NAVORolNGSir>JS'^ITi
“ ' ‘^i-ANTKUN’DE
SfBL.'OrEEK.
1 S. t 1S57
CATALOGUE NO.
L B R A R Y
botanical REEEA--H l-iSTf
Digitized by the Internet Archive
in 2016 with funding from
South African National Biodiversity Institute Libraries
https://archive.org/details/bothaliavolume808unse
JBotbaUa
A RECORD OF
CONTRIBUTIONS
FROM THE
NATIONAL HERBARIUM
REPUBLIC OF SOUTH AFRICA
PRETORIA
VOL. VIII
EDITED BY
L. E. CODD, D.Sc.
CHIEF, BOTANICAL RESEARCH INSTITUTE, DEPARTMENT OF AGRICULTURAL TECHNICAL
SERVICES AND DIRECTOR OF THE BOTANICAL SURVEY OF THE REPUBLIC OF SOUTH AFRICA.
THE GOVERNMENT PRINTER, PRETORIA
G.P.-S.0I40996— 1965-66— 1,000.
CONTENTS TO VOL. YIII.
Part 1. Published May, 1962 Page
1. The Cucurbitaceae of Southern Africa.
A. D. J. Meeuse 1
2. Notes and New Records of African
Plants. Various Authors 113
Part 2. Published December, 1964.
1. The South African Species of Dipcadi.
A. A. Obermeyer 117
2. The South African Species of Laga-
rosiphon. A. A. Obermeyer 139
3. The South African Species of Antheri-
cum, Chlorophytum and Trachyandra.
Addenda et Corrigenda. A. A. Ober-
meyer 147
4. The South African Species of Ortho-
siphon. L. E. Codd 149
5. Notes and New Records of African
Plants. Various Authors 163
6. Species of Plocamium on the South
African Coast. R. H. Simons 183
7. Notes on the Species of Zonaria in
South Africa. R. H. Simons 195
Part 3. Published May, 1965.
1. The South African Stipeae and Aris-
tideae (Gramineae). B. de Winter. . . 201
Part 4. Published December, 1965.
1. The Cycads of Southern Africa. R. A.
Dyer 405
Supplement No. 1. Published May, 1966. Page
1. Botanical Research Institute 1.7.64
to 30.6.65:—
Stair SI
Systematic Botany Section S5
Economic Botany SIO
Botanic Garden Sll
Botanical Survey Section SI 5
Visitors S18
Overseas Visits, Conferences, etc... SI 8
Overseas Study Tour. D. Edwards S19
Collecting Expeditions, Excursions,
etc S22
A Cycloptychis Hunt. H. C. Taylor S25
Notes on Some Kniphofias in the
Eastern Orange Free State. L. E.
Codd S27
Publications S31
Personalia S32
2. The Flowers on the New Decimal
Coins. Cythna Letty S35
3. Field Work is Not all Fun. R. A.
Dyer S43
4. Plants which Cause “ Gousiekte ”
Poisoning. L. E. Codd and S. Vooren-
dyk S47
5. Contemporary Botanists and Botanical
Collectors:
1. Dr. F. Z. van der Merwe. I. C.
Verdoorn S59
2. Mr. L. C. Leach. L. E. Codd.. S61
6. Note on Vegetative Propagation of
Proteaceae and Ericaceae. J. Admiraal S65
The Cucurbitaceae of Southern Africa
By
A. D. J. Meeuse
A few years ago the National Chemical Laboratory, Pretoria, in co-operation with
the Division of Horticulture, started an investigation into the nature of the bitter
principles contained in the fruits and other parts of many species of the Cucurbitaceae
and the genetical implications(^) in connection with the incidence and occurrence of
these substances in certain horticultural varieties and their related wild forms. Later
biochemical studies of the bitter substances and the enzymes associated with them
were undertaken. (^)
The Division of Botany was requested to co-operate by assisting with the purely
botanical (taxonomic) side of the study so as to relate a certain bitter substance with
the correct scientific name of the plant it was extracted from.('fi It was decided not
to restrict the taxonomic work to the species or genera that were known to produce
bitter fruits but to revise the whole family as represented in Southern Africa. This
was justified not only because, for reasons explained below, the family was badly in
need of revision, but also because bitter principles appear to be of a much more common
occurrence in the Cucurbitaceae than was originally expected, which widened the scope
of the biochemical and genetical studies considerably, and consequently necessitated
the proper identification of more species.
Previous Taxonomic Studies of the Cucurbitaceae of Southern Africa
If one disregards the few species described by Linnaeus, Berg, Lamarck and some
other early authors, the first important contributions are found in Thunberg’s publi-
cations. Thunberg described 10 species from the Cape Province and, including the
previously described ones, recognised about 15 species from this area. Although
more recent studies reduced most of Thunberg’s names to synonyms, his specific names
were adopted by Seringe in DC., Prodr. 3, published in 1828, who added very little to
our knowledge. Schrader and Naudin later described a few more species but it was
not until 1862 when Sonder treated the family in Vol. 2 of the Flora Capensis that
a more comprehensive survey of all species known up to that time was given. Sonder
described several new species.
Similarly, the Cucurbitaceae of Tropical Africa were treated by J. D. Hooker
in Flora of Tropical Africa, Vol. 2 (1871). Some of the species included in this work
have later been recorded from Southern Africa, notably in South West Africa, Bechu-
analand and the northern Transvaal.
In 1881 A. Cogniaux monographed the whole faimly in A. and C. de Candolle’s
“ Monographiae Phanerogamarum ” and this author remained for many years the
recognised expert. In 1916 his treatment of a part of the family (Fevilleae and
(b Enslin, Joubert and Rehm, J. S. Afr. Chem. Inst. 7: 131-138 (1954).
(^) Enslin, Joubert and Rehm, J. Sci. Food Agric. 7: 646-655 (1956).
F) Rehm, Enslin, Meeuse and Wessels, J. Sci. Food Agric. 8; 679-686 (1957).
2
Melothrieae) was published as Vol. IV, 275, 1, of “ Das Pflanzenreich ” and his manuscript
of another part (Cucurbitaceae — Cucurbiteae — Cucumerinae), after having been edited
and revised by Harms, was posthumously published in 1924 as Vol. IV, 275, II, of the
same work. These last two monographs include all but one of the genera occurring
in the area under discussion.
Present Position
Although comparatively recent monographs were available, preliminary studies
indicated that a new revision was desirable. Not only are the two treatments in “ Das
Pflanzenreich ” (as in most volumes of this work) based on a study of a very limited
number of herbarium specimens of each species recognised therein, so that the specific
limits are often drawn too narrow, but also the nomenclature needed looking into. In
addition it was found that Cogniaux described well-known older species, including
some of his own, again under a different name and sometimes in a different genus.
Although this can be explained in some cases as a result of incomplete specimens or
extreme variability, others among his type specimens are quite adequate and can without
difficulty be identified as belonging to some older and often well-known species. One
naturally feels a great deal of diffidence when criticizing a recognised authority, but after
careful consideration and a study of many types or isotypes, most of which had been
annotated by Cogniaux himself, it is felt that in several genera a drastic reduction
in the number of species is indicated and that a few genera recognised by Cogniaux
cannot be maintained.
There is another reason why the number of species in certain genera treated by
Cogniaux in his monographs must be reduced. As in some other families with a
similar type of vegetative growth and occurring in similar habitats, notably in Con-
volvulaceae, the leaf-shape of many representatives of the Cucurbitaceae varies
enormously, from undissected, faintly lobed or angled to deeply dissected. This
condition is often encountered in one individual, e.g., the leaves of the young plants,
i.e., also the lowermost leaves, are undivided, whereas those formed later become
more dissected as the plant grows older, or in other cases the leaf-shape depends on
the conditions of growth: the leaves on creeping stems sometimes have entire or slightly
dissected leaves whereas the uppermost parts of the climbing stems of the same individual
can be deeply dissected. Specimens taken from different portions of the same individual
or from an old and a young plant may, therefore differ considerably in appearance
and also a creeping specimen need not at once be associated with a specimen taken from
climbing plants of the same species. In doubtful cases cultivation of plants from seed and
regular observation of specimens growing under different conditions provide the final
proof. Many of the species mentioned below were grown from seed at the Division of
Botany, Pretoria, by the present author or on the Experimental Farm of the Division of
Horticulture at Roodeplaat near Pretoria by Dr. S. Rehm and observations on the
living plants complemented the herbarium and field studies.
Apart from the material of the National Herbarium, Pretoria, the specimens of
all the important South African herbaria, the Government Herbarium at Salisbury,
Southern Rhodesia, and the herbaria of Brussels, Coimbra and Leiden were studied.
An important collection containing many types was received on loan from Zurich and
a small set of the more recent Dinter collections was borrowed from Berlin-Dahlem.
Thanks are due to the Directors of all these herbaria for their kind co-operation.
Where quoted in the text the herbaria are indicated by the standard abbreviations of
the Index Herbariarum.
Special thanks are due to Mr. D. J. B. Killick, our Liaison Officer then at Kew,
who compared several types at Kew and the British Museum (Nat. Hist.), and was
most helpful in providing abstracts of the literature not available in Pretoria.
3
Economic Importance.
Apart from the above-mentioned toxic principles found in many genera of the
Cucurbitaceae, vi'hich have in the past been extensively used in medicine and are still
important in native medicine in many parts of the world and which have, at least in
Southern Africa, caused occasional poisoning of cattle and humans, this family has
provided some important economic plants. Forms producing edible fruits which are
extensively grown throughout the world as vegetables or table fruits (Cucumis: cucumber,
gurkins and.inelon; CitruUus: water-melon; Lagenaria: calabash gourds; Cucurbita:
rr~plimpkms7~^u^shes or/gantaloube^ vegetable marrow, etc.; Sechium, Luff a, Tricho-
san thes, Momordicd) .
Native tribes all over the world also eat fruits or sometimes the tuberous roots
of wild species and in some cases, especially in Southern Africa, these products are
a staple diet in certain seasons, such as the dried fruit pulp of Acanthosicyos horrida
for the Hottentot tribes in the Namib area of South West Africa, or they form the main
(and in the dry season sometimes the only) source of water for the Bushmen in the
Kalahari, for example the wild form of CitruUus lanatus or “ Tsamma ” melon. Apart
from these, the fruits and/or roots of species of Cucumis, Coccinia and Coral/ocarpus
and the fruits of CitruUus naudinianus are part of the diet of many tribes. Recent
ethnological explorations among the Bushmen in the Kalahari included studies of their
diet and it goes without saying that a proper identification of the food plants is essential
to prevent confusion. Cucumis hookeri for instance is eaten by Hottentots and Bushmen,
but up to now nearly all herbarium specimens of this species were named “ Cucumis
africanus ”, “ C. dissectifolius ” or “ C. myriocarpus ”. It is now apparent that Cucumis
zeyheri and C. myriocarpus (— C. dissectifolius) do not occur in South West Africa
or most of the Kalahari proper and these two species are invariably very bitter and
extremely poisonous. An identification of the edible form of C. hookeri as C. zeyheri
or C. myriocarpus would of course lead to most contradictory and confusing reports.
Finally, some species occur as weeds on cultivated lands, viz. (in the order of their
frequency and importance), Cucumis myriocarpus, C. melo (wild form), CitruUus
lanatus and Cucumis zeyheri, but as weeds they are not very troublesome and are
easily controlled by cultivation (hoeing, etc).
CUCURBITACEAE
Annual or perennial, mostly herbaceous climbing or creeping plants, occasionally
woody, rarely erect or arborescent, almost invariably with tendrils and often scabrid
on stems and /or leaves. Rootstock sometimes large, tuberous. Leaves alternate,
simple and undivided to deeply palmately or more rarely pinnately dissected with
entire or dissected segments, or sometimes palmately or pedately compound, usually
petiolate, often cordate at the base, varying from small in some genera to large in others,
very rarely reduced or wanting; petiole sometimes with a stipuliform bract in its
axil or biglandular at its apex. Tendrils laterally at the same nodes as the leaves, simple,
bifid or occasionally multifid, rarely reduced to straight spine-like organs or wanting.
Flowers unisexual, monoecious or dioecious (in monoecious species often protandric),
borne in the leaf-axils and, in monoecious species, male and female flowers in different
axils (often the male ones in the upper axils, the female ones in the lower axils) or
more rarely male and female flowers produced in the same axil; male flowers solitary
or in fascicles, umbels, racemes or cymes; female shallow to cup-shaped or sometimes
flowers usually solitary. Male flowers: Calyx and corolla united below to form a
tubular receptacle; calyx-lobes 5, usually free; corolla-lobes free or united at the
base, white, cream-coloured, greenish or yellow, rarely orange or red, often small and
inconspicuous, but in some genera rather large but rarely showy; stamens often 3,
rarely 5 or 2, occasionally 4 plus one staminode, inserted on the receptacle; filaments
usually short, free or united; anthers free or cohering in a head, 1- or 2-thecous
4
(often 1- and 2-thecous ones occur in one flower); connective often produced above
the thecae, straight, curved, sigmoid or condupUcate; rudiment of ovary often present.
Female flowers'. Calyx and corolla as in the male flowers but occasionally different
in size; staminodes usually present; ovary inferior, more rarely free at the apex,
connate with the receptacle to form a globose to fusiform or tubular structure which is
often separated from the rest of the flower by a constriction ; ovary-chambers 3, more
rarely 2 or 1 or (by spurious septa) 4-6 ; ovules usually several to many, rarely only
1-2; style terminal, simple or divided at the apex, sometimes surrounded at the base
by a disc or cup-like structure; stigma various but usually with 2 or 3 fleshy lobes.
Fruit a usually indehiscent, more rarely dehiscent, soft or hardshelled berry (pepo),
which is fleshy inside, or occasionally fruit dry, leathery or corky. Seeds often
numerous, often compressed, smooth or variously pitted, occasionally hairy, often
margined, rarely with membranous wings; testa leathery, crustaceous or bony,
tegmen membranous or hyaline; cotyledons conforming to the shape of the seed,
thick or flattened; radicle usually short, conical; endosperm very scanty or 0.
Mainly circumtropical in distribution, rare in boreal climates. Of the about 100
genera, 16 are represented in Southern Africa.
The generic limits in this family, as first defined by D. J. Hooker in Benth. & Hook.,
Gen. PI. 1 (1867) and amended by Cogniaux in his monograph 1881 have found general
recognition and are without any essential changes used in botanical works, such as
regional or local floras, all over the world. It is not to be expected that any drastic
changes in the generic taxonomy of this family will take place, at least not in the near
future. One may differ in minor points such as the delimitations of two or three closely
related taxa, and in the present paper the only important changes in the circumscription
of the genera are the reductions of Hymenosicyos to Oreosyce, of Pisosperma and
Toxanthera to Kedrostis, of Raphanistrocarpus to Raphanocarpus and of Sphaerosicyos
to Lagenaria. The reasons are given under the genera in question.
The definition of the various genera is based on characters taken from both the male
and the female flowers and occasionally, in addition, from the fruits and seeds, more
rarely from vegetative characters. The use of characters of both male and female
flowers in one key (as is usually done) often makes it difficult to establish the genus
if the specimen is dioecious or, if monoecious, is protandric and bears flowers of only
the one sex. Fruiting specimens without flowers are even more difficult to place, as
in many genera the fruit is of the same type.
It is tried here to avoid this difficulty by using a key with subdivisions, in each
of which the vegetative characters, the morphology of the male flowers and of the
female flowers and the structure of the fruits and seeds are exclusively used in this
sequence, so that if special parts are lacking an alternative part of the key can be used.
In most cases types, isotypes or otherwise authenticated specimens of each species
were available for study. The type specimens originally deposited in the Berhn
Herbarium, such as several of the “ older ” Dinter specimens, were destroyed by fire,
but fortunately in most cases isotypes were located in other herbaria.
Key to Genera
A rigid leafless shrub armed with spines; tendrils 0 10. Acanthosicyos
Plant with well-developed leaves; tendrils present, very rarely reduced to spines or
absent (occasionally flowers appearing on the bare stems before the leaves
but, if so, stems without spines and tendrils well developed):
Petioles biglandiilar at the apex (rarely glands absent in all leaves of the
specimens) 13. Lagenaria
5
Petioles never glandular at the apex (sometimes sessile glands present on the
leaf-blades) :
Petioles (at least the majority) with a conspicuous, sessile, stipuliform, toothed
or fimbriate leafy bract at the base:
Flowers monoecious, minute; fruit subglobose, about the size of a cherry or
slightly smaller; seeds few, compressed, marginate:
Male and female flowers borne in different axils; style surrounded by a
disc ; plants usually drying a light greyish green (S.W. Africa, Angola)
2. Dactyliandra
Male and female flowers borne in the same axil; style without a disc;
plant usually drying dark (Bechuanaland, Transvaal, tropical Africa)
3. Blastania
Flowers dioecious, not very small, with tubular receptacle; fruit usually
ovoid and apiculate or at least somewhat pointed; seeds but slightly
compressed, emarginate 14. Trochomeria
Petiole without conspicuous stipuliform bract at the base:
Flowers present A
Only fruits present AA
A
Receptacle with 3 incurved scales at the base of the petals (and often with a
few additional smaller ones):
Flowers monoecious ; ovary smooth, ovules few (often 2) ; fruit dry, narrowly
fusiform, few-seeded; leaves simple 8. Raphanocarpus
Flowers monoecious or dioecious; ovary often muricate or with soft thick
spines; fruit fleshy, not narrowly fusiform, many-seeded. ... 9. Momordica
Receptacle without incurved scales at the base (though sometimes in female
flowers with minute staminodes):
Male flowers present (with or without female flowers) B
Only female flowers present BB
B
Anthers straight or more or less curved but not conduplicate or sigmoid:
Stamens 4, with a rudimentary fifth stamen, anthers small, usually hori-
zontal 1. Gerrardanthus
Stamens 3-5, but if 5, all stamens fertile; anthers not horizontal:
Rudiment of pistil in male flowers none or small, gland-like:
Flowers monoecious or dioecious; fruit not circumscissile at the
base 6. Kedrostis
Flowers always monoecious, the male and female ones usually borne
in the same axils ; fruit circumscissile at the base .... 7. Corallocarpus
Rudiment of pistil in male flowers evident:
Receptacle rather wide, campanulate to shortly and broadly tubular;
anthers usually not linear, basi-fixed, almost invariably with distinct
filaments ; connective rather broad ; flowers monoecious or dioecious ;
ovary in female flowers not setose 4. Melothria
Receptacle rather narrowly subcylindric; anthers linear, dorsifixed in
the middle on very short filaments; connective narrow; flowers
monoecious; ovary in female flowers setose 5. Oreosyce
Anthers conduplicate or sigmoid:
Receptacle short, rotate, campanulate, cup-shaped or shortly tubular, not
much longer than wide:
Filaments connected at least at their apices to form a distinct tube; anthers
cohering; plants dioecious, flowers rather large 16. Coccinia
Filaments free; anthers free or sometimes cohering:
Rudiment of pistil in male flower absent or minute; connective not
produced above the anthers; tendrils usually bifid to multifid:
6
Flowers racemose, rarely solitary, large, white, green-veined; tall
perennial climber or occasionally prostrate {L. mascarena) 13. Lagenaria
Flowers solitary or fascicled, not racemose, yellow; prostrate plants
11. Citrullus
Rudiment of pistil in male flower evident; tendrils simple:
Plants monoecious or dioecious; connective produced beyond the
anthers; ovary and, as a rule, also the fruit often muricate or
covered with soft spines or bristles, or densely hairy. ... 12. Cucumis
Plants dioecious; connective not produced beyond the anthers but with
papillae or glands at the apex {T. sagittata) 14. Trochomeria
Receptacle elongate, tubular, funnel-shaped or cylindric, distinctly longer
than wide:
Corolla-lobes usually narrow, tapering into an acute apex; rudimentary
pistil evident; connective at its apex with glands or papillae; dioecious
perennials with tuberous root 14. Trochomeria
Corolla-lobes broad, ovate or oblong to obovate; rudimentary pistil small,
gland-like :
Perennial; anthers cohering 15. Peponium
Annual; anthers free (L. siceraria) 13. Lagenaria
BB
Receptacle short, rotate, campanulate or shortly tubular, not much longer than
wide ;
Style surrounded by a disc at the base:
Ovary smooth (glabrous or hairy):
Fruit smooth or somewhat foveolate, glabrous, usually small, globose
or sometimes somewhat fusiform, rarely attaining 2-5 cm, seeds
marginate 4. Melothria
Fruit usually hairy, more than 2 cm (usually more than 3 cm) long,
seeds not marginate 12. Cucumis
Ovary muricate, or covered with soft thick spines ; fruit with warts, muricate
or with soft spines:
Soft spines on ovary and fruit bearing several bristles and intermingled
with setae; female flowers often with developed, dorsifixed, linear,
sessile anthers ; seeds marginate 5. Oreosyce
Soft spines on ovary and spines or warts on fruit not covered or inter-
mingled with bristles (sometimes ending in a bristle); only minute
staminodes present; seeds emarginate 12. Cucumis
Style not surrounded by a disc at the base:
Flowers small or minute; corolla usually greenish- white or pale yellow:
Fruit a red berry circumscissile at the base 7. Corallocarpus
Fruit not as above:
Tendrils simple (see also 14. Trochomeria sagittata) 6. Kedrostis
Tendrils bifid:
Ovary with 2 placentas and numerous horizontal ovules; fruit a berry
with tumid unwinged seeds 6. Kedrostis
Ovary 3-loculed with few pendulous ovules in each locule; fruit
dry, dehiscent with 3 valves at the apex; seeds winged
1. Gerrardanthus
Flowers larger; corolla yellow, buff, orange or white with green veins,
rarely greenish-yellow :
Ovary subglobose or, if ellipsoid, muricate; plants never climbing;
corolla yellow or yellowish, if white with distinct green veins (see
also 16. Cocci nia rehmannii) 11. Citrullus
Ovary ellipsoid, not muricate; corolla white with distinct green veins
(L. mascarena) 13. Lagenaria
7
Ovary ovoid, narrowly ellipsoid to narrowly cylindric or fusiform-linear,
not muricate; corolla various but often yellow 16. Coccinia
Receptacle tubular, funnel-shaped or subcylindric, distinctly longer than wide:
Leaves usually deeply dissected, rarely over 6 cm in diameter; petals often
narrow, tapering towards the apex; ovary glabrous; 14. Trochomeria
Leaves more or less lobed but usually not deeply dissected, often well over
6 cm in diameter; petals broad (see also 13. Lagenaria siceraria) 15. Peponium
AA
Fruit dry, linear-fusiform, few-seeded, ultimately dehiscent by longitudinal slits ;
seeds not winged 8. Raphanocarpus
Fruit dry, subtruncately 3-valved at the apex, seeds few, winged.. . 1. Gerrardanthus
Fruit fleshy though sometimes hard-shelled, dehiscent or indehiscent, seeds not
winged ;
Fruit muricate or with distinct protuberances or with soft spines:
Fruit ultimately dehiscent, often orange or scarlet ; seeds much compressed,
marginate, usually appearing as if corroded especially along the margin
9. Momordica
Fruit indehiscent, rarely orange or scarlet; seeds not appearing as if
corroded :
Soft spines on fruit bearing several bristles and intermingled with setae;
seeds marginate 5. Oreosyce
Soft spines or warts on fruit not setiferous on the sides and not intermingled
with setae but sometimes terminating in a single soft bristle; seeds
emarginate :
Fruit usually well over 5 cm long, ellipsoid, pale yellow or pale yellowish-
green, covered with coarse blunt conical protuberances; tendrils
rigid, straight, spinescent; leaves deeply palmatisect; plant
prostrate (see also 12. Cucumis ficifolius) 11. Citrullus
Fruits not bearing warts, spines or other protuberances, but sometimes hairy;
Fruit circumscissile at the base, scarlet when ripe; seeds tumid 7. Corallocarpus
Fruit not circumscissile at the base:
Leaves twice trifoholate (M. clematided) 9. Momordica
Leaves not twice trifoholate:
Outer layer of fruit soft, usually easily squashed or, if harder, either fruit
under 2 cm in diam. or geocarpic, or ellipsoid-fusiform, rarely
exceeding 10 cm in length; seeds ovate or elliptic to subglobose,
not subrectangular and not notched on one or both ends:
Seeds distinctly compressed, often marginate:
Fruit small, globose or rarely somewhat fusiform, smooth or finely
foveolate, glabrous, under 3 cm long; seeds marginate
4. Melothria
Fruit larger, rarely under 3 cm long, glabrous or hairy, never
foveolate :
Fruit over 5 cm in diam., often much larger, not geocarpic,
glabrous, never scarlet, globose (in cultivated specimens
often very large, oblong, and seeds usually dark coloured)
11. Citrullus
Fruit not globose or, if so, seeds whitish or fruit geocarpic or
fruit under 5 cm in diam.:
Seeds smooth, more or less regularly elhptic or ovate in
outline (not triangular or produced on one side), white;
fruit hairy or glabrous, but not scarlet when ripe (see also
14. Trochomeria) 12. Cucumis
Seeds more or less triangular in outline, or ovate-oblong and
produced on the one side, white or dark coloured; fruit
scarlet and always quite glabrous when ripe:
Tendrils usually simple; fruit when ripe easily squashed;
seeds white or light coloured 16. Coccinia
Tendrils bifid; fruit with a firm outer layer; seeds dark
coloured when ripe 15. Peponium
Seeds not much compressed, not marginate:
Tendrils forked 6. Kedrostis
Tendrils simple:
Monoecious or dioecious; seeds usually subglobose 6. Kedrostis
Dioecious; seeds usually ellipsoid or ovate in outline, somewhat
compressed 14. Trochomeria
Outer layer of the large fruit (over 6 cm in diam., never geocarpic)
hard and almost bony (see also 1 1 . Citrullus, wliich has almost
invariably dark and rather regularly shaped ovate, elliptic or
oblong seeds); seeds narrowly triangular, notched or emarginate
on the one side, or subrectangular, truncate or bidentate at one
or both ends 13. Lagenaria
1. GERRARDANTHUS
Gerrardanthus Harv. ex Benih. & Hook./., Gen. PI. 1 : 840 (1867); Harv., Gen. S. Afr.
PI. ed. 2: 127 (1868); Cogn., Mon. Cucurb. 935 (1881); Pflanzenreich 275.1: 18
(1916); Baill., Hist. PL 8: 425 (1886); Pax in Pflanzenfam. 4, 5: 12 (1889); Phillips,
Gen. ed. 2: 744 (1951). Type species: G. macrorhizus Harv. ex Benth. & Hook. f.
Atheranthera Mast, apud Hook. f. in FI. Trop. Afr. 2: 519 (1871).
Climbing shrubs; rootstock tuberous. Stems often long. Leaves petiolate,
membranaceous, cordate or cordate-hastate, entire or more or less 3-5-lobed. Tendrils
bifid near the apex. Flowers dioecious, small. Male flowers racemose or paniculate,
sometimes appearing as if fascicled; receptacle small, rotate; sepals 5, small; corolla
rotate or widely campanulate, deeply 5-partite with oblong or linear segments, of which
two are slightly larger than the other 3; stamens 4; filaments remote, short, incurved;
anthers small, often horizontal in the flower, 1-thecous, all or in pairs cohering,
connective sometimes with a dorsal (in the flower erect) subulate spur; staminode
filiform or subulate ; no rudimentary pistil present. Female flowers solitary or racemose ;
receptacle and perianth as in the male; staminodes often absent; ovary elongate,
trigonous, imperfectly 3-locular with 2 or few pendulous ovules per locule; styles 3,
short, patent, subtrigonous and stigmas truncate-bilobed, or style 1, short, thick and
conical with a sessile reniform stigma. Fruit elongate, terete-obconical, dry, coriaceous,
broadly truncate and 3-valved at the apex, few-seeded. Seeds oblong, compressed,
with a large elliptic or oblong wing at the apex; testa crustaceous, finely granulate;
cotyledons oblong, straight; radicle short, straight, conical.
Found in tropical Africa, extending into Angola and through Natal to the eastern
Cape Province. Two species in South Africa.
Leaves glabrous ; connective without a spur-shaped appendix; styles 3 1. G. macrorhizus
Leaves tomentose on lower surface; connective produced into a spur-like appendix; style 1,
short, conical 2. G. tomentosus
1. G. macrorhizus Harv. ex Benth. & Hook, f, Gen. PI. 1: 840 (1867); Cogn.,
Mon. Cucurb. 935 (1881); Pflanzenreich 275.1; 19 (1916).
G. megarhiza Decne. & Harv. in Harv., Gen. S. Afr. PI. ed. 2: 127 (1868).
9
Type: The original description is the same as the generic description in Genera
Plantarum, but no specimens are cited. Harvey mentions under G. megarhiza no
actual specimen, but he states that he received material from W. T. Gerrard, in whose
honour he named the genus. By inference one might select Gerrard’s specimens in
Trinity College, Dublin, as the type material, but it may be possible to establish which
specimens D. J. Hooker has annotated at Kew when drawing up a description, either
from herbarium records or from his correspondence.
Tuber flattened, bitter 30-60 cm long and up to 1 • 50 m in diam. Stems woody,
high-climbing; ultimate portions slender, angular and strongly sulcate when dry.
Leaves glabrous, drying membranous, broadly ovate-cordate, subtriangular-hastate
or pentagonal in outline, more or less angular to 3-7-lobed, 3-8 cm long and wide,
shallowly and broadly cordate, rarely more deeply so at the base; the margin entire;
the lobes unequal with the terminal one the longest, usually triangular and rounded
to subacute, mucronate; the terminal one often more acute and with a longer mucro;
petioles slender, smooth or somewhat sulcate, glabrous, 1 • 5-4 cm long. Tendrils
elongate, slender, glabrous. Male flowers: common peduncle often very short or
wanting and flowers as if fascicled ; pedicels very slender, 0 • 5-3 cm long; sepals oblong,
obtuse, 3 mm long, corolla rotate, glabrous, brownish, about 12 mm in diam, segments
at the apex often somewhat emarginate; filaments filiform, about 1 • 5 mm long; anthers
horizontal, without spur-like appendage. Female flowers solitary; peduncle fihform,
1- 2 cm long, minutely bracteate at the base; ovary narrowly obconical, 1-5 cm long,
2- 2-5 mm thick; styles 3, about 2 mm long. Fruit obconical-subcylindric, narrowed
at the base, drying brownish-yellow, glabrous and smooth, faintly angular and with
obscure longitudinal ribs, 5-6-5 cm long and 15-22 mm in diam. Seeds bright brown,
compressed, linear-oblong, often somewhat oblique with one nearly straight and one
curved lateral side, narrowed and subtruncate at one or both ends, 16-25 mm long
and 4-7 mm wide; wing oblong to elliptic, pellucid, pale brown, 2-3 cm long and
7-14 mm wide.
Found in eastern Cape Province through Natal to tropical East Africa, in lowland
forests.
Cape Province. — East London: Nahoon Island, J. Wood in Herb. Galpin 6269 (GRA,
L, PRE); Galpin 8215 (PRE). Komgha: Flanagan 702 (BOL, GRA, PRE, SAM);
Schlechter 6217 (GRA); Kei Bridge, Galpin 7776 (PRE, GRA). Kentani: Pegler
1213 (BOL). Mqanduli: Pegler 628 (PRE).
Natal. — Umzinto: Mtavalumine (Umtwalumi), Rudatis 2074 (NH). Pinetown:
Marianhill, Forbes 1028 (NH). Nongoma: Mkuzi, Gerstner in NH 22803 (NH).
Zululand, “ Umhlatuzi Valley near Falls ”, Wood 2179 (NH, SAM).
Swaziland. — Lebombo Mts., Keith s.n. (PRE).
Portuguese East Africa. — Sul do Save: Lebombo Range, Goba, Esteves de Souza
114 (PRE).
Although the original specimens collected by Gerrard were not studied, the identi-
fication of the specimens is certain because specimens were compared at Kew and the
gathering Schlechter 6217, cited by Cogniaux in his monograph, was available.
2. G. tomentosus Flook.f. in Curtis, Bot. Mag. t.6694 (1883); Cogn. in Pflanzenreich
275.1: 20 (1916). Type: Natal, Inanda, Wood 451 (K, holo.; NH, iso.!)
Branches somewhat robust, sulcate, greyish, when young shortly and densely
tomentose, when older glabrescent. Leaves: petioles robust, shortly and densely
tomentose, 3-5 cm long; lamina membranous, reniform-suborbicular, slightly 5-7
lobed or almost to the middle, deep green, shortly and sparsely hairy above, greyish-
green, more or less densely tomentose, 6-12 cm long and broad; lobes ovate or
triangular, somewhat acute, entire or slightly undulate ; basal sinus narrowly rectangular,
2-2 - 5 cm deep. Tendrils rather robust, very long, rather densely and shortly puberulous.
10
Flowers dioecious; male flowering branches slender, 10 cm long; flowers fasciculate
at the nodes, fascicles bracteate; pedicels very slender, spreading, 3-12 mm long;
bracts fohaceous, shortly petiolate, triangular-lanceolate, acuminate, 1 cm long; buds
scarcely 2 mm thick; corolla 12-18 mm in diam., its segments ovate-oblong, obtuse,
the margin broadly reflexed; filaments rather short; anthers elUptic-ovoid, connective
produced into an elongate spur; female flowers solitary or in short 3-4 flowered
racemes; sepals broadly triangular, obtuse, 2 mm long; staminodes setaceous; ovary
tubulose-campanulate, 10-nerved, very shortly puberulous, 1-5-2 cm long; style single,
very short, thick, conical; stigma sessile, reniform. Fruit narrowly campanulate,
10-nerved, smooth, the apex obtuse, trivalved, 6-7 cm long. Seeds pale fulvous,
slightly marginate, 20-25 mm long, 5-6 mm broad; wing pellucid, its base shortly
decurrent, obovate-oblong, 3 cm long, 17-18 mm broad, its apex rotund.
Recorded only from Natal.
Natal. — Durban: near Durban, McClean s.n. (PRE). Inanda: Wood 451 (NH).
This apparently very rare species has, as far as can be ascertained, been collected
only once since Wood collected the type specimen nearly 80 years ago, viz. by McClean
in 1941.
2. DACTYLIANDRA
Dactyliandra Hook, f in FI. Trop. Afr. 2: 557 (1871); Cogn., Mon. Cucurb. 626
(1881); Pflanzenreich 275 .1 : 129 (1916); Pax in Pflanzenfam. 4, 5 : 15(1889); Phillips,
Gen. ed. 2: 745 (1951). Type species: D. welwitschii Hook. f.
Slender annual herb, scandent or prostrate. Leaves once or twice tripartite or
palmately 5-7-lobed; petiole with a stipuliform bract-like organ at the base. Tendrils
simple. Flowers minute, monoecious. Male flowers not in the same axils as the female
ones, in a pedunculate few- to 12-flowered subumbellate raceme; receptacle cam-
panulate with a ring of hairs with globose tips in the throat; sepals subulate or triangular;
corolla lobes oblong or oblong-elliptic, rounded or obtuse at their tips, outside glandular-
papillose or pubescent; stamens 3 inserted on the receptacle near the base; filaments
short, free; anthers ovate, cohering, two 2-thecous and the third 1-thecous; thecae
linear, but bent like a horse-shoe; connective not produced beyond the anthers;
rudiment of ovary represented by a depressed gland. Female flowers solitary or
occasionally in pairs, calyx and corolla as in the male; staminodes 0; ovarium oblong
to subglobose, attenuate at the apex; placentas 3; ovules numerous; style columnar,
inserted in a disc; stigma subcapitate, 3-lobed. Fruit a small, globose indehiscent
berry, red when ripe, usually few-seeded. Seed irregularly oblong in outline, dilated
about the middle and truncate at the ends, slightly compressed with 2 flatfish surfaces
and a distinct margin; testa very thick.
A monotypic genus found in Angola and South West Africa. Cogniaux (1916)
recognised two species, but the second is indistinguishable from the type species.
D. welwitschii Hook, f in FI. Trop. Afr. 2: 557 (1871); Cogn., ll.cc. (1881);
(1916); Hiern, Cat. Welw. Afr. PI. 1, 2: 403 (1898). Type: from Angola.
Blastania luederitziana Cogn. apud Schinz in Abh. Bot. Ver. Brandenb. 30: 152 (1888);
Dinter in Fedde, Repert. 15: 349 (1917-1919).
Dactyliandra luederitziana (Cogn.) Cogn. in Pflanzenreich 131 (1916).
Stems up to about 1 m long, usually slender and much branched, sulcate and with
short white bent thick bulbous-based hairs on the ribs ultimately turning scabrid.
Leaves digitately 5-7-lobed nearly to the base, on both sides shortly setulose-scabrid,
ultimately finely white-punctate and usually very scabrous, 3-8 (-12) cm long and wide;
the segments oblong to obovate or lanceolate, acute to obtuse, mucronate, much
11
narrowed towards the base, irregularly and coarsely dentate to pinnatilobed, the middle
one the longest, the lateral ones gradually smaller, lowermost often with a lateral lobule
or bilobed; petioles slender, scabrid-setose, 2-5 cm long. Stipuliform bract sessile,
suborbicular-reniform to ovate, sometimes slightly 2-lobed or asymmetrical, shortly
setose, later white punctate and scabrid, long ciliate-dentate along the margin, 5-15
mm long and 3-13 mm wide. Male flowers: common peduncle slender, often filiform,
sparsely asperulous, 2-5 cm long; pedicels erectopatent, filiform, scabridulous or
shortly setulose, up to 15 mm long, at the base with minute, subulate, early deciduous
bracts; receptacle shortly hirsute, 1-5-2 -5 mm long and wide; sepals patent to
recurved, 0-5-1 -5 mm long; corolla whitish or yellowish papillose-pubescent, 1-5-3
mm long. Female flower: pedicel short; ovary shortly papillose-puberulous, 2-3 mm
in diam. Fruit ultimately glabrous, smooth, about 15 mm in diam.; pedicel up to
2 cm long. Seeds 5-7 mm long, 3 - 5-4 mm wide and about 1 - 5 mm thick.
Type: In the original description Hooker cites: “ Sandy thickets in Loanda,
Dr. Welwitsch ”. According to Hiern and Cogniaux there is only one Welwitsch
number (No. 833) collected in Loanda and this is apparently the type gathering
(Welwitsch also collected this species in Mossamedes, No. 832). The actual type
specimen is apparently in BM.
Angola. — Loanda; Welwitsch 833 (COI, iso.!); Muceque de Calemba, Gossweiler
“ 10” (leg. Sept. 1935, COI). Benguela: Menyhart 218 (Z). Mossamedes: Exell
& Mendoni'a 2152 (COI); Rio Mucungo, Carisso & Sousa 318 (COI); Sierra de Lua,
Gossweiler 10694 (COI).
South West Africa. — Kaokoveld; Kunene banks, Story 5795 (PRE); near Sanitatas,
Story 5683 (PRE); Oruwanjai, de Winter & Leistner 5645 (PRE). Ovamboland:
Oshando,5c/7z«z317(Z). Grootfontein: Tsumeb, ILczrts.n. (PRE); Namutoni, .ffreycrin
TM 20658 (PRE). Outjo: Pamela, Volk 2S9\ (NH); between Franzfontein and Brandberg,
de Winter 3121 (PRE). Omaruru: Brandberg, Liebenberg 5011 (PRE); Merxmiiller
& Giess 1577 (M, PRE). Okahandja: Dinter 12 (COI, GRA, L, SAM); 76 (SAM).
Swakopmund: Bradfleld 586 (PRE). Karibib: Kinges 3449 (PRE). Rehoboth:
between Rehoboth and Uhlenhorst, Wilnian 459 (BOL, PRE); between Kalkrand and
Rehoboth, de Winter 3537 (PRE); Acocks 18160 (PRE); Naukluft Mts., Buellsport,
Strey 2146; 2178; 2336 (PRE). Gibeon; Packriem, Range 1350 (SAM). Liideritz;
Otjisewa, Kinges 2484 (PRE).
Although the original specimens cited by Cogniaux 1888 under Blastania luederit-
ziana were not seen, some specimens cited by Cogniaux in his 1916 monograph {Dinter
12, Schinz 317) leave no doubt about its identity. These specimens are indistinguishable
from the type specimen of Dactyliandra welwitschii (the “ differences ” mentioned by
Cogniaux in his 1916 monograph do not hold). A characteristic of all specimens,
both from Angola and from South West Africa, is that they dry to a rather light,
somewhat greyish green.
3. BLASTANIA
Blastania Kotschy & Peyr., PI. Tinn. 15, t.7 (1865-66); edit. Kanitz 21 (1868); Cogn.,
Mon. Cucurb. 627 (1881); Pflanzenreich 275.1: 133 (1916); Pax in Pflanzenfam.
4, 5: 16 (1889); Phillips, Gen. ed. 2; 745 (1951). Type species: B. garcinii (L.) Cogn.
(fide Phillips).
Ctenopsis sensu Hook. f. ex Naud. in Ann. Sci. Nat. 5me ser. 6: 12 (1867), non De
Notar. (1847).
Ctenolepis Hook. f. in Benth. & Hook. f. Gen. PI. 1: 832 (1867); FI. Trop. Afr. 2:
557 (1871); Clarke in Hook, f., FI. Brit. Ind. 2: 629 (1879).
12
Annual scabrid, prostrate or scandent herbs, often drying dark. Leaves orbicular
in outline, deeply digitately 3-5-lobed or -partite with a large stipuliform, orbicular or
oblong to subreniform, toothed or ciliate bract at the base of the slender petiole.
Flowers monoecious. Male flowers racemose; receptacle short, campanulate; sepals
small, subulate; corolla rotate, deeply 5-partite; stamens 3, inserted on the receptacle
free; filaments very short ; anthers small, one 1-thecous, two 2-thecous; thecae straight ;
connective not apically produced; rudimentary pistil wanting. Female flowers solitary
in the same axils as the male raceme, shortly pedicelled; receptacle, calyx and corolla
as in the male; staminodes 0; ovary ovoid to subglobose with few horizontal ovules
on 2-3 placentas; disc absent; style columnar; stigmas 2, rarely 3. Fruit small,
fleshy, globose, sometimes more or less obhque, 2- or few-seeded, indehiscent. Seeds
ovoid, much compressed, concave or boat-shaped above, convex beneath, smooth,
with an acute margin.
Species 2, the type species in India and Ceylon, the second widespread in tropical
Africa, India and Ceylon.
B. cerasiformis (Stocks) A. Meeuse, comb. nov.
Zehneria cerasiformis Stocks in Hook. Journ. Bot. Kew Card. Misc. 4: 149 (1852).
Lectotype: A syntype of both Zehneria cerasiformis and Blastania fimbristipula is
Kotschy 205 and this is selected here as the lectotype.
Bryonia fimbristipula Fenzl in Flora 19; 312 (1844), nomen tantum.
Blastania fimbristipula Fenzl ex Kotschy & Peyr., PI. Finn. 15, t.7 (1865-66); Cogn.,
ll.cc. (1881); (1916); Hutch. & Dalz., FI. W. Trop. Afr. 1: 178 (1927).
Ctenopsis cerasiformis (Stocks) Naud. l.c. 13 (1867).
Ctenolepis cerasiformis (Stocks) Hook. f. in FI. Trop. Afr. 2: 558 (1871); C. B.
Clarke, l.c. 629 (1879); Andrews, Flow. PI. Anglo-Egypt. Soudan 1; 181 (1950);
Keay, FI. W. Trop. Afr. ed. 2, 1, 1; 208 (1954).
Stems usually slender, elongate, usually branched, angular-sulcate or striate,
glabrous or shortly and sparsely hispid, ultimately often slightly scabrid, occasionally
white-punctate and scabrid, often geniculate at the nodes; internodes sometimes up
to 12 cm long. Leaves palmately 3- or 5-partite, membranous, 2-5-10 cm long and
wide, the segments ovate-oblong or obovate to elliptic or lanceolate, narrowed at both
ends, acute or subobtuse, mucronate, remotely and often coarsely serrate-dentate, on
both surfaces shortly scabrid-setose, ultimately white-punctate and scabrid; the central
one with longer mucro, entire or 3-lobed; lateral ones smaller; basal ones often more
or less deeply 2-lobed; petioles somewhat dorso-ventrally flattened and longitudinally
striate-sulcate, densely shortly setose-hispid, ultimately white-punctate and scabrous,
1-5 cm long. Stipuliform bracts ovate, oblong, subreniform to suborbicular, subsessile
or narrowed and subpetioled at the base, shortly setose-scabrid or scabrid on the abaxial
side, long-ciliate, 4-20 mm long and 3-15 mm wide. Male flowers: common peduncle
filiform, glabrous or puberulous, 0-5-3 cm long, subcapitately 3-10-flowered; pedicels
1-3 mm long; receptacle 0-5-1 mm long and about 1 mm wide; sepals about 0-5
mm long; corolla whitish or yellowish, its segments 0-8-1 mm long. Female flowers:
pedicel short, in fruit up to about 5 mm long; ovary in anthesis about 2 mm in diam.,
glabrous or nearly so. Fruit smooth, glabrous, 10-15 mm in diam. Seeds often flat
or slightly concave on the one side and convex on the other side, smooth, attenuated
to acute at the base, 7-9 mm long, 4-6 mm wide and about 1-5 mm thick.
Bechuanaland. — Between Kachikau and Kasane, nr. Chobe River: Erens 394 in
Hb. Pole-Evans 4195 (PRE, SRGH). N’gamiland: Curson 71 (PRE).
Transvaal. — Kruger National Park: Pilgrims Rest district, near Satara, v.d. Schijff
2224 (PRE); Nelspruit district, between lower Sabie and Skukuza, v.d. Scliijfl' 1772
(PRE).
13
The following specimens often quoted in regional floras and monographs were
also studied:
Portuguese East Africa. — Boruma: Menyliart 645a (Z).
Soudan.— Cordofan: Arasch-Cool, Kotschy 205 (L, M, syntype gathering of Blastaiiia
fimbristipula and Zehneria cerasiformis); Sennaar, Kotschy 331 (L); Togodile, Ehren-
berg 188 (L).
Senegambia. — Dagana: Leprieur (BR, L).
Arabia. — Yemen: Badjil, Schweinfurth 557 (BR).
4. MELOTHRIA
Melothria L., Sp. PI. ed., 1: 35 (1753); Ser. in DC., Prodr. 3: 313 (1828); Benth.
& Hook, f., Gen. PI. 1: 830 (1867); Hook. f. in FI. Trop. Afr. 2: 562 (1871); Cogn.,
Mon. Cucurb. 572 (1881) and in Pflanzenreich 275.1: 75 (1916); Baill., Hist. PI. 8:
446 (1886); Pax in Pflanzenfam. 4, 5: 15 (1889); Phillips, Gen. ed. 2: 745 (1951).
Zehneria Endl., Prodr. FI. Norf. 69 (1833); Gen. 936 (1840); Sond. in FI. Cap. 2: 485
(1862); Benth. & Hook, f., l.c. ; Hook, f., op. cit. 558 (1871).
Pilogyne Schrad., Ind. Sem. Hort. Goett. (1835), nomen; ex Eckl. & Zeyh., Enum.
PI. Afr. Austr. 277 (1834); Naud. in Ann. Sci. Nat. 5me. ser. 5: 36 (1866).
Mukia Arn. in Hook., Journ. Bot. 3: 271 (1841); Naud. in Ann. Sci. Nat. 4me. ser.
12: 141 (1859); Benth. & Hook, f., op. cit., 829; Hook, f., op. cit. 561.
Perennial or sometimes annual, climbing or prostrate often forming annual
herbaceous stems from a perennial rootstock. Stems usually slender. Leaves entire
to more or less distinctly palmately lobed, often thin in texture. Tendrils simple (very
rarely bifid, not in S. Afr.). Flowers monoecious or dioecious, minute or small, male
flowers corymbose, racemose or subumbellate in few- to many-flowered inflorescences,
less often fasciculate or solitary, in monoecious plants usually not in the same axils
as the female flowers ; receptacle campanulate; sepals dentiform, often minute; corolla
white to yellow or greenish deeply 5-partite; stamens usually 3; filaments inserted on
the receptacle, free, sometimes with a ring or zone of hairs; anthers free or occasionally
slightly cohering, two 2-thecous and the third 1-thecous (if stamens 2 both 2-thecous,
if stamens 4 only 1, 2-thecous, if stamens 5 all 1-thecous); thecae straight or rarely
curved; connective rarely produced beyond the anthers, often papillose; rudiment of
pistil globose or annular, rarely 3-lobed; female flowers solitary, fascicled or subumbel-
lately racemose or corymbose; receptacle and perianth as in the male; staminodes 3,
rarely bearing remains of anthers or wanting; ovary globose or ovoid or sometimes
fusiform, with 3 placentas few- to many-ovuled constricted under the receptacle;
style short, at the base surrounded by an annular disc; stigmas 3, linear, very rarely 2,
or one 3-lobed stigma. Fruit an indehiscent usually small berry, globose, ovoid,
ellipsoid to fusiform, few- to many-seeded. Seeds ovate or suborbicular in outline,
much compressed or rarely thick, often marginate, smooth or occasionally pitted or
rugose.
Type species: Melothria pendula L. (being the only species in this genus at its
inception).
About 100 described species, circumtropic, with some species extending into more
temperate zones. This genus, the largest of the Cucurbitaceae, has been divided by
Cogniaux into three subgenera: Eumelothria which should now be called Melothria,
Solena and Mukia. The limits between these subgenera are not very sharply defined,
in fact in his latest monographic treatment of the family, Cogniaux sometimes treated
one species under two different names and in two different subgenera (see under
M. marlothii).
14
Flowers monoecious:
Flowers sessile, clustered; plant roughly hirsute-setose, leaves long triangular-cordate, seeds
areolate 1- M. mademspatana
Flowers usually distinctly pedicellate; pubescence and leaf-shape not as above; seeds smooth:
Glabrous in nearly all its parts 2. M. marlothii
More or less densely pubescent 3. M. cinerea
Flowers normally dioecious:
Vegetative parts and calyx quite glabrous 4. M. parvifolia
Vegetative parts and calyx more or less hairy (sometimes pubescence scanty, but at least on
main veins of lower leaf-surface) 5. M. cordata
l.M. maderaspatana (L.) Cogn., Mon. Cucurb. 623 (1881) and Pflanzenreich
275.1: 126 (1916); Pax in Engl., Pflzw. O.-Afr. C: 396 (1895); Hiern, Cat. Welw.
Afr. PL 1 (2); 403 (1898); Burtt Davy, FI. Transv. 1: 225 (1926); Hutch. & Dalz.,
FI. W. Trop. Afr. 1: 179 (1931); Andrews, Flow. PI. Anglo-Egypt. Soudan 1: 181
(1950). Type: from India.
Cucumis maderaspatensis L., Sp. PI. ed. 1 : 1012 (1753); Willd., Sp. PI. 4; 615 (1805).
Bryonia scabrella Linn, f., Suppl. PI. 424 (1781); Willd., op. cit., p. 619; Ser. in DC.,
Prodr. 3: 306 (1828).
Mukia scabrella (Linn, f.) Arn. in Hook. Journ. Bot. 3: 216 (1841); Naud. in Ann.
Sci. Nat. 4me ser. 12: 142 (1859); Sond. in FI. Cap. 2: 489 (1862); Hook. f. in FI.
Trop. Afr. 2: 561 (1871).
Bryonia micropoda E. Mey. ex Drege, Zw. Pflzgeog. Doc. 158 (1843), nomen tan turn.
Annual. Stems several, prostrate or climbing, usually much branched, up to about
1 m long, the younger portions slender, sulcate, hispid to subaculeate with short
usually curved whitish bulbous-based stiff hairs, older ones much stouter, glabrescent.
Leaves thinly herbaceous to firm, ovate-hastate to oblong-hastate or narrowly triangular,
undivided to 3- or occasionally 5-lobed; 3-15 cm long and 2-5-13 cm wide, the apex
subacute to acute or shortly acuminate, minutely mucronate; the basal sinus usually
wide and shallow, sometimes almost imperceptible, the margin very minutely denticulate
or almost entire to distinctly crenulate-dentate; upper surface setose-scabrid becoming
punctate-scabrid and very rough, somewhat setose on the midrib; lower surface shortly
hispid-setose turning scabrid-punctate, the veins remaining hispidulous; primary and
secondary nerves not conspicous above (except the midrib), prominent below forming
a coarse conspicuous reticulum; petioles shortly setose-hispid or aculeate-hispid,
sometimes densely so, 0-75-8 cm long. Tendrils sparsely aculeate-hispid like stems and
petioles, finely sulcate at least in lowermost portion. Flowers monoecious, fascicled
or the females occasionally solitary, sessile or shortly pedicelled. Male flowers:
receptacle long pilose-hirsute, about 2 mm long; sepals erect, 1-1-5 mm long; corolla
yellowish, pilose-hirsute outside, about 2 mm long; connective of stamens produced
into a usually bifid apical portion. Female flowers: ovary subglobose, covered with
usually blackish setae. Fruit globose, ultimately glabrous and red when ripe, juicy.
Seeds white, elliptic, elliptic-oblong or ovate-oblong, rather thick, verrucose or scrobi-
culate, distinctly margined (the margin usually with finely raised central portion),
about 4 X 2-5 X 1-5 mm.
Type : There is no specimen in the Linnaean Herbarium which is labelled “ Cucumis
maderaspatensis”. Accordingly, the figure cited by Linnaeus in Pluk., Almag. t.l70,
f.2 (1696) must be taken to represent the type.
Found throughout tropical Africa to Angola, Transvaal and Natal and northwards
to Senegal and Egypt. Also from India to China, Formosa, Philippines, Malesia and
Australia.
Recorded from the following districts: Transvaal: Soutpansberg, Pietersburg,
Tzaneen, Nelspruit, Barberton; Natal: Mtunzini, Lower Tugela, Inanda, Durban,
Pinetown, Polela.
15
The following specimens are of special interest: Junod 2622, from Shilovane,
Tzaneen, Transvaal, in Z, erroneously cited by Cogniaux in Pflanzenreich 275 . 1 : 68
(1916) as Oreosyce triangularis Cogn, (see under Oreosyce)-, Schlechter 11845, from
Komatipoort, Barberton, Transvaal, in BOL, COI, GRA, PRE, cited by Cogniaux
(1916) and by Burtt Davy; Drege s.n., sub. nom. Bryonia^ micropoda E. Mey., from
Durban, Natal, in L, cited by Sonder in Flora Capensis and by Cogniaux; Wood 956,
from Polela district. Natal, in GRA, cited by Cogniaux (1916).
2. M. marlothii Cogn. in Abh. Bot. Ver. Brandenb. 30: 152 (29th Sept. 1888);
in Engl. Bot. Jahrb. 10: 270 (9th Oct. 1888); and in Pflanzenreich 275.1: 100 (1916);
Dinter in Fedde, Repert. 19: 188 (1923/24). Type: see below.
M. acutifolia Cogn. in Bull. Herb. Boiss. 3: 419 (1895) and in Pflanzenreich 275.1:
105 (1916); Burtt Davy, FI. Transv. 1: 225 (1926).
Annual. Stems climbing, slender, sometimes filiform, sulcate, glabrous or nearly
so, often much branched, up to several meters long. Leaves thinly herbaceous drying
membranaceous, ovate or narrowly deltoid-cordate to 5-angled, entire, angular or
tricuspidate to shallowly 5-lobed, attenuate to acute or acuminate and terminating in
a mucro at the apex, with a usually rather large, rounded semi-orbicular or subquadrate
to rectangular shallow to fairly deep basal sinus and a remotely and minutely callous-
dentate to subentire margin, 2-8 cm long and 1-5-6 -5 cm wide, glabrous to finely
and minutely setose-scabrid or punctate on both surfaces, but more distinctly so on
upper surface, lower one sometimes smooth; main nerves slender, secondary and finer
nervation inconspicuous; petioles slender, glabrous, 0-5-4 cm long. Tendrils filiform
to capillary, glabrous. Flowers monoecious, minute, greenish yellow, or whitish, the
male and female flowers usually produced in the same axils. Male flowers subumbel-
lately racemose but raceme often contracted so that they appear fascicled, 3-10 (-16)
flowered; common peduncle usually under 1 cm long but occasionally reaching 3 cm,
capillary, glabrous; pedicels very slender, 1-5 mm long; flowers 1-5-2 mm long and
as wide; receptacle hemisphaerical; sepals very minute, subulate, corolla glabrous.
Female flowers solitary or binate; pedicels very slender, in fruit up to about 8 mm
long; ovary spherical, glabrous. F/-w/F spherical, a dirty greenish red when ripe,
finely reticulate, 6-9 mm in diam. Seeds white, much compressed, ovate in outline,
not or obscurely marginate, 3 -8-4 -5 mm long and 2 -8-3 -3 mm wide.
Type'. Cogniaux cited Marloth 957 and a specimen collected by Schinz No. 320,
from Amboland in the original description. The original t3pe specimen in Berlin
was destroyed. A specimen in PRE labelled “ Melothria marlothii ” Cogn. from
Barkly-West, Victoria Farm, Jan. 1886 {Marloth No. 957) is Coccinia rehmannii Cogn.
There is another specimen in PRE labelled “ Melothria, Queipiitz, Bechuanaland, May
1886 ”, numbered Marloth 1381, which is undoubtedly Melothria marlothii Cogn.
I could not find No. 1381 cited anywhere (not in Engler’s “ Plantae Marlothianae ”
in Engl. Bot. Jahrb. 10, nor in Cogniaux’s publications in Schinz’s “ Beitrage zur
Flora von Africa ” published in several journals, or in Cogniaux’s Cucurbitaceae
monographs). As Coccinia rehmannii was described only as late as 1895 and the
specimen in PRE numbered Marloth 957 bears fruit only and no flowers, so that it was
not adequate for description and may have remained unnamed in the Berlin herbarium
(which point unfortunately cannot be established any more), whereas Melothria marlothii
was recognised as an undescribed species in 1888 and, in any case, all unnamed specimens
of this genus in the Berlin herbarium must have been studied by Cogniaux for his
1916 monograph in “ Das Pflanzenreich ”, it would seem that there has never been
a Melothria in the Berlin Herbarium with the number 1381 but possibly a Coccinia
and, consequently, that the labels in Marloth’s own herbarium were interchanged.
However, there is not sulficient evidence that the specimen labelled “ 1381 ” in PRE
can be taken to represent an isotype. The original holotype in B was destroyed.
16
The description of M. niarlothii and all other specimens cited by Cogniaux leave
no doubt that the original type specimen {Marloth 957 in B) was a Melothria and not
a Coccinia. The identity of M. marlothii is not in doubt either, because the Schinz
specimen from Oshiheke, Amboland in Z is still extant. This specimen is designated
here as the lecto-type.
South West Africa to Transvaal, extends through Bechuanaland northwards to
Northern Rhodesia.
South West Africa. — Kaokoveld: Koako Otavi, t/c Winter &Leistner 5518; Ohopoho,
de Winter & Leistner 5332 (K, M, PRE). Ovamboland: Oshiheke, Schinz 320 (Z).
Okavango: Nyangana, de Winter & Marais 4881 (PRE), Runtu, de Winter & Marais
4492 (PRE). Grootfontein : Otavifontein, Dinter 5413 (B); on road to Abenab,
Schoenfelder S 569 (PRE); farm Kumkauas, Kinges 2911 (PRE); Gautscha Pan,
Story 6237 (PRE); Otjiwarongo: Waterberg, Rodin 2584 (BOL); Okanjande, Range
s.n. (BOL). Karibib: Erongo, Pearson 9841 (BOL). Okahandja: BradfieJd 224
(PRE); de Winter 2704 (PRE). Windhoek: Otjiesewa, Wiss <6 Kinges 784 (PRE).
Rehoboth: Naukloof Mts., farm Buellsport, Strey 2179 (PRE, NBG). Locality
unknown: “ Okakarara ”, Liebenberg 4693 (PRE).
Bechuanaland. — Ngamiland: Mabele a Pudi, S'on in T.M. 28796 (PRE); Linyanti
River plains, Mrs. Moss in herb. Moss No. 18566 (J); Chukudu Pan, Story 4940
(PRE).
Cape Province. — Barkly West: Victoria Farm, Marloth 957 (type originally in B,
no isotypes seen); locality not known but most probably near Kuruman; Queiputz,
Marloth 1381 (PRE, see under “ Type specimen ”).
Transvaal. — Brits: Hartebeespoort, ilTogg s.n., Rc/pto/? 586. Pretoria: near Pienaars
River, Meeuse 9542; Roodeplaat Exp. Stn., Leendertz in T.M. 8639, Rehm s.n. (all
PRE). Pretoria or Bronkhorstspruit district: “ Neu Halle”, Elandsrivier, Rehmann
4904 tsyntype of M. acutifolia Cogn., Z). Waterberg: near Ellisras, Meeuse 10541
(PRE), Ons Hoop, Codd 8468a (PRE).
Southern Rhodesia. — Victoria Falls, Rogers 13192 (GRA).
Northern Rhodesia. — Livingstone, Rogers 7140 (BOL); Mambova, Munro ML17
(PRE).
Cogniaux (1916) placed M. acutifolia in the section Solena (Lour.) Cogn., charac-
terised by almost invariably dioecious flowers and usually marginate seeds. However,
M. acutifolia has monoecious flowers and immarginate or obscurely marginate seeds.
When using Cogniaux’s key one would not hesitate to look for this species among the
species of the section Melothria {=Eumelothria Cogn.) which has monoecious flowers
and often immarginate seeds and to which Cogniaux correctly referred M. marlothii.
The original descriptions of the two species, repeated by Cogniaux in his 1916
monograph, are also identical for practical purposes, the only difference being “ Flores
. . . fasciculati ” in M. marlothii against “ Flores masculi racemosi ” in the other
species. The male flowers are actually subumbellately racemose on a short common
peduncle, so that young (and also some older) inflorescences appear as fascicles of male
flowers. Finally, a comparison of the syntype specimen Rehmann 4904 (and other
specimens from the same region such as Leendertz in T.M. 8639, Meeuse 9512) with
Schinz 320 (a syntype of M. marlothii) and other specimens from South West Africa
clearly shows the identity.
The species has not been collected frequently, but it is probably much more common
than the herbarium records would indicate, because this rather inconspicuous plant
may have been overlooked. According to an oral communication of Mr. R. G. Strey
of the National Herbarium, formerly of the Farm Buellsport, Rehoboth, South West
Africa, M. marlothii is fairly common in river beds in the Karibib, Okahandja and
Rehoboth districts. This species seems to prefer deep sandy soil.
17
3. M. cinerea {Cogn.) A. Meeuse, comb. nov.
Kedrostis cinerea Cogn. in Bull. Herb. Boiss., 2me. ser., 1: 883 (1901); and Planzen-
reich 275.1: 142 (1916); Dinter in Fedde, Rep. 18: 434 (1922). Type: South West
Africa, Omaruru, Gifkopje, Dinter 1440 (Z, holo.!).
Melothria breyeri Burtt Davy, FI. Transv. 1: 46, 225 (1926). M. hispidula Burtt
Davy, l.c.
Perennial, prostrate or sometimes climbing usually somewhat canescent. Steins
several to many, produced annually from their thickened and woody corticate bases,
slender, herbaceous, sulcate or costate-striate, usually not over 80 cm long, covered
with short setose hairs often intermixed with longer pilose-hirsute hairs but usually
appearing shortly hispid, ultimately glabrescent. Leaves often more or less secund
on creeping stems, varying from suborbicular-penlagonal or reniform-cordate to
broadly cordate-ovate or cordate-triangular in outline; rounded to acute and sub-
acuminate at the apex and with a usually narrow and fairly deep, rarely broad and
shallow, basal sinus, undivided and often 5-angled or sometimes distinctly to deeply
palmately 3-5 (-7)-lobed, 1 -5-6 cm long and about as wide; the margin subentire to
minutely denticulate or sometimes distinctly to rarely coarsely crenate-dentate, bluntly
dentate or repand, usually shortly ciliate; upper surface usually rather densely and
shortly adpressed strigose-scabrid, as a rule slightly more densely so towards the margin,
lower surface rather densely and shortly setose-scabrid or punctate-scabrid and not
infrequently aculeate-setose on the prominent slender main nerves, both surfaces
ultimately turning punctate-scabrid when the stiff hairs wear off; petioles rather slender,
sulcate-ribbed and coarsely and shortly hispid like the stems, 0-5-3 cm, rarely up to
4-5 cm, long. Tendrils filiform, slender to very slender, hispid as are the stems and
petioles. Flowers monoecious, light yellow. Male flowers in contracted epedunculate
few to many (about 20)-flowered racemes, but by reduction sometimes in few-flowered
fascicles; pedicels slender, usually under 1 cm long, the lowermost occasionally up
to 2 cm, articulated at the apex below the calyx, hispid as are the stems, petioles, tendrils
and the rhachis; receptacle narrowly campanulate, rather laxly hispid, 1 -5-3 mm long;
sepals erect, subulate-triangular, 0-5-1 - 3 mm long; petals somewhat obovate or oblong-
ovate, slightly and shortly hairy, 2-5-4 (-5) mm long. Female flowers solitary rarely
2 or 3-nate; pedicels usually short, in fruit rarely exceeding 1-5 cm; ovary long-
rostrate, shortly and densely hirsute-pubescent. Fruit ovoid-oblong or subglobose-
ovoid to somewhat obovoid, usually abruptly contracted into a narrow rostrum, with
about 8 fine longitudinal ribs, at first densely and shortly hirsute -pubescent, more or
less glabrescent, orange when ripe, 13-19 mm long (with the rostrum) and 8-12 mm
in diam., 6-14 seeded. Seeds ovate-elliptic much compressed, somewhat contracted
at one end, finely rugose, finely or obscurely marginate, 4-6 mm long, 3-4-5 mm wide
and 1-1-5 mm thick.
South West Africa. — Kaokoveld: Ohopoho, de Winter & Leistner 5351 (K, M,
PRE). Outjo: near Otjikondo, de Winter 3061 (L, PRE, SRGH); Omaruru: Ohere-
Oos, Merxmueller & Giess 1584 (M, PRE). Karibib: Dinter 2505 (PRE, SAM); 6871
(B, BOL, PRE); Okomitundu, Seydel 1289 (PRE). Okahandja: Waldau, Dinter 377
(PRE, SAM).
Transvaal. — Waterberg or Potgietersrust : Palala River, Breyer s.n. (PRE). Pieters-
burg: Molepo Reserve near P.O. Boyne, Gerstner 5356 (PRE). Lydenburg: Sekukuni-
land, Driekop, Barnard 238; farm Parys, Barnard & Mogg 715; farm Het Fort,
Barnard & Mogg 1011; Nooitgedacht Mt. near Branddraai, Young A 714; between
Branddraai and Ohrigstad, Young A 6\\-, near Ohrigstad, Fowng A 534; Burgersfort,
Meeuse 10285; Steelpoort, Meeuse 10287 (all PRE). Middelburg: near Loskop
Dam, Meeuse 10328 (PRE). Pilgrims Rest: Graskop, Mogg s.n. (PRE). Nelspruit:
White River, Rogers 25960 (isotype of M. breyeri Burtt Davy, PRE); Mogg s.n. (PRE);
Liebenberg 2383 (PRE); Crocodile Poort, Codd 7770 (PRE, SRGH); Kaap Muiden,
Rogers H. No. 24615 (PRE), Mogg s.n. (PRE).
18
Southern Rhodesia.— Near Zimbabwe, Wild 3045 (SRGH).
It is very strange indeed that Cogniaux placed this species in the genus Kedrostis,
because the holotype is quite adequate and is clearly a Melothria (ovary with a disc
at the style-base!). Burtt Davy quite correctly placed it in Melothria and also rightly
concluded that it had never been described in Melothria before. He had no reason
to look for his species under Kedrostis. However, Burtt Davy found two extreme
specimens which he took for two different species, the one with undissected leaves
and the other one with deeply palmatisect leaves. He also reported a difference in
the female flowers which he described as solitary in M. hispidula (described as having
lobed to palmatisect leaves) and as “ about three-flowered ” in M. breyeri (described
as having undivided leaves). The isotype of M. breyeri in the National Herbarium,
Pretoria, has only one female flower in a capsule, so that the arrangement of the female
flowers can not be verified, but the vegetative parts and numerous male flowers agree
quite satisfactorily with the type of Kedrostis cinerea. As far as could be ascertained
the female flowers are usually solitary but fascicles of 2-3 female flowers as reported
by Burtt Davy were occasionally observed. As the degree of dissection of the leaves
varies a great deal, there can be no doubt that all names cited above refer to one taxon.
The field notes indicate that M. cinerea is usually found on granite outcrops, sometimes
on quartzite or sandstone.
4. M. parvifolia Cogn. in Bull. Herb. Boiss. 3: 420 (1895); Pflanzenreich 275.1:
111 (1916); Burtt Davy, FI. Transv. 1: 225 (1926). Syntypes: Rehmann 8839, 8842,
from Durban, Natal (Zi).
Scandent perennial. Stems slender to filiform, glabrous, smooth but sulcate,
usually much branched, up to several meters long. Leaves membranaceous, broadly
ovate-cordate in outline, sparsely and minutely white-punctate to almost glabrous
and dark green above, slightly paler glabrous and smooth beneath, 2-4 cm long and
broad, 3-5-lobed to the middle or less deeply so with ovate-oblong, angular-lobulate
lobes, usually more or less obtuse, apiculate, basal sinus rather shallow, broadly rounded;
petioles slender, sulcate, glabrous, 1-2-5 cm long. Thnt/nA filiform, glabrous. Flowers
dioecious, the male ones racemose, the female ones solitary (always?). Male flowers:
common peduncle filiform, glabrous, 4-6-flowered, 1-2 cm long; pedicels capillary,
erecto-patent, 1-2 mm long; receptacle campanulate, glabrous, about 1-5 mm high
and 2-5 mm in diam.; sepals subulate, minute; petals narrowly ovate, obtuse, minutely
papillose, about 2 mm long. Female flowers: pedicels capillary 4-6 mm long; ovary
subglobose, glabrous. Fruit globose, smooth, 8-10 mm in diam. Seeds narrowly
ovate in outline, smooth, obscurely marginate, about 3 • 5 mm long and 2-5 mm in diam.
Found in Portuguese East Africa, Transvaal, Natal and probably also tropical
East Africa.
Transvaal. — Barberton: Komatipoort, Schlechter Will (according to Cogniaux and
Burtt Davy, not seen).
Portuguese East Africa. — Sul do Save: Louren^o Marques district, Masinga,
Schlechter 12130 (BOL, COI, PRE, Z); Katembe, Schlechter 11613 (BOL, GRA, Z);
Incanhini, Schlechter 12040 (GRA); Goba, Esteves de Sousa 110 (PRE); Louren?o
Marques, Polana, Borle 428 (PRE).
Natal. — Durban: Rehmann 8839, 8842 (syntypes, Z); Durban Bluff, Meebold 13155
(M); Isipingo North, Ward 832 (NU); Umzinto: Ifhfa, Huntley 199 (NU).
This species is very close to M. minutiflora Cogn. from tropical Africa and may
eventually have to be reduced to the latter. For the time being the available material
suggests that it is distinct. It differs from nearly all the other South African species
of the genus by being quite glabrous in all its vegetative parts, the only other glabrous
species {M. marlothii) belongs to the monoecious section Solena.
19
Burtt Davy, l.c. mentions that “A specimen from Witpoortjie Kloof, Krugersdorp,
collected May 1924, may be an immature stage of this The specimen he obviously
referred to is Moss 10657 in (J) which I refer to Trochomeria macrocarpa.
5. M. cordata (TImnb.) Cogn., Mon. Cucurb. 613 (1881); Pflanzenreich 275.1:
114 (1916). Type: in Herb. Thunberg (UPS, photo, in PRE!).
Bryonia cordata Thunb. in Hoffm., Ph^ogr. Blatt. 5 (1803); FI. Cap. 34 (1807).
B. punctata Thunb., B. scabra Thunb., Prodr. PI. Cap. 13 (1794); FI. Cap. 34 (1807).
B. repanda Bl., Bijd. 923 (1825).
Pilogyne velutina Schrad. in Linnaea 12: 412 (1838).
Zehneria scabra Sond. in FI. Cap. 2: 486 (1862); Oliv. in FI. Trop. Afr. 2: 560
(1871).
Melothria punctata (Thunb.) Cogn., Mon. Cucurb. 615 (1881); Pflanzenreich 275.1:
117 (1916), excl. syn. Bryonia angulata Thunb. (q.e. Kedrostis nana); Burtt Davy,
FI. Transv. 1: 225 (1926) et auct. plur., nomen illeg., non M. punctata Rafin. (1836).
M. tomentosa Cogn., op. cit., 614 (1881); 115 (1916). M. membranifolia Cogn. in
Bull. Herb. Boiss. 3: 420 (1895); Pflanzenreich 275.1: 113 (1916). M. velutina
Cogn., op. cit., 613 (1881); 115 (1916) excl. syn. Bryonia laevis Thunb. (q.e. Ceratio-
sicyos spec, vide infra).
Chmber or occasionally prostrate. Roots perennial, fibrous, producing few to
man> annual herbaceous stems which are usually branched. Stems slender, sulcate
when dried, glabrous to more or less densely covered with rather short and usually
stiff hairs, 1-3 m long, occasionally much longer. Tendrils shortly and usually stiffly
pubescent, rarely nearly glabrous when young, glabrescent and older parts often
becoming quite glabrous. Leaves varying from membranous to firmly herbaceous or
rigid, when fresh green to dark green above, paler and in hairy forms sometimes
canescent below, usually drying a dark brown; in outline usually triangular-cordate
but varying from narrowly deltoid with cordate base to suborbicular-cordate, faintly
3-5-angular or sub-trilobed but occasionally deeply 3-5-lobed to a little beyond the
middle, usually acute to acuminate and mucronate at the apex and cordate at the base
with a usually very distinct and rather deep basffl sinus, usually distinctly undulate-
dentate with callus-tipped acute teeth, more rarely only shallowly or very coarsely
so, as a rule on upper surface sparsely setulose and glabrescent, very often retaining
hair-bases as scabrid white dots, on lower surface more densely and more persistently
pubescent with short usually stiff hairs but in some forms pubescence is only retained
on the principal veins and in others the pubescence is very dense to tomentose, canescent
or fulvous; lamina 3-8 cm by 2-5-6 cm, occasionally up to 12 X 9 cm, petioles stiffly
pubescent, usually rather sparsely so, 1-3 cm or occasionally up to 5 cm long. Flowers
dioecious, the male ones subcapitately to subumbellately racemose, the female ones
solitary, fasciculate or in a short raceme. Male flowers up to about 30 (usually 8-20)
per inflorescence; peduncle usually more or less pubescent, sulcate, up to 7 cm
occasionally up to 12 cm long; pedicels slender to filiform, patent, usually more or less
pubescent; 2-8 mm long; receptacle campanulate, 1-5-5 cm high and 2-3-5 cm in
diam., usually more or less pubescent; sepals broadly subulate to triangular-oblong,
usually acute, more or less pubescent, 0-5-1 mm long ; petals 1 - 5-2 - 5 mm long, shortly
pubescent; stamens with long hairs near the middle. Female flowers: common
peduncle, if present up to 6 cm long; pedicels up to 10 mm; ovary subglobose usually
sparsely hairy. Fruits subglobose or ellipsoid, glabrous, smooth or somewhat foveolate,
orange, yellow, red or brown when ripe, 7-12 mm long, 6-10 mm in diam. Seeds
slightly marginate, smooth, about 4 mm long and 2-5 mm broad.
From Abyssinia along the eastern side of Africa down to the Cape Province,
West Tropical Africa, also in the Comores, Madagascar, the Mascarenes and Java.
20
Recorded from the following districts: Cape Province: Cape Peninsula and all
coastal districts; in eastern Cape from Aliwal North, Tarka, Queenstown, Somerset
East and Uitenhage ; eastwards recorded from practically all districts ; Natal: recorded
throughout; Orange Free State: Harrismith, Bethlehem, Fouriesburg, Ficksburg;
Basutoland: many records; Swaziland; Transvaal: Pietersburg, Letaba, Lydenburg,
Pilgrims Rest, Barberton, Ermelo, Brits.
The following specimens are of special interest:
Cape Province. — Without exact locahty: types of Bryonia cordata, B. punctata and
B. scabra (2 sheets) in herb. Thunberg in UPS (photos.! in PRE). East Cape:
Katriviersberge, Ecklon & Zeyher 1788, type number of M. velutina in SAM, S (photo.,
PRE). Alexandria: Olifantshoek, Ecklon & Zeyher 1783 (BOL, L, M, PRE), original
number cited as “ Pilogyne suavis Schrad.”; 1784 (M, SAM), cited by Cogniaux under
M. cordata. Various Drege gatherings (in L) cited by Cogniaux under M. cordata,
M. velutina and M. punctata.
Natal. — Inanda: Wood 748 (in NH), isotype of M. membranifolia Cogn.
An extremely variable species with, accordingly, a great number of synonyms.
The variation is found in the shape of the leaves, from the common ovate-triangular-
cordate shape to almost reniform-cordate (as in the type) or broadly cordate, in the
texture of the leaves from firm and very scabrid to membranous, much smoother (in
the form described as M. membranifolia Cogn.), in the degree of incision of the leaf-
margins from almost entire to coarsely toothed or subserrate or occasionally somewhat
more deeply incised to nearly palmatilobed, in the degree of pubescence, slight in
certain forms such as those described as M. membranifolia but very dense in those
referred to M. tomentosa by Cogniaux and others; in the length of the male peduncles,
size of flowers and degree of development of the female racemes (on a long stalk or
flowers almost fascicled). All these forms intergrade.
As regards the nomenclature of this species, the epithets Bryonia scabra Thunb.
and B. punctata Thunb. 1794 are the oldest validly published and legitimate ones, but
in Melothria they cannot be applied on account of M. scabra Naud. and M. punctata
Rafin., different (American) species, hence the combination M. punctata (Thunb.) Cogn.
for the species under discussion is illegitimate. The next available name is Bryonia
cordata Thunb. (1803), the type of which unfortunately represents a form with a less
common leaf-shape and short sub-fasciculate female inflorescence (whereas the type
of B. punctata represents the most common form). The older name Bryonia angulata
Thunb. (1794) mentioned in the synonymy of “ M. punctata'' by Cogniaux, cannot
be considered at all because the type specimen clearly belongs to Kedrostis nana.
Melothria velutina Cogn. is a hairy form with a slightly different appearance but
I cannot find anything tangible to separate it from M. cordata. Bryonia laevis Thunb.
tentatively placed as a synonym of M. velutina by Cogniaux, appears to belong to
Ceratiosicyos (Achariaceae) and is identical with C. ecklonia Nees. Thunberg’s name
antedates the latter name, so that a new combination is necessary, and is now effected.
Ceratiosicyos laevis (Thunb?) A. Meeuse, comb. nov.
Bryonia laevis Thunb., Prodr. PI. Cap. 13 (1794); FI. Cap. 35 (1807). Type in Herb.
Thunberg (UPS, holo., PRE, photo.!).
Ceratiosicyos ecklonii Nees in Eckl. & Zeyher, Enum. PI. Cap. 281 (1839); Harv.
in FI. Cap. 2: 501 (1862). Type: Ecklon & Zeyher 1797 = 36.10, from eastern Cape
(PRE, iso.!).
21
Doubtful Species
M. natalensis Cogn. in Pflanzenreich 275-1: 96 (1916) is a plant that does not
agree with any other African species of the genus known to me. The type specimen
was destroyed and a duplicate of this gathering {Rudatis 515) could not be traced.
From the description I am inclined to refer it to Trochoineria sagittata (in fact, there
is nothing in the rather meagre description that does not apply to this species), which
certainly occurs in the area where Rudatis 515 was collected.
Probably to be excluded from the South African Flora
M. tridactyla Hook. f. in Oliv., FI. Trop. Afr. 2: 562 (1871). Cogniaux (1916)
cites a specimen from East London (leg. O. Kuntze) which I have not seen. The nearest
localities of M. tridactyla that have been recorded are in the southern part of Portuguese
East Africa, so that it is most unlikely that this tropical species would occur near East
London and not have been recorded from elsewhere in the Union. I am at a loss to
suggest the true identity of Kuntze’s specimen; it might not even be a Melothria and
it is not at all unlikely that it is Trochoineria sagittata, whose vegetative parts resemble
some species of Melothria sect. Melothria.
Excluded Species of Melothria
M. hederacea (Sond.) Cogn. (Zehneria hederacea Sond.) = Kedrostis nana (Lam.)
Cogn., see p. 30.
5. OREOSYCE
Oreosyce Hook. f. in FI. Trop. Afr. 2: 548 (1871), (“ Oreosycios ”, sphalm. in clavis,
op. cit. 522); Cogn., Mon. Cucurb. 564 (1881); Baill., Hist. PI. 8: 449 (1886); Pax
in Pflanzenfem. 4, 5: 15 (1889); Cogn. in Pflanzenreich 275.1: 67 (1916).
Cucumis Auct., pro parte, exclus. type.
Hymenosicyos Chiov. in Ann. di Bot. 9: 62 (1911); Harms in Notizbl. Bot. Gart.
Berlin 8: 485 (1923); Pflanzenreich 275 .2: 157 (1924).
Scandent herbaceous plants with usually slender stems, often hispid or sparsely
to densely covered with short rather stiff hairs. Leaves petiolate, undivided to more
or less distinctly palmately 3-5-angled, -lobed or -fid. Tendrils simple. Flowers
monoecious, small, yellow, pedicellate; male flowers solitary or fascicled; receptacle
cylindric-campanulate, hispid; sepals dentiform or subulate; corolla rotate to cam-
panulate with oblong segments; stamens 3, subsessile in the middle of the receptacle,
dorsally attached in their middle; anthers slightly cohering or free, linear or oblong,
two 2-thecous and the third 1-thecous.; thecae straight, obtuse; connective narrow,
usually produced at the apex as a small tooth; rudiment of pistil gland-like; female
flowers solitary, usually borne in different axils from those bearing the male ones;
calyx and corolla as in the male; staminodes 3, small or sometimes well-developed
and apparently fertile stamens present; ovary fusiform, oblong or ellipsoid, setulose,
with 3 placentae and numerous horizontal ovules; style columnar, surrounded by a
disc at the base; stigma 3-lobed, the lobes oblong, usually fringed. Fruit rather small
indehiscent, ellipsoid or ovoid, covered with hairy or bristly protuberances intermingled
with long setae, many-seeded. Seeds much compressed, ovate in outline, margined.
Type species: O. africana Hook. f. (the genus was monotypic when described).
Found in tropical Africa, extending into Angola, the Transvaal, Abyssinia and
Somaliland.
22
Oreosyce seems to be a perfectly natural genus, although its author did not realise
that several species he described himself in Cucumis belong here. These species differ
from true Cucumis species in the first place in the straight anther-thecae. It is an
established fact that straight or slightly curved anthers occur in certain subfamilies of
the Cucurbitaceae such as the Melothrieae whereas sigmoid conduplicate thecae are
characteristic of the Cucurbiteae. The fruit of Oreosyce bears both protuberances
and setae, and it contains marginate seeds whereas the fruits of Cucumis are smooth
or hairy or covered with protuberances, but not covered with both soft spines and setae,
and contain emarginate seeds, facts which were noticed by Hooker f. and is evident
from his remarks sub “ Cucumis ” in FI. Trop. Afr. 2: 542. Probably owing to the fact
that of Oreosyce africana no fruits were known at that time. Hooker overlooked the
connection of these “ aberrant ” species of Cucumis with the Melothrieae and with
Oreosyce. As Hooker attributed much importance to the presence or absence of an
apical extension of the connective and Oreosyce as well as the “ aberrant ” species
of Cucumis have, apart from the straight thecae, a small apical appendage of the
connective or none at all, whereas Cucumis proper has a broad and conspicuous
appendage, it is strange that he did not associate these “ aberrant ” species with
Oreosyce.
In 1911 Chiovenda described a genus Hymenosicyos, based on one of Hooker’s
“ aberrant ” species of Cucumis (C. membranifolius Hook. f.). Chiovenda’s work was
critically discussed by R. E. Fries in Wiss. Ergebn. Schwed. Rhodesia-Kongo-Exped.
1911-12, Bot. 1, 2: M0(1916). Fries pointed out that there were some discrepancies
between his own observations and the characters mentioned by Chiovenda and
ventured the opinion that Hymenosicyos Chiov. is very closely related to Oreosyce
Hook. f. However, he did not go so far as actually to reduce the former to the older
genus Oreosyce. Harms (op. cit., 1923, p. 485) did not add much evidence to Fries’s
observations. He agreed that Hymenosicyos hardly differs from Oreosyce, but instead
of taking the final step and sinking Chiovenda’s genus, he transferred a few species
described by Cogniaux in Oreosyce to Hymenosicyos. It is true that in these species
the connective is produced at the apex, whereas in a syntype specimen of Oreosyce
{Mann 2025, K) this is apparently not the case, but the appendage of the connective
is never so obvious to be sufficient for a generic distinction. Mr. Killick informed
me that the specimen Boughey 176 from the same locality as Mann 2025 and an exact
match of it has the connective slightly produced above the anthers as in “ Hymenosicyos'" .
Cogniaux was apparently unaware of the relation between Hymenosicyos (which he
retained in Cucumis) and Oreosyce, because he redescribed Cucumis membranifolius
Hook. f. again as Oreosyce villosa and O. triangularis.
There is very little doubt that Hymenosicyos is nothing but a synonym of Oreosyce,
and all the species referred to the former genus by Chiovenda and by Harms have
to be transferred to or retained in Oreosyce. The re-descriptions of the same species
in the various genera involved have naturally caused a considerable synonymy. If
all the species described as Cucumis, Hymenosicyos, or Oreosyce that are congeneric
are sorted out, the total number of valid species of Oreosyce will most probably not
exceed 4 or 5. One species, which is wide-spread, extends the range of the genus into
the Transvaal. In view of the synonymy involved a number of synonyms relating to
and specimens hailing from tropical Africa are cited.
O. subsericea {Hook. /.) A. Meeuse, comb. nov.
Cucumis subsericeus Hook. f. in FI. Trop. Afr. 2: 545 (1871); Cogn., op. cit. 506
(1881); 153 (1924). Type: Welwitsch 838, Pungo Andongo, Angola (K, holo.).
C. membranifolius Hook, f., I.c. 545 (1871); Cogn., l.c. 506 (1881); 153 (1924). C.
cecili N.E. Br. in Kew Bull. 1906: 104 (1906); Cogn., l.c. 148 (1924).
Oreosyce triangularis Cogn. in Bot. Jahrb. 21: 207 (1895); Pflanzenreich 275.1: 68
(1916); Burtt Davy, FI. Transv. 1: 226 (1926), as to name only. O. villosa Cogn.,
op. cit. 68 (1916). O. aspera Cogn. and O. parvifolia Cogn., op cit. 268 (1916).
23
O. bequaertii De Wild, in Revue Zool. Afr. 9, 3: 90 (1921); PL Bequaert4: 558 (1922).
Hymenosicyos membranifolius (Hook, f.) Chiov. in Ann. di Bot. 9: 63 (1911); R. E.
Fries in Wiss. Ergebn. Schwed. Rhod. Kongo Exped. 1911-12, Bot. 1, 2; 310 (1916);
Harms in Notizbl. Bot. Gart. Bed. 8: 486 (1923); Pflanzenreich 275.2: 157 (1924).
H. subsericeus (Hook, f.) Harms, op. cit. 487 (1923); 158 (1924). H. bequaerti (De
Wild.) Harms, op. cit. 159 (1924); Robyns, Spermat. Nat' Parc Albert 2: 393 (1947).
H. villosus (Cogn.) Harms, H. triangularis (Cogn.) Harms, op. cit. 487 (1923); 159
(1924). H. bryoniifolius Merxm. in Mitt. Bot. Staatssamml. Munchen H 6; 205
(1953).
Stems usually rather slender, longitudinally sulcate-striate, at first shortly setose-
hirsute on the ribs with bulbous-based hairs, glabrescent, the hair-bases often persistent,
rendering the stems finely scabrid or scabrid-muriculate. Leaves thinly herbaceous
varying from distinctly 3-5-lobed to triangular with cordate base, with sub-acute to
acuminate apex, the lobes, if present, rounded when small, acute or acuminate when
conspicuous; the margin sub-entire to sinuous or somewhat dentate, usually minutely
and rather remotely calloso-ciliate ; the blade 2-8 cm long and 1 • 5-7 • 5 cm broad near
the base; on both surfaces with fine stiff" straight and somewhat bulbous-based adpressed
hairs, more densely so beneath, sometimes ultimately becoming finely scabrid if hairs
fall off, but usually pubescence persistent; petioles retrorsely setose-hispid, glabrescent
and becoming scabrid, 1-12 cm long. Tendrils usually somewhat setulose with bulbous-
based hairs. Male flowers: pedicels slender, 0-5-2 cm long, setulose; receptacle
2-5-4 (-5) mm long, more or less densely setulose-hirtellous as are the 1-3 mm long
subulate-linear sepals; corolla yellow; the lobes usually shortly hairy mainly on the
nerves outside, 4-8 mm long. Female flowers usually slightly larger than the male ones
on the same plant, on usually somewhat incrassate peduncles 0-5-4 cm long; ovary
ovoid or elliptic, densely bristly. Fruit yellow or orange when ripe, broadly ellipsoid
or ovoid; in the specimens seen up to 18 mm long and 12 mm in diam. Seeds elliptic-
oblong, attenuate at the one end, 5-6 mm long and 2-3 mm broad.
Found in Africa south of the Sahara, southwards to Angola, S. Rhodesia and the
northern Transvaal, also in Abyssinia.
Transvaal. — Soutpansberg ; near Louis Trichardt, Blenkiron & Young in Hb. Moss
No. 14537 (J, PRE). Letaba: Shilovane, Junod 59 (PRE).
Southern Rhodesia. — Salisbury: near Salisbury, Eyles 4763 (SRGH). Makoni:
Rusape, Dehn R 25/52 (Type of Hymenosicyos bryoniifolius, M, isotype SRGH).
Umtali: Chase Alll; 5214 (SRGH); 6042; 6222 (SRGH, PRE). Urungwe: Msukwe
River, Wild 4209 (SRGH); Distr.? Stapleford, Gilliland 378 (SRGH).
Nyasaland. — Zomba Plateau: Zomba, Brass 16315 (SRGH).
Tanganyika. — ^Nyasa. : Kyimbila, Stolz 335 (L, M). Kilimanjaro: Namui River,
Endlich 291 (M). Moschi: Wallace 1072 (EA, PRE). Usambara: Mlalo, Holst 630
(type gathering of Oreosyce triangularis, fragment of type in BR).
Belgian Congo. — Ruwenzori : Kisuki, Bequaert 4703 (type of Hymenosicyos bequaertii,
BR, holo.); Tshibinda (W. of Lake Kivu), Humbert 7385 (PRE).
As Mr. Killick has pointed out, there is a difference between the type species
Oreosyce africana and the specimens cited above in the shape of the leaves which are
much more acuminate in O. africana. As the fruit of the type species is apparently
still unknown (see Keay, FI. W. Trop. Afr. ed. 2: 210) and may be different from
that of O. subsericea, it seems better not to take up the name O. africana for the cited
specimens.
24
Hooker published Cucumis subsericeus and C. membranifolius on the same page,
but in this order, and the first name was chosen for the new combination made in
Oreosyce. O. triangularis Cogn. is based on Holst 630 which differs slightly in the
leaf-shape, but is otherwise not essentially different. However, the specimen Junod
2622 (in Z!), cited by Cogniaux under this name in Pflanzenreich is not this species
at all, but Melothria maderaspatana. Burtt Davy cites “ Jun. 2622 fide Cogn.” in his
FI. Transv. under O. triangularis and thus repeated Cogniaux’s error. Ironically, the
species does occur in the Transvaal after all.
6. KEDROSTIS
Kedrostis Medik., Phil. Bot. 2: 69 (1791); Cogn., Mon. Cucurb. 632 (1881); Pflan-
zenreich 275.1; 138 (1916); Pax in Pflanzenfam. 4, 5: 17 (1889); Phillips, Gen. ed.
2: 746 (1951).
Coniandra Schrad. in Eckl. & Zeyh., Enum. PI. Afr. Austr. 2: 216 (1836) et Reliq.
in Linnaea 12: 403 (1838); Sond. in El. Cap. 2: 483 (1862).
Cyrtonema Schrad., op. cit. 276 (1836) and 403 (1838).
Rhynchocarpa Schrad., op. cit. 403 (1838); Naud. in Ann. Sci. Nat., 4me. ser., 12:
146 (1859); Benth. & Hook, f., Gen. PI. 1: 531 (1897); Hook. f. FI. Trop. Afr. 2:
563 (1871).
Pisosperma Sond. in FI. Cap. 2: 498 (1862), Benth. & Hook, f., op. cit., 831; Harv.,
Gen. S. Afr. PI. ed. 2: 126 (1868); Cogn., Mon. Cucurb. 631 (1889) and Pflanzenreich
275.1: 136 (1916); Pax, op. cit., 17; Phillips, op. cit., 746.
Toxanthera Hook. f. in Hook., Icon. PI. 15, t.l421 (1883); Pax, op. cit., 17; Cogn.,
op. cit. 137 (1916); Phillips, op. cit., 746.
Type species: Bryonia africana L. (cited by Medikus) = K. africana (L.) Cogn.
Perennials, often with tuberous rootstock forming annual herbaceous prostrate or
scandent stems, often foetid with a smell resembling carbon disulphide. Leaves various,
entire, lobed or more or less deeply palmately or sometimes pinnately dissected, usually
distinctly petiolate. Tendrils simple or occasionally (in the subgenus Toxanthera)
bifid. Flowers monoecious, rarely dioecious, often small to minute, whitish, yellowish
or greensih. Male flowers racemose, subcorymbose or subumbellate on a usually
distinct common peduncle; pedicels often slender; receptacle varying from widely
and shortly campanulate to cyhndric-campanulate; sepals varying from ovate to
linear-lanceolate, often pubescent or glandular-pubescent outside; corolla-lobes ovate,
elliptic or oblong, often pubescent, papillose or glandular outside; stamens 3, rarely
4 or 5, inserted in the receptacle; filaments free, usually short; anthers cohering,
usually one 1-thecous and two 2-thecous (when 5 stamens present all monothecous) ;
thecae straight or more or less curved; connective often produced, entire, bifid or
bi-partite; rudiment of pistil 0 or sometimes present, small, glandlike, very rarely
distinct. Female flowers solitary or occasionally fascicled or clustered, usually shortly
pedicelled to sessile; receptacle, calyx and corolla as in the male flowers; staminodes
0-5, often 3, usually small or linear or filiform; ovary ovoid to subglobose or elliptic;
sometimes (mainly in the subgenus Toxanthera) oblong or narrowing into a beak,
with 3 or sometimes 2 placentas and few or (in the subgenus Toxanthera) rather many
ovules; style elongated or short, columnar, without or with an indistinct disc at the
base; stigmas 2 or 3 or cohering and apparently simple, often thick and each usually
more or less shortly bilobed. Fruit indehiscent, usually smooth, fleshy, small, ovoid
or subglobose or (in the subgenus Toxanthera) rather large, often beaked, few- to
many seeded, usually red when ripe. Seeds globose to ellipsoid, often margined;
testa smooth, crustaceous; tegmen thin, membranous; cotyledons thick and fleshy,
radicle short conical.
25
Mainly an African genus, but extending to India. Cogniaux, in his monographic
treatment of the genera Kedwstis, Toxanthera and Pisosperma recognised 33, 3 and 1
species respectively. One can safely assume that, after a critical re-examination, the
number of species, even after the inclusion of Toxanthera and Pisosperma, will prove
to be considerably less (about 20).
The characters used to distinguish Toxanthera and Pisosperma from Kedrostis
break down altogether. The presence or absence of a produced apical part of the
connective is not a reliable character. A study of several authentic specimens makes
it clear that specimens that are identical with “ Pisosperma capense ” Sond. have been
referred to a variety of Kedrostis zeyheri by Sonder himself and also by Cogniaux and
that specimens which are undoubtedly conspecific with Toxanthera natalensis and
T. lugardae have been described as species of Kedrostis by Cogniaux and others and
in fact agree altogether with the section Cogniauxiana of Kedrostis. The natural
solution is to reduce these genera to Kedrostis. It is true that the species referable to
the section Cogniauxina O. Ktze differ in several respects slightly from typical species
of Kedrostis (the fruits are much larger, the tendrils often bifid), but these differences
are not important enough for generic distinction. The proposed reductions and the
Code of Botanical Nomenclature necessitate the following changes in the nomenclature
of the sub-genera:
1. Kedrostis Medik. subgenus Kedrostis (= Typokedrostis Cogn.). This includes
Pisosperma Sond.
2. Kedrostis Medik. subgenus Toxanthera {Hook.f.) A. Meeuse, stat. nov. Toxan-
thera Hook, f., l.c. (pro gen.). Kedrostis sect. Cogniauxina O. Ktze. in Post & O.
Ktze., Lex. Gen. Phaner. 107 (1903); Cogn. in Pflanzenreich 275 .1 ; 150 (1916). Type
species: Toxanthera natalensis Hook. f. = K. natalensis (Hook, f.) A. Meeuse. Dis-
tinguishing characters: Male flowers racemose, connective not produced at the apex
(often produced in the subgenus Kedrostis). Ovary fusiform or long-cylindric. Fruit
large, oblong-fusiform, smooth, many-seeded. Tendrils usually bifid.
3. Kedrostis Medik. subgenus Gilgina Cogn., op. cit. 155 (1916). This subgenus,
not represented in Southern Africa, is characterised by fascicled male flowers, a large
ovoid fruit covered with thick soft spines, compressed seeds and spinescent tendrils.
Tendrils simple:
Leaves undivided, softly herbaceous drying thin (and often papery); plants softly hairy, monoe-
cious, usually bearing the minute male and female flowers in the same axils, the male
ones on very slender pedicels in short, few-flowered and often contracted racemes, the
female ones sessile; ovary globose, hairy; fruit sessile, globose, with a small apiculus or
(in S. Africa) more rarely long-rostrate 1. K. foetidissima
Leaves more or less distinctly lobed or dissected, or if undivided, plants not as above:
Flowers appearing before the leaves 2. K. capensis
Flowers appearing when leaves are present:
Dioecious* 3. K. nana
Monoecious*:
Male and female flowers borne in the same axils, minute:
Ovary globose, fruit spherical with an abrupt short apiculus or without any apiculus;
male raceme on a distinct common peduncle; leaves herbaceous 4. A', africaita
Ovary more or less elongate; fruit long-rostrate; male raceme reduced so that flowers
appear almost fascicled; leaves somewhat succulent 5. K. crassirostrata
Male and female flowers borne in different axils:
Male flowers with papillose suberect corolla-lobes; ovary ovoid; fruit ovoid to sub-
globose up to 3 cm long; vegetative parts usually covered with soft white curved
hairs; leaves usually deeply dissected with narrow, often toothed pinnatilobed
segments, up to 7 cm long 2. K. capensis
*As the male flowers are usually protandrous, some specimens appear unisexual and suggest
dioecy. If such a specimen does not agree with K. nana, the other part of the key should be tried.
{K. nana is a climber, mainly in scrub and on bushes, occurring in the coastal and some adjoining
districts from the Cape Peninsula to the southern half of Natal, the other species are either manifestly
monoecious or do not occur in the same region).
26
Male flowers with usually spreading corolla-lobes which are often hairy but not distinctly
papillose; ovary linear-fusiform; fruit oblong-fusiform, 4—8 cm long; vegetative
parts usually hairy but not with soft white curved hairs; leaves, if dissected, not
more deeply so than a little beyond the middle and lobes broad, finely denticulate
along the margin, blade up to 16 cm long see 6 Ai. natalensis
Tendrils bifid 6. K. natalensis
1. K. foetidissima (Jacq) Cogn., Mon. Cucurb. 634 (1881); Pflanzenreich 275 .1 ;
140 (1916); Dinter in Fedde, Report. 18: 434 (1922); Burtt Davy, FI. Transv. 1: 226
(1926); Hutch. & Dalz., FI. W. Trop. Afr. 1: 179 (1931); Andrews, Flow. PI. Anglo-
Egypt. Soudan 1: 175 (1950).
Trichosanthes foetidissima Jacq., Collect. 2: 341 (1788); Icon. Rar. 3: t.624 (1793).
Melothria foetida Desr. in Lam., Encycl. Method. 4: 87 (1797); Ser. in DC., Prodr.
3: 313 (1828).
Rhynchocarpa foetida (Desr.) Schrad. in Linnaea 12: 403 (1828); Naud. in Ann. Sci.
Nat., 4me. ser. 12: 146 (1859), 16: 176 (1862); Hook, f., FI. Trop. Afr. 2: 564
(1871).
Type: not seen; the plate in Icon. Rar. 3, t.624 was taken to be representative.
Perennial with tuberous rootstock. Stems annual, herbaceous, climbing (or
occasionally prostrate when no support is available), usually slender, often branched,
reaching 1-2 m, at first long and softly pilose, or sometimes shortly pubescent,
occasionally glabrous, hairy ones glabrescent. Leaves herbaceous, varying from ovate-
cordate to sub-orbicular-cordate or sub-triangular-cordate, undivided or faintly 3-5-lobed
with short obtuse lobes ; the apex varying from obtuse to shortly cuspidate, the cordate
base usually with a narrow and fairly deep sinus, the margin more or less distinctly
calloso-denticulate or undulate to dentate or entire ; both surfaces usually with rather
long, soft to very short stiff hairs turning scabrid especially so above; petioles usually
short, often shortly and densely hairy. Flowers monoecious. Male flowers in short
1-7-flowered racemes; common peduncle usually very slender, filiform, pubescent,
up to 3 cm long; pedicels very slender, usually erecto-patent, up to about 1 cm long;
calyx hairy, rarely glabrous ; petals more or less hairy or papillose-hairy, 2-5 mm long,
rarel}' only about 1 -5 mm long. Female flowers solitary in the same axils as the male
racemes, sessile or sub-sessile (pedicels up to about 5 mm long, shortly hairy); ovary
subglobose, shortly or long-beaked, softly pilose, 2-3 mm in diam. Fruit red when
ripe, few-seeded, covered with long patent hairs usually thinly so, rarely quite glabrous.
Seeds often 4, ovoid, somewhat compressed, distinctly margined, margin usually shortly
winged, truncate or notched on the attenuate end, 4-6 mm x 3-4 mm x 1-5-2 mm;
testa finely granulated or nearly smooth.
Africa south of the Sahara to South Africa, Eritrea and Somaliland. Kedrostis
rostrata (Rottl.) Cogn., op. cit. 636 (1881) from India can most probably not be
separated from this species.
This plant occurs in two different forms which are here treated as sub-species,
because the only constant and important difference is in the shape of the fruits; those
of the typical subspecies are long-rostrate and those of the other subspecies are
subglobose, abruptly and very shortly rostrate or erostrate.
{a) K. foetidissima subsp. foetidissima.
K. foetidissima var. perrottetiana (Ser.) Cogn., op. cit., 635 (1881) and 141 (1916), var.
genuina Cogn., l.c., and var. divergens (Hochst.) Cogn., l.c. (cum syn.).
Leaves 1-5-12 cm long, 0-5-9 cm wide, subsessile or petioles up to 6 cm long;
basal sinus varying from rounded, semi-orbicular about 0-5 cm deep to narrow, up to
2 cm deep, or to more or less rectangular, up to 2 x 2 cm. Fruits ovoid, 15-22 mm long
with a 10-15 mm long sublinear or narrowly conical straight or falcate rostrum, length-
wise marked with thick nerves.
27
Distribution as the species, but rare in southern Africa.
Transvaal. — Soutpansberg, between Chipise and the Nuanetzi River; Gerstner 6072
(BOL, PRE).
Cogniaux (1916) also records it (as var. genuina) from' Lourengo Marques but
the cited specimen {Schlechter 1 1678) as represented in BOL and GRA, is the subspecies
obtusiloba (see below). Another gathering cited by Cogniaux as the var. genuina, viz.
Genard 375 = McKen s.n., most probably also belongs to the subsp. obtusiloba because
all specimens from Natal and Zululand belong to this subspecies as far as can be
ascertained.
{b) K. foetidissima subsp. obtusiloba (Sond.) A. Meeuse, stat. nov.
Zehneria obtusiloba E. Mey. ex Sond. FI. Cap. 2: 487 (1862).
Bryonia obtusiloba E. Mey. ex Drege, Zw. Pflzgeogr. Doc. 156, 169 (1843), nomen
tantum.
Melothria obtusiloba (Sond.) Cogn. op. cit. 616 (1881).
Kedrostis foetidissima var. microcarpa Cogn., op. cit. 635 (1881) and 141 (1916); Burtt
Davy, l.c. K. minutiflora Cogn. in Bull. Herb. Boiss. 2me. ser. 1: 884 (1901), and
in Pflanzenreich 275.1: 144 (1916); Dinter in Fedde, Repert. 18: 434 (1922). K.
obtusiloba (Sond.) Cogn. in Pflanzenreich 275.1: 143 (1916); Burtt Davy l.c.
Type: A specimen collected by Drege near the Umzimkulu River in the eastern
Cape Province or southern Natal in herb. Sonder (now in S). Sonder cited several
specimens in his description in Flora Capensis, but as he took up the manuscript name
{Bryonia) obtusiloba, given by E. Meyer, Drege’s gathering automatically becomes
the type.
Found in Angola, Southern Rhodesia, Portuguese East Africa and South Africa.
Recorded from the following districts: South West Africa: Ovamboland, Okavango,
Outjo, Grootfontein, Okahandja, Otjiwarongo, Windhoek; Transvaal: Soutpansberg,
Sibasa, Letaba, Pietersburg, Potgietersrust, Waterberg, Pretoria, Bronkhorstspruit,
Brits, Lydenburg, Carolina, Marico; Natal: Ingwavuma, Nongoma, Weenen, Estcourt,
Umvoti, Impendhle, Camperdown, South Coast; Cape Province: Vryburg, Kuruman,
Barkly West, Hay, Kimberley.
Some interesting specimens are the following:
South West Africa. — Okahandja: Dinter 428 (cited by Cogniaux 1916 as K. minuti-
flora Cogn., SAM, PRE); Outjo, Etero: Dinter 1444 (holotype of K. minutiflora, Z);
Windhoek, Auas Mts., Northern foothills: Dinter 1895 (cited by Cogniaux 1916 as
K. obtusiloba).
Transvaal. — Magalakwin River (Prob. Waterberg distr.): Schlechter 4275 (BOL,
GRA), cited by Cogniaux 1916 and Burtt Davy as K. obtusiloba', Pretoria or Bronk-
horstspruit distr.: Rehmann 4810 (Z) cited by Cogniaux 1916 as K. foetidissima var.
microcarpa Cogn.
The type gathering of Zehneria obtusiloba Sond. was not studied, but from other
specimens referred here by Cogniaux {Schlechter 4275, Dinter 1895) it is quite clear
that this species cannot be separated from K. minutiflora Cogn. (of which the type,
Dinter 1444, and two specimens of Dinter 428, referred to this species by Cogniaux
were available for study) nor from K. foetidissima var. microcarpa Cogn. The
differences mentioned by Cogniaux (1916) are not satisfactory, at any rate, in his key
(on p. 139) every specimen with subglobose fruits without a long rostrum would key
out as K. obtusiloba or K. minutiflora, but never as K. foetidissima and one would not
arrive at the non-rostrate varieties of the latter. In addition, Cogniaux’s description
of K. minutiflora as “ glaberrima ” does not hold true; the specimen Dinter 1444 in
Z, apparently the holotype, is not quite glabrous on stems, receptacle and fruits as
indicated in the key on p. 139 of Cogniaux's 1916 monograph.
28
2. K. capensis (Sond.) A. Meeuse, comb. nov.
Pisosperma capense Sond. in FI. Cap. 2: 498 (1862); Cogn., Mon. Cucurb. 632 (1881);
Pflanzenreich 275.2: 136 (1916); Phillips, FI. Basutol. [Ann. S. Afr. Mus. 16: 101
(1917)].
Perennial. Stems annually produced from a tuberous subglobose or napiform
up to 10 cm thick rootstock, when young suberect and short, but later prostrate and
attaining a length of not more than 40-50 cm, herbaceous, usually rather slender,
angular-striate, covered like all vegetative parts with short white curved hairs, very
rarely quite glabrous. Leaves usually secund, almost invariably firm, greyish green,
ovate to oblong in outline, the first-formed undivided to pinnatilobed or palmately
lacerated with a truncate-hastate to broadly and shallowly cordate base, up to about
4 cm long and 3 cm wide, but later usually very deeply palmatifid with 5-7 oblong,
lanceolate or linear, entire, toothed or occasionally pinnatilobed segments, up to 7 cm
long and about 5 cm wide, the segments often linear (if so 2-3 mm wide); the lateral
ones shorter than the central ones, the lowermost usually much smaller often bilobed,
the typical pubescence of curved white hairs usually confined to the margin on upper
surface, but thinly dispersed on the lower and usually slightly paler lower surface;
petioles 4-15 mm long. Tendrils reduced or short or wanting. Flowers monoecious,
sometimes appearing before the leaves, the male ones in (sometimes short or contracted)
racemes or fascicled (by reduction rarely solitary), the female ones solitary in different
axils. Male flowers: common peduncle often rather stout, up to 6 cm but usually
under 4 cm long, few-flowered or occasionally up to about 20-flowered; more or less
hairy to subglabrous or hairy towards the apex; pedicels filiform, very slightly thickened
towards the apex, shortly hairy or puberulous, up to 3 cm long, usually with subulate-
filiform bracteoles at the base ; receptacle 3-6 mm long, shortly hairy like the pedicels
or rarely subglabrous, sepals triangular or triangular-ovate, sub-acute to acute, erect
or erecto-patent, 1-2 (-3) mm long, usually hairy; petals pale yellow or greenish yellow,
ovate, oblong, obtuse to subacute, 3-6 (-8) mm long, densely papillose. Female flowers:
pedicels short, in fruit not exceeding 1 cm; ovary ovoid, usually hairy; staminodes
3-5, linear or ligulate. Fruit subglobose to ovoid or oblong, shortly rostrate, ultimately
glabrous, 1-5-2 -5 (-3) cm long and 1-1-5 cm in diam., few- (to 10-)seeded. Seeds
thick, about 5-6 X 4-5 x 3-4 mm.
Type specimen: Sonder cited several gatherings (Zeylier, Drege, Barber). The
only original specimens of which one can be sure that Cogniaux has also studied them,
are the specimens Zeyher Cucurb. No. 1 and Zeyher No. 593, in the Kew herbarium.
A specimen in SAM collected by Zeyher has no original number; a number in pencil
(285) does not agree with the cited number (593), but the locality is Caledon River
and it is almost certainly a duplicate of one of the original gatherings cited by Sonder.
South We>t Africa. — Keetmanshoop: Great Karasberg, Narudas Slid, Pearson
8523; 8524; 8525 (BOL).
Cape Province. — Little Namaqualand: Goodhouse, Marloth 14046 (PRE); grown
from seed collected in the Richtersveld, Doom River: Herre Hb. no. 25599 (BOL);
Vaalheuwel, Acocks 19423 (PRE); Spektakel, Bolus 9509 (BOL). Clanwilliam:
Lamberts Bay, Van Pullensvlei, Pole Evans & van Nouhuys 32. Ceres: Karroo Poort,
Leipoldt 3462; 5712 (BOL). Laingsburg: Whitehall, Compton 3012; 4437 (BOL);
8638; 13927 (NBG). Prince Albert: Zwartberg Pass, Bond 847 (NBG). Beaufort
West: Kromrivier, Goatcher Hb. no. 1016 (BOL); Nelspoort, Courland’s Kloof,
Pearson 1470 (SAM); Rhenosterkop, Burke 143 (SAM). Willowmore: Kommando-
kraal, Zeyher 861 (SAM). Graaff Reinet: Bolus 296 (BOL). Cradock: Acocks
16317 (BOL, PRE). Middelburg: Wapadsberg Plateau, 16214 (PRE). Tarka:
Tarkastad, Wilmot s.n. Bedford: near Cookhouse, Acocks 11910 (PRE). Fort
Beaufort: near Blinkwater, S/ory 1676 (PRE). Somerset East or Uitenhage: Zuurberg,
Holland 296 (GRA). Molteno/Sterkstroom: Andriesberg, Galpin 2181 (GRA, PRE,
29
BOL). Aliwal North; F. Bolus 207 (BOL). Komgha: Flanagan s.n. (SAM). Hay:
Niekerk’s Hoop, Wilman 1410 (KMG, BOL); 2398; 2399 (KMG); Rietkloof, Acocks
8555 (KMG); Blaauwbosputs, Acocks 2009 (KMG); Floradale, Ferrar Hb. no. 6145
(KMG); Esterhuysen 2308 (BOL, PRE). Barkly West: Geluk, Acocks 1469 (KMG).
Orange Free State. — Kroonstad: Moss 7434 (J); Pont 363a (PRE); 365c (PRE,
SRGH). Bloemfontein: Potts 499 (BOL); Mostert 355 (PRE). Without precise
locality, Caledon River, Zeyher 285 (?) in SAM ; Bloemfontein or Fauresmith: Rehmann
3679 (Z).
Basutoland. — Leribe, Thaba Phatsoa, Dieterlen 194a; 194b (PRE).
Apart from the fact that the Zeyher specimen from the Caledon River in SAM
is most probably a duplicate of an original specimen, the identity of this species can
easily be estabhshed from other specimens cited by Cogniaux (1916), such as Rehmann
3679 and Galpin 2181.
K. capensis is much more variable than one would expect from Sonder’s or
Cogniaux’s descriptions. Sonder laid a great deal of stress on the supposed precocious
nature of the flowers, but this phenomenon is incidental and certainly not a generic
character. Cogniaux (1916, p. 132) keys out the genus Pisosperma by using the absence
or presence of an apical connective appendix as a main distinguishing character, but
this is fallacious and this author has cited specimens belonging to K. capensis under
species of Kedrostis instead of under Pisosperma. These specimens with well developed
leaves appear for instance as K. zeyheri var. angustiloba (Sond.) Cogn. (op. cit. 1881,
p. 642, 1916, p. 148), viz. Burke 143 from Rhenosterkop (already referred to Coniandra
zeyheri var. angustiloba by Sonder, 1862, p. 485), or as K. zeyheri, viz. Bolus 296 from
Graaff Reinet (Cogn. op. cit. 1881, p. 642, 1916, p. 148). The specimen Bolus 296
in BOL has an annotation in Dr. H. Bolus’s handwriting; “ Coniandra digitata
Sond. fide Dr. Sonder”. Cogniaux (op. cit. 1881 and 1916) cites Burke 143 under
K. digitata and the same number, as was stated above, under K. zeyheri var. angustiloba.
These are additional examples of the confusion that existed in the genus Kedrostis
(s.l.), some of which can be traced back to Sonder’s treatment in Flora Capensis, but
are mostly due to Cogniaux’s not sufficiently critical treatments of the genus and the
allied genera Toxanthera and Pisosperma (here reduced to the synonymy of Kedrostis).
K. capensis can always be distinguished from related species of Kedrostis by the
following characters:
1. The monoecious flowers ^dioecious in forms of K. zeyheri).
2. The typical hairs which are never quite absent from the leaves (the leaves are
quite glabrous or at least scabrid or white-punctate in K. zeyheri and other related
species).
3. The size of the flowers (which are larger than those of some related monoecious
species) which are, in addition, racemes or fascicled, not subumbellate as in e.g.
K. africana.
3. K. nana {Lam.) Cogn., aggregate species.
This is a complex separable into three varieties:
Leaves broadly ovate to reniform-cordate, entire or angular to lobed (but not beyond the middle);
lobes broad, obtuse, rarely acute, entire or rarely dentate or faintly 3-lobed var. nana
Leaves deeply palmatilobed to palmatisect; lobes usually acute, often dentate, lobulate or dissected,
or lobes narrow, ligulate, rounded:
Leaves usually smooth above (rarely white-punctate-scabrid); segments of leaves usually more
or less rhomboid or cuneate, often coarsely and acutely dentate var. zeyheri
Leaves usually scabrid on upper or on both surfaces with white pustules; segments of leaves
variable, usually more or less ligulate, sometimes again lobed with ligulate lobes var. schlechteri
1517626-2
30
(a) K. nana {Lam.) Cogn. var. nana.
Bryonia nana Lam., Encycl. 1; 497 (1783). B. triloba Thunb. Prodr. PI. Cap. 13
(1794), non Lour. (1790). B. africana Thunb., l.c., non L. Bangulata Thunb.
Kedrostis mollis (Kunze) Cogn., Mon. Cucurb. 637 (1881). K. velutina Cogn. in
Vjschr. Naturf. Ges. Zurich. 53: 492 (1908); Pflanzenreich, 143 (1916). K. angulata
(Berg.) Fourcade in Trans. Roy. Soc. S. Afr. 21: 91 (1932), nomen illeg.
Sicyos angulata Berg., Descr. PI. Cap. 352 (1767), non L. (1753).
Cyrtonema triloba (Thunb.) Schrad. apud Eckl. & Zeyh., Enum. 276 (1834); Linnaea
12: 403, 404 (1838). C. mode Kunze in Linnaea 20: 49 (1847).
Zehneria hederacea Sond. in FI. Cap. 2: 487 (1862).
Coniandra thunbergii Sond., l.c. 484. C. mode (Kunze) Sond., l.c. 485 (1862).
Melothria hederacea (Sond.) Cogn., Mon. Cucurb. 611 (1881); Pflanzenreich 107
0916).
Type: A sterile specimen in Herb. Lamarck (P), grown in the botanical gardens
of Paris. Mr. W. Marais kindly compared the type and reported that although the
specimen is sterile, it undoubtedly agrees with the general conception of K. nana.
Foetid perennial with thick tuberous roots and long slender climbing or more
rarely prostrate stems, varying from nearly glabrous to more or less densely and shortly
pubescent. Stems slender, longitudinally sulcate, usually much branched and usually
thinly hairy when young, glabrescent. Leaves fleshy drying subcoriaceous rarely
thinner in texture, in outline cordate-reniform or suborbicular-cordate, usually distinctly
angular sometimes somewhat lobed, rarely as far as the middle, deep green when alive
but usually drying greyish green (often paler and more greyish on lower surface) 2-7
cm long and broad with a shallow rounded to subtruncate, occasionally broadly triangular
basal sinus, an obtuse or rounded rarely acute, sometimes mucronate apex and lobes
and a subentire to crenulate or somewhat crenate-dentate margin; petioles 1-4 cm long,
usually rather slender. Male plants: flowers racemose; common peduncle slender to
filiform, up to about 20-flowered, almost invariably shortly and sparingly pubescent,
2-10 cm long; pedicels erect-patent, filiform to capillary, 3-16 mm long, often with
minute bracts at the base; receptacle narrowly campanulate to subcylindric 2-4 mm
long, 1-3 mm wide, usually more or less hairy; sepals narrowly triangular, erect,
1-2 mm long; petals light yellow, ovate-oblong or ovate-triangular, subacute, densely
papillose-puberulous on the outside 4-9 mm long. Female plants: flowers solitary,
on a filiform peduncle up to 8 mm long in fruit; ovary oblong-fusiform, more or less
beaked, sometimes distinctly but finely longitudinally striate. Fruit ovoid, conical-acute
to rostrate, orange to red, smooth and glabrous when ripe, few-seeded, 1 • 5-2 cm
long, about 8 mm in diam. Seeds smooth, 6-7 mm long, 3-4 mm broad and 2-3 mm
thick.
A rather variable taxon, described under several names, varying from almost
completely glabrous to rather densely hairy and varying in the leaf-shape. The identity
of various synonyms was established by studying authentic specimens (or photographs
of these), as indicated under cited specimens.
Cape Province. — Cape Peninsula: Bolus 4015 (BOL); Compton 21860 (NBG);
Williamson 23 (GRA); near Cape Town, Rehmann 1548 (BR, L); Camp’s Bay, A Prior
s.n. (PRE); Galpin 27 (GRA, PRE); Esterhuysen Mill (PRE); Barker 5582 (NBG);
Llandudno, Compton 8892 (NBG); Houtbay, Compton 13229 (NBG); Table Mountain,
Stinkwater Gorge, Moss 7443 (J); Muizenberg, MacOwan 3107 (NBG); Eyles 6503
(SRGH); Brain 5884; 6049; Meebold 11867 (M); Esterhuysen 12848 (BOL).
Slangkop Station, Marloth 1 1941 ; East Head, Williamson 14 (GRA); near Simonstown,
Wodey Dod 1176 (BOL); Rodin 3293 (BOL, PRE); Lion’s Head, Wolley Dod 2325
(BOL); near Ronde Vlei, Leipoldt 4153. Stellenbosch: Krdusel 474 (M). Caledon:
Onrust Rivier, van der Riet in Hb. Marloth 11941 (PRE); van Niekerk 337 tBOL);
Fsterhuysen Hb. no. 25604 (BOL); Kleinmond, de Fos 387 (BOL); Hermanns, Burtt
31
Davy 18493 (PRE). Bredasdorp : Melkbosch, Lotsy & Goddijn 1689 (L); Brand-
fontein, Schlechter 10579 (BOL). Riversdale: Muir s.n., 105 in Hb. Galpin 5201
(PRE). Mossel Bay: Mossel Bay, Tay/o/" 1 1 14; 1115; 1116 (PRE). George: Kaai-
mans River, Wilman Hb. no. 25159 (BOL, PRE); Wilderness, Compton 10648 (NBG).
“ George-Knysna ” : Martin 4423; 4458. Knysna: Breyer TRY no. 23694 (PRE);
Keet 700 (GRA, PRE); Taylor 1154; 1335. Humansdorp: Krom River, Drege 8186
(type of Zehneria hederaceae Sond). (S, holo., photo. PRE, L, iso.!); Groot Rivier
Forest, 736 (NBG). Humansdorp: Fowrcut/e 1 193 (NBG). Uitenhage: Zwart-
kopsrivier, Ecklon & Zeyher 1781 (NBG); Zeyher 2480 (iso.! of Kedrostis mollis in
NBG); Port Elizabeth; Humewood, Paterson 589 (BOL, GRA); Port Elizabeth,
/. L. Drege s.n. (GRA). Alexandria: Olifantshoek, Ecklon & Zeyher (NBG).
Bathurst : Port Alfred, Salisbury 12 (NBG); Fish River Mouth, Davies s.n. (GRA)
Komgha; Kei Mouth, Elanagan 609 (NBG, PRE). Cape, without precise locality :
Tbunberg (holotypes of Bryonia angulata Thunb., B. africana Thunb., non L. and
B. triloba Thunb. in UPS, photo, in PRE!).
Natal. ^ — Eshowe: Lawn 2263 (NH).
{b) K. nana {Lam.) Cogn. var. zeyheri (Schrad.) A. Meeuse, stat. nov.
Coniandra zeyheri Schrad. apud E. & Z., Enum. PI. Cap. 275 (1834); Linnaea 12;
403 (1838); Sond. in FI. Cap. 2: 485 (1862), inch var. angustiloba Sond.
Cyrtonema sphenoloba Schrad., op. cit., 276 (1834); 403 (1838).
Kedrostis zeyheri (Schrad.) Cogn., Mon. Cucurb. 641 (1881); Pflanzenreich 147 (1916),
inch var. angustiloba (Sond.) Cogn., op. cit. 642 (1881); 148 (1916). K. digitata
(Thunb.) Cogn. sensu Cogn., Mon. Cucurb. 639 (1881) and in Pflanzenreich 146 (1916),
ex parte, inch var. major Cogn., op. cit. 640 (1881); 147 (1916).
Type: Ecklon & Zeyher 1775 (L, NBG, iso.!).
Differs from K. nana var. nana only in the shape of the leaves, which are dissected
beyond the middle with rhomboid to oblong-cuneiform, often 3-lobed acute or obtuse
lobes and are usually glabrous.
Cape Province. — Worcester: Hex River Mts., Lam & Meeuse 4569 (L). Prince Albert:
Bolus 10463 (BOL). Uitenhage: Zwartkopsrivier, Ecklon & Zeyher 280 (BOL, NBG);
Zeyher 1775 (L, NBG); 2472 (NBG); Enon, Thode A1104 (NH); Addo, Ecklon &
Zeyher 761 (BOL, NBG). Port Elizabeth: Redhouse, Paterson 229 (GRA); 526
(GRA, BOL); 626 (GRA); Humewood, Johns s.n. (NBG); Alexandria: Coega,
Rogers 113 (GRA); 4 miles S. of Zuurberg Hotel, Story 2302 (PRE); Boschmans
River, Ecklon & Zeyher 845 (NBG); Zeyher 2471 (NBG). Bathurst; Kariega River,
Acocks 18331; 18331A; 18355 (PRE). Albany: Howieson’s Poort, Britten 2867
(GRA); Pluto’s Valley, Dyer 1835 (GRA, PRE). Fort Beaufort: near Fort Beaufort,
VmrjF 2214 (PRE). King William’s Town: S/m 20234b (PRE). East London: Galpin
3181 (GRA, PRE). Kentani: Pegler All (PRE).
Natal. — Camperdown: Eranks in Hb. Wood 11682 (NH). Eshowe: Lawn 552
(NH). Lower Umfulosi: near Empangeni, Lawn 2173 (NH); Umhlatuzi Valley,
Lawn 2206 (NH). Hlabisa; Ward 1658 (NH). Ingwavuma; Ndumu Game Reserve,
Ward 3167 (PRE).
(c) K. nana {Lam.) Cogn. var. schlechteri (Cogn.) A. Meeuse, stat. nov.
K. schlechteri Cogn. in Bull. Herb. Boiss. 2me. ser. 6; 829 (1906); Pflanzenreich 148
(1916). K. digitata (Thunb.) Cogn. Mon. Cucurb. 639 (1881) and in Pflanzenreich 146
(1916), ex parte.
Bryonia digitata Thunb., Prodr. PI. Cap. 13 (1794).
Type: Schlechter 6053 from Gamtoos River (Z, holo.!; PRE, photo, BOL, GRA,
iso.!).
32
Differs from the other two varieties in the usually deeply dissected leaves with
narrow (ligulate) lobes or ultimate lobes; upper or both surfaces of the leaves with
distinct white pustules, rarely without.
Cape Province. — Riversdale: Muir 2160 (PRE). Humansdorp or Port Elizabeth:
Gamtoos River, Schlechter 6053. Uitenhage or Port Elizabeth: “ Koega River and
Karroid ground, Winterhoeksberge ”, Ecklon & Zeyher 79.2 (PRE). Port Elizabeth:
Addo Game Reserve, Brynard 427 (PRE); 36 miles from P.E. on Steytlerville Road,
Long 1210 (GRA, PRE). Alexandria; Archibald 7360 (PRE).
The three varieties as defined here are not sharply separable, but the extremes
are so different in habit that 1 hesitate to unite them all. Culture experiments with
seeds of all forms grown side by side may clear up their relationships. There is certainl}-
some relation between the distribution (ecology) and the morphology of the three
forms, var. schlechteri occurring in the driest regions, var. nana in the coolest and
dampest, whereas var. zeyheri is more or less intermediate in this respect. The holotype
of K. nana is in the Lamarck herbarium (P) and was seen by Mr. W. Marais who reported
that, although it consists of sterile pieces, it undoubtedly represents K. nana as understood
by Cogniaux and others (here called var. nana). K. mollis is merely a hairy form
(isotype, Zeyher 2489 in NBG, SAM), as is K. velutina Cogn. {Schlechter 2596 in
Z!, holo.).
The specimen Drege 8186 on which Sonder based his Zehneria hederacea (S, holo.!
L, iso.!) is rather “ typical ” K. nana var. nana. Cogniaux transferred it to Melothria,
1 do not know on what grounds, but presumably because he never saw the type
and transferred it “ automatically ” from Zehneria to Melothria. Thunberg collected
the var. nana several times and described it as Bryonia africana (non L.), B. angulata
and B. triloba. B. angulata was placed by Cogniaux (1881, 1916) as a synonym of
'''Melothria punctata''' (= M. cordata, see p. 19) but here again he did not study
the type which is clearly Kedrostis nana var. nana. Bryonia digitata Thunb. is represented
in his herbarium by a poor specimen which is most probably the same as Kedrostis
schlechteri Cogn. Coniandra digitata Sond. and Kedrostis digitata Cogn. are nomen-
claturally based on Thunberg’s name, but as far as T can see both Sonder and Cogniaux
included mostly specimens which are actually K. capensis (= Pisosperma capense
Sond.) and perhaps some which are K. africana, so that the name is somewhat confused.
For these reasons I do not take up digitata as the epithet for the variety of nana to which
1 refer Thunberg’s Bryonia digitata, but the unambiguous epithet schlechteri.
4. K. africana (L.) Cogn., Mon. Cucurb. 643 (1881); Pflanzenreich 275.1: 149
(1916); Dinter in Fedde, Repert. 18: 434 (1922); Burtt Davy FI. Transv. 1; 226
(1926).
Bryonia africana L., Sp. PI. ed. 1; 1013 (1753). B. dissecta Thunb., Prodr. PI. Cap.
13 (1794); FI. Cap. 154 (1811); FI. Cap. ed. Schultes .36 (1823); Ser. in DC, Prodr.
3: 308 (1828). B. digitata Thunb., op. cit. 13 (1794); 154 (1811); 35 (1823); Ser.,
op. cit. 309. B. pinnatifida Burch., Trav. S. Afr. 547 (1822); Ser., op. cit., 308.
B. grossulariaefolia E. Mey. ex Drege, Zw. Pflgeog. Doc. 132, 169 (1843), nomen tantum.
B. multifida E. Mey. ex Drege, op. cit., 54, 147, 149, 169, nomen tantum.
Coniandra grossulariaefolia E. Mey. ex Arnott in Hook., Journ. Bot. 3; 273 (1841).
C. glauca Schrad. apud E. & Z., Enum. PI. Cap. 277 (1834); Linnaea 12; 403 (1838).
C. dissecta (Thunb.) Schrad., C. pinnatisecta Schrad. apud E. & Z., op. cit., 276. C.
africana (L.) Sond., C. digitata (Thunb.) Sond., Fl. Cap. 2; 483 (1862). C. punctulata
Sond., op. cit., 484.
Kedrostis digitata (Thunb.) Cogn., op. cit. 639 (1881); 146 (1916). K. glauca (Schrad.)
Cogn., op. cit., 640 (1881); 147 (1916). K. punctulata (Sond.) Cogn., op. cit. 642
(1881); 148 (1916); Burtt Davy, l.c.
33
Type: Linnaeus originally based this species on “ Bryonia africana lacinata,
tuberosa radice, floribus luteis” (Linnaeus, in 1753, cited erroneously “ herbaceis ”
instead of “luteis”), Herm. Farad. Bot. 107 (1698) and t. “ 108” (recte; t. 23, see
below). In addition, Linnaeus cited “ Hort. Cliff. 453 ”.
Herman cites under the plate cited by Linnaeus “ Bryonia africana glabra foliis
profunde sectis, teruiroribus (sic!). This plate is the 23rd and should be cited as
“ t. 23 ” but, as is done in many contemporary works, the relevant text page is also
indicated, which in this case is “ 108 ”. However, there is no reference to a Bryonia
africana glabra on this page 108, and this phrase name does not occur anywhere in the
text. The phrase name cited by Linnaeus and subsequent authors, viz., “ Bryonia
africana laciniata, etc.” is not found on any of the other plates in the book. Linnaeus
assumed that the plate belongs to the text on p. 107 (and not on p. 108 as indicated
on t. 23), and this is most probably what Herman intended. Plate 22 of Herman’s
publication refers to p. 107, but this plate represents a plant from Ceylon and is the
basis of Bryonia laciniosa L., which is now generally accepted to be Bryonopsis laciniosa
(L.) Naud., which does not occur in Southern Africa. For nomenclatural purposes,
at any rate, t. 23 (= t. “ 108 ”) is the basis of Bryonia africana L.
In the second edition of Species Plantarum (p. 1013) Linnaeus cites “ Bryonia
africana laciniata . . . herbaceis ” again, repeating the error in the first edition.
Willdenow, in his edition of the Species Plantarum 4: 624 (1805), cites the same phrase
under “ Bryonia africana ” but this is B. africana Thunb. non L. (= Kedrostis nana)
so that this citation is absurd, because the cited plate does not agree in the least with
Thunberg’s description of B. africana (non L.) as having angular 5-lobed leaves and,
in addition, because “ B. (africana), Sp. PI. 1438 ” is cited by Willdenow under Bryonia
dissecta Thunb. |= K. africana (L).] Cogn.; and the Herman reference belongs to the
Linnaean B. africana, not to Thunberg’s.
Herbaceous climber, occasionally prostrate. Rootstock tuberous, perennial.
Stems annual, slender, angular or angular-sulcate; glabrous or nearly so, much
branched, up to 6 m long but usually considerably shorter. Leaves sessile to shortly
petioled, herbaceous drying membranous or occasionally thicker or tougher to slightly
coriaceous, glabrous and smooth to scabrid-punctulate mainly on upper surface,
sometimes scabrid with distinct white scaly dots, orbicular, cordate or triangular in
outline, varying from deeply pinnately or digitately dissected to occasionally (and
also the first leaves formed) only pinnately or palmately lobed, 2-10 cm in diam.;
the segments varying from narrow, filiform to lanceolate or elliptic or somewhat obovate
to cuneiform; the ultimate lobes filiform to rather wide (up to 12 mm broad), acute,
acuminate, obtuse or rounded, usually mucronate, entire or occasionally dentate the
margins flat (in the thin, herbaceous leaves) or more or less recurved (in the firmer
leaves which are usually also more punctate-scabrid); petioles slender, sulcate, glabrous
or slightly pilose, 1-12 mm long. Tendrils simple, slender, glabrous. Flowers monoe-
cious, minute, with a yellowish green, whitish or cream-coloured corolla, almost
invariably the subumbellately racemose flowers in the same axils as the sessile or
subsessile and solitary or occasionally fascicled female flowers. Male flowers: common
peduncle filiform, glabrous, finely striate, few- to 12-flowered, 2-8 cm long; pedicels
capillary, usually patent or erecto-patent, 2-5 mm long, usually with minute (0-5-1
mm long) bracts at the base; receptacle 0-5-1 -5 mm long and 1-2 mm in diam.,
sepals narrowly triangular to linear or subulate, 1-2 mm long; petals finely papillose,
1-2 mm long. Female flowers: peduncles (even in fruit) not exceeding 6 mm; ovary
subglobose or ovoid, constricted to shortly rostrate at the apex, glabrous. Fruit
subglobose, often shortly and abruptly pointed, red when ripe, glabrous, few-seeded,
8-15 mm in diam. Seeds ovoid-oblong or ovoid, somewhat attenuate and truncate
at the end; finally granulated, 4-5-6 mm long, 3 -5-4 -5 mm wide and 2-5-3 mm
thick.
34
Recorded from the following districts : South West Africa: Rehoboth, Maltahohe,
Luderitzbucht, Bethanien, Keetmanshoop, Warmbad; Cape Province: Vryburg,
Kuruman, Barkly West, Hay, Kimberley, Herbert, Colesberg, Murraysburg, Graaff-
Reinet, Beaufort West, Laingsburg, Prince Albert, Swellendam, Uniondale, Uitenhage,
Port Elizabeth, Albert, Cradock, Queenstown, Fort Beaufort, Cathcart, Albany,
Bathurst, Komgha, Umtata; Natal: Utrecht, Weenen, Estcourt, Pietermaritzburg,
Camperdown; Orange Free State: Kroonstad, Boshoff, Senekal, Bloemfontein,
Fauresmith, Ladybrand; Transvaal: Rustenburg, Marico, Potchefstroom, Krugersdorp,
Brits, Pretoria, Lydenburg.
The following specimens are of special interest: Zeyher 603 (NH, PRE), cited
by Sonder as Coniandra punctulata, by Cogniaux as K. punctulata, locality not certain
(Sonder in FI. Cap. and Cogniaux cite different localities, but both seem to be in
Bushmanland) ; Zeyher 602 (SAM) and Burke s.n. (PRE), from Rhenosterkop Beaufort
West, also cited by Sonder and by Cogniaux as C. {K.) punctulata; Zeyher 600 (PRE)
and Burke 288 (SAM), from Mooi River, Potchefstroom, Transvaal, cited by Cogniaux
and by Burtt Davy as K. africana; Zeyher 600 cited by Sonder as C. africana; Burke
287 from the Sneeuwberg, probably collected in Graalf-Reinet (PRE), and Ecklon &
Zeyher Ml 6 from the Gauritz River, Swellendam (SAM) and 1778 from same locality
(M), cited by Sonder and by Cogniaux as C. (K.) africana; Ecklon & Zeyher Mil
(SAM), probably from Uitenhage, cited by Sonder as C. glauca var. dissecta and by
Cogniaux as K. glauca var. dissecta (this is probably the basis of Coniandra dissecta
Schrad., see below); Ecklon & Zeyher 1774 from Couga, Winterhoeksberge, Uitenhage,
isotype of C. glauca Schrad. (SAM); Drege 8188a and 8189 from Beaufort West (L),
both cited by Cogniaux as K. punctulata {Drege 8189 cited by Sonder as C. punctulata,
Drege “ 8188 ” is cited by Sonder as probably being C. glauca)', Rehmann 4193 (BOL,
GRA) and 4708 (BR, L), from Pretoria, cited by Cogniaux (1916) and by Burtt Davy
as K. africana; Bolus 146 from Graaff-Reinet (BOL, GRA, SAM), cited by Cogniaux
under K. africana; Flanagan 1412 from Kimberley (BOL, PRE), erroneously cited
as “no. 3650 ” by Cogniaux under K. punctulata (“ 3650” is the altitude in feet!);
Galpin 2134 from Queenstown (BOL, PRE) cited by Cogniaux as K. africana; Schlechter
2528 from Uitenhage (BOL, COI, GRA, J, Z), cited by Cogniaux (1916) as K. punctulata;
Dinter 2686 from the Schaaprivier (SAM), cited (1916) as K. africana; Dinter 2249
from Rehoboth-Aus (SAM) cited (1916) as K. punctulata; Dinter 1158 from Jakalls-
kuppe, Luderitzbucht, and Dinter 1905 from Schaaprivier (probably Rehoboth distr.)
both in PRE and SAM, cited by Cogniaux (1916) as 7^. punctulata var. tenuiloba (Sond.)
Cogn. ; Leendertz 464 from Pretoria (L, PRE) cited by Burtt Davy as K. africana;
Schlechter 6145 from Gonubi River, East London (GRA), erroneously referred to
K. zeyheri var. angustifolia by Cogniaux (1916).
The various described species lumped here are only forms of one variable species.
The variation in leaf shape and the presence or absence of the white scabrid scale-like
dots on the vegetative parts account for the considerable synonymy. Young plants
form undivided to palmatilobed or palmately dissected leaves which occasionally
persist in older plants, and such specimens have been described as Bryonia digitata
Thunb. and Coniandra glauca Schrad. The forms with pinnatisect leaves occur in two
different types, one with narrow leaf-segments with recurved margins which sometimes
appear almost filiform and are often also distinctly white-punctate (described as
Coniandra punctulata Sond.; specimens with very narrow leaf segments as var.
tenuiloba Sond.); and one with fiat, broader, more herbaceous and usually less white-
punctate segments which is Coniandra africana as understood by Sonder, and K. africana
sensu Cogniaux. There are many intermediates, however.
The latest monographer of the genus (Cogniaux) obviously had some difficulty
in separating some of the four “ species ” he upheld (but which are reduced to one
here), because the gathering Ecklon & Zeyher Mil appears in his treatments of 1881
35
and 1916 as “ K. glauca var. B. dissecta (Sond.) Cogn.”, but this is certainly wrongly
cited, for the varietal name is based on Coniandra glauca var. dissecta Sond. in FI. Cap.
2; 484, and Sonder cites “ C. dissecta Schrad.” (apud E. & Z., Enum. PI. Cap. 276)
as a synonym. Schrader and Sonder both cite “ Ecklon & Zeyher 1777 ” under the
name “ dissecta ” but the latter author and Cogniaux apparently overlooked the fact
that Schrader cited Bryonia dissecta Thunb. as a synonym of his Coniandra dissecta,
so that he actually made a new combination. Now both Sonder and Cogniaux cite
Bryonia dissecta Thunb. as a synonym of K. africana, but the Ecklon & Zeyher
gathering appears in their publications as a var. dissecta of K. glauca. Such inconsis-
tencies are easily explained if one accepts the synonymy as indicated in the present
paper.
5. Kedrostis crassirostrata Breinek. in Ann. Transv. Mus. 15, p. 260 (1933).
Type: Bremekamp & Schweickerdt 40 in PRE (holo.!).
Most probably perennial with an underground tuberous rootstock, glabrous in all
vegetative parts or nearly so. Stems annual, climbing, herbaceous, up to about 2 m
long, slender, geniculate at the nodes, sulcate; internodes usually 4-5 cm long. Leaves
subsessile, somewhat fleshy drying thin and papery, the lamina suborbicular in outline,
white-punctate above, smooth below, 2-7 cm in diam., digitately 3-5-lobed with
pinnatifid segments; the ultimate lobes oblong to linear, mucronate, with the margin
usually more or less reflexed below, up to 25 mm long and 0 • 5-2 mm, rarely 3 mm wide.
Male flowers subumbellately racemose; common peduncle rather stout, up to about
5 mm long, usually ebracteolate at the base, articulate at the apex; receptacle widely
campanulate, glabrous, about 1 mm long and 2 mm in diam. ; sepals patent, triangular-
subulate or lanceolate, usually a little longer than the receptacle, very acute; petals
ovate or oblong, acute, about 1 • 5 mm long, papillose-hairy on the back. Female
flowers solitary in the same axils as the males, subsessile; ovary from a lageniform or
ellipsoid basal portion long-rostrate, densely papillose-hairy, about 4 mm long; calyx
and corolla as in the male. Fruit on an incrassate up to 2-5 mm long peduncle, ovoid-
lageniform or obpyriform, rounded at the base, glabrous and red when ripe, rostrate
with a conical acute or obtuse rostrum, with longitudinal faint ribs, 1-5 seeded, 15-25
mm long (of which a little less than half is taken up by the rostrum) and 8-15 mm
in diam. Seeds brown, broadly ellipsoid with a slightly attenuate and truncate base,
very slightly rugose, laterally margined, 4-5 mm long.
South West Africa. — Gobabis: Sandfontein, Bleek Hb. No. 26241 (SAM); Wilman
Hb. No. 26701 (SAM); Babi-babi: Wilman Hb. No. 3085 in (KMG) = Hb. No.
15300 (BOL).
Cape Province. — Vryburg: Palmyra, 60 miles N.W. of Vryburg, Rodin 3535 (BOL,
PRE); near Vryburg, Barkhuizen 100 (PRE). Taungs: Home Rule, Brueckner 596
(KMG, PRE); Vryburg: Ffenrici 48 (PRE); Tipperary, Brueckner 1090 (KMG,
PRE); Dry Harts, Mogg 8961 (PRE). Barkly West: Waldeck’s Plant, Acocks 1443
(KMG, PRE).
Transvaal. — Pietersburg: Brak River, Bremekamp & Schweickerdt 40 (PRE, holo.!).
Potgietersrust : Roedtan, Meeuse 9605 (PRE). Pretoria: near Pienaars River, Codd
818 (PRE); Gomes Pedro 710. Bloemhof: Kameelpan, Christiana, Theron 3471
(PRE).
Bechuanaland Protectorate. — Mochudi, Rogers 6361 (BOL, PRE).
This species occurs in dry sandy areas (Kalahari sand). As was already pointed
out by Bremekamp, this species is closely related to K. africana and “ K. punctulata ”
(= africana), but is quite distinct by its short male peduncles and long-rostrate fruits.
