32, Ellis 276 100. Fig. 33-34, Paspalum paspalodes. Fig. 33, Ellis 432 X 100.

Fig. 34, Ellis 254 X 100.

R. P. ELLIS

241

transverse section might not be so distinct and a

complete range of intermediates may be found from

papillate P. paspalodes, through narrower and thicker

cuticular P. paspalodes to typical P. vaginatum. This

development of an abaxial cuticle, and other associ-

ated differences, on P. paspalodes is not correlated

with a saline environment as was expected. P. vagina-

turn is a maritime species and always grows in a

saline habitat, but the cuticular specimens of P.

paspalodes were collected at Hartebeespoort Dam

(Ellis 432), a pond in the National Botanical Gardens,

Pretoria (Ellis 146) and at Port Shepstone (Ellis 726).

The reason for this cuticular development in moist

habitats is obscure.

The capable technical assistance of Miss E. V. C.

Taylor and Miss L. Breytenbach is acknowledged.

For the photographic enlargements the author is

grateful to Mr S. R. van Jaarsveld.

REFERENCES

Loxton, A. E., 1974. The taxonomy of Paspalum paspalodes

and P. vaginatum Swartz as represented in South Africa.

Bothalia 11: 243-245.

Metcalfe, C. R., 1960. Anatomy of the monocotyledons. 1

Gramineae. Oxford: Clarendon Press.

Turpe, A. M., 1966. Histotaxonomia de las especies argentinas

del genero Paspalum. Lilloa 32: 35-299.

Bothalia 11,3: 243-245 (1974)

The taxonomy of Paspalum paspalodes and P- vaginatum

as represented in South Africa

A. E. LOXTON*

ABSTRACT

The taxonomy and distribution of two species of the subgenus Disticha of Paspalum, as re-

presented in South Africa, are presented. Two distinct, closely related species are recognized, namely

Paspalum paspalodes (Michx.) Scribn. (previously known as P. distichum L., the name being rejected

as a nomen confusion), and P. vaginatum Swartz.

Authors of floras and various taxonomic works

which include the geographical areas occupied by

Paspalum paspalodes (Michx.) Scribn. and P. vagi-

natum Swartz, differ in their taxonomic and nomen-

clatural treatment of these species.

Stapf (1900) recognized two species of the subgenus

Disticha from South Africa, namely P. distichum L.

and P. vaginatum Swartz with its variety nanum

Doell. Stapf in his own hand altered his copy of

FI. Cap. 7: 371 (1900), changing P. distichum to

P. vaginatum and placing P. digitaria Poir. in syno-

nomy under P. distichum.

Chase (1929) and Chippindall (1955) recognized

both P. distichum and P. vaginatum as distinct species,

but not variety nanum Doell of the latter species.

Launert (1970) considered P. vaginatum to be a

synonym of P. distichum. The material which he

cites, does not, in my opinion, agree with either the

material of P. paspalodes or that of P. vaginatum as

defined below.

Clayton (1972), in a more recent work, suggests

that, “the name P. distichum be rejected as a nomen

confusion due to a mistake in typification”. Confusion

arose when it was found that there are two species

represented on the type sheet of P. distichum in the

Linnean Herbarium. The specimen which fits modern

usage of the name P. distichum, i.e. with pubescent

upper glumes, cannot be regarded as the type because

it was collected by Patrick Browne about ten years

after P. distichum was described. The other specimen

matches the material of P. vaginatum, i.e., with gla-

brous upper glumes.

In following Clayton, the name P. distichum is

rejected as a nomen confusion , the correct name for

the plants with pubescent upper glumes being P.

paspalodes (Michx.) Scribn. The plants with glabrous

upper glumes remain as P. vaginatum , and P. distichum

L. is regarded as a synonym of this species.

Morphologically these species may be distinguished

on spikelet characters, the differences being sum-

marized in Table 1. The most important character

used to distinguish them is whether or not the upper

glume of the spikelet is pubescent or glabrous. In

P. paspalodes the upper glume is finely appressed

pubescent, (Fig. 1 C), while in P. vaginatum the upper

glume is glabrous, (Fig. 2 B).

In P. paspalodes the lower glume is usually present

in many of the spikelets of a raceme. The lower glume

is usually present as a small triangular scale, (Fig. 1 B),

more rarely half as long as the spikelet (Fig. 1 D) or,

in some, reduced to a small rim, (Fig. 1 A). The lower

glume of P. vaginatum, however, is absent, (Fig. 2 A).

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria.

Fig. I. — Paspalum paspalodes: A, lower glume reduced to a

rim; B, lower glume a triangular scale; C, upper glume ap-

pressed pubescent; D, lower glume more than half the length

of the spikelet.

Fig. 2. — Paspalum vaginatum: A, lower glume absent; B. upper

glume glabrous.

244

THE TAXONOMY OF PASPALUM PASPALODES AND P. VAGINATUM

Table 1. — Comparison of spikelet characters of Paspalum paspalodes and P. vaginatum.

Anatomically the two species are readily separable,

the major differences lying in the distribution of the

marginal sclerenchyma, the shape and size of the leaf

margins and the shape of the adaxial ribs and fur-

rows. Ellis ( 1974) discusses their anatomical differences

in greater detail.

The distributions and habitat preferences of these

species are quite different. From Fig. 3, it can be seen

that P. paspalodes is the more widespread of the two

and is usually found growing under fresh-water

conditions. P. vaginatum is restricted to saline con-

ditions along the coast, (Fig. 4). The only evidence

of this species moving inland is indicated by a plant

( Ward 2009), which was found growing at Inyamiti

pan in northern Zululand.

Fig. 3. — Distribution of Paspalum paspalodes in South Africa.

Fig. 4.— Distribution of Paspalum vaginatum in South Africa.

Paspalum paspalodes (Michx.) Scribn. in Mem.

Torr. Bot. Club, 5: 29 (1894). “paspaloides”.

Digitaria paspalodes Michx. in FI. Bor. Amer. 1: 46 (1803).

Paspalum digitaria Poir. in Lam. Encycl. Suppl. 4: 316 (1816).

P. distichum auct., non L.

A loose, mat-forming perennial, creeping by means

of slender rhizomes and many-noded stolons; culms,

subcompressed, sparingly branched, erect or decum-

bent up to 50 cm high; leaves well developed, 2-7 mm

wide, up to 14 cm long, usually glabrous except for

a few hairs at the sheath-mouth; ligule a short mem-

brane about 0,5 mm long, with white hairs from

behind; inflorescence of 2 (3) racemes, 1 ,5-7 cm long,

widely expanded to deflexed; spikelets solitary, rarely

in pairs, imbricate, ovate-acuminate, tapering towards

the base, 2, 5-3, 5 (4mm) long, 1,3-1, 5 mm wide

(the variation in size often found on the same plant);

lower glume frequently present as a minute triangular

scale, rarely up to half as long as the spikelet or, in

some, reduced completely; the upper glume minutely

appressed-pubescent, equal in length to the lower

lemma; the midnerve of the lower lemma prominent.

Material Examined:

S.W.A. — 2517 (Gibeon): Wittendorf on the Hardap Dam

(-BB), Giess, Volk & Bleissner 5598.

Transvaal. — 2330 (Tzaneen): 20 km on Witvlag Road from

Louis Trichard t (-AA), Stephen 284 = Oakes 1581. 2527

(Rustenburg): Brits, Hartebeestpoortdam (-DB), de Winter

113; 2528 (Pretoria): Sewage works (-CA), Repton 3475;

Wonderboom, Abercrombie sub PRE 19604; Caledonian

giounds. Repton 3113; Constantia Park extension (-CB), Ellis

146. 2529 (Witbank): Middelburg (-CD), Du Plessis 1170.

2627 (Potchefstroom): Nooitgedacht (-CA), Louw 1631. 2628

(Johannesburg): Rietfontein (-AA), Cohen 878; Springs, Stryd-

pan (-AB), du Toit 8212. 2629 (Ermelo): Nooitgedacht (-DB),

Potter 1453. 2726 (Odendaalsrus): Maquassie, Kommar do-

drift on Vaal River (-CA), Morris 1066.

Orange Free State. — 2726 (Odendaalsrus): Bothaville on

Vais River (-BC), Goossens 1131. 2727 (Kroonstad): Midden-

spruit, 11,3 km north of Kroonstad (-CA), Scheepers 1324.

2826 (Brandfort): Theunissen, Erfenis Dam Reserve (-BC),

Wipplinger 2. 2828 (Bethlehem): Suzanna 16,1 km north of

Bethlehem (-AB), Werger 295. 2925 (Fauresmith): Flood plain,

north of Fauresmith (-CB), Leistner 1091 . 2926 (Bloemfontein):

Botanical gardens (-AA), van Heerdert 95. 3025 (Colesburg):

Springfontein (-BC), Pole-Evans 1649.

Natal. — 2732 (Ubombo): Mkuze (-AC), Strey 5664. 2830

(Dundee): Vantsdrift (-AA) Codd 163; Wee.ien district, Tuaela

estates (-CC), Edwards 2057. 2831 (Nkandla): Black Umfolozi

River ( -3D), Moll 5249. lower Tugela (-CC), Edwards 188);

Mtunzini, Twinstreams (-DD), Moll 5396. 2929 (Underberg) :

Estcourt district (-BB), Edwards 2045. 2930 (Pietermaritzburg):

Lions River (-AC), Moll & Mauve 168); Albert Falls (-AD),

Cumins 362; Table Mountain, outside Amatula forest (-CB),

Killick 413; Chase Valley, Gordon-Gray 146; Durban, Isipingo

North (-DD), Ward 5259; Durban, Umbilo River, Ward 5 51

& 6153. 2931 (Stanger): Mapumulo, Umvoti Valley (-AA),

Moll 1531; Tongaat beach (-CA), Hillary 330. 3030 (Port

Shepstone): Amanzimtoti (-BB), Ward 6 44; Euieka farm, 3 km

from Marburg on Izotsha Road (-CB), Ellis 726. 3130 (Port

Edward): Umtavuma Pont (-AA), Ward 5266.

A. E. LOXTON

245

Cape. — 2820 (Gordonia): Aughrabies Falls (-CB), Acocks

18817. 2824 (Kimberley): Kalkdrift (-DB), Brueckner 787.

3026 (Aliwal North): Aliwal North (-AD), Coetzee A 18,

Higgins 8213. 3028 (Matatiele): 22 km from Matatiele on

Quachas Nek road (-BD), Ellis 254. 3126 (Queenstown) on

bank of Komani River (-DD), Dickin 37. 3221 (Merweville):

Kookfontein, Riet River (-AA), Acocks 2577. 3227 (Stutter-

heim): Komga, Fort Warden (-DB), Ellis 1106. 3318 (Cape

Town): Rondebosch, Keurboom Park (-DC), Adamson 7 1.

3327 (East Eondon): Selbourne (-BB), Rattray 1368; Lily-

fontein, Dickin 15.

Paspalum vaginatum Swartz, Prodr. Veg. Ind.

Occ. 21 (1788).

P. distichum L. Syst. Nat. ed. 10, 2: 855 (1759).

Perennial, up to 69 cm high, creeping by means of

horizontal rhizomes and stout or slender, wiry stolons;

culms simple or branched, subcompressed, usually

grooved, decumbent or erect, often with overlapping,

keeled sheaths and stiffly ascending blades; ligule

membranaceous approximately (about) 0,5 mm long

with a row of white hairs at the back of it; leaves 2,5-

15 cm long, up to 4 mm wide, rarely more; inflores-

cence of 2 (3) racemes, 1,5-7 cm long, conjugate,

usually spreading or deflexed at maturity; spikelets

solitary, imbricate, commonly 3,5-4 (4,5 mm) long,

1,2-1, 5 mm wide, ovate-lanceolate; lower glume

absent; upper glume glabrous, weakly 5-nerved, the

midnerve of upper glume and lower lemma sup-

pressed, not prominent, the lower lemma often trans-

versely undulate.

Materia! Examined :

Natal. — 2632 (Bella Vista): Ingwavuma district, Inyamiti

lake (-CD), Ward 2009. 2732 (Ubombo): Sordwana Bay (-BC),

Ward 3363. 2831 (Nkandla): Umlalazi Nature Reserve (-DD),

Moll 5401. 2930 (Pietermaritzburg): Congella Beach (-DD),

Medley Wood 11982, Franks 12963. 2931 (Stanger): Tugela

River mouth (-AB), Edwards 1727; Durban Bay (-CC), Strey

64/8. 3030 (Port Shepstone): Umbogintwini lagoon (-BB),

Ward 5276.

Cape. — 3228 (Butterworth): Mazeppa Bay (-BC), Ellis 276;

Haga Haga (-CC), Ellis 1114. 3326 (Port A'fred); Kowie,

Saltvlei (-BD), Britten 2996. 3422 (Mossel Bay): Wilderness,

De Vleie (-BB), Ellis 1299; Swartvlei, Cloverdale Beach, Jacot-

Guillarmod, Shaw & Saenger 7090. 3423 (Knysna): Woodbourne

(-AA), Dnthie 878.

I am most grateful to Dr J. H. Ross at Kew for his

valuable assistance in answering numerous queries,

also to Dr O. A. Leistner for his advice as to the

format of the paper and to Mr S. R. van Jaarsveld

for doing the photographic work.

OPSOMMING

Die taksonomie en verspreiding van twee soorte van die

subgenus Disticha van Paspalum, soos in Suid-Afrika verteen-

woordig, word behandel. Twee onderskeibare, maar naverwante

soorte word erken, naamlik Paspalum paspalodes (Michx.)

Seri bn. (voorheen bekend as P. distichum L., maar die naam

word verwerp as ’n nomen confusum), en P. vaginatum Swartz.

REFERENCES

Chase, A., 1929. The North American species of Paspalum.

Contr. U.S. Nat. Herb. 28, 1 : 41-52.

Chippindall, L.K.A., 1955. Paspalum in Meredith et al.. The

grasses and pastures of South Africa 386-391. Johannes-

burg: C.N.A.

Clayton, W.D., 1972. Gramineae in FI. W. Trop. Afr. 3,2:

445, 446.

Ellis, R. P., 1974. Comparative leaf anatomy of Paspalum

paspalodes and P. vaginatum. Bothalia 11,3: 235- 241.

Launert, E., 1970. Gramineae in Prod. FI. S.W. Afr. 160: 146.

Stapf, O., 1899. Gramineae in El. Cap. 7: 370-371

Bothalia 11, 3: 247-257 (1974)

The genus Elephantorrhiza

J. H. ROSS*

ABSTRACT

The genus Elephantorrhiza is reviewed. Nine species, including one new species, E. praetermissa

J. H. Ross, from the eastern Transvaal, are recognized. A description of each species and a key to the

identification of the nine species are provided. Lectotypes are selected for two species. The known

distribution of each species is plotted and attention is drawn to the more serious deficiencies in our

knowledge of some of the species. Concern is expressed over the present conservation status of

several species

INTRODUCTION

Elephantorrhiza, a small genus of nine species

restricted to Africa south of the equator, was founded

by Bentham, in Hook., J. Bot. 4: 344 (1841), on a

specimen collected by Burchell (No. 2410) in the

northern Cape Province. Burchell 2410 was also the

specimen on which Burchell had earlier based his

Acacia elephantina in his Trav. 2: 236 (1824), and on

which De Candolle had based his Acacia elephantor-

rhiza in his Prodr. 2: 457 (1825). Although De Can-

dolle placed this specimen in Acacia , he was not

altogether satisfied with the decision and questioned

whether it was not perhaps a species of Prosopis.

Sprengel, in Sys. Cur. Post. 4: 165 (1827), transferred

De Candolle’s species to Prosopis and later E. Mey.,

Comm. 1: 165 (1836), transferred Burchell’s A.

elephantina to Prosopis.

In founding Elephantorrhiza as a new genus,

Bentham (l.c.) wrote: “ . . . not only do the more

pedicellate flowers, habit and foliage, remove it in

appearance from Prosopis. but the pod described

above (by E. Mey., l.c.) does not at all agree with that

of Prosopis, which is multilocular, indehiscent, and

does not separate from the sutures. The Elephantor-

rhiza is nearer in character to Adenanthera, but the

habit and the pod appear to me to be sufficient to

warrant its being considered as a distinct genus.

I have seen it only in flower”.

The last fairly comprehensive account of the genus

was by Phillips in Bothalia 1: 187 (1923). Much

material and information have become available

since then but, despite this, the present state of

knowledge of this essentially southern African genus

leaves much to be desired. One of the aims of this

paper is to draw attention to the more serious gaps in

our knowledge in the hope that collectors will make an

effort to collect the additional material and make

the field observations necessary to enable a more

comprehensive account of each species to be compiled.

In keying out species of Elephantorrhiza, emphasis

is usually laid on the habit of the plants, that is,

whether the plants are suffrutices with unbranched

aerial stems or whether they are shrubs or small trees

with branched aerial stems. This is the character

employed in the first dichotomy of the key and, as far

as is known, it is a fairly reliable character. However,

E. elephantina, which typically has unbranched

aerial stems, may prove an exception when the growing

apex has been damaged, because then the stems

sometimes develop lateral branches.

Leaflet shape, size and, in particular, the position of

the midrib in the leaflet, are of prime taxonomic

significance in distinguishing the species. Number of

pinnae pairs and, to a lesser extent, number of leaflet

pairs are also useful characters. Pod size, especially

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

pod width, is of taxonomic significance, but in general

the flowers in all species are fairly uniform. E. rangei is

perhaps the exception in that it has larger flowers than

the other species. The colour of the minute glands at

the base of the pedicels is sometimes a useful character.

The underground root systems of each species need

to be investigated as they may provide useful addi-

tional means of distinguishing some of the species. For

example, there is a suggestion that E. obliqua and

E. elephantina have different root systems, but field

observations are necessary to substantiate this.

ELEPHANTORRHIZA

Elephantorrhiza Benth. in Hook., J. Bot. 4: 344

(1841); Harv. in FI. Cap. 2: 277 (1862); Benth.

& Hook, f., Gen. PI. 1:590 (1865); Harv., Gen.

PI., ed. 2: 91 (1868); Benth. in Trans. Linn. Soc.

Lond. 30: 365 (1875); Taub. in Pflanzenfam. 3, 3:122

(1892); Sim, For. FI. Cape Col. 209 (1907); Harms in

Engl., Pflanzenw. Afr. 3, 1: 398 (1915); Phillips in

Bothalia 1: 187 (1923); Bak. f.. Teg. Trop. Afr.

3: 800 (1930); Burtt Davy, FI. Transv. 2: 331 (1932);

Phillips, Gen. ed. 2: 392 (1951); Torre in Consp.

FI. Angol. 2: 263 (1956); Brenan in FI. Trop. E. Afr.

Legum. — Mimos.: 19 (1959); Hutch., Gen. Flow. PI.

1 : 288 (1964); Schreiber in Prodr. FI. S. W. Afr. 58: 16

(1967); Brenan & Brummitt in FI. Zamb. 3, 1 : 23

(1970). Type species: E. elephantina (Burch.) Skeels

(E. burchellii Benth.)

Prosopis sensu E. Mey., Comm. 165 (1836); Eckl. & Zeyh.,

Enum. : 259 (1836), non Prosopis L., Mant. 1 : 10 (1767).

Unarmed small trees, shrubs or suffrutices, often

with a greatly enlarged underground root-stock or a

number of root-stocks. Leaves bipinnate; petioles

eglandular; pinnae 3-42 pairs: pinnae each with many

pairs of leaflets. Inflorescences of spiciform racemes

which are axillary, solitary or fascicled, often

aggregated. Flowers normally hermaphrodite, 5-

nierous, usually pale yellowish-white, on pedicels

1-2 mm long. Calyx gamosepalous, small, 1-2,5 mm

long, 5-toothed. Petals 5, free or shortly united basally.

Stamens 10, fertile, free among themselves, slightly

adnate to the corolla basally; filaments 4-7,5 mm

long; anthers with a usually rapidly deciduous

apical gland. Ovary usually sessile, glabrous; style

filiform; stigma terminal. Pods straight or somewhat

curved, not spirally twisted, often large and up to

45 cm long, somewhat compressed, without transverse

septa within; at maturity the valves separating from

the persistent margins, but not splitting transversely

into segments; the outer layer (exocarp) of the pod-

wall often peeling off the inner layer (endocarp), the

layers remaining intact or breaking up irregularly.

Seeds often ± compressed.

A genus of 9 species restricted to Africa south of the

equator.

248

THE GENUS ELEPHANTORRHIZA

KEY TO SPECIES

Suffrutex; annual aerial stems unbranched (unless damaged), up to 0,75 (1) m high :

Pinnae (1) 2-6 pairs per leaf; leaflets 4-13 pairs per pinna, 2-6,5 mm wide, ovate to ovate-

oblong, very oblique basally, midrib starting in the distal corner of the leaflet-base and

gradually becoming ± central in the leaflet, lateral nerves and veins usually prominent;

confined to the Transvaal 1. E. obliqua

Pinnae (2) 7-17 pairs per leaf; leaflets (7) 12-45 (55) pairs per pinna, 0,5-2 (2,5) mm wide,

linear to linear-oblong, rarely narrowly oblanceolate, base asymmetric; widespread. . . .

2. E. elephantim

Shrub or small tree with woody, branched, aerial stems, up to 7 m high, very rarely a suffrutex

with branched, procumbent stems:

Leaflets with the midrib marginal throughout, in (17) 27-40 (50) pairs per pinna, 3-7, 5x

0,5-1, 2 mm 9. E. suffruticosa

Leaflets with the midrib central or nearly so, at least towards the apex:

Suffrutex; branched aerial stems procumbent, longitudinally striate, glabrous to densely

puberulous; leaves with (2) 5-10 pinnae pairs 3. E. woodii

Shrub or small tree; branched aerial stems erect, glabrous:

Leaflets only slightly asymmetric basally, with the proximal side cuneate to slightly

rounded, 1,5-3, 5 (4,5) mm wide; leaves with (1) 4-8 (9) pairs of pinnae. .4. E. burke

Leaflets ± strongly asymmetric basally, with the proximal side broadly rounded —

truncate to almost auriculate and the distal side cuneate; leaves with 2-41 pairs of

pinnae:

Leaves with (3) 9-30 (41) pinnae pairs (if less than 10 pairs then leaflets usually 4 mm

or more wide); leaflets 0,7-8 mm wide, usually on very short petiolules; pods

1 ,3-2,4 cm wide, very narrow in proportion to their length, when mature the po-

sition of the seeds usually marked by distinct raised bumps; flowers usually pre-

cocious (except in the Transvaal); widespread, but absent from South West

Africa 6. E. goetzei

Leaves with 2-12 (14) pinnae pairs; leaflets 0,9-3, 5 mm wide; pods 2-3,9 cm wide,

± compressed, position of seeds not marked by distinct raised bumps; restricted

to the eastern Transvaal and South West Africa :

Leaflets 0,9-1 ,5 mm wide, sessile; racemes short, 4-5,5 cm long, usually aggregated

on short lateral branchlets, less frequently solitary or fascicled; pods 12-18 X

2-3,2 cm; restricted to the eastern Transvaal 5. E. praetermissa

Leaflets (1)1 ,5-3,5 mm wide; racemes 5, 5-9, 5 cm long, axillary, solitary or paired;

pods (15) 18,5-40,5 x 2-3, 9 cm; restricted to South West Africa:

Pinnae in (2) 6-14 pairs per leaf; calyx up to 1 ,5 mm long; pods 3-3,9 cm wide

8. E. schinziana

Pinnae in 3-7 (9) pairs per leaf; calyx 2-2,25 mm long; pods 2-2,5 cm wide

7. E. rangei

1 . Elephantorrhiza obliqua Burtt Davy in Kew Bull.

1921: 191 (1921); Phillips in Bothalia 1: 189 (1923);

Burtt Davy, FI. Transv. 2: 332 (1932) pro parte excl.

specim. Rogers 22011. Type: Transvaal, 2630

(Carolina), between Carolina and Oshoek, at an out-

span ±1,6 km from Robinson’s, Burtt Davy 2976

(BM, holo.!; FHO,! K!).

Suffrutex producing at ground level annual her-

baceous stems up to 30 cm high from a number of

underground rhizomes; aerial stems usually un-

branched (rarely branched after damage to the main

apex), longitudinally striate, pubescent or glabrous.

Leaves pubescent or glabrous: petiole 2-6 cm long;

rhachis (0) 1,5-9 cm long; pinnae (1) 2-6 pairs;

rhachillae 2-11 cm long; leaflets 4-13 pairs per pinna,

5,5-15x2-6,5 mm, very oblique, ovate to ovate-

oblong, broadly truncate basally, asymmetric and

attached by one corner, midrib starting in the distal

corner of the leaflet-base and gradually becoming

almost centra! in the leaflet, usually with 2-3 other

prominent veins arising from the leaflet-base, midrib

and lateral nerves prominent above and below,

acute or distinctly mucronate apically, glabrous or

sparingly pubescent on the margins. Racemes axillary,

often solitary, on the lov/er or the apical part of the

stem, 3 , 5-6 cm long (including the peduncle), glabrous

or very sparingly pubescent. Flowers yellowish-white,

on pedicels up to I ,5 mm long, with minute reddish

glands at the base of the pedicels. Calyx campanulate,

up to 2 mm long, shortly 5-toothed, glabrous.

Relals shortly united below, up to 4,5 mm long,

I mm wide, linear-oblong, indexed apically, glabrous.

Stamens free among themselves, slightly adnate to the

corolla basally; filaments up to 7,5 mm long; anthers

up to 0,8 mm long, with a deciduous apical gland.

Ovary up to 2 mm long, glabrous, sessile. Pods

unknown.

Apparently confined to the Transvaal, (see Fig. 1)

Occurs in grassland.

E. obliqua is readily distinguished from all other

species by its large ovate leaflets with prominent

venation. E. obliqua appears to have a different under-

ground root system to E. elephantina but field obser-

vations are required to substantiate this.

KEY TO VARIETIES

Stems pubescent; petioles, rhachides and rhachillae sparingly

pubescent (a) var. obliqua

Stems glabrous; petioles, rhachides and rhachillae glabrous

(b) var. glabra

Unfortunately not enough material is available to

evaluate the taxonomic significance of the degree of

pubescence of the stems and leaves as a means of

distinguishing varieties within this species.

(a) var. obliqua.

Phillips in Bothalia 1: 189 (1923); Burtt Davy,

FI. Transv. 2: 332 (1932) pro parte excl. specim.

Rogers 2201 I . Type as above

Transvaal. — 2630 (Carolina): between Carolina and Oshoek,

at an outspan J- 1 ,6 km from Robinson’s farm, Burtt Davy 297u

(BM, FHO, K).

Known only from the type collection. As var.

obliqua has not been re-collected since January 1905,

concern must be expressed about its current conser-

vation status. A thorough search in the type locality is

necessary to establish whether the plant still survives.

More material of E. obliqua , particularly fruiting

material, is required.

J. H. ROSS

249

Fig. I . — The known distributions of Elephantorrhiza burkei, E. obliqua, E.

praetermissa, E. rangei, E. schinziana, E. suffruticosa and E. woodii.

Rogers 22011 from the Petersburg District of the

Transvaal, cited under E. obliquci var. obliqua by

Burtt Davy l.c. : 332 (1932), is in fact Dichrostachys

cinerea (L.) Wight & Arn. subsp. nyassana (Taub.)

Brenan.

(b) var. glabra Phillips in Bothalia 1 : 189. t. 5 fig. 1

(1923); Burtt Davy, FI. Transv. 2: 332 (1932). Syn-

types: Transvaal, 2529 (Witbank), Botsabelo (-CB),

Eiselen sub PRE 1229 (GRA ! ; K ! ; PRE!); Middelburg

(-CD), Jenkins sub TRY 9128 (PRE!).

E. transvaalensis Phillips ined.

Known only from the type collections. As typical

var. glabra has not been re-collected for over fifty

years, there is also concern about its current conser-

vation status. A thorough search in the type localities

is necessary to establish whether the plant still sur-

vives. More material is desired.

Codd 10119 from Bellevue farm near Twentyfour

Rivers in the Waterberg District of the Transvaal

resembles E. obliqua. The stem, petioles, rhachides and

rhachillae are glabrous or almost so, and the leaves

have up to 6 pinnae pairs and up to 19 pairs of leaflets

per pinna. The leaflets are 9-18x3-4 mm, i oblong,

oblique basally with an excentric midrib and two other

prominent veins arising from the leaflet-base, con-

spicuously venose and distinctly mucronate apically.

Although leaflet shape differs somewhat from the

leaflet shape of the syntypes, Codd 10119 is closer to

E. obliqua than to any of the other species and, for the

present, is referred to E. obliqua var. glabra. Un-

fortunately Codd 10119 is sterile. Further collections

are required to indicate whether or not Codd 10119

falls within the range of variation of E. obliqua.

2. Elephantorrhiza elephantina (Burch.) Skeels in

U. S. Dept. Agric. Bur. PI. Ind. Bull. 176: 29 (1910);

Bak. f., Leg. Trop. Afr. 3: 800 (1930); Burtt Davy,

FI. Transv. 2: 332 (1932); O.B. Miller in J.S. Afr. Bot.

18:31 (1952); Leistner in Mem. Bot. Surv. S.

Afr. 38: 123, t. 14 (1967); Schreiber in Prodr.

FI. S. W. Afr. 58 : 1 6 ( 1 967) ; Van der Schijff & Snyrnan

in J. Arn. Arb. 51 : 114 (1970); Brenan & Brummitt in

FI. Zamb. 3,1: 27 (1970); Ross, FI. Natal: 194 (1973).

Type: Cape Province, 2723 (Olifantshoek), between

Matlowing [Mashowing] River and Kuru (-BB),

Burchell 2410 (K. holo.! ; P!).

Acacia elephantina Burch., Trav. 2 : 236 (1824). Type as

above. A. elephantorhiza DC., Prodr. 2: 457 (1825), nom.

illegit. Type as above.

Prosopis elephantorrhiza (DC.) Spreng., Syst. Cur. Post.

IV: 165 (1827); Eckl. & Zeyh., Enum.: 259 (1836). Type as

above. P. elephantina (Burch.) E. Mey., Comn. 165 (1836).

Type as above.

Elephantorrhiza burchellii Benth. in Hook., J. Bot. 4: 344

(1841), nom. illegit.; Harv. in FI. Cap. 2: 277 (1862); Benth. in

Trans. Linn. Soc. Lond. 30: 365 (1875); MacOwan in Agric. J.

Caoe G. H. 10: 29 (1897); Schinz in Mem. Herb. Boiss. 1:116

(1900); Sim, For FI. Cape Col. 209, t. 16, VIII (1907); Dinter,

Deutsch-Siidwest Afrika FI. Forst und landwirtschaft. Frag.

78 (1909); Harms in Engl., Pflanzenw. Afr. 3, I: 400, t. 229

(1915); Dinter in Feddes Repert. 17: 190 (1921); Hofmeyer in

S. Afr. J. Nat. Hist. 3: 215 (1921); Phillips in Bothalia 1:189,

t. 5 fig. 2 (1923); Marloth, FI. S. Afr. 2; fig. 26 (1925). Type as

above. E. rangei (“ rangeri ”) sensu Phillips in Bothalia 1: 192,

t. 5 fig. 5 (1923) pro parte quoad specim. Dinter 2264 et Herb.

Mus. Austro — Afr. 4485 non Harms sensu stricto. E. dinteri

Phillips ined.

Suffrutex producing at ground level annual stems

20-90 cm high from the woody end of a ± elongate rhi-

zome, aerial stems usually unbranched except for in-

florescences (rarely branched after damage to the main

apex); young stems glabrous or rarely pubescent.

Leaves glabrous or sparingly pubescent: petiole 1,3-

3,6 (8) cm long; rhachis 3,5-13,5 (17,5) cm long;

pinnae 2-4 pairs in lower leaves, increasing to 7-17

opposite or subopposite pairs in upper leaves; rhachil-

lae 3-9 ( 10, 5) cm long; leaflets (7) 1 2-45 (55) pairs per

pinna, (4) 5—10 (15) X (0,3) 0,5-2 (2,5) mm, linear

to linear-oblong, rarely narrowly oblanceolate, glab-

rous or almost so. base nearly always asymmetric,

with the proximal side rounded to cuneate, apex

symmetric to asymmetric, acute, usually mucronate or

rarely obtuse, lateral nerves and veins prominent or

not. Racemes usually confined to the lower part of the

stem, axillary, solitary or clustered, (2) 3,5-8 (12) cm

long (including the peduncle), glabrous or very

rarely pubescent. Flowers yellowish-white, on pedicels

up to 1,5 mm long, pedicels articulated near the

middle, with minute reddish-brown glands at the base

250

THE GENUS ELEPHANTORRH1ZA

of the pedicels. Calyx shortly campanulate, up to

1.75 mm long. 5-toothed, glabrous. Petals free or

slightly connate basally, 2,75-3,75 mm long, up to 1

mm wide, linear-oblong, indexed apically, glabrous.

Stamens free among themselves, slightly adnate to the

corolla basally; filaments up to 6,5 mm long; anthers

up to 1 mm long, with a deciduous apical gland.

Ovary up to 1,75 mm long, linear, shortly stipitate.

glabrous. Pods dark brown or reddish-brown,

(5) 9,5-15 (21) x 3-5,7 cm, straight or slightly

curved, oblong, compressed, usually prominently

transversely venose, often umbonate over the seeds,

at maturity the valves separating from the persistent

margins, the outer layer of the pod-wall peeling off

the inner layer, the layers usually breaking up ir-

regularly. Seeds 1 8—26 x 13-18x6-13 mm, ±

ellipsoid.

Found in South West Africa, Botswana, Rhodesia,

Mozambique, the Transvaal, Orange Free State,

Swaziland, Natal, Lesotho and Cape Province,

(see Fig. 2). Occurs in grassland and open scrub;

often gregarious.

As a selection of specimens from Botswana,

Rhodesia and Mozambique was cited by Brenan &

Brummitt in FI. Zamb. 3, 1 ; 27 (1970), no specimens

from these territories are cited here.

Fig. 2. — The known distribution of Flephantorrhiza elephantina

S.W.A. — 2217 (Windhoek): Bodenhausen, Seydel 2388.

2219 (Sandfontein): Oas, Seydel 3761. Grid ref. unknown:

Lichtenstein, Dinter s.n. (Z); Kalahari, Nosob, Fleck 399a

(Z); Fleck 398a (Z).

Transvaal. — 2330 (Tzaneen): Houtbosch, ReUmann 6280(77).

2429 (Zebediela): Percy Fyfe Nature Reserve, Huntley 1492.

2430 (Pilgrim’s Rest): Pilgrim’s Rest, Rogers 23066. 2431

(Acornhoek): Kruger National Park, 29 km from Satara on

Rabelais road. Van der Schijff 3291. 2527 (Rustenburg):

Rustenburg, Nation 225(K). 2528 (Pretoria): Groenkloof,

Phillips 3051. 2529 (Witbank): ±11 km from Middelburg

on road to Hendrina, Marsh 115. 2531 (Komatipoort): Barber-

ton, Galpin 562. 2627 (Potchefstroom): Vereeniging, Burtt Davy

15084. 2628 (Johannesburg): 12,8 km from Heidelberg on

Brakpan road. Marsh 57. 2629 (Bethal): 1 1 km from Ermelo on

road to Hendrina, Marsh 96.

O.F.S. — 2627 (Potchefstroom): Sasolburg, Theron 569. 2727

(Kroonstad): near Kroonstad, Pont 454 (Z). 2825 (Boshof):

24 km from Kimberley along Boshof road, Badenhorst 86.

2827 (Senekal): Doornkop, Goosens 901. 2828 (Bethlehem):

Bethlehem, Phillips 3186. 2829 ( Harrismith), Sankey 35 (K).

2926 (Bloemfontein): P.O. De Burg, Bloemfontein, Cyrus sub

PRE 8794. 2927 (Maseru): Ladybrand, Rogers sub TRV 5057.

Grid ref. unknown: Olifantsfontein, Rehmann 3512 (K,Z).

Swaziland. — 2631 (Mbabane): Black Mbuluzi Falls,

Compton 28 175 ; Evelyn Baring Bridge, Compton 291 60.

Natal. — 2729 (Volksrust): Laingsnek, 18,4 km from New-

castle, Marsh 65. 2732(Ubombo):4,8 km west of Jozini, Lebom-

bo Mts., Edwards 2914. 2829 (Harrismith): Harts Hill, near Co-

lenso, Strey 9942. 2831 (Nkandla): Mtunzini, Mogg 5803. 2832

(Mtubatuba): Hluhluwe Game Reserve, Ward 2704. 2930

(Pietermaritzburg): Inanda, Wood 634 (K,Z). 2931 (Stanger):

Mount Moreland, Wood 2607 (NH).

Lesotho. — 2828 (Bethlehem): Leribe, Dieterlen 46 (P).

Cape. — 2623 (Morokweng): bank of Matlowing River,

between Takoon [Takun] and Molito, Burchell 3210 (K).

2624 (Vryburg): farm Palmyra, 96 km north-west of Vryburg,

Rodin 3532. 2722 (Olifantshoek): between Matlowing River and

Kuril, Burchell 2410 (K,P). 2723 (Kuruman): between source of

Kuruman River and Kosi Fontein, Burchell 2537(K). 2824

(Kimberley): Kimberley, Marloth 852 (GRA). 3126 (Queens-

town): “Prospect”, Queenstown, Galpin 1917 (GRA). 3226

(Fort Beaufort): Shiloh, Baur 379 (GRA,K); between Klip-

plaatrivier and Swart Kei, Drege (BM,K).

E. elephantina , commonly known as “Elands-

boontjie”, is the commonest and most widespread

species. E. elephantina shows considerable variation in

the number of pinnae pairs and in the number, size

and shape of the leaflets. This variation appears

to some extent to be geographical. There is a tendency

for specimens from South West Africa, Botswana,

the western portion of Rhodesia, the western Trans-

vaal, Orange Free State and northern Cape to have

leaves with fewer than 10 pinnae pairs, fewer than

26 leaflet pairs per pinna and leaflets more than 8x 1

mm. The leaflets in these areas are frequently glaucous

and the midrib is close to the distal margin basally but

gradually becomes ± centric so that the leaflets are

± symmetric apically. In the eastern areas of Rhode-

sia, Mozambique, the eastern Transvaal, Swaziland

and Natal there is a tendency for specimens to have

leaves with more than 10 pinnae pairs, more than 26

leaflet pairs per pinna and leaflets less than 8 mm

long and 1 mm wide. The midrib in these specimens is

very close to the distal margin of the leaflets

throughout their length as in E. suffruticosa

and the leaflets are asymmetric apically. The

Pretoria district of the Transvaal appears to be a

critical area, for to the west the leaves tend to have

fewer than 10 pinnae pairs and fewer pairs of large

leaflets, while to the east the leaves tend to have more

than 10 pinnae pairs and more numerous pairs of

smaller leaflets. The extremes, for example Seydel

3761 from Oas in South West Africa and Wood 634

from Inanda in Natal, look very different but as

there is T continuous variation throughout and the

individual characters often vary independently, no

means has been found of delimiting the two groups

satisfactorily.

The thick, red, underground root-stocks were at one

time used for tanning and dyeing. Burtt Davy, FI.

Transv. 2: 332 (1932), reports having dug up a root-

stock it 8 metres long at Vereeniging.

Although the exocarp of the ripe pod is fairly hard,

it readily absorbs water and soon starts to disintegrate.

Seeds often germinate within the moist disintegrating

pods on the surface of the soil. The interesting and

unusual type of germination of the seeds of E. ele-

phantina is discussed by Hofmeyer in S. Afr. J.

Nat. Hist. 3: 215 (1921) and by Van der Schijff

and Snyman in J. Arm Arb. 51 : 1 14 (1970).

3. Elephantorrhiza woodii Phillips in Bothalia

I : 193, t.5 fig. 6 (1923). Type; Natal, 2829 (Harrismith)

Pieters, near Colenso (-DB), Wood 7958 (NH,

holo.!).

J. H. ROSS

251

Suffrutex (but see below) producing at ground

level annual, procumbent, branched stems up to

60 cm long, from an elongate rhizome; aerial stems

longitudinally striate, glabrous to densely puberulous.

Leaves glabrous to densely puberulous: petiole

0,8-1, 6 cm long; rhachis (1)3,5-8,5(13) cm long,

distinctly sulcate above, pinnae (2)5-10 opposite or

subopposite pairs; rhachillae 1,8-6 cm long; leaflets

12-28 pairs per pinna, 2, 5-6(9) x 1-1 ,8(2,25) mm,

linear to linear-oblong, glabrous, asymmetric basally,

midrib starting in the distal corner of the leaflet-base

and gradually becoming almost central in the leaflet,

proximal side of base rounded, apex symmetric or

asymmetric, acute or obtuse, mucronate, midrib

prominent or not, lateral veins indistinct. Racemes

axillary, usually solitary, 4, 5-9, 5 cm long (including

the peduncle), glabrous to densely puberulous. Flowers

yellowish-white, on pedicels up to 1,25 mm long,

pedicels articulated near the middle, with minute

glands at the base of the pedicels. Calyx up to 1,5

mm long, shortly 5-toothed, glabrous. Petals free or

slightly connate basally, up to 3 , 25 mm long, 1 , 25 mm

wide, lanceolate, slightly inflexed apically, glabrous.

Stamens free among themselves, slightly adnate to the

corolla basally; filaments up to 6 mm long; anthers

up to 1 mm long, with a deciduous apical gland.

Ovary up to 2 mm long, linear, subsessile, glabrous.

Pod (only one ever collected) dark reddish-brown,

9x3,2 cm, falcate, compressed, prominently trans-

versely venose, umbonate over the seeds. Seeds

immature.

Found in Natal and Lesotho (See Fig. 1). Occurs in

grassland.

Phillips, in Bothalia 1: 189 (1923), keyed out

E. woodii under the species which are “shrubs or

small trees with a distinct aerial stem”. As there is no

information about the habit of the plants in Medley

Wood’s collector’s book or on the herbarium sheets,

it is thought that Phillips assumed E. woodii was a

shrub or small tree because the aerial stems are

branched. During recent years several attempts

have been made to re-collect E. woodii in the type

locality and eventually, in 1970, Mr R. G. Strey

succeeded in finding a few plants. However,

the plants were procumbent and not shrubby

or arborescent as indicated by Phillips. The habit

notes in the above description are based on the speci-

mens collected by Mr Strey and subsequently culti-

vated at the Natal Herbarium, Durban. Each year the

branched, procumbent, aerial stems die back and are

replaced by a new set of stems the following season.

The stems are relatively weak and show no signs of as-

suming an erect posture. More field observations are

required to establish whether E. woodii ever does grow

as an erect shrub. At the time of collecting the living

plants, concern was expressed for the future of the

species in the type locality, at least, because the plants

survive only in the narrow strips of uncultivated land

which may themselves be cultivated at any time.

E. woodii is closely related to E. elephantina but

differs in having branched, procumbent, aerial stems

and leaflets with a slightly different venation. It

differs from all the other species with branched

stems in that the stems are procumbent and longi-

tudinally striate.

Key to Varieties

Stems glabrous or almost so; petioles, rhachides and

rhachillae glabrous or almost so; inflorescence

axes ± glabrous (a) var. woodii

Stems pubescent; petioles, rhachides and rhachillae

pubescent; inflorescence axes pubescent, .(b) var. pubeseens

Unfortunately not enough material is available

to evaluate the taxonomic significance of the degree

of pubescence of the stems, leaves and inflorescence

axes as a means of distinguishing varieties within

this species.

(a) var. woodii.

Phillips in Bothalia 1: 193 (1923); Ross, FI. Natal;

194 (1973). Type as above.

Natal.— 2829 (Harrismith): Pieters (-DB), Strey 9760;

Harts Hill, near Colenso (-DB) Strey 10000; Ladysmith

(-DB), Geekie 46 (NU).

More material of var. woodii is required.

The following note written by Phillips on Wood

7958 (NH) indicates that he had initially intended

calling this specimen E. wahlbergii (Harv.) Phillips:

“This is undoubtedly the plant described by Harvey

as Entadal wahlbergii in FI. Cap. 2: 277. I propose

the new combination Elephantorrhiza wahlbergii ,

Phill. 8/6/19”.

(b) var. pubeseens Phillips in Bothalia 1: 193

(1923); Ross, FI. Natal 194 (1973). Type: Natal,

probably 2929 (Underberg), near Little Tugela, 1219

metres (-BA), Wood 2867 (NH, holo. !)

E. pubeseens Phillips ined.

Natal.— probably 2929 (Underberg): near Little Tugela,

Wood 286 7 (NH). Grid ref. unknown: between Pietermaritzburg

and Newcastle, Nov. 1883, Wilms 1973 (BM).

Lesotho?. — no precise locality, Cooper 2279 (K).

Wilms 1973 (BM), collected between Pietermaritz-

burg and Newcastle in 1883, appears to be referable

to E. woodii var. pubeseens. However, the specimen has

wider leaflets and a somewhat different facies to

typical var. pubeseens. More material is desirable

but, as the specimen was not well localized, there

seems little likelihood of further material being

collected.

Cooper 2279, a flowering specimen with immature

leaves, also appears to be referable to var. pubeseens.

Although recorded as being collected in “Basutoland,”

Cooper is not known to have been in present-day

Lesotho.

Mogg sub PRE 9644, an immature and rather frag-

mentary specimen from Charlestown in northern

Natal, is extremely difficult to place with certainty. The

specimen in the Kew Herbarium consists of a single

stem bearing axillary racemes and very young foliage,

the details of which are scarcely discernible. The stem

is 17,5 cm high, longitudinally striate and pubescent.

The leaves have 2 pinnae pairs and up to 1 6 leaflets per

pinna. The immature leaflets are up to 4,5 >< 1,75

mm, the midrib is excentric basally and the leaflet-apex

is distinctly mucronate. This specimen is hesitantly

referred to E. woodii var. pubeseens, but additional and

better material from this area is required to establish

the identity of the plants.

More material of var. pubeseens, and from definite

localities, is required. Pods would be of particular

interest.

4. Elephantorrhiza burkei Benth. in Hook., Lond.

J. Bot. 5: 81 (1846); Harv. in FI. Cap. 2: 278 (1862);

Benth. in Trans. Linn. Soc. Lond. 30: 365 (1875);

Phillips in Bothalia 1: 192, t.5 fig. 4 (1923); Bak.f.,

Leg. Trop. Afr. 3: 801 (1930); Burtt Davy, FI.

Transv. 2: 332 (1932); O. B. Miller in J. S. Afr. Bot.

18: 31 (1952); Wild, Guide FI. Viet. Falls 149 (1953);

Brenan & Brummitt in FI. Zamb. 3, 1 : 27 (1970).

Type: Transvaal, Magaliesberg, Burke & Zevher

(K, holo.!;BM!,TCD!,Z!).

E. elephantina (Burch.) Skeels var. burkei (Benth.) Merr. in

Contr. Gray Herb. 59: 18 (1919). Type as above.

17749-4

252

THE GENUS ELEPHANTORRHIZA

A branched shrub or small tree 1-3(6) m high,

occasionally as small as 0,3 m, but then the stems

distinctly woody and branched and the inflorescences

normally borne on lateral shoots of the current

season’s growth; bark dark grey to reddish; young

branchlets glabrous. Leaves glabrous or almost so:

petiole 2, 6-6, 5 cm long; rhachis 3,6-14,5 cm long;

pinnae (1 )4— 8(9) pairs; rhachillae 3,5-12,5 cm long;

leaflets (9)12-23(32) pairs per pinna, 7—17 x 1,5-3, 5(5)

mm, narrowly oblanceolate to very narrowly elliptic

or linear-oblong, usually glaucous, glabrous, base

slightly asymmetric (less so than in E. elephant ina ),

with the proximal side rounded to cuneate, apex

symmetric, obtuse to rounded, generally mucronate,

lateral nerves and veins prominent or not. Racemes

axillary, solitary or fascicled, often on lateral shoots,

5-10(12,5) cm long (including the peduncle), glabrous.

Flowers yellowish-white, on pedicels up to 2 mm long,

pedicels articulated near the middle, with minute

reddish glands at the base of the pedicels. Calyx

campanulate, up to 2,5 mm long, 5-toothed, the

teeth up to 0,75 mm long, glabrous. Petals shortly

united basally, up to 4,5 mm long, 1 mm wide,

linear-oblong, inflexed apically, glabrous. Stamens

free among themselves, slightly adnate to the corolla

basally; filaments up to 5 mm long; anthers up to

0,75 mm long, with a deciduous apical gland. Ovary

up to 2 mm long, glabrous. Pods dark brown to

reddish-brown, 10-19(28) x 2,5-4 cm, straight or

slightly curved, oblong, compressed, sometimes

prominently transversely venose, at maturity the

valves separating from the persistent margins, the

outer layer of the pod-wall peeling off the inner layer,

the layers remaining intact or breaking up irregularly.

Seeds ±9-13 x 8-12 mm.

Found in Botswana, Rhodesia, Mozambique and

the Transvaal (See Fig. 1). Favours rocky situations,

in woodland, grassland and scrub.

As specimens from Botswana, Rhodesia and

Mozambique were cited by Brenan & Brummitt

in FI. Zamb. 3, 1: 27(1970), it is not considered

necessary to cite specimens from these territories here.

Transvaal. — 2329 (Pietersburg): University College of the

North, 28,8 km from Pietersburg on road to Tzaneen, Van

Vuuren 1293. 2425 (Gaberones): Lekkerlach, Louw 59/ (NH).

2428 (Nylstroom): 15,2 km north of Warmbaths, Marais 1236.

2429 (Zebediela): Chuniespoort, Pole Evans sub PRE 19452.

2431 (Acornhoek): Kruger National Park, 25,6 km north-east

of Skukuza, Codd 5723. 2527 (Rustenburg): Wolhuterskop,

Schweickerdt 1642. 2528 (Pretoria): end of Daspoort range

about 9,6 km east of Pretoria, Phillips 3040. 2529 (Witbank):

Loskop Dam Nature Reserve, Mogg 30609.

E. burkei differs from E. elephantina primarily in

being a shrub or tree with branched, perennial,

aerial stems and not a suffrutex with annual, un-

branched, aerial stems (unless damaged). The leaflet-

base in E. burkei is less asymmetric than in E.

elephantina , and the midrib soon becomes ± centric.

The leaflets of E. burkei are typically larger than those

of E. elephantina. E. burkei appears to have smaller

seeds than E. elephantina , but more fruiting material

is required to confirm this.

5. Elephantorrhiza praetermissa J. El. Ross, sp.

nov., E. goetzei (Harms) Harms typicae affinis,

sed foliorum pinnis constanter paucioribus et legu-

minibus brevioribus et latioribus differt; E. elephan-

tinae (Burch.) Skeels etiam affinis sed caulibus

ramosis; ab utraque insuper foliolis differt.

Frutex 1-2 m altus; ramuli juveniles griseo- vel

rubro-brunnei, glabri. Folia glabra: petiolus 2,2-4 cm

longus; rhachis 4-9 cm longus; pinnae (3)5-10(12)-

jugatae; rhachillae (2, 8)3, 5-6(7) cm longae; foliola

20-40-juga, 5-10 mm longa, 0,9-1, 5 mm lata,

lineari vel lineari-oblonga, sessilia, glabra, basi

asymmetrica, costa basi obliqua, apice ± centralis,

latere proximo rotundata, apice rotundata vel acuta,

fere symmetrica, nervis lateralibus subtus haud

conspicuis vel obscuris. Inflorescentia racemosa,

racemis solitariis, fasciculatis vel in ramulis lateralibus

aphyllis abbreviatis aggregatae, 4-5,5 cm longis,glabris.

Elores ochroleuci, pedicellati, pedicellis 1,5-2 mm

longis, articulatis. Calyx 0,75-1,25 mm longus,

quinquedenticulatus, glaber. Petala basi breviter

connata, 2-3 mm longa, lineari-oblonga, glabra.

Stamina 10; filamenta 4-5 mm longa; antherae apice

glandula caduca coronatae. Ovarium ± 2 mm longum,

lineare, glabrum. Legumina atrobrunnea vel rubro-

brunnea, 12-18 cm longa, 2-3,2 cm lata, oblongo,

recta plus minusve curvata, compressa, obscure vel

prominenter nervosa. Semina ± 15x13x3,5 mm.

Type : Transvaal, 2430 (Pilgrim’s Rest), Steelpoort

valley, near Sarashof (-CC), Codd 9830 (PRE,

holo.; BM, K, iso.).

Shrub 1-2 m high; young branchlets grey- or

reddish-brown, glabrous. Leaves glabrous: petiole

2,2-4 cm long; rhachis 4-9 cm long, sulcate above,

sometimes with minute, scattered, dark glands;

pinnae (3) 5-10(12) opposite or subopposite pairs;

rhachillae (2 , 8) 3 , 5-6 (7) cm long ; leaflets 20-40 pairs,

5-10x0,9-1,5 mm, linear or linear-oblong, sessile,

glabrous, asymmetric basally, midrib starting in the

distal corner of the leaflet-base and gradually be-

coming almost central in the leaflet, proximal side of

base rounded, apex rounded or acute, nearly sym-

metric, lateral nerves not visible or inconspicuous

beneath, sometimes with minute, dark purplish glands,

at the base of the leaflets. Inflorescences racemose,

racemes solitary, fascicled or aggregated on abbre-

viated lateral branchlets, 4-5,5 cm long (including the

peduncle), glabrous. Flowers yellowish-white, pedi-

cellate, pedicels 1,5-2 mm long, articulated near :>r

below the middle, with minute, dark reddish glands

at the base of the pedicels. Calyx 0,75-1,25 mm

long, 5-toothed, glabrous. Petals shortly united basally,

2-3 mm long, linear-oblong, glabrous. Stamens free

among themselves, slightly adnate to the corolla

basally; filaments 4-5 mm long; anthers with a

deciduous apical gland. Ovary ± 2 mm long, linear,

glabrous. Pods dark brown or reddish-brown,

12-18x2-3, 2 cm, oblong, straight or slightly curved,

compressed, obscurely or prominently venose, at

maturity the valves separating from the persistent

margins. Seeds ±15x13x3,5 mm.

E. praetermissa appears ffo have a rather restricted

distribution in the eastern Transvaal (see Fig. 1).

Occurs on dry wooded hillsides.

Transvaal. — 2429 (Zebediela): 25 km NNW of Schoonoord

(-DB), Acocks 20969. 2430 (Pilgrim's Rest): 59 kmfromLyden-

burg on road to Steelpoort via Tweefontein (-CC), Vorster

2129; 74 km from Lydenburg on road to Steelpoort via Twee-

fontein (-CC), Vorster 2128; near Sarashof (-CC), Codd 9830.

The first specimen of this plant was collected by

Dr L. E. Codd in Sekukuniland in the eastern Trans-

vaal some fifteen years ago, and the second by Mr

J. P. H. Acocks two years later. In response to an ap-

peal for fruiting material, Mr P. Vorster collected

some excellent specimens earlier this year.

The specimens, which are fairly uniform vegetative-

ly and appear to represent a distinct entity, cannot be

accommodated satisfactorily within any of the existing

species of Elephantorrhiza. There are a number of

small differences which, collectively, distinguish

the specimens from those of related species and appear

to be of significance.

J. H. ROSS

253

E. praetermissa is most closely related to E. goetzei

and to E. elephanlina. Both E. goetzei and E. elephan-

tina occur in the eastern Transvaal but the specimens

in question differ from the material of both of these

species.

E. praetermissa differs from typical E. goetzei in

having consistently fewer pinnae pairs. Although

from 3-41 pinnae pairs are recorded in E. goetzei ,

specimens with fewer than 12 pinnae pairs usually

have large leaflets which are ±4-8 mm wide, and

these have been separated as E. goetzei subsp. lata. The

fewer pinnae pairs distinguish E. praetermissa from

typical E. goetzei , and the narrow leaflets distinguish

E. praetermissa from E. goetzei subsp. lata. The

leaflets of E. praetermissa differ from those of typical

E. goetzei in having a somewhat thicker texture and

in being ± sessile; the leaflets of typical E. goetzei

usually have distinct petiolules. In E. goetzei the pods

are long and narrow in proportion to their length

(15-44 x 1 ,3-2,4 cm) and, when mature, the position

of the seeds is marked by distinct raised bumps. In

E. praetermissa the pods are shorter and broader

(12-18x2-3,2 cm), ± compressed, and lack distinct

raised bumps over the seeds. The seeds of E. praeter-

missa are ± compressed in constrast to the ellipsoid

or lenticular seeds of E. goetzei , and they are smaller

than those of the latter. Although the length of the

racemes provides no discontinuity between the two

species, the racemes of E. praetermissa are consistently

short and are much shorter than is usual in E. goetzei.

E. praetermissa differs from E. elephanlina in being a

robust shrub 1-2 m high and in having branched aerial

stems. The leaflets of E. praetermissa differ slightly in

texture and lack the ± conspicuous venation of typical

E. elephantina , while the pods tend to be slightly

narrower than is usual in E. elephantina.

Although E. praetermissa is described as locally

common by collectors, very few specimens have been

collected. More material is required.

6. Elephantorrhiza goetzei (Harms) Harms in

Engl., Pflanzenw. Afr. 3, 1: 400 (1915); Bak. f.,

Leg. Trop. Afr. 3: 802 (1930); Brenan, Checklist

Tang. Terr. 344 (1949); Wild, Guide FI. Viet. Falls

149 (1953); Williamson, Useful PI. Nyasal. 52 (1955);

Brenan in FI. Trop. E. Afr. Legum. — Mimos. 19,

fig. 4 (1959); White, For. FI. N. Rhod. 91 (1962);

Brenan & Brummitt in FI. Zamb. 3,1 : 24, t. 4 (1970).

Type: Tanzania, Rufiji District, Goetze 82 (B, holo.f,

BM, drawing!; K!).

Piptadenia goetzei Harms in Bot. Jahrb. 28: 397 (1900).

Type as above.

Shrub or small tree 1-4 (7) m high; bark grey-

brown to dark brown or reddish-brown to purplish,

often becoming blackish; young branchlets glabrous.

Leaves glabrous or nearly so: petiole 1-5 (7,5) cm

long; rhachis 6-20 (45,5) cm long, sulcate above;

pinnae 3-30 (41) opposite or subopposite pairs;

rhachillae 1 ,8-9 (11) cm long; leaflets 9-40 (48) pairs,

3,5-9 (22) x 0,7-8 mm, linear-oblong to narrowly

oblong, midrib starting in the distal corner of the

leaflet-base, gradually becoming almost central in the

leaflet, proximal side of the base rounded and almost

auriculate, apex acute to rounded and mucronate,

nearly symmetric, glabrous, lateral nerves and veins

not or scarcely visible. Racemes solitary, fascicled or

borne on short lateral shoots, (2) 5-20 (23) cm long

(including the peduncle), glabrous. Flowers yellowish-

white, sometimes tinged with pink or purple, on

pedicels up to 1 mm long, pedicels articulated near

the middle, with minute, pale yellowish-white glands

at the base of the pedicels. Calyx 1 ,5-1 ,75 mm long,

with 5 acute teeth, glabrous. Petals shortly united

below, becoming almost free in open flowers, 2,5-3

mm long, linear-oblong, inflexed apically, glabrous.

Stamens free among themselves, slightly adnate to the

corolla basally; filaments up to 4,5 mm long; anthers

up to 1 mm long, with a deciduous apical gland.

Ovary up to 2 mm long, linear, glabrous. Pods dark

brown or reddish- or purplish-brown, 1 5-30 (44) x 1 , 3

-2,4 cm, linear, straight or curved, raised over the

seeds, at maturity the valves separating from the

persistent margins, the outer layer of the pod-wall

peeling off the inner layer, the layers remaining intact

or breaking up irregularly. Seeds 11-20x9-18x7-12

mm, ellipsoid to lenticular.

Key to subspecies

Leaves with (3) 14-41 pairs of pinnae; rhachillae 3, 5-9, 5

cm long; leaflets in (11) 20-48 pairs, 3,5-12 x 0,7-

3 mm (a) subsp. goetzei

Leaves with 4-15 pairs of pinnae; rhachillae 6,5-15 cm

long; leaflets in 9-28 pairs, mostly 12-22 x 4-8 mm

(b) subsp. late,

(a) subsp. goetzei.

Brenan & Brummitt in Bol. Soc. Brot., Ser. 2, 39:

189 (1965); FI. Zamb. 3, 1 : 24 (1970).

E. rubescens Gibbs in J. Linn. Soc. Bot. 37: 441 (1906)

Eyles in Trans. Roy. Soc. S. Afr. 5: 364 (1916). Type: Rhodesia,

Matopo Hills, Gibbs 184 (BM, holo.!). E. cf. peter siana sensu

Gomes e Sousa, PI. Menyharth. 70 (1936). E. cf. goetzei

(Harms) Harms, Torre in Consp. FI. Angol. 2: 263 (1956).

E. sp. sensu Torre l.c. 264 (1956).

Acacia rehmanniana sensu Exell in Bol. Soc. Brot. Ser. 2,

12: 16(1937), non Schinz.

Found in Tanzania, Angola, Botswana, Zambia,

Rhodesia, Malawi, Mozambique and the Transvaal

(See Fig. 3). Occurs in woodland of various types and

scrub; favours rocky places.

As a selection of specimens from Tanzania, Angola,

Botswana, Zambia, Rhodesia, Malawi and Mozam-

bique have been cited in recent regional floras, no

further specimens from these territories are cited here.

Transvaal. — 2331 (Phalaborwa): Kruger National Park,

Shingwedzi, in Lebombo Mts., Van der Schijff 3848. 2429

(Zebediela): Wolkberg, 14,4 km south of Boyne on road to

Welcome Gold Mine, Codd 10393. 2430 (Pilgrim's Rest): 28

km south east of Gravelotte (-BB), Codd 9477; Abel Erasmus

Pass, c. 3 km south of Strydom Tunnel (-BC), Vorster &

Coetzer 2099. 2431 (Acornhoek): Kruger National Park,

±29 km from Satara on Rabelais road (-BC), Van der Schijff

3290\ Kruger National Park, 10,4 km from Nwanedzi [Ngwa-

netsi] on Satara road (-BC), Story 3967.

In the area delimited for Flora Zambesiaca E.

goetzei frequently produces its flowers when the plant

is leafless. In the Transvaal, however, E. goetzei

usually produces flowers together with the leaves.

As leaves are required for establishing which

subspecies specimens belong to, it is not possible to

refer many of the leafless flowering specimens from the

Flora Zambesiaca area to either subspecies with

certainty. These leafless flowering specimens have been

plotted under the category “subsp. uncertain” on the

accompanying distribution map of E. goetzei (see

Fig. 3).

It is possible that E. petersiana Bolle in Peters,

Reise Mossamb. Bot. 1: 9 (1861) is an earlier name

for E. goetzei. If this were ever confirmed, then E.

petersiana would be the correct name for this species.

The holotype of E. petersiana , now destroyed, was a

flowering specimen (without leaves) collected by Peters

at Sena in Mozambique. Unfortunately the type

description is too imperfect to enable the species to be

positively identified. Bak.f., Leg. Trop. Afr. 3: 802

(1930) shed no light on the identity of E. petersiana.

254

THE GENUS ELEPHANTORRHIZA

Burtt Davy, FI. Transv. 2: 332 (1932) based his E.

(?) sp. nov. ? (=£.(?) elongata Burtt Davy ined.)

on Burtt Davy H 2304 collected at Potgietersrust in the

Transvaal. This specimen, which is quite leafless,

has ± straight immature pods up to 21x1,7 cm.

The pods of Burtt Davy H 2304 resemble those of

E. goetzei fairly closely and, although E. goetzei

has not been recorded from Potgietersrust sub-

sequently, it seems likely that Burtt Davy’s specimen is

referable to E. goetzei. E. (?) sp. nov. ? is therefore a

probable synonym of E. goetzei.

(b) subsp. lata Brenan & Brummitt in Bol. Soc.

Brot., Ser 2, 39: 189 (1965); FI. Zamb. 3, 1 : 26

(1970). Type: Zambia, Katombora, Morze 55 (FHO,

holo!.,K fragm.!).

E. sp. 1— White, For. FI. N. Rhod.: 91 (1962).

Known only from Zambia and Rhodesia (See Fig.

3). Occurs in woodland of various types.

An adequate selection of specimens has already been

cited in Bol. Soc. Brot., Ser. 2, 39: 189 (1965) and

in FI. Zamb. 3, 1 : 26 (1970).

7. Elephantorrhiza rangei Harms in Bot. Jahrb. 49:

420 (1913); Dinter in Feddes Repert. 17: 190 (1921);

Bak. f., Leg. Trop. Afr. 3: 802(1930); Range in Feddes

Repert. 30: 148 (1932); Schreiber in Prodr. FI. S. W.

Afr. 58 : 17 (1 967) pro parte quoad specim. Range 455.

Type: South West Africa, 2617 (Bethanie), Naute,

near Keetmanshoop (-DD), Range 455 (B, holo.t, BM,

drawing!; SAM, iso.!)

A branched shrub (? or small tree) to 4 m high;

young branchlets reddish-brown to purplish, glabrous.

Leaves glabrous: petiole 2-4,5 cm long; rhashis

(1,5) 3-7,5 cm long [petiole and rhachis together

described as 2-15 cm long by Flarms]; pinnae in 3-7

[9] opposite or subopposite pairs, sometimes in 1-3

pairs on immature leaves; rhachillae 5-8, 5 [9] cm

long; leaflets 24-36 pairs per pinna, 6-9 [12] xl

-2,75 ]4] mm, linear-oblong to oblong, sometimes

slightly falcate, midrib starting in the distal corner

of the leaflet-base, gradually becoming almost central

in the leaflet, proximal side of the base rounded to

almost auriculate, apex rounded to acute, mucronate,

almost symmetric, lateral nerves inconspicuous,

glabrous. Racemes axillary, solitary or paired, 5,5-8

cm long (including the peduncle), glabrous. Flowers

greenish-yellow, on pedicels 1-1 ,75 mm long, pedicels

articulated just belowthe middle, with minute yellowish

glands at the base of the pedicels. Calyx campanulate,

2-2,25 mm long, glabrous, 5-toothed. Petals shortly

united basally, 3-4 mm long, 1 mm wide, oblong,

indexed apically, glabrous. Stamens free among

themselves, slightly adnate to the corolla basally;

filaments up to 5,5 mm long; anthers up to 0,9 mm

long, with a deciduous apical gland. Ovary up to 4 mm

long, shortly stipitate, linear, glabrous. Pods dark

brown or reddish-brown, 18,5-20 [22] x 2-2, 5 cm,

straight or almost so, oblong, compressed, slightly um-

bonate over the seeds, transverse venation relatively

inconspicuous, at maturity the valves separating

from the persistent margins. Seeds unknown.

J. H. ROSS

255

Known only from the type locality in South West

Africa (See fig. 1). Ecology unknown.

S.W.A. — 2617 (Bethanie): Naute, near Keetmanshoop (-DD),

Range 455 (SAM).

The above description was drawn up from two

isotypes in the SAM Herbarium. The extreme dimen-

sions given in square brackets were recorded by

Harms in his type description.

There is considerable variation in leaflet size even on

a single branch; the upper leaves often have small

leaflets and the lower leaves larger leaflets.

E. rangei bears a superficial resemblance to E.

sujfruticosa but differs in having larger and broader

leaflets in which the midrib is ± centric apically,

slightly longer pedicels, and larger flowers.

The specimens cited by Phillips in Bothalia 1 : 192

(1923) under E. rangei (“ ranged ”), and on which

t.5 fig. 5 was based, are in fact referable to E.

elephantina.

E. rangei is known only from the type collection.

It has never been re-collected since Jan. 1908 and the

possibility exists that it is now extinct. A thorough

search for this plant in the type locality is most

desirable in an attempt to evaluate its present con-

servation status.

8. Elephantorrhiza schinziana Dinter in Feddes

Repert. 17: 190 (1921); Bak. f.. Leg. Trop. Afr.

3: 802 (1930); Schreiber in Prodr. FI. S. W. Afr. 58: 17

(1967). Syntypes: South West Africa, 1917 (Tsumeb),

farm Heidelberg near Tsumeb (-BB), Dinter 1689

(SAM!); Otavi(-CB), Dinter 745 (SAM!)

A branched shrub ( ? or small tree) up to 2 , 5 m high ;

bark grey- to dark or reddish-brown; young branchlets

grey- or reddish- to purplish-brown, glabrous.

Leaves glabrous: petiole 2, 2-3, 5 (5,2) cm long;

rhachis (4,5) 7,5-14,5 (20,5) cm long; pinnae (2)

6- 11 (14) opposite or subopposite pairs; rhachillae

5,5-10 (14) cm long; leaflets (14) 21-40 pairs per pin-

na, (5) 7-14x1,5-3,5 mm, linear-oblong to oblong,

midrib starting in the distal corner of the leaflet-base,

gradually becoming almost central in the leaflet,

proximal side of base rounded, apex rounded and

sometimes distinctly mucronate, nearly symmetric,

lateral nerves scarcely visible, glabrous, somewhat

glaucous. Racemes axillary, solitary or paired,

7- 9,5 cm long (including the peduncle), glabrous.

Flowers yellowish-white, on pedicels up to 0,75 mm

long, pedicels articulated towards the apex, with

minute yellowish glands at the base of the pedicels.

Calyx cupular, up to 1 ,5 mm long, glabrous, shortly

5-toothed. Petals shortly united basally, 3-3,75 mm

long, 1 mm wide, linear-oblong, inflexed apically,

glabrous. Stamens free among themselves, slightly

adnate to the corolla basally; filaments up to 5 mm

long; anthers up to 0,8 mm long, with a deciduous

apical gland. Ovary up to 2,25 mm long, glabrous.

Pods dark brown or reddish-brown, (15) 19-30

(40, 5) x 3-3, 9 cm, straight or slightly curved, oblong,

compressed, umbonate over the seeds, prominently

transversely venose, at maturity the valves separating

from the persistent margins, the outer layer of the

pod-wall peeling off the inner layer, the layers tending

to break up irregularly. Seeds immature, mature seeds

unknown.

Known only from the Grootfontein District in

South West Africa (See Fig. 1) Ecology unknown.

S.W.A. — 1917 (Tsumeb): farm Heidelberg near Tsumeb (-BB)

Dinter 1689 (SAM); Otavi (-CB), Dinter 745 (SAM), Dinter

5300 (PRE, Z). Grid ref. unknown: Grootfontein District,

farm Asis, Volk 767 (M).

The original specimens on which Dinter based his

description are no longer available for study. For-

tunately, however, both syntypes, namely Dinter

745 and 1689, are represented in the South African

Museum collections. There is one sheet of Dinter 745

and three sheets of Dinter 1689. One of the sheets of

Dinter 1689 is a mixed gathering consisting of a vege-

tative shoot of E. sujfruticosa and a pod of E. schin-

ziana, while the flowers in the capsule could belong to

either species. On the second sheet of Dinter 1689

there is a vegetative shoot of E. schinziana with a

mature pod attached and, in addition, there is a

flowering specimen which is leafless apart from an

extremely young leaf or shoot on which no details

are discernible. In view of the mixed gathering of

E. schinziana and E. sujfruticosa on the first sheet of

Dinter 1689, the possibility exists that the flowering

specimen on the second sheet belongs to E. sujfruti-

cosa and not to E. schinziana. It will be recalled that in

South West Africa E. sujfruticosa usually flowers when

leafless. As none of the other specimens of E. schin-

ziana examined were in flower, the details of the

flowers in the above description were taken from this

second specimen of Dinter 1689. It is possible therefore

that the flowers described are those of E. sujfruticosa

and not of E. schinziana. The third sheet of Dinter

1689 consists of a single leaf and two mature pods

(the valves of one pod are mounted separately which

gives the impression that there are three pods).

As the one sheet of Dinter 1689 is a mixed gathering,

and as there is a possibility that a second sheet is also a

mixed gathering, it seems desirable to select a lecto-

type for E. schinziana from the third sheet of Dinter

1689 or from Dinter 745. Dinter 745 is the better

specimen of the two and consequently I now select

Dinter 745 as the lectotype of E. schinziana.

There is a considerable variation in leaflet size on a

single branch; some of the upper leaves often have

distinctly smaller leaflets than the leaflets on the lower

leaves.

E. schinziana has not been re-collected since Jan.

1939 and there is a possibility that it is now extinct.

A thorough search for this plant in the type localities

is most desirable in an attempt to evaluate its present

conservation status. If E. schinziana is re-discovered,

it is important that an effort be made to collect both

flowering and fruiting material. Flowering material

is essential to establish whether or not the flowers

on which the above description was based are in fact

those of E. schinziana.

9. Elephantorrhiza suffruticosa Schinz in Mem.

Herb. Boiss. 1 : 117 (1900); Dinter, Deutsch-Siidwest-

Afrika FI. Forst und landwirtschaft. Frag. 78 (1909);

Harms in Engl., Pflanzenw. Afr. 3, 1: 400 (1915);

Dinter in Feddes Repert. 17: 190 (1921); Phillips

in Bothalia 1 : 193, t. 5 fig. 7 (1923); Bak. f., Leg. Trop.

Afr. 3: 801 (1930); Schreiber in Prodr. FI. S. W. Afr.

58: 17 (1967); Brenan & Brummitt in FI. Zamb. 3, 1 :

26 (1970). Syntypes: South West Africa, 1816 (Namu-

toni), Amutele (-AB), Schinz 2070 (Z!); 2215 (Trek-

kopje), Ebony Mine (-AB), Schenck 457 (Z!); 2317

(Rehoboth), Rehoboth (-AC), Fleck 497a (Z!), Fleck

499a (Z); grid ref. unknown, “sandige stellen bei

Ombalambuenge”, Rautanen 242 (Z!) ; Angola, Huila

district, “Kilevi am Kunene” (south of Humbe),

Schinz 2071 (Z!).

A branched shrub or small tree 1-5 m high;

bark grey-brown to dark or reddish-brown; young

branchlets glabrous, or sometimes puberulous to

shortly pubescent. Leaves glabrous to puberulous or

shortly pubescent: petiole (0,6) 1,5-3, 5 cm long;

rhachis (0,5) 10-17 (25,4) cm long; pinnae (2) 15-27

256

THE GENUS ELEPHANTORRHIZA

(42) opposite or subopposite pairs; rhachillae (1,4)

2-3,5 (6,8) cm long; leaflets (17) 27-40 (50) pairs per

pinna, 3-7,5 x 0,5-1, 2 mm, linear-oblong to linear,

rarely almost falcate, midrib marginal throughout,

proximal side rounded basally, apex asymmetric,

obtuse to acute, often nnicronate, lateral nerves and

veins not or scarcely visible, glabrous or sometimes

sparingly pubescent on the margins. Racemes axil-

lary, solitary or 2-3 together, or borne on lateral

shoots, (4) 6-14 (18) cm long (including the peduncle),

pubescent or sometimes glabrous. Flowers yellowish-

white, on pedicels up to 1 mm long, pedicels arti-

culated near the middle, with minute reddish, reddish-

brown or pale yellow glands at the base of the pedicels.

Calyx cupular, up to 1 mm long, shortly 5-toothed,

glabrous or sometimes very sparingly pubescent.

Petals shortly united basally, 3-3,75 mm long, 1 mm

wide, linear-oblong, indexed apically, glabrous.

Stamens free among themselves, slightly adnate to the

corolla basally; filaments up to 5 mm long; anthers up

to 0,8 mm long, with a deciduous apical gland.

Ovary up to 2 mm long, linear, glabrous. Pods dark

brown or reddish-brown, 8,5-30,5x 1,8-2,25 cm,

straight or slightly curved, linear-oblong to linear,

compressed, usually prominently transversely venose,

umbonate over the seeds, at maturity the valves

separating from the persistent margins, the outer

layer of the pod-wall peeling off the inner layer, the

layers remaining intact or breaking up irregularly.

Seeds 1 3-1 5 x 9-12 mm, roughly ellipsoid.

Found in Angola, South West Africa, Rhodesia and

Mozambique (See Fig. 1). Occurs in woodland,

grassland and in broken country; often among rocks.

The ecological preferences of E. suffruticosa are not

clear and more information is required.

As specimens from Rhodesia and Mozambique

were cited by Brenan & Brummitt in FI. Zamb. 3,1 :

26 (1970), it is not considered necessary to cite speci-

mens from these territories here.

I have not seen the specimen from Huila, Dekindt

536 (LISC), hesitantly referred to E. suffruticosa by

Torre in Consp. FI. Angol. 2: 263 (1956). The follow-

ing specimens from Angola have, however, been

examined :

Angola. — Huila District: 8 km para o Chitado, Azancot de

Menezes & Hewiques 38 (K); proximo do Chilau, Azancot de

Menezes 919 (K); na picada Chimbole — Gambos, Azancot de

Menezes 698 (K); 4 km de Cavalaua a Gambos, Mendes 1692

(BM).

S.W.A. — 1814 (Otjitundua): Otusemba, Story 5920. 1815

(Okahakana): 84,4 km north of Okakeujo on road to On-

dongua, De Winter 3617. 1816 (Namutoni): Amutele, Schinz

2070 (Z). 1817 (Tsintsabis): 48 km north of Tsumeb, Rodin 2605.

1916 (Gobaub): farm Zukov 337, De Winter 3019. 2016 (Otjiwa-

rongo): Otjiwarongo, Liebenberg 4909. 2116 (Okahandja):

Okahandja, Dinter 314 (BM, E, GRA, Z). 2117 (Otjosondu) :

Omupanda, Wtdfhorst s.n. (Z). 2214 (Swakopmund): Swakop-

mund, Seydel 114 > (Z). 2215 (Trekkopje): Ebony Mine, Schenck

457 (Z). 2216 (Otjimbingwe): Auas Mts., pass between Haris and

Aub, Pearson 9658 (K). 2217 (Windhoek): Windhoek, Rogers

29755. 2317 (Rehoboth): Rehoboth, Fleck 499a (Z). 2416

(Maltahohe): Bull’s Mouth Pass, Pearson 8919; Bullsport, Nau-

kluft Mts., Hardy 1982. 2417 (Mariental): Voigtsgrund, Keet

1662. Grid ref. unknown: “sandige stellen bei Ombalam-

buenge”, Rautanen 242 (Z).

The narrow leaflets of E. suffruticosa , with the

midrib marginal throughout the length of the leaflet,

are most diagnostic and enable this species to be

readily distinguished from all of the other Elephantorr-

hiza species.

There appears to be a discontinuity in the distri-

bution of E. suffruticosa between eastern and central

Rhodesia and South West Africa and Angola (See

Fig. 1). No significant morphological differences have

so far been noted between specimens from these two

areas of distribution except for an inconsistent ten-

dency for some leaflets to be more acute and mucro-

nate in Rhodesia than in South West Africa and

Angola. In addition, in specimens from Rhodesia

there are usually ± conspicuous reddish or orange

glands at the base of the leaflets, while in Angola and

South West Africa the glands are inconspicuous and

pale yellow or even absent.

In the Flora Zambesiaca area E. suffruticosa

usually produces flowers together with the leaves,

while E. goetzei usually flowers when leafless. The

leaves which accompany the flowers in E. suffruticosa

enable specimens of these two species to be readily

distinguished in the Flora Zambesiaca area. Other-

wise it is sometimes difficult to differentiate these two

species from flowers alone. The inflorescence axes in

E. suffruticosa are often puberulous whereas in E.

goetzei they are glabrous, and this character sometimes

assists in the identification of specimens.

In South West Africa, however, E. suffruticosa

frequently flowers when leafless, while in the Transvaal

E. goetzei usually produces flowers together with the

the leaves. In the area delimited for the Flora of

Southern Africa, therefore, the reverse situation

tends to prevail; E. suffruticosa flowering when

leafless and E. goetzei producing flowers together with

the leaves. Although E. goetzei has been recorded

from Angola, it has not been recorded from South

West Africa.

Schinz in Mem. Herb. Boiss. 1 : 117 (1900) cited six

specimens of E. suffruticosa and these specimens must

all be regarded as syntypes. The syntypes are: Fleck

497 a and 499 a, Schenck 457, Schinz 2070 and 2071,

and Rautanen 242. An examination of all the material

of E. suffruticosa in the University of Zurich Her-

barium revealed that there are two sheets of Fleck

497 a; one in flower and young pod collected in Octo-

ber 1891, and the other with mature pods and seeds

plus the fragments of a leaf collected in January 1891.

Schinz (l.c.) was uncertain whether all of the above

specimens belonged to the same taxon, and was there-

fore in some doubt about including them all under his

E. suffruticosa. In particular, he expressed doubt

about the specimen of Fleck 491a , with the robust ma-

ture pods, stating that they bore a resemblance to the

pods of E. elephantina. As none of the specimens was -

regarded as complete, Schinz drew up his type

description from all of the syntypes. In view of the

doubt in Schinz’s mind whether all of the specimens

belonged to the same taxon, it seems desirable to

select a lectotype for E. suffruticosa.

Owing to the difficulty of establishing the identity

of leafless flowering specimens with absolute certainty,

Schenck 457 and the flowering specimen of Fleck 491a

have been eliminated as potential lectotypes as neither

has any leaflets. Fleck 499 a and Rautanen 242 are

sterile and are not considered suitable for choice as

lectotype either. The two best specimens are undoubt-

edly the Schinz specimens; one from “Kilevi am

Kunene”, just south of Humbe in Angola, and the

other from Amutele in Amboland, South West Africa.

However, neither of these specimens bears a collector's

number. Both specimens were named E. suffruticosa

by Schinz and the localities of collection correspond

with the localities cited for the specimens of Schinz

2071 and 2070 respectively. There are no other

Elephantorrhiza specimens in Schinz's collection

from these localities, so it is considered safe to assume

that these two specimens are the specimens he had

before him and cited as Schinz 2071 and 2070.

J. H. ROSS

257

Schinz drew attention to the specimen from Kilevi

having immature pods and, as the pods on this speci-

men in question from Kilevi are immature, this

supports the contention that this was the specimen

Schinz was referring to.

The Schinz specimen from “Kilevi am Kunene"

(number 2071), consisting of three leaves and two pods

and having two inflorescences in the capsule mounted

on the sheet, is the better of the two Schinz specimens

of E. suffruticosa. Consequently, I now select Schinz

2071 from Kilevi am Kunene in Angola as the lecto-

type of E. suffruticosa.

Having arrived at this decision to regard Schinz 2071

as the lectotype of E. suffruticosa , I found it more than

a little disconcerting to find that Schinz in Mem.

Herb. Boiss. 1 : 105 (1900) cited Schinz 2071 from

Olukonda-Oshiheke, Amboland, South West Africa

as one of the syntypes of Acacia arenaria Schinz.

Fortunately I have examined this syntype of Acacia

arenaria and can vouch for its identity. In any event, it

seems unlikely that Schinz 2071 from Olukonda-

Oshiheke in South West Africa, the syntype of

Acacia arenaria, would ever be confused with Schinz

2071 from Kilevi am Kunene in Angola, the lectotype

of Elephantorrhiza suffruticosa. It is as well, however,

to draw attention to the existence of these two

Schinz specimens, each numbered 2071.

The flowering specimen, Hornby 2448 (K) from

Poole farm. Hartley Distr., Rhodesia, doubtfully

referred to E. suffruticosa in FI. Zamb. 3, 1 : 26 (1970),

is in fact E. goetzei subsp. goetzei (see the leaves in the

envelope mounted on the sheet).

INSUFFICIENTLY KNOWN SPECIES

10. Elephantorrhiza sp.

Suffrutex producing at ground level unbranched,

longitudinally striate, glabrous, stems 60-80 cm high.

Leaves glabrous or almost so: petiole 3-4,5 cm long;

rhachis 7,5-18 cm long; pinnae 3-8 opposite or

subopposite pairs; rhachillae 7,5-10 cm long; leaflets

13-22 pairs per pinna, 9-11 x3-5 mm, very oblique,

broadly truncate basally or sometimes slightly

auricled on the proximal side, asymmetric and at-

tached by one corner, the midrib starting in the distal

corner of the leaflet-base and gradually becoming

almost central in the leaflet, rounded to acute or

distinctly mucronate apically, glabrous throughout or

with few minute marginal cilia. Flowers and pods

unknown.

Known from one gathering from the eastern Cape.

Cape. — 3128 (Umtata): Umtata aerodrome (-DA), Strey

11073

Strey 11073 does not appear to match material of

any of the existing species. Like E. elephantina and

E. obliqua , the specimen has unbranched aerial stems.

Strey 11073 appears to differ from E. elephantina

in having larger leaflets, and differs from E. obliqua in

having larger leaves with more numerous pinnae and

leaflet pairs, and leaflets without a conspicuous

venation. More material, particularly fertile material,

is required in order to make a positive identification.

EXCLUDED SPECIES

Elephantorrhiza pubescens Phillips, in Bothalia

1: 190, t.5 fig. 3 (1923), based on Rogers 8659 (PRE

holo.!; K!; SRGH) from Zambia, is a synonym of

Entada abyssinica Steud. ex A. Rich.

CONCLUSIONS

The present state of knowledge of several species of

Elephantorrhiza in southern Africa leaves much to be

desired. Only the four most widespread species,

namely, E. elephantina, E. burkei, E. goetzei and

E. suffruticosa are fairly well collected and documented.

Flowering and fruiting material of E. praetermissa has

been collected, but more material is desired. The four

remaining species, however, are inadequately known

and there is concern about the present conservation

status of all of them.

A list of the most serious deficiencies in our know-

ledge has been drawn up in the hope that collectors

will make an effort to collect the material and in-

formation required to enable a more comprehensive

picture to be compiled.

1. Pods of E. obliqua have never been collected.

Both var. obliqua and var. glabra are known only

from the type collections and no authentic material

of either has been collected for over fifty years.

More material of E. obliqua, particularly fruiting

material, and information on the current conservation

status of the species are required.

2. More material of E. woodii var. woodii, parti-

cularly fruiting material, is required. The pods of

E. woodii var. pubescens are unknown. There is a

need for flowering and fruiting material of var.

pubescens and it is essential that the precise localities of

collection are recorded. Ecological notes and infor-

mation on the current conservation status of both

varieties of E. woodii are desirable.

3. The identity of the plants from northern Natal,

which have hesitantly been referred to E. woodii var.

pubescens needs elucidation. Additional and better

material is required.

4. E. schinziana has not been re-collected since

Jan. 1939. As it is not absolutely certain whether the

flowers of E. schinziana have been collected, flowering

material is most desirable. It would also be interesting

to know whether or not E. schinziana flowers when

leafless. More material, ecological notes and infor-

mation on the current conservation status of this

species are required.

5. E. rangei is known only from the type collection

(Jan. 1908). More material, ecological notes and

information on the current conservation status of the

species are required.

6. More material of Strey 11073 from the Umtata

aerodrome in the eastern Cape is required.

Bothalia 11, 3: 259-262 (1974)

Smodingium dermatitis: the intercellular secretory canals of the

aerial axis and their relationship to this toxicity

R. P. ELLIS *

ABSTRACT

Smodingium argutum E. Mey. ex Sond., closely related to the American poison-ivy. Toxicodendron

radicans (L.) Kuntze, is the only known indigenous member of the Anacardiaceae in South Africa

causing dermatitis. The sap of the intercellular secretory canals of the stem, petiole and leaf of both

plants, is responsible for this toxicity. These secretary canals are described, illustrated and compared

for both species, and their formation and the function of the canals and sap, are briefly discussed.

INTRODUCTION

Smodingium argutum E. Mey. ex Sond. is the only

known indigenous South African member of the

Anacardiaceae that causes dermatitis. Numerous

other examples of potentially poisonous members of

the family are found throughout the world and include

the American poison-ivy. Toxicodendron radicans

(L.) Kuntze, the poison oak, Rhus diversi/oba , certain

sumacs, the mango, cashew-nut, Japanese lacquer and

Indian marking-nut trees (Whiting 1971).

Smodingium argutum, which is confined to South

Africa, varies in appearance from a small woody

shrub to a small tree or even a creeper. The leaves are

trifoliate with dentate margins. The natural distri-

bution of the plant appears to be concentrated in

Pondoland and East Griqualand, extending north-

wards through Natal to Barberton (Findlay 1963).

It is now relatively common in parks and gardens in

the Transvaal, being readily obtainable from nurseries.

Findlay (1963) suggested that the dermatitis pro-

duced by S. argutum was of the poison-ivy type —

the plants belong to the same family, both have self-

melanising sap, the skin reactions are similar to those

of poison-ivy and natural and experimental cross

reactions take place between sufferers sensitive to the

two plants (Whiting 1971). This paper shows further

the close relationship between these two poisonous

plants from different continents.

Fresh, living material was obtained from plants

growing in the National Botanical Gardens, Brum-

meria, Pretoria, originating from seed collected by

Dr L. E. Codd near Port St Johns, Transkei. Poison

ivy ( Toxicodendron radicans) was also collected from

these gardens for comparative studies. Leaf, petiole

and young stem tissue of both these plants was fixed

in gluteraldehyde or FAA, embedded in tissuemat,

sectioned at 5/j. and stained in safranin and fast green.

ANATOMY

Intercellular secretory cancds in the Anacardiaceae

The outstanding anatomical feature of the axis of

all members of the Anacardiaceae is the universal

presence of secretory canals or ducts (Metcalfe and

Chalk 1950). These are present in the roots, stems,

leaves, fruit and embryo in the primary phloem

(McNair 1918). In addition, in certain representatives

of the family, they may be found in the secondary

phloem of older roots and stems, while others possess

medullary or pith secretory canals.

In accordance with the classification of inter-

cellular spaces (Stern, 1954) these canals are termed

intercellular secretory canals to avoid any possible

confusion with the term resin canal. This term has

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X10I, Pretoria.

often been used to describe the ducts of the Anacar-

diaceae, but resin canal is now limited to the canals

of the gymnosperms. Gum duct is another alternative

term that has been used in place of secretory canals.

Amongst the dicotyledons secretory canals are

characteristic of the Compositae and Anacardiaceae

only (Fahn 1959). In the Anacardiaceae, various ana-

tomical works (Metcalfe and Chalk 1950) show that

there are no essential differences in the arrangement

and structure of the intercellular secretory canals,

and the following description of the canals of 5.

argutum serves further to confirm this.

Intercellular secretory canals 0/ Smodingium argutum

Stem anatomy

The phloem of the primary vascular bundles of

Smodingium is separated from the parenchymatous

outer cortex by a pericycie of isolated, arc-shaped

strands of sclerenchymatous fibres, having the convex

side towards the exterior (Fig. 1, 2). This fibrous

pericycie is absent in Toxicodendron radicans (Fig.

3, 4). In older stems of Smodingium there is a tendency

towards a composite and continuous ring of scleren-

chyma with the individual arc-shaped strands not

clearly distingiushable. A single, large secretory

canal is situated in the primary phloem of each

vascular bundle on the inside of each sclerenchymatous

arc. Outside the pericycie no canals are found.

Secretory canals are present in concentric circles in

the secondary phloem in most members of the Ana-

cardiaceae (Metcalfe and Chalk 1950), and in Rhus

diversi/oba (McNair 1918) these canals anastomose in

the internodes forming a complete cylindrical network

in the bark. Canals in older bark become displaced and

distorted due to the dilatation of the bark when the

epithelial cells divide in a plane perpendicular to the

duct (Esau 1969). No evidence of canals in the second-

ary phloem was found in the material studied,

but it is not unlikely that this will be the case in older

material. Pyrocatechol tannin in the parenchyma and

collenchyma of the cortex, as in R. ovata (Watkins

1940), was not observed.

A single ring of smaller medullary secretory canals

is present interior to the primary xylem in the pith

of young stems of Smodingium argutum (Fig. 1).

Watkins (1940) noted the presence of small secretory

canals in the pith of Rhus laurnina and R. ovata ,

but McNair (1921) comments on their absence in

R. diversi/oba and R. radicans (Toxicodendron radicans )

(Fig. 3). In other members of the Anacardiaceae

these medullary canals vary in number in different

members of a single genus or even in different inter-

nodes of a single plant.

260

SMODINGIUM DERMATOL, l HE INTERCELLULAR SECRETORY CANALS

Figs. 1-4. — Transverse section of young stem. All bright field. Figs. 1-2, Smodingium argutum.

Figs. 3-4, Toxicodendron radicans. Figs. 1 and 3, x 6,25; Figs. 2 and 4, x 40. SC — secretory canals;

MC — medullary secretory canals; P — pericycle fibres.

Figs. 5-8. — Transverse section of petiole. Figs 5-6, Smodingium argutum. Figs. 7-8, Toxicodendron

radicans. Figs. 5 and 7, 6,25, bright field; Figs. 6 and 8, x 40, interference contrast. SC— secretory

canals.

R. P. ELLIS

261

Petiole anatomy

Metcalfe and Chalk (1950) state that secretory

canals in the phloem of the petiolar vascular system

are presumably present throughout the Anacardiaceae.

In Smodingium there are six secretory canals arranged

in a ring following the outline of the petiole as seen in

transverse section. The largest canal is adaxially

situated. The canals are exterior to the heart-shaped,

cylindrical strand of xylem tissue (Fig. 5). Arc-shaped

pericycle caps of thickened fibres are present exterior

to the secretory canals which are, therefore, located

in the phloem (Fig. 6). The bundles and canals

branch upon reaching the leaflets. In Toxicodendron

radicans there are more numerous canals in the petiole

and the pericycle caps are lacking, as in the stem

(Fig. 7, 8).

Leaf anatomy

The midrib of Smodingium argutum includes a

ventral band of three vascular bundles with secretory

canals surrounded by pericycle arcs (Fig. 9). The canal

arrangement is similar to that of Rhus diversiloba

(McNair 1918) and Toxicodendron radicans (Fig. 1 1).

Dorsally there is a single bundle in Smodingium in

contrast to Rhus diversiloba , which has five to seven

bundles arranged in an arc. Scattered throughout the

midrib are small irregular secretory canals.

All the lateral veins of the leaves of both

Smodingium argutum and poison-ivy contain at least

one secretory canal on their dorsal sides in the phloem

(Fig. 10, 12). Some canals end blindly in the mesophyll

and others anastomose in a reticulate manner as do

the bundles which they accompany.

Formation of the secretory canals

Secretory spaces, in the form of cavities or canals,

are formed by schizogeny or lysigeny or by both

phenomena (Esau 1965). Schizogenous canals are

lined by secretory cells comprising the epithelium,

and lysigenous canals are surrounded by more or less

disintegrated cells, the breakdown of which brought

about the formation of the canal.

According to Sieck (1895) the canals of the Ana-

cardiaceae are of schizolysigenous origin, but Fahn

(1969) states that they are schizogenous and develop

between resin-producing parenchyma cells which

form the duct epithelium. McNair (1918) found the

initial canal development of Rhus diversiloba to be

clearly schizogenous. In Smodingium argutum , the

medullary secretory canals are clearly schizogenous

but, especially in older stems, petioles and leaves,

the epithelial lining of the canals is often disjointed

and indistinct (Fig. 2, 6).

Chemical composition of the sap of the secretory

canals

The abundant canals of the Anacardiaceae contain a

milky sap which turns black on exposure to air. The

oleoresin fraction of the sap is often termed urushiol

and its antigenetic properties, which are retained long

after the plant is dead, are related to the presence of

pentadecylcatechols (Whiting 1971). The American

poison-ivy oleoresin contains a 1 ,2-dihydroxy benzene

(catechol) with a 15 atom side chain in the third

position (Fisher 1967). Proof of the close chemical

relationship between the allergens of poison-ivy and

Smodingium has been obtained (Findlay et al 1973,

Eggers 1973). These findings suggest that both species

manufacture a C15, as well as a C17, side-chain type of

catechol, in which the types and degrees of unsatu-

ration varies. Thus, S. argutum sap showed large

amounts of the C17 diolefin, and poison-ivy shows

a prominent diolefin at C15, as well as traces of C17

side-chain compounds.

Small amounts of fresh sap from these plants,

collected from the secretory canals with capillary

pipettes, for mass spectrometry, showed similar

results. Thus, it is in these secretory canals that the

poisonous principles are harboured.

As mentioned, many workers on the Anacar-

diaceae (Metcalfe and Chalk 1950) have shown that

there are no essential differences in the arrangement

and structure of the intercellular secretory canals in

the family. Why some representatives of the family.

Pigs. 9 and 11. — Transverse section of midrib of leaf. Figs. 10 and 12. Section of leaf lamina. Figs.

9-10, Smodingium argutum. Figs. 11-12, Toxicodendron radicans. Figs. 9 and 1 1, 6,25, bright

field; Figs. 10 and 12, x 40, interference contrast. C — secretory canals.

262

SAIO DING I U M DERMATITIS: THE INTERCELLULAR SECRETORY CANALS

such as Smodingium, and some members of the genus

Rhus , should be poisonous, or why their poisons

should vary, either in physiological action, or chemical

composition, cannot be elucidated from their anatomy

and remains a mystery.

Function of the secretory canals

Various suggestions have been proposed as to the

function of the secretory canals and the resin contained

in them. Most of the theories hinge on the characteris-

tic of this sap to harden when in contact with the air.

When first expressed from the canals the resinous sap

is transparent or light grey in colour, but it rapidly

coagulates and hardens into a brown or black varnish-

like mass. Microscopically the fresh sap can be seen

to be composed of a colourless liquid in part, and in

part of minute globules. Soon these globules become

dark brown and at the same time oblong, rectangular,

colourless crystals separate out.

McNair (1918) suggests that this hardening forms an

efficient covering for any wounds the plant may receive.

The abundance and relatively large size of these canals,

together with their anastomosing, make up an

extensive intercommunicating system. Wherever the

plant is wounded a protective scab is rapidly formed.

The viscosity of the secretory canal contents varies

with the seasons. In spring, when the young leaves and

twigs are fragile, it is very watery, but later in the

summer it is noticeably thicker and slower in

exudation.

In autumn, just before the fall of the leaves of

Toxicodendron radicans , Trecul (1867) noticed the

obstruction of the secretory canals at the base of the

petiole. This obstruction was effected by an enl rge-

ment or increase in the parietal cells of the canals.

Trecul (1867) suggests that the development of these

tvlosoids mav have a part to play in the initiation of

leaf fall.

Toxicity and the secretory canals

It has been conclusively shown that the freshly

exuded, non-volatile, resinous sap is the only part of

the plant capable of causing dermatitis in Rhus

diversiloba (McNair 1916 a) and Smodingium argutum

(Findlay et a / 1973). Poisoning can thus occur

only as a result of direct contact with the sap of the

plant or by contact with this sap on clothing or tools.

There is a contradiction in the literature as to whether

or not the virulency of the sap from the secretory

canals is lost when it hardens into a varnish-like

substance. McNair (1917) is of the opinion that the

irritating properties of the fresh sap are lost, but

Whiting (1971) states that it retains its sensitising

potential long after it has adherd to and hardened on

objects.

McNair (1921) has shown that the sap retains its

toxicity, without much variation in degree of virulency,

throughout the year. He correctly points out that the

tendency of Rhus diversiloba to cause poisoning

varies during the course of the year in accordance

with the stage of growth of the leaves, stems and

flowers. The virulency of the plant is influenced by the

virulency of the sap, the turgescence and ease of

fracture of the leaves and stems, the conditions of

light and humidity and thus turgor and photosyn-

thesis, as well as by the conspicuousness of the plant.

McNair (1921) supports this with clinical statistics.

These influences on virulency are equally applicable

to Smodingium argutum in South Africa.

This study was conducted under the guidance of

Prof. G. H. Findlay of the University of Pretoria,

with the assistance of Mrs H. Vismer and co-operation

of Dr S. H. Eggers of the C.S.I.R. The author is

gratefull to Miss L. Breytenbach for technical assis-

tance and Mr S. R. van Jaarsveld for photographic

assistance.

REFERENCES

Eggers, S. H., 1973. In press.

Esau, K., 1965. Plant anatomy. New York: John Wiley &Sons.

Esau, K., 1969. Encyclopedia of Plant anatomy 5 (2). Berlin:

Gebruder Borntrager.

Fahn, A., 1969. Plant anatomy. London: Pergamon Press.

Findlay, G. H., 1963. Dermatitis of ‘Poison-Ivy’ type from an

indigenous South African plant — Smodingium argutum E.

Mey. S. Afr. Med. J. 37: 883-888.

FrNDLAY, G. H., Whiting, D. A., Eggers, S. H. and Ellis,

R. P., 1973. In press.

Fisher, A. A., 1967. Contact dermatitis. Philadelphia: Lee and

Fibiger.

McNair, J. B., 1916. The pathology of dermatitis venenata

from Rhus diversiloba. J. Infect. Dis. 19: 419-428.

McNair, J. B., 1916 a. The transmission of Rhus poison from

plant to person. J. Infect. Dis. 19: 429-432.

McNair, J. B., 1917. The oxidase of Rhus diversiloba. J. Infect.

Dis. 20: 485^198.

McNair, J. B., 1918. Secretary canals of Rhus diversiloba.

Bot.Gaz. 65: 268-273.

McNair, J. B., 1921. A study of Rhus diversiloba with special

reference to its toxicity. Am. J. Bot. 8 (3): 127-146.

McNair, J. B., 1921 a. The morphology and anatomy of

Rhus diversiloba. Am. J. Bot. 8 (4): 179-191.

Metcalfe, C. R. and Chalk, L., 1950. Anatomy of the dico-

tyledons. Oxford : Clarendon Press.

Sieck, W., 1895. Die schizolysigenen Secretbehalter. Jahrb.

Wiss. Bot. 27: 227.

Stern, W., 1954. A suggested classification of intercellular

spaces. Bull. Torrey Bot. Club 81 : 234-235.

Trecul, M. A., 1867. Des Vaisseaux propres dans les tere-

binthinees. Compt. Rend. 65: 17.

Watkins, K. S., 1940. Stem anatomy of chaparral shrubs.

Bot.Gaz. 101: 391-402.

Whiting, D. A., 1971. Plant dermatitis in the Southern Trans-

vaal. 5. Afr. Med. J. 45: 1 63-167.

Bothalia 11,3: 263-264 (1974)

The identity of some species of Hermannia represented in the

Linnaean, Thunberg, Bergius and Cavanilles Herbaria

B. DE WINTER

A study of the types of Hermannia species in the

above-mentioned herbaria has brought to light that

several names which have fallen into disuse have to be

resuscitated. Most of these names were either dis-

regarded or were not investigated by Harvey for his

treatment of Hermannia and Mahernia in the Flora

Capensis Vol. 1. Some of them are mentioned in his

lists of “Doubtful or little-known species”. This

ommission was unfortunate because the present

investigation has shown that several name changes are

unavoidable. The following species are discussed.

Hermannia confusa Salter in J.S.Afr. Bot. 12:

99 (1946). Type: Cape, between Caledon and

Babylon's Tower, Ecklon & Zeyher Enum. No. 374

(BOL, holo.).

In 1946 Salter described a new species, Hermannia

confusa Salter, to accommodate those specimens cited

by Harvey in the Flora Capensis under Hermannia

tenuifolia Sims. His reason for doing so was that the

identity of H. tenuifolia could not be determined with

certainty, mainly because of the poorness of the plate

in the Bot. Mag. on which it was based. With this

opinion I concur since no specimen is available to

confirm the identity of the plant figured. (The plate

shows an erect plant with an inflorescence different

from those of the specimens cited by Harvey.)

I had the opportunity of studying the specimen of

which Sims said: “there is a defective specimen in the

Banksian Herbarium under the name decomposita,

which appears like it”. The latter specimen, which

Salter did not see, closely matches the types of H.

myrrhifolia Thunb. and of H. pinnatisecta Salter, but

does not agree with the plate in the Botanical Maga-

zine, which plate must be regarded as the type of H.

tenuifolia Sims, no specimen having been preserved as

far as I could ascertain.

Hermannia diffusa L.f, Suppl. 302 (1781). Type:

Cape, Verloren Valley, Thunberg s.n. (LINN Cat.

No. 854. 16, holo.; UPS, Thunb. Cat. No. 15471, iso).

Mahernia diffusa (L.f.) Jacq., Hort. Schoenbr. 2, t. 201 (1797)’

M. biserrata (L.f.) Cav., Diss. 6: 326 (1788). M. pilosula Harv.

(& vars) in Harv. & Sond., FI. Cap. 1 : 212 (1860). Types: Cape,

Swartland, Ecklon & Zeyher, near Paarl, Drege ; Groenekloof,

Drege-, Pappe s.n. (K, S, syns.).

Hermannia biserrata L.f. Suppl. 302 (1781). Type: Cape,

Thunberg s.n., Cat. No. 15466 (UPS, lecto.). H. pilosula (Harv.)

Hochr. in Annu. Conserv. Jard. Bot. Geneve 1 1-12: 3 (1907).

Hermannia diffusa L.f. is not mentioned in the Flora

Capensis and as a result the name went into disuse for

many years. Harvey, apparently unaware of the earlier

name, placed all the material referable to H. diffusa in

Mahernia pilosula Harv. and the latter name has to be

reduced to synonymy. H. diffusa L. is mentioned in

Adamson and Salter’s Flora of the Cape Peninsula,

and is contrasted with H. ciliaris. It seems doubtful

that his description applies to the true H. diffusa.

H. diffusa differs from H. pinnata L. (H . ciliaris L.f.)

in that the leaves vary from ovate to narrowly ovate,

with lacerate to deeply bipinnately divided margins,

and are never linear or whorled as in the latter species.

In floral characteristics H. diffusa is remarkably

constant, but the leaves are extremely variable as is

described above. This characteristic has caused

frequent misidentification and a certain amount of

proliferation of names. Hermannia biserrata L.f.

represents one extreme of the leaf variation having the

lamina of the leaf only shallowly lacerate on the

margin. There is no specimen of the latter species in

the Linnaean Herbarium. It is based on a Thunberg

gathering and the specimen named H. biserrata (Cat.

No. 15466) in the Thunberg Herbarium, which agrees

with the description as well as with the Cavanilles

plate of Mahernia biserrata (L.f.) Cav., should there-

fore be regarded as the lectotype of this name.

Mahernia bipinnata L. var. acutifolia Harv. and M.

bipinnata L. var. g/andulosa Harv. are both referable to

H. diffusa. The status of M. pilosula Harv. var.

latifolia is somewhat dubious but probably also falls

within the variation of the species.

Hermannia glabrata L.f. Suppl. 301 (1781). Type:

Cape, Roggeveld, Thunberg (LINN Cat. No. 854.11,

lecto.; UPS, S, iso-lecto.).

H. linearis (Harv.) Hochr. in Annu. Conserv. Jard. Bot.

Geneve 11-12:3 (1908).

Mahernia glabrata (L.f.) Cav., Diss. 6: 326 (1788). M. linearis

Harv. in FI. Cap. 1 : 217 (1860). Type: Cape, Somerset, Swartkey

on mountain plains, 4 000 ft., Drege (K, holo.; TCD, iso.).

Hermannia glabrata L.f. is mentioned in the Flora

Capensis under the little-known and doubtful species

of Mahernia by Sonder as Mahernia glabrata Cav.

The only material referable to H. glabrata cited in the

Flora Capensis (a single specimen collected by Drege)

was placed by Harvey in a new species which he

described as M . linearis. It is therefore not surprising

that the name Hermannia glabrata (Mahernia glabrata)

fell into disuse. Hermannia glabrata is represented in

the Linnaean Herbarium by two specimens. Both are

annotated by Linnaeus the younger, but the annota-

tion on the second (LINN Cat. No. 1 854 . 1 2) is follow-

ed by a question mark and is a poor specimen. Speci-

men 1854.1 1 was therefore chosen as lectotype of the

species.

Hermannia grossularifolia L., Sp. PI. 673 (1753).

Type: Cape (LINN Cat. No. 854. 14)

H. vesicaria Cav., Diss. 483 (1788). Type: Cape, fruiting

specimen (P, holo.).

Mahernia ovata E. Mey. ex Turcz. in Bull. Soc. Nat.

Moscow 31: 221 (1858). Types: Cape, s.n. (holo.,?;

K, LE, iso.).

This species is represented in the Linnaean

Herbarium by a single specimen, Cat. No. 854.14. This

sheet is written up in Linnaeus’s own hand, the specific

epithet ‘pinnata’ being crossed out and grossularifolia

written above it. The specimen is in full agreement with

the references to Commelin and van Rooyen in the

Species Plantarum. There is no specimen in the Hortus

Cliffbrtianus at the British Museum and, since LINN

Cat. No. 854.14 is the only specimen we have that was

seen by Linnaeus, it must be regarded as the holotype.

The description in the Species Plantarum would in any

case exclude it from actually being confused with H.

pinnata which has linear leaves. It agrees with H.

vesicaria Cav. and matches rather closely the specimen

collected by Drege and deposited in Kew. The latter

is an isotype of Mahernia ovata E. Mey. ex Turcz.

264

THE IDENTITY OF SOME SPECIES OF HERMANN l A

Hermannia pinnata L. Sp. PI. 1: 674 (1753).

Type: Specimen in Hortus Cliffortianus (BM, holo).

H. verticillata Berg., Descr. PI. Cap. 169 (1767), nom. illeg.

Type: Cape, Legit? (SBT). H. ciliaris L.f. Suppl. 302 (1781).

Types: Cape, Groenekloof, Thunberg s.n., Cat. No. 15468 &

15469. (UPS, Cat. 15468, lecto.). H. verticillata (L.) K. Schum.

in Pflanzenfam. 3, 6: 80 (1890); Hochr. in Annu. Conserv.

Jard. Bot. Geneve 11-12: 13 (1907).

Mahernia verticillata L., Mant. 1: 59 (1767). Type: Cape

(LINN Cat. No. 854.17, holo.).

Linnaeus described H. pinnata in his Species Plan-

tarum of 1753 and refers to a plate in Plukenet,

Mantissa 50 t. 344 fig. 3 (1692). This figure is recog-

nizable as representing the species presently known as

H. verticillata or H. ciliaris in most herbaria. There

is no specimen bearing the name H. pinnata in the

Linnaean Herbarium. Linnaeus, however, refers to

the Hortus Cliffortianus in the original description

and Mr Oliver, formerly liaison officer for this

Institute at Kew, kindly checked the Hortus Cliffor-

tianus and found a specimen named H. pinnata which

proved to agree with the Plukenet drawing and with

H. verticillata of modern herbaria. The identity of the

plant which should bear the name H. pinnata, is

therefore not in doubt but the name has hardly ever

been used in botanical works. In 1767 Bergius, in his

Descr. PI. Cap., described a species which he named

H. verticillata. This name is, however, illegitimate since

Bergius cites H. pinnata L., an older name, in

synonymy. A month after Bergius’s Descr. PI.

Cap. appeared, Linnaeus published his Mantissa

Plantarum in which is described Mahernia verticillata

L. Linnaeus supplied no reference to his earlier H.

pinnata which, as already stated, is the same species as

Mahernia verticillata L. Sheet 854.17 in the Linnaean

Herbarium bears the name H. “ verticillata ” in the

younger Linnaeus’s hand as well as an inscription

“pinnata?”, indicating that the latter considered the

possibility that H. pinnata could be the same as

M. verticillata L. The combinations H. verticillata

(L.) K. Schum., made in 1890 in the Pflanzenfamilien

based on Mahernia verticillata L., and H. verticillata

(L.) Hochr., based on the same type, are both invali-

dated by the earlier illegitimate homonym H.

verticillata Berg. (1767). Hermannia ciliaris L.f.,

(1781) based on a Thunberg specimen, differs in a few

unimportant characteristics from the type of H.

pinnata, but falls within the variation of this species

and must therefore be placed in synonymy.

Adamson in J.S.Afr. Bot. 5,2: 54 (1939) apparently

was not aware that H. pinnata L. and Mahernia

verticillata L. were conspecific and his suggestion that

H. ciliaris L.f., Suppl. (1781) is the oldest available

name, is therefore incorrect.

Hermannia procumbens Cav. Diss. 329 (1788).

Type: Cape, Groenekloof infra Leeuestaart, Thunberg

(MD, holo., PRE, photo; UPS, iso.).

subsp. myrrhifolia ( Thunb .) De Wint., comb. nov.

Hermannia myrrhifolia Thunb., Diss. Herm. 16 (1794). Type:

Cape, Swartland, Thunberg s.n.. Cat. No. 15487 (UPS, lecto.; S,

iso-lecto.). H. pinnatisecta Salter in J.S.Afr. Bot. 12: 102 (1946).

Type: Cape, Malmesbury, Mamre Hills, Compton 14928 (NBG,

holo). H. pinnatisecta var. auriculata Salter, l.c. Type: Cape,

Malmesbury, Ysterfontein, Compton 17374 (NBG, holo.; BOL,

iso.).

The type of Hermannia pinnatisecta Salter is a good

match of the type of H. myrrhifolia Thunb. I have not

found it justified to uphold the var. auriculata of

H. pinnatisecta.

The sinking of H. myrrhifolia under H. procumbens

is in line with Pillans’s manuscript notes on these

species which I have had available for study. The

subspecies differs mainly in the dissected leaves and the

less membranous calyx which usually has acute, not

subacute lobes.

Hermannia pulchella L.f., Suppl. 302 (1781).

Type: Cape, Roggeveld, Thunberg s.n. (LINN, Cat.

No. 854.15, holo.).

Mahernia pulchella (L.f.) Cav., Diss. 6: 325, t. 177, fig. 3

(1788). M. vernicata Burch., Trav. 1: 278 (1824). Type: Cape

Province, Fraserburg, Dovaal Fonteyn, Burchell 1461. (K, holo.;

PRE, iso.).

Hermannia pulchella was described by the younger

Linnaeus in his Supplementa in 1781. The Savage

Catalogue of the Linnaean Herbarium lists the sheet

No. 854.15 as collected by “?Thunberg”. Unfortu-

nately the sheet contains two specimens of which the up-

per agrees with the specimen named H. pulchella in the

Thunberg Herbarium (No. 15491) as well as with the

Thunberg specimen of this species in the Bergius

Herbarium. There is, therefore, little doubt that the

upper specimen on the sheet in the Linnaean

Herbarium should be regarded as the holotype of the

species. The lower specimen closely matches the type

specimen of Thunberg’s Hermannia cernua. It seems

likely that both specimens were collected by Thunberg

and that the lower specimen on sheet No. 854. 1 5 in the

Linnaean Herbarium is an isotype of Hermannia

cernua Thunb. Burchell described a new species

Mahernia vernicata in 1824 and this name was added

later by Harvey in the Flora Capensis 2: 589 (1862) as

an additional species which he regarded as probably

the same as Mahernia pulchella (L.f.) Cav. Mahernia

vernicata is here regarded as synonymous with M.

pulchella. The name applied by Linnaeus the younger

to this species eventually fell into disuse in spite of

being much older. The material of the species can be

found under the name Hermannia vernicata in most

herbaria.

Bothalia 11,3: 265- 268 (1974)

A note on Dichrostachys cinerea in South Africa

J. H. ROSS*

ABSTRACT

The nature of the morphological variation within Dichrostachys cinerea (L.) Wight & Arn. in

the area delimited for the Flora of Southern Africa is discussed. An attempt is made to evaluate the

taxonomic significance of some of the differential characters employed by Brenan & Brummitt in

Bol. Soc. Brot., Ser. 2, 39: 61-1 15 (1965) in order to delimit the infraspecific taxa recorded from our

area. Certain modifications to Brenan & Brummitt’s treatment are proposed.

INTRODUCTION

Dichrostachys cinerea (L.) Wight & Arn. is an extra-

ordinarily variable and taxonomically complex species,

widespread in Africa, Asia and reaching Australia.

An analysis of the variation within the species, which

resulted in the recognition of a number of infraspeci-

fic taxa, was the subject of a very detailed paper by

Brenan & Brummitt in Bol. Soc. Brot., Ser. 2, 39:

61-115 (1965). Of the numerous infraspecific taxa

recognized by Brenan & Brummitt, the following

eight are recorded from the area delimited for the

Flora of Southern Africa:

Subsp. nyassana (Taub.) Brenan.

Subsp. africana Brenan & Brummitt var. africana.

Subsp. africana Brenan & Brummitt var. lugardiae

(N.E. Br.) Brenan & Brummitt.

Subsp. africana Brenan & Brummitt var. setulosa

(Welw. ex Oliv.) Brenan & Brummitt.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria.

Subsp. africana Brenan & Brummitt var. pubescens

Brenan & Brummitt.

Subsp. africana Brenan & Brummitt var. plurijuga

Brenan & Brummitt.

Subsp. argillicola Brenan & Brummitt var. hirtipes

Brenan & Brummitt.

Subsp. forbesii (Benth.) Brenan & Brummitt.

Some idea of the distribution of these taxa in our

area may be gained from Fig. I .

Decisions have had to be reached for the account

of D. cinerea being prepared for the Flora of

Southern Africa and, unfortunately, 1 do not find

Brenan & Brummitt’s treatment of the species in our

area altogether acceptable. It seems necessary there-

fore to give reasons.

Fig. 1. — The known distribution of the infraspecific variants of Dichrostachys cinerea in South Africa.

266

A NOTE ON DICHROSTACHYS CINEREA IN SOUTH AFRICA

DISCUSSION

From a study of the abundant material of D.

cinerea in our area it is at once apparent that there is

rt continuous variation throughout, and that the

extremes have been delimited and given formal

taxonomic status. When extreme, each taxon is

usually fairly easily recognizable but, as mentioned

by Brenan & Brummitt, there are intermediates

between most of the taxa. In view of the numerous

intermediates, it is perhaps debatable whether the

creation of an elaborate system of infraspecific taxa

was entirely desirable. The purpose of this note, how-

ever, is to discuss the difficulties encountered in de-

limiting some of the taxa in our area, to attempt to

evaluate the taxonomic significance of some of the

differential characters employed by Brenan & Brum-

mitt, and to find an acceptable taxonomic treatment

in our area.

In its typical form in tropical Africa, subsp. nyas-

sana is distinct and easily recognized by its broad

leaflets, large leaves and usually fascicled peduncles.

However, in our area leaflet width, leaf size and the

arrangement of the peduncles provide no discontinuity

between subsp. nyassana and subsp. africana var.

africana, and some specimens from the eastern Trans-

vaal, Swaziland and Natal, are extremely difficult to

place with certainty. Indeed, it is sometimes a matter

of opinion whether they should be assigned to subsp.

nyassana or to subsp. africana var. africana. What

is particularly disconcerting is that sometimes a speci-

men is assigned to subsp. nyassana because some of

the leaflets exceed 2 mm in width, while a duplicate

of the same collection may be entirely devoid of

leaflets 2 mm or more wide and keys out to subsp.

africana var. africana. The problem of differentiating

depauperate specimens of subsp. nyassana and robust

specimens of subsp. africana var. africana in our area

is a very real and difficult one. All available evidence

indicates that there is no clear morphological discon-

tinuity between the two taxa and that the one merely

grades into the other. The distributional range of

subsp. nyassana in our area is very similar to that of

subsp. africana var. africana.

Subsp. africana var. africana* , on the other hand,

also grades into subsp. africana var. lugardiae and

into subsp. argillicola var. hirtipes in our area. In

FI. Zamb. 3, 1 : 38 (1970), subsp. argillicola var.

hirtipes was distinguished from subsp. africana by

having fewer pinnae pairs and narrower pods. In

our area, however, specimens with few pinnae pairs

which key out to subsp. argillicola var. hirtipes often

have pods up to 1,1 cm wide, so that pod width fails

to provide a discontinuity between the two taxa.

Although on some specimens of subsp. argillicola var.

* Since Brenan & Brummitt’s paper on Dichrostachys was

published in 1965 important changes affecting autonyms (auto-

matically established names) were introduced into Article 26 of

the latest edition of the International Code of Botanical Nomen-

clature (1972). One of these changes is the rejection in certain

circumstances of the previous ruling that autonyms must

always be adopted for taxa which include the type of the

correct name of the next higher taxon. In some instances this

results in a name which was correct when published now being

made retrospectively incorrect, and thus enforcing the adop-

tion of another, often undesirable, name. Dichrostachys cinerea

subsp. africana var. africana is such an example. Included in

this variety was Cailliea dichrostachys Guill. & Perr. var.

leptostachys (DC.) Guill. & Perr., so that the correct name for var.

africana, which was itself correct under the Code when published,

is now var. leptostachys under the new Code. This requires a

new combination for var. leptostachys. However, as an attempt

is to be made at the Leningrad Congress in 1975 to have the

recent changes in the Code affecting antonyms reversed, it is

considered undersirable to effect the new combination until the

outcome of this attempt is known.

hirtipes the pods are loosely coiled, on others the pods

are strongly coiled and no distinction can be drawn

between them and specimens of subsp. africana var.

africana on the grounds of the degree of coiling of

the pods. The number of pinnae pairs likewise pro-

vides no discontinuity between the two taxa. On

occasions it is found that a specimen keys out to subsp.

argillicola var. hirtipes, while another, with an almost

identical facies, keys out to subsp. africana var.

africana because the latter possesses an extra pair

of pinnae. The distributional ranges of subsp. africana

var. africana and subsp. argillicola var. hirtipes in

our area are very similar, and there is little evidence

that subsp. argillicola var. hirtipes is characteristic

of the heavy clay soils of valleys. Brenan & Brummitt

l.c. : 84 (1965) recorded only one specimen of subsp.

africana var. africana from Natal, but from the more

abundant South African material available to me, it

is clear that this taxon is fairly widespread in Natal.

The more abundant South African material availa-

ble also reveals that there is continuous variation in

leaflet width and other morphological characters

between subsp. africana var. africana and var. lugar-

diae. In var. africana there tends to be a stipitate

gland on the rhachis at the junction of all the pinnae

pairs, while in var. lugardiae the glands are usually

only at the junction of the lowermost 1-2 pinnae

pairs and the uppermost 1-5 pairs, but the number

and position of the glands represent an inconsistent

tendency. Depauperate specimens of subsp. africana

var. africana tend to be confused with robust speci-

mens of var. lugardiae on the one hand, and with

robust specimens of subsp. argillicola var. hirtipes on

the other. In our area, at least, subsp. africana var.

africana occupies an intermediate position between

subsp. nyassana, subsp. africana var. lugardiae and

subsp. argillicola var. hirtipes.

Subsp. africana var. setulosa is characterized by

having sessile or very shortly (to 0,3 mm) stipitate

glands on the rhachis at the junction of each pinnae

pair. Var. setulosa and var. lugardiae have a similar

facies and the two taxa are differentiated almost

solely on the type of glands on the leaf-rhachis. Both

taxa have a prominently western distribution in our

area, var. lugardiae extending the further east of the

two. The sessile glands appear to be a fairly reliable

diagnostic character and in general little difficulty has

been experienced in referring specimens to this taxon

except in the central Transvaal where some specimens

appear to grade into var. lugardiae. These problema-

tical specimens have a stalked gland at the junction

of the lowest pinnae pair and ± sessile glands at the

junction of all of the other pinnae pairs. It is perhaps

worth recording here that five of the eight taxa

encountered in our area have been recorded within

a ± ten kilometre radius of Pretoria, namely, subsp.

nyassana, subsp. africana varieties africana , lugardiae ,

setulosa and subsp. argillicola var. hirtipes.

Subsp. africana var. pubescens is very infrequent

and irregular in its occurrence in our area. The densely

pubescent leaflet surfaces appear to be quite distinc-

tive and enable this taxon to be recognized without

difficulty.

Subsp. forbesii, which occurs infrequently along

the Natal coast in our area, is distinguished from the

other subspecies in having subglabrous to densely

puberulous peduncles and glabrous to sparsely

appressed-puberulous young branchlets. There is some

evidence in our area that the degree of pubescence of

the peduncles alters with age; the peduncles of flower-

ing specimens on some plants are ± densely pubescent,

J. H. ROSS

267

while the peduncles on fruiting specimens on the same

plant are glabrescent. Unfortunately there is only one

flowering specimen from our area with subglabrous

peduncles. Most of our specimens with ± glabrous

peduncles and young branchlets, have a somewhat

different facies to the type specimen of subsp .forbesii.

CONCLUSIONS

Because of the ± continuous morphological varia-

tion within D. cinerea, Brenan & Brummitt’s treat-

ment of the species in our area is not altogether

acceptable. However, as it seems convenient, for

example, to distinguish formally those specimens with

broad leaflets and large leaves with stalked glands in

the eastern Transvaal, Swaziland and Natal from the

specimens with narrow leaflets and small leaves with

sessile glands in South West Africa, the western Trans-

vaal and northern Cape, it is felt that some infraspeci-

fic taxa within D. cinerea should be recognized. Since

it is not considered desirable to accept Brenan &

Brummitt’s treatment in its entirety, certain modifica-

tions to it are proposed.

The modifications proposed below are regarded as

provisional until the species has been thoroughly

investigated in the field in an attempt to analyse the

patterns of variation and to assess the taxonomic

significance of the differential characters. At this stage

I am not convinced that an elaborate infraspecific

hierarchial system is necessarily the best way to dis-

play or explain the patterns of variation, or that the

variation patterns within this complex species can

necessarily be successfully analysed. We appear to

be dealing with an aggregate of ecotypes with minor

morphological distinctions and in some areas of distri-

bution each population has its own characteristic

“look”. As no taxon in our area appears to be really

more distinct, and therefore of greater taxonomic sig-

nificance than any other, it is felt that they should all

have the same taxonomic rank. However, as the pro-

posed modifications are regarded as provisional, new

combinations have been avoided.

Although there is no distinct morphological dis-

continuity between subsp. nyassana and subsp.

africana var. africana in our area, it is proposed to

continue to uphold subsp. nyassana for those plants

with leaflets 2 mm or more wide. To maintain subsp.

nyassana will, of course, create certain difficulties and

it will be largely a matter of opinion whether some

specimens should be referred to subsp. nyassana or

to subsp. africana var. africana. However, in reaching

this decision I have been influenced by the fact that

over most of its range subsp. nyassana appears to be a

± distinct taxon. There is some evidence to suggest

that subsp. nyassana has slightly different ecological

preferences in certain areas, but field studies are

required to substantiate this. It would be interesting

to collect specimens over a period of years from cer-

tain selected plants which are ± intermediate between

subsp. nyassana and subsp. africana var. africana to

establish whether the maximum leaflet width and

leaflet size is fairly uniform or whether maximum

leaflet width fluctuates above and below 2 mm over

the years depending upon climatic conditions. Per-

haps the present limits between subsp. nyassana and

subsp. africana var. africana need to be re-defined

slightly.

As the abundant South African material reveals

that there is ± continuous morphological variation

between subsp. africana var. africana and var. lugar-

diae, it is felt that little is to be gained by continuing

to recognize both taxa. Consequently var. lugardiae

is relegated to synonymy within var. africana. The

lack of any well-defined morphological, geographical

or ecological discontinuities between subsp. africana

var. africana and the specimens which key out to

subsp. argillicola var. hirtipes persuades me to include

the latter under subsp. africana var. africana. The

range of morphological variation covered by the plants

included in var. africana in its original sense, was

considerable, so that the inclusion in var. africana of

specimens previously referred to var. lugardiae and

to subsp. argillicola var. hirtipes scarcely increases the

overall range of variation. The limits of subsp.

africana var. africana in our area are therefore en-

larged and it becomes the most widespread and

predominant taxon encountered. Var. setulosa appears

to be a ± distinct taxon with a fairly well defined

distribution and is therefore considered worthy of

retention. Var. pubescens is likewise upheld.

Two specimens collected near Ndumu in northern

Tongaland on the border of Mozambique, namely,

Strey & Moll 4014, 4020, fall within the limits of

subsp. africana var. plurijuga. These specimens have

slightly narrower leaflets and a few more pinnae pairs

than is usual in var. africana but, until more informa-

tion is available, it is intended to include these speci-

mens in var. africana.

Subsp. forbesii is infrequent in its occurrence along

the Natal coast. In the absence of a good selection of

flowering specimens with ± glabrous peduncles it

is difficult to assess the taxonomic significance of this

character. Sparsely appressed-puberulous young

branchlets are found in specimens of subsp. africana

var. africana in Zululand which have densely pube-

scent peduncles. Detailed field studies are required to

establish whether the plants with ± glabrous pedun-

cles have distinct ecological preferences or whether

they are merely variants within populations with

predominantly pubescent peduncles. Useful informa-

tion would result from collecting specimens over a

period of years from selected plants with ± glabrous

penduncles to establish whether any variation in the

degree of pubescence of the peduncles on a plant is

recorded. For the present those specimens with ±

glabrous peduncles are included in either subsp.

nyassana or in subsp. africana var. africana.

It seems opportune to consider the identity of

Acacia spinosa E. Mey., Comm. 170 (1836), which

was based on a specimen collected by Drege at Port

Natal. Oliv. in FI. Trop. Afr. 2: 333 (1871) and

Benth. in Trans. Linn. Soc. Lond. 30: 382 (1875)

cited A. spinosa as a synonym of Dichrostachys nutans

(Pers.) Benth., the latter now being regarded as a

synonym of D. cinerea subsp. africana var. africana.

Brenan & Brummitt l.c. 1 15 (1965) mention not having

seen a type specimen of A. spinosa so it was of great

interest to find an isotype in the Paris Herbarium.

The specimen consists of a sterile twig with a yel-

lowish-brown stem and spines. The leaf-rhachides

are up to 1 ,8 cm long, pubescent, and have a stalked

gland up to 1 mm high at the junction of each of the

4-5 pinnae pairs, the rhachillae are up to 3,8 cm long,

and the leaflets are up to 5,5 X 1,25 mm, glabrous

above, with sparsely ciliolate margins and scarcely

visible lateral veins beneath. The lack of flowers or

fruits is unfortunate but the specimen confirms the

earlier decisions to treat A. spinosa as a synonym of

D. nutans.

D. caffra Meisn. ex Benth. in Hook., J. Bot. 4:

354 (1841) is a nomen nudum which, according to

Bentham l.c., was based on Krauss 166. I have not

seen a specimen but Sonder in FI. Cap. 2: 278 (1862)

and Bentham in Trans. Linn. Soc. Lond. 30: 382

(1875) regarded D. caffra as a synonym of D. nutans.

and there is no reason to doubt their decisions.

17749-5

268

A NOTE ON DICHROSTACHYS C1NEREA IN SOUTH AFRICA

When using the following key, the width of the

largest leaflets must be used; if any leaflets are 2 mm

or more wide, the specimen should be referred to

subsp. nyassana. It is likely that most specimens can

be correctly placed, but intermediates occur between

most of the taxa, and these may cause difficulty. In

particular, it may be difficult to decide whether some

specimens should be assigned to subsp. nyassana or

to subsp. africana var. africana.

Until the species has been thoroughly investigated

in the field and an attempt made to analyse the

patterns of variation, this more conservative treat-

ment is preferred to the one adopted by Brenan &

Brummitt.

Leaflets some or all 2 mm or more wide; leaves often large and up to 18 cm long, with pinnae up

to 7,5 cm long; peduncles usually fascicled subsp. nyassana

All leaflets less than 2 mm wide; leaves smaller than above, pinnae usually less than 4 cm long;

peduncles single or sometimes fascicled:

Surfaces of leaflets (apart from the ciliate margins) glabrous or sometimes with few hairs on the

lower surface only:

Glands on leaf-rhachis stipitate or columnar, 0,5-2 mm tall, present at the junction of each

pinna pair or absent from some (very rarely the gland between the lowest pinna pair

stipitate and glands between the remainder ± sessile) ; leaflets 0 , 6-1 , 75 (2) mm wide ....

subsp. africana var. africana

Glands on leaf-rhachis sessile or very shortly (to 0,3 mm) stipitate, present at junction of all

pairs of pinnae; leaflets 0,5-0, 8 mm wide subsp. africana var. setulosa

Both surfaces of leaflets densely pubescent subsp. africana var. pubescens

Bothalia 11,3: 269-271 (1974)

The identity of Erythrina princeps

L. E. CODD*

ABSTRACT

An examination of available evidence leads to the conclusion that Erythrina princeps A. Dietr.

must be regarded as a synonym of E. humeana Spreng.

Krukoff and Barnaby in Phytologia 25: 17 (1972)

have expressed the opinion that Erythrina princeps

A. Dietr. (1834) and E. lysistemon Hutch, are con-

specific and that therefore the former name, being

the older, should be taken up for this species.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

E. princeps was described from a plant of unknown

origin grown in the Berlin Botanic Garden and

unfortunately the specimens known to have existed

were destroyed during World War II. However, a

photograph of the type specimen was taken in 1929

by the Field Museum of Natural History, Chicago,

and their negative No. 2375 is designated by Krukoff

and Barnaby as the neotype. At the request of Dr

Krukoff, the Field Museum kindly sent me a print

from this negative, which is reproduced here, with

their permission, as Fig. 1.

Fig. 1. — Erythrina princeps

A. Dietr., photograph of

the type taken in 1929 by

the Field Museum of

Natural History, Chicago,

and reproduced with their

permission.

270

THE IDENTITY OF ERYTHRINA PRINCEPS

An examination of this print shows that it does not

agree with E. lysistemon, which has a dense, compact

inflorescence (Fig. 2). In fact, it agrees in all respects

with E. humeana Spreng. Fig. 3 is a photograph of

Bot. Reg. t. 736A (1823), which may be regarded as

typifying E. humeana (see discussion below), and Fig. 4

shows a modern herbarium specimen of this species.

Before World War II, Dr KrukofT examined a

dissected flower from an authentic specimen of

E. princeps grown in Berlin Botanic Garden in 1844,

which was among those destroyed during the war.

He describes it as a narrow, declined flower with

included androecium that characterizes E. lysistemon.

It may be pointed out that this characteristic, while

serving to separate E. lysistemon from E. caffra, as

was Dr Krukoff’s intention, could apply equally well

to E. humeana as to E. lysistemon.

In addition to the markedly elongate inflorescence

of E. humeana, with spaced verticils of flowers, there

is a slight difference in the vexillum shape which

assists in distinguishing this species from E. lysistemon.

In E. humeana the vexillum is broadest towards the

apex producing an obtuse to clavate appearance

while, in E. lysistemon, the vexillum is broadest near

the middle and tapers towards the apex. In this

respect also, the type of E. princeps resembles E.

humeana.

An important character worth noting is the presence

of prickles on the petiole, showing in the type of E.

princeps and mentioned in the original description.

This is a constant characteristic of E. humeana but

in E. lysistemon the presence of an occasional prickle

on the petiole may be regarded as exceptional.

Fig. 2. — Erythrina lysistemon Hutch. ( Code / 7987 from Mqan-

duli) showing the dense, compact inflorescence of this

species.

Fig. 3. — A photograph of Bot. Reg. t.736A (1823), lectotype

of E. humeana Spreng.

Historically it is most unlikely that E. lysistemon

could have been grown to the flowering stage in

Europe by 1834. It has been observed here in Pretoria

that this species takes at least 10 years from sowing

to flowering, which would necessitate its introduc-

tion to Europe before about 1824. E. lysistemon

reaches its southernmost distribution in the Transkei,

a region unexplored botanically before Drege made

his momentous journey from the eastern Cape

Province to Natal in 1832. Although E. lysistemon is

common near “Port Natal”, this territory was also

virtually unexplored until the arrival in 1838 of the

naturalists, Krauss and Wahlberg.

There is no evidence that E. lysistemon was success-

fully cultivated in Europe until the latter part of the

19th century and than possibly only in the Mediter-

ranean region. As stated by KrukofT and Barnaby,

E. princeps (i.e. E. humeana) was widely grown in

European stoves by the mid 19th century.

E. humeana was originally figured in England as

E. caffra in Bot. Reg. 9: t. 736 A & B (1823): “intro-

duced into this country by Sir Abraham Hume, in

whose collection at Wormleybury the drawing was

made”, and in Bot. Mag. t.243 1 (1823): “flowered

for the first time, we believe, in the Count de Vandes

stove.” Both publications appeared on 1 September

1823. In 1826 C. Sprengel realised that the figured

plant was distinct from E. caffra and renamed it

E. humeana. In his protologue Sprengel refers only

to “Brot. Vag.”, which may be interpreted as Bot.

Mag. However, it seems probable that he was aware

of both plates because only in the Bot. Reg. text is

the name Hume mentioned. The Bot. Reg. plate

No. 736A is, therefore, selected as the lectotype.

L. E. CODD

271

Fig. 4. — Erythrina humeana Spreng., a modern herbarium

specimen from the eastern Cape Province ( Codd 9297).

The species is distributed from Albany and Bathurst

districts to Natal and northwards to Swaziland,

eastern Transvaal, Mozambique, Rhodesia and

Malawi. The typical form, with broadly ovate leaf-

lets, occurs in the eastern Cape Province while,

further north, a form with hastate leaflets is found

(Fig. 5). This form has been described as E. raja

Meisn. (after Raja, a genus of fishes, in allusion to

the aculeate petioles) and as E. hastifolia Bertol. f.

but, because there are intermediates, it is felt that this

form is not worthy of separate taxonomic status.

The synonymy relating to E. humeana is shown

below.

Erythrina humeana Spreng., Syst. Veg. 3: 243

(1826); Sim, For. FI. P.E.A. 43 (1909); Phillips in

Flow. PI. S. Afr. 3: t.l 12 (1923); Marloth, FI. S. Afr.,

2,1: 81, t.29 (1925); Collett in Bothalia 4: 225

Fig. 5. — Erythrina humeana Spreng., illustrating tne ionn with

hastate leaflets described as E. raja Meisn. and E. hastifolia

Bertol. f. ( Codd & Dyer 4636 from north-eastern Transvaal).

(1941); Codd, Bot. Surv. S. Afr. Mem. 26: 72, t.67

(1951); Batten & Bokelmann, Wild Flows. E. Cape

78, t.66 (1966); Hennessy, S. Afr. Erythrinas 21, t.6

(1972). Lectotype: Bot. Reg. 9: t.736A.

E. caffra sensu Ker, Bot. Reg. 9: t.736 (1823); sensu Sims in

Bot. Mag. t.243 1 (1823); sensu DC., Prodr. 2: 412 (1825),

partly; sensu Reichb., FI. Exot. 5: t.3 1 2 (1836).

Erythrina princeps A. Dietr. in Otto & Dietr., Allg. Garten-

zeitung 2: 305 (1834).

E. humei E. Mey., Comm. 151 (1836); Harv. in FI. Cap.

2: 237 (1862); Bak. in FI. Trop. Afr. 2: 182 (1871); Bak. f.,

Leg. Trop. Afr. 370 (1929). — var. raja (Meisn.) Harv., l.c.

(1862). — var. hastifolia (Bertol. f.) Bak. f., l.c. (1929).

E. raja Meisn. in Hook. Lond. J. Bot. 2: 96 (1843). Type:

Port Natal, Kraitss.

E. hastifolia Bertol. f. in Mem. Acc. Sc. Bolog. 2: 568, t.38

(1850).

Bothalia 11,3: 273-275 (1974)

Anomalous vascular bundle sheath structure in Al/oteropsis

semialata leaf blades

R. P. ELLIS*

ABSTRACT

Alloteropsis semialata (R.Br.) Hitchc. of the Paniceae, would be expected to exhibit typical

eupanicoid leaf anatomy with a single bundle sheath of large parenchymatous cells with specialized

chloroplasts and radially-arranged chlorenchyma. Specimens from South Africa showed the

following bundle-sheath and mesophyll deviations from the generalized panicoid model: 1. bundle

sheath is double; 2. inner bundle sheath is composed of large cells with specialized chloroplasts; 3.

outer sheath consists of many smaller cells with few, or without chloroplasts; 4. outline of the third-

order bundles is tall and narrow; 5. chlorenchyma is not radially arranged. Other specimens of

A. semialata, however, were found to have a large-celled outer bundle sheath devoid of chloroplasts

and with indistinctly radiate chlorenchyma.

The genus Alloteropsis has been placed in the tribe

Paniceae of the Poaceae. Species of this genus would,

therefore, be expected to exhibit a typical panicoid

leaf anatomy. Several specimens of A. semialata from

South Africa have, however, been observed to exhibit

fundamental differences in the structure and composi-

tion of the vascular bundle sheaths. The bundle-

sheath number is considered to be a diagnositic

character of importance in the basic classification of

the Poaceae (Metcalfe 1954) and thus the anomalous

condition observed in A. semialata warrants mention.

The panicoid grasses are generally characterized

by the presence or absence of an inner bundle sheath

(Avdulov 1931). The single or outer bundle sheath is

composed of relatively few, large, inflated paren-

chymatous cells imparting upon the vascular bundle,

sensu Metcalfe (1960), a distinctly angular outline.

The chlorenchyma cells of panicoid grasses are more

or less radially arranged.

Brown (1961) recognized three panicoid subtypes on

leaf anatomy, the Paniceae falling under the eupani-

coid subtype. This group of tribes is characterized

by the inner bundle sheath usually being absent and

wnh the chlorenchyma cells not radially arranged,

as seen in transverse section. These cells are normaly

tightly packed with few or no air spaces.

Within the panicoid grasses the main line of evolu-

tion appears to be from the typical eupanicoid struc-

ture towards the development of specialized plastids

for starch conversion and storage in the parenchy-

matous sheath cells (Brown 1958). This is accompanied

by a loss of the starch-synthesis function by the

chloroplasts of the chlorenchyma, the cells of which

show a corresponding, more regular, radial arrange-

ment. Associated with this development is the com-

plete loss of the inner bundle sheath.

Some species of Panicum, Tricholaena , Oplismemts ,

Brachiaria and Eriochloa. of the eupanicoid Paniceae,

retain an endodermis-like inner bundle sheath (Brown

1958). Tateoka (1956) found, in some species of

Panicum and other genera, that the cells of the

parenchymatous sheath do not have the specialized

starch plastids typical of the Paniceae. Thus there are

exceptions to the generalized model for the eupanicoid

subtype.

Metcalfe (1960) states that single bundle sheaths,

radiate chlorenchyma and angular, third-order vas-

cular bundles are diagnostic characters of the genus

Alloteropsis. These characters are repeated in the

descriptions given for A. cimicina (L.) Stapf and

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Alloteropsis sp. nov. Specimens of A. semialata

from South Africa showed considerable differences

from these generic characters and from the generalized

eupanicoid anatomy.

The following bundle-sheath and mesophyll

differences were observed in specimens of A. semialata

(Figs. 1-4, 6, 7):

1. the bundle sheath of all orders of vascular

bundle consists of two layers of parenchymatous

cells.

2. the inner bundle sheath is composed of inflated,

parenchymatous cells with a large, centrally-situated

chloroplast. These cells are larger than those of the

outer sheath. The inner bundle-sheath cells are

unthickened except in first-order vascular bundles

where the radial and inner tangential walls are

slightly thickened.

3. the outer bundle sheath is comprised of numerous

smaller, thin-walled cells without, or with a few

normal chloroplasts.

4. the outline of the third-order vascular bundles

is not angular, but tall and narrow.

5. the chlorenchyma cells are not tabular and

radially arranged as seen in section, but are hexagonal

to inflated in shape, with no definite pattern of

arrangement. Furthermore, these cells are distinctly

larger than the outer and inner bundle-sheath cells.

This anomalous structure was observed in the

following specimens, the voucher specimens being

deposited in the National Herbarium, Pretoria.

Transvaal. — 2528 (Pretoria); National Botanical

Gardens, Brummeria (-CA) 5-11-1969 R. P. Ellis 2\

1-12-1972 R. P. Ellis 1244 ; Faerie Glen (-CD)

22-11-1972 R. P. Ellis 721. Natal.— 2729 (Volks-

rust): Clontarf Siding, Newcastle (-DB) 12-11-1969

R. P. Ellis 36.

This structure bears a superficial resemblance to

the bundle sheaths of the genus Aristida of the aristi-

doid subtype of the panicoid grasses (Brown 1958).

Aristidoid grasses are characterized by two paren-

chyma sheaths and no endodermis-like inner sheath,

but the chlorenchyma is distinctly radiate. In the

sheaths of Aristida the larger inner-sheath cells also

possess abundant chloroplasts, more than in the

smaller, outer bundle-sheath cells (Lommasson 1957).

Other specimens of A. semialata were found to

have a large-celled, outer bundle sheath devoid of

chloroplasts, a typical mestome or inner bundle

sheath around the larger bundles and the chloren-

chyma indistinctly radiately arranged (Figs. 5, 8).

274

BUNDLE SHEATH STRUCTURE IN ALLOTEROPSIS SEMIALATA LEAF BLADES

Figs. 1-5. — Leaf blade outline of Alloteropsis semialata in transverse section. All x 6, 25, bright

field. Fig. 1, Ellis 2. Fig. 2, Ellis 721. Fig. 3, Ellis 1244. Fig. 4, Ellis 36. Fig. 5, Ellis 733.

The specimens exhibiting this structure are:

Transvaal. — 2530 (Lydenburg): Belfast (-CA)

6-2-1904 Burtt-Davy 1323. 2531 (Komatipoort):

Ida Doyer Nature Reserve (-CC) 21-1-1971 R. P.

Ellis 378 and Natal. — 3030 (Port Shepstone): Eureka

Farm, Izotsha (-CD) 20-12-1971 R. P. Ellis 733.

These specimens conform reasonably closely to the

Metcalfe (1960) description for the genus except for

the lack of chloroplasts in the sheath cells and the

presence of a double bundle sheath.

The occurrence of two distinct types of anatomical

structure within a single species is unexpected and

further investigation is required. Chippindall (1955)

notes that A. semialata is exceedingly variable in

inflorescence size and form and it is not improbable

that these anatomical differences are correlated with

morphological ones.

REFERENCES

Avdulov, N. P., 1931. Karyo-systematische Untersuchung

der Familie Gramineen. Rail. Appl. Bot. Genet. Plant

Breed., Suppl. 44: 1-428.

Brown, W. V., 1958. Leaf anatomy in grass systematics. Bot.

Gaz. 119: 170-178.

Brown, W. V., 1961. Grass leaf anatomy: Its use in syste-

matics. Recent Advances in Botany. I: 9th International

Bot. Congress 1959: University of Toronto Press.

Chippindall, L. K. A., 1955. In Meredith et al. Grasses and

Pastures of South Africa. Johannesburg: Central News

Agency.

Lommasson, R. C., 1957. Vascular bundle sheaths in the genus

Aristida. Phytomorphology 7 (3-4): 364.

Metcalfe, C. R. 1954. Recent work on the systematic anatomy

of the Monocotyledons. Kew. Bull. 4: 523-532.

Metcalfe, C. R., 1960. Anatomy of the Monocotyledons.

I. Gramineae. Oxford: Clarendon Press.

Tateoka, T., 1956. Re-examination of anatomical charac-

teristics of the leaves in Eragrostoideae and Panicoideae

(Poaceae). J. Jap. Bot, 31; 210-218.

R. P. ELLIS

275

Figs. 6-8. — Transverse sections of the leaf blade of Alloteropsis semialata. Fig. 6, Ellis 36x40 inter-

ference contrast. Fig. 7, Ellis 36x 100 interference contrast. Fig. 8, Ellis 733x40 interference

contrast.

Bothalia 11, 3: 277-290 (1974)

Notes on African plants

VARIOUS AUTHORS

ANACARDIACEAE

A NEW COMBINATION IN OZOROA

Fernandes in Bol. Soc. Brot. 27: 150 (1965) resus-

citated the genus Ozoroa Del. to accommodate all the

species formerly placed in Heeria Meisn. except for

Heeriu argentea (Thunb.) Meisn. which remained as

the only species of this monotypic genus. The main

reasons for resuscitating Ozoroa lie in the differences

in the fruits of the two genera. Since Heeria argentea

is very similar in general facies to some of the other

species formerly placed in Heeria , a special effort was

made to study fresh fruits. The main differences

mentioned by Fernandes could be confirmed and the

resuscitation of Ozoroa is therefore supported. Most

of the combinations in Ozoroa have already been

made in various publications. The species Heeria

concolor (C. Presl) Kuntze, however, is retained

under the genus Heeria in the Prodromus Sudwest-

afrika 74: 5 (1968) by Merxmueller and Schreiber

because mature fruits of this species were not avail-

able for study at that time. Even though they sus-

pected that it would prove to be an Ozoroa they

retrained from making the combination. The affinities

of Heeria concolor are clearly with Ozoroa dispar

(C. Presl) R. & A. Fernandes and O. crassinervia

(Engl.) R. & A. Fernandes. Recently fairly mature

fruits of this species have been collected by Mrs

Eve Jenkins and there is no further doubt that it

belongs in Ozoroa. The new combination is here

effected.

Ozoroa concolor (C. Presl ex Sond.) De Wint.,

comb. nov. Type: Cape, between Natvoet and Gariep,

1 000-1 500, Drege (S, holo.; TCD, PRE, iso.).

Anaphrenium concolor E. Mey. in Drege, Zwei Doc. 163

(1843), nom. nud.

Rhus concolor E. Mey. ex C. Presl, Bot. Bemerk. 42 (1844),

nom. nud.

Rhus concolor C. Presl ex Sond. in FI. Cap. 1 : 521 (1860).

B. de Winter

ASCLEPIADACEAE

A NOTE ON THE IDENTITY OF CELASTRUS FILIFORMIS

This note is the outcome of a request by Dr L. E.

Codd to investigate the identity of Celastrus filiformis

L.f.

Linnaeus fil., Suppl. 153 (1781), based his descrip-

tion of Celastrus filiformis on a Thunberg specimen

from the Cape Province. Linnaeus described the

species as follows: “Celastrus inermis, foliis lanceo-

latis integris, ramis filiformibus, pedunculi axillari-

bus uniflorus”.

Thunberg in his Prodromus 42 (1794) and FI.

Cap. ed. Schult. 217 (1823), repeated the protologue

almost verbatim but, in the latter work, he added a

more detailed description. From this latter descrip-

tion it was clear that the type specimen of C . filiformis

had opposite leaves.

Sonder in FI. Cap. 1: 470 (1860) cited, under

Mystroxylon eudeaeforme , M. filiforme Eckl. &

Zeyh., Enum. 125 (1836), but excluded the synonymy,

namely, “C. filiformis Thunb. FI. Cap. p. 217”,

which suggested that this basionym did not belong

to the same taxon as his M. eucleiforme. Sonder was

quite correct because M. eucleiforme has alternate

leaves, while the type of C. filiformis has opposite

leaves. The true identity of C . filiformis L.f., however,

was not revealed by Sonder in his treatment of

Celastrineae.

Following Sonder’s decision, it is somewhat sur-

prising to find that Davison in Bothalia 2: 327 (1927),

cited “ Celastrus filiformis Thunb. Prodr. (1794),

42; FI. Cap. ed. Schultz (1823), 217” as a synonym

of Cassine eucleiformis (Eckl. & Zeyh.) Kuntze,

particularly so as the latter is a younger name. As

species of Cassine have alternate leaves, it was clear

that Davison's decision was incorrect, and that the

identity of Celastrus filiformis remained unresolved.

Through the courtesy of the Director, Institute of

Systematic Botany, Uppsala, the type of Celastrus

filiformis from the Thunberg Herbarium, number

5606, was sent on loan to Kew (see Fig. 1). Examina-

tion of the specimen revealed that it belongs in

Asclepiadaceae and that it is a Secamone. To be more

precise, it matches what we have been accustomed to

calling 5. frutescens (E. Mey.) Decne. Drege 2230a

from near the Zwartkop River, Uitenhage division,

in the Kew Herbarium, an isosyntype of S. frutescens,

is a good match of the right-hand twig of Thunberg's

type, except that the latter lacks follicles. The basio-

nym of 5. frutescens (E. Mey.) Decne. (in DC.,

Prodr. 8: 501, 1844) is Astephanus frutescens E. Mey.,

Comm. 223 (1837). Celastrus filiformis L.f., Suppl.

153 (1781) is, therefore, a far earlier name and. as this

specific epithet has not previously been taken up in

Secamone , it will have to be adopted for this taxon.

The new combination is effected below.

Secamone filiformis (L.f.) J . H. Ross , comb. nov.

Celastrus filiformis L.f., Suppl. 153 (1781); Thunb., Prodr.

42 (1794); FI. Cap. ed. Schult. 217 (1823). Type: Cape Province,

Thunberg No. 5606 (UPS, holo.!)

Astephanus frutescens E. Mey., Comm. 223 (1837). Syntypes:

Cape Province, 3226 (Fort Beaufort), near Kat River, Drege

(K!); 3325 (Port Elizabeth), near Zwartkop River, Drege (K!);

Beans River (Boontjesrivier), Zuurberg Range, Drege (K,

fragm.!).

Secamone frutescens (E. Mey.) Decne. in DC., Prodr. 8:

501 (1844). Syntypes as for Astephanus frutescens.

Cassine eucleiforme sensu Davison in Bothalia 2: 327 (1927)

pro parte quoad syn. Celastrus filiformis L.f., non (Eckl. &

Zeyh.) Kuntze sensu stricto.

J. H. Ross

278

NOTES ON AFRICAN PLANTS

Fig. 1. — Thunberg 5606, the

holotype of Celastrus

filiformis L.f. (by permis-

sion of the Director,

Institute of Systematic

Botany, Uppsala).

EUPHORBIACEAE

A NEW SPECIES OF EUPHORBIA

Euphorbia celata R. A. Dyer , sp. nov. dioecia,

habitu subterraneo valde distincta. Herba perennis

dioecia tuberosa. Tuber plus minusve napiformis

durus 3-5 cm diam. Cattles subterranei 2 — plures,

erecti vel usque 15 cm rhizomatosi; rami epegaei,

1-2 cm longi, 3-5 mm crassi. Folia pauca alterna

subsessilia obovata vel non-nunquam rotunda,

3-6 mm longa, 3-5 mm lata, apice recurva acuta.

Bracteae 2 (3), foliaceae. hifiorescentia masculina

cymosa; cyathium primum sessile, lateralia 2 (3),

breviter pedunculata, cyathium circiter 4 mm longum,

infra glandulos 2,5 mm diam.; gland ul i transerve

late elliptici 1,5 mm lati concavi, margine leviter

crenulati; lobi oblongi circiter 0,75 mm longi,

fimbriati. Cyathium femineum solitarium; involucrum

mox deciduum. Ovarium sessile glabrum; styli

2,5 mm longi, fere ad medium lobati, apice

bifidi. Capsu/a 5,5-6 mm diam., sessilis obtuse

triangulata vel subglobosa; semina subpyriformia,

3 mm longa rugulosa.

Type: Cape Province, Vanrhynsdorp District,

Moedverloer, 4 km N. of Hoi River railway siding,

E. slope of white quartzite hill in grey clay beneath

quartz pebbles, 100 m alt., 12 May 1973, H. Hall

4272 (PRE, holo.).

VARIOUS AUTHORS

279

Perennial dioecious, glabrous herb with a

tuberous root-stock. Tuber more or less turnip-

shaped, 3-5 cm diam., hard. Stems 2-several, sub-

terranean, short or rhizomatous up to 15 cm long or

longer, slender or becoming tuberous at random

intervals in a tangled mass; branches above ground

1-2 cm long, 3-5 mm thick, succulent, with few

leaves and older parts with an incrustation of latex

exudate. Leaves few, obovate to subrotund, 3-6 mm

long, 3-5 mm broad, alternate, subsessile, somewhat

fleshy, dull grey-green, with acute, recurved apex.

Inflorescence: — Male, 1-few at ends of branchlets,

cymose; cymes subtended by 2(3) foliaceous bracts,

with subsessile central cyathium and 2(3) shortly

pedunculate, lateral cyathia; bracts usually somewhat

narrower than leaves, about equal in height to the

central involucre; peduncles about 4 mm long;

cyathia about 4 mm long, 4 mm diam. across in-

volucre-glands, 2,5 mm diam. below glands; in-

volucre early deciduous, with 5 glands and 5 oblong,

fimbriate lobes about 0,75 mm long; glands approxi-

mate, broadly transversely elliptic, 1 mm broad,

concave with inner margin upturned and outer

margin subcrenate. Female cyathia solitary, terminal;

bracts similar to those of male; involucre subcylin-

dric, 4 mm long, 1 ,5 mm diam., abscissile at base and

soon falling (leaving developing capsule exposed);

glands similar but smaller than in male and with

smooth outer margin; ovary sessile, obtusely 3-

angled; style 2,5 mm long, divided to nearly half-

way; stigmas bifid. Capsule 5,5-6 mm diam., 4-5 mm

high, very obtusely 3-angled or subglobose; seeds

somewhat pear-shaped, slightly angled, 3 mm long,

with rugulose surface. Fig. 2.

Only a botanically-interested person with keen

eyesight would have noticed the presence of this

insignificant-looking species. Such a person is Mr

Harry Hall. He first located it in 1970 and revisited

the site in May 1973 when his record reads very much

as follows: This morning I went to the locality,

exactly 40 km north of Vredendal (Cape Province),

with a blazing sun and sweltering north-east wind for

company. The state of the veld is as bad as ever I have

seen it, for there has been no rain for about a year.

1 knew the precise spot and found a few specimens

making a stout effort to flower: without the prelimi-

nary knowledge of my first discovery it would have

been worse than looking for the proverbial needle in

a haystack. Even so, 1 was on my hands and knees —

no fun on a quartzite gravel — for less than 2 cm of

dull grey-green growth appears above ground. The

plant has a fairly distinctive main tuber and the

branches spread out below ground to emerge at

random, sometimes quite far from the root-stock:

great care is needed when excavating from the iron-

hard clay for one never knows where the parent

tuber is.” The plant has no claim to beauty, is Harry

Hall’s final verdict.

Fig. 2. — Euphorbia celata

R. A. Dyer (H. Hall

4272, holotype, in PRE).

280

NOTES ON AFRICAN PLANTS

Botanically, however, it is of more than passing

interest because of its combination of characters:

it is dioecious, it has a hard, tuberous, storage root-

stock and mostly subterranean stems, thus ensuring

survival of the species in a most inhospitable environ-

ment. When Harry Hall made his first collection,

complete tubers were not obtained, but both male and

female branches were present. In the second collection

there were three complete male plants and one female.

Whether or not one might find a bisexual plant with

further searching is a question for future observation.

At this stage Harry Hall recollected having seen a

similar-looking plant in 1970 near Groenrivier, 39 km

S.W. of Garies, that is not much more than 130 km

north of Vanrhynsdorp, and decided to make another

visit on 6/6/73 on my behalf. His actual travelling

was 400 km from Vredendal and back. Again his

account is of special interest: “I enjoyed myself, all

alone, mining for the plants: remembering its peculiar

habit of virtually living under rocks I went armed with

a steel crowbar and was thus able to explore the full

range of the complicated rootsystem. Now and then

there was a tuber-like root, but mostly it was such an

entanglement of roots and branches and secondary

tuberous growths that one was never sure just how

many individual plants were involved. In the broken-

granite rock-clefts there was more aerial growth than

in the open where only the merest fragments came

above ground. Flowers were present but I saw no

sign of young fruits.” Harry Hall was also impressed

by the curious reticulated surface of the tuberous

growth. This may have been accentuated by a shrivell-

ing due to the very dry conditions.

I am satisfied that the two collections are of a

single species which varies in its underground growth-

form depending largely on the local geological forma-

tion. The characters of the cyathia from the two

collecting sites correspond closely.

Genera in several other families have evolved

different though equally interesting adaptations in

order to survive in the same area, such as, for example,

Crassulaceae, Mesembryanthemaceae, Asclepiadaceae

and even Geraniaceae.

The specific epithet celata, meaning hidden or

concealed, seems appropriate, firstly for a plant which

has escaped earlier detection in a botanically well-

trodden region and secondly it is well supported by the

observations of the collector. In seeking an affinity for

E. celata among the species recorded in the monograph

on the Succulent Euphorbieae of Southern Africa by

White, Dyer and Sloame, 1941, it finds a position

somewhere between such different species as E. garie-

pina Boiss. and E. juttae Dinter.

Cape. — 3117 (Garies): 39 km S.W. of Garies on N. bank of

Groenrivier (-DD), H. Hall 4282. 3118 (Vanrhynsdorp):

Moedverloer, 4 km N. of Hoi River railway siding (-AD),

H. Hall 3553; 4272.

R. A. Dyer

FLACOURTIACEAE

A NOTE ON THE AUTHOR CITATION FOR ERfUDAPHUS MUNDII, BASIONYM OF SCOLOPIA MUNDII

There is disagreement about the correct author

citation for Eriudaplnis mundii , the basionym of

Sco/opia mundii. Some authors, for example Warburg

( 1 893) and Gilg (1908, 1925), have attributed E. mundii

to Arnott. This is quite incorrect: Arnott had nothing

to do with the description of E. mundii in Ecklon &

Zeyher's Enumeratio plantae africae australis extra-

tropicae 2 : 272 ( 1 836).

Sleumer (1972), in his recent monograph of

Scolopia, regards Nees as the author, but the present

author feels that, on the evidence available, Ecklon &

Zeyher should rather be accepted as the authors.

The situation is that the treatment of Eriudaphus in the

Enumeratio is by Nees but, while two of the species

described, E. zeyheri and E. ecklonii, are clearly

attributed to Nees, the third, E. mundii, is attributed

to no one. Elsewhere in the Enumeratio, Ecklon &

Zeyher scrupulously give credit to authors whose

manuscript names are used, usually at the end of the

Latin description, but sometimes after the binomial.

In the case of E. mundii no author is mentioned at all,

which rather suggests that Ecklon & Zeyher inserted

the description of this species into Nees’s manuscript

and are therefore the authors.

Harvey, in his Genera Plantarum 1: 417 (1838),

wrote under Eriudaphus-. “Mr Zeyher has lately sent

me specimens of E. zeyheri and of a new species of

this genus, thus characterized : — 4. E. serratus . . .

which Harvey then proceeded to describe. From this it

is clear that Zeyher was aware of the existence of

E. zeyheri and, if E. mundii had already been described

by Nees, one would have expected Zeyher to use the

name E. mundii when communicating with Harvey

about the new species.

Unfortunately the destruction of Nees’s herba-

rium in Berlin makes it impossible to ascertain

whether Nees annotated the type of E. mundii, Mund

s.n.

Whether to attribute E. mundii to Nees or

Ecklon & Zeyher is, admittedly, a matter of opinion,

but the evidence available does seeem to point to

Ecklon & Zeyher. The full citation for Scolopia mundii

would therefore be S. mundii (Eckl. & Zeyh.) Warb.

This citation will be used in the Flora of Southern

Africa.

SCOLOPIA STOLZII , A NEW RECORD FOR SOUTH AFRICA

In November 1967 Mr R. G. Strey and Dr E. J.

Moll of the Botanical Research Unit, Durban,

collected sterile specimens of a tree growing in swamp

forest at the Sihadla River Crossing in Tongaland,

N. Zululand. The plant was tentatively identified in

the National Herbarium as “cf. Flacourtia in die a" .

A few months later, in April 1968, Mr Strey again

visited the area and collected fruiting material of the

plant. A further collection was made by Dr Moll and

Mr Nel in October 1971. On the author’s return from

Kew in 1971, Dr Moll communicated with the author

and suggested that the plant might be a Scolopia.

However, the material did not match any of the

known species of Scolopia in South Africa. On receipt

of flowering material from Dr Moll in December

1972, it was possible to confirm that the plant was a

Scolopia and further investigations in the National

Herbarium revealed that the plant closely resembled

several sheets of S. stolzii Gilg ex Sleum. var. stolzii

from tropical Africa.

VARIOUS AUTHORS

281

Specimens were then sent to Dr H. Sleumer in

Leiden, who recently revised Scolopia in Blumea

20: 1-64 (1972). Dr Sleumer reported that the

specimens could be referred to S. stolzii sensu lato

and that he wondered whether there were any real

differences between the two varieties stolzii and

riparia (Mildbr. & Sleum.) Sleum., which he had

upheld in his monograph.

In his monograph Sleumer distinguishes the two

varieties on the grounds of differences in leaf size,

texture, margin and pubescence of the ovary. An

examination of a wide range of material of S. stolzii

at Kew and the British Museum (Natural History)

by Dr J. H. Ross, and of isotypes of both varieties

by the present author, showed that these differences

do not hold. Ovary pubescence is a very variable

character in Scolopia, for example in S. flanaganii

(H. Bol.) Sim the ovaries range from completely

glabrous to densely short-pubescent. In view of the

above, the author will not uphold var. riparia in

his treatment of Scolopia for the Flora of Southern

Africa.

S. stolzii has several noteworthy characters. The

two lowermost pairs of nerves are acutely angled and

long-ascending (curved-ascending according to

Sleumer), while the rest are obtusely angled. The

lowermost pair is sometimes so near the base and so

prominent, that at a quick glance the leaf appears

to be digitately nerved. Moll & Nell (5608) describe

the foliage as pendent and the young leaves as deep

pink in colour. The flowers are sessile and solitary

or in pairs in the axils of the upper leaves. In this

character S. stolzii is similar to S. flanaganii which,

however, has up to five flowers.

South African material of S. stolzii in PRE:

Natal.— 2732 (Ubombo): Sihadla River Crossing, Killick &

Vahrmeijer 4061 ; Molt 4724 \ Moll & Nel 5608 ; Strey 8187 ;

Strey & Moll 3890.

D. J. B. Killick

1R1DACEAE

MONTBRETIOPS1S : REDUCED TO SYNONYMY IN TRITONIA (IRIDACEAE)

A diminutive but very attractive species of Iridaceae

was discovered early this century in the vicinity of

Prince Albert, north of the Swartberg Mountains.

This plant, with large, yellow, medianly zygomorphic

flowers is very reduced in size and lacks an aerial

stem at the time of flowering. When the seeds are

ripe, however, a stem may emerge a few centimetres

above ground. Marloth (1912) described the plant as

Gladiolus florentiae in honour of Lady Florence

Philips and he provided an excellent figure of it in his

Flora of South Africa (1915: t. 47a).

The choice of genus has proved incorrect as Louisa

Bolus (1929) pointed out. She acknowledged the

similarity of this plant to both Tritonia and to Mont-

bretia (a genus now regarded as part of Tritonia),

but decided to place it in a new genus, Montbretiopsis ,

on account of its reduced stem and the inflorescence

which she interpreted as cymose in contrast to the

spike in Tritonia. In this she was incorrect for the

inflorescence of her new genus is simply a reduced

spike, often with a single flower, thus the flowers

seem pedicellate. The position of the floral bracts

at the base of the ovary (indicating a sessile flower)

shows, however, that the flower stalk is not strictly

a pedicel, but the reduced scape itself. Thus the main

reason for the recognition of Montbretiopsis falls

away.

A close analysis of the morphology confirms the

contention that Montbretiopsis must be regarded as a

Tritonia. The two genera have almost identical

characteristics; their flowers are similar in shape, even

to the presence of small teeth or calluses in the centre

of each of the three lower perianth segments. The

style branches, usually a characteristic generic feature

in Iridaceae, are also similar, being quite long and

entire except at the very apex which is slightly ex-

panded or barely bifid. The corm tunics too are similar

in colour and texture.

One species of Tritonia, T. flava, occurring in the

arid Calvinia-Sutherland region of the Cape Province,

is in fact strikingly similar to Montbretiopsis in all

features except that T. flava usually has a short aerial

stem, though in a dry season this may not extend

above the ground. The similarity is so great that the

species are easily mistaken and can often be disting-

uished only by the bracts: those in Montbretiopsis

are dry-membranous while in T. flava they are some-

what herbaceous. Both bract types are found in

Tritonia and are not significant from the generic

standpoint.

The occurrence in other Cape Iridaceae of reduced

species in genera that are typically long stalked, is

worth mentioning at this point. In most examples

the reduced species occur in arid karoo or karoo-

margin areas and the reason for the adaptation

seems clear. Stemless species are known in the

following genera: Hesperantha , e.g. H. hantamensis,

H. rosea ; Lapeirousia e.g. L. plicata , L. montana;

Babiana, e.g. B. villosula; Moraea , e.g. M. falcifolia.

These examples all indicate that reduction of the

inflorescence axis alone is no criterion for recognition

of genera.

As a result of this survey, Montbretiopsis is reduced

to synonymy in Tritonia and the following combina-

tion is necessary:

Tritonia florentiae {Mari) Goldbl., comb. nov.

Type: Tygerberg, Prince Albert district, fi. May,

Marloth 4452 (PRE, holotype).

Gladiolus florentiae Marl, in Trans. Roy. Soc. S. Afr. 2: 241

(1912).

Montbretiopsis florentiae (Marl.) L. Bol. in S. Afr. Gard.

19: 215 (1929).

REFERENCES

Bolus, L., 1929. Plants new and noteworthy. .S'. Afr. Gard.

19: 215.

Marloth, R., 1912. Some new South African succulents and

other plants. Part IV. Trans. Roy. Soc. S. Afr. 2: 241.

Marloth, R., 1915. Flora of South Africa IV. Monocotyledons.

Cape Town and London.

Peter Goldblatt

282

NOTES ON AFRICAN PLANTS

LABIATAE

NEW SPECIES OF PLECTRANTHUS

Plectranthus xerophilus Codd, sp. nov. ; species

insignis, P. tomentoso Benth. primo adspectu maxime

simile sed calyce subaequaliter 5-dentato differt.

Suffrutex perennis radicibus tuberosis. Caulis

erectus parce ramosus, 1-1,7 m altus, teres vel

obscure quadrangularis, cano-tomentosus. Folia brevi-

ter petiolata vel subsessilia; petiolus 2-5 mm longus;

lamina ovata vel elliptica, 3,5-9 cm longa, 2,5-7 cm

lata, supra strigosa, subtus dense cano-tomentosa

glanduloso-punctata, valde reticulata, apice rotundato

vel obtuso, basi cuneata vel obtusa, margine profundo

crenato. Inflorescentia gracilis erecta subspicata vel

paniculata, 25-60 cm longa; bracteae lanceolatae

vel lineari-lanceolatae acuminatae deciduae, 1,5-3

mm longae, dense tomentulosae. Verticillastri 0,3-

2,5 cm distantes, 12-20-floribus; pedicelli 2-4 mm

longi glanduloso-tomentulosi. Calyx 2 mm longus

subaequaliter 5-dentatus, dense glanduloso-tomentu-

losus; dentes lanceolati, ± 1,5 mm longi. Corolla

violacea, glanduloso-tomentulosa; tubus 4 mm longus,

basi 1 mm late, fauce 4 mm late; labium posticum

cucullatum, 3 mm longum obscure 4-lobatum cymbi-

forme, 4-6 mm longum. Stamina 4, 7-8 mm longa,

basi breviter connatis. Stylus 9-10 mm fauce exsertus.

Nuculae brunneae, laeves, angulares.

Type. — Transvaal, 2430 (Pilgrims Rest): near

Marone (-CD), Codd & Dyer 7729 (PRE, holo.).

Perennial, slender shrub with thick, horizontal,

tuberous roots; stems erect, semi-woody, terete to

obscurely 4-angled, sparingly branched, 1-1,7 m tall

(including inflorescence), finely grey-tomentose. Leaves

opposite, subsessile to shortly petiolate, subcoriaceous,

ovate to elliptic, 3,5-9 cm long, 2,5-7 cm broad,

upper surface dark grev-green, strigose, lower surface

reticulate, densely grey-tomentose, tomentum of

crisped multicellular hairs, short, gland-tipped hairs

and numerous sessile gland-dots; margin coarsely

crenate; apex obtuse to rounded; base cuneate to

truncate; petiole 2-5 mm long. Inflorescence subspicate

or paniculate with 1-3 pairs of basal branches, carried

on slender peduncles up to 30 cm long; racemes up to

35 cm long; bracts early deciduous, lanceolate to

linear-lanceolate, acuminate, 1,5-3 mm long, densely

crisped tomentulose. Flowers densely clustered in

1 2-20-flowered verticillasters, verticillasters 0,3-2, 5

cm apart; pedicels 2-4 mm long, densely crisped

tomentulose. Calyx 2 mm long at flowering stage,

enlarging to 4 mm long in fruit, subequally 5-toothed,

densely crisped tomentulose and gland-dotted; teeth

lanceolate, up to 1,5 mm long, the uppermost

slightly broader than the rest. Corolla violet to mauve-

purple, crisped tomentulose and gland-dotted without,

glabrous within, bilabiate; tube 1 mm wide for 2

mm and then curved upwards, expanding into the

upper lip, 4 mm wide at the throat; upper lip hooded,

3 mm long, obscurely 4-lobed; lower lip boat-shaped,

4-6 mm long. Stamens shortly connate at the base,

7-8 mm long, curved upwards in the lower corolla

lip. Style curved upwards, exserted from the throat

by 9-10 mm. Nutlets brown, smooth, angled. Fig. 3

and 4.

Found on dry, rocky slopes in the eastern and

northern Transvaal and shows a marked ability to

survive under hot, arid conditions. Its flowering

season is from March to May.

Transvaal. — 2229 (Waterpoort): 6 km N. of Mara (-DC),

Meeuse 10199. 2331 (Phalaborwa): Kruger National Park,

Makadzi area (-CC), Van der Schijff 3853. 2429 (Zebediela):

24 km E. of Malipsdrift (-BB), 31/5/35, Obermeyer & Ver-

(loorn 23; between Chuniespoort and Malipsdrift (-BC), Pole

Evans 4370; near Chuniespoort Police Station (-BD), Meeuse

10351. 2430 (Pilgrims Rest): 3 km S. of Mica (-BB), Leach

11637; 3 km S.E. of Steelpoort (-CA) Codd & Dyer 7712;

50 km N. of Burgersfort (-CA), Leach & Jones 13249 ; 30 km

N. of Steelpoort Bridge on road to Penge Mine (-CB), Codd

10489; near Marone (-CD), Codd & Dyer 7729; Codd 8504;

16 km N. of Orighstad (-DA), Codd 10020; Echo Caves Motel,

30 km S. of Erasmus Pass tunnel (-DA), De Winter 7725;

Blyde River Gorge near Mariepskop (-DB), Marsh s.n.; 48 km

N. of Lydenburg (-DC), Dyer 3959.

P. xerophilus was first collected near Malipsdrift

in 1935 ( Obermeyer & Verdoorn 23) and has been

cultivated in our Botanic Garden for more than

20 years.

The delay in describing it is largely because of

uncertainty regarding its correct generic position.

In certain respects its characteristics would place it in

the genus Coleus , for example, the densely-glomerate

verticils, the bracts being very early deciduous and

the stamens being shortly connate at the base. On the

other hand, the calyx shape, which is subequally

5-toothed, is not found in Coleus, but is reminiscent

of Burnatastrum, a genus usually included in Plectran-

thus.

The modern tendency to take a broad view of

Plectranthus, and to include Coleus within it, removes

some of these difficulties. However, P. xerophilus

possesses characteristics which are unusual even in

this broader concept of Plectranthus, for example the

thick, horizontal roots, the slender, woody stems

which are subterete, not 4-angled, and the peculiar

hooded upper lip of the corolla. Consideration was

given to according it separate generic status, but

there appear to be insufficient grounds for such a

view.

The dense tomentum on the underside of the leaves

of P. xerophilus, and the fact that the bracts are

shed at a very early stage, gives this species a super-

ficial resemblance to Plectranthus Sect. Coleoides

Benth., including species such as P. zatarhendi

Forsk. and P. tomentosus Benth., but these species

have the conventional Plectranthus corolla and 1 + 4

calyx shape (i.e. the upper calyx tooth much larger

than the other four).

Plectranthus hilliardiae Codd, sp. nov., a P. ambiguo

(Bol.) Codd corolla tubo basi expanso 3-4 mm

diametro differt.

Herba erecta perennis 30-40 cm alta; caulis semi-

succulentus 4-angulatus parce ramosus, breviter

antrorsi-pilosus glandulosus, nodis hispidis. Folia

petiolata; petiolus 1,5-3, 5 cm longus; lamina late

elliptica vel obovato-elliptica, 5,5-9 cm longa,

4-6 cm lata, submembranacea, utrinque parce strigosa,

glanduloso-punctata, apice acuta vel obtusa, basi

cuneata attenuata, margine non profundo crenato-

dentata ciliata. Inflorescentia erecta raro ramosa;

racemi 8-15 cm longi, rhachis glanduloso-tomen-

tulosa; bracteae lineari-lanceolatae acuminatae persis-

tentes, 3 mm longae, glanduloso-pubescentes. Verticil-

lastri 1,5-2, 5 cm distantes ± 6-floribus; pedicelli

4-5 mm longi glanduloso-hispiduli. Calyx 4-5 mm

longus glanduloso-hispidulus; lobus posticus ovatus

erectus, 2,5 mm longus; lobi laterales deltoideo-

acuminati; antici lineari-deltoidei. Corolla lilacina,

minute glanduloso-puberula; tubus 2, 3-2, 7 cm

longus prope basin expansus, cylindrica, 3-4 mm

diam.; labium posticum erectum, 5-6 mm longum,

5 mm latum obscure 4-lobatum; labium anticum

cymbiforme, 4 mm longum. Stamina 4, filamentis

liberis, superna 6 mm longa, inferna 8 mm longa.

Stvlus 8-10 mm fauce exsertus.

VAROIUS AUTHORS

283

Fig. 3. — Plectranthus xe

philus Codd ( Codd

Dyer 7729, holotype,

(PRE).

Fig. 4. — Plectranthus xero-

philus Codd, showing the

horizontal, tuberous roots

{Hardy 2259).

17749-6

5' ?

284

NOTES ON AFRICAN PLANTS

UNIVERSITY Of NATAL

HERBARIUM

Type: Natal, 3130 (Port Edward): Umtamvuna

River (-AA), Hilliard & Burt 6767 (PRE , holo. ; NU).

Erect, perennial herb 30-40 cm tall, aromatic;

stems semi-succulent, 4-angled, sparingly branched,

shortly antrorse pilose and gland-dotted, with

occasional long, multicellular hairs and tufts of long,

stiff, multicellular purple-tinted hairs at the nodes.

Leaves: petiole 1,5-3, 5 cm long, densely appressed

glandular pilose, similar to the stems; blade broadly

elliptical to obovate elliptical, 5,5-9 cm long, 4-6

cm broad, semi-fleshy when fresh, drying membra-

naceous, dark green and sparingly strigose above,

under-surface paler, reticulate-veined, sparingly stri-

gose on the nerves and freely dotted with small,

transparent, sessile glands; apex acute to obtuse;

base cuneate, attenuate into the petiole; margin

ciliate, shallowly crenate-dentate above the middle.

Inflorescence erect, simple or sparingly branched,

with a pair of branches near the base; raceme 8-15

cm long; rhachis glandular tomentulose with multi-

cellular and short gland-tipped hairs; bracts linear-

lanceolate, acuminate, 3 mm long, persisting to the

fruiting stage, with gland-dots and long multicel-

lular hairs. Flowers in sessile, usually 3-flowered

cymes forming i 6-flowered vertici I lasters, verticil-

lasters 1,5-2, 5 cm apart; pedicels 4-5 mm long,

finely glandular-hispidulous. Calyx 4-5 mm long at

flowering stage, enlarging in fruit, glandular-hispi-

Fig. 5. — Plectranthus hilli-

ardiae Codd (Hilliard &

, Burtt 6767, holotype, in

PRE).

dulous near the base, subglabous above; upper lip

ovate, erect, not decurrent on the tube, 2,5 mm long;

lower lip 4-toothed, the lower pair longer than the

median; teeth linear- to deltoid-subulate. Corolla

pale bluish, purple-flecked on the lobes, minutely

glandular-puberulous; tube 2, 3-2, 7 cm long, slightly

deflexed and expanding abruptly near the calyx to

4 mm wide and parallel-sided or narrowing to 3 mm

wide at the throat; upper lip erect, 5-6 mm long and

equally broad, bilobed at the apex and with two lateral

ear-like lobes; lower lip shallowly boat-shaped,

4 mm long, horizontal or deflexed. Stamens free,

the lower pair the longer, up to 8 mm long, horizontal

to recurved. Style horizontal to ascending, exserted

by 8-10 mm. Fig. 5.

A softly semi-succulent herb which grows among

rocks near and in the margins of scrub forest and so

far collected only on high ground above the Umtam-

vuna River a few miles inland from Port Edward.

The flowering time is February-March.

Natal. — 3130 (Port Edward): Umtamvuna River (-AA)

26/2/71, Hilliard & Burt 6767; heights overlooking Umtam-

vuna River, 27/2/71, Nicholson s.n.

The species is named in honour of Dr Olive M.

Hilliard of Natal University, who, together with

Mr B. L. Burtt of the Royal Botanic Gardens,

Edinburgh, has made such valuable contributions to

our knowledge of the South African flora.

L. E. Codd

VARIOUS AUTHORS

285

LEGUMINOSAE

TYPIFICATION OF SCHOTIA CAPITATA

Bolle, in Peters Reise Mossamb. Bot. 1 : 18

(1861), based his description of Schotia capitata on

a flowering specimen collected by Peters at Inham-

bane in Mozambique. Unfortunately the holotype

in the Berlin Herbarium was destroyed during the

war, and all efforts to trace an isotype have proved

unsuccessful. In the British Museum (Natural History)

there is, however, a sketch prepared in 1926 of a

portion of the holotype in the Berlin Herbarium.

The sketch consists of one leaflet (2, 6x1,6 cm),

one flower bud, and carries the comment “material

very poor”. Unfortunately this sketch does not

enable S. capitata to be positively identified.

As mentioned by Codd in Bothalia 6 (3): 522

(1956), in establishing the identity of S. capitata ,

circumstantial evidence had. to some extent, to be

relied upon. Only two Schotia species are known

from Mozambique, namely, S. brachypetala Sond.

and the species interpreted as S. capitata. Collections

from the type locality have failed to reveal a third

species. Bolle provided a detailed floral description

of S. capitata and fortunately the reference to five

petals eliminates S. brachypetala from contention.

Although the application of the name S. capitata

is no longer in doubt, in the absence of any type

material it seems desirable to select a neotype. I now

select the flowering specimen Balsinhas 204 in the

Kew Herbarium from Mozambique, Louren^o

Marques Prov., Goba, proximidades do rio Maivavo,

as the neotype of 5. capitata.

J. H. Ross

POACEAE

A NOTE ON A POSSIBLE BIGENERIC HYBRID BETWEEN DIGITARIA AND ANTHEPHORA

In 1932 Goossens described Anthephora aequiglumis ,

distinguishing it from the closely-related Anthephora

pubescens Nees by the distinctly-elongated branches

of the rhachis (at least in the lower part), which are

about 7 mm long, and the subequal glumes which

are shorter than the lemmas which do not form a

false involucre. Later in that same year, Stent (1932)

recombined the species into a new genus, Tarigidia ,

which she described in order to accommodate this

plant. She maintained that “. . . the structure of the

spikelet is very similar to that of Digitaria, but the

elongated lower glume and the type of inflorescence

place Tarigidia beyond the limits of that genus”.

More recently, Launert (1957) described this grass as

Digitaria otaviensis , completely overlooking Tarigidia

aequiglumis. Subsequently in 1970 he sank his species

under T. aequiglumis.

On examination of the type of T. aequiglumis and

all available material, it became evident that this

grass shares characters of Anthephora and Digitaria ,

representing a possible bigeneric hybrid between

these two genera, the putative parents in this case

being Anthephora pubescens and one of the species of

Digitaria sect. Erianthae.

Tarigidia resembles Anthephora in having the same

type of spicate inflorescence, spikelets in clusters, the

production of the lower glumes, which are almost as

long as the lower lemmas, and by the occasional

fusion of the lower glumes. Resemblance to Digitaria

is evident in the nervation of the lower lemma, and

the characteristic flattened upper lemma with broad,

thin margins. One specimen in particular, Dinter

5589 (collected at Otavi in South-West Africa), has

the branches of the panicle produced on the lower

half with the lower glumes of many of the spikelets

reduced as in Digitaria. Here the similarity to that

genus is closer than it is to Anthephora.

It is generally accepted today that Anthephora

belongs to the tribe Paniceae, (c.f. Chippindall 1955;

Simon 1971 and Clayton 1972) however, Tatoeka

(1957) and Brown & Smith (1972) follow Pilger

(1954) in keeping Anthephora in a separate tribe,

Anthephorae.

Up to now T. aequiglumis has only infrequently

been collected within the region of overlap of Digitaria

sect. Erianthae and Anthephora pubescens (Fig. 6).

From the number of collections of T. aequiglumis

and from its disjunct distribution, it is evident that

hybridization, under natural conditions, if occurring,

is a rare phenomenon.

Attempts to hybridize Anthephora pubescens and

one of the species of Digitaria sect. Erianthae artifi-

cially should be undertaken to show that the progeny

resemble material of the presently known Tarigidia

aequiglumis.

Material examined

SWA — 1917 (Tsumeb): Otavi (-CB), Dinter 5589. This

specimen was designated as the type of Digitaria otaviensis

by Launert (1957).

O.F.S.— 2726 (Odendaalsrus): Hoopstad district, Odendaals-

rus, on an empty plot (-DC), Schultz s.n. sub. P RE 8344 (P RE

holo.); 2727 (Kroonstad); Groenebloeme Station (-LA),

Potts 2574 (PRE, para.).

Cape.— 2624 (Vryburg); 35,2 km N.W. of Vryburg on Genesa

Road (-DC), Acocks 12692.

REFERENCES

Brown W. V. & Smith, B. N„ 1972. Grass evolution the

Kranz syndrome 13C/12C ratios and continental drift.

Nature 239: 345-346.

Chippindall, L. K. A., 1955.

The grasses and pastures

Johannesburg: C.N.A.

Paniceae in Meredith et al.

of South Africa. 317-453.

Clayton, W. D., 1972. Gramineae in FI. W. Trop. Afr. 3,2:

354-356.

Goossens A. P„ 1932. The genus Anthephora Schreb. Trans. R.

Sac. S. Afr. 20: 189-200.

Launert, E„ 1957. Neue arten und erstfunde aus Siidwest-

afrik’a. Mitt. Bot. Miinchen 2: 307.

286

NOTES ON AFRICAN PLANTS

Launert, E., 1970. Gramineae in Prodr. FI. S.IV. Afr. 160: Stent, S. M. 1932. Notes on African grasses, XII. Kew. Bull.

202. 1932; 151.

Pilger, R., 1940. Unterfamilie Panicoideae. Gramineae III in Tateoka, T., 1957. Proposition of a new phylogenetic system

Pflanzenfam. Band 14 de. of Poaceae. J. Jap. Bot. 32, 9: 275-287.

Simon, B. K., 1971. Rhodesian and Zambian grass lists. Kirkia

8,1:3-83. A. E. LOXTON

NON-KRANZ (C3) LEAF ANATOMY IN THE PANICOID GRASSES: ADDITIONAL RECORDS FROM SOUTH AFRICA

Anatomically the Kranz syndrome is characterized

by the presence of a parenchymatous bundle sheath

composed of large, thick-walled cells, possessing

specialized chloroplasts and often surrounded by

radially-arranged mesophyll cells. Smith and Brown

(1973) and Brown and Smith (1972) point out that

these anatomical specializations are related to physio-

logical characters associated with the photosynthetic

fixation of carbon. At least two different photosyn-

thetic pathways are present in the Poaceae (=Grami-

neae), each being consistently associated with a

particular form of leaf anatomy (Carolin et al. 1973).

Thus the highly-specialized parenchyma sheath is

restricted to those plants exhibiting the C4 photo-

synthetic pathway and non-Kranz, C3 plants

have a thin-walled parenchyma sheath with a

few plastids, smaller than, but of the same type as

those of the mesophyll. Anatomical evidence is

therefore reliable for predicting the Kranz and non-

Kranz syndromes.

Smith and Brown (1973) give 13C/12C ratios for

198 grass species and correlate this with anatomical

and physiological data from the literature. Their

results show that all the species in the subfamilies

Arundinoideae, Oryzoideae and Festucoideae are non-

Kranz and all species of the subfamilies Eragrostoi-

deae and Aristidoideae are Kranz. Only in the sub-

family Panicoideae do both Kranz and non-Kranz

types occur. The tribes Andropogoneae, Anthepho-

reae, Arundinelleae, Maydeae and Melinideae are

Kranz, and the tribe Isachneae is non-Kranz. Only

in the tribe Paniceae itself are both Kranz and non-

Kranz genera found and the genus Panicum is reported

as being the only genus with both Kranz and non-

Kranz species. Non-Kranz, panicoid genera repre-

sented in South Africa and recorded by Smith and

Brown, are Sacciolepis and Oplismenus. None of the

non-Kranz species of Panicum mentioned by the above

authors are found in South Africa.

Listed below, together with specimen citations,

are further non-Kranz species of the Paniceae as

judged by leaf anatomical structure:

Acroceras macrum Stapf (Fig. 7.)

Transvaal. — 2330 (Tzaneen): Lakeside Holiday Resort

(-CA), Ellis 1595.

Cape. — 3228 (Butterworth): Mazeppa Bay (-BC), Ellis 266.

Alloteropsis semialata ( R . Br.) Hitchc.

Transvaal. — 2531 (Komatipoort): Ida Doyer Nature

Reserve (-CC), Ellis 378.

Natal. — 3030 (Port Shepstone): Eureka Farm, Izotsha

(-CD), Ellis 733.

Panicum aequinerve Nees

Transvaal. — 2531 (Komatipoort): 15 km from Barberton

on the road to Shya la Ngubo dam (-CA), Ellis 1541.

Natal. — 2829 (Harrismith): Cathederal Peak, stream banks

(-CC), Ellis 1453 ; in Rainbow forest, Ellis 1457. 2929 (Under-

berg): 16 km from Underberg on the road to Mondi (-DD),

Ellis 214. 3030 (Port Shepstone): Eureka Farm, Izotsha (-CD),

Ellis 1470.

Cape. — 3227 (Stutterheim): Fort Warden, Komga (-DB),

Ellis 283; Hogsback forest (-CA), Ellis 1476.

Panicum ecklonii Nees (Fig. 8.)

Natal. — 2729 (Volksrust): Majuba pass (-BD), Ellis 30.

2829 (Harrismith): Cathedral Peak (-CC), Ellis 1421 ; Ellis

1449. 2929 (Underberg): Kamberg (-BD), Ellis 53.

Panicum hymeniochilum Nees

Transvaal. — 2530 (Lydenburg): Badplaas (-DC), Ellis

465; 19 km from Nelspruit on the road to Alkmaar (-BD),

Ellis 1560. 2630 (Carolina): 16 km from Amsterdam on the

road to Piet Retief (-DA), Ellis 1523.

Panicum laticomum Nees

Cape. — 3227 (Stutterheim): Fort Warden, Komga (-DB),

Ellis 1101.

Panicum natalense Hochst.

Transvaal.— 2528 (Pretoria): National Botanical Gardens,

Brummeria (-CA), Ellis 417. 2530 (Lydenburg): Long Tom

pass (-BA), Ellis 453; Sabie, Ellis 460. 2531 (Komatipoort):

Ida Doyer Nature Reserve (-CC) Ellis 376.

