BOTHALIA

sag *

PUBLICATIONS OF THE BOTANICAL RESEARCH INSTITUTE

Obtainable from the Division of Agricultural Information, Department of Agricultural Technical Services, Private Bag X144, Pretoria,

Republic of South Africa

FLORA OF SOUTHERN AFRICA

A taxonomic treatment of the flora of the Republic of South Africa, Lesotho, Swaziland and South-West Africa.

Initially in volumes, but subsequently in fascicles.

Now available:

Vol. 1 (1966). Price Rl,75. Overseas: R2,20. Post free.

Vol. 13 (1970). Price RIO. Overseas: R12. Post free.

Vol. 16, 1 (1975). Price R13,50. Overseas: R16,75. Post free.

Vol. 26 (1963). Price R4,60. Overseas: R5,75. Post free.

THE FLOWERING PLANTS OF AFRICA

This publication is issued as an illustrated serial, much on the same lines as Curtis’s Botanical Magazine, and for

imitating which no apology need be tendered.

The desire and object of the promoters of the publication will be achieved if it stimulates further interest in the study

and cultivation of our indigenous plants.

The illustrations are prepared mainly by the artists at the Botanical Research Institute, and the Editor is pleased to

receive living plants of general interest or of economic value for illustration.

Each part contains 10 plates and costs R1 ,50 per part. Two, three or four parts may be published annually, depending

on the availability of illustrations. A volume consists of four parts. From Volume 27, the price per volume is: cloth

binding, RIO, morocco binding, R14.

BOTHALIA

Bothalia is a medium for the publication of botanical papers dealing with the flora and vegetation of Southern Africa

One or two parts of the journal are published annually.

The following parts are available:

Vol. 3 Part 1 Out or print

2 1937 75c

3 1938 75c

4 1939 75c

Vol. 7 Part 1 1958 R2

2 1960 R3

3 1961 R3

4 1962 R3

Vol. 4 Part 1 1941 75c

2 1942 75c

3 1948 75c

4 1948 75c

Vol. 5 1950 R3

Vol. 6 Part 1 1951 Rl,50

2 1954 R2,50

3 1956 R2

4 1957 R2

Vol. 8 Part 1 1962 R3

2 1964 R3

3 1965 R3

4 1965 R3

Supplement

Vol. 9 Part 1 1966 R3

2 1967 R3

3 & 4 1969 R6

Vol. 10 Part 1 1969 R3

2 1971 R3

3 1971 R3

4 1972 R3

Vol. 11 No. 1 & 2 1973 R6

3 1974 R3

4 1975 R3

Vol. 12 No. 1 1976 R5

MEMOIRS OF THE BOTANICAL SURVEY OF SOUTH AFRICA

The memoirs are individual treatises usually of an ecological nature, but sometimes taxonomic or concerned with

economic botany. Forty-one numbers have been published, some of which are out of print.

Republic of

South Africa

Republiek van

Suid-Afrika

BOTHALIA

Volume 12, No. 2

Edited by/Onder redaksie van

D. J. B. Killick

Editorial Committee/Redaksiekomitee: B. de Winter, D. Edwards, D. J. B. Killick and/en O. A. Leistner

Botanical Research Institute

Navorsingsinstituut vir Plantkunde

m. VAN LANDBOU-TEBNIESE DIENSTE

BkLIOTEEK

NAV. INSTITUUT VIR PLAN TK WIDE

H 10 1977

PRETORIA

BOTANICAL RESEARCH INSTITUTE

DB1!. OF ASMWIIURAI TECHWCAl SERVICES

Department of Agricultural Technical Services

Departement van Landbou-tegniese Dienste

South Africa/Suid-Afrika

1977

Digitized by the Internet Archive

in 2016

https://archive.org/details/bothaliavolume1212unse_0

CONTENTS-INHOUD

Volume 12, No. 2

Page

Bladsy

1. Hans Justus Thode (1859-1932), pioneer plant collector in the Natal Drakensberg. D. J. B.

Killick 169

2. The South African species of Teucrium (Lamiaceae) L. E. Codd 177

3. A note on the Stachys aethiopica Complex. L. E. Codd 181

4. New taxa and a new combination in the genus Cotyledon. H. R. Tolken 191

5. The identity of Erica flavisepala. E. G. H. Oliver 195

6. The identity of Eriosema nanum. C. H. Stirton 199

7. Morphological studies of the Ochnaceae. P. C. V. du Ton 205

8. Leaf anatomy of the South African Danthonieae (Poaceae). I. The genus Dregeochloa. R. P.

Ellis 209

9. Cytogenetic studies in the Eragrostis curvula Complex. T. B. Vorster and H. Liebenberg 215

10. A note on the flowers of Halleria lucida. C. H. Stirton 223

11. The pollination of Canavalia virosa. C. H. Stirton 225

12. Broad-spectrum pollination of Plectranthus neochilus. C. H. Stirton 229

13. Freshwater algae of Southern Africa. II. Triplastrum stipulosum from the Transvaal. M. Isabella

Claassen 231

14. Freshwater algae of Southern Africa. IV. Some Micrasteriae from Rhodesia, including a new

species. M. Isabella Claassen 239

15. Asexual nuclear division in Neocosmospora. K. T. Van Warmelo 247

16. Notes on African plants:

Araceae. A. A. Obermeyer 251

Asclepiadaceae. R. A. Dyer 253

Cyperaceae. P. Vorster 257

Fabaceae. J. H. Ross 257

Frullaniaceae. P. Vorster 257

Poaceae. P. C. V. Du Toit 258

Ranunculaceae. D. J. B. Killick 258

Rutaceae. J. H. Ross 258

Selaginellaceae. P. Vorster 259

Thelypteridaceae. P. Vorster 260

17. World climatic patterns in grassland and savanna and their relation to growing seasons. R. Kirk

Steinhorst and J. W. Morris 261

18. Automatic classification of the highveld grassland of Lichtenburg, south-western Transvaal.

J. W. Morris 267

19. Cape Hangklip area. 1. The application of association analysis, homogeneity functions and Braun-

Blanquet techniques in the description of South Western Cape vegetation. C. Boucher 293

20. A preliminary account of aerial plant biomass in fynbos communities of the mediterranean type

climate zone of the Cape Province. F. J. Kruger 301

21. Silene dewinteri, a new species of the Caryophyllaceae from the south-western Cape. G. Bocquet 309

22. The taxonomic status of the genus Rubidgea. Eva Kovacs-Endrody 313

Book reviews 319

Bothalia : 12,2: 169-175 (1977)

Hans Justus Thode (1859-1932), pioneer plant collector in the Natal

Drakensberg

D. J. B. KILLICK*

ABSTRACT

An account is given of the life of Hans Justus Thode (1859-1932) from the time of his arrival in Cape

Town in 1885 or 1886 until his death in Durban in 1932. Thode was the pioneer plant collector of the Natal

Drakensberg, but also collected wider afield in all four provinces of the Republic. His contributions to South

African botany are assessed.

RESUME

HANS JUSTUS THODE (1859-1932), PREMIER BOTANISTE RECOLTEUR DANS LES

DRAKENSBERG DU NATAL

La vie de Hans Justus Thode (1859-1932) est relatee, depuis le moment de son arrivee au Cap en

1885 (ou 1886) jusqu'a sa mort a Durban en 1932. Thode fut le premier a recolter des plantes dans les

Drakensberg du Natal, mais ses activites se sont egalement exercees en d'autres endroits dans les

quatre provinces de la Republique. Un inventaire est dresse de ses contributions a la botanique sud-

africaine.

INTRODUCTION

As a student of the flora of the Natal Drakens-

berg, it was perhaps inevitable that the author should

become interested in the life of Hans Justus Thode

(Fig. 1), pioneer plant collector of the Drakensberg

and the first to describe its vegetation.

The author was fortunate in obtaining a number of

Thode’s personal papers, some from Miss M. D.

Gunn, former librarian of the Botanical Research

Institute, and some from Mr R. G. Strey, Curator

of the Natal Herbarium. Otherwise, research on

Thode would have been extremely difficult because,

from all accounts, he was a retiring and uncom-

municative person, who rarely spoke of himself.

Fig. 1. — Hans Justus Thode. Portrait taken c. 1920.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

BIRTHPLACE

Thode was born in 1859 probably in Germany.

His death notice (Estate No. 18272, Master's Office,

Pietermaritzburg), gives his nationality as Swiss, but

all evidence points to a German origin. All the people

contacted, who knew Thode, accepted him as German

without question. Amongst his manuscripts is a

poem entitled "All Red" obviously written during

or about the time of World War I. The poem is

strongly anti-British and would hardly have been

written by a Swiss, even a German Swiss. In an

article entitled “Journeys in the South African

Mountains”, Thode refers to the “Riesengebirge at

home”. The Riesengebirge or Giant Mountains lie

between Dresden and Breslau (Wroclaw), in what

was formerly Silesia in Germany, but is now in

Poland.

According to Mr H. Struck of Draycott, Natal,

during World War I Thode gave the authorities to

understand that he was Swiss in order to escape

possible internment. This may account for the Swiss

nationality given on his death notice.

Thode had four sisters and one brother. The one

sister Alwine Margarethe Sophie married Carl

Alexander Hammer, a storekeeper who, according

to Mr H. Zunckel of Bergville, lived at Oliviershoek.

In the death notice of this sister, the birthplace is

given as Germany. Another sister, Sylvia, married

a Mr Mummbrauer of Noodsberg. The two remaining

sisters lived in Germany, the one Mrs Kate Putz in

Munich and the other Mrs Suzanna Saphir in

Leipzig. The brother Felix emigrated to Australia.

Mrs Kate Putz’s daughter-in-law, Mrs Emmy Putz,

claims that she remembers seeing a baptism certi-

ficate which indicated that Kate Putz (nee Thode)

was born in Manchester, England (letter from

Dr A. Schreiber, Munich, to Miss M. D. Gunn, 19th

August 1966).

To make tracing of Thode's birthplace even more

difficult there are suggestions that Thode changed

his name on arrival in South Africa. According to

Mrs H. Mummbrauer of San Francisco, U.S.A.,

he was originally called Freylinghausen, while

according to Mr Erich Mummbrauer of Bronkhorst-

spruit, his real name was Von Ettinghausen.

Mr Mummbrauer stated that Thode came of military

stock but, because of his love of botany, rejected a

military carreer and emigrated to South Africa

changing his name at the same time. However, no

evidence to support these statements has been found.

170 HANS JUSTUS THODE (1859-1932), PIONEER PLANT COLLECTOR IN THE NATAL DRAKENSBERG

ARRIVAL AND SOJOURN IN CAPE

In Table 1 Thode’s collections from 1886-1906 are

listed chronologically. The first collection is dated

January 1886-December 1887. From this it would

appear that Thode must have arrived in Cape Town

in 1885 or early 1886. Why he chose to come to

South Africa is not known. While in the south-

western Cape he collected intensively and made

notes on the Cape flora. He published his observations

under the title “Die vier Jahreszeiten am Cap” in

a series of articles in Natuurwissenschaftliche Wochen-

schrift (1892).

TABLE 1. — Thode’s collections between 1886-1906

1. First Cape Series (Jan. 86-Dec. 87), 11 fascicles

(a) First Peninsula Series (Jan. 86-Aug. 87)

(b) 86 Supplement (Du Toit’s Kloof, Mamre, Paarlberg

Caledon) Aug. 86

(c) 87 — (Clarkson, etc.) Dec. 87

2. Kaffrarian Series (Dec. 87-Mar. 99)

(a) First East London Series (Jan. 88-Dec. 89)

(b) 87-88 nr. Supplement (Gamtoos R., Grahamstown,

King William’s Town) Dec. 87-Jan. 88

(c) 99 — (Komga, Tembuland, Pondoland) Feb. -Mar. 99

3. First Natal Coastal Series (May-Oct. 90)

4. First Highland Series (Weenen, Klip R. County, O.F.S.)

Oct. 90-Mar. 91

5. Second Natal Coastal Series (Apr. 91-July 1906)

(a) 91 Durban Series (Apr. -Nov. 91)

(b) 97 Supplement ( Dombeya natalensis, Avoca) June 97

(c) 1901 -(Barringtonia, Zeuxine , etc.). June 97

(d) 1 903— ( Disci polygonoides, Pinetown) Oct. 1903

(e) 1904 -{Habenaria, Mikania , Lower Umkomaas, etc.)

July 1904

6. Second Highland Series (Oliviershoek, ? etc.) Nov. 91-

Dec. 92

7. Third Highland Series (O.F.S., Mont-aux-Sources, etc.)

Jan. 93-June 94

(a) Emberton Supplement ( Cynorchis compacta) Aug. 94

(b) Transvaal Supplement ( Gerbera jamesonii) Dec. 1893

8. First Natal Midlands Series (Noodsberg) Oct. 94-Nov. 95

9. Fourth Highland Series (Lesotho, Witzieshoek). Dec.

1895- Jan. 97

(a) 97 Supplement (Brachycorythis tenuior, Emmaus) Jan. 97

10. Second Natal Midland Series (Umsindusi) Feb.-June 97

11. Second Cape Series (July 97-Jan. 99)

(a) Second Peninsula Series (July 97-Aug. 98)

(b) 98-99 Supplement (Cedarbergen, Bokkeveld) Oct. 98

Jan. 99

12. Fifth Highland Series (Oliviershoek Pass) Aug. 99-

Dec. 1900

13. Sixth Highland Series (Springfontein, Empangweni) June

1902- Jan. 1904

(a) Springfontein Fasc. (June-Nov. 1902)

(b) 1904 Supplement (Empangweni, etc.) Jan. 1904

14. Third Cape Series (Aug. 1903)

(a) Stormberg (Steynsburg) Fascicle

(b) Karoo (Beaufort West) Fascicle

(c) Tulbagh Fascicle

(d) Third Peninsula Series

15. Third Natal Midland Series (Kronsberg, Howick) Jan.

1903- Nov. 1904 (June 1906, Dec.)

(a) 1905 Supplement (Scholia brachypetala, etc.) Aug. 1905

(b) 1906 Supplement ( Eulophia bicolor , Aster erigeroides

etc.) Jan, June 1906

16. Seventh Highland Series (Bushman's Pass Reserve), Jan.

1905

From Table 1 it appears that Thode moved to

British Kaffraria in about December 1887 stopping

en route at the mission station at Clarkson near

Humansdorp. He collected in East London,

Grahamstown and King William’s Town up to

December 1889. It is not known where he was based

while in Kaffraria. His observations on the flora of

British Kaffraria were published in Engler’s Bot

Jahrb. 12: 589-607 (1890).

DRAKENSBERG AND SUBSEQUENT YEARS

In 1890 Thode “first set foot on the blessed soil

of fair Natal.” Between May and October 1890 he

collected along the coast of Natal, but then set off

for the “majestic range of the Kathlambe or Drakens-

berg, which has irresistably attracted and again and

again induced me to climb its lofty heights”. In the

middle of October 1890 he travelled by train to

Estcourt, a 10-hour journey, and saw for the first

time in the distance the still partly snow-clad

Drakensberg. Thode writes: “There it was lying in

front of me, the goal of my long cherished dreams and

expectations, the rarely visited and little known,

almost mythical realm of clouds, once the retreat of

robbing Bushmen, whose curious paintings can still

be found in many a cave below the cliffs of the river

valleys; now a sphere of solitude still unviolated by

the devastating activities of man. What an inex-

pressible incentive to be the first explorer of this

terra incognita, to open this veiled enchanted country

to investigating science and to return with trophies of

its interesting flora in the form of many a plant still

undescribed.” Medley Wood, the Curator of the

Natal Herbarium, tried to dissuade Thode from

exploring the Drakensberg, pointing out that the

mountains were infested with hostile Basothos.

This in no way deterred Thode.

He then travelled to the Empangweni Mission

(maintained by the Berlin Mission Society), which

he had chosen as his “fixed quarters”, because of

its fairly close proximity to the mountains. His first

Drakensberg expedition (29th October 1890) was to

Cathkin Peak and Champagne Castle. Unfortunately

inclement weather and recalcitrant Bantu bearers

prevented him from climbing either of these peaks.

Champagne Castle had been successfully climbed two

years earlier by the Stocker brothers. Thode

described this trip in great detail in a published news-

paper article and mentioned some of the more

striking plants he encountered. Presumably, his haul

of plant specimens was considerable.

A succession of expeditions followed, details of

which are given in Table 2.

To enable Thode to pursue his botanical interests,

he became a tutor at mission schools and on farms,

some situated close to the Drakensberg. He would

advertise his services in newspapers. The following

is an advertisement found amongst his papers:

“Teacher, elderly, ample experience and qualifications

desires re-engagement as tutor on farm (Natal or

Cape). All standards, Afrikaans, music, etc. Salary

£5 p.m. Numerous references.” Thode would spend

some time in one area and having exhausted it

botanically, would take up a post in another area.

This was the pattern of his existence to the end.

Thode sent most of his collections to Dr Adolf

Engler of the Berlin Herbarium, but apparently did

not make much out of his collections financially.

Prof. A. W. Bayer (letter of 3rd June 1965) writes

that “Thode was convinced that science was to serve

mankind and that it was quite wrong for anyone

to receive payment or to make money from scientific

work”. H. Bolus in a letter to A. W. Hill at Kew

(26th August 1908) describes Thode as “too peculiar

to be hired himself”. According to Prof. Bayer,

Thode would bring his specimens to Illings Store

at Ladysmith and this firm would pack and despatch

them to Berlin.

D. J. B. KILLICK

171

TABLE 2. — Thode’s main Drakensberg expeditions

By 1894 he had collected to such an extent and

seen so much of Natal that he was able to publish

an article in Bot. Jahrb. 18, Beibl. 3, 43: 14-45 (1894)

describing very accurately the botanical regions of

Natal.

In 1930 Markotter published a list of plants

collected by Thode at Witzieshoek, Oliviershoek Pass

and Koolhoek. These collections cover the period

1891 to 1914, consequently Thode must have spent

a considerable time in that area. At Oliviershoek he

presumably stayed on occasions with his sister Alwine,

whose husband C. A. Hammer had a store there.

At Witzieshoek it is known that he often stayed

with the Rev. Ross, a missionary, and at “Koolhoek”

(Fig. 2) about 21 km south south-east of Memel,

he was tutor to the Cronje family in 1907. According

to Mr C. J. Cronje, son of Karel Pieter Cronje,

Thode also tutored on the farm of Mrs Moorman

at Witzieshoek for a time.

However, Thode did not spend all his time during

that period in the Witzieshoek-Oliviershoek Pass-

Koolhoek area. From Table 1 it will be seen that he

visited the Cape Peninsula from July 1897 to August

1898 and the Cedarberg and Bokkeveld from

October 1898 to January 1899. He also passed through

the eastern Cape (Komga, Tembuland and Pondo-

land) in February to March 1899 presumably on his

way back to Natal. En route he collected on “Prospect

Farm” (Komga) belonging to H. G. Flanagan. In

August 1903 he visited the Karoo (Steynsburg,

Beaufort West, etc.). In 1904-1906 he was tutor on

the farm “Kronsberg”, Noodsberg, belonging to

Mr Herman Mummbrauer. On this farm he met

H. Rudatis and actually commenced a combined check-

list of birds with Rudatis and Mr G. Mummbrauer.

To gain some idea of Thode’s movements between

1908 and 1920 the writer visited the Stellenbosch

Herbarium, which contains 6309 of his specimens,

and scanned as many of the specimens as possible

in the time available, obtaining localities and dates

from the labels. Judging from the consecutiveness

of some of the dates, it would appear that Thode

had several fixed bases during this period. These

were apparently: “Sweet Home” (Krantzkop), Sept.

1910-Sept. 1911; “Warrock” (26 km west of Lady-

smith), Oct. 1911-March 1913; “Scottspoort” (near

Weenen), Oct. 1913-March 1914; Empangweni Mis-

sion (near Loskop), Nov. 1914-Jan. 1915; Shakas-

Fig. 2. — Homestead on the farm

"Koolhoek” near Memel where

Thode was tutor to the Cronje

family in 1907.

172 HANS JUSTUS THODE (1859-1932), PIONEER PLANT COLLECTOR IN THE NATAL DRAKENSBERG

kraal, April 1914-Sept. 1916 and finally “Yorkshire

Wolds” (near Rosetta), Aug.-Nov. 1918. If more

specimens had been examined, the periods might have

proved to be longer.

Many other localities were noted, some being near

the bases already mentioned. It is clear that Thode

paid regular visits to southern Natal between 1908

and 1920: there are numerous and sometimes annual

collections from “Friedenau”, Kenterton, “Moyeni”,

St Michael’s Mission, Oribi Flats, Marburg, Ivunga

River, Beach Terminus and Park Rynie. It is likely

that Thode visited H. Rudatis and stayed with him

on his farm “Umgai”. There was certainly contact

between Thode and Rudatis, because some of

Thode’s specimens are marked “ex herb. Rudatis”.

In 1911 Thode visited the Transvaal and collected

in Pretoria and Heidelberg and in early 1918 he

travelled to the Cape and collected at Worcester,

Laingsburg, Montagu, Beaufort West and Matjies-

fontein. In 1919 he crossed into Pondoland and

collected in the vicinity of the Mzamba River. In

1920 he again visited the Transvaal collecting in

Pretoria and Rustenburg.

In 1920 Thode was tutor on the farm “Altemooi”

belonging to Mr Adrian Daneel. “Altemooi” is

situated about 19 km SE of Wakkerstroom and

20 km N of Utrecht. He taught the Daneel children,

as well as those of Mr H. Repsold of the farm

“Tweekloof” next door. The portait of Thode (Fig. 1)

was given to the author by Mrs Elizabeth Rachel

Kritzinger (nee Daneel), one of Thode’s pupils.

Thode stayed at “Altemooi” for two years, then

apparently moved to “Tweekloof” for another two

years. Many of Thode’s specimens (1920-1924) were

collected on the farm “Kaffirdrift”, another farm

belonging to Mr Repsold.

Thode’s movements from 1924 to 1931 have been

traced by examining the accession book in the

National Herbarium, Pretoria, and then getting dates

from his specimens in the herbarium (there are 2859

Thode specimens in PRE). In this way Table 3 was

compiled. From Table 3 it will be seen that Thode

spent a considerable time in the Cape and Transvaal.

In 1925-1926 he was tutor on Mr James Trollip’s

farm “Glen Makopo”, Hobhouse, in the district

of Thaba Nchu, Orange Free State.

FINAL DAYS

According to Table 3 Thode must have made his

last collections at Witbank in the Transvaal from

December 1930 to March 1931. No records can be

traced of specimens collected later than this.

It is likely that he spent his last days in Durban

living at 170 Mansfield Road (the address given on

his death notice). It is also possible that he spent

some of his time in the Natal Herbarium naming

his collections. The late Miss Helena M. L. Forbes,

former curator of the Natal Herbarium, wrote to

Miss Gunn (30th October 1953): “Thode had been in

the habit of spending a month or more at the

Herbarium naming his specimens every few years.

I think he came three times after I came here, once

to name up a collection, which he sold to Stellenbosch

University”.

On 30th May 1932 Dr A. P. D. McLean, who was

then stationed at Natal Herbarium, wrote to

Dr I. B. Pole Evans, Chief of the Division of Botany,

stating that Thode was in Addington Hospital in a

critical condition and had been visited by Miss Forbes.

According to Miss Forbes, Thode’s impecunious

state weighed heavily on his mind even when delirious

and he was anxious that his collections be sold.

TABLE 3. — Collecting localities from 1924 to 1931

On 1st June 1932 Thode died and was buried in

a communal grave with no gravestone at Stella

Wood Cemetery, Block N, 1374. Later the grave

site was sold and to-day several gravestones occupy

the site.

David Strachan & Taylor, accountants of Durban,

were asked by the Master of the Supreme Court to

handle Thode’s estate. Thode died insolvent owing

£52. 6s. 6d. It was hoped that the sale of his collection

of 2859 numbers would defray this debt. To this end

negotiations between Strachan & Taylor and the

Division of Botany proceeded protractedly with

Dr McLean pressing for the purchase and the Division

of Botany expressing little interest because, as

Dr E. P. Phillips put it, “The cream of Thode’s

collections i.e. his high mountain gatherings, had

earlier been sold to the University of Stellenbosch”.

Eventually, on 3rd May 1934, the Division of

Botany offered £36 and this was accepted by the

Administrators. There were three or four duplicates

of each specimen, so the collection was made up

into three sets, one for the Natal Herbarium, one

for the National Herbarium, Pretoria, and one for

Kew. The specimens were numbered Al, A2, etc. to

avoid confusion with his other collections. Unfor-

tunately Thode did not adopt a system of continuous

consecutive numbering. This collection was mainly

of Transvaal and Cape plants (see Table 3, which

was based on this collection). The large wooden

trunk in which Thode stored his specimens and other

effects, is still in the Natal Herbarium, Durban.

Thode’s personal effects, beds and odds and ends,

were left to his sister Mrs Alwine Hammer of 340

Vause Road, Durban. Some of his books were sold

by M. Thompson & Co., auctioneers, for 4/6d.

D. J. B. KILL1CK

173

THODE, THE MAN

Fortunately the author was able to contact four

of Thode’s past pupils, namely Mr Erich Mumm-

brauer formerly of the farm “Kronsberg”, Noods-

berg, in Natal, Mr C. J. Cronje formerly of the

farm “Koolhoek” near Memel in the Orange Free

State, Mrs E. R. Kritzinger (nee Daneel) formerly of

the farm “Altemooi” near Utrecht in Natal and

Mrs Lorna Grove (nee Trollip) formerly of the farm

“Glen Makopa”, Hobhouse, in the Orange Free State.

Discussions with these people revealed a

remarkably similar memory of the man. Thode clearly

left a lasting impression on them. He was slightly

built with a moustache and small beard. He always

wore a brown suit buttoned to the top with a wide-

brimmed hat or helmet on his head. His plant press

was slung high up on his shoulders. He was short-

sighted and wore thick spectacles. Thode was an

excellent pianist, being passionately fond of

Beethoven. He showed little patience with his pupils

if they played incorrect notes. Mr Cronje relates how

when the family received visitors the children would

be asked to play the piano. When a mistake was made,

Thode would storm over to the piano and correct

the young pianist. The Cronjes asked him to desist

from this practice, so on similar occasions when

mistakes were made, Thode would walk agitatedly

around the room pretending to put all the pictures

straight.

He seems to have been an accomplished poet (there

are several unpublished poems amongst his papers)

and was very fond of the works of Goethe.

He had a good knowledge of the classics as the

following incident will show. Mr Heinz Lorenz a

former pupil of Thode’s at Empangweni Mission

School told Mr R. G. Strey of the Natal Herbarium

that Thode once taught chemistry and physics at a

school in Pietermaritzburg. In the neighbouring

classroom the headmaster was giving a lesson in

Latin. Thode overhead him incorrectly quoting some

Latin verse or prose. When the headmaster a few

minutes later left the classroom for some reason or

other, Thode quickly nipped into the classroom and

corrected the offending quotation on the blackboard.

On his return, the headmaster was furious that the

quotation had been corrected and promptly sacked

Thode.

Mr E. Mummbrauer recalls that after dinner

Thode would retire to his room and could be heard

walking aroung the room conducting an imaginary

orchestra.

Apparently he never rode on horseback, but used

horses for carrying his equipment.

Thode was fond of good food, but disliked

Mrs Lorenz’s “Mehlklosse”. According to Mr H.

Struck of Draycott, Thode would sometimes outstay

his welcome on the farm and stay for weeks and

weeks. The only way to get rid of him was to put

sago pudding regularly on the table.

Mr Struck tells of an incident when Thode went

visiting the Zunckels in the Bergville area. To get

to the Zunckels, Thode had to cross a river in full

spate. He took off his clothes, tied them into a

bundle and started to ford the river. Unfortunately

the current was strong and he slipped, losing his

clothes in the process. Mr Struck describes the

amusing sight of Thode running from tree to tree to

hide his nakedness and when near the Zunckel

homestead being spotted by the children, who ran

i idoors shouting “There is a naked man outside”.

Thode was no respecter of authority. This is amply

shown in a letter he wrote to the permit secretary

in Durban. He castigated him for his lack of

knowledge of geography and ended off with some

incredibly rude remarks. He was most indignant

when learning from Mr Sydney Barnes, the first

conservator of the Drakensberg Game Reserve that

he would have to pay 6d per day for each of his

party i.e. for himself, Bantu and horse. He wrote:

“The Government, or rather the Forestry Depart-

ment, under whose supervision the Game Reserve

is kept, should know best why and for what purpose

such an absurd system of (taxation) pillaging

travellers has been introduced . . . for our legislators

seem too fond of indulging in measures, the wisdom

of which is shrouded in mystery quite impenetrable

to ordinary common sense”.

He was obviously a man with a short temper.

Prof. A. W. Bayer (letter to author, 3rd June 1965)

tells the story of a collecting trip by ox-wagon made

by Thode with Rudolf Schlechter (probably between

June and October 1893). Schlechter was apparently

rather strict with his Bantu servants and one morning

they found that the servants had deserted. After

No... 3.

Herbarium, University Stellenbosch.

H ERBARIUM I Ex JUSTUS ThODE.

Fig. 3. — Example of the special

label used for Thode’s col-

lection in Stellenbosch Her-

barium. The handwriting is

Thode’s.

174 HANS JUSTUS THODE (1859-1932), PIONEER PLANT COLLECTOR IN THE NATAL DRAKENSBERG

breakfast, they proceeded to break up camp and

to round up and inspan the oxen. Schlechter then

picked up the driver’s whip and announced to Thode

“I will drive; you can voorloop”. Thode was so

annoyed at this brusque instruction that he went up

to the wagon, took his rucksack, turned his back on

Schlechter and disappeared into the bush.

CONTRIBUTIONS TO SOUTH AFRICAN BOTANY

Thode’s contributions to South African botany

lie in the many plant specimens he collected during

his 46 years in the country and in his seven

publications. As already indicated, he was the first

to collect in the Natal Drakensberg, being several

years earlier than H. G. Flanagan and H. Bolus

(summer 1893-94) and Maurice Evans (1894). Thode

was much chagrined when plants like Erica frigida,

E. algida and Gladiolus flanaganii were described

from material collected by Flanagan instead of from

his material collected three years earlier (letter from

Thode to H. Bolus, 5th February 1896).

It is not known how many specimens Thode

collected, because he did not consecutively number

his specimens. He must have sent thousands of

specimens to the Berlin Herbarium; there are 6309

of his specimens in the Stellenbosch Herbarium

(Fig. 3), collected between 1886 and 1920, and

there are 2859 specimens in the National Herbarium,

Pretoria, collected between 1924 and 1931, with

duplicate sets in the Natal Herbarium and at Kew.

The specimens he collected were of good quality, but

often lacked adequate field notes. Oddly enough,

Thode did not collect grasses.

Thode is commemorated in the following species:

Alepidea thodei Diimmer, Athanasia thodei H. Bol.,

Disa thodei Schltr., Erica thodei Guth. & H. Bol.,

Geranium thodei Schltr., Holothrix thodei Rolfe,

Kniphofia thodei Bak., Lessertia thodei L. Bol.’

Manulea thodeana Diels, Osteospermum thodei Mark ’

Rhodohypoxis thodiana (Nel) Hilliard & Burtt’,

Romulea thodei Schltr. and Sebaea thodeana Gilg.

Thode’s papers though few were of a high standard.

The following is a list of his publications:

Thode, J.,* 1890. Die Kustenvegetation von Britisch-

KafFrarien und ihr Verhaltnis zu den Nachbarfloren

Bot. Jahrb. 12: 589-607.

Thode, J„ 1892. Die vier Jahreszeiten am Cap. Naturwissen-

schaftliche Wochenschrift. 7,14: 131-133. 15- 144-147

21:206-207.23:226-228.

Thode, J., 1892. Die botanische Hohenregionen Natals. Bot.

Jahrb. 18. Ill Beibl. 43: 14-45.

Thode, J. 1901. The botanical regions of Natal determined by

altitude. 16 pp. Durban: Durban Field Naturalists’

Society.

Thode, J., 1913. Leucadendron natalense Thode & Gilg.

Notizbl. Bot. Gart. Bert. 5: 290.

Thode, J., 1922. Bowkeria citrina Thode. Kew Bull. 1922- 31

(1922).

Thode, J., 1924. A new Pyrenacantha from Natal. J. Bot.

42: 115-116 (1924).

In his first paper, “Die Kustenvegetation von

Britisch-Kaffrarien und ihr Verhaltnis zu den Nach-

barfloren” (1890), Thode compared the coast

vegetation of British Kaffraria with that of the

south-west Cape and Natal. He showed that although

Compositae and Leguminosae were the two dominant

families in all three regions, the remaining families

differed significantly in their order of importance.

Illustrating his broad knowledge and interest in

floristics and plant geography, Thode compared the

flora of East London with the floras of Ethiopia and

India.

* Thode always omitted the initial of his Christian name, Hans.

TABLE 4. — Location of some Thode localities mainly in Natal

D. J. B. KILLICK

175

His second paper, “Die vier Jahreszeiten am

Cap” (1892) showed a surprising knowledge of the

south-west Cape flora after only a short period of

residence there. Thode described the four seasons

of the Cape and related them to the flora; he com-

mented on the effect of fire on fynbos; he mentioned

the exploitation of the flora and urged that con-

servation measures be adopted.

In 1894 he published “Die botanische Hohen-

regionen Natals”, an accurate picture of altitudinal

zonation of vegetation in Natal. This paper stamps

Thode as a competent plant geographer. His

description of the vegetation of the Natal Drakens-

berg is the first on record and is extremely accurate

as to detail.

In 1901 he published “The botanical regions of

Natal determined by altitude”. This represents a

condensed version of his 1894 paper, translated into

English and read before the Durban Field Naturalists’

Society on 16th May 1901.

The remaining papers were descriptions of new

species.

In conclusion, it may be said that Thode’s

contributions to South African botany were

considerable, especially if it is realized that he was

an amateur botanist, attached to no institution, with

no library facilities and collecting in remote areas

often never collected before. Bolus in writing to

A. W. Hill at Kew (26th August 1905) described

Thode as “wanting in systematic effort and regularity,

and so has not done all he might have done”. This

is a somewhat harsh judgement, but Bolus does

concede that “he has written some valuable papers

on the flora of that region”.

LOCATION OF CERTAIN THODE COLLECTING

LOCALITIES

Taxonomists have frequently found difficulty in

locating some of Thode’s localities. This is partly

due to Thode’s frequent use of farm names. To

assist taxonomists, the exact location of some of the

“problem” localities is given in Table 4.

ACKNOWLEDGEMENTS

Thanks are due to Miss M. D. Gunn for having

done some of the spadework in this research and for

her continued interest in its progress. Thanks are

also due to Miss E. Hagedorn for translating some of

the original manuscripts from German to English.

UITTREKSEL

V i Verslag word gegee van die lewe van Hans Justus

Thode (1859-1932), vanaf s& aankoms in Kaapstad

in 1885 o/" 1886 tot sy dood in Durban in 1932. Thode

was die baanbreker-plantversamelaar van die Natalse

Drakensberge, maar hy het ook verder weg in al vier

provinsies van die Republiek plante versamel. 'n

Waardasie van sy bydraes tot Suid-Afrikaanse plant-

kunde word gemaak.

REFERENCES

Markotter, E. I., 1930. 'n Plantgeografiese skets en die flora

van Witzieshoek, O.V.S.; Oliviershoekpas, Natal en Kool-

hoek, O.V.S. Ann. Univ. Stell. 8, Al : 1-50.

t

Bothalia 12,2:177-179 (1977)

The South African species of Teucrium (Lamiaceae)

L. E. CODD*

ABSTRACT

While writing up the three South African species of Teucrium for the Flora of Southern Africa it became

necessary to replace two well-known names as follows: T. Irifidum Retz. (1779) (=T. capense Thunb., 1800)

and T. kraussii Codd (=T. riparium Hochst., 1845, non Rafin., 1838).

RESUME

LES ESPECES SUD-AFRICAINES DE TEUCRIUM ( LAMIACEAE )

En etudiant les trois especes sud-africaines de Teucrium pour inclusion dans la Flore d'Afrique

australe, il s'est avere necessaire de remplacer comme suit deux noms largement connus: T.trifidum

Retz. (1779) (=T.capense Thunb., 1800) et T.kraussii Codd (= T.riparium Hochst., 1845, non Rafin.,

1838).

Teucrium is a large genus of perhaps 300 species

widely distributed over the temperate and warmer

regions of the world, chiefly in the northern

hemisphere, but with a few species in Australia, South

America and Africa, mainly on the mountains of

north-east tropical Africa and into South Africa.

Three species are recognized in South Africa and,

because the small white flowers are very similar and

provide no diagnostic characters, the species are

distinguished on vegetative and inflorescence

characters. As may be expected, misidentifications

have occurred in herbaria, resulting in miscon-

ceptions regarding distribution.

Considering the fact that the names commonly

applied to these three species have been in use for

well over a century, it came as a surprise to find

that two name changes were required: (1) T. irifidum

Retz. (1779) must replace T. capense Thunb.; and

(2) T. riparium Hochst. (1845) is a later homonym

of T. riparium Rafin. (1838), an American species,

and is replaced by a new name, T. kraussii

Codd.

All three species are used medicinally for stomach

disorders and haemorrhoids, as well as for treating

snake-bite, while meat suspected of being infested

with anthrax is boiled with the plant to render it

harmless to eat. Common names such as Aambeibossie

and Maagbossie refer to these properties, while

Paddaklou and Akkedispoot refer to the variously

lobed leaves.

Key to species

Peduncles usually 1-flowered (rarely 3-flowered) and much shorter than the internodes; shrub-

lets 10-30 cm tall; leaves narrowly 3-fid, 0,5-2, 5 cm long 1. T. africanum

Peduncles 3-7-flowered, often as long as or longer than the internodes; soft shrubs 30-110

cm tall ; leaves 3-5-fid or -lobed, or subentire with a few teeth towards the apex, 2-6

cm long:

Leaves more or less 3-5-fid or -lobed, if subentire then drying greyish 2. T. trifidum

Leaves entire or few-toothed towards the apex, usually drying dark brown 3. T. kraussii

1. Teucrium africanum Thunb., Prodr. 2: 95

(1800); FI. Cap. ed Schult. 445 (1823); Benth.,

Lab. 669 (1835); in E. Mey., Comm. 243 (1837); in

DC., Prodr. 12: 577 (1848); Skan in FI. Cap. 5, 1 : 384

(1910). Type: Cape, without locality, Thunberg s.n.

(UPS, holo.).

Ajuga africana (Thunb.) Pers., Syn. PI. 2: 109 (1807).

Greyish, bushy shrublet 10-25 (-30) cm tall,

branching freely from the base; stems erect to

decumbent, simple or sparingly branched, rather

densely leafy, greyish glandular-tomentose. Leaves

3-fid, 0,8-3 cm long, grey-green; lobes linear to

linear-oblong, 0,5-2, 5 cm long, 1-3 mm broad,

occasionally the median lobe again 3-fid; basal

portion of leaf narrow, up to 3 mm broad. Inflores-

cence simple, occupying the upper third of half of

the stem; flowers solitary or rarely 2 or 3 per peduncle;

peduncles 3-8 mm long, usually distinctly shorter

than the internodes.

Found under fairly arid conditions in macchia,

karoo, coastal or thorn scrub from Bredasdorp to

near Grahamstown and, inland, to Middelburg and

Graaff Reinet.

May be separated from the next species, T. trifidum,

by its smaller stature, rarely exceeding 30 cm tall,

and the usually solitary flowers on short peduncles.

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Bentham, Lab. 670 (1835) and in DC., Prodr.

12: 577 (1848) placed T. trifidum Retz. (1799) as

being possibly conspecific, while Skan in FI. Cap,

5,1: 384 (1910) added T. trifidum Wendl. (1798) with

a question mark. These names are discussed under

the next species.

2. Teucrium trifidum Retz., Obs. 1: 21 (1779).

Type: Cape, without locality, right-hand specimen

on sheet in Hb. Retzius (LD, lecto.).

T. trifidum Wendl., Bot. Beobacht. 50 (1798); nom. illegit.

Type- not indicated. T. capense Thunb., Prodr. 95 (1800); FI.

Cap. ed Schult. 445 (1823); Benth., Lab. 667 (1835); in E.

Mey., Comm. 243 (1837); in DC., Prodr. 12: 577 (1848); Skan

in FI. Cap. 5,1: 385 (1910); Wilman, Checklist Griq. W. 231

(1946); Jacot-Guill., FI. Lesotho 236 (1971); Ross, FI. Natal

302 (1972). Type: Cape, near “Zeekoerivier” (Humansdorp

district), Thunberg s.n. (UPS, holo.). T. africanum sensu Wilman,

l.c. (1946).

Ajuga capensis Pers., Syn. PI. 2: 109 (1807).

An erect soft undershrub 30-1 10 cm tall, branching

freely from the base; stems virgate, branching freely

in the upper half or third, woody below, herbaceous

above, thinly and shortly greyish tomentose. Leaves

usually deeply 3-fid or 3-5-partite, rarely almost entire

or shortly toothed towards the apex, 2-6 cm long,

drying greyish-green to grey-brown, thinly tomentose

to almost canescent; lobes linear to lanceolate,

1-3,5 cm long, 3-8 mm broad, often again shortly

lobed or toothed; basal portion of leaf 3-8 mm

broad. Inflorescence a leafy panicle occupying the

178

THE SOUTH AFRICAN SPECIES OF TEUCRIUM (LAMIACEAE)

DOTANISKA

MC't.tT

LUND

N« I lerb , Pretoria

i’RC Nee. No.. 629/

Fig. 1. — Type of Teucrium trifi-

dum in the Retzius Herba-

rium, LD (right-hand speci-

mens selected as the lecto-

type).

upper third of the stem; flowers usually in 3-7-

flowered pedunculate cymes, rarely solitary; peduncles

5-20 (25) mm long, usually longer than the internodes.

Common in the central and south-western

Transvaal, apparently not extending beyond the

Soutpansberg, but extending westwards to the

northern Cape Province and just entering Botswana,

southwards to northern Natal, central Orange Free

State and eastern Cape Province, reaching its south-

westerly limit about Humansdorp. Characteristically

found in dry woodland where it is often gregarious

under thorn trees or in bush groups, particularly on

overgrazed or disturbed places.

The leaves of S’, trifidum vary considerably from

lanceolate and almost entire to deeply 3-5-partite.

It overlaps with T. afrieanum in the southern Cape

from about Grahamstown inland to Queenstown and

occasional specimens, from the latter area particularly,

may be difficult to place with certainty. T. afrieanum

can usually be separated on the basis of the shorter

stature, rarely exceeding 30 cm tall, and the usually

solitary flowers on short peduncles. In the northern

Cape, under semi-arid conditions, T. trifidum tends to

produce plants shorter than usual with somewhat

shorter, but 3-flowered, peduncles. These have often

in the past been identified as T. afrieanum, e.g. in

Wilman, Checklist Griq. W. 231 (1946), but their

affinity lies with T. trifidum.

In eastern Cape, about Komga, T. trifidum overlaps

with T. kraussii and here again occasional specimens

may need careful study to be sure of their identity.

In such cases, specimens of T. trifidum with subentire

leaves can be recognized by the leaves being some-

what smaller and greyish-green, as against the longer

and broader leaves of T. kraussii, which tend to dry

dark brown.

As mentioned under the previous species, the

names T. trifidum Retz. (1779) and T. trifidum Wendl.

(1798) were tentatively included in T. afrieanum by

earlier authors. The type of T. trifidum Retz. is

present in the Retzius Herbarium, Lund, and was

kindly sent on loan to me by the Curator, Dr Ove

Almborn, whose assistance is gratefully acknowledged.

This was examined in relation to the two names

published by Thunberg in 1800, T. afrieanum Thunb.

and T. capense Thunb.

L. E. CODD

179

The type sheet, annotated as Teucrium trifidum N.

by Retzius on the reverse side, has two specimens

mounted on it (see Fig. 1). The left-hand specimen

corresponds with T. africanum Thunb. and the right-

hand specimen with T. capense Thunb. It is, therefore,

necessary to scrutinise the protologue carefully and

a copy of the significant parts is given below:

65. TEUCRIUM trifidum foliis lanceolatis trifidis, pedunculis

axillaribus trifloris.

Habitat ad Cap. B. Spei, unde habui, dono D. BLADH.

Caulis tetragonus, hirtus, ramosus, foliosus.

Folia opposita, lanceolata, majora lobo, minora dente

utrinque acuto incisa, ceterum integra, supra scabra,

subtus tomentosa; tomento flavescente vix conspicuo.

Pedunculi oppositi, axillares, triflori, longitudine inter-

nodiorum.

The epithet trifidum would apply more correctly

to the left-hand specimen, but the short diagnosis is

applicable mainly to the right-hand specimen. The

rest of the protologue describes the significant

characteristics of the right-hand specimen, eg.:

“ Caulis . . . ramosus . . “ Folia lanceolata, majora

lobo, minora dente utrinque acuto inciso, ceterum

integra . . “ Pedunculi . . . triflori, longitudine

internodiorum.” In fact no mention is made of the

characters exclusive to the left-hand specimen (the

deeply trifid leaves and single flowers on short

peduncles), while all the characteristics mentioned

above apply exclusively to the right-hand specimen.

There is thus nothing in the description to compel

one to the view that it is based on a mixture of the

two specimens but, assuming that Retzius had both

before him, it is considered advisable to make the

right-hand specimen the lectotype. This means that

T. capense Thunb. goes into synonymy.

One may speculate how the specimens were acquired

by Bladh, who was a “supercargo” in the Swedish

East Indian Company from 1766 to 1784 and sent

plants from China, Indiafl and South Africa to

several Swedish botanists. He is not known to have

travelled at all in South Africa, but might have

obtained specimens from Thunberg, who had under-

taken an expedition together with Masson in 1773-74

to the Sundays River area, where both species are

found.

T. trifidum Wendl. is an illegitimate homonym.

No specimen is cited but the description suggests that

it is conspecific with T. trifidum Retz.

3. Teucrium kraussii Codd, nom. nov. Type:

Natal, Umlaas River, Krauss 153.

T. riparium Hochst, in Flora 66 (1845); Benth. in DC.,

Prodr. 12: 576 (1848); Skan in FI. Cap. 5,1: 385 (1910);

Compton, FI. Swaz. 66 (1966); Ross, FI. Natal 302 (1972);

nom. illegit., non T. riparium Rafin. (1838). Type: as above.

An erect soft undershrub 50-1 10 cm tall, branching

from the base; stems simple below, branched in the

upper half or third, softly woody below, herbaceous

above, finely to fairly densely tomentose, usually

with spreading hairs. Leaves narrowly lanceolate to

oblong-lanceolate or oblanceolate, 2,5-6 cm long,

remotely 1 — few toothed towards the apex or entire,

subglabrous or sparingly hispidulous above, sparingly

to fairly densely hispid and minutely gland-dotted

below; apex obtuse to acute, base narrowly cuneate.

Inflorescence a leafy panicle occupying the upper

third of the stem, often diffusely branched; flowers

in 2-7-flowered pedunculate cymes; peduncles 6-20

mm long, usually longer than the internodes.

Distributed from Swaziland through semi-coastal

and midland Natal to King William’s Town district

in the Cape, in open bush and grassland.

T. kraussii overlaps with T. trifidum in the eastern

Cape Province and occasional intermediate specimens

may be difficult to place with certainty. The main

distinguishing characters are discussed under T.

trifidum.