36
6. K. natalensis {Hook. /.) A. Meeuse, comb. nov.
Toxanthera natalensis Hook. f. in Hook., Icon. PI. 15: 17, t. 1421 (1883); Cogn. in
Pflanzenreich 275.1: 137 (1916). T. lugardae N. E. Br. in Kew Bull. 1909: 112
(1909); Cogn. in Pflanzenreich 275.1: 138 (1916). T. kwebensis N.E. Br., op. cit.,
113; Cogn., op cit., 137.
Kedrostis longipedunculata Cogn. in Bull. Herb. Boiss. 3: 421 (1895). K. rautanenii
Cogn. in Vierteljschr. Naturf. Ges. Zurich 55: 247 (1910); Pflanzenreich 275.1: 150
(1916). K. gilgiana Cogn., op. cit. fig. 35, 151 (1916). K. eminens Dinter & Gilg.
ex Cogn., op. cit. fig. 36, 152 (1916). K. otaviensis Dinter, Sukkulentenf. S.W. Afr. 2
(in Fedde, Repert., Beih. 53): 116 (1928), nomen subnudum.
Type: Hooker mentioned two gatherings, viz. Gerrard 1192 (excl. fruit) and
Wood 813. The accompanying plate shows a fruit and seeds which of course could
not have belonged to Gerrard 1 192, so that Wood 813 (K ex Herb. Hook.) is proposed
here as the lectotype.
Perennial. Rootstock tuberous, ovoid, to fusiform up to 20 cm long and 10 cm
thick (in very old specimens probably larger). Stems several from the crown of the
root, herbaceous, subterete, longitudinally striate, more or less thinly hairy (the
pubescence varying from shortly scabrid-pilose to setulose, shortly hirsute or puberulous),
prostrate and up to 1 • 50 m long or climbing and attaining to several meters in length.
Tendrils usually bifurcate above a short simple portion, or some or occasionally all
of them simple, often elongate, thinly hairy as are the stems. Leaves herbaceous,
dark green above, lighter beneath, broadly cordate-suborbicular or orbicular-subreniform
in outline, varying from faintly 5-lobed to 5-sect to below the middle, more or less
hairy above usuall> thinly so, thinly to densely hairy below, 3-16 cm long and wide;
the margin faintly to rather coarsely sinuoso-dentate, the teeth usually acute to setulose;
the veins often prominent below forming a coarse network; the lobes varying from
triangular to ovate, obovate, broadly elliptic, oblong or lanceolate with the basal ones
always tending to be falcate in deeply dissected leaves, acute to apiculate; basal sinus
usually deep and subrectangular, less often broadly rounded; petioles subterete, hairy
as are the stems, 15-10 cm long. Flowers monoecious, male and female ones produced
in the same or in different axils; receptacle semiglobose, broadly campanulate or
broadly obconical, shortly pubescent to hirtellous as are the peduncles, bracteoles,
pedicels and sepals, 2-4 mm high and as wide; sepals patent to erect, triangular-
subulate to linear-lanceolate or occasionally triangular-ovate, usually longer than the
receptacle; corolla pale yellow to somewhat greenish pale yellow, usually shortly
pubescent, usually shortly exceeding the sepals, more or less triangular-ovate usually
subacute to obtuse. Male flowers subumbellately racemose; common peduncle varying
from 1 cm (prostrate specimens) to 18 cm long (in climbing ones under favourable
conditions), usually 2-7 cm long; pedicels slender, hairy, usually patent, 2-12 mm
long; bracteoles subulate, deciduous. Female flowers solitary; pedicels 0-5-4 cm
long, rarely longer; ovary narrowly oblong, linear or fusiform, glabrous or more or
less hairy, usually about 2 cm long; staminodes 0-5, small; style columnar, stigmas 2,
at first cohering into an erect digitiform structure, later separating and often reflexed
with truncate to flabelliform tips. Fruit narrowly ellipsoid with long rostrate apex,
linear-fusiform or fusiform, smooth, red and glabrous when ripe, 4-9 cm long and
1-5-2 cm in diam. Seeds rather numerous, broadly ellipsoid-subglobose, 4-6 mm
long and 4-5 mm in diam.
South West Africa. — Ovamboland: Omakunde, Rautanen 703 (Z, type of K. rautanenii
Gilg). Okavango: Sambiu Camp, de Winter 4019 (PRE). Grootfontein : Gautscha
Pan, Story 6229 (PRE); Otavi, Dinter 5316 (B, PRE), 5524, 5548 (B). Okahandja:
Eahero, Dinter 3280 (SAM); Osona: Dinter 62 (isotype of K. gilgiana Cogn., PRE,
SAM). Windhoek: Leutwein, Dinter s.n. (B).
37
Bechuanaland Protectorate.— Mochudi, Rogers 6393 (BOL).
Cape Province.— Mafeking, v.d. Merwe 20 (BOL). Vryburg, Mogg 8961 (PRE).
Hay: Postmasburg, Lewis 5354 (SAM); Asbestos Hills, Esterhuysen 806 (BOL, KMG,
PRE); Dunmurry, Wibnan (BOL, Hb. No. 256013, KMG, H. No. 2371); Niekerk’s
Hope, WUman 1410 (KMG). Barkly West: near “Border”, Acocks 1891 (KMG).
Komgha, Flanagan 641 (BOL, SAM). Kentani : Peg/e>r 1280 (BOL, GRA, PRE, SAM).
Umzimkulu: Clydesdale, Tyson 2558 (BOL, SAM).
Transvaal.— Rustenburg: 15 miles E.S.E. of Rustenburg, Leistner 533 (PRE); 12
miles S. of Zwartruggens, Code! 2662 (PRE). Krugersdorp: Witpoortjie, Mottley
2385; Moss 16189 (J). Pretoria: Groenkloof, Mogg 14214 (PRE); Bon Accord,
Pyramid Hills, v. Niekerk & Wasserfall 32 (PRE, SRGH); de Wildt, Murrav 512
(PRE). Brits: Farm Welgevonden, Mogg 14632 (PRE); Crocodile Drift, Obermeyer in
T.M. 35159 (PRE, NH). Carolina: Waterval Boven, Rogers 14291 (Z). Lydenburg;
Ohrigstad, Young A 579 (PRE). Belfast: Schoemans Kloof, Pole Evans 3930 (PRE);
Young A 295 (PRE). Nelspruit: Ship Mountain, van der Schijjf 1656 (PRE).
Barberton: Pot? 5447 (PRE); Hislops Creek, P/iorncrp/? 862 (NH). Letaba: Ofcolaco,
Story 5403 (PRE). Soutpansberg: Messina, Moss & Rogers 5050 p.p. (J). Sibasa:
Pafuri, van der Schijjf 4096; 4174.
Natal. — Ndwedwe: IFoor/ 7520 (BOL, PRE, L). Camperdown: between Drummond
and Inchanga, Eshuis s.n. (PRE). Verulam: “ Umhloti and Inanda ”, Wood 813
(NH, isotype!). Inanda: Wood 8409 (type of K. longipedunculata Cogn., Z). Durban:
Wood 8540 (NH); Sarnia, Mariott in NH 26174 (NH).
Southern Rhodesia. — Bulalima-Mangwe: Embakwe, Feiertag Hb. No. 45475;
45480(SRGH). Bulawayo: Eyles & Johnstone 5\ {GRA)-, Kolbe A\65 {BOY). Wankie:
Rogers 5829 (BOL). Chipinga: Birchenough Bridge, Obermeyer 2506 (PRE).
A study of the available material, including living specimens cultivated from seeds
or tubers in Pretoria, and of most of the types, necessitates the reduction of many
described species to Toxanthera natalensis, the oldest name available, which at the
same time must be transferred to Kedrostis. Slight differences in habit, size and shape
of leaves, length of peduncles, etc., are easily explained by the variation of the ecological
conditions in the large area of distribution. Some species were described as having
a single digitiform stigma and others as having two reflexed flabelliform stigmas, but
this is merely a matter of age: the two stigmas are erect and cohering in young flowers
but later separate and become reflexed. The number of staminodes varies from 0
to 5 which probably explains the fact that Cogniaux and N. E. Brown described some
forms with 5 staminodes under Toxanthera and others (with 0-3 staminodes) under
Kedrostis. The seeds are white or whitish and N. E. Brown’s statement that the seeds
of Toxanthera lugardae are scarlet must be erroneous; most probably the red fruit
pulp adheres to the seeds or has stained them red.
The known area of distribution suggests that the species extends even further
north and some species described from West- and East Africa such as Kedrostis leder-
mannii and K. rigidiuscula are possibly identical. This would extend the range
considerably.
The actual type specimen of Kedrostis eminens was not available for study (only
fig. 35 in Cogniaux’s 1916 Monograph), but Dinter himself added this name to his
no. 62 (described as K. gilgiana by Cogniaux) and to his number 5316. On his sheets
in the Berlin herbarium Dinter wrote that Toxanthera kwebensis and K. eminens are
most probably identical. It seems fairly safe to reduce K. eminens to a synonym of
K. natalensis. K. otaviensis Dinter, which was published without a proper description,
is another synonym, because Dinter mentioned that he collected this plant under the
number 5524, a sheet of which is represented in Berlin and proved to belong here.
38
Excluded Species
K. bainesH (Hook, f.) Cogn. (Rhynchocarpa bainesii Hook, f.) = Corallocarpus bamesii
(Hook, f.) A. Meeuse, see p. 41.
K. cinerea Cogn. in Bull. Herb. Boiss. 2me. sen, 1: 883 (1901) = Melothria cinerea
(Cogn.) A. Meeuse, see p. 17.
7. CORALLOCARPUS
Corallocarpus Wehv. ex Benth. & Hook, f., Gen. PI. 1: 831 (1867); Trans. Linn. Soc.
27: 32 (1869); Hook. f. in FI. Trop. Afr. 2; 565 (1871); Cogn., Mon. Cucurb. 645
(1881); Pflanzenreich 275.1: 156 (1916); Pax in Pflanzenfam. 4, 5: 18 (1889);
Phillips, Gen. ed. 2: 747 (1951).
Perennials with usually tuberous rootstock forming annual herbaceous scandent
or prostrate stems, rarely (not in S. Afr.) erect shrubs; often scabrid or more or less
hairy. Leaves petiolate, entire or palmately dissected, usually more or less cordate
at the base. Tendrils simple, very rarely bifid or absent. Flowers monoecious, minute,
greenish-yellow, usually the male and female ones produced in the same axil. Male
flowers racemose or subumbellate on elongated peduncles, rarely fasciculate; receptacle
campanulate, sepals short; corolla 5-partite, segments ovate-oblong; stamens 3, free,
inserted on the receptacle; filaments very short, anthers entire or bi-partite, glabrous,
2-thecous and the third 1-thecous; thecae straight; connective often bifid, dilated
or produced at the apex; rudiment of pistil minute. Female flowers sessile or shortly
pedicelled, solitary or fascicled, very rarely racemose; receptacle, calyx and corolla
as in the male; staminodes 0 or minute; ovary subglobose or ovoid, often beaked,
2-3-chambered with few ovules; style straight, not surrounded by a disc at the base;
stigmas 3- or rarely 2-4-lobed. Fruit small, fleshy, subglobose to ovoid or oblong,
obtuse to beaked, smooth, few seeded, when ripe red and circumscissile near the base.
Seeds obovoid, tumid; cotyledons obovate, thick, fleshy; radicle small, conical.
Type species: C. welwitschii (Naud.) Hook. f. ex Welw. (the only species mentioned
by Welwitsch).
Mainly centred in Africa in the drier subtropical and tropical zones, extending
through Arabia to India and with a few species in America. In the latest monograph
(1916), 38 species are distinguished, but a critical re-examination would probably
reduce this number to about 20 or even less.
Corallocarpus is closely related to Kedrostis but differs constantly in the circum-
scissile fruits. The species of Corallocarpus are mostly very similar in general appearance
and indicate a “ natural ” genus.
Fruit usually distinctly rostrate, stipitate to pedunculate; leaves white-punctate below, deeply
3-5-sect with acutely lyrate pinnatifid lobes, the ultimate segments narrow (southern portion
of S.W. Africa, Namaqualand) 1. C. dissectus
Fruit obtuse to rounded or sometimes acute but usually not distinctly rostrate; leaves not white-
punctate below, entire or lobed to 3-5-sect but not with acutely pinnatifid lobes divided into
narrow segments:
Fruit usually stalked, ovoid to oblong, usually more or less acute at the apex 2. C. welwitschii
Fruit sessile, broadly ellipsoid to subglobose, rounded at the apex 3. C. bainesii
1. C. dissectus Cogn., in Bull. Herb. Boiss. 2me. ser. 1 ; 881 (1901); Pflanzenreich
275.1: 160 (1916). Type: Dinter 2 in Z.
39
Stems slender, sulcate, somewhat geniculate at the nodes. Leaves triangular-ovate
or somewhat cordate in outline, 3-6 cm in diam., digitately 3-5-partite to the base with
pinnatifid segments; ultimate segments usually few, often concave, linear, acute;
upper surface as a rule glabrous and more or less smooth; lower surface usually finely
scabrid with small acute whitish tubercles; petioles slender,’ glabrous, smooth, 5-20
mm long. Tendrils filiform, slender, usually elongate, glabrous. Male flowers: common
peduncle slender to filiform, straight, glabrous, sulcate, 3-12 cm long, at the apex
sub-capitately 3-8-flowered; pedicels capillary patent, 1-3 mm long; receptacle cam-
panulate, glabrous, about 1 • 5 mm long, sepals erect, triangular-subulate, nearly 0 • 5
mm long; petals about 1 mm long. Female flowers solitary or in few-flowered fascicles
(rarely more than two together), shortly pedicelled; ovary ovoid or ellipsoid, long-
rostrate. Fruit on a much elongating, up to 3 cm long, much incrassate, usually
slightly to distinctly clavate, glabrous pedicel, subglobose or broadly ellipsoid, rounded
to subtruncate at the base, with a narrowly conical, up to 7 mm long rostrum, glabrous,
7-14 mm long without the rostrum, 7-10 mm in diam. Seeds ovoid, slightly compressed,
smooth, faintly margined, 6 mm X 3 mm x 1 • 5 mm.
South West Africa. — Gibeon: Griindorn, Range 1339 (SAM). Bethanien: Inachab,
Dinter 2 (Z, type!); Sandverhaar, Dinter 1201b (PRE, SAM); Tschaunap Mission
Station, Gerstner 6356 (PRE). Warmbad: Warmbad, Pearson 4282 (BOL).
Cape Province. — Bushmanland: between Wortel and Dabainoris, Pearson 3032
(BOL). Gordonia: Gonsis, Wilman Hb. no. 579 (KMG).
2. C. welwitschii (Naud.) Flook. f. ex Welw., Sert. Angol. in Trans. Linn. Soc.
27: 32, t. 12 (1869); El. Trop. Afr. 2: 566 (1871); Cogn., Mon. Cucurb. 651 (1881);
Pflanzenreich 275.1: 162 (1916).
Rhynchocarpa welwitschii Naud. in Ann. Sci. Nat. 4me. ser. 17: 198, t. 10 (1862).
Corallocarpus schinzii Cogn. in Abh. Bot. Ver. Brandenb. 30: 153 (1888); op. cit.
162 (1916). C. gilgianus Cogn. in Pflanzenreich 275.1: 163 (1916). C. glaucicaulis
Dinter & Gilg. ex Dinter in Eedde, Repert. 16: 241 (1919), nomen. C. scaber Dinter
& Gilg. ex Dinter, op. cit., p. 241, nomen. C. bequaertii De Wild, in Rev. Zool. Afr. 9,
Suppl. Bot. B. 91 (1921); PI. Bequaert. 1: 559 (1922); Robyns, El. Spermat. Parc Nat.
Albert 2: 393 (1947).
Type : Naudin described this plant from cultivated specimens grown in the botanical
garden in Paris from seeds sent by Dr. Welwitsch from the botanical garden, Lisbon.
These plants were obviously of the same origin as the specimens distributed under
the number Welwitsch 799, which was reported to be cultivated in Lisbon. The plates
accompanying Naudin’s and Welwitsch’s descriptions are also representative.
Perennial, often somewhat glaucous. Rootstock tuberous, usually somewhat
fusiform. Stems annual, herbaceous, only becoming woody very late, slender to rather
stout, angular-sulcate, often geniculate, thickened at the nodes, shortly scabrid-setose
on the ridges between the grooves, glabrescent to glabrous, up to 2 m long, rarely
longer; internodes usually 3-10 cm long. Tendrils usually slender to filiform, thicker
and sulcate-ribbed near the base, glabrous or scabrid-setose towards the base. Leaves
cordate to orbicular-cordate, oblong-cordate or occasionally hastate-cordate in outline,
3-8 cm long and 2-7 cm wide, cordate at the base (usually deeply so with a conspicuous
basal sinus), either undivided or more or less distinctly lobed to deeply palmately
5-sect; if lobed, the lobes variable but usually narrowed at the base, sometimes
lobulate; apices of leaves (and lobes) acute or obtuse, mucronate; leaf-margins dentate
to sinuoso-dentate or lobulate to occasionally nearly entire; both surfaces but especially
the paler lower one usually shortly scabrid-hispid; petioles usually sparsely and shortly
scabrid-setose, 2-5 cm long. Male flowers: common peduncle straight, glabrous,
capitately 2-8-flowered at the apex, 1-6 cm long; pedicels filiform, erecto-patent.
40
1-3 mm long; flowers minute. Female flowers solitary or 2 to few together in fascicles,
on short stout pedicels. Fruit on a stout subclavate peduncle up to 1 cm long, ellipsoid
to oblong, usually not very acute at the apex, smooth and ultimately usually quite
glabrous, 5-8-seeded, 1 5-20 mm long and 8-12 mm in diam. Seeds about 3-5 X 2x1-5
mm.
Angola. — Specimens grown in botanical garden, Lisbon, from seeds collected in Angola
(probably in Loanda): Welwitsch 799 (COI, original material from which Welwitsch
collected seeds sent to Naudin in Paris).
South West Afr'ca. — Kaokoveld: Ohopoho, de Winter & Leistner 5213 (K, M,
PRE). Grootfontein: Schinz 298 (Z); Schoenfelder 927 (PRE); farm Kumkauss,
Kinges 2913 (PRE); Auros, Otavi, Dinter 5525 (B); Nama Pan, Story 5315; Aha
Mts., Story 6377 (PRE). Outjo: between Kamanjab and Outjo, de Winter 3065 (PRE).
Omaruru: V. Mahzahn S.W.A. 7 (BOL). Okahandja: Dinter 443 (SAM, isotype
of C. gilgianus Cogn.); 534 (PRE, SAM); 3088 (PRE, SAM); Bradfleki 318A
(PRE); Rapsom 5 (PRE). Karibib: Dinter 6729; 6992a (B); Klein Ameib, Dinter
7095 (B). Okomitundu: Seydel 1346 (PRE). Windhoek: Awas Mts., Pearson 9616
(BOL); Windhoek: de Winter 2700 (PRE); farm Otjisewa, Wiss & Kinges 805;
farm Rietfontein, Strey 2540 (PRE); farm Lichtenstein, Dinter 4578 (B). Rehoboth:
farm Djab, Walter 4510 (M); Naukluft Mts. near Ababis, Pearson 9698 (BOL, probably
this species); farm Buellsport, Strey 2044 (PRE); Rehoboth or Maltahohe:
Gamis, Dinter s.n. (SAM); Maltahohe: Usib River near Nomtsas, Pearson 9303
(BOL). Gibeon: Haribes, Volk 2351/56 (M); Voigtsgrund, /’cari'ow 9363 (BOL).
Liideritzbucht: Aus, Schinz 306 (BR, Z, isotype and type of C. schinzii Cogn.);
Dinter 6270 or 6276 (B); Krause! 866 (M); Kahrnstal, Dinter 8165 (B, M); farm
Gamochab, Kinges 2505 (PRE). Keetmanshoop: near Narubis, Acocks 18020 (PRE);
Griindorn, Pearson 4275 (BOL). Warmbad: Great Karasberg, Kraikluft: Pearson
8112; 8115 (BOL); Narudas-Sud, Pearson 8177 (BOL, SAM); 8522 (BOL).
Exact district unknown: Omaruru or Rehoboth, “Farm Weissenfels ”, Volk 1485/56
(M); without locality: Schaefer 927 in Herb. Dinter 3887 (B).
Bechuanaland Protectorate. — Near Molepolole, Miller B/564 (PRE).
Cape Province. — Hay: Niekerk’s Hoop, fF/Vma/i Hb. No. 1407 (KMG, BOL); Lanyon
Vale, Acocks 474 (KMG); Dunmurry, Wilnmn Hb. No. 2620 (KMG, BOL).
Also recorded from the Belgian Congo (e.g. Gillett 3190! from Moanda in BR,
cited by Cogniaux 1916).
Young plants have, according to observations in the garden of the Division of
Botany, usually undissected leaves. Similarly, young shoots often form first undissected
leaves and later more dissected leaves. These undissected leaves occasionally occur
in older plants or slioots and such plants have been described as a different species,
viz. C. schinzii. Strangely enough, Cogniaux described a var. subintegrifolia of C.
welwitschii in Mon. Cucurb. 652 (1881) and a var. lobatus Cogn. of C. schinzii in
Pflanzenreich 275.1: 163 (1916). This renders the distinction between the two very
obscure, because the other characters used by Cogniauux in his key in the 1916 mono-
graph, viz., verruculose seeds (in C. welwitschii) against smooth seeds (in C. schinzii)
and an acute base to the male receptacle (in C. welwitschii) against a rounded receptacle
(in C. schinzii), do not hold. These two names are clearly synonymous.
C. gilgianus, as is clear from an isotype {Dinter 443 in SAM), is only an extreme form
with very narrow leaf-segments. C. welwitschii is sometimes very similar in appearance
to C. hainesii { — C. sphaerocarpus), but is certainly distinct from the latter. Not only
does it differ in some morphological features (such as the fruits), but it is known that
the Bushmen eat the roots of C. welwitschii and not its leaves, whereas they do not
eat the bitter roots of C. hainesii but eat its leaves as a salad. The Bushmen know the
differences between the two quite well. In addition, the plants grow perfectly true
from seed and retain all the morphological characteristics of the mother plant and
41
their ecology is also quite different. C. bainesii is a typical “ Kalahari plant ” in its
distribution; it occurs in South West Africa only in the Okavango and in the Groot-
fontein, Otjiwarongo, Gobabis, Okahandja and Keetmanshoop districts and is wide-
spread in the arid sandy areas of Bechuanaland, Griqualand-West and the Transvaal.
C. welwitschii is found in most of the remaining districts of South West Africa with the
exception of the Namib desert and extends northwards into the Belgian Congo but not
far to the east. It seems to avoid the arid sandy areas preferred by the other species.
3. C. bainesii {Hook, f.) A. Meeuse, comb. nov.
Rhynchocarpa bainesii Hook. f. in FI. Trop. Afr. 2: 564 (1871). Type: Chapman
& Baines s.n. (K, holo.!) from Norton Shaw Valley, N’gamiland (see notes).
Kedrostis bainesii (Hook, f.) Cogn., Mon. Cucurb. 644 (1881); Pflanzenreich 275.1:
150 (1916).
Corallocarpus sphaerocarpus Cogn. apud Schinz in Abh. Bot. Ver. Brandenb. 30: 151
(1888); Pflanzenreich 275 . 1 : 164 (1916). C. sphaerocarpus scaberrimus, send vslx.
haslatus Cogn. in Bull. Herb. Boiss. 3: 422 (1895); Burtt Davy, FI. Transv. 1: 226
(1926). C. dinteri Cogn., op. cit. 165 (1916).
Stems branched usually from near the base, usually slender except in lower portions,
slightly fleshy, geniculate at the nodes, glabrous or scabrid, in the older parts usually
as if covered with whitish scales, up to 1 m long. Tendrils glabrous, or sometimes
scabrid. Leaves somewhat rigid (slightly fleshy when fresh), in outline ovate or
suborbicular-cordate at the base, 3-8 (-12) cm long and broad, usually distinctly
palmately 3-5-lobed, either with oblong lobes which are somewhat constricted at the
base and rounded and mucronate at the apex, or subhastately 3-lobed with a large
central lobe and somewhat reflexed subrotundate basal lobes; the segments entire,
undulate or sublobulate, usually densely and shortly setose-hairy on upper surface,
glabrous or scabrid below or scabrid throughout; petioles puberulous to shortly
hispidulous or scabrid, rarely glabrous, 1-4 cm long. Male flowers: common peduncle
filiform, straight, glabrous, 1-3 cm long, up to 12-flowered, pedicels capillary, erecto-
patent, 1-2 mm long, flowers minute. Female flowers solitary to few-flowered, sessile
or almost sessile. Fruit sessile or almost so, ellipsoid or ovoid, smooth and ultimately
glabrous or shortly pubescent, rounded and often with a short apiculus at the apex,
often about 6-seeded, 10-16 mm long and 8-10 mm in diam. Seeds 4-5 x 2 '5-3 x
2-2-5 mm.
South West Africa. — Okavango: Schoenfelder 29 (PRE). Grootfontein: Schinz
303; 305 (Z, syntypes of C. sphaerocarpus)', S. of Grootfontein, Schoenfelder 925
(PRE); Tsumeb: Dinter 7583 (B); Otavi: Dinter 630 (PRE, SAM, type gathering
of C. dinteri Cogn.); 5292 (B, BOL, NH, PRE, SAM); ex Otavi, cult, at Okahandja,
Dinter 30^9 (SAM). Otjiwarongo: .finmarr/ 249 (SAM); Klein Waterberg, Okosongo-
mingo, Volk 217 (M); Okahandja: farm Quickborn, Bradfleld 317 (PRE). Gobabis:
Omitara, Liebenberg 4594 (PRE); Sandfontein, Wilman H. No. 1660 (KMG, SAM);
50 miles E. of Gobabis, de Winter 2475 (PRE). Keetmanshoop: 21 miles S.E. of
Aroab, de Winter 3439; Acocks 18087 (PRE).
Bechuanaland Protectorate. — Kachikau, Munro ML 9 (PRE) and grown from its
seeds in Pretoria, Meeuse 9640 (PRE); Ngami, on koppies. Van Son in T.M. 28804
(PRE); Ngamiland, Curson 573; Mochudi, Harbor in Hb. Rogers 6511 (BOL).
Cape Province. — Vryburg: Palmyra (60 miles N.W. of Vryburg), Rodin 3605A (BOL).
Barkly West: near Border, Acocks 1874 (KMG, PRE); Pniel, Acocks Hb. No. KMG
4226 (BOL, KMG); Newlands, Esterhuysen 1226 (BOL), Lewis Hb. No. 53464 (SAM);
Hay: Paauwfontein, Hb. No. 1405 (BOL, KMG); Herbert; Campbell, Wilman
Hb. No. 1405 (BOL, KMG); St. Clair, Douglas, Orpen in Hb. MacOwan 153 (SAM).
Kimberley; Doornlaagte, Power Hb. No. 4117 (BOL, KMG).
42
Transvaal. — Soutpansberg : Messina, Gerstner 5445 (PRE); near Soutpan, Obermeyer,
Schweickerdt & Verdoorn 30 (PRE); between Saltpan and Waterpoort, Stopp Ml
(M); Msekwa’s Poort, Gerstner 5947a (PRE); Bandolierkop, Gerstner 5681 (PRE).
Sibasa: Chi pise, Verdoorn 2009 (PRE). Waterberg: near Monte Christo, Codd 6598
(PRE). Warmbad : Rooiberg, farm Sandfontein, Forssman s.n. (PRE). Potgietersrus :
farm Doornpoort (Rehmann’s “ Klippan ”) near Grass Valley, Meeuse 9481 (PRE).
Lydenburg: neax Moxont, Codd & Dyer 112% {?RE). Pretoria: near Pienaar’s River,
Acocks 12470 (PRE). Kruger National Park: Stonwane, van der Schijff 1918 (PRE).
Barberton: Komatipoort, Schlechter 11812 (BOL, COI, GRA, Z); Burrt Davy 380
(BOL).
Natal. — Ingwavuma: Ndumu, Gerstner 3470 (NH).
In addition. Southern Rhodesia and Portuguese East Africa.
The type sheet (K, photo.! in PRE) contains a mixture of two species, the other
being Coccinia rehmannii Cogn. There can be no doubt that the description of
Rhynchocarpa bainesii Hook. f. applies to the portion that, though poor, is identical
with Corallocarpus sphaerocarpus Cogn., and a new combination is required.
8. RAPHANOCARPUS
Raphanocarpus Hook.f., Icon. PI. 11 : 67, t. 1084 (1871); FI. Trop. Afr. 2: 540 (1871);
Cogn., Mon. Cucurb. 426 (1881); Pflanzenreich 275 .2: 56(1924); Pax in Pflanzenfam.
4, 5: 25 (1889); Chiov., FI. Somal. 1: 181 (1929); Phill., Gen. ed. 2: 718 (1951).
Momordica sect. Raphanistrocarpus Baill. in Bull. Soc. Linn. Paris 1: 309 (1882);
Hist. PI. 8: 407, 442 (1886).
Raphanistrocarpus (Baill.) Pax in Pflzenfam. 4.5: 25 (1889); Cogn., op. cit. 53 (1924).
Perennials with tuberous rootstock forming scandent or prostrate annual herbaceous
or occasionally perennial woody stems, or annual herbs. Leaves petiolate, rarely sessile,
usually ovate or cordate-orbicular, entire or slightly lobed. Tendrils simple. Flowers
rather large, monoecious; male flowers in axillary pedunculate few-flowered racemes;
common peduncle sometimes adnate to the petiole of the subtending leaf and, if so,
usually exserted beyond its lamina ; pedicels usually bracteate at the base ; receptacle
shortly campanulate to shallowly cup-shaped with 5 incurved scales at the base within ;
sepals ovate, elliptic or lanceolate, acute or acuminate; corolla-lobes elliptic, acute
or acuminate, orange or yellow; stamens 3 or 4 or occasionally 2; filaments broadly
linear, free; anthers 3 and, if so, two 2-thecous and the third 1-thecous, or 4 (3 1-thecous,
1 2-thecous) or 2 (1 3-thecous, 1 2-thecous); thecae linear, often flexuous or curved;
connective narrow or broad, not produced beyond the anthers; rudimentary pistil
represented by a gland (nectary) or 0; female flowers axillary, solitary or occasionally
in pairs; pedicel usually not adnate to the petiole of subtending leaf; receptacle calyx
and corolla as in the male but as a rule smaller; staminodes none; ovary slender,
fusiform, sulcate, 1 -chambered, with 2-5 partly pendulous, partly erect (or all erect)
ovules; style short, slender; stigma capitate, entire or bilobed. Fruit dry, corky,
fusiform, slender, 1-locular or sometimes with 2 or more loculi formed by false transverse
septa, indehiscent or irregularyl longitudinally splitting. Seeds few, ellipsoid or linear-
elliptic in outline; testa crustaceous, smooth; tegmen thin, hyaline; cotyledons more
or less obovate, one slightly folded over the other; radicle, conical, acute.
Type species: R. kirkii Hook. f. (the only one originally described with the
genus).
Two species (5 described ones) in southern tropical and subtropical Africa (both
extending into South Africa) and, in addition, one species in Somaliland.
43
Cogniaux and Harms (1924) maintained the two genera Raphanocarpus Hook. f.
and Raphanistrocarpus (Baill.) Pax, although Harms in a note (p. 54) already stated
that they perhaps had better be united. However, the plants referred to these two
genera are very similar in general appearance and it is felt that they cannot be generically
distinct, a conclusion already reached by Chiovenda. Phillips (Gen., p. 748) mentions
only one genus {Raphanocarpus) and the description given there appears to apply to
Raphanocarpus s.s. However, he must have meant to include Raphanistrocarpus,
because the number of species given is 5. The omission of Raphanistrocarpus (even
as a synonym) may have been through an oversight, because the two names are so
very similar.
Male peduncles adnate to the petiole of the subtending leaves; female peduncles very short
1. R. welwitschii
Male peduncles free to the base; female peduncles usually elongate, up to 8 cm long 2. R. boivniii
1. R. welwitschii Hook. f. in FI. Trop. Afr. 2: 541 (1871); Cogn., Mon. Cucurb.
427 (1881); in Bot. Jahrb. 10: 270 (1888) and in Pflanzenreich 57 (1924); Hiern, Cat.
Welw. Afr. PL 1 (2): 393 (1898); Dinter in Fedde, Repert. 23: 133 (1926/1927). Type:
Angola, Mossamedes, Welwitsch 790 (K, holo.); R. humilis Cogn., op. cit. 149
(1888); 57 (1924). R. humilis Cogn. var. prostratus Susseng. in Mitt. Bot. Staatssamml.
Miinchen, H. 8: 342 (Dec. 1953).
Annual, glabrous to sparsely pubescent. Stems herbaceous, slender, terete,
sulcate, up to at least several meters long. Leaves firm to thinly coriaceous or sometimes
membranous, cordate-ovate to cordate-suborbicular, usually somewhat 5-angled to
3- cuspidate, cordate at the base with a narrow sinus and rounded usually overlapping
basal lobes, gradually acuminate into a narrowly triangular tip ending in a long subulate
mucro, entire or sinous to bluntly dentate or crenate; nerves slender, somewhat
prominent below; blade 3-9 cm long, 2-7 cm wide; petioles slender, flattened,
sulcate-striate, 2-9 cm, occasionally up to 12 cm long. Male flowers: free part of
peduncle terete, slender to filiform, usually slightly shorter than the subtending leaf,
much smoother than connate part; pedicels 4—12 mm long; bracts lanceolate, up to
13 mm long, narrowly acuminate; receptacle short; sepals lanceolate, much acuminate,
usually very sparsely pubescent, 10-20 mm long and 2-5-4 mm wide; corolla orange,
petals obovate-oblong, obtuse to subacute, 2 -5-3 -5 cm long. Female flowers: pedicels
short, in fruit somewhat incrassate and up to 20 mm in length; ovary narrowly fusiform,
thinly pubescent, about 15 mm long; sepals linear-subulate or lanceolate-subulate,
4- 5 mm long; petals 12-16 mm long. Fruit striate-ribbed, rostrate, glabrescent, 2- or
3-seeded, up to 6 cm long. Seeds nearly smooth, distinctly attenuate and finely
corrugated near the base, rounded at the apex, 11-12 mm long, 3 -5-4 -5 mm wide
and 1 • 5-2 mm thick.
Found in Angola, extending into Belgian Congo, Southern Rhodesia and South
West Africa.
Angola. — Membassaco: Cubal, Faulkner 132 (PRE); Capacca, Faulkner 361 (PRE).
South West Africa. — Kaokoveld: near Orawanjai, de Winter & Leistner 5643
(K, M, PRE); Ohopoho, de Winter & Leistner 5326 (K, M, PRE); between Franz-
fontein and Khorikasa, Belck 60 (Z). Ovamboland: Oshando, Schinz 314 (Z);
Ombahaka, Rautanen 492 (Z); Kamanyab, Thorne Hb. No. 31873 (SAM); near
Tsuwandes, de Winter 3057 (PRE). Outjo: Thorne Hb. No. 31787 (SAM); Pamela,
Volk 2893 (NH). Grootfontein: Namutoni, Breyer in T.M. 20657 (PRE); Namutoni
& Sandup, Barnard s.n. (SAM); Tsumeb: Marsh s.n. (PRE); Grootfontein, Schoen-
f elder S 407 (PRE). Okahandja: Otjihua, Dinter 460 (PRE, SAM); Okahandja,
Dinter 4617 (B); District unknown, but in northern part: Kaiantes, Thorne Hb. No.
31784 (SAM); Cayimaeis, Thorne Hb. No. 31786 (SAM). Swakopmund or Karibib:
“ Kuiseb ”, Fleck 498 (Z). Karibib: Karibib, Rautanen 512; 520 (Z, two of the numbers
44
cited by Cogniaux sub R. humilis Cogn.); Marloth 1291 (GRA); Dinter 6782 (B);
Kinges 3126 (PRE, cited by Sussenguth as R. humilis var. prostratus)\ Khan River,
Dinter 67 (Z); between Karibib and Okahandja, de Winter 2668 (PRE); probably
Karibib, Fleck 2a (Z); between Swakopmund and Okanhandja or Okahandja: Mrs.
£///o») in Herb. Rehmann (Z). Gobabis: Sfcyn O.P. 1541 (PRE); between
Gobabis and Windhoek, de Winter 2512. Localities not precisely known but in central
and northern areas: “ Hereroland ”: Mar/ot/; 8543 or 8573 (PRE); Nels 27 (Z);
“Tabakstuin ” : Dinter 247 (Z); possibly Windhoek distr. : Dinter s.n. (leg. 1911)
in SAM. Rehoboth: Schinz 315 (Z); between Rehoboth and Uhlenhorst, Wilman
450 (BOL, PRE); Naukluft Mts., Buellsport, Strey 2161 (PRE). Without precise
locality: Fleck 947 (Z); Nortier s.n. (BOL).
The type was not studied, but specimens referred to this species by Cogniaux
leave very little doubt about the identity of the species concerned, for example: Schinz
314, Marloth 1291, Fleck 2a, 498, 947, Rautanen 492, Dinter 247, 460.
Dinter (Neue u. wenig bekannte Pfl. D.S.W. Afr., p. 48 (1914) had already pointed
out that “ Raphanocarpus humilis Cogn.” is only a depauperate form of R. welwitschii.
A comparison of some specimens referred to R. humilis by Cogniaux viz., Rautanen
512 and 520, clearly shows that Dinter correctly reduced this species to a synonym of
R. welwitschii. However, in 1953 Suessenguth described a var. prostrata of “ R. humilis ”
of which a quoted gathering (Kinges 3126) was available for study. Almost needless
to say, this specimen and the supposed variety belong to R. welwitschii — R. kirkii.
2. R. boivinii (Baill.) Chiov., El. Somal. 1: 181 (1929).
Momordica boivinii Baill., Hist. PI. 8: 407, f. 289-291 (1886).
Raphanistrocarpus boivinii (Baill.) Cogn. in Engl., Pflanzenwelt O.-Afr. C. 397 (1895);
Pflanzenreich 54 (1924); Burtt Davy, FI. Transv. 1 : 224 (1926). R. asperifolius Cogn.,
Pflanzenreich 56 (1924).
Raphanocarpus tuberosus Dinter, Neue u. wenig bekannte Pfl. D.S.W. Afr. 48 (1914),
nomen nudum, et in Fedde, Repert. 23 : 133 (1926-1927) nomen tantum. R. asperifolius
(Cogn.) Chiov. l.c.
Perennial forming annual herbaceous stems from a tuberous rootstock. Stems
slender, deeply sulcate, glabrous or thinly hairy, up to 5 m long. Leaves membranous,
cordate-oblong, or triangular-ovate, undivided or slightly 3-5-lobed, often somewhat
angular, triangular-acuminate at the apex with the tip obtuse or subacute, distinctly
mucronate, cordate at the base with broadly rounded, free or overlapping basal lobes
and a narrow, rarely wide, basal sinus, remotely and often somewhat irregularly calloso-
dentate to coarsely dentate with the teeth calloso-mucronate ; 2-7 cm long and 1-2-5
cm wide; on both surfaces more or less densely and shortly setose-scabrid or later
finely punctate, more densely so and hence ciliate near the margin; nerves slender,
not or hardly prominent below; petioles slender, thinly hairy to glabrous, 0-8-2 -5
cm long. Male flowers: common peduncles filiform, terete, firm, smooth, glabrous or
rarely somewhat hairy, 2-8 cm long; pedicels capillary, usually puberulous, up to 12
mm long; bracts small, narrow; sepals lanceolate-triangular, acute, usually thinly
covered with short stiff hairs, 10-14 mm long; petals yellow or light orange, often with
dark centre, narrowly obovate, obtuse to subacute, 1 - 5-2 cm long. Female flowers:
peduncles slender, filiform, short or (in fruit) up to 8 cm long, with a subulate, 4-7
mm long bract at the apex; in fruit incrassate towards the apex and finely sulcate;
ovary linear-fusiform, in anthesis 10-12 mm long, usually shortly hairy mainly in lines
following the faint longitudinal ribs; sepals lanceolate or lanceolate-snbulate, usually
hairy, acute, 3-4 mm long; petals but slightly smaller than in the male flowers. Fruit
fusiform, attenuate at both ends, usually faintly striate-ribbed, ultimately often glabrous,
4-7 cm long and 4-6 mm in diam. Seeds subcylindric-fusiform, not or slightly
compressed, at both ends subtruncate and finely rugose, otherwise smooth, 9-13 mm
long, 3-5 mm in largest diam.
45
Type specimen: This must be a specimen collected by Boivin in Mombasa cited
by Cogniaux (now probably in P) — not seen.
Recorded from East Africa (Mombasa) to South West Africa, Southern Rhodesia,
Transvaal and Portuguese East Africa. Not yet recorded from Bechuanaland Protec-
torate (although to be expected there).
South West Africa.— Grootfontein: Otavi, Dinter 744 (PRE, SAM); 5366 (B,
BOL, NH, PRE, SAM); Tsumeb, Barnard 173 (SAM); Naegelsbach 7D (PRE);
Grootfontein, Boss s.n. (PRE). Otjiwarongo: Waterberg Plateau, Boss s.n. (PRE).
Okahandja: farm Quickborn, BradfieM 307 (PRE); Okahandja, Dimer 505 (SAM).
“Between Windhoek and Walvis Bay”: Esdaile in Herb. Rogers 15325 p.p. (J).
A’.fi.— The numbers Dinter 505, and 744 have been cited by Dinter (1926/1927) as
R. tuberosus and Dinter 5336 has been distributed under that name.
Transvaal. — Warmbaths: Rooiberg Valley, Kwarriekraal, Mogg s.n. Brits: farm
Welgevonden, Mogg 14639. Lydenburg: Griffen Mine, Breyer in T.M. 19725; near
Branddraai, Cw/(i 3268 ; near Ohrigstad, A 535 (all PRE). Letaba: Shilowane,
Breyer in T.M. 21460 (PRE); between Shilouvane and Spelonken, Jiinod 1489 (Z).
Nelspruit: Kaap Muiden, Mogg s.n. (PRE); Klokwene, von c/or S'c/;//^3228; M’Bianide
Spruit, van der Schijff' 2\%0 (all PRE). Sibasa: Punda Maria, van der Schijff 3604.
Portuguese East Africa.— Niasa: nr. Gurue, Hornby 4562 (PRE); Nampula, Torre
1411 (COI); Mandimba, Hornby 4509 (PRE).
Although the type specimen was not studied, Baillon’s figures cited above are
quite adequate to recognise this species.
This plant varies in degree of pubescence and the more densely hairy form described
by Cogniaux (1924) as Raphanistrocarpus asperifolius is not worthy of a separate
specific or even a varietal status.
9. MOMORDICA
Momordica L., Sp. PI. ed. 1: 1009 (1753); Gen. PI. ed. 5: 440 (1754); Ser. in DC.,
Prodr. 3: 311 (1828); Naud. in Ann. Sci. Nat. 4me. ser. 12: 129 (1859); Sond. in
FI. Cap. 2: 491 (1862); Benth. & Hook, f., Gen. PI. 1: 825 (1867); Hook. f. in FI.
Trop. Afr. 2: 534 (1871); Cogn., Mon. Cucurb. 427 (1881); in Pflanzenreich 8 ( 1924);
Bail., Hist. PI. 8: 441 (1886), partim; Pax in Pflanzenfam. 4, 5: 23 (1889); Phillips,
Gen. ed. 2: 748 (1951).
Mostly perennial, sometimes annual, often forming annual scandent or prostrate
stems from a tuberous rootstock. Leaves entire, lobed or palmately or pedately 3-9-
foliate. Tendrils simple, rarely bifid. Flowers monoecious or sometimes dioecious;
often rather large; male ones corymbose, subumbellately racemose or solitary, often
bracteate at the apex of the peduncle or lower down; receptacle shallowly cup-shaped,
campanulate or shortly funnel-shaped with 2-3 incurved scales decurrent from the
petals near the base inside, sometimes with a few additional smaller scales decurrent
from some or all of the stamens; sepals variable; corolla-lobes usually ovate, elliptic
or orbicular, sometimes hairy; stamens 3; filaments free or somewhat cohering;
anthers at first cohering, later free, two 2-thecous and the third 1-thecous; thecae
conduplicate; connective bilobed, not produced at the apex; rudiment of pistil 0
or represented by a gland; female flowers solitary; receptacle, calyx and corolla as in
the male; but scales in base of receptacle usually smaller; staminodes 0 or represented
by 3 glands at the base of the style; ovary oblong in outline or fusiform, with 3
placentas and many ovules; often echinate; style slender; stigmas 3, entire or bifid.
Fruit a dehiscent 3-valved or sometimes indehiscent berry; oblong, ellipsoid, fusiform,
ovoid or cylindric, smooth, echinate, with soft spines or with short protuberances,
few- to many-seeded, usually orange or bright red when ripe. Seeds thick or compressed,
smooth or variously sculptured; testa crustaceous; tegmen thin, membranous; coty-
ledons elliptic; radicle shortly conical.
46
Type species; M. charantia L.
A tropical and subtropical genus found in the Old World, the few species in
America most probably introduced; valid species about 80, mostly African.
Leaves palmately-pedately pluri-foliolate with 3 groups of 3 or more leaflets \. M. clematidea
Leaves undivided to palmatifid but not compound;
Leaves undivided to faintly lobed or angular, ovate-cordate or narrowly cordate to cordate-
triangular, minutely to rather coarsely denticulate or crenate-dentate; male flowers in
a few-flowered subumbellate raceme or solitary; ovary densely setose; fruit ellipsoid or
ovoid, bright orange-yellow, softly and densely setose-aculeate 2. M. foetida
Leaves more or less lobed to palmatifid;
Tendrils bifid; leaves membranous when dry, finely scabrid, scabrid punctate or smooth;
plant glabrous or nearly so 3. M. welwitschii
Tendrils simple or, if divided, leaves not drying membranous and plant pubescent at least
on the younger portions;
Tendrils always simple; petals of male flowers 12-20 (-24) mm long, those of female flowers
usually smaller than in the male flowers;
Bract of male flowers above the middle, usually a little below the calyx;
Petals of the male flowers light yellow, up to 15 mm long; calyx usually green; leaves
drying a pale somewhat yellowish green as a rule; larger lobes of older leaves with
usually 7 or more teeth or lobules 4. M. balsaniina
Petals of male flowers cream-coloured, green veined, the 3 larger ones with a dark
blotch at the base; calyx dark purple to black at least on the receptacle; leaves
dark green drying deep green or rather dark brown; larger lobes of older leaves
with 3-5, rarely up to 9, teeth or lobules 5. M. invohwrata
Bracts of male flowers below the middle 6. M. charantia
Tendrils, at least the older ones, usually bifid; petals of male flowers 20-30 mm long, those of
female flowers as long as or longer than in the male flowers 1. M. repens
]. M. clematidea Sond. in FI. Cap. 2: 491 (1862); Cogn., Mon. Cucurb. 434
(1881); Pflanzenreich 20 (1924); Burtt Davy, FI. Transv. 1: 227 (1926). Syntypes:
Burke 357 and Zeyher 578.
M. cardiospermoides Klotzsch in Peters, Reise Mossamb., Bot. 150 (1864); Flook. f.
in FI. Trop. Afr. 2: 535 (1871).
Perennial. Stems herbaceous, angular, glabrous, up to several meters long.
Leaves biternately multifoliolate; petiole slender, glabrous or bearing tufts of hairs
at the base and the apex, under 2 cm long; petiolules 3; of which the middle one is
1- 5 cm long, its leaflet pinnately 5-7-foliolate, occasionally subpinnatifid, 1-6 cm
long; the lateral petiolules shorter, their leaflets alternately pinnately 3-5-foliolate,
up to 7 cm long including the petiolules; secondary petiolules 2-8 (-25) mm long;
ultimate leaflets membranous, ovate or oblong, acute, rounded at the base, crenate-
dentate with mucronulate teeth, glabrous, dark green and finely punctate-scabrous
above, paler and smooth below, 0 ■ 5-3 cm long and 0 ■ 4-2 cm wide. Tendrils filiform,
glabrous, simple or bifid. Flowers all solitary. Male flowers: peduncles filiform,
glabrous, bracteate at the apex, 5-15 cm long; bract sessile, concave, suborbicular,
entire, emarginate or subcordate at the base, 1-2 cm wide; receptacle subrotate,
glabrous, 6-7 mm in diam., sepals ovate, obtuse to subacuminate, 5-7 mm long, 4-5
mm wide; petals yellow, a dirty green near the base inside, 1 •5-2-5 cm long. Female
flowers: peduncle ebracteate or with a small elliptic or oblong bract at the apex, 1-5
cm long; Ovary ellipsoid-fusiform, soon glabrous, smooth; sepals ovate-triangular,
2- 3 mm long; corolla sometimes smaller than in the male flowers. Fruit fleshy, ovold-
oblong or oblong, acute, rounded at the base, smooth, orange red, 5-9 cm long and
2-5-5 cm in diam. Seeds dark brown, ovate, much compressed, with a few tubercles,
12-13 mm long, 8-9 mm wide and 3-5 mm thick.
Found from Northern Rhodesia and Tanganyika to Zululand, the Transvaal and
Bechuanaland; fairly common, especially on sandy soil.
Recorded from the following districts; Transvaal'. Soutpansberg to Barberton
in the east, and Lydenburg, Bronkhorstspruit, Pretoria, Brits and Rustenburg in the
south and west; Natal: Hluhluwe {Ward 1911, NH), Verulam.
47
The following specimens are of special interest: Burke 357 (SAM), duplicate of
one of the twin-types from “ Crocodile River ” (probably Brits district) in the Transvaal
and several specimens cited by Burtt Davy, namely Burtt Davy 2035 from Warmbaths
(PRE); Galpin 705 from Barberton (PRE); Rogers 22508 from Barberton (PRE).
The reduction of Momordica cardiospermoides Klotzsch to M. clematidea was first
made by Cogniaux in 1881. The type of the first (B) is now destroyed, but specimens
from East Tropical Africa agree in every respect with those from the southern Transvaal
and there can be no doubt that Cogniaux’s reduction was correct.
2. M. foetida Schum. & Thonn., Beskr. Guin. PI. 426 (1827); Cogn., Mon. Cucurb.
451 (1881); in Pflanzenreich 41 (1924); Burtt Davy, FI. Transv. 1 : 227 (1927); Hutch.
& Dalz., FI. W. Trop. Afr. 1 : 181 (1931); Andrews, Flow. PI. Anglo-Egypt. Soudan 1 :
181 (1951). Type: Thonning 85 from the Guinea Coast.
Momordica morkorra A. Rich., Tent. Fl. Abyss. 1: 292, t. 53 (1847); Naud. in Ann.
Sci. Nat. 4me. ser. 12, p. 134 (1859); Hook. f. in Fl. Trop. Afr. 2: 538 (1871). M.
schimperiana Naud., l.c. 23 (1867); Cogn., l.c. 453 (1881); 40 (1924); Hiern, Cat.
Welw. Afr. PI. 1 (2): 394 (1898); Andrews, l.c. (1951). M. cordifolia E. Mey. ex
Sond., Fl. Cap. 2: 492 (1869); Naud., op. cit. (1867), p. 22; Pegler in Ann. Bolus
Herb. 2, p. 117 (1917). M. cucuUata Hook, f., l.c. (1971). M. cordata Cogn. in
Bot. Jahrb. 21: 208 (1895), ex descr. M. angustisepala Harms in Bot. Jahrb. 58:
239 (1923); and in Pflanzenreich 41 (1924).
Cucumisl cordifolius E. Mey. ex Drege, Zw. Pflzgeog. Doc. p. 149, 176 (1843), nomen
tantum.
Eulenburgia mirabilis Pax in Engl. Bot. Jahrb. 39, p. 654 (1907).
Perennial, dioecious or occasionally monoecious. Stems rather stout, glabrous
or occasionally hairy, sulcate very often marked with small linear dark spots, up to
5 m long. Leaves ovate-cordate or narrowly cordate to cordate-triangular, acute or
acuminate, with a usually broad and shallow cordate to truncate base in the centre
decurrent into the petiole, and a minutely to rather coarsely denticulate or crenate-
dentate margin, glabrous to hairy (especially so to subtomentose on lower surface)
6-16 cm long and 4-13 cm wide, petioles usually firm, 3-9 cm long. Tendrils simple
or bifid. Male plant: common peduncle fairly slender, angular-sulcate, subumbellately
few-flowered, 3-15 (occasionally up to 25) cm long, bracteate at the apex; bract varying
from small and inconspicuous to large, up to 3 cm long and 5 cm wide (in S. Africa
usually small), subreniform to oblong, often folded, boatshaped or cucullate, glabrous
or hairy, entire or crenate to dentate; pedicels usually hairy, 0-5-7 cm long; calyx
usually dark; receptacle wide and short, usually somewhat (and variously) hairy;
sepals varying from broadly ovate to lanceolate, obtuse or sometimes acute, with a
thinner and usually ciliate margin, 5-10 mm long, 5-8 mm wide; petals yellow, usually
dark brown at the base inside, 2-3 ■ 5 cm long. Female plants: flowers solitary; peduncle
rather slender, with or without a small bract near the base, middle or apex, 2-12 cm
long (occasionally bearing in addition 1-3 male flowers); ovary ovoid-oblong or
ellipsoid-fusiform, densely and softly muricate-setose, rostrate or narrowed at the
apex, 15-20 mm long; calyx and corolla sometimes smaller than in the male. Fruit
ovoid or ellipsoid, bright orange yellow, softly and densely setose-aculeate, 5-8 cm
long, 3-5 cm in diam., irregularly dehiscent. Seeds dark brown, elliptic-oblong with
a thin prominent margin, faintly rugose, 10-12 mm long, 6-7 mm wide and 3 mm thick.
Found in Africa south of the Sahara, extends southwards to Angola, the Transvaal
and through Natal to the Eastern Cape Province.
Recorded from the following districts: Transvaal: Pietersburg, Letaba, Lydenburg,
Belfast, Pilgrims Rest, Nelspruit, Barberton; Swaziland; Natal: Nongoma, Hlabisa,
Weenen, Estcourt, Eshowe, Lower Tugela, Lions River, Pietermaritzburg, Inanda,
Durban, Richmond, Umzinto; Cape Province: Umzimkulu, Kentani, Komgha.
48
The following specimens are noteworthy: Cape Province. — Between Umtata and
Umsamwubo rivers (probably Port St. Johns distr.), “ Cucumisl cordifolius E. Mey.”
leg. Drege (L); Natal. — Umzinto: Dumisa, Rudatis 1027 (PRE); Transvaal. — Pieters-
burg: Houtbosch, 6312 (BR); Lydenburg: IT/Vmj' 486 (L, PRE) ; Barberton:
GaJpin 647 (BOL, PRE); 784 (PRE), all numbers cited by Cogniaux (1924) sub M.
foetida; Wilms 486; Rehmann 6312 and Galpin 647; and 784 also cited by Burtt Davy
(1926).
There is a considerable variation in the pubescence of this species. The extreme
forms [var. villosa Cogn. in Engl. Bot. Jahrb. 21, p. 208 (1895)] are connected by inter-
mediates with almost entirely glabrous ones so that this variety is not maintained here.
The tendrils can be simple or bifid and this character has no diagnostic value. The
importance of the variation in the development of the bract on the male peduncles
has been over-estimated, and several described species are merely forms which in
larger series of specimens are all connected by intermediates.
3. M. welwitschii Hook. f. in FI. Trop. Afr. 2; 538 (1871); Cogn., Mon. Cucurb.
435 (1881); Pflanzenreich 275.2: 23 (1924); Hiern, Catal. Welw. Afr. PI. 2; 393
(1898). Type; Wehvilsch 787 from Mossamedes (Angola) in BM.
Climbing herb. Stems slender, subterete, glabrous, smooth or longitudinally
sulcate, older internodes 4-7 cm long. Leaves herbaceous drying membranaceous,
in outline cordate-reniform to orbicular-cordate or broadly cordate with a wide and
shallow basal sinus and the leafbase decurrent in the middle along the petiole, the
apex acute to acutely acuminate; the blade 3-9 cm long and 2-9 cm broad, 5-angled
and often subtrilobed to more or less distinctly 5-9-lobed (to about the middle), the
lobes varying from triangular-ovate or elliptic-rhomboid to occasionally oblong,
usually acute, often faintly lobular or coarsely dentate with usually acute and mucronate
apices and lobes (or teeth), the margin minutely ciliate-scabrid; upper surface dark
green, minutely scabrid to smooth, lower surface pale green finely scabrid-punctate ;
venation fine, the secondary and tertiary clearly visible on the lower surface, obscure
on upper surface; sinuses between the lobes usually narrow and acute; petioles slender
usually dorsolaterally flattened and often winged in upper half by the decurrent leafbase,
glabrous, 1-7 cm long. Tendrils slender to filiform, glabrous or nearly so, bifid at
about 1-2 cm from the base. Flowers monoecious, solitary. Male flowers: pedicels
filiform, glabrous or puberulous near the apex, 4-7 cm long, ebracteate or with a minute
bract near the middle or in upper half; calyx dark, black when dried, receptacle shallow
4-6 mm long, 6-10 mm in diam.; sepals oblong or elliptic to ovate-orbicular 5-6 mm
long, minutely ciliate, otherwise glabrous; corolla cream-coloured to pale yellow
with 3 black spots on the base of larger petals inside, nearly glabrous, 3-4 cm in diam.;
the largest petals 15-20 mm long, broadly ovate, obtuse or subacute. Female flowers:
pedicels as in the male, but up to 5 cm (usually under 4 cm) long, mostly ebracteate
and slightly thickened towards the apex in fruit; ovary ellipsoid or ovoid, in open
flower about 6 mm long and 4 mm in diam., sparsely covered with cylindric papillae
about 0-75 mm long, receptacle above the ovary and calyx smaller than in the male,
together about 1 • 5 mm high and 4 mm in diam. ; sepals black, broadly rounded at
the apex 1-2 mm long; corolla as in the male but only its size. Fruit resembling
that of M. halsamina, broadly ovoid-ellipsoid or somewhat angular-ovoid, somewhat
pointed or conical at the apex, glabrous, with a few longitudinal rows of sparse conical
cylindric protuberances, ultimately orange or red when mature; 2-5-3 cm long and
2-2-5 cm in diam. Seeds elliptic-discoid, light brown, with rugulose-wavy ridges
along and near margin on flat surfaces, very shortly produced-stipitate at umbilical
end, 9-10 mm long, about 7 mm broad and about 2 mm thick.
49
Found in southern Angola and Kaokoveld (South West Africa).
South West Africa.— Kaokoveld: Kapupa Valley, Story 5839 (PRE); 30 miles S.
of Kunene on road to Orupembe, de Winter & Leistner 5789 (PRE); near Ohopoho,
de Winter & Leistner 5279 (PRE).
4. M. balsaminaL., Sp. PI. ed. 1 : 1009 (1753); Ser. in DC., Prodr. 3: 311 (1828);
Sond. in FI. Cap. 2: 491 (1861); Hook f. in FI. Trop. Afr. 2: 537 (1871); Cogn.
Mon. Cucurb. 439 (1881); Pflanzenreich 28 (1924); Burtt Davy, FI. Transv. 1: 227
(1926); Hutch. & Dalz., FI. W. Trop. Afr. 1: 181 (1931); Range in Fedde, Repert.
38: 273 (1935); Andrews, Flow. PI. Anglo-Egypt. Soudan 1: 181 (1950).
M. gariepensis E. Mey. ex Drege, Zw. Pflzgeog. Doc. 202 (1843), nomen tantum; ex
Sond., I.C., in syn. M. schinzii Cogn. in Abh. Bot. Ver. Brandenb. 30: 149 (1888);
Pflanzenreich 30 (1924); Range in Fedde, Repert. 38: 273 (1935).
Perennial herbaceous climber. Rootstock tuberous. Stems several to many,
slender, sulcate, usually soon glabrous, up to 2 m long and over, occasionally attaining
10 m. Leaves suborbicular in outline, membranous or thinly herbaceous, glabrous
or slightly hairy mainly on the larger nerves below, smooth or minutely punctate,
sometimes scabridulous, 3-7 (-10) cm in diam., 5-7-lobed to about the middle; the
lobes rhomboid or obovate to eliptic-rhomboid obtuse and long-mucronate to acute,
coarsely and irregularly dentate-lobulate with in the larger leaves usually 7 or more
teeth or lobules; lowermost leaf-segments considerably smaller; petioles slender,
usually glabrous, sometimes shortly setose-pubescent 1-3 cm long. Tendrils slender,
simple, glabrous or somewhat hairy. Flowers monoecious, all solitary. Male flowers:
peduncles filiform or very slender, glabrous or usually somewhat hairy towards the apex,
up to 5 cm occasionally to 9 cm long; bract sessile at or just below the apex of the
peduncle, sub-membranous, whitish, green-veined, ovate to orbicular or reniform,
cordate at the base, 5-10 mm long and 8-15 mm broad, entire or finely dentate, acute to
mucronate-aristate; calyx green or occasionally purplish-black, often softly and rather
densely hairy; receptacle shallow, 2-4 mm high and 5-8 mm in diam., sepals ovate to
elliptic, acute and aristate-mucronate, 5-7 mm long and 3-5 mm broad; corolla pale
yellow; petals unequal (outer 2 slightly larger), ovate to obovate, acute to cuspidate
or rounded to emarginate and mucronulate. 7-15 mm long and 5-12 mm broad.
Female flowers: peduncle ebracteate or bracteate near the base or the middle, 0-5-3
cm long; ovary oblong-fusiform or ovoid-fusiform, rostrate, verrucose, glabrous or
softly hairy with curled woolly hairs; sepals narrowly triangular to ovate-triangular,
acute to acuminate, 2-3 mm long; corolla as in the male but often smaller. Fruit
subglobose to ovoid with a broad conical rostrum and abruptly and shortly attenuate
at the base, bright orange-red to scarlet when ripe, rather sparsely muricate or tubercled,
2-5-5 cm long and 2-4 cm in diam. Seeds with a carmine-red arillus, brown, ovate
or oblong to elliptic in outline, much compressed, finely verruculose on the large faces
and marked with sinuous raised ridges which usually form an elliptic figure and extend
to the edges often making the latter appear finely corrugated, grooved on the lateral
faces, 8-12 mm long, 5-7 mm broad and 1-5-2 mm thick.
Type specimen: Linne (1753), in the original diagnoses, cited several “ pre-
Linnaean ” publications, among them the Pinax 306 by Bauhin. The species is also
represented in the Linnaean Herbarium and this specimen is considered to represent
the type.
Found in tropical and subtropical Africa, extending to north west India; introduced
elsewhere in the tropics.
50
Recorded from the following districts: South West Africa: wide-spread except
in the drier areas along the west coast, recorded from as far south as Gibeon and
Rehoboth; Bechuanaland Protectorate; Transvaal: Soutpansberg, Sibasa, Pietersburg,
Potgietersrust, Waterberg, Rustenburg, Warmbaths, Pretoria, Brits, Bronkhorstspruit,
Marico, Lichtenburg, Potchefstroom, Wolmaransstad, Nelspruit, Barberton; Cape
Province: Namaqualand near the Orange River (Drege), Prieska, Vryburg, Mafeking;
Orange Free State: Hoopstad; Natal: Lower Tugela, Estcourt, Weenen, Lions River.
The following specimens are noteworthy : a specimen leg. Drege at “ Verleptpram ”
near the Orange River in North-west Namaqualand (L, PRE,) labelled “ Momordica
gariepensis E. Mey.” South West Africa, Unkuambi (Ovamboland), Rautanen 491,
686 (Z); Rehoboth, S'c/n’rtz 310 (GRA, Z); Oshiheke, 31 1 (BR, Z); Okahandja,
Dinter 330 (GRA, L, SAM, Z), and sine loco. Fleck 7a and 161 (Z). Transvaal, Phoberg,
Holub (Z) — all these numbers cited by Cogniaux (1924) under M. schinzii Cogn. of
which Schinz 310 and 311 in (Z) are the syntypes. Transvaal, “ Vaal River” and
“ Magaliesberg ” Burke 81 (PRE, SAM), Zeyher 595 (SAM), both cited in Flora
Capensis and by Burtt Davy as M. balsamina.
Momordica schinzii is inseparable from M. balsamina, the distinguishing character
indicated by Cogniaux (1924), i.e. male bracts entire or dentate, is useless.
5. M. involucrata E. Mey. ex Sond. in FI. Cap. 2: 491 (1862); Cogn., Mon. Cucurb.
440 (1881); Pflanzenreich 275.2: 30 (1924); Curtis, Bot. Mag. t. 6932 (1887); Wood,
Natal PI. 6: t. 516 (1909). Type: Drege s.n. from Durban in Herb. Sonder (S, holo.!,
L, iso.!).
M. balsamina L. var. leucantha Ch. Huber, Cat. 6 (1864), nomen. M. huberi Tod.
in Giorn. R. Inst. Agric. Sicilia 165 (1863). M. balsamina L. var. huberi Naud. in
Ann. Sci. Nat. 5 me ser. 5: 21 (1867).
Perennial monoecious herbaceous climber. Stems rather slender, glabrous or nearly
so, up to several meters long. Leaves in outline cordate-suborbicular to pentagonal,
3-5-lobed to about the middle with obovate or subrhomboid and usually long-mucronate
to apiculate lobes, dark green above; slightly lighter below, drying deep-green to dark
brown on both surfaces, glabrous or nearly so, smooth, 3-6 (-9) cm long and broad;
lobes with usually distinctly constricted base, dentate-lobulate with 3-5 or occasionally
up to 9 teeth or lobes, rarely only remotely calloso-dentate; the lowermost leaf-segments
always much smaller than the other ones; petioles thinly pubescent to glabrous,
1-4 cm long. Tendrils simple. Male flowers solitary; peduncles filiform, usually
glabrous below and thinly pubescent towards the apex, 3-7 cm long, bracteate at the
apex by a sessile, broadly reniform-cordate sub-entire, usually emarginate and mucronate
subglabrous ciliate bract 8-12 mm long and 12-20 mm broad; receptacle and sometimes
the sepals blackish, the first 3-7 mm long and 6-10 mm wide at the mouth, the latter
ovate-sub-orbicular, obtuse to rounded sometimes acuminate or abruptly apiculate-
aristate by an up to 3 mm long apiculus, glabrous or puberulous, 5-10 mm long and
3-8 mm broad near the base; corolla white or pale cream, green-veined with 3 dark
blotches on base of longer petals inside, asymmetrical; the lobes obovate or sub-
orbicular-obovate, 15-24 mm long, the two large ones 12-13 mm broad, the other
3, 8-10 mm broad. Female flowers solitary on a 3-4 cm long pedicel which is bracteate
near the base; sepals triangular or elliptic-lanceolate acute, 2-3 mm long; corolla
usually smaller than in the male flower but occasionally up to 21 mm long; ovary
oblong fusiform, rostrate verrucose. Fruit orange-red when ripe, ovoid sub-globose
often somewhat angular (hexa- or octogonal), apiculate, with some conical protuberances,
dehiscent, 3-5 cm long, 2-3 cm in diam. Seeds brown, compressed, on the flat faces
finely pitted-grooved as if corroded, 9-11 mm by 5-5-7 by 1-5-2 mm.
51
Cape Province. — Komgha: Kei Mouth, Flanagan 165 = 1464 in Herb. MacOwan
(GRA, PRE, NBG). Kentani: Mazeppa Bay, Theron 12 (PRE). Port St. Johns:
Mogg s.n. (PRE).
Natal. — Umzinto: Ifafa coastal bush, Handley 56 (NU); Scottburgh, Mauve 1005
(PRE); Umkomaas, Gilmore Hb. no. 20952 (NH). Pinetown: Winkle Spruit, Lansdell
s.n. (NH, PRE); Doonside, Wylie Hb. No. 23307 (NH). Durban: Isipingo Beach,
Ward 729 (NU, PRE); 730 (NU); Forbes & Obermeyer 54 (NH, PRE); “ Umlaas
Height”, Drege s.n. (L. iso.!). Durban, Krauss 90 (PRE); Rehmann 8841 (BR);
Schlechter 1196 {GRK)\ lUoor/ 130 (BOL, GRA, NBG, NH); 655 (NH); 3093 (BOL);
Barker 3564 (NBG); Berea, Wood 11093 (NH, PRE); 11386 (J, NH, NU, PRE);
Forbes 42 (NH); Pillans s.n. (NBG). Inanda: Wood s.n. (GRA). Lions River:
Howick, Shafton 109 (GRA). Mtunzini: Tugela Beach, Johnson 372 (NBG). Lower
Umfolozi: Umsundusi River, Gerstner Hb. No. 23015 (NH). Hlabisa: False Bay,
Gerstner 4124; 4725; 4726 (PRE); E. of Nyalazi River, 3043 (NH, NU, PRE);
St. Lucia, Lansdell 72 (NH). Ubombo: Mkuzi, Galpin 13327 (PRE). Ingwavuma:
Ngamini Pan, Gerstner 3406 (NH), 3466 (NH, PRE).
Also in the coastal areas of the southern part of Portuguese East Africa.
Apparently a species with ecological requirements altogether different from M.
balsamina, the latter being a species of bushveld (savannah) vegetation, generally
occurring over 1,500 ft. altitude, and M. involucrata a species restricted to the coastal
bush and not going far inland.
6. M. charantia Z.., Sp. PI. ed. 1 : 1009 (1753); see Cogn., Pfianzenreich 24 (1924),
for full synonymy.
This plant, which is wide-spread in tropical Africa, apparently does not occur
in a wild state in the area considered in the present paper. The few specimens of the
species seen (from Natal) are obviously escapes from vegetable gardens of Indians
who use this plant in their curries (some of the field-notes even suggest this source,
e.g. on a specimen leg. Gerstner from Umzinto). The specimen Hdpfner 99 from
Damaraland cited by Cogniaux under M. charantia var. abbreviata Ser. I have not seen,
but it is doubtful if this was correctly named. The only species of Mornordia I have
seen from South West Africa is M. balsamina {M. schinzii). There are no Indian
settlements in South West Africa nor were there any in Hopfner’s time, and as far as
can be ascertained M. charantia has never been cultivated in South West Africa, so that
Hopfner’s specimen cannot have been a garden escape. It is more likely that this
specimen belongs to M. balsamina. At any rate, M. charantia is not considered to be
an indigenous element of our South African flora.
7. M. repens Bremek. in Ann. Transv. Mus. 15: 261 (1933). Type: Bremekamp
& Schweickerdt 34 (PRE, holo.!).
M. marlothii Harms in Fedde, Repert. 36: 269 (1934).
Perennial, monoecious, prostrate or occasionally scandent herb. Stems up to
about 2 m long, firm and usually rather stout, in sicco deeply sulcate, more or less
densely covered with a pubescence of rather short pilose to hirsute hairs; internodes
usually 3-5 cm long. Tendrils bifid at the apex when mature pubescent like the stems,
the basal undivided portion usually rather firm, sulcate, the apical branches filiform,
usually slender. Leaves herbaceous, when fresh slightly fleshy, drying papery to a
peculiar yellowish green or a pale olive-green, reniform-orbicular in outline, 3—7 cm
in diam., usually more or less deeply lobed and the lobes again lobulate (the lobules
acute or obtuse, mucronulate) ; basal sinus deep (usually about half the length of the
leaf); both surfaces scabrid or shortly setose-scabrid, the margin ciliate; petioles terete,
densely pubescent, striate, 2-6 cm long. Male flowers solitary or occasionally in few-
flowered subumbellate cymes; peduncle densely pubescent, striate, up to about 6 cm
52
long, with a subreniform, more or less 3- or 5-lobed, 6-9 mm long and 9-14 mm wide,
rarely small, subcordate bract near the middle in solitary flowers and at the apex if
flowers racemose; pedicels resembling the peduncle but usually more slender, up to
3 • 5 cm long; receptacle pubescent, short, about 8 mm in diam.; sepals ovate or oblong,
pubescent, dark, lighter towards margins and apex, apiculate, 7-10 mm long and 4-5
mm wide; petals pale yellow, dark veined, broadly ovate or obovate, 2-3 cm long,
pubescent. Female flowers solitary; peduncles rather stout, 3-7 cm long, in the middle
with a bract like the male peduncles or ebracteate; ovary rostrate from an ellipsoid or
ovoid-fusiform basal portion, densely and shortly pubescent and in widest portion
muricate, 15-25 mm long; corolla often larger than in the male, up to 35 mm long.
Fruit subglobose or somewhat ellipsoid or depressed, rostrate, muricate with large
flattened acutish protuberances and with 10 raised ribs, apparently indehiscent, reddish-
brown when ripe, 4—7 cm in diam. and 5-6 cm long; rostrum up to 1 cm long. Seeds
pale, yellowish, orbicular in outline, compressed with flat, or concave, convex or
irregularly bulging sides and 2 parallel peripheral, usually irregularly sinous, angular
or incised, ridges, about 20 mm in diam. and 7-8 mm thick; testa hard and bony.
Endemic in Southern Africa.
Bechuanaland Protectorate. — Palapye, Marloth 3329 (isotype of Momordica mar-
lotliii Harms, PRE); Mahalapye, Mansergh Hb. No. 25586 (BOL); Mochudi, Harbor
in H. Rogers no. 6512 (BOL); nr. Derdepoort (Transvaal border), Codd 8883 (PRE).
Transvaal. — Rustenburg: Matlabas River, 5 miles S. of Limpopo, Mogg 24584
(J, PRE); near Makoppa, Codd 8649 (PRE); near Thabazimbi, Principal of Potchef-
stroom Agr. College (cultivated at Prinshof, Pretoria, PRE). 40 miles N.W. of
Thabazimbi, Story 6048 (PRE). Waterberg district: near Monte Christo, Codd 6597
(PRE); 12 miles E. of Ellisras, Meeuse & Strey 10450 (PRE). Potgietersrust : nr.
Potgietersrust, Leendertz Hb. No. 6016; 6568 (not 6068 as cited by Bremekamp;
PRE); Rogers 1316 (GRA); Thode A 1703 (PRE, NH); Maguire 1297 (NBG); Meeuse
9458 (PRE). Pietersburg; Brakrivier, Bremekamp & Schweickerdt 34 (PRE, type!);
Magalakwin River, Salt Lake, Hutchinson 2655 (PRE). Bronkhorstspruit : farm
Rooikop, Smuts & Gillett 3031 (PRE). Without precise locality: “ Lekker Kraal ”
(most probably Transvaal), P. Krantz Hb. No. 8217 (PRE); N. Transvaal: Pole Evans
2522 (PRE).
The fact that this rather striking plant has remained undescribed for such a long
time can be explained only by its relatively small area of distribution and its ecological
requirements (deep dry sandy soil) so that, although often locally frequent, it is not
very common.
10. ACANTHOSICYOS
Acanthosicyos Welw. ex Hook. f. in Benth & Hook., f., Gen. PI. 1 : 824 (1867); Welw.,
Sert. Angol. in Trans. Linn. Soc. 27: 30 (1869); Hook. f. in FI. Trop. Afr. 2: 531
(1871); Cogn., Mon. Cucurb. 418 (1881); Pflanzenreich 275.2: 4 (1924); Baill.,
Hist. PI. 8: 442 (1886); Pax in Pflanzenfam, 4, 5: 23 (1889); Phillips, Gen. ed. 2:
747 (1951).
Perennial erect or ascending rigid shrub, much branched from the base. Leaves
reduced to minute scales. Tendrils 0. Spines paired at the nodes. Flowers dioecious,
sessile or subsessile. Male plants: flowers fasciculate or solitary; receptacle wide;
sepals firm, often unequal, petals rather coriaceous, broadly ovate; stamens 3 or 5,
inserted in the mouth of the receptacle; filaments short and thick; anthers broad,
one I -locular and two 2-locular or all 1 -locular; thecae strongly flexuous; connective
53
dilated, not produced at the apex; rudimentary pistil rarely present. Female plants:
flowers solitary; perianth as in the male; staminodes 5, elongate; ovary ovoid with
3-5 placentas; ovules numerous, horizontal; style columnar with 3-5 2-horned, flat
or capitate stigmas. Fruit subglobose, indehiscent, many-seeded, covered with hemis-
pherical conical spine-tipped protuberances. Seeds oblong’ or ovate in outline, not
much compressed, immarginate, testa smooth, thick, hard.
Type species: Monotypic.
Found in the dry sandy desert area near the coasts of southern Angola (Mossamedes)
southwards to both banks of the Orange River and north-western Namaqualand.
A. horrida Welw. ex Hook, f., l.c. (1867); Sert. Angol. 31, t. 11, 11a (1869);
Hook, f., l.c. (1871); Cogn., op. cit. 419 (1881); 6 (1924); Marloth in Bot. Jahrb. 9:
173, t. 3 (1887); Flora of S. Afr. 3, 2: 205, f. 88-91, PI. 53, f. A (1932); Engl, in Bot.
Jahrb. 19: 151 (1891); Hiern, Cat. Welw. Afr. PI. 1, 2: 392 (1898); Dinter in Fedde,
Repert. 15: 81 (1917); Fedde, Repert. Beih. 3: 84, 86 (1921); Range in Fedde, Rep.
30, t. 129 (1932); 38: 272 (1935). Type: Wehvitsch 806 (BM, holo., COI, iso.!).
Shrubby, 0-5-1 m tall, forming dense and large bushes on small dunes which
the plant builds up itself. Rootstock woody. Stems robust, terete, pale yellowish
or glaucous to greyish green, sulcate; twigs subvirgate, pubescent in the youngest
parts, the floriferous ones subtomentose. Spines terete, straight or nearly so, 2-3 cm
long, glabrous. Male flowers greyish-tomentose outside; sepals ovate to suborbicular-
obcordate, 4-6 mm long; petals furfuraceous inside, 10-12 mm long, pale yellow.
Female flowers: peduncle in fruit incrassate, ultimately attaining a length of about
1 cm; staminodes erect, linear, dilated near the apex, 5-7 mm long; ovary 15-17
mm long, densely covered with puberulous oblong-conical soft 2-2-5 mm long spines;
style 9-10 mm long, stigma 6-7 mm wide. Fruit pale orange-yellow when ripe, with
orange-yellow pulp, up to about 20 cm in diam. Seeds nearly white, sub-obliquely
truncate near the base, 14—15 mm long, 9-11 mm wide, 6-7 mm thick.
Angola. — Mossamedes: coastal dunes between Porto de Pinda and Banza de Caroca,
Wehvitsch 806 (COI, isotype!) ; Rio Coroca, Capello & Ivans 4 (COI); Cumilunga,
Exell & Mendonca 2225 (COI); Mossamedes, Lapa & Faro s.n. (COI); Tiger Bay
(Bahia dos Tigros), Newton s.n. (COI).
South West Africa. — Kaokoveld: Namib near Sarasas, Hall 408 (NBG). Swakop-
mund and surrounding area: Dinter 2806 (PRE, SAM); Moss 17895 (J); Bradfleld
466 (PRE); Wiss 956 (PRE); Walvis Bay, Marloth 1179 (BOL, BR, GRA, L, PRE,
SAM); Cleverly in Herb. MacOwen 1462 (GRA, PRE, SAM); Galpin & Pearson
7450; 7462; 7472 (PRE); Pole Evans Hb. no. 19333 (PRE), Rodin 2136 (BOL, PRE)
no. 2231 in herb. Strey (PRE), Jensen s.n. (PRE, a beautiful fasciation), Chaplin s.n.
(NBG), de Winter 3183 (PRE), Merxmueller & Giess 1748 (PRE, M). “ Kuiseb ” near
Walvis Bay, Strey “82”, 2583 (PRE). Luederitzbucht : Kovis Mtns., Dinter s.n.
(SAM), 6674 (B, BOL, PRE, SAM); Orange River, N. side, 1-5 miles from mouth,
Pillans 5605 (BOL, PRE); Daberas Drift (on both banks). Range 1569 (SAM).
Cape Province. — Richtersveld : Daberas Drift: Range 1569 (SAM), see above.
Namaqualand: nr. Aughrabies, 15 miles E. of Port Nolloth, Marloth s.n. (PRE).
The first record of this plant is from Alexander in Exp. of Discovery, p. 271, t. 2
(1837) who mentioned its use as a food plant and figures a fruit. The first valid botanical
description was published 30 years later. The ripe fruits are edible when ripe, but bitter
and unpalatable when still green. The seeds contain a considerable quantity of fat and
are exported in large quantities to Cape Town. They are an excellent substitute for
almonds.
54
Acanthosicyos horrida is an example of extreme adaptation to a life on the dry
sand-dunes (which it builds up itself by catching the sand and emerging every time it
is covered by the sand) of the Namib desert with its extreme climatic conditions. The
xerophytic habit (reduced leaves, spines, leathery perianth, etc.), the peculiar shrubby
growth and the absence of tendrils are unusual among Cucurbitaceae.
As regards the citation of the specific name, in Bentham and Hooker, Genera
Plantarum, only a description of the genus is given, but this is a “ genero-specific ”
description because only a single species is mentioned. The species was, therefore,
also validly published in Genera Plantarum and has to be cited as horrida Welw.
ex Hook. f. (1867)”.
11. CITRULLUS
Citrullus Schrad. in Eckl. & Zeyh., Enum. PI. Afr. Austr. 2: 279 (1836), nom. cons.;
Forsk., FI. Aeg.-Arab. 167 (1775), nomen nudum; E. & Z. in Linnaea 12: 412 (1838);
Naud. in Ann. Sci. Nat. 4me. ser. 12: 99 (1859); Sond. in FI. Cap. 2: 493 (1862);
Benth. & Hook, f., Gen. PI. 1 : 826 (1867); Hook. f. in FI. Trop. Afr. 2: 548 (1871);
Cosn., Mon. Cucurb. 507 (1881); Pflanzenreich 275 .2: 102 (1924); Pax in Pflanzenfam.
4, 5: 27 (1889); PhilUps, Gen. ed. 2: 749 (1951); Taxon 10: 125 (1961).
Colocynthis Mill., Gard. Diet. Abridg., ed. 4: 1 (1754); Kuntze, Rev. Gen. 1: 256
(1891); Post & Kuntze, Lexicon 136 (1904), as “ Colocynthis L. (1753) ”; Exell, Cat.
Vase. PI. S. Tome, Suppl. 21 (1956); Taxon 5: 38 (1956).
Type species: C. vulgaris Schrad.
Prostrate annuals or perennials with herbaceous or sometimes more or less woody
stems, often with a musk scent or foetid. Leaves orbicular to triangular-ovate in
outline, always deeply 3-5-lobed with lobulate to pinnatisect lobes. Tendrils 2- or 3-fid,
or sometimes simple, in one species absent, in one species straight, spinescent. Flowers
monoecious or dioecious, medium-sized, on short pedicels. Male flowers solitary,
rarely fasciculate; receptacle broadly campanulate; sepals narrow, remote; corolla
yellow or pale yellow, rotate or shallowly campanulate, segments nearly free, ovate-
oblong, obtuse; stamens 3, inserted near the base of the receptacle; filaments short,
free; anthers free or somewhat cohering, one 1-thecous and two 2-thecous with linear,
sigmoid or conduplicate thecae; connective dilated but not produced into an apical
appendage; rudiment of ovary glandlike. Female flowers solitary; calyx and corolla
as in the male; staminodes 3, short, setiform or ligulate; ovary ovoid to subglobose,
smooth or with soft protuberances; placentas 3; ovules numerous, horizontal; style
short, columnar, not surrounded by a disc at the base; stigmas 3, thick, reniform,
sub-bilobed. Fruit globose to oblong, with a soft to firm or rather hard pericarp and
a usually softer pulp containing the numerous seeds, indehiscent. Seeds ovate to
oblong in outline, much compressed, white- or dark-coloured; testa hard, tegmen
membranous; cotyledons oblong to obovate; radicle conical, subacute.
A genus of 4 species, distributed from the Mediterranean area to India and Ceylon
and throughout the greater part of Africa, extending into the Cape Province.
The name Citrullus Schrad. has been proposed and accepted for conservation
by the Paris Congress. The relevant information does not consider Colocynthis Mill.
(1754) and that is why Exell (see Taxon 5, 1 : 38) has stated that: “ this conservation
should be reconsidered with full knowledge of the facts ”. There is no strong argument
in favour of conservation of Citrullus Schrad. (1836) against Colocynthis Mill.; the latter
name clearly antedates the first and is unambiguous and has, moreover consistently
been used in several recent floras (see the citations under Colocynthis citrullus below).
It is true that some authors do not use Colocynthis “ Mill.” but “ Tournef.” “ L.”,
“ Haller ” or “ Ludw.” as the author, but as the same genus is intended there is no
difficulty. The necessary combinations in Colocynthis have all been made and there
55
is no advantage in taking up 07rM//w5, except (see Taxon 2: 99,134; 4: 198; 5: 15)
as an argument to retain the name frequently used for the water-melon in agricultural
and horticultural circles, i.e., Citrullus vulgaris Schrad. This name is untenable because
it is illegitimate and, in Citrullus, would have to be replaced by a new combination
based on Momordica lanata Thunb. In Colocynthis, the legitimate name is C. citrullus
(L.) Kuntze, which retains the name “ Citrullus ” as the epithet. However, conservation
of Citrullus was confirmed in Taxon 10: 125 (1961).
The genus can be divided into two subgenera which dilfer in the fruits (smooth
in subgenus Citrullus and tubercled in the subgenus Pseudocucumis), in the tendrils
(soft and coiled or wanting in Citrullus, straight and spinescent in Pseudocucumis)
and in the chemical composition of the bitter substances found in the fruits (see Enslin
et al.).
Citrullus Schrad. subgen. Citrullus: Monoecious, tendrils coiled or absent, fruit
smooth. Contains all species except C. naudinianus.
Citrullus Schrad. subgen. Pseudocucumis A. Meeuse subgen. nov. — Dioicus, cirrhi
rigidi, spinescentes, pepo tuberculatum. Type: C. naudinianus (Sond.) Hook. f.
Tendrils present, sometimes straight, spinescent:
Tendrils reduced to straight spines; flowers dioecious; ovary and fruit with protuberances
1. C. naudinianits
Tendrils well-developed, usually 2- or 3-fid; flowers monoecious; ovary and fruit smooth
2. C. I amt us
Tendrils wanting 3. C. ecirrhosus
1. C. naudinianus (Sond.) Hook. f. in FI. Trop. Afr. 2: 549 (1871); Cogn., Mon.
Cucurb. 511 (1881); Verb. Bot. Ver. Brandenb. 30: 150 (1888); Pflanzenreich 275.2:
114 (1924); N.E. Br. in Kew Bull. 1909: 112 (1909); Burtt Davy, FI. Transv. 1:230
(1926):
Cucumis naudinianus Sond. in FI. Cap. 2: 496 (1862). C. dissectifolius Naud. in Ann.
Sci. Nat. 4me. ser. 11: 23 (1859), ex parte.
Colocynthis naudinianus (Sond.) Kuntze, Rev. Gen. 1: 256 (1891).
Type: When Naudin described Cucumis dissectifolius he cited a specimen collected
by Ward near Grahamstown and the Burke gatherings 276 and 488 from Mooi River
(Potchefstroom), all in herb. Hooker (now in K). Sonder referred Burke 276 to Cucumis
dissectifolius and Burke 488 to Cucumis naudinianus. The Ward specimen was not
mentioned, but from the locality alone it can be safely concluded that it does not belong
to Sonder’s C. naudinianus. Cogniaux stated that “ Burke 488 ” partim belongs to
C. naudinianus but referred Burke 276 and the Ward specimen to C. dissectifolius.
Accordingly, the holotype of C. naudinianus is Burke 488, at least that part that agrees
with the description given by Sonder, in Herb. Hooker (K.). A duplicate of Burke
488 in SAM represents this species and is clearly an isotype.
Perennial herb. Root tuberous, long fusiform-cylindric, up to about 1 m long and
8 cm in diam. Stems prostrate, several from the crown of the root, up to 3 cm long,
rooting at the nodes, often branched, usually firm, angular-sulcate, originally sub-hirsute,
shortly pubescent, glabrescent. Tendrils short, straight or slightly curved at the apex,
rigid and spinescent, glabrescent up to 3 • 5 cm rarely up to 7 cm long. Leaves secund,
rigid, deeply palmatifid, scabrid above, often setose along the main nerves, setose
(mainly on the veins) below, or scabrid on both sides, 3-8 cm, rarely up to 14 cm, long
and 2-6 cm, rarely up to 12 cm, wide; lobes 5, all more or less irregularly and lyrately
lobed, usually acute, the middle lobe the longest, the lateral ones smaller, the lowermost
much smaller, often with a lobule or sub-bilobed, all lobes and their secondary lobes
entire to dentate or lobed, midribs rather conspicuous and, like the petiole, often
brownish, purplish or reddish; petioles shortly setose-hispid, turning scabrous, 0-5-4
56
cm, rarely up to 7 cm, long. Flowers dioecious, solitary; receptacle cup-shaped to
shortly campanulate, 5-6 mm long, 7-8 mm in diam., shortly hispid like the linear-
lanceolate, subacute or obtuse, 6-7 mm long and about 1-5 mm wide sepals; petals
light yellow, ovate-elliptic, subacute, strongly veined, shortly hispid-pubescent, 1 • 5-2
cm long and 0-8-1 cm wide. Male flowers: pedicels 0-5-2 cm long, hispidulous,
filament dorso-laterally flattened, widening at the base, 4-5 mm long; anthers
subquadrate-orbicular, 4-5 by 4-5 mm; connective very broad and flattened. Female
flowers: pedicels longer than in the male, 3-8 cm long; ovary ellipsoid to ellipsoid-
fusiform, 12-15 mm long, covered with thick short soft conical spines tipped by short
setae; staminodes ligulate or cylindric-setaceous, small. Fruit ellipsoid, greenish
yellow or somewhat greyish light green, covered with large blunt conical spines, 6-12
cm long and 4-8 cm in diam. Seeds white, smooth, 7-5-10 mm long, 4-6 mm broad
and about 3 mm thick.
From southern Angola and Barotseland (Northern Rhodesia) to South West
Africa, Griqualand West and the northern and western Transvaal, also recorded from
the Matopos (Bulawayo distr.) in Southern Rhodesia and from Portuguese East
Africa.
Recorded from the following districts ; South West A frica: Ovamboland, Okavango,
Grootfontein, Okahandja, Gobabis, Otjiwarongo, Windhoek, Gibeon; Bechuanaland
Protectorate: widespread and common in many places; Cape Province: Gordonia,
Kuruman, Barkly West, Mafeking, Postmasburg, Vryburg, Kimberley; Transvaal:
Lichtenburg, Yentersdorp, Potchefstroom, Pretoria, Warmbaths, Waterberg, Pot-
gietersrust, Soutpansberg, Sibasa, Rustenburg; Orange Free State: Hoopstad, Boshoff;
Portuguese East Africa: Lourenco Marques.
The following specimens are of special interest:
South West Africa. — Okahandja: Dinter 308 (COT, GRA, L, SAM). Gibeon: near
Kalkfontein, Range 1383 (SAM). Gobabis: Range 780 (SAM), all cited by Cogniaux
(1924).
Bechuanaland. — Kwebe Hills, Lugard 155 (GRA), cited by N. E. Brown (1909).
Transvaal. — Potchefstroom: Mooi River, Burke 488 (K, holo, PRE, photo of holo.!.
SAM, iso.!). Potgietersrust: near Grass Valley (“ Klippan ”), Rehmann 5171 (BR,
cited by Cogniaux 1924 and by Burtt Davy).
Portuguese East Africa. — Sul do Save: Lourengo Marques, Schlechter 11672 (BOL.
BR, COl, GRA, L), cited by Cogniaux 1924.
Southern Rhodesia. — Bulawayo: Rogers 5739 p.p. ; 5747 (BOL).
Northern Rhodesia. — Barotseland: Watt & Brandwyk, 345; Setula, Borle 257
(PRE).
Angola. — Kuando, near Santa Cruz Mission, Codd 7551 (PRE); R. Kunzumbra,
Pocock 525 (BOL).
This species is a typical Kalahari species which prefers deep sandy soil. In
Bechuanaland the fruits are eaten on a large scale, raw or roasted, but occasionally
specimens with bitter fruits are encountered. The bitter substances in these fruits
when chemically analysed show a composition similar to those found in species of
Cucumis and altogether different from those in species of Citrullus sect. Citrul/us. The
simple spine-like tendrils, tuberculate fruits and the chemical composition of the bitter
substances in the fruit of Citrullus naudinianus clearly indicate a close relationship to
Cucumis, but in its floral characters (no apical appendages of the connective; rudimentary
ovary in the male flowers; style without a disc at the base in the female flowers) it is
clearly a species of Citrullus, so that this species forms a link between these two genera.
The best solution seems to be the retention of the true species of Citrullus as the section
Citrullus and the inclusion of C. naudianus in a separate section Pseudocucumis. The
alternative, a complete reduction of Citrullus to Cucumis, would certainly cause incou'
venience and would not improve matters.
57
2. C. lanatus {Thunb.) Mans/, in Kulturpfl. Beih. 2: 421 (1959).
Momordica lanata Thunb., Prodr. PI. Cap. 13 (1794). Type: in Herb. Thunberg,
Uppsala.
Cucurbita citrullus L., Sp. PI. ed. 1: 1010 (1753). C. angaria Duch. in Lam., Encycl.
158 (1786), nom. illeg.
Citrullus battich Forsk., FI. Aeg.-Arab. 167 (1775). C. vulgaris Schrad. apud Eckl.
& Zeyh., Enum. 279 (1836); Linnaea 12: 412 (1838); Naud. in Ann. Sci. Nat. 4me.
ser. 12: 100 (1859); Sond. in FI. Cap. 2: 494 (1862); Hook. f. in FI. Trop. Afr. 2:
549 (1871); Cogn., Mon. Cucurb. 508 (1881); Pflanzenreich 275.2: 103 (1924);
Burtt Davy, FI. Transv. 1 : 229 (1926). C. amarus Schrad. in Linnaea 12: 413 (1838);
19: 651 (1845).
Colocynthis citrullus {L.) Kuntze, Rev. Gen. 1: 256 (1891); Exell, Vase. PI. S. Tome
185 (1944); Andrews, FI. Anglo-Eg. Soudan 1: 168 (1950); Keay, FI. W. Trop.
Afr. ed. 2, 1: 213 (1954). Hiern, Cat. Afr. PI. Welw. 1: 397 (1898), sphalm. “ C.
amarissimus ” (for full synonymy, cf. Cogniaux 1924).
A prostrate monoecious annual, producing several herbaceous, rather firm and
stout stems up to 3 m long. Young parts more or less densely lanate-villous (with
curved hairs), older parts glabrescent. Leaves ovate or triangular-ovate in outline,
cordate at the base, herbaceous but rigid, soon somewhat scabrous on both sides,
6-20 cm long and 4-15 cm broad, deeply trifid with lobulate to pinnatilobed, obovate,
oblong or linear to lanceolate segments of which the central one is always the longest
and usually acute or acuminate; the lateral ones usually rounded to obtuse at the
apex often more or less bilobed; petioles usually with some curved hairs, 4-12 cm long.
Tendrils rather robust, usually bifid, pubescent especially towards the apex at least
when young. Male flowers: peduncle usually elongate, villous; receptacle broadly
campanulate, villous; sepals about as long as the receptacle; corolla usually greenish
on the outside, pale yellow inside, up to 3 cm in diam. Female flowers: ovary lanate;
style 4-5 mm long. Fruit in the wild form globose (3-) 6-20 cm in diam., in the cultivated
forms globose to ellipsoid or oblong and up to 60 cm long and 30 cm in diam., soon
quite glabrous and smooth; the pericarp hard but not woody, in the wild forms pale
green or greyish green usually mottled with longitudinal irregular bands of dark green
or dark greyish green, in cultivated forms often concolorous, very pale green to dark
green or yellowish, or mottled with darker green; or marmored with a darker shade;
the flesh in the wild form and some cultivated forms (citron-watermelon) used for
preserve, firm and rather hard, almost invariably white, in cultivated forms somewhat
spongy in texture but very juicy and soft, usually pink or bright reddish-pink. Seeds
usually very numerous, in wild forms black or dark brown as a rule; in cultivated forms
also white or mottled, varying in size but usually 6-12 mm long.
An old cultigen cultivated and in a semi-wild state in the warmer parts of the
whole world, but truly native most probably only in the more or less sandy drier areas
of Southern Africa, chiefly in the area of the so-called Kalahari desert. The cited
specimens were all growing wild or were apparently wild.
South West Africa. — Okavango: Diyona near Nyangana, de Winter & Marais
4613. Karibib: Usakos-Otjimbingue, Marloth 1281. Okahandja: Dinter 127; Quick-
born, Bradfield 200, 201. Otjiwarongo: Waterberg, Volk 1098; Boss s.n. Gobabis:
de Winter 24SS. Windhoek: Avisdamm. Liebenberg 4464. Marienthal: Dinter 2024.
Bechuanaland. — Ngamiland, Curson 177; Molepole, Story 4937; 4999; Kaotwe,
Van Son in T.M. 28802; between Kuke and Gomodimo, Van Son in T.M. 28794.
Cape Province. — Little Namaqualand: Richtersveld, Sendlingsdrift, Herre s.n. in
Hb. Marloth. Gordonia: 12 miles N. of Witkop, Po/c 2146. Calvinia: Schmidt
500. Prieska: Bryant J 2S3. Kuruman: Marloth 1137. Vryburg: Armoedsvlakte,
Henrici 116; Foley s.n. Kimberley: Vaal River, Marloth 825. Middelburg: Groot-
fontein, Verdoorn s.n.; Theron 593. Albany: near Grahamstown, Schbnland 4357.
58
Transvaal. — Soutpansberg: near Saltpan, Smit s.n.; Schweickerdt & Verdoorn 455.
Potgietersrust : Zebediela, Quin s.n. Waterberg: Mosdene, near Naboomspruit, Galpin
M 135; M 139. Warmbaths: Leendertz s.n. Brits: Welgevonden, Mogg s.n. Pretoria:
Chippindall 117; Fountains Valley, Verdoorn 746. Wolmaransstad: Strydpoort,
Sutton 240. Potchefstroom : 7 miles E. of Rysmierbult, Louw 1382. Johannesburg:
Turffontein, Bryant D 91. Bronkhorstspruit : Rooikop Farm, Smuts & Gillett 2020;
2078. Heidelberg: Uitgevallen 197, Burtt Davy 13676. Middelburg: King “ G ”.
Nelspruit: Kruger National Park, Pretoriuskop, van der Schijff 2416.
The original (pre-Linnaean and Linnaean) descriptions of this plant refer to
cultivated forms, as the wild forms became known to botanists at a much later date.
Many varietal names for the cultivated forms have been proposed (see e.g.. Harms
in Pflanzenreich 275.2: 107-108 and also Kew Bull., Add. Ser. 9, 2: 366, 1911) but
to my mind one should regard all forms as one species, because some cultivated forms
(such as the “ citron water melon ” with very firm fruit pulp) are so close to the wild
“ Tsamma ” of the Kalahari that one cannot draw a line anywhere to separate them.
Experiments by Dr. S. Rehm of the Horticultural Research Station, Pretoria, in which
he crossed the wild form and cultivars prove that the wild and cultivated water melons
behave completely as if they are one and the same species, producing completely fertile
progeny for several generations. Spontaneous “ hybrids ” between cultivated water-
melons and wild or semi-wild forms have been sent in by farmers several times. It
cannot be denied that there are several distinct “ types ” or “ forms ”, differing in fruit
size, colour of fruit, skin and pulp (red, pink, white or yellow), colour and size of
seeds, etc., but they are to be classified as “ cultivars ”.
The wild “ Tsamma ” occurs in two biochemically different forms, the one without
the bitter cucurbitacin E-glucoside (elaterinid), which provides food and water for the
Bushmen in the semi-arid regions of the Kalahari, and the other with elaterinid. Bitter
specimens of cultivated water-melons occasionally appear spontaneously in the fields
(the bitter substance is again elaterinid in this case), but these are constant (homozygous
for “bitter”) and must be mutants; these genetically constant, bitter variants must
not be confused with the crosses between cultivated water-melons and bitter wild
Tsammas which segregate in the F2 into bitter and non-bitter forms.
The bitter Tsammas have caused confusion with the bitter Colocynthis vulgaris
(colocynth or bitter apple). Reports on the occurrence of the true colocynth in Southern
Africa are, therefore, suspect. I have not seen any specimens that are referable to
C. vulgaris from the area covered by this revision.
3. C. ecirrhosus Cogn. in Schinz in Verb. Bot. Ver. Brandenb. 30: 151 (1888);
Bot. Jahrb. 10: 270(1888); Pflanzenreich 275 .2: 115(1924); Dinter in Fedde, Repert.
15: 432 (1919). Type: Luederitz 136 (Z, lecto.!), see below.
Colocynthis ecirrhosus (Cogn.) Chakrav. in Sci. & Cult. 15: 32 (1949).
Monoecious prostrate perennial with woody rootstock. Stems several to many
fairly slender to stout, canescent, not much branched, in the young parts puberulous
to shortly hispid, later glabrous, usually scabrid, up to at least 2 m long. Leaves rigid
and fragile, ovate to broadly suborbicular in outline, with strongly recurved margins,
canescent, usually glabrous, rather smooth above, except on the shortly setose-hispid
main nerves, at first shortly setose-scabrid, later callous-scabrid and white-punctate
mainly near the margins below, 2-6 cm in diam., 3-5-fid nearly to the base; segments
strongly lobulate, more or less rounded at the apex, crisped along the margin; petioles
fairly stout, greyish, shortly and densely setose, 0-5-2 cm long. Tendrils 0. Male
flowers: pedicel finely setose-hispid to glabrous, 0-5-3 cm long; receptacle shortly
and rather sparsely setose-hispid, 5-6 mm long; sepals shortly setose, erect, narrowly
59
triangular, 1-5-3 -5 mm long; petals yellow, obovate, rather densely and shortly
pubescent outside, 7-10 mm long. Female flowers: peduncles rather stout, 1-3 cm
long; ovary narrowly ovoid or ellipsoid, shortly and thinly pubescent, 15-25 mm long.
Fruit subglobose, green mottled with lighter green turning concolorous, yellow, glabrous,
8-15 cm in diam., with whitish very bitter pulp. 'S>eeds black or brown, ovate in outline
compressed, 8 mm long.
Type: Cogniaux cited specimens collected by Schinz, Luecleritz, Pechuel-Loesche
and Mar loth 1192. Of these some have been destroyed in the Berlin Herbarium, but
some of the specimens or duplicates are extant. The only remaining original specimen
annotated by Cogniaux that was available for study is apparently Luederitz 136 in Z
and thus automatically becomes the lecto-type.
South West Africa. — Omaruru: near Brandberg, Rodin 2755 (BOL, PRE); Merx-
mueller & Giess 1651 (M, PRE). Swakopmund: near Swakopmund, Marloth 1192
(BOL, GRA, L, PRE, SAM); Bradfleld s.n. (GRA, PRE); 560 (PRE); Lam & Meeuse
4086 (L); de Winter 3193 (PRE); near Goanikontes, Rodin 2157 (BOL, PRE); near
Rossling, Esdaile in Hb. Rogers 15365 (Z); Richthofen, Dinter 2807 (SAM); Haigam-
chad, Galpin & Pearson 7452 (PRE, SAM); 60 Km E. of Swakopmund, Dinter 61 \5
(B). Walvis Bay: near Walvis Bay, Jensen s.n. (PRE). Karibib: Kubas, Dinter 225
(SAM). Luederitz: Koichab dunes, Kinges 2345 (PRE); near Aus, Dinter 6117 (B,
BOL); Gerstner 6316 (NBG, PRE). Exact localities unknown: “ Garieb ”, Lindner
in Z (Lindner collected mainly near Otjimbingue); inter Walfischbay et Otjitambi,
Luederitz 136 (Z) = prob. s.n. (BR), this can be in the Swakopmund or Omaruru
distr. (Lectotype in Z, isotype BR).
Cape Province. — Richtersveld : Herre (seed only, grown in Pretoria, but died when
still young so that no specimens were made).
C. ecirrhosus is almost entirely restricted to the “ Namib ”, a desert area with
a very low rainfall and usually sandy soil, and seems to be well adapted to the extreme
climatic conditions in this habitat. Although limited in its distribution, this plant is
locally often frequent. This is one of the very few Cucurbitaceae without tendrils.
The fruit is invariably very bitter, in contra-distinction to that of C. lanatus which
(even in the wild form) is usually not bitter.
12. CUCUMIS
Cucumis L., Sp. PI. ed. 1 : 1011(1753); Gen. PI. ed. 5: 442 (1754); Ser. in DC., Prodr.
3: 299 (1828); Naud. in Ann. Sci. Nat., 4me ser. 1 1 : 9 (1859); 12: 108 (1859); Sond.
in FI. Cap. 2: 494 (1862); Benth. & Hook, f., Gen. PI. 1: 826 (1867); Hook. f. in
FI. Trop. Afr. 2: 542 (1871); Cogn., Mon. Cucurb. 479 (1881); Pflanzenreich 275 .2:
116 (1924); Pax in Pflanzenfam. 4, 5: 27 (1889); Phillips, Gen. ed. 2: 749 (1951).
Annual or perennial herbs, usually scabridly or setosely pubescent, mostly prostrate,
sometimes scandent, rarely suberect. Rootstock of perennial species a thick and
woody taproot or main roots fibrous, occasionally fibrous roots forming tubers with
spongy tissue. Leaves usually angular to pinnatilobed or palmately divided, the lobes
not frequently again lobulate or dissected, generally petiolate. Tendrils simple, occasion-
ally reduced, straight or spinescent, very rarely absent. Flowers usually small,
monoecious or less often dioecious, the male ones fasciculate or clustered, less often
solitary, usually distinctly, but generally shortly pedicellate, rarely sessile, the female
ones solitary, very rarely fascicled on usually short pedicels which are often thickened
and often more or less elongate in fruit. Male flowers: receptacle campanulate, usually
narrowly so, or turbinate; sepals 5, usually subulate or linear; corolla generally light
yellow, sometimes pale yellow or bright yellow, usually rotate; the 5 lobes nearly free,
obovate, oblong or ovate; stamens 3, free, with usually short linear filaments, two
60
2- thecous and the third 1-thecous, the thecae conduplicate, the connective produced at
the apex into an appendage; rudiment of pistil present and evident. Female flowers:
calyx and corolla as in the male but corolla not infrequently somewhat larger;
staminodes 3, setiform or ligulate; ovary globose to fusiform or subcylindric, smooth,
tubercled or with soft spines, glabrous or hairy, with 3-5 placentas and numerous
horizontal ovules; style short, columnar, surrounded by a disc at the base; stigmas
3- 5, globose, obovate or flattened and somewhat dilobed, sometimes connivent. Fruit
various, rather small to large, globose, ovoid, ellipsoid, oblong or trigonous, when
mature concolorous or longitudinally variegated with more or less mottled bands of a
darker or a different colour, green, yellow, white, orange or red, smooth, pustulate,
tuberculate or covered with protuberances or soft spines, glabrous or hairy, the pericarp
usually firm but not hard, the interior juicy, pulpy or gelatinous, usually white or almost
colourless, rarely green, orange or yellow. Seeds numerous, white, generally small,
ovate, oblong or obovate, much compressed, smooth, immarginate but the margin
more or less acute; cotyledons confoiming to the outline of the seed; radicle conical,
obtuse or subacute.
Type species: C. sativus L.
An Old World genus of about 30 species, the majority wide-spread in Africa,
extending to the Middle East, Arabia and India, one African species introduced and
naturalised in America, a few occurring as weeds in Europe and Australia, the commonly
cultivated species (C. sativus L., cucumber; C. melo L., melons) sometimes, as escapes
from cultivation, ruderal in many parts of the world.
The latest comprehensive treatment of the genus (Cogniaux, 1924) is not very
satisfactory because several species are included which are referable to Oreosyce
(= Hymenosicyos in the Melothrieae!, not Cucurditeae).
Most species are found in arid and semi-arid regions, in deserts, steppes and
savannah forests, very few prefer moister conditions and deep shade. One species
(C. humifructus) is the only geocarpic member of the Cucurbitaceae.
Tendrils wanting II. C. rigidus
Tendrils present at least at the base of most of the leaves, although sometimes short and stiff or
occasionally more or less spinescent:
Leaves pentagonal in outline, 2-A cm long and broad, palmately 5-lobed with usually acute
triangular lobes, firm in texture, finely scabrid; all other vegetative parts at least when
young, pedicels, calyx, median veins of corolla and young fruit with very short, often curved,
bulbous-based aculeate hairs; fruit asperulous, longitudinally banded, up to 3 cm long,
on a 3-6 cm long stalk; dioecious prostrate perennial with woody whitish basal parts of
stems (S.W. Africa) 6. C. asper
Leaves either longer than broad, often over 4 cm long, or leaf-shape, leaf-size and/or pubescence,
or fruit not as above:
Leaves merrbranous when dry, about as long as broad, palmately 5-lobed with broad lobes;
stems slender; fruits oblong-ellipsoid, longitudinally variegated with dark green mottled
bands on a pale dull greenish-white to pale yellow background, smooth and glabrous
except for a few perpendicular very thick long stiff multicellular hairs; green (not canercent)
climbing to prostrate perennial with the habit of a Melothria (Transvaal). . . 7. C. umbrosus
Leaves not membranous when dry; plant and fruit not as above:
Fruit not spiny or tubercled when ripe (smooth and glabrous or occasionally hairy);
ovary hairy or glabrous but not with soft spines or tubercles:
Annual, monoecious, not canescent nor glaucous; leaves usually large, about as broad
as long, 6-8 cm in diam. and over; fruit usually over 5 cm long:
Flowers sessile, several male ones and a female one clustered; ovary after fertilization
pushed into the ground and ultimately fruit geocarpic on a 10-25 cm long
peduncle 2. C. humifructus
Flowers not clustered; fruit borne on a stout peduncle up to about 3 cm long above
ground 1. C. melo
Perennial, ovary and young fruit hairy; fruit when ripe up to about 4 cm long, rarely
larger :
61
Plant, especially young parts and lower surface of leaves covered with soft adpressed,
usually yellowish to brownish, more rarely greyish hairs, leaves longer than broad,
not distinctly cordate at the base, often somewhat 3-lobed, sometimes dissected,
drying dark green to dark brown; fruit red when ripe; flowers dioecious (not
in S.W. Africa) 3. c. hirsutus
Plant canescent or glaucous, not with soft adpressed hairs; leaves usually about as
long as broad; distinctly cordate at base, drying grey or greyish-green; fruit
greenish yellow to orange when ripe; flowers monoecious:
Fruit ellipsoid, pale (greenish)-yellow, with about 10 paler longitudinal bands, leaves
sometimes lobed or subhastate 4. C. angolensis
Fruit subglobose, deep yellow to orange, concolorous when ripe, leaves not lobed or
subhastate 5. C. dinteri
Fruit with soft spines, tubercles or protuberances; ovary with soft spines or tubercles:
Leaves distinctly lobed to deeply pinnati- or palmatisect:
Segments of almost completely palmatisect leaf 5-7, linear, many times longer than
broad, usually entire and usually under 4 mm broad; dioecious perennial with
variegated fruit 9. C. heptadactylus
Segments of leaf not linear, as a rule over 4 mm broad, often coarsely dentate to
lobulate :
Female pedicels usually over 3 cm long; in fruit often much longer (6-18 cm);
fruits not variegated or banded, usually very densely covered with long subulate
soft spines 15. C. anguria var. longipes
Female pedicels usually under 3 cm long and in fruit not much lengthened ; if pedicels
over 3 cm long, fruit variegated or with very short sparse spines:
Fruit distinctly variegated at least when young with longitudinal usually darker
bands, or, if indistinctly variegated when mature, globose, under 3 cm in
diam. and species annual:
Annual, monoecious; stems usually slender; leaves usually dissected, but rarely
coarsely and more or less acutely incised-dentate :
Spines on fruit laterally compressed, large, dense; fruit narrowly ellipsoid to
oblong, usually over 4 cm long; bands on fruit brownish to dirty purple
on a paler, dull pale yellow or greenish-white background ; corolla usually
over 1 cm long. 12. C. africanus
Spines on fruit terete or very small; fruit ellipsoid or subglobose, under 4 cm
long, corolla under 1 cm, usually under 7 mm long:
Fruit ellipsoid, distinctly longitudinally banded with dark green on a greenish-
white background when immature, ultimately bands of a dirty orange to
rusty-brown colour; spines rather dense, often over 2 mm long
13. C. myriocarpus
Fruit subglobose, when mature pale yellow with or without indistinct
narrow pale longitudinal bands; spines sparse, often under 2 mm
long (sometimes very sparse and/or very short) 14. C. leptodennis
Perennial, dioecious; stems usually stout; roots forming tubers 50-100 cm below
ground; flowers over 1 cm long; leaves usually coarsely and more or less
acutely incised-dentate 10. C. kalahariensis
Fruit not variegated; perennials:
Fruit under 5 cm long, covered with subulate spines over 2 mm long, usually
densely so, rarely with shorter spines (not in S.W. Africa). . .... 16. C. zeyheri
Fruit usually over 5 cm long, sparsely covered with short conical spines or
pustulate 17. C. ficifolius
Leaves shallowly or hardly lobed:
Leaves thin in texture, pubescence of semi-pungent but not aculeate patent long hairs;
annual climber; fruits ellipsoid to subcylindric, sometimes more or less trigonal,
7-20 cm long, with coarse blunt protuberances, orange when ripe and filled with a
green pulp 8. C. metidiferus
Leaves, pubescence and fruit not as above:
Fruit concolorous; densely covered with soft spines. . . . see 15. C. anguria var. longipes
Fruit longitudinally variegated, not as densely covered with soft spines. . . . (young
plants of) 13. C. myriocarpus
1. C. meloL., Sp. PI. ed. 1: 1011 (1753); Hook. f. in FI. Trop. Afr. 2: 546 (1871);
Cogn., Mon. Cucurb. 482 (1881); Pflanzenreich 275.2: 120 (1924); Burtt Dav>,
FI. Transv. 1: 228 (1926). Type; The name was originally based on pre-Linnaean
authors, but the specimen “ Cucumis melo 4 HU ” in the Linnaean Herbarium (No.
1152/8) can be taken as the lecto-type.
1517626-3
62
Annual creeper. Stems branched, hirsute, glabrescent, in the wild forms sometimes
short, but in cultivated forms up to 2 m long. Leaves sub-orbicular to reniform,
usually angular or faintly 3-7-lobed, cordate at the base, as a rule denticulate, softly
hairy to subhirsute on both surfaces, in cultivated forms up to 15 cm long and wide,
but in the truly wild forms sometimes considerably smaller; petioles as a rule about as
long as or longer than the corresponding blades, subhispid. Tendrils usually well
developed. Flowers monoecious, light yellow. Male flowers fasciculate on slender
pedicels up to 2 cm long; receptacle hairy, 4-8 mm long; sepals subulate, shorter
than the receptacle; corolla in cultivated forms up to 2 cm long, but in the wild forms
usually much smaller (often about 8 mm long). Female flowers solitary on a robust,
up to 2 cm long pedicel; perianth as in the male; ovary various but in the wild forms
ellipsoid or faintly turbinate, hairy. Fruit in the wild oblong or ellipsoid to obpyriform
or somewhat turbinate, smooth, usually green mottled in bands with lighter greyish
green but often ultimately concolorous, light to rather deep yellow, 4-6 cm long and
2-4 cm in diam. with white pulp (in cultivated specimens much larger with various types
of skin and colour pattern; smooth or rough, yellowish, greenish, greyish or brownish,
etc., and with yellow, orange or green pulp). Seeds in cultivated forms 10-12 mm by
5- 7 mm, in wild forms sometimes considerably smaller.
The wild form, which is often referred to as the var. agrestis Naud., is a native of
Africa south of the Sahara and extends southwards to Southern Rhodesia and the
Transvaal and has also been recorded (in a truly “ wild ” state?) from south-west Asia.
Cultivated melons are grown in all the warmer countries of the world, and elsewhere
in greenhouses. Specimens that have run wild or are obviously culture relics often
“ degenerate ” into forms approaching the original wild specimens, so that it is not
always easy to see from a herbarium specimen if it is such a depauperate cultigen or
a truly wild form. As far as can be ascertained, the wild form occurs in the Transvaal
(usually in cultivated or old lands), but not in South West Africa or elsewhere in the
Republic of South Africa.
As there is no sharp dividing line between the wild and the cultivated forms, it
is not advisable to retain the wild ones as “ var. agrestis ”. “ Forma culta ” is to be
preferred as a designation for the cultivated melon races, if one does not refer to a
particular race by its horticultural varietal name.
2. C. humifructus Stent (sphalm. humofructus) in Bothalia 2: 356, 358 cum icon.
(1927); Meeuse in Farming in S. Afr., June 1955; Arch. Neerl. Zool. 13, suppl. 1:
314-318 (1958). Type: Stent Hb. No. 2866 (indicated on the sheet by the author,
PRE, holo.!).
Cucumis sp. nov. ?, Burtt Davy, FI. Transv. 1: 229 (1926).
Annual prostrate herb. Stems one to few from the apex of the rootstock, trailing,
often rooting at the nodes, firm, stout, angular, pilose, up to 2 m, rarely 3 m, long.
Leaves orbicular or broadly ovate-pentagonal in outline, obtuse to apiculate or subacute,
cordate at the base, undivided, angular or faintly lobed, 5-15 cm in diam.; the margin
minutely ciliate and callous-toothed, both surfaces scabrid with sharp bulbous-based
hairs; main veins 5; petioles firm, terete, hispid, up to 6 cm long. Tendrils short,
slender. Flowers in contracted racemes or fascicles, usually a few male ones together
with a female one; male ones: pedicels slender, pilose, articulate below the receptacle,
6- 20 mm long; receptacle about 5 mm long, pilose, 10-nerved; sepals 2-3 mm long,
linear-subulate, obtuse, thinly pilose-hirsute; petals pale yellow, 2-7 mm long, papillose
with short gland-tipped hairs on both sides, intermingled with pilose hairs on the veins
outside, mid-vein protruding as a gland-like mucro; female ones: pedicel very short
and stout, after pollination bending downwards, lengthening considerably and thickening
into a long terete root-like organ reaching a length of 10-25 cm and a diam. of 3-5
mm, carrying the ovary into the soil; calyx and corolla as in the male, ovary ellipsoid
63
or narrowly ovoid, contracted into a narrow neck at the apex, densely covered with
retrorse adpressed setae which act as barbs when the peduncle grows into the soil.
Fruit geocarpic, subglobose, depressed globose or ellipsoid to ovoid, whitish, yellowish
or pale greenish, glabrous, finely rugose, 5-9 cm long and in diam. Seeds white,
much compressed, 16-20 mm long, 7-9 mm broad, embedded in a subhyaline jelly-like
pulp.
South West Africa.— Grootfontein: Aha Mts., Story 6364 (PRE). “ Otjisondu ”
(= Otjosandu) or Otjizondo?, district Karibib or Otjiwarongo ?, Piischel in Herb.
Dinter 579 (SAM, this specimen bears a tentative name “ Cucumis pueschelii ” given
by Dinter).
Transvaal. — ^Waterberg: near Sandrivierspoort, Stratford s.n.; Meeuse 9631 (both
PRE); Mosdene near Naboomspruit, Galpin M 146 (PRE, BOL). Potgietersrust
(teste Stent). Pretoria: Elandsfontein, v.d. Wal s.n. (PRE); Cullinan (teste Stent).
Cultivated specimens ex Galpin M 719 (which was a fruit only): Stent Hb. No. 2866
(type, PRE); Galpin M 719” (PRE).
Southern Rhodesia. — Gwelo: plants grown at Salisbury from seed collected by
Arnold at Gwelo, Hb. 3551, 6154 (SRGH). Hartley: Poole Farm, Hornby 3343
(SRGH).
Also in Tanganyika and the Belgian Congo.
This plant is interesting being the only geocarpic Cucurbitacea known. As a
consequence of its geocarpy, the plant is only found on soils consisting of deep fine
sand, obviously because the resistance of the soil must not be too great. Still, the
peduncles often get damaged during their growth and do not bear fruit (field observations
by the present author). The fruit, which develops only after the peduncle has reached
its ultimate length, forms a waxy substance on its surface, so that it is water-repellent.
This is obviously an adaptation to geocarpy — it serves to prevent early decay if the soil
becomes water-logged.
Another interesting biological feature is the possible relation between the occurrence
of antbears (or aardvarks), Orycteropus afer (Pallas) and of Cucumis humifructus.
It is a fact that this plant is almost exclusively found near old ant-bear holes — this
phenomenon was independently reported by several observers — and in the Waterberg
district the name “Aardvark Cucumber ” was used for this plant. It is an established
fact that the animal eats the fruits and disperses the seeds with its dung (which it buries).
For a detailed discussion see Meeuse 1955, 1958.
3. C. hirsutus Sond. in FI. Cap. 2: 497 (1862); Cogn., Mon. Cucurb. 489 (1881);
Pflanzenreich 275.2: 133 (1924); Burtt Davy, FI. Transv. 1: 228 (1926); Andrews,
Flow. PI. Anglo-Egypt. Soudan 1: 172 (1950). Type: Sonder mentioned specimens
collected by Burke (no. 297) and Zeyher (no. 581) from near Potchefstroom, Transvaal
and Krauss 91 from “ Port Natal ”. Burtt Davy (l.c.) designated the Burke & Zeyher
gathering as “ twin types ” (Herb. Sonder in S).
C. sonderi Cogn., op. cit. 489 (1881); 134 (1924); Burtt Davy, l.c. (1926).
Dioecious, prostrate, very rarely scandent, perennial herb, with woody rootstock.
Stems annual, varying from rather slender to fairly robust, sulcate-angular to subterete,
usually not much branched, when young densely, later more sparsely, pilose or subhirsute,
up to about 2 m long. Leaves rigid to firmly herbaceous, often secund, cordate,
pentagonal, elliptic, ovate-oblong, ovate or oblong-lanceolate in outline, entire or
shallowly, rarely deeply, 3-5 (-7) lobed, palmately 5-7-nerved from the base, 2-10
cm long and 1-6 (-9) cm wide; the base truncate or rounded to shallowly and broadly
cordate or occasionally subhastate; the apex of the leaf or its lobes acute or obtuse to
rounded, usually minutely mucronate; the margin entire to sinuous or finely, rarely
coarsely, denticulate; both surfaces more or less densely strigose-tomentose or setose, on
64
the upper surface often glabrescent and, if so, frequently turning white-punctate and
scabrid, pubescence on the lower surface more persistent; petioles rather stout to slender,
0 • 5-2 • 5 cm long, hispid. Tendrils usually very short and reduced, more or less densely
pilose or hispid. Male plants: flowers solitary or fasciculate or in a short raceme;
common peduncle, if present, up to 25 mm long; pedicels slender, more or less densely
pubescent, 0-5-3 cm or occasionally up to 7 cm long; receptacle narrowly campanulate,
densely villous-hirsute or more or less densely pilose, 4-6 mm long, 3-4 mm wide
at the apex; sepals linear to subulate or filiform, erecto-patent, 3-5 mm long; corolla
light yellow, hairy, 6-13 mm long. Female plants: peduncle up to 3 cm long, in fruit
incrassate and glabrescent; ovary subglobose to broadly ovoid or ellipsoid, densely
covered with stiff hairs, about 7 mm long; calyx and corolla as in the male. Fruit
globose or broadly ellipsoid, at first pale greenish yellow with irregular longitudinal
dark green zones, when ripe red or brownish, concolorous, ultimately very sparsely
hairy to glabrous, smooth, 4-6 cm long, 2-5-4 cm in diam. Seeds slightly compressed,
7-9 (-13) mm long, 5-7 (-10) mm broad.
Plants agreeing with the description given here occur from southern Soudan to
the Rhodesias, Bechuanaland, Transvaal, Natal and the eastern Cape Province. It
remains to be seen if the species C. wildemanianus Cogn. from the Belgian Congo is
really distinct; if it is not, the area of distribution covers most of west tropical Africa
as well.
The species is widespread and, judging by the numerous herbarium records, fairly
common throughout its area of distribution. In the area under discussion it has been
recorded from Bechuanaland: near Gaberones and Mochudi; Transvaal: practically
all districts except a few in the extreme south-west, and Standerton, Wakkerstroom
and Piet Retief; Swaziland (teste Burtt Davy); Natal: most districts, not recorded
from some lowland districts in Zululand; Cape Province: Mount Currie, Umtata,
Kentani, Komgha and Mafeking. The following specimens are of special interest:
Transvaal. — Potchefstroom : “ Mooi River et Wonderfontein ”, Burke 291 \ Zeyher
581 (syntypes in SAM). Potgietersrust : Makapansberge, Rehmann 5403 (BR). Lyden-
burg: Wilms 494 (PRE).
Natal. — Durban: Krauss 91 (PRE, one of the original gatherings cited by Sonder).
Biggarsberg: Wood 4548 (NH). All these specimens are cited by Sonder, Cogniaux
or Burtt Davy.
The species Cucumis sonderi Cogn., which was supposed to differ in leaf shape,
cannot be maintained. The degree of dissection varies considerably, often on one
specimen. A closely related group of species is centred around (and most probably
is identical with) C. welwitschii Cogn., viz., C. seretii De Wild., C. homblei De Wild,
and C. seretioides Suesseng. This group of species is supposed to differ from C. hirsutus
mainly in the prominent nervation on the lower surface of the leaves (cf. Cogniaux’s
key in Pflanzenreich, 1924, p. 117-118), but this character seems to break down.
Larger series of specimens from tropical Africa than available at the National Herbarium
are required to establish the limits of variation and to decide if these species are really
distinct. At any rate, the South African specimens form a homogeneous group and
are undoubtedly all referable to Cucumis hirsutus. The “ varieties ” distinguished by
Cogniaux (1924, etc.), and Burtt Davy (1926) are only extreme forms linked up by
series of intermediate specimens and not worthy of varietal rank.
4. C. angolensis Hook. f. ex Cogn., Mon. Cucurb. 487 (1881); Pflanzenreich
275.2: 131 (1924). Type: Welwitsch 831 from Moss^medes, Angola (K, holo.!;
COI, iso.!).
65
Perennial, glaucous and somewhat canescent. Stems prostrate to climbing,
slender to stout but usually more slender than those of C. dinteri, usually not much
branched, angular-sulcate, shortly hispid or with short bulbous-based setae along the
angles, later often glabrescent and/or scabrid, as a rule over 50 cm long (in cultivated
specimens up to 1 • 50 m). Leaves rather fleshy, varying in shape from reniform-
suborbicular to cordate-ovate or sometimes triangular-cordate, sometimes more or
less 5-lobed and subhastate, 4-8 cm long and broad, sometimes larger, with a rounded
but usually abruptly and shortly apiculate apex and distinct basal sinus, if lobed, with
usually somewhat angular and sub-acute, rarely rounded lobes; the margin varying
from subentire to rather coarsely crenulate-dentate or undulate-sublobate; teeth
usually separated by obtuse shallow sinuses; both surfaces at first densely and shortly
setose-scabrid later often glabrescent turning scabrid, but main veins on lower surface
remaining setose-hispid; petioles shortly hispid or setose with bulbous-based hairs,
2-10 cm long. Tendrils glabrescent. Flowers monoecious, light yellow. Male flowers
solitary, fasciculate or very shortly racemose; pedicels slender to filiform, finely setulose,
2-20 mm long; receptacle more or less densely setulose to sericeo-setulose, 5-7 mm long;
sepals erect, subulate, 4-6 mm long; corolla finely hairy, 5-9 mm long. Female
flowers solitary on setulose or somewhat hispid 13-20 mm long pedicels; ovary oblong,
densely covered with long stiff white hairs; corolla sometimes larger than in the male
flowers. Fruit ellipsoid, smooth, glabrous, yellowish green to pale lemon-yellow and
longitudinally variegated with about 10 narrow paler bands, 4-5 cm long and 3-4
cm in diam. Seeds elliptic, usually with the widest part in the middle and both ends
tapering into a subacute point, 5-6 mm long and about 2 mm broad.
Angola. — Mossamedes : Castro 136 (COI); Exell & Mendonca 2157 (BM, COI).
South West Africa. — Kaokoveld: Anabib (Orupembe), Story 5699; 5747 (PRE);
Karos Fountains, de Winter 3086 (PRE, K, M); Sesfontein, de Winter £ Leistner
5863 (PRE, K, M); 5786a (only fruits collected; specimens grown at Roodeplaat
near Pretoria as 5786a bis in PRE); Otjihu, de Winter & Leistner 5661 (PRE,. K, M).
Outjo: Franzfontein, 4925 (PRE); Witklipp, Fu/Ar 2895 (PRE); 2897 (M).
This species is closely related to C. dinteri and hybridizes with the latter as was
shown by Dr. S. Rehm of the Division of Horticulture. These two species cannot
always with certainty be separated in the herbarium, but Dr. Rehm has grown them
side by side and there are quite a number of differences which, however, cannot always
be noticed in herbarium specimens.
In the first place there is a difference in habit. C. dinteri grows as a short suberect
to decumbent bush with stems up to 50 cm long. C. angolensis forms much longer
prostrate to climbing stems well over 50 cm long. C. angolensis is glaucous and not
so canescent as C. dinteri. Other differences are in the size of the leaves (larger in
C. angolensis) and in leaf-shape (leaves not lobed in C. dinteri, sometimes lobed in
C. angolensis), but none of these are absolutely reliable. The best and apparently
reliable differences are in the fruits and seeds. C. dinteri has subglobose and concolorous
yellow to orange fruits and obovate seeds, whereas C. angolensis has larger ellipsoid
and longitudinally variegated greenish yellow to pale yellow fruits and longer elliptic
seeds. There are also important chemical differences between the two, C. dinteri
having extremely bitter fruits rich in cucurbitacin D and very bitter leaves, and C.
angolensis having only slightly bitter fruits with a mere trace of cucurbitacin D and
almost non-bitter leaves. There is also a difference in distribution, C. angolensis
being restricted to southern Angola and the Kaokoveld region and C. dinteri having
a more southern distribution.
5. C. dinteri Cogn., in Bull. Herb. Boiss. 2 me ser. 1; 882 (1901); Pflanzenreich
275.2: 131 (1924). Type: Dinter 1436 from Giftkopje, Omaruru, S.W. Afr. (Z, holo.!;
BR, iso.!).
66
C. cogniauxianus Dinter ex Cogn., op. cit. 131 (1924). Syntypes: Aus, Dinter 1115;
Satansplatz, Dinter 2078; Buellsport, Dinter s.n.
Perennial, canescent. Stems suberect to decumbent, usually under 50 cm long,
angular-sulcate, often very woody and grey to white at the base, the younger parts
shortly and usually sparsely hispid or setose-scabrid on the angles. Leaves canescent,
as a rule suborbicular-cordate or reniform-cordate, sometimes faintly 5-angled rarely
very indistinctly 5-lobed, rounded at the apex or shortly and abruptly apiculate, deeply
cordate at the base, firm and rather thick in texture, usually distinctly crenate-dentate
along the margin, on both surfaces at first rather densely and shortly setose-strigose,
glabrescent, but veins on lower surface persistently setulose-hispid, blade usually under
5 cm long and broad ; petioles shortly hispid or setose, rarely over 5 cm long. Tendrils
usually short or more or less reduced, shortly and sparsely hispid. Flowers monoecious.
Male flowers solitary or in fascicles of 2-3; pedicels slender to filiform, finely setulose,
2-10 mm long; receptacle more or less densely setulose to sericeo-setulose, 4-6 mm
long; sepals erect, subulate, 1-5-2 mm long; corolla bright light yellow, finely hairy,
5-7 mm long. Female flowers solitary on 3-12 mm long pedicels, ovary ovoid-globose
to oblong, densely covered with rather long stiff, white hairs. Fruit subglobose when
ripe, smooth, glabrous, yellow to orange and concolorous, 2-3-5 cm long and 2-3
cm in diam. Seeds obovate with narrow base and rounded to subacute apex, 4-4-5
mm long and about 2-5 mm broad.
South West Africa. — Omaruru: Erongo, Giftkopje, Dinter 1436 (Z, holo.!; BR,
iso.!). Karibib: Usakos, Dinter 5932 (B); Klein Ameib, Dinter 7088 (B, BOL, M,
PRE, Z); Kubas (near Ababis), Dinter 223 (COI, PRE, SAM); Karibib, Kinges
3637 = Krdusel 730 (M); Donkerhoek, Volk 2769 (M). Rehoboth; Buellsport,
D/rttc;- s.n. (PRE, SAM); S'rrcy 2084 (PRE) ; farm Djab, IKu/to- 4512 (M). Rehoboth
or Maltahohe: Naukluft Mts., Rodin 2886 (BOL); Naukluft, Dinter 8313 (B), Volk
819 (M). Gibeon; Haribes, Volk 2450/1956 (M); Satansplatz, Dinter 2078 (SAM);
Leber River, Ganaams, 9374; 9375 (BOL). Luederitzbucht : farm Weissenborn,
Kinges 2383 (PRE); Aus, Dinter 1115 (PRE, SAM); 6114 (B, BOL, M, PRE, SAM);
Merxmueller & Giess 2366 (M, PRE). Bethanien: Tschaunaup Mission, Gerstner
6328 (PRE). Keetmanshoop: Keetmanshoop, Dinter 3558 (B); about 25 miles from
Aroab on road to Rietfontein, de Winter 3406 (PRE, K, M); about 19 miles from
Ariamsvlei on road to Warmbad, de Winter 3574 (PRE, K, M); Karas Mts., Boss
TRV 35901 (PRE); Klein Karas, Dinter 5086 (B); s.n. (PRE); Oertendahl 70 (PRE);
Holoog, Bed of Great Fish River, Pearson 9705 (BOL); Seeheim, Dinter “ 3897 ”,
may be Schaefer 446 (B). Warmbad: Garius, Dinter 4233 (B, SAM).
Cape Province. — Hay: Lelykstad, Acocks 223 (KMG, BOL, NH). Kenhardt: 12
miles S.E. of Kakamas, Acocks 14382 (PRE).
For a discussion of the differences between C. dinteri and C. angolensis, see under
the latter.
6. C. asper Cogn., in Bull. Herb. Boiss 2me. ser. 1: 882 (1901); Pflanzenreich
275.2: 135 (1924); Dinter in Fedde, Repert. 16: 364 (1920). Type: Dinter 1447
from Ababis (distr. Karibib), S.W. Afr. (Z, holo.!).
Perennial. Stems rather numerous from the apex of the perennial woody rootstock,
prostrate to climbing, branched mainly near the base, up to about 75 cm long; when
young green, longitudinally sulcate, wiry but slender, densely covered with very short
bulbous-based setose-hispid hairs, soon becoming grey or canescent, ultimately very
woody and covered with a longitudinally fissured whitish bark and up to about 8 mm
in diam. Leaves more or less pentagonal in outline, firmly herbaceous, 2-4 cm long and
broad, 5 (-7)-lobed nearly to the middle or less deeply so with usually 5 triangular
lobes of which the central one is somewhat larger, and sometimes two very small
67
additional basal ones; all lobes usually acute, distantly and irregularly dentate or
subentire, often mucronulate, rarely rounded and/or lobulate; lamina finely scabrid-
setulose on both surfaces, but especially along the margin and on veins on lower surface,
with bulbous-based short aculeate hairs, soon becoming finely scabrid (the aculeate
hairs remaining longest on the margins and the veins below), with a triangular, rather
shallow to rather deep basal sinus; petioles slender, 1-5-4 cm, occasionally up to
6 cm long, densely and finely aculeate-hispid with bulbous-based thick aculeate hairs
which are partly retrorse (mainly in upper half of petiole) partly curved upwards
(mainly in basal part of petiole), or nearly all pointing upwards. Tendrils slender,
rather long, glabrescent. Flowers dioecious, rarely monoecious, pale yellow. Male
flowers solitary, fascicled or in short few-flowered racemes; peduncles and pedicels
filiform, finely and shortly aculeate-hispid, the first up to 8 mm, the last up to 25 mm
long; receptacle finely aculeate-hispid, 3-5 mm long; sepals triangular-subulate with
linear not very acute tip, finely setulose-aculeate, 1 -5-3 mm long; corolla-lobes sparsely
and very shortly aculeate-scabrid on median veins, 4-5 mm long. Female flowers
solitary, pedicels considerably thicker than the male ones, with the same scabrid
pubescence, in flower up to about 3 cm, in fruit up to about 6 cm long; ovary oblong-
fusiform, densely setose-puberulous. Fruit ellipsoid, somewhat obovoid or somewhat
obpyriform, rather densely covered with very short retrorse aculeae on a thick hard
base, when ripe greyish-green with longitudinal mottled dark green bands, 2-5-3 cm
long and 2-2-5 cm in diam. Seeds obovate-elliptic, 5-6 mm long, 2-5-3 mm broad.
South West Africa. — Kaokoveld: 30 miles S. of Kunene on road to Orupembe,
de Winter & Leistner 5796 (PRE, K, M); Ombepera, de Winter & Leistner 546S (PRE,
K, M). Omaruru: Brandberg, Tsisab Gorge, MerxmuUer & Giess 1657 (M, PRE).
Karibib: Ababis, Dinter 1447 (Z, holol).
Apparently a very local endemic growing in rock crevices. The asperulous fruits
are unique in the genus and very characteristic. Cogniaux described this species as
monoecious, but plants grown in Pretoria (Division of Botany Garden) and at Roode-
plaat Horticultural Research Station were all monosexual, so that Dinter 1447 must
have been an abnormal plant.
7. C. umbrosus A. Meeuse & Strey, sp. nov., a fructu oblongo-ellipsoideo levi
sparse setis patentibus subpungentibus pluricellularibus armato demum glabrescenti
et habitu Melothriae facile distinguendus.
Perennis herbaceus monoicus. Radices 1-2 m longae, divergentes, in partibus
superioribus incrassatae, fusiformes, ad 1 cm diam. Caules pauci vel plurimi, scandentes
vel interdum prostrati, graciles, angulati, pilis brevibus unicellularibus et pilis longioribus
3-4-cellularibus hirsuto-pilosis, ad 5 m longis. Folia herbacea ambitu 5-angulata vel
late ovata basi cordata, 5-lobata (interdum 5-secta) utrinque setuloso-strigosa, 3-6 cm
longa et lata, sinu basilari profundo angusto, lobis latis subangulatis vel subrotundatis
(interdum angustioribus) obtusis mucronatis margine leviter subremoteque crenato-
denticulatis, petiolis gracilibus breviter pilosis 3-6 cm longis. Cirrlii gracillimi, hirtius-
culi, glabrescenti, 3-9 cm longi. Flores masculi solitarii, pedicellis ad 3 cm longis,
receptaculo campanulato ca. 5 mm longo, sepalis subulatis vel filiformibus, 4-5 mm
longis, corolla pallide flava papillosa ca. 8 mm longa. Flores fenunei solitarii, pedicellis
demum 5-8 cm longis, corolla 9-12 mm longa, ovario, ellipsoideo vel fusiforme, dense
piloso. Pepo oblongo-ellipsoideus, setis patentibus pluricellularibus subpungentibus
sparse hirsutus, glabrescens, fasciis irregularibus alternim pallentibus (demum flavescen-
tibus) et atroviridibus longitudinaliter variegatus, ad 4 cm longus et 2 cm latus. Semina
numerosa, elongato-ovata, ca. 3 mm longa.
Transvaal. — Soutpansberg: Wylliespoort, Meeuse 10612. Middelburg: Kloofs on
northern side of Loskop Dam, Farm Rietvallei 92, Strey 2851 ; Meeuse & Strey 10317
(PRE, holo.l; B, BOL, BM, BR, EA, K, L, M, P, S, SRGH, isos.!).
68
A perennial herbaceous climber with the habit of a Melothria, growing in shady
places. Roots several, up to 2 m long, divergent, in upper portions swollen, more
or less fusiform, up to 1 cm in diam., crown of the roots with buds which give rise to
several stems. Stems few to many, usually branched, weak and slender, climbing
or occasionally prostrate, pubescent with short unicellular and longer 3-4-cellular
rather stiff hairs, especially in youngest parts, up to 5 m long. Leaves softly herbaceous,
in outline pentagonal or broadly ovate with cordate base, 5-lobed to a little less than the
middle, or in Rhodesian specimens 5-fid, on both sides rather densely covered with
adpressed bristly hairs, 3-6 cm long and as wide; basal sinus deep and narrow; lobes
rounded or somewhat angular, in Rhodesian specimens narrower, oblong-lanceolate,
usually obtuse, mucronate, the margin shallowly and rather remotely denticulate with
the veins protruding at the teeth as minute mucros; basal lobes usually considerably
smaller than the upper 3 ; petioles slender, shortly hairy, 3-6 cm long. Tendrils slender,
finely hirtellous, glabrescent, up to 9 cm long. Flowers all solitary. Male flowers:
pedicels slender, finely setulose-pubescent, up to 3 cm long; receptacle campanulate,
setulose-pubescent, about 5 mm long; sepals subulate or filiform, 4-5 mm long;
corolla pale yellow, papillose, about 8 mm long. Female flowers: pedicels slender,
ultimately 5-8 cm long; corolla lobes 9-12 mm long; ovary broadly fusiform or
ellipsoid, densely setose-pilose. Fruit oblong-ovoid, smooth with sparse stiff patent
bristly hairs, glabrescent, banded with mottled bands of a dark green colour on a pale
green to yellowish background, up to 4 cm long and 2 cm in diam. Seeds numerous,
elongate-ovate, in fruits studied up to 3 mm long.
This species is rather aberrant in that it grows in the shade, sometimes in deep
shade, whereas all other species known to me are found in open country or at best in
forest edges, and also because in habit it resembles species of Melothria (slender climbing
stems, leaf-shape). The first specimen collected {Strey 2851), with male flowers only,
was without hesitation referred to Melothria spec., until the flower dissection unmis-
takably indicated the genus Cucumis. Material with female flowers and fruits was
especially collected in the same locality to enable a description of this remarkable plant.
A second locality was discovered later in Wylliespoort which seems to point out
that this species must have been overlooked in the past. A search in suitable localities
(rather moist ravines and kloofs on granite, quartzite and other acid rocks, in the
shade) will certainly reveal that this plant is more widespread.
A study of some Rhodesian specimens showed that they differ from the Transvaal
ones only in the more deeply dissected leaves and I do not hesitate to include these
specimens in C. umbrosus.
Southern Rhodesia. — Umtali distr. : Umtali Commonage, Chase 769, 5979 (PRE,
SRGH); Limunya’s Reserve, Chase 6008 (PRE, SRGH).
8. C. metuliferus E. Mey. ex Schrad. in Linnaea 12: 406 (1838); Naud. in Ann.
Sci. Nat. 4me. ser. 11: 10 (1859); Sond. in FI. Cap. 2: 495 (1862); Hook. f. in FI.
Trop. Afr. 2: 543 (1871); Cogn., Mon. Cucurb. 499 (1881); Pflanzenreich 275.2:
146 (1924); Hiern, Cat. Welw. Afr. PI. 1, 2: 397 (1898); Burtt Davy, FI. Transv. 1:
228 (1926); Hutch. & Dalz, FI. W. Trop. Afr. 1: 182 (1931); Andrews, Flow. PI.
Anglo-Egypt. Soudan 1 : 172 (1950). Type: A Drege specimen from “ Omsamwubo ”
in Herb. Sonder (S, holo.).