Natal. — 2829 (Harrismith): Cathedral Peak (-CC), Ellis

1442. 3029 (Kokstad): Franklin (-AD), Ellis 223.

Fig. 7. — Acroceras macrum.

Transverse section of leaf

blade showing paren-

chyma sheath cells devoid

of chloroplasts and dis-

tinctly radiate chloren-

chyma. Ellis 266. X40.

VARIOUS AUTHORS

287

\ i 4)' * J.

s-s ■'j i:

- . ?

/*. s Kjft, , ,u«i;-S»a # ^ T ’

V c~;

: . i

* '

l V » '

Fig. 8. — Panicum ecklonii.

Transverse section of the

leaf blade. £7/w 30.

x6,25.

Panicum parvifolium Lam.

Transvaal. — 2528 (Pretoria): Cultivated in greenhouse.

Seed probably originally from Madagascar (-CA), Ellis 412.

Pseudechinolaena polystachya ( H.B.K .) Stapf

Natal. — 3030 (Port Shepstone): Eureka Farm, Izotsha

(-CD), Ellis 1465.

All specimens examined, of the above species, had

non-Kranz structure except for Alloteropsis semialata

where the majority of specimens possessed typically

Kranz leaf structure (Ellis, 1974). This is therefore

the first record of both Kranz and non-Kranz speci-

mens within a single species.

REFERENCES

Brown, W. V. & Smith, B. N., 1972. Grass evolution, the

Kranz syndrome, 13C/12C ratios and Continental drift.

Nature 239: 345-346.

Carolin, R. C., Jacobs, S. W. L. & Vesk, M., 1973. The

structure of the cells of the mesophyll and parenchymatous

bundle sheath of the Gramineae. Bot. J. Linn. Soc. 66:

259-275.

Ellis, R. P., 1974. Anomalous vascular bundle-sheath structure

in Alloteropsis semialata leaf blades. Bothalia 11: 273-275.

Smith, B. N. & Brown, W. V. 1973. The Kranz syndrome in

the Gramineae as indicated by carbon isotopic ratios. Amer.

J. Bot. 60: 505-513.

R. P. Ellis

THE CORRECT AUTHOR CITATION FOR TR1STACHYA SUPERB A

The author citation for Tristachya superba is

usually given as (De Not.) Schweinf. & Aschers.

(1867), the basionym being Loudetia superba De

Not. (1852). However, according to Hubbard in

FI. Trop. Afr. 10: 13 (1937), the generic name Loudetia

had not yet been validly described in 1852, having

originated on Hochstetter’s printed labels, on which

it was given without a description. Only in 1854

was a description to Loudetia provided by Steudel

(Syn. PI. Glum. 238). Loudetia superba was therefore

not validly published and has no standing as basionym

of Tristachya superba. The starting point for the

nomenclature of the species here concerned is

therefore 1867, and the correct citation is Tristachya

superba Schweinf. & Aschers.

P. VORSTER

POLYPODIACEAE

PYRROSIA SCHIMPERANA, A NEW FERN RECORD FOR SOUTH AFRICA

The first mention of Pyrrosia schimperana (Mett.)

Alston occurring in South Africa was by Van der

Schijff in Fauna & Flora 14: 51 (1963). In this he cited

Van der Schijff 5505 and Schweickerdt 2610 from

Mariepskop. In 1971 this information was repeated by

Van der Schijff & Schoonraad in Bothalia 10: 478,

482. These records seem to have escaped the attention

of taxonomists. No specimens were sent to either PRE

or BOL, and Schelpe, in FI. Zamb. Pterid. 147 (1970),

did not include South Africa in the distribution area

of P. schimperana.

Recently a further collection of P. schimperana has

been made, namely Scheepers & Engelbrecht 1999 from

the Blyderivierspoort Nature Reserve, downstream of

the Belvedere Creek confluence. It was collected in

gallery forest, on litter on a boulder. These specimens

matched material of P. schimperana from tropical

Africa. The identification was confirmed by Prof.

E. Schelpe, who also confirmed its novelty to the

South African flora. In a personal communication,

Prof. Schelpe told the author that he had observed the

species in this locality in the winter of 1972. He wrote:

“The largest, but inaccessible, specimens there are

much larger than any material I have seen from

Rhodesia.”

P. schimperana can be recognized by the combina-

tion of only one kind of stellate hairs with short,

flattened, white branches on the leaf surfaces, the non-

ciliate rhizome scales, and the closely-spaced leaves.

According to Schelpe (1971, l.c.) P. schimperana is

epiphytic and lithophytic in forest and tall woodland,

400-1 100 m above sea level.

P. schimperana is widely distributed throughout

tropical Africa, to Ethiopia in the north, Nigeria in the

west, and Mozambique in the east. Previous to the

records reported on above, the southernmost limit of

its distribution was in eastern Rhodesia. This shows a

disjunct distribution, with the 500 km-wide Limpopo

Valley separating the Transvaal and Rhodesian locali-

ties. A rather similar distribution is shown by Leucos-

permum saxosum S. Moore (Proteaceae), which occurs

abundantly on the Chimanimani Mountains in

Rhodesia, with small outlying colonies in the Pilgrims

Rest district of the Transvaal.

Transvaal. — 2430 (Pilgrims Rest): Blyderivierkloof,

Mariepskop (-DB), Van der Schijff 5505 (PRE, PRU); Schweic-

kerdt 2610 (PRE, PRU); Blyderivierspoort Nature Reserve

(-DB), Scheepers & Engelbrecht 1999 (PRE, BOL); Abel

Erasmus Pass (-BC), Van der Schijff 5638 (PRE, PRU).

P. VORSTER

288

NOTES ON AFRICAN PLANTS

STERCULIACEAE

A NEW SPECIES OF HERMANNIA

Hermannia helicoidea Verdoorn, sp. nov., ab

aliis speciebus pariter foliis cuneatis et infra albo-

tomentosis, cymis gracilissimis helicoideis differt.

Fruticulus ramosus, ramis gracilibus incon-

spicuis stellato-pubescentibus. Stipulae basi latae,

amplexicaules, decurrentes, apice mucronatae vel

attenuatae caudatae, 1 , 5-5 mm longae, basi 2-6 mm

latae. Folia petiolata; lamina anguste vel late oblongo-

cuneata, 3-10 mm longa, 2-6,5 mm lata, tenuiter

stellato-tomentosa, discolor, marginibus crenatis atque

undulatis; petioli 1,5-3, 5 mm longi. Inf/ore scent iae

axillares ac terminales cymas helicoideas ac gracilis-

simas formantes. Calyx campanulatus tenuiter stel-

lato-pubescens, c. 4 mm longus, 5-lobatus in dimidio

superiore sinubus latis. Petala c. 6,5 mm longa.

Stamina c. 4 mm longa filamentis obovato-oblongis

hyalinis. Capsula c. 4 mm longa stellato-pubescens.

Type: Clanwilliam District, Hex River Valley,

Pillans 9063 (PRE, holo.)

Shrublet, branched; branches slender, at first

minutely stellate pubescent, glabrescent. Stipules

broad-based, amplexicaul, decurrent, mucronate at

the apex or attenuate and caudate, 1,5-5 mm long,

2-6,5 mm broad. Leaves petioled; blade narrowly or

broadly oblong-cuneate, 3—10 mm long, 2-6,5 mm

broad, finely stellate-tomentose, discolorous with

margins crenate and undulate; petiole 1,5-3, 5 mm

long. Inflorescence of axillary and terminal, very

slender, helicoid cymes. Calyx campanulate, finely

stellate-pubescent, about 4 mm long, 5-lobed in the

upper half, with wide sinuses. Petals about 6,5 mm

long, more or less orbicular in the upper half, narrowed

below into a claw with infolded margins. Stamens

about 4 mm long, with obovate-oblong hyaline

filaments. Capsule about 4 mm long, stellate-pubes-

cent.

Recorded from the Clanwilliam district in the Hex

and Olifants River Valleys.

To date this species is represented by very few

specimens in herbaria, and all of the specimens are

from a restricted area. It appears to be a distinct

species, readily distinguishable from its allies with

similarly cuneate, dorsally albo-tomentose leaves

(e.g. H. repetenda , H. rigida and H. multiflora),

by the slender, recurved cymes which are both axillary

and terminal. It differs from two other species,

H. muricata and H. aspera which it resembles super-

ficially, not only by its characteristic inflorescence,

but also by the fact that the leaves of H. muricata are

more oblong in shape and more coarsely crenate,

and that H. aspera has sessile leaves and is roughly

pubescent.

Cape. — 3218 (Clanwilliam): without precise locality Stokoe

s.n.; hills in the Hex River Valley (-BD), Pillans 9063; Olifants

River Valley (— B ?), Schlechter 5092.

I. C. Verdoorn

TURNERACEAE

NEW TAXA IN THE TURNERACEAE

Wormskioldia lacerata Oberm., sp. nov., W,

longipedunculata Mast., affinis sed foliis profunde

laceratis echinulatis flavo-viridibus; caulibus sine setis

purpureis differt.

Planta perennis tota dense echinulata. Radix

lignescens crassa. Caules annui erecti vel ascendentes.

Folia linearia ad 12 cm longa profunde lacerata lobis

patentibus marginibus dentatis costa mediana inferiore

juventute albo-pubescentia. Racemus demum folia

longe superans; pedunculus erectus accrescens ad 20

cm longus. Flores majores ad 4 cm longi; petala

rubro-aurantiaca (capucina). Capsula reflexa teretia

ad 7 cm longa breve echinulata, rostrata.

Type: Transvaal, 2431 (Acornhoek): Newington

(-CD), Rogers 22578 ( PRE, holo.).

Chamaephyte with a long, strong, thick, woody

tap-root, soboliferous, the annual, erect or ascending,

densely leafy stems up to about 30 cm long; plant

echinulate all over, the patent sclerotic setae up to

3 mm long, not swollen below. Leaves pale green,

linear-lacerate, up to 12 cm long, 2-3,5 cm broad, the

narrow patent lobes unequal in size and shape, up to

c. 2 cm long, diminishing in size towards the acuminate

apex, margin dentate, the teeth with a scabrid tip,

the wide midrib setiferous below and, at least when

young, covered with a white, curly pubescence which

may also extend to the side nerves and margin.

Inflorescence erect in upper leaf axils, 1-4-flowered,

the buds close together; the rhachis at first bent to one

side, lengthening during anthesis and straightening

in fruit, the raceme with the reflexed capsules then

much overtopping leaves. Flowers developing one at

a time, opening in the morning, closing at night,

on short, erect pedicels which reflex in fruit, bracts

minute. Ccdyx-tube narrowly tubular, c. 12 mm long,

the lobes short, acute, erect, echinulate and sparsely

pubescent. Corolla rotate, with broadly obovate

orange-red petals narrowed into a claw below, which is

adnate to the calyx-tube above and fused to it below,

where the tube is woolly-pubescent around the ovary;

the claw bears a minute ligule or pustule just above the

fused area. Stamens 5, equal or with 2 shorter,

anthers just exserted from mouth of corolla. Ovary

terete, densely pubescent with erect hairs, styles 3,

just overtopping the stamens, stigmas apical, small,

minutely multipartite. Capsule reflexed, silique-like,

terete, up to 7 cm long, echinulate, rostrate, many-

seeded ; seeds straight, oblong-terete, 4 mm long,

reticulate, pale yellow, aril about half as long as seed,

unilateral. Fig. 9.

Recorded from the eastern Transvaal Lowveld in

open spaces or in disturbed areas.

Transvaal. — 2330 (Tzaneen): Hans Merensky Nature

Reserve (-DA), Oates 77; 5 km S. of Gravelotte (-DC), Leach

11559', Werdermann & Oberdieck 1922. 2430 (Pilgrims Rest):

10 km from Mica on the road to Hoedspruit (-BB ), Ihlenfeldt

2303 ; 16 km from Mariepskop on the road to Klaserie, (-DB),

Coetzer 151 . 2431 (Acornhoek): Klaserie (-CA), Killick & Strey

2511 ; Newington (-CD), Rogers 22578 (PRE, holo.); Bmtendag

909; Skukuza (-DC), Cholmondeley s.n. 2531 (Komatipoort):

Pretoriuskop (-AB), Van der Schijff 1075, 1217 , Codd & Winter

4918 ; Malelane (-BC), Lang sub TRV 31644; De Kaap Valley

(-CB), Thorncroft 54; Barberton (-CC), Galpin 747.

Swaziland. — 2531 (Komatipoort): Tshaneni, near Sand

River reservoir (-DC), Edwards 213.

This species was incorrectly identified as Worm-

skioldia schinzii Urb., (e.g. Burtt Davy, FI. Transv.

1 : 119, 1926) which is a western species based on a

plant collected by Newton in Angola. Urban in his

original description erroneously placed Newton’s

locality (Gambos, Caculovar River) in Mozambique.

W. juttae Dinter & Urban may prove to be a sub-

species of W. schinzii, as suggested by Schreiber in the

Prodromus of S.W. Africa, 88 : 4 (1968). W. lacerata is

confined to the Lowveld of the eastern Transvaal

VARIOUS AUTHORS

289

Hfl. "i; \\>nu Ml"; I'.i.

- ' ■ • - * ** * U'^iqtfCiU < - v: cu a * '' . '

' - ri

Fig. 9. — YVormskioldia lace-

rata Oberm. ( Rogers

22578, holotype, in PRE).

(with one record from northern Swaziland). It can be

distinguished from the western W. schinzii by the

long, echinulate setae (c. 2 mm long), the lacerated

leaves and the long capsules. In W. schinzii the plants

are scabrid and glandular-pubescent with the short

setae swollen below; the leaves are coarsely, doubly

serrate and the capsules about 3 cm long.

W. lacerata is closely related to If7, longipedunculata,

but apart from a more southern, fairly restricted

distribution, it is readily distinguished by its lacerate,

more yellow-green (not glaucous) leaves and its coarse,

echinulate pubescence present on stem, leaves, pedun-

cle, calyx and capsule. However, the long, red to purp-

lish setae found on the stems of W. longipedunculata

are absent.

290

NOTES ON AFRICAN PLANTS

Fig. 10. — Turnera oculata

Story var. paucipilosa

Oberm. (Giess 9377 , holo-

type, in PRE).

These perennial species with their deep, strong

root systems and soboliferous habit are not easily

eradicated from ploughed lands and road verges and

may persist in these surroundings for many years, the

bright orange-red flowers attracting attention. W.

lacerata appears to possess the largest flowers of the

seven species found in southern Africa. Unfor-

tunately little is known about the suspected heterostyly

in the genus. Field observations on stamen- and style-

lengthening are needed to throw light on this pheno-

menon*.

Turnera oculata Story var. paucipilosa Oberm .,

var. nov., a var. typica planta pauce pilosa dentibus

foliis profunde serratis bracteis 1 ,5-2 cm longis differt.

Type: South West Africa. 1713 (Swartboois-

drift): 32 km W. of Etengua near Otjitanda ( — AC),

Giess 9377 (PRE, hoi©.; WIND).

* In February 1974 Mr J. P. Nel, of the National Botanic

Garden, Nelspruit, collected a number of flowers of Worm-

skioldia lacerata at the Lekasi Township, 24 km E. of Nelspruit

(2531 CC, Komatipoort). From these it could be established

that heterostyly does occur for in some the styles were exserted

above the stamens, in others they were shorter than the stamens,

reaching only about halfway. In the long styled flowers the

styles were 9 mm long, the stamens 4 mm. In flowers with short

styles these were 4 mm long and the stamens 8 mm.

Shrublets 30-60 cm high, with a sparse, stellate,

apprcssed pubescence and white, simple strigose hairs

on younger parts and leaf-veins, the epidermis clearly

visible (in the typical variety it is hidden by the dense

pubescence). Leaves glaucous, variable, ovate to obo-

vate, up to 3 cm long, acute to obtuse, margin coarsely

serrate. Bracts linear c. 1,5-2 cm long. Flowers

similar to the typical variety. Capsules unknown.

South West Africa. — 1712(Posto Velho): Otjinungua, 12km

S. of Kunene River (-AB), De Winter & Leistner 5751. 1713

(Swartbooisdrift): 32 km W. of Etengua near Otjitanda (-AC),

Giess 93/7 (PRE, holo.; WIND); 8 km W. of Otjitanda (-AC),

Kotze 118.

This taxon is mentioned in the Prodromus Flora ot

South West Africa 88: 3 (1968) by Schreiber, but no

decision was reached about its status. The three

collections mentioned above come from three different

localities, close to one another and close to localities

where the typical variety occurs. In the material

available the two taxa can be easily distinguished.

A study of the South African representatives of this

family (Turneraceae) gives one the impression that the

species have become well stabilized and this, I believe,

supports the recognition of this taxon as a variety.

When more material and results of research become

available, its status can be reviewed.

A. A, Obermeyer

Bothalia 11,3: 291-294 (1974)

Notes on miscellaneous Acacia species from Tropical Africa

J. H. ROSS*

ABSTRACT

Information relating to a number of miscellaneous Acacia species from tropical Africa, is pre-

sented. Two new species, A. manuhensis and A. pseudonigrescens, are described.

While continuing studies on the tropical African

Acacias, attention was drawn to several matters

which required investigation. A number of decisions

requiring explanation in print form the subject of this

paper.

ACACIA ANDONGENSIS WELW EX HIERN

Welwitsch 1814 from Pungo Andongo in Angola,

the type specimen of A. andongensis Welw. ex Hiern,

Cat. Afr. PI. Welw. 1:314 (1896), is a fruiting speci-

men. The isotype in the British Museum (Natural

History) reveals that the young branchlets are armed

with paired, broad-based, recurved prickles and that

the leaves have up to 9 pinnae pairs and up to 19

pairs of leaflets per pinna. The leaf-rhachides are

unarmed and there is a small gland at the junction of

the top, or the top 3, pinnae pairs. The leaflets are

oblong or linear-oblong, up to 7,5 x 2 mm, discolo-

rous, with a basal tuft of pubescence on the lower

surface to one side of the midrib. The inflorescence

axes are fairly densely pubescent and the remains of

the calyces (just visible at the point of attachment of

some of the pods) are sparingly pubescent. The pods

are up to 17 cm long and 2, 5-2, 7 cm wide, glabrous

apart from some indumentum on the stipe.

A. andongensis is clearly very closely related to A.

goetzei Harms subsp. microphylla Brenan, but it

apparently differs from the latter solely in having

sparingly pubescent calyces. Pubescent inflorescence

axes do occur in A. goetzei, for example, Tweedie

1800 (K) from 30 km north of Kitale on the road to

Suam bridge, Trans-Nzoia district, Kenya. Similarly,

pubescent stipes also occur, for example. Batty 929

(K) from 16 km from Morogoro along the Iringa road

in Tanzania. So far, however, sparingly pubescent caly-

ces have not been definitely recorded in A. goetzei.

Although A. andongensis and A. goetzei are ex-

tremely closely related and are probably not really

specifically distinct, it is considered prudent at this

stage to maintain the two as distinct species, particu-

larly as A. andongensis is the earlier name and must

be adopted if A. andongensis and A. goetzei are

shown to be conspecific. The alleged distinctions

recorded between A. andongensis and A. goetzei in the

key to species in Consp. FI. Angol. 2: 269 (1956)

unfortunately do not prevail. It is perhaps significant

that A. andongensis is still known only from the type

collection. More material of A. andongensis from the

type locality is desired to enable the identity of the

species to be positively established. It seems reason-

ably certain, however, that further material will reveal

that the sparingly pubescent calyces of Welwitsch

1814 represent a local variant which is not worthy of

formal taxonomic recognition, and that the name

A. andongensis will have to supplant the more familiar

A. goetzei.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X 1 01 , Pretoria.

ACACIA ETBA1CA SCHWEINF. VAR. HIRTA A. CHEV.

Chevalier 8992 and 9050 from the Central African

Republic, the syntypes of A. etbaica Schweinf. var.

hirta A. Chev., in Bull. Soc. Bot. Fr. 74: 959 (1927),

are housed in the Paris herbarium and both are

flowering specimens. The young branchlets are pubes-

cent, the bark is reddish-brown and flaking minutely,

and the paired stipular spines are straight and up to

1 cm long. The leaves have up to 14 pinnae pairs and

the petioles, rhachides and rhachillae are sparingly to

fairly densely pubescent. The leaflets are in up to 24

pairs per pinna, up to 3,75 x 1 mm, discolorous,

glabrous apart from the conspicuous marginal cilia.

The flowers are in round heads on fairly densely pubes-

cent peduncles, and the involucels are up to one-third

of the way up the peduncle. The calyx is pubescent

apically, while the corolla is glabrous.

Although bearing a superficial resemblance to A.

etbaica , Chevalier 8992 and 9050 are in fact both

referable to A. gerrardii Benth. var. gerrardii. A.

etbaica is not recorded from west Africa. This oppor-

tunity is now taken of reducing A. etbaica var. hirta

to synonymy under A. gerrardii var. gerrardii.

A. gerrardii Benth. var. gerrardii.

Brenan in Kew Bull. 12: 369 (1958). Type:

Natal, Gerrard 1702 (K, holo.!, BM!, TCD!).

A. etbaica Schweinf. var. hirta A. Chev. in Bull. Soc. Bot.

Fr. 74: 959 (1927) synon. nov. Syntypes: Central African

Republic, Lac Iro, Chevalier 8992 ( P! ) ; Chevalier 9050 (P!).

ACACIA FISCHERI HARMS

The two syntypes of A. fischeri Harms, in Bot.

Jahrb. 51: 365 (1914), namely, Fischer 157 and

Stuhlmann 672 from Tanzania, were both destroyed

in the Berlin herbarium during the last war, and all

attempts to trace isosyntypes have so far proved

fruitless. In the British Museum (Natural History)

there is a sketch of the Berlin specimen of Fischer 157,

but unfortunately the sketch is rather poor and does

not really enable the species to be identified. In the

British Museum, however, there is also a flowering

specimen of A. fischeri, namely, B. D. Burtt 1131

from near Salia in the Kondoa district of Tanzania.

It carries the comment “Det. in Dahlem Herb. 5/12/

1928 by Dr Harms & B.D.B.” In the absence of any

isosyntypes, Burtt 1131 in the British Museum is

selected as the neotype of A. fischeri.

ACACIA GOSSWEILER1 BAK.F.

E. G. Baker, in J. Bot. 66, Suppl. Polypet: 156

(1928), based his description of A. gossweileri on

Gossweiler 1740 from Angola. The young branchlets

of Gossweiler 1740 are armed with paired, broad-

based, recurved prickles and the lower surfaces of the

leaf-petioles, rhachides and some of the rhachillae are

armed with numerous scattered, yellowish, recurved

prickles up to 2,5 mm long. The petioles, rhachides

and rhachillae are sparingly to densely clothed with a

292

NOTES ON ACACIA SPECIES FROM TROPICAL AFRICA

slightly golden, spreading indumentum. The petioles

have a small gland on the upper surface and the rhachi-

des have a small gland at the junction of the top 1-4

pinnae pairs. The leaves have up to 10 pinnae pairs and

up to 16 leaflets per pinna. The leaflets on the holotype

in the British Museum (Natural History) are all

fairly uniform and are up to 6 x 2 mm, but on the

isotype in the Kew herbarium, one leaf is much larger

than the others and bears larger leaflets which are up

to 10 x 3,5 mm. The leaflets are mostly linear-oblong

or oblong, slightly falcate, asymmetric basally,

rounded to subacute apically, discolorous, with a

somewhat prominent venation beneath, with ciliate

margins and pubescence on the proximal side of

the midrib on the lower surface, especially basally.

The inflorescence axes are up to 9 cm long (including

the peduncle) and are sparingly clothed with spreading

hairs. The flowers are sessile or almost so and the

calyces and corollas are glabrous.

Gossweiler 1 740 falls within the range of variation

of A. goetzei Harms. A. goetzei is extremely variable

vegetatively, especially in the indumentum and in

leaflet shape and size, and the armature of the leaf-

rhachis is also variable. As the leaflets in Gossweiler

1740 are mostly less than 3 mm wide and are oblong

or linear-oblong, as the leaf-rhachides are armed,

and as there is a small gland at the junction of the top

1-4 pinnae pairs, A. gossweileri is referred to synony-

my under A. goetzei subsp. microphylla Brenan.

A. goetzei Harms subsp. microphylla Brenan in

Kew Bull. 11: 204 (1956). Type: Malawi, Mombera

district, Njakwa to Fort Hill, Greenway 6393 (K,

holo. !).

A. gossweileri Bak. f. in J. Bot. 66, Suppl. Polypet: 156

(1928) synon. nov. Type: Angola, Benguela, Anha, rio Lutombi,

Gossweiler 1740 (BM, holo.!; K!).

ACACIA LATHOUWERSII STANER

Brec/o 1395 from Ishwa — Lac Albert in Zaire, the

holotype of A. lathouwersii Staner, in Ann. Soc. Sci.

Brux. 55 B: 311 (1935), is housed in the Jardin Natio-

nal de Belgique Bruxelles and consists of two twigs.

The left-hand twig, which is the larger, bears flowers

and pods, while the right-hand specimen bears flowers

only. Some of the spinescent stipules on each twig

are fused basally into deep reddish-brown to purplish,

± round “ant-galls” up to 3,5 cm in diameter. The

bark on the twigs is yellowish and split transversely

irregularly. The leaves are glabrous, have up to 12

pinnae pairs and the leaflets are glabrous throughout

or minutely ciliate. The peduncles are glabrous and

the involucels are basal. The flowers are whitish, the

calyces are glabrous and up to 1,5 mm long, and the

corollas are glabrous and up to 4 mm long. The pods

are falcate, finely longitudinally venose, glabrous,

attenuate at both ends and up to 5 mm wide.

Bredo 1395 is undoubtedly conspecific with A.

drepanolobium Harms ex Sjostedt. A. drepanolobium

is a rather variable species particularly in the degree

of pubescence, and A. lathouwersii appears to be

simply a glabrous form of this species. This opportu-

nity is now taken of reducing A. lathouwersii to syn-

onymy under A. drepanolobium.

A. drepanolobium Harms ex Sjostedt, Schwed.

Zool. Exped. Kilimanjaro 8: 116-117, t. 6 fig. 7-8,

t. 7 fig. 2-3 (1908). Type: Tanzania, Kiliman-

jaro, between Kwagogo and Moshi, Engler 1688 (B,

holo. f, K, drawings!).

A lathouwersii Staner in Ann. Soc. Sci. Brux. 55 B: 311

(1935). synon. nov. Type: Zaire, Ishwa — Lac Albert, Bredo

1395 (BR, holo.!).

ACACIA MACALUSOI MATTEL

Mattei, in Boll. Orto Bot. Giard. Col. Palermo 7:

94 (1908), based his description of A. macalusoi on

Macaluso 65 from Giumbo in the Somali Republic.

Unfortunately it has not been possible to locate

Macaluso 65. The specimen is thought to be housed

in the Instituto Botanico di Palermo, but attempts

to trace it have been unsuccessful. Consequently the

identity of A. macalusoi remains in some doubt.

Chiovenda, in Ann. Bot., Roma 13: 392 (1915),

mentions having received the type specimen on loan.

He found that there were two specimens mounted on

the sheet of Macaluso 65, a flowering twig and a fruit-

ing twig, and noted that the two twigs belonged to

two different species, the flowering twig being very

closely allied to A. Senegal (L.) Willd.. Because of this

mixed gathering, Chiovenda excluded the flowering

twig from A. macalusoi and amended the description

of the species accordingly, the fruiting twig thus

becoming the lectotype of the species. Owing to the

paucity of the fruiting twig, Chiovenda was unable

to establish to which species A. macalusoi was most

closely allied, but he concluded that it was not

allied to A. Senegal.

Subsequently, Chiovenda, FI. Somala 2: 186, fig.

113 (1932), referred Senni 135 and 152 to A. maca-

lusoi, fig. 1 13 being based on the latter specimen. These

two specimens were received on loan from the Her-

barium Universitatis Florentinae. Both are very poor

fruiting specimens but they are apparently members

of the A. goetzei Harms complex. As the remains of

the calyces (just visible at the point of attachment of

the pods) are sparingly puberulous, the specimens are

thought possibly to be referable to A. rovumae Oliv.,

but the material is too poor for one to come to any

definite conclusion. If the two specimens matched the

type specimen of A. macalusoi, then there is a distinct

possibility that A. macalusoi is not a good species.

ACACIA MANUBENSIS J. H. ROSS

Acacia manubensis J. H. Ross, sp. nov., A. goetzei

Harms affinis, sed leguminibus minoribus differt;

affinis etiani A. nigrescenti Oliv., sed foliolis numero-

sioribus et minoribus differt; ab utraque insuper

cortice flavido papyraceo desquamato differt.

Arbor parva, usque 5 m alta; cortice flavido papy-

raceo desquamato. Ramuli juveniles rubrobrunnei vel

atrogrisei, brevissime incano-puberuli, lenticellis

crebris notati. Stipulae baud spinescentes. Aculei

infrastipulares, ut videtur geminati, usque 3 mm

longi, cum ramulis concolores. Folia: petiolus 0,5-

1,4 cm longus, puberulus, supra interdum glandula

ornatus; rhachis 0-2,8 cm longus, puberulus; pinnae

1 -4-j ugatae ; rhachillae 0 , 8-3 , 8 cm longae, puberulae ;

foliola 3-7-juga, 7-12 mm longa, 3,5-9 mm lata,

oblique obovato-elliptica vel ± obovato-orbicularia,

apice rotundata usque obtusa, matura supra appresse

puberula, subtus appresse puberula, costa et nervis

lateralibus subtus satis obviis. Inflorescentiae spicatae;

spicis axillaribus solitariis vel fasciculatis, 1 ,8-3,5 cm

longi; pedunculis ± 1 cm longis, puberulis. Flores

sessiles. Calyx 2-2,5 mm longus, glaber. Corolla

3- 4 mm longa, glabra. Stamina numerosa; filamenta

4- 4,5 mm longa; antherae apice glandula caduca

coronatae. Ovarium breviter stipitatum, ± 1,2 mm

longum, glabrum. Legumina 5,8-8 cm longa, 1 ,7-2 cm

lata, oblonga, recta vel subrecta, purpurascenti-brun-

nea, minutissime puberula, subtiliter nervosa, apice

rotundata vel subtiliter mucronata. Semina haud

matura.

J. H. ROSS

293

Type: Somali Republic, Manub, 49° 50' E, 11°

05' N, J. G. B. Newbould 1080 (K, holo.).

Small tree to 5 m high; bark yellowish, papery,

peeling; young branchlets reddish-brown or dark

grey, very shortly and finely puberulous, with nume-

rous somewhat transversely-elongated lenticels, flak-

ing minutely. Stipules not spinescent. Prickles infra-

stipular, apparently in pairs, up to 3 mm long, the

same colour as the branchlets. Leaves: petiole 0,5-

1,4 cm long, puberulous, eglandular above or with

a small, slightly-raised gland; rhachis 0-2,8 cm long,

puberulous; pinnae 1-4 pairs; rhachillae 0,8-3, 8 cm

long, puberulous; leaflets 3-7 pairs, 7-12 x 3,5-

9 mm, oblique, obovate-elliptic or ± obovate-orbicu-

lar, apex rounded to obtuse, appressed puberulous on

upper and lower surfaces, midrib and lateral nerves ±

conspicuous beneath. Inflorescence spicate; spikes

1,8-3, 5 cm long, axillary, solitary or fascicled;

peduncles ± 1 cm long, puberulous. Flowers sessile.

Calyx 2-2,5 mm long, glabrous. Corolla 3-4 mm

long, glabrous, tinged with red. Stamens numerous;

filaments 4-4,5 mm long; anthers with a deciduous

apical gland. Ovary shortly stipitate, ±1,2 mm long,

glabrous. Pods 5,8-8 x 1 ,7-2 cm, oblong, straight or

almost so, purplish-brown, minutely puberulous, fine-

ly venose, apex rounded or minutely mucronate. Seeds

immature.

On account of the paired prickles, spicate inflores-

cences and flowers with glabrous calyces, A. manu-

bensis falls within the A. goetzei Harms — A. nigres-

cens Oliv. complex. It differs from A. goetzei in having

different leaflets and smaller pods, from A. nigrescens

in having more numerous and smaller leaflets and

smaller pods, and from both of these species in having

yellowish, papery, peeling bark.

A. manubensis is known only from the type collec-

tion. More material is desired.

I am grateful to Mr H. K. Airy Shaw for checking

the Latin description.

ACACIA PSEUDONIGRESCENS BRENAN & J. H. ROSS

Acacia pseudonigrescens Brenan & J. H. Ross>

sp. now; A. nigrescenti Oliv. affinis, sed cortice laevi

griseo pulverulento, calyce dense pubescente, corolla

adpresse pubescente, petiolis glandula magna appla-

nata vel ± depressa brevi spatio supra pulvinum sita

instructis, foliolis utrinque puberulis, nervis com-

pluribus conspicuis e basi ortis praeditis differt.

Arbor parva, gracilis, usque ad 5 m alta, ramulis

suberectis, cortice laevi griseo pulverulento. Ranuili

juveniles brunnei vel rubrobrunnei, juniores dense et

brevissime incano-puberuli, demum glabrescentes,

glandulis numerosis minimis conspicuis inter pilos

immixtis, epidermide desquamante, inermes. Stipulae

haud spinescentes. Folia: petiolus 1-2 cm longus,

dense et brevissime puberulus, supra glandula appla-

nata vel ± depressa magna rotundata vel elongata

usque 3,5 X 2,25 mm super basim petioli sita orna-

tus; pinnae 1-jugatae; rhachillae 2-3,2 cm longae,

dense et breviter puberulae; foliola 2-juga, 18-35 mm

longa, 11-31 mm lata, oblique obovato-orbicularia

vel late obovato-elliptica, basi asymmetrica, apice

rotundata et interdum plus minusve emarginata. coria-

cea, glauca, supra et subtus subtiliter appresse pube-

rula, venulis compluribus conspicuis e basi ortis

instructa. Inflorescentiae spicatae, spicis 2, 5-3, 2 cm

longis, solitariis vel in ramulis lateralibus abbreviatis

aggregatis; pedunculi 0,6-1 cm longi; pedunculi et

axes dense breviter pubescentes. Flores ochroleuci,

sessiles. Calyx cupularis, 2 mm longus, extra dense

pubescens, breviter 5-lobatus, lobis triangularibus.

Corolla 3 mm longa, extra adpresse pubescens, lobis

5, ovato-triangularibus, 0,5-1 mm longis, 0,75-1 mm

latis, apice subacutis. Stamina numerosa; filamenta

libera, usque 4 mm longa; antherae ± 0, 1 mm latae,

apice glandula caduca coronatae. Ovarium stipitatum,

± 1 mm longum, pubescens; stylus glaber, ± 2,5

mm longus. Legumen ignotum.

Type: Ethiopia, 8 km west of Mustahil on western

track to Kelafo, 5° 15' N, 44' 40' E, low limestone or

gypsum hillocks with cover of predominantly succu-

lent shrubs, 305 m, 21 June 1971, M. G. Gilbert 2129

(K, holo.).

Small, slender tree up to 5 m high, branches rather

erect, bark smooth, grey, powdery. Young branchlets

brown or reddish-brown, in places appearing as

though whitewashed over a purplish background,

densely and shortly puberulous when young, but

becoming glabrescent with age, with numerous minute,

conspicuous glands scattered amongst the hairs,

flaking minutely, unarmed. Stipules not spinescent.

Leaves: petiole 1-2 cm long, densely and shortly

puberulous, with a large, rounded or elongate-

flattened or depressed gland, up to 3,5 x 2,25 mm,

situated at the base of the petiole; pinnae 1 pair;

rhachillae 2-3,2 cm long, densely and shortly puberu-

lous, with a gland at the junction of the top leaflet

pair; leaflets 2 pairs per pinna, 18-35 mm long,

11-31 mm wide, obliquely obovate-orbicular or

broadly obovate-elliptic, asymmetric basally, apex

rounded and sometimes slightly emarginate, coria-

ceous, glaucous, finely appressed-puberulous above

and below, with several conspicuous veins arising

from the base. Inflorescences spicate, spikes 2,5-

3,2 cm long, solitary or aggregated in short, lateral

shoots; peduncle 0,6-1 cm long; peduncles and axes

densely and shortly pubescent. Flowers pale yellowish-

white, sessile. Calyx cupular, 2 mm long, densely

pubescent externally, shortly 5-lobed, lobes triangular.

Corolla 3 mm long, appressed-pubescent externally,

lobes 5, ovate-triangular, 0,5-1 mm long, 0,75-1 mm

wide, subacute apically. Stamens numerous: filaments

free, up to 4 mm long; anthers ± 0,1 mm wide,

with an apical, caducous gland. Ovary stipitate, ± 1

mm long, pubescent; style glabrous, ±2,5 mm long.

Pod unknown.

A. pseudonigrescens bears a strong superficial re-

semblance to A. nigrescens Oliv., but is readily

distinguishable from the latter in having a smooth,

grey, powdery bark, a densely pubescent calyx, an

appressed-pubescent corolla, a pubescent ovary, a

petiole with a large, flattened or ± depressed, discoid

or elongate gland situated a short distance above the

pulvinus, and finely appressed-puberulous leaflets

with several conspicuous basal nerves arising from the

point of attachment. In A. nigrescens the leaflets have

a distinct midrib and prominent lateral nerves, al-

though occasionally there are also a few relatively

inconspicuous basal nerves. In A. pseudonigrescens,

however, there is often no conspicuous midrib, but

rather a number of prominent basal nerves arising

from the point of attachment of the leaflet. The leaflets

in A. pseudonigrescens are finely appressed-puberulous

on both surfaces, while in A. nigrescens the leaflets are

either glabrous throughout or sometimes they are

clothed with semi-erect hairs above and/or below.

Leaflet size and shape in the two species are similar,

but the texture differs, the leaflets in A. pseudonigres-

cens being more coriaceous.

Although Gilbert 2129 is unarmed, it is anticipated

that further collections of this species may be armed

with paired prickles. It will be recalled that A. nigres-

cens is occasionally unarmed.

294

NOTES ON ACACIA SPECIES FROM TROPICAL AFRICA

In addition to the above morphological differences

between A. pseudonigrescens and A. nigrescens, a

large geographical discontinuity separates the two

species, A. nigrescens not being recorded further north

than Tanzania.

More material of A. pseudonigrescens, particularly

fruiting material, is desired.

J.P.M.B. & J.H.R.

ACACIA SEYAL DEL.

In the past A. seyal has often been confused with

A. hockii De Wild., but, although the two are closely

related, they are considered to be distinct species.

A. hockii differs from A. seyal chiefly by having a

non-powdery bark. The twigs are usually (but not

always) more elongate and slender, with a reddish or

brownish bark which does not peel to expose the inner

layer as, so characteristically, does the bark of

A. seyal, while the young branchlets are usually

clothed with a more or less dense puberulence which

is absent in A. seyal.

De Wildeman described three varieties of A. seyal

from Zaire, namely, var. lescrauwaetii, var. seretii

and var. kassonionga, but an examination of the

type specimens of each of these varieties revealed

that all are in fact referable to A. hockii and not to

A. seyal. A. hockii, which is widespread in tropical

Africa and occupies a wide range of habitats, is a very

variable species. All of the above variants fall within

the overall range of variation of A. hockii and none is

considered worthy of retention under A. hockii. The

three varieties are now relegated to synonymy.

A, hockii De Wild, in Feddes Repert. 1 1 : 502

(1913). Type: Zaire, Katanga, Luafu valley. Hock

s.n. (BR, holo. !).

A. seyal Del. var. lescrauwaetii De Wild, in Ann. Mus. R.

Congo Beige, Bot., Ser. V, 2: 128 (1907) synon. nov. Type:

Zaire, entre Lulua et Kanda-Kanda, Lescrauwaet 338 (BR,

holo.!).

A. seyal Del. var. seretii De Wild. l.c. : 128 (1907) synon.

nov. Type: Zaire, region du chef Guago, Seret 290 (BR, holo.!).

A. seyal Del. var. kassonionga De Wild., PI. Bequaert. 3:

65 (1925) synon. nov. Syntypes: Zaire, Kikosa, Delevoy 91

(BR!); nord de la Lukuga, Delevoy 224 (BR!).

ACACIA STENOCARPA HOCHST. EX A. RICH.

A. stenocarpa Hochst. ex A. Rich., Tent. FI.

Abyss. 1: 238 (1847), has frequently been misinter-

preted in the past and often confused with A. hockii

De Wild. It was therefore of great interest to find two

sheets of Schimper 1948 from Ethiopia, the type num-

ber of A. stenocarpa, in the Paris herbarium. Un-

fortunately the material on each of the sheets is rather

poor, one sheet consisting of a small flowering twig

and a fruiting twig, and the other of a small fruiting

twig, two pods and sterile fragments. However, they

do enable the species to be identified.

The young branchlets are sparingly puberulous and

armed with slender, paired, stipular spines up to 6

cm long. The epidermis of the branchlets is brown,

flaking off to reveal a powdery, yellowish inner layer,

and the leaves have 2-6 pinnae pairs and 9-14 leaflets

per pinna. The leaflets have inconspicuous lateral

nerves and are glabrous except for the ciliolate mar-

gins. The peduncles are glabrous or have few scattered

hairs, and the involucels are almost at or below the

middle of the peduncle. The calyx is inconspicuously

pubescent above and the corolla is glabrous. The pods

are linear, falcate, up to 1 1 cm long and 0,6 cm wide,

finely longitudinally striate, glabrous apart for some

sessile glands, and longitudinally dehiscent.

Schimper 1948 is undoubtedly conspecific with

A. seyal Del., and A. stenocarpa is now reduced to

synonymy under A. seyal var. seyal.

A. seyal Del, FI. Egypte Expl. Planches: 286, t.

52 fig. 2 (1813). Type: Egypt, Delile (?MPU, holo.).

A. stenocarpa Hochst. ex A. Rich., Tent. FI. Abyss. 1: 238

(1847) synon. nov. Type: Ethiopia, Schimper 1948 (P, iso.!).

ACACIA TORTILIS ( FORSK .) HAYNE VAR. LENT1-

CELLOSA CHIOV.

A. tortilis var. lenticellosa Chiov., FI. Somala 2:

197, fig. 121 (1932), was based on Senni 480 from the

Somali Republic. Although var. lenticellosa was not

validly published, the identity of Senni 480 is neverthe-

less of interest.

Senni 480, which is housed in the Herbarium

Universitatis Florentinae, consists of two sterile twigs,

each armed with a mixture of short, recurved and long,

straight, stipular spines, and two pods in a capsule

mounted on the upper left-hand side of the sheet. The

young branchlets are glabrous and lenticellate, the

leaves have up to 5 pinnae pairs, the leaf-rhachides are

sparingly pubescent and the leaflets are glabrous. The

pods are falcate, 5,5 mm wide, and glabrous apart

from a few hairs at the base of the stipe. Senni 480 is

referable to A. tortilis subsp. raddiana (Savi) Brenan

var. raddiana.

ACACIA TORTILIS (FORSK.) HA YNE VAR. PUBESCENS

A YLMER EX BURTT DA VY

Burtt Davy, in Kew Bull. 1930: 404 (1930), stated

that the type specimen of A. tortilis var. pubescens

Aylmer ex Burtt Davy, namely, Aylmer X8 from kilo

8 on Medani railway, Khartoum, Sudan, was housed

in the Kew herbarium, but all efforts to trace this

specimen have failed. Of the other three specimens

cited and determined by Burtt Davy as var. pubescens,

only Muriel S/9 and Letourneux 257 have been located.

Muriel S/9 from the Blue Nile near Wadi Medani is

now selected as the neotype of var. pubescens. Choice

of a neotype for var. pubescens is to some extent

academic, because var. pubescens Aylmer ex Burtt

Davy is a later homonym and synonym of A. tortilis

var. pubescens A. Chev. in Bull. Soc. Bot. Fr. 74:

960 (1927), the latter now being A. tortilis subsp.

raddiana (Savi) Brenan var. pubescens A. Chev.

ACACIA VERRUGERA SCHWEINF. VAR. SUBINERMIS

A. CHEV.

Chevalier, in Bull. Soc. Bot. Fr. 74: 959 (1927),

based his description of A. verrugera var. subinermis

on Chevalier 7598 from the Central African Republic.

The holotype in the Paris herbarium is a flowering

specimen. The young branchlet is yellowish, glabrous

or nearly so, and, as the varietal epithet implies,

unarmed. The leaves have up to 18 pinnae pairs and

up to 35 leaflets per pinna. The leaflets are narrowly

oblong, up to 5 x 1 mm, rounded apically, glabrous

throughout or with few small, marginal cilia. The

involucels are apical.

Chevalier 7598 is clearly referable to A. sieberana

DC. The absence of stipular spines is not considered

to be worthy of taxonomic recognition, and, as the

young branchlet is ± glabrous, Chevalier 7598 is

referred to A. sieberana var. sieberana. A. verrugera

var. subinermis is now reduced to synonymy.

A. sieberana DC., Prodr. 2: 463 (1825). Type:

Senegal, Sieber 43 (G, holo.:K!).

/I. verrugera Schweinf. var. subinermis A. Chev. in Bull.

Soc. Bot. Fr. 74: 959 (1927) synon. nov. Type: Central African

Republic, Ndelle, Chevalier 7598 (P, holo.!).

Bothalia 11, 3: 295-297 (1974)

Comments on the treatment of the Poaceae in the Prodromus einer

Flora von Siidwestafrika (1970)

B. DE WINTER- & P. VORSTER*

INTRODUCTION

The latest available revision of the Poaceae

(Gramineae) of South West Africa appeared in

Merxmiiller’s “Prodromus” (1970) under the

authorship of E. Launert. During the compilation of

the Master Index of South African Grasses, being

undertaken by the National Herbarium, Pretoria, a

number of species not recorded in the Prodromus came

to light. Some of these have, so far, been found only

in the Caprivi Strip, an area specifically excluded from

the Prodromus since it is covered by the Flora

Zambesiaca. For the purposes of the preparation

of the Flora of Southern Africa it is useful, however, to

bring the record up to date.

In a number of instances the species delimitation

adopted in the Prodromus is at variance with our

views. Since some of these concern species with a

mainly South African distribution, it was regarded as

necessary to elaborate on these differences even though

complete revisions of the groups concerned could at

present not be undertaken. Unnecessary name

changing and confusion will thus be avoided.

COMMENTS ON SPECIES DELIMITATION, NOMEN-

CLATURE AND IDENTIFICATIONS

Chloris radiata (L.) Swartz. S.W.A. — 2117

(Waterberg): Waterberg (-?), Giess, Volk & Bleissner

6615. Launert (p. 48) cites this number as C. virgata

Swartz.

Echinochloa frumentacea ( Roxb .) Link. S.W.A. —

1821 (Andara): Kake Camp between Andara and

Bagani (-BA), De Winter & Wiss 4373. (Cited by

Launert p. 71 as E. crusgalli (L.) Beauv. This material

matches the type of E. frumentacea which we regard as

a distinct species.

Eragrostis trichophora Coss. & Dur. De Winter

2378 is E. trichophora, not E. lehmanniana Nees as

quoted by Launert (p. 108). Note the whorled lower

branches of the inflorescence and the grey mem-

branous glumes. S.W.A. — 2217 (Windhoek): Neu-

damm Experimental Farm (-AD), De Winter 2378.

Paspalum paspalodes (Michx.) Scribn. Launert (p.

146) sinks P. vaginatum Swartz as a synonym under

P. distichum L. Clayton in FI. W. Trop. Afr. ed 2,3:

446 (1972) proposes to reject P. distichum as a nomen

confusum, as the name has constantly been misapplied.

We concur with his proposal, and if accepted, this

means that the species erroneously regarded as P.

distichum by most authors, needs a new name.

According to Clayton (personal communication) the

earliest name available is P. paspalodes (Michx.)

Scribn. S.W.A. — 2417 (Mariental): Hardapdamm,

GIB 1 10 (-DB), Giess, Volk & Bleissner 5598.

Paspalum vaginatum Swartz. This species is synony-

mous with P. distichum L. as was recognised by

Launert (p. 146). We regard P. distichum as a nomen

confusum (see above). The earliest alternative name

available for this entity is P. vaginatum Swartz.

S.W.A. — 2214 (Swakopmund): Swakopmund (-DA),

Research Station Gobabeb 00102; Van Vuuren 987.

* Botanical Research Institute, Private Bag X10I, Pretoria.

Setaria tenuiseta De Wit. S.W.A. — 1718 (Kuring-

kuru): Makambu Camp 32,8 km west of Kuringkuru

on the way to Katwitwi (-BC), De Winter & Marais

5023. This specimen is cited, incorrectly in our

opinion, as Setaria finita Launert (Launert, l.c. p.

173).

Setaria ustilata De Wit. S.W.A — 1821 (Andara):

between Bagani Camp and Mahango (-BA), De

Winter & Wiss 4403. 1917 (Tsumeb): ca. 16 km

southeast of Tsumeb (-BD), Basson42. 21 15 (Karibib):

Ohere-Ost (-BA), Merxmiiller & Giess 1594. Merx-

miiller & Giess 1594 is cited by Launert (l.c. p. 173) as

S. pallide-fusca (Schumach.) Stapf & C. E. Hubb. We

consider S. ustilata to be distinct from S. pallide-fusca.

The specimens cited above match the type of S.

ustilata in PRE. Ecologically they occupy a habitat

distinct from that of S. pallide-fusca. S. pallide-fusca

is a weed of agriculture and, where the rainfall is low,

prefers moist habitats. In areas of higher rainfall it is

common in disturbed areas as a ruderal weed. S.

ustilata, on the other hand, is confined to semi-

disturbed areas which are more arid, where it usually

grows in shady places, particularly in the humus-rich

soil under trees. Whereas it seems to grade into S.

pallide-fusca in the tropics, and is apparently included

in the latter in the Flora of Tropical Africa, it is

sufficiently distinct in our area to be separated as a

species. See Stapf in FI. Trop. Afr. 9: 819 (1930).

There may be an older name for this taxon, but, until

this is established, there seems to be no alternative to

using De Wit’s name.

Sporobolus acinifolius Stapf, S. glaucus Mez, S

salsus Mez and S. tenellus (Spreng.) Kunth

No material which we would refer to 5. tenellus

sensu stricto was seen from South West Africa or

Botswana, nor do we regard S. acinifolius as being

synonymous with S. tenellus. S. acinifolius is, however,

fairly widely distributed in all these territories.

Although occasionally difficult to distinguish, these

two species are in our opinion justifiably retained as

distinct species by Chippindall in Grasses and Pastures

of South Africa (1955).

The sinking of S. salsus and S. glaucus into

synonymy under S. tenellus can also not be supported.

However, S’, salsus and S. glaucus, in our opinion,

represent one species in spite of the difference in size

of the spikelets. In habit the latter two agree closely,

being distinctly tufted with fairly long leaves. By

contrast, S. acinifolius has a branched rhizome

producing short leaves, the basal parts being mat-

forming. The spikelets are 1,4-1, 6 mm long in S.

acinifolius and 2, 0-2, 9 mm in S. salsus (syn. S.

glaucus). The Southern African specimens were iden-

tified by us as shown below.

S. acinifolius Stapf. S.W.A. — 1815 (Okahakana):

Gemsbocklaagte, Etoshapfanne, Giess 2213. 1816

(Namutoni): Namutoni (-DD), Volk 2612\ east of

Namutoni (-DD), De Winter 2969. 2015 (Otji-

horongo): district Outjo, Wittklipp (-AC) Volk 2873.

2115 (Karibib): Okaukuejo, bei Okandeka (-DA).

296

COMMENTS ON POACEAE

Giess, Volk & Bleissner 6108. 2217 (Windhoek):

Neudamm Experimental Farm (-AD), De Winter

2375. 2517 (Gibeon): 2,2 km north of Asab (-BD),

De Winter 3478. 2519 (Koes): near Koes (-CC),

Acocks 15599. Cape. — 2624 (Vryburg): Vryburg

(-DC), Rodger sub PRE 32349: Henrici 174. 2723

(Kuruman): Kuruman (-AD), Mogg 7635 ; Carding-

ton (-BA), Esterhuysen 2215 ; 38 km east of Kuruman

on the road to Vryburg (-BD), Codd 1294. 2822 (Glen

Lyon): Lucas Dam (-BB), Wilman sub PRE 32350.

2824 (Kimberley): Koopmansfontein (-AA),

Brueckner 187 ; Barkly West (-DA), Acocks 155 .

S. salsus Mez (Syn.: S. glaucus Mez). S.W.A. —

1815 (Okahakana): E.N.P. Ondongab (-?), Du Toit

& Le Roux 456 ; 457: between Namutoni and

Okakwea (-DC), Volk 2413. 2416 (Maltahohe): 48 km

north of Mariental (-BA), Basson 268. 2417

(Mariental): Farm Narib GIB 106 (-BA), Giess, Volk

& Bleissner 5624.

The specimens De Winter & Marais 4666 from 13

km east of Tamso Camp in Omuramba Khaudum and

Le Roux 278 from Fischers Pan, Etosha National

Park, are best placed under S. salsus for the present.

More materia! is needed to determine whether they

form a part of the variation of this species.

S. tenellus ( Spreng .) Kunth. No material of this spe-

cies, sensu stricto , is known to us from South West

Africa. The following South African specimens are

referred to S. tenellus by us : O.F.S. — 2728 (Frankfort) :

Naauwpoort (-BC), Sim s.n. 2825 (Boshof): Boshof

(-CA), Burtt Davy 12399: 21 km N.N.E. of Dealesville

(-DB), Acocks 14012. 2924 (Hopetown): Luckhoflf

(-DB), Henrici 2156: Bergrivier, Fauresmith (-DD),

Smith 5429. 2925 (Jagersfontein): Fauresmith (-CB),

Smith & Pole Evans sub PRE 8831: Verdoorn 896:

Henrici 3069. 2926 (Bloemfontein): Bloemfontein

distr. (-AA), Acocks 12511. Cape. — 2824 (Kimber-

ley): Rooipoort (-CA), Leistner 1229. 3023 ( Brits-

town): Omdraaisvlei (-AA), Bryant 1190: Britstown

(-DA), Von Broembsen 81 . 3025 (Colesberg): Knoffel-

fontein (-AA), Smith 5385: Vrede, Philippolis (-AD),

Smith 4347 A. 3026 (Aliwal North): Groenvlei (-AD),

Henrici 3905: 3905 A: 3905B: 3905C: Smith 3889. 3122

(Loxton): 61 km S.E. of Loxton (-DD), Acocks 14377.

3124 (Hanover): 24 km S.S.E. of Richmond (-AA),

Acocks 15849: Hanover (-AB), Acocks 16350. 3322

(Beaufort West): Driefontein (-BC), Van der Merwe

2775. 3324 (Steytlerville): 34 km from Steytlerville on

the road to Uitenhage (-DA), Story 2337 .

Sporobolus natalensis (St end.) Dur. & Schinz.

Clayton in FI. W. Trop. Afr. ed 2, 3: 410 (1972)

points out that S. natalensis probably belongs in the

synonymy of 5. africanus (Poir.) Robyns & Tournay or

may be of hybrid origin. De Winter 3980 cited by

Launert (p. 184) under S. natalensis is a fairly typical

specimen of S. pyramidalis Beauv. S.W.A. — 1718

(Kuring-Kuru): junction of Mpungu Omuramba and

Okavango River between Tondoro and Lupala (-DD),

De Winter 3980.

Sporobolus subtilis Kunth, S. conrathii Chiov. and

S. welwitschii Rendle

We regard the first two species as distinct, not only

from one another, but also from S', welwitschii

Rendle, for the following reasons: S. subtilis has the

rhachilla produced and, like S. welwitschii, has slender,

oblique rhizomes terminating the culms (or clusters

of culms) at the base. It favours wet places such as

vleis. S. conrathii has no produced rhachilla and the

habit is caespitose with a dense covering of fibrous old

sheaths at the base (unlike S. welwitschii). No rhizomes

have been noted. It prefers open grassland but is

occasionally found in wetter places.

S. subtilis Kunth. Transvaal. — 2528 (Pretoria):

Rayton (-DA), De Winter 276: Schweickerdt 1756.

2630 (Carolina): Athole Pasture Research Station

(-CA), Norval 69. Natal. — 2732 (Ubombo): Mpan-

gazi Lake (-DA), Strey & Huntley 5013. Not recorded

from S.W.A.

S. conrathii Chiov. S.W.A. — 1724 (Katima Mulilo):

3 km southwest of Katima Mulilo (-CB), De Winter

9189. 1820 (Tarikora): 24,9 km. S. of Kapupahedi

on the track to Tamso (-DC), De Winter & Marais

4639. Transvaal. — 2428 (Nylstroom): Sterkrivier

Dam (-BC), Jacobsen 2614. 2528 (Pretoria): Premier

Mine (-DA), Menzies 16: Rayton (-DA),

Schweickerdt 1826. Grid reference unknown : Rusten-

burg distr., Crocodile River, Schlechter 3977 ;

Frankenwald, HBG 25042. O.F.S. — Grid reference

unknown : Maccau Vlei, Brandmuller 36.

S. welwitschii Rendle (syn.: S. baumianus Pilg.,

S. macrothrix Pilg.). These three species are

indistinguishable morphologically except by the length

of the spikelets which are 0, 8-1,0 mm long in S.

baumianus, 1 ,2-1 ,3 mm in S. welwitschii and 1 ,9-2, 1

mm in S. macrothrix. The bases of the type specimens

of these three species agree in that the new shoots are

enveloped in cataphylls. Long, fibrous, basal sheaths,

typical of S. conrathii, are always absent. In habit as

well as in the characteristics of the bases, the specimens

cited below resemble S. subtilis Kunth, but the

rhachilla is never produced.

S.W.A. — 1724 (Katima Mulilo): Katima Mulilo

area (-AD), Killick & Leistner 3110. 1820 (Tarikora):

13 km east of Tamso in Omuramba Khaudum (-DA),

De Winter & Marais 4681. De Winter & Marais 4639

named S. welwitschii by Launert (p. 186) is 5. con-

rathii Chiov.

NEW RECORDS FOR THE AREA COVERED BY THE

PRODROMUS

Alloteropsis cimicina ( L .) Stapf. S.W.A. — 1720

(Sambio): 31,6 km west of Nyangana Mission on road

to Rundu (-DC), De Winter & Marais 4582. 1724

(Katima Mulilo): 4,8 km east of Katima Mulilo

(-AD), De Winter 9141: ca. 1 1 ,2 km south of Katima

Mulilo on the road to Ngoma (-CB), Killick &

Leistner 3027.

Bromus diandrus Roth. S.W.A. — 2416 (Maltahohe):

Buellsport (-AB), Strey 988. Introduced species,

probably an escaped weed of cultivation.

Diplachne biflora Hack, ex Schinz. S.W.A. — 1720

(Sambio): 3,8 km east of Masari Camp on road to

Nyangana (-CC), De Winter & Wiss 4086.

Diplachne eleusine Nees. S.W.A. — 1913 (Sesfontein):

4 km east of Sesfontein (-BA), De Winter & Leistner

5584.

Eragrostis habrantha Rendle. S.W.A. — 1718 (Groot-

fontein): Grootfontein (-CA), Engler 6256. 2518

(Tses): Asis (-BC), Volk A. 39 in NH 30761.

Flyparrhenia dregeana (Nees) Stapf ex Stent.

S.W.A. — 1917 (Tsumeb): Auros, Hochtalboden

(-DA), Volk 624.

Monocymbium ceresiiforme (Nees) Stapf. S.W.A.—

1819 (Karakuwisa): Omuramba Omatako 66,9 km

south of Rundu (-DA), De Winter & Marais 5048.

1920 (Tsumkwe): Gautscha Pan (-DC), Watt S.W.A.

2395.

Paspalidium cf. P. platyrrhachis C. E. Hubb.

S.W.A. — 1719 (Rundu): Rundu (-DD), Volk 1795.

(This genus is in need of revision.)

Poa annua L. S.W.A. — 2116 (Okahandja): Ondon-

ganji, 24 km east of Omaruru (-AC), Maguire 2051.

(Common, weedy, introduced species).

B. DE WINTER AND P. VORSTER

297

Setaria cana De Wit. S.W.A. — 2117 (Waterberg):

Waterberg (-AC), Boss sub TRV 36343. This specimen

is represented in PRE by the holotype only. It seems to

be related to S. phragmitoides Stapf and may prove to

be synonymous with this species.

NEW RECORDS FOR THE CAPRIVI STRIP

Dinebra retroflexa ( Vahl ) Panz. S.W.A. — 1725

(Livingstone): Mpilila Island (-CC), Killick & Leistner

3342 ; Mpilila Island (-CC), De Winter 9242.

Echinochloa frumentacea (Roxb.) Link. S.W.A. —

1724 (Katima Mulilo): 12,9 km west of Katima

Mulilo (-DB), De Winter 9132.

Entolasia imbricata Stapf. S.W.A. — 1723 (Singalam-

we): Singalamwe (-CB), Killick & Leistner 3240.

Leptochloa uniftora Hochst. ex A. Rich. S.W.A. —

1725 (Livingstone): Mpilila Island (-CC), Killick &

Leistner 3347.

Loudetiopsis glabrata (K. Schum.) Conert. S.W.A. —

1724 (Katima Mulilo): Katima Mulilo (-AD), Ellis

344 ; Kabe floodplains (-DA), Brown s.n.

Monocymbium ceresiiforme (Nees) Stapf. S.W.A. —

1724 (Katima Mulilo): ca. 80 km from Singalamwe on

road to Katima Mulilo (-CA), Killick & Leistner

3289.

Sporobolus stapfianus Gancl. S.W.A. — 1724 (Katima

Mulilo): Ngoma area (-DC), Killick & Leistner 3003.

Bothalia 11, 3: 299-303 (1974)

Notes on Acacia species from north-east Tropical Africa

J. H. ROSS*

ABSTRACT

Information relating to a number of miscellaneous Acacia species described by E. Chiovenda

from north-east tropical Africa is presented.

Although an excellent account of the Acacia species

of tropical east Africa was published some years ago

(Brenan in FI. Trop. E. Afr. Legum.-Mimos., 1959),

doubt still surrounds the identity of a number of

species described by Emilio Chiovenda from north-

east tropical Africa, particularly from the Somali

Republic. Through the kindness of Prof. Dr C. H.

Steinberg, Conservator of the Elerbarium Universitatis

Florentinae, the type specimens of most of these

species were received on loan. Those decisions

requiring explanation in print form the subject of this

paper.

ACACIA BRICCHETTIANA Chiov.

Robecchi Bricchetti 533 from the Ogaden, Somali

Republic, the type specimen of A. bricchettiana Chiov.,

in Ann. Bot., Roma 13: 396 (1915), consists of two

flowering twigs. The older twigs of the previous season

are reddish- to dark brown, sparingly puberulous and

with numerous, somewhat transversely-elongated

lenticels, while those of the current season are reddish-

brown, with fewer lenticels but more densely

puberulous. The epidermis has split longitudinally in

places to reveal a yellowish inner layer. The paired

stipular spines are slender and up to 4,5 cm long.

The leaves are fairly densely puberulous throughout,

with minute, scattered glands along the rhachis, up to

11 pinnae pairs and 10-17 leaflet pairs per pinna.

The leaflets are up to 5 x 1,3 mm, the lower surface

is paler than the upper, the margins have conspicuous

cilia, while scattered hairs sometimes also occur on the

lower surface. The midrib (and occasionally a few of

the lateral nerves) is conspicuous and slightly raised

on the lower surface.

The peduncles are very short, up to 6 mm long,

densely pubescent and glandular; the involucel is

basal. The calyces and corollas appear to be reddish or

purple, but this impression may be the result of the

drying process. The calyces are up to 2 mm long and

are glabrous or subglabrous except for the lobes which

are pubescent. Likewise, the corollas, which are ±

3 mm long, are glabrous except for the lobes which are

pubescent. The corolla-lobes are up to 1 ,2 mm long,

so the corolla is divided almost down to the calyx.

As the description of Acacia g/overi Gilliland in

Kew Bull. 6: 139, t.4 (1951) seemed very close to that

of A. bricchettiana, Glover & Gilliland 388, the holotype

of A. gloveri from the Ogaden, Somali Republic, was

compared with Bricchetti 533. Vegetatively Glover &

Gilliland 388 is a good match of Bricchetti 533 agreeing

in that the older twig is dark ± purplish-brown,

transversely lenticellate and with the epidermis peeling

slightly to reveal a yellowish inner layer, while the

younger twig is reddish-brown and densely pube-

rulous. Similarly, the leaves agree in being fairly

densely puberulous throughout and glandular, and

the leaflets are of a similar size although some in

Glover & Gilliland 388 are marginally larger than those

of Bricchetti 533. Leaflet colour, venation and pubes-

scence in Glover & Gilliland 388 matches that of

Bricchetti 533.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

17749-7

Although Glover & Gilliland 388 is a fruiting speci-

men, the remains of an inflorescence is visible on the

holotype in the Kew Herbarium and also on the

specimen in the carpological collection. The peduncle

is puberulous and the remains of a basal involucel are

evident. The remains of dark-coloured spatulate

bracts, pubescent at the apex, are visible on the

inflorescence and these agree well with those on the

flowering specimen of Bricchetti 533.

In view of this general agreement it seems certain

that A. gloveri and A. bricchettiana are conspecific.

Consequently, I now reduce A. gloveri to synonymy

under A. bricchettiana, the earlier name.

Acacia bricchettiana Chiov. in Ann. Bot., Roma

13: 396 (1915). Type: Somali Republic, Ogaden,

Robecchi Bricchetti 533 (FI, holo.!).

A. gloveri Gilliland in Kew Bull. 6: 139, t.4 (1951), synon,

nov. Type: Somali Republic, Ogaden, between Wardere and

Walwal, Glover and Gilliland 388 (K, holo.!; BM, iso.!).

ACACIA CHEILANTHIFOLIA CHIOV.

Chiovenda, FI. Somala 1: 168, t. xvii, fig. 1 (1929),

based his description of A. cheilanthifolia on five

syntypes from the Somali Republic, namely, Puccioni

& Stefanini 479 (531), 509 (562), 663 (727), 762 (843)

and 1010 ( 1 1 1 5). Of the five syntypes, only one, namely

Puccioni & Stefanini 663, carries both flowers and

pods. The following description is based on Puccioni &

Stefanini 663 and any dimensions recorded other than

those from this specimen are placed in parenthesis.

The bark on the older twigs is ashen to greyish-

brown, often with a somewhat mottled effect, flaking

off here and there to reveal a yellowish-brown inner

layer, while the young branchlets are reddish-brown

and sparingly puberulous. The prickles are in threes

near the nodes, typically with the two laterals pointing

upwards and the median one pointing downward.

However, there is some variation in the degree of

curvature of the prickles and often they are ± straight

and spreading. In Puccioni & Stefanini 1010 and, to a

lesser extent and only in some instances, in Puccioni

& Stefanini 479 an 762, the median prickle also points

upwards so that all three prickles point upwards. The

prickles are up to 6 (12) mm long.

The leaves are small. The petiole is up to 6 mm long,

with a large gland up to 2 mm long, elongated along

the length of the petiole (sometimes rounded and only

0 , 5 mm in diameter), glabrous to sparingly puberulous.

The rhachis is up to 1,5 (2,4) cm long, glabrous to

sparingly puberulous, with recurved prickles on the

lower surface and with a fairly large discoid gland at

the junction of the top 1-2 pinnae pairs. The pinnae

are in (3-) 6 pairs and the rhachillae are up to 4(7) mm

long and are without a terminal, recurved prickle. The

leaflets are in 3-4 pairs per pinna, 1, 5-2(5) X 1(2,5)

mm, oblong to elliptic or rotund-ovate or obovate-

oblong, glabrous throughout.

The inflorescence axes are glabrous or subglabrous

and up to 4 cm long. The flowers are sessile and the

calyx and corolla are glabrous throughout. The pods

are yellowish-brown, 3-4(4, 5) X 1 cm, straightish,

rounded or mucronate apically, flattened, dehiscent'

venose and densely puberulous throughout.

300

NOTES ON ACACIA SPECIES FROM NORTH-EAST TROPICAL AFRICA

Puccioni & Stefanini 695 (769), the type specimen of

A. cheilanthifolia var. hirtella Chiov., differs from

typical A. cheilanthifolia in that the young stems,

petioles, rhachides, rhachillae, leaflets, inflorescence

axes and calyces are densely puberulous, and in that it

lacks pods.

On account of the prickles in threes at the nodes,

A. cheilanthifolia is clearly a member of the A. senega!

(L.) Willd. complex. A. cheilanthifolia is a distinct

species, differing from A. Senegal chiefly in having

fewer leaflet pairs and smaller pods. A. cheilanthifolia

bears a superficial resemblance to some specimens of

A. somalensis Vatke, but the latter differs in having

solitary prickles, usually 1 pinna pair per leaf, 2

leaflets per pinna and reddish-brown pods 1,5-1, 7

cm wide.

ACACIA CUFODONTII CHIOV.

Cufodontis 114 from Borana Prov., Ethiopia, the

type specimen of A. cufodontii Chiov. in Miss. Biol.

Borana Race. Bot. 4: 55, fig. 5 (1939), consists of a

single flowering twig. The greyish-brown, puberulous

twig is armed with recurved prickles which occur

singly near each node. The leaves have 2-4 pairs of

pinnae and the petioles, rhachides and rhachillae are

± densely pubescent. The leaflets are in 9-13 pairs per

pinna, up to 3 X 1 mm, and have conspicuous,

spreading, marginal cilia, while some leaflets are also

pubescent on the lower surface. The inflorescence axes

are ± densely pubescent, especially basally.

A. cufodontii clearly falls within the A. senega!

complex and is not specifically distinct from A.

senega!. Whether the prickles occur singly or in threes

near a node in A. Senegal seems to be of no significance

as both arrangements may occur on one and the same

shoot, although some gatherings may show all or

nearly all the prickles arranged singly. Although A.

cufodontii is not specifically distinct from A. senega!,

in the absence of pods and adequate field notes, it is

not possible to refer it to an existing variety of A.

Senegal with certainty. As the inflorescence axes are

pubescent, A. cufodontii is not referable to var.

leiorhachis Brenan. However, it could possibly be

either var. Senegal, var. kerensis Schweinf. or var.

rostrata Brenan. Lack of material and information

make the present knowledge of A. Senegal in north-

east tropical Africa most unsatisfactory.

Acacia Senegal ( L .) Willd., Sp. PI. ed. 4, 4:

1077 (1806) sens. lat.

A. cufodontii Chiov. in Miss. Biol. Borana Race. Bot. 4: 55,

Fig. 5 (1939), synon. nov. Type: Ethiopia, Borana Prov.,

Malca Guba sul Daua Parma, Cufodontis 1 1 4 (FI, holo.!).

ACACIA GORINII CHIOV.

Chiovenda, FI. Somala 2: 194, fig. 118 (1932),

based his description of A. gorinii on three syntypes

from the Somali Republic, namely, Gorini 221 and

Senni 435 and 449.

Gorini 221 consists of three twigs, the right-hand one

of which is in fruit while the two others are sterile.

There are four pods on the specimen, one attached to a

twig, two mounted on the sheet and one in the capsule

on the top left-hand corner of the sheet. The twigs are

yellowish-brown to greyish-white, glabrous or sub-

glabrous and lenticellate. The paired stipular spines

are ± straight and up to 1 ,3 cm long. The leaves have

up to 9 pinnae pairs, up to 13 leaflet pairs per pinna

and the leaflets are up to 9 x 2,5 mm. The pods are

chestnut-brown, up to 11,8 X 1,4 cm, dehiscent,

longitudinally striate and with flattened, ± wing-like

margins.

Senni 435 from Lac Badana consists of four almost

leafless, flowering twigs. The peduncles are short, up to

6 mm long, pubescent, and the involucel is in the lower

half of the peduncle. The calyx and corolla are

pubescent apically. The capsule mounted above the

collector’s label contains a single old pod together

with several inflorescences and leaf fragments.

Senni 449 from Obe, the third syntype, consists of

two flowering twigs, while the capsule mounted in the

bottom left-hand corner contains the remains of two

ancient pods and some seeds. Apart from having fewer

pinnae pairs and smaller leaflets, Senni 449 is

essentially similar to Gorini 221.

As it is quite clear that A. gorinii is conspecific with

A. nubica Benth., this opportunity is now taken of

reducing A. gorinii to synonymy.

Acacia nubica Benth. in Hook., Lond. J. Bot.

1: 498 (1842). Type: the Sudan, Kordofan, Kotschy

407 (K, holo.!; FI!; OXF!; PI; Z!)

A. gorinii Chiov., FI. Somala 2: 194, fig. 118 (1932), synon.

nov. Syntypes from the Somali Republic.

ACACIA HUMIFUSA CHIOV.

Chiovenda, FI. Somala 1: 163 (1929), based his

description of A. humifusa on Puccioni & Stefanini

416 (468) from the Obbia province of the Somali

Republic. Puccioni & Stefanini 416 is a poor, sterile

specimen in young leaf. The young branchlets are

greyish to greyish-brown, puberulous and lenticellate.

The paired, stipular spines are straight or almost so

and up to 1 cm long. The leaves are pubescent, with up

to 9 pinnae pairs and up to 11 leaflet pairs per pinna.

The leaflets have conspicuous marginal cilia and are

pubescent or glabrous beneath. The sole contents of

the capsule mounted on the right of the specimen,

appears to be some entomological curiosity.

Fortunately it is possible to match this sterile type

specimen against material of A. edgeworthiiT. Anders,

in the Kew Herbarium. Consequently, this opportunity

is now taken of reducing A. humifusa to synonymy

under A. edgeworthii.

Acacia edgeworthii T. Anders, in J. Linn. Soc. 5,

suppl. 1: 18 (1860). Syntypes from Aden, Edgeworth,

Hooker & Thomson (K!) and T. Anderson (K !).

A. humifusa Chiov., FI. Somala 1 : 163 (1929), synon. nov.

Type: Somali Republic, Obbia province, Garbauen e Durgale,

Puccioni & Stefanini 416 (FI holo.!).

ACACIA NERVOSULA CHIOV.

Chiovenda, FI. Somala 2: 192, Fig. 117 (1932),

based his description of A. nervosula on Senni 194

from Mogadiscio in the Somali Republic. Senni 194

is a rather poor specimen consisting of a sterile twig

with two detached pods and the remains of a third in a

capsule mounted in the top right hand corner. The

twigs are unarmed, grey, elongate, and along them are

numerous closely-arranged cushions representing

abbreviated lateral shoots from which the usually

clustered leaves arise. The leaves are tiny and have but

a single pair of pinnae; the petiole is up to 3 mm long

and the rhachillae are up to 1 ,4 cm long and fairly

densely pubescent. The leaflets are in 5-8 pairs per

pinna, up to 4,8 X 1,8 mm, obovate-oblong or

obovate-elliptic, sometimes slightly falcate, with or

without sparse marginal cilia, while some are

appressed-pubescent below. A few of the basal lateral

nerves of the leaflets are fairly conspicuous on the

lower surface. Only the smaller of the two whole pods

has a peduncle and this is ± I cm long and

puberulous. The pods are flattened and densely

puberulous; the larger is 11,6 x 2,2 cm and the

smaller 6,6 X 1,65 cm.

J. H. ROSS

301

Although there are no flowers, it is quite apparent

that Senni 194 is not an Acacia at all, but an Albizia.

The specimen agrees well with the description and very

limited material of Albizia obbiadensis (Chiov.) Brenan

and, consequently, Acacia nervosula is to be treated as

a synonym of that species.

Albizia obbiadensis (Chiov.) Brenan in Kew Bull.

17: 166 (1963). Type: Somali Republic, Obbia,

Biomal, Puccioni & Stefanini 553 (FI, holo.).

Acacia nervosula Chiov., FI. Somala 2: 192, fig. 1 17 (1932),

synon. nov. Type: Somali Republic, Mogadiscio, Senni 194

(FI, holo.!).

ACACIA OXYOSPRION CHJO V.

Senni 651 from the Somali Republic, the type

specimen of A. oxyosprion Chiov. var. oxyosprion,

FI. Somala 2: 188, fig. 115 (1932), consists of two

twigs. Each twig bears the remains of a few leaves and

inflorescence axes, while the one on the left also carries

three pods. The capsule mounted on the bottom left-

hand corner contains two pods, flowers, leaflets and

the fragments of some non-leguminous plants.

The bark on the old twigs is greyish- to purplish-

brown, while that on the younger stems is yellowish-

to reddish-brown. The prickles are in threes near the

nodes, the two laterals pointing slightly upwards or

almost at right angles to the twig and the median one

pointing downward. The only leaflets attached to the

specimen are 1 ,2 mm wide. The inflorescence axes are

sparingly pubescent basally and the pods are

yellowish-brown, up to 7 x 2,2 cm, flattened, papery,

venose, sparingly pubescent throughout and distinctly

rostrate apically.

There is no reason at all why A. oxyosprion should

not be regarded as conspecific with A. Senegal (L.)

Willd. The basally pubescent inflorescence axes and

the pubescent pods with distinctly rostrate apices in

A. oxyosprion var. oxyosprion , match those of A.

senega! var. rostrata Brenan from southern Africa.

There is, of course, a large geographical discontinuity

between the specimens with rostrate pods in the Somali

Republic and those in southern Africa. The Somali

material of A. oxyosprion var. oxyosprion sometimes

has slightly larger leaflets than is found in material of

A. Senegal var. rostrata in southern Africa but, apart

from this, there is little difference between the

specimens and the differences do not appear worthy

of any formal taxonomic recognition. Typical leaflets

and some larger leaflets sometimes appear together on

the same twig of A. oxyosprion var. oxyosprion in the

Somali Republic, for example Senni 822 (FI). This

opportunity is now taken of reducing A. oxyosprion

var. oxyosprion to synonymy under A. Senegal var.

rostrata.

Acacia Senegal (L.) Willd. var. rostrata Brenan

in Kew Bull. 8: 99 (1953). Type: Transvaal, Zoutpans-

berg district, Dongola Reserve, Verdoorn 2264 (K,

holo.! PRE!).

A. oxyosprion Chiov. var. oxyosprion in FI. Somala 2: 188,

fig. 115 (1932), synon. nov. Type: Somali Republic, Pozzi di El

Meghet, Senni 651 (FI, holo.!).

Unfortunately, Guidotti 21, the type of A. oxyosprion

var. pubescens Chiov., has not been available for

examination. From the description, var. pubescens

apparently differs from var. oxyosprion solely in having

a denser indumentum. If this is indeed the case then

there is a possibility that var. pubescens also belongs

to the same taxon as A. Senegal var. rostrata. How-

ever, as it would be unwise to pass judgement on var.

pubescens without first examining the type specimen,

no decision on its identity has been taken.

This uncertainty over the identity of A. oxyosprion

var. pubescens is emphasized, because if var. pubescens

and A. Senegal var. rostrata prove to belong to the

same taxon, then the varietal epithet “ pubescens ” has

priority over “ rostrata' ' and must be adopted for this

taxon once the new combination in A. Senegal has

been effectively published (but see comments under

A. Senegal var. pseudoglaucophylla). The type speci-

men of A. oxyosprion var. pubescens is thought to be

housed in the Institute Botanico dell’ Universita,

Modena, but efforts to borrow the specimen on loan

have failed.

ACACIA PARADOXA CHIOV.

Chiovenda, FI. Somala 1: 165, t.17/2 (1929),

based his description of A. paradoxa on Puccioni &

Stefanini 553 (585) from Ilbehla in the Obbia Province

of the Somali Republic. The specimen consists of

three twigs, each of which has young fruits. The

young branchlets are ashen and the prickles are in

threes at the nodes; the two laterals pointing upward

and the median one downward. The leaves appear

slightly glaucous, have 2-3 pinnae pairs and 5-6

leaflets per pinna. The leaflets are up to 4x1,6 mm,

linear- to obovate-oblong and glabrous. The petiole

and rhachis have few spreading hairs, while on the

lower surface of each rhachilla, a short distance

from the apex, there is a single recurved prickle.

The pods are immature, the largest being 6, 5x2,6

cm, apparently papery in texture and pale brown.

On account of the prickles occurring in threes at the

nodes, A. paradoxa Chiov. falls within the A. Senegal

complex. Several species of Acacia have prickles on

the rhachis, but the presence of a single prickle to-

wards the apex of each rhachilla is characteristic of

only one species, namely, A. hamulosa Benth. A

comparison of Puccioni & Stefanini 553 with material

of A. hamulosa , revealed that A. paradoxa Chiov.

cannot be distinguished from A. hamulosa. Conse-

quently I have no hesitation in reducing A. paradoxa

Chiov. to synonymy under A. hamulosa.

A. paradoxa Chiov. is, of course, an illegitimate

name, being a later homonym of A. paradoxa DC.,

Cat. Hort. Monsp.: 74 (1813). However, as A.

paradoxa Chiov. is now regarded as a synonym of

A. hamulosa , this obviates the necessity of giving

the taxon a new name.

Acacia hamulosa Benth. in Hook., Lond. J. Bot.

1: 509 (1842). Type: Arabia, hills near Gedda,

S. Fischer 72 (K, holo.!).

A. paradoxa Chiov., FI. Somala 1: 165, t.17/2 (1929) synon

nov., non A. paradoxa DC., Cat. Hort. Monsp.: 74 (1813).

Type from the Somali Republic, Puccioni & Stefanini 553

(FI, holo.!).

ACACIA PUCCIONI AN A CHIOV.

Puccioni & Stefanini 158 (177), the type specimen

of Acacia puccioniana Chiov., FI. Somala 1: 164,

t. 14 Fig. 2 (1929), consists of a single flowering twig.

The presence of spines terminating short lateral,

spreading shoots which bear leaves and flowers,

and of flowers with only 10 stamens, indicates, that

the specimen is not an Acacia at all, but a species of

Dichrostachys.

The young stems are pubescent and the leaves are

very small, up to 8 mm long in all. The leaves have

1 or 2 pairs of short pinnae, each of which has up to

7 pairs of tiny, pubescent leaflets which are up to

2 mm long and 0,7 mm wide. Unfortunately there are

no pods.

302

NOTES ON ACACIA SPECIES FROM NORTH-EAST TROPICAL AFRICA

The specimen is not referable to Dichrostachys

cinerea (L.) Wight & Arn. and does not appear to

match material of D. kirkii Benth. or of any other

species represented in the Kew herbarium. Further

material from the Somali Republic may reveal the

identity of Puccioni & Stefanini 158. However, as the

identity of Puccioni & Stefanini 158 is not clear, I am

refraining from making a new combination of the

specific epithet in Dichrostachys in case the specimen

does prove to belong to an existing taxon.

ACACIA SENEGAL (L.) WILLD. VAR. PLATYOS-

PRION CHIOV.

Chiovenda, FI. Somala 2: 187, Fig. 114 (1932),

based his description of A. Senegal var. platyosprion

on six syntypes from the Somali Republic, namely,

Senni 87, 116, 191, 201, Gorini 421 and Guidotti 3.

Five of the syntypes have been examined, but unfor-

tunately the sixth, namely Guidotti 3, has not been

available for examination.

Three of the syntypes, namely Senni 87, 116 and

191 are fruiting specimens, while Senni 201 consists

of two leafless, flowering twigs. The peduncles are

sparsely puberulous basally and the pods are sparsely

appressed-pubescent, rounded apically and up to

2,4 cm wide. As the syntypes examined are indis-

tinguishable from specimens of A. Senegal var.

Senegal , this opportunity is now taken of reducing

var. platyosprion to synonymy.

Acacia Senegal (L.) Willd. var. Senegal.

A. Senegal var. platyosprion Chiov., FI. Somala 2: 187, Fig.

114 (1932), synon. nov. Syntypes from the Somali Republic.

ACACIA SENEGAL (L.) WILLD. VAR. PSEUDOGLAU-

COPHYLLA CHIOV.

Chiovenda, in Stefanini-Paoli Miss. Somal. : 72

(1916), based his description of A. Senegal var.

pseudoglaucophylla on six syntypes from the Somali

Republic, namely, Paoli 83, 253, 682, 875, 883 and

1052. Examination of these syntypes revealed that

they almost certainly do not all belong to the same

taxon.

Paoli 83 has glabrous or subglabrous leaf-rhachides

and rhachillae, and up to 5 pinnae pairs per leaf.

The inflorescence axes are glabrous or subglabrous

and the pods are up to 8,3x2, 7 cm, puberulous and

rounded to subacute apically. Although it is a little

unusual in having glabrous or subglabrous inflore-

scence axes, this specimen is in all other respects

indistinguishable from material of var. Senegal and,

consequently, is referred to var. Senegal. It should be

mentioned that glabrous inflorescence axes do occur

in var. Senegal in other areas of distribution.

Paoli 253, which is sterile apart from a fragment of

a pod in the capsule mounted on the bottom left

corner, appears to be similar to Paoli 83 and therefore

it too is referred to var. Senegal.

Paoli 883 consists of two leafless, fruiting twigs and

a couple of detached pods. The inflorescence axes are

iglabrous and the pods are up to 7, 5x1,8 cm,

puberulous and rounded apically. Paoli 883 probably

belongs to the same taxon as Paoli 83 and 253 and is

also referred to var. senega!.

Paoli 1052 consists of two twigs, the right-hand one

of which is in flower and the left-hand one in young

fruit. The leaf-rhachides are pubescent and there are

up to 4 pinnae pairs per leaf. The inflorescence axes

are pubescent throughout or basally only and the

young pods are pubescent throughout and distinctly

acuminate or rostrate apically. On account of the

pods, and in the absence of any field notes, the

specimen is hesitantly and provisionally referred to

var. rostrata Brenan.

Paoli 875, which consists of three twigs, is similar

to Paoli 1052 in having pubescent leaf-rhachides and

inflorescence axes. The solitary young pod in the

capsule mounted in the lower left corner is pubescent

throughout and distinctly acuminate apically. Once

again, in the absence of field notes, the specimen is

hesitantly and provisionally referred to var. rostrata.

The varietal epithet pseudoglaucophylla has priority

over rostrata (and over A. oxyosprion var. pubescens)

but, as the identity of Paoli 875 and 1052 is not

beyond all doubt, it would seem unwise at this stage

to adopt the name pseudoglaucophylla for the speci-

mens with rostrate pods hitherto referred to var.

rostrata. Clearly, better and much more material

with adequate field notes is required from north-east

tropical Africa to establish the correct varietal epithet

for the plants with distinctly rostrate pods.

Paoli 682, the sixth syntype, is a miserable, sterile

specimen of which the precise identity remains in

doubt. The prickles near the nodes are solitary or in

threes, and the leaves have one or two pinnae pairs.

To sum up, three of the syntypes of var. pseudo-

glaucophylla are referred to var. Senegal , two are

hesitantly and provisionally referred to var. rostrata ,

and one cannot be identified.

ACACIA STEFANINI CHIOV.

Paoli 844 from the Somali Republic, the type

specimen of A. stefanini Chiov. in Ann. Bot., Roma

13: 395 (1915), consists of three leafless, flowering

twigs. The young branchlets are greyish- to reddish-

brown, subglabrous to very shortly puberulous, and

the paired, stipular spines are short and distinctly

hooked. The peduncles are short, up to 6 mm long,

fairly densely puberulous, and the involucel is at

the base of the peduncle, or a short distance above it.

The calyces are cupular and minute, up to 0,5 mm

long. The corollas are 2-2,5 mm long, that is, at least

four times as long as the calyces. The capsule in the

bottom right-hand corner contains a small twig,

flowers, and the remains of one rhachilla. This

rhachilla reveals that the leaflets are in 7 pairs, ±

2 X 0,5 mm and slightly puberulous.

Paoli 844 is a fairly good match of Hildebrandt 1394,

the isotype of A. reficiens Wawra subsp. misera (Vatke)

Brenan in the Kew herbarium. It agrees with

Hildebrandt 1394 in the colour and indumentum of the

young branchlets, the short, hooked spines, the short,

puberulous peduncles, the small, cupular calyces, and

the tubular corolla which is ± 4 times as long as the

calyx. Consequently, I have no hesitation in reducing

A. stefanini to synonymy under A. reficiens subsp.

misera.

Acacia reficiens Wawra subsp. misera {Vatke)

Brenan in Kew Bull. 12: 90 (1957). Type: Somali

Republic, Meid, Hildebrandt 1394 (B, holo.f; BM!;

K!).

A. stefanini Chiov. in Ann. Bot., Roma 13: 395 (1915),

synon. nov. Type from the Somali Republic.

Some years after describing A. stefanini, Chiovenda,

in FI. Somala 2: 197 (1932), referred the specimens

Senni 711 and 813 to this species and provided a

description of the pods. A pod of Senni 813 is

illustrated in Fig. 119. The pods were described as

turgid and cylindrical with strongly convex valves. As

the description of the pods and the illustration are at

variance with the pods of typical A. reficiens subsp.

misera, Senni 7 1 1 and 8 1 3 were borrowed on loan from

Florence.

J. H. ROSS

303

Senni 711 and 813, which are both flowering speci-

mens, are referable to A. reficiens subsp. misera.

However, the capsule mounted on the sheet of Senni

813 contains eight old pods. All of the pods have been

extensively galled, which accounts for their atypical

appearance and Chiovenda’s description.

ACACIA UNISPINOSA ( FIORl ) CHIOV.

When Fiori described A. asak (Forsk.) Willd. var.

unispinosa in L’Agric. Coloniale 5: 93, Fig. 67/3

(1911), he omitted to cite any specimens or to

nominate a type specimen, citing merely “Nel Samhar

alia stazione di Mai Atal e Uakiro (F)”. Under the

circumstances, all of the specimens in FI collected

prior to the date of publication and determined by him

as var. unispinosa, are regarded as syntypes.

Among the material received on loan from Florence,

were three specimens collected by Fiori and deter-

mined by him as ‘‘‘'Acacia asak W. var. unispinosa

Nob.”. On two of the specimens “form, unispinosa

Nob.” appears instead of var. unispinosa, but this is of

no consequence. The three specimens concerned are

Fiori 135b, 7 Feb. 1909, from Uakiro, Samhar region

(2 sheets) and Fiori 135b, 30 Mar. 1909, from stazione

di Mai Atal, Samhar region. Of these three syntypes

I now select the specimen of Fiori 135b, 7 Feb. 1909,

from Uakiro with flowers (and five pods plus leaflets

and inflorescences in the capsule mounted on the lower

right-hand corner) as the lectotype of A. asak var.

unispinosa.

The branchlets are greyish to purplish-grey or

greyish-brown, glabrous or sparingly pubescent on the

young extremities, and armed with a solitary recurved

prickle near each node. All leaves have one pinna pair,

sparingly puberulous petioles up to 7 mm long,

rhachillae up to 1,5 cm long, and 4-7 pairs of

glabrous glaucous leaflets per pinna which are up to

6 x 2,5 mm. Some of the leaflets are galled. The

inflorescence axes are sparingly pubescent throughout

or basally only and are up to 5 cm long. The calyces

are glabrous or very sparsely puberulous. The pods are

yellowish-brown, up to 4,8 x 1,1 cm, ± straight,

glabrous or very sparingly pubescent, especially on the

margins and near the stipe, venose and longitudinally

dehiscent.

Two other specimens, namely, Baldrati 4147, 4148,

were also determined by Fiori as var. unispinosa but,

as they were collected subsequent to the publication of

var. unispinosa, they are of lesser interest. They are

both appalling specimens and their chief interest must

surely be the numerous galls on the inflorescences and

leaflets. The prickles on each of the specimens are

either solitary or else in threes near the nodes.

It is clear that var. unispinosa is specifically distinct

from A. asak, as the latter has much larger leaves with

3-6 pinnae pairs, 6-17 pairs of leaflets per pinna and

usually larger pods. Var. unispinosa is also specifically

distinct from A. Senegal, differing chiefly in having

fewer pinnae pairs, fewer pairs of larger leaflets and

smaller pods. Chiovenda. FI. Somala 1: 169 (1929),

was therefore correct in giving var. unispinosa specific

rank.

When Chiovenda elevated var. unispinosa to A.

unispinosa he referred other specimens to this taxon.

Although not important as far as typification is con-

cerned, the identity of these specimens is nevertheless

of interest. Puccioni & Stefanini 236 [264] from Tra

Tigieglo e Bur Cal, Somali Republic, which consists

of a single flowering branchlet, is most probably A.

unispinosa.

Senni 822, however, clearly belongs to a different

taxon. The prickles are in threes near the nodes, or

solitary. The leaves have a spreading indumentum,

3-4 pinnae pairs, and each pinna has up to 12 pairs

of leaflets which have spreading marginal cilia. The

leaflets are of two different sizes and both occur on the

same twig. The smaller leaflets attain a size of 3 x 0,8

mm and the larger 5 x 2,25 mm. The inflorescence

axes are pubescent throughout and the pods are up to

4.7 x 1,4 cm, densely puberulous, and vary from

rounded to acute or rostrate apically. Senni 822 is

clearly referable to A. Senegal and, as the specimen was

collected from a small tree and some of the pod apices

are rostrate, it is hesitantly and provisionally referred

to var. rostrata Brenan.

Now that it is established that A. unispinosa is

specifically distinct from A. asak and from A. Senegal,

it remains to discuss its relationship to A. oliveri

Vatke. A. oliveri Vatke in Oest. Bot. Zeit. 30: 274

(1880), was based on Hildebrandt 729c from the

Danakil in Ethiopia. Despite repeated attempts I have

not succeeded in tracing a type specimen. However,

the description of A. oliveri is fairly detailed and agrees

well with the material of A. unispinosa except for one

important discrepancy, namely, the pods of A. oliveri

were described as "legumen adultum 1 ,2 dm longum,

ad 3 cm latum”. The pods of A. unispinosa are much

smaller and only attain a size of ±4,8 x 1,1 cm.

There is, however, some significant evidence which

suggests that the pod described by Vatke did not in

fact belong to the specimen of Hildebrandt 729c.

This evidence is in the form of a sketch in the British

Museum of the Berlin specimen of Hildebrandt

729c, drawn during E. G. Baker’s visit to the Berlin

Herbarium in 1926. The life-size drawing, which

consists of a vegetative shoot, one inflorescence and

three pods, carries the comment “pods 2-3-seeded,

flat”. The pods are up to 3,5 cm long and 1,1 cm

wide and these dimensions are in keeping with the pods

of A. unispinosa. On the strength of the drawing of the

Berlin specimen of Hildebrandt 729c in the British

Museum, it seems reasonable to assume that the adult

pod described by Vatke did not in fact belong to

Hildebrandt 729c and that it inadvertently came from

some other source.

Vatke recorded the colloquial name of A. oliveri as

“Tikible”. On Bally 7044 (a fairly good match of

Fiori 135b, the lectotype of A. asak var. unispinosa)

in the Kew Herbarium, from 15 miles south of

Massawa in Ethiopia, the colloquial name is recorded

as “Tikkiville”. Although spelt somewhat differently,

the two names are phonetically similar and the

differences in spelling could perhaps be partially

explained by the lapse of some seventy years between

the time when Hildebrandt 729c and Bally 7044 were

collected.

As the material at present referred to A. unispinosa

matches the drawing of the Berlin specimen of Hilde-

brandt 729c in the British Museum, agrees with Vatke’s

description of A. oliveri (with the exception of the

mature pod), and has a similar colloquial name to the

one recorded by Vatke, I am reasonably satisfied that

A. unispinosa and A. oliveri are one and the same taxon.

A. oliveri, being the earlier name, must be adopted for

this taxon and A. unispinosa is now reduced to

synonymy.

Acacia oliveri Vatke in Oest. Bot. Zeit. 30: 274

(1880) excl. descr. leguminis. Type: Ethiopia, the

Danakil, Hildebrandt 729c (BM, drawing!).

A. unispinosa (Fiori) Chiov., FI. Somala 1 : 169 (1929), synon.

nov. Type: Ethiopia, Samhar, Uakiro, 7 Feb. 1909, Fiori 135b

(FI, lectol).

Bothalia 11,3: 305-308 (1974)

Southern African grasses with foliage that revives after dehydration

D. F. GAFF* and R. P. ELLISt

ABSTRACT

A brief survey in Southern Africa revealed 1 1 grass species and a number of sedges with foliage

that recovers from air-drying (equivalent to equilibrium with air at 20-40 % relative humidity at 28°

C.). The desiccation tolerance limits were extremely low being equivalent to approximately 0-5 %

relative humidity. Some species may have a potential use in agriculture.

INTRODUCTION

Until recently only about nine angiosperm species

were considered to have foliage that would tolerate

dehydration to the point of air-dryness in a manner

similar to lichens on exposed rock surfaces. It has since

become clear that this phenomenon is of wider

occurrence among the angiosperms than previously

thought, particularly in southern Africa where 11

new angiosperm examples were recently reported

(Gaff 1971). The fact that some of these were grasses,

raised the possibility that some “resurrection” plants

might be found that had a potential use in agriculture

in arid areas. Consequently, a search was made for

further desiccation-tolerant grasses.

METHOD

Grasses were marked while in flower during the

moist summer season. Identical voucher specimens

were collected and deposited for identification at the

National Herbarium in Pretoria. Identifications were

also made by the herbaria at Salisbury, Windhoek,

Tananarive, and at the Sugar Research Institute of

Mauritius, where appropriate. The marked plants

were re-examined towards the end of the dry season

which, in the regions traversed, extends from about

April to September. Wherever possible, plants were

collected in a completely air-dry condition — in many

cases from shallow soil pans on otherwise bare rock

slopes fully exposed to the sun, i.e. extremely xeric

sites. If the marked plant had not dried completely,

the clump was removed from the soil, placed in an

open paper bag, and allowed to dry in the prevailing

conditions of low air humidity and moderately high

air temperatures.

The air-dry plants were submerged in water to

rehydrate for 24 hours in diffuse laboratory light.

Plants which became crisp in texutre and healthy in

appearance were subjected to various tests to establish

survival of the leaves (neutral red uptake, Sullivan

and Levitt 1959; evans blue exclusion, Gaff and

Okong’o-Ogola 1961; and formation of chlorophyll

in nonchlorophyllous tissue).

Water stress levels were assessed as: (a) Relative

water content, i.e. the water content of the tissue as a

percentage of its water content at full turgor. A 70 °C

oven was used to obtain dry mass since higher

temperatures caused errors due to decomposition

(Paech and Tracey 1955; Loomis and Shull 1937).

Twenty-four-hour immersion in water was usually

required for attainment of full turgor, (b) Water

potential, determined by the gravimetric vapour-

exchange method (Slatyer 1958). Errors due to respira-

tory loss of C02 (Kreeb and Onal 1961) were assumed

to be negligible in the dry leaves. Drought tolerance

levels were determined by allowing the air-dry foliage

to equilibrate to air of various relative humidities,

* Botany Department, Monash University, Clayton, Victoria

3168, Australia.

t Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

which were maintained constant by saturated salt

solutions in contact with the solid salt at 28° C. Con-

centrated sulphuric acid was used for relative humidi-

ties of approximately zero, and various dilutions of this

were used for intermediate humidities to 10%. Rela-

tive humidity values were based on the data in Robin-

son and Stokes 1955. The course of equilibration was

followed gravimetrically until consecutive masses

were constant.

The regions traversed were Rhodesia (Salisbury,

Rusape, Matopos, Fort Victoria, Lundi), Transvaal

(Messina to Pretoria and the Magaliesberg), the

northern Cape Province (Kimberley west to Port

Nolloth) and South West Africa (Karasburg to

Okahandja, Swakopmund, Maltahohe), the Malagasy

Republic (Tananarive to Ambalavao), and Mauritius.

These regions have predominantly summer rainfull

with the exception of the Port Nolloth-Karasburg area

where rainfall is sporadic rather than seasonal.

Unfortunately, it was not possible to examine the

winter rainfull areas of the Southern Cape and the

succulent karoo.

RESULTS AND DISCUSSION

Air-dry foliage from 1 1 of the 80 grass species

examined revived on rehydration. The water potentials

of the air-dry foliage were mainly in the 30-40% RH

range (Table 1), i.e. approximately the same range of

values as the field-dry “resurrection” plants reported

earlier (Gaff 1971). The drought tolerance limits

(Table 2) were considerably lower than the water

stress levels; most species had tolerance limits of

approximately 0% RH. Again this is consistent with

the earlier study, even though the times allowed here

to ensure full equilibration were some months longer.

Some of the remaining species examined, parti-

cularly Aristida spp., gave a superficial appearance of

revival in the young tiller shoots. However, the

healthy green appearance deteriorated after a further

day of full hydration, and the crispness of the tissue

could be attributed to a high fibre content rather

than to restoration of cell turgor. The tests with

evans blue and neutral red confirmed that these

tissues had not survived. Most of the species listed in

the appendix grew in extremely xeric sites, and

appeared superficially to be dry and dead during the

dry season. However, closer examination usually

revealed a few well-hydrated immature tillers persist-

ing in the centre of the otherwise defunct clump. The

possibility that some of these species may yet prove

desiccation tolerant cannot be ruled out entirely,

since the root system of the plant was often extensively

damaged as the plant was removed from the soil,

and artificial dehydration may have been more rapid

than desiccation under field conditions.

Desiccation-tolerant grasses appear to be restricted

to the subfamily Eragrostoideae sensu Tateoka

(1957). Even here they occur in three tribes which

Hutchinson, 1959, considers are related, viz the

Eragrostideae ( Eragrostis and Tripogon), Sporoboleae

306

SOUTHERN AFRICAN GRASSES WITH FOLIAGE THAT REVIVES AFTER DEHYDRATION

(. Sporobolus ) and the Chlorideae ( Oropetium , Bra-

chyachne, Microchloa). All these tribes have the

eragrostoid type of mesophyll and parenchymatous

bundle-sheath arrangement (Carolin, Jacobs & Vesk,

1973) and are Kranz species (Brown and Smith

1972), which presumably exhibit the C4 photosyn-

thetic pathways. It would seem that these groups have

some predisposition toward desiccation tolerance

that is realized in only a small number of genera,

some of which possess a predominance of tolerant

species (e.g. Microchloa *), others only a minor pro-

portion (e.g. Eragrosti s, Sporobolus). This pattern

suggests a relatively recent appearance of the

adaptation during the evolution of the grass tribes.

There is evidence that young, meristematic tissue is

more drought tolerant than mature, differentiated

tissue on the same plant (reviewed in Levitt 1972).

It seems likely that the basal-meristematic pattern of

growth in the grass and sedge leaves might allow a

gradual extension of extreme drought tolerance

developed in the basal meristem to progressively

older areas of the leaf. That is, species in which

desiccation tolerance is confined to the basal centi-

metre of the leaf (e.g. Eragrostis hispida, Cyperus bellus)

or the immature leaves (e.g. Sporobolus lampranthus),

are in an earlier stage of evolutionary adaptation than

species in which the full length of the mature foliage is

tolerant (e.g. Eragrostis nindensis, Coleochloa setifera ,

etc.). Most grasses with tolerant mature leaves retain

the chlorophyll in the dry state, and are fully green

after imbibition is complete (24 hrs or less). Eragrostis

nindensis and the two species of Coleochloa lose their

chlorophyll as they dehydrate, consequently a further

24 hrs is required for regreening after full imbibition.

As this would seem a somewhat less efficient level

of ecological adaptation, it is tempting to assume that

this also is an intermediate evolutionary stage.

However, markedly different response patterns in the

fine structure of drying cells in the two types (chloro-

phyll-retainers as opposed to chlorophyll-losers),

suggest caution is necessary in accepting this assump-

tion, (unpublished data. Gaff, N. Hallam, S. Zee and

T. P. O’Brien).

In general the desiccation-tolerant grasses and sedges

appear to be pioneer plants colonizing shallow soil

pans. Whereas the two sedges Coleochloa setifera and

C. pallidior commonly grow to 30-50 cm, the grasses

are usually low-growing. To a certain extent this is a

consequence of the shallowness of the soil (often 1-2

cm deep) and the severity of their habitat; all the

grasses show a significant increase in height when

grown in pots regularly supplied with a complete

mineral nutrient in a greenhouse at 25 °C. In most

cases, the increase would be 50-100%, but a threefold

increase from about 6 cm to 20 cm occurs in Micro-

chloa cajfra. Their low growth habit probably places

them at a disadvantage when competing with taller

species in moister deep soils. Nevertheless, Eragrostis

nindensis, and Sporobolus lampranthus commonly

grow as broad tussocks, obviously of considerable age,

interspersed between taller grass species on deeper

soil flats.

* It seems that Microchloa indica ( 1.1.) Beauv. is a marginal

case. Plants grown from seed collected in Rhodesia, showed

survival of a significant proportion of cells in the young leaves

following artificial drying of potted plants. On the other hand,

field-dry plants collected in Northern Australia were dead;

however the species is clearly annual and these plants had

probably senesced prior to dehydration. We were unable to

examine Microchloa ensifotia Rendle, but from the description

of the plant and its habitat in Angola (Launert 1966), it is

highly probable that this, too, is a resurrection grass,

The low height of the grasses is in part traceable to

their common growth habit; all are perennials which

tiller repeatedly to form a dense, compact tussock;

very little elongation of the internodes occurs except in

the inflorescence. None are stoloniferous or rhizo-

matous. Possibly the vascular structure of grasses, i.e.

largely two metaxylem vessels and a protoxylem canal

arranged in widely-separated bundles rather than

in a firm, compact mass of tracheids, is more liable to

damage as the dehydrating plant shrinks. Con-

sequently re-establishment of water continuity in the

xylem from the root to the transpiring leaf on rehy-

dration would be impaired (c.f. Gaff 1972). This factor

might have restricted the evolution of desiccation

tolerance in grasses which have extensive development

of the internodes.

The taller resurrection grasses may be useful

components of pastures subject to prolonged drought.

Although there has been no systematic assessment

of how long the grasses can endure desiccation,

dry material of one resurrection plant, Borya nitida

Labill., has survived 2\ years to date. There seems

to be no reason why the grasses should not be equally

durable. Rehydration is rapid in all species; most

grasses expand to their original dimensions within

4-5 hours and are fully turgid well within the 24-hour

period allowed here. This means that green feed is

available within a day of a rainfall of 10 mm or more.

A further day is required by E. nindensis for regreening,

since this species loses its chlorophyll on dehydration.

Apart from any fodder value, their persistence over

long periods without rain should minimize wind

erosion.

At present, E. nindensis would appear to have the

most potential for introduction to semi-arid pastures

beyond its present distribution. Although of only

moderate height (approximately 12 cm.), its value as a

pasture plant has been commented on by Walter and

Volk (1954) who drew attention to its rapid resumption

of growth after rain, but who do not appear to have

realized the desiccation-tolerant nature of the foliage.

Their analyses revealed a generally high content of

protein and phosphate, which they considered con-

firmed the good reputation of the species in S. W.

Africa. E. nindensis tussocks possess a strong, firm

root system, which resists wrenching of the plant from

the soil by grazing animals. Although the Coleochloa

species are taller-growing, there was no evidence of

grazing. Presumably a high fibre content of the leaves

renders them unpalatable or possibly the rock sur-

faces on which they grow may not readily be grazed

by domestic animals.

ACKNOWLEDGEMENTS

This investigation would not have been possible

without the generous help received from many

quarters. As it is not possible to thank everyone in-

dividually, we hope that the staff of the institutes

mentioned below will accept our thanks as being in-

tended personally for each member of the staff who

has assisted.

The authors are indebted to: The Anglo-American

Corporation of South Africa and Mr H. F. Oppen-

heimer; The Australian Research Grants Commission,

D66/ 1 6 1 37 ; The Botanical Research I nstitute, Pretoria ;

BP South Africa; Mr R. R. Cohen; The Departments

of General Botany and Plant Physiology at the

University of Pretoria; Mrs I. R. Dick; Direction des

Eaux et Forets et de la Conservation des Sols, Mala-

gasy Republic; Dr D. Friend; Mrs L. M. Padevetova;

Mr A, Razafindratsira; Rural Credit Development

D. F. GAFF AND R. P. ELLIS

307

Fund, Reserve Bank of Australia; The Salisbury Her-

barium; The South African Council for Scientific and

Industrial Research; The Sugar Research Institute,

Mauritius; The Tananarive Herbarium (ORSTOM);

The Veterinary Diagnostic Laboratories, Windhoek;

The Water Research Foundation of Australia;

The Windhoek Herbarium.

The first author wishes to thank Mr & Mrs J.

Vahrmeijer, Dr & Mrs B. de Winter, and Mr & Mrs N.

Graham for the kind hospitality they extended to him.

OPSOMMING

'n Beperkte ondersoek in Suidelike Afrika het 1 1 grassoorte

en ’n aantal biesies opgelewer, met blare wat herstel van lug-

uitdroging (gelykstaande aan 'n ewewig met lug van 20-40%

relatiewe vogtigheid by 28° C). Die grense van die uitdrogings-

verdraagsaamheid was besonder laag, naamlik ongeveer 0-5 %

relatiewe vogtigheid. Sommige van die soorte het dalk moont-

likhede vir gebruik in die landbou.

REFERENCES

Brown, W. V. and Smith, B. N., 1972. Grass evolution, the

Kranz syndrome, 13C/12C ratios and Continental Drift.

Nature 239 : 345-346.

Carolin, R. C., Jacobs, S. W. L. and Vesk, M., 1973. The

structure of the cells of the mesophyll and parenchymatous

bundle sheath of the Gramineae. Bot. J. Linn. Soe. 66: 259-

275.

Gaff, D. F., 1971. Desiccation-tolerant flowering plants in

southern Africa. Science 174: 1033-1034.

Gaff, D. F., 1972. Drought resistance in Welwitschia mirabilis

Hooker fil. Dinteria 7: 3-8.

Gaff, D. F. and Okong’o-Ogola, O, 1971. The use of non-

permeating pigments for testing the survival of cells.

J. Exp. Bot. 22:342-346.

Hutchinson, J., 1959. The families and genera of flowering

plants. Vol. II Monocotyledons. London: Oxford Uni-

versity Press 2nd. Ed.

Kreeb, K. and Onal, M., 1961. Uber die gravimetrische

Methode zur Bestimmung der Saugspannung und das

Problem des negativen Turgors. Planta 56: 409-416.

Launert, E., 1966. A taxonomic revision of the genus Micro-

chloa R. Br. Senckenberg Biol. 47 : 291-301 .

Leith, H. and Ashton, D. H., 1961. The light compensation

points of some herbaceous plants inside and outside a

deciduous wood in Germany. Can. J. Bot. 39: 1255-1259.

Levitt, J., 1972. Responses of plants to environmental stresses.

New York & London : Academic Press.

Loomis, W. E. and Shull, C. A., 1937. Methods in plant

physiology. New York.

Paech, K. and Tracey, M. V., 1955. Modern methods of plant

analysis. Vol. I. Berlin: Springer Verlag.

Robinson, R. A. E. and Stokes, R. H., 1955. Electrolyte

Solutions. London: Butterworths.

Slatyer, R. O., 1958. Measurement of DPD in plants by a

method of vapour equilibration. Aust. J. Biol. Sci. 1 1 :349-

365.

Sullivan, C. Y. and Levitt, J., 1959. Drought tolerance and

avoidance in two species of oak. Plant Physiol. 1 2 : 299-305.

Tateoka, T., 1957. Proposition of a new phylogenetic system for

the Poacea e.J.Jap. Bot. 32: 275-287.

Walter, H. and Volk, O. H., 1954. Grundlagen der Weide-

wirtschaft in Siidwestafrika. Stuttgart: Eugen Ulmer Verlag.

Table 1. — Species with foliage that survives air-drying; and the water stress in their leaves immediately prior to rehydration

for the survival tests.

misidentified as Ficinia filiformis Schrad. in Gaff 1971 .

308

SOUTHERN AFRICAN GRASSES WITH FOLIAGE THAT REVIVES AFTER DEHYDRATION

Table 2. — Drought Tolerance of foliage. The minimum relative humidity (at 28 °C.) which allows survival of 50% or more of

the foliage equilibrated to it.

Species

GRASSES

Brachyacfme patentiflora

Eragrostis hispida leaf bases

E. nindensis

E. paradoxa leaf bases

Microchloa cajfra

M. kunthii

Oropetium capense

Sporobolus festivus

S. lampranthus leaf bases

S. stapfianus

Tripogon minimus

SEDGES

Coleochloa setifera

C. pallidior

Cyperus bellus

Mari sc us sp

leaf base

APPENDIX

Species examined in which the air-dry foliage failed to

survive. Collector’s (R. P. Ellis) numbers are given together with

abbreviated collection localities: MAUR — Mauritius MLG —

Malgasy Republic NC — Northern Cape Province RH —

Rhodesia SWA — South West Africa TVL — Transvaal

Anthephora argentea Goosens, 877, NC.

A. pubescens Nees, 927, SWA.

Aristida adscensionis L., 800, 845, 864, NC & RH.

A. aequiglumis Hack., 1036, TVL.

A. congesta Roem. & Schult. subsp. barbicollis (Trin. & Rupr.)

De Wint., 841, NC.

A. decaryana A. Camus, MLG.

A. effusa Henr., 1064, SWA.

A. engleri Mez, 879, NC.

A. meridionalis Henr., 865, NC.

A. spectabilis Hack., 833, TVL.

A. transvaalensis Henr., 1027, TVL.

A. vestita Thunb., 844, NC.

Asthenantherum glaucum (Nees) Nevski, 872, NC.

Brachiaria nigropedata (Munro ex Fical. & Hiern) Stapl, 834,

TVL.

Bulbostylis burchellii C. B., 1038, TVL.

B. oritrephes (Ridl.) C. E. Hubb., TVL.

B. boekleriana (Schweinf.) Beetle, TVL.

Craspedorachis rhodesiana Rendle, 817, RH.

Cymbopogon excavatus (Hochst.) Stapf ex Burtt-Davy, TVL.

C. marginat us (Steud.) Stapf ex Burtt-Davy, TVL.

Cyperus exilis Willd. ex Kunth., MAUR.

C. obtusiflorus Vahl, TVL.

C. smithii McClean, TVL.

Dolichochaetae biseriala (Stapf) Phipps, TVL.

Danthoniopsis ramosa (Stapf) W. D. Clayton, 1054, SWA.

Digitaria monadactyla (Nees) Stapf, TVL.

Diheleropogon amplectens (Nees) Clayton, TVL.

Diplachne biflora Hack., TVL.

Enneupogon desvauxii Beauv., 856, 842, NC.

E. pretoriensis Stent, 835, TVL.

E. scaber Lehm., 861, NC.

Eragrostis annulata Rendle ex S. Elliot, 921, SWA.

E. brizantha Nees, 910, SWA.

E. capensis (Thunb.) Trin., NC.

E. chapelieri (Kunth )Nees, 816, RH.

E. curvula (Schrad.) Nees, 860, NC.

E. lappula Nees, 815, RH.

E. lehmanniana Nees, 840, 867, NC.

E. patens Oliver, 808, RH.

E. porosa Nees, 871, 923, NC, SWA.

E. pseudo-obtusa de Wint., 839, NC.

E. racemosa (Thunb.) Steud., 1023, TVL.

E. rigidior Pilg., 1052, RH, TVL.

E. stapfii de Wint., 1037, TVL.

E. tenuifolia Hochst., UGANDA.

E. tenel!a( L.) Beauv. ex Roem. & Schult., MAUR.

Fimbristylis exilis Roem. & Schult. ,SW A.

F. glomerulata Nees, MAUR.

F. obtusifolia Kunth, MAUR.

Heteropogon contortus (L.) Beauv. ex Roem. & Schult., 849, NC.

Loudetia flavida (Stapf) C. E. Hubb., 798, RH.

L. ramosa (Stapf) C. E. Hubb., TVL.

L. simplex (Nees) C. E. Hubb., TVL.

Monelytrum luderitzianum Hack, emend Schweick., SWA.

Panicum arbusculum Mez, 878, NC.

Rhynchelytrum brevipilum (Hack.) Chiov., 1056, SWA.

R. repens (Willd.) C. E. Hubb., TVL.

R. setifolium (Stapf) Chiov., 1055, SWA.

Schmidtia kalihariensis Stent, 862, NC.

S. pappophoroides Steud., 837, NC.

Setaria appendiculata (Hack.) Stapf, 905, SWA.

5. lindenbergiana (Nees) Stapf, 1029, TVL.

S. perennis Hack., TVL.

Sporobolus coromandelianus (Retz.) Kunth, 912, SWA.

5. fimbriatus Nees var. latifolius Stent, 914, SWA.

5. pectinatus Hack., 1076, TVL.

S. pyramidalis Beauv., 846, NC.

Stipagrostis ciliata (Desf.) de Wint., 863, NC.

S. hirtigluma (Steud.) de Wint., 908, SWA.

5. hochstetterana (Boeck. ex Hack.) de Wint., 909, SWA.

S', namaquensis (Nees) de Wint., 881, SWA.

S. obtusa (Del.) Nees, 854, NC.

S. uniplumis (Licht.) de Wint. var. neesii (Trin. & Rupr.) de

Wint., SWA.

Sutera canescens (Benth.) Hiern., 1041, NC.

Themeda triandra Forsk., TVL.

Tricholaena monocline (Trin.) Stapf et C. E. Hubb., TVL.

Trichoneura grandiglumis (Nees) Ekman, TVL.

Urochloa panicoides Beauv., 916, TVL.

Zoysia tenuifolia Willd., MAUR.

Bothalia 11, 3: 309-323 (1974)

On concepts and techniques applied in the Ziirich-Montpellier

method of vegetation survey

M. J. A. WERGER *

ABSTRACT

Because a wider use of the Zurich-Montpellier method in the study of vegetation in South Africa

is envisaged, an outline of concepts and techniques applied in that method is given.

TABLE OF CONTENTS

page

Introduction 309

1 . Sampling 310

1 . 1 Site selection 310

1 .2 Representativeness 310

1 . 3 Homogeneity 310

1.4 Minimal area and plot size; plot form 311

1.5 Structure 312

1 .6 Floristic lists 313

1 .7 Cover-abundance 313

1 . 8 Sociability 314

1 . 9 Other analytic characters of the vegeta-

tion 314

1 . 10 Habitat characteristics 314

1.11 Relev e 315

1 . 12 Number of releves 315

2. Synthesis 315

2. 1 Tabulation 315

2.2 Synthetic characters, character species,

differential species 316

2.3 Ecological confirmation of the table

pattern 317

2.4 The association and other syntaxa; the

constancy table 317

2.5 Nomenclature 318

2.6 Homotoneity; affinities between syntaxa 318

2.7 Community structure 319

Acknowledgements 320

Opsomming 320

References 320

INTRODUCTION

The science of plant ecology has been practised in

South Africa since the beginning of the present

century. Of the early ecological studies in South

Africa, the comprehensive treatise on the vegetation

of the Cape Province by Marloth (1908) is perhaps

the most significant. For more than half a century

thereafter South African ecology was predominantly

inspired bv the writings of Clements (e.g. Bews,

1918; Phillips, 1931; Bayer, 1955; Kiilick, 1963)

and Tansley (e.g. Adamson, 1938: Story, 1952;

Edwards, 1967). The studies of this period, discussed

in more detail by Kiilick (1967), were mainly of a

non-formal, descriptive nature and often admirably

accurate and informative. During the early fifties two

ecological studies were of considerable local signi-

ficance. In 1953 Acocks published his “Veld Types of

South Africa”, a classification of South African

vegetation into 70 veld types and 75 variations, based

on a floristic comparison of stand data. Acocks

selected well-developed, extensive patches of vege-

tation and recorded all species encountered. Abun-

dance of each species in each stand was estimated and

from comparison of these lists the veld types were

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria; now at:

Botanisch Laboratorium, Afd. Geobotanie, Toernooi-

veld, Nijmegen, Netherlands.

extracted. Since its publication, Acocks’ Veld Types

has formed the basis of the majority of ecological

studies carried out in South Africa. In 1955 Tidmarsh

& Havenga published their statistical wheelpoint

method for surveying and measuring the cover of

vegetation, which has since been used extensively in

Southern Africa in its original as well as in modified

forms.

The development of statistical techniques in ecology

received attention in South Africa soon after the intro-

duction of these techniques in Europe and America

and, mainly from 1960 onwards, several studies were

undertaken using Goodall’s (1953) interspecific cor-

relation analysis (Van Vuuren, 1961), association

analysis (e.g. Grunow, 1965; Roberts, 1966; Scheepers

1969; Taylor, 1969; Coetzee, 1972), information ana-

lysis (Grunow & Lance, 1969), hierarchical syndrome

analysis (Coetzee & Werger, 1973) and various kinds

of ordination techniques (e.g. Woods & Moll, 1967;

Morris, 1969; Moll, 1969; Grunow & Morris, 1969;

Louw, 1970).

The Zurich-Montpellier or Braun-Blanquet ap-

proach to the study of vegetation has, since its

origin in Europe shortly after the turn of the century,

proved to be an efficient and reliable method for

vegetation survey and classification in most countries

of Europe and also in other continents (Whittaker,

1962; Braun-Blanquet, 1968; Tiixen, 1969, 1970a;

Doing, 1970). However, until 1969 this approach

remained virtually unknown in Africa south of the

equator, except for work in Zaire, Rwanda and

Burundi by Belgian phytosociologists, a vegetation

survey in southern Mocambique by Myre (1960, 1962,

1964), and a small survey in the Windhoek area of

South West Africa by Volk & Leippert (1971) (cf.

Werger, 1973a; Werger et al. , 1972). Language

difficulties, amongst others, were suggested as a main

reason for this remarkable omission by Werger et al.

(1972). But basic differences between the scientific

approaches of Anglo-American and continental Euro-

pean ecologists should also not be underestimated.

Although English accounts of several aspects of the

method were given by Braun-Blanquet as early as

1932, and later by Poore (1955, 1956), Becking (1957),

Moore (1962) and later authors, lack of training as

well as the local psychological climate that was by then

so influenced by the attitude of general non-acceptance

of the method, prevailing amongst the English-

speaking ecologists, prevented its application until

1969. In 1969 a phytosociological survey of the

Upper Orange River valley, in which the Zurich-

Montpellier method was applied, was initiated (Wer-

ger, 1973a) and since then a number of smaller

surveys, in which this method was successfully applied,

were carried out (Van Zinderen Bakker. 1971;

Werger eta!., 1972; Werger, 1973b; Leistner & Werger

1973; Van Zinderen Bakker & Werger, 1973; Coetzee.

1 974 ; Coetzee & Werger, 1 974). Because of the success

of these studies a wider acceptance of the Zurich-

310

ZURICH-MONTPELL1ER METHOD OF VEGETATION SURVEY

Montpellier method in South Africa is envisaged. A

discussion of the concepts and techniques applied in

this method, particularly of those aspects not covered

by the previously-mentioned as well as the more recent

publications in English (Kiichler, 1971; Shimwell,

1971), was therefore thought to be useful. In this

presentation no pretence to completeness is made.

According to the Ziirich-Montpellier method,

selected, representative, homogeneous plots of a

certain minimum size are sampled in the phytocoe-

noses (stands) making up the vegetation of the area

to be surveyed, recording all species and rating them

on a cover-abundance and, optionally, a sociability

scale. Some other analytical characters of the vege-

tation in the plot might also be recorded. The samples

are entered in a table from which the vegetation

units are extracted. The units are interpreted ecologi-

cally and ranked in a hierarchy. Thus, the method

consists of an analytical, sampling phase and a

synthetic phase, which will be discussed separately.

7. SAMPLING

1 . 1 Site selection

In the Ziirich-Montpellier School the selection

of a site for a plot is generally carried out subjectively.

Although much and severely criticized, this procedure

is in accordance with the community-unit theory,

which postulates that vegetation consists of natural

entities usually in contact with each other along

narrow boundaries (cf. Werger, 1973a; Whittaker,

1956, 1962, 1967). The most efficient way to find the

units or associations, which are types abstracted from

actually-occurring stands or phytocoenoses, is to

select for sampling those stands which may possibly

be examples of such an association. Stands which are

obviously heterogeneous in habitat, structure or

floristic composition, and might therefore logically be

expected to represent parts of two or more units or

associations, should be avoided because they do not

contribute information which can be used to describe

the two or more community-types that they represent.

Thus, the subjective decision to avoid mixed stands

does not imply the circularity suggested by Whittaker

(1956): “associations are being studied in terms of

samples taken in terms of associations”. The real

procedure is that associations are being studied in

terms of plot samples taken in terms of phytocoenoses

which are possible association-individuals. It is

obvious that the question of subjective selection of

phytoceonoses for sampling is closely related to the

question of representativeness, homogeneity and total

number of plots to be sampled. The area of investi-

gation must be well known in all its variety before

the study is started, since a good knowledge of the

variation in the study area allows a more representative

sampling of the various stands (phytocoenoses). Pre-

judice about the expected results of the study should

be avoided in this selection. The subjective selection

of stands for sampling, guarantees an optimal samp-

ling efficiency, because obviously heterogeneous

plots are avoided as far as possible (Ellenberg, 1956;

Becking, 1957; Dahl 1957; Barkman, 1958; Gounot,

1961, 1969; Braun-Blanquet, 1964; Daubenmire, 1968;

Moore et a/., 1970; Knapp, 1971 ). The strength of the

arguments against the procedure of subjective

selection of plot sites is often overestimated, since it

seems to be difficult, according to the experience of

Ivimey-Cook & Proctor (1966) to “collect any

substantial body of phytosociological data to support

a conclusion seriously at variance with the facts”.

On the other hand, this question need not be crucial

in the decision to apply the Ziirich-Montpellier

method, since there is no fundamental objection

against stratified random, random and systematic

sampling. These sampling strategies can also yield

good results. Particularly in the last two cases,

however, a considerable amount of plot data will be

too heterogeneous to be used in the extraction of

community-types (cf. Taylor, 1969). A test for homo-

geneity of the plots might prove useful when these

strategies are used. When they are used in an area

where different communities occur in a mosaic of

patches of small dimensions, the possibility exists that

virtually no plot may be usable, or that so many plots

are situated on similar transitions that these transitions

are also extracted from the tables and can wrongly be

interpreted as types. Great care should thus be taken in

evaluating the plot data and the synthetic results

based on random and systematic sampling strategies.

1 . 2 Representativeness

Stands for sampling should be selected in such a

manner that each is representative of the vegetation

of which it is part and that each plot sampled therein

should yield a more or less typical description of

that vegetation in terms of both floristic composition

and structure. Each plot should represent only one

entity of vegetation (Ellenberg, 1956; Dahl, 1957;

Gounot, 1961; Braun-Blanquet, 1964; Knapp, 1971).

In an open tree or shrub vegetation, for example,

the situation of plots should be such that the relative

importance of the tree or shrub component within the

plot is comparable with that of the surrounding

vegetation to be represented.

The requirement of representativeness of the

vegetation in the sample plot should not be confused

with the question of the degree of development or

disturbance of the vegetation. Although samples taken

in optimally developed vegetation are most suitable

for the first description of a new syntaxon (Westhoff,

1967, 1968; Westhoff & Den Held, 1969), it will often

be useful to sample stands representative of below-op-

timally developed vegetation, such as overgrazed

grassland. Such a sample is important in the establish-

ing of the status of the grassland concerned and of the

successional trends under the influence of a specific

treatment of the vegetation to which the sample

belongs. When random or systematic sampling is

carried out, the question of the representativeness of

each individual plot is, of course, by-passed. This

disadvantage is particularly important in open vege-

vegetation.

The representativeness of a plot is closely related to

its homogeneity.

1 . 3 Homogeneity

According to Goodall (1952), “homogeneity has

bulked large in ecological literature, because most

schools of plant sociology specify it as a first desidera-

tum for a ‘stand’ or area of vegetation, which can

serve as a unit of classification”.

The approaches based on the community-unit

theory and those based on the individualistic hypo-

theses both require homogeneous sample plots (Bray

& Curtis, 1957; Curtis, 1959; Grunow & Morris,

1969). Workers using the various ordination techni-

ques often assess only floristic homogeneity of the

sample plot, relating homogeneity to the distribution

of the species present in the sample plot. In the Ziirich-

Montpellier School floristic, structural and environ-

mental homogeneity are usually assessed visually.

Homogeneity is important in the community-unit

theory approaches because information per sample

of one vegetation entity only, and not of mixtures, is

wanted (Ellenberg, 1956; Dahl, 1957; Barkman,

1958; Gounot, 1961, 1969; Braun-Blanquet, 1964;

Daubenmire, 1968; Tlixen, 1970b; Knapp, 1971;

M. J. A. WERGER

311

Shimwell, 1971). Most techniques for testing homo-

geneity of the sample plot start tacitly from one of two

assumptions; either that if the plot is floristically

homogeneous it is also structurally and environ-

mentally homogeneous and thus only floristic homo-

geneity needs to be tested, or that structural and

environmental heterogeneity not expressed in the

floristic component is irrelevant.

Homogeneity was first discussed by Nordhagen

(1923) and soon became the object of a polemic

between Kylin and Romell (cf. Kylin, 1923, 1926;

Romell, 1925, 1926). Kylin, assuming that, on the

average, the species in a community are distributed

randomly, regarded homogeneity as depending on

variations in density among the species and suggested

that species-area curves and frequency-distribution

curves are useful tools for determining the homo-

geneity of a patch of vegetation. Kylin (1926) dis-

tinguished homogeneity of vegetation based on the

distances between the individuals of a species, which

is homogeneity in the sense of Nordhagen (1923),

from homogeneity of vegetation based on distances

between all the various species, which deals with the

composition of the vegetation. Romell (1925, 1926),

who also distinguished homogeneity of distribution

from homogeneity of composition, pointed out that,

with respect to the former, homogeneity is a matter

of scale. Although a patch of vegetation might

be heterogeneous if studied on a small scale,

as a consequence of the fact that plant-indivi-

duals occur as discrete units, the same area

might be homogeneous when studied on a larger scale.

This matter of scale also applies to plants occurring in

clumps (compare also Goodall, 1954a, 1961, 1970;

Van der Maarel, 1966a). Since plant individuals are

more often under- or overdispersed (cf. Ashby,

1948; Goodall, 1952, 1970; Augarde, 1957; Greig-

Smith, 1964; Kershaw, 1964; Segal, 1969) than

randomly distributed, the question of scale is of

crucial importance in statistical tests for homogeneity.

Thus, Tuomikoski’s (1942) proposal to determine

homogeneity on lack of correlation between the

occurrence of different species is meaningless unless

the question of scale is regarded at the same time.

Dahl & Hadac (1949) gave a definition of homo-

geneity that is often regarded as fairly satisfactory:

“A plant species is said to be homogeneously distri-

buted within a certain area if the probability of finding

an individual of a plant species within a test area of

given size is the same in all parts of the area. A plant

community is said to be homogeneous if the in-

dividuals of the plant species which we use for the

characterization of the community are homogeneously

distributed”. They further point out that plant

communities are never fully homogeneous, and that

one should be satisfied with more or less homogeneous

plant communities. They also state that “the human

eye, badly adapted to measurement, but well to

comparison, rapidly gives the trained sociologist an

impression of whether a plant community he has

before his eyes is highly homogeneous or not”. As

pointed out by Goodall (1961), the definition by Dahl

& Hadac (1949) does not fit vegetation with a mosaic

pattern. Goodall (1961) then suggested the following

definition: “the distribution of a species in an area of

vegetation is homogeneous if there exists some sample

size for which the variation between replicate samples

is independant of the distance between them”. Hence,

the question of homogeneity is linked to the concept of

minimum area. If homogeneity in the sense of Goodall

(1961) could be proved, then a statistically valid

minimum area would exist (cf. Goodall, 1961;

Kershaw, 1964). Testing Goodall's definition involves,

however, a considerable amount of cumbersome

field work and computation. Besides, this and

other tests usually do no more than give an indication

of the degree of heterogeneity.

Since many ecologists agree with Greig-Smith

(1964), that “empirical description of vegetation

cannot wait for clarification of theoretical concepts”,

in particular not if these concepts bear relatively

little importance in many fields of ecological interest,

the suggestion by Dahl & Hadac (1949) to assess

homogeneity subjectively has often been accepted

(e.g. Ellenberg, 1956; Dahl, 1957; Braun-Blanquet,

1964; Daubenmire, 1968; Knapp, 1971). For instance

Daubenmire (1968) defines a homogeneous patch of

vegetation as “one in which variations are attri-

butable to chance, rather than to instrinsic habitat

factors”, and comments that one should not try to

find homogeneity but rather try to eliminate as much

heterogeneity as possibe. This should be done sub-

jectively, because “the results of lengthy computation

often do no more than verify vegetation discontinui-

ties that are evident to a trained synecologist by

careful inspection”. An ecological definition of

homogeneity is given by Godron (1968): “une station

est homogene lorsque chacque espece pent y trouver

des conditions de vie equivalentes d'une extremitee

a l’autre, et non pas en tons les points de la station”

(a stand is homogeneous when every species experiences

equivalent living conditions throughout the entire

extent of the stand and not necessarily in every point

of it). Information theory tests to determine homo-

geneity have been applied by Godron (1966). Since

statistical testing of homogeneity is cumbersome,

many ecologists take the practical approach of sub-

jectively assessing homogeneity of the plot in terms of

the least possible obvious heterogeneity in floristics,

structure and environmental features.

1 .4 Minima! area and plot size; plot form

Another much-discussed and criticized concept is

that of minimal area, dealt with recently by Werger

(1972) and Moravec (1973). Two basic approaches to

the concept of minimal area can be observed :

(1) The approaches based on the community-unit

theory try to determine a certain minimal size of

area in which the community can be represented.

The determination of this size of area is important,

since communities can be sampled most efficiently

with plots the size of minimal area or slightly larger.

If a community is sampled with plots smaller than the

minimal area then the community-type cannot

easily be extracted from the data; if it is sampled with

plots larger than the minimal area then much effort is

wasted.

(2) One aim of the statistical approach to vegetation

is to establish whether plant communities exist or not.

This involves pattern studies that lead automatically to

studies of minimal area, again as the minimum size of

area in which the community, if it exists, can be

represented.

In the community-unit theory approaches, however,

the analytic and synthetic concepts of minimal area

are often confused. The analytic concept of minimal

area implies determination of the size of area on

which a phytocoenosis (community, in Whittaker's

(1956, 1962) terminology) is fully represented, in

order to ascertain a suitable plot size for sampling

that and similar phytocoenoses. The synthetic

concept of minimal area implies determination of the

average size of area on which a particular association

(community-type, in Whittaker's (1956, 1962)

terminology) is represented. To determine this area it

is, amongst other requirements, necessary to know the

312

ZURLCH-MONTPELL1ER METHOD OF VEGETATION SURVEY

total number of species of an association (compare

Raabe's remarks following a paper by Van der Maarel,

1970). Due to the varying distance between the species

within each of several phytocoenoses of the same

association, Raabe regarded it as impossible to deter-

mine the minimal area of an association. Only the

minimal area of a particular phytocoenosis can be

determined.

Since the problem of a suitable plot size for sampling

vegetation is more important in vegetation survey

than the possible determination of the minimal area

of an association, the analytic concept of minimal

area will be further considered here.

Braun-Blanquet (1913) originally determined mini-

mal area as the area above which no new species occur

in the association (probably meaning the phytocoeno-

sis). This size of area is determined with the species-

area curve (cf. Werger, 1972). Later Braun-Blanquet

(1928, 1951, 1964) modified this to the area at which

the species-area curve becomes more or less horizontal

(cf. also Ellenberg, 1956). Recently, Tiixen (1970c)

reinterpreted this curve, regarding it as consisting of

three phases: (a) a strongly curved phase; (b) a

slanting straight line; (c) a horizontal line. It seems

that this reinterpretation deserves further investigation

to establish to what degree the three phases are a

result of scale and a neglect of mathematical inter-

polation.

Relative scales of abscissa and ordinate axes are

important in determining the position of the inflection

point of the species area curve (Cain & Castro, 1959;

cf. Werger, 1972).

The Uppsala School defined minimal area as the

area above which there was no increase in constant

species, constant species being those with a percentage

frequency over 90% on an area of sufficient size, that

is, above the minimal area. The constancy-area

curve is supposed to show a step above which only

extremely large areas would add new constant

species (Du Rietz et a!., 1920; Du Rietz, 1921).

Poore (1964, 1968) constructed species-area curves

for stands of a tropical rain forest in Malaysia and

found that only the curves for constant tree species

flattened out. Hopkins (1955, 1957), however, could

neither find a break in the species-area curves nor a

step in the constancy-area curves from data of several

British stands of vegetation. Goodall (1970) argued

that the asymptotic form of the species-area curve is an

illusion due to the use of linear scale for a measure of

area.

After he had made a detailed study of Dutch dune

vegetation, Van der Maarel (1966a) concluded that

minimal area cannot satisfactorily be defined as an

absolute intrinsic character of the vegetation and

advocated a pragmatic approach, defining minimal

area as the minimal size of area to be analysed to get

a representative view of a phytocoenosis.

Several other definitions of minimal area were

attempted (cf. Goodall, 1952; Werger, 1972) without

solving the matter.

Calleja (1962), Segal (1969), Van der Maarel (1970),

Werger (1972) and others emphasized that, apart from

floristic richness, minimal area also depends on struc-

ture of the phytocoenosis. Thus, the species-area

relation is an interwoven expression of floristic rich-

ness and structure of a patch of vegetation.

Statistical tests, usually based on homogeneity

assumptions in a phytocoenosis, failed to determine a

minimal area objectively (cf. Goodall, 1954b, 1961;

Greig-Smith el al. , 1963; Greig-Smith, 1964; Kershaw,

1964), with the possible exception of some vegetation

types in Western Australia (Goodall, 1963).

Werger (1972) concluded that no convincing mini-

mal-area definition had been formulated and regarded

the concept unsuitable for ascertaining a suitable plot

size for sampling the vegetation. He suggested that

optimal plot size be determined rather than minimal

area, and described a procedure for doing so based on

the regression equation of Gleason (1925) and obser-

vations by Williams (1943) that the average uniform

stand or phytocoenosis seems to be manifested within

an area of about a half to one hectare. Optimal plot

size determined according to this procedure appeared

to depend on floristic richness and structure of

various phytocoenosis (cf. Werger, 1972, 1973a).

This agrees with the previously mentioned arguments

of Raabe (see Van der Maarel, 1970) and indicate that

suitable plot size is likely to differ slightly from phyto-

coenosis to phytocoenosis, even if they belong to the

same community.

Moravec (1973) has recently proposed using the

relationships between area and similarity- and homo-

geneity-coefficients to determine minimal area. After

taking a number of nested quadrats (more than five)

in a phytocoenosis, he calculated the similarities

and homogeneity-coefficients between quadrats of the

same size. To calculate similarity he used Ceska’s

(1966) formula and to calculate homogeneity-

coefficients, Moravec’s (1971b) formula. Plotting the

mean similarity- and homogeneity-coefficient values

against increase of area, resulted in quickly rising

curves which, after reaching certain values, fluctuated

around these values or even decreased. He regarded as

the minimal area of the phytocoenosis the area corres-

ponding to the value reached by the curves after

which they started to fluctuate or decrease. Although

this seems to be a promising approach, Moravec

(1973) found that it did not give unequivocal results in

phytocoenoses showing a regular increase in species

with regularly decreasing density.

In the Ziirich-Montpellier method one is neither

bound to a fixed plot size, nor to a fixed plot form in

sampling the vegetation of a region, because species

are rated on a cover-abundance scale with relative

values. It is important that plot size be adapted

to give a more or less typical description of the phyto-

coenosis that is represented by the vegetation in the

plot, and that the vegetation in the plot should

represent an example of one vegetation type only

(compare Chapters 1 . 2 and 1 . 3 above).

1 . 5 Structure

In sampling, an important decision with regard to

the structure of a phytocoenosis has to be taken,

namely, the decision on structural homogeneity.

Particularly in semi-arid, open shrubby vegetation,

such as the hillside vegetation of the Karoo, low

trees and shrubs are often scattered, singly or in small

groups, over an open vegetation of dwarf shrubs and

grasses. Under these low trees and shrubs certain

smaller species occur, which are usually not found in

the open dwarf-shrub layer (cf. Werger, 1973a). One

can regard such a vegetation as consisting of a mosaic

of two vegetation types, a dwarf shrub and grass

community, and a low tree and shrub community.

However, one can also regard it as consisting of one

vegetation type. In this example, the latter view should

be taken for the following reason. If it were a mosaic

of two communities one would expect to find locally

more extensive homogeneous patches of either of the

two communities. This is, however, not the case.

The hillside vegetation of the Karoo represents,

apparently, a stable ecological-sociological equili-

brium of different growth forms that constitute one

inseparable entity. That certain species occur only

M. J. A. WERGER

313

under the shrubs or low trees should be interpreted as

a consequence of their specific ecological amplitude,

and the occurrence of low trees and shrubs in the

community, which have an impact on microclimate

and local soil moisture conditions. Moreover, the

absence of more extensive homogeneous patches of

such a vegetation of low trees or shrubs with a

specific understory, strongly suggests that the trees or

shrubs do not constitute a vegetation type on their

own, and that the hillsides of the Karoo are not

covered with a close mosaic of two communities but

with one open shrub community. A similar structure

was recently reported by Oberdorfer (1970) from the

Canary Isles, and Dahl & Hadac (1949) interpreted,

according to the same principles, a superficially similar

structure from South Norway, that was properly

considered a mosaic of two communities.

In general, samples should be taken in such a

manner that each plot adequately represents the

structure of the surrounding vegetation. Thus, the

plot line can often cut halfway through a small group

of shrubs, rather than including either a complete

group of shrubs or of the dwarf shrubs and grass

vegetation only.

Another feature of structure in a phytocoenosis is

stratification. Various vegetation layers can often be

recognized in a phytocoenosis, sometimes clearly

distinguishable from each other, at other times hardly

recognizable as separate strata. In the Ziirich-Montpel-

lier School it is believed that these strata are in mutual

ecological interaction and cannot be considered as

separate, independent ecological units (Braun-Blan-

quet, 1964; Westhoff, 1967, 1968; Boerboom, 1960).

There are exceptions to this rule, such as in the case of

crustal lichen communities on tree trunks (cf. Barkman

1958), because such communities do not root in the

same substratum as the trees and other layers, and in

the case of aquatic communities (cf. Den Hartog &

Segal, 1964; Segal, 1965; Westhoff, 1967; Westhoff &

Den Held, 1969), although there is not unanimity in

the latter case (De Lange, 1972). However, the ques-

tion of scale also plays a role. In order to have a record

of the structure of a sampled phytocoenosis, Ztirich-

Montpellier workers traditionally keep notes at each

plot of the various vegetation layers distinguished,

their ranges in height and an estimation of the aerial

cover of each layer. An estimation of total aerial

cover of the entire vegetation of the sample plot,

including all strata, is also recorded.

Tiixen (1957) proposed a simple formula for com-

parison of the cover values of the various vegetation

strata and thus the structure of vegetation samples.

However, since the various layers of a phytocoenosis

are not always clearly separable from one another this

procedure is of only limited value.

1 . 6 Floristic lists

At each sample site complete lists of the species

occurring in the plot should be drawn up. These can

include vascular plants as well as mosses, hepatics,

lichens, fungi, algae, and other organisms. Since the

later extraction of syntaxa in phytosociological tables

is based on these floristic lists, it is of great importance

that all species of the taxonomical groups considered,

which are present in the plot, be recorded and that

the plant individuals should be correctly identified.

The requirement of complete species lists cannot

easily be met in arid and semi-arid areas, since,

depending on irregular rainfall, some geophytes and

annuals might not be visible every year or at any

particular time of the year (cf. Werger & Leistner,

1973). Complete species lists would thus require

multiple visits to each plot site. The extent of a study

area and the time available for study often do not

allow this, so that one has to utilize permanently

recognizable species in such cases (cf. Werger et a!.,

1972; Werger, 1973 a, b). In fioristically rich areas

particularly this is usually not a serious drawback and

the communities can be clearly characterized floristi-

cally on the bases of floristic lists in which only

permanently recognizable species are entered. Further-

more, Batanouny & Abu El-Souod ( 1972) concluded

that in arid regions perennials are generally better

indicators of specific habitat factors than ephe-

merals. The results of the survey by Leistner & Werger

(1973) of the southern Kalahari, seem to comfirm this

conclusion. Also Grunow (1965), Ayyad & Ammar

(1973) and Zohary (1973) indicate the relatively

unimportant role of annuals in vegetation classification

in dry regions, but Garcia-Moya (1972) found them

very useful in a phytosociological classification of arid

vegetation of Arizona.

A species which occurs in more than one vegetation

layer of the plot is sometimes recorded for every

layer in which it occurs and each time rated on the

cover-abundance and sociability scales (see below).

This is especially important in, for example, rejuve-

nation studies in forests and in succession studies.

Some workers have included animal species in

their plot lists in an attempt to delineate types of

biocoenoses (see Braun-Blanquet, 1964; Tiixen,

1969). Studies that have taken smaller animals like

Arthropoda or Mollusca into account, have been

markedly successful (e.g. Molinier, 1960).

Although the Ziirich-Montpellier system is based on

floristics, this is not necessarily the only criterion for a

detailed classification system of vegetation (cf.

Langford & Buell, 1969). In an ordination study

Knight & Loucks (1969) used structural-functional

characters, such as life form, leaf size, pollination

mechanism, bark thickness, seed-dispersal mechanism,

etc., regardless of species, to group stands of forest.

Limitations of such an approach are due to structural-

functional characters usually not being unanimously

acceptably defined as are species. Further, such an

approach often requires a better autecological know-

ledge of species in the plot, than has been gained

of the species of most areas.

1 . 7 Cover-abundance

The relative importance of each species in a quadrat

is assessed on the well-known cover-abundance

scale used by the Ziirich-Montpellier School (Braun-

Blanquet, 1928, 1951; Ellenberg, 1956; Becking,

1957; Schmithiisen, 1968; Knapp, 1971). This scale,

which is based partly on cover and partly on abun-

dance, has received a good deal of criticism (Barkman

et al. , 1964). Cover refers to aerial cover, or the area

covered when the canopy is vertically projected onto

the ground. Cover and abundance are estimated.

Although this is basically a subjective procedure, the

estimates of different phytosociologists analysing the

same patch of vegetation, prove to be amazingly

similar, so that the procedure is not entirely hap-

hazard (Dahl, 1957; Van der Maarel, 1966b). The

common cover-abundance scale, which is easily

comparable with other cover-abundance scales such

as the Domin-scale (cf. Becking, 1957), reads as

follows:

r Very rare and with a negligible cover, (usually a

single individual).

-(-Present but not abundant and with a small

cover value (less than 1 % of the quadrat area).

314

zOrich-montpellier method of vegetation survey

1 Numerous but covering less than 1 % of the quad-

rat area, or not so abundant but covering 1-5% of the

quadrat area.

2 Very numerous and covering less than 5 % of the

quadrat area, or covering 5-25% of the quadrat area

independent of abundance.

3 Covering 25-50% of the quadrat area independent

of abundance.

4 Covering 50-75 % of the quadrat area independent

of abundance.

5 Covering 75-100% of the quadrat area indepen-

dent of abundance.

As shown, scale-units 3, 4 and 5 refer to cover,

whereas the other scale-units can refer to cover and

abundance. Barkman et al. (1964) argued that the

steps in the scale are not proportional to one another

and that the combination of abundance and cover in

one and the same scale is in principle illogical.

They criticized particularly the broad definition of

scale-unit 2. Without altering the basic units of the

cover-abundance scale, they advocated modifications

by adding secondary symbols to the above-listed ones.

This had the advantage of enabling more precise and

narrower definitions of the scale-units, and floristic

lists rated on their scale allowed direct comparison

with older ones rated on the traditional Ziirich-

Montpellier scale by simply disregarding the additional

symbols. The scale of Barkman et al. (1964) appears,

however, to be more suitable for detailed survey of

permanent quadrats, and the like. Only their modi-

fication of scale-unit 2 was adopted in some vegetation

surveys in South Africa (Werger, 1973a, b), reading

as follows:

2m Very numerous, covering less than 5% of the

quadrat area.

2a Covering between 5 and 12% of the quadrat

area independent of abundance.

2b Covering between 13 and 25% of the quadrat

area independent of abundance.

A useful and practical way of estimating cover is in

terms of the average canopy diameter of individuals

of a species. Assuming a circular form to species,

it can be readily calculated that the following relation-

ship exists independently of the size of the particular

diameter (data from the Botanical Research Institute,

In the latest edition of his Pflanzensozio/ogie ,

Braun-Blanquet (1964) slightly changed the cover-

abundance scale without giving reasons for doing so.

Instead of the traditionally used lower limit of

5% of scale-unit 2, he now takes 10% as the lower

limit. This is very disadvantageous as the traditional

scale was used for about half a century and an exten-

sive data collection has been built up. The new defi-

nition of scale-unit 2, which does not constitute an

obvious improvement, can thus only be confusing and

should be avoided.

I .8 Sociability

All species occurring in a quadrat are often also

rated on the sociability scale which is devised to

indicate the grouping of individuals of the same

species. For this purpose, any sprout coming out of the

soil is regarded as a separate individual, regardless of

its subterranean connections with other “individuals”.

The scale, which is “circular” rather than linear

(cf. Barkman et al., 1964), reads as follows:

1 Single individuals.

2 Grouped or tufted.

3 In troops, small patches or cushions.

4 In small colonies or extensive patches, or forming

carpets.

5 In extensive crowds or pure populations.

As clearly pointed out by Barkman (1958), it is not

the sociabilities of different species that are compared,

but the sociabilities of the same species in different

habitats.

Many workers consider that sociability is a specific

feature of each species and they regard it an un-

necessary recording (e.g. Ellenberg, 1956; Barkman

et al., 1964), whereas others emphasize its importance

and document how certain species have varying

degrees of sociability under different ecological

conditions and in different associations (Meltzer &

Westhoff, 1942; Pfeiffer, 1962; Braun-Blanquet, 1964).

Van der Maarel (1966a, 1970) emphasized the

correlation of the sociability of a species with its

growth form, and used a modified sociability scale for

studying vegetation structure.

If the sociability of the species in a community

remains largely the same, its summarized values are

often listed in the end column of the phytosociological

table in order to avoid enlarging the table unneces-

sarily.

1 .9 Other analytic characters of the vegetation

Many other analytic characters of the flora and

vegetation in a quadrat can be recorded, such as

local frequency, density, vitality, fertility, phenology,

growth and life form of the species, and yield of the

vegetation in the quadrat. Some of these characters

are more commonly used in detailed studies of perma-

nent quadrats, or are suitable for special purpose

studies.

For vitality, a number of scales have been developed

(Barkman, 1958; Braun-Blanquet, 1964; Knapp, 1971)

Life forms are usually determined following the system

of Raunkiaer (1934), or modified versions of it

(e.g. Ellenberg, 1956; Braun-Blanquet, 1964; Ellenberg

& Mueller-Dombois, 1967). For growth forms,

comprehensive systems have been worked out (e.g.

Schmithtisen, 1968).

1.10 Habitat characteristics

At each quadrat site, habitat characteristics are

noted. Such notes may be rather superficial or reach a

high degree of detail, and should generally be in

correspondence with the purpose and scale of the

survey and the time available. For instance, statistical

ecologists, who are often more interested in plant-

habitat correlations than merely in a multi-purpose

classification system (cf. Lambert & Williams, 1962;

Lambert & Dale ,1964), often aim for a rather detailed

habitat characterization. The phytosociological school

of Emberger in Montpellier records habitat features in

great detail. Standard checksheets were developed

for this purpose (Godron, 1968). The importance of

standardization of habitat observations is emphasized

by Godron & Poissonet (1970).

M. J. A. WERGER

315

It is obvious that the more precisely habitat obser-

vations are made the more clearly associations

extracted from the data can be characterized eco-

logically later. Apart from the usual identifying

information such as locality, date of sampling and

size of the plot, the following can also be important:

(a) height above sea level;

(b) position in the landscape;

(c) slope angle and slope direction;

(d) details on soil profile, soil depth, rockiness of the

soil, etc.;

(e) water table;

(f) geological substrate;

(g) climatic information.

(h) biotic and other influences;

These data are often entered in the headings of

phytosociological tables, together with information

on vegetation layers, total cover and total number of

species in each plot sample.

1.11 Re/eve

Such a completed site record of species, their

relative importance, other analytic characters of flora

and vegetation, and habitat notations is generally

called a releve.

An appropriate definition of the term releve is

given by Godron (1968): “Le releve est un ensemble

d’observations ecologiques et phytosociologiques

qui concernent un lieu determine” (The releve

constitutes the total of ecological and phytosociolo-

gical observations at a certain point).

1.12 Number of releves

There is no fixed rule concerning the exact number

of releves needed to survey an area adequately.

The number depends entirely on the scale of the

survey, the variety in the survey area and the degree of

precision which one wants to achieve. At the “Centre

d’Etudes Phytosociologique et Ecologique" at Mont-

pellier the number of releves to be taken in an area

is determined by the size and heterogeneity of the

area and the cost of sampling (Godron, pers. comm.).

Statisticians agree that a higher sampling intensity

reduces the variance in the data more effectively

than improved sampling precision (Goodall, 1970).

The total set of releves taken should as adequately as

possible reflect the total variety in the study area. Even

distribution of releves over the study area usually en-

sures this. In exceptional cases, such as for example

that of the dune area of the southern Kalahari (Leist-

ner & Werger, 1973), there is so little variation in the

study area, and homogeneous p'nytocoenoses cover

such extensive areas that an even spread of the releves

is not strictly necessary.

The statement by Shimwell (1971) that the number

of releves collected is “usually less than the manage-

able number of 40” is meant to illustrate when a start

with tabulation can be made and should not be

interpreted as binding advice.

For the compilation of a sound phytosociological

classification system, a large number of releves from

a wide area is necessary (cf. Tiixen, 1 970b).

2. SYNTHESIS

The synthetic stage of the survey follows the

sampling of the field data and should finally lead to

interpreted results. Several concepts and technical

procedures used in the process of synthesis are briefly

elucidated below.

2 . ! Tabulation

The compilation of phytosociological tables from

field data has been described in great detail by Ellen-

berg (1956), and English and French translations of

that account have been published by Kiichler (1967)

and Gounot (1969), respectively. Shimwell (1971)

has also discussed the tabulation procedures in

English so that only a brief outline need be given here.

The field data are entered into a matrix in which

the rows represent species and the columns releves.

When this matix is completed, the result is called the

‘raw table’. The next step is a visual study of the asso-

ciation between species. Positive as well as negative

association is important (Tuomikoski, 1942). The

matrix is rearranged so that positively associated species

are grouped together, apart from the general and the

most infrequent species that do not show clear dis-

criminant floristic associations and are listed in the

lower part of the table. A species’ membership of a

species group is not necessarily determined by

association with any individual species in that group;

it can also be based on the restriction of the species to

the general distribution range in the table of a group of

associated species. A second rearrangement of the

matrix places the releves with a strongly similar species

content together. These successive rearrangements of

rows (species) and columns (releves) in the matrix

should be continued until a clear pattern of mutually

discriminant noda of species-releve groups is obtained.

This way of grouping species and releves into species-

releve groups is in theory similar to taxonomical

classification procedures in which, according to Hull

(1964-65) “usually no one particular property or set of

properties is necessary and any one of numerous sets is

sufficient”. Wittgenstein (1960) originally discussed

this classification process under the term “Familien-

ahnlichkeiten”, or family resemblances (cf. Gilmour

& Walters, 1964; Bambrough, 1966; Khatchadourian,

1966; Werger, 1973a).

Compilation of a final phytosociological table from

a raw table may be facilitated by the use of certain

expedients, such as the apparatus described by Muller

et al. (1972). The latter authors also give a brief review

of similar useful tools.

The process of rearranging a raw table into a final

phytosociological table is largely objective, as has

been emphasized by Ellenberg (1956) and demonstra-

ted by Spatz & Siegmund (1973) and Coetzee &

Werger (1973). The latter authors carried out hier-

archical syndrome analysis on Fynbos data, which

had previously been used to construct phytoso-

ciological tables. The results of the analysis matched

well with the structure of the tables. Spatz & Siegmund

(1973) developed a computer programme to prepare

phytosociological tables. They took data from Ellen-

berg (1956) and were able to compile a computer-

made table, which was nearly identical to that made by

hand by Ellenberg. This and several other computer-

based methods might yield even better results when

quantitative measures of absence similar to those

described by Swan (1970) are used.

It is common practice to place the individual

releves within a nodum in a specific sequence. This

sequence can be according to decreasing total number

of species, according to any environmental gradient

observed, or to any other varying character of the

vegetation or habitat observed (Braun-Blanquet.

1964). Species are usually placed in order of presence

in each nodum. Table units which differ greatly from

one another in floristics are often presented in separate

tables.

17749-8

316

ZURICH-MONTPELLIER method of vegetation survey

The question of whether or not to leave unrepre-

sentative releves out of the final table has been dis-

cussed at great length. Both sides have convincing ar-

guments for their points of view. Tiixen (in lit.), for in-

stance, argues that in initial surveys the object is not so

much to show all possible transitions and fragmen-

tary forms in which an association can occur, but to

describe syntaxa as clearly and definitely as possible.

Heterogeneous, transitional and fragmentary releves

should thus be omitted from the table. Zonneveld

(1960), Segal (1969), Gounot (1969) and Klotzli (1972),

on the other hand, strongly reject the practice

of leaving “bad” releves out of the table. Ac-

cording to Gounot it is possible in such cases

that the table is an illustration of the author's

preconceived ideas rather than a means of syn-

taxonomical research. He adds, however, that such

practices do not invalidate the method as such, but

merely throw doubt upon certain results. The present

author considers that, in addition, the retention of

releves that are atypical representatives of a nodum

has definite advantages in initial studies of unde-

scribed areas, since there releves may well prove valu-

able data for the later recognition of previously unre-

cognized and undersampled syntaxa. In this context, it

is worthwhile referring to Westhoff & Den Held (1969),

who argue that once a proper syntaxonomical system

is available, intermediate or transitional releves are

valuable in that they give information on the habitat

of these transitional vegetation stands. They show that

the transitional habitat is a mixture of the habitat

requirements of the communities that constitute the

transitional stand. The relative importance of the

various habitat factors is in direct correspondence

with that of the constituent communities. Therefore,

releves should not be rejected on prejudice. Each

releve should be interpreted and a releve should,

at most, be omitted from the table when it appears to

be clearly heterogeneous.

2.2 Synthetic characters, character species, diffe-

rential species

A finalized phytosociological table contains far

more information than the sum of the information in

each releve (Tiixen, 1970b). It displays the main

synthetic characters of a community. Since the species

are arranged according to presence within the noda,

the phytosociological table shows, at a glance, the

constant, or dominant species. A number of species

will occur constantly over the entire table, whereas

others will be limited, completely or largely, to one or

a number of noda or table-units, that may, after ecolo-

gical confirmation, represent communities. These spe-

cies that are restricted in their occurrence are faithful to

one or more noda or possible communities. They may

be called differential species, because they differentiate

between noda (cf. Ellenberg, 1956; Damman, 1964).

This need not coincide with a high presence value

within the nodum. Faithfulness to a community or

certain communities is called fidelity and is a synthetic

character. Besides presence, fidelity can also be based

on cover-abundance, sociability and vitality values

(Pfeiffer, 1962; Braun-Blanquet, 1964; Segal, 1969;

Westhoff & Den Held, 1969). The Ziirich-Montpellier

School, following Szafer & Pawlowski, recognizes

five degrees of fidelity —

species that are exclusive to a community in a

particular geographical region, are called exclusive

( treu ) ;

species that are found mainly in one particular

community and seldom in others, are called

selective (fest);

species that are found optimally in one commun-

ity, although they also occur in several others, are

called preferential (hold) ;

species that are indifferent to a particular com-

munity, are called indifferent or vague (vag); and

species that are uncommon or rare in a parti-

cular community, are called strange ( fremd ).

Only exclusive, selective and preferential species are

character species (Ellenberg, 1956; Becking, 1957;

Braun-Blanquet, 1928, 1951, 1964; Knapp, 1971;

Shimwell, 1971). Originally it was thought that most

communities would contain exclusive species (cf.

Braun-Blanquet, 1915). When more data became

available it appeared that more species were selective

or preferential rather than exclusive. Then, also, the

concept of differential species was first introduced by

Koch (1925). This is a species that is preferential for a

certain community and, therefore, distinguishes this

community from floristically related communities,

but it is at the same time equally well or even better

represented in still other communities. Differential

species were initially used only to distinguish sub-

associations and variants of associations, but gradually

it became common practice to use differential species

also to characterize associations and higher syntaxo-

nomical units (Becking, 1957). Tuomikoski (1942)

pointed out that character species are but a special

case of differential species. With the accumulation of

data from various parts of Europe it became clear

that few species are faithful to a specific community

over its entire distribution area. Those species that

are, are called absolute character species (Braun-

Blanquet, 1951; Becking, 1957; Westhoff & Den

Held, 1969; Knapp, 1971 ; Westhoff & Van der Maarel

1973). When the distribution area of a species exceeds

that of the association in which it is found, it often

occurs that the species is faithful to a specific asso-

ciation in one part of its distribution area and to

another vicarious association in another part. Such

species are called territorial or regional character

species (Braun-Blanquet, 1951; Becking, 1957;

Westhoff & Den Held, 1969; Westhoff & Van

der Maarel, 1973). It often occurs, however,

that certain species are faithful only in a lim-

ited part of the range of the association. These

species are called local character species (Becking,

1957; Westhoff & Den Held, 1969; Oberdorfer,

1968; Knapp, 1971; Westhoff & Van der Maarel,

1973). Meijer Drees (1951) and Becking (1957)

elucidated these geographically-based concepts of

character species in clear diagrams. Becking (1957)

also took the relative size of the range common to

association and species into account and pointed

out that a species on the fringe of its distribution area

can characterize an association significantly whereas

the same species may be without diagnostic value in

its optimal range. This is a common phenomenon

with an obvious ecological explanation. In its optimal

range a species will not be strongly restricted by

specific less favourable ecological conditions, while

on the fringe of its distribution area, where the

environment is rather unfavourable, it can survive

only in a narrowly defined habitat (cf. also Klotzli,

1972).

In a later paper Becking (1961) suggested dis-

tinguishing between faithful presence, faithful socia-

bility and faithful vitality. He tried to define these

concepts quantitatively, but this did not prove to be

useful.

Schwickerath (1944, 1954, 1968) has repeatedly

attacked the concept of local character species and

argued that from comprehensive studies covering the

M. J. A. WERGER

317

entire range of the association, these species would

appear to be either true character species or geo-

graphical differential species. According to Schwicke-

rath, geographically-varying character species will

still have a clear optimum over the entire range of the

of the association, although this optimum may vary

geographically, Geographical differential species occur

only in a limited range of the association. They can

either be largely restricted to a specific association,

or they can be important in other associations as well.

Still another type of character species is the trans-

gressive character species. Such a species is characteris-

tic of a syntaxon above the association level, for

example an alliance, but at the same time it is more

characteristic of a specific association within that

alliance than of the other associations in the alliance

(Becking, 1957; Westhoff & Den Held, 1969).

Species which do not differentiate between com-

munities are called accompanying or companion

species. They can be constant as well as rare.

With the gradual broadening of the fidelity concept,

it became more usual to characterize an association by

differential species, and even by characteristic species

combinations, although these practices are used with

some restriction (Barkman, 1958; Westhoff & Den

Held, 1969). Fidelity of species to associations can

be determined only on the basis of phytosociological

tables, using the method described by Schwickerath

(1931) in less evident cases. Associations are not

“determined in the field because species of known

narrow amplitude are present”, reducing the later

determination of faithful species to a circular argument

as Poore (1955, 1956) suggested. As pointed out by

Barkman (1958), Moore (1962) and Gounot (1969),

this suggestion was based on a poor understanding of

the Zurich-Montpellier approach and is entirely

invalid.

The East German School of Scamoni and Passarge

has diverged slightly from the theory of the Ziirich-

Montpellier School, and uses the concept of sociolo-

gical species groups. These are formed of species in

the same vegetation layer, which show a similar

sociological behaviour owing to similar ecological

requirements. The species in one sociological group

often possess similar growth-forms. In certain com-

munities they have their maximum or optimum

together, whereas they are simultaneously absent from

other communities (Passarge & Hofmann, 1964, 1968;

Scamoni et al., 1965; Passarge, 1968, cf. Doing, 1962,

1969; Janssen, 1967; Klotzli, 1972). The theoretical

and practical value of this approach has still to be

shown (cf. Hohenester, 1970).

The average number of species in the releves of

a community and the community structure can also

be regarded as synthetic characters of the vegetation.

These topics will be dealt with in sections 2.6 and

2.7.

2.3 Ecological confirmation of the table pattern

The arrangement of species and releves in the

phytosociological table is not an end in itself, but

should lead to a comprehensive classification system

of syntaxa, that can serve as the basis for further

ecological studies (Braun-Blanquet, 1964; Ivimey-

Cook & Proctor, 1966; Tiixen, 1970b). When the noda

in the phytosociological table are established, their

biological reality should be confirmed by demon-

strating coincidence between the community-types

they possibly represent and specific habitat conditions.

The checking of the ecological meaning of a supposed

community can be done in the field, or from the

habitat characterization of the releves on the field

sheets. Should this latter approach be used, final

checking in the field is still necessary.

Confirmation of coincidence between the pattern on

the phytosociological table and specific habitat

conditions is necessary for the following reason:

All taxa have their ecological amplitude (which,

for the purpose of this discussion, includes the

effect of competition) within which they can survive

and outside which they cannot exist. Species that have

a high presence over the entire phytosociological

table presumably have an ecological amplitude that

is at least so wide as to include all habitats of the

releves making up the table. Character and differen-

tial species, however, have a smaller ecological ampli-

tude, at least in the area where the releves were taken.

Their ecological amplitude is such that these species

are unable to exist in some of the habitats of the

releves comprising the table. Because the central

concept of all approaches based on the community-

unit theory implies that vegetation consists of basic

natural entities which are generally in contact with one

another along narrow boundaries, it is supposed

that more or less coincident ecological amplitudes of

some species, at least to the one side, result in such a

narrow boundary. Hence, if the coincidence of certain

habitat conditions and a nodum section in the table

can be confirmed, a basic vegetation unit, a community

is delineated. Species, as indicators of a certain habitat

and typical of a community, are thus indicators for

the habitat typical for a community. This is also

expressed by Von Glahn (1968) when he says that the

vegetation type is “das maximale korrelative Konzen-

trat wiederholt vorkommender Artverbindungen und

bestimmter Merkmalskombinationen seines Standort-

tes”. In this way it can be shown that the pattern in the

table is not an artificial one due to a clever mani-

pulation of the matrix rows and columns, but re-

represents real, fioristically- and environmentally-

characterized natural entities. In the Zurich-Mont-

pellier approach it is thus empirically determined that

patterns in floristic composition correspond with

patterns in the environment (cf. Goodall, 1953;

Gounot, 1961, 1969; Moore, 1962; Segal, 1969;

Garcia-Moya, 1972).

2.4 The association and other syntaxa ; the constancy

table

The definition, proposed by Schroter and Flahault

and accepted by the Third International Botanical

Congress in Bruxelles in 1910, of the basic natural

community of the Zurich-Montpellier system, the

association, says that it is a plant community that has a

definite floristic composition, a uniform physiognomy

and is bound to uniform habitat conditions. Although

special weight was and is given to character species,

and later to differential species, the association is not

based solely on these species, but on its total floristic

composition (Braun-Blanquet, 1915, 1959, 1964;

Westhoff & Den Held, 1969).

Associations are subdivided into subassociations,

variants and facies. Subassociations and variants are

characterized by their own differential species and

mostly it is only in subordinate features that their

specific habitats differ from the habitat of other

subassociations in the same association. A facies is

characterized by the strong dominance of a specific

species and possesses no character or differential

species of its own (Ellenberg, 1956; Braun-Blanquet,

1964).

Associations are combined into alliances, alliances

into orders and orders into classes. Alliances and

orders have their own characte jand differential spe-

318

ZURICH- MONTPELLIER METHOD OF VEGETATION SURVEY

cies; classes are distinguished on the basis of character

species only (Braun-Blanquet, 1964). A useful tool in

the procedure of hierarchization is the so-called

Roman or constancy table (although based on

presence). The presence of each species within each

community is rated on a five-point scale and the range

of its cover-abundance value is noted:

I Species present in 1-20% of the releves of a

community.

II Species present in 21-40% of the releves of a

community.

III Species present in 41-60% of the releves of a

community.

IV Species present in 61-80% of the releves of a

community.

V Species present in 81-100% of the releves of a

community.

Each community is thus reduced to a single column

in the Roman table, so that it is easy to assess which

species are faithful to the various communities.

With the development of the concepts of character

and differential species, the concept of the association

was also somewhat modified. Whereas, according to

certain authors, the association should still possess

character species (Schwickerath 1954, 1968; Braun-

Blanquet, 1964; Oberdorfer, 1968) others recognize,

in exceptional cases, associations that have only

differential species or only a characteristic species

combination (Doing Kraft, 1956; Maas, 1959;

Barkman, 1958; Damman, 1964; Westhoff & Den

Held, 1969; Werger, 1973a). Many association

character species are only local or regional character

species. According to Oberdorfer (1968), every

association is therefore at the same time a regional or

geographical association (Gebietsassoziation). If, in

different areas of the regional association, differences

occur in the species combinations of accompanying or

also order and class character species, while the floristic

composition of the group of association and alliance

character species remains unaltered, one should

distinguish geographical races of the association.

If, however, the floristic composition of the group of

alliance and association character species alters,

different geographical associations should be dis-

tinguished (Oberdorfer, 1968). Borhidi (1963) sug-

gested the combination of regional associations into

synassociations, which would rank between regional

associations and suballiances. Schwickerath (1954,

1968) wanted to recognize only geographic subasso-

ciations of an association and, as pointed out above,

argues that most species, which are called local

character species at present, will finally prove to be

geographical differential species, whereas a few others

will prove to be true character species. Knapp (1971)

accepted the idea of geographical associations and

suggested the combination of them into chief associa-

tions (Hauptassoziationen) or association groups

(Assoziationsgruppe). As pointed out by Westhoff

& Den Held (1969), several authors have recently

regarded geographical associations as the real

associations and promoted the chief associ-

ation simply to alliance. This is one of the practices

which can lead directly to what Pignatti (1968)

called “the inflation of higher syntaxonomical units”.

2 . 5 Nomenclature

No internationally accepted nomenclatural code for

syntaxonomical use exists as yet, although, particularly

since the early fifties (cf. Moravec, 1969), several

attempts have been made to achieve this aim. Westhoff

& Den Held (1969) point out that one of the main

reasons for the persistent failure to reach an inter-

national agreement is that the type-method as used in

taxonomy, and advocated for syntaxonomical use by

Barkman (1958) and Moravec (1968), is not suitable,

in that it is hardly possible to designate generally

acceptable type-releves. Another difficulty results

from a taxon name being intended as a label without

diagnostic function, whereas the name of a syntaxon

refers to one or more species occuring in it. Although a

strict priority rule is often difficult to apply, priority is

generally recognized as far as possible. Tentative rules

were published by Meijer Drees (1954) and were

commented upon by several workers in the same issue

of Vegetatio. More recently Moravec (1968) published

a renewed design for a nomenclatural code, which

comprised several of Meijer Drees’ proposals. The

suggestions of Moravec (compare also Moravec,

1969, 1971a), which seem to have received consider-

able approval, deal with :

(a) conditions for the effective and valid publication

of a name;

(b) sufficiency of the added diagnosis;

(c) typification of the name;

(d) priority;

(e) retention, rejection and alteration of names;

(f) citation of authors’ names.

In the Zurich-Montpellier School syntaxa are

traditionally named after one or two taxa, adding an

appropriate suffix to indicate the rank of the syntaxon

concerned, to the name of the genus (Braun-Blanquet,

1928, 1951, 1964). These suffixes are -etosum, -etum,

-ion, -etalia and -etea for subassociations, associations,

alliances, orders and classes respectively. If desireable,

the specific epithet can be added in genitive form.

Detailed linguistic guides for the correct formation of

syntaxonomical names have been published by Bach

et al. (1962) and in particular by Rauschert (1963).

Originally it was suggested that only character

species be used for naming the associations, but

this proved untenable (cf. Moravec, 1969). The only

requirement that can be made at present is that the

name-giving taxon must at least be present in the

syntaxon concerned. It is common practice, however,

that syntaxa are named after one or two of the taxa

listed in the group of character and differential

species. Sometimes a prominent or constantly abun-

dant species is used in combination with a species of the

character and differential species group to designate

a syntaxon. The terms “inops” and “typicum” for

subassociations should be used following Westhoff

& Den Held (1969).

2 . 6 Homotoneity; affinities between syntaxa

As stated above, the average number of species

in the releves of a community can be regarded as a

synthetic character of the community (Raabe, 1952).

Comparison of such figures for various communities

indicates the relative floristic richness of a community.

The value for the average number of species in the

releves of a community is based on the total number of

species in each releve and is, therefore, a standard for

judging the relative floristic agreement between the

releves and the communities to which they belong.

When the total number of species in each releve is

more or less similar for all releves within one com-

munity, this is an indication that the community-table

is more or less homogeneous. Table homogeneity is

called homotoneity in recent literature (Dahl, 1957;

cf. Westhoff & Van der Maarel, 1973). Hence, the

absolute value of these figures is, as far as homo-

toneity is concerned, theoretically unimportant.

M. J. A. WERGER

319

Only the deviation of the value for each releve from

the average is important, in that it indicates to what

extent each releve is floristically representative of

the entire community (Knapp. 1971). A complicating

factor concerning the matter of average number of

species is, of course, the number of species that are

common to all releves in a community or to only a

limited number of releves, or that are present in only

one releve in a community (cf. Tiixen, 1970c). At

present there is no agreement on the theoretical impli-

cations of the homotoneity of a community-table [see,

for example, the discussion following Tiixen's (1970c)

expose], but a number of procedures have been sug-

gested to determine homotoneity. Jaccard’s commu-

nity coefficient is probably the oldest way of determi-

ning the similarity between two stands or releves (cf.

Ellenberg, 1956; Dahl, 1957; Becking, 1957; Barkman,

1958; Braun-Blanquet, 1964; Ceska, 1966). Jaccard’s

coefficient takes the form c / (a+b+c), where releve

A is compared with releve B, and a is the number of

species occuring only in A, b the number of species

occuring only in B, and c the number of species

common to A and B. Poore (1955c) proposed a

modification of this formula 200c/(a+b+2c).

Other often-used coefficients are those of Kulczinsky

and of Sorensen (cf. Dahl, 1957; Becking, 1957;

Braun-Blanquet, 1964; Ceska, 1966). Kulczinsky’s

coefficient is similar to the one given by Poore

(1955c), except that a, b, and c are presence degrees

(cf. Barkman, 1958). Also, Sorensen’s coefficient is

similar to the others: 2c/ (a+b+2c), in which a, b and

c are again numbers of species (cf. Dahl, 1957;

Becking, 1957; Braun-Blanquet, 1964). Ellenberg

(1956) proposed the formula c/ (2a+2b+c), where

a, b and c represent total cover values calculated

according to Tiixen & Ellenberg (1937). Pfeiffer

(1957) suggested using only the most constant species

of a syntaxon, as defined by Raabe (1952) (see below),

in homotoneity calculations. The homotoneity

value of a syntaxon equals, according to the Pfeiffer’s

formula, 100 (Vi +V2 Vz)/n.g, where

V2 ....... Vz are the number of species in each

individual releve, n is the total number of releves,

and g the total number of species in the syntaxon.

Barkman (1958) modified Jaccard's coefficient to

c/V ab (in which a, b and c are again numbers of spe-

cies). Recently, Moravec (1971b) proposed a procedure

to determine the homotoneity of a set of releves based

on species with a constancy of over 60 per cent in

the species composition of an “average” releve,

corrected by a factor which is proportional to the

difference in species number between the richest and

poorest releve. Slightly more complicated ways of

determining homotoneity were proposed by Looman

& Campbell (1960) and Dahl (1960). Looman &

Campbell (1960) used Sorensen’s coefficient, but

determined statistically the minimum value for which

two releves can still be regarded similar. Dahl (1960)

proposed an index of uniformity defined as the ratio

between the mean number of species per releve and

the index of diversity. Gounot & Calleja (1962) and

Ceska (1966) studied the various possibilities of using

Sorensen’s coefficient and the latter author suggested a

formula derived from it to calculate the mean floristic

similarity within a set of releves. Yet another way of

determining the homotoneity of a community table

is the graphical representation of the number of

species in the various presence classes, based on the

Roman table. If the diagram of the five presence

classes is U-shaped, which means that presence

classes I and V are more frequent than the

adjacent in-between ones, the community is considered

to be homogeneous (Matuszkiewicz, 1948; Dahl &

Hadac, 1949; Becking, 1957; Dahl, 1957; Gounot,

1961, 1969; Goodall, 1970; Bjornstad, 1971; cf.

Etter, 1948).

Similarity coefficients of the types discussed above,

or derived ones, are also used in computer-based pro-

grams for the compilation of phytosociological tables

(e.g. Spatz & Siegmund, 1973). Other computer-based

programs use entirely different distance coefficients,

however (e.g. Schmid & Kuhn, 1970).

Many of the above-mentioned coefficients for

calculating similarity between two releves are also used

to calculate the affinities between two syntaxa (cf.

Barkman, 1958; Braun-Blanquet, 1964). Here again,

values for total numbers of species, total cover

values (Tiixen & Ellenberg, 1937), or “Gruppena-

bundanz” values (Schwickerath, 1931) can be used.

Raabe (1952) suggested that only the most constant

species be used in the calculations. The group of most

constant species would comprise the “characteristic

species combination” and its number would be as

large as the average number of species of the syntaxon.

Barkman (1958) suggested a more complicated method

based on total cover values.

2.7 Community structure

The definition of the association as accepted at

Bruxelles in 1910 (see previous), and also more modern

versions, imply, amongst other requirements, that the

association should be of a uniform physiognomy. If

one uses a fairly rigid physiognomic classification

system such as that of Fosberg (1967), however, this

requirement can often not be met (cf. Werger, 1973a,

b). Owing to the rigidity of the physiognomic classifi-

cation system, communities with a considerable

similarity in structure can fall into completely different

formation classes.

A second reason for structural heterogeneity

within some associations and other syntaxa, is that

due to environmental circumstances some phyto-

coenoses possess an additional vegetation layer as

compared with other floristically similar phyto-

coenoses (cf. Van Donselaar, 1965; Werger, 1973a).

This is a feature that is apparently related to the

phenomenon of twin formations as discussed by West-

hoff (1967, 1968), who adopted the term from nine-

teenth century publications by Hult and by Katz.

Westhoff (1967, 1968) elaborates on a number of

examples where structurally different variants, sub-

associations or associations are combined into the

same higher syntaxon because their floristic com-

position is strongly related. According to the floristic

principle of the Zurich-Montpellier School these

syntaxa cannot be equally well classified in other,

structurally more similar syntaxa. Westhoff (1967,

1968) observed that these twin formations occur only

where extreme habitat factors predominate. Such ex-

treme habitat factors are usually abiotic, for example,

extreme temperatures, frequent burning and a

strongly-fluctuating water table, and are to such an

extent overriding that vegetation structure in these

places is of relatively low ecological importance.

Structure should thus not be over-estimated as a

diagnostic criterion in such situations and floristic

criteria should normally prevail (Westhoff, 1967, 1968).

In South Africa, where mismanagement of the

natural vegetation has occurred on a large scale,

overgrazing and severe trampling can also be regarded

as extreme habitat factors (cf. Werger, 1973a).

In such cases it is clear that the floristic principle

should prevail and that differences in structure should

be regarded as of secondary importance in compiling a

classification of vegetation.

320

ZURICH-MONTPELL1ER METHOD OF VEGETATION SURVEY

ACKNOWLEDGEMENTS

The following persons kindly made critical com-

ments on the manuscript: Mr B. J. Coetree, Dr D.

Edwards, Mr J. W. Morris and Prof. Dr V. Westhoff.

Their assistence is gratefully acknowledged.

OPSOMMING

Omdat die Zurich-Montpellier metode by die

bestudering van plantegroei in Suid-Afrika in die

toekoms waarskynlik baie meer gebruik gaan word,

word hier ’n kort beskrywing van die begrippe en

tegnieke, waarvan by hierdie metode gebruik gemaak

word, gegee.

REFERENCES

Acocks, J. P. H., 1953. Veld Types of South Africa. Mem.

Bot. Surv. S. Afr. 28: 1-192.

Adamson, R. S., 1938. The Vegetation of South Africa. London:

British Empire Vegetation Committee.

Ashby, E., 1948. Statistical ecology. II. A reassessment. Bot.

Rev. 14:222-234.

Augarde, J., 1957. Contribution a l’etude des problemes de

l’homogeneite en son application a la phytosociologie.

Bull. Serv. Carte Phytogeogr. Ser. B 2:1 1-23.

Ayyad, M. A. & Ammar, M. Y., 1973. Relationships between

local physiographic variations and the distribution of com-

mon Mediterranean desert species. Vegetatio 27: 163-176.

Bach, R., Kuoch, R. & Moor, M., 1962. Die Nomenklatur der

Pflanzengesellschaften. Mitt. Flor.-Soz. Arbeitsgem. N. F.

9: 301-308.

Bambrough, R., 1966. Universals and family resemblances.

in Pitcher, G. (ed.): Wittgenstein, pp. 186-204. London:

MacMillan.

Barkman, J. J.. 1958. Phytosociology and ecology of crypto-

gamic epiphytes. Assen : Van Gorcum.

Barkman, J. J., Doing, H. & Segal, S., 1964. Kritische

Bemerkungen und Vorschlage zur quantitativen Vegeta-

tionsanalyse. Acta Bot. Neerl. 1 3 : 394-419.

Batanouny, K. H. & Abu El-Souod, S., 1972. Ecological and

phytosociological study of a sector in the Lybian desert.

Vegetatio 25: 335-356.

Bayer, A. W., 1955. The ecology of grasslands, in Meredith,

D. (ed.): The grasses and pastures of South Africa, pp.

539-550. Parow: Central News Agency.

Becking, R. W., 1957. The Zurich-Montpellier School of

phytosociology. Bot. Rev. 23: 41 1-488.

Becking, R. W., 1961. Quantitative evaluation of phytosocio-

logical data. Rec. Adv. Bot. 2: 1346-1350.

Bews, J. W., 1918. The grasses and grasslands of South Africa.

Pietermaritzburg: Davis.

Bjornstad, A., 1971. A phytosociological investigation of the

deciduous forest types in Sogne, Vest-Agder, South Norway.

Norw. J. Bot. 18: 191-214.

Boerboom, J. H. A., 1960. De plantengemeenschappen van de

Wassenaarse duinen. Meded. Landbouwhogeschool Wage-

ningen 60 (10): 1-135.

Borhidi, A., 1963. Die Zonologie des Verbandes Fagion

illvricum. I. Allgemeiner Teil. Acta Bot. Acad. Sc. Hung.

9: 259-298.

Braun(-Blanquet), .1., 1913. Die Vegetationsverhaltenisse der

Schneestufe in den Ratisch-Lepontischen Alpen. N.

Denkschr. Schweiz. Naturf. Ges. 48: 1-347.

Braun(-Blanquet), J., 1915. Les Cevennes meridionales

(Massif de T Aigoual). Geneve: Soc. Gen. Impr.

Braun-Blanquet, J., 1928. Pflanzensozio/ogie. Wien: Springer.

1. Aufl.

Braun-Blanquet, J., 1932. Plant sociology, (transl. G. D.

Fuller & H. S. Conard). New York: McGraw Hill.

Braun-Blanquet, J., 1951. Pflanzensoziologie. Wien: Springer.

2. Aufl.

Braun-Blanquet, J., 1959. Grundfragen und Aufgaben der

Pflanzensoziologie in Turrill, W. B. (ed.): Vistas in

Botany pp. 145-171. London: Pergamon Press.

Braun-Blanquet, .1.. 1964. Pflanzensoziologie. Wien: Spiinger.

3. Aufl.

Braun-Blanquet, .1., 1968. L’ecole phytosociologique Ziiricho-

Montpellieraine et la S.I.G.M.A. Vegetatio 16: 1-78.

Bray, J. R. & Curtis, J. T., 1957. An ordination of the upland

forest communities of southern Wisconsin. Ecol. Monogr.

27: 325-349.

Cain, S. A. & De Oliveira Castro, G. M., 1959. Manual of

vegetation analysis. New York : Harper.

Calleja, M., 1962. Etude de la courbe aire-espece et Pure

minimale. Bull. Serv. Carte Phytogeogr., Ser. B. 7: 161-179.

Ceska, A., 1966. Estimation of the mean floristic similarity

between and within sets of vegetational releves. Folia

Geobot. Phytotaxon. 1 : 93-100.

Coetzee, B. J., 1972. 'n Plantsosiologiese studie van die Jack

Scott-Natuurreservaat. M.Sc. thesis, Pretoria: Univ.

Pretoria (Unpubl.)

Coetzee, B. J., 1974. A phytosociological classification of the

vegetation of the Jack Scott Nature Reserve. Bothalia

11(3): 329-347.

Coetzee, B. J. & Werger, M. J. A., 1973. On hierarchical

syndrome analysis and the Zurich-Montpellier table

method. Bothalia 11 : 159-164.

Curtis, J. T., 1959. The vegetation of Wisconsin. An ordination of

plant communities. Madison: Univ. Press.

Dahl, E., 1957. Rondane. Mountain vegetation in South

Norway and its relation to the environment. Skr. Norske

Vid.-Akad. i Oslo. I. Mat.- Naturv. Kl. 1956(3): 1-374.

Dahl, E., 1960. Some measures of uniformity in vegetation

analysis. Ecology 41 : 805-808.

Dahl, E. & Hadac, E., 1949. Homogeneity of plant commu-

nities. Stud. Bot. Czechoslov. 10: 159-176.

Damman, A. W. H., 1964. Some forest types of central New-

foundland and their relation to environmental factors.

Forest Science Monogr. 8 : 1-62.

Daubenmire, R., 1968. Plant communities. A textbook of plant

synecology. New York : Harper & Row.

De Lange, L., 1972. An ecological study of ditch vegetation in

the Netherlands. Ph. D. thesis. Univ. Amsterdam.

Den Hartog, C. & Segal, S., 1964. A new classification of the

waterplant communities. Acta Bot. Neerl. 13: 367-393.

Doing, H., 1956. Het assaciatiebegrip en de systematiek van

plantengemeenschappen volgens de methode van Biaun-

Blanquet. Vakbl. v. Biol. 36 ( 1 1 ): 1-13.

Doing, H., 1962. Systematische Ordnung und floristische

Zusammensetzung niederlandischer Wald- und Gebiisch-

gesellschaften. Wentia 8: 1-85.

Doing, H., 1969. Sociological species groups. Acta. Bot. Neerl.

18:398-400.

Doing, H., 1970. The use of phytosociological methods in

various parts of the world. Belmontia 4 (12): 53-61.

Du Rietz, G. E., 1921. Zur methodologischen Grundlage der

modernen Pflanzensoziologie. Wien: Holzhausen.

Du Rietz, G. E., Fries, T. C. E, Osvald, H. & Tengwall,

T. A., 1920. Gesetze der Konslitution natiirlicher Pflanzen-

gesellschaften. Vetensk. Prakt. Unders. Lapp!. Flora och

Fauna 7 : 1-47.

Edwards, D., 1967. A plant ecological survey of the Tugela

River Basin, Natal. Mem. Bot. Surv. S. Afr. 36: 1-285.

Ellenberg, H., 1956. Aufgaben und Methoden der Vegetations-

kunde. in Walter, H.: Einfuhrung in die Phytologie.

IV, 1. Stuttgart: Ulmer.

Ellenberg, H. & Mueller-Dombois, D., 1967. A key to Raun-

kiaer plant life forms with revised subdivisions. Ber.

Geobot. Inst. ETH, Stiftung Rube! 37 : 56-73.

Etter, H., 1948. De l'analyse statistique des tableaux de

vegetation. Vegetatio 1 : 147-154.

Fosberg, F. R., 1967. A classification of vegetation for general

purposes, in Peterken, G. F. (ed.): Guide to the checksheet

of conservation areas. IBP Handbook 4: 73-120.

Garcia-Moya, E., 1972. A preliminary vegetation classification

of the Tombstone , Arizona, vicinity. Ann Arbor: Univ.

Microfilms.

Gilmour, J. S. L. & Walters, S. M„ 1964. Philosophy and

classification, in Turrill, W. B. (ed.): Vistas in Botany

4: 1-22. Oxford: Pergamon.

Gleason, H. A., 1925. Species and area. Ecology 6: 66-74.

Godron, M., 1966. Application de la theorie de l’information

a l’etude de homogeneite et de la structure de la vege-

tation. Oecol. Plant. 1 : 187-197.

Godron, M., (ed.), 1968. Code pour le releve methodique de hi

vegetation et du milieu. Principes et transcription stir cartes

perforees. Paris: Centre Nat. Rech. Scient.

M. J. A. WERGER

321

Godron, M. & Poissonet, J., 1970. Standardization and treat-

ment of ecological observations, in Frankel, O. H. &

Bennett, E.: Genetic resources in plants — their explo-

ration and conservation. IBP Handbook 11: 189-204.

Goodall, D. W., 1952. Quantitative aspects of plant distribution

Biol Rev. 27: 194-245.

Goodall, D. W., 1953. Objective methods for the classification

of vegetation. I. The use of positive interspecific correlation.

Aust. J. Bot. 1 : 39-63.

Goodall, D. W„ 1954a. Vegetational classification and

vegetational continua. in Janchen, (ed.): Festschrift fur

Erwin Aichinger. 1. Angewandte Pflanzensoziologie. pp.

168-182. Wien: Springer.

Goodall, D. W., 1954b. Minimal area: a new approach.

VIII Congr. bit. Bot. Paris 1954, Sect. 7. Rapp, et Comm.

pp. 19-21.

Goodall, D. W., 1961. Objective methods for the classification

of vegetation. IV. Pattern and minimal area. Austr. J. Bot.

9: 162-196.

Goodall, D. W., 1963. Pattern analysis and minimal area —

some further comments. J. Ecol. 51 : 705-710.

Goodall, D. W., 1970. Statistical plant ecology. Ann. Rev.

Ecol Syst. 1 : 99-124.

Gounot, M., 1961. Les methodes d'inventaire de la vegetation.