UITTREKSEL

Tydens die beskrywing van drie Suid-Afrikaanse

soorte Teucrium vir die Flora of Southern Africa,

het dit nodig geblyk om twee goedbekende name as

volg te vervang: T. trifidum Retz. (1779) (=T. capense

Thunb., 1800) en T. kraussii Codd (=T. riparium

Hochst., 1845, non Rafin., 1838).

Bolhalia 12,2: 181-189 (1977)

A note on the Stachys aethiopica Complex

L. E. CODD

ABSTRACT

The Stachys aethiopica Complex is examined and a key provided to the species recognized. The com-

bination 5. natalensis Hochst. var. galpinii (Briq.) Codd is effected and two new species, S. reticulata Codd

and S. arachnoidea Codd are described.

RESUME

NOTE SUR LE COMPLEXE DES STACHYS AETHIOPICA

Le complexe des Stachys aethiopica est examine et une cle est fournie pour les especes reconnues.

Une nouvelle combinaison est operee: S. natalensis Hochst. var. galpinii (Briq.) Codd; et deux especes

nouvelles sont decrites: S. reticulata Codd et S.arachnoidea Codd.

Among the South African species of Stachys one

of the main taxonomic problems concerns the

delimitation of species in the S. aethiopica Complex.

These comprise mainly soft, nondescript, somewhat

straggly perennial herbs with small ovate cordate

leaves and relatively small, inconspicuous flowers.

The complex extends from the south-western extremity

of the Cape Province, through the eastern summer-

rainfall parts of southern Africa northwards into

tropical Africa.

S. aethiopica L., Mant. 1: 82 (1767) was the first

to be described and was presumably based on

material from the south-western Cape Province.

Further species have been distinguished on the basis

of differences in tomentum, whether the leaves are

petiolate or sessile, on the number of flowers per

verticil and, to some extent, on leaf shape.

Our species were reviewed by Skan in FI. Cap-

5,1: 336-367 (1910), who upheld 19 species in this

complex, namely: S. rehmannii Skan, S. malacophylla

Skan, S. sessilifolia E. Mey. ex Benth., S. galpinii

Briq., S. transvaalensis Guerke, S. rudatisii Skan,

S. parilis N.E. Br., S. aethiopica L., S. scabrida Skan,

S. cooperi Skan, S. harveyi Skan, S. serrulata Burch,

ex Benth., 5. attenuata Skan, S. fruticetorum Briq.,

S. leptoclada Briq., S. flexuosa Skan, S. cymbalaria

Briq., S. priorii Skan, and 5. sublobata Skan, while the

following were treated as “imperfectly known ’’species:

S. capensis Presl, S. graciliflora Presl, S. hispidula

Hochst. and S. natalensis Hochst. The only species

added since then is H. villosissima H. M. Forbes from

Natal.

Most of these entities represent local variations of

restricted extent which may or may not warrant the

rank of species. Due to inadequate collecting, many

of them are still very poorly known and a good deal

more material is required before their status can

be assessed.

In the present study the above 24 “species” are

reduced to 11, while two additional species are

described from the eastern Transvaal and Swaziland.

In few cases are there clear-cut differences between

species and it is often necessary to rely on a com-

bination of characters.

The picture which emerges is of two widespread

and variable species, from which a number of

segregate species have arisen or are in the process

of developing. The two widespread “parent” species,

which can usually be separated on the basis of

pubescence, are:

(1) S', aethiopica L., with a mainly southern

distribution from the south-western Cape to Natal,

Lesotho and the Orange Free State, with the lower

surfaces of the leaves and calyces thinly tomentulose

to hispid or densely glandular;

(2) a more northerly species, to which the name

S. natalensis Hochst. is applied, extending from

Natal to Swaziland, Transvaal and Rhodesia, with

the lower surfaces of the leaves and calyces hispid-

villous (sometimes only on the nerves) to densely

villous (often obscuring the surface), and with

scarcely any glands.

The two species meet in Natal, where a gradation

in pubescence occurs. In this critical area, however,

one can separate them on another character, the

number of flowers per verticil. In S', natalensis the

verticils are 2-flowered throughout its entire range;

in S. aethiopica the verticils appear to be always

4-6-flowered in Natal, Lesotho and the Orange Free

State. In the Cape occasional specimens, some of

them rather depauperate, are found with 2 or 3

flowers per verticil, but these do not have villous

pubescence and would not be confused with S.

natalensis. In this way, the two species can be

maintained as distinct and some of the confusion

of the past concerning the limits of S. aethiopica can

be clarified.

With regard to the number of flowers per verticil,

it must be remembered that there are certain segregate

species related to S. aethiopica which have 2-flowered

verticils (5. rudatisii, S. cymbalaria and S. sublobata),

in the same way that there are species with villous

to densely tomentose pubescence related to S.

natalensis but with 4-6-flowered verticils (5. malaco-

phylla and the newly described S. reticulata). In

other species the number of flowers per verticil is

not constant and may very from 2-6 (S. rehmannii,

S. scabrida and S. flexuosa).

From the foregoing it will be realized that the

construction of a key that will work satisfactorily

for every specimen is not an easy matter. The following

key must, therefore, act mainly as a guide. With the

increase in knowledge as further material comes in,

it is quite likely that our species concepts will be

modified.

182

A NOTE ON THE STACHYS AETHIOPICA COMPLEX

Key to species in the S. aethiopica Complex

Leaves and calyx villous, often densely so, the lower surface of the leaves sparingly hispid-villous to

densely tomentose:

Pubescence stellate 3. S. rehmannii

Pubescence not stellate:

Verticils 2-flowered:

Leaves discolorous, sparingly pubescent and greenish-brown above, white tomentose below: bracts

leaf-like 2. S. arachnoidea

Leaves concolorous; bracts usually differentiated from the leaves:

Leaves densely matted-tomentose below obscuring the surface 6. S. sessilifolia

Leaves densely to sparingly strigose or hispid below, often mainly on the nerves:

Inflorescence compact, 3-6 cm long; corolla purple, the lower lip up to 8 mm long, longer

than the tube 13. S. flexuosa

Inflorescence fairly compact to lax, (4-) 6-15 cm long; corolla whitish, the lower lip 5-7 mm

long, shorter than the tube 7. S. natalensis

Verticils normally 4-6-flowered (an occasional 2-flowered verticil may be present):

Stem with longish appressed retrorse hairs; leaves often discolorous and reticulate-veined

below 4. S. reticulata

Stem spreading-pilose to villous or with short retrorse hairs; leaves concolorous:

Leaves densely matted-tomentose below, obscuring the surface:

Leaves petiolate; uppermost bracts shorter than the calyx 5. S. malacophylla

Leaves usually sessile or subsessile; uppermost bracts longer than the calyx.... 6. S. sessilifolia

Leaves strigose to hispid below, mainly on the nerves:

Leaves drying greenish, truncate to subcordate at the base; inflorescence usually

compact 13. S. flexuosa

Leaves usually drying brownish, deeply cordate at the base; inflorescense usually lax. .8. S. aethiopica

Leaves and calyx subglabrous to hispid or glandular-tomentulose, the lower surface of the leaves glabrous

to hispid (mainly on the nerves) or glandular-puberulous:

Verticils 2-flowered:

Lower surface of leaf and calyx gland-dotted:

Leaves 2-4,5 x 1 ,5-2,5 cm, fairly thick-textured; margin finely and regularly crenulate. .1.5. rudatisii

Leaves less than 2 cm long or, if longer, margin rather coarsely toothed :

Leaves broadly ovate to suborbicular, shallowly crenate 11. 5. cymbalaria

Leaves ovate, usually thin-textured, margin rather coarsely toothed 8. 5. aethiopica

Lower surface of leaves and calyx eglandular or nearly so :

Stem glabrous to sparingly retrorse-scabrid or with scattered longish hairs:

Leaves drying dark brown, ovate-deltoid; hairs on upper leaf surface, when present, bulbous-

based 10. 5. scabrida

Leaves drying greenish-brown; hairs on upper leaf surface not bulbous-based:

Leaf blade broadly ovate to suborbicular, 8-15x6-12 mm; margin shallowly crenate

11. 5. cymbalaria

Leaf blade narrowly triangular, 10-15x4-6 mm; margin deeply crenate 12. 5. sublobata

Stem variously pubescent but not as above:

Stem and rhachis usually hispid-villous, fairly rigid, eglandular (Natal, Swaziland, Transvaal)

7. 5. natalensis

Stem and rhachis shortly retrorse-pubescent to softly pilose often glandular (western Cape)

8. 5. aethiopica

Verticils 4-6-flowered (an occasional 2-flowered verticil may be present):

Stem subglabrous to sparingly retrorse-scabrid or with scattered multicellular retrorse hairs; hairs

on upper leaf surface, when present, bulbous-based 10. 5. scabrida

Stem variously pubescent but not as above:

Inflorescence of 1-4 (rarely more) verticils, often subcapitate to somewhat spaced below; calyx,

rhachis and leaves eglandular; stem shortly and softly retrorse-pubescent 9. 5. graciliflora

Inflorescence usually slender of few to several spaced verticils (rarely subcapitate); glandular hairs

or gland-dots often present on calyx, rhachis and lower surface of leaves; stem variously

pubescent 8. 5. aethiopica

1. Stachys rudatisii Skan in FI. Cap. 5, 1: 347

(1910); Ross, FI. Natal 304 (1972). Type: Natal,

Dumisa, Rudatis 405 (K, holo. ; NH!, PRE!).

Stems glandular hispidulous with retrorse multi-

cellular hairs and copious short glandular hairs.

Leaves petiolate, rather large; blade ovate, (2-)

3-4,5 x(l ,5-)2-2,5 cm, the lower surface densely

glandular-hispidulous; margin finely crenulate. Bracts

leaf-like to somewhat differentiated. Verticils 2-

flowered. Calyx glandular hispid.

A little-known species from southern Natal where

it apparently grows in damp grassy places among

rocks and in shady thickets.

Natal.— 3029 (Kokstad): Ngeli Mt. (-DA), Hilliard &

Burn 3488. 3030 (Port Shepstone): Dumisa (-AD), Rudatis 405.

S. aethiopica (p. 186) often has densely glandular

stems and leaves but the leaves are usually smaller,

broadly ovate and rather coarsely toothed, while the

verticils are normally 4-6-flowered, not 2-flowered

as in S. rudatisii. In the type, Rudatis 405, the bracts

are large and leaf-like, while in Hilliard & Burtt

3488 the bracts are reduced, making it somewhat

intermediate with S. aethiopica. A specimen from the

Hogsback, Rattray sub BOL 14275, may belong here,

but the leaves are broadly ovate and subsessile.

If the specimen were included here, the known area

of the spesies would be extended considerably but

more material, especially from the type area, is

required before the limits of the species can be

assessed.

2. Stachys arachnoidea Codd, sp. nov., a S.

natalensis Hochst. foliis discoloribus, subtus dense

albo-tomentosus et minute flavo-punctatis differt.

L. E. CODD

183

Herba, perennis, aromaticus; caules procumbens

vel subscandens, ramosus, usque ad 1 m longus,

quadrangularis, sulcatus, dense et molliter albo-

tomentosus. Folia subsessilia vel breviter petiolata;

petiolus usque ad 1 cm longus; lamina late ovato-

deltoidea vel subrotunda, 1,8-4 cm longa, 1,5-

3 cm lata, chartacea, supra olivacea tenuiter

pubescens, subtus dense albo-arachnoidea et minute

flavo-punctata, apice obtuso vel rotundato, basi

cordata, margine crenulata. Inflorescentia laxa, 5-15

cm longa; bracteae foliiformes versus apicem leviter

parviores; verticillastri 2-floribus; pedicelli 1,5 mm

longi. Calyx tubuloso-campanulatus, 8-9 mm longus,

dense et molliter albo-tomentosus et minute punc-

tatus; dentes deltoideo-lanceolati, 3-4 mm longi.

Corolla alba, malvino-maculata, subtiliter glanduloso-

pubescens; tubus 8-9 mm longus, 2 mm diam.;

labium posticum ascendens, concavum, 3-4 mm

longum, 3 mm latum, apicem rotundato; labium

anticum horizontale, 8-9 mm longum, lobo medio

obcordato, 4 mm longo, lobis lateralibus rotundatis,

1,5 mm longis. Stamina 4, 2-3 mm fauce exserta.

Stylus bifidus, 2 mm exsertus.

Type. — Swaziland, near Mbabane, Compton 25890

(PRE, holo.!).

Perennial aromatic herb; stems procumbent to

subscandent, branched, up to 1 m long, 4-angled,

grooved, densely and softly white tomentose.

Leaves subsessile to shortly petiolate; petiole up to

1 cm long; blade broadly ovate-deltoid to subrotund,

1 ,8-4 cm long, 1 ,5-3 cm broad, chartaceous, greenish

and thinly pubescent above, the lower surface with a

dense white web-like tomentum and freely supplied

with minute yellowish pustule-like gland-dots; apex

obtuse to rounded, base cordate; margin crenulate.

Inflorescence lax, 5-15 cm long; bracts leaf-like,

slightly smaller towards the apex; verticils 2-flowered;

pedicels 1,5 mm long. Calyx tubular-campanulate,

8-9 mm long, densely and softly white tomentose

and minutely gland-dotted; teeth deltoid-lanceolate,

3-4 mm long. Corolla white spotted with mauve,

finely glandular-pubescent; tube 8-9 mm long,

2 mm in diameter; upper lip ascending, concave

3-4 mm long, 3 mm broad, apex rounded; lower lip

horizontal, 8-9 mm long, median lobe obcordate,

4 mm long, lateral lobes rounded, 1,5 mm long.

Stamens 4, exserted by 2-3 mm. Style bifid, exserted

by 2 mm. Fig. 1.

Found in forest margins and on grassy slopes in

the mountains of eastern Transvaal and northern

Swaziland at altitudes of 1 300-2 000 m.

Transvaal. — 2329 (Pietersburg): near Haenertsburg (-DD),

Codd9453. 2330 (Tzaneen): Westfalia Estate (-CA), Scheepers

729; Woodbush Forest (-CC), Mogg 20300; Agatha (CC),

McCallum s.n. 2430 (Pilgrims Rest): Mariepskop (-DB),

Van der Schijff 4401 ; 5139. 2530 (Lydenburg): 18 km W. of

Sabie (-BA), Codd 9858.

Swaziland. — 2531 (Komatipoort): Havelock area (-CC),

Miller 5858; Compton 30650. 2631 (Mbabane); Forbes Reef

(-AA), Compton 30492; near Mbabane (-AC), Bolus 12242

BOL); Compton 25890; 26481.

Although first collected as early as 1905 by Bolus

near Mbabane in Swaziland, this species has been

confused with S. transvaalensis (now included in

S. natalensis, p. 185) and S. rudatisii. It was listed

under both these names by Compton, FI. Swaz. 66

(1966), but can be distinguished by the rounder,

discolorous leaves with white cobwebby tomentum

and minute yellow gland-dots on the lower surfaces.

It shows no close relationship to any other species

and is probably the most clear-cut member of the

S. aethiopia complex.

Fig. 1. — Stachys arachnoidea, near Mbabane, Swaziland

{Compton 25890 in PRE, holotype), x

3. Stachys rehmannii Skan in FI. Cap. 5,1: 345

(1910). Type: Transvaal, Houtbosch, Rehmann 6178

(K, holo.!).

Stem sparsely to densely stellate-hispid. Leaves

petiolate, blade ovate-deltoid to subrotund, 1,6-

2, 2 x 1,5-2 cm, densely grey stellate-hispid on both

surfaces. Brads differentiated from the leaves towards

the apex. Verticils usually 2-flowered but sometimes

4-6-flowered. Calyx densely and shortly stellate-

villous.

Found among rocks in mountain grassland at

altitudes of 1 300-2 200 m in the northern and

north-eastern Transvaal.

Transvaal. — 2328 (Baltimore): Blouberg (-BB), Codd &

Dyer 9022; Van der Schijff 5399 ; Strey & Schlieben 8515. 2329

(Pietersburg): Houtboschberg (-DD), Schlechter 4454. 2330

(Tzaneen): 2 km E. of Steilkop, New Agatha (-CC), Muller &

Scheepers 73. 2430 (Pilgrims Rest): The Downs (-DD), Rogers

22027.

Readily distinguished from all other members of

the S. aethiopica complex by the presence of dense

stellate pubescence on all parts of the plant.

4. Stachys reticulata Codd, sp. nov., a 5.

aethiopica L. foliis discoloribus, subtus reticulatis

dense tomentosis, dentibus calyce patentibus differt.

Herba, perennis, aromaticus; caules decumbens vel

procumbens, ramosus, usque ad 60 cm longus,

inflorescentiis adscendentibus, dense retrorso-pilosus.

Folia petiolata; petiolus 5-12 mm longus; lamina

ovato-deltoidea, late ovata vel subreniformis,

1-2,5 x 0,8-2, 2 cm, subcoriacea, plerumque discolor.

184

A NOTE ON THE STACHYS AETHIOPICA COMPLEX

supra hispida, subtus reticulata, dense glanduloso-

tomentosa, basi late cordata, apice obtuso vel

subacuto, margine crenulata. Inflorescentia laxa,

4-8 cm longa, interdum subcapitata; bracteae

reductae, superis calyce breviores; verticillastri 4-6

floribus; pedicelli 1 mm longi. Calyx tubuloso-

campanulatus, 7-8 mm longus, dense glanduloso-

hispidus; dentes lineari-lanceolati, acuminati, 2,5-3

mm longi, patentes. Corolla alba vel lilacina, subtiliter

glanduloso-pubescens; tubus 8-9 mm longus, 2 mm

diam.; labium posticum horizontale, oblongum,

4-5 mm longum, concavum; labium posticum

deflexum, 6-7 mm longum, lobo medio subrotundo,

4 mm longo, lobis lateralibus 2 mm longis. Stamina

4, 4 mm fauce exserta; filamenta puberula. Stylus

bifidus, 4 mm exsertus.

Type. — Transvaal, Mariepskop, Werdermann &

Oberdieck 1868 (PRE, holo. !).

Perennial aromatic herb; stems decumbent to

procumbent, branched, up to 60 cm long with the

inflorescences ascending, 4-angled, fairly densely

retrorse pilose. Leaves petiolate; petiole 5-12 mm

long; blade ovate-deltoid to broadly ovate or

subreniform, 1-2,5 x 0,8-2, 2 cm, fairly firm textured,

usually discolorous, upper surface hispid, lower

surface reticulate-veined, densely glandular-tomentose ;

base broadly cordate, apex obtuse to subacute;

margin crenulate. Inflorescence lax, 4-8 cm long,

occasionally subcapitate; bracts reduced, the upper

ones shorter than the calyx; verticils 4-6-flowered;

pedicels 1 mm long. Calyx tubular-campanulate,

7-8 mm long, densely glandular-hispid; teeth linear-

lanceolate, acuminate, 2,5-3 mm long, spreading.

Corolla white to pale mauve, finely glandular-

pubescent; tube 8-9 mm long, 2 mm in diameter;

upper lip horizontal, oblong, 4-5 mm long, concave;

lower lip deflexed, 6-7 mm long median lobe

subrotund, 4 mm long, lateral lobes 2 mm long.

Stamens 4, exserted by 4 mm; filaments puberulous.

Style bifid, exserted by 4 mm. Fig. 2.

Found among rocks in exposed situations in

mountain grassland in the Pilgrims Rest District of

the eastern Transvaal at altitudes of 1 500 to 2 200 m.

Transvaal. — 2430 (Pilgrims Rest): Mariepskop (-DB),

Meeuse 9963; Killick & Strey 2391 ; Werdermann & Oberdieck

1868; Bos 1028; Van der Schijff 4456 ; 4823; near Pilgrims Rest,

Black Hill, Ga/pin 14342; W. of Pilgrims Rest, Meeuse 10047;

God’s Window, 10 km N. of Pilgrims Rest, Davidson & Mogg

33093.

Although confused in the past with S. aethiopica,

5. reticulata may be distinguished by the usually

discolorous leaves, reticulate and densely tomentose

below, and by the spreading calyx teeth, The

tomentum on the stems is also characteristic being

fairly densely appressed pilose. The stem pubescence

of S. aethipica varies from retrorse to antrorse or

spreading, but is usually short and hispid.

From S. natalensis it differs in the 4-6-flowered

verticils, the reticulate venation on the undersides of

the leaves and in the spreading calyx teeth, while

from S. rehmannii it can readily be separated by the

simple, not stellate, pubescence.

It was first collected on the exposed summits of

hills near Pilgrims Rest by Galpin in 1937 and

appears to be restricted to the mountains of the

Pilgrims Rest District, where it usually grows among

rocks of the Black Reef quartzite formation.

5. Stachys malacophylla Skan in Kew Bull.

1909: 421 (1909); FI. Cap. 5, 1 : 345 (1910). Lectotype:

Cape, Queenstown, Galpin 1955 (K, lecto.; PRE!).

Fig. 2. — Stachys reticulata , Mariepskop, Transvaal ( Werder-

mann & Oberdieck 1868 in PRE, holotype), X F

Stem fairly densely spreading-pilose. Leaves

petiolate, blade broadly ovate, 1, 4-3 x 1-2,2 cm,

densely appressed pilose above, matted velvety pilose

below. Bracts differentiated from the leaves. Verticils

usually 4-6-flowered but sometimes 2- or 3-flowered.

Calyx densely hispid-villous.

A little-known species from mountains in the

eastern Cape Province.

Cape. — 3126 (Queenstown); near Queenstown (-DD), Galpin

1955; 5891. 3227 (Stutterheim): King William’s Town (-CD),

Sim 19590.

Closely related to S. sessilifolia from which it is

separated by the petiolate leaves and usually 4-6-

flowered verticils, while the tomentum on the stems

and leaves is less densely woolly. Both species are

very poorly known and further collecting is required

in order to establish how reliable these distinctions

are. See also notes under S. reticulata from the

eastern Transvaal.

6. Stachys sessilifolia E. Mey. ex Benth. in E.

Mey., Comm. 239 (1837); Drege, Zwei Doc. 151

(1843); Benth. in DC., Prodr. 12: 476 (1848); Skan

in FI. Cap. 5,1: 345 (1910) Type: Cape, between

Umzimvubu and Umsikaba Rivers, Drege 4752 (K,

holo. !).

S. bachmannii Guerke in Bot. Jahrb. 26: 75 (1898). Type:

Cape, Pondoland, Bachmann 1169.

Stem densely and softly villous. Leaves shortly

petiolate below, sessile above; blade ovate to ovate-

deltoid or narrowly ovate, 1-2,2 x 0,7-1 , 1 cm,

L. E. CODD

185

densely appressed villous above, densely matted-

villous below. Bracts leaf-like, becoming smaller

towards the apex. Verticils usually 2-flowered (but

up to 6-flowered according to FI. Cap.). Calyx

densely shaggy-villous.

Found in dense grassland in the eastern Cape

Province.

Cape. — 3029 (Kokstad): Mt. Currie (-AD), Tyson 1331. 3129

(Port St Johns): between Umzimvubu and Umsikaba Rivers

(-BD ?), Drege 4752 (K).

A little-known species closely related to S. natalensis

and which may be only an extreme form of that

species, distinguished by the densely matted-villous

lower surface of the leaves. Further material is

required in order to determine how meaningful this

distinction is.

A specimen from northern Natal, Grid 2730

(Vryheid): near Luneburg (-BC), Galpin 9870, has

this type of tomentum and 2-flowered verticils but

differs in having petioles up to 7 mm long. It is

tentatively included in S. sessilifolia until more

material is forthcoming in order to assess the

importance of subsessile versus petiolate leaves.

It may be noted that this character varies a good

deal in the material included in S. natalensis below.

The type of S. bachmannii has not been seen; the

species was included in S. sessilifolia by Skan and,

judging by its description, this decision appears to

be correct.

7. Stachys natalensis Hochst. in Flora 28:65

(1845). Type: Natal, Table Mt., Krauss 1139.

Stems variously pubescent from densely to sparingly

villous or shortly scabrid to softly pubescent. Leaves

subsessile or petiolate; blade ovate to ovate-deltoid,

variable in size, 1-4x0, 6-2,4 cm shortly and

sparingly appressed pubescent to densely strigose

above, less dense, more spreading and mainly on the

nerves below often with some glandular hairs but not

conspicuously glandular. Bracts differentiated towards

the apex. Verticils 2-flowered. Calyx densely villous

to densely and shortly pubescent, usually without

glands.

A widespread and variable species found in grass

on stony hillsides, in semi-shady kloofs and wooded

places in the mountains of northern central and

eastern Transvaal, Swaziland, northern and coastal

Natal as far south as Durban, with an occasional

record from the eastern Cape. Also recorded from

Rhodesia.

No material of the type, Krauss 1139 from Table

Mt., near Pietermaritzburg, has been traced, but the

description is considered adequate to identify it

with the present concept, though Skan, l.c., lists it

as an “imperfectly known species.” The specimen

Schlechter 2894 from near Verulam, Natal, may be

regarded as representative and, if no material of

the type can be found, would be a suitable specimen

to select as a lectotype.

Two varieties, based mainly on the degree of

pubescence, are recognized.

Key to varieties

Pubescence on stem, leaves and calyx sparingly villous

to shortly scabrid or tomentose (a) var. natalensis

Pubescence on stem, leaves and calyx densely

villous (b) var. galpinii

(a) var. natalensis.

Stachys natalensis Hochst. in Flora 65 (1845); Skan in FI.

Cap. 5,1: 367 (1910); Ross, FI. Natal 303 (1972). Type: Natal,

Table Mt., Krauss 1139. S. transvaalensis Guerke in Bot. Jahrb.

28: 316 (1901); Skan, l.c. 346 (1910). Type: Transvaal, Lyden-

burg District, Wilms 1136 (BM). S. leptoclacla Briq. in Bull

Herb. Boiss. ser. 2,3: 1084 (1903); Skan, l.c. 351 (1910); Ross

FI. Natal 303 (1972). Type: Natal, Bluekrantz River, Schlechter

6865 (Z, holo.! ;BOL! ). S. aethiopica sensu Letty, Wild Flow.

Transv. 284, t. 141: 3 (1962); sensu Compton, FI. Swaz. 66

(1966).

Stems shortly scabrid or tomentose to sparingly or

fairly densely villous. Leaves petiolate, blade hispid-

villous above, less so and mainly on the nerves below.

Calyx hispid-villous.

Distribution and ecology more or less as for the

species but not prevalent in the mountain grassland

of the eastern Transvaal. Also occurs in Rhodesia.

Transvaal. — 2329 (Pietersburg): Louis Trichardt (-BB),

Breyer sub TRY 20926; Spelonken (-BD), Junod 123. 2330

(Tzaneen): Tshakoma (-AB), Obermeyer sub TRY 31569. 2428

(Nylstroom): Geelhoutkop (-AD), Breyer sub TRY 18105;

Sterkrivierdam Nature Reserve (-BC), Jacobsen 2004; Moord-

drift (-BD), Leendertz sub TR Y 7330; N. of Warmbaths (-CD),

Repton 792; Sidey 1339; Smuts & Gillett 3089; Strey 3947;

Naboomfontein (-DA?), Galpin 13481. 2430 (Pilgrim’s Rest);

Ohrigstad Dam Nature Reserve (-DC), Jacobsen 1436; 1766;

2909. 2527 (Rustenburg): Scheerpoort (-DD), Leendertz sub

TRY 27126; farm Uitkomst 499 (-DD), Coetzee 654. 2528

(Pretoria): Wonderboom Reserve (-CA), Pole Evans 340;

C. A. Smith 104; 240; Koedoespoort (-CB), Leendertz sub TR Y

8570; Obermeyer sub TRY 27657; Botanic Garden (-CB),

Codd 855 ; 35 km E. of Pretoria (-CD), Repton 582; Trigaarts-

poort (-DB), Bruce 97 ; 18 km N.E. of Bronkhorstspruit (-DB),

Codd 2694. 2529 (Witbank): Bundu Inn (-AC), Mauve 4916;

Loskop Dam (-AD), Theron 1376. 2530 (Lydenburg); Crocodile

Valley (-AD), Burtt Davy 7662; Nelspruit (-BD), Liebenberg

2596; Cythna Letty Nature Reserve (-DD), Muller 2219. 2531

(Komatipoort): Barberton (-CC), Edwards 46.

Swaziland. — 2631 (Mbabane): Komati Bridge (-AA),

Compton 26835; 26931; near Mbabane (-AC), Hardy 27;

Hlatikula (-CD), Compton 26265 ; 30228.

Natal. — 2730 (Vryheid): Utrecht (CB), Wahl sub TRY 15536.

2732 (Ubombo): Pongola Poort (-AC), Repton 6020. Mpangazi

(-DA), Strey 5090; Mbazwane (-DA), Yahrmeijer 1133. 2829

(Harrismithj, 34 km N. of Ladysmith (-BB), Codd 8622; Blou-

krans River (-DD), Schlechter 6865 (Z). 2830 (Dundee): Tugela

Ferry (-CD), Galpin 14780; Kranskop (-DD), Galpin 14764.

2832 (Mtubatuba): Hluhluwe Game Reserve (-AA), Hirchins

366; 526; 643. 2929 (Underberg): Estcourt (-BB), West 1332;

Pentz 497. 2930 (Pietermaritzburg): near Durban (-DD),

Medley Wood 59; 6437. 2931 (Stanger): near Verulam (-CA),

Schlechter 2894; Moll 1990.

Cape. — 3228 (Butterworth): Kentani (-CB), Pegler 187.

In general the pubescence is sparingly to fairly

densely villous and grades somewhat into the very

densely villous pubescence of var. galpinii. In Natal

some specimens have hispid or shortly tomentose

pubescence, resembling the condition often found in

S. aethiopica, but may be separated from that species

by the 2-flowered verticils. The type of S. leptoclada

Briq., Schlechter 6865 (Z) from Bloukrans River near

Ladysmith is such a plant. Schlechter 6865 may be

a mixed gathering because the specimen in PRE

with this number is S. grandifolia E. Mey. ex Benth.

or, more likely, the material may have become mixed

when it was distributed. The type of S. transvaalensis

Guerke is more densely pubescent than typical var.

natalensis.

(b) var. galpinii (Briq.) Codd, stat. nov.

S. galpinii Briq. in Bull. Herb. Boiss. ser. 2,3: 1082 (1903);

Skan in FI. Cap. 5,1: 346 (1910); Compton, FI. Swaz. 66 (1966);

Ross, FI. Natal 303 (1972). Type: Transvaal, near Barberton,

Galpin 681 (K!, PRE!, SAM!). S.lupulina Briq., l.c. 1082 (1903).

Type: “Natal, near Claremont, Schlechter 4651” (Z, holo.!;

BOL!). 5. parilis N.E. Br. in Kew Bull. 1901 : 131 (1901); Skan,

l.c. 347 (1910); Ross. l.c. 303 (1972). Type: Natal, Drakensberg,

Tiger Cave Valley, Evans 387 (K, holo.; NH!). S. villosissima

H. M. Forbes in Bothalia 4: 38 (1941); Ross, l.c. 304 (1972).

Type: Natal, Entumeni, Forbes 783 (NH, holo.; PRE!).

Stems erect, branching from the base, 12-20 cm

tall or decumbent to straggling 30-40 cm long; stem

densely villous. Leaves subsessile or petiolate, densely

villous on both surfaces. Calyx densely villous.

186

A NOTE ON THE STACHYS AETHIOPICA COMPLEX

Found in dense grass, often among rocks at high

and medium altitudes in central and eastern Transvaal

and northern Swaziland, extending to northern Natal.

Transvaal. — 2428 (Nylstroom): Moorddrift (-BD), Leen-

dertz sub TRV 7330. 2528 (Pretoria): Wonderboompoort (-CA),

Mogg 15860; near Zoo (-CA), C. A. Smith 3272; Hennops River

(-CC), Repton 1115. 2529 (Witbank): Middelburg (-CD),

Young A 20. 2530 (Lydenburg): 27 km W. of Lydenburg (-AB),

Codd 8063; Dullstroom (-AC), Galpin 13051; near Belfast

(-CA), Leendertz sub TRV 9178; Leistner 510; Machadodorp

(-CB), Galpin 13109; 18 km S.E. of Sewefontein (-CD), Codd

8114. 2531 (Komatipoort): near Barberton (-CC), Galpin 681;

Thorncroft sub TRV 4337; Mauve 4424. 2630 (Carolina): near

Carolina (-AA), Bolus 12241.

Swaziland. — 2531 (Komatipoort): near Havelock Mine

(-CC), Miller 3033. 2631 (Mbabane): near Mbabane (-AC),

Codd 9511; Compton 24559; 25270; 25379; 26707; 26754.

Natal. — 2731 (Louwsburg): Pongola Experimental Farm

(-BC), Nel 199. 2829 (Harrismith): Willowford Station (-DD),

Acocks 10599. 2831 (Nkandla): 6 km S. of Mtonjaneni (-AD),

Codd 1808; near Melmoth (-CB), Mogg 4549 ; 6163; Entumeni

(-CD), Forbes 783. 2832 (Mtubatuba): Dukuduku (-AD),

Strey 5511; Richard’s Bay (-CC), Venter 4975.

This variety can be recognized by the combination

of densely villous pubescence and 2-flowered verticils;

it grades into var. natalensis and many specimens

may be regarded as transitional between the two.

In var. galpirtii the leaves may be petiolate or sub-

sessile and the latter specimens come near to S.

sessilifolia, in which the lower surfaces of the leaves

are densely matted-villous and the stems are softly

tomentose. In S. malacophylla the lower surfaces of

the leaves are also densely pubescent, but the verticils

are usually 6-flowered. The types of S. lupulina Briq.,

S. parilis N.E. Br. and S. villosissima H. M. Forbes

are not appreciably distinct from that of S. galpinii.

Skan, l.c., draws attention to the confusion

concerning the type of S. lupulina. Briquet cites the

specimen as “Natal, Claremontplats prope Claremont,

Schlechter 4651, ann. 1892.” On the type sheet in Z

and an isotype in BOL the label reads: “Claremont

flats prope Cape Town, Schlechter 465, 9. III.

1892”. It is undoubtedly conspecific with S. natalensis

var. galpinii, which does not occur in the Cape, but

could have been collected in northern Natal or the

eastern Transvaal. Skan concluded that it had

probably been introduced at the Cape but there is

no evidence to support this.

8. Stachys aethiopica L., Mant. 1: 82 (1767);

Burm. f., FI. Cap. Prodr. 16 (1768); Ait., Hort. Kew.

2: 302 (1879); Willd., Sp. PI. 3: 102 (1800); Thunb.,

FI. Cap. ed. Schult. 447 (1823); Benth., Lab. 548

(1834); in E. Mey., Comm. 239 (1837); in DC.,

Prodr. 12: 476 (1848); Bol. & Wolley-Dod in Trans.

S. Afr. Phil. Soc. 14: 310 (1904): Skan in FI. Cap.

5,1: 348 (1910); Marloth, FI. S. Afr. 3,2: 180, t.

47B (1932); Salter in FI. Cape Penins. 697 (1950);

Jacot-Guill., FI. Lesotho 237 (1971); Ross, FI.

Natal 303 (1972). Type: Cape Province, LINN 736.

13 (lecto.).

Betonica capensis Burm. f., FI. Cap. Prodr. 16 (1768). Type.

Pluk., Almagest. Bot. t. 315, f. 3 (1696).

S. aethiopica var. grandiflora Burch, ex Benth. in E. Mey.,

Comm. 239 (1837). Type: Cape, Klein Winterhoek, Drege 75d

(K, holo.). — var. hispidissima Benth., l.c. 239 (1837); Skan in

FI. Cap. 5,1 : 348 (1910). Type: Cape, Hex River Kloof, Drege

75h (K, holo.). — var. gtandulifera Skan, l.c. 348 (1910); Phill. in

Ann. S.A. Mus. 16:244 (1917); Jacot-Guill., FI. Lesotho 237

(1971). Syntypes: several, inch Zwartkei River, Baur s.n. (K;

PRE!). — var. parviflora Skan, l.c. 348 (1910); Salter in FI. Cape

Penins. 697 (1950). Syntypes: several, inch Cape Peninsula,

Signal Station, Wolley-Dod 3048 (K; BOL!). S. pulchella

Salisb., Prodr. 83 (1796), nom. illegit. Type: based on S.

aethiopica L. S. serru/ata Burch, ex Benth., Lab. 549 (1834);

in DC., Prodr. 12: 477 (1848); Skan, he. 350 (1910). Type:

Cape, near Knysna, Burchell 5155 (K, holo.!). S. capensis

Presl, Bot. Bemerk. 100 (1844); Skan, l.c. 366 (1910). Type:

Cape, without locality, Krebs 273. S. hispidula Hochst. in Flora

66 (1845); Skan, l.c. 367 (1910). Type: Cape, Humansdorp

District, Krauss 1125. 5. fruticetorum Briq. in Bull. Herb Boiss

ser. 2,3: 1083 (1903); Skan, l.c. 351 (1910). Type: Cape Sir

Lowry’s Pass, Schlechter 1179 (Z, holo.!; BOL!). S', harveyi

Skan, l.c. 350 (1910). Type: Cape, near Cape Town, Harvey

s.n. (TCD, holo.). S. attenuata Skan, l.c. 351 (1910). Syntypes:

Cape, near Bainskloof, Bolus 2896 (K, BOL!), Paarl Mt.

Drege 75b (K).

Stems decumbent or ascending, up to 50 cm long

or more, rather slender, sparingly to freely branched,

variously pubescent with short antrorse hairs, short

or long retrorse hairs often mixed with longish

spreading hairs or glandular hairs, or (especially the

rhachis) densely glandular-tomentulose. Leaves

petiolate; blade broadly ovate to ovate-deltoid,

usually about 0,8-3, 5x0, 6-2, 5 cm, but up to

6x4 cm (Natal coastal form), upper surface usually

shortly hispid to pilose, lower surface hispid to

pilose, mainly on the nerves, often with some glands,

to densely glandular-puberulous, rarely almost

glabrous. Inflorescence usually tapering, of many

whorls, occasionally with few whorls and subcapitate.

Bracts differentiated from the leaves towards the

apex. Verticils usually 4-6-flowered occasionally

2-3-flowered (mainly depauperate specimens). Calyx

sparingly to densely hispid and often glandular.

Distributed from Clanwilliam District to the

Peninsula and along the coast to the eastern Cape

Province and to Natal, Lesotho and the Orange Free

State; found in a variety of habitats from fynbos to

dry woodland and coastal dune bush, extending to

mountain grassland where it is usually found among

sandstone rocks.

O.F.S. — 2827 (Senekal): near Senekal (-BC), Goossens 918;

near Ficksburg (-DD), Potts 3184; Galpin 13831. 2828 (Beth-

lehem): near Bethlehem (-AB), Flanagan 2114; Phillips s.n.;

Scheepers 1412; Werger 244; Slabberts (-AC), Stam 179;

Fouriesburg (-CA), Potts 3265; Clarens (CB), Van Hoepen sub

TRV 18163. 2926 (Bloemfontein): Thaba Nchu (-BB), Roberts

2246.

Natal. — 2830 (Dundee): 14 km S.W. of Nqutu (-BA), Codd

2400; Kranskop (-DD), Strey 4288 (NH). 2831 (Nkandla):

Babanango (-AC), King 255; Melmoth (-CB), Mogg 6167.

2929 (Underberg): Tabamhlope Pasture Research Station (-BA),

West 550; near Sani Pass (-CB), Marias 1434. 2930 (Pieter-

maritzburg) : Tweedie (-AC), Mogg 6764; Dargle (-CA), Hilliard

and Burtt 3194; Pietermaritzburg (-CB), Goossens 131; Mogg

2235; Byrne (-CC), Galpin 11929; Durban (-DD), Strey 4347;

11318; Isipingo (-DD), Ward 3775. 2931 (Stanger): Mt. Edge-

combe (-CA), Medley Wood 1126. 3030 (Port Shepstone):

Dumisa (-AD), Huntley 166 (NH); Port Shepstone (-CB),

Dimock-Brown 475 (NH).

Lesotho. — 2828 (Bethlehem): Leribe (-CC), Dieterlen 101;

Maluti Mts. (-DD), Staples 85. 2927 (Maseru): Teyateyaneng

(-BA), Collett 477; Mamathes (-BB), Jacot-Guillarmod 866;

869; Tebetebeng River, Jacot-Guillarmod 359; Roma (-BC),

Ruch 1689; Mafeteng (-CC), Gerstner 226. 2929 (Underberg):

Sehlabathebe (-CC), Jacot-Guillarmod, Getliffe and Mzamane

154.

Cape. — 3027 (Lady Grey): Barkly East (-DC), Gerstner 206.

3028 (Matatiele): 6 km E. of Kenegha (-BC), Story 524. 3029

(Kokstad); 19 km N. of Swartberg(-AB), Codd 8533; near Kok-

stad (-CB), Coleman 467; Bizana (-DD), Acocks 12225. 3127

(Lady Frere): Cala (-DA), Pegler 1646. 3129 (Port St. Johns):

Port St. Johns (-DA), Howlett 10. 3219 (Wuppertal): Cedarberg

(-AC), Galpin 10546; Citrusdal (-CA), Hanekom 1164. 3224

(Graaff Reinet): Graaff Reinet (-BC), Sister Francis 37. 3225

(Somerset East): National Mountain Zebra Park (-AB),

Liebenberg 6431 ; Boschberg (-DA), MacOwan 559 partly. 3226

(Fort Beaufort); Mungo Mt. (-CB), Galpin 11550; 16 km N. of

Fort Beaufort (-DC), Story 2224; Garfield (-DD), Acocks

23513. 3227 (Stutterheim) : Windvoelberg (-AC), Roberts 1798;

near Happy Valley (-AD), Johnson 1298 ; Kabaku Hills (-CB?),

Acocks 9271; King William’s Town (-CD), Sim 19593; near

Komga (-DB), Flanagan 414. 3228 (Butterworth): near Mazeppa

Bay (-BC), Hilner 422; Kentani (-CB), Pegler 231. 3318

(Cape Town): near Hopefield (-AB), Lettybl; Darling (-AD),

Bolus sub TR V4657; Cape Town (-CD), Young sub TR V 26635;

Mrs. Southey sub Galpin 7849; Lions Head (-CD), Ecklon

s.n.- Marloth 65; Camps Bay (-CD), Thode A143; Sea Point

(-CD), C. A. Smith 2889; Paarl Mt. (-DB), Kruger M 118;

near Stellenbosch (-DB), Parker 3895; Bos 135; Kerfoot 6081.

3319 (Worcester): near Gouda (-AC), Esterhuysen 18787;

L. E. CODD

187

Prospect Peak (-BC), Esterhuysen 15902 ; Gydouw (-AD),

Leipoldt s.n.; Veld Reserve (-CB), Olivier 35; French Hoek Pass

(-CC), Boucher 2343; near McGregor (-DD), Van Breda &

Joubert 1973. 3322 (Oudtshoorn) : between Wilderness & Knysna

(-DC), Wells 3723. 3323 (Willowmore): near Joubertina (-DD),

Van Breda 1168. 3325 (Port Elizabeth): near Kenkelbosch

(-DB), Archibald 5995; Paterson 55; Port Elizabeth (-DC),

Borle 15. 3326 (Grahamstown): Alicedale (-AC) Rogers 12016;

Port Alfred (-DB), Tyson s.n.; Rogers 16654; ' Kowie (-DB),

Britten 810. 3327 (Peddie): East London (-BB), Galpin 1879.

3418 (Simonstown) : Hout Bay Nek (-AB), Hutchinson 103;

Muizenberg (-AB), Pillans 3678. 3419 (Caledon): Enon (-AB),

Thode A2741. 3420 (Bredasdorp): National Bontebok Park

(-AB), Liebenberg 6431 ; Kompanjies River (-AB), Marsh 581 ;

De Hoop (-AD), Van der Merwe 1158; Fort Beaufort (-BD),

Marsh 820; Nachtwacht (-CA), C. A. Smith 3053. 3423

(Knysna): Keurboomstrand (-AB), Codd3565.

Skan attempted to reduce the confusion which

existed in this complex by isolating certain discordant

elements as separate species, and by upholding four

varieties within 5. aethiopica. With the advent of a

good deal of modern material, the position is still

confused, but it appears that a broader concept of

the species should be adopted, while the separation of

varieties within S. aethiopica seems scarcely justified.

Several of the peripheral species upheld by Skan are

still maintained, such as S. rudatisii, S. gracilifiora

(-S. cooperi), S. scabrida ( = S. priorii ), S. flexuosa,

S. cymbalaria and S. sublobata, and the distinguishing

characters are discussed under the respective headings.

Others, such as S. harveyi, S. serrulata, S. attenuata

and S. fruticetorum are now included in S. aethiopica.

Regarding the typification of S. aethiopica, Linnaeus

refers to Pluk., Aim. 245, t. 315, f. 3, while the

description indicates that he had a specimen before

him. Of these two elements, the specimen in the

Linnaean Herbarium No. 736.13 may be regarded

as the type and is a plant with retrorse to spreading

pubescence and lacking glandular hairs. In both

elements the plant referred to has 2-flowered verticils,

but this must be regarded as exceptional as most

specimens which match the type have 2-6-flowered

verticils. The typical form occurs mainly in the

south-western Cape Province, but it grades into the

varieties upheld by Skan.

Bentham based his var. hispidissima on a Drege

gathering with denser and more stiffly spreading

hairs on stems and leaves, but this is scarcely distinct

from the typical form. The other two varieties,

described by Skan, are possibly more distinct but,

due to the many intermediates, it is preferred to

regard them as forms rather than give them taxonomic

rank.

In the plants placed by Skan in var. parviflora, the

hairs on the stem are antrorse and shortly scabrid,

the petioles tend to be shorter and the flowers smaller.

Salter in FI. Cape Penins. 697 (1950) upheld this

variety but, as one moves away from the Peninsula,

the distinction becomes less clear.

The form separated as var. glandulifera is usually

recognizable under magnification by the glandular-

puberulous pubescence on most parts of the plants,

especially the under-surface of the leaf, the rhachis

and the calyx. It appears to be the prevalent form

in the eastern, summer-rainfall part of South Africa,

especially along the Drakensberg escarpment.

Towards the west it grades into typical 5. aethiopica

and the type of S. fruticetorum, Schlechter 1179 from

Sir Lowry’s Pass, is an intermediate specimen.

Another form has been observed along the coast

of Natal in which the leaves are much larger than

usual, up to 6x4 cm. It appears to be a semi-weed

in disturbed areas, and serious consideration was

given to according it separate status of some sort.

However, there is a gradation in leaf size and many

intermediate specimens cannot be allocated with

certainty. At present it is regarded as a form growing

under favourable subtropical conditions. The

pubescence is somewhat similar to the plants discussed

under var. glandulifera.

To summarize, the pubescence in S. aethiopica is

mainly hispid, either antrorse, retrorse or spreading,

sometimes with scattered longish hairs and often

finely glandular, especially on the calyx and under-

surface of the leaves. This can usually be distinguished

from the villous pubescence of S. natalensis but there

should be no confusion between the two because,

in the eastern summer-rainfall areas, S. aethiopica

always has 4-6-flowered verticils, while S. natalensis

is strictly 2-flowered. The two overlap in Natal and

the extreme eastern Cape Province. S. aethiopica is

absent from the Transvaal and rare in Swaziland,

but a very glandular form occurs in Lesotho and

the eastern Orange Free State.

As it often happens in variable species, S. aethiopica

tends to be a repository for specimens which cannot

satisfactorily be placed elsewhere. For example, a

few specimens of rather distinct facies, with stiff

stems and very densely glandular-puberulous stems,

leaves and calyces, but with 2-flowered verticils,

have been collected in the eastern Cape Province,

namely: Stutterheim District, Acocks 2971 and Sim

s.n.; Port St. Johns District, Flanagan 2595. If

S. aethiopica were divided into formal varieties, these

specimens would probably be worthy of varietal

rank.