Annual. Stems climbing or prostrate, sulcate, with long stiff patent hairs, up to
several meters long. Leaves herbaceous, dark green, cordate or cordate-ovate, usually
with a large basal sinus, shallowly 3-5-lobed or -angled, acute or shortly acuminate,
with irregularly dentate margin, at first rather densely hairy mainly on the veins later
glabrescent and scabrid, 4-15 cm long and about as wide; petioles with long hairs
like the stems, 3-10 cm long. Flowers monoecious. Male flowers fascicled or solitary;
69
pedicels filiform, pilose, 0-5-2 cm long; receptacle pilose, about 4 mm long; sepals
about 2 mm long; corolla hairy, light yellow, 5-6 mm long. Female flowers solitary,
ovary ellipsoid, oblong or somewhat fusiform, covered with robust soft spines ending
in a thin bristle, 15-20 mm long. Fruit subcylindric, ovate or ellipsoid, to subcylindric,
often somewhat trigonal, at first a finely mottled dark green, when ripe orange-red,
with green pulp, thinly covered with 1-1 • 5 cm long blunt or bristle-tipped stout
protuberances (5-) 8-16 cm long and 4-9 cm in diam. Seeds numerous, much com-
pressed, ellipsoid, attenuate at the base, 6-8 mm long.
Found in Africa south of the Sahara, extending in the south-west to the Okavango
area and, in the east, along the coast just reaching the eastern Cape Province. It has
been recorded from the following: South West Africa: Okavango area; Bechuanaland;
Transvaal: Soutpansberg, Sibasa, Pietersburg, Letaba, Lydenburg, Pilgrim’s Rest,
Nelspruit, Pretoria; Swaziland; Natal: fairly widespread in Zululand but, in recent
years, not collected near Durban and further south, where it was collected a century
ago; Cape Province: Port St. Johns. The following specimens are noteworthy: Zulu-
land, Gerrard cfe McKen s.n. (NH); Port St. Johns, Umzimvubu (“ Omsamwubo ”),
Drege (L, isotype!).
The attractive reddish-orange fruit which contains a large number of seeds embedded
in a light emerald green “ jelly ” occurs in two types, bitter (and unpalatable) or non-
bitter and reputed to be an excellent vegetable hke the cucumber; in South Africa
locally known as “ jelly-melon ” and occasionally cultivated.
9. C. heptadactylus Naud. in Ann. Sci. Nat. 4me. ser. 11: 24 (1859); Sond. in
FI. Cap. 2: 497 (1862); Cogn., Mon. Cucurb. 497 (1881); Pflanzenreich 275. 2: 142
(1924); Burtt Davy, FI. Transv. 1: 229 (1926).
Perennial. Rootstock woody. Stems prostrate, usually rather slender, branched,
geniculate at the nodes, angular, striate or sulcate, shortly pilose or hirsute, up to about
1 • 50 m long. Leaves rather stiff, greyish-green or canescent, 5-7-parted nearly to
the base; the lobes linear, acute, entire or with one or a few teeth and revolute margin,
thinly setose to nearly glabrous above, setose-scabrid and white-punctate beneath,
the terminal one 4-10 cm long and 1-4 mm wide (rarely wider), the lateral ones shorter,
the lowermost ones much shorter; petioles slender, hairy, 3-40 mm long. Tendrils
usually very short. Flowers dioecious. Male plant: flowers fasciculate or shortly
racemose (occasionally by reduction solitary); peduncle 1-10-flowered, slender, hairy,
up to about 2 cm long; receptacle obconic, thinly hispid, 4-6 mm long; sepals narrowly
triangular, 1 -5 mm long; corolla pale yellow, shortly pubescent, 3-4 mm long, petals
shortly cuspidate. Female plant: flowers solitary on short pedicels; ovary covered
with soft thick bristles, ellipsoid. Fruit ovoid or ellipsoid, with dark longitudinal
bands, 3-5 cm long, 2-4 cm in diam., covered with thick subcylindric callous-tipped
and minutely setose-mucronate blunt spines. Seeds oblong, 6-7 x 3-4 x 1 • 5 mm.
Cape.- — ^Vryburg: Rcxigpr Hb. no. 26976 (SAM); Henricil6-, 176 (PRE). Kuruman:
24 miles E. of Korannaberg on road to Tsenin, Rogers 12569 (BOL); Esperanza,
Esterhuysen 2146 (BOL, PRE). Hay: Griquatown, Wilman s.n. (PRE); Papkuil,
Wilman Hb. no. 1298 (KMG); Asbestos Hills, Clifton, Acocks 2132 (PRE, KMG);
Bernouilli, Wilman Hb. no. 26950 (SAM); Witwater, Wilman Hb. no. 1408 (KMG).
Barkly West: Asbestos Mts., Hutchinson 3020 (BOL); Daniels Kuil, Wilman Hb. no.
2994 (KMG, BOL, PRE); Lewis Hb. no. 54221; 53488 (SAM); Leistner 621 (PRE,
KMG); Hoekplaas, Acocks 1856 (PRE); Pniel, Acocks 1579 (PRE, KMG); 1538
(PRE); near Vaal Hartz, v.d. Linde s.n. (PRE). Kimberley: Kolbe 4001 (BOL);
Riverton, Wilman Hb. no. 570 (KMG); Kenilworth, Wilman Hb. no. 22012 (PRE);
Spytfontein, Marloth 842 (PRE). Herbert: Honeynest Kloof, Wilman Hb. no. 569
(KMG, PRE). Hopetown: Muskett in Hb. Bolus no. 2561 (BOL). De Aar: Quagga-
fontein. Story 1085 (PRE). Victoria West or Richmond: Winterveld, Drege 8183 (L.).
70
Colesberg: Burke 139 (SAM); Zeyher 591 (SAM). Middelburg; Leistner 629 (PRE);
Grootfontein : Tlieron 719 (PRE). Albert: Burghersdorp, Pocock 133 (GRA).
Transvaal. — Christiana: Burtt Davy 12996 (PRE). Yentersdorp: Goedgedacht,
Sutton 551 (PRE).
Orange Free State. — Jacobsdal: Grysbank, Schweickerdt 1139 (PRE). Fauresmith:
Smith 5277a (PRE); Pont in Herb. Henrici 2942 (PRE); Codd 3410 (PRE). Bloem-
fontein: near Besters Put, Burtt Davy 11773 (PRE). Bethulie: near Bethulie, Flanagan
1501 (BOL, PRE, SAM). Rouxville: Commissie Drift, Acocks 12517 (PRE). “ Caledon
River” (probably Rouxville): Zeyher 590 (PRE).
This species is quite distinct by its deeply dissected leaves with very long and
narrow linear lobes unlike those of all other species, and its dioecious flowers. A
specimen collected by Bolus (no. 13110, BOL) is labelled “ Hex River, De Dooms ”.
This locality does not seem to link up with the area of the species and the locality or
the label may be wrong.
10. C. kalahariensis A. Meeuse, sp. nov., aff. C. heptadactylo et C. africano, sed
radicibus tuberiferibus praecipue differt.
Dioicus, perennans. CauJes procumbentes, elongati, ramosi, robusti, angulato-
sulcati, minute sparseque hirsuti demum scabri vel subglabri ad 3 m longa et 0-5
cm diam., ad nodos geniculati radicantes, radicibus simplicibus fibrosis bipedalibus,
extrema radice tubere fusiform! vel subcylindrato spongiosi-carnoso 5-20 cm longo
2-3 cm diam. ; internodia ad 20 cm longi. Folia subcanescentia, rigida, ambitu ovata
vel ovato-oblonga, profunde 3-7-fida, utrinque minute setoso-hispidula demum scabra,
5-15 cm longa 3-9 cm lata, lobis oblongis vel linearibus acutis vel subacutis vel
subapiculatis crasse dentatis vel pinnatilobatis 1-5 cm longis. Cirrhi robustiusculi,
satis breves, basin versus distincte incrassati minute aculeato-setosi demum scabri.
Flores masculi solitarii vel interdum fasciculati, pedicellis gracilibus breviter setosis
1-2 cm longis, receptaculo anguste campanulato, dense setoso ca. 7 mm longo, sepalis
lineari-subulatis, setosis, ca. 5 mm longis, corolla flava, 6-7 mm longa. Flores feminei
solitarii, pedicellis brevibus 4-6 mm longis ca. 1-5 mm crassis, sepalis et corolla ut
in mare sed majoribus, ovario ellipsoideo vel ovoideo dense setoso. Pepo ellipsoideus
vel oblongus, aculeatus fasciis alterne albo-viridibus et fusco-purpureis longitudinaliter
variegatus, 3 • 5-5 cm longus 2-3 cm crassus, aculeis crassulis compressis subuncinatis
obtusis vel subacutis minute mucronatis ca. 5 mm longis. Semina ovato-elliptica,
albida, 6-7 mm longa, 3-4 mm lata 1 • 5-2 mm crassa.
Type: Story 5320 (male and female plant), grown from seed at the Roodeplaat
Horticultural Research Station (seed originally collected by Dr. R. Story at Nama Pan,
South West Africa), in National Herbarium, Pretoria.
Perennial dioecious creeper. Stems annual, branched, rather stout, angular-
sulcate, minutely and sparsely hirsute when young, later becoming scabrid to subgla-
brous, geniculate and rooting at the nodes, up to 3 m long and 0 • 5 cm in diam., internodes
up to 20 cm long. Roots fibrous, 50-60 cm long, towards the base dilated to form
cylindric, fusiform or sausage-shaped tuberous portions 5-20 cm long and 2-3 cm in
diam., which are covered with a thin light-brown bark and contain a fleshy-spongy
soft white tissue inside. Leaves somewhat greyish-green, rather rigid, secund, in outline
ovate or ovate-oblong, deeply palmately 3-7-fid, on both sides at first finely setose-
hispidulous, ultimately scabrid, 5-15 cm long and 3-9 cm wide; the lobes oblong to
linear, acute, subacute or subapiculate, coarsely dentate to somewhat pinnatilobed;
petioles rather stout, sulcate, at first finely aculeate-setose, ultimately scabrid, 1-5
cm long. Tendrils stoutish, rather short, distinctly thickened towards the base, minutely
aculeate-setose, glabrescent, turning scabrid. Male flowers solitary or occasionally
fasciculate, on slender, shortly setose-hispid pedicels 1-2 cm long; receptacle narrowly
71
campanulate, densely setose, about 7 mm long; sepals linear-subulate, setose, about
5 mm long; corolla yellow, 6-7 mm long. Female flowers solitary on pedicels 4-6
mm long and 1 -5 mm thick; sepals and corolla as in the male flower, but larger, 7-9
mm and 9-15 mm respectively; ovary ellipsoid or somewhat ovoid, densely setose.
Fruit ellipsoid or oblong, aculeate, variegated by irregularly bordered brownish-purple
longitudinal bands on a greenish-white background, 3-5-5 cm long and 2-3 cm in
diam., spines soft and rather thick, laterally compressed, somewhat uncinate, obtuse
to subacute, finely mucronate, about 5 mm long. Seeds white, ovate-elliptic, 6-7
mm long, 3-4 mm wide and 1 • 5-2 mm thick.
South West Africa. — Grootfontein; Karokowisa, Schoenfelder 10; Nama Pan,
Story 5320 (cultivated at Roodeplaat, PRE, holotype!).
Bechuanaland. — 272 miles N.W. of Molepolole on way to Ghanzi, Story 4986 (fruits
and underground storage organs only; also specimens cultivated in Pretoria); 132
miles N.W. of Molepolole, Story 4895; Kaotwe, van Son in TRY 28803 (all in PRE).
This interesting species is characterised by its dioecious flowers, underground
storage organs and spinose banded fruits with seeds which are rather large for the
genus. In several respects C. kalahariensis resembles C. hookeri, but the latter is an
annual with monoecious flowers and fruits with relatively more spines and smaller
seeds. There is also an affinity with the dioecious C. heptadactylus, but the shape of
the leaf and the different root system of the latter are quite sufficient for a clear
distinction, apart from other details such as pubescence, size of flowers and morphology
of the fruit. Vegetatively, C. kalahariensis also resembles C. flcifoUus (= C. pustulatus)
and if a specimen of the latter (which is monoecious and bears concolorous fruits)
only bears flowers of one sex and no fruits it is almost impossible to distinguish between
them. However, these two species do not appear in the same regions : C. kalahariensis
is a species of deep Kalahari sand soils, whereas C. flcifolius has not been found in
similar localities, and any doubtful specimens can be referred to the one species or the
other if the locality is known.
The curious underground tuberous swellings of the roots are edible and are used
by the Bushmen tribes in north-eastern South West Africa and northern Bechuanaland.
Unfortunately, such appropriate epithets as “ edulis ” and “ tuberosus ” are pre-
occupied in Cucumis, so that the name “ kalahariensis ” is given here, which refers
to its apparently specific habitat.
Dr. Story, previously of the Botanical Survey, has taken a great deal of trouble
to obtain good specimens by growing plants from tubers at Prinshof Experimental
Station, but unfortunately only male plants appeared. Dr. S. Rehm of the Division
of Horticulture raised some male and female plants from Dr. Story’s fruits collected
under no. 5320. From these plants a number of flowering male and female specimens
and some ripe fruits were obtained. These specimens constitute the only complete
material of this species and they were selected as the type, although one would, as a rule,
prefer specimens taken from plants grown under natural conditions. A comparison
between the cultivated specimens and wild specimens suggests that the size of the
leaves as given in the description may be on the large side, but otherwise there is no
essential difference.
11. C. rigidus E. Mey. ex Naud. in Ann. Sci. Nat. 4me. ser. 11: 85 (1859); Sond.
in FI. Cap. 2: 497 (1862); Cogn., Mon. Cucurb. 507 (1881); Pflanzenreich 275.2:
154 (1924). Type: A specimen leg. Drege in north-west Namaqualand near the Orange
River (“ Gariep ”), presumably the actual type in P (iso.! in L).
C. rigidus E. Mey. ex Arnott in Hook., Journ. Bot. 3: 278 (1841), nomen tantum;
Drege, Zw. Pflzgeogr. Doc. 92: 176 (1943), nomen nudum.
72
Perennial monoecious prostrate to erect or suberect canescent suffrutex. Stems
firm, robust, branched, angular, densely covered with adpressed thick stiff hairs, up
to about 0-5 m rarely 1 m, long, the older parts in the grooves between the angles
with thin adpressed white hairs and aculeate with thick hard white setae, ultimately
glabrescent but persistently white or canescent. Leaves stiff and thick, canescent, ovate
in outline, truncate or slightly decurrent at the base, rounded at the apex, with an
irregularly and sparsely dentate or dentate-lobulate margin, very scabrid-setose turning
punctate-scabrid, 3-6 cm long and wide, entire to crenulate-lobulate to more or less
3-lobed with ovate obtuse lobes of which the central one is the longest, separated by
rounded sinuses; petioles rigid, firm setose-scabrid, 1-5 cm long. Tendrils 0. Male
flowers subsessile or on pedicels up to about 1 cm long, usually solitary; receptacle
obconic-campanulate, densely canescent-villous, 4-6 mm long, sepals triangular, about
2 mm long, hairy like the receptacle; petals hirsute outside, about 4 mm long. Female
flowers solitary, peduncles up to 1 cm long; ovary oblong, densely covered with rigid
thick bulbous-based setae. Fruit ovoid to ellipsoid, concolorous, 4-6 cm long, 2-5-4
cm in diam., rather sparsely covered with soft flattened, 5-6 mm long spines which end
in a bulbous-based mucro. Seeds 6-7 mm long.
South West Africa. — Luederitz: Kahanstal, Dinter 8128 (B, BOL, PRE). Warmbad:
S. of Warmbad, Pearson 4021; 4023 (BOL).
Cape. — Namaqualand: near Orange River, Drege s.n. (L, iso.!); Bethany Drift,
Pearson 6049 (BOL); Anisfontein, Pillans 5300 (BOL); Sendlingsdrift, Pillans 5501
(BOL); Doornpoort, Pearson 6911 (BOL); Drieklip, Tucker Hb. no. 13235 (BOL);
Richtersveld, Marloth 12330 (BOL, PRE).
12. C. africanus L.f., Suppl. PI. 423 (1781), non sensu auct. Type: specimen in
Hermann Herb. (BM), corresponding to Herm., Parad. Bot. t. 134.
C. hookeri Naud. in Card. Chron. 30: 1503 (1870); Illustr. Hort. 1871: 239 (1871);
Journ. Bot. 9: 58 (1871); Cogn., Mon. Cucurb. 503 (1881); Pflanzenreich 151 (1924).
Type: cultivated specimens in Paris (grown from seed) in P, holo.l; photo.! in PRE.
Annual herb. Stems prostrate, slender to stout, branched, sulcate, shortly hirsute
or pilose or occasionally aculeate on the raised lines between the grooves, turning
scabrid, up to about 1 m long. Leaves ovate-cordate in outline, deeply 3-5-(7-) lobed,
2-10 cm long and a little less in width, on both surfaces rather sparsely setose-strigose,
usually more densely so below, or in plants from very arid localities setose-scabrid
to aculeate-scabrid with thicker bulbous based white stiff hairs, often glabrescent and
turning punctate-scabrid; the lobes broadly elliptic to lanceolate obtuse to acute
lobulate or coarsely dentate, sometimes finely so to subentire, separated by usually
very distinctly rounded sinuses; central lobe the largest, often pinnatilobed, basal
lobes much smaller, petioles hispid to aculeate, 2-8 cm long. Flowers monoecious,
bright yellow. Male flowers solitary, fascicled or in a contracted raceme; receptacle
narrowly campanulate, hirsute, usually 4-6 mm long; sepals subulate or filiform,
1-5-4 mm long; corolla 10-15 mm long (in specimens from arid localities, such as
Little Namaqualand, calyx and corolla smaller). Female flowers: pedicels short or
somewhat elongated, occasionally up to 6 cm long; calyx and corolla as in the male;
ovary fusiform or oblong to narrowly ellipsoid, densely covered with bulbous-based
setae soon developing into blunt conical protuberances carrying terminal thick setae.
Fruit rather variable in size and shape, ellipsoid to oblong or sometimes subcylindric,
more or less densely covered with flattened stout blunt conical-cylindric 5-10 mm
long spines which are usually mucronate by the base of the setae originally terminating
them, when ripe pale greenish white with broad purplish brown bands, consisting of
a number of more or less completely fused dots and/or longitudinal streaks, 3-9 cm
long and 2-4-5 cm in diam. Seeds 4 -5-5 -5 mm long, 2-2-5 mm wide and 0-75 mm
thick.
73
Type specimen; Judging by the original description, the sole basis of the species
seems to be “ Cucumis africanus echinatus minor. Hyslrix vegetabilis vulgo ” Herm.,
Parad. Bot. 133, t. 134. Strangely enough the plate cited by the younger Linnaeus
(t. 36, not 134) corresponds with p. 134, i.e. the description of “ Cucumis echinatus
colocynthidis folio ”, not that of “ Cucumis africanus echinatus minor ”, which name
with the correct page number (133) was also cited by Linnaeus. One might contend
that Linnaeus’s name was based on heterogeneous elements and, therefore, a nomen
confusum. On the other hand it is known that several of the plants dealt with in the
“ Supplementum Plantarum ” were shown, or lent, to Linnaeus by Thunberg, but in
all or the majority of such cases Linnaeus expressly mentions Thunberg’s name and
there is no reference to Thunberg in the original description of Cucumis africanus.
As there is no actual type specimen, the specimen preserved in Hermann herbarium
in BM — SLOANE must be regarded as a typotype and this specimen is, according
to Mr. C. Jeffrey of Kew, to whom I am indebted for the information, undoubtedly
the stripe-fruited plant known as C. hookeri Naud. Thunberg who was the first to use
the name Cucumis africanus again had collected a plant and this specimen is still in
the Thunberg herbarium at Uppsala, but his plant is the same as C. zeyheri. Naudin
in his monograph of the genus Cucumis (1859) gave the first critical discussion and he
was already aware of the uncertainties of Linnaeus’s name. He declared, among
other things, that this species is synonymous with Cucumis africanus echinatus major
(sic! recte minor), vulgo hystrix vegetabilis Herm. Parad. Bot. p. 138 and, as far as
he could see, not with “ C. echinatus Colocynthidis folio Herm. op cit., tab. 134”.
Naudin of course rightly assumed that Linnaeus had wrongly associated the plate
(again incorrectly cited as “ 134”, incidentally) with the description of a different
plant, (Naudin’s quotation is very inaccurate though, p. “ 138 ” instead of 133, “major”
instead of “ minor ” and “ vulgo hystrix vegetabilis ” instead of “ Hystrix vegetabilis
vulgo ”). However, this did not clear up the case. Naudin pointed out that several
specimens collected in South Africa agree with Herman’s description of C. africanus
echinatus major which plant was said to have been grown from seed ex Africa, whereas
C. echinatus colocynthidis folio was a plant that appeared spontaneously and whose
origin was unknown to Herman. Naudin suggested that the latter species might have
been Cucumis anguria L., known in Europe for a long time and frequently cultivated
(hence a possible subspontaneous appearance). All subsequent authors like Sonder
in Flora Capensis and Cogniaux in his monographs follow Naudin’s interpretation,
which is unfortunate because practically all authors used the name C. africanus for the
species with concolorous yellow fruits for which the valid name is now C. zeyheri.
South West Africa. — Kaokoveld: 22 miles S. of Ohopoho, de Winter & Leistner
5628 (PRE). Outjo: between Kamanjab and Outjo, de Winter 3070; Franzfontein,
Liebenberg A95Q^. {fWE). Grootfontein; near Grootfontein, S’c/iocn/cWcr S413; S447
(PRE). Omaruru; Brandberg, Rodin 2721 (PRE); also in (BOL) but wrongly labelled
(“ Otjiwarongo ”). Okahandja: 15 miles from Okahandja on Windhoek road, de Winter
2698 (PRE). Swakopmund; Haikamchab, Go/p/n <6 Pcoraon 7508 (SAM). Windhoek:
Kuiseb River Bed near Harris, Pearson 9510 (BOL); near Steinhausen (Farm Aurora),
de Winter 2425 (PRE); halfway Windhoek-Okahandja, de Winter 2701 (PRE).
Gobabis: Voolsangom, N.W. of Gobabis, de Winter 2450 (PRE); Sandfontein,
Wilnian Hb. no. 27030 (SAM); Omitara, Liebenberg 4570 (PRE). Rehoboth: Reho-
both. Fleck 703 (Z). Luederitzbucht ; Aus, Dinter 6237 (B, BOL); Farm Landsberg,
Kinges 2128 (PRE); Bethanien, Dooms, Range 1301 (SAM); Inachab, Dinter 973 (Z).
Keetmanshoop: Lowenfluss (Chamob), Fenchel 41 (Z); Holoog, Pearson 9704 (BOL).
Warmbad: Great Karasberg, Narusdas-Sud, Pearson 8219 (BOL); 15 miles N. of
Karasberg, Wilman 270 (BOL, PRE). Locality not traceable : “ Churumanus ” (in
Rehoboth distr.?): Fleck 505 (Z); without locality: Fleck 43 (Z).
Bechuanaland.— Ngamiland: Curson 486 (PRE); Mabele a Pudi (N’Gami) van Son
Hb. no. 28798 (PRE); Kaotwe, van Son Hb. no. 28803; Mochudi: Rogers 2352
(BOL).
74
Cape. — Namaqualand: Goodhouse on Orange River, Wilman 256 (BOL); Aggenys,
Pearson 12931 (BOL); Klipplaat, Pearson 3302; 3940 (BOL). Gordonia: Kalahari
Gemsbok Nat. Park, Story 5579; Leistner 1133; Brynard 409 (PRE); Auob River
in Gemsbok Game Reserve, Kinges 2004 (PRE); Gordonia, Lewis Hb. no. 53327
(SAM). Kenhardt: near Kenhardt, Comins 641; Acocks 18821 (PRE); Jagbult,
40 miles N.W. of Marydale, Story 1132 (PRE). Kuruman: 165 miles from Kuruman
on road to Witdraai, Story 5478 (PRE); Kuruman, Esterhuysen 802 (BOL, PRE);
Lewis Hb. no. 53505 (SAM); Pole Evans 2081 (PRE). Vryburg: Palmyra, 60 miles
N.W. of Vryburg, Rodin 3677 (BOL, PRE); Brueckner 1103 (PRE). Barkly West:
Daniel’s Kuil, Lewis Hb. no. 54022 (SAM); Boetsap, Acocks & Hafstrdm H1389
(PRE). Hay; Floradale, Esterhuysen 2432 (BOL). Kimberley: Magersfontein,
Wilman Hb. no. 22011 (PRE); Herbert: Ramah, Wilman Hb. no. 576 (KMG); Honeynest
Kloof, Wilman s.n. (PRE). Prieska: near Prieska, Bryant J. 282 (BOL, PRE, J); 285
(PRE). Hopetown: Muskett m Hb. Bolus no. 2407 (BOL). De Aar: Quaggafontein,
Acocks 12606 (PRE). Calvinia: Schmidt 570; 575 (PRE). Williston: near Williston,
Theron 1630 (PRE). Fraserburg: Bolus 10386 (BOL). Victoria West: Hazel Halt,
near “Three Sisters”; Smith 2488 (PRE). Prince Albert: Merweville, Marloth
13668 (PRE). Graaff-Reinet : Farm Rietvlei, Galpin 11511 (PRE). Somerset East:
/. L. Drege 510; 679 (GRA).
Transvaal. — Soutpansberg : Messina, Rogers 20797; 20999 (PRE); 21525 (GRA);
Moss & Rogers 5050 (J, partly); s.n. (J, partly); Tshakoma, Obermeyer Hb. no. 30159
(PRE). Sibasa ; Sibasa, Junod s.n. (PRE). Barberton : Komatipoort, Moss & Rogers
519 (GRA).
Also recorded from Southern Rhodesia.
This species has been confused in the more recent treatments of this family and in
the South African herbaria with C. zeyheri and C. myriocarpus (— C. dissectifolius).
It is quite distinct by its larger flowers and the morphology of the fruit. It is, for
instance, extremely doubtful if Cucumis zeyheri or C. myriocarpus occur in South
West Africa at all and in Bechuanaland Protectorate they are found only in the extreme
south-east (Mochudi, etc.). All records of these species from South West Africa are,
therefore mostly referable to “ Cucumis hookeri ” and partly to C. ficifolius (= C.
pustulatus), a few perhaps to C. kalahariensis. The fruits of C. africanus occur in two
not very sharply distinct forms, in a large oblong sybcylindric shape and in a smaller,
more ellipsoid one. These two shapes are rather well correlated with the taste; the
large ones are a relished refreshment and source of water among South West African
natives, whereas the smaller ones are usually bitter and poisonous. The actual condition
is somewhat more complicated, because there are three types of fruits with increasing
amounts of the bitter substance and one non-bitter form. The genetical implications
of this variation are very interesting and are at present being studied by the Division
of Horticulture, Pretoria. A similar case is Cucumis metuliferus which species also
occurs in two forms, one with bitter unpalatable fruits and one with non-bitter edible
fruits.
13. C. myriocarpus Naud. in Ann. Sci. Nat. 4me. ser. 11: 22 (1859); Rev. Hort.
106 (1860), cum ic. ; Sond. in FI. Cap. 2; 496 (1862); Cogn., Mon. Cucurb. .502 (1881);
Pflanzenreich 275.2: 150 (1924), emend. Schweick. in S. Afr. Journ. Sci. 30: 459
(1933); Adamson and Salter, FI. Cape Penins. 739 (1950). Type: Burke, Vet River
(K, holo.!).
C. prophetarum sensu Jacq., Hort. Vindob. 1, t. 9 (1770); Thunb., FI. Cap. ed. Schultes
36 (1823); Ser. in DC, Prodr. 3; 301 (1828); Schrad. in Linnaea 12: 415 (1838);
non L. (1755). C. dissectifolius Naud., op. cit. 23 (1859); Sond., op. cit. 496 (1862);
Cogn., op. cit. 492 (1881); 135 (1924); Burtt Davy, FI. Transv. 1: 229 (1926). C.
naudianus Sond., l.c., pro parte, excl. type. C. africanus L.f. var. myriocarpus (Naud.)
Burtt Davy, op. cit., 229. C. merxmuelleri Suesseng. in Trans. Rhodesia Sci. Assoc.
43: 61 (1951).
75
Annual. Stems prostrate or occasionally twining at the ends, fairly slender,
branched, angular-sulcate, at first pilose, later on the angles covered with short thick
curved spinose setae later turning rough or scabrid, 0-5-2 m (usually 0-75-1 -25 m)
long. Leaves herbaceous to somewhat rigid, green or faintly canescent, in outline
ovate or suborbicular to oblong, with a wide and shallow basal sinus, nearly glabrous
and slightly scabrid or somewhat setose-pilose above, or sometimes densely so, more
or less densely hirsute-setose or later scabrid below, 4-10 cm long and 3-7 cm wide,
deeply or rarely shallowly palmately 3-7-lobed; the lobes variously dissected or lobulate
but the ultimate segments usually rather narrow, rarely broad, often denticulate;
petioles shortly and often densely aculeate-hispid, later usually scabrid, 2-10 cm long.
Tendrils slender, usually short. Flowers monoecious. Male flowers fasciculate or by
reduction solitary; pedicels filiform, shortly pilose, 0-5-3 cm long; receptacle hirsute,
3- 5 mm long; sepals subulate, erect 1 -5-2 mm long; petals pale yellow, slightly hairy,
4- 6 (-8) mm long. Female flowers solitary; pedicels 1-5 cm long; ovary usually
rather dark (blackish) when dry, sparsely to rather densely covered with soft spines.
Fruit broadly ellipsoid to subglobose, when nearly ripe green with longitudinal dark
blackish or purplish brown bands turning a bright brownish orange or rusty orange
and less conspicuously striped when quite mature, covered with soft curved spines,
2-4-5 cm long and 1 -5-3 cm in diam. Seeds numerous, pale yellow to white, oblong,
4-5 mm long, 2-2-5 mm wide, 1-1-5 mm thick.
Type: The status of Cucumis myriocarpus Naud. has been discussed by Schweickerdt
who selected a type, viz., Jacquin’s plate in Hort. Vindob. 1, t. 9 (1770), but this is not
permissible because Naudin also cited a specimen retained in C. myriocarpus by
Schweickerdt, hence the type is Burke, Vet River.
The distribution is not as one would expect from previous monographic treatments.
The actual area is from Tanganyika and Southern Rhodesia through the Transvaal,
southern Bechuanalahd and Griqualand West to the eastern Cape Province and Natal.
Not in South West Africa. Occurs as an introduction in the Cape Peninsula and
occasionally in Europe and Australia.
Recorded from the following areas or districts: Bechuanaland: Mochudi; Trans-
vaal: wide-spread, already recorded from practically all districts; Orange Free State:
Boshoff, Hoopstad, Kroonstad, Bethlehem, Harrismith, Bloemfontein; Natal: Utrecht,
Klip River, Bergville, Estcourt; Basutoland; Cape Province: Yryburg, Kimberley,
Barkly West, Albert, Maclear, Sterkstroom, Queenstown, Fort Beaufort, Mount Currie,
Umzimkulu, Graaflf-Reinet, Uitenhage; also recorded from Albertinia (introduced?)
and as an introduction on the Cape Flats.
The following specimens are of special interest: Bu ke, Yet River (Orange Free
State, most probably Hoopstad district, fragment of type in PRE ex Herb. Hooker,
K); Burke 276 from Mooi River (Potchefstroom, Transvaal, in SAM), isotype of
C. dissectifolius Naud. The following gatherings were cited by Cogniaux (1924):
Cape Province, Albert: Cooper 668 (BOL); Queenstown: Gwatyn, Galpin 2040
(GRA, PRE); Uitenhage: Zwartkops River, Ecklon <& Zeyher 1793 (BOL); Cape
Town: Rehmann 2189 (BR) (all cited as C. myriocarpus)-, Basutoland: Junod 2622
(Z); Transvaal, Johannesburg: Modderfontein, Conrath (Z) both as C. dissectifolius.
Cited by Schweickerdt: Vet River, Burke s.n.; Queenstown, Galpin 2049. Orange
Free State: Bethlehem, Thorrold 11405; Steyn 14275; Boshoff, Schweickerdt 1101;
Cape Province: Sterkstroom, Galpin 7726; Basutoland: Watt & Brandwijk 1825;
Griqualand West, Kimberley, Riverton, Wilman s.n. (all PRE).
The species Cucumis dissectifolius and C. myriocarpus cannot be separated.
Cogniaux (1924) keys them out as a perennial against an annual species, but there is
no proof that “ C. dissectifolius ” is always a perennial (if ever). Of the two names,
C. myriocarpus and C. dissectifolius, the first was selected for the aggregate because
76
it has page priority, it has been discussed in great detail in connection with the
segregated species C. leptodermis by Schweickerdt, and it is the name that has been most
generally used, not only in recent taxonomic works such as the “ Flora of the Cape
Peninsula ”, but also in other publications, e.g., in chemical papers. A substance
extracted from the bitter fruits of this species has been referred to as “ myriocarpin
It is felt that the adoption of C. myriocarpus would cause the least inconvenience.
Cucumis merxmuelleri Suesseng. is, judging by the specimens from the type area,
including topotypes compared with the type of C. merxmuelleri at Munich, perfectly
good C. myriocarpus. The specimens referred to the species C. myriocarpus or C.
dissectifolius in various monographs and other publications do not all belong here.
The specimens from South West Africa cited under these two species by Cogniaux,
for instance, are referable to other species (C. hookeri Naud., C. ficifolius A. Rich.
= C. pustulatus Hook.f.). The following specimens from South West Africa erroneously
cited by Cogniaux in 1924 were studied: Range 1052 (cited as “ 1082”, under C.
dissectifolius) is C. ficifolius {= C. pustulatus); Schlechter 2279 and 9911 are probably
C. leptodermis. Pearson 9570 (cited in Ann. Bolus Herb. 3, 1 : 22 (1920) and by Cogniaux
as C. myriocarpus is C. africanus. According to a note on a sheet in PRE this specimen
agrees with Schaefer 375, which makes Schaefer 375 not “ C. dissectifolius’’' but C.
africanus L.f. (= C. hookeri Naud.).
It is interesting that Rehmann collected this species as early as 1881 in the Cape
Flats where it still occurs as a weed.
14. C. leptodermis Schweick. in S. Afr. J. Sci. 30: 359 (1933). C. myriocarpus
Naud. et Auct. plur., ex parte, exclus. type. Type: Schweickerdt 1244 from De Aar
(Cape Province) in PRE (by original designation).
Annual monoecious prostrate herb. Stems several, branched, shortly scabrid,
angular-striate, up to 1 m long. Leaves thinly herbaceous, dark green, nearly glabrous
on the upper surface, hirsute-setose turning scabrid below, with a shallow rounded
basal sinus, 3-5 (-7) cm in diam., 5- or sometimes 3- or 7-lobed; the lobes and the
sinuses between them rounded; central lobe longer; margins dentate; petioles slender
to robust, angular-striate, scabrid or shortly hispid, 1 • 5-9 cm long. Tendrils 3-5 cm
long, scabrid. Male flowers solitary or fascicled; peduncle filiform, pilose, about
1 cm long; receptacle sparsely hirsute, 2-5-3 mm long; sepals subulate, 1-1-5 mm
long; corolla slightly pubescent. Female flowers: ovary subglobose or broadly ellipsoid,
with remote setae. Fruit on a 1-2-5 cm long peduncle, globose, sparsely covered with
short soft spines, longitudinally zoned with darker and lighter green bands (the dark
green bands with 1-4 soft spines, the lighter ones normally espinose), ultimately pale
yellow, concolorous. Seeds pale yellow, ovate-oblong, 5-6 mm long, 2-2 - 5 mm broad,
1-1 • 5 mm thick.
Restricted in its distribution to the area indicated by the following citations.
Cape Province. — Kenhardt: Jagbult, 40 miles W. of Marydale, Story 1133 (PRE).
Prieska: Prieska, Bryant J. 284 (PRE). Calvinia: Calvinia, Smith 2475 (PRE); farm
Diepdrift, Gill 48 (PRE). Barkly West: Silverstream, Esterhuysen 799 (BOL, KMG,
PRE). Kimberley: Power Hb. no. 22013 (PRE). Britstown: Britstown, Schweickerdt
1234 (PRE). De Aar: near De Aar, Schweickerdt 1244 (PRE, type); Story 1076
(PRE). Wellington: Wellington, Marloth 11869 (PRE, reported here as a weed).
Worcester: Hex River near De Dooms, Bolus 13109 (BOL). Caledon: Yilliersdorp,
Schlechter 9911 (GRA, PRE). Victoria West: near Victoria West reservoir. Smith
2405. Murraysburg: Murraysburg, Tyson 290 (SAM). Middelburg: Middelburg,
Gill 91; Comins 693 (PRE); Grootfontein, Verdoorn 1476; Theron 152 (PRE). Graaff-
Reinet: 10 miles S. of Graalf-Reinet, Bolus 666 (BOL); probably Graaff-Reinet :
Sneeuwberg, Bolus 666 (BOL). N.B. — ^This number “ 666 ” is found on two specimens
77
with different though approximate localities; Cogniaux (1924) also cites Bolus 666
as “ Sneeuwberg ”, under “ Cucumis myriocarpus". Aliwal North: Aliwal North,
F. Bolus 39 (BOL); Forbes 540 (NH); Steyn Hb. no. 14272 (PRE); cultivated from
seed ex Aliwal North in Pretoria, Becker 11418 (PRE). Adelaide: Adelaide, Rogers
4496 (GRA). The following two specimens probably also belong here. Albany:
Klipdrift, Schlechter 2279 (GRA, J); Brakkloof, White 39 (GRA).
Orange Free State. — Fauresmith: Fauresmith, Smith 4336A; Hartebeestfontein,
Verdoorn 1398; Grapfontein, Enslin Hb. no. 28482; between Fauresmith and Philip-
polis, Schweickerdt 1284; Kafferfontein, Kies 309. Philippolis: Philippolis, Enslin
Hb. no. 28483 (all PRE).
Cucumis leptodermis is very closely related to C. myriocarpus, both morphologically
and biochemically. However, if complete specimens are available (with fruits) separation
is apparently always possible by the characters mentioned in the key (taken from
Schweickerdt’s publication). C. leptodermis also has a somewhat different appearance —
the lobes of the leaves are usually somewhat rounded and broad (often acute and
narrow in C. myriocarpus), the stems and petioles are often stout (usually slender in
C. myriocarpus), the petioles are relatively longer and the setae on stems and petioles
are short, thickened at the base and usually very sparse on the stems and older petioles.
This combination of characters may serve to distinguish non-fruiting specimens.
C. leptodermis hybridizes quite easily with C. myriocarpus, apparently not only
after artificial cross-pollination, but also in nature. This also emphasizes the close
relationship between the two and perhaps C. leptodermis is not more than a variety
or a subspecies of C. myriocarpus. Quite typical fruiting specimens are always clearly
separable and that is why, for the time being, C. leptodermis is kept up here as a species.
15. C. anguria L., Sp. PI. ed. 1: 1011 (1753).
This species as originally circumscribed is apparently derived from a wild African
species described as C. longipes Hook, and the two forms, differing slightly in the shape
of the fruits and the development of the soft spines of the fruit, can for convenience
be separated as varieties. For a discussion see A. Meeuse in Blumea Suppl. 4: 196-205
(1958).
{a) C. anguria L. var. anguria.
C. anguria L., l.c.; Ser. in DC., Prod. 3: 301 (1828); Naud. in Ann. Sci. Nat. 4me.
ser. 11: 11 (1859); 12: 108 (1859); Griseb., FI. Br. W. Ind. Isl. 288 (1860); Hook. f.
in Bot. Mag. 96: t. 5817 (1870); Cogn., Mon. Cucurb. 501 (1881); Pfianzenreich
275.2: 148 (1924); A. Meeuse, l.c. 200 (1958).
Known only in a cultivated and semi-wild state in America.
{b) C. anguria L. var. longipes {Hook, f.) A. Meeuse, tom. cit., 200 (1958).
C. longipes Hook, f., FI. Trop. Afr. 2: 543 (1871); Cogn., op. cit. 491 (1881); 135
(1924); Hiern, Cat. Afr. PI. Welw. 1, 2: 396 (1898). C. chrysocomus sensu Welw.,
Apontam. FI. Angola 586, 589 {‘'‘' chrysocarpa”, sphalm.); Hiern, op. cit., 396, non
Schumach. & Thon. C.figarei Del. (Cat. Hort. Monspel.) ex Naudin var. echinophorus
Naud. in Ann. Sci. Nat. 4me. ser 11: 17 (1859); C. figarei Hook, f., op. cit. 543, ex
parte, excl. syn. C. ficifolius A. Rich, and C. abyssinicus A. Rich. ; C. ficifolius sensu
Cogn. var. echinophorus (Naud.) Cogn., op. cit. 493 (1881); 139 (1924); sensu Hutch,
et Dalz., FI. W. Trop. Afr. 1: 182 (1927); sensu Andrews, Flow. PI. Anglo-Egypt.
Soudan 1: 174 (1950); sensu Keay, FI. W. Trop. Afr. ed. 2, 11: 213 (1954).
Type of variety: The only specimen in Hooker’s own herbarium named Cucumis
longipes is Welwitsch 848 from Loanda, Angola, which was kindly sent on loan by the
Kew herbarium. This is not such a good specimen as the two sheets of Welwitsch
848 (with fruits!) in BM.
78
Annual monoecious creeper. Stems up to 3 m long, branched, sulcate, shortly
aculeate-setose or hispid. Leaves ovate or broadly ovate in outline 3-8 cm, occasionally
up to 12 cm long and 2-6 (-10) cm wide, 3-5- or occasionally 7-lobed beyond the middle
or rarely shallowly lobed, cordate at the base, the margin denticulate or callose-
denticulate ; the basal sinus as a rule broad and rather shallow, with the blade cuneately
decurrent in the centre; both surfaces scabrid or shortly setose-scabrid, lower surface
often more densely so and, in addition, setose-hispid on the slightly prominent main
veins; the lobes usually broad, ovate or obovate to sub-orbicular, obtuse or rounded,
often apiculate or mucronate, usually distinctly contracted at the base; the terminal
lobe the longest (but not so distinctly longer as in C. africanus or C. myriocarpus),
often lobulate; lateral lobes in 5-lobed leaves often somewhat spathulate or obcuneate,
less often lobulate ; basal lobes asymmetrical, somewhat falcate or with basal auricles
rarely lobulate (in 3-lobed leaves the lateral lobes are more markedly falcate and
auriculate); sinuses between the lobes rounded at the base; petioles usually rather
stout, sulcate and setose-hispid, a little shorter to distinctly longer than the blade.
Tendrils often stoutish and sulcate in lower portion, more slender and curvaceous in
upper portion, setose-hispid. Male flowers solitary or fascicled; pedicels slender to
almost filiform, up to 3 cm long, setose-hispid; receptacle obconical, narrowly campanu-
late, 4-6 mm long, shortly hispid, as are the filiform or subulate, 1 • 5-3 mm long sepals ;
corolla 6-9 mm long. Female flowers solitary on a 2-5 cm long, setose-hispid slender
pedicel; receptacle and perianth as in the male; ovary ovoid to ellipsoid or subglobose,
densely covered with soft setae each with a transparent pungent tip. Fruit on an
incrassate, up to 18 cm long pedicel, ellipsoid to subglobose, concolorous, when ripe
light yellowish green to light yellow (usually very densely), covered with long, thin soft
spines with transparent tips, 3-5 cm long and 2-4 cm in diam. Seeds 5-6 mm long,
2-2-5 mm wide.
Occurs in west and central Africa, extending to South West Africa, the Transvaal
and Natal.
South West Africa. — Kaokoveld: Ohopoho, de Winter & Leistner 5156 (PRE).
Ovamboland; Rodin 2685 (BOL). Grootfontein: Okavango, Bunja (grown from
seeds collected by Dr. Rapsom, in Pretoria), Meeuse 9618 (PRE); near Sambusi,
de Winter 4944; near Grootfontein, Schoenfelder S618 (PRE); 30 miles from Groot-
fontein on road to Tsumeb, de Winter 2901 (PRE) ; farm Nutsas, Volk 400 (M). Karibib :
Klein Ameib, Dinter 7093 (B). Okahandja: Okahandja, Marloth 1357 (PRE); Dinter
137 (SAM). Swakopmund: Haikamchab, Go/p/n <6 Pearjow 7508 (PRE). Windhoek:
farm Otijisawa, Krdusel 564 (M). Locality unknown but most probably Gobabis or
Windhoek: “Farm Erichsfelde ”, Volk 1769/1956 (M); Eckenberg-Erichsfelde, Volk
918/1956 (M).
Bechuanaland. — Chukulu Pan between Ghanzi and Molepolole: Story 4938 (PRE).
Transvaal. — Soutpansberg; Dongola Reserve, Codd 4075 (PRE, SRGH). Pieters-
burg: Blaauwberg near Leipzig Mission Station, Codd 8715; Haenertsburg, Moss &
Rogers 895 (J). Letaba: Shilouvane, Junod 2202 (PRE); Magoebaskloof, Gerstner
5393 (PRE). Potgietersrus: Ysterberg, N. of Potgietersrus, Meeuse 9752 (PRE).
Waterberg: Naboomspruit : Mosdene, Galpin M134 (PRE). Lydenburg: Malipsdrift
(cultivated in Pretoria from seed), Meeuse 9615 (PRE); Driekop, Barnard & Mogg
629 (PRE); Meeuse s.n. (from seeds, cult, in Pretoria); Farm Eersterecht, Barnard
& Mogg 418 (PRE). Ermelo: farm Nooitgedacht 10, Potter in Hb. Henrici no. 1569
(PRE). Kruger National Park: The Gorge, vo/? r/cr 2288 (PRE) ; near Rabelais,
van der Schyff 2349 (PRE); Ship Mountain, van der Schyff 1639 (PRE). Nelspruit:
Plaston, Holt 202 (PRE, NH). Barberton: Komatipoort, Burtt Davy 374 (BOL).
Swaziland. — Mbabane, Bolus 11896 (BOL).
Natal. — Newcastle: Charlestown, Wood 5737 (NH). Nongoma: Gerstner 4688
(PRE); “Zululand”: Gerstner 2414 (NH). Mtunzini: Mogg 5913. Inanda: Wood
425 (NH). Durban: Isipingo Beach, Ward 1014 (PRE).
79
Also recorded from many places in Southern Rhodesia, from Bulalima-Mangwe
in the south-west to Umtali in the east and to Urungwe in the Zambesi valley, from
Mazabuka in Northern Rhodesia and several localities in Portuguese East Africa,
its range apparently extending through the Belgian Congo to tropical Africa and to
Nyasaland.
This plant was confused with C. ficifolius A. Rich., which latter name must be
used for the plants better known as C. pustulatus Hook. f. This confusion started
probably because Naudin took up an old invalidly published name (C. figarei) and
included several forms of which Naudin’s variety echinophorus at least partly agrees
with C. longipes. Cogniaux took over Naudin’s arrangement including the var.
echinophorus but used the oldest validly published synonym for the complex. Others
have also used the mcme ficifolius to denote the annual form with long peduncles and
very densely softly spinose fruits. C. figarei is clearly a synonym of C. ficifolius A. Rich,
(non alior.!) and falls away. The name C. longipes is available and is used here in
preference to the var. echinophorus of Naudin.
Cucumis chrysocomus Schumach. & Thonn. was cited by Welwitsch and by Hiern
for a species of Cucumis which is referable to C. anguria var. longipes. Mr. Killick
kindly informed me that the photo of the type in K (also studied by Mr. Hemsley,
at the time working on African Cucurbitaceae) shows clearly that C. chrysocomus is
not a Cucumis but a species of Raphidiocystis {R. caillei). The combination in the
latter genus is, therefore, now made.
Raphidiocystis chrysocoma {Schumach. & Thonn.) Killick & A. Meeuse, comb,
nov. — Cucumis chrysocomus Schum. et Thonn., Beskr. Guin. PI. 427 (1827-1829); in
Dansk. Vid. Selsk. Skrift. 4; 201 (1829); non alior. Raphidiocystis caillei Hutch.
& Dalz., FI. W. Trop. Afr. ed. 1, 1 ; 180 (1927), nomen tantum, in Kew Bull. 1928:
215, descr. et FI. W. Trop. Afr. ed. 2 (Keay), 1, 1: 215 (1954).
16. C. zeyheri Sond. in FI. Cap. 2; 496 (1862); Cogn., op cit., 505 (1881); 152
(1924). Type: Zeyher 582, 583 (S, syns.).
C. africanus sensu Thunb., Prodr. PI. Cap. 13 (1794); FI. Cap. 156 (1811); Ser. in DC.,
Prodr. 3: 301 (1828); sensu Schrad. in Linnaea 12: 415 (1838); sensu Naud. in Ann.
Sci. Nat. 4me. ser. 11: 20 (1859); sensu Sond. in FI. Cap. 2: 495 (1862); sensu Cogn.,
Mon. Cucurb. 504 (1881); Pflanzenreich 275.2: 151 (1924); sensu Burtt Davy, FI.
Transv. 1: 228 (1926); non L. f. C. arenarius Schrad. in Linnaea 12: 416 (1838);
Naud. op cit., 83 (1859); non Schum. & Thonn. C. africanus Linn. f. var. acutilobus
Cogn. in Bull. Herb. Boiss. 3: 418 (1895); Pflanzenreich 152 (1924). C. africanus
Linn. f. var. zeyheri (Sond.) Burtt Davy, op. cit., 229.
Herb with woody perennial thin roots. Stems several, annual, prostrate, very
rarely climbing, usually slender, branched, angular, scabrid or coarsely setose to aculeate
mainly on the angles, up to about 2 m long. Leaves usually green, herbaceous but often
firm, ovate or elliptic to oblong in outline, truncate to subcordate or broadly rounded
at the base, deeply palmately 3-5-lobed, 2-5-10 cm long and 2-7 cm wide; the lobes
acute to obtuse, lanceolate, broadly linear or broader, rotundate to obovate or ovate-
oblong, the margins finely denticulate to coarsely dentate or lobulate, the middle lobe
usually distinctly longer than the others, often lobulate; both surfaces scabrid or shortly
setose-aculeate often ultimately scabrid-punctate, if so, often white-punctate, usually
shortly hispid-setose on the nerves below; petioles shortly setose-hispid to finely
aculeate, often turning scabrid, 0 • 5-8 cm long. Flowers monoecious. Male flowers:
pedicels solitary or fascicled, filiform, shortly hispid, 0-5-2 cm long; receptacle
narrowly c mpanulate, shortly hirsute-setose, 3-4 mm long; sepals subulate, erect,
1-4 mm long; corolla slightly pubescent, 3-5 mm long. Female flowers: pedicels
80
almost invariably under 3 cm long; ovary narrowly ellipsoid to oblong, densely covered
with thick soft setae. Fruit ellipsoid, when ripe concolorous, light yellow, covered
with soft rather short (under 5 mm long) to very short soft spines, 4-6 cm long, 2-4
cm in diam. Seeds 4-5-5 mm long, about 2-5 mm broad, slightly over 1 mm thick.
Occurs in South Africa, from Griqualand West southwards and south-eastwards
to the eastern Cape Province, Orange Free State, Natal and extends northwards to
Southern Rhodesia and Portuguese East Africa, perhaps even to tropical East Africa,
but up to now not found in South West Africa or Angola, as wrongly stated by
Cogniaux in his monographs. Not recorded from Bechuanaland Protectorate but to
be expected in the extreme south and eastern parts.
Recorded from the following: Cape Province: Clanwilliam ?, Vryburg, Taungs,
Barkly West, Kimberley, Hay, Aliwal North, Graaff-Reinet, Uitenhage, Port Elizabeth,
Albany, Victoria East, Queenstown, East London, Stutterheim, Komgha, Kentani,
Mount Currie; Transvaal: practically all districts except the extreme south-west and
extreme south-east; Swaziland; Orange Free State: recorded only from the Senekal
and Fauresmith districts, but it must be more widespread; Zululand (wide-spread)
and Natal: Bergville, Klip River, Estcourt, Weenen, Lions River, Pietermaritzburg,
Camperdown, Richmond, Pinetown, Durban, Port Shepstone; Basutoland. The
following specimens are of special interest;
Cape Province. — Ecklon & Zeyher 1795, said to be from Onderbokkeveld, Clanwilliam
(SAM, isotype of C. arenarius Schrad. non Schum. & Thonn.). Uitenhage: Ecklon
<& Zeyher 1794 (SAM, cited by Sonder & Cogniaux). Aliwal North: Wildschutsberg,
Drege (L.). Graaff-Reinet: Zondags River, Drege (PRE).
Transvaal. — Letaba; Shilovane, Junod 2556 (Z, cited by Cogniaux & Harms 1924
as C. zeyheri); Pietersburg: Houtbosch, Relimann 6311 (BR). Barberton: Galpin
942 (PRE), these numbers cited by Burtt Davy under C. africanus. Lydenburg; Wilms
497 (PRE, cited by Burtt Davy as C. africanus, by Cogniaux & Harms, erroneously
as “ 494 ” as C. zeyheri).
Natal. — Durban: Gueinzius (probably no. 398, cited by Cogniaux 1881 and Cogn.
& Harms as C. zeyheri). Lions River: Howick, Junod 182 (Z); Umkomanzi (Umko-
maas) River, Schlechter 6684 (GRA), both cited by Cogn. & Harms as C. zeyheri.
Portuguese East Africa. — Sul do Save; Ressano Garcia, Schlechter 11894 (GRA),
cited by Cogniaux & Harms as var. acutilobus Cogn. (together with Rehmann 6311
from the Transvaal, see above).
Although the actual syntypes of Cucumis zeyheri were not available for study, a
Gueinzius specimen from Durban, referred to C. zeyheri by Sonder with some doubt,
was seen and several other specimens referred to “ C. zeyheri ” by Cogniaux leave
no doubt about the identity with C. africanus sensu auct. non L.f. As a matter of
fact, Cogniaux, p. 119 (1924), keys out the two on the shape of the lobes of the leaves,
(those of “ africanus ” with broad lobes rounded at the apex, those of zeyheri with
lanceolate acute lobes), but at the same time upholds a var. acutilobus of “ C. africanus ”
(p. 152), which is, therefore, indistinguishable from C. zeyheri. The fruit of C. zeyheri
was described as “ pyriform-globose ” by Sonder which description was copied by
Cogniaux (1881, 1924), but this is fallacious and it must have been either an abnormal
fruit or the way it was preserved. Plants agreeing with “ C. zeyheri ” bear the same
ellipsoid yellow fruits as “ C. africanus ” sensu auct. non L.f.
The fruits are very bitter and are used in native medicine as a (very drastic)
purgative, but this has led to several cases of poisoning through taking an overdose.
C. zeyheri is without any doubt a perennial. Specimens in cultivation have been
growing for nine years already. Reports in the literature that it is an annual species
(e.g., Cogniaux 1924) are erroneous or are based on related annual species confused
with it (C. myriocarpus, C. anguria var. longipes and the true C. africanus L.f. = C.
hooker i Naud.).
81
17. C. ficifolius^. Rich., Tent. FI. Abyss. 1 : 294, t. 53 bis (1847), non alior. Type;
Quartin Dillon, from Abyssinia, in P (photo, in PRE!).
C. pustulatus Hook. f. in FI. Trop. Afr. 2: 544 (1871); Cogn., Mon. Cucurb. 495
(1881); Pflanzenreich 141 (1924); Hutch. & Dalz., FI. W. Trop. Afr. 1; 182 (1931);
Andrews, Flow. PI. Anglo-Egypt. Soudan 1: 172 (1950). ,C. abyssinicm Schimp. ex
Hook, f., I.C., in syn., non A. Rich.
Perennial prostrate monoecious herb. Stems several from a woody rootstock,
rather firm and rigid, elongate, branched, sulcate, armed with short white thick spines.
Leaves rather rigid, cordate-ovate to cordate-oblong in outline, canescent, scabrid,
pedately-palmately 3-5-nerved from the base, 4-11 cm long and 3-9 cm wide, entire
or 3-lobed to deeply 5-fid, lobes obtuse, entire or denticulate, sometimes lobulate;
petioles rather stout, very scabrid, 2-5 cm long. Tendrils usually stout, shortly hispid-
setose. Male flowers solitary or 2-3 together, pedicels filiform shortly hispid-setose,
about 1 cm long; receptacle sparsely hirsute, about 5 mm long; sepals subulate or
narrowly triangular to lanceolate, about 2 mm long; petals puberulous, ovate-oblong,
obtuse, 6-7 mm long. Female flowers solitary, peduncles rather short, 1-3 cm long,
in fruit ultimately attaining 5-10 cm; ovary with thick short spines. Fruit broadly
ovoid to subglobose, concolorous, in sicco turning yellowish-brown, glabrous when
mature, sparsely covered with short hemispherical protuberances representing the bases
of the spines of the ovary, 5-8 cm long and 4-6 cm in diam. Seeds whitish to pale
yellow, oblong, 6 mm long, 2-5-3 mm wide and 1-5 mm thick.
Found in Mauretania, Sudan and Nubia to Somaliland and Arabia, extends into
tropical East Africa, also in Angola and South West Africa.
South West Africa. — Kaokoveld: Ohopoho, de Winter & Leistner 5157 (PRE).
Ovamboland: Olukonda, S'c/i/nz 308 (Z). Grootfontein : Karakowisa, 10;
de Winter 3713 (PRE). Okahandja: Dinter 138 (PRE, SAM). Karibib: Kinges
3451 (PRE). Windhoek: Voigtskirch, 35 miles N.E. of Windhoek, de Winter 2392
(PRE): 8 miles W. of Seeis, Codd 5812 (PRE); Windhoek, Gillman 2 (SAM); Avis-
damm, Liebenberg 4505 (PRE); Osona, Dinter 126 (PRE, SAM); farm Hohenwarte
near Windhoek, Steyn Hb. no. 22544 (PRE) ; between Haris and farm Hoffnungsfelde,
Pearson 9526 (BOL); Hoffnungsfelde, Pearson 9586 (BOL); Rehoboth, Buellsport,
Dinter 8349 (B). Bethanien: Kuperas Range 1052 (SAM). Liideritzbucht : Aus,
Schinz 299 (Z). Hereroland: Luederitz 134 (Z).
Cape. — Barkly West: Newlands, Wilman Hb. no. 1552 (KMG). Probably Hay:
Paardekloof, Cooke Hb. no. 6656 (KMG). De Aar; near Houtkraal Station, Acocks
18831 (PRE).
This species is better known under the name C. pustulatus Hook. f. The confusion
started with Naudin who took up an invalidly published name by Delisle (“ Cat. hort
Monspel.”), and published a description as Cucumis flgarei Del. ex Naud. in Ann.
Sci. Nat. 4me ser. 11: 16 (1869). C. flgarei" includes many forms among them
A. Richard’s Cucumis flcifolius, the latter as v&r. flcifolius. Cogniaux and others used
the name “ C. flcifolius ” as a substitute for Naudin’s name C. flgarei and for a long
time the current application of the name “ C. flcifolius ” was to a plant which was
described by Hooker as Cucumis longipes, an annual. If one goes back to Richard’s
description, it is quite clear that the type of C. flcifolius is characterised by short fruiting
peduncles and short (reduced) tendrils. Richard described the fruits as “ piloso-
echinatis ”, but the plate shows a fruit with short warty protuberances. Moreover,
Richard, in a note under the description, clearly pointed out the differences between
“ C. flcifolius ”, C. prophetarum and C. africanus. “ Elle differe de la . . . seconde
(i.e., C. africanus) par les lobes de ses feuilles tres-obtus, par ses pedoncules courts
et ses fruits non herisses de piquants ”. Naudin, when discussing C. flgarei, mentioned
that he did not see any mature fruits on the specimens named C. flcifolius by Richard,
but he stated that he thought the mature fruits would not have long soft spines when
82
mature and he made a variety “ C.figarei cyrtopodus ” for this form. Hooker may have
been confused by some of Naudin’s conclusions and described the plant again as
Cucumis pustulatus. The type (Quartin Dillon) of C. ficifolius appears to me a good
match of Hooker’s C. pustulatus and must be taken up. This is rather unfortunate,
because the name “ C. ficifolius ” has been used for other species (at least for Cucumis
longipes — C. anguria var. longipes, see no. 15b) by Cogniaux and others, but there
is no other course. Richard’s name C. abyssinicus is of the same date as C. ficifolius
and Hooker, when describing C. pustulatus, mentioned that this might be the same as
his C. pustulatus (he also mentioned Cucumis striatus A. Rich, but, according to
Cogniaux, this is Coccinia adoensis and can be disregarded). However, Hooker also
cited “ C. abyssinicus Schimp. H.C. Abyss, n. 412 (1853 ex herb. Mus. Par.) in A. Rich.
FI. Abyss. I, 294”, which is incorrect because Richard, l.c., only cites Cucumis abys-
sinicus Nobis and does not mention Schimper nor a Schimper specimen (he mentions
Quartin Dillon as the collector). Schimper 412 from Abyssinia is cited by Cogniaux
as Melothria punctata, another Schimper 412 from Yemen, Arabia, is C. pustulatus
(K, teste Killick). At any rate, C. abyssinicus A. Rich, apparently has striped fruits
and, therefore, cannot be the same as C. ficifolius A. Rich.
13. LAGENARIA
Lagenaria Ser. in Mem. Soc. Phys. Geneve 3, 1: 25, t. 2 (1825); in DC., Prodr. 3:
299 (1828); Naud. in Ann. Sci. Nat. 4me. ser., 12: 91 (1859); Sond. in FI. Cap. 2:
489 (1862), ex parte; Benth. & Hook, f., Gen. PI. 1: 823 (1868); Hook, f., FI. Trop.
Afr. 2: 529 (1871); Cogn., Mon. Cucurb. 417 (1881); Pflanzenreich 275.2: 200
(1924); Baill., Hist. PI. 8: 443 (1886); Pax in Pflanzenfam. 4, 5: 29 (1889).
Sphaerosicyos Hook. f. in Benth. & Hook, f., op cit., 824 (1868); op cit., 532 (1871);
Cogn., op. cit., 466 (1881), 99 (1924); Pax, op. cit., 26; Phillips, Gen. ed. 2: 748
(1951). Sphaerosicyus Post. & O. Ktze., Lex. 528 (1903).
Annual or perennial scandent or prostrate plants with musk scent. Stems long,
branched, firm, sometimes rooting at the nodes. Tendrils bifid or very rarely simple.
Leaves entire, angular or lobed to deeply palmately dissected, the margin dentate;
petioles with 2 sessile or stalked glands at the apex which are rarely wanting. Flowers
monoecious or dioecious, rather large, all solitary or the male racemose. Male flowers:
receptacle shallow to narrowly campanulate or turbinate; sepals small, remote; corolla
rotate to saucer-shaped; petals free or nearly so, oblong-obovate; stamens 3, free;
filaments short; two anthers 2-thecous, the third 1-thecous; thecae very much folded;
connective not produced at the apex beyond the anthers, usually broad and flat;
rudimentary pistil 0 or represented by a gland. Female flowers: perianth as in the male;
staminodes 3, small or minute; ovary ovoid to ellipsoid or subglobose, tomentose
or hairy, with 3 placentas and numerous horizontal ovules; style columnar or cylindric,
short or very short; stigmas 3, fleshy, bilobed. Fruit indehiscent, subglobose to oblong,
pyriform, ellipsoid or irregularly shaped with usually a long, often bent, “ neck” with
a leathery to bony pericarp and many seeds embedded in a spongy pulp. Seeds
triangular-oblong to obovate or elliptic, compressed, truncate to sub-bidentate, rarely
rounded at the apex, smooth, marginate; testa tough, leathery to bony; tegmen
membranous; cotyledons, elliptic, radicle conical, subacute.
Type species: Lagenaria vulgaris Ser. — Cucurbita lagenaria L. — Lagenaria
siceraria (Molina) Standi.
Two species, both occurring in southern Africa. The type species, now circum-
tropical, is most probably also of African origin.
83
The monotypic genus Sphaerosicyos Hook, f., here reduced to Lagenaria, differs
from the type species of Lagenaria in its perennial habit and the normally dioecious
(as against monoecious) flowers. The other morphological characters agree very well,
even in rather unusual details such as the glands at the apices of the petioles and the
shape of the seeds. Additional arguments for this reduction are in the first place the
fact that fertile hybrids between Lagenaria siceraria and “ Sphaerosicyos ” have often
been recorded (see, for instance Cogniaux & Harms in Pflanzenreich 275.2; 101).
At the Roodeplaat Experimental Station of the Division of Horticulture near Pretoria
Dr. Rehm observed spontaneous hybridization between the two and found in addition
that artificial cross-pollination yielded about as many successful fertilisations as
artificial pollination of the flowers of one species with its own pollen. The Fi-generation
produced fertile seeds which germinated as readily and as quickly as those of the parent
plants, whereas most interspecific crosses between members of one genus in the
Cucurbitaceae are already sterile in the Fi-generation and do not produce viable seeds.
Another interesting fact discovered by Dr. Rehm is that the same bitter substances
in approximately the same relative quantities occur in the fruits of Sphaerosicyos and of
bitter forms of Lagenaria siceraria. This combination of substances and relative
quantities is rather typical of Lagenaria and Sphaerosicyos whereas this constellation
is not found in the other Cucurbitaceae examined in this respect.
Monoecious; male flowers always solitary; leaves undissected, glands on petiole small; whole
plant densely and softly pubescent; usually cultivated annual, sometimes running wild but
probably also indigenous 1- L. siceraria
Normally dioecious; male flowers usually racemose; leaves usually palmatisect, glands on petiole
conical or subcylindric; plant usually not densely pubescent all over; not cultivated,
perennial 2. L. mascarena
1. L. siceraria {Molina) Standi, in Field Mus. Publ., Bot. Ser. 3: 435 (1930);
Andrews, Flow. PI. Anglo-Egypt. Soudan 1 : 175 (1950).
Cucurbita lagenaria L., Sp. PI. ed. 1, 1010 (1753). C. siceraria Molina, Sagg. Chil.
133 (1782), ed. 2, 316 (1810); Ser. in DC., Prodr. 3; 318 (1828). C. leucantha Duch.
in Lam., Enc. Meth. 2: 150 (1786).
Lagenaria vulgaris Ser. in Mem. Soc. Phys. Geneve 3, 1: 25, t. 2 (1825), and op cit.
299 (1828); Sond. in FI. Cap. 2; 489 (1862); Hook. f. in FI. Trop. Afr. 2: 529; Cogn.,
Mon. Cucurb. 417 (1881); Pflanzenreich 275.2; 201 (1924); Hutch. & Dalz., FI.
W. Trop. Afr. 1; 176 (1931); Robyns, FI. Spermat. Parc. Nat. Albert 2; 399 (1947).
For full synonymy see Cogniaux (1881, 1924).
Type ; The description by Molina was based on material from South America.
It is doubtful whether a type specimen exists, but the name Lagenaria siceraria is
nowadays generally accepted as being correct.
Circumtropical (but apparently of African origin), usually as a cultigen.
Annual, prostrate or climbing softly hairy herb. Stems angular, thick. Leaves
suborbicular-cordate, softly herbaceous, angular or faintly 3-lobed, obtuse or acute
at the apex and with a broad basal sinus, 10-40 cm long and as wide; the margin
dentate; the 5-7 pedately arranged nerves prominent below; petioles rigid, straight,
thick, cylindric, often hollow, 5-30 cm long with sessile glands at the apex. Flowers
solitary. Male flowers: peduncle usually exceeding the petioles; receptacle narrowly
campanulate-funnelshaped, 2-3 cm long; sepals narrowly triangular; petals crisped,
pubescent or tomentose, thickly 5-nerved, 3-4 cm long and 2-3 cm wide. Female
flowers: peduncle usually shorter than in the male flower; ovary ovoid to cylindric,
densely long-villous. Fruit variable in shape and size, ultimately glabrous, green at
first, turning whitish or yellowish at maturity, 10-80 cm long and up to 20 cm in diam.,
often with a narrow “neck” and/or a constriction near the stalk. Seeds 7-20 mm
long.
84
Cultivated by the natives in South Africa for its fruits (Calabashes) of which the
hard outer layers are used for containers and to make dishes, spoons, etc.; the fruits
of some forms with non-bitter fruits are also used as a vegetable when young (local
names of these forms are “ doody ” and “ maranka ”, which names may be of Indian
origin).
A few records of L. siceraria from areas where escapes from cultivation are most
unlikely suggest that this plant also occurs wild in Southern Africa. Examples are,
for instances, Codd & de Winter 5583 from near Letaba Camp, Kruger National Park,
Letaba district; Codd 4283 from near Olifants River Camp, Kruger National Park,
Pilgrims Rest district; Rogers 367 from Nelspruit; and Burtt Davy 10649 from Swaziland
(all in PRE).
2. L. mascarena Naiid. in Ann. Sci. Nat. 4me. ser. 18: 187 (1862).
L. sphaerica E. Mey. ex Drege, Zw. Pflzgeogr. Doc. 197 (1843), nomen tantum; ex
Naud. in Ann. Sci. Nat. 5me. ser. 5: 9 (1866). L. sphaerocarpa E. Mey. ex Arnott
in Hook, f., London J. Bot. 3: 277 (1841), nomen tantum.
Luffa sphaerica E. Mey. ex Sond. in FI. Cap. 2: 490 (Oct. 1862); Wood, Natal PL 3:
t. '289 (1902).
Sphaerosicyos meyeri Hook. f. in FI. Trop. Afr. 2: 532 (1871). S. sphaericus (E. Mey.
ex Naud.) Cogn., Mon. Cucurb. 466 (1881); Pflanzenreich 275.2: 99; Engl., Pflan-
zenwelt O.-Afr., C: 398 (1895); Robyns, FI. Spermat. Parc Nat. Albert 2: 397 (1947)
Type: Naudin did not mention any specimens in his original description, but
stated “ In insulis Mayotte, Nossi-Be . . . The description starts with “ Planta
in Horto parisiensi 10-metralis . . . ” and most probably Cogniaux was correct
when he stated (Pflanzenreich 275.2: 100): Lagenaria mascarena wurde begrundet
auf in Paris kultivierte mannliche Exemplare aus Mayotte ” . The type material was
not available for study, but there is no reason to question Cogniaux’s reduction of
Lagenaria mascarena to Sphaerosicyos sphaericus. At any rate, specimens from the
Comores and Madagascar are indistinguishable from South African material (typified
by a Drege gathering from southern Natal).
Perennial. Stems angular-sulcate, nearly glabrous to subtomentose, up to 10 m
long and over. Leaves rather rigid, pergamaceous when dried, ovate-cordate to sub-
orbicular in outline, varying from shallowly 5-lobed-5-angled to deeply (more than
half-way) palmatisect, 5-18 cm long and as wide, on both surfaces shortly setose-
scabrid and on lower surface sometimes subtomentose; upper surface dark green,
lower surface paler; lobes usually irregularly and coarsely dentate or dentate-serrate
to somewhat pinnati-lobed, the teeth callous-mucronate; apices of lobes usually
acuminate, ending in a long slender mucro; basal lobes sometimes obliquely bilobed;
sinuses between the lobes rounded; basal sinus usually wide and shallow with the
blade broadly cuneately decurrent in the middle; petioles firm, striate-sulcate, 2-8
cm long; glands firm, up to 2 mm long; an axillary bract-like organ sometimes
developed, linear, up to 2 cm long. Male plant: common peduncle glabrous, up to
10-flowered, 5-15 cm long; pedicels 7-30 mm long; bracts minute or small, petiolulate,
triangular; receptacle 3-4 mm high and 9-12 mm in diameter, pubescent; sepals
varying from lanceolate-subulate to subquadrate, 2-4 mm long and 1-2-5 mm wide,
usually acute and distant, petals white, green-veined, papillose, 2 -5-4 -8 cm long;
filaments up to 8 mm long; anthers 6-8 mm long and 5-8 mm wide, yellow. Female
plant: peduncle 2-5 cm long; ovary subglobose to ovoid or ellipsoid, densely tomentose,
12-15 mm long; style about 1 cm long. Fruit dark green mottled with lighter green,
greyish green, white or greenish-yellow, subglobose to subglobose-oblong, or broadly
ovoid, obovoid or ellipsoid, 7-11 cm long and 6-10 cm in diam. Seeds whitish to
yellowish, 11-14 mm long, 4-6 mm wide and 2-3 mm thick.
85
Occurs in east tropical Africa, from Tanganyika to the northern Transvaal and
through Natal to the coastal districts of the Cape Province as far as Knysna, and extends
to the Rhodesias and Angola. Also in the Comores and in Madagascar.
Transvaal. — Soutpansberg: Dongola, near Limpopo River, Verdoorn 2120 (PRE).
Sibasa: van Warmelo 5337/21 (PRE); near Makonda, Codd 6829 (PRE, SRGH);
Munro s.n. (PRE, SRGH); Pafuri, Kruger National Park, van der Schijff 3042 (PRE);
van der Schijff & Marais 3720 (PRE). Barberton: Komatipoort, Rogers s.n. (GRA).
Natal. — Eshowe: Lawn 406; 442 (NH). Lower Tugela: Stanger Beach, Pentz &
10419 (NH). Pietermaritzburg: Smith s.n. (VRE). Durban: Redhill, Prospect,
Forbes 274 (NH); near Durban, Rehmann 8845 (BR, Z); Conrath 754 (Z); Wood
953 (GRA) and most probably the same number (same date on label) in BOL, GRA,
PRE and no. 1461 in Herb. MacOwan (PRE); 3269 (NH); Berea, Wood 5269; 5270
(NH); Galpin 12117 (BOL, PRE); Umbilo, Marriott Hb. no. 22586; 22587 (NH);
Isipingo Beach, Ward 842 (NU, PRE). Pinetown: Umkomaas, van Oosterwijk &
Bruin 234 (PRE). Umzinto: Dumisa, Ifafa Valley, Rudatis 1003 (L); Gerstner 6810
(PRE). Port Shepstone: Shelly Bay, Mogg 12738 (PRE). Southern Natal, probably
Umzinto (between the Umzimkulu and the Umkomaas): Drege s.n. (L, isotype of
Luffa sphaerica Sond. = Sphaerosicyos sphaericus Cogn.).
Cape Province. — Port St. Johns: Isnuka, Galpin 3442 (BOL, GRA, PRE); Port St.
Johns, Schonland 4033 (GRA); Bolus 8911 (BOL); Leighton 2969 (BOL, PRE); Paid
Hb. no. 25385 (BOL, PRE). Willowvale: Willowvale, Rayment Hb. no. 6141 (KMG).
Kentani: near Kentani, Pegler 154 (PRE). Komgha: Kei Mouth, Flanagan 1155
(BOL, GRA, PRE); s.n. (PRE, SAM). East London: Rattray 1356 (PRE). Alexan-
dria: Langebosch Forest Reserve, Story 3244 (GRA, PRE); Archibald 4479 (b) (PRE).
Knysna: Theron 989 (PRE); Duthie 845 (BOL, GRA); 15 miles W. of Knysna,
Peacock Hb. no. 66636 (SAM, PRE); near Goukamma, Fourcade 3929 (BOL).
Portuguese East Africa. — Sul do Save: Lourenco Marques, Junod 371 (PRE);
Gomes & Sousa 3751 (COI, PRE); near Licifo, Pedro & Pedrogao 1414 (PRE); also
recorded from the northern provinces.
14. TROCHOMERIA
Trochomeria Hook. f. in Benth. & Hook, f., Gen. PI. 1: 822 (1867); in FI. Trop. Afr.
2: 524 (1871); Harv., Gen. S. Afr. PL, ed. 2: 124 (1868); Cogn., Mon. Cucurb. 394
(1881); Pflanzenreich 275.2: 184 (1924); Pax in Pflanzenfam. 4, 5: 29 (1889); Burtt
Davy, FI. Transv. 1: 224 (1926); Phill. Gen. ed. 2: 750 (1951).
Heterosicyos Welw. ex Benth. & Hook, f., l.c. and in Trans. Linn. Soc. 27: 33 (1896);
Hook. f. in FI. Trop. Afr. 2: 525 (1871).
Gymnopetalum sensu Bail!., Hist. PI. 8: 445 (1886), ex parte, exclus. type species.
Type species: T. hookeri Harv.
Perennials with tuberous, sometimes large roots and, in the section Trochomeria
(— Eutrochomeria Cogn.), slender prostrate or climbing usually annual stems with
simple tendrils, in the section Heterosicyos (Welw. ex Benth. & Hook, f.) Cogn., erect
stems without tendrils. Leaves usually shortly petiolate, palmately lobed or dissected,
less often undivided, at the base of the petiole sometimes bearing a stipuliform dentate
to fimbriate sessile or subsessile bract. Flowers dioecious, often greenish-yellow,
sometimes precocious; the male ones solitary, fascicled or in racemes, the female
ones solitary. Male flowers: receptacle tubular, funnel-shaped, cylindric or cam-
panulate, very often elongated; sepals very small, ovate, linear, or subulate; corolla
usually rotate; sometimes funnel-shaped in lower portion and the remainder spreading;
petals ovate-triangular to linear-lanceolate or linear-subulate; stamens 3, inserted at
the middle or the base of the receptacle, two 2-thecous and the third 1-thecous;
filaments terete about as long as or longer than the anthers, rarely (in T. sagittata)
86
very short; connective papillose or ciliate at the apex; anthers longitudinally con-
duplicate; rudimentary pistil conical or sometimes {T. sagittatd) depressed, gland-like.
Female flowers: receptacle, calyx and corolla as in the male; staminodes 3, setiform
or linear and elongated, rarely (in T. sagittata) short, thick and fleshy; ovary ovoid
to oblong m outline, often separated from the receptacle by a marked constriction;
placentas 3; ovules horizontal usually numerous; style columnar usually rather long,
but occasionally shorter than the stigma, stigma either 3-lobed and dilated or of two
large flat lobes each usually bilobed, so that the stigma appears to be 4-Iobed. Fruit
ovoid, subglobose or oblong, often pointed, fleshy, red when ripe, usually small, few-
seeded. Seeds white, ellipsoid to subglobose, usually thick, rarely compressed; testa
crustaceous, smooth; tegmen membranous; cotyledons elliptic, flattened or thick
and fleshy, radicle evident or small, conical and blunt.
Found in Africa south of the Sahara to South West Africa and the eastern Cape
Province and in Madagascar.
The morphology of the flowers of T. sagittata is slightly different from those of
the majority of the species, so that the generic description as given by Bentham and
Hooker, Cogniaux and Phillips has to be slightly amended. As will be pointed out
in the discussion of the aberrant species concerned, it certainly belongs in Trochomeria.
Although Trochomeria is reported to be sometimes monoecious, the normal
condition appears to be dioecious. Monoecious specimens (if they occur at all) must
be freaks as also occur in other normally dioecious species {Melothria spp. Kedrostis
spp. and Lagenaria mascarena).
Harms (in Cogniaux et Harms, Pflanzenreich 1924, p. 185) already pointed out
that the leaves in many species are extremely variable so that it is difficult to name a
species with certainty (“ wodurcli die sichere Bestimmung oft erschwert wird ”). This
is quite true, but one should go one step further and reduce several species still
distinguished in the 1924 monograph to synonymy. Apart from the variation in leaf-
shape, the constancy and hence the diagnostic value of the following characters has
been grossly over-estimated;
(a) the absence or presence of a stipuliform bract-like organ at the base of the
petiole;
(b) the relative length of the male peduncle (in relation to the length of the petiole
or the leaf);
(c) the presence of the solitary or fascicled male flowers as against a racemose
sometimes several-flowered inflorescence;
(d) the relative lengths of receptacle and petals.
(Ad a). There are no doubt some species in which the stipuliform leafy organ
is almost always developed (as in the type species) and some in which it is persistently
lacking (as in the subgenus Heterosicyos and in T. sagittata), but Cogniaux already
realised in 1881 that T. macrocarpa occurs in two forms, one without “ bracts ” and one
in which they are developed (“ var. bracteata ” Cogn.). The same phenomenon occurs
in several forms of T. debilis which received different specific names {T. wyleyana,
T. vitifolia, T. baumiana). Specimens showing some leaves with and some without
“ bracts ” or leaves with some small “ bracts ” are frequently encountered.
(Ad b & d). The lengths of the organs concerned vary to such an extent that the
relative lengths, if not strikingly different, have no diagnostic value at all.
(Ad c). The male flowers sometimes occur solitary, in fascicles or in racemes on
one specimen and the arrangement has, therefore no diagnostic value.
87
It is therefore not surprising that a drastic reduction in the number of species
is indicated. The 1924 Pflanzenreich Monograph mentions 19 species and Phillips in
1951 has recorded a total of 26 described species. A preliminary (and necessarily
superficial) analysis of all described species suggests that only about a dozen species
should be retained.
Of the 10 species previously recorded from South Africa (i.e., 7 in the 1924 Mono-
graph and an additional 3 described by Burtt Davy in 1926), for instance, only 4 are
recognised in the present paper, 5 of them are reduced to synonyms and one is excluded
from the genus altogether.
The flowers are not infrequently precocious (i.e., they appear before the leaves)
although this can vary within one species. Precocious flowers are, for instance, not
rare in T. debilis, T. macrocarpa, and T. brachypetala R. E. Fries. One of the species
reduced to a synonym of T. macrocarpa, T. nudiflora Burtt Davy, is based on such a
precociously flowering specimen.
Male flowers with a short funnel-shaped to widely campanulate receptacle not much longer than
wide, 3^ mm long; petals 4-5 mm long; style shorter than the stigma; leaves sagittate,
entire; a diminutive plant 1. T. sagittata
Male flowers with a subcylindric elongate receptacle, distinctly longer to several times longer than
wide; receptacle and petals usually much more than 4 mm long; style much longer than
the stigma; leaves various but usually not sagittate:
Petals very acute, long-acuminate or long-tapering, usually well over 20 mm long 2. T. macrocarpa
Petals acute or subacute but as a rule not long-acuminate or tapering into a narrow acute point,
usually well under 16 mm long:
Leaves variously dissected, but usually very deeply so, the lobes under 1 cm wide; both
surfaces scabrid to puberulous or glabrous, rarely pilose 3. T. debilis
Leaves palmately lobed to about the middle, rarely nearly to the base; the lobes usually
over 1 cm wide (often much wider); both surfaces usually shortly and rather sparsely
pilose-hirsute 4. T. hooker i
1. T. sagittata (Harv. ex Sond.) Cogn., Mon. Cucurb. 400 (1881); Pflanzenreich
275.2: 192 (1924); Burtt Davy, FI. Transv. 1: 225 (1926).
Lagenaria sagittata Harv. ex Sond. in FI. Cap. 2: 489 (1862); Harv., Thes. Cap. 2;
t. 183 (1863).
Type: Sandersdii from Durban (Port Natal) in Herb. Hooker (K), also in TCD;
the Kew specimen proposed here as the actual holotype.
Tuber subglobose, attaining the size of an apple, covered with a brownish bark.
Stems several from the base, prostrate, occasionally climbing, almost invariably
unbranched, slender to filiform, striate, glabrous, rarely over 50 cm long. Leaves
rather rigidly herbaceous, 4-7 cm long and 1-4 cm wide at the base, only in exceptional
cases reaching 13 X 8 cm (with the basal lobes 6-8 cm long), varying from triangular
with a subsagittate base to narrowly sagittate with long basal lobes, or occasionally
5-lobed with two longitudinal parallel basal lobes and two perpendicularly sideways
spreading lateral lobes; margin entire, finely scabrid-setulose, the lobes acute, scabrid
to glabrous or shortly setulose-hairy; petioles slender to filiform, usually glabrous,
1-2 cm long, ebracteate. Tendrils filiform, glabrous or nearly so. Male flowers
usually fasciculate, 1-3, rarely more, per axil, occasionally in a few-flowered subumbellate
raceme; peduncles nodding, capillary, usually shortly hairy, 1-4 cm long; pedicels
of racemose flowers short; receptacle funnel-shaped to narrowly campanulate, 3-5
mm long and 3-4 mm wide at the throat, usually with a few short appressed stiff hairs;
sepals ovate-triangular, up to about 0-5 mm long; petals white to cream or greenish-
white, somewhat fleshy, triangular-ovate, obtuse to subacute, 4-6 mm long, densely
and finely papillose; filaments short, connective broadened and papillose at the apex;
rudiment of ovary represented by a depressed gland-like structure at the base of the
receptacle. Female flowers soWia-ry, pedicels and perianth as in the male; staminodes 3,
short, thick and fleshy, oblong or somewhat obconical; ovary narrowly ovoid or
fusiform-oblong, glabrous and smooth, narrowly apiculate; disc none; style shorter
than the laterally flattened, sub-bilobed and more or less flabelliform stigma. Fruit
on a peduncle up to 4 cm long, ovoid, acute or shortly acuminate, glabrous, faintly
marked with a few fine longitudinal ridges, 16-21 mm long and 11-14 mm in diam.
Seeds in the fruits examined 3-4, subglobose-ellipsoid, smooth, 5-6 mm long, 4-5 mm
wide and nearly 4 mm thick.
Transvaal. — Barberton: Duivelskantoor, Bolus 7772 (BOL).
Natal. — Mahlabatini : Gerstner 4189 (NH). Eshowe: near Eshowe, Lawn 1140
(NH); Gerstner 4086 (NH, leaves exceptionally large, 12-13 cm long, 6-8 cm wide
at the base, basal lobes 4-6 cm long). Kranskop: near Kranskop, Acocks 11620
(PRE, NH). Weenen: Muden, Wylie Hb. no. 28006 (NH, PRE). Umvoti: Greytown,
Meebold 13153 (M). Lion’s River: Karkloof, Wylie Hb. no. 10904 (NH). New
Hanover: Krantzkloof, Schlechter 3197 (BOL, GRA, PRE, also in SAM, where the
number given is 3195, but is probably a mistake for 3197); near Appelbos, Acocks
11828 (NH). Pietermaritzburg: Hawthorn’s Hill, Allsopp 890 (NH). Camperdown:
Botha’s Hill, Wood Hb. no. 1378 (NH). Inanda: Wood 285 (NH, SAM); 7527 (M).
Durban: near Durban, Sanderson 707 (GRA, isotype!); Mogg 11020 (PRE). Ixopo:
near Ixopo, Maxwell Evans 284 (NH). Umzinto: Ifafa, Handley 57 (NU); Dumisa,
Campbellton, Rudatis 1711 (PRE).
Cape. — Umzimkulu: Clydesdale, Tyson 2145 (BOL, SAM). Mount Currie: about
17 miles E. of Kokstad, Killick & Marais 2022 (PRE), Marais 943 (PRE); Kokstad,
Tyson 1827 (BOL). Ngqeleni: Encokos, between Umtata and Port St. Johns, Flanagan
2496 (PRE). Mqanduli: near Mqanduli, Pegler 562 (PRE).
This species differs in some respects from the type species. The flowers are small
and the various floral parts relatively shorter than in typical Trochomeria flowers,
but these differences are all relative and T. sagittata has all essential characters of the
genus such as dioecious flowers, the absence of a disc and the presence of staminodes
in female flowers, the papillose connective and the rudiment of an ovary in the male
flowers, the few-seeded pulpy fruits with tumid emarginate seeds and tuberous roots.
Although it does not resemble the other members of the subgenus Trochomeria
(= Eutrochomeria Cogn.) very much in habit (at a first glance one would sooner take
it for a species of Melothria or Kedrostis), there is no reason to exclude this species from
Trochomeria. A slight emendation of the generic characters mainly pertaining to
relative sizes and shapes of the floral parts is all that is necessary.
As a rule T. sagittata grows among grass and this may be the reason why this small
plant is not so well represented in herbaria through being easily overlooked.
2. T. macrocarpa (Sond.) Hook. f. in El. Trop. Afr. 2: 524 (1871); Cogn., Mon.
Cucurb. 398 (1881); Pflanzenreich 275.2: 188 (1924); Burtt Davy El. Transv. 1: 225
(1926); Meeuse in Flow. PI. Afr. 30: t. 1168 (1954).
Zehneria macrocarpa Sond. in El. Cap. 2: 488 (1862). Syntypes: Transvaal, Potchef-
stroom, Mooi River, Burke 290, Zeyher 579 (S, K).
Trochomeria nudiflora Burtt Davy, op. cit. 57, 225 (1926). Type: Junod 652 (K, holo.l)
from Letaba distr., Transvaal.
Tuber napiform, up to 60 cm long and 25 cm in diam. Stems annual, prostrate
or climbing, more or less pilose, usually sparsely so, up to 1-5 m long. Leaves
suborbicular in outline, palmately digitate or deeply 5-7-lobed, up to 6 cm long and
broad, the lobes (ob) lanceolate to oblong, acuminate, mucronate, entire or somewhat
pinnatisect, sparingly shortly hirsute and scabrid on both sides; petioles up to 25 mm
long, sparsely shortly hirsute. Stipuliform bract suborbicular, pectinate-incised, up to
15 X 15 mm. Male flowers frequently appearing before the leaves, solitary, fascicled
or occasionally shortly racemose (racemes up to 8-flowered); pedicels and common
89
peduncles up to 5 cm long; bracteoles early deciduous; receptacle tubular to faintly
obconical 18-22 mm long, thinly pubescent; sepals triangular-subulate, 1-2 mm long;
petals greenish yellow (“ citrine ”) often tinged with red, spreading with reflexed tips,
linear-subulate from 2-2-5 mm broad base, 18-24 mm long. Female plant: flowers
solitary, pedicels up to 2-5 cm long, ovary 6-8 mm long, separated by a constriction
from the 10-15 mm long receptacle; calyx and corolla as in the male flower; fruit
ellipsoid-oblong, bright red when ripe, 3-4 cm long and 2-3 cm in diam.; seed 8-10
mm long, 5-6 mm broad, 4-5 mm thick.
Recorded from tropical Africa, Angola, Bechuanaland Protectorate, Southern
Rhodesia, Portuguese East Africa and from the following districts. — South West Africa:
Windhoek, Karibib, Okahandja, Outjo, Grootfontein, Okomitundu, Okavango, Ovam-
boland, Kaokoveld; Transvaal: Pietersburg, Potgietersrust, Warmbaths, Rustenburg,
Bloemhof, Ventersdorp, Potchefstroom, Brits, Pretoria, Johannesburg, Heidelberg,
Groblersdal, Nelspruit, Barberton; Swaziland; Natal: Greytown.
This plant is also rather variable, hence the varieties distinguished by Cogniaux
in his 1924 monograph, but they are not clear cut and should not be maintained.
T. nudiflora Burtt Davy is nothing but an early flowering stage, as the flowers often
appear before the leaves.
3. T. debilis /foofc. /. in FI. Trop. Afr. 2: 525 (1871); Ccgn., Mon. Cucurb.
399 (1881); Pflanzenreich 275 . 2 : 189 (1924); Burtt Davy, FI. Transv. 1 : 225 (1926).
Zehneria debilis Sond. in FI. Cap. 2: 488 (1862). Syntypes: Burke 141 and Zeyher
577 (K, S). Z. pectinata Sond., op. cit. 487 (1862). Type: Namaqualand, Buffels
River, Drege s.n. (S, lecto.!; PRE, photo.!). Z. wyleyana Sond., op. cit. 489 (1862).
Type: Namaqualand, Wyley s.n. (S, holo.!; TCD, iso.!).
Trochomeria pectinata (Sond.) Cogn., op. cit. 397 (1881); 187 (1924), excl. var.
subintegrifolia Cogn. in Bull. Herb. Boiss. 3: 418 (1895); Burtt Davy, op. cit. 224
(1926). T. wyleyana (Sond.) Cogn., op. cit. 396 (1881); 187 (1924).
Stems prostrate or climbing up on grasses or bushes, longitudinally sulcate, glabrous
or somewhat hairy, up to about 1 m long, occasionally longer when prostrate, often
much longer when climbing. Leaves: blade 1-3-6 cm by 1-8-11 cm, very variable
in shape and size but usually of 5-7-, rarely 3- or 9-, nearly free, linear or subfiliform to
broadly (ob) lanceolate, entire or coarsely pinnatilobed, acute, acuminate to subaristate
or obtuse, lobes which are 2-5 cm, occasionally up to 8 cm long, and usually under 1 cm
broad, usually more or less lepidote-scabrid with minute broad-based stiff sub-aculeate
hairs especially on lower surface and along the margins; petioles always much shorter
than the blades, usually under 2 cm long; stipuliform bract absent or suborbicular to
flabellate, coarsely and more or less bluntly toothed or with very acute lobes with
sometimes a subulate point, sometimes deeply and irregularly dissected, 0-1 cm long
and broad but occasionally larger. Male Plant: pedicels solitary or fascicled, slender,
under 2 cm long; receptacle narrowly obconical, 1-2 cm long; sepals subulate 1-2
mm long; petals greenish-yellow to olive, triangular to elongate-triangular, acute
to more or less attenuate at the apex, usually 5-12 mm long. Female Plant: ovary
subglobose to ellipsoid 6-17 mm long; calyx and corolla as in the male. Fruit
ovoid-oblong to ellipsoid, usually more or less narrowed into the somewhat pointed
apex, bright red when ripe, 2 -5-3 -5 cm long and 2-2-5 in diam. Seeds few, white,
more or less 8 mm long and more or less 5 mm broad.
Recorded from Angola and Bechuanaland and also the following.— Cope Province:
Namaqualand, Prieska, Hay, Barkly West, Kimberley, Herbert, Phillipstown, Murrays-
burg. Fort Beaufort; Transvaal: Barberton, Nelspruit, Middelburg, Bronkhorstspruit,
Warmbaths, Rustenburg, Waterberg, Potgietersrust; South West Africa: Warmbad,
Keetmanshoop, Windhoek, Gobabis.
90
This variable plant has been described under several names because of the variation
in the degree of development and the dentition of the stipuliform bracts, the differences
in leafshape and in some other characters, none of which are constant.
Zelmeria pectinata was described as monoecious, but the lectotype specimen
(selected by me) in the Sonder herbarium does not show any female features and I
am of the opinion that some mistake was made. This lectotype is undoubtedly the same
as T. debi/is. The other specimen cited by Sonder {Owen s.n. from Durban*) I have
not seen, but it must belong to a different species, because T. debilis does not occur in
Natal. Additional citations under “ T. pectinata ” by Cogniaux in his monographs
refer to specimens from areas where T. debilis has never been collected. For this reason
I prefer to regard Zelmeria pectinata as a somewhat confused entity and although
it has page priority over Z. debilis take up the latter name for the complex. T. pectinata
var. integrifolia is a perfectly typical specimen of T. hookeri (q.v.)
4. T. hookeri Harv., Gen. S. Afr. PI., ed. 2: 125 (1868); Cogn., Mon. Cucurb.
397 (1881); Pflanzenreich 275 .2; 188 (1924); Burtt Davy, FI. Transv. 1 : 224 (1926).
Pilogyne garcini Harv., Thes. Cap. 1: 60 t. 96 (1859), non (L.) Arn.
Zelmeria garcini Sond. in FI. Cap. 2 : 487 (1862), excl. syn., non (L.) Stocks. Z. pectinata
Sond., op. cit. 487 (1862), ex parte.
Trochomeria pectinata (Sond.) Cogn., op. cit. 397 (1881); 187 (1924); Burtt Davy,
I.C., ex parte. T. pectinata var. subintegrifolia Cogn. in Bull. Herb. Boiss. 3: 418
(1895); Pflanzenreich 275.2: 188 (1924). T. subintegrifolia (Cogn.) Burtt Davy,
l.c. T. rotiindata Burtt Davy, op. cit., 52, 221. Type: Galpin 1181 from Barberton
(K, holo. ; PRE, iso.!).
Type: Harvey’s original description and plate in Thes. Cap. 1: 60 t. 96 though
published under the wrong name, were based on a Sanderson specimen (in TCD) and
when he later corrrected himself he did not cite any other specimens, so that this
Sanderson specimen has to be taken as the holotype. Isotypes are present in Herb.
Sonder (according to FI. Cap.) and K.
Stems longitudinally sulcate, shortly hispid when young, glabrescent, up to 2 m
long Leaves herbaceous, broadly triangular-cordate or suborbicular-cordate to
broadly cordate or pentagonal, usually with a wide and shallow basal sinus, more or
less thinly and shortly hispid-setose, 4-9 cm long and 5-9 cm wide, palmatilobed with
usually 5, rarely 3 lobes, usually incised to about the middle with broad ovate or
obovate to obcuneate lobes, occasionally (in the form described as T rotundata Burtt
Davy) more deeply so with oblong or lanceolate lobes, rarely nearly undivided, 5-angled;
lobes usually obtuse or rounded, often distinctly mucronate to apiculate, distinctly
dentate to coarsely pinnatilobed; petioles rather Arm, shortly hairy, 2-3-5 cm long,
almost invariably with a large suborbicular-cordate, long dentate-ciliate stipuliform
bract up to 2-5 x 2-5 cm at the base. Male flowers: pedicels solitary or fascicled
or occasionally racemose on a very short common peduncle. Aliform, hairy, 1-2 (-4)
cm long; receptacle subcylindric, usually distinctly rounded to subtruncate at the base,
slightly widening upwards, 10-16 mm long and about 3 mm in diam., sepals reflexed,
triangular-subulate, distant, about 2 mm long; petals patent to somewhat reflexed,
triangular-oblong or triangular-lanceolate, acute and usually recurved at the apex,
6-10 (-15) mm long, 2-3 mm wide at the base. Female flowers: peduncle slender,
up to 2 cm long, incrassate in fruit; ovary broadly ovoid to subglobose, rostrate
glabrous or puberulous; perianth as in the male but receptacle and petals slightly
wider. Fruit broadly ellipsoid or ovoid, 2-3 cm long and in diam., red when ripe.
Seeds 10-12 x 5 x 3 mm.
* It is, however, known that Miss Owen collected also in the Transvaal and these gatherings are
frequently assigned to Natal.
91
Transvaal.— Soutpansberg; Louis Trichardt, Breyer Hb. no. 24189 (PRE). Sibasa:
Junod s.n. (PRE). Pietersburg or Letaba; Magoebaskloof, Gerstner 5816; 5817
(PRE). Pietersburg: Houtbosch, Rehmann 6304 (BR); 6309 (type of T pectinata
var. subintegrifolia = T. subintegrifolia, BR). Letaba: Shilovane, Junod 1351 (Z);
The Downs, Junod 4243 (PRE). Pilgrims Rest: Mariepskop, Fitzsimons & van Dam
Hb. no. 26255 (PRE); Pilgrims Rest, Rogers 18256 (PRE). Middelburg: Tautesberg,
Young A224 (PRE). Belfast: Machadodorp, Bruce 471 (PRE); Schoemanskloof,
Young A360 (PRE). Carolina: Waterval Onder, R. Guy in Herb. Moss 14977 (J);
Carolina Radermacher Hb. no. 7474 (PRE). Nelspruit: White River, Rogers 23556
(J); Nelspruit, Liebenberg 2536 (PRE). Barberton: Pott 5449 (PRE); Thorncroft
53868 (NH); Umvoti Creek, Galpin 837; Queens River, Galpin 1181. Ermelo;
Mavieriestad, Pott 4890 (PRE).
Natal — “Zululand”: Haygarth in Herb Wood 11018 (NH) Nongoma: near
Nongoma: Gerstner 4693 (PRE). Hlabisa: near Hlabisa, Gerstner 1994 (NH);
Hluhluwe Game Reserve, Ward 1906 (NH, PRE). Entonjaneni: near Biyela Store,
Codd 1899 (PRE). Eshowe: Nkwaleni, Acocks 12955 (PRE). Mtunzini: Woof
11087 (NH, NU); 11387 (BOL, NU, PRE). Mapumulo: Nonoti, Wood 11391 (SAM).
Verulam: Umhlanga, HN no. 8879 (NH). Camperdown: in Hb. Wood
11800 (NH). Umzinto: Umpambinyoni, Rudatis 1980 (NH). Alfred: Harding,
Oliver Hb. no. 18431 (NH). District unknown: Oakfird, Wood 91% (SAM). Natal,
without exact locality: Mrs. Saunders Hb. no. 2311 (NH).
Cape Province. — Prob. Qumbu distr. : Tina River, Schlechter 6396 (COI, GRA,
PRE). Mqanduli: Pegler 585 (PRE). Kentani: Pegler 1197 (BOL). Komgha:
Flanagan 96 (BOL, PRE, SAM). King William’s Town: Buffalo River Valley, Galpin
5915 (GRA, PRE).
Some specimens referred to T. pectinata by Cogmaux, viz., those from Natal and
Mozambique, and certainly those referred to the var. subintegrifolia [= T. subintegrifolia
(Cogn.) Burtt Davy] are referable to T. hookeri. See also the notes under T. debilis
relating to T. pectinata.
Excluded Species
T. rehmannii Cogn. in Pflanzenreich 275.2: 194 (1924). The type specimen, Rehmann
3247 (Z, holo.!) is poor, but it is probably a young plant of Lagenaria siceraria,
at any rate it is not a Trochomeria. The area where the Rehmann specimen was
collected “ Hogeveld, Rietpoort ” is most probably in the Standerton or Volksrust
district, an area very poor in Cucurbitaceae (except cultivated species) and the
chance of finding an undescribed cucurbitaceous plant in that area was remote.
The only species that might occur in that area is Peponium caledonicum and it
might be a seedling of this species if it is not a Lagenaria. In either case the name
can be disregarded.
15. PEPONIUM
Peponium Engl, in Engl. & Prantl, Pflanzenfam. Nachtr. 318 (1897); 2: 75 (1900);
Cogn., in Pflanzenreich 275.2: 212 (1924); Phillips, Gen. ed. 2: 750 (1951).
Peponia Naud. in Ann. Sci. Nat., 5me. ser. 5: 29 (1867); Benth. & Hook, f., Gen.
PI. 1: 823 (1867); Cogn., Mon. Cucurb. 405 (1881); Baill., Hist. PI. 8: 444 (1886);
Pax in Pflanzenfam. 4, 5: 31 (1889); non Grev. (1863).
Peponiella Kuntze, Rev. Gen. 3: 131 (1898).
Type species: Peponia mackenii Naud. = Peponium mackenii (Naud.) Engl.
92
Perennial stout herbaceous climbers or creepers with fibrous or somewhat tuberous
roots. Main stems perennial or annual. Leaves rather large, petiolate, usually
palmatilobed or -sect, rarely entire, dentate; petiole without glands at the apex.
Tendrils usually unequally bifid, rarely 3-5-fid or undivided. Flowers rather large,
white to yellow, monoecious*; the male ones racemose and usually bracteolate, or
solitary, the female ones always solitary. Male flowers: receptacle subcylindric or
narrowly campanulate to narrowly obconical, attenuate but sometimes with a small
subglobose swelling at the base; sepals 5, usually erect to patent, subulate to lanceolate;
petals 5, free, usually patent, obovate, entire, often shortly pubescent or papillose-
puberulous; stamens 3, inserted in the tube of the receptacle, filaments free, anthers
included, cohering into an oblong to subcylindric capitulum, one 1-thecous and two
2-thecous, the locules lengthwise triplicate; connective narrow, not produced at the
apex; rudiment of pistil gland-like or not evident. Female flowers: calyx and corolla
as in the male; staminodes 0, very rarely 3, linear; ovary oblong to fusiform, usually
long-attenuate at the apex, triplacentiferous ; style columnar, stigma tripartite; ovules
numerous, horizontal. Fruit narrowly ovoid or ellipsoid to subcylindric-fusiform,
fleshy with a firmer outer layer and a soft pulp containing the seeds. Seeds numerous,
obovate in outline, much compressed, dark, distinctly margined; testa crustaceous,
smooth.
An African genus of 21 described species (one of which occurs in Madagascar),
with the main centre in east tropical Africa.
It is clear from a study of only a few species of this genus that probably too many
species were maintained by Cogniaux in his 1924 monograph. His main key character
“ Flores masculi solitarii ” as against “ Flores masculi racemosi ” breaks down. In at
least two of the species of Cogniaux’s group with supposedly solitary male flowers
occasional specimens with racemose flowers occur (F. vogelii from West Africa and
P. mackenii from Natal) and in Cogniaux’s group with racemose male flowers, specimens
with solitary male flowers are sometimes found (as in the South African P. caledonicum).
The value of other key characters used by Cogniaux, such as the degree of pubescence
of the vegetative parts and the receptacle and the degree of lobing of the leaves seems
at least doubtful. A revision of the genus as a whole appears to me highly desirable
(an opinion shared by Mr. Killick after a cursory examination of the Kew material)
and would certainly lead to a reduction of the species recognised by Cogniaux by about
a third but, to undertake this, a study of all the African material is necessary and this
is beyond the scope of the present revision. There are in South Africa two species
which are clearly identifiable and which were described so long ago that their names
will probably stand, but additional synonyms among species described from tropical
Africa can be expected. There is, for instance, a strong relationship between P. mackenii
and the tropical P. usambarense (Engl.) Engl.
However, there are in South Africa a few specimens which cannot be quite satis-
factorily named. They are undoubtedly related to some of the species from tropical
Africa. Because of the paucity of the material and the urgent need of a revision of the
whole genus these few gatherings will not be described as new or definitely assigned
to a species recorded from tropical Africa; only their affinities will be indicated.
Male flowers usually solitary on long pedicels; lobes of leaves not constricted at the base; pube-
scence of stems, petioles and sepals usually of long, more or less crinkly and articulate
hairs \. P. mackenii
Male flowers usually racemose on short individual pedicels (sometimes a long-pedicelled male
flower at the base of the raceme in same axil); lobes of leaves usually constricted at the base:
* Cogniaux (1881, 1924) states: “Flores monooici vel dioici ”. However, there is no proof of
a truly dioecious species in this genus (I have never seen one nor a description of a dioecious species)
and I expect that all species are monoecious.
93
Common peduncle (below lowermost male flower) under 12 cm long; pubescence on stems
usually of short hairs, sometimes on young parts forming a short tomentum 2. P. caledonicum
Common peduncle (below lowermost male flower) over 12 cm long:
Plant glabrous or glabrescent 3. P. sp. cf. P. chirindense
Plant persistently more or less hairy 4. P. sp. cf. kilimandscharicum
1. P. mackenii (Naud.) Engl, in Pflanzenfam., Nachtr. 318 (1897); Cogn. in
Pflanzenreich 215 (1924).
Peponia mackenii Naud., tom. cit. 29, t. 3, 4 (1867). Type: cultivated specimens from
Algeria (P, holo., PRE, photo.!, K, iso., PRE, photo.!), see below.
Stems normally climbing, occasionally prostrate, rather stout, longitudinally
sulcate, up to at least 10 m long, at least in the young parts rather densely covered with
multicellular soft curved or crinkly hairs of a drab or pale fawnish colour, less often
pubescence nearly absent. Leaves herbaceous drying thin but not quite membranous,
more or less pentagonal in outline, 6-13 cm long, 6-16 cm broad, palmatilobed to about
the middle or less deeply so with the upper three lobes large, distinct, triangular in
outline, gradually acute or acuminate to cuspidate, the central one slightly larger;
two lowermost lobes often much smaller and oblique or somewhat irregular, sometimes
bilobed; basal sinus usually broad and often rather shallow; upper leaf surface thinly
covered with appressed long or short hairs, glabrescent or becoming smooth, or scabrid
from the sometimes persistent bases of the hairs appearing as minute whitish pustules;
lower surface more densely and more persistently pubescent mainly on the larger veins;
leaf margin rather finely crenate-dentate with usually acute to apiculate-mucronate
teeth; petioles usually densely pilose, rather stout, 3-8 cm long. Male flowers usually
solitary and m this case on 4-18 cm long pubescent pedicels or rarely in few-flowered
racemes on a common peduncle up to about 8 cm long on short usually pubescent
bracteolate pedicels; bracteoles obovate, membranous, sometimes stipitate, about
1 cm long; receptacle obconical to narrowly campanulate, much attenuate towards
the base but a small basal portion again dilated, nearly glabrous, 16-18 mm long and
8-9 mm in diam. at the apex; sepals erect, usually thinly pilose, 5-6 mm long; petals
light clear yellow, papillose, 3-3-5 cm long. Female flowers: pedicels 1-2 cm long;
ovary narrowly fusiform, somewhat pilose. Fruit ovoid-oblong, somewhat narrowed
but rounded at the base, conical-attenuate at the apex, green mottled with white when
young, ultimately glabrous, smooth, red, 6-9 cm long and 3-4 cm in diam. Seeds
dark-brown to dull black, 9-10 mm long, 5-6 mm broad and 1-2 mm thick.
Type: The original material consisted of cultivated specimens grown in Algeria
and sent to Naudin in Paris; authentic material sent to Hooker by Naudin (in K,
photos in PRE) was compared by Mr. Killick with a sheet received from Paris as the
type and judged identical, probably even from the same plant. The seeds were originally
received from Macken who collected them somewhere in Natal.
Natal. — Ngotshe: Ngome Forest, Gerstner 4474; 4506; 4832 (PRE). Nkandhla:
Wood 11426 (NH); Pole Evans 4719 (PRE). Eshowe: Lawn 1860 (NH). Durban:
near Durban, Wood 7913; s.n. (L); Durban Bluff, Marriott Hb. no. 24330; 24331
(NH); Stella Bush, Marriott Hb. no. 36821 (NH); Berea, Wood s.n. (J); Berea Bush,
Wood A5A6 (NH, PRE); 5255 (NBG); 11181 (NH, NU, PRE).
Cape Province. — Port St. Johns: Galpin 3434 (PRE); Mogg 13111 (PRE); Paid
Hb. no. 25382 (BOL). Komgha: Flanagan 1733 (PRE, NBG).
This plant seems to be a forest margin or light forest dweller in the lowland forest
areas in frost-free regions.
Although it has not to my knowledge been recorded outside the area indicated
above, this species has so many features in common with P. usambarense from East
Africa, that the latter may prove to be at best only a variety of P. mackenii. I find
it extremely difficult to separate specimens of P. mackenii with racemose male flowers
1517626-4
94
from specimens in PRE received under the name P. usambarense. The differences that
1 find are in the size of the leaves (they are larger in the East African form), the length
of the male peduncle (longer in P. usambarense) and perhaps in the flowers (larger
in the tropical form), but these do not seem to carry much weight.
2. P. caledonicum (Sond.) Engl., l.c. (1897); Cogn., op. cit. 218 (1924).
Luffa caledonica Sond. in FI. Cap. 2; 490 (1862). Type: Burke 305 (K, holo.; PRE,
photo.!; NBG-SAM, iso.!), see below.
Peponia caledonica (Sond.) Cogn., Mon. Cucurb. 410 (1881); Burtt Davy FI. Transv. 1 ;
230 (1926).
Stems prostrate, stout, sulcate, usually more or less pubescent with usually short
hairs, glabrescent, up to at least 5 m long. Leaves firmly herbaceous drying somewhat
pergamaceous or papyraceous, in outline cordate-orbicular, usually with a broadly
rounded shallow basal sinus and decurrent in the middle on the petiole, 5-partite to
about or somewhat beyond the middle, when mature glabrous above or nearly so,
occasionally somewhat scabrid, finely pubescent beneath but glabrescent and pubescence
ultimately only persistent on main veins, 6-12 cm long and broad; lobes usually
distinctly contracted at the base and approximate to somewhat overlapping, oblong
or obovate to suborbicular, usually rounded or obtuse, rarely (and in this case often
only the central one) acuminate or apiculate; basal ones and rarely the lateral ones
with additional lobule; the margin rather coarsely crenate-dentate with usually acute
and calloso-mucronate teeth; petioles densely and shortly hairy, glabrescent, 2-3 cm
long. Male flowers racemose, rarely solitary; common peduncle densely pubescent,
up to 12 cm long, whole raceme up to 25 cm long and up to about 12-flowered;
pedicels of racemose flowers usually erect, pubescent, articulated at the apex, up to
2 cm long, those of solitary flowers (sometimes also found at the base of a raceme
in same axil) up to 10 cm long; bracteoles ovate or oblong, usually dentate, herbaceous,
hairy, 3-7 mm long, often stipitate; receptacle obconical, pubescent, 16-18 mm long,
6- 8 mm wide at the apex, sepals erect, subulate, 7-8 mm long; petals 24-30 mm long,
puberulous. Female flowers subsessile; ovary fusiform, densely and shortly hairy.
Fruit on a very stout pedicel under 1 cm long, oblong-fusiform or subcylindric with
conical-acuminate apex, green when immature, glabrous, smooth and red when ripe,
7- 10 cm long and 2-3-5 cm in diam. Seeds black, 7-8 mm long, 4-5 mm broad and
1-1-5 mm thick.
Type: Burtt Davy (l.c.) cites Burke 305 and Zeyher 589 as syntypes, and they
are indeed the only two specimens cited by Bonder. The specimen Burke 305 in Herb.
Hooker (K; photo, in PRE) is selected here as the lectotype (iso.! in SAM nunc NBG).
Transvaal. — Rustenburg: Bospoort Dam, Codd 6363 (PRE); Turner s.n. (PRE).
Potchefstroom: v.d. Westhuizen 890; Louw 1430 (PRE); Venterskroon, van Dam in
TRV no. 16934 (PRE). Krugersdorp: Krugersdorp, Jenkins in TRV no. 10106 (PRE);
Muldersdrift, Webster 1 (PRE); Witpoortjiekloof, Moss 5040 bis (J). Brits: Castle
Gorge, Meeuse 9260 (PRE, SRGH). Johannesburg: Johannesburg, GUflllan 102 in
Herb. Galpin no. 6114 (GRA, PRE); Moss 9645; 9778; 16544 (J); Gerstner 6528
(PRE). Vereeniging: Klipriviersberg, Mogg Hb. no. 21566 (PRE, J); Vereeniging:
Mogg 21006 (J); Phillips s.n. (PRE). Bronkhorstspruit : 11 miles N.E. of Bronk-
horstspruit, Codd 2697 (PRE). Lydenburg: Lulu Mts., farm Hoogstepunt, Barnard
495 (PRE). Belfast: 8 miles from Belfast on Stofberg Road, Story 6521 (PRE).
Cape Province. — Hay: Asbestos Mts., Marloth 2065 (PRE); Bergenaars Pad, Acocks
2447 (BOL, KMG, PRE); Paardekloof, Cooke Hb. no. 6653 (KMG). Kimberley:
Spytfontein, Marloth 762 (PRE). Queenstown: near Queenstown, Galpin 2557 (BOL,
PRE); 8285 (BOL, GRA, PRE).
Orange Free State. — Bloemfontein: Mostert 842 (PRE). Rouxville: Wolwekop,
Burke 305 (K, lecto.; PRE, photo.!; NBG, iso.!).
95
In addition, according to Cogniaux (1924) and to Burtt Davy in Kew Bull. 196
(1921), recorded from Bethlehem and Winburg in the Orange Free State, from
Basutoland and from Cradock and Graaff-Reinet in the Cape Province.
P. caledonicmn is usually found in rather exposed open grassy slopes of rocky
ravines or on rocky outcrops in places with rather low winter temperatures and can
apparently stand fairly severe frosts. The specific epithet refers to the Caledon River,
a tributary of the Orange, because Burke and Zeyher collected the plant at Wolwekop
which is not far from the river; it does not refer to the Caledon District of the Cape
Province or any other “ Caledon ” or anything related to Scotland, and this is likely
to be somewhat confusing.
3. P. sp. cf. P. chirindense {Bak. /.) Cogn. in Pflanzenreich 218 (1924).
There is one gathering of a Peponium (Transvaal, Soutpansberg: Wylliespoort,
Smuts 2075 in PRE) which is almost completely glabrous in all its vegetative parts and
has glabrous peduncles. The leaves are larger and thinner in texture than those of
P. caledonicmn and the male peduncle much longer (about 20 cm below the flowers).
This specimen agrees reasonably well with the description of P. chirindense (except
in such details as larger leaves which are also 5-fid rather than 3-fid, less undulate
leaf-margin) and with specimens from the type locality of the latter {Obermeyer 2186
from Chirinda forest, Mt. Selinda, S. Rhodesia). It is for the time being tentatively
referred to this species.
4. P, sp. cf. P. kilimandscharicum (Cogn.) EngL; for synonymy and description,
vide Cogniaux, 219 (1924).
There are three gatherings from the Transvaal which are tentatively referred here;
Gerstner 5902 (PRE) from Soutpansberg, Hangklip Mt. near Louis Trichardt; Story
4089 (PRE, SRGH) from Lydenburg, 18 miles from Spekboom River bridge on road
to Penge Mine; and Codd & Dyer 9089 (PRE) Pietersburg; Blaauwberg.
A rather characteristic feature of P. kilimandscharicum is said to be the (usually
not very dense) pubescence of long, more or less patent flexuous articulate hairs. This
kind of pubescence is found in the three Transvaal specimens. The leaves in these
specimens are larger and thinner in texture, the common male peduncle much longer
and the male bracteoles more membranous than in P. caledonicum and in these respects
they agree with East African material received as P. kilimandscharicum. There is
however, a considerable gap in the distribution from Nyasaland to the Transvaal and
additional gatherings in the Transvaal and in the “ gap ” are necessary before more
definite conclusions can be drawn.
16. COCCINIA
Coccinia Wight & Arn., Prodr. FI. Pen. Ind. Or. 1: 347 (1834); Cogn., Mon. Cucurb.
528 (1881); Pax in Pflanzenfam. 4, 5: 35 (1889); Phillips, Gen. ed. 2: 751 (1951).
Cephalandra Schrad. apud Eckl. & Zeyh., Enum. PI. Afr. Austr. 2: 280 (1836); Linnaea
12: 407 (1838); Sond. in FI. Cap. 2: 492 (1862); Benth. & Hook, f., Gen. PI. 1 : 827
(1867); Hook. f. in FI. Trop. Afr. 2: 550 (1871).
Dioecious perennial climbing, occasionally prostrate, often tall herbs with tuberous
roots. Stems usually angular or sulcate. Leaves often angular to deeply palmatifid
usually cordate at the base, with entire or dentate margins, sometimes with a few black
glands on lower surface near the base between the main veins. Tendrils simple, rarely
bifid. Male flowers subumbellately racemose (often with a solitary male flower at the
base of the inflorescence in same axil) or solitary; receptacle short, cup-shaped to
turbinate; sepals 5, linear or lanceolate; corolla-tube campanulate, lobes 5, broadly
elliptic to broadly ovate, sometimes apiculate, submembranous, veined, often hairy
96
or papillose; stamens 3, inserted in the throat of the receptacle; filaments united or
occasionally free, anthers cohering or connate to form a broad flattened structure,
two 2-thecous, the third 1-thecous, locules conduplicate-sigmoid, connective narrow,
not produced at the apex; rudiment of pistil none. Female flowers solitary; very
rarely racemose; calyx and corolla as in the male; staminodes 3, oblong to subulate;
ovary ovoid or subglobose to fusiform, with 3 placentas and many ovules; style
columnar, stigma of 3 fleshy or flattened lobes. Fruit soft, ellipsoid or subglobose to
oblong-fusiform, many-seeded. Seeds numerous, obliquely obovate to oblong, much
compressed, conspicuously margined, smooth or nearly so.
Type species; Coccinia indica W. & A. (= C. cordifoUa Cogn. non Bryonia
cordifolia L.); the correct name is most probably C. grandis (L.) J. O. Voigt.
A genus of about twenty described species, found in Africa and Asia. The latest
monographic treatment is by Cogniaux in 1881, but several species have been described
since and a revision of the genus as a whole is highly desirable. The species treated
here seem to be clear-cut. In my opinion, this genus belongs in the sub-family
Cucurbiteae, series Cucumerinae, as it is evidently related to the genera Cucurbita,
Lagenaria, Eureiandra, Peponium, Physedra and Adenopus (Cogniaux included it in
Sicyoideae)
Tendrils bifid; plant glabrous 1. C. palmata
Tendrils simple or some bifid, but if so, plant distinctly pubescent:
Leaves sessile, their bases often more or less stem-clasping 2. C. sessUifoUa
Leaves distinctly petiolate:
Whole plant (including calyx except corolla) glabrous or nearly so:
Leaves without black glands at the base, the lobes often oblong or linear, if dentate, rather
regularly so, usually not coarsely lobed or dissected, often rather wide to broadly
rounded at the apex; basal sinus usually narrow; male flowers usually solitary
(E. Cape) 3. C. quinqueloba
Leaves almost invariably with a few black glands near the base on lower surface, the lobes
usually more or less tapering towards the apex (though not necessarily acute); entire
or dentate to lobed or coarsely pinnatisect; basal sinus usually broad and shallow
(at any rate basal lobes never approximate or overlapping); male flowers usually
racemose (Transvaal) 4. C. variifolia
Plant more or less densely hairy on stems, petioles, leaves and/or calyx:
Stems persistently hairy with soft curved or curly hairs; leaves without glands, not scabrid
with raised pustules; male flowers solitary on pedicels usually exceeding 2 cm; ovary
and fruit fusiform, tapering at the apex 5. C. hirtella
Stems often glabrescent or, if retaining the pubescence, hairs stiff; leaves often with glands
on the blade near the base beneath and/or scabrid with small raised pustules on upper
surface or on both sides:
Male flowers often solitary of fascicled; calyx and the subglobose ovary usually rather
densely hairy with curved, rather soft, multicellular somewhat articulate hairs; fruit
ellipsoid or subglobose, rounded at the apex; lobes of leaves generally lobulate to
pinnatisect but margin not rather regularly calloso-dentate (often scabrid-subciliate)
6. C. rehmannii
Male flowers often racemose; calyx and the fusiform-oblong ovary thinly hairy with
short stiff hairs to glabrous; fruit oblong-fusiform to oblong-ovoid, pointed at the
apex; lobes of leaves usually not lobulate, but almost always rather regularly
calloso-dentate 7. C. adoensis
1. C. palmata (Sond.) Cogn., Mon. Cucurb. 540 (1881); M. Wood, Handb. FI.
Natal 54 (1907); Bews, FI. Natal & Zululand 202 (1921); Burtt Davy, FI. Transv.
1 : 231 (1926). Type: Since Sonder took up E. Meyer’s specific epithet, the basionym
is the nomen nudum Momordica palmata E. Mey., typified by a Drege gathering
(near Durban) in Herb. Sonder (nunc S).
Momordical palmata E. Mey. ex Drege, Zw. Pflgeog. Doc. 156, 159, 202 (1843),
nomen tantum. Type: D/cgc s.n. from near Durban, Natal (S, holo., H, iso.!).
97
Cephalandra palmata (E. Mey.) ex Sond. in FI. Cap. 2: 493 (1862). C. mackenii Naud.
in Ann. Sci. Nat. 5me. ser. 5: 17 (1866). Type: Cultivated specimens raised from
seed in France and Algeria, in P.
Coccinia mackenii (Naud.) Cogn., op. cit., 541.
Perennial glabrous or occasionally in young parts somewhat hairy climber. Stems
up to 8 m long, branched, slender, sulcate. Leaves suborbicular to ovate-oblong in
outline, firm, herbaceous to somewhat coriaceous, dark-green above, paler below,
sometimes slightly glaucescent, smooth on both surfaces or finely punctate-scabrid on
upper one, often with a few black glands near the base on lower one, deeply palmately
5-sect, 4-12 cm long and broad; the lobes ovate to oblong-lanceolate, acute or
acuminate, the terminal one slightly larger than the two lateral ones, basal ones
distinctly smaller; the margins from minutely and remotely denticulate to occasionally
lobulate with the dentitions callous-toothed; sinuses between lobes usually subacute
to rounded; basal sinus subrotundate to narrow; petioles slender, striate, 1-6 cm long.
Tendrils bifid, often unequally so. Male flowers: peduncles 1- to racemosely 8-flowered,
slender, 2-10 cm long, one-flowered ones articulated at the apex; pedicels nearly
filiform, articulated at the apex, 8-25 mm long; calyx glabrous, receptacle 3-6 mm long,
5-7 mm in diam., sepals 3-4 mm long; corolla pale yellow, 1-2 cm long; segments
acute. Female flowers: peduncle up to 6 cm long; staminodes oblong, long-pubescent,
2-3 mm long, 1-5-2 mm broad; ovary oblong-linear to fusiform, glabrous, about
1 • 5 cm long. Fruit oblong-fusiform to narrowly ellipsoid, acute, red when ripe, 5-8
cm long, 2-3-5 cm in diam. Seeds dirty white, finely rugulose, 6-5-8 mm long, 3-4
mm broad and about 1 - 5 mm thick.
Portuguese East Africa. — Sul do Save: Lourenco Marques, Earthy (1 ?) = Hb.
no. 18612 (BOL).
Transvaal. — Barberton: Barberton, Highlands Creek, Galpin 785 (BOL, GRA, NBG,
PRE); Williamson 139 (PRE).
Natal. — Utrecht: Glen Atholl farm near Charlestown and Volksrust, Smith 5752
(PRE). Nkandhla: Qudeni forest, 6704 (PRE); Cor/c/ 6991 (PRE). Eshowe:
Lawn 408; 472 (NH). Lion’s River: Dargle, Taylor 2021 (PRE). Pietermaritzburg:
Canham 3S (NV); Randles 2', 34 (NU). Verulam: Umhloti Beach, Cra/iom 12 (NU).
Durban: near Durban, Drege s.n. (sub. nom. Momordica palmata E. Mey. b. in L,
iso.!), Rehmann 8843; 8844 (BR); Moss 5054 (J); F. Roberts in Herb. Moss 18437
(J); Pahl Hb. no. 25383 (BOL); Berea, Wood 8446 (NH); 8579 (BOL, NBG, PRE);
Isipingo Beach, Ward 289 (NU). Pinetown: “ Illovo Valley”, Wood 1863 (NH);
Amanzimtoti, Fisher 496 (NU); Winkle Spruit, Lansdell Hb. no. 34284 (NH).
Umzinto: Ifafa, Handley 63 (NU). Bergville: “Tiger Bush”, Baver & McClean
203 (PRE).
Cape Province. — Port St. Johns: Mogg s.n. (PRE); Pahl Hb. no. 25386 (BOL).
Umtata: Baziya, Baur 127 (BOL). Engcobo: Lewis Hb. no. 66631 (NBG, PRE).
Mqanduli: between Mqanduli and Coffee Bay, Lewis Hb. no. 63402 (NBG).
Kentani: Fegler 436 (PRE).
The identity of Cephalandra mackenii Naud. is quite clear. Mr. Killick compared
authentic material from the Paris Herbarium and reported that, apart from the degree
of lobing of the leaves, it is not distinct from the type of Coccinia palmata. The
distinguishing character given by Cogniaux (1881), viz., male flowers solitary in
C. mackenii and racemose in C. palmata is equally insufficient, because the flowers
are often solitary or racemose on one specimen. The distribution is as indicated above
by the cited specimens, i.e., mainly Natal and the Eastern Cape Province, extending
to the extreme southern part of Portuguese East Africa and to the Barberton district
of the Transvaal. All records from outside this area are, most probably, based on
wrong identifications. The distribution and field notes indicate that C. palmata inhabits
light forests, thickets and forest margins in lowland below 6,000 ft. in frost-free areas.
98
2. C. sessilifolia (Sond.) Cogn., Mon. Cucurb. 534 (1881); Dinter in Fedde, Repert.
16: 168 (1919-1920); Burtt Davy, FI. Transv. 1: 231 (1926).
Cephalandra sessilifolia Sond. in FI. Cap. 2: 493 (1862). Syntypes: Burke 289, Zeyher
580 (K).
Bryonial lagenaria E. Mey. ex Drege, Zw. Pflzgeogr. Doc. 54, 169 (1843), nomen
tantum.
Coccinia schinzii Cogn. in Bull. Herb. Boiss. 3: 419 (1895); Burtt Davy, l.c. (1926).
Type: Transvaal, Klippan, Rehmann 5162 (Z, holo.).
Perennial, glabrous, herbaceous climber. Rootstock tuberous. Stems slender,
branched, angular-sulcate, smooth, up to at least 5 m long. Leaves sessile to subam-
plexicaul, glaucous, herbaceous but firm, smooth or minutely punctate-scabrid, usually
without any glands, deeply palmately 5-lobed, 3-12 cm in diam.; the lobes oblong-
lanceolate, linear or oblong, usually acute and terminating in a mucro, coarsely dentate-
lobulate or trifid, sinuses between the lobes rounded, basal sinus very narrow (basal
lobes often overlapping at the opposite side of the stem), 1-4 cm deep. Tendrils simple.
Male plants: flowers racemose or by reduction solitary; peduncles rather slender,
terete, 1-6 cm long, up to 7-flowered; pedicels of racemose inflorescence articulate
below the calyx, up to 2 cm long; receptacle glabrous, 3-4 mm high and 4-6 mm in
diam.; sepals 1-3 mm long, 1-2 mm wide; corolla pale yellow to nearly white or
sometimes pale dull orange-buff, strongly veined, sublanate to glabrous or papillose
only. Female plants: pedicels stoutish, up to 1-5 cm long; staminodes ovoid-triangular
or oblong-triangular, white or pale cream, pubescent at the base and on the sides with
rather long stiff hairs, glabrous towards the obtuse apex, 2-3 mm long; style columnar,
glabrous, 7-8 mm long; ovary oblong-fusiform, glabrous, about 1-8 cm long and
5 mm in diam. Fruit oblong-fusiform or elongated ellipsoid, acute, red when ripe,
6-9 cm long and about 2-3 cm in diam. Seeds compressed, ovate, attenuate at the one
end, 6-9 mm long.
Type: Sonder, in the original diagnosis, cites three specimens, viz., Burke 289,
Zeyher 580 and the Drege gathering of this species in Herb. Hooker and Herb. Sonder
(S) without selecting a type. Burtt Davy (l.c.) cites the Burke and the Zeyher specimens
as the types and the Kew specimens ex herb. Hooker are to be taken as the syntypes.
Recorded from the following districts. — South West Africa: Outjo, Grootfontein,
Otjiwarongo, Omaruru, Okahandja, Windhoek, Gobabis, Rehoboth; Bechuanaland:
Mochudi, Kanye; Transvaal: Marico, Lichtenburg, Bloemhof, Klerksdorp, Potchef-
stroom, Brits, Johannesburg, Pretoria, Waterberg, Potgietersrus, Pietersburg, Lyden-
burg. Pilgrim’s Rest, Letaba, Soutpansberg; Orange Free State: Kroonstad, Boshoff,
Fauresmith; Cape Province: Vryburg, Barkly West, Hay, Kimberley, Herbert, Hope-
town, Colesberg and Graaff-Reinet.
Some noteworthy specimens are Burke 289 from Mooi River, Potchefstroom,
Transvaal (in PRE and SAM, one of the syntype numbers); Burke s.n. from “ Orange
River ” (in PRE, the corresponding sheet in K has been annotated by Sonder); Drege
s.n. (labelled “ Bryonial lagenaria E. Mey.”) from “ Nieuwe Hantam ” (most probably
Colesberg distr.. Cape) in L, PRE; Burchell 2661 from Hopetown (in L) cited by
Cogniaux; also cited by Burtt Davy as having the male flowers both in racemes and
solitary; Nelson 217 and Rogers Herb. no. 2377 (PRE), both also cited by Burtt Davy;
Rehmann 5161 (in BR) from “ Klippan ” was probably collected together with Rehmann
5162, the type of C. schinzii.
C. schinzii cannot be maintained, because the only difference is in the solitary or
racemose male flowers which difference does not hold. Burtt Davy already mentioned
a specimen Burchell 2661 (K) annotated by Cogniaux (as C. sessilifolia) which has
both solitary and racemose male flowers. Rehmann 5161 from “ Klippan” (in BR)
99
is certainly C. sessilifblia. In addition, specimens observed near the historical
“ Klippan ” (between Roedtan and Grass Valley in the Potgietersrus district) by the
present author, like most specimens, may produce both types of male inflorescences
on one branch. The same variation misled Cogniaux in other cases (see under C.
palmata = “ C. mackenii ”).
3. C. quinqueloba {Thunb.) Cogn., Mon. Cucurb. 533 (1881).
Bryonia quinqueloba Thunb., Prodr. PI. Cap. 13 (1794); FI. Cap. 35 (1807); Lindley,
Bot. Reg. t. 82 (1815); Curtis, Bot. Mag. 43; t. 1820 (1816); Ser. in DC., Prodr. 3;
307 (1828). Type; Cape, Thunberg (UPS, holo.).
Cephalandra quinqueloba (Thunb.) Schrad. apud Eckl. & Zeyh., Enum. PI. Cap. 280
(1834); Sond. in El. Cap. 2: 492 (1862); Naud. in Ann. Sci. Nat. 5me. ser. 5; 16
(1866).
Perennial herbaceous nearly glabrous climber. Stems slender, branched, angular-
sulcate or smooth, up to 10 m long. Leaves glaucous above, paler beneath, smooth
or minutely punctate-scabrid, sometimes minutely hispid, on the veins above, sub-
orbicular in outline, deeply palmately (3-) 5-sect, 4-12 cm long and 3-10 cm wide;
the lobes oblong, often somewhat wider near the usually rounded apex, entire or callose-
denticulate, less often more coarsely dentate or with a few lobules, mucronate, the
central one 2-7 cm long, 0-5-2 -5 cm wide, the lateral ones gradually shorter; sinuses
between the lobes rotundate; basal sinus very narrow (basal lobes of leaf sometimes
overlapping) up to 2-5 cm deep; petioles slender, striate, usually slightly hairy, 1-3
cm long, occasionally shorter. Tendrils simple, striate, glabrous. Male plant: flowers
solitary; pedicels subfiliform, 2-4 cm long or occasionally flowers in a few-flowered
raceme; receptacle glabrous, 4-5 mm high, 6-7 mm in diam. ; sepals triangular,
triangular-lanceolate to ovate, 2 mm long, 1-5-2 -5 mm wide near the base, patent
or erecto-patent; corolla about 2 cm long, shortly pubescent. Female plant: pedicels
stoutish, 1-2 cm long; staminodes oblong, densely lanate-shaggy, about 2 mm long;
ovary oblong-linear, glabrous or nearly so, smooth, attenuate at the apex, 2-2-5 cm
long, 3-4 mm in diam., style 8 mm long, glabrous. Fruit ellipsoid, bright red when
ripe, acute at the apex, more or less rounded at the base, 4-6 cm long and 2-5-4 cm
in diam. Seeds obovate, somewhat attenuate at the base, smooth, 6-7 X 4 X 1-2 mm.
Cape Province. — Humansdorp; Taylor 1289 (PRE). Somerset East: MacOwan s.n.
(GRA); Boschberg, Scott Elliot in Herb. Galpin no. 127 (PRE); Glen Avon, Brown
s.n. (PRE); Zuurberg, Barker 39491 (NBG); Zuurberg Range (= Somerset East or
Uitenhage), Holland 308 (GRA). “ Woods in Uitenhage, Albany and Kaffirland ”,
Ecklon & Zeyher 1796 (BOL, NH, SAM). Port Elizabeth: Van Staadens Mts., Paterson
996 (GRA, PRE); Marloth 1304 (PRE); Theron 570 (PRE); Kraggakamma, Zeyher
579 (BOL, SAM); West 300 9BOL); Enon. Thode A2666 (NH, PRE); Drege s.n.
Momordica quinqueloba E. Mey. a” in L, PRE, SAM); Long 1340 (PRE); Powrie
Hb. no. 25595 (BOL). Alexandria: Gledhill s.n. (GRA), Archibald 6085 (PRE);
Comins 1373; 1374; 1376 (PRE). Albany: near Grahamstown, Dyer 1379 (GRA,
PRE); Story 2346 (PRE); Howison’s Poort, Schbnland 664 (GRA); Blauwkrantz,
Leighton 3135 (BOL). Bathurst: Port Alfred, Galpin 3041 (GRA, PRE); Becker s.n.
(GRA); K.o'Nit, Lotsy & Goddijn 21% {V). Stockenstrom : Katberg, 5’cw//p 316 (NH).
East London: Gane 253 (PRE); Breyer Hb. no. 23225 (PRE); Munro Hb. no. 23161
(PRE); Smith2>61>\ \ 3813; 3861 (PRE); 16982 (BOL); Cambridge, fForwo/c/
128 (GRA). Komgha or Kentani (“ Kei Mouth”): Schlechter 6245 (GRA); Paid
Hb. no. 25387 (BOL). Komgha: Flanagan \1\2 (PRE). Queenstown: near Queens-
town, Galpin 8118 (PRE). King William’s Town: Pirie, Sim 20238 (PRE).
Natal. — Port Shepstone: Umtentweni, Blake s.n. (J).
100
The specimens from tropical Africa cited under this name by Hook. f. in FI. Trop.
Afr. 2, 551, and by Cogniaux (1881) are most probably not conspecific. The plant
hardly extends into southern Natal and the gap from this area to central Africa is
too large to make it probable that the same species is found in the tropical zones.
This species is obviously closely related to C. variifolia but differs in a few
apparently constant characters such as the absence of glands on the leaves which also
have broader, rounded or obtuse and rarely acute leaf-segments and a more ellipsoid,
shorter fruit. It is, in addition, geographically and ecologically separated from the
other species. It occurs in coastal bush or slightly more inland, whereas C. variifolia
is an inland form found in the more arid regions.
C. quinqueloba can usually easily be distinguished from the remaining S. African
species but occasionally specimens with narrow leaf-segments resemble certain specimens
of C. palmata with broad leaf-segments. The simple tendrils (bifid in C. palmata)
make a clear distinction possible.
Sonder (in FI. Cap.) already mentioned that C. quinqueloba can have solitary or
subumbellately racemose flowers. Cogniaux (1881), p. 529, on the other hand, keyed
it out in his group with solitary flowers. Cogniaux made the same mistake with
C. sessilifolia (of which he described a specimen with racemose flowers as a dilferent
species, although Sonder had already reported “ solitary or subumbellate flowers ”)
and with C. palmata (he recognised C. mackenii, described by Naudin from a specimen
of the latter with solitary flowers).
4. C. variifolia A. Meeuse, sp. nov., a foliis nunc angulatis vel leviter lobatis nunc
profunde palmatifidis facile distinguenda.
Perennis, alte scandens, subglabra. Radix tuberosus carnosus fusiformis vel plus
minusve napiformis ad 6 cm diam. Caules satis robusti, in siccitate longitudinaliter
sulcati, ad 5 m longi. Folia flrmiter herbacea vel subcoriacea ambitu late cordato-ovata
vel cordato-suborbicularia vel ovata, 5-angulata vel leviter ad profunde palmatisecta
3-11 cm longa et lata, basi cordata apice obtusa vel acuta minute apiculata, supra
viridia squamis albidis subsparse obtecta, infra pallidiora glauca pustulis minutis
subsparsis et prope basim glandulis paucis nigricantibus obtecta, margine leviter revoluta
subciliata, lobis 5 interdum 3 vel 7 triangularibus vel ovatis vel rhomboideis vel lineari-
oblongis subintegris vel remote leviterque sinuato-dentatis obtusis vel acutis mucrona-
tisque, angustioribus interdum crasse pinnatilobatis, petiolis satis robustis longitudi-
naliter sulcatis 8-35 mm longis. Cirrhi simplices. Flores masculi racemosi interdum
solitarii, pedunculis communibus ad 8 cm longis, pedicellis 2-4 cm longis interdum
(in floribus solitariis) ad 9 cm longis, receptaculo campanulato vel obconico striato
3-5 mm longo 4-6 mm lato, sepalis lineari-lanceolatis acutis 4-7 mm longis 0-5-1
mm latis, corolla pallide ochracea 18-22 mm longa lobis acutis mucronatisque extus
praecipue ad nervos sparse minute pubescentibus apicem versus ciliatis. Flores feminei
solitarii, pedicellis brevibus, ovario subcylindrato vel fusiform! glabro. Pepo oblongo-
ellipsoideus vel oblongo-fusiformis, immaturus viridis albo-variegatus demum coccineus,
5-8 cm longus, 2 -5-3 -5 cm diam. Semina 7-8 mm longa, 4-5-5 mm lata, 1-3 mm
crassa.
Transvaal. — Waterberg: Vaalwater, Meeuse & Strey 10413 (male plant, PRE, holo.!,
B, BM, BOL, BR, EA, K, L, M, P, SRGH, isos.!); 10413 bis (PRE, female plant);
“ Palala River”, Breyer TRV no. 25226; Rietspruit near Nylstroom, van Dam TRV
23372; Naboomspruit, Mosdene, Galpin s.n. Warmbaths: 11 Km from Warmbaths
on Nylstroom Road, Story 1525; near Warmbaths, Bolus 11893; Leendertz TRV
no. 7579; Aeocks 13903.
101
A tall perennial climber, nearly glabrous in all its parts except the very young
portions of the shoots and the corolla, producing several branched annual stems from
the apex of a fusiform to turnip-shaped fleshy perennial tuberous root. Stems rather
stout, firm, branched, longitudinally sulcate, up to 5 m long. Leaves varying in one
plant from undivided, pentagonal-cordate to shallowly or very deeply palmatilobed
with usually 5 broadly triangular to rhomboid or oblong to linear-oblong lobes, of
which the lowermost are the smallest and sometimes so small that the leaf is almost
3-lobed, or they have occasionally a basal lobule so that the leaf is apparently 7-lobed;
blade 3-11 cm long and broad, lobes subentire or more or less remotely calloso-denti-
culate (in shallowly lobed leaves) to coarsely pinnatilobed (in deeply dissected leaves)
obtuse or subacute with minutely apiculate or mucronate apex, with a slightly reflexed
subciliate margin, green and rather sparsely covered with small flat round scales above,
occasionally nearly smooth; paler, glaucous and sparsely punctate with minute pustules
below and in addition often with a few black adpressed glands between the main veins
near the base; quite glabrous or rarely hirtellous-scabrid on lower surface near base
on main veins; basal sinus usually broad and shallow or blade nearly truncate at the
base. Petioles longitudinally sulcate, in older leaves firm, often with a few longitudinal
lines of scabrid minute stiff hairs, 8-35 mm long. Tendrils simple. Male flowers
usually in subumbellate racemes with a solitary flower at the base of the inflorescence
in the same axil; common peduncle rather slender, up to 8 cm long; pedicels of solitary
flowers often equalling the peduncles, those of the raceme much shorter; calyx cam-
panulate or obconical, 3-5 mm long, 4-6 mm in diam. at the apex; sepals linear-
lanceolate, acute, 4-7 mm long, 0-5-1 mm broad, corolla pale buff-yellow, 18-22 mm
long, its lobes acute, mucronate, shortly and sparsely pubescent, ciliate mainly near the
apex. Female flowers solitar> on short pedicels; ovary cylindric to fusiform, glabrous.
Fruit ellipsoid-oblong or somewhat fusiform, green mottled with white when immature,
turning scarlet when ripe, 5-8 cm long and 2 -5-3 -5 cm in diam. Seeds numerous,
7-8 mm long, 4-5-5 mm broad and 1-3 mm thick.
This species was collected for the first time over 50 years ago but had always been
wrongly identified as C. palmata or C. quinqueloba. It differs from the first in the simple
tendrils, among other things, and from the second in several minor points: the
variable leaf-shape, longer inflorescence, larger fruit, paler glaucous lower leaf-surface,
etc. Mr. Killick kindly compared specimens with the material at Kew and gave as
his opinion that it is a distinct species, not matched at Kew except by some of the
cited gatherings (e.g. Bolus 11893). It has apparently a restricted area of distribution.
The striking variation in leafshape is a character not present to such an extent in any
other species, hence the epithet chosen.
5. C. hirtella Cogn. in Bull. Herb. Boiss. 4: 821 (1896); M. Wood, Handb. FI.
Natal 54 (1907); Phillips, Ann. S. Afr. Mus. 16: 101 (1917); Bews, FI. Natal and
Zululand 202 (1921). Type: Schlechter 6775 from Howick, Natal (Z holo.; BR,
GRA, isos.!).