Bull. Serv. Carte Phytogeogr., Ser. B. 6: 7-93.

Gounot, M., 1969. Methodes d' etude quantitative de la vege-

tation. Paris: Masson.

Gounot, M. & Calleja, M., 1962. Coefficient de communaute

homogeneite et aire minimale. Bull. Serv. Carte Phyto-

geogr. Ser. B. 1: 181-200.

Greio-Smith, P., 1964. Quantitative plant ecology. London:

Butterworth. 2nd ed.

Greig-Smith, P., Kershaw, K. A. & Anderson, D. J., 1963.

The analysis of pattern in vegetation: a comment on a paper

by D. W. Goodall. /. Ecol. 51 : 223-229.

Grunow, J. O., 1965. Objective classification of plant com-

munities: a synecological study in the sour-mixed bushveld of

Transvaal. D.Sc. (Agric.) thesis. Pretoria: Univ. Pretoria.

(Unpubl.)

Grunow, J. O. & Lance, G. N., 1969. Classification of savanna

by information analysis. S. Afr. J. Sci. 65: 341-348.

Grunow, J. O. & Morris, J. W., 1969. Preliminary assessment

of ecological status of plant species in three South African

Veld Types. J. S. Afr. Bot. 35 : 1-12.

Hohenester, A., 1970. Idiotaxonomie und syntaxonomie.

Ber. Dtsch. Bot. Ges. 83 : 641-646.

Hopkins, B., 1955. The species-area relations of plant communi-

ties. J. Ecol. 43 : 409-426.

Hopkins, B., 1957. The concept of minimal area. J. Ecol.

45 : 441-449.

Hull, D. L., 1964-65. The effect of essentialism on taxonomy —

two thousand years of stasis. I. Brit. J. Phil. Sci. 15: 314—

326. II. Brit. J. Phil. Sci. 6: 1-18.

Ivimey-Cook, R. B. & Proctor, M. C. F., 1966. The application

of association-analysis to phytosociology. J. Ecol. 54: 179-

192.

Janssen, C. R., 1967. A floristic study of forests and bog vege-

tation, Northwestern Minnesota. Ecology 48: 751-765.

Kershaw, K. A., 1964. Quantitative and dynamic ecology.

London: Arnold.

Khatchadourian, H., 1966. Common names and “family

resemblances”, in Pitcher, G. (ed.): Wittgenstein, pp.

205-230. London: MacMillan.

Killick, D. J. B., 1963. An account of the plant ecology of the

Cathedral Peak area of the Natal Drakensberg. Mem. Bot.

Surv. S. Afr. 34: 1-178.

Killick, D. J. B., 1967. Fifty years of plant ecology in South

Africa. J. S. Afr. Biol. Soc. 7-8 : 1 1-28.

Klotzli, F., 1972. Grundsatzliches zur Systematik von Pflanzen-

gesellschaften. Ber. Geobot. Inst. ETH . Stiff. RiibelAl : 35-47.

Koch, W., 1925. Die Vegetationseinheiten der Linthebene

unter Beriicksichtigung der Verhaltnisse in der Nordost-

schweiz. Systematisch-kritische Studie. Jb. St. Gall.

Naturw. Ges. 61 : 1-146.

Knapp, R., 1971. Einfiihrung in die Pflanzensoziologie. Stutt-

gart: Ulmer. 3. Aufl.

Knight, D. H. & Loucks, O. L., 1969. A quantitative analysis of

Wisconsin forest vegetation on the basis of plant function

and gross morphology. Ecology 50: 219-234.

Kuchler, A. W., 1967. Vegetation mapping. New York:

Ronald Press.

Kylin, H., 1923. Vaxtsociologiska randanmarkningar. Bot.

Notiser 1923: 161-234.

Kylin, H., 1926. fiber Begriffsbildung und Statistik in der

Pflanzensoziologie. Bot. Notiser 1926: 81-180.

Lambert, J. M. & Dale, M. B., 1964. The use of statistics in

phytosociology. Adv. Ecol. Res. 2: 59-99.

Lambert, J. M. & Williams, W. T., 1962. Multivariate methods

in plant ecology. IV. Nodal analysis. J. Ecol. 50: 775-802.

Langford, A. N. & Buell, F. M., 1969. Integration, identity

and stability in the plant association. Adv. Ecol. Res.

6:83-135.

Leistner, O. A. & Werger, M. J. A., 1973. Southern Kalahari

phytosociology. Vegetatio 28: 353-399.

Looman, J. & Campbell, J. B., 1960. Adaptation of Sorensen’s

K (1948) for estimating unit affinities in prairie vegetation.

Ecology 41 : 409—416.

Louw. A. J., 1970. ji Ekologiese studie van Mopanie-veld noord

van Soutpansberg. D.Sc. (Agric.) thesis. Pretoria: Univ. Pre-

toria (Unpubl.).

Maas, F. M., 1959. Bronnen, bronbeken en bronbossen in

Nederland, in het bijzonder van de Veluwezoom. Meded.

Landbouwhogeschool Wageningen 59 (12): 1-166.

Marloth, R., 1908. Das Kapland, insonderheit das Reich der

Kapflora, das Waldgebiet und die Karroo, pflanzengeo-

graphisch dargestellt. Wiss. Ergebn. Dtsch. Tiefsee-Exped.

“ Waldivia' ’ 1898-1899. II, 3. Jena: Fischer.

Matuszkiewicz, W., 1948. Roslinnosc lasow okolic Lwowa

(The vegetation of the forests of the environments of Lvov).

Ann. Univ. Mariae Curie-Sklodowska, Sect. C. 3: 120-193.

(engl. summ.).

Meijer Drees, E., 1951. Enkele hoofdstukken uit de moderne

plantensociologie en een ontwerp voor nomenclatuurregels

voor plantengezelschappen. Rep. For. Res. Inst. Bogor,

Indonesia 51 : 1-218.

Meijer Drees, E., 1954. A tentative design for rules of phytoso-

ciological nomenclature. Vegetatio A: 205-214.

Meltzer, J. & Westhoff, V., 1942. Inleiding tot de planten-

sociologie. ’s Graveland: Breughel.

Molinier, R., 1960. Etude des biocoenoses du Cap Corse.

I. Vegetatio 9: 121-192. II. Vegetatio 9: 217-312.

Moll, E. J., 1969. An investigation of the plant ecology of the

Hawaan forest, Natal, using an ordination technique.

Bothalia 10:121-128.

Moore, J. J., 1962. The Braun-BIanquet system: a reassessment.

J. Ecol. 50: 761-769.

Moore, J. J., Fitzsimons, P., Lambe, E. & White, J., 1970.

A comparison and evaluation of some phytosociological

techniques. Vegetatio 20: 1-20.

Moravec, J., 1968. Zu den Problemen der pflanzensoziologi-

schen Nomenklatur. in Tiixen, R. (ed.): Pflanzensoziologi-

sche Systematik. Ber. Int. Symp. Stolzenau 1964. pp.

142-154. Den Haag: Junk.

Moravec, J., 1969. Die Vorbedingungen der pflanzensoziolo-

gischen Nomenklaturregeln. Acta Bot. Croatica 28 : 259-265.

Moravec, J., 1971a. Bericht liber die Tatigkeit der Nomen-

klaturkommission der Internationalen Vereiniging fur

Vegetationskunde 1969-1970. Vegetatio 22: 251-254.

Moravec, J., 1971 b. A simple method for estimating homo-

toneity of sets of phytosociological releves, Folio Geobot.

Phytotaxon. 6: 147-170.

Moravec, J., 1973. The determination of the minimal-area of

phytocoenoses. Folia Geobot. Phytotaxon. 8: 23—47.

Morris, J. W., 1969. An ordination of the vegetation of

Ntshongweni, Natal. Bothalia 10: 89-120.

Muller, P. J., Werger, M. J. A., Coetzee, B. J., Edwards,

D. & Jarman, N. G., 1972. An apparatus for facilitating the

manual tabulation of phytosociological data. Bothalia

10: 579-581.

Myre, M., 1960. Alguns tipos de graminais do sul da pro-

vincia de Mozambique. Anais Serv. Agric. Flor. Moq. 1960.

Myre, M., 1962. A grassland type of the south of the Mozam-

bique province. C. R. IV. Reun. AETFAT. pp. 337-362.

Myre, M., 1964. A vegetaqao do extremo sul da provincia de

Mozambique. Junta Inv. Ultramar , Est., Ens. e Doc.

110: 1-145.

Nordhagen, R., 1923. Om homogenitet, konstans og minimum

areal. Bidrag til den plantesociologiske diskussion. Nyt.

Mag.f. Naturvid. 61: 1-51.

322

ZURICH-MONTPELLIER METHOD OF VEGETATION SURVEY

Oberdorfer, E., 1968. Assoziation, Gebietsassoziation, geo-

graphische Rasse. in Tuxen, R. (ed.): Pflanzensoziologische

Systematik. Ber. int. Symp. Stolzenan 1964. pp. 124-141.

Den Haag: Junk.

Oberdorfer, E., 1970. Pflanzensoziologische Strukturprobleme

am Beispiel kanarischer Pflanzengesellschaften. in Tuxen,

R., (ed.): Gesellschaftsmorphologie. Ber. Int. Symp.

Rinteln 1966. pp. 273-281 . Den Haag: Junk.

Passarge, H., 1968. Neue Vorschlage zur Systematik nord-

mitteleuropaischer Waldgesellschaften. Feddes Repert.

77: 75-103.

Passarge, H. & Hofmann, G., 1964. Soziologische Arten-.

gruppen mitteleuropaischer Walder. Arch. f. Forstw.

13:913-937.

Passarge, H. & Hofmann, G., 1968. Pflanzengesellschaften des

nordostdeutschen Flachlandes. II. Pflanzensoziologie 16:

1-298.

Pfeiffer, H. H., 1957. Betrachtugnen zum Homogenitats-

problem in der Pflanzensoziologie. Mitt. Flor.-Soz.

Arbeit sgem. N. F. 6-7 : 103-1 1 1 .

Pfeiffer, H. H., 1962. Uber die Bewertung der Geselligkeits-

zahlen bei pflanzensoziologischen Aufnahmen. Mitt. Flor.-

Soz. Arbeitsgem. N. F.9: 43-50.

Phillips, J. F. V., 1931. Forest-succession and ecology in the

Knysna Region. Mem. Bot. Surv. S. Afr. 14: 1-327.

Pignatti, S„ 1968. Die Inflation der hoheren pflanzensoziologi-

schen Einheiten. in Tuxen, R. (ed.): Pflanzensoziologische

Systematik. Ber. Int. Symp. Stolzenan 1964. pp. 85-97.

Den Haag: Junk.

Poore, M. E. D., 1955. The use of phytosociological methods in

ecological investigations. I. The Braun-Blanquet system.

J. Ecol. 43: 226-244. II. Practical issues involved in an

attempt to apply the Braun-Blanquet system. J. Ecol.

43: 245-269. III. Practical applications. /. Ecol. 43: 606-651.

Poore, M. E. D., 1956. The use of phytosociological methods in

ecological investigations. IV. General discussion of

phytosociological problems. J. Ecol. 44: 28-50.

Poore, M. E. D., 1964. Integration in the plant community.

J. Ecol. 52 (suppl.): 213-226.

Poore, M. E. D., 1968. Studies in Malaysian rainforest. I.

The forest on triassic sediments in Jengka Forest Reserve.

J. Ecol. 56: 143-196.

Raabe, E. W., 1952. Liber den “Affinitatswert” in der Pflan-

zensoziologie. Vegetatio 4: 53-68.

Raunkiaer, C., 1934. The life forms of plants and statistical

plant geography. Oxford: Clarendon.

Rauschert, S., 1963. Beitrag zur Vereinheitlichung der sozio-

logischen Nomenklatur. Mitt. Flor.-Soz. Arbeitsgem.

N. F. 10: 232-249.

Romell, L.-G., 1925. Om inverkan av vaxtsamhallenas struktur

pa vaxtsamhallsstatistikens resultat. Bot. Notiser 1925:

253-308.

Romell, L.-G., 1926. Bemerkungen zum Homogenitatsproblem.

Svensk Bot. Tidskr. 20: 441-455.

Roberts, B. R., 1966. The ecology of Thaba Nchu. A statistical

study of vegetation/habitat relationships. D. Agric. thesis.

Pietermaritzburg: Univ. Natal. (Unpubl.).

Scamoni, A., Passarge, H. & Hofmann, G. 1965. Grundlagen zu

einer objektiven Systematik der Pflanzengesellschaften.

Feddes Repert. Beih. 142: 117-132.

Scheepers, J. C., 1969. A preliminary assessment of association-

analysis in the Kroonstad area. Proc. Grassld. Soc. Sth. Afr.

4:78-83.

Schmid, P. & Kuhn, N., 1970. Automatische Ordination von

Vegetationsaufnahmen in pflanzensoziologischen Tabellen.

Die Naturwiss. 57 : 462.

Schmithusen, J., 1968. Allgemeine Vegetationsgeographie.

in Obst, E. & Schmithusen, J. (ed.): Lehrbuch der

a/lgemeinen Geographic. 4. Berlin: De Gruyter. 3. Aufl.

Schwickerath, M., 1931. Die Gruppenabundanz, ein Beitrag

zur Begriffsbildung der Pflanzensoziologie. Bot. Jahrb.

64: 1-16.

Schwickerath, M., 1944. Die Transformation konstant-

methodisch erfasster Gesellschaften in die Charakter-

artenmethode und die Bedeuting und Gliederung des

Differentialartenbegriffs. Bot. Jb. 73: 361-374.

Schwickerath, M., 1954. Lokale Charakterarten — geograph-

ische Differentialarten. Verojf. Geotot. Inst. ETH , Stiftung

Riibel 29: 96-104.

Schwickerath, M., 1968. BegrifT und Bedeutung der geo-

gra hischen Differentialarten. in Tuxen, R. (ed.): Pflanzen-

soz logische Systematik. Ber. Int. Symp. Stolzenan

1964. pp. 78-84. Den Haag: Junk

Segal, S., 1965. Een vegetatie-onderzoek van de hogere water-

planten in Nederland. Wetensch. Meded. K.N.N. V. 57: 1-80.

Segal, S., 1969. Ecological notes on wall vegetation. Den Haag:

Junk.

Shimwell, D. W., 1971. Description and classification of

vegetation. London: Sidgwick & Jackson.

Spatz, G. & Siegmund, J., 1973. Eine Methode zur tabellaris-

chen Ordination, Klassifikation und okologischen Auswer-

tung pflanzensoziologischen Bestandsaufnahmen durch den

Computer. Vegetatio 28: 1-17.

Story, R., 1952. A botanical survey of the Keiskammahoek

District. Mem. Bot. Surv. S. Afr. 27 : 1-184.

Swan, J. M. A., 1970. An examination of some ordination

problems by use of simulated vegetation data. Ecology

51 : 89-102.

Taylor, H. C., 1969. A vegetation survey of the Cape of Good

Hope Nature Reserve. M.Sc. thesis. Cape Town: Univ. Cape

Town (Unpubl.).

Tidmarsh, C. E. M. & Havenga, C. M., 1955. The wheel-point

method of survey and measurement of semi-open grasslands

and Karoo vegetation in South Africa. Mem. Bot. Surv. S.

Afr. 29: 1-49.

Tuomikosi, R., 1942. Untersuchungen liber die Untervegetation

der Bruchmoore in Ostfinnland. I. Zur Methodik der

pflanzensoziologischen Systematik. Ann. Bot. Soc. Zool.-

Bot. Fenn. Vanamo 17: 1-203.

Tuxen, R., 1957. Die Schichten-Deckungsformel; zur Darstel

lung der Schichtung in Pflanzengesellschaften. Mitt.

Flor.-Soz. Arbeitsgem. N. F. 6-7 : 1 12-1 1 3.

Tuxen, R., 1969. Stand und Ziele geobotanischer Forschung in

Europa. Ber-Geobot. Inst. ETH, Stiftung Riibel 39: 13-26.

Tuxen, R., 1970a. Entwicklung, Stand und Ziele der pflanzen-

soziologischen Systematik (Syntaxonomie). Ber. Dtsch.

Bot. Ges. 83: 633-639.

Tuxen, R., 1970b. Pflanzensoziologie als synthetische Wissen-

schaft. Meded. Landbouwhogeschool Wageningen 12:

141-159.

Tuxen, R., 1970c. Einige Bestandes- und Typenmerkmale in der

Struktur der Pflanzengesellschaften. in Tuxen, R., (ed.):

Gesellschaftsmorphologie. Ber. Int. symp. Rinteln 1966.

pp. 76-107. Den Haag: Junk.

Tuxen, R. & Ellenberg, H., 1937. Der systematische und der

okologische Gruppenwert. Mitt. Flor-Soz. Arbeitsgem.

3: 171-185.

Van der Maarel, E., 1966a. Over vegetatiestructuren , -relaties

en - systemen , in het bijzonder in de Duingraslanden van

Voorne. Ph.D. thesis. Utrecht: Univ. Utrecht. (Unpubl.).

Van der Maarel, E., 1966b. Dutch studies on coastal sand dune

vegetation, especially in the Delta Region. Wentia 15 : 47-82.

Van der Maarel, E., 1970. Vegetationstruktur und Minimum-

Areal in einem Diinentrockenrasen. in Tuxen, R. (ed.):

Gesellschaftsmorphologie. Ber. Int. Symp. Rinteln 1966.

pp. 218-239. Den Haag: Junk.

Van Donselaar, J., 1965. An ecological and phytogeographic

study of northern Suriname savannas. Wentia 14: 1-163.

Van Vuuren, D. R. J., 1961. 7; Ekologiese studie van die

plantegroei van 7; noordelike en suidelike kloof van die

Magaliesberge. M.Sc. thesis. Pretoria: Univ. Pretoria.

(Unpubl.).

Van Zinderen Barker, Jr., E. M., 1971 Ecological investigation

on ravine forests of the Eastern Orange Free State (South

Africa). M.Sc. thesis. Bloemfontein : Univ. O.F.S. (Unpubl.)

Van Zinderen Barker Sr., F,. M. & Werger, M. J. A. 1973.

Vegetation and phytogeography of the high-altitude bogs of

Lesotho. Vegetatio: in press.

Volk, O. H. & Leippert, H., 1971. Vegetationsverhaltnisse im

Windhoeker Bergland, Siidwestafrika. J. S. W.A. Wiss.

Ges. 25 : 5-44.

Von Glahn, H., 1968. Der Begriff des Vegetationstypes im

Rahmen eines allgemeinen naturwissenschaftlichen Typen-

begriffes. in Tuxen R., (ed.): Pflanzensoziologische

Systematik. Ber. Int. Symp. Stolzenan 1964. pp. 1-14.

Den Haag: Junk.

Werger, M. J. A., 1972. Species-area relationship and plot

size: with some examples from South African vegetation.

Bothalia 10: 583-594.

Werger, M. J. A., 1973a. Phytosociology of the Upper Orange

River Valley, South Africa. A syntaxonomical and syn-

ecological study. Ph.D. thesis Univ. Nijmegen. Pretoria:

V. & R.

Werger, M. J. A., 1973b. An account of the plant communities

of Tussen die Riviere Game Farm, O.F.S. Bothalia 11:

165-176.

M. J. A. WERGER

323

Werger, M. J. A. & Coetzee, B. J., 1974. A phytosociological

classification of Aughrabies National Park. Bothalia

in prep.

Werger, M. J. A., Kruger, F. J. & Taylor, H. C., 1972.

Pflanzensoziologische Studie der Fybnos-Vegetation am

Kap der Guten Hoffnung. Vegetatio 24: 71-89. also as:

A phytosociological study of the Cape Fynbos and other

vegetation at Jonkershoek, Stellenbosch. Bothalia 10:

599-614.

Werger, M. J. A. & Leistner, O. A., 1973. Vegetationsdynamik

in der sudlichen Kalahari, in Tuxen, R. (ed.): Sukzessions-

forschung. Ber. Int. Symp. Rinteln 1973. Lehre: Cramer

(in press).

Westhoff, V., 1967. Problems and use of structure in the

classification of vegetation. Acta Bot. NeerL 15: 495-511.

Westhoff, V., 1968. Einige Bemerkungen zur syntaxonomischen

Terminologie und Methodik, insbesondere zu der Struktur

als diagnostischen Merkmale. in Tuxen, R. (ed.): Pflan-

zensoziologische Systematik. Ber. Int. Symp. Stolzenau

1964. pp. 54-70. Den Haag: Junk.

Westhoff, V. & Den Held, A. J., 1969. Plantengemeenschappen

in Nederland. Zutphen: Thieme.

Westhoff, V. & Van der Maarel, E., 1973. The Braun-

Blanquet approach of phytosociology, in Whittaker, R. H.

(ed.): Manual of vegetation Science. 5. Den Haag: Junk.

Whittaker, R. H., 1956. Vegetation of the Great Smoky

Mountains. Ecol. Monogr. 26: 1-80.

Whittaker, R. H., 1962. Classification of natural communities.

Bot. Rev. 28: 1-239.

Whittaker, R. H., 1967. Gradient analysis of vegetation.

Biol. Rev. 42 : 207-264.

Williams, C. B., 1943. Area and number of species. Nature

152: 264-267.

Wittgenstein, L., 1960. Philosophische Untersuchungen. Schrifte

I. Frankfurt: Suhrkamp.

Woods, D. B. & Moll, E. J. 1967. Multivariate analysis of

grasslands in the Three Rivers area. Natal. Proc. Grassld.

Soc. Sth. Afr. 2: 103-111.

Zohary, M., 1973. Geobotanical foundations of the Middle East.

Stuttgart: Fischer, and Amsterdam: Swets & Zeitlinger.

Zonneveld, I. S., 1960. De Brabantsche Biesbosch. Een studie

van bodem en vegetatie van een zoetwatergetijden delta.

Ph.D. thesis. Wageningen: Landbouwhogeschool, Wage-

ningen. (Unpubl.)

Bothalia 11, 3: 325-328 (1974)

An automatic bibliography indexing programme

J. W. MORRIS*

ABSTRACT

A relatively simple FORTRAN IV programme, designed for a small computer, for author and

key-word indexes to bibliographic records is described, and examples of output are given. It is com-

pared with some other systems. Suggested improvements to the programme are given.

INTRODUCTION

In order to make available for reference within the

Institute the 300 reprints and 200 literature references

collected during an overseas tour, some method of

indexing was required. A simple alphabetical list by

author, even within broad subject divisions, was

considered insufficient for efficient use of the collec-

tion. The collection covers in-depth the fields of

ecosystem modelling, multivariate analysis, informa-

tion retrieval and related subjects to a less comprehen-

sive degree.

It was, for several reasons, decided to try a com-

puterized system of indexing. Such a system would

satisfy the needs of author and subject indexes.

It would also serve as an example for a more com-

prehensive computer programme for the whole

Institute, or at least it would form the basis for

discussion of such a bibliographic index. Another

reason is that the system has been seen to work on a

number of similar bibliographies overseas. Other

considerations were that the data set was fairly small,

not open-ended, and the subject catergories were

limited in number although covering a fairly wide

field.

The indexing programme is described and then com-

pared with some other systems before conclusions are

made as to the utility of the programme.

COMPUTER PROGRAMME AND INPUT DATA

LAY-OUT

The programme, called biblo of some 500 cards

was written in standard fortran iv and implemen-

ted on an IBM/1130 computer with 32K words core

memory. At the time no other computer was available

and the programme was written specifically for a

small computer. As the sort routine took too long on

the small computer, the programme was modified

to run on an IBM/360. Even on the large computer

with optimization for speed and a faster sort routine,

the programme was time-consuming. Programme

listings and card decks are available, on request,

from the author.

Five card types, numbered 1 to 5, are used for data

entry and between five and seven cards are punched

for each reference. Examples of input data are given

in Fig. 1.

On each card, the accession number is punched in

the first four columns and the card type number in

the fifth. References and reprints may be accessioned

in any order and given any four-digit number.

In this case, reprints in the collection were numbered

consecutively from 1 000 upwards and references

from 2 000. The accession number was written on the

reprint and the reprints filed in numerical order to aid

retrieval. Column six of each card is left blank

(or a zero punched) except in card types 3 and 4 when

a ‘one’ indicates a continuation card of that type.

The first card (type 1) contains the author’s name

and initials starting in the seventh column. When

there is more than one author, each name is separated

by a comma. Periods between and after initials are

omitted. The second card (type 2) contains the date of

publication (year) punched in the seventh through

tenth columns. The title of the article or book is given

on the third card (type 3), starting in column seven.

One continuation card may be used if the title is too

long to fit on one card. The citation is given on the

fourth card type. A continuation card may be used.

The fifth card type contains title-enriching terms.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

COLUMN 11111111112222222222333333333344444444445555555555666666666677777777778

1234 56 7890 1 234567890 1 2 34 1>6 7890 1 234 567890 1 2 3456789 0 1 23456 78 90 1 2 34567890 12 34567890

123410ROSS JH.MORRIS JW

1234201971

123430PRINCIPAL COMPONENTS ANALYSIS OF ACACIA BURKE I AND ACACIA NIGRESCENS IN NA

123431TAL

12344080THALIA 10 (3), 437-450.

1 2 3450NUMER I CAL -TAXONOMY PCA

147010MATHER PM

1470201970

147030PRINC1 PAL COMPONENTS AND FACTOR ANALYSIS

147040C0MPUTER APPLICATIONS IN THE NATURAL AND SOCIAL SCIENCES NO. 10 UNIVERSITY

147041 OF NOTTINGHAM.

147050PC A FACTOR-ANALYSIS

209410R0NLF FJ,S0KAL RR

2094201965

2094T0C0EFFICIENTS OF CORRELATION AND DISTANCE IN NUMlRICAL TAXONOMY

209440UN IV. KANSAS SCI. BULL. 45 1-27.

209450

Fig. 1. — Example of three references punched on 80-column cards tor input to BIBLO.

326

AN AUTOMATIC BIBLIOGRAPHY INDEXING PROGRAMME

Each term is separated by a blank space and two-word

terms, like numerical taxonomy, are joined by a

hyphen (see Fig. 1).

COMPUTER OUTPUT

An alphabetical list of authors and the accession

numbers of their references is the first output from the

programme (Fig. 2). The words of the title and title-

enriching terms are then listed in alphabetical order.

PRINCIPAL AUTHORS

Fig. ?..— Part of an alphabetical list of principal authors and

reference numbers.

Words are listed together with the senior author of the

reference and the accession number (Fig. 3). Title

words with low information content, such as with,

for, and but, specified prior to running the pro-

gramme, are omitted from the list. Up to 200 words,

chosen by the user, may be excluded in this way.

The word index is a kwoc (Key-Word Out of Con-

text) one as the title must be referred to elsewhere

to determine the context of the word. Finally, listings

of the references in order of accession number and of

principal author are given (Fig. 4).

Fig. 3. — Part of KWOC index of title- and key-words.

SOME OTHER SYSTEMS

Details of some comparable bibliographic systems

are summarized in Table 1. They are described in the

following section.

* later modified for IBM/360.

J. W. MORRIS

327

1263 1971 PHIPPS JB

DENDROGRAM TOPOLOGY

SYST. ZOOL. 20 306-308

KEY WORDS * CLASSIFICAT ION-COMPAR I SCN

1220 1972 PHIPPS JB

STUDIES IN THE AP UNO I NF LLE AE - GRAMINEAE XI. TAXIMETRICS OF CHANGING CLASSIFICATI°NS

CANADIAN J BOT 50,787-802

KEY WORDS * P C A NUM-TAXONOMY

1219 1972 PHIPPS JB

STUDIES IN THF ARUND I NE LLE AE - GRAMINEAE

TR I STACHYO ID , AND DANTHQNCI PSCID GROUPS

CANADIAN J BOT 50,935-948

KEY WORDS * HIERARCHY PCA NUM-TAXONOMY

XIII. TAXIMFTPICS OF THF LOUDETIOIO,

1013 1969 PIELOU EC

ASSOCIATION TESTS VERSUS HOMOGENEITY TESTS - THEIR USE IN SUBDIVIDING QUADRATS

INTO GROUPS

VFGFTATIO IB, 4-18

1046 1972 PO I S SON ET P

RELATIONS DE VOISINAGE ENTRE VEGCTA'JX D UNt FORMATION HER EACEE CENSE - DISPOSITIF

EXPERIMENTAL ET PARAMETRES OE LA PRODUCTION

Ot COL PLANT 7,23-43

KEY WORDS * COMPETITION

1064 1965 PRINGLE JS

HYBRIDIZATION IN GENTIANA - GE NT I AN ACE AE - A RESUME OF JT CURTIS STUDIES

Wise ACA SCI, ARTS AND LFTT 54,283-293

1045 1968

KEY WORDS *

RAPP M, ROMANt F

CONTRIBUTION A L'FTUDE DU BILAN PE L'EAU DANS LES ECOSYSTEMES M E D IT ERR AN FEN S

OECOL PLANT 3,271-284

RAINFALL THROUGHFALL

Fig. 4. — Example of bibliography listed alphabetically by principal author. First number on each

author line is reprint accession number and second is date of publication.

A very simple computer-aided bibliography pro-

gramme was developed by Cedergren (1971). Alpha-

betical lists of principal authors and up to four key-

words per reference are produced along with the

citations and one line of comment. The comment may

be the title or additional key-words. Words in the

title (comment line) and secondary authors are not

indexed.

Bridges (1970) discusses the application of computer

processing to maintain a personal bibliography and

produce a sophisticated set of printed indexes. He

considers that personal bibliographies, especially

when well indexed, are important tools for scientific

research, teaching, administration and writing. Some

benefits of computer processing over index-card

bibliographies which he lists are: ease of making

multiple copies, ease of transport and the small

amount of assistance required by users from the com-

piler as the indexing criteria are based on a consistent

set of procedures. In addition to a straight biblio-

graphy listing and author index (principal and secon-

dary authors), a kwic (Key-Word In Context)

index of title and key-words and an index to

sources is given. The word index is in context as a

few words before and after the indexed term are also

printed. The same type of kwic index is given by

Biological Abstracts publications and other commer-

cial abstracting services.

Complex, automated documentation systems have

been developed by Burton et al. (1969) of the U.S.

Department of Agriculture and Creighton et al.

(1971 & 1972) of the Smithsonian Institution in

Washington. The user is free to design input, content

and format. A wide range of outputs, from searches

for specific authors, words or references, to indexes

of all kinds, is available with each system. As a large

computer is a pre-requisite, these systems, as well as

those of the commercial organizations, are not

discussed further.

CONCLUSION

Biblo is less sophisticated than famulus and

selgem (Table 1) but provides more information

in the way of key-word indexes and lists of secondary

authors than cherche. It is the only one capable of

being run on a small computer (Table 1). The pro-

gramme can be improved in a number of ways. At

present two lists of authors are produced (principal

authors and all authors) whereas a list of all authors

could be given with principal authors marked in

some way. A means of identifying new accessions to

the bibliography (c.f. cherche) could be built in.

A third improvement would be the linking of two-

word terms with a special character (such as an

ampersand) in the place of the hyphen used at present.

The special character would be suppressed in the

printing of the indexes and would improve readability.

More sophisticated indexing is also possible.

In that it made available for reference the literature

collected on my overseas tour in the form of alpha-

betical indexes of authors and key-words, it is con-

sidered a successful computer application. The

possibility of applying the programme to other

bibliographies is being actively considered. As

powerful computer facilities are available, the question

of whether to modify biblo or adopt selgem,

famulus or some other system, such as that of T. J.

Crovello (in Morris, 1973), should be carefully

considered.

328

AN AUTOMATIC BIBLIOGRAPHY INDEXING PROGRAMME

ACKNOWLEDGEMENT

The author is grateful to Dr D. Edwards for his

interest in the project and for his comments on this

paper.

OPSOMMING

'n Betreklik eenvoudige FORTRAN IV program vir outeur-

en sleutelwoordindekse tot bibliografiese gegewens, word vir ’n

klein rekenoutomaat beskryf. Voorbeelde van resultate word

gegee. Dit word met ander sisteme vergelyk. Moontlike ver-

beterings aan die program word aan die hand gedoen.

REFERENCES

Bridges, K. W. 1970. Automatic indexing of personal biblio-

graphies. Bioscience 20: 94-97.

Burton, Hilary D., Russel, R. M. & Yerke, T. B. 1969.

FAMULUS: a computer-based system for augmenting

personal documentation efforts. U.S.D.A. Forest Service

Research No te PS IV- 193 : 1-5.

Cedergren, R. J. 1971. Computer-aided bibliographies for

personal or group use. J. Chem. Documentation 11: 224-

226.

Creighton, R. A. & Crockett, J. J. 1971. SELGEM: A sys-

tem for collection management. Smithsonian Institution

Information Systems Innovations 2 (3): 1-24.

Creighton, R. A., Packard, P. & Linn, H. 1972. SELGEM

Retrieval: A general description. Smithsonian Institution

Procedures in Computer Science 1(1): 1-38.

Morris, J. W. 1973. Overseas advances in quantitative ecology

and computerised data banking. Unpublished Departmental

Report, Botanical Research Institute, Pretoria.

Bothalia 11,3: 329-347 (1974)

A phytosociological classification of the vegetation of the Jack

Scott Nature Reserve*

B. J. COETZEEt

ABSTRACT

The vegetation of the Jack Scott Nature Reserve in the Central Bankenveld Veld Type is classi-

fied chiefly by the Braun- Blanquet Table Method. Habitat features, physiognomy, total floristic

composition, differentiating species, woody plants and prominent grasses and forbs are presented for

each community. Characterizing habitat features, in order of importance for the communities, are :

exposure, soil texture, geology, slope, aspect, degree of rockiness and previous ploughing. The

classification correlates well with the major physiographic and climatic variation in the Reserve and

generally does not cut across main physiognomic types. The communities are potentially homo-

geneous management units.

INTRODUCTION

THE STUDY AREA

The Jack Scott Private Nature Reserve is located

50 km west-northwest of Pretoria in the Central

Variation of Acocks’ (1953) Bankenveld Veld Type

and covers approximately 3 100 ha (Fig. 1). The

Bankenveld covers approximately 2 356 800 ha or

2 % of the area of South Africa, and the Central

Variation contains no state or provincial nature

reserves. Conservation of vegetation in this veld type

variation is therefore at present dependent on manage-

ment policies of private landowners (Edwards,

1972a).

The purpose of this study was to provide a classi-

fication of the vegetation of the Jack Scott Nature

Reserve into ecological units correlated with stable

and permanent habitat conditions, distinguishing,

therefore, areas of uniform potential for manage-

ment purposes. To this end, the classification is

based mainly on the presence and absence of groups of

associated species which show similar response to the

environment (Goodall, 1953). Association-analysis

(Williams & Lambert, 1959 & 1960) was used to

obtain a first approximation of such a classification

and was then supplemented by the Braun-Blanquet

Table Method (Werger, 1973, 1974).

History and management

The study area has been managed as a nature reserve

for almost two decades, prior to which it belonged to

mining companies for more than half a century.

Occasional small diggings, old, deserted stone enclo-

sures and a few abandoned farmlands with charac-

teristic vegetation, bear witness to earlier human

activities. Historical records indicate that fires

occurred annually until the fifties, but are now

restricted by roads and firebreaks. Except for fire-

breaks, no fixed burning programme has been

followed although some land was recently burnt to

remove excess plant litter.

The Reserve is fenced with a game fence and is not

sub-divided into camps. Nevertheless, domestic

animals and game do not exert a homogeneous in-

fluence on the vegetation. About 100 ha of land

surrounding the living quarters of farm labourers are

overgrazed and trampled by their restricted number

of domestic stock.

* Partly based on an M.Sc. thesis, Universitity of Pretoria,

f Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria.

Fig. 1 . Geographical positions of the Jack Scott Nature Reserve and weather stations.

330 A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

Game species occurring mainly in the grasslands

are blesbuck, black wildebeest, zebra, springbuck

and gemsbuck. Areas around salt licks are extremely

trampled and heavily grazed by game. Firebreaks,

burned triennially on the same strip of land, attract

a considerable number of game and are heavily

grazed in some areas. A high dolomite grassland area

where these animals concentrate, is also overgrazed

and trampled, but other grasslands are generally

lightly grazed and in good condition. Giraffe, kudu,

bushbuck and nyala concentrate mainly in the savanna

and forest areas where they have little apparent

influence on the vegetation. Also occurring in the

reserve are eland, impala, blue wildebeest, mountain

reedbuck, ostriches, jackal, leopard and brown

hyaena. The ungulates are completely protected from

hunting and efforts are being made to eradicate

predators in order to increase the ungulate popula-

tion. Blesbuck are the most numerous of the game

species and numbered over 400 in January, 1972

(Mason, personal communication).

Exotic plants in the Reserve include a number of

weeds, mostly annuals, in disturbed areas; a small

number of woody plants, such as Salix babylonica ,

Acacia dealbata and Melia azedarach along streams,

an Opuntia species on some steep slopes and a single

Primus persica tree in a stand of natural vegetation;

a local sward of Pennisetum clandestinum and lawns

of Pennisetum clandestinum and Phylla nodiflora at

two recreation areas.

Apart from a relatively limited area around residen-

tial quarters, overgrazed and trampled areas in dolo-

mite grassland, and a few small, isolated, heavily-

trampled areas, the vegetation over most of the

Reserve is thus uncultivated and little disturbed by

human and animal agency.

Physiography

Geological formations in the Reserve are of the

Transvaal System where it forms the southern rim of

the Bushveld Basin with rocks dipping north at 15°

to 30° (Haughton, 1969). Formations occurring in

the Reserve (Fig. 2) are, from below and from south

to north:

(i) Dolomite and chert of the Dolomite Series;

(ii) shale and quartzite of the Timeball Hill Stage

of the Pretoria Series; and

(iii) shale of the Daspoort Stage of the Pretoria

Series.

Local post-Bushveld Complex diabase intrusions

occur within these formations.

The Reserve is situated at an altitude of between

1 310 m and 1 615 m. The southern part is gently

undulating grassland on dolomite, mainly between

1 510 m and 1 570 m with isolated chert-capped hills.

In the north there is the wide west-east valley of the

Skeerpoort River. The southern slopes of the upper

western part of the valley are formed by chert. In the

east the valley cuts through the chert into dolomite

with chert occurring above 1 430 m. The southern

rim of this valley is a long west-to-east-running chert

ridge that is interrupted by a tributary of the Skeer-

poort River, originating on the dolomite highlands.

Other tributary valleys of the Skeerpoort River,

which cut into the southern slopes of the Skeerpoort

Valley, are sheltered from cold southern winds by the

chert ridge.

The gently undulating floor of the Skeerpoort

Valley is mainly shale with few rock outcrops and

soil that is generally only a few centimetres deep.

On the northern side of the valley are south-facing

shale slopes of the Timeball Hill Stage. This ridge is

the southernmost of three parallel ranges which mark

the transition from Witwatersrand high grassland to

Northern Transvaal wooded vegetation. The northern

boundary of the Reserve runs along the summit of

this ridge with the highest points between 1 520 m

and 1 580 m in the west and between 1 360 m and

1 420 m in the east. The rocky summit and northern

slopes of this range are formed by layered quartzite,

which is more resistant to weathering than the shale

buried underneath south-facing talus slopes.

Diabase intrusions are largely limited to the

northern part of the Reserve where the dykes are

associated with long north-south depressions and

drainage lines in shale and quartzite.

Soils in the Reserve are generally residual and

shallow. Weathering of dolomite by solution of

calcium and magnesium carbonates produces a

silica rich sandy loam soil with manganese and iron

concretions. Chert weathers mechanically to a silica-

rich sandy loam soil. Shale, consisting mainly of

iron-rich aluminium silicates produces a typical

yellowish to reddish clay loam soil with many termite

mounds. Quartzite, being a metamorphic sandstone,

forms a sandy soil mixed to a greater or lesser degree

with finer clay material from the surrounding shale.

Soil from the iron and magnesium-rich diabase is

mixed with shale and quartzite colluvium. Soils

associated with diabase dykes have a clay loam

texture.

Climate

The Reserve is situated in a summer rainfall area

with temperate summers and frosty winters, classified

as Cwb-climate in Koppen’s system (Trewartha,

1954).

Average total radiation measured in Pretoria on a

horizontal surface varies from between 560 and 580

cal/cm2/day in December, January and February to

between 340 and 360 cal/cm2/day in May, June and

July. Direct radiation on north-facing slopes at 26° S

latitude is higher than on south-facing slopes, parti-

cularly during winter months. This difference in

winter radiation becomes even more pronounced

since diffuse radiation, which is not aspect dependent,

accounts for only 20% of the total radiation in

June as against 36% in January (Shulze, 1965).

Data from Pretoria (West End), Pretoria (Forestry)

and Irene (Fig. 1), suggest that temperatures are

lowest in June and July, with average monthly minima

about -5° C and an extreme minimum of about

-8,7° C (Weather Bureau, 1954). The average monthly

maxima in the hottest months from November to

February can be as high as 33,7° C and the extreme

maximum recorded at Pretoria (Forestry) is 37, 8°C

during November. The average period with at least

occasional frost stretches from about April to Sep-

tember with frost usually occurring daily from May

until August (Schulze, 1965).

Considerable variation in micro- and meso-climate

may be expected in the Reserve due to variation in

topography, soils and vegetation structure. The high-

lying southern part of the Reserve and tributary

valleys of the Skeerpoort River that have been cut

through the chert divide, are exposed to cold catabatic

winds flowing from the south during many clear,

calm winter nights. Frost in valley bottoms suggests

the occurrence of thermal belts in sheltered valleys

with summits cooling more than slopes and cold air

collecting in valley bottoms. Temperatures on north-

facing slopes will generally be higher than on south-

facing slopes as a result of differences in amount of

radiation received.

B. J. COETZEE

331

TgdS

T^Q

T 2

Shale -

Quartzite

Shale

Dolomite

— Daspoort Stage

x

> Tirryeba II Hill Stage

)

Chert

‘Pretoria

Series

Do Ionite

Series

>Iransvaa

System

di

Diabase (Post Bushveld Complex)

^ Dip -\P- Fault

SCQ,e 0 (OOO 2 0 00 3000 4000 5000

r ^ -- < i i metres

Fig. 2. Geological map of the Jack Scott Nature Reserve (source: unpublished map. Dept, of Mines,

Pretoria).

17749-9

332 A PHYTOSOCIOLOG 1CAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

Some idea of the wind regime in the Reserve can be

obtained by extrapolating from Pretoria and Jan

Smuts weather stations (Schulze, 1965). Northerly

to easterly winds similar to those experienced in

Pretoria are expected to predominate slightly during

summer, a strong predominating south-westerly

component, during winter. The strongest winds in

the region are usually south-westerly to southerly

gusts of short duration, accompanying thunder storms,

with the strongest winds occurring mainly in spring

(September to November). Whirlwinds due to strong

instability and convection may occur on hot summer

days.

Rain occurs mainly as showers and thunderstorms

of short duration. Average annual rainfall, based on

data from Maryvale, Hartebeeshoek and Kromdraai

(Fig. 1), is between 670 and 745 mm with 85-90%

falling from October to March (Weather Bureau,

1965). Hail storms occur, on the average, on four to

five days per year, mainly during the rainy season

(Schulze, 1965).

Previous botanical descriptions

The Jack Scott Nature Reserve lies near the

boundary between two major vegetation types

recognized by authors of all broad vegetation descrip-

tions of the region. Elements of both types occur in

the Reserve. Pole Evans (1936) describes these two

vegetation types as Mixed Grassland, on the one

hand, and Evergreen and Deciduous Trees and Bush

on the other, Adamson (1938) as Highveld Grass-

land and Bushveld, and Acocks (1953) as Bankenveld

and Sourish Mixed Bushveld.

A more detailed classification and 1 : 30 000 vegeta-

tion map by Wells (1964) divided the vegetation on

a structural and ecological basis into forest patches,

tree communities, grasslands and marshy areas.

Tree communities, separated floristically from one

another on the basis of dominant trees, were Protea

caffra veld. Protea rouppelliae veld. Acacia cajfra

veld, Acacia karroo veld and Burkea africana veld.

Grasslands were divided physiographically into

Valley Grasslands, Rocky Grassland, and Highland

Grassland.

METHODS OF SAMPLING AND SYNTHESIS

Samples were divided pro rata on an area basis

among a number of geomorphological vegetation-

physiognomic units delineated on aerial photos and

were placed at random within these units. A minimum

of ten samples per unit was regarded as sufficient.

After placing 188 samples, only old farmlands were

undersampled (i.e. contained less than 10 samples)

and they were therefore supplemented with four

additional samples.

Subsequent to sampling the 192 sites, an additional

37 sampling sites were placed in the following classes,

shown by a field check of geological-physiognomical

variation to be undersampled: two sites in wooded

vegetation on dolomite; five on wooded vegetation on

shale without quartzite talus, and ten sites each on

diabase, quartzite and magnetite-quartzite.

The total number of samples taken in the reserve

was 229, giving an average sampling intensity of one

per 13,5 ha.

Scheepers (personal communication) found a

rectangular quadrat of 4 m X 4 m the most efficient

for highveld vegetation that is similar to large parts

of the Reserve. The same size and shape were used in

this study.

Physiographic data recorded include geology’

rockiness, geomorphology, aspect and slope. Surface

stones and soil colour, texture and stoniness were also

noted for the top 5 cm of soil. Percentage projected

canopy cover and height for each stratum, as well as

total percentage canopy cover, were recorded.

All vascular plants in quadrats were recorded and

additional species occurring within 2 m of quadrats

were noted separately. For 184 quadrats a Braun-

Blanquet cover-abundance value (Kershaw, 1964)

was estimated for each species. A list of trees and

shrubs occurring within the same stand of vegetation

near sampling sites, was included as part of each

releve.

A first approximation of the final grouping of

quadrats was obtained by normal association-

analysis (Williams & Lambert, 1959, 1960), of all

releves, and species occurring in eight or more

releves. To obtain a polythetic classification as well

as an inverse analysis, releves, and also species

occurring in two or more releves were further arranged

following the Braun-Blanquet approach (Coetzee,

1972, 1974; Muller et al, 1972; Werger, 1973, 1974).

PLANT COMMUNITIES

The hierarchy of plant communities recognized in

the Jack Scott Nature Reserve is based on Braun-

Blanquet tables which are summarized in a Summary

or Roman Table (Table 1.).

In Table 1 each community, based on several

releves, is reduced to a single column. The roman

figures in the matrix are ratings of constancy of each

species in releves of each community on a five point

scale (Werger, 1973):

I 1 to 20% constancy.

II 21 to 40% constancy.

III 41 to 60% constancy.

IV 61 to 80% constancy.

V 81 to 100% constancy.

The number of releves summarized in a column is

indicated in the table.

No attempt is made at fixing the ranks of syntaxa

and the general diagnostic value of species, which are

important considerations when naming syntaxa in

accordance with the nomenclatural principles of the

Braun-Blanquet School (Werger, 1973). Since priority

is generally recognized in the Braun-Blanquet School,

information from a wider area is desirable in order to

avoid premature ranking and naming of syntaxa.

The naming of vegetation types is here based on a

combination of prominent species, vegetation struc-

ture and habitat, and the names are diagnostic within

the reserve when applied in their hierarchical context.

Structural terms are those defined by Fosberg

(1967). Communities are classified structurally and

functionally into Fosberg’s system with the code for

each formation given in brackets after each formation

name.

Lists of woody plants occurring in communities

include plants noted around quadrats in the same

stands of vegetation.

Grass, forb and suffrutescent species with a cover

abundance of two or more (i.e. very abundant or

covering 25% or more of the area) in at least 25%

of the releves representing a community, are listed

in Table 2 as prominent species in the grass stratum

of that community.

FIGURE 3

LEGEND

POSITION OF THE JACK SCOTT

NATURE RESERVE

SCHEMATIC REPRESENTATION OF

COMMUNITIES, PHYSIOGNOMY AND

HABI TAT

PHYSIOGNOMY

grassland mssophyll micro phyll various

.formations

>■

CD

O

_J

O

LU

CD

Chert

Quartzite

Dolomite

Var ious

formations

Diabase

Shale

Old

lands

Sandy loam

Clay loam

lo o

o o

I I Acacia caff ra savanna on diabase

1-2*1 A cocio caffra - Chrysopogon montonus savanna

1-2-2 Acacia caf fro - Ruellio cordoto savanna

•.*.*•

_» » • .

E3

3*1 frotto coffrq evergreen savannas on shale

3- 2 A Grasslands on shale

m

T +

+ +

4 1

B Deciduous savannas on shale

a

Sffl

4-3

Vegetation on quartzite outcrops (unmappable dimensions1 )

4- 2 Steppe savanna and grassland on broken quartzite

5- Grasslands on old farmlands

o°o° ° ol

O o 0° ° < ^

o ° o

p- O O

6

KrantZi riparian and kloof communities

I

AIR PHOTOGRAPHY 1968 (JOB NO- 603/68)

Botanical survey section

BOTANICAL RESEARCH INSTITUTE

PRIVATE BAG X 1 01 , PRETORIA.

SOIL TEXTURE

o

—

's/as A

■ ■

SAM HOI TAT 303V

Previ

Th

boun

reco|

tions

the I

veget

hand

on tl

land

and :

-V3C !l't! : J.. ~

oopi opa oqa QQ£_ .s — Q

( Y m AB3CTHA0) VMM* Al - J M

eveASYe

M

San

amon

physi(

were {

of ter

After

and t

additi

rectanj

for hi£

of the

this st

vegeta

shale

diabas

The

was 2\

per 13

Schf

B. J. COETZEE

333

Table 1. Constancy of Species in Communities

334 A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

Table 1. Constancy of Species in Communities ( continued )

Community Number

Monocymbium ceresiiforme

(Nees) Stapf

Leucas neuflizeana Courb

Digitaria brazzae (Franch.)

Stapf

Diant bus mooiensis F. N.

Williams

Pygmaeothamnus zeyheri (Sond.)

Robyns

Crassula transvaalensis (Kuntze)

K. Schum

Hemizygia pretoriae (Giirke)

M. Ashby

Sporobolus eylesii Stent & Rat-

tray

Babiana hypogea Burch

Hermannia lancifolia Szyszyl. . .

Asclepias stellifera Schltr

Oxygonum dregeanum Meisn . .

Helichrysum galpinii Schltr. &

Moeser

Psammotropha mucronata

(Thunb.) Fenzl

Helichrysum setosum Harv

Bulbostylis oritrephes (Ridl.)

C.B.C1

lndigofera pretoriana Harms . . .

Vernonia natalensis Sch. Bip.

Poly gala amatymbica Eckl. &

Zeyh

Digitaria diagonalis (Nees) Stapf

Lasiosiphon burchellii Meisn. . .

Grader ia subintegra Mast

Senecio erubescens Ait

Polygala sp. (B.J.C. No. 795).. .

Phyllanthus incurvis Thunb. . . .

Cyperus obtusiflorus Vahl

Oldenlandia herbacea (L.)Roxb.

Vernonia monocephala Harv. . .

Vernonia staehelinoides Harv . .

lndigofera comosa N.E. Br

Crassula nodulosa Schonl

Burkea africana Hook

Dichapetalum cymosum (Hook.)

Engl. & Prantl

Pavetta zeyheri Sond

Rhyne hosia nitens Benth

Sphedamnocarpus pruriens (Juss.)

Szyszyl

Diospyros lycioides Desf. subsp.

£Mer&e/(Kuntze)DeWint

S uler a caerulea Hiern

Aristida transvaalensis Henr. . . .

Cymbopogon marginatus (Steud.)

Stapf ex Burtt Davy

Commelina africana L

Tapiphyllum parvifrlium (Sond.)

Robyns

Rhus magalismcntana Sond. . . .

Cynodon dactylon (L.) Pers. . . .

Walafrida densiflora Rolfe

Solatium incanum L

Hyparrhenia hirta (L.) Stapf . . .

Wahlenbergia calendonica Sond.

Sporobolus africanus (Poir.)

Robyns & Tournay

Conyza podocephala DC

Hypericum aethiopicum Thunb.

subsp. sonderi (Bredell)

N. Robson

Hermannia depressa N.E.Br.

Eriosema sa lignum E. Mey.

Verbena officinalis L

Euphorbia sp. (B-J.C. 804)...

Ill

B. J. COETZEE

335

Table 1. Constancy of Species in Communities ( continued )

336 A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

Table 1. Constancy of Species in Communities ( continued )

Table 2. PROMINENT (P) SPECIES IN GRASS STRATUM (EXCLUDING KRANTZ, RIPARIAN AND KLOOF COMMUNI-

TIES)

B. J. COETZEE

337

Table 2. PROMINENT (P) SPECIES IN GRASS STRATUM (EXCLUDING KRANTZ, RIPARIAN AND KLOOF COMMUNI-

TIES) ( continued )

111 the vegetation map of the Reserve (Fig. 3)

mapping units correspond to vegetation types at

various levels of the classification. Numbering of

communities in the legend corresponds to their

numbering in the text.

1. Diabase and dolomite vegetation

This vegetation includes savanna and grass forma-

tions on diabase and dolomite and is characterized by

the presence of the Eustachys mutica species group

(Table 1). Eustachys mutica , which occurs in all

these communities and has a high fidelity for this

vegetation, is associated with soils derived from

diabase and dolomite that are both rich in bases.

Eustachys mutica is negatively associated with the

Loudetia simplex group (Table 1), which is charac-

teristic of the poorer soils on chert, shale and quart-

zite.

1 . 1 Acacia caflfra Savanna on diabase

This community occurs on the clay-loam soils of

long depressions on diabase dykes, where the soil

is moister and presumably, relatively rich in bases

(Fig. 4).

The structure is generally deciduous thorn savanna

(112/3) with Acacia caffra as the most prominent tree.

The Artemisia ajfra species group (Table 1) charac-

terizes this community.

Woody plants of the community, with percentage

presence in releves, are:

1 .2 Acacia caffra Savanna of sheltered valleys

Two types of Acacia caffra Savanna, characterized

by Enneapogon scoparius , Kalanchoe paniculata and

Setaria lindenbergiana , occur in sheltered valleys.

Chrysopogon montanus , which belongs to the same

species group (Table 1) is characteristic of the more

xeric of the two communities, occurring on very steep,

north-facing slopes. The geology in these valleys is

generally a mixture of dolomite and chert. Slopes are

rocky and Pellaea calomelanos and Vellozia retinervis,

typical of rocky slopes, are markedly constant in

both communities.

338 A PHYTOSOCIOLOG ICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

1.2.1 Acacia caffra-Chrysopogon montanus Savanna

This community occurs on warm slopes that are

steeper than 30%, are rocky, north-facing, and

underlain by dolomite and chert in sheltered valleys.

The community belongs to Fosberg’s (1967) micro-

phyllous, deciduous thorn shrub savanna (1K2/5),

with Acacia caffra and Ozoroa paniculosa generally

the more prominent trees, and is characterized by

Chrysopogon montanus (Fig. 5).

The following shrub and tree species were recorded

in releves of the community with percentage presence

indicated in brackets:

1.2.2 Acacia caffra-Ruellia cordata Savanna

This mesic community of sheltered Acacia caffra

valley vegetation, occurs on the lower parts of north-

facing slopes of less than 30% and on south-facing

slopes that have over 12% slope. Structurally, the

community includes deciduous thorn savanna (112/3)

and microphyllous deciduous thorn savanna (1K2/5),

with Acacia caffra as the most prominent woody

species (Fig. 5).

Several species, notably the dominant grasses

Setaria perennis , Eragrostis curvula and Themeda

triandra occur in common with many exposed cool,

higher-altitude communities. These grasses, as well

as the following species, differentiate the cooler

mesic from the warm xeric variation of the sheltered

valleys:

Fig. 4. Diabase dykes, indicated

by arrows, with contrasting

vegetation, in the Timeball

Hill Stage.

Fig. 5. Acacia caffra Savanna of

sheltered valleys. The herb

stratum of the Acacia caffra-

C hrysopogon montanus

Savanna on the right, has

a white appearance. Acacia

caffra - Rtiellia cordata

Savana grows in the rest of

the valley,

B. J. COETZEE

339

1.3 Grasslands and Cussonia paniculata Savanna on

dolomite

Vegetation characterized by the Trichoneura grandi-

glumis species group occurs on the gently undulating

dolomite highlands in the south of the Reserve and on

the steep valley slopes formed by tributaries of the

Skeerpoort River that have cut through the chert.

1.3.1 Grasslands on gently undulating dolomite high-

land

Two grassland communities on virtually pure

dolomite high terrain with slopes of less than 12%

(Fig. 6) are differentiated from grassland on steep,

mixed dolomite-chert slopes by the following species:

Triraphis andropogonoides, Callilepis leplophylla.

Euphorbia sp. cf. E. pseudotuberosa. Aster muricatus

Elephantorrhiza elephantina.

(a) Grassland on extensive dolomite outcrop

The rocky habitat of this community is illustrated

in Fig. 7.

Structurally this community falls into Fosberg's

(1967) seasonal grass steppe (2G2/1). Ozoroa panicu-

losa occasionally occurs as an isolated shrub. The

community is characterized by Fingerhuthia sesleriae-

formis, Euphorbia rhombifo/ia and Sutera burkeana

and is further differentiated from the less rocky

grassland on gently undulating dolomite highland

by the constancy of the following species:

Ipomoea obscura

Polygala amatymbica

Sporobolus pectinatus

Eragrostis capensis

Ruellia cordata

Nidorella hottentottica

Brachiaria serrata

Diheteropogon amplectens

Elyonurus argenleus

Setaria perennis

Eragrostis curvula

Rhynchosia totta

Elephantorrhiza elephantina

Asparagus suaveolens

Diplachne biflora

Vernonia oligocephala

Sporobolus pectinatus

Themeda triandra

Eustachys mutica

Heteropogon contortus

Trachypogon spicatus

Elephantorrhiza elephantina is a prominent forb.

(b) Grassland on non-rocky dolomite areas

Rocks are few and inconspicuous in this community

which in heavily-grazed parts has the appearance of

seasonal orthophyll meadows (short grass) (1M2/1),

and in lightly-grazed areas may be classified as

seasonal orthophyll tall grass (1L2/1).

Fig. 6. Grasslands on gently

undulating dolomite high-

lands.

Fig. 7. Grassland on banks of

dolomite rock.

340 A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

Euphorbia inaequilatera, an apophyte, is charac-

teristic, probably as a result of heavy grazing. Some

of the following species, which differentiate this

community from the previous, are more constant in

this community, probably also as a result of grazing

pressure :

Geigeria burkei subsp. burkei

Anthericum cooperi

Digit aria tricholaenoides

A nthericum fasciculatum

Tristachya rehmannii

Digitaria monodactyla

Aristida congesta subsp.

congesta

Gazania krebsiana subsp.

serrulata

Lotononis tenella

Parinari capensis

Cymbopogon excavatus

Senecio venosus

Eragrostis racemosa

Fimbristylis ovata

Large areas in this community are heavily grazed,

especially firebreaks and the summit areas that are

constantly overgrazed and severely trampled. Elyonu-

rus argenteus increases in heavily-grazed parts of the

grassland. This is clearly evident when grazed areas

within the Reserve are compared with ungrazed

areas outside. On firebreaks in lower parts of the

area where there is hardly any grazing, it can be seen

that within three years litter accumulates to such an

extent that fire becomes necessary to maintain the

present vegetation.

1.3.2 Grassland and Cussonia paniculata Savanna

on steep dolomite slopes

This vegetation type occurs on steeper-than-12%

rocky slopes of dolomite mixed with chert, that are

exposed to cold southerly weather (Fig. 8). It includes

seasonal orthophyll tall grass (1L2/1) and mesophyl-

lous, evergreen sclerophyll savanna [111/2 (a)] with

scattered C. paniculata trees on south-facing slopes

in these exposed valleys.

The following species differentiate grassland and

Cussonia paniculata Savanna of steep dolomite slopes

from the two grassland communities on gently undul-

ating dolomite highland:

(i) Pellaea calomelanos and Vellozia retinervis ,

which occur on rocky slopes;

(ii) Microchloa caffra, which occurs on rocky

dolomite slopes in exposed areas as well as in mesic

areas of sheltered valleys;

(iii) Tristachya rehmannii , which is a dominant grass

in this exposed valley vegetation, and Bulbostylis

burchellii. Both species are found on sandy loam

soils of exposed areas, including highlands;

(iv) Schizachyrium sanguineum and Urelytrum

squarrosum , prominent in this vegetation (Table 2)

and Loudetia simplex , are three species found on acid

soils derived from chert, quartzite and shale; and

(v) Rhynchelytrum setifolium , another prominent

grass in this community, and one which occurs in all

communities of the Reserve, is markedly more

constant on steep slopes of exposed valleys than on

the gently undulating dolomite highland.

2 . Chert vegetation with or without scattered small

rock outcrops

Rolling chert grassland covers most of the thick

chert cap of the high central-western part of the

Reserve whereas Protea caffra Savanna occurs on

chert hills and summits in the more strongly-dissected

central and central-eastern parts where the chert is

thinner (Fig. 9).

The chert grasslands have a number of charac-

terizing species that occur neither in the chert savan-

nas nor in dolomite and shale vegetation. But the

only species that differentiate chert savannas from

chert grassland occur also on diabase, shale and

dolomite, indicating a relationship of the chert

savanna to these latter communities. The difference

between grasslands and Protea caffra Savanna on

chert slopes of less than 30% cannot be related to

any observed differences in geology, soil texture,

rockiness, stoniness, altitude, slope, aspect or disturb-

ance, other than the previously mentioned thickness

of the underlying chert cap. On chert slopes of more

than 30%, Protea caffra Evergreen Savanna occurs

on south-facing slopes, and deciduous savanna on

north-facing slopes. The chert vegetation is charac-

terized by the Monocymbium ceresiiforme species

group (Table 1).

2.1 Grasslands on chert and chert-rich dolomite

Seasonal orthophyll tall grass (1L2/1) on chert

and mixed chert-dolomite soils are characterized by

Digitaria brazzae and Oxygonum dregeanum and

further differentiated from savannas on chert by:

(i) Digitaria monodactyla and Lotononis tenella ,

which also occur on steep dolomite slopes in exposed

valleys with chert-capped summits, where a chert

influence also exists, and in non-rocky dolomite

grasslands, where the effect of the dolomite parent

material has diminished as a result of leaching; and

(ii) Kohautia amatymbica and Pentanisia august i-

folia, which occur also on shale.

Fig. 8. An exposed dolomite

valley originating on the

highland. Summits are chert

capped.

B. J. COETZEE

341

2.1.1 Grasslands on chert-rich dolomite

About 50% of the dolomite region is paricularly

rich in chert and has grasslands characterized by

Hemizygia pretoriae and Hermannia lancifolia. These

grasslands are further differentiated from grasslands

on pure chert by Digitaria tricholaenoides, Geigeria

burkei subsp. burkei and Babiana hypogea.

(a) Grassland on rocky chert-rich dolomite areas

The rocky variation of chert-rich grassland, found

on moderate to steep slopes, is differentiated from the

non-rocky variation by:

Loudetia simplex

Urelytrum squarrosum

Gazania krebsiana subsp.

serrulata

Hermannia lancifolia

Babiana hypogea

Pearsonia sessilifolia

Vellozia retinervis

Dianthus mooiensis

Monocymbium ceresiiforme

Sporobolus pectinatus

Diheteropogon amplectens

Anthospermum rigidum

Sphenostylis angustifolia

North-facing slopes differ floristically from south-

facing slopes.

(i) The community of north-facing slopes is

characterized by Babiana hypogea and differentiated

from the community of south-facing slopes by:

Leucas neuflizeana , which is restricted to chert and

chert-rich dolomite soils;

Polygala hottentotta , which occurs also on dolomite

of the highlands and cooler, exposed valleys;

Crabbea angustifolia , which is restricted to dolomite

and chert; and

Diplachne biflora and Setaria perennis, with wider

distributions.

(ii) The community of south-facing slopes is

characterized by Helichrysum galpinii and Sporobolus

eylesii and differentiated from the north-facing

slope community by:

Helichrysum adscendens, characteristic of the vege-

tation on south-facing shale, chert and chert-rich

dolomite slopes;

Pentanisia angustifolia and Thesium transvaalense ,

which occur mainly on highland chert and shale soils;

and

Dicoma gerrardii, which occurs widely on dolomite

of exposed highlands and on chert

The distributions of the differentiating species of

this community and comparison with those of the

former community, suggest that the two communities

occur on soils with differing nutritional status. The

community of south-facing slopes is differentiated by

species that are associated with poor soils and cool,

mesic conditions, whereas differentiating species of

the community on north-facing slopes have wider

distributions on dolomite.

(b) Grassland on non-rocky, chert-rich dolomite

areas.

The non-rocky, chert-rich grassland is differen-

tiated from the two communities on rocky, chert-

rich dolomite by Anthericum cooperi , which charac-

terizes grassland on non-rocky dolomite and non-

rocky, chert-rich dolomite, and Digitaria brazzae ,

which is otherwise restricted to chert-derived soils.

2.1.2 Grasslands on chert

Grasslands on chert are differentiated from grass-

lands on chert-rich dolomite by Nolletia rarifolia ,

which is characteristic of the vegetation on soils

derived from chert and shale.

(a) Grassland on rocky chert areas.

This grassland is distinguished from grasslands on

non-rocky chert areas by Loudetia simplex , Vellozia

retinervis and Schizachyrium sanguineum.

(b) Grassland on non-rocky chert areas.

Grassland on non-rocky chert areas is differen-

tiated from grassland on rocky chert by Crabbea

angustifolia , Cassia biensis , Dicoma gerrardii and

Diplachne biflora.

2.2 Savannas on chert.

Savannas on chert are differentiated from grass-

lands on chert and chert-rich dolomite by:

(i) Ruellia cordata, which also occurs in wooded

vegetation of chert areas containing many boulders,

on diabase dykes and in all dolomite communities;

(ii) Heteropogon contortus, which also occurs on

diabase dolomite and shale;

(iii) Pellaea calome/anos, which is very constant on

rocky dolomite slopes and in chert areas with many

boulders; and

(iv) Oxalis obliquifolia and Senecio serra , occurring

also on diabase and in some dolomite and shale

communities.

2.2.1 Deciduous savanna on chert

Deciduous, broad, sclerophyll shrub savanna

(1K2/2) occurs on north-facing chert slopes that

have a slope angle of more than 30%. The community

is distinguished from evergreen savannas on chert by

Ochna pulchra and Sphenostylis angustifolius. The

Fig. 9. Rolling chert grassland

in the foreground with

Protea caff'ra savanna on

chert hills in the back-

ground.

342 A PH YTOSOCIOLOGICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

following woody species were noted in and near

quadrats with the percentage presence given in

parentheses:

F agar a c open sis

Of the four samples representing this community,

three are very fragmentary stands, which explains the

low percentage presence of woody species.

2.2.2 Protea caflfra Evergreen Savanna on very steep

south-facing chert slopes

Evergreen broad sclerophyll shrub savanna ( 1 K1 /2)

occurs in exposed areas on south-facing chert slopes

with more than 30% slope angle. The community,

represented by only three releves, is characterized by

Helichrysum setosum, Mohria caffrorum and Senecio

serra and also differentiated from other chert savan-

nas by:

. (i) Crassula transvaalensis, Monocymbium ceresii-

■’ or me and Hemizygia pretoriae , otherwise restricted to

grasslands; and

(ii) Helichrysum adscendens and Alloteropsis semia-

lata , which are most constant in communities of

steep to very steep, south-facing slopes underlain by

shale and chert-rich dolomite.

Protea caffra is the constant dominant woody

species and Loudetia simplex the dominant grass.

Other woody species recorded are Vangueria infausta ,

Bequaertiodendron magalismontanum , Nuxia congesta.

Pavetta zeyheri and Rothmannia capensis.

2.2.3 Protea caffra Evergreen Savanna on chert

slopes of less than 30%

Species that differentiate this community from

Protea caffra Evergreen Savanna that occurs on very

steep, south-facing slopes, are: Chaetacanthus burchel-

lii , Setaria perennis, Themeda triandra, Parinari

capensis.

Protea caffra , Diplachne biflora and Indigofera

burkeana differentiate this evergreen savanna from

deciduous savanna on chert.

Two variations of the community occur, the major

variation and an ecotonal variation occurring on the

transition to Acacia caffra Savanna of sheltered

valleys. Both variations are evergreen broad sclero-

phyll shrub savanna (1K1/2) with Protea caffra

forming the upper stratum and Ochna pulchra, which

is an abundant shrub in the grass layer.

(a) Protea caffra — Bulbostylis burchellii Savanna

This main variation of the community is differen-

tiated from the ecotonal variation by:

Leucas neuflizeana Loudetia simplex

Crahbea angustifolia Panicum natalense

Bulbostylis burchellii Parinari capensis

Tristachya rehmannii

Woody plants, with percentage presence, are:

Protea caffra (90) Vangueria infausta (20)

Ochna pulchra (60) Ficus ingens (10)

Ozoroa paniculasa (30)

(b) Ecotonal Protea caffra-Rhynchosia totta Savanna

This ecotonal variation, represented by three

releves, occurs in more sheltered and dolomite-

rich habitats than the major variation, and borders

on the Acacia caffra Savanna of sheltered dolomite

valleys. The following species, which do not occur

elsewhere on chert, differentiate the ecotonal from the

major variation of the community:

(i) Rhynchosia totta , which occurs also on dolomite

and shale;

(ii) Cymbopogon excavatus , which occurs also on

dolomite, shale and diabase; and

(iii) Asparagus suaveolens , occuring also on dolo-

mite and on north-facing slopes with abundant

quartzite boulders.

Woody plants are:

Protea caffra (100% Maytenus heterophylla (33)

presence)

Ozoroa paniculasa (100) Mundulea sericea (33)

3 . Shale vegetation

Vegetation on shale includes evergreen savannas on

steep south-facing slopes and on summits, deciduous

savanna on steep, rocky slopes, at low altitudes, and

grassland on gentle slopes and steep, non-rocky

north-facing slopes (Fig. 10). These communities are

characterized by the Bulbostylis oritrephes species

group (Table 1 )

3 . 1 Protea caffra Evergreen savannas on shale

The Protea evergreen broad sclerophyll shrub

savannas (IK 1/2) on shale include three communities

on summits and south-facing slopes with more than

12% slope of the Timeball Hill Stage. The upper

stratum consists only of low Protea caffra trees. The

Fig. 10. Protea caffra Savanna on

steep south facing shale

slopes (A) and grassland

on gently undulating shale

slopes (B).

B. J. COETZEE

343

characteristic species of these savannas is Vernonia

natalensis, and they are differentiated from other

communities on shale by Protea caffra. Pearsonia

sessilifolia and Rhynchosia totta.

3.1.1 Protea caffra — Alloteropsis semialata Savanna

of non-rocky, upper pediment slopes and non-

rocky summits

This community occurs on very steep, upper

pediment slopes and summits on soil mainly derived

from shale but usually mixed with quartzite colluvium.

The community is characterized by Senecio erubescens

and a Polygala sp. (BJC795) and is differentiated from

other evergreen savannas on shale by:

(i) Alloteropsis semialata and Helichrysum adscen-

dens, both occuring also on very steep, south-facing

chert slopes and south-facing, chert-rich dolomite

slopes; and

(ii) Thesium transvaalense, which has a wide distri-

bution on chert and chert-rich dolomite.

Protea caffra Savanna of non-rocky, upper slopes

and summits is also differentiated from Protea caffra

Savanna on lower pediment slopes by Polygala

amatymbica and Loudetia simplex.

3.1.2 Protea caffra — Digitaria diagonalis Savanna of

non-rocky , lower pediment slopes

Digitaria diagonalis is characteristic of this savanna

which is differentiated from other evergreen savannas

on shale by Setaria flabellata , characteristic of deep

clay-loam soils on diabase and shale. The community

is also distinguished from the Protea- Alloteropsis

Savanna of upper pediment slopes and summits, by

Cymbopogon excavatus and Setaria perennis.

Prominent grasses are:

Brachiaria serrata Diheteropogon amplectens

Themeda triandra Elyonurus argenteus

Traehypogon spicatus Panicum natalense

Cymbopogon excavatus Setaria perennis

Eragrostis racemosa

3.1.3 Protea caffra — Rhynchelytrum setifolium Savan-

na of rocky summits and steep , rocky, south-

facing slopes

Species that differentiate this community from

other evergreen savannas on shale are:

(i) Lasiosiphon burchellii, Phyllanthus incurvis and

Cyperus obtusiflorus, all three being virtually restricted

to this community, to shale grasslands and to decid-

uous savannas on shale;

(ii) Peltaea calomelanos and Ve/lozia retinervis

which occur widely on rocky areas; and

(iii) Anthospermum rigidum and Rhynchelytrum seti-

folium, two widely-occurring species.

3 . 2 Grasslands and deciduous savanna on shale

Deciduous shrub savanna (1K2) on shale is repre-

sented by only three releves from steep rocky slopes

in low areas of the north-eastern part of the Reserve

and includes microphyllous deciduous thorn shrub

savanna (1K2/5) with Acacia caffra as the most

prominent woody plant, and deciduous, broad

sclerophyll shrub savanna (1K2/2) with Mundulea

sericea. Seasonal orthophyll tall grass (1L2/1),

represented by 19 releves, occurs on non-rocky,

north-facing slopes steeper than 12% and on all

slopes with a less-than-12%-slope angle, with or

without shale outcrops.

Grasslands and deciduous savanna on shale are

characterized by Phyllanthus incurvis and further

differentiated from evergreen savannas on shale by:

Crabbea angustifolia Tristachya rehmannii

Bidbostylis burchellii Heteropogon contortus

Dicoma gerrardii

The savanna and grassland formations are not

separable on the species content of the 4 x 4 m samples

taken. Woody species in the savanna formations

are:

Mundulea sericea (100% presence)

4. Vegetation with abundant large boulders , on

quartzite outcrop and on massive chert outcrops

This vegetation is characterized by the Aristida

transvaalensis species group (Table 1). Quartzite

communities, which occur on the summits and north-

facing slopes of the Timeball Hill Stage, are less

exposed to cold winds from the south than the chert

communities.

4 . 1 Wooded and grass formations on steep quartzite

sheet outcrop

Soils on the very narrow quartzite outcrops with

slopes of 12-30%, have a clay-loam texture due to

the influence of the surrounding shale. Stands of

vegetation in this habitat are fragmentary. Physio-

gnomic variations include deciduous orthophyll

shrub steppe savanna (2E2/2), Protea caffra ever-

green, broad sclerophyll scrub (2B1/2), and seasonal

Fig. 1 1 . Broken rocky quartzite

landscape with steppe

savanna and steppe ragss-

land.

344 A PH YTOSOCIOLOG1CAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

grass steppe (2G2/1). The community is negatively

differentiated by the absence of differentiating species

of other communities on sheet outcrops, of massive

outcrops or with abundant large boulders.

Woody plants include:

Tapiphyllum parvifolium Diospyros lycioides (11)

(78% presence) Rhus pyroides (11)

Pro tea caffra (44) Strychnos pungens (11)

Bequaertidendron mag a- Vangueria infausta (11)

lismontanum (22) Ximenia caffra (11)

Combretum mode (22)

Burkea africana (11)

4.2 Steppe savanna and grass steppe on broken

quartzite

Broken, rocky quartzite with abundant boulders is

a conspicuous feature of the landscape in which this

community occurs (Fig. 1 1 ). The soil is a sandy loam.

Structural variations include Protea caffra evergreen

sclerophyll shrub steppe savanna (2E1 /2) and seasonal

grass steppe (2G2/1), and are characterized by the

Crassula nodulosa species group (Table 1). Woody

species recorded include Protea caffra and Ximenia

caffra , both present in only one out of five releves.

4 . 3 Steppe scrub and scrub on very steep quartzite

sheet outcrop

Stands of the community are small and fragmentary

and occur on slopes that are steeper than 30% and

have a clay-loam soil. The vegetation includes

Burkea africana deciduous sclerophyll steppe scrub

(2B2/2); Tapiphyllum parvifolium deciduous sclero-

phyll scrub (1B2/3); and Protea caffra evergreen,

broad sclerophyll steppe scrub (2B1/2). The Burkea

africana species group (Table 1 ) is characteristic of

this community.

Woody species are:

Dichapetalum cymosum , a sufifrutescent plant, is

prominent in the grass stratum and has a constancy

of 75%.

4.4 Steppe scrub , steppe savanna and savanna with

abundant boulders on chert

The vegetation of very rocky, broken chert out-

crops includes Protea caffra evergreen sclerophyll

shrub savanna (1K1/2), deciduous sclerophyll shrub

steppe savanna (2E2/2) and deciduous sclerophyll

steppe scrub (2B2/2). The community is charac-

terized by Oldenlandia herbacea and is differentiated

from other very rocky communities by:

4 . 5 Steppe scrub and scrub on massive chert outcrops

The community is characterized by Sutera caerulea

and includes deciduous sclerophyll steppe scrub

(2B2/2) and deciduous sclerophyll scrub (1B2/3),

with plants growing in crevices and very shallow

humic sandy loam soils amongst massive solid chert

rocks.

The following woody species were recorded:

Tapiphyllum parvifolium

(75% presence)

Bequaertiodendron ma-

galismontanum (50)

Canthium gilfillanii (50)

Dombeya rotundifolia (50)

Nuxia congest a (50)

Vangueria infausta (50)

Acacia caffra (25)

Acokanthera oppositi-

folia (25)

5 . Grasslands on abandoned farm lands

These secondary grasslands include seasonal ortho-

phyll tall grass (1L2/1) and seasonal orthophyll

meadows (short grass) (1M2/1) and are characterized

by the Cynodon dactylon species group (Table 1).

Twelve of the thirteen quadrats in abandoned lands

were on clay-loam soils. These quadrats do not show

a very localized variation on sandy-loam soils ob-

served to be dominated by Eragrostis gummiflua ,

which is a prominent grass on such soils. The two

variations that were sampled, are:

5 . 1 Cynodon dactylon — Themeda triandra Grassland

This community, representing a later stage in the

secondary succession, is differentiated from Cynodon

dactylon-Conyza podocephala grassland on abandoned

lands, by:

Themeda triandra Diheteropogon amplectens

Eragrostis racemosa Schizachyrium sanguineum

Rhynchelytrum repens Senecio venosus

Elephantorrhiza elephantina Pygmaeothamnus zeyheri

5.2 Cynodon dactylon — Conyza podocephala Grass-

land

This earlier serai community is differentiated from

the previous by:

Hyparrhenia hirta Sporobolus africanus

Conyza podocephala Eriosema salignum

Hermannia depressa Verbena officinalis

Hypericum aethiopicum subsp. sonderi

6. Krantz, riparian and kloof communities

The vegetation in these habitats was not sampled

by quadrats and is not included in Table 1. The

following is a general description of this vegetation.

6 . 1 Krantz vegetation

Krantz communities include a range of formations,

namely evergreen, succulent steppe scrub (2B1/5)

and evergreen, succulent shrub steppe savanna (2E1/4)

in which Aloe mutabilis is a prominent species;

deciduous sclerophyll steppe scrub (2B2/2); and

deciduous sclerophyll shrub steppe savanna (2E2/2).

A chasmophyte, Ficus ingens, is conspicuous on

krantzes, and Aloe mutabilis occurs generally in

rock crevices and in shallow soil against shale krantzes

along the Skeerpoort River and its tributaries (Fig.

12). Abundant shrubs and trees include:

Bequaertiodendron maga- Maytenus tenuispina

lismontanum Nuxia glomerulata

Dombeya rotundifolia Pappea capensis

Combretum molle

Less frequent or locally abundant trees and shrubs

are:

Acacia ataxacantha

Brachylaena rotundata

Buddleia saligna

Canthium gilfillanii

Cassine burkeana

Dovyalis zeyheri

Fagara capensis

Ficus soldanella

Grewia occidentalis

Heteromorpha arborescens

Lannea discolor

Osyris lanceolata

Pittosporum viridiflorum

Rhus leptodictya

Rhynchosia nitens

Tapiphyllum parvifolium

Urera tenax

B. J. COETZEE

345

Fig. 12. Ficus ingens , front left, and Aloe mutabilis growing on

shale krantzes.

Grasses and forbs include:

Aristida transvaalensis

Commelina africana

Crassula setulosa

Dichapetalum cymosum

Hibiscus lunariifolius

Kalanchoe rotundifolia

Kyllinga melanosperma

Pellaea calomelanos

Pellaea viridis

Setaria lindenbergiana

Tragia rupestris

6.2 Riparian herbaceous vegetation

Riparian herbaceous vegetation includes the

following communities:

(a) Broad-leafed seasonal submerged meadows

(1P2/2) in which Berula thunbergii , a rooted plant

with floating leaves, is abundant.

(b) Evergreen orthophyll graminoid marsh (1M1/2)

occurring as dense narrow stands of Juncus oxycarpus

fringing moving water.

(c) Tall evergreen graminoid marshes (1L1/2)

composed of dense stands of either Cyperus sexangu-

laris or Juncus punctorius , occur as narrow, fringing

zones along water courses (Fig. 13).

(d) Seasonal fern meadows (1N2/2) which occur

as dense stands of Pteridium aquilinum.

(e) Seasonal orthophyll tall grass (1L2/1), in which

dense stands of Hyparrhenia tamba occur on flat

areas, as well as on slopes and along streams. In

some places isolated individuals of riparian-forest

tree and shrub species, or of Acacia karroo occur in

this tall grass. In localized places Hyparrhenia tamba

also forms a dense undergrowth in riparian forest

and in Acacia karroo deciduous thorn scrub (thorn-

hush) (1B2/4).

6 . 3 Riparian scrub

This scrub occurs on flat areas and gentle slopes

in and along streams, generally in less sheltered areas

than riparian forest. The vegetation includes:

(a) evergreen orthophyll swamp scrub [1B1/3

(a)], with Salix woodii as a prominent shrub or small

tree.

(b) evergreen reed swamp [IB1/3 (b)], where

Phragmitis australis is dominant;

(c) evergreen narrow sclerophyll swamp [1B1/3

(d )], with Rhus lancea as prominent shrub or small

tree, particularly on low islands;

(d) broad-leafed evergreen orthophyll scrub [1B1/1

(a)], with Buddleia salviifolia, along streams; and

(e) deciduous thorn scrub (thornbush) (1B2/4),

in which Acacia karroo is dominant, along streams.

Herbaceous undergrowth species in these scrub

communities include:

Bromus willdenowii

Carex cernua

Cyperus denudatus

Ehrharta erecta

Hyparrhenia tamba

Juncus ejfusus

Kyllinga melanosperma

Leersia he.xandra

Scirpus inclinatus

6.4 Riparian and kloof forests

These forests are found along streams that are

sheltered by high, steep slopes, on deep soils at the

foot of krantzes, or on steep slopes in sheltered

kloofs. The vegetation includes evergreen broad

sclerophyll forest (1A1/6) with Olea africana as

dominant tree species, and dry-season deciduous

Fig. 13. A seam of riparian herbaceous communities with

Cyperus sexangularis and Juncus punctorius and riverine

forest of Acacia ataxacantha, Celtis africana and other

trees behind.

346 A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE VEGETATION OF THE JACK SCOTT NATURE RESERVE

forest (1A2/3) (Fig. 13) with Acacia ataxacantha ,

Celtis africana and Combretum erythrophyllum among

the dominant tree species. Other trees and shrubs in

the forests are:

Acacia caff r a

Acokanthera oppositifolia

Brachylaena rotundata

Buddleia saligna

Cassine burkeana

Cassinopsis ilicifolia

Diospyros lycioides subsp.

guerkei

Diospyros whyteana

Dovyalis zeyheri

Ehretia rigida

Euclea crispa

Fagara capensis

Grewia occidentalis

Hex mitis

Kiggelaria africana

Maytenus heterophylla

Maytenus tenuispina

Maytenus undata

Mimusops zeyheri

Myrica serrata

Myrsine africana

Olinia emarginata

Osyris lanceolata

Pittosporum viridiflorum

Rhamnus prinoides

Rhus dentata

Rhus eckloniana.

Rhus pyroides

nities. Table 1 is incomplete in respect of woody

plants because of the small size of the quadrats, but

the percentage presence of woody plants in commu-

nities is provided in the text. Structural, functional

formations of krantzes, along streams and in kloofs

are largely specific for different habitats. Species

lists for these physiognomic-ecological communities

show them to be floristically well differentiated.

Finally, from this study of a small area in the

Central Bankenveld, it is concluded that this Veld

Type Variation includes a large number of plant

communities, which represent natural ecosystems of

which none are officially conserved (Edwards, 1972a

& b).

ACKNOWLEDGEMENTS

Herbs and lianes include:

Achyranthes sicula

Asparagus asparagoides

Asparagus buchananii ( liane )

Asparagus laricinus

Asparagus setaceus

Asparagus virgatus

Bidens pilosa

Bowiea volubilis {liane)

Clematis brachiata (liane)

Commelina africana

Cyperus albostriatus

Cyphostemma cirrhosum

Ehrharta erecta

Galinsoga par vi folia

Helinus integrifolius (liane)

Hyparrhenia tamba

Kalanchoe rotundifolia

Priva meyeri

Rhoicissus trident at a (liane)

Rhynchosia caribaea (liane)

Setaria chevalieri

Setaria lindenbergiana

Sida rhombifolia

Stipa dregeana

Teucrium capense

DISCUSSION AND CONCLUSIONS

The sampling strategy used in this study adequately

represented floristic types that cover large areas in the

Reserve. Important types that cover small areas and

are related to obvious physiographic variation are

also usually well sampled. The floristic classification

presented accommodates all but two releves of the

sampled variation. The classification is related to con-

spicuous habitat differences. At a broad level, vegeta-

tion differences are associated with soil texture, geol-

ogy, exposed highlands or low, sheltered valleys, very

rocky areas, and abandoned farm lands. Subdivisions

into lesser communities are associated with slope,

aspect, degree of rockiness and geology.

The plant communities recognized are distinguished

by their total floristic composition and are poly-

thetically separable on the presence and absence

of groups of associated species that are related to

specific environmental factors. The main vegetation

types that include a number of communities are

distinguished on the basis of their total floristic

composition, which includes less conspicuous but

good ecologically-differentiating species and are

also fairly well differentiated by prominent species as

shown in Table 2. With the exception of fragmentary

stands of very rocky areas, basis communities of

the hierarchy do not cut across the structural-func-

tional vegetation formations.

Due to their floristic as well as structural-functional

homogeneity, the communities are basic units that

should be taken account of for management of the

wild life. Descriptions and tables of communities

include basic plant ecological information required

for management units such as main habitat features,

physiognomy, a fairly complete list of species,

differentiating species, woody species and prominent

species in the grass stratum. Typical sites for further

observation and extrapolation over mapped units

should be selected with reference to this basic infor-

mation on communities. Distributions and habitat

preferences of particular grass and forb species of

significance for wild life management can be obtained

by referring to Table I and to descriptions of commu-

The Department of Agricultural Technical Services

made this study possible by seconding me to the

University of Pretoria. The study was undertaken

under supervision of Dr G. K. Theron of the Uni-

versity's Department of General Botany and Dr D.

Edwards from the Botanical Research Institute.

I wish to thank Col. J. Scott for the opportunity of

studying the vegetation of his nature reserve and for

facilities provided.

To the staff members of the Botanical Research

Institute my appreciation for their indispensible

assistance, in particular: Dr M. J. A. Werger for his

keen interest and valuable discussions; Mrs E. van

Hoepen and other members of the herbarium for

naming plant specimens; Messrs N. Jarman, J. W.

Morris, J. C. Scheepers and J. Vahrmeijer for helpful

comments; and Mrs M. L. Jarman, Mrs G. E. Thomas

and Miss L. Teversham for technical assistance,

including drawing of figures.

I also gratefully acknowledge the contributions

made by the following persons: Mr D. R. Mason

for his unpublished geological and physiognomic

maps of the reserve which he put at my disposal;

Dr H. Jansen and Messrs B. Meinster and W. C. P. de

Vries for geological information; and Dr P. J. Gouws

for help on the soils.

OPSOMMING

Die plantegroei van die Jack Scott-natuurreservaat is hoof-

saaklik met behulp van die Braun-Blanquet-metode geklassifi-

seer. Die habitatkenmerke, fisionomie, totale floristiese samestel-

ling, differensierende spesies, borne en struike en prominente

grasse en kruide van elke gemeenskap word verstrek. Blootstel-

ling, grondtekstuui, geologie, helling, aspek, mate van rotsagtig-

heid en vroeere bewerking van die grond is, in volgorde van be-

langrikheid, onderskeidende habitatkenmerke. Die klasifikasie

korreleer goed met die uitstaande fisiografiese en klimatologiese

variasie in die Reservaat en individuele gemeenskappe sluit oor

die algemeen nie meer as een van die hoof fisionomiese tipes in

nie. Die gemeenskappe is potensieel homogene bestuurseenhede.

REFERENCES

Acocks, J. P. H. 1953. Veld types of South Africa. Mem. Bot.

Sttrv. S.Afr. 28: 1-192.

Adamson, R. S. 1938. The vegetation of South Africa. London:

British Empire Vegetation Committee.

Coetzee, B. J. 1972. 'n Plantsosiologiese studie van die Jack

Scott-natuurreservaat. M.Sc. thesis, University of Pretoria.

Coetzee, B. J. 1974. Improvement of association-analysis

classification by Braun-Blanquet technique. Bothalia

11: 365-367.

Edwards, D. 1972a. Conservation areas in relation to Veld

Types. Unpubl. ms., Bot. Res. Inst., Pretoria, (in press

Kudu).

Edwards, D. 1972b. Botanical survey and agriculture. Proc.

Grassld. Soc. Sth. Afr.l: 15-19.

Fosberg, F. R. 1967. A classification of vegetation for general

purposes in C. F. Peterken (ed.) IBP , Handbook 4. Guide to

the checksheet for IBP areas p. 73-120. Oxford: Blackwell

Scientific Publications.

B. J. COETZEE

347

Goodall, D. W. 1953. Objective methods for the classification

of vegetation. I. The use of positive interspecific correlation.

Aust. J. Bot. 1 : 39-63.

Haughton, S. H. 1969. Geological history of Southern Africa.

Cape Town: Geol. Soc. S. Afr.

Muller, P. J., Werger, M. J. A., Coetzee, B. J., Edwards,

D. & Jarman, N. G. 1972. An apparatus for facilitating

the manual tabulation of phytosociological data. Bothalia

10:579-582.

Pole Evans, I. B. 1936. A vegetation map of Southern Africa.

Mem. Bot. Surv. S. Afr. 15: 1-23.

Schulze, B. R. 1965. W.B. 28. Climate of South Africa. 8.

General survey Pretoria. Govt. Printer.

Trewartha, G. T. 1954. An introduction to climate. New York.

McGraw-Hill.

Weather Bureau. 1954. W.B. 19. Climate of South Africa .

Climate Statistics. Pretoria. Govt. Printer.

Weather Bureau. 1965. W.B. 29. Climate of South Africa. 9.

Average monthly rainfall up to the end of 1960. Pretoria.

Govt. Printer.

Wells, M. J. 1964. Die plantegroei van die Jack Scott-natuur-

reservaat. Fauna Flora, Transv. 15: 17-25.

Werger, M. J. A. 1973. Phytosociology of the upper Orange

River valley , South Africa. A syntaxonomical and syneco lo-

gical study. Pretoria: V & R.

Werger, M. J. A. 1974. On concepts and techniques applied in

the Zurich-Montpellier method of vegetation survey.

Bothalia 1 1 (3) :

Williams, W. T. & Lambert, J. M. 1959. Multivariate methods

in plant ecology. I. Association analysis in plant communi-

ties. /. Ecol. 47: 83-101.

Williams, W. T. & Lambert, J. M. 1960. Multivariate methods

in plant ecology. II. The use of an electronic digital com-

puter for association-analysis./. Ecol. 48: 689-710.

17749-10

Bothalia II, 3: 349-353 (1974)

Progress in the computerization of herbarium procedures

J. W. MORRIS*

ABSTRACT

Herbarium automation projects at Cape Town (A. V. Hall), Notre Dame (T. J. Crovello),

Harvard (L. I. Nevling), Ottawa (J. H. Soper), Brisbane (S. L. Everist) and the British Antarctic

Survey (D. M. Greene) as well as the proposed system at New York (G. T. Prance) are described in

detail. It is found that data are coded for projects involving small numbers of specimens while, for

large systems, data are entered uncoded where possible. It is noted that not one automation project

has failed and that all users were enthusiastic about the future of such operations.

The need for a large system in South Africa is outlined and the uses to which it could be put are

listed. The system planned for use is briefly described.

INTRODUCTION

Herbaria are data banks in that they store inform-

ation about plant specimens in a retrievable form,

in much the same way that reference libraries store

information on their shelves. The specimens and

associated labels in herbaria constitute an inventory

of botanical information and a history of the plants

from the times of the first collectors until the present.

With the traditional method of filing specimens it is,

for practical reasons, possible to retrieve only limited

kinds of information. Theoretically it is possible to

compile from specimens a generally valuable map of

the distribution of each plant species, to prepare

floristic lists for any given region, to plot changes in

distribution and abundance which have taken place

and to collate a vast amount of information about the

habitats, flowering time and other properties of the

plants. The major difficulty, however, has been the

time needed to retrieve information from each label

and to extract the elements needed for any particular

purpose. With the advent of electronic data processing

(EDP) and, more important, the growing availability

of computers to botanists, much more in the way of re-

trieval is now possible. Since Crovello & MacDonald's

(1970) pioneer survey of biological data-processing

much progress has been made. The present survey is

not nearly as complete as that by Crovello & Mac-

Donald, as far as coverage is concerned, but, where

appropriate, more detail of each application is given-

This survey is concerned only with specimen data,

and taxon data banks, such as that envisaged by the

Flora North America Program (Shetler, 1971), are

not considered.

Following a survey of herbarium automation,

motivation for and the main objectives of a proposed

data bank for South African Herbaria are set out,

and the means by which these objectives can be

obtained are considered.

REVIEW OF SOME OPERATIONAL AND PROPOSED

SYSTEMS

Hall (1972 a & b) in Cape Town was the first known

curator in South Africa to apply EDP to a herbarium

collection. His main design criteria were flexibility of

output and simplicity of input; the programme to be

compatible with a wide range of computers and to be

operational on fairly small ones. He found that as a

number of elements on specimen labels were common-

ly repeated for many specimens, most elements

could be coded before capture with one- to four-

digit numbers, or alpha-numeric codes. Captured

data, in the sense used in this paper, are those words

and phrases accessioned by the computer’s storage

device in such a form that they may be retrieved in a

variety of ways by programming the computer.

The words and phrases may be entered in full, or

represented by code symbols or numbers as done by

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria, South Africa.

Hall. He considered it easier to edit a list of all pos-

sible elements once, give elements codes and then use

the codes when accessioning. The work of repeatedly

punching and editing precise citations written out in

full suggested little reward for much effort. Data were

entered by standard computer card, taxon by taxon,

each taxon being headed by one card with the

taxonomic information for the species, followed by

one specimen card for each herbarium sheet. In

addition to the data fields listed in Table 1, Hall

coded seven hierarchical supra-generic levels. Eleven

programme modules, written in fortran iv (except

for the plotter routines), control transactions within

his bank. The programme was initially written for

an IBM/ 1130 16K computer with 500K disc store

and a plotter, but has been expanded recently (Hall,

1972b) for operation on a Univac 1 106 131K computer

under remote terminal control. Retrieval by virtually

any combination of coded elements is possible.

Hall’s system was designed primarily as an aid, at

the specimen level, to research in systematics. Likely

benefit from such a bank include rationalisation

of specimen collection by pin-pointing poorly-

collected areas and taxonomic groups. Important data

can readily be assembled on the travels of the major

early collectors and the places they visited, and of

herbarium type-specimen holdings. The task of writing

floras of specially circumscribed areas will be assisted

(A project with this aim has been initiated; A. V. Hall,

pers. comm. 1973). Hall considers that, for the

immediate future, the chief users of a data banking

system such as he described, are likely to be those

making regular use of certain special collections.

Other users are those concerned with the balanced

expansion of collections of critical groups, perhaps as

part of a monographic study. The costs involved

prohibit data-banking on a major scale at present

(Hall, 1972a). The software and hardware used by

Hall limit the size of his data bank to its present order

of magnitude.

The label data from 65 000 specimens have been

captured by Crovello (1972) in the first, large, opera-

tional herbarium computerisation project in the

Greene Herbarium (NDG). A Friden 7102 flexowriter

was used over a period of two years to code the data

fields listed in Table 1 and provide machine-readable

paper-tape input for each specimen. The computer in

use is an IBM 370/158 and all programmes are written

in pl/1, although IBM’s sort/merge routine is

used as well.

In a detailed motivation for computerization,

Crovello mentioned the richness of NDG in type

material and the taxonomic and environmental

importance of the pre-urbanization era during which

most specimens were collected. As Greene, the princi-

pal collector, left no notebooks, information as to

where and when he collected was not available but

can be retrieved from a data bank based on his

350

COMPUTERIZATION OF HERBARIUM PROCEDURES

Table 1. Summary of label-data captured for various projects ( P=proposed , U= captured unabbreviated, C= captured in code form,

T = typed but not captured).

specimens. As it is a closed collection, data accumula-

tion for the project would not continue indefinitely

but could be a pilot project for the use of EDP in

biology. Naturally, corrections and new annotations

on specimens being returned from loan will be input

as received. Greene’s collection was too large for

manual retrieval, but not so large as to make the pro-

ject unwieldy and too expensive. A complete list of

types in the Greene herbarium would aid taxonomists

at other herbaria who were seeking types. Crovello

concluded his motivation by stating that the special

attributes of the Greene herbarium and realisation of

some of the ways that the application of EDP to this

collection would assist interested biologists, led him

to the conclusion that the project was so rich in

actual and potential results as to justify the effort.

Gomez-Pompa & Nevling (1970 & 1973) are

collaborating in the preparation of a machine-aided

Flora of Veracruz (Mexico). Their first attempt at

data-capture involved recording all information from

herbarium labels so that it could be computer-

retrieved. Elaborate “dictionaries” were drawn up for

coding the information (for example, 130 flower-

colour classes in two languages). The attempt was

abandoned after a trial run as it took too long to code

a herbarium specimen (L. E Nevling, pers. comm.

1972). The present system uses a standard collector's

label as base (Table 1 ). In addition, associated species,

ambient weather conditions, flower colour, local

names and whether a plant is annual or perennial,

are recorded, where available, in free format. Data

are punched onto computer cards and herbarium

labels are considered a useful by-product of the system.

The computer being used for this project is a Bur-

roughs B6500 and the programmes are written in

Algol. Many of the computer programmes have, in

the absence of more highly-trained programmers,

been written by students. The data have been used to

produce a provisional check-list of the flora. It is

planned to retrieve other data from the data bank and

351

J.W. MORRIS

to correlate it with environmental data. Distributions

are mapped by means of a Calcomp plotter. Very

useful suggestions for curators intending to use EDP

for the preparation of a flora are given by Gomez-

Pompa & Nevling (1973).

Data processing for an automated flora of Ontario

was started in 1963 by Soper. His intention was to

study, firstly the distribution and affinities of the flora

of Ontario, and then to extend the project to the whole

of Canada (Crovello & MacDonald, 1970; Soper,

1969; Soper & Perring, 1967). Although this was

probably the first herbarium EDP project to become

operational, it is still only a pilot study (T. J.

Crovello, pers. comm. 1973).

One of the main reasons Soper cited for indexing a

herbarium was that specimens suffered appreciably

from use and were largely irreplaceable. Although

the actual specimens were required, on many occasions

they could be spared all or most of their handling by

first consulting an index. Label data captured by

Soper are given in Table 1. Initially, a Friden flexo-

writer was used for coding data, but presently he is

using an IBM 2741 typewriter on-line to the computer

for data capture (Morris, 1973). An even more

sophisticated procedure is being developed (McAl-

lister et al., n.d.). Apart from the production of labels,

output will consist of distribution maps drawn by a

plotter, catalogue cards and tabulated listings of label

data. Label data considered essential, desirable or

optional for capture are given by Beschel & Soper

(1970).

G. T. Prance (pers. comm. 1972) has motivated

strongly the provision of computer facilities at the

New York Botanical Garden. The system envisaged

will eventually automate the entire Garden, from

salaries, and journal mailing lists to garden-plant

accessions. As far as herbarium labels are concerned.

Prance will start with clearly-defined sub-sets of data

rather than trying to work systematically through the

entire herbarium. This approach will provide manage-

able-sized projects and obtain useful results fairly

quickly. It will be possible to obtain specimen labels

as a by-product of data capture done to reshuffle data

to assist with phytogeographical and ecological

research. Data to be captured are indicated in Table I.

Everist (1972) is using a Ricoh 5000 paper-tape

punching typewriter with two paper-tape readers and

two punches primarily for preparing herbarium labels

and obtaining machine-readable output as a by-

product. Label data, as set out in Table 1, are typed.

Duplicate labels are produced automatically by

feeding the paper-tape through the typewriter as many

times as required. Data common to more than one

consecutive specimen are pre-punched and auto-

matically inserted.

Labels of all the specimens of one family in the

herbarium have been encoded, as a trial, in addition

to the data captured for new specimens. They have

made slow progress on the specimen backlog and

conclude that a magnetic tape encoder or computer

card punch would be more satisfactory than the

typewriter. Their main difficulty is finding geographi-

cal co-ordinates for old specimens. Everist concludes

that the results are already justifying his expenditure.

Within a few years he expects meaningful answers to

many questions on distribution, floristic and habitat

data and flowering times for much of the flora of

Queensland.

The ultimate necessity for the preparation of a

catalogue during taxonomic and phytogeographical

studies was pointed out by Greene (1972) of the

British Antarctic Survey. In common with many

other institutes, he faced staff shortage, limited

support facilities and rapidly growing collections.

He decided to set up a data bank and retrieval system

to expedite the administration of his material, con-

sisting of about 30 000 specimens. Required from the

system were: (a) a catalogue of all data for every

specimen with facilities for incorporating new data

and updating records; (b) immediate availability of all

data in a variety of sequences; and (c) a rapid method

of duplicating information, such as preparing herba-

rium labels. As it was considered too expensive to

have all these tasks carried out by a computer, a

paper-tape typewriter was used for (a) and (c) and an

ICL 1900 computer was used for storing the data

(Table 1) and performing retrieval tasks. A normal-

language card-file and a coded computer file of speci-

mens have been made.

Greene’s data bank is proving to be really valuable

in speeding up the production of regional floras, and

the chore of checking duplicate herbarium labels and

index cards is eliminated.

The only data fields captured by all systems (Table

1), to sum up extant and proposed systems, are the

binomial/trinomial and collector’s name fields.

Other commonly-recorded fields include accession

number, family, identifier, state of the specimen (SOS),

country, state or province, locality description, degree

reference or co-ordinates, altitude, habitat and other

notes. The detail and form of recording vary from one

system to another. For example, some workers record

genus and species separately and others together, some

record only the state of the flower in the SOS field,

while others note flowers, fruit, seeds, leaves and roots

in this field. An important finding from the literature

survey is that the two projects with fairly small inputs

(Hall and Greene) pre-code information while organ-

isations with larger data sets prefer to enter data

uncoded as far as possible.

Six herbarium data-capture projects are known to be

in operation already and at least one more (New York

Botanical Garden) is due to start soon. Although

problems have been encountered with all the projects,

it is encouraging to note that not one has failed and,

furthermore, that all officers with systems, when

approached, were enthusiastic about the future of

herbarium automation.

THE NEED FOR A SYSTEM FOR SOUTH AFRICA

Soper & Perring (1967) state that a satisfactory

EDP system should be a by-product, or bonus,

of the initial need to label accurately all specimens kept

in the herbarium. This statement holds for South

Africa but we also have other motivation. Crovello

(1967) considers that EDP will never replace the

taxonomist but, rather, it will do away with some of

the routine tasks that an intelligent scientist should be

relieved of. The writer sees this, too, as the aim of our

system, but would take Crovello’s statement further

and contend that EDP brings within the range of

possibility answers to a number of taxonomic,

ecological and environmental questions in need of

answering.

Labels for the 20 000 specimens accessioned annu-

ally by the Botanical Research Institute are typed and

the specimens filed by families at present. By typing

labels in such a way that machine-readable copy is

produced as a by-product, no extra cost per label is

involved. Although the equipment is more expensive

than an ordinary typewriter (as little as R50 per

month hire), shortcuts in the form of abbreviations

and constant fields, entered automatically by the

computer, increase the speed of accession to balance

out the increased cost to a greater or lesser extent

352

COMPUTERIZATION OF HERBARIUM PROCEDURES

Two or twenty top-quality copies of a label are

equally easy to produce and, furthermore, all are

identical.

In addition to the production of neat, multiple-

copies of specimen labels, there are other uses for data

captured in machine-readable form in the process of

making labels. An important use is for the production

of check-lists of large or small areas. Examples are:

(a) Highveld Agricultural Region — for the present

ecological survey; (b) Lesotho or Natal — for a

National or Provincial list; (c) a quarter-degree grid

square — a list is needed for the square in which

Pretoria is situated for the Pretoria Flora at present in

preparation; (d) poorly-collected areas like Tongaland

or the Richtersveld, to give a plant collector an idea of

what is known of the flora before setting out on a

general collecting trip; and (e) endemic distribution

data for assessing conservation priorities of different

specific areas.

By using the computer to count the number of

specimens and number of species collected in each

quarter-degree square, poorly-collected-from areas

should be generally evident and attention can be

devoted to collecting from these areas.

On a recent field trip the author needed to know

the distribution of Terminalia sericea in South Africa

to test a hypothesis. A distribution map immediately

produced by the computer would have spared him a

search through the specimens in the herbarium.

Although it can be argued that the effort of drawing

one distribution map by hand is not very great, the

advantage of being able to draw distributions auto-

matically as required, for any of the 10 000 most

common plants in South Africa, should be obvious.

Once the specimens of the early collectors, such as

Burchell, have been accessioned it will be easy to

print out an itinerary of the places they visited.

A similar service to present-day collectors would

produce collectors’ registers, compiled by computer.

With a quick method of data capture a great deal of

information which is presently for all practical

purposes “lost” becomes usable. For example, “spot”

identifications, voucher specimens and sight records

could be added to the data bank and used in it, admit-

tedly with a lower, but defined, degree of confidence.

At present, this valuable information is discarded.

In addition to their value within the Botanical

Research Institute, the herbarium data will be

available for inclusion in the proposed agricultural

data bank of the Department of Agricultural Techni-

cal Services.

SYSTEM PLANNED FOR THE BOTANICAL RESEARCH

INSTITUTE, PRETORIA

For data capture we propose to use cathode-ray

tube (CRT) screens with keyboards linked by tele-

phone line to the Burroughs B6700 computer run by

the Department of Agricultural Technical Services.

A computer programme will flash the headings, like

genus:, species:, collector:, date:, on

the screen and the operator will key in data directly

from the label. A number of short-cuts designed to

streamline accessioning will be used. For example,

extensive use will be made of abbreviations which the

computer will expand, and conversions of British

standard units of measurement to metric values

will be carried out by computer. All data fields that

are repeated on more than one consecutive label will be

inserted automatically after the first entry. Prelimi-

nary editing of the data will be done by computer and

will be followed by manual checking.

We are presently negotiating for the use of sel-

gem to satisfy our information retrieval needs.

Selgem was developed by the Smithsonian Institution

(Creighton & Crockett, 1971 ; Creighton et al., 1972)

for collections similar to our own. Unfortunately, we

are not even able to consider the Gis package advo-

cated for use by Krauss (1973) as it is incompatible

with the computer available to us. Other current

data-banking systems were considered. In particular,

the comprehensive Cambridge system developed by

J. L. Cutbill (Cutbill & Williams, 1971), and taxir

developed by D. J. Rogers (Rogers, Flemming & Esta-

brook, 1967), were investigated.

Various CRT screen lay-outs are being experi-

mented with at present and we are considering what

information is worth capturing from the specimen

backlog and how new accessions will be treated.

Standards for each data field are being drawn up.

CONCLUSION

The merits and procedures of herbarium automation

that have been discussed by six herbarium curators

who have applied EDP to their collections to improve

the service given by their herbaria, have been con-

sidered for a herbarium data bank for the Botanical

Research Institute in Pretoria. It is concluded that for

our purposes such a data bank includes the uncoded

data indicated in Table 1. Such a system can be

realized practically and will provide immediate

access to information needed for a variety of pur-

poses and which was hitherto for practical purposes

unavailable.

ACKNOWLEDGEMENTS

The author is grateful to the Secretary of the Depart-

ment of Agricultural Technical Services for making

possible the study tour which laid the foundation for

this survey. Grateful acknowledgement is made for the

time and resources made available to the author by all

the overseas and local data bankers who were visited

during this tour. For kindly commenting on the

manuscript I thank B. de Winter, S. L. Everist,

T. J. Crovello and A. V. Hall.

OPSOMMING

Huidige projekte vir herbarium automatisasie in Kaapstad

(A. V. Hall), Notre Dame (T. J. Crovello), Harvard (L. I. Nev-

ling), Ottawa (J. H. Soper), Brisbane (S. L. Everist), die

“British Antarctic Survey” (D. M. Greene) en New York se

voorgestelde sisteem (G. T. Prance), word volledig bespreek.

Daar word gevind dat die inligting vir take met 'n klein aantal

plant monsters gekodeer word, terwyl die inligting vir

groot sisteme, so ver moontlik sonder kodering verwerk word.

Daar word genoem dat geen een van hierdie automatisasie

projekte misluk het nie en dat die gebruikers almal baie entoe-

siasties oor die toekoms van hierdie metodes is.

REFERENCES

Beschel, R. E. & Soper, J. H. 1970. The automation and

standardization of certain herbarium procedures. Canad.

J. Bot. 48 : 547-554.

Creighton, R. A. & Crockett, J. J. 1971. SELGEM: A sys-

tem for collection management. Smithsonian Institution

Information Systems Innovations 2(3): 1-24.

Creighton, R. A., Packard, P. & Linn, H. 1972 SELGEM

Retrieval: A general description. Smithsonian Institution

Procedures in Computer Science 1(1): 1-38.

Crovello, T. J. 1967. Problems in the use of electronic data

processing in biological collections. Taxon 16: 481-494.

Crovello, T. J. 1972. Computerization of specimen data from

the Edward Lee Greene Herbarium (ND-G) at Notre

Dame. Brittonia 24: 131-141.

Crovello, T. J. & MacDonald, R. D. 1970. Index of EDP-IR

projects in systemalics. Taxon 19: 63-76.

Cutbill, J L. & Williams. D. B. 1971. A program package for

experimental data banking, in: J. L. Cutbill (editor)

Data processing in biology and geology. Systematics

Association Special Volume 3: 105-113.

J.W. MORRIS

353

Everist, S. L. 1972. Computer processing of labels in the Queens-

land Herbarium. Unpublished paper read at annual Con-

ference of the Museums Association of Australia, Sydney,

October 1972.

Gomez-Pompa, A. & Nevling, L. I. 1970. La flora de Veracruz.

An. Inst. Biol. Univ. Nal. Anton. Mexico 41. Ser. Botanica

(1): 1-2.

Gomez-Pompa, A. & Nevling, L. I. 1973. The use of electronic

data processing methods in the flora of Veracruz program.

Contr. Gray Herb. 203 : 49-64.

Greene, D. M. 1972. A taxonomic data bank and retrieval

system for a small herbarium. Taxon 21 : 621-629.

Hall, A. V. 1972a. Computer-based data banking for taxono-

mic collections. Taxon 21 : 13-25.

Hall, A. V. 1972b. The use of a data-banking system for

taxonomic collections. Contr. Bolus Herb. 5: 1-78.

Krauss, H. M. 1973. The use of generalized information-

processing systems in the biological sciences. Taxon 22:

3-18.

McAllister, D. E., Leere, A. B. & Sharma, S. P. n.d. A

batch process computer information retrieval and cata-

loguing system in the fish collections of the National

Museums of Natural Science. Syllogeus 1 : 1-20.

Morris, J. W. 1973. Overseas advances in quantitative ecology

and computerised data banking in biology. Unpublished

Departmental Report, Botanical Research Institute,

Pretoria.

Rogers, D. J., Flemming, H. S. & Estabrook, G. F. 1967.

Use of computers in studies of taxonomy and evolution.

in: Th. Dobzhansky, M. K. Hecht & W. C. Steere (eds)

Evolutionary Biology, New York.

Shetler, S. G. 1971. Flora North America as an information

system. Bio Science 21 : 524-532.

Soper, J. H. 1969. The use of data-processing methods in the

herbarium. An. Inst. Biol. Univ. Nal. Auton. Mexico. 40, Ser.

Botanica (1): 105-1 16.

Soper, J. H. & Perring, F. H. 1967. Data processing in the

herbarium and museum. Taxon 16: 13-19.

Bothalia 11, 3: 355-364 (1974)

The Ecological Profiles Technique applied to data from Lichtenburg,

South Africa

J. W. MORRIS- AND J.-L. GUJLLERMf

ABSTRACT

The method of ecological profiles and information shared between species and ecological varia-

bles, developed in France, is described for the first time in English. Preliminary results, using the

technique on Bankenveld quadrat data from Lichtenburg, Western Transvaal, are given. It is con-

cluded that the method has great potential value for the understanding of the autecology of South

African species provided that the sampling method is appropriate.

INTRODUCTION

Presently, one of the most commonly used ecolo-

gical synthesis techniques at the Centre d’Etudes

Phytosociologiques et Ecologiques (C.E.P.E.), Mont-

pellier, is that of ecological profiles and information

shared between species and ecological variables

(Profils ecologiques et information mutuelle entre

especes et facteurs ecologiques). As the writers

consider that the technique deserves trial by other

than French-speaking ecologists, it is described below

in detail. This account is the first known in English.

South African data are used to illustrate its utility.

The first account of the method was given by

Godron (1965) and since then more detailed accounts

have been prepared by Godron (1968), Daget et al.

(1972), Guillerm (1969 a, b and c) and by Guillerm

(1971), all of which have been in French. Both

floristic and habitat data from samples are used in

this univariate technique. According to Daget et al.

(1972), and others, its main application is sampling

improvement, but this is by no means its only use.

SAMPLE AREA AND DATA USED

The 110 x 4 m quadrats (in this account ‘quadrat’

is used for ‘sample’, ‘releve’, etc.) used in the study,

were laid out in a stratified random manner near

Lichtenburg, Western Transvaal, within the 2626AA

quarter-degree square. Details of the area and field

sampling method are given in Morris (1973). The

vegetation is classed as Bankenveld by Acocks (1953).

Floristic data consisted of presence within the

quadrats of 229 species. The 12 ecological variables

listed in Table 1 were coded for analysis. The first

eight are acceptable habitat variables. Soil colour

Table 1. Ecological variables used in the analysis and the

number of classes of each.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private BagXIOl, Pretoria.

t Centre d’Etudes Phytosociologiques et Ecologiques

L. Emberger, B.P. 5051, Montpellier 34, France.

(the ninth variable) was included as a control; it was

not expected that it would be ecologically meaningful

in its own right. Total basal cover was that of all

grass and herb strata combined and the strata of

variables 11 and 12 are those defined by Godron

et al. (1968). For the purpose of this paper, basal

cover was considered an environmental variable

although it is usually considered as a structural

property of the vegetation. Cover was used in this way

as so few ecological variables were available. As the

variables were not originally recorded for analysis

by this technique, there was not always enough

information available to code according to the

specifications of Godron et al. (1968). Topographic

position, apparent biotic influence, soil pH and soil

HC1 reaction were coded according to Godron et al.

Aspect, soil depth and soil colour were coded in the

same way, although the total number of classes was

later reduced in the analysis. As there is little topo-

graphy in the Lichtenburg area, the following special

scale was used for slope: 0 = no slope, 1= slight

(less than 1 %), 2 = 1 % or more. Amount of surface

rock was coded as: l=none, 2=few, 3=moderate,

4=many rocks. Total basal cover was recorded on the

following scale: 2 = 5-9, 9%, 3 = 10-14,9%, 4=15—

19,9%, 5=20-24,9%. Basal cover for individual

strata was recorded as: 0=0%, 1=1-2%, 2=3-4%

3 = 5-6% etc. Two strata, corresponding

with strata II and IV of Godron et al., were recognized.

ECOLOGICAL VARIABLES AND EQUITABILITY OF

SAMPLING

The first step is the calculation of the comprehensive

profile (CP) for each ecological variable. The CP is a

list of all the classes of a variable and the frequency of

occurrence of quadrats in each class. An example of a

CP is given in Table 2 where there are a total of N

quadrats.

Table 2. Comprehensive profile for a variable with K classes.

Comprehensive profiles for the 12 variables of this

study are given in Table 3. The CP gives the distribu-

tion of absolute frequency. Relative frequency

is obtained by dividing each frequency by N. If the

absolute frequency is not too low, a good estimate of

probability of occurrence may be obtained from the

relative frequencies. If the environmental gradient has

been well sampled, quadrats will be equally distributed

through the classes of the CP and, hence, the probabi-

lities of occurrence will be approximately equal.

A variable for which quadrats have equal probabilities

17749-11

356 THE ECOLOGICAL PROFILES TECHNIQUE APPLIED TO DATA FROM LICHTENBURG, SOUTH AFRICA

Table 3. Comprehensive profiles for variables used in the analysis. Class

numbers (from Godron et al., 1968) do not apply to variables marked

with an asterisk.

Class number

of being in each class is said to have a high indeter-

mination (M. Godron, pers. comm., 1973). Indeter-

mination may be estimated by calculating the entropy

(1L) (see Abramson, 1963) of variable L as:

Il=

K

2

i

Ri

N

log2

N

Ri

where the symbols are explained in Table 2. 1 he

formula can also be expressed in the form :

1 K

IL = log2N- j^IRi log2 Rj

IN •

(FORTRAN IV programmes written at C.E.P.E. for

the ecological profiles technique have been modified

for local computers. Listings and card decks are

available from the senior author).

Maximum entropy (Imax,L) of factor L, which is also

the highest indetermination, is found when the

arithmetic mean number of quadrats occurs in each

class of the CP. Then :

R; = -rp for all i and

IV

, K N 1 . NK, ..

Imax, L 1 K. N 1 N 10g2K-

From the entropy of a variable (IL) and maximum

entropy (Imax.r) it is possible to judge equitability of

sampling. Equitability is the degree to which quadrats

cover the variation of a factor with the given class

intervals. The fraction (QL) is used:

VL — I

Amax,L

Usually, the higher the value of Q, the better the

sampling of the total variation within the sample area.

The fraction should, however, be used with caution

as some variables have misleading Q values. The

problem arises when one is dealing with variables

with many classes: for example, the 125 soil types in

‘Code pour le releve methodique’ of Godron et al.

Many classes will be unsampled, or very poorly

represented in the CP. The value of Imax should be

based on the number of classes sampled adequately.

After studying the CP’s for the 12 variables used

here, the number of classes was reduced for five of

them (see Table 1) as certain class frequencies were

very low. In the cases of soil depth and basal cover

adjacent classes were grouped. South, south-west,

east and south-east aspects were grouped and north

and north-west aspects were grouped. The two

soil-colour codes with high frequencies, 6/32 and

6/33, were retained. The other 6/ codes and the 7/,

5/, and 4/ codes were grouped into four classes on the

basis of the ‘simplified’ wavelength code of Godron

et al. It has been suggested by M. Godron (pers.

comm., 1973) that inspection of corrected frequency

profiles (see later) may assist in deciding which classes

should be grouped. The observed and maximum

entropy for each variable and values for Q are given

in Table 4.

Table 4. Observed and maximum entropy for each variable and equitability

of sampling.

J. W. MORRIS AND J.-L. GUILLERM

357

Variables which are equitably sampled according to

the Q criterion, include, in order of decreasing

importance, biotic influence, surface rock, basal

cover for strata IV and II, and soil pH.

Observed entropy has been plotted against maxi-

mum entropy for each variable in Fig. 1. The nearer a

variable is to the diagonal line, connecting points of

maximum entropy, the more equitable is its sampling.

Biotic influence, surface rock and cover stratum IV

are most equitably sampled while soil HC1 reaction,

total basal cover and soil colour are, by this criterion,

poorly sampled.

To improve equitability of sampling, Daget et al.

(1972) suggest resuming stratification of the vegetation

where it was stopped, starting with the variables

known to be poorly sampled. In the choice of varia-

bles, one is also guided by the calculation of mutual

information between species and variables, which

permits the detection of variables of ecological

importance.

MUTUAL INFORMATION BETWEEN SPECIES AND

ECOLOGICAL VARIABLES

The frequency of occurrence of species E in each

class of variable L forms the ecological profile of

species E for variable L. A modified ecological

profile results if relative frequency or corrected

frequency are used instead of absolute frequency

(see Gounot 1958, 1961, 1969, Godron 1965, Guillerm

1969a). The absolute frequency profile is the number

of times species E occurs in each class of variable L.

It gives the information for species which is given for

each variable by the CP. Absolute frequency may yield

misleading results as it is directly proportional to the

total number of occurrences, therefore it is better

to use relative frequency for the species ecological

profile (Daget et a/. 1972). If there are quadrats in

a class of variable L and quadrats in that class

contain species E, relative frequency is given by

Ux/Ri-

Further to smooth out variations caused by

differences in total absolute frequency, corrected

frequency is used to form the ecological profile.

Corrected frequency for class i (Q) is found by multi-

plying relative frequency by the inverse of average

relative frequency over all quadrats:

r _Ui JN

1 Rj ' UT

(see Table 5 for explanation of symbols)

Table 5. Example of ecological profile for species X in K

classes of variable Y.

Information about species behaviour, which is not

apparent from the absolute frequency profile, may be

obtained from the corrected frequency profile.

Topographic

Position

OBSERVED ENTROPY

Fig. 1 . — Relation between observed and maximum entropy for each ecological variable.

SPECIES ENTROPY

RELATIVE FREQUENCY

Fig. 2. — Relation between species entropy and relative frequency.

The presences and absences of each species in an

ecological profile may be used to calculate species

entropy (Is) which is defined as:

Is

UT

N

log.

Ul+1^ ,og2

_N

V^r

K. K

where Uj= EUj and VT— EVj (see Table 5)

i i

A species that is either present or absent in all

quadrats will have Is equal to zero, the minimum

value, and a species present in half the quadrats will

have Is equal to one, the maximum value. The rela-

tionship between relative frequency and species

entropy is shown graphically in Fig. 2.

The information carried by a species relative to an

environmental variable may be calculated. It is known

as the mutual information between species and varia-

ble. For species E and variable L (in K classes) it is

given by :

K

Ui

Il,e= I ^ log2

i

Ui

R,

N

UT

log.

U; N

Ri ' VT

(Symbols as in Table 5)

As mutual information can be calculated for every

species with every variable, a species by variables

matrix of entropy values is available. To reduce the

volume of results, only certain species and ecological

variables are selected, although there is no reason why

all the values should not be calculated if this were

required.

The calculation of mutual information between

species and variables allows the determination of

which variables play an important role in the distribu-

tion of species, in other words, the ‘active’ variables.

J. W. MORRIS AND J.-L. GUILLERM

359

The most convenient way to find active variables is to

calculate mean mutual information and plot the

values against the entropies of the variables. Mean

mutual information is plotted along the ordinate and

the entropy of the variable along the abscissa to give a

two-dimensional ordering of variables. Study of the

distribution of variables within the graph allows

choice of those variables which will improve sampling.

On such a graph, variables placed to the right and

left are, respectively, well-sampled and under-

sampled. Variables placed at the top of the graph are

more ‘active’ than those placed below them. To

choose variables to be re-sampled, Godron (1968)

suggested that variables for re-sampling should be

chosen from among those which are insufficiently

sampled (to the left of graph) and should be those to

which the vegetation appears most sensitive (top of

graph). In his application, Godron re-sampled and

analyzed twice to obtain a satisfactory sample.

MUTUAL INFORMATION BETWEEN LICHTENBURG

SPECIES AND VARIABLES

(a) Overall relationships

In Fig. 3 relationships between mean mutual

species information and variable entropy are given

for the 12 variables. For each variable, mean mutual

information was calculated over the 100 species

(out of a total of 229) with highest entropies. From this

graph, topographic position, cover stratum II, and

stratum IV, soil depth, biotic influence and surface

rock are, in that order, most equitably sampled.

The most active factors are topographic position,

cover stratum II, soil colour, cover stratum IV, soil

depth and biotic influence, in that order. With these

data there is a marked positive correlation between

mean mutual information and variable entropy,

illustrated by the clustering along the 5 % line.

The diagonal lines in Fig. 3 give the value of mean

mutual information divided by the variable entropy

as a percentage. It is an expression of the indicator

value of variables.

7%

VARIABLE ENTROPY

Fig. 3. — Relation between mean mutual information for the species with the highest information

content and the entropy of the ecological variables,

360 THE ECOLOGICAL PROFILES TECHNIQUE APPLIED TO DATA FROM LICHTENBURG, SOUTH AFRICA

The variables are discussed separately in the sub-

sections which follow, but their relationships to each

other should not be forgotten; for example, soil depth

and biotic influence can be related and soil depth is

usually related to slope angle.

(b) Topographic position

Profiles for topographic position present difficulties

in interpretation even though, as a variable, it is

very favourably placed in Fig. 3. Firstly, there is a very

asymmetric distribution of quadrats within the classes

with class 5 being very poorly sampled (see compre-

hensive profile. Fig. 4). Secondly, topography is a

Classes

Fig. 4. — Comprehensive profile for the variable topographic

position (l=flat, 2=round summit, 3= waxing slope,

4 mid-slope, 5 ledge, 6 waning slope, 7 open depres-

sion, 8 =closed depression).

discontinuous variable which means that a great deal

of inspection is necessary to determine trends and

patterns in distribution. Thirdly, the Lichtenburg

area has such little relief that large differences in

floristic composition related to this variable are not

expected. Related to this difficulty is that of the coding

of this variable. As coding according to Godron et al.

( 1 968) was done from field notes, a number of quadrats

could have been misclassified. For example, the

difference between an open and a closed depression,

or between a flat area and the crest of a very large,

rounded hill, was not always clear from the field notes.

At the suggestion of M. Godron (pers. comm.,

1973) product-moment correlation coefficients were

calculated between each pair of corrected profiles of

the 100 species with highest mean mutual information

(4 950 calculations) and the 100 highest positive values

obtained. Correlations were first calculated over all

eight classes of the profile and then over the seven

classes remaining after the fifth class had been

excluded. As corrected frequency profiles for species

over the eight classes were markedly distorted by the

low CP value for class 5, the following discussion refers

only to correlations over the seven remaining classes.

The 100 correlation values were all significant at

p=0,01 % and the first 14 at p=0,001 %. Unfortuna-

tely, these values were not as satisfactory as expected

since mean mutual information level also has to be

taken into account. The following are illustrations.

The highest correlation was between Dicoma anomala

and Loudetia simplex that favour flat areas, rounded

summits and waxing and waning slopes, according to

the profiles. D. anomala had the eighth highest mean

mutual information value for topographic position

while L. simplex was fortieth in the ranked order,

indicating a low importance. Thus the profile for

L. simplex is not nearly as significant as that of D.

anomala even though the profiles are highly correlated.

Gazania krebsiana, Digitaria argyrograpta and

Eragrostis gummiflua profiles are highly correlated,

these species occurring mainly on flat areas and in

open depressions. As all three have low mean mutual

information values, however, the high correlations are

again of not much value.

With data specially collected for an ecological

profiles study, use of the correlation coefficient to

compare species profiles may be of use.

(c) Aspect

Aspect was not equitably sampled and is not active

in the quadrats (Fig. 3). This to be expected as the

Lichtenburg area is very flat. In the following dis-

cussion it must be remembered that out of a total of

seven classes only synthetic classes were retained.

Species favouring flat ground (no aspect) include

Senecio coronatus, Gazania krebsiana and Nidorella

hottentotica.

Species favouring north and north-west aspects

include:

1 . Sporobolus africanus

2. Cymbopogon plurinodis

3. Eragrostis superba

4. Chascanum pinnatifida

5. Eragrostis stapfii

6. Stipagrostis uniplumis

7. Euphorbia pseudotuberosa

8. Eragrostis lehmanniana

9. Hermannia tomentosa

10. Bulbine sp.

1 1 . Anthephora pubescens

12. Turbina oblongata

(In the above and following lists, species are ordered

by decreasing mutual information content. Figured

species have the highest mutual information contents.)

Species favouring the south, south-west, south-east

and/or east aspects include :

1 . Dicoma macrocephala

2. Diheteropogon amp l ec tens

3. Talinum arnottii

4. Coleus neochUus

5. Trachypogon spicatus

6. Blepharis angusta

7. Rhynchelytrum repens

8. Hypoxis sp.

9. Ursinia nana

1 0. Menodora africana

1 1 . Pentanisia sp.

These lists must be considered very tentative as

much more intensive sampling of all aspects will be

necessary to obtain a reliable picture of the influence

of aspect and topographic position on species distri-

butions.

(d) Biotic influence

Results which appeared very good were obtained

from this environmental variable. The problem with

interpreting the results was, however, that the qua-

drats, as regards biotic influence, were classed on the

basis of a subjective appreciation whereby the species

composition of the quadrat influenced the assessment

of the degree of biotic influence and may have

distorted the ecological profiles.

As may be expected, the main patterns of distribu-

tion along the biotic influence profile are an increasing

or decreasing corrected frequency with increase in

biotic influence. Good examples of the former are:

Ursinia nana, Eragrostis tricophora , Cynodon dactylon

and E. lehmanniana (Fig. 5). Other species also having

this trend include:

1. Dicoma macrocephala 5. Heteropogon contort us

2. Corchorus asplenifolius 6. Zornia milneana

3. Blepharis integrifolia 7. Euphorbia inequilatera

4. Koliautia omahekensis 8. Soianum supinum

A related distribution is shown by Eragrostis

superba. This species as well as:

1 . Hermannia tomentosa 4. Stipagrostis uniplumis

2. Indigofera daleoides 5. Eustachys mutica

3. Chascanum pinnatifida

are found in approximately equal proportions in the

three classes of disturbed vegetation but are not

common in undisturbed areas.

J. W. MORRIS AND J.-L. GUILLERM

361

Fig. 5. — Comprehensive profile (CP)

and distributions of (a) Cynodon

dactylon, (b) Eragrostis trico-

phora, (c) E. lehmanniana, (d)

Ursinia nana, (e) Eragrostis

superba, (f) Diheteropogon am-

plectens and (g) Clematopsis

stanleyi, representing the two

main distributions within the

biotic influence profile (Class 1

least influenced, Class 4 most

influenced).

Species found on undisturbed areas and having

successively lower corrected frequencies as the degree

of biotic influence increases, include, notably, Clema-

topsis stanleyi and Diheteropogon amp/ectens (Fig. 5)

as well as:

1. Thesium cost alum 8. Bulbostylis burchellii

2. Eragrostis racemosa 9. Sporobolus pectinatus

3. Schizachyrium sanguineum 10. Brachiaria serrata

4. Acalypha sp. 11. Elephantorrhiza elephantina

5. Cymbopogon excavatus 12. Diplachne fusca

6. Heteropogon contort us 13. Senecio venosus

7. Barleria pretoriensis

(e) Soil depth

The comprehensive profile for soil depth (Fig 6)

shows that deeper soils were not well sampled. As the

whole study area is a dolomitic lithosol with only

Fig. 6. — Comprehensive profile (CP) and distributions of (a)

Zornia milneana, (b) Kohautia omahekensis, (c) Ipomoea

obscura var. fragilis and (d) Oropetium capense, within the

soil depth profile.

pockets of deep soil, additional random sampling will

not produce a more even distribution. Equitability of

sampling will be achieved only if quadrats are placed

at random within the deep-soil pockets,

The ecological profile for Zornia milneana (Fig. 6)

shows it to occur on deep soil, whereas Ipomoea

obscura var. fragilis and Oropetium capense are most

frequent on very shallow soil. It was observed during

fieldwork that the latter species usually occurred in

small sand pockets in extensive dolomite sheets.

The profile confirms the observations. The profile for

Kohautia omahekensis suggests that within the range

included in the study it has a wide soil depth amplitude

below 40 cm, but is most frequent on soils about 1 5 cm

deep.

(f) Soil pH

Even though this factor was not equitably sampled

(Fig. 3), valuable information about the ecology of

certain species can be gained from a study of the

corrected ecological profiles. Certain species have been

plotted in Fig. 7 to illustrate the four main trends.

The most remarkable trend is shown by Finger-

huthia africana and Oropetium capense (Fig. 7),

which are restricted to soil with a pH of 8,0 or higher.

Other species showing this trend, but to a less

marked extent, include:

1 . Stipagrostis uniplumis 4. Vernonia oligocephala

2. Turbina oblongata 5. Brachiaria serrata

3. Sporobolus africanus 6. Euphorbia inequilatera

All these species occur on neutral and basic soils

and not acid ones.

An example of a species rarely found on soils with

a pH above 7,0 is Justicia anagaloides. It grows

equally well in acid or neutral soils (Fig. 7). Other

species with the same distribution include Chascanum

hederaceum, Eragrostis racemosa and Sporobolus

pectinatus.

Species found on acid soils but which also grow in

neutral pH include Eragrostis tricophora and Dicoma

anomala (Fig. 7). Other species which exhibit this

trend of decreasing corrected frequency with increas-

ing pH are:

1. Oxygonum dregeanum 5. Hermannia betonicifolia

2. Pygmaeothamnus zeyheri 6. Lasiosiphon capitatus

3. Raphionacme burkei 7. Zornia milneana

4. Rhynchelytrum repens

Species showing a peak of corrected frequency in the

centre of the range, in other words growing in a neutral

or slightly acid soil, include Barleria macrostegia and

Lightfootia denticulata (Fig. 7). Other species with

similar distributions include:

1. Helichrysum caespititium 5. Diheteropogon amplectens

2. Dicoma macrocephala 6. Cyphocarpha angustifolia

3. Acalypha sp. 7. Schizachyrium sanguineum

4. Chaetacanthus costatus

(g) Basal cover

Total basal cover was inadequately sampled

(Fig. 3), and is not discussed further. Within the

second and fourth strata, however, three patterns of

362 THE ECOLOGICAL PROFILES TECHNIQUE APPLIED TO DATA FROM LICHTENBURG, SOUTH AFRICA

Fig. 7. — Comprehensive profile (CP) and corrected ecological profiles for (a) Justicia anagaloides,

(b) Fingerhuthia africana, (c) Oropetium capense, (d) Eragrostis tricophora, (e) Dicoma anomala,

(f) Barleria macrostegia and (g) Lightfootia denticulata, plotted aginst soil pH classes.

species behaviour emerge. Corrected frequencies of

certain species increase as cover increases while

frequencies of others decrease or, as in the third

pattern, increase to a peak and then decrease. It is

stressed that cover is used as an ecological variable

and not a structural property of the vegetation. It is

assumed that the degree of cover influences the

behaviour of certain species.

Species in which peaks of corrected frequency are

50

40

30

20

10

0

0 12 3 4 5 6

CP stratum II

a

while species with peaks in the second stratum only

include:

I 2 3 4 5 6

CP stratum IV

1 . Chaetacanthus costatus 4. Helichrysum caespititium

2. Diheteropogon amplectens 5. Dicoma anomala

3. Corchorus asplenifolius 6. Bulbine sp.

Species with peaks in the fourth stratum only,

include: Helichrysum cerastiodes, Clematopsis Stanley i

and Oxygonum dregeanum.

Species which increase in frequency as basal cover

in the second stratum increases include Ursinia nana.

Dicoma macrocephaia , Eragrostis stapfii, Eustachys

mutica and Brachiaria serrata , while Andropogoii

appendiculatus, Dipiachne fusca , Ipomoea obscura

var. fragi/is and Eragrostis racemosa exhibit the

opposite trend. In the fourth stratum species which

increase include Thesium cost atom, Raphionacme hir-

suta , Turbina oblongata , Pygmaeothamnus zeyheri and

Nolletia ci/iaris while :

1 . Btepharis integrifotia 6. Cynodon dactylon

2. Eragrostis tehmanniana 7. Lippia scaberrima

3. Set aria nigrirostris 8. Cyperus capensis

4. Digitaria argyrograpta 9. Sida chrysantha

5. Hibiscus microcarpus

decrease in frequency as basal cover in the fourth

stratum increases.

In addition to the expected trends described above,

certain species show rather odd distributions when

both strata are considered. These are illustrated in

Fig. 8. Zornia milneana, a short, creeping herb

Fig. 8. — Comprehensive profile (CP) and corrected ecological

profiles for (a) Zornia milneana, (b) Acalypha sp., (c)

Dipiachne biflora and (d) Elephantorrhiza elephantinal

plotted against the basal cover classes of the second and

fourth strata.

J. W. MORRIS AND J.-L. GU1LLERM

363

shows a marked 1-shaped curve in the second stratum

and decreases sharply in the fourth stratum as cover

increases. Acalypha sp., another short herb, increases

in frequency in the second stratum as cover increases

but reaches a peak in the fouth class of the fourth

stratum. Species which decrease as cover in the

second stratum increases, and increase as cover in the

fourth stratum increases, are Diplachne biflora and

Elephantorrhiza elephantina.

Although basal cover in both strata II and IV are

active and well-sampled variables (Fig. 3), inspection

of the compreoensive qrofiles (Fig. 8) shows that the

distribution of quadrats through the classes is far

from regular, which may account for the odd patterns

described above.

(h) Other variables

Other variables are not discussed in detail. Surface

rock is closely correlated with soil depth in this

study. As the few slopes within the study area were so

gentle, the slope profile does not carry much informa-

tion. HC1 reaction was very poorly sampled and trying

to attach ecological significance to soil colour was not

considered worthwhile.

SPECIES INDICATOR VALUES

For each ecological variable the species may be

ranked by decreasing mutual information content.

The rank then gives the species indicator value for

that variable. While study of the corrected ecological

profiles allows specification of the ecology of the

species, it is possible to identify the species of which the

ecological requirements are most similar by means of

the indicator values. Species with the 20 highest

indicator values for each of the 12 variables were

listed and species which were listed four or more

times are given in Table 6. These species are most active

over all the environmental variables.

Instead of looking at the species by variables

matrix of mutual information, variable by variable,

as was done above, it may be studied species by

species. In Table 7 mutual information values of five

species are given for the twelve variables. To obtain a

Table 6. Species ranked in one of the first 20 positions for at least four variables

and the variables for which they were so ranked. Names of variables

corresponding to numbers are given in Table 7.

Table 7. Mutual information between five species and 12 variables _(X = mean mutual

information, 1 =Stipagrostis uniplumis, 2=Schizachyrium sanguineum, 3=Turbina

oblongata, 4=Dicoma macrocephala, 5=Justicia anagaloides).

Not calculated, less than 0,054.

364 THE ECOLOGICAL PROFILES TECHNIQUE APPLIED TO DATA FROM LICHTENBURG, SOUTH AFRICA

datum, the mean mutual information for the 100

species with highest mutual information values is

given for each variable. The relative importance of

each species may be obtained by comparing the mean

mutual information value with individual values.

The five species occurred in one of the first 20 ranked

positions of the highest number of variables. Any

species in which one was interested could be included.

For detailed study of these values it will probably be

necessary to use either the species rank or. at least, a

corrected value as datum.

INDICATOR GROUPS

According to Daget et at. (1972), species with

similar ecological profiles and carrying a high infor-

mation content for the same variables form ‘ecolo-

gical groups’. To avoid confusion with ‘ecological

groups' in the community sense, M. Godron (pers.

comm., 1973) suggested the term ‘indicator group'. An

indicator group is a collection of species with the

same, or similar, ecological requirements. From past

work, according to the Daget et a! ., it appears that the

number of groups of species or of isolated species

stabilizes rapidly. The succeeding groups confirm

those that have been established before, or only

modify them slightly. As many variables are usually

correlated (for example, slope angle and soil depth)

and the active variables are analyzed first, the remain-

ing variables usually add little new information.

Any number of indicator groups may be established

for a variable as the distribution of species along a

continuous environmental gradient is continuous, or

nearly so. The species may be ordered in a series of

groups that are scale-imbricated. Ordering may be

done automatically with the aid of a card sorter

(Daget & David, 1970).

In this paper, species that have the same, or a similar

pattern of response to a variable have been discussed

together, but no attempt has been made to derive

indicator groups as the data were considered incom-

plete.

CONCLUSIONS

For the determination of ecological profiles it is

necessary to calculate the entropy, or information

content of species, of variables, or mutually between

species and variables. The calculations assume the

frequency distribution to be related to the probabili-

ties of species occurrence. The collection of quadrats

is considered a “population” and is treated as such.

This equivalence has its limitations in that it assumes

the number of quadrats is ‘large’. By use of a more

complex entropy formula, it is possible, however, to

overcome this drawback and use relatively ‘small’

samples. Experience gained from other analyses

carried out at Montpellier shows that, with the

entropy formula described above, about 100 quadrats

are necessary for a reasonable first approximation.

With fewer than 100 quadrats the results should only

be used as a guide although the conclusions concerning

the necessary improvements to sampling are still

useful.

Montpellier ecologists (Daget et al., 1972) stress

that the main application of the ecological profiles

technique is sampling improvement. Deficiencies in

sampling do show up (for example HC1 reaction.

Figs. I & 3), but the ecological profiles produced

have another important use. Valuable quantitative,

although univariate, information about species reac-

tions to enviro amental factors is produced. Because of

inadequate sampling of habitat variables, the ecologi-

cal results of the Lichtenburg analysis should be trea-

ed with caution although they indicate the potential

alue of the technique.

With adequate sampling of habitat variables, deter-

mination of the ecological profiles of common

South African plant species for the most important

environmental variables would provide the kind of

information needed to explain the ecology of South

African vegetation and hence its rational manage-

ment and use.

ACKNOWLEDGEMENTS

We wish to thank M. Godron of C.E.P.E./C.N.R.S.,

Montpellier, for assistance with the translation of

some problematical terms, for commenting on the

manuscript and for encouragement. The senior author

would like to thank the Secretary of the Department of

Agricultural Technical Services, Pretoria, for the

study bursary which made his visit to Montpellier

possible and Dr D. Edwards for encouragement and

valuable comments on the manuscript.

OPSOMMING

Die metode van ekologiese profiele en inligting verdeel deur

spesies en ekologiese veranderlikes, wat in Frankryk ontwikkel

is, word vir die eerste keer in Engels beskryf. Voorlopige

resultate van die gebruik van die tegniek op gegewens in die

Bankenveld van Lichtenburg in Wes-Transvaal, word gegee.

Ten slotte word genoem dat die metode ’n groot potensiele

bydrae kan lewer tot die begrip van die outekologie van Suid-

Afrikaanse spesies, mits monsterneming by die tipe van analiese

aangepas word.

REFERENCES

Abramson, N., 1963. Information theory and coding. London:

McGraw Hill.

Acocks, J. P. H., 1953. Veld types of South Africa. Mem. Bot.

Surv. S. A fr. No. 28. Government Printer, Pretoria.

Daget, Ph. & David, P. 1970. Obtention automatique des

groupes ecologiques en ecailles. CNRS-CEPE note NoA/H

(roneo).

Daget, Ph., Godron, M., Guillerm, J.-L., Drdos, J., Ruzic-

kova, H. & Urvichiarova, E., 1972. Profils ecologiques

et information mutuelle entre especes et facteurs ecologi-

ques. 14° Symposium, Association Internationale de Phyto-

sociologie, Rintelnj fVeser. (1970). Den Haag: Verlag

DrW. JunkNV: 121-149.

Godron, M., 1965. Les principaux types de profil ecologique.

CNRS-CEPE note ( roneo).

Godron, M., 1968. Quelques applications de la notion de

frequence en ecologie vegetale ( Recouvrement, informa-

tion mutuelle entre especes et facteurs ecologiques, echantil-

lonnage). Oecol. Plant. 3: 185-212.

Godron, M., Daget, Ph., Emberger, L., Long, G., Le Floc’h,

E., Poissonet, J., Sauvage, Ch. & Wacquant, J. P., 1968.

Code pour le releve methodique de la vegetation et du milieu.

Paris: CNRS.

Gounot, M., 1958. Contribution a l'etude des groupements

vegetaux messicoles et ruderaux de la Tunisie. Ann.

S.B.A.T., Tunisie 31: 1-152.

Gounot, M., 1961. Les methodes d’inventaire de la vegeta-

tion. Bull. Serv. Carte phytogeogr. B 1 : 7-73.

Gounot, M., 1969. Methodes d' etude quantitative de la vegeta-

tion. Paris: Lie Masson.

Guillerm, J.-L., 1969a. Relations entre la vegetation spontanee

et le milieu dans les terres cultivees du Bas-Languedoc,

These 3° cycle. Ecologie, Fac. Sci. : Montpellier.

Guillerm, J.-L,, 1969b. Procedures d' interpretation des donnees

recueillies ci V aide des formulaires precodes de releves. C.R.

Seminaire sur les methodes d'inventaire phyto-ecologique

et agronomique des prairies permanentes. Montpellier:

CNRS-CEPE, Doc. No. 56: 19-38.

Guillerm, J.-L., 1969c. Une methode de mise en evidence des

groupes ecologiques, appliquec aux terres cultivees du Bas-

Languedoc. Comm. IIP colloque sur la biologie des

mauvaises herbes. E.N.S. A., de Grignon, France. 3 : 86-108.

Guillerm, J.-L. 1971. Profils ecologiques et information

mutuelle entre especes et facteurs ecologiques. Oecol.

Plant. 6: 209-225.

Morris, J. W., 1973 Automatic classification and ecological

profiles of South-western Transvaal Highveld Grassland.

Ph D thesis, University of Natal: Durban,

Bothalia 11,3: 365-367 (1974)

Improvement of association-analysis classification by Braun-Blanquet

technique

B. J. COETZEE*

ABSTRACT

Normal association-analysis was carried out on data collected in the Jack Scott Nature Reserve

in the Central Bankenveld of the Transvaal. As the method was found inadequate for obtaining

optimal definition and arrangement of plant communities, it was supplemented by the Braun-Blanquet

Table Method, which served as a substitute for inverse and nodal analyses. This led to a better

understanding of the vegetation of the Reserve.

Because association-analysis is strictly hierarchical, presentation of inter-group relationships and

interpretation of vegetation-habitat relationships are limited. It is argued that the monothetic

character of normal and inverse association-analyses is a further limitation and although this is com-

pensated for by nodal-analysis, valuable information is discarded as peripheral in the latter process.

A total of 229 4 x4m, stratified, randomly-placed

releves were collected in the Jack Scott Nature

Reserve, an area of 3 100 ha, 50 km west-north-west of

Pretoria, in the Central Variation of Acocks's (1953)

Bankenveld Veld Type. Releves were classified by

normal association-analysis (Williams & Lambert,

1959; 1960) to form a hierarchy with 49 final groups

(Coetzee, 1972).

Each association-analysis final group could gener-

ally be distinguished from other final groups by the

distinctive habitat features which its constituent

releves had in common. There were, however, a

number of releves which did not have the distinctive

habitat features of the groups into which they had been

classified. Furthermore, some groups, which were

widely separated in the classification, appeared to

be closely related because of their similarity in habitat

features which were considered to be important for

discerning inter-group relationships.

The Braun-Blanquet Technique, as described by

Werger (1974), was employed as a means for possibly

improving ,he classification. Floristic composition

of all the releves was presented in a two-way table,

with releves in columns and species in rows. Cover-

abundance estimates were entered in the matrix.

Species with similar distributions were grouped

together in the table and association-analysis final

groups were re-arranged to consolidate patterns in

the Table as much as possible. This consolidation

resulted in a major improvement in the interpretation

of vegetation-habitat relationships. The association-

analysis and semi-Braun-Blanquet classifications are

compared in Figure 1 .

The Braun-Blanquet classification excluded releves

which were classified by association-analysis into

final groups which were very heterogeneous in floristic

and habitat features. The classification accommodating

87 per cent of all releves, comprised a total of 29

groups at various levels in the hierarchy. As indicated

in Figure 1, 15 of these groups correspond to asso-

ciation-analysis groups without re-arrangement of the

association-analysis hierarchy. Groups 4.1.2 and

4.2 in Figure 1 may be combined to form a group

corresponding to association-analysis groups 36 to 41.

Of the remaining 13 groups, another seven correspond

approximately to association-analysis groups and

six cannot be recognized in the association-analysis

classification. These six are: 1.1, 2.1, 2.1(a), 2.2,

4.1.1 and 6.

A clearer concept of the plant communities and their

floristic and habitat relationships was then obtained by

arranging individual releves completely according to

the Braun-Blanquet Method (Coetzee. 1974). This

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

classification is very similar to the re-arranged

association-analysis hierarchy. Differences between

the two are: (i) groups are more homogeneous in

floristic composition and distinctive habitat features;

(ii) the hierarchical arrangement of plant communities

is slightly improved; (iii) the classification accommo-

dates all but two of the 229 releves, left out because

they are obviously very heterogeneous in terms of

floristic composition and habitat features.

Association-analysis and the Braun-Blanquet Met-

hod are similar in the following respects :

(i) both classificatory methods are intended to

show discontinuities in the floristic response of

vegetation to environment;

(ii) both methods take into account all species in

order to establish which ones respond to environ-

mental discontinuities by being non-randomly dis-

tributed and can therefore contribute to a meaningful

classification of releves;

(iii) both association-analysis and the Braun-

Blanquet Method consider association between species

as an indication of non-randomness of distribution;

and

(iv) both methods then classify releves primarily on

their similarities with respect to groups of associated

species.

Despite these points of agreement, the Braun-

Blanquet Method gave better results.

Where vegetation-habitat relationships are multi-

dimensional, only very homogeneous groups and

vegetation types at the most detailed level of classi-

fication have no equally or more plausible alternative

classificatory possibilities. For more heterogeneous

groups and vegetation types several classificatory pos-

sibilities may exist if the system is reticulate. A hier-

archical arrangement merely draws attention to one

such possibility, which has restricted value in interpret-

ing vegetation-habitat relationships, as illustrated by

the following example:

Releves of the Acacia caffra-Chrysopogon montanus

Savanna (group 1.1(a), Figure 1) and the Acacia

caffra-Ruellia cordata Savanna (group 1 . 1(b), Figure

1) appear as two distinct groups in the association-

analysis hierarchy, the hierarchy resulting from re-

arranging association-analysis end groups, and the

classification obtained by the Braun-Blanquet Method

(Coetzee, 1974). Both savannas occur in sheltered

valleys in dolomite and have in common a number of

species which distinguish the sheltered valley vegeta-

tion from vegetation in the rest of the Reserve. The

two savannas, therefore, belong to a single vegetation

type with respect to a particular environmental

discontinuity. In the association-analysis hierarchy,

however, this relationship is obscured because the two

communities are separated from one another by the

Fig. 1— COMPARISON OF ASSOCIATION-ANALYSIS CLASSIFICATION WITH CLASSIFICATION RESULTING FROM RE-ARRANGEMENT OF ASSOCIATION-ANALYSIS

366 IMPROVEMENT OF ASSOCIATION-ANALYSIS CLASSIFICATION BY BRAUN-BLANQUET TECHNIQUE

Floristically-defined and ecologically-interpreted communities that correspond to association-analysis groups at various levels of the hierarchy without rearrangement

B. J. COETZEE

367

first division separating different vegetation types

following another environmental discontinuity. In the

association-analysis hierarchy, the Acacia caffra-

Ruellia cordata Savanna of mesic habitats in sheltered

valleys is thus more closely related to dolomite

communities exposed to cold weather than to the

xeric Acacia caffra-Chrysopogon montanus Community

of sheltered valleys. The summarizing Roman Table

presented by Coetzee (1974) shows both disconti-

nuities. Apart from its hierarchical arrangement,

therefore, the Roman Table shows multidimensional

relationships between plant communities. It presents

the full floristic composition of each group of releves

as well as the overall distribution of all groups of

species separating releve-groups. Normal association-

analysis is, as indicated by Lambert & Williams (1962),

not a complete method for classifying vegetation.

The method is intended as the first of three stages of a

more complete method for studying community-

habitat relationships. A two-way table represen-

tation of noda derived from normal and inverse

association-analyses, also shows multidimensional

classificatory relationships.

The nature of the vegetation types, apparent from a

Braun-Blanquet Table, explains why sufficiently

homogeneous groups cannot be obtained by asso-

ciation-analysis. Groups of releves in a Braun-

Blanquet table are each distinguised by a number of

species. A single releve need not contain a particular

species or a particular combination (set) of species,

but it must have a certain combination of species,

which may be one of several sets of species, for it

to be placed in a group (cf. Hull, 1964). For two

distinct groups of releves no single species need be, or

usually is 100 per cent constant in and absolutely

restricted to either of the two groups. Only approxi-

mations of such releve groupings can be obtained

with association-analysis, where one species is

selected to represent a discontinuity in which two or

more species are involved. Subsequent misclassifi-

cations make it difficult to relate groups to habitat

and render them less suitable for describing plant

communities. Inverse association-analysis by Williams

& Lambert (1961), suffers from the same limitations as

normal association-analysis. Although nodal analysis

compensates for misclassifications of individual

releves and species, such releves and species, which

may be very typical of other groups, are discarded

as being peripheral information (Lambert & Williams,

1962). The Braun-Blanquet Table Method provides a

polythetic classification of species and releves with

no waste of information.

It is concluded that it would be desirable to classify

releves and species into fairly homogeneous groups by

a polythetic method, to construct a two-way table for

interpreting the reticulate vegetation-habitat relation-

ships, and then to erect a hierarchical classification.

Generally, for two-way table presentation of informa-

tion, a hierarchy consolidating the distribution

patterns in the table as much as possible is desirable.

ACKNOWLEDGEMENTS

The Department of Agricultural Technical Services

made possible this study by seconding me to the

University of Pretoria. The study was undertaken

under supervision of Dr G. K. Theron of the Uni-

versity’s Department of General Botany and Dr D.

Edwards of the Botanical Research Institute.

I wish also to thank Dr M. J. A. Werger and Mr

J. W. Morris for valuable comments.

OPSOMMING

Gegewens wat in die Jack Scott-natuurreservaat in die

Sentrale Bankenveld van Transvaal versamel is, is met behulp

van normale assosiasie-analise verwerk. Aangesien dit geblyk het

dat die metode onvoldoende is om optimale definiering en

rangskikking van plantgemeenskappe te verkry, is dit aangevul

deur die Braun-Blanquet-metode wat as plaasvervanger vir

inverse- en nodale-analise gedien het. Dit het tot beter begrip

van die plantegroei gelei. Die streng hierargiese aard van asso-

siasie-analise verberg intergroepsverwantskappe en het beperkte

waarde by die interpretasie van plantegroei-habitat-verhoudings.

Die monotetiese aard van normale en inverse assosiasie-analise

is ’n verdere beperking en alhoewel nodale-analise daarvoor

kompenseer, word waardevolle inligting in laasgenoemde

proses as perifeer verwerp.

REFERENCES

Acocks, J. P. H. 1953. Veld types of South Africa. Mem. Bot.

Saw. S. Afr. 28: 1-192.

Coetzee, B. J. 1972. ’n Plantsosiologiese studie van die Jack

Scott-natuurreservaat. M.Sc. Thesis, University of Pretoria.

Coetzee, B. J. 1974. A phytosociological classification of the

vegetation of the Jack Scott Nature Reserve. Bothalia ,

11 : 329-347.

Hull, D. L. 1964. The effect of essentialism on taxonomy —

two thousand years of stasis (I). Brit. J. Phil. Sc. 15: 3 14—

326.

Lambert, J. M. & Williams, W. T. 1962. Multivariate methods

in plant ecology. IV. Nodal analysis. J. Ecol. 50: 775-802.

Werger, M. J. A. 1974. On concepts and techniques applied in

the Ziirich-Montpellier method of vegetation study.

Bothalia 1 1 : 309-323.

Williams, W. T. & Lambert, J. M. 1959. Multivariate methods

in plant ecology. I. Association-analysis in plant communi-

ties. J. Ecol. 47 : 83-101.

Williams, W. T., & Lambert, J. M., 1960. Multivariate methods

in plant ecology. II. The use of an electronic digital com-

puter for association-analysis. J. Ecol. 48: 689-710.

Williams, W. T. & Lambert, J. M. 1961. Multivariate methods

in plant ecology. III. Inverse association-analysis. J.

Ecol. 49:717-729.

B. J. COETZEE

369

Book Reviews

Epiphytic Orchids of southern Africa by E. R. Harrison.

Durban: Natal Branch of the Wildlife Protection and Conservation

Society of South Africa. 1972. 107 pp. 48 black and white

figures, 48 distribution maps. Price R4 , 50.

This is a field guide to the indigenous epiphytic orchids of

Southern Africa. It is written and illustrated by Mr E. R.

Harrison, farmer and well-known naturalist of Zululand. The

42 epiphytic orchids are distributed along the southern and east-

ern seaboard from Swellendam in the Cape through Natal and

Swaziland to Messina in the Transvaal. The greatest concentra-

tion of species in South Africa is found in Zululand with 32

species. Mr Harrison was therefore ideally situated for his studies

of this group of plants.

Apart from Bolus’s monumental Orchids of South Africa

(1893-1913), Rolfe’s treatment in the Flora Capensis (1912,

1913) and Schelpe’s recent semi-popular work. An introduction

to the South African orchids (1966), there is very little infor-

mation published on the epiphytic orchids of southern Africa.

This book will serve to satisfy the demand for such information.

The author travelled many thousands of miles visiting known

localities and collecting a vast amount of data. He soon found

that photographing the plants was not satisfactory and so de-

cided to illustrate the book himself.

The book starts off with “Hints for readers". Here the author

explains, with the aid of figures, the vegetative and floral struc-

ture of epiphytic orchids. Then follows a mini-glossary of

botanical terms and a vegetative key to the 18 groups which he

recognizes.

For each species a distribution map, a brief non-technical

description, a discussion of affinities, and a description of the

habitat and origin of the specific epithet are given. In addition,

there is a full-page plate in black and white made up of a habit

sketch and drawings of the flower, fruit and sometimes portions

of the leaf and root. The drawings are simple and unpretentious,

but effective.

Mr Harrison obviously has a very intimate knowledge of the

plants he is writing about. This book is recommended to all

who are interested in orchids.

D. J. B. Killick

Common Trees of the Highveld by R. B. Drummond and

Keith Coates Palgrave. Salisbury: Longman Rhodesia.

1973. pp. 99, 54 black and white photographs, 54 colour plates.

Price R2,50.

In a southern African context the choice of the title "Common

Trees of the Highveld” for a book dealing only with the Rho-

desian Highveld is perhaps unfortunate, because there is also a

highveld in the Republic of South Africa where the name had its

origin. Confusion could have been avoided by simply prefixing

Highveld with Rhodesian.

In this book 54 trees are described and illustrated. The trees

are arranged alphabetically according to family and then genus

and species. The text for each species covers the common names

in English and/or Afrikaans, Shona and Ndebele, distribution,

habitat, a non-technical description and finally, considerable

information about economic uses. It is rather a pity that these

different aspects were not dealt with under separate headings.

For reference purposes this would have been extremely useful.

There is a black and white habit photograph of each species,

also a colour plate by the second author’s late mother. The

colour plates were used in Palgrave’s well-known "Trees of

Central Africa” (1956).

This well-presented and reasonably-priced book will un-

doubtedly prove popular with tree enthusiasts. The A5 format,

which is practically pocket-sized, and the serviceable cover,

make it suitable for field use. One very minor criticism : on page 4

it is stated "This is a synonym, or a name which was applied to

this tree in the past, but which is not now considered valid.”

Surely, in view of Article 6 of the International Code of Botani-

cal Nomenclature (1972), “valid” should read “correct”.

D. J. B. Killick

Geobotanical Foundations of the Middle East by Michael

Zohary. Stuttgart: Gustav Fischer Verlag. Amsterdam:

Swets & Zeitlinger. 1973. 2 vol.,pp. xxxi — 738, 279 fig., 8 colour

plates, 1 maps. Price R106.

This monumental work in two volumes is published as part

3 in the series Geobotanica Selecta which is under the general

editorship of R. Tiixen. In this book Zohary aims at pointing out

the general and specific features and problems characteristic of

the plant ecology and phytogeography of what he calls the Middle

East. In this he has been particularly successful. Within the

“Middle East”, Zohary includes Turkey, Crete, Cyprus, Syria,

Lebanon, Israel, Jordan, Egypt, the Arabian peninsula and

Iran, but possibly it would have been more orthodox to call

this the Near East. In this book an overall review of phyto-

geographical territories, vegetation units and the ecological

relationships of these units, is given for the first time. In twenty

chapters the author gives a comprehensive account of flora,

habitats, plant geography, plant communities and plant

utilization of the area, mainly based on his own investigations

carried out during the last forty-five years. Information provided

by his Israeli colleagues as well as data from the very scattered

literature on this area has been integrated. The author carried

out extensive field work in Israel, Jordan, Sinai, Turkey and

Iran, whereas his studies in the rest of Egypt, Crete, Cyprus,

Lebanon, Syria and Iraq were restricted to a limited length

of time, and in some cases were performed more than thirty

years ago. His knowledge of the Arabian peninsula is entirely

based on literature.

In the first chapter the environmental features, geology,

landforms, climates and soils, are outlined in an ecological

perspective. Examples are frequently quoted to demonstrate a

particular feature. Correctly, the author emphasizes that in the

arid and semi-arid regions, which constitute by far the largest

part of the area under discussion, not only the amount of

precipitation, but particularly also the regularity and the

seasonal distribution of the precipitation are extremely im-

portant ecological factors. Climatic diagrams of many localities

(but not of Cyprus and Crete), illustrate the variety in climates

in the area. A generalized climate map and also a generalized

soils map for the entire area, would have been worthwhile

additions to this chapter.

In Chapter 2 Zohary discusses the rich flora as well as the

historical aspects connected with the botanical knowledge of the

area. A chronological list of travellers, collectors and other

contributors towards the botanical knowledge of the area is

given. Important families and large genera ( Astragalus ) are

briefly discussed. Then the composition of the floras of the

individual countries is reviewed. It is interesting to learn how

poor the Arabian peninsula is floristically when compared with

the other areas, in particular Turkey and Iran, the two countries

richest in flora.

The next chapter gives an outline of phytogeographical

concepts such as the “phytogeographical region” and “specia-

tion". A peculiar manner of expression is found in this section

(p. 79), where it is stated that "phytogeographical regions are

often speciation centres of certain taxonomic groups”. Cause

and effect seem to have been confused here.

Five phytogeographical regions meet in the area under

discussion; the Euro-Siberian, the Mediterranean, the Irano-

Turanian, the Saharo-Arabian and the Sudanian regions. A map

shows their geographical positions. The regions and the sub-

divisions are each critically discussed, reviewing the literature

concerned. Zohary differs from several other phytogeographers

in his exclusion of the Sindian area from the Saharo-Sindian

region and inclusion of it within the Sudanian Region. It is

thus removed from the Holarctic Kingdom to Palaeotropis.

Also, his inclusion of the North African Mauritanian Steppe

province into the Irano-Turanian region is not conventional.

Such decisions Zohary does not take lightly, but discusses

critically. His arguments for the rejection of the Afro-alpine

region in the mountains of the Arabian peninsula (p. 97) are,

however, not convincing.

In the next five chapters (160 pages) the territories occupied by

various phytogeographical regions are discussed in more

detail. The most important communities in the different regions

are mentioned, and their ecology is briefly outlined, sometimes

illustrated by altitudinal transects. The statement in this section,

that “one of the causes of the lack of a Saharo-Arabian Territory

in Iran is that the local Irano-Turanian element displays

an extremely wide ecological differentiation and is capable of

populating the most desertic habitats by plants of its own”

(p. 239), is perhaps too teleological. The five chapters are

illustrated by nearly a hundred pen drawings as well as by

eight colour plates of species mentioned in the text. The necessity

for so many of these illustrations in a book like this is doubted by

the reviewer.

370 IMPROVEMENT OF ASSOCIATION-ANALYSIS CLASSIFICATION BY BRAUN-BLANQUET TECHNIQUE

The first volume concludes with a description of nine longi-

tudinal and latitudinal, long-distance transects through the

entire area, and a chapter on endemism, origin of the flora,

and migration. An outline of concepts is given, and Zohary

distinguishes between diffuse versus compact and neo- versus

palaeo-endemism. Endemism in the various countries is dis-

cussed, from which it becomes apparent that endemics are not

evenly distributed over the area, but that the Irano-Turanian

group of species comprises by far the most endemics in all

kinds of habitats in its territory. Many endemics also occur

on isolated mountain peaks. The importance of the Arcto-

Tertiary, the Indo-Malesian, the Palaeo-African and the Meso-

gean elements in the genesis of the present flora of the area is

outlined, as well as floristic development since the Cretaceous.

Volume 2 opens with a discussion on trees and shrubs as

"edificators” of the vegetal landscape. Distribution records,

which are only accurate within the area under discussion,

and the ecological preferences of most of the trees and shrubs

are described. A table listing the altitudinal ranges of selected

trees and shrubs is included here.

Chapter 12 follows, providing a short, general, theoretical

introduction to the phytosociological description of the area.

Zohary, who follows in general the principles of the Zurich-

Montpellier School, states that although its methods “are

very helpful as a general approach, (they are) delusive when ap-

plied to vegetation in arid zones” (p. 400). He deems modifica-

tions and deviations, mainly in the criteria of delimitation

and characterization of the basic plant ecological units, the

associations, a necessity. Because “the distribution ranges of

species making up the associations are strikingly unequal in size,

there will be a componental overlap between associations and the

merging of communities into one another” (p. 400). Therefore,

one of the modifications Zohary thought necessary is the appli-

cation of dominance as the diagnostic marker of associations. As

Zohary states, however, the vegetation in these Middle Eastern

countries has been exposed to uncontrolled human interference

for millennia, which in many cases will have led to selective

eradication of several plant species. For this reason, and because

facies often occur in the vegetation of arid and semi-arid

regions, the application of dominance for the delimitation of

plant communities is certainly not unequivocally desirable.

It can easily lead to arbitrary decisions in recognizing the

communities and to syntaxonomical difficulties. These recog-

nized units would, perhaps, have better been called communi-

ties instead of associations, as long as they have not been deli-

neated on the basis of comprehensive phytosociological tables.

It is perhaps worth mentioning here, that several ecologists

working in arid and semi-arid regions, including the reviewer,

have experienced no particular difficulty in delineating associ-

ations, as meaningful floristic and ecological entities, on the

basis of total floristic composition. It is, therefore, a puzzling

statement which is made on page 400, saying, “to date, nobody

established associations on the basis of floristic properties

alone”. Zohary possibly encountered difficulties owing to the

long-term human disturbance of the vegetation and to the very

application of the dominance criterion.

In seven chapters, altogether over 200 pages, the plant

ecological units of the deserts, semi-deserts and steppes, the

Mediterranean vegetation, the Euxino-Hyrcanian forests, the

temperate steppe forests (both Euxinian and Irano-Turanian),

the Sudanian vegetation and the hydrophytic vegetation are

outlined. Notes on life and growth forms are also included.

Phytosociological tables are omitted, but selected representative

sample records are given. Most records are originally from

Zohary himself, but those from some of his colleagues are

also utilized. Many communities are illustrated by instructive

photographs. The Arabian countries are, for understandable

reasons, strikingly under-represented in sample records as well as

photographs. Short remarks on the habitats of the communities

are often added. Although Zohary expects his newly-described

associations in the majority of cases to be affirmed by future

phytosociological studies, he explicitly claims no nomenclatural

validity for them. Minor points of criticism on his nomenclature

include the following: he uses the suffix “-eto” after the first

genus name, when an association is named after two species,

instead of the use of binding vocal “o”, which has already

for more than ten years been the recognized practice in the

Ziirich-Montpellier tradition. He sometimes uses the phrase

“association of . . . -eturn”, which is a pleonasm. Stipagros-

tetum should be Stipagrostietum. Sometimes authors are added

to names of phytosociological units of higher rank, but more

often they are omitted. It would acutally have been preferable to

include them in all cases. Together these chapters represent a

most valuable and useful account of the communities and their

syntaxonomical and ecological interrelationships in the enor-

mous area that is dealt with, and Zohary’s statement, that “the

Middle East as a whole is one of the most backward areas as far

as botanical investigations is concerned” (p. 399), is certainly no

longer true. The book concludes with a chapter on the inter-

ferences between man and plants through the ages. This section

comprises short essays on the presegetal and the segetal eras

as well as an anticipation on the neo-segetal and the near-

future “era". Lists of plants used by man, grouped according to

usage, are presented. The nature of the cultivated flora and of

the segetal and ruderal flora and vegetation is outlined, and

short notes on the syntaxonomy as well as some sample records

of these vegetation types are included. The chapter also includes

discussions on the origin of cultivated plants in the Middle

East and on pastoralism and its selective effect on flora and

vegetation.

A literature list of 22 pages, and two indices, a subject

index and an index of plant and community names, of altogether

62 pages, are appended.

This work which comprises such an enormous wealth of

information on plant ecology and phytogeography in general, as

well as that of the area in particular, deserves a wide distribution

and should be available in every library dealing with phyto-

geography or with plant ecology of arid and semi-arid areas.

But also to other scientists, Zohary’s book will prove to be

extremely stimulating. As may be expected from the price,

binding, paper quality, reproduction of photo's and lay-out are

of a high quality. Printer’s errors are few. Unfortunately the

price (R106) will probably put this book beyond the reach of all

but the larger libraries.

M. J. A. Werger

Bothalia 11, 4: 371-442 (1975)

Plectranthus (Labiatae) and allied genera in Southern Africa

L. E. CODD*

ABSTRACT

A revision is presented of the 40 species of Plectranthus , 1 species of Rabdosia and 3 species of Solenostemon

which are indigenous, semi-naturalized or widely cultivated in Southern Africa. Descriptions, illustrations, keys

and distribution data are provided. The following new names are published: P. mutabilis Codd, P. psammo-

philus Codd, P. rubropunctatus Codd, P. unguentarius Codd, P. ornatus Codd (nom. nov. for Coleus comosus

Hochst. ex Guerke), P. zatarhendi (Forsk.) E. A. Bruce var. tomentosus (Benth.) Codd, — var. woodii (Guerke)

Codd, P. madagascariensis (Pers.) Benth. var. aliciae Codd, Solenostemon scutellarioides (L.) Codd and S.

shirensis (Guerke) Codd.

Introduction

History

References

Key to Genera

Plectranthus

Key to Species

Subgen. Nodiflorus Codd

1. P. tetragonus Guerke

2. P. esculentus N.E. Br

Subgen. Xerophilus Codd

3. P. xerophilus Codd

Subgen. Burnatastrum (Briq.) Codd

4. P. candelabriformis Launert

5. P. mirabilis (Briq.) Launert

6. P. hereroensis Engl

7. P. spicatus E. Mey. ex Benth.

8. P. cylindraceus Hochst. ex Benth.

Subgen. Coleus (Lour.) Codd

9. P. unguentarius Codd

10. P. amboinicus Lour

Subgen. Calceolanthus Codd

11. P. tetensis (Bak.) Agnew

12. P. caninus Roth

13. P. neochilus Schltr

14. P. ornatus Codd

15. P. barbatus Andr

Subgen. Plectranthus

Sect. Coleoides Benth

16. P. dinteri Briq

17. P. grandidentatus Guerke

18. P. zatarhendi (Forsk.) E. A. Bruce

(a) var. zatarhendi

(b) var. tomentosus (Benth.) Codd

(c) var. woodii (Guerke) Codd ...

19. P. madagascariensis (Pers.) Benth.

(a) var. madagascariensis

(b) var. aliciae Codd

(c) var. ramosior Benth

CONTENTS

Page

372

441

374

376

377

378

380

381

382

383

385

387

388

389

390

392

393

394

395

396

398

399

401

402

403

404

Page

20. P. mutabilis Codd 404

21. P. psammophilus Codd 405

Sect. Plectranthus 406

22. P. verticillatus (L.f.) Druce 407

23. P. strigosus Benth 409

24. P. purpuratus Harv 410

25. P. oertendahlii Th. Fries jun 411

26. P. elegantulus Briq 412

27. P. ciliatus E. Mey. ex Benth 414

28. P. fruticosus L’Herit 415

29. P. grallatus Briq 418

30. P. rubropunctatus Codd 420

31. P. rehmannii Guerke 421

32. P. swynnertonii S. Moore 422

33. P. dolichopodus Briq 423

34. P. zuluensis T. Cooke

35. P. saccatus Benth.

(a) var. saccatus

(b) var. longitubus Codd 428

36. P. hilliardiae Codd 428

37. P. ambiguus (Bol.) Codd 429

38. P. ecklonii Benth. - 431

39. P. petiolaris E. Mey. ex Benth 431

40. P. laxiflorus Benth 434

Rabdosia 436

Subgen. Pyramidium (Benth.) Codd 436

R. calycina (Benth.) Codd 436

Solenostemon 437

Subgen. Solenostemoides (Briq.) Codd 438

1. S. rotundifolius (Poir.) J. K. Morton ... 438

2. S. latifolius (Hochst. ex Benth.) J. K.

Morton 439

3. S. scutellarioides (L.) Codd 439

Index 441

Technical Services, Private Bag XI 01, Pretoria.

Botanical Research Institute, Department of Agricultural

372

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

INTRODUCTION

The genera dealt with in this revision are Plec-

tranthus L’Herit. (in which Coleus Lour, and certain

other genera are included), Rabdosia (Bl.) Hassk.

{Isodon Kudo) and Solenostemon Schumach.

The clarification of the species problems in these

groups has been made possible largely by the oppor-

tunity to study field populations and by the cultivation

of a wide range of plants in the garden of the Botanical

Research Institute, Pretoria. In addition it has been

possible to consult type material of nearly all the

species during visits to various European herbaria

and by means of loans from several overseas herbaria,

whose kind assistance is gratefully acknowledged.

In this way the nomenclatural confusion in the

literature has largely been cleared up.

The delimitation of genera is, however, far from

clear and is likely to remain so until a thorough

study of the tropical central and east African species

has been undertaken. The Southern African species

now placed in Plectranthus fall into four or five main

groups, each of which could be regarded as a separate

genus. However, the distinctions appear to become

blurred in tropical Africa and so it is considered

inadvisable to erect new genera at this stage. The

treatment of the West African Labiatae by Morton

(1962) has been largely followed, while observations

by Launert (1968) and Blake (1971) are of particular

interest. An outline of my views was recently published

(Codd, 1971) and forms the basis of the present

treatment.

One of the main problems in the Plectranthinae

has centred round the generic limits of Plectranthus

and Coleus. Until recently, Bentham’s treatment

(1832, 1848), based simply on the stamens being free

in Plectranthus and united in Coleus, has been

followed with the result that these two genera have

grown to be the largest in the group. A rough count

in Index Kewensis yielded about 490 species names

in the former and about 312 in the latter, though it

must be remembered that a number of names are dupli-

cated through transfer from one genus to the other.

Unfortunately the stamen character has proved to

be unreliable in certain groups and a different

approach to the classification of these species, based

mainly on calyx shape, has been sought (Morton,

1962; Codd, 1971; Blake, 1971). If the present

tendency is followed, about 100 names will go

from Plectranthus to Rabdosia (91 were transferred

by Hara, 1972) and about a similar number should

be transferred from Coleus to Solenostemon as

emended by Morton (see Codd, 1971). If the remainder

of the Coleus species are incorporated in Plectranthus,

the resulting group, though somewhat heterogeneous,

is not excessively large. A classification along these

lines would simplify the allocation of species to their

respective genera.

HISTORY

Plectranthus L'Herit., Stirp. Nov. fasc. 4: 84

(March 1788) was based on two species, P. punctatus

(L.f.) L’Herit. and P. fruticosus L’Herit. About the

same time Germanea Lam., Encycl. 2: 690 (April

1788) was described, also with two species, G. urtici-

folia Lam. and G. maculosa Lam. P. fruticosus and

G. urticifolia have proved to be conspecific and the

two generic names have generally been treated as

synonymous. Until recently the exact dates of

publication were unknown, so that Plectranthus was

formally conserved. With the dates established as

above, of Plectranthus by Britten & Woodward,

J. Bot. Lond. 43: 268 (1905) and of Germanea by

Rothmaler, Chronica Bot. 5: 439 (1939), conservation

becomes unnecessary.

P. punctatus was proposed as lectotype species by

Miss M. L. Green in Prop. Brit. Bot. 107 (1929),

probably because it is the older epithet. However,

P. fruticosus was cited by Phillips, Gen. ed. 2: 650

(1951) and was supported by Bullock & Killick in

Taxon 6: 239 (1957) on the grounds that the characters

of this species agree better with L’Heritier's generic

description, especially as to the spur on the corolla

which suggested the generic name. This proposal was

accepted (Taxon 9: 85, 1960).

Bentham (1832, 1848) monographed the species

then known, dividing Plectranthus into seven sections:

Germanea (Lam.) Benth., Coleoides Benth., Hetero-

cylix Benth., Melissoides Benth., Isodon Schrad. ex

Benth., Pyramidium Benth. and Amethystoides Benth.

In Benth. & Hook.f., Gen. PI. 2: 1175 (1876),

Bentham revised this arrangement, recognizing two

primary groups: Sect. Germanea and sect. Isodon.

In sect. Germanea, with Germanea and Coleoides as

subsections (in which the great majority of con-

ventional Plectranthus spp. are placed), the calyx is

2-lipped with the upper lip consisting of a single

broad tooth and the lower lip of four narrower

acute or acuminate teeth; the cymes are usually (not

always) sessile with the pedicels arising from the

axis of the inflorescence. Sect. Isodon, with Isodon,

Pyramidium, Amethystoides and Melissoides as sub-

sections, was distinguished by the calyx being equally

5-toothed, in some groups more or less 2-lipped

with the upper lip composed of 3 teeth and the

lower lip of 2 teeth; the cymes are pedunculate and

branched.

A similar classification was adopted by Briquet

(1897, pp. 352-357), though Germanea and Isodon

were treated as subgenera. In each subgenus the

largest sections, Coleoides and Isodon, respectively,

were subdivided into a number of series. A separate

genus, Burnatastrum Briq., was based on three species

previously in Bentham’s Sect. Isodon, namely,

Plectranthus lanceolatus Benth. and P. lavanduloides

Bak. (both from Madagascar) and P. spicatus E. Mey.

ex Benth. (from South Africa). The genus was

separated on grounds of the circinnate calyx and the

flowers being arranged in dichasia. However, when

other species are taken into account, it will be seen

that these characters tend to grade into Plectranthus.

In the present treatment, Burnatastrum is placed

as a subgenus of Plectranthus, while the concept is

enlarged to include P. hereroensis Engl., P. cylin-

draceus Hochst. ex Benth., P. mirabilis (Briq.) Launert

and P. candelabriformis Launert. Certain tropical

species may well belong here also.

L. E. CODD

373

The African species were dealt with by Baker in

FI. Trop. Afr. 5: 398-420 (1900), in which 78 species

were reviewed, and Cooke in FI. Cap. 5, 1 : 266-288

(1910), who records 42 species. Both follow basically

Briquet’s classification into the two groups Germanea

and Isodon, though Cooke omits mention of Burna-

tastrum Briq. In fact, Cooke’s treatment is unsatis-

factory in many respects, with closely related species

being widely separated and several species misinter-

preted.

Briquet’s concept of Isodon is now generally

regarded as being generically distinct from Plec-

tranthus. The species occur mainly in Asia and

Malesia, with a few in Africa (see p. 436 for further

discussion). Asiatic species were separated under the

generic name Isodon Schrad. by Kudo in Mem. Fac.

Sci. & Agr. Taihoku Imp. Univ. 2: 118-141 (1929),

and as Amethyst anthus Nakai in Bot. Mag. Tokyo

48: 785 (1934) for Bentham’s section Amethystoides.

The African species were given separate status in the

genus Homalocheilos J. K. Morton. Blake (1971)

points out that there is an earlier name for the genus,

Rabdosia (Bl.) Hassk. in Flora 25, Beibl. 2: 25 (1842),

based on Elsholtzia Willd. sect. Rabdosia Bl., Bijdr.

FI. Ned. Ind. 825 (1825), the type of which is E.

javanica BL, l.c. It may be noted, however, that

Keng in Gard. Bull. Singapore 24: 13-180 (1969)

largely follows Bentham’s concept of Plectranthus

and Coleus (see Solenostemon, p. 439) in his revision

of the Malesian species.

Mention should also be made of Englerastrum

Briq. in Bot. Jahrb. 19: 178 (1894), based on a

single tropical African species, E. schweinfurthii

Briq. This is a depauperate soft herb with leaflike

bracts subtending slender racemose inflorescences;

the calyx is subequally 5-toothed and the stamens

are shortly united at the base. Fries in Notizbl. Bot.

Gart. Berl. expanded the concept to includes species

such as Plectranthus tetragonus Guerke (see p. 376)

and P. floribundus N.E.Br. (see p. 377). This view

was supported by Alston in Kew Bull. 1926: 295

(1926) but Hutchinson & Dandy in Kew Bull. 1926:

479 (1926) reduced the genus to about seven species

which conform with the original concept. The genus

seems scarcely distinct from Rabdosia, judging by

the leaflike bracts along the stems.

Coleus Lour., FI. Cochinch. 372 (1790), was

based on C. amboinicus Lour., l.c., a species probably

of African origin but widely cultivated in the tropics

(see p. 388). It was pointed out by Loureiro that the

stamens are united at the base and Bentham (1832,

1848) came to regard this characteristic as being of

over-riding importance. Thus, although he acknow-

ledged that he had created an artificial assemblage,

he enlarged Coleus to include three very different

sections:

Sect. 1. Calceolus, containing species such as C.

spicatus Benth. and C. barbatus (Andr.) Benth.

Flowers in dense verticillasters, calyx villous in

the throat, upper tooth of calyx large, broadly

ovate.

Sect. 2. Aromaria, containing the single species

C. amboinicus Lour, which he included under

C. aromaticus Benth. Flowers in dense verti-

cillasters, calyx glabrous in the throat, upper

tooth of calyx large, oblong, horizontal.

Sect. 3. Solenostemon, containing species such as

C. scutellarioides (L.) Benth., C. blumei Benth.

and C. latifolius Hochst. ex Benth. Flowers in

loose to compact dichasia, calyx glabrous within,

upper lip ovate, erect, lateral lobes obsolete or

short and truncate, lower lip united into an

oblong lobe, entire or forked at the apex.

Sect. Aromaria has remained monotypic (though a

second species is added in this revision, p. 387).

The question arises whether to retain it as a separate

genus, when the calyx shape differs only slightly from

Plectranthus, or to follow Morton (1962) and Launert

(1968) who include it in Plectranthus. The latter

course is consistent with my view of variation in

Plectranthus and, in the present revision I have

placed Coleus as a subgenus of Plectranthus (cf. the

treatment of Burnatastrum, p. 380).

Many species names, running into hundreds, have

been added to sections Calceolus and Solenostemon

so that a completely distorted view of the genus

Coleus came to be adopted.

Blake (1971) suggested that Coleus could be upheld

as a genus with sect. Calceolus included in it. However,

sect. Calceolus differs as much from true Coleus as

it does from Plectranthus. An alternative would be

to accord generic rank to sect. Calceolus and, if only

the South African species are considered, this would

seem to be fully justified. However, in tropical Africa

the distinction is far from clear. It is, therefore,

included in Plectranthus as subgen. Calceolanthus

(p. 389).

Sect. Solenostemon derives its name from the genus

Solenostemon Schumach. (see p. 437), based on the

West African S. ocymoides Schumach. & Thonn.

When Bentham (1832) included it in Coleus he

renamed it C. africanus Benth. and enlarged the

concept of the section to include certain species from

Madagascar, Asia and Malesia in which the calyx

has a somewhat different shape. In typical Solenoste-

mon the two lowest calyx teeth are fused into an

ovate to oblong, entire or emarginate lip which curves

upwards closing the mouth of the tube, while the

lateral teeth are reduced or obsolete. In the species

which Bentham added, the lower lip of the calyx is

strap-shaped and forked at the apex, while the

lateral calyx teeth are short and rounded to deltoid.

In his later treatment, Bentham (1848) added the

African species C. latifolius Hochst. ex Benth., and

many more species from Africa and elsewhere have

subsequently been described (Codd, 1971). Benth. &

Hook, f., Gen. PI. 2: 1175 (1876) reinstated Sole-

nostemon Schumach. in its strict sense while the

additional species were retained in Coleus. Briquet’s

(1897) treatment was essentially similar, though here

the additional species were given sectional status in

Coleus as sect. Solenostemoides Briq.

With the transfer of true Coleus (C. amboinicus

Lour.) to Plectranthus, Morton (1962) included

sect. Solenostemoides Briq. in the genus Solenostemon,

but gave it the name sect. Coleoidea, pointing out

that it is, in some respects, intermediate between

true Solenostemon and Plectranthus. However, the

calyx is easily recognizable and separation from

Plectranthus appears to be fully justified.

The genus Solenostemon, as amended by Morton,

is common in tropical Africa and Asia, extending

to Malesia. The species are, however, very variable

and considerable difficulty in determining species

limits will be encountered when the group comes to

be revised. Probably one of the best known is the

commonly cultivated “Coleus” with variegated leaves

known usually as Coleus blumei. This may have

influenced Keng (1969) to retain Coleus in the

Benthamian sense for the Malesian species.

374

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

KEY TO GENERA

Calyx 2-lipped to subequally 5-toothed, if 2-lipped then the lower lip consisting of 4

distinct lanceolate-deltoid or subulate teeth:

Bracts small, abruptly differentiated from the leaves 1. Plectranthus

Bracts leaf-like, becoming gradually smaller towards the apex 2. Rabdosia

Calyx 2-lipped, the lowest pair of teeth united for more than two-thirds their length

forming an oblong, emarginate lobe, the lateral pair of teeth much shorter,

truncate 3. Solenostemon

1. PLECTRANTHUS

Plectranthus L'Herit., Stirp. Nov. fasc. 4: 85

(March 1788); Willd., Sp. PL 3: 168 (1800); Benth.,

Lab. 29 (1832); in DC., Prodr. 12: 62 (1848); Benth. &

Hook, f., Gen. PI. 2: 1175 (1876); Briq. in Pflanzen-

fam. 4, 3a: 352 (1897); Bak. in FI. Trop. Afr. 5: 398

(1900); Cooke in FI. Cap. 5,1: 266 (1910); Phillips,

Gen. ed. 2: 649 (1951); Morton in J. Linn. Soc. (Bot.)

58: 231 (1962); Launert in Mitt. Bot. Miinchen 7: 295

(1968); Codd in Mitt. Bot. Miinchen 10:245 (1971);

Blake in Contr. Queensl. Herb. 9: 1-120 (1971).

Lectotype: P. fruticosus L’Herit.

Germanea Lam., Encycl. 2: 690 (April 1788); Hiern,

Cat. Afr. PI. Welw. 1 : 865 (1900). Type: based on two

species, G. urticifolia Lam. and G. maculosa Lam.

Coleus Lour., FI. Cochin. 372 (1790); emend.

Benth., Lab. 47 (1832), partly; in DC., Prodr. 12: 70

(1848); Benth & Hook.f. 2:1176 (1876); Briq. in

Pflanzenfam. 4,3a: 359 (1897); Bak. in FI. Trop. Afr.

5: 422 (1900); Cooke in FI. Cap. 5,1: 289 (1910);

Phillips, Gen. ed. 2: 649 (1951), all partly. Type: C.

amboinicus Lour.

Neomuellera Briq. in Bot. Jahrb. 19: 186 (1894).

Type: N. welwitschii Briq.

Burnatastrum Briq. in Pflanzenfam, 4, 3a: 358

(1897). Lectotype: B. spicatum (E. Mey. ex Benth.)

Briq.

Ascocarydion G. Tayl. in J. Bot. Lond. 69, Suppl.

2: 162 (1931). Type: A. mirabile (Briq.) G. Tayl.

Annual or perennial herbs or subshrubs; stems

and leaves often succulent or semi-succulent.

Inflorescence a lax panicle or racemose cyme, often

congested or subspicate; flowers in verticils, few-

flowered cymes or dichasia, occasionally solitary;

bracts small, abruptly differentiated from the leaves.

Calyx 2-lipped to subequally 5-toothed; when

2-lipped, upper tooth the largest, lower lip of 4

lanceolate-deltoid or subulate teeth; tube often

villous within, sometimes gibbous below. Corolla

bilabiate, tube usually bent and variously expanded

near the base, occasionally expanding gradually or

straight; upper lip usually 4-lobed, shorter than the

lower lip; lower lip entire, boat-shaped. Stamens 4,

rarely 2 abortive ( P . zuluensis), attached at the mouth,

of the corolla tube, free or united in a sheath at the

base, didynamous, declinate in the lower lip of the

corolla; anthers 1-thecous. Style lying with the

stamens in the lower lip of the corolla; stigma shortly

2-lobed.

A genus of about 300 species, common in Africa

South of the Sahara and extending to southern Arabia,

India and Australasia. Of the 40 species dealt with

below, 38 are indigenous to Southern Africa, the two

semi-naturalized species being P. ornatus Codd

(= Coleus comosus Hochst. ex Guerke) and P.

barbatus Andr.