S. aethiopica normally has a slender tapering

inflorescence of several to many verticils. On the

other hand some specimens in the southern Cape

Province have few verticils with an almost subcapitate

inflorescence, resembling the condition found in

S. gracilifiora ( =S . cooperi ) and to some extent in

S', scabrida. The plants previously included in S.

cooperi usually have larger leaves, while those

described as S. scabrida have subglabrous to sparingly

retrorse-scabrid stems and subglabrous leaves. The

type of S. gracilifiora is somewhat intermediate but

it seems possible to maintain all three species, though

occasional specimens, especially from the Knysna-

Tsitsikama area, may be difficult to place with

certainty (see next species).

9. Stachys gracilifiora Presl, Bot. Bemerk. 100

(1844); Benth. in DC., Prodr. 12: 496 (1848); Skan

in FI. Cap. 5,1: 366 (1910) Type: Cape, without

locality, Krebs s.n. (PRC, holo.!, as to left-hand

specimen on sheet labelled S. gracilifiora Presl; PRE,

photo.).

5. cooperi Skan in Kew Bull. 1909: 420 (1909); FI. Cap. 5,1:

343 (1910); Ross, FI. Natal 303 (1972). Syntypes: Cape, Albany

Division, Cooper 15(K1); Kentani, Pegler 908, collected April

1909 (Kl, PRE!).

Stem shortly and often sparingly retrorse-

pubescent. Leaves petiolate, blade ovate to broadly

ovate, often thin-textured, 2-6, 5 x 1,4-5 cm,

eglandular, sparingly hispidulous or shortly and

sparingly pilose, the hairs on the upper leaf surface

soft and not bulbous-based. Inflorescence of 1-4

(rarely more) verticils, somewhat lax below or often

subcapitate, verticils 4-6-flowered; bracts differen-

tiated towards the apex. Calyx sparingly or softly

hispid, eglandular.

A soft straggling herb of moist places in forest

margins, in grass, fynbos or coastal scrub from

southern Natal to Knysna in the Cape.

188

A NOTE ON THE STACHYS AETHIOPICA COMPLEX

Natal. — 3030 (Port Shepstone): Horseshoe Farm (-CC),

Strev 6 169.

Cape. — 3129 (Port St. Johns): Umgazi River Mouth (-CB),

Wells 3494. 3226 (Fort Beaufort): Katberg (-BC), Galpin 2069 .

3227 (Stutterheim): Pirie Forest (-CC), Sim A; Acocks 9284;

near Komga (-DB), Flanagan 1752. 3228 (Butterworth) :

Kentani (-CB), Pegler 908. 3322 (Oudtshoorn): Wilderness

(-DC), Van Niekerk 236. 3325 (Port Elizabeth): Addo National

Park (-BC), Brvnard 363. 3423 (Knysna): Knysna: (-AA), Keet

635; Keurboomstrand (-AB), Fourcade 6119; Taylor 4914;

Storms River Mouth (-BB), Liebenberg 7765. 3424 (Humans-

dorp): Humansdorp (-BB), Galpin 4427.

There is a gradation in leaf size from the specimens

with larger and softer leaves, described as S. cooperi

Skan, to those occurring further west with smaller

and firmer leaves, which match the type of S. gracili-

flora Presl. The stem pubescence of the latter tends to

be slightly more scabrid and thus approaches the

condition found in S. scabrida Skan (see below).

However in S. scabrida the leaves are somewhat

thicker in texture, dry dark brown and the hairs on

the upper leaf surface are thicker with distinctly

swollen bases. On these grounds S. scabrida is kept

distinct, but the two overlap from southern Transkei

to Knysna and further study in this area is required.

As mentioned under the previous species, S.

aethiopica sometimes has few, fairly condensed

verticils which resemble those of S. graciliflora.

However, these plants usually have a coarser

pubescence on the stems, while gland-dots are often

present on the calyx and on the lower surface of the

leaves. No glandular hairs are found in S. graciliflora.

Specimens with large leaves resemble S. tubulosa

MacOwan, a species not included in the present

treatment, but may be separated, when flowers are

available, by the shorter corolla tube which does

not exceed 10 mm in length. In S. tubulosa the corolla

tube is 12-18 mm long and the species has a more

northerly distribution from East Griqualand to

Swaziland.

10. Stachys scabrida Skan in FI. Cap. 5,1: 349

(1910). Lectotype: Cape, Bruintjieshoogte, Burchell

3037 (K, lecto.!; PRE!).

S. priori i Skan, l.c. 353 (1910). Type: Cape, Algoa Bay,

Prior s.n. (K, holo. !).

Stems subglabrous to scabrid with few strong

retrorse-scabrid hairs to longer scattered multi-

cellular retrorse hairs. Leaves petiolate, firm-textured

and usually drying dark brown, ovate to ovate-

deltoid, 1, 2-3x0, 8-2 cm, subglabrous to sparingly

hispid, the upper surface with scattered short to

longish bulbous-based hairs. Inflorescence fairly

slender, of few verticils, spaced below; bracts

differentiated towards the apex. Calyx sparingly

hispid, eglandular.

A straggling herb in grass, fynbos or coastal scrub,

extending from the southern Transkei to Knysna and,

inland, to Steynsburg and Somerset East Districts.

Cape. — 3125 (Steynsburg): Suurberg (-BD), Schonland 3177.

3225 (Somerset East): Bruintjieshoogte (-CB), Burchell 3037.

3227 (Stutterheim): Hanover (-CD), Sim s.n. 3228 (Butter-

worth): “Gekau” River (-BC?), Drege s.n. ( K ). 3324 (Steytler-

ville): Assegaibosch (-CD), Breyer sub TRV 23323; 20 km W.

of Cambria (-DA), Story 2445. 3325 (Port Elizabeth): Van

Stadens Mts. (-CC), Zeyher 831 (K); Perseverance (-DC),

Long 816. 3326 (Grahamstown): Coldspring (-AD), Gane sub

TRV 25270; near Grahamstown (-BC), MacOwan 559, partly.

3423 (Knysna): (-AA), Breyer. sub TRV 23365.

S. scabrida appears to be related to S. graciliflora,

but the pubescence is coarser and more scabrid,

while the leaves are thicker-textured and the hairs

on the upper leaf surface tend to be bulbous-based.

The inflorescences tend to be more slender, rather

than subcapitate, as in S. graciliflora. The distinction

is by no means clear-cut, as indicated in the discussion

of the latter species, and specimens such as Story

2445 tend to be somewhat intermediate.

11. Stachys cymbalaria Briq. in Bull. Herb.

Boiss. ser. 2,3: 1088 (1903); Skan in FI. Cap 5,1 : 352

(1910); Ross, FI. Natal 303 (1972). Type: Cape,

Cradock, Cooper 416 (K, holo.!; W!).

S. aethiopica L. var. tenella Kuntze, Rev. Gen 3,2: 262 (1898).

Type: Cape, Cradock, Kuntze s.n. S. cymbalaria var. alba

Skan, l.c. 352 (1910). Type: Natal, Richmond, Medley Wood

1846 (K, holo.; NH!).

Stems slender, wiry, subglabrous or with a few

long slender spreading hairs or occasionally shortly

glandular-pubescent. Leaves petiolate, blade small,

broadly ovate-deltoid to suborbicular, 8-15x6-12

mm, subglabrous to thinly appressed-pubescent or

glandular-puberulous. Inflorescence lax of 1-5 verticils;

verticils 2-flowered; bracts differentiated, the upper

shorter than the calyx. Calyx puberulous to

hispidulous.

Found among rocks in exposed mountain grassland,

at altitudes of 1 000-1 800 m; recorded from a few

disjunct localities in southern Natal and the eastern

Cape Province.

Natal. — 2930 (Pietermaritzburg): Zwartkop (-CB), Medley

Wood 11120; Richmond (-CD), Medley Wood 1846 (NH).

3029 (Kokstad): Ngeli area (-DA), Hilliard & Burtt 5834 (NU);

7554 (NU); Strey 6337 (NH).

Cape. — 3039 (Kokstad): Mt. Insizwa (-CB), Schlechter 6467;

Hilliard & Burtt 6568. 3224 (GraafF Reinet): Reitvlei (-AB),

Galpin 10011; Lootsberg Pass (-BD), Acocks 13546. 3225

(Somerset East): Mountain Zebra National Park (-ADJ,

Brynard 295; Liebenberg 7107.

The species is closely related to S. aethiopica but is

characterized by the many wiry stems radiating from

the apex of a slender taproot, the small ovate to

roundish leaves and the 2-flowered verticils.

12. Stachys sublobata Skan in FI. Cap. 5,1: 354

(1910). Lectotype: Cape, Swellendam District, Barry-

dale, Galpin 4425 (K, lecto.; PRE!).

Stems slender, sparsely hispidulous, occasionally

with some longish spreading hairs. Leaves petiolate,

blade small, narrowly triangular, 10-15x4-6 mm,

thickish in texture, sparingly hispidulous on both sur-

faces, sometimes with gland-dots below, margin deep-

ly crenate or almost lobed with 3-5 lobe-like teeth.

Inflorescence lax of 2-5 verticils; verticils 2-(or

rarely 6-) flowered; bracts leaf-like below, differen-

tiated towards the apex. Calyx thinly glandular-

hispid.

Found on hillsides in fynbos at altitudes of 300-

900 m in the south-western Cape Province; known

from a few localities between Caledon and Mossel

Bay and, inland, to Ladismith and Oudtshoorn

Districts.

Cape. — 3320 (Montagu): Barrydale (-DC), Galpin 4425. 3321

(Ladismith): Rooiberg Pass (-DA), Acocks 20779. 3322

(Oudtshoorn): Cango Caves (-AC), Bolus 12244. 3419 (Cale-

don): 10 km W. of Rietpoel Station (-BD), Acocks 22582.

3421 (Riversdale): Klein Berg (-BB), Galpin 4426, partly.

This species resembles S. cymbalaria and, like that

species, is closely related to S. aethiopica, but can be

recognized by the narrowly triangular leaves, rather

thickish in texture and with deeply crenate, almost

lobed margins. It normally has 2-flowered verticils,

but one portion of Galpin 4426 in PRE, which

matches S. sublobata, has 6-flowered verticils. The

other specimen mounted on the same sheet has

larger leaves and is S. aethiopica.

L. E. CODD

189

13. Stachys flexuosa Skan in FI. Cap. 5,1: 352

(1910). Type: Cape, Stockenstroom District, old

Katberg Pass, Galpin 2393 (wrongly listed in FI.

Cap. as 2093) (K, holo.; PRE!, SAM!).

Stems fairly densely villous-pilose with long

spreading hairs and some short glandular hairs.

Leaves petiolate, blade ovate, 1-2x0, 6-1, 5 cm,

somewhat appressed-villous with some bulbous-based

hairs above, hispid mainly on the nerves below, base

not deeply cordate. Inflorescence fairly dense, of few

to several 2-6-flowered verticils; bracts leaf-like

below, becoming smaller towards the apex. Calyx

hispid, often with some glands.

Known from a few localities in the eastern Cape

Province, among rocks in mountain grassland.

Cape. — 3126 (Queenstown): near Hopewell (-DC?), Galpin

8379. 3129 (Port St. Johns): Port St. Johns District (CB?),

Swinny & Baker sub TRV 14143. 3226 (Fort Beaufort): old

Katberg Pass (-BC), Galpin 2393 ; Hogsback (-DB), Rattray 403.

The species is included here because the small

ovate leaves with a rather wide sinus at the base are

reminiscent of the S. aethiopica complex, but the

relationship appears to be with another complex

which includes S. obtusifolia and S. tysonii. From

these it differs mainly in the smaller leaves, and

less densely villous stems and leaves, in contrast to the

markedly glandular pubescence of S. tysonii. As in

many other species of Stachys, a good deal more

material is required before the status of S. flexuosa

can be reliably determined.

UITTREKSEL

Die Stachys aelhiopica-kompleks word ondersoek

en 'n sleutel voorsien tot die soorte wat erken word.

Die kombinasie S. natalensis Hochst. var. galpinii

( Briq .) Codd word gemaak en twee nuwe soorte, S.

reticulata Codd en S. arachnoidea Codd word beskrvf.

Bothalia 12,2: 191-194 (1977)

New taxa and a new combination in the genus Cotyledon

H. R. TOLKEN*

ABSTRACT

Five new species and one variety are described and the combination Cotyledon tomentosa Harv. subsp.

ladismithiensis (V. Poelln.) Toelken is effected. The following new names are published; C. sulphurea Toelken,

C. oceultans Toelken, C. similis Toelken, C. viridiflorum Toelken, C. hallii Toelken and C. pygmaea Barker

var. tenuis Toelken.

RESUME

NOUVEA U TAXA ET NOUVELLE COMBINAISON DANS LE GENRE COTYLEDON

On decrit cinq especes et une variete nouvelles. La combinaison Cotyledon tomentosa Harv.

subsp. ladismithiensis ( V. Poelln.) Toelken est operee. Les nouveaux noms publies sont: C. sulphurea

Toelken, C. oceultans Toelken, C. similis Toelken, C. viridiflorum Toelken, C. hallii Toelken et

C. pygmaea Barker var. tenuis Toelken.

Recent collecting has added more information to

the knowledge of some rare species of Cotyledon

from the north-western Cape Province, so that these

can now be described. This has also allowed a

re-evaluation of the delimitation of other species,

which were previously differently interpreted.

1. Cotyledon sulphurea Toelken, sp. nov. a

C. occultante foliis glabris, florescentia praecociore

et ovulis ovatis differt.

Herbae perennes tuberibus subterraneis ramosis

usque ad 10 cm longis, ramis succulentis laevibus

pauciramosis et rare 2 cm longioribus. Folia alterna,

oblanceolata vel spathulata, 0,5-2, 5 cm longa, 0,2-

0,5 cm lata, cuneata, plerunrque obtusa, dorsoven-

traliter, complanata, leviter sulcata supra et convexa

subtus, glabra sed cellulis pustuliformibus, succulenta,

viridia. Inflorescentia thyrsus corymbiformis 1-4

dichasiis patentibus, pedunculo rigido 1-3 (-5) cm

longo et tecto pilis glandulosis, pedicello 0,4-0, 8 cm

longo et tecto pilis glandulosis, bracteis lanceolatis

deciduis. Sepala triangulari-lanceolata, 0,35-0,45 cm

longa, acuta, tecta sparsim pilis glandulosis patentibus,

succulenta, flavo-virentia. Peta/a oblonga, 1,3-1, 5

cm longa, acuta, connata in tubo 0,8-1 cm longo,

tecta plus minusve etiam pilis glandulosis patentibus

extus et pilis aliquot tenuis intus praecipue tubi

petalorum in partibus inferis, sulphurea. Stamina

1-1,2 cm longa et aequilonga, antheris 0,1-0,12 cm

longis et filamentis rectis pilis patentibus praecipue

in partibus quibus connatis petalorum tubo. Squamae

oblongae 0 , 1-0, 1 3 x 0,02-0,04 cm, plus minusve emar-

ginatae, leviter succulentae, flavae. Carpella ovariis

gracilibus gradate constricta ad stylos erectos et

stigmatibus terminalibus; ovarium 10-14 ovulis ovatis

sine porcis verticalibus.

Type: Cape: near Pofadder, Tolken 3676A (BOL,

holo. !).

Perennial with branched underground tubers up to

10 cm long, with fleshy glabrous stems little branched

and rarely longer then 2 cm. Leaves alternate,

oblanceolate to spathulate, 0,5-2, 5 cm long, 0,2-

0,5 cm broad, cuneate, usually obtuse, dorsiventrally

compressed and slightly grooved above and convex

below, succulent, glabrous but with some large

bulging epidermal cells, green. Inflorescence a flat-

topped thyrse with one to several spreading mono-

chasia, with wiry peduncle 1-3 (-5) cm long and

covered with glandular hairs, with pedicels 0,4-

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag XI 01, Pretoria

0,8 cm long and covered with glandular hairs, with

lanceolate bracts deciduous. Sepals triangular-lanceo-

late, 0,35-0,45 cm long, acute, sparsely covered with

glandular hairs, succulent, yellowish-green. Petals

oblong 1,3-1, 5 cm long, acute, fused into a tube

0,8-1 cm long, more or less evenly covered with

spreading glandular hairs on the outside and with

few fine hairs on the inside mainly where the filaments

are fused to the petal-tube, sulphur-yellow. Stamens

1-1,3 cm long and all of equal length, with anthers

0,1-0,12 cm long, with straight filaments, with

scattered hairs mainly where fused to the petal-tube.

Squama oblong 0,1-0,13x0,02-0,04 cm, more or

less emarginate, slightly succulent, yellow. Carpels

with slender ovaries gradually constricted into erect

styles with terminal stigmas; ovary with 24-40 ovate

ovules without distinct vertical ridges.

This species is usually found on quartzite gravel

on the lower slopes of hills and has been recorded

only from the vicinity of Pofadder, but here it is

locally common. It flowers during November and

December, a distinguishing feature between this

species and C. oceultans.

Cape. — 2919 (Pofadder): 4 km N.E. of Pofadder (-AB),

Horn 1011 (NBG); Horn sub NBG 873/61 (NBG); near Pofadder

(-AB), Tolken 3676 A (BOL); Wisura 1961 (NBG); 1 1 km S.E. of

Pofadder (-AB), Tolken 3679A (BOL).

2. Cotyledon oceultans Toelken, sp. nov. a C.

sulphurea foliis paene orbicularibus quibus tectis pilis

glandulosis et florescentia serotinioribus ; a C. pygmaea

ramo brevissimo et trichotomatibus unicellularibus.

Herbae perennes tuberibus subterraneis paucira-

mosis usque ad 6 cm longae, ramis laevibus paucira-

mosis et usque ad 1 cm longis. Folia alterna, spathulata

vel orbicularia, 0,4— 1,2 cm longa, 0,3-1, 4 cm lata,

cuneata vel paene petiolata, obtusa, dorsiventraliter

complanata et leviter sulcata supra, tecta dense tri-

chotomatibus erectis obtusis, succulenta, atrovirentia.

Inflorescentia thyrsus 1 (-3) monochasiis 1-3 floribus,

pedunculo rigido 0,8-1, 5 cm longo et tecto pilis

glandulosis, pedicello 0,5-1, 3 cm longo et tecto

pilis glandulosis, bracteis lanceolatis caducis. Sepala

triangulari-lanceolata, 0,2-0, 3 cm longa, acuta, tecta

pilis glandulosis, succulenta, viridia. Petala oblonga,

1-1,2 cm longa, acuta, connata in tubo 0,6-0, 8 cm

longo, plus minusve etiam tecta pilis glandulosis

extus et pilis aliquot tenuibus intus praecipue in

partibus inferis tubi petalorum. Stamina 1,1-1, 4 cm

longa et aequilonga, antheris 0,1-0,12 cm longis,

filamentis rectis pilis tenuibus sparsis praecipue in

partibus quibus connatis petalorum tubo. Squamae

oblongae, 0,11-0,13x0,02-0,04 cm, plerumque plus

192

NEW TAX A AND A NEW COMBINATION IN THE GENUS COTYLEDON

minusve emarginatae, succulentae, flavae. Carpella

ovariis gracilibus gradate constrictis ad stylos erectos

stigmatibus terminalibus; ovarium 10-14 ovulis

ovatis sed lateriobus ad lateres et porcis verticalibus.

Type: Cape, near Bitterfontein, H. Hall 4289

(PRE, holo. !; NBG).

Perennial with little branched underground tubers

up to 6 cm long, with short glabrous stems little

branched and up to 1 cm long. Leaves alternate,

spathulate or orbicular, 0,4-1 ,2 cm long, 0,3-1 ,4 cm

broad, cuneate or almost petiolate, obtuse, dorsi-

ventrally compressed and slightly grooved above,

densely covered with blunt erect trichomes, deep

green. Inflorescence a thyrse with 1 (-3) monochasia,

each with 1-3 flowers, with wiry peduncle 0,8-1 ,5 cm

long and covered with glandular hairs, with pedicels

0,5-1, 3 cm long and covered with glandular hairs,

with lanceolate deciduous bracts. Sepals triangular-

lanceolate, 0,2-0, 3 cm long, acute, covered with

glandular hairs, succulent, green. Petals oblong,

1-1,2 cm long, acute, fused into a tube 0,6-0, 8 cm

long and more or less evenly covered with spreading

glandular hairs on the outside and with a few fine

hairs on the inside mainly where the filaments are

fused to the petal tube, pale yellow or yellowish-

green. Stamens 1,1-1, 4 cm long and about equally

long, with anthers 0,1-0,12 cm long, with straight

filaments with fine hairs mainly where fused to the

petal tube. Squamae oblong, 0,11-0,13x0,02-0,04

cm, usually more or less emarginate, fleshy, yellow.

Carpels with slender ovaries gradually constricted

into erect styles with terminal stigmas; ovary with

8-20 ovules almost dumbell-shaped and with distinct

vertical ridges.

So far this small Cotyledon has been recorded

from quartzite gravel near Bitterfontein and Steinkopf,

but the fact that these two localities are so far away

from each other suggests that the species is probably

widely distributed over the little known eastern

foothills of the Kamiesberg. Mr H. Hall, the very

observant collector, was the first to draw attention

to this small plant. In addition, the species flowers

in February and March, the driest period in the

Namaqualand, so that it is rarely seen in flower.

Cape. — 2917 (Springbok): 11 km S. of Steinkopf (-BD),

Tolken 4272 (BOL). 3118 (Vanrhynsdorp); near Bitterfontein

(-AB), H. Hall 3960 (PRE): 4289 (NBG, PRE).

3. Cotyledon pygmaea Barker in Flow. PI. S.

Afr. 10: t. 396 (1930). Type: Cape, Vanrhynsdorp,

Vigne sub NBG 2267/29 (K, holo!).

var. pygmaea.

Stems robust and usually (0,4-) 0,5-1 cm in

diameter, with pale yellow rarely greyish bark which

is smooth or peeling in older plants. Leaves densely

covered with heart-shaped trichomes. Squamae trans-

versely oblong to square, 0,08-0,1x0,1-0,13 cm.

Characteristic of C. pygmaea are the multicellular

trichomes which consist of two larger basal cells and

a smaller glandular cell at the apex. The basal cells

are bulging which gives them the heart-shaped

appearance at least on the leaves of var. pygmaea.

The petals are covered with trichomes with oblong

basal cells and the terminal cell is very small. The

latter type of trichomes are found on all parts of

the plants of var. tenuis.

Although extreme forms of the two varieties may

differ in several characters such as the diameter of

steins and their bark, a considerable range of variation

has been recorded and both varieties were found near

Koekenaap. The plant of the typical variety (Hall

sub NBG 587/54) from that area shows somewhat

intermediate characters, but its pollen shows no

abnormality to indicate that the two varieties

hybridize in this region.

var. tenuis Toelken, var. nov., a var. pygmaea

squamis oblongis differt.

Caules tenues 0,2-0, 3 (-0,4) cm in diametro,

cortice griseo plus minusve porcata et non chartaceo.

Folia trichotomibus oblongis sparsim tecta. Squamae

oblongae 0,09-0,12x0,04-0,06 cm.

Type: Cape, near Holrivier Station, H. Hall 3925

(PRE, holo.!; NBG!).

Stems slender 0,2-0, 3 (-0,4) cm in diameter, with

grey bark more or less ridged and not peeling. Leaves

sparsely covered with erect oblong trichomes.

Squamae oblong 0 , 09-0 ,12x0, 04-0 ,06 cm .

Cape. — 3118 (Vanrhynsdorp): near Holrivier Station (-AD),

Hall 3925 (NBG, PRE); 4273 (PRE); Koekenaap (-CB), Hall

2786 (NBG); Hall sub NBG 72/56 (NBG); Vanrhynsdorp (-DB);

Compton et al sub NBG 842a/48 (NBG).

4. Cotyledon similis Toelken , sp. nov. a C. pygmaea

foliis glabris et pilis longissimis in fauce tubi petalorum

differt.

Herbae perennes tuberibus subterraneis paucira-

mosis usque ad 5 cm longis, caulibus erectis rare remo-

sis et usque ad 12 cm longis et usque ad 0,4 cm in

diametro, griseis et porcis nigris verticalibus. Folia

alterna, elliptica vel oblonga, 0,5-1, 3 cm long, 0,2-

0,4 (-0,5) cm lata, acuta vel obtusa, teretia vel

leviter complanata, glabra sed cellulosis pustuliformi-

bus, viridia et saepe rufo-striata. Inflorescentia mono-

chasium 1 (-3) floribus, pedunculo glabro 0,8-3 cm

longo, pedicello glabro 0,4-0, 6 cm longo, bracteis

lanceolatis deciduis. Sepala triangulari-lanceolata,

0,15-0,25 cm longa, acuta, glabra, succulenta, viri-

dia. Petala oblonga, 0,8-1 cm longa, acuta, connata

in tubo 0,6-0, 8 cm longo, extus glabra et intus pilis

longis tenuis praecipue in fauce tubi petalorum,

pallide flavo-virentia vel eburnea. Stamina 0,7-1 cm

longa et aequilonga, antheris 0,08-0,1 cm longis,

filamentis rectis pilis patentibus praecipue in partibus

quibus connatis tubo petalorum. Squamae anguste

oblongae, 0,12-0,14x0,02-0,03 cm, plerumque

penitus emarginatae, leviter succulentae, croceae ad

flavam. Carpella ovariis gracilibus gradate constricta

ad stylos erectos et stigmatibus terminalibus; ovarium

1 0—14 ovulis ovatis sed latioribus ad lateres et porcis

verticalibus.

Type: Cape, N. of Grootmis, Wisura 1303

(NBG, holo!).

Perennial with branched underground tubers up to

5 cm long, each with an erect stem; stems rarely

branched, up to 12 cm long and up to 0,4 cm in

diameter, grey with black vertical ridges. Leaves

alternate, elliptic to oblong 0,5-1 ,3 cm long, 0,2-0, 4

cm broad, acute or obtuse, terete or almost so, gla-

brous but with irregularly bulging epidermal cells,

green, often with reddish-brown stripes. Inflorescence

a monochasium with 1 (-3) flowers, with glabrous

peduncle 0,8-3 cm long, with glabrous pedicels

0,4-0, 6 cm long, with deciduous lanceolate bracts.

Sepals triangular-lanceolate, 0,15-0,25 cm long,

acute, glabrous, succulent, green. Petals oblong,

0,8-1 cm long, acute, fused into a tube 0,6-0, 8

cm long, glabrous on the outside and with long

fine hairs in the throat of the petal tube, pale

yellowish-green to cream. Stamens 0,7-1 cm long and

all about the same length, with anthers 0,08-0,1 cm

long, with straight filaments with spreading hairs

mainly where fused to the petal tube. Squamae nar-

rowly oblong, 0,12-0,14x0,02-0,03 cm, usually

deeply emarginate, slightly fleshy, orange to yellow.

H. R. TOLKEN

193

Carpels with slender ovaries gradually constricted into

erect styles and terminal stigmas; ovary with 10-14

ovules almost dumbell-shaped and with vertical

ridges.

Plants of this species were found either in east-

facing crevices on exposed rock ledges or on quart-

zite gravel on hills mainly west of the mountains

between Komaggas and the Orange River.

The slender grey stems bear leaves only for a very

short period, so that this delicate plant is easily over-

looked. The slender stems are superficially similar to

those of C .pygmaea var. tenuis.

Cape. — 2816 (Oranjemund) : Numees mountain (-BD),

Tolken 3312 (BOL); Holgat River (-DB), Tolken 1857 (BOL).

2917 (Springbok): Augrabies (-AA), Hal! 1362A (NBG); N. of

Grootmis (-AC), Wisura 1303 (NBG); 48 km S.E. of Port

Nolloth (-AC), Acocks 14235 (PRE); 46 km E. of Port Nol-

loth (-AD), M. Schlechter sub SUG 6870 (BOL); Ratelpoort

(-BD), Littlewood sub (NBG) 434/69 (NBG); Komaggas (-CD),

Tolken 2026 (BOI).

5. Cotyledon viridiflorum Toelken, sp. nov. a C.

fragile petalis pubescentibus intra et extra et foliis

dorsoventraliter complanatis glabrescentibus differt.

Fruticuli basis tuberosis, ramis paucis succulentis

usque ad 35 cm longis, erectis, tectis pilis grandulosis

glabrescentibus. Folia alterna, oblanceolata vel ellip-

tica, 2-4 cm longa, 0,8-1 cm lata, cuneata, plerumque

apicibus obtusis, dorsiventraliter complanata et

leviter sulcata supra, tecta pilis glandulosis vel glabres-

centia, succulenta, viridia. Inflorescentia thyrsus spici-

formis (1-) 3-5 floribus, pedunclo 0,8-4 cm longo

et pedicello glabro 0,5-1 (-2,5) cm longo, bracteatis

squamosis. Sepala triangulari-lanceolata, 0, 5-0,6 (-1)

cm longa, acuta, tecta pilis glandulosis patentibus,

succulenta, viridia. Petala oblonga, 2, 4-2, 6 cm longa,

acuta, connata in tubo 1,8-2 cm longo, apicibus

recurvis, tecta pilis plus minusve glandulosis intra et

extra, leviter succulenta, viridia vel flavovirentia.

Stamina 1 ,4-1 ,9 cm longa, antheris nigris 0,4-0, 6 cm

longis, filamentis pilis patentibus et saepe flexibus infra

antheras. Squamae oblongae, 0,1-0,12x0,03-0,05

cm, leviter emarginatae, succulentae, flavae. Carpella

1 ,2-1 ,5 cm longa, ovariis gracilibus gradate constric-

tis ad stylos erectos stigmatibus terminalibus; ovarium

8-12 ovulis elongato-ovatis porcis verticalibus.

Type: Cape, near Modderfontein, Tolken 5327

(PRE, holo!).

Shrublet with tuberous bases, with few brittle fleshy

branches up to 35 cm long, erect, covered at first with

glandular hairs, later glabrescent. Leaves alternate,

oblanceolate to elliptic, 2-4 cm long, 0,4-1 cm broad,

each with cuneate base and obtuse apex, dorsiventrally

compressed and slightly grooved above the central

vein, covered with glandular hairs when young becom-

ing glabrous later. Inflorescence a spike-like thyrse

with (1-) 3-5 flowers, with peduncle 0,8-4 cm long and

pedicels 0,5-1 (-1,5) cm long, with scale-like bracts.

Sepals triangular-lanceolate, 0,5-0, 6 (-1) cm long,

acute covered with spreading glandular hairs, succu-

lent, green. Petals oblong 2, 4-2, 6 cm long, acute,

fused into a tube 1 ,8-2 cm long, with apices recurved

or recoiling, covered with recurved hairs, more or

less glandular on the outside and inside, slightly

fleshy, yellowish-green. Stamens 1,4-1, 9 cm long,

with black anthers 0,4-0, 6 cm long, filaments with

spreading hairs and often with right-angled outward

bend below anthers. Squamae oblong 0,1-0, 12 X

0,03-0,05 cm, slightly emarginate fleshy, yellow.

Carpels with slender ovaries gradually constricted into

erect styles with terminal stigmas; ovary with 8-12

ovules with vertical ridges. Like C. fragilis and C.

vetricosa, this species grows in the shade of overhang-

ing rocks and usually on the south aspect of moun-

tains.

Cape. — 2817 (Vioolsdrif): near Modderfontein (-CC), Tolken

5327 (PRE).

6. C. hallii Toelken , sp. nov. a C. pearsonio habi-

tato erecto-fastigiato et usque ad 40 cm alto, foliorum

cicatricibus paene indistinctis et floribus erectis differt.

C.pearsonii sensu Jacobsen, Hand, Succ. PI. 1 : 289, fig. 286

(1960); Sukk. Lex. 135, t. 40,5 (1970), non Schonl. C. racemosa

sensu Friedr. in Prodr. FI. S.W. Afr. 52: 10 (1968), non Harv.

Fruticuli perennes caulibus succulentis usque ad 5

cm in diametro et leviter ventricosis ad basim, multi-

ramosi et ramis erectis vel paene fastigiata. Folia

alterna, lineari-lanceolata, 0,8-3, 5 cm longa, 0,3-

0,5 cm lata, acuta, teretia vel interdum sulcata supra,

sparsim tecta pilis glandulosis sed glabrescentia,

viridia vel glaucescentia. Inflorescentia thyrsus \-4

monochasiis, pedunculo lignescente 1,5-3, 5 cm

longo, pedicello tenace 1-3 cm longo tecto pilis plan-

dulosis, bracteatis foliiformibus. Sepala triangularia,

0,4-0, 9 cm longa, acuta, tecta pilis glandulosis paten-

tibus, succulenta, viridia. Petala oblonga, 1,7-2 cm

longa, plerumque acutissima, connata in tubo 1-1,4

cm longo, apicibus recurvis sed erectis in fructu, tecta

pilis glandulosis patentibus extus et pilis sparsis tenui-

bus intus, leviter succulenta, viridia vel flavo-viridia.

Stamina 1 ,6-2 cm longa, antheris 0,3-0, 5 cm longis,

filamentis paene glabris et plerumque flexibus infra

antheras. Squamae oblongae 0,12-0,15x0,04—0,06

cm, emarginatissimae, succulentae ad basim, flavae.

Carpella ovariis gracilibus gradate constrictis ad

stylos erectos stigmatibus terminalibus; ovarium 30-36

ovulis ovatis sed latioribus ad lateres et papillosis

plerumque serialis.

Type: Cape, De Hoop, H. Hall 1300 (NBG, holo!).

Perennial shrublet with fleshy stems up to 5 cm in

diameter and slightly swollen towards the base, much

branched and branches erect and almost fastigiate.

Leaves alternate, linear-lanceolate, 0,8-3, 5 cm long,

0,3-0, 5 cm broad, acute, terete or sometimes with

a groove on the upper surface, sparsely covered with

glandular hairs but becoming glabrous, green or yel-

lowish-green. Inflorescence a thyrse with one to several

monochasia, with glandular rarely glabrescent pedun-

cle 1,5-3, 5 cm long and glandular pedicels 1-3 cm

long, with leaf-like bracts. Sepals triangular 0,6-0, 9

cm long, acute, covered with stout spreading glandular

hairs, fleshy, green. Petals oblong, 1,7-2 cm long,

usually very acute, fused into a tube 1-1,4 cm long,

with apices recurved but straightened out again when

fruiting, covered with glandular hairs along the out-

side and in particular the centre of each petal, with

scattered fine hairs or rarely papillose on the inside,

slightly fleshy, green to yellowish-green. Stamens

1,6-2 cm long, with anthers 0,3-0, 5 cm long, with

filaments almost glabrous and usually with a right-

angled outward bend below the anthers. Squamae

oblong, 0,12-0,15x0,04-0,06 cm, strongly emargi-

nate, succulent towards the base, yellow. Carpels with

slender ovaries gradually constricted into erect styles

and with terminal stigmas; ovary with 30-45 elongate

ovules which are usually dumbell-shaped and with

papillae often arranged in irregular rows.

This species occurs on rocky slopes in the mountains

on either side of the Orange River of Vioolsdrif.

S.W. A. — 2717 (Chamaites): near Aiais (-CD), Pillans 65 8

(BOL). 28 16 (Oranjemund) : Kahanstal (-BB), Dimer 8152 (BOL,

K, PRE, SAM); De Winter & Giess 6360 ( PRE).

Cape. — 2817 (Vioolsdrif): De Hoop (-AA), Hall /DO (NBG);

20 km E. of Eksteenfontein (-CD), Leistner 3373 (PRE); 32 km

S.W. of Vioolsdrif (-CD), Werger 405 (PRE).

194

NEW TAX A AND A NEW COMBINATION IN THE GENUS COTYLEDON

7. Cotyledon tomentosa Harv. in FI. Cap. 2: 373

(1862); Schonl. in Rec. Albany Mus. 3: 146 (1915).

Type: Cape, between Groot River and Trumpeter’s

Poort, Zeyher 1085 (S, holo!).

subsp. tomentosa.

C. tomentosa Harv. in FI. Cap. 2: 373 (1862). C. ladismithiensis

sensu Boom in Succulenta, Amst. (1958) 17; sensu Jacobsen,

Handb. Succ.PI. 1 : 284, fig. 278 (I960); Sukk. Lex. 134, T.39, 1

(1970), non V.Poelln.

Shrublets with slender branches, much branched.

Leaves oblanceolate 1,5-2, 5 (-3,5) cm long, 0,8-1, 4

(-1 ,7) cm broad, with 3-5 (-10) apical teeth, more or

less dorsiventrally compressed.

C. tomentosa is a little known species, which is poor-

ly represented in herbaria and at known localities one

finds only a few plants. However, the main obstacle

in interpreting the identity of C. tomentosa was that

Harvey did not mention in his description the three

distinct teeth at the apex of the leaves. The type locality

given by Zeyher was rather vague, but it must be

sought somewhere east of Steytlerville, which lies on

the Gamtoos River (alias Groot River). The plant was

recently collected in the mountains just north of

Steytlerville {Tolken 1778).

Then C. heterophylla Schoenl. non Roxb. was

described from the vicinity of Ladismith, but being a

later homomym it must be replaced by the name

C. ladismithiensis V.Poelln. The latter taxon is here

considered to be a separate subspecies of C. tomentosa,

which is distinguished from the typical subspecies by

its longer almost terete leaves as described below. The

number of teeth on the leaves varies greatly on the

same and different plants, and juvenile plants of both

subspecies seem to have leaves with 1-3 teeth. The

leaves of juvenile plants and of plants of the subsp.

tomentosa growing in deep shade are villose, while

those of the subsp. ladismithiensis are tomentose, but

no seedlings have been seen.

Subsp. tomentosa has been recorded from near

Calitzdorp. Willowmore and Steytlerville, where it

occurs on the slopes of sheltered kloofs. Plants of the

subsp. ladismithiensis occur on rock outcrops between

Laingsburg and the northern slopes of the Lange-

berg near Muiskraal.

subsp. ladismithiensis {V.Poelln.) Toelken, comb,

nov. et stat. nov.

C .ladismithiensis V. Poelln. in Jahrb. Deutsch.Kakt. Ges.

1: 94 (1936); in Fedde Repert. 42: 40 (1937). Type: Cape,

between Ladismith and Laingsburg, E.Pillans sub N.S. Pillans

968 (GRA, holo!; BOL). C. heterophylla Schoenl. in Rec.

Albany Mus. 2: 150 (1907), non Roxb. (1814); in Rec. Albany

Mus. 3: 146 (1915). Type: same as for C. ladismithiensis.

Shrublet with rigid spreading branches, little

branched. Leaves oblong-elliptic (3-) 3, 5-5, 5 (-8) cm

long, 0,8-1, 2 cm broad, with 1 (-3) apical teeth,

usually almost terete and only occasionally the upper

surface is compressed.

Cape. — 2033 (Montagu): Anysberg (-DB), Stokoe 8708

(BOL); western Touwsberg (-DB), Warts 1467 (NBG). 2133

(Ladismith): Ladismith (-AD), Joubert s.n. (BOL); Huisrivier-

pas (-BC), Marloth 13128 (PRE); Ockertskraal (-CA), Hall

1224 (NBG); 16 km S. of Ockertskraal (-CA), Tolken 1699

(BOL); Adamskraal (-CC), Muir sub BOL 30962; Muiskraal

(-CC), Muir sub Marloth 12185 (PRE); between Laingsburg and

Ladismith, E. Pillans sub N.S. Pillans 968 (BOL, GRA).

U1TTREKSEL

Vyf nuwe species en een varieteit word beskryf en die

nuwe kombinasie Cotyledon tomentosa Harv. subsp.

ladismithiensis {V.Poelln.) Toelken word teweeggebring.

Die volgende nuwe name is gepubliseer: C. sulphurea

Toelken, C. occultans Toelken, C. similis Toelken, C.

viridiflorum Toelken, C. hallii Toelken en C. pygmaea

Barker var. tenuis Toelken.

Bothalia 12,2: 195-197 (1977)

The identity of Erica flavisepala

E. G. H. OLIVER*

abstract

The recording of a few scattered plants of E. flavisepala Guth. & Bol. among sympatric populations of

two other species led to a comparison of their morphological characters. From this comparison a putative

hybrid origin was indicated, thus E. X flavisepala Guth. & Bol^E. thunbergii Montin x E. sphaerocephala

Wendl.

RESUME

L'lDENTITE D’ERICA FLAVISEPALA

L' observation de quelques plants <7’E. flavisepala Guth. & Bol., disperses parmi des populations

sympatriques de deux autres especes, a conduit a une comparison de leurs caracteres morphologiques.

Cette comparaison suggere une origine hybride, vraisemblablement E. X flavisepala Guth. & Bol.

= E. thunbergii Montin X E. sphaerocephala Wendl.

In Flora Capensis (1905) Guthrie and Bolus

described E. flavisepala from material that had been

sent to them in a consignment of E. thunbergii Montin

from an unspecified locality in the Cold Bokkeveld.

The species were reported as growing together.

The authors stated that their material was difficult

to place satisfactorily in any of the sections of the

genus and that it constituted a well-marked species

with a general aspect strikingly similar to that of E.

thunbergii. But they noted that it did not have the

peculiar globose corolla-tube of E. thunbergii and,

although anomalous, placed it in the same section,

Cyatholoma.

Since the original collection of E. flavisepala no

material has been collected until recently. Mr G.

Kirsten, a keen collector of ericas, while examining

populations of E. thunbergii in the Loch Lynne area

of the Cold Bokkeveld, came across a few isolated

plants which I was able to confirm were E. flavisepala.

I decided to visit the area myself to study the species

in the wild.

The fine populations of E. thunbergii were located

without any difficulty in the Hartebeeskloof growing

in seepage zones on south-facing slopes being clearly

visible from a distance. Growing in the same seepage

zones were groups of the pink-flowered E. sphaeroce-

Fig. 1. — Erica thunbergii. 1, flo-

wer, x5; la, corolla, x5; 2,

bracteoles, xlO; 3, sepals,

X 10; 4, androecium and

gynoecium, X5; 5, anther,

front, side and back views,

x 20; 6, ovary, x 10; 7, leaf,

X 10 (from Oliver 5139).

* Botanical Research Institute, Department of Agricultural Technical Services, Private Bag X101, Pretoria.

196

THE IDENTITY OF ERICA FLAVISEPALA

T | Fig. 2. — Erica x flavisepala. 1,

I 1 flower, x5; 2, bracteoles,

■fi x 10; 3, sepal, x 10; 4, an-

! ji droecium and gynoecium,

| | x5; 5, anther, front, side

\ and back views, x 20, 6,

\ i ovary, x 10; 7, leaf, x 10.

1 • / (from Oliver 5140).

Fig. 3. — Erica sphaerocephala 1,

flower, x5; 2, bracteoles,

xlO; 3, sepal, xlO; 4, an-

droecium and gynoecium,

x 5 ; 5, anther, front, side and

back views, x20; 6, ovary,

x 10; 7, leaf, x 10. (from

Oliver 5141).

Fig. 4. — Branches of 1, Erica

sphaerocephala ( Oliver 5141);

2, EricaX flavisepala ( Oliver

5140); 3, Erica thunbergii

( Oliver 5139)

E. G. H. OLIVER

197

TABLE 1.— Comparison of morphological characters of E. thunbergii, E x flavisepala and E. sphaerocephala

phala. While investigating E. thunbergii , a few plants

of E. flavisepala were found. It was immediately

suspected that these few scattered plants were hybrids

between the frequently occurring E. thunbergii and

E. sphaerocephala.

E. thunbergii (Fig. 1) is a very distinct species with

yellow, orange and white flowers with globose corolla-

tubes. It belongs to the section Cyatholoma , which it

shares with E. corydalis Salisb. and E. flavisepala

(Fig. 2), the characteristic feature being the globose

corolla-tube with a large dilated cup-shaped limb.

E. corydalis with flowers all white, is also a very dis-

tinct species occurring in the coastal mountains of

the Caledon district. E. sphaerocephala (Fig. 3) belongs

to a completely different section, Pseuderemia, which

is characterized by having the flowers arranged in a

many-flowered capitate head. (See Fig. 4).

A comparison of morphological characters of my

three collections, Oliver 5139, 5140 and 5141 which

are illustrated in the accompanying figures, clearly

shows the intermediate nature of E. flavisepala. The

basic shapes and sizes of the bracteoles, pedicel,

sepals, corolla, anthers and ovary of the two parent

species are very distinct, while those of E. flavisepala

are intermediate in most respects.

E. thunbergii is almost entirely glabrous except tor

a few hairs very occasionally found on the petioles,

whereas in E. sphaerocephala the branches, leaves,

pedicels, bracteoles and sepals are all puberulous and

ciliate, the cilia being plumose. The anthers of E.

thunbergii are distinctly prognathous and are dis-

tinctly aristate. The ovary in E. thunbergii is stipitate

and in E. sphaerocephala it is sessile.

The flower colours are very distinctive. In E. thun-

bergii the large bracteoles and sepals are bright canary-

yellow, the corolla-tube white and the lobes brillian

orange-red. In E. sphaerocephala the bracteoles and

sepals, which are relatively small, are green to reddish

in colour and the corolla is pink. In E. flavisepala the

bracteoles and sepals are yellow tinged with red and

the corolla is bright pinkish red.

The pollen of E. thunbergii and E. sphaerocephala is

typical of the genus Erica , namely in tetrads. The

pollen of the collection Oliver 5140 is abnormal in

consisting of an equal quantity of tetrads and single

grains mixed with some shrivelled grains. This sup-

ports the view of the hybrid nature of E. flavisepala.

The occurrence of only a few scattered plants of

E. flavisepala amongst the populations of the other

two species indicates that the plants were a chance

cross that is not reproductively established. Unfor-

tunately it has not been possible to study the chromo-

somes. The few species of Erica which have so far been

investigated all have n= 12 and their chromosomes are

extremely small. It is therefore unlikely that taxono-

mically useful information could be derived from a

brief investigation of the three taxa.

The facts as set out (in Table 1 ) indicate the putative

origin of E. flavisepala Guth. & Bol. (E. x flavisepala

Guth. & Bol.=£. thunbergii Montin x E. sphaero-

cephala Wendl.).

UITTREKSEL

Die vermelding van 'n paar verspreide plante van E.

flavisepala Guth. & Bol. tussen simpatriese populasies

van twee ander soorte het gelei tot 'n vergelyking van

hid morfologiese eienskappe. Uit hierdie vergelyking

kan 'n veronderstelde hibridiese oorsprong aangedui

word, dus E. x flavisepala Guth. & Bol. = E. thunbergii

Montin X E sphaerocephala Wendl.

Bothalia 12,2: 199-203 (1977)

The identity of Eriosema nanum

C. H. STIRTON*

ABSTRACT

An examination of available evidence leads to the conclusion that Eriosema nanum Burtt Davy must be

regarded as a synonym of E. ellipticifolium Schinz. The relationship between E. ellipticifolium and E. uniflorum

Burtt Davy is discussed.

RESUME

L'IDENTITE D’ERIOSEMA NANUM

Eriosema nanum Burtt Davy doit etre considere comme un synonyme t/’E. ellipticifolium Schinz.

A cette conclusion basee sur I'examen des preuves disponibles est jointe une discussion de la relation

entre E. ellipticifolium et E. uniflorum Burtt Davy.

INTRODUCTION

Eriosema nanum Burtt Davy, together with E.

populifolium Harv. and E. angustifolium Harv., are

among the rarest of the Fabaceae in South Africa.