Perennial, forming annual stems from a fusiform rootstock. Stems several to
many, climbing mainly over low bushes or occasionally prostrate, usually rather stout,
herbaceous, longitudinally sulcate, when young densely, later thinly covered with
somewhat bent or wavy, flattened, drab, fawn or sometimes whitish hairs, up to at
least 3 m long. Leaves herbaceous, soft, dark green, slightly paler beneath, suborbicular-
cordate to ovate-cordate in outline, palmately lobed or 5-7-sect to usually a little
beyond the middle, on both surfaces thinly covered with the same hairs as found on
the stems or a little more densely so on lower surface and always more densely so on
main veins on lower surface, 4-10 cm long and 3-9 cm broad ; basal sinus usually shallow
and wide; lobes ovate, obovate, elliptic or ovate-lanceolate, acute or shortly apiculate
to obtuse or rounded or sometimes attenuate into an acute point at the apex, usually
contracted at the base, somewhat irregularly denticulate to rather coarsely dentate
1517626-5
102
or serrate-dentate, not infrequently some of them or all pinnately lobulate; the middle
lobe slightly longer, the other ones gradually shorter towards the base ; sinuses between
the lobes usually narrow, rounded at their base; petioles covered with the same hairs
as the stems but much more densely so, somewhat flattened above, 1-4-5 cm long.
Tendrils simple or bifid, thinly covered with the same hairs as stems, leaves and petioles,
glabrescent, striate-sulcate at least in lower portion. Male plants: flowers solitary
or rarely geminate; pedicels long and slender, finely sulcate, up to 15 cm long; receptacle
broadly cup-shaped, 4-5 mm high, 6-8 mm in diam. at the mouth, thinly to more or
less densely hairy like pedicels and vegetative parts; sepals lanceolate-subulate, 4-7
mm long, hairy; corolla apricot yellow to pale buff with green venation on outside,
19-23 mm long; shortly hairy mainly on the veins outside; petals acute to acuminate.
Female plants: flowers solitary pedicels short, slightly thicker than in the male plants,
in fruit incrassate, strongly sulcate, glabrescent, up to about 4 cm long; ovary fusiform,
densely villous-lanate with the same type of hairs as found on the other parts, 2-2 • 5
cm long and 4-5 mm in diam. Fruit oblong-fusiform or ovoid-fusiform, more or less
rounded at the base and attenuate to apiculate into an acute apex, at first longitudinally
marked with green bands more or less anastomosing over intervening greyish white
or yellowish bands, when ripe bright orange-red, nearly glabrous, 7-8 cm long and
2-5-4 cm in diam. Seeds nearly smooth, 6-7 mm long and 3-4 mm wide.
Transvaal. — Wakkerstroom: Beeton 60 (SAM).
Natal. — Utrecht: Kafflrdrift, Thode A272 (NH, PRE); Tweekloof, Altemooi, Thode
A 187 (NH, PRE). Weenen: Wood 1001 (GRA, PRE). Estcourt: near Mooi River,
Wood 4105 (NH, BOL); Rietvlei, Greenwich Earm, Fry in Hb. Galpin no. 2773 (PRE);
Kamberg, Gordon Gray 58 (NH). Umvoti: Greytown, Wylie Hb. no. 20464 (NH).
Lion’s River: 8 miles S. of Nottingham Road, N. R. Smuts 1031 (PRE); Howick,
Schlechter 6775 (GRA, BR, isotypes). Impendhle: Huntley 412 (NU, PRE), 482a
(NU). Pietermaritzburg: S?nitli s.n. (PRE); Zwartkop, Rogers 1 139 (GRA). Pieter-
maritzburg or Camperdown: near Hilton Road, Dimock Brown 218 (NH). Camper-
down: between Inchanga and Drummond, Eshuis (PRE); Drummond, Galpin 10277
(PRE). Umzinto: Dumisa, Kenterton, Rudatis 2003 (NH).
Orange Free State. — Bethlehem: Bolus 8162 (BOL), Flanagan 1828 (PRE, SAM).
Ficksburg: Molenspruit, Galpin s.n. (BOL); Fouriesburg, Gemmell 6150 (PRE).
Basutoland. — Leribe, Dieterlen 145 (BR, NH, PRE, SAM); Memanieng Stream,
Jacot-Guillarmod 2204 (PRE); Mamathes, Jacot-Guillarmod 1969 (PRE).
Cape Province. — Wittebergen (Herschel, Lady Grey or Barkly East): Mrs. Barber
751 (GRA). Mount Currie: Tyson 1432 (BOL, NH, SAM); New Amalfi, Vielsalm,
Forbes 1124 (NH). Xalanga: Cala, Flanagan 2840 (PRE, SAM), Bolus 8910 (BOL).
6. C. rehmannii Cogn. in Bull. Herb. Boiss. 3: 418 (1895), aggregate species.
This species is, generally, an inland form, mainly found in dry sandy areas, but
some specimens of a slightly different habit and occurring in somewhat moister coastal
areas seem to constitute a variant of this typical form. They differ in a number of
minor points and are treated here as a variety of the typical C. rehmannii.
Stems usually soon glabrous, in older parts covered with greyish or silvery white adpressed scaly
specks; leaves firm; male flowers almost invariably solitary on pedicels which are rarely
over 2 cm long, rarely fasciculate to racemose with a very short common peduncle (a) var. rehmannii
Stems usually remaining hirsutulous much longer, in older parts not bearing whitish scaly specks;
leaves thinner in texture, drying membranous; male flowers often fasciculate on pedicels
usually 3-9 cm long, occasionally racemose, the common peduncle attaining a length of
8-5 cm (b) var. littoralis
(a) C. rehmannii Cogn. var. rehmannii.
C. rehmannii Cogn., l.c., Burtt Davy, El. Transv. 1: 231 (1926). Type: Rehmann
5168 from “ Klippan ”, Potgietersrust, Transvaal (Z, holo., BR!, K, isos., photos in
PRE). C. ovifera Dinter et Gilg ex Dinter in Vegetab. Veldkost D.S.W. Afr. 16 (1912),
nomen subnuidum, et in Fedde, Repert. 16: 168 (1919), nomen tantum.
103
Stems usually climbing, older ones perennial, several meters long, up to 2 cm in
diam., not forming a corky bark but retaining a greyish or greyish-green rather smooth
outer layer, covered mainly on the angles with whitish or greyish longitudinally arranged
elliptic or oblong scaly specks; younger stems firmly herbaceous, rather slender,
angular and sulcate, when quite young usually rather densely covered with a short
pubescence but as a rule soon glabrous. Leaves rather rigidly herbaceous, paler on
lower surface, pentagonal-suborbicular or pentagonal-cordate in outline, generally
deeply palmately 3 (-5)-lobed; rarely shallowly 5-lobed, on both surfaces hairy when
quite young but usually soon glabrescent and becoming scabrid with raised small
pustules, or on the upper surface with adpressed whitish scales; the main nerves
usually retaining some short setose hairs; the margin slightly thickened, somewhat
cartilaginous and scabrid, ciliate with short curved aculeate hairs; the lobes of the
lamina usually coarsely and jaggedly lobulate to coarsely pinnatisect with acute or
subacute narrowly oblong to linear, mucronate lobes and lobules, more rarely the lobes
broader, somewhat rhomboid or oblong-cuneate, with coarsely dentate to subentire
margins and sometimes rounded, often abruptly cuspidate-mucronate apex, very rarely
(in shallowly palmatilobed leaves) the lobes ovate-triangular to broadly triangular,
dentate, acute; the basal sinus usually deep and rather narrow; lamina usually 3-6
cm long and as wide, rarely attaining 11x11 cm; petioles somewhat dorso-laterally
flattened, angular-sulcate and bearing on the ridges (mainly on the lateral sides) rows
of stiff short bristly hairs, sometimes ultimately glabrescent, usually under 2 cm long but
occasionally attaining 4 cm. Tendrils simple, usually slender, at first somewhat hispi-
dulous mainly towards the somewhat incrassate base, but as a rule soon quite glabrous.
Male flowers solitary or occasionally fasciculate, rarely racemose on a short (up to
about 1 cm long) common peduncle; pedicels under 3 cm, but almost always under
2 cm long, slender to almost filiform, subterete or somewhat angular, somewhat hairy,
at the apex under the calyx with an articulation which appears somewhat peltate after
the flower has fallen off. Calyx hairy with curved multicellular, articular hairs, usually
rather densely so, rarely subglabrous; receptacle obconical to somewhat cup-shaped,
4-6 mm high and 6-9 mm in diam., sepals linear-lanceolate, tapering into a subulate
apex, 6-9 mm long and 0-5-1 mm wide. Corolla cream to pale yellow, often with
a buffy tinge, green-veined, hairy on the main veins, 22-28 mm long; the lobes somewhat
triangular and cuspidate-mucronate at their tips. Female flowers solitary; pedicels
usually under 1 cm but occasionally up to 2 cm long, somewhat sulcate, rather slender
but incrassate and up to 4 mm thick in fruit. Ovary fusiform-ellipsoid to broadly
ellipsoid, usually densely pubescent with curved articulated multicellular hairs but
occasionally glabrous, about 1 cm long. Sepals linear-lanceolate or somewhat oblan-
ceolate, 3-6 mm long. Corolla as in the male flower. Fruit subglobose or broadly
ellipsoid, sometimes contracted at the base just above the stalk, when young white-
and-green mottled, the white spots in rather vaguely defined longitudinal bands, when
ripe quite smooth and glabrous, turning scarlet, 3-4-5 cm long and 2-3-5 cm in diam.
Seeds a dirty white, oblong or somewhat falcate, rounded at the apex, usually somewhat '
contracted above the truncate to faintly bilobed base, 6-7 mm long, 2-5-3 mm broad
and about 1 • 5 mm thick.
This typical form is very wide-spread in South West Africa, Bechuanaland,
parts of Griqualand-West and the Transvaal and extends into Southern Rhodesia
and Northern Natal. The type specimen was collected at “ Klippan ”, which is on
a farm now called “ Doornpoort ”, near Grass Valley, in the Potgietersrust district.
The plant is very common in that area. It has been recorded from the following
districts.— 5om//2 West Africa: Ovamboland, Outjo, Okavango, Grootfontein, Otji-
warongo, Gobabis, Rehoboth, Warmbad; Bechuanaland: almost the whole area;
Transvaal: Marico, Warmbaths, Brits, Pretoria, Bronkhorstspruit, Groblersdal, Lyden-
burg. Pilgrims Rest, Nelspruit, Barberton and districts to the north of these; Cape
104
Province: Kimberley, Barkly West, Hay, Herbert, Kenhardt, Prieska, Little Nama-
qualand; Portuguese East Africa. Lourenco Marques; Southern Rhodesia: Melsetter,
Rupisi, Fort Victoria, Ndanga, West Nicholson, Bulalima-Mangwe.
The plants described as C. ovifera are not clearly separable from C. rehmannii
as was evident from a comparison of specimens named C. ovifera by Dinter himself;
the leaves are only slightly lobed and the fruits larger than in “ typical ” C. rehmannii,
but only cultivation of both forms from seed may decide the status of C. ovifera; for
the moment it is retained in C. rehmannii.
The juicy fruits and the tuberous roots of this plant are edible and are used for
food by Hottentots, Damaras, Bushmen and presumably other native tribes in South
West Africa and Bechuanaland.
This plant apparently prefers sandy soils in a fairly dry to dry but not very arid
climate and is resistant to light frosts. Its distribution clearly indicates a preference
for sandy regions without severe frosts. Especially in the Transvaal one can clearly
see that it does not occur in the real highveld districts but practically everywhere else
in suitable localities.
{b) C. rehmannii Cogn. var. littoralis A. Meeuse var. nov., a typo pedicellis masculis
saepius fasciculatis longioribusque foliis membranaceis differt. Type: Flanagan 457
from Komgha, Kei Mouth, in National Herbarium, Pretoria (male and female plant
on one sheet).
Stems shortly setose-hispid, ultimately glabrescent but apparently not becoming
covered with adpressed scaly specks with age. Leaves usually larger and more thinly
herbaceous than in the var. rehmannii, drying membranaceous; its lobes triangular
or oblong to ovate, sometimes suborbicular-cuneate, callose-dentate, sometimes subentire
or irregularly dentate to pinnatilobulate, but not so distinctly pinnatisect as is usually
the case in the var. rehmannii. Male flowers fasciculate or racemose, rarely solitary;
common peduncle of raceme up to 8-5 cm long; pedicels up to 9 cm long. Female
flowers: pedicels generally longer than in the var. rehmannii, 1 • 5-3 ■ 5 cm long. Otherwise
as the var. rehmannii.
Portuguese East Africa. — Distr. unknown (Sul do Save?): “ Mussaril e Cabeceira ”,
Rodriques de Carvalho s.n. (COI). Sul do Save: Lourenco Marques, Inhaca Island,
Breyer TRV 20506 (PRE); Mogg s.n. (PRE); Noel s.n. (PRE); Lourenco Marques,
Schlechter 11555 (BOL, GRA, COI); Hornby 4599 (PRE); Junod 20 (BR); Borle
253; 427; 442 (PRE); Katembe, Schlechter 11614 (GRA); Inhachingo, Exell,
Mendonca <& Wild 630 (SRGH); Massinga, Exell, Mendonca & Wild 645 (SRGH).
Natal. — Ingwavuma: Ndumu Game Reserve, IFo/'<7 3169; 3170; Out/cy C 15 (PRE).
Ubombo: Manaba Store, Gerstner 3407 (NH); Ubombo coastal veld. Tosh 28 (NU).
Hlabisa: 10 miles N.W. of Mtubatuba, Codd 9620 (PRE). Mahlabatini: Dhlebe,
Gerstner 4261 (PRE, NH). Umvoti: Thorns near Greytown, Wood 5318 (NH).
Mtunzini: Johnson 612 (NBG). Verulam: Umhlanga Rocks, Dohse & de Winter
223 (NH, PRE). Durban: near Durban, Jenkins TRV no. 7092 (PRE); Wood 6350
(BOL, L, NBG, NH, PRE); Stella Bush, Marriott Hb. no. 24341; 27143 (NH);
Berea, Small Hb. no. 34714 (NH). Pinetown: Doonside, Wylie Hb. no. 23299 (NH).
Cape Province. — Mqanduli: Coffee Bay, Tyson 24 (PRE, GRA). Komgha: Kei
Mouth, Flanagan 457 (PRE, holo.!, BOL, GRA, NGB, isos.!). East London: Nahoon,
Nanni 151 (PRE); Shelly Beach, Mogg 11941; 12070 (PRE).
The specimens from the dry thornveld of Northern Zululand are more or less
intermediate between this variety and the var. rehmannii (e.g., ILo/y/3169, 3170, Oatley
C 15, Codd 9620), but the forms from the wet coastal bush are more typical of the var.
littoralis (e.g., Tyson 24, Flanagan 457, Mogg 11941, 12070).
105
In a list of plants collected by Junod, Schinz in Mem. Herb. Boiss. 10: 69 (1900)
mentions “ Cocinia jatrophaefolia (A. Rich.) Cogn. var. australis Cogn.” (nomen
tantum!), citing Junod 20 and 463. A sheet of Junod 20 in BR is C. rehmannii var.
liltoralis and it has no relationship with C. jatrophaefolia which I regard as a synonym
of C. adoensis.
7. C. adoensis (Hochst. ex A. Rich.) Cogn., Mon. Cucurb. 538 (1881); Burtt Davy,
Transv. 1: 231 (1926).
Momordica adoensis Hochst. ex A. Rich., Tent. FI. Abyss. I: 293 (1847). Type:
Schimper 166 from Abyssinia (P, holo.; K, L, M, isos.).
Bryonia convolvuJoides A. Rich., l.c. B. jatrophaefolia A. Rich. op. cit. 289.
Cephalandra pubescens Sond. in FI. Cap. 2: 493 (1862); Hook. f. in FI. Trop Afr
2: 551 (1871).
Coccinia jatrophaefolia (A. Rich.) Cogn., l.c. C. parvifolia Cogn. in Vtjschr. Natiirf.
Ges. Zurich 52: 419 (1907); Burtt Davy l.c. C. pubescens (Sond.) Cogn. ex Harms
in Notizbl. Bot. Gart. Berlin 8: 491 (1923). C. roseiflora Suesseng. in Trans. Rhodesia
Sci. Assoc. 43: 60 (1951).
Tall climber. Stems usually slender, branched, longitudinally sulcate, when young
densely and shortly hispid-pubescent, more or less glabrescent, up to at least 6 m long.
Leaves varying from ovate-cordate and undivided to deeply digitately 3-7-lobed; deep
green drying green or dark brown above, pale greyish to glaucous green below,
herbaceous, when dry thin but rather firm, usually more or less shortly hirsutulous-
pubescent on both surfaces, usually only persistently so on the main veins (especially
on lower leaf-surface), rarely quite glabrous when old, but very often the leaf-bases
of the short stiff hairs persistent as fine punctations and leaf more or less scabridulous;
black glands near leaf-base on lower surface sometimes present; overall length of blade
4-12 (-16) cm, width 3-10 (-17) cm; basal sinus usually shallow or leaf truncate to
subcordate at base, rarely sinus narrow and deeper; margin finely and rather regularly
calloso-dentate to subentire; lobes of lobed leaves varying from oblong or triangular
to linear-lanceolate or oblanceolate, in more deeply dissected leaves almost invariably
contracted at the base and acute or acuminate, in less dissected (and apex of entire)
leaves sometimes obtuse to rounded, lobes usually entire, rarely lobulate; veins usually
distinctly prominent on lower surface; petioles longitudinally striate in dry specimens,
at first densely later more thinly covered with short subsetulose or hirsutulous hairs,
rarely becoming quite glabrous, 0-5-4 cm long. Tendrils simple, usually slender, at
first covered with the same type of short stiffish hairs as stems, leaves and petioles,
later glabrescent. Ma/c^ovrerv usually racemose ; common peduncle up to 12-fIowered,
usually slender, longitudinally sulcate in dried specimens, more or less hairy with
short stiffish hairs, up to 10 cm long; pedicels in raceme erect-patent, somewhat hairy,
articulated at the apex, up to about 20 mm long, those of solitary flowers (and solitary
flowers in same axil as the raceme) attaining about 7 cm; calyx glabrous or hairy
with short stiff hairs, receptacle broadly campanulate to semi-globose, narrowed at
the base, 5-7 mm high and 7-9 mm across at the mouth, sepals linear-subulate, usually
recurved, 2-4 mm long; corolla light buff or pale ochre-yellow to dull orange or a
salmon-yellow, 1-2 cm long, finely and sparsely papillose-pubescent, segments apiculate.
Female flowers: peduncles up to about 2 cm in fruit; ovary fusiform, glabrous or nearly
so; calyx and corolla shghtly smaller than in the male; receptacle very short. Fruit
ovoid-acute or elhpsoid to oblong, red when ripe, rounded at the base and usually
conical-acute at the apex, 4-5 • 5 cm long and 2-3 cm in diam. Seeds obovate in outline,
dull white, smooth, 4-6 mm long, 3-4-5 mm broad and 1-5-2 mm thick.
Type: The name “ Momordica adoensis Hochst.” appeared for the first time on
the herbarium labels of Schimper 166. The holotype is in P, isotypes in K (photo,
in PRE!), L! and M!
106
This plant has the widest distribution in Africa of all the South African species.
Specimens have been seen from Abyssinia, East Africa, Northern Rhodesia to the
extreme N.E. corner of South West Africa, Northern Bechuanaland and the Transvaal,
but apparently this species prefers semi-arid conditions and is absent from the wetter
areas in East Africa, Natal, etc., but is, e.g., found in the Estcourt area in the drier
thornveld of central Natal.
Recorded from the following districts. — South West Africa: Grootfontein; Bechu-
analand: as far south as Mochudi; Transvaal: Rustenburg, Krugersdorp, Johannes-
burg, Pretoria, Middelburg, Lydenburg, Pilgrims Rest, Belfast, Barberton and districts
to the north of these; Natal: Estcourt.
As regards the synonyms, the type of C. Jatrophaefolia {Quartin Dillon & Petit
s.n. in P) is indistinguishable from some forms of C. adoensis. Mr. Killick matched
several S. African specimens to this specimen.
Cephalandra pubescens, based on the gatherings Burke 408 (K, holo, photo. PRE!,
NBG, iso.!) and Zeyher 588 is indistinguishable from C. adoensis; the types are a good
match of the type of the latter.
Coccinia parvifolia Cogn. based on Junod 2491 (type in Z!) is a form with undissected
leaves. Such forms are linked through intermediates with the usual form with palmately
dissected leaves. C. roseiflora Suesseng. is based on a water-colour made by Mrs.
G. Dehn (no specimen!) in M, which I have seen. The plate is rather poor and the
identification from it almost impossible. The practice of describing nowadays species
from plates must be strongly deprecated. All the evidence points to C. adoensis (whose
flowers sometimes have a pink}- tinge) as its identity, the only alternative being
C. sessilifolia.
INDEX.
Page
Acanthosicyos Welw. ex Hook, f 52
horrida Welw. ex Hook. f. 53
Adenopus Benth 96
Atheranthera Mast 8
Blastania Kotschy et Peyr 11
cerasiformis {Stocks) A. Meeuse 12
fimbristipula Kotschy et Peyr 12
garcinii (h.) Cogn 11
luederitziana Cogn. ex Schinz 10
Bryonia africana L 32
africana Thunb. non L 30
angulata Thunb 31
convolvuloides A. Rich 105
cordata Thunb 19
digit at a Thunb 32
dissecta yhunh 32
fimhrist ipida ¥enz\ 12
grossulariifolia E. Mey. ex Drege 32
jatrophaefolia A. Rich 105
laciniosa L 33
/aev/s Thunb 20
lagenaria E. Mey. ex Drege 98
micropoda E. Mey. ex Drege 14
midtifida E. Mey. ex Drege 32
nana Lam 30
ohtusiloha E. Mey. ex Drege 27
pinnatifida Burch 32
punctata Thunb 19
quinquetoha Thunb 99
repanda BI 19
.scabra Thunb 19
.scabrella L.f. 14
trUoba Thunb 30
107
Page
Bryoniopsis laciniosa (L.) Naud 33
Cephalandra Schrad 95
mackenii 'Maud 97
palmata Sond 97
pubescens Sond 105
quinqueloba (Thunb.) Schrad. ex E. & Z 99
sessiUfoUa Sond 98
Ceratiosicyos Nees 20
ecklonii Nees 20
laevis {Thunb.) A. Meeuse 20
Citnillus Forsk 54
CitruUus Schrad 54
Schrad 57
battich Forsk 57
ecirrhosus Cogn 58
lanatus (Thunb.) Mans/. 57
naudinianus (Sond.) Hook. f. 55
vulgaris Schrad. ex E. & Z 57
Coccinia W. et A 95
adoensis (A. Rich.) Cogn 105
hirtella Cogn 101
jatrophaefolia (A. Rich.) Cogn 105
mackenii (Naud.) Cogn 97
ovifera Dinter et Gilg 102
palmata (Sond.) Cogn 96
parvifolia Cogn 105
pubescens (Sond.) Cogn. ex Harms 105
quinqueloba (Thunb.) Cogn 99
rehmannii Cogn 102
rehmannii Cogn. var. littoralis A. Meeuse 104
roseiflora Suesseng 105
schinzii Cogn 98
sessilifolia (Sond.) Cogn 98
variifolia A. Meeuse 100
Colocynthis Mill 54
citrullus (F.) O. Ktze 57
ecirrhosus (Cogn.) Chakrav 58
naudinianus (Sond.) O. Ktze 55
vulgaris Schrad 58
Coniandra Schrad 24
africana (L.) Sond 32
digitata (Thunb.) Sond 32
dissecta Schrad 32
glauca (Thunb.) Schrad 32
grossulariaefolia E. Mey. ex Arnott 32
777o//e (Kunze.) Sond 30
pinnatisecta Schrad 32
punctulata Sond 32
thunbergii Sond 30
zeyheri Schrad 31
Corallocarpus Welw. ex Benth. & Hook, f 38
bainesii (Hook, f.) A. Meeuse 41
bequaertii de Wild 39
dinteri Cogn 41
dissectus Cogn 38
gilgianus Cogn 39
glaucicaulis Dinter 39
scaber Dinter
schinzii Cogn 39
sphaerocarpus Cogn 41
welwitschii (Naud.) Hook. /. ex Welw 39
Ctenolepis Hook, f H
cerasiformis Hook, f
Ctenopsis Hook. f. ex Naud H
cerasiformis (Stocks) Naud 12
108
Page
Cucumis L 59
abyssinicus Schimp. ex Hook, f 80
africanus Z../. 72
africanus L.f. var. acutilobus Cogn 79
africanus L.f. var. myriocarpus (Naud.) Burtt Davy 74
africanus L.f. var. zeyheri (Sond.) Burtt Davy 79
angolensis Hook. f. ex Cogn 64
anguria L. var anguria 77
anguria L. var. longipes (Hook.f.) A. Meeuse 77
arenarius Schrad 79
asper Cogn 66
chrysocomus Schum. & Thonn 79
cogniauxianus Dinter ex Cogn 66
cordifolius E. Mey. ex Drege 47
dinteri Cogn 65
dissectifoliiis'Haud 74
ficifolius A. Rich 80
figareiOcl 81
figarei Hook, f 77
heptadactylus Naud 69
hirsutus Sond 63
homblei de.'NWd 64
hookeri 'Naad 72
humifructus Stent 62
kalahariensis A. Meeuse ■ 70
leptodermis Schweick 76
longipes Hook, f 77
maderaspatensis L 14
melo L 61
ineinbranifolius Hook, f 22
merxniuelleri Suesseng 74
metuliferus E. Mey. ex Schrad 68
myriocarpus Naud 74
naudinianus Sond 55
prophetarum Jacq 74
pueschelii Dinter : 63
pustidatus Hook, f 80
rigidus E. Mey. ex Naud 71
sativus L 60
seretii de Wild 64
seret hides Suesseng 64
sonderi Cogn 63
subsericeus Hook, f 22
umbrosus A. Meeuse & Strey 67
welwitschii Cogn 64
wildemanianus Cogn 64
zeyheri Sond 79
Cucurbita citruUus L 57
anguria Duch 57
lagenaria L 83
leucantha Duch 83
siceraria Mol 83
Cyrtoneina Schrad 24
molle Kimze 30
sphenoloba Schrad 31
triloba (Thunb.) Schrad 30
Dactyliandra Hook, f 10
luederitziana (Cogn.) Cogn 10
welwitschii Hook. f. 10
Eulenburgia mirabilis Pax 47
Eureiandra Hook, f 96
109
Page
Gerrardanthus Harv. ex Benth. & Hook, f 8
macrorhizus Harv. ex Benth. & Hook. f. 8
megarhiza Decne. & Harv 8
tomentosus Hook. f. 9
Gymnopeta/um sensu Baill 85
Heterosicyos Welw. ex Benth. et Hook, f 85
Hymenosicyos Chiov 21
(de Wild.) Harm.s 23
bryoniifoUus Merxm 23
meinbranifolius (Hook, f.) Chiov 23
subsericeus (Hook, f.) Harms 23
triangularis (Cogn.) Harms 23
villosiis (Cogn.) Harms 23
Kedrostis Medik 24
africana (L.) Cogn 32
angulata (Berg.) Fourcade 30
bainesii (Hook, f.) Cogn 41
capensis (Sond.) A. Meeuse 28
cineria Cogn 17
crassirostrata Breniek 35
digitata (Thunb.) Cogn 32
eminens Dinter et Gilg 36
foetidissima (Jacq.) Cogn 26
foetidissima {Jacq.) Cogn. var. divergens (Hochst.) Cogn 26
foetidissima (Jacq.) Cogn. var. genuina Cogn 26
foetidissima {Jacq.) Cogn. var microcarpa Cogn 27
foetidissima {Jacq.) Cogn. subsp. obtusiloba (Sond.) A. Meeuse 27
foetidissima (Jacq.) Cogn. var. perrottetiana (Ser.) Cogn 26
gilgiana Cogn 36
glauca (Schrad.) Cogn 32
longipedunculata Cogn 36
minuti flora Cogn 27
w/o/fo (Kunze ) Cogn 30
nana (Lam.) Cogn 29
nana (Lam.) Cogn. var. nana 30
nana (Lam.) Cogn. var schlechteri (Cogn) A. Meeuse 31
nana (Lam.) Cogn. var. zeyheri (Schrad.) A. Meeuse 31
natalensis (Hook, f.) A. Meeuse 36
obtusiloba (Sond.) Cogn 27
otaviensis Dinter 36
punctulata (Sond.) Cogn 32
rautanenii Cogn 36
rostrata (Rottl.) Cogn 26
schlechteri Cogn 31
velutina Cogn 30
zeyheri (Schrad.) Cogn 31
Lagenaria Ser 82
mascarena Naud 84
sagittata Harv 87
siceraria (Mol.) Standi 83
sphaerica E. Mey. ex Drege 84
sphaerocarpa E. Mey. ex Arnott 84
vulgaris Ser 83
Luffa caledonica Sond 94
sphaerica E. Mey. ex Sond 84
Melothria L 13
acutifolia Cogn
breyeri Burtt Davy
cineria (Cogn.) A. Meeuse 17
cordata {Thunb.) Cogn 19
foetida Lam 26
hederacea (Sond.) Cogn 30
hispidula Burtt Davy 17
maderaspatana (L.) Co^n 14
marlothii Cogn 15
membranifolia Cogn 19
minutiflora Cogn 1 ^
110
Page
natalensis Cogn 21
obtusHoba (Sond.) Cogn 27
parvifolia Cogn 17
pendula L 13
punctata (Thunb.) Cogn 19
tomentosa Cogn 19
tridactyla Hook. f. 21
velutina (Schrad.) Cogn 19
Momordica L 45
adoensis Hochst. ex A. Rich 105
angustisepala Harms 47
balsamina L 49
boivinii Bai\l 44
cardiospennoides Klotzsch 46
charantia L 51
clematidea Sond 46
cordata Cogn 47
cordifoUa Sond 47
ciiciillata Hook. f. 47
foetida Schun. & Thonn 47
gariepensis E. Mey ex Drege 49
buberi Tod 50
involucrata E. Mey. ex Sond. 50
lanata Thunb 57
luarlothii Harms 51
niorkoira A. Rich 47
pahnata E. Mey 96
repens Breinek 51
schimperiana Naud 47
schinzii Cogn 49
welwitschii //ook. /. 48
Miikia Arn 13
scabrella Arn 14
Oreosyce Hook, f 21
africana Hook, f 21
aspera Cogn 22
beguaertii de Wdd 22
parvifolia Cogn 23
subsericea (Hook, f.) A. Meeuse 22
triangularis Cogn 22
villosa Cogn 22
Peponia ITaud 91
caledonica (Sond.) Cogn 94
inackenii Naud 93
Peponiella O. Kotze 91
Peponium Engl 91
caledonicum (Sond.) Engl 94
chirindense (Bak. fS Cogn 95
kilimandscharicum (Cogn.) Engl 95
mackenii (Naud.) Engl 93
usambarense (Engl.) Engl 92
Physedra Hook, f 95
Piiogyne Schrad. ex E. & Z 13
garcinii Harv 90
suavis Schrad 20
Schrad 19
Piso.sperma Sond 24
capen.se Sond 28
Raphanistrocarpus (Baill.) Pax 42
asperifolius Cogn 44
boivinii (Baill.) Cogn 44
Kaphanocarpus Hook, f 42
asperifolius (Cogn.) Chiov 44
boivinii (Baill.) Chiov 44
bumilis Cogn 43
humilis Cogn. \ar prostratus Suesseng 43
Ill
kirk a Hook, f 42
tuberosus Dinter 44
welwitschii /. 43
Raphidiocystis chrysocoma {Schumach. & Thonn.) Killick & A. Meeiise 79
caillei Hutch, et Daiz 79
Rhynchocarpa Schrad ’ 24
hainesii Hook, f 41
foetida Schrad 26
welwitschii 'H&ud 39
Sky os angulata Berg 30
Sphaerosicyos Hook, f 80
meyeri Hook, f 84
sphaericus (E. Mey. ex Naud.) Cogn 84
Sphaerosicyus Post, et O. Ktze 80
Toxanthera Hook, f 24
kwebensis N. E. Br 36
lugardae N. E. Br 36
natalensis Hook, f 36
Trichosanthes foetidissima Jacq 26
Trochomeria Hook, f 85
debilis (Sond.) Hook. f. 89
hookeri Harv 90
macrocarpa (Sond.) Hook. f. 88
nudiflora Burtt Davy 88
pectinata (Sond.) Cogn 89
pectinata (Sond.) Cogn. var. subintegrifolia Cogn 90
rehmannii Cogn 91
rotundata Burtt Davy 90
sagittata (Harv.) Cogn 87
subintegrifolia (Cogn.) Burtt Davy 90
wyleyana (Sond.) Cogn 89
Typokedrostis Cogn 25
Zehneria Endl 13
cerasiformis Stocks 12
debilis Sond 89
garcinii Sond 90
hederacea Sond 30
macrocarpa Sond 88
obtusiloba Sond 27
pectinata Sond 90
scabra Sond 19
wyleana Sond 89
1
'i
II
!
,£
A
113
Notes and New Records of African Plants
GRAMINEAE
Andropogon lacunosus J. G. Anderson, sp. nov., A. pratensi Hochst. ex Hack, affinis,
sed flore superiore spicularum sessilium staminibus ad staminodia redactis, spiculis
pedicellatis 8-9 mm longis, gluma inferiore spicularum pedicellatarum 11-nervata in
medio nervis 3 in marginibus utrinque 4, glumis inferioribus spicularum sessilium et
pedicellatarum dorsaliter lacunis inconditis differt.
Gramen perenne caespitosum ad 60 cm altum rhizomate brevi. Culmi erecti. Folia
ad 20 cm longa 5 mm lata sparse ad dense pilosa pilis e tuberculis minutis ortis. Vaginae
pilosae vel villosae pilis e tuberculis minutis ortis; ligula membranacea ciliolata.
Racemi 2-3 spiciformes ad 7 cm longi. Spiculae sessiles staminibus ad staminodia
redactis. Gluma inferior circiter 8 mm longa 7-9 nervata canaliculata utrinque canalis
lacunis 5-8 inconditis dispositis. Gluma superior circiter 6 mm longa 3-nervata nervo
medio distincto et in setam scabridium 1-3 mm longam producto nervis obscuris.
Flos inferior 5 mm longum lemma redactus. Flos superior bisexualis, staminibus ad
staminodia redactis, palea nulla. Spiculae pedicellatae 8-9 mm longae. Gluma inferiore
1 1 -nervata in medio nervis 3 in marginibus utrinque 4 canalibus 2 et dorsalites lacunis
5-12 inconditis brevibus. Gluma superior circiter 8 mm longa 3-5 nervata. Flos
inferior lemma ad 7 mm longum redactus, palea nulla. Flos superior masculinus
lemmate circiter 4-5 mm longo palea nulla. Antherae ad 4 mm longae.
Transvaal. — Groblersdal: 28+ miles north-east by east of Groblersdal, Acocks
20908. Belfast: Dullstroom, vlei or stream above Suikerboskop, Hector s.n. Lydenburg:
15 miles west of Sabie, c. 6,800 ft., Codd 6441. Ermelo: Athole Pasture Research
Station, c. 5,800 ft., Norval 8, 122; Gericke A49 (PRE, holo.).
Perennial, caespitose with a short rhizome, up to 60 cm tall. Culms erect, slender,
glabrous, striate, often grooved, terete or the basal internodes slightly compressed,
2-4 noded; nodes glabrous. Leaves expanded, up to 20 cm long and 5 mm wide,
narrowed towards the base and tapering to an acute apex, broadest at about the middle,
7-nerved with the midnerve prominent abaxially, sparsely to densely pilose with bulbous-
based hairs on both surfaces but especially on the lower surface, margins scaberulous;
sheaths with a prominent midnerve, pilose to villous with bulbous-based hairs between
the nerves, often densely pilose near the ligule, basal sheaths often tinged with purple;
ligules membranous, ciliolate, truncate, up to 1 mm long. Inflorescence of 2-3 spike-like
racemes; rhachis up to 18 mm long, pubescent; racemes up to 7 cm long. Spikelets
light green or suffused with purple, paired, lower sessile, upper pedicelled; pedicels
linear, 3-4-5 mm long, margins pubescent. Sessile spikelets 6-8 mm long, linear.
Glumes unequal ; lower about 8 mm long, narrowly lanceolate, deeply grooved on the
back, the groove densely pilose in the lower third, and with 5-8 variably shaped,
shallow pits on both sides of the groove, chartaceous, 7-9 nerved, keeled with the keels
narrowly winged, ciliate in the upper half, apex obscurely bifid; upper about 6 mm
long, narrowly boat-shaped, thinly membranous, 3-nerved with the midnerve distinct
and produced into a 1-3 mm long scabrid bristle, lateral nerves obscure, margins
narrowly winged upwards, wings pilose along the margins, apex obtuse. Lower floret
reduced to the approximately 5 mm long lemma; lemma lanceolate, hyaline, nerveless,
margins inrolled and pilose in the upper half, apex obscurely bifid, palea absent. Upper
floret bisexual with the stamens reduced to staminodes; lemma about 5 mm long,
linear, nerveless, hyaline, lobed to about the middle, lobes acute, pilose along the
114
margins in the upper half, awned from between the lobes; awn slender, 10-15 mm long;
palea absent. Pedicelled spikelets 8-9 mm long, lanceolate. Glumes unequal; lower
8-9 mm long, lanceolate, chartaceous, keeled, keels and margins pilose, 1 1 -nerved with
the nerves grouped together, one group of 3 in the middle and 2 marginal groups
of 4 each, with 2 grooves and 5-12 variably shaped, shallow pits on the back, midnerve
excurrent into a very short bristle or apex acute; upper about 8 mm long, lanceolate,
membranous, 3-5-nerved, obscurely winged, pilose along the margins, apex acuminate.
Lower floret reduced to the 7 mm long lemma; lemma hyaline, obscurely 2-nerved,
pilose along the margins; palea absent. Upper floret male; lemma about 4-5 mm
long, hyaline, nerveless, pilose along the margins, apex emarginate; palea absent.
Anthers linear, up to 4 mm long.
Andropogon lacunosus is apparently a rare species which has so far been found only
in the north-eastern and eastern Transvaal where it occurs in vleis and near streams,
usually in sandy soils, at altitudes ranging between 4,700-6,800 ft. It is allied to
Andropogon pratensis Hochst. ex Hack., recorded from the Nile land and central east
Africa also at high altitudes (7,200-9,300 ft.). Mr de Winter, when comparing
A. lacunosus with other species of Andropogon at Kew found that it is unique in the
genus because of the shallow, variably shaped pits on the lower glumes of both the
sessile and pedicelled spikelets.
J. G. Anderson
HIPPOCRATEACEAE
Salada wardii Verdoorn, sp. nov., S. gerrardii Harv. ex Sprague affinis, sed habitu,
folds tenuioribus, gradatim acuminatis, costis lateralibus vix prominentibus, mar-
ginibus integris vel obscure denticulatis, floribus minoribus (alabastris ad 2 mm non
4 mm longis) differt.
Arbuscula vel suffrutex, nonnunquam scandens. Folia opposita vel subopposita
elliptica, basi cuneata, gradatim acuminata, 4-9 cm longa, 1-5-3 -6 cm lata; costae
laterales vix prominentes; margines integri vel obscure denticulati; petioli 3-7 mm
longi. T/orcj' pluri axillares fasciculati ; alabastra ad 2 mm longa; pedicelli graciles,
2-9 mm longi. Calyx 4-5 lobatus; lobi ad 1 mm longi, c 1-5 mm lati, interiores
latiores, nonnuquam minute ciliolatis. Petala 4-5-6, oblonga, ad 3 mm longa, 2-5 mm
lata, imbricata, deinde patentia, marginibus revolutis apicibus rotundatis minute cilio-
latis. Discus crassus, 1 mm altus. Stamina 3, c 1-5 mm longa; filamenta linearia
primus erecta deinde valde reflexa; antherae thecis subdivergentibus confluentibus.
Ovarium 3-loculatum, in disco partim immersum, apice in stylo breve attenuatum; ovula
c 2 pro loculo. Fructus subglobosus c 1 • 5 cm diam.
Natal. — Hlabisa: False Bay Park, Ward 3837 (PRE, holo.); 3855; False Bay area
Gerstner 4951 ; 5239; Ward5\\0. Ubombo: Mkuzi, Gwy 84; Masinga, T/n/ey 602.
Mr. Ward writes as follows about this species : “ A small, slender tree or shrub up to
approximately 9 ft. in height, often exhibiting a scandent habit; common in Sand
Forest, sometimes occurring as the dominant in the shrub layer ”. According to his
present observations he states further, that, where the forest is short and the soil very
sandy, this plant occurs as a small tree or shrub with a single main stem; but where the
forest is taller (25 to 55 ft.) the plant may develop a scandent habit and although common
does not exhibit quite the same dominance as in the shorter forest.
Among the South African species of Salacia, S. wardii is nearest to S. gerrardii Harv.
ex Sprague. In nature they are distinguished mainly by the habit and habitat. S.
gerrardii is a robust climber found in forests at an altitude of about 1,500 ft. in Natal and
the eastern Cape, bordering on Natal, while S. wardii, so far recorded only from Zulu-
land, is usually a small tree and occurs in Sand Forest at lower altitudes (300 ft. or less).
115
Differences that can be seen on herbarium specimens are that the leaves in S. wardii are of
a thinner texture, elliptic instead of oblong and gradually narrowing towards the apex;
the margins are entire or obscurely dentate as opposed to the obtusely serrate-dentate
margins of S. gerrardii; and the veins are not as conspicuous or prominent as those of
S. gerrardii. The species is named in honour of Mr. C. J. Ward, Ecologist, Natal Parks
Board. It is one of several novelties collected by him in northern Zululand.
Small tree or shrub, sometimes scandent. Leaves opposite or subopposite, more or
less elliptic, cuneate at the base, sometimes broadly so, and gradually narrowing to the
apex, 4-9 cm long, 1-5-3 -6 cm broad, veins scarcely prominent, margins entire or
obscurely dentate; petiole 3-7 mm long. Flowers several to many in axillary fascicles;
buds up to about 2 mm long; pedicels slender, 2-9 mm long. Ca/yx 4-5-lobed ; lobes
up to 1 mm long, about 1 -5 mm broad, the inner broader, sometimes minutely ciliate.
Petals 4 or 5, rarely 6, oblong, up to 3 mm long and 2-5 mm broad, imbricate, ultimately
spreading with the margins revolute, apex rounded, sometimes minutely ciliate. Disc
thick about 1 mm high. Stamens 3, about 1 -5 mm long; filaments linear, erect at first,
becoming strongly reflexed; anthers with the cells sub-divergent, confluent. Ovary 5-
celled, partly sunken in the disc, attenuated at the apex into a short style; ovules about 2
to a cell. Fruit subglobose, about 1 -5 cm diam., reddish brown when ripe.
I. C. Verdoorn.
IRIDACEAE
Watsonia latifolia Oberm. sp. nov., W. merianae Mill, affinis sed folds latioribus floribus
maioribus differt.
Cormus crassus fibrosus. Folia lanceolata vel lineari-lanceolata ad 70 cm longa ca
4-8 cm. lata margine et nervis nec incrassatis. Spica simplex multiflora, spathae exteriore
2 cm longae amplexicaulae; perianthium rubrum curvatum tubo basi angusto superiore
ampliore.
Natal. — Ngotshe district: near Ngome at the foot of bare granite slopes, in pockets
of loose soil, in clumps, Codd & Robertson 6403 (PRE, holo., K, SRGH, iso.). Utrecht
district: Donkerhoek, hillsides amongst rocks, sometimes seen in large colonies, Devenish
545 (PRE).
Swaziland. — Mbabane, Nicholson (PRE, 17691).
Transvaal. — Carolina, m. S.E. of The Brook, in rock crevices, Codd 9497 (PRE).
Govt. Forestry Station near Vossmann’s Beacon, Mentzel (PRE).
Plants 70-100 cm tall, gregarious. Conn thick about 6 cm in diam. covered with
coarse fibres. Leaves 5-7 distichous, linear-lanceolate to lanceolate, 30-70 cm long,
4 cm wide or more, apex acute, base narrowed, sheathing, lamina flat, thin, fairly soft in
texture, yellow-green, nerves and margin yellow, thin, not raised. Spike about 35 cm
tall, simple, overtopping the leaves, many-flowered; the flowers situated fairly far apart;
scape bracteate, sturdy; outer spathe valve ovate, about 2 cm long, apex acute, base
amplexicaul; 2 inner lateral valves slightly smaller and narrower. Perianth red, curved,
6-8 cm long; tube with a basal erect narrow part about 2 cm long and a widened and
arched part above it (where it is exserted from the spathe valves) about 2-5 cm long and
8 mm wide; lobes ascending or semi-spreading, lanceolate, about 2 cm long, apex
apiculate; stamens situated dorsally with the filaments attached to the tube where it
becomes wider, decurrent; anthers 8 mm long, attached near the base; the locules
diverging below, reaching to the tips of the perianth lobes above; ovary with 8 biseriate
ovules situated in the lower half of the locule; style filiform, stigmas bifid, fimbriate,
falcate, overtopping the anthers when mature. Capsule and seeds not seen.
116
This handsome Watsonia was first given the name W. latifolia by N. E. Brown,
but he did not publish it. The species is apparently restricted to the south eastern part
of the Transvaal and the neighbouring districts of Natal and Swaziland. No other
northern Watsonia has such broad leaves. Both Professor R. H. Compton and Mr.
N. J. Devenish have found it a handsome garden plant and easy to cultivate.
Watsonia watsonioides (Bak.) Oberm. comb, nov., Tritonia watsonioides Bak., Handb.
Irid. 195 (1892); FI. Cap. 6: 127 (1896). Type: Transvaal, Barberton, Saddleback Mt.,
Galpin 814 (K, syn., PRE, iso.!); Mts. of Swaziland, Thorncroft {Wood 4113: K, syn.,
NH, iso.!).
Watsonia flavida Bol., nom. nov. in Trans. Roy. Soc.S .A. 1: 162 (1909); FI. PI. Afr.
Plate 296 (1928)
When Bolus transferred this species from Tritonia to Watsonia, he gave it the more
appropriate but unnecessary name, W. flavida and it cannot be accepted according to the
international rules of nomenclature. Although the flowers are usually cream to yellow
in colour. Professor R. H. Compton observed one area of several square miles in the
Mankaiana district in Swaziland (alt. 3500^000 ft.) where a form with exclusively dark
maroon flowers occurred. It is apparently confined to the south eastern Transvaal
and Swaziland where it is very common and forms large clumps.
A. Amelia Obermeyer.
117
The South African Species of Dipcadi
by
A. A. Obermeyer
In the Flora Capensis 6:445 (1897) Baker enumerated 14 species for South
Africa. Since then over 50 more “ new ” species have been described for southern
Africa, including a large number from South West Africa. As some are known to be
poisonous and others are eaten by Bushmen and wild animals, it is essential to bring
order into the classification of this genus. It has been necessary to reduce to synonymy
a large number of names. Many of the unnecessary “ new ” species were based on
variable characters and others resulted because some species flower hysteranthously
in spring and later synanthously. It seems also that several hybrids were given specific
rank.
Baker divided the species into two sections; those with the perianth segments
of equal length were classified in the §Tricharis and those with caudate appendages
to the outer segments in the ^Uropetalum. When the appendages are well developed
sectional classification is easy but in some specimens the appendages are very short.
Bentham & Hooker, Gen. PI. Ill, 2: 809 (1883) pointed out that the length of the
appendages varied even in flowers on one raceme and Baker, when describing the
Madagascar sjaecies, D. heterocuspe, also mentions that short and long appendages
occurred on one raceme. As the appendages are formed at an early stage in the bud,
even the short ones may be seen to protrude as three short apical teeth. Apparently
the species with appendages to the outer perianth-segments are restricted to Africa
and Madagscar. Those from Europe and India do not have them. However, the
sections are not used here.
An example of a variable species is D. viride (L.) Moench, the first species to be
described for South Africa, a widespread almost weed-like species extending into
tropical Africa. It appears to be extremely adaptable ard shows great variation in
size, also in the length of the outer appendages which can vary from 2 mm to 3 cm.
Flowers from near the base of the raceme usually have shorter appendages than those
from the apex.
Examples of species flowering hysteranthously are D. ciliare (Zeyh. ex Harv.)
Bak. and D. marlothii Engl, from the drier areas. Later in the season they may flower
synanthously and intermittently. The early spring forms were usually wrongly identified
as D. hrevifolium from the Cape.
Of interest too is the dimorphism of D. gracillimum from the borders of the
Kalahari and the Transvaal Highveld. In early spring, after a dry winter, the reserves
stored in the bulb produce a few-flowered raceme and a few thin leaves which, if water
is insufficient, wither from the tip downwards. The capsules are small. The same
plant looks very different after good rains. The numerous succulent leaves are then
flabby and bend over or lie flat on the ground. Several many-flowered inflorescences
may be produced in which the flowers are bigger and the transformation of the capsule
is especially striking. It may be at least six times the size of the capsule of the depauperate
form. Around Kimberley the older farmers call it “ Oupa Groot Toon ” alluding
to its size. Large and small capsules were also observed in D. hrevifolium Fourc. and
in D. longifolium Bak. Hooker, FI. Brit. India 6: 346 (1892), describes the capsule of
D. serotinum as “very variable, 1/3-2/3 inch in diam.”.
118
A study of D. bakerianum Bol. and D. glaucum Bak. and several related “ species ”
indicates that they hybridize in nature.
The first species described from the Cape was classified as Hyacinthus viridis by
Linnaeus. It does not grow on the Cape Peninsula as one might expect but is found
further east. Thunberg later collected the solitary species of the Peninsula and named
it Hyacinthus brevifolius. Subsequently these two species were transferred several
times to different genera before Baker in 1871 placed them both in Dipcadi but not
under the presently accepted names. The genus Dipcadi was described by Medicus
in 1790 for a species from the Mediterranean region, Hyacinthus serotinus Linnaeus.
Botanists working on the Cape flora overlooked this northern genus and placed the
Cape species variously in Hyacinthus, Lachenalia, Zuccagnia, Polemannia, Uropetalum
and Tricharis. With the exception of Hyacinthus and Lachenalia, these genera have
become synonyms of Dipcadi.
Economic Importance
D. glaucum Bak., “ Dronk ui”; the Bushman’s “ igwashe ”. This species has
been known to be poisonous to cattle and sheep since 1867, often with fatal results,
but the active principle has not been isolated. Steyn in “ The Toxicology of Plants
in South Africa ” 1934, found that it was less poisonous in the flowering stage than
at a later stage. It is widespread in the north-western parts of the Cape Province,
South West Africa and the Transvaal and rare in the Federation where it occurs as
a form with shorter pedicels. (The Central Research Station at Mazabuka reported
on a specimen that it was not toxic to stock.)
D. crispum Bak. Marloth recorded on a specimen collected at Kalkfontein in
Namaqualand by Meyer {Marloth 6430) that it was “ very poisonous ” ; called
“ gifbol ” in the Beaufort West district.
D. platyphyllum Bak. (“ Horinkies ”). Some farmers in the Kimberley district
claim that it is poisonous.
D. marlothii Engl. (syn. D. polyphyUum Bak.) and D. viride (L.) Moench (syn-
D. umbonatum Bak.), are Sotho remedies for gonorrhoea and pimples according to
Watt & van Brandwyk in “ The Medicinal and Poisonous Plants in South Africa ”
(1932).
D. dinteri Bak. It is recorded by Dinter on the herbarium specimen that the
Hottentots eat the bulbs.
D. gracUHmum Bak. On a specimen from Prieska, Bryant (J 188) wrote the
following: “Yeiy common, bulbs said to be edible but have an insipid taste”.
D. longifoliuin (Lindl.) Bak. and D. marlothii Engl. The bulbs are eaten by
Bushmen according to Story and Schoenfelder, either fresh or roasted in ash.
In India the bulbs of this genus are non-poisonous and are eaten in times of
scarcity.
Dipcadi Medicus in Acta Palat. 6: Phys. 431, ex eod. Ust. Ann. ii: 13 (1790);
Baker in Journ. Linn. Soc. 11: 401 (1871); Benth. & Hook, f., Gen. Plant, hi: 809
(1883); Bak. in FI. Cap. 6: 445 (1897); FI. Trop. Afr. 7: 516 (1898); Phillips, Gen.
S.A. FI. PI. ed 2: 190 (1951).
Zuccagnia Thunb., Nov. Gen. Plant. 9: 127 (1798), in error “ Zuccangia ”; in Roem.
Archiv. ii: 2 (1799); FI. Cap. ed. Schult. 328 (1823).
Uropetalon Burch, ex Ker in Bot. Reg. t. 156 (1816). The name was spelt Uropetalum
by subsequent authors.
119
Polemamia Berg, ex SchltdI. in Linnaea 1: 250 (1826), not of Ecklon whose name
is rejected, cf. Taxon 9: 74 (1960).
Tricharis Salisb., Gen. PI. Fragm. 24 (1866).
The name. Dipcadi is said by Adamson & Salter in FI. Cap.' Pen. (1950) to be an
oriental name for the grape hyacinth.
Roots many, thin. Bulb small, ovoid, with thin outer tunics, sometimes produced
into a neck; occasionally forming runners or bulbils. Leaves one, several or many
per flowering shoot, their number more or less constant for each species; lamina
narrow to wide, flat or folded, straight or spirally twisted, glabrous or hairy dorsally;
margin smooth, rarely undulate, ciliate or papillate, occasionally prominent, base
free or clasping, apex acute or acuminate; often the innermost leaf much smaller and
narrower, hysteranthous and/or synanthous. Inflorescence a simple, central raceme,
often secund, with a few to many flowers; scape terete, glabrous or slightly ribbed and
papillate, rachis drooping in bud, becoming erect during anthesis; bracts small,
caducous or persistent; pedicels very short to long, usually erect in bud, recurved
during anthesis, becoming erect again and lengthening slightly in fruit. Flowers green
or brown, sometimes with some yellow, red, brown or white near the tips on the inner
surface and along the margin of the perianth lobes; appendages occasionally reddish;
emitting scent at night and therefore probably pollinated by moths. Perianth deciduous,
at first closing over the ovary when fading; segments forming a tube, fused at the base
or rarely 2/3 of the way, connivent above; the 3 outer segments spreading outwards
from near the middle or erect with only the tips recurved, apex cucullate with a group
of papillae on the inside; tips sometimes extended into a short to long, caudate, terete
appendage; 3 inner segments similar to the outer (but never with caudate appendages),
often connivent to form a tube inside the outer segments with only the tips curved
outwards. Stamens 6, included, uniform, filaments flat, flimsy, fused to the tube below,
appressed to the connivent segments above it; anthers with the apex just reaching
the mouth of the tube, introrse, bilocular, attached dorsally near their base; locules
long and narrow, often slightly diverging and obtuse or sagittate at the base, the
connective produced into a short, apical appendage in one species {D. glaucuni).
Gynoecium with a turbinate, sessile or stipitate ovary, rarely with 3 apical knobs; with
about 20 biseriate ovules per loculus, style short or as long as the ovary, papillate,
hollow (often mucilaginous inside), shorter than the stamens; stigma of 3 lingulate,
papillate lobes* or rarely capitate and sub-persistent {D. marlothii, D. papillatum).
Capsule trigonous in cross-section or rounded or oblong with the side-walls in this
case at first adnate near the centre, splitting later (i.e. D. marlothii); in one species
(D. bakerianum) with an apical knob on each valve; shortly stipitate in some species
{D. longifolium and some Indian species). Seeds flat, round, black, shiny, papillose.
Type Species: D. serotinuni (L.) Medic.
Chromosome numbers
D. serotinum (L.) Medic, 2n == 8, according to Levan (1943).
D. viride (L.) Moench, 2n = 12, according to de Wet [Cytologia, 22: 145-149 (1957)].
D. glaucum Bak., 2n = 18 according to de Wet, l.c. and La Cour (Darlington & Janaki
Ammal, 1945).
Distribution. — Mediterranean region, Africa, Madagascar Socotra and India.
About 100 spesific epithets have been used but only about 30 may eventually be upheld
From southern Africa 13 species are recognised here.
* In dried specimens the lobes are stuck together. Much pollen from the anthers falls on the
stigmas. Self pollination may take place but we do not know if the flowers are protogynous or not.
120
From collectors’ records it is apparent that these relatively unattractive plants
have been collected in South Africa mainly by botanists who have concentrated on
an area, as for instance Fauresmith, Kimberley and Pretoria. The species are probably
more widely spread than our present records show them to be. The large number
of species and the extensive pure stands recorded from around Kimberley are note-
worthy. Here 11 of the 13 species have been recorded. Most of these species extend
to the western side of the Kalahari sandveld in South West Africa and further northwards
to Southern RJiodesia and north-eastwards to the Transvaal. D. gracillimum, D.
ciliare and D. marlothii also penetrate further east and south. Absent from the
Kimberley area are D. brevifolium Bak., a winter rainfall species extending to the
Namib in South West Africa, and D. longifolium Bak., a subtropical, hygrophytic
species.
1. Capsule knobbed; ovary with 3 distinct apical bulges; leaves many, rosulate, lanceolate,
long-tapered in upper half; outer and inner segments equal 2. bakerianiim
Capsule not knobbed; ovary oblong to obovoid, sometimes obscurely umbonate but then
segments unequal 2
2. Style short, stigma capitate, sub-persistent 3
Style and stigmatic lobes (erect and closed in dried flowers) forming a pencil-shaped organ
about as long as the ovary 4
3. Leaves 4-6, linear-lanceolate, acuminate, glaucous, glabrous, margin yellow, prominent,
papillate; scape often minutely ribbed and papillate 3. papiUatum
Leaves many, grass-like, the outer with short white, straggling hairs on the margin and dorsally
at the base (these sometimes absent later in the season); scape smooth 5. marlothii
4. Anthers with a short apical appendage; large plants with many rosulate, glaucous leaves
up to 7 cm wide; pedicels 2-7 cm long 9. glaucum
Anthers without an apical appendage; plants usually smaller and with narrower leaves;
pedicels short 5
5. Leaf 1 (occasionally 2 when the bulb produces 2 flowering racemes simultaneously), synan-
thous, clasping the scape for about half its length below the surface, with the lamina flat on
the ground at right angles to the scape 6
Leaves more than 1, or, if hysteranthous and at a certain stage perhaps showing a small
immature leaf or bract, then not as above 7
6. Lamina folded, linear, margin smooth; caudate appendages of outer segments long 12. vaginatum
Lamina flat, lanceolate, margin crisped-ciliate; caudate appendages of outer segments short
13. platyphyllum
1. Ovary and capsule shortly stipitate; segments fused only near the base, spreading, obtuse,
not caudate; subtropical, glabrous, usually hygrophytic plants 6. longifolium
Ovary and capsule sessile (if occasionally shortly stipitate in D. ciliare then leaves ciliate) 8
8. Outer perianth segments with short to long caudate appendages;* leaves straight or spirally
twisted 9
Outer perianth segments not caudate, more or less equal in length to inner; leaves with a
lax spiral twist in upper half when young II
9. Leaves crisped-undulate, setaceous to glabrescent; flowers campanulate, orange to brown
7. crispum
Leaves glabrous, smooth; flowers tubular, green, occasionally with red, brown or orange
margins 10
10. Leaves strap-shaped, coriaceous, apex obtuse, apiculate, margin prominent, often red
1 1 . rigidifolium
Leaves linear to lanceolate, soft, long tapered towards the apex, margin not prominent 10. viride
11. Leaves 2-3, filiform to linear, glabrous; flowers 12-18 mm; south-western Cape, Namib
1. brevifolium
Leaves many (sometimes few in early spring) 12
12. Flowers about 1 cm long; leaves (at least after rains) many, not rosulate, filiform, equalling
more or less inflorescences in length, glabrous, margin entire 4. gracillimum
Flowers 2-2 • 8 cm long; leaves many, rosulate, linear, flat, long-attenuated to the apex (usually
hysteranthous in spring, later synanthous), shorter than the inflorescence, margin crisped
and ciliate 2. ciliare
1. D. brevifolium (T/iunb.) Fourc., Bot. Surv. Mem. S.A. 20: 99 (1941).
Hyacinthus brevifolius Thunb., Prodr. 63 (1794). Type: Cape, “ Carro et Cannaland,
Oct.-Nov.” Thunberg (UPS, holo., PRE, photo).
* In cases where the outer appendages are short the buds will show them as 3 short apical teeth.
121
D. hyacinthoides (Spreng.) Bak. in J. Linn. Soc. 11: 398 (1871); FI. Cap. 6: 446
(1897). D. spirale Bak. in Engl. Bot. Jahrb. 15: 7 (1892); FI. Cap. 6: 447 (1897).
Type: Cape, Calvinia, Hantam Mountains, Meyer (B, holo., PRE, photo). As “ D.
hyacinthiflora (Berg, ex Schltdl.) Hutch.”, A Botanist in S.A. 200 (1946); Adams.
& Salter, FI. Cape Penin. 197 (1950).
Polemamia hyacinthiflora Berg, ex Schltdl. in Linnaea 1: 250 (1826). Type: Cape,
Camps Bay, Bergius (B, holo., probably destroyed).
Uropetalum hyacinthoides (Berg, ex Schltdl.) Spreng., Syst. 4, 2: 135 t'1827); Cur.
Post. 16: 135 (1827); Roem. & Schult., Syst. Veg. 7: 618 (1829); Kunth, Enum.
4: 378 0843).
Scilla brevifolia t,Thunb.) Roem. & Schult., Syst. Veg. 7: 574 (1829).
Periboeal brevifolia (Thunb.) Kunth, Enum. 4: 294 (1843).
Plants 20-40 cm tall, glabrous. Bulb 15-30 mm in diam. with firm grey outer
tunics sometimes produced into a short neck. Leaves hysteranthous or synanthous,
2-4, filiform to linear or terete and deeply channelled, 6-50 cm long, 1-5 mm wide,
glabrous, sinuous, apex long attenuated, often dying back, frequently forming lax
spirals when young, base free, rarely vaginate. Raceme variable in length, 10-40 cm
long, secund, 8-20-flowered ; scape terete, firm; bracts deltoid, acuminate, persistent;
pedicels about 2 mm long during anthesis, up to 10 mm in fruit. Flowers 12-20 mm
long, in shades varying from brown to green, greenish-yellow to cream, the darker
colours being found on the outside and the lighter inside; perianth-segments fused in
lower half, outer narrower and slightly longer than inner, recurved from the middle;
inner forming a connivent tube with only the tips spreading. Ovary with about 13
ovules in each cell; style about as long as the ovary with a 3-lobed stigma. Capsule
rounded in outline, trigonous, 15 mm long, 18 mm broad; seeds 8 mm in diam.
Distribution. — South-western Cape, eastwards as far as the Middelburg and
George districts; South West Africa, Namib; rocky habitats.
Cape. — Peninsula: Cape Town, Signal Hill, Marloth 1592; Lions Head, Schlechter
1284 (BOL). Worcester: Hex River Mts., Buffelshoek Twins, summit, Esterhuysen
8499 (BOL). Caledon: Appelskraal, Rivier Sonder Einde Mts., Stokoe (SAM 62219,
NBG). Uniondale: Georgida, Esterhuysen 6370. George: Kleinswart River, Gillett
1190 (BOL). Middelburg: Bangor farm, on lower koppies, common. Bolus (BOL
14044). Sutherland: Roggeveld, Farm Fransplaas, Marloth 9833. Ceres: near
Karroopoort, Barker (BOL 26944). Clanwilliam: Cedarberge, shale land below
Tafelberg, Esterhuysen (BOL 18138); near Citrusdal, Stokoe (SAM 63660, NBG).
Vanrhynsdorp: Klaver, Lewis 1457, 2964, Compton 17430 (NBG). Namaqualand:
near Steinkopf, Acocks 19535. Richtersveld, Arrisdrift, on stony flats, Marloth 12303,
12646.
South West Africa. — Luderitz: stony, granite flats 15 miles west of Aus, Giess &
van Vuuren 847; Halenberg, Dinter 6610.
The combination D. hyacinthiflora (Berg.), which Hutchinson made in his book,
A Botanist in Southern Africa, must be considered invalid as he did not mention the
basionym (Polemannia hyacinthiflora Berg, ex Schltdl.). Sprengel arbitrarily changed
the epithet to hyacinthoides.
This is the only species from the Cape Peninsula. The specimens cited under this
species in the Flora Capensis from the Transvaal, Natal and eastern Cape do not belong
here. They are hysteranthous forms of D. ciliare or D. marlothii.
It flowers from September-November; “ faintly scented ” {Foley).
2. D. ciliare (Zeyher ex Harv.) Bak. in Journ. Linn. Soc. 11: 398 (1871); FI.
Cap. 6: 447 (1897). Benth & Hook., Gen. Plant. Ill, 2: 809 (1883).
122
Uropetahim ciliare Zeyher ex Harv., Thes. Cap. II; 45, t. 170 (1863). Type: Cape,
Uitenhage, fields near Zwartkops River, Ecklon & Zeyher 48 (K, lecto, TCD, NBG,
iso. , PRE, photo); Cradock, Cooper 493 (K).
Dipcadi setosum var readii Bak. in Journ. Linn. Soc. 11; 398 (1871). Type: Cape,
Fort Beaufort dist.. Fuller’s Farm, Reade 94, pro parte, excluding left hand specimen.
(TCD, holo. , PRE, photo.). D. readii Bak. in FI. Cap. 6: 447 (1897). D. rytegalanlhiim
Zahlbr. in Ann. Hofmus. Wien 15: 21 (1902), ex descr. Type: Cape, Albany Dist.,
Grahamstown, Penther 511 (W, holo., destroyed). D. adelae Beauv. in Bull. Soc.
Bot. Geneva, ser. II, 5: 326 (1914). Type: Cape, Matatiele district, near Mafube,
Adele Berthoud (G, holo. , PRE, photo.).
Small plants up to about 40 cm high. Bulb ovoid to globose, about 2 cm in diam.
with thin, dark, outer tunics produced into a short neck. Leaves about 6 per shoot,
hysteranthous at first, later synanthous; outer leaves usually twisted spirally, narrowed
in upper half to a long, acuminate apex, broadest below and there laxly vaginate, 5-18
cm long, up to 8 mm broad below, margin usually undulate, often prominent, yellow,
ciliate with long scattered hairs, lamina dorsal ly at the base sparsely to densely covered
with bristles on the prominent nerves; inner leaves 1-2, linear to filiform, glabrescent.
Raceme 20-30 cm high, secund, laxly 4-10 flowered; scape erect, glabrous, faintly
ridged; bracts ovate, acuminate, membranous, about 7-nerved, persistent; pedicels
short, up to 1 cm in fruit. Flowers green, brownish- or yellowish-green, 16-28 mm
long; perianth-segments fused 1/3 of the way, with a broad 5-nerved central band,
about equal, spreading somewhat near the apex (often with one segment bent outwards
below), outer slightly narrower than inner. Stamens with the anthers obtuse or apiculate
at the apex and diverging at the base. Ovary with about 20 biseriate ovules per cell;
style ultimately reaching to the base of the anthers, the 3 stigmatic lobes triangular,
papillate. Capsule varying in size, 9-18 mm high, quadrate to oblong in outline,
trigonous, often very shortly stipitate; seeds 6 mm in diam.
Distribution. — Eastern Cape, Orange Free State, Transvaal. On grassy slopes,
apparently widespread but never common.
Cape. — Oudtshoorn: near Oudtshoorn, fio/wi' (BOL 12398, PRE). Steytlerville: Steyt-
lerville. Barker 5064 (NBG). Graaff-Reinet; near Graaflf-Reinet, Barker 7098 (NBG).
Cradock: Mortimer, Kensit (BOL 9241). Queenstown: plains near Queenstown,
Galpin 2203. Middelburg: Bangor Farm, Bohis (BOL 14108). Murraysburg: near
Murraysburg, Tyson 85 (GRA). Somerset East: MacOwan (BOL). Port Elizabeth:
12 miles N.E. of Port Elizabeth, Barker 6916 (NBG). Alexandria: Kaba, limestone
slope, Johnson 250. Albany: Grahamstown, Daly & Sole 340 (GRA, BOL). King
William’s Town: Kei Road, Glenifler, Ranger 33. East London: Fort Jackson near
East London, Galpin 3123 (GRA, PRE); East London, Comins 1361. Stutterheim:
Kubusie River valley at St. John’s, rocky grass-slope, rare, Acocks 9426. Kentani:
near Kei Mouth, Flanagan 1152. Mount Currie: Clydesdale, Tyson 2124 (GRA,
NBG). Aliwal North; Ruigtefontein, Thode A 1872.
Natal. — Lions River: Culvers near Rosetta, Rogers 28149 (GRA). Pietermaritzburg:
Pietermaritzburg Race Course, Barker 4438 (NBG).
Orange Free State. — Fauresmith; on Reserve, foot of hill. Smith 5570, Henrici 4407,
4198. Kroonstad; Pont 648, 638.
Transvaal. — Bloemhof: “ Kaffraria ”, Bur tt Davy (PRE 12774), in leaf but sterile,
March. Pretoria: Hammanskraal, Mauve 4028, in flower, leafless, July; in leaf but
sterile, December 1959.
Harvey adopted Zeyher’s manuscript name, Uropetalum ciliare under which the
specimens Fcklon & Harvey 48 were distributed. There is a type specimen at Dublin
and one at Kew. Since Harvey’s hand-writing is on the Kew sheet and since he quotes
123
the Cooper specimen mounted next to it, I agree with Kew that it should be made
the lectotype. The material of the isotype at NBG is excellent. Bentham and Hooker
state that Harvey’s figure (t. 170) is not accurate in that the filaments are drawn free
from the tube, and the anthers appear basifixed. The filaments are fused to the tube
below, where the perianth-segments are also fused. Higher up where the segments
become free (although still connivent), the weak filaments are also free but are pressed
to the segments. They can be pulled away without disturbing the tissues of the
segments. The anthers are dorsifixed near the base; the locules often diverge or are
separated below the insertion. The figure was drawn from dried material.
The locality of D. readii Bak. is cited in the Flora Capensis as Fuller’s Farm,
Natal, but Miss M. D. Gunn, Librarian, points out that the farm was situated near
Fort Beaufort in the eastern Cape and is at present part of a forest reserve. The type
specimens have either lost their outer leaves or were hysteranthous. When Zahlbruckner
described D. megalanfhum he admitted that it might prove to be a variety of D. readii
Bak. In the eastern Cape the species flowers from November to January, at first
hysteranthously, later synanthously.
On the highveld it was seldom collected for its flowers so early in the season;
it is difficult to spot because of its small size and moreover its drab colour merges into
the background of dry earth. Here it apparently ffowers only once hysteranthously.
However, in cultivation, bulbs collected from the same area {Mauve 4028 from Ham-
manskraal near Pretoria) produced 2 leaves with the inflorescence, probably as a result
of receiving more water.
The ffowers of the highveld specimens are shorter than those from the eastern
Cape and the dorsal pubescence on the ribs of the leaf was absent.
When leaffess the species can be distinguished from D. marlothii by its glabrous
prophylls and its long style and Ungulate stigmas. In D. marlothii the prophylls are
shortly scabrid, the style is very short and the stigma capitate.
3. D. papillatum Oberm., sp. nov., affine D. ciliari, sed foliis glabris, scapo papillato,
fforibus minoribus, stylo brevi, stigmate capitato differt.
Planta minora ad 40 cm alta. Bulbus ovoideus ca 2 cm latus. Folia circ. 4 rosulata
primo lax spiralia glabra glauca. Racemus 4-7 fforus scapus costatus basi papillis.
Segmenta perianthii aequalia. Stylus brevis, stigmate capitato.
Small plants up to 40 cm high. Bulb avoid, ca 2 cm broad with thin brown outer
skins, produced into a neck, often producing several shoots simultaneously. Leaves
about 4 per flowering shoot, often hysteranthous in spring, forming an erect rosette
with the lamina in lax spirals, becoming straight with age, somewhat coriaceous,
glaucous, glabrous, except for the yellow, minutely ciliate margin, and occasionally
dorsally minutely papillate on the ribs, which are situated very close together. Racemes
up to 40 cm, secund, 4-7 flowered, overtopping the leaves; scape terete, ridged, with
the ribs minutely papillate, occasionally papillae absent; bracts minute, subulate,
deciduous; pedicels 3-5 mm in fruit. Flowers green, 1 cm long, segments fused about
1/3 of the way, constricted above ovary, equal, outer recurved from near the apex.
Stamens with the filaments fused to the perianth-tube below and pressed to the segments
above it, anthers basifixed, slightly sagittate at the base. Ovary oblong, obtusely
triangular, ending in a short, stout style; stigma forming a 3-lobed, papillate, apical
disc. Capsule round, trigonous, 7 mm long; seeds 3 mm in diam.
Distribution. — Transvaal, Orange Free State, northern Cape, Southern Rhodesia,
usually in sand. Common where found.
Transvaal. — Pretoria: N.E. of Hammanskraal on road to Rust de Winter, Codd,
3488 (PRE, holo.); Rust de Winter Research Station, on loose reddish sandy soil‘
124
C. A. Smith (PRE 28284); Eerste Fabrieke, Young 1950; Pretoria, Hatfield, Obermeyer
68. Soutpansberg : Messina, Pole Evans (13121); Bandolierkop, Riley. Pietersburg:
Pietersburg commonage, Hafstrom & Acocks 214.
Orange Free State. — ^Fauresmith, Groenvlei, Kies 382.
Cape. — Gordonia: 14 miles west of Kuruman-Gordonia boundary, Kuruman River,
compact white sand, Leistner 2222 (KMG). Barkly West: Pniel near Barkly Bridge,
Wilman (KMG 4061). Kimberley: Kimberley Golf Course, red sand, Brueckner
513 (KMG).
Southern Rhodesia. — Bulalima-Mangwe, Plumtree, McLeod 11, pro parte (PRE).
Flowering in the summer months. The flowers smell sweetly like honeysuckle,
in the evening.
4. D. gracillimum Bak. in FI. Cap. 6: 446 (1897). Type: Cape, Colesberg, Shaw
(K, holo.! PRE photo.).
D. polyphyllum Bak. in Bull. Hei'b. Boiss. ser. 2: 4: 1001 (1904); not of FI. Cap.
6: 446 (1897). Type: Transvaal, Germiston, Modderfontein, Conrath 673 (K, holo.,
GZU, sheet 041987, iso. ! PRE, photo.). D. involutum Suess. in Mitt. Bot. Staatssamml.
Miinchen 2: 53 (1950). Type: South West Africa, Karibib, sandy grasslands, Dinter
6829 (M, holo., B, BOL, PRE, iso.!).
Ornithogalum dipcadioides Bak. in Bull. Herb. Boiss. ser. 2:4:999 (1904). Type:
South West Africa, near Karibib, Rautanen 445 (Z, holo.!, PRE, photo.).
Plants glabrous, 5-30 cm high. Bulbs ovoid, about 3 cm in diam., attenuated
into a short neck; outer tunics greyish-brown, leathery, inner membranous, transparent;
producing 1-5 flowering shoots simultaneously. Leaves 4-12 per shoot; at first
hysteranthous, later synanthous, flaccid after rains, lamina filifonn or narrowly linear
and canaliculate, up to 30 cm long (usually about 18 cm long), base dilated, mem-
branous, upper part of leaf often spirally twisted when young. Raceme simple, secund,
as long as, or, in spring usually longer than the leaves; scape terete, slender, often
arcuate near the base; bracts varying in size, lanceolate, acuminate, persistent or
caducous; pedicels up to 8 mm in fruit. Flowers 3-12 on a raceme, green sometimes
with a reddish tinge, small, up to 14 mm long, segments fused 1/3 of the way, outer
slightly narrower, recurved when fully opened, inner erect with only the tips recurved;
stamens typical; ovary cylindrical with about 20 biseriate ovules in each loculus, style
hollow, stigma 3-lobed, rounded, papillate. Capsule globose, varying in size, 8-18
mm in diam. trigonous, walls thin, leathery; seed 2-3 mm in diam. to twice as large
in larger capsules.
Distribution. — Northern Cape, Orange Free State, Natal, Swaziland, Transvaal,
Southern Rhodesia, South West Africa, in grassveld, often in sand.
Cape. — Barkly West, Bennie 595 (GRA). Herbert: Valschfontein, Esterhuysen (BOL).
Kimberley: Ronaldsvlei, Pocock (BOL 26949); Soutpan, open alluvial sandveld
above river in pure stands, Brueckner 1008, 1010 (KMG, BOL). Hay: Gordonia
boundary, rare, low quartz and limestone ridges, Acocks 2036; Rietkloof, yellowish
sand, locally common, Esterhuysen (KMG 5402); Ongeluk, frequent on hard red sand,
Wilman (KMG 5387). Namaqualand: sand between hills 4 miles S. of the Orange
River near its mouth, Pillans 5613 (BOL). Prieska: Bryant J188 (SRGH). Angeliers-
pan, adjoining Jagbult, Karroid Broken Veld, Acocks 12675.
Orange Free State. — Parys: Parys, Rogers 740 (GRA, BOL, PRE). Fauresmith:
Reserve, Verdoorn 2305.
125
Transvaal. — Warmbad; Leendertz 1520 (GRA, PRE). Lydenburg: MacMac Falls,
Prosser 2048. Piet Relief: 11 miles S. of Amsterdam, Codd 4720. Ermelo: Mavierie-
stad, Pott 5252. Germiston: Modderfontein, Gilliland (J 26773). Witbank: Gilfillan
(Hb. Galpin 7265).
Swaziland. — Mbabane: Ukutulu, Compton 25228.
Bechuanaland. — Molotoana Siding, woodland on sandy flats, Codd 8946.
South West Africa.^ — Omaruru: Omaruru, Schmidt 8; Namib, Brandberg, south
side, Merxmuller 1678 (M).
Southern Rhodesia. — Umtali: Sabie Drift, Wild 4661 (SRGH, M).
This species was described from a depauperate specimen flowering in spring and
using its stored resources in the effort. It remains small, few flowered and with the
leaves soon withering from the tips downwards during the usually dry early summer
months. In late summer after good rains, however, the plant shows a very different
form. The now numerous leaves become bright green and succulent but so weak
that they bend over or even lie prostrate on the sand. The subsequent scapes are of
about the same length as the leaves. The capsules of these plants are 2 to 3 times the
size of those from the early depauperate specimens. In fact the older farmers call
these plants “ Oupa Groot Toon ” an allusion to the large capsules. It flowers from
September to March. “ Flowers expand from about 6 p.m.-6 a.m. and then begin to
close again ”. H. Bolus.
5. D. marlothii Engl, in Engl. Bot. Jahrb. 10: 3 (1889); FI. Cap. 6: 446 (1897).
Type: Cape, Bechuanaland, Kuruman, Marlotli 1041 (B. holo.!, PRE, iso.).
D. bakerianum Schinz in Verb. Bot. Ver. Brand. 31: 217 (1890), non Bolus. D.
durandianum Schinz in Dur. & Schinz, Consp. FI. Afr. V: 374 (1895), nom. nov. for
D. bakerianum Schinz. Type: South West Africa, Ovamboland, near Olukonda,
Schinz 410! (Z, syn. 1, PRE, photo., GRA, iso.); Rautanen 103 (Z, syn.l, PRE, photo.).
D. polyphyllum Bak. in FI. Cap. 6: 446 (1897). Type: Natal, Groenberg, Wood 1166
(K, holo., NFI, iso.!, PRE, photo.). D. oligotrichum Bak. in Bull. Herb. Boiss. Ser. 2,
4: 1000 (1904); FI. PI. Afr. 24, t. 934 (1944). Type: Transvaal, Germiston, Modder-
fontein, Conrath 676 (K, holo., GZU, iso.!, PRE, photo.). D. ernesti-ruschii O'mXQX
in Fedde, Rep. 19: 185 (1924). Type: South West Africa, Lichtenstein, Auas Mts.,
Rusch sub Dinter 4346 (B, holo.).
Plants 50-100 cm high. Bulb globose produced into a neck and with a brown
leathery skin. Leaves hysteranthous at first, synanthous later in the season, many, often
in lax spirals when young, linear to filiform, 10-45 cm long, 1-4 mm wide, upper half
attenuate into a long filiform point, dilated into a white membranous base, sparsely
to densely bristly on the margins and ribs dorsally in lower half, often with patches
of short white curly retrorse pubescence; inner leaves longer, filiform and glabrescent.
Raceme longer than the leaves, subsecund, 10-20 flowered; scape terete; bracts
lanceolate, aristate, early caducous; pedicels up to 1 cm in fruit, arcuate at the base.
Flowers about 18 mm long (shorter early in the season) segments subequal, fused in
lower third, outer segments spreading slightly in upper half, inner connivent, forming
a tube with the apex slightly recurved. Filaments with a few transverse wrinkles and
a small curve where they become free from the tube, appressed to the inner se^ents
above it; anthers 6 mm, hastate at the base. Ovary narrow ovoid, narrowed into a
semi-persistent short style; stigma capitate, situated below the anthers, 3-grooved.
Capsule oblong to globose, about 15 mm long, with dorsal ridges and the sidewalls
splitting near the axis; seeds typical, 2 mm in diam.
Distribution. — ^Northern Cape, Orange Free State, Basutoland, Natal, Swaziland,
Transvaal, Bechuanaland, South West Africa, Southern Rhodesia; usually solitary
plants in savanna-grassveld, often on sand but also on firmer soils. Fairly common.
126
Cape. — Barkly West: Hai'develd, Acocks 2124. Kuruman: Ferrar (KMG 6681).
Vryburg: Duffield, Brueckner 1157. Postmasburg, 26 m. E.S.E. of Postmasburg,
Leistner 1649. Taungs: Andalusia, Cartwright (KMG 7010). Mafeking: Moshesh
near Mosita, 537 (KMG); Hay: Groenwater, 8569. Mount Currie:
Mount Currie, Goossens 300.
Orange Free State. — Heilbron: Heilbron, Goossens 578. Bethlehem: Bethlehem,
Rogers (TRV 2954). Bloemfontein: Bestersput, Acocks 8499. Bosholf: Smitskraal,
Biirtt Davy (PRE 12887). Kroonstad: near Vais River, Pont 317.
Basutoland. — 'Leribe, Dieterlen 237; Maletsunyane Falls, Jacot-GuiUarmod 1767.
Natal. — Dundee: 7 miles E. of Dundee, Co(7(7 5607. Estcourt: Tabamhlope Research
Station, West 587. Paulpietersburg: Galpin 9294.
Swaziland. — Goedgegun, Nel.
Transvaal. — Wolmaransstad: Boskuil, Sutton 191. Christiana: “ Kaffraria ”, Rurtt
Davy (PRE 12826). Pretoria: Pretoria, Smith 1161; Witfontein 53, Hardy 426.
Bronkhorstspruit : Rooikop, Smuts & Gillett 2518. Nelspruit: Kruger National
Park, Skukuza, von 1801. Barberton: Upper Moodies, Co/p/n 583. Pieters-
burg: Blouberg, Mohlakeng Plateau, Codd & Dyer 5194. Waterberg: 9 miles N.E.
of Thabazimbi, Codd 4800. Ermelo: Nooitgedacht 10, Potter 1761. Witbank:
Sondagsfontein, Thode A 2860. Belfast : Donkerhoek, between Belfast and Lydenburg,
Codd 8032. Johannesburg: Jeppestown Ridge, GilfiUan 91a. Heidelberg: Schlechter
3518.
Bechuanaland. — Gaberones, van Son (TRV 28670).
South West Africa.^ — 'Windhoek: Bingenheim, Gauchape, Volk 1119. Gobabis:
farm Nico, MerxmUIler 1167 (M, PRE). Grootfontein : Rehm (M); 30 miles N. of
Gautscha Pan, Story 6430, 6449; Karakovisa, Schoenfelder 16. Otjiwarongo:
Okahandja, Bradfield 312.
Southern Rhodesia. — 'Matobo: farm Besna Kobila, Miller 2080 (SRGH); Bulawayo,
Brain (SRGH). Salisbury: Enterprise, Wild 3899 (SRGH). Inyanga: Mare River,
Wild 1581 (SRGH). Nyamandhlovu : Pasture Research Station, Plowes 1758 (SRGH).
Bulalima Mangwe: Plumtree, McLeod 10.
The synonym, D. polyphyllum Bak. from Natal, was described as glabrous. It
is possible that the outer leaves were destroyed leaving only the inner new leaves which
are usually glabrous. The isotype collection (NH) is leafless.
This species can be recognized by its long flowers, characteristic pubescence, short
style and swollen stigma. It flowers from September to April. At times a somewhat
pungent smell is emitted at night.
6. D. longifolium (Lindl.) Bak. in Journ. Linn. Soc. 11: 397 (1871). FI. Trop.
A. 7: 519 (1898).
Uropetalum longifolium Lindl. in Bot. Reg. t. 974 (1826); Kunth, Enum. 4: 378 (1843).
Type: Portuguese East Africa, Mozambique, Mafamale Island, 10° 25' S., 39° 45'-43;
E.* growing in fine loose sand, Forbes (t. 974, icono, CGE, para., PRE, photo.).
Dipcadi venenatum Schinz in Bull. Herb. Boiss. 4, App. 3: 42 (1896); El. Trop. Afr.
7: 518 (1898). Type: South West Africa, Ovamboland, near Olukonda, Rautanen
126 (Z, holo. !, PRE, photo.). D. zambesiacum Bak. in FI. Trop. Afr. 7: 517 (1898).
* This is the position given by Exell and Hayes in Kirkia I, 136. The Director, C.I.C.A., Louren^o
Marques, suggested that it could be Mafamede Island (uninhabited) opposite Antonio Enes, more
to the South.
127
Type: Portuguese East Africa, Zambesi Delta, between Mambucha and Vicenti, Scott
(K, holo.). Bak. in FI. Trop. Afr. 7: 519 (1898). Type: Bechuanaland,
Ngamiland, Kwebe, Lugard 57 (K, holo., GRA, iso.!, PRE, photo.). D. anthericoides
Engl. & Gilg in Warb., Kunene-Sambesi Exped. 193 (1903). Type: Angola, Bondo,
on the Kubango and Kuito, on clay near reeds, Baum 499 (B, holo.j, M, iso. !, PRE,
photo.).
Glabrous plants up to 60 cm high. Bulb ovoid, 1-3 cm in diam., outer tunics
dark, leathery; occasionally producing suckers or bulbils. Leaves 4—7, linear to
linear-lanceolate, up to 40 cm long and 9 mm broad, tapering towards the apex and
base, not vaginate, firm to coriaceous, closely ribbed. Raceme simple, overtopping
the leaves, scape terete, tapered at the base, often arcuate just above it; bracts small
lanceolate, subulate, caducous; pedicels lengthening in fruit to about 15 mm. Flowers
4-12, subsessile to shortly stalked, subsecund, ultimately laxly arranged on rhacliis,
green, brownish, yellowish or whitish-green; outer and inner segments fused 1/3 of
the way, equal, linear, about 13 mm long, with about 5 central nerves, cucullate, the
outer recurved from below the middle, the inner broader than the outer, slightly
spreading. Stamens with anthers about 5 mm long, obtuse at the base. Ovary contracted
into a stipe at the base, with about 14 biseriate ovules per cell; style about as long
as the ovary, hollow, papillate ending in a 3-lobed stigma reaching halfway the anther
cells. Capsule on a pedicel 1-2 cm long, sub-quadrate, broader than long, either small
(9 by 12 mm) or large (15 by 20), trigonous, base stipitate; seed 5 mm in diam.
Distribution. — Mozambique, the Federation, Angola, northern South West
Africa, Bechuanaland.* A hygrophyte. Sometimes appearing as a weed in cultivated
lands and lawns where it propagates itself by runners.
Mozambique. — Lourenqo Marques, Sul do Save, near the sea, Pedro 3860 (LM.)
Mozambique: Mocuba, Namagoa often in damp ground, Faulkner 259.
Southern Rhodesia. — Darwin: Umvukwes, Umsengedzi R. on sandy river bank.
Wild 3969 (SRGH). Melsetter: Nyshama farm. Crook 458 (weed in cultivated lands,
SRGH). Salisbury: Prince Edward Dam, in a vlei, Wild 3693 (SRGH). Bikita:
Devuli Bridge 7 miles W. of Birchenough Bridge, Obermeyer 2494. Bulalima Mangwe :
Embakwe, Feiertag (SRGH 45420).
Northern Rhodesia. — Livingstone: Livingstone-Victoria Falls road. Young 1087.
Mongu: between Mongu & Lealui, grassy flood plain, Drummond & Cookson 6266
(SRGH, PRE).
South West Aerica. — Caprivi Strip: Katima Mulilo area, about 5 miles from Katima
on road to Linyanti, occasional on sandy moist flats, Killick & Leistner 3091. Okavango
Native Territory: 5 miles E. of Masari Exp. Farm on road to Nyangana, in low lying
areas, black, sandy clay, de Winter & Wiss 4113. Grootfontein: Namutoni, Breyer
in TRV 20686; Gautscha Pan, 70 (M); Simkue, 6204. Okahandja:
Quickborn, along riverbank, Bradfield 142.
Bechuanaland. — Ngamiland, Kwebe, Lugard 57 (K, GRA).
The cultivated plant which was figured and described must be regarded as the
holotype but apparently it was not preserved. Specimens of the original Forbes
collection are to be found at Cambridge and since Lindley’s handwriting is on it,
Mr. John Lewis suggests it should be made the paratype. Accordingly the sheet at
the British Museum would be a paratype collection.
The plate shows a plant with much wider and flabby leaves, probably the result
of hothouse conditions. But Lindley mentions in his description that the outer leaves
* Baker mentions Barter 3441 from Nupe, Niger Territory in Flor. Trop. Afr. 7: 519 (1898). Not
seen by me.
128
were linear. Baker in his description of it in the Flora of Tropical Africa 7: 519,
stated that the capsule is sessile but this was incorrect for the specimens I have examined;
the pedicel is very short during the flowering stage but lengthens after fertilization and
is up to 2 cm long in fruit.
Schinz called his species D. venenatum because of a note by Rautanen which
said that it was poisonous, for goats after eating it, started swelling and died. The
colour of the flower was said to be white-yellow by Rautanen. Either the note does
not belong to this specimen (the colour of the flower seems to indicate another plant)
or the swelling was caused by something else. As far as we know, the bulbs are edible.
It was experimentally fed to cattle in large quantities at the Central Research Station
at Mazabuka, Northern Rhodesia in December 1931, but was found to be non-toxic
(cf. note on sheet CBS 533 at PRE). It flowers during the summer months from
November to April.
7. D. crispum Bak. in Journ. Linn. Soc. 11: 399 (1871); FI. Cap. 6: 449 (1897).
Type: Hopetown, near the Orange River between Puffadder Halt and Bare station,
Burchell 2682 (K, holo!, PRE, photo).
D. setosum Bak. in Journ. Linn. Soc. 11: 398 (1871). FI. Cap. 6: 447 (1897). Type:
Cape, Namaqualand, Modderfontein, Whitehead (TCD, holo.!, PRE, photo.). D.
volutum Bak. in FI. Cap. 7: 448 (1897). Type: Cape, Namaqualand, Scully 214 (K,
holo.!, PRE, iso., photo.). D. undulatifolium Schinz in Viert. Nat. Ges. Zurich, 57: 534
(1912). Type: South West Africa, Ovamboland, Ondonga, Kestila 4 (Z, holo.!, PRE,
photo.). D. tortile Dyer in FI. PI. Afr. t. 956 (1944). Type: Cape, Vanrhynsdorp,
Rood (PRE 27198, holo.!). D. crispociliatum Suesseng. in Mitt. Bot. Staatssamm.
Miinchen, 2: 51 (1950). Type: South West Africa, Karibib, sandy flats with Aristida
obtusa, Dinter 6830 (M, holo.!, PRE, iso.).
Uropetalum crispum Burch, ex Gawl. in Bot. Reg. sub t. 156 (1816), nom. nud.
Plants up to 30 cm high, variable, usually hairy. Bulb ovoid with yellowish scales
forming a neck (as it is usually situated some distance below the surface). Leaves
about 4, usually in a rosette, variable, linear to linear-lanceolate, 6-20 cm long, 4-16 mm
broad, glaucous, soft (under moist conditions), straight or spirally twisted; margin
straight or much crisped, usually sparsely to densely ciliate with white, long flat, hairs;
ventral surface glabrous, closely ribbed when dry, dorsal glabrous or bristly with a
few to many long and short hairs on the ribs especially near the base. Raceme up
to 30 cm long, laxly 6-14 flowered; scape terete, smooth or ridged with papillae on
the ridges to densely and shortly pubescent (when the papillae lengthen into hairs);
bracts small, lanceolate, acuminate, membranous, about 5 nerved, persistent, margin
entire or ciliate; pedicels short, 2-5 mm, glabrous or densely and patently setose.
Flowers brown to orange, 15-20 mm long; segments fused at the base into a bulbous
tube for about one half of their length, outer slightly longer with a short caudate tip,
narrower than inner, all cucullate (the inner segments not surrounded by outer). Stamens
with the filaments fused to the tube, the free apical part short. Ovary turbinate, style
smooth or papillate above, about as long as the ovary, with three stigmatic lobes
which form a shaggy ridge at the apex, just reaching the base of the anthers. Capsule
large (i.e. Comins 1183) globose, brown, 2 cm in diam. leathery or smaller in others;
seeds {Comins 1183) 9 mm in diam. or smaller.
Distribution.— Western Cape, South West Africa; in stony areas or on sandy
flats. Rare.
Cape. — Namaqualand; Richtersveld, Kalkfontein, Meyer {suh. Marloth 6430); Stein-
kopf, Meyer (sub. Marloth 6449); Eenriet, Herre 13 (BOL). Calvinia: near Hantams
River, 30 miles N.W. of Calvinia, Lewis 2746 (NBG); 26 miles from Calvinia on
Loeriesfontein road. Western mountain karoo, Comins 1 183. Clanwilliam: Welbedacht,
129
Lewis 2554 (NBG). Vanrhynsdorp: Vredendal, Smith 4670 (NBG); Knersvlakte
11 miles N. of South River Bridge, Salter 5473 (BOL). Herbert: Douglas, Marloth
9473; near Hopetown, Muskett 30 (BOL). Britstown: Pole Evans.
South West Africa. — Bethlehem: Sandverhaar, Pearson 4439 (BOL); Halenberg,
Dinter 6613 (B!). Keetmanshoop: Great Karasberg, Nariida Slid, Pearson 8297
(BOL). Gobabis: farm Nico, flooded limestone pan, Merxmiiller 1156 (M). Rehoboth:
Ababis, Strey 2623. Karibib: sandy Aristida obtusa flats, Dinter 6830. Ovamboland,
Odonga, Kestila 4 (Z).
In the Flora Capensis Baker noted the variability of the leaf, probably caused
by local climatic conditions. When cultivated it is long, straight and glabrous; under
dry conditions it is small, spirally twisted and the margin is densely ciliate and undulate.
Some specimens fall between these two extremes. D. tortile Dyer is a fairly good
match of the type of D. crispum (the wild collection) but is more hairy. D setosum
is a good match of D. volutum\ these have elongated leaves and few hairs. D. crispo-
ciliatum Suess. has very small and hairy leaves; it is near D. undulatifolium.
It flowers during the summer months in the summer rainfall area and during the
winter months in the winter rainfall area. “ Sweet smelling, strongly scented ” (Marloth).
According to Marloth it is poisonous.
8. D. bakerianum Bol. in Journ. Linn. Soc. 18: 394 (1881); FI. Cap. 6: 450
(1897). Type: Cape, Murraysburg, Bolus 2059 (BOL, holo.!, PRE, photo.).
D. brevipes Bak. in Kew Bull. 1901: 136 (1901). Type: Bechuanaland, Kwebe Hills,
Mrs. Lugard s.n. (K, holo.). D. rautanenii Bak. in Bull. Herb. Boiss. ser. 2, 4: 1000
(1904). Type: South West Africa, Ovamboland, Omakunde, Rautanen 437 (Z, holo.,
PRE, photo.). D. gonocarpum Suesseng. in Mitt. Bot. Staatssamml. Miinchen, 2: 52
(1950). Type: Southwest Africa, Grootfontein, Rietfontein, Gesundbrunnen, Relini,
22 Nov. 1939 (M, holo.!, PRE, photo.).
Plants glabrous, 25-40 cm high. Bulb ovoid, about 2-5 cm in diam. outer tunics
thin. Leaves rosulate, usually 4 per flowering shoot, ovate-lanceolate, 15-30 cm long,
vaginate and up to 3 cm broad below, upper half tapering into a long slender tip, dark
green, shiny, soft; innermost leaf much narrower. Raceme 10-15-flowered, the flowers
at first close together, not secund, ultimately lax, overtopping the leaves; scape terete,
somewhat swollen; bracts small, lanceolate, acuminate, membranous, thin, persistent,
shrivelling in a concertina-like way, or spirally from the tip downwards; pedicels very
short, up to 3 mm in fruit. Flowers about 18 mm long, green to yellow-green, segments
equal, fused for about 2/3 of the way (inner not surrounded by outer). Filaments
fused to the tube, free and flattened at the tip, anthers protruding from the mouth
of the tube, versatile with the locules diverging below. Ovary truncate with an apical
swelling on each outer corner; style longer than the ovary, stigma 3-lobed, papillate.
Capsule oblong, about 15 by 12 mm, knobbed; seeds oblong, 7 mm.
Distribution. — -Cape (Karoo, northern Cape), Bechuanaland, South West
Africa, sandy areas.
Cape. — GraafF-Reinet : near Graaff-Reinet, Bolus (GRA). Prieska : Prieska com-
monage, on soft sand, fairly common, Bryant J 187. Kimberley: Koedoesrand,
Panheuwel, Jacobs.
Bechuanaland. — Pole Evans & Erens 1226. Ghanzi, de Winter 7409.
South West Africa. — Keetmanshoop: Great Karasberg, Nariida Slid, Pearson 8319,
7979, 7980 (BOL). Mariental: Swartrand, 40 miles W., Basson 164. Windhoek:
Auas Mts., farm Krumhoek, Merxmiiller 795 (M); Neudam College, van Vuuren 1006;
26 miles S. of Windhoek on road to Rehoboth, Gies & van Vuuren 954. Omaruru:
Otjihorongo, on the Ugab, S.W. Anigab, Merxmiiller 1626 (M, PRE). Grootfontein:
130
Klein Namutoni, Breyer in TRV 20685 ; Tsumeb, Naegelsbach 44 (M). Gobabis ;
farm Uitsig, Merxmuller 1109 (M).
This species is easily distinguished by its knobbed capsule and wide dark green
leaves. It flowers from November to February.
9. D. glaucum (Ker-Gawler) Bak. in Journ. Linn. Soc. 11: 401 (1871); FI. Cap.
6: 450 (1897); Steyn, The Toxicology of Plants in South Africa, 503 (1934).
Uropetalon glaucum Burch, ex Ker-Gawler in Bot. Reg. 6: 156 (1816). Type: Cape
Griqualand West between Asbestos Mts. and Wittewater, Burchell 2066 (K, holo.).
Dipcadi longibracteatum Schinz in Verb. Bot. Ver. Prov. Brand. 31 : 218 (1890). Type :
South West Africa, Ovamboland, between Uukuambi and Ondonga, Schinz (Z, holo.,
PRE, photo.). D. magnum Bak. in FI. Trop. Afr. 7: 522 (1898), e descr. Type:
Bechuanaland, Ngamiland, Kwebe, Lugard 88 (K, holo.). D. gracilipes Krause in
Engl. Bot. Jahrb. 48: 357 (1912). Type: South West Africa, Bethanien, Sandverhaar,
Schaefer 376 (B, holo.!, PRE, photo.).
Urginea lorata Bak. in Bull. Herb. Boiss. ser. 2, 3: 664 (1903). Type: South West
Africa, Ovamboland, Ondonga, Rautanen (Z, holo.!, PRE, photo.).
Plants up to 120 cm tall. Bulb ovoid, flattened at the base, 3-6 cm in diam.
Leaves 6-9 in a rosette, glaucous, shiny, sub-fleshy, narrow lanceolate, up to 35 cm
long, 10 mm broad, attenuated and clasping at the base, apex acute or acuminate,
closely but faintly ribbed, innermost leaves narrower than outer. Raceme with flowers
arranged in a spiral (not secund), congested at first, lax afterwards with up to 40 flowers;
scape terete, lengthening during anthesis; bracts linear, membranous, whitish, persistent,
shrivelling at an early age; pedicels variable in length, at first short, thin and ascending,
becoming patent and sturdy and up to 8 cm in length with the apex curved upwards
in fruit. Flowers up to 2-5 cm long (excluding appendages of outer segments), fleshy,
green with a brownish margin, segments fused about halfway, outer with caudate
appendages varying in length from 4-8 mm, inner Ungulate with a yellowish to orange
margin along the slightly recurved tips. Stamens with the connective produced into a
short papillate appendage above the locules which are rounded at the base. Ovary
oblong, style about as long as the ovary with the stigmatic apex indistinctly 6-lobed.
Capsule globose, 2 cm in diam. skin membranous; seeds 8 mm in diam.
Distribution. — Northern Cape, Bechuanaland, northern South West Africa,
Southern and Northern Rhodesia, Transvaal, Orange Free State.
Cape. — Richmond: collector unknown. Prieska: Bryant J 347 (SRGH). Barkly
West: Victoria Farm, Marloth 960. Kuruman: between Hopetown and Kuruman,
Strekdale (BOL. 2030). Kimberley: 30 miles W. of Kimberley on Douglas road, Lewis
3610 (NBG). Gordonia: slopes of Saalskop between Groblershoop and Upington,
Hardy.
Orange Free State. — Fauresmith: Luckoff, Henrici; Fauresmith, Smith 5331. Har-
rismith: Harrismith, in damp places in catchment area, Phillips 3501.
Transvaal. — Soutpansberg: Zoutpan, Galpin 13785; Obermeyer, Schweickerdt &
Verdoorn 291; 259. Pietersburg: Brakriviervlei, Bremekamp & Schweickerdt 54.
Sibasa: Kruger National Park, Baiandbai, Lang in TRV 32151. Lydenburg: Sekukuni,
Farm Middelin, Barnard 444; 207.
Bechuanaland. — Mochudi, Harbor in TRV 17026; near Derdepoort, Codd 8865.
South West Africa. — Rehoboth: Gravenstein, sandstone slope, Volk 11563 (M).
Windhoek: on road to Okahandja 6 miles N., Merxmuller 1003. Otjiwarongo:
Okosongomingo, Volk 462 (M). Grootfontein: Tsumeb, Naegelsbach 71 (M).
131
Okavango Native Territory, Runtu near Okavango River, de Winter 4045. Kaokoveld:
Etoshapan, farm Onguna, Walter 511 (M).
Southern Rhodesia. — Nyamandlovu : near Umfusa Bridge on Victoria Falls Road,
West 2558 (SRGH). Victoria Falls, Rogers 13059 (BOL). Melsetter: Hotsprings
near Odzi River, Chase 4323 (SRGH); Birchenouah Bridge, Sabi River, Obermever
2495.
Northern Rhodesia. — Mazabuka, Central Research Station 539.
D. magnum Bak. from Ngamiland agrees with D. glaucum. The few specimens
seen from the Federation belong to a shade-loving form with shorter pedicels (about
2-5 cm in length) and somewhat smaller in stature. The Central Research Station,
Mazabuka, Northern Rhodesia, found it to be non-toxic to stock (cf. note on sheet
CRS 539, PRE).
D. glaucum seems to hybridize with D. bakerianum cf. p. 135. It flowers from
November to May.
10. D. viride (L.) Moench, Meth. Suppl. 267 (1791); Bak. in Journ. Linn. Soc.
11: 401 (1871); FI. Cap. 7: 449 (1897); FI. Trop. Afr. 8: 523 (1898).
Hyacinthus viridis L., Sp. PI. ed. 2, 454 (1762). Jacquin, Ic. Rar. t. 66 (1798). Red.,
Lil. t. 203 (1808). Type: Cape, collector unknown (LINN, holo., PRE, photo.).
Dipcadi filamentosum Medic, in Ust. Ann. 2, 13 (1791). D. umbonatum (Bak.) Bak.
in Journ. Linn. Soc. 11: 400 (1871); FI. Cap. 6: 448 (1897). D. comosum Welw.
ex Bak. in Trans. Linn. Soc. ser. 2, Bot. 1 : 247, t. 34, fig. 1-3 (1878); Engl., Hochge-
birgsfl. Trop. Afr. 167 (1892); Bak. in FI. Trop. Afr. 7: 521 (1898). Type: Angola,
near Lopollo, on sandy hills in the subtemperate region, Welwitsch 3732 (BM, holo.).
D. lateritium Welw. ex. Bak. in Trans. Linn. Soc. ser. 2, Bot. 1 : 247 (1878); FI. Trop.
Afr. 7: 521 (1898). Syntypes: Angola, Huilla, Welwitsch 3716 (BM, holo., B, iso.!,
PRE, photo.) 3717. D. datum Bak. in FI. Cap. 6: 449 (1897). Type: Cape, Mount
Currie, Clydesdale, Tyson 2107 (K, holo., NBG, BOL, iso.!). D. lividescens Engl. &
Gilg in Warb. , Kunene-Sambesi Exp. 1 94 ( 1 903), ex descr. Type : Angola, left Kubango
bank near Kuimaroa, Baum 459 (B, holo.?). D. palustre Bak. in Bull. Herb. Boiss.
ser. 2, 4: 1000 (1904). Type: Transvaal, Germiston district, Modderfontein, Conrath
675 (K, holo., GZU, iso.!, PRE, photo.). D. conrathii Bak. in Bull. Herb. Boiss. ser. 2,
4: 1000 (1904). Type: Transvaal, Germiston district, Modderfontein, Conrath 677
(K, holo., GZU, iso.!, PRE, photo.). D. longicauda Engl. & Krause in Engl. Bot.
Jahrb. 45: 150 (1910). Type: South West Africa, Okahandja, hard sandy ground,
Dinter 425 (B, holo.!, PRE, photo.). D. helenae Beauv. in Bull. Soc. Bot. Geneve,
ser. 2, 5: 326 (1914). Type: Cape, Matatiele, Mafube, Helene Jacottet (G, holo.!,
PRE, photo.). D. stenophyllum Dinter in Fedde Rep. 19: 185 (1924). Type: South
West Africa, Windhoek, Auas Mts., Lichtenstein, Dinter 4371 (M, holo.!, PRE, photo.).
D. cinnabarinum Suesseng. in Mitt. Bot. Staatssamm. Munchen, 2: 50 (1950). Type:
South West Africa, Grootfontein, Waldorf-Siid, Rehm 6789 (M, holo.!, PRE, photo.).
D. sulcatum Suesseng. in Bot. Staatssamm. Munchen 2: 53 (1950). Type: South
West Africa, Grootfontein district, Rietfontein on exposed limestone, Rehm (M, holo!,
PRE, photo.). D. geniculatum Dinter & Suesseng. in Mitt. Bot. Staatssamm. 1: 51
(1950). Type: South West Africa, Karibib, Dinter 6789 (M, holo., PRE, iso.).
Lachenalia viridis (L.) Thunb., Prod. 64 (1794). Ait., Hort. Kew. 1 : 462 (1789); ed. 2,
ii, 285 (1811). Wind., Sp. Plant. 2: 174 (1799).
Zuccagnia viridis (L.) TTiunb., Nov. Gen. Plant. 9: 127 (1798); in Roem. Archiv. 2: 2
(1799); FI. Cap, ed. Schult. 328 (1823).
Phormium viride (L.) Thunb., Nov. Gen. Plant. 9: 127 (1798), in syn. nom. nud.
132
Uropetalum viride (L.) Gawl. in Bot. Reg. sub t. 156 (1816); Roem. & Schult., Syst.
Veg. 7: 620 (1829); Kunth, Enum. 4: 379 (1843). U. umbonatum Bak. in Saund.
Ref. Bot. t. 17 (1868). Type: Natal, Hutton (K, holo.).
Glabrous plants 15-120 cm high. Bulb firm (when dormant) or loose and scaly
(during growth period), occasionally bulbiferous or with runners. Leaves 1-4 per shoot
(occasionally hysteranthous) variable in size, linear to linear-lanceolate, 6-60 cm long,
3-20 mm broad, usually clasping at the base, attenuated in upper half to an acuminate
point, shiny, flaccid, somewhat succulent in well developed specimens, indistinctly
veined. Raceme central, as long as or longer than the leaves, 15-120 cm long, few to
many flowered; flowers close together at first, lax and subsecund when mature; some-
times the uppermost sterile, the perianth then remaining small but the filiform, often
red appendages of the outer segments developing fully; together with the filiform
bracts they give the raceme of the tall forms its characteristic plumose appearance
(D. comosum form); scape terete, smooth, shiny, sometimes arcuate at the base; bracts
lanceolate, acute or acuminate, apex often long accuminate to filiform, 4-30 mm long,
early deciduous or absent; pedicels 8-18 mm in fruit. Flowers with unequal perianth-
segments about 12 mm long (excluding appendages) fused 1/3 of the way; outer green
with a cream or orange or reddish or brown tinge, usually recurved from the middle,
linear-lanceolate, with a short to long, filiform appendage, 2-30 mm long; inner
forming a connivent tube, lanceolate, acute, with the tips curved outwards. Stamens
typical. Ovary turbinate, sometimes slightly umbonate near the apex of the valves;
about 25 ovules per cell, style about as long as the ovary, sparsely papillate, stigma
shortly 3-lobed, papillate. Capsule oblong in outline, 14 mm long, 9 mm broad,
trigonous; seeds 5 mm.
Distribution. — -Found all over southern Africa, except for the south western
Cape districts.
Cape. — -Ladismith: Huis River Pass, Lewis 753 (NBG). Calitzdorp: top of Rooiberg
Pass, rocky false Fynbos, locally frequent, Acocks 20364. Victoria West: Compton
21656 (NBG). Port Elizabeth: West (BOL 26953). Kentani: valleys, Pegler 1381
(GRA, BOL). Warrenton: Adams {KAAG WIT). Mafeking: Moshesh near Mosita,
among rocks on dam wall, common, eaten, Brueckner 636, 536 (KMG). Kimberley:
Rooipoort, 33 miles W. of Kimberley, near homestead, reddish sandy loam on lime,
occasional, Leistner 1205 (KMG).
Natal. — Utrecht: 6 miles S.E. of P.O. Groenvlei, Codd & Dyer 6274. Umvoti:
Greytown, town commonage, railway bank, Galpin 14753. Mooi River: Mooi River,
Wood 5623 (NBG).
Orange Free State. — Bethulie: Orange River, Flanagan 1643 (NBG). Bloemfontein:
near Rhenoster Spruit, Potts (UCOFS 2869). Fauresmith: Verdoorn sub Henrici
22>22.
Transvaal. — Belfast: Leendertz '\n TPy 9\69 . Pretoria: 4 miles N.E. of Hammans-
kraal, moist sandy soil beside stream, Codd 4781. Waterberg: 18 miles W. of Villa
Nora, sandveld, Codd 6587. Carolina: Bolus 12399 (BOL).
Swaziland. — Mbabane: Dense mountain grassveld 11 miles N.W. of Mbabane,
Codd 4756; 4757. Forbes Reef Road, swamp, Compton 27159.
Mozambique. — Lebombo Marsh, Junod 398. Inhaca Island, Moss (J 27417).
Southern Rhodesia. — Umtali: S.E. portion of Rangehill, Chase 5419 (SRGH).
Northern Rhodesia. — Mankoya: 50 miles E. of Mankoya on road to Kafue Hoek,
Drummond & Cookson 6720 (SRGH, PRE). Mazabuka, Central Research Station,
Vet. Officer (CRS 522).
133
South West Africa. — Grootfontein : near Grootfontein, Schoenf elder 133. Windhoek;
Anas Mts. at Anas Pass, MerxmuUer 796 (M).
The species appears variable and adaptable. We get the depauperate forms D.
lateritium, D. cinnabarinum, D. stenophyllum, etc.; the optimal form, D. comosum,
possibly a polyploid ; the Kalahari form with its deep seated bulb and arcuate scape,
D. geniculatum, etc. The type in the Linnaean Herbarium consists of part of a raceme,
lax and few-flowered with long-tailed outer perianth-segments. The old specimens
found in the Munich herbarium and that pictured by Jacquin are similar to it. It thus
shows the typical form (cultivated in Europe) to be medium sized. In the optimal forms
the plants produce racemes every few months during the summer; they are up to
4 ft tall and often have a comose appearance because of the sterile, long-tailed upper
flowers and long filiform bracts. All gradations are found from this tall form to the
small depauperate form which is only about 10 inches high and few flowered, the
flowers often with short “ tails ” and the plants flowering only once in spring and
then becoming dormant; the latter have comparatively large bulbs in relation to the
scanty parts above ground.
These forms are not confined to geographical areas and the plants are found on
different kinds of soil, on sand near the sea, heavy clay, on sour grassveld, etc. The
red colour sometimes present in the flowers (i.e. D. lateritium, D. cinnabarinum) may be
due to a chemical in the soil.
The “ tails ” lengthen to 3 cm in some cases or remain small (2 mm) in others.
They vary in length even on the same raceme. Hallier, with Arber concurring [cf
Monocotyledons by Arber, p. 1 14 (1925)] put forward the theory that the petal represents
the leaf-sheath and the tip the vestigial petiole. This theory appears to fit the segments
of Dipcadi very well, for the segment is concave but the appendage terete. Within
limits the vestigial petiole may vary in length.
The plants flower either once in early spring or intermittently during the summer
months. Sometimes a faint scent is emitted at night.
11. D. rigidifolium Bak. in Journ. Linn. Soc. 11: 399 (1871); Fl. Cap. 6; 448
(1897). Type: Transvaal, Pretoria dist., Apies River, Burke (K, holo.!, PRE, photo.).
D. rhodesiacum Weim. in Bot. Notiser 1937: 439, fig. 6. Type: Southern Rhodesia,
Makoni near Maidstone, Norlindh & Weimarck 4078 (LD, holo.).
Plants 20-50 cm high (60-80 cm fide Weimarck). Bulbs globose, 2-5 cm in
diam. outer tunics thin, light coloured. Leaves 2-3 per flowering shoot (often 2—3 shoots
per bulb), linear-lanceolate, up to 30 cm long and 25 mm broad, glabrous, coriaceous,
glaucous, closely ribbed, margin prominent, yellow, apex acute, apiculate, base clasping.
Raceme up to 50 cm long, secund, lax, 8-20 flowered; scape arcuate at the base; bracts
narrow-ovate, acuminate, persistent, occasionally the lower sterile; pedicels short, up
to 1 cm in fruit. Flowers about 15 mm long (excluding appendages) green; segments
fused about half way, outer spreading, or recurved in upper half with the caudate
appendages 5-14 mm long; inner segments obtuse, cucullate. Stamens with the anthers
acute or minutely mucronate. Ovary oblong, gradually narrowed into the style; style
about as long as the ovary, stigmas 3, lingulate. Capsule broadly oblong, 18 mm,
trigonous; no ripe seed seen.
Distribution. — Transvaal, northern Cape, Southern Rhodesia; rare.
Transvaal. — ^Pretoria: 1 mile N. of Hammanskraal, Codd 5620, 5621, Mauve 4048;
Rooikop, Smuts 382. Waterberg: Nylstroom, van Dam (TRV20949). Rustenburg:
Zwartruggens, Sutton 792. Bronkhorstspruit : Marais 6. Witbank: Witbank,
Rogers 24300.
134
Cape. — Hay; pale sandveld to the W. of Padkloof, y4cocA:j' 2206. Mafeking: Graham
Green (GRA).
Southern Rhodesia. — Salisbury, Eyies 1884 (SRGH; NBG); Ruwa River, Wild
2261 (SRGH).
It flowers during the summer months, November to April.
12. D. vaginatum Bak. in Fl. Trop. Afr. 6: 523 (1898). Type: Bechuanaland,
Ngamiland, Kwebe, Lugard 47 (K, holo.).
D. baiimii Engl. & Gilg in Warb., Kun. Sam. Exp. 194 (1903), ex descr. Type: Angola,
left bank of the Kubango below Kabindere, Baum 349 (B, holo.?).
Plants 10-40 cm high. Bulb ovoid, about 2-5 cm in diam. Leaf 1, clasping the
scape at the base, forming a long, flexuose, subterranean sheath, lamina spreading
above, folded, linear, 12-20 cm long, 3-10 mm broad, smooth, glabrous. Raceme
laxly 3-10-flowered, secund; scape terete, ribbed, shortly and minutely pubescent
with patent papillae or bristles; bracts small, membranous, persistent; pedicels up to
2 cm in fruit (fide Baker), with a pubescence similar to that of the scape. Flowers
green, about 12 mm long (excluding appendages); segments fused 1/3 of the way;
outer spreading with filiform appendages 8-20 mm long; inner lanceolate with a
cucullate apex. Anthers 5 mm long, filaments fused to the tube except for the tips.
Ovary turbinate, style somewhat shorter than the ovary, ending in 3 rounded papillate
stigmatic lobes. Capsule not seen; seed not seen.
Distribution. — Northern Cape, South West Africa, Bechuanaland, apparently
rare.
Bechuanaland. — Ngamiland, Kwebe, Lugard 47 (K, holo.).
Cape. — Kimberley, 4 miles S. of Kimberley, red loamy sand at foot of dolerite koppie,
rare, Leistner 1202 (KMG).
South West Africa. — 'Okahandja: Quickborn, Bradfield 311, 99c.
A specimen, Quarre 4343 from Elizabethville, Belgian Congo, seems to be near
this species but it has a glabrous scape. D. thollonianum Hua is probably related to
this species but bigger.
It flowers in November-December.
13. D. platyphyllum Bak. in Fl. Trop. Afr. 7: 518 (1898). Type: Bechuanaland,
Ngamiland, Kwebe, Lugard 44 (K, holo.).
D. ciliatum Engl. & Krause, in Engl. Bot. Jahrb. 45: 147 (1910). Type: South West
Africa: Brackwater, amongst small stones, Dinter s.n. (B, holo., PRE, photo.).
Plants 7-20 cm high. Bulb round, about 2-5 cm in diam. covered with leathery,
thin, light brown tunics. Leaf 1 (occasionally 2 if a second shoot is produced), clasping
the scape at the base, forming a long, flexuose, subterranean sheath; lamina lying
flat on the ground, ovate-lanceolate to linear-lanceolate, 5-15 cm long, 1-3 cm broad,
upper surface glabrous except for the crisped, ciliate margin, lower surface with short
bristles on the ribs, base obtuse, clasping, apex acute. Raceme up to 20 cm high, secund,
4-20 flowered, the flowers close together; scape terete, densely pubescent with short,
patent bristles or papillae; bracts small, ovate, acuminate, membranous; pedicels up
to 1 cm in length, with the same pubescence as the scape. Flowers green to brownish-
green, to yellowish-green, about 14 mm long (excluding appendages), segments fused
nearly half way, outer with a very short caudate appendage about 3 mm long; inner
lingulate, cucullate. Anthers 4 mm, filaments fused to the tube except for the tips,
pollen ellipsoid, large. Ovary turbinate with about 20 biseriate ovules, style somewhat
135
shorter than the ovary, ending in 3 papillate, short, stigmatic branches. Capsules
lounded in outline, tiigonous, circ. 15 mm in diam., thin-walled; seeds 5 mm in diam.
Distribution.— Northern Cape, Bechuanaland, South West Africa, Transvaal and
the Federation.
Cape. Kuruman: Batiharos, S/VA' 266 (KMG). Gordonia; 25 miles S.E. of Moenieput,
compact calcareous clay in a depression between red sand dunes, Leistner 1916 (KMG).
Bechuanaland. — Ngamiland, Kwebe, Lugard 44 (K, holo.).
South West Africa. — Rehoboth; between Gamis and Bullsmouth Pass on shale
plains about the headstreams of the Great Fish River, Pearson 8943 (BOL). Windhoek :
sandy flats 26 miles S. of Windhoek on Rehoboth road, Giess & van Vuuren 952.
Omaruru; Namib, Brandberg, Oritsaub, vlei-like sandflats, Merxmiiller 1686 (M).
Karibib: sandy flats, Karibib, Dinter 6788 (M). Gobabis: Sandfontein, WUman
(BOL 15232). Grootfontein: Rietfontein on loamy soil, Volk 2572 (M), Rehm (M,
fruiting). Okavango: farm Eckenberg, Okumuamakamp, Walter 222 (M).
Transvaal.— Soutpansberg: Messina, Moss & Rogers 206 (BOL).
Southern Rhodesia.— Victoria Falls, Rogers 5424 (BOL); in shade of mopani. Wild
3133 (SRGH). Urungwe, Sanyati— Chiroti Junction, in Brachystegia boehnii woodland
Wild 4240 (SRGH).
Northern Rhodesia. — ^Mapanza, Choma, open mopani bush, Robinson 2510 (M,
SRGH).
Nyasaland.— Zomba, Jackson 2092 (SRGH).
The species is closely related to D. vaginatum which has been found growing next
to it on several occasions. Whether D. vaginatum is perhaps a mutant is an open
question.
D. platypliylluin, called “ horinkies ” by farmers around Kimberley, may become
very common in overgrazed areas, according to Mr. O. A. Leistner.
It flowers November-January.
Uncertain Species
D. dinteri Bak. in Bull. Herb. Boiss. ser. 2, 1 : 788 (1901). Type; South West Africa,
Great Namaqualand, Awichab, Dinter 1 : 1038 (Z, holo.!, PRE, photo.). The material
is too poor for a definite conclusion. It may be a synonym of D. crispum Bak. The
long white hairs are absent, however. Baker describes the plant as 2-leaved but there
are about 4 on the type specimen.
Suspected Hybrids of D. glaucum Bak. X D. bakerianum Boi.
D. clarkeanum Schinz in Verh. Bot. Ver. Brand. 31: 218 (1890). Type; South West
Africa, Luderitz Bay, Angra Pequena, near the coast, Schinz (Z, holo.!, PRE, photo.).
D. juttae Engl. & Krause in Bot. Jahrb. 45: 150 (1910). Type: South West Africa,
Okahandja, riverine forest, Dinter 376 (B, holo.!, PRE, photo.).
D. bakerianum is a very characteristic species that can be easily recognized by its
knobbed ovary and capsule, the rosette of long tapered leaves, the swollen scape,
persistent bracts, short pedicels, equal perianth-segments, etc. D. glaucum has a set
of different but also characteristic features, for example, its large size, wide, glaucous
leaves, long pedicels, longer outer perianth-segments and appendaged anthers. A
number of specimens showed intermediate forms combining characters of each and
I therefore suspect them to be hybrids between these two species. Some of these were
described as new species, i.e. D. clarkeanum Schinz and D. juttae Engl. & Krause;
136
others were placed in either one or the other species. Dominant characters appear to
be the development of caudate appendages on the outer perianth-segments and the
connective being produced above the locules.
The knobs were absent from ovary and capsule; this seems to be a recessive
character. The size of the plants, leaf-shape and length of pedicel were intermediate
between those of D. glaucum and D. bakerianum.
Excluded Species
D. Dinter in Fedde, Rep. 19; 186 (1924). Type: South West Africa, Luderitz’
Aus, on quartzite flats, about 2 Km. N. of the railwayline, rare, Dinter 4344. Cult,
at Lichtenstein, Jan. 1923. (B, holo. ?).
D. avasimontanum Dinter in Fedde, Rep. 19: 186 (1924). Type: South West Africa,
Windhoek, Auas Mts., farm Lichtenstein, Dinter 4345 (B, holo.?).
The type specimens of both the above mentioned are sterile. They were placed
by Soelch in his thesis on the Liliiflorae of South West Africa, 1961 (Miinchen), under
ALBUCA NAMAQUENSIS Bak. In Fedde, Rep. Beih. 53: 74 (1928), Dinter himself
already transferred them to Albuca.
A CKNO WLEDGMENTS
For loan of material and kind co-operation, my sincere thanks are due to the
directors and staff of the following institutions : B, BOL, GRA, K, KMG, M, NBG
NH, SRGH, Z. In the citation of specimens, no herbarium is indicated if the specimen
is preserved at PRE.
I am much indebted to Mr. O. A. Leistner of Kimberley, who has given me much
information on the species growing around Kimberley, to Mr. W. Marais, liaison
officer at Kew, and Mr. J. Lewis of the Herbarium, British Museum (Natural History),
for their valuable assistance.
137
INDEX
I-
:[
I.
!
J
139
The South African Species of Lagarosiphon
by
A. A. Obermeyer
An examination of the southern African species of this genus revealed that each
had a typical leaf-structure, a useful character since plants are often sterile.
Very little was known of the life-history of this genus until Wager’s work appeared
in 1928.* Afterwards Dr. M. Ernst-Schwarzenbach also studied the flower-biology
of L. muscoides.f Wager observed 2 species common on the Highveld, L. muscoides
and L. major. As the other three South African species appear closely related, it is
likely that in general their behaviour will be similar.
Examining dried material, especially of such small dimensions, is not very satis-
factory. Living plants must be studied in the manner in which it was done by Wager,
to give us a more complete picture of their life-history. As so little was known about
these species when Wright made his compilation for the Flora of Tropical Africa, his
treatment of them is unsatisfactory. Some are no better than generic descriptions
and without having seen the types, it is impossible to recognize the plants concerned.
Lagarosiphon Harv. in Hook., Journ. Bot. Lond. 4: 230 (1841). Benth. & Hook,
f., Gen. Pi. iii: 450 (1883). Phill., Gen. ed. 2: 58 (1951).
Submerged, dioecious, rooted herbs from a perennial rhizome with long terete
leafy stems ; branches arising from inside a spathe next to a flower. Leaves widely spaced
below, dense above, alternate, subopposite or whorled; with 2 minute narrowly ovoid
intravaginal squamulae; margin toothed (each tooth unicellular); lamina with variation
in cell structure. Male injlorescence axillary; spathes 2 united bracts, obovate or
ovate to lanceolate, compressed or cup-shaped, dentate; the axis (rudimentary inflore-
scence) producing numerous pedicellate flowers (up to 50 were counted in L. major)
consecutively; the buds (in South African species) become detached and rise to the
surface before anthesis where they expand ; perianth in 2 whorls of 3 segments each
(the outer segments slightly narrower), recurving suddenly to form, a bell-shaped structure
that floats on the water; stamens 3, the filaments when expanded situated parallel to
the water-surface with the anthers at right angles to them; each pollen-sac containing
4 tetrads, grains large; staminodes 3, longer than the stamens, papillate and usually
coloured above, joined at the top, acting as a sail. Female inflorescence axillary;
spathes 2, fused, narrowly oblong, ovoid or cylindrical, entire or toothed, containing
1-2-3 flowers (when the capsule develops the inner spathe, which is usually larger,
splits centrally along the midrib); perianth-tube exserted laterally near the apex of
the spathes, lengthening and giving off gas-bubbles continuously inside the delicate
tube to buoy it up so that it will reach the surface (ultimate length of flower depends on
the species: in L. muscoides up to 25 cm, in L. major 15 cm; if it has not reached the
surface by then, it will die); limb 6-lobed; staminodes 3, small, filiform; ovary
1-locular with 3 parietal placentas; styles 3, adnate to the perianth-tube, each divided
above into 2 long, papillate, often brightly coloured stigmas; ovules 6-30, orthotropous,
funicle short or long, straight or bent. Capsule ovoid or cylindrical, attenuate at the
apex into a beak (the persistent base of the perianth), protruding from torn spathe-
valves; becoming mucilaginous and bursting irregularly and so dispersing the seeds;
* Wager, V. A. in Trans. Roy. Soc. S.A. 16: 191-204 (1928).
T Ernst-Schwarzenbach, M. in Ber. Schweiz. Bot. Ges. 5: 56-58 (1945).
140
seeds cylindrical, with a short stipe at the base, pointed at the apex, size varying with
each species, closely ribbed or honeycombed, buoyant at first, sinking later (viable only
after about 6 months in L. muscoides and L. major, unknown in other species).
About 16 species found in Africa, 2 endemic to Madagascar. L. major is reported
as an adventive in Europe and New Zealand.
Leaves thin, transparent, 0-5-2 mm broad, cells usually longer than broad:
Leaves with a hyaline margin consisting of at least 3 layers of cells without chlorophyll 1 . L. muscoides
Leaves green to the margin, teeth on small excrescences:
Leaves in verticils of 6-8; central band usually narrow and the cells not much different
from rest of lamina 2. L. verticillifolhis
Leaves alternate or sometimes in verticils of 2-5; central band usually wide, at least near
the base, with the cells much larger and with transverse septa 3. L. crispus
Leaves more solid, opaque, 2-3 mm broad, cells small, somewhat papillose-rhomboid:
Teeth of leaves minute, blunt, ascending 4. L. major
Teeth of leaves on triangular excrescences, sharp, patent 5. L. ilicifoliiis
1. L. muscoides Harv. in Hook. Journ. Bot. Lond. 4: 230, t. 22 (1841). Wright
in F.T.A. 7: 3 (1897); F.C. 5, 3: 1 (1912). Marloth, Flora of S.A. 4: 15, fig. 6 t. 2
(1915). Wager in Trans. Roy. Soc. S.A. 16: 191 (1928). Ernst-Schwarzenbach in
Ber. Schweiz. Bot. Ges. 55: 56 (1945). Syntypes: Cape, Albany, Zeyher $ (TCD);
without locality, Drege $ (TCD).
Hydrilla dregeana, Presl, Bot. Bemerk. 112 (1844). Type: Cape, Port Elizabeth
district, swamp near the mouth of the Swartkops River, Drege 2276c. H. muscoides
Planch, in Ann. Sci. Nat. Ser. 3, Bot. 11: 79 (1849). Haloragea Drege, Zwei Pfl.
Doc. 130 (1843). Fluvialea Drege, l.c. 136, 140.
Stems 0-5-2 mm in diam., usually weak. Leaves alternate, somewhat laxly spaced
on the stems, spreading, soft, thin, transparent, light green, linear, average size
10 by 1-5 mm, apex acute, margin with 3 rows of hyaline cells, teeth many, acicular,
pointing upwards, leaf-tip with 2 slightly larger teeth, cells of lamina elongate-oblong,
with 2-3 rows of cells on each side of midrib which are slightly ligher in colour, outer
row with thickened cell walls. Male inflorescence: spathe-valves flat, obovate, about
4 by 2 mm with about 23 teeth on each side ; with up to 40 flower-buds inside, flowers
white to pink. Female inflorescence: spathe-valves, ovoid, about 4-5 by 1-5 mm,
with about 25 long, acute teeth on each side; perianth white, maximum length up to
25 cm; ovary flask-shaped, stigmas purple, ovules about 12, funicles straight or bent
in narrow ovaries. Capsule narrow-ovoid to tubular, about 5-10 mm long; about
9 seeds in each capsule, breaking away from the placenta and pushed upwards one
behind the other, about 2-2-5 mm long, surface honeycombed.
Widespread in the summer rainfall areas of southern Africa, from Uitenhage in
the eastern Cape to Natal, Orange Free State, northern Cape, South West Africa,
Bechuanaland, Transvaal, Southern Rhodesia, extending to tropical east Africa.
Cape. — Port Elizabeth, Gamtoos River, Schlechter 6059. Albany: Collingham near
Grahamstown, Dickens. Hay: Papkuil sloot, Wilman 1247. Barkly West: in vlei
at Holpan, Acocks 2308.
Orange Free State. — Kroonstad: near convent, 602. Fauresmith: near Luck-
hoff, farm Klein Gryskop, 445, Smith 499.
Basutoland. — Banen, boggy water of stream to Pela Tseon River, Jacot-Guillarmod
2066. Leribe: Makokoane, Dieterlen 1031.
Natal. — Bergville: Rustenberg Road from Bergville, Edwards 2409.
141
Fig. 1. — Leaf margins of Lagarosiphon species showing spines and location of chloroplasts (dotted
lines). 1. L. muscoides Harv. (Burtt Davy, PRE 12792) Transvaal, Christiana. 2. L. verticillifoliiis
Oberm. (Sim 436) Mozambique, Lourenzo Marques. 3. L. crispus Rendle (Erens 351) Bechuana-
land, Tsotsoroga Pan about 150 m. North of Maun. 4. L. major (Ridley) Moss ex Wager (Cooper
17) Cape, Victoria East, Umdizine River. 5. //(cz/o/nw Oberm. (Galpin 7027) Southern Rhodesia,
Zambesi River above Victoria Falls.
142
Transvaal. — Christiana: Burtt Davy 12792. Potchefstroom: Witpoort, Louw 1708.
Benoni: Moss 11175. Pretoria: Kaallaagte 5 miles E. of Hammanskraal, Mauve
& Schlieben 9599. Belfast: pan 5 miles W. of Belfast, Obermeyer in TRV 36067.
Kruger National Park: Bobubu, van der Schijjf 5671. Pietersburg: Blauwberg,
Nacarene, Strey & Schlieben 8579. Soutpansberg: Duwenhage’s Pan, Schweickerdt
& Verdoorn 621 .
South West Africa. — ^Gibeon: Haribes, Volk 12405. Rehoboth: Bullspoort, Strey
2118; Rodin 2940. Okahandja: Bradfield 390. Grootfontein: Kumkaus, Kinges
2836; Grosshuis, in vleiwater, Schoenfelder S648 $; Dinter 7384 $.
Bechuanaland. — •Northern part, Mumpswe Pan, 25 miles N.N.W. of mouth of Nata
River, Drummond & Seagrief 5165; 5166.
Southern Rhodesia.— Salisbury, 6 Mile Spruit, Wild 3799; 3800; 3803 (SRGH);
Avondale, Eyles 5841 (SRGH). Wankie Game Reserve, 8 miles from Ngomo, Weir 9
(SRGH).
Tanganyika. — ^Mbulu: Mbulumbula, Greenway 6799 (SRGH).
The following specimens cannot be named satisfactorily: — ■
Cape. — ^Kornga: near Komga, Flanagan 1335.
Natal. — Dundee: Tim Worthington Dam, //a/mue;-. Estcourt: Tabamhlope Research
Station, Little Bushmans River, West 900. Newcastle: west of Laingsnek-Quaggasnek
Road, Edwards 2369.
Transvaal. — Standerton: Farm Beginsel, Duvy 2287. Potchefstroom: Frederik-
stad, Mooi River, Louw 1709.
In appearance these plants resemble an overgrown L. muscoides plant. Wager
noted that he found this large form usually together with L. muscoides in Natal. There
is only one fertile collection {Flanagan 1335). As L. muscoides and L. major are both
recorded from these regions, it was thought that they could be hybrids but Ernst-
Schwarzenbach, who crossed these two taxa, found them to be incompatible [Oesterr.
Bot. Zeitschr. 100: 420 (1953)].
2. L. verticillifolius Oberm., sp. nov.
Planta tenuis. Caulis ca 0-5-1 mm diam. Folia 5-7 verticillata, verticillis ad apicem
congestis; lamina pellucida patens linearis, margine chlorophyllacea dentibus multis
brevibus acicularibus e excrescentibus triangularibus ortis, cellulis centralibus ampli-
oribus parte media anguste formantis. Spatharum valvae, S', lanceolatae paucidentatae
apice acutae, $, ad basin ovoidea ad apicem tubulosae tantum pauci-dentatae ; ovarium
9-12 ovulis funiculo recto brevi. Capsula ovoidea attenuata. Semina 1-5 mm longa.
Type: Natal, Hlabisa, Hluhluwe Game Reserve, Ward 2551 $ (PRE, holotype).
Stems 0-5-1 mm in diam., fairly firm, densely leafy. Leaves in verticils of 7
(sometimes 5) occasionally sub-whorled, closely spaced on the stems (internodes up
to 2 cm below, 4-2 mm above), spreading, soft, thin, transparent, linear, average size
10 by 0-5 mm, but occasionally up to 20 mm; margin chlorophyllous, teeth many,
short, acicular, pointing out- and upwards, emerging from a triangular excrescence;
leaf-tip 1 -2-toothed similar to the marginal teeth, with a few rows of cells beside the
midrib that are slightly larger, lacunar, the outer row with thickened walls. Male
inflorescence with the spathe-valves lanceolate, convex, about 2-5 by 1 mm, pointed
at the apex, with a few small marginal teeth. Female inflorescence with the spathe-
valves ovoid below, tubular above, apex bifid, occasionally with a few marginal teeth;
perianth white; stigma-branches long, with large papillae; ovary with 9-12 ovules.
143
funicle straight, shorter than ovule. Capsule ovoid 5 by 2 mm, elongated at the apex;
seeds 1-5 mm long, narrowly grooved and with some protrusions (glandular?).
Recorded from Southern Rhodesia, eastern Transvaal, Zululand and Mozam-
bique.
Southern Rhodesia.— Nuanetsi, 30 miles S. of Chipinda Pools, Phipps 2909; Malangwe
River, S.W. of Mateka Hills, Drummond 5627; 5628.
Mozambique. — Lourengo Marques, Sim 436, Rogers 21378. Rikatla, Junod in TRV
25432.
Transvaal. — Pilgrim's Rest: Kruger National Park, 19V miles N.E. of Skukuza,
Codd 6069; Letaba: K.N.P., Shangoni, Mathlambanthlova Pan, van der Schijff' 2%51 .
Natal. — Hlabisa: Hluhluwe Game Reserve, Ward 2518; 2549 o'; 2551 2 (PRE,
holotype). Ubombo: Mkuzi Game Reserve, Tinley 443.
Easily distinguished by the many-leaved whorls.
3. L. crispus Rend/e in J. Linn. Soc. Bot. 30: 381, pi. 31, fig. 8-17(1895); F.T.A. 7:
4 (1897). Type: Tanganyika, between Uyui and coast, Taylor (BM, holo.).
L. tsotsorogensis Brem. & Oberm. in Ann. Tvl. Mus. 16: 401 (1935). Type: Beehuana-
land, Tsotsoroga Pan, van Son TRV 28853 (PRE, holo.).
Stems filiform, 1-2 mm in diam. Leaves alternate, sub-opposite to verticillate
above, usually densely spaced on the stems, spreading, soft, thin, transparent, linear,
average size 14 by 0 - 5 mm, marginal cells chlorophyllous, the many, sharp curved teeth
pointing upwards, situated on small, triangular excrescences; leaf-tips bi-dentate;
central area narrow to very broad, consisting of larger lacunar cells without or with
very little chlorophyll, usually with transverse septa, midrib distinct. Male inflorescence:
spathe-valves cup-shaped, obovoid, about 2 mm long, acute at first becoming broad
and gaping during anthesis, with about 8 teeth on each side. Female inflorescence:
spathes forming a narrow ovoid tube 2-5 mm long, entire or with a few small teeth,
apex entire at first, becoming bilobed; perianth white, limb about 2 mm in diam.;
ovary with 24-30 ovules, funicles about as long as the ovules; stigmas with short
papillae. Capsule ovoid, acuminate, about 3 mm long, many-seeded; seeds 1 mm
long, turbinate, with raised ridges which have transverse septa (in the dried state).
Recorded from East Africa, Southern Rhodesia, Bechuanaland, South West
Africa, eastern and northern Transvaal.
Kenya. — Northern Province, Sololo, Gilleti 13683.
Tanganyika. — Kivingo, Greenvay 1979.
Southern Rhodesia. — Urungwe, Zambesi Valley, Menswa Pan, Wild 4069; 17 miles
E.S.E. of Chirundu Bridge, Drummond 5404 (SRGH). Lower Sabie, Wiki 2318
(SRGH). Subungwe, Marawa road, Davies D1968 (SRGH).
Transvaal. — Soutpansberg: Limpopo River, farm Weipe, in pan, Codd & de Winter
348; Verdoorn 2273. Pilgrim’s Rest: Leeupan near Tshokwane, van der Schiiff 2166;
4198.
Bechuanaland. — Tsotsoroga Pan about 150 miles N. of Maun, Frens 351.
South West Aerica. — Caprivi Strip: Kabuta Village in Chobe River, Munro. Oka-
vango: Niangana, Okavango River, Dinter 7246. Grootfontein : West of Andara,
Merxmiiller 2096.
Charaeteristic of I . crispus Rendle are its small flowers (limb 1 -5-2 mm in diam.),
small, broadly ovoid capsules about 2-5 mm long and many seeds which are 1 mm long
144
(ovary with 24-30 ovules). In the South African specimens the leaves show the charac-
teristic lacunar broad central band with transverse septa at least in lower half. However,
some specimens from tropical Africa have the leaves longer (2 cm instead of 1) and the
band indistinct. Whether they belong to another species could not be decided here.
In van der Schij^ 4198, three setaceous appendages were noticed at the base of
the ovary similar to those seen in Schoenfelder 5648 (L. muscoides Harv.) from South
West Africa. They may be the rudiments of aborted flowers. Rendle noted that he
found no intrafoliar stipules. They were present in some specimens, but others appeared
to be without them; they may be early deciduous; such delicate organs easily disinte-
grate during pressing and drying.
Sometimes the plants are much oranched, foliose, with the leaves rather long
(2 cm). This could be its habit when, through external circumstances, natural sexual
reproduction is hampered. The species is apparently related to L. tenuis Rendle.
Fig. 2. — Lagarosiphon crispus Rendle. Part of leaf seen from above showing midrib (m) and
network of veinlets (v); ci, large central cells; co smaller outer cells (Erens 35 IV
145
4. L. major {Ridley) Moss apud Wager in Trans. Roy. Soc. S.A. 16: 193 (1928).
Mason in New Zeal. Journ. Sci. 30: 384 (1960). Wild in Kirkia 2: 39, t. 13, d (1961).
L. muscoides Harv. var. major Ridley in J. Linn. Soc. Bot. 22: 233 (1886).* Syntypes:
Cape, Victoria East, Umdizine River, Cooper 17 (K, PRE); Transvaal, Bronkhorst-
spruit, Rehmann 6559 (K); Bethal, Trigardtsfontein, Rehmann 6678 (K).
Elodea crispa hort. non Lagarosiphon crispus Rendle.
Stems about 3 mm in diam., usually firm. Leaves alternate, laxly spaced below,
dense above forming obtuse, broad apices, spreading-recurved, soft, usually fairly
firm, opaque, linear, average size 16 by 2 mm (occasionally up to 3 cm); apex obtuse
to acute to accuminate, margin with the outermost row of cells hyaline, producing
minute, blunt, closely-spaced teeth; leaf-tips with teeth similar to marginal; midrib
sunk above, prominent below; lamina with small papillose, rhomboid cells when
viewed from above, those of the midrib slightly longer and larger. Male inflorescence:
spathe-valves oblong-lanceolate, 5 by 3 mm, with about 8-9 teeth on each side, bearing
about 50 flowers consecutively, perianth pinkish. Female inflorescence: spathe-valves
ovoid, up to 5-5 by 2-5 mm, with about 12 teeth on one side and about 6 on the other;
perianth pink, maximum length 15 cm ; ovary flask-shaped, stigmas red, ovules about
12. Capsule ovoid, about 5 mm long; seeds about 2 mm long.
Widespread in southern Africa from the eastern Cape, Orange Free State, Transvaal
to Southern Rhodesia. Recorded as an adventive from Europe and New Zealand.
Cape. — King William’s Town: near King William’s Town, Sim 1561.
Natal. — Newcastle: W. of Laingsnek-Quaggasnek Rd., Edwards 2369.
Orange Free State. — BoshoflF: Smitskraal, Burtt Davy in PRE 10731.
Transvaal. — Piet Retief: Kleber Standerton: Schlechter 3464. Belfast: Wager
NH 49186 $; Franks in PRE 9770. Pretoria: Rietvlei Dam, Repton 2001. Germiston:
Birc Weigh, Wager NH 49187 $. Potchefstroom : Frederikstad, Louw 1709.
Southern Rhodesia. — Umtali, Tsungwesi River, Wild 4649. Salisbury, Hunyani
Poort Dam, Wild 4146. Matobo: Matopos Dam, Rattray $ in SRGH 82664.
Usually sterile. Often an obnoxious water pest, choking pans in the Republic
and Southern Rhodesia.
5. L. ilicifolius Oberm., sp. nov.
Planta robusta. Caulis ca 3 mm diam. albidus. Folia alterna interdum sub-
opposita lamine opaca reflexa lanceolata 8 X 2-5 mm margine dentibus patentibus
Ilici simili basi triangulari. Cellulae quadrangulare sparse papillosae. Spatharum
valvae anguste ovatae dentatae; ovarium ovoideum attenuatum 9-ovulatis. Capsula
ovoidea. Semen 2 mm longum.