Key to Species

Flowers yellow, in pseudoracemes borne terminally as well as from the upper nodes of the usually leafles

stems :

Plants annual; stems with conspicuous bristles; corolla 4-5 mm long 1. P. tetragonus

Plants perennial with edible tuberous rootstock; stem without bristles; corolla 14-16 mm long. .2. P. esculentus

Flowers white or shades of blue, violet or purple (rarely yellow), disposed in cymes, verticils or dichasia;

inflorescence usually terminal, paniculate, racemose or subspicate, borne on leafy stems:

Mature calyx subequally 5-toothed, often erect or finally circinnate (in P. cylindraceus the uppermost

calyx tooth is slightly larger than the other 4 but is difficult to see because of the dense covering

of hairs):

Flowers in 10-20-flowered sessile cymes; racemes slender up to 35 cm long, peduncle up to 30cm long;

plants with long horizontal tuberous roots 3. P. xerophilus

Flowers in pedunulate or sessile paired cincinni (often compact and glomerate in P. cylindraceus and

P. spicatus) or in 3-flowered pedunculate cymes; roots not tuberous:

Flowers in 3-flowered pendunculate cymes, forming a diffusely branched panicle 30-40 cm

long 4. P. candelabriformis

Flowers in pedunculate or sessile paired cincinni:

Leaves broader than 3 cm, chartaceous or leathery; inflorescence a lax or dense panicle less than

30 cm long, flowers blue:

Leaves thick-textured, densely grey velvety-tomentose below; robust plants with erect, sparingly

branched tomentose stems up to 2 m tall 5. P. mirabilis

Leaves thin-textured, subglabrous to sparingly pubescent below; herbaceous, branched plants

usually less than 1 m tall 6. P. hereroensis

Leaves less than 3 cm broad, semi-succulent; inflorescence subspicate, flowers in clusters, mauve,

purple or, rarely, whitish:

Stems decumbent ; flowers purple, subglabrous, in loose clusters ; corolla 7-8 mm long . .7 . P. spicatus

Stems usually erect; flowers mauve (rarely whitish or pale yellow), tomentose, in densely

glomerate clusters; corolla 4-5 mm long 8. P. cylindraceus

Mature calyx with upper tooth distinctly broader than the rest, oblong to ovate or subrotund, remaining

4 teeth deltoid to subulate; calyx finally horizontal, teeth spreading:

Upper tooth of calyx horizontal, oblong to ovate, usually rounded at the apex; flowers in glomerate,

densely tomentose clusters:

Leaves obovate, cuneate at the base; corolla 4-5 mm long 8. P. cylindraceus

Leaves ovate, broadly truncate to cordate at the base; corolla more than 5 mm long:

Stems erect, woody at the base; corolla white, 10-12 mm long 9. P. unguentarius

L. E. CODD

375

Stems procumbent, succulent; corolla mauve to whitish, 7-9 mm long 10. P. amboinicus

Upper tooth of calyx erect, ovate-deltoid to broadly ovate or subrotund; inflorescence paniculate,

racemose or subspicate:

Mature calyx densely villous in the throat; stamens united at the base; inflorescence subspicate

with pedicels erect, appressed to the rhachis:

Bracts rounded at the apex, subpersistent; stems procumbent, slender, sparingly

branched 11. P. te tens is

Bracts acute to abruptly acuminate, early deciduous, forming a conspicuous imbricate coma at the

apex of the inflorescence; stems erect to procumbent, sometimes mat-forming:

Erect or spreading semi-succulent herbs up to 60 cm tall; leaves ovate-lanceolate to obovate,

2- 5x1, 5-3, 5 cm:

Corolla less than 1 cm long; annual plants 12. P. caninus

Corolla exceeding 1 cm long; perennial or weakly perennial plants:

Corolla 1-2 cm long; inflorescence elongate, 7-15 cm long with 5-12 spaced fruiting

verticils below the flowers; indigenous 13. P. neochilus

Corolla 2-2,5 cm long; inflorescence compact, 3-5 (-9) cm long with 1 or 2, rarely more,

spaced fruiting verticils below the flowers; cultivated or semi-naturalized. .14. P. ornatus

Erect bushy herbs up to 2 m tall; leaves not succulent, ovate to broadly elliptical, 5-9 cmx

3- 5 cm; cultivated or semi-naturalized 15. P. barbatus

Mature calyx glabrous in the throat; stamens free to the base; inflorescence usually paniculate, if

subspicate, then pedicels spreading to ascending:

Bracts deciduous before the flowering-stage (occasionally persisting in abnormal cases); fruiting

calyx gibbous ventrally; flowers in dense verticils, (3-) 4-12 to each bract scar:

Stems erect or decumbent; flowers mauve to purple (rarely white):

Leaves deeply dentate; rhachis coarsely glandular-hispid, pubescence often yellowish (S.W.

Africa) 16. P. dinteri

Leaves crenate-dentate; rhachis sparsely to densely glandular-tomentose, pubescence greyish :

Leaves 4-10 cm long, if less then sparingly to fairly densely strigose:

Leaves densely tomentose on both surfaces:

Stems 30-60 cm tall; inflorescence 8-30 cm long, simple or with a pair of branches

near the base 18a. P. zatarhendi var. zatarhendi

Stems 50-150 cm tall; inflorescence 20-60 cm long, usually with 1 or 2 pairs of

branches near the base 18b. P. zatarhendi var. tomentosus

Leaves sparingly to fairly densely strigose 18c. P. zatarhendi var. woodii

Leaves 2,5-4 cm long; stems 1-several often from a burnt perennial base

19c. P. madagascariensis var. ramosior

Stems procumbent; flowers white, mauve or blue:

Corolla 7-18 mrn long:

Leaves deeply dentate or deeply and coarsely crenate-scalloped :

Leaves deeply dentate, densely tomentose; flowers white 17. P. grandidentatus

Leaves deeply and coarsely crenate-scalloped, medium to densely strigose; flowers

purple-blue to lilac 20. P. mutabiiis

Leaves crenate-dentate :

Leaves 4-10x3,2-10 cm, densely tomentose; flowers usually mauve (rarely white)

18b. P. zatarhendi var. tomentosus

Leaves 1,5-4 (-4, 5) X 1 ,2-3,5 (-4) cm, sparingly to densely short tomentose; flowers

usually white (rarely mauve) 19a. P. madagascariensis var. madagascariensis

Corolla 5-6 mm long:

Leaves coarsely crenate with 3-4 pairs of rounded teeth; corolla white (Transkei)

19b. P. madagascariensis var. aliciae

Leaves obscurely crenate-dentate with 5-7 pairs of shallow teeth; corolla blue-mauve

(Zululand) 21. P. psammophilus

Bracts (often very small) persisting to the flowering stage; fruiting calyx enlarged and often oblique

but not conspicuously gibbous ventrally; flowers in lax verticils with 1-3 flowers to each

bract or in pedunculate 3-8-flowered cymes:

Corolla tube expanding abruptly at or near the base and declinate at or near the mouth of the

calyx, often saccate or spurred dorsally:

Corolla tube less than 10 mm long (occasionally up to 10 mm long in P. zuluensis and then

stem and leaves velvety pubescent):

Fertile stamens 2, staminodes 2; leaves softly velvety pubescent below.. 34. P. zuluensis

Fertile stamens 4; leaves subglabrous to strigose pubescent:

Leaves usually not exceeding 4 cm in length:

Leaves dotted below with minute red gland-dots (also on calyx and corolla):

Corolla tube scarcely narrowed near the throat; upper corolla lip 5-8 mm long;

leaves glabrous to strigose 22. P. verticillatus

Corolla tube narrowed near the throat; upper corolla lip 3-5 mm long; leaves rusty-

strigose or densely grey-tomentose :

Stems and leaves rusty-strigose; petioles up to 2,5 cm long 23. P. strigosus

Stems and leaves densely grey-tomentose; petioles up to 1 cm long.. .24. P. purpuratus

Leaves dotted below with minute colourless or honey-coloured gland-dots;

Straggling herb up to 20 cm tall ; corolla white with a few purple spots . . 26. P. elegantulus

Erect to straggling herb 25-100 cm tall; corolla sky-blue 33. P. dolichopodus

Leaves usually exceeding 4 cm in length (sometimes smaller in P. ciliatus but then leaves

and calyx ciliate with multicellular purple-striped hairs):

Flowers in 3-8-flowered, often pedunculate cymes; corolla white, tomentose; upper

lip of corolla 2 mm long 31. P. rehmannii

376

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Flowers in 1-3-flowered sessile cymes; corolla glabrous or sparsely hairy:

Leaf margin with few (6-14) pairs of large teeth 8-10 mm long which bear small

secondary teeth; corolla whitish with a fringe of hairs on the lower lip

32. P. swynnertonii

Leaf margin not as above, if coarsely toothed with small secondary teeth then

usually more than 16 pairs of teeth; corolla glabrous to minutely pubescent:

Leaves dotted below with minute colourless or honey-coloured gland-dots:

Corolla white, mauve or pink and speckled with purple; corolla tube slightly

narrowing towards the throat; upper lip of corolla 3-7 mm long:

Leaf margin and calyx ciliate with purplish multicellular hairs; stems decum-

bent; corolla whitish freely speckled with purple 27. P. ciliatus

Leaf margin not ciliate, calyx with occasional multicellular hairs; stems

erect; flowers mauve or pink with darker markings 28. P. frutieosus

Corolla sky-blue; corolla tube widening from 1,5 mm at the base to 2,5 mm at

the throat; upper lip of corolla 1,5-2 mm long 33. P. dolichopodus

Leaves dotted below with red gland-dots:

Leaves often thin-textured, apex acute, base abruptly cuneate, margin with

coarse teeth often bearing small secondary teeth 29. P. grallatus

Leaves usually thick-textured, apex obtuse to rounded, base truncate, shortly

attenuate or markedly decurrent on the petiole, margin regularly

crenate 30. P. rubropunctatus

Corolla tube more than 10 mm long or, if less, then upper lip of corolla about 10 mm or more

long and equally broad ( P . saccatus ):

Corolla white, tube 10-13 mm long distinctly narrowing towards the throat; stamens 3 mm

long 25. P. oertendahlii

Corolla usually blue, if white then over 20 mm long, not or scarcely narrowing towards

the throat; stamens 8-10 mm long:

Leaves broadly ovate to ovate-deltoid, subglabrous, base truncate; upper lip of corolla

10-16 mm long and equally broad:

Corolla tube 8-15 mm long 35a. P. saccatus var. saccatus

Corolla tube 20-26 mm long 35b. P. saccatus var. longitubus

Leaves broadly elliptical to obovate-elliptical, sparingly strigose, base cuneate; upper lip

of corolla 5-6 mm long, tube 23-27 mm long 36. P. hilliardiae

Corolla tube expanding gradually from the calyx mouth or nearly parallel-sided for entire length,

straight or curved:

Corolla tube straight, more than 12 mm long:

Corolla tube 20-25 mm long, nearly parallel-sided 37. P. ambiguus

Corolla tube 12-16 mm long, widening slightly towards the throat 38. P. ecklonii

Corolla tube curved, 7-10 mm long, rather like a miniature “Dutchman’s Pipe”:

Leaves coarsely dentate, truncate at the base: corolla purple 39. P. petiolaris

Leaves regularly crenate-dentate, cordate; corolla white with vertical mauve stripes

on upper lip 40. P. laxiflorus

Subgen. Nodiflorus Codd, subgen. nov., a subgen.

Plectrantho structura inflorescentiae, floribus in

pseudoracemis brevibus dispositis, pseudoracemis

nodis superis caulium exorientibus differt.

Type: Plectranthus tetragonus Guerke.

The species included in this subgenus tend to

flower after the leaves have fallen. The flowers are

borne singly or in pairs in short pseudoracemes which

arise at the nodes along the upper half or two-

thirds of the stem. In the two species included in the

present treatment, flowers are yellow which is an

unusual colour in the genus Plectranthus. If a narrow

view were taken of generic limits it is probable that

these differences would warrant separate generic

rank. In this connection it may be noted that P.

defoliatus Hochst. ex Benth. shows a distinct relation-

ship to P. esculentus N.E.Br., though the flowers are

1-3 in the axils of each bract and the calyx is more

or less equally 5-toothed. The two species now dealt

with are as follows:

1 . P. tetragonus Guerke

2. P. esculentus N.E.Br.

P. tetragonus is an annual while P. esculentus is a

perennial with underground tubers which are edible,

with the result that it is cultivated in various parts of

Africa.

I. Plectranthus tetragonus Guerke in Bot. Jahrb.

19; 109 (1894); Bak. in FI. Trop. Afr. 5: 401 (1900);

Launert & Schreiber in Prodr. FI. S.W. Afr. 123: 26

(1969). Type: Tanzania, Usambara, Mashena, Holst

3573.

P. melanocarpus Guerke, l.c. 109 ( 1894); Bak., l.c. 402 (1900).

Type: Tanzania, Massai steppe, Fischer 511. P. biflorus Bak.,

l.c. 402 (1900). Type: Malawi, between Kondowe and Karonga,

Whyte s.n. (K).

Annual, erect, 30 cm-1 m tall; stems solitary from

a fibrous root system, 4-angled, unbranched or

sparingly branched near the base, leafless at flowering,

subglabrous to minutely tomentulose below and with

conspicuous patent bristles 2-5 mm long above.

Leaves “oblong, 8-10 cm long, 5-6 cm broad,

decurrent on the 5-7 cm petiole, glabrous on both

surfaces or sometimes pubescent on the nerves below”

(fide Guerke). Inflorescences produced terminally

and in the axils of the leaves often from near the base

to the apex of the stem, simple or branched, racemose,

4-12 cm long; rhachis densely glandular-hispidulous;

bracts ovate to elliptic, 1-2 mm long, persisting beyond

the flowering stage. Flowers solitary or occasionally

in pairs in the axils of each bract, opposite, sub-

opposite or alternate, 2-3 mm apart; pedicels 3-4 mm

long, minutely glandular-hispidulous. Calyx 2 mm

long at flowering enlarging to 8-9 mm long in fruit,

tubular and slightly bent, glandular-hispid; upper lip

horizontal, ovate-rotund, 2 mm long, obtuse to

apiculate at the apex; lower lip subequally 4-toothed,

2 mm long, teeth linear-lanceolate, subulate, the lower

pair slightly the longer. Corolla yellow, 4-5 mm long,

puberulous; tube 2-2,5 mm long, slightly geniculate

near the base, I mm deep; upper lip erect, 0,75 mm

long, emarginate; lower lip shallowly boat-shaped

2-2,5 mm long. Stamens free at the base, curved

and enclosed in the lower lip, 2-2,5 mm long. Style

coinciding with the stamens. Fig. 1.

L E. CODD

377

Fig. 1. — Plectranthus tetragonus, Grootfontein District, South

West Africa (Nordenstam 2587 in M), x-j.

A tropical African species occurring in Tanzania,

Mozambique, Rhodesia, Zambia and Angola and

entering the northern part of South-West Africa,

where it is found in dry woodland, usually in rocky

places.

S.W.A. — 1917 (Tsumeb): Gross Otavi Mts., between farms

Gross Otavi and Auros (-DA), Nordenstam 2587 (M); Otavi

valley, Asis (-DA), Volk 658 (M).

This is a distinctive species with several remarkable

features, namely, the flowers being solitary (rarely

two) in the axils of the bracts (no doubt by the

reduction of the usual cymes) resulting in short

inflorescences of racemes (often branched), often

from low down on the leafless stems to the apex, the

conspicuous bristles on the stems and the long,

slightly curved calyx tube, producing a distinctive

facies.

Type material of P. tetragonus has not been seen.

The present identification is based on the specimens

so named in the Prodr. FI. S.W.Afr., which are

conspecific with specimens named P. biflorus Bak. at

Kew. Giirke distinguishes P. tetragonus from P.

melanocarpus on the basis of its larger rotund-ovate

bracts and the shape of the upper calyx tooth. On this

basis, all the material seen would be classified as

R. melanocarpus but no doubt these characters vary

and we are dealing with one species, so the name

adopted by Launert and Schreiber is maintained.

2. Plectranthus esculentus N.E.Br. in Kew Bull.

1894: 12 (1894); in Hook., Icon. PI. 25: t. 2488 (1896);

Cooke in FI. Cap. 5,1: 185 (1910); Codd in Mitt.

Bot. Munchen 10: 249 (1971); Ross, FI. Natal 305

(1972). Type: Natal, cult. Botanic Garden, Durban,

Medley Wood 3633 (K !, holo.).

P. ftoribundus N.E.Br., l.c. 12 (1894); in Hook., Icon. PI. 25:

t. 2489 (1896); Cooke in FI. Cap. 5,1: 273 (1910). Lectotype:

(Robyns & Lebrun, 1928): Natal, Inanda, Medley Wood 646

(K !). — var. longipes N.E.Br., l.c. 13 (1894); Bak. in FI:

Trop. Afr. 5: 403 (1900). Lectotype (Robyns & Lebrun, 1928).

Rhodesia, Umzingwani River, Baines s.n. (K).

Coleus dazo A. Chev., Veg. Utiles de l'Afr. Trop. Franc. 1,1:

106 (1905). Type: from West Africa. C. ftoribundus (N.E.Br.)

Robyns & Lebrun in Rev. Zool. & Bot. Afr. 16: 359 (1928);

Ann. Soc. Sci. Brux. ser. B. 49: 96 (1929), nom. illegit., non

C. ftoribundus Bak. (1900). — var. longipes (N.E.Br.) Robyns &

Lebrun, l.c. 360 (1928). C. esculentus (N.E.Br.) G. Tayl. in

J. Bot.Lond. 69, Suppl. 2: 158 (1931).

Englerastrum floribundus (N.E.Br.)Th. Fries jun. in Notizbl.

Bot. Gart. Berlin 9: 73 (1924). — var. longipes (N.E.Br.) Th.

Fries jun. l.c. 77 (1924).

Erect to decumbent, perennial, aromatic herb or

suffrutex from a tuberous-rooted base, 60-120 cm

tall; stems 1 — several from the base, unbranched or

sparingly branched, 4-angled, leafless at flowering,

shortly retrorse pubescent. Leaves drying fairly thick

in texture, subsessile; blade oblong-elliptic to

oblanceolate, 5-8 cm long, 1 ,3-2,5 cm broad, shortly

scabrid hispidulous on both surfaces, dotted with

brown gland-dots below; apex obtuse to rounded;

base obtuse to cuneate; margin obscurely denticulate.

Inflorescences occupying the apical 20-60 cm of the

stem consisting of 2-4 racemes (occasionally branched)

arising at each node; racemes 3-5 cm long; rhachis

shortly and fairly densely hispidulous; bracts ovate

to obovate 2-3 mm long, hispidulous, persisting

beyond the flowering stage. Flowers appearing after

the leaves are shed, solitary in the axil of each bract,

opposite to sub-opposite, 2-5 mm apart; pedicels

3- 5 mm long, minutely hispidulous. Calyx 4-5 mm

long at flowering enlarging to 9-10 mm long in fruit,

campanulate, glandular-hispidulous and gland-dotted;

upper lip horizontal, 3 mm long and equally broad,

obtuse to rounded at the apex; lower lip subequally

4- toothed, teeth linear-lanceolate, the median pair

3 mm long, the lower pair 3,5 mm long. Corolla

yellow, 14-16 mm long, puberulous; tube 6-8 mm

long, 1 mm deep at the base, slightly geniculate and

expanding about the middle to 2,5 mm at the throat;

upper lip erect, 2 mm long, emarginate and with

2 small ear-like lateral lobes; lower lip deeply boat-

shaped, 7-8 mm long. Stamens usually united at the

base, curved and enclosed in the lower lip, the lower

pair the longer, 6-7 mm long. Style exceeding the

stamens by 1-2 mm and finally exserted from the

lower lip by 2 mm. Fig. 2.

Distributed from Equatorial Africa southwards

to Angola, the eastern Transvaal and coastal Natal,

in dry, wooded country, often spread by cultivation.

Transvaal. — 2230 (Messina): near Witvlag (-CC), Pole

Evans 3709 ; Entabeni (-CC), Taylor 803. 2330 (Tzaneen):

Tshakoma (-AB), Van Warmelo sub TRV 36173; Modjadji's

Reserve (-CB), Krige 223; Velcich s.n.; Magoebaskloof (-CC),

Gerstner 5436; 5821. 2430 (Lydenburg): Nelspruit (-BD),

Rogers 21441; Waterval Boven (-CB), Rogers 4775; Mason s.n.;

Godwan River (-DA), Prosser 1259; Kaapsche Hoop (-DB),

Wager sub TRV 15575. 2531 (Komatipoort): White River

(-AC), Wessels s.n.; Plaston (-AC), Holt 323; Barberton (-CC),

Galpin 591.

Swaziland. — 2631 (Mbabane): Bremersdorp (-AD),

Fleishack sub PRE 15166; Stegi (-BD), Keith s.n.; Hlatikulu

(-CD), Stewart sub TRV 8899.

Natal. — Cultivated: Cult. Botanic Gardens, Durban.

Medley Wood 3633 (K); 5620 (K). 2731 (Louwsburg): Ngome

(-CD), Tustin s.n. 2831 (Nkandla): Ngoya (-CD), Sim 2945.

2930 (Pietermaritzburg): Inanda (-DB), Medley Wood 646;

3843 (K). 3030 (Port Shepstone): Umgaye (-AD7), Rudatis 1105.

378

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 2. — Plectranthus esculentus, show-

ing the edible roots (Magoebas-

kloof, Gerstner 5821), X 1.

P. esculentus is related to P. tetragonus, as shown

by the flowers appearing in short pseudo-racemes

after the leaves have been shed and the yellow corolla

colour, which is unusual in the genus Plectranthus.

The chromosome number 2n = 24 recorded by De

Wet (1958) is also a deviation from the basic number 7

usual for Plectranthus subgen. Plectranthus.

Taylor, I .c. was the first to unite P. esculentus and

P. florihundus and his choice of the former epithet

(which is unavoidable in the genus Coleus) must be

followed. The inclusion of C. dazo in synonymy also

seems to be correct judging by a specimen of Chevalier

7257, so named, seen in G. It is probable that P.

densus N.E.Br. and P. primulinus Bak. should also be

included but an investigation of these two species

has not been undertaken.

Subgen. Xerophilus Codd, subgen. nov., a subgen.

Plectrantho calyce subaequaliter 5-dentato, labio

antico corollae cucullato et staminibus basi breviter

connatis differt.

Type species: Plectranthus xerophilus Codd.

The single species in this subgenus, P. xerophilus,

is very distinctive and was at first considered to be

worthy of separate generic status. If a narrow view

were taken of Plectranthus , such a step would be

justified. The relationships of the species are discussed

after its description below.

3. Plectranthus xerophilus Codd in Bothalia 1 1 :

282 (1974). Type: Transvaal, Lydenburg District, near

Marone, Codd & Dyer 7729 (PRE!, holo.).

Perennial slender shrub with thick horizontal

tuberous roots; stems erect, semi-woody, terete to

obscurely 4-angled, sparingly branched, 1-1,7 m

tall (including inflorescence), finely grey tomentose.

Leaves subsessile to shortly petiolate, petiole 2-5 mm

long, finely grey tomentose; blade subcoriaceous,

ovate to elliptic, 3,5-9 cm long 2,5-7 cm broad,

upper surface dark grey-green, strigose, lower surface

reticulate, densely grey tomentose, tomentum of

crisped multicellular hairs, short gland-tipped hairs

and numerous sessile dark reddish-brown gland-dots;

apex obtuse to rounded; base cuneate to truncate;

margin coarsely crenate. Inflorescence terminal, sub-

spicate or paniculate with 1-3 pairs of basal branches,

carried on slender peduncles up to 30 cm long;

racemes up to 35 cm long; bracts early deciduous,

lanceolate to linear-lanceolate, acuminate, 1,5-3 mm

long, densely crisped tomentulose. Flowers densely

clustered in 1 2-20-flowered verticillasters, verti-

cillasters 0,3-2, 5 cm apart; pedicels 2-4 mm long,

densely crisped tomentulose. Calyx 2 mm long at

flowering stage enlarging to 4 mm long in fruit,

subequally 5-toothed, densely crisped tomentulose

and gland-dotted; teeth lanceolate, up to 1,5 mm

long, the uppermost slightly broader than the rest.

L. E. CODD

379

Corolla violet to mauve-purple, crisped tomentulose

and gland-dotted without, glabrous within, bilabiate;

tube 1 mm wide for 2 mm and then curved upwards,

expanding into the upper lip, 4 mm wide at the

throat; upper lip hooded, 3 mm long, obscurely

4-lobed; lower lip boat-shaped, 4-6 mm long.

Stamens shortly connate at the base, 7-8 mm long,

curved upwards in the lower corolla lip. Style curved

upwards, exserted from the throat by 9-10 mm.

Fig. 3, 4.

Fig. 3.— Plectranthus xerophilus, Lydenburg District ( Codct 8504), X

380

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 4. — Plectranthus xero-

philus, Lydenburg District

{Hardy 2259).

Found on dry rocky slopes in the eastern and

northern Transvaal and shows a marked ability to

survive under hot, arid conditions. Its flowering

season is from March to May.

Transvaal. — 2229 (Waterpoort): 6 km N. of Mara (-DC),

Meeuse 10199. 2231 (Phalaborwa): Kruger National Park,

Makadzi area (-CC), Van der Schijjf 3853. 2429 (Zebediela):

24 km E. of Malipsdrift (-BB), 31/5/35, Obermeyer <5 Verdoorn

23; between Chuniespoort and Malipsdrift (-BC), Pole Evans

4370; near Chuniespoort Police Station (-BD), Meeuse 10351.

2430 (Pilgrim’s Rest): 3 km S. of Mica (-BB), Leach 11637 ;

3 km S.E. of Steelpoort (-CA), Codd & Dyer 7712; 50 km N. of

Burgersfort (-CA), Leach & Jones 13249; 30 km N. of Steelpoort

Bridge on road to Penge Mine (-CB), Codd 10489; near Marone

(-CD), Codd & Dyer 7729; Codd 8504; 16 km N. of Ohrigstad

(-DA), Codd 10020; Echo Caves Motel, 30 km S. of Erasmus

Pass tunnel (-DA), De Winter 7725; Blyde River Gorge near

Mariepskop (-DB), Marsh s.n.; 48 km N. of Lydenburg (-DC),

Dyer 3959.

P. xerophilus was first collected near Malipsdrift

in 1935 ( Obermeyer & Verdoorn 23) and has been

cultivated in our Botanic Garden for more than 20

years.

The delay in describing it may be largely attributed

to uncertainty regarding its correct generic position.

In certain respects its characteristics would place it

in the genus Coleus, for example, the densely glomerate

verticils, the bracts being very early deciduous and

the stamens being shortly connate at the base. On

the other hand the calyx shape, which is subequally

5-toothed, is not found in Coleus, but is reminiscent

of Burnatastrum, a genus usually included in Plec-

tranthus.

The modern tendency to take a broad view of

Plectranthus, and to include Coleus within it, removes

some of these difficulties. However, P. xerophilus

possesses characteristics which are unusual even in

this broader concept of Plectranthus, for example

the thick horizontal roots, the slender woody stems

which are subterete, not 4-angled, and the peculiar

hooded upper lip of the corolla. Consideration was

given to according it separate generic status, but

there appears to be insufficient grounds for such a

view if a broad concept of Plectranthus is adopted.

The dense tomentum on the underside of the

leaves of P. xerophilus and the fact that the bracts are

shed at a very early stage gives this species a super-

ficial resemblance to Plectranthus Sect. Coleoides

Benth., including species such as P. zatarhendi

Forsk. and P. tomentosus Benth., but these species

have the conventional Plectranthus corolla and

I -f 4 calyx shape (i.e. the upper calyx tooth much

larger than the other four).

Subgen. Burnatastrum ( Briq .) Codd, stat. nov.

Burnatastrum Briq. in Pflanzenfam. 4,3a: 358

(1897).

Plectranthus Sect. Isodon Schrad. ex Benth., Lab.

40 (1832); in DC., Prodr. 12: 55 (1848), partly.

Lectotype: P. spicatus E. Mey. ex Benth.

Briquet based his genus Burnatastrum on the

South African P. spicatus E. Mey. ex Benth. and two

Madagascar species, P. lanceolatus Benth. and

P. lavanduloides Bak. In separating the genus from

Plectranthus, Briquet stressed the remarkable

circinnate, equally 5-toothed calyx. The structure

of the inflorescence is also different, the flowers being

arranged in scorpioid dichasia, not in small, shortly

pedunculate or sessile cymes as in Plectranthus.

If Burnatastrum were upheld as a distinct genus,

one would almost certainly include in it certain allied

species, e.g. P. cylindraceus Hochst. ex Benth. (calyx

slightly circinnate), P. hereroensis Engl, and P.

mirabilis Briq. (calyx erect, not circinnate, markedly

ventricose) on the grounds of the similar inflores-

cence and equally 5-toothed calyx, even though these

were excluded by Briquet.

The question would then arise whether one should

go further and include other species with more-or-less

subequally 5-toothed calyx, such a P. candelabriformis

Launert and Neomuellera welwitschii Briq. in which

the inflorescence has a much simpler structure.

Because of this gradation from Burnatastrum to

Plectranthus it has been decided to give it subgeneric

rank and the following five species are dealt in this

group:

4. P. candelabriformis Launert

5. P. mirabilis (Briq.) Launert

6. P. hereroensis Engl.

7. P. spicatus E. Mey. ex Benth.

8. P. cylindraceus Hochst, ex Benth.

4. Plectranthus candelabriformis Launert in Mitt.

Bot. Miinchen 7: 300 (1968); Launert & Schreiber in

Prodr. FI. S.W.Afr. 123: 24 (1969). Type: S.W. Africa,

16 km E. of Runtu, Merxmiiller & Giess 1912 (Ml,

holo.).

Erect, perennial, aromatic herb or suffrutex up to

I m tall, branched; branches ascending, 4-angled,

striate, sparingly pubescent with longish multicellular

hairs. Leaves drying thin-textured; petiole 2,3-6 cm

long, pilose; blade ovate, 6-15 cm long, 3,5-8

(-11) cm broad sparingly pilose on both surfaces.

L. E. CODD

381

dotted below with minute orange gland-dots; apex

acute to subacuminate; base rounded to subcordate;

margin regularly crenate-dentate with 15-20 pairs of

teeth. Inflorescence terminal and on lateral branches,

paniculate with 1 or 2 main branches near the base,

30-40 cm long; rhachis hispidulous; bracts sessile or

shortly petiolate, ovate-lanceolate to narrowly

elliptical, 5-10 (-12) mm long, 5 mm broad, ciliate,

gland-dotted below, usually persisting beyond the

flowering stage, each subtending a pedunculate cyme.

Flowers in 3-flowered cymes on a slender peduncle

2 cm long; pedicels 5-10 mm long, puberulous.

Calyx 3-5 mm long at flowering enlarging to 9 mm

long in fruit, ventricose, hispidulous and freely

orange gland-dotted; teeth subequal, horizontal,

linear-lanceolate, up to 4,5 mm long, the uppermost

tooth slightly wider than the others. Corolla violet,

6, 5-7, 5 mm long, sparingly pubescent and orange

gland-dotted on the lips; tube 3-3,5 mm long,

slightly kneed near the base and 1 mm deep, expanding

slightly to the throat; upper lip erect 3,5-4 mm long

and equally broad, emarginate and with 2 lateral ear-

like lobes; lower lip boat shaped, 3,5 mm long, curved

upwards. Stamens free at the base, curved and enclosed

in the lower lip, 3 mm long. Style co-equal with the

stamens. Fig. 5.

Found on sandy soil in thickets, grassy depressions

and disturbed soil in Tanzania, Zambia and the

extreme north of South West Africa.

S.W.A. — 1719 (Runtu): 16 km E. of Runtu (-DD), Merx-

miiller & Giess 1912 (M).

Fig. 5. — Plectranthus candelabriformis, near Runtu, South

West Africa (holotype: Merxmiiller & Giess 1912 in M),

x*.

Launert indicated that the subequally 5-toothed

calyx would place this species in Burnatastrum if

this were separated at generic level, but here the

resemblance ends and there are more differences than

similarities. For example, the flowers are arranged in

3-flowered cymes, not in many-flowered scorpioid

dichasia and the bracts are small and persistent, as

against the early deciduous bracts of Burnatastrum.

It is now included in Subgen. Burnatastrum as a

matter of convenience, rather than erect another

monotypic subgenus to accommodate it. Its nearest

relationship appears to be with P. equisitiformis

(E. A. Bruce) Launert, which has its flowers in

1-3-flowered sessile cymes and in which the upper

calyx tooth is more distinctly broader than the

others.

A close examination of the base of the calyx shows

the presence of a very short stipe about 0,5 mm

long, at which point abscission of the calyx takes

place. This is reminiscent of Holostylon spp. but the

style in P. candelabriformis is shortly bilobed, not

entire as in that genus.

5. Plectranthus mirabilis ( Briq .) Launert in Mitt.

Bot. Miinchen 7: 299 (1968); Launert & Schreiber

in Prodr. FI. S.W.Afr. 123: 25 (1969); Codd in Mitt.

Bot. Miinchen 10:248 (1971). Lectotype: Angola,

Malange, Mechow 489 (Z).

Coleus mirabilis Briq. in Bot. Jahrb. 19: 183 (1894); Bak.

in FI. Trop. Afr. 5: 440 (1900); Codd in Flow. PI. Afr. 36: t.

1417 (1963). — var. mechowianus Briq., l.c. (1894). Type:

Angola, between Malanga and Cuango Rivers, Mechow s.n. —

var. poggeanus Briq., l.c. (1894). Type: Upper Congo, Lulua

River, Pogge 350. — var. hypisodontus Briq., 1. c. (1894).

Type: Angola, Malange, Mechow 489. — var. buchnerianus

Briq., l.c. (1894). Syntypes: Angola, Moma, near Malange,

Buchner 81, 82, 83, 84, 85. C. leucophvllus Bak. in Kew Bull.

1895: 292 (1895); FI. Trop. Afr. 5: 442 (1900). Type: Malawi,

Mivero, Carson 26.

Ascocarydion mirabile (Briq.) G. Tayl. in J. Bot. Lond. 69,

Suppl. 2: 162 (1931).

Erect, perennial, aromatic woody herb or suffrutex.

1-3,5 m tall; stems 1 — several from the base,

unbranched or sparingly branched, 4-angled,

appressed grey-tomentulose with dense, short multi,

cellular hairs and scattered orange gland-dots-

Leaves drying fairly thick in texture; petiole 1-2 cm

long, pubescence as for the stem; blade ovate to

ovate-lanceolate, 6-12 cm long, 3-6 cm broad, dull

green and tomentulose especially on the nerves above,

white-felted below, freely dotted with orange gland-

dots on both surfaces; apex acute; base obtuse to

cuneate, attenuate on the petiole; margin regularly

and finely crenate-dentate except in the lower third

with 25-30 pairs of teeth. Inflorescence terminal,

paniculate, dense, 10-30 cm long, 4-7 cm in diameter,

unbranched or with 1 or 2 branches near the base;

rhachis densely and shortly grey-tomentulose and

gland-dotted; bracts broadly ovate, abruptly

acuminate, 9-12 mm long, 7-8 mm broad, glandular

appressed tomentulose, early deciduous, present only

at the immature apex of the inflorescence. Flowers

arranged in opposite and decussate, pedunculate many-

flowered dichasia; peduncle 4—10 mm long, pedicels

1 mm long. Calyx purple-tinged, 4-5 mm long at

flowering, enlarging to 7-8 mm long in fruit, becoming

erect, ventricose, glandular-hispid; teeth subequal,

horizontal, lower 4 triangular-lanceolate, 3,5 mm

long, uppermost tooth slightly broader, up to 4,5 mm

long. Corolla deep blue, 1,3-1, 5 cm long, pubescent

and gland-dotted over the whole outer surface; tube at

first narrowly cylindrical, ascending for 3^1 mm

and 1 mm in diameter, then sharply decurved and

expanding for 3 mm to 4 mm deep at the throat ; upper

lip erect, 4 mm long, obscurely 4-lobed; lower lip

382

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

deeply boat-shaped, 8-9 mm long. Stamens united

at the base, curved and enclosed in the lower lip or

slightly exserted, 7-9 mm long. Style slightly exceeding

the stamens by 1-2 mm.

Found in moist peaty soil in depressions and along

river banks in Zaire, Malawi, Zambia, Angola and

northern South West Africa.

5. W.A. — 1819 (Karakuwisa): Omuramba Omatako (-DC),

Schoenfelder W49; Marsh s.n.; Giess 10088; Merxmiiller &

Giess 2155; Le Roux 183. 1821 (Andara): Popa Falls, near

Andara (-AB), Maguire 1700.

With its tall, erect stems, grey-white foliage and

deep blue flowers, this is one of the most striking

members of the genus. Taylor, l.c., separated it as a

monotypic genus on the basis of the so-called winged

seeds. However, the seeds are flattened on the margin,

not strictly winged and this differs only in a matter of

degree from other species. In floral characters it is

closely allied to P. hereroensis and would fall within

the compass of Burnatastram, if this were separated

as a distinct genus, although the calyx is not circinate

as in P. spicatus (see p. 383).

Briquet, l.c. described four varieties without

designating one of them as the typical. Type material

has not been seen but Taylor, l.c., is followed in

placing them together in synonymy and in including

C. leucophyllus Bak. also.

Chromosome number 2n =42 (De Wet, 1958, as

Coleus mirabilis).

6. Plectranthus hereroensis Engl, in Bot. Jahrb.

10: 267 (1888); Dinter in Fedde Repert. 22: 380

(1926); Taylor in J. Bot. Lond. 69, Suppl. 2: 160

(1931); Launert & Schreiber in Prodr. FI. S.W. Afr.

123: 25 (1969). Type: S.W. Africa, Hereroland, Kaiser

Wilhelmsberg near Okahandja, Marloth 1350 (B,

holo.t; G!; GRA!; K!, lecto.; Ml; PRE!; SAM!).

P. matabelensis Bak. in FI. Trop. Afr. 5:417 (1900). Syntypes:

Rhodesia, Matabeleland, Shasha River, Holub 1403-1406 (K).

P. myrianthus Briq. in Bull. Herb. Boiss. ser. 2, 3: 1001 (1903);

Cooke in FI. Cap. 5,1 : 271 (1910); Codd in Mitt. Bot. Munchen

10: 248 (1971). Type: Witwatersrand, Hutton 877 (Z, holo.;

GRA!; K!; NH!). P. otaviensis Dinter in Fedde Repert. (Beih.)

53: 116, 1 17 (1928), nomen subnudum based on Dinter 5699 from

Otavi (B!; PRE!; SAM!). P. aurifer Dinter ex Launert in Mitt.

Bot. Munchen 2: 312 (1957); Dinter in Fedde Repert. (Beih.)

53: 117 (1928), nomen subnudum. Type: S.W. Africa, Nossib,

Dinter 7367 (M!, holo.; K!).

Neomutlera damarensis S. Moore in J. Bot. Lond. 39: 265

(1901). Type: Damaraland, Eens.vx. (BM, holo.).

Coleus myrianthus (Briq.) Brenan in Mem. N.Y. Bot. Gard.

9: 43 (1954).

Erect, annual or weakly perennial, aromatic herb,

50-100 (-120) cm tall; stem usually solitary, branching

above, 4-angled, striate, finely and sparsely to densely

glandular-puberulous or tomentulose. Leaves drying

thin to medium in texture; petiole 2-7 cm long,

pubescence similar to the stem; blade ovate to ovate-

triangular, subglabrous to tomentulose, freely dotted

with reddish to brownish gland-dots below; apex

acute; base truncate; margin regularly and con-

spicuously finely or coarsely crenate-dentate with

5-20 pairs of teeth. Inflorescence terminal on the main

stem and side branches, paniculate, lax or dense,

10-25 cm long, simple or with 1 or 2 pairs of branches

near the base; rhachis glandular-puberulous to

minutely hispidulous, often gland-dotted; bracts

linear-oblanceolate to spathulate, 2-3 mm long,

persisting beyond the flowering stage. Flowers

arranged in opposite and decussate, pedunculate, lax

or dense dichasia; peduncle 8-20 mm long, pedicels

1-2 mm long, glandular-puberulous. Calyx 2,5 mm

long at flowering, enlarging to 5-7 mm long in fruit,

becoming erect, ventricose, glandular-scabrid ; teeth

subequal, horizontal, triangular-lanceolate to

subulate, 1-2 mm long, the upper 3 slightly longer

and broader than the lower 2. Corolla pale to deep

blue (rarely white), 6-11 mm long pubescent and

with scattered gland-dots; tube at first narrowly

cylindrical, ascending for 2,5-3 mm and 0,5 mm in

diameter, then sharply decurved and expanding for

2,5 mm to 3 mm deep at the throat; upper lip erect,

2 mm long, emarginate and with 2 narrow lateral

ear-like lobes; lower lip deeply boat-shaped, 4-6 mm

long. Stamens free or united at the base for 1-1,5 mm,

curved and enclosed in the lower lip, 4-5 mm long.

Style subequal to the lower stamens or exserted by

1 mm. Fig. 6.

Found in shady, south-facing kloofs, usually on

dry, rocky, wooded hillsides at medium to high

altitudes, in northern South West Africa and central

and northern Transvaal, extending to Angola, Zambia

and Rhodesia.

S.W.A. — 1712 (Posto Velho): Otjihipaberg, Kaokoveld

(-BC), Davies, Thompson & Miller 66; 17 km W. of Entanza on

road to Ombepera (-BD), De Winter & Leistner 5463. 1812

(Sanitatas): Sanitatas (-BA), Hall 430; 13 km W. of Okonjombo

(-BD), Giess & Leippert 7550. 1915 (Okankuejo): farm Cauas

Okava (-CA), Giess & Leippert 7310. 1916 (Gobaub): farmNi-

mitz(-DC), De Winter 3023. 1917 (Tsumeb) : Nosib (-BD), Dinter

7367 (K,M); Schoenfelder S64I; farm Goab-Pforte (-CA),

Giess, Volk & Bleissner 6391 ; Otavi (-CB), Dinter 5699; 5756;

Kinges 2974; Cole T7 ; Guchab (-CB), Dinter 7655. 2014

(Welwitschia): near Welwitschia, Volk sub Giess 6120. 2017

(Waterberg): Waterberg (-AC), Liebenberg 4760; Giess, Volk &

Bleissner 6618. 2114 (Uis): Brandberg (-BA), Giess 3686;

Wiss s.n. 2116 (Okahandja): near Okahandja (-DD), Marloth

1350; Dinter 564 (G, GRA, K, M, SAM).

Transvaal. — 2229 (Waterpoort): Hanglip (-DD), Meeuse

10165. 2230 (Messina): Entabeni Forest Station (-CC), Loock 2;

Codd 4193; 8399; Munro s.n. 2328 (Baltimore): Blouberg (-BB),

Leeman 112; Codd 8755; Van der Schijff 5387 ; Strey & Schlieben

8508. 2329 (Petersburg): near Louis Trichardt (-BB), Young sub

TRV 26636; Matoks (-BC), Hutchinson & Gillett 4470 (K);

30 km E. of Pietersburg (-DC), Van Vuuren 1452. 2330

(Tzaneen): Westfalia Estate (-CA), Scheepers 972. 2427 (Thaba-

zimbi): Krantzberg (-BC), Codd & Erens 3975. 2429 (Zebediela):

Pyramid Estate (-AA), Galpin 9063. 2527 (Rustenburg):

26 km N. of Brits (BD), Codd 6555; near Rustenburg (-CB),

Rose lnnes 202; Breedts Nek (-CD), Vahrmeijer 1745; Jacksons-

tuin (-DA), Repton 879; Van Vuuren 183; near Jacksonstuin

(-DA), Leistner 185; Hornsnek (-DB), Codd 4182; Scheerpoort

(-DD), Van Vuuren 499. 2528 (Pretoria): near Pretoria (-CA),

Leendertz 3735; Mogg 12435; 15144. 2627 (Potchefstroom):

Witpoortje (-BB), Moss sub TRV 19664. 2628 (Johannesburg):

near Johannesburg (-AA), Saunders 7 (K); Hutton 877 (G,

GRA, K, Z); Gilmore 488; Suiderbosrand, Wolwekloof (-CB),

Bredenkamp 815.

P. hereroensis falls naturally in subgenus Burna-

tastrum. It varies a good deal in characters such as

pubescence, size and number of leaf-margin teeth,

and size of calyx and corolla, and this variation is

reflected in the number of species names which have

been allocated. However, no clear pattern of variation

can be seen and it seems best at this stage to take a

broad view of the species and not to uphold infra-

specific groups.

An interesting feature is the variation in the degree

of union of the stamens. These may be free, or the

two lower ones united at the base, or all four united

in a sheath for 2-3 mm. This has resulted in the

species being placed in Coleus as well as in Plectranthus

and, in addition to the synonyms listed, the following

may also prove to be synonyms on closer investigation :

C. polyanthus S. Moore, C. matopensis S. Moore

and C. gazensis S. Moore.

In this respect, C. hereroensis is intermediate

between P. spicatus (stamens always free) and P.

mirabilis (stamens always united at the base).

In addition to the names listed by Dinter without

adequate description, P. otaviensis Dinter and

P. aurifier Dinter, he distributed specimens of his

No. 7655 annotated P. grandidentatus Dinter.

L. E. CODD

383

Fig. 6. — Plectranthus hereroensis, near Pretoria (Codd 8627), x ] .

P. hereroensis comes nearest to P. mirabilis but is

not likely to be confused with the latter. It is a smaller,

more bushy plant occupying shady, rocky places,

annual or weakly perennial, lacking the dense grey

tomentum of P. mirabilis. The attractive blue flowers

(a white-flowered population has also been recorded)

make it a promising garden plant and, in this con-

nection, it has been noted that plants grown from

seed collected in the Matoppas (= C . polyanthus?)

are particularly successful in cultivation. Specimens

showing this affinity have been grown in the Botanic

Gardens in Durban and Kirstenbosch and have been

recorded as a garden escape at St. Michaels-on-Sea

( Nicholson 792).

Chromosome number 2n =28 (De Wet, 1958).

7. Plectranthus spicatus E. Mey. ex Benth. in E.

Mey, Comm. 230 ( 1837); Drege, Zwei Doc. 133, 141

(1843); Benth. in DC., Prodr. 12: 60 (1848); Cooke in

FI. Cap. 5,1; 270 (1910); Compton, FI. Swaz. 67

(1966); Codd in Mitt. Bot. Miinchen 10: 248 (1971);

Ross, FI. Natal 305 (1972). Fectotype; Cape, Glen-

filling, Drege (4731b in Herb. Benth., K!, lecto.;

MO!; P!; S!).

P. subspicatus Hochst. in Flora 67 (1845). Type: Uitenhage,

Krauss 1112.

Burnatastrum spicatum (E. Mey. ex Benth.) Briq. in Pflan-

zenfam. 4, 3a: 358 (1897).

26700-2

384

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Procumbent, succulent, aromatic herb with

ascending inflorescences up to 60 cm high; stems

usually several from the base, branching, 4-angled,

up to 1 m long, finely crisped tomentulose and

gland-dotted to subglabrous. Leaves fleshy, drying

fairly thick in texture, subsessile to shortly petiolate;

petiole 0-8 mm long; blade obovate 1,5-5 cm long,

0,8-2, 5 cm broad, subglabrous to minutely hispi-

dulous, veins indistinct, dotted with reddish gland-

dots below; apex obtuse; base cuneate; margin

coarsely crenate with 3-4 pairs of irregular teeth,

mainly in the upper half. Inflorescence terminal,

elongate, dense or interrupted, 9-30 cm long, 1,5-

3 cm in diameter, simple or occasionally with a pair

of branches near the base; rhachis minutely crisped

tomentulose and red gland-dotted; bracts linear-

lanceolate to ovate-lanceolate, acuminate, 4-6 mm

long, deciduous at about the early flowering stage.

Flowers arranged in opposite and decussate, com-

pact, several- to many-flowered, subsessile dichasia;

pedicels 0,5-1, 5 mm long, puberulous and gland-

dotted. Calyx 2 mm long at flowering, enlarging to

5 mm long in fruit, slightly ventricose, circinate with

the mouth erect, scabrid and freely red gland-dotted ;

Fig. 7. Plcctranthus spicatus, Alexandria, Cape Province ( Archibald 5981), X 1.

L. E. CODD

385

teeth subequal, linear-lanceolate, the uppermost tooth

distinctly longer than the rest at flowering stage,

only slightly longer and up to 2 mm long at fruiting

stage. Corolla purple, 7-8 mm long, minutely pubes-

cent and with scattered gland-dots; tube at first nar-

rowly cylindrical, 0,5 mm in diameter and ascending

for 2,5 mm, then sharply decurved and expanding for

2,5 mm to 2 mm deep at the throat; upper lip erect,

2,5 mm long and equally broad, emarginate and

with 2 narrow ear-like lateral lobes; lower lip deeply

boat-shaped, 2,5-3 mm long. Stamens free at the

base, curved and enclosed in the lower lip, 2,5-3 mm

long. Style subequal to the lower stamens or slightly

exserted by 1 mm. Fig. 7.

Found in dry woodland, often associated with

other succulent plants, in rocky places or brackish

flats, from the Humansdorp District along the

coastal areas of the eastern Cape Province and Natal

to eastern Swaziland and the eastern Transvaal

Lowveld.

Transvaal. — 2531 (Komatipoort): 11 km S. of Pretoriuskop

-AB), Codd 6500; White River (-AC), Rogers 23127 ; Kruger

National Park, Bukwenene (-AD), Van der Schijff 3019;

Van der Schijff & Marais 3738. 2731 (Louwsburg): Pongola

Flats (-BC), Verdoorn 1728.

Swaziland. — 2631 (Mbabane): Komati Pass (-BA), Compton

31523; near Stegi (-BD), Compton 32320.

Natal. — 2632 (Bella Vista): Ndumo Game Reserve (-CD),

Pooley 626 (NH). 2731 (Louwsburg): Pongola Valley (-BC),

Hall sub NBG 575/56; Magut (-DA), Gerstner 3415 (NH).

2732 (Ubombo): 6 km N. of Ingwavuma (-AA), Codd 7036;

Lebombo Mts. (-AA), Strey 8149; Bombopoort, E. of Gollela,

Hardy 1795; False Bay Park (-CD), Ward 3665. 2831 (Nkandla):

Nkwaleni Valley (-DA), Lawn 364 (NH); Umhlatuzi Valley

(-DD), Lawn 559 (NH). 2832 (Mtubatuba): Hluhluwe Game

Reserve (-AA), Codd 9630. 3030 (Port Shepstone): Oribi Flats,

Umzimkulu Gorge (-CB), McClean 402.

Cape. — 3228 (Butterworth): Xobo Valley (-AB), Van Breda

871 ; near Butterworth (-AC), Pegler 2026; Komga, Kei River

(-CA), Flanagan 2308. 3324 (Steytlerville): Kouga River (-CC),

Fourcade 3228. 3325 (Port Elizabeth): near Enon (-BC),

Drege s.n. (P); Bethelsdorp (-CD), Zeyher 3542; near Uitenhage

(-CD), Leach 8152; Port Elizabeth — Addo road, (-DA),

Long 986. 3326 (Grahamstown): Glenfilling (-BD), Drege b

K, MO, P, S); Alexandria (-CB), Archibald 4276; 5981.

The species is based on two Drege gatherings and

the sheet labelled 4731b ex Herb Benth. (K.) is

chosen as the lectotype.

Briquet based the genus Burnatastrum on P.

spicatus and two Madagascar species, P. lanceolatus

Benth. (1832) and P. lavanduloides Bak. P. spicatus is

very similar to P. lanceolatus, the main difference

being the more lanceolate leaves of the latter. In

separating the genus. Briquet stressed the remarkable

circinate, equally 5-toothed calyx. However, there is a

clear affinity between P. spicatus and P. cylindraceus

(excluded from Burnatastrum by Briquet) which has

a less markedly curved calyx, and with P. hereroensis

and P. mirabilis in which the calyx is erect and

markedly ventricose. From these the affinity extends

to P. candelabriformis and P. welwitschii ( Neomullera

welwitschii Briq.) in which the structure of the

inflorescence begins to resemble typical Plectranthus.

It is interesting to note the change in shape of the

calyx as it matures. At the flowering stage the upper

tooth is distinctly larger than the lower four but,

as it becomes older it becomes more curved and,

at the fruiting stage, all five teeth are erect and the

uppermost tooth is only slightly longer than the

other four.

The type of P. subspicatus Hochst. (Krauss from

near Uitenhage) has not been seen but Bentham

(1848) included it under P. spicatus and this decision

appears to be correct.

8. Plectranthus cylindraceus Hochst. ex Benth. in

DC., Prodr. 12: 60 (1848); A. Rich., Tent. FI. Abyss.

2: 182 (1851); Briq. in Pflanzenfam. 4, 3a: 354 (1897);

Bak. in FI. Trop. Afr. 5: 414 (1900); Compton, FI.

Swaz. 66 (1966); Launert & Schreiber in Prodr.

FI. S.W.Afr. 123: 24 (1969); Codd in Mitt. Bot.

Miinchen 10: 248 (1971); Ross, FI. Natal 305 (1972).

Type: Ethiopia, Samen, near Gapdia, Schimper 1113

(Kl, holo. ; BM!; G!; P!).

P. marrubioides Hochst. ex Benth. in DC., Prodr. 12: 63

(1848); A. Rich., l.c. 181 (1851); Briq., l.c. 354 (1897); Bak.,

I.c. 414 (1900). Type: Ethiopia, Samen, near Jaja, Schimper

1925 (Kl, holo.; BM!; G!; P ! ). P. moschosmoides Bak., l.c.

414 (1900). Type: Angola, Huilla, Welwitsch 5489 (Kl; holo.;

BM!). P. villosus T. Cooke in Kew Bull. 1909: 378 (1909);

FI. Cap. 5,1 : 275 (1910). Type: Natal, Entumeni, Medley Wood

3955 (K!, holo.; NH). P. densi floras T. Cooke in Kew Bull.

1909: 378 (1909); FI. Cap. 5,1 : 276 (1910). Type: Natal, near the

Mooi River, Medley Wood 4475 (K!, holo.; GRA!; NH!;

SAM!). P. glomeratus R. A. Dyer in Flow. PI. S. Afr. 24: sub t.

946 (1944), nom. illegit. Type: as for P. villosus T. Cooke.

P. spiciformis R. A. Dyer in Flow. PI. S. Afr. 24: t. 946 (1944).

Type: Transvaal, Hammanskraal, Mogg sub PRE 27138

(PRE!, holo.).

Perennial, succulent, aromatic herb forming a

dense cluster from which arise annually several to

many stems 60-150 cm tall; stems woody at the base,

usually branching, obscurely 4-angled, finely tomen-

tulose and usually with translucent gland-dots.

Leaves sessile or shortly petiolate, petiole up to

5 mm long; blade broadly obovate to oblong-

obovate 2-5,5 cm long, 1,5-4 cm broad, fleshy,

shortly and softly tomentulose on both surfaces

with pale to yellow gland-dots below, veins indistinct;

apex obtuse; base cuneate: margin coarsely crenate

with 3-5 pairs or irregular teeth, mainly in the upper

half. Inflorescence terminal, short and dense or

elongate and interrupted, usually crowded towards

the apex, 8-35 cm long, 8-20 mm in diameter,

occasionally simple, usually with 1-several pairs of

branches; rhachis shortly and densely crisped tomen-

tulose; bracts ovate-lanceolate to broadly ovate or

suborbicular, 4-6 mm long, 3-5 mm broad, abruptly

acute, minutely tomentulose, persisting to the late

flowering stage. Flowers in dense, glomerate, many-

flowered, opposite and decussate, subsessile dichasia,

densely villous, making the structure difficult to

discern; pedicels 0,5-1, 5 mm long, glandular-

puberulous. Calyx 1,5-2 mm long at flowering,

enlarging to 3 mm long in fruit, slightly curved

upwards, villous and usually gland-dotted; teeth

more or less subequal, the uppermost distinctly

larger than the rest, lanceolate-deltoid, 1,5 mm long,

the lower 4 teeth deltoid-subulate 0,5-1 mm long,

the lowest pair shorter than the lateral. Corolla pale

386

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 8. — Plectanthus cylindraceus, Soutpansberg {Codd 8365), x 1.

mauve and white, whitish or yellow (P. densiflorus),

4-5 mm long; tube straight 0,5 mm deep at the base,

expanding to 1,75 mm at the throat, pubescent on the

lobes; upper lip erect, 1 mm long and equally broad,

emarginate and with 2 narrow ear-like lateral lobes;

lower lip concave, 2 mm long. Stamens free at the

base, declinate, 2,5-4 mm long, exserted. Style

coinciding with the lower pair of stamens. Fig. 8.

Growing communally in dry woodland, on ter-

mita ia under thorn trees and on brackish soil, as-

well as in crevices on exposed rocky slopes; widespread

in Africa from Ethiopia southwards to Angola and

northern South West Africa in the west and, in the

east, to northern, central and eastern Transvaal,

eastern Swaziland and Natal.

S.W.A. — 1918 (Grootfontein): 32 km W. of Grootfontein

(-CA), De Winter 2882; Spitskoppe (-CA), Dinter 7412a;

Schoenfelder S804.

Transvaal. — 2229 (Waterpoort): Wylliespoort (-DD), Codd

8365; Soutpansberg (-DD), Leach 8050. 2328 (Baltimore):

Blouberg (-BB), Codd 8700; Strey & Schlieben 8501; Van der

Schijff 5369; Kwarrihoek (-CD), Steyn 81. 2329 (Petersburg):

Soutpansberg, Lejuma (-AB), Schlieben 10637; between Louis

L. E. CODD

387

Trichardt and Petersburg (-BD), Schweickerdt & Verdoorn 447 ;

Pietersburg (-CD), Pole Evans 2663; Boyne (-DD), Gerstner

5323. 2428 (Nylstroom): Sterkrivierdam Nature Reserve (-BC),

Jacobsen 2389; near Warmbaths (-CD), Leislner 144; Naboom-

spruit (-DA), Galpin M 288; 34 km N.E. of Nylstroom (-DA),

Bruce 165. 2429 (Zebediela): Percy Fyfe Nature Reserve (-AA),

Huntley 1991; Potgietersrus (-AA), Rogers sub TRV 4818;

Maguire 2519; Ganspoort (-BB), Dyer 5409; Sekukuniland

(-DB), Barnard 420. 2430 (Pilgrim’s Rest): Mica (-BB), Hardy

908; 26 km N.W. of Lydenburg (-CD), Codd 8780. 2527

(Rustenburg): 22 km S. of Northam (-AA), Codd 8679; on road

from Thabazimibi to Beestekraal (-AD), Coetzee 1174; Jack-

sonstuin (-DA), Obermeyer 408; near Brits (-DB), Rehm s.n.

2528 (Pretoria): near Hammanskraal (-AD), Mogg s.n.;

sub PRE 27138; Pole Evans 4821 ; Repton 4200; Meintjieskop

(-CA), Schlieben 7981; Daspoort (-CA), Leendertz 154

Wonderboom Reserve (-CA), Repton 1620; Swing Bridge,

32 km N.E. of Pretoria (-CB), Repton 1266; Six Miles Spruit

(-CC), Van Niekerk sub TRV 13222. 2529 (Witbank): 27 km

N.W. of Middelburg (-CB), Mogg 17313; farm Doornkop,

N. of Middelburg (-CB), Du Plessis 1407. 2531 (Komatipoort):

Nyamazane Bantu Trust (-AC), Buitendag 51 1 ; Kruger National

Park, Klokwene(-AD), Van der Schifff'870; Bukwenene (-AD),;

Van der Schijff 3016; Van der Schijff & Marais 3739; Ship Mt.

(-AD), Codd 6037; Van der Schijff 385; 3774; 7 km W. of

Malelane Camp (-AD), Codd 5516; Komatipoort (-BD),

Dyke sub Marloth 5516; Crocodilepoort Mt. (-CA), Codd 7769;

Boulders (-CB), Van der Schijff 2533 ; Barberton (-CC), Thorn-

croft 807 (NH); Rogers sub TRV 23803; Pole Evans 4683.

Swaziland. — 2631 (Mbabane): near Stegi (-BD), Lavranos

27; Compton 30859; 31530.

Natal. — 2632 (Bella Vista): Ndumo Game Reserve (-CC)>

Pooley 521 (NH). 2732 (Ubombo): Lebombo Mts., Jozini area

(-AC), Ward 4185; Ubombo (-CA), Gerstner sub NH 22917

(NH). 2830 (Dundee): Weenen, Mooi River (-CD), Medley

Wood 4475 (GRA, K, NH, SAM); 11696; near Muden (-CD),

Edwards 2769. 2831 (Nkandla): Middledrift (-CC), Edwards

2067; Entumeni (-CD), Medley Wood 3955 (K, NH). 3030

(Port Shepstone): Gibraltar (-CB), Strey 9729.

Throughout its wide range in Africa, P. cylindraceus

is remarkably uniform and infraspecific categories

are not evident. The type, Schimper 1113, from

Ethiopia, is at a young stage but is recognizable as

the same as our plant in Southern Africa. There is

some variation in density of inflorescence, while

P. densiflorus is stated by Medley Wood, in both of

his gatherings from near Weenen (Nos. 4475 and

11696), to have yellow flowers. The only other

gathering from this area, Edwards 2769, is described

as having blue flowers, and shades of blue or mauve

are usual for this species. In no other respect does

the type of P. densiflorus differ, and varietal rank

merely for a deviation in flower colour seems hardly

justified. The flower colour cannot be discerned on

dried specimens.

The glomerate flower clusters are so densely pubes-

cent that structure is difficult to ascertain, but it is

evident that the flowers are arranged in compact

dichasia as in P. spicatus. It differs from P. spicatus

in the denser pubescence, the calyx being smaller

and less markedly circinate, and the smaller, paler

corolla. Although the calyx may be described as

subequally 5-toothed, the uppermost tooth is distinctly

longer and broader than the rest.

Chromosome number 2n=28 (De Wet, 1958, as

P. spicatus and P. villosus).

Subgen. Coleus {Lour.) Codd , stat. nov.

Coleus Lour., FI. Cochin, 372 (1790).

Coleus Sect. Aromaria Benth., Lab. 51 (1832); in

DC., Prodr. 12: 72 (1848); Briq. in Pflanzenfam.

4,3a: 359 (1897).

Type species: C. amboinicus Lour., l.c.

The typification of C. amboinicus, and thus of

Subgen. Coleus, is complicated and is discussed on

p. 388.

The genus was separated mainly on the basis of

the stamens being united at the base, not free as in

Plectranthus. This characteristic led Bentham (Lab.,

1832 and in DC., Prodr. 12, 1848) to enlarge the

genus to include three sections:

Sect. Aromaria : containing the single species C.

amboinicus Lour, (which he renamed C. aroma-

ticus Benth.).

Sect. Calceolus: containing species such as C.

barbatus (Andr.) Benth. and C. spicatus Benth.

This section is now treated as Subgen. Calceolan-

thus (see p. 389).

Sect. Solenostemon: containing species such as

C. scutellarioides (L.) Benth. and C. latifolius

Hochst. ex Benth. Following Morton (1962,

1963), this section is now included in an emended

concept of Solenostemon Schumach., as a separate

genus (see p. <37).

While Sect. Aromaria has remained monotypic,

many species names have been added to Sections

Calceolus and Solenostemon, so that a completely dis-

torted view of Coleus came to be accepted.

As pointed out by Morton (1962) and Launert

(1968) the degree of union of the stamens in “ Coleus ”

species varies and the only character, apart from the

union of the stamens, that can be used to separate

C. amboinicus from Plectranthus is the oblong upper

tooth of the calyx. Thus there seems little justification

for upholding Coleus as a separate genus and it is

now included in synonymy but given subgeneric

status. The remaining “ Coleus ” sections will then be

distributed mainly to the genera Plectranthus and

Solenostemon respectively.

It is probable that P. amboinicus originated in

Africa and was dispersed by the early Portuguese

voyagers. Specimens such as Welwitsch 5556 from

Angola are a very good match of the plant from

India described as C. aromaticus Benth., and a good

deal of variation occurs in this area. A second species,

P. unguentarius is now described from South West

Africa, making two species in the Section, both of

which are recorded from Southern Africa, namely:

9. P. unguentarius Codd

10. P. amboinicus (Lour.) Spreng.

9. Plectranthus unguentarius Codd, sp. nov.,

P. amboinico (Lour.) Spreng. affinis, sed habitu erecto

robustiore, pedicellis longioribus differt.

P. amboinicus sensu Launert & Schreiber in Prodr. FI.

S.W.Afr. 123: 24 (1969).

Suffrutex perennis. Caulis erectus, parce ramosus,

1-1,5 m altus, obscure quadrangularis, cano-tomen-

tosus. Folia subcarnosa, breviter petiolata; petiolus

1-1,5 cm longus; lamina late obovata vel subro-

tundata, 4-6 cm longa, 4-6 cm lata, supra dense

strigosa, subtus dense cano-tomentosa glanduloso-

punctata, valde reticulata, apice rotundato, basi

cuneata vel attenuata, margine profundo crenato.

Inflorescentia subspicata, robusta, erecta usque ad

35 cm longa; rhachis dense lanata; bracteae late

ovatae acutae, 3 mm longae, subpersistentes.

Verticillastri 1-3 cm distantes, dense glomerati,

toinentosi, ca 40-floribus; pedicelli 4-5 mm longi,

dense tomentosi. Calyx demum 5-6 mm longus,

glanduloso-tomentosus; lobus posticus ovatus, acutus,

horizontalis, 2,5 mm longus, 2 mm latus; 4 dentes

antici breviter subulati. Corolla alba, 1-1,2 cm

longa, glanduloso-pubescens; tubus leviter genicu-

latus, prope basin 1 mm diam., ad fauce 2 mm

diam.; labium posticum erectum, 1,5-2 mm longum,

obscure 4-lobatum; labium anticum cymbiforme,

5-7 mm longum. Stamina 4, filamentis ad basin

breviter connatis, superna 5 mm longa, inferna

7 mm longa. Stylus in labio antico inclusus vel

breviter exsertus.

388

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Type. — South West Africa, Kaokoveld, 17 km S.

of Kaoko Otavi on road to Sesfontein, 21/4/57,

De Winter & Leistner 5595 (PRE, holo.).

Perennial erect robust semi-succulent suffrutex

1-1,5 m tall; stems woody at the base, obscurely

4-angled, sparingly branched, densely cano-tomentose

and with minute gland-dots. Leaves semi-fleshy,

slightly aromatic, drying fairly thick-textured ; petiole

1-1,5 cm long, pubescence as for the stems; blade

broadly obovate to subrotund, 4-6 cm long and

equally broad, densely appressed pubescent and veins

indistinct above, woolly-tomentose and distinctly

reticulate veined below, freely dotted with minute

reddish-brown gland-dots; apex rounded; base

cuneate to abruptly attenuate on to the petiole;

margin coarsely crenate except in the lower third with

8-12 pairs of rounded teeth. Inflorescence a terminal

spike-like raceme of interrupted densely glomerate

verticillasters, verticillasters 1-3 cm apart, simple or

with a pair of branches near the base up to 35 cm

long; rhachis densely woolly-tomentose; bracts

broadly ovate, 3 mm long, 2,5 mm broad, acute,

densely tomentose, persisting to early flowering stage.

Flowers in very dense ±20-flowered sessile cymes,

densely tomentose, producing ±40-0 owe red verti-

cillasters; pedicels erect to ascending 4-5 mm long,

densely tomentose. Calyx 2,5-3 mm long at flowering,

enlarging to 5-6 mm in fruit, glandular-tomentose;

tube short, campanulate; uppermost tooth by far the

largest, horizontal, broadly ovate, 2,5 mm long and

2 mm broad, abruptly acute at the apex, the lower

4 teeth short, subulate. Corolla white, 1-1,2 cm

long, glandular-pubescent; tube slightly bent about

the middle, 1 mm deep at the base, expanding to

2 mm at the throat; upper lip short, erect, 1 ,5-2 mm

long and equally broad and somewhat hooded,

emarginate and with 2 obscure lateral lobes; lower lip

boat-shaped, 5-7 mm long, horizontal. Stamens

united at the base for 1-2 mm, curved upwards in

the lower lip, 5-7 mm long, enclosed or slightly

exserted. Style exserted by about 1 mm. Fig. 9.

Recorded from the Kaokoveld, northern South

West Africa, in dry Mopane woodland, on high rocky

situations up to 2 300 m altitude, with broken frag-

ments of limestone and weathered dolomite.

S.W.A. — 1712 (Posto Velho): Kaokoveld, Baynes Mts.,

Otjihipaberg, Davies, Thompson & Miller 88. 1813 (Ohopoho):

Kaokoveld, 17 km S. of Kaoko Otavi on road to Sesfontein,

De Winter & Leistner 5595.

In the herbarium, P. unguent ctr ins can be distin-

guished from P. amboinicus by several small

differences: the stouter stems and inflorescences, the

reddish-brown gland-dots on the underside of the

leaves, the more coarsely crenate leaf margin, the

more numerous flowers in each verticil (40 or more,

as against about 20 in P. amboinicus) and the longer

pedicels. In the field the two would not be confused

because of the difference in habit, P. unguent arius

having robust, erect, semi-woody stems, sparingly

branched, as against the creeping, succulent, much-

branched stems of P. amboinicus. In P. amboinicus

the leaves are fleshy and highly aromatic, being used

medicinally or to flavour food; in P. unguentarius the

leaves are not fleshy and are described as slightly

aromatic. On the other hand, the roots are aromatic

and are used as one of three ingredients of a pomade

used by ladies of the Kaokoveld. Leaves of Fagara

sp. are said to be one of the other ingredients, but the

third ingredient is unknown.

10. Plectranthus amboinicus (Lour.) Spreng., Syst.

Veg. 2: 690 (1825), as “Amboinensis”; Launert in

Mitt. Bot. Miinchen 7: 298 (1968); Ross, FI. Natal

Fig. 9. — Plectranthus unguentarius, Kaokoveld, South West

Africa (De Winter & Leistner 5595), x

305 (1972). Type: a specimen in BM which is prac-

tically unrecognizable; representative specimen chosen

by Launert, l.c. : Siam, A. F. G. Kerr s.n. (BM).

Coleus amboinicus Lour., FI. Cochin. 372 (1790); Briq. in

Pflanzenfam. 4, 3a: 359 (1897); Merrill in Addisonia 20: 11

(1937); Compton, FI. Swaz. 67 (1966); Codd in Mitt. Bot.

Miinchen 10: 248 (1971). — var. violaceus Guerke in Bot. Jahrb.

19: 210 (1894); Bak. in FI. Trop. Afr. 5: 434 (1900); Hiern, Cat.

Afr. PI. Welw. 1: 865 (1900). Type: Tanzania, Lake Chala,

Volkens 321 (K ! ; BM!). C. aromaticus Benth. in Wall., PI. As.

Rar. 2: 15 (1831); Lindl. in Bot. Reg. 18: t. 1520(1832); Benth.,

Lab. 51 (1832); in DC., Prodr. 12: 72 (1848); Hook. f„ FI. Brit.

India 4: 625 (1885); Trimen, Handb. FI. Ceylon 3: 374 (1895);

Hiern, Cat. Afr. PI. Welw. 1: 865 (1900). Type: India, Patna,

Buchanan- Hamilton (in Herb. Wallich, K !). C. crassifolius

Benth. in Wall., PI. As. Rar. 2:15 (1831); Lab. 52 (1832). Type:

India, Wight (in Herb. Wallich, K !).

Plectranthus aromaticus Roxb., Hort. Beng. 45 (1814), nom,

nud. P. aromaticus (Benth.) Roxb., FI. Ind. ed. 2,3: 22 (1832).

Majana amboinica (Lour.) Kuntze, Rev. Gen. 2: 524 (1891),

nom. invalid.

Perennial, succulent, strongly and pleasantly

aromatic herb with many spreading stems up to 1 ,5 m

long, ascending at the inflorescences to 50 cm: stems

somewhat woody at the base, obscurely 4-angled,

freely branched, fairly to densely villous. Leaves fleshy,

strongly aromatic, drying thick-textured; petiole 4-10

mm long, usually more villous than the stems; blade

broadly ovate to ovate-deltoid, densely appressed

pubescent above, woolly-tomentose and with distinct

nerves below, both surfaces dotted with pale to

brownish gland-dots; apex obtuse to rounded; base

truncate, sometimes abruptly attenuate; margin finely

crenate with 7-10 pairs of teeth. Inflorescence ter-

minal, slender, elongate, 10-30 cm long, of inter-

rupted, densely glomerate verticillasters, simple or

L. E. CODD

389

occasionally with a pair of branches near the base;

verticillasters 1-3 cm apart; rhachis shortly and

densely crisped tomentulose and gland-dotted; bracts

broadly ovate to subrotund, shortly acute, 2-3 mm

long and equally broad, persisting to the flowering

stage. Flowers in dense, 4-10-flowered sessile cymes,

densely villous making the structure difficult to discern;

pedicels erect to ascending, 1 ,5-2,5 mm long, densely

glandular-villous. Calyx 2,5-3 mm long at flowering,

enlarging to 5-6 mm long in fruit, glandular-villous,

tube short, campanulate; uppermost tooth by far the

largest, horizontal, oblong to broadly oblong, up to 3

mm long and 2 mm broad, rounded to abruptly apicu-

late at the apex, the lower 4 teeth subulate, 1 ,5-2 mm

long. Corolla lilac, mauve or whitish, 7-9 mm long;

tube slightly bent about the middle, 1 mm deep at the

base, expanding slightly to 1,5 mm at the throat,

pubescent; upper lip erect, 1 ,5-2 mm long and equally

broad, emarginate and with 2 lateral ear-like lobes;

lower lip boat-shaped, 4 mm long, horizontal.

Stamens united at the base for 1-2 mm, curving

upwards in the lower lip, enclosed or slightly exserted,

4-5 mm long. Style coinciding with the lower pair of

stamens. Fig. 10.

Fig. 10. — Plectranthus ambainicus, Zululand ( Ward 3182),

X}.

Distributed from Kenya to Angola in the west and,

in the east, to Mozambique, Swaziland and northern

Natal; in Southern Africa found on rocky, wooded

slopes, loamy flats with succulent plants or the margin

of sand forest, near the sea or up to 150 m altitude

(in Southern Africa).

Swaziland. — 2731 (Louwsburg): Ingwavuma Poort (-BB),

Compton 28621 ; 29081.

Natal. — 2632 (Bela Vista): Ndumo Game Reserve (-CD),

Ward 3182; Tinlev 459. 2732 (Ubombo): Mkuzi Game Reserve

(-CA), Ward 3983.

The typification of the species presents a problem

because Loureiro cites several discordant elements in

his protologue, but his description is detailed and

there has not been any doubt as to the plant described.

Launert, l.c., mentions the presence in BM of a

Loureiro specimen which may be regarded as the

type, but it is in such a poor state or preservation

that it would be scarcely recognizable but for a few

floral fragments. He therefore clarifies the concept by

designating a specimen collected in Siam by A. F. G.

Kerr as a representative specimen.

P. amboinicus has become widely distributed by

cultivation, first in the Far East and, later, in the

West Indies and tropical America. Its strong, not

unpleasant aroma has led to its use medicinally (see

Watt, Indian Medicinal Plants 2: 1017, 1918) and

as a popular spice for flavouring food, being known as

Soup Mint, French Thyme, Spanish Thyme, Country

Borage (Launert) and Indian Mint (Hiern). According

to Trimen, l.c., it is employed in Ceylon as a medicine,

especially for cattle, and a plant is always to be found

growing in a little box suspended on the side of the

native carts.

There has been speculation as to its country of

origin but it undoubtedly occurs naturally in Africa

from Kenya southwards to Angola and Natal. It is

common near Lourengo Marques (e.g. Inhaca Island)

where it would have been accessible to the early

voyagers to the East.

Subgen. Calceolanthus Codci, subgen. nov., subgen.

Plectrantho affinis sed fauce calycis villosa, filamentis

basi connatis differt.

Coleus sect. Calceolus Benth., Lab. 49 (1832); in

DC., Prodr. 12: 71 (1848); Briq. in Pflanzenfam. 4,

3a: 359 (1897).

Type species: P. caninus Roth.

In southern Africa one could regard subgen.

Calceolanthus as a clear-cut genus with no inter-

mediate species linking it with Plectranthus. However,

in tropical Africa there are several species of very

similar facies in which the stamens are united at the

base, but in which the calyx is not villous in the throat.

It therefore seems best to include the group in

Plectranthus at the present stage in our knowledge.

The section Calceolus was included in Coleus by

Bentham because of the filaments being united at

the base. Although this character breaks down in

certain groups, e.g. in subgen. Burnatastrum and the

genus Solenostemon , in subgen. Calceolanthus the

union of the filaments is a constant character.

Three species belonging to the subgenus occur

naturally in our region, namely:

11. P. tetensis (Bak.) Agnew ( =Coleus vagatus

E. A. Bruce)

12. P. caninus Roth

13. P. neochilus Schltr.

In addition, two tropical species are widely

cultivated in South Africa and have become semi-

naturalised. They are, therefore, included in the

present treatment, namely:

14. P. ornatus Codd (= Coleus comosus Hochst. ex

Guerke)

15. P. barbatus Andr.

390

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

11. Plectranthus tetensis (Bak.) Agnew, Upland

Kenya Wild Flow. 635 (1974).

Coleus tetensis Bak. in FI. Trop. Afr. 5: 431 (1900). Type:

Mozambique, near Tete, Kirk s.n. (K!, holo.). C. decumbens

Guerke in Bot. Jahrb. 19: 211 (1894); Bak., l.c. 431 (1900);

Compton, FI. Swaz. 67 (1966). Syntypes: Kenya, Duruma

district, Hildebrandt 230; Tanzania, Kilimanjaro, Volkens 327

(BR!), non Plectranthus decumbens Hook. f. (1864). C. vagatus

E. A. Bruce in Bothalia 6: 227 (1951); Codd in Mitt. Bot.

Munchen 10: 248 (1971). Type: Transvaal, 2,5 km E.

of Skukuza, Codd 5489 (PRE!, holo. ; K !).

Plectranthus vagatus ( E. A. Bruce) Codd in Ross, FI. Natal 305

(1971), non rite publ.

Perennial, trailing, semi-succulent, not unpleasantly

aromatic herb, branching sparingly at the base;

stems 4-angled, procumbent, up to 70 cm long with

ascending inflorescences up to 20 cm, shortly and

fairly densely tomentulose with scattered multicellular

hairs and usually dotted with reddish glands. Leaves

softly succulent; petiole 5-10 mm long, pubescent as

for the stems; blade ovate to obovate, 1,5-2, 5 cm

long, 1,2-2 cm broad, shortly appressed pubescent,

usually with orange-red gland-dots below; apex

obtuse to rounded, base obtuse to cuneate; margin

obscurely crenate-dentate with 4-5 pairs of teeth.

Inflorescence a terminal spikelike raceme, dense and

enclosed in the bracts but not markedly 4-angled at

the apex, often laxer and interrupted below; rhachis

fairly densely tomentulose with tufts of long multi-

cellular hairs below the verticillasters; bracts oblong-

ovate, rounded at the apex, 6-9 mm long, 3-6 mm

broad, villous on the margin, persisting to about the

flowering stage. Flowers in sessile 4-6-flowered cymes,

forming 8-12-flowered verticillasters; verticillasters

densely placed above, up to 5 mm apart below;

pedicels erect, appressed to the rhachis, 3 mm long,

villous. Calyx 1 ,5 mm long at flowering enlarging to

4-5 mm long in fruit, pilose and red gland-dotted

without, densely villous inside; upper lip large,

broadly ovate, abruptly acute, 2,5-3 mm long and

3-4 mm broad; lower 4 teeth subequal, deltoid-

subulate, 2 mm long. Corolla mauve-purple, 1,5-

1,8 cm long, sparingly pubescent and with scattered

red gland-dots; tube at first ascending and narrowly

cylindrical, 0,5 mm in diameter, then geniculate and

expanding about the middle to 3,5 mm deep at the

throat; upper lip erect 3,5-4 mm long, 3 mm broad,

emarginate at the apex and with 2 lateral ear-like

lobes; lower lip deeply boat-shaped, ascending,

9-11 mm long. Stamens united at the base for 2 mm

and curved within the lower lip, of two lengths,

enclosed or shortly exserted, 7-9 mm long. Style

shortly exserted. Fig. 11.

Distributed from Tanzania through Mozambique

and Rhodesia to the northern and eastern Transvaal

Lowveld and northern Zululand, usually associated

with dry thorn-scrub and woodland on brackish flats.

Transvaal. — 2230 (Messina): Messina (-AC), Rogers 2070S.

2231 (Pafuri): Klopperfontein (-CA), Obermeyer sub TRY

28402. 2330 (Tzaneen): Great Letaba (-CC), Breijer sub TRY

17613. 2331 (Phalaborwa): Tendi River (-CB), Yan der Schijff

3536. 2431 (Acornhoek): 2,5 km E. of Skukuza (-DC), Codd

5489. 2531 (Komatipoort): Kruger National Park, Sigaas River

(-AD), Brynard & Pienaar 4431 ; Bukwenene (-AD), Yan der

Schijff 3020; Yan der Schijff & Marais 3736; 8 km N.E. of

Malelane (-BC), Codd 6083; Komatipoort (-BD), Dyke sub

Marloth 5516; Kaapmuiden (-CB), Mogg s.n.; Malelane (-CB),

Leach 8361 ; Barberton (-CC), Buitendag 853; Lomati River

(-DA), Jenkins sub TRY 9909.

Swaziland. — 2631 (Mbabane): on road to Komatipoort

(-BA), Pole Evans 3461.

Natal. — 2632 (Bela Vista): Ndumo Game Reserve (-CD),

Tinley 440.

P. tetensis is sometimes confused with P. neochilus

in the herbarium but it can be readily distinguished

from the latter by the slender, trailing, less villous

stems and the rounded, flattish bracts (distinctly

Fig. 1 1 . — Plectranthus tetensis, Kruger National Park ( Van

der Schijff & Marais 3736), X 1 .

pointed in P. neochilus , P. caninus and P. ornatus)

which persist until the flowering stage, and the tufts

of villous hairs on the rhachis and bracts.

The name Coleus decumbens Guerke cannot be

transferred to Plectranthus because of the earlier

P. decumbens Hook. f. (1864).

Chromosome number 2n =32 (De Wet, 1958, as

“ Coleus vagatus”).

12. Plectranthus caninus Roth, Nov. PI. Spec.

279 (1821); Launert & Schreiber in Prodr. FI.

S.W.Afr. 123; 24 (1969). Type; India, Heyne s.n.

(Herb. Wall., K!, iso.).

Coleus spicatus Benth. in Wall., PI. As. Rar. 2: 15 (1831);

Lab. 49 (1832); in DC., Prodr. 12: 71 (1848); Wight, Ic. t.

1431 (1849); Hook, f., FI. Brit. India 4: 624 (1885). Type: India,

Wight s.n. (Herb. Benth., K!, holo.). C. caninus (Roth) Vatke in

Linnaea 37: 318 (1871), partly, excl. Schimper 622; Guerke in

Bot. Jahrb. 19: 212 (1894); Briq. in Pflanzenfam. 4, 3a: 359

(1897); Codd in Bothalia 7: 433 (1961). C. flavovirens Guerke

in Engl., Pflanzenw. Ost.-Afr. C 347 (1895). Type: East Africa,

Volkens 1771 (BR!, iso.). C. omahekense Dinter in Feddc

Repert. (Beih.) 53: 123 (1928). Syntypes: S.W. Africa, Groot-

fontein, Etemba, Dinter 3265; Otjikuara, Dinter 3265.

Annual, erect, branching, semi-succulent, some-

what unpleasantly aromatic herb, 15-40 cm tall;

stems 4-angled, villous with short and long hairs and

dotted with orange-red gland-dots. Leaves slightly

fleshy; petiole 4-20 mm long, shortly tonientose;

L. E. CODD

391

blade oblanceolate to obovate-oblanceolate or long-

elliptic or ovate-lanceolate 3-5,5 cm long, 2, 5-3, 5

cm broad, sparingly hispidulous on both surfaces,

especially on the nerves below, gland-dotted and

often with short glandular hairs; apex acute to obtuse;

base cuneate; margin subentire to obscurely crenate-

dentate with 4-7 pairs of shallow teeth. Inflorescence

a terminal dense spikelike raceme 2,5-6 (-9) cm

long, simple or occasionally with a pair of branches

near the base, 4-angled and enclosed in large bracts

in the bud stage; rhachis hidden by the flowers;

bracts densely imbricate, broadly ovate, concave,

apiculate, 4-6 mm long, 2-3 mm broad, whitish-green,

purple-tipped, villous, early deciduous. Flowers in

sessile 3-4-flowered cymes forming 6-8-flowered

expanding to 1 mm deep about the middle; upper lip

erect to recurved, 1-1,5 mm long, obscurely 4-lobed;

lower lip boat-shaped, horizontal, 5-6 mm long.

Stamens united at the base for 1,5 mm and curved

within the lower lip or slightly exserted, 5-6 mm

long. Style coinciding with the stamens. Fig. 12, 13.

Tends to grow communally under trees in dry, open

woodland or on rocky outcrops; found in India and

from Ethiopia southwards through tropical East Africa

to Rhodesia, Zambia and northern South West

Africa.

S.W.A. — 1817 (Tsintsabis): Farm Kakuse, east of Etosha

Pan (-DA), Sachs s.n. 1917 (Tsumeb): Farm Sissekab (-AC),

Giess, Volk & Bleissner 6418 ; 19 km from Otavi on Grootfontein

road (-CB), Hardy 2114. 2218 (Gobabis): near Gobabis (-BD),

Tolken 5420.

Fig. 12. — Plectranthus cani-

nus, Gobabis, South West

Africa ( Totken 5420).

verticillasters; verticillasters closely appressed;

pedicels erect, 2-2,5 mm long, puberulous. Calyx

2 mm long at flowering enlarging to 5 mm long in

fruit, glandular-pubescent without, densely villous

inside; upper lip large, broader than long, 2 mm long,

4 mm broad, abruptly apiculate, decurrent on the

tube; lower 4 teeth subequal, linear-subulate, 2 mm

long. Corolla blue-purple, 8-10 mm long, slightly

puberulous on the lips, tube slightly geniculate and

P. caninus is a plant of semi-desert areas and

varies a good deal in stature, branching and leafiness

according to the growing conditions; otherwise, the

floral characters are fairly constant (see also notes on

p. 393). In South West Africa it is an annual but

whether this applies throughout its entire range is

not known. It is related to P. neochilus but can

usually be distinguished by the shorter, compact;

(not interrupted) inflorescences and shorter corolla

392

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 13. — Plectranthus caninus (Tolken 5420), xl.

(8-10 mm long as against 12-18 mm long in P.

neochilus). P. neochilus is a perennial but may behave

as an annual under periodic very dry conditions in

South West Africa.

Briquet, l.c., includes C. heynei Benth. as a synonym

of C. caninus (Roth.) Vatke and it is possible that

P. monadelphus Roxb. may also be conspecific, but

it has not been possible to investigate the typification

of these species in the present study.

It is doubtful if the combination Coleus caninus

can be attributed to Vatke as he does not cite the

basionym. The plant which he was discussing,

Schimper 622, is not P. caninus but P. ornatus

( — Coleus comosus) (see p. 393).

13. Plectranthus neochilus Schltr. in J. Bot. Lond-

34: 394 (1896); Cooke in FI. Cap. 5,1: 285 (1910);

Launert & Schreiber in Prodr. FI. S.W.Afr. 123: 25

(1969); Ross. FI. Natal 305 (1972). Type: Transvaal,

Barberton, Rimers Creek, Galpin 968 (K!, holo.;

GRA!; NH!).

Coleus schinzii Guerke in Bull. Herb. Boiss. 6: 555 (1898);

Bak. in FI. Trop. Afr. 5: 430 (1900). Type: S.W. Africa, Ovam-

boland, Tsumeb, Schinz 56 (Z!, holo.; PRE, photo.). C. pentheri

Guerke in Ann. Naturhist. Hofmus. Wien 20: 48 (1905);

Cooke in FI. Cap. 5,1: 289 (1910); Bruce in Hook. Ic. PI. 34:

t. 3375 (1938); Type: Cape Province, Albany District, Breakfast

Vlei, Krook in Hb. Pcnther 1716 (W!; holo., PRE!). C. neochilus

(Schltr.) Codd in Bothalia 7: 432 (1961); Letty, Wild Flow.

Transv. 288, t. 143,2 (1962). C. carnosus Dinter, ined.; Dinterex

Eliovson, S. Afr. Flow, for the Gard. 165 (1955), illustr. only.

Perennial, or sometimes annual, decumbent, to

erect, often much branched and bushy, succulent,

unpleasantly aromatic herb 12-50 cm tall; stems

solitary to many often first forming a loose or dense

mat, ascending, obscurely 4-angled, branched, finely

tomentulose to sparingly or fairly densely villous,

often with a mixture of long and short hairs, with

scattered orange gland-dots. Leaves succulent, viscid,

tending to fold along the midrib; petiole 5-15 mm

long, pubescent like the stems; blade obovate to

elliptic-ovate, 2-5 cm long, 1,5-3, 5 cm broad,

sparingly to fairly densely appressed pubescent

especially on the nerves below, with orange gland-dots

below; apex obtuse to rounded; base cuneate to

attenuate; margin obscurely crenate with 4-6 pairs of

teeth mainly in the upper half. Inflorescence a ter-

minal spikelike raceme 7-15 cm long, 4-angled and

with a coma of large imbricate bracts in the young

stage, dense at the apex, laxer and interrupted below;

rhachis fairly densely glandular-tomentulose; bracts

imbricate, ovate, acuminate, concave, 6-10 mm long,

4-6 mm broad, greenish-white tipped with purple,

tomentulose and gland-dotted, early deciduous.

Flowers in sessile 3-flowered cymes forming 6-flowered

verticillasters; verticillasters densely placed above

becoming laxer below and spaced 5-15 mm apart;

pedicels erect, appressed to the rhachis, 3-4 mm

long, glandular-puberulous. Calyx 3 mm long at

flowering enlarging to 6 mm long in fruit, glandular-

scabrid and gland-dotted, densely villous in the

throat; upper lip large, broader than long, 2,5 mm

long, 5-6 mm broad with a short median apicule;

lower 4 teeth subequal, lanceolate-subulate, 2 mm

long. Corolla mauve purple, rarely whitish, the upper

lip paler and bluish, 1,2-2 cm long, slightly pube-

rulous on the lips and with scattered orange gland-

dots; tube ca 0,75 mm in diameter at the base, at

first horizontal, slightly geniculate about the middle

and expanding to 1 ,5 mm deep at the throat; upper

lip erect to recurved, 2 mm long and equally broad,

emarginate and with 2 obscure lateral lobes; lower lip

boat-shaped, horizontal, 8-1 1 mm long. Stamens

united at the base for 2-3 mm and curved within the

lower lip, 8-1 1 mm long, slightly exserted. Style

finally exserted by 4-5 mm. Fig. 14, 15.

Fig. 14. — Plectranthus neochilus, Barberlon ( Codd 9531).

L. E. CODD

393

Fig. 15. — Plectranthus neochilus, Lebombo Mts. (Codcl 7798),

Xl.

Found under trees in dry, open woodland and

among rocks (especially dolomite) in grassland, in the

eastern Cape Province, central and northern Natal,

central and eastern Transvaal, and northern South

West Africa, extending to Botswana, Zambia and

Rhodesia.

S.W.A. — 1917 (Tsumeb): Tsumeb (-BA), Schinz 56 (Z);

Nosib (-BD), Schoenfelder 1021 ; 5643; 6 km S.W. of Otavi

(-CB), De Winter 2853; Auros ( DA), Dinter 5634. 2016

(Otjiwarongo): farm Cleveland, 6 km N. of Otjiwarongo

(-BC), Giess, Volk & Bleissner 6334.

Transvaal. — 2329 (Petersburg): 43 km from Pietersburg on

road to Dendron, Coetzee 1280. 2428 (Nylstroom): Towoomba

Pasture Research Station (-CD), Sidey 1401; Mosdene, near

Naboomspruit (-DA), Galpin M 287; Schlieben 9176. 2429

(Zebediela): Potgietersrus (-AA), Rogers sub TRY 4817;

Percy Fyfe Nature Reserve (-AA), Huntley 1110. 2430 (Pilgrim’s

Rest): Lulu Mts., Sekukuniland (-CC), Barnard & Mogg 704.

2526 (Zeerust): 24 km S.W. of Zeerust (-CA), Leistner 549.

2527 (Rustenburg): near Silikaatsnek (-DB), Pole Evans 4761 .

2528 (Pretoria): Pienaars River (-AD), Pedro 696; Werdermann

& Oberdieck 1610; Rooikop (-BA), Smuts & Gillett 2008;

21 km S.E. of Pretoria (-CD), Codd 2570; Rietvlei Research

Station (-CD), Acocks 11262. 2529 (Witbank): Doornkop, near

Middelburg (-CB), Du Plessis 1137; 1403. 2530 (Lydenburg):

near Lydenburg (-AB), Atherstone s.n. 2531 (Komatipoort) :

Louws Creek (-CB), Van Dam sub TRY 21 150; near Barberton

(-CC), Galpin 968 (K, GRA, NH); Thorncroft 109 (K>; 13 km

S.E. of Barberton (-CC), Codd 9531.

Natal. — 2830 (Dundee): Mooi River (-CD), Medley Wood

4340 (K); near Muden (-CD), Codd 8602; Tugela Valley (-CD),

Sidey 3670. 2930 (Pietermaritzburg): Byrne (-CC), Medlev

Wood 3199 (K).

Cape. — 3129 (Port St. Johns): Lusikisiki (-BC), Flanagan

2886. 3227 (Stutterheim): near Dohne (-CB), Acocks 9547;

near Komga (-DB), Flanagan 557. 3228 (Butterworth): Cintza

River Mouth (-CC). Galpin 6554. 3326 (Grahamstown):

Breakfast Vlei (-BB), Krook sub Penther 1716 (W).

P. neochilus varies a good deal in vegetative

characters. In the eastern Cape Province the stems are

semi-prostrate to decumbent, about 10-20 cm long,

with numerous glands and short appressed pubescence

on leaves and stems, and with scattered multicellular

hairs on the stems. In the Transvaal, South West

Africa and Rhodesia, the stems are usually more

ascending, 30-50 cm tall, and usually more con-

spicuously villous. It appears to behave sometimes

as an annual in South West Africa; elsewhere the

plants are perennial, often with somewhat tuberous

roots, especially those found in grassy places. A low-

growing form, up to about 15 cm tall, less pubescent

and which flowers a few weeks later than the tall

form, is known in cultivation and may be worthy of

varietal rank. Its origin is unknown, but it makes a

dense, mat-like growth and is useful as a ground

cover. It is said to have been brought to the Voor-

trekker Monument, Pretoria, in response to a general

request for succulent plants with which to plant the

surrounding garden.

The floral characters are relatively constant. In

the bud stage the inflorescence is a 4-angled spikelike

raceme 3-4 cm long, composed of 4 rows of densely

imbricate, ovate, acuminate bracts. The bracts are

shed as each verticil of flowers starts to open. As

flowering proceeds, the rhachis elongates with the

result that the verticils become separated by inter-

vals of 5-15 mm, producing an interrupted spikelike

raceme of up to 15 cm long. Depending on the

length of the inflorescence, 5-12 spaced verticils may

be seen below the coma of bracts, the uppermost

3 or 4 still flowering and the lower ones in fruit, with

the rhachis easily visible between the verticils. The

corolla varies in length from 1,2 to 2 cm.

Chromosome number 2n=32 (De Wet, 1958, as

“ Coleus pentheri").

The three species P. caninus, P. neochilus and P.

ornatus ( =Co!eus comosus) form a closely related

group and were discussed in Bothalia 7: 432 (1961).

The small amount of tropical material received since

then supports the conclusion reached that the three

may be regarded as distinct on the basis mainly of the

floral characters, as outlined below.

P. caninus : plants annual; inflorescence 2-5, rarely

up to 9 cm long, dense throughout; corolla 0,8-1 cm

long. Distribution-. S.W. Africa, through tropical

Africa to Ethiopia and in India.

P. neochilus: plants perennial (occasionally

behaving as annual in S.W. Africa); inflorescence

7-15 cm long, lax below with 5-12 spaced fruiting

verticils below the flowers; corolla 1,2-2 cm long.

Distribution: eastern Cape Province, Natal, Trans-

vaal, S.W. Africa, Botswana, Zambia and Rhodesia.

P. ornatus: plants perennial; inflorescence 3-5,

rarely up to 9 cm long with one or two, rarely more,

spaced fruiting verticils below the flowers; corolla

2-2,5 cm long. Distribution: chiefly Kenya and

Ethiopia.

14. Plectranthus ornatus Codd, nom. nov. Type:

Ethiopia, Schimper II. 1328 ( P !, holo.).

Coleus comosus Hochst. ex Guerke in Bot. Jahrb. 19: 212

(1894); Bak. in FI. Trop. Afr. 5: 426 (1900); Bruce in Hook.

Ic. PI. 34: t. 3374 (1938); non Plectranthus comosus Sims in

Bot. Mag. t. 2318 (1822). C. spicatus sensu A. Rich., Tent, FI

Abyss. 2: 183 (1851), as to syn. and spec, cited. C. caninus sensu

Vatke in Linnaea 37: 318 (1971); sensu Engl., Hochgebirgsfl.

Trop. Afr. 359 (1892).

Perennial, decumbent to trailing, unpleasantly

aromatic succulent herb, branching freely at the base,

up to 30 cm tall; branches obscurely 4-angled,

glandular-tomentulose. Leaves succulent, drying thick

textured; petiole 2-10 mm long; blade obovate to

broadly obovate, 2-3 cm long, 1,5-2, 5 cm broad

(up to 5x4 cm in cultivated specimens), sparingly

to fairly densely appressed pubescent especially on

the nerves below and usually with orange gland-dots;

apex rounded; base cuneate; margin finely crenate-

dentate in the upper half, with 4-6 pairs of small

394

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

teeth. Inflorescence a terminal dense spike-like raceme

4-6 (-9) cm long, 4-angled and with a coma of large

imbricate bracts in the bud stage, with one or two

spaced verticils (rarely more) at the base; rhachis

glandular-tomentulose; bracts imbricate, ovate,

acuminate, 8-12 mm long; 6-8 mm broad, greenish-

white tipped with purple. Flowers in sessile 3-flowered

cymes, forming 6-flowered verticillasters; verti-

cillasters densely placed with the exception of 1-3

in the lower part; pedicels erect, appressed to the

rhachis, 3-4 mm long, glandular-puberulous. Calyx

3 mm at flowering enlarging to 6 mm long in fruit,

glandular-hispid and gland-dotted without, densely

villous in the throat; upper lip large, broader than

Closely related to P. neochilus but it seems as if

the two can be kept distinct. For the main differences,

see notes under the latter on p. 393. Both species

are in cultivation and are sometimes confused by

gardeners.

Although the name Coleus comosus was proposed

by Hochstetter for a Schimper specimen (11.1328)

at an early stage, it was included under Coleus

spicatus by A. Richard and under P. caninus by

Vatke and Engler (see literature citations). P. caninus

Roth (= Coleus spicatus Benth.) is an annual with

much smaller flowers (see p. 393). Giirke upheld C.

comosus as distinct and provided sufficient descriptive

Fig. 16.— Plectranthus orna-

tus, cultivated in Pre-

toria (Codd 8238), X 1.

long, up to 2,5 mm long, 6 mm broad with a median

apicule; lower 4 teeth subequal, linear-subulate,

3-4 mm long. Corolla bluish mauve with dark purple

mottling on the upper lip, 2-2,5 cm long, slightly

puberulous and with occasional orange gland-dots;

tube ca 0,75 mm in diam. for about 1/3 its length,

then slightly geniculate and expanding to 3 mm deep

at the throat; upper lip 6 mm long, 4 mm broad,

deeply emarginate and with 2 narrow lateral lobes;

lower lip boat-shaped, horizontal, 1,2-1, 5 cm long,

sometimes split longitudinally. Stamens united at the

base for 3-4 mm and curved within the lower lip,

12-14 mm long, anthers sometimes exserted. Style

finally exserted by 4-5 mm. Fig. 16.

Grows over rocks in semi-shade at altitudes of

I 000 to I 500 m, often above the forest zone, from

Ethiopia to Tanzania. Cultivated and semi-naturalized

in South Africa.

Culti vatkd. — Transvaal : from Mrs. Eliovson's garden in

Johannesburg, Codd 8238. Natal: Dundee, from a garden, Aug.

I960, collector not recorded. Cape: cult. Kirstenbosch National

Botanic Garden, Barker 2.

matter contrasting it with C. caninus to validate the

name. Unfortunately on transfer to Plectranthus it

must receive a new epithet as there is an earlier

P. comosus Sims (1822) which is a synonym of

P. barbatus Andr.

Chromosome number 2n =32 (De Wet, 1958, as

“ Coleus comosus").

15. Plectranthus barbatus Andr., Bot. Rep. t.

594 (1809). Type: the Bot. Rep. plate based on a

plant cultivated in England, raised from seed sent

from Abyssinia by Lord Valentia.

P. forskohlaei sensu Ait. f., Hort. Kew. ed. 2,3: 425 (1811);

sensu Sims in Bot. Mag. t. 2036 (1819). P. comosus Sims in

Bot. Mag. t. 2318 (1822). Type: the Bot. Mag. plate based on a

plant cultivated in England originally from India.

Coleus barbatus (Andr.) Benth. in Wall., PI. As. Rar. 2: 15

(1831); Lab. 49 (1832); in DC., Prodr. 12: 71 (1848); Hook, f.,

FI. Brit. India 4: 625 (1885); Trimen, Handb. FI. Ceylon 3:

373 (1895); Bak. in FI. Trop. Afr. 5: 429 (1900); Bruce in Kew

Bull. 1935: 322 (1935); Andrews, Flow. PI. Sudan 3: 208, t. 53

(1956). C. forskohlii sensu Briq. in Pflanzenfam. 4, 3a: 359

(1897).

L. E. CODD

395

Erect, bushy, semi-succulent, somewhat un-

pleasantly aromatic herb or shrub up to 3 or 4 m

tall; stems 4-angled, densely grey woolly tomentose

and freely dotted with reddish gland-dots. Leaves

semi-succulent; petiole 1-2 cm long; blade ovate to

broadly elliptical, 4-9 cm long 2,5-5 cm broad,

densely woolly tomentose on both surfaces and

copiously gland-dotted below; apex obtuse to

rounded; base obtuse to cuneate; margin regularly

crenate-dentate with 10-24 pairs of smallish teeth.

Inflorescence a terminal spike-like raceme 20-32 cm

long, enclosed in large imbricate bracts in the bud

stage, elongating and becoming lax with age; bracts

broadly ovate, acuminate, concave, 12-16 mm long,

10-12 mm broad, whitish green, shortly tomentose,

freely gland-dotted and with ciliate margins, early

deciduous. Flowers in sessile 3-4-flowered cymes,

verticillasters 1-2,5 cm apart; pedicels erect, 4-5 mm

long, glandular-puberulous. Calyx 4 mm long at

flowering enlarging to 7 mm long in fruit, glandular-

hispidulous without, densely villous in the throat;

upper lip broadly ovate, abruptly acuminate, 3-4 mm

long and 2-3 mm broad; lower 4 teeth subequal,

linear-subulate, 3-4 mm long. Corolla pale blue-

mauve, 1,7-2 cm long, sparingly pubescent and

gland-dotted; tube geniculate about the middle and

expanding from 2 mm at the base to 3 mm at the

throat; upper lip erect, 3 mm long, emarginate and

with 2 narrow lateral lobes; lower lip boat-shaped

1-1,3 cm long. Stamens united at the base for 3 mm

and curved within the lower lip with the anthers

slightly exserted. Style finally exserted by 4-5 mm.

Occurs at forest margins and among rocks in

shady places at fairly high altitudes in Ethiopia and

Sudan to Tanzania and also in India. Cultivated in

various parts of the World, including South Africa,

where it has become semi-naturalized in places.

Cultivated. — Transvaal: Pretoria, Repton 3477; Codd 6631;

Schlieben & Mendelsohn 12754. Natal : Pietermaritzburg, Cron-

wright s.n.; Durban North, Strey 3872. Cape: Grahamstown,

Britten s.n.; Troughton s.n.; Kirstenbosch, Barker 1 ; foothills of

Stellenbosch Mts., Bos 8.

This attractive plant was first introduced into

cultivation in England from Ethiopia in 1806 and was

named P. barbatus Andr. in 1809. Later it was again

introduced from India and received the name

P. comosus Sims (1822). The application to it of the

epithet “forskohlaei” is based on a misidentification,

according to Miss Bruce in Kew Bull. 1935: 322 (2935).

P. forskohlaei Vahl is, in any case, an illegitimate

name based on Ocimum hadiense Forsk., which is,

apparently, correctly placed in Ocimum.

The plants grown in South Africa rarely exceed

2 m in height but collectors indicate that it grows to

4 m tall in forest margins in tropical East Africa.

Subgen. Plectranthus.

Plectranthus L’Herit. sect. Germanea (Lam.) Benth.,

Lab. 32 (1832); in DC., Prodr. 12: 62 (1848). —sect.

Coleoides Benth., 11. cc. — subgen. Germanea (Lam.)

Briq. in Pflanzenfam. 4, 3a: 354 (1897).

Herbs or subshrubs usually softly succulent;

flowers in sessile or shortly pedunculate cymes, not

in paired dichasia; calyx clearly 2-lipped, the upper

lip broad, scarcely decurrent on the tube and approxi-

mately equal in length to the lower lip, the latter of

4 narrow acuminate teeth, the lowermost pair usually

slightly the longer and often shortly united at the

base; corolla tube variously gibbous to ±sigmoid,

sometimes with a basal spur; stamens free.

The ideal treatment of Plectranthus would be to

restrict the genus to ihe above circumscription which

agrees, with certain exceptions, to Briquet’s subgen.

Germanea.

Two sections can be recognized, though Bentham's

allocation of species needs to be revised. In his original

work of 1832 he restricted his sect. Germanea to

P. fruticosus L’Herit. on the basis of its spurred

corolla adding, in 1848, P. saccatus Benth. and

P. ciliatus Benth. in which the corolla tube is distinctly

saccate at the base. As there is considerable variation

in the degree of swelling of the corolla tube, this is

not a reliable basis for separating the two groups and

has resulted in something of a mixture in Bentham’s

sect. Coleoides.

A more meaningful grouping is obtained by using

calyx shape and whether the bracts are early

deciduous or not. These characters are also associated

with the number of flowers per cyme, as indicated

below.

Floral bracts usually early deciduous; calyx distinctly

gibbous ventrally; flowers in up to 15-flowered

sessile cymes sect. Coleoides

Floral bracts persisting to beyond the flowering stage;

calyx not markedly gibbous ventrally; cymes, when

sessile, 1-3-flowered sect. Plectranthus

Sect. Coleoides Benth., Lab. 32 (1832); in DC.,

Prodr. 12: 63 (1848), partly; Briq. in Pflanzenfam, 4,

3a: 354 (1897), partly; Codd in Mitt. Bot. Miinchen

10: 247 (1971). Lectotype: P. madagascariensis (Pers.)

Benth.

As here outlined, the section is restricted to those

species in which the floral bracts are shed before the

flowers open (sub-persistent in P. zatarhendi var.

tomentosus, see p. 399); the base of the mature calyx

is distinctly gibbous ventrally and is attached to the

pedicel at a sharp angle; and the sessile cymes are

usually many-flowered (ocassionally 3-flowered in

some forms).

This section is somewhat transitional between true

Plectranthus (i.e. sect. Plectranthus, see p. <06)

and subgen. Calceolanthus (p. 389) but, whereas the

stamens are united in the latter, they are always free

in sect. Coleoides.

An interesting point is that the chromosome

numbers recorded in this section by De Wet (1958)

were usually 2n=42 (P. tomentosus, P. woodii)

though occasionally 2n=28 (P. pachyphyllus), while

in sect. Plectranthus the number was, with rare

exceptions, 2n = 28 (P. fruticosus, P. ciliatus, P.

grallatus, P. saccatus, P. dolichopodus, etc.). In subgen.

Calceolanthus, on the other hand, the numbers he

recorded were 2n = 32 ( Coleus pentheri, C. comosus,

C. vagatus). The species names are those used by

De Wet and may be consulted by looking in the

index at the end of this paper.

Members of the section occur from South Africa

and the Mascarenes, through east tropical Africa to

southern Arabia, Malesia and Australia.

Species limits in this group are difficult to determine

and differences of opinion may well occur regarding

the number of species which should be upheld. In

the present treatment the Southern African material

is classified into the following components:

16. P. dinteri Briq.

17. P. grandidentatus Guerke

18. P. zatarhendi (Forsk.) E. A. Bruce

(a) var. zatarhendi ( =P . pachyphyllus Guerke ex

Cooke)

(b) var. tomentosus (Benth.) Codd ( =P . tomen-

tosus Benth., P. zeylanicus Benth.)

(c) var. woodii (Guerke) Codd (=P. woodii

Guerke)

396

PLECTRANTHUS (LAB1ATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

19. P. madagascariensis (Pers.) Benth.

(a) var. madagascariensis ( = P. hirtus Benth.)

(b) var. a/iciae Codd

(c) var. ramosior Benth.

20. P. psammophilus Codd

21. P. mutabilis Codd

It is admitted that the distinctions are often blurred

and that intermediates are found. However, the fact

that most specimens can be allocated with some

certainty seems to indicate that distinct species are

in the process of evolving.

The oldest name in the complex is P. zatarhendi

(Forsk.) E. A. Bruce, based on Ocymum zatar-

hendi Forsk. from southern Arabia. Christensen in

Dansk Bot. Arkiv. 4: 21 (1922) throws doubt on

Forskal's intention to use “zatarhendi” as a specific

epithet, but Bruce in Kew Bull. 1935: 590 (1935)

concludes that the method of presentation is con-

sistent with the legitimate publication of O. zatarhendi

as a species name and this view has since been

followed. The unfortunate discrepancies found in the

FI. Aegypt.-Arab. may be attributed to the

posthumous publication of Forskal’s notes by

Niebuhr. If a very broad view were taken, the

majority of our specimens (with the possible

exception of P. madagascariensis and P. psammophilus)

could be accommodated in that species. It is

considered that such a treatment would be less

satisfactory than the present one and would not

reflect the great variation which occurs, nor would

it indicate the presence of several natural groups.

Among our described “species”, P. pachyphyllus

probably comes closest to P. zatarhendi in having

short densely tomentose stems, a short simple

inflorescence, and thick-textured, broadly ovate,

crenate-dentate leaves (see figs. 18, 19). The concept

is enlarged to include var. tomentosus with a more

robust stature and long, branched inflorescences, and

var. woodii with more sparsely tomentose stems and

leaves.

16. Plectranthus dinteri Briq. in Bull. Herb.

Boiss. 2 ser. 3: 1070 (1903). Type: South West

Africa, Hereroland, Waterberg, Dinter 336 (B. holo.t).

P. zatarhendi sensu Launert & Schreiber in Prodr. FI.

S.W.Afr. 123:26(1969).

Erect annual or perennial semi-succulent aromatic

herb up to 40 cm (rarely to 1 m) tall; stems 4-angled,

simple or sparingly branched, densely glandular-

yellowish-tomentose with long and short multi-

cellular hairs, gland-tipped hairs and sessile gland-

dots. Leaves softly succulent drying thickish to thin-

textured; petiole 2,5-4 cm long, tomentose like the

stems; blade ovate to broadly ovate, 3-9 cm long,

2,5-7 cm broad, coarsely glandular-strigose to

glandular-tomentose on both surfaces, freely dotted

with small red gland-dots below; apex acute to

obtuse; base truncate; margin coarsely to deeply

dentate with 3-6 pairs of triangular teeth 3-8 mm

long. Inflorescence terminal, simple or with a pair of

branches near the base, 10-25 cm long; rhachis

glandular-hispid with coarse candle-like gland-tipped

hairs, shorter hairs and gland-dots; bracts broadly

ovate, 1,5-2 mm long, early deciduous (before

flowering). Flowers in sessile 3-6-flowered cymes

forming 6-1 2-flowered verticillasters; verticillasters

1-2,5 cm apart; pedicels 3-5 mm long, glandular-

puberulous. Calyx 2 mm long at flowering enlarging

to 4 mm long in fruit with a distinctly gibbous base,

purple-tinged, glandular-scabrid and red gland-

dotted; upper lip suberect, ovate, obtuse to abruptly

acute, scarcely decurrent on the tube, up to 2 mm

long; lower lip ^equally 4-toothed, up to 2 mm

long, teeth linear-lanceolate, subulate. Corolla mauve

to purple, pubescent and red gland-dotted especially

on the lips, 8-10 mm long; tube 4-5 mm long,

expanding gradually to 2,5 mm at the throat, bent

about the middle; upper lip erect to recurved, 2 mm

long and equally broad, apex emarginate and with

2 lateral ear-like lobes; lower lip boat-shaped,

equal to or slightly longer than the tube, 4-5 mm long,

horizontal. Stamens free at the base, 4-5 mm long,

curved and more or less enclosed within the lower lip.

Style coinciding with the upper stamens.

Found in the north-central part of South West

African in sandy places and rock crevices, particularly

on the dolomite formation.

S.W.A. — 1917 (Tsumeb): 13 km from Tsumeb on Otavi

road (-BA), Hardy 2130 ; Nosib (-BD), Dinter 2426 (SAM);

Sehoenfelder S805 ; Auros (-DA), Dinter 5606; Kombat (-DA),

Cole T12. 2017 (Waterberg): Waterberg Plateau (-AC), Giess,

Volk & Bleissner 6579. 2217 (Windhoek): 16 km E. of Wind-

hoek, Leach & Bayliss 12931 .

No material of the type, Dinter 366, has been

seen but there is no doubt concerning the concept

and Dinter 2426 and 5606 may be regarded as fully

representative.

With its deeply dentate leaves it is reminiscent of

P. grandidentatus Guerke but there are several

differences, mainly in the pubescence. The distinction

can best be seen on the rhachis which, in P. grandi-

dentatus is shortly tomentulose often with fine gland-

tipped hairs, while in P. dinteri it is coarsely glandular-

hispid with characteristic stout candle-like gland-

tipped hairs. This character distinguishes P. dinteri

from all other members of the complex.

Launert & Schreiber, l.c., included P. dinteri in

P. zatarhendi and, naturally, if such a wide concept

is adopted, the oldest epithet, zatarhendi must be

used.

17. Plectranthus grandidentatus Guerke in Bull.

Herb. Boiss. 6: 554 (1898), partly: Cooke in FI. Cap. 5,

1: 278 (1910); Compton, FI. Swaz. 66, 157 (1966);

Van der Schijff, Check List Kruger Nat. Park 82

(1969); Codd in Mitt. Bot. Miinchen 10: 248 (1971);

Ross, FI. Natal 305 (1972). Lectotype: East Griqua-

land, Enyembe Mt., Tyson sub Herb. Austr. Afr.

1517 (K!) (see note below).

Perennial semi-succulent aromatic herb; stems

procumbent, 4-angled, up to 2 m long, fairly dense to

densely shaggy tomentose, usually greyish. Leaves

softly succulent, often drying thick-textured; petiole

1,5-4, 5 cm long, tomentose like the stems; blade

ovate to broadly ovate, 2-7 cm long, 1,8-7, 5 cm

broad, sparingly hispid to densely strigose-pubescent

above, medium to densely floccose-tomentose below

and freely dotted with minute sessile red to brown

gland-dots; apex acute; base truncate; margin deeply

dentate with 4-7 pairs of triangular teeth 3-7 mm long.

Inflorescence terminal often on short lateral shoots,

usually simple, occasionally with 1-2 pairs of branches

near the base, 9-24 cm long; rhachis densely glandular-

tomentulose; bracts broadly ovate to subrotund,

3 mm long, early deciduous (before flowering).

Flowers in sessile 3-6-flowered cymes forming 6-12-

flowered verticillasters; verticillasters 5-10 mm apart;

pedicels 2-3 mm long, densely glandular-puberulous.

Calyx 2 mm long at flowering enlarging to 4 mm long

in fruit with a distinctly gibbous base, purple tinged,

glandular-scabrid; upper lip erect, broadly ovate,

obtuse to rounded, scarcely decurrent on the tube,

up to 2 mm long; lower lip ±equally 4-toothed, up to

2 mm long, teeth linear-lanceolate, subulate. Corolla

L. E. CODD

397

Fig. 17. — Plectranthus grandUentstus,

Tabankulu, eastern Cape Province

( Story 4209), X 1 .

white (rarely purple) pubescent and red gland-dotted,

especially on the lips; tube 3-5 mm long expanding

at the base to 1 ,5 mm deep and enlarging to 2,5 mm

at the throat, slightly bent; upper lip erect, recurved

towards the apex, emarginate and with 2 lateral

ear-like lobes; lower lip boat-shaped, longer than the

tube, 4-8 mm long, horizontal. Stamens free at the

base, 5-8 mm long, curved and enclosed within the

lower lip or shortly protruding. Style coinciding

with the upper stamens. Fig. 17.

Found in relatively dry, rocky places and forest

margins from about Queenstown in the eastern Cape

through central and northern Natal to Swaziland

and eastern Transvaal, reaching the Soutpansberg.

Transvaal. — 2329 (Pietersburg): Hanglip (-BB), Meense

10161; Louis Trichardt (-BB), Breyer sub TRV 22721; 10 km S.

of Pietersburg (-CD), Codd 10458; Boyne (-DD), Gerstner

5381. 2430 (Pilgrim's Rest): near Vaalhoek (-DB), Meeuse

10014. 2530 (Lydenburg): Lunsklip Waterfall (-AD), Strey

3195; Schagen (-BD), Liebenberg 2349; Nelshoogte Forestry

Station (-DD), Codd 9558. 2531 (Komatipoort) : Ship Mt., 16

km S.E. of Pretoriuskop (-AB), Codd 5149; 6044; Van der

Schijff 3773; near Barberton (-CC), Codd 8168; 8190. 2630

(Carolina): Iswepe (-DC), Sidey 1601. 2729 (Volksrust): near

Volksrust (-BD), Mogg 7040.

Swaziland. — 2631 (Mbabane): Mukusini Hills (-AC),

Compton 31604; Millers Falls (-AC), Compton 25866; Black

Umbuluzi Falls (-AD), Compton 26587; near Mankaiana

(-CA), Compton 27709.

Natal. — 2730 (Vryheid): Paul Pietersburg (-BD), Gilmore

1710; near Utrecht (-CB), Breyer sub TRV 17005. 2829 (Harri-

smith): Biggarsberg (-BB), Rehmann 7065 (Z); Codd 8621;

near Colenso (-DB), Rehmann 7179 (Z); Acocks 10251 ; Pole

Evans 4777. 2830 (Dundee): near Helpmekaar (-AD), Codd

5933; 13 to 16 km from Kranskop on Mambula road (-DD),

Dyer 4352; 4357. 2929 (Underberg): near Estcourt (-BB),

Pentz 307; Umkomaas Valley between Boston and Bulwer

(-DD), Dyer 4870. 2930 (Pietermaritzburg): Mt. Ashley (-AC),

Moll 1316; Karkloof (-AC), Molt 3519; Byrne (-CC), Strey

10846.

Cape. — 3029 (Kokstad): Zuurberg, near Kokstad (-BC),

Tyson 1177; Emyembe (-BD), Tyson sub. Herb. Austr. Afr.

1517 (K); Tyson 2163 (K). 3126 (Queenstown): near Queens-

town (-DD), Galpin 2090b; 8163; Shiloh (-DD), Baur 797 (K).

3129 (Port St. Johns): 10 km S. of Tabankulu, Story 4209.

P. grandidentatus is a fairly clear-cut entity charac-

terised by trailing or straggly stems, floccose-

tomentose stems and leaves, deeply dentate leaves

which are broadly truncate at the base and usually

white flowers. It has relatively thickish, succulent

leaves and stems which are no doubt associated

with its ability to grow in fairly dry habitats. There is

a good deal of variation in leaf size. At one end of the

scale it approaches P. madagascariensis and occasional

specimens may be difficult to classify with certainty

while, at the other end, there is some tendency to

grade into P. zatarhendi, possibly as a result of

hybridization. This may result in the occasional

mauve or purple flowers noted by collectors, whereas

usually the white flowers stand out clearly from the

purple calyces.

Giirke based his species on one Tyson and nine

Rehmann specimens, not all of which are included in

the present concept of P. grandidentatus. The Rehmann

specimens listed below, annotated by Giirke, are in

Z. They are now classified as follows:

Rehmann 4508 4509, from Wonderboompoort

near Pretoria, are Iboza brevispicata N.E. Br.

Rehmann 6149, 6160, 6161, from Houtbosch, are

nearer to P. zatarhendi than to P. grandidentatus.

Rehman 7038, 7040, from near Newcastle, have

rather small, less deeply dentate leaves and are

somewhat intermediate between P. grandidentatus

and P. madagascariensis var. ramosior.

398

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Rehman 7065 from Biggarsberg and Rehman 7179

from near Colenso fit in fairly well with the

present concept of P. grandidentatus.

Tyson sub Herb. Austr. Afr. 1 5 1 7 in Kew Herbarium

is selected as the lectotype. This gathering does not

appear to be represented in Zurich or in Berlin.

Chromosome number 2n 42 (De Wet, 1958).

18. Plectranthus zatarhendi (Forsk.) E. A. Bruce

in Kew Bull 1935: 590 (1935). Type: Arabia Forskal

348 (C!, holo.).

Perennial aromatic semi-succulent herb; stems

erect to decumbent, 50-150 cm tall, 4-angled, sparsely

to very densely tomentose, often with glandular hairs,

long and short hairs intermingled. Leaves medium to

thick textured; petiole 1-4 cm long, tomentose like

the stems; blade ovate to subrotund (3,5-) 4-10,5 cm

long, (2,5-) 3-10 cm broad, sparingly strigose to

densely appressed tomentose above, sparingly

pubescent (mainly on the nerves) to very densely

woolly-tomentose below and usually gland-dotted,

sometimes gland-dots absent or obscured by the

tomentum; apex acute to rounded; base cuneate to

subcordate; margin shallowly to fairly distinctly

crenate-dentate with 4-15 pairs of teeth. Inflorescence

terminal or with 1-2 pairs of branches near the base,

12-60 cm long; rhachis sparsely to densely and

shortly glandular-tomentose; bracts broadly ovate to

subrotund, up to 4 mm long, usually early deciduous

but frequently persisting to the flowering stage (in

var. tomenlosus). Flowers in sessile 4-15-flowered

cymes forming 8-25-flowered verticillasters; verti-

cillasters 1-3 cm apart; pedicels 2-4 mm long sparsely

to densely and shortly glandular-hirsute. Calyx 2 mm

long at flowering enlarging to 5 mm long in fruit

with a distinct gibbous base ventrally, often purple-

tinged, glandular-scabrid and usually gland-dotted;

upper lip erect, broadly ovate, obtuse, scarcely

decurrent on the tube, up to 2 mm long; lower lip

± equally 4-toothed, up to 2 mm long, teeth linear-

lanceolate, subulate. Corolla 8-13 mm long usually

in shades of mauve to purple, rarely white, pubescent

and gland-dotted on the upper and lower lips; tube

expanding gradually from the base and bent about

the middle; upper lip erect, often recurved towards the

apex, emarginate and with 2 lateral somewhat obscure

ear-like lobes; lower lip boat-shaped usually longer

than the tube, horizontal. Stamens free at the base,

curved and enclosed within the lower lip or shortly

protruding. Stvle coinciding with the upper stamens.

Fig. 18.

According to the present concept, the species

occurs in forest margins, dry woodland and rocky

places in grassland, and is distributed from the Cape

Province through Natal, Swaziland, Transvaal,

tropical East Africa and Somalia to southern Arabia.

The form described as P. zeylanicus Benth. is evidently

cultivated in Ceylon.

See note on p. 396 regarding the validity of the

name P. zatarhendi.

Judging from the few specimens seen from Arabia

and Somalia, the typical form has relatively short,

somewhat decumbent stems, large, densely tomentose,

rather shallowly crenate-dentate leaves (Fig. 18) and

a short, simple or rarely branched inflorescence with

4-8 flowers in the axil of each bract. In South Africa

this is fairly well matched by the entity described as

P. pachyphyllus Guerke ex T. Cooke and no satis-

factory grounds can be found for separating it from

typical P. zatarhendi. Two further groups are separated

at varietal level, namely: (a) var. tomentosus (Benth.)

Codd, which includes P. zeylanicus (Benth.), con-

sisting of plants of robust stature up to 1,5 m tall

with large, normally branched inflorescences and

5-15 flowers in the axil of each bract and (b) var.

woodii (Guerke) Codd in which the stem and leaf

pubescence is more sparsely strigose. There are

intermediates connecting these groups and thus

varietal status appears to be appropriate.

Key to varieties

Stems and leaves densely to very densely tomentose:

Stems 1 or more from a perennial often burnt base, up to

60 cm long; inflorescence simple or frequently with a

pair of branches near the base (a) var. zatarhendi

Stems robust, usually solitary and erect, branching above

(occasionally decumbent), up to 150 cm tall inflore-

scence large, with 1 or 2 pairs of branches near the

base (rarely simple) (b) var. tomentosus

Stems and leaves sparsely to fairly densely strigose

(c) var. woodii

(a) var. zatarhendi.

Ocymum zatarhendi Forsk., Fi. Aegypt. Arab. 109 (1775).

Type: Arabia, Forskal 348 (Cl, holo.).

Plectranthus zatarhendi (Forsk.) E. A. Bruce in Kew Bull.

1935: 590 (1935). P. crassifolius Vahl, Symb. 1:44 (1790);

Willd., Sp. PI. 3: 169 (1800); nom. illegit. Type: as for O.

zatarhendi Forsk. P. pachyphyllus Guerke ex T. Cooke in FI.

Cap. 5,1:285 (1910). Type: Natal, Inchanga, Rehmann 7878

(Z !, holo.). P. aegyptiacus C. Chr. in Dansk Bot. Arkiv. 4,3:

22 (1922), nom. illegit. Type: as for P. zatarhendi.

Gennanea crassifolia (Vahl) Poir. in Lam., Encycl. Suppl.

2: 764 (1812), nom. illegit. Type: as for O. zatarhendi.

Coleus zatarhendi (Forsk.) Benth. in DC., Prodr. 12: 71

(1848); Schwartz, FI. Trop. Arab.

Stems 1-few arising from a perennial base, erect or

decumbent, 30-60 cm long, sparingly branched,

densely glandular-tomentose; leaves broadly ovate,

3, 5-8 x 2-5, 5 cm, densely tomentose on both

surfaces, shallowly crenate-dentate: inflorescence 8-30

cm long, simple or, occasionally, with a pair of

branches near the base, flowers 4-8 in the axil of each

bract, bracts early deciduous. Fig. 18.

Found in the midlands and central coastal Natal

and in mountainous parts of eastern and central

Transvaal, among rocks in dry woodland or on

exposed rocky places in grassland where it is subjected

to frequent burning. It extends through east tropical

Africa to southern Arabia.

Transvaal. — 2328 (Baltimore): about 80 km N.W. of

Potgietersrus (-DA), Maguire 2576. 2428 (Nylstroom): Zand-

rivierpoort (-AC), Coetzee 981; Twenytfour Rivers (-AD),

Rogers 23627; Palala River (-BC), Breyer sub TRV 17783;

poort on Palala Road (-BC), Smuts & Gillett 3431 ; 21 km N.W.

of Warmbaths (-CC), Acocks 23573. 2429 (Zebediela): Sekuku-

niland, Schoonoord (-DD), Barnard & Mogg 874. 2529 (Wit-

bank): near Waterval (-CA), Repton 1204; 2530 (Lydenburg):

near Izaak Siding, Galpin 13300.

Natal. — 2830 (Dundee); Ntunjambile Mt. (-DD), Codd

9660. 2831 (Nkandla): Babanango (-AC), King 262; on road

from Hluhluwe to Nongoma (-BB), Wells 2068; 13 km S. of

Nkandla (-CA) Edwards 1396; Codd 9684. 2930 (Pietermaritz-

burg): near The Dargle (-AC), Marais 815; Groot Noodsberg

(-BD), Strey 6043; 8 km from Merrivale on Boston Road

( CA), Moll 768; near Inchanga (-DA), Rehmann 7878;

Eshuis s.n.; Inanda (-DB), Strey 5164; near Botha's (-DC),

Medley Wood 4775; Krantzkloof (-DD), Haygarth sub TRV

24983; Kloof Nature Reserve (-DD), Dohse 258; Westville

(-DD), Moll 2343; Emberton (-DD), Strey 5432. 3030 (Port

Shepstone): Umgaye (-BC), Rudatis 685.

The above description is based largely on the

plants occurring in South Africa which apparently

cannot be separated from typical P. zatarhendi.

Further study is, however, required. For example,

little is known of the root system. The only complete

specimen seen is Dohse 258 which shows several

horizontal fleshy roots. Whether this will provide a

diagnostic character remains to be seen.

L. E. CODD

399

Fig. 18. — Plectranthus zatarhendi (holo-

type: Forskahl in C), Xl.

Chromosome number 2n =28 (De Wet, 1958, as

“P. pachyphyllus").

The type of P. pachyphyllus is included here

although the leaves ( 1 1—14 mmx9-12 mm) are

somewhat smaller than the other specimens cited.

They may be at an immature stage. The thick texture

and dense tomentum suggest that the specimen

belongs in this group rather than in P. madagasca-

riensis var. ramosior , which occupies a similar ecolo-

gical niche mainly on the Magaliesberg and Wit-

watersrand formations.

As indicated on p. 404 it is probable that the

smaller-leaved var. ramosior was derived from P.

zatarhendi rather than from P. madagascariensis.

In addition to being smaller, the leaves of var.

ramosior are not as thick-textured and are usually

less densely tomentose than those of P. zatarhendi.

At the other end of the scale is var. tomentosus

with usually larger leaves, more robust stature and

larger, branched inflorescences.

Although P. dinteri Briq. was included in P. zatar-

hendi by Launert & Schreiber in Prodr. FI. S.W. Afr.

123: 26 (1969). it is now kept distinct on the grounds

of the more deeply dentate leaves and markedly

glandular-hispidulous rhachis.

(b) var. tomentosus ( Benth .) Codd, stat. nov.

P. tometosus Benth. in E. Mey., Comm. 229 (1837); Drege,

Zwei Doc. 159, 160 (1843); Benth. in DC., Prodr. 12:67 (1848);

Briq. in Pflanzenfam. 4,30: 357 (1897); Wood, Natal PI. 4:t,316

(1906); Cooke in FI. Cap. 5,1 :286 (1910), partly; Dyer in Flow.

PL S. Afr. 24:t.960 (1944); Compton, FI. Swaz. 67 (1966);

Codd in Mitt. Bot. Miinchen 10:248 (1971); Ross, FI. Natal

305 (1972). Type: Port Natal, Drege (Kl. holo.; MO!; P!;

S!). P. zeylanicus Benth., Lab. 36 (1832); FI. Brit. India 4:622

(1885): Trimen, Flandb. FI. Ceylon 3:371 (1895); Blake, Contr.

Queensl. Herb. 9: 11, 44, t .28 (1971). Type: Ceylon, Macrae

(CGE, holo.).

Stems densely to shaggily tomentose, usually

solitary, erect and branched above, up to 1,5 m tall

or decumbent to 70 cm long, roots fibrous; leaves

broadly ovate to subrotund, 4-10,5x3,2-10 cm,

densely tomentose on both surfaces shallowly to

26700-3

400

PLECTRANTHUS (LABI AT AE) AND ALLIED GENERA IN SOUTHERN AFRICA

!

fairly distinctly crenate-dentate with usually 6-15

pairs of teeth; inflorescence 18-60 cm long usually

with 1 or 2 pairs of branches near the base,

occasionally simple, flowers 5-15 in the axil of each

bract, bracts often persisting to flowering stage.

Fig. 19.

Found mainly in semi-coastal areas from about

Komga in the eastern Cape through coastal Natal,

extending inland to Swaziland and the eastern

Transvaal, in dry woodland and rocky grassland.

Cultivated in Ceylon and India.

Transvaal. — 2230 (Messina): near Lake Funduzi (-CD),

Meeuse 9431 ; Makonde (-DA), Van Warmelo 5116/15. 2330

(Tzaneen): Wesfalia Estate (-CA), Scheepers 954. 2430 (Pil-

grim’s Rest): The Downs (-AA), Codd 9466; 18 km S.W. of

Ofcolaco (-AB), Joubert 2; 8 km S.E. of P.O. Steelpoort (-CC),

Codd 7718; Ohrigstad Dam Nature Reserve (-DC), Jacobsen

2364; 29 km N. of Lydenburg (-DC), Joubert 1. 2530 (Lyden-

burg): near Lydenburg (-AB), Wilson 1125 (K); near Badplaas

(-DC), Leach 8857. 2531 (Komatipoort): Crocodilepoort Mt.

(-CA), Codd 7768; near Barberton (-CC), Holt 46.

Swaziland. — 2531 (Komatipoort): Piggs Peak (-CC),

Compton 27621. 2631 (Mbabane): Komati Bridge (-AA),

Compton 26862; Komati Pass (-AA), Compton 31511; Usutu

Canal (-AC), Compton 27760; near Mankaiana (-CA), Compton

27709.

Natal. — 2731 (Louwsburg): near Ngome (-CD), Codd

7015; 9587. 2830 (Dundee): near Nqutu (-BA), Codd 8964;

14 km W. of Muden (-CD), Codd 8608; about 15 km from

Krantzkop on Mambula Road (-DD), Dyer 4351; 4352. 2831

(Nkandla): Ntonjaneni (-AD), Andrews 14; 4 km N. of Hlabisa

(-BB), Codd 9613; near Mtunzini (-DC), Mogg 4416; Ngoye

(-DC), Medley Wood 5752; 10 km S.E. of Eshowe (-DC),

Codd 9647. 2832 (Mtubatuba): Hluhluwe Game Reserve (-AA),

Ward 2993; 3202; 20 km S. of Mtubatuba (-CA), Strey 5611.

2929 (Underberg): Estcourt Research Station (-BB), West 712;

8 km E. of Estcourt (-BB), Acocks 10140. 2930 (Pietermaritz-

burg): 6 km E. of Greytown (-BA), Admiraal & Drijfhout

2898; Ahrens (-BB), Fisher 926; Richmond, Tala Farm (-CD),

Moll 3055; Inanda (-DB), Medley Wood 488 (K); near Durban

(-DD), Drege s.n. (K, MO, P, S); Krauss 75 (K); Sanderson

126 (K); 550 (K); Rehmann 8820 (Z); 3030 (Port Shepstone):

Uvongo (-CD), Liebenberg 8022; Nicholson 864; St. Michaels-

on-Sea (-CD), Nicholson 1115; Shelly Bay, 11 km S. of Port

Shepstone, Mogg 12163.

L. E. CODD

401

Cape. — 3128 (Umtata): near Old Morley (-DD), Codd 9270.

3129 (Port St. Johns): 5 km S.W. of Ludalasi Store (-CD),

Codd 9288. 3227 (Stutterheim): near Komga (-DB), Flanagan

740. 3228 (Butterworth): near Kentani (-CB), Pegler 488.

Both Hooker f., FI. Brit. India 4: 622 (1885), and

Trimen, Handb. FI. Ceylon 3: 371 (1895) state that

P. zeylanicus is commonly cultivated in native

gardens for medicinal purposes and that it is not

known in the wild state. It cannot be separated from

typical P. tomentosus which grows along the Natal

Coast and it may be assumed that the Asiatic plants

came originally from Natal. A parallel is found in

P. amboinicus, now widely cultivated in the East,

which also originated in Africa. The varietal epithet

tomentosus has been adopted as it is already familiar

in South Africa.

In its typical form it is a robust, erect plant with

one or two stems arising from a strong, fibrous root

system, reaching 1,5 m tall when in flower with a

large, branched, inflorescence of pale mauve flowers

(see Flow. PI. S. Afr. 24: t. 760, 1944).

However, similar plants but with shorter, decum-

bent stems are found, particularly in Swaziland and

Transvaal, and the distinction between this form and

var. zatarhendi is far from clear. Some of these

specimens have fairly deeply crenate-dentate leaves

and there may be difficulty in distinguishing them

from P. grandidentatus. Usually in such cases the

dense shaggy white tomentum of P. grandidentatus

is used as a diagnostic character. Thus specimens

such as Codd 8168 and 8190 from Barberton are

included in P. grandidentatus, while Meeuse 9431 and

Van Warmelo 5116/15 from the Soutpansberg area,

with finer tomentum, are placed in var. tomentosus.

Chromosome number 2n =42 (De Wet, 1958).

The plants separated as var. woodii usually have

smaller, less tomentose leaves and less branched

inflorescences, but here again it is not always easy to

allocate plants which show intermediate charac-

teristics.

(c) var. woodii ( Guerke ) Codd, stat. nov.

Plectranthus woodii Guerke in Bot. Jahrb. 26:76 (1898)

(sphalm. "Wodii”); Cooke in FI. Cap. 5,1:287 (1910); Codd in

Mitt. Bot. Munchen 10: 248 (1971); Ross, FI. Natal 305 (1972).

Type: Natal, Ipolweni, Wood s.n. (GRA., lectotype). P. draconis

Briq. in Bull. Herb. Boiss. ser. 2,3: 1071 (1903); Cooke in FI.

Cap. 5,1:288 (1910). Type: Natal, Biggarsberg, Rehmann 7092

(Z!, holo.).

Stems 1-few, arising annually from a perennial base,

decumbent or suberect, 30-60 cm long sparingly

branched, glandular-puberulous to sparsely or fairly

densely strigose; leaves ovate-elliptical to broadly

ovate (3-)3,5-6x2,5-4,8 cm, shortly hispid to

sparingly or fairly densely strigose, shallowly to fairly

distinctly crenate-dentate with 4-7 pairs of teeth;

inflorescence 10-35 cm long, simple or with a pair of

branches near the base, flowers 3-8 in the axil of each

bract, bracts early deciduous. Fig. 20.

Fig. 20. — Plectranthus zatarhendi var.

woodii (lectotype: Ipolweni, Natal,

Medley Wood s.n. in GRA).

5'$ 02

T~nrn

402

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Found among rocks in dry Euphorbia and thorn

scrub, in grassland and on wooded krantzes, mainly

in central Natal, extending to the Transvaal and

across the southern border into the eastern Cape

Province.

Transvaal. — 2529 (Witbank): 67 km N.E. of Bronkhorst-

spruit (-AC), Lavranos s.n.; 11 km E. of Loskopdam wall

(-AD), Codd 9842; near Laerdrift (-BD), Codd 9844. 2531

(Komatipoort): Kaapmuiden (-CB), Mogg s.n.; Barberton

(-CC), Pott 5435.

Natal. — 2829 (Harrismith): Biggarsberg (-BB), Rehmann

7092 (Z); Codd 8620. 2830 (Dundee): near Tugela Ferry (-CD),

Codd 5937; 5939; between Tugela River and Greytown (-DC),

Werdennann & Oberdieck 1357; 16 km N.W. of Greytown

(-DC), Codd 8596; 8597. 2930 (Pietermaritzburg): Ipolweni

(-AD), Medley Wood s.n. (GRA); Ahrens (-BB), Fisher 974;

near Harburg (-BC), Marais 799; Inanda (-DB), Pole Evans

s.n.; Appelsbosch (-DD?), Strey 6420. Fields Hill (-DD),

Rehmann 8032 (Z); Isipingo (-DD), Forbes & Obermeyer 10.

Cape. — 3029 (Kokstad): Umtamvuma Waterfall (-DD?),

Strev 4472. 3030 (Port Shepstone): near Punzi Drift (-CC),

Codd 9344.

The type of P. draconis Briq. is matched by Codd

8620 from the same area, the Biggarsberg. These two

specimens are markedly less pubescent than the others

cited above and are included with some hesitation.

Further study may indicate that this form is worthy

of separate rank.

Many specimens above appear to come from a

perennial rootstock as is the case with var. zatarhendi,

while others have leaves reminiscent of var.

tomentosus. Yar. woodii is distinguished from both

these by the less dense tomentum on stems and

leaves. Some specimens may be confused with

P. madagascariensis but the leaves of var. woodii

are usually larger although there is some overlapping

is this respect.

Chromosome number 2n =42 (De Wet, 1958).

P. woodii is based on two specimens from Natal:

Impolweni (“Ipolweni”), Wood s.n. collected on the

8th April 1891, and Fields Hill (“Vildshill”), Pine-

town, Rehmann 8032. No material of either gathering is

extant in Berlin, but the Rehmann specimen, anno-

tated by Giirke, is present in Zurich. Unfortunately it

is a scrappy specimen and its correct identity is a

little uncertain. In order, therefore, to ensure that

the name for the present concept of the group is

preserved, an isotype of the Wood gathering in GRA

is selected as the lectotype.

19. Plectranthus madagascariensis (Pers.) Benth.,

Lab. 37 (1832). Type: Madagascar, Commerson

(Herb. Juss.!, holo.; P!).

Perennial aromatic herb; stems procumbent, up

to 1 m long, or erect, up to 30 cm (excluding inflores-

cence), 4-angled, sparsely to densely short tomentose,

often with longer hairs and glandular hairs inter-

mingled. Leaves slightly succulent, drying thin to

thickish in texture ; petiole 0 , 5-3 , 5 cm long, tomentose

like the stems; blade ovate to subrotund, sometimes

broader than long, 1,5-3 (sometimes to 4,5) cm

long, 1,2-4 cm broad, sparingly to densely and

shortly strigose above, medium to densely and

shortly tomentose below and freely dotted with

reddish to brown gland-dots; apex obtuse to rounded;

base truncate to cuneate; margin obscurely crenate to

shortly crenate-dentate with 3-7 pairs of teeth.

Inflorescence terminal and on lateral branchlets,

usually simple, occasionally with 1-2 pairs of

branches near the base, 9-25 cm long; rhachis

sparsely to fairly densely and shortly glandular-

hirsute; bracts broadly ovate to obovate, 3 mm long,

early deciduous (before flowering). Flowers in sessile

3-8-flowered cymes forming 6-16-flowered verticil-

lasters; verticillasters 0,5-2 cm apart; pedicels

2-3 mm long, shortly glandular-hirsute. Calyx 2 mm

long at flowering enlarging to 5 mm long in fruit with

a distinct gibbous base, often purple-tinged, glandular-

scabrid and gland-dotted; upper lip erect, broadly

ovate, obtuse, scarcely decurrent on the tube, up to

2 mm long; lower lip ± equally 4-toothed, up to

2 mm long, teeth linear-lanceolate, subulate. Corolla

white or various shades of mauve to purple,

pubescent and reddish gland-dotted on the upper

and lower lips, 0,5-1, 8 cm long; tube expanding

gradually from the base and bent below the middle;

upper lip erect, recurved towards the apex, emar-

ginate and with 2 lateral somewhat obscure ear-like

lobes; lower lip boat-shaped, longer than the tube,

horizontal. Stamens free at the base, curved and

enclosed within the lower lip or shortly protruding.

Style coinciding with the upper stamens.

The species, as here outlined, is found in forest

margins, dry woodland and rocky places in grassland,

and occurs in the Cape Province, Natal, Swaziland

and Transvaal. It has also been recorded from

Mozambique, Madagascar (?) and Mauritius.

The main criterion for separating P. madagasca-

riensis from other species in this complex is leaf size.

In this species the leaves rarely exceed 3,5 cm long

but some overlapping may occur with the lower

limits of leaf size in P. zatarhendi. In such cases (which

are not very common), a certain amount of discretion

must be used and a few specimens with leaves up to

4,5 cm long are considered to belong in P. madagas-

cariensis rather than in P. zatarhendi. Some confusion

may also occur with P. grandidentatus, but this species

usually has larger leaves, thicker in texture with

denser and more shaggy tomentum and, of course,

more deeply dentate leaves. Occasional specimens

may be difficult to allocate because of a certain

amount of gradation of leaf size and margin character,

especially in Natal. See also discussion under P.

mutabilis (p. 405) for the reasons for separating this

species, with its variable, crenate leaves, often lacking

the glands dots characteristic of the other species in

this complex.

A good deal of variation in habit, pubescence,

colour and size of corolla is included in this concept

of P. madagascarensis and three varieties are recog-

nized. The typical variety has procumbent stems,

rooting at the nodes, and white flowers (sometimes

shades of mauve or lilac) and is largely concentrated

in the eastern Cape Province, extending to coastal

Natal with a somewhat doubtful record from Swazi-

land. A form with variegated leaves is commonly

cultivated, but its place of origin is not known.

Var. ramosior occurs in the Transvaal, Swaziland

and the more inland parts of Natal and eastern Cape

Province. It has ascending (rarely procumbent)

stems arising from a perennial rootstock and is

adapted to drier and colder situations among rocks

in grassland subjected to frequent burning. The burnt

base of the plant is evident in many specimens.

Incomplete specimens are not always easy to dis-

tinguish from var. madagascariensis while there is

some gradation of leaf size in Natal with P. zatarhendi

var. zatarhendi , which grows under essentially the

same ecological conditions.

Var. aliciae is known from only two gatherings by

Miss Alice Pegler near Kentani. It has thin textured

leaves which are sparingly pubescent, and very small

white flowers. It seems worthy of separate rank,

though more collecting in the area is required in

order to assess its status satisfactorily.

L. E. CODD

403

Key to varieties

Corolla 7-18 mm long; leaves medium to thick textured,

teeth usually small and pubescence fairly dense to

dense :

Stems decumbent to procumbent, sparingly branched;

flowers usually white (rarely mauve or purple). . . .

(a) var. madagascariensis

Stems erect or ascending usually from a thick perennial

(often burnt) base, usually branched with several

inflorescences; flowers usually mauve or purple

(rarely white) (c) var. ramosior

Corolla 5 mm long; leaves thin textured, sparsely pubescent

with few fairly large rounded teeth (b) var. aliciae

(a) var. madagascariensis.

Ocimum madagascariensis Pers., Syn. PI. 2:135 (1807). Type:

“Madagascar”, Commerson (Herb. Juss.l, holo.; P ! ). O.

tomentosum Thunb., Prodr. 96 (1800); FI. Cap. ed. Schult. 448

(1823), non Plectranthus tomentosus Benth. Type: "Houte-

niquas”, Thunberg (UPS!, holo.).

Plectranthus madagascariensis (Pers.) Benth., Lab. 37 (1832);

in E. Mey., Comm. 230 (1837); Drege, Zwei Doc. 153, 160

(1843); Benth. in DC., Prodr. 12:68 (1848); Briq. in Pflan-

zenfani. 4,3a ;357 (1897); Compton, FI. Swaz. 67 (1966); Ross,

FI. Natal 305 (1972). P. hirtus Benth., Lab. 38 (1832); in E.

Mey., Comm. 229 (1837); Drege, l.c. 140, 153 (1843); Benth. in

DC., Prodr. 12:68 (1848); Briq.. l.c. 357 (1897); Cooke in FI.

Cap. 5,1:284 (1910), partly; Blake, Contr. Queensl. Herb.

9:39 (1971); Codd in Mitt. Bot. Miinchen 10:248 (1971). Type:

Cape, Masson (BM1). P. mauritianus Boj., Hort. Maurit. 254

(1837). Type: Sieber, FI. Maurit. exs. 152 (K!, P!, G!, Ml).

Coleus madagascariensis (Pers.) A. Chev. in Rev. Bot.

Appliq. 33:338 (1953).

Stems decumbent to procumbent, sparingly

branched, rooting at the nodes, ascending at the ends,

fairly densely to densely tomentose; leaves drying

thin to fairly thick in texture, somewhat sparsely to

densely and shortly appressed tomentose; inflores-

cence usually solitary; corolla varying in length from

7-18 mm long, usually white, rarely mauve or bluish.

Fig. 21.

Found in semi-coastal districts from about Knysna

along the southern and eastern Cape to Zululand

and Swaziland. Also in Mozambique, Madagascar!?)

and Mauritius. It occurs in dry woodland and bush,

usually among rocks or in sandy soil.

Swaziland. — 2631 (Mbabane): Evelyn Baring Bridge, near

Mankaiana (-CA), Compton 28680.

Natal. — 2832 (Mtubatuba): Hluhluwe Game Reserve (-AA),

Ward 2623. 2930 (Pietermaritzburg): near Durban (-DD),

Drege b (K,S). 2931 (Stanger): Ndulindi (-AB), Medley Wood

3980 ; 3030 (Port Shepstone): Umtwalume Waterfall (-AD),

Strey 4408; Oribi Gorge (-CA), Nicholson 1011; 1043; 13 km

N.W. of Port Shepstone (-CB), Codd 9357 ; near Mbeni (-CC),

Codd 9348; Horseshoe farm (-CD), Strey 6421.

Cape. — Without locality, Ekeberg s.n. (SBT); Masson s.n.

(BM). 3126 (Queenstown): near Queenstown (-DD), Galpin

2090a. 3129 (Port St. Johns): between Umzimvubu and Umsi-

kaba Rivers (-BC), Drege a (G, K, MO, P, S); Isilimela Mission

(-CB), Codd 9748. 3227 (Stutterheim): Pirie (-CD), Kuntze

s.n. (K); King William's Town (CD), Sim s.n.; Kei Road (-DA),

Ranger s.n. near Komga (-DB), Flanagan 741; near East

London (-DD), Comins 1509. 3228 (Butterworth): near Butter-

worth (-AC), Pegler 2027; near Kentani (-AD), Pegler 1516;

Qora Mouth (-BC), Meeuse 9687. 3322 (Oudtshoorn): Oute-

niquas (-DC), Thunberg s.n. (UPS). 3323 (Willowmore):

Uniondale (-CA), Schonland 3159; Keurbooms River Nature

Reserve (-CD), Heinecken K86. 3325 (Port Elizabeth): Zwart-

waterpoort (-BB), Burchell 3397 (K); Doom Nek, Zuurberg

(-BC), Story 2373; Longmore Forest Reserve (-CC), Marais

394; Schlechter 2599 (K); “Aloes”, Uitenhage (-CD), /. L.

Drege 3151 ; Addo and Zwartkops River (-DA), Ecklon &

Zeyher 519; Ncanaha (-DB), Archibald 5929 A; near Port

Elizabeth (-DC), Galpin 6466. 3326 (Grahamstown): 3 km

S. of Collingham (-BC), Story 2144; near Grahamstown (BC),

Rogers 4618 (Z); Alexandria (-CB), Galpin 10829; Bushmans

River (-DA), Zeyher 898 (K); 1359; Port Alfred (-DB), Rogers

16627 (Z); Kowie (-DB), Britten 428. 3327 (Peddie): East

London (-BB), Galpin 3234; Munro PS 87.

This entity includes P. hirtus Benth., based on a

Masson specimen from the Cape (holotype in BM)

which, in its typical form, has somewhat trailing stems,

small leaves and whitish flowers. No good reasons

Fig. 21. — Plectranthus madagascariensis var. madagascariensis,

Alexandria District, Cape Province ( Archibald 5929A),

Xl.

can be found for keeping this distinct, even at varietal

level, from P. madagascariensis (Pers.) Benth.,

judging from type material seen in the Jussieu and

general herbarium, Paris. Unfortunately the localities

given on the labels were not recorded at the time but

Blake in Contr. Queensl. Herb. 9:39 (1971) notes

that the plants were collected in Reunion and Mauri-

tius and not in Madagascar.

The flower colour of plants occurring in the

Mascarenes is not known, but this is not a very

important character. Several specimens included in the

above citations have flowers described as various

shades of mauve, lilac or blue, for example: Compton

28680, Edwards 3010, Galpin 2090a, Pegler 1516 and

Meeuse 9687. Some also have leaves rather large for

P. madagascariensis , e.g. Galpin 2090a, Meeuse 9687,

Flanagan 741 and Ranger s.n. (flower colour of last

two not known) and the possibility of hybridization

with P. zatarhendi (large leaves and mauve to purple

flowers) cannot be excluded.

Further material is required of the plant repre-

sented by Story 2373 from Doom Nek, Zuurberg,

Alexandria District. Not only is this outside the

general distribution range, but the flowers are much

larger than usual: nearly 2 cm long, as against the

average for the species of about 1,3 cm. The leaves

are less deeply dentate and of thinner texture than

is expected for P. grandidentatus (which often has

large flowers) and the specimen is now placed tenta-

tively in P. madagascariensis var. madagascariensis.

404

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

(b) var. aliciae Codd, var. nov., a var. madagas-

cariensis sed foliis tenuioribus marginibus profunde

crenato-dentatis, floribus minoribus 5 mm longis

differt.

Type: 3228 (Butterworth): near Kentani (-AD),

1909, Pegler 909 (PRE, holo.).

Stems decumbent, 30 cm long, glandular hispid.

Leaves: petiole 2-4 cm long, glandular-hispid;

blade broadly ovate to subrotund, 2,5-4 cm long,

2,2-4 cm broad, sparingly appressed hirtellous on

both surfaces, gland-dotted below; apex obtuse to

rounded, base truncate; margin crenate-dentate with

3-4 pairs of large, rounded teeth, Inflorescence simple

or occasionally branched at the base, 12-25 cm long;

flowers in 3 — 4-flowered sessile cymes, forming 6-8-

flowered verticillasters; bracts early deciduous. Corolla

5 mm long, white.

Known from only two gatherings in 1903 and 1909

from along streams in damp shady forest near Kentani

by Miss Alice Pegler, after whom the variety is

named. As was her custom when she collected the

same species more than once, she gave the gatherings

the same number.

Cape. — 3228 (Butterworth): near Kentani (-AD), 1903,

Pegler 909; 1909, Pegler 909.

Further material is desired for study. The thin-

textured leaves with 3-4 pairs of large, rounded teeth

along the margins, together with the very small

corolla would indicate that this form is worthy of

separate status.

(c) var. ramosior Benth. in DC., Prodr. 12: 68

(1848). Type: “Macalisberg”, Burke (K ! ; BM!).

Stems erect to decumbent or rarely trailing, sparingly

to freely branched, fairly densely to densely tomentose

with dense short hairs and scattered long hairs;

leaves medium to thick in texture, scabrid to densely

and shortly appressed tomentose; usually several

inflorescences; corolla 0,8-1, 2 cm long, usually

mauve or bluish, rarely white.

Concentrated in central Transvaal, extending to

Swaziland and the inland districts of central and

northern Natal and eastern Cape Province usually

in grass among rocks.

Transvaal. — 2428 (Nylstroom): Sterkrivierdam Nature

Reserve (-BC), Jacobsen 2005. 2430 (Pilgrim’s Rest): Dublin

Mine (-AA), Miller 4265; 21 km S.W. of Ofcolaco (-AB),

Codd 9451. 2527 (Rustenburg): Rustenburg Nature Reserve

(-CB), Jacobsen 717; Castle Gorge (-CD), Meeuse 9254;

Breedtsnek (-CD), Codd 1052; Jacksonstuin (-DA), Van Vuuren

481; 530; near Hartebeestepoort Dam (-DB) Young sub TRV

26950. 2528 (Pretoria): Magaliesberg (-CA), Repton 1328;

Hornsnek (-CA), Hutchinson 2566; Codd 4183; 8630; Union

Buildings (-CA), Van der Wal s.n.; Sixmilespurit (-CC), Van

Niekerk sub TRV 13221 ; Doornkloof (-CC), Pole Evans 1034;

Tygerpoort (-CD), Strey 2823; Trigaardtspoort (-DB), Repton

985; Brucee 96. 2530 (Lydenburg): 14 km E. of Sewefontein

(-CB), Codd 8093; 18 km E. of Sewefontein (-CB), Codd 8107.

2626 (Klerksdorp): near Ventersdorp (-BD), Sutton 662.

2627 (Potchefstroom): near Potchefstroom (-CA), Louw 1310;

1672. 2628 (Johannesburg): Linksfield Ridge (-AA), Gilliland

sub J 27004; Kensington (-AA), Dekker 2; near Heidelberg

(-AD), Leenclertz 8130; Thode A453; Repton 5000; Mogg

22955; Suikerbosrand (-CB), Bredenkamp 585.

O.F.S. — 2627 (Potchefstroom): Parys (-CD), Obermeyer

sub TRV 31683.

Swaziland. — 2631 (Mbabane): 5 km S.W. of Mbabane

(-AC), Codd 9512; Stroma (-AC), Compton 25321 ; Ngotshane

(-AC), Compton 26786; Little Usutu River (-AC), Karsten s.n.

Natal. — 2930 (Pietermaritzburg): near Howick (-AC),

Young 2325; Nagle Dam (-DA), Wells 1281 .

Cape. — 3126 (Queenstown): near Queenstown (-DD),

Galpin 8162. 3227 (Stutterheim); near Cathcart (-AC), Roberts

1734. 3327 (Peddie): Peddie (-AA), Sim 19592.

In its typical form, var. ramosior has several erect

or ascending stems arising from a perennial rootstock.

It is common on the rocky wooded and grassy

ridges of the Witwatersrand and Magaliesberg which

were subjected to frequent burning in earlier days.

Complete specimens frequently show the burnt base

of the plant from which the stems arise annually.

The stems are usually branched and each branch

bears a simple inflorescence, resulting in a number of

inflorescences per plant. However, herbarium speci-

mens do not always show the base of the plant,

making the distinction between this variety and var.

madagascariensis less clear. There is also a tendency

for some plants, especially in the eastern Transvaal,

to have straggly stems and these are difficult to

allocate with any degree of certainty. In the eastern

Cape the distinction is also not very clear. For this

reason the group is treated as a variety of P. madagas-

cariensis whereas its affinity seems to lie near to

P. zatarhendi var. zatarhendi. In fact, the main

difference between these two is in size of leaf. Thus

some specimens from Natal are intermediate between

the two and the distinction is sometimes rather

arbitrary. Var. ramosoir tends to have mauve flowers

(as does P. zatarhendi var. zatarhendi) while var.

madagascariensis usually has white flowers, but this

is not a reliable basis for separating the varieties as

flower colour can vary in both.

Bentham bases var. ramosior on “regione Macalis-

berg, Burke”. Two Burke specimens are at Kew

labelled “Sand River” and “Vaal and Mooy Rivers”

respectively. The latter is selected as the lectotype.

20. Plectranthus mutabilis Codd, sp. nov., affinis

P. madagascariensis (Pers.) Benth., sed foliis profunde

crenatis chartaceis vel membranaceis, amplitudine

admodum variabilibus usque ad 5x5 cm differt.

Herba perennis, semi-succulenta; caules graciles,

procumbentes, usque ad 40 cm longi, parce ramosi,

breviter tomentosi vel villosi. Folia submembranacea,

petiolata; petiolus 1,4-3 cm longus; lamina late

ovata vel subrotunda, 1,5-5 cm longa, pariter lata,

supra sparse vel dense strigosa, subtus sparse strigosa

vel dense tomentosa, punctata vel impunctata, apice

obtusa vel rotundata, basi truncata vel cordata,

margine profundo crenata. Inflorescentia erecta, raro

ramosa, 10-25 cm longa: rhachis minute glanduloso-

hispidulus; bracteae subrotundatae, 1 ,5-2 mm longae,

mox deciduae. Verticillastri 1-2 cm distantes, 6-12-

floribus; pedicelli 2-3 mm longi, glanduloso-his-

spiduli. Calyx 2-4 mm longus, basaliter gibbosus,

glanduloso-scabridus; lobus posticus ovatus, sub-

erectus, acutus vel obtusus, 2 mm longus; lobus

anticus subaequaliter 4-dentatus; dentes lineari-

lanceolati, 2,5 mm longi. Corolla azurea vel lilacina,

8-12 mm longa, pubescens; tubus 4-6 mm longus,

deflexus, basi 2 mm late, fauce 3 mm late; labium

posticum erectum vel recurvum 3 mm longum,

4-lobatum; labium anticum cymbiforme, 4-5 mm

longum. Stamina 4, 4-6 mm longa, inclusa vel breviter

exserta, filamentis liberis. Stylus stamina breviter

excedens.

Type. — Transvaal, Pietersburg District, farm Bulbul

at eastern end of Blouberg, Codd 7953 (PRE, holo.).

Perennial aromatic semi-succulent herb; stems

slender, procumbent to hanging from rock ledges,

up to 40 cm long, sparingly branched, shortly tomen-

tose to villous-tomentose often with very long multi-

cellular hairs intermingled with short hairs, especially

at the nodes. Leaves softly semi-succulent, drying

fairly thin in texture; petiole 1,4-3 cm long, shortly

pubescent to villous; blade broadly ovate to sub-

rotund, 1,5-5 cm long and equally broad, sparsely

to densely strigose above, sparsely strigose on the

nerves below to densely woolly tomentose, with or

L. E. CODD

405

without reddish or yellowish sessile gland-dots;

apex obtuse to rounded; base truncate to cordate;

margin deeply scalloped with 3-5 pairs of large

rounded teeth. Inflorescence terminal, usually simple,

occasionally with a pair of branches near the base,

10-25 cm long; rhachis minutely shortly glandular-

hispidulous; bracts subrotund 1,5-2 mm long, early

deciduous, before flowering stage. Flowers in sessile,

3- 6-flowered cymes, forming 6-12-flowered verti-

cillasters; verticillasters 1-2 cm apart; pedicels

2-3 mm long, glandular-puberulous Calyx 2 mm

long at flowering enlarging to 4 mm long in fruit with

a distinctly gibbous base, glandular-scabrid with

yellowish or red gland-dots or sometimes without

gland-dots; upper lip suberect, ovate, obtuse to

acute at the apex, up to 2 mm long; lower lip ^equally

4- toothed up to 2,5 mm long, teeth linear-lanceolate,

the lower pair slightly the longer. Corolla blue,

purple-blue or lilac, 8-12 mm long, pubescent and

usually, not always, with coloured gland-dots on the

upper and lower lips; tube 4-6 mm long, 2 mm deep

at the base expanding to 3 mm deep at the throat,

distinctly bent below the middle; upper lip erect to

strongly bent back, 3 mm long and equally broad,

deeply emarginate at the apex and with 2 lateral

ear-like lobes; lower lip boat-shaped, horizontal or

slightly upcurved, 4-5 mm long. Stamens free at the

base, curved and enclosed in the lower lip or slightly

exserted, 4-6 mm long. Style slightly exceeding the

stamens. Fig. 22.

Fig. 22. — Plectranthus mutabilis, Blouberg, northern Transvaal

(holotype: Codd 7953 in PRE), x^.

On rocky hillsides in the shade of forest or in

exposed places, in rock crevices or hanging from

rock ledges, found mainly on the Blouberg and

Soutpansberg in northern Transvaal, but extending

southwards to the Pretoria District.

Transvaal. — 2229 (Messina): Soutpansberg, Dandy Farm

(-DC), Meeuse 10220; Soutpansberg, 10 km W. of main road

(-DD), Rodin 4011; Punch Bowl Hotel (-DD), Codd 8340;

Wylliespoort (-DD), Bruce 66A. 2328 (Baltimore): Blouberg,

above Leipzig Mission (-BB), Codd 8692; Strey & Schlieben

8473; Blouberg, on route to Trig. Beacon, Codd 8753; Blouberg,

at Trig. Beacon, Codd & Dyer 9040. 2329 (Pietersburg): farm

Bulbul at eastern end of Blouberg (-AA), Codd 7953. 2428

(Nylstroom): Sterkrivierdam Nature Reserve (-BC), Jacobsen

2005. 2430 (Pilgrim’s Rest): 10 km S.W. of Penge Mine (-AD),

Codd 8789. 2528 (Pretoria): Trigaardtspoort (-DB), Repton

985; Bruce 96. 2529 (Witbank): Loskopdam, Kloppersloop

(-DA), Theron 1370.

This is a variable species and not always clearly

recognizable, though its typical form suggests that

the entity is worthy of specific rank. The great varia-

tion in leaf size from 1, 5x1,5 cm to over 5x5 cm

would lead to difficulty in allocating it to either of

the two large species, P. zatarhendi or P. mada-

gascariensis. From these two species it can be dis-

tinguished by the usually deep crenate scalloping of

the leaf margin with 3-5 pairs of large rounded teeth

and the usually thin texture. However, the texture and

size of teeth vary according to the exposure of the

situation. There is also considerable variation in the

pubescence and degree of gland-dotting on the under-

sides of the leaves, the calyx and corolla. The flower

colour shows some variation from deep blue to bluish-

mauve or bluish-white.

21. Plectranthus psammophilus Codd, sp. nov.,

affinis P. madagascariensis (Pers.) Benth., sed corollis

parvulis differt.

Herba perennis, semi-succulenta; caules graciles,

decumbentes vel procumbentes, usque ad 50 cm

longi, ramosi, glanduloso-hirsuti, punctati. Folia sub-

membranacea, petiolata; petiolus 1-2 cm longus;

lamina deltoideo-ovata, 2-4 cm longa, pariter lata,

supra glanduloso-strigosa, subtus sparse glanduloso-

strigosa copiose rufo-punctata, apice obtusa, basi

truncata, margine obscure crenato-dentata. Inflores-

centia erecta, gracilis, raro ramosa, condensata, 10-20

cm longa; rhachis glanduloso-hispidulus; bracteae

late obovatae, 1,5-2 mm longae, persistentes. Verti-

cillastri 1-4 mm distantes, 6-12-floribus; pedicelli

1 ,5 — 3 mm longi, breviter hispiduli. Calyx 1 ,5-3 mm

longus, basalitergibbosus,glanduloso-scabridus, punc-

tatus; lobus posticus ovato-rotundus, ± horizontalis,

1.5 mm longus; lobus anticus subaequaliter 4-den-

tatus, dentes lineari-lanceolati, 1-1,5 mm longi.

Corolla lilacina, 5 mm longa, pubescens, rubropunc-

tata; tubus horizontalis, 2,5 mm longus, prope

basin expansus, cylindricus, 1 mm diam.; labium

posticum erectum, 1 mm longum, 4-lobatum; labium

anticum cymbiforme, 2,5 mm longum. Stamina 4,

2.5 mm longa, inclusa vel breviter exserta, filamentis

liberis. Stylus stamina breviter excedens.

Type. — Natal, Ubombo District, Makatini Flats

and cult. Natal Herbarium garden, Strey 5779 (PRE,

holo. ; NH).

Perennial aromatic semi-succulent herb; stems slen-

der, decumbent to procumbent, up to 50 cm long,

4-angled, branching, glandular-hirsute with long mul-

ticellular hairs, shorter gland-tipped hairs and sessile

red gland-dots, especially at the nodes. Leaves soft,

drying fairly thin in texture; petiole 1-2 cm long

glandular-hirsute; blade ovate-triangular, 2-4 cm

long and equally broad, fairly densely glandular-

strigose above, more sparingly and mainly on the

nerves below, the under surface freely dotted with

406

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

reddish-brown gland-dots; apex obtuse; base truncate;

margin obscurely crenate-dentate with 5-7 pairs of

shallow teeth. Inflorescence slender, fairly condensed,

terminal and on lateral branches, simple or with a

pair of branches at the base, 10-20 cm long; rhachis

glandular-hispidulous; bracts broadly obovate, 1,5-2

mm long, persisting until the flowering stage. Flowers

in sessile, 3-6-flowered cymes forming 6-12-flowered

verticillasters; verticillasters 1-4 mm apart; pedicels

1,5-3 mm long, shortly hispidulous. Calyx 1,5 mm

long at flowering enlarging to 3 mm long in fruit

with a distinct gibbous base, glandular-scabrid and

gland-dotted; upper lip more or less horizontal,

ovate-rotund, rounded at the apex, up to 1,5 mm long;

lower lip dz equally 4-toothed, the lower pair slightly

the longer, 1-1,5 mm long, teeth linear-lanceolate.

Corolla blue-mauve, 5 mm long, with a patch of

pubescence and red gland-dots on the upper and lower

lips; tube horizontal, 2,5 mm long, expanding in the

calyx to 1 mm deep and parallel-sided; upper lip

erect, 1 mm long and broad, emarginate at the apex

and with 2 small lateral ear-like lobes; lower lip

boat-shaped, horizontal, 2,5 mm long. Stamens free

at the base, curved and enclosed in the lower lip or

slightly exserted, 2,5 mm long. Style exserted by

about 0,5 mm beyond the stamens. Fig. 23.

Fig. 23. — Plectranthus psammophilus, Makatini Flats, northern

Natal (holotype: Strey 5779, PRE), xF

Known from several gatherings in northern Zulu-

land where it grows at forest margins on sandy

flats.

Natal. — 2632 (Bela Vista): Ndumu Game Reserve (-CD)’

Ward 3100. 2732 (Ubornbo): Makatini Flats (-AD) and cult.

Natal Herbarium garden, Strey 5779; Wearne s.n.; Sivingu-

vungu (-DA), Vahrmeijer & Drijfhout s.n.

Allied to P. madagascariensis, this is a distinctive

plant with slender, branched stems, slender, fairly

dense inflorescences and extremely small flowers up

to 5 mm long. From P. madagascariensis var. aliciae

it differs in the more compact inflorescence and the

finely toothed leaves.

Sect. Plectranthus.

Sect. Germanea (Lam.) Benth., Lab. 32 (1832); in

DC., Prodr. 12:62 (1848). — sect. Coleoides Benth.,

1 l.cc., partly.

In this section the bracts persist to beyond the

flowering stage; the calyx, although enlarging with

maturity and somewhat oblique at the base, is not

markedly gibbous ventrally; and the flowers are in

few-flowered pedunculate cymes or in 1-3-flowered

sessile cymes (in contrast to sect. Coleoides, p. 395).

No advantage can be seen in attempting to maintain

a number of formal series in this section, as was done

by Briquet in Pflanzenfam. 4, 3a:354 (1897). However,

the corolla varies a good deal in South African species

and a separation can be made into those in which the

corolla is abruptly enlarged to saccate or spurred at the

base (nos. 22-36 below) and those in which the corolla

tube does not immediately widen on emerging from

the calyx and is either straight (nos. 37 and 38) or

somewhat sigmoid (nos. 39 and 40).

A deviation unique in the genus is shown by

P. zuluensis in which the upper two stamens are

abortive, resulting in flowers with 2 fertile stamens and

2 staminodes. Otherwise, species are delimited on

variation in corolla shape, stamen length, leaf size,

shape and margin, and pubescence of leaves and

corolla. Flower colour is usually not discernible in

dried specimens but is helpful in some cases when

fresh flowers or collectors’ notes are available.

The minute gland-dots on the underside of the

leaves and often on calyx and corolla are useful in

separating certain superficially similar species, espe-

cially when faced with sterile or scrappy specimens.

Thus P. fruticosus (honey-coloured droplets) can be

separated from P. grallatus and P. rubropunctatus

(reddish to reddish-brown, flattened gland-dots). Si-

milarly, among the smaller-leaved species, P. elegan-

tulus and P. oertendahlii have colourless to honey-

coloured dots like P. fruticosus, while P. verticillatus,

P. strigosus and P. purpuratus are copiously dotted

with red gland-dots.

Species in sect. Plectranthus are distributed from

South Africa to tropical East and West Africa and

India. In Southern Africa they occur in the eastern

high rainfall part of the territory, with a concen-

tration in Natal, being absent from the northern

Cape and South West Africa.

The species dealt with are listed below. In contrast

to sect. Coleoides, little difficulty is experienced in

delimiting species in the typical section:

22. P. verticillatus (L.f.) Druce

23. P. strigosus Benth.

24. P. purpuratus Harv.

25. P. oertendahlii Th. Fries jun.

26. P. elegantulus Briq.

27. P. cilliatus E. Mey. ex Benth.

28. P. fruticosus L’Herit.

29. P. grallatus Briq.

30. P. rubropunctatus Codd

3 1 . P. rehmannii Guerke

32. P. swynnertonii S. Moore

33. P. dolichopodus Briq.

34. P. zuluensis T. Cooke

L. E. CODD

407

35. P. saccatus Benth.

(a) var. saccatus

(b) var. longitubus Codd

36. P. hilliardiae Codd

37. P. ambiguus (Bol.) Codd

38. P. ecklonii Benth.

39. P. petiolaris E. Mey. ex Benth.

40. P. laxiflorus Benth.

22. Plectranthus verticillatus (X./) Druce in Rep.

Bot. Exch. Cl. Brit. Isles 1916: 640 (1917); Ross,

FI. Natal 305 (1972). Type: Cape (“India, Montin” —

see note below) (LINN 749.4, holo.).

Ocimum verticillatum L.f., Suppl. 276 (1781), as “Ocymum";

Willd., Sp. PI. 3:163 (1800). O. racemosum Thunb., Prodr.

96(1800), as "Ocymum”; FI. Cap. ed. Schult. 448(1823). Type:

Cape, "Houteniquas”, Thunberg (UPS!, holo.; SBT!).

Plectranthus thunbergii Benth., Lab. 37(1832); in E. Mey.,

Comm. 229(1837); Drege, Zwei Doc. 125, 147(1843); Benth.

in DC., Prodr. 12:67(1848); Schinz in Mem. Herb. Boiss.

10:60(1895); Briq. in Pflanzenfam. 4,3a; 357(1897); Cooke in

FI. Cap. 5,1:280(1910); Codd in Mitt. Bot. Miinchen 10:247

(1971); nom. illegit. Type: based on Ocimum verticillatum L.f.

and O. racemosum Thunb. P. nummularius Briq. in Bull Herb.

Boiss. ser. 2,3: 1072 (1903); Cooke l.c. 284 (1910), partly; Letty,

Wild Flow. Transv. 289, t. 144: 2 (1962); Compton, FI. Swaz.

67, 158 (1966); Codd, l.c. 247 (1971); Ross, FI. Natal 305 (1972).

Type: Natal, Camperdown, Rehmann 7702 (Z!, holo.).

Perennial semi-succulent aromatic herb; stems pro-

strate to ascending, brittle, softly succulent, 4-angled,

up to 120 cm long, often rising to 25 cm above the

ground, rooting at the nodes, subglabrous to shortly

antrorse strigulose, with short multicellular hairs

(longer tufts at the nodes), and usually sessile red

gland-dots. Leaves softly to distinctly succulent, often

glossy above and purple-tinged below; petiole 6-30

mm long, finely strigulose to subglabrous; blade

ovate to rotund, 1,6-4 cm long, 1,2-4 cm broad,

drying thin-textured, subglabrous to strigulose on

both surfaces, usually thinly above and more densely

on the nerves below, freely and conspicuously dotted

with red to brownish sessile gland-dots; apex acute to

rounded; base truncate to cuneate; margin crenate-

dentate to shallowly crenate with 3-6 pairs of teeth.

Inflorescence terminal, simple or with a pair of

branches at the base, 4-22 cm long (usually about

10-15 cm); rhachis subglabrous to finely (occasionally

fairly densely) strigulose or finely glandular-tomen-

tulose, often red gland-dotted; bracts ovate-lanceolate

to linear, 2-4 mm long, usually ciliate and red gland-

dotted, persisting beyond the fruiting stage. Flowers

in sessile 1-3-flowered cymes forming 2-6-flowered

verticillasters 6-15 mm apart; pedicels 3-5 mm long,

finely strigulose or glandular-puberulous. Calyx 2,5

mm long at flowering enlarging to 7 mm long in fruit,

purple-tinged, with short appressed hairs and red

gland-dots; upper lip suberect, ovate, acute, 2-2,5

mm long, not decurrent on the tube; lower lip 4-

toothed up to 3 mm long, teeth linear-subulate, the

lower pair the longer, 2mm long, the lateral 1 ,5 mm

long. Corolla white to pale mauve with a few mauve

spots on the upper lip or freely speckled with purplish

spots, pubescent and with scattered red gland-dots;

tube 4-6 mm long, expanding and slightly deflexed

at or near the base forming a swollen to slightly

saccate base 1,75-2 mm deep, scarcely narrowing to

the throat; upper lip erect to bent backwards, 5-8

mm long, 3-4 mm broad, apex emarginate and with 2

lateral ear-like lobes; lower lip shallowly boat-shaped,

5-7 mm long, horizontal. Stamens free at the base,

didynamous, curved within or slightly exceeding the

lower lip, lower pair 6-7 mm long, upper pair about

5 mm long. Style coinciding more or less with the

upper stamens. Fig. 24.

Occurs in forest margins, scrub forest, dry wood-

land and semi-shady, stony places from the Knysna

area through the semi-coastal parts of the eastern

Cape Province, eastern Natal, Swaziland and eastern

Transvaal to the Soutpansberg, extending to southern

Mozambique.

Transvaal. — 2229 (Waterpoort): 18 km W. of Mountain

Inn (-DD), Meeuse 10239; Soutpansberg, on crest 10 km W. of

Main road, Rodin 4010. 2329 (Pietersburg): Louis Trichardt

(-BB), Letty 255; Breyer sub TRV 22726. 2330 (Tzaneen):

Elim ( AA), Obermeyer 727 ; Westfalia Estate (-CA), Scheepers

640; Wolkberg (-CC), Meeuse 9922. 2331 (Phalaborwa):

Kruger National Park, Boulders (-CB), Van der Schijff 2546.

2430 (Pilgrim's Rest): Shiluvane (-AB), Junodsub TRV 10216;

Mariepskop (-DB), Van Son sub TRV 31562; Blyde River

Nature Reserve (-DB), Van der Schijff 6136. 2530(Lydenburg):

Kemps Heights (-AB), Codd 8197; 8320; Rosehaugh (-BD),

Smuts 72; 77; Nelspruit Botanic Garden (-BD), Buitendag 466.

2531 (Komatipoort): Kruger National Park, Numbi (-AA),

Van der Schijff 2664; Bukwenene (-AD?), Van der Schijff &

Marais 3737; near Malelane (-AD), Codd 6100; Barberton

(-CC), Williams sub TRV 8241; Cythna Letty Nature Reserve

(-CC), M idler 2259. 2730 (Vryheid): Piet Retief, Leipoldt s.n.

Swaziland. — 2631 (Mbabane): near Komati Bridge (-AA),

Compton 28831 ; 9 km S. of Forbes Reef (-AA), Bruce 308;

Manzini (-AD), Compton 28805; Stegi (-BD), Compton 26583:

Mankaiana (-CA), Compton 28678; Hlatikulu (-DC), Stewart

sub TRV 10384; Compton 28764.

Natal. — 2632 (Bela Vista): Ndumu Game Reserve (-CD)

Ward 3099; Tinley 441 ; Oat ley D5; Kosi system. Lake Shlangi

(-DD), Vahrmeijer & Tiilken 943. 2732 (Ubombo): 32 km

from Jozini to Bazwana (-AC), Strey 5294; Jozini Dam (-AC),

Repton 5978. 2831 (Nkandla): 5 km N. of Hlabisa (-BB), Codd

1987 ; Nhlwati (-BB), Ward 3212; Eshowe, Mandawe Valley

(-CD), Gerstner 3396; Umhlatuzi Valley (-DC), Lawn 2069

(NH). 2832 (Mtubatuba): Hluhluwe Game Reserve (-AA),

Ward 2624; Hitchins 57; Dukuduku (-AC), Moll 2743; near

Richards Bay (-CC), Oatley 35. 2929 (Underberg): Estcourt

(-BB), Acoclcs 10239; 11344. 2930 (Pietermaritzburg): Camper-

down (-DA), Rehmann 7702 (Z); Isipingo (-DD), Ward 136;

5299. 2931 (Stanger): Zinkwazi (-AD), Moll 2910; 6 km N. of

Umhlanga Rocks (-CA), Watmough 403; Brighton Beach

(-CC), Forbes 509 (NH). 3030 (Port Shepstone): Horeshoe

Farm (-CC), Strey 6155; Uvongo River (-CD), Whellan 1092;

Wichmann's farm (-CD), Strey 11063.

Cape. — Without precise locality: Sparrmann s.n. (S); Ver-

ve cmx s.n. (G, P); Zeyher 274 (M, P). 3029 (Kokstad): Clydesdale

Tyson sub Herb. Norm. 1296. 3129 (Port St. Johns): Lusikisiki

(-BC), Galpin 10956; Umgazi Mouth (-CB), Wells 3456;

Mqanduli District, near sea (-CC), Theron 1507. 3227 (Stutter-

heim): Stutterheim Commonage (-CB), Acocks 9530; near

Komga (-DB), Flanagan 1722. 3228 (Butterworth): between

Bashee River and Morley (-BB), Drege b (G, K, P, S); Qora

Mouth (-BC), Flanagan 1210. 3322 (Oudtshoorn): Outeniquas

(-CD?), Thunberg s.n. (UPS, SBT). 3325 (Port Elizabeth):

Uitenhage, Zuurberg (-BD), Fries Norlindh & Weimark 431

(LD); Zuurberg Sanatorium (-BD?), Long 1256; Zwarkops-

rivier (-CB?), Ecklon (P); Zeyher (P); Ecklon & Zeyher (S);

Winterhoek Mts. (-CB), Fries, Norlindh & Weimark 945 (LD);

near Uitenhage (-CD), Schlechter 2501 ( K); near Port Elizabeth

(-DC), Fries , Norlindh & Weimarck 431 (LD). 3326 (Grahams-

town): Coldspring (-AD), Gone sub TRV 2527 1 ; near Grahams-

town (-BC), Burchell 3579 (K); Britten 461: Zuurberg Pass

(-CB), Archibald 5849 b; Alexandria Forest (-CB), Galpin

10672. 3327 (Peddie): East London (-BB), Nanni 133. 3422

(Mossel Bay): Goukamma (as “Doukamma”) (-BB), Drege a

(K, P). 3423 (Knysna): Knysna (-AA), Breyer sub TRV 23642;

Plettenberg Bay (-AB), Rogers 27114 (Z); Keurbooms River

(-AB), Burchell 5150 (K).

The species exhibits a good deal of variation in

pubescence, leaf-shape and flower colour, and cer-

tain of these differences are associated with geographic

distribution. However, it was not found possible to

subdivide the material satisfactorily into infraspecific

groups.

Towards the south of the range, in the Cape and

Natal coastal regions as far north as northern Zulu-

land, the leaves tend to be ovate, somewhat cuneate

at the base with crenate-dentate margins, while the

flowers are white to pale mauve, with pale mauve

vertical marks on the upper lip. In the Transvaal the

plants are, in general, more robust, with larger,

rounder leaves which are shallowly crenate, while the

flowers are white, freely speckled with purple spots

408

PLECTRANTHVS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 24. — Plectranthus verticillatus, Ngqeleni, Cape Province (Code/ 9285), X I.

on the upper and lower lips. There is a considerable

amount of transition in the more inland Natal locali-

ties and in Swaziland.

The pubescence is more marked in some of the more

northern plants, especially in Swaziland, where some

specimens may be confused with P. strigosus, but

variation in degree of pubescence is found throughout

the distribution range. The semi-succulent leaves arc

often dark green and glossy and, as the plants are

relatively drought- and disease-resistant, they are

popular for ground cover in shade and for hanging

baskets.

Regarding the typification of the epithet Ocimum

verticillatum, our Liaison Officer at Kew in 1961

(Dr D. J. B. Killick) reported that the specimen

No. 749.4 in the Linnaean Herbarium is conspecific

with the material cited above. It is annotated in very

small letters “C.B.S.” (writing not recognisable) and

in pencil in Smith’s writing: “verticillatum”. Mr

Sandwith, who was consulted in October 1961,

considered that this specimen may be accepted as the

one dealt with by Linn, fil . In favour of this view is the

statement by Juel, PI. Thunb. p. 7, last paragraph,

where Linnaeus is quoted as writing to Thunberg

that he has named Montin’s Cape plants as well as

those which Montin had received from Thunberg.

The Linnaean specimen 749.4 could have been one

of the latter, in fact, a similar specimen in the Thun-

berg Herbarium, Uppsala, is the type of Ocimum

racemosum Thunb.

L. E. CODD

409

Willdenow altered the locality, cited as India by

Linn, fil, to “C.B.S.”, but does not state a reason for

doing so. The species as interpreted at present does

not extend to India but could, of course, have been

cultivated there. Montin did not collect in either of

these regions and it is most likely that the plant

came from the Cape and that the information on its

country of origin became confused.

The type of P. nummularius Briq., Rehmann 7702

from Camperdown (in Z), is a scrappy specimen and

it is not certain whether it belongs to the southern or

the northern form of the species. The leaf shape tends

to resemble the latter.

P. verticillatus is most closely related to P. stri-

gosus and P. purpuratus, species with denser pube-

scence than P. verticillatus. Some specimens may be

difficult to classify on pubescence alone, but there are

good floral characters for separating them. In P.

verticillatus the corolla is larger and the tube is more

or less parallel sided while the stamens are 5-7 mm

long; in P. strigosus and P. purpuratus the corolla is

smaller and the tube has a distinct constriction near

the throat while the stamens are much shorter,

reaching only 1-3 mm in length.

The distinctions between P. verticillatus and P.

oertendahlii are discussed under the latter (p. 411).

Although P. verticillatus is often confused with P.

madagascariensis {=P. hirtus) in herbaria, the two

are not closely related. Both have trailing stems and

small, ovate to subrotund leaves. P. madagascariensis

is normally more conspicuously pubescent than P.

verticillatus and, although sterile specimens are some-

times superficially similar, the conspicuous red

gland-dots on the underside of the leaves of P.

verticillatus can be used to separate the two species.

Flowering specimens should present no difficulty as

the bracts in P. madagascariensis are shed before the

flowers open (persistent in P. verticillatus), the corolla

lacks red gland-dots and the mature calyx has a diffe-

rent shape, being distinctly more declinate and

saccate on the underside.

Among the specimens cited under P. nummularius

by Cooke in Flora Capensis, Wood 3980 is P. mada-

gascariensis while the Kew specimen of Wood 3981 is a

mixture of P. verticillatus and Aeollanthus parvifolius.

Chromosome number; 2n=28 (De Wet, 1958).

23. Plectranthus strigosus Benth. in E. Mey.,

Comm. 229 (1837); Drege, Zwei. Pfl. Doc. 153

(1843); Benth. in DC., Prodr. 12:68 (1848); Briq. in

Pflanzenfam. 4,3a:357 (1897); Cooke in FI. Cap.

5,1:280 (1910); Compton, FI. Swaz. 67 (1966); Codd

in Mitt. Bot. Miinchen 10:247 (1971). Lectotype:

Cape, Olifantshoek Forest (Alexandria), Ecklon (K!)

(see note below).

P. strigosus var. lucidus Benth. in DC., Prodr. 12:68 (1848);

Cooke, l.c. 280 (1910). Type: Cape, Bathurst, Burchett 3924

(K!, holo.). P. parviflorus Guerke in Kuntze, Rev. Gen. 3, 2:

261 (1898); Cooke l.c. 281 (1910); non Willd. (1806). Type:

East London, Kuntze s.n. (NY, holo.; PRE, photo.). P. kunt-

zeanus Domin., Biblioth. Bot. 89: 1118 (1928). Type: as for

P. parviflorus Guerke.

Perennial semi-succulent aromatic herb; stems de-

cumbent to ascending, brittle, softly succulent, 4-

angled, up to 30 cm long, forming small mats 30 cm

across and up to 10-25 cm tall, rusty-hispid with

multicellular purplish hairs and red gland-dots. Leaves

softly succulent, fairly thick textured, purple-tinged

below; petiole 5-15 mm long, rusty-hispid like the

stems; blade broadly ovate to subrotund, 1,3-3, 5

cm long, 0,8-3 cm broad, subglabrous to strigose on

both surfaces, denser on the nerves below with long

grey to rusty multicellular hairs and numerous red

gland-dots on the under surface, apex obtuse to roun-

ded; base truncate to shortly cuneate; margin ob-

scurely crenate with 4-6 pairs of shallow teeth, ciliate.

Inflorescence terminal, usually simple, occasionally

with a pair of branches near the base, 4-15 cm long

(usually about 6-10 cm); rhachis densely glandular-

hispidulous, often with red gland-dots; bracts ovate to

lanceolate, 2-3 mm long, usually ciliate and red

gland-dotted, persisting beyond the fruiting stage.

Flowers in sessile 1-3-flowered cymes forming 2-6-

flowered verticillasters; verticillasters 5-12 mm apart;

pedicels 3-5 mm long, densely glandular-puberulous.

Calyx 2,5 mm long at flowering enlarging to 6 mm

long in fruit purple-tinged, sparingly strigulose and

with red gland-dots; upper lip suberect, ovate, acute,

scarcely decurrent on the tube, up to 2 mm long;

lower lip 4-toothed, 2-3 mm long, teeth linear subu-

late, the lower pair the longer, up to 2 mm long.

Corolla whitish to mauve with a few darker markings

in vertical rows on the upper lip, pubescent and with

red gland-dots; tube 3-5 mm long, expanding in the

calyx to form a swollen base 1 ,5 mm deep, narrowing

to 1 mm deep at the middle or near the throat; upper

lip erect 4-5 mm long, 3 mm broad, apex emarginate

and with two rounded lateral lobes projecting for-

ward; lower lip shallowly boat-shaped 3-4 mm long,

horizontal. Stamens free at the base, didynamous,

1,5-3 mm long, scarcely visible. Style coinciding

with the upper stamens. Fig. 25.

A species with a disjunct distribution, being found

from the Alexandria to Lusikisiki Districts in the

Cape and again in Swaziland, occurring in shady,

rocky places and in scrub forest.

Swaziland. — 2631 (Mbabane): near Forbes Reef (-AA),

Compton 26001 ; Hilliard & Burtt 3565; Mbabane (-AC),

Ihlenfeldt 2466; 3 km N. of Mbabane (-AC), Schlieben 9498;

Mpalaleni (-AC), Compton 32207; Mankaiana (-CA), Compton

27725.

Cape. — 3128 (Umtata): near Old Morley (-DD), Codd 9269.

3129 (Port St. Johns): between Umzimvubu and Umsikaba

Rivers (-BC), Drege 4779 c (K). 3228 (Butterworth): near

Kentani (-AD), Pegler 910. 3326 (Grahamstown): near Gra-

hamstown (-BC), Galpin 278; Blaauwkrantz (-BC), Britten 861 ;

Marloth 10922; Alexandria (Olifants Hoek) Forest (-CB),

Ecklon (K); Kariega Mouth (-DA), Britten 2370; Acocks

18351; Kowie West (-DB), Britten 742. 3327 (Peddie): Wool-

ridge (-AB), Bavliss 3349; East London (-BB), C. A. Smith

3674.

In addition to the above, two Drege specimens in

Kew, annotated as coming from “Zuurbergen” and

“Van Stadens River” respectively, and cited under

P. hirtus by Bentham in E. Mey., Comm. 229 (1837)

and Cooke in FI. Cap. 5,1: 284 (1910), appear to

belong rather in P. strigosus. The two species are

often confused but in P. strigosus the bracts persist

beyond the fruiting stage while in P. madagascariensis

(=P. hirtus) they are shed before the flowers open.

It may be accepted from Bentham’s method of

presentation that he based the species on Ecklon

material as well as the specimen cited as being collected

by Drege “ad Cataractam magnam” between the

Umzimvubu and Umsikaba Rivers, probably what

we now know as the Magwa Falls. This gathering,

represented in Kew by the specimen Drege 4779c,

is a rather odd specimen with leaves larger than any

of the others cited above and its identity is not

altogether certain in relation to our modern mateiial.

It looks a bit like P. ciliatus, a species known from

Magwa Falls, but is considered to belong rather

to our concept of P. strigosus. In view of the unusual

character of this specimen, it is preferred to select the

Ecklon specimen ex Herb. Benth. from Olifants

Hoek Forest (our present Alexandria Forest) as the

lectotype. It is mounted on the same sheet as Drege

4779c.