Recent intensive field studies of Eriosema (Stirton,

1975) showed the widespread occurrence of hybridi-

zation and of morphological variation in the genus

in South Africa. This paper assesses various aspects

of phenotypic plasticity and interspecific hybridi-

zation as they affect the identity and delimitation of

E. nanum.

HISTORY

Schinz, in Vjschr. Naturf. Ges. Zurich 66: 229,

1921, based E. ellipticifolium on Junod 1411 from

Shilouvane and Junod 2534 (Fig. 1) from Marovunge,

both in the Eastern Transvaal. Owing to the absence

of ripe or almost ripe fruits, he was unable to

establish clearly whether he was dealing with a

Rhynchosia or an Eriosema species (“Solange keine

reifen oder nahezu reifen Friichte vorliegen, ist es

vorlaufig ein aussichtsloses Bemiihen, feststellen zu

wollen, ob es sich um eine Rhynchosia — oder eine

Eriosema- Art handelt, sicher ist, dass sie sich mit

keiner der mir bekannten Arten dieser oder jener

Gattung deckt”). Fie commented, however, that the

plant was reminiscent of E. salignum, but differed

Fig. 1. — Junod 2534, lectotype of Eriosema ellipticifolium.

‘Botanical Research Institute, Department of Agricultural

Technical Services. Private Bag X101, Pretoria.

in its lower surface indumentum. The name E.

ellipticifolium has never been taken up.

In 1932 Burtt Davy treated E. ellipticifolium very

briefly under the heading “species not seen”. In this

work, however, he published Eriosema nanum based

on Galpin 1139 from the summit of the Saddleback

Mountain in the Barberton area (Fig. 2). As its

allies he noted E. rufescens Schinz and E. burkei

Benth. and commented also that it was perhaps

closest to E. uniflorum Burtt Davy.

Fig. 2 — Galpin 1139, isotype of Eriosema nanum.

A study of type specimens leaves no doubt that

E. nanum is synonymous with E. ellipticifolium.

Population studies showed that the wide leaf variation

encountered is the result of a dine of decreasing

width in a northerly direction (Stirton, 1975). The

type localities of E. ellipticifolium occur in the

northern extremities of this range. It is unfortunate

that the more descriptive name nanum must be

superseded.

200

THE IDENTITY OF ERIOSEMA NANUM

FIELD OBSERVATIONS

Plants growing in full sun tended to be shorter

and more compact than plants of the same species

found growing under different intensities of shade.

In Fig. 3 plants of E. ellipticifolium that were found

growing in short open grassland (plant 1) are con-

trasted with those growing along a road bank in a

pine forest (plant 2). Plant 1 can be seen to be smaller

and more compact than plant 2. Other obvious

differences, not all observable from the photographs,

are the longer inflorescences, the thinner leaves and

the less prominent secondary and tertiary venation

in plant 2. Plants that grew through a thick canopy

of grass had the same facies as the plants that grew

in or along forest margins. These features can be

clearly seen on Stirton 1431. These various expressions

of morphological form could possibly be the result

of a competition balance of mineral nutrients, degree

of light intensity, extent of water availability, or a

combination of these and other factors. The plants

shown in Fig. 3 were chosen as representative of

the range of variation within the species after a critical

inspection of their populations.

Fig. 3 — Eriosema ellipticifolium. The different facies of plants

growing: 1, in short open grassland (Stirton 1467); 2, on a

roadbank in a pine forest ( Stirton 1486); 3, in burnt veld

( Stirton 1040); 4, in unburnt veld adjacent to 3 (Stirton 1041).

Plants that grew in burnt veld (Fig. 3: 3) had a

more stunted form than plants of the same species

which grew in adjacent unburnt veld (Fig. 3: 4). These

plants were collected on the same day. The various

populations were observed subsequently and it was

found that once the new grass had come away the

“dwarf plants” began to grow lank so that by the

end of the season they were similar, but still shorter,

than plants collected in the adjacent tall unburnt

grassveld. The plants that grew in the burnt veld had

almost completed their flowering period by the time

the plants which grew in the unburnt veld had begun

to flower.

The observations in open veld and in forest, and

in burnt and unburnt veld, were found to be consistent

throughout the range of distribution of E. elliptici-

folium. The results lead me to suspect that once the

type specimen of E. urtiflorum Burtt Davy is compared,

it will probably have a facies similar to plant 2 in

Fig. 3 and hence synonymous with E. ellipticifolium.

Compton (1974, pers. comm.) has indicated that he

intends to incorporate E. urtiflorum under E. nanum

in his revised Flora of Swaziland. Judgement is here

reserved until the type is traced, since the description

of E. urtiflorum has features linking E. ellipticifolium

and Stirton 1482 ( E . sp. nov.) allied to E. cordatum.

On the Galpin 1139 sheet from the Bolus Herbarium

there is a specimen Bolus 11854 that has three Bolus

manuscript names; uniflorum, pumilum and cryptantha.

Appended to the specimen is a note, apparently in

N. E. Brown’s handwriting, that says “We afterwards

decided that this was probably a stunted form of

E. burkei" . The specimen is E. ellipticifolium. Field

studies (Stirton, 1975), however, indicated that

stunted forms of E. pauciflorum Klotzsch, rather than

E. burkei, are deceptively similar to E. ellipticifolium.

The most noticeable variation observed in the field

was the range in number of flowers per inflorescence.

Plants of the same population were found to have

inflorescences bearing from one to ten flowers.

Although most specimens have been reasonably

easy to place in this species there remain a few

problems. Hybridization cannot be ruled out as the

cause of some complexing difficulties encountered in a

recent field trip. Two populations in particular

require a detailed study as in both areas E. elliptici-

folium grows sympatrically with the rare E. angusti-

folium Burtt Davy and an undescribed taxon. A

feature of these populations is their marked

geographical separation viz. Magoebaskloof (N.

Transvaal) and Havelock (Swaziland) and also the

wide range of “intermedates” found. Flower colour,

pubescence and the shape of stipules have been

shown in preliminary hybridization studies (Stirton,

1975) on other species to be reliable indicators of

putative hybrids. These three characters varied

markedly in “intermediate” plants in both popu-

lations. E. angustifolium has a striking, stiff, rufous,

patent indumentum, linear leaves, erect habit, and

yellow flowers, and is not readily confused with

either E. ellipticifolium or Stirton 1445 (E. sp.) The

latter occurs from the N. Transvaal southwards to

Swaziland and although fairly common in isolated

areas has not been previously collected. This multi-

stemmed plant is a pink and yellow flowered perennial

with prostrate habit, and small unifoliolate leaves.

Its closest affinity is E. ellipticifolium. The intermediate

plants at Magoebaskloof seem to be hybrids between

E. angustifolium and E. ellipticifolium ( Stirton 1442),

and between Stirton 1445 (E. sp.) and E. ellipticifolium

( Stirton 1446). Further field studies are necessary

before all three putative parents are clearly delimited.

The inter-relationships of these three species remain

obscure.

Three specimens named E. ellipticifolium in this

study are doubtful: these are Jacobsen 1587, Coetzer

150, and Vahrmeijer 2433. The last two of these,

although very close to E. ellipticifolium, differ in

pubescence and their very acute leaves. All these

will no doubt be easier to place once all the montane

species of Eriosema have been studied.

E. ellipticifolium begins flowering in early September

and reaches a peak in December, occasionally

extending into January. The Natal populations

flower much earlier than those in the Transvaal.

TAXONOMY

Eriosema ellipticifolium Schinz in Vjschr. Naturf.

Ges. Zurich 66: 229 (1921). Syntypes: Transvaal,

Shilouvane, near Sanatorium, Junod 1411 (Z!);

Marouvunge, Junod 2534 (Z!). Lectotype: Junod

2534 (Z!).

Eriosema nanum Burtt Davy, FI. Transv., 2: XXII (1932);

Ross, FI. Natal 208 (1972). Type: Transvaal, Saddleback

Mountain, Galpin 1139 (PRE, holo. !; BOL, NBG, NH).

C. H. STIRTON

201

Fig. 4. — Eriosema ellipticifolium. 1, habit, 1: 2, perennial rootstock, x 1; 3, flower bract, x 74; 4, flower, x 24;

5 calyx opened out, x 74; 6a, standard opened out x 54; 6b, standard closed, x 54; 7, wing, x 54; 8, keel,

x 54; 9, vexillar stamen, x 74; 10, staminal sheath, x 74; 11, discoid floral nectary, x 15; 12, gynoecium,

x 74; 13, stigma, x 33; 14, fruit pod, x 1; 15a, seed with strophiole, face view, x 54; 15b, seed with strophiole,

marginal view showing hilum, x 54-

202

THE IDENTITY OF ERIOSEMA NANUM

Perennial herb (4) 9-12 (25) cm high with one to

six stems from short stylopodium of carrot-like,

usually constricted, underground rootstock. Stems

erect or ascending, 3-10 (23) cm long; terete or

angular, often ribbed, glandular or eglandular,

clothed with short white pubescent and long spreading

pale yellow or white, hairs. Stipules free, 6-15 mm

long, 2, 0-3, 5 mm wide; narrowly ovate to ovate,

often falcate, striate, white or ferruginous hairy.

Leaves trifoliolate, basal leaves unifoliolate; leaflets

4. 0- 7,0 cm long, 1,4-3, 5 cm wide, laterals smaller,

elliptic to narrowly elliptic or narrowly oblong, base

truncate, mucronate, middle nerve slightly sunken

above, prominent and subreticulate below, short

appressed pubescent on both sides, denser on nerves,

densely orange gland-dotted below, lightly above,

terminal leaflet symmetrical, laterals asymmetrical

with bases slightly oblique; young leaves glandular,

long appressed tawny haired on veins with short

curly dense pubescence between; Petiole 4-11 mm

long; Petiolules ca. 2 mm long; rachis 5-8 mm long.

Racemes 1-10-flowered; peduncle (1 ,0) 2,0-3, 5 (7,5)

cm long, canaliculate, with long spreading tawny,

or long white hairs interspersed with short white

deflexed hairs; rachis 0,7-2, 1 cm long; pedicels

ca. 3 mm long. Flowers 6-10 (13) mm long, 3-5 mm

wide, reflexed, corolla barely exceeding calyx, yellow;

bracts 6-8 mm long, 1,0-1, 5 mm wide, lanceolate,

shallow boat-shaped, 3-6-veined, thinly pilose, more

than half length of flower. Calyx 6-9 mm long with

long and short admixed pale yellow or white hairs,

glandular outside, hairy inside at apex of lobes with

smaller hairs scattered on either side of veins; tube

2.0- 2, 5 mm long, slightly longer between horn

lobes; lobes unequal, lanceolate, 3-4 times longer

than tube, horn lobes 7, 5-8,0 mm long curving

inwards, lateral lobes 7, 0-8,0 mm long curving

upwards; keel lobe longest, 8, 0-9, 2 mm long.

Standard yellow (7,5) 9-10 (13,0) mm long, (4,5)

6. 0- 7,0 mm wide, narrowly obovate, often narrowed

towards base, white pubescent and glandular on

back; claw 2, 0-2, 6 mm long; width between auricles

3. 0- 4,0 mm; appendages free of auricles, being

upcurled flaps joined slightly in the middle, (1,5)

3. 5- 4,0 mm from base of claw. Wings 7,5-10,1 mm

long, 2,0-3, 3 mm wide at maximum, longer than

keel blades, eglandular, occasionally very glandular;

claw 1,5-2, 5 mm long, attenuate; auricle (1,2)

1 .5- 1 ,6 mm high, straight. Keel blades 6, 5-7, 5 (8,9)

mm long, 3, 0-4,0 (4,6) mm wide at maximum,

auricle 1,0-1, 7 (2,0) mm high, claw 1, 7-3,0 mm

long, attenuate, apex rostrate. Staminal sheath 5,7-

7.0 mm long, 2, 7-4,0 mm wide at maximum, free

stamen 5, 0-6,0 mm long, basal knee 1,0-1, 3 mm

long. Gynoecium 6, 0-6, 6 mm long; ovary 2,4-

3.0 mm long, subsessile; clothed with long stiff hairs

as far as point of style swelling or flexure; gynophore

0, 1-0,5 mm; style unevenly thickened, widest at

middle of thickening, curvature 2,0-2, 2 (2,7) mm

high; stigma globose, inserted. Nectary 0,4-0, 5 mm

high, discoid, margin erose. Fruits broadly ovate,

1,4-1, 6 cm long, 0, 9-1,0 cm wide, clothed with

appressed light yellow hairs, denser along sutures,

with small stiff hairs between; seeds 0, 5-0,6 cm

long, 0,3 cm wide, black, or greenish yellow with

small faint reddish blotches. (Fig. 4).

Eriosema ellipticifolium is restricted to isolated

mountain “islands” in Swaziland, the Transvaal and

Natal (Fig. 5). This species exhibits a disjunct distri-

bution over a wide area and over diverging ecological

conditions and veld types. If is found between

1 600-2 500 m growing predominantly amongst rocks

in short grassland on dry ridges with a northwest

aspect. A number of populations has been located

on forest margins and along forest roads at the

Witklip, Mariepskop and Woodbush Forest Reserves.

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

11601 (PRE).

Transvaal. — 2329 (Pietersburg): Haenertsburg (-DD), Pott

4798 (PRE). 2330 (Tzaneen): Magoebaskloof Hotel (-CC),

Stirton 1443 (PRE). 2429 (Zebediela): Donkerfkloof near Chu-

niespoort (-BA), Vahrmeijer 2443 (PRE). 2430 (Pilgrim’s Rest):

The Downs (-AA), Rogers 22057 (PRE); near Reitz’s grave on

Mariepskop (-DB), Stirton 1449 (PRE); Ohrigstad Nature

Reserve (-DC), Jacobsen 1587 (PRE). 2530 (Lydenburg)

Dullstroom (-AC), G at pin s.n. (24-12-1932, BOL); Witklip

Forest Station (-BD), Kluge 195, 626 (PRE); summit of

Saddleback Mountain (-CC), Galpin 1139 (BOL, GRA, NBG,

NH, PRE). 2531 (Komatipoort): Havelock Border Post (-CC),

Stirton 1464 (PRE).

Natal. — 2731 (Louwsburg): Mount Inyati (-CC), Hilliard &

Burtt 5889 (NH, NU). 2830 (Dundee): The Kop (-DD), Stirton

1033, 1040, 1041 (PRE).

Fig. 5. — Eriosema ellipticifolium. Known distribution in

Southern Africa.

Junod 2534 is chosen as lectotype, since the

quantitative data given in the protologue indicates

that Schinz must have based most of his description

on the two specimens on this sheet.

This little-known species has proved to be more

common and widespread than was previously

accepted. The available herbarium material had been

placed under no less than six species. It had been

most commonly confused with Eriosema cordatum

var. cordatum , but is readily separated from this

and all other species by its very long calyx lobes

which almost equal the length of the flower.

Despite its widespread distribution (Fig. 5) this

species is remarkably uniform and distinctive in the

field. Its poor representation in herbaria is probably

attributable to its dwarf habit (Fig. 6) and not to its

scarcity in the field as, during a recent trip to the

eastern Transvaal, it was found to be locally common

throughout the moist highlands.

C. H. STIRTON

203

Fig. 6. — Eriosema ellipticifolium: dwarf habit.

ACKNOWLEDGEMENTS

This article is an extension of part of a thesis

submitted for the M.Sc. degree at the University of

Natal, Pietermaritzburg. I am most grateful to

Dr. K. D. Gordon-Gray for her guidance and advice

during the initial stages of this study and for kindly

providing funds for Fig. 4. I am grateful to the

Director, Botanisches Museum der (Jniversitat Zurich,

for the loan of the types of E. ellipticifolium Schinz.

UITTREKSEL

'n Ondersoek van al die beskikbare gegewens dui

daarop dat Eriosema nanum Burtt Davy 'n sinoniem

van E. ellipticifolium Schinz is. Die verwantskappe

tussen E. ellipticifolium en E. uniflorum Burtt Davy

word ook bespreek.

REFERENCES

Burtt Davy, J., 1932. A manual of the flowering plants and

ferns of the Transvaal with Swaziland, South Africa. London :

Longmans, Green & Co. 529 pp.

Stirton, C. H., 1975. A contribution to knowledge of the genus

Eriosema ( Leguminosae-Lotoideae ) in Southern Africa

(. excluding Mozambique and Rhodesia). Unpublished M.Sc.

thesis, University of Natal, Pietermaritzburg. 182 pp.

APPENDIX

This appendix lists all specimens considered by the author to

belong to Eriosema ellipticifolium Schinz:

Bolus 11854 (BOL); Coetzer 150 (PRE); Galpin s.n. (BOL),

1139 (BOL, GRA, NBG, NH, PRE); Gilmore 2284 (PRE);

Hilliard 2942 (NU); Hilliard & Burtt 3358 (NU), 5889, 5908

(NH, NU); Junod 1411, 2534 (Z); Kluge 195, 626 (PRE); Muller

2012 (PRE); Pott 4798 (PRE); Rogers 1160, 22057 (PRE);

Stirton 1033, 1040, 1041, 1346, 1431, 1436, 1439, 1441, 1443,

1449, 1464, 1467, 1471, 1474, 1475, 1483, 1486, 1487 (PRE);

Vahrmeijer 2433 (PRE); Van der Schijff 6440 (PRE).

Bothalia 22, 2: 205-208(1977)

Morphological studies of the Ochnaceae*

P. C. V. DU TOITj

ABSTRACT

The cause of the characteristic sloughing of the bark of certain Ochna species is described, while the mor-

phology of the growth buds, leaves and stipules of South African genera of the Ochnaceae is discussed. The

taxonomic significance of these features is indicated where relevant.

RESUME

ETUDES MORPHOLOGIQUES SUR LES OCHNACEAE

La cause de la desquamation de Tecorce, caracteristique de certaines especes du genre Ochna,

est dec rite ; on discute la morphologic des bourgeons de croissance, des feuilles et des stipules dans les

genres sud-africains d'Ochnaceae et, quand il y a lieu, on indique la valeur taxonomique de ces carac-

teres.

INTRODUCTION

During a recent morphological study of the family

Ochnaceae, it was found that relatively little was

known of the organography and anatomy of the

family. A wealth of taxonomic literature was avail-

able, but only a few works gave a synoptic treatment

of the anatomy. The aim of this paper is to draw

attention to these taxonomically important and

interesting characteristics and to stimulate the extra-

polation of these finding to other families.

BARK

An interesting characteristic of some of the indi-

genous species of the Ochnaceae is the rhytidome or

outer bark, i.e. all the tissues on the outside of the

periderm, that peels off in strips or scales.

According to Esau (1965) this manner of rhytidome

formation is common in those species where the

phellogen is initiated superficially. While working on

the anatomy of the family it was noticed in two

species, viz. Ochna arborea and O. pulchra, that the

phellogen was initiated in the epidermis, the epidermal

and phellogen cells lying on the same radial lines, but

at the same time, it was also initiated quite deep in

the cortex. In this instance, the phellogen forms an

interrupted undulating band running through the

epidermis and cortex right round the stem. In time

this undulating band breaks up into outward curving

sectors, like shells. These outward curving shells,

which overlap and are superimposed over yet others

like it, are shed at different times, leaving a smooth,

mottled layer of bark (i.e. all the tissues on the outside

of the vascular cambium). In these two species it was

found that the phellogen gives rise to a very thin layer

of phellem, coinciding with the smoothness of the

bark when the outer layers slough off (Fig. 1). In all

the other indigenous species of the Ochnaceae, the

phellogen is initiated relatively deep in the cortex

with, in addition, a thick layer of phellem. The outer

layers slough away gradually and the bark, exposed

after sloughing, is rough and evenly coloured (Fig. 2).

Making use of this character the species can be

separated into two groups, the one group coinciding

exactly with the section Renicarpus to which O.

arborea and O. pulchra belong.

* Forms part of a M.Sc. thesis submitted at the University

of Pretoria.

t Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

BUDS AND GROWTH PATTERN

Two interesting features regarding the growth

pattern of the Ochnaceae were observed that have

hitherto escaped notice. The first is the two different

kinds of buds encountered, and the second, the

growth pattern. The first type of bud found in the

family is responsible for the annual growth increment.

The growth buds are laterally compressed and taper

to a sharp point (Fig. 3). The second type of bud is

the flower bud of which two kinds are found. Species

where the flowers are borne singly have easily recog-

nizable pear-shaped and globose buds (Fig. 4). Where

the flowers are borne in clusters or cymes the buds

are produced in a club-shaped structure, similar to the

growth buds (Fig. 5).

Fig. 1.— The smooth, peeling bark of Ochna pulchra .

206

MORPHOLOGICAL STUDIES OF THE OCHNACEAE

Fig. 4. — The single flower buds of Ochna serrulata.

Fig. 5. — The club-shaped compound flo ver bud of Ochna

pulchra.

P. C. V. DU TOIT

207

Fig. 6. — The cyme of Ochna

pulchra; note the two rows of

boat-shaped scales.

Both the growth buds and the compound flower

buds are protected by a series of scales. The scales

are arranged in alternating series. One large serrated

or entire scale, covering the axillary bud. On the

opposite side of the stem, the space between the two

margins of the large scale is covered by a lanceolate

or subuliform scale, which seems to be homologous to

the two stipules of the foliage leaves.

At the end of the growing season, the flower and

leaf buds are already fully developed and open

(unfold) with the first spring rains of the following

season. In both the growth and flowering buds, a

typical feature of the Ochnaceae is displayed, viz. the

growthflush, which takes place in early spring when

the temperature rises and just after the first spring

rains. During the growthflush the scales break away

in two or three series and the stem and peduncle

elongates (Fig. 6). The scales mostly break away in

two series due to their pseudo-distichous arrangement.

The angle of divergence is approximately 150°, but

due to the stem being twisted, the leaves and scales

take on the appearance of being distichous. The

spirally arranged nature of the leaves can be seen

when the young leaves unfold. During the growth-

flush, which lasts for approximately a week, the

annual growth increment of the plant is completed

and the axillary buds of most species are quiescent

for the rest of the growing season. The annual growth

increment can be recognized by the amount of new

growth between the series of scale scars at the tip of

the previous season’s stems and the growth bud. The

quiescent axillary buds in the series of scale scars give

rise to the flower buds and/or sideshoots of the

following season. In some species stems may also be

terminated by flowers.

LEAVES AND STIPULES

The family Ochnaceae in Southern Africa is

represented by two genera, viz. Ochna and Bracken-

ridgea. In distinguishing between these two major

taxa, the morphology of the leaves and stipules was

found to be taxonomically useful. According to

10mm

Fig. 7. — Diagram to illustrate the stipules of A, Ochna and B,

Brackenridgea.

208

MORPHOLOGICAL STUDIES OF THE OCHNACEAE

Fig. 8. — Diagram to illustrate the

types of veins in A, Ochna

and B, Brackenridgea. PV — -

Primary-vein; SV — Seconda-

ry-vein. Arrows point to the

angle of emergence of the

different veins.

Sinnott and Bailey (1914), there is a relationship

between the occurrence of stipules and the dentations

on the leaf margins. In the indigenous species of the

Ochnaceae, the leaf margin is excurrent onto the

adaxial side of the petiole and joined to the stipules.

In these instances the stipules can then be regarded

as the most basally situated teeth of the leaf margin.

In all the species of Ochna the stipules are caducous,

especially in Ochna pulchra , in which the stipules are

shed just after the growthflush and the teeth of the

leaf margin are deciduous with age. The stipules are

lanceolate to spathulate, undivided and entire (Fig.

7A). The stipules of Brackenridgea remain on the

plant for the whole growing season, and are divided

into a number of linear lobes, which may be entire or

dentate (Fig. 7B).

The venation is a useful taxonomic character and

the veins can be divided into primary and secondary

veins (Fig. 8). The primary veins are conspicuously

prominent on the adaxial surface, with the secondary

veins less so. In the case of most of the Ochna species

both the primary and secondary veins arise from the

midrib at an angle of 60°-70°. The veins run parallel

to one another with the secondary veins gradually

lessening in diameter until they disappear. The primary

veins run to the leaf margin, turning distally with

their diameter gradually lessening (Fig. 8A). In the

case of Brackenridgea arenaria and O. confusa, the

primary veins arise from the midrib, at an angle of

30°-40°, with the primary veins anastomosing near

the distal end of the leaf. The secondary veins arise

from the midrib at an angle of 60°-70° and run

parallel to one another, gradually lessening in

diameter until they end at the primary veins (Fig. 8B).

UITTREKSEL

Die oorsaak van die kenmerkende wyse waarop

die bas van sommige Ochna soorte afdop, asook die

morfologie van die groeiknoppe, loofblare en steun-

blare word bespreek. Die taksonomiese toepaslikheid

van die eienskappe word aangedni waar ter sake.

REFERENCES

Esau, K., 1965. Plant anatomy. 338-350. New York: John Wiley

and Son.

Sinnott, E. W. & Bailey, I. W., 1914. Investigations on the

phylogeny of the Angiosperms. Am. J. Bot. 1,9: 441-453.

Bothalia 12,2: 209—213 (1977)

Leaf anatomy of the South African Danthonieae (Poaceae).

I. The genus Dregeochloa

R. P. ELLIS*

ABSTRACT

The anatomical structure of the leaf blade as seen in transverse section, and the abaxial epidermis of

Dregeochloa pumila and D. calviniensis is described and illustrated. A generic description is included and the

relationships of the genus are briefly discussed.

RESUME

ANATOMIE DE LA FEUILLE DES DANTHONIEAE ( POACEAE ) SUD-AFR/CAINES I LE

GENRE DREGEOCHLOA

La structure anatomique du limbe foliaire vu en coupe transversale, ainsi que Vepiderme abaxial

de Dregeochloa pumila et de D.calviniensis son t decrits et il/ustres. On y joint une description du

genre avec une breve discussion de ses relations.

INTRODUCTION

The genus Dregeochloa Conert was described to

accommodate the species Danthonia pumila Nees and

a newly described species Dregeochloa calviniensis

Conert (Conert, 1966). These two species have certain

distinct characteristics of spikelet morphology, leaf

anatomy and especially the structure of the mature

caryopsis, which indicates that this genus occupies a

unique and somewhat isolated position in the Dantho-

nieae. Plants of the genus were previously assigned to

the genus Danthonia DC. (Chippindall, 1955), and

certain authors (De Wet, 1954, 1956; Jacques-Felix,

1962) have suggested their inclusion in Asthenatherum

Nevski. However, they can readily be distinguished

from both genera by the development of the leaves,

the insertion of the central awn and the structure of the

caryopsis (Conert, 1966). In addition, these species

exhibit characteristic leaf anatomy which tends to

confirm their placement together in a separate genus,

but throws little light on the phylogenetic position of

this genus. In the anatomical descriptions which

follow, the following abbreviations will be used:

vb/s — vascular bundle/s

l 'vb/s — first order vascular bundle/s

2'vb/s — second order vascular bundle/s

3'vb/s — third order vascular bundle/s

ibs — inner bundle sheath; mestome sheath

obs — outer bundle sheath; parenchyma sheath

ANATOMICAL DESCRIPTION OF THE GENUS

DREGEOCHLOA

Leaf in transverse section

Leaf outline : U-shaped or canaliculate with two halves

of lamina curved upwards to varying degrees on either

side of median vascular bundle; forming no angle.

Leaves narrow (between 1 , 60 mm and 3 , 50 mm wide),

proportionately very thick (between 0,30 mm and

0,70 mm). Ribs and furrows ; adaxial furrows present

between all vbs; equally sized, rounded, adaxial ribs

present over all vbs irrespective of order. Abaxial ribs

and furrows well developed; furrows equal to or

deeper than adaxial furrows; ribs in form of flat-

topped sclerenchyma caps with horizontal prickles at

edges interlocking over cleft-like furrow. Median

vascular bundle: present; midrib or keel not developed.

Vascular bundle arrangement: 5 1 vbs in blade;

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

no 2'vbs present; all bundles located nearer abaxial

surface. Vascular bundle structure: all bundles usually

rounded in section although D. pumila exhibits a

tendency for bundles to be vertically elongated and

elliptical; phloem of l'vbs adjoins inner, thickened

sheath; no lysigenous cavity or protoxylem vessels

developed; metaxylem vessels extremely narrow

(diameter less than adaxial ibs cells), angular and only

slightly thickened. Vascular bundle sheaths: round in

shape in all 3'vbs and most l'vbs, but in D. pumila

l'vb sheaths may be slightly elliptical. Obs always

entire with no extensions; comprised of 11-15 cells

around 3'vbs and 12-17 cells around l'vbs; all cells

similar in shape, thin-walled and conspicuous but

not larger than mesophyll cells. Sclerenchyma:

abaxial strands well developed in wide, straight, deep

bands forming caps on flat abaxial ribs; no scleren-

chyma between vbs; fibres thick-walled with lumen

narrow; walls stain blue-green with safranin and fast

green. Margin: sclerenchyma a well developed cap,

somewhat pointed; crescent-shaped or fibres may

extend abaxially. Sclerenchyma of margin never

extends further than directly above and below ultimate

vb. Margin without angular prickles. Mesophyll:

indistinctly radiate; chlorenchyma of elongated,

tabular cells relatively loosely and irregularly

arranged; air spaces frequent. Colourless cells: entirely

lacking except for bulliform cells. Adaxial epidermis:

costal prickle hairs with bases not larger than, and

situated between, normal epidermal cells; frequent;

hairs slender and elongated, those of D. pumila

(±0,10 mm) about twice the length of those of

D. calviniensis (±0,05 mm); no papillae. Abaxial

epidermis: outer tangential wall of cells flattened or

slightly arched; not papillose.

Abaxial epidermis

Intercostal zones: cleft-like, intercostal, abaxial

furrows make study and illustration of cells of this

zone difficult. Intercostal long cells: length slightly

greater than width; parallel-sided; moderate undula-

tions of anticlinal walls. Stomata: appear low-dome

shaped; 1-2 rows on each vertical wall of cleft-like

furrow. Papillae: absent. Prickle-hairs: interlocking

prickles present overlying stomatal furrow; single row

of prickles on edge of costal zones. Micro-hairs:

none observed. Macro-hairs: absent. Silica-bodies:

present throughout costal zone; silica cells alternate

with costal short cells.

210 LEAF ANATOMY OF THE SOUTH AFRICAN D ANTHONIE AE (POACEAE). I. THE GENUS DREGEOCHLOA

ANATOMICAL DESCRIPTIONS OF THE SPECIES

1. Dregeochloa pumila (Nees) Conert

Leaf in transverse section

Leaf outline: U-shaped leaves with base and sides

about equal in length or more commonly less inrolled

with wide, broad base to U. Width from 3,00 mm to

3,45 mm and thickness from 0,63 mm to 0,70 mm in

freshly fixed leaves. Dried herbarium specimens

narrower (2,00 mm wide) and thinner (0,30 mm

thick). Ribs and furrows: adaxial furrows very shallow

relative to leaf thickness; furrows narrow and cleft-

like in sections from dried herbarium material but

with broad bases with relatively steep sides in freshly

fixed material. Abaxial ribs taller than adaxial ribs;

interlocking prickles well-developed and common.

Vascular bundle arrangement : 1 ' vbs separated by

single 3'vb except on either side of median vascular

bundle where there are two 3 'vbs. Vascular bundle

structure: 3'vb diameter almost same as l'vbs.

Vascular bundle sheaths: obs cells rounded or with

radial walls straight and tangential walls inflated;

translucent, without chloroplasts or with few chloro-

plasts similar to those of mesophyll. Ibs not distinct

around 3 'vbs; probably absent; present around l'vbs;

entire, composed of small thickened cells; thickening

may be on inner tangential and radial walls or equally

on all walls when lumen may become more or less

excluded. Adaxial cells of ibs larger than lateral cells;

relatively thin walled. Adaxial sclerenchyma: small

strands developed; shallow sub-epidermal strip or

deeper groups of fibres. Strands similar in size over

all vbs. Mesophyll : indistinctly radiate with chloren-

chyma surrounding each individual bundle divided

by abaxial furrows and large adaxial bulliform cells.

Adaxial chlorenchyma groups narrow and composed

of palisade-like cells between adjacent bulliform cell

groups. Adaxial epidermis: bulliform cells resemble

fan-shaped groups with exceptionally large central

cell with straight sides occupying more than \ the leaf

thickness. Costal epidermal cells unthickened or with

thin cuticle. Abaxial epidermis: bulliform cells small,

in fan-shaped groups situated at base of furrows;

largest bulliform cell not larger than obs cells. Costal

epidermal cells with outer wall slightly thickened.

Located on sides of adaxial ribs are horizontally

projecting prickles with bases not larger than, and

situated between normal epidermal cells; barbs thick

and relatively very long (0, 10 mm), interlocking over

furrows.

Abaxial epidermis

Prickles: short, hook-like base; barb long (±0,10

mm); barb developed from apex of base; barbs lie at

right-angles across abaxial furrows. Silica bodies:

Round, circular to oblong in shape; sometimes shortly

dumb-bell shaped with a wide middle portion (Fig. 8).

Specimens examined

S.W.A. — 2615 (Luderitz): 1,6 km south of Luderitz lagoon

(-CA), Giess & van Vuuren 653. 2715 (Bogenfels): Pomona

(-AB), Din ter 4018.

Cape. — 2816 (Oranjemund): 2 km west of Beauvallon on road

to Alexander Bay (-DA), Ellis 2173, 2174: 9 km west of Beau-

vallon, Tolken 5278.

Discussion

D. pumila is a short (3-10 cm tall), tufted perennial,

often forming small, dense, flat cushions up to 15 cm

in diameter. It grows in crevices or in loose sand

amongst shale-like, schist rocks in a region of extre-

mely low winter rainfall (±50,0 mm), but where

heavy sea mists occur frequently. It appears to be

restricted to the coastal belt, less than 1 5 km from the

sea, from just south of the Orange River mouth to the

Luderitz area in South West Africa (Fig. 13).

This region undoubtedly is a harsh habitat with

extremes of temperature, insolation and precipitation.

The proximity to the sea has a moderating effect on

seasonal temperature range and frequent mists

probably offset the lack of rainfall somewhat, but an

extremely large diurnal temperature range is to be

anticipated on the dark rocks which this plant prefers.

However, xerophytic adaptions are not well reflected

in the anatomy of the leaf of D. pumila. The leaf is

not permanently infolded and the usual copious

sclerenchyma and thickened cuticle, often correlated

with dry localities, is lacking (Figs. 1 and 2). The

stomata are, however, restricted to abaxial furrows

which are further protected by numerous interlocking

prickles. This arrangement must limit excessive water

loss to a certain degree. In addition the leaves are

very short with a leathery, succulent texture and the

plant forms a low, dense cushion. These growth

characteristics are commonly found in xerophytes but

are exceptional in the Poaceae.

No other members of the Danthonieae are found

in the same habitat as D. pumila although Asthena-

therum glaucum (Nees) Nevski and Merxmuellera

rangei (Pilg) Conert are found in the adjoining sandy

areas but they extend further into Namaqualand, the

Namib Desert and even the Kalahari in the case of

A. glaucum. Anatomically these species are easily

seperated.

Sections taken from herbarium specimens have a

somewhat more xerophytic appearance. The adaxial

and abaxial furrows are more narrow and cleft-like

and the bulliform cells appear reduced in size. In fact,

these specimens superficially resemble D. calviniensis

sections (of which only herbarium material was

studied) more closely than fully turgid, freshly fixed

specimens of D. pumila. This should serve as a caution

when anatomical comparisons are made between

sections of herbarium and fresh material. Thus, Dinter

4018 (Fig. 1) resembles the sections of D. calviniensis

(Figs. 9 and 10) and it is only the sequence of the

different orders of vascular bundle and the silica

bodies and prickle hairs (Figs. 5 and 6) of the

epidermis that reveal it to be a typical D. pumila

specimen.

Remarkably little variation in leaf anatomy was

evident in the specimens studied which may be

expected from a species inhabiting a limited area with

uniform climate, and occupying such a specialized

niche. This species can readily be recognised by the

elongated, deep bulliform cells associated with narrow

columns of adaxial chlorenchyma cells, the thickness

of the leaf and the abaxial furrows with interlocking

prickles.

2. Dregeochloa calviniensis Conert

Leaf in transverse section

Leaf outline: U-shaped with base and sides equal

in length to hollow cylindrical condition where

margins almost meet. Generally more tightly inrolled

than D. pumila; width from 1 ,60 mm to 2,30 mm in

dried material. Ribs and furrows: adaxial furrows of

medium depth and at least \ of leaf thickness. (This

greater depth may be due to shrinkage in the herba-

rium material examined). Furrows narrow, cleft-like.

Adaxial ribs same height as abaxial ribs although

shaped differently; not moniliform; interlocking

prickles small and infrequent. Vascular bundle

arrangement : one 3'vb on either side of the median

R. P. ELLIS

2)1

Figs 1 & 2. — Dregeochloa pumila: transverse sections of the leaf blade; 1, Dinter 4018; 2, Ellis 2174.

Figs 3-8.— Dregeochloa pumila: abaxial epidermis; 3 & 4, Ellis 2173; 5 & 6, Dinter 4018; 7 & 8, Giess & Van Vuuren 653.

212 LEAF ANATOMY OF THE SOUTH AFRICAN DANTHONIEAE (POACEAE). I. THE GENUS DREGEOCHLOA

Figs 9-12. — Dregeochloa calviniensis. 9 & 10, transverse sections of the leaf blade; 11 & 12, abaxial epidermis. 9, Erasmus s.n.; 10

11 & 12, Dekkers s.n.

vascular bundle; lateral two l'vbs adjacent to one

another with no 3'vbs present between them. Vascular

bundle sheath: obs cells inflated and rounded; chloro-

plasts appear specialized and concentrated near inner

tangential wall (Figs. 9 and 10). (This observation

needs confirmation using freshly fixed material).

Ibs well developed and distinct around both 1' and

3'vbs, complete, composed of small thickened cells

often with inner tangential and radial walls markedly

thickened. Around l'vbs the walls of cells adjacent

to the phloem very thick with narrow lumen. Adaxial

cells of ibs larger than lateral cells of sheath. Adaxial

sclerenchyma: well developed strand; usually arched

following shape of rib; similar over all vascular

bundles. Mesophyll: indistinctly radiate with the

chlorenchyma surrounding each individual bundle

being divided by abaxial and adaxial furrows.

Adaxial epidermis: bulliform cells appear absent,

or small and present on sides and bases of~

furrows (Fig. 10). Costal epidermal cells all with

outer wall slightly thickened. Abaxial epidermis:

bulliform cells absent or only very small; situated on

the sides and bases of furrows. Costal epidermal cells

with distinct, thick cuticle on outer wall; thickness of

cuticle and outer wall being equal to, or greater, than

depth of epidermal cell. Interlocking prickles with

very short barbs at sides of costal zones.

Abaxial epidermis

Prickles: short, hook-like base; barbs short (±0,05

mm), developed from apex of base; all barbs point

irregularly in direction of leaf apex. Silica-bodies :

Tall and narrow to square; outline smooth, regular

(Fig. 12).

Specimens examined

Cape. — 3019 (Loriesfontein): Handelkraal, E.N.E. of Lories-

fontein (-CD), Dekkers s.n. (PRE, iso). 3022 (Carnarvon):

Brandewynskuii, between Vosburg and van Wyksvlei (-DB),

Erasmus s.n.

Fig. 13. — Distribution of Dregeochloa pumila and D. calviniensis

compiled from specimens at the National Herbarium,

Pretoria (PRE). X indicates localities of specimens studied

anatomically.

Discussion

This rare, perennial species is known only from the

above two localities and collections. At Handelkraal,

on the margin of the Western Mountain Karroo, it

is locally frequent on limestone. In contrast to

D. pumila , this species forms small, dark green, dense

and erect tufts from 15-25 cm high.

Vegetatively D. calviniensis bears little resemblance

to D. pumila, but the leaf anatomy is strikingly

similar as seen in transverse section. Small differences

occur in the shape and size of the bulliform cells

(this may have been accentuated by the fact that only

herbarium specimens were studied), the size and thick-

ening of the epidermal cells and the prickles and

possibly the arrangement of the different orders of

R. P. ELLIS

213

vascular bundles. The abaxial epidermis, however,

shows important differences, especially in the shape

of the silica bodies. Although D. calviniensis has tall

and narrow silica bodies (Fig. 12) and D. pumila

circular or oblong bodies (Figs. 4 and 6), both these

shapes, and the fact that the silica cells and short

cells are solitary or paired, and not in long rows, are

typical festucoid characters (Metcalfe, 1960). This

difference, although striking, can thus be expected

even in two closely related taxa.

There appears to be a specialization and a centri-

petal concentration of chloroplasts in the parenchyma

sheath cells of both D. calviniensis specimens examined

(Figs. 9 and 10) and it is unfortunate that no fresh

material was obtained to confirm this. In addition, the

mesophyll, which is also poorly preserved, appears to

consist of two layers of more or less radially arranged

chlorenchyma cells. This configuration is typical of

the Kranz syndrome but the 12C/13C ratios are

typical non-Kranz (<5=-26, 32; Dekkers s.n.) as are

those for D. pumila (<5=-25, 93; de Winter & Giess

6235). The possibility that D. calviniensis represents

a very early stage in the aquisition (or late stage in

the loss) of the Kranz syndrome still exists, and this

warrants further investigation. If this is in fact the

case then the relationship of Dregeochloa to Asthena-

therum — a Kranz genus in the Danthonieae — becomes

much clearer.

DISCUSSION AND CONCLUSIONS

In the past the leaf anatomy of this genus has, to

a large extent, been misinterpreted. Thus, de Wet

(1954) grouped D. pumila with the species now inclu-

ded in the genus Asthenatherum — all of which have

“chloroplasts localized around the bundles.” This,

together with morphological criteria, he interpreted

as indicating that D. pumila belonged to the genus

Asthenatherum. Jacques-Felix (1962) stated that D.

pumila had typical panicoid structure and also included

this species with Asthenatherum. Conert (1962) noted

that D. pumila did not have elongated radially

arranged mesophyll cells. His observations led him to

disagree with the above authors and he suggested

that D. pumila occupied an isolated position in the

Danthonieae near Danthonia.

In a later paper, de Wet (1960) grouped D. pumila

with those species where the chlorophyll is not loca-

lised around the bundles but uniformly distributed

throughout the mesophyll and which possess a thin-

walled outer bundle sheath. These new observations

led him to postulate that D. pumila formed a direct

link between Danthonia and Asthenatherum. He was

still of the opinion that in all other respects these

two genera resembled one another very closely but

Conert (1962) noted that the formation of the callus

differed in addition to anatomical differences. This

influenced Conert’s (1962) proposal of an isolated

position for Dregeochloa ( =D . pumila ).

Certain epidermal features have also been confused.

De Wet (1954) described the silica bodies of D. pumila

as being spherical and stated that micro-hairs were

absent. In the 1960 publication de Wet described the

silica bodies as dumbell-shaped and micro-hairs were

observed. In the present study no micro-hairs were

found but it must be stressed that, in this genus, the

intercostal zones are difficult to preserve and study

due to the deep adaxial furrows. On all preparations of

D. pumila the silica bodies are horizontally elongated

and oblong in shape although some silica bodies on

Giess and van Vuuren 653 (Fig. 8) have indentations

in the centre and could be described as being dumbell-

shaped.

The observations of this study, based solely on leaf

anatomy, confirm that these two species closely

resemble one another and constitute a unique, dis-

tinct group. Their structure is unique amongst the

Danthonieae and they show little anatomical resem-

blance to any other South African members of this

tribe. This supports Conert’s (1966; 1971) conclusions

that this genus occupies an isolated position within

the tribe. However, Dregeochloa has certain similari-

ties in common with Asthenatherum (abaxial fur-

rows and prickles, spikelet structure and distribution)

which may indicate that these two genera had a com-

mon ancestor which gave rise to Kranz and non-

Kranz groups. The region where these genera occur

is a zone where Kranz genera ( Aristida , Stipagrostis,

Eragrostis etc.) and non-Kranz genera ( Schismus ,

Karoochloa, Chaetobromus and Merxmuellera ) are

equally abundant. Dregeochloa shows no anatomical

affinities with the genus Merxmuellera {— Danthonia )

and its removal from Danthonia appears justified.

ACKNOWLEDGEMENTS

The author is indebted to Miss L. Breytenbach

and Miss A. Vermeulen for capable technical assis-

tance and to Dr J. Vogel and Mrs M. Greyling of the

Natural Isotopes Division, N.P.R.L., C.S.I.R. for

the 13C/12C determinations.

U/TTREKSEL

Die anatomiese struktuur van die blaar in dwarssnee

en die abaxiale epidermis van Dregeochloa pumila en

D. calviniensis word beskryf en geillustreer. 'n Genus

beskrywing is ingesluit en die verwantskappe van die

genus word kortliks bespreek.

REFERENCES

Chippindall, L.K.A., 1955. In Meredith, The grasses and

pastures of South Africa. Johannesburg: C.N.A.

Conert, H. J., 1962. Uber die Gramineen-Gattung Asthena-

therum Nevski. Senckenberg Biol. 43:239-266.

Conert, H. J., 1966. Dregeochloa, eine neue Gattung der

Gramineen. Senckenberg Biol. 47:335-343.

Conert, H. J., 1971. The genus Danthonia in Africa. Mitt. Bot.

Staatssamml. MiXnchen. 10:299-308.

De Wet, J. M. J., 1954. The genus Danthonia in grass phylogeny.

Am. J. Bot. 41:204-211.

De Wet, J. M. J., 1956. Leaf anatomy and phylogeny in the tribe

Danthonieae. Am. J. Bot. 43:175-182.

De Wet, J. M. J., 1960. Leaf anatomy and morphology in

South African species of Danthonia. Bothalia 7:303-310.

Jacques-Felix, H., 1962. Les Graminees d’ Afrique Tropicale.

I. Generalites, Classification, Description des Genres.

Inst, de Recherches Agronomiques Tropicales et des Cultures

Vivrieres, Paris. 1-345.

Metcalfe, C. R., 1960. Anatomy of the Monocotyledons. I.

Gramineae. Oxford: Clarendon Press.

Bothalia 12,2: 215-221 (1977)

Cytogenetic studies in the Eragrostis curvula Complex

T. B. VORSTER* and H. LIEBENBERGf

ABSTRACT

Cytogenetic studies were undertaken in the Eragrostis curvula Complex. Three plants were studied at each

of 16 collecting points. The overall morphology and embryo sac development of all plants were evaluated, while

the chromosome number and microsporogenesis of some of the plants were also studied. The collecting points

were chosen so ^s to represent a variable environment extending from the bushveid to the highveld regions of

the Transvaal. It was found that the embryo sac development of the plants from the bushveid and the highveld

were, for all practical purposes, obligate diplosporic apomicts, whereas the transition area contained obligate as

well as facultative diplosporic apomicts. The same pattern also held as far as the plant morphology, chromosome

number and microsporogenesis were concerned.

RESUME

ETUDES CYTOGENETIQUES DANS LE COMPLEXE ERAGROSTIS CURVULA

Des etudes cytogenetiques out ete entreprises dans le complexe Eragrostis curvula. Trois plants

ont ete etudies pour chacun des 17 points de recolte. La morphologic generate et le developpement du

sac embryonnaire ont ete evalues partout, tandis que le nombre chromosomique et la microsporogenese

de certains plants ont aussi ete etudies. Les points de recolte ont ete choisis de maniere d representer

un milieu variable s'etendant des steppes boisees ("bushveid") aux prairies de montagnes ("highveld")

du Transvaal. On a constate que, dans les steppes comme dans les prairies, le developpemcnt du sac

embryonnaire comporte obligatoirement line apomixie diplosporique, pour tout ce qtT on pent en obser-

ver; tandis que les zones de transition presentent aussi bien des apomictes diplosporiques obligatoires

que des facultatifs. La me me distribution s'observe aussi en ce qui concerne la morphologic de la plante,

le nombre chromosomique et la microsporogenese.