Type: Bechuanaland, Bridge at Toteng on the N.E. tip of Lake Ngami, Story
Ain (PRE, holotype).
Stems about 3 mm in diam. (thinner in fertile plants), usually firm, whitish. Leaves
alternate, occasionally subopposite, regularly spaced below with the internodes about
3-6 mm long, densely leafy above forming broad obtuse apices to the shoots, reflexed
or recurved, fairly firm, opaque, lanceolate, average size 8 by 2-5 mm (smaller in
fertile plants) apex acute, margin with the outer row of cells hyaline, with the patent
strong teeth situated on broad triangular excrescences, closely spaced, leaf-tip with
2 teeth similar to marginal ones; all cells of the lamina small, similar, squared and
somewhat papillose viewed from above, midrib faint depressed, prominent-below,
purplish in dried specimens, its cells somewhat longer and narrower. Male inflorescence
unknown. Female inflorescence: spathe-valves narrow-ovate, 4 mm long, toothed or
* F. C. 5, 3: 1 (1912).
146
entire, acute, pale wine red; perianth pale lilac (Story); ovary ovoid, attenuate, ovules
about 9, funicle erect, short. Capsule ovoid, tapered towards the apex, 5-7 mm long;
seeds 2 mm.
Recorded from Northern South West Africa (Okavango River), northern Bechu-
analand (Lake Ngami), Southern and Northern Rhodesia (Zambesi River).
SotjTH West Africa. — Okavango Native Territory, lagoon in Okavango River at
Kapako Camp, 4 miles W. of Mupini Mission, de Winter & Marais 4516; Runtu,
bog and riverbed of Okavango River, MerxmuUer 1890; Niangana, Okavango River,
Dinter 7202a.
Bechuanaland. — Bridge at Toteng on the N.E. tip of Lake Ngami, Story 4727 (PRE,
holo.); McConnell in SRGH 68901; Maun, Thamalakane River, McConnell in SRGH
68903.
Southern Rhodesia. — Zambesi River just above Victoria Falls, Galvin 7027; Wilde
in TRV 9066; Young 1592; 5/m 19176; IL/W 31 1 1 (SRGH).
Northern Rhodesia. — Livingstone, Zambesi River, Greenway 6252. Mangu, Luan-
ginga River, about 4 miles N.W. of Sandaula Pontoon, Drummond & Cookson 6586.
Usually sterile in which case it is much coarser than the female plants collected
at lake Ngami.
Note. — The tropical African types of L. schweinfurthii Casp., L. steudneri Casp.,
L. cordofanus Casp. and L. fischeri Gurke were not found at Berlin: probably destroyed
in 1943'.
Acknowledgments
For loan of material I sincerely thank the Keepers of the Southern Rhodesian
Herbarium, Salisbury, the Bolus Herbarium, Cape Town, the Natal Herbarium, Durban
and that of the Eastern Province at the Albany Museum. From Dr. M. Ernst-
Schwarzenbach we gratefully received reprints of her work on this genus. I ast but not
least my sincere thanks to Mr. J. E. Dandy, Keeper of Botany, British Museum
(Natural History) for his helpful comments.
147
The South African Species of Anthericum,
Chlorophytum and Trachyandra
Addenda et Corrigenda
by
A. A. Obermeyer
Shortly after the above revision appeared in Bothalia 7: 669-768 (April, 1962),
the Director of the Berlin-Dahlem Herbarium informed me that several missing types
of Chlorophytum had recently been found amongst the late Prof. Dr. Werdermann’s
papers. The latter had recovered them from the ruins of Baron von Poellnitz’s house
after it had been bombed in November 1945 during the Second World War. The
sheets were sent on loan to PRE and as a result the following observations were made.
C. dregei Poelln. in Port. Acta Biol. 1: 228 (1945). Type: Cape Peninsula, Drege
(B, holo., PRE, photo.). The number 1827 on the sheet appears to be the year it was
collected or incorporated in the herbarium. It is without roots but is undoubtedly
a synonym of C. triflorum (Ait.) Kunth.
C. leipoldii Poelln. in Ber. Deutsche Bot. Ges. 61: 207 (1943y Type: Cape Peninsula,
Leibold (B, holo., PRE, photo.). Von Poellnitz named this plant after its collector
Leibold but misspelt the name leipoldii. I then wrongly assumed it to have been
collected by Dr. L. Leipoldt. It is a synonym of C. triflorum (Ait.) Kunth.
C. namaquense Schltr. ex Poelln. in Ber. Deutsche Bot. Ges. 61: 207 (1943).
Type: Cape, Namaqualand, Vogelklip, Schlechter 11295 (B, holo., PRE, photo.).
The isotype at PRE is similar to the holotype.
C. pulchellum Kunth, Enum. 4: 605 (1843); Bak. in El. Cap. 6: 398 (1897). No. 7
(p. 698) of revision. Type: Cape of Good Hope, Lalande (B, holo., PRE, photo.).
It is a synonym of C. rigidum Kunth (No. 12 of the revision). With this development
the neotype Galpin 8333 falls away. Eor the present this Galpin collection is regarded
as a xerophytic form of the widespread and variable C. comosuin (Thunb.) Jacques.
C. schlechterianwn Poelln. in Ber. Deutsche Bot. Ges. 61: 208 (1943). Type: Cape,
Schlechter s.n. ex hortus Berlin-Dahlem, pressed for herb. 1900 (B, holo., PRE, photo.).
It is a synonym of C. krookiamim Zahlbr.
C. wilmsii Engl. & Krause in Bot. Jahrb. 45: 138 (1910). Type: Transvaal, Lydenburg,
near the town, Wilms 1522 (B, holo., PRE, photo.). It is a synonym of C. bowkeri
Bak. as indicated in the revision, p. 700.
Corrigenda
On page 699 the synonyms Anthericum longituherosum Poelln. and A. vallistrappii
Poelln. were inadvertently quoted as Chlorophytum species,
p. 675, for Galpin 7652 read 7692.
p. 679, for Maguire 21966 read 21962.
p. 696, for Codd 2225 read 6225.
4846586—3
148
p. 697, after Marais add 843.
p. 725, for Sidey 1159 read 1359.
p. 737, T. revoluta (L.) Kunth should be No. 22 (not 23).
p. 744, for Marloth 2666 read 12666.
p. 746, for Marloth 1280 read 12801b.
p. 748, for MerxmuUer 1714 read 1719.
p. 744. The type locality of T. thyrsoidea (Bak.) Oberm. is in the Ceres district
not Tulbagh; cf. Trans. Roy. Soc. S.A. 1931: 274. Yuk River Hoogte is about 30
miles from the Whitehill Karoo Garden.
149
The South African Species of Orthosiphon
by
L. E. Codd
When M. Ashby reviewed the African species of Orthosiphon Benth. in Journ.
Bot. Lond. 76; 1 (1938) he did not deal with the genus Nautochilns Brem., described
in Ann. Transv. Mus. 15: 253 (1933). Later, in reply to an enquiry from the National
Herbarium, the Keeper of Botany of the British Museum, Natural History, quotes
him as saying: “As to the value of this genus, I had not sufficient material to draw
any definite conclusions. . . . There seems some doubt as to whether it should
not be included with Orthosiphon, there being an obvious affinity with the O. serratus
group ”. Phillips in his Gen. S. Afr. Flow. PI. ed. 2: 653 (1951) reduces Nautochilns
to synonymy under Orthosiphon without comment. Although still more material of
certain species is desirable, the evidence now available supports the reduction of
Nautochilns.
Bremekamp typified his genus with N. labiatus (N. E. Br.) Brem. and described
three additional species, namely, N. urticaefolius Brem., N. breyeri Brem. and N.
amabilis Brem. The first two of these are now regarded as being forms of N. labiatus.
The last mentioned has smaller leaves and flowers and, as shown below, forms a link
with Orthosiphon pseudoserratus M. Ashby.
In evaluating Nautochilns, it must be compared with typical Orthosiphon and
with Orthosiphon Sect. Serrati M. Ashby. The latter Section consists of three species,
O. serratus Schltr., O. pseudoserratus M. Ashby and O. tubiformis R. Good, and appears
to be restricted to southern Africa. A character that seems to be of importance in
separating Sect. Serrati from typical Orthosiphon is the place of attachment of the
upper pair of stamens. In typical Orthosiphon, as represented by O. australis Vatke,
the two upper stamens are attached about 1-2 mm from the throat and the filaments
are filiform and glabrous. In Sect. Serrati, the two upper filaments are attached near
the base of the corolla tube, while the filaments are slightly swollen and pubescent in
the lower part. Another character which may have significance is the shape of the
disc. In the few typical species of Orthosiphon examined, the disc is flat except for the
development of an anterior lobe equal to the height of the ovary. In Sect. Serrati,
the disc is annular, surrounding the ovary, and reaching to the lower half or the
top of the ovary. Nautochilus is similar to Sect. Serrati in both these characters.
The superficial floral characters of Nautochilus labiatus which strike one as being
unusual are the well-exserted stamens, the declinate corolla tube and the large upper
and lower lips of the corolla. Taking the stamen character first, in typical Orthosiphon
the anthers are scarcely exserted from the throat of the corolla tube and usually lie
in the lower lip or in the throat of the corolla; this situation is also found in two species
of Sect. Serrati, O. serratus and O. tubiformis but, in O. pseudoserratus, the stamens
are exserted as in Nautochilus. Regarding the corolla tube and lobes, in N. amabilis
the corolla tube is slightly declinate and the corolla lobes are only slightly longer than
in O. pseudoserratus. In O. pseudoserratus, the corolla tube is almost straight, with
only a slight bend near the base. Thus in corolla characters there are intermediates
linking Nautochilus with Orthosiphon.
It therefore seems clear that Nautochilus Brem. cannot be upheld in its present
form. A more natural demarcation from Orthosiphon Benth. would be one based
on (a) the attachment of the upper pair of stamens being near the base of the corolla
150
tube, not near the throat; {b) the filaments being slightly swollen and pubescent in
the lower part; and (c) the disc being annular, not flat with an enlarged anterior lobe,
as in Orthosiphon.
On this basis. Sect. Serrati Ashby would be included in Nautochilus Brem. Genera
in the Labiatae have been erected on even more slender grounds than those listed
above. No advantage can be seen in such a treatment and it is preferred to place
Nautochilus as a subgenus of Orthosiphon.
Orthosiphon Benth. subgen. Nautochilus {Brem.) L. E. Codd, subgen. nov.
Nautochilus Brem. in Ann. Transv. Mus. 15: 253 (1933).
Orthosiphon Benth. sect. Serrati Ashby in J. Bot. Lond. 76: 1 (1938).
Upper pair of stamens attached about 2 mm from the throat of the corolla tube; bracts up to
4 mm long; disc flat except for an enlarged anterior lobe (Subgen. Ortho siphon):
Rhachis pubescent but lacking stipitate glands; leaves glabrous to pubescent, not conspicuously
dotted with red sessile glands below nor appressed canescent:
Petioles 4-20 mm long; leaves ovate to broadly ovate, obtuse to truncate at the base 1. australis
Petioles 0-2 mm long; leaves ovate to ovate-lanceolate, cuneate 2. rubicundus
Rhachis with numerous short stipitate glands; leaves either dotted with red sessile glands below
or appressed canescent:
Leaves 2-3 cm long, scabrid pubescent with numerous red sessile gland-dots below; stems
simple, less than 50 cm long 3. vernalis
Leaves 8-12 mm long, densely appressed canescent especially on the lower surface; stems
muwh branched, usually exceeding 50 cm 4. fruticosus
Upper pair of stamens attached near the base of the corolla tube; bracts usually exceeding 4 mm
long (occasionally less in O. pseudoserratus and O. amabitis)-, disc annular (Subgen. Nuuto-
chiliis) :
Leaves lanceolate or elliptic to obovate, if broadly ovate then margin distinctly serrate; lower
lip of corolla less than 8 mm long;
Corolla tube exceeding 2 cm in length; leaves usually less than 2 cm long and 1 cm broad
5. tubiformis
Corolla tube less than 2 cm in length ; leaves usually exceeding 2 cm long and 1 cm broad :
Leaves usually exceeding 4 cm in length and 2 cm in breadth; stamens exceeding the throat
of the corolla by 4-6 mm 6. serratus
Leaves usually less than 4 cm in length and 2 cm in breadth; stamens exceeding the throat
of the corolla by 7-8 mm 7. pseudoserratus
Leaves broadly ovate to subrotund, distinctly petiolate; margin crenate; lower lip of corolla
more than 8 mm long:
Leaves less than 2-5 cm long and 2 cm broad 8. amabiUs
Leaves more than 2-5 cm long and 2 cm broad 9. labiatus
1. O. australis Vatke in Linnaea 40: 179 (1876); Linnaea 43: 86 (1881-82);
Bak. in FI. Trop. Afr. 5: 373 (1900); Ashby in Journ. Bot. Lond. 76: 40 (1938).
Type: Portuguese East Africa, Rios de Sena. Peters.
O. glabratus var. africanus Benth. in DC., Prodr. 12: 51 (1848). Type: Magaliesberg,
Burke 162 (K, holo., PRE!). O. wihnsii Guerke in Bot. Jahrb. 26: 81 (1898); N. E.
Br. in FI. Cap. 5, 1: 255 (1910); Ashby in Journ. Bot. Lond. 76: 44 (1938). Type:
near Lydenburg, Wilms 1115 (K, BM). var. komghensis N. E. Br., l.c. 256
(1910). Type: Komgha, Flanagan All (K, holo., PRE!). O. neglectus Briq. in Bull.
Herb. Boiss. Ser. 2, 3: 988 (1903). Type: Pretoria, Wonderboompoort, Rehmann
4510. O. inconcinnus Briq., l.c. 991 (1903); N. E. Br., l.c. 256 (1910). Type: Natal,
Camperdown, Wood 4963 (K, NH).
Plectranthus bolusii T. Cooke in Kew Bull. 1909: 377 (1909), partly, as to Bolus 1101 1.
Syntypes: Potgietersrus, Bolus 11011 (BOL!); Houtbosch, Rehmann 6167.
151
Perennial herb, often branching from the base; stems semi-woody, decumbent or
ascending, simple or sparingly branched, 12-60 cm long or even longer, subglabrous
to pubescent with simple and multicellular hairs, rarely densely pubescent. Leaves
opposite, petiolate; lamina ovate to broadly ovate, 1-5-5 cm long, 1-3 cm broad,
subglabrous to pubescent, especially on the nerves below, with simple and multi-
cellular hairs and scattered to numerous sessile pale brownish gland-dots; apex acute
to obtuse, base truncate to abruptly and shortly cuneate; margin coarsely and distantly
serrate to subentire; nerves fairly distinct below, indistinct above; petiole 0-5-2 -5
cm long, slender, pubescent. Inflorescence simple, 4-15 cm long; verticils 1-2 cm
apart; rhachis densely pubescent with simple, often crisped, and long multicellular
hairs; bracts subrotund, abruptly acuminate, 2-5-3 mm long, pubescent below, glabrous
above, persistent, subtending 2-3 flowers; pedicel 2-3 mm long, pubescent. Calyx
campanulate, 5-6 mm long at flowering, enlarging to 7-9 mm in fruit, pubescent with
simple and multicellular hairs intermingled with pale yellowish sessile gland-dots;
upper tooth erect, broadly ovate, rounded to apiculate, markedly decurrent; lower
4 teeth triangular-subulate, the lower two often distinctly longer, especially in fruit,
the lateral two broad-based, usually with a distinct shoulder above. Corolla white
to mauve, pubescent without, bilabiate; tube 7 mm long, slightly compressed laterally,
straight, 1-5 mm deep at the base, 2-2-5 mm deep at the throat; upper lip erect to
slightly reflexed, 3-3-5 mm long and equally broad, 3-lobed, the central lobe the
largest, distinctly emarginate ; lower lip horizontal, concave, 4 - 5-5 - 5 mm long. Stamens
didynamous, exserted from the throat by 2 mm; lower two filaments attached in the
throat, glabrous; upper two filaments attached 1-5-2 mm from the throat, glabrous;
anthers 1-thecous. Disc flat or crenate with a distinct anterior lobe exceeding the
ovary. Ovary 4-lobed; style filiform, 8-9 mm long; stigma thickened, minutely
bifid.
Distributed from Sudan and Ethiopia through east tropical Africa as far south
as Komga in the Cape Province. It inhabits dry, wooded country and is often common,
especially on overgrazed veld. Flowering extends, in South Africa, from November
to April or even later.
Transvaal. — Marico: Zeerust, Leenrfo-rz in TRV 11524; A1487; Lekkerlach,
Lomw 628. Rustenburg; 15 miles N. of Rustenburg, Co<r/r/ 9216. Pretoria; Magalies-
berg, Burke 162; Onderstepoort, Theiler in TRV 12363; Swing Bridge, Repton 3328.
Potgietersrus : Bolus 11011 (BOL); Leendertz 1439; Tliode A 1761; 5 miles E. of
Potgietersrus, Codd 8974; 8 miles E. of Potgietersrus, Comins 903. Groblersdal:
Loskop Dam, Codd 8436. Lydenburg: Krugers Post, Bunt Davy 7276. Letaba;
3+ miles W. of Letaba Camp, Codd & Dyer 4695. Pilgrim’s Rest: Kruger National
Park, 5 miles N.E. of Skukuza, Codd & de Winter 5059. Nelspruit: Kruger National
Park, 10 miles E. of Skukuza, Codd & de Winter 5038; 17 miles N.E. of Pretorius
Kop, Codd 5204; Numbi, van der Schijff 125; 1597; Pretorius Kop, van der Schijfl
1205; 2040; Sigaas, van der Schijff 3222. Barberton: Komatipoort, S'c/z/ec/z/cr 1 1755;
Kaapmuiden, Rogers 25044; 8 miles N. of Barberton, Codd 8154.
Swaziland. — -Stegi: Malinda, Compton 30798. Ellatikulu: near Gollel, Compton
30798.
Natal.— Hlabisa; 10 miles N.W. of Mtubatuba, Codd 9621 ; Hluhluwe Game Reserve,
Ward 2215, 2276; Tinley 588; 590. Weenen: Muden Valley, Repton 1195; 1 mile
W. of Muden, Co(7<7 8604. Camperdown: Medley Wood \\912; Marais S33; Botha’s
Hill, McClean 137. Ixopo: 20 miles N. of Highflats, Codd 9370.
Cape. — Komga: Flanagan All.
The present concept of O. australis is based on Ashby’s work, with the addition
of O. wilmsii and its synonyms. It is probable that O. dissimilis N. E. Br. should also
be included and very likely other tropical African species as well. Typical O. australis
152
occurs in the eastern Transvaal lowveld. O. wilmsii is the inland form subjected to
colder conditions and periodic burning. It branches more freely from the base and
has shorter stems, smaller and thicker-textured leaves with the gland-dots more densely
placed in the somewhat wrinkled under surface. There are, however, many inter-
mediates linking it with O. australis, and thus O. wilmsii could, at most, be regarded
as a variety of O. australis. It is not known if the type of O. australis (a Peters specimen,
in Berlin) still exists.
Regarding Plectranthus bolusii T. Cooke, it was noted by N. E. Brown in FI.
Cap. 5, 1: 256, 282 (1910) that Bolus 11011 belongs in Ortliosiphon wilmsii Guerke
(now included in O. australis Vatke), and this is confirmed by an examination of the
specimen in Bolus Herbarium. The other syntype, Rehmann 6167, and the specimen
of Wood 4488 cited by Cooke as probably a robust specimen of P. bolusii, are both
Orthosiphon labiatus N. E. Br.
2. O. rubicundiis (Z). Don) Benth. in Wall., PI. As. Rar. 2: 14 (1831); Lab. 26
(1832); Ashby in Journ. Bot. Lond. 76: 41 (1938). Type: from India.
Perennial herb, often branching from the base; stems semi-woody, erect or decum-
bent, simple or sparingly branched, 25-60 cm long, glabrous to sparingly pubescent
with scattered multicellular hairs. Leaves opposite, subsessile to sessile; lamina ovate
to ovate-lanceolate, 3 • 5-8 cm long and 1 • 8-4 cm broad, glabrous to sparingly pubescent
on both surfaces or, sometimes, only on the nerves below, with scattered sessile yellowish
gland-dots; apex acute to obtuse, base obtuse to cuneate; margin distinctly and
somewhat distantly serrate to obscurely crenate, often entire in the lower third; nerves
distinct below, indistinct above; petiole 0-2 mm long. Inflorescence simple or
occasionally with a pair of branches at the base, 5-20 cm long, congested or with
verticils up to 1 - 5 cm apart; rhachis densely pubescent with long multicellular hairs;
bracts broadly ovate, abruptly acuminate, 2-5-3 mm long, pubescent below, glabrous
above, subpersistent, subtending 2-3 flowers; pedicel 2 mm long, pubescent. Calyx
campanulate, 4-5 mm long at flowering, enlarging to 8-9 mm long in fruit, hispidulous
with simple and multicellular hairs and scattered sessile pale yellowish gland-dots;
upper tooth erect, broadly ovate, apiculate, decurrent; lower 4 teeth triangular-subulate,
subequal at flowering stage, the two lateral broader-based and shorter in fruit. Corolla
white to mauve, pubescent without, bilabiate; tube 6 mm long, slightly compressed
laterally, straight, 1 -5 mm deep at the base, 2 mm deep at the throat; upper lip erect,
3-5 mm long, 3 mm broad, 3-lobed, the central lobe the largest, emarginate; lower
lip horizontal, concave, 4 mm long. Stamens didynamous, exserted from the throat
by 2 mm; lower two filaments attached in the throat, glabrous; upper two filaments
attached 1-1-5 mm from the throat, glabrous; anthers 1-thecous. Disc flat with an
anterior lobe equal to the ovary. Orary 4-lobed ; style filiform, 8-9 mm long; stigma
thickened, minutely bifid.
Recorded from China, India and tropieal Africa, reaching its southernmost limit
in northern Transvaal, where it grows in grassy places in fairly dense woodland and
flowers from October to January.
Transvaal. — Soutpansberg: Elim, Obermeyer in TRV 29238; Tshakoma, Obermeyer
in TRV 31570.
The synonymy given by Ashby, l.c., has not been investigated and his identification
of the above two specimens, together with other material from tropical Africa, is
accepted.
3. O. vernalis L. E. Codd, sp. nov., O. rubicundo Benth. afflnis sed floribus solitariis,
foliis calyce corolla rubro-punctatis dilfert.
153
Herba perennis; caules 1-3 e basi orti, simplices erecti subquadrangulares 20-30
cm alti pubescenti. Folia opposita ovata vel ovato-lanceolata vel elliptica 2-3 cm
longa 1-1 -5 cm lata scabrido-pubescentia infra rubro-punctata, margine involuto
obscure crenato-dentato, apice obtuso, basi obtusa vel truncata, petiolo 1-2 mm longo.
Inflorescentia simplex 8-15 cm longa; rhachis glanduloso-pubescens rubro-punctata;
bracteae lanceolatae acuminatae 3-4 mm longae persistentes; flores solitarii, pedicellis
3-5 mm longis. Calyx 5-dentatus campanulatus bilabiatus scabrido-glandulosus
rubro-punctatus; dens superior late ovatus acutus vel obtusus suberectus decurrens;
dentes inferiores 4 subaequales triangulo-subulati. Corolla bilabiata purpurea pubescens
rubro-punctata; tubus cylindratus 7-8 mm longus ad faucem 2-5 mm latus; labium
superiorius erectum 5-6 mm longum 2-5 mm latum 4-lobatum; labium inferiorius
horizontale concavum 6-7 mm longum. Stamina didynama, 2 mm exserta; antherae
1-thecae. Discus lobo anteriore. Ovarium 4-Iobatum 0-5 mm longum, stylo filiformi
8-9 mm longo, stigmate minute bilobato. Nucellae suborbiculares 2 mm longae.
Swaziland.— Manzini District: Malkerns, 11th October, 1961, FOns 60/43 (PRE,
holotype); 26th October, 1960, FOns 60/43. Mankaiana District: Evelyn Baring
Bridge over Usutu River, Compton 29167; 31098.
Perennial herb 20-30 cm tall; stems 1-3 arising annually from a woody rootstock,
erect, simple or sparingly branched, subquadrangular, pubescent with simple hairs
and long multicellular hairs intermingled with sessile red gland-dots. Leaves opposite,
ovate to ovate-lanceolate or elliptic, 2-3 cm long and 1-1 • 5 cm broad, scabrid-pubescent
on both surfaces especially on the nerves below, pubescence of short simple hairs and
long multicellular hairs with sessile red gland-dots densely placed below, sparsely
scattered on the upper surface; margin slightly thickened, obscurely crenate-dentate
to almost entire; apex obtuse to rounded, base rounded to truncate; petiole 1-2 mm
long. Inflorescence simple, 8-15 cm long; rhachis glandular-pubescent with long
multicellular hairs scattered among densely placed shorter gland-tipped hairs and
sessile red gland-dots; verticils 3-10 mm apart; bracts lanceolate, acuminate, 3-4
mm long, pubescent and gland-dotted below, subglabrous above, persistent, each
subtending a solitary flower; pedicels 3-5 mm long. Calyx campanulate, bilabiate,
purple, scabrid-pubescent with multicellular hairs, short gland-tipped hairs and sessile
red gland-dots, 5-6 mm long at flowering stage, enlarging to 9-11 mm long in fruit;
upper tooth broadly ovate, acute to obtuse, decurrent on the tube, suberect or horizontal,
3-5-4 mm long and 2-5 mm broad at fruiting stage; lower 4 teeth subequal, triangular-
subulate, the two lateral broader-based and shorter than the two lower. Corolla
bilabiate, purple, pubescent with white appressed crisped hairs intermingled with
numerous sessile red gland-dots without, minutely puberulous within; tube cylindrical,
slightly compressed laterally, straight, 7-8 mm long, 2-5 mm broad at the throat;
upper lip erect, 5-6 mm long, 4-5 mm broad, 4-lobed; lower lip horizontal, concave,
6-7 mm long. Stamens didynamous, exserted by 2 mm from the throat; the two
lower attached in the throat, glabrous; the two upper attached 1-5 mm from the
throat, curving upward, glabrous; anthers 1-thecous, 2-lobed. Disc flat with an
enlarged anterior lobe exceeding the ovary. Ovary 4-lobed, 0-5 mm long; style filiform,
8-9 mm long; stigma minutely bilobed. Nutlets 2 or 3 developing, suborbicular,
slightly compressed, 2 mm long. Plate 1.
Recorded so far from a restricted area in Swaziland. It grows on grassy slopes
where it is apparently subjected to regular burning, the slender stems arising annually
from the base and producing their flowers in spring. Its nearest affinity appears to be
O. ruhicundus Benth., from which it is readily distinguished by the solitary flowers
in the axils of the bracts, the densely stipitate-glandular rhachis and the numerous
sessile, red gland-dots on the undersides of the leaves and on the calyx and corolla.
4. O. fruiticosus L. E. Code!, sp. nov. distincta, in subgenus Orthosiphone, habitu
fruticoso, foliis parvis subintegris canescentibus.
154
Frutex virgatus 60-90 cm altus, caulibus subquadrangularibus glabrescentibus.
Folia opposita lanceolata elliptica vel oblanceolata 8-12 mm longa 2-5-5 mm lata
dense canescentia, apice acuta, basi cuneata, petiole 1-1 • 5 mm longo, margine involute
integro vel subintegro. Inflorescentia simplex 4-9 cm longa; rhachis dense glanduloso-
puberula; bracteae ovatae vel lanceolatae 2-3 mm longae glanduloso-puberulae persis-
tentes, flores 1-3 subtentae; pedicelli 3-5 mm longi. Calyx 5-dentatus campanulatus
bilabiatus glanduloso-puberulus; dens superior late ovatus acuminatus suberectus
decurrens; dentes inferiores 4 subaequales triangulo-subulati. Corolla bilabiata
purpurea pubescens glanduloso-punctata; tubus cylindratus 7-8 mm longus ad faucem
2-5 mm latus; labium superiorius erectum 5-6 mm longum 3-4 mm latum 4-lobatum;
labium inferiorius horizontale concavum 6 mm longum. Stamina didynama, 2 mm
exserta; antherae 1-thecae. Discus lobo anteriore. Ovarium 4-Iobatum 0-5 mm
longum, stylo filiform! 8-9 mm longo, stigmate minute bilobato. Nucellae suborbi-
culares 2 mm longae.
Transvaal. — Lydenburg District: 6^ miles S.W. of Steelpoort Station, Codd 9777
(PRE, holotype); 74 miles N.W. of Maartenshoop, Codd 8797; Steelpoort River
Valley, Pole Evans 4695; 22 miles N. of Schoonoord, Acocks 20952.
Twiggy small shrub 60-90 cm tall; branches subquandrangular and appressed
grey tomentulose when young, becoming terete and glabrescent with age; bark on
old stems greyish, often splitting off in thin strips. Leaves opposite or fasciculate on
short shoots, lanceolate to oblanceolate, 8-12 mm long, 2-5-5 mm broad, densely
canescent especially on the lower surface, pubescence of dense, short, appressed grey
hairs; apex acute, base cuneate into a short petiole 1-1-5 mm long; margin involute,
entire or subentire, rarely obscurely dentate. Inflorescence simple, 4-9 cm long; rhachis
densely glandular-puberulous; verticils 1-1-5 cm apart; bracts ovate to lanceolate,
2- 3 mm long, glandular-puberulous, persistent, subtending 1-3 flowers; pedicels 3-5
mm long. Calyx campanulate, bilabiate, glandular-puberulous, 4-5 mm long at
flowering stage, enlarging to 7-8 mm long in fruit; upper tooth the largest, purple-
tinged, broadly ovate, acuminate, suberect, decurrent, 2-5-3 mm long at fruiting stage;
lower 4 teeth subequal, triangular-subulate, the two lateral broader-based than the
two lower. Corolla bilabiate, purple, sparingly pubescent with scattered orange sessile
gland-dots without, glabrous within; tube cylindrical, slightly compressed laterally,
straight, 7-8 mm long, 2-5 mm broad at the throat; upper lip erect, 5-6 mm long,
3- 4 mm broad, 4-lobed; lower lip horizontal, concave, 6 mm long. Stamens didy-
namous, exserted by 2 mm from the throat; the two lower attached in the throat,
glabrous; the two upper attached 1 - 5 mm from the throat, glabrous; anthers 1-thecous.
Disc flat with an anterior lobe exceeding the ovary. Ovary 4-lobed, 0-5 mm long;
style filiform, 8-9 mm long; stigma minutely bilobed. Nutlets usually 3 developing,
suborbicular, somewhat compressed, 2 mm long. Plate 2.
Recorded so far only from Sekukuniland in the Lydenburg District, where it grows
in dry bushveld, often on stony slopes. It flowers in autumn from February to April.
The shape of the disc and the point of attachment of the upper pair of stamens places
it in the subgenus Orthosiphon, in which subgenus it appears to have no close relative.
Its shrubby habit and small leaves are reminiscent of O. tubiformis in subgenus Nauto-
chilus, but its corolla is much shorter and the leaves are appressed grey pubescent, not
scabrid-pubescent.
5. O. tubiformis R. Good in Journ. Bot. Lond. 63: 173 (1925); Ashby in Journ.
Bot. Lond. 76: 10(1938); Letty, Wild Flows. Tvaal. 285, t. 142 (1962). Type: Pilgrim’s
Rest, Vaalhoek, Rogers 25104 (BM, holo., PRE!).
Virgate shrub 30-70 cm tall; stems erect, branched, subglabrous to hispid with
multicellular hairs. Leaves opposite, shortly petiolate, lanceolate-elliptic to obovate,
1 -4-2 cm long, 7-10 mm broad, usually scabrid pubescent on both surfaces, rarely
155
subglabrous above or appressed pubescent on both surfaces, pubescence of multi-
cellular hairs and simple, often bulbous-based hairs; apex acute to obtuse, base cuneate;
margin slightly thickened, serrate to subentire; nerves strongly reticulate below,
indistinct above; petiole 2-4 mm long. Inflorescence simple, 5-17 cm long; rhachis
shortly glandular-hispidulous; bracts ovate, acuminate, 5-8 mm long, glandular-
puberulous below, glabrous above, persistent, subtending 1-3 flowers; pedicels 4-6
mm long, densely glandular-puberulous. Calyx campanulate, purple-tinged, 8-10
mm long at flowering, enlarging to 1 - 6 cm long in fruit, densely glandular-puberulous
with short gland-tipped hairs; upper tooth subrotund, acute to abruptly apiculate,
decurrent; lower 4 teeth subequal, triangular-subulate, the two lower usually slightly
longer and narrower, the two lateral broader-based. Corolla whitish to pale mauve,
sparingly pubescent without, bilabiate; tube long-cylindrical, 2-3-5 cm long, straight,
somewhat compressed laterally, 1 -5 mm deep near the base, 2 mm deep at the throat;
upper lip erect, 7 mm long, 5 mm broad, 3-lobed, the central lobe the largest, subentire;
lower lip horizontal, concave, 6 mm long. Stamens didynamous, exceeding the corolla
tube by 5-6 mm; lower two filaments attached in the throat, glabrous, 5-6 mm long;
upper two filaments attached near the base of the tube, pubescent below, slightly
exceeding the corolla tube; anthers 1-thecous. Disc annular, obscurely crenate,
almost reaching the apex of the ovary. Ovr/ry 4-lobed; style filiform; stigma exserted
by 6 mm, shortly bilobed, lobes spreading.
Inhabits wooded, stony slopes in relatively dry parts at medium altitudes in the
Lydenburg, Pilgrim’s Rest and Letaba Districts. Easily recognized by the shrubby
habit combined with the exceptionally long, narrowly cylindrical corolla tube.
Transvaal. — Letaba: The Downs, Junod 4341. Pilgrim’s Rest: Vaalhoek, Rogers
25014. Lydenburg: Pole Evans 4687; 1+ miles W. of Driehoek, Codd & Dyer 7696;
5 miles S.E. of Steelpoort P.O., Codd <& Dyer 7714; 4 miles N. of Mogaba P.O., Codd
10022.
6. O. serratus Scliltr. in Journ. Bot. Lond. 35: 431 (1897); Ashby in Journ. Bot.
Lond. 76: 9 (1938). Type: Barberton, Galpin 499 (K, holo., PRE!).
Shrub 30-90 cm tall often branching from the base; stems erect, sparingly branched
or simple, densely hispid with simple and multicellular hairs, sometimes also with
gland-tipped hairs. Leaves opposite or ternate, shortly petiolate, broadly ovate to
obovate-elliptic, 4-9 cm long, 2-3-5 cm broad, densely appressed pubescent to sub-
glabrous on both surfaces, pubescence mainly of long multicellular hairs with scattered
sessile, often somewhat sunken gland-dots, rarely with gland-tipped hairs also present;
apex acute to obtuse, base cuneate to obtuse; margin distinctly and regularly serrate;
nerves reticulate to subobscure below, indistinct above; petiole 3-8 mm long. Inflore-
scence simple, 8-32 cm long; verticils 1-3 cm apart; rhachis glandular-hispidulous;
bracts ovate, long-acuminate, 6-10 (rarely-16) mm long, glandular-hispidulous below,
glabrous above, persistent, subtending 2-6 flowers; pedicels 3-4 mm long, glandular-
hispidulous. Calyx campanulate, purple, 7-8 mm long at flowering, enlarging to 1 - 5
cm long in fruit, glandular-hispidulous with scattered multicellular hairs, short gland-
tipped hairs and sessile, yellowish gland-dots; upper tooth subrotund, abruptly acute,
decurrent; lower 4 teeth subequal, triangular-subulate, the two lateral broader based.
Corolla mauve to purple, glandular-puberulous without, bilabiate; tube (6-) 9-16 mm
long, compressed laterally, straight, 1-5 mm deep at the base, enlarging to 2-5 mm
at the throat; upper lip erect, 6-7 mm long, 3 mm broad, obscurely 3-lobed, the
central lobe the largest, subentire; lower lip horizontal, concave, 5-6 mm long. Stamens
didynamous, exceeding the corolla tube by 4—5 mm; lower two filaments attached
in the throat, glabrous, 5 mm long; upper two filaments attached near the base of the
tube, pubescent in the lower part, 1-1 to 1-2 cm long; anthers 1-thecous. Disc annular,
crenate, reaching the apex of the ovary. Ovm-v 4-lobed; style filiform; stigma exserted
by 5 mm, bilobed, lobes spreading, 0-5 mm long.
156
Found in dense grass on stony hillsides, where it is usually subjected to regular
burning, from eastern Transvaal and Swaziland to northern Zululand. A striking
species when in flower with its numerous fairly large purple flowers. It varies a good
deal in length of corolla tube, size and pubescence of leaves.
Transvaal. — Lydenburg: Kantoorbos Forest Station, Codd 9786. Belfast: Eland-
spruitberg, Scldechter 3866; Elandshoek (Belfast District ?), Rogers in TRV 4736;
Schoemanskloof, ■Swwts' <5: 1056; 2146; 2158; 2200; vc»« IFo/^ in TRV 34869;
34870; Waterval Boven, Rogers 14304; Pole Evans 2620; L. Britton 4810; Mogg
in PRE 15101; Waterval Onder, Galpin 13806; Rogers in TRV 4737; Airlie, Prosser
1254. Nelspruit: Liebenberg 2679; Noe! 71. Barberton: Galpin 499; Williamson
118; Rogers mA\-, Williams in TKW 1656 ■, 10105; C. A. Smith 1Q56\ Clarke
52; Rimers Creek, Williamson 247; Lomati Falls, de Beer in TRV 4939.
Swaziland. — Mbabane, Usutu Canal, Malkerns, Compton 27103; Mbabane, Komati
Bridge, Compton 26932; Hlatikulu, Compton 26309; Pierce 60; 11 miles N.E. of Hluti,
Acocks 15333; 9 miles N. of Mankaiana, Codd 4121 Komati Bridge, Prosser 1954;
Bremersdorp, Leach & Bayliss 10618; without locality. Wells 2032; Stewart in TRV
12795.
Natal. — Hlabisa: 44 miles S. of Hlabisa, Acocks 13064; 2\ miles N. of Hlabisa,
Codd 9614; Hluhluwe Game Reserve, Ward 1681.
7. O. pseudoserratus M. Ashby in Journ. Bot. Lond. 76: 8 (1938). Type: “ Water-
berg District, Moorddrift ”, Leendertz 2243 (BM, holo., PRE!).
Shrublet 30 cm or more tall; stems erect, branched or, apparently, sometimes
simple, hispid-glandular with simple and multicellular, usually gland-tipped hairs.
Leaves opposite, petiolate to subsessile, ovate to broadly elliptic, 2-4 cm long, 1 • 2-2
cm broad, densely glandular-pubescent on both surfaces, pubescence of simple, multi-
cellular and short gland-tipped hairs intermingled with sessile yellowish gland-dots;
apex obtuse, base rounded; margin serrate; nerves reticulate to subobscure below,
indistinct above; petiole 2-5 mm long. Inflorescence simple, 3-15 cm long; rhachis
glandular-hispidulous; verticils 0-7-2 cm apart; bracts ovate-lanceolate, acuminate,
3-6 mm long, glandular-pubescent below, subglabrous above, persistent, subtending
1-3 flowers; pedicels 3-4 mm long, glandular-hispid. Calyx campanulate, 5 mm
long at flowering, enlarging to 1-1 cm long in fruit, glandular-hispidulous with multi-
cellular hairs and short gland-tipped hairs intermingled with sessile, yellowish gland-dots;
upper tooth broadly ovate, abruptly acuminate, decurrent; lower 4 teeth subequal,
triangular-subulate, the two lateral broader based. Corolla white to mauve or pink,
bilabiate, sparsely pubescent without, bilabiate; tube 5-7 mm long, laterally compressed,
slightly declinate or almost straight, 1 mm deep at the base enlarging to 3 mm deep
at the throat; upper lip erect, 3-4 mm long and 2-3 mm broad, 3-lobed, the central
lobe the largest, emarginate at the apex; lower lip horizontal, occasionally reflexed
later, boat-shaped, 5-6 mm long. Stamens didynamous, exserted from the throat
by 7-8 mm; lower two filaments attached in the throat, glabrous, 7-8 mm long;
upper two filaments attached near the base of the tube, slightly thickened and pubescent
below, 1-1-1 -2 cm long; anthers 1-thecous. Disc annular, crenate, almost equal to
the ovary or shorter. Ovary 4-lobed; style filiform, 1 cm long; stigma not thickened,
minutely bilobed.
Recorded so far only from the Potgietersrus District, where it grows on rocky,
wooded slopes at medium altitudes. Differs from O. serratus in the smaller, less serrate
leaves and the more conspicuously exserted stamens. In corolla characters it resembles
O. amabilis and differs from this species in leaf shape and shorter petioles.
Transvaal. — Potgietersrus: Thode A 1762; Leendertz 1494; Moorddrift, Leendertz
2243; Pyramid Estate, Galpin 9065; 9154; Naboomfontein, Galpin 13455; between
Potgietersrus and Zebediela, Pole Evans 3094 (13).
157
8. O. amabilis (Brem.) L. E. Codd, comb. nov.
Nautochilus amabilis Brem. in Ann. Transv. Mus. 15: 254 (1933). Type: Potgietersrus,
Swerwerskraal, Bremekamp in TRV 1220 (PRE, holo.!).
Twiggy shrub 60-90 cm tall; branches erect or spreading-ascending, pubescent,
with numerous multicellular bulbous-based hairs interspersed with yellowish sessile
gland-dots. Leaves opposite, petiolate; lamina broadly ovate to subrotund, 1-1-5
(rarely -2) cm long and 1-1 -4 (rarely-1 -8) cm broad, pubescent on both surfaces, lower
surface greyish with long interwoven multicellular hairs intermingled with yellow
sessile gland-dots, upper surface dark green with shorter and less dense pubescence;
apex rounded, base truncate; margin regularly and finely crenate; nerves indistinct;
petiole 5-12 mm long. Inflorescence simple, 7-14 cm long; verticils 1-1-5 cm apart;
rhachis densely glandular-hispidulous with short gland-tipped hairs and scattered
multicellular hairs and yellow sessile gland-dots; bracts ovate to subrotund, acuminate,
4-6 mm long, glandular-pubescent on both surfaces or on the lower only, subpersistent,
subtending 1-3 flowers; pedicel 3-5 mm long, glandular-hispidulous. Calyx cam-
panulate, purple-tinged, 4-5 mm long at flowering, enlarging to 9 mm long in fruit,
glandular-hispidulous with short gland-tipped hairs, scattered multicellular hairs and
numerous yellow sessile gland-dots; upper tooth more or less horizontal, sut rotund,
apiculate, markedly decurrent; lower 4 teeth triangular-subulate, subequal at flc waring
stage, the two lateral becoming broader-based and shorter than the two lower at
fruiting stage. Corolla mauve or pink, sparsely puberulous without, bilabiate; tube
8-9 mm long, laterally compressed, slightly declinate, 1 mm deep at the base enlarging
to 3 mm deep at the throat; upper lip at first erect, later recurving, 5 mm long and
2-5 mm broad, 3-lobed, the central lobe the largest, emarginate at the apex; lower
lip at first horizontal, later reflexed, boat-shaped, 8-9 mm long. Stamens didynamous,
exserted from the throat by 9 mm ; lower two filaments attached in the throat, glabrous,
9 mm long; upper two filaments attached near the base of the tube, slightly thickened
and pubescent near the base, 1-3 cm long; anthers 1-thecous. Disc annular, slightly
shorter than the ovary. Ovary 4-lobed; style filiform, 1-4 cm long; stigma not
thickened, entire.
A small, much-branched shrub, recorded from dry, wooded places and stream
banks in the Potgietersrus and Lydenburg Districts. Closely related to O. labiatus
but has smaller leaves and flowers. May be separated from O. pseudoserratus by the
more rotund leaf blade, the long slender petioles and the longer corolla lobes.
Transvaal. — Potgietersrus: Swerwerskraal, Bremekamp in TRV 1220. Lydenburg:
Sekukuniland, Barnard 339; 421; near Branddraai, Meeuse 10020.
The specimen Meeuse 10020 differs from the others in having much less pubescent
stems and leaves. The relatively small leaves and flowers match O. amabilis, but there
is also a possibility that it is a depauperate specimen of O. labiatus.
9. O. labiatus N. E. Br. in FI. Cap. 5, 1 : 245 (1910). Type: Transvaal, Woodbush
Mts., Schlechter 4434 (K, holo., PRE!).
Nautochilus labiatus (N. E. Br.) Brem. in Ann. Transv. Mus. 15: 253 (1933); Verdoorn
in Flow. PI. S. Afr. 23: t. 901 (1943); Letty, Wild Flows. Tvaal. 288, t. 143, 1 (1962).
N. breyeri Brem., l.c. 254 (1933). Type: Louis Trichardt, Breyer in TRV 19400 (PRE,
holo. !). N. urticaefolia Brem., 1 .c. 254 (1933). Type : Blaauwberg, Leipzig, Bremekamp
& Schweickerdt 131 (PRE, holo.!).
Plectranthus bolusii sensu FI. Cap. 5, 1 : 282 (1910), partly, as to Rehmann 6167 and
Wood 4488.
Soft shrub 0-6-1 -8 m tall, branching from the base; stems ascending, freely
branched, sparingly pubescent with long multicellular hairs, denser at the nodes, and
usually scattered yellowish sessile gland-dots. Leaves opposite, petiolate; lamina
158
broadly ovate to subrotund, soft, 3-8 cm long and 2-6 cm broad, sparingly pubescent
above, denser below with long multicellular hairs freely intermingled with sessile pale
yellow gland-dots; apex acute to rounded, base truncate to abruptly and shortly
cuneate; margin regularly and coarsely crenate; nerves fairly distinct below, indistinct
above ; petiole slender, 0 ■ 5-3 cm long. Inflorescence simple or, occasionally, with a
pair of short branches at the base, 5-18 cm long; verticils 1-2 cm apart; rhachis
glandular-pubescent with dense short gland-tipped hairs, scattered multicellular hairs
and yellowish sessile gland-dots; bracts ovate, acuminate, 8-10 mm long, purplish,
glandular-pubescent below, almost glabrous above, persistent, subtending 2-4 flowers;
pedicel 5-6 mm long, glandular-puberulous. Calyx campanulate, purple, 5 mm long
at flowering, enlarging to 1-1 • 5 cm long in fruit, glandular puberulous with short gland-
tipped hairs, numerous yellowish sessile gland-dots and occasional multicellular hairs;
upper tooth suberect, ovate-orbicular, abruptly apiculate or rounded, decurrent;
lower 4 teeth triangular-subulate, subequal at flowering, the two lateral broader-based
and shorter at fruiting stage. Corolla pale mauve to pink, sparsely pubescent without,
bilabiate; tube 1-1-2 cm long, declinate, laterally compressed, 1-5 mm deep at the
base, expanding to 4 mm deep at the throat; upper lip at first erect, later recurving, 7-8
mm long, 3-3-5 mm broad, 3-lobed, the central lobe the largest, subentire at the apex;
lower lip at first horizontal, boat-shaped, later recurving below the tube, 8-12 mm long.
Stamens didynamous, exserted beyond the throat by 9-12 mm, the longer (lower)
usually just exceeding the lower lip; lower two filaments attached in the throat,
glabrous, 10-12 mm long; upper two filaments attached near the base of the tube,
slightly thickened and pubescent below, 1-4-1 -6 cm long; anthers 1-thecous. Disc
annular, slightly shorter than the ovary. Ovary 4-lobed; style filiform, 1-1-1 -4 cm
long; stigma minutely bifid.
A distinctive species with its bushy habit, large leaves on long slender petioles
and its large corolla with declinate tube and long, usually reflexed, lips. Found on
dry, rocky, wooded hillsides and wooded watercourses from northern and eastern
Transvaal, through Swaziland to northern Natal. Widely cultivated for its attractive
flowers and long flowering season which, in nature, extends from January to May
and, in cultivation, is even longer.
Transvaal. — Soutpansberg: Smuts s.n.; Louis Trichardt, Breyer in TRV 19400;
Wylliespoort, Schlieben 9264; Ihlenfeldt 2188; near Hangklip, Meeuse 10153; 11
miles W. of Louis Trichardt, Prosser 2005; Sand River Bridge, Meeuse 10205; Lejuma,
Meeuse 10314. Pietersburg: Blaauwberg, Smuts & Pole Evans 905; Bremekamp &
Schweickerdt 131; Codd & Dyer 9138; Houtboschberg, Schlechter 4434; Rehmann
6167 (K). Letaba: Shilouvane, Junod 2400; 4333; Gravelotte, van Dam in TRV
25003. Potgietersrus: Riebeek West, McDonald 208. Rustenburg; Kransberg,
Codd 3740. Belfast: Schoemanskloof, Smuts & Gillett 2203. Lydenburg: 6 miles
S.W. of Penge Mine, Codd & Dyer 7736. Pilgrim’s Rest: Erasmus Pass, Strey 3226.
Nelspruit: Kruger National Park, 9 miles N.W. of Pretorius Kop, Codd 5146; Numbi,
van der Schiflf 3446; Ship Mt., Acocks 16700; near Malelane, Codd 5254; van der
Sclujfl 4209; Brynard & Pienaar 4435. Barberton: Thorncroft in TRV 3260; near
Joe’s Luck Siding, Codd 9534.
Swaziland. — Mankaiana, Compton 27511; Lebombos, Hornby 2857.
Natal. — Ubombo: Pole Evans 2659. Mapumulo: Oqaqeni, Edwards 1815. Krans-
kop: 5 miles S. of Jameson’s Drift, Codd 9669. Weenen: Medley Wood 4488 (NH);
Acocks 10151; 34 miles W. of Weenen, Codd 8616; Nkasine, Edwards 714.
The type of Nautocliilus breyeri Brem. has smaller leaves with more finely crenate
margins than in typical O. labiatus while in N. urticaefolia Brem. the leaves are more
softly tomentose on the lower surface. These are considered to be forms of O. labiatus,
which is a fairly variable species but, nevertheless, an easily recognizable entity.
159
Little Known Species
In his review of the African species of Orthosiphon, Ashby places the following
among the doubtful or excluded species: O. ambiguus Bolus, O. holusii N. E. Br. and
O. varians N. E. Br. The holotypes of these species were kindly sent to me on loan
by the Curator of the Bolus Herbarium, while additional material of certain species
was forthcoming from the Compton Herbarium and the Albany Museum Herbarium.
Two of the species are now transferred to other genera, while the third is placed in
synonymy. In addition, the tropical species, Plectranthus allenii C. H. Wr., is trans-
ferred to Orthosiphon.
Plectranthus ambiguus (Bolus) L. E. Cocld, comb. nov.
Orthosiphon ambiguus Bolus in Journ. Linn. Soc. 18: 394 (1881). Type: near
Grahamstown, MacOwan 987 (BOL, holo.!, SAM!).
Plectranthus coloratus E. Mey. ex Benth. in E. Mey., Comm. PI. 228 (1837); Cooke
in FI. Cap. 5, 1: 279 (1910); non Don (1825). P. dregei L. E. Codd in Flow. PI. Afr.
32: t. 1244 (1957).
Although MacOwan 987 is in an immature state, it is clearly recognizable as the
same species as Plectranthus coloratus E. Mey. ex Benth., non Don, which was renamed
P. dregei L. E. Codd.
Hemizygia bolusii (TV. E. Br.) L. E. Codd, comb. nov.
Orthosiphon bolusii N. E. Br. in FI. Cap. 5, 1: 258 (1910). Type: Natal, Cathkin
Peak, A. Bolus in Herb. Guthrie 4894 (BOL!).
This appears to be a distinct species of Hemizygia and, although the type was
collected on 6/11/1897, no further material has been seen in South African Herbaria.
It comes nearest to H. foliosa S. Moore, but has smaller, more pubescent leaves, simple
racemes, 4-6-flowered verticils and the bracts are shed at a very early stage. In leaf
character it resembles Syncolostemon macranthus (Guerke) M. Ashby, but differs
again in the simple racemes, the more widely spaced verticils and the presence of
villous hairs on the stem and rhachis. Although N. E. Brown reports that the lower
pair of stamens may have free or united filaments, it is felt that this point requires
further investigation, for which additional gatherings of the species are required. In
floral characters it agrees with Hemizygia.
Hemizygia petrensis (Hiern) M. Ashby in Journ. Bot. Lond. 73: 353 (1935).
Orthosiphon petrensis Hiern, Cat. Afr. PI. Welw. 1: 859 (1898). O. varians N. E. Br.
in FI. Cap. 5, 1: 256 (1910). O. holubii N. E. Br., l.c. 258 (1910). O. engleri Perkins
in Bot. Jahrb. 54: 344 (1917). O. mossianus Good in Journ. Bot. Lond. 63: 174
(1925).
Hemizygia mossianus (Good) M. Ashby, l.c. 356 (1935).
Ashby, l.c., already placed O. holubii and O. engleri as synonyms of Hemizygia
petrensis. I have not seen the types of these species but, after examining the types of
O. varians and O. mossianus, I have no hesitation in adding them to Ashby’s concept
of H. petrensis. In O. varians the leaves are ovate-lanceolate; by including this species
in O. petrensis, the leaf shapes now accepted for O. petrensis vary from narrowly
lanceolate to ovate-lanceolate. Ashby separates O. mossianus on the basis of its petiolate
leaves, but this character breaks down and individual specimens may even be found in
which the upper leaves are sessile and the lower ones petiolate. The specimens now
included in the enlarged concept of O. petrensis show some variation in the density
of pubescence, but are characterized by the presence of villous hairs, as agam:t the
short, dense, appressed tomentum of the closely related species, H. canescens (Guerke)
Ashby. There are a few specimens with intermediate pubescence but, as the great
majority can be classified on the basis of their pubescence, it is considered that both
H. canescens and H. petrensis may be upheld as distinct though closely related species.
160
Orthosiphon allenii (C. H. Wr.) L. E. Codd, comb. nov.
Plectranthus allenii C. H. Wr. in Kew Bull. 1908: 437 (1908), as “ alleni ”. Type:
Rhodesia, Lukanda River, C. E. E. Allen 471.
Ashby placed Plectranthus allenii C. H. Wr. as a synonym of O. rubicundus (D,
Don) Benth. I am grateful to Mr. W. Marais, our Liaison Officer at Kew, for reporting
on the type of P. allenii and confirming that it is distinct from O. rubicundus. It is a
characteristic species with stems arising annually from a perennial rootstock with
tuberous roots. The leaves are relatively large, 8-20 cm long and 3-8 cm broad, obovate,
tapering towards the petiole, and are situated in a basal rosette or on a short stem
up to 4 cm long. The habit thus resembles certain species of Gerbera. The inflorescence
is usually unbranched and is borne on a slender peduncle 8-18 cm long, while the flowers
have all the characteristics of a true Orthosiphon. The species has been recorded from
Northern and Southern Rhodesia, Nyasaland, the Congo and Mozambique but not,
as yet, from South Africa.
161
Uvl'TJMf.NT iF iJtia
tC:..' !‘Ai' ' VK
./rr
v",‘ro' ii-io-ui V, ■jO/4i
Plate 1. — Orthosiphon vernalis L. E. Codd (FOns 60/43, holotype
4846586
162
Plate 2. — Ortliosiplion fruticosus I-. E. Codd {Codd 9777, holotype).
163
Notes and New Records of African Plants
by
Various Authors
AMARYLLIDACEAE
Cyrtanthus eucallus R. A. Dyer, sp. nov., C. sanguineo (Lindl.) Walp. affinis, sed
perianthio parviore tubo et filamentis brevioribus dilfert.
Bulbus 2 -5-3 -5 cm diam., in collum ad 10 cm longum contractus. Folia 2-4, ad 50
cm longa, supra medium 2-3 cm lata, coaetanea, dorsaliter carinata. Scapus c. 30 cm
altus, 1-6-florus; bracteae 4-6 cm longae, 1 cm latae; pedicelli 1 -5-4 cm longi. Corolla
rubra, 4 -5-5 -5 cm longa, vix ad infra medium divisa; tubus 2-2-5 cm longus, supra
infundibuliformis; lobi 2-5-3 cm longi, patentes, 1-2-1 -3 cm lati. Stamina 2-seriata,
filamentis 4-5 mm longis.
Transvaal. — Barberton: Aurora Mine, in wooded kloof in moist loose soil, Ising &
McNeil 31a in PRE 29282, holotype; Louws Creek, Hall 2 in PRE 29283.
Bulb 2 -5-3 -5 cm diam. contracted into a neck up to about 10 cm long, covered
by a dark brown tunic. Leaves 2, sometimes 3 or 4 to a bulb, contemporary with
the inflorescence, up to 50 cm long and 2-3 cm broad above the middle, tapering to
base and apex, dark green, somewhat glossy, prominently keeled down back. Scape
solid towards base and with narrow cavity above, up to about 30 cm tall, 5 mm diam.,
1-6-flowered; bracts 2, boat-shaped 4-6 cm long, suberect or spreading. Perianth
red, lighter within throat, 4 -5-5 -5 cm long; tube 2-2-5 cm long, slender for 1 cm
and then gradually expanded to about 2 cm at mouth; lobes elliptic-oblong, 2-5-3 cm
long, 1-1-3 cm broad, spreading. Stamens in 2 series, inserted within the throat of the
tube, those opposite the outer lobes inserted lower in the tube and with slightly shorter
filaments; filaments 4 and 5 mm long respectively; anthers dorsifixed and divided
from the base to the point of attachment.
Some specimens of Cyrtanthus sanguineus (Lindl.) Walp., in the National Her-
barium, Pretoria, from the eastern Cape show a strong likeness to our plant in the
luxuriant leaves and 3-4-flowered scape. The flowers are, however, larger, with a
longer tube and the filaments are longer with the anthers exserted from the throat in
C. sanguineus. These differences appear to justify species-status for C. eucallus, which
has not been recorded from outside the Barberton district.
The first specimen submitted to the National Herbarium for identification was
cultivated and brought by Mr. J. P. Hall of Johannesburg in 1960 and it was only after
further collections were made by Dr. G. Ising of Sweden and Mr. G. McNeil of
Ofcalaco, that it was decided to publish the new name and description. Mr. E. T. E.
(Tom) Andrews of Barberton, who has known of the plant for many years, must be
commended for his successful efforts to protect the natural habitat of the species from
vandalism. The name chosen for the species refers to its natural beauty, a feature of
most of the other species of Cyrtanthus.
Dr. Ising has for some years carried out cytological and genetical research on
the genus Cyrtanthus and his journeys in South Africa in 1962-63 with Mr. McNeil
were made with the specific object of recording and collecting more material in the
wild state.
R. A. DYER
164
ASCLEPIADACEAE
Ceropegia estelleana R. A. Dyer, sp. nov., C . finibriatae E. Mey. affinis, sed floribus
majoribus, corallae lobis supra medium margine pilis clavatis vibratilis pendulis indutis
differt.
Herba perennis tenuis glabra sparse ramosa tortulosa radicibus fusiformibus.
Folia minima lineari-lanceolata usque 5 mm longa, 1 mm lata. Pedunculus 1-2-florus,
pedicellis circiter 1 cm longis. Sepala lineari-lanceolata 5-6 mm longa. Corolla
curvata glabra 5-7 cm longa ; tubus 4 ■ 5-5 • 5 cm longus, basi 1 cm diam. ovoideo-inflatus,
intus papillosus deinde 3 mm diam. superne sensim in fauce 1-5 cm diam. ampliatus;
lobi unguiculati 4 mm lati, 4-5 mm longi, superne in umbraculum 1 ■ 3-1 • 5 cm diam.
connati marginibus pilis clavatis vibratilis pendulis ciliatis. Coronae exterioris lobi
3 mm longi ad basin versus anguste bilobulati; coronae interioris lobi incumbenti-
erecti 3 mm longi.
Cape. — Albany: Queens Pass in Fish River Valley, karroid scrub, Bayliss 1114, cult.
PRE, holotype.
Perennial succulent herb with fusiform roots. Stem slender, up to about 3 ft.
tall, somewhat twining and moderately branched. Leaves rudimentary, linear-lanceolate,
up to 5 mm long and 1 mm broad. Flowers 1-2 from each rudimentary peduncle,
developed successively on pedicels about 1 cm long. Calyx divided to base with sepals
linear-lanceolate, 5-6 mm long. Corolla glabrous, about 6 cm long with light green
inflated base 1 • 25-1 ■ 3 cm long, 1 cm diam. ; tube green at first turning white with
green veins, 3 mm diam. at base, gradually expanded to 1 - 5 cm diam., slightly incurved
at mouth and with a minute lobe in the sinuses between the main lobes; lobes narrow
in basal 4 mm with incurved margin, expanded in upper portion with long pendulous
clavate-vibratile hairs, then united for apical 7 mm into a parachute-like canopy 1 • 7 cm
diam. Corona purple arising about 1 mm above base of staminal column; outer
lobes about 3 mm long, erect, divided almost to base into two slender lobes with only
the tips slightly diverging; inner lobes about 3 mm long, incumbent-erect and diverging
above.
This species is named after Mrs. Estelle Bayliss, an ardent and successful collector
of Ceropegia plants, who kindly supplied field notes to the following effect: The
plants were collected in xerophytic bush off the road from Grahamstown to Fort
Beaufort in the Fish River Valley at an altitude of 2,400 ft. The area is rocky with
outcrops of sandstone and shale and supports a predominately succulent or semi-
succulent flora in which species of Euphorbia and Aloe are conspicuous. Three other
species of Ceropegia have been collected in the same area; C. ampliata E. Mey., C.
carnosa E. Mey. and C. stapeliiformis Haw. It seems likely that the present species
has remained undetected for so long because of its close resemblance to Cynanchum
tetrapterum (Turcz) R. A. Dyer (=C. sarcostemmatoides K. Schum.) with which it
was growing on the occassion of its discovery.
R. A. Dyer.
CAMPANULACEAE
Wahlenbergia monotropa Killick, sp. nov. distincta, nulla aflinitate manifesta
obvia.
Herba erecta, 4-12 cm alta, simplex vel basi ramosa, caule scabrido. Filia alternata;
superiora oblanceolata, 5-8 mm longa, 0-5 mm lata, integra vel interdum l-dentata;
inferiora obovata, 4-5 mm longa, 1-3 mm lata, integra vel l-dentata. Flores terminales,
solitarii, subsessiles, caerulescentes. Calycis tubus obconicus, 1-25 mm longus; lobi
subulati, 2-5 mm longi, marginibus puberulis. Corolla subcampanulata, tubo 2-5
165
mm longo; lobi tubo aequilongi, acuto-rotundati, apice nonnihil cucullati. Stamina 5;
filamenta basi lanceolato-oblonga pilosa, 1-5 mm longa; antherae 1 mm longae.
Ovarium 3-locuIare, ovulis numerosis. Stigma 3-fidum, lobis linearibus recurvatis.
Basutoland. — Drakensberg Area: in alpine grassveld between Indumeni Dome and
Cleft Peak, 9,800 feet, Killick 2334 (PRE, holotype).
Erect herb, 4-12 cm high, simple or branched at the base, stems scabrid. Leaves
alternate; upper oblanceolate, 5-8 mm long, 0-5 mm wide, entire or sometimes
1- toothed; lower obovate, 4-5 mm long, 1-3 mm wide, entire or 1-toothed. Flowers
terminal, solitary, subsessile, pale blue. Calyx tube obconical, 1-25 mm long; lobes
2- 5 mm long, margin puberulous. Corolla subcampanulate; tube 2-5 mm long,
lobes equalling the tube in length, acute-round, apex somewhat cucullate. Stamens 5;
filaments lanceolate-oblong at the base, 1-5 mm long; anthers 1 mm long. Ovary
3- locular, with numerous ovules. Stigma 3-fid, lobes linear recurved.
This is yet another new species from the summit of the Drakensberg between
Indumeni Dome and Cleft Peak: since 1954 ten new species from this area have been
described in Bothalia and elsewhere. Wahlenbergia monotropa is an inconspicuous
constituent of alpine grassland at 9,800 feet and flowers in February.
The affinites of W. monotropa are not obvious and it is difficult to assign this species
to any of von Brehmer’s sections in his treatment of Wahlenbergia in Engl. Bot. Jahrb.
53 (1915). The epithet monotropa refers to the apparently isolated taxonomic position
of the species.
D. J. B. Killick
CAPPARIDACEAE
A Note on Capparis transvaalensis Schinz
Recently a few specimens of Capparis were sent from the National Herbarium
for checking. In order to name them I had to get some type specimens on loan from
Berlin. After comparing the material in Kew Herbarium with the types, I came to
the conclusion that six of the species recognized by Gilg and Benedict in their mono-
graph of the family were conspecific. There seemed to be three recognizable varieties
apart from some intermediate specimens. Dr. Codd, after studying the material in
Pretoria, came to the conclusion that only two varieties are worth upholding, one of
them rather variable.
Capparis transvaalensis Schinz in Vjschr. Naturf. Ges. Zurich 57: 556 (1912).
Type: Transvaal Mailaskop, Schlechter 4510 (B, iso.!).
var. transvaalensis.
Branches shortly grey-tomentose; leaves elliptic-lanceolate to ovate, mostly
dark-green; sepals 5-7 mm long; stamens 18-21.
Recorded from the northern and eastern Transvaal, Swaziland, and Portuguese
East Africa.
var. calvescens (Gilg & Bened.) Marais, stat. nov.
CopparA ca/ve’i'ce/25 Gilg & Bened. in Bot. Jahrb. 53 : 195 (1915). Type: Natal, Tugela,
Wood 8472 (B, holo.l). C. schlechteri Schinz, op. cit.: 555 (1912). Type: Cape,
Tsitsa River, Schlechter 6385 (B, iso.!). C. flanaganii Gilg & Bened., op. cit.: 197
(1915). Type: Cape, Komga, Flanagan 809 (B, holo.!). C. rudatisii Gilg & Bened.,
op. cit.: 198 (1915). Types: Natal, Port Shepstone, Friedenau, Rudatis 1388 (K,
iso.!); Weenen, Wood 4438 (K, iso.!).
166
Branches soon becoming glabrous, often yellow-green; leaves narrowly elliptic
to broadly ovate, often drying yellow-green; sepals 4-5 mm long; stamens 5-14.
Recorded from the eastern Cape, Transkei and Natal.
From northern Natal, Zululand and Swaziland there are some specimens which
agree well with var. transvaalensis except for the number of stamens which varies from
8-15.
C. solanoides Gilg & Bened., op. cit. : 197 (1915), based on a Medley Wood
specimen from Little Noodsberg, belongs to this species, but the holotype is too poor
to assign to either variety.
The specimens from Portuguese East Africa, which are given as C. rudatisii in
Flora Zambesiaca, have more than 20 stamens per flower, not c.8 as stated in the
description, and belong to var. transvaalensis.
W. Marais
CRUCIFERAE
Notes on the Genus Heliophila
My study of the genus Heliophila at the Kew Herbarium has shown the necessity
for several nomenclatural changes and the description of a new species. I wish to
thank the Director of the Royal Botanic Gardens, Kew, for the facilities afforded me
to carry out the work and I thank also Directors of other institutions who have sent
material on loan.
Heliophila minima {Stephens) Marais, comb. nov.
Cleome minima Stephens in Ann. S. Afr. Mus. 9: 35 (1921).
H. pearsonii Schulz in Bot. Arch. 31: 528 (1931). H. pearsonii var. prageri Schulz,
l.c. H. pearsonii var. edentata Hainz in Mitt. Bot. Staatss. Miinchen 2, 11 : 39 (1954).
H. smithii Schulz in Notizbl. Bot. Gart. Berk 11: 227 (1931).
Heliophila scoparia Burch, ex DC., Syst. 2: 693 (1821).
var. scoparia.
Stems and leaves, though ridged or wrinkled when dry, mostly quite smooth;
flowers described as purple, pale lilac, pale mauve or pink; ovules 24-36. Flowering
period mainly from June to July.
var. aspera (Scliltr.) Marais, stat. nov.
Heliophila aspera Schltr. in Bot. Jahrb. 49: 410 (1913).
Stems and leaves muricately striate or papillate-denticulate ; flowers white, produced
from November to February; ovules 10-16. Intermediate specimens between the two
varieties do occur.
Heliophila seselifolia Burch, ex DC., op. cit.: 684.
De Candolle cites two specimens, Burchell 1318 and 1391. His diagnosis agrees
completely with No. 1318. Burchell 1391 is H. crithmifolia Willd. I take the specimen
of Burchell 1318 in the Kew Herbarium as lectotype.
var. seselifolia.
Petals each with a small appendage at the base, and each of the two short filaments
with a large appendage.
167
var. nigellifolia (Schltr.) Marais, stat. nov.
H. nigellifolia Schltr. in Bot. Jahrb. 49; 416 (1913).
Petals without appendages or with a small appendage each; filaments always
without appendages; valves of fruit undulate and more or less inflated over the seeds.
var. marlothii {Schulz) Marais, stat. nov.
Heliophila marlothii Schulz in Bot. Arch. 31; 533 (1931).
Schulz cites two specimens, Mar loth 9904 and 10405. The flowers of Mar loth
10405 have petals with small appendages, and short filaments with large appendages.
This does not agree with the description, and I choose the specimen of Marloth 19904
in the National Herbarium, Pretoria (from Marloth’s own herbarium) as lectotype.
This variety is distinguished in having both petals and filaments without appendages,
and in having broad fruits and large seeds.
Heliophila cornuta Sond. in Abh. Ges. Naturw. Hamb. 1; 246, t. 28 (1846).
Sonder based this species on Leptormus longifolius Eckl. & Zeyh., Enum. No. 59,
excl. syn., and on Heliophila scoparia (c) herb. Drege. The Kew herbarium specimen
of Zeyher (L. longifolius E. & Z.) is from Sonder’s own herbarium; it is in good
condition and I choose it as lectotype.
var. cornuta.
Sepals conspicuously horned; ovules (14-) 18-24; flowering period from July
to August.
var. squamata {Schltr.) Marais, stat. nov.
H. squamata Schltr. in Bot. Jahrb. 27; 139 (1899).
Specimens mostly quite glabrous but some, including the type, with puberulous
pedicels and sepals; sepals thickened at the tip, but not conspicuously horned; ovules
(18-) 28-38; flowering period mainly from September to November.
In the past there has been a considerable amount of misapplication of names.
A few of the cases which have been clarified are given below.
Heliophila pinnata L.f, Suppl. ; 297 (1781). Type; Thunberg in Herb. linn. No.
840-11!
H. trifida Thunb., Prodr. PI. Cap.; 108 (1800), nom. illeg., non auct. H. exilis Schltr.
in Bot. Jahrb. 27; 133 (1899). Type; Schlechter 8636 (Bl). H. oreophila Schltr. in
Bot. Jahrb. 49; 416 (1913). Type; Schlechter 10951 (Bl). H. clavuligera Schulz in
Bot. Arch. 31; 531 (1931). Type; Schlechter 390 (Bl).
Thunberg applied his own illegitimate epithet “ trifida ” to specimens from the
Tulbagh, Ceres, Clanwilliam, Calvinia, and Vanrhynsdorp districts, which I now
recognize as H. pinnata L.f. However, it appears that the name H. trifida Thunb.
has consistently been misapplied to a plant from the Cape Peninsula which appears
still to be without a name. Nevertheless, it would be premature to give it a new name
at this stage.
Heliophila elongata {Thunb.) DC., Syst. 2; 697 (1821).
Cheiranthus elongatus Thunb., Prodr. PI. Cap.; 109 (1800).
De Candolle made the combination, but did not identify any other plants with
Thunberg’s specimen. In the same work he described H. stylosa Burch, ex DC on
page 692 and H. virgata Burch, ex DC. on page 693. They are conspecific with H.
elongata. H. dolichostyla Schltr. in Bot. Jahrb. 49; 414 (1913) is another synonym.
168
Adamson in FI. Cape Penins.: 413 (1950) applied the name H. elongata (Thunb.)
DC. to the species which formerly went under the name H. linearifoUa Burch, ex DC.
This was presumably done on the basis of Sonder citing H. elongata (Thunb.) DC.
and the basionym in synonymy under H. linearifoUa Burch, ex DC. in Flora Capensis.
After studying Thunberg’s type there is no doubt that this is a wrong application
of the name H. elongata.
So far true H. elongata (Thunb.) DC. is the only species in which sessile axillary
resinous glands have been observed. Similar glands can sometimes be seen on pedicels
with reduced sepals in the lower part of the inflorescence below the normal flowers.
Heliophila carnosa {Thunb.') Steud., Nom. ed. 2, 2; 742 (1840).
Cheirantlius carnosus Thunb., Prodr. PL Cap.: 108 (1800). Type: Hb. Thunberg
no. 15141! C. gramineus Thunb. l.c. Type: Hb. Thunberg No. 15157! H. tripartita
Thunb. l.c. Type: Hb. Thunberg No. 15227! //. grommea (Thunb.) DC., Syst. 2: 697
(1821). H. abrotanifolia Banks ex DC. op. clt.: 690. Type not seen. H. platysiliqua
R. Br. in Hort. Kew. ed. 2, 4: 99 (1812). Type: Masson s.n. (BM!). H. abrotanifolia
Banks ex Sond. in Abh. Ges. Naturw. Hamb. 1: 249 (1846). Type as H. platysiliqua.
H. abrotanifolia var. heterophylla Sond. l.c., excl. syn. Thunberg. H. abrotanifolia
var. tripartita (Thunb.) Sond. l.c. Type as H. tripartita. H. abrotanifolia var. tenuiloba
Sond. in FI. Cap. 1: 49 (1860). Type: Zeyher 1901 (K, iso.!). H. sulcata Conrath
in Kew. Bull. 1908: 219 (1908). Type not seen. H. subcornuta Beauv. in Bull. Soc.
Bot. Geneve 1913, ser. 2, 5: 328 (1914). Type: Jacottet 86 (G!). H. grandiflora
Schltr. in Bot. Jahrb. 49: 415 (1913). Type: Schlechter 11353 (B!).
This is the most widespread species of the genus and also the most variable. Apart
from an occasional distinctive plant there is a gradation over the whole range, and it
has proved impossible to sort out any clear-cut varieties.
Heliophila laciniata W. Marais, sp. nov. ; a H. carnosa Steud. planta annuali;
a H. crithmifolia Willd. foliis omnino laciniatis, floribus majoribus, seminibus par-
vioribus plerumque longis latioribus differt.
Herba annualis glabra, foliis rosulatis. Folia 4-15 cm longa pinnato-5-14-lobata
omnino laciniata alterna vel opposite brevia vel longa ascendentia descendentia vel
patentia; stipulis duabus minutis subtenta; folia inferiora interdum simplicia. Pendu-
culus 20-50 cm altus nudus simplex vel ramosus. Pedicelli sub anthesi 8 -8-1 -3 cm
longi erecti, in fructu 1-1-6 cm longi patentes vel recurvato-patentes ; folia bracteis
duabus minutis subtenta. Sepala 4-5 mm longa 1 -4-1 • 5 mm lata oblonga, exterioribus
duobus cucullatis apice brevissime cornutis, interioribus duobus basi inflatis. Petala
8-8-11 mm longa 4 - 3-6 mm lata obovata vel valde late obovata rotundata unguiculata.
Filamenta 3 -2-3 -8 mm et 3-4-4 mm longa. Anther ae 1- 7-2-1 mm longae. Ovarium
lineare, stylo brevi, ovulis 20-28. Siliqua 2-5-4 cm longa, 2-2-7 mm lata linearis,
margine recto; valvae 1-nervatae subtiliter reticulatae; stylus 1-2-5 mm longus tenuis
rectus. Semina l-7-l-8x 1-5-1 -6 mm, reniformi-rotundata anguste alata.
Cape. — Namaqualand: Springbok, Acocks 19573 (K; PRE, holotype); Acocks 19335
(K, PRE); between Spektakel and Komaggas, Bolus 6515 and in Herb. Mus. Austro-
Afric. No. 483 (BM, BOL, K).
Glabrous annual herb with the leaves crowded in a basal rosette. Leaves 4-15
cm long, pinnately 5-14-lobed, the lobes spaced over the whole length of the leaf,
alternate or opposite, short or long, projecting backwards, at right angles or upwards;
the lower few (juvenile) leaves sometimes simple; each leaf subtended by two minute
stipules. Peduncle 20-50 cm tall, naked, simple or branched. Pedicels 0-8-1 - 3 cm long
in flower, erect; 1-1-6 cm long in fruit, spreading or recurved-spreading; each
subtended by two minute bracts. Sepals 4-5 mm long, 1-4-1 -5 mm broad, oblong;
the outer two cucullate, very shortly horn-tipped; the inner two inflated at the base.
Petals 8-8-11 mm long, 4-3-6 mm broad, obovate or very broadly obovate, rounded,
169
clawed. Filaments 3 -2-3 -8 and 3-4-4 mm long. Anthers 1- 7-2-1 mm long. Ovary
linear; style short; ovules 20-28. SiUqua 2-5-4 cm long, 2-2-7 mm broad, linear,
margin straight; valves 1-nerved, finely net-veined; style 1-2-5 mm long, slender,
straight. Seeds 1-7-1 -8 by 1-5-1 -6 mm, reniform-circular, narrowly winged.
Fleliophila laciniata is allied to but differs from H. carnosa. 'm being an annual and
it differs from H. crithmifolia in the leaves, which have the lobes spaced along the
whole length of the leaf, the larger flowers, and the smaller seeds which tend to be
broader than long. Our species is known from only a few collections from an area
south-west of Springbok. They were collected during July and September. The flowers
are white, becoming mauve.
W. Marais
CYCADACEAE
Encephalartos inopinus R. A. Dyer sp. nov., foliolis lineari-attenuatis 14-21 cm
longis 8-13 mm latis 13-21-nervatis infra glaucis facile distinguitur.
Planta basi ramosa; rami cylindrici usque 2 m longi, 17-23 cm diam. Folia
80-110 cm longa, petiolis 10-20 cm longis; rhachis rectus nonnunquam superne leviter
tortus, glaber, pulvino dense bruneo-lanato usque 3-5 cm longo et 4-5 cm lato,
foliolis infimis spinosis reductis; foliola media patenti-recurva lineari-attenuata, 14-21
cm longa, 8-13 mm lata, infra glauca, pungentia, marginibus integris vel rariter
margino inferne apicem versus 1-2 dentibus minutis instructis. Strohili ignoti.
Transvaal: Lydenburg; between Penge and Kromellenboog, cult. Johannesburg,
van Hoepen PRE 29663; Kromellenboog about 1 mile east of Asbestos mine on south
facing krans, Els photo; Dyer 5788 (PRE, holotype).
Rootstock freely branched from base; stems up to about 2 m long, 17-23 cm diam.
(including the persistent leaf bases); bracts lanceolate, 4 - 5 cm, long tomentose. Leaves
80-110 cm long including petiole 10-20 cm long; rhachis nearly straight, sometimes
slightly up-curved and twisted towards apex, glabrous except for woolly back of
pulvinus, flat or ridged on upper surface towards base, rounded on under surface,
8-13 mm thick; pulvinus up to 3-5 cm long, 4-5 cm broad, with dense brown wool
on back readily rubbed off, shrinking in size and wearing down considerably with
age to about 1 -5-2 cm long: leaflets glabrous, glaucous on under surface when young,
becoming green with age, base or foot inserted parallel to axis, pungent, with entire
margin or occasionally with 1-2 minute teeth on the lower margin near, and directed
towards apex, those leaflets above middle of leaf directed slightly upwards and
outwards and sometimes recurving, not overlapping, reduced in size near apex of
rhachis, those about the middle spreading more or less at right angles from rhachis,
those below directed slightly downwards and slightly falcate, 1-2 cm distant from each
other and reduced in size rather abruptly to 2-6 prickles on either side of the rhachis;
median leaflets linear-attenuate, 14-21 cm long, 8-13 mm broad in the lowest
I and from there gradually tapering to the slender pungent apex, only slightly narrowed
at base, not very rigid, with 13-21 veins evident but not raised on the lower surface.
Cones unobserved.
This species is established on two specimens from the Eydenburg district of the
Transvaal. One of these is now in cultivation in Johannesburg, collected in 1955
by Dr. Helmoed van Hoepen from a parent plant some distance from Penge and which
appears to have since perished in the wild state. The other is in the wild state, discovered
more recently by Mr. C. A. Els about one mile east of Kromellenboog on a krans
rising sheer from the Steelpoort river. This plant was investigated in August 1964
by an expedition from the Botanical Research Institute consisting of Dr. E. E. Codd,
170
Miss I. C. Verdoorn, Mrs. E. van Hoepen and the writer, guided by Mr. Els. A local
Bantu claims to know of the presence of about 3 more plants on the mountains a few
miles further east of Kromellenboog. No other species is known within many miles
of this area.
To which of the known species of Encephalartos this one is most nearly related
is not a simple matter to decide. It is remarkable in how many inconspicuous ways
it differs from those species to which it bears a general resemblance. The free branching
into relatively slender trunks is noteworthy but not of specific importance and it is
mainly the leaf characters which attract critical attention. The pulvinus is broad for
its length; the rhachis is glabrous even at an early stage and it often has a slight twist
towards the apex; the comparatively long linear-attenuate leaflets are gracefully
spreading recurved and deflexed towards the base, glaucous on the under surface at
first and becoming green with age, and the basal ones are reduced to prickles. Its
foliage is certainly amongst the most graceful in the genus. Cones are unrecorded.
The coincidence of the discovery of an “ unexpected ” new species of Encephalartos
and on a farm with the name “ Onverwagt ”, prompted the choice of the specific
epithet.
R. A. Dyer
GRAMINEAE
Eriochrysis brachypogon Stapf subsp. australis J. G. Anderson, subsp. nov. a typo
robustiore (ad 1 • 3 m longa) differt.
Eolia ad 8 mm lata, undique dense pilosa. Spiculae sessiles 5-6 mm longae, apice
glumae inferioris inaequaliter 3-lobata vel acuta, apice glumae superioris cuspidata
vel breviter aristata. Lemma floris superioris 3 mm longa. Spiculae pedicellatae
4-5 mm longae. Antherae 3 mm longae.
Swaziland. — Mbabane district: Forbes Reef, swamp, Compton 30488 (PRE, holo-
type).
Transvaal. — Ermelo district: Athole Pasture Research Station, vlei, Norval 100.
Pietersburg district: Woodbush, Wager 22960; Worsdell s.n. no locality.
This subspecies is apparently rare and has so far been collected only in the Mbabane
district of Swaziland and the Ermelo and Pietersburg districts of the Transvaal, where
it grows in swampy areas at an altitude of approximately 4,500-6,000 ft. Professor
R. H. Compton, who collected it in Swaziland, reports that, although occurring only
as scattered individuals, they are conspicuous plants, standing high above the general
level of the vegetation in one of the several small swamps between Forbes Reef and the
top of the Komati Pass.
It is readily distinguished from typical Eriochrysis brachypogon by its more robust
habit, the wider culm leaves which are densely and softly hairy on both surfaces and
by the spikelets which are larger in all parts.
Danthonia aureocephala J. G. Anderson, sp. nov. D. davyii C. E. Hubbard affinis,
sed paniculis contractis plerumque interruptis, spiculis maioribus plumis lanceolatis
differt.
Gramen perenne dense caespitosum ad 90 cm altum. Folia ad 40 cm longa stride
involuta vel canaliculata purgentia; vaginae intus prope ligulam sericeae. Paniculae
contractae plerumque interruptae. Glumae lanceolatae ad 2 cm longae aureae nitidae,
nervo medio prominenti, nervis lateralibus 3-5 obscuris glumae aequantibus conver-
gentibus nervo medio confluentibus. Lemma valde bilobatum ad 13 mm longum
171
(lobi inclusi), inter lobos aristatum, obscure lO-nervatum nervis convergentibus in
aristam ductis, extus fimbria media obliqua transversa interrupta instructum, saepe
margine (ab medio deorsum) pilosa; lobi anguste lanceolati acuminati scaberuli ad
7 mm longi ; arista ad 1 5 mm longa geniculata contorta ; callus 1 ■ 5 mm longus truncatus
vel obtusus latere pilosus. Palea ad 9 mm longa, apice breviter 2-lobata. Antherae
4 mm longae.
Natal. — Bergville; Cathedral Peak Forest Research Station, fairly frequent on ridges
of spur leading to Organ Pipes Pass, Killick 1727 (PRE, holotype); 1831; very locally
common on Organ Pipes Ridge, Killick 3450; N.E. slopes on the Camel, Esterhuyzen
12874; Mhlwazeni Valley, Cathkin Peak Area, Edwards 2453.
A wiry, densely tufted perennial up to 90 cm tall. Culms erect, simple, glabrous,
obscurely striate, 1-noded; nodes mostly basal with the upper internodes long. Leaves
mostly basal, glabrous; sheaths striate, sericeous on the adaxial surface near the
ligule especially on and near the mid-nerve, glabrous abaxially, basal sheaths com-
pressed, short, stramineous, hard, shiny, upper sheaths long, rounded or slightly
compressed, pale green or tinged with purple; ligule a ciliate fringe up to 1-0 mm long;
blades up to 40 cm long, linear, rigid, tightly involute or canaliculate, pungent, smooth.
Inflorescence a contracted, usually interrupted panicle up to 17 cm long, rigid; rhachis
subangular, striate, glabrous; branches solitary or binate, lower up to 5 cm long,
upper shorter; pedicels 3-20 mm long, subterete or angular, minutely scaberulous,
sparingly hairy at the apex. Spikelets up to 23 mm long, golden-yellow, shiny, 3-4-
flowered; rhachilla internodes up to 2 mm long, glabrous or ciliolate on the margins,
apex explanate, rounded. Glumes lanceolate, subequal, up to 2 cm long, scarious,
glabrous, golden-yellow, shiny, apex acuminate; midnerve prominent, lateral nerves
3-5, obscure, ^ the length of the glume, usually converging and fusing with the
midnerve. Lemma membranous, lanceolate, deeply bi-lobed, up to 13 mm long
(including the lobes), awned from between the lobes, obscurely 10-nerved with the
nerves converging and passing into the awn, glabrous adaxially, abaxially with 2-3
tufts of hair forming a median oblique transverse interrupted fringe on either side of
the midrib, often pilose on the margin from the middle downwards, otherwise glabrous,
tufts of hair up to 5 mm long; lobes narrowly lanceolate, acuminate, sometimes shortly
mucronate, scaberulous, up to 7 mm long, awn up to 15 mm long, flattened, narrowing
upwards, scaberulous, geniculate, twisted in the lower half; callus 1-5 mm long,
truncate to obtuse, laterally pilose, hairs up to 1-5 mm long. Palea up to 9 mm long,
narrowly lanceolate, membranous, sharply keeled with a nerve in each keel, keels
minutely ciliate; apex shortly 2-lobed. Anthers 4 mm long.
Danthonia aureocepliala is a winter-flowering grass occurring on ridges in mountain
grassveld at high altitudes in the Cathedral Peak area of the Natal Drakensberg. It
is a hard wiry grass characterized by the shiny, golden-yellow appearance of the
spikelets.
This species is very closely related to Danthonia davyii C. E. Hubbard described
from Mt. Mlanje, Nyasaland. When compared with the description of D. davyii
(Kew Bull. 1939, 645) and with Drummond and Robson 5835, deposited in the National
Herbarium, Pretoria, the spikelets of D. aureocepliala were found to be larger in all
parts with the panicle contracted and dense instead of lax, and the glumes lanceolate
instead of narrowly lanceolate.
D. aureocephala is also related to three other species of Danthonia, all of which
occur in the Drakensberg area of Natal. They are D. drakensbergensis Schweick., D.
macowanii Stapf and D. stereophylla J. G. Anderson. Together with D. davyii C. E.
Hubbard, they form a more or less distinct group in the genus. In order to facilitate
172
the identification of this new species a short key which includes all the closely allied
species is given:
1. Lobes of the lemma adnate, less than 2 mm long, acute, awnless, median awn up to 6 mm
long, straight, rarely somewhat bent and twisted (i) D. macowanii
Lobes of lemma well developed, acute, acuminate or awned, median awn 8-20 mm long, geni-
culate and with a twisted base 2
2. Base of the leaf-blades hairy above the ligule; lobes of the lemma awned, lobes including the
awns 3-9 mm long 3
Base of the leaf-blades glabrous or scabrid above the ligule, lobes of the lemma acute or
acuminate, rarely with awns up to 2 mm long 4
3. Lemma (including lobes and awns) up to 1 cm long; leaf-blades usually breaking above the
ligule, the remaining portion ultimately splitting and recurving outwards, hairs on the lemma
arranged in distinct tufts (ii) D. drakensbergensis
Lemma (including lobes and awns) 1 • 1-1 -6 cm long; leaf-blades not breaking off and splitting
above the ligule; hairs on lemma not in distinct tufts but arranged in a more or less con-
tinuous band along the margins ; (iii) D. stereophylla
4. Lemma (including the lobes) up to 8 mm long; palea up to 7 mm long; glumes narrowly
lanceolate, panicle lax, yellow or yellow-brown; spikelets dull (v) D. davyii
Lemma (including the lobes) 1 • 1-1 -3 cm long; palea 8-9 mm long; glumes lanceolate; panicle
laxly contracted, golden-yellow ; spikelets shiny (iv) D. aureocephala
(i) D. macowanii Stapf in FI. Cap. 7: 527 (1899). Type: Cape. — ^Somerset East:
on the banks of streamlets near the summit of Bosch Berg, and rarely in the bed of
the Little Fish River, 4,800 ft., MacOwan 1986.
Distribution. — From the south-eastern and eastern Cape, Basutoland, Natal and
the eastern Transvaal. Usually found along streambanks in mountainous areas at
altitudes ranging from 3,000 to 9,000 ft.
(ii) D. drakensbergensis Schweickerdt in Fedde Rep. XLIII: 88 (1938). Type:
Cape. — Barkly West: summit of Doodmans Krans Mountain, 2,900 m, Galpin 6903.
Distribution. — Barkly East and Maclear districts of the north-eastern Cape, the
Drakensberg area of Natal and Basutoland and at Mariepskop in the north-eastern
Transvaal. Usually growing on streambanks and seepage areas at altitudes higher
than 6,000 ft.
(iii) D. stereophylla J. G. Anderson in Bothalia 7, 2: 419 (1960). Type: Natal. —
Bergville: Cathedral Peak Forestry Station, 7,400 ft, Killick 1184.
Distribution. — Occurs in the Drakensberg area of Natal and Basutoland usually
on rocky outcrops or on dry basalt cliffs at altitudes ranging from 6,000-11,000 ft.
(iv) D. aureocephala J. G. Anderson, sp. nov. Type: Natal. — Bergville:
Cathedral Peak Forest Research Station, fairly frequent on ridges of spur leading
to Organs Pipes Pass, Killick 1727.
Distribution. — Found only in the Bergville district and usually growing on ridges
in mountain grassveld at altitudes ranging from 3,450-9,000 ft.
(v) D. davyii C. E. Hubbard in Kew Bull. 501 (1936); FI. of Trop. Afr. 10,
1 : 137 (1937). Type: Nyasaland. — Mt. Mlanje, on steep grass slopes of the Lichenya
Gorge, 1950 m, Burtt Davy 2068/29.
Distribution. — Nyasaland and Southern Rhodesia on steep grass slopes and in
rocky situations at altitudes higher than 6,000 ft.
J. G. Anderson
173
New Combinations and New Species in the Aristideae
Stipagrostis sect. Schistachiie {Fig. & De Not.) de Winter, comb. nov.
Schistachne Fig. & De Not. in Mem. Acad. Sci. Tor. 2, 12: 252 (1851).
Stipagrostis uniplumis (Licht.) de Winter
var. intermedia (Schweick.) de Winter, comb. nov.
Aristida gradUor (Pilger) var. intermedia in Bothalia 4, 1: 124 (1941).
Aristida congesta R. & S.
subsp. congesta.
subsp. barbkollis {Trin. & Rupr.) de Winter, comb. nov.
Aristida barbkollis Trin. & Rupr., Gram. Stip. 135 (1842).
Stipagrostis namibensis de Winter, sp. nov., S. hermannii (Mez) de Winter affinis,
sed aristis brevioribus ( 1 • 4-2 ■ 5 cm), glumis acutis (non acuminatis) firmioribus latioribus
differt.
Gramen annuum ad 20 cm altum. Culmi simplices 2-3 (4)-nodosi, geniculato,
ascendentes. Panicula nonnihil contracta. Spiculae pallidae. Glumae inaequales-
glabrae, chartaceae, lanceolatae; inferiores superioribus longiores, c. 8-9 mm longae‘
3-4 (5)-nervatae, apice subobtuso vel acuto, emarginato; superiores 3-nervatae, 6-7
mm longae. Lemma glabrum, laeve, 3 • 5-4 mm longum (callum inclusum), inter apicem
lemmatis basin columnae articulatum; columna 2-5-5 mm longa; aristae 3, centrali
pluraosa 1-4-2 -5 cm longa, lateralibus glabris tenuissimis; callus acutus pilosus.
Falea parva squaraiformis. Lodiculae 2, membranaceae, nervis paucis. Antherae 3,
valde anguste oblongae. Ovarium glabrum, stylis distinctis, stigmatibus plumosis.
Caryopsis anguste elliptica 2-2-5 mm longa; embryo caryopseos c. ^ aequans; iiilum
caryopseos ^ aequans atro-brunneum.
South West Africa. — ^Omaruru Distr., 11 miles E. of Omaruru River mouth, in
depressions where water collects after rain on bare gravel flats, de Winter & Hardy
8119 (PRE, holotype; K, M, BOL, WINDHOEK, SRGH, US).
Stipagrostis ramulosa de Winter, sp. nov., S. sabuUcola (Pilger) de Winter affinis,
sed planta parviori, glumis brevioribus, aristis longioribus, infiorescentiis laxioribus,
spiculis paucioribus differt.
Gramen perenne, valde dense caespitosum, multo ramosum ad 30 cm aitum et
1 m diam., foliis tenuibus. Culmi graciles, a nodis fasciculati, multo nodosi. Foliorum
vaginae laeves vel carina scaberulae. Ligula floccosa. Foliorum laminae filiformes,
ad 10 cm longae. Panicula laxa elongata, spiculis paucis (6-20). Spiculae 1-florae,
pallidae. Glumae subaequales, lanceolatae, acuminatae, 3-nervatae, inferiores 8-9 mm
longae, superiores 7-8 mm longae. Lemma tubulosum, 3-nervatum, glabrum, laeve,
2-5 mm longum (callum inclusum), prope apicem articulatum; columna conica inanis,
c. 1 mm longa; aristae 3, plumosae, excurrentes apice nudo, subaequales vel Jaterale,
leviter breviores, centrali 1-2-1 -5 cm longa, lateralibus 1-0-1 -3 cm longis. Palea
squamiformis, parva, c. 1 mm longa. Lodiculae 2, membranaceae, obtusae vel acutae,
nervis paucis. Stamina 3; antherae anguste oblongae, c. 2 mm longae. Ovarium
glabrum, stylis distinctis, stigmatibus plumosis, angustis. Caryopsis anguste elliptica,
2-2-5 mm longa embryo caryopseos f aequans; hilum caryopseos aequans.
South West Africa. — Outjo Distr., Unjab River Mouth 12-9 miles N. of Torra Bay
on coast road, sandy river bed, de Winter & Hardy 8197 (PRE, holotype; K, M, BOL,
WINDHOEK, SRGH. US).
B. DE Winter
174
GUTTIFERAE
In the Flora of Southern Africa 26: 250 (1963), a specimen from Swaziland,
Ben Dlamini s.n., was listed as Acokanthera sp. The receipt of further material of this
plant has led to its correct identification as Garcinia gerrardii Harv. ex Sim. The name
Garcinia gerrardii was not validly published by Harvey in his Genera of South African
Plants, ed. 2: 26 (1869). It was, however, taken up and provided with a description
and illustration by Sim in The Forests and Forest Flora of the Colony of the Cape
of Good Hope, 141, t. 13 (1907). In the same year, G. natalensis Schltr. was described
and, as the two are generally considered to be synonyms, it becomes necessary to settle
the question of priority. The date of publication of the part of Engler’s Botanische
Jahrbiicher containing G. natalensis is given as 21 May, 1907. Evidence collected by
Mr. W. Marais, our Liaison Officer at Kew, indicates that Sim’s publication was
available by May 1907 and probably has priority. I am grateful to Mr. Marais for the
trouble he has taken to obtain this information, and to Dr. Steam of the British Museum
(Natural History) who provided him with useful assistance and advice. The pertinent
information gathered by Mr. Marais is summarized below:
1. The Copyright Receipt Office of the British Museum, Bloomsbury, reports
that Sim’s book was received on 14 May 1907.
2. The University Library, Cambridge, reports that their copy is stamped 27
May, 1907.
3. The National Library of Scotland also received their copy on 27 May, 1907.
4. The Library of the Trinity College, Dublin, received their copy during May,
1907, but cannot give the exact date.
5. The book was reviewed by the Natal Agricultural Journal for 28 June, 1907.
The review could hardly have appeared in less than two months after the copy was
distributed.
6. A query sent to the printers in Aberdeen was referred to the Aberdeen Public
Library, who stated that the date was given in the English Catalogue of Books as
March, 1907. This date is not altogether reliable as the Catalogue is in part compiled
from “ dates of intended publication ”.
The weight of evidence indicates that Sim’s book was available by the middle
of May. On this basis, the correct name for the Garcinia is as follows, together with
its synonyms.
Garcinia gerrardii Harv. ex Sim, Forest FI. Cape Col. 141, t. 13 (mid-May 1907).
G. natalensis Schltr. in Bot. Jahrb. 40: 93 (21 May, 1907); Robson in FI. Zamb. 1 : 398
(1961).
G. transvaalensis Burtt Davy in Kew Bull. 229 (1924); FI. Tvaal. 1: 252 (1926).
Acokanthera sp. Codd in FI. S. Afr. 26: 250 (1963).
Robson, I.C., draws attention to the similarity between this species and G. volkensii
Engl. (1895) which has the outer floral whorls usually 5-merous, while in G. gerrardii
they are 4-merous. He considers that the difference may be only of subspecific value,
but there are also slight differences in texture and venation which give the impression
that the two are distinct species.
L. E. Codd
175
MYRICACEAE
A Note on the Identity of Myrica conifera Burm. f.
While working on the genus Myrica for the Flora of Southern Africa, the author
discovered that the name Myrica conifera Burm. f. cannot be applied to the species
known by that name in South Africa. M. conifera was described by Nicholas Burman
in his Prodr. Cap. 31 (1768), the species being based on Pluk. Phytog. t. 48, f. 8 (1691).
The figure is of a twig with male inflorescences. The relative legend, though it indicates
that the plant comes from Africa, includes two pre-Linnean citations which refer to
the American species, M. cerifera L. (1753).
Lamarck, in his Encycl. 2: 593 (1786), was apparently the first to suggest that
Fig. 8 represented not an African species, but the American species, M. cerifera. In
a note on M. cerifera he states that M. aethiopica L. (Linnaeus regarded M. conifera
as a synonym of this species) is probably conspecific with M. cerifera. According
to Lamarck, neither Fig. 8 cited by Linnaeus under M. aethiopica in Linn. Mant. Alt.
298 (1771) nor the description of that species, reveals any significant differences between
the two species. Moreover, Ray Hist. 1800 (1704) cited by Plukenet under Fig. 8 refers
to a plant from Carolina and not from Africa.
Rendle, in Journ. Bot. 41 : 85 (1903), points out that Fig. 8 is obviously a male
specimen of the same species as Fig. 9, which represents a fruiting specimen of M.
cerifera. He states that the leaves are exactly alike in the two specimens and have the
serrations in the upper part of the blade, which is characteristic of M. cerifera. He
concludes that M. conifera is a synonym of M. cerifera and does not apply to the
African species.
The present author has compared the two figures and agrees that they represent
M. cerifera rather than the African species. Of probably greater diagnostic value than
the leaf margin mentioned by Rendle, are the short and broad male inflorescences
(Fig. 8) and the clustered fruits (Fig. 9) — characters which clearly refer the figures to
M. cerifera.
According to Rendle, the valid name of the African species is M. aethiopica.
The specimen of M. aethiopica in Herb. Linn, (1169-4) certainly represents the African
species. However, M. aethiopica is illegitimate, because it is nomenclaturally super-
fluous: in the original description of M. aethiopica, the earlier M. conifera is cited as
a synonym. See Article 63 of the International Code of Botanical Nomenclature
(1961).
The oldest name available is M. serrata Lam. The type, a Sonnerat specimen
(photo., PRE) in the Lamarck Herbarium, Paris, agrees with the species which up to
now has been known as M. conifera.
The synonymy as given above may be summarized as follows: —
1. Myrica cerifera L., Sp. PI. 1024 (1753).
Myrica conifera Burm. f.. Prodr. Cap. 31 (1768).
Myrica aethiopica L., Mant. Alt. 298 (1771), nom. illegit., pro parte quoad syn. excl.
spec, in Herb. Linn.
2. Myrica serrata Lam., Encycl. 2: 593 (1786).
Myrica conifera auct. non Burm. f.
Myrica aethiopica L., nom. illegit., pro parte, quoad spec, in Herb. Linn, tantum.
D. J. B. Killick
176
SCROPHULARIACEAE
Graderia linearifolia L. E. Codd, sp. nov., G. subintegrae Mast, affinis, sed foliis
linearibus glabris, plantis parvioribus compactioribus differt.
Herba compacta perennis 7-14 cm diam. Caules prostrati glabri tenues 4-11 cm
longi. Folia opposita vel subopposita secunda linearia glabra 1 -0-1 -6 cm longa 1 mm
lata, margine revoluto integro. Flores in axillis solitarii, pedicillis 2-3 mm longis,
bracteolis 2 linearibus 4-5 mm longis. Calyx infundibuliformis aequaliter 5-dentatus
6-7 mm longus, tubo 3-3-5 mm longo, dentibus ovato-triangulis 3-3-5 mm longis.
Corolla infundibuliformis rosea, siccitate nigra, 1-6-2 cm longa ad osculum 1-4-1 -8
cm lata; tubus 1 - 1-1 -3 cm longus, lobis subpatentibus late rotundatis 5-7 mm longis
5-7 mm latis. Stamina 4 inclusa didynama; antherae 2-loculares, loculis 2 mm longis.
Ovarium ovoideum compressum 2 mm longum 2-loculare, ovulis numerosis; stylus
filiformis 1-2 cm longus leviter exsertus, stigmate minute bifido. Capsula coriacea
compressa ad apicem latior, apice dehiscens. Semina numerosa 1-25 mm longa, testa
reticulata.
Transvaal. — Lydenburg District; on top of Steenkampsberg, between Lydenburg
and Draaikraal, flowering 18/10/57, Codd 9832 (PRE, holotype); fruiting 11/3/58,
Codd 9855.
A compact, semi-woody, perennial herb with a thin, woody rootstock, branching
from the base, forming small cushions 7-14 cm in diameter. Stems mainly prostrate,
glabrous, slender, somewhat quadrangular, 4-1 1 cm long. Leaves crowded, opposite
or subopposite or in whorls of 3 or 4, secund, erect, glabrous, linear, scarcely narrowed
at the base, 1 -0-1 -6 cm long and up to 1 mm broad, drying blackish; margin narrowly
revolute, entire; median nerve prominent below, slightly impressed above. Flowers
crowded towards the centre of the plant, solitary in the leaf axil; pedicels 2-3 mm
long; bracteoles 2, linear, erect or subspreading, 4-5 mm long, inserted at or below
the base of the calyx. Calyx funnel-shaped, equally 5-toothed, 10-nerved, 6-7 mm
long at flowering stage, glabrous; tube 3-3-5 mm long; teeth ovate-triangular, acute,
3-3-5 mm long, each with a distinct median nerve. Corolla funnel-shaped, pink to
rose-red, drying black, minutely and sparsely puberulous without, 1-6-2 cm long and
1-4-1 -8 cm wide at the mouth; tube 1-1-1 -3 cm long; limb somewhat spreading,
5-lobed; lobes subequal, broadly rounded, 5-7 mm long and 5-7 mm broad. Stamens
included, didynamous; filaments slender, minutely puberulous below, the two shorter
4 mm long attached 5 mm above the base of the corolla tube, the two longer 7-8 mm
long attached 2-2-5 mm above the base of the corolla tube; anthers 2-celled; cells
diverging, oblong-falcate, 2 mm long, mucronulate, one cell narrower than the other
Ovary ovoid, compressed, glabrous, 2 mm long, 2-celled; ovules numerous; style
filiform, glabrous, 1 -2 cm long, exceeding the stamens, slightly exserted; stigma some-
what thickened, minutely bifid. Capsule included in the persistent calyx, woody,
compressed, widening obliquely towards the apex, obscurely bilobed, one lobe acute
with often the base of the style persisting, the other lobe rounded, dehiscing along the
apex. Seed numerous, oblong, 1-25 mm long; testa reticulate. Plate 1.
The genus Graderia was reviewed by Melchior in Notizbl. Bot. Gart. Berk 1 1 : 603
(1932), where four species were recognized, namely: G. scabra (L.f.) Benth., distributed
from about George in the Cape through the eastern Cape Province, Natal, Swaziland,
eastern Transvaal to eastern Southern Rhodesia; G. subintegra Mast., restricted to
southern and eastern Transvaal; G. iringensis Melch., based on Schlieben 1414 from
Tanganyika; and G. fruticosa Balf. f., recorded from Socotra.
To these a fifth species is now added, known as yet only from the top of the
Steenkampsberg, on the road from Lydenburg to Draaikraal. Here it grows in sparse
mountain grassland on poor, gravelly, quartzitic soil. It is readily distinguished from
the other species by the dwarf, compact habit and the narrow, linear, glabrous leaves.
177
The flowers are smaller and, because of the short internodes, appear close together
near the centre of the small cushions. The plants are inconspicuous and are easily
overlooked except during their short flowering period in spring.
L. E. CODD
STERCULIACEAE
Meihania Integra Verdoorn, sp. nov., stylis longis ad M. acwninatam accedens sed
foliis angustioribus marginibus integris differt.
Fruticulus, 20-45 cm altus, basi lignosus, rhizomatus. Caules plures, graciles,
teretes, lignosi, parce ramosi, ramulis breviter tomentosis. Stipulae subulatae, 6-15
mm longae. Petioli 0-4-1 cm longi. Folia anguste lanceolato-elliptica, 2-6-5 cm
longa, 0-4-1 -2 cm lata, Integra, supra velutino-tomentosa, infra albo-tomentosa
plerumque squamis minutibus et stellato-piliis evidentibus. PeduncuU axillares, 1-3
flori, 4-20 mm longi, pedicellis 3-14 mm longis. Bracteolae ovatae acuminatae, basi
subobtusae vel cordatae, c. 12 mm longae, 6 mm latae, utrinque tomentosae. Calyx
5”partitus, lobis lanceolatis acuminatis, c. 15 mm longis, 3-5 mm latis extus tomentosis.
Petala lutea, c. 16 mm longa, 10-14 mm lata, calyce longioria. Stamina filamentis
1 mm longis antheris 3-4 mm longis; staminodii ligulati, 9 mm longi. Ovarium
subglobosum, c. 3 mm diam., albo-tomentosum, stylo 8 mm longo, ovulis 7 pro loculo.
Capsula c. 10 mm longa 8 mm lata tomentosa.
Transvaal. — Pietersburg; Wolkberg, Van Vuuren 1316; Codd & Verdoorn 10407
(holotype, PRE); 10433.
Shrublet 20-45 cm tall with several slender, woody stems from a woody rootstock’
laxly branched; branches shortly tomentose, sometimes minute, reddish-brown, stellate
scales are obvious in the tomentum of the new growth. Stipules subulate, 6-15 mm
long. Leaves on flowering branches clustered towards the apex, narrowly lanceolate-
elliptic, mostly 2-6-5 cm long and 0-4-1 -2 cm broad (lower leaves or leaves on sterile
shoots larger and lanceolate-elliptic to ovate-elliptic), entire, very rarely lower leaves
shallowly toothed in part, finely and densely velvetty tomentose above, greyish-
tomentose, with reddish-brown lepidote scales, bearing clustered hairs, obvious on
lower surface, nerves obscure above, prominent beneath, basal nerves 3, rarely 5,
subrounded at the base, slightly narrowing towards the subacute or subrounded apex,
mucronate; petiole 0-4-1 cm long, tomentose and sparsely to densely lepidote with
minute reddish-brown scales. Inflorescence axillary in the upper leaves, 1-3-flowered;
peduncle 4-20 mm long, pedicels 3-14 mm long, both tomentose and with scattered
reddish-brown scales which bear clustered hairs. Epicalyx-bracts ovate acuminate,
rounded to cordate at the base, about 12 mm long, 6 mm broad near the base, tomentose
on both surfaces, also lepidote on the outer. Calyx-lobes lanceolate acuminate, about
15 mm long, 3-5 mm broad near the base, tomentose and lepidote without. Petals
canary yellow, turning brownish at maturity, broadest at the apex, about 16 mm long,
10-14 mm broad at apex. Stamens united at the base for about 1 mm, filaments 1 mm
long, anthers 3-4 mm long, staminodes 9 mm long. Ovary subglobose about 3 mm
diam., white tomentose; style 6-8 mm long, branches 1-5-3 mm long; ovules 7 in
each locule. Capsule about 10 mm long, 8 mm diam., shortly tomentose and scally,
the scales with groups of short hairs; seeds with numerous minute raised dots and
ridges.
This species resembles M. acuminata in the long style and petals, but differs among
other things, in the narrowly elliptic leaves with entire margins and the habit of the
plant. In. M. Integra several slender stems arise from a woody rootstock whereas
178
M. acuminata has, usually, one erect stem, which is branched, and a comparatively
slender tap root. The only other species of Melhania in South Africa known to have
entire leaves is M. prostrata which differs principally in the upper surface of the leaf
being glabrous, or pilose with subappressed single hairs, instead of densely velvetty
tomentose as in M. integra.
To date M. integra has been recorded only from the Wolkberg area where it was
found quite plentifully growing in mountain grassland on shallow soil over dolomite.
Melhania polygama Verdoorn, sp. nov., ad M. forbesii et M. acuminatam accedens,
sed ab ilia brateolis ad apicem subulatum attenuatis ab hac brateolis calyce longiore
sensim (non abrupte) attenuatis, ab ambabus foliis ovato-trullatis differt.
Fructiculus, c. 35 cm altus, basi lignosus. Caules plures, crassiusculi, siccitate
3-5 mm crassi, parce ramosi, ramulis cinereo-sulphureis stellato-tomentosis. Stipulae
subulatae, 5-10 mm longae. Petioli 1-2 cm longi, tomentosi. Folia utrinque stellato-
tomentosa, late ovato-trullata vel ovato-elliptica, 3-9 • 5 cm longa, 1 • 5-6 cm lata,
basi late cuneata vel subrotundata, ad apecim leviter angustata, obtusa vel subtruncata,
crenato-dentata. Pedimculi axillares, 6 cm longi, tomentosi, 1-3-flori, pedicelli 0-3-1 -5
cm longi. Flores polygami. Bracteolae ovatae, basi, cordatae vel subrotundatae,
ad apicem subulatum attenuatae, calyce longiores, c. 14 mm longae, 4-12 mm latae^
utrinque tomentosae. Calyx 5-partitus, lobis deltoideo-acuminatis, c. 1 1 mm longis,
extus stellato-tomentosis. Petala lutea, masculo 14 mm longis, femineo 10 mm longis.
Stamina masculo antheris 4 mm longis, femineo abortivis, 2 mm longis; staminodii
ligulati, 6-5 mm longi. Ovarium dense stellato-hirsutum ; stylus masculo c. 7 mm
longus ramulis abortivis vel brevibus; stylus femineo c. 2-5 mm longus, ramulis stylo
longioribus; ovula 5 pro loculis. Capsula late oblonga, c. 9 mm longa, tomentosa
et stellato-hirsuta.
Natal. — Lower Umfolozi, Umfolozi Game Reserve, Ward 3324; Dengeseni Beacon
Hill, Strey 4957a; 4957b (holotype, PRE); 4957 c and d.
Shrublet about 35 cm tall, with a woody rootstock. Stems many, rather robust, in
dried specimens 3-5 mm diam., laxly branched, new growth with a sulphur-grey tomen-
tum densely interspersed with rather long hairs of the same colour grouped on scales.
Stipules subulate, 5-10 mm long. Leaves finely and densely sulphur-grey tomentose
on both surfaces, upper surface somewhat darker than the lower, the tomentum densely
interspersed with grouped hairs; lamina broadly ovate-trullate or ovate-elliptic, 3-9-5
cm long, 1 - 5-6 cm broad, broadly cuneate or subrounded at the base, narrowing slightly
to the obtuse or subtruncate apex, shallowly crenate-dentate except at the base; petiole
1-2 cm long, densely sulphur-grey tomentose, fairly stout. Inflorescence axillary in
the upper leaves; peduncle up to 6 cm long, tomentose, 1-3-flowered at the apex;
pedicels 0-3-1 -5 cm long. Flowers polygamous. Epicalyx-bracts ovate, cordate at
the base, gradually attenuating to a subulate upper portion and acute apex, longer
than the calyx, about 14 mm long, 4-12 mm wide, tomentose on both surfaces. Calyx
with lobes deltoid-acuminate, about 1 1 mm long, outer surface tomentose interspersed
with scales bearing grouped hairs. Petals yellow, broadest at the apex, in essentially
male flowers about 14 mm long, in female shorter, about 10 mm long. Stamens in
male flowers, with anthers about 4 mm long, in female abortive, about 2 mm long;
ligulate staminodes about 6-5 mm long. Ovary densely beset with long, grouped
hairs on scales; style in male flowers long and slender, about 7 mm long, branches
aborted or short, style in female flowers thick and short, about 2-5 mm long, branches
longer than the style, about 5 mm long, sometimes decurrent on the style; ovules 5
in a cell. Capsule broadly oblong, about 9 mm long, tomentose and with groups of
rather long, sulphur-coloured hairs; seeds with numerous raised dots and ridges.
This is the first record of polygamous flowers in a species of Melhania. The
■ essentially male flowers can be recognized by the larger petals which are as long as the
179
epicalyx-bracts, while those of the female flowers are shorter than the bracts. Invariably
the anthers in the long-petalled flowers (male) were found to be well developed and the
style long with abortive branches. In the short-petalled flowers (female) the anthers were
small and abortive while the styles were short and had 5 well developed, long branches.
Only a very few flowers among the scores dissected appeared to be hermaphrodite.
To date M. polygama is known only from the type locality where it is common
in grassveld on the hill-slopes. The bushy plants, about 35 cm high, are formed by
many stems which arise from a woody rootstock. This habit of growth is another
feature in which this species differs from M. forbesii and M. acuminata, the two species
mentioned in the Latin diagnosis as resembling M. polygama in some respects. In
these species the plants are weakly perennial with a comparatively slender tap root
and one or two erect stems from which the seasonal growth arises. The shape of the
leaf, which is mentioned as a distinguishing feature in the Latin diagnosis, is usually
ovate-trullate in M. polygama that is the broadest portion is about a third of the length
above the base. In the other two species the leaves are ovate-oblong with the broadest
portion at or neai the base. The styles in the male flower of our species are long, like
those of M. acuminata, while in the female they are short, like those of M. forbesii.
The shape of the epicalyx-bracts, is distinct in all three species and can be used to
distinguish between them. In M. forbesii they are consistantly broadly acuminate,
in M. acuminata abruptly acuminate into a more or less caudate apex and in M.
polygama gradually attenuate to a more or less subulate apical portion.
1. C. Verdoorn.
181
Plate 1. — Graderia lineanfolia L. E. Codd (Codd 9832, holotype).
1846586
183
Species of Plocamium on the South African Coast
hy
R. H. Simons
Species of the Red Algal genus Plocamium are a common and an attractive feature
of the South African intertidal marine flora. This paper collates information on this
genus from literature and the collections of Tyson, Becker, Isaac, Simons, the University
of Cape Town Ecological Survey, Pocock and the Botanic Station, Durban. Specimens
of the latter two collections, although seen, are not cited here. The Ecological Survey
referred to above was carried out by Professor Stephenson assisted by various members
of the University staff between 1931 and 1940 and their collections are cited below
with the designation “ Ecol. SurvP, after the manner of Silva (1959).
The coast south of Swakopmund in South West Africa in the west and of Xai-Xai,
near the mouth of the Limpopo River in the east, is included in the geographical region
referred to here as South Africa.
Morphology
Plocamiums develop sympodially and the last-formed laterals of individual sym-
podia together constitute the more or less zig-zag axes of the erect fronds. The terminal
portion of each sympodium is usually displaced to one side and appears as a lateral.
The latter is referred to as the “lowest of the pinnae (in a secund group of laterals)”,
and is most often simple. All the true laterals of a sympodium are compound, unilateral
and in one plane; their number is limited and characteristic of the species. Because
the laterals arise adaxially only they occur as secund groups of pinnae on alternate
sides of the zig-zag axes.
Alternating secund groups of laterals consisting of more than two, at least in the upper parts:
Laterals in pairs below, in threes above 4. P. rigidum
Laterals seldom in pairs:
Lowest of laterals in each secund group usually branched, fertile branches (sporophores)
appear different from sterUe ones P. glomeratwn
Lowest of laterals in each secund group simple, sporophores scarcely different from sterile:
Simple pinnae much longer than wide, much less than 0-5 mm wide at base 3. P. becked
Simple pinnae about twice as long as broad, about 1 mm or more at base, plant broadly
flattened 2. P. membranacewn
Alternating secund groups of laterals mostly of two:
Plant foliose, apically somewhat membranous:
Lower of paired pinnae very narrowly triangular, less than 0 • 5 mm wide at base 8. P. telfairiae
Lower of paired pinnae broadly triangular, often more than 1 mm wide at base:
Plant about 2 mm wide, pinnae not obviously toothed, sporophylls modified pinnules 6. P. suhdi
Plant 5 mm or wider, leathery below, somewhat membranous above, pinnae frequently obviously
toothed on outer margins, teeth often compound, sporophylls frequently intermixed
with sterile bracts in axils of pinnae 7. P. corallorhiza
Plant not foliose, pinnae narrowly awl-shaped, fleshy 5. P. cormitum
1. P. glomeratum J. Ag. in Sp. Alg. 2: 397 (1852); Epicr. : 340 (1876). Fig. 1.
P. subfastigatum Kiitz. in Tab. Phyc. XVI: 18, t. 51, fig. e and f (1866).
Holdfast: Loosely intertwined, terete, slender, prostrate axes with short,
unbranched, rigid haptera and erect fronds. Erect system: Fronds up to 10 cm high.
4846586—4
184
Fig. 1 and 2. Fig. 1. — Plocamium glomeratum J. Ag. : (a) Portion of a well-developed plant, X 2,
(6) portion of a fertile plant showing sporophores, x 4; (c) portion of plant in detail, X n.
Fig. 2. — Plocamium membranaceum Suhr: (a) portion of a well-developed tetiasporic plant, X 2,
ib) portion of plant in detail showing axillary sporophylls, X 50.
185
subdichotomously branched; axes slender, more or less terete below, complanate
and somewhat membranous above but remaining narrow; up to six pinnae in each
secund group, the lowest frequently branched; sporophores aggregated, ultimately
ball-like, sporophylls bipinnately branched, modified pinnules; colour dark red; male
and female plants not seen.
South West Africa.— Swakopmund, Isaac I. 1717; Simons 354; 497.
Cape Province.— Namaqualand : Port Nolloth, Isaac I. 1716. Cape Town: Sea
Point, s.n. (BOL 27222); False Bay, ^'co/. iSwrv. F. 87. Caledon: Cape Hangklip,
Simons 676. Bredasdorp: Arniston, Ecol. Surv. A.R.3.F. Riversdale: Still Bay,
Ecol. Surv. S.S. 36, in part.
Agardh (1852) comments on three characteristics of this species which, in South
Africa at least, seem unique: first, the branching of the pinna lowest in each secund
group; secondly, the number of pinnae in each of such groups — three to six; thirdly,
the much-branched sporophylls aggregated on sporophores different from sterile
fronds — “ as in some ferns ”. The Ecological Survey records and specimens of this
species are based on a wrong identification by Kylin (1938) and have been transferred
to P. beckeri Simons.
2. P. membranaceum Suhr in Alg. Beitr., Flora 23: 261 (1840); J. Ag. in Sp.
Alg. 2: 399 (1852); Epicr.: 341 (1876); de Toni in Syl. Alg. 4: 593 (1900). Fig. 2.
P. latiusculum Kiitz. in Tab. Phyc. XVI: 17, t. 47 (1866).
Holdfast: Relatively stout branched haptera. Erect system: Broadly flattened
throughout, leathery below, somewhat membranous above where mid-rib sometimes
clearly visible; pinnae patent, deltoid or triangular, obtuse at tips, broad, about 2 mm
at base, lowest of each secund group simple, pinnules sometimes lobed; sporophylls
modified pinnules and axillary, widely distributed, shortly, compactly and dichotomously
branched. Male and female plants not seen.
Cape Province. — Cape Town: Table Bay, Tyson s.n. (BOL 27216); Camps Bay,
Simons 656.
The name given to this plant belies its true nature as it is quite as leathery as a well-
developed specimen of P. corallorhiza. As with the latter species, however, on being
pressed it loses its true texture and appears thinly membranous. The width of this
plant is stated to be as much as 8 mm although Kiitzing’s (l.c.) figure does not indicate
anything like this. The plants cited here are up to 6 mm wide and the detailed structure
is as indicated in Kutzing’s Figure “e”, t. 47 (l.c.). The appearance of this plant is
distinctive mainly because of its breadth and the clearly defined laterals in secund groups
of three or four. P. membranaceum appears to be a deep water form as it has only been
found washed up. The plants are up to 20 cm high.
3. P. beckeri Simons, sp. nov. Fig. 3.
P. glomeratum auct. non J. Ag. ; Kylin in Lunds Univ. Arsskr. N.F. Avd. 2, 34 (8):
12, t. 3, Fig. 7 (1938).
Thallus usque ad 15 cm altus, delicatulus, ramosissimus; axibus primariis com-
pressis, seriebus secundis alterne pinnatis; pinnis imis in qua serie fere subulatis,
simplicibus, secundis paulum ramosis, tertiis plus ramosis, quartis (rarissime) multo
ramosis, pinnulis gracillimis, submembranaceis, subcostatis; sporophyllis pinnularum
transformatione formatis, clavatis simplicibus vel dichotomo-decompositis ; cystocarpiis
non visis; colore coccineus.
Type: Natal, Port Edward, Simons 623 (PRE).
186
Fig. 3 and 4. Fig. 3. — Plocamium beckeri Simom: (o) portion of a well-developed plant, X 1; (6) portion of plant in detail, X 15, (c)
sporophylls as modified pinnules, X 30
Fig. 4. — Plocamium rigidum Bory: (a) Portion of a well-developed plant, X 1; (6) portion of a plant in detail, X 15; (c) sporophylls
as modified pinnules, x 30.
187
Holdfast: Short, intertwined, slender, rhizomatous axes with short, branched
haptera and erect fronds. Erect system: Fronds up to 15 cm high, delicately and
much-branched, compressed, slender; almost always three laterals in each secund
group, occasionally four, the lowest simple, almost awl-shaped, the second little-
branched, the third more-branched and, when present, the fourth most-branched;
ramuli very slender, somewhat membranous and faintly ribbed ; sporophylls transformed
pinnules, clavate, simple or dichotomo-compound, very often cruciform; bright pink
in colour.
Cape Province.— Bredasdorp : Arniston, lyoac I. 868, 1. 1538. Mossel Bay: Tergniet,
Isaac I. 297, I. 1568, I. 1569, I. 1715. Knysna: Robberg, Isaac I. 1428, I. 1429. Port
Elizabeth, Ecol. Surv. E. 60, E. 62, E. 78, E. 79. Bathurst: Kowie, Becker s.n. (BOL
27219, 27220, 27221).
Natal. — Port Shepstone: Port Edward, Simons 623 (PRE, type); Ramsgate, Simons
624. Durban: Reunion Rocks, Isaac I. 1023; Durban, Ecol. Surv. D. 64. Inanda:
Umhloti, Simons 630. Umfolosi: Umpangazi, Ecol. Surv. G. 1. G.
Mozambique. — Maputo: Ponto Abril, Isaac I. 379, I. 700. Inhaca Island: Light-
house, Isaac I. 200, B. 255. Gaza: Xai-Xai, Isaac 1. 710.
The name P. beckeri Schm. appears in manuscript on Becker mountings of this
species. Attempts to find a published reference to the name have failed, so it is assumed
to have no standing. In describing this as a new species it seems fitting to retain the
epithet “ beckeri ” to honour Dr. H. Becker who contributed so materially to South
African phycology in the late 19th century.
This species has in the past been confused with P. coccineum (Huds.) Lyngb.
from the northern hemisphere. P. beckeri differs in that its pinnae almost always
alternate in threes and it lacks marginal sporophylls. Another closely related species
is P. delicatulum Baardseth (1940), but the latter produces simple or once-forked
sporophylls compared with the frequent cruciform or sometimes much branched
sporophylls of P. beckeri.
4. P. rigidum Bory in Belanger, Voy. Bot. Crypt., 164 (?); Kiitz., Sp. Alg., 885
(1849); J. Ag., Sp. Alg. II, pt. 2: 397 (1852); J. Ag., Epicr. Syst., 340 (1876); J. B.
de Toni, Syl. Alg. IV, sect. 2: 592 (1900); H. Kylin, Rhodoph. von Siidafrika, 12,
t. 3, fig. 8 (1938). Fig. 4.
Holdfast: Creeping, compressed rhizomatous axes, branched at intervals, up to
1 mm thick; strongly recurved hook-like pinnae interlock with the same from adjacent
rhizomes. Erect system : Fronds arise from upturned apices and other parts of rhizomes,
much branched, somewhat leathery, rigid; pinnae alternately paired below, in threes
above; sporophylls more or less palmately branched, branches of variable length;
cystocarps sessile, marginal.
South West Africa. — Swakopmund, Simons 352, 355, 374, 496; Isaac I. 1351, I. 1352,
I. 1353, I. 1539, I. 1540, I. 1541. Liideritz, Simons 406, Isaac I. 1356, I. 1357, I. 1358.
Cape Province. — Namaqualand: Port Nolloth, Simons 357; Isaac I. 1168, I. 1260.
I. 1537, I. 1543, I. 1544, I. 1545, I. 1546; Ecol Surv. N. 6. A.; Hondeklip Bay, Ecol,
Surv. H.B.4.G. Malmesbury: Paternoster, Ecol. Surv. P.ll.E; Simons 481. Cape
Town: Oudekraal, Ecol. Surv. A. 87. Caledon: Cape Hangklip, Ecol. Surv. C.H.3.F.
Bredasdorp: Dyer Island Tyson s.n. (BOL 27213); Cape Agulhas, Ecol. Surv. A.G.4.F;
Tyson s.n. (BOL 27210, 27211, 27215). Locality unstated: Tyson s.n. (BOL 27212,
27214).
This species is very variable and its limits are difficult to establish. Generally,
the thallus is fairly rigid. In its more typical form it resembles P. cornutum, but.
188
Fig. 5 and 6. Fig. 5.—Plocamiwn cornutum (Turn.) Harv.; (a) Portion of a well-developed plant,
X 1; (b) portion of plant in detail, X 2; (c) sporophylls as modified pinnules, X 40.
Fig. e.—Plocamium suhrii Kiitz. : (a) Portion of a well-developed plant, X 1 ; {b) portion of plant
in detail, X 10; (c) axillary sporophylls X 50; (d) marginal, sessile cystocarp, X 10.
189
whereas in the latter species all the pinnae alternate in pairs and are more or less awl-
shaped, in P. rigidum the pinnae are somewhat more triangular and in the upper parts
occur in threes. There are some rather more delicate forms which seem otherwise
to be indistinguishable from the type. It is possible such forms were referred by
Grunow (1867) to P. rigidum var. tenuior. There is too a rather more membranous
form with somewhat more triangular pinnae but attempts to find any distinguishing
character from P. rigidum have not succeeded.
5. P. cornutum {Turn.) Harv. in Ner. Austr. : 123 (1847); J. Ag. in Sp. Alg. 2;
404 (1852); Epicr.: 346 (1876); de Toni, Syl. Alg. 4: 598 (1900). Fig. 5.
Fucus cornutus Turner in Hist. Fuc. : t. 258 (1811).
Thamnocarpus cornutus Kiitz. in Phyc. Gen.: t. 59, fig. 3 (1843); Tab. Phyc. XVI: 19,
t. 55, fig. a-c (1866).
Holdfast : Relatively stout, shortly branched, decurrent axes with knob-like
attaching suckers at their extremities; colour a dilute pink, the whole occupying little
space. Erect system: Ascending main axes comparatively little-branched; pinnae in
alternating pairs, the lower simple, awl-shaped, the upper somewhat branched, all
crowded, overlapping, appearing to arise from all sides; sporophylls pedicellate,
apparently axillary, shortly and furcately branched; cystocarps not seen.
Cape Province. — Namaqualand: Port Nolloth, Isaac I. 1205; Simons 311. Malmes-
bury: Paternoster, Simons 480; Ecol. Surv. P.4.F. ; Saldanha Bay, Simons 566. Cape
Town: Table Bay, Tyson s.n. (BOL 27164, 27165, 27172); Three Anchor Bay, Tyson
s.n. (BOL 27173); Oudekraal, Ecol. Surv. A. 54; Buffel’s Bay, Isaac I. 1224; Sunny
Cove, Isaac I. 378; Kalk Bay, Tyson s.n. (BOL 27166); False Bay, Ecol. Surv. F. 85.
Caledon: Hermanns, Simons 487, 552; Isaac 1. 377. Knysna: Ecol. Surv. K.KN.l.A.
Bathurst: Kowie, Becker s.n. (BOL 27167, 27168, 27169, 27170, 27171); Kleinemonde,
Ecol. Surv. X.l.V. East London: Kidd’s Beach, Isaac 1. 1307; East London, Ecol.
Surv. L. 90. Kentani: Qolora, Ecol. Surv. Q.Q.l.R; Locality unstated: Becker s.n.
(BOL 27174, 27175, 27176).
P. cornutum is comparatively easy to recognize because of its crowded pinnae
which appear to arise on all sides of the somewhat terete axis which is sparingly branched.
Occasionally forms approach the habit of P. rigidum but generally the latter can be
distinguished by their terminal arrangement of pinnae in secund groups of three.
6. P. suhrii Kiitz. in Sp. Alg.: 886 (1849); Tab. Phyc. XVI: 19, t. 54 (1866)
Fig. 6.
P. procerum Suhr, Alg. Eckl. : 726 (1834), nom. illegit.
P. nobile J. Ag., Sp. Alg. 2: 397 (1852); Epicr.: 341 (1876).
P. fuller ae Schm. ex Mazza in Sag. Alg. Ocean.: 221 (1908); de Toni, Syl. Alg. VI:
315 (1924).
P. membranaceum. Kylin in Rhodoph. von Siidafrika, t. 3, fig, 9 (1938). Auct. non
Suhr.
Holdfast: Slender rhizomes of limited length from which groups of short haptera
arise at intervals. Erect system: Compressed fronds of about 2 mm width arise on
short terete stalks from the holdfast, up to 15 cm high, not much branched; pinnae
in alternating pairs, sometimes strictly alternate or, occasionally, in threes, the lower
simple, triangular and somewhat falcate, margins entire, crenate or obsoletely toothed,
the upper laterals more or less branched; mid-rib of single row of cells often clearly
visible as a thin median line in young parts ; sporophylls modified pinnules, branched,
almost digitate, sometimes in the axils of pinnae; cystocarps marginal, sessile, on
pinnae.
190
7b.
Fig. 7 and 8. Fig. 7. — Plocamium corallorhiza (Turn.) Harv. : (a) Portion of a well-developed plant
showing axillary sporophylls, X 1 ; (b) sporophylls with toothed bract, X 40.
Fig. 8. — Plocamium telfairiae Harv. ex Kiitz. : (a) Portion of a well-developed plant, X 1 ; {b) portion
of a plant in detail, X 7; (c) sporophylls as modified pinnules, X 30.
191
Cape Province.— Knysna : Robberg, Ecol. Surv. R.R.2.C; Plettenberg Bay, Isaac
I. 1548. Humansdorp : Storms River, Ecol. Surv. T.2.G, T.T.4.D. Bathurst: Kowie,
Becker s.n. (BOL 27141, 27142, 27143, 27144, 27145, 27146). Locality unstated: Tyson
s.n. (BOL 27217, 27218).
Natal.- — Inanda: Umhloti, Simons 631.
Agardh (1876) by retaining his name of P. nobile for this taxon rejected the prior
claim of P. suhrii published by Kiitzing in 1849. The justification for this was apparently
Agardh’s refusal to acknowledge that Kiitzing’s two publications of P. suhrii are
synonymous and both referred to P. procerum Suhr. However, there seems to be no
good reason for allowing this suggestion of Agardh’s and the Kiitzing name must,
therefore, be accepted as correct, since P. procerum Suhr, on which this taxon is based,
is a homonym.
The main feature of this species is the grouping of laterals in pairs as is found in
P. corallorhiza. It differs from the latter in its smaller size, both in length and width,
the absence of teeth on the margins of the pinnae, and the occurrence of sporophylls
as modified pinnules. Agardh’s statement that the laterals occur most often in secund
groups of three is misleading, unless he had the wrong plant in hand, as both Suhr
and Kiitzing stated quite clearly that the laterals are paired or even strictly alternate.
Suhr also pointed out that his plant was akin to P. corallorhiza rather than to P.
coccineum. This being the case, Agardh’s statement must be viewed with caution.
Plants, which might be of this species, with laterals occasionally in threes have been
seen.
7. P. corallorhiza (Turn.) Harv. in Hook, et Harv. Alg. New Zeal.: 542 (1845);
Harv. in Ner. Austr.: 121 (1847); J. Ag. in Sp. Alg. 2: 402 (1852); Epicr.: 348
(1876); de Toni in Syl. Alg. 4: 602 (1900). Fig. 7.
Eucus corallorhiza Turn, in Hist. Fuc. t. 96 (1811).
P. robertiae Schmitz ex Mazza in Sag. Alg. Ocean. Nuova Not. 19: 18 (1908).
Eucus cirrhosis Turn, in Hist. Fuc. t. 63 (1811).
Thamnophora corallorhiza C. Ag. in Sp. Alg.: 225 (1821); Syst. Alg. 240 (1824); Suhr
in Alg. Eckl.: 7, t. 1, fig. 10 (1834); Kiitz. in Sp. Alg.: 887 (1849); Tab. Phyc. XVI:
20, t. 56, fig. f-k (1866).
Holdfast'. Relatively thick, somewhat flattened, branched rhizomes with short
strongly recurved simple pinnae and short haptera attaching plant to rock or other
plants. Erect system'. Fronds from holdfast narrowly complanate at base rapidly
expanding to 10 mm or more in width, compressed, leathery below, slightly membranous
ultimately, up to 30 cm high; pinnae alternate or more commonly alternating in pairs,
the lower simple, the upper more or less branched, broadly triangular, outer margins
frequently arcuate and toothed, teeth often compound ; sporophylls axillary or superficial,
often intermixed with minute, toothed bracts; cystocarps marginal on axillary bracts,
frequently closely clustered.
Cape Province. — Malmesbury: Yzerfontein, Ecol. Surv. Y.l.F. Cape Town: Table
Bay, Tyson s.n. (BOL 27179, 27181, 27182, 27183, 27184, 27186, 27187, 27190);
Kommetje, Isaac 1. 826; Oliphantsbos, Isaac 1. 270; Buffel’s Bay, Isaac 1. 1221;
Muizenburg, Isaac 1. 266, 1. 267, 1. 261 A', False Bay, Ecol. Surv. F. 4; Tyson s.n.
(BOL 27185, 27188, 27189, 27191, 27192, 27197). Caledon : Hermanns, Isaac 1. 268.
Bredasdorp: Arniston, Isaac 1. 864, 1. 927. Riversdale: Still Bay, Ecol. Surv. S. 81.
Mossel Bay: Tergniet, Isaac 1. 296, 1. 1562. Knysna: Plettenberg Bay, Isaac 1. 1407.
Humansdorp: Storms River, Ecol. Surv. T.T.4.C. Port Elizabeth: Ecol. Surv. E. 22.
192
Bathurst: Kowie, Becker s.n. (BOL 27140, 27147, 27148 and consecutive numbers up
to 27161, 27194, 27195, 27196); Shand s.n. (BOL 27180); Kleinemonde, Ecol. Surv.
X.5.A. East London: Kidd’s Beach, Isaac 1. 265, 1. 265A; East London, Ecol. Surv.
L. 38; Eastern Beach, Isaac 1. 269. Komgha: Cape Morgan, Flanagan s.n. (BOL
27177). Willowvale : Udwesa Point (Dwessa), /raac 1. 271, 1. 1272. EUiotdale: The
Haven, Ecol. Surv. H.l.C. Port St. Johns: Isaac 1. 1448, Simons 657 and consecutive
numbers up to 675. Locality unstated: Becker s.n. (BOL 27193).
Natal. — Pinetown: Isipingo, Isaac 1. 274. Durban: Reunion Rocks, Isaac 1. 467;
Durban, Ecol. Surv. D. 30. Lower Tugela: Umhlali, Ecol. Surv. U.8.A, U.22.A,
U.U.4.A; Salt Rock Beach, Isaac I. 273.
Mozambique. — Gaza: Xai-Xai, Isaac I. 713.
Locality unstated: Becker s.n. (BOL 27198 and consecutive numbers up to 27205).
There is a fair degree of variation in the form of the simple pinnae of this species
especially regarding their shape and the presence or absence of marginal teeth. The
outer margins of these pinnae are anything from strongly incurved to slightly recurved
and with or without obvious teeth. Considering the range of variation in these
characters, there seems to be no justification for the recognition of a separate species,
P. robertiae, with entire margins.
Structures resembling axillary clusters of sporangia, but which consist instead of
what seem like miniature juvenile forms of the plants bearing them, are of fairly frequent
occurrence. The outer margins of these tiny pinnae are usually markedly toothed.
These bracts are sometimes found intermixed with sporophylls, or they contain
spermatangia.
8. P. telfairiae Harv. ex Kiitz. in Sp. Alg. : 885 (1849); J. Ag. in Sp. Alg. 2: 400
(1852); Epicr.: 342 (1876). Fig. 8.
P. abnorme Okam. in Icon. Jap. Alg. 3: 1 and 4, t. 101 (1913), non Hook, et Harv.
HoUfast: Slender, creeping, compressed axes with scattered, short, recurved
pinnae and haptera at intervals. Erect system : Fronds slender, compressed, somewhat
membranous, expanding slightly towards apices ; pinnae in alternating pairs, the lower
simple, entire, elongated, narrowly triangular, the upper branched, pinnae patent
below, crowded and delicate above; sporophylls axillary, or modified pinnules, branches
variable in length and form; cystocarps not seen.
Mozambique. — Inhaca Island: Lighthouse, Isaac 1. 913, B. 427.
This is a new record for southern Africa. Yendo (1915) suggested that this species
is synonymous with P. nobile J. Ag.
Literature Cited
Agardh, j. (1852) Species Algarum. Lund.
Agardh, j. (1876) Species Algarum. Epicrisis Systematis Floridearum. Lip-
siae.
Baardseth, E. (1941) The Marine Algae of Tristan da Cunha — Results of Nor-
wegian Scientific Expedition to Tristan da Cunha, 1937-
1938. No. 9. Oslo.
Barton, E. S. (1893) A Provisional List of the Marine Algae of the Cape of Good
Hope. J. Bot. 31.
Delf, E. M. and M. R. Michell (1921) The Tyson Collection of Marine Algae. Ann. Bol. Herb. 3:
110-111.
De Toni, J. B. (1900) Sylloge Algarum 4. Patavii.
De Toni, J. B. (1924) Sylloge Algarum 6. Patavii.
193
tw •
^ ja '*^1
rs^y,
fr m
a<!4^:
m
Ifl.’si
V-',. :--^a^.’VMii
* - Wir
- P'
IS L
J\^- ' I'-i'V '^'
195
Notes on the Species of Zonaria in South Africa
by
R. H. Simons
Summary. — An account is given of the anatomy of the three species of Zonaria
recorded from South Africa by Stephenson (1948). It is pointed out that Zonaria
harveyana (Pappe ex Kiitz.) Aresch. is dorsiventrally constructed, whereas Z. tournefortii
(Lamour.) Mont, and Z. subarticulata (Lamour.) Papenf. are not. The desirability of
a generic separation is acknowledged but, because a nomenclatural problem may arise
therefrom a name for the new genus is not given.
Observations. — During a routine check on the identity of a Brown Alga collected
in Natal — Zonaria harveyana (Pappe ex Kiitz.) Aresch. (Papenfuss, 1944) — it was
found that this species is distinctly dorsiventral in structure. This dorsiventrality is
indicated by the following: —
(i) The cells of one surface (the dorsal) are more or less cuboid but sometimes
only half the width of the elongated cells making up the medullary tissue, whereas
the ventral surface has cells almost indistinguishable from those of the medulla (Fig. 1 ;
Plate 2A) and are parallelepiped-like.
Fig. 1, 2 and 3. Fig. 1. — L.S. of Zonaria harveyana thallus. Fig. 2. — T.S. of Zonaria harveyana
thallus. Fig. 3. — L.S. of Zonaria harveyana passing through a transverse ridge.
d.s. — dorsal layer; f.c. — fertile cell; i — indusium; m — medulla; N.F. — new frond; O.F. —
old frond; s.c. — supporting cell; v.s. — ventral layer.
196
(ii) Zones of new growth originate only from the ventral surface of the frond
(Fig. 3; Plate lA).
(iii) Sori arise only on the ventral surface (Fig. 3; Plate IB).
The structure of Zonaria tournefortii (Lamour.) Mont, and Z. subarticulata
(Lamour.) Papenf. (Papenfuss, 1951), both from Natal, were then studied for com-
parison, and neither plant was found to have any dorsiventral characteristics. Z.
tournefortii, however, is sometimes confusingly like Z. harveyana in that transverse
sections of their thalli are scarcely different from each other; both may have superficial
cells of the same width as, or half the width of, the medullary cells (Fig. 2), but these
two species differ in respect of the distribution, and relative number, of half and full-
width cells in the surface layers. In the specimens examined the relative number of
half-width cells was far higher in Z. tournefortii than in Z. harveyana furthermore,
in the latter species, only full-width cells are present in the ventral surface layer, but
in Z. tournefortii half-width cells are found equally in both surfaces. The differences
between these two species are accentuated in longitudinal section. The contrast between
the dorsal and ventral surfaces of Z. harveyana is then obvious, whereas it is equally
obvious that there is no such contrast in the surfaces of Z. tournefortii (Fig. 4). In
Z. subarticulata (Fig. 5) there is no surface differentiation visible either.
Fig. 4 and 5. Fig. 4. — L.S. of part of thallus of Zonaria tournefortii. Fig. 5. — L.S. of part of thallus
of Zonaria subarticulata passing through a transverse ridge.
N.F. — new frond; O.F. — old frond.
197
Many species of the Dictyotales have in their fronds narrow transverse bands
of rather heavily pigmented tissue often associated with a growth periodicity, each dark
band marking the termination of a period of continuous growth. The three species
at present being discussed have such markings, but in Z. harveyana they are constituted
differently from those of the other two species. In Z. harveyana the tissue of a new
period’s growth has its origin in a narrow transverse zone of ventral surface cells lying
immediately behind the anterior margin of the tissue of the previous period’s growth.