410

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Chromosome number 2n 28 (De Wet, 1958).

It is strange that between the two areas where

P. strigosus occurs one finds the closely related species

P. purpuratus (i.e. around Camperdown in Natal).

In corolla characters the two are very similar and the

main distinction is in the dense, velvety pubescence

of the stems and leaves of P. purpuratus , which

species also forms more compact plants with shorter

stems than P. strigosus. No intermediates between the

two species are known and so the two are maintained

as distinct.

24. Plectranthus purpuratus Harr., Thes. Cap. 1 : 53,

t. 83 (1859); Cooke in FI. Cap. 5,1: 282 (1910);

Codd in Mitt. Bot. Miinchen 10: 247 (1971); Ross,

FI. Natal 305 (1972). Type: a plant cultivated at Kew,

without date, from seed sent by Mr. Vause (“Vance”)

of Port Natal (TCD, holo; K!).

Perennial succulent herb forming small mats up

to 20 cm tall ; stems decumbent to ascending, obscurely

4-angled, brittle with short internodes, densely velvety

tomentulose and red gland-dotted. Leaves succulent,

drying thick-textured, purple below; petioles 3-8 mm

long, velvety tomentulose; blade broadly ovate,

subrotund or broadly obovate 5-15 mm long and

equally broad, densely and finely appressed grey

velvety on both surfaces nerves indistinct above,

copiously red gland-dotted below; apex obtuse to

rounded, base obtuse to shortly cuneate; margin

obscurely crenate to subentire with 3 or 4 pairs of

shallow teeth. Inflorescence terminal, simple or

occasionally branched below, 4-10 cm long, rhachis

finely glandular-velvety with occasional red gland-

dots; bracts ovate-lanceolate 1-1,5 mm long, per-

sisting beyond the fruiting stage. Flowers in sessile

1-2 (rarely 3)-flowered cymes forming 2-4-flowered

verticillasters; verticillasters 4-8 mm apart; pedicels

3-4 mm long densely glandular-puberulous. Calyx

2,5 mm long at flowering enlarging to 5-6 mm long

in fruit, puberulous and freely red gland-dotted;

upper lip suberect broadly ovate, acute, decurrent on

the tube, up to 2 mm long; lower lip 4-toothed,

2,5 mm long, teeth deltoid-subulate, subequal, the

lower pair slightly longer. Corolla white with a few

blue-mauve marks in the throat, pubescent and with

orange gland-dots; tube 3 mm long, expanding in the

calyx to form a swollen base 2 mm deep, narrowing

to I mm deep about the middle and expanding again

slightly towards the throat; upper lip erect, 3 mm

long and equally broad, emarginate and with

2 rounded lateral lobes projecting forward; lower lip

slightly concave, 3 mm long, horizontal. Stamens

L. E. CODD

41 1

free at the base, didynamous, obscured by the lateral

corolla lobes, 0,5 and 1,5 mm long. Style coinciding

with the longer stamens, stigma deeply bifid. Fig. 26.

Fig. 26. — Plectranthus purpuratus, Nchanga, Natal ( Eshuis

s.n.), X 1 .

Found in dry, rocky places and on krantzes in a

restricted area between Pietermaritzburg and Durban.

Natal. — Without precise locality: Hort. Kew, seed sent by

"Mr. R. Vanse”, Port Natal, without date (K); Jan. 1862 (K);

Feb. 1863 (K); Nov. 1939, Miss Bruce's handwriting on label

(K); Cooper 3106 (K). 2930 (Pietermaritzburg) Table Mt.

(-CB), Killick 504; Nchanga (-DA), Eshuis s.n.; Umzinyati,

Inanda (-DB), Medley Wood 1223 (NH, SAM); s.n. (K);

Monteseel (-DC), Strey 5208; Shongweni Dam (-DC), Morris

895; Westville, near Durban, Ward 6393.

Considering the restricted distribution of the

species and its relatively insignificant appearance, it

is of interest that it found its way into cultivation in

England at such an early stage. Seed was sent to

Kew by Mr. R. Vause (not Vanse or Vance as has

appeared on labels or in literature) of Durban.

Miss M. D. Gunn informs me that Mr. Vause was a

member of the Durban Botanic Garden Committee

during the time of McKen and was elected Ffon.

Secretary in 1858.

As discussed under P. strigosus (p. 409), the main

difference between P. strigosus and P. purpuratus lies

in the dense velvety greyish tomentum of the latter as

against the more strigose rusty pubescence of P.

strigosus. P. purpuratus also has a more compact

habit with more marked purple coloration on the

underside of the leaves. In floral characters the two

are almost identical though the stamens are even more

inconspicuous in P. purpuratus. Another species

with very shortly exserted stamens is P. oertendahlii

(see below).

A remarkable character shown by P. xerticillatus ,

P. strigosus and P. purpuratus is the presence of a red

gland-dot situated between the anther cells. It is

apparent that these three species form a closely related

group.

Chromosome number: 2n = 28 (De Wet, 1958),

2n = 30,56 (Morton, 1962).

25. Plectranthus oertendahlii Th. Fries jun. in

Acta Hort. Gothoburg. I :253 (1924). Type: Cult.

Uppsala (UPPS, holo.).

Perennial semi-succulent herb, freely branched, up

to 20 cm tall; branches decumbent, 4-angled, up to

20 cm long, rooting at the lower nodes, shortly

appressed glandular-tomentulose with short antrorse

multicellular, purple-tinged hairs and subsessile glands,

the hairs longer at the nodes. Leaves : petiole 1-3 cm

long, glandular-tomentulose and purple-tinged as in

the stems; blade broadly ovate to suborbicular 3-4

(-4,5) cm long, 2,5-4 (-4,5) cm broad, semi-succu-

lent, drying thin-textured, hispid with forward-poin-

ting bristles on both surfaces, green with pale nerves

above, purple-tinged below with raised nerves, freely

dotted with transparent gland-dots; apex obtuse to

rounded, base obtuse to shortly cuneate; margin

shallowly crenate-dentate with 5-7 pairs of teeth,

shortly hispid. Inflorescence terminal, simple or pani-

culate, 7-20 cm long, with up to 2 pairs of branches

near the base; rhachis densly and shortly glandular-

hispidulous; bracts lanceolate, acuminate, 5-6 mm

long, persisting beyond the fruiting stage, glandular-

hispidulous, shortly cilate. Flowers in sessile, usually

3-flowered cymes forming usually 6-flowered verti-

cillasters, verticillasters 1,2-1, 5 cm apart; pedicels

3-5 mm long, densely glandular-hispidulous. Calyx

4 mm long at flowering, enlarging to 8 mm long in

fruit, finely glandular-hispidulous, purple-tinged; up-

per lip erect, ovate, acute, up to 3 mm long, slightly

decurrent on the tube ; lower lip 4-toothed, up to 5 mm

long at maturity, teeth lanceolate-subulate, 1,5-2, 5

mm long, the lower pair distinctly longer. Corolla

suffused with pale mauve, paler towards the lobes,

the upper lip with 4 erect purplish stripes, minutely

puberulous outside; tube 1,2-1, 3 cm long, laterally

compressed, expanding in the calyx forming a saccate

base 4 mm deep and narrowing to 1,75 mm at the

throat; upper lip erect, 5 mm long, 4-5 mm broad,

bilobed at the apex and with 2 small, lateral ear-like

lobes; lower lip shallowly boat-shaped, horizontal to

slightly deflexed, 4-5 mm long. Stamens free, of two

lengths, the lower the longer, 3 mm long. Style

exserted by 3 mm. Fig. 27.

Described in 1924 from plants that had been cul-

tivated in Scandinavia for more than 20 years and

recently located in wooded river valleys near the coast

in the Port Shepstone District of Natal.

Natal. — 3030 (Port Shepstone): Oribi Gorge (-CA), Nichol-

son 1055; 1060; 1191; cultivated in Grahamstown in 1936,

Britten 6539 (GRA); Uvongo River (-CD), Nicholson 1025;

1202; Strey 11063.

Cultivated. — Plant received from the Botanical Institute,

Fund, and cultivated in Pretoria, Codd 10782.

Although cultivated in Sweden since the early part

of this century, its country of origin remained unre-

corded. The next gathering known was by Miss

Lilian Britten in 1936 who grew a plant in Grahams-

town, collected earlier in the Oribi Gorge, Natal.

Fortunately she made a herbarium specimen before

her plant was lost to cultivation, but its identity was

not recognized at the time. It is only recently, due to

the enthusiasm and assiduous collecting around Port

Shepstone of Mr H. B. Nicholson, and later by Mr

Strey, that its occurrence in the wild state has been

established. It appears to have a limited distribution,

although often locally very common, and one can

only speculate how it reached Scandinavia so many

years ago. Mr Nicholson points out that there are

strong historical links between this part of the country

and Sweden, the Hermansburg Mission Society having

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 27. — Plectranthus oertendahlii, cul-

tivated, Pretoria (ex hort. Lund,

Sweden), x|.

settled in the area in 1897 while a Swedish surveyor

was responsible for the development of the township

of Uvongo and insisted on the retention of large

park areas.

The flowers on naturally occurring plants are often

smaller than on the cultivated type plant, while the

nerves on the upper surface are not always as pallid,

but there is little doubt that the gatherings listed from

Port Shepstone District represent P. oertendahlii

and that the wild origin of this species has been

cleared up.

P. oertendahlii is related to P. ciliatus but lacks the

long, purple-striped multicellular hairs of the latter

species (except at the nodes), while the smaller,

suborbicular leaves with few teeth are reminiscent of

P. verticillatus. In the herbarium, P. oertendahlii

is most likely to be confused with P. verticillatus

but the red gland-dots shown by the latter species are

lacking, while the corolla tube does not narrow

towards the throat as in P. oertendahlii. There are

also differences in stem pubescence and flower colour.

The species name commemorates I. A. Oertendahl,

chief gardener in the University Botanic Garden,

Uppsala, when the species was described by Prof.

Fries.

26. Plectranthus elegantulus Briq. in Bull. Herb-

Boiss. ser. 2,3:1005 (1903); Cooke in FI. Cap. 5,1;

286 (1910); Ross, FI. Natal 305 (1972). Type: Natal,

Karkloof, Rehmann 7368 (Z!, holo. ; PRE, photo.).

Soft, straggling herb up to 20 cm tall, pleasantly

aromatic; branches softly succulent, 4-angled, spread-

ing-erect, up to 30 cm long, finely glandular-stri-

gulose with short, appressed antrorse multicellular

hairs. Leaves : petiole 1,5-4 cm long, glandular-stri-

gulose as in the stems; blade broadly elliptic to

broadly ovate, 2,5-4 cm long, 2-3,5 cm broad,

thin textured, subglabrous to sparingly hispidulous

above, thinly strigulose on the nerves below, both

surfaces dotted with minute simple pale honey-

coloured droplets, the underside of the leaf not

suffused with purple; apex acute to obtuse, base

abruptly cuneate for 3-5 mm; margin finely ciliate,

regularly crenate-dentate with 5-8 pairs of teeth.

Inflorescence terminal, suberect, usually simple, rarely

with a pair of branches from near the base; racemes

4-12 cm long; rhachis finely antrorse strigulose to

hispidulous; bracts ovate-lanceolate to lanceolate,

long acuminate, 2,5-4 mm long, persisting to the

fruiting stage, gland-dotted as in the leaves and

with finely ciliate margins. Flowers in sessile 1-3-

flowered cymes forming 2-6 flowered verticillasters;

verticillasters 1-1,5 cm apart; pedicels 4-6 mm

L. E. CODD

413

long, minutely glandular-puberulous. Calyx 4 mm

long at flowering stage enlarging to 7-8 mm in fruit,

finely glandular-puberulous with a few fringing mul-

ticellular hairs; upper lip erect, ovate, acute, not

decurrent on the tube, 3 mm long, lower lip 4-toothed,

4 mm long at maturity, the lower pair the longer,

linear-subulate, median pair deltoid-subulate. Corolla

whitish with a few purple spots on the upper and

lower lip, minutely puberulous outside; tube 4-5

mm long, scarcely deflexed, expanding in the calyx

forming a saccate base, narrowing slightly towards

the throat; upper lip erect, 4-5 mm long and equally

broad, bilobed at the apex and with 2 small lateral

inconspicuous earlike lobes; lower lip boat-shaped

3 mm long, horizontal to slightly deflexed. Stamens

free, the lower pair the longer, up to 2 mm long.

Style horizontal, 3 mm long. Fig. 28.

Fig. 28. — Plectranthus elegantulus, Karkloof, Natal ( Codd 8582), X 1.

414

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

A small straggly herb on forest floors, usually in

moist places, restricted to the Natal Midlands.

Natal. — 2929 (Underberg): near Deepdale (-DD), Pole

Evans 4863; Strey 4817. 2930 (Pietermaritzburg): Karkloof

area (-AC), Rehmann 7368 (Z); Dimock-Brown 349; Codd

8582; Mol! 3446; Byrne (-CC), Galpin 12018. 3029 (Kokstad):

Ingeli Mt. (-DA), Hilliard & Burn 5782 (NH).

P. elegantulus is closely related to P. ciliatus but

is a smaller plant with smaller leaves and flowers.

The pubescence on the stems is finer and shorter,

consisting of short, appressed hairs and usually

lacking the long, multicellular, purple-striped hairs

characteristic of P. ciliatus.

In poorly prepared herbarium specimens there

may be a tendency to confuse P. elegantulus with

certain other species with small, few-toothed leaves,

such as P. verticillatus and P. saccatus. The first can

be readily separated by the presence of red gland-dots

on the undersides of the leaves. P. saccatus is a more

erect shrublet with strikingly large, usually blue-

mauve flowers and has subglabrous to minutely

puberulous stems and leaves.

Chromosome number: 2n=28 (De Wet, 1958).

27. Plectranthus ciliatus E. Mey. ex Benth. in

E. Mey., Comm. 227 (1837); Drege, Zwei Doc. 150

(1843); Benth. in DC., Prod. 12:62 (1848); Cooke in

FI. Cap. 5, 1:275 (1910); Verdoorn in Flow. PI. Afr.

27: t. 1051 (1949); Compton, FI. Swaz. 66 (1966);

Ross, FI. Natal 305 (1972). Type: “Omsamwubo”

(Umzimvubu, i.e. near Port St. Johns), Drege s.n.

(K, numbered 4777!, lecto.; MO;P;S).

P. natalensis Guerke in Bull. Herb. Boiss. 6: 552 (1898);

Cooke in FI. Cap. 5, 1: 283 (1910), partly, excl. Tyson 1793,

Wood 558. Type: Natal, Camperdown, Rehmann 7701 (Z!,

holo.; PRE, photo.).

Soft, straggling herb or shrublet up to 60 cm tall,

pleasantly aromatic; branches softly succulent, 4-

angled, spreading-erect, up to 60 cm long, glandular

pilose with few to many long multicellular spreading

to forward-pointing hairs and short stiff hairs, the

hairs having maroon-purple sap, giving a purplish

colour to the branches. Leaves : petiole 1,5-3, 5 cm

long, densely glandular shaggy as in the stems; blade

broadly elliptic, broadly ovate or, rarely, subrotund,

(3,5-)4-8 cm long, 3-5,5 cm broad, thin-textured and

smooth to thickish and rugose, sparingly hispid above,

nerves prominent and strigose below, both surfaces

freely dotted with minute simple honey-coloured

droplets, the underside of the leaf usually suffused

with purple; apex acute to obtuse, base attenuate to

abruptly cuneate for 6-12 mm; margin regularly and

shallowly crenate with 8-14 pairs of teeth, conspi-

cuously ciliate. Inflorescence terminal, erect, simple or

with a pair of branches near the base; racemes 6-20

cm long; rhachis antrorse-strigose with multicellular

and shorter hairs; bracts ovate-lanceolate to narrowly

lanceolate, long-acuminate, 3-6 mm long, persisting

to the fruiting stage, gland-dotted as in the leaves and

with conspicuously ciliate margins. Flowers in sessile,

usually 3-flowered cymes forming it 6-flowered verti-

cillasters, verticillasters 1-2 cm apart; pedicels 5-8

mm long, minutely strigulose. Calyx 4-5 mm long at

flowering stage enlarging to 8-10 mm long in fruit,

appressed hispid to glandular puberulous, fringed

with multicellular hairs; upper lip erect, ovate, acute,

not decurrent on the tube, up to 4 mm long; lower

lip 4-toothed, 4 mm long at maturity, the lower pair

much the longer, linear-subulate, median pair deltoid-

subulate. Corolla whitish background freely speckled

with purple on the inner surface of upper and lower

lips, sparingly glandular outside; tube 5-8 mm long

slightly deflexed and expanding in the calyx forming a

saccate base, narrowing slightly towards the throat;

upper lip erect, 5-7 mm long and equally broad,

bilobed at the apex and with 2 small, lateral, acute,

ear-like lobes; lower lip boat-shaped, 3-6 mm long,

horizontal to deflexed. Stamens free, the lower pair

the longer, up to 8 mm long, horizontal to recurved.

Style horizontal to ascending, slightly exceeding the

stamens. Fig. 29.

A soft straggling plant on forest floors and moist,

shady places, distributed from Uniondale and Knysna

along the eastern Cape semi-coastal areas to eastern

Natal and the mountains of eastern Transvaal.

Transvaal. — 2430 (Pilgrim’s Rest): Mount Sheba, near

Pilgrim’s Rest (-DD), Jones & Leach 27. 2531 (Komatipoort):

4 km W. of Havelock Mine (-CD), Codd 9528.

Swaziland. — 2631 (Mbabane): Hlatikulu, Stewart s.n. (K).

Natal. — 2731 (Louwsburg): near Ngome Mission Station

(-CD), Codd 7016. 2830 (Dundee): Qudeni Forest (-DB),

Fisher & Schweickerdt 135. 2831 (Nkandla): Nkandla Forest

(-CA), Codd 6968; Entonjaneni (-CB), Andrews 12; Eshowe

(-CD), Lawn 129 (NH); Gerstner 2211 (NH); Mecbold s.n.

(M); Entumeni (-CD), Medley Wood 3997 (K); Gerstner 2781.

2832 (Mtubatuba): Mpate River (-AA), Ward 3393; Duku-

duku Forest (-AC), Strey 7586. 2930 (Pietermaritzburg): 8

km N.W. of York, Codd 8577; near Greytown (-BA), Galpin

14743; Noodsberg (-BD), Edwards 3001; Table Mt., near

Pietermaritzburg (-CB), Killick 329; Camperdown (-DA),

Rehmann 7701 (Z); 7703 (Z); Imnda (-DB), Johnson 1311;

Medley Wood 63 (K, NH); Inchanga (-DC), Rehmann 7916 (Z);

Fields Hill, near Pinetown, Rehmann 8033 (Z); 8034 (Z);

Nkutu Falls (-DD), Redshaw 12; Molweni Kloof (-DD),

Hilliard & Burn 3811; Westville (-DD), Ward 4949. 2931

(Stanger): Durban (-CC), Jenkins sub TRY 7712. 3030 (Port

Shepstone): near Dumisa (-AD), Hilliard & Burtt 3815;

Amanzimtoti (-BB), Franks sub NH 14574 (NH); Umgayi

(-BC), Ward 5460; Oribi Flats (-CB), McClean 548; Uvongo

River (-CD), Nicholson 352 (NH); 1023; Izotsha (-CD),

Strey 8085. 3130 (Port Edward): Umtamvuma River (-AA),

Nicholson 997; Beacon Hill (-AA), Strey 7234; Port Edward

(-AA), Strey 4938.

Cape. — 3029 (Kokstad): near Tabankulu (-CD), Story 4199;

11 km S.S.W. of Bizana (-DD), Acocks 13409. 3128 (Umtata):

Baziya (-CB), Baur 37 (K). 3129 (Port St. Johns): Magwa

Falls (-BC), Codd 9313 A; 9313B; Port St. Johns (-DA), Galpin

2842; between Umtata and Umzimvubu Rivers (-DA), Drege

4777 ( K ); s.n. (P, S). 3130 (Port Edward): Umtamvuna Water-

fall (-AA), Strey 4471. 3226 (Fort Beaufort): Alice (-DD),

Sidey 606. 3227 (Stutterheim): Amatola Mts. (-CA), Erens

2225; Pirie (-CD), Kuntze s.n. (K); Sim s.n.; Kei Road (-DA),

Rogers 3201 (Z); near Komga (-DB), Flanagan 739; 1721.

3228 (Butterworth): near Kentani (-CB), Pegler 352. 3323

(Willowmore): near Stone’s Hill (-CA), Schonland 3146.

partly; Knysna Forest (-CC), Keel 736. 3326 (Grahamstown):

near Grahamstown (-BC), Burchell 3580 (K, P); Atherstone 17;

Britten 5219; Story 2124.

P. ciliatus is allied to P. fruticosus and has similar

honey-coloured gland-dots throughout the plant. It

differs from P. fruticosus in being a smaller, straggling

plant, the stems are more densely beset with purplish,

multicellular hairs, the leaves are more cuneate at the

base and the bracts and calyx are more conspicuously

fringed with multi-cellular hairs. The flowers are also

different, having a whitish background, freely speckled

on the upper and lower lips with purple spots.

P. grallatus may readily be separated from P.

ciliatus by its red gland-dots and more spreading,

hispid pubescence on the stems, as against the antrorse

strigose pubescence of P. ciliatus. P. grallatus also has

whitish flowers but with only a few purple spots on

the upper lip.

P. elegantulus is in all ways a smaller plant than

P. ciliatus, with smaller leaves and shorter, finer

pubescence.

Several specimens of Drege from “Omsamwubo”

(Umzimvubu or Port St. Johns River) have been

seen and the specimen ex Herb. Benth., numbered

4777 in Kew, is selected as the lectotype. The type of

P. natalensis Guerke, Rehmann 7701 (Z) falls within

the range of variation of P. ciliatus.

Chromosome number 2n =28 (De Wet, 1958).

L. E. CODD

415

Fig. 29. — Plectranthus ciliatus, Magvva Falls, eastern Cape Province ( Codd 93 1 3 A), x 1.

28. Plectranthus fruticosus VHerit., Stirp. Nov.

fasc. 4: 85, t. 41 (March 1788); Ait., Hort. Kew.

ed.l, 2: 322 (1789); Willd., Sp. PI. 3: 168 (1800);

Thunb., FI. Cap. ed. Schult. 448 (1823); Benth.,

Lab. 32 (1832); in DC., Prodr. 12: 62 (1848); Briq.

in Pflanzenfam. 4,3a: 354 (1897); S. Moore in J. Bot.,

Lond. 41 : 406 (1903); T. Cooke in FI. Cap. 5, 1 : 271

(1910); Burtt in Bot. Mag. t. 9616 (1940); Dyer &

Bruce in Flow. PI. Afr. 28: t. 1101 (1951); Bailey,

Stand. Cycl. Hort. 3: 2712 (1963); Bullock & Killick

in Taxon 6: 239 (1957); Compton, FI. Swaz. 66 (1966);

Courtenay-Latimer et <?/., Flow. PI. Tsitsikama t. 54

(1967); Blake in Contr. Queensland Herb. 9: 3 (1971);

Codd in Mitt. Bot. Staatssamml. Miinchen 10: 247

(1971); Ross, FI. Natal 305 (1972). Type: t. 41 of

L’Herit., Stirp. Nov. fasc. 4 (1788).

Germanea urticifolia Lam., Encycl. 2: 690 (April 1788);

Tabl. Encycl. Ill: t. 514 (1819). Type: a cultivated plant as

illustrated in Tabl. Encycl. Ill: t. 514 (1819).

Plectranthus urticifolius (Lam.) Salisb., Prodr. Stirp. Hort.

Chapel Allerton Vig. 88 (1796). P. galpinii Schltr. in J. Bot.,

Lond. 34: 393 (1896); T. Cooke in FI. Cap. 5, 1: 282 (1910);

Phillips in Flow. PI. S. Afr. 8: t. 294 (1928). Type: Transvaal.

Barberton, Rimer's Creek, Galpin 939 (B, holo.; GRA!; NH!,

PRE!). P. arthropodus Briq. in Bull. Herb. Boiss. ser. 2, 3:

1073 (1903); T. Cooke in FI. Cap. 5, 1: 273 (1910). Type:

Transvaal, Houtbosch, Rehmann 6151 (Z !, holo.; PRE, photo.).

P. charianthus Briq. in Bull. Herb. Boiss. ser. 2, 6: 824 (1906).

Type: Transvaal, Houtbosch, Rehmann 6157 (Z!, holo.; PRE,

26700-4

416

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

photo.). P. peglerae T. Cooke in Kew Bull. 1909: 378 (1909);

FI. Cap. 5, 1: 283 (1910); Bews, Plant Forms and Evol. in S.

Afr. 98 (1925). Type: Eastern Cape, Kentani, Pegler 377 (K!,

holo.; GRA!.; PRE!). P. behrii Compton in J. S. Afr. Bot. 11 :

122 (1945); Lewis in FI. PI. Afr. 28: t. 1109 (1951); Batten &

Bokelmann, Wild Flow. E. Cape 126, t. 100 (1966). Type:

Pondoland, Lusikisiki, Behr sub NBG 1252/31 (NBG!, holo.;

PRE, clono.).

Soft shrub 60 cm-2 m tall, freely branched,

aromatic; roots fibrous; branches ascending, 4-angled,

usually purplish, glandular-pilose with few to many

antrorse multicellular hairs and minute sessile gland-

dots, rarely (at high altitudes) the stems densely

beset with short gland-tipped hairs, hairs longer and

spreading at the nodes. Leaves : petiole 2-5 cm long,

densely antrorse pilose; blade broadly ovate to ovate-

elliptic or rarely, lanceolate-elliptic, 4-14 cm long,

3,5-11 cm broad, thin-textured and smooth to

thickish and rugose, subglabrous to sparingly pilose

above, more densely so on the nerves below, rarely

(at high altitudes) densely glandular-hispidulous,

sparingly to densely dotted with minute, simple,

honey-coloured droplets, the underside of the leaf

usually suffused with purple to a greater or lesser

extent; apex obtuse to acute; base obtuse or truncate,

then abruptly attenuate to the petiole for usually

less than 1 cm (longer and more cuneate at the base

in specimens from southern Cape Province); margin

with 12-22 pairs of short broad obtuse teeth, teeth

regular, entire or rarely with an occasional secondary

small tooth. Inflorescence terminal with 1-3 pairs of

branches at the base forming a pyramidal panicle

8-25 cm long; rhachis glandular-puberulous with

dense short hairs, occasional long antrorse multi-

cellular hairs and sessile glands; bracts lanceolate to

ovate-lanceolate, acuminate, 2-6 cm long, persisting

to the fruiting stage, glandular-pilose with a few

fringing multicellular hairs. Flowers in sessile, opposite

(l-)3-flowered cymes forming 2-6-flowered verti-

cillasters; verticillasters 0,5-2, 5 cm apart; pedicels

4-9 mm long, glandular-hispidulous, with few

scattered multicellular hairs. Calyx 4 mm long at

flowering stage enlarging to 7 mm long in fruit,

appressed hispid with numerous honey-coloured

gland-dots and scattered multicellular hairs; upper

lip erect, ovate, acute to acuminate, slightly decurrent

on the tube, 2,5 mm long; lower lip 4-toothed, 4 mm

long at maturity, the lower pair the longer, subulate,

median pair deltoid-subulate. Corolla bluish- mauve,

rarely pink ( P . behrii) or pale blue, speckled with

darker spots on the upper lip, sparingly glandular

outside; tube 3-8 mm long, deflexed and expanding

in the calyx, saccate to distinctly spurred at the base,

narrowing slightly towards the throat; upper lip

erect, 2,5-6 mm long and up to 7 mm wide, bilobed

at the apex and with 2 small lateral ear-like lobes;

lower lip boat-shaped, 2-5 mm long, at first horizontal

and later strongly recurved below the tube. Stamens

free, the lower pair the longer, up to 8 mm long, at

first pointing forward and curved upwards at the

tips, later coiled below together with the lower

corolla lip. Style horizontal, exserted by 6-8 mm.

Fig. 30.

Found in forest and scrub forest and in shady

places among rocks from Caledon district in south-

western Cape, along the south coast to eastern Cape,

eastern Natal at medium to low altitudes, Swaziland,

and in the mountains of eastern central and northern

Transvaal.

Transvaal. — 2229 (Waterpoort): 32 km N.E. of Louis

Trichardt (-DD), Gerstner 5996. 2328 (Baltimore): Blouberg

(-BB), Codd 8747; 9170. 2329 (Pietersburg): near Louis Trich-

ardt (-BB), Breyer sub TRV 19519; Young sub TRV 26638;

Houtboschberg (-DD), Nelson 43 1. 2330 (Tzaneen): 32 km E.

of Louis Trichardt, Meeuse 9786; Westfalia Estate (-CA),

Scheepers 239; Woodbush Forest (-CC), Rehmann 6155 (Z);

6157 (Z); Codd 9413; Prosser 1907; Letty 457; Magoebaskloof

(-CC), Codd 8412; 8414; New Agatha Forest Reserve (-CC),

Muller & Scheepers 6. 2427 (Thabazimbi): Krantzberg (-BC),

Erens 2100. 2428 (Nylstroom): Palala River, Breyer sub TRV

18031. 2430 (Pilgrim’s Rest): The Downs (-AA), Codd & Dyer

7753; 20 km S.W. of Ofcolaco (-AA), Codd 9447 ; Mariepskop

(-DB), Van Son sub TRV 31560, partly; Codd 7857; Meeuse

9991; Van der Sehijff 4409; 4462; 4464; 4822; 4990; 5525;

5633; Mount Sheba (-DD), Jones & Leach 5; near Pilgrim's

Rest (-DD), Galpin 14500, partly. 2528 (Pretoria): Trichardts-

poort (-DB), Bredell 8; Bruce 95. 2529 (Witbank): Loskopdam

(-AD), Theron 1120; 1791 ; 24 km N.W. of Middelburg (-CB),

Codd 10104. 2530 (Lydenburg): near Lydenburg (-AB), Wilms

1127 (G); 1182 (K); between Lydenburg and Dullstroom (-AC),

Pole Evans 4291; Lunsklip Waterfall (-AD), Strey 3196;

Spitskop (-BA), Wilms sub TRV 6448; Sabie Valley (-BB),

Galpin 13750; Rosehaugh (-BD), Holt 360; Waterval Boven

(-CB), Rogers 14480; 14486; near Kaapsche Hoop (-DB),

Bruce 336. 2531 (Komatipoort): Barberton (-CC), Rogers

25526; Liebenberg 2353; Rimers Creek (-CC), Galpin 939;

Codd 8181; 8183. 2630 (Carolina): 10 km W. of P.O. The Brook

(-BA), Codd 9495. 2730 (Vryheid): Piet Retief (-BB), Leipoldt

s.n.

Swaziland. — 2531 (Komatipoort): near Havelock Mine

(-CC), Miller 7245; Codd 7836; Piggs Peak, Kings Forest

(-CC), Compton 29977. 2631 (Mbabane): near Mbabane (-AC),

Bolus 12247 (BOL, K); Compton 25785; Black Umbuluzi

valley (-AD), Compton 26792; near Mankaiana (-CD), Stewart

s.n. (K).

Natal. — Without locality, Gerrard 1235 (K). 2730 (Vryheid):

Kaffirdrift (-CB), Thode A341 . 2731 (Louwsburg): near Ngome

(-CD), Gerstner 4568; Codd 9570. 2830 (Dundee): near Help-

mekaar (-AD), Codd 5934; Qudeni (-DB), Gerstner 663; Codd

6993; Krantzkop (-DD), Dyer 4350; Codd 9656; Krantzkop-

Middeldrift road (-DD), Edwards 2102. 2831 (Nkandla):

Nhlwati (-BB), Ward 3438; near Eshowe (-CD), Codd 9650;

Admiraal & Drijfhout s.n.; Ngoye (-DC), Wylie sub Wood

8521 ; Twinstreams (-DC), Venter 3388; 24 km N.W. of Port

Durnford (-DD), Codd & Dyer 2823. 2930 (Pietermaritzburg):

near Hylton (-AC), Marais 809; Karkloof forest (-AC), Moll

3498; Karkloof area (-AD), Codd 8580; 34 km from Howick

on Mt. Alida road (-AD), Ross 2067; Tala farm (-CD), Moll

3057; Nchanga (-DA), Codd & Dyer 2804; Drummond (-DD),

McClean 140; Umbilo River Falls (-DD), Rehmann 8150 (Z).

2931 (Stanger): Oqaqeni (-AA), Edwards 1816. 3029 (Kokstad):

(-DA), Ingeli Forest, Codd 8562; Ward 181. 3030 (Port Shep-

stone): Oribi Gorge (-CB), Nicholson 1049.

Cape. — Without locality Thunberg s.n. (UPS); Sparmann

s.n. (S); Ekeberg s.n. (SBT); Ecklon 668 (SAM). 3029 (Kokstad):

Tonti Forest (-CD), Forest Officer 578. 3129 (Port St. Johns):

10 km W. of Port St. Johns (CB), Codd 9299; Isnuka Forest

(-DA), Galpin 2841; Ntafufu (-DA), Strey 8533. 3226 (Fort

Beaufort): Fort Fordyce (-CB), Story 2112. 3227 (Stutterheim):

Pirie (-CD), Kuntze s.n. (K); Sim 19589. 3228 (Butterworth):

near Kentani (-AD), Pegler 377; Codd 9248; Dwessa Forest

(-BD), Britten 7009. 3322 (Oudtshoorn): Montagu Pass (-CD),

Schweickerdt s.n.; Rehmann 276 (Z); near George (-DC),

Prior s.n. (K); Burchell 6051 (K); Schlechter 2415; Hutchinson

1279a; Witfontein Forest Reserve (-DC), Marais 691; Wilder-

ness (-DC), Mogg s.n. 3323 (Willowmore): Outeniqua (-CD),

Krauss 194 (M); Drege s.n. (P); Blaauwkrantz River (-DC),

Penther 1708 (M); Tsitsikama Forest (-DD), Pappe s.n. (SAM);

Galpin 4423; Schlieben 10229; Bokelmann s.n. 3419 (Caledon):

Genadendal (-BA), Prior s.n. (K); Rivier Sonder Einde Range,

Esterhuysen 5069. 3420 (Bredasdorp): Grootvadersbosch

(-AB), Marloth 3500; Voormansbosch (-AB), Zeyher 3545

(SAM); near Swellendam (-AB), Ecklon 70.10 (K); near

Heidelberg (-BB), Esterhuysen 10367; National Bontebok Park

(-CA), Liebenberg 6359. 3421 (Riversdale): Langeberg (-AB),

Muir 2920. 3423 (Knysna): near Knysna (-AA), Theron 1013;

Werdermann & Oberdieck 932; The Heads (-AA), Williamson

120; Deepwalls Forest (-AA), Keet 419; Schonland 3587;

Marsh 1324; near Garden of Eden (-AA), Kapp 108; Bruce 234;

Plettenberg Bay (-AB), Smart sub TRV 26685. 3424 (Humans-

dorp): Clarkson (-AB), Thode A979.

There is a certain amount of variation in this species

according to habitat conditions. Under forest con-

ditions the plants are taller and more branched, the

stems less pubescent and the leaves larger, thinner and

smoother. Flower colour is normally pale to deep

mauve with purple flecks on the upper lip of the

corolla. A form with pinkish-lilac flowers was described

as P. behrii Compton but it is known from only one

or two gatherings in the Lusikisiki district and differs

in no other respect from P. fruticosus. It is a popular

horticultural subject and should rather have been

given a cultivar name than separate taxonomic status.

L. E. CODD

417

W

Fig. 30. — Plectranthus fruticosus, Karkloof, Natal ( Codd 8580), X 1.

In typical P. fruticosus there is a distinct dorsal

spur at the base of the corolla tube, from which the

generic name Plectranthus is derived (Greek: spur

flower), but this is not a constant character, parti-

cularly in Natal and the eastern Transvaal, where

plants in some populations have corollas spurred to

a greater or lesser degree, while in other populations

the corolla is saccate or bulged at the base. For this

reason P. galpinii Schltr. is now included in

synonymy. Its corolla is saccate at the base and tends

to be paler in colour than typical P. fruticosus, but it

differs in no significant respect. The type of P. peglerae

T. Cooke has a spurred corolla tube and cannot be

differentiated from P. fruticosus. P. arthropodus Briq.

and P. charianthus Briq. are both based on Rehmann

specimens from the Woodbush area, north-eastern

Transvaal. Even though the type of P. charianthus

consists of inflorescences only, both this and the type

of P. arthropodus are recognisable as P. fruticosus,

(see following page).

418

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

The two species most likely to be confused with

P. fruticosus are P. grallatus Briq. and P. rubropunc-

tatus Codd, and both occur in the Woodbush area.

However, these two may readily be distinguished from

P. fruticosus by the presence of conspicuous red to

brownish gland-dots on the undersides of the leaves,

bracts, calyx and corolla. These are best seen at a

magnification of about 20x and with a bright

artificial light source. They are larger, somewhat

flattened to depressed in the centre and often appear

as a ring of minute red gems, which are very different

from the honey-coloured, simple gland-dots of P.

fruticosus. These ruby-coloured gland-dots are found

in many other species, but are diagnostic for separating

P. fruticosus from P. grallatus and P. rubropunctatus.

These three species have been studied in the eastern

Transvaal and further distinctions have been noted.

Flower colour in P. grallatus and P. rubropunctatus is

usually white, flushed to some extent with mauve or

pink, while the stems are glandular-hispid with

spreading, not antrorse, multicellular hairs. At high

altitudes on Mariepskop in the eastern Transvaal,

both P. rubropunctatus and P. fruticosus apparently

occur and here the distinctions between the two

species are not as clear-cut as elsewhere. Here P.

fruticosus develops thicker leaves with a more marked

glandular-hispid pubescence of leaves and stems, while

the red gland-dots of P. rubropunctatus are less

marked or become obscured by the more dense

pubescence. Elsewhere in its sympatric distribution

from Swaziland to the Soutpansberg, the distinction

between P. rubropunctatus and P. fruticosus presents

no difficulty.

P. grallatus has a wide distribution from the eastern

Cape to north-eastern Transvaal, but it occurs more

inland, at higher altitudes and more exposed situations

than P. fruticosus. In addition to the differences noted

above, the leaves of P. grallatus have more coarsely

dentate margins and the base of the leaf is longer

attenuate into the petiole than in P. fruticosus. The

distinctions between P. grallatus and P. rubropunc-

tatus are discussed under these two species.

Chromosome number 2n =28 (De Wet, 1958).

29. Plectranthus grallatus Briq. in Bull. Herb.

Boiss. ser. 2,3: 1004 (1903); Cooke in FI. Cap.

5,1: 287 (1910); Trauseld, Wild Flow. Drakensberg

160 (1969); Jacot-Guillarmod, FI. Lesotho 239 (1971);

Ross, FI. Natal 305 (1972). Type: Cape, Mount

Frere, Sclilecter 6415 (Z,! holo.; GRA!; PRE!).

P. transvaalensis Briq., I.c. 1005 (1903); Cooke, l.c. 288

(1910); Phillips in Ann. S.A. Mus. 16: 241 (1917). Type:

Transvaal, Houtbosch, Rehmann 6154 (Z!, holo.). P. krookii

Guerke ex Zahlbr. in Ann. Naturhist. Hofmus. Wien 20: 48

(1905); Cooke, l.c. 274 (1910). Type: East Griqualand, between

Insizwa Range and Umzimhlava River, Krook in PI. Penther

1698 (W!, holo., K!).— var. grcindifolia T. Cooke, l.c. 275

(1910). Type: East Griqualand, near Kokstad, Tyson 1793

(K!, holo.). P. praetervisus Briq. in Bull. Herb. Boiss. ser. 2,

6: 825 (1906). Type: Natal, Mt. Prospect, Rehmann 6965 (Z!,

holo.). P. cooperi T. Cooke, l.c., 278 (1910), partly, as to

Cooper 2982 (K!). P. natalensis forma glandulosa Phillips in

Ann. S.A. Mus. 16: 241 (1917). Syntypes: several, including

Leribe, Dieterlen 417 (PRE!); near Witzies Hoek, Flanagan

1927 (PRE!). P. ciliatus sensu Jacot-Guillarmod, FI. Lesotho

239 (1971); P. fruticosus sensu Jacot-Guillarmod, l.c. (1971).

Aromatic herb 40-150 cm tall; stems 1-3 arising

from a tuberous rootstock, 4-angled, simple or

sparingly branched, glandular-hispid; pubescence

patent to crisped, mainly of short spreading hairs,

gland-tipped hairs and some long, multicellular

articulate hairs and scattered red gland-dots. Leaves:

petiole 2-10 cm long, sparingly to densely glandular-

hispid with pubescence similar to the stems; blade

broadly ovate 5-16 cm long, 3,5-14 cm broad,

relatively thin textured, smooth to slightly rugose,

thinly pilose above, denser below and often densely

glandular-hispidulous, especially on the nerves,

sparingly to densely gland-dotted, especially towards

the margin with minute reddish to brownish gland-

dots which are flattened to depressed in the centre

and usually situated in a slight depression of the leaf

surface, underside of leaf not suffused with purple;

apex acute to abruptly acuminate, base abruptly

to gradually cuneate, rarely truncate, usually attenuate

to the petiole for 1-1,5 cm; margin with 14-20 pairs

of rather pronounced, acute, irregular teeth, usually

with secondary teeth. Inflorescence paniculate with

1-3 pairs of branches near the base, pyramidal,

10-26 cm long; rhachis puberulous to minutely

or coarsely patent glandular-hispid with short gland-

tipped hairs, short to longish multicellular hairs and

usually some red gland-dots; bracts lanceolate to

ovate-lanceolate, acuminate, 2-6 mm long, persisting

to the fruiting stage, glandular-hispid with usually

some fringing hairs and red gland-dots. Flowers in

sessile, opposite 3-flowered cymes; verticillasters

0,6-2 cm apart; pedicels 4-8 mm long hispidulous.

Calyx 4 mm long at flowering enlarging to 7-8 mm

long in fruit, glandular hispid usually with red gland-

dots; upper lip erect, ovate, obtuse to acute, slightly

decurrent on the tube, 2,5 mm long, lower lip

4-toothed, 4 mm long, teeth deltoid-subulate, the

lower pair longer than the median. Corolla white

with a faint flush of pink or mauve on the upper lip.

which usually has a few spots as well, sparingly

pubescent outside and usually with some red gland-

dots; tube 5-8 mm long slightly deflexed, expanding

and saccate at the base, narrowing slightly towards the

throat; upper lip erect, obovate, 4-6 mm long and up

to 7 mm wide with a few purple dots near the base,

bilobed at the apex and with two projecting, acute,

ear-like lateral lobes; lower lip boat-shaped, 4-5

mm long, horizontal to slightly deflexed. Stamens free,

the lower pair the longer, up to 7 mm long. Style

exserted by 6-7 mm. Fig. 31.

Found in forest and scrub forest and in shady

places among rocks from the eastern Cape Province,

mainly along the Drakensberg escarpment and neigh-

bouring areas in Natal, Lesotho and the eastern

Orange Free State to central and eastern Transvaal.

Transvaal. — 2329 (Petersburg): near Haenertsburg (-DD),

Codd 9424 : 9427 A; Houtbosch (-DD), Rehmann 6154 (Z).

2427 (Thabazimbi): Krantzberg (-BC), Codd & Erens 3980;

Codd 7869. 2430 (Pilgrim’s Rest): Mamatzeeri (-AA), Junod

291 ; The Downs (-AA), Junod 4330; Shilouvane (-AB), Junod

4334; Mariepskop (-DB), Codd 7869. 2527 (Rustenburg):

Hartebeestpoort Dam, Pole Evans 4774. 2528 (Pretoria): Jack

Scott Nature Reserve (-CC), Wells 2457. 2530 (Lydenburg):

Dullstroom (-AC), Galpin 13358; Berlin (-DB), Hofmeyr s.n.

2531 (Komatipoort): Agnes Mine, near Barberton (-CC),

Codd 9546. 2627 (Potchefstroom): near Krugersdorp (-BB),

Jenkins sub TRV 10022; Mogg 23237; Witpoortjie (-BB),

Gilmore sub TRV 26374; Young sub TRV 26637; Murray s.n.

2730 (Vryheid): near Wakkerstroom (-AC), Van Dam sub TRV

24313; Galpin 9795; Devenish 401; 406.

O.F.S. — 2828 (Bethlehem): near Bethlehem (-AB), Potgieter

sub TRV 21896; Golden Gate National Park (-AD), Liebenberg

6825; 7475; Roberts 3239; 27 km E. of Clarens (-BC), Marais

1279; Fouriesburg (-CA), Potts 3107; Witzieshoek (-DB),

Flanagan 1927; Junod sub TRV 17316. 2829 (Harrismith):

near Harrismith (-AC), Sankev 231 (K); near Swinburne

(-AD), Codd 10525.

Natal. — 2729 (Volksrust): near Charlestown (-BD), C. A.

Smith 5661. Laingsnek (-BD), Codd 8632; Mt Prospect (-DB),

Rehmann 6965 (Z). 2730 (Vryheid): 24 km E. of Utrecht (-CA),

Codd 49A. 2731 (Louwsberg): Ngome Forest (-CD), Codd

9569. 2828 (Bethlehem): Natal National Park (-DB), Schelpe

1250 (NH); Edwards 543; Mont-aux-Sources, Schelpe 1512

(NH); Bayer & McClean 59. 2829 (Harrismith): Van Reenens

Pass (-AD), Medley Wood 8925 (NH); Phillips s.n.; Codd 8513;

Cathedral Peak Forest Station (-CC), Killick 1328; 2366;

Tabamhlope (-DC), West 1096. 2830 (Dundee): 14 km S. of

Fig. 31. — Plectranthus grallatus, Mt. Frere, eastern Cape Province (isotype: Schlechter 6415 in GRA), Xl.

Dundee (-AA), Codd 2419; near Weenen (-CC), Medley Wood

4378 (K); Rogers 28192. 2929 (Underberg): Cathkin Park,

Galpin 11840; Howlett 113; Champagne Castle (-AB), Bayer

1262; Acocks 10075 (NH); Strey 7841; Giants Castle (-AD),

Symons 193; Bos 991 ; Traitseld 374; 533; 583; near Otterburn

Store (-BB), Edwards 2682; Mooi River (-BB), Mogg 7206;

Loteni Nature Reserve (-BC), Killick 3856; near Underberg

(-CD), McClean 639. 2930 (Pietermaritzburg): 13 km W. of

Nottingham Road (-AC), Codd 8525; near Howick (-AC),

Medley Wood 7198; Balgowan (-AC), Mogg 3819; Dargle,

Moll 2677 .

Lesotho. — 2828 (Bethlehem): Leribe (--CQ, Dieterlen 417;

Moteng (-CC), Jaeot Guillarmod 3665. 2928 (Marakabia):

Mahlatsas (-?), Jacot Guillarmod 460; Ja;ot Guillarmod 2208.

Cape. — 3028 (Matatiele): near Mount Frere (-DD), Schlech-

ter 6415 (GRA, Z); Acocks 12546. 3029 (Kokstad): near Kok-

stad (-CB), Tyson 1793 (K); between Insizwa Mt. and Umzim-

hlava River, Krook 1698 (K, W). 3127 (Lady Frere): 25 km

N.E. of Engcobo (-BD), Marais 1384; Glen Grey (-CA),

Galpin 8042. 3128 (Umtata): 10 km N.E. of Halcyons Drift

(-BD), Marais 869. 3226 (Fort Beaufort): Kaga Mts. (-CA),

Fitchett 32. 3227 (Stutterheim): Hogsback (-CA), Johnson 1289;

1308; Pirie Forest (-CD), Dodd sub Galpin 8024.

Although P. grallatus is nearly as widespread in

South Africa as P. fruticosus and occupies similar

habitats, the two rarely occur together, one of the

few exceptions being in the Woodbush area of the

420

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

north-eastern Transvaal. The two are frequently

confused in the herbarium but are, in fact, quite

distinct. The distinguishing characters are discussed at

length under P. fruticosus.

P. grallatus and P. rubropunctatus are closely related

and occasional specimens from the Mariepskop area

are not easy to allocate with certainty. P. rubro-

punctatus is restricted to the eastern Transvaal and

Swaziland from Mbabane to Woodbush and may be

distinguished from P. grallatus mainly on leaf charac-

ters. In P. rubropunctatus the leaf margin has rounded

teeth which are entire, while in P. grallatus the

margin is more deeply dentate with larger, acute

teeth which are again minutely denticulate. There is

also a difference in the base of the leaf : in P. grallatus

the leaf is abruptly attenuate for about 1-3 cm at the

base but not markedly decurrent on the petiole;

P. rubropunctatus exists in two forms, in the typical

being attenuate and distinctly decurrent on the petiole

and in the other being truncate to subcordate at the

base.

P. grallatus shows a good deal of variation in

texture, pubescence and size of leaves according to the

degree of protection or exposure of the habitat. This

has led to the somewhat extensive synonymy. The

types of all synonyms have been seen and they do not

deviate in any significant way from typical P. grallatus.

Chromosome number 2n =28, 56 (De Wet, 1958).

A distinct variant has been found in the eastern

Cape Province and the Weza Forest area of southern

Natal. This form is smaller and more delicate than

the typical P. grallatus and the upper lip of the

corolla is marked with several vertical purple lines

of rows of dots. In other respects the plants agree with

P. grallatus and so separate status is not considered

justified. Specimens representative of this form are not

included in the citations above and are as follows:

Natal. — 3029 (Kokstad): Ingeli Forest (-DA), Codd 8561;

Strey 6308; Hilliard & Burtt 3477.

Cape. — 3128 (Umtata): Pot River Berg (-AB), Galpin 6818.

3227 (Stutterheim): Cata Forest Reserve (-CA), Story 3265.

30. Plectranthus rubropunctatus Codd , sp. nov.,

affinis P. grallato Briq., sed foliis crassioribus, mar-

ginibus aequaliter et non profunde crenatis differt.

P. aff. arthropodus sensu Compton, FI. Swaz. 66 (1966).

Suffrutex vel herba perennis, erectus vel procum-

bens; caules semi-succulenti, 4-angulati, ramosi, 1-2 m

longi, glanduloso-hispidi, rubropunctati. Folia petio-

lata; petiolus 2-8 cm longus; lamina late elliptica vel

late ovata vel subrotunda, 4-15 cm longa, 3-10 cm

lata, chartacea vel coriacea, supra hispidula, subtus

reticulata, hispidula vel strigulosa, rubropunctata,

apice obtusa vel rotundata, basi subcordata vel

truncata vel decurrens, margine aequaliter et non

profunde crenata. Inflorescentia erecta plerumque

paniculata, 10-25 cm longa; bracteae ovatae vel

obovatae, acuminatae, persistentes, 2-6 mm longae,

glanduloso-hispidae, rubropunctatae. Verticillastri

0,5-1 ,2 cm distantes. ± 6-floribus; pedicelli 4-6 mm

longi, glanduloso-hispiduli. Calyx 4-7 mm longus,

glanduloso-hispidus, rubropunctatus; lobus posticus

late ovatus, erectus, 2,5 mm longus, acutus vel

obtusus; lobi laterales deltoideo-subulati; antici

lineari-subulati, 4 mm longi. Corolla alba vel pallido-

lilacina, minute pubescentia, rubropunctata; tubus

3-4 mm longus, prope basin leviter deflexus, saccatus;

labium posticum erectum, 2,5-4 mm longum, pariter

latum, obscure 4-lobatum; labium anticum cym-

biforme, 2,5-4 mm longum. Stamina 4, filamentis

liberis, supera 5 mm, infera 6 mm longa. Stylus

6-8 mm fauce exsertus.

Type: Transvaal, Barberton District, Nelshoogte

Forest Station, Strey 4081 (PRE, holo.).

Perennial aromatic herb or sub-shrub up to 2 m

tall or procumbent with stems up to 2 m long (Sout-

pansberg, Blouberg); stems semi-succulent, woody

below, branched, 4-angled, usually purplish, glandular

hispid, usually with dense short pubescence and

gland-tipped hairs with few to many spreading multi-

cellular hairs and scattered red gland-dots. Leaves

chartaceous to somewhat coriaceous; petiole 2-8 cm

long, finely to coarsely glandular-hispid ; blade broadly

elliptic or broadly ovate to subrotund, 4-15 cm long,

3-10 cm broad, upper surface minutely hispidulous,

lower surface paler, reticulate-veined, hispidulous to

strigulose especially on the veins, sparingly to

copiously dotted with red to brownish flattened gland-

dots; apex obtuse to rounded, base subcordate or

truncate to shortly attenuate or markedly decurrent

on the petiole; margin regularly and shallowly crenate

with 12-30 pairs of teeth, teeth rounded, entire or

rarely again minutely toothed. Inflorescence terminal,

usually paniculate, with 1-3 pairs of branches near the

base, 10-25 cm long; rhachis densely glandular-

hispidulous; bracts ovate to obovate, acuminate,

2- 6 mm long, persisting to the fruiting stage, glan-

dular-hispid with scattered red gland-dots and often

a fringe of multicellular hairs. Flowers in sessile

3- flowered cymes forming 6-flowered verticillasters,

verticillasters 0,5-1, 2 cm apart; pedicels 4-6 mm

long, glandular-hispidulous. Calyx 4 mm long at

flowering enlarging to 7 mm long in fruit, glandular-

hispid with scattered multicellular hairs and few to

many red gland-dots; upper lip erect, broadly ovate,

2,5 mm long, acute to obtuse; lower lip 4-toothed,

about 4 mm long, median teeth deltoid subulate, lower

pair slightly longer linear-subulate. Corolla white,

scarcely or noticeably suffused with a pinkish-mauve

to bluish tinge, pubescent and with red gland-dots;

tube 3-4 mm long, slightly deflexed, expanding within

the calyx and saccate at the base, narrowing slightly

towards the throat; upper lip erect, 2,5-4 mm long

and equally broad in the upper part, emarginate at

the apex and with 2 small lateral ear-like lobes;

lower lip boat-shaped, 2,5-4 mm long, at first

horizontal, later deflexed. Stamens free, the lower pair

the longer, up to 6 mm long. Style exserted by 6-8 mm.

Fig. 32.

Found as an under-storey soft shrub in forest,

scrub-forest and in shady places among rocks and

grass at relatively high altitudes from 1 200 to 2 000 m

in the northern and eastern Transvaal and in Swazi-

land.

Transvaal. — 2229 (Waterpoort): between Mountain Inn

and Hanglip (-DD), Meeuse 10 19 1. 2230 (Messina): Entabeni

Forest Reserve, Muchindudi Falls (-CC), Codd 4187. 2328

(Baltimore): Blouberg (-BB), Strey & Schlieben 8510. 2330

(Tzaneen): Westfalia Estate (-CA), Scheepers 1106; Politsi

(-CA), Bayliss BA 2756; Magoebaskloof, Debingini Falls

(-CC), Obermeyer sub TRY 31862; Codd 9421; New Agatha

Forest Reserve. Midler & Scheepers 218; 2430 (Pilgrims Rest):

Marovuni (-AA), Junod 72; Wolkberg (-AA), Meeuse 9874;

9912; The Downs, (-AA), Codd & Dyer 7751 ; 16 km S.W. of

Ofcolaco (-AB), Codd 9446; Joubert 4; Mariepskop (-BD),

Codd 7863; 7888; Werdermann & Oberdieck 1889; Meeuse

9946; 9953; 9979; Van der Schijff 4455; 4457; 4458; 4460;

4687; 4848; 4870; 4900; 4901; 4902; 7326; near Pilgrims Rest

(-DD), Galpin 14500, partly; Graskop (-DD), Galpin 4484;

Kowyns Pass (-DD), Codd 9478. 2530 (Lydenburg): between

Lydenburg and Sabie (-BA), Meeuse 10048; Sabie (-BB),

Rogers 23681 ; Nelshoogte Forestry Station (-DB), Codd 8143;

Strey 4080; 4081. 2531 (Komatipoort): near Barberton (-CC),

Codd 9543; Story 5994; Liebenberg 2353.

Swaziland. — 2531 (Komatipoort): near Piggs Peak (-CC),

Codd 9526; Compton 27672; 28725; 29986; 32246; Karsten

s.n. 2631 (Mbabane): between Mbabane and Piggs Peak.

Hilliard & Burtt 3555; 4 km N. of Forbes Reef (-AA), Ihlen-

feldt 2493; Schlieben 9547; near Mbabane (-AC), Compton

'26748; 31996; 32251.

L. E. CODD

421

NATIONAL

Pretoria.

Fig. 32. — Plectranthus rubropunctatus, Nelshoogte, eastern

Transvaal (holotype: Strey 4081 in PRE), x)-.

A good deal of variation in size and texture of leaves

is associated with the habitat. Plants growing in dense

forest have large, thin-textured leaves while those

growing in exposed, drier situations e.g. among rocks

above the forest zone, have smaller and thicker

leaves.

There is also some interesting variation which is

associated with geographic distribution. At the

extreme north of the range, on the Blouberg and

Soutpansberg, the stems are procumbent and trailing,

while from Magoebaskloof southward to its

southerly limit in Swaziland, the plants are erect

and bushy. Further investigation is required to see

if the procumbent form should be accorded separate

status of some sort.

In the more southerly distribution area there is

variation in the shape of the leaf base. In Magoebas-

kloof and on Mariepskop the leaf base is truncate or

subcordate and is not decurrent on the petiole.

Further south, at Nelshoogte and in Swaziland, the

leaf blade tends to be distinctly decurrent on the

petiole often giving the petiole a winged appearance.

There are, however, intermediates in which the

base is shortly cuneate as in P. grallatus; such plants

can be distinguished from P. grallatus mainly by

their more regular and shallower crenate-dentate

leaf-margins and, usually, by the somewhat thicker

texture of the leaf of P. rubropunctatus.

P. grallatus is its neareast affinity and the two

species have similar reddish gland-dots on the under-

side of the leaf and other parts of the plant, and the

flowers of both are whitish, flushed with pink or

mauve. They meet only in the Woodbush area of

north-eastern Transvaal.

P. fruticosus often occurs together with P. rubro-

punctatus and the two may be confused, especially

in the herbarium. They can, however, almost

invariably be distinguished by the presence, in P.

rubropunctatus, of red or brownish gland-dots on the

underside of the leaves, especially towards the margins,

as well as on the calyx and often on the corolla; in

P. fruticosus the gland-dots are minute, pale to honey-

coloured droplets. There is also a difference in flower

colour: as mentioned above, the flowers of P. rubro-

punctatus are basically white with a flush of pink or

mauve or, rarely, more or less uniformly bluish,

while in P. fruticosus the flowers are bluish-mauve

freely speckled with purple on the upper lip.

The main difficulty in distinguishing between the

two is experienced at high altitudes on Mariepskop,

where both P. rubropunctatus and P. fruticosus occur

in exposed, rocky situations above the forest zone.

Due to the more severe conditions, both species

here exhibit smaller and thicker leaves and more

markedly glandular-hispid pubescence on stems and

leaves, with the result that the diagnostic gland-

dots are not always discernible on herbarium

specimens. A field study in this area would be of value

to determine the degree of intergradation between

the two species.

P. arthropodus Briq. and P. charianthus Briq. were

both described from Rehmann specimens collected in

the Woodbush area of the Transvaal. In recent years

the name “arthropodus” has been wrongly applied in

the National Herbarium to the plants now placed as

P. rubropunctatus (see Compton, FI. Swaz. pp. 66,

157, 1966). A re-examination of both type specimens

revealed that both have the pale-coloured gland-dots

of P. fruticosus so that these two names must be

added to the synonymy of this species, while the

plants with reddish gland-dots must receive a new

name.

31. Plectranthus rehmannii Guerke in Bull. Herb.

Boiss. 6: 553 (1898); Cooke in FI. Cap. 5,1: 274

(1910), partly, excl. Thorncroft 3259; Ross, FI. Natal

305 (1972). Type: Natal, Karkloof, Rehmann 7359

(Z.!, holo. ; K!).

Erect, branched herb or subshrub 60-120 cm tall,

aromatic; branches semi-succulent, somewhat woody

at the base, 4-angled, finely appressed tomentulose

with short, retrorse multicullular hairs and small

subsessile glands. Leaves: petiole 1,5-6 cm long,

appressed tomentulose like the stems; blade ovate to

oblong-ovate, 8-14 cm long, 5-8 cm broad, thickish-

textured, upper surface with sparsely scattered short

multicellular hairs, lower surface paler, reticulate-

veined, hispidulous along the nerves, freely dotted

with orange-brown to dark gland-dots and small

subsessile glands; apex acute to acuminate; base

rounded to truncate, not or shortly attenuate into the

petiole; margin finely and regularly crenate-serrate

with over 50 pairs of teeth, ciliate. Inflorescence

terminal, erect, 25-35 cm long, paniculate with 1-3

pairs of branches near the base; rhachis densely and

finely glandular-tomentulose; bracts ovate-lanceolate,

accuminate, 7-10 mm long, persisting to the early

fruiting stage, later breaking off near the base, finely

glandular-puberulous. Flowers in few-flowered cymes

consisting of a central sessile flower and 2 lateral

branches in each bract; peduncles of lateral cyme-

branches up to 7 mm long, producing 8-16-flowered

verticillasters, verticillasters 1,2-2 cm apart; pedicels

3-6 mm long, glandular-puberulous. Calyx 2,5 mm

422

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

long at flowering stage enlarging to 9 mm long in

fruit and then at right angles to the pedicel, finely

glandular-puberulous with scattered reddish glan-

dots; upper lip suberect, broadly ovate, acute to

obtuse, up to 3 mm long; lower lip sub-equally

4-toothed, 5 mm long, teeth deltoid-subulate,

2-2,5 mm long, the lower pair slightly the longer.

Corolla white, covered with a tomentum of white

multicellular hairs; tube 4 mm long, deflexed and

A clear-cut species with no near relationships,

characterized by the finely crenate-serrate leaves and

the small, saccate-based, white flowers covered with

a white tomentum.

Thorncroft 3259 from Barberton, cited by Cook in

Flora Capensis under P. rehmannii , is P. laxiflorus,

a species which has finely toothed, cordate-based

leaves and a very different corolla shape (see p. 434).

Fig. 33. — Plectranthus rehmannii, Kar-

kloof, Natal {Codd 8587), Xl.

much expanded at the base, 3 mm in depth at the

base, narrowing somewhat towards the throat; upper

lip very short and inconspicuous, 2 mm long, 3 mm

broad, apex bilobed and with two minute lateral

ear-like lobes; lower lip boat-shaped, curved upwards,

4 mm long. Stamens free, up to 2,5 mm long, the

upper pair of anthers usually visible in the throat of

the open corolla. Style slightly exceeding the stamens.

Fig. 33.

A soft shrubby plant with a limited distribution

in the Natal Midlands where it is often locally

common in forest margin scrub.

Natal. — 2930 (Pietermaritzburg): Karkloof ( AC), Rehmann

7359 (K, Z); Mott 3497 (NH); Curry’s Post (-AC), Schlechter

6804 (NH); Hilliard 4852; Shafton, Howick (-AC), Evans sub

NH 19990 (NH); near Lidgetton (-AC), Medley Wood 6313;

13 km N.W. of York (-AD), Codd 8587; Swartkop (-CB),

Medley Wood 10268; Byrne (CC), Medley Wood 3167 (K).

32. Plectranthus swynnertonii S. Moore in J. Linn.

Soc. Bot. 40: 176 (1911). Type: Rhodesia, Chirinda

Forest, Swynnerton 337 (Kl).

Soft herb 40-75 cm tall pleasantly aromatic (“Cress-

like”); branches slender, softly succulent, semi-

prostrate, spreading or erect, finely glandular-

puberulous with some scattered multicellular hairs

and a fringe of longer hairs at the nodes. Leaves:

petiole 4-13 cm long, slender, minutely glandular-

puberulous with occasional multicellular hairs; blade

broadly ovate to subrotund, 5-15 cm long, 4,5-

14 cm broad, very thin textured, dark above with

scattered appressed multicellular hairs, paler and

reticulate-veined below, sparingly strigose on the

veins and freely dotted with sessile yellowish to

brownish gland-dots, often slightly sunken or

flattened; apex obtuse to acute, base truncate to

L. E. CODD

423

subcordate; margin coarsely and deeply serrate-

dentate, teeth 6-16 mm long usually with small

secondary teeth. Inflorescence terminal, simple or

sometimes with a pair of branches at the base;

racemes 6-15 cm long; rhachis finely glandular-

puberulous; bracts broadly ovate to obovate, acumi-

nate, 2-4,5 mm long, persisting to the fruiting stage,

finely glandular with a few fringing multicellular

hairs towards the apex. Flowers in sessile or almost

sessile (l-)3-flowered cymes forming 2-6-flowered

verticillasters; verticillasters 1-3 cm apart; pedicels

3-5 mm long, glandular-puberulous. Calyx 4 mm

long at flowering stage enlarging to 8 mm in fruit,

finely puberulous; upper lip erect, broadly ovate,

abruptly acute, slightly decurrent on the tube, 2,5 mm

horizontal to recurved. Stamens free, the lower pair

the longer, up to 6 mm long. Style horizontal, exserted

by 6 mm. Fig. 34.

A ground-layer herb to soft shrub in moist forest,

usually on humus-rich well-drained soil, found in the

mountain forests of northern and north-eastern

Transvaal, and in similar situations in eastern

Rhodesia.

Transvaal. — 2230 (Messina): Entabeni Forest Station,

Soutpansberg (-CC), Gatpin 10249 ; Codd 8388. 2330 (Tzaneen):

Westfalia Estate near Duiwelskloof (-CA), Scheepers 947;

Bos 1168; Woodbush Forest, Debingini Falls (-CC), Codd

9420; Bruce & Kies 79; 4 km N.W. of Magoebaskloof Hotel

(-CC), Leistner, Thom & Gilliam 3284.

Fig. 34. — Plectranthus swynnertonii,

Soutpansberg ( Codd 8388), xl.

long; lower lip 4-toothed, 4-5 mm long at maturity,

the lower pair much the longer, linear-subulate,

median pair deltoid-subulate. Corolla white, flushed

with mauve-pink and with a few purple dots on the

upper corolla lip inside, sparingly glandular and with

an occasional yellow gland-dot outside and fringed

with white hairs; tube 4-6 mm long, scarcely deflexed,

expanding abruptly in the calyx forming a somewhat

saccate base, narrowing towards the throat; upper lip

erect, 4-6 mm long and up to 6 mm broad, bilobed

at the apex and with 2 prominent, rounded ear-like

lateral lobes; lower lip boat-shaped, 4-5 mm long.

Among South African species with persistent

bracts, P. swynnertonii can readily be distinguished

by its almost round leaves, which are very thin in

texture, somewhat cordate at the base and with

coarsely dentate margins, the teeth being 1-1,5 cm

long with small secondary teeth. The gland-dots on the

underside of the leaf are also distinct, being yellowish

to pale brown, often rather flattened or sunken.

33. Plectranthus dolichopodus Briq. in Bull. Herb.

Boiss. ser. 2,3: 1069 (1903); Cook in FI. Cap. 5,1:

287 (1910), partly, excl. Flanagan 740; Ross, FI.

424

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Natal 305 (1972). Type: Natal, Karkloof, Rehmann

7383 (Z!, holo; PRE, photo.).

P. cooperi T. Cooke in Kew Bull. 1909: 377 (1909), partly as

to Wood 1843 and Gerrard 1673.

Erect or straggling herb 25 cm-1 m tall, sparingly

branched, arising from a mass of fine fibrous roots;

branches softly semi-succulent to slightly woody,

4-angled, glandular-pilose to hispidulous, with long

multicellular purple-striped spreading to antrorse

hairs, short stiff unicellular hairs and gland-tipped

hairs. Leaves: petiole slender 2-6 cm long, sparsely

glandular-pilose; blade broadly ovate to ovate-deltoid,

3-10 cm long, 2,5-8 cm broad, chartaceous, upper

and lower surfaces subglabrous with sparsely scattered

multicellular hairs and minute colourless gland-dots,

nerves inconspicuous above, slightly raised below;

apex acute to abruptly acute, base broadly truncate,

shortly attenuate for 3-6 mm into the petiole; margin

coarsely crenate-dentate with 6-12 pairs of teeth,

sparingly ciliate. Inflorescence terminal, erect, simple

or paniculate 10-20 cm long; with 1 or 2 pairs of

branches near the base; rhachis glandular-hispidulous;

bracts obovate, acuminate 1 , 5-2 mm long, persistent

beyond the fruiting stage, sparingly hispidulous,

ciliate. Flowers in sessile, usually 3-flowered cymes

(often fewer towards the apex), forming 2-6-flowered

verticillasters, verticillasters 1-3 cm apart; pedicels

6-10 mm long, glandular-hispidulous, persistent, often

with the basal part of the calyx still attached. Calyx

purple tinged; 3 mm long at flowering stage, enlarging

to 6 mm long in fruit, sparsely glandular-puberulous;

upper lip erect, ovate, acute to acuminate, up to

2,5 mm long: lower lip 4-toothed, 3 mm long, the

lower pair slightly the longer, linear-subulate, 2 mm

long, the median pair deltoid-subulate, 1 ,5 mm long.

Corolla sky-blue to deep blue, sparingly pubescent

without; tube 5 mm long, deflexed and expanding

in the calyx, 1,5 mm in depth at the base expanding

to 2,5 mm at the throat; upper lip very short, 1,5-

2 mm long suberect, 3 mm broad, bilobed at the

apex; with 2 small ear-like lateral lobes; lower lip

boat-shaped, curved upwards, 4-5 mm long, enclosing

the stamens and style. Stamens free, up to 4 mm long,

curved upwards in the keel. Style exceeding the

stamens by 1 mm. Fig. 35.

A soft herb of cool forests with a somewhat dis-

junct distribution from Stutterheim District in the

Cape, Natal Midlands and foothills of the Drakens-

berg, and the Mariepskop-Woodbush area in the

Transvaal.

Transvaal. — 2330 (Tzaneen): Westfalia Estate (-CA),

Scheepers 923; Woodbush Forest (-CC), Schlechter 4722;

Codd 9427 ; Letty 458; Taylor 695, New Agatha Forest Reserve,

Midler & Scheepers 174. 2430 (Pilgrim’s Rest): The Downs

(-AA), Codd 7752; Cyprus Point (-AB), Junod 4328; Marieps-

kop ( DB), Van Son sub TRV 31563; sub TRV 31651; Codd

7862; Meeuse 9995; Van der Schijff 4376; 4452; Bos 1056.

Natal.- Without locality: Gerrard 1673 (K). 2731 (Louws-

burg): Ngome Forest (-CD), Gerstner 4513; Codd 9593. 2829

(Harrismith): Cathedral Peak (-CC), Killick 1676; 1965. 2830

(Dundee): Qudeni Forest (-DB), Gerstner 6711 ; Fisher 803;

Codd 9674; Krantzkop ( DD), Dyer 4356; Edwards 2097;

Strey 4232. 2929 (Underberg): Cathkin Park ( AB), Galpin

11839; Xumeni Forest ( DD), Moll 3001; Bulwer (-DD),

McClean 158. 2930 (Pietermaritzburg): Lidgetton (-AC).

Mogg 6684; 13 km N.W. of York, Codd 8581; Byrne (-CC),

Medley Wood 1843.

Cape. — 3029 (Kokstad): Ingeli Pass (-DA), Acocks 13564.

3227 (Stutterheim): near Toise River, Flanagan 2270.

The last mentioned specimen, Flanagan 2270,

represents a considerable extension of the distribution

range and confirmation of its occurrence in this

area is desirable. There is, however, no sound reason

for doubting the authenticity of the information on

the label.

Although herbarium specimens are not always

easy to identify, a mention on the collector's label of

“sky-blue flowers” should be sufficient to eliminate

the possibility of almost any other South African

species. Other characteristics are the purple-tinged

calyx, the membranous, subglabrous leaves on long

slender petioles and the bunch of thin, fibrous roots.

The type, Rehmann 7383 in Z, is a rather scrappy

immature specimen with only one half-open flower

in a capsule, but the leaf texture, petioles and

pubescence, together with the fibrous roots enable it

to be linked without doubt to the modern gatherings

cited above.

Cooke, in FI. Cap., misunderstood this species.

Flanagan 740, which he cites as P. dolichopodus is, in

fact, a member of the P. tomentosus-P . madagas-

cariensis complex, while the two specimens Wood

1843 and Gerrard 1673, included by Cooke in P.

cooperi are, in fact P. dolichopodus.

Chromosome number 2n =28 (De Wet, 1958).

P. dolichopodus has no close relative in South

Africa but is closely related to material named

P. sylvestris Guerke (1894) at Kew. The type of P.

sy/vestris , Volkens 965 from Mt Kilimanjaro, has

not been seen, but the specimens in the P. sylvestris

cover differ only slightly in the leaves being more

acuminate at the apex and distinctly cuneate at the

base. It seems also to be a more robust plant than

P. dolichopodus with larger flowers. However, Giirke

relates P. sylvestris to a different group, namely,

Section Coleoides Benth.

Another apparently related species is P. sphaero-

phyllus Bak. but this species seems less close to P.

dolichopodus than is P. sylvestris.

34. Plectranthus zuluensis T. Cooke in Kew Bull.

1909: 379 (1909); FI. Cap. 5,1: 281 (1910); E. A.

Bruce in Flow. PI. Afr. 28: t. 1110 (1951); Codd in

Mitt. Bot. Munchen 10: 247 (1971); Ross, FI. Natal

305 (1972). Type: Natal, Gerrard 1675 (K!, holo.).

Erect soft shrub 1-2 m tall, freely branched:

branches ascending, softly woody below, semi-

succulent above, obtusely 4-angled, shortly and finely

tomentulose, especially the younger parts which are

whitish and velvety to the touch, minutely dotted

with dark gland-dots. Leaves : petiole 2,5-6 cm long,

finely velvety pubescent; blade ovate to broadly ovate,

3-7 cm long, 2, 5-5, 5 cm broad, softly semi-succulent,

drying thin-textured, thinly and shortly pubescent

on both surfaces, denser on nerves and margins,

gland-dots on lower surface colourless, nerves slightly

raised below; apex acute, base truncate to shortly

L. E. CODD

425

Fig. 35. — Plectranthus dolichopodus,

Karkloof, Natal ( Codd 8581), xl.

cuneate; margin regularly and coarsely crenate with

7-12 pairs of teeth. Inflorescence terminal, simple or

rarely branched near the base, 4-8 cm long; rhachis

finely glandular-tomentulose; bracts ovate to lanceo-

late, 2-4 mm long, persisting beyond the fruiting

stage. Flowers in sessile 3-flowered cymes forming

6-flowered verticillasters; verticillasters 5-12 mm

apart; pedicels 2-5 mm long, glandular-tomentulose.

Calyx 3 mm long at flowering, enlarging to 7 mm long

in fruit, purple-tinged, glandular-hirsute, becoming

subglabrous; upper lip suberect, ovate, obtuse, 2 mm

long, slightly decurrent on the tube; lower lip

4-toothed, up to 4 mm long, the lower pair the longer,

linear-subulate, 2 mm long, the lateral deltoid 0,5-1

mm long. Corolla pale blue-mauve to almost white

with usually six rows of mauve dots on the upper lip;

tube 5-10 mm long, laterally compressed, expanding

in the calyx forming a saccate base 2,5-4 mm deep

at the base, narrowing to 2 mm at the throat; upper

lip erect, 5-6 mm long, 4-5 mm broad, apex emar-

ginate and with 2 obscure lateral lobes; lower lip

shallowly boat-shaped, 5-6 mm long, at first

426

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 36. — Plectranthus zuluensis, Nkan-

dla Forest (Codd 9691), X 1.

horizontal soon deflexed. Stamens free at the base,

only the lower pair fertile with filaments 5-7 mm

long, upper pair abortive, reduced to stami nodes,

1-2 mm long. Style finally exserted by 5 mm. Fig. 36.

Restricted to Natal where it occurs in semi-coastal

forests, often common along streams, from Port

Shepstone District to Ngoye and Nkandla Forests

in the north.

Natal.- Without precise locality: Gerrard 1675 (K). 2830

(Dundee): near Krantzkop Mt. ( DD), Burn 3043 ; Dyer 4355.

2831 (Nkandla): Nkandla Forest (-CA), Wylie sub NH 9382

(NH); Codd 6969; 21 km S. of Nkandla (-CA), Codd 9691;

Eshowe (-CD), Thode A 1255; Lawn 113 (NH); between

Nkwaleni and Empangeni (-DA) or ( DB), Gerstner sub NH

22235 (NH); Ngoye Forest (-DC), Wylie sub NH 12830 (NH);

Huntley 234; Wells & Edwards 84; Strey 6110; Venter 2183.

2930 (Pietermaritzburg): Noodsberg (-BD), Medley Wood

5305 (NH); Little Noodsberg (-BD), Strey 6242; Nkutu Falls

(-DD), Johnson 1491. 2931 (Stanger): near Mapumulo (-AA),

Strey 9467; Moll 2972. 3030 (Port Shepstone): Umgai (-AD),

Strey 10582; near Dumisa (-AD), Hilliard & Burn 3818; Ifafa

(-BC), Rudatis 323; Mgai Farm (-BC), Nicholson 1154; Oribi

Gorge (-CA), Nicholson 1010; 1044 ; 1046; Uvongo River

(-CD), Nicholson 1024; Izotsha ( CD), Strey 8087 (NH).

P. zuluensis possesses the unusual characteristic of

having only two fertile stamens (the lower pair)

while the upper two are reduced to short staminodes.

In all other respects the species lits into the genus

Plectranthus. The species can usually be recognized

also by the softly velvety young branchlets.

The holotype, Gerrard 1675 from Natal, without

details of locality, is in Kew Herbarium together

with an immature specimen of the same species,

Gerrard 1229.

Chromosome number: 2n — 28 (De Wet, 1958).

35. Plectranthus saccatus Benth. in E. Mey., Comm.

227 (1837). Type: Cape, “Omsamwubo” (Umzim-

vubu River, near Port St Johns), Drege (K, in Herb.

Benth.!, lecto.; Gl; PI; SI).

Erect to spreading soft shrub 50-120 cm tall, freely

branched; branches woody below, semi-succulent

above, purple-tinged, somewhat 4-angled, minutely

glandular-puberulous with occasional longer multi-

cellular hairs at the nodes. Leaves : petiole 1,5-5 cm

long, glandular-puberulous, purple-tinged; blade

broadly ovate to ovate-deltoid, 2-7 cm long, 1,5-

5 cm broad, softly semi-succulent, drying thin-tex-

tured, sparingly glandular-puberulous to subglabrous,

gland-dots colourless, nerves raised below; apex

acute; base truncate to obtuse or shortly cuneate into

the petiole; margin coarsely dentate with 3-6 pairs of

teeth, purple-tinged and fringed with short hairs.

L E. CODD

427

Inflorescence a short terminal raceme 5-12 cm long

with relatively few but large flowers; rhachis densely

and shortly glandular-puberulous; bracts lanceolate,

acuminate, 1,5-2 mm long, persisting beyond the

fruiting stage. Flowers very large, in sessile 1-3-

flowered cymes, forming 2-6-flowered verticillasters;

verticillasters 1-2 cm apart; pedicels 4-6 mm long,

slightly thickened and apparently articulated near the

apex, purple-tinged, minutely and densely glandular-

puberulous. Calyx 3 mm long at flowering, enlarging

to 8 mm long in fruit, minutely puberulous to sub-

glabrous, with or without a few long multicellar hairs

on the margins of the teeth; upper lip erect, ovate,

obtuse, up to 2,5 mm long, slightly decurrent on the

tube; lower lip 4-toothed, up to 5 mm long, the

lower pair the longer, linear-subulate, 3 mm long,

the lateral pair deltoid-subulate, 2 mm long. Corolla

large, mauve to pale blue or, rarely, white, minutely

puberulous, the upper lip with or without purple

blotches; tube 0,8- 1,5 cm (typical) or 2-2,6 cm

long (var. longitubas), laterally compressed, expanding

in the calyx forming a saccate base 4-6 mm deep

at the base, parallel-sided or narrowing slightly

towards the throat; upper lip erect, 1-1,6 cm long

and equally broad, notched at the apex and with

2 small pendulous lateral lobes; lower lip boat-shaped,

8-12 mm long, at first horizontal, later deflexed.

Stamens exserted or recurved within the lower lip of

the corolla, subequal, 8-10 mm long, free or shortly

connate at the base. Style finally exserted by about

1 cm.

Occurs in forests or shady, rocky places not far

from the coast, from the Kentani District in the

eastern Cape Province to the Ingwavuma District in

northern Natal.

P. saccatus may be distinguished from all other

species by the large corolla, the upper lip, which is

1-1,6 cm long and equally broad, far exceeding

that of any other species. There is a good deal of

variation in corolla size and shape, the plants occurring

at the extreme north end of the range at Gwalaweni

Forest having exceptionally long corolla tubes. In

view of the marked discontinuity in corolla-tube

length shown by these northern plants, they are

separated as a distinct variety.

Key to varieties

Corolla tube 0,8-1, 5 cm long (a) var. saccatus

Corolla tube 2-2,6 cm long (b) var. longitubus

(a) var. saccatus.

Plectranthus saccatus Benth. in E. Mey., Comm. 227 (1837);

in DC., Prodr. 12: 62 (1848); Wood & Evans, Natal PI. 1 : t. 85

(1899); Hook. f. in Bot. Mag. t. 7841 (1902); Cooke in FI. Cap.

5, 1 : 273 (1910); Batten & Bokelman, Wild Flow. E. Cape 127,

t. 101 (1967); Codd in Flow. PI. Afr. 41 : t. 1601 (1970)- Ross'

FI. Natal 305 (1972).

The leaf-blade tends to be smaller (2-5 cm long)

and the corolla tube is distinctly shorter (0,8-1 ,5 cm

long) than in var. longitubus. Fig. 37.

Fig. 37. — Plectranthus saccatus var.

saccatus, Port St. Johns ( Codd

9298), xl.

428

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Distribution as for the species with the exception

of the Gwalaweni Forest at the south end of the

Lebombo Range.

Natal. — Without locality: Gerrard s.n. (K); Sanderson

s.n. (K). 2830 (Dundee): Qudeni Forest (-DB), De Winter 8200.

2831 (Nkandla): Nkandla Forest (-CA), Wells 2490; Hlinza

Forest (-CD), Moll 4990; near Eshowe (-CD), Codd 9651 ;

Ngoye Forest (-DC), Wells & Edwards 80. 2832 (Mtubatuba):

Hluhluwe Game Reserve (-AA), Ward 3061. 2930 (Pieter-

maritzburg): Inanda (-DB), Medley Wood 323 (K, NH); near

Durban (-DD), Medley Wood 7382; 10208; Rogers 15005.

3030 (Port Shepstone): Oribi Gorge (-CA), Nicholson 1056;

Izingolweni (-CC), Medley Wood 3037 (K, NH); Izotsha

(-CD), Strey 8084. 3130 (Port Edward): Beacon Hill (-AA),

Strey 7223; 10929.

Cape. — 3129 (Port St. Johns): Mkambati Leper Institute

(-BD), Marais 968; Goss Point (-BD), Strey 10148; 10 km W.

of Port St. Johns (-CB), Codd 9298; Isinuka (-DA), Galpin

2840; 11460; Flanagan 2501; near St. Johns River, Drege s.n.

(G, K, MO, P, S). 3227 (Stutterheim): Tsomo River (-BA?),

Barber s.n. (K). 3228 (Butterworth): near Kentani (-AD),

Pegler 338.

P. saccatus var. saccatus shows some variation

in corolla colour from mauve to blue, while the

degree of marking of the upper lip varies from pale

to deep purple; De Winter 8200 from Qudeni Forest is

particularly striking. The plant collected on the

Mkambati Leper Institute ( Marais 968) is distinctly

more woody with more leathery leaves than other

members of the species.

Chromosome number 2n =28 (De Wet, 1958).

The species is based on a gathering by Drege near

the “Osamwubo” (Umzimvubu or St. Johns) River

and several specimens have been seen. The specimen

ex Herb. Bentham in Kew is selected as the lectotype.

(b) var. longituhus Codd , var. nov. a var. saccato

tubo corollae 2-2,6 cm longo differt.

Type: 2732 (Ubombo); Ingwavuma District, Gwala-

weni Forest (-AC), Edwards 2930 (PRE, holo.).

Leaves tend to be larger, 3-7 x 2,5-5 cm, as

against 2-5x1, 5-4 cm in var. saccatus ; the corolla

tube is distinctly longer (2-2,6 cm long), but tends

to be narrower (4-5 mm deep at the base as against

5-6 mm in var. saccatus). Fig. 38.

Recorded as yet only from the Gwalaweni Forest

at the southern end of the Lebombo Range.

Natal. — 2732 (Ubombo): Gwalaweni Forest, Edwards 2930;

Vahrmeijer & Hardy 1699; Vahrmeijer 1913.

The markedly longer corolla tubes shown by these

plants appears to warrant their separation as a

distinct variety. The leaves also tend to be larger

and the corolla tube tends to be narrower, but

there is overlapping in these characters. Although the

flower colour is normally pale blue-mauve, occasional

white-flowered plants are known from this locality.

36. Plectranthus hilliardiae Codd in Bothalia 1 1 :

282 (1974). Type: Natal, near Umtamvuna River,

Hilliard & Burt 6767 (PRE!, holo.; NU).

Erect perennial herb 30-40 cm tall, aromatic; stems

semi-succulent, 4-angled, sparingly branched, shortly

antrorse pilose and gland-dotted with occasional

long multicellular hairs and tufts of long stiff multi-

cellular purple-tinted hairs at the nodes. Leaves:

petiole 1 ,5-3,5 cm long, densely appressed glandular

pilose similar to the stems; blade broadly elliptical

to obovate-elliptical, 5,5-9 cm long, 4-6 cm broad,

semi-fleshy when fresh, drying membranous, dark

green and sparingly strigose above, under surface

paler, reticulate-veined, sparingly strigose on the

nerves and freely dotted with small transparent sessile

glands; apex acute to obtuse; base cuneate, attenuate

into the petiole; margin ciliate, shallowly crenate —

dentate above the middle. Inflorescence erect, 8-15 cm

Fig. 38. — Plectranthus saccatus var. longitubus, Ubombo

District, northern Natal ( Vahrmeijer & Hardy 1699), X 1.

long, simple or with a pair of branches near the base;

rhachis glandular-tomentulose with multicellular and

short gland-tipped hairs; bracts linear-lanceolate,

acuminate, 3 mm long, persisting to the fruiting stage,

with gland dots and long multicellular hairs. Flowers

in sessile, usually 3-flowered cymes forming de-

flowered verticillasters, verticillasters 1,5-2, 5 cm

apart; pedicels 4-5 mm long, finely glandular-

hispidulous. Calyx 4-5 mm long at flowering stage,

enlarging in fruit, glandular-hispidulous near the

base, subglabrous above; upper lip ovate, erect,

not decurrent on the tube, 2,5 mm long; lower lip

4-toothed, the lower pair longer than the median;

teeth linear- to deltoid-subulate. Corolla pale bluish,

purple-flecked on the lobes, minutely glandular-

puberulous; tube 2, 3-2, 7 cm long, slight deflexed

and expanding abruptly near the calyx to 4 mm

wide and parallel-sided or narrowing to 3 mm wide

at the throat; upper lip erect, 5-6 mm long and

equally broad, bilobed at the apex and with two

lateral ear-like lobes; lower lip shallowly boat-shaped,

4 mm long, horizontal or deflexed. Stamens free, the

lower pair the longer, up to 8 mm long, horizontal to

L. E CODE)

429

Fig. 39. — Plectranthus hilliardiae, Umtamvuna River, Natal

(Hilliard & Burtt 6767, PRE, holotype), X%.

recurved. Style horizontal to ascending, exserted by

8-10 mm. Fig. 39.

A softly semi-succulent herb which grows among

rocks near and in the margins of scrub forest, known

only from a restricted area on each side of the

Urrtamvuna River in southern Natal and Pondoland.

The flowering time is February-March.

Natal. — 3130 (Port Edward): Umtamvuna River (-AA),

Hilliard & Burtt 6767 ; heights overlooking Umtamvuna River

(-AA), Nicholson s.n.

Cape. — 3130 (Port Edward): Pondoland, Impahlane River,

Thode 3187 (STE).

The species is named in honour of Dr. Olive M.

Hilliard of Natal University, who, together with

Mr B. L. Burtt of the Royal Botanic Gardens,

Edinburgh, has made such valuable contributions to

our knowledge of the South African flora.

Its nearest affinity is P. ambiguus which it resembles

in the long corolla tube. However, unlike P. ambiguus,

the corolla tube of P. hilliardiae expands at the base

to 4 mm deep and remains this width (or narrows

slightly to 3 mm) at the throat. The tube is also

slightly deflexed at the base, while the colour is a

pale blue. The leaves of P. hilliardiae tend to be

more softly succulent in texture and more elliptical

in shape with a distinctly cuneate base than in P.

ambiguus.

37. Plectranthus ambiguus (Bol.) Codd in Bothalia

8:159 (1964); Batten & Bokelmann, Wild Flow. E.

Cape 125 (1966); Ross, FI. Natal 305 (1972). Type:

near Grahamstown, MacOwan (BOL!, holo. ; SAM!).

P. colorants E. Mey. ex Benth. in E. Mey., Comm. 228 (1837);

Cooke in FI. Cap. 5, I : 279 (1910), partly, excl. Gerrard 1671 ;

Wood 3036; 3977; Gueinzius s.n.; non Don (1825). Syntypes:

between Umgazana and Umzimvubu Rivers, see note below,

Drege a (BM!, K!; MO!; Hb. Drege, P ! ; S ! ) ; between Umzim-

kulu and Umkomaas Rivers, Drege h (Hb. Drege, P!). P.

dregei Codd in Flow. PI. Afr. 32: t. 1244 (1957). Syntypes: as for

P. coloratus E. Mey. ex Benth.

Erect to decumbent herb or soft shrublet 40-120 cm

tall, branching from near the base; branches ascending

to spreading, softly woody below, semi-succulent

above, 4-angled, shortly and densely to sparingly

strigose-pubescent, with short antrorse multicellular

hairs and gland-dots and tufts of long multicellular

hairs at the nodes. Leaves very variable in size,

texture and pubescence; petiole 1-7 cm long antrorse

pubescent; blade ovate to broadly ovate, 2,5-12 cm

long, 2-9 cm broad, thin to thickish and slightly

rugose in texture, thinly pubescent to subglabrous

above, more dense especially on the nerves below

with longish multicellular hairs and numerous sessile

honey-coloured to brown simple gland-dots; apex

obtuse to acute; base abruptly cuneate to somewhat

decurrent on the petiole; margin shallowly crenate

with 6-12 pairs of teeth. Inflorescence a congested

terminal panicle, rarely simple, 4-17 cm long, usually

with 1-3 pairs of branches near the base; rhachis

densely to sparingly antrorse glandular-pubescent;

bracts obovate, 2-5,5 mm long, persisting to beyond

the fruiting stage. Flowers in sessile 3-flowered cymes

(often fewer on the branches of the inflorescence)

forming usually 6-flowered verticillasters; verti-

cillasters 2-6 mm apart; pedicels 3-5 mm long,

glandular-tomentulose. Calyx 4 mm long at flowering

enlarging to 8 mm long in fruit, usually suffused

with purple, glandular-hispidulous (less so with age);

upper lip erect, ovate to ovate-oblong, acute, up to

3 mm long; lower lip 4-toothed, up to 5 mm long, the

lower pair the longer, linear-subulate, 2 mm long,

the lateral deltoid-subulate, 1-1,5 mm long. Corolla

violet to purple; tube long and straight, 2-2,5 cm

long, not expanded and 1 mm deep at the base,

almost parallel sided, expanding to 2 mm deep at the

throat; upper lip erect, 4-5 mm long, 4 mm broad,

broadly bilobed at the apex and with 2 obscure

lateral ear-like lobes; lower lip shallowly boat-

shaped, 3-5 mm long, horizontal to deflexed. Stamens

free at the base didynamous, horizontal or recurved

in the lower lip, upper pair 4-5 mm long, lower pair

6-7 mm long. Style finally exserted by 6-7 mm.

Fig. 40.

Found in forest margins and on shady, rocky

hillsides from the Albany and Bathurst Districts of

the Cape along the eastern Cape, usually not far

from the coast, to the Ngoye Forest in Natal.

Natal. — 2831 (Nkandla): near Eshowe (-CD), Wylie sub

NH 7993 (NH); Ngoye Forest (-DC), Garland s.n.. Cooper 117.

2930 (Pietermaritzburg): Inanda (-CB), Medley Wood 480

(K, NH); Durban (-DD), Franks sub Wood 11809; Isipingo

(-DD), Ward 820; 4903. 3030 (Port Shepstone): without exact

locality, Dimock-Brown 436; Dumisa (-AD), Rudatis 314; 16

km W. of Umzinto (-BC), Codd 8574; Oribi Gorge (-CA),

Codd 9351; Nicholson 1041; 1048; Mehlomnyama (-CB),

Acocks 13311; Shelley Beach (-CD), Strey 8437 ; Uvongo (-CD),

Liebenberg 8082; Nicholson 1021; Izotsha Falls (-CD), Strev

7430; 8086.

Cape. — (or Natal) between Umzimkulu and Umkomaas

Rivers, Drege b (P). 3129 (Port St. Johns): 16 kmW of Port St.

Johns (-CB), Codd 9296; between Umgazana and Umzimvubu

Rivers (-DA), Drege a (BM, K, MO, P, S). 3227 (Stutterheim):

Keiskammahoek (-CA), Sidey 617 ; Killick 893; Amatola Mts.

(-CA), Erens 2229. Pirie Forest (-CD), Sim s.n.; Galpin 3280;

King William's Town (-CD), Barber s.n. (K); Sim 19590, • Kei

430

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SCUTHERN AFRICA

Fig. 40. — Plectranthus ambiguus, Umzinto, Natal ( Codd 8574), X 1.

Road (-DA), Ranger 238; near Komga (-DB), Flanagan 738.

3228 (Butterworth): near Kentani LAD), Pegler 907. 3326

(Grahamstown): near Grahamstown (-BC), MacOwan 987

(BOL, SAM); Kowie (-DB), Britten 1418. 3327 (Peddie):

near Peddie (-AA), Sim s.n.; Fort Grey ( BB), Galpin 7826.

P. ambiguus is characterized by the long straight

cylindrical corolla tube which is not saccate near the

base. On the other hand, the leaves show a good deal

of variation in size, texture and amount of pubescence,

tending to be smaller and less pubescent towards the

southern part of the distribution range.

The only other species with a straight, non-saccate

corolla tube is P. ecklonii, a more robust species in

which the corolla tube is slightly shorter (1 ,2-1,5 cm

long) and widens noticeably towards the mouth.

The corolla also tends to be a paler blue in colour.

Its nearest relative is the recently described P.

hilliardiae, in which the corolla tube, although

parallel-sided, expands at the base and is somewhat

deflexed.

L. E. CODD

431

The name P. color atus E. Mey. ex Benth. cannot be

used because of the earlier P. color atus Don (1825).

It is based on two Drege specimens a and b. The

locality of specimen a given in Drege, Zwei Pfl. Doc.

150 (1843), viz. “V, b, 50: zwischen Omtata und

Omsamwubo”, is not quite the same as the more

restricted locality “inter Omgaziana et Omsamwubo”

given by Bentham in E. Mey., Comm. 228 (1837).

The type of P. ambiguus, MacOwan 987, although

rather immature, is recognizable as being the same

species as P. coloratus E. Mey. ex Benth. The new

name given for the latter, P. dregei Codd was,

therefore, not necessary.

Chromosome number 2n =28 (De Wet, 1958, as

“P. coloratus”).

38. Plectranthus ecklonii Benth. in DC., Prodr.

12: 64 (1848); Cooke in FI. Cap. 5,1: 279 (1910);

Batten & Bokelmann, Wild Flow. E. Cape 126, t. 101

(1966); Ross, FI. Natal 305 (1972). Type: Cape,

slopes of Katberg, Ecklon s.n. (K!, holo.).

P. fruticosus sensu Marloth, FI. S. Afr. 3, 2: t. 46 (1932).

Erect soft shrub 70 cm-2,5 m tall, freely branched,

aromatic; branches ascending, woody below, semi-

succulent above, 4-angled, sparingly strigose-

pubescent with short antrorse multicellular hairs and

often with tufts of longer hairs at the nodes, glab-

rescent with age. Leaves: petioles 2-5 cm long with

short multicellular hairs; blade ovate to oblong-

elliptic, 6-17 cm long, 4-10 cm broad, firm-textured,

often slightly rugose, thinly pubescent to subglabrous

above, more dense especially on the nerves below with

appressed multicellular hairs and numerous sessile

reddish-brown gland-dots, rarely gland-dots appa-

rently absent; apex acute; base cuneate to rarely

obtuse; margin conspicuously crenate-dentate with

16-25 pairs of teeth, ciliate. Inflorescence a terminal

panicle 12-25 cm long with usually 2-4 pairs of

branches near the base; rhachis fairly densely to

sparingly provided with multicellular hairs, often

denser at the verticillasters ; bracts ovate to obovate,

4- 8 mm long, cilate, otherwise subglabrous, usually

persisting to the fruiting stage though sometimes shed

earlier. Flowers in sessile usually 3-flowered cymes

producing usually 6-flowered (sometimes fewer)

verticillasters; verticillasters 5-15 mm apart; pedicels

5- 7 mm long sparingly pubescent or puberulous to

subglabrous. Calyx 5 mm long at flowering enlarging

to 10-11 mm long in fruit, often suffused with purple,

glandular-puberulous towards the base with scattered

reddish gland-dots; upper lip erect, ovate, obtuse,

3 mm long; lower lip 4-toothed, up to 5 mm long,

the lower pair the longer, linear-subulate, 3 mm long,

the lateral deltoid-subulate, 2 mm long. Corolla

pale blue or mauve to bluish-purple (rarely white),

puberulous with red gland-dots; tube long and

straight 1,2-1, 5 cm long, not expanded and 1 mm

deep at the base, expanding gradually to 3 mm deep at

the throat; upper lip erect, 5-6 mm long, 4-5 mm

broad, broadly bilobed at the apex and with 2 obscura

lateral ear-like lobes; lower lip shallowly boat-shaped,

4-5 mm long, horizontal to deflexed. Stamens free

at the base, horizontal to recurved in the lower lip,

upper pair 6-8 mm long, the lower 12-16 mm long.

Style finally exserted by 12-16 mm. Fig. 41.

Locally common understorey soft shrub at forest

margins, in scrub forest or wooded stream banks,

from Somerset East and Albany Districts in the Cape,

mainly in the coastal or midland areas, to near Ing-

wavuma in Natal and to Barberton in the Transvaal.

Transvaal. — 2531 (Komatipoort): on road to Shiya-lo-

Ngubo Dam (-CC), Net 367.

Natal. — Without locality: Gerrard 1671 (K); cultivated,

Durban Botanic Garden, Medley Wood 5736. 2732 (Ubombo):

Gwalaweni Forest (-AC), Vahrmeijer & Drijfhout 1916; Mkuze

Pass (-AC), Vahrmeijer 1412. 2831 (Nkandla): Nkandla Forest

(-CA), Codd 6973; Entumeni (-CD), Medley Wood 3977 (K);

Eshowe (-CD), Lawn 188 (NH); Ngoye Forest (-CD), Wells

& Edwards 6; Hlinza Forest (-CD), Moll 4993. 2832 (Mtuba-

tuba): Flluhluwe Game Reserve (-AA), Ward 2177. 2930

(Pietermaritzburg): 8 km N.W. of York, Codd 8578; near

Durban (-DD), Gueinzius s.n. (K); Pinetown (-DD), Wells

I860. 3029 (Kokstad); Weza, Ingeli Forest (-DA), Strey

10898. 3030 (Port Shepstone): Oribi Gorge (-CA), McClean 45;

Nicholson 1059; near St. Michaels-on-Sea, Nicholson 1035.

3130 (Port Edward): Umtamvuna River (-AA), Nicholson 744.

Cape. — 3029 (Kokstad): near Tabankulu (-DC), Story 4206.

3128 (Umtata): Baziya (-CB), Baur 187 (K). 3225 (Somerset

East): Boschberg (-DA), Burchell 3139 (K); Somerset East,

Bowker s.n. (K). 3226 (Fort Beaufort): Slopes of Katberg

(-DA), Ecklon s.n. (K); near Alice (-DD), Henrici 3882.

3227 (Stutterheim): Amatola Mts. (-CA), Erens 2227; Pirie

Forest (-CD), Sim 204; Dontsa Pass (-DA?), Acocks 9742;

near Komga (-DB), Flanagan 737. 3228 (Butterworth): Colly-

Wobbles (-AB), Van Breda 877; near Kentani (-AD), Pegler

376; Codd 9246; 9247; The Haven (-BB), Theron 1461. 3323

(Willowmore): Stones Hill, Schonland 3141. 3326 (Grahams-

town): near Grahamstown (-BC), MacOwan 500.

P. ecklonii is one of the most robust of the South

African species forming a soft shrub up to 2 m or

more tall.

Its closest relative is P. ambiguus but it is readily

distinguished by the shorter corolla tube, paler in

colour, which widens towards the throat, the larger

and laxer inflorescence, and the conspicuous reddish-

brown gland-dots on the undersides of the leaves.

Chromsome number 2n ==28 (De Wet, 1958).

The specimen Ecklon s.n. ex Herb. Benth. in Kew

may be regarded as the holotype.

39. Plectranthus petiolaris E. Mey. ex Benth. in

E. Mey., Comm. 228 (1837); in DC., Prodr. 12: 66

(1848); Cooke in FI. Cap. 5,1: 272 (1910). Type:

Cape, between “Omtata et Omsamwubo”, Drege

(K!, numbered 4773b, lecto.).

P. kuntzei Guerke in Kuntze, Rev. Gen. 3, 2: 260 (1898);

Cooke, l.c. 277 (1910). Type: Natal, Clairmont, Kuntze s.n.

(K!).

Straggly branching herb up to 1 m tall; branches

ascending to spreading, semi-succulent, 4-angled,

pubescence a fairly dense mixture of short gland-

tipped hairs, long retrorse multicellular hairs and

thick broad-based hairs with tufts of longer hairs at the

nodes. Leaves very variable in size; petiole 2-15 cm

long, glandular retrorse pubescent as for the stems;

blade broadly ovate-deltoid, 4-14 cm long, 3,5-

1 1 cm broad, thin-textured, thinly and shortly strigose

above, denser especially on the nerves below and

with minute sessile colourless gland-dots; apex

obtuse to acute; base truncate to subcordate; margin

coarsely crenate-dentate with 6-12 pairs of teeth,

the large teeth often again minutely toothed.

Inflorescence slender, terminal, 10-25 cm long, simple

or with 1-2 pairs of branches near the base; rhachis

glandular-puberulous, sometimes with intermixed

multicellular hairs; bracts broadly ovate to sub-

rotund, ciliate, 2-3 mm long, persisting to fruiting

stage. Flowers in sessile 1 -3-flowered cymes forming

2-6-flowered verticillasters; verticillasters 1-3 cm

apart; pedicels 4-7 mm long, glandular-puberulous.

Calyx 2 mm long at flowering, enlarging to 8 mm

432

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Fig. 41.— Plectranthus ecklonii, Tabankulu, eastern Cape Province (Story 4206), x 1.

long in fruit, glandular-puberulous especially towards

the base; upper lip ovate, suberect, obtuse, up to

3 mm long, decurrent on the tube; lower lip -toothed,

up to 4 mm long, the lower pair slightly the longer,

lanceolate-subulate, 2 mm long, the lateral deltoid,

1,5 mm long. Corolla deep violet-purple, often with

bluish lips; tube not expanded at the base, 0,75 mm

deep and ascending for 3 mm then deflexed an4

starting to expand for 4-5 mm, reaching 3 mm deep

at the throat; upper lip erect, 6-8 mm long, 4 mm

broad, deeply notched at the apex and with 2 obscure

lateral ear-like lobes; lower lip shallowly boat-shaped,

7-9 mm long, horizontal. Stamens free at the base,

curved within the lower lip, subequal, 4-5 mm long.

Style scarcely exceeding the stamens, usually re-

maining within the lower corolla lip. Fig. 42.

L. E. CODD

433

Fig. 42. — Plectranthus petio-

laris, Port St. Johns ( Codd

9295), xl.

Found on shady damp forest floors and rocky

hillsides from the Port St. Johns area, usually not

far from the coast, to Ngoye and the forests of the

Makatini flats in northern Natal.

» ?,ATal-T:2732 (Ubombo): near Mbazwane (-BC), De Winte

i W*riei'jer 8543; Sordwana Bay, Vahrmeijer & Tolken 3Ji

2°3' (Nkandla): Ngoye Forest (-DC), Wells & Edwards 8(

2832 (Mtubatuba): Baheni Stream (-AB?), Ward 3049; Duki

duku(-AC), Strey 5594. 2930 (Pietermaritzburg): Camperdowr

(-DA), Franks sub NH 12832 (NH); lnanda, Groenberg (-DB

Johnson 1313; Ismont (-DC), Strey 8400 (NH); Port Nata

Orege c (P); Clairmont (-DD), Kuntze s.n. (K); Berea, Durba

(-DD), Medley Wood 3390; 5754 (K, NH); Burman Bus)

f"uJ Huntley 112 (NH). 3030 (Port Shepstone

Ifafa (-BC), Rudatis 339; Oribi Gorge (-CA), Codd 9341

A icholson 1047; 1054; Wichman's farm (-CD), Strey 8317.

Cape.— 3029 (Kokstad): Mtentu Bridge (-CD), Strey 1Q649

(NH). 3129 (Port St. Johns): between Umtata and Umzimvubu

Rivers (-CB?), Drege b (MO, P, S); 4773b (K); 16 km W. of

Port St. Johns (-CB), Codd 9295.

There are only two South African Plectranthus

species which have the characteristic corolla shape

resembling a miniature “Dutchman’s pipe”, namely

P. petiolaris and P. laxiflorus (see below), but the

two are easily distinguished. In P. laxiflorus the

leaves are finely and regularly crenate in contrast to

the coarsely and irregularly dentate leaves of P.

petiolaris, while the corolla of P. laxiflorus is white

with faint mauve stripes on the upper lip as against the

violet purple corolla of P. petiolaris. One of the most

434

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

useful distinguishing characters is that in P. laxiflorus

the cymes are nearly always shortly pedunculate

while in P. petiolaris they are sessile.

Three Drege gatherings are cited with the original

description: (a) “prope Sparrbosch”; (b) “inter

Omtata et Omsamwubo” and (c) “Port Natal”.

Specimens of Drege (b) have been seen in several

herbaria (K, MO, P, S) and the one in Kew,

numbered 4773 b, is selected as the lectotype. A

specimen of Drege (c) has been seen in the Drege

Herbarium in Paris, but no specimen of _ Drege (a)

has been encountered. Judging from the locality,

which is near Swellendam, it is most unlikely that the

specimen is P. petiolaris. The only species known to

occur in this area is P. fruticosus. It seems probable

that the specimens sent to Bentham, no doubt

annotated with E. Mey. manuscript names, were

returned to E. Meyer and their subsequent fate has

been the subject of speculation.

40. Plectranthus laxiflorus Benth. in E. Mey.,

Comm. 228 (1837); Drege, Zwei Doc. 145, 149, 157

(1843); Benth. in DC., Prodr. 12: 63 (1848); Cooke

in FI. Cap. 5, 1: 276 (1910); Hulme, Wild Flow.

Natal t. 26, f.2 (1954); Compton, FI. Swaz. 67, 157

(1966); Van der Schijfif, Check List Kruger Nat.

Park 82 (1969); Ross, FI. Natal 305 (1972). Type:

between Umzimkulu and Umkomaas Rivers, Drege

(ex Herb. Benth. numbered 3586, K!, lecto.).

P. hylophilus sensu Cooke in FI. Cap. 5, 1 : 277 (1910).

Germanea laxiflora (Benth) Hiern, Cat. Afr. PI. Welw.

1,4: 861 (1900)

Erect to spreading aromatic herb, freely branched,

70-150 cm. tall; branches ascending to spreading,

often rooting at the nodes, herbaceous, 4-angled,

sparingly to densely tomentose especially towards

the nodes with long retrorse white multicellular hairs,

often with short gland-tipped hairs intermingled.

Leaves: petiole 2,5-8 cm long, tomentose; blade

broadly ovate-deltoid, 6-10 cm long, 4-6 cm broad,

thin-textured to somewhat rugose, thinly appressed

tomentulose above, denser especially on the raised

nerves below, usually with reddish gland-dots

(sometime absent); apex acute to acuminate; base

cordate; margin regularly crenate-dentate with 14-20

pairs of teeth. Inflorescence simple or laxly branched

12-35 cm long with 1-3 pairs of branches near the

base; rhachis glandular-tomentose with long spreading

multicellular hairs and short gland-tipped hairs;

bracts ovate to ovate-lanceolate 4-6 mm long,

persisting beyond the fruiting stage. Flowers in

usually 3-flowered pedunculate (rarely subsessile)

cymes forming usually 6-flowered verticillasters ;

verticillasters 1,5-2, 5 cm apart; penduncles about

5 mm long; pedicels 4-6 mm long, glandular-

tomentulose. Calyx 2,5 mm long at flowering

enlarging to 7 mm long in fruit, glandular-tomentulose

especially towards the base and with scattered red

gland-dots; upper lip scarcely larger than the other

teeth, suberect, ovate-lanceolate, 2,5 mm long;

lower lip 4-toothed, up to 3,5 mm long the lower

pair slightly the longer, ovate-subulate, 2 mm long,

the lateral deltoid, 1,5 mm long. Corolla whitish

to pale mauve with 4-5 dark vertical lines on the

upper lip, shortly pubescent and gland-dotted; tube

not expanded at the base, 0,75 mm deep and ascending

for 2,5 mm then curved downwards and starting to

expand for 4-5 mm, reaching 2,5 mm deep at the

throat; upper lip erect, oblong, 6-7 mm long, 4 mm

broad, emarginate at the apex and with 2 obscure

lateral ear-like lobes; lower lip boat-shaped, some-

what ascending. Stamens free to the base, curved

within the lower lip, subequal, about 5 mm long.

Style slightly exceeding the stamens, usually remaining

within the lower corolla lip. Fig. 43.

In forest margins and on shady stream banks,

often locally common forming a dense growth,

from about Humansdorp along the eastern Cape to

Natal, Swaziland and eastern and northern Transvaal,

extending into tropical Africa.

Transvaal. — 2229 (Waterpoort): Wylliespoort (-DD), Haf-

strom & Accocks 1334. 2230 (Messina): Entabeni (-CC),

Obermeyer 1192. 2329 (Pietersburg) : Hanglip (-BB), Meeuse

10162; Louis Trichardt (-BB), Breyer sub TRV 22110; Young

sub TR V 27246; Woodbush (-DD), Van Warmelo 218; Hout-

boschberg (-DD), Schlechter 4762. 2330 (Tzaneen): Tshakoma

(-AB), Obermeyer 970; Westfalia Estate (-CA), Scheepers 170;

Bos 1356; Duiwelskloof (-CA), Galpin 10109; 10110; Modjad-

je’s Reserve (-CB), Krige 49; Woodbush (-CC), Wager sub

TRV 23067; Magoebaskloof (-CC), Mogg s.n.; Bruce 193;

New Agatha (-CC), MacCallum 137; Letty 370; Muller &

Scheepers 27. 2430 (Pilgrim’s Rest): The Downs (-AA), Crundall

s.n.; Wolkberg near Malopetsi, Meeuse 9914; Shiluvane (-AB),

Junod 777 ; Mariepskop (-DB), Fitzsimons & Van Dam sub TR V

22672; Van Son sub TRV 31565; Codd 7865 ; Verdoorn 2452;

Van der Schijff 4448; De Hoek (-DD), Taylor 649; near Graskop

(-DD), Strey 3727. 2530 (Lydenburg): between Lydenburg and

Dullstroom (-AB or -AC), Pole Evans 4292; Lunsklip Falls

(-AD), Codd 10014; Sabie (-BB), Rogers sub TRV 20558. 2531

(Komatipoort) : Barberton (-CC), Thorncroft 3259; Codd 8185.

Swaziland. — 2531 (Komatipoort); near Havelock Mine

(-CC), Codd 7829; Piggs Peak (-CC), Compton 30041. 2631

(Mbabane): Forbes Reef area (-AA), Compton 32333; Mbabane

(-AC), Compton 25056; 25589; Mankaiana (-CA), Compton

27706; Hlatikulu (-CD), Stewart 91 (K).

Natal. — Without locality: Gerrard 2844 (K); Sanderson

392 (K); Medley Wood 4237 (K). 2730 (Vryheid): Dumuka Mt.

(-?), Gerstner 4581; 4650. 2731 (Louwsburg): 11 km W. of

Ngome, Codd 9567; Ngome Forest, Gerstner 4523. 2830

(Dundee): Krantzkop-Middeldrift (-DD), Dyer 4354; Edwards

2099; 2100. 2831 (Nkandla): Babanango (-AC), King 342;

Ntonjaneni (-AD), Andrews 7; Nkandla Forest (-CA), Codd

1388; 6964; Lawn 500 (NH); Melmoth (-CB), Mogg 6164;

Eshowe (-CD), Lawn 154 (NH); Ngoye Forest (-DC), Venter

2283. 2929 (Underberg): 11 km W. of Nottingham Road

(-BD), Codd 8521; near Lundy’s Hill (-DB ?) Marais 825;

Deepdale (-DB), Strey 4816; 13 km from Donnybrook to

Mondi (-DD), Mauve 4856. 2930 (Pietermaritzburg): Umvoti

(-AB), Nicholson 873; Karkloof (-AC), Moll 3497; Huntley

377; Karkloof Range (-AD), Ross 2074; 13 km N.W. of York

(-AD), Codd 8579; Town Hill (-CB), Fisher 336 (NH); Town

Bush Valley (-CB), Fisher 1028 (NH); Ross 998 (NH); Byrne

(-CC), Galpin 11963; Inanda (-DB), Medley Wood 1047 (K,

NH); Everton (-DD), Hilliard 5021. 2931 (Stanger): near

Stanger (-AD), Pentz & Acocks 10427; Amatikulu River

(-BA), Strey 4204.

Cape. — 3029 (Kokstad): Insizwa (-CC), Strey 10781; Tonti

Forest (-CD) Forest Officer 576. 3129 (Port St. Johns): Port

St. Johns (-DA), Galpin 2844; Theron 1586. 3226 (Fort Beau-

fort): Katberg (-DA), Ecklon & Zeyer 91 (S); Shaw s.n. (K);

on path to Big Tree (-DB), Jacot-Guillarmod 5601. 3227

(Stutterheim): Hogsback (-CA), Johnson 1282; Amatola Mts.

(-CA), Erens 2228; Cata Forest Reserve (-CA), Story 3266;

Keiskamahoek Forest (-CA), Wells 3152; Pirie Forest (-CD),

Sim 19581; 19584; B. H. Dodd sub Galpin 8023; near Komga

(-DB), Flanagan 104. 3228 (Butterworth): near Kentani (-AD),

Pegler 161; between Quku and Kei Rivers, Drege a (G, MO,

P). 3326 (Grahamstown): near Grahamstown (-BC), Zeyher

876 (K); Ecklon & Zeyher 196 (SAM); Britten 931; Howiesons

Poort (-BC), Zeyher 3544; Britten 5249. 3424 (Humansdorp):

Witte Elsbos (-AA), Fourcade 1217 (K).

The corolla shape of P. laxiflorus is similar to

that found in P. petiolaris and the characteristics

on which the two species may be separated are

listed under the latter (see p. 433). A useful field

character is that P. laxiflorus has a sharp citronella-

like scent, unlike that of any other South African

species.

Chromosome number 2n =28 (De Wet, 1958).

Bentham based his species on an Ecklon specimen

(unspecified) and three Drege gatherings: “(a) inter

Key et Gekau, (b) inter Omtata et Omgeziana,

(c) inter Omsamculo et Omcomas.” There is a

specimen of Ecklon s.n. from near Grahamstown

at Kew while in several herbaria specimens of Drege a

(G, P, MO) and Drege c (K, P, S) have been seen.

The specimen of Drege c ex Herb. Bentham in K

L. E. CODD

435

is selected as the lectotype. No material of Drege b

has been seen.

A superficial examination of tropical African

species revealed several that are closely related to

P. laxiflorus and some of them are likely to be placed

in synonymy on closer scrutiny. The species which

fall in this relationship are:

P. johnstonii Bak. in FI. Trop. Afr. 5: 411 (1900).

Type: Tanzania, Kilimanjaro, Johnston 69 ( K !).

P. triflorus Bak., l.c. 417 (1900). Type: Tanzania,

Kilimanjaro, Thompson s.n. (K !).

P. kondowensis Bak., l.c. 417 (1900). Type: Malawi,

between Kondwe and Karonga, Whyte s.n. (K!).

P. urticoides Bak., l.c. 412 (1900). ( =P . laxiflorus

var. genuinus Briq.). Type: Angola, Pungo Andongo,

Welwitsch 5545 (K !).

P. hylophilus Guerke in Bot. Jahrb. 19: 203 (1894).

Type: Cameroons, Preuss 815 (no material seen).

P. violaceus Guerke, l.c. 201 (1894). Type: Tan-

zania, Lutindi, Holst 3317 (K !).

P. albus Guerke, l.c. 201 (1894). Type: Tanzania’

Kilimanjaro, Volkens 744 (no material seen, but

specimens so named in K have somewhat smaller

flowers than P. laxiflorus).

P. fraternus T. C. E. Fries in Notizbl. Bot. Gart.

Berlin 11: 26 (1930). Syntypes: Kenya, R. E. &

T. C. Fries 604; 643; 884 (K!); 911 (Kl); 1208 (K !).

Similar to above.

Cooke in FI. Cap. l.c. 277 (1910) cited Junod 111

from Shiluvane, Eastern Transvaal as P. hylophilus

Guerke. It is, however, P. laxiflorus, as is Thorncroft

3259 from Barberton, cited by Cooke as P. rehmannii.

EXCLUDED SPECIES

Plectranthus bolusii T. Cooke is a mixture of

Orthosiphon sujfrutescens (Thonn.) J. K. Morton

( Bolus 11011) and O. labiatus N.E. Br. ( Rehmann

6167 and Medley Wood 4488).

P. succulent us Dyer & Bruce =Thorncroftia suceu-

lenta ( Dyer & Bruce) Codd.

P. thorncroftii S. Moore =Thorncroftia thorncroftii

(S. Moore) Codd.

436

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

2. RABDOSIA

Rabdosia ( Bl .) Hassk. in Flora 25, Beibl. 2:25 (1842);

Blake in Contr. Queensl. Herb. 9: 4 (1971); Hara in

J. Jap. Bot. 47: 193 (1972). Type species: Rabdosia

javanica (Bl.) Hassk.

Elsholtzia Willd. sect. Rabdosia Blume, Bijdr. FI.

Ned. Ind. 825 (1825).

Plectranthus sections Isodon Schrad. ex Benth.,

Pyramidium Benth. and Amethystoides Benth., Lab.

29: 40 (1832); in DC., Prodr. 12: 55-61 (1848), partly;

Plectranthus subgen. Isodon (Benth.) Briq. in

Pflanzenfam, 4,3a: 352 (1897), partly.

Isodon (Benth.) Kudo in Mem. Fac. Sci. & Agr.

Taihoku Imp. Univ. 2: 118 (1929); J. Soc. Trop, Agr.

Taiwan 2: 145 (1930); Codd in Taxon 17: 239 (1968).

Amethystanthus Nakai in Bot. Mag. Tokyo 48:

785 (1934).

Homalocheilos J. K. Morton in J. Linn. Soc. (Bot.)

58: 249, 268 (1962); FI. West, Trop. Afr. ed. 2,2:

460 (1963).

Perennial herbs or erect or straggling subshrubs;

stems woody, not succulent. Inflorescence paniculate

with flowers arranged in pendunculate dichasia;

bracts leaflike at the base becoming progressively

smaller towards the apex, persistent. Calyx ± equally

5-toothed, the uppermost sometimes slightly smaller

than the others, if somewhat 2-lipped then the upper

lip 3-toothed and the lower 2-toothed. Corolla

bilabiate, tube expanded and deflexed near the base;

t4>per and lower lips relatively small. Stamens 4, free,

attached at the mouth of the corolla tube, didynamous,

declinate in the lower lip of the corolla. Style lying

with the stamens in the lower lip of the corolla.

Species about 100, distributed mainly in Asia and

Malesia with a few in Africa. One South African

species of rather distant affinity is now included but

is placed in a separate subgenus. See also p. 373 for

historical notes.

Subgen. Pyramidium (Benth.) Codd, stat. et comb,

nov.

Plectranthus sect. Pyramidium Benth., Lab. 44

(1832); in DC., Prodr. 12: 61 (1848); Briq. in Pflanzen-

fam. 4,3a: 354 (1897). Type species: Plectranthus

ternifolius D. Don from India.

Erect shrublets; flowers in dense, pyramidal

panicles; ripe calyx erect, somewhat urceolate.

There are two species in the subgenus, R. ternifolia

(D. Don) Hara and R. calycina (Benth.) Codd.

Although the two are widely separated geographically,

they are remarkably similar in appearance. R. terni-

folia, which occurs in India, has narrower, more

lanceolate-ovate leaves and a more markedly striate

calyx with shorter teeth than R. calycina , which is

restricted to the eastern part of South Africa.

The subgenus is in some respects intermediate

between Rabdosia and Plectranthus but, because of the

leaf-like bracts, it is best removed from Plectranthus.

The alternative, to make it a separate genus, does

not appear justified.

Rabdosia calycina (Benth.) Codd in Bothalia 11:

117 (1973); in Ross, FI. Natal 305 (1972), without

basionym. Lectotype: Cape, between St. Johns and

Umsikaba Rivers, Drege 3584 (K!, lecto.; = Drege b

in G!; MO!; P! ; S!).

Plectranthus calycinus Benth. in E. Mey., Comm. 230(1837);

Drege, Zwci Pfl. Doc. 148, 152 (1843); Benth. in DC., Prodr.

12: 61 (1848); Briq. in Pflanzenfam. 4, 3 a: 352 (1897); Cooke

in FI. Cap. 5, I; 270 (1910); Compton, FI. Swaz. 66, 157

(1966); Trauseld, Wild Flow. Drakensberg 160(1969); Codd in

Mitt. Bot. Munchen 10: 250 (1971).— var. pachystachyus (Briq.)

T. Cooke, l.c. 271 (1910). P. pyramidatits Guerke in Bull.

Herb. Boiss. 6: 522 (1898). Type: Transvaal, Houtbosch,

Rehmann 6179 (Z!). P. pachystachyus Briq. in Bull. Herb.

Boiss. ser. 2, 3: 1003 (1903). Type: Natal, Umkomaas, Medley

Wood Ab2\ (K!).

Erect, perennial, aromatic soft shrub or woody

herb, 60-150 cm tall; stems 1-several arising

annually from a perennial rootstock, woody below,

simple or sparingly branched, markedly ribbed below,

4-angled above, glandular-puberulous to densely

velvety-tomentose. Leaves opposite or ternate, sub-

sessile or shortly petiolate, coriaceous; blade ovate

lanceolate to broadly ovate, 4-10 cm long, 2-4,5 cm

broad, subglabrous to strigose above, paler, reticulate-

veined, subglabrous to densely velvety-tomentose and

copiously punctate with minute orange to reddish-

brown gland-dots below; apex acute to acuminate;

base obtuse to shortly cuneate; margin regularly

crenate-dentate with 12 to 22 pairs of teeth. Inflores-

cence a dense, cylindrical to pyramidal, terminal'

panicle, 10-30 cm long; rhachis sparsely to densely

and shortly glandular-tomentose; bracts ovate, per-

sistent, leaf-like at the base, becoming progressively

smaller towards the apex. Flowers in pendunculate

dichasia, densely placed; pedicels up to 4 mm long,

glandular-hispidulous. Calyx equally 5-toothed, glan-

dular-hispid, villous within, 2 mm long at flowering,

enlarging to 7-9 mm long in fruit; mature calyx erect,

somewhat urceolate, tube cylindrical, 10-ribbed,

teeth lanceolate-deltoid, 2 mm long. Corolla 8-11 mm

long, white to cream with a mauve margin on the

lower lip (described by collectors as mauve, pinkish,

white and mauve, or white and pink), densely

tomentose and gland-dotted; tube 5-6 mm long,

expanding abruptly and saccate dorsally at the base,

sharply deflexed, laterally compressed, 4 mm deep

at the base, narrowing to 3 mm at the throat; upper

lip erect, 2 mm long and equally broad, notched at

the apex and with a pair of lateral ear-like lobes;

lower lip usually curved upwards, shallowly boat-

shaped, up to 5 mm long. Stamens didynamous,

2,5-3 mm and 3, 5-4, 5 mm long, curved and enclosed

within the lower corolla lip. Style enclosed in the

lower lip. Fig. 44.

Found usually among rocks in dense grassland and

at forest margins, from the Amatola Mountains in the

eastern Cape Province, throughout Natal, especially

in the midlands and Drakensberg area, the mountains

of Swaziland and eastern Transvaal, reaching its

northern limit on the Soutpansberg and Blouberg.

Transvaal. — 2229 (Waterport): Farm Buckworth (-DC),

Meeuse 10243; north of Louis Trichardt (-DD), Breyer sub

TRV 22103; Rodin 4041; Meeuse 10164. 2230 (Messina):

Entabeni Forest Station (-CC), Obermeyer 1235. 2329 (Peters-

burg): Blouberg (-AA), Leeman 118; Codd 8759; Van der

Schijff 5416; Strey & Schlieben 8517; Houtbosch (-DD),

Rehmann 6179 (Z); Bolus 10982; Codd 9428; Meeuse 9813;

near Haencrtsburg (-DD), Codd 9433; Magoebaskloof (-DD),

Codd &. Muller 358; Burger 126; Iron Crown Mt. (-DD),

Meeuse 9853. 2330 (Tzaneen): Duiwelskloof (-CA), Galpin

10113; Scheepers 657; New Agatha (-CC), McCallttm s.n.;

Miiller & Scheepers 22. 2430 (Pilgrims Rest): The Downs

(-AA), Crundall s.n.; near The Downs (-AC), Vahrmeijer

2366; Mariepskop (-DB), Van der Schijff 5100; Peach Hill

(-DC), Galpin 14364; Ohrigstad Dam Nature Reserve (-DC),

Jacobsen 2345; Graskop (-DD), Galpin 14357; 14414. 2530

(Lydcnburg): Steenkampsberg (-AA), Hardy 904; Dullstroom

(-AC), Galpin 12471; Werdermann & Oberdieck 2049; Mt.

Anderson (-BA), Meeuse 10070; Strey 3544; Long Tom Pass-

(-BA), Leistner & Mauve 3215; farm Zwagershoek (-BA),

Obermeyer sub TRV 28033; between Belfast and Dullstroom

(-CA), Pole Evans 3991 ; Belfast (-CA), Pole Evans H 11564;

Hutchinson 2742; Machadodorp (-CB), Young sub TRV 26639;

Nelsberg (-DD), Taylor 1879. 2531 (Komatipoort): Krokodil-

poort (-AD), Nel 233; near Barberton (-CC), Galpin 820;

Liebenberg 2431; Lomati Valley (-CC), Thorncroft 2051; near

L. E. CODD

437

Fig. 44. — Rabdosia calycina, Umzimkulu, eastern Cape Pro-

vince ( Ward 6279), x)r.

Angle Station (-CC), Clarke 40. 2630 (Carolina): Vossmans

Beacon (-BA), Bruce 269; near Lochiel (-BB), Repton 894.

2729 (Volksrust): Amersfoort (-BB), Sidey 3510; Volksrust,

Haagner sub TRV 10288. 2730 (Vryheid): near Wakkerstroom

(-AC), Galpin 10215; Repton 916; Oshoek (-AC), Devenish

461.

O.F.S. — Without locality, Cooper 1016 (K). 2729 (Volksrust):

near Mt. Pelaan (-DC), Muller 916. 2828 (Bethlehem): Wit-

zieshoek, Liebenberg 8127. 2929 (Harrismith): near Van Reenen,

Phillips s.n.

Swaziland. — 2531 (Komatipoort): Havelock Mine (-CC)’

Miller 5190. 2631 (Mbabane): Mbabane (-AC), Compton

25023; 25600; 26715; 27873; Usutu Forest (-AC), Compton

27766; Hlatikulu (-CD), Stewart sub TR V 9628.

Natal. — Without locality: Gerrard s.n. (K); Gerrard &

McKen s.n. (K). 2729 (Volksrust): Langsnek (-BD), Rehmann

6961 (K); Majuba (-BD), Rogers sub TRV 3313; near Charles-

town (-BD), C. A. Smith 5646; 5665. 2730 (Vryheid): near

Utrecht (-CB), Breyer sub TRV 16951. 2828 (Bethlehem):

Drakensberg National Park (-DB), Codd & Dyer 2775; Edwards

540; Sidey 1640; Mont-aux-Sources (-DD), Bayer & McClean

29. 2829 (Harrismith): Van Reenen (-AD), Medley Wood

13130; Cathedral Peak Forestry Station (-CC), Killick 1399;

Admiraal & Drijfhout 2868; near Ladysmith (-DB), K untie

s.n. (K); Wilms 2201 (K). 2830 (Dundee): Qudeni Forest

(-DB), Gerstner 6795. 2831 (Nkandla): Inhlazatshe (-AA),

Gerstner sub NH 22747 (NH); Babanango (-CA), King 276;

Sidey 3701 ; Ulundi (-AD), Evans 485 (NH); Melmoth (-CB),

Mogg 4590; Eshowe (-CD), Lawn 461 (NH); Ngoye Forest

Reserve (-DC), Venter 2343. 2929 (Underberg): Cathkin Park

(-AB), Galpin 11859; Champagne Castle (-AB), Acocks 10091

(NH); Giants Castle Reserve (-AB), Symons 384; Trauseld

574; 729; Nicholson 491 (NH); Tabamhlope Mt. (-BA), West

101 ; Dalton Bridge (-BB), West 772; Mooi River (-BB), Mogg

7149; Bamboo Mt., McClean 706; Underberg (-CD), McClean

705 (NH); Coleford (-CD), Moll 5151; between Boston and

Bulwer (-DB), Dyer 4871; Everglades-Boston road (-DB),

Moll 632. Mpendhle (-DB), C. A. Smith 8305B; Marwaga Mt.

(-DC), McClean 222; Bulwer (-DD), Allsopp 831 ; Donnybrook

(-DD), Medley Wood 13093. 2930 (Pietermaritzburg): Tweedie

(-AC), Mogg 6747; Nottingham Road (-AC), McClean 815

(NH); near Dargle (-AC), Edwards 3078; 3079; Karkloof (-AC),

Huntley 374 (NH); Greytown (-BA), Wylie sub NH 21943;

Town Bush Valley (-CB), Rump sub NH 20592 (NH); Byrne

(-CC), Galpin 12019; Strey 10943; Richmond (-CD), Bayliss

2194; Inanda (-DB), Medley Wood 489 (K). 3030 (Port Shep-

stone): Ellesmere (-AD), Rudatis 638; near Umkomaas (-BB)',

Medley Wood 4621 ; Oribi Flats (-CA), McClean 354; Izotsha

Falls (-CB), Strey 7584; Umtamvuna Forest Reserve (-CC),

Strey 6969. 3130 (Port Edward): near Port Edward (-AA),

Bayliss 552.

Cape. — 3028 (Matatiele): near Mt. Frere (-DD), Lewis 4338

(SAM); Theron 2194. 3029 (Kokstad): Clydesdale (-BD),

Tyson 2749; Ibisi Cuttings (-BD), Ward 6729; Mr. Frere-

Cedarville road (-CA), Strey 10803; Kokstad (-CB), Pegler

1771; near Mt. Ayliff (-CD), Lewis 4336 (SAM); Harding

(-DB), Olivier 121 (NH). 3126 (Queenstown): near Queenstown

(-DD), Galpin 8161. 3127 (Lady Frere): Engcobo (DB),

Britten 7033. 3128 (Umtata): Baziya (-CB), Baur 97 (K, SAM);

near Mqanduli (-DD), Codd 9266. 3129 (Port St. Johns): near

Magwa Store (-BC), Strey 8550; near Libode (-CA), Theron

2181; between Morley and Umtata River (-CC), Drege a

(G, K, P); between St. Johns and Umsikaba Rivers (-DA?),

Drege 3584 (K); b (G, MO, P, S). 3226 (Fort Beaufort): Ama-

tola Mts. (-BD), Dyer 1980; Peffers Kop (-DB), Acocks 9760;

Hogsback (-DB), Rattray 14; Johnson 1142; Killick 917;

Comins 1459. 3227 (Stutterheim): Stutterheim (-CB), Story

3441; Pirie (-CC), Sim 19583; 19587; 3228 (Butterworth):

Kentani (-CB), Pegler 162; Kei Mouth (-CB), Flanagan 1889.

There is a good deal of variation in pubescence

with a gradient from south to north. Some Cape

specimens are almost glabrous, while in the northern

Transvaal many specimens are densely tomentose,

especially on the under surfaces of the leaves. An

extreme form is represented by Vahrmeijer 2366 from

mountains near The Downs, with very densely

tomentose leaves, while the rhachis, pedicels and

calyx are markedly glandular-villous.

Chromosome number 2n =28 (De Wet, 1958).

A few specimens from high altitudes (over 1 800 m)

in the Soutpansberg and Blouberg differ in being

freely branched shrubs with leaves densely tomentose

below and smaller inflorescences, calyx and corolla.

However, the distinctions are not clear cut. The

branched habit may be because the plants grow in

protected, rocky places which are not easily reached

by fire.

3. SOLENOSTEMON

Solenostemon Schumach. in Schumach. & Thonn.,

Beskr. Guin. PI. 271 (1827); Benth. & Hook. f. 2: 1 175

(1876); Briq. in Pflanzenfam. 4,3a: 359 (1897); Bak.

in FI. Trop. Afr. 5: 420 (1900); Hutch. & Dalz., FI.

W. Trop. Afr. 2: 289 (1931); emend. Morton in J.

Linn. Soc. (Bot.) 58: 251 (1962); Codd in Mitt. Bot.

Munchen 10: 249 (1971); Blake in Contr. Queensl.

H?rb. 9: 6 (1971). Type: S', ocymoides Schumach. &

Thonn. from West Tropical Africa.

Coleus sect. Solenostemon (Schumach.) Benth., Lab.

52 (1832); in DC., Prodr. 12: 72 (1848).

Coleus sect. Solenostemoides Briq., l.c. 360 (1897).

Solenostemon sect. Coleoidea J. K. Morton, l.c. 253

(1962) descr. angl.

Annual or perennial, softly woody or subsucculent,

erect or straggling herbs or subshrubs. Inflorescence

paniculate with flowers arranged in dense or lax

pedunculate or sessile dichasia; bracts sharply

differentiated from the leaves, early deciduous. Calyx

bilabiate, the uppermost lobe usually broadly ovate,

the lateral teeth short and truncate to deltoid or

obsolete, the lowermost lobe ovate to oblong and

entire or emarginate, or strap-shaped and forked,

formed from the union of the two lowest teeth for

438

PLECTRANTHUS (LABI AT AE) AND ALLIED GENERA IN SOUTHERN AFRICA

their entire, or greater part of their length. Corolla

usually violet coloured, tube expanding gradually

from the base to the throat, twice curved, the lower

lip large and deeply boat-shaped. Stamens 4, free or

shortly united at the base, attached at the mouth of

the corolla tube, didynamous, declinate in the lower

lip of the corolla. Style lying with the stamens in the

lower lip of the corolla.

The genus as amended by Morton is common in

tropical Africa and extends to Asia and Malesia, with

two species recorded from South Africa. More than

60 names have been published, mainly in Coleus (see

Codd, l.c., 1971) but, due to the variation in vegetative

characteristics, species limits are difficult to determine,

and it is likely that the number of species eventually

recognized will be considerably less than this.

The historical background is outlined on p. 373.

Solenostemon Schumach. was placed as a section of

Coleus by Bentham. To the typical species he added

certain species from Madagascar, Asia and Malesia

with a somewhat different calyx shape. In typical

Solenostemon the two lowest calyx teeth are fused

into an ovate to oblong, entire or emarginate lip which

is bent upwards closing the mouth of the tube when

mature, while the lateral teeth are reduced or obsolete.

In the species which Bentham added, the lower lip

of the calyx is strap-shaped and forked at the apex,

while the lateral calyx teeth are short and rounded

to deltoid.

Benth. & Hook, f., Gen. PI. 2: 1175 (1876), rein-

stated Solenostemon Schumach. in its strict sense,

while the additional species were retained in Coleus.

Briquet (1897) also treated Solenostemon as a distinct

genus and the additional species were placed in

Coleus sect. Solenostemoides Briq.

With the transfer of true Coleus (i.e. the solitary

species C. amboinicus Lour.) to Plectranthus, Morton,

l.c., included sect. Solenostemoides Briq. in the

genus Solenostemon, but gave it the name sect.

Coleoidea, pointing out that it is, in some respects,

intermediate between true Solenostemon and Plec-

tranthus. However, the calyx is easily recognizable and

separation from Plectranthus appears to be fully

justified. Whether sect. Solenostemoides should be

included in Solenostemon is open to question and the

alternative would be to make a separate genus of it.

In the present treatment it is retained in Solenostemon

but given subgeneric status.

Subgen. Solenostemoides {Briq.) Codd, comb, et

stat. nov. Lectotype: S. latifolius (Hochst. ex Benth.)

J. K. Morton.

Coleus sect. Solenostemoides Briq. in Pflanzenfam.

4, 3a: 360 (1897).

Solenostemon sect. Coleoidea J. K. Morton in J.

Linn. Soc. (Bot.) 58: 253 (1962).

One of the best known species in this subgenus is

the commonly cultivated “Coleus” with variegated

and often incised leaves, usually referred to as Coleus

blumei Benth. This possibly persuaded Keng, when

dealing with the Malesian species in Gard. Bull.

Singapore 24: 51 (1969), to retain the genus Coleus

in a broad sense. He also found it necessary to take a

broad view of species limits and included C. blumei

and several others in C. scutellarioides. This species

is widely cultivated and is frequently grown in South

Africa, in gardens in the warmer areas, and as a

pot-plant. It is, therefore, included in the key below

and is transferred to Solenostemon.

A broad view is also taken of the indigenous

material which was described as Plectranthus tysonii

Guerke and Coleus rehmannii Briq., both of which

are now included in S. latifolius (Hochst. ex Benth.)

J. K. Morton.

The third species included in the key, S. rotundi-

folius (Poir.) J. K. Morton, is cultivated in native

gardens for the edible, potato-like tubers and is not

known in the wild state in South Africa.

Key to species

Roots tuberous; flowers in densely glomerate clusters, corolla 5-7 mm long 1. S. rotundifolius

Roots fibrous; flowers in dense or lax clusters, corolla 8-15 mm long:

Leaves not variegated, occasionally with a dark V-shaped mark near the base of the blade; indigenous

2. S. latifolius

Leaves variegated: cultivated plants 3. S. scutellarioides

1 . Solenostemon rotundifolius {Poir.) J. K. Morton

in J. Linn. Soc. (Bot.) 58: 272 (1962). Type: Mauritius,

Commerson (P, holo.).

Germanea rotundifolia Poir. in Lam., Encycl. 2: 763 (1812).

Plectranthus rotundifolius (Poir.) Spreng., Syst. 2: 690 (1825).

Coleus dysentericus Bak. in Kew Bull. 1894: 10 (1894); FI.

Trop. Afr. 5: 437 (1900). Type: Niger region. Barter 846 (K,

holo.). C. rotundifolius (Poir.) A. Chev. & E. Perrot., Veg.

Util. Afr. Trop. Franc. 1 : 101, 119 (1905).

Perennial, aromatic, semi-succulent herb; rootstock

producing ovoid, potato-like tubers; stems erect to

decumbent, 30-60 cm tall, 4-angled, puberulous to

shortly pubescent. Leaves fairly thick-textured, petio-

late; petiole 2-3 cm long, puberulous; blade ovate,

2,5-5 cm long, 2-3 cm broad, puberulous to strigose,

gland-dotted below; apex acute, base cuneate;

margin crenate-dentate. Inflorescence terminal,

slender, simple 6-10 cm long; bracts ovate,

acuminate, 2 mm long, deciduous. Flowers in com-

pact, sessile dichasia; pedicels 0-1 mm long. Calyx

1,5 mm long, enlarging to 3 mm long in fruit,

glandular-hispid; upper lip ovate; lateral teeth

truncate; lower teeth fused for most of their length

forming an oblong, emarginate lobe. Corolla bluish-

purple, 5-7 mm long, pubescent, gland-dotted; tube

curved in an inverted U; upper lip erect, 1,5 mm

long, 4-lobed; lower lip boat-shaped, 2,5 mm long.

Stamens curved within the lower lip 2-2,5 mm long,

shortly united at the base. Style slightly exceeding

the stamens. Fig. 45.

Cultivated for the potato-like tubers in eastern

Transvaal and Zululand; origin uncertain, probably

tropical Africa.

Transvaal. — Cult. Skinners Court, Pretoria, Mundy sub

Tvl. Dept, Agr. 3457; cult. Division of Botany, Pretoria, Native

Affairs Dept. 2330 (Tzaneen): Westfalia Estate (-CA), Scheepers

931. 2531 (Komatipoort): Barberton (-CC), Clarke 65.

Natal. — Zululand, without precise locality, Curson s.n.

2731 (Louwsburg): Ngome (-CD), cultivated, Gerstner 4454.

There are probably further synonyms among the

species listed by Chevalier & Perrottet, l.c., but no

attempt has been made to sort these out. Three

varieties of S. rotundifolius are also maintained by

these authors.

L. E. CODD

439

Fig. 45. — Solenostemon rotundifolius,

cultivated in Pretoria.

2. Solenostemon latifolius (Hochst. ex Benth.)

J. K. Morton in J. Linn. Soc. (Bot.) 58: 271 (1962);

Ross, FI. Natal 305 (1972). Syntypes: Ethiopia,

Schimper 825; 1828 (K).

Coleus latifolius Hochst. ex Benth. in DC., Prodr. 12: 74

(1848); A. Rich., Tent. FI. Abyss. 2: 184 (1851), Bak. in FI.

Trop. Afr. 5: 437 (1900). C. rehmannii Briq. in Bull. Herb.

Boiss. ser 2, 3 : 1075 (1903); Cooke in FI. Cap. 5, 1 : 289 (1910).

Type: Transvaal, Houtbosch, Rehmann 6156 (Z !, holo.).

Plectranthus tysonii Guerke in Bot. Jahrb. 24: 77 (1898);

Cooke, l.c. 276 (1910). Type: Cape, Griqualand East, Clydes-

dale, Tyson 2769 (K!; G!; PRE!).

Perennial aromatic herb 20-70 cm tall; stems

slightly succulent, suberect to procumbent, 20-150 cm

long, 4-angled, sparingly hispidulous to shortly

crisped tomentose. Leaves thin to medium-thick in

texture; petiole 1,5-6 ( — 10) cm long, pubescence

similar to the stems; blade broadly ovate-deltoid,

2,5-8 cm long, 2-6,5 cm broad, sparingly pubescent

to densely hispidulous, sometimes with a dark V-

shaped mark, and dotted below with minute reddish-

brown gland-dots; apex acute, base truncate; margin

crenate to crenate-dentate with 6-12 pairs of somewhat

rounded teeth. Inflorescence terminal, usually simple,

10-35 cm long; rhachis sparsely to densely

hispidulous; bracts broadly ovate, acuminate,

3-5 mm long, early deciduous. Flowers in sessile or

rarely pedunculate, usually compact, dichasia, or in

reduced few-flowered sessile cymes in depauperate

specimens; pedicels 2-5 mm long, hispidulous.

Calyx 2 mm long at flowering, enlarging to 7 mm

long in fruit, somewhat gibbous ventrally, glandular-

hispidulous, often purple tinged; upper lip suberect,

ovate, obtuse to rounded, 2 mm long, apex truncate

to rounded; lower pair of teeth united, forming a

strap-shaped lobe up to 4 mm long, forked at the

apex. Corolla 11-14 mm long, blue-violet to purple,

usually paler on the upper lip, sparingly pubescent

and gland-dotted; tube twice bent, 1 mm deep and

bent upwards for 1,5-2 mm then bent or curved

downwards and expanding for 2-3 mm, reaching

2-3 mm deep at the throat; upper lip 2 mm long and

equally broad, emarginate and with 2 small lateral

ear-like lobes; lower lip large, boat-shaped, 7-9 mm

long, horizontal. Stamens 7-8 mm long, united at the

base for 2 mm or occasionally free to the base. Style

exceeding the stamens by about 1 mm, often shortly

exserted from the lower lip. Fig. 46.

Found in forests, open woodland and among rocks

from East Griqualand to Natal, Swaziland and the

mountains of eastern and northern Transvaal,

extending to Ethiopia and west tropical Africa.

Transvaal. — 2230 (Messina): between Sibasa and Lake

Funduzi (-CD), Story 4845. 2329 (Pietersburg): Houtbosch

(-DD), Rehmann 6156 (Z); Obermeyer sub TRY 31863. 2330

(Tzaneen): Westfalia Estate (-CA), Scheepers 340; 598; Bos

1339; Magoebaskloof (-CC), Codd 8408; Letty 456. 2430

(Pilgrims Rest): Mariepskop (-DB), Van Son sub TRY 34759;

Codd 7915; Van der Schijff 5140; God’s Window (-DD),

Davidson 252. Kowyns Pass (-DD), Ranh & Schlieben 9727.

Swaziland. — 2531 (Komatipoort): near Piggs Peak (-CC),

Codd 7820; 9522; Clarke 59. 2631 (Mbabane): near Mbabane

(-AC), Compton 25030; 25612; 26793; 27680; Codd 9517,

Schlieben 9520.

Natal. — 2831 (Nkandla): Ngoye (-DC), Wylie sub Wood

5638 (K). 2930 (Pietermaritzburg): Inchanga (-DA), Eshuis

s.n. 3030 (Port Shepstone): Dumisa (-AD), Rudatis 262;

Umkorqaas Valley (-BA?), Ward 150; Mehlomnyama (-CB),

Codd 9363.

Cape.— 3029 (Kokstad): Clydesdale (-BD), Tyson 2769;

near Rode (-CC), Acocks 22111; Umzimvubu Cutting (-CC),

Strey 11147 (NH).

There is a good deal of variation in size, texture

and pubescence of leaves, depending mainly on the

degree of dryness and exposure of the habitat, from

moist, shady, forest conditions to more open valley

bushveld. Some specimens from Swaziland (e.g. Codd

9517) have long, trailing stems and were at first

considered worthy of separate status. The curvature

of the corolla tube also varies considerably and could

not be used as a criterion for separating the South

African material into distinct groups. There is also

variation in the degree to which the stamens are

united at the base, most specimens having the filaments

united for 1-2 mm while, in a few Natal specimens,

the filaments are free to the base. This gives some

indication of the difficulties which will be expe-

rienced when a revision of the tropical African

material is undertaken.

Chromosome numbers 2n =24, 48, 60 (De Wet,

1958, as “Coleus tysonii”, Morton, 1962).

3. Solenostemon scutellarioides ( L .) Codd, comb,

nov. Type: from Asia.

Ocimum scutellarioides L., Sp. PI. ed. 2: 834 (1763); Sims in

Bot. Mag. t. 1446(1812).

Plectranthus scutellarioides (L.) R. Br., Prodr. FI. Nov. Holl.

506 (1810). P. blumei (Benth.) Launert in Mitt. Bot. Miinchen

7: 301 (1968).

440

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

Coleus scutellarioides (L.) Benth. in Wall., PI. As. Rar. 2:

16 (1830); Lab. 53 (1832); in DC., Prodr. 12: 73 (1848); Hook,

f., FI. Brit. India 4: 626 (1885); Keng in Gard. Bull. Singapore

24: 51 (1969). C. blumei Benth., Lab. 56 (1832); Hook, in Bot.

Mag. t. 4754 (1853). Type: from Java.

Perennial aromatic herb, erect or procumbent, up

to 1 m tall; stems semi-succulent, 4-angled, finely

tomentose. Leaves very variable, petiolate; blade

membranous, usually large, ovate-deltoid or broadly

ovate to ovate-oblong, usually brightly coloured or

blotched; margin crenate dentate or variously

incised. Inflorescence terminal, simple or with a pair

of branches near the base; rhachis finely tomentose;

bracts ovate, long acuminate, early deciduous.

Flowers in sessile, several-flowered dichasia. Calyx

about 7 mm long when mature; upper lip erect,

ovate, 2 mm long, rounded at the apex; lateral teeth

broadly oblong, 2 mm long, rounded to truncate;

lower teeth united forming a strap-shaped lobe 4 mm

long, forked at the apex. Corolla 8-10 mm long,

blue-violet, paler on the upper lip, sparingly

pubescent; upper lip short, 1-1 ,5 mm long; lower lip

boat-shaped, 4-5 mm long. Stamens usually united

at the base for 1-2 mm.

Originally from eastern Asia and Malesia, the

many forms or hybrids with variegated leaves are

widely grown in gardens in the tropics and as pot-

Fig. 46.— Solenostemcn iaCfolius, Ma-

riepskep ( Codd 7915), x 1.

plants in cooler areas. Various forms are cultivated

in South Africa.

Transvaal. — Cultivated' in Pretoria, Codd 9835.

This is the well known “Coleus” of gardens and

this may have influenced Keng, l.c., to uphold the

genus Coleus in the broad, Benthamian sense for the

Malesian species. Keng places C. blumei as a synonym

of C. scutellarioides and upholds four varieties which

are, however, of doubtful value. Probably more

synonyms will be added when the group comes to be

revised as a whole.

Chromosome number 2n=48 (Morton, 1962).

Another species which has been grown in South

Africa is the species described as Coleus shirensis

Guerke from East Tropical Africa. It is a more robust

plant than C. latifolius with larger leaves, more

floriferous inflorescence and more villous rhachis and

pedicels. The transfer to Solenostemon, with S.

zambesiacus Bak. as a synonym, is effected below.

Solenostemon shirensis ( Guerke ) Codd , comb,

nov. Syntypes: Nyasaland, Buchanan 376 (K!); 602b;

Mt. Mlanje, Whyte s.n. (K!).

Coleus shirensis Guerke in Bot. Jahrb. 19: 216 (1894);

N.L.Bi. in Bot. Mag. t. 8024 (1905).

Solemostemon zambesiacus Bak. in FI. Trop. Afr. 5: 421

(1900). Syntypes: Nyasaland, Blantyre, Last s.n.; between

Shibisa and Tshinmuzo, Kirk s.n. (K!).

L. E. CODD

441

REFERENCES

Bentham, G., 1832. Labiatarum Genera et Species. London.

29-58.

Bentham, G., 1848. Labiatae. In A. P. De Candolle, Prodr.

12: 55.

Blake, S. T., 1971. A revision of Plectranthus in Australia.

Contr. Queensl. Herb. 9: 1-120.

Briquet, J., 1897. Labiatae. Pflanzenfam. IV, 3a: 348-363.

Codd, L. E., 1971. Generic limits in Plectranthus, Coleus and

allied genera. Mitt. Bot. Miinchen 10: 245-252.

De Wet, J. M. J., 1958. Chromosome numbers in Plectranthus

and related genera. S. Afr. J. Sci. 54: 153-6.

Hara, H., 1972. On the Asiatic species of the genus Rabdosia

(Labiatae). J. Jap. Bot. 47: 193-203.

Keng, H., 1969. Flora Malesianae Precursores. XLVIII. A

revision of the Malesian Labiatae. Gard. Bull. Singapore

24: 13-180.

Launert, E., 1968. Miscellaneous notes on Labiatae. Mitt. Bot.

Miinchen 7 : 295-307.

Morton, J. K., 1962. Cytogenetic studies on the West African

Labiatae. J. Linn. Soc. (Bot.) 58: 231-283.

INDEX

Amethystanthus Nakai 373, 436

Ascocarydion G. Tayl 374, 381

Burnatastrum Briq 372, 374, 380

spicatum (E. Mey. ex Benth.) Briq 374, 383

Coleus Lour 373, 374, 387

sect. Aromaria Benth 373, 387

sect. Calceolus Benth 373, 387, 389

sect. Solenostemoides Briq 373, 438

sect. Solenostemon Briq 373, 387, 437

amboinicus Lour 373, 374, 388

var. violaceus Guerke 388

aromaticus Benth 387, 388

barbatus (Andr.) Benth 394

blumei Benth 373, 438, 440

caninus (Roth) Vatke 390

caninus sensu Vatke 393

carnosus Dinter ms 392

comosus Hochst. ex Guerke 393

crassifolius Benth 388

dazo A. Chev 377

decumbens Guerke 390

dysentericus Bak 438

esculentus (N.E. Br.) G. Tayl 377

flavovirens Guerke 490

floribundus (N.E. Br.) Robyns & Lebrun 377

var. longipes (N.E. Br.) Robyns & Lebrun 377

forskohlii sensu Briq 394

gazensis S. Moore 382

latifolius Hochst. ex Benth 439

leucophyllus Bak 381

madagascariensis (Pers.) A. Chev 403

matopensis S. Moore 382

mirabilis Briq 381

var. buchnerianus Briq 381

var. hypisodontus Briq 381

var. mechowianus Briq 381

var. poggeanus Briq 381

myrianthus (Briq.) Brenan 382

neochilus (Schltr.) Codd 392

omahekense Dinter 390

pentheri Guerke 392

polyanthus S. Moore 382

rehmannii Briq 439

rotundifolius (Poir.) A. Chev 438

schinzii Guerke 392

scutellarioides (L.) Benth 387, 440

shirensis Bak 440

spicatus Benth 390

spicatus sensu A. Rich 393

tetensis Bak 390

vagatus E. A. Bruce 390

zatarhendi (Forsk.) Benth 398

Elsholtzia Willd 373, 436

sect. Rabdosia B1 373, 436

javanica B1 373

Englerastrum Briq 373

floribundus (N.E. Br.) Th. Fries jun 377

var. longipes (N.E. Br.) Th. Fries jun 377

schweinfurthii Briq 373

Germanea Lam 372, 374

crassifolia (Vahl) Poir 398

laxiflora (Benth.) Hiem 434

maculosa Lam 372, 374

rotundifolia Poir 438

urticifolia Lam 372, 374, 415

Homalocheilos J. K. Morton 373, 436

Isodon (Schrad. ex Benth.) Kudo 373, 436

Majana amboinica (Lour.) Kuntze 388

Neomullera Briq 374

damarensis S. Moore 382

welwitschii Briq 374

Ocimum madagascariensis Pers 403

racemosum Thunb 407

scutellarioides L 439

tomentosum Thunb

verticillatum L.f

zatarhendi Forsk."

Plectranthus L'Herit

sect. Amethystoides Benth

sect. Coleoides Benth

sect. Germanea Benth

sect. Heterocylix Benth

sect. Isodon Schrad. ex Benth

sect. Melissoides Benth

sect. Pyramidium Benth

subgen. Burnatastrum (Briq.) Codd ....

subgen. Calceolanthus Codd

subgen. Coleus (Lour.) Codd

subgen. Isodon (Schrad. ex Benth.) Briq.

subgen. Nodiflorus Codd

subgen. Plectranthus

sect. Coleoides Benth

sect. Plectranthus

subgen. Xerophilus Codd

aegyptiacus C. Chr

ambiguus (Boh) Codd

amboinicus (Lour.) Spreng

aromaticus (Benth.) Roxb

arthropodus Briq

aff. arthropodus sensu Compton

aurifer Dinter ex Launert

barbatus Andr

behrii Compton

biflorus Bak

blumei (Benth.) Launert

bolusii T. Cooke

calycinus Benth

var. pachystachyus (Briq.) T. Cooke

candelabriformis Launert

caninus Roth

charianthus Briq

ciliatus E. Mey. ex Benth

coloratus E. Mey. ex Benth

comosus Sims

cooperi T. Cooke

crassifolius Vahl

cylindraceus Hochst. ex Benth

densiflorus T. Cooke

densus N.E. Br

dinteri Briq

dolichopodus Briq

draconis Briq

dregei Codd

ecklonii Benth

elegantulus Briq

esculentus N.E. Br

floribundus N.E. Br

var. longipes N.E. Br

forskohlaei sensu Ait. f

fruticosus L'Herit

fruticosus sensu Marloth

galpinii Schltr

glomeratus R. A. Dyer

grallatus Briq

grandidentatus Dinter ms

grandidentatus Guerke

hereroensis Engl

hilliardiae Codd

hirtus Benth

Iiylophilus sensu Cooke

krookii Guerke ex Zahlbr

var. grandifolia T. Cooke

kuntzei Guerke

kuntzeanus Domin

lanceolatus Benth

lavanduloides Bak

laxiflorus Benth

403

407

398

372, 374

372, 436

372, 395

372, 395, 406

372

372, 380, 436

372

372, 436

380

389

387

436

376

395

406

378

398

429

388

415, 421

420

382

394

416

376

439

435

436

380

389, 390, 393

415, 421

414

429

393, 394

418, 424

398

385

378

396

423

401

429

431

412

377

394

372, 374, 415, 412

431

415

385

418, 421

382

396, 401

382

428

403

434

418

W 431

409

372, 385

434

442

PLECTRANTHUS (LABIATAE) AND ALLIED GENERA IN SOUTHERN AFRICA

madagascariensis ( Pers .) Benth 402, 409

var. aliciae Codd 402, 404

var. madagascariensis 403

var. ramosior Benth 399, 402, 404

marrubioides Hochst. ex Benth 385

matabelensis Bak 382

mauritianus Boj 403

melanocarpus Guerke 376

mirabilis (Briq.) Launert 381

moschosmoides Bak 385

mutabilis Codd 404

myrianthus Briq 382

natalensis Guerke 414

forma glandulosa Phillips 418

neochilus Schltr 392

nummular ius Briq 407

oertendahlii Th. Fries jun 411

ornatus Codd 393

otaviensis Dinter 382

pachyphyllus Guerke ex T. Cooke 398

pachystachyus Briq 436

parviflorus Guerke 409

peglerae T. Cooke 416

petiolaris E. Mey. ex Benth 431

praetervisus Briq 418

primulinus Bak 378

psammophilus Codd 405

punctatus ( L.f .) L'Herit 372

purpuratus Harv 410

pyramidatus Guerke 436

rehmannii Guerke 421

rotundifolius (Poir.) Spreng 438

rubropunctatus Codd 418,420

saccatus Benth 426

var. longitubus Codd 428

var. saccatus 427

scutellarioides (L.) R. Br 439

spicatus E. Mey. ex Benth. . , 372, 383

spiciformis R. A. Dyer 385

strigosus Benth 409

var. lucidus Benth 409

subspicatus Hochst 383

succulentus Dyer & Bruce 435

swynnertonii S. Moore 422

ternifolius D. Don 436

tetensis (Bak.) Agnew 390

tetragonus Guerke 376

thorncroftii S. Moore 435

thunbergii Benth 407

tomentosus Benth 399

transvaalensis Briq 418

tysonii Guerke 439

unguentarius Codd 387

urticifolius (Lam.) Salisb 415

verticillatus {L.f.) Druce 407

villosus T. Cooke 385

woodii Guerke 401

xerophilus Codd 378

zatarhendi {Forsk.) E. A. Bruce 396,398,402

var. tomentosus {Benth.) Codd 399

var. woodii {Guerke) Codd 401

var. zatarhendi 398

zeylanicus Benth 399

zuluensis T. Cooke 424

Rabdosia (Bl.) Hassk 373, 374,437

subgen. Pyramidium {Benth.) Codd 436

calycina {Benth.) Codd 436

javanica {Bl.) Hassk 436

ternifolia {D. Don) Hara 436

Solenostemon Schumach 373, 374,437

sect. Coleoidea J. K. Morton 373, 438

subgen. Solenostemoides {Briq.) Codd 438

latifolius {Hochst. ex Benth.) J. K. Morton 438

ocymoides Schumach. & Thonn 373, 437

rotundifolius {Poir.) J. K. Morion 438

scutellarioides {L.) Codd 439

shirensis {Guerke) Codd 440

zambesiacus Bak 440

Bothalia 11, 4: 443-447 (1975)

Notes on African Acacia species

J. H. ROSS*

ABSTRACT

Information concerning miscellaneous African Acacia species is presented. A. dekindtiana A. Chev. and

A. hirtella E. Mey. var. inermis Walp. are relegated to synonymy under A. karroo Hayne, the misapplication of

the name A. giraffae Willd. is discussed, reasons for rejecting the names Mimosa reticulata L. and Mimosa

capensis Burm. f. are elaborated, and the identity of Mimosa senegalensis Forsk. is disclosed.

A number of decisions arising from a continuation

of studies on the African Acacias require explanation

in print. These form the subject of this paper.

ACACIA DEKINDTIANA A. CHEV.

A. Chevalier, in Rev. Bot. Appliq. 27: 509 (1947),

based his description of A. dekindtiana on Dekindt

431 from Huila in southern Angola. The holotype,

which consists of both flowering and fruiting material,

is housed in the Paris Herbarium. The paired stipular

spines are straight or almost so, the stems are dark

brown with minutely flaking bark, and the young

branchlets are sparingly pubescent. The petioles,

leaf-rhachides and rhachillae are fairly densely clothed

with short spreading hairs and the petioles and

rhachides are distincltly sulcate adaxially. The leaves

have 2-4 pinnae pairs and there is a slightly columnar

gland on the rhachis at the junction of each pinna

pair. The pinnae have up to 12 pairs of leaflets which

have short marginal cilia. The inflorescences are

capitate, on axillary peduncles and fascicled; the

peduncles are glandular and fairly densely pubescent ;

the involucels are ±2 mm long and one-third to

halfway up the peduncle. The corolla lobes are

slightly reflexed. The pods are reddish-brown, falcate,

up to 12 cm long and 7-9 mm wide, irregularly

constricted between some of the seeds, longitudi-

nally dehiscent; the valves are brittle, have a fine

± longitudinal venation and very sparse short

indumentum. The seeds are elliptic, ±7,5x5 mm.

Dekindt 431 matches several specimens of A. karroo

Hayne from Botswana, South West Africa and

southern Angola, for example, Barbosa 9727 (K)

from the Huila district in Angola. As it is clear that

A. dekindtiana is not specifically distinct from

A. karroo, the species is now reduced to synonymy.

Acacia karroo Hayne, Arzneyk. Gebr. Gewachse

10: t.33 (1827). Type: South Africa, Cape Province,

Herb. Willdenow 19184 fol. 2 (B, lecto.).

A. dekindtiana A. Chev. in Rev. Bot. Appliq. 27: 509 (1947);

Torre in Consp. FI. Angol. 2: 285 (1956). Type: Angola, Huila

Distr., Huila, Dekindt 431 (P, holo.l).

A. robusta sensu Oliv. in FI. Trop. Afr. 2: 349 (1871), non

Burch.; Benth. in Trans. Linn. Soc. Lond. 30: 510 (1875)

pro parte quoad specim. Welwitsch; Hiem, Cat. Afr. PI.

Welw. 1: 314 (1896); Bak.f., Leg. Trop. Afr. 3: 841 (1930)

pro parte quoad specim. Angola.

A. horrida sensu Gossweiler in Agron. Angola 7: 249 (1953),

non (L.) Willd.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

ACACIA GIRAFFAE WILLD. AND ACACIA ERIOLOBA

E. MEY.

Willdenow, when describing Acacia giraffae, Enum.

Hort. Berol.: 1054 (1809), recorded that the species

had been discovered in the interior of the Cape

Province by the celebrated traveller Lichtenstein who

sent him seeds and a dried specimen without flowers.

The description of A. giraffae was based on a sterile

specimen in the Willdenow Herbarium (No. 19171)

in Berlin and the name A. giraffae has been applied

subsequently to one of the dominant and, in many

areas, most characteristic trees of the dry interior

areas of southern Africa, particularly in the dry

Kalahari thorn veld.

Fig. 1. — The holotype of Acacia giraffae Willd. (Wildenow

Herbarium No. 19171). Reproduced by permission of

the Director of the Botanischer Garten und

Botanisches Museum, Berin-Dahlem,

444

NOTES ON AFRICAN ACACIA SPECIES

Fig. 2. — Enlargement of por-

tion of the holotype of

Acacia giraffae Willd. Re-

produced by permission of

the Director of the

Botanischer Garten und

Botanisches Museum,

Berlin-Dahlem.

Examination on microfiche of the holotype of

A. giraffae suggested that the name A. giraffae had

been misapplied as the holotype appeared to differ

significantly from the present concept of the species.

In response to my request Dr. H. Scholz, Botanischer

Garten und Botanisches Museum, Berlin-Dahlem,

to whom I am extremely grateful, compared two

specimens sent to him with the holotype of A. giraffae.

The comparison confirmed that the name A. giraffae

has been widely misapplied as the holotype of A.

giraffae is actually a specimen of what has until now

been called the A. giraffae x A . haematoxylon hybrid

[Ross in Bothalia 10(2): 359-362, 1971], A sterile twig

of Acocks 13190 from 9,6 km WSW of Abrahams

Dam in the northern Cape Province was pronounced

by Dr Scholz to be a good match of the holotype of

A. giraffae.

It seems extraordinary that Lichtenstein collected

a specimen of this convincing Acacia hybrid between

1803-1806 because, although relatively widespread

irt the northern Cape, the hybrid is nowhere common

and it is only as recently as 1946 that it was re-

collected. The unfortunate consequence of the

holotype of A. giraffae being a twig of what has until

now been called the A. giraffae x A. haematoxylon

hybrid is that the name A. giraffae applies to this

hybrid and not to the well-known and widespread

plant for so long known under this name. The latter

plant must now be re-named.

The next available name for the plant previously

known as A. giraffae is A. erioloba E. Mey., Comm. 1 :

171 (1836), which was described from a specimen

collected in Little Namaqualand. The whereabouts

of the type specimen, if it is still extant, is unknown

but E. Meyer’s description clearly identifies the plant.

His reference to the leaves being glabrous is significant

as the leaves of the hybrid plants are clothed with a

fairly dense grey indumentum. As no type of

A. erioloba appears to be extant it is considered

desirable to select a neotype and so preserve the

application of the name. In seeking for a neotype 1

had hoped to find a fruiting specimen collected in

Namaqualand but unfortunately all of the Cape

material at present available is either sterile or in

flower. Consequently, I now select Morris 1042 in the

Kew Herbarium from between Kommandodrif and

Makwassie in the western Transvaal as the neotype

of A. erioloba.

The hybrid plants must now be re-named A.

eriolobaxA. haematoxylon as it is considered un-

desirable to use the name A. giraffae for them since

this may perpetuate some confusion.

The necessary changes in nomenclature are

summarized below:

1. Acacia erioloba E. Mey., Comm. 1:171 (1836),

non A. erioloba Edgw. in J. Asiat. Soc. Beng. 16:

1215 (1847); Harv. in FI. Cap. 2: 280 (1862); Engl, in

Bot. Jahrb. 10: 22 (1888). Type from Namaqualand

(whereabouts unknown); Transvaal, 2726 (Odendaals-

rust), between Kommandodrif and Makwassie

(—AC), J. W. Morris 1042 (K, neo.!).

A. giraffae sensu auct. mult., non A. giraffae Willd., Enum.

Hort. Berol.: 1054 (1809) sensu stricto: Burch., Trav. 2:240

(1824); DC, Prodr. 2: 472 (1825); Harv. in FI. Cap. 2: 280

(1862); Benth. in Trans. Linn. Soc. Lond. 30: 503 (1875);

Marloth in Trans. S. Afr. Phil. Soc. 5: 271 (1889); Schinz in

Mem. Herb. Boiss. 1 : 108 (1900); Sim, For. FI. Cape Col.: 213,

t.58 (1907); Burtt Davy in Kew Bull. 1908: 157 (1908); Glover

in Ann. Bolus Herb. 1: 148, t.18/1 (1915); Harms in Engl.,

Fflanzenw. Afr. 3 (1): 352 (1915); Dinter in Feddes Repert. 15:

79 (1917); Pole Evans in S. Afr. J. Sci. 17: figs. 35, 36 (1920);

Burtt Davy in Kew Bull. 1922: 327 (1922); Marloth, FI. S. Afr.

2: 54, tt.l8D, 19 (1925); Bak.f., Leg. Trop. Afr. 3: 835 (1930);

Burtt Davy, FI. Transv. 2: 340, fig. 59 (1932); Hutch., Botanis,

in S. Afr.: 17?, 341, 386, 412, 418, 424, 425, 481, 543, 547t

cum photogr. (1946); West in Rhod. Agric. J. 47: 206 (1950);

O. B. Miller in J. S. Afr. Bot. 18: 21 (1952); Pardy in Rhod.

Agric. J. 50: 4 (1953); Torre in Consp. FI. Angol. 2: 281 (1956);

Story, Mem. Bot. Surv. S. Afr. 30: 23 (1958); Leistner in Koedoe

4: 101 (1961); Palmer & Pitman, Trees S. Afr.: 153, tt. vi, 34,

35 (1961); F. White, For. FI. N. Rhod.: 84, fig. 17L (1962);

von Breitenbach, Indig. Trees S. Afr. 2: 292 (1965); De Winter

ei a!., 66 Transv. Trees: 46 (1966); Leistner, Mem. Bot. Surv.

S. Afr. 38: 67, 123, tt. 21, 23, 25, 28, 30, 36, 38, 44, 48 (1967);

Schreibcr in FI. S.W. Afr. 58: 8 (1967); Brenan in FI. Zamb.

3,1: 93, 1. 15/10 (1970); Ross in Bothalia 10(2): 359 (1971);

in Bothalia 10(4): 547 (1972); Palmer & Pitman, Trees S. Afr.

2: 769 (1973); Schrciber in Mitt. 'Bot. Staatssamml. Munchen

11: 1 17 (1973).

2. Acacia erioloba E. Mey. x Acacia haematoxylon

Willd.

A. giraffae Willd., Enum. Hort. Berol.: 1054 (1809). Type:

Interior of the Cape Province, Herb. Willdenow 19171 (B, holo.).

A giraffae Willd. X A. haematoxylon Willd., Leistner, Mem.

Bot Surv. S. Afr. 38: 67, 123, t.24 (1967); Ross in Bothalia

10(2): 359 (1971).

J. H. ROSS

445

ACACIA HIRTELLA E. MEY VAR. INERMIS WALP.

Walpers, in Linnaea 13: 542 (1839), based his

description of A. hirtella var. inermis on a specimen

collected by Mund somewhere in the Cape Province.

Unfortunately the type specimen has not been traced.

There is a Mund specimen from the Cape Province

in the Kew Herbarium, but, as the flowering branches

are armed with paired spines, it is assumed that it

cannot be an isotype. A. hirtella is, of course, now

regarded as a synonym of A. karroo Hayne. As

flowering twigs of A. karroo are fairly often devoid

of spines no justification is seen for upholding var.

inermis and it is now relegated to synonymy.

Acacia karroo Hayne, Arzneyk. Gebr. Gewachse

10: t.33 (1827). Type: South Africa, Cape Province,

Herb. Willdenow 19184 fol. 2 (B, lecto.).

A. hirtella E. Mey. var. inermis Walp. in Linnaea 13: 542

(1839). Type: Cape Province, locality unknown, Mund (where-

abouts unknown).

ACACIA ROBECCHII PIROTTA

Mention was made by Pirotta in Bull. Soc. Bot.

Ital. 1893: 61 (1893) of some specimens of two species

of Acacia collected by Robecchi-Bricchetti on his

travels in north-east Africa. The first species collected

in 1889 on the road from Zeila to Gildessa in the

country of the Danakil appeared new and Professor

Pirotta called it A. robecchii. He failed to provide a

description and A. robecchii has remained a nomen

nudum. The second species was found in the Ogaden

and was said to be related to A. fistula Schweinf.

[A. seyal Del. var. fistula (Schweinf.) Oliv.], but

quite distinct. Both species were said to be armed

with swollen spines or “ant-galls”.

In an attempt to establish the identity of A. robecchii

all of the specimens collected by Robecchi-Bricchetti

in the above-mentioned localities were received on

loan through the courtesy of Prof. Dr C. H. Steinberg,

Conservator of the Herbarium Universitatis

Florentinae.

Among the specimens received were two collected

in 1889 on the road from Zeila to Gildessa.

Unfortunately both are appalling specimens. One,

Robecchi-Bricchetti 266, consists of an almost leafless

sterile twig armed with paired straight stipular spines.

The absence of a peeling epidermis suggests that the

specimen is probably referable to A. ehrenbergiana

Hayne. The other specimen, Robecchi-Bricchetti 269,

consists of a sterile young branch. Sterile branches

bearing young foliage are extremely difficult to place

with certainty but the specimen is probably referable

either to A. edgeworthii T. Anders, or to A. tortilis

(Forsk.) Hayne subsp. spirocarpa (Hochst. ex A.

Rich) Brenan. As all of the paired spines are long

and straight, and as there are none of the short

recurved spines usually associated with A. tortilis,

I am inclined to the view that the specimen is referable

to A. edgeworthii but the material is too poor to

come to a definite decision. Neither of these specimens

has inflated “ant-galls” and so neither could be the

specimen on which the name A. robecchii was based.

The identity of A. robecchii therefore remains

unknown.

The second species mentioned by Pirotta is

represented by the specimen Robecchi-Bricchetti 540

from the Ogaden. The powdery, flaking, yellowish

epidermis on the young branch, the deeply bilobed

“ant-galls”, and leaflet venation indicate that the

specimen is referable to A. zanzibarica (S. Moore)

Taub. var. microphylla Brenan.

MIMOSA CAPENSIS BURM. F.

Mimosa capensis Burm.fi, Prodr. FI. Cap.: 31

(sphalm. 27) (1768), which was based on a figure

published by Plukenet in his Phytographia, 1. 1 23

fig. 2 (1692), has previously been rejected as a name

of uncertain application as the plant illustrated (see

Fig. 3) cannot be positively identified (Verdoorn in

Bothalia 6: 411, 1954; Ross in Bothalia 10: 386,

1971). It seems desirable, however, to elaborate on

the reason for doing so.

The MS Journal (‘Dag Register') of Simon van

der Stel’s Expedition to Namaqualand in 1685-6

was discovered at Trinity College, Dublin, Ireland,

and is thought to be the original by Waterhouse who

published a book on it in 1932. The artist Heinrich

Claudius accompanied the Simon van der Stel

Expedition to Namaqualand and a plant illustrated

on that journey (TCD No. 807) is reproduced here

in black and white as Fig. 4. A translation of the

notes accompanying the drawing TCD No. 807

(Waterhouse, Simon van der Stel’s Journal of His

Expedition to Namaqualand 1685-6: 166, 1932)

reads as follows: —

“This tree grows in such abundance in Namaqua-

land that almost all the forests are composed of it.

On account of its multitude of hurtful thorns we

call it Thorn Tree, whereas the natives call it Choe .

It is moderately tall and large but crooked, and it

has good, hard, useful wood. It is found only along

rivers and brooks. Its flowers have a remarkably

pleasant smell and they are followed by a pod

containing a few flat seeds, the effects of which

are so far unknown.”

Along the route followed by the van der Stel

expedition Claudius would certainly have encountered

the plant that is now known as Acacia karroo Hayne.

The only other Acacia species armed with paired

stipular spines and with flowers in round heads that

he may possibly have encountered was A. erioloba

E. Mey. However, the illustration attributed to

Claudius (Fig. 4) bears little actual resemblance to

A. karroo, to A. erioloba or to any other South

African Acacia species. The leaves are shown to be

consistently imparipinnately compound whereas in

all of the indigenous South African Acacia species

the leaves are always paripinnately compound, and

the pods illustrated are at variance with those of

A. karroo and of A. erioloba. Father Tachard, who

visited the Cape in 1685, is quoted by Karsten, The

Old Company’s Garden: 89 (1951), as having said

of Claudius that “He draws and paints animals and

plants to perfection.” As Claudius was an artist of

such high repute it seems odd that his illustration is

inaccurate in several obvious and significant respects

and bears so little actual resemblance to any of the

Acacia species. That is, of course, if the painting was

executed by Claudius and at present there is no reason

to doubt that it was.

Plukenet’s 1. 1 23 fig. 2, referred to by Burm.fi, is

almost identical to the illustration executed by

Claudius on the Namaqualand expedition. Plukenet’s

illustration differs chiefly in that it has been reversed

from left to right, i.e., the leaves, inflorescences and

pods are depicted facing in the other direction. In

addition, Plukenet has added a loose inflorescence,

a loose pod and two more loose seeds. The Claudius

drawings are known to have been copied and the

copies copied and a set of drawings was presented

to D. H. Compton, the Right Reverend the Bishop

of London from 1675-1713, while his lordship was

attending a Congress in Amsterdam in 1691. Both

Petiver and Plukenet has acccess to the drawings in

446

NOTES ON AFRICAN ACACIA SPECIES

Fig. 3. — Photograph of Plu-

kenet, Phytographia 1. 123

figs. 1 and 2 (1692).

Bishop Compton’s possession. The close similarity

between the Claudius and Plukenet illustrations

suggests that Plukenet copied Claudius’s drawing:

no specimen on which Plukenet could have based

the illustration has been located in the Sloane

Herbarium in the British Museum (Natural History),

although this does not, of course, provide proof that

Plukenet copied the Claudius illustration. It does,

however, strengthen the argument that Plukenet

copied an illustration and not an actual specimen.

Aloe and Gladiolus paintings prepared by Claudius

are known to have been copied by Petiver and by

Plukenet (Reynolds, Aloes of South Africa: 18, 27,

29, 1950; Lewis et al in J. S. Afr. Bot. Suppl. 10:

xxii, 1972) and it is therefore a reasonable assumption

that the Plukenet figure reproduced here as Fig. 3

was also copied.

The significance of this is that Plukenet’s 1. 123

fig. 2 is a copy of a painting of a Cape plant. It could

perhaps be argued that the plant depicted by Claudius

was in all probability the Cape Acacia (A. karroo).

Indeed, the painting was identified as such in J. S. Afr.

Bot. 13: 10 (1947). However, as the illustration

cannot be positively identified it seems most

undesirable to accept that the plant depicted is the

Cape Acacia (A. karroo) on the strength of circum-

stantial evidence alone and thereby supplant the

unquestionable A. karroo, a species typified by a

specimen in the Willdenow Herbarium in Berlin,

with the questionable epithet “capensis”. Mimosa

capensis, which was based solely on Plukenet 1. 1 23

fig. 2, is therefore rejected as a name of uncertain

application.

It is interesting and perhaps significant that the

plant depicted by Plukenet in his Phytographia 1. 1 23

fig. 1 (see Fig. 3 above) is A. karroo. The figure was

based on a sterile twig of A. karroo, Herb. Sloane

Vol. 99, fol. 3 in the British Museum (Natural

History), and is a good representation of it. Burm.f.,

Prodr. FI. Cap.: 31 (1768), quoted Plukenet 1. 123

fig. 1 under the name Mimosa nilotica but this was

an incorrect identification by Burman, and Linnaeus,

Sp. PI. 1: 521 (1753), cited this same Plukenet figure

in synonymy under Mimosa scorpioides.

MIMOSA RETICULATA L.

Analysis of the protologue of Mimosa reticulata

L., Mant. 1: 129 (1767), indicates that it is based on

discordant elements. The diagnostic phrase-name and

the fairly comprehensive description that follows were

taken from a plant in cultivation in the Botanic

Garden in Uppsala but in the synonymy reference is

made to Boerhaave and to Plukenet’s Phytographia

1. 123 fig. 2, the latter being the same figure that

Burman, Prodr. FI. Cap.: 31 (1768), referred to under

his Mimosa capensis. Linnaeus’s description of the

pod of M. reticulata (“Fructus ovalis, palmaris,

latitudine semipalmaris, compressus, seminibus sparsis

magnis”) is altogether at variance with the pod figured

by Plukenet and although Linnaeus was under the

impression that his living plant and the plant depicted

by Plukenet represented the same species this was

clearly not the case. Unfortunately no specimen of

the plant that Linnaeus had in cultivation in Uppsala

appears to be extant and the species cannot be

positively identified from the description alone. The

name M. reticulata, which applies to the plant culti-

vated in Uppsala, must therefore be rejected as a

name of uncertain application. The epithet

“reticulata” was apparently taken from the descrip-

tion of the pod in the Boerhaave synonymy.

J. H. ROSS

447

Fig. 4. — Photograph of TCD

No. 807 (Reproduced

from Waterhouse, Simon

van der Stefs Journal

of his Expedition to

Namaqualand 1685-6).

MIMOSA SENEGALENSIS FORSK.

The binomial Mimosa senegalensis was first

published by Houttuyn, Nat. Hist. 3: 614 (1774).

It is not known whether Houttuyn's new binomial,

although not indicated as such, was published

inadvertently or deliberately, but, as M. senegalensis

Houtt. was based on M. Senegal L., Sp. PI. 1: 521

(1753), the name was superfluous when published

and therefore illegitimate. M. senegalensis Forsk.,

FI. Aegypt. Arab.: 176 (1775), a later homonym of

M. senegalensis Houtt. and therefore illegitimate,

referred to quite a different plant. Owing to its

illegitimacy, however, the name M. senegalensis

Forsk. cannot be taken up and so the correct name

for the plant described under this name by Forskal

is Acacia hamulosa Benth. in Hook., Lond. J. Bot. 1 :

509 (1842).

Roberty, in Candollea 11: 120 (1948), discussed

the idea of making a new combination with M.

senegalensis Houtt.: — “L’on pent done etre tente

de creer un nouveau binome: Acacia senegalensis

(Houtt.) x . . ., comb. nov. Mais cette solution ne

presente aucun avantage pratique et ne correspond

guere qu’a unde subtilite, assez vaine, d’erudition.”

Of course such a new combination would also be

illegitimate.

26700- 6

Bo th alia 11, 4: 449-451 (1975)

The typification of Mimosa Senegal

J. H. ROSS"

ABSTRACT

Analysis of the protologue of Mimosa Senegal L. in Species Plantarum ed. 1: 521 (1753) indicated that

it embraced two different elements. The absence of a type specimen or the existence of an illustration from

which Linnaeus could have drawn up his diagnostic phrase-name necessitated the selection of a neotype to

preserve the application of the name M. Senegal.

The protologue of Mimosa Senegal L., Sp. PI. ed. 1 :

521 (1753), the basionym of Acacia Senegal (L.)

Willd. in L., Sp. PI. ed. 4, 4: 1077 (1806), is as follows:

“29. MIMOSA spinis ternis: intermedio reflexo, Senegal.

foliis bipinnatis, floribus spicatis.

Mimosa aculeata, floribus polyandris spicatis,

legumine compresso laevi elliptico. Adanson ,

ex B. Jussieu.

Mimosa spinis geminis distinctis, foliis dupli-

cato-pinnatis: partialibus utrinque quino-

pluribus. Hort. cliff'. 209. Hort. ups. 146. Roy.

higdb. 471 (sphalm. 411). Mat. med. 261.

Acacia. Batih. pin. 392. Alp. aegypt. 9. 1. 9.

Acacia altera vera, siliqua longa villosa, cortice

candicante donata. Pluk. aim. 3. p. 251./. 1.

Habitat in Arabia.

Cortice a/bo distinguitur liaec species primo

intuitu.

Spinae ad folii exortum ties."

The reference in the diagnostic phrase-name to a

plant armed with three spines,! with the central one

recurved, and in the synonymy to a plant with paired

spines, makes it manifestly clear that Linnaeus's

concept of M. Senegal in the Species Plantarum

embraced two quite different elements.

The diagnostic phrase-name, namely, “MIMOSA

spinis ternis: intermedio reflexo, foliis bipinnatis,

floribus spicatis”, together with the Adanson syn-

onymy, namely, “Mimosa aculeata, floribus poly-

andris spicatis, legumine compresso laevi elliptico.

Adanson, ex B. Jussieu ”, appear for the first time in

the Species Plantarum and did not originate in any of

Linnaeus’s earlier works. Much of the synonymy,

however, originated in and was taken in a modified

form from Hortus Cliffortianus: 209 (1738). It is

quite clear therefore that to his earlier basic concept

of a species armed with paired spines in Hortus

Cliffortianus, Linnaeus subsequently added in the

Species Plantarum the diagnostic phrase-name of a

species armed with three spines. The description in

Hortus Cliffortianus is as follows:

“10. MIMOSA spinis geminatis, foliis duplicato-pinnatis.

Acacia altera vera f. Spina mazcatensis vel arabica, foliis

angustioribus, flore albo, siliqua longa villosa plurimis

isthmis & cortice candicantibus donata. Pluk. aim.

3. /.251. /. 1.

Acacia, Sant & Akakia. Alp. aegyp. 6. t. 6.

Acacia vera. Bauh. hist. 1. p. 429.

Crescit in Arabia forte et ad Caput bonae spei, unde

semina habuimus varia et hanc iis immixtam; enata

etiam fuit e seminibus virginianis per D. Gronovium

communicatis.

Folia quatuor, quinque vel sex paria singulo petiolo com-

muni insident pinnatim, singulapinnata numerosis pinnis;

ad exortum petioli communis spinae duae oppositae.

Differt a 5 ta specie, cum qua confunditur a plurimis spinis

in hac minoribus minu que rigidis, et foliis partialibus

quam quinque paribus

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

t In Mimosa Senegal the plants are actually armed with

prickles and not with stipular spines. However, for the purpose

of this discussion the term spine is retained to obviate any

confusion which might arise from a change in terminology.

Plukenet’s excellent illustration in Almagestum

Botanicum 4: t.25 1 fig. 1 to which Linnaeus referred

depicts a plant armed with paired stipular spines,

with bipinnate leaves, flowers in round heads and

distinctly moniliform pods. This illustration makes it

perfectly clear that the plant Linnaeus had in mind in

Hortus Cliffortianus was Mimosa nilotica [Acacia

nilotica (L.) Willd. ex Del.]. The descriptive phrase in

Hortus Cliffortianus, namely, “Acacia altera vera f.

Spina . . .”, was copied by Linnaeus, without

additional information, from Plukenet’s Almagestum

Botanicum 2: 3 (1696). It is most unlikely that

Linnaeus saw the actual specimen drawn, but although

he gives a diagnostic phrase, there is nothing in it that

could not have been obtained from a study of

Plukenet’s figure. The plant described and illustrated

by Alpini, referred to by Linnaeus in his synonymy

in Hortus Cliffortianus, is Acacia nilotica, as is the

plant described by Bauhin in his Hist. Plant. 1 : 429.

Analysis of the protologue of M. Senegal in the

Species Plantarum reveals that the third descriptive

phrase, namely, “Mimosa spinis geminis . . .”,

is based on the diagnostic phrase-name in Hortus

Cliffortianus. This descriptive phrase, together with

the additional synonymy (with the exception of the

Adanson synonymy) cited by Linnaeus in Species

Plantarum, refers to Acacia nilotica. Therefore, with

the exception of the diagnostic phrase-name and the

Adanson synonymy, all of the protologue of Mimosa

senega! in Species Plantarum refers to Acacia nilotica.

For the purpose of the typification of M. Senegal

the diagnostic phrase-name is the most important of

the constituent elements in Linnaeus's protologue.

Analysis of this diagnostic phrase-name indicates that

Linnaeus had before him a specimen armed with

three spines, the central one of which was recurved,

and bearing bipinnate leaves and flowers in spikes.

The assumption that Linnaeus had before him a speci-

men and not an illustration is based on the fact that

no illustration of such a plant was published prior

to 1753, or for many years subsequently. Unfortu-

nately, however, there is no specimen of M. Senegal

preserved in the Linnaean Herbarium in London, or

in the Linnaean collections in Stockholm.

It seems almost certain that Linnaeus based the

diagnostic phrase-name of M. Senegal on a specimen

collection by Michel Adanson in Senegal between

1749 and 1753. Bentham, in his revision of the

Mimoseae in Trans. Linn. Soc. Lond. 30: 516 (1875),

was of the same opinion: “It appears to me evident

that Linnaeus, in characterizing his M. Senegal,

had in view the plant brought by Adanson from

Senegal, as furnishing the best gum arabic of com-

merce, and which also constituted the M. senegalensis

of Lamarck. Linnaeus’s reference to Adanson and

to the three spines, with the central one recurved,

identify the species, notwithstanding the confusion

thrown on it by the various synonyms applying to

almost as many different plants, and his note that it

was easily known by its white bark, which has induced

the false reference to A. albida."

450

THE TYPIFICATION OF MIMOSA SENEGAL

<u„

0 to

M»Wi M.

ADA VS

F'G' n eotype° of8 * Mimosa (Herb- AdanS°n No‘ 16899) in the Paris Herbarium elected as the

J. H. ROSS

451

Linnaeus was a correspondent of B. Jussieu, but

not of Adanson, and it is quite possible that B.

Jussieu sent Linnaeus some notes and a specimen of

M. Senegal that he had received from Adanson.

Lamarck, Encycl. Meth. Bot. 1: 19 (1783), gives all

of the credit for this species to Adanson, mentioning

that the latter had provided the only good description

of this interesting plant, and that Linnaeus had

wrongly added other discordant elements to Adanson's

good description.

As the specimen on which Linnaeus based his

diagnostic phrase-name of M. Senegal is no longer

extant, and as there is no illustration from which

Linnaeus could have drawn up this phrase-name, the

selection of a suitable specimen (neotype) to preserve

the application of the name is desirable. In searching

for a representative specimen of M. Senegal it was

considered desirable to select a specimen collected

in Senegal, and, if possible, by Adanson. Fortunately

an excellent specimen of M. Senegal collected by

Adanson (No. 59c) in Senegal in 1749 is preserved

in the Adanson Herbarium (No. 16899) in Paris

(see Fig. 1), together with several duplicates of the

same gathering.

This Adanson specimen shows quite clearly the

three spines with the central one recurved, the

bipinnate leaves, and the spicate inflorescence to

which Linnaeus referred in the diagnostic phrase-name

of M. Senegal in the Species Plantarum. The leaves

have up to 5 pinnae pairs. This excellent specimen

(Herb. Adanson No. 16899), which will preserve the

traditional and current application of the name

M. senega/, is now selected as the neotype of M.

senega!.