INTRODUCTION

Cytogenetic studies on E. curvula (Schrad.) Nees

and its close relatives have made it clear that they are

probably all part of a large agamic complex. The

mechanism of apomixis is diplospory, as defined by

Gustafsson (1946), followed by parthenogenesis

(Hakansson, 1943) and pseudogamy as the main

means of reproduction (Brown and Emery, 1958;

Liebenberg, 1961; Liebenberg and Pienaar, 1962;

Streetman, 1963; Voight, 1971; Voight and Bashaw,

1972 and Brix, 1974). This is probably the reason

why the taxonomy of this complex has been so

difficult to unravel.

The E. curvula group has a wide distribution

throughout Southern Africa and the existence of

polymorphism within species and intermediate types

between species led De Winter (1955) to treat E.

robusta Stent as a synonym of E. curvula and to

maintain E. chloromelas Steud. as a closely related

species in what is generally known as the E. curvula

Complex.

Unfortunately most of the cytogenetic studies

undertaken on E. curvula and its relatives have been

carried out, for understandable reasons, on those

sections of the complex that are economically

important in pasture breeding. Inevitably these

sections might well comprise robust-growing high

polyploid apomicts. This has created the impression

that the complex was an old one consisting of mostly

obligate apomicts.

It was considered therefore that cytogenetic studies

(especially embryo sac studies) of a naturally growing

and unselected E. curvula Complex population might

be well worth studying in an attempt to elucidate

this problem.

MATERIAL

A region was selected in the Transvaal (South

Africa), which included highveld, a transition area and

lower bushveid. Seventeen collecting points were

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria,

t Department of Genetics, University of Pretoria.

selected with three collections at each point in order

to evaluate variation among as well as within

collecting points over a reasonably varied area

(Fig. 1).

Collecting points were selected at regular intervals

on the lower bushveid and highveld, while those over

the transition area were selected (with the help of

Fig. 2) so that the bottom, slopes and summits of

ridges would be included.

Herbarium specimens of each collection were

made and identified at the National Herbarium in

Pretoria, where they are now housed (Table 1).

It was decided that a minimum of fifty embryo sacs

should be investigated for every plant. Additional

embryo sacs of a specific plant were thus sectioned or

new plants were collected when this number was not

available from the collected material. More plants

were therefore collected than were actually used in

the study. All plants collected are reflected in Table 1

and given herbarium numbers. Those plants used

in this study were, however, given additional numbers

(Plant Numbers — Table 1) to identify those plants

originating from one collecting point.

METHODS

The inflorescences for microtome studies were

fixed in Navashin fixative (Stockholm modification —

Maheshwari, 1939). Inbedding and staining techniques

outline in Johansen (1940) were followed using

Heidenheins Haematoxylin as nuclear stain and

Orange G as counter stain.

The propionic carmine squash method of Pienaar

(1955) was slightly modified for the study of micro-

sporogenesis and chromosome numbers.

RESULTS

Although the study of embryo sac development was

the main object ot the present work, micro-

sporogenesis and chromosome numbers were also

investigated as far as possible. One big problem was,

that by the time that microsporogenesis could be

investigated (only after the embryo sacs were studied),

the material had been left in the fixative too long.

HEIGHT ABOVE SEA LEVEL IN METRES

216

CYTOGENETIC STUDIES IN THE ERAGROSTIS CURVULA COMPLEX

JOHANNESBURG

Fig. 1. — Map of the roads between Brits and Johannesburg indicating the collecting points A to Q as well as the distance

kilometres between them; collecting points A to E are defined as the lower bushveld. F to M as the transition area and N to

as the highveld.

DISTANCE IN KILOMETRES

Fig. 2. — A graph indicating the collecting points with regard to the height above sea level and distance.

05'

T. B. VORSTER AND H. LIEBENBERG

217

Fig. 3. — Microsporogenesis stages of some of the collections. A: Diakineses of a rather “regular”

tetraploid with a few distinct heteromorphic bivalents (H) -collection L-3. B: Anaphase I of a rather “regular”

hexaploid -collection H-2. C: Anaphase I of a rather “regular” hexaploid (two laggards with one undergoing

chromatid segregation) as well as one paracentric inversion bridge (P) -collection J-l. D: Anaphase I of a very

abnormal hexaploid (approximately 12 laggards, some of them undergoing chromatid segregation, can be

seen) -collection G-3.

218

CYTOGENETIC STUDIES IN THE ERAGROSTIS CURVULA COMPLEX

Because of the small and compact spikelets of

Eragrostis, fixative penetration is poor with the result

that the material deteriorates soon after fixation.

Analyses were made of the number of monovalents

and laggards during metaphase I and anaphase I

respectively and are recorded with the chromosome

numbers in Table 2. The quality of the slides did not

permit multivalent or any other more complex

analyses.

Paracentric inversion bridges (Fig. 3C) were also

encountered in collections C-2, C-3, H-3 and J-l.

These bridges, however, occurred at a very low

frequency (except for plant J-l which had chromatid

bridges in 16% of the anaphase I cells studied). Apart

from indicating that the polyploids are probably

segmental-allopolyploids, this information was con-

sidered too scanty to be of any further use. Another

phenomenon that was conspicuous in a few

collections, e.g. L-3, and that would support the

segmental-allopolyploid nature of the plants was

the occurrence of heteromorphic bivalents, where

one chromosome was much larger than its partner

(Fig. 3A).

The meiotic analysis showed that plants with low

chromosome numbers tend to have “regular” meioses,

whereas higher polyploids were inclined to have

abnormal meioses although some variation occurred.

A few groups could roughly be distinguished as

regard chromosome number, microsporogenesis

abnormalities and the locality from whence they came:

(1) Hexaploids with abnormal meisoses (all these

plants are from the lower bushveld and the

beginning of the transition area) (Fig. 3D).

(2) Hexaploids with rather “regular” meioses (these

plants are from the transition area and the

beginning of the highveld) (Fig. 3B).

(3) An octoploid with a relative “regular” meiosis

(from the transition area).

(4) A pentaploid with, as expected, a very abnormal

meiosis (from the transition area).

(5) Two tetraploids with relatively “regular”

meioses (from the transition area).

A total of 3 902 embryo sacs was examined and the

results are summarized in Table 1. It was found that

3 306 of these embryo sacs were apomictic by means

of diplospory followed by parthenogenesis and

pseudogamy while 99 were sexual monosporic Poly-

gonum type (Maheshwari, 1950) embryo sacs. One

hundred and nineteen embryo sacs were abnormal or

divergent, while 378 were degenerated (Table 1).

DISCUSSION

Four factors were considered in attempting to

evaluate the natural variation in the E. curvula

Complex in the area of this study.

1 . Morphological variation in collected material

It was quite striking that only true E. curvula was

collected on the lower bushveld as well as the high-

veld, while true E. curvula as well as types which

could not be judged as true were collected in the

transition area (Table 1).

All the E. chloromelas types (true and not true

types) were collected in the transition area. This

indicates a great deal of variation, not only between,

but also within collection points in the transition

area.

2. Variation in chromosome numbers

The E. curvula Complex has a wide range of

chromosome numbers, i.e. 2n=20, 40 and 60 (Pienaar,

1953) and 2n = 50 (De Wet, 1954) for E. curvula,

2n = 40 and 60-63 (Pienaar, 1953) for E. chloromelas

and 2n-=70 and 80 (Pienaar, 1953) for E. robusta.

In this study an additional number of 2n = 80 was

found for E. curvula in the transition area (Table 1,

plant H-3).

Although chromosome counts for only a few

plants were made, it is clear that the numbers in the

transition area are very variable (2n=40, 50, 60 and

80). The few plants in the lower bushveld and highveld,

which were studied in this respect, all had 60

chromosomes and, although it is risky to draw any

definite conclusions, it seems as if there might be a

greater chromosome number variation in the

transition area.

3. Variation in microsporogenesis

Table 2 shows clearly that the meioses of plants are

inclined to become more regular as the transition

area is approached from the lower bushveld. The

two hexaploids on the highveld have rather “regular”

meioses, but it would again be dangerous to draw

any conclusions as to the significance of this

phenomenon, because of the low number of plants

investigated.

4. Variation in embryo sac development

Considerable difficulty was at first experienced

with the classification of some of the sexual embryo

sacs and, because it was important to calculate the

exact degree of sexuality for every plant, this problem

has been attended to in some detail (Vorster &

Liebenberg, in press).

Sexual embryo sacs were sporadically found over

the whole sampling area but they occurred at a higher

frequency in the transition area (Table 1). Not only

was there variation in the embryo sac development

over the area as a whole, but also within collecting

points.

This occurrence of a high percentage sexuality and

obligate apomixis within one collecting point stresses

the discontinuous variation that occurs in the embryo

sac development of the complex.

CONCLUSIONS

The method used in this study to try and evaluate

the natural variation in embryo sac development in

a specific area, proved to be very satisfactory,

especially because it showed that previous studies

did not reflect a true picture of the occurrence of

sexuality in this complex. It would seem that the

group is far from being an old obligate agamic

complex. Since several specimens in the transitional

zone were E. chloromelas types and showed, together

with one E. curvula type, fairly high sexuality, it may

be that E. chloromelas is supplying the source of

sexuality in this area.

Sexual embryo sacs occurred most frequently in

the transition area. Microsporogenesis studies showed,

furthermore, that the meioses were inclined to

become more regular as the transition area was

approached from the lower bushveld, while the

transition area also contained plants with the most

variable morphology as well as the most variable

chromosome numbers.

From these results, one seems justified in expecting

that one or more of the original sexual ancestors of

this agamic complex may still be present in or near

the transition area (east or west from it) as sampled

by this study. According to Stebbins (1950) the

variation pattern in the Crepis Complex indicated

that in the vicinity of diploid sexual ancestors,

variable facultative apomicts occurred, whereas

obligate apomics with great morphological uniformity

T. B. VORSTER AND H. LIEBENBERG

219

TABLE 1. - Number and percentage embryo sac types found for the different specimens

220

CYTOGENETIC STUDIES IN THE ERAGROSTIS CURVULA COMPLEX

x

Diakinesis monovalents

T. B. VORSTER AND H. LIEBENBERG

221

were found far from them. The diploid E. curvula

reported by Pienaar (1953) as well as the diploid

sexual found by Voight (1971) in the variety conferta

lends further evidence that at least some of the diploid

sexual ancestors of the complex still exist.

In conclusion, it is clear from this study, however

limited its scope, that the E. curvula Complex is

inadequately known and understood. If any sense

is to be made out of its problematic morphological

variation pattern, it is clear that a complete cyto-

genetic study is called for. Such a study should

include all related types and even species, over as

much of their distribution area as possible. The

present authors have already begun this task and

it is hoped that in this way the relationships and

therefore the taxonomy of the complex will be

unravelled or at least better understood.

UITTREKSEL

Sitogenetiese studies is onderneem op die Eragrostis

curvula kompleks. Daar is drie plante by elk van die

17 versamelpunte geneem en bestudeer. Hierdie plante

is geevalueer ten opsigte van hul klassifikasie (morfo-

logie) en kiemsakontwikkeling. Sommige van die

plante se mikrosporogenese en chromosoomaantal kon

ook bestudeer word. Die versamelpunte is uitgekies

om 'n variabele omgewing, wat strek vanaf ft laer-

liggende bosveld na die hoeveld, te verteenwoordig.

Dit is gevind dat die kiemsakontwikkeling van die

plante afkomstig vanaf die laerliggende bosveld sowel

as die hoeveld vir alle praktiese doeleindes verpligte

diplosporiese apomikte is, terwyl die plante afkomstig

vanaf die oorgangsgebied verpligte sowel as fakultatiewe

diplosporiese apomikte opgelewer het. Dieselfde patroon

kon waargeneem word ten opsigte van die plantmorfolo-

gie, chromosoomaantal en mikrosporogenese .

REFERENCES

Brix, K., 1974. Sexual Reproduction in Eragrostis curvula

(Schrad.) Nees. J. Plant Breeding 71 : 25-32.

Brown, W. V. & Emery, W. H. P., 1958. Apomixis in the

Gramineae: Panicoideae. Amer. J. Bot. 45: 253-263.

De Wet, J. M. J., 1954. Chromosome numbers of a few South

African grasses. Cyto/ogia 19: 97-103.

De Winter, B., 1955. Eragrostis Beauv. In Meredith, Grasses

and pastures of South Africa. Johannesburg: Central News

Agency. 132-184.

Gustafsson, A., 1946. Apomixis in higher plants. I. The

mechanism of apomixis. Lunds. Univ. Arsskr. N.F. Avd.

„ 2. 42: 1-66.

Hakansson, A., 1943. Die Entwicklung des Embryosacks und

die Befruchtung bei Poaalpina. Hereditas 29: 25-61.

Johansen, D. A., 1940. Plant microtechnique. New York: Me

Graw-Hill.

Liebenberg, H., 1961. Apomiksie by Eragrostis en Themeda.

Ongepubliseerde M.Sc. verhandeling, Universiteit van

Stellenbosch.

Liebenberg, H. & Pienaar, R. de V., 1962. Apomiksie by

Eragrostis en Themeda. Verrigtinge van die 2de Kongres

van die Suid-Afrikaanse Genetiese Vereniging. 150-153.

Maheshwari, P., 1939. Recent advances in microtechnique. II.

The paraffin method Cytologia 10: 251-281.

Maheshwari, P., 1950. An introduction to the embryology of

the angio sperms. New York: Me Graw-Hill.

Pienaar, R. de V., 1953. Cyrological studies in some South

African species of the genus Eragrostis. Unpublished Ph.D.

thesis, University of the Witwatersrand.

Pienaar, R. de V., 1955. Combinations and variations of

techniques for improved chromosome studies in the Grami-

neae. J.S. Afr. Bot. 21 : 1-8.

Stebbins, G. L., 1950. Variation and evolution in plants. New

York: Columbia Univ. Press.

Streetman, L. J., 1963. Reproduction of the lovegrasses, the

genus Eragrostis , i.e. E. chloromelas Steud., E. curvula

(Schrad.) Nees, E. lehmanniana Nees and E. superba Peyr.

Wrightia 3 : 41-51.

Voight, P. W., 1971. Discovery of sexuality in Eragrostis

curvula (Schrad.) Nees. Crop Sci. 1 1 : 424—425.

Voight, P. W. & Bashaw, E. C., 1972. Apomixis and sexuality

in Eragrostis curvula. Crop Sci. 12: 843-847.

Bothalia 12,2: 223-224 (1977)

A note on the flowers of Halleria lucidci

C. H. STIRTON*

ABSTRACT

Studies of sunbirds {Cinnyris spp.) feeding on the nectar of flowers of Halleria lucida L. suggest that

partial protandry may be operative in the breeding system of this cauliflorous tree. Attention is drawn to

certain anomalies depicted in published botanical drawings. These anomalies are discussed in relation to the

sequential development of the androecium and gynoecium in live flowers. Colins colius , the speckled coly, is

reported to eat the fruits of Halleria lucida. This bird also feeds on nectar after piercing the base of the corolla

tube.

RESUME

NOTE SUR LES FLEURS D’HALLERIA LUCIDA

Des observations sur la consommation de nectar par des souimangas du genre Cinnyris ex ploitant

les fleurs rf’Halleria lucida L. suggerent qu'une protandrie partielle pent intervenir dans le systems

reproducteur de cet arbre cauliflore. On attire T attention sur certaines anomalies figurant dans des

dessins botaniques qui ont ete publies et on discute ces anomalies par rapport au developpement sequen-

tiel de Tandrocee et du gynecee dans des fleurs vivantes. Colius colius, la veuve tachetee, est signale

comme mangeant les fruits c/'Halleria lucida. Cet oiseau se nourrit egalement de nectar apres avoir

perce la base du tube de la corolle.

Halleria lucida L. is one of the few cauliflorous

trees in South Africa (Marloth, 1932). The massed

red tubular flowers are visited regularly by various

species of sunbirds ( Cinnyris ) and sugarbirds

( Anthobaphes ) and this has led to the belief that

pollination is ornithophilous (Marloth, 1932. Vogel,

1954). This note has arisen from observations of

Cinnyris spp. feeding on the flowers of a solitary tree

of H. lucida growing in the National Botanical

Gardens, Pretoria.

A close inspection of massed flower clusters revealed

that each cluster was composed of flowers at different

stages of development. Sequential dissections from

bud stages to fully mature flowers suggested that

partial protandry might operate as an outbreeding

mechanism. It was found that the didynamous

stamens expanded at different rates and that the

stigma was always below the lowest pair of these

stamens until their anthers had dehisced. This

sequential development is shown in Fig. 1. Prior to

the opening of the corolla lobes all four anthers

were held at the same level, even though the filaments

of the two pairs were adnate to the corolla tube at

different levels. Steps 1-5 in Fig. 1 show the emergence

of the anthers from the corolla mouth. The outer

pair of anthers (deepest inserted) developed faster

than the inner pair of anthers, so that by the time

the inner pair had dehisced the outer pair had

completely shed their pollen. The style expanded

rapidly once dehiscence of the inner pair of anthers

had begun and eventually either overtopped or

attained the level of the outer pair of anthers. It was

difficult to establish accurately the time of receptivity

of the stigma, but it appears to have occurred once

the stigma had reached the level of the inner pair of

dehisced anthers.

Any interpretations of the phenomenon discussed

above must be made cautiously for a number of

reasons. Firstly, different authors have commented

on different flower types in Halleria lucida and

secondly, different patterns of development have

been reported for the androecium and gynoecium.

Sims (1815), who commented on a rudimentary fifth

filament, compared this species to the genus Lonicera

(Caprifoliaceae) in which Burman (Rar. Afric. 244,

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

t. 89, fig. 2) had placed it. Verdoorn (1946) drew

attention to anomalous flowers and commented as

follows: “While the majority of flowers on this

shrub are funnel-shaped with an oblique, somewhat

2-lipped mouth, several were found to be trumpet-

shaped with the lobes equally spreading and 4 or 5

in number. In these examples the stamens were also

equally disposed and not unilateral.” Flowers observed

in Pretoria approximated those in t. 961, fig. 4

(Flow. PI. Afr., 1946) of Verdoorn’s article. No

abnormal flowers were found. Vogel (1954), however,

has commented that the stamens and stigma hardly

exsert from the corolla mouth and his Fig. 132: 4

closely approximates t. 961, fig. 8 described as an

abnormal flower by Verdoorn. It is difficult to assess

from these drawings, as well as from the t. 1744

(Bot. Mag. 41, 1815) described by Sims, whether the

relative position of anthers and stigma can be ascribed

to the artist having drawn flowers at different stages

of development. If the flowers drawn were indeed

mature, it must still be decided whether such flower

types are abnormalities or whether they are normal

and may possibly play an important role in the

pollination strategy of the plant. This will require

further field studies.

My assessment of partial protandry is derived

from the observations outlined above. If this is

correct, it would mean that a bird flying from flower

to flower could conceivably transfer pollen from a

flower with only one pair of anthers dehisced to a

flower in which the stigma was completely exerted.

This would constitute a measure of outbreeding if

more than one plant was involved. On the other

hand, the late emergence of the second pair of anthers,

with their dehiscence occurring as it does with the

elongation of the stigma, could guarantee inbreeding.

Limited emasculation and bagging studies of flower

clusters on the Pretoria plant indicated that autogamy

was operative.

An important feature of this cauliflorous tree is

that the flower clusters, consisting as they do of

flowers at all levels of development, ensure a constant

visitation over a long period by potential pollinators.

As Vogel (1954) has indicated, the flowers are rich

in nectar, but non-odoriferous. The fruits of this

plant are relished by the speckled coly, Colius colius,

which also pierces the base of the corolla tube to

feed on the nectar.

Fig. 1. — Sequential developmental stages of the flower of Halleria lucida once the corolla mouth has opened. 1, outer

pair of stamens beginning to emerge from mouth, note stigma well below inner pair of stamens; 2, outer pair of

stamens emerged from corolla mouth; 3, right hand anther of outer pair has begun dehiscence, while the inner

pair has begun to emerge; 4, and 5, the outer pair has dehisced, the inner pair has emerged, as has the stigma;

6, inner pair has dehisced; 7-9, the stigma gradually elongates until it reaches the upper pair of anthers or overtops

them.

An interesting feature of the dehisced anthers is

that they are completely flattened, with the pollen-

shedding surface facing downwards (i.e. towards the

viewers of Fig. 1). This provides a maximum surface

area to brush against any bird’s bill or tongue inserted

into the flower.

Detailed studies on a tree growing in the

Kirstenbosch Botanical Gardens in Cape Town gave

the same results as the studies carried out in Pretoria.

UITTREKSEL

Waarnemings op suikerbekkies (Cinnyris spp.) wat

op die nektar van blomme van Halleria lucida L. voed,

skep die indruk dat gedeeltelike protrandrie werksaam

mag wees in die voortplantingstelsel van hierdie boom,

waar die blomme op die stam gedra word. Aandag word

gevestig op sekere afwykings wat aangedui is in

gepubliseerde plantkundige tekeninge. Hierdie

afwykings word bespreek in verhouding tot die opeen-

volgende ontwikkeling van die androecium en gynoecium

in lewende blomme. Dit word berig dat Colius colius,

die gespikkelde muisvoel, die vrugte van Halleria lucida

eet. Hierdie voel voed ook op nektar nadat dit die basis

van die kroonbuis deurboor het.

REFERENCES

Marloth, R., 1932. The flora of South Africa. 3,1: 133-134,

t.35. Cape Town: Darter.

Sims, J., 1815. Halleria lucida L. Bot. Mag. 41 : 1. 1 744.

Verdoorn, I. C., 1946. Halleria lucida L. Flow. PI. Afr. 25:

t.96 1 .

Vogel, S., 1954. Bliitenbiologische Typen als Elemente der

Sippengliederung, pp. 67, 242. Jena: Veb. Gustav Fischer

Verlag.

Bothalia 12,2: 225-227 (1977)

The pollination of Cancivalia virosa by Xylocopid and

Magachilid bees

C. H. STIRTON*

ABSTRACT

The floral morphology of Canavalia virosa (Roxb). Wight & Arn. is discussed in relation to pollination by

Xylocopa flavorufa De Greer and Megachile combusta Sm. It was found that the relationship between size of

flower and bee influenced the type of pollinating strategy and its success. Bees smaller than M. combusta proved

ineffective pollinators.

RESUME

LA POLLINISA TION DE CANAVALIA VIROSA PAR LES ABEILLES XYLOCOPIDES ET

MEGA CHILI DES

La morphologie florale de Canavalia virosa (Roxb.) Wight & Arn. est discutee en rapport avec

sa pollinisation par Xylocopa flavorufa De Greer et Megachile combusta Sm. On a constate que la

relation entre dimensions de I'abeille et de la fleur influence la technique de pollinisation et son succes.

Des abeilles plus petites que M.combusta se sont averees inefficaces pour la pollinisation.

INTRODUCTION

Canavalia virosa (Roxb.) Wight & Arn. has been

commonly known in South Africa as Canavalia

ferruginea Piper. Although often confused with

C. ensiformis (L.) DC. and C. glochidiata (Jacq.) DC.,

it is usually kept separate (Sauer, 1964, Verdcourt,

1971). However, it has been recently suggested that

these three species may belong to one species. C.

ensiformis, which includes wild and cultivated types

(Westphal, 1974). The existence of this complex is

not surprising if one considers the widespread

cultivation of these plants in modern times as

vegetables, fodder and cover crops. C. virosa is used

here in the sense of Westphal (1974).

GEOGRAPHICAL DISTRIBUTION

C. virosa extends southwards from Arabia, Socotra

and India through tropical Africa into north-east

South Africa (Westphal 1974). Its distribution south

of the Limpopo River is shown in Fig. 1.

FLORAL MORPHOLOGY

The general morphology of C. virosa has been

adequately reviewed by Piper & Dunn (1922), Sauer

(1964), Verdcourt (1971), and Westphal (1974).

Piper et al. (1922) and Scott Elliot (1891) have written

generalized accounts of the structural relationships

of the various flower parts of C. maritima (Aubl.)

Thours. and C. ensiformis (L.) DC. respectively.

Vogel (1954) drew attention to analogies between the

flowers of C. maritima and Salvia africana L. and

also discussed pollination in both species. No account

appears to have been given of the pollination of

C. virosa. However, since, as Sauer (1964) has

commented, Canavalia spp. generally have mono-

tonous corollas and broad flowering schedules, it is

to be expected that the genus is probably minimally

adapted to specific pollinators. This implies that

pollination is similar in most cases.

The flower structure of C. virosa hardly differs

from the majority of known Canavalia spp. Flowers

are generally held in an upright position with the

standard facing downwards (e.g. Fig. 4 in Westphal,

1974). But, as Piper et al. (1922) pointed out, the

flowers are commonly inverted i.e. standard below

and facing upwards (e.g. Fig. 2 in Vogel, 1954). So

far only inverted flowers have been observed in the

* Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

Transvaal. A front view of such a flower is shown

in Fig. 2: 1. An outstanding feature of this flower

is the longitudinal bulging along the upper surfaces

of the two wings (arrowed in Fig. 2: 1). This feature

plays a prominent role in pollination and it is

surprising that very little attention has previously

been drawn to it. In Fig. 2: 2 and 3 the bulges are

arrowed relative to their position to the standard (S)

and keel (K). The two peglike callosities at the base

of the standard (S) rest on the portion of wing tissue

between the bulges and auricles of the wings.

The auricles of the wings articulate with the

standard at the point marked * in Fig. 2: 4. This

articulation allows the keel and wings of both upright

and inverted flowers to return to their original

position after being depressed by a visiting insect.

Fig. 1. — Known distribution of Canavalia virosa south of the

Limpopo River (based on collections housed at the National

Herbarium, Pretoria).

226

THE POLLINATION OF CANAVALIA VIROSA BY XYLOCOPID AND MEGACHILID BEES

3

Fig. 2. — Canavalia virosa

showing various views of

the flower: 1, front view

of flower with an arrow

indicating the opening

caused by the bulging of

the upper wing margins;

2, side view of flower

with one wing and keel

blade removed to show

relative position of stan-

dard, wing and keel; 3,

enlargement of centre

portion of 2 showing

wing pulled aside from

the upper margin of the

keel; 4, side view of

opening flower to show

relative position of wing

and standard auricles.

(K. = keel, S = standard,

W = wing). Arrows indi-

cate longitudinal bulging

of upper wing margins.

POLLINATING VECTORS AND THEIR STRATEGIES

In 1964 Sauer wrote that the available information

on pollinating insects in Canavalia was meagre.

Xylocopid bees have been reported to pollinate

C. maritima e.g. Xylocopa violacea (Scott— Elliot, 1891)

and X. aestuans (Piper et al. 1922). Vogel (1954)

reported that Anthophora sp. also pollinates this

species. Lepidoptera may also visit C. maritima

(Scott— Elliot, 1891).

Recent field studies in the Nelspruit District of the

eastern Transvaal, revealed that only tw'o types cf

bees were able to depress the keel of C. virosa

effectively. These bees differed markedly in size and

adopted different pollinating strategies.

Xylocopa flavorufa De Greer (Fig. 3: 1) is a noisy

fast-flying solitary bee found in coastal bush, montane

savanna woodland, fynbos and dry savanna woodland

(Watmough, 1974). Both male and female bees were

caught feeding on the flowers of C. virosa in the

Nelspruit Botanical Gardens. The bee lands on the

inverted standard with some force causing the

whole inflorescence to shudder. This insect fits snugly

into the area presented by the standard and maintains

a hold by grasping the standard mainly near the

peg-like callosities. Its head lies near the opening

formed by the longitudinal bulges of the wings of the

flower (Fig. 4: 1). To trip the mechanism the bee

raises its abdomen and thorax and then thrusts its

head against the opening. This movement forces the

keel and wings away from the staminal sheath thereby

exposing the stigma and anthers, which dust the back

of the bee’s thorax. This sequence is identical to

that shown in Fig. 2 of Vogel (1954) in which he

depicts Anthophora sp. visiting flowers of C. maritima.

Megachile combusta Sm. (Fig. 3: 2) is a smaller

and lighter bee than Xylocopa flavorufa. It is incapable

of landing and maintaining its position as does

X. flavorufa. This bee lands more or less on the side

of the standard and after repeatedly slipping on the

shiny area above the callosities, it eventually

manoeuvres its body into a position that brings its

abdomen into line with the tip of the keel (Fig. 4: 2).

When it pushes its head against the opening (Fig.

2: 1) it depresses the keel and wings and causes its

abdomen to brush against the exposed stigma and

anthers. The wings and keel retract when the insect

alights.

Apis mellifera L. and other bees smaller than

M. combusta were common visitors to flowers of

C. virosa. These bees were unable to depress the keel.

They collected loose pollen and in flowers damaged

by beetles that fed on nectar. Lepidoptera were

common visitors.

DISCUSSION

In his discussion of the process of speciation in

Canavalia Sauer (1964) noted that almost nothing

was known about pollination or sterility barriers

that could protect the divergent adaptations of

subpopulations of diverging species. From his

observations that isolated greenhouse plants could

set seed, he deduced that facultative self-pollination

1

t i

1 cm

2

i 1

1 cm

Fig. 3. — Photographs of bees. 1, Xylocopa flavorufa ; 2,

Megachile combusta.

C. H. STIRTON

227

Fig. 4.— Positions adopted by X. flavorufa (1) and M. combusta (2) when operating the pollination mechanism of Canavalia virosa.

Arrows indicate that pollen is dusted onto the thorax of X. flavorufa and onto the underside of the abdomen of M. combusta.

was possible, but he pointed out that the attractiveness

of flowers to bumble bees and butterflies indicated

cross-pollination.

Watmough (1974) noted that many indigenous

Papilionaceae in Southern Africa apparently have no

regular effective pollinators other than carpenter bees,

which alone possess the power and weight necessary

to operate the floral mechanisms. Canavalia appears,

from the available evidence, to fall into such a class

of plants. Plants of C. virosa in the eastern Transvaal

tend to set between 2-4 seed pods per inflorescence,

a factor which could indicate outbreeding as bees

did not forage in inclement weather. If inbreeding

was operative, one would expect a much higher seed

set. It may, therefore, be advantageous to breed

and release xylocopid bees in those areas where

C. virosa and C. ensiformis are being cultivated.

Watmough (1974) cites a number of references,

which indicate that the introduction of these bees

has improved the pollination and/or yield of a number

of other agricultural crops.

Of the bees studied in this paper, X. flavorufa

seems a more “natural” pollinator, whereas M.

combusta and other bees smaller than it show every

sign of being opportunists. M. combusta has

“developed” a method whereby it can utilize nectar

and pollen, but one which is not always successful.

Since few bees appear to be large enough to

pollinate C. virosa successfully, it would appear that

the large inverted flowers of this species constitute

a possible isolating mechanism. M. combusta is the

smallest known bee, which can operate the floral

mechanism. A comparative study of sympatric

Canavalia spp. with respect to pollinating vectors

could give a better insight into the mechanisms which

protect the divergent adaptations of subpopulations

of diverging species, particularly those in complexes

such as that of C. ensiformis.

ACKNOWLEDGEMENTS

My sincere thanks to Mrs E. Buitendag and

Mr E. van Jaarsveld of the Lowveld Botanic Gardens,

Nelspruit, for their help and hospitality; to Mr R. H.

Watmough of the Plant Protection Research Institute,

Pretoria, for naming the insect collection; to Miss M.

Scott for the line drawings and to Mrs A.

Romanowski for the photographs.

UITTREKSEL

Die blommorfologie van Canavalia virosa ( Roxb .)

Wight & Am. word bespreek in verhouding tot

bestuiwing deur Xylocopa flavorufa De Greer en

Megachile combusta Sm.. Dit is gevind dat die ver-

houding tussen grootte van blom en by die tipe

bestuiwingstrategie en die sukses daarvan beinvloed.

Bye kleiner as M. combusta het ondoeltreffende

bestuiwers blyk te wees.

REFERENCES

Piper, C. V. & Dunn, S. T., 1922. A revision of Canavalia.

Bull. Misc. Inf. Kew 4: 129-145.

Sauer, J., 1964. Revision of Canavalia. Brittonia 16: 106-181.

Scott Elliot, G. F., 1891. Notes on the fertilization of South

African and Madagascar flowering plants. Ann. Bot. 5,14:

332-405.

Verdcourt, B., 1971. Leguminosae 4. Papilionoideae 2. In E.

Milne-Redhead & R. M. Polhill, Flora of Tropical East

Africa. London: Crown Agents for Overseas Governments

and Administrations.

Vogel, S., 1954. Bliitenbiologische Typen als Elemente der

Sippengliederung. Jena: Veb Gustav Fischer. 338 pp.

Watmough, R. H., 1974. Biology and behaviour of carpenter

bees in Southern Africa. J. Ent. Soc. Sth. Afr. 37,2: 261—

281.

Westphal, E., 1974. Pulses in Ethiopia, their taxonomy and

agricultural significance. Belmontia, New Series 3,9:

1- 363.

Bothalia 12,2: 229-230 (1977)

Broad-spectrum pollination of Plectranthus neochilus

C. H. STIRTON*

ABSTRACT

The pollination ecology of Plectranthus neochilus Schltr. is discussed and compared with that of another

garden plant, Plectranthus barbatus Andr.. Pollinators and flower visitors of P. neochilus include members of

the Megachilidae, Anthophoridae, Syrphidae, Bombyliidae, Sphingidae, Apidae.

RESUME

EVENTAIL DE POLLINISATION DE PLECTRANTHUS NEOCHILUS

L'ecologie de la pollinisation de Plectranthus neochilus Schltr. est discutee et comparee avec cede

d'une autre plante de lardin, Plectranthus barbatus Andr. On a releve, parini les insectes qui pollinisent

et visitent les fleurs de P. neochilus, des representants des Megachilidae, Anthophoridae, Syrphidae,

Bombyliidae, Sphingidae, Apidae.

INTRODUCTION

The insect-plant relationships of plants under

cultivation are often very different from those of the

same species growing under natural conditions. The

“wild” pollinators may be absent in gardens and the

successful transference of an outbreeding plant will

then depend on either its potential for vegetative

reproduction or on the plasticity of its pollination

strategies.

This paper assesses the plant-insect relationships of

Plectranthus neochilus Schltr., an indigenous species

cultivated in gardens in the Pretoria area.

GENERAL MORPHOLOGY

P. neochilus is mostly a perennial, decumbent to

erect, often much branched and bushy (Fig. 1),

unpleasantly aromatic, 12-50 cm tall herb (Codd,

1975). Although it is variable in vegetative characters

it is relatively constant in floral characters. The inflo-

rescence is a 4-angled spikelike raceme 3-4 cm long

when in the bud stage. It is composed of 4 rows of

densely imbricate, ovate, acuminate bracts (Fig. 2A).

Flowering proceeds by the gradual opening of each

verticil of flowers. This is accompanied by the shedding

of the bracts surrounding each verticil and by the

elongation of the rhachis between the opened verticils

(Fig. 2C). Thirteen spaced verticils can be seen in

Fig. 2C, with the rhachis also cleany visible between

the verticils. Open flowers can be seen in Fig. 2B.

POLLINATION ECOLOGY

Over a period of three weeks plants of P. neochilis

were observed in the Pretoria National Botanic

Gardens, Brummeria. They cover large areas in massed

displays, a feature seldom found in the wild. These

displays are rich sources of nectar and attract

numerous insect visitors such as bees, hoverflies and

butterflies. The following insects represent the

commonest visitors:

Megachilidae Megachile maxillosa Guer.

Megachile chrysorrhoeae Gerst.

Megachile sp. (leaf cutter)

Megachile combusta Sm.

Megachile sp. cf. combusta Sm.

Anthophoridae Xylocopa caffra (L.)

Xylocopa senior Vachal

Xylocopa sichele Vachal

Anthophora sp.

Apidae Apis mellifera L.

Syrphidae . Asarcina cf. rostrata Wied.

Bombyliidae unidentified flies

Sphingidae Macroglossum trochilus (Hbn.)

Of these insects, only the bees effectively worked the

pollination mechanism every time. The syrphids,

bombylids and sphingids insert their probosci without

touching either the androecium or gynoecium.

In all cases, the bees landed on the horizontal lower

boat-shaped lip of the flower, depressing it and

exposing the stamens and stigma (which then rubbed

against the lower abdomen as the insect began

Fig. 1. — Plectranthus neochi-

lus growing among rocks

8 km S.E. of Barberton.

Note the large number

of inflorescences. Photo:

L. E. Codd.

Botanical Research Institute, Department of Agricultural Technical Services, Private Bag X101, Pretoria.

230

BROAD-SPECTRUM POLLINATION OF PLECTRANTHUS NEOCHILUS

Fig. 2. — Plectranthus neochilus : A, region of densely imbricate,

ovate, acuminate bracts; B, region of opening flowers;

C, region of opened verticils showing gaps between the

verticils (numbered 1 — 1 3). Pole Evans 4775, cultivated at

Prinshof (1954-01-19).

feeding). The most active feeding times were between

1 lhOO and 14h00., especially on calm, bright and sunny

days.

Scott Elliot (1891) recorded the following insects

on P. ecklonii Benth.: Apis mellifera collecting pollen,

a bombylid sucking and effecting cross-fertilization, as

well as two lepidoptera.

Although the flower beds could not be observed

continuously, it was clear that, as the inflorescences

gradually opened and expanded, there were definite

changes in the numbers and types of insects seen

visiting the plants over a three week period. It is

apparent that the gradual opening of the inflorescence

verticil by verticil, the abundant nectar production,

the small flowers, and many inflorescences on a single

plant all ensure a steady visitation by pollinators over

a long period.

During two months of observations in the wild,

however, only megachilid bees have been observed

visiting P. modulus.

Plectranthus barbatus Andr. is an attractive garden

plant cultivated in various parts of the world and,

according to Codd (1975), it has become semi-

naturalized in parts of South Africa. This plant is

grown in the Pretoria National Botanic Gardens in

the vicinity of P. neochilus, which is a less robust

plant with smaller flowers. Of the insects visiting

P. neochilus, only Xylocopa sichele Vachal, X. senior

Vachal and Macroglossum trochilus (Hbn.) also

visited P. barbatus, the latter two very rarely. Both

species of Plectranthus are new feeding records for

M. trochilus (Hbn.).

It may be concluded that P. neochilus possesses

inherent features, which have enabled it to be visited

by a broader spectrum of insect visitors in cultivation

than in the wild. The limited “garden strategy” of

P. barbatus needs further investigation, particularly

with reference to its original habitat.

ACKNOWLEDGMENTS

My sincere thanks go to Dr R. H. Watmough for

identifying the insects and to Dr L. E. Codd for his

interest and identification of the Plectranthus material.

UITTREKSEL

Die bestuiwings-ekologie van Plectranthus neochilus

Schltr. word bespreek en vergelyk met die van ’n

ander tuinplant, P. barbatus Andr.. Bestuiwers en

blombesoekers van P. neochilus sluit lede van die

Megachilidae, Anthophoridae, Syrphidae, Bombyliidae,

Sphingidae en Apidae in.

REFERENCES

Codd, L. E., 1975. Plectranthus (Labiatae) and allied genera in

Southern Africa. Bothalia 11: 371-442.

Scott Elliot, G. F., 1891. Notes on the fertilization of South

African and Madagascar Flowering Plants. Ann. Bot. 5:

333-405.

Bothalia 12, 2: 231-237 (1977)

Freshwater algae of Southern Africa. II. Triplastrum spinulosum

from the Transvaal

M. ISABELLA CLAASSEN*

ABSTRACT

The presence of Triplastrum spinulosum (Kisselev) Gauthier-Lievre (Desmidiaceae) in South Africa is

reported for the first time. The characters of T. spinulosum varieties spinulosum, indicum (Iyengar & Ramanathan)

Gauthier-Lievre and africanum Gauthier-Lievre were found to be exhibited by specimens collected at Ottosdal

in the south-western Transvaal. The differences among specimens of the populations studied are ascribed to

variation within one variable species and therefore the varieties indicum and africanum are relegated to syn-

onymy.

RESUME

ALGUES DULCICOLES D'AFRIQ UE AUSTRALE. II. TRIPLASTRUM SPINULOSUM

DU TRANSVAAL

La presence de Triplastrum spinulosum ( Kisselev ) Gauthier-Lievre ( Desmidiaceae ) en Afrique du

Sud est signalee pour la premiere fois. On a constate que les caracteristiques des varietes spinulosum,

indicum ( Iyengar & Ramanathan) Gauthier-Lievre et africanum Gauthier-Lievre de Vespece T. spinu-

losum se rencontrent dans les specimens recoltes a Ottosdal dans le sud-ouest du Transvaal. Les diffe-

rences entre specimens des populations etudiees sont attribuees a la variation intraspeciftque d'une

espece variable et les varietes indicum et africanum sont des lors placees en synonymie .

INTRODUCTION

Samples containing Triplastrum spinulosum were

collected by the author during April 1972 and

February 1975 from a small pan 16 km east of Ottosdal

in the south-western Transvaal and by Mr G. Germis-

huizen, Botanical Research Institute, Pretoria, during

February 1974 from a small pool on the banks of the

Nyl River on the farm Mosdene near Naboomspruit,

central Transvaal.

Both habitats contain water during the rainy

season, but may dry up partially or completely during

the dry season (winter). During December 1974,

which was an exceptionally dry summer, the Ottosdal

pan was completely dry.

The hydrogen-ion concentration and the tempera-

ture of the water at the time of collection were pH 7,7

and 24°C, and pH 7,5 and 27°C for April 1972 and

February 1975 respectively at Ottosdal and pH 6,1

and 26°C at Mosdene.

METHODS

The material was preserved with 4% formalin.

Slides were made by mounting a sample droplet in a

drop of glycerine. The methods used for the drawings

and photomicrographs were the same as those given

in the first paper of this series (Claassen, 1973).

OBSERVATIONS AND DISCUSSION

Iyengar and Ramanathan (1942) found a desmid

in material collected during December 1940 in a paddy

field near Madras, South India, which resembled the

species Triploceras abbreviatum Turner (Krieger, 1937;

Turner, 1892) and Triploceras simplex Allorge

(Krieger, 1937). One of the main characters of the

genus Triploceras Bailey is the presence of numerous

whorls of knotlike projections (verrucae). As this

character was lacking in both Turner’s and Allorge’s

species as well as in their own taxon, Iyengar and

Ramanathan (1942) decided to establish the new genus

Triplastrum to accommodate these taxa, the consti-

tuent species being Triplastrum indicum Iyengar &

Ramanathan, the desmid collected near Madras,

Triplastrum abbreviatum (Turner) Iyengar & Rama-

nathan and Triplastrum simplex (Allorge) Iyengar &

Ramanathan. However, they were unaware of the

existence of Triploceras spinulosum Kisselev (Krieger,

* Department of Botany, University of Pretoria, Pretoria.

1937), the description of which was based on a single

specimen collected in Turkestan in 1930. Triploceras

spinulosum closely resembles Triplastrum indicum and

in 1960 Gauthier-Lievre reduced the latter to a variety

of the former, thus Triplastrum spinulosum (Kisselev)

Gauthier-Lievre var. indicum (Iyengar & Ramanathan)

Gauthier-Lievre.

Gauthier-Lievre (1960) also described a new variety

based on material collected in Africa, namely Triplas-

trum spinulosum var. africanum. The name of this new

variety proposed by Gauthier-Lievre was not validly

published, however, since it was not accompanied by a

Latin description or diagnosis as stipulated by Article

36 of the International Code of Botanical Nomencla-

ture (Stafleu, 1972).

Turner (1892) described the apices of Triploceras

abbreviatum as three- or four-lobed. Iyengar and

Ramanathan (1942) incorporated this into their

description of the genus Triplastrum. Krieger (1937)

mentioned only three-lobed apices in his description

of Triploceras abbreviatum Turner, and Gauthier-

Lievre mentioned only three-lobed apices in her

description of the genus Triplastrum. In the material

collected in the Transvaal the Ottosdal specimens have

either three-lobed apices (Figs 1-3, 16 - upper cell,

17-22, 31, 32, 33 - cell on left, 34-38) or four-lobed

apices (Figs 4, 9, 16 - lower cell, 25, 33 - cel! on right)

while some specimens have one semicell with a

three-lobed apex and the other semicell with a four-

lobed apex (Figs 10, 11, 26-28, 39). The Mosdene

specimens have only four-lobed apices (Figs 5-8, 14).

Gauthier-Lievre (1960) described Triplastrum spinu-

losum var. spinulosum as having weakly dilated apices,

divided into three hardly divergent lobes which are

bi-, tri- or quadridenticulate. Krieger's figure of this

variety (1937, PI. 53, Fig. 7), one of Gauthier-Lievre’s

figures of her var. africanum (1960, Fig. 2t, cell on

left), and Figs 1 and 13 (upper semicell) in the present

paper as well as some of the specimens with four-

lobed apices (Figs 7, 8, 14) resemble one another in

that the apices are weakly dilated, and the polar lobes

hardly divergent and bi- or tridenticulate and seldom

quadridenticulate (Fig. 13).

According to Gauthier-Lievre (1960), Triplastrum

spinulosum var. indicum differs from the typical variety

in its larger dimensions and longer more divergent

polar lobes. Iyengar and Ramanathan’s figures for

232 FRESHWATER ALGAE OF SOUTHERN AFRICA. II. TRIPLASTRUM SP1NULOSUM FROM THE TRANSVAAL

this variety agree with Figs 2 (lower semicell) and 17

(cell on left) of the present paper as well as with some

of the cells with four-lobed apices (Figs 4-6). The

dimensions of these cells overlap both those of the

type and var. indicum.

Triplastrum spinulosum var. africanum (Gauthier-

Lievre, 1960) differs from the typical variety in that the

cells are slightly narrower below the apices and the

polar lobes are strongly divergent, bidenticulate and

seldom tridenticulate. The majority of the Ottosdal

specimens (Figs 2 - upper semicell, 3, 19-22, 31, 32,

34-38) agree with Gauthier-Lievre’s figures for this

taxon (1960, Figs 2q-t), but bidenticulate and tridenti-

culate polar lobes are approximately equal.

Iyengar and Ramanathan (1942) observed two

chloroplasts each with a central pyrenoid in each

semicell. Gauthier-Lievre (1960) reported 1 or 2

pyrenoids per semicell for the type and 1-3 pyrenoids

for her var. africanum. In the material collected in

Transvaal, the majority of the specimens observed had

two chloroplasts per semicell (Figs 1 , 2, 4-8, 11, 12, 14,

19, 21, 23, 24, 26-28, 31, 32, 41), some had three

chloroplasts per semicell (Figs 9, 25) or one semicell

contained two chloroplasts and the other three

(Figs 3, 20, 22). Each of these chloroplasts contained

a central pyrenoid. Only a few specimens were

observed in which the delineation of the chloroplasts

was indistinct and there appeared to be only one

chloroplast per semicell, but each contained two pyre-

noids.

In the Ottosdal material one specimen (Figs 9, 25)

was found which was considerably larger than the

others: length without spines 141 /am, with spines

144 pm; maximum width 14 pm; width of isthmus 11

pm; width of apices 17-18, 5 pm. If this specimen is

also taken into consideration, the dimensions of the

Transvaal specimens overlap those of Gauthier-

Lievre’s three varieties of Triplastrum spinulosum. The

dimensions of the other Transvaal specimens are as

follows:

Ottosdal Mosdene

length without spines 61-97 pm

length with spines. . . 63-102,5 pm

maximum width 9-11 ^m

width of isthmus. .. . 8,5-10,5 /xm

width of apex 11-18 pm

56-66 pm

59-68 pm

9,3-10 fim

8, 8-9, 3 pm

10, 5-14 pm

Zygospores, found in the Ottosdal material col-

lected during April 1972 (Figs 17, 18, 34-39), resemble

those found by Iyengar and Ramanathan (1942) but

are slightly smaller. The dimensions are 34-38 x

36^40 /am.