This meristem gives rise to a ventral hump of tissue by periclinal division of its cells
which first become greatly eleyated above the surrounding surface (Fig. 3). Transverse
division in this tissue ensues and then becomes confined to the cells in its anterior
margin, thereafter extending the length of the frond. At this stage the anterior margin
of the previous period’s growth plays no part in the extension of the frond and it
survives as a dorsal transverse band or ridge (Plate lA) which is rendered visible by the
accumulation of pigment.
In Z. subarticulata a new period’s growth is initiated and maintained by transverse
divisions of part of the tissue of the anterior margin of the old growth. There is no
further activity of the anterior cells of either surface in this region, however, and on
becoming heavily pigmented these cells fomi the transverse bands (Fig. 5) seen on
both surfaces. The transverse bands of Z. tournefortii are visible only from the
surface; in longitudinal section they cannot be recognized at all.
Lastly, consistent with its dorsiventral character, the sori of Z. harveyana occur
on only one surface, namely the ventral. Both Z. tournefortii and Z. subarticulata,
however, show no such surface discrimination in their fertile fronds.
Other observations made on the sori of Z. harveyana were that each is covered
by an indusium (Plate IB) and that the fertile cells seen had undivided contents. The
latter observation suggests that these cells might have been oogonia. A further feature
of the fertile cells was that each had a special stalk cell (Fig. 3; Plate IB).
Discussion. — The genus Zonaria in South Africa, as constituted by the three species
recorded by Stephenson (l.c.) exhibits two types of basic structure, the one dorsiventral,
the other not. Obviously the two groups should have separate generic status, as in
fact they once had when Z. harveyana was included in the genus Homoeostrichus J.
Ag. The latter genus, however, contains four species and the type species was not
indicated in the original diagnosis (J. Agardh, 1894). Of the four species only Z.
harveyana {Homoeostrichus harveyana) occurs in South Africa. Because no mention
of dorsiventrality or any other character for a valid separation of Zonaria and
Homoeostrichus was made in their diagnoses no conclusion can at present be reached
as to whether Z. harveyana should be included in Homoeostrichus.
Literature Cited
Agardh, J. G. (1894) Anal. Algol. Contin. 1: 15.
Papenfuss, G. F. (1944) Notes on Algal Nomenclature. III. Farlowia, 1(3);
337-346.
Papenfuss, G. F. (1952) Notes on South African Marine AlgaeTII. J. S. Afr. Bot.,
17 (4); 167-188.
The Constitution of the Intertidal Fauna and Flora of South
Africa. III. Ann. Natal. Mus., 11 (2): 207-304.
Stephenson, T. A. S. (1948)
199
Plate !. — Zonaria hatveyana. A: L.S. of frond through junction of old and new growtn, B. L.S. of
frond through a sorus. Scale: 1 div. = lO/x.
4846586
200
201
INTRODUCTION
The Problem
In 1929 Henrard published an excellent Monograph of the genus Aristida in which
many of the difficulties previously encountered with the identification of the species
were solved. Henrard also gave an improved subdivision of the genus into sections
which facilitated the identification and gave a much clearer picture of relationships
of the species. In 1936 Theron studied the anatomy of a large number of the South
African species of Aristida and divided the genus into a number of groups based on
anatomical characters. His grouping differs considerably from the sections recognized
by Henrard. Schweickerdt in 1941 aligned all the available South African material
of the genus with Henrard’s Monograph and introduced a number of corrections and
improvements. In Schweickerdt’s review of the subdivisions of the genus the results
obtained by Theron are discussed and it is remarked that if the anatomy leads one to
put such diverse elements together, a classification based on anatomy produces “ a
highly artificial classification which shows no advantage over a natural classification
based solely on organographic characters ”. In recent years practically all workers
who have used anatomical characters as additional aids in producing a “ natural”
classification for the grasses have achieved gratifying results. Therefore, a re-assessment
of the anatomical features of Aristida seemed to be necessary to determine whether
there is a stronger correlation than is indicated by Theron between organographic and
anatomical characteristics.
In the course of a previous study it was noticed that some of the species occupy
an anomalous position in the sections in which they have been placed, e.g. A. sericans
is the only species with plumose awns in the section Chaetaria.
Since some species of the genus Stipa, as treated in the Flora Capensis (Stapf,
1898) and in the recent Grasses and Pastures of South Africa (Chippindall, 1955)
strongly resemble members of the genus Aristida except for the single awn, a fact
already pointed out by Stapf (1898, p. 574) and by Chippindall (1955, p. 291), the latter
on information supplied by the present author, it was decided to include the South
African Stipeae in this study.
Hubbard and Vaughan in Grasses of Mauritius and Rodrigues (1940, p. 23) removed
Aristida from the Stipeae. In order to accommodate this genus they created a new
tribe, the Aristideae. A study of the differences between the tribes Stipeae and Aristideae
has, therefore been undertaken.
In addition to the organographic features, the anatomy of the embryo and leaf-
blade, as well as cytological features have been used, wherever possible, in an attempt
to determine the relationships of the tribes and genera.
Reliable keys for the identification of vegetative specimens of grasses are urgently
needed. The difficulties of drawing up such keys are well known. Since some of the
tribes, and even genera and species, of the Gramineae can often be distinguished
readily on anatomical grounds it was decided to explore the possibility of combining
vegetative and anatomical characteristics in such keys.
The results of these investigations are set out in the following pages.
202
Delimitations
This study was limited mainly to the investigation of the South African species.
However, where more information was required to establish the characteristics of
taxa of higher rank, such as tribes and sections, a selection of extra-South African
species was also investigated. Wherever possible the type species of genera and sections
were studied. These species are not necessarily referred to in the text.
The anatomy of the leaf-blade of all the known South African representatives
was examined and in addition, where desirable, also the embryo, awn and shoot.
The karyology of all species of which living material could be obtained was
investigated.
Definitions
The following anatomical concepts are not in general use and for this reason
are defined here.
Vascular bundle unit: the vascular bundle as viewed in cross section including the
ad- and abaxial stereome strands, chlorenchyma and epidermis. The outline of these
units is determined ad- and abaxially by the epidermis and laterally by imaginary
vertical lines through the motor cell groups to the abaxial epidermis. In leaves where
the units do not protrude the outline would be more or less rectangular.
First order vascular bundle: possessing proto- and metaxylem vessels; lysigenous
cavity usually present.
Second order vascular bundle: possessing only metaxylem, with 2 large vessels, neither
protoxylem nor lysigenous cavity present.
Third order vascular bundle: xylem and phloem not clearly demarcated, or at least
large metaxylem vessels absent.
Silicified cell zone: longitudinal area of the leaf epidermis as seen in surface view,
situated above a stereome strand, and containing silicified cells.
Stomatal zone: longitudinal area of the leaf epidermis containing stomata, and bounded
laterally by silicified cell zones.
Explanation of Symbols
The symbols used in the anatomical drawings to indicate different types of cells
and tissues are explained in the following table. They are arranged alphabetically.
AB, abaxial epidermis
AD, adaxial epidermis
BH, bicellular linear hair
C, chlorenchyma
CC, suberized cell
CS, constriction
ICS, inner chlorophyll-bearing bundle
sheath
IS, inner bundle sheath
LC, long ripple-walled cell
MB, bundle of the midrib
MC, motor cell
OH, one-celled hair
OS, outer chlorophyll-bearing bundle
sheath
RB, retrorse barb
S, mechanical tissue or stereome
SC, silicified cell
SP, supporting parenchyma
ST, stoma
T, multicellular epidermal gland
IVB, first order vascular bundle
2VB, second order vascular bundle
3VB, third order vascular bundle
203
Citation of Specimens
A selection of typical representatives are cited for most species. These were chosen
mainly from collections made after the publication of Schweickerdt’s work on Aristida
(1941) and can be regarded as amplifying citations given in the- latter work. For the
species not mentioned by Schweickerdt the specimens studied are cited in full. Similarly
type specimens are indicated only when information about these is not available in
Schweickerdt’s work. Except where indicated otherwise, specimens cited are deposited
in the National Herbarium, Pretoria (PRE).
Acknowledgement
The author wishes to express his grateful appreciation to Prof. H. G. Schweickerdt
for guidance and encouragement throughout his studies. He is indebted to the Chief
of the Botanical Research Institute, Department of Agricultural Technical Services,
Pretoria, for providing the facilities that made this study possible, as well as to those
members of Iris staff who rendered assistance in various ways. To his wife he is deeply
indebted for her most valuable assistance in the preparing of drawings, reading of the
text and constant encouragement. The author is grateful to Mr. I. van der Westhuizen
for technical assistance, to Mr. J. P. H. Acocks for the use of a number of unpublished
distribution maps, to Miss A. Walters for a number of the line drawings and to Dr.
D. J. B. Killick for reading the text.
MATERIALS AND METHODS
Anatomy
Anatomical work on the various species was limited to the leaves, shoots and
caryopses since it is mainly these organs which afford information of taxonomic value.
The material utilized was taken from herbarium sheets, but wherever possible fresh
materia] was used to check on results obtained with the dry material. The dry material
was first soaked in a mixture of equal parts of 50 per cent alcohol and pure glycerine
for a few days, washed in water, and the silica removed by immersion in 25 per cent
hydrofluoric acid for about 1 hour. Sections of the leaf-blade and shoots were cut
by hand in 70 per cent alcohol, other methods having proved to be either too time-
consuming or not very successful due to the hardness of the material. After sectioning,
the material was taken down to water in two stages, 50 per cent, 25 per cent alcohol,
and then transferred to full strength commercial “ Javel ” for a minute or two, which
had the effect of bleaching the cell contents and stretching the sections. After wasning
thoroughly, the sections: were stained by the double-stain technique developed by
Evans (1949, p. 191), and made permanent by the usual procedure using Canada
Balsam as mounting medium. For determining the distribution of chlorenchyma in
the leaf, the bleaching in Javel was omitted, but otherwise the same procedure was
followed. Epidermal preparations were made by placing a leaf-blade with its lower
epidermis in contact with a glass slide, and scraping away the tissue until only the lower
epidermis remained. Prior removal of the upper epidermis, followed by a gentle
maceration in “ Javel ” for about 30 minutes, not only facilitates scraping but results
in cleaner preparations. Staining with safranin yielded good results. Caryopses were
sectioned by freezing microtome. In order to orientate the grains, a small thin square
of agar was placed on the holder and the grain manipulated into the desired position,
i.e. with the grain lying on its side and the hilum parallel to the surface of the agar.
Thus orientated on, and held in position by the agar, the grain may be frozen and
sectioned. By this method reasonably good longitudinal sections of the embryo wree
obtained. The sections were double-stained with safranin and haematoxylin (after
Evans, 1949, p. 191) dehydrated and permanently mounted in Canada Balsam.
204
The silica inclusions of leaves of Aristida and Stipagrostis species were prepared
for study by boiling leaves in concentrated sulphuric acid, while adding drops of nitric
acid, until all organic matter was dissolved, i.e. until the liquid became quite clear.
This liquid was subsequently centrifuged, the acid poured off, and the residue washed
in order to remove the acid. Permanent mounts of the silica residue in Euparal were
microscopically examined and photographed, under phase contrast.
For all material used in the preparation of drawings the collector’s name and
number is cited in the relevant text. Except where otherwise stated all specimens
used are preserved in the National Herbarium, Pretoria (PRE).
Karyology
Only root tip material was used for determining the chromosome numbers. This
was collected mainly from plants growing under natural conditions. The best results
were obtained with root tips gathered on sunny days from vigorously growing plants,
between the hours of 10 a.m. and 1 .00 p.m., one to two days after soaking rains had
fallen. Tips from thick young roots 1-3 cm long were found to be superior to those
from older and longer roots. Due to the toughness of even young roots, squashing
techniques proved to have no advantages over the standard techniques of embedding
in wax and sectioning. Sections were cut at a thickness of 14/x. The best fixative
was found to be Randolph’s chromo-acetic formalin (CRAF) (Conn & Darrow, 1946,
p. I2-I2). Stockwell’s modification of Flemming’s Triple Stain gave good results
(Stockwell, 1934, p. 121).
Well-spread metaphase plates were drawn with the aid of a Camera Lucida and
for each plant investigated several counts were made.
Herbarium specimens of all the individuals examined karyologically are deposited
in the National Herbarium, Pretoria (PRE).
205
PART I
THE TRIBE STIPEAE
1.1. History
The tribe Stipeae was described by Nees von Esenbeck (1829, p. 371) based on
the genus Stipa L. and including Chaetaria, of which he considered Aristida to be a
synonym. In Stipa, Nees included Oryzopsis, Piptatherum and Urachne as synonyms.
Four years later Kunth (1833, p. 187) enlarged the tribe by including Streptachne but
redivided the comprehensive genus Stipa sensu Nees into five genera : Stipa, Lasia-
grostis, Macrochloa, Oryzopsis and Piptatherum. Trinius & Ruprecht (1842, pp. 1
& 97) still further enlarged the concept of the tribe by including Dichelachne and
Anisopogon. They divided Stipa into five sections and regarded Urachne and Lasia-
grostis as distinct genera, each with two sections; in addition, Aristella and Ortho-
raphium were upheld as genera distinct from Stipa. Further additions were made
by Bentham (1882) and this tendency culminated in Hackel’s treatment of the tribe
(1896, p. 100-104). Here twelve genera were included although many genera formerly
recognized were relegated to synonymy and thus a very wide concept of Stipa was
reintroduced. In conformity with the recent trend of dividing the grasses into smaller,
more “ natural ” tribes, most present-day authors have again severely restricted the
number of genera included in the Stipeae. In the introduction to his monograph
of Aristida, Henrard (1929, p. 13), summarized his concept of the Stipeae in a key
to the genera namely Stipa, Oryzopsis, Nasella, Piptochaetium, Aciachne, Milium,
Aristida and Timouria. Hubbard (1934, p. 214) treated the tribe essentially in the same
manner, but did not mention Timouria (apparently including it in Stipa).
Avdulov (1931 , p. 406) in his classic work on the karyosystematics of the Gramineae
pointed out that the genera Stipa and Oryzopsis are more closely allied to genera such
as Phaenosperma, Oryza and Ehrharta than to the genera of the tribe Agrosteae in
which they are placed by some authors. He came to this conclusion mainly, not only
on the basis of karyological similarity, viz. small chromosomes with basic number 12,
but also on the basis of the presence of certain primitive features exhibited by the
florets, e.g. 3 lodicules in Oryza and Stipa, and 6 stamens in Phaenosperma, Oryza
and Ehrharta. Anatomically the leaf-blades of these genera also show a great deal
of similarity. Avdulov pointed out, however, that the genera probably evolved
independently and thus represent distinct evolutionary groups.
The fact that Aristida occupied an anomalous position in the Stipeae was realized
by Hubbard (Hubbard & Vaughan, 1940, p. 23) who removed this genus from the
Stipeae to accommodate it in his new tribe, the Aristideae. (Latin diagnosis published
by Hubbard in Bor, 1960, p. 685). Eight years later Hubbard (1948, p. 325) transferred
Milium from the Stipeae to the Agrosteae, but did not list the genera to be retained
in the Stipeae since only British grasses were dealt with in the relevant work.
An important, though lesser known, work on the Stipeae, is that of Elias (1942).
His work is unique in the Gramineae in that he made a comparative study of the fossil
and living forms of the Stipeae and by this method was able to indieate certain evolu-
tionary trends of the tribe. In order to enable him to compare fossil remains with
living species of Stipa, Oryzopsis, Piptochaetium and Nasella, Elias made a thorough
study of the morphological characteristics of the “hulls” of these grasses. Where
possible he subdivided the genera into sections, each section being composed of a
group of closely related species. In addition to the four genera mentioned above,
Elias devoted several pages to a discussion of the relationship of Stipa to Aristida
both from a morphological and an anatomical point of view. He arrived at the
conclusion that Aristida represents a type more advanced than that found in any of the
206
other genera under consideration. He stated that Aristida probably evolved from
Stipidium (a fossil genus) through Stipa, summing up his views as follows : “ The
comprehensive modem genera Stipa and Aristida of Stipeae, which between themselves
differ primarily in the division of the single strong awn of Stipa into three slender
awns in Aristida, are apparently descendants of a single fossil genus Stipidium with a
non-indurated awn The reasons why the present author cannot support Elias’s
views will be elaborated under the discussion of the relationship of Aristida to Stipa,
as seen in the light of recent advances in the knowledge of their anatomy, karyology
and embryology.
Parodi & Freier (1945), who studied the Stipeae both anatomically and taxonomi-
cally, include Aristida, as a distinct natural genus, in the Stipeae. Furthermore they
regard the genera Stipa, Nasella and Oryzopsis to be so closely related that the
possibility of accommodating them together as taxa of infrageneric rank cannot be
precluded. Aciachne is regarded not to be a close relative of the foregoing and the
view is expressed that it probably represents a distinct subtribe of the Stipeae whereas
the taxonomic position of Milium and Relchella in relation to the above is not yet
clear.
In a recent account of the Stipeae, Roshevitz (1951) divided the tribe into three
subtribes and included Streptachne and Amphipogon, as well as resuscitating several
genera such as Lasiagrostis, Macrochloa, Orthoraphium, Timouria and Ptilagrostis
which are placed in synonymy under Stipa by most present-day workers. On account
of the existence of a large number of species somewhat intermediate in character most
of these genera are doubtfully distinct from Stipa. In addition Amphipogon and
Streptachne could be placed more satisfactorily in other tribes. Their taxonomic
position will be discussed in the course of this paper.
The most recent account of the tribe is that of Pilger (1954, p. 330-332) who again
included Dichelachne, Milium and Aciachne, recently removed from the tribe by other
authors. His reason for doing so was probably motivated by a distrust of the
present-day tendency to over-emphasize the taxonomic value of anatomical features
of the vegetative organs thus to some extent under-estimating the importance of the
morphology and internal structure of the reproductive organs. Pilger’s classification
consequently is more artificial than one would have expected, i.e. if all data available
had been evaluated in determining relationships. Finally a new genus of the Stipeae,
Trikeraia, based on Stipa hookeri Stapf, was described by Bor (1954, p. 555).
1.2. Delimitation
From the foregoing outline of the historical evolution of the Stipeae it is evident
that, with the exception of Pilger, modern authorities have tended to support a rather
narrow delimitation of this tribe. It is the present author’s view that the tribe should
be limited to the genera Stipa, Oryzopsis, Piptochaetium and Nasella. Trikeraia Bor,
a segregate from Stipa obviously also belongs to the Stipeae. These genera agree
closely in anatomical, karyological and embryological characters as well as in the
general morphology of the spikelet. These characters are; the diffusely arranged cells
of the chlorenchyma which is continuous between the bunoles (Prat, 1936); the lack
of bicellular hairs of the epidermis (Prat, 1936); medium-sized chromosomes with
basic numbers 6, 11, 12 (Brown, 1951; Stebbins & Love, 1941; Darlington & Wylie,
1935); the presence of an epiblast; the characteristic course of the vascular strands
in the embryo; the absence of a cleft between the coleorrhiza and base of the scutellum;
the first leaf (sheathed by the coleoptile) with the margins not overlapping (Reeder,
1951); the one-flowered spikelet with a usually 5-7 (rarely 3)-nerved, single-awned
(rarely with a long central awn and two smaller lateral awns), indurate or rarely hyaline,
lemma clasping the caryopsis; the large membranous lodicules, usually 3 (rarely 2)
in number; the linear hilum, the usually small embryo and the membranous ligule.
207
The following genera were, at some time or other, included in the Stipeae; they
all, however, possess certain characters which do not conform to those of the typical
representatives of the tribe and have therefore been excluded.
Milium. — (Henrard, 1929; Bentham, 1882; Darlington & Wylie, 1955; Hubbard,
1959; Hackel, 1896). Lemmas awnless, chromosomes in multiples of 4, 9 or 14 and
much larger than in the Stipeae. Probably — as done by Hubbard (1948) — best placed
jn the Agrosteae on account of the large chromosomes.
Aciachne. — (Henrard, 1929; Bentham, 1882; Hubbard, 1959; Hackel, 1896).
On morphological and anatomical grounds regarded as probably representing a subtribe
of the Stipeae (Parodi & Freier, 1945). This genus is in need of further investigation.
Muehlenbergia, Lycunis, Perieilema. — (Hackel, 1896); spikelets 1- or 2-flowered;
glumes persistent, shorter than the lemmas; lemmas 3-nerved, practically awnless;
hilum punctiform and basal in Muehlenbergia and Perieilema and probably likewise
in Lycurus (no fruiting material seen); embryo as in the Paniceae except for the
embryonic leaf (i.e. sheathed by the coleoptile) which has the margins overlapping
(Reeder, 1957); chlorenchyma of the leaf-blade radially arranged round the bundles
(Avdulov, 1931; Schwabe, 1949); two-celled epidermal hairs recorded for Muehlen-
bergia and Lycurus (Tateoka, 1959); chromosomes small, in multiples of 8, 10 or 14
(Darlington & Wylie, 1955). Darlington & Wylie (1955) place Muehlenbergia and
Lycurus in the Sporoboleae, while Reeder (1957) places them in the Chloridoid-
Eragrostoid group (which includes Sporobolus), mainly on the characters of the embryo.
Pilger (1954) places all three genera in the tribe Eragrosteae; Perieilema and Lycurus
in the subtribe Lycurinae, and Muehlenbergia in the subtribe Muehlenbergiae. Schwabe
(1949) likewise suggests that Lycurus and Muehlenbergia should be placed in the
Eragrosteae. On the available evidence this seems an excellent way of expressing
their affinities.
Brachyelytrum. — (Bentham, 1882; Hackel, 1896). Rhachilla produced beyond
the base of the floret; grains flattened and grooved ventrally; embryo “ festucoid ”
except for the cleft present between the base of the scutellum and the coleorrhiza;
chromosomes always in multiples of 11. Placed by Pilger in the Festuceae. Reeder
(1957) however, doubted whether it belongs to this tribe and suggested possible
affinities with Oryza and Stipa.
Dichelachne. — (Trin. & Rupr. 1842; Pilger 1954, 1956). Very little information
regarding this genus is available. The awn, however, originates from the back of the
lemma as in the Aveneae and Agrosteae and, for this reason, the genus should not be
included in the Stipeae. Hubbard (1959) and Gardner (1952) place Dichelachne in
the Agrosteae.
Aristida and Amphipogon. — For discussion see page 221.
Streptachne. — (Henrard, 1929; Kunth, 1833; Trin. & Rupr., 1842; Roshevitz,
1951). Transferred by Henrard (1929) to Aristida as the section Streptachne which
accommodates the single-awned species.
1.3. Phylogenetic Position
At the end of the sub-family Festucoideae, Pilger (1954) lists several tribes, which
he apparently regarded as having no clear affinities. Here he placed the Stipeae between
the Pappophoreae and the Nardeae. Studies of the anatomy and embryo of the Pappo-
phoreae have shown that the latter probably is closely related to the Eragrosteae, and
cannot, therefore, be regarded as allied to the Stipeae. It is possible that the Nardeae
is correctly placed in the Festucoideae, but the Stipeae is probably misplaced in this
sub-family. In his scheme of relationships of the tribes of the Gramineae, which is
208
based largely on leaf anatomy, Brown (1958, p. 173) placed the Danthonieae and
Stipeae adjacent to each other, and indicated that the affinities of these tribes are still
uncertain. Stebbins (1956, p. 894), in a diagram showing the evolutionary inter-
relationships of the principal sub-families and tribes, places the Stipeae as an isolated
group and regards it as probably an offshoot of the Danthonioid-Arundinoid complex.
The Stipeae differs mainly from the members of the Danthonioid-Arundinoid complex
in having strictly one-flowered spikelets, an indurate lemma tightly enclosing the grain,
a terete not flattened awn, a linear hilum and in the internal structure of the embryo.
The Stipeae differs from the Festuceae in the one-flowered spikelet, indurate lemma
tightly clasping the grain and the small chromosomes usually in multiples of 6, 11 or
12; and from the Aveneae in the indurate lemma tightly clasping the grain, the small
chromosomes and the awn which originates from the apex and not from the back of
the lemma.
From the foregoing it is evident that until more detailed information concerning
the characteristics of the different tribes is available, the relationship of the Stipeae
to other tribes cannot be determined and established with certainty. As indicated by
Stebbins (1956) and by Brown (1958), it will probably prove to be an isolated group
with no close affinity to any of the other tribes.
In an excellent paper on the tertiary prairie grasses of the High Plains of North
America, Elias (1942, p. 68) traced the relationship of some of the “ Stipoid ” genera
by medium of fossils found in Miocene and Pliocene deposits. The genus Stipa, it
is postulated, has evolved from two different fossil genera namely Parastipidiwn and
Stipidium, the former having given rise to the 6 crownless sections of Stipa, whereas
the latter is considered to be ancestral to the 3 sections possessing crowns. The tribe
Stipeae thus presents one of the rare cases in the Gramineae where the pre-liistory is
known to some extent.
1.4. Distribution
The Stipeae in the concept of the present author is restricted mainly to the temperate
and warm temperate regions of both hemispheres, only rarely occurring in the tropics.
The majority of the species, are typical dry grassland inhabitants, however, a few are
found in forests or moist habitats. In the Northern Hemisphere the centres of greatest
concentration of species occur in North America, the Mediterranean area, and the
Russian Steppes. In the Southern Hemisphere the most important areas of distribution'
are South America, Australia and Tasmania.
Among the component genera of the Stipeae the genus Stipa is the largest and also
the most widespread. Many species, particularly those of the Southern Hemisphere,
are typical constituents of desert grassland. The Australian representatives, however,
form a rather distinct group, differing in a number of minor characters from the species
occurring in other parts of the world.
Oryzopsis is a small genus confined to the Northern Hemisphere. It is closely
related to Stipa.
Nasella, regarded by Parodi as a section of Stipa, occurs in Peru, Chile and the
Argentine.
Piptochaetium, a genus of about 21 species, is confined to southern South America
with the exception of one species found in the United States. This genus was revised
recently by Parodi (1944, pp. 244-310) who described six phyto-geographical areas of
distribution for South America, ranging from the high mountain grassland of the
Andes to the lowland steppes of the Argentine.
209
1.5. STIPEAE
Agrost. Bras. 371 (1829)
Annual or perennial herbs, frequently with tough, rigid cuims. Ligule membranous
or membranous and fimbriate, obtuse. Leaf-blade narrow, rolled and wiry, or fairly
wide and expanded. Spikelets all similar, bisexual, \-flowered, arranged in contracted
or open panicles. Rhachilla disarticulating above the glumes, not produced beyond
the fioret [“ occasionally present as a wart-like projection at the base of the palea ” :
Elias (1942)]. Glumes persistent, the upper (and often the lower) as long as or longer
than the florets. Lemma usually terete, with convolute or involute margins, rarely
hyaline and thin, 3-7 (usually 5)-nerved with the nerves anastomosing at the apex,
awned from the entire or minutely 2-lobed tip; the awn apical, single, or rarely with
two additional, shorter, lateral awns, three of the nerves of the lemma passing into
the awn; callus short or long, obtuse or acute, bearded. Lodicules 3 (rarely 2).
Caryopsis tightly embraced by the lemma and palea; hilum linear, at least half tlie
length of the grain; embryo the length of the grain (Fig. 2, F and 4, D).
Anatomy
Leaf-blade expanded, folded or semicircular. Chlorenchymd continuous between the bundles,
not arranged in a definite pattern, cells circular or angular. Bundle sheaths two: inner of small cells
with thickened walls; outer or large more or less circular thin-waUed cells furnished with chloroplasts,
gradually merging into the adaxial stereome or interrupted by the fibres of the stereome both ad- and
abaxially. Epidermis of a relatively simple type : long ripple-walled elements generally rather shallowly
undulate, alternating with paired short elements; silicified cells variable, often dumb-bell-shaped
(many species with elongate, square or sub-circular silicified cells), always accompanied by a suberized
cell; unicellular hairs of various types present, two-celled hairs absent; stomata present on the adaxial
surface only, or present on both surfaces.
Awn. — Column angular to sub-circular in cross-section, occasionally grooved on one side, containing
a large central vascular bundle flanked by 2 groups of chlorenchyma cells, each subtended by a smaller,
lateral vascular bundle; central bundle surrounded by small or large much-lignified cells; small vascular
bundles in close contact with the chlorenchyma on one side, and supported by usually small much-
lignified cells on the other; chlorenchyma groups situated just beneath the lateral epidermis; epidermis
opposite the chlorenchyma groups containing stomata; seta agreeing anatomically with the column,
but the lateral vascular bundles, as well as the layers of lignified cells surrounding the central bundle
disappearing towards the apex, and the groups of chlorenchyma cells converging and finally contacting
the central bundle.
Caryopsis. — Starch grains compound; embryo usually with a well-developed epiblast (occasionally
small but always present); no cleft present between the coleorrhiza and the lower part of the scutellum ;
vascular strand diverging into the scutellum directly below the coleoptile; first leaf (sheathed by the
coleoptile) with the inflexed margins not overlapping (Fig. 2, E and 4, E, F).
References. — Duval-Jouve, 1875; Kennedy, 1899; Holm, 1901; Hughes, 1921; Prat, 1936;
Parodi, 1944; Reeder, 1957.
Karyology
The chromosomes are small and occur in multiples of 9, 10, 11, 12, 14, 16 and 17. Stebbins (1941 ,
p. 379) suggests that most of the numbers found in the aneuploid series of chromosome numbers in
Stipa probably could have been derived from the basic numbers X = 11 and 12. Consequently these
two basic numbers probably could also account for the numbers found in the other genera as well,
most of them being multiples of 11 and 12.
Brown & Emery (1957) report for Stipa leucotricha Trin. & Rupr. that in the mitotic division the
nucleolus disappears before metaphase. They found the presence or absence of nucleoli persisting
to metaphase to have some bearing on classification in the Gramineae, since the nucleolus disappears
before metaphase in the subfamily Festucoideae and in a part of the group Phragmitiformes Avdulov,
whereas it persists to metaphase at least in a large percentage of cases in the subfamily Panicoideae
as well as in the Oryzeae, Unioleae and Bambuseae of the group Phragmitiformes.
Genera: Stipa, Trikeraia, Oryzopsis, Nasella, Piptochaetium.
References. — Avdulov, 1931; Hunter, 1934; Stebbins & Love, 1941; Brown, 1951; Darlington
& Wylie, 1955; Brown & Emery, 1957.
210
1.6. Delimitation of the Genera
Stipa is the largest of the genera included in the tribe and also the most variable,
as testified by the nine sections into which it is divided by Elias. Althougn Oryzopsis
is retained as a distinct genus by practically all authors, it is closely allied to Stipa
and some species are intermeaiate between these two genera [Clements ex Elias (1945,
p. 71)]. It is for this reason that, at different times authors have placed one and the
same taxon either in Oryzopsis or in Stipa.
Nasella a genus with relatively few species is regarded as distinct by Hitchcock
(1950, p. 443) and others, but as a section of Stipa by Parodi. Parcdi & Freier (1945)
studied the genera anatomically as well as organographically and at present their key
to the genera appears to be the best available. Parodi & Freier found that Nasella,
Oryzopsis and Stipa could not be separated solely on anatomical grounds, whereas
Piptochaetium is a genus both organographically and anatomically distinct.
Trikeraia Bor (1954, p. 555) differs from Stipa in the hyaline, not indurated lemma,
furnished with two short lateral bristles, formed by the elongation of the lateral nerves
of the lemma, in addition to the stout median awn typical of the Stipeae. Most likely
Trikeraia is correctly placed in the Stipeae. Metcalfe (1960, p. 504) described the
anatomical features of the genus.
Though it seems fairly certain that a greater degree of uniformity has been obtained
by the exclusion of anomalous genera from the Stipeae, it is, nevertheless evident
that in the classification of the genera retained in the tribe progress is improbable
unless a monograph of the group as a whole is undertaken. There are indications
that anatomical studies, particularly of the awn, will be of considerable value in arriving
at a classification more natural than the one existing at present. In future work this
aspect should be taken into account.
1.7. THE GENUS STIPA
1.7.1. Delimitation
It was decided that for the purpose of this study it would be more satisfactory
to follow Elias in retaining Stipa as a comprehensive genus subdivided into sections,
than to accept the large number of ill-defined genera upheld by Roshevitz and, to some
extent by Pilger. The South African representatives are therefore retained in Stipa
and provisionally placed in the appropriate sections as defined by Elias (1942). It is
realized that a re-assessment of Elias’ treatment of the sections may necessitate nomen-
clatural changes, but such a study does not fall within the scope of the present paper.
It may be pointed out here that Roshevitz (1951), in spite of dividing the Stipeae
sensu stricto (i.e. with the exclusion of Amphipogon and Streptaclme) into eleven
“ genera ”, did not achieve a natural grouping in all respects. For instance, he included
the African Oryzopsis keniensis Pilg., a taxon indistinguishable from the South African
Stipa dregeana Steud. var. elongata (Nees) Stapf, in the genus Piptatherum. Undoubtedly
this plant is more closely related to some Asian and North American species of Stipa
placed by Roshevitz in the subtribe Stipiinae, than to Piptatherum coerulescens (Desf.)
Beauv., the type of the genus Piptatherum, which belongs to the subtribe Timouriinae.
Only two sections, Ptilagrostis and Tortilia, both possessing crownless lemmas,
occur in South Africa. Ptilagrostis consisting of several species is widely distributed
in North America, Europe and Asia with one representative, viz. Stipa dregeana Steud.,
in Africa. The monotypic section Tortilia is based on Stipa tortilis Desf. described
from North Africa, but also occurring in South Africa. For reasons of priority
S. capensis Thunb. supersedes S. tortilis so that the latter must be placed in synonymy.
211
« 1) 2
s.s ^
^<oi>5
a T3 ^
, .9 «
•S^ =
2 9
•o'ts
Me®
.9 =« U O. u,
so "a 3.-e
■®.9
^•4-1
x: o
a> JS
-5)
J u g
r
p > 1 =
>0 i:
TS D ^
0)
^00
c c c
□ c3*z;
to.S2
S? D.Ci
.S o M
■g Jj o
o
^ o O.
i"® V-
“■g ">' 3?
.S Cl'S
c/5 g G
® e-®'
C IH
, O g g -
: g-G ^
: S 2_g
.G ^ g
D. O
C u
3 <L) ^
= s-s
-2i o‘>
3 c/3 i>
Sx-o O
^ (L) ^ <L>
^ X.X X
G G
J> ^ §
o o o
G G
X
„ 6 u 2^^
g 3
e o.® CK.b
u C G r/K o
ag S — -®
® 5 2 ®
va a A GO
So ®
.9 £
q 1)
9
e m
® M
oj S
Sag
_c C °
do ^ ®
U TD
® > O
■- 1 s
Tj <u
X) ■“
£ - 15 ■« !
9: O q.c O-^
^ U ® T-J 77
u
^00
§ £•£
'§1.^
9'°’^
G <U
'5 n
X» o
fS -n cd c/3
X O
W 0->
3 c/3 >
g p
H ^ ^
^ (D <U
1) fS «U N
G’S'O^
•so ®
U CD <L)
cej c/3 X
35 g p
®S)T) ®
g) ^
.G C ^ ^
c3 G 5
OJ i:go
.S ^ o
iH
®= i ^
®.2 e"
Sh "O ^
"•1.
. rG-— qj
<1> ^ G.
o, a
cS ■
>»
G 0)
' X
G T3
C « c
s S §
o 0 P
ri ® 'f
G (L) C/3 '—'
iiXg
2
^ <L> 2 £
: c« G
X *0
.9°
CJ) a>
c • ^ ^
•— ^ X
G 2^^ ..
X tjv Qd ^
O M §“ s
g g p ^ ^
G G3 > TD ^ fli
X>c!" O c
go”® O'
g’5'3'
Q . bO ^ g
g "O C ^ C 33 o
2 S 4- G O ^
g c S ® 3734-
9.Am“^.ta2^
-C c
3 .
o c/3
G qj
G ^
•-G C-O
C G p,
OX
q^ -o
O qj
xS
bO
G
a .
g
. D CL
"X S'
X C« q> .
r' S -S ^
.B ^ ^
•X o e g:
00 g c g 05
x.ti a w;o
^ ^ o > 2
2 X
<L> GS
^ C3 r
G C«-^
•S s E
^ E G
X X
bfl<^
G
bb (D
S c§
.^X
^ cu
O . GX
^ ^_-G .G -*-'
5 a"^ ®
o p )=: X
: X
• 9
a> X
S'§ tB
g G X — '
0*0
G
OX c
^ X G . .
S j3^X3 _g
9 60 a
c >
3 i. ®rii
ww ^ <u
i a a>
9 •
'
C <N
=:£
a a
CD
G o O
G ^
-2 X
P D 'w
S G '
O* w
2 1> u
9 ■£ as
0^3
c3 X
X (D
_5 o
CL ..2
a> X
l-S E
^ g G
X X
^ -2 ^
'^5-3 rs Q ^ -C3
Il1|||^
.s ‘-j >-) ^
05
212
1.7.2. The South African Representatives
Stipa is the only genus of the Stipeae represented in South Africa. In the most
recent account of the South African Gramineae by Chippindall (1955) four species
are mentioned. A careful study of the organography and anatomy of these species
has, however, revealed that only two of them should be retained in the tribe Stipeae.
The use of the single awn as a generic criterion has proved to be quite unnatural. As
early as 1898 Stapf pointed out that Stipa parvula Nees probably represented a reduced
Aristida, in which the lateral awns have disappeared. Henrard (1929), however, over-
looked this and omitted the species from his monograph. Elias (1942) investigated
Stipa parvula organograpliically and concurred with Stapf’s view, but did not transfer
the species to the genus Aristida. Chippindall (1955) and de Wet (1960) likewise
remarked on the presence of misplaced species of Aristida in the genus Stipa (on infor-
mation supplied by the present author).
In Fig. 1 the four species under consideration are contrasted: S. parvula and
S. namaquensis exhibit all the characters regarded as typical of the Aristideae except
for the single awn, whereas S. dregeana and S. capensis are typical representatives of
the Stipeae. Stipa parvula and Stipa namaquensis have therefore been transferred to
the genera Aristida and Stipagrostis (a section of Aristida now given generic rank)
respectively (pp. 242 and 375). Stipa namaquensis is the sole representative of a new
section viz. Anomala of the genus Stipagrostis whereas Stipa parvula is placed in the
monotypic section Schizaclme transferred from the genus Stipa to Aristida (p. 236).
1.7.3. STIPA L.
Sp. PI. 78 (1753)
Spikelets solitary, pedicelled, borne in terminal contracted or open panicles.
Rhachilla disarticulating above the glumes, not produced beyond the base of the floret.
Floret 1, hermaphrodite, usually shorter than the glumes, linear in outline. Glumes
persistent, narrow, acute, acuminate or tipped by a bristle (rarely obtuse), 1-3 (rarely 5)-
nerved, equal or unequal. Lemma convolute, or with margins not overlapping at the
base, cylindrical or slightly dorsally compressed, indurated at maturity, glabrous or
hairy, 3-7-(usually 5)-nerved, nerves anastomosing at the apex, apex acute or minutely
2-lobed; awn persistent, single, straight, flexuous or once to twice-geniculate, column
twisted, glabrous or hairy, bristle glabrous, pubescent or distinctly plumose; callus
pungent or obtuse, short, or long bearded. Palea as long as or shorter than the lemma,
completely enclosed by, or visible between, the margins of the lemma near the base,
membranous or somewhat indurated, 2-nerved. Lodicules 2-3, equal and large, or
posterior reduced, usually rounded at the apex, glabrous, indistinctly nerved. Stamens
2-3, anthers elongated, sometimes penicillate at the apex. Ovary glabrous; styles
free; stigmas plumose, apically exserted. Caryopsis tightly enclosed by the lemma
and palea, free; hilum linear, slightly shorter than the grain; embryo (seldom ;^)
the length of the grain.
Densely tufted perennials or rarely annuals. Culms usually erect, branched or
simple; internodes hollow or solid. Leaf-blades usually long and narrow, often
convolute, more rarely flat and fairly wide. Ligule membranous, lacerated, or a mem-
branous ciliate rim. Panicle narrow and contracted, to eifuse and open.
Anatomy. — As described for the tribe.
Karyology. — As deseribed for the tribe.
Type Species: Stipa pennata L.
Name from the Greek “ stupe” meaning “ tow ” alluding to the plumose awns of
the type species.
213
A very large genus widely distributed in temperate and warm-temperate regions
ot the world.
1.7.4. Key to the South African Species based on Organographic Characters
Perennial with knotty rhizomatous base; callus of the floret very short, obtuse; awn 2-3 times
as long as the lemma, arcuate or once geniculate-, palea as long as the lemma, sparsely pube-
scent 1. S. dregeana
Annual; callus of the floret elongated, tapering to an acute point; awn at least eight times as
long as the floret, twice geniculate-, palea much shorter than the lemma, glabrous 2. S. capensis
1.7.5. Key Based on Anatomical Characters
Midrib of leaf-blade much enlarged and protruding abaxially, furnished with colourless parenchyma
surrounding the bundle; first order bundle units wider than deep, hardly protruding adaxially
1. 5. dregeana
Midrib of the same size as the other first order bundle units, without additional colourless paren-
chyma; first order bundle units broadly ovate in outline and protruding adaxially 2. S. capensis
1.7.6. Description of the South African Species
1. S. dregeana Steud., Syn. PI. Glum. 1: 132 (1855). Type: Albany, Hills near
Grahamstown, Drege s.n. (Bf; PRE, fragment of holo.!).
Fig. 2. — Stipa dregeana: A, diagram of leaf-blade in cross section; B, cross section of vascular bundles
of the leaf-blade; C, abaxial epidermis of the leaf-blade; D, cross section of the young
shoot; F., longitudinal section of the embryo; F, caryopsis showing the embryo; G. seedling
showing first leaf; H, chlorenchyma cells in longitudinal section (de Winter 7607).
214
Perennial with a short, knotty, branched system of rhizomes. Culms erect, usually
simple, 3-noded, smooth, glabrous, 90-120 cm high; intemodes enclosed in, or exserted
from the sheaths, terete, hollow, glabrous, striate; nodes conspicuous, often dark in
colour, glabrous. Lower leaf-sheaths reduced to elongated bracts tightly enclosing
the base of the culms, increasing in size upwards, usually smooth, striate. Ligule
membranous, up to 5 mm long, obtuse or lacerated, often with the margins produced
into auricle-like outgrowths. Leaf-blades hnear, flat, widest near the middle, up to
12 mm wide and 60 cm long, narrowed towards the base and acuminate, glabrous
or sparsely villous, firm in texture, smooth below and shghtly rough on the nerves
above, finely nerved with nerves numerous. Panicle effuse and wide, or contracted
and narrow, erect or nodding, when fully developed 15-40 cm long; branches very
finely scabrid, single, 2-3 nate or fascicled, short to long and flexuous, up to 15 cm long,
filiform and smooth. Pedicels scabrid, shorter or longer than the spikelets. Spikelets
pale green, 6-7 mm long, one-flowered, linear-lanceolate. Glumes sub-equal, 3-nerved,
lanceolate in outline, the lower broader than the upper, finely acuminate, menbranous
and hyaline in the scabridulous upper half. Lemma convolute, narrowly lanceolate-
cylindric, about 5-5 mm long, appressed hairy all over, 5-nerved, obscurely bi-lobed
with a stiff, straight or arcuate awn from the apex; awn persistent, scabrid, slightly
twisted below, usually bent about 2-4 mm above the base, 10-15 mm long; callus
very short, obtuse, shortly bearded. Palea subequal to the lemma, lanceolate, sub-
obtuse, not keeled, appressed hairy on the back, 2-nerved. Lodicules 3, oblong, obtuse,
the two anterior membranous, hyaline, the posterior smaller. Stamens 3, anthers
yellow, about 3-3-5 mm long, glabrous. Caryopsis spindle-shaped, brown, 3-4 mm
long; hilum linear, black, about | the length of the grain; embryo small, about ^
the length of the grain, no constriction present beteen the coleorrhiza and the coleoptile
(Fig. 2, F.).
Anatomy (Fig. 2, A, B)
Leaf-blade flat, very thin, wide, tapering gradually towards the margins; abaxial surface flat,
with very few hairs; adaxial surface flat except for the very slightly projecting first order bundle units,
with short unicellular hairs flanking the motor cell groups. Abaxial epidermis: Stomalal zones with
1-2 rows of stomata along each margin, very wide (12-15 rows of cells), long ripple- walled cells with
walls only slightly undulate, elongate and tapering towards the ends, alternating with pairs of short
215
elements ; silicified zones: silicified cells dumb-bell-shaped, often with a double constriction and usually
accompanied by a short or elongate thin-walled suberized cell with slightly undulate or almost smooth
walls, a few unicellular broad-based retrorse barbs present in this zone, bicellular hairs absent. Vascular
bundle units poorly differentiated, wider than deep, more or less rectangular; first order bundles 9-13
or more, alternating with groups of 2-4- second order bundles; bundle sheaths variously interrupted
or complete, outer sheath of large more or less circular thin-walled cells,, inner sheath of smaller cells
of which the inner tangential walls are strongly thickened. Chlorenchyma cells large, circular or angular,
forming a continuous tissue between the bundles ; in longitudinal section somewhat irregular in arrange-
ment and with intercellular spaces. Stereome strands present ad- and abaxially opposite all bundles,
strongly developed opposite the first order bundles and less so opposite second order bundles, strands
narrow abaxially, wider adaxially especially opposite first order bundles; all strands consisting of
strongly thickened fibres with fairly large cell-lumen. Motor cells fairly well differentiated, all epidermal
in origin. Midrib strongly projecting abaxially; midrib bundle surrounded by large thin-walled paren-
chyma cells (containing no or very few chloroplasts) additional to the parenchymatous outer sheath;
inner sheath consisting of two layers of cells with strongly thickened walls whereas the other bundles
have a single inner sheath. Margin subacute.
Awn: (Fig. 8, A & B). Column roughly triangular in outline with rounded angles, ventral side with
a shallow depression in sections cut near the foot, furnished with three vascular bundles; central bundle
large and surrounded by large cells with very strongly asymetrically lignified walls, of which those
situated on the ventral and dorsal sides of the bundle much exceed the others in size; lateral bundles
small and situated near the base of the central bundle, each subtending a large somewhat elongated
group of chlorenchyma cells which flank the central bundle and lie directly below the epidermis;
epidermis with the outer tangential walls very strongly lignified, furnished with stomata; seta essentially
similar in structure to the column but cells surrounding the bundles less lignified and smaller, the lateral
bundles disappearing towards the apex of the seta, the groups of chlorenchyma cells lying in close
contact with the central bundle which in turn is surrounded by relatively thin- walled cells.
Caryopsis: starch grains compound, consisting of numerous granules; embryo in sagittal section
showing the presence of a well-developed epiblast; the lower part of the scutellum fused to the coleorrhiza
or absent, no cleft being present at the base of the coleorrhiza; the scutellum bundle diverging directly
below the coleoptile (Fig. 2, E); cross-section of the coleoptile showing the first embryonic leaf with
margins touching but not overlapping, and with three vascular bundles, one median and two lateral,
the coleoptile with two lateral bundles and the scutellum with one median bundle. Anatomically no
differences could be detected characterizing the varieties.
Karyology
The chromosomes are of the small type and 2n = 48 (Fig. 7 No. 21).
Inflorescences rather dense, usually at least 4 times as long as broad, the branches short, densely
clustered and bearing spikelets from near the base var. dregeana
Inflorescences effuse and open, width about equal to the length, branches long and flexuous bearing
spikelets towards their ends var. elongata
(a) var. dregeana.
5. t/regeawa Steud., Syn. PI. Glum. 1 : 132 (1855); Dur. & Schinz, Consp. FI. Afr. 5: 811
(1895); Stapf, FI. Cap. 7: 572 (1898); Chippindall, Grasses and Pastures S.A. 1: 289
(1955).
Lasiagrostis capensis Nees, FI. Afr. Austr., 167 (1841); Trin. & Rupr., Gram Stip., 88
(1842).
Inflorescence dense, much elongated.
Cape. — Riversdale: Oakdale, Liebenberg 5617 {a). Oudtshoorn: near Cango Caves,
Bolus 12433. Uitenhage: Redhouse, Paterson 15946; Addo, Glennie s.n. Port Eliza-
beth: Koega, Fa/rt/o/p 252. Alexandria; Boschhoek, .^coc/cy 12089; Sandflats,
bald 5433. Cradock; Mountain Zebra Park, Barnard 518; Brynard 30; Cradock,
Holland s.n. Albany: Grahamstown, Godfrey & Story, S.H. 1351; Britten 2964.
Bathurst: Port Alfred, Tyson s.n. Dep. Agric. No. 12601; Kowie West, Britten 5926;
Kowie, Tyson, Tvl. Mus. 17070. King William’s Town: Keiskama Valley, Story 3690.
The distribution of this variety is limited to the Cape Province where it occurs
from the Peninsula to the King William’s Town district (Fig. 3). It occurs mainly
in shady places but often in habitats which are much more arid than those of the var.
elongata. The large, hard tufts are found along forest margins as well as in xerophytic
karroid scrub. It is well-grazed by game and stock.
216
(6) var. elongata (Nees) Stapf 'm FI. Cap. 7: 573 (1898); Chippindall in Grasses
and Pastures S.A. 1: 289 (1955). Type: Uitenhage, Primeval forest at Krakakamma,
Drege s.n. (Bf; PRE, fragment of holo.!).
Lasiagrostis elongata Nees, FI. Afr. Austr. 168 (1841). L. capensis var. elongata (Nees)
Trin. & Rupr., Gram. Stip., 88 (1842).
Stipa elongata (Nees) Steud., Syn. Plant. Glum. 1: 132 (1855).
Oryzopsis keniensis Pilger, Notizbl. Bot. Gart & Mus. Berlin 9: 509 (1926).
Piptatherum keniense (Pilg.) Roshev. in Bot. Mater Herb., Kamarov, Bot. Inst., U.S.S.R.
Acad. Sci., 14 (1951).
Inflorescence elfuse and divaricate, not much longer than broad.
Cape. — Cape Peninsula: Camps Bay, 772; Orange Kloof, Ai/a/ni'on 881 ; Kirsten-
bosch. Bolus s.n.; Skeleton Ravine, Marloth 6048; Bolus s.n. Alexandria: Olifants-
hoek, Johnson 909. Victoria East: Hogsback, Johnson 1139. Cathcart: Liebenberg
5389. Stutterheim: Kologha Forest, Acocks 9020; Dohne Mtn., Galpin 2455. East
London: near river, Go/p/n 6533. King William’s Town: Sim 20251. Komga: Grass
valleys near Komga, Flanagan 909; Kei Road Station, Ranger s.n. Kentani: Chippen-
dall 325; Along stream, Pegler 1106. Mt. Frere: Story 939. Umzimkulu: Insikeni
near Hoha, van Fonder 4.
Natal. — Lion’s River: Nottingham Road, Maclean 955. Estcourt: Little Tugela
River, Pentz 178. Bergville: Cathedral Peak Forest Res. Station, Indumeni River
valley, Killick 1675; The Cavern, in fern forest, Gemmel 5347. Mont, aux Sources,
Mogg 5314.
Transvaal. — Krugersdorp: Magalies River, Codd 522; Waterval 74, Mogg 23285.
Soutpansberg : Malta Gorge, Junod 4432; Blaauwberg, Leipzig Mission, Schweickerdt
1818.
British East Africa: Kenya. — Ngong, van Someren AH 9548; Muguga, Verdcourt
1841; Nakuru, Eastern Mau Forest Reserve, Maas-Geesteranus 6102; Londiani,
Tonderet Forest Reserve, Maas-Geesteranus A9\5.
The variety elongata occurs in the coastal areas of the Cape from the Peninsula
to the Transkei, along the Drakensberg escarpment in Natal and in a few areas of the
Transvaal (Fig. 3). It is also found in the cool highlands of Kenya at altitudes of over
6,000 feet. Personal observation and collectors’ notes indicate that it prefers cool,
shady and moist conditions. These factors explain, to some extent, its very interrupted
distribution; suitable habitats being rare and occurring at wide intervals. It is possible
that it has been over looked in the cool mountainous areas of the Rhodesias, Uganda
and Tanganyika where one would expect it to occur. The fact that it is usually not
abundant in its known habitats is a point in favour of the latter view.
Stipa dregeana was described originally by Nees as Lasiagrostis capensis FI. Afr.
Austr., 167 (1841). The genus Lasiagrostis is regarded by most authors as synonymous
with Stipa, but Roshevitz (1951) retains it as distinct. If regarded as synonymous
with Stipa and retained as a section, the name of the section must be Achnatherum
for nomenclatural reasons, as pointed out by Elias (1942, p. 62) and Hitchcock (1950,
p. 216). Investigation, however, has shown that S. dregeana does not belong to this
section sensu stricto, but probably, in agreement with Elias (1942, p. 64) is provisionally
best placed in the section Ptilagrostis. Some modification here may eventually be
desirable since the two lobes at the apex of the lemma, otherwise typifying the section,
are lacking in S. dregeana. Oryzopsis keniensis Pilger, which is conspecific with Stipa
dregeana var. elongata, is placed in the genus Piptatherum Beauv. of the subtribe
Timouriinae Roschev. by Roshevitz (1951). The type species of Piptatherum Beauv.
is Piptatherum coerulescens Desf. to which S. dregeana is not closely related. Pipta-
therum coerulescens has a glabrous, ovate lemma and an untwisted deciduous awn,
whereas the palea is distinctly exposed at the base of the floret, the lemma only over-
lapping upwards: a typical characteristic of Oryzopsis sensu lato which includes
217
Piptatherum. In S. dregeana the lemma is cylindric, and in the fruiting spikelet the
palea is slightly exposed towards the base of, and between the practically parallel margins
of the lemma. In mature specimens the awn is often twisted at the base and does not
disarticulate. The following species are apparently close relatives of S. dregeana: Stipa
sibirica from Europe, Stipa vaseyii from N. America, Achnatherum pekenense (Hance)
Ohwi (= Stipa extremiorientalis) from Japan, and Oryzopsis aequiglumis Duthie from the
North-West Himalayas. All these species technically belong to the genus Oryzopsis
according to Bor’s delimitation (1960, p. 641) of Stipa and Oryzopsis. Bor separates
Oryzopsis from Stipa as follows; “ Lemmas elliptic, plump; awn eventually deciduous;
not twisted but curved; callus short, obtuse, never long and pointed”. If Bor’s
delimitation of these genera is accepted S. dregeana, although possessing a twisted
persistent awn, could possibly be placed in Oryzopsis. Another fact which may be of
significance is the shortness of the epiblast in the embryo of S. dregeana (Fig. 2, E).
According to Reeder (1957, p. 764) Stipa has a very long epiblast which reaches to the
tip of the coleoptile (cf. S. capensis. Fig. 4E) whereas Oryzopsis has a “ rather small
epiblast ”. The rather short epiblast of S. dregeana would again point to a closer
affinity with Oryzopsis. The delimitation of the genera by different authors, however,
varies to such a degree that at present a transfer of S. dregeana to Oryzopsis seems
unjustified.
A new combination in Stipa based on Lasiagrostis capensis Nees, the oldest name
for this species, is not possible due to the existence of an earlier homonym, Stipa
capensis Thunb., and S. dregeana Steudel therefore is the valid name for the taxon
when placed in the genus Stipa.
2. S. capensis Thunb., Prodr. 19 (1794); Thunb., FI. Cap. ed Schult. 106 (1823);
Nees, FI. Afr. Austr. 1; 170 (1841); Trin. & Rupr., Gram. Stip. 63 (1842); Steud,
Syn. PI. Glum. 1 ; 129 (1855); Tackholm & Drar, FI. Egypt, 1 : 354 (1941); Chippindall
in Grasses & Pastures S. Afr. 1: 290 (1955). Type: Cape, Thunb. s.n. (UPS, holo.).
Stipa tortilis Desf., FI. Atlant. 1: 99 t. 31, Fig. 1 (1798); Kunth, Enum. 1 Suppl.:
134 (1835); Trin. & Rupr., Gram. Stip. 64 (1842); Steud., Syn. PI. Glum. 1: 130
(1855); Boiss, FI. Or. 5: 500 (1884); Durand & Schinz, Consp. FI. Afr. 5: 812 (1895).
Stapf in FI. Cap. 7: 372 (1898).
Caespitose annual. Culms erect or geniculate, slender, hollow, 10-100 cm high,
unbranched or branched from lowermost nodes, glabrous ; internodes usually enclosed
by the sheaths, or the upper exserted; nodes glabrous. Leaf-sheaths rather loose,
often slipping ffom the culms, the upper wider and sheathing the base of the young
panicles, sparsely pubescent to pilose or glabrous. Ligule a membranous ciliolate
rim. Leaf-blades linear, tapering to a fine point, expanded or somewhat rolled, 5-20
mm long and up to 3 mm wide, glabrous or sparsely pubescent to pilose below, often
scabrid above. Panicle spike-like, narrow, 3-15 cm long; rachis terete, slightly ribbed,
scabrous with appressed hyaline spines; branches fascicled, scabrous, very unequal,
branched or simple, up to 7 cm long; pedicels shorter than the spikelets. Spikelets
linear in outline, silvery, 1 -2-1 -6 cm long. Glumes very narrowly lanceolate to linear,
hyaline, silvery, tapering to a fine apex, 3-nerved, the lateral nerves short; lower usually
exceeding the upper in length, up to 1-6 cm long; upper up to 1-5 cm long. Lemma
cylindrical, margins convolute, tightly clasping the grain, indurated, 4-5 mm long
excluding the callus, sparsely bristly, 5-7-nerved, dorsally constricted just below the
awn; awn solitary, deciduous, articulated at apex of lemma, 5-8 cm long, column
twice geniculate, tightly twisted, hairy; bristle short, scabrid; callus pungent, bearded,
up to 2 mm long. Palea much shorter than the lemma, 1-25-1 -5 mm long, obtuse,
somewhat indurated, 2-nerved, glabrous. Lodicules 2, oblong, membranous, hyaline,
up to 1-5 mm long. Stamens 3, anthers yellow, 2 -0-2 -3 mm long, penicillate. Ovary
glabrous, styles distinct, stigmas plumose. Caryopsis spindle-shaped, shortly stalked,
3-3-5 mm long, brown; hilum linear, nearly as long as the grain; embryo about
the length of the grain (Fig. 4, D).
218
0-25fTim
Fig. 4. — Stipa capensis: A, cross section of vascular bundles of the leaf-blade; B, diagram of leaf-
blade in cross section; C, abaxial epidermis of the leaf-blade; D, caryopsis showing the
embryo; E, longitudinal section of the embryo; F, cross section of the embryo {Acocks
18591).
Anatomy (Fig. 4, A & B)
Leaf-blade thin and flat, tapering towards the margins, abaxial surface with shallow depressions
between the bundle units which are slightly raised, both surfaces with a few hairs; midrib only slightly
protruding abaxially, not distinctly keeled; margins obtuse to acute. Abaxial epidermis: stomatal
zones with a few stomata or stomata absent and zones consisting of elongate rectangular ripple-walled
cells usually alternating with short elements; short elements single or paired, wider than long,
occasionally giving rise to short unicellular papilla-like hairs, occasionally much enlarged, and producing
very long soft unicellular hairs; bicellular hairs absent; silicified cells small, dumb-bell-shaped.
Vascular bundle units: first order units about 9, trapezoid in outline, alternating with the second order
units; second order bundle units broadly ovate in outline. Bundle sheaths: outer of large thin-walled
parenchymatous cells, in the larger first order bundles supplemented by large cells both ab- and adaxially
and these cells gradually decreasing in size as the walls increase in thickness; thus merging into the
fibres of the stereome; inner sheaths of much smaller cells with the inner tangential walls strongly
thickened, the outer walls, however, thin-walled. Chlorenchyma of irregular thin-walled cells forming
a continuous tissue between the bundles. Stereome strands composed of small fibres, not very strongly
developed, present both ab- and adaxially opposite all bundles or occasionally absent adaxially opposite
the second order bundles, but always absent opposite the motor cells. Motor cells rather poorly
differentiated, thin-walled, 5-8 in number, occupying only about i the thickness of the leaf.
Awn: column more or less square in outline, shallowly grooved ventrally, furnished with three
vascular bundles; central bundle large and surrounded by large cells with very strongly asymmetrically
lignified walls, those situated on the ventral and dorsal sides of the bundle much exceeding the others
in size; lateral bundles small and situated near the base of the central bundle, each subtending a large
somewhat elongated group of chlorenchyma cells which flank the central bundle and lie directly below
219
the lateral epidermis; epidermis with the outer tangential walls strongly lignified, furnished with
stomata opposite the chlorenchyma groups. Seta similar to the column but all cells with thinner walls,
towards the apex the lateral bundles disappear and the chlorenchyma cells lie in close contact with the
central bundle which is then surrounded by one or two layers of thin-walled cells.
Caryopsis: starch grains compound, composed of numerous granules; embryo in sagittal section
showing the presence of an extremely elongated epiblast, the lower part of the scutellum fused to the
coleorrhiza, no cleft present at the base of the coleorrhiza and the scutellum bundle diverging below
the coleoptile. The cross-section of the first embryonic leaf (sheathed by the coeloptile) with margins
touching but not overlapping; the coleoptile with two lateral vascular bundles and the scutellum with
one median vascular bundle (Fig. 4 E & F).
Cape. — George: Karoo in vicinity of Gauritz River, Ecklon s.n. Clanwilliam;
Schlechter 8588; Welbedacht, Maguire 1855. Vanrhynsdorp: Ebenezer at Olifants
River, Drege 8105; Zandkraal locally abundant in Marginal Strandveld, Acocks 14736;
2 m. N. of Vanrhynsdorp, Schweickerdt 2558; Wiedou River, Barker 9483. Calvinia:
Groot Toring, Acocks 18591. Namaqualand: Doom River Bridge, Taylor 997;
Wallekraal, between Kamieskroon and Hondeklipbaai, Schweickerdt 2565.
Iraq. — Isbel Hamrin near Injana, Guest s.n. (N.H. 15797); Isbel Daramish-kan,
Guest s.n. (N.H. 16003).
It occurs in the arid parts of the south western Cape in areas with winter rainfall
(Fig. 3), and is widely distributed in North Africa and the Middle East. The interrupted
distribution of S. capensis follows a pattern similar to that of many species in the
Gramineae and various other families (see page 304). This annual species is adapted
to desert conditions.
Guenzel (1921, p. 6) investigated the anatomy of the leaf-blade of S. capensis
(as S. tortilis Desf.) and his observations agree fairly closely with the description given
above. The specimen investigated by Guenzel was collected at El-Kantara in North
Africa.
PART 2
THE TRIBE ARISTIDEAE
2.1. History
2.1.1. Taxonomic Studies
For a detailed discussion of the history of the nomenclature of the genus the
reader is referred to Henrard’s introduction to his Monograph (1929) as well as
Schweickerdt’s introductory remarks in his treatment of the South African species
(1941).
In the past, Aristida has almost invariably been placed in the tribe Stipeae which,
in turn, is treated either as a distinct tribe or as a subtribe or a group of the Agrosteae.
Since 1875, when Duval-Jouve studied the anatomy of the leaf-blade of two
Aristida species, it has been known that the anatomy of this genus differs from that of
Stipa, the type genus of the Stipeae. These anatomical differences were, however,
not regarded to be of sufficient value to exclude Aristida from the Stipeae, and, should
there have been no other differences between Stipa and Aristida, one would have been
inclined to agree. Avdulov (1931), Prat (1936) and others have, however, shown that
the type of anatomy of the leaf-blade is usually closely correlated with karyological
and organographical characteristics. Probably influenced by these observations,
Roshevitz (1937) transferred the genus Aristida to the tribe Sporoboleae. Roshevitz’s
concept of the Sporoboleae is a wide one and includes very heterogeneous elements.
Anatomically Aristida and Sporobolus, however, agree fairly closely and a distant
affinity between Aristida, Sporobolus and the genera of the Eragrosteae, which have a
generally similar anatomy, seems probable. A discussion of the relationship of the
Aristideae with other tribes of the Gramineae is given later in this paper.
220
In 1940 Hubbard & Vaughan created a new tribe, the Aristideae, to accommodate
the single genus Aristida but gave no description. In most recent works on the
classification of the Gramineae this tribe has been accepted. A valid description by
Hubbard appeared in Bor’s Grasses of Burma, Ceylon, India and Pakistan (1960).
The latest account of the Aristideae was prepared by Pilger (1956, p. 118) edited by
Eva Potztal, and posthumously published in the Pflanzenfamilien. In this treatment
Pilger enlarged the concept of the tribe by including the two Australian genera
Amphipogon and Diplopogon. The reasons why the latter two genera are excluded
from the Aristideae by the present author are given on p. 221.
2.1.2. Anatomical Studies
Duval-Jouve (1875, p. 319 and PI. 17, fig. 10) was one of the first to investigate
and describe the anatomy of a member of the Aristideae. His drawing of Arthratherum
pungens (Stipagrostis pungens) gives an excellent representation of the structure of the
leaf-blade, and agrees closely with that of Stipagrostis namaquensis studied by the
present author. In both taxa the presence of a single chlorophyll-bearing bundle sheath
can be seen clearly.
The most valuable study of Aristida undertaken up to the present is probably
that of Holm (1901, p. 101-133), who came to many conclusions similar to those arrived
at independently by the author of this study.
Holm studied 32 species of the section Chaetaria, two species of section Streptachne
and three of section Arthratherum sensu stricto. All these species have naked awns
and were found by Holm to be characterized by having two chlorophyll-bearing bundle
sheaths, or as he described them “ a double parenchyma sheath ”, as well as having
the chlorenchyma arranged in radial “ palisades ”. All Aristida species possessing
plumose awns, some of which had been referred to the section Arthratherum by various
authors until Bentham and Hooker and Hackel transferred them to the section Stipa-
grostis, were found by Holm to possess a single chlorophyll-bearing sheath while the
inner sheath was usually composed of thick-walled cells devoid of chloroplasts.
On basis of these observations Holm (p. 129) suggested that the species with
plumose awns and a single chlorophyll-bearing sheath should be removed from Aristida
proper, the latter having naked awns and a double chlorophyll-bearing sheath. Henrard
obviously also considered this possibility but for various reasons decided otherwise.
Species of this group investigated by Holm are A. plumosa, A. acuti flora, A. brachyathera,
A. ciliata, A. pungens var. pennata and A. pennata. Holm’s observations were fully
verified in the present investigation and were shown to be applicable to the South
African species. In addition Holm studied many other genera either purported to be
related to, or ecologically associated with, the genus Aristida. None of these genera,
however, exhibited a double chlorophyll-bearing sheath, a characteristic up to the
present met with only in Aristida sensu stricto. Holm was one of the first authors to
point out the difference between the chlorenchyma organization of Stipa, where the
cells show no regular radial pattern, and that of Aristida, Muehlenbergia and Lycurus,
where the chlorenchyma cells are radially arranged in a single row of “ palisades ”.
Theron (1936, p. 4) apparently misinterpreted the structure of the bundle sheaths,
so clearly described by Holm, and stated that he had not found constant differences
in this respect between the species with plumose awns and those with glabrous awns.
Theron investigated the anatomy of the leaf-blades using material taken from herbarium
specimens only and probably did not have access to living or preserved wet material.
This may explain the incorrect detail in many of his drawings. Failure to recognize
the difference in the bunole sheaths of the sections Stipagrostis and Schistachne on the
one hand, and the remaining sections on the other, led him to divide the species into
“ anatomical groups ”, which have very little bearing on taxonomic relationship based
mainly on organography.
221
Jelenc (1950) made a general anatomical study of the North African species of
Aristida but, as in the case of Theron (1936), was apparently not aware of the work of
Holm (1901). Jelenc, however, correctly interpreted the differences in the bundle
sheaths of the plumose-awned and glabrous-awned species, but could not, as he states
himself, interpret the structure of the chlorenchyma on basis of the slides prepared
from dry material, nor could he detect any bicellular hairs as observed by Prat (1936).
The presence of bicellular hairs was likewise not observed by Theron. In the present
paper these structures are clearly illustrated (Fig. 157E). Jelenc pointed out the
differences in the structure of the silicified cells in the epidermis of the two groups;
being (1) circular in the species with plumose awns and (2) dumb-bell-shaped in the
species with glabrous awns. Although the shape of the silicified cells is not of specific
diagnostic value throughout, it nevertheless is usually a good indication of the affinities
of the species.
Several other authors have studied Aristida species in more generalized papers
on grass anatomy but none of these has expressed opinions on the taxonomy of Aristida
or its sections. Practically all the information contained in these papers strongly
supports the evidence gathered by Holm and Jelenc and that presented by the present
author.
Papers mentioning only one or two species are the following: Volkens, 1887,
A. ciliata; Guenzel, 1921, p. 3, A. pungens; Sabnis, 1921, p. 225, A. funiculata and
A. hirtigluma', and Zemke, 1938, p. 399, A. ciliata.
Several South West African species were investigated by Guenzel (1913, pp. 16-22).
He reported on the presence of bicellular hairs in addition to unicellular long hairs,
furthermore on the differences in the bundle sheaths of the plumose and non-plumose
species. For the non-plumose group A. adscensionis and A. barhicollis and for the
plumose group A. uniplumis, A. ciliata, A. obtusa and A. namaquensis are mentioned.
Lommasson (1947) offered a critical appraisal of the interpretation of the homology
of the “ double parenchyma sheath ” in Aristida. He concluded that there was no
obvious reason for regarding the two “ parenchyma sheaths ” of the naked awned
species as being different from (i.e. not homologous with) the two sheaths found in
most grasses [as indicated by Holm (1901, p. 103)]. He suggested that the enlargement
of the inner sheath cells was a result of the acquisition of chloroplasts by the cells of
that layer. He pointed out that practically in all grasses the parenchyma cells on the
chlorophyll-vascular boundary (the outer-sheath) were enlarged in a similar way. In
Aristida this boundary has shifted to the inner sheath thus occupying the same position
as the “ mestome sheath ” present in other grasses, the cells being large and containing
chloroplasts, whereas the outer sheath cells have become smaller and are functionally
part of the chlorenchyma surrounding the bundle. This change (that has taken place
in Aristida) should be regarded, therefore, as one of function: the inner sheath is
homologous with the “ mestome sheath ” present in other grasses.
Finally, Kinges (1961, pp. 50-93) studied the structure of the embryos of a large
number of grasses including some Aristida species and Caceres (1961, p. 1) investigated
eight Argentinian Aristida species.
2.2. Delimitation
Until Pilger (1956, p. 118) widened the concept of the tribe to include Diplopogon
and Amphipogon, Aristida was the monotypic representative of the Aristideae. The
former two genera are organographically very different from Aristida and an anatomical
study of the leaf-blade of Amphipogon has revealed additional supporting differences.
Diplopogon, of which no material was available for investigation, is closely allied to
222
Amphipogon (and probably also agrees with the latter in anatomical characteristics).
In the anatomy of the leaf-blade, Amphipogon somewhat resembles that of some
Danthonia species. Even though the leaf-blades of Plectrachne and Triodia (Burbidge,
1946) differ considerably from Amphipogon in anatomical detail, they are nevertheless
basically similar. The presence of two-celled hairs observed in Plectrachne and the
agreement of the characteristics of the embryo with that met with in the Danthonieae,
raises the question as to whether these genera could be included in the Danthonieae.
Whereas the lemma of Amphipogon resembles that of Triodia, not only in general
texture but also in the tri-lobed apex, and occasionally in indumentum, this organ is
three-nerved in Amphipogon, but often several nerved in Triodia. Nevertheless there
is a tendency towards a reduction in the number of nerves in Triodia, some species
having only three nerves with vestiges of additional ones at the base of the lemmas
(Burbidge, 1946), so that this character seems less important. Finally the structure
of the embryos of Amphipogon and Triodia agree with that of Danthonia. Thus, in
spite of possessing one-flowered spikelets Amphipogon appears to be related to Triodia,
which has many-flowered spikelets. The subtribe Triodiinae created, and probably
misplaced, by Pilger (1956) in the tribe Festuceae should therefore be reinvestigated
on the lines suggested above in order to determine whether it should be transferred
to the tribe Danthonieae.
The differences between Amphipogon and Aristida are summarized below.
1 . Chlorenchyma continuous between the bundles
and several to many layers thick
2. Embryo inserted very obliquely at the base of
the caryopsis; embryo about i- of the length
of the grain; hilum punctiform and basal;
pericarp loose, occasionally slightly crustaceous
3. Lemma loosely enveloping the grain, often
with vertical rows of hairs between the nerves
dorsally
4. Palea well-developed and produced into two
bristles apically
1 . Chlorenchyma, a single layer of tabular cells
radially arranged around the bundles.
2. Embryo inserted on the ventral face of the
caryopsis; embryo to L the length of the
grain; hilum linear, at least half the length
of the grain; pericarp tightly adhering to the
grain.
3. Lemma tightly enclosing the grain, scabrid or
glabrous or with very sparse, scattered,
appressed hairs.
4. Palea a small chartaceous scale never much
more than half the length of the lemma.
The inclusion of Amphipogon (and Diplopogon) in the Aristideae therefore does
not seem justifiable and for this reason the tribe should be limited to the genus Aristida
sens. lat.
2.3. Phylogenetic Position
As may be seen from the historical account, Aristida and Stipa with its allies have
almost invariably been closely associated taxonomically. For this reason a careful
re-examination of the characteristics of the Aristideae and of representatives of several
genera of the Stipeae was undertaken.
The remarkable superficial similarity of the florets and grains of the Aristideae
and Stipeae presents one of the most striking examples of convergent evolution in the
Gramineae, a family in which this phenomenon is particularly well-developed. In
spite of this resemblance the characteristics separating the two tribes are profound
and easily detected, even on a purely organographic basis. After a thorough study
of the South African, as well as a number of non-indigenous, representatives the
following table, contrasting the organographical, anatomical and cytological charac-
teristics of the two tribes was drawn up. Information drawn from the considerable
literature on the anatomy and cytology of these tribes, especially the more recent papers
by authors such as Reeder (1957), Brown (1958), Lommasson (1957) et alia, was also
made use of.
223
Stipeae
1. Ligule membranous.
2. Lemmas 3-7 (usually 5)-nerved.
3. Awn, including the column, a solid structure
with a distinct demarcation from the body of
the lemma; seta of the awn with three vascular
bundles and groups of diffusely arranged
chlorenchyma cells
4. Lodicules 3 (rarely 2), large, membranous,
fleshy at the base, obtuse
5. Embryo -4 the length of the grain (rarely
6. Embryo characterized by —
(a) a large well-developed epiblast;
(b) the absence of a cleft between the coleor-
rhiza and base of the scutellum;
(c) the vascular strand diverging into the
scutellum directly below the base of the
coleoptile
7. Chlorenchyma of leaf-blade diffuse and con-
tinuous between the bundles, composed of
several layers of irregular or circular cells
8. Bundle sheaths 2; the outer of large thin-walled
cells containing chloroplasts and often not well
differentiated from the chlorenchyma; the inner
of smaller cells with at least the tangential walls
thickened, without chloroplasts
9. Bicellular hairs absent from the epidermis.
Aristideae
1. Ligule ciliate.
2. Lemmas 3- or rarely 1 -nerved.
3. Awn consisting of a grooved column formed
by the attenuated upper part of the lemma,
column divided at the apex into 1-3 awnlike
lobes, or, when the column is absent, the apex
of the lemma divided into three very narrow
awn-like lobes; setae of the awns each with
a single vascular bundle with radially arranged
chlorenchyma cells.
4. Lodicules 2 or 0, fleshy at the base, obtuse,
5. Embryo -J— 2- the length of the grain.
6. Embryo characterized by —
(a) the absence of an epiblast ;
(h) the presence of a deep cleft between the
coleorrhiza and the base of the scutellum.
(c) the vascular strand diverging some distance
below the base of the coleoptile and with a
distinct thickened internode leading up to
the base of the coleoptile.
7. Chlorenchyma of leaf-blade radially arranged
in a single layer of tabular cells around the
bundles, this layer separated from the adjacent
by irregular groups of parenchyma cells which
contain no or few chloroplasts.
8. Bundle sheaths 2; the outer of larger or
smaller cells than the inner, well differentiated
from the chlorenchyma; both sheaths thin-
walled or the inner with thickened walls; both
sheaths, or only the outer, containing chloro-
plasts.
9. Bicellular hairs always present on the epider-
mis.
10. Basic chromosome number 6, 1 1 or 12 (or their 10. Basic chromosome number 1 1 (no other basic
derivatives); chromosomes small or of medium number so far determined with certainty);
size chromosomes small.
1 1 . Distribution mainly in the temperate areas of 11. Distribution pan-tropical with extensions into
both hemispheres with extensions into the warm temperate areas.
warm temperate areas, rarely into the tropics
From the foregoing it follows that the Aristideae and Stipeae cannot be regarded
as being closely allied. Hence in any “ natural ” classification they should be widely
separated, as was actually done by Brown (1958) and by Stebbins (1956). In a diagram-
matic presentation of the relationship of the main grass groups based primarily on the
anatomy of the leaf-blade. Brown placed the Aristideae between the Pappophorea and
Maydeae but indicated that the first mentioned was an isolated group; whereas the
Stipeae was placed amongst isolated tribes like the Danthonieae, Oryzeae and Arun-
dineae. Brown’s diagram agrees in most respects with an earlier diagram published
by Stebbins (1956) setting out the evolutionary interrelationships of the grasses.
Of the various tribes of the Gramineae, the Eragrosteae probably exhibits the
greatest number of characters in common with the Aristideae. The Sporoboleae, the
Chlorideae and the Pappophoreae are likewise related to the Eragrosteae and thus the
placing of Aristida in the Sporoboleae (which dilfers from Eragrosteae only in the
one-flowered spikelets) by Roshevitz (1937), is not as anomalous as it may appear to
be on first sight. Examination of the following table contrasting the features of the
Eragrosteae and the Aristideae leaves no doubt, however, as to the distinctness of
these two tribes. Differences between the tribes are printed in italics.
224
Aristideae
1. Ligule a fringe of hairs.
2. Spikelets strictly one-flowered.
3. Lemmas 3-nerved, indurate and tightly clasping
the grains
4. Apex of the lemma always produced into 1-3
awns; the awn(s) formed by much narrowed
apical lobe(s) of the lemma
5. Lodicules 2. rarely absent; usually rather
elongate, fleshy only at the base and with a
membranous obtuse apex
6. Embryo the length of the grain.
7. Internal structure of embryo characterized
by—
(a) the absence of an epiblast;
(b) the presence of a deep cleft between the
base of the scutellum and the coleorrhiza;
(c) the vascular strand diverging some distance
below, and with a distinct thickened inter-
node leading up to, the base of the coleoptile
8. Hilum of the embryo linear, more than half the
length of the grain
9. Chlorenchyma of the leaf-blade radially
arranged in a single layer of tabular cells
around the bundles, each layer separated from
the next by irregular groups of parenchyma
cells containing a few or no chloroplasts
10. Bundle sheaths 2, both containing chloroplasts
(Aristida) or the inner without chloroplasts
(Stipagrostis)
11. Bicellular hairs present on the epidermis.
1 2. Silicified cells of the epidermis dumb-bell-shaped
or subcircular, paired with variously shaped
suberised cells
13. Basic chromosome number apparently always 1 1
14. Distribution mainly in tropics and subtropics
including desert areas
Eragrosteae
1 . Ligule a fringe of hairs or membranous.
2. Spikelets usually many-flowered, rarely 1-2-
flowered.
3. Lemmas 3-nerved, membranous to chartaceous,
loosely enclosing the grains.
4. Apex of the lemma not awned or with 1-3
awns or lobes; the awn(s) when present
formed by much narrowed apical lobe(s) of
the lemma.
5. Lodicules 2, rarely absent ; usually short, fleshy,
and truncate, or truncate with a small lateral
horn.
6. Embryo -3— 1 the length of the grain.
7. Internal structure of embryo characterized
by—
(a) the presence of an epiblast;
ib) the presence of a deep cleft between the
base of the scutellum and the coleorrhiza;
(c) the vascular strand diverging some distance
below, and with a distinct thickened inter-
node leading up to, the base of the coleop-
tile.
8. HHum of the embryo punctiform and basal.
9. Chlorenchyma of leaf-blade radially arranged
in a single layer of tabular cells around the
bundles, each layer separated from the next
by irregular groups of parenchyma cells
containing few or no chloroplasts.
10. Bundle sheaths 2, the inner without chloro-
plasts.
1 1 . Bicellular hairs present on the epidermis.
1 2. Silicified cells of the epidermis usually kidney-
shaped; paired with subcircular suberised
cells.
13. Basic chromosome number 1, 8, 9, 10, 12.
14. Distribution in the tropics and subtropics
including desert areas.
The main features in which the Aristideae differs from the Eragrosteae may be
summarised as follows: spikelets always with one floret; lemma indurate, tightly
clasping the grain; lodicules with an obtuse membranous apex, fleshy only at the
base; epiblast absent; hilum linear.
Pilger (1956), in his revised Gramineae II, recognizes amongst others the tribes
Eragrosteae, Aristideae and Chlorideae and places them next to each other in the
subfamily Eragrostoideae whereas the Sporoboleae is regarded as a subtribe of the
Eragrosteae. This classification is in conformity with modern data and probably
gives a good reflection of the natural afiinities of these groups. The placing of the
Pappophoreae in the subfamily Eestucoideae, however, does not conform with the
anatomical, cytological and embryological data available for this tribe, and it should
probably be placed near the Eragrosteae and Aristideae in Pilger’s subfamily Eragros-
toideae. De Wet (1956, p. 7) expressed surprise at Pilger’s views on the afiinities of the
tribe Aristideae, as well as at the fact that it is placed far from the Stipeae in Pilger’s
classification. De Wet is of the opinion that the two tribes are closely allied and were
probably both derived from “ the primitive Arundineae complex Whereas it is
225
possible that the Stipeae could have been derived from “the primitive Arumlineae
complex ” this seems unlikely for the Aristideae, because of its strong resemblance to
those of the subfamily Eragrostoideae.
In his discussion of the anatomical features of Aristida Metcalfe (1960, p. 40)
favours affinities with the Stipeae. This view is, however, no longer tenable.
In 1961 Caceres published an anatomical investigation of eight Argentinian species
of Aristida, and concurred with the opinion expressed by other authors, that Aristida
has an “ Eragrostoid ” anatomy and should be included in the subfamily Eragros-
toideae.
2.4. ARISTIDEAE C. E. Hubbard
In Bor, Grasses of Burma, Ceylon, India and Pakistan 685 (1960)
Annual or perennial herbs or occasionally suffrutices. Ligule a fringe of hairs.
Leaf-blades terete to expanded, narrow, linear. Spikelets all alike, bisexual, 1 -flowered,
arranged in contracted or open panicles. Rhachilla disarticulating above the glumes,
not produced beyond the floret. Glumes persistent, usually exceeding the lemmas.
Lemmas terete, becoming rigid and indurated, scabrid or smooth, or very rarely with
a few appressed hairs, 1-3-nerved, the nerves converging towards the apex but not
anastomosing, and each, or only the central nerve, produced into an awn, or upper
part of lemma produced into a caniculate 3-nerved twisted column, each nerve produced
into an awn at the apex; callus usually well developed, obtuse, pungent or bifid, bearded.
Palea much shorter than the lemma, indurate. Lodicules 2 or 0. Stamens 3 or 1.
Caryopsis tightly embraced by the lemma; hilum linear, more than half the length
of the grain, embryo the length of the grain.
Anatomy
Leaf-blade variously shaped in cross section, usually without a distinct keel; keel rarely developed
but always without colourless parenchyma. Silicified cells circular to dumb-bell-shaped, usually
accompanied by short to long, thin-walled suberized cells with undulate walls. Stomata usually present
in the ab- and adaxial epidermis. Hairs: one-celled and acute, as well as two-celled, cylindrical hairs
present, the latter with the apical cells usually longer than the basal, thin-walled and deciduous. The
bases of these cells persist, but the apical cells which are extremely thin-walled usually wither or are
rubbed off when epidermal scrapes are prepared. Even when present they are not easily detected.
Usually only the basal cell is depicted in the drawings. Vascular bundles with a double sheath: outer
sheath of small or large cells containing chloroplasts and always relatively thin-walled; inner sheath
of larger or smaller cells than the outer, with or without chloroplasts, and the cells thin-walled or with
regularly thickened walls. Chlorenchyma consisting of a single layer of tabular cells, radially arranged
around the bundles. Irregular parenchyma cells often present between the bundles below the motor
cells. Stereome consisting of typical thick-walled fibres or more rarely adaxially of large thick-walled
parenchyma cells. Motor cells consisting of more or less triangular groups of colourless parenchyma
occupying I to the whole width of the leaf.
Awn with column grooved, or rarely subcircular, containing three vascular bundles; setae each
with a single bundle. Embryo without an epiblast, with a deep cleft between the lower part of the
scutellum and the coleorrhiza; the margins of the first leaf of the embryo (sheathed by the coleoptile)
not overlapping.
Genera: Aristida, Stipagrostis. (Sartidia, a genus of uncertain affinity, is here
tentatively placed near Aristida. Because of its divergent anatomical features the
characteristics of this genus have not been included in the general description of the
tribe).
Type species: Aristida adscensionis L.
The Aristideae are mainly sub-tropical in distribution occurring in the drier areas
with relatively high winter temperatures.
226
2.5. The Genera of the Aristideae
Investigation of the South African representatives of Aristida sensu lato has shown
that the genus can be divided into three well defined groups. It has long been known
that the sections Stipagrostis and Schistachne differ in anatomical structure from the
other sections. All previous authors have either regarded the anatomical features as
insufficiently important to create two genera, or else, due to the incompleteness of the
data available, have refrained from creating two genera.
For the purpose of this paper all the South African species, and in addition a
number of the type species, or representatives, of the various sections from other areas,
were studied anatomically. The agreement between the grouping of the species based
on anatomy and the classification based exclusively on external morphology, is strong
and shows that the sections give on the whole, a fair reflection of the true relationship
of the species. This is in direct contrast with the conclusions arrived at by Theron
(1936) who investigated about 40 of the South African species (see p. 201).
In the course of this investigation a third well-defined group was discovered in
which the structure of the chlorenchyma differs strongly from that of the two groups
already discussed above. On anatomical features Aristida can therefore be divided
into three groups as follows: —
(1) Sections Stipagrostis and Schistachne.
(2) Sections Chaetaria, Pseudochaetaria, Arthratherum, Pseudarthratherum and
Streptachne.
(3) The divergent species A. jucunda, A. angolensis and A. vanderijstii.
To facilitate comparison of these three groups, the differences are contrasted as
follows: —
Group A
Stipagrostis
and
Schistachne
1. Lemmas with one or three
awns, at least the central awn
plumose, or if solitary and not
plumose, with a pencil of
hairs at the junction of awn
and lemma
2. Column present or absent
3. Embryo the length of the
grain
4. A cleft present between the
coleorrhiza and the base of
the scutellum
5. Starch grains compound, com-
posed of numerous granules
6. Hilum linear, on the surface
of the grain, or lying in a
shallow groove
7. Coleoptile 2-nerved
8. Chlorenchyma cells radially
arranged around the bundles
in a single row, the groups
separated by groups of paren-
chyma cells
Group B
Chaetaria, Pseudochaetaria,
Arthratherum, Pseudarthra-
therum, Streptachne
1. Lemmas with one or three
awns, awns scabrid, never
plumose nor pencilled at the
base
2. Column present or absent
3. Embryo FT the length of
the grain
4. A cleft present between the
coleorrhiza and the base of
the scutellum
5. Starch grains compound,
composed of numerous
granules
6. Elilum linear, rarely lying in
a groove
7. Coleoptile 2-nerved
8. Chlorenchyma cells radially
arranged around the bundles
in a single row, the groups
separated by groups of
parenchyma cells
Group C
A. jucunda, A. angolensis
and
A. vanderijstii
1. Lemmas 3-awned, awns
scabrid never plumose, not
pencilled at the base.
2. Column present or absent.
3. Embryo not more than i
the length of the grain.
4. A cleft absent between the
coleorrhiza and the base of
the scutellum.
5. Starch grains compound,
composed of 3-4 granules.
6. Hilum linear, lying in a
groove.
7. Coleoptile 3-nerved.
8. Chlorenchyma cells not
radially arranged around
the bundles; several cell
layers thick and continuous
between the bundles.
227
Group A
Stipgrostis
and
Schistachne
9. Two bundle sheaths present,
the outer usually of larger
cells than the inner, and only
the outer containing chloro-
plasts ; well differentiated from
the tabular chlorenchyma cells
10. Silicified cells of the epidermis
usually circular or more or less
square in outline, occasionally
dumb-bell-shaped ; bicellular,
cylindrical hairs present
1 1 . Column of the awns furnished
with three vascular bundles,
the central bundle without a
layer of chlorenchyma cells, the
lateral bundles partially
sheathed by a single layer of
chlorenchyma cells
12. Confined to the Old World,
mainly in desert areas (250 mm
or less precipitation per annum)
13. Chromosomes usually with
2n = 44 (rarely 2n = 22)
Group B
Chaetaria, Pseudochaetaria,
Arthmtherum, Pseudarthra-
therum, Streptachne
9. Two bundle sheaths present,
the outer usually of smaller
cells than the inner and both
containing chloroplasts ;
well differentiated from the
tabular chlorenchyma cells
10. Silicified cells of the epider-
mis usually dumb-bell-
shaped, rarely more or less
circular or substituted by the
paired elements of the sto-
matal zone; bicellular,
cylindrical hairs present
11. Column of the awns fur-
nished with three vascular
bundles, all bundles partially
sheathed by a single layer
of chlorenchyma cells
12. Pantropical, often in low
rainfall areas but not con-
fined to these
13. Chromosomes with 2n = 22
(rarely 2n = 44 or 55)
Group C
A. jucunda, A. angolemis
and
A. vanderijstii
9. Two bundle sheaths present,
the outer of larger cells than
the inner, and only the outer
containing chloroplasts;
poorly differentiated from
the surrounding chloren-
chyma cells which are cir-
cular or irregular in shape.
10. Silicified cells of the epider-
mis dumb-bell-shaped or
cylindrical; bicellular hairs
present.
11. Column of the awns fur-
nished with three vascular
bundles, each bundle
flanked by two many-celled
groups of diffuse chloren-
chyma cells.
12. Three species only: 'the
Congo Republic, Angola and
South Africa.
13. Chromosomes with 2n = 22
(S. jucunda).
Constant diflferences between groups A and B are the plumose awns and single
large-celled outer chlorophyll-bearing sheaths surrounding the vascular bundles in
group A, and the glabrous awns and double chlorophyll-bearing sheaths, of wnich
the outer sheath usually consists of much smaller cells than the inner, in group B.
Furthermore the two groups differ in distribution: group A being confined to the
Old World and, with very few exceptions, occurring only m areas with a rainfall of
less than 250 mm, whereas group B has a pantropical distribution and the species
are adapted to a wide range of habitats (see chapter on distribution).
Other significant differences which do not hold for all species are the usually
circular silicified cells of group A (3 of the annual species have dumb-bell-shaped cells)
and the dumb-bell-shaped silicified cells of the great majority of the species of Group B.
Most of the species of the section Stipagrostis and Schistachne (Group A) inves-
tigated cytologically have 2n = 44 while Chaetaria and related sections (Group B)
frequently have 2n = 22. In addition the general facies of the species belonging to
the sections Stipagrostis and Schistachne is different from that of the species belonging
to the other sections of Aristida.
All these characteristics, taken together, are, in the author’s opinion, sufficient
reason for the establishment of two distinct genera, namely (1) Stipagrostis, composed
of the sections Schistachne and Stipagrostis, and (2) Aristida composed of the sections
Aristida {Chaetaria), Pseudochaetaria, Arthralherum, Pseudarthratherum and Streptachne.
That Henrard had contemplated a similar decision is indicated by a remark in his
Monograph (1929, p. 25): “From the geographical distribution of both sections
treated here, one might conclude that we have here a distinct genus before us, a hypothesis
we cannot wholly reject if we consider the morphological and anatomical characters
of these sections Jelenc (1951) in a study of the stem and leaf of the Aristida species
of North Africa, also noticed some of tne distinguishing characteristics mentioned
228
above, and in his summary expressed surprise that Henrard did not divide the genus
into two subgenera on basis of the presence or absence of a double parencnyma sheath
(bundle sheaths). Jelenc investigated about 20 species, but regarded the information
as insufficient to bring about such a fundamental change in the classiffication of the
genus. Henrard mentions 46 species for the section Stipagrostis and Schistachne and
of these 34 (29 by the present author) have now been studied anatomically. So far all
species investigated conform to the general pattern described above. Since the external
morphological characters are quite striking, even though not easily expressed in words,
no difficulty should be experienced in identifying a single lemma as that belonging
to a member of this group even after the plumose awns have dropped.
The importance attached by some authors, such as Holm (1901), to the unique
feature of double chlorophyll-bearing bundle sheaths found in Arisdda sensu stricto
is, as pointed out by Lommasson (cf. p. 27), over-emphasized. As a generic difference
it is certainly of importance but it cannot be regarded as of sufficient phylogenetic
significance to exlude a genus such as Stipagrostis lacking this characteristic, from the
Aristideae.
A full account of the genus Aristida and its sections can be found on pages 234-238.
The third group, comprising the species A. angolensis, A. jucunda, and A.
vanderijstii, agrees closely with the section Aristida (Cliaetaria) of Aristida in the
external morphology of the spikelets, with the exception of certain characteristics of
the caryopsis. However, the very different leaf-anatomy taken in conjunction with
the differences in the caryopsis, indicates that these species should not be retained in
Aristida. These species constitute the genus Sartidia (described on p. 381).
2.6.1. Key to the Genera based on Organographic Characters
Awns three or one, glabrous scabrid or smooth:
Embryo the length of the caryopsis; caryopsis terete or compressed (rarely
grooved); glumes usually 1 -nerved (rarely 3-nerved) Aristida
Embryo up to \ the length of the caryopsis; caryopsis always deeply grooved
ventrally; glumes 3-5-nerved; awns spirally contorted at the base Sartidia
Awns three or one, at least the central awn plumose, or if glabrous and single,
then with a pencil of hairs at the base of the column Stipagrostis
2.6.2. Key to the Genera based on Anatomical Characters
Chlorenchyma radially arranged around the vascular bundles in a layer only one cell
thick:
Bundle sheaths two, both containing chloroplasts ; the outer sheath consisting
of smaller (or at least not larger) cells than the inner sheath, cells of both
sheaths relatively thin-walled; silicified cells of the epidermis usually dis-
tinctly dumb-bell-shaped or rarely irregular or sub-circular in shape. . Aristida
Bundle sheaths two, only the outer containing chloroplasts; the outer sheath
consisting of larger (or at least not smaller) cells than the inner; cells of
the inner sheath usually fairly thick-walled or at least more strongly lignified
than those of the outer sheath which are thin-walled, or rarely slightly
lignified ; silicified cells usually sub-circular in shape, rarely dumb-bell-shaped
(in the annual or sub-perennial species) Stipagrostis
Chlorenchyma continuous between the bundles, diffusely arranged and several to
many cells thick; bundle sheaths two, only the cells of the outer sheath
containing chloroplasts, the inner sheath consisting of strongly lignified cells,
the outer sheath of thin-walled cells; sheaths not very well differentiated from
the surrounding chlorenchyma; cells rather variable in size and shape. . Sartidia
229
2.7. THE GENUS ARISTIDA
2.7.1. History (see history of the Tribe on p. 219)
2.7.2. Distribution and Ecology (Fig. 5 and 6)
The genus Aristida occurs in the tropics and subtropics of both hemispheres. The
areas of highest concentration of species lie mainly along the Tropic of Cancer in the
northern, and along the Tropic of Capricorn in the southern hemisphere. The genus
is well developed in both hemispheres with a slight bias in the number of species in the
southern hemisphere.
Fig. 5. — World distribution of the section Aristida of the genus Aristida.
Fig. 6. — World distribution of the sections of the genus Aristida excluding the section Aristida.
5752152-2
230
The distribution of the different sections of the genus varies considerably. Of
these the section Aristida (Chaetaria) has both the widest distribution and the least
advanced floral structure and should be regarded as the most primitive of the sections.
It shows four areas of high concentration of species: (1) the southern part of the United
States including Mexico, (2) Eastern and Central South America particularly Brazil,
(3) the Rhodesias, Transvaal, Angola and South West Africa on the continent of
Africa and (4) Queensland and New South Wales in Australia. All these areas fall
within the regions which have a 50°F mean temperature for the midwinter month
(after Hartley, 1958). The influence of winter temperatures on the distribution of
Aristida is therefore well-marked. In many respects the distribution of Aristida sensu
stricto, and especially the section Aristida {Chaetaria), agrees in general features with
that of the tribe Eragrosteae as set out by Hartley (1958). The area of maximum
development of the Eragrosteae in North America is, however, not paralleled by the
section Aristida. For the section Aristida, the northern boundary of the “ peak area ”
in North America lies further south inside the 50°F. mean temperature of winter month
isotherm. A similar discrepancy is found in the distribution of the tribe Paniceae
when compared with that of the section Aristida, but here the “ peak area ” in North
America lies further east than for the Eragrosteae. The inclusion of all the sections
in a general distribution map of Aristida does not bring about an extension of the
distributional area, nor has it an effect on the positioning of the areas of “ peak
abundance ”. Should, however, the genus Stipagrostis be included the centres of
“ peak abundance ” would be strongly affected in the Old World, to which the latter
genus is confined. A secondary “ peak area ” adjacent to the one occupying the centre
of Southern Africa and a new one in North Africa would then be created. For this
reason the distribution of Stipagrostis is discussed separately (Fig. 81). From the
maps it is evident that the majority of Aristida species grow in areas with relatively
high winter temperatures and, that no species are found in regions which have very
cold winters. Rainfall seems to have less effect on the distribution, but few species
occur in areas with more than 1,500 mm per year while the number of species tolerating
less than 250-500 mm annually is also relatively small. Even though there seems
to be a certain correlation between the occurrence of grasslands and deciduous forests
and the occurrence of areas with high numbers of Aristida species, discrepancies occur,
the most striking of which is the high incidence of species found in Florida, in areas
covered mainly by evergreen coniferous forests and with an annual rainfall of between
1,000-1,500 mm. Records, however, show that in Florida most of the species occur
in open pine barrens on poor stony or sandy soil or on the sandy coastal plains. This
is in agreement with the occurrence of Aristida species in Africa in ecologically similar
habitats. It can be assumed therefore, that low humidity (or low available soil moisture)
is a secondary factor determining the distribution, while relatively high midwinter
temperatures are of primary importance.
The section Arthratherum is fairly small, containing about 40 species, and occurs
in both the New and Old Worlds. About five species occur in the United States and
Mexico but none is found in the West Indies or South America. Three species occur
in India and Pakistan, one in the Phillipines and about seven in Australia, mainly
on the eastern side of that Continent. Over 20 species occur in Africa. There are
two areas with high concentrations of species in Africa: Somaliland has six species
and South Africa 8 species. In both these areas the species are found mostly in areas
of low rainfall, on sandy, poor soils, or on rocky outcrops, usually in areas free from
frost.
The section Pseudarthratherum has 21 species some of which are doubtfully distinct
so that the number of species may have to be reduced. No members of this section
occur in North America, two are found in South America, while five occur in the
Galapagos Islands. The remainder of the species are found in Africa and Arabia
one of which extends into India. Due to the small number of species and their
231
scattered distribution, no distinct areas of high concentration are discernable. Eritrea
on the one hand, and central South Africa and Southern Rhodesia on the other, have
four species each. Most of the species occur in arid areas.
The section Streptachne, as clearly stated by Henrard, is not a uniform group,
but consists of species related to members of the section Arisiida and are only classified
together on basis of the reduction or absence of the lateral awns. The distribution
of the species confirms Henrard’s observation that it would probably be better to
unite them with the section Aristida. Eight species occur from the Southern United
States and Mexico to Panama; one occurs in Colombia in South America; one species
occurs in North East Africa; another in India and three in Australia. It seems likely
that they represent advanced types, derived from species belonging to the section
Aristida, and that they developed independently in or near the areas where they occur
today. The same applies to A. parvula the type of a monotypic new section occurring
only in South West Africa and the Northwestern Cape.
The last section, namely Pseudochaetaria contains three species and is confined
to Africa. All three species occur in the western half of the African continent: one
is limited to the Cape Verde Islands another to Senegambia; the third occurs from
Senegambia to South West Africa and eastwards to Southern Rhodesia and Abyssinia.
The majority of the species of Aristida occur in areas with an annual rainfall of
250 to 750 mm. Even those occurring in areas where the rainfall is high usually grow
in localities where the available moisture is limited. A few species, however, prefer
habitats where the soil is water-logged for at least part of the year: A. recta is found
mainly on boggy slopes where seepage water is available during the rainy seasons. In
the areas of the Transvaal, South West Africa and Angola, with an annual rainfall
of less than 750 mm A. jmciformis is found exclusively in depressions and gullies where
water collects and remains for long periods during the rainy season. It is increasing,
however, in the mountain sourveld [Ngongoni Veld, Acocks (1953, p. 35)] of Natal
where over-grazing and burning have caused a deterioration of the veld. In these
areas it forms large stands in situations which, apparently, are not very wet. The Natal
mountain sourveld has an annual rainfall of between 750 and 1,500 mm, which is
probably high enough to support growth even in apparently dry habitats. This predis-
position for high rainfall, or habitats where moisture collects during the rainy seasons
in areas of lower rainfall, may explain why there is no tendency for A. junciformis
to extend its distribution and form pure stands in over-grazed areas in drier regions.
A. junciformis subsp. galpinii grows at high altitudes in the mountains of Natal,
Basutoland and the Cape where the annual precipitation is high. A. monticola, a species
of very limited range in the Natal Drakensberg, shows a preference for wet habitats
occurring in great abundance along mountain streams and in seepage areas along this
escarpment.
In general, Aristida species are regarded as poor pasture and animals graze them
mainly in the very young stages. In the semi-desert areas of South West Africa and
the North Western Cape species such as A. engleri, A. dasydesmis and others, are
grazed fairly extensively but this is probably due to the scarcity of better fodder. The
Aristida species are of the most important pioneers in over-grazed or denuded veld
and, as such, are of economic value in the reclamation of these areas (Bews, 1929,
p. 209). On the other hand they are regarded as a pest in sheep-farming areas since
the “ seeds ” get entangled in the wool. Cases have been reported of the sharp calli
of the florets penetrating the skin of animals, causing irritation and pain, resulting
in a general fall-olf in the condition of the animals.
As could be expected the anatomy of Aristida shows less typical adaptations to
extreme arid conditions than that of Stipagrostis, which occurs mainly in desert areas.
It is striking that of the few species of Aristida growing in areas with less than 1 50 mm
of rain, at least one shows adaptations of the leaf structure similar to that found in
oi'^
^^irc
K'%''
y^r/c
*>>1
I ''VU'*
s^y\
nr
lIVTV
9
10
11
12
L
13
'.•fe
^l5:,rwcr
wnfo\ 5*»
17
18
Fig. 7. — Chromosomes of: 1, Aristida curvata, 2n = 22; 2, /4. effusa, 2n = 22; 3, A. meridionalis,
2n — 22; 4, A. traiisvaalensis, 2n = 22; 5, A. Junciforniis, 2n = 44; 6, /I. aequiglumis,
2n = 22; 1, A. .stipitata var. stipitata, 2n = 22; 8, .4. spectabilis, 2n = 22; 9, /4. congesta,
2n = 22; 10, /I. stipitata var. graciliflora, 2n = 22; l\, A. hordeacea, 2n = 22; \2, A. scabri-
vahis, 2n = 22; 13, A. diffusa var. burkei, 2n = 22; 14, A. canescens, 2n = 55; 15, Stipa-
grostis namaquensis, 2n = 44; 16, do; 17, 5. hochstetteriana var. hochstetteriana, 2n = 44;
18, S. obtusa, 2n = 44; 19, S. uniplumis var. uniplumis, 2n = 44; 20, Sartidia jucunda,
2n = 22; 21, Stipa dregeana, 2n = 48.
233
Stipagrostis. This species, A. dasydesmis, is endemic in the north-western coastal
areas of the Cape, and has terete, folded leaves with the bundle units strongly projecting
adaxially and very well developed stereome tissue (Fig. 51).
2.7.3. Karyology (Fig. 7: 1-14)
Fifteen species of Aristida sensu stricto were investigated by the author and in
all of them the basic number was found to be II. Most of the species proved to be
diploids, one a tetraploid and one a pentaploid. In the table below the species
investigated are given with a reference to the locality, and the collector and number
of the specimen preserved. In the different sections the species are arranged alpha-
betically.
2n
% Aristida —
A . aequiglumis 22
A. canescens 55
A. curvata 22
A. effusa 22
A. junciformis 44
A. scabrivalvis 22
A. transvaalensis 22
§Pseudochaetaria —
A . hordeacea 22
rthratherum — •
A. dijfusa vai-. burkei 22
A. meridionalis 22
A. spec tab ills 22
A . stipitata 22
var. graciliflora 22
§Pseudarthratherum —
A. congest a 22
Locality
Pretoria: South Africa
Pretoria: South Africa
Windhoek: South West Africa
Windhoek: South West Africa
Pretoria: South Africa
Pretoria: South Africa
Pretoria: South Africa.
Otavi: South West Africa
Pretoria: South Africa
Windhoek: South West Africa
Pretoria: South Africa
Pretoria: South Africa
Pretoria: South Africa
Pretoria: South Africa
Collector and Number
de Winter 7515
de Winter 7561
de Winter 7112
de Winter 7132
de Winter 7766
de Winter 7512
de Winter 5982
de Winter 6799
de Winter 795
de Winter 6711
de Winter 7516
de Winter 7562
de Winter 5981
de Winter 7514
Of the twenty-nine species of which the chromosome number is known, only six have been reported
to have the chromosomes not in multiples of 11. These are as follows: —
^Aristida —
A. i unciform’s 2n = 24 de Wet, 1954
2n = 36 de Wet, 1958
A. canescens 2n = 48 de Wet, 1954
A. transvaalensis 2n = 24 de Wet & Anderson, 1956
A. rhiniochloa 2n = 38 Thomas, ex Dari. & Wylie, 1955
§Pseudarthratheruin —
A. barbicollis 2n = 24 de Wet, 1954
^Arthratherum —
A. dijfusa var. burkei 2n = 36 de Wet & Anderson, 1956
It will be noticed that three sections are represented in these species so that n = 12
cannot be regarded as characteristic of any particular section. Of the above mentioned
species four were re-investigated (A. junciformis, A. canescens, A. transvaalensis and
A. diffusa var. burkei). In the material examined the basic number proved to be
n= 11 for all four species. A. rhiniochloa and A. congesta subsp. barbicollis (A.
barbicollis) could not be re-examined because of lack of material. It therefore seems
likely that the basic number of 1 1 is typical for Aristida.
The basic chromosome number is of value in distinguishing the tribe Aristideae
from related tribes such as Eragrosteae and Sporoboleae, but offers no additional
characteristics in the delimitation of genera, and sections of genera. Polyploidy is
uncommon and as far as is known is limited to the section Aristida which is regarded
234
by the author as the most ancient of the sections on basis of the primitive floral features
exhibited by many of the species.
The chromosomes of Aristida are of the same small type as those met with in the
Eragrosteae and Chlorideae. Due to their small size no morphological studies of the
individual chromosomes were undertaken.
In the course of the study of a large number of karyological preparations many
cases were observed where the nucleolus persists to the early metaphase. This, in
agreement with the findings of Brown & Emery (1957, p. 587), is an additional character
which the Aristideae has in common with the Eragrosteae, Pappophoreae and the
Sporoboleae.
2.1 A. ARISTIDA L.
Aristida L., Sp. PI. 1: 82 (1753); Kunth, Enum. Plant. 187 (1833) pro parte;
Trin. & Rupr., Gram. Stip. 99 (1842) exclud. §Stipagrostis; Bentham in J. Linn. Soc.
Bot. 19: 30 (1855) exclud. §Stipagrostis Nees and Schistachne Fig. & De Not.; Steud.,
Syn. PI. Glum. 1: 132 (1855) exclud. §Stipagrostis; Hack., True Grasses 101 (1896)
exclud. §Stipagrostis and ^Schistachne', Henrard, Mon. Gen. Aristida 1: 13 (1929)
exclud. \Stipagrostis, ^Schistachne, and A. sericans', Hubbard in Hutch. Fam. FI.
PI. 2, 2: 214 (1934) exclud. ^Stipagrostis and ^Schistachne', Pilger in Bot. Jahrb. 76,
3: 330 (1954) exclud. ^Stipagrostis and ^Schistachne.
Spikelets solitary, pedicelled, borne in terminal contracted or open panicles.
Rhachilla disarticulating above the glumes, not produced beyond the base of the floret.
Floret 1, hermaphrodite, equalling, or shorter than, or exceeding the glumes, narrowly
oblong to linear in outline, terete or laterally compressed. Glumes persistent, narrow,
acuminate to obtuse, or emarginate and mucronate, or shortly to distinctly awned,
1-3 {5)-nerved, nerves anastomosing or evanescent. Lemma cylindrical or laterally
compressed, convolute, or with margins involute forming a shallow ventral groove,
indurated at maturity, glabrous or finely to coarsely scabricl, rarely sparsely hairy,
3-nerved, nerves converging towards the apex but not joining, and each produced into
an awn, or the upper part of the lemma narrowea into a canaliculate, 3-nerved, usually
twisted, long or short column, each of the nerves, more rarely only the central nerve,
excurrent into an awn at the apex of the column or lemma, lemma not articulated,
or articulated between the apex of the lemma and the base of the column, or between
the apex of the column and the base of the awns; awns glabrous or scabrid, never
plumose, usually 3, or more rarely the lateral awns much reduced, or quite absent
and then only the central awn remaining; callus well-developed, obtuse, truncate,
emarginate, bifid or acuminate, bearded with usually short stiff" bristly hairs, or rarely
callus hairs up to half as long as the lemma and very fine, rarely callus almost glabrous.
Palea usually less than half as long as the body of tne lemma, indurated or membranous,
nerveless or 2-nerved, not keeled, the margins incurved, glabrous. Lodicules usually
two, rarely 3 or absent, fleshy at the base, membranous upwards, with several nerves,
obtuse. Stamens 1 or 3, anthers elongated. Ovary glabrous; styles free, stigmas
plumose, laterally exserted. Caryopsis tightly enclosed by the lemma, usually free,
rarely more or less adhering to the lemma ; hilum linear, slightly shorter than the grain ;
embryo usually about the length of the grain; starch grains compound, composed
of numerous granules.
Densely tufted perennials or annuals, usually without a rootstock or, rarely, a
well-developed rootstock present. Culms erect, simple or branched, usually solid
Leaf-blades linear, expanded or folded, mostly well-developed in a basal tuft, in annual
species borne mainly on the culms. Ligule a dense fringe of hairs. Panicles very
dense and spike-like, to effuse and divaricate.
235
Anatomy
Shoots long attenuated or rarely short and thick when very young; circular in cross section or
rarely compressed; leaf-blades conduplicate in the bud; sheaths tightly enrolling the young blades,
tapering towards the free margins (Fig. 10, D, E, F).
Leaf-blades flat, V-shaped or horseshoe-shaped in cross section. Silicified cells dumb-bell-shaped
or rarely circular in shape {A. dasydesmis), occasionally rather irregular (circular, kidney- or dumb-bell-
shaped in A. diffusa), short or much elongated, very narrow in the constricted middle section, the
236
distended ends usually containing a number of crystal-like bodies; usually large in the wider zones and
smaller in the narrower zones. Suberized cells variable in shape, thin-walled and usually accompanying
a silicified cell; from broader than long, to as long as the silicified cells, and with walls undulate. Hairs
of two kinds: (1) unicellular sharp hairs which vary from short broad-based retrorse barbs to long
thin hairs from a relatively narrow base, and (2) bi-cellular hairs arising from a narrow basal cell
usually situated in or flanking the stomatal zone; basal cell of bi-cellular hair thin- walled or with
slightly thickened walls, usually widening towards the truncate apex of the cell which is topped by a
very thin-walled, linear, obtuse cell, the latter deciduous and thus only the basal cell persisting. These
hairs usually lie longitudinally, closely appressed to the surface. Bundle sheaths: outer sheath usually
consisting of small, thin- walled, parenchymatous cells containing chloroplasts; inner sheath usually
of much larger cells with somewhat thickened walls and likewise containing chloroplasts, sheaths
complete or variously interrupted by the stereome strands. Chlorenchyma consisting of a single row of
tabular cells radially arranged around the bundles, occasionally with a few irregular cells below the
adaxial stereome flanking chlorenchyma, and with a few cells surrounding the air spaces below the
stomata.
Awn (Fig. 8, C & D): Column of the awn triangular or rarely subcircular in cross section, deeply
grooved ventrally, furnished with three vascular bundles of approximately equal size, all partly
surrounded by a single layer of radially arranged chlorenchyma cells; stomata present in the groove
opposite the chlorenchyma. Setae with a single vascular bundle surrounded by a usually complete,
single-layered sheath of radially arranged chlorenchyma cells. Lignified cells surrounding the bundles
and chlorenchyma usually strongly asymmetrically thickened, likewise the epidermal layer. In transi-
tional areas such as the junction between column and lemma, and at the branching point of the awns
at the apex of the column, the bundles are all devoid of a chlorenchyma sheath.
Embryo large, from slightly over i to 1 the length of the grain, usually with a slight constriction
between the coleoptile and the coleorrhiza (Fig. 9, R, S, T, U) in sagittal section showing the absence
of an epiblast, the free lower part of the scutellum which is separated from the coleorrhiza by a deep
cleft and the distinct elongation of the vascular tissue between the point of divergence of the scutellum
bundle and the base of the coleoptile (Fig. 9, M, N, O). The cross section of the coleoptile shows the
first embryonic leaf with the margins meeting, not overlapping and with five vascular bundles; the
coleoptile sheath with two lateral bundles, and the scutellum with one median bundle (Fig. 9; P, Q).
Type species: Aristida adscensionis L.
A genus consisting of about 260 species, widely distributed in the tropics and
subtropics of both hemispheres. Characteristic of poor dry soils in areas of relatively
low rainfall but a few moisture-loving species are known. Although most species
prefer a drier climate they are not true desert grasses such as most species of Stipa-
grostis.
2.7.5. Enumeration of the South African Sections
Aristida is divided into six sections, one of which viz. Schizachne was described
originally by Trinius and Ruprecht as a monotypic section of the genus Stipa.
1. §Schizachne
^Schizachne (Trin. & Rupr.) de Winter in Kirkia 3: 132 (1963).
Stipa ^Schizachne Trin. & Rupr., Gram. Stip. 53 (1842). Type species: A. parvula
(Nees) de Winter (=Stipa parvula Nees).
Annual; lemma with a distinct column and an articulation between the apex of
the lemma and the base of the column; callus obtuse and furnished with very long,
fine hairs more than half the length of the body of the lemma; lemma convolute;
awn single, twice geniculate.
2. §Pseudochaetaria
§Pseudochaetaria Henrard, Rev. Gen. Aristida 34 (1929). Type species: A. hordeacea
Kunth.
Annuals; lemma with an articulation situated at the base of the awns; column
absent; callus obtuse; margins of the lemma somewhat involute and spikelets shallowly
ventrally grooved; awns three.
237
Fig. 9, — A, seedling of S', imiplumis; B, seedling of S. hochstetterlana; C, S. anomala, longitudinal
section of the embryo; D, S. obtusa, do; F, S. obtiisa, cross section of the embryo; G, S.
ciliata, do; H, S. anomala, ventral view of caryopsis; T, S. ciliata, do; J, S. obtusa, do;
K, L, young seedlings of A. stipitata var. graciUflora and A. curvata; M, N, O, longitudinal
sections of the embryos of A. hubbardiana, A. hordeacea and A. scabrivalvis ; P, Q, cross
sections of the embryos of A. hubbardiana and A. hordeacea; R, S, T, U, ventral views of
grains of A. stipitata var. graciUflora, A. congesta subsp. barbicollis, A. hordeacea, A. scabri-
valvis.
Fig. 10. — Cross sections of shoots: A, S. amabilis {de Winter 3398); B, S. anomala (de Winter 3407);
C, S. obtusa {de Winter 3204); D, A. congesta {de Winter 7565); E, A. canescens {de Winter
795); F, A. hordeacea {de Winter 2729).
238
3. §Aristida
Chaetaria P. Beauv., Essai Agrost. 30 (1812). Aristida ^Chaetaria (P. Beauv.) Trin.,
Gram. Unifl. et Sesquifl. 175 (1824). Type species: A. adscensionis L.
Annuals or perennials with or without a column; articulation always absent;
callus obtuse or acuminate; lemmas convolute, or the margins involute, and spikelets
shallowly ventrally grooved, avms three, always glabrous.
4. §Arthratherum
Arthratherum P. Beauv., Essai Agrost. 32 (1812). Aristida ^Arthratherum (P. Beauv.)
Reichenb., Consp. 50 (1828). Type species: A. hygrometrica R. Br.
Annuals or perennials with a distinct column and an articulation between the apex
of the lemma and the base of the column, if articulation absent then callus bifid;
callus truncate, emarginate, bifid or acute; lemmas convolute; awns three.
5. §Pseudarthratheruin
§Pseudarthratherum Chiov. ex Henrard, Mon. Gen. Aristida 34 (1929). Type species:
A. astroclada Chiov.
Annuals or perennials with a distinct, sometimes short, column, and an articulation
between the apex of the column and the base of the awns; callus sub-obtuse, acute
or acuminate; lemmas convolute; awns three.
2.7.6. Key to the South African Sections
1. Awn single; articulation situated between the summit of the lemma and the foot of the column;
callus furnished with very long fine hairs more than half the length of the body of the lemma
1. Schizachne
la. Awns three, the lateral occasionally weakly developed; articulation absent or present; hairs
of the callus bristly never half as long as the body of the lemma:
2. Lemma without an articulation; callus never bifid 3. Aristida
2a. Articulation present, or if absent then callus bifid:
3. Articulation situated between the apex of the lemma and the foot of the column; column
of the awns usually well developed; if column or articulation absent then callus bifid
4. Arthratherum
3a. Articulation situated between the summit of the column and the foot of the awns or, when
column absent, between the summit of the lemma and the foot of the awns; caUus acute
or rounded:
4. Articulation between the summit of the column and the foot of the awns; callus acute
to obtuse 5. Pseudarthratherum
4a. Articulation between the summit of the lemma and the foot of the awns; column com-
pletely lacking; callus rounded 2. Pseudochaetaria
2.7.6. 1. Key to the South African Species Based on Organographic Characters
1. §Schizachne
Awn single; hairs of the callus more than half the length of the body of the lemma. . 1. A. parvula
2. §Pseudochaetaria
Articulation situated between the apex of the lemma and the branching point of the awns; body
of the lemma somewhat furrowed ventrally; only South African species known. . 2. A. hordeacea
3. § Aristida
1. Internodes glabrous or scabrous, rarely with some short hairs below the nodes:
2. Annuals or sub-perennials with soft, easily compressible internodes, if sub-perennial then
inflorescence effuse and very large; lemmas often coarsely scabrid, never with a twisted
column, at most narrowed into a beak:
239
3. Spikelets small, usually 2-5 cm or less in length including the awns, if up to 3 cm then
lemma narrowed into a beak and inflorescence short, dense and spike-like:
4. Panicle contracted and spike-like, or contracted and interrupted:
5. Panicle spike-like, very dense and short, up to 6 cm long; body of the lemma very
scabrous, furrowed ventrally and narrowed into a short beak: .glumes shortly awned
3. A. hubbardiana
5a. Panicle contracted, usually elongate and interrupted; lemma not narrowed into
a beak:
6. Panicle dense, but much interrupted, the branches usually bearing spikelets to near
the base; body of the lemma included in, or only shortly exceeding, the upper
glume 7. A. curvata
6a. Panicle contracted but not very dense, the branches usually with a fairly long,
naked basal part; lemmas usually strongly exserted from the glumes when mature
5. A. adsceiisionis subsp. guineensis
4a. Panicle effuse and open with the branches remote and divaricate:
7. Body of the lemma included in the glumes, not exserted beyond them:
8. Glumes distinctly awned, the lower glume usually shorter than the upper; delicate
annuals 8. A. scabrivalvis
8a. Glumes attenuate or shortly awned, the lower glume usually exceeding the upper
in length; short-lived perennials usually with a basal tuft of leaves 9. A. bipartita
7a. Body of the lemma exserted beyond the glumes at least in the mature spikelets;
glumes variable in length, awnless:
9. Spikelets clustered together at the ends of the branchlets; lemmas usually scabrid
all over 6. A. effusa
9a. Spikelets diffusely scattered on the branchlets, not densely clustered on them;
lemmas usually scabrid on the keel only 5. A. adscensionis subsp. guineensis
3a. Spikelets large and coarse, scabrid all over, or rarely the lemma scabrid only on the keel,
3-5 cm long including the awns; glumes reddish-brown or purplish in colour 4. A. rbiniocbloa
2a. Perennials usually with firm hard culms, or with a dense fibrous base; body of the lemmas
usually produced into a beak, or with a twisted column; glabrous or finely scaberulous, rarely
without a beak or column:
10. Lower glume longer than the upper; culms much branched; base rhizoraatous 13. A. monticola
10a. Lower glume shorter than the upper, or glumes subequal but the lower at least slightly
shorter than the upper:
11. Lemma without a beak or column, only slightly narrowed upwards:
12. Culms terete; leaves expanded with the marginal nerves raised and prominent
10. A. canescens
12a. Culms compressed, especially towards the base; leaves conduplicate, the marginal
nerves occasionally large, but not raised above the others 14. A. junciformis
11a. Lemma with a long or short column which is usually somewhat twisted:
13. Glumes more or less subequal the upper only 1-1-5 mm longer than the lower,
awnless, occasionally mucronate; internodes of culm not, or only slightly com-
pressed :
14. Spikelets small; glumes shorter than 10 mm, sub-glabrous; lateral awns almost
absent, or at least more weakly developed than the central awn; culms usually
much branched from the nodes 12. A. transvaalensis
14a. Spikelets larger; glumes about 10 mm long or longer, the lower frequently
slightly recurved from the apex and either smooth, or scaberulous to minutely
hairy, especially towards the tip; lateral awns well-developed; culms rarely
branched from the nodes 11. A. aequiglumis
13a. Glumes very unequal: the lower about to f as long as the upper; awned or
awnless :
15. Spikelets small, about 1 cm long including the awns, or slightly longer, usually
purplish; awns usually somewhat spirally contorted; base of plant covered
with fibres derived from the old leaf-sheaths 15. A. recta
15a. Spikelets larger, usually about 2 cm long including the awns; base of plant
not fibrous 14. A. junciformis & vars.
la. Intemodes covered with a fugacious, adpressed wool below the nodes, or lanate-woolly all
over, sometimes the lower internodes only woolly, or pubescent-lanate and the upper glabrous;
very robust tall plants 16. A. sciurus
240
4, §Arthratheruin
1. Callus truncate, obliquely truncate, rounded or even slightly emarginate; culms fascicled from
the lower nodes 17. A. dasydesmis
la. Callus bifid, or conical and acute; culms branched or unbranched:
2. Callus bifid:
3. Auricles and ligule pubescent or ciliate, not woolly; culms glabrous; glumes very unequal,
apices never membranous:
4. Lemmas smooth, glabrous or slightly scaberulous on the keel upwards; culms unbranched
(very rarely somewhat branched) 18. A. diffusa
4a. Lemmas scabrid on the back in the upper half; culms branched or unbranched
20. A. engleri
3a. Auricles with a dense flake of wool, or lower internodes woolly to pubescent; glumes
subequal or unequal, apices membranous or firm:
5. Lower internodes pubescent to woolly, the upper often glabrous; column of the awns
rather short, about 6-5 mm long 19. A. vestita
5a. Lower and upper internodes glabrous; column of the awns 1-0-3 -5 cm long:
6. Glumes subequal or somewhat unequal, usually straw-coloured, with the apices mem-
branous, hyaline and delicate, often torn or broken 21. A. spectabilis
6a. Glumes very unequal, lower not more than |, usually about half the length of the
upper, often purplish, the apices firm, not distinctly hyaline:
7. Column of the awns usually exceeding 2 cm long; annual plants.. 23. A. stipoides
7a. Column of the awns usually not exceeding 2 cm; perennial plants 22. A. meridionalis
2a. Callus conical and acute, never bifid:
8. Internodes woolly or densely tomentose especially towards the base of the culms:
9. Panicle spike-like, long, contracted and narrow 24. A. mollissima
9a. Panicle often narrow but effuse and divaricately branched 25. A. argentea
8a. Internodes quite glabrous or minutely scaberulous only, never woolly or pubescent
26. A. stipitata
5. §Pseudarthratheruin
1, Panicles contracted but much branched, the branches erect or slightly spreading, never spike-
like; robust, erect perennials, with inflorescence up to 20 cm long 27. A. pilgeri
la. Panicles spike-like and dense, or somewhat interrupted, the branches short, usually bearing
spikelets to the base, or panicles lax and open, the branchlets ending in a dense cluster of
spikelets 28. A. congesta
2. 7. 6. 2. Key Based on Anatomical and Vegetative Characters
Due to the great difficulty in distinguishing many of the species on anatomical or vegetative grounds
it was not possible to draw up a full key. The guide given below may, however, be useful in limiting
the number of species to be considered in the identification of sterile material of any particular
species.
1. Annuals with soft, compressible culms; midrib usually protruding abaxially; stereome strands
of the third order bundles rather weakly developed, often consisting of a few fibres only:
2. Culms densely scabrous or scabridulous :
3. Culms scabrous with short spines 4. A. rhiniochloa
3a. Culms densely covered with spreading or reflexed, short, stiff hairs 2. A. hordeacea
3. A. hubbardiana
2a. Culms quite smooth, or only very sparsely puberulous:
4. Many short innovation shoots present at the base of the plant; auricles and margin of the
collar of the sheath often bearded 9. A. bipartita
4a. Innovation shoots absent or very few 5. A. adscensionis subsp. guineensis
7. A. curvata
6. A. effusa
8. A. scabrivalvis
1. A. parvula
241
la. Perennials with firm, hard culms; midrib of the blade not protruding abaxially:
5. First order bundles, except for midrib, concentrated on the margins of the blade, usually
only one, or occasionally two, on each margin but then closely associated, much exceeding
the 3rd order bundles, as well as the midrib ; the midrib flanked by 2-A third order bundles
and only very slightly exceeding these in size; stereome strongly developed opposite the
motor cell groups, but absent or weakly developed opposite the, third order bundles:
6. Culms unbranched or occasionally branched:
7. Leaves erect, not curled when dry; 3rd order bundles without abaxial stereome strands
11. A. aequiglumis
7a. Leaves curled when dry; 3rd order bundles with abaxial stereome strands 10. A. canescens
6a. Culms much branched especially from the upper nodes:
8. Plants with a rhizomatous base, each culm arising from a fairly long rhizome; base
not very dense and fibrous 13. A. monticola
8a. Plants without well developed rhizomes, densely caespitose; base tough, fibrous
12. A. transvaalensis
5a. First order bundles more or less uniformly spaced over the whole width of the leaf, or at
least a well developed first order bundle present between the marginal bundles and the
midrib; if first ordei bundles marginal, then all bundles subequal in size:
9. Adaxial stereome strands of the first order bundles consisting of cells which are much
larger and much thinner-walled than those of the abaxial strands; basal intemodes of
culms woolly or pubescent 16. A. sciurus
9a. Adaxial stereome strands of all bundles consisting of typical small, very thick-walled
fibres; internodes of culms glabrous, woolly or pubescent:
10. Leaf-blades narrowly to widely V-shaped in cross-section; first and third order bundles
subequal, and gradually decreasing in size towards the midrib; stereome strands
usually absent abaxially opposite the 3rd order bundles, but strongly developed opposite
the motor cell groups:
1 1 . Epidermal cells much smaller than the cells of the inner bundle sheaths 14. A. junciformis
11a. Epidermal cells exceeding the cells of the inner bundle sheath in size 15. A. recta
10a. Leaf-blades expanded, or if horseshoe- or crescent-shaped then with the bundle units
very unequal in size; stereome strands occasionally weakly developed, but present
opposite the motor cell groups as well as the bundles:
12. Auricles of the leaf-blade with a flake of wool:
13. Annual plants 23. A. stipoides
13a. Perrenial plants 22. A. meridionalis
21. A. spectabilis
12a. Auricles of leaf-blade without a flake of wool:
14. Basal internodes of the culms woolly or pubescent 25. A. argentea
19. A. vestita
24. A. mollissima
14a. Basal internodes of the culms glabrous:
15. First order bundles much exceeding the 2nd and 3rd order bundles in height:
16. Adaxial stereome elements of first order bundles much larger and thinner-
walled than the abaxial fibres; culms fascicled from the lower nodes; leaves
subterete in cross section 17. A. dasydesmis
16a. Adaxial and abaxial stereome strands both consisting of typical, thick-
walled fibres 18. A. diffusa
15a. First order bundles subequal to or only slightly exceeding the third order bundles
in height:
17. Marginal first order bundles usually with a third order bundle between them:
18. Very robust unbranched culms, up to 2 m high 21. A. spectabilis
26. A. stipitata
27. A. pilgeri
18a. Smaller plants with branched or unbranched culms, usually not exceeding
1 m in height 26. A. stipitata
20. A. engleri
27. A. pilgeri
17a. Marginal first order bundles usually adjacent 28. A. congesta & subspp.
242
2.7.7. Description of the South African Species
1. A. parvula (Nees) de Winter in Kirkia 3: 132 (1963)
Stipa parvula Nees, FI. Afr. Austr. 169 (1841). Type; Great Namaqualand ; stony
Karoo hills, Drege 2551 (Bf; PRE, fragment of holo.!).
Annual, very small and compact (6 cm high) to loose, spreading and up to 40 cm
high. Culms erect, or sprawling and semi-prostrate, geniculate, simple or much
branched from the upper nodes (in older plants, wiry and brittle) pallid to deep purple ;
nodes glabrous; internodes smooth and usually glossy. Leaf-sheaths usually much
shorter than the internodes, the bare culms being a feature of robust plants, glabrous
except for the sparsely villous, membranous margins. Leaf-blades narrowly linear.
Fig. 11. — A. parvula: cross section of the leaf-blade (Schweickerdt 2220).
0-1 mm
Fig. 12. — A. parvula: abaxial epidermis of the leaf-blade.
243
convolute or expanded, up to 2 mm wide, up to 12 cm long. Panicle effuse, 4-12 cm
long, longer than wide, the branches and spikelets spreading when mature; branches
binate, clustered or solitary, angular, scabrid on the angles. Spikelets usually purple
or rarely pallid, spreading. Glumes unequal, boat-shaped, lanceolate, 1 -nerved strongly
awned, glabrous except on the prominent scabrid midrib; lower shorter than the
upper, 4 -5-5 -5 mm long, bifid, awn about 2 mm long; upper 6 -5-7 -5 mm long,
bifid, awn about 2 mm long. Lemma linear, about 7 mm long including the callus,
coarsely scabrid upward especially on the keel, pallid, purple or mottled with purple,
articulated at the apex of the lemma; column grooved, about 5 mm long, 3-nerved,
twisted, scabrous, kneed at the articulation and at the junction of column and awn;
awn solitary 6-9 mm long, finely scaberulous; callus truncate, bristly with bristles
half as long as the body of the lemma. Palea a firm, broadly oblong, sub-truncate,
2-nerved scale, about 1 mm long. Lodicules fleshy in young flowers later membranous,
lanceolate, 5-8-nerved, about 1 • 5 mm long. Anthers linear about 1 • 5 mm long. Ovary
glabrous, styles distinct, stigmas elongate, narrow plumose. Caryopsis linear; hilum
linear; embryo about \ the length of the grain.
Anatomy (Fig. 11, 12 & 159; 1)
Leaf-blade V-shaped or more or less expanded in transverse section, upper and lower surface
slightly undulate the bundle units hardly projecting; lower surface almost glabrous, upper hispidulous;
midrib slightly projecting; margins obtuse. Abaxial epidermis: stomatal zones with 1-3 rows of
stomata and 4-8 rows of elongate, rectangular, ripple-walled cells with strongly undulate walls;
silicified cell zones with 3-5 rows of silicified cells alternating with long narrow ripple-walled cells
opposite the first order bundles and 1-3 rows opposite the third order bundles, silicified cells dumb-
bell-shaped, often with a thin, elongate central part, suberized cells square to narrowly oblong and with
undulate walls, bicellular hairs present flanking the zones, unicellular retrorse barbs very few. Vascular
bundle units: first order units usually 5, groups of 3^ third order units flanking the midrib, 1-3 third order
units between the marginal first order bundles, and single or paired on the margin; units all more or
less of same size and square in outline; midrib only slightly larger than the other first order bundles.
Stereome strands: abaxially fairly well developed opposite the first order bundles, and to a lesser degree
adaxially, in contact with the bundles but often not interrupting the bundle sheaths, very weakly
developed opposite the third order bundles and not interrupting the chlorenchyma or bundle sheaths.
Motor cells well developed, present alternating with all the bundles in Y-shaped or triangular groups
occupying ^ to the full width of the leaf.
Cape. — Kenhardt: Acocks 17601; Namaqualand: Schweickerdt 2536.
South West Africa.— Warmbad: Dinter 5128; Pillans 6372. Keetmanslioop : de
Winter 3474; Acocks 18109; Luderitz; Kinges 2691; Oertendahl 193. Omaruru:
de Winter 3158; Boss s.n. (Tvl. Mus. No. 36372). Swakopmund: Marloth 1202;
Schweickerdt 2220, 2223 and 2239; Boss s.n. (Tvl. Mus. No. 36411). Karibib; Boss
s.n. (Tvl. Mus. No. 36372); Boss 4. Outjo: Galpin & Pearson 7420. Kaokoveld:
de Winter & Leistner 5874 and 5699.
Distribution. — (see Fig. 43).
This species prefers rocky situations in desert and semi-desert areas.
Stapf in the Flora Capensis (1896, p. 572) lists Stipa parvula as a doubtful Stipa
and suspected it to be a single-awned Aristida. Elias (1942, p. 64) concurred with
Stapf’s view after a careful morphological examination. Stapf’s suspicion has been
confirmed by an investigation of the anatomy and external morphology (see Fig. 1
and discussion on p. 212).
The radial arrangement of the chlorenchyma, the presence of two bundle sheaths,
the cells of which contain chloroplasts, as well as the grooved column leaves no doubt
that this species is a true Aristida. The solitary awn would seem to indicate that
A. parvula should be placed in the section Streptachne, which contains those species
of Aristida where the lateral awns are much reduced or absent. In Streptachne, however,
the awn is not deciduous, even though an articulation is present, the awns have a single
bend and the callus bristles are relatively short. In A. parvula the callus bristles are
244
half as long as the body of the lemma, the awn twice geniculate ana deciduous. For
this reason the monotypic section Schizachne, created by Trinius & Ruprecht to
accommodate this species in Stipa, was transferred to Aristida.
Even though A. parvula has an articulated awn, a characteristic not found in any
of the annual species of Sect. Aristida (Chaetaria) it resembles species such as A. ejfusa
and A. scabrivalvis so closely both organographically and in the anatomy of the
leaf-blade, that it could have originated from an ancestor with features similar to these
species.
2. A. hordeacea Kunth, Rev. Gram. 2, t. 173 (1830)
A. hordeacea Kunth var. longiaristata Henr., Crit. Rev. Aristida 2: 244 (1927). A.
steudeliana, Trin. & Rupr., Gram. Stip. 155 (1842).
Fig. 13. — A. hordeacea: cross section of the leaf-blade {de Winter 3097).
r
0’1 mm
Fig. 14. — A. hordeacea: abaxial epidermis of the leaf-blade.
245
Annual, usually erect, branched from the base and lower nodes, 10-90 cm high.
Culms erect to geniculate, several-noded; nodes pubescent; internodes compressed,
densely pubescent, glabrescent. Leaf-sheaths keeled, pubescent; auricles bearded.
Leaf-blades expanded or folded, keeled, scabrous below, hirtellous above, up to 30 cm
long and 1 cm wide. Panicle exserted, spike-like and very dense, or somewhat interrupted
at the base, up to 26 cm long, linear-oblong to sub-ovate. Spikelets greenish or pallid.
Glumes lanceolate, awned, 1 -nerved, keeled, dorsally scaberulous, lower 6-11 mm
long excluding the 2-5-5 mm long awn; upper 7-12 mm long excluding the 1-4 mm
long awn. Lemma linear, fusiform, very scabrid all over, ventrally furrowed, up to
7-5 mm long including the callus, distinctly articulated; callus rounded, bearded,
about 5 mm long; awns subequal, scabrid, 35-50 mm long. Caryopsis linear-lanceolate,
distinctly furrowed ventrally.
Anatomy (Fig. 13, 14 & 159 ; 2)
Leaf-blade expanded and more or less moniliform in transverse section: both surfaces undulate
due to the slightly protruding bundle units; adaxial surface covered with short bristly unicellular
hairs mainly flanking the stereome strands; silicified cells strongly protruding; midrib projecting
strongly abaxially, forming a distinct keel; margins sub-obtuse. Abaxial epidermis: stomatal zones
with one or two rows of stomata, short elements present, often paired, the one cell undulate the other
smooth-walled and produced into a small retrorse barb, ripple-walled cells strongly undulate, elongate-
rectangular, walls thin; bi-cellular linear hairs present; silicified cell zones with silicified cells dumb-
bell-shaped, often very large opposite the first order bundles, suberized cells rectangular, walls undulate
and thin, many retrorse broad-based barbs present flanking the silicified cell zones, or within them in
narrow zones. Vascular bundle units: first order units 7, more or less square to sub-circular; second
order units sub-circular in outline. Stereome strands ad- and abaxially well developed only opposite
the first order bundles, abaxially present in small groups opposite the third order bundles and most
of the motor cell groups; adaxially often present in minute groups opposite the third order bundles,
usually subtending retrorse barbs. Motor cell groups one to two rows of cells wide and occupying
the whole thickness of the leaves.
South West Africa. — Otjiwarongo: Volk 2818; de Winter 2729; Bradfield 440.
Outjo: Volk 2859. Grootfontein : Schweickerdt 2298; Schoenfelder 511; Kinges
2820; 6799. Okavango: de Winter & Marais 5026; Maguire 2328.
Distribution. — (see Fig. 55)
This species is found mainly in shallow depressions and on the edge of small pans
where water collects after rain. Almost invariably it is limited to those depressions
where the soil is heavy clay.
Its closest relative amongst the South African species seems to be hubbardiana,
with which it agrees in spikelet characteristics as well as general habit. It is distinguished
from the latter by the presence of an articulation between the apex of the lemma and
the base of the awns, as well as the generally larger spikelets. The leaf anatomy is also
rather similar (see Fig. 13 and 15).
3. A. hubbardiana Schweickerdt in Notizbl. Bot. Gart. Mus. Berol. 14, 122; 196
(1938)
Annual, slender and rather lax, up to 50 cm high. Culms erect, geniculate several-
noded; nodes glabrous; internodes exserted, glabrous. Leaf-sheaths glabrous, slightly
keeled; auricles usually glabrous. Leaf-blades expanded or folded, almost glabrous
on lower surface, scaberulous on the upper up to 10 cm long, 2 mm wide. Panicle
well exserted when mature up to 6 cm long, obovate in outline. Spikelets subsessile.
Glumes about 7 mm long, linear lanceolate lower awned, scaberulous on the keel and
awn, upper shortly awned, keel glabrous. Lemma up to 7 mm long, very scabrid in
its upper two-thirds, shallowly furrowed ventrally; callus short, rounded, densely
bearded; awns scabrous up to 30 mm long. Caryopsis linear-lanceolate, furrowed
ventrally.
246
Anatomy (Fig. 15, 16 & 159: 3)
Leaf-blade widely V-shaped in transverse section; abaxial bundle units slightly protruding; adaxial
surface with the bundles protruding a little less than one-third of their height; bristle-like unicellular
sharp hairs present, adjacent to the stereome strands; midrib protruding abaxially to form a slightly
projecting keel; margins obtuse. Abaxial epidermis: stomatal zones with very few short elements
and usually one row of stomata, ripple-walled cells shallowly undulate; silicified cell lanes with
sdicified cells dumb-bell-shaped often large and elongate, suberized cells of usual type, unicellular
sharp hairs as well as bicellular linear hairs present. Vascular bundle units: first order bundle units 5;
second order bundle units 3 on either side of the midrib and alternating with the other first order
bundles. Stereome strands abaxially present opposite all bundles as well as the motor cell groups;
adaxially present opposite the second order bundles as minute groups of fibres, but well developed
opposite the first order bundles. Motor cells in Y-shaped groups 1-3 rows wide at the base, and
occupying the whole thickness of the leaf.
I 1
0-1 mm
Fig. 16. — A. hubbardiana: abaxial epidermis of the leaf-blade.
247
Fig. 18. — A. rhiniochloa: cross section of the leaf-blade (Codd 5977).
248
0*1 mm
Fig. 19. — A. rhiniochloa: abaxial epidermis of the leaf-blade.
South West Africa. — Tsumeb: de Winter 2936 ; Dinter 1600 (isotype). Grootfontein .
Schweickerdt 2054; 2059; 2094; 2111; 2143; Kinges 2^51. Outjo: Fo/fc 2850.
249
Distribution. — (see Fig. 17).
A relatively rare species with a limited distribution in the Outjo, Tsumeb and
Grootfontein districts of South West Africa. Like its relative, A. hordeacea it occurs
mainly in seasonally flooded depressions in calcareous clayey soils (see previous species).
4. A. rhiniochloa Hochst. in Flora 38: 200 (1855)
A. andoniensis Henrard, Grit. Rev. Aristida 3: 691 (1928).
A. rigidiseta Pilger in Engl. Bot. Jahrb. 51: 413 (1914).
Annual, erect, very scabrid and coarse, up to 60 cm high. Culms erect, rarely
geniculate, 3-many noded; nodes minutely pubescent; internodes retrorsely scabrid.
Leaf-sheaths rather lax, keeled, scabrous; auricles long-bearded. Leaf-blades expanded,
up to 20 cm long and 4 mm wide, keeled, very scabrous on both surfaces. Panicle
effuse, or contracted and much interrupted, up to 30 cm long, rhachis and branches
very scabrid, axils shortly hairy. Spikelets coarse, pallid or purplish-brown. Glumes
broad, very acute, scaberulous to almost smooth, 1 -nerved, awned; lower up to 17 mm
(usually about 14 mm) long, gradually tapering into the awn; upper up to 15 mm long,
with two lateral teeth flanking the awn. Lemma up to 13-5 mm (usually about 1 1 mm)
long with lines of coarse antrorse barbs from base to apex, very rarely almost smooth;
keeled on the back, deeply grooved ventrally with in-rolled margins ; callus sub-obtuse,
bearded; awns rigid, coarsely scabrid 18-30 mm (rarely 40 mm) long. Caryopsis
narrowly lanceolate deeply grooved ventrally.
Anatomy (Fig. 18, 19 & 159: 4)
Leaf-blade expanded in transverse section; abaxial surface flat or slightly undulate, covered with
short, sharp, bristly hairs flanking, or opposite, the stereome strands; adaxial surface strongly undulate,
due to the projecting bundle units, covered with short bristly hairs; midrib strongly projecting abaxially
forming a distinct keel; margins obtuse. Abaxial epidermis: stomatal zones with one or two rows
of stomata short elements present but few, these occasionally produced into short sharp hairs; ripple-
walled cells elongate-rectangular with strongly undulate, rather thin walls, numerous linear bicellular
hairs present; silicified cell zones with silicified cells dumb-bell-shaped, larger in the wider zones,
suberized cells rectangular with undulate thin walls; many broad-based retrorse barbs present flanking
the wider, and present within the narrower, zones. Vascular bundle units: first order bundle units 5,
subcircular in outline, somewhat flattened ab- and adaxially; second order units subcircular. Stereome
strands ad- and abaxially well developed only opposite the first order bundles, abaxially present in small
groups opposite the third order bundles as well as the motor cell groups; adaxially in minute groups
usually subtending unicellular hairs. Motor cell groups one to two rows of cells wide and occupying
the whole thickness of the leaf.