ACKNOWLEDGMENT

I am extremely grateful to Dr H. Heine, Labora-

toire de Phanerogamie, Museum National d’Histoire

Naturelle, Paris, for conducting a search in the

Adanson Herbarium in response to my request, for

arranging for the specimens of M. Senegal collected

by Adanson to be photographed, and for supplying

some detailed information.

Bothalia II, 4: 453-462 (1975)

The Acacia Senegal complex

J. H. ROSS*

ABSTRACT

A brief account of the present state of knowledge of the species in the Acacia Senegal complex is given. A

short description of each species is provided together with a key to the identification of the species. Attention

is drawn to the taxonomic difficulties encountered within the complex.

INTRODUCTION

The Acacia senega I complex is taken to include all

of those Acacia species with spicate inflorescences and

armed at or near the nodes with prickles either in

threes, with the central one typically hooked down-

wards and the two laterals curved upwards or at

times spreading laterally, or else solitary, the two

laterals being absent. Occasionally all three prickles

may point up or down but the presence of the three

prickles at or near the nodes will identify the specimen

as a member of this complex. Very rarely and in only

one species (A. ankokib) are specimens sometimes

unarmed. Whether the prickles occur singly or in

threes near the nodes in some species appears of no

significance as both arrangements may occur on one

and the same shoot, although some gatherings may

show all or nearly all the prickles arranged singly.

The species currently recognized within this taxonomi-

cally difficult complex, in chronological order, are:

A. Senegal (L.) Willd., A. asak (Forsk.) Willd., A.

hamulosa Benth., A. oliveri Vatke, A. soma/ensis

Vatke, A. hunteri Oliv., A. dudgeoni Craib ex Hoik,

A. thomasii Harms, A. condyloclada Chiov., A.

ogadensis Chiov., A. caraniana Chiov., A. cheilanthi-

olia Chiov., A. zizyphispina Chiov. and A. ankokib

Chiov.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Apart from A. Senegal which is widespread through-

out Africa and extends into Arabia, and A. dudgeoni

which is confined to tropical West Africa, the re-

maining species in the complex are concentrated in

north east tropical Africa and Arabia. The difficulties

within this complex are aggravated by the paucity of

the material of several of the species and the inade-

quate field notes on many of the existing specimens.

Not only are the limits of some of the species not

very clearly defined, but fhe range of morphological

variation within certain species is poorly understood.

Field studies in north east tropical Africa are essential

to clarify many of the problems encountered in this

complex.

The aim of this paper is to provide an account of

the present state of knowledge of the species in this

complex and to draw attention to the gaps in our

information in the hope that future collectors will

concentrate on gathering the required information

and material. A short description of each species is

provided and the difficulties within each species are

discussed. Only the most important synonyms and

brief literature references are included as it is the

intention to deal with these in detail elsewhere. As

both flowering and fruiting material is often not

available, in the key to the identification of the

species the dichotomies are based, as far as is possible,

on vegetative characters.

KEY TO SPECIES

All or most leaflets more than 3 mm wide:

Leaves with 1 pinna pair:

Calyx 3-4 mm long; corolla 4-6 mm long; stamen-filaments 13-18 mm long; slender straggling virgately

branched tree; prickles in threes; leaflets (2)3-4 pairs per pinna 12. A. ogadensis

Calyx 1-1 ,5 mm long; corolla 1 ,75-2,25 mm long; stamen-filaments up to 5 mm long; shrub or small

tree with a rounded crown; prickles solitary or in threes; leaflets 4-9 pairs per pinna. ... 8. A. oliveri

Leaves with 2-6 pinnae pairs:

Leaflets 1 or 2 pairs per pinna, 8-19 x 5-13 mm, obovate, minutely appressed-puberulous on both

surfaces; pods (2, 6)3, 2-4, 8 cm wide 15. A. zizyphispina

Leaflets 3-20 pairs per pinna:

Leaves with 2 pinnae pairs 8. A. oliveri

Leaves with 3-6 pinnae pairs:

Petiole with a large flattened ±discoid gland 1-3 mm in diameter at the base of the petiole or at

the point of attachment of the lowest pinna pair; internodes 5-9 cm long:

Petiole 1 ,3-2,8 cm long, the gland basal ; leaflets 3-9 mm wide, minutely appressed-puberulous on

bothsurfaces or the lower only, rarely glabrous; pods 1 ,8-2, 1 cm wide, densely puberulous

13. A. condyloclada

Petiole 2-6 mm long, the gland at the point of attachment of the lowest pinna pair; leaflets

3^1 mm wide, glabrous; pods 0,9-1 cm wide, glabrous 14. A. caraniana

Petiole with a small slightly raised gland < 1 mm in diameter which is variable in position but

seldom situated as above; internodes mostly <2,5 cm long, seldom up to 4,5 cm long

7. A. asak

All or most leaflets less than 3 mm wide:

All leaves with 1 or 2 pinna pairs:

Leaflets 2 (rarely 3) pairs per pinna, glabrous; prickles solitary 9. A. soma/ensis

Leaflets 3-15 pairs per pinna, if only 3 pairs present then lower surface fairly densely clothed with

appressed hairs and not glabrous as above; prickles solitary or in threes:

454

THE ACACIA SENEGAL COMPLEX

Leaflets less than 1 mm wide:

Leaves with 2 pinnae pairs; leaflets 6-8 pairs per pinna; prickles in threes; calyx puberulous;

confined to Aden 1. A. hunteri

Leaves with 1 (rarely 2) pinna pairs; leaflets 8-15 pairs per pinna; prickles solitary ; calyx glabrous;

confined to Somalia 4. A. sp.

Leaflets 1-2, 5(3) mm wide:

Calyx 3^4,5 mm long; corolla 6,5-7 mm long; stamen-filaments 13-15 mm long; leaflets 7-15

pairs per pinna; straggling shrub or slender tree with elongate whippy branches 11. A. thomasii

Calyx 1-2,25 mm long; corolla 1 ,75-3,25 mm long; stamen-filaments up to 6 mm long; leaflets

3-9 pairs per pinna; shrub or small tree with a rounded crown:

Leaflets 3-5 pairs per pinna, the lower surface fairly densely clothed with conspicuous whitish

appressed-hairs; inflorescence axis clothed with spreading hairs basally at least; petiole

and leaf-rhachis clothed with spreading hairs; leaves with 1 pinna pair

10. A. sp. near somalensis

Leaflets 4-9 pairs per pinna, glabrous; inflorescence axis sparingly to densely puberulous

throughout or basally only; leaves with 1 or 2 pinna pairs 8. A. oliveri

All or most leaves with 3-16 pinnae pairs:

Leaves with 3-6 pinnae pairs:

Leaf-rhachillae each with a recurved prickle on the lower surface at or just below the apex. . . .

5. A. hamulosa

Leaf-rhachillae without a recurved prickle on the lower surface but sometimes with a prickle

terminating the leaf-rhachis:

Pinnae with 3-4 (rarely 5) pairs of leaflets 6. A. cheilanthifolia

All or most pinnae with 5-25 pairs of leaflets:

Leaves very small, petiole and rhachis together <1 cm long 1. A. hunteri

Leaves larger than above, petiole and rhachis together >1 cm long:

Corolla 6,5-7 mm long; calyx 3-4,5 mm long; stamen-filaments 13-15 mm long; straggling

shrub or slender tree with elongate whippy branches; leaves with 3 pinnae pairs..

11. A. thomasii

Corolla 1,5-4 mm long; calyx 1-3,5 mm long; stamen-filaments up to 8 mm long; leaves

with 3-6 pinnae pairs:

Petiole with a large flattened + discoid gland 1-3 mm in diameter at the point of attach-

ment of the lowest pinna pair; leaves with 3 pinnae pairs; internodes 5-8,5 cm long;

pods 0,9-1 cm wide, glabrous 14. A. caraniana

Petiole with a small often slightly raised gland < 1 mm in diameter which is variable in

position but seldom situated as above, or eglandular; leaves with 3-6 pinnae pairs;

internodes mostly <4,5 cm long:

Bark on young branchlets greyish, the outer layer flaking away to expose a papery

peeling yellowish inner layer; prickles often absent; leaflets (4)6-9 pairs per pinna,

glaucous, minutely puberulous on both surfaces; corolla 1,5-2 mm long; pods

0,9-1 cm wide; confined to Somalia 16. A. ankokib

Epidermis of young branchlets not flaking away to reveal a papery peeling yellowish

inner layer; prickles invariably present; leaflets (3)7-25 pairs per pinna, not minutely

puberulous on both surfaces; corolla 2,5-4 mm long; pods (0,9)1, 1-3, 4 cm wide:

Leaflets 1,25-3 mm wide, glaucous, mostly glabrous throughout or sometimes with

a small basal tuft of hairs on the lower surface; pinna widely spaced, >(5) 7 mm

apart giving the leaf an “open” look; petiole and leaf-rhachis glabrous or

sparsely pubescent but hairs not spreading conspicuously; pods (0,9)1 , 1 —

1 ,5(1 ,8) cm wide, glabrous; confined to Ethiopia, Sudan and Arabia. .. . 7. A. asak

Leaflets 0,5-1,75 mm wide, very rarely wider, glabrous or appressed-pubescent

beneath, with or without spreading marginal cilia; pinna usually < 5 mm apart;

petiole and leaf-rhachis mostly sparingly to densely clothed with spreading

hairs; pods (1,2) 1,5-3, 4 cm wide, sparingly to densely appressed puberulous

or pubescent, seldom glabrous; widespread 3. A. senega!

Leaves with 7-16 pinnae pairs:

Leaves with 7-16 pinnae pairs; pods almost glabrous when mature; confined to tropical West

Africa 2. A. dudgeoni

Leaves with 7-10(12) pinnae pairs; pods sparingly to densely appressed puberulous or pubescent

when mature; absent from tropical West Africa 3. A. senega!

1. Acacia hunteri Oliv. in Hook., Icon. PI. 14: 36,

1. 1 350 (1881). Type: Neighbourhood of Aden, F.

Hunter 156 (K, holo. !).

Small shrub; young branchlets greyish, as though

whitewashed over a purplish background. Prickles

in threes, all three pointing upwards. Leaves very

small: petiole up to 4 mm long, pubescent; rhachis

i:p to 6 mm long, pubescent; pinnae 2-3 pairs;

rhachillae up to 8 mm long; leaflets 6-8 pairs per

pinna, up to 2x0,75 mm, glabrous. Inflorescence

axis up to 1,5 cm long, pubescent. Calyx ±1,5 mm

long, puberulous. Corolla ±2 mm long. Pods yellow-

ish- or reddish-brown, up to 3x0,9 cm, compressed,

valves brittle, inconspicuously venose, sparingly

pubescent throughout or on margins and stipe only.

A. hunteri is known only from the type collection.

It differs from A. Senegal in being very small in all of

its parts and the suggestion has been advanced that

perhaps A. hunteri is no more than a diminutive

variant of A. senega!. More material is desired.

2. Acacia dudgeoni Craib ex Holl. in Kew Bull.,

Add. Series ix: 291 (Dec. 1911); Craib in Kew Bull.

1912: 98 (1912); Keay & Brenan in Kew Bull. 1949:

129 (1949). Syntypes: Nigeria, Borgu, Dudgeon

58 (K!); Kontagora, Dalziel 41 (K!).

A. samoryana A. Chev. in Bull. Soc. Bot. Fr., Mem.viiiD:

167 (2nd March 1912). Type: Dahomey, between Firou and

Konkobiri, Chevalier 24326 (P, holo.!).

A. Senegal subsp. semgalensis (Houtt.) Roberby var. samor-

yana ( A. Chev.) Roberty in Candollea 11: 157 (1948). Type as

above.

Shrub or small tree to 4 m high, seldom up to

8 m high; bark grey or brown, rough; young branch-

lets brown or reddish-brown, densely pubescent, bark

on older branches grey to yellowish-brown or brown,

flaking minutely. Prickles in threes, the central one

hooked downwards and the two laterals curved

upwards, or occasionally solitary. Leaves: petioles,

rhachides and rhachillae densely clothed with spread-

ing hairs; petiole 0,4-1, 5 cm long, glandular;

J. H. ROSS

455

rhachis 2,2-5 cm long, usually with a small gland

at the junction of the top 1-3 pinnae pairs, often with

scattered recurved prickles on the lower surface;

pinnae 7-16 pairs; rhachillae 1-2,4 cm long; leaflets

13-25 pairs per pinna, 1, 5-5x0, 5-1 mm, linear-

oblong, oblong or slightly broader apically, apex

rounded, glabrous or with marginal cilia. Inflorescence

axis densely pubescent, up to 9 cm long. Calyx

1-1 ,75 mm long, glabrous to fairly densely pubescent.

Corolla 1,5-2,75 mm long, glabrous. Stamen-fila-

ments up to 6 mm long. Pods yellowish-brown or

brown, 5-10 x 1 ,7-2,5 cm, straight or slightly curved,

rounded to acute apically, venose, glabrous when

mature.

Found in Mali, Ivory Coast, Ghana, Dahomey and

Nigeria. Occurs in moister regions than A. Senegal.

The following is a selection of specimens examined

in the Kew Herbarium:

Ivory Coast. — Bouna Wildlife Reserve, 43 km east of

Ouangofetini on road to Bouna, Wilde & Leeuwenberg 3497.

Ghana. — Wenchi-Sunyani district, Ashanti, Hepper & Mor-

ton A3200; near Tamale, Lloyd Williams 115; north Gambaga

district, Lloyd Williams 490.

Nigeria. — Bauchi Prov., between Dindima and Gar, Keay

FHI 37900; Bornu Prov., Kojofa district, Kojofa, Rosevear

FHI 26604.

In tropical West Africa A. dudgeoni is distinguished

from A. Senegal in having 7-16 pinnae pairs per leaf

and 13-25 pairs of leaflets per pinna as opposed to

the 3-6 pinnae pairs and 8—1 5 pairs of leaflets of

A. Senegal. The pinnae tend to be more crowded on

the rhachis in A. dudgeoni giving the leaf a somewhat

different “look” to the leaves of A. Senegal.

In southern Africa (Natal), however, A. Senegal

sometimes has up to 12 pinnae pairs per leaf and

up to 24 pairs of leaflets per pinna so that when

viewed on a continental basis the difference in the

number of pinna and leaflet pairs provides no clear

discontinuity between A. dudgeoni and A. Senegal.

In Natal the greater number of pinnae and leaflet

pairs are recorded from plants in more mesic habitats

which is interesting because in West Africa A.

dudgeoni grows in moister habitats than does A.

Senegal. However, whereas the largest leaflets in

Natal are found on plants in mesic habitats, in West

Africa A. dudgeoni tends to have smaller leaflets

than A. Senegal.

The geographical distributions of A. dudgeoni and

A. Senegal in West Africa were reported to be quite

distinct and were plotted by Aubreville, FI. Forest.

Soudano-Guin. : 255, t.23 (1950). Recently, however,

A. dudgeoni has been recorded from Senegal (Berhaut,

FI. Senegal, ed. 2: 46, 1967). I have not seen a specimen

of A. dudgeoni from Senegal and have been unable

to substantiate this reported occurrence of the species

in Senegal. Confirmation of the existence of A.

dudgeoni in Senegal is desired.

3. Acacia Senegal (L.) Willd. in L., Sp. PI. ed.

4, 4: 1077 (1806) excl. icon, et syn. fere omnibus cit. ;

Benth. in Trans. Linn. Soc. Lond. 30: 516 (1875);

Brenan in FI. Trop. E. Afr. Legum.-Mimos. : 92

(1959); in FI. Zamb. 3, 1 : 79 (1970). Type: Senegal,

Herb. Adanson No. 16899 (P, neo.).

Mimosa Senegal L., Sp. PI. 1: 521 (1753) excl. icon, et syn.

fere omnibus cit; Ross in Bothalia 11 : 449 (1975). Type uncer-

tain, presumably a specimen collected by Adanson in Senegal.

Shrub or tree to 15 m high with a slightly rounded

or flattened and somewhat spreading crown, or a

slender spindly tree with irregular virgate branches;

bark yellowish- or greyish-brown to purplish-black,

rough or smooth and papery and peeling off in strips;

young branchlets yellowish- or greyish-brown to

purplish-black, sometimes as though whitewashed

over a darker background, glabrous to densely

pubescent. Prickles in threes, the central one hooked

downwards and the two laterals ± curved upwards,

or else solitary. Leaves: petioles, rhachides and

rhachillae sparingly to densely clothed with spreading

hairs, seldom glabrous; petiole glandular or not,

gland di 0,5-0,75 mm in diam.; rhachis 0,7- 7 cm

long, with a gland at the junction of the top 1-5

pinnae pairs, between each pinna pair or absent from

some, with or without scattered recurved prickles on

the lower surface; pinnae (2)3-8(12) pairs; rhachillae

0,5-4 cm long; leaflets 7-25 pairs per pinna, 1-5(9)

X 0,5-1 ,75(3) mm, linear to elliptic-oblong, sparingly

appressed-pubescent on both surfaces, or glabrous.

Inflorescence axis densely pubescent or glabrous, up

to 12 cm long. Calyx 2-2,75(3,5) mm long, glabrous

to somewhat pubescent. Corolla 2,75-4 mm long,

glabrous or subglabrous. Stamen-filaments up to

7 mm long. Pods yellowish- or greyish-brown to brown,

(1,8)4-19 X 1,2-3, 4 cm, rounded to acuminate

apically, venose, sparsely to densely appressed-

pubescent or puberulous.

A. Senegal is widespread in tropical Africa from

Senegal in the west to Ethiopia and Somalia in the

north-east, southwards to the Transvaal, Swaziland

and Natal, and it extends to India. The most wide-

spread and unquestionably the most variable and

taxonomically difficult species within the entire

complex.

The present delimitation of infraspecific taxa

within A. senega! which of necessity relies partly on

growth form, is most unsatisfactory as numerous

specimens cannot be referred to a particular taxon

with certainty. This is largely the result of inadequate

information concerning the exact nature of the

morphological variation in parts of the species range,

particularly in north-east tropical Africa. When more

information is available the varietal limits may need

to be re-defined. Until such time it is hoped that

the accompanying key to the currently recognized

varieties will enable most specimens to be identified.

Key to varieties

Inflorescence axis glabrous throughout or sometimes with

some basal hairs:

Slender spindly tree with irregular straggling branches or

a well-grown tree with a rounded crown; bark yel-

low, papery and peeling; young branchlets glabrous

or subglabrous; inflorescence axes often purplish

with yellow mottling; flowers often produced before

or with the young leaves (c) var. leiorhachis

Tree with a flattened and spreading or lax rounded

crown; bark not papery and peeling; young branch-

lets sparingly to densely pubescent; inflorescence

axes olive-green; flowers usually produced after

the leaves (a) var. senega!

Inflorescence axis sparingly to densely pubescent:

Tree:

Apex of pod rounded to acute, seldom acuminate

(a) var. Senegal

Apex of pod usually strongly acuminate or rostrate

(d) var. rostrata

Shrub or bush:

Apex of pod rounded to acute, seldom acuminate;

leaves with up to 4 pinnae pairs. ... (b) var. kerensis

Apex of pod usually strongly acuminate or rostrate;

leaves with up to 12 pinnae pairs. . (d) var. rostrata

The four varieties currently recognized within A.

Senegal and the difficulties encountered are discussed

below.

(a) var. Senegal.

Brenan in FI. Trop. E. Afr. Legum.-Mimos.: 93

(1959); Ross in Bol. Soc. Brot., Ser. 2, 42: 226 (1968);

Brenan in FI. Zamb. 3,1: 79 (1970). Type: Senegal,

Herb. Adanson No. 16899 (P, neo.).

456

THE ACACIA SENEGAL COMPLEX

Mimosa Senegal L., Sp. PI. 1: 521 (1753) excl. icon, et syn.

fere omnibus cit. Type uncertain, presumably a specimen

collected by Adanson in Senegal.

M. senegalensis Houtt., Nat. Hist. 3: 614 (1774); Lam.,

Encycl. Meth. Bot. 1 : 19 (1783) excl. ref. Forsk., Aegypt. p. 176

et icon, cit.; non M. senegalensis Forsk., FI. Aegypt. Arab.:

176 (1775). It is not known whether Houttuyn's new binomial,

although not indicated as such, was published inadvertently

or deliberately. However, as M. senegalensis Houtt. was based

on M. senega! L., the name was superfluous when published

and therefore illegitimate. M. senegalensis Forsk., a later

homonym of M. senegalensis Houtt. and therefore illegitimate,

referred to a quite different plant (A. hamalosa).

Acacia verek Guill. & Perr. in FI. Seneg. Tent. 1 : 245, t. 56

(1832). Syntypes from Senegal (PI).

A. rnpestris Stocks ex Boiss., FI. Or. 2: 638 (1872). Type:

Beloutchistan, Stocks 502 (K, holo.l).

A. virchowiana Vatke & Hildebr. in Oesterr. Bot. Zeit. 30:

275 (1880) pro parte quoad fol. et flor. tantum. Type: Kenya,

Teita Distr., Voi River and elsewhere, Hildebrandt 2486 (BM,

iso. !).

A. Senegal var. platyosprion Chiov., FI. Somala 2: 187,

fig. 114 (1932). Syntypes: Somalia, Mogadiscio, Senni 191

(FI!); Chisimaio, Senni 116 (FI!), Senni 201 (FI!); da Bur Gao

e Cu Daio, Senni 87 (FI!); Gobuin, Gorini 421 (FI!); between

Bug Berde and Ato, Guidotti 3 (whereabouts unknown).

Var. Senegal grows as a tree up to 14 m high with

a flat and spreading or lax and rounded crown. The

inflorescence axes are typically pubescent throughout

although occasionally the axes may be glabrous or

subglabrous, and the pods are rounded to acute or

acuminate apically. Var. senega I shows a wide range

of variation in general habit, indumentum, leaflet-

size and flower-size but the arborescent growth form,

pubescent inflorescence axes and pods with rounded

to acuminate apices distinguish the variety. Var.

Senegal is widespread in tropical west, north-east

and east Africa and extends as far south as Mozam-

bique.

(b) var. kerensis Schweinf. in Bull. Herb. Boiss.

4, app. 2: 216 (1896); Brenan in FI. Trop. E. Afr.

Legum.-Mimos. : 93 (1959); Ross in Bol. Soc. Brot.,

Ser. 2, 42: 229 (1968). Syntypes: Ethiopia, near

Keren, Schweinfurth 745 (B|, K!); Bogu valley,

Schweinfurth 741 ( Bt) ; near Djuffa, Schweinfurth

998 (Bf).

As indicated by Brenan, in FI. Trop. E. Afr.: 93

(1959), the application of the name var. kerensis

is not certain and considerable reliance is placed on

Schweinfurth’s reference to the shrubby growth form.

At present var. Senegal and var. kerensis are dis-

tinguished from one another entirely on growth form.

All shrubby specimens of A. senega! in north-east

Africa are referred to var. kerensis which is most

unsatisfactory since it is not known whether var.

kerensis represents a good taxon or whether it merely

embraces a heterogeneous assemblage of shrub-like

growth forms. Occasional specimens described by

collectors as, for example, “a spreading bush or

sometimes a small tree” cannot be referred to a

variety with certainty, while other specimens, which

are morphologically indistinguishable from typical

var. kerensis but which have been described as

“small straggling trees”, are often referred to var.

senega!.

Of the three syntypes of var. kerensis, only one,

namely Schweinfurth 745 from the Afbaron valley

near Keren, has been traced. The specimen of Schwein-

furth 745 in the Kew Herbarium has small leaves

with up to 3 pinnae pairs, pubescent inflorescence

axes, and large pods 9-11,5 X 2-2,6 cm with acute

or acuminate apices. Schweinfurth drew attention

to the frequent presence of a small recurved prickle

on the lower surface of the leaf-rhachis either at or

near the apex even when the rhachis is otherwise

without prickles but this character is often found

in other variants of A. senega!.

Var. kerensis , as delimited at present, accommodates

a heterogeneous assemblage of shrubs and does

not appear to represent a good taxon. As the only

available syntype of var. kerensis has large pods with

acute or acuminate apices, it is proposed that those

plants in north-east Africa with rostrate pod apices

should be referred to var. rostrata Brenan. As men-

tioned above, the necessity to distinguish var. kerensis

from var. senega! on growth form alone is unsatis-

factory because, in the absence of field notes, it is

often not possible to refer a specimen to either

variety with certainty. Var. kerensis does often have

smaller leaves than in var. senega!, but this is an

inconsistent tendency on which no great reliance

can be placed. Until the nature of the variation in

north-east Africa is carefully investigated in the field

and properly understood this unsatisfactory situation

will prevail. The significance of growth form as a

means of delimiting infraspecific categories within

A. senega! requires critical evaluation and it would

be interesting to know whether var. senega! ewer grows

as a shrub. The effect on plants of tapping for gum

is not clear.

Var. kerensis is recorded from Ethiopia, Somalia,

Uganda, Kenya and Tanzania.

(c) var. leiorhachis Brenan in Kew Bull. 8: 98

(1953); Ross & Brenan in Kew Bull. 21: 69 (1967);

Ross in Bol. Soc. Brot., Ser. 2, 42: 231 (1968); Brenan

in FI. Zamb. 3, 1: 80, 1. 1 5 fig. 5 (1970). Type: Tan-

zania, Tanga Prov., Pare Distr., Same, Greenway

2192 (K, holo.l, EA, FHOl, iso.).

A. circummarginata Chiov. in Ann. Bot., Roma 13: 394

(1915); Brenan in FI. Trop. E. Afr. Legum.-Mimos.: 94, fig.

14/18 (1959). Syntypes: Ethiopia, Ogaden, Paoli 794, 913 bis,

920, 1010 (FI!).

A. Senegal var. Senegal sensu Brenan, FI. Trop. E. Afr.

Legum.-Mimos.: 93 (1959) pro parte quoad syn. A. Senegal

var. leiorhachis.

Var. leiorhachis grows either as a slender spindly

tree with irregular straggling branches or as a well-

grown tree with a rounded crown, and the bark is

usually yellow, papery and peeling. The inflorescence

axes are normally glabrous apart from some basal

hairs or glabrous throughout, and the pods have

rounded or acute apices and are several times longer

than they are broad. The glabrous inflorescence

axes distinguish var. leiorhachis from the other

varieties of A. senega!, although it must be borne in

mind that occasional specimens of var. senega! have

glabrous inflorescence axes. Indeed, the possibility

exists that the type specimen of var. leiorhachis is

an atypical specimen of var. senega! with glabrous

inflorescence axes.

In the southern part of the range of distribution

of var. leiorhachis the slender irregular straggling

growth form is sufficient, in many instances, to

distinguish it in the field from the other varieties of

A. senega!. However, in Kenya the situation is more

complicated because var. leiorhachis embraces two

quite different growth forms. In addition to the

slender straggling growth form reminiscent of A.

thomasii, there is a form with non-virgate branching

which grows into a substantial tree with a rounded

crown (A. circummarginata). Although the two

growth forms are strikingly different in the field,

herbarium specimens are extremely difficult to sepa-

rate. The decision to include these two different

growth forms in var. leiorhachis was taken because

of the inability to find any constant distinguishing

morphological characters. Field observations on the

variation in habit assumed by var. leiorhachis in

Kenya, together with notes on ecological preferences,

J. H. ROSS

457

would be welcome. It may ultimately be found desir-

able in the light of new evidence to accord the two

different growth forms separate status.

The slender straggling growth form of some speci-

mens of var. leiorhachis is reminiscent of that exhibited

by A. thomasii and by A. ogadensis. Both A. thomasii

and A. ogadensis differ in having larger flowers and,

in addition, A. ogadensis differs in having leaves

with only 1 pinna pair and (2)3-4 pairs of large leaflets

per pinna.

Var. leiorhachis occurs in Ethiopia, Kenya, Tan-

zania, Zambia, Rhodesia, Mozambique and the

Transvaal.

(d) var. rostrata Brenan in Kew Bull. 8: 99

(1953); Ross in Bol. Soc. Brot. Ser. 2, 42: 233 (1968);

Brenan in FI. Zamb. 3, 1 : 79 (1970). Type: Transvaal,

Zoutpansberg Distr., Dongola Reserve, Verdoorn 2264

(K, holo. !, PRE, iso.!).

A. spinosa Marloth & Engl, in Bot. Jahrb. 10: 20 (1888) nom.

illegit., non A. spinosa E. Mey., Comm.: 170 (1836). Type:

South West Africa, Hereroland, near Usakos, Marloth 1257

(K!, PRE!, iso.).

A. trispinosa Marloth & Engl. [Marloth in Trans. S. Afr.

Phil. Soc. 5: 269 (1893) nom. nud.] ex Schinz in Mem. Herb.

Boiss. 1: 115 (1900) sphalm, nom. illegit., non A. trispinosa

Stokes in Bot. Mat. Med. 3: 168 (1812). Type as for A. spinosa.

A. rostrata Sim, For. FI. P.E. Afr.: 55, t.37a (1909) nom.

illegit., non A. rostrata Humb. & Bonpi. ex Willd. in L., Sp.

PI. ed. 4, 4: 1060 (errore typogr. 1054) (1806). Type: Mozam-

bique, Lourenco Marques & Maputo, Sim 6263 (whereabouts

unknown).

A. oxyosprion Chiov. var. oxyosprion , FI. Somala 2: 188,

fig. 115 (1932). Type: Somalia, Pozzi di El Meghet, Senni 651

(FI, holo.!).

A. volkii Suesseng. in Mitt. Bot. Staatssamml. Munchen 2:

40 (1954). Type as for A. spinosa Marloth & Engl.

Var. rostrata differs from the other varieties in

having pods with rostrate or strongly beaked apices,

although on occasional specimens the apices may be

acute or even rounded. The latter, however, are

usually accompanied by pods with rostrate apices so

that there is little difficulty in identifying the specimens.

Var. rostrata grows either as a shrub branching at

or near the base or a small tree up to 6 m high with

a flattened or rounded crown, and yellowish- or

greyish-brown i papery bark. The inflorescence

axes are pubescent.

Var. rostrata was initially thought to be confined

to Angola, South West Africa, Botswana, Rhodesia,

Mozambique, Transvaal, Swaziland and Natal, but

specimens with rostrate pod apices from Uganda,

Kenya and Somalia are also referable to this variety.

Some of the specimens from north-east Africa some-

times have slightly larger leaflets than those recorded

in southern Africa but, apart from this, there is little

difference between the specimens from the two areas

and the differences do not appear worthy of formal

taxonomic recognition.

The possibility exists that rostrata is not the earliest

epithet available at varietal rank for this taxon.

A. oxyosprion Chiov. var. oxyosprion is regarded as

a synonym of var. rostrata , but the identity of A.

oxyosprion var. pubescens Chiov. remains in doubt.

Unfortunately, Guidotti 21, the type of var. pubescens

has not been traced. From the description, var.

pubescens apparently differs from var. oxyosprion

solely in having a denser indumentum. If this is

indeed so, then there is a distinct possibility that var.

pubescens belongs to the same taxon as A. senega I

var. rostrata in which case the varietal epithet

pubescens would have to be adopted.

Chiovenda described several other taxa, which are

very closely related to or conspecific with A. senega /,

which is unfortunate as it has only served to compli-

cate an already confused situation. A. cufodontii

Chiov., in Miss. Biol. Borama Race. Bot. 4: 55,

fig. 5 (1939), from Ethiopia is conspecific with A.

senega / but, owing to the absence of pods and field

notes, it has not been possible to assign it to a

particular variety. The densely pubescent inflorescence

axes exclude it from var. leiorhachis.

Chiovenda based his description of A. Senegal var.

pseudoglaucophy/la, in Stefanini-Paoli Miss. Somal.:

72 (1916), on six syntypes from Somalia. As discussed

in Bothalia 11: 302 (1974) these syntypes belong to

three different taxa. As Chiovenda apparently had

no clear concept of his var. pseudoglaucophylla , and

as uncertainty over its circumscription remains, it

seems desirable to select one of the specimens as the

lectotype of the variety. I now select Paoli 83 from

Mogadiscio as the lectotype of var. pseudoglauco-

phylla. Paoli 83 is a fruiting specimen with glabrous

or subglabrous inflorescence axes and puberulous

pods up to 8, 3x2, 7 cm with rounded or subacute

apices. The leaf-rhachides and rhachillae are glabrous

or almost so and the leaflets are dark on the upper

surface and pale beneath. The decision to be taken

when more information is available is whether Paoli

83 represents a distinct variety within A. Senegal or

whether, as intimated in Bothalia l.c., var. pseudo-

glaucophylla should be assimilated into var. Senegal.

Several other flowering specimens with glabrous

inflorescence axes and discolorous leaflets from

Somalia in the Kew Herbarium, namely, Gillett 4396

from Duwi [ 10°5'N x 44 1 5'E], Drake-Brockman 51,

Thomson 27, may possibly belong to the same taxon

as Paoli 83 but this requires investigation.

In addition, there remain a few other specimens

from Somalia and Ethiopia that cannot be satis-

factorily placed. At present the status of these speci-

mens is not clear. It is not known whether they

represent distinct local races or whether they represent

the response to unusual or extreme habitats.

4. A. sp.

Represented in the Kew Herbarium by Ironside

Wood S/73/46 from Hudiso near Sheikh in Somalia.

This may well prove to be only a variant of A. Senegal

but, until more material is available, it seems prudent

to keep it separate. It differs from A. senega! in having

only 1 pinna pair per leaf although the occasional leaf

sometimes has 2 pinnae pairs. The 8-15 pairs of

leaflets per pinna are up to 4x1 mm. The prickles

are solitary, the inflorescence axes are very sparingly

pubescent, and the immature pods are acute apically.

The I pinna pair per leaf gives the specimen quite a

different facies to the material of A. senega!.

Gillett 4427 pro parte (one twig of Gillett 4427 is

referable to A. sp. near somalensis) from Duwi in

Somalia probably also belongs here.

5. Acacia hamulosa Benth. in Hook., Lond. J.

Bot. I: 509 (1842); in Trans. Linn. Soc. Lond. 30:

516 (1875) pro parte excl. ref. Mimosa asak et Acacia

asak; Brenan in Kew Bull. 8: 100 (1953); Ross in

Bothalia 11: 301 (1974). Type: Arabia, hills near

Gedda, S. Fischer 72 (K. holo.!).

Mimosa senegatensis Forsk., FI. Aegypt.-Arab. : 176 (1775)

nom. illegit., non M. senegalensis Houtt., Nat. Hist. 3: 614

(1774). Syntypes from Arabia, Forskal (C!).

Acacia asak sensu auct. mult., non (Forsk.) Willd.

A. paradoxa Chiov., FI. Somala 1: 165, 1. 1 7 fig. 2 (1929)

nom. illegit., non A. paradoxa DC., Cat. Hort. Monsp. : 74

(1813). Type: Somalia, Obbia, Ilbehla, Puccioni & Stefanini

533 (FI. holo.!).

458

THE ACACIA SENEGAL COMPLEX

Shrub to 2,75 m high; young branchlets whitish,

grey or grey-brown to purplish, sometimes as though

whitewashed over a purplish background, glabrous

or sparingly pubescent at first but becoming glabrous

later. Prickles in threes, the central one strongly

hooked downwards and the two laterals curved

upwards. Leaves glabrous: petiole 0,3-1, 3(2) cm

long; rhachis 0,6-1 ,5(2,5) cm long; pinnae 3-4 pairs;

rhachillae 0,3-1 cm long, with a recurved prickle

on the lower surface either at or just below the apex;

leaflets (3)5-8 pairs per pinna, 2-5x0,75-2 mm in

Africa (up to 10x3,25 mm in Arabia), glabrous.

Inflorescence axis up to 4 cm long, glabrous or

pubescent. Calyx 1,5-2, 5 mm long, glabrous or

sparingly pubescent. Corolla 2,5-3,75 mm long.

Pods greenish-yellow to brown, (3 , 5)4-8 , 5 X (2)3-5

cm, 1-4-seeded, valves thin and ± papery, venose,

glabrous.

Found in Ethiopia, Somalia, Kenya (Northern

Frontier Province) and Arabia. Appears to occupy

a diverse range of habitats.

The following is a selection of the specimens

examined :

Ethiopia. — Ogaden, Danot-Mersin road, Simmons S22 (K).

Somalia. — Eastern Al Madu Range at Baditir near Dana,

Bally 10913 (K); Einad Newbould 1010 (K); 8 km south of

Galkayu, Hemming 1407 (K); boundary pillar 93 [45° 9'E x

8°37'N], Gillett 4178 (K).

Kenya. — Northern Frontier Province, 10 km north of

Isiolo-Wajir road on the Garba Tula-Merti road, Ballv &

Smith B14697 (K).

Arabia. — 15 km north of Jidda on Medina road, Schwan 24

(K); behind Zamuk, Popov, Tillin & Gilliland 4211 (K); around

Jedda, Trott 32 (K); upper parts of Khusaf valley, Aden,

Waring 102 (K).

A. hamulosa is readily distinguished from all of

the other species in this complex by the presence

of a recurved prickle either at or near the apex on

the lower surface of each rhachilla. There is some-

times a recurved prickle terminating the rhachis in

some of the other species in the complex but never

the rhachillae.

6. Acacia cheilanthifolia Chiov., FI. Somala 1 :

168, t.xvii fig. 1, t. xliv fig. 2, t. xlv fig. 1 (1929); Ross

in Bothalia 11:299 (1974). Syntypes: Somalia,

Migiurtini, between Ereri Jelleho e Martisor Dinsai,

Puccioni & Stefanini 663 [727] (FI!); valle di Bacba,

Puccioni & Stefanini 762 [843] (FI!); Bacino del

Darror, El Uncud, Puccioni & Stefanini 1010 [1115]

(FI!); Obbia, Magghiole, Puccioni & Stefanini 479

[531] (FI!); Obbia, between Uarandi e Scillin-

Bilhelli, Puccioni & Stefanini 509 [562] (FI!).

Shrub or small tree; young branchlets reddish-

brown, sparingly to densely puberulous; older twigs

ashen to greyish-brown, often somewhat mottled,

the epidermis flaking off here and there to reveal a

yellowish-brown inner layer. Prickles in threes, the

central one curved slightly downwards and the two

laterals upwards or else all three prickles pointing

upwards. Leaves small: petiole up to 6 mm long,

with a rounded or elongate adaxial gland; rhachis

up to 1, 5(2, 4) cm long, with a fairly large discoid

gland at the junction of the top 1-2 pinna pairs,

sometimes with recurved prickles on the lower

surface; pinnae 3-6 pairs; rhachillae up to 7 mm long;

leaflets 3-4 (rarely 5) pairs per pinna, 1, 5-2(5) X

1-2,5 mm, oblong, elliptic, obovate-oblong or rotund-

ovate, glabrous or densely puberulous. Inflorescence

axis up to 4 cm long, glabrous to densely puberulous.

Calyx glabrous to densely puberulous. Pods yellowish-

brown, 3-4, 5x1-1, 1 cm, straightish, rounded or

mucronate apically, compressed, inconspicuously

venose, densely puberulous.

A. cheilanthifolia is endemic in Somalia.

Young branchlets, petioles, rhachides, rhachillae, leaflets,

inflorescence axes and calyces glabrous to sparingly

puberulous (a) var. cheilanthifolia

Young branchlets, petioles, rhachides, rhachillae, leaflets,

inflorescence axes and calyces densely puberulous

(b) var. Iwt el la

(a) var. cheilanthifolia.

Chiov., FI. Somala 1: 169 (1929).

Apart from the five syntypes only one other speci-

men, namely Popov GP/57/18 (BM) from El Gal,

has been seen.

(b) var. hirtella Chiov., FI. Somala 1 : 169 (1929).

Type: Somalia, Migiurtini, between Dhur and Hossa

Uein, Puccioni & Stefanini 695 [769] (FI holo.!).

Known only from the type collection.

A. cheilanthifolia is closely related to A. Senegal

and to A. hamulosa. It differs from A. Senegal in

having 3-4 (rarely 5) pairs of leaflets per pinna and

narrower pods, and from A. hamulosa in lacking a

recurved prickle at or near the apex on the lower

surface of each rhachilla, and in having narrow pods.

More material of both varieties of A. cheilanthifolia

is desired.

7. Acacia asak ( Forsk .) Willd. in L., Sp. PI. ed.

4, 4: 1077 (1806), non auct. al., quae = A. hamulosa

Benth. ; Brenan in Kew Bull. 8: 97 (1953). Type:

Arabia, Forskal (C, holo.!).

Mimosa asak Forsk., FI. Aegypt.-Arab. : 176 (1775). Type

as above.

Acacia glaucophylla Steud. ex A. Rich., Tent. FI. Abyss. 1:

243 (1847); Benth. in Trans. Linn. Soc. Lond. 30: 516 (1875);

Bak. f.. Leg. Trop. Afr. 3: 827 (1930). Syntypes: Ethiopia,

mountain sides near Takazze River below Djeladjeranne,

Schimper 725 (BM!, K!, OXF!); depression near Takazze

River, Schimper 1710 (BM!, K!).

A. triacantha Hochst. ex A. Rich., Tent. FI. Abyss. 1 : 244

(1847). Type: Ethiopia, Modat, Schimper 1746 (P, holo.!;

BM, FI, K, iso.!).

Shrub or slender tree to 10 m high; bark on young

stems yellow, papery and peeling, on older stems

dark grey and deeply fissured; young branchlets

dark grey- or reddish-brown to purplish, lenticellate,

glabrous or sparsely puberulous; internodes mostly

<2,5 cm long but occasionally up to 4,5 cm long.

Prickles solitary or in threes, the central one slightly

hooked downwards, the laterals spreading almost at

right angles to the stem or pointing slightly upwards.

Leaves glabrous throughout or sparingly pubescent:

petiole 0,4-2, 4(4) cm long, with a small slightly

raised gland <1 mm in diameter which is variable

in position; rhachis 0,9-5, 5 cm long; pinnae 3-6

pairs; rhachillae 0,8-4, 5(6, 8) cm long; leaflets

(3)7-20 pairs perpinna, 3-9x 1 ,25-3,75 mm, glaucous,

glabrous throughout or lower surface sparingly pubes-

cent, sometimes with a small basal tuft of hairs to one

side of the midrib on the lower surface, inconspicuous

marginal cilia sometimes present. Inflorescence axis

up to 1 1 cm long, glabrous or very sparingly pubescent.

Calyx 1,5-2 mm long, glabrous. Corolla 2,5-3 mm

long. Stamen-filaments up to 5 mm long. Pods brown

or reddish-brown to purplish, (2,8)3,8-12x0,9-

1,5(1, 8) cm, straight or almost so, compressed,

inconspicuously venose, glabrous.

Found in the Sudan, Ethiopia and in Arabia.

Occurs in rocky areas and on stream banks.

The following is a selection of the specimens

examined in the Kew Herbarium:

Sudan. — Kassala, Broun 1039; Karora Hills, Crowfoot s.n.;

between Suakin and Berber, Schweinfurth 89.

J. H. ROSS

459

Ethiopia. — 80 km west of Dire Dawa along road to Addis

Ababa, Burger 2531 ; below Filfil, Mooney 8079: 6,4 km north

of Barentu, Hemming 1015; Mont Zibo, near Saati, Schwein-

furth 537.

Arabia. — Dhofar plain. Bent 157; Yemen, Gebel Melhan,

Schweinfurth 610.

Although a few specimens of A. asak sometimes

show an approach to some specimens of A. senega! ,

the two species have quite a different facies. The

differences between the two species are difficult to

express concisely in words but A. asak may be

distinguished by its normally larger glaucous leaflets,

larger leaves with widely spaced pinnae, and usually

narrower glabrous pods.

8. Acacia oliveri Vatke in Oesterr. Bot. Zeit. 30:

274 (1880) excl. descr. leguminis; Ross in Bothalia 1 1 :

303 (1974). Type: Ethiopia, Danakil territory, HUde-

brandt 729c (BM, drawing!).

A. asak (Forsk.) Willd. var. unispinosa Fiori in Agric.

Colon. 5: 93, fig. 67/3 (1911). Type: Ethiopia, Samhar, Uakiro,

7 Feb. 1909, Fiori 135b (FI, lecto. !).

A. unispinosa (Fiori) Chiov., FI. Somala I: 169 (1929); FI.

Somala 2: 186, fig. 112 (1932). Type as for A. asak var. unispi-

nosa.

Shrub or small tree to 4 m high with a rounded

crown; young branchlets reddish-brown, glabrous

or sparingly pubescent, older stems greyish or

greyish-brown to purplish-grey; internodes 0,8-2, 5

cm long. Prickles solitary or in threes, recurved.

Leaves: petiole 2,5-10 mm long, sparingly to densely

pubescent; pinnae 1-2 pairs; rhachillae 0,5-1 ,5 (2,5)

cm long, glabrous to densely pubescent; leaflets 4-9

pairs per pinna, 3-5(9) x 1 ,25-1 ,75(3,75) mm, oblong

or obovate-oblong, glaucous, glabrous. Inflorescence

axis sparingly to densely puberulous throughout or

basally only. Calyx 1-1,5 mm long, glabrous or

very sparsely puberulous. Corolla 1 ,75-2,25 mm long.

Stamen-filaments up to 5 mm long. Pods olive-brown,

3,2-6xl-l,6 cm, ± straight, acute apically, 2-5-

seeded, compressed, relatively inconspicuously venose,

glabrous or very sparingly pubescent on margins

and near stipe.

A. oliveri is endemic in Ethiopia. Its ecological

preferences are not clear but it apparently favours

rocky areas.

The following is a selection of the specimens

examined :

Ethiopia. — 24 km south of Massawa, Bally 7044 (K); near

Afta, Hemming 1182 (BM); Airuri, Popov 1377 (BM); Mont

Zibo, near Saati, Schweinfurth 538 (K); Mabra Plain, Hemming

1263 (BM).

The 1 or 2 pinna pairs per leaf and 4-9 pairs of

glabrous leaflets per pinna enable A. oliveri to be

easily identified. It differs from A. somalensis in

having a greater number of leaflets per pinna, and

from A. sp. near somalensis in having glabrous leaflets.

Specimens of A. oliveri often have galled leaflets

and inflorescences.

9. Acacia somalensis Vatke in Oesterr. Bot. Zeit.

30: 274 (1880). Type: Somalia, near Meid, HUde-

brandt 1396 (BM, K, iso.!).

Shrub to 3 m high; young branchlets reddish or

reddish-brown, glabrous, older stems grey or grey-

brown, sometimes as though whitewashed over a

darker background. Prickles solitary (never in threes),

typically ± straight and pointing upwards, the tip

alone being slightly reflexed, or, recurved and hooked

downwards. Leaves small: petiole to 2 mm long;

pinna 1 pair; rhachillae to 4 mm long; leaflets 2

(very rarely 3) pairs per pinna, 3, 5-4, 5(5) x 1 ,75-

2(3) mm, obovate-oblong, glaucous, glabrous. Inflo-

rescence axis glabrous. Calyx 1,5-2,25 mm long,

glabrous. Corolla 2,5-3,25 mm long. Stamen-filaments

up to 5 mm long. Pods dark reddish-brown, 2,5-

3,5(5) x 1 ,5-1 ,8 cm, 1-3-seeded, compressed; valves

thin, venose, glabrous.

A. somalensis is endemic in Somalia. Found mostly

at low altitudes near the coast on gravel, rocky

stony plains or hillsides.

In addition to the isotype, the following specimens

have been examined:

Somalia. — Heis, a few kilometres along the coast from Meid,

Glover & Gilliland 711 (BM, K); Mait, Bally 11231 (K);near

Mait, Popov GPj57!7 (BM, K.); 57 km from Erigavo to Mait,

17 km from Mait, Hemming 2040 (K); coast east of Berbera,

Popov 1183 (K), Popov 1184 (BM).

A distinctive species which is usually easily recog-

nized by the solitary prickles, leaves with 1 pinna

pair, and the 2 (rarely 3) pairs of glabrous leaflets

per pinna.

10. Acacia sp. near somalensis.

Superficially very similar to A. somalensis with

which it agrees in habit, in the prickles being solitary,

± straight and ascending or recurved, and in the

leaves having only 1 pinna pair. It differs from

A. somalensis , however, in the following characters:

1. the young extremities are fairly densely pubes-

cent;

2. the petioles and leaf-rhachillae are clothed with

spreading hairs;

3. the rhachillae are 4-9 mm long;

4. there are 3-5 pairs of leaflets per pinna;

5. the leaflets are 3-5x1-2,25 mm, linear-oblong

to obovate-oblong, fairly densely clothed with

conspicuous whitish appressed hairs on the

lower surface;

6. the inflorescence axes are clothed with spreading

hairs basally at least.

This taxon is endemic in Somalia. It is represented

in the Kew Herbarium by the following specimens:

Somalia. — 112 km from Borama, Durkahamaya in Marahle

Hills, Glover & Gilliland 877; Afard, Gillett 4458; Duwi,

Gillett 4427 pro parte; near Bulhar, Drake-Brockman 669.

With the exception of the last specimen, the speci-

mens were all collected at considerably higher alti-

tudes than the specimens of A. somalensis.

The sheet of Gillett 4427 in the Kew Herbarium

consists of mixed gathering and it is not possible

to establish whether the collector’s notes refer to the

specimen of this taxon mounted in the upper left

hand corner of the sheet or to the other entity (A. sp.)

mounted on the sheet. The capsule above the col-

lector's label contains loose pods, but, once again,

it is not possible to establish with certainty to which

taxon they belong although it seems probable that

they belong to the other entity (A. sp.). It would be

reassuring to have confirmation that the loose pods

in the capsule mounted above the collector's label

of Gillett 4458 definitely belong to A. sp. near soma-

lensis.

In Glover & Gilliland 877 the bark on the trunk is

yellowish and flakes off to reveal a greenish-yellow

inner layer, and that of Gillett 4458 is described as

“yellow scaling off”. No details concerning the

bark of A. somalensis are available but there may

well be a difference between the bark of this taxon

and that of A. somalensis. Field observations are

required to establish whether any such difference

does exist.

The pubescent young branchlets, petioles, leaf-

rhachillae and inflorescence axes, and the more

numerous pairs of leaflets with densely appressed-

pubescent lower surfaces are sufficiently distinctive

460

THE ACACIA SENEGAL COMPLEX

to indicate that this taxon almost certainly requires

formal taxonomic recognition. However, as the

available material is rather inadequate, and as there

is an element of doubt whether the loose pods of the

two Gillett specimens actually belong, it is considered

premature to decide precisely what rank should be

accorded to this taxon. The decision on whether to

give this taxon infraspecific status within A. somalensis

or specific status must wait until more material,

particularly fruiting material, is available for exami-

nation.

A. oliveri is readily distinguished by its glabrous

leaflets from specimens of this taxon.

11. Acacia thomasii Harms in Bot. Jahrb. 51:

366 (1914); Brenan in FI. Trop. E. Afr. Legum.-

Mimos.: 95, fig. 14/20 (1969). Type: Kenya, Kitui

District, Ikutha, F. Thomas III 127 (B, holo.f, BM,

drawing!).

Straggling shrub with elongate branches or slender

virgately branched tree up to 7(12) m high; young

branchlets usually grey- or yellowish-brown, some-

times as though whitewashed over a darker back-

ground, densely pubescent; internodes 1,2-4 cm

long. Prickles in threes, the central one hooked

downwards and the laterals curved upwards or

sometimes nearly straight, or occasionally prickles

solitary. Leaves: petiole 0,2-1, 2 cm long, glandular;

rhachis 0-0,8 cm long, pubescent; pinnae 1-2(3)

pairs; rhachillae (1)1 ,5-4 cm long; leaflets 7-15 pairs

per pinna, 3-9 x 1,5-3 mm, obliquely oblong or

elliptic-oblong, rounded to acute or mucronulate

apically, glabrous or sometimes pubescent on lower

surface. Inflorescence axis sparingly pubescent below,

glabrous above, up to 9,5 cm long. Calyx 3-4,5 mm

long, glabrous. Corolla 6,5-7 mm long, glabrous.

Stamen-filaments 13-15 mm long. Pods yellowish-

brown or brown, 5-10x1,5-2,3 cm, straight or

nearly so, subacute to acuminate apically, coriaceous,

venose, puberulous. Seeds ± subcircular-lenticular,

10-12 mm in diam.

A. thomasii is confined to Kenya.

The following is a selection of specimens examined

in the Kew Herbarium:

Kenya. — Northern Frontier Province, 66 km east of Isiolo

near the end of the Nyambeni Range, Bally & Smith B14466;

Machakos district, mile 138 and 129 from Mombasa on the

main Nairobi road, near Kenani, Verdcourt 2390; Masai

district, Laitokitok, Seret River, Vesey-Fitzgerald 29; Teita

district, Voi-Mwatate road, Drummond & Hemsley 4281 .

The distinctive growth form and large flowers

distinguish A. thomasii from all of the other species

in this complex except A. ogadensis. It differs from

A. ogadensis in having more numerous pairs of

narrower leaflets per pinna, and a different range of

distribution.

12. Acacia ogadensis Chiov. in Ann. Bot., Roma

13: 393 (1915); FI. Somala 1: 167 (1929). Syntypes:

Somalia, Ogaden, Rohecchi Bricchetti 594 (FI!);

between Bardera and Marda, Paoli 811 (FI!), Paoli

812 (FI!).

Alhizia ogadensis (Chiov.) Bak.f. ex Chiov., FI. Somala 2:

182, fig. 110 (1932). Syntypes as above.

Slender virgately branched straggly tree up to 5 m

high, sometimes (fide Glover & Gilliland 1037) with

two tiers of foliage, one at ± 1 m high through which

the stem grows to produce a second tier at 2,75- 5 m;

young branchlets grey or reddish-brown to purplish,

sometimes as though whitewashed over a purplish

background. Prickles in threes, the central one

hooked downwards and the two laterals spreading

almost at right angles to the stem or sometimes also

slightly hooked downwards. Leaves: petiole up to

4 mm long, glandular; pinna 1 pair; rhachillae

0,8-2 cm long; leaflets (2)3-4 pairs per pinna,

5- 15 x (2, 5)3—9 mm, oblong, obovate-oblong or obo-

vate, glabrous or minutely appressed-pubescent on

lower surface. Inflorescence axis puberulous or gla-

brous, up to 5 cm long. Calyx 3-4 mm long, pubescent

or glabrous. Corolla 4-6 mm long. Stamen-filaments

up to 18 mm long, lilac. Pods olive- or yellowish-

brown, 6-7 x 1 , 3-1 , 9 cm, compressed, acute apically,

venose, densely puberulous.

A. ogadensis is endemic in Somalia, where it

apparently favours limestone outcrops.

A. ogadensis is represented in the Kew Herbarium

by the following specimens:

Somalia. — South border at Dalaliso, Glover & Gilliland

1037; 104 km south of Belet Wein, Hemming CFH 93; Galkayu

Plains, Bally 9606; 100,8 km north of Dusa Mareb, Bally 9574;

49,6 km north of Gardo, Hemming 1579.

All three syntypes are leafless flowering specimens

in which the inflorescence axes are puberulous and

the calyces fairly densely pubescent. The flowering

specimens in the Kew Herbarium cited above agree

with the syntypes in flower size and in overall facies

but differ in having glabrous inflorescence axes and

calyces. The taxonomic significance of the presence

or absence of the indumentum is not clear and

requires investigation.

The slender straggling virgate growth form of

A. ogadensis is reminiscent of A. thomasii and of

A. senega! var. leiorhachis. A. ogadensis differs from

both A. thomasii and A. Senegal var. leiorhachis ,

however, in having fewer pairs of broader leaflets

per pinna. The large flowers also help to distinguish

A. ogadensis from all of the other species in the

complex except A. thomasii. A. ogadensis appears to

be most closely related to A. thomasii.

Flower colour in A. ogadensis needs clarification:

it is not clear whether the stamen-filaments alone are

lilac or whether the corolla too is lilac.

13. Acacia condyloclada Chiov. in Ann. Bot.,

Roma 13: 391 (1915); Brenan in FI. Trop. E. Afr.

Legum.-Mimos. : 94, fig. 14/19 (1959). Type:

Ethiopia, Ogaden, between Sassaban and Carbaden,

Riva & Rnspoli 1079 (FI, holo.!).

Tree 2-1 1 m high with white or yellow peeling bark;

young branchlets reddish-brown, purplish or blackish,

puberulous; internodes 5,5-9 cm long, often ± en-

larged towards the apex. Prickles solitary or in threes,

slightly hooked or ± straight and spreading. Leaves:

petiole 1 ,3-2,8 cm long, with a large flattened ± dis-

coid gland 1,5-3 mm in diameter near the base;

rhachis 3-6,5 cm long, eglandular, puberulous;

pinnae 3-4 pairs; rhachillae 2-7,5 cm long; leaflets

6- 9 pairs per pinna, 9-20x3,5-9 mm, obliquely

oblong to slightly ovate or obovate, rounded apically,

puberulous above and below. Inflorescence axis

glabrous, up to 13 cm long. Calyx ±2,5-2,75 mm

long, glabrous, with a dark stripe running from the

base of the calyx to the apex of each lobe. Corolla

±3,5 mm long, glabrous. Pods brown or reddish-

brown, 6-10,5x1,8-2,1 cm, straight, rounded or

apiculate apically, venose, densely puberulous.

Found in Ethiopia, Somalia and Kenya (Northern

Frontier Province). Appears to favour limestone

outcrops.

The following specimens have been examined:

Somalia. — Bugshenleh, Peck Y102 (K).

Kenya. — Northern Frontier Province, Lag Ola, 45 km west

of Rarnu on Banessa road, Gillett 13279 (K).

J. H. ROSS

461

Burger 3396(K) and 3397(K), a sterile and a

flowering specimen respectively, from limestone slopes

west of Daletti above the Gobelli river valley in the

Harar Province of Ethiopia [42°7'Ex 8°32'N] are

almost certainly referable to A. condyloclada. Unfortu-

nately there is very little material of A. condyloclada

available for comparison, but the two above specimens

have the overall facies of this species. They agree in

having smooth white peeling bark, puberulous

purplish to blackish young branchlets with long

internodes, a large gland at the base of the petiole,

eglandular rhachides, 3-5 pinnae pairs, 6-12 pairs of

mostly puberulous large leaflets per pinna, glabrous

inflorescence axes, and sesssile flowers with glabrous

calyces and corollas. However, one twig of Burger

3397 is unusual in having two leaves with slightly

smaller leaflets (6x2 mm), although the other twig

has the typical larger leaflets (up to 10x4,5 mm).

Some of the leaflets are ± glabrous. In addition,

the flowers tend to be slightly smaller than in the

other flowering specimen examined, i.e. the calyx is

± 2 mm long as opposed to 2,5-2,75 mm. As in

A. condyloclada , in many of the flowers a dark stripe

is evident running up each calyx-lobe. These two

specimens clearly match A. condyloclada more closely

than any other species and the present circumscription

of the species could easily be amended to make

provision for the slight differences noted above.

More material is required.

The white peeling bark, long internodes, large

basal petiolar gland, and few pairs of large leaflets

per pinna distinguish A. condyloclada from the other

species in this complex.

14. Acacia caraniana Chiov., FI. Somala 1: 166,

1. 1 8 (1929). Type: Somalia, Migiurtini, Behen,

Puccioni & Stefanini 704 (FI, holo.).

Slender tree to 7 m high with pale grey or whitish

bark; young branchlets reddish-brown to purplish

or as though whitewashed over a reddish or purplish

background, glabrous; internodes 5-8,5 cm long.

Prickles solitary or in threes, central prickle not

strongly recurved, the laterals (when present) spread-

ing laterally and almost at right angles to the stem.

Leaves: petiole 2-6 mm long, with a large flattened

gland 1-2 mm in diameter at the point of attachment

of the lowest pinna pair; rhachis 1,5-4 cm long,

glabrous; pinnae 3 pairs; rhachillae 2-7,5 cm long;

leaflets 5-13 pairs per pinna, 5-13x1,25-4 mm,

oblong or obovate-oblong, glaucous, glabrous. Inflor-

escences on abbreviated axillary shoots, axes up to

12,5 cm long, glabrous. Flowers? (only remnants

present but corolla apparently <4 mm long). Pods

olive, 6-15,5 cm long (including a stipe 1,8-4 cm

long), 0,9-1 cm wide, straight or almost so, com-

pressed, margins slightly constricted between some

of the seeds, 3-7-seeded, finely venose, glabrous.

A. caraniana is endemic in Somalia, where it

apparently favours gypsum outcrops.

The following specimens have been examined:

Somalia. — Hills between Garoe and Galcaic, Hemming

CFH75 (K); 17 km SW of Garoe, Hemming 1376 (K); SE

Nogal area, Glover & Gilliland 1047 (BM, K).

A. caraniana is a distinctive species which is easily

recognized by the long internodes, large petiolar

gland at the point of attachment of the lowest pair of

pinna, the 3 pinnae pairs, and the narrow, long-

stipitate pods.

15. Acacia zizyphispina Chiov., FI. Somala 1:

167 (1929). Syntypes: Somalia, between Garass-

Hebla-Aden and Jesomma, Puccioni & Stefanini

152 (FI!); between Avorrei and Bulo-Burti, Puccioni

& Stefanini 171 (FI!)

A. impervia Gilliland in Kew Bull. 6: 139, t.3 (1951). Type:

Somalia, Ogaden, between Wardere and Walwal, Glover &

Gilliland 386 (K, holo.!; BM, FHO, iso.!).

Shrub to 2,75 m high, often with a short trunk

0,3-0, 6 m high and then branching to give a flattish

or rounded crown; young branchlets glabrous or

very sparingly puberulous, dark reddish-brown,

becoming dark grey-brown to purplish, sometimes as

though whitewashed over a purplish background.

Prickles in threes, the central one strongly hooked

downwards, the two laterals strongly curved up-

wards. Leaves: petiole 0,6-1, 8 cm long; rhachis

0,7-2, 5 cm long; pinnae 2-3 pairs; rhachillae 0,3-

1 ,3 cm long; leaflets 1 or 2 pairs per pinna, 8-19 x

5-13 mm, obovate, asymmetric basally, minutely

appressed-puberulous on both surfaces. Inflorescence

axis glabrous, up to 4 cm long. Calyx up to 2,5 mm

long, glabrous. Corolla up to 4 mm long. Stamen-

filaments up to 7 mm long. Pods yellowish-brown,

brown or reddish-brown to purplish, 4,5-9,3x

(2, 6)3, 2-4, 8 cm, rounded apically, compressed,

valves thin, venose, glabrous.

Found in Ethiopia and Somalia. Ecological pre-

ferences not clear.

The following is a selection of specimens examined

in the Kew Herbarium:

Ethiopia. — Ogaden, near Wardere, Popov 1124: 25,6 km

east of Wardere, Hemming 1463; 82 km east of Wardere,

Hemming 1514.

Somalia. — 27 km north of Giuale, Hemming CFH8I

28,8 km south of Ghelinsor, Hemming 1429.

A. zizyphispina is a distinctive species, which is

easily distinguished by the 1 or 2 pairs of large

leaflets per pinna, 2 or 3 pinnae pairs per leaf, and

the broad thin-valved pods.

16. Acacia ankokib Chiov., FI. Somala 2: 190,

fig. 116 (1932). Syntypes: Somalia, unlocalized,

Robecchi-Bricchetti 529 (FI!); Migiurtinia, Bender

Merajo, Guidotti 35 (whereabouts unknown).

Tree to 6 m high, sometimes several-stemmed from

the base; bark yellowish, papery, peeling off in large

thin pieces; very young branchlets greyish or reddish-

brown, fairly densely puberulous at first but becoming

glabrous, the greyish epidermis on older branchlets

peeling off to expose a papery yellowish inner layer.

Prickles present or often absent (see note below).

Leaves: petiole 0,6-1 ,9 cm long, with a small slightly

raised gland which is variable in position; rhachis

2-5 cm long; pinnae (2)3-6 pairs; rhachillae 1,2-3, 2

cm long; leaflets (4)6-9 pairs per pinna, 4-10 x

0,75-1,75(2,5) mm, linear-oblong or oblong, obtuse

apically, distinctly petiolulate, often bent basally so

that the apex of the leaflet tends to point upwards

towards the apex of the pinna, glaucous, minutely

puberulous on both surfaces. Inflorescence axis

minutely puberulous, up to 6 cm long. Calyx 1-1,5

mm long, glabrous or almost so. Corolla 1,5-2 mm

long. Stamen-filaments up to 4 mm long. Pods brown,

4, 5-7, 5x0, 9-1 cm, compressed, acute apically,

inconspicuously venose, minutely puberulous.

A. ankokib is endemic in Somalia. Ecological

preferences unknown.

The following specimens have been examined:

Somalia. — Mijertein, Everard 3 (BM); 64 km from Scuciuban

on El Gal road. Hemming 1771 (K); El Gal, Popov GP/57/23

(K).

The true nature of the armature in A. ankokib

is not clear and requires clarification. In Robecchi-

Bricchetti 529 the prickles are mostly paired with a

small, sharp or blunt outgrowth in the position

462

THE ACACIA SENEGAL COMPLEX

normally occupied by the third prickle, while occa-

sionally a well-developed third prickle is present but

situated some distance below the paired prickles. It

is not clear whether the outgrowths are reduced

prickles or whether they merely simulate prickles.

Guidotti 35, the other syntype, was said by Chiovenda

to be unarmed, and all of the specimens cited above

are unarmed. Until it has been established con-

clusively whether the prickles are basically paired or

in threes, it is uncertain whether A. ankokib is a

member of the A. senega! complex or whether it

should be excluded.

A. ankokib is readily distinguished by the yellowish

papery bark on the bole which peels off in large pieces,

the greyish epidermis on the older branchlets which

peels away to expose a papery yellowish inner layer,

the few pairs of distinctly petiolulate, glaucous,

minutely puberulous leaflets, and the narrow pods.

Bothalia 11, 4: 463-470 (1975)

The naturalized and cultivated exotic Acacia species in

South Africa

J. H. ROSS*

ABSTRACT

The first species of Acacia from Australia are thought to have been brought to the Cape in 1845. Some

of the factors which are believed to have contributed to the success of the exotic Acacia species are enumerated

briefly. A key to the identification of the naturalized Acacia species is provided, together with descriptions of

each species and an indication of their range of distribution in our area.

INTRODUCTION

Seeds of Acacia cyclops A. Cunn. ex G. Don and

A. sa/igna (Labill.) Wendl., the first Australian

species introduced into our area, are thought to have

been brought to the Cape in about 1845 (Roux in

S. Afr. J. Sci. 57: 99-102, 1961). These species, which

were introduced initially to stabilize the shifting sand

on the Cape Flats, proved highly successful for this

task and their achievement in reclaiming large areas

of drifting sand has been of lasting value. However,

their spread beyond the confines of the Flats and the

fact that they would in time invade, suppress and

displace the indigenous vegetation in many areas

was never envisaged by those who went to such great

lengths to propagate them. So successful are some of

these introduced species in the south-western Cape

that there is a grave concern over the future of many

of our indigenous species. Indeed, the stage has

now been reached when some form of biological

control of these introduced invading species is being

contemplated.

Some of the factors which are thought to account

for the success and spread of these introduced species

were discussed by Roux & Middlemiss in S. Afr.

J. Sci. 59: 286-294 (1963). They are briefly:

1. Fire assists the spread of the introduced Acacias

by exposing bare ground, raising soil pH,

destroying the seeds of some indigenous species

and inducing the germination of the seeds of

the Acacias.

2. The introduced species have a much more rapid

and sustained growth-rate than the indigenous

species.

3. A. cyclops and A. saligna, at least, form root-

nodules and are capable of fixing atmospheric

nitrogen.

To these may be added:

4. Schutte in S. Afr. J. Bot. 26: 45-49 (1960)

revealed that micronutrient deficiency symptoms

occur in nature in some species of the indigenous

Cape vegetation, and suggested that a factor

favouring the spread of the introduced species

may be their lower requirements for certain

micronutrients.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria.

5. Jones et al. in J.S. Afr. Sci. 59: 295-296 (1963)

found that substances which had a delaying

effect on the germination of seedlings of certain

indigenous species could be extracted from the

dried roots of A. cyclops, but the ecological

significance of this is not known.

6. The introduced species have escaped from their

natural pests and predators and therefore,

unlike the indigenous species, are free from pest

pressure.

None of the introduced Acacia species, except

perhaps A. farnesiana (L.) Willd., is likely to be

confused with any of the indigenous Acacia species.

The introduced species can be readily distinguished

as they either have the leaves modified to phyllodes

(entire, leaf-like flattened organs), or, if the leaves

are bipinnate, are unarmed (except for A. farnesiana).

Although A. farnesiana is armed with paired stipular

spines, it is easily distinguished from all of the

indigenous species by the following combination of

characters: absence of “ant-galls”, leaflets with the

lateral nerves raised and somewhat prominent

beneath, apical involucels, bright golden-yellow

flowers in non-paniculate heads, and dark brown to

blackish subterete or turgid pods. A further very

significant diagnostic feature is that the anthers lack,

even in bud, the small deciduous apical gland which

is present in all of the indigenous capitate-flowered

Acacias occurring in our area.

Examination of available herbarium material sug-

gests that the species dealt with in this account are

the only Acacia species which have become naturalized

in our area. Most herbaria contain very little material

of these introduced species and it appears that in

general collectors tend to ignore them and concentrate

almost solely on the indigenous vegetation. Most of

the species are known to be more widely distributed

than indicated in this account, but unfortunately

this knowledge is not substantiated by herbarium

specimens. For example, A. mearnsii De Wild, is

widespread throughout most of Natal and yet speci-

mens have been collected from only three degree

squares. This is a regrettable state of affairs and a

plea is made to collectors to pay more attention

to these exotic species and to collect specimens from

throughout the distributional ranges of these species.

26700—7

464

THE NATURALIZED AND CULTIVATED EXOTIC ACACIA SPECIES IN SOUTH AFRICA

Key to the naturalized and cultivated exotic Acacia species

Leaves bipinnate:

Plant armed with paired stipular spines 1. A. famesiana

Plant unarmed:

Leaflets large, mostly 2-5 cm long, 0,4-1 cm wide, lanceolate to linear-lanceolate, often somewhat

falcate; leaves large, 30-40 cm long 13. A. elata

Leaflets 1,5-15 mm long, <2 mm wide; leaves smaller than above:

Leaves with 1-4 pinnae pairs:

Petiole and rhachis together <2,5 cm long; pinnae crowded, glaucous; midrib ± central in the

leaflet or inconspicuous; inflorescence an axillary raceme or panicle, longer than the

leaves 5. A. bailey ana

Petiole and rhachis together >3 cm long; pinnae neither crowded nor glaucous; midrib almost

marginal throughout the length of the leaflet, pubescent; inflorescences axillary, solitary,

paired or fascicled 14. A. visite

Leaves with (5)8-26 pinnae pairs:

Leaflets 1,5-5, 5 mm long:

Leaf-rhachis with a gland at the junction of each pinna pair and usually also with additional

glands between the pinnae pairs; pods ± moniliform 2. A. mearnsii

Leaf-rhachis with a gland at the junction of all or most pairs of pinnae, but not between the

pinnae pairs; pods not or only slightly moniliform 3. A. dealbata

Leaflets 6-15 mm long:

Young Irranchlets prominently angled, sometimes with wing-like ridges 1-2 mm high; pinnae up

to 15 pairs per leaf ; leaflets 6-15 X 0,3-0,75 mm, midrib ± central, glabrous; inflorescence

an axillary raceme or panicle 4. A. deeurrens

Young branchlets not prominently angled; pinnae up to 7 pairs per leaf ; leaflets 6-9 x 0,8-1 ,25(2)

mm, midrib almost marginal throughout its length and pubescent; inflorescences axillary,

solitary, paired or fascicled 14. A. visite

Leaves apparently simple, modified to phyllodes by dilation of the petiole and rhachis:

Plant armed with stipular spines 6. A. armata

Plant unarmed:

Phyllodes each with one main longitudinal nerve:

Phyllodes < 4 cm long, 0,5-2 cm wide, seldom up to 4,5 cm long but then <0,5 mm wide and

linear-oblong:

Phyllodes 2-5 mm wide, linear-oblong, the margins typically densely ciliate 17. A. fimbriata

Phyllodes 6-20 mm wide, not linear-oblong:

Young branchlets and phyllodes densely grey-pubescent, especially when young; phyllodes ovate

to elliptic or elliptic-oblong, mostly 10-20 mm wide; pods 1,5-2 cm wide

7. A. podalyriifolia

Young branchlets and phyllodes glabrous; phyllodes obliquely obovate-Ianceolate to ovate-

triangular, 5-11 mm wide; pods 5-7 mm wide 15. A. cultriformis

Phyllodes 5-22 cm long:

Phyllodes 1,5-3 mm wide 18. A. adunca

Phyllodes >5 mm wide:

Phyllodes ± straight or sometimes slightly falcate, linear-lanceolate to linear-oblong or

oblanceolate, mostly 0,6-1, 4 cm wide:

Flower-heads >6,5 mm in diameter; peduncles 6-22 mm long; naturalized species

8. A. saligna

Flower-heads < 6 mm in diameter; peduncles 3-6 mm long; cultivated only. . 16. A. ret inodes

Phyllodes distinctly falcate, obovate-Ianceolate, mostly 1,3-3 cm wide 9. A. pycnantha

Phyllodes each with 2-7 or more longitudinal nerves:

Flowers in spikes:

Pods straight or slightly curved, glabrous; inflorescence axis glabrous or almost so; naturalized

species 10. A. longifolia

Pods variously coiled or twisted, pubescent; inflorescence axis pubescent; cultivated only

19. A. maidenii

Flowers in round heads:

Phyllodes <4 mm wide:

Phyllodes green; inflorescences axillary, solitary or paired, rarely fascicled; pods 3-3,5 mm

wide, sparingly to densely pubescent, margins not winged 20. A. viscidtda

Phyllodes greyish or glaucous; inflorescence a short axillary raceme; pods 8-18 mm wide, the

margins distinctly winged 21. A. pendula

Phyllodes >5 mm wide:

Phyllodes 0,6-2, 5 cm wide, green, longitudinal nerves conspicuous; pods falcate or variously

coiled or spirally twisted, margins not winged; naturalized species:

Flowers bright yellow; petals free; phyllodes usually ± straight, sometimes slightly falcate,

with anastomosing longitudinal venation between the main longitudinal nerves

11. A. cyelops

Flowers pale yellowish-white; petals united to above the middle; phyllodes usually ± falcate,

sometimes ± straight, with conspicuous reticulate venation between the main longi-

tudinal nerves 12. A. melanoxylon

Phyllodes 0, 5-0,9 cm wide, greyish or glaucous, longitudinal nerves not prominent; pods

straight or slightly curved, the margins distinctly winged; cultivated only... 21. A. pendula

J. H. ROSS

465

1. Acacia farnesiana ( L .) Willd. , Sp. PI. ed. 4, 4:

1083 (1806); Benth. in FI. Austral. 2: 419 (1864);

Oliv. in FI. Trop. Afr. 2: 346 (1871); Benth. in Trans.

Linn. Soc. Lond. 30: 502 (1875); Bak. f., Leg. Trop.

Afr. 3: 835 (1930); Brenan, Checklist Tang. Terr.:

334 (1949); Gilbert & Boutique in FI. Congo Belg. 3:

164 (1952); Torre in Consp. FI. Angol. 2: 278 (1956);

Keay in FI. W. Trop. Afr. ed. 2, 1 : 499 (1958); Brenan

in FI. Trop. E. Afr. Legum.-Mimos. : 111, fig. 16/38

(1959); in FI. Zamb. 3, 1: 111 (1970); Ross, FI.

Natal 193 (1973). Type: Aldinus, Exact. Descr. PI.

Romae Hort. Farnesiano 4 (1625) (lecto.!).

Mimosa farnesiana L., Sp. PI. 1 : 521 (1753). Type as above.

Shrub or small tree to 4 m high; young branchlets

grey to reddish-brown or purplish, epidermis not

obviously peeling off, with numerous somewhat

transversely elongated lenticels, glabrous or almost

so. Stipules spinescent, in pairs, usually short, up to

1,5 (3) cm long, straight, slender, never inflated;

“ant-galls” and other prickles absent. Leaves bipin-

nate: petiole sparingly to ± densely pubescent,

usually with a small gland; rhachis sparingly to ±

densely pubescent, often with a small gland below

the junction of the top pinna pair; pinnae 2-7 pairs;

leaflets 10-21 pairs, 2-7x0,75-1,75 mm, midrib and

lateral nerves visible and somewhat raised beneath,

glabrous throughout or with few inconspicuous mar-

ginal cilia. Inflorescences capitate, on axillary pe-

duncles, solitary or in pairs or three’s. Flowers bright

golden-yellow, sweetly scented; peduncles sparingly

to ± densely pubescent basally, sparingly glandular;

involucel apical. Calyx and corolla glabrous except

for the apices of the lobes. Pods dark brown to

blackish, 4-7, 5x0, 9-1 ,5 cm, straight or curved,

subterete and turgid, glabrous, indehiscent, finely

longitudinally striate. Seeds chestnut-brown, 7-8 x 5 , 5

mm, elliptic, thick, only slightly compressed; areole

6-7 x4 mm.

Probably a native of tropical America, doubtfully

so in Africa. Widely introduced in the tropics and

often becoming naturalized. Only planted or an

escape from cultivation in our area.

Transvaal. — 2528 (Pretoria): Wonderboom Poort, at the

footpath of stones laid across the Aapies River, Gerstner 5519.

2531 (Komatipoort): Komatipoort, Pole Evans sub PRE 18281.

Natal. — 2930 (Pietermaritzburg): Durban Botanic Gardens,

Ross 1714.

Cape. — 2824 (Kimberley): Kimberley, Wilman sub BOL

15701.

A. farnesiana is grown for ornament and for its

fragrant flowers which are used to make perfume.

The pods of A. farnesiana are very distinctive and

enable the species to be easily recognized. In the

absence of pods, it will be helpful to recall that no

other African Acacia has the following combination

of characters: absence of “ant-galls”, leaflets with

the lateral nerves raised and somewhat prominent

beneath, apical involucels, and bright golden-yellow

flowers in non-paniculate heads.

A further very significant distinguishing feature of

A. farnesiana is that the anthers lack, even in bud,

the small deciduous apical gland which is present in

all of the indigenous capitate-flowered acacias occur-

ring in our area.

2. Acacia mearnsii De Wild., PI. Bequaert.

3: 61 (1925); Brenan in FI. Trop. E. Afr. Legum.-

Mimos.: 95, fig. 15/21 (1959); Brenan & Melville

in Kew Bull. 14: 37 (1960); Tindale in Beadle, Evans

& Carolin, Handb. Vase. PI. Sydney Distr. & Blue

Mts. 231 (1962); Brenan in FI. Zamb. 3, 1: 111

(1970); Ross, FI. Natal 193 (1973). Type: Kenya,

near Thika, Mearns 1092 (BR, lecto.; BM!).

A. clecurrens sensu Bak. f.. Leg. Trop. Afr. 3: 853 (1930)

saltern pro parte, non Willd. sensu stricto. A. mollissima sensu

auct. mult., Benth. in Hook., Lond. J. Bot. 1: 385 (1842);

Burtt Davy, FI. Transv. 2: 345 (1932); Brenan, Check-list

Tang. Terr. 333 (1949); Salter in Adamson & Salter, FI. Cape

Penins. 454 (1950); F. White, For. FI. N. Rhod. 82 (1962),

non Willd.

Unarmed tree up to 15 m high with a conical or

rounded crown; bark grey-brown to blackish, smooth

or rough on very old trunks; young branchlets

angular; all parts (except flowers) ± densely pubescent

or puberulous, indumentum on young parts often

golden. Leaves bipinnate: petiole 1-2,5 cm long,

often with a gland above; rhachis usually 4-12 cm

long, with numerous raised glands all along its upper

surface both at and between the junctions of the

pinnae pairs; pinnae 8-21 pairs; leaflets 15-70 pairs,

1,5-4x0,5-0,75 mm, linear-oblong, appressed-

pubescent or glabrous beneath, margins usually with

cilia. Inflorescences capitate, in terminal panicles.

Flowers pale yellow, fragrant, on peduncles 2-6 mm

long. Calyx sparingly pubescent especially towards

the apices of the lobes. Corolla glabrous or almost so.

Pods (1 ,6) 3-10x0,5-0,8 cm, jointed, almost monili-

form, ± grey-puberulous, dehiscing longitudinally

along one margin only, straight or slightly curved.

Seeds black, ±5x3,5 mm, elliptic, compressed,

smooth; caruncle conspicuous; areole 3,5x2 mm.

Introduced from Australia and now widespread in

parts of the Transvaal, Swaziland, Natal and the

Cape Province.

Transvaal. — 2430 (Pilgrim's Rest): Belvedere 26N forest,

±29 km from Pilgrim's Rest, Davidson & Mogg 33515. 2528

(Pretoria): Waterkloof, Pretoria, Scblieben 10090. 2530 (Lyden-

burg): 26,4 km SSE. of Lydenburg, D. Morris 58.

Swaziland. — 2731 (Louwsburg): 3,2 km east of Goedgegun,

Ross 1767.

Natal. — 2929 (Underberg): Cathkin Peak area, Strey 7809.

2930 (Pietermaritzburg): Winterskloof, Ross 2129. 3030 (Port

Shepstone): Mtwalume, Wood 10589.

Cape. — 3318 (Cape Town): Ida's Valley, bottom of Hell's

Hoogte Pass, Stellenbosch, Thompson 837. 3319 (Worcester):

Bain's Kloof, While 5657. 3225 (Somerset East): Glen Avon

Falls area, P. T. van der Wait 190. 3326 (Grahamstown):

Grahamstown, Troughton 49.

A. mearnsii is the well-known Black Wattle, which

is economically important for the tannin content

of the bark. The wood is used for firewood and for

building.

By a strange mischance, A. mearnsii, the earliest

valid name for this Australian species, is based on

a specimen collected in Kenya and thought by De

Wildeman to be endemic there.

3. Acacia dealbata Link, Enum. Hort. Berol. 2:

445 (1822), non A. dealbata A. Cunn. (1825); Benth.

in FI. Austral. 2: 415 (1864); Trans. Linn. Soc.

Lond. 30: 497 (1875); Burtt Davy, FI. Transv. 2: 346

(1932); Brenan, Check-list Tang. Terr. 332 (1949);

in FI. Trop. E. Afr. Legum.-Mimos.: 50 (1959);

Salter in Adamson & Salter, FI. Cape Penins. 455

(1950); Tindale in Beadle, Evans & Carolin, Handb.

Vase. PI. Sydney Distr. & Blue Mts. 231 (1962);

Brenan in FI. Zamb. 3, 1: 112 (1970); Ross, FI.

Natal 193 (1973). Type a plant cultivated at Berlin.

Unarmed shrub or tree up to 15 m high with a

conical or rounded crown; bark grey-brown to

blackish, smooth or rough on very old trunks;

young branchlets usually densely short-pubescent,

rarely subglabrous, ± grey-pruinose, indumentum

grey, or sometimes yellowish at first and then grey.

Leaves bipinnate, often glaucous: petiole 0,5-2 cm

long, eglandular; rhachis 2,5-10 cm long, with a

raised gland on its upper surface at the junction of

each pair of pinnae, but without other glands in

466

THE NATURALIZED AND CULTIVATED EXOTIC ACACIA SPECIES IN SOUTH AFRICA

between the pinnae pairs as in A. mearnsii; pinnae

(5) 10-26 pairs; leaflets in 17-50 pairs, 2-5, 5x0, 4-0, 7

mm, linear-oblong, sparingly to ± densely pubescent

or glabrous beneath, margins with or without cilia.

Inflorescences capitate, panicled or racemose. Flowers

bright yellow, on densely pubescent peduncles up to

6 mm long. Calyx and corolla glabrous except for

apices of the lobes. Pods 3-8x0, 7-1, 3 cm, not or

only slightly moniliform, dehiscing longitudinally

along one margin only, straight or slightly curved.

Seeds brown to blackish-brown, 5-6 x 3-3, 5 mm,

elliptic, compressed, smooth; caruncle conspicuous;

areole 3,5-4x0,75-1,5 mm.

Introduced from Australia.

Transvaal. — 2526 (Zeerust); Swartruggens, Sutton 1031.

2528 (Pretoria): east of Pretoria, Kinges 1781. 2529 (Witbank):

Loskop Dam Reserve, Theron 1752. 2531 (Komatipoort):

Kruger National Park, Pretoriuskop-Seekoeigat, Van der Schijff

3177. 2628 (Johannesburg): Melville Koppies, Johannesburg,

MacNae 1161 (BOL). 2629 (Bethal): Ermelo, Burtt Davy 594.

Natal. — 2730 (Vryheid): near Grootspruit, Strey 8053.

2930 (Pietermaritzburg): Hilton Road, Ross 2105; farm Moun-

tain Glen, Dargle, Taat 1025.

Lesotho. — 2927 (Maseru): Roma, Ruch 16; Mamathe’s,

Jacot-Guillarmod 1426.

Cape. — 3318 (Cape Town): Cape Town, Gerstner 6147-

3326 (Grahamstown): 1820 Settlers Nature Reserve, Troughton

227.

A. dealbata, commonly known as the Silver Wattle,

is sometimes confused with A. mearnsii. It differs

from the latter in lacking the glands in between the

pinnae pairs along the upper surface of the leaf-

rhachis, in being more pruinose, and in having

wider usually less moniliform pods.

4. Acacia decurrens Willd.. Sp. PI. ed. 4, 4:

1072 (1806); Benth. in FI. Austral. 2: 414 (1864);

Trans. Linn. Soc. Lond. 30: 496 (1875); Burtt Davy,

FI. Transv. 2: 345 (1932); Tindale in Beadle, Evans

& Carolin, Flandb. Vase. PI. Sydney Distr. & Blue

Mts. 230 (1962). Ross, FI. Natal 193 (1973). Type

from Australia, unknown.

Mimosa decurrens Donn, Hort. Cant. 1: 114 (1796), nomen

nudum.

Unarmed tree up to 12 m high with a conical or

rounded crown; young branchlets prominently angled,

sometimes with wing-like ridges 1-2 mm high,

glabrous or the very young shoots slightly tomentose-

pubescent. Leaves bipinnate, green, decurrent: petiole

angular, 1,5-2, 5 cm long, often eglandular; rhachis

3-10 cm long, with a raised gland just below the

junction of each pinna pair; pinnae (5) 8-15 pairs;

leaflets 15-35 pairs, 6-12x0,3-0,75 mm, linear,

usually glabrous throughout. Inflorescences capitate,

panicled or racemose. Flowers bright golden-yellow,

on peduncles 2-5 mm long. Calyx sparingly pubescent

on apices of lobes. Corolla glabrous or almost so.

Pods brown or dark brown, 3,5-10x0,4-0,7 cm,

not or only slightly moniliform, dehiscing longitu-

dinally along one margin only, straight or slightly

curved. Seeds brown to blackish-brown, ±5 x 3 , 5 mm,

elliptic, compressed, smooth; caruncle conspicuous;

areole ±3,5x2 mm.

Introduced from Australia.

Natal. — 2930 (Pietermaritzburg): Winterskloof, Ross 2130.

A. decurrens , commonly known as the Green Wattle,

is readily distinguished by its long narrow leaflets

from all of the other introduced Acacia species with

bipinnate leaves in our area.

A. decurrens is usually attributed to “Wendl.,

Willd.” with Mimosa decurrens Wendl., Bot. Beob. :

57 (1798), being taken as the basionym. However,

Willdenow cited only Mimosa decurrens Donn,

Hort. Cant. 1: 114 (1796), which is a nomen nudum.

As he provided no reference to Wendland, either

direct or indirect, Willdenow’s binomial must be

treated as a new name.

5. Acacia baileyana F. Muell. in Trans. & Proc.

Roy. Soc. Victoria 24: 168 (1887); Brenan in FI.

Trop. E. Afr. Legum.-Mimos. : 50 (1959); Tindale

in Beadle, Evans & Carolin, Handb. Vase. PI. Sydney

Distr. & Blue Mts. 231 (1962); Brenan in FI. Zamb.

3, 1: 112 (1970). Type from Australia.

Unarmed shrub or tree up to 5 m high; young

branchlets subglabrous to sparingly pubescent. Leaves

bipinnate, glaucous: petiole very short, 2-8 mm long;

rhachis 0-1,2 cm long, with a gland at the junction

of each or only the top few pinnae pairs; pinnae

(1) 2-4 pairs, crowded; leaflets 12-20 pairs, 3-7 x

0,8-1, 5 mm, linear-oblong, often slightly falcate,

glabrous throughout or with few marginal cilia.

Inflorescences capitate, in axillary racemes or panicles

longer than the leaves. Flowers bright yellow, on

peduncles 2-5 mm long. Calyx and corolla glabrous

or almost so. Pods brown, 4-10x0,8-1,4 cm,

straight or slightly curved, margins entire or only

slightly and irregularly constricted between some of

the seeds, dehiscing longitudinally along one margin.

Seeds blackish, ±6x3 mm, smooth; caruncle con-

spicuous; areole ±5x2 mm.

Introduced from Australia.

Transvaal. — 2528 (Pretoria): Prince’s Park, Rep ton IB.

2626 (Klerksdorp): near Rooijantjiesfontein, Kinges 1475.

Natal. — (Pietermaritzburg): Manderston, Ross 2103.

Cape. — 3125 (Steynsburg) : Grootfontein, Theron 612. 3326

(Grahamstown): Grahamstown, Troughton 65.

6. Acacia armata R. Br. in Ait.f., Hort. Kew,

ed. 2, 5: 463 (? Dec. 1813); DC., Prodr. 2: 449 (1825);

Benth. in FI. Austral. 2: 347 (1864); Trans. Linn.

Soc. Lond. 30: 461 (1875); Salter in Adamson &

Salter, FI. Cape Penins. 453 (1950); Beadle, Evans &

Carolin, Handb. Vase. PI. Sydney Distr. & Blue Mts.

224 (1962). Type: South Australia, Bay IX Memory

Cove, R. Brown (BM, holo.!).

Shrub up to 3,5 m high; young branchlets reddish-

brown or brown, angular-striate, usually hirsute-

pubescent, seldom glabrous. Stipules spinescent, in

pairs, slender, divaricate, up to 1 cm long. Leaves

phyllodic, apparently simple, 0, 5-1(1 , 5) x 0-2-0, 6

cm, obliquely-ovate to oblong or narrowly lanceolate,

undulate, with a single nearly centric midrib, apex

obtuse or distinctly mucronate, glabrous throughout

or sometimes with hairs on the margins and on the

midrib. Inflorescences capitate, on axillary peduncles

which are about as long as the phyllodes. Flowers

bright yellow. Calyx lobed but not separating into

sepals, ± half as long as the corolla. Petals narrow,

glabrous. Pods straight or ± falcate, 2, 5-6 x 0,2-0, 6

cm, dehiscent, villous, rarely glabrous or hispid.

Seeds dark brownish-black, ±7x2,5 mm, smooth;

caruncle conspicuous.

Introduced from Australia.

Cape. — 3318 (Cape Town): Cape Peninsula, Rhodes Estate,

Salter 7619 (BOL); above Rhodes Memorial and Groote

Schuur Hospital, Gerstner 6141.

It is quite probable that A. paradoxa DC., Cat.

Hort. Monsp. : 74 (Feb.-Mar. 1813), is an earlier

name for this species. However, until this has been

definitely established, the name A. armata is retained.

7. Acacia podalyriifolia A. Cunn. ex G. Don,

Gen. Syst. 2: 405 (1832); Benth. in FI. Austral. 2:

374 (1864); Trans. Linn. Soc. Lond. 30: 474 (1875);

Brenan, Check-list Tang. Terr. 332 (1949); in FI.

Trop. E. Afr. Legum.-Mimos.: 51 (1959); Beadle,

J. H. ROSS

467

Evans & Carolin, Handb. Vase. FI. Sydney Distr. &

Blue Mts. 225 (1962); Brenan in FI. Zamb. 3,1:

113 (1970); Ross, FI. Natal 193 (1973). Type:

Australia, Queensland, Birnam Range, Brisbane

River, A. Cunningham 157/1828 (K, holo.!).

Unarmed shrub or small tree up to 6 m high;

young branchlets densely grey-pubescent. Leaves

phyllodic, apparently simple, glaucous, mostly 1,5-

4x1-2 cm, ovate to elliptic or elliptic-oblong, often

oblique, with a single midrib and finely but distinctly

penninerved, sparingly to densely pubescent, with 1

or 2 marginal glands. Inflorescences capitate, in

axillary racemes which are usually longer than the

phyllodes, mostly terminal. Flowers bright yellow,

on pubescent peduncles up to 7 mm long. Calyx

less than half as long as the corolla, pubescent

apically. Petals ± free, hirsute. Pods brown, glabrous

or pubescent, 4-8, 5 x 1,5-2 cm, straight or almost

so, flattened, margins often ± undulate, dehiscing

longitudinally. Seeds dark brownish-black, 6-7 x

±3,5 mm, smooth, compressed; caruncle conspicu-

ous; areole 3, 5-4 x ±1,5 mm.

Introduced from Australia.

Transvaal. — 2528 (Pretoria): Riviera, Pretoria, Schlieben

10083.

Natal. — 2930 (Pietermaritzburg): slopes below World's

View, Ross 2104.

Cape. — 3318 (Cape Town): Stellenbosch, Louw 5. 3326

(Grahamstown) : Grahamstown, Troughton 44.

8. Acacia saligna ( Labill .) Wendl. , Comm. Acac.

26 (1820); Benth. in FI. Austral. 2: 364 (1864); Trans.

Linn. Soc. Lond. 30: 469 (1875); Maslin in Nuytsia

1: 334 (1974). Type from Western Australia,

Labillardiere (FI, lecto.).

Mimosa saligna Labill., PI. Nov. Holl. 2: 86, t. 235

(1806). Type as above.

Acacia cyanophylla Lindl., Bot. Reg. 25: Misc. 45

(1839); Benth. in FI. Austr. 2: 364 (1864); in Trans.

Linn. Soc. Lond. 30: 469 (1875); Salter in Adamson &

Salter, FI. Cape Penins. 454 (1950); F. White, For.

FI. N. Rhod. 82 (1962); Roux & Middlemiss in S.

Afr. J. Sci. 59: 286 (1963); Henderson & Anderson

in Mem. Bot. Surv. S. Afr. 37: 170, fig. 84a, b, c

(1966); Brenan in FI. Zamb. 3, 1: 112 (1970); Ross

FI. Natal 193 (1973). Type from Australia.

Unarmed shrub or tree up to 10 m high; young

branchlets slightly angular, glabrous. Leaves phyllodic,

apparently simple, glabrous, mostly 8-22x0,5-1,4

cm (the lower ones sometimes much longer and 4 cm

or more wide), usually narrow, linear-lanceolate to

linear-oblong or oblanceolate, straight or slightly

falcate, with a basal gland, much narrowed basally,

with a single midrib and finely but distinctly penni-

nerved, sometimes glaucous (on young plants and

coppice shoots bipinnate leaves are sometimes pro-

duced at the apex of the phyllode). Inflorescences

globose, in short axillary racemes. Flowers bright

yellow, on peduncles up to 2,2 cm long. Calyx

slightly pubescent apically. Corolla glabrous, Pods

brown, 5,5-15x0,5-0,6 cm, straight or slightly

falcate, flattened, margins slightly constricted between

some of the seeds, dehiscing longitudinally. Seeds

dark brown, 5-7x2,75-3,5 mm, smooth, com-

pressed; caruncle conspicuous; areole 3, 5-5 x± 1.5

mm.

Introduced into the Cape Province from Australia

and now fairly widespread from the Cape Peninsula

to the eastern Cape; also introduced into Natal

more recently.

S.W.A. — 2615 (Luderitz): Luderitz, Kinges 2736.

Transvaal. — 2528 (Pretoria): Sunnyside, Repton 1861.

Natal. — 2930 (Pietermaritzburg): Botha's Hill, Ross 2132.

2931 (Stanger): Virginia Airport, Ross 2139.

Cape. — 3318 (Cape Town): Ida’s Valley, bottom of Hell’s

Hoogte Pass, Stellenbosch, Thompson 836. 3325 (Port Eliza-

beth): Port Elizabeth, Begg s.n. (GRA). 3326 (Grahamstown):

road from Port Elizabeth to Grahamstown, Wells 2603. 3418

(Simonstown): Tokai, Burn Davy sub FHO 20021 (K). 3422

(Mossel Bay): Sedgefield, farm Karawater, bank of Karatara

River, Ross 2410.

A. saligna, commonly called the “Port Jackson

Wilow” on account of its pendulous branches and

phyllodes, was introduced on the Cape Flats in the

1870’s in an attempt to stabilize the shifting dune

sands. It proved highly successful for this purpose

and soon started spreading by natural means. A.

saligna is now found far beyond the area of the

Cape Flats and has become a serious menace in many

parts of the Cape Peninsula and on the mainland

by invading and displacing the indigenous vegetation.

In many areas A. saligna occurs in dense stands.

A. saligna coppices when cut down near the ground.

The wood is relatively soft and the branches are

brittle.

9. Acacia pycnantha Benth. in Hook., Lond. J.

Bot. 1 : 351 (1842); FI. Austral. 2: 365 (1864); Trans.

Linn. Soc. Lond. 30: 469 (1875); Salter in Adamson

& Salter, FI. Cape Penins. 455 (1950); Beadle, Evans

& Carolin, Handb. Vase. PI. Sydney Distr. & Blue

Mts. 226 (1962). Type from Australia.

Unarmed shrub or tree up to 10 m high; young

branchlets terete or almost so, glabrous. Leaves

phyllodic, apparently simple, glabrous, 10-20 x (1)

1,5-3 cm, obovate-lanceolate, distinctly falcate,

mostly obtuse apically, narrowed basally, with a

single midrib and finely but distinctly penninerved,

margin nerve-like, with a fairly large marginal gland

near the base (on young plants and coppice shoots

bipinnate leaves are sometimes produced at the apex

of the phyllode). Inflorescences globose, in axillary

racemes or panicles. Flowers bright yellow, on stout

peduncles up to 7 mm long. Calyx about 2/3 as long

as the corolla, pubescent apically. Corolla ± glabrous.

Pods brown, 6-12x0,4-0,7 cm straight or slightly,

curved, flattened, margins slightly constricted between

some of the seeds, dehiscing longitudinally. Seeds

dark brownish-black, 5-7x2,75-3,5 mm, smooth,

compressed; caruncle conspicuous.

Introduced from Australia.

Cape. — 3318 (Cape Town): Pinelands, Salter 8767 ; Cape

Town University, Leighton sub BOL 25537. 3418 (Simonstown):

Somerset West, Parker 3517 (K). 3420 (Bredasdorp): Potteberg,

Van Niekerk sub BOL 23359.

10. Acacia longifolia (Andr.) Willd. in L., Sp. PI.

ed. 4, 4: 1052 (1806), non A. longifolia Paxt. (1846);

Benth. in FI. Austral. 2: 397 (1864); Trans. Linn.

Soc. Lond. 30: 487 (1875); Salter in Adamson &

Salter, FI. Cape Penins. 454 (1950); Beadle, Evans &

Carolin, Handb. Vase. PI. Sydney Distr. & Blue

Mts. 228 (1962); Henderson & Anderson in Mem.

Bot. Surv. S. Afr. 37; 170, fig. 84 d, e, f (1966);

Ross, FI. Natal: 193 (1973). Type from Australia.

Mimosa longifolia Andr., Bot. Rep. t. 207 (1802). Type as

above.

Unarmed shrub or tree to 8 m high; young branch-

lets angular, glabrous or the young shoots minutely

pubescent. Leaves phyllodic, apparently simple, gla-

brous, 6-18x0,7-2 cm, linear-lanceolate or narrowly

oblong to oblanceolate, straight or almost so, mucro-

nate apically, sometimes obliquely so, narrowed

basally, with 2-5 prominent longitudinal nerves and

faintly or conspicuously anastomosing almost longi-

tudinal veins between the nerves. Inflorescences

468

THE NATURALIZED AND CULTIVATED EXOTIC ACACIA SPECIES IN SOUTH AFRICA

spicate, axillary, solitary or in pairs; spikes up to

4 cm long. Flowers bright yellow, sessile. Calyx very

short. Corolla glabrous. Pods brown, 7-14x0,4-0,6

cm, cylindrical, straight or slightly curved, margins

constricted between the seeds, dehiscing longitudinally

along both margins, valves longitudinally wrinkled

or striate, acuminate apically, glabrous. Seeds dark

brownish-black, 4-7 x ±2,5 mm, more or less

oblong, smooth, compressed; areole ±3,5x 1 ,5 mm;

funicle not much folded, thickened almost from the

base into a small ± cupular aril enclosing the apex

of the seed.

Introduced from Australia.

Transvaal. — 2627 (Potchefstroom): Randfontein, Barnard

sub PRE 32122. 2628 (Johannesburg): Johannesburg, Moss

5258 (BM).

Natal. — 2930 (Pietermaritzburg): Town Bush Valley, 1 ,6 km

west of Cascade Falls, Ross 1281 (NU); Hilton Road, Ross 2106.

Cape. — 3318 (Cape Town): Rondebosch, lower slopes of

Devil's Peak behind University, White 5002. 3319 (Worcester):

Franschhoek, Van der Merwe 1209. 3325 (Port Elizabeth):

24 km up Elands River road, Acocks 21275. 3326 (Grahams-

town): road from Port Elizabeth to Grahamstown, Wells 2602.

3418 (Simonstown): near Wynberg, Schlechter 1061 (GRA).

3419 (Caledon): Kogelberg Reserve, Paardeberg, Grobler 17140.

3422 (Mossel Bay): Mossel Bay, Hutchinson s.n. (K).

A. longifolia is commonly known as the Golden

Wattle. Like several of the other introduced Aus-

tralian species, A. longifolia is also invading and

displacing the indigenous vegetation in some areas.

A. longifolia is a variable species. Although some

of the extremes look very different, they are con-

nected by an almost continuous range of inter-

mediates and consequently cannot be separated satis-

factorily. Bentham l.c. 397 (1864) enumerated six

forms of A. longifolia.

Beadle, Evans & Carolin l.c. 228 recognized two

varieties, namely, var. longifolia and var. sophorae

(Labill.) F. Muell. ex Benth. Var. sophorae has mostly

obovate-oblong, oblanceolate or oblong-elliptic phyl-

lodes 1 ,2-3,6 cm wide and 5-12 cm long, in contrast

to the linear or linear-lanceolate phyllodes 0,3-1 cm

wide and 7,5-13 cm long of var. longifolia. Although

there is no distinct morphological discontinuity be

tween the two, in Australia var. sophorae has some-

what different ecological preferences and tends to

occur as a low plant along the coastal sand dunes,

while var. longifolia grows into a larger plant. Speci-

mens from our area are often difficult to place in one

variety or the other with certainty.

11. Acacia cyclops A. Cunn. ex G. Don, Gen.

Syst. 2: 404 (1832); Benth. in FI. Austral. 2: 388

(1864); Trans. Linn. Soc. Lond. 30: 481 (1875);

Salter in Adamson & Salter, FI. Cape Penins. 454

(1950); Roux in S. Afr. J. Sci. 57: 99 (1961); Roux &

Middlemiss in S. Afr. J. Sci. 59: 286 (1963); Middle-

miss in S. Afr. J. Sci. 59: 419 (1963); Henderson &

Anderson in Mem. Bot. Surv. S. Afr. 37: 172, fig. 85

(1966). Syntypes: Western Australia, King George’s

Sound, A. Cunningham 104/1818 (K!), 328/1821 (K!).

Unarmed shrub or small tree up to 6 m high;

young branchlets usually angular and glabrous.

Leaves phyllodic, apparently simple, glabrous, 3-9 x

0,6-1, 5 cm, narrowly-oblong, usually ± straight,

sometimes slightly falcate, obliquely mucronate api-

cally, narrowed basally, with 3-5 prominent longi-

tudinal nerves and anastomosing almost longitudinal

veins. Inflorescences globose, solitary or two or three in

short axillary racemes. Flowers bright yellow, on

peduncles up to 7 mm long. Calyx pubescent apically,

more than half as long as the corolla. Petals free.

Pods brown, 5-15x0,8-1,3 cm, oblong, falcate or

variously coiled or spirally twisted, flattened, margins

not constricted between the seeds, dehiscing longi-

tudinally along both margins. Seeds dark brown,

5-7 x 3-4 mm, smooth, compressed; areole ±4x2

mm; funicle thickened, bright red or orange, en-

circling the seed in a double fold.

Introduced into the Cape Province from Australia

and now widespread in coastal areas from Lambert’s

Bay in the north-west to Kidd’s Beach in the north-

east.

S.W.A. — 2615 (Luderitz): Luderitz, Kinges 2732.

Cape. — 3318 (Cape Town): Hell’s Hoogte, Stellenbosch,

Taylor 7298. 3325 (Port Elizabeth): Port Elizabeth, Theron

1142. 3326 (Grahamstown): Kowie River, Wells 2580. 3418

(Simonstown): Cape Peninsula, Rodin 3287A. 3419 (Caledon):

near Caledon, Gilliland A 62 (BM). 3422 (Mossel Bay): Sedge-

field, farm Karawater, bank of Karatara River, Ross 2408.

3423 (Knysna): bank of Lagoon, road to Knysna Heads,

Bos 935.

Like A. saligna , A. cyclops was introduced on the

Cape Flats in the 1870’s in an attempt to stabilize

the shifting dune sands. It proved highly successful

for this purpose and soon started spreading by

natural means. A. cyclops is now found far beyond

the area of the Cape Flats and has become a serious

menace in many parts of the Cape Peninsula and on

the mainland by invading and displacing the indi-

genous vegetation. In many areas A. cyclops occurs

in dense almost impenetrable stands.

Unlike A. saligna, A. cyclops does not usually

coppice when cut down. The wood of A. cyclops

provides a useful firewood.

A. cyclops is commonly known as “Rooikrans”

on account of the bright red funicle which encircles

the seed. The pods usually remain attached to the

plant long after the ripe seed have been shed.

A number of species of birds feed on the conspi-

cuous funicles and assist in the distribution of A.

cyclops (see Middlemiss in S. Afr. J. Sci. 59: 419,

1963).

12. Acacia melanoxylon R. Br. in Ait. f., Hort.

Kew ed. 2, 5: 462 (1813); Benth. in FI. Austral. 2:

388 (1864); Trans. Linn. Soc. Lond. 30: 481 (1875);

J. Phillips in Mem. Bot. Surv. S. Afr. 14: 291 (1931);

Salter, FI. Cape Penins. 454 (1950); Beadle, Evans &

Carolin, Handb. Vase. PI. Sydney Distr. & Blue Mts.

227 (1962); Ross, FI. Natal: 193 (1973). Type:

Tasmania, Port Dalrymple, R. Brown (BM, holo.!).

Unarmed tree up to 20 m high; young branchlets

angular, glabrous or the young shoots minutely

pubescent. Leaves phyllodic, apparently simple, gla-

brous, mostly 6-12x0,6-1,2 (2,5) cm, linear-

lanceolate to oblanceolate or narrowly obovate,

straight to falcate, narrowed basally, with 3-7 promi-

nent longitudinal nerves and a conspicuous reticulate

venation between the longitudinal nerves (on young

plants bipinnate leaves are sometimes produced at

the apex of the phyllode). Inflorescences globose,

solitary or in short axillary racemes. Flowers pale

yellowish-white, on peduncles up to 6 mm long.

Calyx more than half as long as the corolla. Corolla

glabrous. Pods brown, 5-15x0,6-0,8 cm, oblong,

falcate or variously coiled or spirally twisted, flat-

tened, margins thickened, not constricted between

the seeds, dehiscing longitudinally along both margins.

Seeds dark brownish-black, 4-5 x ±2,5 mm, smooth,

compressed; areole ±3x1 mm; funicle very long,

thickened, almost encircling the seed in a double fold.

Introduced from Australia.

Transvaal.- 2528 (Pretoria): Wonderboom Reserve, Repton

1871. 2627 (Potchefstroom): Krugersdorp, Webster sub PRE

32118. 2628 (Johannesburg): around Johannesburg, Moss

7082 (BM).

J. H. ROSS

469

Swaziland. — 2631 (Mbabane): 1,6 km from Hlatikulu on

Sitobela road, Ross 1759.

Natal. — 2730 (Vryheid): Donkerhoek, Devenish 1020.

2929 (Underberg): farm Vergelegen, Umkomaas River near

Lesotho border, Rissik s.n. 2930 (Pietermaritzburg): slope

below World's View, Ross 2128.

Lesotho. — 2927 (Maseru): Masoeling, Jacot-Guillarmod

2605.

Cape. — 3219 (Wuppertal): Cedar Mts., Algeria forest reserve.

Bos 516. 3318 (Cape Town): Rondebosch, near University of

Cape Town, White 5066. 3326 (Grahamstown): Grahamstown,

Roux sub PRE 32121. 3422 (Mossel Bay): Sedgefield, farm

Karawater, banks of Karatara River, Ross 2409.

A. melanoxylon , the well-known Blackwood, yields

a good timber which is used in the manufacture of

furniture. Like several of the other introduced

Australian species, A. melanoxylon is also invading

and displacing the indigenous vegetation in some

areas.

In addition to the species dealt with in some detail

above, several species are cultivated in our area.

At present, however, there is no evidence to suggest

that any of them have become naturalized. The

species cultivated are: —

13. Acacia elata A. Cunn. ex Benth. in Hook,,

Lond. J. Bot. 1: 383 (1842), non A. elata R. Grah.;

Benth. in FI. Austral. 2: 413 (1864); in Trans. Linn.

Soc. Lond. 30: 495 (1875); Summerh. in Bot. Mag.

154: t.9214 (1930); Brenan in FI. Trop. E. Afr.

Legum.-Mimos. : 50 (1959); in FI. Zamb. 3, 1: 111

(1970); Tindale in Beadle, Evans & Carolin, FI.

Sydney Region: 272 (1972). Type: Australia, New

South Wales, shaded ravines, Cunningham (K, holo.!).

Unarmed tree. Leaves bipinnate, large, 30-40 cm

long; pinnae 3-5 pairs; leaflets 8-15 pairs per pinna,

mostly 2-5x0, 4-1 cm, lanceolate to linear-lanceolate,

often somewhat falcate, usually finely pubescent at

least on the lower surface. Flowers pale yellow, in

round heads, arranged in axillary racemes or panicles.

Pods ±9-15x0,9-1,3 cm, linear-oblong, straight

or curved, the margins irregularly constricted between

the seeds, compressed, dehiscing along both margins.

A. elata is easily distinguished from all of the other

species with bipinnate leaves by its large leaflets.

Recorded from Krugersdorp in the Transvaal,

Gerstner 6671, but much more widely cultivated.

The name A. terminalis (Salisb.) Macbride, in Contr.

Gray Herb. 59: 7 (1919), has been applied incorrectly

to this species.

14. Acacia visite Griseb. in Abh. K. Ges. Wiss.

Gottingen 19: 135 (1874). Type from Argentina.

Unarmed tree. Leaves bipinnate; pinnae 2-7 pairs;

leaflets 24-38 pairs per pinna, 6-9x0,8-1,25(2) mm,

linear or linear-oblong, acute apically, midrib almost

marginal throughout its length and usually pubescent.

Flowers in round heads; inflorescences solitary, paired

or fascicled in the axils of the leaves. Pods 7-12 x

1,4-1, 9 cm, valves thin, dehiscing longitudinally.

Recorded from Capital Park, Pretoria, Repton

1880; Grounds of Division of Botany, Pretoria,

Verdoorn s.n., Schlieben 10106; Bloemfontein, Potts

3219.

15. Acacia cultriformis A. Cunn. ex G. Don , Gen.

Syst. 2: 406 (1832); Benth. in FI. Austral. 2: 375

(1864); in Trans. Linn. Soc. Lond. 30: 474 (1875);

Brenan in F.Z. 3, 1: 113 (1970). Type from New

South Wales, Australia.

Unarmed shrub or small tree; young branchlets

angular, glabrous. Leaves phyllodic, apparently simple,

0,8-3 xO, 6-1 , 1 cm, obliquely obovate-lanceolate to

ovate-triangular, glaucous, glabrous, with a single

main longitudinal nerve and finely penninerved,

usually with 1 marginal gland, sometimes on a

prominent angle. Flowers in small round heads,

arranged in axillary racemes which are longer than

the phyllodes and are often ± aggregated terminally.

Pods 5-9x0, 5-0,7 cm, linear-oblong, glabrous,

longitudinally dehiscent.

Recorded from Stellenbosch, Garside 1248 (K).

A. cultriformis differs from A. poclalyriifolia in

being glabrous and in having narrower pods.

16. Acacia retinodes Schlechtend. in Linnaea 20:

664 (1847); Benth. in FI. Austral. 2: 362 (1864); in

Trans. Linn. Soc. Lond. 30: 468 (1875): Stapf &

Ballard, Bot. Mag. 153: t.9 1 77 (1929); Brenan in

FI. Trop. E. Afr. Legum.-Mimos.: 51 (1959); Court

in Willis, Hand. PI. Victoria 2: 227 (1972). Type

from Australia.

Unarmed glabrous shrub or small tree. Leaves

phyllodic, apparently simple, linear-lanceolate to

-oblong or oblanceolate, straight or slightly curved,

4,5-17 cm long, up to 1,5 cm wide, narrowing

gradually towards the base, with a single main

longitudinal nerve and finely but distinctly penni-

nerved. Flowers in round heads up to 6 mm in dia-

meter; inflorescences on peduncles 3-6 mm long,

arranged in short axillary racemes. Pods 7-12 x

0,5-0, 7 cm, linear-oblong, flattened, longitudinally

dehiscent; funicle encircling the seed in a double fold.

Recorded from Roodeplaat near Pretoria, Du Toit

105, 151, Schlieben & Mendelsohn 12717.

A. retinodes differs from A. saligna in having

smaller flower-heads, shorter peduncles and funicles

which encircle the seeds in a double fold.

17. Acacia fimbriata A. Cunn. ex G. Don, Gen.

Syst. 2: 406 (1832); Beadle, Evans & Carolin, FI.

Sydney Region: 267 (1972). Type from New South

Wales, Australia.

Unarmed shrub or small tree. Leaves phyllodic,

apparently simple, linear to narrowly oblong-elliptic,

2-4,5 cm long, 2-5 mm wide, narrowed basally,

with a single main longitudinal nerve, margins

typically densely ciliate, usually with a rounded

gland near the base. Flowers in small round heads,

arranged in axillary racemes. Pods linear-oblong,

straight, flattened, up to 7 cm long and 7 mm wide,

dehiscent.

Recorded from the grounds of the Union Buildings-

Pretoria, Repton 2640, Schlieben 10084: Grahams-

town, Troughton 228.

18. Acacia adunca A. Cunn. ex G. Don, Gen.

Syst. 2: 406 (1832); Maiden, For. FI. New South

Wales 5, part 46: 113-118, t.173 (1911). Type:

Australia, New South Wales, Hunters River, Cun-

ningham 79/1827 (K, holo.).

A. accola Maiden & Betche in Proc. Linn. Soc. New South

Wales 31(4): 734 (1907). Syntypes from Australia.

Unarmed small tree; young branchlets angular,

glabrous. Leaves phyllodic, apparently simple, 5-12

cm long, 1,5-3 mm wide (in our area), linear, with

a single main longitudinal nerve, usually with an

oblique slightly recurved point apically, a fairly

conspicuous marginal gland situated a short distance

above the base. Flowers in small round heads, arranged

in short axillary racemes which are mostly aggregated

terminally. Pods reddish-brown when mature, 7-10 x

0,8-1 cm, oblong, margins often irregularly con-

stricted, valves thin, umbonate over the seeds,

longitudinally dehiscent.

470

THE NATURALIZED AND CULTIVATED EXOTIC ACACIA SPECIES IN SOUTH AFRICA

Recorded from the Groot Drakenstein in the Cape

Province, Voorligtingsbeampte C4.

19. Acacia maidenii F. Muell. in Linn. Soc.

New South Wales Macleay Mem. Vol. 222: t.29

(1893); Court in Willis, Handb. PI. Victoria 2: 240

(1972); Beadle, Evans & Carolin, FI. Sydney Region:

271 (1972). Type from New South Wales, Australia.

Unarmed small to medium-sized tree. Leaves

phyllodic, apparently simple, 6-15x0,8-1,5 cm,

with 3-7 main longitudinal nerves and almost anasto-

mosing longitudinal veins. Flowers in elongate spikes

up to 4 cm long, spikes axillary, solitary or in twos

or threes. Pods 4-12 cm long, 3-5 mm wide, variously

coiled or twisted, pubescent.

Recorded from the Caledonian Grounds, Pretoria,

Repton 3766.

Differs from A. longifolia in having pubescent

coiled pods.

20. Acacia viscidula A. Cuim. ex Benth. in Hook.,

Lond. J. Bot. 1: 363 (1842); in FI. Austral. 2: 387

(1864); in Trans. Linn. Soc. Lond. 30: 480 (1875).

Type: Australia, New South Wales, banks of Lachlan

River, Fraser (K, holo.).

Unarmed shrub or small tree; young branchlets

angular, mostly sparingly pubescent, viscid. Leaves

phyllodic, apparently simple, 4,5-10 cm long, 1,25-

3 mm wide, linear, narrowed basally, with several

longitudinal nerves. Flowers in small round heads,

on axillary peduncles, solitary or paired, rarely

fascicled; peduncles up to 5 mm long, pubescent.

Sepals free or shortly united basally. Corolla

pubescent. Pods 4-7 cm long, 3-3,5 mm wide,

linear, sparingly to densely pubescent, longitudinally

dehiscent.

Recorded on the Cape Peninsula on the slopes below

the ruins of Lady Anne Barnard's cottage, Salter 9044.

21. Acacia pendula A. Cunn. ex G. Don, Gen.

Syst. 2: 404 (1832); Benth. in FI. Austral. 2: 383

(1864); in Trans. Linn. Soc. Lond. 30: 479 (1875);

Court in Willis, Handb. PI. Victoria 2: 238 (1972).

Type from New South Wales, Australia.

Unarmed tree or shrub. Leaves phyllodic, appa-

rently simple, linear to linear-oblong or lanceolate,

4,5-8 cm long, 3-9 mm wide, narrowed towards

the base, coriaceous, with several inconspicuous

longitudinal nerves, often greyish or glaucous.

Flowers in small round heads, usually arranged in

very short axillary racemes. Pods oblong, flattened,

4-8x0, 8-1 ,8 cm, the margins bordered by a narrow

wing 0,5-2 mm wide.

Recorded from a Johannesburg park, Hobson sub

PRE 32341: Middelburg, Cape, Loock sub PRE 32340.

ACKNOWLEDGMENT

I am most grateful to Mr L. Pedley, Senior Botanist,

Queensland Herbarium, Indooroopilly, for confirming

the identity of the specimens of A. adunca and A.

viscidula.

Bothalia 11, 4: 471-472 (1775)

The typification of Mimosa farnesiana

J. H. ROSS*

ABSTRACT

The protologue of Mimosa farnesiana L. in Species Plantarum ed.l: 521 (1753) was based on several

elements. In the absence of a type specimen it was considered desirable to select a lectotype in order to preserve

the current application of the name M. farnesiana.

The protologue of Mimosa farnesiana L., Sp. PI. ed.

1: 521 (1753), the basionym of Acacia farnesiana

(L.) Willd. in F., Sp. PI. ed.4,4: 1083 (1806), is

as follows:

“30. MIMOSA spinis geminis distinctis, farnesiana

foliis bipinnatis: partialibus octojugatis.

Hort. ups. 146.

Acacia indica farnesiana. Aid. fames. 2.

Raj. hist. 977.

Habitat in Domingo. h“

Analysis of the protologue indicates that it was based

on several elements, namely, a plant in cultivation

in the Botanic Gardens in Uppsala, and the Aldinus

and Ray synonymy.

Aldinus, Exactissima descriptio rariorum plan-

tarum Romae in Horto Farnesiano: 2-7 (1625),

provides, under the name Acacia Indica Farnesiana,

a very detailed description and two illustrations of

a plant in cultivation in the garden of Cardinal

Farnese in Rome. The illustration on p. 2 shows the

habit of the plant and the illustration on p. 4 (repro-

duced here as Fig. 1) is of a twig bearing flowers

and fruits. Aldinus records that seed of the plant

were received from the island of St. Domingo and

were germinated in the year 1611. John Ray, Hist.

Plant. 1 : 977 (1686), provides a summary of the

Aldinus diagnosis.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria.

Fig. 1. — Illustration of Acacia

Indica Farnesiana In Aldi-

nus, Descr. Rar. PI. Rom.

Hort. Farnesiano 4 (1625).

472

THE TYPIFICATION OF MIMOSA FARNESIANA

The diagnostic phrase-name of M. farnesiana in

Species Plantarum is essentially a synopsis of the

entry in Hortus Upsaliensis: 146 (1748). This entry in

Hortus Upsaliensis: 146 is as follows:

“6. MIMOSA spinis geminis distinctis, foliis

duplicato pinnatis, partialibus utrinque octo.

Acacia indica farnesiana. Aid. hort. 2. Raj.

hist. 977.

Acacia indica, foliis scopioidis leguminosae,

siliquis fuscis teretibus resinosis. Herm.

lugdb. 5.

Habitat in Domingo.

Hospitatur in Caldario, Arbor.

Obs. Flores sessiles, capitati, lutei, suaveolentes,

polyandri. Legumina teretia, crassa, utrinque

angustiora, obtusa."

The reference in the phrase-name to leaves with

8 pinnae pairs, “partialibus octojugatis” is odd

because M. farnesiana seldom, if ever, has 8 pinnae

pairs and this observation could not have come from

the Aldinus plate as suggested by Isely in Sida 3:

376 (1969) as the maximum number of pinnae

illustrated is seven.

Although it is quite clear that Linneaus had a

plant in cultivation in the Botanic Garden in Uppsala

the origin of this plant is not known and unfortunately

there is no specimen of M. farnesiana in the Linnaean

Herbarium in London. There is, however, a Linnaean

specimen of M. farnesiana in the Museum of Natural

History in Stockholm. Through the courtesy of

Prof. B. Nordenstam, Curator of the Botany Section

of the Museum of Natural History, Stockholm, this

Linnaean specimen was received on loan (see Fig. 2).

The specimen is not annotated by Linnaeus in any

way and it seems unlikely therefore that it was in

Linnaeus’s possession when Species Plantarum was

written. On the back of the sheet, at the bottom and

in the centre, is an annotation by Dahl, namely

“Dahl a Linne P”. According to Lindman, in Arkiv

for Bot. 7: 6 (1907), Dahl received many specimens

from Linnaeus during the latter’s later years and

during the years immediately after his death. It is

assumed therefore that this specimen of M . farnesiana

was in Linnaeus’s possession during his later life and

after Species Plantarum was written so that it could

not have been the specimen on which the phrase-

name of M. farnesiana was based, if the phrase-name

was indeed based on a specimen.

As no specimen on which Linnaeus could have

based his phrase-name of M. farnesiana is extant, and

as the phrase-name itself is not sufficiently diagnostic

to positively identify M. farnesiana , it is desirable to

have a lectotype to preserve the traditional and

current application of the name, particularly as

M. farnesiana is one of a complex of fairly closely

related species. The absence of a specimen necessitates

the selection of a lectotype from one of the other

elements in the protologue. It is quite clear that

Linnaeus relied to some extent at least on the Aldinus

synonymy for his concept of M. farnesiana. For

example, the specific epithet “ farnesiana ” was

obviously taken from Aldinus because there is

otherwise no explanation of why Linnaeus would

have used this epithet. The habitat recorded by

Linnaeus, namely, “Domingo”, is also that recorded

by Aldinus. The Aldinus plate reproduced here as

Fig. 1 is therefore taken as the lectotype of M.

farnesiana, and this accords with the view expressed

by Isely in Sida 3: 376 (1969).

Fig. 2. — The Finnaean specimen of Mimosa farnesiana L,

received on loan from the Museum of Natural History,

Stockholm.

As M. farnesiana is one of a complex of closely

related species it could perhaps be argued that the

Aldinus plate does not establish the identity of the

species beyond all possible doubt. With this in mind

it is proposed that the lectotype is backed up by the

representative Linnaean specimen in the Museum

of Natural History, Stockholm, and reproduced here

as Fig. 2. The specimen does not, of course, have any

status as far as typification is concerned, but it was

in Linnaeus’s possession and it is conceivable that it

may have come from the plant that was in cultivation

in the Botanic Garden in Uppsala. The conspicuous

venation of the leaflets of this Linnaean specimen

distinguish it from Acacia pinetorum F. J. Hermann

and from A. smal/ii Isely, and accord with the

traditional and current application of the name

M. farnesiana.

It is perhaps as well to mention that there is some

controversy over the authorship of the work here

attributed to Aldinus. Pritzel, Thesaurus Lit. Botanicae

ed. 2: 58 (1871) attributes the work to Castellus and

notes “Operis “Exactissima descriptio” autor est

Petrus Castellus, atque falso sibi vindicavit Aldinus;

typographus enim hisce etiam verbis: “In gratiam

Tobiae Aldini scripsi cuncta” profitetur, Aldinum

auctorem non esse. Seguier.” Aldinus was Cardinal

Farnese’s physician and so the work may well have

been dedicated to him. Saccardo, La botanica in

Italia: 12 (1895), credits Aldinus with the work.

In the Catalogue of the Library of the British Museum

(Natural History) 1: 26 (1903) the work is attributed

to Aldinus but there is a note reading “By some

this has been considered to be really the work of

P. Castelli.”

Bothalia 11, 4: 473-479 (1975)

The genus Acacia MilJer in South Africa. I. Stipules and spines

P. J. ROBBERTSE*

ABSTRACT

A large number of seedlings and young twigs of South African Acacia species was sectioned and the

vascularization of the nodes and internodes studied. The nodes of all the species examined are trilacunate

and the vascular tissue of the stipules originates from the lateral leaf traces. The Gummiferae species all have

spinescent stipules, while stipules of the Vulgares species are membranous. Prickles containing no vascular

tissue are found on the nodes and in some species also on the internodes of the Vulgares species. These prickles

always occur on the ridges formed on the stem by leaf traces.

INTRODUCTION

The genus Acacia in South Africa is represented by

only two of the six subgenera or series recognized

by Bentham (1875), viz. the Vulgares and the Gummi-

ferae. The Gummiferae species all have spinescent

stipules and no other spines or spinescent outgrowths

occur on the stem, while the Vulgares species have

membranous stipules together with prickles at the

nodes and/or on the internodes.

Vassal (1969) describes the external morphology of

the stipules and spines in seedlings of African Acacia

species. Nothing, however, is mentioned about the

nodal anatomy or the vascularization of the seedlings.

Vassal also bases the phyllotaxy of these seedlings on

their external morphology alone.

A study of the stipules and spines of the South

African Acacias cannot be regarded as complete

without a good knowledge of the vascularization of

the nodes and leaf bases. It was therefore necessary to

study the anatomy of the stem, in particular the

anatomy of the nodes and the vascularization of the

whole seedling axis.

MATERIALS AND METHODS

Seeds of all the South African Acacia species (see

p. 476-179 for list of all specimens used in this and

following studies, series I-V) were germinated and

seedlings were fixed in F.A.A. after they had reached

the four- to eight leaf stages. Serial sections were

made of the root, hypocotyl, cotyledonary node and

nodes of the first two vegetative leaves. For the

study of the nodal anatomy of young twigs on mature

trees, material of A. caffra ( Robbertse 851, PRU)

was used.

RESULTS AND DISCUSSION

Sketches of the serial sections of the seedling axis

of A. kraussiana from the root to the second vegetative

leaf are shown in Fig. I. Apart from differences in

length of the internodes and other small differences,

the basic anatomy of seedlings of the different species

is the same.

In Fig. la, the first vegetative leaf is sectioned

through the petiole, while the second leaf is sectioned

through the pulvinus. In the pulvinar region, the

vascular tissue of the leaf forms a closed cylinder.

Proximally, the cylinder breaks up to form the three

leaf traces. Fig. lc-e. From the lateral leaf traces

vascular tissue branches off to the stipules (Fig. If

and g).

Fig. 1 . — Line diagrams of a series of cross sections of the seedling axis of Acacia kraussiana, taken from attachment of

the first two vegetative leaves, down to the root. bl. 1, first vegetative leaf; bl. 2, second vegetative leaf; st, stem,

st. bl., stipules; w, x, y, z, vascular bundles of the stem.

Department of General Botany, University of Pretoria.

474

THE GENUS ACACIA MILLER IN SOUTH AFRICA. I. STIPULES AND SPINES

In all the Vulgares species, the ground tissue of the

stipules, as seen in the serial cross sections, fuses with

that of the stem before it fuses with the ground tissue

of the leaf base. In all the Gummiferae species the

ground tissue of the stipules fuses with the leaf base

before the latter joins the stem.

The nodes of all the leaves are trilacunar and the

“stele” of the stem consists of four collateral vascular

I

bundles. The central leaf tra:e of leaf No. 2 enters

the “stele” of the stem between vascular bundles

y and z (Fig. lg), while the two lateral leaf traces

enter through the leaf gaps between vascular bundles

w and y, and x and z respectively. The central leaf

trace of leaf No. 1 enters the “stele” of the stem

between vascular bundles xw and xz, while the

lateral traces enter through the leaf gaps between

bundles yz and xz, and xw and y respectively (Fig. 2).

It is clear from this investigation that the first and

second leaves are alternate, even though, organo-

graphically, they may appear to be opposite or sub-

opposite.

Below the nodes of insertion of the cotyledons, the

vascular tissue of the axis divides into two sections,

forming two lacunae through which the vascular

tissue of the cotyledons enters the stele. In Fig. 1 m-q

the cotyledons have been removed, but the two leaf

traces from each cotyledon can be seen entering the

lacunae. This region is, in fact, a compound node

consisting of two monolacunar nodes, each with

two leaf traces. Vassal (1969) has dealt with the

ramifications of the vascular tissue in the cotyledons

and this will therefore not be repeated in this paper.

In the lower part of the hypocotyl the vascular

cylinder breaks up into four vascular bundles. The

primary xylem of each bundle again divides into two

groups. Each of these primary xylem groups turns

through 180° in the direction of the nearest medullary

ray where it pairs with a second group of primary

xylem originating from an adjacent vascular bundle.

Thus a tetrarch stele is formed (Fig. 1 r-u).

Fig 2. — The phyllotaxy and vascularization of a

yong stem of Acacia caffra I, schematic diagram

of the primary vascular bundles; II, diagram of

the phyllotaxy. 1-6, trilacunar nodes.

The same type of vascularization as in the nodes

of the first and second vegetative leaves is found in

the nodes of the third, fourth and succeeding leaves

of the seedling and in nodes of young leaves on

mature trees. The resulting phyllotaxy is 2/5 (Fig. 2).

In this type of vascularization there are five prominent

vascular bundles in the internodes and, apart from

the leaf traces, four prominent bundles in the nodes.

The five prominent bundles in the internodes form

five ridges on the surface of young stems. Three of

these ridges below each node lead to three leaf

traces (Figs. 2 and 3). In the Vulgares species where

prickles are found, they occur only on the ridges.

In Acacia schweinfurthii, A. kraussiana and A. brevi-

spica the prickles occur in five longitudinal rows

along the five ridges on the internodes. The same type

of distribution of prickles is also found on seedlings

of A. hereroensis, young coppice shoots of A. here-

roensis and A. caffra. In A. caffra and A. hereroensis,

however, prickles are also found on the nodes.

In most other Southern African Vulgares species

two prickles are found on the nodes (Fig. 3), while

in the case of A. Senegal, three prickles are found

at each node. In species where two prickles are

P. J. ROBBERTSE

475

found, they originate on the ridges leading to the two

lateral leaf traces (Figs. 3 and 4). In A. Senegal with

three prickles at each node, they originate on the

ridges leading to the three leaf traces.

Fig. 3. — Line diagram of a series of cross sections through the petiole and node of Acacia caffra. C, section of

proximal part of petiole; D, E, section of pulvinus; F, G, H, section of node, fl, phloem; gb, glandular body;

lvb, lateral vascular bundle; oks, vascular tissue of axillary bud; svb, central vascular bundle; v, fibre cap;

vts, vascular tissue of stipule; x, xylem.

Fig. 4. — Schematic reconstruc-

tion of the course of the

primary vascular tissue in

the node of the stem of

Acacia caffra. Small letters,

a-h, refer to Fig. 1 and

capital letters, A-Ef, to Fig.

3.

476

THE GENUS ACACIA MILLER IN SOUTH AFRICA. I. STIPULES AND SPINES

In no South African Acacia species have prickles

ever been found between ridges as they are always

situated on the ridges. Prickles on the nodes do not

always appear at the same level, but one may be placed

lower down on the internode and in some cases one

or both may be lacking. Prickles are often found on

the leaf rhachis and rhachillae of certain species,

but even in these cases they occur opposite the

extension of the central leaf trace into the rhachis

or ramifications of the central leaf trace.

No vascular tissue is found in the prickles (Fig. 4).

They consist only of ground tissue covered by an

epidermis.

In the Gummiferae species and A. a/bida the spines

are modified stipules. As in the stipules of the Vulgares

species, stipules of Gummiferae species are supplied

by vascular tissue branching from the lateral leaf

traces (Fig. 3f and 4). The stipules of the Vulgares

species are caducous, but those of the Gummiferae

species develop into long, straight or curved persistent

spines.

Spines of A. erio/oba ( — A . giraffae sensu auct. mult,

fide Ross, 1975) are inflated at their bases and A. lue-

deritzii var. retinens also develops occasional inflated

spines. Monod & Schmidt (1968) have commented

on the pseudo-galls found in African Acacias , but in

this study, no further attention has been given to

this aspect.

The vascularization of the leaf in South African

Acacia species will be dealt with in a subsequent paper.

ACKNOWLEDGEMENTS

I am indebted to the Council for Scientific and

Industrial Research and the Research Committee of

the University of Pretoria for financial assistance.

UITTREKSEL

Snee van 'n groot aantal kiemplante en Jong stingels

van Suid-Afrikaanse Acacia-spesies is gemaak en die

verloop van die vaatweefsel in die knope en litte is

bestudeer. Die knope van al die ondersoekte soorte is

trilakuner en die vaatweefsel van die steunblare ont-

spring uit die later ale blaar spore. A l die Gummiferae-

soorte se steunblare is dorings terwyl die steunblare

van die Vulgares-soorte membraanagtig is. Haakdorings

wat geen vaatweefsel bevat nie kom op die knope en

by sekere soorte ook op die litte van die Vulgares-

soorte voor. Laasgenoemde dorings word steeds op

riwwe op die stam, wat deur die blaar spore veroorsaak

word, aangetref.

REFERENCES

Bf.ntham, G., 1875. Revision of the suborder Mimosaceae.

Trans. Linn. Soc. 30: 335-664.

Monod, T. & Schmidt, C., 1968. Contribution a l'etude des

pseudo-galls formicaires chez quelques acacias africains.

Bull, del' t.F.A.N. 30, 3 : 953-958.

Ross, J. H., 1975. Notes on African Acacia species. Bothalia

1 1 : 443-447.

Vassal, J., 1969. Contribution a l’etude de la morphologie des

plantules d’ Acacia. Acacias africains. Bull. Soc. Hist. Nat.

Toulouse. 105-111.

SPECIMENS USED

A. albida

South West Africa. — 1714 (Ruacana Falls): Ruacana Falls

on the Kunene River (-AC), Robbertse 327.

Transvaal. — 2230 (Messina): Tshipise rest camptseed only)

(-CA), Robbertse 2! 5; 2430 (Pilgrim's Rest): at tomato canning

factory, near Hoedspruit (-BD), Robbertse 139 ; 2528 (Pretoria):

between Rooiwal power station and Hammanskraal (-CA),

Robbertse 153 A.

A. arenaria

South West Africa.- 1816 (Namutoni): 66 km north of

Namutoni in Ovamboland ( BC), Robbertse 321 .

A. ataxacantha

Transvaal. — 2427 (Thabazimbi): Waterberg, Nylstroom

district (-DD), Robbertse 797: 2428 (Nylstroom): between

Nylstroom and Vaalwater (-CB), Robbertse 250 (a): 2527

(Rustenburg): Hartebeespoort Dam (-DD), Robbertse 387;

2528 (Pretoria): Springbok Park, Pretoria (-CA), Robbertse

237: 2529 (Witbank): western bank of Loskop Dam (-AD),

Robbertse 246: 2530 (Lydenburg): Lydenburg (-AB), Robbertse

134: 2530 (Lydenburg): Schoeman's Kloof (-AD), Robbertse

106: 2530 (Lydenburg): at 16,9 km mark between Machadodorp

and Nelspruit (-CB), Robbertse 805.

Cape Province.— 3129 (Port St John’s): 16 km from Port

St. John’s on road to Umtata (-CB), Robbertse 876.

A. borleae

Transvaal. — 2431 (Acornhoek): between Ngwanedzi and

Letaba, 11 km from Ngwanedzi-Satara cross roads (-BB),

Robbertse 225; 2431 (Acornhoek): 11km from Ngwanedzi on

road to Letaba, Kruger National Park (-BD), Robbertse 276;

2531 (Komatipoort): 8 km from Pretoriuskop on road to

Skipberg, Kruger National Park (-AB), Robbertse 281 ; 2531

(Komatipoort): Skipberg near Pretoriuskop (-AB), Robbertse

807; 2531 (Komatipoort): 20 km outside Malelane Rest Camp

at turn-off to Pretoriuskop, Kruger National Park (-BC),

Robbertse 173.

A. brevispica

Fixed material provided by Dr J. H. Ross of Pietermaritz-

burg.

A. burkei

Transvaal. — 2229 (Waterpoort): Johanna Uys Nature

Reserve, beyond Soutpansberg (-DC), Robbertse 212; 2426

(Mochudi): Derdepoort, Rustenburg district, 16 km from the

Limpopo River (-CB), Robbertse 188; 2427 (Thabazimbi):

45 km from Warmbad on road to Beestekraal (-DD), Robberste

164; 2428 (Nylstroom): at the foot of hill, north of Warmbad

(-CD), Robbertse 203; 2430 (Pilgrim's Rest): near Echo Caves

(-DB), Robbertse 136; 2431 (Acornhoek): between Tsokwane

and Satara, Kruger National Park (-DB), Robbertse 123;

2431 (Acornhoek): Leeupan, Kruger National Park (-DC),

Robbertse 228; 2431 (Acornhoek): 30,5 km from Skukuza on

road to Tsokwane near Leeupan, Kruger National Park (-DD),

Robbertse 175; 2526 (Zeerust): 77 km from Rustenburg via

Mabieskraal (-BB) Robbertse 184; 2527 (Rustenburg): 6 km

south of Northam (-AB), Robbertse 795; 2527 (Rustenburg):

Hartebeespoort Dam (-DD), Robbertse 386; 2529 (Witbank):

Loskop Dam (-AD), Robbertse 245; 2531 (Komatipoort):

34 km from Pretoriuskop on road to Skukuza, Kruger National

Park (-BA), Robbertse 813.

A. caffra

Transvaal. — 2327 (Ellisras): Ellisras (-DB), Robbertse and

v. d. Walt 879; 2428 (Nylstroom): 26 km from Nylstroom on

road to Ellisras (-CB), Robbertse and Van der Walt 882; 2429

(Zebediela): Zebediela (-AD), Robbertse 738 ; 2526 (Zeerust):

77 km from Rustenburg via Mabieskraal (-BB), Robbertse 182;

2528 (Pretoria): at 33 km mark north of Pienaarsrivier (-AB),

Robbertse 333; 2528 (Pretoria): Campus, University of Pretoria

(-CA), Robbertse 179; 2528 (Pretoria): Brummeria, Pretoria

(-CA), Robbertse 823; 2528 (Pretoria): Pienaarsrivier Dam,

near Pretoria (seed only) (-CB), Robbertse 17.

Natal. — 2729 (Volksrust); southern foothills of Drakensberg,

26 km from Newcastle on road to Volksrust (-DB), Robbertse

367.

Cape Province. — 3227 (Stutterheim): Fort Cox (-CC),

Robbertse 851 ; 3227 (Stutterheim): King William's Town

(-CD), Robbertse 781 .

A. clavigera (see A. robust a)

A. davyi

Transvaal. — 2230 (Messina): 40 km north of Louis Trichardt

on road to Sibasa (-CD), Robbertse 803; 2330 (Tzaneen): at

26,6 km mark north of Louis Trichardt, on road to Sibasa

(-AA), Robbertse 274 ; 2521 (Komatipoort): mountain pass

between Nelspruit and Malelane (-CB), Robbertse 103; 2530

(Lydenburg): at 49,9 km mark between White River and

Nelspruit (-BD), Robbertse 806.

A. erioloba

Transvaal. — 2229 (Waterpoort): 8 km north of Vivo (-CD),

Robbertse 802 ; 2526 (Zeerust): 77 km from Rustenburg via

Mabieskraal (-BB), Robbertse 183; 2527 (Rustenburg): between

Hartebeespoort Dam and Rustenburg (-DA), Robbertse 144(a);

2528 (Pretoria): At 4,8 km mark, north of Pienaarsrivier (-AB),

Robbertse 332.

Cape Province. — 2722 (Olifantshoek): 80 km north of

Kuruman (-BB), Robbertse 340.

P. J. ROBBERTSE

477

A. erioloba X A. haematoxylon

Cape Province. — 2623 (Morokweng): 80 km north-north-

west of Kuruman (-CC), Robbertse 339.

A. erubescens

South West Africa. — 2116 (Okahandja): 10 km west of

Okahandja (-DD), Robbertse 305.

Transvaal. — 2229 (Waterpoort): Johanna Uys Nature

Reserve, beyond Soutpansberg (-DC), Robbertse 211; 2426

(Mochudi): Derdepoort Hospital on banks of Limpopo River

(-CB), Robbertse 189; 2426 (Mochudi): Dwarsberg district of

Swartruggens (-DC), Robbertse 185; 2427 (Thabazimbi) :

14,5 km north-east of Thabazimbi on road to Thabazimbi

(-CB), Robbertse 196; 2528 (Pretoria): 22,5 km from Warmbad,

on road to Beestekraal, Brits (-AA), Robbertse 160; 2528

(Pretoria): 101 km from Warmbad on road to Brits (-CA),

Robbertse 165.

A. exuvialis

Transvaal. — 2426 (Mochudi): 32 km from Dwarsberg on

road to Derdepoort (-CB), Robbertse 187; 2431 (Acornhoek):

Olifants Rest Camp, Kruger National Park (-BB), Robbertse

224; 2431 (Acornhoek): 5 km from Ngwanedzi-Satara cross

roads, on road to Tshokwana, Kruger National Park (-DB),

Robbertse 226; 2431 (Acornhoek): Skukuza Rest Camp, Kruger

National Park (-DC), Robbertse 174; 2431 (Acornhoek):

Skukuza Rest Camp, Kruger National Park (seed only) (-DC),

Robbertse 278; 2531 (Komatipoort): Between Pretoriuskop

and Skukuza, Kruger National Park (-AB), Robbertse 112;

2531 (Komatipoort): Skipberg near Pretoriuskop, Kruger

National Park (-AD), Robbertse 809; 2531 (Komatipoort):

between Malelane gate and Rest Camp, Kruger National

Park (-CB), Robbertse 171.

A. fleckii

South West Africa. — 1816 (Namutoni): 11 km north of

Namutoni, Etosha Pan Game Reserve (-DB), Robbertse 320

(a).

Transvaal.— 2327 (Ellisras): between Ellisras and Stockpoort

(-DA), Robbertse and Van der Walt 878; 2426 (Mochudi):

32 km north of Dwarsberg Post Office on road to Derdepoort

(-CB), Robbertse 186; 2426 (Mochudi): 30,5 km mark between

Derdepoort and Rooibokkraal (-DC), Robbertse 191; 2427

(Thabazimbi): at 65,8 km mark between Northam and Makoppa

(-AC), Robbertse 285.

A. galpinii

Transvaal. — 2427 (Thabazimbi): Thabazimbi, in the town

(-CB), Robbertse 289; 2428 (Nylstroom): at 33 km mark north

of Nylstroom (Bad-syn-loop) (-CB), Robbertse 334; 2528

(Pretoria): Campus, University of Pretoria (-CA), Robbertse 34 :

2528 (Pretoria): Campus, University of Pretoria (-CA),

Robbertse 389.

A. gerrardii

Transvaal. — 2428 (Nylstroom): at 30,5 km mark north of

Nylstroom (-CB), Robbertse 2i8; 2428 (Nylstroom): Bad-syn-

loop, 30,5 km north of Nylstroom (-CB), Robbertse 798;

2429 (Zebediela): Potgietersrust (-AA), Robbertse 392; 2529

(Witbank): Loskop Dam (-AD), Robbertse 242; 2531 (Komati-

poort): Pretoriuskop, Kruger National Park (-AB), Robbertse

109; 2531 (Komatipoort): between Skukuza and Pretoriuskop

on road to Skipberg, Kruger National Park (-AB), Robbertse

229; 2531 (Komatipoort): 24 km from Pretoriuskop on road

to Skukuza, Kruger National Park (-AB), Robbertse 811.

A. giraffae (see A. erioloba )

A. grandicornuta

Transvaal. — 2327 (Ellisras): at junction of road from Villa

Nora between Vaalwater and Tom Burke (-DB), Robbertse 256;

2426 (Machudi): between Derdepoort and Rooibokkraal (-DA),

Robbertse 190; 2531 (Komatipoort): 47 km from Pretoriuskop

on road to Skukuza, Kruger National Park (-BA), Robbertse

815; 2531 (Komatipoort): between Skukuza and Lower Sabie,

Kruger National Park (-BB), Robbertse 118.

A. haematoxylon

Cape Province. — 2623 (Morokweng): 80 km north-north-

west of Kuruman (-CC), Robbertse 341; 2721 (Tellery Pan):

37 miles from Van Zyl’s Rust on road to Aroab (-AA), Robbertse

296(a); 2722 (Olifantshoek): 63 km from Kuruman on road to

Van Zyl's Rust (-BB), Robbertse 295 ; 2821 (Upington): 8 km

outside Upington on road to Keimoes (-CA), Robbertse 198.

A. hebeclada subsp. hebeclada

Transvaal. — 2329 (Pietersburg): 34,6 km mark between

Pietersburg and Vivo (-CB), Robbertse 800; 2528 (Pretoria):

32 km north of Pretoria on road to Warmbad (-AC), Robbertse

330; 2528 (Pretoria): 8 km from Rooiwal power station on

road to Warmbad (-CA), Robbertse 145; 2528 (Pretoria):

between Rooiwal power station and Hammanskraal, nearby

Pretoria. (-CA), Robbertse 154.

Cape Province. — 2722 (Olifantshoek): 63 km from Kuruman

on road to Van Zyl's Rust (-BB), Robbertse 294.

A. hebeclada subsp. tristis

South West Africa. — 1715 (Ondangua): 140 km from

Namutoni in Ovamboland (-DD), Robbertse 324.

A. hereroensis

South West Africa. — 2217 (Windhoek): 21 km south of

Windhoek on southern slope of hill (-CA), Robbertse 301 .

Transvaal. — 2526 (Zeerust): Lindley’s Poort (-BC)>

Robbertse 881; 2527 (Rustenburg): at 34,5 km mark north of

Rustenburg on road to Thabazimbi (-AD), Robbertse 794;

2527 (Rustenburg): 33,7 km north of Rustenburg on road to

Thabazimbi (-AD), Robbertse 284; 2528 (Pretoria): Zeiler

Street between Mitchell and Church Streets Pretoria (-CA),

Robbertse 153; 2528 (Pretoria): Zeiler Street between Mitchell

and Church Streets, Pretoria (-CA), Robbertse 236.

Cape Province. — 2624 (Vryburg): at 146 km mark between

Mafeking and Vryburg (-DD), Robbertse 292.

A. karroo

Transvaal. — 2527 (Rustenburg): Tweerivier, Rustenburg

district (seed only) (-CA), Robbertse 29; 2528 (Pretoria):

between Pienaarsrivier and Warmbad (-AB), Robbertse 158;

2528 (Pretoria): between Hammanskraal and Pienaarsrivier

(-AC), Robbertse 148; 2528 (Pretoria): Brummeria, Pretoria

(-CA), Robbertse 824; 2528 (Pretoria): Brummeria, Pretoria

(-CA), Robbertse 234; 2529 (Witbank): Loskop Dam (tall

drooping growth habit) (-AD), Robbertse 247 ; 2530 (Lydenburg)

Schoeman’s Kloof, Nelspruit (-AD), Robbertse 105; 2530

(Lydenburg): between Nelspruit and Barberton (-DB),

Robbertse 168.

Cape Province. — 2817 (Vioolsdrif): Vioolsdrif on banks of

Orange River (-DC), Robbertse 199; 3118 (Vanrhynsdorp):

at foot of Van Rhyn's Pass, Vanrhynsdorp (-DB), Robbertse 203;

3227 (Stutterheim): 8 km north of King William's Town (-CD),

Robbertse 782; 3228 (Butterworth): Kei Mouth (-CB), Robbertse

875.

A. kirkii

South West Africa. — 1817 (Tsitsib): 8 km from the gates of

Etosha Pan Game Reserve, Between Namutoni and Tsumeb

(-CC), Robbertse 329.

A. kraussiana

Natal. — 2732 (Ubombo): Makatini Flats (seed only)

(-CA), Robbertse 24; 2930 (Pietermaritzburg): Pietermaritzburg

(-CB), Robbertse and Reyneke 880.

A. luederitzii var. retinens

Transvaal. — 2426 (Mochudi): 39,6 km mark between

Derdepoort and Rooibokkraal (-DC), Robbertse 192: 2528

(Pretoria): between Pienaarsrivier and Warmbad (-AB),

Robbertse 149; 2528 (Pretoria): between Pienaarsrivier and

Warmbad (-AB), Robbertse 156: 2528 (Pretoria): 27,5 km

from Warmbad on road to Beesterkraal (-AB), Robbertse 163:

2528 (Pretoria): at 57 km mark between Pretoria and Pienaars-

rivier (-AD), Robbertse 249.

A. luederitzii var. leuderitzii

South West Africa. — 1715 (Ondangua): Oshikango, Ovam-

boland (-BD), Robbertse 325; 1816 (Namutoni): Namutoni

Rest camp, Etosha Pan Game Reserve (-DD), Robbertse 319.

A. mellifera

Transvaal. — 2228 (Maasstroom): 16 km from Mogalakwena

River on road to Alldays (-DD), Robbertse 262 ; 2428

(Nylstroom): At 12 km mark between Vaalwater and Villa

Nora (-AA), Robbertse 253; 2428 (Nylstroom); at 49 km mark

between Pienaarsrivier and Warmbad (-CD), Robbertse 202;

2527 (Rustenburg): 24 km outside Rustenburg on road to

Zwartruggens (-CA), Robbertse 239; 2528 (Pretoria): Pienaars-

rivier (-AB), Robbertse 331; 2528 (Pretoria); Pienaarsrivier.

near Warmbad (-AB), Robbertse 155; 2528 (Pretoria) : at 49,8 km

mark between Pretoria and Pienaarsrivier (-AD), Robbertse 248.

Cape Province. — 2723 (Kuruman): 1 1 km west of Kuruman

(-AD). Robbertse 342.

A. montis-usti

South West Africa.— 2014 (Welwitschia): on the farm

Goedgegun on road to Twyfelfontein (-DB), Robbertse 316;

2114 (Uis): Witvroukloof, Brandberg (-BA), Robbertse 315.

478

THE GENUS ACACIA MILLER IN SOUTH AFRICA. I. STIPULES AND SPINES

A. nebrownii

South West Africa. — 2618 (Keetmanshoop): 80 km west

of Aroab on road to Keetmanshoop (-DB), Robbertse 298.

Transvaal. — 2228 (Maasstroom): 16 km from Mogalakwena

River on road to Alldays (-DD), Robbertse 261 ; 2229 (Water-

poort): at 4.7 milestone between Vivo and Alldays (-CD),

Robbertse 208 ; 2229 (Waterpoort) : 13 km north of Vivo on

road to Alldays (-CD), Robbertse 337; 2229 (Waterpoort):

western edge of Soutpan, north of Soutpansberg (-DC),

Robbertse 213; 2230 (Messina): Farm Adieu on the Ngwanedzi

River, near Tshipise (-CA), Robbertse 217.

A. nigrescens

Transvaal. — 2231 (Pafuri): 0,8 km from Punda Milia

Rest Camp, Kruger National Park (-CA), Robbertse 818;

2431 (Acornhoek): Satara Rest Camp, Kruger National Park

(-BD), Robbertse 124; 2528 (Pretoria): Campus, University of

Pretoria (-CA), Robbertse 201; 2528 (Pretoria): Campus,

University of Pretoria (-CA), Robbertse 388.

A. nilotica

Transvaal. — 2528 (Pretoria): Derdepoort, Pretoria (-CA),

Robertse 391 ; 2528 (Pretoria): Wonderboom, Pretoria (-CA),

Robbertse 791; 2528 (Pretoria): Pienaarsrivier Dam, Pretoria

(seed only) (-CB), Robbertse 18, 2529 (Witbank): Loskop Dam

(-AD), Robbertse 243.

A. permixta

Transvaal. — 2329 (Petersburg): at 52,3 km mark between

Pietersburg and Vivo (-AD), Robbertse 205; 2329 (Petersburg) :

between Vivo and Pietersburg (-AD), Robbertse 264; 2329

(Pietersburg) : at 43,5 km mark north west of Pietersburg on road

to Vivo (-CA), Robbertse 270; 2329 (Pietersburg): at 42,6 km

mark between Pietersburg and Vivo (-CA), Robbertse 801;

2427 (Thabazimbi): between Rooibokkraal and Thabazimbi

(-AC), Robbertse 196; 2427 (Thabazimbi): at 106,2 km mark

between Sentrum and Thabazimbi (-AD), Robbertse 288;

2527 (Rustenburg): 56 km from Brits on road to Beestekraal

(-BC), Robbertse 166.

A. polyacantha

Transvaal. — 2230 (Messina): 50 km from Louis Trichardt

on road to Punda Milia (-CD), Robbertse 804; 2230 (Messina):

32 km from Punda Milia on road to Louis Trichardt (-DD),

Robbertse 220 ; 2230 (Messina): 66,7 km mark between Sibasa

and Punda Milia (-DD), Robbertse 275 ; 2231 (Pafuri): Punda

Milia Rest Camp, Kruger National Park (-CA), Robbertse

87 9- 2231 (Pafuri): Punda Milia, Kruger National Park (-CA),

Robbertse 177.

A. rehmanniana

Transvaal. — 2329 (Pietersburg): Louis Trichardt (north of

town) (-BB), Robbertse 178; 2329 (Pietersburg): Louis Trichardt

(north of town) (-BB), Robberste 821; 2329 (Pietersburg):

8 km outside Pietersburg on road to Vivo on banks of Blood

River (-CD), Robbertse 204; 2329 (Pietersburg); Blood River,

just outside Pietersburg (-CD), Robbertse 265; 2329 (Pieters-

burg): 16 km mark from Pietersburg on road to Potgietersrus

(-CD), Robbertse 269; 2329 (Pietersburg): Pietersburg (-CD),

Robbertse 338; 2329 (Pietersburg): 4 km west of Pietersburg

(-CD), Robbertse 799.

A. reficiens

Southwest Africa. — 2114 ( U is) : 3,2 km east of Uis

(-BB), Robbertse 312; 2116 (Okahandja): 3,2 km north of

Wilhelmstal (-CD), Robbertse 307; 2116 (Okahandja): 4,8 km

west of Okahandja (-DD), Robbertse 304; 2217 (Windhoek):

22,5 km north of Windhoek (-AC), Robbertse 302.

A. robusta

Transvaal. — 2528 (Pretoria): Campus, University of

Pretoria (-CA), Robbertse 35; 2528 (Pretoria): Campus,

University of Pretoria (-CA), Robbertse 180; 2530 (Lydenburg):

Schoemanskloof, near Nelspruit (-AD), Robbertse 100.

Cape Province. — 3227 (Stutterheim): near bridge over

Kei River (-DB), Robbertse 869; 3227 (Stutterheim): near

bridge over Kei River (-DB), Robbertse 873 ; 3228 (Butterworth) :

1 6 km from Kei Mouth (-CB), Robbertse 870.

A. robynsiana

South West Africa. — 2014 (Welwitshia): 38,6 from Goed-

gegun on road to Twyfelfontein (-DB), Robbertse 317 and

317 (a).

A. schweinfurthii

Transvaal. — 2230 (Messina): Farm Adieu beside Ngwanedzi

River near Tshipise (-CA), Robbertse 218; 2231 (Pafuri):

Punda Milia Rest Camp, Kruger National Park (-CA), Rob-

bertse 221; 2231 (Pafuri): Punda Milia Rest Camp, Kruger

National Park (-CA), Robbertse 820; 2430 (Pilgrim's Rest):

petrol station on banks of Blyde River near Hoedspruit (-BD),

Robbertse 138; 2431 (Acornhoek): Skukuza, Kruger National

Park (-DC), Robbertse and Reyneke 877.

A. Senegal var. leiorhachis

Transvaal. — 2229 (Waterpoort): at 51,2 km mark between

Louis Trichardt and Messina (-DB), Robbertse 214; 2230

(Messina): at foot of hill near Tshipise Rest Camp (-CA),

Robbertse 216; 2329 (Pietersburg): Vivo (-AB), Robbertse 207;

2329 (Pietersburg): Vivo (-AB), Robbertse 238; 2329 (Pieters-

burg): 11 km south of Vivo between Pietersburg and Vivo

(-AB), Robbertse 335.

A. Senegal var. rostrata

South West Africa. — 2217 (Windhoek): 22,5 km north

of Windhoek (-AC), Robbertse 303.

Transvaal. — 2328 (Baltimore): 14,5 km from Villa Nora

on road to Marnitz (-AC), Robbertse 257; 2426 (Mochudi):

60 km mark between Derdepoort and Rooibokkraal (-DA),

Robbertse 193; 2531 (Komatipoort) : between Pretoriuskop and

Skukuza, Kruger National Park (-AB), Robbertse 115; 2531

(Komatipoort): 3,2 km outside Malelane Rest Camp, on road

to Skukuza, Kruger National Park (-BC), Robbertse 172;

2531 (Komatipoort): 32 km from Pretoriuskop on road to

Skukuza, Kruger National Park (-BA), Robbertse 812.

A. sieberana var. woodii

Transvaal. — 2529 (Witbank): Dennilton near Loskop Dam

(-AC), Robbertse 241; 2528 (Pretoria): Campus, University

of Pretoria (-CA), Robbertse 31; 2528 (Pretoria): Campus,

University of Pretoria (-CA), Robbertse 235; 2530 (Lydenburg):

Schoemanskloof, near Nelspruit (-AD), Robbertse 101.

A. swazica

Transvaal. — 2531 (Komatipoort): east of Skipberg, Kruger

National Park (-AD), Robbertse 230; 2531 (Komatipoort):

14.5 km from Pretoriuskop at Skipberg (-AD), Robbertse 282;

2531 (Komatipoort): Skipberg near Pretoriuskop, Kruger

National Park (-AD), Robbertse 810; 2531 (Komatipoort):

45 km outside Barberton on road to Malelane (-CB), Robbertse

170; 2531 (Komatipoort): 5 km outside Barberton on road to

Malelane (-CC), Robbertse 169.

A. stuhlmannii

Transvaal. — 2228 (Maasstroom): at Groblersdrif across

Limpopo River near Swartwater (-CC), Robbertse 260; 2328

(Baltimore): 21 km from Villa Nora on road to Marnitz (-AC),

Robbertse 240; 2328 (Baltimore): 21 km from Villa Nora on

road to Marnitz (-AC), Robbertse 259; 2329 (Pietersburg):

3,2 km north of Vivo (seed only) (-AB), Robbertse 271 ; 2329

(Pietersburg): At 3,2 km mark north of Vivo on road to Alldays

(-AB), Robbertse 336.

A. tortilis subsp. heteracantha

South West Africa. — 2116 (Okahandja): 19 km from

Wilhelmstal on road to Omaruru (-CA), Robbertse 308; 2116

(Okahandja): Wilhelmstal (-CD), Robbertse 306.

Transvaal. — 2431 (Acornhoek): Skukuza Rest Camp between

camp and living quarters, Kruger National Park (-DC), Robbertse

816; 2527 (Rustenburg): 29 km north of Rustenburg on road to

Thabazimbi (-AD), Robbertse 793; 2527 (Rustenburg): Twee-

rivier, Rustenburg district (-CA), Robbertse 144; 2528 (Pretoria):

22.5 km from Warmbad on road to Beestekraal (-AB), Robbertse

162; 2528 (Pretoria): 8 km from Rooiwal power station on

road to Warmbad (-CA), Robbertse 147; 2528 (Pretoria): at

47.5 km mark between Pretoria and Hammanskraal (-CA),

Robbertse 266; 2529 (Witbank): Loskop Dam (-AD), Robbertse

244.

A. tortilis subsp. spirocarpa

South West Africa. — 1714 (Ruacana Falls): Ombalantu,

Ovamboland (-DB), Robbertse 326.

A. tenuispina

Transvaal. — 2328 (Baltimore): Turflaagte, 17,5 km from

Villa Nora (-AC), Robbertse 258; 2426 (Mochudi): Rooibok-

kraal (-BB), Robbertse 286; 2426 (Mochudi): at Rooibokkraal

in pan (-BB), Robbertse 195; 2426 (Mochudi): at 30,5 km mark

between Derdepoort and Rooibokkraal (-DC), Robbertse 194;

P. J. ROBBERTSE

479

2527 (Rustenburg): 10,5 km north of Rustenburg on road to

Thabazimbi (-CB), Robbertse 283; 2527 (Rustenburg): 5 km

north east of Rustenburg on road to Thabazimbi (-CB),

Robbertse 792; 2528 (Pretoria): between Pienaarsrivier and

Warmbad (-AB), Robbertse 150; 2528 (Pretoria): between

Pienaarsrivier and Warmbad (-AB), Robbertse 159; 2528

(Pretoria): 9,7 km mark between Pienaarsrivier and Warmbad

(-AB), Robbertse 250; 2528 (Pretoria): at 9,7 km mark between

Pienaarsrivier and Warmbad (-AB), Robbertse 267.

A. welwitschii

Transvaal. — 2431 (Acornhoek): Gomondwane, Kruger

National Park (seed only) (-DC), Robbertse and Van Wyk 232;

2431 (Acornhoek): Leeupan, Kruger National Park (-DC),

Robbertse 227; 2431 (Acornhoek): Leeupan, Kruger National

Park (-DC), Robbertse 277; 2431 (Acornhoek): Leeupan,

Kruger National Park (-DC), Robbertse 817 .

A. xanthophloca

Transvaal. — 2230 (Messina): on farm of Department of

Agricultural Technical Services, west of Messina on banks of

Limpopo River (-AA), Robbertse 273 ; 2230 (Messina): Messina,

west of bridge over Limpopo River (-AC), Robbertse 222;

2231 (Pafuri): Pafuri, Kruger National Park (-AD), Robbertse

223; 2431 (Acornhoek): 8 km outside Satara on road to Letaba,

Kruger National Park (-BD), Robbertse 176; 2528 (Pretoria):

Pretoria North (cultivated) (-CA), Robbertse 390.

26700—8

Bothalia 11, 4: 481-489 (1975)

The genus Acacia in South Africa. IV. Hie morphology of

the mature pod

P. J. ROBBERTSE*

ABSTRACT

The external morphology and the anatomy of the pods of all the South African Acacia species are discussed.

It was found that the South African Acacia species can be grouped into a number of distinct categories on the

basis of the anatomy of their pods. There is quite a marked difference in the anatomy of the pod of the series

Vulgares and that of the series Gummiferae.

INTRODUCTION

Most authors who provide keys to the South African

and other African Acacia species (Oliver, 1871;

Bentham, 1875; Harvey, 1862; Keay, 1958; Baker,

1930; Burtt Davy, 1932; Young, 1955; Verdoorn,

1957; Brenan, 1959 and Von Breitenbach, 1965) use

the shape and external structure of the pod as taxono-

mically important characters. No previous instance is

known where the anatomy of the pod was used as a

taxonomic character to distinguish between Acacia

species. Kraus (1866), Hildebrandt (1873-1874) and

Majewsky (1873) made a general study of the anatomy

of the pod, and Fahn and Zohary (1955) devised a

scheme according to which they classified the pods of

a number of Leguminosae genera on an anatomical

basis.

In the course of this study the external morphology

and anatomy of the pods of all South African Acacia

species were examined. Only mature preserved and

dried pods were used. For the external and anatomical

examinations at least 10 pods of each species derived

from two or more localities were intensively studied,

and general observations were made in the field and

on samples from the National Herbarium in Pretoria.

For the anatomical examination, microtome sections

were made of one pod of each species as well as hand

sections of at least two other pods per species.

EXTERNAL MORPHOLOGY OF THE POD

The pods of South African Acacia species vary

considerably in form and texture. They may be linear,

elongated, straight, crescent-shaped or curled into a

spiral, flattened, round, sometimes constricted (Fig. 1).

They have a membraneous, leathery or woody texture

and the surface may be smooth, hairy or covered with

glands. The texture of the valves is a character

determined by the anatomy of the pods. This character

is accordingly discussed under anatomy.

Fig. 1. — Pods of some southern African Acacia species. 1. A. albida; 2, A. schweinfurthii ; 3, A. ataxacantha, 4, A. htn-

roensis; 5, A. caffra; 6, A. senega! var. rostra, a; 7, A. senega! var. leiorhachis; 8 A melhfera subsp. de,mens,9A

erubescent 10, A.galpinii; 11, A. nigrescens; 12, A. burkei ; 13, A.montis-usti ,1 4 , A. enoloba; 1. S A . haematoxy on

16, A. sieberana vltwoodii; 17, A. robusta subsp. clavigera; 18, A luedentz,, var .retmens; 19. A reficuns 20,

A. hebeclada subsp. hebeclada; 21, A. hebeclada subsp. chobiensis; 22, A. stuhlmannu ; 23 A rehmanniana , 24 A

grandicornuta; 25, A. tor, ills subsp. spirocarpa; 26, A. tor, ills subsp. he te mean, ha; 27 A kirku 28, A. nilotica

subsp. kraussiana; 29, A. gerrardii var. gerrardii; 30, A. xanthophloea; 31, A. arenana; 32, A. davyu 33, A. kanoo,

34, A. tenuispina; 35, A. exuvialis; 36, A. borleae; 37, A. nebrownti; 38, A. permixta, 39, A. swazica.

* Department of General Botany, University of Pretoria.

482

THE GENUS ACACIA IN SOUTH AFRICA. IV. THE MORPHOLOGY OF THE MATURE POD

Most species are characterized by dehiscent pods,

but the pods of A. schweinfurthii, A. kraussiana, A.

albida, A. erioloba (—A. giraffae sensu auct. mult, fide

Ross, 1975), A. haematoxylon, A. stuhlmannii, A.

xanthoploea and A. nilotica do not normally open.

A. hebeclada pods usually open along one side,

whereas A. sieberana and A. tortilis pods are sluggish

in opening and usually remain closed. According to

Fahn and Zohary (1955) the ability of a pod to dehisce

is determined by its anatomy and by the fibrils in the

fibre walls.

Pods exhibit three types of epidermal appendages:

hairs, multicellular glands and minute, reddish-brown,

multicellular organs. The hairs are unicellular and

their length varies from a few millimetres in A.

stuhlmannii to papilliform hairs in A. tortilis subsp.

heteracantha. Species with hairy ovaries do not

necessarily bear hairy pods, e.g. A. xanthophloea and

A. reficiens. The same applies to species with gland-

covered pods. A. karroo, like A. nebrownii, A. borleae,

A. tenuispina, A. swazica, A. exuvialis and A. permixta,

has glands on the ovaries but the mature pods of

A. karroo possess few or no glands whereas the

mature pods of a species such as A. permixta are

covered with glands. In respect of the number of

glands on the mature pod, A. exuvialis lies between

A. karroo and A. permixta and the other species

between A. exuvialis and A. permixta (see also Ver-

doorn, 1957 and Ross, 1972).

All the species with glands on the mature pods are

shrub-like. A. borleae and A. exuvialis can grow to

some 3 m in favourable conditions, but most of

these shrubs vary from 1 to 2 m. A dwarfed form of

A. karroo is also found on the Springbok Flats and

near Rustenburg. These variations of A. karroo require

intensive study and the problem will have to be passed

over here for lack of further data.

Few data are available on the presence of the

gland-like bodies (Fig. 2A) found in some Acacia

species. Brenan (1959) writes in connection with A.

brevispica of the . . reddish or dark glands with

which the inflorescence-axes are sprinkled, . . The

same type of gland-like bodies are found on the pods

of A. erioloba, A. stuhlmannii, A. sieberana, A.

haematoxylon and A. nilotica, all belonging to the

series Gummiferae, and A. schweinfurthii, A. polya-

cantha, A. hereroensis, A. fleck ii and A. Senegal of the

series Vulgares. The number of glands per surface

unit not only varies considerably according to species,

but also within the same species. Fig. 2B shows young

pods of A. erioloba from two different trees growing

close to the Pienaars River. The dark colour of the

three pods is caused by a dense concentration of

glandular bodies; the concentration on the light

coloured pod is less dense. All the pods on one tree

have the same colour and the colour difference in the

pods from the two trees is conspicuous, even at a

distance of 80 m. The glands can also occur on the

vegetative parts, e.g. in the seedlings of A. erioloba,

where they occur in the axils of leaves and cotyledons.

A

i

Fig. 2. — A. gland-like body on part

of section through Acacia erio-

loba pod; B, twigs and pods of

A. erioloba with dark pods

crowded with gland-like bodies,

gl, gland-like body.

P. J. ROBBERTSE

483

These minute, reddish-brown, gland-like bodies

consist of epidermal cells with the walls thickened

and permeated with tannin. The epidermal cells

enclose a number of thin-walled packing cells.

These organs with their elongated stalks and incrassate

cellular walls are very similar to the sporangium of a

fern, but their function is unknown. Their presence

may be of taxonomic importance, but too little is

known on the subject at present for any comment

to be possible.

ANATOMY OF THE POD

The abaxial epidermis ( outer epidermis)

In the discussion of the external morphology

reference was made to the structure of the abaxial or

outer epidermis of the pod and the glands

characteristic of certain species. These glands are

shield-shaped, with a diameter of approximately

0,5 mm. They are solid while young, but as soon as

they enlarge, a hypodermal cavity is formed at the

tip. The cavity is formed schizogenously, since the

cells do not disintegrate but merely separate (Fig. 3 A).

The gland cavity is filled with a glutinous fluid which

is secreted immediately the epidermal cells are

mechanically damaged, or else through a glandular

aperture which is schizogenously formed. This

glandular secretion appears to be a good nourishing

medium for certain fungi (unidentified), since older

glands are constantly infected with them. The glan-

dular stalks vary from species to species. A. exuvialis

and A. nebrownii have thin stalks (some 3 to 5 cells in

diameter), whereas A. permixta has a considerably

thicker stalk (10 cells or more in diameter). The

glands on A. bor/eae pods are slightly submerged

below the pod surface, whereas those of A. permixta

are usually found on a protrusion.

Fig. 3.— Segments of cross sections through the pericarp. A, B, Acacia swazica; C, A. schweinfurthii ; D, A. poly-

cantha. ae, adaxial epidermis; en, endocarp; fs, fibre stratum; gl, gland; ops, outer parenchyma stratum, sc,

schizogenous cavity; scl, sclereid stratum.

484

THE GENUS ACACIA IN SOUTH AFRICA. IV. THE MORPHOLOGY OF THE MATURE POD

The adaxial epidermis ( inner epidermis)

The adaxial epidermis of the pods is usually only

one cell layer thick. Fig. 3C shows local periclinal

divisions in the epidermal cells of A. schweinfurthii to

form more layers of cells. In certain species the

epidermal cells may grow out to form multicellular

hairs, as shown by Fahn and Zohary (1955). Several

species, e.g. A. niloticci and A. rehmanniana , form

a spongy tissue around the seeds: in a cross-section

of the pod this tissue resembles chains of thin-walled

intertwined parenchymatous cells (Fig. 3D). The cell

walls of adjacent cellular chains fuse to form a

pseudo-parentchymatous tissue.

To trace the origins of this pseudo-parenchymatous

tissue, an ontogenetical study of the pods was

necessary. Only the pods of A. karroo of the series

Gummiferae and A. caffra of the series Vulgares were

used for this study. The young ovaries of both types

have papilliform epidermal cells around the base of

the funiculus (“transmitting tissue”, Fahn, 1967).

These papillaceous epidermal cells (Fig. 4) are most

conspicuous and grow out to form multicellular hairs,

which grow into the loculus, early in the develop-

ment of the pod — apparently immediately after

fertilization. These multicellular hairs or cellular

chains are largely responsible for the formation of the

pseudo-parenchyma in A. karroo and A. caffra and

there is reason to believe that this is the case with

most other species. Further discussion will refer to the

pseudo-parenchymatous tissue as a spongy endocarp

to distinguish it from the outer portion of the pericarp

and the fibre stratum of the endocarp. In A. caffra the

pseudo-parenchymatous tissue occupies only a small

portion of the loculus near the placenta: the rest of

the epidermis consists of one layer, occasionally of two

ayers of cells.

Fig. 4. — Part of cross section through the flower of Acacia

karroo to show the structure of the ovary wall, ov,

ovule; ow, ovary wall ; tt, transmitting tissue.

In species such as A. fieckii it is difficult to determine

to what extent the rest of the adaxial epidermal cells,

away from the placenta, contribute to the formation

of pseudo-parenchymatous tissue, since the walls of

the cellular chains derived from the placenta will also

merge with cell walls of other epidermal cells.

It is clear, however, that the adaxial epidermal cells

can grow out into multi-cellular hairs in various

places.

In A. erioloba and A. haematoxylon virtually all

the adaxial epidermal cells grow out to become part

of the spongy pseudo-parenchyma, and no original

epidermal cells can be observed in the mature pod.

In A. haematoxylon even subepidermal tissues may

contribute to the formation of the spongy endocarp,

but since there is no sclerenchymatous stratum

(Fahn and Zohary, 1955) in this case, it is impossible

to determine the boundaries of the epidermis in the

mature pod. In the other species which do possess a

sclerenchymatous zone, it extends to adjoin the

adaxial epidermis.

The mesophyll of the pericarp

According to Fahn and Zohary (1955) the mesophyll

of the pericarp is divided into an outer parenchyma-

tous stratum and an inner sclerenchymatous stratum.

The parenchymatous stratum extends from the

abaxial epidermis up to the sclerenchymatous stratum

and includes the vascular bundles and the various

sclereid layers. The sclerenchymatous stratum adjoins

the adaxial epidermis and consists of longitudinal

and latitudinal fibres and sclereid layers (Fig. 5E)

but one, two or all three of these layers may be

partially or completely absent (Fig. 5A, H and I).

This classification of the pericarp is useful for anato-

mical purposes, but is not ontogenetically correct.

The wall of the ovary, as in A. karroo , consists of

two clearly distinguishable strata (Fig. 4A). The

outer stratum consists of the abaxial epidermis and a

hypodermal layer. The cells of both these layers

contain a secretion-like substance, and the hypodermal

layer is responsible for the formation of a portion

of the parenchymatous stratum outside the vascular

bundles (Fig. 6A — C). The other stratum of the

ovary wall extends from the hypodermis up to the

adaxial epidermis and consists of smaller, colourless

cells. The vascular bundles, sclerenchymatous stratum

and sclereid cells originate in this layer.

For the purpose of this discussion, the classification

of the pericarp made by Fahn and Zohary (1955)

will not be used, since it makes no provision for a

clear delimitation of the sclereid layers. Instead, the

terms used will be parenchymatous stratum which

includes the vascular bundles with their sclerenchyma

caps (fibre caps) a fibre stratum adjacent to the

adaxial epidermis and consisting of longitudinal or

else of longitudinal and latitudinal fibres, and a

sclereid stratum which may be situated in the paren-

chymatous or next to the fibre stratum.

Using the data collated during this study, a key

can be drawn up for the identification of the pods of

South African Acacia species based on the anatomy

of the pod (Fig. 5). For this classification, the arrange-

ment of pericarpal tissues is schematically shown for

each of the eleven groups (Fig. 5). Pods of the series

Gummifera (sensu Rob.*) and series Vulgares (sensu

Benth.) can be distinguished by the fact that the

fibre stratum of the Gummiferae consists of longi-

tudinal fibres only (Fig. 5H, J and K) or is absent

(Fig. 5G and I), whereas those of the Vulgares

consists of longitudinal and latitudinal fibres. In

some Vulgares species, too, sclereid cells may be

found on top of the fibre stratum (Fig. 5E and F).

According to this classification, A. albida belongs to

the series Vulgares (sensu Benth.), since the pericarp

of the latter taxon also contains longitudinal as well

as latitudinal fibres.

Fahn and Zohary (1955) made an anatomical study

of the pods of 78 species of 54 genera of the Legumi-

nosae and were able, on the basis of this study, to

show developmental tendencies amongst certain

groups. They showed, for instance, that from the

Mimosoideae through the Caesalpinioideae to the

* Gummiferae sensu Benth., excluding A. albida.

P. J. ROBBERTSE

485

Fig. 5.— Key to the grouping of southern African Acacia species, based on the anatomy of the pod.

486

THE GENUS ACACIA IN SOUTH AFRICA. IV. THE MORPHOLOGY OF THE MATURE POD

Fig. 6. — Parts of sections through the margins of Acacia pods. A, A. kirkii; B, A. karroo; C, A.

ataxacantha. art, articulation; ops, outer parenchymatous stratum; fs fibre stratum.

Papilionoideae the sclerenchymatous stratum tends

to diminish. According to this view the series Gummi-

ferae is considered more advanced than the Vulgares,

since the latter’s fibre stratum consists of latitudinal

as well as longitudinal fibres whereas in the Gummi-

ferae it is reduced to longitudinal fibres only or is

absent altogether.

Unfortunately Fahn and Zohary did not include

many examples from the genus Acacia and come to

some faulty conclusions, such as the following: “Thus

in the subfamily Mimosoideae, the most primitive

subfamily of the Leguminosae, no species was found

without any sclerenchymatous stratum”, whereas this

fibre stratum is absent in two South African Acacia

species alone, viz. A. haematoxylon and A. stuhlmannii.

In their description of their “ Acacia type" pericarp

they state that the sclerenchymatous stratum consists

of latitudinal and longitudinal fibres. At the same

P. J. ROBBERTSE

487

time they illustrate a cross-section of part of the

pericarp of A. raddiana (—A. tortilis subsp. raddiana)

in which only longitudinal fibres occur. They were,

therefore, unable to distinguish between the two

types of pericarpous composition shown in Fig. 5.

The fact that A. decurrens and A. neriifolia , which they

call “ Acacia types ”, belong to the series Bothry-

cephalae and Phyllodinae respectively, stresses the

relationship between the series Vulgares and the two

above-mentioned subgenera. This relationship is

discussed in greater detail in Robbertse (1973) and

two papers still in press.

(a) The parenchymatous and sclereid strata

As already mentioned, the parenchymatous stratum

is heterogeneous in origin. The portion between the

vascular bundles and the abaxial epiderm consists

of 5-10 cell layers and derive principally from the

hypodermal layer of the ovary wall. This portion

differentiates relatively early and does not change

significantly. The cells are larger than those in the

inner portion of the parenchymatous stratum and in

most species it contains a secretion clearly visible in

unstained preparations and very chromatic (Fig. 3D).

In A. tortilis the outer portion of the parenchymatous

stratum is collenchymatous and in A. schweinfurthii

it is replaced by a sclereid zone.

The inner portion of the parenchymatous stratum,

i.e. the portion on the adaxial side (inside) of the

vascular bundles, generally consists of smaller cells

than the outer portion. The vascular bundles are

embedded in the parenchymatous zone and are

bounded on the abaxial side by fibre caps. In species

such as A. nilotica and A. montis-usti, where the

pericarp is succulent when young, only a few isolated

fibres are found on the abaxial side of the widely

spaced vascular bundles. In species with woody

pods, such as A. hebeclada, A. erioloba and A. robust a,

there are large groups of fibres and the vascular

bundles lie close together. In certain species (Figs. 5C,

F, G and J) there are sclereid cells between the fibre

bands of adjacent vascular bundles. An interesting

phenomenon here is that the parenchymatous cells

directly adjacent to the fibre caps of the vascular

bundles can become meristematic in order to give

rise to new cells. Since the fibre caps, on account of

their position, represent pericycle cells and since a

pericycle is potentially meristematic (Esau, 1965), this

phenomenon cannot be considered unusual. Young

A. hebeclada pods have secondary parenchymatous

pericycle cells adjacent to the fibre caps, and these

later become sclereid cells. The meristematic activity

of the pericycle cells is so pronounced in A. erioloba

and A. haematoxylon pods that sclereid cells are

forced from between the vascular bundles to form a

sclereid zone around them (Fig. 5H and I). The latter

can therefore be considered a derived type which may

have originated from the type of structure found in

A. hebeclada and other species (Fig. 5J). In species

without any sclereid cells among the vascular bundles,

(Fig. 5K) no meristematic activity has been noted in

the pericycle.

The position of A. albida (Fig. 5B), where the

sclereid cells are located adaxially with respect to the

vascular bundles, is basically the same as for groups G

and J (Fig. 5). The difference is that the cells of the

outer portion of the parenchymatous stratum in

A. albida are enlarged to form a more or less fleshy

portion. The pericycle cells adjoining the outer

parenchymatous stratum (=cortex) extend anticlinally

to contribute to the fleshiness of this zone and this

extension of the cells displaces the sclereid cells

deeper into the parenchymatous stratum.

In the pods of certain species of the series Vulgares,

the parenchymatous cells directly adjacent to the

fibre zone become meristematic. The secondary cells

formed in this way differentiate into sclereid cells,

resulting in a sclereid layer on the abaxial edge of

the fibre stratum, which may be confined to the median

longitudinal axis of the valves or may reach to the

edges. This is the structure found in group 6 (Fig. 5)

of the series Vulgares. It is never found amongst the

Gummiferae.

Compared with A. ataxacantha pods, A. hereroensis

pods have more longitudinal fibres and even possess

occasional single sclereid cells on top (i.e. at the

abaxial edge) of the fibrous stratum. In A. polyacantha ,

A. caffra and A. galpinii there are also more sclereid

cells on top of the fibre stratums, so that we have here

a sclerenchymatous stratum consisting of fibres and

sclereid cells.

The sclereid cells have conspicuously large nuclei

and are therefore alive. Live sclerenchymatous

elements have also been described by Van Vuuren

(1970) in Adenia. The pods of A. nigrescens, A.

burkei and A. welwitschii (Fig. 5C) have sclereid cells

amongst the adjacent vascular bundles. The entire

portion of the parenchymatous stratum between the

sclereid cells linking the vascular bundles and the

sclereid cells adjacent to the fibre stratum is tem-

porarily meristematic. The secondary parenchymatous

cells formed as a result of this meristematic activity

also have conspicuous nuclei like the sclereid cells of

A. caffra. Although some of the secondary paren-

chymatous cells differentiate into sclereid cells, most

of them retain their thin walls. The notable

meristematic activity of the adaxial portion of the

parenchymatous stratum is similar to a phenomenon

found in the pods of A. erioloba and A. haematoxylon.

The arrangement of tissue types in the pericarp of

A. erubescens and A. mellifera (Fig. 5F) look like a

combination of the arrangements shown in Fig. 5C

and E.

The arrangement shown in Fig. 5Fa is found near

the edges of the valves, whereas Fig. 5Fb shows the

tissue arrangement along the median axis of the

valves. The tissue organization found in A. galpinii

(Fig. 5E) actually lies between types C and F, since

the succession of the tissue layers at the sides of the

valves corresponds with the arrangement shown in

Fig. 5C.

It has already been pointed out that the sclereid

cells found in the pericarp of A. schweinfurthii and

A. kraussiana (Fig. 3C) are derived from hypodermal

cortex tissue. These sclereid cells are therefore

ontogenetically totally different from the other

sclereid cells found amongst vascular bundles or at

the abaxial edge of the fibre stratum. The latter two

kinds of sclereid cells differ because those amongst

the adjacent vascular bundles originate from the

pericycle and those adjoining the fibre stratum from

an inner tissue layer which may be considered homo-

logous with a pith. In the ovary wall, however,

‘pericycle’ and ‘pith’ cells cannot yet be differentiated.

Strictly speaking, the pods of the series Vulgares show

sclereid cells of three different origins whereas the

series Gummiferae possesses only one kind of sclereid

cell — the kind originating from the pericycle.

At the edges of the valves, in most species, the fibre

caps of adjacent vascular bundles merge lengthwise

to form a continuous fibre layer (Fig. 6C). In some

species such as A. xanthophloea, A. kirkii (Fig. 6A)

and A. stuhlmannii, where the pods do not dehisce,

the fibre layer extends well past the vascular bundles,

imparting a characteristic edge structure to the pods.

488

THE GENUS ACACIA IN SOUTH AFRICA. IV. THE MORPHOLOGY OF THE MATURE POD

Next to the fibre layers in the valve edges sclereid

cells occur amongst the vascular bundles, so that the

edge structure of these pods (e.g. A. erubescens and

A. mellifera of the series Vulgares and A. nilotica

and A. xanthophloea of the series Gummiferae)

differs from the portion along the central longitudinal

axis of the valves.

(b) The fibre stratum

The fibre stratum originates directly underneath

the adaxial epidermis but does not extend to the

edge of the pod valves. In most Vulgares species the

fibre stratum extends virtually up to the vascular

bundles in the valve edges (Fig. 6C) whereas in

most Gummiferae species it ends before the edge of

the husk. In A. karroo and certain other species the

fibre band is macroscopically visible as a leathery

layer at the adaxial side of the valves. As already

noted, the fibre stratum may consist of both

longitudinal and latitudinal fibres. The latitudinal

fibres directly adjoin the adaxial epidermis (Fig.

5A-F) and vary from one to approximately four layers.

Longitudinal fibres occur in most South African

Acacia species and are found on the abaxial side of

the latitudinal fibres or, where these are absent as in

the Gummiferae (Fig. 5H, J and K), against the

adaxial epidermis.

The greatest variation in the layers of longitudinal

fibres is found in species of the series Gummifera.

In A. karroo it may consist of up to ten layers of

fibres, whereas in A. haematoxylon and A. stuhlmannii

it is completely absent (Fig. 5G and I). The fibre

stratum in A. erioloba differs from that of the other

species. In this species we find reduced vascular

bundles with sclerenchymatous fibres on the adaxial

side (inside) of the larger, outer vascular bundles.

These reduced vascular bundles appear to be branches

of the outer vascular bundles and where they are

thickly clustered they form a fibre band which,

though interrupted, occupies the same position as the

fibre stratum found in species such as A. karroo

(Fig. 5K).

THE RELATION BETWEEN THE DEHISCENCE OF

PODS AND THE ANATOMY OF THE PERICARP

Pods which do not dehisce are found inter alia in

A. kraussiana, A. schweinfurthii, A. erioloba and A.

haematoxylon. In all these species there is a sclereid

stratum in the abaxial portion of the parenchymatous

stratum, a character found in no other South African

species. A. haematoxylon and A. stuhlmannii were the

only two species examined which had no fibre stratum

and both possess non-dehiscent pods. This seems to

show that the presence of a sclereid stratum in the

abaxial part of the parenchymatous stratum and the

absence of a fibre stratum are correlated with non-

dehiscent pods.

Fahn and Zohary (1955) found that crossed fibres

in the fibre stratum and a differentiating layer at the

edge of the pod are prerequisites for the dehiscence

of pods. In every taxon of the series Vulgares (sensu

Rob.*) examined, crossed fibres occur in the fibre

stratum and the pods dehisce, but in A. kraussiana and

A. schweinfurthii , which have the same characteristic,

the pods do not dehisce. It seems then that the

presence of a hypodermal sclereid stratum (Fig. 3C)

in this case is the dominant factor precluding the

pods from dehiscing.

* Vulgares sensu Benth., excluding A. kraussiana, A. brevis-

pica and A. schweinfurthii.

In species with pods that do dehisce, the articulation

layer is usually well developed (Fig. 6B and C).

However, an articulation layer is usually absent in the

closed pods of A. xanthophloea, A. kirkii (Fig. 6A)

and, to a lesser extent, A. nilotica. The latter three

species belong to group K (Fig. 5) where the remaining

species all have dehiscent pods although all the fibres

are longitudinal.

To sum up, it can be stated that a well-developed

fibre stratum and articulation layer are probably

prerequisites for the dehiscence of Acacia pods and

that the presence of a sclereid stratum in the abaxial

portion of the parenchymatous stratum, the absence

of a fibre stratum and an undeveloped articulation

layer at the edge of the pod may be reasons why a

pod does not dehisce.

THE ORIENTATION OF SEEDS IN THE PODS

In general, the pods of South African Acacia

species have a single, unpartitioned loculus. As

already noted in the discussion of the adaxial epider-

mis, some species exhibit a spongy endocarp which

may fill up the loculus partially or entirely. In the

species where an extensive spongy endocarp is found,

as in A. erioloba and A. haematoxylon, pseudo-

partitions are formed between the seeds, because the

endocarp fills up all the open spaces in the husk not

occupied by the seeds themselves.

In most species the seeds are arranged in a single

row in the pod (Fig. 7A-C), but in A. erioloba

a longitudinal section of the pod may show two or

even three rows of seeds ranged one above the other.

(Fig. 7 D-F). This arrangement basically corresponds

with that of A. cavenia (Mel.) Hook. & Arn. where

two adjacent rows (one row in each valve) is found

with a false septum of spongy tissue between the two

rows of seeds (Fig. 7G).

In species with flattened pods, the seeds are also

flattened and can be found in three different positions.

In most of the Vulgares species (sensu Rob.) the hilum

is aligned diagonally backwards (Fig. 7B), whereas

in most of the Gummiferae species (sensu Rob.)

with flattened pods the hilum points in the direction

of the stylar end of the pod (Fig. 7A). The seeds in the

pods of A. schweinfurthii, A. kraussiana and A. albida

have the hilum pointing towards the placenta. In

cases where the pods are thicker and more or less

rounded, as in A. haematoxylon, A. stuhlmannii and

A. erioloba, each seed (as against the location of

seeds in Fig. 7A-C) is rotated against its own

longituginal axis through an angle of 90° so that the

raphe-antiraphe (Robbertse 1973) is directed vertically

to the valves (Fig. 7D and E). In A. erioloba the

hilum of the seeds in the same pod can be oriented

in different directions since the developing seeds crowd

and displace one another.