When all the characters found in the Transvaal

specimens were taken into consideration it was

decided that the specimens studied belonged to one

species, namely Triplastrum spinulosum and that the

two varieties indicum and africanum should be regarded

as synonyms.

Diagnosis

Cellulae parvae, circiter 6-10-plo longiores quam

latiores, in medio leviter at manifeste constrictae;

semicellulae rectae, cylindricae, lateribus paene paral-

lels vel leviter a basi paene usque ad apicem attenua-

te, apices aliquatenus inflati, trilobi vel quadrilobi;

qui lobi singuli 2-4 spinas breves ferunt; cellulae paries

incolor, pori perparvuli. Chloroplasti axiales bini vel

terni in semicellulis singulis, in seriem mediam dis-

positi, quorum unicuique laminae radiantes vel

stellatae cum pyrenoide medio. Zygospora sphaerica

(-subsphaerica) pariete spisso, margine crenato.

Longitudo sine spinis 56-99, 5 (141) pm, cum spinis

59-105 (144) pm; latitudo ad basim semicellulae

9-14 pm; isthmi latitudo 8, 5-12 pm; apicis latitudo

11-18 (18,5) ^.m; zygospora 36-42x38-42 ^m.

Cells small, about 6-10 times longer than broad,

with a slight but well-defined median constriction;

semicells straight, cylindrical, with sides nearly parallel

or slightly attenuated from base to just below the

apex; apices more or less inflated, three- or four-lobed,

lobes slightly to strongly divergent, each lobe bearing

2-4 short spines; cell wall colourlesss, pores minute.

Chloroplasts axial, 1 (?) or 2-3 in each semicell,

arranged in a median series, each chloroplast with

radiating plates (stellate) and a central pyrenoid.

Zygospore spherical to subspherical, thick-walled,

margin crenate. Length without spines 56-99,5 (141)

pm. with spines 59-105 (144) ^m; maximum width

near base of semicell 9-14 ^m; width of isthmus

8,5-12 pm; width of apices 11-18 (18,5) pm;

zygospore 38-42x36-42 pm.

The geographical distribution of this species is:

Asia. — Turkestan (in a rice-swamp near Weljko-Alexewskoje).

India (in a paddy field near Madras, December)

Africa. — Sudan (in a swamp of the River Niger near Gao,

December). French Equatorial Africa (Ubangi-Shari, February).

Uganda (in a swamp on the road from Masaka to Kampala,

August). Transvaal (in a small pool on the banks of the Nyl

River near Naboomspruit, February; in a small pan near Ottos-

dal, February and April).

It seems that this species is very rare and that

usually only a few specimens are found in samples

collected. In the samples collected near Ottosdal

during April 1972, however, it was fairly abundant.

As could be expected in a fairly large population

anomalous specimens were also found. In the Ottosdal

material one specimen was found where, in each

semicell, one of the polar lobes was under-developed,

without spines and somewhat subapical (Figs 12,

41). In another specimen only one of the semicells

was like this (Fig. 13, lower semicell). Several asym-

metrical cells were found (Figs 1 1, 29, 40) and one very

narrow semicell was observed (Fig. 30). The most

anomalous specimen was a single semicell with a

prominent basal inflation encircled by a whorl of

eight lobes similar to the polar lobes (Figs 15, 42-44).

In the Mosdene material one anomalous specimen

was found with the polar lobes of one semicell

undeveloped (Fig. 14).

The two varieties are thus placed into synonymy

and the circumscription of the species is amplified.

Triplastrum spinulosum ( Kisselev ) Gauthier-Lievre

in Revue Algol., 5, 1960, p. 64.

Syn.: Triploceras spinulosum Kisselev in Trans. Uzbekistan

Inst. Tropic. Med. 1, 1930, p. 50, PI. 4, Fig. 5; Krieger in

Rabenhorst’s Kryptogamenflora 13(1), 1937, p. 449, PI. 53, Fig.

7.

Triplastrum indicum Iyengar & Ramanathan in J. Indian Bot.

Soc., 21, 1942, p. 228, Figs 1-5. T. spinulosum (Kisselev)

Gauthier-Lievre var. indicum (Iyengar & Ramanathan)

Gauthier-Lievre in Revue Algol., 5, 1960, p.64. - var. africanum

Gauthier-Lievre, l.c. PI. 64, Fig. 2, q-t, non rite publ.

ACKNOWLEDGEMENTS

I wish to thank the Council for Scientific and

Industrial Research and the Research Committee

of the University of Pretoria for financial assistance,

Mr G. Germishuizen for collecting the Mosdene

samples and Prof. Dr H. L. Gonin, who kindly

prepared the Latin diagnosis.

UITTREKSEL

Triplastrum spinulosum ( Kisselev ) Gauthier-

Lievre (Desmidiaceae) is vir die eerste keer in Suid-

Afrika aangetref. Daar is gevind dat verteenwoordigers

M. ISABELLA CLAASSEN

233

wat naby Ottosdal in die suid-westelike Transvaal

versamel is die kenmerke van T. spinulosum varieteite

spinulosum, indicum ( Iyengar & Ramanathan) Gaut-

hier-Lievre cn africanum Gauthier-Lievre weerspieel.

Verskille tussen verteenwoordigers van hierdie popidasie

word as variasie in die self de takson aanvaar en die

varieteite indicum cn africanum word dus as sinonieme

beskou.

REFERENCES

Claassen, M. Isabella, 1973. Freshwater algae of Southern

Africa. 1. Notes on Gloeotrichia ghosei R. N. Singh. Br.

Phycol. J., 8: 325-331.

Gauthier-Lievre, L., 1960. Les Genres Ichtyocercus* , Triploce-

ras et Triplastrum en Afrique. Revue Algol., 5: 55-65.

Iyengar, M. O. P. & Ramanathan, K. R., 1942. Triplastrum,

a new member of the Desmidiaceae from South India.

J. Indian Bot. Soc., 21 : 225-229.

Krieger, W., 1937. Die Desmidiaceen Europas mit Berucksich-

tigung der aussereuropaischen Arten. 1. Teil. In Raben-

horst’s Kryptogamenflora von Deutschland, Osterreich und

der Schweiz, 13(1): 1-712.

Stafleu, F. A. (Chairman), 1972. International Code of Botanical

Nomenclature. Utrecht: A. Oosthoek’s Uitgeversmaat-

schappij NV.

Turner, W. B., 1892. The freshwater algae of East India.

K. Svenska Vetensk.-Akad. Handl., 25: 1-187.

* The original spelling of the name of this genus by its

authors was Ichthyocercus (West, W. & G. S. West, 1897,

Welwitsch’s African freshwater Algae, J. Bot., 35: 80).

234 FRESHWATER ALGAE OF SOUTHERN AFRICA. II. TRIPLASTRUM SPINULOSUM FROM THE TRANSVAAL

F,GS I “18. —Drawings of Triplastrum spinulosum. 1, cell with three-lobed apices, lobes tridenticulate; 2, cell with three-lobed apices,

obes bi- or tridenticulate; 3, cell with one semicell containing three chloroplasts and the other semicell two; 4-8, cells with four-

lobed apices, lobes bi- or tridenticulate; 9, large specimen with four-lobed apices, lobes tridenticulate, each semicell with three

chloroplasts; 10, cell with one three-lobed and one four-lobed apex, lobes bi- or tridenticulate; 11, asymmetrical cell with one

three-lobed and one four-lobed apex, lobes bident iculate; 1 2, anomalous cell with one polar lobe in each semicell under-developed,

without spines and slightly subapical, normal lobes bidenticulate; 13, anomalous cell with one polar lobe in one semicell under-

developed, without spines and slightly subapical, normal lobes bi-, tri- or quadridenticulate; 14, anomalous cell with polar lobes

of one semicell undeveloped; 15, anomalous semicell with a prominent basal inflation encircled by a whorl of eight lobes; 16,

conjugating cells showing protuberances; 17, conjugating cells with young zygospore; 18, conjugating cells with mature

zygospore. (Figs 1-4, 9-12 and 15-17: scale B. Figs 5-8, 13, 14 and 18: scale A).

M. ISABELLA CLAASSEN

100/um

Figs 19-31. — Photomicrographs of Triplastrum spinulosum. 19, 21, cells with three-lobed

apices and two chloroplasts per semicell; 20, 22, cells with three-lobed apices, one

semicell with two chloroplasts and the other semicell with three; 23, 24, cells with

four-lobed apices; 25, large specimen with four-lobed apices and each semicell

with three chloroplasts; 26-28, cells each with the upper apex four-lobed and the

lower apex three-lobed; 29, asymmetrical cell with upper semicell broader than

lower semicell; 30, abnormally narrow semicell; 31, two cells after division re-

maining attached to each other, (c, chloroplast; p, pyrenoid; n, nucleus).

236 FRESHWATER ALGAE OF SOUTHERN AFRICA. II. TRIPLASTRUM SPINULOSUM FROM THE TRANSVAAL

'•**•*-

0

3

;

Figs 32-36. — Photomicrographs of Triplastrum spinulosum 32, two

cells just before conjugation, the dirt particles represent the margin

of the mucilaginous envelope; 33, conjugating cells showing protuber-

ances; 34, conjugating cells with young zygospore; 35, 36, further

stages in developement of zygospore. (Fig. 32: scale A. Figs. 33-36:

scale B).

M. ISABELLA CLAASSEN

237

Figs 37-44. — Photomicrographs of Triplastrum spinulosum. 37, immature zygospore

showing crenate margin; 38, 39, mature zygospores; 40, asymmetrical cell; 41,

anomalous cell with one polar lobe in each semicell under-developed, without

spines and slightly subapical; 42-44, anomalous semicell with a prominent basal

inflation encircled by a whorl of eight lobes, photographed at various focussing

adjustments.

Bothalia 12, 2: 239-245 (1977)

Freshwater algae of Southern Africa. IV. Some Micrasteriae from

Rhodesia, including a new species

M. ISABELLA CLAASSEN*

ABSTRACT

Micrasterias ambadiensis (Gronblad & Scott) Thomasson, M. crux-melitensis (Ehrenberg) Hassall forma

minor Turner, M. decemdentata (Nageli) Archer, M. pinnatifida (Kutzing) Ralfs var. incudiformis West & West,

M. radiata Hassall var. brasiliensis Gronblad sensu lato, and M. tropica Nordstedt var. tropica are discussed.

A new species, M. schweickerdtii Claassen, is described.

RESUME

ALGUES DULCICOLES D'AFRIQUE AUSTRALE. IV. QUELQUES MICRASTERIAE DE

RHODESIE, INC LU ANT UNE NO UVELLE ESP EC E

On discute Micrasterias ambadiensis (Gronblad & Scott) Thomasson, M. crux-melitensis (Ehren-

berg) Hassall forma minor Turner, M. decemdentata (Nageli) Archer, M. pinnatifida (Kiitzing)

Ralfs var. incudiformis West & West, M. radiata Hassall var. brasiliensis Gronblad sensu lato, et

M. tropica Nordstedt var. tropica. Une nouvelle espece est decrite: M. schweickerdtii Claassen.

INTRODUCTION

A sample containing a rich assemblage of desmids

was given to me by Prof. Dr H. G. W. J. Schweickerdt,

formerly head of the Department of General Botany,

University of Pretoria. He collected the material

during July 1957 from a small pool on dolomite rocks

near Rusape, Rhodesia, but provided no physical or

chemical data.

The sample contained seven different Micrasterias

taxa amongst a number of other desmids. Although

this sample was poorer in taxa, the taxon-composition

showed a remarkable resemblance to that of Lake

Ambadi, Sudan (Gronblad, Prowse & Scott, 1958).

METHODS

The material was preserved in 4% formalin.

Slides were made by mounting a sample droplet

in a drop of glycerine.

Drawings were made with a Zeiss binocular bright

field/phase contrast microscope using a Leitz micro-

meter-net-ocular and specially printed squared paper,

the squares being 2x2 cm. The lenses used were a

12, 5 x eye-piece and a 40 x objective.

Photomicrographs were taken on Adox KB 14

film using a 35 mm Willd microscope camera on a

Zeiss Nomarski contrast microscope. A 6x eye-piece

and a 40 x objective were used.

All dimensions are given in micrometres (^m).

OBSERVATIONS AND DISCUSSION

The following taxa were identified in the sample

collected by Schweickerdt:

1. Micrasterias ambadiensis ( Gronblad & Scott)

Thomasson (1960, p. 22, Figs 4: 10 & 10: 4; Lind,

1971, p. 542, PI. 2, Fig. 3).

Syn.: M. radians Turner var. ambadiensis Gronblad & Scott

(Gronblad, Prowse & Scott, 1958, p.21, PI. 11, Fig. 119 & PI. 26,

Fig. 363; Gronblad, 1962, p.7, PI. 1, Fig. 9).

Some variation with regard to the delineation of

the lateral lobes occurred in the Rusape specimens

(Figs 1 1-14 and 26-29). In the type figures by

Gronblad & Scott (1958, Figs 119 and 363) as well

as Thomasson’s figure (1960, Fig. 4: 10) the inferior

lateral lobes have three spines each and the superior

lateral lobes either three or four. In the specimen

depicted in Gronblad’s Fig. 9 (1962) both inferior

and superior lateral lobes have four spines each and

* Department of Botany, University of Pretoria, Pretoria.

in Lind’s specimen (1971, PI. 2, Fig. 3) both inferior

and superior lateral lobes have three spines each.

In the Rusape specimens the number of spines in

both inferior and superior lateral lobes varies from

two to four. In a single specimen one inferior lateral

lobe was completely absent and another consisted of a

single spine (Fig. 12, bottom semicell). An abnormal

coalescence of the cell wall in the inner portion of the

sinus was observed in another specimen (Fig. 26).

This specimen also showed abnormal cell wall

thickenings at the base of each incision between the

inferior and superior lateral lobes and between the

superior lateral lobes and the polar lobe (upper

semicell) as well as a large granule in the centre of the

apical margin of the polar lobe. Length without

spines 114-136, with spines 140-168; width without

spines 98-1 13, with spines 1 12-131 ; width of isthmus

23-24; maximum width of polar lobe without spines

44-48, with spines 60-72. Very common.

2. Micrasterias crux-melitensis ( Ehrenberg ) Hassall

forma minor Turner (1892, p. 92, PI. 5, Fig. 4c;

Gronblad, Prowse & Scott, 1958, p. 19, PI. 11,

Fig. 124 & PI. 25, Fig. 358; Hinode, 1969, p. 199,

Fig. 7: 11; Krieger, 1939, p. 65, PI. 115, Fig. 1).

In some of the Rusape specimens the basal portion

of the polar lobe is noticeably broader than in the

typical form (Fig. 15). Length without spines 74-76,

with spines 87-89; width without spines 80-82, with

spines 86-89; width of isthmus 15-17; maximum

width of polar lobe 42-44. Rare.

3. Micrasterias decemdentata ( Nageli ) Archer

(Borge, 1918, p. 66, PI. 5, Fig. 23; Gronbald &

Croasdale, 1971, p. 11, PI. 4, Fig. 47; Gronblad,

Scott & Croasdale, 1964, p. 15, PI. 9, Fig. 219;

Krieger, 1939, p. 34, PI. 104, Figs 9-11; Lind, 1971,

p. 542, PI. 1, Fig. 29).

Syn.: M. decemdentata (Nag.) Arch. var. galpinii Claassen,

1961, p. 585, PI. 17, Figs 4-7.

Specimens with 10 spines per semicell were rather

rare; more common were specimens with 6 (Figs 16

and 17, upper semicells) or 9 (Fig. 17, lower semicell)

spines per semicell. Fig. 16 represents a specimen

in which one of the polar lobes is undeveloped (lower

semicell). Length 47,5-50; width without spines

46-50, with spines 56,5-62; width of isthmus 10-12;

maximum width of polar lobe without spines 30-31,

with spines 36-40,6. Rare.

240 FRESHWATER ALGAE OF SOUTHERN AFRICA. IV. SOME MICRASTERIAE FROM RHODESIA, INCLUDING A

NEW SPECIES

4. Micrasterias pinnatifida ( Kutzing ) Ralfs var.

incudiformis West & West (1895, p. 48, PI. 6, Fig. 5;

Gronblad & Croasdale, 1971, p. 11, PI. 4, Fig. 49;

Krieger, 1939, p. 19, PI. 100, Fig. 3).

These specimens (Figs 18-21) more closely resemble

the plant represented in Gronblad & Croasdale’s

Fig. 49 than the type specimen but their dimensions

agree better with that of the type. Length 56-60;

width without spines 57-62, with spines 61-67;

width of isthmus 11-12; maximum width of polar

lobe without spines 38-46,8, with spines 44-52.

Very common.

5. Micrasterias radiata Hassall var. brasiliensis

Gronblad sensu lato, 1945, p. 15, PI. 4, Figs 82 & 83;

Forster, 1969, p. 41, PI. 11, Fig. 5 & PI. 12, Fig. 1;

Thomasson, 1960, p. 24, Figs 4: 8 & 6: 12.

Syn.: M. radiata Hassall pro parte, in Krieger, 1939, p. 68,

PI. 117, Fig. 4 (non PI. 116, Figs 4-6 & PI. 117, Figs 1-3);

M. radiata Hassall forma Nordstedt (1869, in Borge, 1925, p. 29,

PI. 2, Fig. 7) in Forster, 1974, p. 156; M. radiata Hassall var.

groenbladii sensu Forster, 1974, p. 156, PI. 7, Fig. 1 (non PI. 5,

Fig. 80 in Scott, Gronblad & Croasdale, 1965); M. radians

Turner var. brasiliensis (Gronblad) Krieger in Forster, 1964, p.

380, PI. 18, Figs 1 & 2; Krieger & Scott (1957, p. 135) in Forster,

1974, p. 156.

The plant referred to here and all the cited figures

seen by the present author more closely resemble

Gronblad’s Fig. 83 than his Fig. 82. A broad inter-

pretation of M. radiata var. brasiliensis has been

adopted because, as the following discussion reveals,

there is no name available for the plant illustrated in

Fig. 83, and it is felt that, in view of the rather confused

nomenclature, it would be unwise to describe a new

taxon at this stage.

When Gronblad (1945) created M. radiata var.

brasiliensis he illustrated it with two figures. In his

Explanation of Plates he said “Fig. 82 is the most

frequent form, Fig. 83 is a larger form with more

slender processes'’ (Scott, Gronblad & Croasdale,

1965, p. 40). Croasdale (Scott, Gronblad & Croasdale,

1965) felt that the plant represented in Gronblad’s

Fig. 83 should be distinguished from the plant

represented in Fig. 82. She included this plant (Fig.

83) in the new variety M. radiata var. groenbladii

Croasdale (Scott, Gronblad & Croasdale, 1965, p. 39,

PI. 5, Fig. 80) and thereby it could possibly be con-

strued that she indirectly made Gronblad’s Fig. 82

the lectotype of M. radiata var. brasiliensis. As

Croasdale failed to indicate a type, the name M.

radiata var. groenbladii was not validly published

(Stafleu, 1972, p. 41, Art. 37). Forster (1969, p. 41)

decided that the plants depicted in Gronblad’s Fig. 83

and Croasdale’s Fig. 80 represented different varieties.

He transferred the plant represented in Gronblad’s

Fig. 83 back to M. radiata var. brasiliensis and

incorporated the plant represented in Croasdale’s

Fig. 80 in his new variety M. radiata var. croasdaleae

Forster (1969, p. 41, PI. 12, Fig. 2) as a synonym.

The name M. radiata var. croasdaleae was also not

validly published as no type was indicated, although

Forster mentioned on the first page of his paper

where slides containing new described taxa are

preserved.

When Forster (1969, p. 41) discussed M. radiata

var. brasiliensis he cited only Gronblad’s Fig. 83

and not Fig. 82 and it is not clear whether he meant

that his plant was M. radiata var. brasiliensis Gron-

blad, pro parte, quoad Fig. 83.

Later Croasdale (Gronblad & Croasdale, 1971,

p. 40) indicated her Fig. 80 in Scott, Gronblad &

Croasdale, 1965, as the type for M. radiata var.

groenbladii Croasdale. This validated the publication

of the variety and means that the name M. radiata

var. croasdaleae Forster is superfluous and illegitimate

and Forster’s concept of M. radiata var. groenbladii

Croasdale in Forster, 1974, p. 156, is incorrect.

When Forster (1974, p. 156) recognized Croasdale’s

variety groenbladii, he excluded Fig. 80 and cited

Gronblad’s var. brasiliensis, pro parte, quoad Fig. 83

as one of the synonyms. Thomasson (1960, p. 24,

Figs. 4: 8 and 6: 12) recorded M. radiata var.

brasiliensis Gronblad, 1945, cf. Fig. 83 for Lake

Bangweulu and although this paper appears in

Forster’s list of references (1974, p. 195) Thomasson’s

figures were not cited amongst the above-mentioned

synonyms.

Fig. 24 of the Rusape specimens resembles

Thomasson’s Fig. 4; 8 (1960, p. 15) and Fig. 9

resembles Thomasson’s Fig. 6: 12 (1960, p. 19) and

Forster’s PI. 7, Fig. 1 (1974, p. 213).

In the Rusape specimens there is a noticeable

variation in the width of the basal portions of the

polar lobes of mature semicells (Figs 9, 10, 24 and

25). An abnormal coalescence of the cell wall was

observed in the inner portion of the sinus (Fig. 24)

between the superior lateral and polar lobes (Fig. 24)

and between the lobules of the inferior and superior

lateral lobes (Fig. 25, mature semicell). The latter

also showed abnormal cell wall thickenings in the

diverging processes of the polar lobe. Length with

spines 180-208; width with spines 138-150; width of

isthmus 20-22; maximum width of polar lobe 80-100.

Common.

6. Micrasterias schweickerdtii Claassen, sp. nov.

(Figs. 1-8).

DIAGNOSIS

Inter species descriptas nulla affinitas obvia.

Cellulae amplae circiter 1 , 2— 1,3-pIo longiores quam

latiores, ellipticae, penitus constrictae, sinus apertus;

semicellulae trilobae. Lobus polaris magnus, sub-

cuneatus, anguli laterales deorsum curvati, apex

incisura mediana signatus, margini apicis spinae sunt

plerumque curvatae, 6 vel 7 utrimque ab incisura

mediana. Lobi laterales incisuris non profundis in

lobulos quattuor inaequales divisi; lobulus superior

sursum curvatus, tribus vel quattuor spinis

marginalibus ac curvatis instructus; quarum ima

maxima, vel in formam lobuli parvi et bidenticulati

delineata; lobuli mediani aut binis spinis marginalibus

instricti, aut in binos lobulos minores et bidenti-

culatos subdivisi, aut ex singula spina magna et

lobulo parvo et bidenticulatu constant; lobulus

inferior deorsum curvatus, tribus, spinis marginalibus

instructus. Cellulae paries porosus, plurimis instructus

spinis quae in series subradiatas in lobis polaribus et

lateralibus sunt dispositae; modo aliquot spinae in

semicellula media admodum supra isthmum

inveniuntur, modo nullae. Longitudo sine spinis

256-307 //m, cum spinis 284-337 //m; latitudo sine

spinis 201-234 /^m, cum spinis 227-262 /rm; isthmi

latitudo 41-45,5 /*m; lobi polaris latitudo maxima

sine spinis 124-146 /urn, cum spinis 139-161 ptm.

Iconotypus: fig. mihi 5.

This species does not correspond to any other

Micrasterias in literature available to the author.

Cells large, about 1,2-1, 3 times longer than

broad, elliptic, deeply constricted, sinus open;

semicells 3-lobed. Polar lobe large, subcuneate,

lateral angles curved downwards, apex with a median

notch, apical margin with 6-7, generally curved,

spines on each side of the median notch. Lateral lobes

divided into 4 unequal lobules by shallow incisions;

superior lobule curved upwards, furnished with 3-4

M. ISABELLA CLAASSEN

241

marginal curved spines, lowermost spine largest or

delineated as a small 2-denticulate lobule; median

lobules furnished with 2 marginal spines each or

subdivided into 2 smaller 2-denticulate lobules or

consists of one large spine and a small 2-denticulate

lobule; inferior lobule curved downwards, furnished

with 3 marginal curved spines. Cell wall porose,

furnished with numerous spines arranged in sub-

radiate rows within the polar and lateral lobes, with

or without a few spines in die middle of each semicell

just above the isthmus. Length without spines 256-

307, with spines 284-337; width without spines

201-234, with spines 227-262; width of isthmus

41-45,5; maximum width of polar lobe without

spines 124-146, with spines 139-161. Very rare.

Several unsuccessful attempts were made to turn

the cells over, so that they could be studied in lateral

or apical views.

The specimen depicted in Figs 3, 4 and 8 was

slightly anomalous in that one side of the sinus was

linear and not open.

7. Micrasterias tropica Nordsledt var. tropica

(Krieger, 1939, p. 56, PI. 112, Fig. 4).

In the polar lobes of the Rusape specimens (Figs

22, 23, 30 and 31) the diverging processes are longer

than in the type. These specimens fall rather in

between the plants represented in Krieger’s Fig. 4

(1939, PI. 112) and his Fig. 7 (1939, PI. 113) for

M. tropica var. elegans West & West as the lateral

and polar lobes are more slender than in the typical

variety and less slender than in var. elegans. The

semicell depicted in Fig. 23 could be compared with

var. elegans but for the shorter polar lobe. Length

92-104; width 80-106; width of isthmus 14-17,7;

maximum width of polar lobe 58-63,5. Rare.

ACKNOWLEDGEMENTS

I am indebted to the University of Pretoria for

research facilities, to Prof. Dr H. G. W. J.

Schweickerdt for the sample provided and to Prof.

Dr H. L. Gonin, who kindly prepared the Latin

diagnosis.

UITTREKSEL

Micrasterias ambadiensis ( Gronblad & Scott )

Thomasson, M. crux-melitensis ( Ehrenberg ) Hassal

forma minor Turner, M. decemdentata ( Ncigeli )

Archer, M. pinnatifida ( Kutzing ) Ralfs var. incudi-

formis West & West, M. radiata Hassall var. brasi-

iiensis Gronblad sensu lato, en M. tropica Nordstedt

var. tropica word bespreek. 'n Nuwe spesie, M.

schweickerdtii Claassen, word beskryf.

REFERENCES

Borge, O., 1918. Die von Dr. A. Ldfgren in Sao Paulo gesam-

melten Susswasseralgen. Ark. Boi. 15: 1-108.

Claassen, M. I., 1961. A contribution to our knowledge of the

freshwater algae of the Transvaal Province. Bothalia

7: 559-666.

Forster, K., 1964. Desmidiaceen aus Brasilien. 2. Teil: Bahia,

Goyaz, Piauhy und Nord-Brasilieri. Hydrobiologia 23 :

321-505.

Forster, K., 1969. Amazonische Desmidieen. 1. Teil: Areal

Santarem. Amazomiana 2: 5-116.

Forster, K., 1974. Amazonische Desmidieen. 2. Teil: Areal

Maues-Abacaxis. Amazoniana 5: 135-242.

Gronblad, R., 1945. De Algis Brasiliensibus, praecipue

Desmidiaceis, in regione inferiore fluminis Amazonas a

Professore August Ginzberger (Wien) anno MCMXXVII

collectis. Acta Soc. Sci. Fenn. 1 1 : 1-43.

Gronblad, R., 1962. Sudanese desmids II. Acta Bot. Fenn.

63: 2-19.

Gr5nblad, R. & Croasdale, Hannah, 1971. Desmids from

Namibia (S.W. Africa). Acta Bot. Fenn. 93: 2-40.

Gronblad, R., Prowse, G. A. & Scott, A. M., 1958. Sudanese

desmids. Acta Bot. Fenn. 58: 2-82.

Gronblad, R., Scott, A. M. & Croasdale, Hannah, 1964.

Desmids from Uganda and Lake Victoria collected by Dr

Edna M. Lind. Acta Bot. Fenn. 66: 1-57.

Hinode, T., 1969. On some Japanese desmids (6). Hikobia

5: 196-201.

Krieger, W., 1939. Die Desmidiaceen Europas mit Berucksich-

tigung der aussereuropaischen Arten. 2. Teil. In Raben-

horst’s Kryptogamenflora von Deutschland, Osterreich und

der Schweiz 13(2): 1-117.

Lind, Edna M., 1971. Some desmids from Uganda. Nova

Hedwigia 22: 535-585.

Scott, A. M., Gronblad, R. & Croasdale, Hannah, 1965.

Desmids from the Amazon Basin, Brazil. Acta Bot. Fenn.

69: 1-94.

Stafleu, F. A. (Chairman), 1972. International Code of Botanical

Nomenclature. Utrecht: A. Oosthoek's Uitgeversmaat-

schappij N.V.

Thomasson, K., 1960. Notes on the plankton of Lake Bangweulu

Part 2. Nova Acta R. Soc. Scient. Upsal. 17: 1-43.

Turner, W. B., 1892. The freshwater algae (principally Des.

midieae) of East India. K. Svenska Vetensk. -Akad. Hand-

25 : 1-187.

West, W. & West, G. S., 1895. The freshwater algae of Mada-

gascar. Trans. Linn. Soc. Lond. (Bot.) 5: 41-90.

242 FRESHWATER ALGAE OF SOUTHERN AFRICA. IV. SOME MICRASTERIAE FROM RHODESIA, INCLUDING A

NEW SPECIES

Figs 1-4. — Micrasterias schweickerdtii ; 4, semicell showing spines in surface view.

M. ISABELLA CLAASSEN

243

Figs 5-8. — Micrasterias schweickerdtii.

244 FRESHWATER ALGAE OF SOUTHERN AFRICA. IV. SOME MICRASTERIAE FROM RHODESIA, INCLUDING A

NEW SPECIES

Figs 9, 10. — Micrasterias radiata var. brasiliensis; 9, mature cell; 10, cell in divisional stage. 11-14, M. ambadiensis,

various specimens showing variation of lateral lobes.

Fig. 15. — Micrasterias crux-melitensis forma minor. 16, 17, M. decemdentata. 18-21, M. pinnatifida var. incudiformis ;

18, 19, front view of cell; 20. apical view of semicell; 21, basal view of semicell. 22, 23, M. tropica var. tropica.

M. ISABELLA CLAASSEN

245

Figs 24, 25. — Micrasterias radiata var brasiliensis; 24, mature cell; 25, cell after division with anomalous upper semicell. 26-29, M. ambadiensis,

various specimens showing variation of lateral lobes. 30, 31, M. tropica var. tropica.

Bothalia 12, 2: 247-250 (1977)

Asexual nuclear division in Neocosmospora*

K. T. VAN WARMELOf

ABSTRACT

A fungus isolated from soybean stem material showed marked similarities to two existing species of

Neocosmospora , i.e. N. vasinfecta E. F. Smith and N. africana von Arx. The processes of somatic nuclear division

in authentic cultures of these two species and the new isolate were examined to determine the nature of the

mechanism of nuclear division and whether there were any differences among the cultures. No significant

differences were observed. The divisions showed asynchronous anaphase disjunction but there was no evidence

of aneuploidy or irregular reconstitution of daughter nuclei. Nuclear division was interpreted as being strictly

mitotic.

RESUME

CA R YOCINESE ASEXUEE CHEZ NEOCOSMOSPORA

Une moisissure isolee de tiges de soya a montre des ressemblances marquees avec deux espices

existantes de Neocosmospora, so it N. vasinfecta E. F. Smith et N. africana von Arx. Les processus de

caryocmese somatique ont ete examines dans des cultures authentiques de ces deux especes et dans le

nouvel isolat pour determiner la nature du mecanisme caryocinetique et noter les differences entre cultures

s'il y en a. On n'a pas observe de differences significatives. Les divisions ont montre une dislonction

asynchrone a V anaphase , mais il ny avait aucun signe d'aneuploidie ou de reconstitution irreguliere

de noyaux-fils. La caryocinese a ete interpretee comme etant strictement mitotique.

REVIEW

Reported mechanisms of somatic nuclear division

in the fungi seem to fall into two basic categories,

i.e. divisions which are essentially mitotic, differing

only slightly from the classical concept, and non-

mitotic divisions in which completely different

processes are operative.

Divisions in a large variety of fungi were shown to

differ from mitosis in higher plants only in that the

nuclear membrane was persistent throughout the

division (Thyagarajan, Conti & Naylor, 1962; Moor,

1966; Robinow & Marak, 1966; Heath & Greenwood,

1968; McManus & Roth, 1968; Aist, 1969; Aldrich,

1969; van Winkle, Biesele & Wagner, 1971). The

spindle was thus fully intranuclear (Ichida & Fuller,

1968; Motta, 1969; Zickler, 1970).

Persistence of the nuclear membrane was, however,

not an invariable condition, as fungi were reported

in which the membrane broke down during division

(Robinow, 1963; Namboodiri & Lowry, 1967; Motta,

1969; Brushaber & Jenkins, 1971).

An additional difference between somatic divisions

in fungi and higher plants was the variable presence

of centrioles. Several fungi, representative of widely

differing taxonomic groups, were reported as lacking

centrioles (Namboodiri & Lowry, 1967; McManus &

Roth, 1968; Aldrich, 1969) whereas their presence was

demonstrated in others (Moor, 1966; Robinow &

Marak, 1966; Brushaber & Jenkins, 1971; van

Winkle et al., 1971).

Despite the differences referred to above, however,

the divisions in a large number of fungi have been

described as essentially mitotic (Robinow, 1963;

Hosford & Gries, 1966; Knox-Davies, 1966, 1967;

Finley, 1970; Brushaber & Jenkins, 1971, van

Warmelo, 1971).

The actual arrangement of chromosomes imme-

diately prior to, and during, separation has been the

subject of intensive investigation and a large number

of different mechanisms of chromosomal separation

have been proposed as a result.

The formation of typical metaphase plates was

reported (Robinow, 1963; Finley, 1970) and also the

separation of discrete chromatids on a well-defined

* Based on a Ph.D. (Agric.) thesis submitted to the University

of Stellenbosch, 1973. Promoter: Prof. P. S. Knox-Davies,

Ph.D. (Wis.).

t Rand Afrikaans University, P.O. Box 524, Johannesburg.

spindle (Knox-Davies, 1966, 1967; Ichida & Fuller,

1968; McManus & Roth, 1968; Motta, 1969;

Brushaber & Jenkins, 1971). Aist (1969), however,

maintained that instead of forming a true metaphase

plate the chromosomes became attached to the

spindle at different points. During anaphase separation

the chromatids thus became strung out on the

spindle.

Instead of separate chromosomes, the presence of a

single coherent strand, formed by the linked chromo-

somes has been reported (Weijer, Koopmans &

Weijer, 1965; Dowding, 1966; Weijer & Weisberg,

1966; Brushaber, Wilson & Aist, 1967; Laane, 1967;

Namboodiri & Lowry, 1967; Heale, Gafoor & Raja-

singham, 1968; Brushaber & Jenkins, 1971). The

single nuclear strand is thought to replicate to form

a double strand which then separates in a number of

ways. Movement of each replicated strand on a well

defined spindle was reported by Heale et al. (1968).

Spindles were, however, apparently absent in the other

fungi showing filamentous nuclei. Shearing of chro-

matids by cytoplasmic streaming was suggested by

Weijer et al. (1965) but, whereas this proposal received

some support (Laane, 1967), it has not been

generally accepted. In the absence of a spindle

the exact mechanism of chromatid disjunction and

movement is not yet fully understood.

Judging from the conflicting descriptions given

above it seems possible, as suggested by Bracker

(1967), that there may be a number of mechanisms of

somatic nuclear division in the fungi, instead of one

common process.

MATERIALS AND METHODS

Three cultures were examined in this investigation:

1. Neocosmospora vasinfecta E. F. Smith; CBS

237.75.

2. Neocosmospora africana von Arx; CBS 237.

3. Neocosmospora sp. isolated from soybean

stem, Pietermaritzburg, South Africa.

Referred to in the following text as “isolate

P.”

Somatic divisions were studied using the macerated

mycelium technique described by Ward & Ciurysek

(1962). After being stained using the HCl-Giemsa

technique (van Warmelo, 1971) the hyphae were

flattened under a coverglass which was then sealed

with paraffin oil.

248

ASEXUAL NUCLEAR DIVISION IN NEOCOSMOSPORA

Figs 1-8. — Mitosis in Neocosmospora vasinfecta, x3000. 1, early prophase; 2, prophase

with clearly visible chromosomes; 3, late prophase with short chromosomes; 4, mid

metaphase with six distinguishable chromosomes; 5, metaphase/anaphase; 6, mid

anaphase showing chromosomes between clearly visible centriolar plaques; 7, early

telophase with spindle bridge apparently enclosed within nuclear membrane;

8, early telophase.

Figs 9-16. — Mitosis in Neocosmospora africana, x 3000. 9, early prophase with visible

chromosomes; 10, prophase; 11, prometaphase in polar view with visible centriolar

plaques (arrowed); 12, very early metaphase with forming spindle (arrowed);

13, early metaphase; 14, anaphase; 15, early telophase with spindle bridge; 16,

telophase.

Figs 17-24. — Mitosis in Neocosmospora isolate P, x 3000. 17, early prophase with visible

chromosomes and nuclear membrane; 18, prophase in large hypha with clearly

visible chromosomes; 19, prophase with shortened chromosomes; 20, mid metaphase

with formed spindle; 21, early anaphase; 22, late anaphase with chromosome

disjunction almost completed; 23, late anaphase showing residual interzonal fibres;

24, late telophase.

K. T. VAN WARMELO

249

RESULTS

As the mechanisms of somatic division in the

three cultures did not show any significant differences,

the processes of division will not be described

separately. Individual isolates will be highlighted

only where necessary.

The interphase nuclei were large and typical. The

onset of prophase was shown by the delimitation of

chromosomes within the nuclei (Figs 1, 9 and 17).

Initially long and thin, the chromosomes were

shorter and more distinct at mid-prophase (Figs 2,

10 and 18). Nucleoli which were clearly distinguishable

at early prophase, were usually absent at mid-

prophase. The size of the somatic cell appeared to

affect the size of the nucleus. In isolate P (Fig. 18)

cells of large diameter were found in which the

nuclei were much larger, and the chromosomes much

clearer, than in the more common narrow hyphae.

At late prophase the chromosomes were appreciably

shorter than at early prophase, but could not be

counted (Figs 3 and 19). Occasional persistent

nucleoli were seen at this stage.

At prometaphase (Fig. 11) centriolar plaques were

sometimes distinguishable.

During early metaphase the chromosomes had

condensed further and were shorter than at late

prophase (Fig. 12). At metaphase they were situated

midway between the two centriolar plaques and

incipient spindles were observed (Figs 12 and 13).

At mid-metaphase the spindle was clearly distin-

guishable (Figs 5 and 20).

Occasionally individual chromosomes were distin-

guished at metaphase (Figs 4 and 20) and a count

at this stage gave an estimated number of six.

Anaphase was a variable stage and chromosome

separation appeared to be asynchronous, as a variety

of nuclear figures were found (Figs 6, 14, 21 and 22).

At late anaphase the interzonal fibres between the

terminal chromosome aggregations were occasionally

observed (Fig. 23).

At early telophase the two daughter nuclei became

reconstituted but were frequently connected by a

chromosome or spindle bridge (Figs 7 and 15)

apparently still enclosed within the nuclear membrane.

The bridge then broke and shortened (Fig. 8). The

divisions thus appear to be entirely intra-nuclear.

Mid- and late-telophase nuclei (Fig. 16) were dense

and smaller than post-division interphase nuclei

(Fig. 24) which were identical with pre-division

interphase nuclei.

At no stage were nuclear strands observed or

stages seen which would have suggested linking-up

of individual chromosomes.

DISCUSSION

No significant differences in nuclear division were

found to occur in these three cultures. This agrees

with their close similarity in other respects.

Due to the small diameter of most somatic hyphae

there is probably insufficient space available for the

formation of a typical metaphase plate, as found in

higher organisms, and for the simultaneous disjunction

and movement of chromatids on the spindle. The only

mechanism which would ensure an equal distribution

of chromatids under such conditions would be a

sequential or asynchronous separation. Pairs of

chromatids would then separate and move away from

the central spindle area before separation of the next

chromatid pair, thus ensuring non-interference by

non-sister chromatids destined for different daughter

nuclei.

Although anaphase separation was asynchronous

there was no evidence of irregular chromosome

disjunction or aneuploidy. This was deduced from the

absence of lagging chromosomes or fragments at late

telophase or interphase, indicating complete recon-

stitution of daughter nuclei, the constancy of chromo-

some numbers and the phenotypic stability of the

cultures.

The consistent presence of observably separate

chromosomes during nuclear division, coupled with

the absence of stages showing linking of chromosomes,

indicates that somatic nuclear division in Neocosmos-

pora can be interpreted as strictly mitotic.

ACKNOWLEDGEMENTS

This study was supported by research grants from

the Rand Afrikaans University, Johannesburg and

the Council for Scientific and Industrial Research,

Pretoria.

UITTREKSEL

n Swamsoort afkomstig van sojaboon weefsel het

duidelike ooreenkomstes getoon met twee bestaande

Neocosmospora soorte, nl. N. vasinfecta E. F. Smith

en N. africana von Arx. Die somatiese kerndelings-

prosesse in outentieke kidture van hierdie twee

organismes en in die nuwe isolaat is ondersoek om

vas te stel wat die delingsmeganismes was en of daar

verskille tussen die kulture aanwesig was. Geen

wesenlike verskille tussen die kulture is opgemerk nie.

Die delings het asinkroniese anafaseskeidings getoon

maar daar was geen tekens van aneuploiedie of onreel-

matige vorming van die dogterkerne nie. Die meganisme

van kerndeling is suiwer mitoties.

REFERENCES

Aist, J. R., 1969. The mitotic apparatus in Fungi, Ceratocystis

fagacearum and Fusarium oxysporum. J. Cell Biol. 40:

120-135.

Aldrich, H. C., 1969. The ultrastructure of mitosis in myxamoe-

bae and plasmodia of Physarum flavicomum. Am. J. Bor.

56: 290-299.

Bracker, C. E., 1967. Ultrastructure of fungi. A. Rev.

Phytopath. 5: 343-374.

Brushaber, J. A. & Jenkins, S. F., (Jr.) 1971. Mitosis and

clamp formation in the fungus Porta monticola. Am. J. Bot.

58: 273-280.

Brushaber, J. A., Wilson, C. L. & Aist, J. R., 1967. Asexual

nuclear behavior of some plant pathogenic fungi. Phyto-

pathology 57 : 43-46.

Dowding, E. Silver, 1966. The chromosomes in Neitrospora

hyphae. Can. J. Bot. 44: 1121-1125.

Finley, D. E., 1970. Somatic mitosis in Ceratobasidium flaves-

cens and Pellicularia koleroga. Mycologia 62: 474-485.

Heale, J. B., Gafoor, A. & Rajasingham, K. C., 1968. Nuclear

division in conidia and hyphae of Verticillium albo-atrum.

Can. J. Genet. Cytol. 10: 321-340.

Heath, I. B. & Greenwood, A. D., 1968. Electron microscopic

observations of dividing somatic nuclei in Saprolegnia .

J. Gen. Microbiol. 53 : 287-289.

Hosford, R. M. & Gries, G. A., 1966. The nuclei and para-

sexuality in Phymatotrichum omnivorum. Am. J. Bot. 53:

570-579.

Ichida, A. A. & Fuller, M. S., 1968. Ultrastructure of mitosis

in the aquatic fungus Catenaria anguillulae. Mycologia 60:

141-155.

Knox-Davies, P. S., 1966. Nuclear division in the developing

pycnospores of Macrophomina phaseoli. Am. J. Bot. 53:

220-224.

Knox-Davies, P. S„ 1967. Mitosis and aneuploidy in the vegeta-

tive hyphae of Macrophomina phaseoli. Am. J. Bot. 54:

1290-1295.

Laane, M. M., 1967. The nuclear division in Penicillium expan-

sum. Can. J. Genet. Cytol. 9: 342-351.

McManus, Sister M. A. & Roth, L. E„ 1968. Ultrastructure

of the somatic nuclear division in the plasmodium of

the Myxomycete Clastodenna debaryanum. Mycologia 60:

426-436.

250

ASEXUAL NUCLEAR DIVISION IN NEOCOSMOSPORA

Moor, H., 1966. Ultrastrukturen im Zellkern der Backerhefe

J. Cell Biol. 29: 153-155.

Mot'ta, J. J., 1969. Somatic nuclear division in Armillaria

mellea. Mycologia 61 : 873-886.

Namboodiri, a. N. & Lowry, R. J., 1967. Vegetative nuqear

division in Neurospora. Am. J. Bot. 54: 735-748.

Robinow, C. F., 1963. Obervations on cell growth, mitosis, and

division in the fungus Basidiobolus ranarum. J. Cell Biol.

17: 123-152.

Robinow, C. F. & Marak, J., 1966. A fiber apparatus in the

nucleus of the yeast cell. J. Cell Biol. 29: 129-151.

Thyagarajan, T. R., Conti, S. F. & Naylor, H. B., 1962.

Electron microscopy of Rhodotorula glutinis. J. Bad. 83:

381-394.

Van Warmelo, K. T., 1971. Somatic nuclear division in

Stemphylium botryosum. Bothalia 10: 329-334.

Van Winkle, W. B., Biesele, J. J. & Wagner, R. P., 1971. The

mitotic spindle apparatus of Neurospora crassa. Can. J.

Genet. Cytol. 13: 873-887.

Ward, E. W. B. & Ciurysek, K. W.. 1962. Somatic mitosis in

Neurospora crassa. Am. J. Bot. 49: 393-399.

Weijer, J., Koopmans, A. & Weijer, D. L., 1965. Karyokinesis

of somatic nuclei of Neurospora crassa. III. The juvenile

and maturation cycles (Feulgen and crystal violet staining).

Can. J. Genet. Cytol. 7: 140-163.

Weijer, J. & Weisberg, S. H., 1966. Karyokinesis of the somatic

nucleus of Aspergillus nidulans. I. The juvenile chromosome

cycle (Feulgen staining). Can. J. Genet, Cytol. 8: 361-374.

Zickler, Denise, 1970. Division spindle and centrosomal

plaques during mitosis and meiosis in some Ascomycetes.

Chromosoma 30: 287-304.

Bothalia 12,2: 251-259 (1977)

Notes on African plants

VARIOUS AUTHORS

ARACEAE

A NEW SPECIES OF GONATOPUS FROM SOUTHERN AFRICA

Gonatopus rhizomatosus Bogner & Oberm. sp. nov.,

G. boivinii ( Decne.) Engl, affinis, sed rhizomate

cylindraceo longo horizontali non globoso et petiolo

non geniculato differt.

Planta rhizomatosa 20-40 cm alta. Folium unum

annuum decompositum proteranthum; petiolus tenuis

15-25 cm longus; lamina a basi ternata supra 1-2

divisa; pinnae oppositae vel suboppositae decurrentes.

Spadix breve pedunculata c. 10 cm longa. Spatha

angusta c. 7 cm longa margine ad basi libero. Flores

masculi superiores; tepala paribus oppositis semi-

prismatica truncata 3 mm longa. Staminum filamenta

ponnata. Flores steriles in medio masculini et feminei.