Transvaal. — Kruger National Park: Brynard & Pienaar 4270; de Winter and Codd
647; van der Schijjf 1879. Waterberg: Codd 8503. Rustenburg: Acocks 18754;
Codd 8678.
South West Africa. — Okahandja: Gaerdes 16. Otjiwarongo: Liebenberg 4906;
Dinter 5754; de Winter 2723. Tsumeb: de Winter 2927. Outjo: de Winter and
Leistner 5138. Kaokoveld: de Winter and Leistner 5378.
Distribution. — (see Fig. 20).
A very coarse, hard species occurring on rocky slopes, gravelly flats, sand or loam,
in dry to very dry, often disturbed, situations.
A. rhiniochloa is a distinct species with no obvious close relatives. Sturgeon
(1954; p. 13) states that A. serrulata described by Chiovenda (and which he later
regarded as a variety of A. rhiniochloa) is, in her opinion, a closely related, but distinct,
species. No material of this plant was available for study. After examining type
material of A. andoniensis it became clear that this species represents a form of A.
rhiniochloa with rather glabrous lemmas. The wide range of material of A. rhiniochloa
now available for study has shown that the scabridity of the lemma varies considerably.
250
It is probable that A. hordeacea, A. hubbardiana and A. rhiniochloa are related
even though the latter is a much coarser plant than the others, with a different panicle.
All three species have deeply ventrally grooved caryopses. In addition they are all
annuals, the lemmas are ventrally grooved, scabrid and lack a column. It seems likely
that A. hordeacea should be regarded as allied to the other two species in spite of the
presence of an articulation. Anatomically they also agree by having expanded leaves
with abaxially protruding midribs and a very similar type of bundle unit.
5. A. adscensionis L. subsp. guineensis {Trin. & Rupr.) Henrard, Crit. Rev. Aristida
1: 216 (1926)
A. guineensis Trin. & Rupr., Gram. Stip. 137 (1842).
Annual, forming erect to lax sprawling tufts, up to 75 cm high. Culms erect,
sometimes strongly geniculate; nodes glabrous; internodes glabrous, smooth, often
purplish, the basal ones bare. Panicle narrow but rather laxly branched and interrupted,
never spike like. Spikelets pallid or purphsh. Glumes unequal, 1 -nerved, keeled, at
least upwards, usually tinged with purple, minutely scaberulous on the keels, otherwise
glabrous or lower minutely hispidulous; lower 5-7 mm long erose-bifid or acute;
upper 6-8-5 mm long erose-bifid and occasionally mucronulate. Lemma 8-14 mm
long, well exserted from the glumes when mature, pallid, purplish, or mottled with
purple, glabrous and smooth, keels coarsely scabrid in the upper half; callus rounded,
shortly bearded; awns scabrid, central 12-17 mm long, lateral 8-12 mm long. Caryopsis
linear, terete.
Fig. 21. — Distribution of adscensionis subsp. guineensis; MA. stipoides.
251
Anatomy of the leaf-blade as in A. ciirvata (Fig. 159: 5, p. 253)
Transvaal. — Pretoria; Repton 1738; 2120. Pietersburg: Briggs 10. Waterberg:
Carver 8. Barberton; Thorncroft 43. Kruger National Park; Obermeyer 36088;
van der Schijff 396; 1643. Soutpansberg; Codd & Dyer 3904; de Winter & Codd
648; Codd 5801; Acocks 17850.
Distribution. — ^(See Fig. 21.)
Closely allied to A. curvata (cf. notes, p. 254) and A. ejfusa. Differs from A. effusa
in the smaller, more graceful, spikelets and, particularly, in the lemmas which are
scab rid on the keels only. A. effusa occurs only in South West Africa and A. adscensionis
subsp. guineensis in the Transvaal, from Pretoria northwards into tropical Africa.
Anatomically these species are indistinguishable.
6. A. effusa Henrard, Crit. Rev. Aristida 1; 155 (1926)
A. caerulescens Desf. var. breviseta Hack, in Engl. Bot. Jahrb. 11; 400 (1889). A.
waibeliana Henrard, Crit. Rev. Aristida 3; 679 (1928).
Fig. 22. — A. effusa: cross section of the leaf-blade {de Winter 3526).
Fig. 23. — A. effusa: abaxial epidermis of the leaf-blade.
252
Fig. 24.— Distribution of MA. effusa; %A. bipartita.
Annual, slender, erect, up to 90 cm high. Culms branched from most nodes;
nodes glabrous; internodes glabrous, usually exserted. Leaf-sheaths lax, slipping from
the culms- auricles glabrous to ciliate. Leaf-blades up to 30 cm long and 3 mm wide,
expanded. Panicles effuse and open, 20-30 cm long and 15 cm wide, or much smaller;
branches spreading, flexuous, bearing the spikelets at the apex in short spike-hke clusters.
Spikelets usually purplish. Glumes equal to slightly unequal; lower lanceolate up to
8-5 mm long, acute or mucronulate, scaberulous towards the apex, scabrid on the
keel, 1 -nerved; upper to 8 mm long, truncate or very obtuse, 1 -nerved, keel smooth.
Lemma linear up to 1 1 • 5 mm long or much shorter, very scabrous upwards, shortly
to much exserted beyond the glumes; callus rounded, bearded; awns eiect, scabrid,
central awn about 16 mm, lateral awns about 13 mm long. Caryopsis linear, terete.
Anatomy as for A. curvata except that the leaves of A. curvata are somewhat more robust (Fig. 22,
159: 6).
South West Africa. — ^Gobabis; de Winter 2467; 2437. Windhoek: Liebenbetg
4455' 4432; de Winter and Giess 1\26\ 7114; 7132; 7134; de Winter 3526. Oka-
handja; Bradfieldl3%. Otjiwarongo: de Winter 2125 Volk 2925. Karibib: Kmges
3055- de Winter 2661. Tsumeb: Volk 1672. Grootfontein : Kmges 2882; 2780;
de Winter and Giess 6784; Schoenfelder 99; de Winter 2930. Outjo: Volk 2868.
Kaokoveld: de Winter and Leistner 5174.
Distribution. — ^(See Fig. 24.)
A species distinguished by its effuse panicle, with the spikelets contracted at the
ends of the branches, and by the scabrid lemmas which are exserted from the glumes
when mature. Related to A. adscensionis subsp. guineensis, which has the lemma scabrid
only on the keel and the spikelets not congested at the ends of the branches. Dis-
tinguished from A. scabrivalvis by the awnless glumes and exserted lemma.
7. A. curvata (Nees) Trin. & Rupr., Gram. Stip. 133 (1842).
Chaetaria curvata Nees, FI. Afr. Austr. 186 (1841). Chaetaria curvata Nees ^ minor
Nees, l.c. p. 187. Chaetaria mauritiana (Kunth) Nees var. nana Nees, l.c. p. 188.
Aristida confusa Trin. & Rupr., Gram. Stip. 134 (1842). A. pusilla Trin. & Rupr.,
l.c. p. 140. A. pygmaea Trin. & Rupr., l.c. p. 133. A. strictiflora Trin. & Rupr., l.c.
p. 134. A. adscensionis L. var. pygmaea (Trin. & Rupr.) Dur. and Schinz, Consp. 5:
800 (1894). A. adscensionis L. var. strictiflora (Trin. & Rupr.) Dur. and Schinz, Consp.
l.c. A. curvata (Nees) Trin. & Rupr. var. nana (Nees) Henrard, Crit. Rev. Aristida 3:
487 (1928).
Annual, erect or sprawling, 10-100 cm high. Culms erect, geniculate, often branched
from the base and lower nodes; nodes glabrous, internodes glabrous. Leaf-sheaths
tight or lax, keeled, smooth or scaberulous; auricles pubescent. Leaf-blades linear,
expanded or folded, up to 15 cm long and 1-2-5 mm wide, glabrous on the lower
surface, scabrid on the upper. Panicle narrow and dense, or interrupted, up to 24 cm
long and 1-2-5 cm broad. Spikelets pallid, green, or tinged with purple. Glumes
more or less keeled, unequal, linear-lanceolate to lanceolate; lower 4-7 mm long,
emarginate and mucronate from the sinus, scabrous on the keel and strigose on the
margins ; upper up to 6-8 mm long, scaberulous upwards, bifid with a mucro from the
sinus. Lemma usually mottled with purple, finely punctulate and scabrous on the keel
up to 10 mm long, usually as long as, or slightly shorter than, the glumes, rarely slightly
exceeding them; callus rounded, bearded; column absent; awns unequal scabrid,
central awn about 20 mm long, lateral awns about 15 mm long. Caryopsis linear,
subterete.
Anatomy (Fig. 25 and 159: 7)
Leaf-blade flat in transverse section; adaxial surface more or less undulate; midrib usually somewhat
protruding forming a keel; adaxial surface shallowly grooved due to the slightly protruding bundle
units, and with few to many hairs opposite the bundles. Abaxial epidermis: stomaial zones with one
to two rows of stomata, and a few to many rows of elongate-rectangular ripple-walled cells with
strongly undulate walls; many two-celled linear hairs present; silicified cell zones: silicified celts
dumb-bell-shaped, short or elongate with a long, thin, middle portion; suberized cells more or less
square to elongate-rectangular, thin-walled with undulate walls; long ripple-walled cells narrower
and with finer more compactly undulate walls than those of stomatal zones; short broad-based retrorse
barbs present. Vascular bundle units: first order units usually 5 (occasionally 3), mostly larger than the
bundles of lower rank, midrib flanked by 4-7 second or third order bundles, marginal first order bundles
approximate, alternating with 1-3 second or third order bundles, bundle units subcircular to ovate
(second or third order) or more or less square to subcircular (first order) in outline. Stereome strands
usually weakly developed in all bundles or relatively well developed ab- and adaxially opposite the first
order bundles; present ab- and adaxially, often as a few fibres only, opposite most of the second and
third order bundles. Motor cells in ob-triangular or oblong groups of thin-walled cells, 1-2 cells wide
and usually occupying the whole thickness of the leaf.
Cape. — Campbell: Brueckner 1028; Kenhardt: Codd 1194. Hay; Esterhuysen 2372.
Swellendam: Esterhuysen 1819. Namaqualand; Schlechter 11226.
Basutoland. — Quthing: Dieterlen 1190.
Orange Free State. — Boshof: Pole Evans HI 1603. Fauresmith; Smith 5571.
Transvaal. — Potchefstroom : Louw 1435; 1273. Pretoria: Schweickerdt 1326; Bar-
berton: Thorncroft 43. Potgietersrust ; Galpin 8892. Ermelo: Henrici 1473.
South West Africa. — Warmbad: Oertendcihl \\%. Keetmanshoop; de Winter 3415.
Gibeon: van Vuuren and Giess 1094. Luderitz: van Vuuren and Giess 802; 2340.
Rehoboth; Strey 2102. Windhoek: de Winter and Giess 7143; 7125; 7131; 7112.
Grootfontein ; Kinges 2%%\. Kaokoveld: de Winter and Leistner 5324.
254
Distribution. — (See Fig. 26.)
Examination of a large number of specimens has shown that the sheets designated
by Schweickerdt as A. submucronata can be distinguished from A. curvata only by their
generally more robust habit. Due to a lack of specimens from tropical Africa it was
not possible to decide whether A. curvata should be regarded as synonomous with
A. submucronata and, for this reason, A. curvata is retained as a distinct species. The
key character used by Henrard and Schweickerdt, i.e. the acute, minutely awned lower
glume of A. submucronata and the erose-bifid mucronate lower glume of A. curvata
does not seem to hold for the range of specimens now available. A. curvata is also
closely allied to A. adscensionis subsp. guineensis from which it can be distinguished
by the smaller spikelets and more open inflorescence of the latter. This whole group
{A. curvata, A. adscensionis and A. submucronata) should be re-examined, taking into
consideration the entire distribution, as well as the more localized distribution of the
various forms. In this connection the reader is referred to the discussion on the inter-
relationship of the South African species (pages 301-303). Anatomically the species
of this group are indistinguishable except for slight variations in the width of the
leaf-blade and, to some extent, the development of the stereome strands, characteristics
of too little importance to make the anatomy of value in the delimitation of the various
species.
O • 1 mm
Fig. 25. — A. curvata: abaxial epidermis of the leaf-blade (Theron 103).
Guenzel (1913, p. 18) described the leaf anatomy of Aristida adscensionis L. The
specimens examined by him, however, came fiom South West Africa where this species
sensu stricto does not occur. He probably described A. curvata since this species has
in the past often been mis-identified as A. adscensionis. Guenzel’s description and
diagramatic drawing of the cross section of the leaf-blade agree closely with the present
author’s observations on A. curvata.
8. A. scabrivalvis Hack, in Bull. Herb. Boiss, 2, 6: 708 (1906)
Annual, erect or ascending, up to 85 cm high. Culms much branched, 3-4-noded,
nodes glabrous; internodes glabrous. Leaf-sheaths lax, keeled, minutely scaberulous
or smooth; auricles ciliolate or glabrous. Leaf-blades linear up to 30 cm long and
3-5 mm wide, glabrous on the lower surface, hirtellous on the upper. Panicle open
255
and divaricate up to 30 cm long or much smaller, ovate in outline. Spikelets purple
or purplish brown. Glumes linear-lanceolate, strongly awned, subequal or unequal,
hirtellous, or at least with a scabrid keel; lower up to 9 mm long including the awn,
awns up to 2-5 mm long; upper strongly bifid, up to 9 mm long, awn up to 1-5 mm
long. Lemma linear, as long as or shorter than the glumes, keeled, 5-8 mm long,
coarsely scabrid in the upper part or at least on the keel; callus about 1 • 5 mm, rounded,
bearded; awns scabrid about 14 mm long. Caryopsis linear, terete.
SC
Fig. 27. — A. scabrivalvis: cross section of the leaf-blade (de Winter 775).
256
0 • 1 mm
RB
ST
BH
SC
CC
Fig. 2S.—A. scabrivalvis: abaxial epidermis of the leaf-blade.
Anatomy as for A. curvata (Fig. 27, 28 and 159: 8)
Fig. 27 is not entirely typical: the two well developed marginal bundles usually alternate with
a third order bundle and are not adjacent as in the figure.
257
A distinct species recognized by the very large inflorescence, awned usually
hirtellous glumes, and scabrid lemma, which is subequal to or shorter than the glumes.
Allied to A. effusa and A. bipartita. Differing from A. effusa in the awned glumes, and
in the lemma which equals, or is shorter than the glumes; and from A. bipartita in
the upper glume exceeding the lower in length. Two varieties can be recognized.
Panicle branches which arise from the main axis effuse, and bearing several secondary branchlets
at least towards the base of the panicle, branchlets effuse or somewhat contracted var. scabrivalvis
Panicle branches arising from the main axis contracted, somewhat spike-like and stiffly spreading
var. contracta
(a) var. scabrivalvis.
A. scabrivalvis Hack, in Bull. Herb. Boiss. 2, 6: 708 (1906).
Cape. — Barkly West: Acocks 1809; Brueckner 1272; Kimberley: Leistner 2274.
Transvaal. — Bloemhof: Leistner 28. Vereeniging: Burtt-Davy 4850. Pretoria:
Smith 6101. Waterberg : Cot/t/ 904. Potgietersrust: 427; 8891. Kruger National
Park: van der Schijff 16^2,-, 1720; de Winter 528.
Natal. — Estcourt: Acocks 10130; Perry 148. Lower Umfolosi: Letley 22.
South West Africa. — Otavi: Dinter 5752. Grootfontein : de Winter and Giess 6794;
de Winter 2929. Okavango: de Winter 4068.
Distribution. — (See Fig. 29.)
{b) var. contracta de Winter in Kirkia 3: 132 (1963). Type: Transvaal, Sout-
pansberg, de Winter and Codd 321 (PRE, holo.!).
Transvaal.^ — ^Letaba: de Winter and Codd 690; 321.
Natal. — Weenen: Acocks 10530; Edwards 1237.
Distribution. — (See Fig. 29.)
9. A. bipartita (Nees) Trin. & Rupr., Gram. Stip. 144 (1842)
Chaetaria bipartita Nees, FI. Afr. Austr. 187 (1841).
Perennial or weak perennial up to 65 cm high, with a short erect or oblique rhizome.
Culms erect or ascending; nodes glabrous, internodes glabrous. Leaf-sheaths firm,
glabrous or minutely puberulous between the nerves ; auricles long bearded. Leaf-blades
up to 20 cm long and up to 2 mm wide, rigid, smooth oi scaberulous on the lower
surface, scabrid on the upper. Panicle effuse, up to 30 cm long and nearly as wide
but usually much smaller; branches spreading, solitary, ffexuous, naked for most of
the length, and bearing clusteis of 1-4 spikelets at the apex. Spikelets pallid oi purplish.
Glumes unequal, the lower exceeding the uppei in length, 1 -nerved, mucronate or
shortly awned; lower smooth or scabrous on the keel upwards, up to 11 mm long,
uppei smooth up to 9 mm long. Lemma linear, included in the glumes, about 8 mm
long, smooth, often mottled with purple; callus obtusely rounded, bearded; awns
subequal, central slightly longer scabrous, 7-13 mm long. Caryopsis linear, terete.
Anatomy (Fig. 30 and 159: 9)
Leaf-blade flat or very widely V-shaped in transverse section; abaxial surface flat with very few
hairs; adaxial surface grooved between the protruding bundle units which are furnished with few
to many short bristly unicellular hairs. Abaxial epidermis: stomata! zones with usually one row of
stomata; long ripple-walled cells elongate-rectangular with strongly undulate walls; two-celled linear
hairs present; silicified cell zones: silicified cells dumb-bell-shaped; suberized cells usually shorter
than the silicified cells, thin-walled, when very short often irregularly dumb-bell-shaped, when longer
with undulate walls; long ripple-walled cells with finer undulations than those of the stomatal zones;
short retrorse barbs flanking the silicified zones usually accompanied by a short element (suberized
cell). Vascular bundle units: first order units 5, only slightly larger than the third order units; midrib
smaller than the other first order units; broadly ovate in outline in contrast with the others which are
more or less square in outline; third order bundle units ovate to subcircular. Stereome strands present
ab- and adaxially to all bundles or absent adaxially to some of the third order bundles, usually present,
even though small, opposite the motor cell units; strands often fused together abaxially opposite the
first order bundles and adjacent motor cell groups. Motor cells in groups, one to two cells wide and
occupying the whole thickness of the leaf. See also Theron (1936, p. 25).
258
O' 1 mm
Fig. 30. — A. bipartita: abaxial epidermis of the leaf-blade {Acocks 9563).
Distribution. — fSee Fig. 24.)
Cape. — Komga; Flanagan 1175. Queenstown: Galpin 2578. Stutterheim : Acocks
9563. King William’s Town: Story 917.
Basutoland. — Mohales Hoek: Dieterlen 1208.
Orange Free State. — Heilbron: BrandmuUer 108. Senekal: Goossens 930. Beth-
lehem: Goossens 1123.
Transvaal.— Lichtenburg: Kinges 1631. Standerton: Marais 10. Pretoria: Mogg
13,727. Waterberg : de Winter 736. Rustenburg : Acocks 8840. Potgietersrust :
de Winter 2275.
Natal. — Hlabisa: Ward 2319. Estcourt: Acocks 9869. Dundee: Codd 200.
Fig. 31. — A. canescens subsp. canescens: cross section of the leaf-blade (de Winter 195K).
259
Fig. 32. — A. canescens subsp. canescens : abaxial epidermis of the leaf-blade.
Fig. 33.— Distribution of A. canescens: #subsp. canescens; Osubsp. ramosa; A. engleii: Mvar.
engleri; Qvar. ramossissima.
260
A species recognized by its very lax spreading inflorescence with long flexuous,
bare branches bearing a few spikelets at the ends, and by the inverse position of the
glumes, i.e. the lower exceeding the upper in length.
Though related to annual species such as A. effusa and A. scabrivalvis, A. bipartita
is a perennial with a stronger development of the stereome strands of the leaf-blade.
Anatomically it can, therefore, be distinguished from its close relatives.
10. A. canescens Henrard, Crit. Rev. Aristida, Suppl. 708 (1933)
Perennial up to 1 m high, densely caespitose, erect. Culms erect, simple or
branched; internodes terete or slightly compressed, glabrous ; nodes glabrous. Leaf-
sheaths tight, glabrous or scaberulous; auricles bearded or shortly pubescent. Leaf-
blades linear, hard and coarse, expanded near the base and curling when old, margins
thickened, glabrous or scaberulous on the lower, and scabrid on the upper surface.
Panicle erect, lax or contracted, but much interrupted, up to 20 cm long. Spikelets
pallid, yellowish green or tinged with purple. Glumes unequal, 1-nerved, lanceolate,
emarginate or subtruncate and shortly muconate, only the lower scaberulous on the
keel and occasionally on the flanks, 5-5-8 mm long; upper 8-11 mm long. Lemma
7-11 mm long, smooth or scaberulous on the keel, without a column; callus obtuse,
bearded ; awns scabrid, subequal, central 9-1 5 cm long, lateral 7-1 3 mm long. Caryopsis
linear, terete; embryo ^ the length of the grain.
Anatomy (Fig. 31, 32 and 159: 10)
Leaf-blade flat in transverse section with the midrib small (equal to third order bundles) and the
lateral first order bundle units enlarged, together forming a strong marginal rib; abaxial surface
undulate with very few hairs visible in cross section; adaxial surface shallowly grooved between the
slightly protruding third order bundles and with a deep narrow groove between the marginal first
order bundles; covered with short bristly hairs opposite the bundles; margins obtuse. Abaxial
epidermis: stomatal zones with one to two rows of stomata; short elements few; ripple-walled cells
elongate-rectangular or with the end walls oblique, walls very strongly undulate and rather thick;
linear bicellular hairs numerous; silicified cell zones: silicified cells dumb-bell-shaped, in the wider
zones larger and with elongate, much constricted, central portion; suberized cells rectangular, undulate,
thin-walled; few broad-based, unicellular retrorse hairs present, fairly long and thin, or short and
barb-like. Vascular bundle units: first order units 5, midrib equal in size to the third order bundles
but marginal first order bundle units much enlarged, with the inner more or less square in outline,
the other forming a unit with the marginal second order bundle and tapering obliquely into the obtuse
margin; third order units subcircular or ovate. Stereome strands: most strongly developed abaxially
opposite the first order units of the marginal ribs, less strongly so adaxially; the remaining bundle
units, including the midrib, with smaller strands present both ab- and adaxially; motor cells supported
by abaxial groups of fibres. Motor cells in groups 1-2 rows wide, occupying almost the whole thickness
of the leaf, stereome strands present abaxially.
Culms mostly unbranched and erect, often robust subsp. canescens
Culms geniculate and often branched from the nodes subsp. ramosa
(fl) subsp. canescens.
A. canescens Henrard, Crit. Rev. Aristida Suppl. 708 (1933).
Erect perennial forming usually dense tufts; culms unbranched.
Cape. — Taungs: Henrici 39. Vryburg; Theiler 226. Barkly West: Brueckner 837.
Natal. — -Weenen: Edwards 1238. Estcourt: Acocks 11424.
Orange Free State. — Kroonstad: Pont 149. Heilbron: Goossens 425.
Transvaal. — Bloemhof; Burtt-Davy 13057. Potchefstroom : Story 755. Bethal:
Kinges 1379. Pretoria: de Winter 795A ; 7561. Waterberg: Straker 11. Pot-
gietersrust : Galpin 426.
{b) subsp. ramosa de Winter in Kirkia 3: 132 (1963). Type: Cape, Cathcart,
Kei Valley at Kei View, Acocks 9691 (PRE, holo!).
261
Perennial, usually geniculate and branched from the nodes, rarely some cuhns
unbranched.
Cape. — King William’s Town: Sim 2835. Cradock : Brynard 181. Bedford: Story
4554; Acocks 12804.
Distribution. — (See Fig. 33.)
The subsp. ramosa has a distribution quite distinct from that of subsp. canescens.
It has so far been recorded only from the central areas of the eastern Cape Province.
Fig. 34. — A. aequiglumis: cross section of the leaf-blade {de Winter 7515).
0’1 mm
5752152-3
Fig. 35. — 1. aequiglumis: abaxial epidermis of the leaf-blade.
262
A distinct species occasionally confused with A. junciformis but distinguished
from the latter by the terete culms and the coarser spikelets, the lack of a column and
the less tough basal parts with thinner fibrous roots. Anatomically readily distinguished
from A. junciformis by the much enlarged marginal bundle units which in size well
exceed the units adjacent to the midrib. In A. junciformis, however, all the bundles
when viewed in cross section approximately reach the same height, even though the
third order bundle units are narrower. Though superficially similar, these two species
are not closely related.
Fig. 36. — Distribution of 0A. aeqiiigliimis ; mA. vestita.
11. A. aequiglumis Hack, in Bull. Herb. Boiss. 3; 381 (1895)
Perennial, densely caespitose usually about 40 cm high. Culms slender simple
or rarely branched from the upper nodes; nodes glabrous; internodes glabrous.
Leaf-sheaths glabrous, persistent. Leaf-blades erect, setaceous, convolute, up to 15 cm
long. Panicle linear-oblong, contracted or lax and open, about 10 cm long. Glumes
1 -nerved with a prominent glabrous or scabrid midrib, scabrid to shortly hairy on the
sides, apices usually somewhat recurved ; lower 7-1 1 mm long ; upper 7-1 1 • 5 mm
long. Lemma including the column and callus 8-16 mm long, minutely scaberulous
upwards; column twisted 2-8 mm long; awns somewhat unequal, central 15-30 mm
and lateral 14-26 mm long; callus bearded obtuse to sub-obtuse. Caryopsis linear,
terete.
Anatomy (Fig. 34, 35 and 159: 11)
Leaf-blade more or less flat to somewhat curved in transverse section; bundle units except for
the marginal ones only slightly protruding adaxially, abaxial surface very slightly undulate; marginal
units very strongly projecting adaxially and forming distinct marginal ribs; long unicellular hairs
present opposite the adaxial stereome strands; midrib smaller than the marginal bundles and not
forming a keel; margins obtuse or subacute. Abaxial epidermis: stomatal zones with one to three
rows of stomata, and six to twelve rows of more or less rectangular long ripple-walled cells with a
few short elements scattered amongst them; bicellular hairs with rather short and broad basal cells
present in, or flanking, this zone; silicified cell zones consisting of one to seven files of silicified cells,
alternating with one or two rows of long ripple-walled elements; silicified cells dumb-bell-shaped,
those opposite the first order bundles very large; suberized cells thin-walled with undulate walls,
slightly shorter to much shorter than the silicified cells, and usually alternating with the latter; unicel-
lular hairs often slender and sharp-pointed and usually found flanking or present in the silicified cell
263
zones, occasionally absent. Vascular bundle units: first order units 3, mid-vein small, the marginal
units on each side formed by a very large first order unit combined with a third order unit; mid-vein
flanked by 3 to 5 small second order bundles. Stereome strands consisting of small thick-walled fibres;
well developed both ad- and abaxially opposite the first order bundles, but less so in the midrib than
in the marginal bundles; absent abaxially opposite the second order bundles, but present opposite the
motor cell groups; small groups present adaxially opposite the second order bundles. Motor cells in
Y-shaped groups, occupying the whole width of the leaf and alternating with all bundles.
Transvaal. — ■Potchefstroom : J. F. v. P. 616. Krugersdorp ; Acocks 18718. Johan-
nesburg: Leendertz 6115. Pretoria: de Winter 15] 5; Schweickerdt 1343; 1328 and
1753. Rustenburg : Coddl041. Waterbeig: Galpin M 1\3. Middelburg : Schlechter
4129.
Distribution. — (See Fig. 36.)
A reasonably well defined species usually found growing in rocky situations such
as on hill slopes. A. aequiglumis greatly resembles A. transvaalensis and often occurs
in association with it. As pointed out by Schweickerdt it can, however, be distinguished
by the less contracted inflorescence and larger spikelets. Even though A. aequiglumis
IS occasionaly somewhat branched the habit is a useful characteristic to distinguish
it from the copiously branched A. transvaalensis which, moreover, has the leaves
spreading at more or less right angles to the culms, while those of A. aequiglumis are
erect resulting in a more “ bunched ” appearance.
Anatomically the two species are very similar; the marginal bundle units being
very strongly developed and exceeding the others in size. The marginal units project
towards the midrib and in A. aequiglumis are more gibbous (in cross-section) than in
A. transvaalensis.
Fig. 37. — A. transvaalensis: cross section of the leaf-blade {de Winter 7513).
12. A. transvaalensis Henrard, Crit. Rev. Aristida Suppl. 742 (1933)
Densely caespitose perennial, up to 70 cm high. Culms erect, branched from
most of the upper nodes; nodes and internodes glabrous. Leaf-sheaths usually lax
and slipping from the culms, or culm-sheaths tights. Leaf-blades convolute, setaceous,
those on the culms spreading, up to 15 cm long, branches solitary, congested. Glumes
264
lanceolate, unequal, 1 -nerved, acute, glabrous or scabrid on the keels, lower 5-9 mm
long, shortly awned; upper 6-5-10 mm long, minutely awned from a subtruncate
apex. Lemma smooth, minutely granular upwards, with the column up to 11-5 mm
long; column well developed or almost absent, 1-6 mm long, scabrid, twisted; awns
unequal, the central divergent and 6-13 mm long; lateral very short (almost absent)
to 8 mm long. Caryopsis linear, terete.
Fig. 38. — A. transvaalemis: abaxial epidermis of the leaf-blade.
Anatomy: as for A.aequiglumis except for the marginal units which do not project inwards as strongly
as in A. aequiglumis [i.e. they are less gibbous in cross section (Fig. 37, 38 and 159: 12, pp. 263 and
395].
Transvaal. — Johannesburg: Hitchcock 24116. Pretoria: Moss 14259 (isotype);
de Winter 7513; Schweickerdt 1327; 1331 and 1332. Letaba : Scheepers 933. Nel-
spruit: v.d. Schijff 3073. Barberton: Thorncroft 21.
Swaziland. — Meeuse 10135.
Natal. — New Hanover: Acocks 13774. Kranskop: Fisher & Schweickerdt 85.
Distribution. — (See Fig. 17.)
Fig. 39. — A. nwnticola: cross section of the leaf-blade (Mogg H. 20634).
265
Related to A. aequiglumis and A. canescens.
A species found on rocky outcrops and hillsides, usually mainly on dry northern
slopes.
As pointed out by Schweickerdt the lateral awns of this species are often weakly
developed so that some spikelets may appear to have a single awn. The column also
is very variable, even in spikelets of the same individual.
01 mm
Fig. 40. — A. monticola: abaxial epidermis of the leaf-blade.
13. A. monticola Henrard, Crit. Rev. Aristida 2: 355 (1927)
Perennial usually rather sprawling and forming large masses, often rooting at the
lower nodes up to 70 cm high. Culms erect, geniculate and usually branched at the
nodes; nodes glabrous; internodes glabrous, thin, wiry. Leaf-sheaths tight, shorter
than the internodes; auricles long-bearded. Leaf-blades linear, expanded or convolute
upwards, up to 12 cm long and 2 mm wide, glabrous on the lower, and scaberulous
on the upper, surface with occasionally a few long hairs, margins thickened. Panicle
erect, rather open, lanceolate, up to 14 cm long. Spikelets often purplish brown.
Glumes linear-lanceolate, 1 -nerved, lower 6-8-5 mm long, shortly awned, glabrous
with a scabrid keel; upper glabrous, 5-6 mm long, bidentate with a short mucro.
Lemma linear, smooth, 6-6-5 mm long and produced into a short twisted column;
callus obtuse, bearded; awns unequal, central up to 18 mm, lateral up to 13 mm long.
Cary apsis not seen.
Anatomy: as for A. aequiglumis but usually 4-5 third order bundle units flanking the midrib and with
the marginal units somewhat less strongly developed (Fig. 39, 40 and 159: 13).
Natal. — Estcourt; Acocks 11431. Bergville: Edwards 570 and 642; Killick 1258;
1382; 1951; . 1915; 1000 and 1056; MeeboM s.n. (NH 15728); Mogg s.n. (H. 20634)
(isopleth). Weenen: Lilians 15487.
Distribution.— (See Fig. 73.)
Related to A. aequiglumis and A. transvaalensis but distinguished from both these
species by the creeping rhizome. The presence of a well developed rhizome in Aristida
is very rare. A species of local occurrence in the Drakensberg in the Bergville and
Estcourt districts of Natal and remarkable for its preference for moist situations such
as stream-banks and seepage areas on steep slopes.
266
In anatomical characteristics very close to A. transvaalensis and A. aequiglumis,
this character together with many organographic similarities indicating a close rela-
tionship.
Theron (1936; p. 21) investigated this species anatomically but gives no drawing.
His description agrees in main details with the author’s observations.
0 -1 mm
Fig. 42. — A. jimciformis subsp. junciforniis: abaxial epidermis of the leaf-blade.
14. A. junciformis Trin. & Rupr., Gram. Stip. 143 (1842)
Perennial usually densely caespitose up to 60 cm high, with a fascicled base and
very thick almost spongy roots. Calms erect, simple or sparingly branched, wiry, 3-4-
267
noded; nodes glabrous, internodes glabrous; compressed below the nodes. Leaf-sheaths
glabrous or sparsely woolly. Leaf-blades narrow, gradually passing into the sheaths,
sub-setaceous, folded, up to 30 cm long and 1 mm wide. Panicle narrow, contracted
or rather lax, up to 20 cm long and 1-8 cm wide, erect or somewhat nodding; branches
with spikelets congested towards the ends. Glumes unequal, linear-lanceolate; glabrous
or minutely pubescent, subacute, or acute and awned; lower 4-9 mm long: upper
8-12 mm long. Lemma up to 9 cm including the column, smooth or scaberulous
upwards; column well developed or practically absent; central awn 12-35 mm long;
lateral awns 9-28 mm long; callus bearded, obtuse. Caryopsis linear, terete.
Fig. 43.— Distribution of juncifonnis subsp. juncifonnis; mA. pai vula.
Anatomy (Fig. 41, 42 and 159: 14).
Leaf-blade V-shaped in transverse section, keel not protruding; adaxial surface with the bundle
units projecting strongly, furnished with short bristly hairs adjacent to the steieoine stiarids, abaxial
surface, smooth and practically hairless; margins subacute to obtuse. Abaxial epidermis, stomatal
zones of 6-9 rows of rectangular narrowly-oblong ripple-walled cells alternating with transversely
narrowly-oblong short elements and with one to four rows of stomata; silicified cell zones with few
to many rows of dumb-bell-shaped silicified cells alternating with short, thin-walled, transversely
oblong to oblong suberized cells with undulate walls; unicellular hairs not observed; bi-cellular hairs
present in, or flanking the stomatal zones, or occasionally even in the silicified cell zones. Vascidar
bundle units more or less oblong in outline; first order units 3, midrib flanked by three 1“'™
bundles, all units approximately equal in height, gradually diminishing in size towards the midiib
which usually is slightly smaller than the other bundles, and ovate to circular in outline. Stei eome
strands present abaxially opposite the motor cell groups as well as the bundles except for the two third
order bundles on either side of the midrib; the strands of the motor cells flanking the third older
bundle units inside the marginal first order bundle, usually fused to form a single strand subtending
the two motor cell groups as well as the bundle; the adaxial strand fairly strongly developed opposite
268
the bundles, or consisting only of small groups of fibres. Triangular or oblong motor cell groups
present and alternating with all the bundles; the groups on either side of the third bundle from the
midrib at times fused laterally to form a compound unit enclosing the bundle abaxially.
Culms 2^-noded, forming strictly erect tufts, often quite tall and not very densely leafy at the
base; panicles usually rather robust subsp. junciformis
Culms usually 1-noded rarely some 2-noded forming small leafy tufts; panicles rather small
subsp. galpinii
(a) subsp. junciformis.
A. junciformis Trin. & Rupr., Gram. Slip. 143 (1842). A. welwitschii Rendle, Cat.
Pl. Welw. 2, 1: 202 (1899). A. angustata Stapf, in F.C. 7: 556 (1899).
Culms two or more-noded ; panicles usually contracted ; peduncles of the spikelets
short; often tall, strictly erect tufts, with fine leaves and short branched rhizomes.
Cape. — ^George: Baker 19. Somerset East; Acocks 15742. Ceres: Adamson D 25.
Somerset West: van Rensburg 192. Fort Beaufort; Killick 864.
Orange Free State. — -Senekal; Goossens 889. Bethlehem: Potts 4517. Kroonstad:
G 00 s sens 1179.
Natal. — Underberg; McCleanbOl. Maritzburg: iV/cC/can 198. Durban: Schweickerdt
1351. Melmoth; Tinley 14. Hlabisa: Michelmore 15.
Transvaal. — Ermelo; Burtt Davy 9320. Heidelberg: Burtt Davy 9185. Pretoria:
Pole Evans H 17587. Waterberg: de Winter 735. Pietersburg: v.d. Schijff 5396;
Strey & Schlieben 8524. Soutpansberg; Codd 6565.
South West Africa. — Okavango: de Winter 5043.
Distribution. — (See Fig, 43.)
Fig. 44. — A. junciformis subsp. galpinii: cross section of the leaf-blade (Galpin 6900),
269
Fig. 45. — A. junciformis subsp. galpinii: abaxial epidermis of the leaf-blade.
Fig. 46. — A. recta: cross section of the leaf-blade {Schweickerdt 1677).
A variable and very widely distributed subspecies. An aggressive plant in areas
where the rainfall is high and selective overgrazing takes place, such as in the sour
mountain grasslands in parts of Natal. In areas with a lower rainfall it occurs mainly
in depressions where water collects during the rainy season. Its very tough roots,
covered by a spongy white outer layer, makes it very difficult to remove from the soil.
270
The bases of the erect, hard tufts are branched and often produce short rhizomes.
(b) subsp. galpinii (Stapf) de Winter in Kirkia 3: 132 (1963).
A. galpinii Stapf in Kew Bull. 1910: 130 (1910).
Culms one-noded or rarely 2-noded; panicles contracted but laxly branched;
peduncles often longer than the spikelets. Low very densely leafy tufts with a distinctly
rhizomatous base.
Cape. — Port Elizabeth: Liebenberg 5388. King William’s Town: Dyer 257. Keis-
kamahoek: Story 3429; 3789; 3287. Mt. Currie: Acocks 13283. Barkly East:
Acocks 20148. Cathcart: Galpin 2417; Liebenberg 5367. Kokstad: Mogg 5013.
Basutoland. — Pole Evans 14; 19; 66; 2; 98; Guillarmod 695; Staples 112.
Natal. — Bergville: Killick 1246; 1247; 1235; Bayer & McClean 24; Nkand'a:
Fisher and Schweickerdt 122.
This subspecies is conhned to high altitudes in the mountainous areas of the Cape,
Basutoland and Natal. No specimens have been collected below 3,000 feet and in
Basutoland it occurs up to 8,500 feet and probably at even higher altitudes.
The two subsp. of A. junciformis can be distinguished by the habit as given in
the key. Anatomically they are very similar and only distinguishable by the more
robust development of the leaf-blades of subsp. junciformis (cf. Fig. 41, 44 and 42,
45). Anatomically A. junciformis is similar only to A. recta (see p. 273) and is easily
distinguished from all the other species examined as set out in the key on p. 241.
Fig. 47. — A. recta: abaxial epidermis of the leaf-blade.
15. A. recta Francli. in Bull. Soc. d’Autun 8: 365 (1896)
A. atroviolacea Hack, apud Schinz in Bull. Herb. Boiss 2, 6: 707 (1906). A. gossweileri
Pilger in Engl. Bot. .lahrb. 39: 598 (1907). A. hockii De Wild, in Bull. Jard. Bot.
Brux. 6: 39 (1919).
Perennial, caespitose, 10-30 cm high with the persistent basal leaf-sheaths breaking
up into hbres, forming a dense tuft at the base of the culms. Culms erect, l-(2)-noded.
Leaf-sheaths striate glabrous; auricles bearded. Leaf-blades mostly basal, setaceous.
271
folded, up to 20 cm long and 1 mm wide. Panicle ovate-lanceolate in outline, up to
8 cm long and 4 cm wide, lax but not much branched or diffuse. Spikelets contracted
towards the ends of the branches, brownish-purple to deep purple. Glumes subequal
to unequal, 1-nerved, glabrous or scabrid on the keels; lower lanceolate shortly awned,
3-keeled, 2-5-6 mm long; upper linear-lanceolate, shortly awned 5-7-5 mm long.
Lemma 4-5 mm long narrowed into a very short column, scaberulous below the awns;
awns subequal scabrid, usually loosely spirally contorted at the base; callus 0-25
mm long, rounded obtuse, bearded. Caryopsis not seen.
Fig. 48. — Distribution of 9A. recta; MA. spectabilis.
Anatomy (Fig. 46, 47 and 159: 15)
Leaf-blade V-shaped in transverse section, keel not protruding; adaxial surface with the bundles
strongly protruding, furnished with very short almost papillose hairs opposite the stereome strands;
abaxial surface smooth and practically glabrous; margins subacute. Abaxial epidermis: stomatal
zones of 4-9 rows of rectangular, narrowly-oblong ripple-walled cells alternating with transversely
narrowly-oblong short elements and with 1-2 (rarely 3-4) rows of stomata; silicified cell zones with
1-3 files of dumb-bell-shaped silicified cells alternating with short thin-walled rectangular or very
broadly oblong suberized cells with undulate walls; unicellular hairs very rare, short and acute;
bicellular hairs short and thin-walled present in the transitional area between the zones. Vascular
bundle units more or less oblong in outline; first order units 3; midrib flanked by three third order
units; all units more or less equal in height, gradually diminishing in size towards the midrib which
usually is somewhat smaller than the other units. Stereome strands present abaxially opposite the
motor cell groups as well as the bundle units except for the pairs of third order bundles flanking the
midrib; the strands subtending the motor cell groups flanking the bundle on the inside of the marginal
first order units, fused to form a single strand subtending the bundle as well as the two motor cell
groups; adaxially replaced by a single row of large parenchyma cells opposite the bundles, except
in the margin where a small group of fibres is present. Motor cell groups alternating with all bundles.
272
Fio. 49. — A. sciurus: cross section of the leaf-blade (Penlz & Acocks 10314).
0-1 mm
Fig. 50. — A. sciurus: abaxial epidermis of the leaf-blade.
Transvaal. — Bronkhorstspruit : Scinveickerdt 1677; 1656. Ermelo : Norval 24.
Carolina; Codd 511. Pilgrim’s Rest: de Winter and Codd 222. Soutpansberg : Codd
3047.
Swaziland. — Mbabane: Dlamini s.n.
273
Distribution. — (See Fig. 48.)
A distinct species recognized by its small stature, dense fibrous base, and small
brown or purple spikelets with usually laxly spirally contorted awns. The probably
distantly related A. junciformis is a more robust plant which tends to develop short
rhizomes and lacks the fibrous basal leaf-sheaths typical of A', recta. A. recta is found
mainly in seepage areas in sour mountain grassland and, together with A. monticola
and A. junciformis, is exceptional for the genus in preferring moist habitats.
It is widely distributed in tropical Africa from the Rhodesias to Ghana and East
Africa. The specimens from tropical Africa agree very well with the South African
material.
Anatomically similar to A. junciformis but differing from this species in the very
large epidermal cells which equal or even exceed the bundle sheath cells in size.
16. A. sciurus Stapf in FI. Cap. 7: 557 (1899)
Perennial, very robust and erect, up to 140 cm high. Culms arising from a short
rhizome, 4-5-noded, simple; internodes adpressedly woolly just below the nodes,
glabrescent, terete; nodes glabrous inconspicuous. Leaf-sheaths crowded near the
base, glabrous or more or less fugaciously woolly ; ligule a fringe of short hairs hidden
by a tuft of wool at the mouth of the sheath which may disappear with age. Leaf-
blades linear, folded, up to 60 cm long or rarely longer, up to 5 mm wide at the base.
Panicle dense, contracted to almost spike-like, up to 45 cm long. Spikelets pallid or
tinged with purple. Glumes 1-nerved, very unequal, thin; lower lanceolate, acute,
minutely scaberulous upwards and on the keels, 5-7 mm long; upper linear-oblong,
11-13 mm long mucronate. Lemma glabrous, up to 12 mm long; column absent;
awns subequal central up to 23 mm, lateral up to 20 mm long; callus 1-1-5 mm long,
rounded-obtuse, bearded. Caryopsis linear, terete.
Anatomy (Fig. 49, 50 and 159: 16)
Leaf-blade V-shaped in transverse section; abaxial surface smooth, glabrous; adaxial surface
with deep grooves between the strongly projecting first order bundle units, covered with short papiUa-Iike
unicellular hairs opposite the bundles; margins subacute to obtuse. Abaxial epidermis consisting
almost exclusively of long elongate-rectangular cells with deeply and compactly undulate walls,
alternating with transversely elongate short elements (suberized cells); stomata very few or absent.
Silicified zones few and usually reduced to 1-2 files of silicified cells which are solitary, or accompanied
by suberized cells alternating with short ripple- walled cells; silicified cells variable in shape, circular,
transversely-oblong, kidney-shaped or dumb-bell-shaped; no hairs observed in the abaxial epidermis.
Vascular bundle units: first order units obovate in outline; the ovate third order units being only one
third to half of the height and much narrower than, and completely overtopped by the first order units;
midrib smaller than the other first order bundles, not projecting abaxially. Stereome strands abaxially
fused to form a practically continuous hypodermal band ± 1-5 cell-layers thick, multi-layered opposite
the bundles and forming rather narrow projections which enter the bundles, consisting of typical fibres
(very strongly thickened walls) ; adaxial stereome consisting of large cells with large lumen and relatively
thin though distinctly lignified walls, large anchor-shaped strands opposite the first order bundles
and few-celled groups opposite third order bundles. Motor cells in many-celled triangular groups
alternating with all the bundles.
Transvaal. — Potgietersrust : de Winter 2321. Carolina; Mundy H. 4292 A.; Burtt-
Davy 1418; Pole Evans 1967. Belfast: Schweickerdt 1914; Pole Evans 3688. Lyden-
burg: Burtt-Davy 1449; Story 3897.
Swaziland. — Mbabane; Compton 30594; 25712; Codd and Muller 330.
Natal. — Estcourt; Edwards 2049; 2051. Ngutu; West 755.
Distribution. — (See Fig. 20.)
A robust perennial vegetatively strongly resembling species belonging to the section
Arthratherum in stature and in the woolly indumentum of the internodes and mouth
of the sheaths. The spikelets are, however, fairly typical of the section Aristida. The
anatomy resembles that of A. diffusa and A. dasydesmis. The leaf-blade is V-shaped
274
in transverse section and the first order bundle units protrude very strongly, charac-
teristics not found in any other species of the section Aristida. This species may
represent a link between the sections Aristida and Arthratherum and has no obvious
close relatives.
SC
CC
ST
LC
I 1
O’ 1 m m
Fig. 52. — A. dasydesmis: abaxial epidermis of the leaf-blade.
275
A species found mainly in mountain sourveld in the eastern Transvaal, Swaziland
and Natal. A specimen was also collected in the mountains near Potgietersrust in
the Transvaal bushveld.
17. A. dasydesmis Mez in Fedde Rep. 17: 148 (1921)
A. vestita var. parviflora Trin. & Rupr., Gram. Stip. 158 (1842)
Perennial, much branched, caespitose up to 70 cm high. Culms many-noded,
fascicled from the lower nodes ; nodes and internodes glabrous. Leaf-sheaths glabrous,
lax; auricles pubescent. Leaf-blades stiffly erect, semiterete, smooth sub-pungent,
about 1 mm wide. Pan/c/e somewhat contracted in young specimens but lax and effusely
branched up to 10 cm long and 6 cm wide. unequal, glabrous 1 -nerved, lower
about 5 mm long; upper up to 8 mm long. Lemma glabrous, including the callus
up to 8 mm long; column of awns twisted, about 5 mm long; awns slightly unequal;
central up to 25 mm, the lateral up to 20 mm long. Caryopsis not seen.
Anatomy (Fig. 51, 52 and 159: 17)
Leaf-balde horseshoe-shaped in transverse seetion; abaxial surface smooth and practically glabrous ;
adaxial surface deeply grooved due to the strongly projecting bundle units, with numerous hairs opposite
the bundles. Abaxial epidermis: stomatal zones with one row of stomata, ripple-walled cells elongate-
rectangular, with thickened, strongly and compactly undulate walls, usually with a few solitary short
elements scattered amongst them; silicified cell zones with the silicified cells subcircular and usually
accompanied by a very short, more or less square to elongate-rectangular, thin-walled suberized cell
with undulate walls, ripple-walled cells narrower, and with thinner more finely undulate walls. Vascular
bundle units: first order units 5, midrib slightly smaller than the adjacent first order bundle units, more
or less ovate in shape, the other units oblong with rounded adaxial surfaces; third order units ovate,
half or less than half the length of those of first order. Stereonie strands very well developed abaxially
opposite all the bundles, as well as the motor cell groups; adaxially very strongly developed opposite
the first order bundles and less so opposite the second order bundles, consisting of small very thick-
walled fibres, or of rather large cells with large lumen, or of a mixture of both types. Motor cells in
oblong or obtusely triangular groups, occupying the whole thickness of the leaf.
Cape. — Namaqualand ; Schlechter 11228 (lectotype !) ; Pearson 2952 and 3487; Drege
s.n. (Zilverfontein) ; Acocks 19518. Vanrhynsdorp : Stokoe 8399.
Distribution. — ^(See Fig. 29.)
A species closely related to A. vestita and A. diffusa but with smaller spikelets
and branched fascicled cuhns, an obtuse truncate or emarginate not distinctly bifid
callus, and wiry, erect and sub-terete leaf-blades. It may have arisen as an adaptation
of a form of A. vestita to extremely at id conditions. Its distribution is restricted to
the desert areas of the north western Cape. Anatomically similar to A. diffusa var.
diffusa but the leaf-blade sub-terete. Also similar to A. sciurus in leaf anatomy (cf.
Fig. 159: 16 and 17) the latter species, however, always lacks an articulation in the
lemma and has a subacute or obtuse callus, while the column is absent. Whereas
in some specimens of A. dasydesmis the articulation of the lemma is very inconspicuous,
it is clearly visible in others. Theron (1936; p. 23) investigated the leaf anatomy of
one specimen of this species, namely Schlechter 11228. His description agrees fairly
well with the one given above but no drawing is available for comparison.
18. A. diffusa Trin. in Mem. Acad. Imp. Sc. Petrop. 6, 1 : 86 (1830)
Perennial, densely caespitose, up to 75 cm high. Culms simple or rarely branched,
erect, often wiry, smooth, one to several-noded, internodes terete, glabrous; nodes
glabrous. Leaf-sheaths glabrous or lower ones covered with a fugacious wool; auricles
pubescent or glabrous. Leaf-blades convolute, rather robust to setaceous up to 30 cm,
rarely longer. Panicle up to 30 cm, rarely longer, very effuse and open, or somewhat
dense but much branched. Spikelets yellow-brown to pallid, nodding. Glumes unequal,
glabrous; lower up to 4-10 mm long, obtuse; upper 11-18 mm long; somewhat
involute and curved. Lemma including the callus 10-14 mm long, almost smooth to
276
scaberulous upwards; column from almost absent to 6 mm long; awns subequal to
unequal, central up to 35 mm long but usually shorter; callus bifid, densely bearded.
Caryopsis linear, terete.
Fig. 53. — A. diffusa var. diffusa: cross section of the leaf-blade {Schlechter 9138).
Fig. 54— a. diffusa var. diffusa: abaxial epidermis of the leaf.
277
Anatomy (Fig. 53, 54 and 159: 18)
Leaf-blade somewhat crescent-shaped in transverse section; abaxial surface usually quite flat or
only slightly undulate, midrib not protruding, and leaf without a keel; adaxial surface deeply grooved
between the very strongly protruding first order bundle units, usually densely covered with hairs opposite
the bundles. Abaxial epidermis: stomatal zones with one row of stomata and a few rows of long
ripple-walled cells, elongate-rectangular in shape and with finely and compactly undulate walls (in
var. diffusa stomata sometimes absent and abaxial epidermis consisting of a series oT silicified zones);
silicified zones: silicified cells small, dumb-bell-shaped, shorter than or exceeding the stomata in length
(in var. diffusa some cells circular, each accompanied by a kidney-shaped or rectangular suberized cell),
long ripple-waUed cells with compactly and finely undulate walls. Stereome strands strongly developed
adaxially opposite the first order bundle units; abaxially also present opposite, and in contact with the
second order bundles as well as the motor cell groups, or forming a continuous band below the epidermis;
adaxially well developed opposite the first order bundles and in the var. diffusa also in small groups
laterally situated on the flanks of these units, present also opposite the second order bundles. Motor
cells in triangular to broadly oblong groups occupying most of the leaf thickness.
Culms 1-2-noded; lemma usually about 12-14 mm long; southern Cape var. diffusa
Culms several-noded, very rarely 2-noded; lemma usually 12 mm or less; widespread in the
summer rainfall areas var. burkei
(fl) var. diffusa.
A. hystrix sensu Thunb., Prod. 19 (1794) non L.f. (1781). A. diffusa Trin. in Mem.
Acad. Imp. Petrop. 6, 1 : 86 (1830). A. vestita Thunb. var. brevistipitata Trin. & Rupr.,
Gram. Slip. 158 (1842). A. vestita Thunb. var. densa Trin. & Rupr., Gram. Slip.
158 (1842). A. vestita Thunb. var. pseudo-hystrix Trin. & Rupr., Gram. Slip. 158
(1842). A. vestita Thunb. var. schraderiana Trin. & Rupr., Gram. Stip. 158 (1842).
A. vestita Thunb. var. diffusa (Trin.) Walp., Ann. Bot. 3: 747 (1852). A. pseudo-hystrix
(Trin. & Rupr.) Steud., Syn. PI. Glum, 1: 142 (1854). A. diffusa Trin. vat. brevis-
tipitata (Trin. & Rupr.) Henrard, Grit. Rev. Aristida 3: 665 (1928). A. diffusa Trin.
var. densa (Trin. & Rupr.) Henrard, Grit. Rev. Aristida 3: 665 (1928). A. diffusa
Trin. var. eckloniana (Tiin. & Rupr.) Henrard, Grit. Rev. Aristida 3; 666 (1928).
A. diffusa Trin. var. schraderiana (Trin. & Rupr.) Henrard, Grit. Rev. Aristida 3: 668
(1928). A. diffusa Trin. var. genuina Henrard, Monogr. Aristida 1: 97 (1929).
Densely caespitose plants; culms usually one- rarely two-noded. Panicle effuse
or somewhat contracted, up to 15 cm long. Spikelets purple, purplish-brown or brown.
Glumes unequal; lower obtuse, 4-10 mm long; upper up to 18 mm long but usually
shorter. Lemma including the callus up to 14 mm long; column of the awns 2-5 mm
long.
Gape. — Paarl: Schlechter 9138. Galedon; de Villiers 16. Uitenhage: Zeyher 4004;
Ecklon & Zeyher 447. Malmesbury: Liebenberg 4240. Stellenbosch: van Rensburg
193; Sandwith 154. Tulbagh: Zeyher s.n., NH. No. 8685.
Distribution.— (See Fig. 55.)
This variety is limited to the southern Gape, but partly overlaps with var. burkei
in distribution. Intermediates between these two varieties are not uncommon. Ana-
tomically it differs very little from var. burkei. Fig. 159: 18 and 18a depict a very
robust and a less robust leaf, in cross-section, showing the variation possible within
this variety.
{b) vai. burkei (Stapf) Schweikerdt in Notizbl. Bot. Gart. & Mus. Berlin 14: 122
and 195 (1938).
A. burkei Stapf in FI. Gap. 7: 577 (1899).
Densely to laxly caespitose plants. Culms 3 to several-noded, rarely 2-noded.
Panicle lax and effuse, up to 30 cm long but often much smaller. Spikelets yellow-
brown to palhd. Glumes unequal; lower obtuse, up to 6 mm long; upper usually
about 12 mm long. Lemma about 12 mm long including the callus; column of the
awns 2-5 mm long.
278
Cape.— Queenstown : Galpin 2383; Everett 50. Middelburg ; Comins 699; Theron
720. Somerset East; Theron 1X6. Komga; F/anagan 1780. Aliwal North: Hanogan
2305. Tarka: Story 4501. Colesberg: Liebenberg 9.
Basutoland.— Leribe : Dieterlen 199a. Mamathes; Guillarmod 858. Malutis:
Staples 29 and 131.
Orange Free State.— Heilbron: GoossensSAl. Senekal; Goossens 95^. Bethlehem:
Potts 4519. Fauresmith: Smith 4168; Pole Evans 1573.
Natal.— Utrecht; Codd 2533. Estcourt; West Xm. Pietermaritzburg; Comins 205.
Weenen: Acocks 10115. Umvoti: Edwards 1255.
Transvaal.— Heidelberg: Burtt-Davy 9161. Ermelo: Henrici 1524. Pretoria: de
Winter 195 ■, Mogg 15, 979; Pole Evans 369. Waterberg: Codd 910. Potgietersrust ;
Codd 6578.’ Pietersburg: Gerstner 5354. Lydenburg: Strey 3728.
Distribution. — (See Fig. 55.)
Fig. 55. — Distribution of A. diffusa; ®var. diffusa; Ovar. burkei; mA. hordeacea.
The var. burkei grades into the var. diffusa in certain areas of the Cape but on
the whole is fairly distinct, the culms usually being more than 2-noded and the spikelets
smaller and more graceful. The extremes of both varieties are very easily distinguished
and for this reason var. burkei is retained. Geographically var. dijfusa has a rather
limited distribution while var. burkei is very widespread, covering most of South Atnca^
Anatomically it is very similar to var. diffusa but the leaves are usually less robust.
The stereome consists of separate strands which abaxially never form a continuous
layer, as is the case in some specimens of var. diffusa.
279
19. A. vestita Thunb., Prodr. 19 (1794)
A. lanuginosa Burchell, Travels 2: 226 (1824). A. vestita forma amplior Hack, in
Engl. Bot. Jahrb. 11: 400 (1889). A. floccicuhnis Mez, in Fedde Rep. 17: 147 (1921).
Chaetaria vestita (Thunb.) Beauv. Agiost. 30 (1812). Arthratherum vestitum (Thunb.)
Nees, FI. Afr. Austr. 174 (1841).
Perennial densely caespitose, up to 85 cm high. Culms 3-5-noded rigid, simple
or rarely branched; internodes, or at least the lower, densely to sparsely pubescent;
nodes glabrous. Lower leaf-sheaths woolly or glabrescent; auricles with a flake of
wool. Leaf-blades linear, up to 24 cm long and 4 mm wide, expanded or rolled. Panicles
pyramidal, effuse and open, up to 20 cm long and 12 cm wide, laxly contracted when
young. Spikelets light brown or pallid. Glumes fii'm, 1-nerved, unequal; lower
lanceolate 4-5-7 mm long, rounded at the apex; upper narrowly lanceolate 9-13 mm
long. Lemma glabrous, 7-11 mm long including the callus, often mottled with purple;
column variable 2-7 mm long, twisted, awns unequal, the central up to 35 mm long, the
lateral shorter; callus bifid about 1-5 mm long, bearded. Caryopsis linear, terete.
Anatomy (Fig. 57 and 159: 19)
Leaf-blade in transverse section expanded; adaxial surface with the bundles protruding about
one third of their height; numerous unicellular bristle-like hairs present opposite the stereome strands;
abaxial surface flat, smooth and almost glabrous. Abaxial epidermis cells with very strongly thickened
outer walls; anticlinal walls much thinner; stomatal zones usually with one row of stomata and one
to four rows of long ripple-walled cells with strongly undulating walls, the undulations approximate,
and viewed under low magnification the walls often apparently thick and smooth; a few paired short
elements present; silicified cell zones consisting of 1 to 9 files of silicified cells alternating with rows
of narrow finely ripple-walled cells; silicified cells dumb-bell-shaped and usually accompanied by
oblong to rectangular suberized cells with slightly undulate walls; very few unicellular sharp-pointed
and linear bicellular hairs present. Vascular bundle units: first order units 5-7, square in outline, only
slightly exceeding the third order units in height; third order units ovate in outline, alternating with the
first order bundles in pairs or as single units; midrib similar to other first order bundles (occasionally
somewhat smaller). Stereome strands usually well developed abaxial ly, and present opposite all the
bundles as well as at least some of the motor cell groups. Motor cells in oblong to triangular groups
occupying about two thirds of the leaf width, alternating with all except occasionally the marginal
bundles.
Fig. 56. — A. diffusa var. burkei: cross section of the leaf-blade (de Winter 795).
280
Fig. 57. — A. vestita: abaxial epidermis of the leaf-blade {de Winter 3584).
Cape. — Sutherland: Acocks 16911. Hanover: Sim (Galpin Herb. No. 6260). Ken-
hardt: Pole Evans 2182; Liebenberg 5363. Hay: Esterhuyzen 2293; Brueckner 1042.
Gordonia: cie Winter 3584. Kuruman: Pole Evans 2097. Kimberley: Acocks 2346.
Barkly West: Brueckner 834.
Orange Free State. — ^Boshof: Wolff 17.
Transvaal. — Christiana: Theron 599; Burtt-Davy 13030. Lichtenburg: Kinges 1412.
Distribution. — (See Fig. 36.)
Occurs on limestone, dolorite and ironstone formations as weU as on red loamy
and sandy soils.
Fig. 58. — A. engleri: cross section of the leaf-blade {Pole Evans 80).
281
Closely allied to A. dijfusa which it resembles very strongly in spikelet characters,
the main differences being the sMghtly smaller spikelets, the moie scabrid lemmas,
and the woolly lower internodes of the culms. Anatomically it can be distinguished
by the more or less square, less protruding first order bundle units which do not much
exceed those of the third order. In A. diffusa the first order units protrude strongly
and much exceed those of the third order in height. Difficulties experienced in dis-
tinguishing this species from A. diffusa var. burkei in the veld may indicate that it
could be regarded as a variety of A. diffusa but the anatomical differences, together
with the difference in spikelet size, seem to indicate a higher category, particularly
since they overlap in distribution.
Fig. 59. — A. engleri: abaxial epidermis of the leaf-blade.
20. A. engleri Mez in Fedde Rep. 17: 147 (1921)
Perennial, caespitose and erect to lax and sprawling up to 70 cm high. Culms
many-noded, simple or much branched from the nodes, nodes glabrous; internodes
glabrous, thin and wiry to fairly robust. Leaf-sheaths glabrous; auricles pubescent.
Leaf-blades expanded or involute, narrow, up to 15 cm long. Panicle effuse and open,
or rather narrow and much longer than broad, the branches short and somewhat
contracted. Spikelets brownish or pallid. Glumes unequal, glabrous or scaberulous,
up to 6 mm long, sub-obtuse; upper 9-10 mm long. Lemma 5-8 mm long, scabrid
upwards; column about 4-5 mm long; awns slightly unequal; central about 20 mm
long, lateral about 15 mm long; callus bifid, bearded. Caryopsis linear, terete.
Anatomy of the leaf-blade is mainly as described for A. vestita (see also Fig. 58, 59 and 159: 20).
Culms not branched or rarely branched from the upper nodes; panicle much longer than wide
and rather contracted var. engleri
Culms much branched ; panicle not much longer than wide, rather open var. ramosissima
(a) var. engleri.
A. engleri Mez in Fedde Rep. 17: 147 (1921).
Erect tufted perennials, usually with simple cuhns or rarely branched from the
upper nodes ; panicles much longer than wide or if rather wide the branches contracted
and dense.
Cape. — Kenhardt: Acocks 16366; 14278, and 16374; Leistner 2422; Wasserfall 1051.
Gordonia: de Winter 3585.
282
South West Africa. — Warmbad: Theron 1972. Keetmanshoop : de Winter 3416;
3274 ; 3307 and 3369. Luderitz : Kinges 2268 ; Giess and van Vuwen 749 ; Giess
2382. Bethanien : Kinges 2194. Maltahohe: Louw 283. Rehobotb : Volk 2522
and 2541.
Distribution. — (See Fig. 33.)
Related to A. diffusa and A. vestita but differing from A. diffusa by the narrow
somewhat contracted panicles and somewhat smaller spikelets, and from A. vestita
by the glabrous lower internodes of the culms. Distinguished from A. dasydesmis
by the rolled or expanded, not stiff subterete leaves. Found mainly on rocky outcrops
near the Orange River and further northwards in South West Africa. Anatomically
very similar to A. vestita.
(b) var. ramosissima de Winter in Kirkia 3: 132 (1963). Type: Cape, Gordonia,
il5 m. N.N.W. of Winstead, Acocks 2053 (PRE, holo.!).
A. engleri Mez sensu Schweickerdt Bothalia 4, 1: 165 (1941) pro minore parte.
A variety characterized by the much branched slender often fascicled cuhns, and
small usually elongate but rather lax and open panicles. Indistinguishable from var.
engleri on floral characteristics.
Cape. — Hay: Esterhuyzen 2359; Acocks 2053. Gordonia: Leistner 2042. Kuruman:
Cooke s.n. (McGreg. Mus. No. 6355). Postmasburg: Leistner 1599; 1719.
South West Africa. — Keetmanshoop: Liebenberg 5228. Luderitz: Kinges 2365.
Distribution. — (See Fig. 33.)
This variety is not always clearly distinguishable from var. engleri. Both occur
in the Karasberg in South West Africa. The typical forms are, however, very easily
separated: the stereome is usually less developed and hence the leaf-blades tend to
be softer than those of the var. engleri.
Fig. 60. — A. spectabilis: cross section of the leaf-blade (de Winter 111).
Due to the mixed gathering on sheet 2 of Pearson 3487 (PRE), collected at “Alewyns
Fontein ” (Aalwynsfontein) in Namaqualand this collection is quoted under both
A. engleri and A. dasydesmis by Schweickerdt. The loose inflorescence on sheet 2
283
clearly belongs to A. engkri whereas the other two fragments agree with sheet 1 of
Pearson 3487. This explains Schweickerdt’s remark under A. engkri where he states
that the facies of the two species are identical. A. dasydesmis, with branched and
fascicled culms and stiffly erect subterete very firm leaf-blades, is easily distinguished
from A. engkri, which is a densely tufted plant often with branched but not fascicled
culms, and with expanded or rolled, but never subterete leaves. A. engkri also occurs
further northwards and more inland, and does not overlap in distribution with A.
dasydesmis (as far as known at present).
A. engkri is very closely allied to A. diffusa var. burkei and A. vestita and inter-
mediates are not uncommon. These intermediates may represent hybrids.
0-1mm
Fig. 61. — A. spectabilis: abaxial epidermis of the leaf-blade.
21. A. spectabilis Hack, in Bull. Herb. Boiss. 3, 8; 380 (1895)
Perennial, densely caespitose, robust up to 1-75 m high. Culms stout, erect,
simple terete. Leaf-sheaths well developed and exceeding the internodes, glabrous, often
with a flake of wool at the mouth of the sheath, and somewhat woolly on the margins
downwards. Leaf-blades linear, involute or expanded, up to 60 cm long and 4—5 mm
wide at the base. Panicle lax and effuse, nodding, 30-40 cm long and up to 25 cm
wide. Spikekts glabious, pallid. Glumes linear lanceolate, acute, firm below and
with membranous hyaline tips, 1 -nerved, subequal, about 10-12 mm long. Lemma
linear, 6- 5-7-5 mm long including the callus, smooth or minutely scaberulous upwards;
column twisted, 5-6 mm long; awns unequal, central up to 44 mm long, lateral up
to 28 mm long; callus up to 1 mm long, bearded, bifid. Caryopsis not seen.
Anatomy (Fig. 60, 61 and 159: 21)
Leaf-blade in transverse section expanded; adaxial surface with the bundle units protruding,
and with grooves extending about one third of the leaf thickness between the bundles; bristle-like,
unicellular sharp hairs present adjacent to the stereome strands; abaxial surface flat, unicellular hairs
practically absent; margins obtuse or obliquely truncate. Abaxial epidermis cells with very strongly
thickened walls; stomatal zones usually with a single row of stomata and one to three rows of long
ripple-walled cells, undulations very deep and approximate, clearly visiable only when strongly
magnified; short elements more or less transversely oblong in outline; silicified cell zones of one to
three files of short elements alternating with files of ripple-walled elements opposite the third order
bundles and motor cells, and six to seven files of short elements alternating with files of long ripple-
walled cells opposite the first order bundles; silicified cells dumb-bell-shaped, as long as, or shorter
than the stomata and usually accompanied by a transversely oblong smooth-walled suberized cell or
an oblong suberized cell with undulate thin walls; both the unicellular and bicellular type of hair
very rare or absent. Vascular bundle units: first order units 5-9, square in outline, the second and
284
third order units projecting only slightly less than those of the first order and ovate to ovate-oblong
in outline, usually in pairs between the first order units, but single adjacent to the marginal first order
unit; midrib like the other first order bundles. Stereome strands abaxially well developed opposite
all bundles as well as the motor cell groups, slightly projecting into the first order bundles and in contact
with the other bundles; adaxially in contact with the first order bundles and free from the other bundles.
Motor cells in more or less equally wide, elongate groups of thin-walled cells, occupying almost the
whole width of the leaf-blade.
Cape. — Mafeking: Brueckner 381.
Transvaal. — Potchefstroom : Louw 1390. Pretoria: Schweickerdt 1339; de Winter
773. Brits; Mogg 14985. Rustenburg: Codd 1046.
Distribution. — (See Fig. 48).
A species which occasionally superficially resembles A. diffusa var. burkei but
much more robust and with subequal glumes which are hyaline at the tips. Related
to A. meridionalis but distinguished from this species by the denser pallid inflorescence,
and subequal glumes with hyaline tips.
A. spectabilis exhibits a marked preference for acid soils derived from quartzite
and occurs mainly on the summits and stony slopes of mountains, usually only on
northern aspects. Apparently a rather rare species.
22. A. meridionalis Henrard, Crit. Rev. Aristida 2: 344 (1927)
A. stipoides Lam. var. meridionalis Stapf in FI. Cap. 7: 562 (1899)
Perennial compactly caespitose, very robust, up to 2 m high. Culms erect, simple
or somewhat branched, 2-3-noded; internodes glabrous; nodes often dark coloured,
glabrous. Leaf-sheaths densely woolly or glabrescent those of the culms glabrous;
auricles with a dense flake of wool. Leaf-blades narrow, involute, up to 50 cm long
and 5 mm wide. Panicle large, lax and effuse, often nodding, up to 80 cm long and
over 20 cm wide. Glumes unequal, glabrous; lower 5-7 mm; upper 10-15 mm.
Lemma 7-9 mm long, mottled with purple, smooth or shghtly punctulate and scaberulous
upwards; column slender, twisted, up to 2 cm long; awns subequal, up to 5 cm long;
callus 1 • 5-2 mm long, bearded, distinctly bifid. Caryopsis linear terete.
Anatomy (Fig. 62 and 159: 22)
Anatomy of the leaf-blade in general details similar to that of A. spectabilis but with a distinct
tendency towards the development of more second order bundles which alternate with those of the
first order.
Cape. — Mafeking: Pole Evans 2390. Kimberley: Esterhuyzen 1048.
Orange Free State. — Boshof; Wolff 25.
Transvaal. — Waterberg: Codd 4879. Potgietersrust : Galpin M617. Zoutpansberg :
de Winter & Codd 336.
South West Africa. — Otavi: Dinter 5767. Okavango: Maguire 2492.
Distribution. — (See Fig. 63.)
A species occurring mainly in the warm, dry, sandy areas of South Afiica. The
large divaricate open inflorescence together with the flake of wool at the mouth of the
leaf-sheath makes this species easily recognizable. A. meridionalis is related to A.
stipoides from which it can be distinguished by the perennial habit and, on the whole,
shorter column of the awns. A. spectabilis which may be confused with A. meridionalis
has usually more dense inflorescences, pallid spikelets and membranous-tipped glumes.
A number of specimens in the National Herbarium collected in Southern Rhodesia
viz. Goldsmith 138/36, Chase s.n. and Fisher 1402 agree in most respects with the above
species, except for the densely lanate internodes of the culms, and could be described
as a new variety of A. meridionalis (see also notes under A. mollissima).
285
^<vwwTnrr'w^~r-"~"o--^
— ^.JuifWStStS>K__y — \_yt/UUULk,^/ ^ 'CU
WsisinszrumJijKnjinJtfiswuinMnswisuuuwwmrisinnfinsinMSUt^^
— — Vw.X~..x-sA,._/ — \,,._>virLrirLfiA>^ \..._>On-''nA_-/ v.A-'-'A.— ■'v.-X^uz/iiltiX^
LC
I
0-1 mm
1
Fig. 62. — A. meridionalis: abaxial epidermis of the leaf-blade [Pole Evans 3307 (3)].
The tendency to develop second order bundles which alternate with first order
bundles, is the only feature by which this species can be distinguished from A. spectabilis
on anatomical grounds. This tendency is also noticeable in the related though annual,
A. stipoides. In the other species of the section Arthmtherum the bundles (or groups
of bundles) alternating with those of the first order are invariably of the third order.
286
23. A. stipoides Lam., Encycl. Method. 157 (1791).
A. gracillima Oliv. in Trans. Linn. Soc. 29 ; 173 (1875). A. lamarkii (R. & S.) Steud.,
Nom. Bot. 1; 69 (1821). Chaetaria lamarkii R. & S., Syst. Veg. Ed. 2: 393 (1817).
A. fontismagni Schweick. in Bot. Jahrb. 76, 2; 220 (1954).
Annual, laxly caespitose, up to 1 - 5 m high; internodes and nodes glabrous. Leaf-
sheaths glabrous; auricles with a dense flake of wool. Leaf-blades narrow, expanded
or involute, up to 30 cm long and 3-5 mm wide. Panicle effuse and very lax, often
nodding, up to 50 cm long, the spikelets spreading. Glumes usually purplish, very
unequal, glabrous, 1-nerved; lower lanceolate, up to 7 mm long; lower linear-lanceolate,
15-20 mm long, mucronate from a bifid apex. Lemma including the callus up to
9 mm long, smooth below, punctulate upwards; column 2-3 cm long; awns subequal,
scabrous, up to 5-5 cm long; callus slender, about 2 mm long, bearded, distinctly
bifid. Caryopsis linear, terete.
Fig. 64. — A. stipoides: cross section of the leaf-blade (Schweickerdt 2129).
I 1
0 1 mm
Fig. 65. — A. stipoides: abaxial epidermis of the leaf-blade.
287
Anatomy (Fig. 64, 65 and 159: 23)
Anatomically very similar to A. meridionalis and distinguishable only by the slightly weaker
stereome strands; only distinguishable from A. spectabilis, and some other species of the section, by
the development of second order bundles alternating with the first order bundles especially next to the
midrib.
South West Africa. — Ovamboland: de Winter and Giess 10A\. Okavango: Maguire
2307; de Winter and Marais A65A; /e 7?om.y 1007; Fo/Zc 1790 and 1203. Grootfontein ;
Schweickerdt 2129 and 2131. Gobabis: Liebenberg 4630.
Distribution. — (See Fig. 21.)
Though described by Schweickerdt as a robust perennial in his revision of Aristida
(1941; p. 161) specimens from Senegal studied by the author, as well as the two
Barnard specimens cited by Schweickerdt, are all very robust annuals lacking, or rarely
with a few, innovation shoots. Henrard also makes use of the annual habit of A.
stipoides in order to distinguish between this species and A. meridionalis. Barnard 42
and 815 moreover agree with the type of A. fontismagni which is here placed in synonymy
under A. stipoides. A. stipoides and A. meridionalis are undoubtedly closely related
and are both organographically, and anatomically very similar, the only anatomical
difference being the weaker development of the stereome in the annual A. stipoides.
This species, like A. meridionalis, is found mainly in sandy, dry and warm areas
but is less widely spread in southern Africa than the latter. It occurs in suitable
habitats from South West Africa northwards to Senegal and apparently is not found
in the eastern parts of tropical Africa where, however, A. meridionalis is fairly common.
For the discussion of the anatomy the reader is referred to the previous species.
24. A. moUissima Pilger in Engl. Bot. Jahrb. 40: 80 (1908)
A. elymoides Mez in Fedde Rep. 17: 148 (1921).
0 1mm
Fig. 66. — A. moUissima: abaxial epidermis of the leaf-blade (Murray 2).: CC, suberized cell; OH
one-celled hair; LC, long ripple- walled cell; SC, silified cell; ST stoma.
Perennial, densely caespitose, up to 85 cm high. Culms erect, many-noded simple
or rarely branched upwards; nodes glabrous; internodes densely lanate-tomentose.
Leaf-sheaths lanate tomentose to glabrescent; auricles ciliate or furnished with a flake
of wool. Leaf-blades linear to filiform, up to 30 cm long and 3 mm wide. Panicle
narrow, very dense and spike-like, up to 20 cm long and 5 cm wide, axis lanate.
288
Spikelets almost sessile, glabrous. Glumes very unequal, shortly awned, narrowly
lanceolate, 1-nerved, scabrid on the keel, and finely scaberulous upwards; lower 10-14
mm long; upper 17 -5-24 mm long. Lemma, including the callus and up to the
articulation, 8-10 mm long, minutely granular; column 16-27 mm long, scabrid,
twisted; awns subequal 2 -8-5 -5 cm long; callus 2-2-5 mm long, acuminate, bearded.
Caryopsis linear, terete.
Anatomy (Fig. 66 and 159: 24)
Practically indistuinguishable from A. spectabilis in the anatomy of the leaf-blade. The slight
difference which can be detected in individual leaves are not constant.
Cape. — Kimberley: Wilman s.n. (McGr. Mus. 4089). Mafeking: Pole Evans 2435-
Vryburg: Murray 2. Kuruman: Codd 1285; Esterhuyzen 2236; Leistner 1065.
Orange Free State.— Kroonstad: Goossens 1175. Hertzogville : Serfontein 9.
Transvaal. — Christiana; Strey 407; Schweickerdt 1465. Lichtenburg: Kinges 1410.
South West Africa. — Otjiwarongo: Volk 2337. Okavango: Volk 1741-, Story 6474.
Distribution. — (See Fig. 67.)
• Fig. 67. — Distribution of ^A. inollissima; mA. argentea.
A. mollissima is virtually indistinguishable from A. stipitata var. stipitata except
for the densely woolly internodes of the culms. It could be regarded as a “ form ”
of the latter and has a distribution which coincides almost completely with that of the
var. stipitata. The same may apply to A. argentea which has woolly culms, and A.
stipitata var. graciliflora, with glabrous culms. A number of specimens indistinguishable
from A. meridionalis, except for the densely lanate internodes, have been collected
289
in Rhodesia but have not yet been described. The forms differing from each other
solely on basis of indumentum as far as the internodes are concerned should not be
regarded as independent species, but preferably as taxa worthy only of infraspecific
rank. Before deciding on such changes the problem should, however, be investigated
thoroughly on a much wider scale than was possible here. ,
25. A. argentea Schweick. in Engl. Bot. Jahrb. 76, 2: 218 (1954). Type: Trans-
vaal, 50 m N.W. of Potgietersrust to rd. to Maalstroom, Schweickerdt 1806. (holo!
UPR; iso!, PRE).
Fig. 68. — A. argentea: cross section of the leaf-blade {Schweickerdt 1806).
LC
SC
CC
O'l mm
Fig. 69. — A. argentea: abaxial epidermis of the leaf-blade.
290
Perennial, caespitose, up to 90 cm high. Culms erect or somewhat geniculate,
branched from the nodes or simple, 3-many-noded; nodes glabrous; internodes
densely to sparsely lanate-tomentose. Leaf-sheaths lanate to glabrescent, rather lax;
auricles ciliate to somewhat woolly. Leaf-blades linear to filiform, up to 30 cm long
and 4 mm wide. Panicle erect, lax, narrow, up to 30 cm long, occasionally contracted.
Spikelets glabrous, pallid or purplish. Glumes very unequal, glabrous; lower 1-2-
nerved, 10-13 mm long, keel scabrid; upper 1-nerved, 18-20 mm long, awned, keel
scabrid. Lemma minutely granular, 9-11 mm long; column twisted, 13-16 mm long;
awns more or less unequal, up to 6 cm long but usually shorter; callus about 2-5
mm long, very acute, bearded. Caryopsis linear, terete.
Anatomy (Fig. 68, 69 and 159: 25)
Only differing from A. spectabilis in the first order vascular bundle units which protrude slightly
more than in the latter species.
Transvaal. — Warmbad: Schweickerdt 2035; de Winter 693. Soutpansberg: de
Winter and Codd 664. Nelspruit; van der Schijff 4189.
Distribution. — (See Fig. 67.)
Related to A. stipitata var. graciliflora but with the internodes of the culms densely
woolly and a more openly branched inflorescence. (See also note under A. mollissima.)
Usually found in sandy soil in open bushveld.
26. A. stipitata Hack, apud Schinz in Verb. Bot. Ver. Prov. Brandenburg 30 .
143 (1888)
Perennial, laxly to densely caespitose, 30-150 cm high. Culms erect, usually
branched from the upper nodes, rarely some culms simple; internodes and nodes
glabrous. usually lax, glabrous ; auricles pubescent, long bearded. Leaf-
blades variable in length, up to 30 cm long but often much shorter, involute or expanded
and up to 4 mm wide, borne mainly on the culms. Panicles usually erect and stiff
or rarely somewhat drooping, several borne at the ends of fascicled branches or solitary
at the ends of robust culms, spike-like, many-flowered and dense, or lax open and
few-flowered, often interrupted when spike-like. Glumes unequal, chartaceous, narrowly
lanceolate, 1-nerved, glabrous and smooth with scabrid or scaberulous keels; lower
up to 13 mm long, tapering to a fine point, midnerve often excurrent into a short awn;
upper up to 22 mm long tapering to a very fine acuminate apex, often shortly awned
from between two hyaline lacerate lobes. Lemma including the callus up to 11 mm
long, mottled or pallid, finely punctulate, distinctly articulated; column very variable
12-40 mm long, usually about 16-25 mm long; awns very variable in length, subequal,
25-60 mm long; callus acuminate, densely bearded, 2-3 mm long. Caryopsis linear,
terete.
Anatomy (Fig. 159: 26 and 26a)
In leaf anatomy practically identical with A. spectabilis.
Tall strictly erect robust plants 70-150 cm high; culms usually simple or rarely branched from
some of the upper nodes:
Panicles dense and spike-like, usually pallid, the branches appressed; column of the awns
20-40 mm long var. stipitata
Panicles long and narrow, with the branches spreading when mature; spikelets usually purplish;
column usually not exceeding 20 mm in length var. robusta
Smaller, erect or somewhat geniculate plants, usually not exceeding 60 cm in height, culms branched
from the upper nodes (rarely unbranched); column 15-25 mm long:
Panicles dense and spike-like var. spicata
Panicles narrow but loosely branched and rather lax and interrupted var. graciliflora
291
(fl) var. stipitata.
Very tall, strictly erect, robust perennials with dense spike-like inflorescences and
mostly pallid spikelets. Culms usually simple or with some culms branched from the
upper nodes, 70-150 cm high, in very tall plants almost reed-like and up to 3-5 mm
thick.
Fig. 70. — Distribution of A. stipitata: ^ var. stipitata; Ovar. robusta; @ var. spicata, ®var. grad-
I i flora.
Orange Free State. — Bothaville; Acocks 19120.
Transvaal. — Pretoria: Smith 6573; Schweickerdt 1695. Warmbad:
Smuts 679. Waterberg: de Winter 752; Codd 971. Potgietersrust;
South West Africa. — Gobabis; de Winter 2458. Ovamboland:
Giess 7038. Okavango: de Winter and Wiss 4400.
Pole Evans and
de Winter 2280.
de Winter and
Distribution. — (See Fig. 70.)
This variety occurs mainly in areas characterized by deep white sandy soil. The
type, collected by Schinz at Omatope in South West Africa, has the columns ot the
awns up to 40 mm long but in most other gatherings they rarely exceed 30 mm. Hie
length of the column is not correlated with other characteristics that could be used
to distinguish var. stipitata from the other varieties of A. stipitata and must be regarded
as variable.
(6) var. robusta {Stent & Rattray) de Winter in Kirkia 3. 133 (1963).
A. graciliflora var. robusta Stent & Rattray in Proc. Rhod. Sci. Ass. 32. 44 (
292
Practically indistinguishable from var. stipitata except for the branches of the
narrow panicle which are spreading and rather lax, i.e. not dense and appressed. Inter-
mediates between these two varieties are not uncommon.
Bechuanaland. — de Winter 7394.
South West Africa. — Okavango: de Winter & Marais 4607; Maguire 2333 and
2453; Story 6421 and 6477. Otjiwarongo: Volk 1087 and 1002.
Distribution.— (See Fig. 70.)
Maguire, who collected this variety in the Okavango Native Reserve, reports
that it grades into the var. stipitata. This observation supports the author’s view that
A. stipitata and A. graciliflora cannot be regarded as distinct species. Schweickerdt
cited several specimens of var. robusta under A. graciliflora, from which it can be
distinguished only on size. Occurs on deep sandy soils in the Kalahari.
(c) var. spicata de Winter in Kirkia 3: 133 (1963). Type: Bechuanaland, Tsabong,
de Winter 7485 (PRE, holo.!).
Erect, lax to dense, usually much-branched plants rarely exceeding 60 cm in height,
with dense pallid spike-like panicles often borne in terminal fascicles or pairs.
Orange Free State. — ’Botha ville: Acocks 18711. Heilbron: Brandmuller 118.
Cape. — Gordonia: Leistner 2033; 2164 and 1136; Barnard Hay: Acocks 2159.
Barkly West: Ferrar 2. Kuruman: Acocks 458.
South West Africa. — Gibeon: van Vuuren & Giess 1119.
Bechuanaland. — Tsabong: de Winter 7485.
Distribution. — (See Fig. 70.)
Prefers sandy soils associated with rocky outcrops. Intermediates between var.
spicata and var. graciliflora occur fairly frequently.
id) var. graciliflora (Pilger) de Winter in Kirkia 3: 133 (1963).
A. graciliflora Pilger in Engl. Bot. Jahrb. 53: 599 (1907).
Erect, rather lax, much branched plants only occasionally exceeding 60 cm in
height, with narrow but lax, often effuse, purpUsh panicles.
Cape. — Kimberley: Acocks 1605. Barkly West: Acocks 1432.
Orange Free State. — ^Hoopstad: Goossens 1250. Kroonstad: Goossens 1152.
Transvaal. — ^Wolmaransstad: Liebenberg 2355. Bloemhof: Burtt-Davy 13030A.
Ventersdorp: Pole Evans 3144. Potchefstroom : Theron 4. Pretoria: Smith 6567.
Brits: de Winter 2194. Rustenburg: van Nouhuys 22. Waterberg: de Winter 752.
Warmbad: de Winter 6017. Potgietersrust: Galpin 425. Nelspruit: van der Schijff
213 and 1332.
Swaziland. — Acocks 15339.
Natal. — Ubombo : Tinley 609.
Bechuanaland. — Liebenberg 5485; Hillary & Robertson 565.
Distribution. — (See Fig. 70.)
The var. graciliflora is the most widespread and most common of the varieties
and, although frequently associated with sand, can also be found on red loamy soils,
frequently in rocky situations.
The varieties now included in A. stipitata show no anatomical differences in the
structure of the leaf-blade and differ only in size:, the leaf-blades of the larger plants
being more robust. The curved leaf-blade of var. graciliflora depicted in Fig. 159:
26a is not characteristic since in the majority of cases the leaves are fully expanded.
293
Fig. 71. — A. pilgeri: cross section of the leaf-blade {de Winter 27.10).
r — I
0’1 mm
Fig. 72. — A. pilgeri: abaxial epidermis of the leaf-blade.
27. A. pilgeri Henrard, Crit. Rev. Aristida 2: 443 (1927)
Perennial, caespitose, robust, up to 1-5 m high. Culms erect, simple, 3-6-noded;
internodes and nodes glabrous. Leaf-sheaths tight, pallid, margins ciliolate; amides
glabrous or bearded. Leaf-blades linear, expanded at the base and usually laxly
involute upwards, up to 35 cm long. Panicle usually dense, more or less contracted,
at times subspicate, up to 40 cm long and 5 cm wide; branches up to 11 cm long,
solitary. Spikelets congested, shortly pedicelled. Glumes slightly unequal, lanceolate
to linear-lanceolate, awned; lower 7-5-9 mm long including the 1 -5-2 mm long awn;
5752152-4
294
the upper 10-12 mm long including the 2 mm long awn. Lemma 6-7 mm long including
the column, finely punctulate, scaberulous upwards; column about 1-5 mm long
twisted; awns subequal, 10-12 mm long; callus 1-5 mm long, bearded, acuminate.
Caryopsis linear, terete.
Anatomy (Fig. 71, 72 and 159; 27)
Leaf-blades flat or slightly curved in transverse section; abaxial surface flat; adaxial surface
grooved between the shortly protruding bundle units, covered with short bristle-like hairs opposite
the vascular bundles; midrib not projecting abaxially. Abaxial epidermis: stomatal zones with one
to two rows of stomata, ripple-walled cells elongate-rectangular, with strongly and deeply undulate,
slightly thickened walls with short, paired or unpaired elements scattered amongst them as well as some
bi-cellular hairs: silicified zones: silicifled cells dumb-bell-shaped, kidney-shaped or circular, usually
accompanied by a square to elongate-rectangular thin-walled suberized cell with undulate walls;
ripple-walled cells mostly very strongly thickened, the undulate walls apparently almost solid due
to the approximate ripples, this zone flanked by unicellular very sharp-pointed retrorse hairs projecting
obliquely over the stomatal zones. Vascular bundle units: first order units 7, square to very broadly
oblong with truncate apices; second order units only very slightly smaller than the first order units;
the leaf of approximately the same width throughout; midrib similar to the other first order units.
Motor cells in Y-shaped groups one to two cells wide, and occupying the whole width of the leaf, present
and alternating with all but the two marginal bundles.
Transvaal. — •Pretoria: Schweickerdt 2036. Potgietersrust; Liebenberg 4418.
South West Africa. — -Otjiwarongo : de Winter 11 ZQ', Liebenberg Grootfontein :
Giess 225Q', Schweickerdt 2Q55. Okavango: Fo/A: 1706; 1831 and 2156; Story 52A6.
Distribution. — ^(See Fig. 73.)
Fig. 73. — Distribution of 0A. pilgeri; MA. monticola.
295
Of the South African species this is the most distinctive taxon in the section
Pseudarthrathemm. It is more robust than most of the other species, often being up
to 150 cm high and differing in the structure of the inflorescence from all the others.
The panicle in this species is contracted but much branched, the branches solitary,
up to 11 cm long and with the spikelets appressed. The other. species have spike-hke,
very dense, or very narrow and interrupted inflorescences with short branches, or if
the inflorescence is divaricate then the branches bear spike-hke clusters at their apices.
In A. pilgeri the column of the lemma is relatively short and not much twisted. Less
robust plants often resemble A. canescens so closely that an examination of the
articulation of the lemma and the nervation of the leaves is necessary before a correct
determination is possible.
Fig. 74. — A. congesta subsp. congesta: cross section of the leaf-blade {de Winter 5986).
I I
O'l mm
Fig. 75. — A. congesta subsp. congesta: abaxial epidermis of the leaf-blade.
296
Anatomically it agrees so closely with the other species of the section that it cannot
be distinguished without the use of additional organographic characteristics. It is
not particularly closely related to any of the other species included in the section
Pseudarthratherum.
28. A. congesta R. & S., Syst. Veg. 2: 401 (1817)
Perennial, densely tufted, up to 75 cm high, or in desert areas reduced to a small
annual. Culms simple or branched from the lower nodes, erect or geniculate, several-
noded; internodes usually somewhat compressed, glabrous; nodes glabrous. Leaf-
sheaths compressed, glabrous or scaberulous upwards, lower strongly keeled the upper
less so; auricles bearded or glabrous. Leaf-blades linear, up to 20 cm long, folded
or convolute. Panicle dense, spike-like, often interrupted towards the base, or composed
of many peduncled, false spikes in a spreading and divaricate inflorescence; up to
20 cm long. Glumes unequal, sub-hyaline; lower lanceolate, scaberulous on the keel,
and often minutely scaberulous on the flanks upwards, up to 8 mm long, awned;
upper 10 mm long, scaberulous on the keel upwards, shortly awned. Lemma 7-10
mm long including the callus and column ; scabrid to coarsely scabrid upwards; column
up to 5 mm long; awns subequal, up to 25 mm long but usually shorter; callus 1 mm
long, bearded, acuminate. Caryopsis linear, terete.
Anatomy (Fig. 74, 75 and 159: 28)
Anatomically practically indistinguishable from A. pilgeri.
Inflorescence subspicate, very dense and contracted, occasionally with one or two subspicate
peduncled branches at the base subsp. congesta
297
Inflorescence composed of several to many peduncled spike-like branches arranged in an open
divaricate panicle subsp. barbicollis
(a) subsp. congesta.
A. congesta R. & S., Syst. Veg. 2: 401 (1817). A. congesta var. genuina Choiv., FI.
Erit. 383 (1899). A. rangei Pilger in Engl. Bot. Jahrb. 48: 344 (1912). A. longicauda
Hack. ap. Henriques Bol. Soc. Brot 6: 143 (1883). A. alopecuroides Hack, in Verb.
Bot. Ver., Prov. Brandenb. 30: 114 (1888). A. congesta var. megalostachya Henrard,
Mon. Aristida 1: 126 (1929).
Chaetaria congesta (R. & S.) Nees, FI. Afr. Austr. 189 (1841).
Inflorescence dense and spike-like, occasionally with one or two peduncled spike-
like branches at the base; lemma scaberulous upwards.
Fig. 11.— a. congesta subsp. barbicollis: cross section of the leaf-blade {de Winter 791).
0-1 mm
Fig. 78. — A. congesta subsp. barbicollis: abaxial epidermis of the leaf-blade.
Cape. — Komga: Flanagan 2317. Graaff-Reinet : Galpin 10588. Gordonia . Pole
Evans 2173.
298
Basutoland. — Staples 142.
Orange Free State. — Senekal: Goossens 933.
Transvaal. — Pretoria: Codd 860. Waterberg: Burtt-Davy 1112.
Natal. — Estcourt: Pentz 79. Dundee: Acocks 10273.
South West Africa. — Gobabis: de Winter 2439. Windhoek: de Winter & Giess
7142. Ovamboland: Schinz (H. Vindob. No. 15847). Okavango: Maguire 2480.
Distribution. — (See Fig. 76.)
In the drier areas of the Cape and South West Africa this variety tends to behave
as an annual. This annual form was described by Pilger as A. rangei. Since a gradual
transition from the annual to the perennial habit can be traced in specimens from these
areas, A. rangei cannot be regarded as a distinct species.
The South African specimen cited by Schweickerdt as A. longicauda (Moss &
Rogers 608) cannot, in the author’s opinion, be regarded as specifically distinct from
A. congesta sens. lat. It is likely that A. longicauda also merely is a form of A. congesta
since the inflorescence fragment of the type in PRE (from Mozambique) dilfers so
little from the South African material as to be indistinguishable.
{b) subsp. barbicollis {Trin. & Rupr.) de Winter in Bothalia 8 : 173 (1964).
A. barbicollis Trin. & Rupr., Gram. Stip. 135 (1842).
A. barbicollis var. conglomerata Henrard, Crit. Rev. Aristida, Suppl. 705 (1933).
Inflorescence composed of several to many peduncled spikes arranged in an open
divaricate panicle. Otherwise similar to subsp. congesta both in floral and vegetative
characteristics.
Anatomy practically identical with that of the suhsp. congesta (Fig. 77, 78 and 159: 28a)
Cape. — Albany: Dyer 1365. Mount Frere: Story 935.
Transvaal. — Bloemhof: Leistner 125. Pretoria: de Winter 7511. Potgietersrust:
Galpin 422. Nelspiuit: v.d. Schijjf 309.
Swaziland. — Pole Evans 3446.
Natal. — Pietermaritzbuig: Goossens 153. Estcourt: West 1560. Hlabisa: Ward
2435.
Distribution. — (See Fig. 76.)
No difference in the structure of the spikelets of subsp. barbicollis and subsp.
congesta could be found and some plants cannot be referred with certainty to either
subspecies, not even on the characteristics of the inflorescence. The majority of
specimens are comparatively easily distinguished. Subspecies barbicollis is much more
limited in distribution than subsp. congesta.
2.7.8. Relationship and Derivation of the Sections
The probable derivation of the sections of Aristida and Stipagrostis is graphically
presented in Fig. 79.
Of the six sections of Aristida, the section Aristida (Cliaetaria) has the widest
distribution and the simplest floral stiucture. As pointed out by Pilger (1956, p. 119
Footnote III) in a discussion of the awns of the genus, those lemmas possessing no
column can be regarded as having a simple tri-lobed apex, the lobes being very narrow
and awn-like. The development of a column by the elongation and twisting of the
apex of the lemma, as well as the acquisition of an articulation, must be regarded as
progressive steps in the evolution of the group. The section Aristida consists of species
in which the lemmas either lack a column or possess a column. The column may be
weakly to well developed. None of these species, however, has articulated lemmas.
It seems reasonable, therefore, on basis of the relatively primitive structure of its
299
lemmas, to regard it as the oldest or most primitive of the sections. It is probable
that most of the other sections were derived, by slight modifications, from this section.
The section Pseudarthratherum, for instance, may be distinguished from the section
Aristida by the presence of an articulation at the summit of the column just below
the branching point of the awns. A. burkei was placed in the section Aristida {Chaetaria)
by Henrard. Schweickerdt (1941, p. 169) however, found that in A. burkei an articulation
was sometimes present, and also pointed out that the callus was bifid, and not obtuse
or acute as in the section Aristida. On this evidence he placed it in the section
Arthratherum as a variety of A. dijfusa. This decision is supported by the anatomical
evidence presented in the present paper. A. sericans, a species with plumose awns,
was placed by Henrard and Schweickerdt in the section Aristida {Chaetaria) because
of the absence of an articulation. A. sericans is extremely similar to A. capensis var.
dieterleniana (Stipagrostis zeyheri var. sericans). Examination of a wide range of
material of both these taxa has proved that intermediates between non-articulated
and distinctly articulated lemmas exist. Anatomical investigation has furnished
additional evidence that A. sericans must be considered to be a representative of the
genus Stipagrostis which, with the exception of the species just mentioned, invariably
has a pronounced articulation. The presence of plumose awns must, therefore, be
regarded as taxonomically more important than the absence or presence of an
articulation. Another good example of relationships which cut across the division into
sections based on the presence or absence, as well as the position of the articulation,
is found in the species characterized by having ventrally furrowed florets. A. hordeacea,
the only South African member of the section Pseudochaetaria, has a ventrally-furrowed
lemma with an articulation situated at the base of the awns. A. hubbardiana, a South
West African species, and A. ellipticc from South America, are closely related to A.
hordeacea but lack an articulation, yet possess ventrally-furiowed florets. The two
species A. hordeacea and A. hubbardiana likewise show similarity in the anatomical
structure of the leaf-blade. It therefore seems probable that, in spite of having been
placed in two separate sections, these species are more closely related to each other
than to many of the other species in the sections to which they were relegated. The
exclusive use of the absence or presence (combined with the position) of the articulation
for subdividing the genus Aristida into sections, although of great piactical value,
cannot be regarded as entirely satisfactory from the taxonomic point of view.
The section Streptachne occurs in most areas where the section Aristida is
represented, except for South America, West Africa, Southern Africa and China. A
pronounced articulation is lacking and the known species exhibit all stages from short
lateral awns to a complete suppression of the lateral awns. The species of the section
Streptachne do not form a natural coherent group and the probable explanation seems
to be, that they developed independently from different members of the section Aristida
in (or in the vicinity of) the areas in which they occur today. A tendency towards the
reduction of the lateral awns can be seen in a number of species placed in the section
Aristida, e.g. in A. transvaalensis where the lateral awns vary greatly in length and some
of the individual spikelets in which the lateral awns are practically absent could,
technically, be referred to the section Streptachne.
A South West African species which was previously referred to Stipa is transferred
to Aristida on both organographic and anatomical grounds; Aristida parvula (Stipa
parvula) has lemmas with a single awn, but this awn is deciduous, conspicuously twice
geniculate and the articulation is situated between the foot of the column and the apex
of the lemma; in addition to the above the rounded callus is conspicuously long-
bearded. None of the species of the section Streptachne has deciduous or conspicuously
twice geniculate awns, even though signs of an articulation may be present. For this
reason A. parvula is placed in a separate section Schizachne. Aristida parvula is an
annual and seems to be allied to annual species of the section Aristida such as A. effusa,
which it resembles greatly in habit. It does not seem to be related to any of the species
300
of the section Arthratherum, which have similar articulations, but possess three awns,
larger spikelets, acute or bifid calli, and are usually all perennials. In Aristida parvula
the callus is obtuse as in the species of the section Aristida. It seems most likely that
it was derived from one of the annual species of the latter section by reduction of the
lateral awns and the acquisition of an articulation.
Section Pseudar- “ Diffusae ” “ Graciliflorae "
The section Pseudochaetaria is composed of three species: only A. hordeacea is
represented in Southern Africa. This section is probably also a derivative of the section
Aristida. As stated earlier in this discussion, A. hordeacea is undoubtedly related to
A. hubhardiana, a member of the section Aristida. The latter would become a typical
member of Pseudochaetaria by the acquisition of an articulation between the base of
the awns (no column present) and the apex of the lemma.
The derivation of the section Arthratherum presents more difficulty. The species
in this section have, on the whole, larger spikelets and are often more robust plants
than those in the section Aristida. The calli are either bifid, or very acute and long.
A. sciurus, the only very robust South African representative of the section Aristida,
strongly resembles some of the species of Arthratherum in vegetative characteristics. In
leaf anatomy A. sciurus approaches A. diffusa (section Arthratherum) more closely than
301
any of the South African representatives of the section Aristida. The spikelets, however,
agree with those of species of the latter section in most details. Some of the species
at present placed in the section Aristida eventually may prove to belong to the section
Arthratherum, as was found to be the case in A. dijfusa var. burkei. Except for A.
dijfusa var. burkei, the only other species with a bifid callus placed in the section
Aristida by Henrard is A. riparia Trin. : the large spikelets and 15 mm long column
suggests that A. riparia may also belong to the section Arthratherum. Should this
suspicion prove to be correct, this will be the first record for the section Arthratherum
from South America. It seems unlikely that Arthratherum could have been derived
from Pseudarthratherum which is a smaller, less widespread, section than the former.
The section Pseudarthratherum is absent from North America, where five species
belonging to the section Arthratherum occur, and is represented by two species in
South America where the section Arthratherum is absent. All the characteristics
combined in the section Arthratherum are present in the section Aristida, except for
the bifid callus found in a number of African species, and the articulation. It therefore
seems probable that Arthratherum was also derived from some ancestor belonging to
the section Aristida and was probably not derived from any of the other sections.
It is fully realized that the placing of related species in different sections is
taxonomically inaccurate, but for practical reasons, i.e. the ease of determination and
reference, it is probably advisable to retain the basic arrangement presented by Henrard
although in a somewhat modified form.
2.7.9. Interrelationship of the South African Species
The affinities of the various species are discussed in the descriptions.
In Fig. 80 the possible interrelationship of the species is graphically presented.
Unbroken double lines indicate very close affinity, unbroken single lines, close affinity
and dotted lines more distant affinity.
Certain points are evident when this diagram is studied. The annuals form groups
of rnter-related species; links with the perennial species, even within the same section,
appear to be absent. The absence of a column seems to be associated mainly with
an annual habit, but exceptions occur, such as the perennial A. canescens, which has
no column, and A. parvula, an annual with both a column and an articulation.
The species of the section Aristida fall into five groups; 1. The annual group
typified by A. adscensionis, including the sub-perennial A. bipartita, consists of relatively
small plants occurring mainly in disturbed soils. All of these are without a column
and agree with each other closely in leaf anatomy, shape of grain and general charac-
teristics of the spikelet. 2. The second group contains two species: A. rhiniochloa
and A. hubbardiana. These two species are not very closely related but agree in the
rather broad ventrally furrowed grain which adheres to the palea, and the scabrid
surface of the whole lemma, which is narrowed towards the apex, as well as in the
anatomical structure of the leaf. In all these characteristics they agree with A. hordeacea
which, however, is placed in a distinct section due to the presence of an articulation
at the apex of the lemma. 3. A. transvaalensis, A. aequiglumis and A. monticola are
grouped together on the characteristic anatomy of the leaf-blades, the similarity in
the structure of the spikelet, as well as the branched culms. A. canescens, although
somewhat divergent, should be linked with this group. 4. The fourth group consists
of two species typified, anatomically, by leaf-blades which are V-shaped in cross-
section; A. recta, although very distinct from A. junciformis, is probably best placed
near it. Both species show a marked preference for habitats which provide moist
summer conditions. 5. A. sciurus occupies a somewhat intermediate and isolated
position. In the characters of the spikelet it belongs to the section Aristida, but
anatomically, and in habit it approaches species such as A. diffusa and A. dasydesmis
of the section Arthratherum.
302
Fig. 80. — Interrelationship of the species of Sartidia and the South African species of Aristida.
303
The section Arthratherum falls roughly into three groups; —
(1) A. diffusa and its allies, including those species with a bifid callus, a rather
short column to the awns, and strongly to slightly projecting bundle units
as seen in cross-sections of the leaf-blades.
(2) A. meridionalis, A. spectabilis and A. stipoides form a second group, charac-
terized by large open graceful inflorescences with numerous spikelets, a tuft
of wool at the mouth of the sheath, as well as a bifid callus; anatomically
they are very similar, the bundle units being nearly equal in height as seen
in cross-section.
(3) A. moUissima and its allies, are anatomically indistinguishable from group 2
above but the inflorescences are less copious, the spikelets rather lobust,
the calli acute, pungent and well developed, and the culms often branched.
The section Pseudarthratherum consists of a number of taxa so closely allied that
specific identification is often very difficult. A. pilgeri is practically the only species
where intermediate links with the other species are absent. Due to its robustness it
often resembles, in habit, species belonging to the section Arthratherum.
As discussed previously the genus Sartidia is very distinct anatomically and no
evidence of close affinities with any of the sections of Aristida could be found.
This anatomical study has, therefore, confirmed that the classification based on
organographic characters, is, with a few exceptions basically sound. Furthermore on
the basis of anatomical structure, it is now possible to recognize groups of related
species within the sections more clearly than was hitherto possible on purely organo-
graphic grounds. Moreover anatomical study indicated the existence of an anomalous
group of species, i.e. the genus Sartidia, which shows little affinity with Aristida in
spite of a great organographic similarity.
From the above it may be inferred that an anatomical study of all the species
of Aristida should provide valuable data for a refinement of the classification as well
as the clarification of possible relationships within and between the sections.
Fig. 81. — World distribution of the genus Stipagrostis.
304
2.8. THE GENUS STIPAGROSTIS
2.8.1. History
The genus Stipagrostis, described by Nees in Linnaea 7: 290 (1832), was based
on a single species, S. capensis Nees. Since S. capensis is a synonym of Aristida obtusa
Del. the latter must be regarded as the type species of the genus. Kunth (1833, p. 197)
accepted Nees’s genus and included only the type species.
In 1841 Nees enlarged the genus to include two more species: S. dregeana and
S. geminifolia. Apparently he did not regard the plumose awn(s) as an important
characteristic, since all the other plumose species then known were accommodated
by him in the genus Arthratherum. The diagnostic character used to distinguish
Stipagrostis was the presence of two small lobes flanking the column at the apex of
the lemma. This character has subsequently proved to be of little value. Trinius and
Ruprecht in 1842 reduced Stipagrostis to a section of Aristida including in it all the species
with plumose awns known to them, totalling seventeen species. Nine years later Figari
and Denotaris dscribed the genus Schistachne based on Aristida ciliata Desf. Henrard
in his monograph of the genus Arisida regarded both Stipagrostis and Schistachne
as sections of Aristida: Schistachne containing all the species which have plumose
awns and an articulation at or near the middle of the lemma, and Stipagrostis all those
with plumose awns articulated at the apex of the lemma, except for A. sericans which
has no articulation and therefore was placed in Chaetaria.
Schweickerdt’s account of the South African species is basically the same as that
of Henrard except for the clarification of the status of some of the species. The latest
account is that given by Chippendall (1955) which is an abbreviated account based on
Schweickerdt’s revision.
2.8.2. Distribution and Ecology
In contrast to Aristida, which is pantropical in distribution, Stipagrostis occurs
only in the Old World where it is limited to the drier areas of Africa and the Middle
East (Fig. 81). The majority of the species of Stipagrostis occur in desert or semi-desert
areas and the general distribution of the genus coincides with areas with an average
annual rainfall of less than 250 mm. The only deviation from this pattern is found
in Asiatic Russia which has vast areas with a rainfall of under 250 mm, but where
no species of Stipagrostis occur. The scant data from which the distribution m tnis
area was plotted may be the reason for this apparent deviation. It seems as if
temperature, which in many cases plays such an important role in the distribution
of the Gramineae (Hartley, 1958), is not of great importance here. This is clear from
a study of the map, where many of the areas which support no species of Stipagrostis
are subject to conditions of temperature similar to those areas supporting many species.
This phenomenon probably is explained by the higher annual rainfall (500 to 1,500
mm) of the areas where no species occur.
The disjunct distribution of Stipagrostis (Fig. 81) is not unique and other examples
liave come to the attention of the present author, additional to those which have already
been noted by Engler in Die Pflanzenwelt Afrikas (1921, p. 851). It is interesting
to note that a large number of species are disrupted in distribution, many without
showing any signs of having undergone changes recognizable in the taxonomic sense,
while others have developed varieties or subspecies in the different areas. As pointed
out by Engler, there evidently must have been a connection between these widely separated
floras in the past, probably during a drier geological period in Africa.
It is reasonably certain that the Namib Desert along the West Coast of South-
West Africa and Angola is geologically an extremely old area. In this desert belt,
averaging 60 to 80 miles in width and about 900 miles in length, the majority of the
305
species of Stipagrostis are concentrated, many of them endemics, occupying extremely
limited distributional areas. It seems possible that many of these species originated
in this area, since they do not occur in the northern areas of the generic distribution,
although suitable habitats exist there. S. hirtigluma, S. ciliata and S. obtusa occur
both in Northern and Southern Africa, while all the other species are either southern
or northern in distribution.
As could be expected, the whole group is characterized by adaptation to extremes
of temperature and low rainfall. These adaptations can be divided, into three cate-
gories
{a) avoidance of the driest periods by the adoption of an annual habit;
{b) reduction of vulnerable parts (leaves) and transfer of function of these parts
to less vulnerable organs (culms);
(c) protection of the vulnerable parts by extreme thickening of the cell walls.
The annua! habit is found in only a few species and probably has to be regarded
as a secondary development in this predominantly perennial genus. Species such as
S. hermannii and 5. subacaulis, which occur in the drier parts of the Namib Desert,
are dwarf annuals only a few centimetres high, capable of reaching maturity and
producing seed a few weeks after germination. All the parts are soft, the cells relatively
thin-walled and the plants die off as soon as conditions become unfavourable. The
seeds of these plants are apparently very resistant to desiccation and remain viable
for long periods. No experiments have, however, been carried out to test the longevity
of the seeds of grasses occurring in the Namib areas. That a source of viable seed
is normally available in the soil after several years of drought can be seen after good
rains, when otherwise bare flats are covered by waving grass within a few weeks. Sight
should not be lost, however, of the fact that after very slight showers small plants,
easily overlooked, are present, which although very depauperate, will produce a few
viable seeds and thus replenish the supply.
It is striking that the annual species of Stipagrostis are found mainly in the drier
areas, and that, even within a species, the annual varieties are desert forms, while the
longer-lived forms are found further inland where the rainfall is higher. Examples
of these transitions are found in species such as S. hirtigluma, where the variety patula
is more or less perennial and requires a higher rainfall, whereas the var. pearsonii is
annual, and occurs in the more arid areas. In 5. uniplumis the forms of the typical
variety occurring near the dry coast are short-lived, and the forms of the var. neesii
are progressively longer-lived as they occur further inland in areas with a higher
rainfall. This transition is reflected in the degree of development of the stereo me
strands and number of first order vascular bundles as seen in cross sections of the
leaf-blade.
The second adaptation takes the form of a severe reduction in the size of the
leaf-blades which soon disarticulate leaving the partly bare stems to perfonn the
photosinthetic activity. The leaves of these plants are usually expanded, the stereome
poorly developed, so that they resemble the leaves of the annual species to some extent.
Examples of this adaptation are found in S. fastigiata, S. geminifolia and S. brevifolia.
The vast majority of the species rely on simple thickening of the epidermal cell
walls and reduction in leaf-surface, to conserve moisture. Practically all these species
have very narrow leaves, curved like a horseshoe in cross section, or circular with
a deep longitudinal median groove adaxially. The abaxial epidermis is strongly
thickened while the adaxial epidermis is relatively thin-walled and usually partly
consists of a dense cover of long hairs, which may help to retain a damp atmosphere
in the groove. Except in the few species where the stereome forms a contiimous band
below the abaxial epidermis, stomata occur on both surfaces. In several species they
306
are situated in deep grooves, even on the abaxial surface, and in a few species these
grooves are completely filled with long, woolly, longitudinally arranged hairs which
protect the stomata.
A number of species are psammophytes, for example S. lanipes occurs at the foot
of or between shifting dunes near the coast of South West Africa. If covered by
wind-blown sand, it grows out again and may form hummocks held by the extensive
system of rhizomes. Only the apical few centimetres of the plants appear above the
sand. Herbarium specimens give an entirely erroneous impression, since usually only
the flowering tops are represented. S. omabilis and S. sabulicola are found typically
on the crest of dunes. Both these species are tall and reed-like and play an important
part in stabilizing the dunes. S. sabulicola is found only in areas of the Namib desert
with an average rainfall of 75 mm or less. S. amabilis, however, is a typical constituent
of the dune-crest vegetation of the Kalahari further inland. Most of the other species
are less exacting in their ecological requirements and will grow in a great variety of
habitats and soils.
Fruit dispersal in the genus is highly specialized and effective. The relatively
light fruits are enclosed in the lemmas, which are furnished with 1-3 plumose awns,
and can be transported easily by wind over great distances. The awns are often twisted
into a column articulated to the lemma. The lemma is furnished with a very sharp
callus, which rarely is bifid; the latter bears a ling of antrorse stiff hairs above its
naked, indurate base. These fruits become attached to animals and ate probably
dispersed often in this way. The lemma, awns and callus form a very efficient mechanism
^'or burying the fiuits in the soil. In most of these arid aieas heavy dews are expeiienced
at night. The untwisting of the hygroscopic awns, followed by a retwisting under
drier conditions during the day, drives the sharp callus into the soil. Observations
along the roads in South West Africa have proved that dense masses of awns blown
into hollows are only the ends of the awns appearing above the soil surface while the
fruits are buried deep in the soil. The awns eventually disarticulate leaving the grains
buried at a suitable depth for germination. This enhances the chances of survival
of the young seedlings. The soils of most of these desert areas are sandy or gravelly,
and obviously ideal for the effective operation of this mechanism. Nevertheless similar
adaptations are found in many grasses not subjected to such rigorous conditions.
2.8.3. Karyology (see Fig. 7).
Previous to this study, only two species of Stipagrostis had been investigated
karyologically; namely S. obtusa (as Aristida obtusd) and A. gmcilior var. intermedia,
here regarded as a putative hybrid of S. hirtigluma var. pearsonii and S. uniplumis
var. uniplumis.
The table below summarizes the information available on the chromosome numbers
of Stipagrostis. The locahties and collector’s numbers of specimens prepared from the
plants investigated by the author are indicated and these specimens are preserved in
the National Herbarium, Pretoria. For
publication is cited.
S. uniplumis var. intermedia (as A. gracilior var.
intermedia)
S. uniplumis var. uniplumis
S. uniplumis var. neesii
S. uniplumis (as Aristida uniplumis)
S. hochstetteriana
Y. namaquensis
S. obtusa
S. obtusa (as Aristida obtusd)
S. pungens (as Aristida pungens) .
the Other species the authority and yeai of
307
The basic chromosome number in Stipagrostis, like that of Aristida sensu stricto,
is 11. Most of the species investigated are tetraploid, only one record of a diploid
having been found (Reese 1957, p. 604) in North African material of S. obtusa. The
South West African specimens investigated by the author proved to be tetraploid
\vith 2n = 44. De Wet and Anderson (1956, p. 7) reported 2n = 36 for Stipagrostis
uniplumis var. intermedia {Aristida gracilior var. intermedia). The high incidence of
tetraploids in species of Stipagrostis, which occur mainly in areas of extreme aridity,
conforms with the findings of those workers who have remarked on the high incidence
of polyploidy found in plants from areas with very extreme climatological and edaphic
conditions. Reese (1957, p. 631), who made a thorough survey of the occurrence of
polyploids in north Saharan plants could, however, find no evidence to support this
theory.
As in the case of Aristida sensu stricto, the chromosomes are very small and
morphological studies of the individual chromosomes were not undertaken. The
chromosome numbers are so uniform that, as in Aristida, there appeared to be no
evidence for the delimitation of sections within the genus. The nucleus in stipagrostis
frequently persists to the mitotic metaphase (see also remarks on p. 234 under Aristida).
2.8.4. STIPAGROSTIS Nees
Stipagrostis Nees in Linnaea 7: 290 (1832); Kunth, Enum. Plant. 175 (1833);
Nees, FI. Afr. Austr. 171 (1841). Aristida sect. Stipagrostis, Trin. & Rupr., Gram.
Stip. 163 (1842); Bentham & Hooker, Gen. plant. 3, 2: 1141 (1883); Hackel, True
Grasses, 101 (1896); Henrard, Mon. Gen. Aristida 1; 45 (1929); Schweickerdt in
Bothalia 4, 1; 95 (1941); Chippindall in Grasses & Pastures S. Afr. 291-306 (1955);
Pilger in Nat. Pflanzenfam. 14d, 11; 123 (1956). Schistachne Fig. & De Not in Mem.
Acad. Sci. Tor. 2, 12: 252 (1851). Aristida sect. Schistachne (Fig. & De Not.) Henrard,
Gen. Aristida 1: 35 (1929); Schweickerdt, l.c. p. 93; Chippindall, l.c. p. 291; Pilger,
l.c. p. 124.
Spikelets solitary, pedicelled, borne in terminal, contracted or open panicles.
Rhachilla disarticulating above the glumes, not produced beyond the base of the floret.
Floret 1, hermaphrodite, equalling or shorter than the glumes, the awns usually long
exserted, body narrowly oblong, cultrate or linear in outline. Glumes persistent,
narrow, acuminate to obtuse, occasionally emarginate, 1-5-nerved (7-9- or 11 -nerved
in a few species) mostly both 3-nerved; nerves anastomosing or evanescent. Lemma
cylindrical, or margins involute and lemma grooved ventrally, indurated at maturity,
glabrous or scabrid, 3-nerved, nerves converging towards the apex but not fusing and
each produced into an awn, or upper part of lemma narrowed into a canaliculate,
3-nerved, usually twisted, long or short column, each of the nerves of the column
excurrent into an awn at the apex of the column, lemma articulated at or above the
middle, or articulation situated between the base of the column and body of the lemma ;
awns all plumose or only central awn plumose, if awns single and not plumose, then
with a pencil of hairs from the column; callus well developed, pungent or minutely
bifid, usually oblique, bearded or practically glabrous. Palea usually much less than
half the length of the lemma, indurated, 2-nerved, not keeled, the margins incurved,
glabrous. Lodicules usually 2, occasionally absent, fleshy at the base, membranous
upwards, with several nerves, obtuse. Stamens 3, anthers elongated. Ovary glabrous:
styles free; stigmas plumose, laterally exserted. Caryopsis terete, tightly enclosed
by the lemma but free, frequently shallowly grooved ventrally; hilum finear, slightly
shorter than the grain; embryo the length of the grain; starch grains compound,
composed of numerous granules.
308
Anatomy
Shoots long attenuated or abbreviated; in cross section circular or somewhat flattened and the
leaf-blades conduplicate in the bud. The sheaths which taper towards the free margins tightly enrol
the young leaves, and have groups of large thin-walled cells between the bundles (Fig. 10, A, B, C,
p. 237).
Leaf-blades in transverse section sub-circular with a deep narrow groove adaxially, horse-shoe-,
crescent-, V-shaped or expanded and flat. A marked tendency towards a circular to V-shaped outline
is evident (Fig. 160). Silicified cells circular, depressed circular, rounded rectangular or dumb-bell-
shaped in S. subacaulis and S. hermannii; occasionally dumb-bell-shaped in immature leaves of many
other species. The silicified cells which have thickened walls are usually accompanied by thin-walled
suberized cells, which they may resemble, but the latter are usually more or less square with somewhat
undulated walls. Two types of hair occur. One-celled broad-based, usually retrorse, hairs vary from
long attenuate and needle-pointed to very short retrorse barbs and are found mainly adjacent to, or
flanking the stereome strands. In the species with abaxial grooves long woolly hairs are occasionally
present in the grooves. The second type of hair is two-celled and tubular in shape, the terminal cell
being very thin-walled and deciduous. The 2-celled hairs are found in the stomatal area between the
narrow ends of the long ripple-walled cells and in strongly xerophytic leaves are often difficult to detect.
They occupy positions homologous to the short elements. Chlorenchyma consisting of a single row
of rectangular cells radially arranged around the bundles and separated adaxially by the motor cells,
abaxially by one to two layers of circular or irregularly shaped very thin-walled cells containing a few
chloroplasts.
Awns: column of the awns deeply grooved in most species, more rarely shallowly grooved or
almost solid, furnished with three vascular bundles; central vascular bundle large, surrounded by
colourless, usually strongly lignified cells; lateral bundles somewhat smaller, with a well developed
single layer of radially arranged chlorenchyma cells often interrupted by lignified cells ab- and/or
adaxially. Setae with a single central bundle surrounded by a single layer of radially arranged
chlorenchyma. The bundles in both the setae and column lie embedded in a matrix of cells usually
with strongly asymmetrically thickened walls. At the branching point of the awns and in the basal
part of the setae, the lateral vascular bundles are devoid of a chlorenchyma sheath, as is the case with
the central bundle throughout the column. Above the lower third of the setae all the bundles have a
chlorophyll-bearing sheath which is frequently complete (Fig. 8, E and F).
Embryo in longitudinal section showing the absence of an epiblast, but with a free lower part of
the scutellum, which is separated from the coleorrhiza by a deep oblique cleft; vascular tissue distinctly
elongated between the point divergence of the scutellum bundle and the base of the coleoptile (Fig. 9,
C, D, E, F, G). The margins of the first embryonic leaf (sheathed by the coleoptile! meeting but not
overlapping and with 5 vascular bundles; the coleoptile sheath with two lateral vascular bundles and
the scutellum with one median bundle.
Densely tufted perennials, delicate annuals or suffrutices with a strongly developed
knotty rhizomatous base. Culms erect, simple or branched, hollow oi solid. Leaf-blades
long and narrow, subterete or folded, rarely more or less expanded, occasionally poorly
developed and soon deciuous, only sheaths and the chlorophyll-bearing culms persistent.
Ligule a dense fringe of hairs. Panicles narrow and spike-like, or effuse and open.
Type species: Stipagrostis obtusa (Del.) Nees {=S. capensis Nees).
Name from “ stupe ” meaning tow, and “ agrostis ” meaning grass, an allusion
to the plumose awns of the type species.
A genus consisting of about 50 species. Confined to the desert or semi-desert
areas of the Old World mainly in the western part of South Africa, in Somaliland,
North Africa, the Middle East and extending to Western Tibet. One species occurs
in the southern desert areas of Russia.
2.8.5. Enumeration of the Sections
Stipagrostis is divided into three sections one of which, viz. Anomala, is mono-
typic.
1 . § Schistachne
Stipagrostis sect. Schistachne (Fig. & De Not.) de Winter in Bothalia 8 : 173 (1964)
Schistachne Fig. & De Not. in Mem. Acad. Sci. Tor. 2, 12: 252 (1851). Aristida
sect. Schistachne (Fig. & De Not.) Henrard, Mon. Gen. Aristida 1: 35 (1929). Type
species: S. ciliata Desf.
309
2. § Stipagrostis
Stipagrostis Nees in Linnaea 7: 290 (1832). Aristida sect. Stipagrostis (Nees)
Trin. & Rupr., Gram. Slip. 163 (1842). Type species: S. obtusa (Del.) Nees.
3. § Anomala
Stipagrostis sect. Anomala de Winter in Kirkia 3: 133 (1963). Type species:
S. anomala de Winter in Kirkia 3: 133 (1963). {=Stipa namaquensis Pilger).
2.8.6. Key to the Sections
1. Awns three, all or only the central awn plumose:
2. Body of the lemma with an articulation situated at or just above its middle; at maturity
the awns and column break off with the conical, hollow upper part of the lemma attached
1. Schistachne
2a. Articulation present or rarely absent, if present then situated at or slightly below the
point of insertion of the awns or column; column, together with the awns, disarticulating
on maturity without, or with only a very short, apical hollow part of the lemma attached
2. Stipagrostis
la. Awns solitary, glabrous upwards and with a well developed pencil of hairs on the basal part of
the column; articulation situated between the apex of the lemma and the base of the awn
(column) 3. Anomala
2. 8. 6.1. Key to the South African Species Based on Organographic Characters
1. § Schistachne
1. Only the central awn of the lemma plumose:
2. Glumes membranous to fairly firm or chartaceous, narrow, acute or acuminate with a fine
delicate apex; nodes glabrous, sheaths never woolly:
3. Tufted plants with mostly unbranched culms up to 90 cm high; central awns 4-10 cm
long, lateral awns usually less than half as long as the central:
4. Panicle dense and spike-like or branched, if branched the branches strongly contracted
and with many spikelets; glumes glabrous or only the lower glume long hairy
1. S. hochstetteriana & vars.
4a. Panicle narrow and elongate but lax and interrupted, spikelets borne singly on the
rhachis in the upper part of the inflorescence, lower down on lax short branches each
bearing few (2-7) long-pedicelled spikelets; both glumes softly (and usually shortly)
pilose, at least on the margins 4. S. dinteri
3. Suffrutescent plants up to 2 m high; branches fascicled from the nodes; central awn up
to 3-5 cm long but usually much shorter, lateral awns more than half to nearly as long
as the central 6. S. amabilis
2a. Glumes of a very firm texture (cartilaginous), glabrous or rigidly ciliate, subequal, rather
broad, obtuse or emarginate with slightly ciliate tips; nodes bearded, or if glabrous, the basal
sheaths woolly:
5. Nodes conspicuously bearded with a ring of spreading white hairs, very rarely glabrous;
lower sheaths glabrous or villous; plume of awns silvery white 2. S. ciliata
5a. Nodes glabrous and smooth, lower sheaths sparsely to densely woolly; plume of awns
yellowish 3. S. schaeferi
la. All the awns plumose, usually equally strongly so, or the lateral awns not as densely plumose
as the central awn:
6. Internodes glabrous or scaberulous but never woolly-pubescent; panicle branches scabrous
or glabrous 7. S. namaquensis
6a. Internodes woolly or pubescent just below the nodes; panicle branches woolly-pubescent
5. S. proxima
2. § Stipagrostis
1. All the awns plumose with long, spreading hairs, the lateral at times very scantily plumose:
2. Culms much branched; branches often fascicled at the nodes; plants suffrutescent, tufted or
reedlike; leaves mostly borne on the culms:
310
3. Inflorescence exserted beyond the leaves, effuse and open; leaves usually rather short
and pungent; subwoody suffrutex 11. S. lutescens
3a. Inflorescence overtopped by the leaves, spiciform or narrow and lax; leaves long, setaceous
or filiform; plants reedlike or small and tufted:
4. Plants very robust and reedlike, up to 2 m or more high, inflorescences 10-30 cm long
with numerous spikelets 9. S. sabulicola
4a. Plants densely tufted and fine-leaved usually not exceeding 40 cm in height ; inflorescences
usually less than 10 cm long and with few spikelets (6-20) 10. S. ramulosa
2a. Culms not or rarely branched and then only from near the base; leaves borne mainly on
basal innovation shoots:
5. Inflorescence very narrow, not less than 8 times as long as wide, contracted and interrupted;
glumes glabrous; plume of the awns obtuse 8. S. damarensis
5a. Inflorescence effuse and open or somewhat contracted but not interrupted, if narrow
then glumes pilose; plume of awns acute and excurrent into a naked tip or obtuse:
6. Lateral awns only scantily adpressed ciliate, obtuse or sub-obtuse; glumes glabrous;
inflorescence effuse and open 16. S. dregeana
6a. Lateral awns always conspicuously plumose, excurrent into a naked tip ; plumes acute in
outline; inflorescence open or narrow; glumes glabrous or pilose 17. S. zeyheri & subspp.
la. Only the central awn plumose with long spreading hairs, the lateral awns quite naked, or rarely
very scantily, appressedly ciliate pubescent:
7. Suflfruticose perennials with pubescent or woolly nodes; spikelets grouped together in spici-
form clusters and glumes hairy, or spikelets not conspicuously clustered and glumes
glabrous, but then vegetative parts densely covered with glandular tubercles:
8. Glumes glabrous; nodes woolly; vegetative parts with glandular tubercles 12. S. brevifolia
8a. Glumes hairy, nodes pubescent to woolly; vegetative parts without glandular tubercles:
9. Peduncle of mature panicle usually longer than the upper internodes and well exserted
from the upper sheath; leaf-blades fairly well developed 13. S. fastigiata
9a. Peduncle of mature panicle and upper internodes about equal in length; panicle often
sheathed at the base or at least not well exserted; leaf-blades very weakly developed
14. S. geminifolia
7a. Caespitose perennials or annuals, if suffruticose, then nodes quite glabrous and glumes
glabrous and smooth over the whole surface (occasionally with a few marginal or apical
hairs) :
10. Glumes glabrous or finely scaberulous:
11. Branching point of the awns glabrous or hairy, never with a distinct pencil of long
fine hairs:
12. Lower glume shorter than the upper or subequal, quite glabrous and smooth, or
very sparsely scaberulous on the nerves or near the apex only (never all over);
perennials with culms much branched downwards, or not branched and panicles
effuse:
13. Lateral awns minutely appressedly ciliate near the apices:
14. Caespitose plants, occasionally shortly branched at the very base, leaves mainly
basal; glumes brownish purple; apex of central plumose awn obtuse in outline
16. S. dregeana
14a. Suflfruticose plants with much branched culms, leaves borne on the culms,
no basal tuft present; glumes usually straw-coloured; apex of central plumose
awn obtuse to acute in outline 11. S. lutescens
13a. Lateral awns quite glabrous; culms suflfruticose and much branched in the lower
third only, slender and simple upwards 15. S. garubensis
12a. Lower glume exceeding the upper in length, if subequal then the lower densely,
minutely scaberulous over its whole surface; dwarf annuals or perennials with very
fine basal leaves and narrow, somewhat contracted panicles:
15. Lower glume very finely scaberulous all over; perennials with a dense basal tuft
of fine leaves:
16. Lower leaf-sheaths woolly all over or on the margins only 19. S. lanipes
16a. Lower leaf-sheaths occasionally villous but never woolly:
17. Vegetative parts densely covered with minute papillae, leaves long hairy in
the longitudinal grooves 18. S. gonatostachys
311
17a. Vegetative parts smooth, or blades and sheaths villous, leaves without hairs
in the longitudinal grooves 20. S. obtusa
15a. Lower glume glabrous and smooth or sparsely scaberulous, but then dwarf annuals
not exceeding 10 cm in height:
18. Plants campactly caespitose, culms poorly developed and very short; inflorescence
almost hidden amongst the leaf-blades; column glabrous 25. S. subacaulis
18a. Plants spreading with the culms geniculately ascending and well developed;
inflorescence sheathed at the base but well exserted from the tufts; columns hairy
at least at the branching points of the awns:
19. Glumes thinly membranous, narrow and tapering to a very fine elongate apex,
awns 1-A cm long, column hairy at the branching point of the awn and
occasionally at the base 23. S. hermannii
19a. Glumes chartaceous lanceolate, tapering somewhat abruptly into an acute
or emarginate apex, awns 1 ■ 4-2 • 5 cm long, column hairy all over 24. S. namibensis
11a. Branching point of the awns with a distinct pencil of long fine hairs; glumes glabrous
21. S. uniplumis
10a. Glumes hirsute or pilose over most of their surface, sometimes glabrous at the tips and
on the sides; perennials or annuals:
20. Inflorescences contracted and very narrow, the branches, where present, spiciform
and subsecund; leaves conspicuously scabrid all over, ribbed, abaxially with long crisped
hairs originating from between the ribs 18. S. gonatostachys
20a. Inflorescences usually effuse and divaricate or occasionally somewhat contracted, but
then the branches not spiciform:
21. Branching point of the awns with a distinct pencU of long fine hairs; callus bristles
gradually increasing in length upwards, not interrupted; apex of lemma with obtuse
lobes passing abruptly into the awn (best observed when anthers are enlarged but not
yet exserted); central awn up to 3 cm long 21. S. uniplumis var
21a. Branching point of the awns without a distinct pencil of hairs; the central awn
(and column) variously hairy, 4-5-7 cm long; callus bristles arranged in two tufts;
apex of the lemma gradually narrowing into the awn 22. S. hirtigluma & vars.
3. §Anomala
Awn single, glabrous upwards ; with a well developed pencil of hairs above the articulation between
column and lemma 26. S. anomala.
2. 8. 6. 2. Key Based on Anatomical and Vegetative Characters
1. Suffruticose or caespitose perennials; culms firm, usually not compressible; silicified cells of
epidermis subcircular, more or less square or transversely oblong never dumb-bell-shaped
in outline:
2. Adaxial stereome of the first order bundles consisting of large relatively thin-walled cells with
large cell cavities, or of a layer of typical thick-walled fibres below the epidermis which
towards the bundles gradually merge into larger, relatively thin-walled cells with large cell
cavities :
3. Tall, tufted plants with a branched system of rhizomes; culms reed-like, bearing the
branchlets in fascicles at the nodes; abaxial surface of leaves smooth (not deeply and
narrowly grooved), or with shallow depressions between the ribs, glabrous:
4. Adaxial stereome strands consisting of a layer of typical thick-walled fibres, gradually
merging into larger cells with large cell cavities and relatively thin walls towards the
bundles; bundle units elongate, twice as deep as wide 9. S. sabulicola
4a. Adaxial stereome strands consisting exclusively of large cells with large ceU cavities and
relatively thin walls; bundle units less than twice as deep as wide:
5. Leaves terete, stereome very well developed 6. S. amabilis
5a. Leaves horseshoe-shaped in cross section, stereome well developed . . 7. S. namaquensis
3a. Smaller plants with unbranched culms; the abaxial surface of the leaves distinctly grooved,
or if not grooved the plants much branched but the culms not reed-like, the whole plant
forming a bushy suffrutex or much branched fine-leaved low tuft:
312
6. Abaxial surface of the leaves not grooved or with only slight depressions between the
ribs, glabrous or scabrid in the grooves:
7. First order bundles 5, bushy sulTrutex 11. S. lutescens
7a. First order bundles 3, fine-leaved, soft and densely tufted 10. S. ramulosa
6a. Abaxial surface of the leaves deeply grooved, glabrous, scabrid or hairy in the grooves:
8. Abaxial grooves furnished with long hairs; culms woolly or pubescent below the nodes
5. S, proxima
8a. Abaxial grooves glabrous or scabrid; culms glabrous 8. S. damarensis
2a. Adaxial stereome consisting of typical thick-walled fibres with small to very small cell
cavities :
9. Suffrutices with culms much branched at least at the base:
10. Bundle units of first order strongly projecting adaxially ; stereome strands well developed
and interrupting the bundle sheaths at least abaxially 15. S. garubensis
10a. Bundle units of first order not strongly projecting adaxially, only slightly more than
bundles of lower order; stereome strands rather weakly developed usually not inter-
rupting the bundle sheaths, or by a very narrow intrusion of fibres:
11. Vegetative parts of plants covered with fine glandular tubercles 12. S. brevifolia
11a. Vegetative parts of plants smooth or scabrid, not distinctly gladular-tuberculate :
12. Internodes very unequal: the lower short, longer upwards and very short at the
apex; leaves usually very poorly developed 14. S. geminifolia
12a. Internodes varying in length but not distinctly and constantly unequal; leaves
usually fairly well developed 13. S. fastigiata
9a. Caespitose perennials never suffruticose with unbranched or somewhat branched culms;
leaves long and well developed:
13. Leaf-blades sub-circular in cross section with only three first order vascular bundles;
motor cells in groups flanking or “capping” the midrib, absent elsewhere:
14. Leaves very scabrid with fine, short, broad-based, hyaline spines 4. S. dinteri
18. S. gonatostachys
14a. Leaves glabrous, smooth, woolly or villous:
15. Leaves not deeply and narrowly grooved abaxially 26. S. anomala
15a. Leaves distinctly and narrowly grooved abaxially:
16. Leaves with long soft appressed hairs in the longitudinal abaxial grooves
19. S. lanipes
16a. Leaves scabrid in the grooves but without long soft hairs 20. S. obtusa
13a. Leaf-blades sub-circular, horseshoe- or crescent-shaped, usually with 5 first order
vascular bundles (or very rarely with 3 in weak specimens); motor cells present alter-
nating with most bundles or at least not only flanking the midrib, or if only flanking
the midrib, consisting of one or two large circular cells supported by a few much
smaller cells:
17. Nodes of the culms with a ring of spreading white hairs 2. S. ciliata
17a. Nodes glabrous or scabrid:
19. Abaxial surface of the leaf-blade deeply and rather narrowly grooved; basal sheaths
woolly at least on the margins, glabrescent 3. S. schaeferi
18a. Abaxial surface of the blade not distinctly grooved but often with shallow wide
depressions between the ribs:
18. Abaxial stereome continuous or partially continuous 17. S. zeyheri
19a. Abaxial stereome strands consisting mainly of separate strands or occasionally
2 adjacent strands united:
20. Culms somewhat branched from the upper nodes 21. S. uniplumis
20a. Culms never branched from the nodes 1 7. S. zeyheri
16. S. dregeana
1 . S. hochstetteriana
la. Annuals a few cm to over 60 cm high, culms soft and easily compressible; silicified cells of
epidermis dumb-bell-shaped or sub-circular in outline:
313
21. Leaf-blade moniliform in cross section: the bundle units projecting both ad- and abaxially;
keel distinctly projecting abaxially and furnished with a strong stereome strand; both
surfaces scabrid with prominent hyaline spines:
22. Culms fairly well developed, geniculate:
23. Leaf-blade V-shaped in cross section 23. S. hermannii
23a. Leaf-blade curved in cross section 24. S. namibensis
22a. Culms practically absent or very short 25. S. subacaulis
21a. Leaf-blade not moniliform in cross section: the bundle units projecting mainly adaxially;
no distinct keel present:
24. Leaf-blade V-shaped in cross section 22. S. hirtigluma
24a. Leaf-blade horseshoe-shaped, somewhat expanded or flat in cross section 21 S. uniplumis
2.8.7. Description of the South African Species
1. S. hochstetteriana {Beck, ex Hack.) de Winter in Kirkia 3: 134 (1963)
Perennial forming dense erect tufts up to 90 cm high. Culms 2-4-noded, scaberulous
or glabrous; internodes glabrous. Leaf-sheaths glabrous or sparsely pilose. Leaf-
blades convolute or setaceous, scabrous on the upper surface, glabrous or sparsely
pilose below. Panicle dense, spike-like, up to 9 cm long and 1 cm broad, or branched
with the branches dense and contracted. Spikelets congested on short branches.
Glumes narrowly lanceolate or linear, acuminate, 3-nerved, glabrous or softly hairy.
Lemma spindle-shaped, articulated just above the middle; column well developed;
central awn plumose in the upper part, lateral awns glabrous, fine ; callus acute,
densely hairy. Caryopsis very narrowly oblong-elliptic about 3 ■ 5 mm long ; hilum
the length of the grain.
O’l mm
Fig. 82. — S. hochstetteriana var. secalina: cross section of the leaf-blade (de Winter 3408).
314
0*1 mm
Fig. 83.— 5. hochstetteriam var. secalina: abaxial epidermis of the leaf-blade.
Fig. 84.— Distribution of 5. hochstetteriana; #var. hochstetteteriana; Ovar. secahna; M5. proxima.
315
Anatomy (Fig. 82, 83 and 160: 1 and la)
Leaf-blade crescent-shaped in transverse section, abaxial surface slightly undulate; adaxial surface
deeply grooved; numerous long, one-celled hairs on the abaxial surface, keel absent; margins obtuse
to sub-acute. Abaxial epidermis: stomatal zones with two rows of stomata and 5-8 rows of ripple-
walled cells with strongly undulate walls, interspersed with paired or single, short elements, 2-celled
hairs, and indistinct one-celled sharp-pointed hairs; silicified cell zones with 3^ files of silicified cells
adjacent to the third order bundles and 5-6 files adjacent to the first order bundles, long elements
in this zone with finely but strongly undulate walls, silicified cells sub-circular, usually accompanied
by a rectangular, or broadly oblong, thin-walled suberized cell with undulate walls. Vascular bundle
units: first order bundle units 5, very broadly oblong in outline gradually reduced in size towards the
margin; third order bundle units 2-3 next to midrib, the others alternating with the first order units,
about half the size of first order units and rectangular in outline ; midrib similar to the other first
order units. Bundle sheaths interrupted ad- and abaxially in first order bundles and usually abaxially only
in third order bundles. Chlorenchyma a single layer of tabular cells radially arranged around the
bundles; a few irregular chlorophyll-bearing cells present adaxially in the first order bundles on the
outer perimeter of the regular chlorenchyma and adaxially below the stereome in the third order bundles.
Stereome strands well developed and consisting of fibres; abaxially present opposite all bundles,
adaxially in anchor-shaped strands opposite the first order bundles, present only as few-celled groups
opposite third order bundles. Motor cells present alternating with all bundles, composed of Y-shaped
groups of 6-9 cells, occupying i— | of the thickness of the leaf. [See also Theron (1936; p. 12)].
Lower glume pilose var. hochstetteriana
Lower glume, as well as upper, quite glabrous var. secalina
var. hochstetteriana.
Aristida hochstetteriana Beck, ex Hack in Verb. Bot. Ver. Prov. Brandenb. 30 : 144
(1888).
South West Africa. — ^Keetmanshoop : Leistner 1791; de Winter 3254. Gibeon:
Schweickerdt 2268. Okahandja: de Winter & Giess 7118. Karibib: de Winter 2660;
de Winter 6041.
var. secalina {Henrard) de Winter in Kirkia 3: 134 (1963).
Aristida secalina Henrard, Crit. Rev. Aristida 3: 552 (1928).
Cape. — Roberts 1057.
South West Africa. — Keetmanshoop : de Winter 2 Windhoek: de Winter 2644.
Omaruru: Liebenberg 5035. Outjo: de Winter 3048. Kaokoveld: de Winter &
Leistner 5708.
Distribution. — (See Fig. 84.)
Typical specimens of this species are distinguished from all other species of the
genus by the dense spike-like inflorescences combined with usually erect, unbranched
culms, and a caespitose habit.
Organographically the two varieties differ only in the hairy or glabrous lower
glume, the habit of the plants and characteristics of the spikelet being identical.
Anatomically the varieties do not differ, neither do they occupy different ecological
regions. To retain A. secalina and A. hochstetteriana as separate species on basis of
the slender character of the indumentum of the glumes unsupported by any structural
differences, seems quite unjustifiable, and for this reason A. secalina has been reduced
to the status of a variety under S. hochstetteriana.