It is difficult to assess the taxonomic value of the

seed orientation in the pod, since the variation in

South African species is too minute to warrant a

conclusion. Brenan (1957) used this character to

differentiate between A. robusta and A. clavigera,

but the local material examined shows a transitional

phase between the two orientations. It will be

necessary, therefore, to examine species from all the

genera of the subfamily Mimosoideae to determine

whether the orientation of seeds in the pod is

taxonomically significant or not.

ACKNOWLEDGEMENTS

This work was supported by the Research and

Publication Committee of the University of Pretoria

and the C.S.I.R.

P. J. ROBBERTSE

489

Fig. 7. — Diagram of seed arrangement in the pods of certain

Acacia species. A, Gummiferae; B. Vulgares; C, A. albida ,

A. schweinfurthii, A. kraussiana and A. brevispica; D-F,

A. erioloba ; G, A. cavenia.

UITTREKSEL

Die uitwendige morfologie en anatomie van die

peule van al die Suid-Afrikaanse Acacia -soorte word

bespreek. Dit is gevind dat die Suid-Afrikaanse Acacia-

soorte op grond van die anatomie van die peul in 'n

aantal groepe verde el kan word. Daar is opvallende

verskille tussen die anatomie van die peule van die

series Vulgares en Gummiferae.

REFERENCES

Baker, E. G., 1930. Leguminosae of Tropica l Africa, 3. Ostend:

Erasmus Press.

Bentham, G., 1875. Revision of the suborder Mimoseae.

Trans. Linn. Soc. Lond. 30: 335-664.

Brenan, J. P. M., 1957. Notes on Mimosoideae: IV. Kew Bull.

1957: 357-372.

Brenan, J. P. M., 1959. Leguminosae, subfam. Mimosoideae.

In C. E. Hubbard & E. Milne-Redhead, Flora of Tropical

East Africa. London: Crown Agents.

Burtt davy, J., 1932. A manual of the flowering plants and

ferns of the Transvaal with Swaziland, South Africa. London :

Longmans, Green.

Esau, K., 1965. Plant anatomy. New York: John Wiley.

Fahn, A. & Zohary, M., 1955. On the pericarpial structure of

the legumen, its evolution and relation to dehiscence.

Phytomorphology 1 : 99-1 1 1 .

Fahn, A., 1967. Plant anatomy. London: Pergamon.

Harvey, W. H., 1862. Leguminosae. In W. H. Harvey & O. W.

Sonder, Flora Capensis, ,2. Kent: Reeve.

Hildebrand, F., 1873-1874. Die Schleuderfruchte und ihr

im anatomischen Bau begrundeter Mechanismus. Jahrb.

Wiss. Bot. 9: 235-276.

Keay, R. W. J., 1958. Mimosaceae. In J. Hutchinson & J. M.

Dalziel, Flora of West Tropica! Africa, ed. 2, 1,2. London:

Crown Agents.

Kraus, G., 1866. Ober den Bau trockener Pericarpien. Jahrb.

Wiss. Bot. 9:235-276.

Majewsky, P., 1873. Kurtze Notiz. Ober die gewebeartigen

Plazenhaute. Bot. Ztg. 33 : 24.

Oliver, D., 1871. Mimoseae. In D. Oliver, Flora of Tropica!

Africa , 2. London: Reeve.

Robbertse, P. J., 1973. Die genus Acacia in Suid-Afrika — 3

(met spesiale verwysing na die morfologie van die saad).

Tydskr. Natuurwet. 13,2: 72-95.

Ross, J. H., 1972. The Acacia species with glandular glutinous

pods in Southern Africa. Bothalia 10,2: 351-354.

Ross, J. H., 1975. Notes on African Acacia species. Bothalia

1 1 ; 443-447.

Van Vuuren, D. R. J., 1970. ’// Morfologiese studie van die

stingels van die houtagtige verteenwoordigers van die genus

Adenia Forsk. in Suid-Afrika. D.Sc.-verhandeling, Univ.

van Pretoria.

Verdoorn, I. C., 1957. South African species of Acacia with

glandular glutinous pods. Bothalia 6: 153-413.

Von Breitenbach, F., 1965. The indigenous trees of Southern

Africa, 2. Pretoria: Department of Forestry.

Young, R. G. N., 1955. The Acacia species with spicate in-

florescence in the Transvaal. Candollea 15: 79-123.

Bothalia 11, 4: 491-493 (1975)

The typification of Lycium inerme

J. H. ROSS*

ABSTRACT

As the protologue of Plectronia ventosa L., Mant. 1 : 52 (1767), was based on discordant Oliniaceous and

Rubiaceous elements, and as P. ventosa has been typified in an Oliniaceous sense, it is no longer possible to use

the name Canthium ventosum (L.) Kuntze for a Rubiaceous plant. The correct name for the Rubiaceous plant,

previously but incorrectly called C. ventosum, is Canthium inerme (L.f.) Kuntze, with Lycium inerme L.f., Suppl.

150 (1781) as the basionym. The holotype of Lycium inerme is housed in the Thunberg Herbarium in Uppsala.

The protologue of Plectronia ventosa L., Mant.

1: 52 (1767), was based on two discordant elements:

(1) Burm., Rariorum Africanarum Plantarum X: 257,

t. 94 (1739), which is what has until now been called

Canthium ventosum (L.) Kuntze, and (2) a specimen

(No. 277.2) in the Linnaean Herbarium in London.

As the latter is an Olinia, it is clear that the protologue

of P. ventosa was based on a Rubiaceous element and

an Oliniaceous element.

In the International Code of Botanical Nomen-

clature: 339 (1972), the genus Olinia Thunb. is

conserved over Plectronia L. in an Oliniaceous sense,

the specimen in the Linnaean Herbarium referred to

above being regarded as the lectotype of Plectronia and

the genus Plectronia as a synonym of Olinia. The

specimen in the Linnaean Herbarium is not annotated

by Linnaeus but it is annotated “Plectronia ventosa”

by Linnaeus filius. As it is not possible to date the

specimen and therefore establish whether or not

Linnaeus definitely saw it, it seems that one must

assume that Linnaeus could have seen the specimen.

Since it is difficult to be absolutely certain which

of the two elements typifies P. ventosa it seems

advisable to follow the choice of the specimen in the

Linnaean Herbarium already made in the Code, and

there is every reason to believe that this choice is

correct. (It is as well to bear in mind that the specimen

in the Linnaean Herbarium is a flowering specimen,

while both flowers and fruits are depicted in Burman’s

t. 94). Analysis of the protologue of P. ventosa

reveals that there is certain information in it which

Linnaeus could not have obtained from Burman’s

t. 94. For example, Linnaeus mentioned that the

stems were tetragonal: this is not apparent from

Burman’s t. 94 but the stem of the specimen in the

Linnaean Herbarium is clearly tetragonal. The

generic description of Plectronia in Mant. 1 : 6 pro-

vides even better evidence. In the generic description

the perianth is described thus: “ Perianthium mono-

phyllum, turbinatum, obsolete quinquedentatum,

clausum sinubus S, squamis 5 villosis : persistens.” This

reference to 5 scales seems very significant as Linnaeus

could not have obtained this information from

Burman’s t. 94, and, in any case, the plant depicted (a

Canthium ) does not have 5 scales in the mouth of the

perianth. Olinia, on the other hand, does have 5

scales and these are present in flowers of the specimen

in the Linnaean Herbarium. Analysis of the generic

description of Plectronia suggests that it was probably

based very largely on a specimen with the exception of

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria.

details of the fruit and seed and for these Linnaeus

clearly makes reference to Burman:

“Per. Bacca oblonga, bilocularis. Burm.

Sem. Solitaria, oblonga, compressa. Burm.”

De Candolle, Prodr. 4: 475 (1830), typified Plec-

tronia in a Rubiaceous sense but his decision need

not be followed, particularly as it appears that the

decision was in conflict with Linnaeus’s concept.

Plectronia could only now be used in a Rubiaceous

sense if it was proven that the Burman illustration was

actually the type of P. ventosa, and, in view of the

above facts, this seems unlikely.

Cufodontis, in Osterreich Bot. Zeitschr. 107: 106

(1960), published the new combination Olinia ventosa

(L.) Cufod., specimen No. 277.2 in the Linnaean

Herbarium being regarded as the lectotype of the

basionym Plectronia ventosa L. The name Olinia

ventosa must be adopted for the plant that has until

now been called O. cymosa (L.f.) Thunb.

As Plectronia ventosa is typified in an Oliniaceous

sense, it is obvious that the name Canthium ventosum

(L.) Kuntze can no longer be applied to a Rubiaceous

plant. The next available name for the Rubiaceous

plant hitherto but wrongly called C. ventosum is

Canthium inerme (L.f.) Kuntze, with Lycium inerme

L.f., Suppl. 150 (1781) as the basionym.

The protologue of Lycium inerme is as follows:

“ inerme . Lycium inerme, glabrum, foliis oblongis

glabris, floribus aggregatis pedunculatis,

stipulis barbatis. Habitat in Cap. bonae

spei. Thunberg .”

There is no Thunberg specimen named Lycium

inerme in the Linnaean Herbarium in London or in

the Linnaean collections in Stockholm, and the name

Lycium inerme does not appear in the microfiche

index to the Thunberg Herbarium in Uppsala. No

specimen named Canthium inerme or Plectronia

inerme appears either. However, in H.O. Juel, Plantae

Thunbergianae 430 (1918), there is a cross reference

under Lycium inerme to Serissa capensis Thunb.,

Gen. Nov. PI. 9: 131 (1798). There are two specimens

named Serissa capensis in the Thunberg Herbarium

in Uppsala, namely, Nos. 5314 and 5315, and, through

the courtesy of the Director, Institute of Systematic

Botany of the University, Uppsala, these two

specimens were received on loan.

I had hoped to find the name Lycium inerme L.f.

written somewhere on at least one of the Thunberg

specimens named Serissa capensis but unfortunately

this is not the case. However, it is quite clear that there

492

THE TYPIFICATION OF LYC1UM 1NERME

was a name in the bottom right-hand corner of

Thunberg 5314 (reproduced here as Fig. 1) at some

stage but the name was subsequently erased and

Serissa capensis written in its stead. Unfortunately the

name was so carefully erased that it is impossible to

form any idea of what was written there previously.

So, there is the following circumstantial evidence

that one or both of these Thunberg sheets now named

Serissa capensis may have been the one(s) on which

Linnaeus filius based his description of Lycium

inerme :

1. Juel's cross reference under Lycium inerme to

Serissa capensis in his Plantae Thunbergianae 430

(1918) and his citation on page 419 of both L. inerme

and 5. capensis as synonyms of Plectronia ventosa L.

2. The arrangement of the specimens in Thunberg's

Herbarium, i.e. Thunberg 5314 and 5315 follow

consecutively after the specimens of Lycium.

3. The fact that L. inerme has always been accepted

to be Rubiaceous. For example, in Index Kewensis

and C. H. Wright in FI. Cap. 4 (2): 109 (1904).

HtW NtOATIVL

: 4H00

\L.- • ) Kj'fjC

e ' Z-tU /i.,74

. /• ,» £ /v . st

Fig. 1. — Thunberg 5314, one

of the specimens named

Serissa capensis in the

Thunberg Herbarium,

Uppsala.

Analysis of the protologue of Lycium inerme

reveals that it was almost certainly based entirely

on Thunberg 5314. Some of the details in the proto-

logue, for example, “stipilus barbatis”, could not

have been obtained from Thunberg 531 5 as the stipules

in 5315 are not as described, while the “floribus

aggregatis pedunculatis” is unlikely to have been

taken from 5315 as the specimen has only a few

rather inconspicuous young inflorescences in amongst

the leaves. In addition, the petioles in 5315 are clearly

pubescent above and there are “pockets” of hairs on

the lower surface at the points where the main veins

J. H. ROSS

493

depart from the midrib. Thunberg 5314, on the other

hand, agrees well with the protologue in every respect,

although it is perhaps odd that there is no mention in

the protologue of fruits. As Thunberg 5314 clearly

agrees with the protologue of Lycium inerme I am

persuaded to regard this specimen as the holotype.

It remains to add that Thunberg 5314 matches material

of what has until now but wrongly been called

Canthium ventosum. The correct name for this taxon

is therefore Canthium inerme (L.f.) Kuntze, Rev.

Gen. 3: 545 (1898), with Thunberg 5314 (UPS) as

the holotype of the basionym, Lycium inerme. The

specific epithet “inerme” is rather inappropriate as

the plant often has spinescent branchlets.

It is clear that Thunberg 5314 and 5315 belong to

two different taxa and that 5315 is in fact a specimen

of Canthium mundianum Cham. & Schlechtd. The

protologue of Serissa capensis Thunb., Gen. Nov.

PI. 9: 131 (1798), seems to have been based very

largely, if not entirely, on Thunberg 5314 as there is

certain information in it that could not have been

gleaned from 5315. For example, once again the

description of the leaves and stipules matches those of

Thunberg 5314 rather than those of 5315. In the

generic description of Serissa there is reference to

fruits which quite clearly must have been taken from

Thunberg 5314 as 5315 lacks fruits. However,

S. capensis is an illegitimate name because Thunberg

cited the earlier Lycium barbatum Thunb., Prodr. FI.

Cap. 37 (1794) in synonymy, and because it was based

in part at least, if not entirely, on the type specimen

Lycium inerme L.f. This is stressed because if it was

ever argued that 5. capensis was based entirely on

Thunberg 5315 the illegitimacy of the name 5. capensis

would prevent the adoption of the specific epithet

“capensis” for the taxon to which Thunberg 5315

belongs.

'

Bothalia 11, 4: 495-503 (1975)

Euphorbia gummifera , E. gregaria and a new species from

Damaraland

L. C. LEACH*

ABSTRACT

The identity of Euphorbia gummifera Boiss. and E. gregaria Marl, is discussed and a new species,

E. damarana Leach, is described. A key to the species including E. carunculifera Leach from southern Angola

and a map showing their distribution are provided. The possible significance of these species in the phylogeny

of the genus is also discussed.

Some morphological deviations from the norm for

the genus might, at first, seem to justify the exclusion

of Euphorbia gummifera Boiss. E. gregaria Marl, and

E. damarana Leach (described on p. 000) from

Euphorbia. The most important deviations are found

in the relatively large, 3-6-merous, fleshy fruits with

delayed modes of dehiscence and, in E. damarana , a

relatively thick, hard, woody septate endocarp; this

species also lacks the usual partial partitions at the

base of the inside wall of the involucre.

However, some at least of these deviations are to

be found in other quite different and apparently

scarcely related species which are, and it seems

correctly, universally accepted in Euphorbia , for

example, the large, 3-6-locular, fleshy fruits of E.

virosa Willd. and delayed patterns of dehiscence in

such widely divergent species as E. monteiri Hook, f.,

E. lathyris L. and an as yet undescribed geophyte

from Zambia. In view of this and the apparently close

affinities existing between the three species under

discussion and such species as E. hamata Sweet,

E. gariepina Boiss. and E. tirucalli L., I have little or

no hesitation in accepting them as belonging in

Euphorbia as presently constituted.

Of some historical interest in this connection are a

letter signed by Radlkofer and a note by Pax which

are attached to a specimen of E. damarana (Rautanen

208) in the herbarium of Zurich University. In the

former, Radlkofer (replying, it seems probable, to

Pax) makes the point that the 5-locular fruit does not

necessarily exclude Rautanen’s plant from Euphor-

biaceae, but he nevertheless refrains from offering

any suggestion regarding its correct generic placing.

According to Pax’s note the material is excluded

from Euphorbia by its anastomosing laticiferous

ducts. However, Boodle & Fritsch (Solereder's

Systematic Anatomy of the Dicotyledons, 1908)

question the accuracy of Pax’s concept of the laticifers,

and their doubts seem to receive justification in sub-

sequent literature, from which it appears that no

anastomoses are known in the non-articulated

laticiferous cells which occur in Euphorbia.

These three species, together with E. carunculifera

Leach from southern Angola, appear to form a

closely related group which may be of considerable

importance in relation to the phylogeny of the genus.

The dioecious habit and multi-locular fruits (“multi-”

in the context of Euphorbia) of these species are

among characters usually considered to be primitive

and it may be significant that these and similar

* c/o National Herbarium, P.O. Box 8100, Causeway,

Salisbury, Rhodesia.

assumedly primitive traits are by no means uncommon

in the same general geographical area (the Namib

and its environs).

One is prompted to ask whether these characters

have in fact survived because of suitable environmental

conditions or have developed in response to these

factors, as have geophytes and other life forms;

certainly it seems probable that delayed dehiscence

(not considered to be primitive) may be an adaptation

to prolonged drought periods and appears to be

directly comparable to the arrested development of

the follicles in Stapelieae. It may also be noteworthy

in this connection that several taxa possessing

assumedly primitive characters (e.g. E. virosa , E.

damarana and E. carunculifera ) are often closely

associated with Welwitschia.

Although apparently forming a distinct evolutionary

group, no attempt is made at this stage to fit these

species into the formal nomenclature of the subgeneric

hierarchy, as it is considered that this should be left

in abeyance until a much larger proportion of the

species, their distribution, variation, synonymy etc.

are known in greater detail than at present. Many

species are still to be described, while others are so

incompletely known that they are little more than

names, of which even the identity is, in some instances,

still subject to considerable doubt; whole areas,

particularly in Africa, are virtually unknown

euphorbia-wise and many important aspects (e.g.

palynology), especially in relation to the succulent

species, remain almost entirely unstudied. It is felt

that any formal classificatory decisions reached on

such inadequate information add nothing to our

knowledge and lead almost inevitably to unnecessary

nomenclatural complication and instability.

In large areas of their respective distributions, which

extend from just south of the Orange River to north

of Benguela in Angola, these gregarious plants form

large colonies in which they are not only the dominant

but often almost the only perennial species. The

colonies sometimes extend for many kilometres and

often determine the character of the vegetation over

vast areas.

Yegetatively the four species are very similar in

appearance and from a distance are scarcely to be

distinguished from one another. However, in fruiting

characteristics they are very distinctive and their

distributions are sufficiently disjunct for identifications

made from minor characters usually to be reliably

confirmed by the locality when only sterile material is

available.

26700-9

496

EUPHORBIA GUMM1FERA, E. GREGARIA AND A NEW SPECIES FROM DAMARALAND

Fig. 1. — Approximate areas of distribution of Euphorbia

carunculifera, (1), E. damarana (2), E. gummifera (3) and

E. gregaria (4).

A key to the species, based ( 1 ) on fruiting characters and (2) on vegetative characters follows:

Gregarious, dioecious, unarmed shrubs, 1-3 m high with distinctly ribbed or terete, rod-like branches and

branchlets.

(1) Key to species based on fruiting characters

Capsules very shortly exserted from the involucre:

Capsules obtusely 3-4-lobed, truncate at the base, 10-12 mm diam. Seed ecarunculate, ±5,5 mm long

and broad, ±flat on the ventral side E. gummifera

Capsules subglobose, 4-6-locular, ±20 mm diam. Seeds ecarunculate, ±8 mm long x 6 mm broad

somewhat angular on the ventral side from being compressed on each side of the suture. .E. damarana

Capsules far exserted from the involucre:

Capsules subglobose 3-5-locular, ±20 mm diam.; pedicel ±15 mm long abruptly decurved. Seeds ecarun-

culate, ±6,5 mm long x3,5 mm broad, usually somewhat angular on the ventral side..£. gregaria

Capsules ellipsoid or ovoid, obtusely 3-4-lobed ±10 mm long x 8 mm diam.; pedicel 7-9 mm long,

erect and straight. Seeds carunculate, obtusely 4-angled, ±4 mm long x 3mm diam E. carunculifera

(2) Key to species based on vegetative characters

Branches relatively freely rebranched, hard and woody, usually somewhat curved, decurrently 3-ribbed

from the slightly prominent but insignificant leaf scar; leaves minute, ovate, recurved, tomentulose.

Shrubs averaging 1 m high; winter rainfall areas of the southern Namib, usually less than 80 km

from the coast E. gummifera

Branches rather sparingly rebranched above, terete (not ribbed), tough and fibrous rather than woody:

Leaf scar and inflorescence bracts provided with a pair of dark glands at their base; leaves narrowly

elliptic, erect (neither recurved nor channelled), almost glabrous. Shrubs averaging 2,25 m high;

Namib and adjacent areas of the southern Kaokoveld E. damarana

Leaf scar and inflorescence bracts without dark glands;

Branchlets alternately or randomly produced; leaves narrowly obovate, strongly recurved, channelled,

densely tomentose. Shrubs averaging 1,5 m high; Namaqualand mainly along the Orange and

Fish rivers, usually more than 100 km from the coast E. gregaria

Branchlets and branches often dichotomously forked on each side of an old terminal inflorescence.

Shrubs averaging 1,75 m high; southern Angola, semi-desert between Mo^amedes and

Lucira E. carunculifera

L. C. LEACH

497

Difficulty may sometimes be experienced in the

identification of leafless sterile specimens when

branches become shrivelled in the course of drying,

as a pattern of decurrent wrinkles, somewhat similar

to that of the ribs in E. gummifera, sometimes appears,

especially in specimens of E. gregcria. It has, unfor-

tunately, not been possible to find a “key” character

to cover such instances.

Euphorbia gummifera Boiss ., Cent. Euphorb.: 26

(1860); in DC. Prodr. 15, 2: 97 (1862); Marloth in

Trans. Roy. Soc. S. Afr. 1: 316 (1910); N.E. Br. in

Thistleton-Dyer, Flora Cap. 5, 2: 295 (1915); Dinter

in Feddes Repert. 17: 264 (1920-21); Range in

Feddes Repert. 36: 254(1934); White, Dyer & Sloane,

Succ. Euphorb. 1: 130 (1941); Jacobsen, Elandb.

Succ. PI. 1: 437 (1960); P. G. Meyer in Merxm.

Prodr. FI. S.W. Afr. 67: 26 (1967), p.p. excl. specirn.

De Winter & Leistner 5709 et Merxmiiller & Giess

1465; Jacobsen, Sukk. Lexikon: 191 (1970); Giess in

Dinteria 4: 60 (1971). Type: S. Africa, N. Cape.

Drege 2944 (P! S!).

Euphorbia sessiliflora E.Mey., Drege Zwei PIT Docum. —

184 (1843) nom. nud., non Roxb. (1814).

Type: as above

Plant an unarmed, probably unisexual shrub, up

to ±1,5 (1,8) m high x 2 m diam., freely branched

from the base, relatively freely alternately branched

and rebranched above. Branches and branchlets hard

and woody with a succulent grey-green bark, usually

shortly decurrent at their junctions with the main

branches but seldom articulated, characteristically

decurrently 3-ribbed from the slightly prominent but

minute leaf-scar, ultimate branchlets ±7-10 mm

diam. Leaves minute, ovate, subacute, recurved,

minutely tomentulose. Inflorescence sessile, terminal

or lateral from near the apex of the branchlets, cyathia

clustered; bracts scale-like red-brown, ovate, acute,

±1,5 cm long x 1 mm wide. Involucre red, minutely

whitish tomentose, ±2,5-3 mm long x 4-5 mm

diam. including the glands, more or less obconic,

truncate at the base; glands 5-6, dark red, convex,

transversely broadly elliptic or very broadly obovate,

spreading or strongly deflexed, 1,0-1, 5 mm wide x

0,75-1 ,0 mm long; lobes 5-6, erect, broadly obovate

or subquadrate, entire or dentate, densely woolly

tomentose outside, woolly ciliate, glabrous within;

bracteoles filamentose, finely divided and woolly at

the dark reddish apex, 2,5 mm long; ^pedicels

2,0-2, 5 mm long ±0,35 mm diam.; filaments

noticeably thicker than the pedicels, ±0,45 mm diam.,

2 mm long; anther cells and pollen bright yellow.

Capsule more or less broadly ovoid, obtusely 3-4-

lobed, truncate at the base, broadly subacute at the

apex, 10-12 mm diam., 7-10 mm high, sparingly,

minutely white puberulous; barely exserted from the

involucre on a pubescent pedicel 1, 5-2,0 mm long;

perianth distinct but rudimentary (not merging with

the capsule base). Seed more or less broadly ovoid,

slightly concave at the truncate base, broadly sub-

acute at the apex, dorsi-ventrally slightly compressed

or somewhat 3-angular in section from being slightly

compressed on each side of the suture, 5, 5-6,0 mm

long, 5-6 mm broad, 4-5 mm thick, smooth, pale

creamy brown with patches of darker colour and a

brown suture. (Description based partly on Giess,

Volk & Bleissner 5371 and partly on Leach & Canned

15072).

South West Africa. — 2615 (Luderitz): Rots Koppie (-CB),

Range 72 (SAM); Haalenberg (-CB), st. 283.1929, Dinter 6697

(BOL; K; Z); ibid. cult. Kirstenbosch, st. July 1930, NBG

1781/27 (BOL); ibid. fr. 2.viii.l973, Leach & Connell 15072

(K; LISC; M; MO; PRE; SRGH; WIND); ibid. st. Sept. 1927.

Pillans 5934 (BOL); ibid. 3 Sept. 1927, Pillans 5960 (BOL);

Fig. 2. — Euphorbia gummifera. Branch showing charac-

teristic pattern of three ribs decurrent from the leaf scar.

Leach & Cannell 15072.

st. 27. viii. 1919, Pole Evans sub PRE No. H19371 ( PRE)',

Tschaukaib, 3 fl., Nov. 1908, Marloth 4636 (PRE); Namib,

st. 16. ix. 1913 Range 1868 (SAM). 2616 (Aus): Garub(Garup)

(-CA), fr. 83.1910, Dinter 1046 (SAM); 21 .ii. 1909, Pearson

4174S, 4175<2 (K). 2715 (Bogenfels): Pomona (-A), fr. May,

1928, Kolle s.n. (PRE); fr. Jan., 1926, Wagner sub Marloth 6804

(PRE); Namib near Pomona, fl. Nov., 1908, Marloth 4636

(PRE). 2716 (Witputz): Namuskluft (-DD), fr. 21. ii. 1963,

Giess, Volk & Bleissner 5371 (M; WIND).

Cape. — 2816 (Oranjemund) : between Groot Derm & Arris

Drift (-B), fr. Oct., 1926, Pillans 5234 (BOL; K); “Zwischen

Verleptpraam und der Munching des Gariep, unter 1000' ”

Drege 2944 (P; S). 2817 (Vioolsdrif): ± 27 km from Vioolsdrift

towards Stinkfontein (-CD), lvs. ll.v.1969, Werger 392

(PRE; SRGH).

E. gummifera is the least widely distributed of the

trio, being restricted to the winter rainfall areas of the

Namib from Luderitz Bay to just south of the Orange

River, usually less than 80 km from the coast.

The shrubs are the smallest of the group but are

more freely rebranched than those of either of its

relatives; from both of these it differs in its smaller

cyathia, differently shaped, very much smaller,

3-4-lobed capsules and smaller seeds, while its hard,

woody, darker coloured, ribbed branches and

branchlets, which are usually decurrent at their

junctions, are quite different from the terete, more

rod-like branches of E. gregaria and E. damarana.

Plants appear generally to be unisexual but the

remains of an involucre attached to a partially

dehisced capsule ( Giess et a! 5371 ) bears a number of f

pedicels surrounding the $ pedicel; from this it seems

that at least some individuals may be bisexual.

The capsules appear to separate first between the

lobes, while dehiscence appears often to be either

498

EUPHORBIA GUMMIFERA, E. GREGARIA AND A NEW SPECIES FROM DAMARALAND

Fig. 4. — Euphorbia gummifera.

Shrub ± 1 m high, showing

freely rebranched rather

twiggy habit. Leach & Con-

nell 15072.

very retarded or perhaps sometimes incomplete,

judging from the detached (or ready to fall) capsules

found at the Haalenberg locality. The nauseous odour

mentioned by Marloth was not noticed and it is

thought that this may prove to be of a seasonal

nature.

A specimen with prominent leaf bases, originally

collected at Haalenberg, S.W. Africa, cultivated at

Kirstenbosch, NBG 1781/27, “young branches pubes-

cent”. shows the relationship with E. hamata Sweet

and E. gariepina Boiss. in its atypical attenuated

twigs, even more clearly than in normal wild

specimens.

Euphorbia gregaria Marloth in Trans. Roy. Soc.

S. Afr. 2: 36, t. 1, fig. 7 (1910); N.E.Br. in Thistleton

Dyer, Flora Cap. 5, 2: 295 (1915); in Ann. Bolus

Herb. 1: 102 (1915), “ E . gregarica ”; Dinter in

Feddes Repert. 17: 264 (1920-21), p.p. excl. specim.

Dinter 2823; Marloth, FI. S. Afr. 2 (2): 139 (1925);

Range in Feddes Repert. 36: 254 (1934); White,

Dyer & Sloane, Succ. Euphorb. I: 133 (1941);

Jacobsen, Handb. Succ. PI. 1: 436 (1960), p.p. excl.

distrib. “between Jakalswater and Usakos”; P. G.

Meyer in Merxm., Prodr. FI. S.W. Afr. 67: 26 (1967),

p.p. excl. distrib. “Swakopmund Distr.”; Jacobsen,

Sukk. Lexikon: 190 (1970). Type, S.W. Africa,

Marloth 4683 (B|; PRE!).

“Aggenys Euphorbia”, Pearson in Ann. S. Afr. Mus. 9:

9(1911).

Plant an unarmed, unisexual, rounded shrub

averaging about 1,5 m in height, densely branched

from the base and rather less densely, alternately

branched above, with the ultimate branches and

branchlets 6-12 mm in diam. Branches ascending,

erect, terete, neither ribbed nor striate in the live

state, fibrous and tough, with a succulent, greyish

green bark, usually somewhat tawny-tomentose

towards the apex, especially on younger parts.

Leaves caducous, sessile, narrowly obovate, subacute,

rather fleshy, densely tomentose, strongly recurved,

usually somewhat channelled, ±7 mmx2,5 mm.

Inflorescence a densely tawny-tomentose cluster of

very shortly pedunculate cymes, terminally or laterally

L. C. LEACH

499

produced on or towards the apex of the branches and

branchlets; bracts broadly ovate, rather fleshy;

peduncles and cyme branches stout, up to ±4 mm

long (usually less), all densely tomentose. Male

cyathia usually crowded, deciduous; involucre

glabrous inside, cupshaped, ±3 mm deep, 6-6,5 mm

diam. including the glands; glands 5, distant, fleshy,

spreading, sometimes slightly deflexed, more or less

semicircular, lightly rugulose on the glabrous upper

surface, 1, 5-2,0 mm wide; lobes 5, erect, more or

less subquadrate, irregularly truncate or coarsely

dentate, ±1,5 mm wide. Male flowers ±60, arranged

in 5 fascicles opposite the lobes and closely grouped

around an aborted female flower borne on a short

pentagonal pedicel; fascicles each subtended on the

inside by a pair of more or less obovate cuneate,

relatively fleshy, tomentulose woolly bracts about

4 mm long; bracteoles numerous, mostly filamentose

or a few membranous, more or less cuneate, ±3 mm

long, and rather woolly at the more or less truncate

apex; pedicels glabrous ±3 mm long; filaments

±1 mm long and with the anthers glabrous. Female

involucre densely tawny-tomentose, narrowly obconic,

truncate at the base, ±6 mm diam. including the

glands, ±5 mm long, otherwise more or less as in

the male. Ovary densely tomentose, more or less

ellipsoid or ovoid, very soon exserted from the

involucre on a stout pedicel; ovule suspended under

the cross section varying somewhat, dependent on the

number of locules, ±7 mm long, 6-6,5 mm wide,

5-6 mm thick, brownish cream or creamy brown with

darker patches and markings and a dark brown

suture. (Description based mainly on Leach & Cannell

14049, with seed from Pearson 8719).

South West Africa. — 2517 (Gibeon): UitkomsFarm (-CD),

fr. 26.ii.1963, Giess, Volk & Bleissner 5536 ( WIND). 2616 (Aus):

Kuibis, st. Nov. 1908, Marloth4683 (K; PRE). 2617 (Bethanie):

Tschaunaup Mission (-DA), fr. 25.ii.1947, Gerstner 6362 (PRE).

2618 (Keetmanshoop): ± 8 km N of Narubis (-DC), fr.

17.iv. 1933, Lang sub Transv. Mas. 31754 (PRE); ± 3 km N of

Narubis (-DC), fl. & fr. 22. ix. 1967, Leach & Cannell 14049

(SRGH). 2716 (Witpiitz): ± 40 km S of Aus (-AB), fr. 24. ii.

1963, Giess, Volk & Bleissner 5460 (M; WIND). 2717 (Chamai-

tes): near Gawachab (-AB), fl. & fr. 8.xi.l970, Leach & Cannell

14715 (BOL; LISC; M; PRE; SRGH). Holoog (-BD), fr.

19. i. 1916, Pearson 9719 (BOL); fl. c? 9 8.xi.l970, Leach &

Cannell 14717 (PRE); Fish River Canyon, “bei Logerplatz”

(-DA), fr. 1952-53, Walter 2328 (M; WIND); 2718 (Grunau):

Wasserfall (-BA), <J fl. 14.xii.1912, Pearson 8084 (BOL; K;

SAM); Krai Kluft Ravine, Gt. Karasburg (-BA), 19. i. 1913,

Pearson 8564 (BOL; K; SAM); near Klein Karasburg, fr.

Sept. 1909, Mar loth 4683 (PRE); Klein Karas, fr. 22.iv.1931,

Ortendahl 147 (S); W slopes of Great Karasberg, 1 2.xii. 1912,

Pearson 8075 (K). 2817 (Vioolsdrif): “Driekoppen Mtn.”

SE of AiAis (-AB), st. Sept. 1931, Pillans 6567 (BOL); ± 16

km N of Vioolsdrift (-DA), dry inflor. 5.viii. 1 973, Leach &

Cannell 15079 (K ; M ; PRE ; SRGH ; WIND). 28 1 9 ( Ariamsvlei) :

near Nakob (-BB), <? fl. 10. iv. 1956, Theron 1949 (PRE);

between Nakop & Ariamsvlei (-BB), <$ fl. 23. ix. 1967, Leach &

Cannell 14051 (BOL; M; S; WIND).

Fig. 5. — Euphorbia gregaria. Left :

inflorescence of 3 plant with

crowded cyathia. Right: $

plant showing long-pedicelled

fruits. Leach & Cannell 14049.

a fleshy, pad-like obturator. Styles 3-5, fleshy,

tomentose on the outside, glabrous on the deeply

channelled rugulose inner face, ±2 mm long, free

nearly to the base, widely spreading, strongly recurved

at the much dilated, bifid apices. Capsule tawny-

tomentose, more or less obovoid, becoming some-

what pointed at the apex when dry, 3-5 locular,

slightly obtusely angular, ±20 mm diam. and almost

as long, very tardily dehiscent, separating first between

the locules, far exserted on a sharply decurved,

densely tomentose pedicel up to 20 mm long x5 mm

diam.; perianth more or less circular or obtusely

pentagonal, ±8 mm diam. Seed more or less broadly

ovoid, flat or slightly concave at the truncate base,

somewhat broadly subacute at the apex with a beak

on the dorsal side of a small depression at the end

of the suture, more or less obtusely 4-angled from

being slightly compressed on each side of the suture

with a slight obtuse ridge down the back, often

somewhat dorsiventrally compressed, then often

almost flat on the ventral side (cf. E. gummifera),

Cape. — 2820 (Kakamas): between Nakop & Naroegas (-AA),

fl. & fr. 23. ix. 1967, Leach & Canned 14053 (SRGH), $ fl.

21. ix. 1970, Leach & Canned 14491 (SRGH); ± 112 km W of

Upington (-AA), fr. Apr. 1933, Lang sub Transv. Mus. 31855

(PRE; SRGH); ± 45 km E of Naroegas (-BC), 3 fl. 19.V.1955,

De Winter 3586 (M; PRE; WIND); Aughrabies (-CB), st.

Esterhuysen 733 (BOL). 2918 (Gamoep): Aggenys (-BB),

fl. 7. i. 1909, Pearson 3338 & 3538 (BOL; K); Wortel (-BB),

fl. 10. i. 1909, Pearson 3339 (BOL; K). 2919 (Pofadder): near

Pella (-AA), st. 20. vi. 1927, Hornemann sub Marloth 13383

(PRE); fr. only, 8,i. 1 909, Pearson 5037 (K).

E. gregaria is widely distributed, mainly along the

Fish and Orange rivers and usually more than

100 km from the coast. Large numbers are to be seen

to the west of Seeheim, often associated with E. virosa

Willd., and frequently forming large colonies in which

it is the dominant species; eastwards its distribution

extends as far as Kakamas, plants being common

on both sides of the Orange River.

It was feared that the large numbers of plants to

be inundated at the site of the Naute Dam to the

south-east of Seeheim might so taint the water

500

EUPHORBIA GUMMIFERA. E. GREGARIA AND A NEW SPECIES FROM DAMARALAND

(toxicity tests had been made at Onderstepoort) as to

render it unsuitable for human or animal con-

sumption. However, further tests carried out by

Dr M. Wolf at the Veterinary Diagnostic Centre,

Windhoek, disclosed that no such effect was evident,

even at 20-25 times the concentration expected. It is

thought that the two related species may also be

similarly innocuous since no irritating effects on the

mucous membranes, such as smarting nostrils, have

been noted despite extensive handling of both dry

and living specimens.

In habit E. gregaria most closely resembles E.

damarana but is rather more freely rebranched above

and of a somewhat smaller stature, while its densely

tawny tomentose inflorescence and leaves, which are

strongly recurved and lack the characteristic dark

glands at their base, are both quite different from

those of the Damaraland plants. When in fruit

E. gregaria is immediately distinguished from either of

its congeners by the long deflexed pedicel on which

its large rather fig-like fruits are borne. In addition

to being far exserted from the involucre these fruits

are very much larger and differently shaped from

those of E. gummifera, from which the inland species

is further distinguished by its terete (not ribbed)

branches and branchlets. Production of fruits and

seeds appears to be very erratic and is thought

probably to be dependent on seasonal conditions.

Euphorbia damarana Leach , sp. nov.

E. gregaria sensu Dinter in Feddes Repert. 17:264(1920-21),

р. p. quoad specirn. Dinter 2823; sensu White, Dyer &

Sloane, Succ. Euphorb. 1: 133 (1941), p.p. quoad fig. 121;

sensu Jacobsen, Handb. Succ. PI. 1: 436 (1960), p.p. quoad

distrib. “between Jakalswater and Usakos". E. gummifera

sensu P. G. Meyer dubitationc in Merxm., Prodr. FI. S.W. Afr.

67: 26 (1967), p.p. quoad specim. De Winter & Leistner 5709

et Merxmiiller & Giess 1465. E. sp. aff. E. gregaria , Giess in

Dinteria 4: 69 (1971).

Euphorbicie gregariae Marl, affinis sed folia anguste

elliptica, fere glabra, haud recurva nec canaliculata;

glandulis nigris stipulaneis basi foliorum bractearum-

que instructa; cyathiis masculis usque ad duplo

grandioribus et pedicellis bracteolisque tholum pro-

minentem formantibus; glandulis transverse oblongis

saepe contiguis; capsula brevissime ex involucro

exserta, loculis 4-6 valde distincta. E. gummiferae

Boiss. etiam affinis sed ab ilia (praeter capsulam

breviter stipitatam) characteribus simillimis atque

aliis statini dignoscenda.

Frutex erectus, rotundatus, inermis, apparenter

unisexualis, usque ad 3 m altus, saepe 6 m diam.

attingens, e basi ramosissimus, superne parce temere

ramosus; ramis ramulisque ultimis 6-12 mm diam.

Rami bacillares, teretes, plus minusve recti, fibrosi et

tenaces cortice succulento, nec costati nec striati,

pallide flavovirens, scabridiusculi, glabri vel raro

ramulis juvenissimis ad apicem flavido-tomentulosis.

Folia caduca, sessilia, anguste elliptica, subacuta,

usque ad 9 mm longa x3,5 mm lata, fere glabra vel

interdum sparse tomentosa, basi in quoque latere

glandula nigra instructa. hiflorescentia terminalis in

ramis ramulisque disposita; pedunculi et rami

cymarum brevi, crassi, glabri vel interdum flavido

tomentosi; bracteae parvae, plus minusve late ovatae,

с. I mm longae, porphyreae, saepissime glandulis

nigris basi instructae. Cyathia mascula decidua;

involucrum flavido-tomentosum, infundibuliforme,

c. 4 mm longum, 6-9 mm diam. glandulis inclusis;

glandulis 5, carnosis, transverse oblongo-ellipticis,

c. 3 4 mm 2 mm, saepe contiguis plerumque valde

dcflexis, flavis vel fuscis; margine plerumque leviter

crispato; lobis 5 tomentosis, erectis, transverse

oblongis, irregulariter truncatis vel grosse dentatis,

c. 2 mm latis; flores masculi numerosi 5-fasciculati et

cum bracteolis filiformibus pilosis tholum prominen-

tem formantes; fasciculi bracteolis partim connatis

intra subtenti sed involucri septis carentibus; flos

femineus vestigialis. Inflorescentia feminea similis sed

pedunculis paucioribus vel cyathiis interdum sub-

sessilibus, raro plus quam 3 capsulis maturescentibus;

involucrum flavido-tomentosum, plus minusve urceo-

latum, c. 6 mm longum x5,5-6 mm diam. glandulis

inclusis; glandulis 5, carnosis, transverse oblongo-

ellipticis, 2-2,5 mm latis, c. 1,5 mm longis, patulis

vel deflexis, plerumque separatis, fuscis vel forsan

interdum flavis; lobis 5, illis cyathii masculi similibus

sed extus intusque densius tomentosis; bracteolis

filiformibus vel aliquanto spathulatis brevius fim-

briato-laciniatis, c. 2 mm longis. Ovarium ovoideum,

flavido-tomentosum, mox ex involucro partim

exsertum; stylis crassis, c. 1,5 mm longis, libris fere

ad basim, patulo-recurvis, profunde sulcatis, apice

dilatatis biff dis rugosis. Capsula loculis 4-6, sub-

sphaerica, levissime angulata et apice in sinubus

brevissime leviter sulcata, flavo-virens demum flavis,

tomentulosa, c. 20 mm diam., 16 mm alta, endocarpio

lignoso aliquanto crasso, breviter ex involucro

pedicello tomentoso, c. 4 mm longo exserta; perianthio

rudimentali vel carenti. Semen plus minusve ovoideum,

basi truncatum, apice subacutum parum oblique trun-

catum, in quoque latere suterae compressum, c. 8 mm

longum, 6-6,5 mm latum, 5,5 mm crassum, laeve

brunneo-cremeum vel pallide brunneum saepe fuli-

gineo-maculatum.

Type: S.W. Africa, Damaraland, Leach & Canned

15064 A $ (K; LISC; M; PRE, holo.; SRGH;

WIND).

Plant an erect, unarmed, apparently unisexual,

rounded shrub up to 3 m high and often as much as

6 m in diam., densely branched from the base, rather

sparingly and randomly rebranched above, with the

ultimate branches and branchlets 6-12 mm in diam.

Branches more or less straight and rod-like, terete,

neither ribbed nor striate, fibrous, and tough with a

succulent bark (occasionally grazed, a short tuft of

fibres then being left at the apex), pale yellowish

green, often somewhat whitish from the resinous

deposit on the surface, very slightly rough to the touch,

glabrous or rarely slightly yellowish tomentulose on

very young apical new growth. Leaves apparently

fleeting, sessile, narrowly elliptic, subacute, up to 9 mm

long x3.5 mm wide (in cult, up to 22 mm x8 mm),

almost glabrous or sometimes with a sparse tomentum,

erect or spreading (when erect with a peculiar incurved

set to the base), with a dark gland (presumably

stipular in origin) on each side at the base. Inflores-

cence a cluster of very shortly pedunculate cymes,

terminal on the branches and branchlets or

occasionally on short lateral flowering spurs; bracts

scale-like, more or less broadly ovate, about 1 mm

long, red-brown, often with a dark gland on each

side at their base; peduncles and cyme branches short

( ±2 mm) and stout, glabrous or sometimes yellowish

tomentose. Male cyathia deciduous; involucre

yellowish tomentose, shallowly funnel-shaped, ±4 mm

long with the basal portion solid and somewhat

stalk-like, 6-9 mm diam., including the glands;

glands 5, fleshy, transversely oblong-elliptic, usually

slightly crisped on the margins, about 3-4 mm x2 mm,

contiguous or separate, usually strongly deflexed so

that the upper surface is turned outward; lobes 5,

tomentose. transversely oblong, ±2 mm wide, erect,

irregularly truncate or coarsely toothed at the rather

woolly apex; male flowers numerous, intermingled

with a dense mass of woolly bracteoles, which with

the pedicels eventually form a prominent woolly

L. C. LEACH

501

Fig. 6. — Euphorbia damarana.

Plants at the type locality,

± 64 km west of Khorixas

(Welwitschia), ± 2 m high.

Fig. 7. — Euphoroia damarana. Erect, almost glabrous

leaves with dark stipular glands.

dome, arranged in 5 fascicles, not separated by

partitions or pockets on the inside of the involucral

wall but each subtended on the inside by a bract

formed from partially united bracteoles; the fascicles

are closely grouped around what appear to be the

vestiges of an undeveloped female flower. Female

inflorescence similar to the male but with fewer, more

frequently tomentose peduncles and cyme branches;

it is seldom that more than one or two capsules reach

maturity (4 seen only once); involucre yellowish

tomentose, somewhat urceolate, ±6 mm long, 6 mm

diam. including the glands, ±3,5 mm diam. at the

base of the 3,5 mm long, solid basal portion (this

solid portion, similar to that in the male involucre,

is here treated as being part of the involucre in view

of the supporting bracts at its base; however, it

could perhaps be considered as being integral with

the cyme branch as in some instances there are traces

of a junction with the base of the involucral cup);

glands 5, fleshy, dark coloured or perhaps sometimes

yellow, transversely oblong-elliptic, 2-2,5 mm wide,

±1,5 mm long, spreading to strongly deflexed,

slightly crenulate and crisped on the margins, usually

separate; lobes 5, more or less as in the male cyathium

but densely tomentose inside and out; bracteoles

filiform or somewhat spathulate, densely woolly at the

apex, ±2 mm long, closely arranged around the

ovary. Ovary ovoid, yellowish tomentose, very soon

partially exserted from the involucre. Styles, equal in

number to the locules of the ovary, very stout,

±1,5 mm long, free almost to the base, spreading

recurved, deeply grooved down the inner face, much

dilated at the divergently bifid, rugose apices. Capsule

subspherical, slightly angular, yellowish green, tomen-

tulose, more densely so towards the base, averaging

20 mm diam. x 16 mm high, 4-6-locular, with a

relatively thick, hard, woody, 4-6 septate endocarp,

very tardily dehiscent, perhaps sometimes incompletely

so, opening first between the locules with the styles

persistent, very shortly exserted from the involucre

on a stout, tomentose pedicel about 4 mm long;

perianth rudimentary or lacking. Seed more or less

oblong-ovoid, with a central depression at the

truncate base, slightly obliquely truncate at the

broadly subacute apex, somewhat 3-angled in cross-

section from being compressed on each side of the

suture, up to 8 mm long, 6,5 mm wide x 5, 5 mm

thick but varying somewhat depending on the number

of seeds in the capsule and the degree of compression

to which they were consequently subjected, smooth,

brownish cream to pale brown with a dark brown

suture and blackish brown blotches on the flattened

areas on each side of the suture, around the apex and

often at the base, with a whitish area around the

hilum.

Fig. 8. — Euphorbia damarana. 1, hard woody endo-

carp with two locules removed; 2, fleshy outer

covering of fruit partially removed to show woody

interior.

502

EUPHORBIA GUMMIFERA, E. GREG ARIA AND A NEW SPECIES FROM DAMARALAND

Fig. 10. — Fruiting branches of Euphorbia damarana.

Leach & Canned 15064.

South West Africa. — 1812(Sanitatas): 56 km W of Otjihu

on Orupembe road (-BB), fr. 4.V.1957, De Winter & Leistner 5709

(M; PRE; WIND); Otjikongo (-BD), fl. 294.1958, Merxmuller

& Giess 1465 (M). 2013 (Unjab Mouth); 1 155 km W of

Khorixas (Welwitschia) (-B), fr. l.iv. 1963, Hardy & De Winter

1471 (PRE). 2014 (Welwitschia): ' 64 km W of Khorixas

(Welwitschia) FAD), fl. & fr. 27.vii.1973, Leach & Canned

I5064A . B<J (K; LISC; M; PRE; SRC.H; WIND); Sorris

Sorris ( DD), st. I l.iv. 1955, De Winter 3148 (PRE; WIND);

S of Ugab River, on Uis road (-DD), fl. & fr. 29.vii.1973,

Leach A Cannell 15068 (BOL; K; MO; PRE; Z). 2114 (Uis): S

of Ugab River on Sorris Sorris road (-BB), fl. & fr. 29.vii.1973,

Leach A Canned 15067 (BM; M; NBG; SRGH; WIND);

near Brandberg, fr. 7.xii.1947, Rodin 2752 (BOL; PRE). 2115

(Karibib): Klein Spitzkoppe (-CC), <Jfl. 1 S.iii. 1 963, Giess,

Folk A Blcssner 5811 (M). 2215 (Trekkopje): Aukas ( AB)

: fl. _20.viii.1919, Pole Evans sub PRE H 19308 <4 H 19309,

(PRE); Trckkoppie (-AC), st. Sept. 1920, Reuning sub Marloth

9920. (PRE); Jakalswater (-CB), 3 fl. Apr. 1913, Dinter 2823

(SAM); Tinkas Flats, <5 fl. 8.ii.l970, Jensen 473 (PRE); along the

railway line between Usakos and Swakopmund, fr. 22. vi. 1942,

Turvey sub PRE 32336 (PRE); between Usab and Ubib, fr.

6.viii. 1 892, Rautenen 208 { Z).

This, the tallest of the three closely related species,

has a distribution restricted almost entirely to desert

areas to the north of the tropic of Capricorn. The

southernmost record is from Tinkas Flats in the

Namib Desert Park at approximately 22° 50' S; this

locality is more than 300 km from the most northerly

known for E. gummifera and almost as distant from

the nearest recorded occurrence of E. gregaria.

The main concentration known to me lies on the

southern side of the Ugab River, where it constitutes

not only the dominant species, but at the time of year

when fruiting material was collected virtually the only

species to be seen over a vast tract of country

extending to the lower slopes of the Brandberg.

At the type locality west of Khorixas (Welwitschia),

colonies are rather smaller but again the species is

dominant where present. Regeneration at both

localities appeared to be very poor, but fruits with

apparently viable seed were produced in large numbers

in July 1973; it would therefore be interesting to learn

the seedling and subsequent juvenile position in the

following few seasons.

E. damarana appears to be most closely related to

E. gregaria but differs, most significantly, in its

shortly pedicelled, often 6-locular fruits with hard

woody endocarp and larger seeds. The almost glabrous

erect leaves of E. damarana are quite different from

those of E. gregaria, in that species heavily tomentose

and strongly recurved, while the dark stipular glands

at the base of the leaves and bracts allow most (even

sterile) specimens to be distinguished from similarly

sterile material from either of its close relatives.

The larger male cyathia with bracteoles and pedicels

forming a prominent woolly dome, and transversely

oblong, often contiguous glands are also characteristic

of the new species. A relationship with E. gummifera

is also evident but here there is much wider divergence

in flowering and fruiting characters, while the ribbed,

woody branches and branchlets of the smaller, more

freely rebranched shrubs of the southerly based

species are quite different from those of E. damarana.

Affinities with E. carunculifera and E. tirucalli, albeit

L. C. LEACH

503

less close, are evidenced by the vegetative characters

and gregarious colony forming habit of the former

and the dark stipular glands and peculiar inflexing

of the base of the erect leaves of the latter.

Ripe fruits of E. damarana , already dry and

beginning to split between the locules when collected,

have shown no further signs of dehiscing, their whole

character, with fleshy outer covering and hard woody

endocarp, being reminiscent of that of the indehiscent

fruits of Elaeophorbia.

ACKNOWLEDGEMENTS

Mr and Mrs lan Canned for their companionship

and assistance in the course of the expedition to the

Kaokoveld and elsewhere in South West Africa.

The Directors of: The Bolus Herbarium, University

of Cape Town (BOL); The Royal Botanic Gardens,

Kew (K); Botanische Staatssammlung, Miinchen

(M); The National Botanic Gardens, Kirstenbosch

(NBG & SAM); Museum National d’Histoire

Naturelle, Paris (P); The Botanical Research Institute,

Pretoria (PRE); Naturhistoriska Riksmuseum, Stock-

holm (S); The National Herbarium, Windhoek,

S.W. Africa (WIND); the Institut fur Systematische

Botanik der Universitat, Zurich (Z), for the loan of

much valuable material.

The Director, Dept, of Nature Conservation and

Tourism, Windhoek for the necessary collecting

permits.

Mr W. Giess, Curator of the National Herbarium,

Windhoek, for the facilities of the herbarium &

other personal assistance.

Dr L. Jaklitsch of Salisbury, for translations of the

letters and notes attached to the Rautenen specimen

of E. damarana.

Dr B. de Winter, Director, Botanical Research

Institute, Pretoria (PRE), for the facilities of the

herbarium and library and much appreciated assis-

tance in connection with the Kaokoland expedition.

Bothalia II, 4: 505-510 (1975)

Notes on Euphorbia mauritanica , E. gossypina and some related

species with an amplified description of E. berotica

L. C. LEACH*

ABSTRACT

Some aspects of the possible evolutionary development in the Euphorbia mauritanica — lateriflora — gossypina

complex are discussed, and the seed morphology is considered probably to be of considerable significance in

the taxonomy of the group. The relevance of the recently discovered relic population of E. gossypina is also

considered, and a detailed description of this species, based on the Rhodesian material, is provided.

The identity of the little known E. berotica N.E. Br. from Angola is established and an amplified

description drawn up. Attention is also given to the confusion concerning the type locality and distribution of

E. mauritanica L. var. lignosa White, Dyer & Sloane.

These notes, although relating mainly to the

problems of identification and classification of the

complex containing Euphorbia mauritanica L., E.

lateriflora Schum. & Thonn., E. gossypina Pax, etc.,

deal also with the possible evolution of the group.

Presently included in Sect. Tirucalli , it seems to me

that the complex may well represent part of a separate

phylogenetic line. However, it is considered that

clarification of the relationships and distribution of

the genus as a whole is needed before such speculations

should be accorded formal nomenclatural recognition

in the infrageneric classification which, in the writer’s

opinion, already suffers from a number of doubtfully

justifiable “sections”.

The discovery, near Mangula in Rhodesia, of a

colony of plants which clearly belongs to the complex,

poses some interesting questions since no member

of the group was previously known from the Flora

Zambesiaca area nor, in fact, from any locality within

a considerable distance from its boundaries.

It seems very probable that this disjunct, isolated

population is indicative of a more “blanket” distri-

bution of the complex in the past, and perhaps

provides concrete evidence of the link between its

southern, western and north-eastern relatives. Such

a connection need not necessarily have existed during

an arid period nor via an “arid corridor” (as is

sometimes suggested in discussions of the similarities

between the floras of the arid areas of south-western

and north-eastern Africa), as the species concerned

(E. gossypina ) is not only tolerant of, but appears to

thrive under “pluvial” conditions. In fact, it may well

be that the Mangula population has been enabled

to survive arid periods only because of the

exceptionally high water availability obtaining in its

particular habitat (Jacobsen, 1973). A specimen of

this species, growing in rather heavy red soil and in

shade in my garden near Salisbury, has received

between 1 500 mm and 1 600 mm of rain this 1973-74

season (almost double the average for Mangula, and

considerably above the maximum recorded for that

area). This plant is thriving under these conditions

whereas, during the semi-drought of the previous

season, most specimens of the species then in culti-

vation were decidedly unhappy and some died.

When first investigating plants of this group

occurring in southern Angola it seemed that they

belonged in the affinity of E. mauritanica , and that

they were possibly identical with var. lignosa White,

Dyer & Sloane. However, although seemingly less

likely from a geographical viewpoint, it is apparent

*c/o National Herbarium, P.O. Box 8100, Causeway,

Salisbury, Rhodesia.

from the morphology of the seeds (Fig. 1), that the

Angolan plants may be rather more closely related

to the Rhodesian plants than to the southern E.

mauritanica complex, and that related species of the

north-eastern and south-western arid areas may well

be vicariants which have arisen under the high rate of

speciation which seems usually to be associated with

desert conditions.

Fig. 1. — Comparison of seeds in Euphorbia mauritanica —

gossypina Complex, a, E. mauritanica ( Leach & Bayliss

15095, eastern Cape); b, E. berotica ( Leach & Cannell

14026, southern Angola); c, E. gossypina (Leach et at.

14236, Mangula, Rhodesia). Photo: by courtesy of Ministry

of Information, Rhodesia.

506

NOTES ON EUPHORBIA MAURITANIA, E. GOSSYPINA AND SOME RELATED SPECIES

The relic population of E. gossypina now surviving

at Mangula would have been included in the closed

forest area envisaged (Wild, 1968) under conditions

of 150% of present day rainfall (conditions apparently

eminently suitable for the species concerned) thus

providing a direct connection with that area of west

Africa where the closely related E. lateriflora is

endemic today. Unfortunately I have not seen seeds

of this latter species, but from the evidence offered by

a plant in cultivation, as well as by flowering material

in liquid, it seems probable that it is distinct from the

previously mentioned taxa.

Also included in this grouping are E. schimperi

Presl, E. bottae Boiss., E. nubica N.E. Br., E. con-

sobrina N.E. Br., and E. merkeri N.E. Br.; other

possible relationships, such as that with E. cameronii

N.E. Br. require further investigation. Characteristic

of these succulent and subsucculent glabrous shrubs

are: slender, terete, exstipulate branches with, as far

as is known, a thin, more or less odourless, apparently

innocuous latex; alternate caducous leaves; a brac-

teate umbellate inflorescence, with cyathia with entire

glands, emarginate or bidentate ciliate lobes and an

exserted capsule bearing carunculate seeds.

Relevant herbarium material examined but not

cited elsewhere in this paper includes, in addition to

numerous examples of E. mauritanica complex from

southern Africa: Arabia: Schimper 468 (K). Ethiopia

(Eritrea): Schweinfurth & Riva 1083 (K). Sudan:

Schweinfarth 207 (K); 265 (K). “Nubia” Th. Bent s.n.

(K). Somalia: Gillett 4521 (K); Peck 321 (PRE).

Ghana: Chipp 188 (K); Newton 1195 (SRGH).

Nigeria: Barter 3309 (K); Gunn s.n. 31065 in PRE

(PRE).

The taxonomy of the widely scattered populations

involved appears to be as difficult to unravel as that

of the closely related E. mauritanica Complex, and

seems likely to remain a problem until adequate

material becomes available. Although habit is

generally distinctive, identification of the mostly

incomplete herbarium material is usually difficult

and sometimes well nigh impossible on any basis

other than the locality of origin. However, from the

comparisons now made, it seems that seed characters

may be diagnostic at specific level and may con-

sequently prove to be of considerable assistance in

the classification of this taxonomically most difficult

group.

Euphorbia gossypina Pax in Bot. Jahrb. 19: 119

( 1895); N. E. Brown in FI. Trop. Afr. 6, 1 : 553 (1911);

Brenan, Check-list Tang. Terr. 2: 211 (1949). Type:

Tanzania, Tl, Mwanza District, Kagehi, Fischer 514.

E. implexa Stapf in Kew Bull. 1908: 408 (1908). Type:

Uganda, U4, Mengo District, Mawokota, E. Brown 414.

E. sp„ W. B. G. Jacobsen in Kirkia 9,1: 164, 193 (1973).

Plant a laticiferous, extipulate, unarmed, glabrous,

succulent shrub, about I m high when self-supporting,

but up to 2,5 m cr more when semi-scandent on

shrubs and trees, much branched from the base and

above; branches at first erect, soon arching over if not

supported (sometimes becoming pendent over cliffs)

with flowering branchlets more or less erect; latex

rather thin but copious, odourless and apparently

innocuous. Branches terete, alternately or randomly

produced, marked with alternate leaf-scars: succulent,

becoming rather woody with age, green; main

branches up to 2 cm diam. at ihe base, flowering

branchlets usually 6mm diam. tapering to

3 4 mm at the apex. Leaves alternate, caducous,

sessile, often with an axillary bud, strongly deflexed.

glabrous, green, fleshy, entire, usually concave above,

obtusely convex or somewhat keeled beneath, linear-

elliptic tapering into the base, or narrowly ovate to

ovate, acute, obtuse or truncate at the base, those

towards the base of the branches distant, up to 6 cm

long x 6,5 mm, becoming shorter and less distant

above; those of the ultimate branchlets shorter and

proportionally wider, becoming more so and more

crowded towards the apex of the branchlets, then

±8- 10 mm long x 4,5 mm wide; leaf scar transverse,

usually more or less straight (seldom lunate), slightly

prominent, particularly on flowering branchlets, with

a slight subtuberculate swelling just below, brownish,

becoming corky and eventually blackish and con-

spicuous towards the base of the branchlets.

Inflorescence a glabrous, terminal cymose-umbel, with

a whorl of 4-7 (usually 5, up to 9 seen) spreading

rays (peduncles) surrounding the initial cyathium,

each subtended by a bract at its base and bearing a

pair of bracts and a single bisexual cyathium at its

apex. Bracts of the whorl glabous, sessile, strongly

deflexed, usually deciduous in fruit, ovate, acute or

subapiculate, obtusely truncate at the base, up to

16 mm long and 8 mm wide, entire or occasionally

with 1-3 relatively large, distant teeth, green, often

somewhat yellowish towards the base and occasionally

minutely red flecked. Initial cyathium male deciduous

or sometimes bisexual, supported on a short, stout,

prominently ribbed, pale green stipe; involucre

yellowish, with 5-8 glands or occasionally eglandulose;

glands rather variable as to shape and size, mostly

more or less obovate, up to 4 mm long, 2,5-3 mm

wide, entire or irregularly crenate or emarginate,

spreading, only occasionally deflexed, often concave,

yellow becoming deep orange; lobes 5-8, erect,

obtuse, emarginate or bidentate, ciliate with long,

slightly crisped, silky white hairs, otherwise glabrous,

greenish yellow. Rays pale green, glabrous, variable

in length (up to 25 mm), usually ± 2 mm diam.;

bracts paired, usually deciduous in fruit, widely

spreading, broadly elliptic or broadly ovate to very

broadly ovate, up to 8 mm long and 7 mm wide

(sometimes wider than long), acute or subapiculate,

often shortly canaliculate at the apex, slightly keeled

beneath, entire or occasionally sparingly toothed as

those of the whorl; cyathium terminal, very shortly

stipitate (perceptible only in fruit); involucre glabrous,

pale green, cup-shaped, ± 4 mm deep, ± 6,5 mm

diam. including the glands; glands 4, separate,

spreading to deflexed, obovate or subcircular or

occasionally transversely elliptic, 2-2.5 mm long,

2-3 mm wide, entire or occasionally crenulate or

emarginate, usually convex from being recurved

on the margins, pale green becoming dull yellow on the

minutely and shallowly pitted upper surface; lobes 5,

pale green erect, ovate or subquadrate, obtuse,

±1,25 mm x 1 ,25 mm, irregularly denticulate, emar-

ginate or bidentate, conspicuously ciliate with long,

slightly crisped, silky, white hairs. Male flowers ± 35,

arranged in 5 fascicles with numerous filamentose

bracteoles, densely white hairy at the apex; pediclse

± 3 mm long; filaments pale green, 1-1,5 mm long;

anther cells pale buff, ±0,5 mm diam., pollen yellow.

Ovary pale green, far exserted to one side (over the

gap caused by the absence of the fifth gland) on a

glabrous, pale green pedicel, slightly expanded at its

apex into a rudimentary rim-like perianth. Styles

±2,5 mm long, free almost to the base, spreading,

bifid, with the lobes diverging recurved, slightly

enlarged, obtuse and lightly rugulose at the apex.

Ovule suspended under an entire, fleshy obturator.

Capsule obtusely 3-lobed, 7-8 mm diam., ±5,5 mm

high, eventually held erect on a pedicel 6-7,5 mm

long, pale green, becoming pale brown shortly before

dehiscence. Seed more or less ovoid, slightly obliquely

truncate on the ventral side, more or less truncate at

the base with a small gibbosity on each side of the

L. C. LEACH

507

suture towards the ventral side, smooth, often very

slightly pusticulate, brown, sometimes mottled with

darker colour, ± 3 mm long x 2,5-2,75 mm, some-

times somewhat laterally compressed; with a very

small, whitish caruncle. Figs. 2-6.

Kenya. — K4, Nairobi Distr., Mbagathi, Gilbert Rogers 232

(K); “Dugoretti” near Nairobi, Tweedie 1417 (K); “Karichwa

Kubwa Stream”, Nairobi, A. G. McLoughlin 984 (PRE);

ibid. Ballv “E12” (PRE). K6, "Narok Distr., Lemek Valley”,

R. M. Hornby 3088 (SRGH). K7, “Ribe to Galla Country”;

T. Wakefield (K); Teita Distr., Mwatate, Bally “El 3" (PRE).

Tanzania. — T3, Pare Distr., “Kihurio to Ndungu” “locally

dominant both sides of the S. Pares, 1800ft-4000ft”, Greenway

6476 (PRE); ibid., Leach & Brunton 10210 (K; SRGH).

Rhodesia. — N, Lomagundi Distr., Mangula, Whindale

Farm, alt. ±1185 m, restricted to a small area of chert cliffs,

fl. 18. ii. 1968, Jacobsen 3372 (PRE; SRGH); ibid., fl. 14.ii. 1968,

Wild 7686 (SRGH); ibid., fl. 24.iii.1969, Leach , Biegel & Pope

14236 (BM; BR; K; LISC; SRGH).

Fig. 2. — Euphorbia gossypina, plant from near Mangula,

Rhodesia, showing typically random branching ( Leach ,

Biegel & Pope 14236).

The specimens from the Pare District are included

here on the basis of the available material, although

from this habitat some divergence would not be

surprising.

The widely disjunct and apparently relic nature of

the small, discrete population of this species at

Mangula in Rhodesia, seems to indicate its isolation

over a very long period. Combined with its restricted

and specialized habitat this would lead one to expect

that the plants would prove to be taxonomically

distinct. However, a very close relationship with

E. gossypina certainly exists and as no significant

differences have been found between specimens from

the Mangula population and those from various

East African localities, there seems no doubt that they

are conspecific which, in view of the apparent capacity

Fig. 3. — Euphorbia gossypina, normal inflorescence with male

initial cvathium (Leach, Biesel & Pope 14236).

Fig. 4. — Eupnorbia gossypina, typical fruiting cyme with

deciduous initial cyathium (Leach, Biegel & Pope 14236).

Fig. 5. — Euphorbia gossypina, inflorescence with bisexual initial

cyathium with eight glands (Leach, Biegel & Pope 14236).

508

NOTES ON EUPHORBIA MAURITANICA. E. GOSSYPINA AND SOME RELATED SPECIES

Fig. b. — Hiupnurma gossypina, typically vanaPle inflorescences

{Leach, Biegel & Pope 14236).

for variation and speciation displayed by other

portions of the complex, is somewhat surprising.

There also seems little doubt (ex ciescr.) that N.E.

Brown was correct when placing E. implexa in

synonymy with Pax’s species.

Euphorbia berotica N.E.Br. in FI. Trop. Afr. 6,

1: 600 (1912). Type: Angola, Mogamedes District,

Welwitsch 633 (BM, holo.!; LISU!).

E. tirucalli sensu Hiern, Cat. Afr. PI. Welw. 1 ,4: 949 (1900),

pro parte excl. Col. Carp. 916.

Shrub glabrous, yellowish green, spineless, much

branched from the base, up to ± 75 cm high; root

rhizomatose, thick, horizontal, often giving rise to

adventitious plantlets at some little distance from the

parent plant. Branches alternate or occasionally forked

or subverticillate (possibly due to damage or die-

back of the growing point), spreading-ascending or

suberect, initially succulent, soon becoming woody,

terete, slender (flowering branches tapering from

3,5-5 mm diam. to 1-2,5 mm at the apex). Leaves

alternate, spreading to spreading-deflexed, sessile,

fleshy, glabrous, concave above, convex beneath,

very narrowly ovate or linear-ovate, ±1,5 mm x

10 mm to 2,5-3 mm x30 mm, quickly caducous;

leaf-scar transverse, ± straight, often slightly pro-

minent but scarcely conspicuous. Inflorescence a

cymose umbel (occasionally compound) with the

initial cyathium male deciduous or sometimes

bisexual, and 2-5 spreading rays (peduncles)

developing from below the initial terminal cyathium,

each with a bract at its base. Bracts unequal, varying

in size and shape, 2-4,5 mm wide, 2, 5-8, 5 mm

long (when more than 3 in the whorl often wider than

long), sessile, subquadrate or very broadly ovate to

narrowly ovate, acute or attenuate, often apiculate,

eciliate, most dissimilar when 3 (cf. those of E.

monteiri Hook.f.), usually deflexed, deciduous or

subpersistent. Rays (peduncles) 4-10 mm long, I ,5-2

mm diam. (sometimes varying considerably in length

in the same cyme), with a pair of bracts at their apex

and bearing a shortly stipitate bisexual cyathium,

occasionally with secondary pedunculate cyathia

arising from the axils of the bracts. Bracts sub-

persistent (persistent at least to ± mature fruiting

stage), more regular in size and shape than those at

the base of the cyme, subquadrate, subcircular or

broadly ovate, apiculate, eciliate, often erose, widely

spreading usually somewhat convex, 2-3 mm long,

3-4 mm wide. Cyathia shortly stipitate; stipe ± 2 mm

long and thick, initially merging imperceptibly into

the involucre but the junction eventually quite

distinct: involucre glabrous, cup-shaped, 2,5-3 mm

deep, 5,5-7 mm diam. including the glands; ptands

3-5, mostly 4 (usually 5 in the initial, somewhat

larger terminal cyathium, rarely 5 in secondary

cyathia, the 5th then relatively much reduced in size),

widely spreading, separate, fleshy, minutely pitted on

the concave upper surface, convex beneath, subcircular

or transversely broadly elliptic, ± 2-3 mm wide, 2 mm

long, entire or slightly crenulate-crisped on the outer

margin, yellow; lobes 5, erect, bidentate, irregularly

(sometimes obsoletely) woolly ciliate: male flowers

± 35, glabrous, with numerous, filamentose-laciniate,

very variable woolly bracteoles; pedicels 2-2,5 mm

long; filaments ± 1,5 mm long. O vary glabrous, far

exserted, curved to one side, on a 4-5 mm long pedicel

with a rudimentary rim-like perianth; styles 2-2,5 mm

long, free almost to the base or shortly connate,

deeply bifid with the lobes widely diverging, recurved.

Capsule obtusely 3-lobed or sometimes 2-lobed

(rarely with 1 aborted), 5,5-6 mm diam., 4,5-5 mm

high, eventually held erect on a pedicel ± 7 mm long.

Seed more or less broadly ovoid, somewhat obliquely

truncate at the apex, subtruncate at the base, lightly

obtusely rugulose or somewhat pusticulate, brownish

cream or brown to almost black; with an almost

sessile, somewhat laterally compressed, creamy yellow,

subtranslucent caruncle. Figs. 6-9.

Angola. — Mogamedes Distr., “Suffrutex 2-pedalis, radice

crassa . . . tota stirps . . . flavescenti-viridi-glauco. Hab. in

rupes arenoso-rubris . . . anti Boca do Rio Bero.” st. July 1859

Welwitsch 633 (BM; LISU); yellow-green subsucculent shrub

± 0,75 m high, in association with Cissus sp., in sand on rocky

slope ± 37 km from Mogamedes on Vila Arriaga road, cult.

Nelspruit, fl. & fr. Dec. -Jan. 1968-69, Leach & Cannell 14026

(BM; K; LISC; LUAI; M; MO; PRE; SRGH); subsucculent

shrubs scattered on stony bare hills, in association with Euphor-

bia virosa, ± 24 km from Mogamedes on road to Lucira, st.

1 1 .ix. 1967, Leach & Cannell 14033 (BOL; BR; COI; G; LISU;

LUA; WIND); shrubby plants with thick horizontal root, st.

26.x. 1970, Leach & Cannell 14659 (BM; K; LISC; M ; MO; PRE;

SRGH); shrubs scattered on sandy flats ± 19 km S of Curoca

drift on Espinheira road, cult. Salisbury, fl. 6.viii. 1 971 , Leach

& Cannell 14679 (SRGH); scattered shrubs in close association

with Euphorbia congestiftora in scattered “mopane” association

on sandy Hate ± 25 km S of Curoca drift, st. 29.x. 1970. Leach

& Cannell 14681 (BM; LISC), idem cult. Salisbury, fl, 6. viii.1971

(SRGH). Huila Distr., “entre Pocolo e Quihita, subarbusto de

1-1,7 m, formando colonias entre pedras”, fl. 29.viii.1963,

G. Barbosa & P. Gouveia 10720 (LUAI).

South West Africa. — 1713 (Swartbooisdrif): sandy flats,

± 17 km E of Ombepera ( — AC), shrubs ± 0,6 m high, fr.

12. iv. 1957, De Winter & Leistner 5513 (WIND).

Fig. 7. — Euphorbia berotica, plant with thick rhizomatous roots

bearing adventitious shoots ( Leach & Cannell 14659).

L. C. LEACH

509

Fig. 8. — Euphorbia berotica, plant gro-

wing in association with E. virosa

near Mogamedes ( Leach & Con-

nell 14033).

Fig. 9. — Euphorbia berotica a, plant flowering and b, fruiting

in cultivation at Nelspruit ( Leach & Connell 14026).

The heading “Imperfectly known species” under

which N. E. Brown described this species seems to be

particularly appropriate. The type specimen in the

herbarium of the British Museum (BM) (the only

specimen seen by N. E. Brown), comprises a single

sterile twig, and it is obvious that the original

description was, to a very large extent, compiled from

the information given in Welwitsch’s field notes;

without these it is doubtful if the specimen could

have ben identified, even as to the family. There is

another specimen from the same gathering in the

herbarium of Lisbon University (LISU), which is

similarly unidentifiable.

The species does not appear to have been again

collected until 1967, when plants conforming exactly

with Welwitsch’s “aff. E. rhipsaloidei . . . radice

crassa horizontali . . . flavescenti-viridi” (this is a

peculiar hue characteristic of plants of the E. mauri-

tanica group, one of the common names for which is

“the yellow milk-bush”), were collected some 37 km

from Mogamedes towards Vila Arriaga ( Leach &

Cannell 14026). To the best of my knowledge there is

no other species in southern Angola to which Wel-

witsch's description could be applied. As plants were

not in flower at the time, a few plantlets were collected

and placed in cultivation in the author’s garden at

Nelspruit in the eastern Transvaal, where both flowers

and fruits were produced during the summer of

1968-69, so enabling adequate herbarium material to

be prepared and the amplified description to be

drawn up.

E. berotica appears to be most closely related to

E. mauritanica, but differs in being a smaller shrub

with more slender branches which soon become quite

woody, while its thick rhizomatose horizontal root,

which is not known to occur in E. mauritanica , often

gives rise to plantlets at some little distance from the

parent plant and is to some extent reminiscent of the

habit of the related but quite different E. stolonifera

Marl. The leaves of E. berotica are longer and pro-

portionally narrower than those found in E. mauri-

tanica., and its eciliate, more nearly persistent bracts,

particularly those supporting the lateral cyathia,

contrast quite sharply with those of its relative which

are ciliate, concave and very quickly caducous. The

3-4, widely separate and concave, glands of the

lateral cyathia (normally 5 in E. mauritanica) are also

distinctive; however, it is possibly in the smaller

capsule and quite different seeds that the most signi-

ficant difference is to be seen.

It is interesting to note that in the narrow leaves,

more nearly persistent bracts and four glands there

appears to be a connection with the recently dis-

covered and strangely disjunct Rhodesia population

of the related E. gossypina Pax, a description of which

is given on p. 506. Although in habit apparently more

closely related to E. mauritanica , it is noteworthy that

the seeds of E. berotica and E. gossypina are more

similar to each other than to those of E. mauritanica.

However, although obviously falling into the same

group, in stature, habit and slender branches, as well

510

NOTES ON EUPHORBIA MAURITANIA, E. GOSSYPINA AND SOME RELATED SPECIES

as in some details of the inflorescence, the Angolan

species is altogether different from the Rhodesian.

Although E. berotica has a limited if somewhat

scattered distribution, it seems probable that its

coastal and inland populations should be recognized

as taxonomically distinct. However, it is considered

that no formal decision should be reached on this

until an adequate quantity of material allows the

extent of variability to be more reliably assessed.

At present the apparently significant differences

between the coastal and inland populations include

the more erect, rather taller but less freely rebranched

habit of the latter as well as their less yellowish colour;

while in the few flowering specimens seen, their leaves

and bracts are considerably longer and proportionally

narrower than those found in plants from the coastal

area. It also seems noteworthy that the liquid used

for preserving specimens remains clear when material

from inland populations is preserved, but becomes

blackish when used for similar specimens from near

Mogamedes.

Any discussion of this species would be incomplete

without reference to E. mauritanica var. lignosa

White, Dyer & Sloane to which, in slender lignescent

branches, it superficially appears to be very closely

related. However, this variety of E. mauritanica

appears to differ from var. namaquensis N.E. Br.

only in the extent, or perhaps the earlier development,

of its woodiness and, on the evidence of the herbarium

material examined, is represented by a number of

geographically widely scattered individuals.

The southern Naniib distribution, recorded by its

authors for var. lignosa , appears to refer to that of the

quite different E. chersina N.E. Br. of which one of the

syntypes is Marloth 4638 ; since var. lignosa is also based

on a specimen labelled Marloth 4638 it seems that there

has been some confusion of specimens or labels.

During a recent visit to Luderitz Bay, made to

establish, if possible, the identity of this variety, no

plants of E. mauritanica could be found, nor are

there any herbarium records of var. lignosa from the

coastal areas of the southern Namib where it is reputed

to be “the most frequent shrub”. On the other hand,

E. chersina is extremely common around Luderitz

Bay and there are numerous records of its occurrence

southward to as far as the Richtersveld to the south of

the Orange River. It seems certain, therefore, that the

material of var. lignosa labelled Marloth 4638 is from

some locality other than Luderitz Bay and that the

collector's notes attached thereto refer, in fact, to

E. chersina.

Incidentally, the fact that var. lignosa does not occur

in the southern Namib clears up the apparently rather

anomalous situation in which var. foetens (Dinter)

White, Dyer & Sloane, by far the most succulent

variety of E. mauritanica , appeared to share a large

part of its distribution with the least succulent (var.

lignosa).

Acknowledgements

The author wishes to thank the following:

The Directors of the herbaria of: the British

Museum, London (BM); the Royal Botanic Gardens,

Kew (K); the Centro de Botanica da Junta de In-

vestigates do Ultramar, Lisbon (LISC); the Univer-

sidade de Lisboa, Lisbon (LISU); the Instituto de

Investigagao Cientifica, Sa da Bandeira (LUAI); the

Botanical Research Institute, Pretoria (PRE), and

the State Herbarium, Windhoek (WIND), for the

loan of much valuable material.

Mr J. Lavranos of Johannesburg and Mr L. E.

Newton of Kumasi for live plants of the E. maurita-

nica-gossypina-schimperi Complex, and flowering

material of E. lateriflora preserved in fluid.

Dr E. J. Mendes, Centro de Botanica, Lisbon, for

much information concerning Angolan localities and

other related matters.

The South African Council for Scientific and

Industrial Research for generous financial support in

respect of the 1967 Angolan expedition.

Dr W. T. Stearn for his most helpful discussions

concerning nomenclatural problems.

Dr B. de Winter, Director, Botanical Research

Institute, Pretoria, for the facilities of the herbarium

and library, Dr L. E. Codd, formerly Director of that

Institute and Mr R. B. Drummond of the National

Herbarium, Salisbury, (SRGH), for many valued

discussions.

REFERENCES

Brown, N. E., 191 1-1912. Euphorbia. In W. T. Thiselton-Dyer,

Flora of Tropical Africa 6,1. London: Reeve & Co.

De Winter, B., 1971 . Floristic relationships between the northern

and southern arid areas in Africa. Mitt. Bot. Staatssamml.

Munchen 10: 424-437.

Jacobsen, W. B. G., 1973. A check list and discussion of the

flora of a portion of Lomagundi District, Rhodesia. Kirkia

9,1: 139-207.

Meyer, P. G., 1967. Euphorbiaceae. In H. Merxmiiller,

Prodromes einer Flora von Sudwestafrika 67.

White, A., Dyer, R. A. & Sloane, B. L., 1941. The succulent

Euphorbieae. California: Abbey Garden Press.

Wild, El., 1968. Phytogeography in south central Africa.

Kirkia 6, 2: 197 222.

bothciHa 11, 4: 511-513 (1975)

A new species of Drechslera on Trihulus terrestris

W. J. JOOSTE*

ABSTRACT

Drechslera multiformis Jooste isolated from Tribulus terrestris Linn, hay is described and its morphological

characteristics in different cultural conditions are discussed.

An interesting species of Drechslera Ito was

isolated from Tribulus terrestris hay in the course of

an investigation on toxigenic fungi in sheep pastures

(Kellerman & Marasas 1971). After intensive study

this isolate was found to have a unique combination

of characteristics which warranted description of a

new species.

Drechslera multiformis Jooste, sp. nov.

Coloniae in vitro plerumque himantoideae flocco

delicato obsitae. Mycelium olivaceo-griseum, viridi-

olivaceum vel olivaceo-nigrum. Hyphae submelleae

vel atro-mellea, lucidae, laeves vel exique verrucosae,

septatae, ramosae, 4-7 /tm diametro. Conidiophora

septata, atro-mellea ad umbrina basaliter fere flexuosa,

conidiifera parte geniculata, plerumque 2-7 geniculis,

3-7 pm diametro et usque ad 300 pm longitudine,

saepe ramosa ramis imparibus, saepe conidiis holo-

blasticis pseudoseptatis obovatis vel clavatis ad

conidiophora affixa terminantia. Cellulae conidio-

genae exigue ad cicatricem conidicam inflatae,

verruculosae et circa porum atro-umbrinae. Conidia

32,0-80,0 x 15,0-30,0 pm, solitaria, laevia variabilia,

obovata, ellipsoidea, clavata vel teretiuscula, recta

vel exigue curvata, latitudine maxima fere aut in

media aut parte superiore sua, plerumque leniter ad

apicem rotundatum vel rostellatum attenuata, inter-

dum geniculata vel apicaliter furcata, 1-7 pseudo-

septis, isabellina ad umbrina, extremis basilaribus

apicalibisque aut similiter aut dilute coloratis, raro

pseudorostrum conidiophoro secundario habent, hilo

inconspicuo.

Habitat in Tribulus terrestris Linn, foeno, Bloem-

fontein, Africa australis.

Typus ut siccata cultura PREM 44794 et 1M1

165250.

Colonies in vitro mostly velvety with delicate

floccose overgrowth. Mycelium olivaceous grey,

greenish olivaceous or olivaceous black. Hyphae pale

to dark honey, bright, smooth or slightly roughened,

septate, branched, 4-7 pm in diameter. Conidiophores

septate, dark honey to umber, basally mostly flexuous,

3-7 pm in diameter, and up to 300 pm long, conidium

bearing part geniculate mostly with 2-7 geniculations,

often branched with branches unequal, often ter-

minating in holoblastic pseudoseptate, obovoid or

clavate conidia remaining attached to the conidio-

phore. Conidiogenous cells slightly swollen at the

conidial scar, verruculose and dark umber around

the pore. Conidia 32,0-80,0 x 15,0-30,0 //m, smooth,

variable, obovoid, ellipsoidal, clavate, or somewhat

cylindrical, straight or slightly curved, maximum

width mostly in the centre or upper part of the

conidium, mostly tapering gently to a rounded or

somewhat beaked tip, occasionally forked or

geniculate, pseudoseptate with 1-7 pseudosepta,

isabelline to umber, basal and apical ends similarly or

lighter coloured, rarely with a false beak serving as a

secondary conidiophore, hilum inconspicuous.

Transfers of the holotype have been deposited in the

Potchefstroom University Culture Collection as No.

1215 and in the Centraalbureau voor Schimmel-

cultures, Baarn, Netherlands as No. CBS 480.74.

METHODS

Owing to the known variability of members of the

genus, the morphological characteristics of this species

were studied following growth of the fungus in

different cultural conditions. A culture originating

from a single spore isolate was subcultured on Difco

malt agar (MA), Difco potato-dextrose agar (PDA)

and oatmeal agar (CBS-recipe),* (OA). Cultures

were incubated at 20 °C, 25 °C, and 30 °C. Colony

diameter was measured after 10 days and permanent

mounts were made in Cunninghams mounting fluid

(Cunningham, 1972). All microscopic observations

were made on the fungus mounted in this medium.

The conidiophore and conidium dimensions are

summarized in Table 1. All colour designations are

according to Rayner (1970).

* 30 g oatmeal boiled for lh on low heat in 1 / water, strained

and the filtrate made to 1 /, autoclaved for lh at 100 kPa

before addition of 15 g agar per / and final autoclaving for 20

min at 100 kPa.

TABLE 1 — Colony diameter (mm) and conidium dimensions (jim) of Drechslera multiformis on different media after 10 days

* Institute for Botanical Research, Potchefstroom University

for C.H.E., Potchefstroom.

26700-10

512

A NEW SPECIES OF DRECHSLERA ON TRIBULUS TERRESTRIS

RESULTS

The colony growth at different temperatures is

variable although fairly constant for each medium.

On MA and PDA growth is somewhat restricted and

the mycelium mainly olivaceous to greenish black,

velvety and the colonies have an irregular edge.

An olivaceous black pigment diffuses into the agar.

On OA the growth is rapid and luxurious. The

cottony mycelium is olivaceous grey and the colony

shows an olivaceous black centre. No pigment was

observed in the agar.

Ig- . i Th-eelislera multiformis. I, young conidiophore with conidium with false beak- 2

blanched conidiophore; 3, conidiophore; 4, conidium with false beak bearing secondary

con i did . 5. branched conidiophore with holoblastic conidium; 6, terminal holoblastic

conidium 7 detached holoblastic conidium; 8 10, forked conidia; II, ellipsoidal and clavate

collid'd 12, beaked and clavate conidia; 13, ellipsoidal and somewhat cylindrical conidia; 14

obovoid and ellipsoidal conidia; 15, geniculate conidia.

W. J. JOOSTE

513

The hyphae are light to dark honey, smooth,

septate and have the same range of dimensions on all

the media. They are often slightly roughened on

PDA. The conidiophores are generally more darkly

pigmented than the mycelium. Branching occurs

regularly, but not predominantly (Figs. 2, 5). A

feature of the conidiophore is the slight inflation near

the conidial pore and the slightly verruculose and

darkly pigmented area around the pore. The coni-

diophore varies at different temperatures. On PDA

about 3-7 geniculations were observed (Fig. 3),

while only 1-2 occurred at 30 °C. A similar

observation was made on OA, although at lower

temperatures the geniculations were more evenly

spaced. A singular characteristic of this fungus is

the production of single holoblastic terminal conidia

on some conidiophores (Fig. 6). These conidia have

their outer walls continuous with that of the conidio-

phores and generally remain attached to the

conidiophore. However, it can become detached by

breaking at the septum, which separates it from the

conidiophore (Fig. 7). These terminal holoblastic

conidia are observed in old as well as young cultures

and is predominant on PDA at 30 °C.

The conidia are extremely variable in shape

(Figs. 8-15). Typically they are obovoid or ellipsoidal

and with the apical end tapering to a somewhat

beaked tip. However, many variations of this shape

occur on the various media and at the different

temperatures. On OA the conidia are generally

ellipsoidal and tapering to a rounded tip. Forked or

geniculate conidia (Figs. 8, 9, 10, 15) occur at

20-25 °C, while the terminal holoblastic conidia are

common at 30 °C. On MA at 20-25 °C the conidia

are generally obovoid to clavate and conidia with

false beaks bearing secondary conidia (Figs. 1, 4)

occur occasionally. On PDA at 20 °C the conidia

are mainly obovoid. However, terminal holoblastic

conidia occur commonly at 25-30 °C, but forked

conidia are rare at these temperatures.

CONCLUSIONS

Although this species has some features in common

with other Drechslera spp., it has distinctive

characteristics. It differs considerably in conidial

morphology from other Drechslera spp. with

branched conidiophores such as D. miyakei and

D. ravenelii (Ellis 1971). Bipolaris indica (Wadhami &

Tewari, 1969) has obovoid, apically flattened and

occasionally forked conidia. However, it has a

prominent hilum which is absent in this species.

Forked or geniculate conidia also occur in Helmin-

thosporium atypicum (Deshpande & Deshpande, 1966).

In H. atypicum this is a predominant characteristic

compared to the relative rare occurrence of this type

of conidia in D. multiformis. Furthermore, the

terminal holoblastic conidia are absent in H. atypicum

and its conidia are up to 10-septate compared to the

max'mum of 7 found in D. multiformis.

It may be concluded that D. multiformis is

characterized by the branched conidiophores, the

pigmented verruculose area surrounding the conidial

scar and the terminal holoblastic conidia which pre-

dominates at temperatures above 25° C.

ACKNOWLEDGEMENTS

Dr W. F. O. Marasas, Plant Protection Research

Institute, Pretoria, is thanked for making available

the culture of the fungus to the author. Prof. M. C.

Papendorf is thanked for his encouragement and

the critical reading of the manuscript and Dr M. B.

Ellis, Commonwealth Mycological Institute is thanked

for his valued opinion on this fungus. The financial

assistance of the South African C.S.I.R. and the

Department of Agriculture Technical Services is

gratefully acknowledged. Mr P. J. W. Jones of the

Department of Latin, Potchefstroom University for

C.H.E. is thanked for the preparation of the Latin

diagnosis.

REFERENCES

Cunningham, J. L., 1972. A miracle mounting fluid for per-

manent whole mounts of microfungi. Mycologia 64:

906-911.

Deshpande, K. S. & Deshpande, K. B., 1966. Contribution

to the taxonomy of the genus Helminthosporium I. Sydowia

Ser. 2, 20: 39-45.

Ellis, M. B., 1971. Dematiaceous Hyphomycetes. Kew: Com-

monwealth Mycological Institute.

Kellerman, T. S. & Marasas, W. F. O., 1971. The possible

role of fungi in the aetiology of geeldikkop. Unpublished

paper presented at the Congress of the S.Afr. Vet. Med.

Ass., East London, S. Afr., Sept. 1971.

Rai, J. N., Wadhami, K. & Tewari, J. P„ 1969. Bipolaris

indica sp. nov. Sydowia Ser. 2, 23: 8-10.

Rayner, R. W., 1970. A mycological colour chart. Kew: Com-

monwealth Mycological Institute.

BPS

Bothalici 11, 4: 515-521 (1975)

Notes on African plants

VARIOUS AUTHORS

ARALIACEAE

A NEW SPECIES OF CUSSONIA FROM THE SOUTHERN CAPE

Cussonia gamtoosensis Strey, sp. nov., C. spicatae

Thunb. affinis, sed arboribus parvioribus multi-

caulibus, foliis bidigitatis, pinnulis lanceolatis integris

cuspidatis, spicis minoribus differt.

Arbores parvae multicaules ad 4 m altae, radicibus

tuberosis. Folia in semi-fasciculis terminalibus dis-

posita; foliola digitata, pinnulis (3) 5-7, simplicis vel

articulatis. Pseudo-inflorescentia biennis, in anno

primo ex fasciculo ramis sparse foliosis constans, ad

apicem fasciculo bractearum imbricatarum 10-15 mm

longarum terminans. Rami primarii 4-8, anno secundo

folia exuti, spicas 4-8 penduculatas producentes.

Spicae partibus floriferis 2-3 cm longis, 2-2,5 cm

latis. Flores dense spiraliter dispositi, sessiles, virides.

Bracteae 4 mm longae, 3 mm latae, ciliatae, chartaceae,

bracteolis simulares. Calyx ovario inferiore adnatus,

oram dentatam 0,5 mm altam formans. Petala 5,

valvata, deltoidea. Fructus cuneiformes, sessiles;

exocarpium carnosum. Semina 1 vel 2, albumine

ruminato.

Type: Cape, 3325 (Port Elizabeth): Gamtoos Ferry

(-CC), 17/5/1972, Strey & Nicholson 10957 (PRE,

holo. ; NH; K; PMB).

Evergreen small trees, with up to 24 stems, up to

4 m high, trunks 3-10 cm thick, rarely branched;

wood coarse, long-fibrous, pithed; bark smooth,

light grey; roots tuberous. Leaves bidigitately com-

pound, suborbicular in outline, without scales at

articulations, up to 30 cm in diameter; petiole terete,

10-30 cm long, 2-5 mm in diameter, ribbed, glabrous;

stipules intrapetiolar, adnate to base of petiole and

fused at base, about 1 ,5 cm long; leaflets (3) 5 to 7,

digitately compound with 2 to 5 pinnules, simple or

occasionally with a vertebrate articulation (see Strey,

R.G. in Bothalia 1 1 : 191-201 ; 1973), spreading at an

angle to the petiole, thick leathery, mottled dull

greyish olive-green, petiolule 2-6 cm long, not, or

only slightly winged, pinnules 4-10 cm long, 1,5-

2,5 cm broad, entire or sometimes with 1-2-dentate

margin, lanceolate, cuspidate, nerves connecting near

the revolute margin. Pseudo-inflorescence biennial,

during first year consisting of 4-8 sparsely leafy

inflorescence branches, 14-30 cm long, 8-12 mm in

in diameter, longitudinally ribbed, with apical cluster

of triangular bracts about 12 mm long and 20 mm

broad; during second year primary inflorescence

branches drop leaves and bracts and develop a

pseudo-umbel of 4-8 spikes with peduncles 2-3 cm

long, about 1 cm in diameter, terete, with 2-4 turgid

triangular bracts at base about 1 cm long, spikes

2-3 cm long, 2-2,5 cm in diameter. Flowers spirally

arranged, sessile, bracts 3 mm broad when fresh,

4 mm long with hyaline tip, carinate, ciliate, nearly

as long as ovary; bracteoles equivalent but slightly

smaller. Calyx rim 1 mm, 5-pointed, accrescent in

fruit. Petals 5, valvate, 5 mm long, 3 mm broad at

base. Stamens 5, alternating with petals; filaments

5 mm long; anthers 2 mm long, oblong, introrse,

versatile, spreading, caducous. Ovary inferior, bilo-

cular; styles about 0,25 mm long, persistent, connivent

on minute stylopod; disc flat with radiating ridges.

Fruit 8 mm long, conical, mauve and fleshy when

fertile. Seed one per locule, 6 mm long, plano-

convex, rugose; areole orbicular under short obtuse

crest; endosperm ruminate.

Cape. — 3325 (Port Elizabeth): Gamtoos Ferry (-CC), Strey

10359 (NH, PRE, K). Strey & Nicholson 10 57 (PRE holo.;

NH, K, PNB); Von Breitenbach s.n. (Saasveld Herb.)

Due to its habit and leaf colour this species blends

well with its surroundings and is hard to spot in the

Euphorbia brush on the mountain slopes on which it

occurs. So far it has only been recorded from the

type locality.

Its habit is the same as that of C. zuluensis Strey,

but overall it is nearest to the type of C. spicata

Thunb. For the following reasons it is regarded as

specifically distinct from C. spicata. It is a small,

many-stemmed tree, the leaves are bidigitate, more

reduced, have a different colour and surface (see

Reyneke, W.F. in Proc. Electron Microscopy Soc.

Sthn. Afr. 3: 33-34; 1973, who refers to the species

here described as Cussonia sp. nov.), while bracts,

bracteoles and flower characters also show differences.

R. G. Strey

FLACOURTIACEAE

A CHANGE OF STATUS IN SCOLOPIA

An examination of type and other material of

Scolopia flanaganii (H. Bol.) Sim var. flanaganii and

var. oreophila Sleum. in herbaria and investigations

in the field have led the author to the conclusion that

var. oreophila is sufficiently distinct to constitute a

species. It differs from var. flanaganii in that the

leaves have an acuminate instead of acute-blunt

apex, 5-6 instead of 7-13 pairs of veins, and the

flowers are sessile and solitary instead of pedicellate

and 2 or 3 per leaf axil. In addition, the leaves of

var. oreophila tend to be somewhat rhombic in

shape with usually reddish petioles.

The change of status is made as follows:

Scolopia oreophila (Sleum.) Killick, stat. nov.

S. flanaganii (H. Bol.) Sim var. oreophila Sleum.

in Blumea 20,1: 51 (1972).

D. J. B. Killick

516

NOTES ON AFRICAN PLANTS

Fig. 1. — Cussonia gamtoosensis. 1, habit; 2, pseudo-inflorescence, x 1/6; 3a, b, seed, front and back view, x 2;

3c, surface of seed enlarged, x 5; 4a, g, petal, front and side view, x 4; 5, stamen, x 4; 6, anther, x 8; 7, flower

bud, x 4; 8, flower at anthesis, x 4; 9, longitudinal section of flower at anthesis, x 4; 10, longitudinal section

of mature flower, x 4; 11, buds showing arrangement in spike, x 2; 12, flower viewed from above, x 4; 13,

arrangement of bracts, x 4; 14a, b, c, bracts, front view, x 4.

VARIOUS AUTHORS

517

OCHNACEAE

THE IDENTITY OF OCHNA ATROPURPUREA

In 1811 De Candolle described Ochna atropurpurea

basing it on a description and plate in Plukenet’s

Almagestum (1694/96). He derived his epithet from

the words “floribus pentapetalis, atro-purpureis”. In

the Almagestum it is stated that Alexander Brown

collected the plant (now in the Sloane Herbarium,

BM) at the Cape of Good Hope, but this seems

unlikely since it does not match any of the species

known from South Africa. According to Robson in

Taxon 11: 49 (1962), Brown may have visited

Mozambique while collecting in the Comoro Islands.

This view is supported, as the Brown specimen agrees

very closely with the type of a Mozambique species,

Ochna mossambicensis Klotzsch, Peters s.n.

Later, in his Prodromus 1 : 736 (1824), De Candolle

cites both Plukenet, t. 263, fig. 1 and 2, and Burchell

4126 under O. atropurpurea DC. Subsequently all the

South African material corresponding to the Burchell

specimen has been referred to Ochna atropurpurea

DC. The Burchell specimen from the eastern Cape,

however, does not match the Plukenet figure nor the

Brown specimen, but agrees very closely with the type

of Ochna serrulata (Hochst.) Walp. All the South

African material previously known as Ochna atro-

purpurea DC. should therefore be referred to O.

serrulata (Hochst.) Walp.

The nomenclatural consequences are summarized

below:

Ochna atropurpurea DC. in Ann. Mus. Paris 17:

412 (1811). Type: The illustration in Plukenet, Aim.

41, t. 263, fig. 1 and 2 (1694/96), based on material

collected by Alexander Brown (Sloane Herbarium,

BM, typotype).

O. jabotapita var. (] L., Sp. PI. 513 (1753). O.

squarrosa var. /?L., Sp. PI. ed. 2: 732 (1762). O. mossa-

bicensis Klotzsch in Peters, Reise Mossamb. Bot.

1: 88, t. 16 (1861); Robson in Taxon 1 1 : 49 (1962);

FI. Zamb. 2, 1 : 233 (1963). Type: Mozambique: Sena,

Peters (B, holof ; K; PRE, photo).

For full synonymy, See Robson, l.c. The species

occurs in Kenya, Tanzania and northern Mozam-

bique.

Ochna serrulata {Hochst.) Walp., Rep. 400

(1846). Type: Natal, Natal Bay, Krauss 473 (K; PRE,

photo.).

Diporidium serrulatum Hochst. in Krauss in Flora

27: 304 (1844). Type: as above.

Ochna atropurpurea sensu auct. plur. as to South

African material, non DC.

The distribution of this species extends from

Knysna in the eastern Cape, through Natal to the

north-eastern Transvaal.

A CHANGE IN STATUS AND A NEW NAME FOR OCHNA ATROPURPUREA VAR. ANGUSTIFOLIA

The plant described as Ochna atropurpurea DC.

var. angustifolia Phill. is considered to be worthy of

specific rank. When its status is raised, the epithet

angustifolia cannot be used because of the existence

of Ochna angustifolia Engl. & Gilg published in Bot.

Jahrb. 32: 135 (1902). Therefore a new name, Ochna

gamostigmata, has been chosen. In view of the

inadequate diagnosis provided by Phillips, a full

description is given here and the change in status is

formally made.

Ochna gamostigmata Du Toit, stat. et nom. nov.

Lectotype: Transvaal, Barberton, Upper Moodies,

Galpin 963 (PRE, lecto.).

Ochna atropurpurea DC. var. angustifolia Phill. in

Bothalial : 95 (1922), non O. angustifolia Engl. & Gilg

(1902).

Shrubs 0,5-1 m high, deciduous; bark rough,

brown, young stems reddish-brown, lenticeliate.

Leaves alternate; petiolate; blade linear, lanceolate

or ovate, 16-25 (-45) x3-8 (-9) mm, acute or rounded

at the apex, often mucronate, rounded or cuneate at

the base, margin inconspicuously denticulate to con-

spicuously serrulate, lateral nerves numerous, pro-

minent above, smooth below, tertiary reticulate

venation inconspicuous, midrib prominent above,

chartaceous or subcoriaceous uniformly dark green;

petiole 1-1,5 mm long, sometimes slightly swollen at

the base. Flowers solitary, terminating short axillary

shoots up to 0 , 5 mm long, yellow ; pedicel 6-8 (-9) mm

long, articulated at the base. Sepals 5, 4-6 x2-3 mm,

green, elliptic or ovate in flower; 6-7 x4-5 mm, dark

ferruginous, reflexed in fruit. Petals 5,5-7 x4-5 mm,

rhombic or obovate, with claw 2 mm long, deciduous.

Stamens numerous, irregularly arranged in a dense

whorl, filaments 1-1,5 mm long, straight, anthers

yellow, 2-3 mm long, dehiscing by apical pores,

deciduous. Carpels 4-5, free; styles completely

united; stigma capitate, lobed. Drupelets globose,

3,5-4 x 3, 5-4 mm, attached at the base.

Recorded from the Natal Drakensberg, Swaziland

and eastern Transvaal above 1 000 m, on loam or

clayey soils; warm summers, rainfall up to 2 000 mm,

frost in winter; flowering October-January.

Transvaal. — 2430 (Pilgrim's Rest): The Downs, Madeira

(-AA), Crandall sub PRE 32976; 17 km E. of Graskop (-DD),

Codd & De Winter 3119. 2531 (Komatipoort): Shelangubu

Valley, S. E. of Barberton (-CC), Pole-Evans 4680; Barberton,

Upper Moodies (-CC), Galpin 963. 2630 (Bethal): Ermelo,

Nooitgedacht (-CD), Pott 5096. 2730 ( Vryheid): Wakkerstroom,

farm Oshoek (-AD), Devenish 1264.

Swaziland. — 2531 (Komatipoort): Havelock, Mtutusiriver

(-CC), Compton 29126. 2631 (Mbabane): Hlatikulu (-CD),

Galpin 9622.

Natal. — 2730 (Vryheid): Hlobane, forest margin at top of hill

(-DB), Du Toit 33. 2829 (Harrismith): hills north of Bergville

(-CB), Gillett 1188. 2929 (Underberg): Estcourt Research

Station (-BB), West 817; Estcourt, at top of Griffin's Hill

(-BB), Du Toit 9. 2930 (Pietermaritzburg): 22 km N.W. of New

Hanover (-AB), Codd 1465; Karkloof, Benvie (-AD), Medley

Wood 7813; Karkloof Falls (-AD), Rycroft 58.

Ochna gamostigmata differs from O. serrulata in

being much smaller, under 1 m high, with smaller,

narrower, acuminate rarely rounded leaves and

smaller flowers. O. gamostigmata has the pedicels

articulated at the base, the styles completely united

and the capitate stigma lobed, whereas in O. serrulata

the articulation is situated 1-3 mm above the base,

the styles free and recurved at the apex for 1-2 mm,

and the individual stigmas capitate. O. gamostigmata

is apparently never very abundant, usually only one

or two specimens being found at each locality. It

occurs in a distinct habitat, namely exposed forest

margins on heavy soils at an altitude of 1 000-2 000 m.

In this it differs from its nearest ally, O. serrulata ,

which is usually found in forests on sandy or loamy

soils, at altitudes below 1 000 m.

51 S

NOTES ON AFRICAN PLANTS

A CHANGE IN STATUS FOR OCHNA OCONNORI1

Ochna arborea Burch, ex DC. var. oconnorii ( PhiU .)

Du Toit, stat. nov.

O. oconnorii Phill. in Bothalia 1 : 92 (1922); Robson

in FI. Zamb. 2: 230 (1963). Type: Transvaal, Wood-

bush forest, O'Connor sub PRE 1257 (PRE, lecto!).

Examination of a wide range of material in PRE

reveals that O. oconnorii is merely a luxuriant meso-

phytic variety of O. arborea with longer, narrower,

sharply serrate leaves and well developed flowering

side branches with many-flowered branchlets. The

type of O. arborea was collected in the south-east

Cape in sclerophyllous coastal vegetation. Its leaves

are comparatively small and leathery and it is few-

flowered. This typical variety is found intermingled

with the more luxuriant variety along its range from

the Cape to the northern Transvaal and intermediates

showing characters of both taxa are frequently found.

P. C. V. du Toit

RUTACEAE

A NOTE ON VEPRIS LANCEOLATA

For many years now the White Ironwood or

Witysterhout has been known as Vepris undulata

(Thunb.) Verdoorn & C. A. Smith. However, Boscia

undulata Thunb., Prodr. PI. Cap.: 32 (1794), the

basionym of Vepris undulata (Thunb.) Verdoorn &

C. A. Smith in J. S. Afr. For. Assoc. 20: 35,50 (1951),

is a nomen nudum. The name Boscia undulata was only

validly published by Thunberg in his FI. Cap. I : 576

(1813). The epithet “ undulata ” therefore dates from

1813 and not from 1 794.

During the intervening years between 1794 and the

valid publication of B. undulata in 1813, Toddalia

lanceolata Lam., Tab. Encycl. Meth. Bot. 2: 117

(1797), was published. The earliest specific epithet for

the White Ironwood is therefore “ lanceolata ” and the

species must now revert to the name under which it

was formerly known, namely, Vepris lanceolata (Lam.)

G. Don.

Vepris lanceolata (Lorn.) G. Don, Gen. Syst. 1 : 806

(1831); Harv. in FI. Cap. 1: 447 (1860); Verdoorn in

Kew Bull 1926: 395 (1926); Burtt Davy, FI. Transv.

2: 479 (1932). Syntypes from the Mascarenes, Herb.

Lamarck (P-LA, IDC microfiche 114/15 and 114/16).

Toddalia lanceolata Lam., Tab. Encycl. Meth.

Bot. 2: 117 (1797); Sim, For. FI. Cape Col.: 157,

t. 22 (1909). Syntypes as above.

Boscia undulata Thunb., [Prodr. PI. Cap.: 32 (1794)

nomen nudum] FI. Cap. 1 : 576 (1813). Syntypes from

the Cape of Good Hope, Herb. Thunberg 3791, 3792

(UPS).

Scopolia lanceolata (Lam.) Spreng., Syst. Veg. 1 :

779 (1824).

Asphades undulata (Thunb.) DC., Prodr. 2: 90

(1825).

Vepris undulata (Thunb.) Verdoorn & C. A. Smith

in J. S. Afr. For. Assoc. 20: 35,50 (1951); Palmer &

Pitman, Trees S. Afr.: 274 (1961); Mendonca in FI.

Zamb. 2 (2): 200 (1963); Ross, FI. Natal: 214 (1973);

Palmer & Pitman, Trees S. Afr. 2: 989 (1973).

J. H. Ross

STERCULIACEAE

A NEW SPECIES

Hermannia liioralis Verdoorn sp. nov., inter species

filamentis cruciformibus et foliis verticillatis ad

H. elliotianam fioribus angustis accedens sed ab ea

pedicellis brevibus bracteis tectis, petalis luteis et

calycibus villosis recedens.

Fruticulus effusus; ramuli glandulis breve stipitatis

instructi. Stipulae foliaceae simplices vel ad basim

lobulatae. Folia breve petiolata, glandulis breve

stipitatis instructa; lamina lineare-lanceolata, lanceo-

lata vel anguste oblanceolata, irregulatim crenato-

lobulata. Cymae terminales, 2-florae; pedunculi longi

(c. 7 cm. longi), graciles, glandulis stipitatis instructi;

pedicelli breves, bracteis tecti; bracteae latae, pal-

mati-vel pedati-lobatae. Floris c. 13 mm longi aliquan-

tum angustis. Calyx c. 8 mm longus, turbinatus vel

urceolatus, 5-lobatus; lobi 5 mm longi villosi. Petalci

lutea, contorta. Filamenta supra medium dilatata, lobis

villosis. Ovarium breve stipitatum, dense stellate-

pubescens, pilis brevibus.

Type: Cape, 3217 (Vredenburg): Brittania Bay

(-DB), Acocks 15205 (BOL. holo. ; PRE, iso.).

Low suffrutex much branced at the base, lower

branches straggling; branches with short glandular

hairs, gland-tipped hairs and, sometimes, short

stellate hairs also present. Stipules leaflike, shorter

than the leaves, sessile, simple or lobed to the base,

lobes linear, 5-9 mm long, entire, pubescent with

fine, mostly gland-tipped hairs, subglabrescent. Leaves

shortly petiolcd; blade, lanceolate linear-lanceolate or

narrowly oblanceolate, somewhat unevenly lobed

on the margins, pubescent with mostly glandtipped

hairs especially along the margins, 7-20 mm long;

petiole up to 5 mm long, pubescent with gland-tipped

hairs. Inflorescence of 2-llowered cymes, terminal on

the branches (some appearing opposite the upper

OF HERMANNIA

leaves as the main branch turns aside); peduncles

long, 2-7 cm long, slender pubescent with gland-tipped

hairs; pedicels short, 1,5-5 mm long; bracteoles

broad concealing the pedicels, palmately or pedately

lobed 5-10 mm long, minutely glandular pubescent

but villose on the margins. Calyx about 8 mm long,

urceolate to turbinate, lobed to beyond the middle,

lobes narrowly oblong, sinuses narrow, minutely

glandular pubescent but villose along the margins

without and within. Petals yellow, tightly twisted

giving the flowers a narrow appearance, about 12 mm

long, oblong-obovate, narrowing gradually to a short

claw with narrowly infolded margins microscopically

pubescent below. Stamens about 7,5 mm long,

filaments dilated above the middle, lobes villous;

anthers obscurely ciliate. Ovary about 3,5 mm long,

densely stellate with short hairs; stipe 0,75 mm long;

styles 9,5 mm long. Capcule oblong at maturity,

about 10 mm long, subdensely pubescent with short

stellate hairs glabrescent, perianth persisting at base.

Cape. — 3217 (Vredenburg): Brittania Bay, Acocks 15205.

3218 (Clanwilliam): St. Helena Bay, Lavranos 10991.

The late Mr Pillans recognized Acocks 15205 as

belonging to an undescribed species and gave it the

manuscript name Hermannia litoralis. His notes did

not indicate the relationships or the features by

which to distinguish the species. Attempts to find

more material of this species were fruitless until late

in 1973 when Mr John Lavranos collected it at St

Helena Bay, that is close to Britannia Bay, the type

locality. With this further material it has been possible

to distinguish and define the species. Mr Pillans apt

name has been retained, but his rough description

does not exactly tally with the present concept of

the species.

VARIOUS AUTHORS

519

TYPIFICATION OF HERMANNIA RUGOSA

A study of the type specimens of H. rugosa

Adamson reveals that they do not constitute a single

definable species but represent several distinct

elements. Mixed gatherings of Hermannia species are

not infrequent for the plants often grow in mixed or

hybrid populations.

The type specimens of H. rugosa Adamson are as

follows:

1. Salter 8772 from Newlands House Estate, holo-

type (BOL). There are two specimens on this sheet

representing two distinct elements. The lower repre-

sents a species with leaves which are crenate and

crisped on the margin and the calyx somewhat salver-

shaped, angled in the upper half and densely stellate

pubescent. The upper specimen is nearest H. multi-

flora Jacq., differing from the lower in the leaves not

being crisped as well as crenate on the margin. In

addition, it shows one of the narrow entire, petioled

leaves between the bracts near the base of the

inflorescence, a characteristic of H. multiflora. This

suggests that it is one of the few putative hybrids seen

between H. rugosa in the restricted sense and H.

multiflora.

The lower specimen with crisped leaves agrees with

specimens from Paarl Mountain, specially collected

for this investigation by Miss M. F. Thompson of

the Stellenbosch Herbarium, all with the typical leaves

and calyx. The first specimen from that area brought

to my notice, Kruger M 22. is also a mixed gathering,

three of the specimens on the sheet agreeing with the

lower specimen of the holotype, and the fourth

nearest H. alnifolia L. having the characteristic small

flowers, the calyx with stellate hairs mainly on the

obvious veins and the subrotund leaves. A very

interesting find is a Drege specimen in the Stockholm

Herbarium collected ‘Zwischen Paarlberg und

Paardeberg” which agrees with the present day

Paarl Mountain specimen and the lower specimen on

the type sheet of H. rugosa, Salter 8779.

These findings have shown that we are dealing with

a definte species, but that it is necessary to select the

lower specimen on the type sheet as the lectotype

of H. rugosa Adamson and to exclude the upper

specimen and some of the paratype material.

2. Salter 8669, also from Newlands Estate, agrees

with the lectotype.

3. Salter 8780 from Green Point, not a very good

specimen but could be H. rugosa sensu stricta, having

crisped leaves and the correct calyx.

4. Adamson 2511 from Green Point is most

probably a hybrid between H. rugosa and H. alnifolia

but nearest H. rugosa.

5. Pillans 4761 (cited as 4701) from the Tiger Berg,

is H. prismatocarpa E. Mey. ex Harv. with leaves

not strongly crisped and the calyx cup-shaped, wide

at the mouth.

6. Gillet 4162, from Malmesbury district is H. pris-

matocarpa E. Mey. ex Harv. with the diagnostic

capsules much exceeding the calyx.

7. H. Bolus 12619 from Saldanha Bay, is also

H. prismatocarpa.

8. Gillett 3731, Piquetberg, appears to be H.

repetenda Yerdoorn ( = H . hirsuta Schrad. non. Mill.).

The inflorescence is immature.

9. H. Bolus 9948 from near Moorreesberg in the

Malmesbury district, appears to be a hybrid nearest

H. scabra Cav., which is fairly common in that area,

and possibly with some H. multiflora influence.

To date, Salter 8900, Adamson 2922 and 3526, and

PHlans 8621 have not been traced.

Adamson’s description of H. rugosa is rather

wide and not detailed enough. It does not mention

the features which have been found to characterize

the species, namely the crisped leaves and the some-

what salver-shaped, densely stellate pubescent calyx.

But there is nothing of significance in it to exclude

the lectotype.

I. C. Yerdoorn

STRELITZIACEAE

THE STATUS OF STRELITZIA JUNCEA

At the time of the publication of the first edition of

Hortus Kewensis by W. Aiton in 1789 only one

species of Strelitzia (S. reginae Ait.) was in cultivation

at Kew. It was recorded as having been introduced

by Sir Joseph Banks in 1773. It is clear from con-

temporary writings that Banks had received it from

his sponsored collector, Francis Masson, who had

accompanied the Swedish botanist Carl Peter Thun-

berg in that year on a botanical expedition to the

eastern Cape. They travelled as far east as Coega a

few miles north east of Port Elizabeth and must have

collected plants or seed or both along their route.

When the second edition of the Hortus Kewensis

was published in 1811 the number of “species” of

Strelitzia in cultivation at Kew was six, including the

arborescent S. alba (L.f.) Skeels {=S. augusta Thunb.

also sent to Banks by Masson). The other five were

all within the S. reginae complex (S. reginae, S.

ovata Ait., S. farinosa Ait., 5. angustifolia Ait. and

S. parvifolia Ait.) and all introduced from the Cape

without exact origin or location. S. parvifolia, which

concerns us most, was introduced by Sir Joseph Banks

and probably collected by Masson during his second

visit to the Cape. Aiton described his 5. parvifolia ,

the small-leaved Strelitzia, as having a scape equal in

length to the petiole, which is 20 times longer than the

linear lanceolate leaf. There is no type preserved at

Kew.

Sir James Smith writing under the heading of

S. reginae in Rees’s Cyclopaedia 1819, comments that

S. angustifolia is a mere variety of S. reginae and that

offsets of an original root gradually diminished in

size and breadth of their leaves, became first S.

angustifolia and then S. parvifolia of Hort. Kew.

Similar varieties may have been freshly imported from

the Cape, but in his view that did not prove their

specific difference. “In some specimens the leaf

dwindles to a point”. This last observation clearly

refers to S. juncea Link.

This passage was quoted in Feb. 1821 when Ker-

Gawler described and illustrated S. parvifolia var.

juncea in the Botanical Register 6, t. 516. Shortly

520

NOTES ON AFRICAN PLANTS

afterwards in the same year this concept was given

specific status as S. juncea Link in Enum. PI. Hort

Reg. Berol. 1 : 150, whether or not on the same plant

is uncertain. While there is appreciable leaf variation

within the S. reginae complex and between it and

5. juncea, no structural difference has been observed

in their flowers, which does tend to reduce the

importance placed on the difference in the leaves.

In 1893 J. G. Baker at Kew published a synopsis

of the Genera and Species of the Musaceae in Ann.

Bot. 7, in which he recognized two acaulescent species

of Strelitzia, S. reginae with four varieties and S.

parvifo/ia with variety juncea. Thus, when Wright

monographed the genus Strelitzia for Flora Capensis,

5,3: 318 (1913), he followed Baker closely and added

only one further variety under S. reginae.

The position remained relatively unchanged until

Moore & Hyypio in the the U.S.A. published their

comments on Strelitzia in Baileya 17: 65 (1970).

Since distinctions from S. reginae , other than the

lack of a leaf-blade, are not apparent in S. juncea and,

since in their view its distribution in South Africa

was not well documented, they reduced S. juncea to

S. reginae and added that, if desired, it could be

designated as a cultivar — juncea. To this proposal 1

objected. But writing from Pretoria, with a hazy

memory of 40 years earlier, 1 had written to Professor

Moore stating erroneously that 5. reginae occurred

with S. juncea near Uitenhage in the eastern Cape

and that intermediate forms corresponding to what

had been called S’, parvifolia and S. angustifolia

occurred there and this suggested to me that

hybridization was the cause of the intermediate forms.

My wishful thinking was corrected by Mrs Nod

Urton of Port Elizabeth, who has studied the local

Strelitzias for 30 years. While S. juncea was common

in the vicinity of Uitenhage, no S. reginae occurred

there. She was uncertain of the application of the

name parvifolia, because of some variation in the

leaves of S. juncea near Uitenhage. In the great majority

of plants the leaves had practically no lamina, others

had a small lamina and yet a few others a lamina up

to 15 cm long and 3,75 cm broad. I went back to

square 1.

It was at this juncture that 1 raised the interest of

Prof. J. G. C. Small and Dr El. A. van de Venter of

the University of Port Elizabeth in the problem.

It was for his thesis on the distribution, leaf morpho-

logy, embryology and germination of the acaulescent

species of Strelitzia Ait. that Dr van de Venter was

awarded his Ph.D. in 1974.

Dr van de Venter plotted the distribution of S.

reginae from the farm Ventershoek near Patensie in the

Humansdorp District eastwards at intervals in coastal

districts to southern Natal and thence with a wide

break to Zululand. S. juncea was found in abundance

north-west of Uitenhage and in two other relatively

small populations, the one north of Port Elizabeth

and the other near Patensie in the presence of S.

reginae. It was this natural association of 5. reginae

and S. juncea at Patensie, which I had hoped to find,

to help explain the intermediate forms which had

reached Europe in the early days before any artificial

hybridization had been possible in gardens. Patensie

was on the direct route taken by Thunberg and

Masson and was the most likely site from which

Masson shipped his first specimen of S. reginae to

Banks, and here there occurs naturally a complete

range of forms between S. reginae and S. juncea.

Beginning with seed from segregated wild stock of

S. reginae and 5'. juncea. Dr van de Venter observed

no difference in the germination of their seedlings

until the development of the second lamina Those of

S. juncea were somewhat smaller. After the eighth

leaf of S. juncea, successive leaves showed a progressive

reduction in the size of the lamina in contrast to the

increasing size in S. reginae. After the eighth leaf of

S. juncea. there was also a progressive lengthening of

successivepetioles. In his view, the distinctive

development patterns between the two plants

were genetically governed.

With this background, he suggested that the inter-

mediate forms between S. reginae and S. juncea,

which he studied in the wild at Patensie, were the

result of hybridization between the two species, but

that it might seem tempting to speculate that S.

juncea arose in the first place as a mutation of

S. reginae.

It has long been accepted in plant taxonomy that

hybridization and mutations are factors in the

evolution of new taxa worthy of specific status, so in

either case S. juncea could retain its specific identity

and status.

Relative to the status of S’, parvifolia, Dr van de

Venter states that a few isolated wild clones could be

identified as such, but that nowhere did he find a pure

stand conforming to the description. On anatomical

features, his material showed a close relationship to

S. reginae and could be regarded as an extreme

form of it, but there appeared no direct correlation

with S’, juncea. Thus, in his view, one can only

conclude that the application of the old, epithet

parvifolia is ambiguous and should not be allowed

to confuse the application of the name S. juncea for

an extensive and relatively uniform wild population

in the district of Uitenhage.

Where typical specimens are grown together either

singly or in mass, for example at Kirstenbosch

Botanical Garden and in various parks and other

gardens, they retain their characteristics and those of

S. juncea are invariably taller than those of S. reginae.

It is also interesting, but not surprising, that if the

two species are grown together with inadequate

light for normal photosynthesis, S. juncea is the first

to show signs of dejection.

It is relevant to remark here on the name S. tereti-

folia Barrow ex Steud., Nom. Bot. ed. 2, ii: 645 (1840),

which Wright l.c. referred to under the heading

“Imperfectly known species, but suggested that it

was equal to 5. parvifolia var. juncea. It seems

certain that the trivial epithet teretifolia was derived

from its use by John Barrow in his Account of

Travels into the Interior of Southern Africa in the

years 1797 and 1798 ‘published in London 1801.

On p. 189 we read: “The Strelitzia regina also, now

in full and beautiful bloom, grew everywhere in wide-

spreading patches in the vicinity of the Great Fish

River, but not one of the new species, discovered about

twenty mile to the northwards of Swartkop’s river,

could be found among them. The cerulean blue

nectarium of the reginae seemed to be uniformly faded,

and it lost its colour by a short exposure to the

weather, which did not appear to be the case with

that of violet-blue of the teretifolia."

There is little doubt that this is a reference (possibly

the earliest) to S. juncea Link as we see it, and it is

of considerable interest to have Barrow’s early assess-

ment of the problem we have discussed above.

In a letter received from Prof. Harold E. Moore,

since this article went to press, he proposes to give

varietal rank to S. juncea ander S. reginae and

there is mutual regret that oue views do not coincide.

R. A. Dyer

VARIOUS AUTHORS

521

Fig. 2. — Photograph of Thunberg 7191, syntype of

Conium chaerophylloides (Thunb.) EckL & Zeyh.

housed in the Institute of Systematic Botany of

the University, Uppsala.

The two syntypes were compared with the material

of C. maculatum L. housed in the Kew Herbarium

in an attempt to establish which characters

differentiate these allegedly distinct species. The

comparison revealed that the syntypes of C. chaero-

phylloides and the material of C. maculatum agree in

all essential characters examined, although two small

apparent differences were noted:

1. The stems in the syntypes of C. chaerophylloides

(and in all of the South African material seen) lack

the purplish spots, i.e. they are not maculate,

which are a fairly characteristic feature of much of the

material of C. maculatum. However, C. maculatum

is a variable species and does not always have maculate

stems: there is a variant in Europe with non-maculate

stems.

Fig. 3. — Photograph of Thunberg 7192, syntype of

Conium chaerophylloides (Thunb.) Eckl. & Zeyh.

housed in the Institute of Systematic Botany of

the University, Uppsala.

2. The petals in C. chaerophylloides are slightly

narrower and smaller although there is no discon-

tinuity in petal size between C. chaerophylloides and

C. maculatum. Petal size in C. maculatum appears to

be fairly variable and petals the size of those on the

syntypes of C. chaerophylloides do occur in European

material.

These slight differences do not appear to be of

sufficient taxonomic significance to enable C. chaero-

phylloides and C. maculatum to be maintained as

distinct species. 1 am persuaded, therefore, to follow

the decision taken by Hiern in FI. Trop. Afr. 3: 9

(1877) to treat C. chaerophylloides as a synonym of

C. maculatum. The occurrence in South Africa of

plants with unspotted stems could perhaps be

accounted for by the introduction from Europe of

the variant with non-maculate stems.

Conium maculatum L., Sp. PI. ed. 1: 243 (1753);

Hiern in FI. Trop. Afr. 3: 9 (1877).

Seseli chaerophylloides Thunb.. Prodr. PI. Cap. 1: 51 (1794);

Willd., Sp. PI. ed. 4, 1: 1461 (1798); Thunb.. FI. Cap. ed.

Schult. 259 (1823); DC., Prodr. 4: 148 (1830). Syntypes: Cape

Province, Thunberg 7191 (UPS!), Thunberg 7192 (UPS!).

Conium chaerophylloides (Thunb.) Eckl. &Zeyh., Enum.: 355

(1836). Syntypes as above.

EJMBELLIFERAE

A NOTE ON THE IDENTITY OF CONIUM CHAEROPHYLLOIDES

The description of Seseli chaerophylloides Thunb.,

Prodr. PI. Cap. 1:51 (1794), the basionym of Conium

chaerophylloides (Thunb.) Eckl. & Zeyh., Enum. 355

(1836), was based on Thunberg 7191 and 7192 from

the Cape Province. Through the courtesy of the

Director, Institute of Systematic Botany of the

University, Uppsala, the two syntypes of C. chaero-

phylloides were received on loan (see Figs. 2 and 3).

.

Bothalia 11, 4: 523-537 (1975)

A quantitative description of some coast forests of Natal

D1LWYN J. ROGERS* and EUGENE J. MOLLf

ABSTRACT

Ten stands of subtropical forest in four areas along the coast of Natal, South Africa, were sampled using

five 1/10 acre circular plots in each stand. A total of 101 woody species over one inch d b.h. was encountered

with a range of 20 to 40 species per stand. Quantitative results, including numbers and sizes, are given for

canopy, subcanopy, and understorey species as they occurred in these various layers. Specific size limits were

used to recognize the three layers, thus putting all stands on a comparable basis. Relative basal area per

acre figures were used as an indication of the relative dominance of the various species and layers, and the

possibility of a biological principle to justify such usage is mentioned.

Although the stands are complex and are seemingly heterogeneous, there are definite patterns of species

behaviour, and trends are indicated for the 10 stands and for the four forest areas. Relatively few species are

dominant in each stand and the apparent diversity is mainly due to the high percentage of species that are

relatively uncommon in each stand. The methods described in this paper should be applicable in a study of a

broader range of Natal forests, an area from which quantitative studies have heretofore been virtually absent

INTRODUCTION

Few quantitative studies of subtropical forest

vegetation have been made in South Africa. These

forests have been described subjectively and profiles

have been drawn to show the different layers. More

specifically, virtually no quantitative studies have been

made of the coasts forests in Natal until recent

studies in which Moll (1968a; 1968b; 1969) gathered

density and frequency data for various synusiae in

large quadrats.

On the other hand, rapid techniques to sample

large area of forest vegetation have been used in the

temperate zone, for example the quarter method

(Cottam and Curtis, 1956), where quantitative analytic

techniques have been developed (Curtis, 1959).

In temperate zone forests, however, the vegetation is

more homogenous and the number of woody species

is smaller compared to subtropical forests. The

quarter method was briefly investigated in Natal

forests, but was not considered suitable because it

requires a larger homogenous area than was found to

exist. Instead of the quarter method, this paper

describes an area-method of sampling, which was

applied in 10 stands of forest in Natal, and the

subsequent quantitative analysis of these stands.

STUDY AREA

The 10 stands of forest, from four areas of Coast

Forest, were studied in December 1967. Table 1

gives a summary of various physical characteristics of

the stands. Stands 1 to 3 were located in the Krantz-

kloof Nature Reserve, 25 km northwest of Durban,

and are shown as area A on the map of Natal (Fig. 1).

Stand 1 was on a Table Mountain Sandstone slope

above a cliff. Stand 2 was just below the 100 m high

cliff, and stand 3 was closer to the valley bottom

than stand 2 (Fig. 2). The underlying rocks of stands

2 and 3 were granite, but many large sandstone rocks

Fig. 1. — Map of Natal, South Africa, showing forests in which

stands were studied. Latitude, longitude, and the location

of Durban are shown for purposes of orientation. See

Table 1 for location of stands in areas A to D. The few

remaining similar natural forests along the Natal Coast are

located in areas marked X.

TABLE 1. — Characteristics of the ten stands of Natal Coast Forest which were sampled

* Department of Biology, Augustana College, Sioux Falls,

South Dakota, U.S.A.

t Department of Botany, University of Cape Town, South

Africa.

524

A QUANTITATIVE DESCRIPTION OF SOME COAST FORESTS OF NATAL

had fallen from the cliff, and weathered sandstone was

probably the main soil parent material. Stands 9

and 10 were located in Stainbank Nature Reserve,

16 km southwest of Durban (area B on the map),

and both stands were on Table Mountain Sandstone.

Stands 4 to 6 were in the Hlogwene Forest, which is

80 km north of Durban near the town of Stanger

(area C on the map). Tne soil parent material was

dune sand. This forest is owned and protected by

Hulett’s Sugar Co., and is pictured in Fig. 3. Stands

7 and 8 were in the Hawaan Forest, 19 km north of

Durban (area D on the map). This forest is owned

by Natal Estates (Pty) Ltd and is likewise on dune

sand.

Fig. 2. — Krantzkloof Nature

Reserve, looking east.

Stand 1 is above the cliff

on the left, stand 2 at the

base of the cliff, and

stand 3 to the left of the

valley bottom. Note arid

northwest-facing slopes

in distance.

I ig. 3. Aerial view of Hlogwene Forest, location of stands 4-6, looking south. The Indian Ocean

is on the left, and the Tugela River in the foreground. Most dune forests of this type along the

coast have been replaced by sugarcane plantations, or, near Durban, by housing development.

(Photo: Natal Mercury).

DILWYN J. ROGERS AND EUGENE J. MOLL

525

Each stand was situated well within the forest to

avoid any edge effect. The total forest in each of the

four areas ranges from 50 to 250 ha.

The forests discussed here are classified by Acocks

(1953) as Coastal Tropical Forest, although the

Krantzkloof stands are at the inland and altitudinal

limits of this type. Edwards (1967) classifies the

forests in this part of Natal from sea level to 500 m

as Coast Lowlands Forest, and from 500 to I 200 m

above sea level as Semi-Coast Forest.

The four forest areas discussed in this paper are

representative of the few remaining remnants of a

once much more extensive and perhaps continuous

forest along the coast dunes, in the lower valleys, and

on east- and south-facing slopes inland to about the

500 m contour level. Approximately 90 percent of

this forest type has been destroyed since the arrival of

Bantu and European settlers in the last 400 to 500

years (Brookes and Webb, 1965).

The best climatic data available for the stands are

statistics from Durban (Anonymous, 1954). The mean

daily temperatures are 24 °C in February and 16,5 C

in July, with an annual mean daily temperature of

20,5 °C. The absolute minimum and maximum

temperatures are 4 °C and 42 °C, respectively. Light

frost is known to occur occasionally in the Krantz-

kloof Forest, but the other forests may be considered

frost-free. Average precipitation for Durban ranges

from 28 mm in July to 121 mm in November, with a

yearly average of 1 008 mm. These precipitation

figures do not take into account the high humidity

throughout the year, resulting in heavy dews in winter.

COMPLEXITY AND HETEROGENEITY

The number of species of vascular plants in a

region gives an indication of the potential com-

plexity of its plant communities. South Africa has

an estimated 16 000 to 18 000 species of which some

5 000 species occur in Natal (Ross, 1972). Six hundred

to 650 of the Natal species are woody, i.e., trees,

shrubs and lianes (Ross, 1972; Ross and Moll, 1972).

Natal has rugged and diverse topography, rising

from the Indian Ocean to over 3 300 m on the

Drakensberg Range within a distance of 160 km.

The many slopes and aspects, and the diverse climatic

conditions associated with mountainous terrain lead

to many microclimates. With the many species from

which to select, the plant communities are extremely

complex.

Since Natal forests are virtually unstudied quan-

titavely, and yet occur in such an interesting situation

regarding topography, climate, and numbers of woody

species, it was considered desirable to attempt a

quantitative study. The 10 stands of Subtropical

Coast Forest studied probably represent some of the

most complex forests in Natal. If rapid quantitative

methods could be devised which work in this region,

a broader comparative study of the forest vegetation of

Natal could be undertaken. The aim of these quanti-

tative methods was to sample and describe a larger

area of forest, rather than to describe only one plot at

a time, as has usually been done in studies of tropical

and subtropical forest. A further aim was to assess

the degree of complexity and heterogeneity. The

samples consequently represented the vegetation of

one or two hectare, which is still much smaller than

those Curtis (1959) and his associates sampled in the

temperate forests of Wisconsin where their samples

usually represented vegetation of 5 to 8 hectare.

FIELD METHODS

The following criteria were followed in the selection

of stands. The stands were in upland situations free

from standing water at any time of the year, and were

free from any recent disturbance by man or domestic

animals. The prime criterion for selection was

topographic homogeneity; an area of one to two

hectare in size with the same slope, aspect and soil

parent material was selected. The vegetation had

“visual homogeneity” from a structural standpoint, in

that the age or size classes of species, the physiognomy,

and the density were relatively uniform throughout

the stand. It was also determined that certain of the

prominent tree species did occur throughout the

stand. However, an actual estimate of vegetational

homogeneity could not be determined until analysis

of the samples. There was no attempt to study the

full range of successional stages and all the stands

sampled were relatively mature.

Each stand was sampled with five 1/10 acre circular

plots. A 1/10 acre plot has an area of 4 356 square

feet, and therefore a radius of 37,2 feet.* The worker

who recorded the data (Rogers) stood in the centre

of the plot and determined the outer perimeter of the

circle using a field rangefinder. A rope marked with

the exact radius was also used in dense situations, and

for trees located near the perimeter. The other worker

(Moll) identified the species, and measured them

with the aid of an assistant, Bernard Mkhize.

Two categories of woody species were recorded.

Saplings and large shrubs were those individuals

having a diameter at breast height (4,5 feet) of one

to four inches. Trees, measured with a basal area

tape giving readings in square inches, were those

individuals having a d.b.h. of more than four inches,

i.e., a basal area of more than 12 square inches.

The basal areas of occasional trees with buttresses at

4,5 feet were measured above the butresses. Basal

area of the tree trunk is a common measure of

size or dominance (Curtis, 1959), and is an easier

and more objective measure than height or canopy

spread in a closed forest. Herbs, lianes, epiphytes

and small shrubs were not sampled.

The information recorded in the field for each

stand, therefore, was altitude, slope, aspect, soil

parent material, and measurement or listing of woody

species of over one inch in d.b.h. In addition, a soil

sample was collected in each of the five plots per

stand; these were then combined to give one soil

sample for each stand.)

RESULTS AND DISCUSSION

It would have been possible to assess each tree

regarding its presence in the canopy or other sub-

ordinate layers. However, whether a tree is in the

canopy or not is relative and varies with site, topo-

graphy, age of the stand, which other species are

present and interpretation of the observer. Therefore,

canopy individuals were characterized arbitrarily as

those having a basal area of 100 square inches or

more. This size was not only convenient, but gave a

reasonable separation. It was empirically determined

in the stands that trees exceeding this size class

formed the canopy layer a remarkable number of

times. Trees having a basal area of less than 100

square inches were classified as subcanopy. Those of

less than 12 square inches were listed as understorey,

a category which includes saplings of canopy and

subcanopy trees and large shrubs. Using these

criteria we found a good correlation in our study with

the same species subjectively assigned to the various

layers by Henkel (1934) and Moll (1967).

In Table 2, the data recorded for the canopy,

subcanopy and understorey species are shown for

various layers. The order in which the stands are

listed is based on an ordination which will be discussed

later.

* Feet and acres have been retained here because the work was

completed before metrication, and because comparisons of

basal areas (in square inches) were to be made with previously-

published data from temperate forests.

527

^.-Number of individuals per acre and basal area per acre, m square inches, for 77 species, by layer, in the 10 stands (numl*:r) r „„„ .

X signifies that a species was present in the stand, but was not recorded in the sample. Not included in this table 24 <b“*‘ *"* des,goated by C- s and u respectively. Various totals and averages for layers as well

I each stand, is indicated by parentheses p 01 ,nc,uaea in this table are 24 species which occurred i

Species

canopy species

Podocafpus latifolius.

Ficus natalensis

j, Sapium ellipticum

I Drypetes gerrardii

j, Commiphora woodii

■ Maytgnus acuminata .....

7s. M. peduncu/ariy. ~ \

8, Brachylaena uniflora

7. Combretum kraussii. ......

!l). Erythroxyluni pictum

11. Protorhus longifolia

'■ Harpephyllum caffrum

d Manilkara discolor

•4. Rhus chirindensis

*'■ Apodytes dimiata

^ Canthium ventosum

17' Scolopia zeyheri

Trichilia dregeana

I9' Chrysophyllum viridiflorum.

Strychnos henningsii

2*' Co,a natalensis

i

Chaetacme aristata ....

Celtis africana

24‘ Mil,ettia grandis

25' Cordia caffra

26' Canthi“m obovatum.

26700

2

340

(22

(4492

4

720

in one stand only, although they are included in the totals. These species are listed in the appendix. For each of the three categories of species,

as percentage relative dominance are also shown.

the dominant in each layer.

Stand 9

U

Stand 10

2

520

2

236

J8L

(2274)

2

214“

2

120

8

360

8

520

2

200

6

118

8

298

2

90

2

460

2 (38

1320 (2472

2

112

4

156

2

10'

2

10

2

10

2

10“

_2 1

10)

2

236

(24)

(4822)

2

268

2

1600

10

732

6

446

6

180

24

1062

6

350

2

174

2

66

2

28

2 (44)

380 (2158)

U

Stand 4

S

U

Stand 5

S

U

Stand 6

S

U

Stand 7

S U

_2_

10

4

20

2

10

(40)

2

10

4

670

2

122

22 (68

3546 (3976

2

748

8

3000

2

10

12)

60)

Stand 8

U

2

236

(18

(2968

6

1850

2

76

88

6786

8)

40)

o

52*

TABLE 2.— (Continued)

Species

Stand 1

c s u

Stand 3

27. Albizia adianthifolia

28. Croton sulvatious

29. Strychnos usambarensis. . . .

30. S. madagascariensis

31. S. decussata

32. Dovyalis longispina

33. Ziziphus mueronata

34. Cavacoa aurea

Totals: Canopy species..

SUBCANOPY SPECIES

1. A I lop hy l us dregeanus

2. Eugenia zu/uensis

3. Ochna arborea

4. Pavetta lanceolata

5. Schrebera alata

6. Clerodendrum glabrum

7. Raw sonia lucid a

8. Anastrabe integerrima

9. Eugenia natalita

10. Cryptocarya woodii

11. Gardenia amoena

12. Psycho tria capensis

13. Vitellariopsis marginata

14. Maerua racemulosa

15. Bequaertiodendron natalense.

16. Rothmannia globosa

17. Xylo theca kraussiana

18. Euclea na tale ns is

28

5552

134

6176

118

590

2

28

2

50

6

126

(32)

(1774)

16

426

_2

10

30

150

10

50

_8

40

12

60

_2

10

2

10

40

200

32

160

2 (54)

30 "(570)

34

170

2

30

16

80

2

30

40

200

10

50

6

30

2

10

60

12442

26

900

6

216

2

46

14

398

(132

Stand 4

2_

10

6 10

1700 516

2

580

22 56 34

110 10892 1926

6

TABLE l.— (Continued)

Species

% Taberaemontana ventricosa.

2a Cassipourea genardii

i| Mimusops o bo vat a

2

10

JS_

40

ii

90

jl Baphia racemosa

. 23, Dry petes arguta

24. Cassine aethipoica

25, Oricia bachmannii

1 26, Cussonia sphaerocephala

\ 27. Turraea flloribunda

28. Cassipourea gummiflua . .

29. Deinbollia oblongifo/ia. . .

30. Diospyros na tale ns is. . . .

11. Linociera peglerae

12. Ochna natalitia

33. Drypetes natalensis

34. Sapium integerrimum.. . .

35. Vangueria chart ace a

36. Teclea gerrardii

Stand 1

c s u

Stand 3

c s u

16

584

2 2

52 10

6 2

146 10

4

20

Stand 2

c s u

1_

10

c

Stand 9

S

10

262

14

368

16

280

6

136

6

142

2

252

24

120

8

40

20

100

(88)

(440)

16

80

22

110

37. Suregada africana

Totals: Subcanopy species.. . .

understory species (shrubs)

U Tricalysia capensis

^ Cassine papillosa

Erythroxylum emarginatum

^ Cussonia nicholsonii

Dracaena hooker iana

74 348

2664 1740

(372)

(1860)

_1_

10

180 248_

4522 1240

(128)

(640)

2

10

100 _l_46^

3372 730

(50)

(250)

136 498

4240 6490

28

140

4

20

24

120

(84)

(420)

C Pane

°via golungensis .

Totals: Understorey species. .

Stand 10

C

U

Stand 4

c s u

4

48

29

284

12

60

48

240

(186)

(930)

6

30

6

30

2

128

2

To"

(50)

(1146)

2

28

78

1914

426

2130

16

80

X

112

2480

18

90

(24)

(120)

100

500

2

10

46

230

4

20

4

20

4

20

2

10

8

40

212

1060

(61

(30)

Stand 5

C S u

6

30

Stand 6

c s u

Stand 7

U

Stand 8

C S

(2)

(200)

2

194

2

38

2

104

8

536

U

2

To"

04)

(70)

12

60

2

10

6

30

2

10

2

10

40

200

(20)

(100)

20

100

TABLE 2. — ( Continued )

DILWYN J. ROGERS AND EUGENE J. MOLL

531

The number of trees per acre in the canopy ranged

from 26 to 82 in the 10 stands. Trees per acre for the

canopy species in the subcanopy layer ranged from

22 to 134, while trees per acre of subcanopy species

in the subcanopy layer ranged from eight to 180,

etc.

Summary totals and averages for the layers, and

for each stand as a whole, are listed at the bottom of

Table 2. The total number of species appearing in

the various layers ranges from 20 to 40 species per

stand.

Basal area as a measure of relative dominance

Relative dominance, based on basal area figures, is

used as a measure of importance because it bears a

fairly direct relationship to the amount of light

intercepted or shade cast, the amount of water and

nutrients taken up, photosysthesis, litter dropped,

etc., which are underlying reasons for attempting to

assess importance of the individual species.

Note then at the bottom of Table 2 that although

the total number of stems per acre ranges from

268 to 1 116 (with an average of 658 ±60%), the

total basal area per acre ranges only from 16 438 to

24 514 square inches (an average basal area of

19 714 ±20%). This same pattern was found in a

study of 86 stands of deciduous forest in southern

Wisconsin which showed all degrees of cutting from

heavy to light (Rogers, 1959), where the number of

stems per acre ranged from 231 to 1 454, while the

average basal area was 18 391 ±25%. It appears that

there may be some phenomenon similar to “carrying

capacity” operating in temperate and subtropical

hardwood forests, i.e., as a forest matures from n

large number of small stems to a smaller number of

larger stems, the amount of ground actually occupied

by the trees does not change nearly as much as the

number of trees. Furthermore, the approximate

average of 20 000 square inches is measured at

breast height. Measurement of the basal area of a

tree at ground level is usually a little over twice as

great as at breast height, i.e., a basal area figure at

ground level would approximate 40 000 square

inches per acre. With approximately 6,3 million

square inches to an acre, this means that less than

one per cent of the ground is actually occupied by the

trunks of saplings and trees. These statistics are

mentioned to illustrate that there may be some

biological principle involved in basal area figures.

If this is true, the use of basal area figures as an

indication of relative dominance for species would be

more than a convenience, and, in fact, might have

some absolute significance.

The canopy species vary in relative importance

from 61% in stand 9 to 97% in stand 7. In five of

the 10 stands (1,9, 10, 7, 8). canopy species dominate

in the subcanopy layer. Only in one stand (stand 1)

do understorey species, i.e., shrubs, dominate the

understorey layer. In every other stand it is the

subcanopy species which predominate in the under-

storey layer. Using basal area as an indication of

importance, the canopy layer is shown to be the

dominant layer in all but stands 1 and 9. If more

stands covering a broader successional range of

forests were studied, it is possible that these relative

dominance values could be used to indicate

successional trends. On the basis of relative dominance

of canopy layers and canopy species, for example,

stands 1, 5, 6, 9 and 10 may be more immature or

pioneer stands, whereas stands 2, 3, 4, 7 and 8 may

be more mature or climax stands. It is not implied,

however, that these stands are part of the same

successional series.

In Table 3 the dominant species of each layer for

each category are shown. The figures indicate impor-

tance as a percentage of the total basal area for the

stand, and are determined as shown in the following

example : Protor/ws longifolia is the dominant canopy

layer tree in stand 1 , having a basal area of 4 492 of

the total 5 552 square inches (see Table 2); 4 492 is

23,9% and 5 552 is 29,5% of 18 792, the total basal

area for stand 1. By comparison of Tables 2 and 3,

the importance of each species in relation to the

layer it dominates and to the stand as a whole can

be seen.

TABLE 3. — Dominant species of each category for each layer. The basal area of each species as a percentage of the total basal area for

each stand is used as the indication of importance. The species indicated by initials can be determined by reference to Table 2. Also

listed are the three leading species for each stand with their relative importance values totalled. To indicate trends for the most important

species the fourth leading species is shown for stands 4, 6, 7 and 8; the fifth leading species is shown for stand 5 and the sixth for stand 2

26700-1 1

532

A QUANTITATIVE DESCRIPTION OF SOME COAST FORESTS OF NATAL

TABLE 3. (continued)

Also listed in Table 3 are the three leading species

of each stand, with a fourth leading species sometimes

shown to emphasize certain trends. Only the first

three are totalled in each case, however, to keep the

figures on an equivalent basis. Note that Cola nata-

lensis in stand 8 with a relative importance of 48,8%

is the most dominant species of any stand. Note also

the totals for the three leading species range from 38%

in stand 9 to 71% in stand 8.

Ordering of the stands

The index of similarity of Czekanowski, as discussed

(1959), was used to compare the similarity of each of

the 10 stands with every other stand. The comparison

was done on the basis of species present in the stands

using the 101 species which were found in the sample.

The formula used is ■■ . where a and b are the

a+b

numbers of species present in each stand, and w is

the number common to both stands (Curtis, 1959).

Stands I and 8 had the lowest similarity and were,

therefore, the end stands. The other stands were

arranged according to their similarities to the end

stands, and the resulting order is used in Tables 2

and 3 (see Table 4).

Not enough stands were studied for a valid two-

dimensional or three-dimensional ordination. How-

ever, it can be seen that the groups of stands which

have the highest indices of similarity with one another

are those within one forest area. For example, stands

1, 2 and 3 of forest area A have more in common

with each other than with any other stands, as do

stands 9 and 10 of area B, etc.

The values for the stands within each forest area

compared to every other forest area were then

averaged, and an average similarity within and

between the forests is shown in Table 5. Stands 1,

2 and 3 of area A average 52% similarity with one

another, stands 9 and 10 of area B are 61% similar,

etc., whereas forest area A averages 40% similarity

with area B, etc. Table 6 shows that whereas areas

A and B average 40% similarity, and areas C and D

average 37% similarity, areas A and B average only

22% similarity with areas C and D.

TABLE 4. — Percentage similarity calculated by - — , of the

a H- b

ten stands with one another using 101 species

DILWYN J. ROGERS AND EUGENE J. MOLL

533

TABLE 5. — Average percentage similarity of the four forests

with one another

slopes, compared to (CD), the stands along the coast

on dune sand and generally on more gradual north-

and east-facing slopes (see Table 1).

Patterns of species distributions

Strychnos madagascariensis is one of three leading

species in forest areas C (stands 4-6) and D (7-8),

and Celtis africana is one of the four leaders in these

two areas, except for stand 5 in which it was fifth.

Cola natalensis and Cavacoa aurea are dominants in

area D (Fig. 4), but are absent from area C. Con-

versely, Trichilia dregeana and Bec/uaertiodendron

natalense are dominants in area C (Fig. 5), but are

absent from area D.

Species patterns are less distinct in areas A (stands

1-3) and B (9-10), although Protorbus longifo/ia is

first or second in each stand, except stand 2 in which

it is sixth. Tricalysia capensis, a shrub present in

seven of the 10 stands, actually achieved a rank of

third leading dominant in stand 1 (Fig. 6). Millettia

Fig. 4. — View inside Hawaan

Forest (area D). The two

dominant species having

a combined relative im-

portance of 55 to 60%

in stands 7 and 8 were

Cavacoa aurea (right) and

Cola natalensis (left

centre).

Fig. 5. — View inside Hlog-

wene Forest (Area C).

Ekebergia capensis (left)

has a some what buttressed

base. Most of the smaller

trees are Bequaertioden-

dron natalense, one of the

four leading dominants

in this forest, having an

average relative impor-

tance in the three stands

(4, 5, 6) of approximately

15%.

We may therefore conclude that the first order of

similarity is within one forest area (1, 2, 3) (9, 10)

(4, 5, 6) (7, 8). The second order of similarity is (AB),

the stands farther inland, altitudinally higher, on

rocky soil, and in a topographically more mesic

situation with generally fairly steep southwest-facing

grandis appears as a dominant in stands 9 and 10,

the only stands in which it is ever present. Podocarpus

latifolius, a very conspicuous tree in the forest, is

present only in stands 1-3, and is a dominant only

in stand 2 (Fig. 7).

534

A QUANTITATIVE DESCRIPTION OF SOME COAST FORESTS OF NATAL

Fig. 6. — View inside Krantz-

kloof Forest (area A,

stand 1). The dark-boled

trees are Protorhus longi

folia, perhaps the most

characteristic dominant

species of areas A and B.

Most of the shrubs are

Tricalysia capensis, which

was very common in this

stand.

Fig. 7. — View inside Krantz-

kloof Forest (area A).

The three leading

dominants in stand 2 are

shown. The large tree is

Podocarpus latifolius, the

smooth-barked trees are

Drypetes gerrardii and

the small trees with fluted

trunks are Bequaertio-

dendron natalense. Note

the presence of sandstone

rocks fallen from cliffs to

the left (see Fig. 2) and

the cycad, Encephalartos

villosus, in the ground

layer.

Strychnos madagascariensis is one of the three

leading dominants in six of the 10 stands studied and

may therefore be considered as the single species

most likely to be a dominant in these Coast Forests.

Bequaertiodendron natalense is present in eight of the

10 stands, the highest frequency for any species.

Only two species appear in all four forest areas how-

ever: Chaetacme aristata and Tricalysia capensis.

The distribution patterns of other species may be

seen in Table 2. The number of species of the various

layers as appearing in the two main groups of stands,

the inland, rocky areas (A and B) and the coast dunes

(C and B), may be summarized as follows:

Although there are definite patterns in the distri-

bution of species in the various stands and forest areas,

no definite groups of species with identical patterns are

obvious. It appears likely that the behaviour

indicates a continuum rather than an association

interpretation. Further studies covering a wider

range of successional conditions are desirable before

we can fully understand these distribution patterns.

Soil analysis

From an analysis of the soil samples there were a

few trends apparent within the groups of stands.

The soils of areas A and B had more coarse sand, more

clay, lower pH (4,2 to 5,0), and higher water-

retaining capacities, while those of areas C and D

had more medium and fine sand, higher pH (6,0 to

7,0), and lower water-retaining capacities. The soils

of forest area D had the highest total bases, but the

results for the other areas varied with the stands.

Area D, likewise, had the highest cation exchange

capacity, followed by areas A, C and B. Results

concerning specific nutrients varied from stand to

stand with no real pattern. If a larger number of

stands were studied, it is possible that presence or

absence of certain species might be correlated with

certain nutrients.

Heterogeneity and homogeneity of stands

The terms heterogeneity and homogeneity are

relative. It is possible to prove that a stand is hetero-

geneous, but it is unlikely that a stand of any size can

be demonstrated to be homogeneous. Consequently,

we can speak only of higher or lower degrees of

heterogeneity.

DILWYN J. ROGERS AND EUGENE J. MOLL

535

Table 7 gives an indication of the relative degree of

heterogeneity of each stand. Some stands may be

considered less heterogeneous than others. The left

hand columns show the decreasing rate at which new

species were sampled in the successive plots of a

stand. While the figures are similar to a species area

curve, it is realized that chance may have influenced

limits to recognize the layers, rather than deter-

mining each individual subjectively in the field, the

time spent in sampling a stand is shortened. Likewise,

the relative contribution of the canopy, subcanopy

and understorey species to the various layers can be

readily assessed and is on a comparable basis for

all stands.

TABLE 7. — Relative degrees of heterogeneity for the stands on the basis of species appearance in the sample plots (see text for

discussion)

the results. However, the results are intended to show

that there is probably some degree of vegetational

unity in each stand. Completely heterogeneous

stands would perhaps show a relatively constant

rate of encountering new species in successive plots.

With this interpretation, stand 7 may be more

heterogeneous than the others.

The right hand columns of Table 5 illustrate the

frequency with which species appeared in the plots

of each stand. Relatively more species having high

frequencies indicates relatively less heterogeneity.

Stands 1, 9 and 10 could be considered as the least

heterogeneous stands by this method. A high per-

centage of species occurring in one plot only would

imply that a stand is relatively more heterogeneous, as

is stand 2 by this interpretation.

A relatively large number of species may indicate

complexity for a stand, but does not necessarily

indicate heterogeneity. For example, stand 1 had

40 species, yet seems less heterogeneous than most

stands as 55% of its species have a frequency of

60% or more. Conversely, stand 7 has only 22 species

yet may be more heterogeneous than most stands.

CONCLUSIONS

The aim of this paper was to demonstrate the

feasibility of applying quantitative phytosociological

techniques in fairly complex subtropical forest. From

the results obtained it is apparent that such a quan-

titative description is feasible and yields useful

results.

The phenomenon of layering, or stratification, is

generally considered to be more pronounced in

tropical and subtropical than in temperate forests.

Although discontinuous layers were not looked for

in this study, the species have inherent growth

properties which permit or do not permit them to be-

come large canopy trees. Stratification, as a structural

characteristic of these forests, was described by

using specific size limits to classify species as to those

commonly found in the canopy, subcanopy, or

understorey. In the 10 stands, a total of 101 species was

found, of which 38 were classified as canopy species,

49 as subcanopy species, and 14 as understorey

species (i.e., shrubs). By using objective, indirect

Basal area in square inches per acre was used as an

indication of the relative dominance of the species

and of the various layers. Basal area is an objectively-

determined measure which reflects both the size and

the number of trees, and each species and layer can

be expressed as a percentage of the total basal area of

the stand. The relative basal area figures are con-

sidered to give an approximation of the relative

influence on the environment of the various species

and layers.

The subtropical forests of the Natal coast were

found to be complex, having a large number of species,

genera, and families, and this complexity is enhanced

by layering (Table 2 and Appendix). The forests were

al o found to be fairly heterogeneous (Tables 2 and 3).

There are definite trends in the distribution of species,

with certain species tending to occur together with

other species in similar habitats (Table 2). There is,

therefore, a pattern in the variations from stand to

stand and from area to area as illustrated in the

similarity figures (Tables 4, 5 and 6). Only a few species

are relatively important in each stand, and many of

the same species are dominant in several stands (Table

3).

The 10 stands studied were in four forest areas.

In each case the stands within one forest area had the

highest degree of similarity with each other. Two of the

forest areas were located on coast dunes and the

other two were farther inland, altitudinally higher, and

on steeper, rocky slopes. As would be expected, the

dune forests were more similar to one another, as were

the inland forests to each other. We may, therefore,

conclude that the first order of similarity was geo-

graphical proximity, and that the second order was

topographic and soil similarity (Tables 4, 5 and 6).

Protorhus longifolia was the dominant canopy species

most characteristic of the inland forests, while

Strychnos madagascariensis and Celtis africana were

most characteristic of the dune forests. There is an

indication that the inland and dune forests represent

two different communities, and that the five stands

in each community represent some sort of successional

trends. However, of the 77 species occurring in more

than one stand, 43% of them occurred in both inland

and dune stands, and no indication of discontinuous

species groups or associations was noted.

535

A QUANTITATIVE DESCRIPTION OF SOME COAST FORESTS OF NATAL

An interpretation of the behaviour of the species

is that of all the species present in a stand only a small

percentage of them is dominant and present

throughout the stand, while a large percentage

occurs in very small numbers. The relatively large

number of individuals appearing only occasionally

in the stands leads to the interpretation of the stands

as being heterogeneous. However, as seen in Table 3,

only two or three species are required to attain more

than 50°/ dominance in eight of the stands, com-

pared with the total of 20 to 40 species found in each

of the 10 stands. Four species are necessary in stand 2

and six species in stand 9 to attain 50% dominance.

It is likely that this same situation regarding dominants

prevails in temperate zone forests, with two or

three species being dominant. However, in temperate

forests, only a few other species would constitute the

remainder of the stand, compared with these sub-

tropical forests in which a large number of species

constitute the remainder.

The sampling method described in this paper is

relatively rapid in the field and requires a minimum

of field equipment (a range-finder and a basal area

tape). The analytic methods are relatively simple, yet

much information is derived. The use of several

smaller plots is preferable to one large one in that

this avoids basing judgement on one atypical area

and allows for variability within a stand. Plots of 1/10

acre are not too large to use and simplify the analytic

methods. Circular plots have a smaller perimeter-

to-area ratio than any polygon with the same area,

which helps reduce heterogeneity. By use of a range-

finder to delimit the plots, the difficulty of staking out

quadrats is avoided. Sampling of smaller shrubs and

herbs could easily be incorporated into the study by

using several small quadrats per plot. With data from

a larger number of stands, a statistical treatment of

the forest communities of a larger region could be

achieved.

The first criterion of stand selection was topo-

graphical homogeneity, and then visual vegetational

homogeneity from a structural standpoint. Sampling

and analysis of these stands then followed, and we

believe that floristically defined vegetation units have

been demonstrated to exist despite some hetero-

geneity. We therefore conclude that the analysis

was justified and that the methods described offer the

possibility of a broader comparative study of the

relatively complex subtropical forests of Natal.

ACKNOWLEDGEMENTS

This study was conducted while the senior author

was visiting lecturer in botany at the University of

Natal and visiting biologist with the Natal Parks,

Game and Fish Preservation Board. Financial

assistance, and the use of office space and facilities,

are gratefully acknowledged. Appreciation is expressed

by the authors to Mkhize for field assistance and to

Hulett’s Sugar Co., Natal Estates (Pty) Ltd, and the

Natal Parks Board for permission to carry out these

studies in areas under their jurisdiction.

UITTREKSEL

Tien subtropiese worn! stande is in vier gebiede

Icings die kus van Natal, in die Republiek ran Suid-

Afrika, geinonster. Vyf 1/10 acre sirkelvormige persele

binne elke stand is hiervoor gebruik. 'n Groottotaal

van 101 houtagtige spesies met stamdeursnee groter

as een duim d.b.h. is aangeteken, met 7/ variasie van

20 tot 40 verskillende spesies per stand. Kwantitatiewe

resultate, met inbegrip van aantal en grootte, word

gegee vir die kroon, sub-kroon en onderkroon spesies

soos Indie in die verskillende strata voorkom. Spesifieke

grootteklasse is gebruik om die verskillende strata te

identifiseer om sodoende al die stande op 'n vergelykbare

basis te plaas. Getalle van relatiewe basala bedekking

per acre is gebruik as 7? aanduiding van die relatiewe

dominansie van die verskillende spesies en strata.

Die moontlikheid van 7? biologiese beginsel om hierdie

gebruik te regverdig word vermeld.

Ten spyte van die komplekse struktuur van die

stande asook die feit dat hulle heterogeen voorkom,

toon die spesies van die verskillende stande nogtans

soortgelyke gedragspatrone. Die ontwikkelingstendense

vir die 10 stande en vier woudgebiede word aangedui.

R latief min spesies kom dominant in elke stand voor

en die oenskynlik hoe floristiese verskeidenheid is

grotendeels toe te skryf aan die hoe persentasie spesies

wat relatief skaars is. Die metodes wat hier beskryf

word, behoort bruikbaar te wees in "n studie van die

breer verspreidings-gebied van woude in Natal; 7?

gebied waar kwantitatiewe studies tot non toe feitlik

ontbreek het.

REFERENCES

Acocks, J. P. H., 1953. Veld types of South Africa. Mem. Bot.

Surv. S. Afr. 28.

Anonymous, 1954. Climate of South Africa. Part 1. Climate

Statistics. Pretoria: Weather Bureau.

Brookes, E. H. & Webb, C. DE B, 1965. A History of Natal.

Cape Town: Oxford University Press.

Cottam, G. & Curtis, J. T., 1956. The use of distance measures

in phytosociological sampling. Ecology 37: 451-460.

Curtis, J. T., 1959. The vegetation of Wisconsin. Madison:

University of Wisconsin Press.

Edwards, D., 1967. A plant ecological survey of the Tugela

basin. Mem. Bot. Surv. S. Afr. 36.

Henkel, J. S., 1934. A field book of the woody plants of Natal and

Zululand. Durban and Pietermaritzburg: Natal University.

Moll, E. J., 1967. Forest trees of Natal — a field guide to 200

species. Durban: Wildlife Protection and Conservation

Society of South Africa.

Moll, E. J., 1968a. A quantitative ecological investigation of

the Krantzkloof Forest, Natal. /. S. Afr. Bot. 34: 15-25.

Moll, E. J., 1968b. An account of the plant ecology of the

Hawaan Forest, Natal. J. S. Afr. Bot 34: 61—76.

Moll, E. J., 1969. An Investigation of the plant ecology of the

Hawaan Forest, Natal, using an ordination technique.

Bothalia 10: 121-128.

Phillips, E. P., 1951. The genera of South African flowering

plants. Mem. Bot. Surv. S. Afr. 25.

Rogers, D. J., 1959. Ecological effects of cutting in southern

Wisconsin woodlots. Ph.D. thesis. University of Wisconsin

(Unpublished).

Ross, J. H., 1972. The flora of Natal. Mem. Bot. Surv. S. Afr. 39.

Ross, J. H. & Moll, E. J., 1972. A List of Natal trees. Durban:

Natal Branch of the Wildlife Protection and Conservation

Society of South Africa.

APPENDIX

A total of 101 species in 78 genera and 39 families was sampled

in the 10 stands and the taxa are listed in systematic order

(Phillips, 1951 ). The 24 species which were omitted from Table 2

are included here, and in parentheses after each of these species

is the letter C, S, or U, indicating the layer, and a number

referring to the stand in which it occurred.

podocarpaceae: Podocarpus latifolius (Thunb.) R. Br. ex Mirb.

Agavaceae: Dracaena hookeriana K. Koch

Ulmaceae: Celtis africana Burm.f., C. gomphopyhlla Bak.

(C-2), Chaetacme aristata Planch.

Moraceae: Ficus capensis Thunb. (C-2), F. natalensis Hochst.

Lauraceae: Cryptocarya woodii Engl.

Capparaceae. Maerua racemulosa (A. DC.) Gilg & Ben.

Leguminosae: Albizia adiantliifolia (Schumach.) W. F. Wight,

Baphia racemosa (Hochst.) Bak., Millettia grandis (E.

Mey.) Skeels.

Erythroxylaceae: Erythroxylum emarginatum Thonning,

E. pictum E. Mey. ex Sond.

Rutaceae: Fagara davyi Verdoorn (S-l), Vepris undulata

(Thunb.) Verdoorn & C. A. Sm. (S— 10), Teclea gerrardii

Verdoorn, Oricia bachmannii (Engl.) Verdoorn.

Burseraceae: Commiphora harveyi (Engl.) Engl. (S-9), C.

woodii Engl.

Ptaeroxylaceae: Ptaeroxylon obliquum (Thunb.) Radik. (C-9).

Meliacf.ae: Turraea florihunda Hochst., Ekebergia pterophylla

(C. DC.) Hofmeyr (S-l), Trichilia dregeana Sond.

DILWYN J. ROGERS AND EUGENE J. MOLL

537

Euphorbiaceae: Drypetes arguta (Muell. Arg.) Hutch., D.

gerrardii Hutch., D. natalensis (Harv.) Hutch., Antidesma

venosum E. Mey. ex Tul. (S— 6), Croton sylvaticus Hochst.,

Acalypha glabrata Thunb. (U-3), Suregada africana (Sond.)

Kuntze, Sapinm ellipticum (Hochst. ex Krauss) Pax,

S. integerrimum (Hochst. ex Krauss) J. Leon., Cavacoa

aurea (Cavaco) J. Leon.

Buxaceae: Buxus natalensis (Oliv.) Hutch. (U-7).

Anacardiaceae : Harpephyllum caffrum Bernh., Protorhus

longifolia (Bernh.) Engl., Rhus chirindensis Bak.f. forma

legatii (Schonl.) R. & A. Fernandes.

Celastraceae: Maytenus acuminata (L.f.) Loes., M. pedun-

cularis (Sond.) Loes., M. undata (Thunb.) Blakelock (U-l),

Cassine aethiopica Thunb., C. papillosa (Hochst.) Kuntze.

Icacinaceae: Apodytes dimidiata E. Mey. ex Arn.

Sapindaceae: Allophylus dregeanus (Sond.) De Wint.,

A. melanocarpus (Sond.) Radik. (S-3), Deinbollia oblongi-

folia (Sond.) Radik., Pancovia golungensis (Hiern) Exell &

Mendonca.

Rhamnaceae: Ziziphus mucronata Willd.

Sterculiaceae : Cola natalensis Oliv.

Ochnaceae: Ochna arborea Burch, ex DC., O. natalitia Engl. &

Gilg.

Clusiaceae: Garcinia gerrardii Harv. ex Sim (U-2).

Violaceae: Rinorea angustifolia (Thouars) Baill. (S-9).

Flacourtiaceae: Rawsonia lucida Harv. & Sond., Xylotheca

kraussiana Hochst. var. glabrifolia Wild, Scolopia zeyheri

(Nees) Harv., Homalium dentatum (Harv.) Warb. (S-9),

Dovyalis longispina (Harv.) Warb.

Thymelaeaceae: Peddiea africana Harv. (U-l).

Rhizophoraceae : Cassipourea gerrardii (Schinz) Alston,

C. gummiflua Tul.

Combretaceae: Combretum kraussii Hochst., C. molle R. Br.

ex G. Don (S— 10).

Myrtaceae: Eugenia natalitia Sond., E. zuluensis Duemmer.

Araliaceae: Cussonia sphaerocephala Strey, C. nicholsonii

Strey.

Sapotaceae: Chrysophyllum viridi folium Wood & Franks,

Bec/uaertiodendron natalense (Sond.) Heine & Hemsl.,

Mimusops obovata Sond., Vitellariopsis marginata (N.E. Br.)

Aubrev., Manilkara discolor (Sond.) J. H. Hemsl.

Ebenaceae: Eudea natalensis A. DC., Diospvros natalensis

(Harv.) Brenan, D. scabrida (Harv. ex Hiern) De Wint.

(U-l).

Oleaceae: Schrebera alata (Hochst.) Welw., Linociera peglerae

(C.H. Wr.) Gilg & Schellenb., Olea woodiana Knobl. (S-4).

Loganiaceae: Strychnos decussata (Pappe) Gilg, S. henningsii

Gilg, S', madagascariensis Poir, S. usambariensis Gilg.

Apocynaceae: Carissa bispinosa (L.) Desf. ex Brenan (U-9),

Tabernaemontana ventricosa Hochst. ex A. DC., Rauvolfia

caff r a Sond. (C-3).

Boraginaceae: Cordia cajfra Sond.

Verbenaceae: Clerodendrum glabrum E. Mey.

Scrophulariaceae: Anastrabe integerrima E. Mey. ex Bcnth.

Rubiaceae: Xeromphis obovata (Hochst.) Keay (S— 10). Gardenia

amoena Sims, Rothmannia globosa (Hochst.) Keay,,

Oxyanthus latifolius Sond. (U-10), Tricalvsia capensis

(Meisn.) Sim, Vangueria chartacea Robyns, Canthium

ventosum{ L.) S. Moore, C. mundianum Cham. & Schlechtd.

(S-9),C. obovatum Klotzsch, Pavetta lanceolata Eckl.,

Psychotria capensis (Eckl.) Vatke.

Compositae: Brachylaena uniflora Harv.

Bothalia 11, 4: 539-560 (1975)

A west-east vegetation transect through Africa south of the

Tropic of Capricorn

B. J. COETZEE* and M. J. A. WERGERf

ABSTRACT

Changes in predominant vegetation physiognomy, prominent species and physiography along the route

following an 1 800 km transect across southern Africa near the Tropic of Capricorn, are described. Eleven main

discontinuities in the structure and floristic composition of the vegetation along the transect are related to a

climatic gradient across the Continent. Floristic variation within the main structural types is largely related to

rainfall, severity of frost, soil conditions, exposure, slope and aspect. The main vegetation classes distinguished

coincide largely with major differences in carrying capacity of the vegetation.

INTRODUCTION

Southern Africa is mainly a fairly flat, high plateau

rising steeply from a narrow coastal strip. These

topographical features, together with the cool and

warm ocean currents flowing on the west and east

coasts respectively, have a strong bearing upon the

climate and consequently upon the vegetation of the

area. Although in climate the western coastal strip

differs markedly from the eastern one, both are

strongly oceanic, whereas the plateau is distinctly

continental. On the plateau, the climate also changes

markedly in the west-east direction. Obviously, these

climatic and physiographic differences influence

floristic composition and struc ure of the vegetation.

On a general scale close correlations between

climatic features, such as precipitation, temperature

and humidity on the one hand and differences in

physiognomy and prominent species of the vegetation

on the other, may be expected. In order to establish

these correlations an 1 800 km transect was taken

across the continent 1° 30' to 4° south of the tropic

of Capricorn, starting in the coastal Namib desert in

the west and proceeding through the steppes, grass-

lands, savannas and woodlands of the plateau, down

to the tropical dense woodlands near the east coast.

The path of the transect shown in Figs. 1-12 was

followed by road and conspicuous changes in pre-

dominant physiognomy, prominent species and

physiography noted. Physiognomic fea'.ures recorded

were cover, height and dominant growth forms of

the different strata.

The physiognomic-floristic vegetation classes thus

distinguished were related to climatic features recorded

in the literature, to topographic maps and to physio-

graphic features recorded along the transect. Habitat

features which were related to more detailed variations

within the vegetation classes described here were

obtained from reports of local studies.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private BagXIOl, Pretoria.

t Botanical Laboratory, Sect. Geobotany, University of

Nijmegen, Toernooiveld, Nijmegen, Netherlands.

TOPOGRAPHY

Approximately seven-eights of the 1 800 km

transect across the continent passes through the

interior plateau and the western and eastern escarp-

ment zones, all of which are above 920 m altitude

(Figs. 1 and 2). The altitude decreases rapidly below

920 m to sea level over a distance of 80 km in the

west and over 130 km in the east. The western

escarpment and the central interior plateau attain

1 560 m and are separated by a low, virtually flat area

of the western interior, 380 km wide and approxi-

mately 925 m above sea level. The eastern interior

plateau lies between 925 and 1 250 m. The highest

altitudes are on the eastern escarpment which rises

to 2 350 m above sea level.

CLIMATE

The climates according to Koppen’s classification

(Fig. 3) are arid over the larger part of the transect

with the exception of a temperate rainy region (Cwa

and Cwb) occurring in the eastern escarpment zone

and a tropical rainy region (Aw) along the east coast

(Schulze, 1947). Several arid climatic types are

traversed, the main division being between the desert

climates, occurring west of a line crossing the transect

at approximately 700 km from the west coast, and the

Steppe climates occurring east of this line. A narrow

zone along the west coast, stretching approximately

40 km inland, is classified as BWk or BWkn'. The

latter is cold desert with frequent fog and with the

mean temperature of the hottest month exceeding

18 °C. Further inland, but still west of the escarp-

ment, the climate (BWh) is hot and dry with the

mean annual temperature exceeding 18 °C. As shown

in Fig. 4 the region west of the western escarpment

falls within the winter rainfall area. The western

escarpment and western interior plateau kave a cooler

desert climate and fall )within the summer rainfall

area (BWkw'). The remainder of the transect also

has summer rainfall. The upper, central interior

plateau (Fig. 1) falls in the regime of a cold and

dry steppe climate, but with the mean temperature of

the hottest month exceeding 18 °C (BSkw'). The

lower part of the eastern interior plateau and a

larger part of the area east of the eastern escarpment

have a hot and dry steppe climate, classified as

BShw (Figs. 1 and 3).

540 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

Fig. 1 .—Structural changes in the vegetation and altitudinal profile along the transect from Liideritz to Vila Luiza.

B. J. COETZEE AND M. J. A. WERGER

541

542 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

Fig. 4. — Winter rainfall (shaded) and summer rainfall areas of southern Africa (from Weather Bureau,

1957).

Fig. 5 shows that the percentage number of years

with frost increase rapidly from the coastal regions

(less than 25 per cent) towards the escarpments (100

per cent). Virtually the whole of the interior plateau

has frost occurring with greater or lesser severity

every year. Schulze (1965) divided southern Africa

into a number of regions, based on the temperature

regime over the country. From this work it may

be seen that the lowest minimum temperatures along

the transect occur on the upper parts of the interior

plateau between Van Zylsrust and Derby and between

Groblersdal and the eastern escarpment. The low

temperatures that occur in a narrow zone along the

eastern escarpment are not shown in Fig. 6. As shown

in Fig. 6, absolute minimum temperatures over most

of the interior plateau between Keetmanshoop and

Derby are between — 5 °C and — 10 °C. Frost is less

severe west of Keetmanshoop and north-east of

Derby (excluding the eastern escarpment region) and

frost-free zones occur on the lowlands bordering the

east and west coasts. Differences in frequency and

severity of frost between the Rustenburg-Groblersdal

area of the interior plateau on the one hand and the

lowland east of the high escarpment on the other,

mark an important distinction between these two hot

and dry steppe regions (BShw).

Fig. 5. — Percentage number

of years with frost

(minimum temperature

less than 0° C) (from

Schulze, 1965).

B. j. COETZEE AND M. J. A. WERGER

543

Fig. 6. — Absolute minimum

temperatures (° C) in

southern Africa (from

Schulze, 1965).

Mean daily maximum temperatures of the hottest

month (Fig. 7), show an inverse pattern to that of

frequency of frost and absolute minimum tem-

peratures. Mean daily maximum temperatures in

January are highest in the interior plateau areas where

frost is most severe and frequent, and they decrease

rapidly as one descends towards the coast. Mean

daily maximum temperatures for January near the

east coast exceed 20 °C and are higher than along the

west coast where this figure is 12,5 °C. These tem-

peratures exceed 27,5 °C over most of the interior

plateau (Fig. 7), and absolute maximum temperatures

exceed or approach 40 °C over the entire transect

(Fig. 8).

Fig. 7. — Mean daily maxi-

mum temperatures (° C)

for the warmest month

(January) in southern

Africa (from Jackson,

1961).

Fig. 8. — Absolute maximum temperatures (°C) in southern

Africa (from Schulze, 1965).

The normal annual rainfall of the desert regions

between the west coast and escarpment is less than

100 mm (Fig. 9). It increases gradually from 100 mm

on the western escarpment to 200 mm near the

eastern boundary of the desert climatic types, including

the western interior plateau. The reliability of the

rainfall, expressed as the deviation from the average

annual rainfall (Fig. 10), decreases rapidly from the

drier parts of the cold steppe region towards the

lower rainfall desert areas in the west so that in these

more arid regions rainfall not only becomes lower, but

is also less reliable. Further eastward the rainfall

gradually increases and becomes more reliable.

In the easternmost part of the cold steppe region of the

upper, central interior plateau and over the hotter

steppe climate of the lower, eastern interior plateau

544 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

the rainfall is between 600 and 700 mm. Precipitation

increases rapidly to approximately 2 000 mm in the

temperate rainy area of the eastern escarpment and

decreases again to between 600 and 700 mm in the

hot steppe lowland region east of the escarpment.

Another rapid increase occurs eastward towards the

tropical rainy region and the rainfall at the coast is

between 1 000 and 1 250 mm.

ON

2

Ph

■Mean annual rainfall in southern Africa (from Jackson, 1961).

B. J. COET2EE AND M. J. A. WERGER

545

Fig. 10. — Mean deviation as

a percentage of the

average annual rainfall

(from Weather Bureau,

1957).

There is an evident increase in rainfall from the

west coast to the east coast in which the effect of

altitude above sea level is apparent from local rainfall

peaks. Altitude above sea level also influences to a

certain extent the occurrence and severity of frosts.

In the desert climatic regions (BW) west of

Kuruman, where the rainfall is below 200 mm, all

12 months are dry, as defined by Walter & Lieth

(1960). East of this line the number of dry months

per year drop to five or six.

Humidity over any part of the transect, including

the western dry winter-rainfall area, is higher in

midsummer (Fig. 11) than in midwinter (Fig. 12). In

summer, as well as in winter, both coastal regions

have a higher humidity lhan the interior and the

humidity along the tropical rainy east coast is higher

than the humidity of the cooler desert along the west

coast at any particular time of the year.

15° 20° 25° 30° 35°

Fig. 11. — Mean humidity

mixing ratio for the

warmest month (January)

in southern Africa (from

Jackson, 1961).

Fig. 12. — Mean humidity

Mixing ratio for the col-

dest month (July) in

southern Africa (from

Jackson, 1961).

35°

20’

25°

35°

546 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

PHYTOGEOGRAPHY

The boundary between the Karoo-Namib Region

and the Sudano-Zambezian Region (Volk, 1966;

Werger, 1973) largely coincides with the boundary

between the desert climatic types (BW) and the

steppe climatic types (BS). On the eastern escarp-

ment the transect crosses through narrows bands of

the Afro-montane and Austro-afro-alpine Regions

(Chapman & White, 1970; White, 1971; Coetzee,

1967). In the tropical rainy area (Aw), near the east

coast, the transect crosses through the Usambara-

Zululand Domain of the Guineo-Congolian Region

(Chapman & White, 1970; White, 1971). All these

Regions are part of the African Palaeotropis.

VEGETATION

The true Namib desert extends from Liideritz for

about 100 km inland. Altitude increases from sea

level to about 1 250 m over this distance. The

vegetation in the Namib is generally extremely sparse.

Near the coast, where strong seaward and landward

prevailing winds alternate, the desert is a mosaic of

deep, coarse sandy patches, with shallow runnels and

eroded granitic outcrops. The outcrops are virtually

bare, but in the drainage lines a sparse vegetation of

dwarf shrubs and xerophytic grasses occurs. Main

species are Salsola aphylla, Psilocaulon marlothii,

Drosanthemum floribundum, Osteospermum crassi-

folium, Lycium decumbens, Sarcocaulon spinosum and

the grass Eragrostis cyperoides. Apart from the grass,

which is hard and coarse, all species are succulent in

either their leaves or their stems. Total cover of this

vegetation is below one per cent with plants up to

30 cm tall (Fig. 13).

On dune sand a sparse growth of Stipagrostis

sabulicola, around which small dunes are formed,

occurs locally (cf. Giess, 1962, 1971). On rocky places

a few kilometres inland, open stands of the virgate

shrub Lebeckia sp. are occasionally encountered.

Shrubs may be up to 3 m tall and cover less than one

per cent. The shrub Ectadium virgatum var. latifolium

with hard, coriaceous leaves, is also found here.

Further away from the coast, in the hot, dry desert

( B Wh ), where strong winds and fog gradually decrease,

extensive open stands of Sarcocaulon spinosum, up

to 0,10 m high and covering less than one per cent

are found (Fig. 14). On the slopes of the mountains

and sometimes on the gravelly plain open stands of

Fig. 13. — Rocky outcrops

near Liideritz showing

very sparse vegetation of

succulent dwarf shrubs.

Fig. 14. — Extremely open

stand of Sarcocaulon

spinosum in the southern

Namib, approximately 20

km from the coast.

R. J. COETZEE AND M. J. A. WERGER

547

Fig. 15. — Southern Namib

approximately 70 km in-

land, entirely bare of

vegetation.

the stem succulent Euphorbia gummifera, up to 1,5 m

in height, occur. About 50 km inland a plain which is

absolutely devoid of vegetation, is reached. This is

interrupted by occasional very sparsely vegetated

mountainous outcrops of coarse, gravelly, reddish

sand, which often have a hard surface layer (Fig. 15).

About a hundred kilometres inland, approximately

15 km west of Aus, the Vornamib starts with a low

grass steppe of about 5 per cent total aerial cover

(Fig. 16). The main species are Stipagrostis lanipes,

S. obtusa, S. ciliata, Ehrharta pusil/a and Eragrostis

nindensis. Sarsocaulon spinosum and Eriocephalus

pubeseens also occur occasionally. The soil consists

mainly of a rather coarse, sometimes gravelly sand.

Further eastward the shrub Dyerophytum africanum

becomes locally abundant and grass cover increases

to about 8 per cent. Acacia erioloba trees start to

occur here. Near Aus the total aerial cover of the

vegetation which, except for Stipagrostis species,

now also contains many other herbs, geophytes and

dwarf shrubs, has increased to about 15 per cent with

local patches reaching 25 per cent. In this vicinity,

120 km inland from Luderitz, a small, low escarp-

ment brings the general level of the surface to 1 500 m

(Fig. 1 ). The slopes of the escarpment and the isolated

mountains are vegetated mainly by dwarf shrubs,

about 0,40 m high, covering up to 20 per cent, with

isolated small trees of up to 3 m. The plains between

the mountains are covered with the 0,25 m high

grass steppe vegetation described above. The soil of

the plains consists mainly of yellowish and reddish

sand. This Vornamib or desert margin vegetation

falls within the cooler desert climate (BWlcw') with

sporadic summer rainfall up to approximately

100 mm, and continues to about 50 km east of Aus,

170 km from Luderitz. The countryside becomes

gradually more rocky with ridges and table moun-

tains covered with dwarf shrub vegetation and the

grass-covered plains become more restricted in extent.

The dominant grass is now Stipagrostis ciliata. which

locally covers up to 40 per cent.

Fig. 16. — Vornamib near Aus

showing lew grass steppe

with mainly Stipagrostis

spp.

26700—12

548 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

A vegetation type, which Giess (1971) calls dwarf

shrub savanna and which differs considerably from

the Namib and Vornamib vegetation, starts 170 km

inland from Luderitz. The climate is of the cool

desert type with summer rainfall between 100-200 mm.

The soil of this plateau area is usually stony or

gravelly with stones usually large (about 0,10-

0,20 m in diameter and occasionally reaching 0,50m).

Calcrete concretions are abundant and sometimes

thick layers of calcrete are present. Total cover of the

vegetation varies between 10 and 25 per cent (Fig.

17). Dwarf shrubs and shrubs are dominant. Grasses,

which otherwise do not cover more than 5 per cent,

are more abundant where sandy patches occur. The

most important grasses are Stipagrostis obtusa and

S. ciliata. The shrubs Rhigozum trichotomum and

Calophractes alexandri, which are often between

0,40-2 m tall, and the dwarf shrubs Zygophyllum cf.

meyeri and Z. sujfruticosum, up to 0,50 m high, are

dominant in patches which can be up to 5 km in

diameter. Shrubs of Acacia mellifera subsp. detinens,

Phaeoptilum spinosum, Boscia foetida and Parkin-

sonia africana and low trees of Boscia albitrunca and

Aloe dichotoma also occur and can locally cover up

to two per cent and reach 2 or 3 m in h ight. Drainage

lines are occasionally fringed by open woodlands

with trees up to 6 m tall and covering about 15 per

cent, mixed with shrubs, up to 4 m tall, which cover

15 to 25 per cent. The tree species are usually

Acacia erioloba, Ziziphus mucvonata and Euclea

pseudebemis and the shrubs predominantly Acacia

mellifera subsp. detinens. Over vast stretches the

vegetation described above alternates with open

shrubland, in which Euphorbia gregaria is completely

dominant (Fig. 18). The latter type of vegetation

occurs particularly in the vicinity of Goageb, where

the rocky plateau is undulating and has an occasional

butte rising above it. Where the plains are stone-

strewn and rocky they are interrupted by low, slightly

raised ridges. The granite and quartzite stones are

angular, usually 0,10-0,20 m in diameter and often

cover 75 per cent or more of the surface. The soil

between these stones is a sandy loam. Euphorbia

gregaria is a glaucous, leafless, succulent shrub, with

a globular form, measuring 1-2 m in diameter.

A number of other shrubs occur, of which Rhigozum

trichotomum is most abundant, but Phaeoptilum

spinosum, Acacia mellifera subsp. detinens and some

Boscia spp. also occur. Stipagrostis uniplumis is

present although not abundant. The total cover of

this vegetation type is almost entirely made up by

Euphorbia gregaria and varies between 5-20 per cent.

Fig. 17. — Dwarf shrub steppe

of mainly Zygophyllum

spp. on stony soil

approximately 60 km east

of Aus.

Fio. 18. — Extensive open

stand of Euphorbia

gregaria approximately 80

km west of Keetmans-

hoop.

B. J. COETZEE AND M. J. A. WERGER

549

Near Seeheim the transect crosses a few shale

ridges, on which an open shrub community occurs.

The main shrub and low tree species up to 2 m

tall, are Boscia cilbitnmca, Aloe dichotoma, Cadaba

aphylla, Euphorbia virosa, Adeuolobus gariepina.

Acacia mellifera subsp. de tineas, Phaeoptilum spino-

sum and Rhigozum trichotomum. Conspicuous grass

species include Stipagrostis uniplumis, S. hoch-

stetterana var. seca/ina and Panicum arbusculum.

Total cover reaches about 20 per cent, of which

about 4 per cent is contributed by the grass and

dwarf shrub layer.

Towards Keetmanshoop the vegetation over large

stretches is an open shrubland with usually a low

ground cover. It is a slightly sloping, sometimes

undulating plateau covered with a sandy loam,

containing fine gravel in places. The shrubs, up to 1,5

m tall and covering about 10-15 per cent and

occasionally reaching 25 per cent, are mainly

Rhigozum trichotomum (dominant), Parkinsonia

africana, Phaeoptilum spinosum, Acacia mellifera

subsp. detinens, Boscia foetida, Cadaba aphylla,

Boscia albitrunca and Zygophyllum suffruti cosum.

A few small trees of Pappect capensis also occur.

Grasses are sparse with a total cover of about one per

cent, mainly made up by Stipagrostis uniplumis. In

drainage lines shrub growth is often somewhat more

dense and Acacia erioloba and Acacia karroo occur

occasionally. In deep, sandy drainage lines Stipa-

grostis namaquensis is encountered.

shrub steppe, reaching 0,60 m height. The landscape

is flat and the soil changes from a rather deep sandy

loam to being very shallow with thick calcrete out-

crops. Main species include Salso/a tuberculata and

Zygophyllum dregeanum, which are dominant in

places. Stipagrostis ciliata, S. uniplumis and Ennea-

pogon brachystachyus are important grasses. Total

cover varies between 10-25 per cent. The dwarf

shrubs often cover about 10-15 per cent and the

grasses 5-10 per cent. In places patches of Rhigozum

trichotomum are encountered, and low trees of

Boscia albitrunca are rare. Sometimes this karroid

dwarf shrub steppe is interrupted by a pan or a patch

of savanna with shrubs up to 3 m tall, but covering

only 2 per cent or less. The main shrubs are Boscia

foetida and Parkinsonia africana , but scattered low

trees of Boscia albitrunca are found. On exposed

calcrete, patches of Catophractes alsxandri, up to

2 m tall and covering about 2 per cent, occur.

About 100 km east of Keetmanshoop, nearly

500 km inland of Liideritz, this vegetation type of

karroid dwarf shrub steppe interrupted by open

shrub savanna, comes to an end, when along a very

narrow boundary the dune sand area of the southern

Kalahari is reached. Although the main climatic

conditions do not change, the presence of a thick

sand substrate brings about a major change in the

vegetation.

For the next 350 km, stabilized red sand dunes

dominate the landscape. Sometimes a pan with a

surface of very compact, loamy soil or less often

calcrete and dry riverbeds, with calcrete or pink to

white sandy river banks interrupt the monotony of

Fig. 19. — Open shrub and

dwarf shrub vegetation

approximately 40 km east

of Keetmanshoop with

drainage line in middle

distance.