Flores feminei inferiores. Gynoecium ampulliformum

bilocularum; ovulum anatropum. Bacca oblongo-

globosa c. 15 mm longa; semen anguste ellipsoideum,

12 mm longum.

Type: Transvaal, 2531 (Komatipoort): 5 km S. of

Kaalrug Police Station near Swedish Mission (-DA),

Codd 7814 (PRE, holo.; M, iso.).

Herbaceous plant producing annually one leaf and

later beside it 1 or rarely 2 inflorescences. Rhizome

horizontal, 20 cm long and 3 cm in diameter, or longer,

cylindrical, tuberous, somewhat contracted irregularly

at intervals, forming much-branched mats, brown

outside with white flesh; roots scattered, fairly thick.

Fig. 1. — Gonatopus rhizomatosus.

Eastern Transvaal ( Codd

7814), x}.

252

NOTES ON AFRICAN PLANTS

Fig. 2. — Gonatopus rhizomatosus. 1, male perianth, x 6; 2,

stamens and sterile ovary from above, x 10; 3, stamens

seen from side, xlO; 4, female perianth, x6; 5, outer

tepal x 6; 6, inner tepal x 6; 7, ovary, X 6; 8,

longitudinal section through ovary, x 10; 9, fruits, x^.

( Moll 4974A).

Fig. 3. — Pollen grains (x500) of Gonatopus rhizomatosus

Photo: W. Barthlott.

VARIOUS AUTHORS

253

Leaf e rect, with a petiole 15-25 cm long, dull grey with

lighter mottling, surrounded by cataphylls at the base;

lamina somewhat fleshy, decompound, 3- branched

at the base, once or twice branched above it, pinnae

opposite or subopposite, linear to oblong, attenuate

above and below, 4-8 cm long, all or only the terminal

decurrent on rhachis, often with a small side-lobe.

Spadix on a short peduncle c. 10 cm long, grey,

mottled; spathe constricted, green in bud, turning

pinkish buff with black stripes but soon drying up, c.

7 cm long, closely enveloping spadix, apiculate, not

fused below, the upper part reflexed at anthesis, the

lower enclosing the female part for some time,

dropping off when fruit is set. Male flowers compressed

in spiral series on upper half of cylindrical, obtuse

spadix; tepals in 2 opposing free pairs, semi-prismatic,

thick, truncate, c. 3 mm long; stamens 4, with the

filaments fused into a tube, surrounding the sterile,

clavate gynoecium; anthers arranged in a circle,

thecae opening apically towards the centre. Pollen

grains depressed globular, 95-1 10// x 30-40//, meri-

dionosulcate, exine sparsely covered with very small

foveolae. Female flowers on lower part of spadix, with

sterile flowers in between male and female flowers;

tepals similar to male flowers; stamens absent;

gynoecium flask-shaped, laterally compressed, 2-

locular with an anatropous basal ovule in each locule;

style short; stigma discoid, large, green, just protruding

from perigon. Berries bilocular, oblong-globose, c.

15 mm long, in a globose cluster surrounded basally

by the persistent tepals and unfertilized flowers, walls

soft, thin. Seed narrowly ellipsoid, 12 mm long, coat

thin, without endosperm. Chromosomes: 2n=c.68

(N. Jacobsen). Figs 1, 2 & 3.

Eastern Transvaal, northern Natal, Mozambique,

Rhodesia; common in undergrowth of dune forest in

northern Natal, often in shady rock pockets in low-

lying bushveld areas of eastern Transvaal; flowering

November to February.

Transvaal. — 2331 (Phalaborwa): Gorge (-DD), Van der

Schijff 2387. 2431 (Acornhoek): 21 km E. of Skukuza (-DC),

Codd 5045. 2531 (Komatipoort) : 4 km W. of Lower Sabie

Camp (-BB), Codd 5707 (M, PRE); Kaap Muiden (-CB),

Mogg 13687 ; 5 km S. of Kaalrug Police Station near Swedish

Mission (-DA), Codd 7814 (PRE, holo.; M. iso.).

Natal. — -2632 (Bela Vista): Mtombeni Forest, slopes facing

Nhlange Lake (-CD), Tinley 396. 2731 (Louwsburg): Josini

Dam, slopes (-BD), Strey 5278. 2732 (Ubombo): Ngwavuma,

on road to Ndumu (-AA), Venter 5215; Lebombo foot hills on

road to Pongola River (-AC), Strey 6580; Makatini flats near

Mbazwane (-BC), Vahrmeijer & Toelken 208. 2831 (Nkandla):

Umfolozi Game Reserve (-BD), White 10357; 16 km W. of

Empangeni on Melmoth Road (-DD), Moll 4974a. 2832

(Mtubatuba): Hluhluwe Game Reserve (-AA), Ward 1967;

near Mtubatuba (-AC), Codd 9619.

Mozambique. — Inhaca Island, M. Moss sub. J. 28428.

Rhodesia. — Beitbridge district, 16 km S. of Lundi River,

Wild 7610.

Two species of Gonatopus occur in southern Africa,

namely G. boivinii (Decne.) Engl., a tropical eastern

African species, which reaches its southernmost limit

in northern Natal, and G. rhizomatosus, the species

here described, which has been recorded from northern

Natal, eastern Transvaal Lowveld, Mozambique and

Rhodesia.

In G. boivinii , a taller plant, the underground part

consists of a globose often depressed tuber, the petiole

forms a swollen articulation in the middle and the

decompound lamina has sessile or petiolate paired

pinnae, which are not decurrent. In G. rhizomatosus

the underground part consists of a long horizontal

tuberous rhizome, the petiole is without a swollen

central articulation and the decompound lamina has

the pinnae opposite or subopposite and decurrent.

There is also a difference in the size of the pollen

grains for they are much larger than those of G. boivinii

and the exine is more sparsely covered withf oveolae.

Mr Josef Bogner of the Botanical Garden of

Munich, co-author of G. rhizomatosus , is an authority

on the Araceae. He assisted me most generously in

clarifying the position of this species and the tropical

ones described by A. Pet?r under the latter’s genus

Microculcas. Mr Bogner will deal with these taxonomic

problems at a future date. He assures me that our

southern rhizomatous tpecies differs from Microculcas

marattioides described by Peter from Tanzania. The

latter also possesses a horizontal rhizome but, in

addition, develops long stolons which swell up to

about the thickness of a finger and produce a leaf at

their ends. The petiole too, is swollen basally, but

becomes attenuated and thin above.

A. A. Obermeyer

ASCLEPIADACEAE

A NEW SPECIES OF CARALLUMA FROM THE CAPE

Caralluma multiflora R. A. Dyer sp. nov., C.

mammilliari (L.f.) N.E. Br. affinis, corolla alabastro

plusminusve oblonga, obtusa; corollae lobis replicatis

carinatis, plusminusve oblongis, apicem versus

incurvis, carina infra medium et fauce dense papillosis

distinguitur.

Planta carnosa ramosa, usque ad 15 cm alta;

caules 4-5-angulati, circiter 2 cm crassi, tuberculati;

tubercula triangulata, 6-7 mm prominentia, apicem

versus indurata, acuta. Flores multi, fasciculati.

Corolla alabastro plusminusve oblonga, obtusa,

5-angulata, circiter 1,5 cm longa, glabra, atropur-

purea, breviter tubulata; tubus campanulatus, circiter

4 mm longus, intus pallidus, sparse pubescens et

maculatus; faux dense papillosa; lobi plusminusve

oblongi, erectopatentes, apicem versus incurvi,

replicati, carinati; carina infra medium papillosa.

Corona breviter campanulata; lobi exteriores suberecti,

profunde bifidi, dentibus 0,5 mm longis; lobi

nteriores incumbenti-erecti, basi in lobulum sub-

quadratum vel ovatum producti. Pollinia circiter

0,25 mm longa. Folticuli 2 (1), 13-14 cm longi,

7-8 mm diam, longistrorsum maculati.

Type. — Cape, 3119 (Calvinia): Kareeboomfontein,

west of Rebunie (-DA) 17th October 1975, W. J.

Hanekom 2475 (PRE, holo.).

Succulent plant up to about 15 cm tall, freely

branching and rooting from the basal branches;

branches 4-5-angled, prominently tuberculate along

the angles, with a spread of 2,5-2,75 cm, bluish-green

to purplish red-brown; tubercles triangular in side

view, spreading 6-7 mm, with a sharp hard apex.

Flowers mainly from flowering eyes towards the apex

of branches between or along the angles, in fascicles

of up to 12 flowers together, opening more or less

together; pedicels 4-5 mm long. Sepals linear-

lanceolate, about 3 mm long. Corolla in bud more or

less oblong in outline, but broadest slightly above

middle, 5-angled, 1 ,5 cm long, glabrous outside, dark

maroon to black with short yellowish-green to light

254

NOTES ON AFRICAN PLANTS

coloured tube; tube campanulate, about 4 mm long,

with scattered hairs within and maroon-marked,

densely papillate round mouth of tube; papillae with

acute apex bent at right-angles; lobes more or less

oblong in outline, suberect, with incurving tips and

recurving margins, obtusely keeled down inner face,

with few papillae down keel towards base. Corona

shortly cup-shaped or campanulate at base; outer

lobes deeply bifid and alternating with spreading basal

lobule of inner corona-lobes, teeth of outer lobes about

0,5 mm long, suberect; inner corona-lobes about

2 mm long, incumbent-erect into a cone-shaped apex

above the staminal column, with a spreading ovate or

subquadrate basal lobe (variable) alternating with the

outer corona-lobes and about the same length.

Pollinia about 0,25 mm long with very short weak

connectives; carrier very shortly winged. Follicles

Fig. 4. — Caralluma multiflora. 1, open flower, x3; 2, corolla

opened out, x3; 3, papilla with sharp tip turning at

right-angles, xlOO; 4, corona, xl5; 5, pollinia and

carrier, x 60. ( Hanekom 2335).

(1) 2 on a separate plant without flowers, towards base

of branches, upright, 13-14 cm long, 7-8 mm thick,

longitudinally striped and mottled. Fig. 4.

Cape. — 31 19 (Calvinia): Kareeboomfontein, west of Rebunie-

kop (-DA), rocky ground with some protection from shrublets,

24/8/74, Hanekom 2335 ; 17/10/75, Hanekom 2475.

This is a relatively robust succulent which, like so

many members of the Stapelieae, is found on rocky

slopes in the halt-shade of shrublets. Under favourable

conditions in the field it may be so floriferous that the

flowers almost obscure the branches. Up to 12 flowers

are produced in fascicles from each flowering eye, all

developing together. It is presumed that under normal

conditions of warmth and sunshine, all flowers would

be open together, but this has not actually been

observed. The first plants collected by Mr W. J.

Hanekom flowered after a short period in cultivation

at the Botanical Research Institute in Pretoria and

were not considered to be entirely normal. In 1975

Mr Hanekom brought two specimens direct from the

veld, one in the initial stages of flowering and the

other with several follicles. A few flowers were open

with the lobes erect-spreading, whereas those not open

showed a tendency for the corolla-lobes to remain

slightly cohering at the tips. The papillae-like hairs on

the inside of the corolla have an unusual shape, being

bent approximately at right angles about the middle

and acuminate. The corolla is dark maroon to black.

In spite of enquiries among Cape collectors, no

record of the species other than the present one seems

to have been made. One wonders how it happens that

such an interesting species should remain unheralded

for so long.

The fruiting specimen brought by Mr Hanekom

showed the follicles arising from flowering eyes

towards the base of the branches, whereas the flowers

on the other specimen were from younger flowering

eyes towards the tips of the branches. This illustrates

a feature of Stapelieae, where the fertilized ovary

undergoes a protracted resting period, after which the

follicles elongate fairly rapidly to maturity. The

number of follicles is usually very small by comparison

with the large number of flowers. This is a result of

the highly specialized and frequently unsuccessful

mechanism for insect pollination. But nature has

introduced compensations by providing an abundance

of seeds in each follicle: each seed is attached to a

very light and delicate cluster of hairs, which act as

parachutes for wide seed dispersal.

R. A. Dyer

TWO NEW SPECIES OF BRACHYSTELMA

Brachystelma canum R. A. Dyer, sp. nov., B.

macropetalo (Schltr.) N.E. Br. affine, foliis oblongo-

ellipticis breviter pubescentibus canis, floribus minori-

bus lobis intus glabris, corona breviter campanulata,

lobis interioribus parvis pulvinatis differt.

Herba erecta, usque 30 cm alta, breviter pubescens,

basi 1-2 ramosa, (? radicibus fasciculatis incrassatis),

internodiis 2-4 cm longis. Folia elliptico-oblonga,

2-4 cm longa, 5-10 mm lata, sensim basin versus

attenuata, molliter pubescentia cana, marginibus

leviter undulatis, plus minusve plicatis. Flores (1)

2 extra axillares producti; pedicelli circiter 1 cm longi.

Sepala lineari-lanceolata, circiter 4 mm longa. Corolla

9-11 mm longa, basi 1,5 mm connata; lobi lineari-

lanceolati, extus pubescentes, intus glabri, leviter

replicati. Corona 0,25 mm supra corollae basin

expansa, campanulata, 1,75-2 mm diam., 5-saccata

margine integro; coronae lobi interiores parvi

pulvinati.

Type: Transvaal, 2625 (Delareyville): 25 km N.E.

of Setlagoli, on road to Mafeking, sandy Terminalia-

veld (-AB), rare, about 1300 m alt., 16th February

1956 /. P. H. Acocks 18774 (PRE, holo.).

Herb erect, up to about 30 cm tall, with underground

stem up to 6 cm long or more (roots not recorded but

assumed to be succulent, fascicled, more or less

cylindric or fusiform); branches 1-2 from the base,

unbranched above, shortly and softly pubescent,

basal internodes 2-4 cm long becoming gradually

shorter as growth continues. Leaves elliptic-oblong,

2-4 cm long, 5-10 mm broad, gradually tapering to

the base with no distinct petiole, finely pubescent, grey,

margins slightly undulate inclined to fold upwards.

Flowers usually 2 together, extra axillary, opening in

succession. Calyx divided to base, with segments

linear-lanceolate, about 4 mm long, pubescent.

Corolla 9-11 mm long, lobed almost to the base

(united for 1,5 mm); lobes linear-lanceolate, with

VARIOUS AUTHORS

255

margins somewhat replicate, finely pubescent on outer

surface, glabrous on inner surface. Staminal column

1 ,5 mm high. Corona appearing to arise 3- up staminal

column 1,5 mm high, more or less campanulate,

1,75-2 mm broad, forming 5 pockets with outer

margin entire; inner corona lobes represented by 5

swellings at base of filaments; filaments conspicuous,

oblong with slight narrowing at the middle. PoUinia

subpyriform, about 0,25 mm long with very short

connectives to carrier. Fig. 5.

Fig. 5. — Brachystelma canum. 1, staminal column and

corona, Xl5; la, outer margin of outer corona-pocket;

lb, rudimentary inner corona lobs pressed on base of

anther; lc, anthers; 2, pollinia, with carrier, X60.

( Acocks 18774).

This is a further species known only from a single

specimen collected by J. P. H. Acocks nearly twenty

years ago. Because of the sandy nature of the soil

the root system was comparatively deeply buried

below ground level and the collector did not record

whether it was a single tuber or a cluster of fleshy roots.

The habit of growth and floral structure indicate a

close relationship with B. macropetalum (Schltr.)

N.E. Br., and for this reason it is suggested that the

plant has a cluster of fleshy roots. This was one of the

characters on which Schlechter in Bot. Jahrb. 20,

Beibl. 51:50 (1895), established the genus Brachystel-

maria as distinct from Brachystelma R. Br., but N.E.

Brown in FI. Cap. 1908 did not uphold Schlechter’s

segregations.

Brachystelma ngomense R. A. Dyer, sp. nov., herba

tuberosa prostrata; corolla circiter 10 mm longa, infra

medium connata tubulata; corona late campanulata,

4 mm diam. 2,5-3 mm alta, margine incurvata

distinguitur.

Tuber subglobosa, 2-3 cm diam.; rami 1-3 annui

prostrati, usque 10 mm (cult. 20 mm) longi, sparsim

ramulosi, glabri vel minute pubescentes. Folia ovata

vel lanceolata, usque circiter 1,5 cm longa (cult.

2 cm x 1,5 cm), breviter petiolata, glabra. Flores 1-2

extra axillares producti; pedicelli 5-10 mm longi.

Sepala lanceolata, 2-3 mm longa. Corolla circiter

10 mm longa, infra medium connata, late campanu-

lata: tubus circiter 5 mm diam., intus albidus rubro-

maculatus; lobi ovato-triangularcs, ± 5 mm longi,

patentes, leviter replicati, atropurpurei, apice albido.

Corona exterior campanulata, 4 mm diam., 2,5-3 mm

alta, 5-marsupiiformis, margine incurvata, lobis

interioribus confluens; lobi interiores breves. PoUinia

±0,25 longa, subpyriformes, caudiculis brevissimis.

Type: Natal, 2731 (Louwsburg), Ngotshe district,

Ngome (-CD), near sawmill on rock-sheets, 7

December 1975, Hilliard & Burtt 8441 (PRE, holo.;

NU).

Tuber subglobose, 2, 0-3,0 cm diam., with 1-3

annual stems. Stems slender, prostrate, glabrous or

minutely pubescent, sparsely branched, about 10 cm

long (up to 20 cm under cultivation). Leaves ovate to

lanceolate, up to about 1 ,5 cm long (2 cmx 1 ,5 cm

under cult.) glabrous. Flowers 1-2, subaxillary;

pedicels ±1 cm long, ascending. Sepals lanceolate,

2-3 mm long. Corolla with the tube slightly bulging

in bud, 1 cm long, divided to about way; tube

campanulate, ±5 mm diam. with flatfish base, slightly

bulged inwards below the sinuses, making the mouth

obtusely 5-angled; white within and purple mottled;

lobes triangular-ovate, ±5 mm long, spreading-

recurved, concave on upper side towards base,

replicate towards apex, dark red or maroon, with the

apex a small white mucro. Corona campanulate,

4 mm diam., 2, 5-3,0 mm high, overtopping the

staminal column, forming 5 deep pockets, with

minutely toothed incurved outer margin, confluent

with the short incurved inner corona-lobes; inner

corona-lobes pressed to base of anthers. Pollinia

±0,25 mm long, subpyriform, with translucent upper-

inner margin, attached by very short caudicles to

carrier with minute lateral bulges. Fig. 6.

Fig. 6. — Brachystelma ngomense. 1, twig, natural size; 2,

corona, x7,5; 3, pollinia, with carrier, x60. ( Hilliard &

Burtt 8441).

This species was discovered by Dr Olive Hilliard

and Mr Bill Burtt on a collecting expedition in

December 1975, when they were asked to make records

of any form of Brachystelma pulchellum (Harv.)

Schltr. which they might encounter. B. ngomense

seems to have no close relationship to any known

species. Several tubers were collected in Natal near

Ngome in a small area on the edge of sheets of exposed

rock, but no plants were seen in the open veld.

The corolla-lobes were described as rich dark satiny-

red, while inside the tube was white with crimson

mottling; the corona was dull mustard, overlaid red.

A tuber flowered at the Botanical Research Institute in

February 1976, when it was figured for eventual

publication in Flowering Plants of Africa.

R. A. Dyer

CEROPEGIA MAFEKINGENSIS, A NEW COMBINATION

In Flora Capensis 4,1:854 (1908) N. E. Brown

described the species Brachystelma mafekingense. He

based it on a small branch preserved in formalin

received from Dr S. Schonland at the Albany Museum,

Grahamstown. Dr Schonland, later Professor of

Botany, Rhodes University, had received a plant from

Mr Graham Green at Mafeking in November 1906

and had divided it between the Albany Museum

256

NOTES ON AFRICAN PLANTS

collection and the Royal Botanic Gardens Kew. A

scrap of material of less than 2 cm high has remained

in a capsule for nearly 70 years in the Albany Museum

with no record of its distribution other than near

Mafeking in the north-western Cape on the border of

Botswana.

My work on the genus Brachystelma has been

disjointed, because it has been a sporadic time filler

between many other projects. Nevertheless, from time

to time in recent years I have been specially curious to

know what relationship B. mafekingense has to other

species, since from Brown’s description it seemed to

be an odd man out.

Without this problem in mind, however, the Dinter

material of Brachystelma in the South African

Museum Herbari am, was borrowed from the National

Botanic Gardens, Kirstenbosch in 1969. Two dwarf

specimens puzzled me particularly at the time,

Dinter 2701 and an unnumbered one. Dinter 2701

from near Grootfontein in South-West Africa had

first been named in manuscript by Dinter as Ceropegia

sp. which he later changed to Blepharanthera nigra.

The second specimen although manifestly the same

species, bears the name B/epharoste/ma avasmontanum

Dinter. This was from the Auasberge south of Wind-

hoek. A note on the sheets indicated that P. G. Green-

way had examined the specimens in 1924 and had not

been able to name them nor was E. P. Phillips prepared

to do so when he studied them 1940. Because of my

own indecision in classifying them in 1969, no mention

of them was made in my article, including Dinter’s

other specimens, in Bothalia 10,2:373 (1971).

Belatedly in 1974, 1 became aware of the publication

of a species under the name Ceropegia patriciae

W. Rauh and G. Buchloh in Kakteen und andere

Sukkulenten (1964). This species was based on a

solitary specimen collected near Hammanskraal some

40 kilometres north of Pretoria on a joint expedition

by Dr Rauh and Dave and Patricia Hardy. In spite of

the 1600 km odd gap in the distribution records,

I had no hesitation in identifying Dinter 2701 and its

companion as specifically equal to Ceropegia patriciae.

The recorded distribution of C. patriciae from

northern and central South-West Africa via the

north-eastern Cape to the northern Transvaal is

by no means unique. Species such as Brachystelma

dinteri Schltr.have a very similar recorded distribution

pattern.

In 1975 all the material of Brachystelma was

borrowed from the Albany Museum, Grahamstown.

It did not take long to establish that B. mafekingense

N.E. Br. is conspecific with Ceropegia patriciae

Rauh & Buchloh, with the former specific epithet

taking priority. The question arose into which of the

two genera the species should be classified. Both

Brachystelma and Ceropegia embrace a wide range of

corolla and corona forms and the relative lengths of

the corolla-tube and -lobes are the main distinguishing

features. The distinction between the genera is not

always clear-cut such as in the present case but, in my

opinion, the weight of evidence is in favour of

Ceropegia , with C. pygmaea Schinz being the nearest,

yet distant, known relative. This makes it necessary to

validate the new combination Ceropegia mafekingensis ,

which is done below. Among the material of the genus

Brachystelma borrowed in 1975 from the Bolus

Herbarium were two specimens collected by Dr

F. Z. van der Merwe, Nos 46 and 92. They were sent

to the Bolus Herbarium in 1942 from a bushy plain

near Zeerust only about 50 km from the type locality

of B. mafekingense, which they were found to match

very closely.

Ceropegia mafekingensis (N.E. Br.) R. A. Dyer

comb. nov. Type: Cape, near Mafeking, Green sub

GRA 1683 (K, holo.; GRA1).

Brachystelma mafekingense N.E. Br. in FI. Cap. 4,1 : 854

(1908).

Ceropegia patriciae Rauh & Buchloh in Kakteen und andere

Sukkulenten 8:151 (1964). Type: Transvaal, near Hammans-

kraal, Rauh 12369 (HEID). C. sp. Dinter ms. (No. 2701).

Blepharanthera nigra Dinter ms. (No. 2701).

Blepharostelma avasmontanum Dinter ms. (specimen without

number).

Tuber ±8 cm diam., more or less depressed, 0-7 cm

below ground, producing one or mo;e annual stems.

Stems branching from near base and appearing tufted,

1,5-5 cm high, puberulous. Leaves tapering into a

short petiole or rounded at the base, oblong-lanceolate,

spreading, 7-30 mm long, up to 7 mm broad, with

margins half folded upwards, ± undulate, puberulous

below, glabrous above. Flowers in 10-20-flowered

umbellate cymes, the last terminal; pedicels 3-8 mm

long, puberulous. Sepals narrowly lanceolate, 2-4 mm

long. Corolla 10-12 mm long, divided to more than \

way; tube 4-6 mm long, 3-3,5 mm diam., obtusely

5-angled, slightly narrowed to the mouth, outside

minutely puberulous or papillate on the upper part,

maculate with dark purple, glabrous within, yellow

and purple punctate; lobes 6-7 mm long, free at tips,

connivent at first, becoming slightly spreading,

obovate-oblong, ±2 mm broad, auriculate at base,

inflexed, acute at apex, margins replicate, minutely

puberulous outside, glabrous, verrucose, and blackish

on inner surface. Corona campanulate at base;

outer lobes forming 5 pockets with the outer margin

extending into lobules behind the base of the inner

corona-lobes; inner corona-lobes arising from within

the base of the outer corona, linear, obtuse, 2 mm

long, incumbent-erect, much overtopping the staminal

column with obtuse connivent tips minutely papillate.

Pollinia globose-pyriform, ±0,25 mm long, with

translucent beaked upper margin, with short, delicate

caudicles and small carrier.

Cape. — 2525 (Mafeking): near Mafeking, Nov. 1906, Green

sub GRA 1683 (GRA).

Transvaal. — 2526 (Zeerust): bushy plain near Zeerust, Van

der Merwe 46; 92 (BOL). 2528 (Pretoria): Hammanskraal,

Rauh 12369 (HEID).

S.W.A. — 1918 (Grootfontein): near Grootfontein Dinter

2701 (SAM); 2217 (Windhoek) Auasberge, Dinter s.n. (SAM).

Two statements in Brown’s original description

helped to confuse my concept of the species. Firstly,

the statement that the flowers were in a terminal,

10-12-flowered umbel and secondly, that the stems

were “richly branched from near the base 4-7 cm

above the tuber (Schonland)”. The flowers are falsely

terminal and the 4-7 cm length refers to the depth of

the tuber below ground level. The height of the stem

above ground of the type was only 1,5-2 cm. Rauh

and Buchloh state that the stem may be up to 5 cm

long in C patriciae, but in the single specimen they

cite the tuber was more or less at ground level.

In conclusion, for the historical record. Miss M.

Gunn was able to supply the following information

about Mr Graham Green, who collected the type

material of C. mafekingensis. Mr Green was appointed

as a clerk in the Cape Colonial Service in 1887.

Most of his appointments were in the eastern Cape,

including Grahamstown. He was promoted to the

post of magistrate of Mafeking in 1903, whence in

1906 he despatched the type material to the Albany

Museum in Grahamstown.

I wish to express my appreciation of the loan of the

material, which made this investigation possible.

R. A. Dyer

VARIOUS AUTHORS

257

CYPERACEAE

CYPERACEAE NEW TO THE FLORA OF NATAL

Since the publication of Ross’s Flora of Natal ( 1 972)

two species of Cyperaceae new to the province have

been recorded:

Courtoisia cyperoides (Roxb.) Nees ( Vorster 2702;

2703) was found in April 1975 in the Mkuze Game

Reserve along the borders of pans, growing on wet

mud.

Material of Mariscus squarrosus (L.) C.B. Cl.

( Vorster 2700) was collected at the same time and

place. Another specimen, Arnold 865, was collected

about 20 km from Cathedral Peak on the road to

Estcourt in March 1975. In naming these specimens 1

follow Hooper in FI. W. Trop. Afr. ed. 2, 3:294

(1972), who reduced M. aristatus Rottb. to a synonym

of M. squarrosus.

Both these species are widely distributed throughout

the hot, dry areas of Africa and Asia. Although they

occur mostly in areas of low rainfall, they never occur

far from seasonal pans usually growing intermingled

with each other on wet mud along the shores. Both

species are known from Botswana, South West Africa

and the drier parts of Transvaal, and M. squarrosus

also from the high, dry parts of Lesotho.

P. Vorster

FABACEAE

SOME OBSERVATIONS ON ELEPHANTORRHIZA BURKEI AND E. ELEPHANT1NA

In a revision of the genus Elephantorrhiza Benth., in

Bothalia 11:252 (1974), it was stated that “£. burkei

appears to have smaller seeds than E. elephantina, but

more fruiting material is required to confirm this.”

In an attempt to glean further information on this

matter, populations of these two species growing

naturally in the Pretoria National Botanic Garden

were kept under observation from spring 1974 until

the end of winter 1975.

In E. elephantina the flowering racemes are usually

confined to the lower part of the stem with the result

that the pods are normally suspended just above the

ground or else rest on the ground where they tend to

be inconspicuous. In E. burkei, however, the flowering

racemes are borne on the branched stems some

distance from the ground and so the pods are conspi-

cuous, especially when mature or approaching

maturity.

Seeds of both species were collected, particularly

those of E. burkei about which less was known and

which were far more numerous. Considerable variation

in the shape of the seeds was noted in E. burkei,

seeds varying from elliptic to almost quadrate, the

latter as a result of the seeds being tightly compacted

and laterally compressed in the pods. The seed

dimensions recorded for E. burkei were 9-15 X 8-12 X

5-7 mm and those for E. elephantina 17-26X 1 3—1 8 x

6-13 mm. These measurements confirmed the

suspected existence of a difference in seed size between

the two species, the seeds of E. burkei apparently

being consistently smaller than those of E. elephantina.

As indicated in Bothalia l.c., the pods of E. burkei

tend to be longer and narrower than those of E.

elephantina.

At maturity the valves of the pods in both species

separate from the margins, the margins persisting as

an almost continuous but empty frame in a manner

reminiscent of most species of Entada (see illustration

in Palmer & Pitman, Trees S. Afr. 2:827, 1973). This

is particularly conspicuous in E. burkei where the

valves roll back and together with the margins usually

persist on the plant for many months. In E. elephantina

the pods tend to disintegrate and disappear more

rapidly.

A considerable amount of variation in the size of

individual plants was noted in the population of

E. elephantina. Scattered in amongst the population

were several plants that were substantially larger

than the surrounding plants, almost as though they

had been the recipients of a heavy application of

fertilizer. A cytogenetical study of this population

may yield interesting results and would give an

indication of whether or not polyploidy is involved.

J. H. Ross

FRULLANIACEAE (HEPATICAE)

TYPIFICATION OF FRULLANIA SYLVESTRIS

When Sim described Frullania sylvestris in Trans.

Roy. Soc. S. Afr. 15:40 (1926), no specimens were

cited. In the only subsequent work dealing with this

species, Arnell, Hepaticae of S. Afr. 138 (1963), no

type was designated or any specimens cited, although

the distribution is cited as “ Knysna. Natal,

several localities.”

In the main collection of Sim’s specimens, which is

housed at PRE, there is only one specimen of this

species, Sim s.n. in PRE Cryptogamic Herbarium

(Hepaticae) 1740, annotated as follows in Sim’s own

hand:

“ Frullania sylvestris Sim

Giants Castle, Natal. 5000'

Sept 1903, T.R. Sim.”

I conclude that Sim’s description was probably

based on this specimen, and consequently regard it

as the holotype.

P. Vorster

258

NOTES ON AFRICAN PLANTS

POACEAE

ODONTELYTRUM: A NEW GENERIC RECORD FOR SOUTH AFRICA

During an aquatic weed survey in December, 1974,

Mr F. de W. Brand, a member of the Transvaal

Provincial Administration, collected an aquatic grass,

Odontelytrum abyssinicum Hack., in the municipal

dam at Belfast (c. 2 000 m) in the eastern Transvaal.

This gathering, Brand 44, represents a new generic

record for South Africa. It has subsequently been

re-collected by the author at Belfast, Du Toit 1083.

The monotypic genus Odontelytrum , first described

by Hackel, was based on a Schimper specimen,

collected in Ethiopia (Abyssinia) in stagnant water

between Paffat and Dewra Tabor (c. 2700 m). The

nearest known locality outside Ethiopia is Oljoro

Nyuki on the Ngorongoro crater floor (c. 1900 m), in a

Cypm/s-swamp in Tanzania. The grass seems to prefer

a cool climate in stagnant or slow-flowing water. It is

possible that this grass spread to South Africa via

the eastern mountain ranges being dispersed by

waterbirds. This view is strengthened by the

observations of Mr Brand, who stated that waterbirds

relish the seeds and young shoots.

Although superficially resembling Oryza longistami-

nata, the wild rice, and Oryzidium barnardii, Odonte-

lytrum is closely allied to Pennisetum, a genus with

which it cannot be confused. On close examination,

the long “awn” of Odontelytrum proves to be an

elongate bristle of the involucre surrounding the

spikelet. In the case of Oryza and Oryzidium, the awn

arises from the lemma.

The specimen from the eastern Transvaal was

identified in time for the genus to be included in

Volume 2 of The Genera of Southern African

Flowering Plants by R. A. Dyer (1976). Readers are

referred to this publication for a full description of the

genus. To facilitate identification and to encourage

further observations on its distribution, a brief

diagnostic description is given here.

Odontelytrum abyssinicum Hack, in Oestr. Bot

Zeitschr. 48: 86 (1898); Stapf in Hook.Ic.P1.31.t. 3074

(1922). Type: Aethiopia, between Paffat and Dewra

Tabor, Schimper 1121.

Aquatic perennials, with soft spongy culms, leaves

light green or purplish, minutely scabrid. Inflorescence

a stout raceme, sheathed at the base. Spikelets

solitary, purple or straw-coloured, and enveloped by

an involucre of partially fused bristles. Spikelets

two-flowered, the lower male, the upper bisexual.

At maturity spikelets fall entire with the involucre.

P. C. V. Du T oi

RANUNCULACEAE

THE CORRECT NAME FOR ANEMONE CAPENSIS

In his write-up of Anemone capensis in Flower.

PI. Afr. 40: t. 1569 (1969), Oliver attributed the

binomial to Lamarck (1783), pointing out that

Lamarck made no mention at all of Atragene capensis

L. (1753) in his description of Anemone capensis in

Encycl. 1 : 164 (1783) and therefore Lamarck was the

sole author of Anemone capensis. Oliver also stated

that the combination Anemone capensis (L.) Harv.

(1860) based on Atragene capensis L. (1753) was a

later homonym of Anemone capensis Lam. and there-

fore illegitimate.

This is true, but unfortunately Oliver overlooked

Atragene tenuifolia L.f. (1781), which is conspecific

with Anemone capensis Lam. and antedates it by two

years. I have not been able to trace the holotype of

Atragene tenuifolia, but Thunberg specimen No. 12999

in Herb. Thunberg, Uppsala, which is probably an

isotype, agrees with our concept of Anemone capensis.

The correct name to use, therefore, is Anemon

tenuifolia (L.f.) DC. (1824).

The relevant synomy is as follows —

Anemone tenuifolia (L.f.) DC., Prodr. 1 : 18 (1824).

Type: Cape, Thunberg (holotype not traced, but

isotype probably specimen No. 12999 in Thunberg

Herbarium, UPS; PRE, photo.!).

Atragene capensis L., Sp. PI. 543 (1753). Type: Cape (LINN

711.3; PRE, photo). A. tenuifolia L. f., Suppl. 270 (1781), as

above. A. tenuis Thunb., FI. Jap. 239 (1784). Type: none cited.

Anemone capensis Lam., Encycl. 1:164 (1783); Oliver in

Flower. PI. Afr. 40 :t. 1 569 (1969). Type: Cape (specimen in

Herb. Jussieu, P). A. arborea Hort. in Steud. Nom. ed. 95 (1840),

nomen nudum. A capensis (L.) Harv. in FI. Cap. 1 : 3 (1859-60),

nom. illeg. ; Adamson in FI. Cap. Penins. 401 (1950). A capensis

(L.) Harv. var. tenuifolia Harv., l.c.

Pulsatilla africana Spreng., Syst. 2: 664 (1825). Type: none

cited. P. tenuifolia (L.f.) Spreng., l.c.

D. J. B. Killic

RUTACEAE

ZANTHOXYLUM LEPRIEURII*, A NEW RECORD FOR SOUTHERN AFRICA

The first specimen of this Zanthoxylum from

Southern Africa was collected by Drs Codd and Dyer

(No. 4564) in the Sibasa district of the Transvaal

near Punda Milia in the Kruger National Park in

October 1948. However, in the absence of fertile

material, this specimen, like each of those collected in

the Kruger Park and in north-eastern Natal (Tonga-

land) during the next twenty years, remained unidenti-

fied. In December 1969 good flowering material of

both male and female plants was collected (A/o//4887,

4889) near Makanes Drift in Tongaland and it is

these specimens which finally enabled the plants to be

identified. There is still no fertile material from the

Transvaal, although there is a photograph in the

National Herbarium, Pretoria, of a fruiting branchlet

collected by Mrs E. Jenkins north-east of Sibasa in the

northern Transvaal.

’•‘The taxonomic position of the genera Zanthoxylum L. and

Fagara L. has long been a source of controversy. Most early

taxonomists, for example, Benth. & Hook.f., Gen. PI. 1 : 297

(1862), regarded the two as congeneric, but later opinion

favoured the argument advanced by Engler in Pflanzenfam.

3 , 4 :95— 102 (1 896), which supported the retention of both genera.

In recent years evidence has accumulated which supports the

view that the genera are not distinct. Following Hartley, in

J.Arn.Arb. 47:171-221 (1966), and Waterman, in Taxon

24:361-366 (1975), Fagara and Zanthoxylum are here regarded

as congeneric, Zanthoxylum being the correct name for the

combined genera.

VARIOUS AUTHORS

259

The specimens are referable to Z. leprieurii Gu ill. &

Perr., a species described from Senegal, but widespread

in the forests of west tropical Africa to the Equatoria

Province of the Sudan, Uganda, Tanzania, northern

Mozambique, Zaire and Angola. Although not

specifically distinct, the Southern African plants do

nevertheless differ from typical Z. leprieurii in certain

characters.

In typical Z. leprieurii the inflorescence is a very

robust structure up to 30 cm long with the axis 2-3 mm

in diameter basally, and the axis and its many lateral

branches are sparingly to densely glandular-puberu-

lous, particularly the lateral branches. It is not clear

whether the inflorescence is carried erect or whether it

is pendulous as none of the collector’s notes refer to

this feature. The inflorescence in the Tongaland plants

(and presumably in the Transvaal) is smaller, seldom

attaining a length of 20 cm and the axis is slender

and ±1 mm in diameter basally. The inflorescence

in these plants is pendulous and glabrous, and the

lateral branches are fewer and more widely spaced so

that the inflorescence forms a more open lax panicle.

The flowers are more consistently aggregated into

fascicles along the lateral branches, particularly in the

male, and there are fewer flowers per lateral branch.

This contrasts with the tropical material where the

flowers are more often solitary.

The Southern African specimens also differ from

material of typical Z. leprieurii in having consis-

tently smaller leaves with fewer pairs of smaller

leaflets, but there is no discontinuity in any of these

characters between specimens from the two areas.

Typical Z. leprieurii is a forest species which grows

to a height of 27 m, while the Southern African plants

grow on sandy soils as shrubs or small trees and

seldom attain a height of 8 m. Plants are known from

only two localities in Tongaland, namely, False Bay

Park and the Makanes Drift-Sihangwane area where

they form one of the constituents of the sand-forest

and grow in association with species such as A/bizia

forbesii Benth., Newtonia hildebrandtii (Vatke) Torre

and C/eistanthus schlechteri (Pax) Hutch., to name but

a few. Only six male plants have been found at False

Bay Park, but in the Makanes Drift-Sihangwane

area plants are locally quite common.

In Tongaland the specimens of Z. leprieurii are

entirely leafless in winter. In spring the young leaves

and inflorescences appear together, but while the

leaves develop the young flower buds remain closed

for several months before finally opening. The male

inflorescences are relatively long and conspicuous and

project beyond the leaves, while the female inflores-

cences are usually smaller and are sometimes partially

hidden in amongst the leaves.

Zanthoxylum leprieurii Guill. & Perr., FI. Seneg.

Tent. 141 (1831). Type: Senegal, Leprieur (P, holo.!).

Fagara leprieurii (Guill. & Perr.) Engl, in Pflanzenfam-

3,4:118 (1896). F. angolensis Engl, in Bot. Jahrb. 23 : 148 (1896).

Type: Angola, without locality, Welwitsch 4575 (K, iso.!).

F. nitens (Hiern) Engl., Pflanzenw. Afr. 3,1 :749 (1915). F. sp.,

Ross, FI. Natal 213 (1973); Palmer & Pitman, Trees S. Afr.

2:982-3 (1973).

Zanthoxylum nitens Hiern, Cat. Afr. PI. Welw. 1:112 (1896).

Type: Angola, Golungo Alto, road to Bango, Welwitsch 4558

(K, iso. !).

The following specimens from Southern Africa

have been examined: —

Transvaal. — 2231 (Pafuri): Kruger National Park, 5 km

N.E. of Punda Milia, Codd & Dyer 4564; Kruger National Park,

Punda Milia, Van der Schijff 2)30; 3803.

Natal. — 2632 (Bela Vista): 8 km W. of Muzi border post,

Moll 5654 (NH). 2732 (Ubombo): 3,2 km N.E. of Makanes

Drift, Moll 4887; 4889; Ross 2368; near Jobe’s Kraal, Garland

442; 10 km Pongolo bridge/Maputa, Moll 4382; Sihangwane,

Tinley 901 (E,NU); Pooley 727 (E, NU); 2 km W. of Sihangwane

Store, Moll 4880; False Bay Park, Ward 3850; Ross 2229.

J. H. Ross

SELAGINELLACEAE (LYCOPSIDA)

SELAGINELLA IMBRICATA IN THE FLORA OF SOUTHERN AFRICA AREA

During August 1956 a specimen of Selaginella

imbricata ( Story 5841) was collected in the Kapupa

Valley (12°34' E, 17°21' S) in the northwest of South

West Africa. This species was not included in Prodr.

FI. S.W. Afr. The presence of S. imbricata in South

West Africa is rather unexpected, as this species

usually occurs in rather moist forest situations. The

specimen was reported as occurring “along fissures in

rocks in mountain gorge”. Judging from the geo-

graphical situation, the area where the specimen

was collected, must be extremely dry.

The find is of interest, firstly because it adds another

species to the flora of South West Africa, and secondly,

because it extends the known range of the species

westwards by 1300 km. The nearest point from which

it was previously known, is the Victoria Falls.

This is the first time the genus has been recorded

from South West Africa, although a number of species

occur in adjacent territories.

Two other specimens in the National Herbarium,

also proved to be S. imbricata (Forsk.) Spring ex

Decne. Apparently these two specimens, Galpin s.n.

in PRE Cryptogam ic Herbarium (Pteridophyta)

3532 from Devil’s Hoek in the Royal National Park

in 1928, and Esterhuysen 10823 from the “Turret”

and “Amphlet” in the Cathkin Peak area in 1956,

represent the only two collections from which this

species is known in South Africa. This is about

1500 km south of the nearest known locality near

Melsetter in Rhodesia.

Selaginella imbricata is widespread throughout the

eastern part of Africa, from the escarpment between

Mozambique and Rhodesia in the south to Arabia in

the north. It occurs westwards along the Zambesi

valley to at least as far as the Victoria Falls and, as

already mentioned, is found in the extreme north-

western part of South West Africa.

It would seem that the plant must be very rare in

South Africa to have escaped more frequent collection,

as it is quite a distinct and conspicuous species. It

can easily be distinguished from other South African

species because of the rhomboid outline of the

branched stem system, which stands erect, as it is

rather stiff, and is arranged in one plane like a fern

frond, instead of being procumbent and closely

adpressed to the substrate. It occurs in moist, shady

forest situations against earth banks. Under dry

conditions the plant rolls up into a series of ball- or

clockspring-like structures, but unrolls again after

rains.

While some species of Selaginella occur under moist

conditions and cannot withstand drought, many

species are xerophytes of the “resurrection” type.

Such species are dry and inactive and more or less

rolled up during the greater part of the year. As soon

as they are moistened, however, the leafy branches

unfold and the tiny leaves attain a green colour once

again. S. imbricata is intermediate in this respect.

Its habitat is generally not dry at all, but it is capable

of surviving dry spells by going into quiescence.

P. VORSTER

260

NOTES ON AFRICAN PLANTS

THELYPTERIDACEAE

NEW COMBINATIONS IN THELYPTER1S

Holttum in J. S. Afr. Bot. 40: 123—168 (1974)

stated that the estimated 1 000 species in the family

Thelypteridaceae can either be regarded as belonging

to a single genus, Thelypteris, or they can be placed in

a number of different genera. In Blumea 19: 17-52

(1971) he adopted the latter course by grouping the

450-500 Asian-Malesian species into various genera.

“For comparative purposes”, Holttum (1974) also

classified the 55 African species of the family Thelyp-

teridaceae according to this scheme. He added,

however, that “. . . .for persons who are concerned

only with Africa, or a part of Africa, it is reasonable to

adopt one genus Thelypteris, as Prof. Schelpe has

done in his account of the ferns in Flora Zambesiaca

(1970)”.

I agree with this approach, with the possible excep-

tion of Ampelopteris , a monotypic genus, which can

easily be recognized by the unicellular forked hairs

and numerous gemmae on the fronds.

The characters used to differentiate between the

genera of Thelypteridaceae (pinna veins anastomosing

or not, reduction of lower pinnae, hairiness or scaliness

of fronds) are, in my opinion, not sufficient in the

African context to justify division into genera. The

presence or absence of indusia may be useful for

subgeneric division, but will have to be studied in

conjunction with other characters.

Holttum (1974) pointed out that some of the

African taxa described in that paper still lack names

in Thelypteris. For two of the South African species

these combinations are made below, as they have not

been proposed by Schelpe (in press):

Thelypteris altissima ( Holtt .) Vorster comb. nov.

Christella altissima Holtt. in J. S. Afr. Bot. 40: 141 (1974).

Thelypteris prolifera ( Retz .) Vorster comb. nov.

Hemionitis prolifera Retz., Obs. Bot. 6:38 (1791).

Ampelopteris prolifera (Retz.) Copel., Gen. Fil. 144 (1947).

P. Vorster

Bothalia 12, 2: 261-265 (1977)

World climate patterns in grassland and savanna and their relation

to growing seasons

R. KIRK STEINHORST* and J. W. MORRISf

ABSTRACT

The climate at eleven IBP savanna or grassland study sites from five continents are described and principal

components analysis is used to compare them. A multivariate linear discriminant function based on mean

monthly precipitation, mean monthly temperature, latitude and altitude, is used to predict the length of the

growing season at each site. At most sites, the actual and predicted start and end of the growing season agreed

closely. It is concluded that growing season on a world-wide basis may be predicted fairly reliably from a

small number of abiotic variables by means of a multivariate discriminant function.

RESUME

CYCLES CLIMATOLOGIQUES MONDIAUX EN PRAIRIE ET SA VANE ET LEU R RELATION

AVEC LES SAISONS DE CROISSANCE

Les climats en onze points d' etude de la savane on de la prairie, choisis par le Programme Biologique

International sar les cinq continents, sont decrits et compares par le moyen de V analyse des principaux

composants. Une fonction discriminante lineaire multivariee, basee sur les moyennes mensuelles de la

precipitation et de la temperature , sur hi latitude et T altitude , est utilisee pour predire la longueur de la

saison de croissance en chaque endroit. Dans la plupart des cas, il y a un excellent accord entre le calcul

et T observation pour le debut et la fin de la saison de croissance. On en conclut que cette saison peut,

d l echelle mondiale, faire Vobjet d'une prevision assez sure a partir d'un petit nombre de donnees abioti-

ques et en utilisant une fonction discriminante multivariee.

INTRODUCTION

In this article we consider a variety of savanna and

grassland sites and one tundra site, characterize their

climates and relate climate to the conditions for

growth of the natural vegetation of the sites. By so

doing, we hope to better understand the factors

involved in growing season determination. By growing

season, we mean that period of the year from the

time when environmental factors cease to limit

growth, usually in spring, until they limit it again,

usually in autumn.

With the present world population and limited food

supplies it is important that we use all available land

in an efficient manner for either agriculture, grazing,

wood production, or conservation. As recently

pointed out by Newman & Pickett (1974), the one

uncontrollable factor which limits land productivity

is weather. Success or failure of crops and of forage

production are dramatically affected by whims of

weather. Newman & Pickett suggest that knowledge

of world climate will enable us to predict productive

uses for each area. To apply this strategy, however,

knowledge of how crops and other vegetation respond

to climatic factors is required. For crops, and

monospecific communities in general, physiological

studies give us the required information concerning

responses to light, temperature, precipitation, and

other factors. For savanna, grassland, or tundra

communities of many species, physiological studies

give us no single answer. Another solution is required

if we are to utilize such communities efficiently on a

world-wide basis for grazing or fodder production.

The International Biological Program (IBP) gives

us a unique opportunity to study these communities

of many species. In this paper we include data from

IBP grassland sites and one tundra site (eutrophic

mire) falling in a range from 6° to 60° latitude,

north and south of the equator from five continents

(Table 1). Each site represents typical savanna,

grassland or tundra vegetation under grazing, haying,

or fallow conditions. The data include short-term

specific weather records on at least a monthly basis

and the associated growing seasons as determined

subjectively by the investigators at each site. In order

to predict conditions for growth of the local

vegetation, we consider monthly precipitation and

average temperature along with latitude and altitude.

* Natural Resource Ecology Laboratory, Colorado State

University, Fort Collins, Colorado, U.S.A. (Now at: Institute

of Statistics, Texas A & M University, College Station, Texas).

t Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria, 0001.

TABLE 1. — Study sites and sources of data

262 WORLD CLIMATE PATTERNS IN GRASSLAND AND SAVANNA AND THEIR RELATION TO GROWING SEASONS

CHARACTERIZATION OF CLIMATE TYPES BY

PRINCIPAL COMPONENTS ANALYSIS

There is a wide range of climates, both tropical

and temperate, represented in the sites included in

the study. Of the tropical climate sites, Lamto,

Ivory Coast, has uniform temperatures with heavy

precipitation from March to October with a brief

respite in mid-summer. The two Indian sites have

three seasons; a rainy season from mid-June to

September, a cool dry season from October to

February, and a hot dry season from March to mid-

June. Welgevonden, South Africa, is a dry sub-

tropical savanna site with summer rainfall. The

temperate sites range from 30 to 60° from the equator.

They have seasons dictated by the angle of the sun’s

incoming rays. Armidale, Australia, is a cool

temperate rangeland with moderate precipitation,

which is fairly uniform but heaviest in summer.

Temperate sites w'hich are relatively humid are

Lanzhot, Czechoslovakia: Glenamoy, Ireland:

Terschelling, Netherlands; Hardangervidda, Norway;

and Moor House, U.K. The season of rainfall varies

greatly among temperate sites. For example, Ireland’s

heavy precipitation occurs in November, December

and January, whereas Terschelling has uniform

rainfall throughout the year. The Pawnee Site in the

United States is very dry relative to the other sites.

A number of these sites have climates which are

influenced by geographical features. The Netherlands,

Norway, Ireland, Ivory Coast and United Kingdom

sites are influenced by proximity to the ocean. Both

the Czechoslovakian and the United States sites are

continental. The Pawnee (U.S.) site's weather is

furthermore modified by the north-south Rocky

Mountain Range, a few kilometres to the west.

Czechoslovakia on the other hand is bounded by the

major east-west mountain ranges of Europe. Another

site affected by mountains is the northern India site

(Kurukshetra), whose climate is influenced to some

extent by the Himalayas.

All in all, these sites have a fairly predictable

pattern of monthly temperature based on incoming

radiation, but they present a mosaic of precipitation

regimes. Principal component analysis (PCA) (Seal,

1964) was used to illustrate this contrast. For

temperature, using the twelve monthly values as

attributes over the eleven sites (entities), the first

component (an average temperature component)

accounts for 94% of total variability in the data and

the second component brings the cumulative per

cent to 99%. The second component contrasts

summer and winter temperatures, yielding a range

from very uniform monthly temperatures (Ivory

Coast, Ireland and U.K.) to strongly seasonal ones,

including U.S. A. and Czechoslovakia (Fig. 1). The

first component separates warm sites, such as the

Indian and the Ivory Coast sites, from cold ones such

as U.K. and Norway. As a hierarchical cluster

analysis (Johnson, 1967) on temperature correlation

or seasonal patterns grouped the sites in the same

manner as component two of the PCA, except for

the U.S. site which, for these data, showed a monthly

temperature pattern similar to the Moor House site,

these results are not presented. A hierarchical cluster

analysis using euclidian distance, that is, temperature

magnitudes rather than patterns, yielded clusters

predictably similar to the first component of the

PCA and is also not presented.

For mean monthly precipitation values over the

eleven sites, the first three principal components

account for only 52%, then 75%, then 86% of the

total variability. There is no discernable pattern of

rainfall along latitudinal or altitudinal gradients.

This result follows from the above characterization

of climate types since the precipitation at these sites

is variously affected by oceans, mountains, and other

non-regular causes. In contrast to the sun’s orbit,

which is the principal regulator of temperature,

there is no single unifying phenomenon that controls

precipitation.

DETERMINATION OF GROWING SEASON BY

DISCRIMINANT ANALYSIS

If one were to take a plant physiological approach

to the specification of conditions for growth, one

would consider available solar radiation, maximum

and minimum air temperature, soil temperature,

number of daylight hours (related to latitude),

available moisture and phenological stage. We have

taken four variables as integrators of all these factors,

namely: mean monthly temperature, monthly pre-

cipitation, latitude and altitude. The growing seasons

for the 1 1 sites do not simply commence and end as

the sun progresses to and from the equator. If this

were so, either average monthly temperature or

latitude would suffice to predict growing seasons.

Fig. 1. — First and second princi-

pal components of monthly

mean temperature data for the

eleven sites.

First Principal Component

R. KIRK STEINHORST AND J. W. MORRIS

263

A step-wise discriminant algorithm (Dixon, 1971) was

applied to determine if some combination of these

four variables was a good indicator of growth. Each

month, at each site, was classified as either a growing-

or non-growing-season month according to the

dates defining the length of the growing season

supplied by each investigator. The means for the

four response variables are given in Table 2, and the

pooled covariance matrix is given in Table 3.

TABLE 2.— Mean values (F) over the eleven sites for the four

responses

TABLE 3. — Pooled covariance matrix(E) of variances and

covariances among the four responses

Data from South Africa and Australia were

displaced by six months for the calculations so that

“summer” and “winter” seasons would coincide

with northern hemisphere seasons. Growing season

months clearly have more rainfall and are warmer

(Table 2). The latitude and altitude means differ in the

two groups only in so far as the sites at different

latitudes or altitudes have greater than or less than

6 months of growth. Curiously, there are relatively

more growing season months at greater latitudes

than non-growing months at those latitudes (group

means of 38,9 and 38,1 respectively), although the

difference is not significant.

The most important single variable is precipitation

(t-test between growing and non-growing groups

yield a significance of p <0,0005). However, when

used as an indicator of growth, this variable by itself

incorrectly labels 35 growing season months as

non-growing months while correctly categorizing

84 months. The remaining 13 months were non-

growing months mis-specified as growing months.

The best discriminant function uses either all four

variables, in which 29 of 132 months are mis-specified,

or all variables excluding altitude for which 29 of,

132 months are also mis-specified. Latitude i ; needed

in addition to precipitation and temperature in

order to predict growth more accurately for some

cool sites. Using all four variables, the multivariate

t-test for groups yields a significance of p <0,0005.

While it is clear that the largest difference between

the growing and the non-growing months is their

precipitation, precipitation alone is not sufficient to

predict suitable conditions for growth.

Using all four variables the discriminant function

for group i is

— i/rh T /di £_1;c=ai-f-b1iX

where //, is the four-variate mean for group i in

Table 2, Z is the pooled estimate of the variance-

covariance matrix (Table 3), and x is the four variate

response for a particular site and month. For the

growing season group, ax= — 27,6731 and b\=

(0,02709, 1,20478, 0,71537, 0,00781); for the non-

growing season months one obtains a2= — 19,60077

and b\=(0, 01511, 0,99291, 0,61318, 0,00681).

After cancelling constant terms the posterior

probability of being in group i is:

aj + bfix

n\t

ai ~\-b'l1x a 2+b12x

nxe +ft2e

where n \ and n2 are the prior probabilities of being

in the growing and non-growing groups, respectively

(Dixon, 1971). Both n1 and n2 are taken to be one-half

in this case. Prior probabilities are those based on the

assumption or hypothesis under investigation, in this

instance, that a month has an equal chance of being

either a growing or a non-growing month. The two

posterior probabilities, in this instance, are the

chances of a month being a growing and non-growing

month, calculated by substitution in the above

equation of the month’s precipitation, temperature,

latitude and altitude values using (al5 blx) and

(a2, b 12), respectively. Posterior probabilities were

calculated for each month at each site and each

month as classified as growing or non-growing,

depending on which posterior probability was greatest.

In Fig. 2 the growing seasons as given by the

investigators are compared with the calculated growing

seasons from the four-variable discriminant analysis.

There are a total of 103 of 132 months correctly

specified, whereas 17 growing season months were

incorrectly classified as non-growing months and

12 non-growing season months were specified as

suitable for growth. Half of the misclassified months

were for tropical countries where growth would

appear to be regulated by something in addition to

rainfall and temperature. Predictions for both Indian

sites and the Ivory Coast were particularly poor while

those for South Africa and Australia are acceptable.

On the other hand, the growing season was precisely

predicted at Hardangervidda. Other temperate

locations were treated relatively well, except that the

linear discriminant technique was not able to

accurately predict initiation of growth at Glenamoy,

Ireland.

DISCUSSION AND CONCLUSIONS

Although the 1 1 sites represent vastly differing

climates, including various combinations of tem-

perature and moisture regimes, use of multivariate

linear discriminant analysis suggests that, over a large

portion of the globe, growing season is basically

a function of precipitation and temperature with

some modification due to latitude. Given these three

variables, altitude may be of slight additional signi-

ficance. Growth in tropical areas is dependent upon

other or additional factors, including possibly mean

monthly minimum temperature. The unsatisfactory

predictions could also be accounted for by the use of

one or two years’ records instead of long-term

averages in the calculations. The success in correctly

classifying conditions conducive to growth using a

multivariate linear function of three or four variables

suggests that a plant community approach, as opposed

to species-specific physiological approaches, can be

successful.

It is possible to predict conditions suitable for

growth at the community level on a global basis using

a few abiotic variables. Future research can be directed

at finding particular response curve relationships

between initiation and termination of growth and

264 WORLD CLIMATE PATTERNS IN GRASSLAND AND SAVANNA AND THEIR RELATION TO GROWING SEASONS

_____ discriminant analysis

actual

S.lndia

USA

UK

S- Africa

Norway

Netherlands

Ivory Coast

Ireland

N. India

Czechoslovakia

Australia

Fig. 2 — Comparison of actual and predicted length of growing season at each site.

"i m r

J T J

Months

i a r

"I O I N I D~

these factors. Until such research is done, the dis-

criminant function used here gives us a reasonable tool

for determining growing season. Management strate-

gists can use this function to predict the possible

outcomes of various strategies under conditions which

have not occurred previously. For example, if one

anticipates a cool wet spring, representative pre-

cipitation and temperature figures can be supplied

and the effect on the start and end of the growing

season may be predicted.

ACKNOWLEDGMENTS

Authors who contributed data for this study are

thanked. We also wish to thank J. Dodd, K. P.

Haddock, J. G. W. Jones, S. Jonsson and J. J. P. van

Wyk for stimulating discussions during the IBP

Grassland and Tundra Modelling Workshop in August

1972 at Fort Collins, Colorado, from which this paper

developed. This paper reports on work supported in

part by U.S. National Science Foundation Grants

GB — 31862X and BMS73— 02027 A02.

UITTREKSEL

Elf IBP savanne- en grasveldstudiegebiede ( vanaf

vyf vastelande) se klimaat word beskryf en met be/ndp

van lioofkomponentanalise vergelyk. 'n Veelvoudige

lineere diskriminantfunksie word gebruik om die lengte

van die groei-seisoen vir elke gebied te bereken. Die

gemiddelde maandelikse reenval, gemiddelde maande-

likse temperatuur , lengtegraad en hoogte bo seespieel

word gebruik om hierdie funksie te bereken. Vir die

meeste van die gebiede was daar min verskil tussen die

werklike en die berekende aanvang en einde van die

groeiseisoen. Daar word tot die gevolgtrekking gekom

dat die groeiseisoen redelik betroubaar, wereldwyd

bereken kan word met behulp van 'n veelvoudige

diskriminantfunksie en 7z klein aantal abiotiese

gegewens.

REFERENCES

Cesar, J. & Menaut, J. C., 1974. Le peuplement v6g6tal. In

Analyse d'un ecosysteme tropical humide: la savane de

Lamia ( Cote-d'Ivoire ). Lamto: Bulletin de liaison des

chercheurs de Lamto, numero special 1974, fascicule II.

R. KIRK STEINHORST AND J. W. MORRIS

265

Dixon, W. J., 1971. Biomedical Computer Proprams. Berkeley:

Univ. of Calif. Press.

Johnson, S. C., 1967. Hierarchical clustering schemes. Psycho-

metrika 32: 241-254.

Ketner, P., 1972. Primary production of salt-marsh communities

on the island of Terschelling in the Netherlands. (Disserta-

tion) Nijmegen: Thoben Offset.

Lamotte, M., 1967. Recherches ecologiques dans la savane de

Lamto (Cote-d’Ivoire). Presentation du milieu et pro-

gramme de travail. La Terre et la Vie 21 : 197-215.

Lecordier, Ch., 1974. Le climat. In Analyse d'un ecosysteme

tropical humide: la savane de Lamto (Cote-d Ivoire).

Lamto: Bulletin de liaison des chercheurs de Lamto,

numero special 1974, fascicule I.

Newman, J. E. & Pickett, R. C., 1974. World climates and

food supply variations. Science 186: 877-881.

Pandeya, S. C., Mankad, N. R., Jain, H. K., Vyas, N. L., &

Wala, G., 1972. An approach to system analysis of a

grassland biome at Rajkot, India, during the growing

months. Biol. Land Plants 41 3-433.

Seal, H., 1964. Multivariate statistical analysis for biologists.

London: Methuen.

Singh, J. S., & Yadava, P. S., 1974. Seasonal variation in

composition, plant biomass and net primary productivity

of a tropical grassland at Kurukshetra, India. Ecol.

Monogr. 44: 351-376.

Skarveit, A., Ryden, B. E., & Karenlampi, L., 1975. Climate

and hydrology of some Fennoscandian tundra ecosystems.

In F. E. Wielgolaski (ed.), Fennoscandian tundra ecosy-

stems. Part /, Plants and micro-organisms. Berlin: Springer-

Verlag.

Smi'd, P., 1972. Fundamental climatological and hydrological

characteristics of grassland ecosystems in the Lanzhot

areas. In M. Rychnovska (ed.), Ecosystem study on Grass-

land Biome in Czechoslovakia. Brno: Czechosl. IBP/

PY-PP Report No. 2.

Bothalia 12,4: 267-292 (1976)

Automatic classification of the highveld grassland of Lichtenburg,

south-western Transvaal*

J. W. MORRIS!

ABSTRACT

A qualitative, semi-detailed plant ecological study of the area between 25° 54' and 26' 22' E and 26‘ 00'

and 26° 20' S, situated around the town of Lichtenburg in the south-western Transvaal, South Africa, is

reported. Mean annual temperature of the study area is 17 °C and annual rainfall is about 600 mm. A basic

difference is recognised between the Bankenveld Land System and the CT Grassland Land System. The former

is underlain by dolomite with lithosolic soils, Bankenveld vegetation and cattle ranching as the chief land-use,

whereas the latter is underlain by granite, Ventersdorp lavas, Dwyka tillite and surface limestone with Shorrocks,

Mangano and Lichtenburg series soils, Cymbopogon-Themeda Veld vegetation and extensive cultivation of maize

as the chief land-use. One hundred and ten 16 m2 quadrats were placed within each Land System by means of a

stratified-random strategy. Of the 247 species encountered, nearly 100 occurred in less than six quadrats.

Themeda triandra, Aristida congesta, Elionurus argenteus, Anthospermum rigidwn and Justicia anagaloides were

common throughout. Two association analyses were carried out and 15 final groups were interpreted out of a

total of 21 groups.

RESUME

CLASSIFICATION QUANTITATIVE DE LA PRAIRIE DE IIAUTES TERRES DE LICrlTEN-

B'JRG, SUD-OUEST DU TRANSVAAL

On expose les resultats d’une etude phyto-ecologique quantitative et semi-detaillee de !a region

comprise entre 25° 54' et 26° 22' E. et 26°00 et 26° 20' S., autour de la ville de Lichtenburg dans le sud-

ouest du Transvaal, Afrique du Sud. La moyenne annuel le de la temperature y est de 17 C et la hauteur

de pluie d' environ 600 mm. On recommit une difference fondamentale entre le Bankenveld Land System et

le CT Grassland Land System. Le premier est supporte par de la dolomite avec des lithosols, une vegeta-

tion du type Bankenveld et Televage du betail const ituant la principale utilisation du terrain : tandis qu'au

second sont sousjacents du granite, des laves de Ventersdorp, de la tillite de Dwyka et des calcaires de sur-

face avec des sols des series de Shorrocks, Mangano et Lichtenburg sur lesquels pousse une vegetation de

prairie d Cymbopogon — Themeda et dont la principale utilisation est la culture extensive du mats. Cent

et dix carres de 16 m de cote ont ete repartis sur chaque systeme selon la technique de stratification

aleatoire. Des 247 especes rencontrees, pres de 100 ont ete trouvees sur moins de six carres. Themeda

triandra, Aristida congesta, Elionurus argenteus, Anthospermum rigidum et Justicia anagaloides

etaient communes partout. Deux analyses d' association ont ete executees et d'un total de 21 groupes,

15 groupes finals ont ete interpretes.

CONTENTS

Page

Introduction 267

1. Description of study area 268

1.1 Location 268

1.2 Climate 268

1.3 Geomorphology and geology 270

1.4 Soils. 271

1.5 Vegetation 272

1.6 Land-use 273

1 .7 CT Grassland and Bankenveld Land

Systems 274

2. Procedures used in this study 274

2.1 Sampling strategy 274

2.2 Association analysis 274

2.3 Presence and indicator value 276

2.4 Re-allocation 276

3. Classification of the vegetation by

association analysis 276

3.1 Summary statistics of data 276

3.2 Total association analysis 276

3.3 Bankenveld association analysis. .. . 282

4. Discussion and conclusions 289

4.1 Sampling strategy 289

4.2 Methodological aspects 289

4.3 Vegetational aspects 290

Acknowledgements 290

Uittreksel 291

References 291

* Forms part of a Ph.D thesis submitted to the University

of Natal (Morris, 1973).

t Botanical Research Institute, Department of Agricultural

Technical Services, Private Bag X101, Pretoria.

INTRODUCTION

The overall model for the ecological survey of the

natural and semi-natural vegetation of the Highveld

Agricultural Region of South Africa was designed by

D. Edwards and J. C. Scheepers of the Botanical

Research Institute. At an early stage, they decided

that instead of spreading surveying efforts equally

over the whole Region of approximately 1 1 655 000

ha, the Key Area approach, whereby efforts would

initially be concentrated in the Lichtenburg, Maquas-

sie, Kroonstad, Bethlehem and Villiers Key Areas,

would be used. Each Key Area is a quarter degree

square in extent and covers approximately 69 000 ha.

Individual surveys of these five areas would be

followed by extrapolatory surveys linking the areas.

The aim of the survey reported here, a contribution

towards the Highveld Region survey, was an account

of the plant ecology of the Lichtenburg Key Area.

A report on the Kroonstad and Bethlehem Areas

(Scheepers, 1975) has already been produced.

The basic sampling strategy of 4 by 4 m quadrats,

with a stratified-random distribution, was chosen

during preliminary studies for the Kroonstad Area

and the Lichtenburg and Bethlehem surveys followed

suit to aid later comparison and extrapolation. The

technique of association analysis was chosen to

synthesize the results of each survey.

Two Land Systems (Dowling, 1968; Mabbutt,

1968) occur in the Lichtenburg Key Area. In this

communication they are referred to, for convenience

and brevity, as CT Grassland and Bankenveld Land

Systems. The former includes Dry Cymbopogon-

Themeda Veld and Sandy Cymbopogon-Themeda

Veld and the latter covers part of Bankenveld (Acocks,

1953). The main features of and differences between

the two Land Systems are detailed later. Briefly,

Bankenveld lies on an area of dolomitic lithosol

268

AUTOMATIC CLASSIFICATION OF THE HIGHVELD GRASSLAND OF LICHTENBURG, S.W. TRANSVAAL

north of Lichtenburg where a great deal of the

natural and semi-natural vegetation (physiognomi-

cally a grassland) remains and the main land-use

activity is cattle and sheep farming. CT Grassland is

an extensively-cultivated area south, east and west

of Lichtenburg where maize is the chief crop. Natural

and semi-natural vegetation is rare but, where still

encountered, the physiognomic structure is grassland

or savanna. In general, it is only small rock outcrops

and poorly-drained areas which are not ploughed in

the CT Grassland Land System.

1. DESCRIPTION OF STUDY AREA

1 . 1 Location

The location of the study area and its position

relative to towns in the South-western Transvaal,

South Africa are indicated in Fig. 1. Lichtenburg is

the only town situated in the area. Initially, the area

for study was the 2626AA quarter degree square but

as parts were extensively cultivated the area was

enlarged, as illustrated in Fig. 1, to enable placement

of enough sampling points in relatively undisturbed

vegetation. A rectangular shape was retained even

though much of the enlarged area was not sampled.

The quarter degree square covers about 69 000 ha

and the whole study area about 177 000 ha.

1 . 2 Climate

Apart from precipitation and air temperature

records, no climatic data are available from within

the study area. Data from Potchefstroom, and even

Pretoria, are used to complete the description of

climate given below. Where data from outside the

study area are used, they are intended as indications

of conditions prevailing within the area and no more.

1.2.1 Radiation and sunshine

Few meteorological stations in South Africa record

radiation. The station nearest to Lichtenburg is at

Pretoria, half a degree further north and 200 km east.

The instrument at Pretoria is a Kipp solarimeter and

all measurements are from a horizontal surface. The

percentage radiation received at the top of the atmos-

phere, but not reaching the surface of the earth at

Pretoria, averages 39 percent annually with a maxi-

mum attenuation of 45 percent in December and

minimum of 31 percent in June (Schulze, 1965).

Fig. 1. -Locality of study area. Dark shading indicates 2626AA quarter degree square and lighter shading the enlarged

study area. Inset: position in South Africa.

J. W. MORRIS

269

The monthly march of radiation at Pretoria is

given by Weather Bureau (1968). The highest mean

daily sums occur in summer from October to February

when the global radiation exceeds 530 cal/cm2/day

and the lowest mean daily sums in June and July

(less than 360 cal/cm2/day). Radiation increases

sharply from July to September and then stays fairly

constant during summer before starting to decrease

in February. The levelling off of radiation in summer

is due to an increase in cloudiness compensating for

radiation increases during this season. Diffuse radia-

tion varies from over 150 cal/cm2/day in summer to

less than 80 cal/cm2/day in winter. The highest sum of

radiation recorded on a single day (800 cal/cm2)

occurred during the month of January and the lowest

(27 cal/cm2) during July (Schulze, 1965).

1.2.2 Air temperature

Although air temperature data have been recorded

at Lichtenburg since 1905, observations until 1950

are the most recently published statistics (Weather

Bureau, 1954). During that period the mean of daily

maximum temperature varied from over 28 °C in

December and January to under 20 °C in June and

July. Mean of daily minimum was about 15 °C in

January and February and less than 2 °C in June and

July. Of more interest ecologically are the extreme

maxima and minima (Walter & Leith, 1960). Extreme

maximum temperatures in summer (September to

February) exceed 35 °C while in winter the maximum

recorded temperatures exceed 25 °C (July). Extreme

minima below -5°C occur from May to August. In

mid-summer (December to February) the extreme

minimum temperature is above 3 °C.

1.2.3 Grass minimum temperature and frost

Mean grass minimum temperature at Potchef-

stroom (Schulze, 1965) varies from 11 °C in December,

January and February to below -5 °C in June and

July. The maximum exceeds 16 °C from November

to March and drops to 8 °C in July while the minimum

temperatures throughout the year (except February)

are below 1 °C. The extreme minimum of -17 °C

was recorded during July.

In the absence of observations on frost occurrence,

climatologists have used various air temperatures to

indicate the occurrence of frost. If the criteria of the

publication Weather Bureau (1954), namely a mini-

mum temperature less than 0 °C in a Stevenson

screen 1,2 m above the surface, are accepted, Lichten-

burg averages 106 days per annum (30-year records)

with a possibility of experiencing frost. The average

first date is 19th May and the average last date is

2nd September. As the average number of days on

which frost actually occurs is 26, frost may be expected

one night in four from the middle of May until the

end of August. The extreme first date was 16th April

and the extreme last date on which frost has been

recorded was 26th September.

1.2.4 Soil temperature

Soil temperatures have been recorded at Potchef-

stroom for five years (Schulze, 1965). The highest

monthly mean temperatures have been recorded at

1400 hours in November, December, January and

February when the temperature exceeded 29 °C at

10 cm and 24 °C at 20 cm. The lowest temperatures

(0800 hours) in June and July at these depths, where

most plant roots are located, were 8,3 °C at 10 cm

and 10,2 °C at 20 cm.

1.2.5 Surface wind

No information on surface wind is available from

within the study area but the general features may be

determined from a knowledge of air circulation

patterns over South Africa and from observations of

wind speed and direction at nearby meteorological

stations. In general, there seems to be little seasonal

change in wind direction or force. Winds with a

northerly component predominate, but it is difficult

to assign any one prevailing direction according to

Schulze (1965). The northerly component results from

the normal anticyclonic circulation of air around a

high pressure cell located over the interior of South

Africa throughout the year.

TABLE 1.— Mean monthly rainfall in mm (r) and mean number of days with rain (d) at the seven stations with recording periods

of over 20 years from Weather Bureau (1965)

270

AUTOMATIC CLASSIFICATION OF THE HIGHVELD GRASSLAND OF LICHTENBURG, S.W. TRANSVAAL

1.2.6 Precipitation

Precipitation within and near the study area is

almost entirely in the form of rain, most of which

falls during summer. While severe hail storms are

experienced with a fairly high frequency, snow is

rare. Data from a number of meteorological stations

measuring rainfall within the study area are given in

Table 1. Mean annual rainfall at Lichtenburg is just

over 600 mm, with 85 percent falling during the six

summer months from October to March. Rain may be

expected to fall on ten days each month during Decem-

ber and January and on only one day per month in

winter (Table 1), with an average of about 60 days per

year.

1.2.7 Climate classification and summary

According to Schulze (1947), the Lichtenburg study

area is in the BSkw class of Koppen. The climate is

arid (steppe), cold and dry with a mean annual

temperature below 18 °C. Mean temperature of the

hottest month exceeds 18 °C and the dry season is

during winter. Following the Thornthwaite classifica-

tion (Schulze, 1947; 1958), the study area lies on the

border between DB’d to the west and CB’d to the

east. DB’d is a semi-arid warm (steppe) climate and

CB’d is subhumid warm. In both climates, moisture

is deficient throughout the year. Following the clas-

sification of UNESCO-FAO (1963) the climate is

accentuated temperate tropical. In other words, there

is a dry period of between one and eight months

coinciding with the period of shortest day-length, the

mean temperature of the coldest month is between

0 °C and 10 °C and the xerothermic index is between

150 and 200.

The climate of the study area is summarized in Fig.

2 (modified after Walter & Leith, 1960) and is

described by Schulze (1965), on which the following

description is based. The average daily maximum

temperature is about 28 °C in January and 18 °C

in July but in extreme cases temperatures may rise to

37 °C and 25 °C, respectively. Average daily minima

range from about 15 °C in January to 2 °C in July,

whereas extremes may drop to 6 °C and -10 °C,

Lichtenburg (1477m) 17° 602mm

Fig. 2. — Climate diagram for Lichtenburg (a = altitude,

b == mean annual temperature, c = mean annual

precipitation, d = mean daily minimum temperature

of coldest month, e = absolute minimum temperature,

f = mean daily maximum temperature of hottest

month, g = absolute maximum temperature, h = mean

range of temperature, i = curve of mean monthly

temperature, j = curve of mean monthly precipitation,

k = dry season, 1 = wet season, m = mean monthly

precipitation over 100 mm, n = months with absolute

daily minimum temperature below 0°C, p = mean dura-

tion of frost-free period in days).

respectively. The period during which frost is likely

to occur lasts, on average, for 106 days from May to

September, during which period frost occurs on about

26 days. Sunshine duration in summer is about 60

percent and in winter 80 percent of the possible.

Mean annual precipitation is about 600 mm.

Rainfall is almost exclusively due to showers and

thunderstorms. Winter months are normally dry and

about 85 percent of the annual precipitation falls

during the summer months from October to March. A

small rainfall peak occurs in January. Heavy falls of

125 mm to 150 mm are occasionally recorded in a

few hours. The average annual number of thunder-

storms is about 75. These storms are often violent with

severe lightning and strong, but short-lived, gusty,

south-westerly winds and are sometimes accompanied

by hail. The area has a high hail frequency with an

average of four to seven occurrences annually.

1 . 3 Geomorphology and geology

1.3.1 Relief

The study area lies between 1460 m and 1520 m

above sea level. It consists of a large, undulating

plain characterised by the absence of any marked

topographical features. Gentle rises and shallow

hollows throughout, and shallow valleys of the Harts

River and its tributaries in the south are the most

noteworthy topographical features.

Even in this very monotonous landscape a clear

relationship between topography and geology is

evident. The south-eastern corner is underlain by

by Archaean granite, which forms dome-shaped hills.

This slightly-raised portion is part of the divide

between the Harts River to the west and Schoon

Spruit in the east. The area occupied by Ventersdorp

lava, south and south-east of Lichtenburg, is usually

devoid of physiographic features. Volcanic brecchia

appears in dome-shaped outcrops and volcanic tuff

usually forms a featureless topography. The area

covered by dolomite, to the north of Lichtenburg, is

very flat, being relieved by occasional chert ridges,

shallow depressions, dry watercourses and, more

frequently, by sink-holes. Surface limestone, west of

Lichtenburg, builds extensive flat plains.

1.3.2 Geological strata

The geology of the area has been described by Von

Backstrom et al. (1953), on which the following

account is based.

Although Archaean granite occurs south-east of

Lichtenburg, outcrops are rare. It is usually overlain

by sandy surface drift.

Deposits of the Ventersdorp system are found in a

belt along the southern edge of the area from the

south-western corner to a point south of Lichtenburg

and then occur again along the eastern boundary of

the area. The system consists mainly of andesitic lava.

Intercalated are agglomerate and volcanic conglome-

rate, pyroclastic sediments or tuff and clastic sediments

including boulder conglomerate.

Rocks of the Dolomite series of the Transvaal

system are, for the most part, covered by more recent

deposits, particularly of gravel and surface limestone.

Small outcrops occur north and north-east of Lichten-

burg, on the boundary of the study area. The Dolomite

series consists mainly of blue-grey, massive, dolomitic

limestone, with intercalated lenses and layers of

chert and shale, the latter being developed more

particularly near the base of the Series. In many

places dolomite has been weathered chemically and

numerous sink-holes are found as a result.

J. W. MORRIS

271

Dwyka series tillite and shale of the Karoo

system occur over a fairly wide area south and east

of Lichtenburg. Owing to fast weathering and the

softness of the formation, rocks are seldom exposed

at the surface. Tillite is composed of a soft, clayey,

unstratified matrix in which unsorted fragments are

spread at random. The fragments are mostly of chert

and various quartzites. Examples of glaciated pave-

ments and other distinct signs of glaciation are found.

Economically, the Tertiary and recent deposits are

the most important in the area. They consist of gravel,

surface limestone, sand and alluvium.

Gravel is found mainly overlying Dolomite series

rocks in a belt along the northern edge of the area. The

deposits, which vary in depth from a few centimetres

to over 50 m, are made up of rounded alluvial material

with which is mixed angular, eluvial chert. The alluvial

material consists mainly of white, cream-coloured or

light grey pebbles of chert and chalcedony although a

variety of pebbles derived from other geological

systems are also encountered. The eluvial material

is composed of angular and irregularly-shaped

fragments of chert and concretionary grains of iron,

or manganese-iron, derived from weathering of the

dolomite floor. Dolomite series rock includes a

proportion of very hard, white or grey chert in

layers, lenses and veins. As the matrix is dissolved by

chemical action, chert remains behind as angular

rubble.

Surface limestone covers a large area west of

Lichtenburg. It is not generally exposed at the surface,

but is covered by a thin, sandy, overburden usually

from 20 cm to one metre thick. The thickness of the

limestone itself is difficult to determine but it is

known to occur as a thin, hard, crust although it is

usually much thicker. Depths between 10 m and 38 m

have been recorded. The quality of the limestone

varies from some of great purity to some that can

be described as calcrete. At the surface, the limestone is

usually hard and massive while at greater depths it

is softer, granular, friable and slightly stratified.

In the area north-west of Lichtenburg the older

formations are overlain by red and yellowish Kala-

hari sand, consisting in the main of slightly rounded

grains of quartz, less than one mm in diameter.

South-east of Lichtenburg, coarse-grained sand covers

a fairly extensive area. It consists of subangular

grains of quartz and felspar with flakes of mica and

is mainly derived from underlying Archaean granite.

1.3.3 Erosion surfaces

Erosion surfaces of the Highveld Agricultural

Region have been mapped by Harmse (1967). Dolo-

mite, and gravels overlying dolomite, north of Lichten-

burg are of the Karst structural phase of the African

surface. Sand overlying the rest of the study area is

of the aggradational phase (aeolian sand) of the

African surface. Sand was transported from the

north-west.

1.3.4 Economic geology

As the surface mining of diamonds and limestone

within the study area influence the natural vegetation,

these activities are described briefly.

The first diamond in the Lichtenburg district was

discovered during 1921, although diamonds were

being mined in the South-western Transvaal before

1914 (Wagner, 1914; Du Toit, 1951; Draper, 1928;

Williams, 1930). As diamond discoveries increased at

Lichtenburg, more and more fortune hunters arrived

and many spectacular ‘rushes’ were held, culminating

in the famous ‘charge’ on the farm Grasfontein in

March, 1927 when some 25 000 runners participated.

In August, 1926, an estimated 56 000 non-Europeans

were employed. According to Williams (1930), a

population of over 100 000 was resident on the new

diamond fields shortly after their discovery. The

presence of this large population, concentrated in a

small area north of Lichtenburg, must have had a

marked effect on the natural vegetation. In addition

to direct disturbance by trampling, digging and

clearing, the keeping of livestock, gathering of fire-

wood and cultivation of crops probably caused rapid

deterioration of vegetation near the diggings. Very

little activity is evident on the diggings now and only

an occasional miner is encountered.

A large proportion of the area underlain by surface

limestone has been purchased by two cement com-

panies with factories in the area. The thin, sandy

overburden is bulldozed away and surface limestone

is mined by open-cast methods.

1 . 4 Soils

Soils of the 2626AA quarter degree square have

been described and mapped by Van der Bank (1968).

His map of 17 soil series, four soil complexes and

three land classes has been simplified in collaboration

with E. Verster (pers. comm.) to eight soil series to

conform with present soil nomenclatural concepts of

the Soils and irrigation Research Institute, Pretoria.

The soil survey by Van der Bank (1968) clearly

illustrates from one side the differences in emphasis

placed by pedologists and plant ecologists in the

same area. He describes, in detail, soils which are

under heavy cultivation while the Dolomite lithosol

receives scant mention. To the botanist, on the other

hand, the natural vegetation of the ploughed area is

solely of historical interest while the vegetation of the

unploughed areas (the lithosols) has great economic

and academic value. These differences in emphasis

naturally hinder close correlation between results of

these and other soil and vegetation surveys of the

same area.

1.4.1 Shorrocks, Mangano and Lichtenburg Series

Shorrocks series soils occur in small patches north-

west of Lichtenburg. Mangano series soils occur very

locally within the Dolomite region north of Lichten-

burg, while Lichtenburg series soils cover relatively

small areas near the Harts River in the south-eastern

corner of the study area.

Shorrocks, Mangano and Lichtenburg Series are

well-drained soils. They are characterised by good

internal drainage and weak structure. They have

high, fine sand and low clay content, prominent red

colours and an absence of mottling. Horizons that

impede drainage are absent. Kaolinite is the principal

clay mineral and the exchange capacity varies between

two and eight milli-equivalents percent. Calcium and

magnesium are the dominant exchange cations

throughout the profile.

1.4.2 Soetmelk Series

Soetmelk is the non-lithosol series covering the

largest area in the quarter degree study area. It

occurs throughout the area not underlain by dolomite.

Restricted drainage is a characteristic of Soetmelk

soils. With increasing depth the clay content increases

and the structure becomes weak medium blocky.

Red and yellow mottles associated with uncsmented

iron concretions are distinctive features of the lower

horizons. Principal clay minerals are kaolinite and

illite. Montmorillonite tends to increase in the lower

272

AUTOMATIC CLASSIFICATION OF THE HIGHVELD GRASSLAND OF LICHTENBURG, S.W. TRANSVAAL

horizons. Exchange capacities vary between four and

eight milli-equivalents percent. Calcium and magne-

sium are the dominant cations.

1.4.3 Rensburg, Bonheim and Lindley Series

The Rensburg, Bonheim and Lindley series are

poorly-drained soils. Poor permeability, high clay

content, high exchange capacity and strong blocky or

prismatic structure are the salient features of these

Series.

The occurrence of these soils, either collectively

or singly, is limited in extent. Rensburg series occurs

in the bed of the Harts River while Bonheim series

soils are restricted to slopes of the Harts River valley,

where the river leaves the dolomite substrate. Lindley

series occurs in small patches north and north-west of

Lichtenburg. It is usually underlain by dolomite.

1.4.4 Kalkbank Series

The feature common to Kalkbank sandy type and

Kalkbank loamy type soils is the underlying surface

limestone formation encountered at shallow depths.

Where the limestone is buried deeper the soil is

usually Soetmelk series. The parent materials are

either aeolian sand, drift materials, or both.

These soils cover areas of moderate size, usually in

complexes, to the west of Lichtenburg.

1 . 5 Vegetation

The vegetation of the study area is described by

Acocks (1953) in his survey of the veld types of

South Africa (Fig. 3).

In the South-western Transvaal, Dry Cymbopogon-

Themeda Veld (Veld Type 50) occurs in an arc from

Lichtenburg to south-west of Delareyville and then

south to Wolmaransstad. This arc, comprising the

northern variation of the Veld Type, is found on

sandy soils. The dominant species is Themeda triandra.

Cymbopogon plurinodis is the tallest grass, but is

usually not common. This Veld Type lies between

altitudes of 1 280 m and 1 370 m on flat, sandy

country with a summer rainfall of 430 to 580 mm per

annum and with frosty winters. Species of general

occurrence are given in Table 2.

Species of less general occurrence include : Cynodon

dactylon, Digitaria argyrograpta, D. eriantha, Panicum

coloration and Stipagrostis uniplumis.

Sandy Cymbopogon-Themeda Veld (48) occurs in

the rough square bounded by Lichtenburg, Venters-

dorp, Klerksdorp and Ottosdal, to the immediate

east of the arc of Dry Cymbopogon-Themeda Veld.

The square comprises the northern variation of the

Veld Type where the altitude ranges from 1 310 m to

25°

26°

2 7‘

J. W. MORRIS

273

TABLE 2. Relative abundance values (in thousands) of those species of general occurrence, according to Acocks 0975) with

relafye abundances of over 1 000 in Dry Cymbopogon-Themeda Veld (50), Sandy Cymbopogon-Themeda Veld (48) and Baker,-

vela (oi )

1 520 m and summer rainfall is from 510 to 690 mm.

Winters are frosty. Species of general occurrence are

listed in Table 2. Species of less general occurrence

include Cynodon dactylon, Cynodon incompletus,

Digitaria argyrograpta and Helichrysum rugulosum.

According to Acocks, this Veld Type merges into the

western variation of Bankenveld and needs more

study. Dry Cymbopogon-Themeda Veld usually occurs

at a slightly lower elevation than Sandy Cymbopogon-

Themeda Veld and usually receives slightly less rain.

Sandy Cymbopogon-Themeda Veld and Dry Cym-

bopogon-Themeda Veld are both Pure Grassveld

Types while Bankenveld (61) is a False Grassveld

type. Bankenveld occurs in a belt from east of Venters-

dorp, to Klerksdorp and Parys. The Western varia-

tion occurs near Lichtenburg. It is found on sandy

plains and low, rocky ridges, ranging in altitude from

1 370 m to 1 680 m and receives 560 to 690 mm rain,

mostly in summer. It is a rather sparse, sour, strongly

tufted vegetation and, in the nature of its grasses, is

clearly transitional, according to Acocks, from Sandy

Cymbopogon-Themeda Veld to Sour Bushveld (20),

which occurs north of the study area. The climax was

possibly an open savanna with Acacia caffra. Species

of general occurrence are given in Table 2. Species

of less general occurrence include Anthospermum

rigidum, Digitaria eriantha, D. monodactyla, Eus-

tachys mutica, Kohautia amatymbica and Pygmaeo-

thamnus zeyheri.

Tentative boundaries to Veld Types drawn by

Acocks (1953) for the South-western Transvaal may

be re-drawn after more intensive study. Limits ot

Bankenveld near Potchefstroom and Ventersdorp,

east of the study area, have been accurately mapped

by Grunow (1959). Photographs from NASA/

ERTS-1 flights will greatly aid the further mapping

of vegetation in this area at reconnaissance scales.

Bankenveld is actually found as far south as Lichten-

burg, as far west as a line from Lichtenburg to Mafe-

king and as far east as a line from Lichtenburg to

Rustenburg, although this is not apparent from

Acocks’s map.

1 . 6 Land-use

Broad land-use categories are closely tied to

geology, soils and vegetation in the study area. The

parts with deep, well-drained soils overlying all

geological systems except dolomite (Dry Cymbopogon-

Themeda- and Sandy Cymbopogon-Themeda Veld

Types) are extensively ploughed. Van der Bank

(1968) comments that, in practice, all soils that are

ploughable have already been ploughed. The principal

crop is maize and limited quantities of sunflowers,

grain sorghum, groundnuts and cattle fodder are

grown. It is estimated by Van der Bank (1968) that

the average yield of the maize crop is about 1 170

kg/ha although some farmers reap 1 550 or more

kg/ha.

Cattle ranching is the chief occupation of farmers

on the dolomitic lithosols (Bankenveld Veld Type)

north of Lichtenburg. Occasional sink-holes in the

area, which have become filled with aeolian sand, are

invariably cultivated for cash crops like maize. Some

irrigation, using the good supplies of water from

boreholes, is undertaken.

An alarming activity is the open-cast mining of

surface limestone in the area by two cement companies.

Relatively large tracts of land have been acquired and

274

AUTOMATIC CLASSIFICATION OF THE HIGHVELD GRASSLAND OF LICHTENBURG, S.W. TRANSVAAL

are being systematically laid desolate as no attempt

appears to have been made to stabilise and re-

vegetate stripped areas.

1.7 CT Grassland and Bank env eld Land Systems

According to Dowling (1968) and Mabbutt (1968),

the land of a region is conceived as a series of Land

Systems, each of which contains typical land facets,

which have similar features wherever they occur and

which can be readily identified on aerial photographs.

When relief, geology, geomorphology, soils, vegeta-

tion and land-use are considered, two Land Systems

may be recognised in the study area. These two

Land Systems were treated separately in the sampling

and analysis of vegetation reported in this account.

One Land System, named CT Grassland, is under-

lain by granite, Ventersdorp lavas, Dwyka tillite and

surface limestone on the aggradational phase (aeolian

sand) of the African erosion surface. Soils of Shor-

rocks, Mangano and Lichtenburg series are generally

sandy and well-drained. The natural vegetation was

Dry Cymbopogon-Themeda Veld or Sandy Cymbopo-

gon-Themeda Veld but most of it has been destroyed

for the cultivation of maize, the chief land-use.

The second Land System, named Bankenveld, is

underlain by dolomite, variously covered by alluvial

gravels on the Karst structural phase of the African

erosion surface, and has lithosolic soils. The Veld

Type is Bankenveld and cattle ranching is the main

land-use.

2. PROCEDURES USED IN THIS STUDY

2 . 1 Sampling strategy

The most important consideration in designing the

vegetation sampling strategy was ensurance of

compatibility with other component projects of the

ecological survey of the Highveld Agricultural Region.

A common sampling strategy will enhance possibili-

ties for integration with and extrapolation between

individual surveys, including some by other officers.

As J. C. Scheepers of the Botanical Research Insti-

tute, Pretoria, who is responsible for surveying the

eastern half of the Region, had started field sampling

before the study reported here was initiated, his

general strategy, as outlined below was followed.

2.1.1 Sample placement and size

Initially, the 2626AA quarter degree square (Fig. 1)

was taken as the study area. All lands under culti-

vation, or showing signs of past cultivation, as well

as the town of Lichtenburg and the bed of the Harts

River were excluded from the area to be sampled.

The remaining area was stratified into Bankenveld

and CT Grassland Land Systems (see 1.7) and phy-

siognomic-physiographic units were delimited on

1 :36 000-scale aerial photographs. Morris (1973)

details the stratified-random sampling strategy that

was used to mark 220 sample positions within the

study area and gives reasons for the study area

having to be enlarged to that illustrated in Fig. 1.

Briefly, insufficient suitable sampling sites were

found within the quarter degree square.

The sample size of 4x4 m, chosen during pilot

studies for the Highveld Survey Project, was retained

for uniformity. Pilot studies had shown that a sample

area of 16 m2 was both the smallest adequate sample

for grassland vegetation of this kind and the largest

area that could be sampled in an economically-

justifiable time (J. C. Scheepers, pers. comm.). Species

present in a belt, approximately two metres wide,

around the perimeter of the sample were also recorded

to aid re-allocation of samples mis-classified by asso-

ciation analysis owing to the chance absence of a

positive dividing species. Details of quadrat location

in the field are given in Morris (1973).

2.1.2 Habitat data

A limited amount of habitat information was

gathered. Physical factors recorded for each site

included geology, geomorphology, aspect, angle of

slope and exposure. Soil series (Van der Bank, 1968)

and depth were recorded and then for each horizon,

soil pH, soil reaction to dilute hydrochloric acid

(HC1), moist soil colour (Munsell) and soil texture

were noted. Biotic influence was noted by animal type

and by degree on a four-point scale from absent to

very intense.

2.1.3 Vegetation data

Total basal cover and height and basal cover of

each constituent stratum were estimated for the 16 m2

sample and the presence of all permanently-recognis-

able plant species was recorded. Cover-abundance on

the Braun-Blanquet scale (Werger, 1973) was also

noted for each species.

2 . 2 Association analysis

At the time of sampling, the only objective method

for classifying vegetation known to be suitable was

the hierarchical technique of the Southampton-

Canberra school (Williams & Lance, 1958; Williams

& Lambert, 1959, 1960, 1961; Lambert & Williams,

1962). The original technique, association analysis,