About 20 km east of Keetmanshoop the transect

crosses some oddly-weathered rocky outcrops on

which Aloe dichotoma is very abundant. Other shrubs

and dwarf shrubs also occur, covering about 5 per

cent, as well as an open grass growth in which various

Stipagrostis spp. and Panicum arbusculum are con-

spicuous. Total cover of this vegetation reaches about

25 per cent. Having passed through these rocky

outcrops, the transect runs again on a loamy plain

carrying mainly the same open shrub vegetation as

in the vicinity of Keetmanshoop (Fig. 19). However,

approximately 50 km east of Keetmanshoop, just

over 200 km after this open shrub vegetation started,

the vegetation changes to an open, karroid dwarf

this area. The vegetation on the sand consists largely

of a very open tree or shrub savanna (Fig. 20),

although owing to overstocking, denuded areas and

drifting dunes occur frequently. On pan surfaces and

in dry riverbeds open grassland communities, some

entirely ephemeral, are found, and on the calcrete

an open dwarf shrub steppe usually occurs. Trees of

Acacia erioloba and A. haematoxylon are most

abundant in or near the dry river beds, but can

occur throughout the area. The vegetation on the

sand usually consists of a catena of two or three

communities (cf. Leistner & Werger, 1973). On the red

dune tops occurs a community with the tall shrubby,

1,5 m high grass Stipagrostis amabilis and the tuft

550 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

grasses Eragrostis lehmanniana, E. trichophora,

Asthenatherum glaucum, Aristida meridionals, Stipa-

grostis uniplumis, Brachiaria glomerata and Schmidtia

kalahariensis. The virgate shrub Crotalaria spartioides

is also typical of the dune tops. Other woody plants

on dune tops include Acacia erioloba, A. haematoxylon

and Boscia albitrunca that are up to 6 m high. Several

shrub Lebeckia linearifolia occur in the river bed,

although the latter also occurs on the sand dunes.

Other frequently-encountered plants in the river beds

include Geigeria pectidea, Amaranthus dinteri, Psoralea

obtusifolia, Platycarpha carlinoides, Panicum

impeditum, Deverra aphylla and several others (cf.

Leistner & Werger, 1973). Large trees of Acacia

Fig. 20. — Dune country in the

southern Kalahari show-

ing open tree savanna of

Acacia erioloba with tall

tufts of Stipagrostis ama-

bilis on dune crests.

other annual and perennial forbs and dwarf shrubs

occur on dune tops as well as in dune valleys, including

Acanthosicyos naudinianus, Citrullus lanatus, Chas-

canum pumilum, Oxygonum spp., Limeum spp.,

Hermannia spp. and Tribulus zeyheri. Total aerial

cover usually varies between 5—1 5 per cent, but can

be nearly zero in heavily grazed areas.

In the dune valleys Asthenatherum glaucum is

particularly abundant. Other frequently encountered

species include Crotalaria sphaerocarpa, Hirpicium

echinus, Dicoma schinzii, Stipagrostis uniplumis,

Acanthosicyos naudinianus, Eragrostis lehmanniana,

Brachiaria glomerata, Schmidtia kalaharensis, Gisekia

africana and Limeum sulcatum. Total cover of this

vegetation varies between 10-25 per cent. Trees and

shrubs are less common than on dune crests, but

Acacia haematoxylon , A. erioloba and Grewia flava

are regularly encountered, although they usually

cover less than one per cent.

In deeper dune valleys, where a calcrete layer comes

close to the surface and the sand is of a pinkish

colour, a community with more dwarf shrubs and

low shrubs up to 1,20 m in height, covering about

10-15 per cent, occurs. Common species include

Monechma incanum, Aptosimum albomarginatum, Tn-

digoferci alternans, Chrysocoma polygalifoHa , Stipa-

grostis ciliata and particularly Rhigozum trichotomum.

A similar community is found on the pink sand

dunes fringing some of the dry riverbeds and pans,

but Acacia mellifera subsp. detinens, Grewia flava,

Acacia erioloba, Boscia albitrunca and Lycium

austrinum are more common here. The trees are up to

6 m tall and the total aerial cover usually varies

between 10 and 15 per cent. In the dry river beds

Stipagrostis obtusa and Rhigozum trichotomum fre-

quently dominate alternately on white sandy fringes,

whereas in the clayey central part Panicum coloratum,

Chloris virgata, Geigeria pectidea, Enneapogon

brachystachyus and others are locally dominant.

Often the dwarf shrub Galenia secunda and the

erioloba and A. haematoxylon are also found and

under these trees Setaria verticillata is usually

dominant.

Where the river is fringed by calcrete banks, an

open dwarf shrub community, up to about 0,60 m

tall, and with an aerial cover between 5 and 15 per

cent, is found. Important plants on these calcrete

banks include Aizoon schellenbergii, Barleria rigida,

Zygophyllum pubescens, Sylitra biflora, Indigofera

auricoma, Enneapogon scaber, Rhigozum trichotomum

and Stipagrostis obtusa. Sometimes a few individuals of

Boscia albitrunca and Lycium austrinum are en-

countered. In pans extensive stands of the annual

grass Sporobo/us coromandelianus or the perennial

Sporobolus rangei can be seen. Cover in these stands

can vary greatly, but it is seldom more than 30 per

cent.

Near Hotazel, about 850 km from Liideritz, where

the precipitation reaches 300 mm, the Kalahari ends.

This boundary is, however, not as sharp as where

the Kalahari started, 350 km further westward.

Even as far as 100 km before the Kalahari ends, in

the vicinity of Van Zylsrus, at the 200 mm isohyet,

the vegetation starts to change gradually. Trees and

shrubs gradually play a more important role in the

landscape. Acacia haematoxylon, A. mellifera subsp.

detinens and Boscia albitrunca are the most common

trees. They are up to 4 m tall and cover up to 30 per

cent of the surface. Rhigozum trichotomum is locally

also important. Grasses cover about 20-25 per cent.

On dune crests Terminalia sericea is now the most

common tree, being 3 to 6 m tall and covering up to

20 per cent.

About 50 km west of Hotazel the first rocky

outcrops occur, which carry a vegetation with shrubs

and grasses. In the sandy areas total cover gradually

increases and can reach values of about 40 per cent.

B. J. COETZEE AND M. J. A. WERGER

551

Jn this vegetation type, shrubs, which are mainly

Rhigozum trichotomum, Lycium tenue, Acacia melli-

fera subsp. detinens and Grewia flaxa , and the trees

Boscia albitrunca and Acacia erioloba, vary in cover

between 5 and 20 per cent. The trees are up to 7 m

tall. Dwarf shrubs, particularly Gnidia polycephala,

and grasses, such as Eragrostis lehmanniana, Stipa-

grostis uniplumis, S. ciliata and 5. obtusa, often cover

about 25 per cent.

Near Hotazel the dune sand starts to disappear

from the landscape and the surface rises to the

upper, central plateau (Fig. 1). The climate changes

from a desert type to a cold, dry steppe climate

(BSkw'), with a summer rainfall of over 300 mm.

In the shrub layer Tarchonanthus camphor atus

becomes more important. East of this area the

“white” grasses (mainly Stipagrostis ) of the Karoo-

Namib Region are rapidly replaced by “purple”

grasses of the Sudano-Zambezian Region.

Apart from local sandy patches, the soil in the

area between Hotazel and Kuruman, a distance of

about 50 km, is usually shallow, loamy sand, and

very stony. The landscape is mainly slightly

undulating, regularly interrupted by mountainous

ridges. The vegetation consists of an open to a fairly

dense shrub savanna. The shrubs are 1 - 3 m tall,

covering between 5 and 35 per cent. The most

important shrubby species are Rhigozum trichotomum,

Rhigozum obovatum, Acacia haematoxylon, A. hebe-

clada var. stolonifera, A. mellifera subsp. detinens,

Euclea ovata, Grewia flaxa and, in particular, Tarcho-

nanthus camphoratus. Occasional low trees of Boscia

albitrunca occur and where the soil consists of

somewhat deeper sand the shrub layer is sparse and

trees of Acacia erioloba, covering 1-2 per cent, are

present. In open patches, grasses and herbs cover up

with the vegetation of the southern Kalahari, although

they represent different subtypes. The vegetation on

the Ghaap Plateau alternates from nearly pure

grassland to a dense shrub savanna. Where the

calcrete or dolomite is overlain by a thin layer of

reddish Kalahari sand, grasslands occur (Fig. 21).

The grasses, which are up to 1 m high, reach total

cover values of 60 or 80 per cent. The most important

species are Eragrostis lehmanniana, Themeda triandra,

Chrysopogon montanus, Cymbopogon plurinodis,

Heteropogon contortus, Digitaria eriantha, Sporobolus

fimbriatus, Aristida curxata, A. congesta, A. diffusa

var. burkei, Enneapogon brachystachyus and several

others. Where calcrete or dolomite is exposed on the

surface, shrubby patches occur which are usually

dominated by Tarchonanthus camphoratus with or

without Grewia flaxa, but other woody species, such

as Acacia mellifera subsp. detinens, A. karroo, Rhus

lancea, R. ciliata and occasionally Boscia albitrunca

occur as well. Acacia erioloba sometimes occurs as

polycorms on deeper sand. These woody species are

usually 1-2 m, or occasionally up to 4 m high and

cover between 5 and 40 per cent. Total cover in this

shrub savanna is up to 80 per cent.

About 25 km east of Vryburg, more than 1 000 km

inland from Luderitz, an important boundary is

reached. Agricultural cultivation becomes possible

here and soon becomes one of the main forms of

land use. West of this boundary, agricultural practices

are confined to natural-pasture management except

for the true Namib desert where the carrying capacity

is too low. Where this boundary is reached the climate

is still of the BSkw' type, but the rainfall has increased

to 500 mm. Near Lichtenburg, about 160 km east-

northeast of Vryburg, at the 600 mm isohyet. the

climate changes to a hot and dry steppe climate

(BShw), but the general features of the landscape

Fig. 21. — Grasslands on

Ghaap Plateau, approxi-

mately 15 km east of

Kuruman. In the back-

ground a patch of shrub

savanna dominated by

Tarchonanthus camphora-

tus is visible.

to 25 per cent, but usually their contribution is lower.

Common grasses include Anthephora pubescens,

Themeda triandra, Heteropogon contortus, Eragrostis

lehmanniana, Aristida congesta and Rhynchelytrum

repens. Total cover values vary between 25 and 40

per cent.

At Kuruman, where the rainfall reaches 400 mm,

the Ghaap Plateau starts. The Plateau is a vast flat

plain of calcrete and dolomite, extending eastwards

over a distance of 170 km to beyond Vryburg (cf.

Mostert, 1967). This plateau vegetation is classified

by Acocks (1953) as Kalahari Thornveld, together

remain the same. The landscape is a flat to gently

undulating grassy plain, which extends about 250 km

in an east-northeast direction along the transect (Fig.

22). The soil is sometimes deep and ploughed, other

times shallow and rocky, with calcrete and dolomite

outcrops. Pans occur regularly. The vegetation consists

of almost pure grasslands, which in the western and

central parts are interrupted by occasional thornbush

savannas. The grasses are tufted and about 1 m tall,

with cover values of 70-80 per cent. The species com-

position changes as the precipitation increases from

500 mm in the west to 700 mm in the east. In the same

direction the grass cover also become somewhat

552 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

denser. In the drier western part of this central

plateau grass land, which falls in Acocks’s (1953) Dry

Cymbopogon-Themeda Veld, species with xerophytic

characteristics are prevalent (Roberts et ah, 1972). The

most conspicuous species are Stipagrostis uniplumis,

Eragrostis lehmanniana, Anthephora pubeseens, Cym-

bopogon plurinodis and Themeda triandra. A mixture of

species with low nutritive value during all but the

earlier part of the growing season, called sour grasses,

parts, was found by Morris (1973) and Morris &

Guillerm (1974) to be most strongly related to position

on the gently undulating surface. In the eastern parts

in the Central Variation of Acocks’s Bankenveld,

where the topography is more strongly rolling, Coetzee

(1974) found exposure, slope, aspect and pedological

characteristics to be important factors related to

variation in species composition.

Fig. 22. — Grasslands of the

eastern part of the central

plateau with open woody

vegetation on outcrops

in the distance.

' V .A V A,

•"CSV-..

is characteristic of the higher rainfall area, and pre-

dominate in the central and eastern parts. These

parts fall in Acocks’s Sandy Cymbopogon-Themeda

Veld and Bankenveld respectively. Notable among

these sour grasses are Trachypogon spicatus, Sporo-

bolus pectinatus, Loudetia simplex, Eragrostis race-

mosa, Cymbopogon excavatus and Diheteropogon

amplectens.

The patches of thornbush savanna in the western

parts of this grassland plateau consist of an open

to fairly dense tree and shrub layer, up to 5 m in

height, with cover values of at least 15 per cent, and

a grassy understorey. Total cover values of this

vegetation type are often between 60 and 80 per cent.

The main woody species include Acacia mellifera

subsp. delinens, A. hebeclada var. stolonifera, A.

tort ilis, A. erioloba and A. karroo , but Tarchonanthus

camphoratus, Grewia flava, Ziziphus mucronata,

Ehretici rigida, Rhus ciliata, Rhus pyriodes and Rhus

lancea also occur. Over most of these western and

central parts of the plateau area, however, woody

plants are absent, probably as a result of particularly

severe frosts that occur on 60-90 days per annum

(Acocks, 1953; Roberts et ah, 1972, Fig. 5 and 6).

Acacia caffra and Protea caffra savannas with a

grassy understorey occur regularly on low ridges and

hillsides in the eastern parts. Acocks (1953) considers

regular burning to be an important factor in main-

taining a predominantly grassy vegetation here. It is

only east of the Ghaap Plateau that the cover of the

vegetation becomes sufficiently dense to allow fires

on an extensive scale.

Local variation in florist ic composition in the grass-

lands in the Lichtcnburg area, which is on the

transition from the drier western to the wetter central

Immediately north-west of Derby, just over 100 km

east-northeast of Lichtenburg and more than 1 250 km

from Luderitz, the landscape changes abruptly. The

gently undulating grassland gives way within less

than one kilometre to hilly terrain with grassy and

shrubby savannas and woodlands on hillsides and

denser woodland in valley bottoms. Frost occurs

regularly in winter, but is less severe than on the

slightly higher and more exposed grassland plains to

the south. The rainfall is as high as on the adjoining

grassland plains and the climate is still classified as

BShw. The tree canopy, covering 20-30 per cent, is

between 3 and 8 m tall and is dominated by Rhus

lancea. Acacia karroo, Olea africana and Acacia

robusta. A fairly dense woodland of Rhus lancea,

Buddleia saligna, Olea africana and Ziziphus mucro-

nata, 5-8 m tall and with a total aerial cover of

75 per cent, is typical of the valley bottoms. Several

shrub species form a middle stratum and the grass

cover ranges from 50 per cent in the denser woodland

to 95 per cent in the more open savannas on slopes.

The grasses are up to 1,5 m tall. Dominant grasses

on hotter, west-facing slopes include Themeda

triandra, Cymbopogon plurinodis and Heteropogon

contortus while Setaria lindenbergiana is dominant

on mesic south-facing slopes. This vegetation type,

which falls in Acocks’s Sourish Mixed Bushveld,

occurs as an irregular belt on the foothills and

gentle slopes fringing mountains and highlands and

is crossed for about 20 km by the transect.

The vegetation on the Magaliesberg, which is

encountered south-west of Rustenburg, about 1 300

km from Luderitz, and where the rainfall is between

700 and 800 mm, is together with the rest of the

B. J. COETZEE AND M. J. C. WERGER

553

mountain vegetation in the northern and north-

western Transvaal classified by Acocks (1953) as

Sour Bushveld. Several communities have been

described by Coetzee (1975) from a small area near

Rustenburg on this narrow mountain range. Forest

communities occur in kloofs with tree species up to

about 13 m high and with interlocking canopies.

These forests include Ilex mitis, Pittosporum viridi-

florum, Rothmannia capensis, Bequaertiodendron

magalismontanum , Mimusops zeyheri, Celtis africana

and Ficus pretoriae as common canopy trees. The

predominant vegetation types of the south- to west-

facing mountain slopes are deciduous Acacia caffra

savannas with trees up to 8 m tall and evergreen

Protea caffra savannas in which the trees are up to

5 m tall. Tree cover in both savanna types is between

15 and 60 per cent (Fig. 23). Protea caffra savannas

also occur on top of the mountain where the

vegetation is mainly dense sour grassland and open

shrubland. Prominent species in the grasslands

include Themeda triandra, Trachypogon spicatus,

Eragrostis racemosa, Diheteropogon amplectens ,

Rhynchelytrum setifolium, Brachiaria serrata, Loudetia

simplex, Tristachya biseriata and many other grasses

and forbs, most of which also occur in the grass

layers of the Protea caffra savannas. The shrublands,

storeys. Setaria perennis, S. lindenbergiana, Hetero-

pogon contortus, Trachypogon spicatus and Tristachya

biseriata are dominant grasses in these deciduous

woody vegetation types. Variation in the vegetation of

a northern and southern kloof of the Magaliesberg

have been described by Van Vuuren & Van der

Schijff (1970) who conclude from the species com-

position on the two sides of the mountain that it

forms a clear climatic and floristic boundary.

The landscape between Rustenburg and Groblers-

dal, a distance of 220 km, where the rainfall is again

between 600 and 700 mm and the frost less severe

than south of the Magaliesberg, is mainly flat to

undulating with a number of distinct savanna and

woodland vegetation types that are related to soil and

topographic differences. Considerable parts of this

vegetation have been severely disturbed by townships,

horticulture, removal of trees and overgrazing. The

grasses are a mixture of sour and sweet types, sweet

grasses being those that retain their nutritive value

well into the dry season. A grassy woodland of

Faurea sa/igna and Burkea africana, in which the

tree canopy is up to 9 m tall and covers 20 per

cent, with an understorey of tall grasses of up to 1 ,5 m

high, covering about 60 per cent, occurs in a flat

narrow belt of coarse sandy soil, fringing the

Fig. 23. — Acacia caffra and

Protea caffra Savannas

on the Magaliesberg near

Rustenburg.

which are up to 3 m tall, are confined to broken

rocky outcrops and are characterized by Bequaertio-

dendron magalismontanum and Landolphia capensis.

On extremely shallow litholic soils, a large number of

small plants with xerophytic characters is found,

including the desiccation-tolerant Myrothamnus

flabellifolius, Selaginella dregei and Oropetium

capense and a number of succulents including

Euphorbia schinzii, Adromischus umbraticola, Frithia

pulchra, Khadia acutipetala, Anacampseros sitb-

velutinum and Euphorbia clavarioides. North- and

east-facing slopes of the mountain are dominated by

deciduous savannas and woodlands with Burkea

africana, Faurea saligna, Combretum zeyheri, Acacia

caffra and several other woody species. The trees are

up to 10 m tall and cover between 30 and 60 per cent.

Shrubs and a dense grass layer form the under-

Magaliesberg on the northern side. The dominant

grasses are Eragrostis spp. and Digitaria spp. (Fig.

24).

Further to the north-east, on the same flats but on

black clay soils derived from norite, about 10 km

of thornbush savanna occurs. The savanna is

dominated by 3-8 m tall thorny microphyllous trees

of Acacia tortilis and A. karroo with 15 per cent cover,

occasional shrubs and a dense grass layer forming the

undergrowth. In relatively well preserved stands of

this vegetation type the grass cover is 95 per cent and

up to 1 m tall, with Themeda triandra, Elionurus

argenteus, Aristida spp. and Heteropogon contortus

as dominants. South-facing and north-facing slopes

of a long range of norite hills, which protrude from

the flats and run parallel to the Magaliesberg differ

554 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

considerably in species composition. A woody

vegetation of Acacia caffra and many other low tree

and shrub species occurs on the south-facing slopes of

these hills. The woody species range between 1 and

5 metres in height with cover varying from 15 to

60 per cent. The grasses are 1 m tall, cover 60-70 per

cent and include Themeda triandra, Eragrostis spp.,

Setaria lindenbergiana, Heteropogon contortus, Rhyn-

chelytrum setifolium and Elionurus argenteus. The

north-facing slopes carry a shrubby woodland vege-

tation with up to 7 m tall trees covering 60 per cent

and a grass cover of 75 per cent. Combretum mode

and Pouzolzia hypoleuca are among the dominants,

but many more woody species are common. The

dominant grasses, up to 1,5 m tall, are Heteropogon

contortus and Setaria lindenbergiana.

In the undulating landscape to the north-east of

where the transect crosses the norite hills, the vege-

tation belongs to Acocks’s Sourish Mixed Bushveld

and Mixed Bushveld. A broad-leaved deciduous Com-

bretum apiculatum savanna occurs on the coarse,

gravelly and sandy soils of the convex upper parts of

the undulations (Fig. 25). A mixture of several other

woody species contributes to the 25 per cent total

cover of the upper stratum, which is 3-8 m tall.

Shrubs up to 3 m tall can cover approximately

10 per cent. Grass cover varies between 30 and 90

per cent and dominant grasses are up to 1 ,5 m tall.

In this savanna, clumps of dense woodland with

diameters of about 10 m, are found on termite nests

where the soil is more loamy and moisture and

aeration conditions differ from those of the

surroundings. Dominant woody species in such dense

clumps include Dicrostachys cinerea, Acacia torti/is, A.

nilotica, Rhus leptodictya and Ziziphus mucronata.

The first three species are listed by Grunow (1965)

among the prominent plants of woodland communities

occurring on sandy loam topsoils underlain by

sandy clay, within the Sourish Mixed Bushveld.

Microphyllous thornbush dominated by Acacia melli-

fera subsp. detinens, with a dense grassy understorey,

grows in the clayey lower parts of undulations (Fig.

26). The shrubs are between 3 and 5 m tall and

cover up to 80 per cent and Panicum maximum , the

dominant grass in a 1,5 m tall stratum, covers 80-90

per cent in some places. A savanna vegetation with

Sclerocarya caffra, Terminalia sericea, Combretum

apiculatum, Acacia nilotica, A. tortilis, Dicrostachys

cinerea and Euclea undulata among the prominent

trees and shrubs, is found on flat areas with red

sandy-loam soil (Fig. 27). The tree stratum includes

trees from 2-8 m tall and covers 25 per cent, the

shrub stratum ranges from 0,5-2 m, and covers 2 per

cent and the grass and forb stratum, which reaches

up to one metre, covers approximately 60 per cent.

In a detailed study of a small area within Acocks’s

Sourish Mixed Bushveld, Grunow (1965) found

differences in soil to be the most important factors

related to the occurrence of various communities.

About 100 km east-north-east of Rustenburg lies

an isolated flat area of 20 by 40 km covered by

deep, reddish-brown Kalahari sand with a distinctive

vegetation. The vegetation is a 7 m tall savanna with

trees covering 25 per cent. Shrubs, between 1 and 2 m

tall, also cover 25 per cent. The grass layer is 1 m

high, covering 80 per cent. Dominant trees and shrubs

include Acacia erioloba, Mundulea sericea, Acacia

luederitzii var. retinens, Boscia albitrunca, Terminalia

sericea, Grewia flava and Tarchonanthus camphoratus,

which differentiate this vegetation from that of

surrounding areas. Panicum maximum, Eragrostis

rigidior and Digitaria spp. are among the dominant

grasses. This sandy vegetation has been classified by

Acocks as Kalahari Thornveld.

The short, steep climb from Groblersdal eastward

through hilly terrain, passes through shrubby savanna

with many tree and shrub species, including Ter-

minalia sericea, Combretum zeyheri, Peltophorum

africanum, Acacia nilotica, Sclerocarya caffra, Dicro-

stachvs cinerea, Dombeya rotundifolia, Acacia tortilis,

Pterocarpus rotundifolius and several others. Dombeya

rotundifolia and Croton gratissimus are conspicuous

trees and shrubs of the savannas on rocky outcrops.

On hillsides a savanna is encountered in which Faurea

saligna, Diplorrhynchus condylocarpon, Burkea

africana, Acacia caffra, Pterocarpus rotundifolius,

Combretum zeyheri, Vitex sp. and Ozoroa sp. are

among the predominant woody species. Burkea

africana savanna occurs on deeper soils near the

summits. The woody species in these savanna types

reach heights of up to 6 m and cover about 30 per

cent. These vegetation types fall in Acocks’s Sourish

Mixed Bushveld. From a small area 40 km to the

south of the transect and in a similar part of this

Veld Type, Theron (1973) described a number of

communities in which slope, aspect, altitude, geology,

soil texture and soil depth were among the habitat

features that are related to the variation in vegetation.

B. J. COETZEE AND M. J. A. WERGER

555

Fig. 25. — Combretum apicula-

tum Savanna in Mixed

Bushveld near Jericho.

556 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

Dense, sour grassland predominates over the 15 km

of undulating plateau in the temperate rainy region

(Cwb) west of the Steelpoort Valley. This grassland

is similar in structure and related in prominent

species to the grasslands of the moister parts of the

upper interior plateau, crossed by the transect

between Lichtenburg and Rustenburg and found on

top of the Magaliesberg. The rainfall is between 700

and 800 mm and soils are coarse and sandy. Dominant

grasses include Urelytrum squarrosum, Themeda

triandra, Schizachyrium sanguineum, Eragrostis

chloromelas, Elionurus argenteus, Setaria perennis,

Tristachya rehmannii, Trachypogon spicatus, Era-

grostis racemosa, Panicum natalense, Perotis patens,

Heteropogon contortus and Cymbopogon excavatus.

On granitic outcrops the vegetation is a shrubby

savanna of Acacia caffra, Combretum mode, Dombeya

rotundifolia, Cambium gilfiUani, Bequaertiodendron

magalismontanum and several others, all of which are

common in the Sour Bushveld of the Magaliesberg.

The vegetation on this plateau is classified by Acocks

as North-Eastern Sandy Highveld.

On the slopes of the Steelpoort Valley, 40 km east

of Groblersdal, the vegetation consists of a shrubby

savanna with a large variety of woody species,

including Terminalia prunioides, Rhus engleri, Euclea

undulata var. myrtina, Grewia vernicosa, Cassine

transvaalensis , C. aethiopica, Boscia albitrunca, Croton

menyhartii, Acacia gerrardii, Ormocarpum trichocar-

pum, Croton gratissimus var. gratissimus, Kirkia

wilmsii, Vepris reflexa, Mundulea sericea, Strychnos

madagascariensis, Commiphora angolensis and Pavetta

tristis. Most of these species clearly differentiate the

vegetation of this hot dry valley, which receives

500-600 mm precipitation annually, from the cooler

moister savanna and woodland regions to the west

and east. A savanna dominated by Acacia tortilis,

Terminalia prunioides, Boscia foetida subsp. foetidci

and patches of the low shrub Euclea linearis occurs

on the 2 km wide, flat valley floor.

East of the Steelpoort Valley a vegetation type

similar to that encountered west of the valley

continues for another 40 km. Further eastward the

topography changes to a strongly rolling and moun-

tainous country, with savannas on rocky hillsides and

otherwise sour dense grasslands, which belong to the

Eastern Variation of Acocks’s (1953) Bankenveld.

The rainfall of this 20 km wide isolated patch of

Bankenveld is between 600 and 700 mm. Many of the

dominant woody species and grasses are the same as

those found in the higher rainfall parts of the central

interior plateau, the Sour Bushveld of the Magalies-

berg and the rest of the north-eastern mountain

plateau.

The ascent over 25 km eastwards to the edge of the

Drakensberg escarpment passes through Acocks’s

North-Eastern Mountain Sourveld. The topography is

undulating to rolling, rising in altitude to 2 350 m.

Mean annual precipitation is between 700 mm in

the west and I 750 mm or more on the escarpment.

The area is subject to severe frosts. Here the transect

crosses through a vegetation with a strong Afro-

mcntane element and a considerable Austro-afro-

alpine element in the flora. Van der Schijff &

Schoonraad (1971), in a study of the escarpment

50 km to the north of the transect, list a number of

characteristic sour grasses, geophytes, forbs and

dwarf shrubs for the typical dense mesophytic grass-

land of this region, including the large tussocky

Merxmuellcra drakensbergensis, and Eragrostis

sclerantha, E. caesia, E. capensis, Agrostis barbuligera

var. longipilosa, Harpechtoa falx, Loudetia simplex,

Wat sonia densiflora, M oraca spathulata, Helichrysum

spp., Erica drakensbergensis, Protea gaguedi,

Vaccinium exul, Restio sieberi var. schoenoides and

Tetraria cuspidata. From the highest region of the

escarpment the altitude drops within a few kilometres

down steep mountain slopes to Sabie where the

woodlands of the Eastern Transvaal Lowveld start.

The climate of these slopes is classified by Schulze

(1947) as Cwa. These escarpment slopes are covered

mainly by grassland and tree plantations. Small trees,

shrubs and ferns, including the tree fern Cyathea

dregei, occur in small moist sheltered kloofs. Dense

indigenous evergreen montane forest occurs patchily

in deep kloofs and extensively on mesic south-east-

facing slopes of the eastern escarpment (Fig. 28).

Van der Schijff & Schoonraad (1971) distinguished

four strata in well developed parts of this forest.

The upper tree stratum of up to 18,0 m includes

Podocarpus latifolius, P. falcatus, Cussonia umbellifera,

Ochna oconnorii, Curtisia dentata, Kigellaria africanct

Nuxia floribunda, N. congesta, Apodytes dimidiata and

several others Other strata are a lower shade-tolerant

tree stratum, a shrub stratum and a herbaceous

stratum. The cover of the latter two strata varies

considerably, depending on aspect and amount of

light penetrating the higher strata. The rainfall in such

kloofs at Mariepskop to the north is according to

Van der Schijff & Schoonraad (1971) possibly up to

2 500 mm per annum.

Fig. 28. — Dense montane kloof forest on the eastern escarpment

with Ensete ventricosum in foreground (Photo: J. C.

Scheepers).

The area between the escarpment and the east

coast is practically frost free except for some

occasional frosts occurring along rivers (Van der

Schijff, 1957, 1969). The climate of this lowveld area

is, except for the narrow coastal belt, classified as a

hot and dry steppe climate (BShw), but the belt

next to the escarpment has a higher rainfall than the

central plateau area. Over the first 40 km to the east

of the escarpment the landscape is hilly and rolling

to undulating with sandy soils and rainfall decreasing

B. J. COETZEE AND M. J. A. WERGER

557

from 1 000 mm in the west at the bottom of the

escarpment, to 700 mm in the east. The vegetation

belongs to Acocks’s Lowveld Sour Bushveld. It is a

savanna or woodland with Terminalia sericea among

the dominant woody species. Other common woody

species are Sclerocarya caffra, Albizia versicolor,

Dichrostachys cinerea, C ombre turn mo lie, Pterocarpus

angolensis and others. Evergreen trees, including

Trichilia emetica, which are found only along river

banks on the drier flats to the east, are common in

this higher rainfall area (Van der Schijff, 1969).

Shrubs and sour grasses largely make up the under-

storeys. Woody species are up to about 8 m tall

although they are usually shorter and cover generally

between 20 and 60 per cent.

The 90 km-wide area eastward to the South African

border on the Letombo Mountains has a mean annual

rainfall of 600-700 mm. The western 60 km of this

area is an undulating granitic terrain with sandy soils

and has mainly a deciduous broad-leaved savanna,

dominated by Combretum apiculatum with shrubs and

grasses in the understorey. Large termite mounds of

the genus Macrotermes are common in this savanna,

the old ones carrying a specific community, among

which Diospyros mespiliformis, Schotia brachypetala,

Ziziphus mucronata, Spirostaehys africana and Xantho-

cercis zambesiaca are typical. A denser woodland of

microphyllous thorny species including several Acacia

species, Dichrostachys cinerea and Spirostaehys

africana occurs on clayey soils between the rises. On

dolerite dykes an Acacia nigrescens woodland, which

is up to 10 m high, occurs (Van der Schijff, 1957, 1969).

A savanna, dominated by Acacia nigrescens and

Sclerocarya caffra and including Combretum imberbe

and Lonchocarpus capassa, occurs on the heavy basaltic

soils of the flats immediately west of the Lebombo

Mountains (Fig. 29). The trees are up to 10 m high

and usually cover up to about 25 per cent. The

dominant grasses in the ground layer are sweet and

include Panicum and Digitaria species. Total aerial

cover amounts to about 60 or 70 per cent. On the

rhyolitic, litholitic soils of the Lebombo Mountains a

savanna dominated by Combretum apiculatum with

several other woody species including Pterocarpus

rotundifolius is found. The woody species are up to

4 m tall and total cover of the vegetation amounts to

about 60 per cent. On ridges and in kloofs the woody

species are denser and include Diospyros mespiliformis,

Terminalia phanerophlebia, Antidesma venosum, Afzelia

quanzensis, Euphorbia confinalis, E. cooperi, Ficus

ingens, Androstachys johnsonii and others (Van der

Schijff, 1957; Gomes Pedro & Grandvaux Barbosa,

1955). On the undulating terrain with basaltic soils,

east of the Lebombo Mountains, the Acacia nigrescens-

Sclerocarya caffra savanna, occurs again over a stretch

of about 40 km. Further eastward the annual rainfall

increases to approximately 900 mm. The soil is sandy

and carries a broad-leaved savanna or woodland

similar to the Terminalia sericea savanna described

above (Fig. 30), which occurs on the sandy soils in the

higher rainfall lowveld near the escarpment. This

vegetation type continues up to about 18 km from the

coast and is often destroyed by a primitive type of

cultivation and numerous orchards of Anacardium

occidentale (cashew). Along drainage lines Syzygium

cor datum is common.

About 18 km before the east coast of Africa is

reached at Vila Luiza, the transect enters the coastal

belt with a tropical rainy climate (Aw with annual

rainfall 900 — 1 100 mm) and with deep sandy soils

interrupted by wide, marshy and clayey flood plains

and lagoons. On the clayey and marshy soils extensive

grasslands and swamp vegetation types occur. The

Fig. 29. — Acacia nigrescens-Sclerocarya caffra Savanna on

heavy basaltic soil west of the Lebombo Mountains.

Fig. 30. — Terminalia sericea Woodland on sandy soil in eastern

Lowveld.

grasses are up to 1,5 m tall and cover up to 100 per

cent. Important species are Setaria hoist ii, Ischaemum

brachyatherum, Chloris gay ana, Panicum maximum,

P. deustum and others. Acacia xanthophloea and

Hyphaene crinita are common in these grasslands. In

inundated areas Imperata cylindrica, Cy perns papyrus,

Phragmites spp., Juncus spp., Tvpha sp. and Eich-

hornia crassipes are found. On the sandy soils the

natural vegetation is a dense forest, but in the area

where the transect runs, this forest has largely been

destroyed and has also been replaced by Anacardium

occidentale orchards and primitively-cultivated lands.

Patches of the dense forest with trees up to 15 m high

still occur, and include the following species: Afzelia

quanzensis, Brachylaena discolor, Sclerocarya caffra,

Trichilia emetica, Albizia adiantifolia, Garcinia living-

stonii, and others (Gomes Pedro & Grandvaux

Barbosa, 1955; Tinley, 1971). Closer to the coast a

dune forest, which is on the outermost dunes and is

therefore pruned by the wind to a coastal scrub (Figs.

31 and 32) is found. The dune forest includes Dialium

schlechteri, Scolopia zeyheri, Brachylaena discolor and

others. The coastal scrub includes Mimusops caffra,

Brachylaena discolor, Diospyros rotundifolia, Vepris

lanceolata, Euclea natalensis and Acacia robust a. On

the beach the pioneer sand binders are Scaevola

thunbergii, Ipomoea pes-caprae, Canavalia maritima,

Sophora tomentosa and Tephrosia canescens.

558 A WEST-EAST VEGETATION TRANSECT THRO H AFRICA SOUTH OF THE TROPIC OF CAPRICORN

• '

Fig. 31. — Dense coastal forest

on dune with Indian

Ocean on right and fresh-

water lagoon on left

(Photo: E. J. Moll).

Fig 32. — Wind-pruned dune

scrub near the coast

(Photo: E. J. Moll).

DISCUSSION

Main vegetation discontinuities

Eleven main discontinuities in the structure and

floristic composition of the vegetation along the

transect are strongly related to a climatic gradient

across the continent.

1. The true Namib desert occurs between the west

coast and the escarpment in the winter rainfall desert

climate with a sporadic annual rainfall of less than

100 mm. It is distinguished by a very low vegetation

cover of less than one per cent, consisting mainly of

succulents and other xerophytes and is entirely

unsuitable for grazing of any form (Rattray, 1960).

In the cool desert with frequent fog along the west

coast, plants occur more commonly than in the hot

inland desert where virtually no vegetation is en-

countered over extensive areas.

2. In the cool, summer rainfall, desert climate of the

Vornamib on the western escarpment, the rainfall is

below 100 mm and in the cool desert climate of the

westernmost interior plateau, it is between 100 mm and

300 mm. Here the vegetation cover increases to

between 5 and 25 per cent with local patches of grass

on deep sand covering up to 40 per cent. In the

Vornamib low open grasslands are found on sandy

soils while dwarf shrubs are dominant on litholitic

soils. Trees occur only occasionally in this area. On

the loamy and litholitic soils further east where the

annual rainfall increases to over 100 mm and becomes

less sporadic, dwarf shrubs and shrubs become

dominant with short grasses more abundant in sandy

patches. Trees occasionally cover up to 2 per cent

except for the shale ridges near Seeheim where trees

and shrubs cover up to 15 per cent. Further east on

the deep Kalahari sands, where the rainfall is still

below 200 mm, the vegeta ion is largely a very open

tree or shrub savanna with a tall open grass under-

storey. At the 200 mm isohyet trees gradually become

more important and the species composition changes

even though the Kalahari sand continues for another

100 km.

B. J. COETZEE AND M. J. A. WERGER

559

3. From about 50 km west of Hotazel to Kuruman,

where the climate changes from a desert type to a

cold dry steppe and the rainfall is between 300 and

400 mm, the total cover of the vegetation increases

and is usually between 25 and 40 per cent. A tall tree

savanna with shrubs and grasses occurs on sandy

areas west of Hotazel. The tree cover varies between

5 and 25 per cent and the grasses are mainly “white”

Stipagrostis spp. of the Karoo-Namib Region. The

soils between Hotazel and Kuruman are mainly

litholitic loamy sands and carry an open to fairly

dense shrub savanna with “purple" grasses of the

Sudano-Zambezian Region.

4. On the Ghaap Plateau, a vast plain of calcrete and

dolomite extending from Kuruman over a distance of

170 km eastward, the total cover of the vegetation has

increased to between 60 and 80 per cent. The rainfall

is here between 400 and 500 mm and the climate

belongs to the cold dry steppe type. A dense shrub

savanna occurs on exposed calcrete or dolomite. One

metre tall grasslands occur where the rocks are

overlain by a thin layer of Kalahari sand.

5. From about 25 km east of Vryburg to Derby, a

distance of 250 km. the vegetation is mainly dense

grassland, changing in species composition from

xerophytic grasses to sour mesophytic grasses as the

rainfall increases from 500 mm in the west to 700 mm

in the east. Near Lichtenburg at the 600 mm isohyet,

the climate changes from the cold dry steppe to the

hot dry steppe type. Stands of savanna occasionally

occur on the flat to undulating plains in the west and

become more regular on low ridges and hillsides

towards the east. Trees are absent, however, over

most of the area, probably as a result of severe frost.

Regular burning is also con idered to be an important

factor which restricts the occurrence of trees, parti-

cularly in the eastern higher rainfall area. Local

variation in floristic composition is related to topo-

graphical position on the gently undulating plains

while pedological features as well as exposure, slope

and aspect are important factors in the more broken

terrain.

6. Savannas and woodlands, with a grass cover

varying from sour to a mixture of sour and sweet,

occupy most of the lower eastern interior plateau

from Derby to 25 km east of Groblersdal. The rainfall

is between 600 and 700 mm over most of this area and

the climate is of the hot and dry steppe type. In these

respects it corresponds to the grasslands of the upper

central interior plateau, from Vryburg to Derby, but

frost is markedly less severe. Soil differences again

account for variation in floristic composition in the

gently undulating areas, and pedological features,

exposure, slope, aspect and altitude for the variation

in mountainous terrain.

7. In the temperate rainy climate of the high sandy

plateau bordering on the eastern escarpment zone,

dense sour grasslands occur, which are related to

those of the central interior plateau and to local

patches of grassland on top of the Magaliesberg. The

rainfall is between 700 and 800 mm over most of

this area and severe frosts occur regularly. Occasional

stands of savanna vegetation are found on rocky

outcrops and occur more widely on rocky hillsides in

the strongly rolling to mountainous landscape

towards Lydenburg. These savannas are also related

to those found in the areas with higher rainfall of the

upper central, interior plateau and on the Magalies-

berg.

8. A local, but floristically very distinct type of

savanna occurs in the Steelpoort Valley which, owing

to its lower altitude, has a hot, probably frost-free

climate and interrupts the previously-mentioned

plateau bordering on the escarpment zone. The

rainfall in this valley is between 500 and 600 mm.

9. Dense, mesophytic grasslands with an Austro-

afro-alpine element and Afro-montane forests occur

on the eastern escarpment which rises to 2350 m

altitude. Rainfall on the escarpment is above 1750 mm

and severe frost is a regularly-occurring phenomenon.

10. The area between the escarpment and the east

coast is virtually frost-free and from Sabie as far as

18 km from the east coast a hot and dry steppe climate

prevails with an annual rainfall varying between 600

and 900 mm. This area carries a savanna and wood-

land vegetation, sometimes up to 10 m in height. The

savannas of the higher rainfall areas with sandy soils,

near the escarpment and east of the Lebombo

Mountains, are distinct in their floristic composition.

In the central lower rainfall areas, floristically distinct

types of savanna occur on different soils such as sandy

granitic soils, clayey soils of depressions, termite

mounds and basalt and litholitic soils of the Lebombo

Mountains.

11. The sandy coastal belt, which has a tropical

rainy climate with an annual rainfall between 900

and 1 100 mm, is characterized by a number of distinct

vegetation types including marsh and swamp commu-

nities, coastal dune forests and pioneer sand binding

communities on the beach.

Thus, rainfall, the occurrence and severity of frost

and possibly fire appear to be the most important

differences related to the general physiognomic and

structural variation of the vegetation along the transect

from west to east across southern Africa. The vege-

tation becomes denser as the rainfall increases and

along the same gradient a sequence of structural

changes, related to growth form and cover, is apparent.

These range from desert vegetation of succulents and

other xerophytic growth forms to dwarf shrub steppes

or xerophytic grass steppes, shrublands or xerophytic

grasslands, shrub savannas with an understorey of

xerophytic grasses, tree savannas with an under-

storey of xerophytic grasses, mesophytic grasslands or

tree savannas and forests.

In the lowest rainfall areas vegetation is necessarily

absent or very open and vegetation structure is not a

prominent feature of the landscape. In the lower

rainfall areas soil depth determines whether the

vegetation is grassy or predominantly woody. The

woody vegetation occurs on the shallower soils and

is either dwarf shrub steppe, shrubland, shrub savanna

or tree savanna depending on the rainfall. Walter

(1962), in discussing the antagonistic relationships

between grasses and woody plants, ascribes the

occurrence of woody plants on shallow soils in the

lower rainfall areas to their coarse and extensive root

systems being better adapted than the root systems of

grasses to reach the non-uniformly distributed water

supplies in shallow and stoney soils. On deeper soils

in the more arid regions, the finely branched and

intensive root systems of grasses deplenish available

water, which then does not penetrate below the grass

roots and is thus not available for woody plants during

the dry season. In the higher rainfall areas, some water

is left by the grasses and this enables woody plants

to survive the dry season, their heights and total

cover increasing as rainfall increases.

In the higher rainfall areas, however, whether the

vegetation is grassland or savanna seems to depend

also on the absence or regular occurrence of severe

frosts and fire. Presence of the latter two factors

favours grasslands. Soil depth appears to be an

overriding factor as rocky soils carry a tree growth

560 A WEST-EAST VEGETATION TRANSECT THROUGH AFRICA SOUTH OF THE TROPIC OF CAPRICORN

even in areas where severe frosts and fire regularly

occur. Acocks (1953) suggests that the rocky areas

provide some protection against fire. It may be

speculated that microclimatic conditions between

rocks provide some protection for young stems against

frost and that moisture conditions favourable to the

extensive root systems of woody plants, together with

protection against fire, enhance the growth of woody

plants to bigger and more frost resistant individuals.

Protection against fire may be particularly important

for the growth of woody plants in areas with severe

frost, since fire damage is probably more severe if

preceded by frost damage.

Floristic variation within the main structural types

described above, is largely related to rainfall, severity

of frost, soil conditions, exposure, slope and aspect.

Carrying capacity for animal husbandry

According to data provided by Rattray (1960) the

main vegetation classes distinguished above coincide

largely with major differences in carrying capacity of

the vegetation. Rattray (1960) indicates that the true

Namib is entirely unsuitable for grazing in any form.

The remainder of the area under the desert climatic

regime is suitable for sheep and cattle farming and

has a carrying capacity of 2, 5-6 hectares per sheep or

26-35 hectares per beast. The area within the BSKw'

climate, comprising the transition from the Kalahari

to the grassland vegetation of the Ghaap plateau and

part of the grasslands of the upper central interior

plateau, have a carrying capacity of 8 to 13 hectares

per beast. Within the hotter steppe climates (BShw)

as far as Groblersdal, the carrying capacity is between

2,5 and 5 hectares per beast on sour grass vegetation

and between 5 and 7,5 hectares per beast on mixed-

sweet and sour grass vegetation. Between Groblersdal

and Sabie, with the exception of the Steelpoort Valley,

the carrying capacity on the sour grasslands and

savannas of the eastern escarpment in the Cw climatic

regime amounts to 1-3,5 hectares per beast, and the

vegetation is also suitable for sheep farming. From

Sabie eastwards over the zone with the hot and dry

steppe climate (BShw) the carrying capacity is between

2,5 and 5 hectares per beast. For the coastal belt with

the tropical rainy climate (Aw) Rattray (1960) gives

no carrying capacity figures.

ACKNOWLEDGEMENTS

We gratefully acknowledge the assistance of Mrs

G. E. Thomas who drew the diagram and maps.

UITTREKSEL

Veranderings in die oorheersende plantegroei-

fisionomie, prominente species en fisiografie, icings 'n

roete wat 7? 1800 km transek dear suidelike Afrika

naby die Steenbokskeerkring volg, word beskryf. Elf

hoof diskontinu'iteite in die struktuur en floristiese

samestelling van die plantegroei in die transek, hou

verband met 7 1 klimaatgradient oor die Kontinent.

Floristiese variasie binne die hoof strukturele tipes hou

lioofsaaklik verband met reenval, strafheid van ryp,

grondeienskappe, blootstelling, helling en aspek. Die

hoof plantegroeiklasse wat onderskei is, val saam met

hoof drakragverskille van die plantegroei.

REFERENCES

Acocks, J. P. H.. 1953. Veld Types of South Africa. Mem. Bot.

Surv. S. Afr. 28: 1-192.

Chapman, J. D. & White, F., 1970. The evergreen forests of

Malawi. Oxford: Commonwealth Forestry Institute.

Coetzee, B. J., 1974. A phytosociological classification of the

vegetation of the Jack Scott Nature Reserve. Bothalia 11:

329-347.

Coetzee, B. J., 1975. A phytosociological classification of the

Rustenburg Nature Reserve. Bothalia 1 1 : 561-580.

Coetzee, J. A., 1967. Pollen analytical studies in East and

Southern Africa. In E. M. van Zinderen Bakker (Snr.),

Palaeoecology of Africa 3 : 1-146.

Giess, W., 1962. Some notes on the vegetation of the Namib

desert. Cimbebasia 2: 1-35.

Giess, W., 1971. Eine vorlaufige Vegetationskarte von Siidwest-

afrika. Dinteria 4: 1-114.

Gomes Pedro, J & Grandvaux Barbosa, L. A., 1955. Esbogo

da vegeta?ao de Mozambique. In Esboqo do reconhecimento

ecologia-agricola de Mozambique. Centro Inv. Cient.

Algodeira. Memorias e Trabalhos 23: 67-226.

Grunow, J. O., 1965. Objective classification of plant com-

munities: a synecological study in the sour-mixed bushveld

of Transvaal. D.Sc. (Agric.) thesis. University of Pretoria

(unpublished).

Jackson, S. P. (ed.), 1961. Climatological atlas of Africa.

Lagos — Nairobi: CCTA/CSA.

Leistner, O. A. & Werger, M. J. A., 1973. Southern Kalahari

phytosociology. Vegetatio 28: 355-399.

Morris, J. W., 1973. Automatic classification and ecological

profiles of south-western Transvaal High veld grassland.

Ph.D. thesis, University of Natal (unpublished).

Morris, J. W. & Guillerm, J. L., 1974. The ecological profiles

technique applied to data from Lichtenburg, South Africa.

Bothalia 1 1 : 355-364.

Mostert, J. W. C., 1967. Veld types, problems and pasture

research in the Orange Free State Region. Proc. Grassld.

Soc. Sth. Afr. 2: 31-37.

Rattray, J. M„ 1960. The grass cover of Africa. FAO Agric.

Stud. 49: 1-168.

Roberts, B. R., Opperman, D. P. J. & Van Rensburg, W. L. J.,

1972. Introductory veld ecology and utilization. Bloem-

fontein: Univ. O.F.S.

Schulze, B. R., 1947. The climates of South Africa according to

the classifications of Koppen and Thornthwaite. S. Afr.

Geogr. J. 29: 32-42.

Schulze, B. R., 1965. Climate of South Africa. 8. General

survey. Pretoria: Weather Bureau.

Theron, G. K., 1973. ’n Ekologiese studio van die plantegroei

van die Loskopdam-Natuurreservaat. D.Sc. thesis, University

of Pretoria, (unpublished).

Tinley, K. L., 1971. Determinants of coastal conservation:

dynamics and diversity of the envornment as exemplified

by the Mozambique coast. Proc. S.A.R.C.C.U.S. 1971:

125-153.

Van Der Schijff, H. P., 1957. ’n Ekologiese studie van die flora

van die Nasionale Krugerwildtuin. D.Sc. thesis, University

of Potchefstroom (unpublished).

Van Der Schijff, H. P., 1969. A check list of the vascular plants

of the Kruger National Park. Pub!. Univ. Pretoria, Nuwe

Reeks 53: 1-100.

Van Der Schijff, H. P. & Schoonraad, E., 1971. The flora

of Mariepskop complex. Bothalia 10: 461-500.

Van Vuuren, D. R. J. & Van Der Schijff, H. P., 1970. "n

Vergelykende ekologiese studie van die plantegroei van 'n

noordelike en suidelike kloof van die Magaliesberg.

Tydskr. Wet. Kuns 1970:1 6-75.

Volk, H. O., 1966. Die Florengebiete von Siidwestafrika.

J.S.W.Afr. Wiss. Ges. 20: 25-58.

Walter, H., 1962. Die Vegetation der Erde in oko-physiologischer

Betrachtung. 1. Jena: Fischer.

Walter, H. & Lieth, EL, 1960. Klimadiagramm-Welt atlas.

Jena: Fischer.

Weather Bureau 1957. Rainfall maps. Pretoria: Weather

Bureau.

Werger, M. J. A., 1973. Notes on the phytogeographical

affinities of the Southern Kalahari. Bothalia 11:1 77-1 80.

White, F., 1971. The taxonomic and ecological basis of choro-

logy. Mitt. Bot. Staatssamml. Miinchen 10: 91-1 12.

Bothalia 11,4: 561-580 0975)

A phytosociological classification of the Rustenburg Nature Reserve

B. J. COETZEE*

ABSTRACT

The vegetation of the Rustenburg Nature Reserve, situated on the Magaliesberg in Acocks’s (1953) Sour

Bushveld Veld Type of South Africa, is classified by the Braun-Blanquet Method. Five major vegetation types,

including main subtypes, basic community types, variations and sub-variations are described floristically,

physiognomically and in terms of habitat features. The vegetation is mapped at community type and variation

level, at a scale of 1:30 000.

INTRODUCTION

The Rustenburg Nature Reserve is situated in

Acocks’s (1953) Sour Bushveld, which is listed by

Edwards (1972a) as one of 52 of the South African

Veld Types extremely lacking in conservation. The

Sour Busveld covers 18 306 km2, occurring in

mountainous areas in the Transvaal. The Rustenburg

Nature Reserve, which covers 2 896 ha, i.e. 0,2 per

cent of the Veld Type, falls under the Nature Con-

servation Division of the Transvaal Provincial

Administration and is the only area in this Veld Type

that is managed specifically and reasonably per-

manently for conservation (Edwards, 1972a).

The Reserve is situated between 25° 41' S and

25° 45' S and between 27° 9' E and 27° 13' E, on the

Magaliesberg, 2,5 km southwest of the outskirts of

Rustenburg. The area comprises the farm Rietvallei

824 and a portion of previous town land. Rietvallei

originally belonged to President Paul Kruger of the

Zuid-Afrikaanse Republiek, who used the area as

summer grazing for his horses. Later the farm

belonged to the Rustenburg Town Council from

whom it was obtained by the Transvaal Provincial

Administration.

A botanical survey of the Reserve was carried out

by Mr N. H. G. Jacobsen who supplied the botanical

information for completion of an IBP (International

Biological Programme) check sheet (Von Richter &

Jacobsen, 1970/1) and compiled a check list of 554

plant species occurring in the Reserve (Jacobsen,

1971). Several of the plant communities mentioned

in the IBP check sheet are confirmed as vegetation

types in the classification presented here. The plant

communities listed by Jacobsen were based on general

reconnaissance observation whilst the communities

described here were abstracted hierarchically by

comparing complete floristic lists from sampling

points spread over the Reserve. Inevitably, therefore,

the communities mentioned in the IBP check sheet

comprise vegetation types at different levels of the

present hierarchical classification as well as com-

binations of communities which are here not regarded

as together forming distinct vegetation types. The check

sheet also lists communities that, although distinct

in prominent species, are not so in total floristic

composition. These differences emphasize some

advantages of formalized semi-detailed surveys and

the need raised by Edwards (1972a) to standardize

and co-ordinate the recognition and identification of

plant communities for a reference framework of

South African vegetation.

*Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Other surveys in the Sour Bushveld include those

by: (i) Collett (1956) who gives a general description

of a small nature reserve on the Magaliesberg;

(ii) Van Wyk (1959) who described, at a broad level,

the vegetation of the Pilansberg, part of the Magalies-

berg and some hills in between these two mountains;

and by (iii) Van Vuuren (1961) and Van Vuu.ren &

Van der Schijff (1970). In the latter survey, com-

munities were identified on the basis of total woody

species composition. Woody species proved to be

strongly differentiating between plant communities,

also in the present study, and as Van Vuuren (1961)

also lists grasses and forbs occurring in the

communities recognized, his results can be integrated

with classifications based on total floristic com-

position. Communities that are strongly related to

those identified in the Rustenburg Nature Reserve

have been described from the Central Variation of the

Bankenveld by Coetzee (1972, 1974a) and from the

Sourish Mixed Bushveld by Du Plessis (1973) and

Theron (1973). These affinities stress the need for a

uniform classification system that can be expanded

by a process of integration and revision to include

all Veld types.

The classification presented here is based on the

the Braun-Blanquet method of vegetation survey,

discussed in detail by Westhoff & Van der Maarel

(1973) and Werger (1974a). This procedure is

recommended by the Botanical Research Institute for

making primary inventories of plant communities.

It is methodologically suited to defining plant com-

munities on the basis of considerations such as

those discussed by Edwards (1972b), Werger (1973a &

b, 1974b), Coetzee & Werger (1973, 1974a) and

Coetzee (1974b). The method has been developed

over a long period of extensive and increasingly

successful applications in Europe, outlined in historical

perspective by Werger (1973c), and is specifically

designed to facilitate the integration of results from

different workers in different areas into a com-

prehensive hierarchical reference framework in a

manner explained by Werger (1974a). Presentation of

results in table form has been standardized and the

wealth of information essential for rational use of

natural resources, readily apparent in such tables, is

discussed by Werger (1974b).

PHYSIOGRAPHY AND PHYSIOGNOMY

The part of the Magaliesberg on which the Reserve

is situated comprises recrystallized quartzite with

interbedded hornfels and diabase intrusions (Fig. 1).

The quartzite and hornfels are sedimentary rocks of

the Transvaal System (Magaliesberg Stage of the

SiDi

562

A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

Pretoria Series) and the intrusive diabase is of the

Bushveld Igneous Complex. The Reserve lies on the

summit, eastern slopes and foothills of the mountain

(Figs. 2 and 3).

Fig. 1 Geological map of the Rustenburg Nature Reserve

(from the Dept, of Mines, 1960).

Fig. 2— Topography of the Rustenburg Nature Reserve (from

Trigonometrical Survey Office, 1969).

Fig. 3. — Air photo showing

the boundaries of the

Rustenburg Nature Re-

serve and propos ?d camp-

ing sites (black discs).

The broken line shows

proposed new boundaries.

fi. .1. COETZEE

563

It is intended to enlarge the Reserve to include

part of the western slopes, which have been included

in the classification but not in the vegetation map

(Fig. 8).

Altitudes in the Reserve vary from 1 230 m in the

lowest part, on the eastern side of the mountain, to

1 660 m on the summit.

The western slopes of the Magaliesberg (Fig. 11)

are steep and underlain by diabase alternating with

hornfels, which weather faster than the quartzite

summit. The soils on the western slopes are litholitic,

mainly dark reddish-brown, with sandy clay-loam

texture. Evergreen Protea caffra- dominated woodlands

are found on steep, flat or convex slopes with no

water accumulation, whereas deciduous woodlands in

which Acacia caffra is dominant, co-dominant or sub-

dominant are found on concave slopes or on convex

islands of diabase, overlain by deep r.on-stoney soils

that carry evergreen Protea caffra- dominated wood-

lands, evergreen Protea gaguedi- dominated shrubland

and seasonal grassland, occur on the far northern

part of the plateau. Some of these deeper soils have

dark reddish brown orthic A-horizons and dusky-red,

dark reddish brown, reddish brown or dark red

B-horizons, while others have dark brown, orthic

A-horizons and yellowish red to red B-horizons.

The B-horizons have sandy-loam to sandy clay- loam

and clay-loam textures.

The southern plateau region is a basin with a flat

marshy area in the bottom, vegetated mainly by

dense Phragniites mauritianus- dominated seasonal

reedswamp. The marsh is at 1 425-1 440 m altitude.

Deep soils overlie the quartzite on the gentle slopes

rising from the marsh to the steeper brim where the

Fig. 4. — View from the north over flat to convex plateau area, with exposed quartzite in the foregnurdo

and the plateau basin in the background.

slopes below cliffs, where water accumulation is

considerable. Acacia caffra is strongly dominant in

cool mesophytic areas such as concave south-facing

slopes. On warmer west-southwest-facing convex

slopes directly beneath cliffs, Acacia caffra is also

dominant but with Combretum molle, Combretum

zeyheri, Dombeya rotundifolia and Vangueria infausta

as subdominants. Combretum mode and Pouzolzia

hypoleuca are dominant and Acacia caffra sub-

dominant on a very hot west-northwest-facing

convex slope beneath tall cliffs. Patches of semi-

deciduous forest occur in kloofs of the western

slopes.

Most of the Reserve is situated on a 2-3,5 km

wide summit plateau of quartzite, extending over

8 km in a north-south direction. The plateau contains

two geomorphologically distinct regions (Fig. 4).

The northern region is a predominantly flat to

convex area of exposed quartzite, 1 500-1 650 m in

altitude. At the highest part, the northern plateau

region divides the whole plateau into a northern and

southern catchment area. A considerable amount of

free perennial water originates in each of the catch-

ment areas, forming streams down the eastern side

of the mountain. The northern plateau region is

mainly a mosaic of lithosol and very shallow-litholitic

soils. The soils are gravelly, dark reddish-brown to

black, sandy to sandy-loam, with much decomposed

organic material. Areas of extensive sheet outcrop

carry a seasonal grassland vegetation with scattered

stands of widely spaced Lopholaena coriifolia shrubs

and a number of characteristic xerophytic grasses

and succulents. Semi-deciduous Landolphia capensis-

Bequaertiodendron magalismontanum Shrubland

grows amongst bouldery rocky outcrops. Two small

quartzite is exposed. The brim emerges at 1 440 m

altitude in the south and at 1 500 m altitude in the

north, east and west. The deeper soils of the plateau

basin are well differentiated over most of the area,

becoming gradually more litholitic towards the brim.

Deciduous Acacia caffra- dominated woodland, with

evergreen Protea caffra trees as sub-dominants,

occupy the far northern corner of the basin where the

soil has a dark reddish brown to dusky red clay-loam

orthic A-horizon and a dark reddish brown to dark

red clay-loam B-horizon. Other well differentiated

soils of the plateau basin have dark reddish brown

orthic A-horizons varying from sandy-loam to sandy

clay-loam, and dark red to dark reddish brown, mostly

sandy clay-loam B-horizons. Most of the area carries

seasonal grassland with isolated stands of deciduous

Burkea a/mY7m7-dominated and evergreen Protea

caffra- dominated woodlands and a stand of evergreen

Protea gaguedi-dommaitA shrubland. The litholitic

soils towards the edges are dark reddish brown, with

texture ranging from sand to clay-loam, and carry

grassland. The vegetation of the rocky quartzite brim

of the plateau basin is mainly seasonal grassland and

semi-deciduous Landolphia capensis-Bequaertiodendron

magalismontanum Shrubland as in the northern

plateau area. Faurea saligna trees fringe narrow

drainage lines down the rocky sides of the basin.

A northwest to southeast series of valleys, underlain

by diabase, separate the larger part of the summit

plateau from another quartzite summit area to the

east, which extends as a ridge from the northern

plateau region to the southeast (Fig. 5). The summits

of the western slopes of the series of valleys are

between I 570 m and 1 650 m altitude, the summits

of the eastern slopes between 1 450 m and 1 530 m

26700-13

564

A PHYTOSOCIOLOG1CAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

Fig. 5. — Valleys separating

quartzite ridge on the

right from plateau region,

the edge of which can be

seen on the left.

and the valley bottoms between 1 380 m and 1 440 m.

The valleys are drained along four lines that cut

through the eastern side. Small patches of semi-

deciduous forest occur where drainage lines run

through the kloofs. The four catchment areas are

separated by three transverse saddle-like watersheds.

Soils of this series of valleys are mainly

deep-litholitic and dark reddish brown with sandy

clay-loam texture. An interrupted seasonal grassland

zone is found on the upper slopes of the western

sides of the valleys, which are exposed to the north

and east (Fig. 6). Evergreen Protea cajfra woodlands

grow at the upper end of the series of valleys in

the north, on the concavo-convex watersheds and on

concavo-convex and convex surfaces of valley sides.

Deciduous Acacia coyfra-dominated woodlands occur

in the lower parts of the valleys and concavo-concave

surfaces of the valley sides. Soils of these Acacia

cajfra- dominated woodlands are markedly less stony

than those of the grasslands and Protea caffra-

dominated woodlands.

A lithosol-litholitic complex of sheetlike to broken

quartzite with semi-deciduous Landolphia capensis-

Bequaertiodendron magalismontanum Shrubland pre-

dominates on the steep upper northeast-facing slopes

of the eastern part of the mountain (Fig. 7). The

soils are gravelly, black sand to sandy-loam with much

decomposed organic material. Broad-leaved deciduous

woodlands with Burkea africana, Ochna pulchra,

Combretum zeyheri and Faurea saligna as prominent

trees, occur on the less rocky lower slopes and

foothills of the northeastern side of the Magaliesberg.

Soils here are gravelly, stony, brown to dark brown

Fig. 6. — Upper end of

northernmost of series

of valleys, with Rhyn-

chosia monophylla — Tris-

tachya biscriata Grassland

on the upper slopes of

the western sides of the

valleys, which are higher

than the eastern sides.

B. J. COETZEE

565

Fig. 7. — Steep, rocky, north-

east-facing slopes with

the Croton gratissimus —

Landolphia capensis

Variation of Landolphia

capensis — Be .uaertio-

dendron magalismontanum

Shrubland, viewed from

the lowland in the east

of the Reserve.

sand, sandy-loam and sandy clay-loam, and dark

reddish brown, predominantly sandy clay-loam

and clay-loam.

A low flat area of tertiary to recent alluvium

protrudes into the foothills in the east of the Reserve

at an altitude of 1 250-1 320 m. The deep soil is

differentiated into a dark reddish brown to dusky

red sandy clay-loam orthic A-horizon, and a dark

reddish brown to dusky red, clay-loam to sandy clay-

loam B-horizon. The B-horizon is gravelly in some

areas. Deciduous Acacia caffra- dominated woodland,

mostly with Combretum zeyheri and Dombeya rotundi-

folia as sub-dominants, is found in this part of the

Reserve.

CLIMATE

The following climatic data were recorded over a

period of 42 years at the Rustenburg-51 1/458 weather

station, 10 km northeast of the Reserve, and over a

period of five years at Little Quendon-51 1/432, 3 km

east of the Reserve (Weather Bureau, 1954). Average

monthly maximum temperatures are between 34 C

and 36 °C during the hottest months of October to

February and between 24,8 °C and 25,2 °C during

the coldest months of June and July. Mean monthly

minimum temperatures are highest (11,9 °C-13, 1 °C)

during December to February and lowest ( — 2 °C-

0 °C) during June to August. Ground frost may be

expected to occur on the average at least once per

month from May to September at these weather

stations, employing a Stevenson-screen temperature

of 3 °C as criterion for light ground frost (Schulze,

1965). Light ground frost or near ground frost

conditions may be expected to occur daily at these

stations in June and July when mean daily minimum

temperatures are between 1 ,8 °C and 3,4 °C.

The Rustenburg weather station is, however, at

1 119 m altitude and Little Quendon at 1 200 m

altitude, whereas at the Rustenburg Nature Reserve

the altitude varies from 1 230 m-1 660 m. Van Vuuren

(1961), who recorded temperatures at various altitudes

on the northern and southern side of another part of

the Magaliesberg over a one year period, found that

average weekly maximum temperatures were 1 ,82 C

higher at the northern foot of the mountain than at the

northern summit. Average weekly minimum tem-

peratures recorded by him were lower at the foot

than at the summit. Due to differences in radiation,

discussed by Coetzee (1974), temperatures are

generally higher on north-facing slopes than on

south-facing slopes, as found by Van Vuuren (1961).

Such temperature effects may be modified by dense

vegetation cover, under which Van Vuuren (1961)

recorded on the average less extreme values. Cold

air from the summits accumulates in the bottom of

the series of valleys between the plateau and eastern

summit ridge and concentrates in kloofs draining

these valleys. This was experienced at a camping site

at the bottom of one such kloof where cold gravity

winds, strong enough to be clearly felt, flowed down

the kloof during clear calm autumn nights. Similarly,

cold air from the summit pleateau will flow down

slopes and drainage lines and cold air south of the

plateau divide will accumulate in the bottom of the

plateau basin and escape through an opening in the

southeastern brim of the basin.

Winds are mainly light to moderate and blow

mostly from the northern sector in summer and

winter, except for short periods during thunderstorms

or weather changes when they have a southerly

component (Weather Bureau, 1960; Van Vuuren,

1961 ; Schulze, 1965). The Rustenburg Nature Reserve

falls between the 700 mm and 800 mm per year

rainfall isohyets according to a 1 :250 000 rainfall

map of the Department of Water Affairs (1966).

These figures are confirmed by records of 32-54 years

at Rustenburg-51 1/400 4,5 km northeast, Donker-

hoek-51 1/310, 1,8 km north-northeast, Baviaans-

kranz-51 1/404, 3,5 km east and Buffelshoek-51 1/285,

1 km southwest of the Reserve (Weather Bureau,

1965). The rainfall is reliable, being at least 85°/ of

the normal rainfall during 75-85% of all years, and

falls mainly during the summer months of October

to March when 85-90% of the normal annual rainfall

is received (Weather Bureau, 1957). The rainfall is

almost exclusively due to thunderstorms and instability

showers (Schulze, 1965).

566

A PH YTOSOCIOLOGICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

MANAGEMENT

The Nature Conservation Division of the Transvaal

Provincial Administration, whose policy is to conserve

natural areas, to introduce endemic fauna and at the

same time to provide recreational facilities, have been

allowing limited organized excursions into the

Reserve. A camping site has been provided on the edge

of the Acacia caffra woodland at the lower end of a

kloof on the eastern side of the mountain, and

mountain huts are being built in a number of sites on

the summit. Plans are in hand to extend facilities,

and an office complex and camping site is to be built

in the Acacia caffra woodland in the northern part

of the plateau basin. The house and store of the

Superintendent and the living quarters of his staff

are already situated in this woodland. Apart from

the camping sites, huts, living quarters and a few

concealed sand quarries, excavated to maintain roads,

the Reserve is unscarred by human impact.

The Reserve is fenced with a game fence and is

lightly stocked with a large variety of game species,

utilizing different habitats. There are no signs of

overgrazing and trampling. The following account of

larger game species occurring in the Reserve is based

on observations by the Superintendent, Mr J. de

Klerk. The figures given in brackets are his up to date

census figures (pers. comm.). Species found mainly

in the grasslands of the mountain plateau are

springbuck (38), red hartbeest (33), blesbuck (32),

Burchell’s zebra (23), black wildebeest (17), oribi (2)

and steenbuck (1). Sable antelope (21) are found

mainly in the woodlands of the plateau region, kudu

(10) are observed chiefly in the woodlands of the

series of valleys between the two summit areas, and

mountain reedbuck (73+) occur widespread on the

mountain slopes. Waterbuck (12) concentrate in the

densely wooded areas near water, i.e. in kloof forest

on the eastern side of the far northern plateau, in

Acacia caffra woodland near the marshy part of the

plateau basin and in nearby thickets. Impala (114)

and reedbuck (8) are usually observed in the

woodlands of the flats and foothills on the eastern

side of the mountain and in woodlands on the

plateau. Impala are also frequently observed in grass-

lands near woodlands. Klipspringer (16) are found in

rocky habitats all over the Reserve and Natal duiker

(16) are widespread. Rock rabbits are among the

conspicuous small mammals and live in large numbers

in rock crevices of cliffs. Predators known to occur

in the Reserve are leopard, brown hyaena, black-

backed jackal and caracal.

A burning programme for the Reserve has been

introduced recen ly by the Nature Conservation

Division. This entails periodic rotational burning of

certain areas after the first spring rains to remove

accumulated litter when this is judged to be in excess.

Areas thus burned are the grasslands, shrublands and

woodlands of the plateau basin, excluding the marsh,

and the grasslands and Protea gogz/ez/z'-dominated

shrubland on the deep soils overlying the diabase in

the far northern part of the plateau. Except for

accidental fires which have occurred from time to time

the rest of the Reserve is not burnt and is protected

by fire breaks.

METHODS OF SURVEY AND CLASSIFICATION

The Braun-Blanquet method of sampling and

synthesis followed here is reviewed and described by

Werger (1974a). Some optional sampling procedures,

which fall within the flexibility allowed by the method,

were introduced. The Braun-Blanquet method specifies

that the total sample should show as adequately as

possible the total variety in the study area. To achieve

this, sampling sites were stratified using 1 :360 000

air photos. After having become acquainted with

variation in the field, variation in vegetation structure,

dominant tree species and habitat was mapped on

the air photos. Twenty-one stratification classes were

obtained. The maximum sampling intensity was

approximately one site per 6,5 ha for 14 smaller

classes, each of which covered 64,8 ha or less. This

means that a proportionately larger number of

sampling units for smaller classes were considered

necessary only where the total number of samples in

the class did not exceed ten. These smaller classes,

which covered 496 ha (17% of the Reserve) required

74 sampling units (39% of the sample taken in

the Reserve). The remaining seven classes covering

2 400 ha received 116 sampling units. The minimum

number of sampling units per larger class was ten,

so that the sampling intensity for the larger classes

was approximately one sampling unit per 23,8 ha in

four of these classes and between one per 6,5 ha and

one per 23,8 ha in the remaining three. The final

vegetation map (Fig. 10), based on floristic tables,

virtually corresponds to the initial stratification map

on the air photos. This is due firstly to the prominance

of habitat features related to plant communities in

mountainous terrain, where strong topographic

differences are of major importance (cf. Van Vuuren,

1961; Theron, 1973; Du Plessis, 1973; Coetzee,

1974a); and secondly to the strong differentiating

Fig. 8. — Species-area curves

for: (A) Tristachya bise-

riata — Combretum zeyheri

Variation of Burkea afri-

cana — Ochna pulchra

Woodland (Sect. 3.1 . 1.1b);

(B) Asparagus krebsia-

nus — Landolphia capensis

Variation of Landolphia

capensis — Be uaertio-

dendron magalismonta-

num Shrubland (Sect.

3.2.1b); (C) Grassland

Variation of Digitaria

brazzae — Tristachya

rehmannii Woodlands,

Shrublands and Grass-

lands (Sect. 3.1.2c);

(D) Thesium trans-

vaalense — Eragrostis

nindensis Variation of

Cyperus rupestris — Era-

grostis nindensis Grass-

land (Sect. 3.2.2 b).

B. J. COETZEE

567

character of prominent woody species. Without the

formal floristic analysis, however, it is impossible to

determine the hierarchical level of the reconnaissance

classes and floristic and ecological relationships

between them.

A set of nested quadrats was placed in four of the

stratification classes, chosen for their dissimilarity

and wide occurrence in the Reserve, to obtain some

idea of the relationship between quadrat size and

number of species (Fig. 8). Three of the four resultant

species area curves (Fig. 8: A, B, C) show a marked

levelling off in number of species when exceeding 32 m2

(8 x4m) and the fourth showed a similar levelling

off after 16 m2 (4x4 m). A quadrat size of 50 m2

(5x10 m) was therefore considered efficient for

reducing qualitative floristic variance between samples

of very similar vegetation types. However, because of

the coarse structure of some of the vegetation types,

fication unit (Fig. 9). A quadrat size and shape of

10x10 m was adhered to throughout the survey

even though such rigidity is not prescribed and is in

some instances regarded as undesirable in the Braun-

Blanquet method. Quadrats were nevertheless suffi-

ciently homogeneous and representative to make any

change in size and shape unnecessary, although in

grasslands quadrats were usually unnecessarily large.

Tables 2-6 show the ordered sampling data.

More data from a wider area is essential before

community types can be ranked and before character

species can be distinguished. A capital “D” before

the name of a species in Tables 2-6 means that such

a species differentiates a particular community type

from all other community types in the Reserve,

whereas “d” means that the species is differentiating

for more than one community type which do not

form an exclusive type at a higher level in the

Fig. 9. — Distribution of sam-

ples in the Rustenburg

Nature Reserve with

boundaries of mapping

units (cf. Fig. 10).

a quadrat size of 100 m2 (10x10 m) was chosen to

obtain representative cover values for species. No

new floristic scale of pattern was encountered at this

quadrat size.

Sampling sites were placed randomly in the strati-

fication units on the air photos to obtain a represen-

tative distribution, but these points served only to

indicate the approximate position of the quadrats

which were then placed in the field in a visually

homogeneous stand, representative of the strati-

hierarchy. The other units involved are shown in

parentheses after the “d”. Communities have been

tentatively named, primarily by constant differen-

tiating species which are, wherever possible, also

conspicuous. The names remain, however, merely

symbols ; the floristic-sociological unit (phy tocoenon) to

which a particular stand of vegetation belongs must be

determined on the basis of total species composition

(Westhoff & Den Held, 1969).

568

A PH YTOSOCIOLOGICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

PHYTOCOENA

Each of the mapping units in Fig. 10 belongs to

one of the eight physiognomic types already mentioned

(Table 1). Physiognomic types that correspond with

distinct phytocoena are forests. Acacia caffra-

dominated woodlands, broad-leaved deciduous wood-

lands, Protea gaguedi-dommaizd. shrublands,

Bequaertiodenciron-dom'mated shrublands and reed-

TABLE 2. — Hypoestes verticillaris

swamps. The grassland and Protea co^ra-dominated

phytocoena cut across physiognomic boundaries at

higher levels of the hierarchy.

As indicated in Table 1, the phytocoena of the

Rustenburg Nature Reserve have been hierarchically

grouped into five main vegetation types.

1 . Hypoestes verticillaris — Mimusops zeyheri Forests

(Table 2)

; — Mimusops zeyheri Forests

Community number

Relov6 number

Slope (°)

Aspect

Soil texture (S= sand, sandy; C» clay; L= loam)

Soil colour (Bl« black; 6S3» dark reddish brown)

Outcrop ($)

Boulders

Large ctoncs(o=ab3ent; r= rare; fo frequent; a» abundant)

odium stones

Small stones & gravel

Differentiating species of Ilex mitis-Pittosporum

viridiflorum Forest (l.ll

D Pittosporum viridiflorum

D Halleria lucida

D Cyperus albostriatus

D Rothmannia capensis

D Ilex mitis

d(3«2.1. )Bequaertiodendron magalismontanum

D Blechnum attenuatum

D Secamone alpini

D Tyrsine africana

d(S.2.1bl) Haemanthuc ma{7iificus

Differentiating species of Diospyros whyteana -

Celtia africana Forest (l.2«l.)

D Diospyros whyteana

D Maytenus undata

D Solanum rubetorum

D Celtis africana

d(2.2b;2.2) Rhus pyroides

D Combretum erythrophyllum

D Acacia ataxacantha

D Acocanthera oppositifolia

Differentiating species of Ficus pretoriae -

Urera tan ax Forest (l.2.2. )

D Urera ten ax

D Cyphostemma cirrhosum subsp. cirrho3um

D Hermannia floribunda

D Enteropogon macro etachyus

D Pious pretoriae

D Droguetia woodii

d(S.2.i«) Croton gratiosimus subsp. gratissimus

D Corumelina benghalensis

D Setaria verticil lata

D Tagetes minuta

Differentiating species of Acalypha glabrata -

Dombeya rotundifolia Forests (1» 2. )

D Rhoicissus tridentata

d(2) Dombeya rotundifolia

d(2) RhuQ leptooictya

D Acalypha glabrata

Differentiating species of Hypoestes verticil lari:-

Tinusops zeyheri Forests (l.)

r o a

r o f

+ + +

5 4

3 2

2 + +

+ + +

2 3

2 +

1 1

+ 54

3 2 1

D Aohyranthos sicula

D Hypoestes vorticillaris

d(3*2.1) Fagara capensis

D Mimusops zeyheri

Infrequent Bpeciee

Combretum molle

Grewia occidentalis

Helinus integrifolius

Papoea capennis

Species of single or double oeciLncnce: Euclea crispa (relevS 196: (+); 56:+),

Oplismenus hirtellus (lB9:(+)i 19^:+), Panicum maximum (15:+; 54:+)» Pavetta

acsimilis var. pubescens (l5s+; 56i+), Solanum giganteum (166:+; 196:+),

Solanum nodiflorum (55*+» 54*+)» Acacia karroo (15:+), Brachylaena rotundata (166:+),

Bridelia mollis (15: + ), Duddleia s'alviifolia (166: + ), Cussonia 3picata (55*+)i

Diospyros lycioides subsp. guerkei (l5: + )» Durant a repens (15: + ), Ehrharta erecta (15:1),

Erigcron floribundus (15:+), Euclea natalensis (15:+), Olycino javanica (15:+),

Jrewia monticola (54j + )i Jletcromorpha arboroecone (l5* + )» Hibiscus calyphylluo (15: + ),

Littonia modesta (l5*+)» Tariscus indocomus (53x+)i T'yrioa pilulifora (166:1),

Tux ia congceta (56:+), Oeyrio lancoolata (l5t+), Phyllogeiton zeyheri (l5*+)

Ptericlium aquilinum (l66: + ), Sida r’regoi (15»+), Toucrium oapenoe (54:+)» Vauguerla

infausta (I5s+), Vepris undulata (196:+), Withania somnifera (54s+)» Ziziphus

mucronata (15:+).

table 1.

Relationship between the phytosociological hierarchy and physiognomic classes

PHYTOSOCIOLOGICALLY DEFINED VEGETATION TYPES

Hierarchical arrangement of mapping units

2

Eustachys mutica-Acacia caffra Woodlands (shown in Table

2.1 Combretum zeyheri - Acacia caffra Woodland

2.2 Brachiaria serrata - Acacia caffra Woodland

2.3

3 Loudetia simplex - Aristida aequiglumis Woodlands, Shrublands and Grasslands

3*1 Eragrostis racemosa - Diplachne biflora Woodlands, Shrublands and Grasslands

(shown in Table 4)

3.1.1 Sphenostylis angustifolius - Tristachya biseriata Woodlands and Grasslands

3. 1.1.1 Burkea africana - Ochna pulchra Woodland

3-1.1. 2

3-1.1. 3 Tristachya biseriata - Protea caffra Woodland

3-1.2 Digitaria brazzae - Tristachya rehmannii Woodlands, Shrublands and

Grasslands

3.2 Coleochlca setifera - Selaginella dregei Shrubland and Grassland

3.2.1

3.27lf

4

5

Mapping units

PHYSIOGNOMIC CLASS

Hypoestes verticillaris-Mimusops zeyheri Forests

(shown in Table 2)

(a) Not mapped (Kalanchoe paniculata - Acacia

caffra Variation)

(b) Digitaria smutsii - Acacia caffra Variation

FORESTS

ACACIA CAFFRA - DOMINATED WOODLANDS

(a) Blumea alata - Acacia caffra Variation

(b) Protea caffra - Acacia caffra Variation

Setaria lindenbergiana - Acacia caffra

Woodland

(a) Tristachya biseriata - Combretum zeyheri

Vari ation

(b) Silene burchellii - Burkea africana

Variation

Rhynchosia monophylla - Tristachya

biseriata Grassland

(a) Alloteropsis semialata - Protea caffra

Variation

(b) Cryptolepis oblongifolia - Protea

caffra Variation

(a) Elephantorrhiza elephantina - Protea

caffra Woodland Variation

(b) Monocymbium ceresiiforme - Protea

gaguedi Shrubland Variation

(c) Grassland Variation

Landolphia capensis - Bequaertiodendron

magalismontanum Shrubland (shown in Table 5)

Cyperus rupestris - Eragrostis nindensis

Grassland (shown in Table 5)

Aristida junciformis - Arundinella

nepalensis Grassland (shown in Table 6)

Pteridium aquilinum - Phragmites

mauritianus Reedswamp (shown in Table 6)

BROAD LEAVED DECIDUOUS WOODLANDS

GRASSLANDS

PROTEA CAFFRA - DOMINATED WOODLANDS

PROTEA GAGUEDI - DOMINATED SHRUBLANDS

GRASSLANDS

BEQUAERTIODENDRON - DOMINATED SHRUBLANDS

GRASSLANDS

REEDSWAMPS

o

(

I

h

CD

0

r

8

Qfl

o

"O

0)

£ o

(/)

o

0 -C

I

X

0-

_J

o

Q

< Q

Z

_L <

1 1 1 CD

8 Q

— LU

r— L_

a: 5

lu 9

< e

O-Q

0)

Q

Z

<

_i

Q

O

5

<

cc

LL

<

O

<

U

<

a

<

0

H

D

0)

>

1

o

<

H

0)

D

LU

03

0

b

03

a

<

1

co

4-»

D

£

(/)

0

u

03

4-4

‘o>

b

E

3

C

03

4-»

C

o

E

c o

0

o>

03

E

c

S

T3

C

0

T3

O

<D

03

3

a

0

CO

I

0

'0

c

0

a

03

o

I?

Q- to

O J2

■O 3

c *-

03 -C

_l CO

CO

Q

■ II

c

.5 £

4-1 O

CO 4_|

CO

'Z

CO

>

CO

0

o

CO

b

03

0

<

1

CO

4-1

(0

03

CO

0

E

2

CD

c

o

4-4

.CO

03

>

CO

k.

4-

4—

03

0

CO

b

CO

<

1

CO

L_

4-

(0

o

03

0

4->

o

k.

Q.

o

J_

Q

LU

: d

3

o

<

LU

I-

o

cr

a.

c

o

4-1

<0

CO

>

CO

03

O

.CO

o

CO

0

<

1

CO

c

CO

O)

S_

0

D

c

0

T3

c

CO

k.

CO

4->

0

0)

CO

Q

z

<

_l

CO

<

cc

o

CD

3

(X

X

CD

a

LU

5

1 <

CC

<

H

0

Q

1

LU

<

N

<

CC

DO

3

03

0

4->

0

L_

fl.

1

0

E

o

l

0

O

z

<

5

I

LU

*z E

$5 1

I CO £

< cc -

rn ^

o

i- °a

D

Q

O

LU

O

Q

LU

_J

I

Q

<

O

CE

CD

co

T3 oS

C

3

0

T3

c

<0

(0

CO

2

U)

CO

0

c

0

■O

c

to

0

T3

c

2

0

CO

co

2 0>

CO

O)

0 0

CO co

n

N

m CO

03 S-

> n

r

o

co

4-1

(0

c

(0 o

co co

CO

o

a.

<

z

I

o

0

1

<

z

<

o

cc

LL

<

LU

*

CC

D

CO

CD

O

CD

£

o

4 s

O) 0

a >

w

2 =

LU

CO

\_

4->

0

&

3

i_

CO

3

&

>>

a

.c

a

o

c

0

E

co

’co

5

o

c

>

r

.L T3

C c

c 2

0) o

. .

(0

0 p

!■«

(0 i-

•c r

■o

c

2

CO

(0

2

U)

CO

'(0

c

0

a

0

c

0

c

'-6

c

3

<

I

CO

£

i.

O

0

■o

4->

0

3

4->

0

i_

n

E

0

u

1

0

4->

0

'i_

0

>

£.

O

0

4-1

0

CC

Q.

<

H

<

C

o

4-»

0

s-

0

>

0

i_

4—

4-

0

o

CL

I

<

5

CD

3 Q

LU

or

£ ^0 < =

a

E

0

1

0

L.

0

3

C

3

4-i

'Z

3

0

E

0

4-4

E

U)

2

.c

CL

1

E

3

C

3

if

£

3

2

u

0

4->

— CL

C

o

4->

.0

L_

0

>

0

o

0

CL

I

0

U)

c

o

12

O

0

a

0

a

u

o

■O

c

3

0

o

0

4-4

0

u

0.

1

0

c

4-1

c

0

r

a

0

N

Z

c

0

£.

a

0

LU

<1>

>

L.

a>

00

<D

cal

<D

s_

rj

z

txo

X)

c

<u

4->

00

3

<U

Q_

al

E

c

o

4-J

ri

4->

a>

t>o

<u

>

txQ

a i °-

mmm

.

'• ■ ■ • • y • ,

ill

B. J. COETZEE

569

Forests are of small extent in the Reserve and

have a result been poorly sampled, particularly since

the releves cover three distinct forest types, each

represented by only two or three releves. One of these

forest types, the Ilex mitis — Pittosporum viridiflorum

Forest, has few species in common with the other

two forest types but is nevertheless more closely

related to them than to any other syntaxon in the

Reserve. The three forest types are therefore regarded

as belonging to an exclusive phytocoenon, differen-

tiated by a number of species as shown in Table 2.

Of these, Hypoestes verticil laris and Achyranthes sicula

are the most constant in all three forest types.

All three forest types grow in kloofs, which can

be either very hot and dry, or relatively warm and

permanently moist, or cool and less moist. The

marked floristic and structural affinities of these

forests, however, suggest that the kloof habitats should

be very similar. Woody species of the upper stratum,

being largely restricted to a particular forest type,

seem to respond most to the habitat differences

between the kloofs. Virtually only species of the lower

strata are responsible for the floristic affinities,

suggesting that the distinctive habitat similarities

are those that affect mostly these lower strata species.

The most obvious of such similarities are those

resulting from the dense upper canopy layer, such as

poor light penetration, less radiation heat received

and lost by the surface, less drying out of the soil and

air, and a mat of organic material. It appears, however,

that some uniform habitat conditions, independent of

vegetation structure and associated with kloofs, must

primarily exist to determine the tall, dense canopy

cover of all these forest types. This habitat feature

may be a concentration of drainage water deep

enough below the surface to supply the extensive and

relatively deep root systems required to support tall

forest trees. The primary habitat similarities between

the different kloofs thus appear to cause certain

structural similarities in the vegetation, which then

create the necessary habitat conditions for floristic

similarities.

I . I Ilex mitis — Pittosporum viridiflorum Forest

This forest type is strongly differentiated by its

dominant woody species. Hex mitis, Pittosporum

viridiflorum, Rothmannia capensis, Halleria lucida and

Bequaertiodendron magalismontanum. The latter

species also differentiates the Landolphia capensis —

Bequaertiodendron magalismontanum Shrubland (Sect.

3.2) where the species occurs as a shrub. In this

forest it grows into a 13 m tall tree. This forest is

found in narrow east-facing kloofs with perennial

streams or free underground water near the surface.

The dominant tree stratum has a dense uneven

canopy covering 90 to almost 100%, and is between

5 m and 13 m tall. Shrubs and small trees, up to 5 m

tall, cover less than 1% in the denser forest, but can

cover 20% in the more open forest represented by

Releve No. 166. Similarly, the forb layer covers 2%

in the denser tree stand sampled and 30% in the

more open tree stand. Blechum attenuatum is the

dominant forb under the denser tree canopy and

Cyperus albostriatus is dominant under the more open

tree canopy. The woody liane Secamone alpini, which

is also a differentiating species, appears in both

releves of this forest type. The tree fern Cyathea

dregei, which was not recorded in the releves, also

occurs in such forests in the Reserve.

The Ilex mitis — Pittosporum viridiflorum Forest,

although apparently much poorer in species, has

many characteristic species in common with

the Mimusops - C hr y sophy Hum - Apodytes dimidiata

Variation and the Mimusops-Chrysophyllum-Strychnos

usambarensis Variation of the Mimusops-Chryso-

phyllum Community described by Van Vuuren (1961).

The habitats are also very similar. These variations

occupy the most mesic habitats of sheltered parts of

the kloof on the northern side of the mountain.

1.2 Acalypha glabrata — Dombeya rotundifolia

Forests

These two forest types also grow in sheltered but

drier kloofs where there is no surface water or free

ground water near the surface.

(a) Diospyros whyteana — Celtis africana Forest

An example of this forest type is shown in Fig. 1 1.

Virtually all differentiating species are dominant or

sub-dominant woody species (Table 2).

Fig. 1 1 . — Western slopes of

the Magaliesberg with

Diospyros whyteana — Cel-

tis africana Forest in

southwest-facing kloof,

and Landolphia capen-

sis — Bequaertiodendron

magalismontanum Shrub-

land in the foreground.

570

A PHYTOSOCIOLOG ICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

This forest is found in relatively cool kloofs of

various aspects where there is no perennial free

water. Strongly differentiating species such as Celt is

africana, Combretum erythrophyllum and Diospyros

whyteana have also been observed to grow in riverine

forest in the Bankenveld where it is also cool but

moister. Van Vuuren (1961) observed that the

vegetation in a closely related community on the

southern side of the Magaliesberg is less tropical than

on the northern side and it is clear from one of his

diagrams (Fig. 8: Van Vuuren, 1961) that average

minimum temperatures in the community on the

southern side are lower during winter months than

on the northern side of the mountain where vegetation

similar to Ilex mitis— Pittosporum viridiflorum Forest

is found. Further more, two of the predominantly

north-facing releves of the forest type described here

(Releves No. 15 and 55) are situated in kloofs into

which a considerable amount of cold air drains

during winter from catchment areas in the higher

valleys. Low winter temperatures rather than moisture

deficiency appear therefore to be important habitat

features distinguishing Diospyros whyteana — Celtis

africana Forest habitats from those of Ilex ntitis —

Pittosporum viridiflorum Forest.

The upper stratum covers 70-90% and is

5-12 m tall, with Celtis africana , Combretum erythro-

phyllum and Mimusops zeyheri as dominants. A smaller

tree and tall shrub stratum, 2-5 m tall, covers

10-70% and a layer of shrubs and young trees,

0,5-2 in tall, covers 5-25%. A low stratum consisting

of varying proportions of grasses, forbs and small

woody plants covers 1-60%.

The Acalypha glabrata community described by Van

Vuuren (1961) and characteristic of the relatively

dry kloof on the southern side of the mountain, is

closely related to the forest type described here.

These two vegetation types have many distinctive

habitat features and many, though not all distinctive

species in common.

(b) Ficus pretoriae — Urera tenax Forest

Releves of this forest are on steep to very steep

north, northwest and east-northeast-facing talus

slopes. These slopes are hot an dry, probably with

much less cold air accumulation than in Diospyros

whyteana — Celtis africana Forest.

The dominant tree and tall shrub stratum is 2-5 m

tall, covering 85% or more. Dominant species in this

stratum are Mimusops zeyheri, Urara tenax, Rhus

leptodictya, Croton gratissimus, Dombeya rotundifolia

and Lannea discolor. Ficus pretoriae, another domi-

nant, is an emergent tree, up to 10 m tall and covering

5%. A shrub stratum, dominated by Acalypha glabrata

in some places and by Grewia monticola in others, is

0,5-2 m tall, with cover varying from 0,25-25%.

The lowest stratum in some places covers only

I % but in others up to 75%, with Enteropogon

macrostachyus, Droguetia woodii and an unidentified

species of the Malvaceae as dominants.

This forest type has weak affinities with the Croton —

Combretum Variation of the Croton Community,

described by Van Vuuren ( 1 961 ) as an ecotone between

sheltered mesophytic kloof forest and arid shrubby

vegetal' The variation described by Van Vuuren

has a habitat features and distinctive species in

commo vvith the forest type described here, but

seems -row in a more mesophytic habitat and to

contain more mesophytic species.

2. Eustachys mutica — Acacia caffra Woodlands

(Table 3)

Table 3 shows a distinct phytocoenon comprising

a number of closely related woodland types dominated

by Acacia caffra. These woodlands occur on flat

level surfaces with clay-loam soils and on slopes

that are probably nutritionally enriched and in some

places relatively mesic, due to water accumulation.

The latter is inferred from the geomorphology and

topographic position of these slopes, and the high pH

and conductivity of the soils relative to surrounding

areas suggest a higher nutritional status (Table 2,

3 and 4). The Acacia caffra Savannas on diabase and

in sheltered valleys, described by Coetzee (1974a),

belong to the same syntaxon as the woodlands

described here. The Eustachys mutica — Acacia caffra

Woodlands are divided into three main syntaxa,

which form a series from hot and xeric to cool and

mesic shown by their arrangement in Table 3. Some

of the differentiating species of Eustachys mutica —

Acacia caffra Woodlands do not occur in the extreme

xeric Kalanchoe paniculata — Acacia caffra Variation

whereas others are absent from the extreme mesic

Set aria lindenbergiana — Acacia caffra Woodland.

Brachiaria serrata — Acacia caffra Woodland

occupies the centre position in the series. This wood-

land has a number of differentiating species in

common with the moderately xeric Digitaria smutsii —

Acacia caffra Variation of the Combretum zeyheri —

Acacia caffra Woodland on the hot xeric end of the

series. The Protea caffra — Acacia caffra Variation,

which is the more mesic part of the central unit,

shares a number of differentiating species with the

Set aria lindenbergiana — Acacia caffra Woodland on

the cool mesic end of the series. Some of these latter

differentiating species are also shared with com-

munities in the Eragrostis racemosa-Diplachne biflora

syntaxon (Tables 3 and 4). The mesophytic part of

this series is therefore partly differentiated from the

more xerophytic part by floristic affinities with

vegetation found on soils that are probably more

leached.

2.1 Combretum zeyheri — Acacia caffra Woodland

This woodland is the more xeric of the Eustachys

mutica — Acacia caffra woodlands and has two

variations.

(a) Kalanchoe paniculata — Acacia caffra Variation

This Variation occurs below cliffs on steep, convex,

northwest to west-facing slopes on the western side

of the Magaliesberg, outside the present boundaries

of the Reserve (Fig. 11). The soils are litholitic, very

stoney (all sizes) and gravelly. Releve 14 is atypical

of this variation in species composition and habitat

(Table 3). The releve lacks virtually all differentiating

species of Eustachys mutica — Acacia caffra Wood-

lands, but does not fit better elsewhere.

The tallest trees (5-8 m) cover 1% or less. These

include Dombeya rotundifolia and Rhus leptodictya

on west-facing slopes and Pappea capensis and

Combretum zeyheri on northwest-facing slopes. The

dominant tree stratum is between 2 m and 5 m tall

and covers 30-50%. Acacia caffra is the dominant

tree on west-facing slopes and Combretum molle is

the dominant tree on northwest-facing slopes. Cover

of the shrub and small tree stratum, which is 0,5-2 m

tall, is 1-2% on west-facing slopes, but can be 20%

on north-facing slopes where Pouzolzia hypoleuca is

the dominant in this stratum. Grasses and forbs

cover 40-50%.

(b) Digitaria smutsii — Acacia caffra Variation

TABLE 3- “ Eustachys mutica - Acacia caffra Woodlands

•a

x> --j- to as ••

' -H rH r— I rH "

CQ + r+ (UVDVDCO

C • • as cd Q* tA

CD rH P -P *"W

T3 VO cd fl - 4-

a fl:

- aS E >>

»~s CD P .fl '

E CQ -H OJ

w •• fl bO P~>CQ -H P 4- O O'

VO w -H H O' O I) 1 ‘ Tl jj

as KA rH<orH;fl'aiA>,P

& '•do «

CO

3 CO

x: <d

p cd 'P

c S co -H

- fl S c o

•H d O (0

r — t 0.-H -H

1 !>j ft ^ ^

g ffg J

CH p <->

CQ • - cd 43 4- Onc

CQ + -

CO p<CQ P •

ft,D - KN J- E OO- + 3 **

>r4 3 /~> (tf (D CQ c\ •• flvo

fi) § 4* 6 •- CQ -H rH tA fl OJ

c -p +0,0^ fH ^

5 S&H -2^'Sw <0 +5 9

+ rH *rl •• IfN^a

» .. o o cr>

So o\ y .

B O H + ,^

-i •- E «S • *'0O ‘g

• • + O 'H + u\ C _ ”•

ON •• O 5 •• , E o

tTNOO P ® IA P ££

lA w ^ on w d h a

--V-' C*3 i— I P ■* S °

T) S .O

•H CO

M 3 o m

•H o T3 ~ ,

p o P -h flog o > 4-

- O H -H

P ’fl W 'o

. O CQ T3

A 00

•H <JN

,fl rH

P<

ft

- — -H W

5 -H W

bO c + +

' fl 4) •• ;•

-N o rH -=r Lf\

4- rH- CO KN

T3 i — i - c vo d o t ■

•h L3 ^ «3 rA ,fl p ••

o 4- -H — ' Q, P CQ

a) fl •• «h ;>» P on <

rH qiA ft CdrH P« rH

Co CO ON fl rH Cd _ — '

T3 -H H -P U CO

^ co x: c 'H fl -

PhP fl 4-

H O -H •• *

^ . CO P P«ON^

(D cd ON bO W On +

X cn r4 ^ <D -H H ”

+3 o" 4- • o P^a>

q *h •• » -fl c o>

d o E-t O CD -H rH

I — I (1) p

•H P« W

.3

co tfl

CQ rH Cd <D -3" CN-

ON J- OJ

rH VO

c TJ -P (0 uS

-oop< >y

^ q co e 4- co .

4- fl • O fl •• 2

.. P, <D VO

4- O (0 W O (\l - ■

P -H + P» P

cd fl

-:s

to p o

O O -H

4- O <0 rH r^s

> ^ — 1

co cd

C c VO

§ H

— -- CM ••■OP + >>

4- VO ON 4- (0 -H •• Q

— < P CO

OO , O ON •

- On •-

e oo •- ' •• -h x o

+ J- ^ CTN-^v 'K) ft ft

4- C\J vO t O ^ m Ij E (NJ O i P* W >c3

fO CP G CO 05 “H O

as CD p - U C <d

•H •- -W -P O P

Cd^Vn-HCTJ

0 + 4 'H + CD W

x-^P'-'OC -

••CO 4- ^ 4)/^

H ON •• g lA Cd U +

K\ rH OO ON -H W ••

— ON <h H C ^3 H

fe-f 3 1 1 13 « gr. vJ§a 1 t

Cflvo 4) H >> E^ - p

P O CM

cd -J-

xi

10 ON E -o

Z + ..55| 8,5 i g -'S -6P «® ° •§

LfN •• E -^«S2<5S^S® fll^oS - O — r-s cd c “H - CO co O £ O H Cd

4-4- - P -3- O

OHIt v» W “n ^

-P (0-» E p O H H ft H CO P, co

ft (fl ft OH (d ft

• - 4-

C OJ P

CO CD P P CO ’H

CO — fl +

-■3 . -

--n bO O •• P O- l

4- >> -H K\ P fA-

fl rH '

iu vu c n — Cd rH •

o P P co CQ ft o H

X3UMft3+3(H

* H ** P* ’H *"

3 to P - P Cd O OO

«M O 4- O

XZ CD O ON - E

N rH O C

\ a> fX *r

C K HH

• S

P OJ 4-

On •• — . , — _

E H [N fA CO -H

6 •• rH 4-00

.. « P t>- >5 • • ON

LTN X rO ON P f^V rH

w o

' X5 cdoo

• H P tA

i co cd"-'

a -

S S S r^- " '.£• g

o ) g B

- (0 at ta —

4-> -P 3 ^

+ t, <« h m co ■

.. to p Jp 'H 3 to

- J3 bO rH 1— 1 <o

o n. Cu - -P

I >5 O fbO' £ “

_.. J H b3 »r + CQ ■ — I -O t<

^oa-3-OO

01 IB rH -H 0. - •• ^

h ij ‘p

d; X - •- *H - 4- n>

<D -fl

. . - P V* +

«M rH 4- (0 .fl

•H O ■, -ri P ^ ^ ^

... -B ev«ij: I

::

-

■'* !#'. 1C. • 1 *d t la;

tiPJ.

vumam niuic i maupiijr nv

\KJ/ l£l tdl Id vMIIUUlll 1 11.UVIU VUI11U

,

B. J. COETZEE

571

Fig. 12. — View from the

Magaliesberg to the

flats between the north-

eastern foothills of the

mountain, with the

Digitaria smutsii — Acacia

caffra Variation of Com-

bretum zeyheri — Acacia

caffra Woodland.

This Variation occurs on the well differentiated

alluvial soils of the flats between the northeastern

foothills of the Magaliesberg (Fig. 12).

Emergent trees, 5-8 m tall, in some places cover less

than 1% but in other places 5-10%. A 2-5 m tall

tree stratum is always present, covering 15-35%.

Acacia caffra is usually the dominant tree with

Tree Savanna, described by Theron (1973), which

grows at the foot of slopes and in valley bottoms on

stabilized alluvial soils.

h2.2 Braciaria serrata — Acacia caffra Woodland

This phytocoenon which forms the central part of

the hot xeric to cool mesic series also has two

variations.

Fig. 13. — The Digitaria smut-

sii— Acacia caffra Varia-

tion of Combretum zey-

heri — ■ Acacia caffra

Woodland.

Dombeya rotundifolia, Combretum zeyheri and Zizi-

phus mucronata as sub-dominant trees. Shrubs and

young trees, notably Lippia javanica, Psidia punctata ,

Acacia caffra, A. karroo, Dombeya rotundifolia and

Ziziphus mucronata, form a 0,5-2 m tall stratum

covering 1-5%. Grasses and forbs cover 75-90%

(Fig. 13). Eragrostis acraea is a dominant grass in

places, but did not occur in quadrats. This Variation

has affinities with Acacia karroo — Setaria perennis

(a) Blumea alata — Acacia caffra Variation

This variation occurs on litholitic soils of the

lower north-northeast to east-facing slopes of the

series of valleys between the two summit areas of the

Reserve (Fig. 18). Two of the differentiating species

of this Variation also differentiate the Cryptolepis

oblongifolia — Protea caffra Variation, which is the

xeric variation of Tristachya biseriata — Protea caffra

26700—14

572

A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

Woodland and grows mostly on higher east-northeast-

facing slopes of the same series of valleys. Acacia

karroo is prominent, having distinctively high cover

values in the Blumea alata — Acacia caffra Variation,

and Rhynchelytrum setifolium seems to be significantly

constant.

The tallest tree stratum of 5-10 m can be absent,

< over 5-35% or, where the second stratum covers very

little, 65%. A second tree stratum, 2-5 m high usually

covers 20-25%, but can also cover only 2%. Acacia

caffra is mostly the dominant tree, with either A.

karroo or Lannea discolor co-dominant in places, but

Acacia karroo can be dominant with A. caffra sub-

dominant. A young tree stratum. 0,5-2 m tall, of

Acacia caffra and Lannea discolor covers 10% in

Quadrat No. 26 where the total cover is only 50%,

and a tall tree stratum is absent, probably owing to

the particularly steep (27°) slope. A young tree and

shrub stratum, 0,5-2 m tall and including Acacia

caffra, Dombeya rot undifolia and Artemisia afra,

covers 2% in Quadrat No. 27, where tall trees cover

65% and smaller trees cover only 2%. The grass and

forb stratum in the latter two quadrats covers

50-65% whereas in the other quadrats of this variation

this lower stratum covers 75-85%.

(b) Protea caffra — Acacia caffra Variation

This variation is restricted to the flats and gentle

slopes in the northern part of the plateau basin.

Most of the large number of more constant differen-

tiating species and about half of the approximately

40 species contributing significantly to its characteristic

distinctive habitat features between two extensive

vegetation types.

The physiognomy of this variation is shown in

Fig. 14. An upper tree stratum, 5-10 m tall, can be

absent or cover up to 4% and includes Acacia caffra,

A. karroo, Protea caffra, and Faurea saligna. A 2-5 m

tall tree stratum, with most of the trees from 4-5 m

tall and including Acacia caffra, A. karroo and

Protea caffra, covers mostly 45-60%, but can cover

75% where taller trees are absent (Quadrat No. 58).

Acacia caffra is usually the dominant tree with

Protea caffra sub-dominant but P. caffra can also be

co-dominant, or dominant with Acacia caffra sub-

dominant. A young tree and shrub stratum, 0,5-2 m

tall, covers 0,5 to 1% and includes Rhus eckloniana,

R. pyroides, Acacia caffra, Diospyros lycioides, Arte-

misia afra, Ziziphus mucronata and Lippa javanica.

The grass and forb layer covers 70-90%.

2.3 Setaria lindenbergiana — Acacia caffra Woodland

(Fig. 15)

Setaria lindenbergiana, a distinctive differentiating

species of this variation, also differentiates Landol-

phia capensis — Becjuaertiodendron magalismontanum

Shrubland (Table 5), but has distinctively high cover

values in the Setaria lindenbergiana — Acacia caffra

Variation.

The variation is found on cooler aspect slopes on

the western side of the Magaliesberg and in the

valleys between the two summit areas. The slopes are

Fig 14. — Physiognomy of the

Protea caffra — Acacia

caffra Variation of Bra-

chiaria serrata — Acacia

caffra Woodland; Acacia

caffra (A), Acacia karroo

(B) and Protea caffra (C).

species combination, are shared with vegetation types

of the Eragrostis racemosa — Diplaclme biflora syn-

taxon (Table 4). The Protect caffra — Acacia caffra

Variation is therefore transitional between Eustachys

mutica Acacia caffra Woodlands and the Eragrostis

racemosa — Diplaclme biflora syntaxon (Table 4), two

distinct vegetation types at a broad level of the

hierarchy presented here as well as of the hierarchy

described by Coetzee (1974). Both these broad

vegeh ;on types are among the most widespread in

the Sour Bushveld. The homogeneous ecotonal

situation occurring here could be studied more

intensively to gain information on extreme states of

flat to concave, below cliffs or usually low in the

topography where there is probably water accumu-

lation. Soils are litholitic.

Five to 10 in tall trees, including Acacia caffra,

A. karroo, Cussonia paniculata and Faurea saligna, are

usually present, in some places covering less than

2% and in others covering 15-20%. A 2-5 m stratum

covers 20-60%. Acacia caffra is the dominant tree.

A. karroo can be co-dominant and Dombeya rotun-

difolia and Cussonia paniculata can be sub-dominant.

Shrubs and young trees, 0,5-2 m tall, can be absent

TABLE 4. - Eragrostis racemosa - Diplachne biflora Woodlands, Shrublands and Grasslands

+ +

K\

CVJ

+ +

+ 4*

4- +

+ +

4- 4- 4-

4- + CM CVJ

4- 4- CM KV

rH + + CVJ

CVJ 4- 4- KV

+ HN

H 4- 4-

+ + 4-

+ + +

+ +

-f

CV! +

4-4-4-

+ +

+ + +

+ 4-

+ + T“

+ 4-4- +

4- 4-4- + +

4- 4- 4- 4- 4- + +

+ + + + 4-4-4-

4- + + 4- 4- + _+_

+ 4-

CM

+ +

■+ +

+ +

CVJ +

W +

H +

iv: +

CV. +

CVJ +

+ + + +

+ H +

+ + + +

4-4-4-

+ +

4-4-4-

+ rH +

4-4-4-

+ 4- +

+ +

4- 4- + +

rH

4-4-4-

CVJ 4- 4- rH

CM + CVJ CVJ

CM +

CM i

KV

+ +

+ 4- +

+ + +

+ 4- +

+ + +

+ 4-

+ ♦

+ +

rH + CVJ

CM + + +

CM + CM CM

H CM H KV

+ + KV

CM + CM

+ H + KV

+ + KV CM

.. — CM

On + i — I

+ rH CM

4-4-4-

+ 4-

• pv 4- O' P c3 On -h

KV

4- +

4- 4- 4- 4-4-4-

4- 4- + 4- +

4- +

+ 4-

CM 4-4-

CM CM

CM

+ + + -T

+ 4- +-d-

+_+

+ + rH

+ + H H

+ + + rH

4- 4- + rH

+ +

+ + +

+ + rH

+ +

CM +

rH + +

KV

+ + 4-

4- 4-

4- +

4-4-4-

+ 4-

4- +

4- 4-

CM 4-

+ 4-

CM 4- +

_rH +_+_+ rH _

KV + + 4-

KV 4- + 4- +

KV 4- + 4- KV

4- +-+ +

+ 4-

+ +

4-4-4-

+ 4- 4-

4-4-4-

4- 4-

4- + 4- 4- 4-

4-4-4- +

rH + + KV

CM rH + CM

4- 4- 4- KV

4- 4-

4- 4- + 4- rH

4 + + + rH

+ + + + rH

+ + + CM

+ + CM CM

+ + CM

KV +

+

KV +

4- +

4- 4-

4- +

4- rH + rH +

+ +

+ + +

CM + 4-

CM + +

CM + + +

CM + + + +

CM + + +

rH + +

+ + rH

+ +

+ + +

rH + +

CM

+ + +

+ + + +

CM+++ +

CM+rH+H-

KV+rH+H-

CM + + + +

KV + + + +

CM + +

CM + t-

CM + +

+ +

+ + + CM

+ + + +

+ + +

+ + + rH

H + CM rH +

+ + + H-

H + + CM +

+ + CM

+ +

CM ♦

CM + +

+ rH + + +

+ + + + + + +

+ +

+ CM rH + + + +

+ + + +

+ +

rH +

+ +

r* + +

CM

CVJ

KV + +

KV rH +

KV + +

+ CM

+ OJ

•f + +

+ + +

+ +

+ + +

+ +

+ + + +

+ +

+ + + +

+ +

+ + + + + +

+ CM + rH + + +

+ + + + +

+ + + +

+ +

CM + CM

rH + + OJ

+ + +

+ + + rH +

+ + CM +

i — I + + CM +

CM + + rH

+ + +

+ +

+ + +

+ +

+ +'

+ + + +

+ + +

+ + + +

+ +

+ rH

CM +

KV +

CM +

+

+ +

+ + +

+ +CM +

+ CM + +

H + rH

+ + +

+ +

rH + + CM rH

H- + +H +

CM + + CM CM

+ + + CM rH

CM + CM rH CM

+ CM

+ rH

+ +

+ + + +

+ + +

+ + + +

+ + + + +

CM

CM +

+ +

CM CM H-

+ +

+ + + + +

+ + +

+ +

+_+_ +

+ + +

H + +

+ +

+ CM +

+ H +

H + +

+ KV <

CM + +

CM + + + +

KV + + v- +

rH +

+ +

rH +

CM +

+ + CM +

CMCM+^ + H-H-H-*

H- + +

+ +

4- + + + +

4-4- + 4-

4- +

4- 4-

2 % ^

rt o

CO «H rH «H

•H & -H

JH o

'H H (C '

(C 3 -H 1

CD CO

■zs

s &

bdi <D

CQ O > i-l fi

Q T3 ft '*) Q

CO

§ §■

i T3 «

CQ

ft 0 H i

CD rH

CO ft -H

» ft +> s>

ho <d H

•H OH

rH 4-> «H

CO CQ

CQ *H CO ft

ft -H CC

CO O H-> •

^ A £ .

ft U 'O V U'HriH

rJCQ^cO^ftcOS

CO>sOft<DCOOO,

12 -s.ei

'OT3QQQQfiQrCJQQ

o

&

ft cO cO

hO CO

CD CD

-P CD

ft ft!

2 -H *H

& & .O

I ^ rH

I CO CQ ^

cO

X CQ o ft *H CO CD

O hO CD ft H ft -H

C cq H o CQ CQ ft

«H O

.... "H ft

q&.'OflOQflOOftflQ

E ft E ft

ft rH ft O

o cO ft

<H &0 O <H

•H ft ft CQ

<M CQ Cd CD

CO U ° O

•H ft CC CD

• f-r p

i *H »H ft

1 O rl ft

: -h a; o

: ft w s

KV CD

CM ftc ol CM (

OQ^O ftpft 1

n-h

I el ’«!

CO CQ -ft

ft 6> "Eo

w § Tl

H tQ ft ft ft

O E O ■

ft CQ -H E Cd

E -H ft ft _

ft ft |>5 CQ cd

CD CD

ft- rH O ft ft

CQ O CO ft O

O -P N O CQ

Xi Ph *rj .H ft

ft l>iftH Id

•H ftft

ft ft, W M <-» vu O'

EH CQ <: O CQ ft ft

£■3

ftl ft

« Ph

Q Q OJ ^

ft

cO KV

ft JJ CO

«.!

> ho cO O CQ

nd ft > '

-P O O CQ ft -rl

O >> _ >> ft P rft

jft cd > -h g .

CO O ft cO rH ft cO

CD (0 ft'O CO S; CD

O -P £ ftft -f, Xi

E CQ rH CD ft, ® X>

O «H (0 O cO cO

hO*H cO CQ

O -p H o

ft, CQ ft XI

O CD CQ O

(0 B ft cO s, -P

u rl w ft PsfttftftH

<<i;o^oooo«

r-N KV

X X ^ ^N “

- ftD CD

CM CM jp CM *H

CM CM <M CM <D

s y * — ' ' — ^ V — X ft

-0 0 0-0 -OQ-flqQQQ-OcQ

•HH-O M

cO ft cO CD CCS

ft O^ -p A .H

ft H ft to ^

O (0 S ft

ft X CD ft

ft O > H

rft * -

OJ KV ftjr

M* I CM

O ft ft

P CO O CO H ft

CO -P -H -P rH

-ft- CO ft CO CD ftH

ftp CO ft .a O

•H i — I

P, CD i

CO CQ CD 1

° a,1-

cO CO > -H CD

W O C [Q CQ pj

S

CO CQ

H O

>3 ft

O

ft CH

CO CD

CO cO rH ^

CQ O CO f>s

O CQ »rl .ft <D

S <D rO P, N

CD ft CD

§<D

CQ

P CO O P O ft

CO ft CQ CO ft

O CQ g

ft S <D

cO E -P

CQ -H ft *H CQ

O P CD P ft

ft <D P CQ <D

ft CQ ft ft O

CO CD hO CQ

CD CQ

r3-

,0 *H

O cO

d .

•h cr ft ft

- CQ ^ CO

•H kft

§ ‘H

E CO cO P *H

CD E ft -

5 e:s s w-

P (0 ft

ft *H

- ^ „ cO j>sp

eE *H O *ri rl H <D

hO ft O ft P, 0) cO •

_ ft COO <D t>3 CO *rl R> H 1 ^ rt w

CQfqO^>P,Q,QC4QftOW^CQ

•H hO CQ

O -P CD

CM CM CM H CM

• CM CM CM CM CM

ip-OOflO-O-O1

O CQ H

«H ft — »

S -rl P

CO rft ft

CO •

O ft

cO

e - s1 •*

CD CO •.

ft CQ CO

•H O hO CQ

•S'd § g

P XI

<D O

, ;,rft CO

hC .H ft

hO <D

rH _

ft o h

ft CQ

PO»H«v~ — n —

>5 MH ^h^P.

» CO CO <D rH

-O H ft H

•H P P 0)

_ _ _ P CD hO ft

a OX CQ T3 -H o

£> cO rH *rl ft ft ft - -- ■ • . ■

ftftft O ft*H E^r^sCD CD

CD P, P

rft ^ CQ

0^0

-ox

hO CQ -H CQ CO

O -H W O ft

p, P O -H Cd O

n n ft o O "H

2 •£

ft cd p, ft ,

.. , _ . . ft >t, CD CD rH CO

EHffl^PHft^OEHCQCa

CM

CM CM

OJ CM CM CM CM CM CM

OJ OJ KV KV OJ KV KV

, r . § S T. 1

...... -H rs OS O

- O'- 4- O Q -H OO -rJ

4- ^ On r. : » o

- ^ c£V rH ^ •

3 ft

If

4» - 0> >3 O'

rH ^ X ft H

O 4- 7S^X

O KV ^

« O' B

ft On ft r

r- - - C - ^ ■

O X p '

P cc

«C ft ft

ft CD P

> <h CQ CQ

ft 4) 4) O' o

CO P. rH

y 0) bO •- ft

ft CD 4- -H

CO P, •• X

rH 3 r-N ft ft U

+ O ft

x: •• cc a

p« - o cc -h «c

*■ o H -H N C

- •• C 4- w ft C -h

^ o O)

OQ P

o o CQ UV

- ^

CO T3 CQ

,, ..PC

XX CO ft p. + P

•rl .O

w —i X CO

CQ OO ft

cO X

. _ _ _ _ -o

4- + -H -H P ft

.. £ ^ O ft- - r-

CO •- <D rH > + P

-- i — , w . rH 4" rH ft O • " ® H C

C^ftXrHrH--XCD>rtXOOOCNp

Sp o o OO P, CQ 'W >3 rH p

4> X O •- ,-v^ O -rl X — - o

P -P CQ P, ft + 4- X "H ^ CQ y ft

(Q ft 3 O •• •• COO >. 4-

o-. r.i .H O -S' >» -H + P

S O ON X ft ^ JO

W CC O' ft

-X COP • rH . rH -r t-1

-f (0 ft CQ<5v^Efl>

+ £lKo,,-H' ®§H ^cOC»-2cOn£'CX

i x

J X -r

T3 CP

O -

■ X On C

X ft E E

ft P

CQ

N -H

CQ .r( rH

X O

CO ft <H

« ft .H

O rO

> ft CO O CD

CO TJ

ra 'So

ft c rft ft CO

•H X

~ tQ '

ft X

-O ft CD rH CO .rH •

IQ ft E hO E

ft Oft

co cd ft u

CD CD

ft

cO X

o -H 'ft

'P-H C

CQ CQ CD

eg

I «*H •T5

O ft O *H CO E Cd ft 5

I rH CD O >3 ft O

bO rft 2 rM *H P ft

C ? 3 2+J

a) t» cr co c

P, ft CD ft ft

CO ft CO r*r*

P CO rft P -H>3r>3Cfl

rr ^ cd -d

CQ O *H

• ft 0)

cO C ft ft ft 5

•H p T3

rH ft O

O CQ p X

CQ CQ

E E

•H *H Cfl CQ t

p, P, rH P CO

(Q >3 CD CDrP*

<: k ft n 1

s 2 ft

I <D O

? PH'td

I >3 *H

I O 2c

E r-H

0> w <0

a O O CQ CQ -H bO

c p p CQ O P,

§ ft ft cO *H P, r

CQ X ft *

O P

C/Q < ■

ft XrP

ft ft o o o

bO hO rP *H*H

Cd Id P- rHrH

w w ^ &

OP..

CC ft 'ft

-p rH ft

cofcg

bO CO <0

2S.2 S

CO §C «

2 rH O Q

— —oo

- — p, — q ft- ”

(• 4- O 4- rofflwKV

• •• X + •• .O P -H rH

“ •• !>- rH CO (On/

, . irv'-' co o

n vfl w ft a»

(0 O ft rP “

+ ft <o

' Q* <D

IH fl O ffl U •'

LTV *H -rl X O-

ft p ' — ■ - -

l ,o* CO

p; x

KV OO W rr, r,. « ^ ; y ^3

_ _ . . ,cr* co -

(C - ft .H ^ P. P <D «0 ft rP

^J-+ljr^C00004><DrH<D • — '

-5 ^ tfl + H H ^

+ .. y-.pV-.pO P 'ft -H CC 'H H 'h

.. —nr bo lTv co -h 0) ft ' ? 3 o >'7,o O ffl

2 * ^ s s w s m -s § .t §

C KV — 0 >3 OQ + P -COftOOrH-Hrt

— « X - ^ c ^ ^ ^ S ^ 'c

cs + . + -h P ft 4- Co i—i ft

§ ^ P, .. X 4-/> UVP o CO 'ft ® ^

•• CDhT P.OJ*- + OO O w ^ o H ft E

XJftrHrQCT'O O OO •• E -J ft -H P- O X •'

ft rH rft OO LTV — 3 rH

W ft O t>- + Oft

y W p — •• .rl cfl t>

1 Cd CQ .-ovhP.H

s P P -H + + ^ ^

I CQ yr-> CO — ' W X3

Vi/^- o yp4-fcd •• C^V . 4J T r-.

-o y h to O •• VO VP P -H

P CO 'ft p 'ft OJ O-'-' *CJ < P«

O O cfl VO -W -H -H r-' -r • *ri

E E P O H -H H »-J3 + •• h H W ?

py ypi — ” m -CQ rH CQOvS > P P TJ

1 'i * 3 TJ i 55 S r . R 8£!w ■ 3 5 -g

sisals ..a-g s a|g.s

^ --H VO CQ ft CQ -H - P rH O

— y ^

p, p 4- X U

P P,

CQ O

§23 ■

>3 4-

O • X ••

> W POO

g(?3

+ 4-

*> 6 O

(\Jw -ft Cfl

v.pp,idp

bO£ O' ffl E O ffl CQ

CJN CTN

CQ 3

+ — + — W ft |

• • 4- •• — 4- *H H

OO .. -0- ^ P O

CO KVOO 4-00 10 rQ

y CQ

~ P-

•H P -H

ft O "

X CO

p ora

ft - rH

P

CC o

■H bC ^ -

Cfl O P rft

O X

O X X

rH ft

hO Cfl (0

hO CC

CO X

ft • O TO U ft

OX E p P, P

W ^ X O cfl ft

P, ft P, - P, P *- CD

ftOOft^*-1 ^ =»

ON rft ft E W

.w , « . H P-d (0

.. t + KVH S O - CD C -H - O -

o ft E^ • - EH < rH 4-

OJ CM •• 0W04-4- Cfl 4- —

~ -C' .«*» h-V o

ON On 4- 4- E VO <D '

+ — rv oO Cfl

- TJ •• -

^ OO ft -H

P cfl

CQ -

ft +

^ OO ■

- oo

- - *■ E ^

O CM •• ft

I CM ON ft 4- y

•h ra y ft co

ft <2

T3 ft >3 CD

"o 2? 23 ’£

.5 e

« cd ^ ^ +

LTV o •• Cfl W P< CO ~

rH rH rH ft Cfl ft rH

^ <m in y cq p ra^ro

ffl irvo E ft ho co

— ■ - * E ra P, ■ • w >• i r rz

ft P, O ^ X ft VO p, ffl

ft P O H AJr.

SP rH ' -*-

O <Z

Cfl CQ < X

ra Cfl

E 'ft -ft

' CM ft

.. 4-

on Ah

f?S

IP ”

x> oo

° IA E

(0 -H +'

tf hO

i

P X x

-oo

N oo

cfl KV X

P,

« -n « r. • - • • P

-•HEP,E<h<ra*H** - *M ' ft

- » ft -H ft H O H (^J-7 o PCQ

ft E CQ P <0 H ft VO 4- bO • CO P

— r. ft r \ r4 M I . . a • r~| 4J

CM TH IQ ft -H

O E OO P CDft

ni o H 1 ca s '®®wM2a-g

■S^^e -3.

N ^ O 5 §-H 4-1ffl4- CQ

¥ Eh ON ft ft +» ” O • • 3

.. a m •• h 0) ra-HO t‘"N a

- o O E cfl rH « rH p co

H Q

-vO P

4- H ra -H P

ft P, -r4

P, ra rP

&

p, >3 ra — •• ft - cq ffl

cc ft cu^> rH ra^s ra ra co-

+ Cn <H 4- -H ffl O

co - ra K o E P,co

P CV.

'ra +

O r^^HCO Cfl Cfl

y ft ft ft aS ■

•H VO ft T3

3 X

^ ft

' Co bo *o y W X

v H 4) ffl -H 'H O o

-ft P P.-H -H ffl 10,

Cfl X <<

LTV CM

5 •

P *rl

‘H r-C 'y uy -r w

>, CO 4- Cfl— '•••H^ftON

-ft _ O h ^ ra irv

X Cfl VO o 4- P rH

P [v-

Cfl ft VO Cvfl

^ Cfl — H rH -H ft

•H 4- — ' C/3 X

OO 4- Cfl - Cfl •

— ' •• *H

O rH

v ^ ^ rr> m ~ -

Cfl P ON

to S

+”"+ -5 (d rp H 3

.S

C£g

H vO *H ffl + O

co vy o ^ ^ ^ rj.

y_|«M— rft— ;4-ra cfl ^ O ^

^ vH ..CO Cfl.. X • -H ^5 o

_ - +-- Q o p, P X — ft

<d <j\ ft “

cfl

>3 'ft cfl O-v

^ 'S cfl S ^ c

— cfl ft 3 *H E

cfl cq ra S X "o O

§5 l Sftil-g

y P 3 X ft rH ra

*+. *.r. tsttggvjsli

.. .. vp o ra cfl -h •• >. ft ft 4- -h ra

+ VO U> ^ h H ® M CV.

Cfl CO CflOv^O cfl 4-

k o .,-QP ~ X CQ P Cfl x ON . ft ‘rj “

I— . ^ “ rr««<._ir-»_ 4 at .p C Os

„ _ O .

CQ P, CQ v

- X CQ P Cfl X ON

CO 3 O rH

w M CQ ra - COP."

Xoo C C ft Cfl ffl Cfl

y 4- -H ft

tsl •• rH ^ w w

on uv bo • 4- cfl co <♦-, -o

•H X '

- ft

C IA O

CKr-s ffi 10 ^ ^

ra 5 “ + — ft <3 5 « S -H « +

23 g £§03 S.S

3 y NOrHCflCOftftrHft>ra — 'rH

X — . . co py» — y E co ra

® § ^ » 1- 5 -* ra •£ 2 fc 13 .2 5

+ r .ft 4) rap bo ffl be

~w--**^'^rHNX ft P ^

P, ra p ON -H CO 4-P cfl ft | - -H rarH

ID K] 4) ^-r' poo •* •• X cfl ra > ft" •*'

_i 3 oooa-ho 4- idffl

[n. y ft •• cfl o

— y aS -vO -H *

.^.-n O rH ^ rH P 05'^'

(Q n nw + 4-y Cd 4- — • ft ra 4*

y x >s 1

px to ra

cfl - p

& i £ ■£ §§ 0,03 . z 5^

S_H-y y ft ft n aj\q ia ra rax p Q cd ao

o

57

W

fa

kc

va

an

no

( o

lit

co

no,

ka

Ac

do

Ac

Qi

an

the

shi

caj

CO'

65'

for

50-

thi

(

slo

M(

tiai

40

spec

of

taxc

Var

mut

race

dist

hier

desc

vege

the

situ;

intei

5/

w

fa

kc

va

ar

cc

( o

lit

a

' -aA

v •

*•

slo

M<

tia

40

-

‘

spe<

of

tax<

Var

mat

race

dist

hier

des(

vegt

the

situ;

inte

B. .1 COETZEE

573

Fig. 15. — Physiognomy of

Setaria lindenbergiana —

Acacia caffra Woodland.

or cover up to 20% and include Artemisia afra, Rhus

eckloniana, R. discolor, R. leptodictya, Lippia javanica ,

Acacia caffra, A. karroo, Acalypha glabrata and

Diospyros lycioides. Grasses and forbs cover 70-80%.

This variation has affinities with the Acacia caffra —

Setaria lindenbergiana variation described by Van

Vuuren (1961) as growing in temperate and moist

habitats below cliffs with water accumulation, on

the southern side of the Magaliesberg. Another related

vegetation type was described by Coetzee (1974) as

Acacia caffra Savanna on diabase, where the soil is

relatively moist and presumably base-rich. This

Variation also appears to have affinities with Acacia

caffra — Setaria perennis and Faurea saligna — Setaria

perennis Tree Savannas described by Theron (1973).

3. Loudetia simplex — Aristida aequiglumis Wood-

lands, Shrublands and Grasslands (Tables 4 and 5)

This widespread phytocoenon can be classified with

some of the Sourish Mixed Bushveld of the Loskop

Dam Nature Reserve (cf. Theron, 1973), the Chert

Vegetation, Shale Vegetation and the vegetation with

abundant large boulders, on quartzite outcrops and

on massive chert outcrops, described by Coetzee

(1974a) from the Jack Scott Nature Reserve in the

Bankenveld and the Chrysophyllum community

described by Van Vuuren (1961). The vegetation is

differentiated by a large number of species, common

to Table 4 and 5, which have a wide distribution in

the cooler and higher rainfall areas of the Transvaal

(cf. Coetzee & Werger, 1975). In the Rustenburg

Nature Reserve the Loudetia simplex — Aristida aequig-

lumis syntaxon grows on the more leached soils

with low conductivity and pH. This vegetation

includes broad-leaved deciduous woodlands. Protea

cajfra'dominated evergreen woodlands, Protea gaguedi

-dominated evergreen shrublands, Bequaertiodendron-

Landolphia-dommated semi-deciduous shrub-land and

seasonal grasslands. These form two broad types, the

first occurring on deeper litholitic and better developed

soils, and the second on bouldery outcrops and on

a shallow litholitic and lithosol mosaic with

extensive sheet outcrop.

3.1 Eragrostis racemosa — Diplachne biflora Wood-

lands, Shrublands and Grasslands

Table 4 shows the vegetation on litholitic and deeper

soils, excluding the vegetation on very shallow

litholitic soils found in areas of extensive sheet out-

crop.

3.1.1 Sphenostylis angustifolius — Tristachya biser-

riata Woodlands and Grasslands

The vegetation associated with litholitic soils

includes three major syntaxa which seem to grow on

separate parts of a complex gradient associated with

altitude and soil nutrients: the first, which is deciduous

woodland (3. 1.1.1), occurring on the foothills on

the northeastern side of the mountain (with an

exceptional variation growing in deep soils on the

plateau); a second, which is Protea ca/fra-dominated

woodland (3. 1.1. 3), found on slopes in the series of

valleys between the two summit areas and on slopes

on the western side of the mountain; and a third,

which is grassland (3. 1.1. 2), growing on the lower

pH and conductivity soils near the summit and on the

plateau. The deciduous woodland variation on the

plateau (3.1.1.1b) occurs in areas that appear to be

nutritionally richer than the grasslands of litholitic

soils as indicated by the frequency of comparatively

high soil conductivity (Table 4). The pH values of the

deciduous woodland variation on the plateau,

however, are, like the grasslands, generally lower than

those of the Protea caffra- dominated woodlands and

the deciduous woodlands of the foothills. Releve 42

of the deciduous woodlands on the plateau is an

exception with a relatively high soil pH and supports

these suggestions because it includes Protea caffra as

well as Oxalis obliquifolia of the same group of

differentiating species for Protect cq^ra-dominated

woodlands (see also discussion in 3.1.2a).

3 . 1 . 1 . 1 Burkea africana — Ochna pulchra Wood-

land ( deciduous woodland)

(a) Tristachya biseriata — Combretum zeyheri

Variation

574

A PHYTOSOCIOLOG1CAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

Fig. 16. — The Tristachya bise-

riata — Combretum zeyheri

Variation of Burkea afri-

cana — Oehna pulehra

Woodland on the eastern

side of the mountain,

in the foreground.

This variation occurs on the litholitic soils of the

foothills on the northeastern side of the Magaliesberg

(Fig. 16) and is differentiated from the related

variation on the plateau by Combretum zeyheri,

Combretum molle and other species of the same group

as shown in Table 4. The two species mentioned also

differentiate Combretum zeyheri — Acacia caffra

Woodland from the other more mesic Eustachys

mutica — Acacia caffra Woodlands, suggesting that

the variation of the foothills grows in a more xeric

habitat than that of the variation on the plateau.

TapiphyUum parvifolium also differentiates Landolphia

capensis — Bequaertiodendron magalismontanum Shrub-

land, which, like the variation on the foothills and

unlike the one on the plateau, occurs on litholitic

soils. Releves include various hotter and cooler

aspects except the cooler southerly aspects from

southeast to south-southwest.

Trees from 5-10 m tall cover less than one to 20%,

usually less than 6%, and include Combretum zeyheri,

C. mode, Burkea africana, Faurea saligna and Oclma

pulehra. A 2-5 m tall tree stratum covers from less

than one to 15% and includes the same species as the

taller stratum as well as Lannea discolor, Strychnos

pungens, Mundulea sericea, Ximenia caffra and

Ozoroa paniculata. A 0,5-2 m tall stratum of young

trees and shrubs, including all the species of the upper

stratum as well as Cryptolepis oblongifolia, Lannea

discolor, Bequaertiodendron magalismontanum, Tapi-

phyUum parvifolium and Ozora paniculata, usually

covers 5-15% but can also cover only one per cent.

Any one of the woody species in the upper two strata

can be dominant and in some places a number of them

are co-dominant. The grass and forb stratum covers

60-85%.

(b) Silene burchellii — Burkea africana Variation

This variation differs considerably in species com-

position, habitat and appearance from the former

(Fig. 1 7). The position of the Silene burchellii — -

Fig 17. — Physiognomy of the

Silene burchellii — Burkea

africana Variation of Bur-

kea africana — Ochna pul-

chra Woodland growing

on flats in the plateau

basin and dominated by

Burkea africana trees.

B. J. COETZEE

575

Burkea africana Variation in the lower parts of the

plateau basin in areas of water accumulation, indicate

that it is more moist than the Tristachya biseriata-

Combretum zeyheri Variation of the same woodland

type. The former is probably also cooler, as a result

of its position on the top of the mountain and the

accumulation of cold air in the plateau basin during

winter.

Burkea africana trees, 5-8 m tall, are usually present

covering 1% or less on the shallower soils and up to

25% on deeper soils. Smaller Burkea africana trees

(2-5 m high) cover 25-55%, where taller trees are

sparse or absent. A 2-5 m layer of Protea caffra trees

covering up to 20%, occurs in quadrat No. 42

(pH=5,9) under denser stands of the taller trees.

Most of the gentler slopes in this Protea caffra-

dominated woodland are found in the northern upper

end of the series of valleys where three of the quadrats

concerned are situated (Nos. 43, 44 and 45).

Protea caffra is the only species in the tree stratum,

which is 2 to 4 or 5 m tall and covers 15-25%. Young

Protea caffra trees, 0,5-2 m tall were sometimes

present. Grasses and forbs cover 70-85%.

(b) Cryptolepis oblongifolia — Protea caffra

Variation

This is the more xeric of the two variations and

occurs on 9-29° slopes in the valleys between summit

areas. Aspect is mostly east-northeast but east and

west-southwest aspects were also recorded.

Fig. 18. — Tristachya bise-

riata — Protea caffra

Woodland in the fore-

ground, growing in the

series of valleys between

the two summit areas. On

the opposite slope to the

right is the Blumea alata —

Acacia caffra Variation

of Combretum zeyheri —

Acacia caffra Woodland

and, in the background,

an example of Rhynchosia

monophylla — Tristachya

biseriata Grassland.

A 0,5-2 m stratum, including Burkea africana.

Protea gaguedi, Ochna pulchra, Vangueria infausta and

Rhus eck/oniana, usually covers 1-2% and grasses

and forbs cover 65-15%.

3. 1 . 1 .2 Rhynchosia monophylla — Tristachya biser-

riata Grasslands

Grassland belonging to the Sphenostylis angusti-

folius — Tristachya biseriata syntaxon, occurs on non-

rocky litholitic soils of the upper valley slopes between

the two summit areas and of the plateau basin (Fig.

23).

3. 1.1. 3 Tristachya biseriata — Protea caffra Wood-

land

This woodland comprises the Protect caffra-

dominated vegetation of litholitic soils (Fig. 18).

(a) Alloteropsis semialata — Protea caffra Variation

This mesic Variation occurs on 10-35° slopes with

southerly aspects found in the series of valleys

between the summit areas in the Reserve and on the

western side of the Magaliesberg outside the Reserve.

In four of the five quadrats with the lowest slope

angles (10,5° to 17,5°) 75% or more of the un-

vegetated surfaces are covered with accumulated

grass and leaf litter. This condition should be

monitored since the recently introduced burning

programme aims to protect these areas from fire.

Protea caffra trees form a 2-5 m stratum covering

15-25%, exceptionally 5%. Cryptolepis oblongifolia

shrubs and young Protea caffra trees, 0,5-2 m tall,

are usually present, covering one per cent or less.

The grass and forb stratum covers 75-85%.

3.1.2 Digitaria brazzae — Tristachya rehmannii

Woodlands, Shrublands and Grasslands

This vegetation occurs on the deeper, well differen-

tiated soils of the plateau and includes Protea caffra-

dominated woodland, Protea gagwcr/Z-dominated

shrubland and grassland. All these variations are

found on well drained slopes and depressions of the

lower areas on the plateau where water accumulation

can be expected. Distinct differences between the

habitats of these variations and their sub-variations

are characterized by position in the landscape,

geomorphology, soil texture, colour and pH.

(a) Elephantorrhiza elephantina — Protea caffra

Woodland Variation

Soil pH in this variation is not as low as in the

shrubland and grassland variations (Table 4). Higher

pH values also distinguish the habitat of Protea

co^ra-dominated woodlands on litholitic soils from

that of the grasslands on litholitic soils (3.1.1.). These

two Protea cq^ra-dominated types have a group of

differentiating species, including P. caffra, in common.

576

A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

Fig. 19. — The Elephanterrhiza

elephantina — Protea

caffra Woodland Varia-

tion with 5 m tall Protea

caffra trees.

In the strongly related vegetation of the Jack Scott Na-

ture Reserve in the central Bankenveld, the only

difference observed between the habitats of grasslands

and Protea caffra- dominated woodlands on chert, over-

lying dolomite, was a marked difference in thickness

chert cap (Coetzee, 1972, 1974). Protea caffra wood-

lands occur where dolomite is close to the surface

and although soil pH was not measured this difference

does suggest comparatively higher soil pH in the

Protea caffra woodlands. There were no indications

of climatic differences between the two adjacent

areas.

(i) The first sub-variation occurs in flat, low but

well drained drainage belts. One of the differentiating

species, Setaria flabellata, is restricted to the three

quadrats with sandy clay-loam and clay-loam A-

horizons. B-horizons are dusky red. Two to 5 m tall

Protea caffra trees cover 40-60% and grasses and

forbs, up to 2 m tall, cover 75-85% (Fig. 19).

(ii) A second sub-variation is found at the end of

long slopes just before and on the steeper decent to

drainage lines, steep drops or further slopes and on

slightly elevated areas from which water drains

outward in all directions (Fig. 23). B-horizons are

dark red to dark reddish brown. Protea caffra trees,

2-5 m high cover 5-30% and grasses and forbs up

to 2 m tall cover 80-90%.

(b) Monocymbium ceresiiforme — Protea gaguedi

Shrubland Variation (Fig. 20 and 23).

Fig. 20. — One metre tall Pro-

tea gaguedi — dominated

shrubland belonging to

the Monocymbium ceresii-

forme — Protea gaguedi

Variation of Digitaria

brazzae — Tristachya

rehmannii Woodlands

Shrublands and Grass-

lands.

B. J. COETZEE

577

Protea gaguedi-dommatcd shrubland occurs on

concave slopes with deep soils in lower parts of the

plateau. The shrubs, mostly 0,3-1 m tall but up

to 1,75 m, cover 10-15% and grasses and forbs

cover 70-80%.

(c) Grassland Variation

Grasslands belonging to this vegetation occur on

deep soils of flat to convex slopes of the lower plateau

areas, excluding those slopes described for the

Elephantorrhiza elephantina — Protea caffra Woodland

Variation. The total cover is between 60 and 85%

(Fig. 23).

3.2 Coleochloa setifera — Selaginella dregei Shru-

land and Grassland

The vegetation shown in Table 5 includes a shrub-

land type occurring on broken bouldery outcrops and

a grassland type found in areas with extensive sheet

outcrop.

3.2.1 Landophia capensis — Bequaertiodendron

magalismontanum Shrubland

This vegetation occurs on broken outcrops with

plants growing in cracks, fissures and litholitic soil-

pockets. In some places trees are dominant in the

upper stratum. The shrub growth-form is, however,

usually very prominent and many woody species that

occur widely as trees, are often shrubby in this vege-

tation. Trees are often also characteristically stunted.

Two main variations of this community type were

found in the Reserve: (a) one on the very steep,

northeast-facing slopes of the Magaliesberg; and

(b) another on the summit plateau.

(a) Croton gratissimus — Landolphia capensis

Variation

This variation occurs on 27-29° xeric north-facing

slopes of the mountain (Fig. 7). The extensive quartzite

outcrop is broken but low and flat.

An upper stratum of trees and tall shrubs, 2-4 m

tall, including Bequaertiodendron magalismontanum,

Croton gratissimus and Combretum zeyheri, covers

2-25%. Shrubs, 0,5-2 m tall cover 30% and include

Bequaertiodendron magalismontanum, Landolphia

capensis and Tapiphyllum parvifolium. Grasses and

forbs cover 5-55%.

(b) Asparagus krebsianus — Landolphia capensis

Variation

The variation on the cooler plateaux is further sub-

divided :

(i) The Faurea saligna — Landolphia capensis Sub-

variation is found mainly on steep southerly and

easterly-facing slopes along deep drainage lines. Trees

and tall shrubs, 2-5 m tall, usually cover 20-30%

and include Faurea saligna, Bequaertiodendron maga-

lismontanum, Brachylaena rotundata, Canthium

suberosum, Rhus leptodictya, Nuxia congesta, Tapi-

phyllum parvifolium and Combretum mode. A shrub

stratum of 0,5-2 m tall, covers 2-16%. Grasses and

forbs cover 10-15%.

(ii) The other sub-variation occurs on the more

open parts and on slopes with northerly aspects (Fig.

21). Shrubs and trees vary from 0,5-5 m, but are

often lower than 3 m, and cover 15-40%. Lower

shrubs, grasses and forbs cover 5-20%.

3.2.2 Cyperus rupestris — Eragrostis nindensis

Grassland

This grassland is found on the plateau in areas

that are a mosaic of extensive, flat unbroken sheet

outcrop and litholitic soils (Figs. 22 and 23). Cover

of outcrop in quadrats varies from 1-95°% but these

differences do not distinguish the two variations

found :

(a) The Coleochloa setifera— Eragrostis nindensis

Variation, differentiated by Coleochloa setifera, occurs

mainly on black soil. Lopholaena coriifolia shrubs of

up to 2 m are usually present, covering less than 5%.

Grasses and forbs cover from 15 to 55%.

(b) The Thesium transvaalense— Eragrostis nindensis

Variation found mainly on dark reddish brown soil,

is transitional to the Eragrostis racemosa—Diplachne

biflora unit occurring on deeper litholitic dark reddish

Fig 21. — Stand of Landolphia

capensis — Be^uaertioden-

dron magalismontanum

Shrubland amongst out-

crops of large, quartzite

boulders on the plateau.

578

A PHYTOSOCIOLOG ICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

Fig. 22. — Stands of Cyperus

rupestris — Eragrostis

nindensis Grassland grow-

ing on very shallow

litholitic soils and in

cracks in a mosaic of bare

sheet-outcrop and shallow

gravelly soils.

brown soils (Table 4). Grass and forb cover in this

variation is between 40 and 50% and Themeda

triandra has characteristically constant high cover

values as opposed to the former.

4. Aristida junciformis — Arundinella nepalensis

Grassland (Table 6)

This vegetation occurs in slightly elevated areas,

with relatively high water table, fringing the Pteridium

aquilinum — Phragmites mauritianus Reedswamps

(Sect. 5) and small streams, and in shallow submerged

marshy areas (Fig. 23). The vegetation has a few

infrequent species in common with the reedswamps

but species occurring elsewhere in the Reserve are

rarely encountered. As shown in Table 6, the

dominant species vary with different water table

depths and soil characteristics. The Rhynchospora

g/m/ca-dominated quadrat in the submerged area

with stagnant water had a floating layer of iron

bacteria.

5. Pteridium aquilinum — Phragmites mauritianus

Reedswamp (Table 6)

The reedswamp occurs in the bottom of the plateau

basin in a mass of water, humic material and roots,

which reach down to below 2 m depth (Fig. 23).

The Phragmitis mauritianus plants grow 4-5 m above

the surface of the water and cover 30-90%.

Fig. 23. — Part of the plateau

basin with stands of:

Pteridium a uilinum —

Phragmites mauritianus

Reedswamp (A); Aristida

junciformis — Arundinella

nepalensis Grassland (B),

which fringes the reed-

swamp; the Grassland

Variation (C), the Mono-

cymbium ceresiiforme —

Protea gaguedi Shrubland

variation (D) and the

Elephantorrhiza elephan-

tina — Protea cafTra

Woodland Variation (E)

of Digitaria brazzae —

Tristachya rehmannii

Woodlands, Grasslands

and Shrublands; Cyperus

rupestris — Eragrostis

nindensis Grassland (F);

and Rhynchosia mono-

phylla — Tristachya bise-

riata Grassland (G'.

13. J. COETZEE

579

TABLE 6. — Aristida junciformis — Arundinella nepalensis Grassland and Pteridium aquilinum — Phragmites mauritianus

Reedswamp

m.P r • . \ 7 I , nypencum lalandii

*+’ Leersia hexandra (182:+; 185:+), Polygonum strigosum (l8l:+: 182:+),

Adenostemma caff rum (178:+), Albuca setosa (176:+), Aloe davyana (l74:+), Anthoxanthum sp.

(.180 +), Asclepias fruticosa (180:+), Chironia palustris subap. palustris (174* + )

Crassocephalum picridifolium (l78:+), Cyperus sphaerospermus (l76:+), Epilobium

tetragonum (179: + ), Enosema salignum (175:+), Euphorbia striata (l76:+), Gnidia

microcephala (174:+), Haplocarpa scaposa (174:+), Helichrysum kuntzei (174:+), Heteropogon

contortus (176:+), Hibiscus aethiopicus (175:+), Hyparrhenia dregeana (l74:l), Hypericum

aethiopicum subsp. sonderi (l76:+), Indigofera sp. (175:+) Ledebouria sp. (174:+)

Nemesia fruticans (l76 :+) , Paspalum orbiculare (180:+), Pellaea viridis (l74 :+) , Scirpus

GAn i 8 )( 8™l,8scens U74:+), S. isatideus (175 : +) , S. serratuloides

(180. +), benecio sp. (168:+), Sotaria perennis (l76:+), Thelypteris confluens (l78:+).

580

A PHYTOSOCIOLOGICAL CLASSIFICATION OF THE RUSTENBURG NATURE RESERVE

6. Releves No. 175 and 176 (Table 6)

Releves No. 175, dominated by Pteridium aquilinum.

and 176, dominated by Scirpus burkei, are included

in Table 6 because they have a few species, mostly

infrequent, in common with Aristida junciformis —

Arundinella nepalensis Grassland and Pteridium

aquilinum — Phragmites mauritianus Reedswamp, but

are otherwise very poor in species.

Both releves are from the predominantly grassland

fringe around the reedswamp.

ACKNOWLEDGEMENTS

This project was launched by the Botanical Research

Institute preparatory to more extensive surveys of the

Transvaal Bushveld. Permission to study the Reserve

was granted by the Nature Conservation Division of

the Transvaal Provincial Administration, who also

provided accommodation and other facilities, parti-

cularly through the kind co-operation of Messrs

Delmyn Pretorius and J. de Klerk, the conservation

officers on the Reserve. Miss L. Teversham did most

of the technical work, and Mr S. Makena assisted

me in the field. The soil samples were analysed for

pH and resistance by the Soil and Irrigation Research

Institute. Plant identifications were done by the

herbarium staff of the Botanical Research Institute.

Dr J. W. Morris handled the computer work of

accurately printing Tables 2-6 in the desired form

before publication and Mr O. J. H. Bosch made useful

suggestions during a field inspection of the soils.

All these contributions are gratefully acknowledged.

U1TTREKSEL

Die plantegroei van die Rustenburg-natuurreservaat

op die Magaliesberg in Acocks (1953) se Suurbosveld

Veldtipe word met behulp van die Braun- Blanquet-

metode geklassifiseer. Die floristiese samestelling, fisio-

nomie en habitatkenmerke van vyf hoof plantegroei-

tipes, insluitende hoof subtipes, basiese gemeenskaps-

tipes, variasies en subvariasies in die Reservaat, word

beskryf Die plantegroei is op die vlak van gemeen-

skapstipes en variasies teen ft skaal van 1 : 30 000

gekarteer.

REFERENCES

Acocks, J. P. H., 1953. Veld types of South Africa. Mem. Bot.

Surv. S. Afr. 28: 1-192.

Coetzee, B. J., 1972. 'n Ptantsosiolugiese studie van die Jack

Scott-Natuurreservaat . M.Sc. Thesis, University of Pretoria.

Coetzee, B. J., 1974a. A phytosociological classification of the

vegetation of the Jack Scott Nature Reserve. Bothalia

1 1 : 329-347.

Coetzee, B. J., 1974b. Improvement of association-analysis

classification by Braun-Blanquet technique. Bothalia

1 1 : 365-367.

Coetzee, B. J. & Wf.rger, M. J. A., 1973. On hierarchical

syndrome analysis and the Ziirich-Montpellier table

method. Bothalia 11: 159-164.

Coetzee, B. J. & Werger, M. J. A., 1974. On association-

analysis and plant communities. Vegetatio (in press).

Coetzee, B. J. & Werger, M. J. A., 1975. A west-east vegeta-

tion transect through Africa south of the Tropic of Cap-

ricorn. Bothalia 11: 000-000.

Collet, D. G., 1956. Some botanical features of the Wonder-

boom Nature Reserve. Fauna & Flora 7 : 66-87.

Du Plessis, J. C., 1973. 'n Floristies-ekologiese studie van die

plaas Doornkop in die distrik Middelburg, Transvaal. M.Sc.

thesis, University of Pretoria.

Edwards, D., 1972a. Conservation areas in relation to veld

types. Unpubl. MS, Bot. Res. Inst., Pretoria.

Edwards, D., 1972b. Botanical survey and agriculture. Proc.

Grassld. Soc. Sth. Afr. 7: 15-19.

Jacobsen, N. H. G., 1971. A floristic survey of the provincial

nature reserves, Transvaal. Unpubl. Report, Project

T.N.6.3.3, Div. Nat. Cons., Pretoria.

Mines, Dept, of, 1960. Geological Survey 1:125 000. Sheet 4

(Rustenburg). Pretoria: Dept, of Mines.

Schulze, B. R., 1965. Climate of South Africa. 8. General

Survey. Pretoria: Weather Bureau.

Theron, G. K., 1973. Ekologiese studie van die plantegroei

van die Loskopdam-natuurreservaat . D.Sc. thesis, University

of Pretoria.

Trigonometrical Survey Office, 1969. Topo-cadastral map

of South Africa 1: 50 000. 2527CA Rustenburg ( West).

Pretoria: Trigonometrical Survey Office.

Van Vuuren, D. R.J, 1961. 'n Ekologiese studie van die plante-

groei van ~n noordelike en suidelike kloof van die Magalies-

berge. M.Sc. thesis, University of Pretoria.

Van Vuuren, D. R. J. & Van der Schijff, H. P., 1970. ’n

Vergelykende ekologiese studie van die plantegroei van "n

noordelike en suidelike kloof van die Magaliesberge.

Tydskr. Natuurwet. 1970: 16-75.

Van Wyk, J. J. P., 1959. 'n Sistematies-ekologiese studie van die

plantegroei van die plaas Koedoesfontein No. 746 en om-

gewing in die Rustenburgse distrik. M.Sc. thesis, P. U. vir

C. H.O., Potchefstroom.

Von Richter, W. & Jacobsen, N. H. G., 1970/71. IBP/CT

check sheet ( Mark VII) for survey of IBP areas. Unpubl.,

Pretoria.

Water Affairs, Dept, of, 1966. 1: 250 000 Rainfall map —

Sheet 2526 Rustenburg. Pretoria: Dept, of Water Affairs.

Weather Bureau, 1954. Climate of South Africa. I. Climate

statistics. W.B.19. Pretoria.: Weather Bureau.

Weather Bureau, 1957. Climate of South Africa. 4. Rainfall

maps. Pretoria: Weather Bureau.

Weather Bureau, 1960. Climate of South Africa. 6. Surface

winds. Pretoria: Weather Bureau.

Weather Bureau, 1965. Climate of South Africa. 9. Average

monthly rainfall up to the end of 1960. Pretoria: Weather

Bureau.

Werger, M. J. A., 1973a. On the use of association-analysis and

principal component analysis in interpreting a Braun-

Blanquet phytosociological table of a Dutch grassland.

Vegetatio 28: 129-144.

Werger, M. J. A., 1973b. An account of the plant communities

of Tussen-die-Riviere Game Farm, Orange Free State.

Bothalia 11: 165-176.

Werger, M. J. A., 1973c. Phytosociology of the Upper Orange

River Valley: A syntaxonomical and synecological study.

D. Sc., University of Nijmegen.

Werger, M. J. A., 1974a. On concepts and techniques applied

in the Ziirich-Montpellier Method of vegetation survey.

Bothalia 1 1 : 309-323.

Werger, M. J. A., 1974b. Applicability of Ziirich-Montpellier

methods in African and sub-tropical range lands. In R.

Tiixen (gen.ed.), Handbook of vegetation science, W. Krause

(ed.) 13, Application to grassland husbandry. The Hague:

Junk.

Westhoff, V. & Den Held, A. J., 1969. Plantengemeenschappen

in Nederland. Zutphen: Thieme.

Westhoff, V. & Van Der Maarel, E., 1973. The Braun-

Blanquet approach. In R. Tiixen (gen.ed) Handbook of

Vegetation Science, R. H. Whittaker (ed.) 5, Ordination

and classification of vegetation. The Hague: Junk.

581

Book Reviews

Trees of Southern Africa by Eve Palmer & Nor ah Pitman.

Cape Town: A. A. Balkema. 3 vol. 1972-1973. Pp. 2235. Price

R25 per volume.

This work, in three volumes, is a successor to the authors’

Trees of South Africa (1961), which attained such wide

popularity in South Africa. While the earlier book dealt with

only 176 tree species, this deals with all the just over 1000

species in Southern Africa — also, it covers a wider area taking

in South West Africa, Botswana, Lesotho and Swaziland as well.

The book is lavishly illustrated by nearly 2000 black and white

photographs, numerous colour plates and 900 line drawings

by artists Norah Pitman and Rhona Collett. In the preparation

of the book Eve Palmer, who was responsible for the text and

her husband, Geoffrey Jenkins, well-known novelist, who con-

tributed the colour photographs, travelled over 160,000 kilo-

metres collecting specimens and seeing the trees growing in

their natural habitats. The text was written in co-operation with

the Botanical Research Institute, Pretoria. This book un-

doubtedly represents the most comprehensive and detailed

work ever produced on the trees of Southern Africa.

The 298-page introduction covers such topics as prehistory,

trees, men and history, distribution, trees and animals, trees

and magic, poison trees, trees and food and what men make of

trees. This serves as a fascinating background to the pages that

follow.

The tree descriptions comprise the following: — a brief

synonymy, common names in English, Afrikaans and Bantu,

a non-technical botanical description, notes on the distribution,

ecology, ethnobotany, medicinal and economic uses, and

derivation of the scientific name. Each description is accom-

panied by illustrations of the habit, bark, leaves, flowers and

fiuit of the tree.

The illustrations vary considerably in quality. Some keys

to species are provided. These keys have been conceived by the

first author, taken from existing keys or adapted from exis-

ting keys.

A few criticisms can be levelled at the book. Although a

reference bibiography, chronologically arranged, is given (pp.

292-296) there is no tie-up in the text to these references. It

may be argued, of course, that for a book intended chiefly for

the layman, the insertion of source references would interrupt

the continuity and smooth flow of the text. One wonders, too,

why with a page width of 19 cm, 6 cm should be devoted to

a left-hand margin. This seems extravagant, though perhaps

appealing from a lay-out point of view. On p. 1541 it is stated

that Kiggelaria africana was first collected in the Cape in the

mid- 18th century. In actual fact it was probably collected in

the second half of the 17th century, because it was referred to in

the works of Sterbeeck (1682), Hermann (1687) and Plukenet

(1692). Scolopia thorncroftii is treated as a distinct species,

whereas it is clearly a synonym of S', zeyheri (see Sleumer, 1972).

On p. 213 Eve Palmer jumped the gun in using the name

Ozoroa concolor (Presl ex Sond.) De Wint.; this combination,

at the time of going to press of this review, has not been

published. The inevitable gremlin has crept in: the plate facing

p. 289 of Moringa ovalifolia bears the caption “The most valued

trunk in Southern Africa — that of the black stinkwood, Ocotea

ballata Typographical errors are few.

This book is fluently written and readable and is recom-

mended to all those interested in the trees of Southern Africa.

The price, however, will probably put this book beyond the

reach of all but libraries.

D. J. B. Killick

Ferns of the Witwatersrand by F. D. Hancock. & A. Lucas.

Johannesburg: Witwatersrand University Press. 1973. Pp.

xv +94. Price R7,00.

When preparing this work, the authors obviously gave

special attention to the needs of the undergraduate student.

In the introductory chapter life cycles and the general ecology

of ferns in the study area are discussed. This is followed by an

enumeration of the principal features of the four divisions of

Pteridophyta and a short discussion of their evolution.

An identification key with line illustrations showing

diagnostic characters enables the reader to identify the 32

species which have been recorded on the Witwatersrand. The

more serious student will be disappointed that neither generic

nor family diagnoses are provided. Each page of text consists

of a brief description and a section on derivation of the scientific

name, vernacular names, habitat, diagnostic features, distri-

bution outside the Witwatersrand and uses by Man. The line

drawings by Barbara Pike and Patsy-Lynne Edkins are generally

of excellent quality, illustrating the habit or part of the frond,

the sporangium and the spores. Unfortunately no scale is

provided. This can be very misleading, for example in the case

of Selaginella mittenii the habit sketch is larger than life and

the plant looks like S. kraussiana. The four water-colour plates

show much less detail than the line drawings, and merely add

to the cost of the book.

The authors are congratulated on this work, but it is

suggested that they bring out a cheaper, soft-covered edition

which should increase the popularity of the book. The printing

is of excellent quality, the paper is good, and the cloth binding

attractive and durable.

P. Vorster

Grundfragen und Methoden in der Pflanzensoziologie.

Ed. E. van der Maarel & R. Tuxen. The Hague: W. Junk. 1972.

Per. 14th Int. Symp. Ini. Verein. Vegetationskunde, Rinteln,

1970. Pp. x X +533 , figures, tables. Price R25.

This is certainly one of the most interesting and important

volumes in the well-known series of proceedings of the annual

international symposia of the International Association for

Plant Geography and Ecology organized by Tuxen. As always,

the discussions following the papers are as significant as the

papers themselves, and they clearly reflect the cordial, but

sometimes blunt atmosphere which is so typical of Tuxen’s

symposia. At this symposium, the possibilities of integrating

the methods, concepts and results of the mathematically

orientated and the “traditional” ecologists were explored.

The good and often long discussions following the papers

frequently show the marked contrast in thinking and approach

between these two groups of ecologists. The “traditionalists”

often make much effort to be understanding and open to the

ideas of the mathematical ecologists, and both sides repeatedly

express the need for synthesis between these two approaches.

The possibilities and power of mathematical techniques in

ecology are thoroughly discussed, as well as the meaning of the

results they produce in terms of the interpretations of the tradi-

tional ecologists. In this respect the discussions in this volume

may, in the long term, prove to be of historical interest in

illustrating the present period of attempts by ecologists to unify

their efforts in promoting the ecological understanding of

vegetation and so to further the science of ecology.

The number and quality of the contributions in the present

volume are too many to allow a complete review and only a

few topics can be mentioned here. Most papers and large parts

of the discussions are in German, although some are in English

or French. However, virtually all papers have English sum-

maries.

The first half of the book merits its title and deals with

phytosociological methods in general, whereas the second half

discusses some more specific topics. The book opens with a

paper by Westhoff discussing the place of vegetation science

in the biological sciences. Westhoff reviews briefly the Anglo-

American and some continental European opinions regarding

vegetation science. He discusses the various scientific branches

which are part of vegetation science, building on the well-

known work of Schmithusen, and concludes that vegetation

science fully deserves a place as a distinct branch of the biological

sciences. In the discussion following this paper the place of

ecosystem research is debated.

Moore discusses computer-based methods for the analysis

of phytosociological data in historical perspective, with emphasis

on their application in compiling Ziirich-Montpellier phytoso-

ciological tables. Unfortunately Moore wrongly considers that

phytosociologists of the Ziirich-Montpellier School first

“intuitively grasp" the plant communities in the field, checking

them later in their tables. Although this may frequently have

occurred it is certainly not the correct approach of the School

as such. Homogeneous stands of vegetation are recognized in

the field and each stand is sampled as an example of a plant

community yet to be recognized. This is done independently

of any preconceived ideas as to the nature and affinities of

communities. The communities are only compiled and delimited

in a table as abstractions of many similar stand samples. Moore

describes the Ziirich-Montpellier procedure of table making

as divisive. The comparison of releves with one another on

their species contents and the arrangement of the most similar

ones together is, however, an agglomerative procedure.

A most interesting paper is given by Whittaker on conver-

gences of ordination and classification. After pointing out

the differences in "scientific cultures” between the Anglo-

American and continental European plant ecology, he argues

that therefore no “simple fusion in agreement” between the

two is to be expected. He suggests, however, three possible

convergences, namely, in theory of vegetation structure, in

of quantitative classification techniques, and in the application

gradient analysis as an adjunct to classification. He emphasizes

the possibilities of using this third procedure as an aid to clarify

582

and communicate vegetational relationships in a Ziirich-

Montpellier classification. Whittaker regards complete con-

vergence between the two approaches as not even desirable

for the advance of ecological theory and understanding of

vegetation.

The possibilities of more objective phytosociological

methods are considered by Doing, while several authors, e.g.

Orloci, Ivimey-Cook, Fresco, Van Emden, Romane, discuss

quantitative analysis strategies, such as association-analysis

and various types of information and factor analyses. Romane

presents an account of an application of factor analysis of

correspondences, a technique which allows a direct estimation

of the relationships between species occurrences and habitat

factors.

Daget, Godron and Guillerm present their technique of

ecological profiles, being frequency distributions of species

against various classes of habitat factors. From these profiles and

a calculation of the indicator value of each species for various

ecological factors, ecological groups of species can be compiled

which indicate distinctive environmental conditions, it is

unfortunate that there is apparently some editorial mistake in

this paper, which makes it difficult for the reader to follow.

Tiixen's “Critical observations on the interpretation of

phytosociological tables” is also a useful paper, in which he

compares the differences in eleven releves made of the same

quadrat on the same day by eleven phytosociologists, who were

each allowed only fifteen minutes. Tiixen demonstrates how

careful one must be in sampling vegetation and in deliminting

types, and how difficult it is to interpret data collected by others

in vegetation unknown to the interpreter. Fie also warns against

the false idea of exactitude, which is sometimes given by some

analytical techniques. The discussion following this paper is

long and good, touching on themes such as sampling and the

necessity for mutual understanding between the person who

carries out the field work and the mathematically orientated

office ecologist. The exposition by Pignatti, on how he would

detect that there was something wrong with such data, if they

were presented to him, is unconvincing.

Van der Maarel presents his preliminary findings in an

application of principal component ordination of plant com-

munities on the basis of their plant genus, family and order

relationships.

Further articles deal with homotoneity and new computer

programmes for the processing of phytosociological data

(Stockinger & Holzner; Spatz). In a paper by Wagner the

procedure of omitting unrepresentative stands from a phytoso-

ciological table is discussed.

The second half of the volume deals with more specific

topics. There are several papers on phenology and its inter-

pretation in a community context, of which in particular the

ones by Dierschke and by Hartmann are followed by stimulating

discussions. Then follow a number of papers on the phytosocio-

logical classification of forest communities, saline communities

and bog communities, as well as on specific aspects of some

weed communities, waterplant communities and Icelandic

grasslands. Altogether this second half of the volume is as

varied as the first, it is not reviewed here in detail, because it is

mainly, although not entirely, of local interest.

It is fortunate that the present volume has been published

relatively soon after the symposium, since this increases its

value considerably. This publication, with its 38 papers and

long and important discussions should not be absent from any

library specializing in plant ecology. An invaluable book,

indeed, and for a reasonable price.

M. J. A. Werger

Die Vegetation von Afrika by R. Knapp. Stuttgart: Gustav

Fischer Verlag. 1973. Pp. xliv+626, 823 figures, 825 tables &

species lists. Price ca. R60.

This book, whose full title reads “The vegetation of

Africa with references to environment, development, economy,

agriculture and forestry geography”, is published as part III

in the series “Vegetationsmonographien der einzelnen Gros-

sraume” under the general editorship of H. Walter, and includes

a short introduction and a detailed list of contents in English.

It attempts to give an ecologically relevant account of African

vegetation based on nearly twenty years of literature studies

by the author and ten years of study trips to all parts of Africa.

The book consists of eight chapters: first, a general one dealing

with climate, phytogeography, anthropogenic influences, fire

and methods used in the book, followed by seven chapters

consecutively describing the tropical rain forest zone, the

savanna and dry deciduous woody vegetation zone, the montane

vegetation, the afro-alpine vegetation, the vegetation of the

deserts and subdeserts, the vegetation of the evergreen sclerophyl-

lous zones and adjacent winter rainfall regions, and the

vegetation of the Macaronesian areas (Cape Verde and Canary

Islands and Madeira). The description of the vegetation of the

islands in the Indian Ocean (Madagascar, Mascarenes,

Seychelles and Socotra) is included in appropriate sections of

the various chapters. The descriptions and interpretations

offered in the book are to a very large extent based on data

collected by Knapp, and although more than 1 500 literature

references are listed at the end of the book only a few of them

have been used and intergrated in the descriptions. Con-

sequently, a picture of the vegetation of Africa is presented,

that is far more subjective than was necessary. This is particu-

larly apparent in the sections dealing with the more complex

vegetation types. It is also apparent in, for example, the section

on phytogeography, where, without any discussion of literature,

the Osambara-Zululand Domain as part of the Guineo-

Congolian Region, and the Afro-alpine Region in southern

Africa, are not recognized. All the high mountain vegetation in

southern Africa is considered to be montane, but this montane

zone also includes, according to Knapp, virtually the entire

grassland area of the Highveld and the Bankenveld. Thus it

gets a rather wide interpretation.

The description of each zone starts with notes on its

distribution and possible subdivisions, its specific ecological

features such as climate, soils, human and other influences, its

economic uses and possibilities, etc., and then proceeds to a

description of its plant communities. Each description is fol-

lowed by a list of species indicating which are dominant and

which are characteristic according to Knapp. This presentation

of species lists has, however, two serious disadvantages:

(1) The lists of species, indicating which are characteristic

(absolute character species according to Knapp, p.vii and

p.32), suggest that the “plant communities” described are

comparable to associations or other syntaxa as recognized by

Zurich-Montpellier methods, which they are not at all. Knapp's

“plant communities” are rather formations or subformations.

(2) The species lists suggest a far higher degree of accuracy

than actually given. This is obvious when reading sections on

vegetation types with which one is familiar.

Although there are still many minor points in the book

which are not strictly true, or with which one might disagree,

it is, apart from the previous objections, an admirable piece

o" work: a single author succeeding in giving so many useful

facts on such a large and diverse area! The literature list is

probably the largest one existing on African vegetation, and it

is just a pity that it does not include a number of the more

important publications of the last four or five years, particularly

those from South Africa and Angola. The book has a good index

and is well and extremely richly illustrated with many maps,

profiles, diagrams and photographs. Its price is therefore not

excessively high and, although it will be prohibitive for many

a scientist's private library, institutes concerned with the

vegetation of Africa, or with vegetation formations and their

geographical and ecological characteristics, should have it in

their libraries.

M. J. A. Werger

Handbook of Vegetation Science. Chief ed. R. Tiixen.

Ordination and Classification of Communities. Ed. R. H.

Whittaker. The Hague: W. Junk. VoL 5. 1973. Pp. x + 738, 91

figures , 40 tables. Price R40.

Handbook of Vegetation Science. Chief ed. R. Tiixen.

Vegetation Dynamics. Ed. R. Knapp. The Hague: W. Junk.

Vol. 8. 1974. Pp. x + 366, 37 figures, 8 tables. Price R20.

With the publication of these first volumes of the Handbook

of Vegetation Science, a landmark in vegetation science has been

reached. The Handbook, of which R. Tiixen is editor in chief,

is planned to provide in 18 volumes an up-to-date and com-

prehensive summary of concepts, methods and knowledge

acquired in vegetation science. The Handbook is a most

ambitiously planned project that attempts to integrate into one

composite picture the various approaches to the study of

vegetation based on the many philosophies concerning the

nature of vegetation, the insights gained from all the different

theoretical and applied branches of this science, as well as the

historical lines towards the stage of development where this

science stands today.

Volume 5, under the editorship of R. H. Whittaker, is the

first volume to appear in print, and it certainly comes close to

its planned goal of reviewing and integrating all major concept

and methods applied in the analysis of vegetation and the

synthesis of these data. In twenty chapters the book deals with a

wide range of ordination and classification approaches without

much overlap among the chapters. Each chapter has a summary

in English and German and a bibliography of cited works.

From these bibliographies il seems that the chapters, except the

last one, have been completed before or in 1971. In the intro-

duction, Whittaker, who is not only a competent editor of the

book but also author or co-author of eight of its chapters,

distinguishes between community classification (syntaxonomy)

583

versus phytosociology, and between ordination versus gradient

analysis. He states that the book is an account of procedures,

rather than of approaches to understanding of which those

procedures are part; of ordination and classification, rather

than of gradient analysis and phytosociology. He also emp-

hasizes again that the two major approaches, ordination and

classification, are complementary.

The section on direct gradient analysis consists of four

chapters. In the first two chapters Whittaker discusses the

techniques and results of direct gradient analysis. Attention is

paid to sampling and the various ways of analysing vegetation

relationships from the samples. He formulates again his con-

clusions about the spatial structure of vegetation, as he has

explained already in previous publications, particularly in

Gesellschaftsmorphologie (Ber. Int. Symp. Rinteln 1966. The

Hague: Junk, 1970). In Chapter 4 Whittaker and Woodwell

deal with retrogression and the way to measure this. In the final

chapter of this part of direct gradient analysis, Sobolev and

Utekhin give an instructive account on Russian approaches to

community systematization, particularly of Ramensky's ideas

and procedures. They clearly point out the differences and

similarities with American concepts and procedures.

The second section on indirect gradient analysis, consisting

of six chapters, starts with a contribution by Goodall on

measures of similarity and correlation in their own right,

irrespective of the particular ordination and classification

techniques to which they may be put. Goodall's contribution is

a clearly written, useful discussion giving guidelines for the

choice of the indices to be used. In the next chapter McIntosh

discusses and illustrates various matrix and plexus techniques,

most of which are graphical representations of data and results.

Cottam, Goff and Whittaker discuss the Wisconsin (Bray and

Curtis) comparative ordination technique, pointing out its three

important advantages over other ordination techniques: (i) it

is a most versatile and least vulnerable technique; (ii) it is

equally effective for direct and indirect ordination, and (iii) it

has great value in research as a framework for investigating

and understanding vegetational relationships. In Chapter 9

Dagnelie discusses factor analysis and its application in

vegetation studies. Dagnelie writes in French, which is rather

unfortunate: he has published an account of this technique in

French before, and for a wider understanding of the technique

it would have been better if this chapter had been written in

English. In the next chapter, however, Orloci, discussing and

comparing a large number of ordination procedures, again

briefly describes factor analysis. He points out its disadvantages

and says that these “may be the reasons why factor analysis,

despite its effective use by Dagnelie, has not aroused wider

interest among phytosociologists" (p. 280). Orloci concludes

that Kruskal's method of multidimensional scaling has a good

potential in phytosociological ordinations and advises its further

use. The final chapter in this part on indirect gradient analysis

is an evaluation of ordination techniques by Whittaker and

Gauch. It gives, as the authors say, "a classification of ordination

techniques and an ordination of some of these by relative

usefulness", as concluded from tests with simulated oeno-

cline data. The various techniques need to be evaluated

in terms of freedom from distortion of sample positions, ranges

of sample variation that can be handled, clarity of data treatment

and results, computational expenses, and general effectiveness

for research. The authors' conclusion, which has also been

published in Ecology 53 (1972), is that the method of Bray and

Curtis is the best and that principal component analysis is the

worst of the compared methods. Principal component analysis

is most useful for narrow ranges of community variations.

(Compare also Beals, 1973, J. Ecol. 61).

The third section of the book, consisting of nine chapters

and 400 pages, deals with classification. An introductory chapter

by Whittaker gives a perspective that facilitates the under-

standing of the following chapters (compare Whittaker, 1962,

Bot. Rev. 23). Whittaker also explains why a classification of

vegetation is an artificial classification. Beard discusses the

physiognomic approach. His discussion, particularly that on

floristic and physiognomic units, tends to be somewhat dog-

matic, and apparently Beard favours to confuse the issue as far

as the concept “association" is concerned. Then follow chapters

by Whittaker on dominance types, especially as used by

Clements and by Frey of the Finnish School and forest site-types

as used in particular by Cajander. Barkman contributes a

comprehensive and critical review of synusial approaches to

classification, giving clear examples. From this it becomes

apparent that virtually all synusial work is done in the northern

temperate and boreal regions, mostly on cryptogamous epiphytic

and on aquatic vegetation. Very few data have been collected

on the rich vascular epiphytic vegetation of the tropics. There

are also review chapters on Russian approaches to classification

of vegetation by Aleksandrova, on Scandinavian (mainly

Uppsala) and Baltic approaches by Trass and Maimer, and on

numerical classification by Goodall. Goodall evaluates the

various numerical classificatory approaches and points out how

ordination and numerical classification can be efficiently

combined. The final and longest chapter in the book by Westhoff

and Van der Maarel, discusses in detail the concepts and

techniques of the Braun-Blanqet approach. It also reviews

briefly the extensive use of this approach in various parts of the

world. Several references to literature mentioned in this section

are not listed in the 19 pages of references at the end of the

chapter, however, and should be traced in Excerpta Botanica

Sect. B, Sociologica. An outline of the use of numerical tech-

niques as complementary to the Braun-Blanquet approach in

order to save effort and facilitate and improve interpretation

of the results, is added. This chapter is the most comprehensive

account of the Braun-Blanquet approach existing in English,

and is obviously up-to-date.

It may be concluded that the editor (and the authors)

have succeeded in giving in one volume a comprehensive and

integrated picture of the philosophies and methods of modern

vegetation science, and it is to be expected that this will facilitate

the further development of the science, because it makes so

much information that was formerly difficult to get at, readily

available to a wide group of scientists.

Volume 8, edited by R. Knapp, gives a less integrated

picture of the main points of knowledge available on vegetation

dynamics and the methods used to acquire this knowledge

than Whittaker succeeded in integrating the contributions to his

book. The contributions to Volume 8 deal with a wide range of

topics and although all are somehow concerned with the theme

of the book, those contributions that deal with fairly closely

related topics overlap one another more strongly than the

chapters of Volume 5. Nevertheless, Volume 8 is a most useful

source of information on vegetation dynamics, with at the end

of the book 64 pages of literature references on the subject.

The book is subdivided into 7 sections, together consisting

of 27 short chapters by 17 authors. Knapp himself has written

five chapters. Apart from five chapters in German by Tuxen,

Reinhold and Aichinger, with English summaries, none of the

other chapters, which are all in English, have summaries.

Apparently Major contributed one manuscript which was cut

by the editor and published as five different chapters in various

sections of the book.

Major’s contribution on kinds of change in vegetation and

chronofunctions, and Rabotnov's discussion on differences

between fluctuations and successions, form the first section

of the book. Rabotnov considers fluctuations to be characterized

by (i) differently orientated changes in phytocoenoses from year

to year, (ii) reversibility of the changes, and (iii) absence of

invasion of new species.

The second section is devoted to methods of syndynamic

analysis. Tiixen discusses the value of macrofossils, pollen,

spores and subfossil soil profiles for drawing conclusions on the

history of vegetation types. Knapp outlines the value of studies

of newly invaded land areas, of relics of historical records on

maps, photographs and other documentations, of permanent

quadrats and of plant migrations for drawing conclusions on the

syndynamics of vegetation types. Also, in this section, Aleksan-

drova and Karamysheva point out some specific factors causing

vegetation changes in the Eurasian tundras, steppes and

semideserts. In the tundra severe grazing by reindeer can cause

problems, and Aleksandrova reports that it can take 20 to 50

years before the lichen cover is regenerated after complete

destruction. Very interesting are the short chapters by Stearns

on the use of documents of the American General Land Office

that has collected records since 1785, and by Reinhold on the

value of historical forestry records in Central Europe and

France for syndynamic studies in formerly forested areas.

The section concludes with a chapter by Knapp briefly discussing

cyclic (regeneration) succession and its relations to linear

succession and to fluctuations, as well as the value of ecosystem

research, simulation and modelling in the study of vegetation

dynamics. Knapp says that only natural events of catastrophic

dimensions, or actions of man, should cause destruction of

stable communities ( Scliluss - and Datiergese/lsclwften). To this

one should also add the action of indigenous large herbivores,

however.

The third section consits of two chapters, both by Knapp,

summarizing some of the main points of knowledge about

genetic and cytological conditions of plant populations and the

impact of mutual influences between plants, such as competition,

repression, allelopathy and promoting effects, as causes for

vegetation changes.

The fourth section deals with classification of successions.

Dansereau advocates the importance of Huguet del Vilar's

scheme of physiological regimes; Whittaker discusses the

climax concept and points out that mainly because he favours

the concept of the population structure of vegetation, he

prefers a climax pattern hypothesis; Major deals with duration

in successional series; Aleksandrova and Knapp explain many

technical terms from, respectively, Russian and Western

Hemisphere literature dealing with vegetation dynamics;

and, finally, Aichinger illustrates briefly his concept of vegetation

development type.

584

The next section, on productivity and chemical changes

in succession stages, contains a brief introduction by Lieth

to the developing literature on this subject (with a printing

error in a subtitle, p. 185), three papers by Major with many data

on biomass, nitrogen and ash elements accumulation and

pH changes in successions, and a chapter by Beard discussing

vegetational changes on ageing landforms in the tropics and

subtropics. Beard considers such vegetational changes to be

mostly retrogressive: erosion (peneplanation) of afforested land

leads to changes in the soil (desiccation) and this brings about

changes in the vegetation, mostly towards savannas or grass-

lands. Beard also adds another two climax terms to the already

enormous climax vocabulary.

Then follows a section with interesting examples of fluc-

tuations, in which Korchagin and Karpov, Coupland and

Bykov, write on fluctuations in the coniferous Taiga, North-

American grasslands and Turanian semideserts. In the North-

American grasslands the effects of overgrazing are accentuated

by severe drought. Overgrazing of mixed (mid and short)

grasslands leads, according to Coupland, to short grasslands

and in the drier areas to encroachment of desert shrubs, as

found also in hot marginal grassland areas of South Africa.

The final section in the book contains two papers by

Tiixen on synchronology of Central European vegetation.

Tiixen reviews the possibilities and results of palaeosociological

studies of fossils of the Carboniferous, the Tertiary and the

Quarternary. He also reviews the existing palaeosociological

literature on the various Central European syntaxa.

This book edited by Knapp contains a wealth of data and

ideas on vegetation dynamics. It is indeed a pity that Tiixen

has used the same foreword in both volumes, thereby crediting

Whittaker in volume 8 with the editorial work done by Knapp.

Interested vegetation scientists and libraries will undoubtedly

want to purchase these volumes of the Handbook of Vegetation

Science. The volumes are certainly not expensive by to-day’s

standards.

M. J. Werger

585

GUIDE FOR AUTHORS

GENERAL

Bothalia is a medium for the publication of botanical

papers dealing with the flora and vegetation of Southern

Africa. Papers submitted for publication in Bothalia should

conform to the general style and layout of recent issues of

the journal (from Vol. 11 onwards) and may be written in

either English or Afrikaans.

TEXT

Manuscripts should be typed, double-spaced on one side

of uniformly-sized A4 paper having at least a margin of 3 cm

all round. Latin names of plants should be underlined to

indicate italics. All other marking of the copy should be left

to the editor. Metric units are to be used throughout. Manus-

cripts should be submitted in duplicate to the Editor, Bothalia,

Private Bag X10I, Pretoria.

ABSTRACT

A short abstract of 100-200 words preferably in both

English and Afrikaans should be provided. In the abstract

the names of new species and new combinations should not

be underlined.

FIGURES

Black and white drawings, including graphs, should be in.

jet-black Indian ink preferably on bristol board. Lines should

be bold enough to stand reduction. Indicate the desired lettering

lightly in pencil: the printer will insert the final lettering. If

authors prefer to do their own lettering, then use some printing

device such as stencilling, letraset, etc. It is recommended

that drawings should be twice the size of the final reduction.

Photographs submitted should be of good quality, glossy,

sharp and of moderate, but not excessive contrast. Photograph

mosaics should be composed by the authors themselves: the

component photographs should be mounted neatly on a

white card base leaving a narrow gap between each print;

number the prints using some printing device.

Figures should be planned to fit, after reduction, into

a width of 8 cm, 11 cm or 17 cm with a maximum vertical

length of 24 cm.

The number of each figure and the author’s name should

be written on the back of the figure using a soft pencil.

Captions for figures should be collected together and typed

on a separate page headed Captions for Figures. A copy of

each caption should be attached to the base of each figure.

Do not underline plant names in captions — only collectors'

names and numbers.

Authors should indicate in pencil in the text where they

would like their illustrations to appear.

TABLES

Tables should be set out on separate sheets and numbered

in Arabic numerals.

CITATION OF SPECIMENS

In citing specimens the grid reference system should be

used (Technical Note: Gen. 4, 4c). Provinces/countries should

be cited in the following order: S.W. Africa, Botswana,

Transvaal, Orange Free State, Swaziland, Natal, Lesotho and

the Cape. Grid references should be cited in numerical sequence.

Locality records for specimens should preferably be given to

within a quarter-degree square. Records from the same one-

degree square are given in alphabetical order i.e. (-AC)

precedes (-AD), etc. Records from the same quarter-degree

square are arranged alphabetically according to the collectors’

names; the quarter degree references must be repeated for

each specimen cited. The following example will explain the

procedure :

Natal. — 2731 (Louwsburg): 16 km E. of Nongoma (-DD),

Pelser 354; near Dwarsrand, Van der Merwe 4789. 2829

(Harrismith): near Groothoek (-AB), Smith 234; Koffifontein

(-AB), Taylor 720; Cathedral Peak Forest Station (-CC),

Marriott 74; Wilgerfontein, Roux 426. Grid ref. unknown:

Sterkstroom, Strydom 12.

Records from outside Southern Africa should be cited

from north to south i.e. preceding those from Southern Africa.

The abbreviation “distr.” should be added to all district

names, e.g.

Kenya. — Nairobi distr.: Nairobi plains beyond race course,

Napier 485.

GIDS VIR SKRYWERS

ALGEMEEN

Bothalia is 'n medium vir die publikasie van plantkundige

artikels wat handel oor die flora van Suidelike Afrika. Artikels

wat voorgele word vir publikasie in Bothalia behoort ooreen

te stem met die algemene styl en rangskikking van onlangse

uitgawes van die tydskrif (vanaf Vol. It). Dit mag in Engels

of in Afrikaans geskryf word.

TEKS

Manuskripte rnoet getik wees in dubbelspasiering slegs

op een kant van ewegroot A4 papier, met reg random ’n

rand van minstens 3 cm breed. Latynse name van plante moet

onderstreep word om aan te dui dat dit kursief gedruk moet

word. Alle ander merke moet aan die redakteur oorgelaat

word. Metrieke eenhede moet deurgaans gebruik word.

Manuskripte moet in tweevoud ingedien word by die Redak-

teur, Bothalia, Privaatsak XI 01, Pretoria.

UITTREKSEL

’n Kort uittreksel van 100-200 woorde moet voorsien

word, verkieslik beide in Engels en Afrikaans. In die uittreksel

moet die name van nuwe soorte en nuwe kombinasies nie

onderstreep word nie.

AFBEELDINGS

Wit en swart tekeninge, insluitende grafieke, moet met

pikswart Indiese ink geteken word, verkieslik op “bristol

board”. Lyne moet dik genoeg wees om verklein te kan word.

Dui die verlangde byskrifte liggies in potlood aan: die drukker

sal die uiteindelike byskrifte invoeg. Indien skrywers verkies

om hulle eie byskrifte te maak, gebruik dan een of ander

hulpmiddels soos letraset of 'n sjabloon. Dit is wenslik dat

tekeninge tweemaal so groot as die uiteindelike verkleining

sal wees.

Fotos wat ingedien word, moet van hoe kwaliteit wees-

glansend, skerp en van matige maar nie oordrewe kontras.

Fotomosaieke moet deur die skrywer self saamgestel word:

die afsonderlike fotos moet netjies monteer word op 'n stuk

wit karton met ’n smal strokie tussen die fotos; nommer die

fotos met behulp van een of ander druk-hulpmiddel.

Afbeeldings moet so beplan word dat hulle nav erkleining

sal pas in ’n breedte van 8 cm, 1 1 cm of 17 cm met ’n maksimum

vertikale lengte van 24 cm.

Die nommer van elke afbeelding sowel as die skrywer

se naam moet op die rugkant van die afbeelding geskryf word

met ’n sagte potlood.

Onderskrifte vir afbeeldings moet bymekaar getik word

op ’n afsonderlike bladsy met die opskrif Onderskrifte vir

Afbeeldings. 'n Afskrif van elke onderskrif moet aan die

onderkant van elke afbeelding vasgeheg word. Moenie plant-

name in onderskrifte onderstreep nie, slegs versamelaarsname

en -nommers.

Skrywers moet met potlood in die teks aandui waar hulle

graag hulle afbeeldings wil he.

TABELLE

Tabelle moet op afsonderlike velle papier kom en ge-

nommer word met Arabiese nommers.

SITERING VAN EKSEMPLARE

Wanneer eksemplare siteer word, moet die ruitverwysing-

stelsel gebruik word (Tegniese Nota: Gen. 4, 4c). Provinsies/

Iande moet in die volgende volgorde siteer word: Suidwes-

Afrika. Botswana, Transvaal, Oranje-Vrystaat, Swaziland,

Natal, Lesotho en die Kaapprovinsie. Ruitverwysings moet

in numeriese volgorde siteer word. Lokaliteitsrekords vir

eksemplare moet verkieslik tot binne kwartgraadvierkante

gegee word. Rekords uit dieselfde eengraadvierkant word in

alfabetiese volgorde aangebied, nl (-AC) kom voor (-AD),

ens. Rekords uit dieselfde kwartgraadvierkant word alfabeties

gerangskik volgens die versamelaars se name, en die kwart-

graadverwysings moet herhaal word vir elke eksemplaar wat

siteer word. Die volgende voorbeeld sal die metode verduidelik:

Natal. — 2731 (Louwsburg): 16 km O. van Nongoma

(-DD), Pelser 354; naby Dwarsrand, Van der Merwe 4789,

2829 (Harrismith): naby Groothoek (-AB), Smith 234; Koffie-

fontein (-AB), Taylor 720; Cathedral Peak Bosboustasie

(-CC), Marriott 74; Wilgerfontein, Roux 426 ; Ruitverwysing

onbekend: Sterkstroom, Strydom 12.

Rekords van buite Suidelike Afrika moet siteer word

van noord na suid, d.w.s. dit gaan die van Suidelike Afrika

v oraf. Die afkorting "distr.” behoort by alle distriksname

gevoeg te word, bv.

Kenya. — Nairobi-distr. : Nairobivlakte anderkant die ren-

baan, Napier 845.

586

REFERENCES

References in the text should be cited as follows: "Jones

(1955) stated. . or . . (Smith, 1956)" when giving

a reference simply as authority for a statement. The list of

references at the end of the article should be arranged alphabe-

tically and the literature abbreviations used should conform

to the list of Literature Abreviations (Technical Note: Tax.

6/1) issued by the Botanical Research Institute, thus:

HUTCE1INSON, J., 1946. A botanist in Southern Africa.

London: Gawthorn.

MORRIS, J. W., 1969. An ordination of the vegetation of

Ntshongweni, Natal. Bothalia 10: 89-120.

If. as in many taxonomic papers, periodicals or books

are mentioned in the text, usually in the species synopsis, they

should be cited as in the following examples: Gilg & Ben. in

Bot. Jahrb. 53: 240 (1915) and Burtt Davy, FI. Transv. 1:

122 (1926).

REPRINTS

Authors receive 75 reprints gratis. If there is more than

one author, this number will have to be shared between or

among them.

VERWYSINGS

Verwysings in die teks moet as volg siteer word: “Jones

(1955) beweer . . .” of . . (Smith, 1956)” wanneer ’n

verwysing slegs as outoriteit vir ’n stelling gegee word. Die

verwysingslys aan die einde van die artikel moet alfabeties

gerangskik wees en die literatuurafkortings wat gebruik word,

moet in ooreenstemming wees met die lys van Literatuur-

afkortings (Tegniese Nota: Tax. 6/1) wat uitgegee is deur die

Navorsingsinstituut vir Plantkunde, as volg:

HUTCHINSON, J., 1946. A botanist in Southern Africa.

London: Gawthorn.

MORRIS, J. W., 1969. An ordination of the vegetation of

Ntshongweni, Natal. Bothalia 10: 89-120.

Wanneer, soos in baie taksonomiese artikels die geval

is, tydskrifte of boeke in die teks genoem word, gewoonlik

in die soortsinopsis, behoort hulle siteer te word soos in die

volgende voorbeelde: Gilg & Ben. in Bot. Jahrb. 53: 240

(1915) en Burtt Davy, FI. Transv. 1: 122 (1926).

HERDRUKKE

Skrywers ontvang 75 herdrukke gratis. Wanneer daar

meer as een skrywer is, sal hierdie aantal tussen hulle verdeel

moet word.