The robust specimens with rather branched inflorescences cited below were at
first regarded as putative hybrids between S. hochstetteriana and S. ciliata since they
seemed intermediate in inflorescence characteristics, and were found growing together
with the suspected parents. Latei observations showed these plants to be widespread
in marginal desert areas always in association with S. hochstetteriana. Since they are
mainly found on loose sand in watercourses they are now regarded as representing
very robust specimens of S. hochstetteriana.
316
The following are examples of this form: — ■
South West Africa. — Windhoek: Louw spec. B. Swakopmund: Volk 24. Kaokoveld :
de Winter & Leistner 5745; 5702; 5746.
2. S. ciliata (Desf.) de Winter var. capensis {Trin. & Rupr.) de Winter in Kirkia 3 :
133 (1963)
Aristida ciliata Desf. in Schrad. Neues Journ. Bot. 3: 225 (1809). A. ciliata var.
capensis Trin. & Rupr., Gram. Stip. 164 (1842). A. ciliata var. villosa Hack, in Bull.
Herb. Boiss 4, 3: 18 (1896). A. ciliata var. pectinata Henrard, Grit. Rev. 1: 95 (1926).
A. ciliata var. tricholaena Hack, in Bull. Herb. Boiss 4, 3: 18 (1896).
Perennial, forming dense to rather lax tufts up to 85 cm high. Culms 2-3-noded,
glabrous, branched only at the base from short rootstocks; nodes bearded with long
spreading hairs. Leaf-sheaths glabrous, somewhat woolly on the margins or densely
pilose. Leaf-blades scabrous on the upper surface, glabrous or sparsely pilose below,
convolute sub-setaceous, short or long. Panicle up to 30 cm long, narrow but often
open and loose, the branches well developed and usually appressed. Glumes sub-equal,
3-nerved, glabrous or covered with hyaline spines, obtuse or subacute, sub-coriaceous,
often purplish at the base, 8-5-12 mm long. Lemma smooth, articulated near the
middle; column well developed; central awn plumose; lateral awns glabrous fine;
callus long-hairy, acute. Caryopsis very narrowly oblong, about 3-5 mm long; hilum
the length of the grain.
Anatomy (Fig. 85, 86 and 160: 2 and 2a)
Leaf-blade horseshoe-shaped in transverse section with the arms often somewhat elongated;
abaxial surface slightly undulate, with slight depressions between the bundles; adaxial surface deeply
grooved between the bundles, densely covered with long sharp-pointed unicellular hairs; keel absent;
margins subacute. Abaxial epidermis: stomatal zones with 2-3 rows of stomata and 4-6 rows of long,
strongly ripple-walled cells; silicified cell zones with 7-9 files of silicified cells adjacent to first order
bundles and 4-5 files adjacent to the third order bundles; ripple-walled cells shorter than those of
stomatal zone, with finely but strongly undulate walls; silicified cells sub-circular, single or accom-
panied by slightly larger, more or less square suberized cells with somewhat undulate walls; papilla-like
unicellular retrorse hairs present flanking the zones. Vascular bundle units: first order units 5 or
rarely 3, oblong in outline; third order units triangular, much smaller than the first order units, two
on each side of the midrib otherwise alternating with other first order bundles; midrib slightly smaller
than the other first order bundle units but otherwise similar. Bundle sheaths 2; outer sheath of large
thin-walled cells, the inner of much smaller thick-walled cells; outer and inner sheaths of first order
bundles interrupted ab- and adaxially by the stereome strands except for the midrib where the inner
sheath is intact adaxially; in the third order bundles the outer sheaths are interrupted abaxially and
the inner sheath usually intact, except for the marginal bundles, where both sheaths are interrupted
abaxially, and the outer adaxially. Chlorenchyma a single layer of tabular cells arranged radially
around the bundles. Stereome strands well developed, present opposite all bundles abaxially, and
adaxially present opposite the first order bundles and the marginal third order bundles. Motor cells
in triangular groups of cells alternating with all except the marginal bundles, occupying T-f the width
of the leaf-blade.
Cape. — Prince Albert: Bolus 12432. Clanwilliam: Smith 2600. Kenhardt: Codd
1184. Gordonia: Acocks 18820. Mafeking: Pole Evans 2404. Namaqualand:
Mar loth 12414.
South West Africa. — 'Keetmanshoop : Oertendahl 120; de Winter 3407. Luderitz:
Dinter 3922; Kinges 2248. Swakopmund: de Winter 3209.
Distribution. — (See Fig. 87.)
This species is readily distinguished from all others by the conspicuous spreading
hairs on the nodes, the obtuse subcoriaceous glumes and the erect mostly unbranched
culms. In very old specimens the nodes may appear glabrous but in such cases the
bulbous bases of the hairs can still be observed. In a number of specimens collected
between Aus and Luederitzbucht in South West Africa, the upper nodes are sometimes
glabrous and the spikelets much smaller than in the majoiity of specimens. Nevertheless
these specimens can be referred only to this species. Examples of such specimens
are: Kinges 2727; 2293; Giess & van Vuuren 739; 700a.
317
Fig. 85. — S. ciliata var. capensis: cross section of the leaf-blade {de Winter 3407).
0*1 mm
Fig. 86. — S. ciliata var. capensis: abaxial apidermis of the leaf-blade.
318
Fig. 87. — Distribution of S. ciliata var. capensis.
The characters on which the varieties listed as synonyms have been distinguished
show practically no correlation with habit or othei characteristics, and vary considerably
in the degree of development in different plants. For this reason all these have been
included in a rather variable “ complex vaiiety ” which represents the South African
material of the species. The typical variety is North African.
This species was investigated anatomically by Guenzel (1921, p. 17). Guenzel’s
description of the anatomy of the leaf-blade agrees in essentials with the observations
presented here but the diagramatic drawing of the cross section shows a stronger develop-
ment of the stereome than found in most of the specimens studied by the present
author. In Guenzel’s drawing the third order bundles adjacent to the midrib are
adaxially supported by stereome strands, a feature often not present in less strongly
sclerized leaves [see also Theron (1936, p. 11)]. Jelenc (1950) investigated the anatomy
of this species using north African material. His description agrees in essentials with
that given here.
3. S. schaeferi {Mez) de Winter in Kirkia 3: 136 (1963)
Aristida schaeferi Mez in Fedde Rep. 17: 152 (1921). A. schaeferi var. biseriata
Henrard, Monograph Aristida 1: 41 (1929).
Perennial forming dense tufts up to 70 cm high, with the leaf-blades usually rather
short, rootstock somewhat branched and robust. Culms 2-3-noded, nodes quite
319
glabrous. Leaf-sheaths woolly when young, glabrescent, and even glossy, when older
Leaf-blades mostly basal, short, firm, subulate, recurved, villous-pubescent on the
upper surface; those borne on the culms considerably longer (all leaves longer and
softer in very wet seasons). Panicle up to 26 cm long, contracted but loose. Spikelets
usually purplish near the base. Glumes 3-nerved, glabrous or rigidly ciliate, sub-
coriaceous 8-0-12-5 mm long. Lemma smooth, articulated near the middle; column
well developed; central awn plumose, 4-4-5 cm long, the plumose part often yellowish;
lateral awns naked, fine; callus hairy, acuminate. Caryopsis very narrowly oblong-
elliptic, about 4 mm long; hilum linear as long as the grain.
Anatomy (Fig. 88, 89 and 160: 3)
Leaf-blade horseshoe-shaped in transverse sections; abaxial surface grooved between the bundles,
the grooves gutter-shaped; adaxial surface deeply grooved between the bundles, furnished with
numerous long, sharp-pointed, unicellular hairs; keel absent; margins subacute to acute. Abaxial
epidermis: stomatal zones with 2-3 rows of stomata and 4-6 rows of strongly ripple-walled long cells
with retrorse unicellular hairs scattered amongst them; silicified cell zones with 7-8 files of silicified
cells, alternating with long ripple-walled cells adjacent to the first order bundles, and 3-4 files adjacent
to the third order bundles; silicified cells sub-circular, single or usually accompanied by elongate-
rectangular or square (sometimes broader than long) thin-walled suberized cells with undulate walls.
Vascular bundle units: first order units 5, oblong in outline; third order units oval, 2 on each side
of the midrib and alternating with the first order bundles; midrib slightly smaller but otherwise
similar to the other first order bundles. Bundle Sheaths: outer consisting of large, regular, thin-walled
cells, abaxially interrupted by the stereome strands in all bundles except occasionally the marginal
bundle; only adaxially interrupted in the first order bundles; inner sheath of irregular, smaller thin-
walled cells, only interrupted abaxially in the first order bundles by the stereome strands. Chlorenchyma
a single layer of tabular cells radially arranged around the bundles. Stereome strands present and well
developed opposite all bundles abaxially, and adaxially opposite the first order bundles, absent or
represented by a few fibres adaxially in the third order bundles. Motor cells present in triangular
groups alternating with all the bundles except the two marginal bundles and occupying half or more
of the width of the leaf-blade [see also Theron (1936; p. 11)].
Fig. 88. — S. schaeferi: cross section of the leaf-blade {Volk 26).
320
ST
LC
RB
CC
CS
0*1 mm
Fig. 89. — S. schaeferi: abaxial epidermis of the leaf-blade.
South West Africa. — Luderitz; Giess & van Vuuren 662; 737; Giess 2351; Kinges
2586; 2701. Swakopmund: Volk 19; 26; Giess 3044; 3056.
Distribution. — (See Fig. 149.)
A species closely allied to S. ciliata but usually easily distinguished by the lanate
basal sheaths and the absence of spreading hairs on the nodes. In S. ciliata the basal
sheaths may be densely villous but are never woolly. The leaf-blades of S. schaeferi are
harder and thicker than those of S. ciliata and usually rather blunt and crescent-shaped
but may be longer (in very wet seasons) and then strongly resemble those of S. ciliata.
The awns have a tendency to be yellowish in contrast to the silvery awns of S.
ciliata.
The presence or absence of bristles on the glumes has proved to be a variable
characteristic, and is regarded as unreliable for varietal distinction. The variety biseriata
is therefore not upheld and regarded as a form.
Anatomically S. schaeferi can be distinguished from S. ciliata by the deep rather
narrow grooves between the bundles on the abaxial surface of the leaves.
4. S. dinteri {Hack.) de Winter in Kirkia 3: 134 (1963)
Aristida dinteri Hack, in Bull. Herb. Boiss 2, 1 : 767 (1901). A. coma-ardeae Mez
in Fedde Rep. 17: 152 (1921).
Perennial, forming dense tufts up to 40 cm high. Culms simple, 2^-noded, nodes
glabrous, internodes smooth or scaberulous. Leaf-sheaths glabrous, scabrid. Leaf-
blades convolute, setaceous, very scabrous on both surfaces and with long fine hairs
on the upper surface. Panicle narrow, elongate but lax and interrupted, up to 20 cm
long; branches solitary or binate, up to 10 cm long, each with 1-3 spikelets. Spikelets
erect, pale. Glumes linear-lanceolate, chartaceous, subequal or slightly unequal, 12-16
mm long, 3-5-nerved, sparsely hairy with soft spreading hairs. Lemma glabrous,
body including the column 16-18 mm long, articulated 5-5-5 mm from the base of
the callus; central awn plumose in upper part; lateral awn fine, glabrous. Caryopsis
narrowly oblong-elliptic, the apex subtruncate, about 3-5 mm long; hilum equalling
the grain.
321
Fig. 91. — 5. dinteri: abaxial epidermis of the leaf-blade {de Winter and Hardy 8209).
Anatomy (Fig. 90, 91 and 160: 4)
Leaf-blade horseshoe-shaped in cross-section, grooved between the bundles on both surfaces
with shortish very sharp unicellular hairs on both surfaces; keel protruding slightly more than the
other bundles; margins obtuse. Abaxial epidermis cells rather thin-walled; stomatal zones usually
with 2-3 rows of stomata and 4-8 rows of strongly ripple-walled cells; silicified cell zones with 3-4
files of silicified cells opposite the first order bundles and fewer opposite the third order bundles;
silicified cells broadly oblong or more rarely dumb-bell-shaped but short; ripple-walled cells with fine
but distinct undulations; numerous retrorse barbs present; bicellular hairs present flanking the zones;
322
multicellular somewhat sunken glands present in this zone. Vascular bundle units: first order units 3,
subcircular in outline; third order units alternating with those of the first order or sometimes in pairs
flanking the midrib. Bundle sheaths: outer consisting of large regular thin-walled cells; inner about
half the size of the outer and with somewhat thickened walls ; only the outer sheath containing chloro-
plasts. Chlorenchyma a single radially arranged layer of tabular cells round the bundle. Stereome
rather poorly developed, present on both surfaces opposite all bundles except adaxially opposite the
third order bundles flanking the midrib; fibres typically thick-walled abaxially, with very large lumen
adaxially. Motor cells only present opposite the third order bundles flanking the midrib.
South West Africa.— Omaruru: de Winter and Hardy 8209. Outjo: Galpin and
Pearson 7416.
This species is poorly represented in herbaria. It is related to S. obtusa and its
allies, a fact only fully realized when similarities in the anatomical structure of the
leaf-blade are observed. S. dinteri agrees with S. obtusa, S. lanipes and S. gonatostachys
in the following organographical characteristics: only the central awn plumose and
with a naked base; inflorescence narrow, elongate but laxly branched and often
inteirupted; leaf-blades setaceous. It is distinguished from these species by the articu-
lation which is situated in the body of the lemma, and not at the base of the column.
Theron (1936, p. 8) places this species next to S. obtusa on basis of its anatomy
which is in full agreement with the present author’s view. S. dinteri is, technically,
a member of the section Schistachne while S. obtusa belongs to the section Stipagrostis.
There is no doubt, however, that they are closely allied in spite of the different positions
of the articulation on the lemmas of the two species. The prescence of multicellular
glands on the culms and neives (Fig. 91, p. 321) combined with the caespitose habit
distinguish the vegetative parts of this species from all others. The only other species
known to possess glands is S. brevifolia which is, however, suffruticose in habit.
5. S. proxima {Steud.) de Winter in Kirkia 3: 135 (1963)
Aristida proxima Steud., Syn. Plant. Glum. 145 (1855).
Perennial forming erect tufts and furnished with rhizomes; up to 35 cm high.
Culms 2-6-noded, densely woolly at and just below the nodes, otherwise pubescent.
Leaf-sheaths adpressedly pubescent, the base with a ling or tuft of long white hairs.
Leaf-blades convolute, filiform, rigid and sub-pungent, appressedly pilose between the
nerves below and scabrid on the upper surface. Panicle narrow but rather lax up to
15 cm long, branches and peduncle pubescent to woolly, especially in the axils, 2-5-
spiculate upwards. Spike lets yellowish. Glumes unequal, narrow to lanceolate,
glabrous, 3-nerved, with the lateral nerves anastomosing with the central nerve; the
lower 6 • 5-10 mm, the upper 8-1 1 • 5 mm long, scaberulous on the keel above. Lemma
glabrous, body including the callus 5-5-7 mm long; articulation 3-5-5 mm from the
base of the callus; awns plumose, subequal or the central longer; callus acute, hairy.
Caryopsis not seen.
Anatomy (Fig. 92, 93 and 160: 5)
Leaf-blade horseshoe-shaped in transverse section, deeply grooved both abaxially and adaxially;
keel absent; margins acute. Abaxial epidermis: stomatal zones usually completely obscured by long
wavy one-celled hairs lying longitudinally in the grooves between the bundles; where visible with
2-3 rows of stomata and about 4 rows of rather short, ripple-walled cells with very strongly undulate
walls; silicified cell zones with 7-9 files of silicified cells adjacent to the first order bundles and 5-6
files adjacent to the third order bundles; ripple- walled cells short with strongly undulate walls;
silicified cells sub-circular, usually accompanied by short smooth-walled suberized cells, which are
broader than long; numerous papilla-like retrorse hairs present, those flanking the zones especially
well developed. Vascular bundle units: first order units 3, oblong in outline, third order units alternating
in groups of 2 with first order units, ovate, much smaller than those of first order; midrib like other
first order bundles. Bundle sheaths: outer and inner sheaths interrupted abaxially in all bundles except
the marginal third order bundles, and adaxially only in the first order bundles. Chlorenchyma a single
layer of radially arranged tabular cells; irregular cells present on the perimeter of the chlorenchyma
adaxially. Stereome strands wel developed and present opposite all bundles, consisting of fibres
abaxially, and of large, somewhat thickened, parenchyma cells adaxially. Motor cells in Y-shaped
groups of 9-17 cells occupying F- f of the leaf thickness [see also Theron (1936; p. 14)].
323
Fig. 92. — S. proximo, cross section of the leaf-blade {Flanagan 1657).
0-1 mm
Fig. 93. — S. proximo: abaxial epidermis of the leaf-blade.
^ 2
324
Cape. — Aliwal North: Sister Stephanie 212; Flanagan 1657.
Distribution. — ^(See Fig. 84.)
A very distinct species distinguished from the related S. namaquensis and its allies
by the culms which are woolly just below the nodes, the woolly-pubescent panicle
branches, and the simple culms not branched at the nodes.
Apparently a rare species which so far has been collected only in tire Lady Grey
and Aliwal North districts of the Cape Province. Its basal parts are not completely
known but there ai'e indications that, like the other suffrutescent species, it has a well
developed system of rhizomes.
Anatomically it may be distinguished from S. namaquensis, S. sabulicola and
S. amabilis by the deeply grooved abaxial surface of the leaf-blades, and from S.
damarensis, which likewise has the leaf-blades deeply grooved abaxially, by the long,
woolly, longitudinally appressed hairs present in the grooves.
Fig. 94. — Distribution of amabilis; S. lutescens; Ovar. lutescens; 0var. marlothii.
6. S. amabilis (Schweick.) de Winter in Kirkia 3: 133 (1963)
Aristida amabilis Schweick. in Bot. Jahrb. 76, 2: 217 (1954).
Similar to S. namaquensis, but only the central awn plumose and the leaves much
longer, usually drooping and not as pungent. For other characteristics see description
of S. namaquensis.
Anatomy (Fig. 160: 6)
Leaf-blade sub-circular in outline with a deep groove adaxially; bundles flanking the sides of the
groove densely congested ; motor cells small, with their walls fairly strongly thickened ; cells surrounding
the phloem abaxially, very small and strongly thickened, staining much deeper than the fibres of the
stereome strands. Otherwise agreeing in all respects with S. namaquensis.
325
Cape. — Gordonia; Schweickerdt 2279; Leistner 1365. Namaqualand: Leistiier 2510.
South West Africa. — -Keetmanshoop : de Winter 3398. Gibeon: Keet 1641 (Type);
Leistner 1820; Codd 5870; Keet 1658.
Distribution. — (See Fig. 94.)
As stated by Schweickerdt this species is almost exclusively found on the crest
of sand-dunes in the Kalahari, where it acts as a sand-binder. The closely allied
S. namaquensis on the other hand is found mainly in dry river courses and gullies.
According to Mr. O. A. Leistner, who has made an intensive study of the flora of the
Kalahari, S. namaquensis rarely occurs on the slopes and even the top of sand-dunes
but on the whole the two species occupy distinct ecological niches.
7. S. namaquensis (Nees) de Winter in Kirkia 3: 135 (1963).
Arthratherum namaquense Nees, FI. Afr. Austr. 1; 185 (1841).
Aristida namaquensis (Nees) Trin. & Rupr., Gram. Stip. 174 (1842). A. namaquensis
var. vagans (Nees) Walp. in Ann. Bot. 3; 750 (1852).
Fig. 95. — S. namaquensis: cross section of the leaf-blade {Pearson 3883).
5752152-5
326
Fig. 96. — S. namaquensis: abaxial epidermis of the leaf-blade.
Sujfmtescent perennial with a long creeping rhizome forming lax, sprawling to
dense, erect tufts up to 2 m high but often very much smaller. Culms fascicled,
ascending, prostrate or erect, usually with fascicles of branches from the lower and
middle nodes, or rarely practically simple, glabrous. Leaf-sheaths scale-like below,
well developed upwards, glabrous or rarely sparsely hairy. Leaf blades setaceous or
subulate, convolute, the lower rigid and very pungent, usually glabrous below and very
scabrid on the upper surface. Panicle more or less exserted, narrow, often contracted
but loose, up to 30 cm long but usually much shorter; rhachis and branches nearly
smooth or scabrid. Spikelets yellowish or pallid. Glumes chartaceous, unequel,
lanceolate to linear-lanceolate, gcuminate with bluntish tips, 3-nerved, glabrous; the
lower 8-13, the upper 9-5-15 mm long. Lemma smooth, ending in a short twisted
beak, articulated 2-5-5 mm below the branching point of the awns, lemma and beak
(column) together 8-11 mm long, awns all plumose; somewhat unequal and variable
327
in length, the central 11-25 mm long and usually more strongly plumose than the
lateral. Caryopsis very narrowly oblong, about 5 mm long.
Fig. 98. — S. brevifolia x S. namaquensis: cross section of the leaf-blade {de Winter 3266).
Anatomy (Fig. 95, 96 and 160: 7)
Leaf-blade crescent-shaped in transverse section with a slightly undulating abaxial and a deeply
grooved adaxial surface; upper surface covered with long unicellular sharp-pointed hairs; keel absent;
margin obtuse to subacute. Abaxial epidermis cells with strongly thickened walls: stomatal zones
with 2 rows of stomata and 6-8 rows of strongly ripple-walled cells, with or without a single, or double,
file of silicified cells between the rows of stomata adjacent to the stereome strands subtending the
motor cells: silicified cell zones with 6-7 files of silicified and suberized cells adjacent to first order
bundles and 3-4 files adjacent to the third order bundles; silicified cells broadly oblong, single or
accompanied by square or oblong suberized cells with slightly undulate walls, or suberized cells wider
than long and smooth-walled; ripple-walled cells with fine but distinct undulations; papillae but no
long hairs present; 2-celled hairs few. Vascular bundle units: first order units 5, very broadly oblong
in outline; third order units triangular, two on each side of the midrib and alternating with the other
bundles, much smaller than the first order bundles; midrib similar to the other first order units but
slightly smaller. Bundle sheaths: outer consisting of large regular cells with slightly thickened walls,
inner of smaller cells with walls of about equal thickness to those of the outer; inner and outer
interrupted by the stereome strands ab- and adaxially in first order bundles or much reduced in size
opposite the stereome strands, not interrupted adaxially in the third order bundles. Chlorenchyina
a single radially arranged row of tabular cells around the bundles adaxially with a row of irregular
cells on the periphery. Stereome strands well developed; adaxially in groups of large relatively thin-
walled cells opposite all bundles, often with “ enclaves ” of thick-walled fibres between the bundle and
the larger cells below the epidermis especially in older more strongly sclerized leaves; abaxially in
groups of thick-walled fibres opposite all bundles with additional small groups opposite most motor
cell groups. Motor cells present alternating with alt bundles (except the two marginal ones), forming
triangular groups 2 cells wide, or tapering to one cell in width at the apex, occupying more than half
the width of the leaf-blade.
328
Cape. — Namaqualand : Taylor 1083. Prieska: Bryant 650. Hopetown: Leistner
1306. Kimberley: Leistner 1186.
South West Africa. — Waimbad: Keet 1640. Luderitz: van Vuuren 855. Gibeon:
Liebenberg 5155. Keetmanshoop ; de Winter 3255.
Distribution. — (See Fig. 97.)
Distinguished from the related S. amabilis by the plumose lateral awns and the
pungent leaves, and from S. damarensis and S. sabulicola by the articulation situated
near the middle of the body of the lemma.
For anatomical differences compare Fig. 160: 6, 8, 9 for these species.
An easily recognized species, very constant in floral characteristics but variable
in size. Known to hybridize with S. brevifolia producing plants usually somewhat
closer to S. brevifolia in floral characteristics but intermediate vegetatively. Examples
of such specimens are: de Winter 3266 and Liebenberg 5197 from the Keetmanshoop
district in South West Africa and Acocks 15324 and 15299 from Laingsburg in the
Cape.
The anatomy of de Winter No. 3266 was investigated and also found to be inter-
mediate (cf. Fig. 98 and 160: 11a).
Fig. 99. — S. damarensis: cross section of the leaf-blade (Moss 17837).
329
Fig. 109. — S. damarensis: abaxial epidermis of the leaf-blade.
Guenzel’s (1913, p. 21) description of the leaf-anatomy of S. namaquensis agrees
in all important characteristics with the author’s observations. The essential difference
between the latter two species is the large-celled adaxial stereome in S. namaquensis
[see also Theron (1936, p. 13)].
8. S. damarensis {Mez) de Winter in Kirkia 3: 134 (1963)
Aristida damarensis Mez in Fedde Rep. 17: 152 (1921).
Perennial forming usually rather lax tufts, up to 1-2 m high, branched from near
the base and furnished with well developed rhizomes. Culms simple, 3-4-noded,
glabrous; nodes glabrous. Leaf-sheaths loose, glabrous except for the occasionally
hairy margins. Leaf-blades firm, convolute, grooved abaxially between the nerves,
scabrid or scaberulous in the grooves as well as adaxially. Panicle usually sheathed
at the base, up to 45 cm long, narrow, contracted but rather lax (not spike-like);
rhachis and branches smooth or scaberulous, branches solitary to 3-nate, 10-12-
flowered. Spikelets erect, pallid or yellow. Glumes unequal, the lower usually longer
than the upper, almost glabrous, scabrous, or minutely pilose on the margin and apex,
narrowly lanceolate; the lower 3-5-nerved, and 12-17 mm long; the upper 3-nerved
and 12-14 mm long. Lemma glabrous including the callus 5-6-5 mm long; column
5-7 mm long; awns plumose, the central 15-22 mm long, the lateral 9-15 mm long.
Caryopsis not seen.
Anatomy (Fig. 99, 100 and 160: 8)
Leaf-blade sub-circular in transverse section with a deep median longitudinal groove and deep,
narrow grooves between all the bundles adaxially; furnished with numerous rather short unicellular
bristle-like patent or appressed hairs; abaxial surface with narrow grooves between all the bundles;
bristle-like sharp-pointed unicellular hairs originating from the sides of the grooves; margins subacute
to obtuse. Abaxial epidermis with much thickened walls; stomatal zones: due to deep grooves in which
the stomatal zones are situated only narrow parallel strips of the silicified cell zones remain in scrapes;
these zones are bordered by numerous, bristle-like, sharp unicellular hairs lying more or less longi-
tudinally in the grooves; silicified cell zones with 5-6 files of silicified cells alternating with files of
deeply ripple-walled long elements adjacent to the third order bundles, and 8-9 files of silicified cells
alternating with files of ripple-walled cells adjacent to first order bundles; silicified cells sub-circular,
highly refractive, solitary or accompanied by rectangular, broader than long, suberized cells with
smooth or slightly undulate walls (occasionally suberized cells situated between two silicified cells);
because of the hidden position of the stomatal zones in the grooves two-celled hairs were not observed,
although they are probably present. Vascular bundles: first order bundle units oblong; third order
units much smaller than those of first order, ovate, two on each side of the midrib and alternating with
the other bundles; midrib simdar to the other first order bundles. Bundle sheaths: outer of large rectan-
gular, or rounded, thin-walled cells, interrupted ad- and abaxially by the stereome strands in all first order
bundles, and abaxially only, in third order bundles; inner sheath of smaller somewhat thick-walled
330
cells irregular in size and shape. Chlorenchyma a single row of tabular cells radially arranged around
the bundles. Stereome strands abaxially well developed opposite all bundles, consisting of small thick-
walled fibres; adaxially well developed opposite the first order bundles as relatively small, thick-walled
parenchyma with distinct cell cavities near the epidermis, increasing in size towards the bundles ; those near
the bundles being large with only slightly thickened walls ; opposite the third order bundles consisting
of a few cells only; in the two bundles flanking the midrib and in the marginal bundles forming a small
cap of thick-walled parenchyma. Motor cells poorly developed but present alternating with all bundles
in triangular or oblong groups two cells wide at the base and 3-5 cells deep, occupying half or more
of the leaf thickness [see also Theron (1936; p. 14)].
South West Africa. — Omaruru: Liebenberg s.n.; de Winter & Hardy %\2)\. Karibib:
de Winter & Hardy 8067; 8095. Outjo: de Winter & Hardy 8177. Kaokoveld:
de Winter & Leistner 5717.
A species strongly resembling the widespread S. namaquensis in organographical
characteristics and, due to this resemblance, probably often overlooked. Distinguished
from S. namaquensis by the articulation, which is situated near the apex of the body
of the lemma, by the well developed column of the awns and by the culms which are
branched at the base, but are simple upwards.
Rather rarely collected but a fairly common species in the upper Namib and
coastal Kaokoveld in South West Africa.
Anatomically characterized by the narrow grooves between the nerves on the
abaxial surface of the leaf-blade, and thus distinguishable from its close allies such
as S. namaquensis, S. sabulicola and S. amabilis.
Fig. 101. — S. sabulicola: cross section of the leaf-blade {Keet 1612).
331
0*1 mm
Fig. 102. — S. sabiilicola: abaxial epidermis of the leaf-blade.
9. S. sabulicola (Pilger) de Winter in Kirkia 3: 135 (1963)
Aristida sabulicola Pilger in Engl., Bot. Jahrb. 40: 81 (1908).
Robust perennial, stiffly erect, up to 2 m high, laxly to densely caespitose, with
a much-branched vigorous system of rhizomes. Culms reed-like, erect, glabrous,
fasciculately branched from the nodes. Leaf-sheaths scale-like at the base of the
culms, longer upwards, tight, glabrous. Leaf-blades very rigid, 25-60 cm long,
more or less convolute and sub-terete, smooth on the lower surface, scabrid on the
upper, many-neived. Panicle usually overtopped by the blades, usually sheathed at
the base, narrow, densely spiciform, 10-30 cm long and 1-2 cm broad. Spikelets
densely congested. Glumes glabrous, smooth, slightly unequal, acute or sub-obtuse;
lower 3-nerved, 8-11 mm long; upper 1-nerved, 9-12 mm long. Lemma punctulate
to smooth, including the callus 4-75-6 mm long; column 1-1-5 mm long; awns
densely plumose, subequal, 5-5-10 mm long; branching point of awns with two thin,
hairy appendages; callus acute, curved, densely bearded. Caryopsis obovate, 3-3-5
mm long; embryo about half the length of the grain; hilum as long as the grain.
Anatomy (Fig. 101, 102 and 160: 9)
Leafs-blade horseshoe-shaped in transverse section; abaxial surface smooth or only slighy
undulating, glabrous; adaxial surface deeply grooved due to the strongly projecting ribs, densely
covered with bristle-like unicellular hairs, except opposite the motor cells; keel absent, margins
obliquely subacute. Abaxial epidermis: stomatal zones consisting of one or occasionally 2 rows of
stomatal.and 3-5 rows of short ripple-walled cells interspersed with groups of short elements; silicified
cell zone of 2-5 files of short elements alternating with 1-2 files of longer ripple-walled cells, adjacent
to the third order bundles, and 6-8 files adjacent to the first order bundles; silicified cells sub-circular
in ou line, accompanied by 2-4 suberized cells, one on each side, or a pair on each side of the silified
cells; suberized cells wider than long, rather angular, the narrow ends subacute, in groups of 2^t
together, groups alternating with silicified cells and very broadly oblong cells with papilla-like apices;
no two-celled hairs or unicellular hairs observed. Vascular bundle units 9-13; first order units narrowly
oblong-rectangular, alternating with the third order units; third order units much smaller than the first
order units; triangular in shape; midrib smaller than the other first order bundles, ovate-oblong in
outline (occasionally consisting of a second order bundle). Bundle sheaths: outer of large, round or
more or less square, thin-walled cells densely filled with chloroplasts ; inner of large rather irregular
cells containing no chloroplasts, but with thickened walls. Chlorenchyma a single row of tabular cells
radially arranged round the bundles, supported by one or two layers of thin-walled irregular parenchyma
cells containing chloroplasts, situated on the periphery of the true chlorenchyma. Stereome strands
332
well developed ad- and abaxially opposite the first order bundles; adaxially consisting of small fibres
forming a thick strand, partially enveloping the phloem ; abaxially forming a cap of small fibres supported
by large thick- walled parenchyma cells stretching down to the xylem; small strands of sclerenchyma
also present below the epidermis flanking the vascular buncles usually in a single layer; well developed
strands present abaxially opposite the third order bundles and consisting of small fibres; adaxially
in small strands consisting of small fibres; large thick-walled parenchyma cells absent, or present as
a few cells only. Motor cells in elongate very distinct, narrowly-oblong groups, 1-2 cells wide and
up to 9 cells deep, usually occupying i or more of the thickness of the leaf [see also Theron (1936;
p. 17)].
South West Africa. — 'Luderitz: de Winter & Giess 6181; Kinges 3480. Swakop-
mund: Strey 2443; 2120; Giess 3024; Keet 1705; 1612.
Distribution. — ^(See Fig. 109.)
An easily recognized reed-like species inhabiting dune-tops and sandy gullies in
the drier parts of the Namib, where it occupies an ecological niche similar to that
occupied by S. amabilis and S', namaquensis in the more inland, less dry, areas.
Distinguished from S. amabilis by the awns which are all plumose, and from both
S. amabilis and S. namaquensis by the short awns which spread almost at right angles
and do not exceed the glumes, as well as by the spiciform very dense inflorescence
and by the very strongly developed stiff rachis which is overtopped by the leaf-blades.
Anatomically it is quite distinct from all the other species of the genus on account
of the unique structure of the adaxial stereome of the first order bundles which consists
of small fibres below the epidermis grading into large thick-walled parenchyma cells
towards the bundles ; as well as the shape of the first order bundles which are narrowly
oblong in outline.
Fig. 103. — S. ramiilosa: cross section of the leaf-blade (de Winter and Hardy 8187).
333
Fig. 104. — 5. ramiihsa: abaxial epidermis of the leaf-blade (de Winter and Hardy 8197).
10. S. ramulosa de Winter in Bothalia 8 : 173 (1964)
Perennial, very densely tufted, fine-leaved, much branched, up to 30 cm high
and often forming extensive tufts up to 1 m in diameter. Culms very slender and
fascicled from the nodes, often rather wiry, many-noded, glabrous. Leaf-sheaths
smooth or scaberulous on the back, rather loose. Ligule woolly. Leaf-blades firm,
filiform, ribbed, scabrid especially on the ribs, up to 10 cm long. Panicle lax, elongate,
with relatively few spikelets, sparsely branched, up to 15 cm long (usually shorter),
often hidden by the leaves or only shortly exserted beyond the leaves; rhachis and
branches densely scabrid, pedicels somewhat swollen towards the tips and furnished
with small bristles just below the spikelet. Spikelets erect, pallid. Glumes subequal
or the upper slightly longer than the lower, lanceolate, acuminate, scaberulous on the
keels and apex, 3-nerved, the lateral nerves situated near to the midnerve and reaching
up to half the length of the glumes; the lower 8-9 mm long; the upper 7-8 mm long.
Lemma firmly chartaceous, tubular, 3-nerved, glabrous, body of the lemma up to the
articulation and excluding the callus about 2-5 mm long, articulation situated near
the apex of the lemma about 1 mm below the junction of the awns and included in
the glumes; column formed by the conical upper part of the lemma, about 1 mm long;
awns 3, spreading, all plumose for their whole length, plumes acute to acuminate in
outline, awns with a naked excurrent tip; the central 1-2-1 -5 cm long; the lateral
1 ■ 0-1 • 3 cm long; callus very acute, densely hairy. Caryopsis 2-2 • 5 mm long, narrowly
elliptic in outline; embryo f the length of the grain; hilum as long as the grain,
very shallowly grooved and dark brown.
334
Anatomy (Fig. 103, 104 and 160; 25)
Leaf-blade subcircular to horseshoe-shaped in cross section; adaxial surface with a deep median
channel shallowly grooved on the sides, furnished with retrorse barbs as well as unicellular sharp
hairs; abaxial surface more or less smooth; margins obtuse to almost truncate. Abaxial epidermis:
cells fairly thickwalled; stomatal zones with 1-3 rows of stomata and 2-6 rows of long ripple- walled
elements, undulations deep and closely spaded; silicified cell zones consisting of files of silicified cells
alternating with files of long ripple-walled cells with walls distinctly undulate but less deeply than
those of the stomatal zones. Vascular bundle units: first order units 3, very broadly oblong in outline;
third order units suborbicular to ovate in outline, alternating with or occasionally in pairs between
the first order bundles; midrib similar to other first order bundles. Bu idle sheaths: outer of large cells
with somewhat thickened walls and containing chloroplasts ; inner of smaller cells with thicker walls
and devoid of chloroplasts. Chlorenchyma a single layer of tubular cells radially arranged round the
bundle. Stereome well developed opposite all bundles abaxially and opposite the first order bundles
adaxially, absent adaxially opposite the third order bundles, consisting of typical thick-walled fibres
abaxially, adaxially of much larger cells with large cell lumen. Motor cells present on either side of the
two third order bundles flanking the midrib or occasionally as a single unit capping these third order
bundles.
South West Africa. — Outjo: de Winter & Hardy 8197
A very distinct species which cannot be confused with any others. Easily recognised
by the extremely short column, the subequal awns which are all plumose, the narrow
inflorescences which are overtopped by the leaves, and the finely filiform leaf-blades.
In spikelet characteristics it strongly resembles S. sabulicola which is regarded as its
closest affinity. From the latter it differs mainly in the shorter glumes, longer awns,
hairlike leaf-blades and low tufted habit. Anatomically it shows certain similarities
with the suffrutescent species (adaxial stereome of large cells with large lumen) on the
one hand, and the S. obtusa group (sub terete leaves with three first order bundles)
on the other.
11. S. lutescens (Nees) de Winter in Kirkia 3: 135 (1963)
Perennial, suffrutescent in habit, up to 1 m high, much branched and fascicled
in the lower parts and with a strong, branched system of rhizomes. Culms many-noded,
glabrous; nodes glabrous or bearded with long white hairs. Leaf-sheaths glabrous,
the lower densely covering the culms. Leaf -blades subterete, rigid, short, curved and
pungent or long and stiff, glabrous below, scabrid on the upper surface. Panicle effuse,
pyramidal, up to 25 cm long, branches 2-5-nate or solitary, axils of branches glabrous
or with a pencil of hairs. Spikelets yellowish. Glumes lanceolate, acuminate, shortly
awned or narrowed into an obtuse to truncate apex, 3-nerved; the lower 9-11 mm,
the upper 11-5-14 mm long. Lemma glabrous; column 0-5-5 mm long, usually
twisted; central awn plumose, sparsely hairy in lower part, densely hairy upwards,
up to 2-7 cm long; lateral awns scantily plumose or apparently glabrous but then
scantily and appressedly ciliate. Caryopsis not seen.
Anatomy (Fig. 105, 106 and 160: 10)
Leaf-blade horseshoe-shaped in transverse section; abaxial surface undulate, glabrous; adaxial
surface deeply grooved due to the projecting bundle units, densely covered with bristle-like unicellular
hairs adjacent to the stereome, glabrous elsewhere; keel absent; margins subacute to obtuse. Abaxial
epidermis: stomatal zones with one or two rows of stomata and 3-5 rows of rather short ripple-walled
cells with very strongly and compactly undulate walls; suberized cells broader than long, oblong to
very narrowly oblong in outline; silicified cell zones often without silicified cells, only the suberized
cells being developed; 3-4 files of suberized cells, alternating with files of ripple-walled elements with
finely but densely undulate walls adjacent to third order bundles, and 5-7 files of suberized cells
interspersed with a few silicified cells opposite the first order bundles; silicified cells sub-circular,
solitary or accompanied by a transversely elongated suberized cell; suberized cells from almost square,
to narrowly oblong in a transverse direction, solitary or in pairs, wider than those in stomatal zone;
no 2-celled hairs, or long unicellular-hairs, observed, but a few large cells with papilla-like apex present.
Vascular bundle units: first order bundle inits 5, broadly oblong to obovate in outline; third order
bundle units 2 on each side of the midrib, more or less ovate in outline, alternating with the first order
bundles and with 2-3 first or second order bundles on each margin; midrib similar to the other first
order bundles. Bundle sheaths: outer of thin-walled large-cells containing chloroplasts; inner sheath
335
consisting of cells irregular in size and shape, with thickened walls and without chloroplasts. Chloren-
chyma a single layer of tabular cells arranged radially around the bundles. Stereome strands present
abaxially opposite all bundles, as well as opposite the motor cells and composed of small fibres;
adaxially present opposite all bundles but less developed in third order bundles and there composed
of large thin-walled or slightly thickened parenchyma cells. Motor cells grading into the parenchyma
opposite the ster:ome strands, forming continuous bands across the leaf [see also Theron (1936, p. 18-19)
including A. marlothii].
Fig. 105. — S. lutescens var. marlothii: cross section of the leaf-blade {Strey 2445).
O' 1 m m
Fig. 106. — S. lutescens var. marlothii: abaxial epidermis of the leaf-blade.
336
Axils of the panicle-branches quite glabrous; nodes glabrous; plume of the central awn subacute
in outline var. lutescens
Axils of the panicle-branches distinctly bearded; nodes usually hairy; plume of the central awn
obtuse in outline var. marlothii
var. lutescens.
Arthratherum lutescens. Nees, FI. Afr. Austr. 1: 179 (1841).
Aristida lutescens (Nees) Trin. & Rupr., Gram. Stip. 173 (1842); A. corythroides J.
Karl in Mitt Bot, Staatssamml. Muenchen 1, 3:87 (1951).
South West Africa. — -Luderitz : Kinges 2547 ; Dinter 3944 ; de Winter & Giess
6104; Giess & van Vuuren 734; Merxmueller & Giess 2332; Acocks 15653.
var. marlotbdi {Hack.) de Winter in Kirkia 3: 135 (1963).
Aristida marlothii Hack, in Engl. Bot. Jahrb. 11 : 400 (1889). A. lutescens var. marlothii
(Hack.) Stapf in F.C. 7: 567 (1889).
South West Africa. — Luderitz: Kinges 2348; de Winter & Giess 6098; 6134.
Swakopmund: Giess 3070; Strey 2445.
Distribution. — (See Fig. 94.)
The two varieties mentioned above were previously regarded as specifically distinct
but differ so little, and are anatomically so similar, that if their identical distributional
ranges are taken into consideration one must conclude that at most they represent
varieties of a single species.
S', lutescens is related to S. dregeana and S. garubensis but the relationship is not
a very close one. It may be distinguished from all the other species by the effuse
inflorescence, the stiff rigid leaf-blades, combined with stout culms and a suffrutescent
habit.
Anatomically it fits in well with the other suffrutescent species possessing branched
rhizomes.
Fig. 107.— 5. hrevifolia: cross section of the leaf-blade (Adamson 1538).
337
Fig. 109. — Distribution of #5. brevifolia; MS. sabiilicolc.
338
12. S. brevifolia {Nees) de Winter in Kirkia 3; 133 (1963)
Arthratherum brevifolium Nees, FI. Afr. Austr. 1; 183 (1841).
Aristida brevifolia (Nees) Steud., Nom. Bot. 2: 130 (1841).
Perennial, bushy suffrutex up to 1 m high with a much branched woody rootstock.
Culms many-noded, rough and covered with numerous glands ; nodes with an evanescent
flake of wool. Leaf-sheaths gland-dotted, covered with evanescent wool at the mouth
and along the margins, upper sheath less woolly than the lower; auricles with a flake
of wool. Leaf-blades convolute or flat, 3 mm wide, up to 12 cm long, but usually
much shorter, asperulous below, scabrid on the upper surface. Panicle exserted,
contracted, linear and spike-like but loose and interrupted, up to 20 cm long, usually
shorter. Spikelets yellowish often tinged with purple. Glumes lanceolate, acuminate,
unequal, glabrous, glandular, puberulous or ciliate near the margins the lower 8-15
mm, the upper 11-5-17 mm long, 3-5-nerved. Lemma smooth, including the callus
5-5-7 mm long; column twisted, 4—9 mm long; central awn densely plumose upwards,
2-3-4 cm long; lateral awns naked, fine, 1-2 cm long. Caryopsis not seen.
Anatomy (Fig. 107, 108 and 160: 11)
Leaf-blade crescent-shaped in transverse section; abaxial surface strongly undulate with very
slender, retrorse unicellular hairs in the grooves; adaxial surface deeply grooved between the projecting
bundle units, with many, rather long, unicellular hairs opposite the stereome strands; multicellular
gland-like protruberances, similar to those on the culms, present, margins subacute to obtuse. Abaxial
epidermis: stomatal zones with one to four rows of stomata and 3-5 rows of long, strongly ripple-
walled cells; many slender, unicellular, broad-based sharp hairs present, as well as linear two-celled
hairs; silicified cell zones with about 6 files containing silicified cells, alternating with long, finely
ripple-walled cells, adjacent to first order bundles; 3-5 files containing silicified cells alternating with
files of ripple- walled cells adjacent to the third order bundles; silicified cells sub-circular, usually
accompanied by transversely-oblong to oblong thin-walled suberized cells, which have undulate walls;
broad-based but rather slender retrorse barbs flanking the zones. Vascular bundle units: first order
units 5; broadly oblong in outline; third order units ovate to oblong, 2-3 units flanking the midrib
on each side: 1-3 between the lateral first order bundles and 2-3 on the margins; midrib similar to the
other first order bundles. Bundle sheaths: outer sheaths of large rather thin-walled cells containing
chloroplasts, usually not interrupted by the stereome, the cells opposite the abaxial stereome often
rather small; inner sheaths of somewhat smaller cells with slightly thickened walls in the first order
bundles, and of cells much smaller than those of outer sheath in the third order bundles. Chlorenchyma
a single row of tabular cells arranged radially around the bundles. Stereome strands rather weakly
developed in all bundles of fairly well-developed both ad- and abaxially opposite the first order bundles;
poorly developed opposite the third order bundles, the latter having relatively small groups of fibres;
all strands consisting of typical small-celled fibres. Motor cells in triangular groups of thin-walled
colourless cells occupying more than half the thickness of the leaf and alternating with all the bundles
[see also Theron (1936; p. 12)].
Cape. — Vanrhynsdorp ; Liebenberg 5605. Kenhardt : Acocks 18835. Namaqualandj
Zeyher 1813; Pearson 3405
South West Africa. — Keetmanshoop; Dinter 4970; Oertendahl 575.
Distribution. — (See Fig. 109.)
A very characteristic species due to its dense, bushy sulfrutescent habit, very short
leaves and glandular protruberances on the culms and sheaths. It occasionally hybridizes
with S. namacjuensis (see p. 328).
13. S. fastigiata {Hack.) de Winter in Kirkia 3; 134 (1963)
Aristida fastigiata Hack, apud Schinz in Bull. Herb. Boiss. 1 : 768 (1901); A. walteri
Suessenguth in Mitt. Bot. Staatssamml. Muenchen 11: 39 (1954).
Perennial, sub-suflfruticose with much branched thick rhizomes. Culms arising
in fascicles from the rhizomes. 4-5-noded; internodes more or less equal in length,
somewhat scaberulous, nodes densely bearded with long spreading white hairs. Leaf-
sheaths slightly scabrid, with ciUate margins; auricles densely bearded. Leaf-blades
flat, or convolute upwards, 2-3 mm wide, up to 12 cm long but usually much shorter.
339
scaberulous below, scabrid to pubescent above. Panicle usually exserted, more or
less obovate, dense, sub-fastigiate, up to 8 cm long; branches scabrid, bearded in
the axils; pedicels short, hairy. Spikelets clustered, often brownish at the base.
Glumes unequal, 3-nerved, pubescent, with glabrescent tips, the lower 10-14 mm long,
the upper 14-18 mm long. Lemma smooth, 5-5-5 mm long including the callus;
central awn plumose, 3-3-5 cm long; lateral awns naked, fine, 1-2 cm long.
Fig. 110. — S. fasiigiala: cross section of the leaf-blade {cie Winter 3573).
0- 1 rn m
Fig. 111. — S. fastigiata: abaxial epidermis of the leaf-blade.
340
Anatomy (Fig. 110, 111 and 160: 12)
Leaf-blade expanded, flat; abaxial surface undulate with a few short hairs opposite or flanking
the bundles; adaxial surface with shallow grooves between the bundles and densely covered with rather
short, unicellular hairs; keel absent; margins obtusely rounded. Abaxial epidermis cells fairly thick-
walled ; stomatal zones with 2-3 rows of stomata, and 4 to 8 rows of ripple-walled cells with strongly
undulate walls; many uni-cellular retrorse barbs present; silicified cell zones with 1-3 files of silicified
cells, and 3-6 files of ripple-walled cells with distinct but fine undulations; silicified cells sub-circular,
solitary or accompanied by oblong-rectangular, square, or transversely oblong, thin-walled suberLzed
cells; many broadly oblong papilla-like unicellular retrorse barbs present; no two-celled hairs observed.
Vascular bundle units all approximately of the same size, sub-circular in outline; first order bundle
units 5; 2 second order units near the margin; third order bundle units alternating in pairs with the
first order bundles, and singly with the marginal second order bundles; midrib similar to the other
first order bundles. Bundle sheaths: outer of large sub-circular, thin-walled cells containing chloroplasts ;
inner of smaller less regular cells with slightly thickened walls. Chlorenchyma a single layer of tabular
cells radially arranged around the bundles. Stereome strands very weakly developed both ab- and
adaxially, usually consisting of one or two layers of small fibres not in contact with the bundles. Motor
cells present alternating with all the bundles, groups stretching across the whole width of the leaf near
the mibrib, and gradually diminishing in size to less than half the width near the margin.
Cape. — Kenhardt ; Leistner 2360.
South West Africa. — Warmbad: Acocfo 18808; de Winter 3513. Bethanien: Walter
2161. Without locality: Boss s.n. (Tvl. Mus. 36148); Volk s.n. (Natal Herb. 33280).
Distribution. — (See Fig. 123.)
Fig. 112. — S. geminifolia cross section of the leaf-blade (Marloth 12415).
341
O • 1 mm
Fig. 113. — S. geminifoUa: abaxial epidermis of the leaf-blade.
Closely related to 5. geminifoUa from which it differs in the internodes being more
or less uniform in length and by the somewhat smaller spikelets. It is usually a larger
plant than S. geminifoUa and occurs further inland, where the rainfall is slightly higher.
S. geminifoUa has the upper two leaf-sheaths subtending the inflorescence approximate,
and even more strongly reduced leaf-blades than S. fastigiata. S. fastigiala occurs
mainly in the very dry desert near the coast in the Luderitz distiict of South West
Africa and in the Richtersveld of South Africa.
Anatomically the species differ only in minor details. In spite of the fact that
Theron (1936; p. 26) placed this species vei'y near to S. geminifoUa, he compares
it with Aristida spectabilis, a completely unrelated species and a true Aristida.
14. S. geminifoUa Nees, FI. Afr. Austr. 173 (1841)
Aristida geminifoUa (Nees) TTrin. & Rupr. Gram. Stip. 169 (1842).
Perennial, small suffrutex, or sub-suffrutex up to 25 cm high. Culms erect or
ascending, glabrous or hairy below the panicle, few-noded; internodes very unequal,
the lowermost short, the intermediate nodes much longer and exserted, up to 12 cm
long, the upper short to very short, 5-20 mm long; nodes bearded or glabrescent.
Leaf-sheaths short, glabrous, with woolly or ciliate margins, the lower tight, the upper
laxer, often almost spathe-like; auricles bearded. Leaf-blades very short, often rudi-
mentary, subulate or flat, involute, very rigid, spreading, smooth below, pubescent
on upper surface. Panicle erect or nodding, short, ovate to oblong, up to 4 cm long
and 2-5 cm wide; rhachis glabrous or hairy; pedicels very short, hairy. Spikelets
densely clustered, brownish, often tinged with purple. Glumes unequal, linear-lanceolate,
acuminate and awned, hairy at least upwards, the lower 8-12 mm long, 3-nerved,
the upper 10-12 mm long, 1-3-nerved. Lemma including the callus 4-6 mm long,
smooth; central column 1-2 mm long, shghtly twisted; central awn 2-3 cm long
plumose; lateral awns fine, glabrous, 1 - 5 cm long. Caryopsis narrowly elliptic, small,
2 mm long; embryo i to f the length of the grain, hilum as long as the grain.
342
Anatomy (Fig. 112, 113 and 160: 13)
Leaf-blade occasionally horseshoe-shaped but usually expanded in transverse section; keel absent;
abaxial and adaxial surfaces undulate giving the section a moniliform appearance; adaxial surface
furnished with long unicellular hairs; abaxial surface with a few papilla-like retrorse barbs; margins
obtuse. Abaxial epidermis: cells thin-walled; stomatal zones with 2-A rows of stomata and 4—7 rows
of long ripple-walled cells with distinctly undulate walls interspersed with single, or occasionally paired,
short elements, which are transversely elliptic to narrowly elliptic in shape; sparsely furnished with
papilla-like retrorse hairs; silicified cell zones with 1^ files of silicified cells opposite the first order
bundles, and 1-2 files opposite the third and second order bundles; long elements with the walls finely
undulate; silicified cells irregularly scattered in the files, solitary or accompanied by thin-walled
suberized cells; suberized cells similar in shape to those in the stomatal zones, solitary or in pairs;
unicellular papilla-like large retrorse barbs, present usually accompanied by 1 or 2 suberized cells;
two-celled hairs present in the stomatal zone. Vascular bundle units: first order units 3, the two marginal
bundles on each margin usually of the second order; 2 third order bundles on either side of the midrib
and alternating with the other bundles; first and third order units subequal, sub-circular in outline;
midrib similar to the other first order bundle units. Bundle sheaths: outer of large thin-walled cells
containing chloroplasts, inner of more irregular smaller cells with slightly thickened walls; both sheaths
in all bundles complete. Chlorenchyma consisting of a single row of tabular cells radially arranged
around the bundles. Stereome strands present abaxially opposite all bundles, and usually in contact
with these but not intterupting the bundle sheaths; small weakly developed, adaxial groups present;
all strands consisting of small-celled fibres. Motor cell groups alternating with all but the three marginal
bundles; groups elongate, forming bands stretching across the whole width of the leaf.
Cape. — Namaqualand : Drege s.n. (Kunkunnuwub); Marloth 12415.
South West Africa. — Luderitz: Dinter 3829; 4099; 6365; Schaefer (Marloth
Herb. 12987); Kinges 2573; 2699; de Winter & Giess 6107 ; Giess & van Vuuren 670.
Distribution. — (See Fig. 114.)
Fig. 1 14.— Distribution of 5. geminifolia; S. garubensis.
A distinct species unlikely to be confused with any of the others except for the
closely allied S. fastigiata (see page 341).
Anatomically it forms part of the small group of species wliich have more or less
expanded, flat leaf-blades in contrast to the majority of species of Stipagrostis where
the blades are terete, grooved, rolled or folded (see Fig. 160).
The blades have very little sclerenchyma, and in this respect resemble those of
the annual species. The reduced blades are deciduous and in the dry season only
the bare culms remain. The culms contain chloroplasts and when dormant are straw-
coloured but become green soon after rain, even before the new leaves have appeared.
Although this has been observed in only one other species viz. S. uniphimis, it probably
applies to all the suffrutescent species as well as some of the remaining species.
343
Theron (1936, p. 26) places this species between Aristida bipartita and A. spectabilis,
species not related to S. geminifolia and only very superficially similar in the general
shape of the cioss-section of the leaf-blade. In error he designated the specific epithet
as “ geminiflora
15. S. garubensis (Pilger) de Winter in Kirkia 3: 134 (1963)
Aristida garubensis Pilger in Engl. Bot. Jahrb. 118: 343 (1912).
AD
Fig. 115. — S, garubensis: cross section of the leaf-blade {Hinges 2289).
Perennial, bushy suffrutex, the basal part of the culms almost woody, up to 60 cm
high. Flowering culms arising from the woody basal part, thin and elegant, glabrous,
2-3-noded; nodes glabrous. Lower leaf-sheaths scale-like, the upper longer, and
bearing fully developed blades, glabrous. Leaf-blades sub-terete, setaceous and wiry,
up to 12 cm long, glabrous on the latter surface, and finely grooved, hirtellous on
the upper surface. Panicle very lax but often contracted and rather narrow, up to
20 cm long; branches solitary or 2-3-nate. Spikelets pallid, purplish at the base.
Glumes lanceolate, glabrous; the lower up to 12 mm long, sub-obtuse, slightly bidentate;
the upper 12-5-15 mm long, subacute. Lemma smooth, including the callus 5-5 mm
long; column 8-10 mm long, twisted; central awn up to 40 mm long, plumose to
the base or glabrous in the lower third; lateral awns glabrous, fine, up to 15 mm long.
Anatomy (Fig. 115, 116 and 160: 14)
Leaf-blade horseshoe-shaped in transverse section; keel not developed; adaxial surface with
the bundles strongly projecting and with short sharp unicellular hairs adjacent to the stereome strands;
abaxial surface with deep narrow grooves between all the bundles with short retrorse barbs projecting
diagonally across the grooves; margins obtuse. Abaxial epidermis: cells with extremely thickened
tangential walls especially in the grooves, radial walls much less thickened hence epidermis appearing
344
thin-walled in surface view. Stoinatal zones with 2 to 3 rows of stomata and 4-6 rows of long ripple-
walled cells; walls of long cells very strongly undulate with the proximal ends of the undulations often
constricted and showing up as circles in the surface view of the epidermis; a few short elements,
occasionally paired, present; unicellular hairs rare; bicellular linear hairs rare but present; silicified
cell zones with 3-5 files of short elements usually alternating with 1-2 files of long ripple-walled cells
adjacent to the stereome strands of the third order bundles, and 5-7 files adjacent to the first order
bundles ; walls of long elements finely and compactly undulate. Vascular bundle units: first
order units 5, broadly oblong to obovate in outline; third order units more or less ovate in outline,
2 on each side of the midrib and alternating with the other bundles; midrib similar to the other first
order bundles. Bundle sheaths: outer of large parenchymatous cells filled with chloroplasts, interrupted
ad- and abaxially in the first order bundles, and abaxially only in the third order bundles; inner sheath
of smaller thick-walled cells. Stereome strands consisting of small fibres throughout, very strongly
developed abaxially and projecting into all bundles; adaxially projecting only into the first order
bundles, and present as small distinct groups of cells opposite the third order bundles: Motor cells
alternating with all except the marginal bundles, forming Y-shaped groups of thin-walled cells occupying
about half of the leaf thickness [see also Theron (1936; p. 8)].
0-1 m ."■
Fig. 116. — S. gantbensis: aba dal epidermis of the leaf-blade.
South West Africa. — Luderitz: Acocks 15647; Kinges 22S9 ■, 2349; Merxmuel/er
& Giess 2369; Giess & van Vuuren 853; Giess 2329.
Distribution. — (See Fig. 114.)
S. garuhensis is a distinct species without close affinities. It is easily recognized
by its suffrutescent habit; the basal part is robust and almost woody whereas the
flowering culms are very slender and graceful, the inflorescence is lax and open and the
leaves are setaceous and wiry. It has so far been recorded only from the very dry
coastal Namib in the vicinity of Luederitzbucht in South West Africa.
16. S. dregeana Nees, FI. Afr. Austr. 172 (1841)
Aristida dregeana (Nees) Trin. & Rupr., Gram. Stip. 169 (1842).
Perennial, up to 30 cm high, laxly to densely caespitose, branched at ^he base.
Culms slender, erect or geniculate, simple, 1-3-noded; nodes and internodes glabrous.
Leaf-sheaths scale-like at the base of the culms, well developed upwards, scaberulous.
Leaf-blades setaceous, convolute, up to 13-5 cm long, glabrous beneath, hirtellous on
the upper surface. Panicle erect or somewhat nodding; branches mostly bi-nate,
branches and branchlets capillary and somewhat flexuous. Spikelets usually suffused
with purple. Glumes linear-lanceolate, acute, sub-equal, 3-nerved; the lower 11-13
mm long, minutely scaberulous on the keel, when young often sparsely hairy; the
upper 1 1-1 3 mm long, obtuse, emarginate to bifid and mucronulate, minutely scaberulous
345
upwards, otherwise glabrous. Lemma glabrous, smooth, including the callus about
4 mm long; column 4 mm long, twisted, glabrous; central awn 3-4 cm long, plumose
in the upper half; lateral awns naked or almost so, up to 19 mm long; callus acute,
densely hairy.
Fig. 117. — S. dregeana: cross section of the leaf-blade {Kinges 2572).
0‘1 mm
Fig. 118 -S. dregeana: abaxial epidermis of the leaf-blade.
346
Anatomy (Fig. 117, 118 and 160: 15)
Leaf-blade horseshoe-shaped in transverse section; keel not developed; adaxial surface with the
bundles strongly projecting and covered with rather short, sharp unicellular hairs opposite the stereome
strands; abaxial surface undulate, with very few hairs; margins obtuse. Abaxial epidermis: cells
with strongly thickened walls; stomatal zones with 1 to 2 rows of stomata and 3 to 4 rows of very long
strongly ripple- walled cells, often alternating with short elements; short elements transversely oblong
in shape; undulations of the ripple- walled cells often very strongly developed, the ends of the undulations
constricted and forming papilla-like structures which show up as circles in surface view, these structures
are especially well developed near the stomata and form four papillae overhanging and almost completely
obscuring each stoma; silicified cell zones with 4-8 files of short elements opposite the third order
bundles and 9-12 files opposite the first order bundles; silicified cells square to sub-circular in outline,
solitary or accompanied by a thin-walled, transversely-oblong, square to oblong suberized cell with
often somewhat undulate walls; long ripple- walled cells finely and compactly undulate; unicellular,
sharp, rather short hairs present, flanking some of the silicified zones and projecting inwards over the
stomatal zones; no bicellular trichomes observed but probably present. Vascular bundle units well
developed; first order units oblong-rounded; third order units much smaller, more or less ovate in
outline; midrib similar to the other first order units. Bundle sheaths: outer of large thin-walled cells
containing chloroplasts, interrupted abaxially by the stereome strands in all the bundles except
occasionally the marginal third order bundles, interrupted adaxially only in the first order bundles;
inner sheath of smaller cells with thickened walls, not interrupted. Cfilorenchyma a single row of tabular
cells arranged radially around the bundles. Stereome strands well developed and projecting into all
the bundles abaxially, adaxially well developed opposite the first order bundles and projecting into them,
present as small distinct groups opposite most of the third order bundles. Motor cells alternating
with all except the marginal bundles, forming Y-shaped groups of thin-walled cells occupying half
or more of the leaf thickness.
Cape. — -Namaqualand : Drege s.n. (Mouth of Orange River).
South West Africa. — Luderitz: Dinter 6344; Giess & van Vuuren 682; Kinges
2572; Marloth 4664.
Distribution. — So far collected only in the coastal area of South West Africa from
the Orange River to Luderitz.
A species probably best placed near S', zeyheri and S. gariibensis. It differs from
S. zeyheri in having only the central awn plumose and from S. garubensis in lacking
a suffrutescent much-branched base.
For anatomical differences compare the Figures for these species (Fig. 115, 116,
117-122, 160). Theron (1936, p. 20) compares this species with Aristida burkei (reduced
by Schweickerdt to a var. of A. diffusa) from which it, however, differs by the single
large-celled chlorophyll-bearing sheath, and the glabrous awns. Theron’s description
of the anatomy agrees in general with the present author's observations.
17. S. zeyheri (Nees) de Winter in Kirkia 3; 136 (1963)
Perennial, compactly caespitose. Culms simple or rarely branched from the base,
up to 75 cm high, wiry, glabrous. Leaf-sheaths glabrous or with a few scattered hairs;
auricles glabrous or bearded. Leaf-blades filiform, folded, subterete, up to 40 cm
long; the lower surface smooth, the upper scaberulous to hispid. Panicle erect or
nodding, narrow, contracted or open to divaricate. Spikelets pale brown to purplish.
Glumes glabrous or hairy, unequal to sub-equal, 3-nerved; the lower 10-19 mm long;
the upper 11-20 mm long. Lemma including the callus 5-8-5 mm long, smooth or
finely pustulate; column 1-14 mm long, glabrous to hairy; the central awn 16-40 mm
long, plumose; the lateral plumose, 13-35 mm long; callus 2-5 mm long, acuminate
densely hairy. Caryopsis very narrowly elliptic, about 5 mm long; hilum as long as
the grain.
Anatomy (Fig. 119, 120, 121, 122 and 160: 16a-d)
Leaf-blade V-shaped or sub-circular with a deep adaxial median groove; abaxial surface slightly
undulate to almost smooth; adaxial surface on either side of the median groove undulate with a cleft
on each side of the midrib; or deeply grooved between the bundle units, densely covered with short
347
papilla-like unicellular hairs; margins subacute or obtuse. Abaxial epidermis: stomatal zones with
1-2 rows of stomata and 3-4 rows of long, strongly ripple- walled cells with scattered, intermixed paired
short elements; or stomata very few and whole epidermis of short elements alternating with long
ripple- walled cells; silicified cell zones consisting of files of silicified cells alternating with files of
ripple-walled cells with finely but densely undulate walls. Vascular bundle units: first order units 5,
broadly oblong in outline; third order units ovate in outline, flanked by motor cell groups or situated
more or less below them, alternating with first order units or with two bundles flanking the midrib,
sometimes absent between the 2 marginal first order bundles; midrib similar to the other first order
bundles. Bundle sheaths: outer of large rather regular cells, only interrupted adaxially by the stereome
strands in all but occasionally some of the third order bundles; inner sheaths of irregular thick-walled
cells only occasionally interrupted by stereome abaxially in the first order bundles. Chlorenchyma
a singly row of tabular cells radially arranged around the bundles. Stereome well developed opposite
all bundles including the margin, or forming a continuous abaxial layer; adaxial stereome poorly
developed and consisting of large cells with only slightly thickened walls, or rf well developed, consisting
of typical fibres. Motor cells present on either side of the midrib, absent elsewhere, or alternating
with all bundles.
Glumes glabrous;
Column hairy:
Glumes usually purplish, leaves usually not overtopping the inflorescence; winter rainfall
areas of tire Cape subsp. Zeyheri
Glumes usually straw-coloured; leaves usually overtopping the inflorescence, east coast of
the Cape subsp. barbata
Column glabrous subsp. macropus
Glumes hairy subsp. sericans
(a) subsp. zeyheri.
Arthratherum zeyheri Nees, FI. Afr. Austr. 177 (1841). Aristida zeyheri (Nees) Steud.,
Nom. Bot. 2; 132 (1841). A. capensis var. canescens Trin. & Rupr., Gram. Stip. 178
(1842). A. capensis var. zeyheri (Nees) Walp. in Ann. Bot. 3: 751 (1852).
Perennial, forming an erect tuft from a short knotty rhizome. Panicle lax and
open, usually exceeding the leaves. Spikelets often purplish. Column of the awns
hairy, 2-5-12 mm long; plumes of the awns greyish to silvery.
Anatomy (Fig. 160: 16)
Leaf-blade subterete with a deep adaxial groove. Stereome strands usually fused to form a
continuous abaxial sheath of fibres. Deep grooves present between the bundle units on the adaxial
surface. Motor cell groups present alternating with all the bundles.
Cape. — Cape Peninsula: Hitchcock 240S6 & 24083; Adamson 3103; Ecklon & Zeyher
s.n.; Wolly Dod 1969. Paarl; Schlechter 9105. Caledon: Schlechter 9843; Parker
4471. Worcester: van Breda 3. Tulbagh: Ecklon & Zeyher s.n.; Malmesbury:
Acocks 14103; Barenbrug 6. Stellenbosch: Smuts s.n.; Sandwith 153. Clanwilliam;
Pillans 8740.
Distribution. — (See Fig. 123.)
This subsp. is not well differentiated from subsp. barbata, which in its typical form
however, looks entirely different. The lax spreading inflorescences, longer column of
the awns and usually purplish spikelets characterize the subspecies zeyheri which is
limited to the winter rainfall area of the Cape. Anatomically the leaf-blades of subsp.
zeyheri and subsp. barbata are practically identical.
{b) subsp. barbata (Stapf) de Winter in Kirkia 3: 136 (1963).
Aristida capensis var. barbata Stapf in F.C. 7: 565 (1899).
Perennial, forming large erect densely leafy tufts. Panicle usually rather dense
and contracted, overtopped by the setaceous, often curling, leaves. Spikelets usually
pallid. Column of the awns hairy, 2-4 mm long; plumes of the awns greyish or tawny.
348
Anatomy as for subsp. zeyheri (Fig. 160: 16a)
Cape. — Uitenhage: Zeyher 4501; Ecklon & Zeyher 502. Alexandria; Burtt-Davy
14196; Archibald 5506. Bathurst: Hutton 2631; Tyson BH. 14877. East London:
Codd 91AQ. Kentani: Flanagan 1782; Pegler 265; Theron 1188; Acocks 13588.
Natal. — Mtunzini: Guy & Ward 63. Lower Tugela: Edwards 1723.
This subspecies has a well-defined geographic distribution along the coastal areas
of the eastern Cape and northern Natal (see Fig. 123). Recognized by the very well
developed often curling, setaceous leaf-blades which usually overtop the rather narrow
straw-coloured inflorescences. It usually inhabits coastal dunes and forms very robust
leafy tufts.
(c) subsp. macropus (Nees) de Winter in Kirkia 3: 136 (1963).
A. capensis Thunb., Prodr. 19 (1794) non Stipagrostis capensis Nees. Aristida capensis
var. macropus Nees, FI. Afr. Austr. 176 (1841). A. capensis var. fulviberbis Tiin. &
Rupr., Gram. Stip. 179 (1842). A. capensis var. genuina Henrard, Crit. Rev. Aristida
1: 77 (1926).
Chaetaria capensis (Thunb.) Beauv., Agrost. 30 (1812)
Fig. 119. — S. zeyheri subsp. macropus: cross section of the leaf-blade (Bolus 12910).
349
0-1 mm
Fig. 120. — S. zeyJieri subsp. niacropus: abaxial epidermis of the leaf-blade.
Perennial, forming erect tufts from a short knotty rhizome. Leaves not very dense,
usually not exceeding the inflorescence; blades setaceous. Panicle lax and effuse,
the branches occasionally somewhat contracted. Spike lets brown to purplish. Glumes
glabrous. Lemma articulated; column of the awns glabrous, variable in length, 6-12
mm long, shorter than or exceeding the glumes; plumes of the awns dirty white to
distinctly yellow.
Anatomy (Fig. 119, 120 and 160: 16b)
Leaf-blade sub-terete with a deep adaxial groove. Stereome strands well developed but not fused
into a continuous layer. No deep grooves present between the bundle units. Motor cell groups well
developed only adjacent to the midrib.
Cape. — Malmesbury: van Rensburg 144. Worcester: Fairall 333. Piquetberg: Haf-
strom & Acocks 2. Clanwilliam: Liebenberg 4302; Schlechter 8213. Vanrhynsdorp :
Rood 789; Acocks 14747; 19503. Namaqualand: Drege s.n. (Ebenezer); Ester-
huyzen 1329 & 1406; Acocks 16467.
Distribution. — (See Fig. 123.)
This subspecies occurs mainly in the dry western Cape. It is distinguished from
the other subspecies by the long glabrous column and the usually yellowish plumes
of the awns. Its distribution partially overlaps with subspecies zeylieri. Insufficient
information is available to decide whether it is ecologically separated from subspecies
zeyheri in these areas, but this seems probable. Anatomically it differs rather strikingly
from subspecies zeyheri by the interrupted stereome, and from the subspecies barbata
by the absence of deep grooves between the bundle units adaxially, as well as in the
presence of only two well developed motor cell groups flanking the midrib (cf. Fig.
160: 16, 16a and 16b).
{d) subsp. sericans {Hack, apud Schinz) de Winter in Kirkia 3: 136 (1963).
Aristida sericans Hack, apud Schinz in Bull. Herb. Boiss 3: 381 (1895). A. capensis
var. dieterleniana Schweick. in Kew Bull. 1939: 613 (1939).
350
Perennial, forming erect tufts, densely leafy at the base. Leaves shorter than the
exserted inflorescence; blades filiform, erect. Panicle erect stiff spike-like, or narrow
and interrupted. Spikelets palUd to purple. Glumes densely to sparsely villous. Lemma
distinctly articulated, or articulation poorly developed or even absent; column of the
awns 2-7 mm long, hairy; plumes of the awns dirty white to pinkish.
Anatomy (Fig. 121, 122 and 160: 16c and d)
Leaf-blade sub-terete to V-shaped in cross-section. Stereome strands separate, or more or less
united to form a continuous layer abaxially. Deep grooves present between the bundle units adaxially.
Third order bundle units two on either side of the midrib and alternating with the first order bundles,
or absent between the marginal pairs of first order bundles. Motor cell groups present between all
bundles except the marginal first and third order bundles.
Fig. 121. — S. zeyheri subsp. sericans: cross section of the leaf-blade (Story N.H. 24574).
I- 1
0-1 rnm
Fig. 122. — S. zeyheri subsp. sericans: abaxial epidermis of the leaf-blade.
351
Cape. — Steynsburg: Acocks%610. Aliwal North: y4cocA'5 18853. Lady Grey: Acocks
13825. Mount Currie: Acocks 13282.
Basutoland. — ^Maluti Mtn.: Staples 146. Quthing: Dieterlen 1205. Mamathes:
Guillarmod 914.
Orange Free State. — ■Tabanchu Celliers 11. Ficksburg: Dyer 4742. Bethlehem:
Story 1947; Acocks 12341. Flarrismith: Acocks 11206; Acocks 13559.
Transvaal. — 'Vereeniging: s.n. (N.Fl. 24575); F/orj (N.H. 24574); Story 1022.
Krugeisdorp: Mogg 23270. Pretoria: Pole Evans 388. Carolina: Codd 1718;
Acocks 12925.
Distribution. — (See Fig. 123.)
Fig. 123. — Distribution of S. zeyheri: ®subsp. zeyheri; Qsubsp. macropus; 0 subsp. barbata;
©subsp. sericans; MS’, fastigiata.
This sub-species may be distinguished from all the others by the hairy glumes and
the elongate contracted or spike-like panicle. It is the only repiesentative of the genus
in the highveld grasslands of the Cape, the Orange Free State and the Transvaal.
In the past this subspecies has been referred to the section Chaetaria (Aristida)
since an articulation is lacking. The presence of plumose awns, as well as the anatomical
structure of the leaf-blade offers, however, conclusive evidence that it occupied an
anomalous position in the section of true Aristida. It should be accommodated in
Stipagrostis, in spite of the absence of an articulation. The specimens cited are not
quite homogeneous and a gradual transition from the non-articulated A. sericans to
the articulated A. capensis var. dieterleniana can be traced. No other differences could
be detected and since in addition the distribution of these taxa co-incides there seems
to be little doubt that they belong to one subspecies.
The taxa designated by Henrard and Schweickerdt as varieties of Aristida capensis
have been reclassified in an attempt to create more natural units. This has been made
possible by the laiger range of specimens now available. Since more oi less distinct
distributional ranges could be established for the taxa now recognized it was decided
to establish subspecies rather than varieties.
As reported by Henrard and by Schweickerdt, A. ciliaia occasionally hybridizes
with S. zeyheri var. macropus (.4. capensis var. macropus). This hybrid was assigned
the name A. schlechteri by Henrard. (The present author has, however, not adopted
this name as many other hybrids have since come to notice, and to assign specific
names to all these would cause confusion). Only one additional record of this hybrid
has come to the author’s notice since Schlechter’s original gathering, namely Acocks
14817. Thus the hybrid may be regarded as being one of relatively little importance.
352
The anatomy of the leaf-blade supports Henrard’s view that his Aristida schlechteri
is of hybrid origin, the general features being intermediate between those of S. zeyheri
and S. ciliata.
Fig. 124.-5'. gonatostacliys: cross ssction of the leaf-blade {Dinter 1022).
01 mm
Fig. 125. — 5. gjitutoslacliys: abaxial epidermis of the leaf-blade.
353
18. S. gonatostachys (Pilger) de Winter in Kirkia 3: 134 (1963)
Aristida gonatostachys Pilgei in Engl. Bot. Jahrb. 48: 343 (1912).
Perennial densely caespitose dwarf plants, 2-10 cm high; Culms exserted beyond
the small tussocks, naked, geniculate, 1-noded, thickened towards the node, hairy
just below the upper node or glabrescent. Leaf-sheaths reduced to scales near the base,
long upwards, glabrous or scantily villous on the margin, the upper sheath leafless
or with a short blade. Leaf-blades up to 4 cm long, involute, obtuse, deeply grooved
beneath, densely covered with long crisped hairs arising in the grooves, scaberulous
on the upper surface. Panicle shortly peduncled, partially enveloped by the uppermost
sheath, contracted, narrow, bearing 3-6 spikelets; rhachis pubescent. Glumes subequal,
or the lower slightly exceeding the upper; the lower 9-10 mm long, minutely scaberulous;
the upper almost glabrous. Lemma glabrous, purplish or pallid, including the callus
about 4 mm long; column twisted, glabrous, up to 4 mm long; central awn up to
16 cm long, densely plumose in the upper part; lateral awns naked, up to 9 mm long.
Anatomy (Fig. 124, 125 and 160: 17)
Leaf-blade narrowly horseshoe-shaped in transverse section; abaxial surface deeply grooved
between the bundles; adaxial surface undulate, covered with long, sharp unicellular hairs; margins
acute, often terminated by papilla-like hairs. Abaxial epidermis: cells with extremely thickened walls;
stomatal zones often completely hidden by very long, wavy unicellular hairs lying longitudinally appressed
in the grooves; silicified cell zones of 3-5 hies containing silicifled cells, alternating with narrower
ripple-walled cells; silicifled cells square or sub-circular, usually accompanied by thin-walled suberized
cells, the latter square, broadly oblong or transversely-oblong in shape; many papilla-like broad-based
retrorse barbs present, especially along the margins of the zones; two-celled hairs not observed but
probably present. Vascular bundle units: first order units 3, broadly ovate-oblong; third order units
alternating with those of the first order, broadly oblong to oblong-circular; midrib similar to the other
first order bundles. Bundle sheaths: outer of large thin-walled parenchyma cells containing chloroplasts,
interrupted only in the first order bundles by the stereome strands; inner sheaths of much smaller
thick-walled cells lacking chloroplasts; complete in all bundles. Chlorenchyma a single row of tabular
cells arranged radially around the bundles. Stereome well developed abaxially opposite all bundles
but in contact only with the first order bundles; adaxially poorly developed, present in small groups
opposite all the first order bundles and the marginal third order bundles; consisting of small-celled
fibres throughout. Motor cells present only opposite the third order bundles flanking the mibrib,
consisting of large groups of thin-walled parenchyma cells.
South West Africa. — Luderitz: Range 1177; Giess I'i'iA', Kinges llhl', de Winter
& Hardy 7907; 7888. Swakopmund; de Winter & Hardy 8098.
Distribution. — ^So far only recorded from the Luderitz and Swakopmund districts
in the dry central Namib Desert. It occurs on wind-blown sand between rocks on
mountainsides, as well as in depressions on the gravelly desert flats where water collects
during showers.
This species is closely allied to S. obtusa with which it may be confused. Dis-
tinguishing features are the very scabrid leaf-sheaths and blades; the bristles on the
swollen part of the upper internode ; the culm which is geniculate at the upper internode
so that the inflorescence and sheath form an angle with the rest of the culm; the, on
the whole, shorter central awn (±1'5 cm); a longer callus and the narrower more
contracted, subsecund inflorescence which is usually not fully exserted from the rather
broad uppermost sheath. S. gonatostachys forms a connecting link between S. obtusa
and S. hermannii, species which on a superficial examination would not appear to
be related. The latter two species both occur in the Namib in South West Africa
S. hermannii is mainly found in the very dry central and coastal parts of the desert
while S. obtusa has a wide distribution but also inhabits marginal desert areas,
penetrating deeper into the desert in wetter seasons. In this way the two species
occasionally come into contact. A careful examination of the material of S. gonatos-
tachys has led the author to consider the possibility that this species originated as a
result of hybridization between S. obtusa and S. hermannii. S. hermannii has very
scabrid sheaths and leaf-blades, culms which are geniculate at the uppermost node,
spicate inflorescences sheathed at the base, and a well developed callus. It can
354
therefore be seen that most of the characters by which S. gonatostachys is distinguished
from S. obtusa are possessed by S. hermannii. S. gonatostachys can however be regarded
as having become a well stabilized reasonably distinct species.
Fig. 126. — S. Iviipes: cross section of the leaf-blade {de Winter & Giess 6133).
Fig. 127. — S. lanipes: abaxial epidermis of the leaf-blade.
355
The Swakopmund specimens {de Winter & Hardy 8098) have very hairy glumes
but do not differ in other respects from the Luderitz specimens, and probably only
represent a hairy form.
19. S. lanipes {Mez) de Winter in Kirkia 3: 135 (1963)
Aristida lanipes Mez in Fedde Rep. 17: 153 (1921).
Practically the only difference between this species and S. obtusa is the woolly
leaf-sheaths of the former. Only two collections of S', lanipes are known, viz. the
type gathering, Range 1822, apparently a rather depauperate specimen and de Winter
& Giess 6133 both from the Luderitz district in South West Africa. The latter specimen
has inflorescences up to 21 cm long, strongly resembling those of S. obtusa, but more
depauperate inflorescences on the same plants reach a length of only 4 cm and bear
5-7 spikelets, thus greatly resembling the type. In de Winter & Giess 6133 the lower
glume exceeds the upper in length whereas in the type the lower glume is subequal
to or shorter than the upper. This character, however, is known to vary in other species
and, therefore, may be of secondary importance.
Anatomically S. lanipes agrees very closely with S. obtusa (see Fig. 126, 128 and 160: 18 and 19).
From the above it is evident that the status of S. lanipes as a species distinct from
S. obtusa is somewhat doubtful. Until a fuller range of specimens becomes available
a final decision must be withheld.
20. S. obtusa {Del.) Nees in Linnaea 7: 293 (1832)
Aristida obtusa Del., FI. Aegypt 31 (1813); A. bifida J. Karl in Mitt. Bot. Staatssamml.
Muenchen 3: 86 (1951).
Arthratherum obtusum (Del.) Nees, FI. Afr. Austr. 179 (1841).
Stipagrostis capensis Nees in Linnaea 7: 291 (1832).
Perennial, compactly caespitose due to the numerous innovations bearing dense
tufts of leaves. Culms from a few cm up to 60 cm high, slender, 1-noded, inteinodes
and nodes glabrous. Leaf-sheiths glabrous or rarely villous, the basal occasionally
woolly on the margin; auricles bearded. Leaf-blades variable in length, 1-5-25 cm
long, setaceous, convolute, subteiete, occasionally sparsely villous or scaberulous,
hirtellous on the upper surface, apex obtuse. Panicle very narrow, erect, contracted,
but rather loose and interrupted at the base, 2-5-20 cm long, branches solitary, axils
glabrous. Spikelets pallid or flushed with purple at the base. Glumes subequal,
dorsally scaberulous with the very fine spines in numerous, more or less parallel longi-
tudinal rows; the lower obtuse, 3-nerved, usually exceeding the upper in length,
8-5-12 mm long; the upper subacute, 1-3-nerved, 8-11 mm long. Lemma smooth,
about 3 mm long including the callus; column variable in length 4-10 mm long,
stiaight or twisted; central awn scaberulous in the lower half, plumose upwards,
2-3-2 cm long; lateral awns naked, very fine 0-75-1-75 cm long; callus acute or
occasionally minutely or obliquely bifid. Caryopsis narrowly elliptic-oblong, about
2 mm long; embryo about f the length of the grain; hilum as long as the grain.
Anatomy (Fig. 128 and 160: 19)
Leaf-blade sub-circular in transverse section with a smooth-sided deep adaxial median groove
and long, thin unicellular hairs on the epidermis opposite the stereome strands in the adaxial groove;
abaxial surface with deep narrow grooves between all the bundles; margins obliquely obtuse. Abaxial
epidermis, when scraped, breaking up into narrow strips due to the deep grooves, in which the stomatal
zones are situated. The scrapes show only the silicified cell zones which consist of up to 8 files containing
only silicified cells, alternating with files of long ripple-walled cells; silicified cells sub-circular, solitary
or accompanied by thin-walled suberized cells which vary in shape from transversely narrowly-oblong
to oblong with somewhat undulate walls; broad-based papilla-like retrorse barbs are present in these
zones. Vascular bundle units: first order units 3, square or oblong in outline; third order units similar
to those of the first order but smaller, two on either side of the mibrib and one in each margin; midrib
356
similar to the other first order bundles. Bundle sheaths: the outer of large thin-walled cells, interrupted
by the stereome strands ab- and adaxially in the first order bundles, and abaxially in the third order
bundles flanking the midrib, intact in all the other bundles ; inner sheath of smaller cells with thickened
walls interrupted abaxially only in the first order bundles, otherwise intact. Stereome strands well
developed abaxially opposite all bundles forming narrow bands of approximately constant width;
present adaxially opposite first order bundles, and in very small groups opposite the marginal third
order bundles. Motor cells present only in two very large many-celled groups flanking the midrib.
Cape. — Prince Albert; Marloth 2055. Kimberley: Leistner 944. Kenhardt; Leistner
2466. Kuruman: Pole Evans 2089. Hay: Esterlmyzen 2211. Gordonia : Story 5512.
Namaqualand: Zeyher 1815.
Fig. 128. — S. obtiisa: cross section of the leaf-blade (Krapohl PRE 21862).
357
Orange Free State. — -Fauresmith : Pole Evans 1572.
South West Africa. — Luderitz: Giess 2040. Keetmanshoop : Dinter 4952. Gibeon:
Codd 5878. Rehoboth; Strey 2103. Windhoek: de Winter & Giess 7139. Omaruru:
Liebenberg 5031. Swakopmund : de Winter 3204.
Distribution. — ^(See Fig. 129.)
This widespread and common species is one of a small group of four closely
related species, i.e. S. obtusa, S. lanipes, S. gonatostachys and S. dinteri. The relationship
of these species is referred to on page 353 in the general discussion of the affinities
of the species of Stipagrostis. For specific differences the reader is referred to the
discussion under the other species and to the keys.
Aristida bifida J. Karl, described in 1951, is regarded to be synonymous with
S. obtusa, as no differences other than the bifid callus could be detected. This character
is not constant: the callus being acute in the majority of specimens examined, very
obliquely bifid in some specimens, and distinctly bifid in others. Specimens with
bifid and acute calli have been collected even in the same population. S. obtusa is
the only species of Stipagrostis for which a bifid callus has been recorded.
One record of a putative hybrid between S. obtusa and S. dregeana was collected
near Aus in South West Africa. The vegetative parts of this plant resemble S. obtusa
but the abaxial grooves on the leaf-blade are shallower and less hairy. Whereas the
inflorescence is narrower than in S. dregeana, the glumes are brownish and flushed
with purple and the plumes of the awns tend to be yellowish, as in typical S. dregeana.
5752152-6
358
The lower glume is finely scaberulous (this roughness is more like that of S. obtusa
than of S. dregeana which is coarser and sparser) and slightly exceeds the upper glume
(see Giess & van Vuuren 815 in PRE and Windhoek).
The diagrams of the cross-section of the blade of S. obtusa given by Guenzel (1921 ;
Fig. 22b) and Theron (1936; t. 26 Fig. 1) agree closely with Figure 128 in this paper.
Guenzel’s diagram probably depicts a young leaf since the stereome is absent adaxially
opposite the midrib and the marginal third order bundles. Guenzel also comments
on the variability of the shape of the silicified cells, which further indicates that the
AD
Fig. 130. — S. iiniplumis var. iiniphmis: cross section of the leaf-blade {de Winter 3410).
r
0'1 mm
Fig. 131, — S. iiniplumis var. uniplumis: abaxial epidermis of the leaf-blade.
359
material he used most probably was young. In this respect juvenile leaves of Stipa-
grostis species occasionally show variability: the silicified cells vary from dumb-bell-
shaped to circular in some young leaves whereas they are almost invariably circular
in well developed, mature leaves. See also .Telenc (1950) for an account of S. obtusa
{Aristida obtusa) in North Africa.
21. S. uniplumis (Licht.) de Winter in Kirkia 3: 136 (1963)
Perennial or weak perennial, usually densely caespitose, up to 75 cm or more
high. Culms erect, 3-4-noded, simple or branched upwards, scaberulous, or smooth
and glabrous; nodes glabrous. Leaf-sheaths glabrous; auricles densely long-bearded.
Leaf-blades setaceous, convolute, up to 15 cm high, glabrous on the lower surface,
scaberulous on the upper, and with a few long hairs towards the ligule. Panicle exserted,
narrow and contracted or effuse. Spikelets palid. Glumes glabrous or with a few
hairs on the margins, chartaceous, unequal; the lower 8-9 mm long; the upper 9-11
mm long, flat or grooved on the back and conspicuously veined. Lemma abruptly
passing into the awn, including the callus 2 ■5-4-0 mm long; callus about 1 mm long,
acute, densely bearded ; column about 5 mm long with a dense pencil of hairs at the
branching point of the awn, otherwise glabrous, central awn up to 2 • 5 cm long, naked
in the lower third or plumose to the base; lateral awns naked, up to 12 mm long, very
fine. Caryopsis narrowly elliptic-oblong, 2-2-5 mm long; embryo nearly half the
length of the grain; hilum as long as the grain.
Anatomy (Fig. 130, 131 and 160: 20a-c)
Leaf-blade more or less flat to widely V-shaped in transverse section; keel almost absent; abaxial
surface undulate or practically smooth; adaxial surface with first order bundles slightly protruding;
margins sub-acute to obtuse. Abaxial epidermis: cells fairly thin-walled to thick- walled : stomatal
zones with one to three rows of stomata and 3-6 rows of long strongly ripple-walled cells as well as
scattered narrowly-oblong short elements; bicellular hairs present; silicified cell zones with 4—7 files
of silicified cells, alternating with files of finely undulate ripple-walled cells; silicified cells sub-circular,
in young leaves occasionally kidney-shaped or dumb-bell-shaped, solitary or accompanied by square,
oblong or transversely oblong suberized cells with slightly undulate walls; both zones usually flanked
by broad-based, retrorse barbs. Vascular bundle units: first order units 3-7, broadly oblong to broadly
ovate in outline; marginal bundle units of the second or third order; third order units ovate, 1-2 on
either side of the midrib and one between the second and first order marginal units; midrib similar to
the other first order bundles. Bundle sheaths: outer sheath of large thin-walled cells containing
chloroplasts, abaxially interrupted by the stereome strands in the first order bundles and occasionally
interrupted in the marginal second order bundles; complete or interrupted in some of the third order
bundles; inner sheath of smaller thick-walled cells, complete in most bundles. Chlorenchyma a single
row of tabular cells arranged radially around the bundles. Stereome strands developed abaxially
opposite the first order bundles and the marginal second order bundles weakly, developed and usually
in contact with the first order bundles adaxially, small groups of fibres present opposite the third order
bundles abaxially, but absent adaxially. Motor cells forming large triangular groups of large thin-
walled cells alternating with all bundles; occupying f of the width of leaf-blade.
S. uniplumis sens. lat. is closely allied to S. hirtigluma but may be distinguished
by the finer tubereles on the upper part of the lemma which abruptly passes into the
awn ; by the naked column and distinct pencil of hairs at the branching point of the
awns; the shorter central awn with the apex of the plume usually obtuse in outline;
the usually glabrous glumes; and the bristles of the callus which gradually increase
in length upwards but are not arranged in tufts.
The var. intermedia has pilose glumes and occasionaly the plume ends in a naked
tip.
Glumes glabrous; perennial or weakly perennial:
Panicle with numerous spikelets, contracted or open; central awn of lemma naked in the lower
third, or if plumose to the base, then neither rigid and bristle-like nor dark in colour but
diverging at a sharp-angle from the lemma; glumes usually not exceeding 10 mm in length
var. uniplumis
360
Panicle often with rather few spikelets ; central awn of lemma plumose to the base, rather rigid
and bristle-like, frequently dark purple, diverging at nearly a right angle from the lemma;
glumes at least 10 mm long, usually longer var. neesii
Glumes pilose; annual var. intermedia
(a) var. uniplumis.
Aristida uniplumis Licht. in R. & S., Syst. Veg. 2: 401 (1817). Aristida uniplumis var.
pearsonii Henrard, Crit. Rev. Aristida 3: 647 (1928).
Arthratherum uniplume (Licht.) Nees, FI. Afr. Austr. 181 (1841).
Sub-perennial or perennial with simple cuhns, or culms branched upwards. Glumes
usually not exceeding 10 mm in length. Lemma from 2-3-5 mm long excluding the
callus; column glabrous or hairy upwards, central awn variable in length, plumose
to the base or only in the upper two-thirds, usually palhd (rarely purplish), the plumes
silvery.
Cape. — Herbert; Acocks 1938. Kuruman; Pole Evans 2094.
South West Africa. — Bethanien: Kinges 2167. Luderitz; Kinges 2402. Otji-
warongo; Volk 1694. Okavango: de Winter & Wiss 4415; Volk 394. Windhoek:
Liebenberg 4512; de Winter 2610.
Orange Free State. — Boshof; Burtt-Davy 10123.
Transvaal.^ — Waterberg: Pole Evans 8817. Soutpansberg : Pole Evans 906.
Distribution. — (See Fig. 132.)
This variety is rather variable, as may be seen from the description. Specimens
from the drier coastal areas (with less than 150-200 mm rainfall) are usually weakly
perennial; they possess a rather long central awn with a conspicuous, naked basal
part, short lemmas, 2-2-5 mm long excluding the callus, and usually contracted
inflorescences. A gradual increase in stature and the development of a more perennial
base seems to be coupled with an increase in rainfall further inland. On the inland
plateau of South West Africa some specimens have the central awn plumose to the
base but differ in no other respect from the typical specimens. These represent the
former var. pearsonii.
Hybrids between var. uniplumis and S. hirtigluma occur fairly frequently. These
hybrids have very long awns, a pencil of hairs at the branching point of the awn and
very sparsely hairy glumes, thus combining the characteristics of the two species.
Examples of such specimens are: Boss s.n. (Tvl. Mus. No. 35618); Boss s.n. (Tvl.
Mus. No. 36205); Boss s.n. (Tvl. Mus. No. 36357); Bradfield 373; de Winter 5710.
{b) var. neesii {Trin. d Rupr.) de Winter in Kirkia 3: 136 (1963).
Aristida uniplumis var. neesii Trin. & Rupr., Gram. Stip. 173 (1842).
Perennial, forming dense tufts and usually with a short knotty branched rhizome.
Culms simple or branched upwards. Panicle often rather narrow but lax and bearing
few spikelets. Glumes usually exceeding 10 mm in length. Lemma 3-5-4 mm long
excluding the callus, column hairy or glabrous; central awn usually rather short,
plumose to the base, usually dark in colour, stiff and almost bristle-like, diverging
at a right angle fiom the lemma and column.
Anatomy (Fig. 160: 20b)
Leaf-blade more or less V-shaped in transverse section, much more strongly developed than in
var. uniplumis-, stereome very well developed abaxially opposite the 5-7 first order bundles and relatively
weakly developed opposite some of the third order bundles; adaxially fairly well developed opposite
the first, and to a lesser degree opposite the third order bundles; the marginal first and second order
bundles with a fused stereome. Vascular bundle units hardly projecting adaxially, abaxial surface
smooth.
361
Cape. — Vryburg; Stent H. 2\5\5‘, Sharpe lAl'i. Mafeking; Pole Evans 2AA\. Barkly
West; Esterhuysen 1044. Piquet berg: Pole Evans 2057.
Orange Free State. — Hoopstad: Goossens \2A%. Boshof: Wolff \2.
Transvaal. — Pretoria: de Winter 793. Marico: Story 1210; Codd 1326. Wol-
maransstad: Liebenberg 2437.
Distribution. — (See Fig. 132.)
This variety grades into var. uniplumis and many specimens can only be referred
to either variety arbitrarily. In its typical form it is very distinct both organographically
as well as anatomically (see Fig. 160: 20, 20a, 20b).
It occurs in the same areas as var. uniplumis but often on soil which is less sandy
than that frequented by the latter and in this way is to some extent ecologically separated
from it. Anatomically it also grades into var. uniplumis.
(c) var. intermedia (Schweick.) de Winter in Bothalia 8 : 173 (1964)
A. gracilior var. intermedia Schweickerdt in Bothalia 4, 1 : 124 (1941).
Annual with simple culms. Glumes usually not exceeding 8 mm in length (rarely
9 mm long). Lemma from 2 -5-3 -00 mm long excluding the callus; column glabrous,
central awn pallid or purplish, plume obtuse or with a short excurrent naked tip.
Anatomical features very similar to that of the var. uniplumis but the stereome less developed,
the hairs on the adaxial surface longer and the blade somewhat V-shaped in cross section.
South West Africa. — Swakopmund: de Winter & Hardy S030. Karibib: Volk 150.
Omaruru: Merxmueller & Giess 1680, de Winter & Hardy 8211. Outjo: de Winter
3059; Liebenberg 4962; de Winter & Hardy 8141; 8136; 8160; 8172.
Distribution. — (See Fig. 132.)
This variety is mainly confined to the submarginal Namib Desert. Even though
rare in normal years, when little rain falls in the desert it is dominant in many areas
in the Karibib, Omaruru and Outjo districts in favourable years. This probably explains
its poor representation in Herbaria.
As indicated by the varietal epithet this variety is somewhat intermediate in
character between S. uniplumis and S. hirtigluma. This may possibly indicate a hybrid
origin especially since it occupies areas where the parent species are usually not present.
Stebbins (1950, p. 279) has shown that where species hybridize the hybrid, if able to
maintain itself, usually occupies habitats to which the parent species are not well
adapted. The var. intermedia occurs deeper into the desert, i.e. in areas too dry for the
parent species.
Schweickerdt places this variety under (Aristida gracilior) = S. hirtigluma because
of the pilose glumes. Except for this character, however, the spikelets indicate a much
closer affinity to S. uniplumis: the lemmas pass abruptly into the awn, the branching
point of the awns is furnished with a distinct pencil of hairs and the callus bristles
are evenly distributed not arranged in two distinct tufts. The presence or absence
of hairs on the glumes has proved to be diagnostically unreliable in other species of
the genus and not too much weight should be attached to this character.
22. S. hirtigluma (Steud.) de Winter in Kirkia 3: 134 (1963)
Annual or perennial, caespitose, usually erect up to 80 cm high. Culms 2-5-noded ;
internodes terete, scaberulous or almost glabrous; nodes glabrous. Leaf-sheaths
finely scaberulous; auricles densely long-bearded. Leaf-blades 6-20 cm long, setaceous,
convolute, shallowly grooved below, or smooth, scaberulous ; the upper surface scabrid
362
to hirtellous. Panicle loose and much branched, or contracted with short branches,
or very open and effuse with the branches spreading, up to 30 cm long and 20 cm wide
or much shorter and narrower; pinkish in colour. Spikelets pallid to green or purplish.
Glumes unequal to sub-equal, acute to sub-acute, densely pilose all over or at least
on the back, 3-nerved; the lower rounded or obtusely keeled on the back, 6-11 mm
long; the upper grooved on the back or at least flattened, never keeled, 10-13 mm
long. Lemma including the callus about 3-5-4 mm long, conspicuously tuberculate,
especially upwards; column 7-10 mm long, hairy or glabrous or scabrid, twisted;
central awn plumose from base to apex or only in the upper part with a naked excurrent
tip, 5-7 cm long; lateral awns naked, up to 2-5 cm long; callus 0-5-0-75 mm long,
curved, the acute tip glabrous, for the remainder hairs more or less in two series: a
short fringe near the base of the callus and a long fringe at the base of the lemma.
Caryopsis very narrowly elliptic, about 2 mm long; embryo the length of the grain;
hilum as long as the grain.
Fig. 132. — Distribution of S. uniplumis: #var. iniiplumis; Ovar. neesii; Mvar. inleniicclia.
Anatomy
Leaf-blade V-shaped in transverse section; abaxial surface flat or slightly undulating, furnished
with large retrorse barbs; keel slightly protruding; margins obtusely rounded. Abaxial epidermis:
stomatal zones with one or two rows of stomata and 2-5 rows of long ripple-walled cells with strongly
and compactly undulate walls; short elements in these zones usually produced into retrorse barbs;
linear 2-celled hairs present; silicified cell zones with 4-6 files containing silicified cells, and 8-12 files
of ripple-walled cells with finely undulate walls adjacent to the first order bundles; 1-2 files containing
silicified cells alternating with ripple-walled elements adjacent to the third order bundles; silicified
cells sub-circular, usually single or occasionally accompanied by a thin-walled more or less square
363
suberized cell with somewhat undulate walls; broad-based retrorse barbs flanking this zone. Vascular
bundle units: first order bundles 3, with or without second order bundles accompanying the marginal
first order bundles, very broadly oblong and rounded in outline; third order bundles 2-3 on each
side of the midrib and one on each margin; midrib similar to the other first order bundles but broadly
ovate in outline and adaxial stereome strands much smaller and not in contact with the bundle. Bundle
sheaths: outer of large thin-walled cells, interrupted abaxially by the stereome strands in the first and
second order bundles; inner sheaths of smaller cells with slightly thickened walls in the abaxial half
of the bundle, and strongly thickened walls adjacent to the stereome strands. Chlorenchyma a single
layer of tabular cells radially arranged around the bundles. Stereome strands: abaxial strands very
strongly developed opposite the first and second order bundles, the strands of the marginal first and
second order bundles fused but small groups opposite the third order bundles; adaxial strands well
developed only opposite the two marginal first order bundles; small groups of cells present opposite
all the other bundles including the midrib. Motor cells strongly developed flanking the midrib and
occupying i or more of the leaf thickness, groups rapidly diminishing the size outwards to absent near
the margins.
The characteristics of the spikelets of S. hirtigluma were compared critically with
those of A. gracilior and no constant differences could be detected. As stated by
Henrard (1928) the only difference between these species is the denser more contracted
inflorescence of 5. hirtigluma. Although small differences are found in the anatomy
of the leaf-blade their general pattern is remarkably similar. Since the species in
question both occur in South West Africa, it does not seem justifiable to distinguish
them solely on the contraction of the inflorescence. A. hirtigluma was described from
Arabia and the disjunct distribution could be advanced as a reason for separating the
species if forms identical with A. hirtigluma were not present in South West Africa.
Nevertheless this argument would not be a very strong one since many species have
a similar disjunct distribution (see p. 304). A. gracilior and its varieties are here,
therefore, united with S. hirtigluma.
S. hirtigluma sens. lat. is closely allied to S. uniplumis but may be distinguished by
the strongly tuberculate upper part of the lemma which gradually narrows into the
awn; by the longer central awn which is excurrent into a naked tip; and by the
pilose glumes.
Annuals with very few leaves at the base, the cuhns not hidden by the leaves; internodes rather
soft and easily compressible:
Panicle rather narrow, when fully exserted much longer than wide var. hirtigluma
Panicle very lax, open and spreading, when fully exserted not much longer than wide var. pearsonii
Perennials with a dense tuft of setaceous rather tough leaves which more or less hide the culms;
internodes firm and not easily compressible; panicle effuse and open, when fully exserted
not much longer than wide var. patula
(a) var. hirtigluma.
Aristida hirtigluma Steud. ex Trin. & Rupr., Gram. Stip. 171 (1842).
Annual, with few basal leaves. Panicle much longer than wide. Spikelets with
the column of the awns glabrous or hairy, central awn hairy down to the branching
point of the awns.
Anatomy (Fig. 133, 134 and 160: 21)
As described for the species, but the marginal first order bundle units not accompanied by large
second order units as is the case in var. patula.
South West Africa. — ^Swakopmund: Giess 3034; Boss (Tvl. Mus. No. 36191).
Rehoboth: Fleck 802. Bethanien: van Son (Tvl. Mus. No. 31827); Kinges 2166;
2159; 2218; 2099.
Distribution. — (See Fig. 137.)
364
O'l mm
Fig. 133.— 5. hirtigluma var. hirtigluma: cross section of the leaf-blade (Trott s.n.; Arabia).
Fig. 134.— 5. hirtigluma var. hirtigluma: abaxial epidermis of the leaf-blade.
365
This variety has a fairly wide distribution in the Namib Desert from Bethanien
in South West Africa to Loanda in Angola; it also occurs in North Africa and the
desert areas of the Middle East. Relatively few specimens from Southern Africa are
available for study. The original description is based on Schimper 165 collected in
Abyssinia.
Variety hirtigluma does not seem to be selective as to habitat and is found growing
in sandy and gravelly soils as well as in rocky substrata.
{b) var. pearsonii {Hemard) de Winter in Kirkia 3: 134 (1963).
Aristida gracilior var. pearsonii Henrard, Crit. Rev. Aristida 3: 695 (1928).
Differing from var. hirtigluma in the effuse, open and spreading panicle which
when fully exserted from the sheaths usually is only slightly longer than wide. Some
specimens have the central awn of the lemma plumose for its whole length while others
are naked towards the base and plumose only upwards.
The anatomy of the leaf-blade of this variety agrees in detail with that of var. hirtigluma.
South West Africa.— Keetmanshoop ; de Winter 3259 ■, 3411; Oertendahl 140; 158.
Gibeon: Schweickerdt 2261 ; van Vuuren & Giess 1161. Omaruru: de Winter 6035a.;
3167; Kinges 3439, 3213; Schweickerdt 2261; 2257; Boss s.n. (Tvl. Mus. No. 36379;
36371; 363746). Outjo: Liebenberg 4931 ; de Winter 3124; Volk 2%14. Kaokoveld:
de Winter & Leistner 5693; 5621; 5307; 5377.
This variety occurs in desert and semi-desert areas from the Warmbad district
in the South to the Kaokoveld in the North of South West Africa (see Fig. 137) as
well as in Angola. The type specimen was collected in the Mossamedes area of Angola
{Pearson 2249).
The character on which the annual var. pearsonii was distinguished in the past
from the perennial var. patula {Aristida gracilior) i.e. the bare lower part of the central
awn, is unreliable. Both plumose awns as well as awns with a naked base occur in
var. pearsonii. The annual habit and absence of a dense basal tuft of leaves seems
to be a more reliable character for distinguishing var. pearsonii from var. patula, since
the latter is a perennial and possesses a dense basal tuft of leaves.
A. gracilior var. intermedia Schweickerdt agrees with S. hirtigluma sens, lat., only
in the pilose glumes and annual habit. In the characters of the lemma and awns it
matches S. uniplumis var. uniplumis in detail. For this reason it has been transferred
to S. uniplumis.
(c) var. patula {Hack.) de Winter in Kirkia 3: 134 (1963).
Aristida hirtigluma var. patula Hack, in Denkschr. Kais. Akad. Naturw. 78: 401
(1906). Aristida gracilior Pilger in Engl. Bot. Jahrb. 40: 80 (1907).
Densely tufted, strictly erect perennial, with a dense basal tuft of tough setaceous
leaves. Panicle effuse and spreading with the branches well developed, when fully
exserted from the sheaths not much longer than wide. Lemma with the central awns
plumose for its entire length; the column glabrous or hairy.
Anatomical structure of the leaf-blade is as described for the species but the marginal first order
bundle and the stereome of these bundles fused to form a single unit (Fig. 135, 136 and 160: 21a),
Cape. — Vryburg: Brueckner 1180. Kimberley: Leistner 1634.
South West Africa. — Windhoek: de Winter & Giess 1\35; de Winter 3521; SchweL
ckerdt 2264; Liebenberg 4511. Otjiwarongo: Liebenberg 4903. Karibib: de Winter
2680. Grootfontein : Merxmueller & Giess \340; Schweickerdt 2\68; 2294; de Winter
2833; de Winter & Giess 6801; Story 5229. Estosha Game Reserve: Volk 2411;
Breyer s.n. (Tvl. Mus. No. 20614).
366
Bechuanaland. — Mabua Sehoba Pan, de Winter 7464.
Transvaal. — Soutpansberg ; Codd de Winter & Codd 'iXQ. Leydsdorp: Soek-
mekaar: Hafstrom & Acocks 3.
Fig. 135. — S. hirtigluma y^x. patiila: cross section of the leaf-blade (de Winter 3527).
Fig. 136. — S. hirtighima var. patida: abaxial epidermis of the leaf-blade.
Distribution. — (See Fig. 137.)
This variety which occurs in less dry areas than the previous varieties shows a
marked preference for calcareous soils and is common on limestone outcrops on the
367
inland plateau of South West Africa from Windhoek northwards. It is seen occasionally
in sandy areas usually with limestone outcrops occurring in the vicinity.
As in the case of S. uniplumis the habit of the varieties of /i. hirtigluma is strongly
correlated with the annual rainfall; those occurring in the drier coastal areas being
annuals whereas var. patula, the inland variety, is a perennial.
In the Waterberg and Grootfontein areas of South West Africa very robust
specimens, possessing pilose glumes, occur. Because of the hairy glumes these have
previously been referred to the var. patula {A. gracilior) but differ from typical plants
in the taller harder cuhns, the often shorter awns and the less conspicuously tuberculate
lemmas. These specimens are very variable in floral characters: the glumes being
almost glabrous to densely pilose whereas the habit is strongly reminiscent of S.
uniplumis. Volk who collected extensively in the Waterberg area remarked on the
resemblance in habit of some of these specimens to S. uniplumis, occurring in the same
area. There seems to be little doubt that the following specimens represent hybrids
between S. hirtigluma var. patula and S. uniplumis var. neesii: Volk 1080; 1020; 376;
1531; 2338; cle Winter 2778; Giess 2223; Boss (Tvl. Mus. Nos. 36456; 36205).
Esterhuysen 1026; le Roux 1043. The anatomy of the leaf-blade of these specimens
approaches that of S. uniplumis var. neesii much more closely than that of S. hirtigluma
var. patula (see Fig. 138, 139 and 160: 20c).
Fig. 137.— Distribution of 5. %wa.r. hirtigluma; ^var. patula; 0\av. pearsonii.
368
Fig. 138.-S. uniplumis var. neesii X S. hirtigluma var. pawh: cross section of the leaf-blade (rfe
Winter 2778).
SC
I
Fig. 139. — 5. uniplumis var.
1
0-1 mm
neesii X 5. hirtigluma var. patula: abaxial epidermis of the leaf-blade.
369
23. S. hermannii (Mez) de Winter in Kirkia 3: 134 (1963)
Aristida hermannn Mqz in Fedde Rep. 17; 153 (1921).
Annual, laxly caespitose, up to 15 cm high. Culms geniculately ascending, 2-4-
noded, shortly hairy below the uppermost node; internodes densely but minutely
pubescent. Leaf-sheaths loose, chartaceous densely scabrid especially on the ribs,
margins ciliolate; auricles bearded. Leaf-blades much reduced, folded or convolute,
densely scabrid on both surfaces, 1-2 cm long and about 1 mm wide. Panicle partly
enclosed by the uppermost sheath, ovate-oblong, dense, up to 8 cm long; rhachis and
branches scabrid; axils pubescent. Spikelets silvery. Glumes narrowly lanceolate,
acute, shortly awned, 3-nerved, the lower usually exceeding the upper, 9-1 1 • 5 cm long,
densely but minutely scaberulous; the upper 8-10 cm long, scaberulous towards the
apex. Lemma punctuate-scaberulous, including the callus 3-5-4 mm long; callus
very acute, densely bearded, 1-1-25 mm long; column very scabrous, twisted, hairy
below and at the branching point, very variable in length, 1-10-5 mm long; central
awn scabrid in lower third, plumose upwards, 2-4 cm long; lateral awns naked, very
fine, 1-1-5 cm long. Caryopsis very narrowly oblong-elliptic; embryo about ^ the
length of the grain; hilum as long as the grain.
Fig. 140. — S. hermannii: cross section of the leaf-blade (Kinges 2634).
Anatomy (Fig. 140, 141 and 160; 22)
Leaf-blade V-shaped in transverse section, both surfaces strongly undulate giving the section a
moniliform appearance; hairs present both ad- and abaxially, the abaxial longer and thinner; keel
well developed; margins obtuse. Abaxial epidermis thin-walled: stomatal zones with 3-5 rows of
stomata and 4-7 rows of long ripple-walled elements interspersed with long retrorse barbs; bi-cellular
linear hairs present in these zones. Silicified cell zones with 3-4 files of short elements alternating
with usually single files of long ripple-walled cells (undulations less marked than in the stomatal zones)
adjacent to the first order bundles; 2-3 files of short elements present adjacent to third order bundles;
silicified cells variable in shape, but usually dumb-bell-shaped and alternating with oblong to almost
square, slightly, to distinctly, ripple- walled suberized cells, or large broad-based retrorse hairs; bicellular
hairs found mainly in the transitional area between the stomatal and silicified cell zones. Vascular
bundle units: first order units 3, circular in outline; third order units 3 on either side of the midrib
and two on each margin, almost as large as the first order units; midrib better developed than the
370
other first order bundles. Bundle sheaths: outer sheath of large thin-walled cells; inner of smaller
irregular cells with thickened walls, interrupted only abaxially by the stereome in the first order bundles
and in the midrib, but complete in all the third order bundles. Chorenchyma a single row of tabular
cells radially arranged around the bundles. Stereome strands well developed only opposite the first
order bundles adaxially, but present as small groups of fibres, often only one cell layer thick, both
ab- and adaxially in all the other bundles. Motor cell groups present, alternating with all except the two
marginal bundles on either side; 4—6 cells in each group, occupying about half the leaf thickness.
Fig. 141. — S. hermannii: abaxial epidermis of the leaf-blade.
South West Africa. — Luderitz : Dinter 6396; Kinges 2021; 2634; Giess & van
Vuuren 729. Swakopmund ; Boss s.n. (Tvl. Mus. Nos. 36395; 36439).
Distribution. — (See Fig. 142.)
Closely allied to S. subacaulis from which it may be distinguished by the better
developed culms which are somewhat exserted from the tuft of leaves, and by the
sparsely, bristly hairs below, and at the branching point of the awns. Anatomically
these two species are rather distinct, differing from all the other species in the moniliform
appearance of the leaf in cross-section and the pronounced keel projecting abaxially.
It is difficult to distinguish S. hermannii from S. subacaulis anatomically. To some
extent S. hermannii is also related to S. uniplumis particularly the sub-perennial desert
forms of the latter species.
S. hermannii is confined to the drier coastal areas of South West Africa.
24. S. namibensis de Winter in Bothalia 8 : 173 (1964)
Annual forming sprawling lax tufts, up to 20 cm high, but usually smaller. Culms
simple, 2-3 (4)-noded, internodes densely scabrid to sub-pubescent, geniculately
ascending, nodes glabrous. Leaf-sheaths loose, scabrid. Leaf-blades rather thin,
ribbed abaxially and adaxially and scabrid on the ribs. Ligule a fringe of hairs. Panicle
somewhat contracted but much branched and not very dense, rhacbis and branches
densely scabrid, branches usually 2-nate, alternate. Spikelets erect, pallid, usually
purplish at the very base. Glumes unequal, inverse, glabrous, rather firm, chartaceous,
lanceolate; lower usually distinctly exceeding the upper in length, 8-9 mm long,
sub-obtuse emarginate and with a few short hairs at the hyaline apex 3- (4-5)-nerved,
rounded on the back; upper 3-nerved acute, 6-7 mm long and flattened on the back.
371
Lemma 3-nerved, tubular with the margins strongly overlapping, glabrous, smooth,
including the callus 3-5-5 mm long, articulation situated between the apex of the
lemma and foot of the column; column very variable in length 2-5-5 mm long, hairy;
awns 3, with only the central plumose in the upper |-f, shortly hairy downwards
(plumes acute to sub-obtuse) 4-2-5 cm long; lateral awns fine half or usually less
than half the length of the upper, very finely scabrid, strongly diverging from the
central; callus very acute, often somewhat twisted, densely bristly, the bristles at the
base of the lemma longer. Caryopsis narrowly elliptic in outline, 2-2-5 mm long;
embryo about | the length of the grain; hilum as long as the grain, dark brown.
Fig. 142. — Distribution of S. hermannii; S. subacaidis.
Anatomy (Fig. 143, 144 and 160: 22a)
Leaf-blade horseshoe-shaped in transverse section or somewhat expanded ; ad- and abaxial surfaces
somewhat undulated, furnished with retrorse barbs on the abaxial surface; adaxlally mainly with
fairly long unicellular hairs; keel mainly projecting abaxially; margins obtuse. Abaxial epidermis:
cells thin- walled; stomatal zones with 2^ rows of stomata and 4-8 rows of long ripple-walled cells,
undulations moderately deep and close together; silicified cell zones 4-6 files of short elements alter-
nating with 5-7 files of long ripple-walled elements adjacent to the first order bundles; 1-3 files of
short elements present opposite the third order bundles; silicified cells dumb-bell-shaped; bicellular
hairs present but few. Vascular bundles: first order bundle units 3, subcircular; third order bundles
subcircular, 2-3 on each side of the midrib and one on either margin; midrib forming a fairly distinct
keel. Bundle sheaths: outer of large regular thin-walled cells filled with chloroplasts; inner smaller
and lacking chloroplasts. Chlorenchyma of a single layer of tabular cells radially arranged around
the bundles. Stereome strands well developed only abaxially opposite the first order bundles, present
as small groups opposite the third order bundles, very weakly developed or absent adaxiaJly. Motor
cells fairly well developed between all bundles except the marginal two.
372
S
Fig. 143. — S. namibensis: cross section of the leaf-blade {Keel 1698).
O' 1 mm
Fig. 144. — S. namibensis: abaxial epidermis of the leaf-blade.
South West Africa. — Swakopmund : de Winter & Hardy 8025. Omaruru : de Winter
& Hardy 8199. Namib Desert (no precise locality), Keet 1698.
S. namibensis is allied to S. hermannii and S. subacaulis. From S. subacaulis
it can be distinghuished easily by the well developed culms. Its closest affinity is with
S. hermannii which on the whole has longer awns (2-4 cm) and membranous narrow
glumes which taper to a very fine apex. In S. namibensis the awns are usually shorter
373
( 1 • 4-2 • 5 cm) and the glumes broadly lanceolate with acute to sub-obtuse and emarginate
apices. It is not uncommon in the desert near the coast in the Swakopmund and
Omaruru districts, at least in good years. It occurs in the same habitat as the other
two related species, often together with them.
25. S. subacaulis {Nees) de Winter in Kirkia 3: 136 (i963)
Arthratherum subacaulis Nees, FI. Afr. Austr. 1: 180 (1841).
Aristida subacaulis (Nees) Steud., Nomencl. Bot. 2, 1: 132 (1842).
Annual forming small compact tufts up to 10 cm high, but usually much smaller.
Culms minutely hairy; internodes finely hairy; nodes more or less glabrous. Leaf-
sheaths loose, chartaceous, scabrid, above ; auricles bearded. Leaf-blades short,
folded or convolute, up to 3 cm long and about 1 mm wide, strongly striate, densely
scabrid below, hispid on the upper surface. Panicle contracted, usually hidden by the
leaves, enclosed at the base by the upper sheath; rhachis and branches scabrid.
Spikelets silvery. Glumes linear-lanceolate, acuminate, 3-nerved, membranous; the
lower 10-13 mm long; the upper 8-10 mm long. Lemma smooth, glabrous, including
the callus 3-4 mm long; column very variable, 1-13 mm long, twisted; central awn
scaberulous to shortly hairy in the lower part, plumose above the middle, up to 4-3
cm long; lateral awns fine, glabrous, up to 1-8 cm long. Caryopsis very narrowly
oblong-elliptic; embryo about half the length of the grain; hilum as long as the grain.
Anatomy (Fig. 145, 146 and 160: 23)
Leaf-blade V-shaped in transverse section; upper and lower surface strongly undulate, furnished
abaxially with papilla-like cells and retrorse barbs; adaxially mainly with long, thin unicellular hairs;
keel strongly developed and projecting both ad- and abaxially but more strongly so abaxially; margins
truncate to obtuse. Abaxial epidermis: cells thin-walled: stomatal zones with 2-4 rows of stomata
and 4-6 rows of long ripple-walled cells, undulations deep and well-spaced; silicified cell zones with
5-7 files of short elements alternating with 1-2 files of long ripple-walled cells (undulations much less
distinct than in stomatal zones) adjacent to the first order bundles; 2-3 files of short elements present,
adjacent to third order bundles; silicified cells dumb-bell-shaped alternating with narrowly oblong to
alrnost square, slightly to distinctly undulate, suberized cells and large broad-based retrorse hairs;
bicellular hairs present, flanking the stomatal zones. Vascular bundles: first order bundle units 3,
broadly ovate to oblong-circular; third order units subcircular, 2 or 3 on either side of the midrib and
one on either margin ; midrib larger and with a larger stereome strand than the other first order bundles,
forming a conspicuous keel. Bundle sheaths: outer of large, regular, thin-walled cells filled with
chloroplasts; inner of smaller very slightly thickened more irregular cells, lacking chloroplasts; outer
sheath interrupted only abaxially by the stereome strands in the first order bundles, otherwise complete.
Chlorenchyma of a single row of tabular cells radially arranged around the bundles. Stereome strands
well developed only abaxially opposite the first order bundles, small few-celled groups present adaxially
opposite all the bundles, and slightly better developed groups present abaxially opposite second and
third order bundles. Motor cells consisting of two large groups of cells flanking the midrib; smaller
groups alternating with the other bundles except the marginal two.
Cape. — Namaqualand. — Drege s.n. (Verleptpram; isotype); Hardy 551; Marloth
12394.
South West Africa. — ^Luderitz; Giess & van Vuuren 740; 705; Giess 2348; Dinter
4104; Kinges 2633. Swakopmund: Boss s.n. (Tvl. Mus. 36316); Schweickerdt 2222;
2232; 2241.
Distribution. — (See Fig. 142.)
For affinities see notes under S. hermannii (p. 370). The distributional range of
this species is approximately identical to that of .S. hermannii and the two species often
occur together.
Theron (1936) compares S. lanipes with S. subacaulis. The more or less V-shaped
leaves of S. subacaulis, in which the motor cell groups alternate with most of the bundles,
differ markedly from those of S. lanipes where the subterete leaves possess very prominent
motor cell groups only flanking the midrib. The affinity of S. subacaulis no doubt
lies with S. hermannii, a fact that Theron (1936, p. 31) also mentions but only under
374
S. hermannii. S. hermannii and S. subacaulis are, however, not placed next to each
other in Theron’s enumeration. Theron’s Fig. 31 depicting a cross-section of the leaf-
blade of S. subacaulis, agrees reasonably well with Fig. 160: 23 of this paper with the
exception that no stereome strands are indicated abaxially opposite the third order
bundles.
Fig. 145. — -S. subacaulis: cross section of the leaf-blade (Kinges 2633).
Fig. 146. — 5. subacaulis: abaxial epidermis of the leaf-blade.
375
26. S. anomala de Winter in Kirkia 3: 133 (1963)
Stipa namaquensis Pilger in Bot. Jahrb. 51 ; 12; 412 (1914) non Stipagrostis namaquensis,
(Nees) de Winter. Type: Namaqualand, Dinter 2602 (leg. Koppel) (B, holof; fragment
PRE).
Perennial, or weak perennial, forming erect dense to lax tufts 10-60 cm high,
densely to sparsely leafy at the base. Culms erect or somewhat geniculate near the
base, 1-2-noded; nodes glabrous: internodes glabrous, finely striate. Leaf-sheaths
lax, chartaceous shallowly grooved upwards, upper sheaths often firmer and tightly
clasping the culms, glabrous. Leaf-blades setaceous, the basal often short (1 cm) and
curved, or up to 20 cm long and curling at maturity; subterete, deeply grooved and
scabrid on the upper surface, more or less smooth or slightly grooved and minutely
scaberulous below. Panicle erect, stiff, very narrow and elongate, much interrupted,
or more dense when very short, 3-17 cm long; rhachis and branches scabrid and
narrowly grooved. Spikelets densely congested on short branches or in fascicles on
the rhachis. Glumes chartaceous, subequal or unequal, the lower shorter than the upper,
lanceolate, pallid or purplish, 3-nerved, apices erose-truncate or acute; the lower
Fig. 147.— S. anomala: cross section of the leaf-blade {Scott s.n.).
376
rounded on the back to slightly keeled, scabrid to scaberulous on the outer surface,
7-10 mm long; the upper flat to slightly furrowed between the lateral nerves dorsally,
scaberulous to scabrid, 9-12 mm long. Lemma glabrous, finely punctate with scattered
white tubercles upwards, about 4 mm long including the callus; column of the awn
with a pencil of long stiff hairs at the base, glabrous upwards, twisted, grooved, 4-7
mm long; awn solitary, rarely with the rudiments of the lateral awns visible as small
protuberances, glabrous and densely scabrous, when mature diverging at almost a
right angle from the column, pallid or purple; callus acute, densely bristly. Palea
very short, scale-like, about 1 mm long, 2-nerved, glabrous. Lodicules 2, lanceolate,
membranous, many-nerved, 1-1-5 mm long. Ovary glabrous, styles distinct, stigmas
linear, plumose. Caryopsis narrowly spindle-shaped; embryo about ^ the length of
the grain; hilum linear, situated in a narrow shallow groove.
Anatomy (Fig. 147, 148 and 160 : 24)
Leaf-blade sub terete or horseshoe-shaped in transverse section; abaxial and adaxial surfaces
almost smooth or very shallowly grooved; keel absent; margins sub-obtuse. Abaxial epidermis:
stomatal zones with 1-3 rows of stomata and 3-5 rows of elongate-rectangular, long ripple-walled
cells with strongly undulate walls, bicellular hairs present; retrorse unicellular barbs few; silicified
cell zones with 6-9 files of silicified cells alternating with 5-8 files of narrow, elongate long elements with
finely undulate walls; silicified cells sub-circular or weakly dumb-bell-shaped, usually accompanied
by square to oblong suberized cells with undulate walls. Vascular bundle units: first order units 3,
shape not well defined; the third order units flanking the midrib, paired, the marginal units single
or paired, much smaller than the first order units; midrib similar to the other first order bundles.
Bundle sheaths: outer sheath cells much exceeding the slightly thickened inner sheath cells in size,
only the outer interrupted abaxially by stereome, all others complete. Chlorenchyma consisting of
a single layer of tabular cells radially arranged around the bundles. Stereome strands exceedingly
well developed abaxially opposite the first order bundles, present only as small strands adaxially
opposite the third order bundles; a small adaxial strand present between the mid-bundle and motor
cell group. Motor cells forming a single large fused group opposite the nfidrib and the pairs of flanking
third order bundles, otherwise absent.
0- 1 m m
Fig. 148. — S. anomala: abaxial epidermis of the leaf-blade.
Cape. — Swellendam: Esterhuysen 1865. Beaufort West: Acocks 14329; Fairall 325.
Prieska; Liebenberg 4114; Bryant K 35; Scott s.n. (N.H. 20131). Herbert: Leistner
1316. Hay: Acocks 2357 and 2336. Kenhardt: Minnell s.n. (N.H. 20203); von
Broembsen 92; Leistner 2362.
377
South West Africa. — Liebenberg 5165; Schweickerdt 2280; 2272; de Winter 3409;
3258. Luderitz: de Winter & Giess 6135, Kinges 2362; 2638. Bethanien: Kinges
2161; 2165. Mariental: Schweickerdt 2310; 2270.
Distribution. — (See Fig. 149.)
This anomalous species was in the first instance described as a Stipa because it
has a single awn. Investigation of the anatomy of the leaf-blade as well as the organo-
graphy of the spikelet has shown that it belongs to the genus Stipagrostis (see also
Fig. 1). It is related to S. obtusa which it resembles in the anatomy of the leaf-blade
(cf. Fig. 128 and 147) in habit, and in the finely scabrid glumes. It differs from all
the other species of the genus on account of its single glabrous awn and for this reason
is placed in the new section Anomala.
Although the awn is glabrous the base of the column is furnished with a pencil
of long fine hairs resembling the pencil of hairs present at the branching point of the
awns in S. unipluniis. The large outer chlorophyll-bearing bundle sheath, circular
silicified cells and the three-nerved glumes are characters typical of Stipagrostis and
there is no doubt that 5. anomala belongs to the latter rather than to Aristida. The
presence of vestiges of lateral awns in some specimens indicates that S. anomala probably
originated from a three-awned ancester, and provides an additional reason for including
it in a new section of Stipagrostis, rather than placing it in a new genus on account
of the single awn.
2.8.8. Interrelationship of the South African Species and Sections
The closer affinities of the various species have been discussed under the description
and enumeration of the species (2.8.7., p. 313). A summary of these views is given in
diagrammatical form in Fig. 150, p. 378. Here very closely related species are connected
by double lines, closely related species by single unbroken lines, and more distant
affinities by dotted fines.
378
379
This diagram illustrates that subdivision of the genus Stipagrostis into the ^Schis-
tachne and ^Stipagrostis on the basis of the position which the articulation occupies
in the lemma, results in the recognition of two groups of species which in their delimi-
tation are not entirely natural. For example, S. namaquensis {§Schistachne) appears
to be very closely related to both S. amabilis i§Schistachne) and S. damarensis i§Stipa-
grostis), and, to a lesser degree, to S. proximo {^Stipagrostis), and furthermore to S.
brevifolia and S. sabulicola, both of the ^Stipagrostis. S. damarensis (^Stipagrostis),
on the other hand, shows a stronger degree of relationship to S. namaquensis {§Schis-
tachne) than to S. brevifolia (^Stipagrostis). S. sabulicola (^Stipagrostis) shows affinities
both with S. namaquensis (§Schistachne) and S. amabilis (^Schistachne), but little affinity
with the remaining species of the ^Stipagrostis except for S. damarensis, thus occupying
a somewhat isolated position (perhaps unnaturally so) among the members of this
section.
Similarly, S. dinteri (§Schistachne) finds its closest relatives in the ^Stipagrostis,
in particular the species S. obtusa, S. lanipes and S. gonatostacliys, although affinity
with S. hochstetteriana (^Schistachne) is not lacking. S. ciliata, a member typical of
the §Scliistachne, shows a very close degree of relationship to S. schaeferi, less so to
5. hochstetteriana, both members of ^Schistachne, but the sphere of relationship includes
S. zeyheri and S. dregeana, both members of the ^Stipagrostis.
The vertical dotted line in Fig. 150 divides, on the left hand, an interrelated group
of suflfrutescent, rhizomatous species from, on the right hand, a separate interrelated
group of which the species are mainly caespitose (S. garubensis and S. lutescens have
suffrutescent bases). The two groups, individually, aie composed of species of both
§Schistachne and §Stipagrostis.
In view of the above, from a taxonomic point of view, the sections Schistachne
and Stipagrostis should be united. However, the position of the articulation in the
lemma is so useful and convenient a character in the identification of the species, that
for the present the two sections have been retained.
Hybrids occur fairly frequently and of these, at least two are the result of crossing
where one parent belongs to ^Stipagrostis and the other to ^Schistachne (S. brevifolia
X S. namaquensis; S. zeyheri X S. ciliata). No hybrids have been reported involving
on one hand, a species from the suffrutescent group and, on the other, from the
predominantly ceaspitose group.
No karyological evidence could be found for dividing the genus Stipagrostis into
sections, all the South African species investigated being tetraploids with 2n = 44.
2.9. THE GENUS SARTIDIA
2.9.1. Affinities
In the discussion on the generic delimitation of the Aristideae on p. 226, the genus
Aristida sensu lato, is divided on basis of organographic as well as anatomical charac-
teristics into three groups (see Table on p. 226). Groups A and B represent the genera
Aristida and Stipagrostis respectively, whereas group C, which consists of three species,
represents a new genus. Enumeration of the characteristics of the three groups showed
that this new genus, Sartidia, may be distinguished from Aristida and Stipagrostis
by the size and structure of the embryo, and the anatomy of the leaf-blade and awns.
In addition, Sartidia differs from Stipagrostis in the nature of the awns; Sartidia has
stiff rigid spirally contorted awns, while Stipagrostis has very slender finely plumose
awns. Sartidia and Aristida are, however, very similar in organographic charac-
teristics and thus may easily be confused. Sartidia is characterized by the possession
of 3-5-nerved glumes, lemmas which are hardly narrowed towards the apex and are
produced into three stout, spirally contorted bristles, a very small embryo and a deeply
380
ventrally grooved caryopsis, as well as leaves which are rolled in the bud. Although
all these characteristics are present in the genus Aristida, they are never represented
jointly in any of the Aristida species. Should an examination of the spikelet be incon-
clusive in establishing the generic identity of a species, study of the anatomy of the
leaf-blade and embryo will show clearly whether the species belongs to Sartidia or
Aristida. (Fig. 152, 154, 156.)
The question now arises whether it would be sound to create a new genus on the
basis of such seemingly weak organographic differences, quite apart from the fact
that these are supported by distinctive anatomical differences.
For the Gramineae, a family characterized by extreme reduction of the floral
parts, many cases could be cited where, in the past, genera were created on apparently
insignificant organographic differences. For example, one need only mention Poa
and Eragrostis, Eragrostis and Diplachne, Monerma and Lepturus, and many others.
All of these genera are today accepted as being quite distinct; some of them are even
regarded as belonging to different tribes. The present-day tendency in grass taxonomy
is to recognize as genera those groups which, to a degree, are uniform in their combined
organographic, anatomical as well as cytological characteristics, and this approach
undoubtedly has resulted in a greatly improved classification. It is therefore felt that
retention of the species of Sartidia, which differ from Aristida in the anatomical
structure of both the embryo and the leaf-blade, in the otherwise homogeneous
Aristida, would be contrary to the accepted practice.
The two genera Aristida and Stipagrostis together form a relatively homogeneous
tribe, although Aristida has double chlorophyll-bearing bundle sheaths, whereas in
Stipagrostis only the cells of the outer sheath contain chloroplasts. If Sartidia were
to be included in the tribe, the Aristideae would be a tribe heterogeneous in anatomical
constitution. In addition, the structure of the embryo of Sartidia differs from that
of Aristida and Stipagrostis in the absence of a cleft between the coleorrhiza and the
base of the scutellum. For these reasons collectively it was decided to treat Sartidia
as a separate entity, and hence its characteristics were excluded from the description
of the Aristideae.
On the other hand, a chromosome count of Sartidia jucunda revealed a basic
number of n = 1 1 (2n = 22) and no significant size differences could be detected
between the chromosomes of S. jucunda and those of the various species of true
Aristida. The possibility should therefore not be ignored that Sartidia may yet prove
to be an anomalous genus of the Aristideae.
Of many genera investigated, the Australian genera Amphipogon, Diplopogon,
Plectrachne and Triodia show the greatest similarity to Sartidia in the anatomy of the
leaf-blade. In these genera the chlorenchyma cells are arranged diffusely while bicellular
hairs are present, characters agreeing with those of Sartidia. In Amphipogon and
Plectrachne the embryo has a cleft between the coleorrhiza and the base of the scutellum,
a character they thus have in common with members of the Danthonieae, and in
addition, the anatomy of the leaf-blade agrees in general features with that of Danthonia,
so that they probably could be placed in the same tribe. However, the characters
of the embryo and hilum, as well as the structural differences of the lemma, prevent
Sartidia from being placed near these genera and exclude it from the Danthonieae.
Although the embryo of Diplopogon and Triodia have not been examined, other
characters show such similarity to those of Amphipogon and Plectrachne that it is
probable that their embryos are similar in structure.
Since the attempts to establish close affinities for Sartidia have failed, it is felt that
further progress will hardly be possible until a wider knowledge of the anatomical
features of various other genera has been acquired. Thus for the time being the genus
is placed next to Aristida.
381
2.9.2. SARTIDIA de Winter
In Kirkia 3: 137 (1963)
Spikelets solitary, pedicelled. Rhachilla disarticulating above the glumes, not
produced beyond the base of the floret. Floret 1, hermaphrodite, exceeding the glumes
(the awns long exserted), linear in outline. Glumes persistent, narrow, acute to
acuminate, with or without apical awns, the lower usually 3-nerved, the upper 3-5-nerved,
nerves evanescent. Lemma somewhat dorsally compressed, sub-cylindrical, grooved
ventrally, with involute margins, thinly to thickly coriaceous when mature, glabrous
and smooth or scabrid, 3-nerved, the nerves more or less parallel and not anastomosing,
hardly narrowed upwards and without an articulation, passing into three awns, each
furnished with a single nerve, with or without a column; column when present slightly
twisted, as thick as the body of the lemma; awns glabrous or scabrid only, thick and
coarse, spreading, collectively laxly spirally contorted at the base when mature, sub-
equal; callus well-developed, acute, densely but shortly bearded. Palea very small,
scale-like, usually coriaceous below with a sub-membranous apex, glabrous, dot sally
grooved, 2-nerved, apex obtuse. Lodicules 2, membranous, about as long, or slightly
longer than the palea, lanceolate, many-nerved. Stamens 3, anthers broadly linear.
Ovary glabrous; styles free, laterally exserted. Caryopsis tightly enclosed by the lemma
but free, dorso-ventrally compressed, deeply grooved on the ventral face; hilum linear,
nearly as long as the grain and situated in the groove; embryo the length of the
grain; starch grains compound, composed of 3-4 granules; first leaf of the seedling,
narrowly linear in outline (Fig. 152F).
Densely caespitose perennials, without rhizomes. Culms erect, usually simple,
solid. Leaf-blades linear, often involute towards the base, reddish brown when dry.
Ligule a dense fringe of long or short hairs. Panicles erect, narrow, rather contracted,
usually interrupted, with few spikelets, branches solitary or binate.
Anatomy
Shoot elongated and more or less circular in cross section, leaf-blades rolled in the bud (Fig.
152E).
Leaf-blade flat, widely V-shaped or horse shoe-shaped. Silicified cells of epidermis dumb-bell-shaped,
short to very long, usually accompanied by square to elongate-rectangular suberized cells with undulate
walls. Two types of hair present: unicellular broad-based retrorse barbs, which may be very short
to well developed and long; and linear bicellular hairs. Bundle sheaths: outer sheath of large thin-
walled parenchymatous cells containing chloroplasts; inner sheath devoid of chloroplasts, thin- walled
adaxially, gradually becoming more thick-walled towards the abaxial surface, the lower cells often
indistinguishable from the fibres of the stereome. Chlorenchyma of irregular cells forming a continuous
tissue between the bundles.
Awns: Column of the awns, when present, oblong to sub-circular in cross-section, grooved
ventrally with the edges overlapping, forming a thick-walled more or less cylindrical structure furnished
with three vascular bundles. Vascular bundles each flanked by two oblong groups of diffusely arranged
chlorenchyma cells. Chlorenchyma and vascular bundles embedded in a matrix of thick-walled cells.
Groove densely clothed with short bristle-like hairs. Stomata flanking the groove, and opposite groups
of chlorenchyma cells. Setae with only one vascular bundle, subtended by a single large group of
chlorenchyma cells, adjoining the epidermis ventrally. Matrix of setae consisting of small thick-walled
cells (Fig. 8, G & H)
Embryo in sagittal section showing the absence of an epiblast; cleft between the coleorrhiza and
base of the scutellum absent; a distinct elongation of the vascular tissue forming an internode between
the point of divergence of the scutellum bundle and the coleoptile. The cross section of the coleoptile
region shows the first embryonic leaf with five vascular bundles and with the margins meeting but not
overlapping; the coleoptile with three bundles (one median and two lateral) and the scutellum with
a single median bundle only (Fig. 152, H & I).
Karyology (Fig. 7; 20, p. 232)
Chromosomes (of S. jucunda) of the small type and with a basic number of n = 1 1 (2n = 22).
The karyology of the other two species is unknown.
Type species; Sartidia angolensis (C. E. Hubb.) de Winter.
382
The generic name Sartidia is an anagram of Aristida, which Sartidia resembles
closely in organographic characteristics.
A genus consisting of three African species, so far recorded from the Belgian
Congo, Angola, South West Africa and the Transvaal Province of South Africa.
2.9.3. Key to the Species Based on Organographic Characters
Lemma including the awns not exceeding 6 cm in length, body of the lemma finely scabridulous
in the upper part, chartaceous S. jucunda
Lemma including the awns 8-5-13 cm in length, body of the lemma smooth, thickly coriaceous:
Lemma 1-5-2-3 cm long, column absent; branches of the panicle smooth or nearly smooth;
basal leaf-sheaths estriate, smooth or nearly smooth, often somewhat glossy . . S. angolensis
Len-una up to 4 cm long including the well developed loosely twisted column; branches of the
panicle scabrid especially on the angles; basal leaf-sheaths dull and distinctly striate
S. vanderijstii
2.9.4. Key Based on Anatomical and Vegetative Characters
Stereome strands projecting into the first order bundles, partly or completely encircling the phloem
(Fig. 152A & B, p. 383) S. jucunda
Stereome strands in contact with the bundles but not projecting into the bundles or encircling
the phloem (Fig. 154, A & C, p. 386; 156, A & B, p. 388);
First order bundles usually 5; abaxial surface of leaf-blade hairy on the ribs, basal leaf-sheaths
estriate somewhat glossy S. angolensis
First order bundles 5-7; abaxial surface of leaf-blade with numerous scattered short retrorse
barbs, basal leaf-sheaths dull and striate S. vanderijstii
2.9.5. Description of the Species
Fig. 151. — Sartidia jucunda: A, lower glume, X 3; B, upper glume, X 3; C, lemma, X i; D
spirally contorted base of the awns, X 3; E, callus, X 8; F, palea, X 8; G, lodicules,
X 8; H, anthers, X 8; 1, gynoecium, X 8; J, caryopsis, showing hilum, x 7; K, caryopsis
showing embryo, X 7; L, cross section of caryopsis showing the grooved ventral side,
X 7
383
Fig. 152. — Sartidia jucimda: A, cross section of vascular bundles of the leaf-blade; B, diagram of
leaf-blade in cross section; C, abaxial epidermis of the leaf-blade; D, chlorenchyma cells
in longitudinal section; E, cross section of the shoots; F; seedling showing first leaf; G,
caryopsis showing the embryo; H, longitudinal section of the embryo; I, cross section
of the embryo (Codd 8686).
384
1. S. jucunda (Schweickerdt) de Winter in Kirkia 3; 137 (1963)
Aristida jucunda Schweickerdt in Bot. Jahrb. 76, 2: 221 (1954). Type: Transvaal,
Pietersburg distr. Blaauwberg, Schweickerdt 1807 (UPR, hole!, PRE, iso!). (Fig. 152)
Perennial, densely caespitose, up to 1 m high including the inflorescence, with
numerous slender intravaginal innovation shoots. Culms erect, usually simple, 2-3-
noded, finely striate, terete, glabrous; nodes included or exserted, glabrous. Leaf-
sheaths glabrous, striate, lax, the upper pallid or occasionally flushed with purple,
the lower usually a deep rusty brown, or rarely pallid, margins membranous free.
Ligule a fringe of long cilia; auricles long ciliate or glabrous; collar except for the
midrib with a line of very short bristles. Leaf-blade, when dry, a deep rusty brown,
linear, up to 45 cm long, narrowed towards the often convolute base, about 3-4 mm
wide, tapering to fine apex, scabrous on the upper surface, finely scaberulous below,
striate ; blades of innovations very fine, convolute. Panicle strongly exserted, narrow,
laxly contracted, much interrupted, up to 35 cm long with relatively few spikelets,
branches erect, solitary or binate, very unequal, bearing 1-6 spikelets upwards. Spikelets
deep rusty brown, shortly to long pedicillate, pedicels flattened on one side, margins
scabrous. Glumes rusty brown, glabrous, membranous, narrowly lanceolate, subequal,
up to 23 mm long, awn-tipped, finely scaberulous on the nerves and awn; upper 3-nerved,
scaberulous on nerves and awn. Lemma chartaceous, elongate, spindle-shaped, suffused
with purple, about 12 mm long, glabrous near the base, finely scaberulous all over in
the upper part, ventrally longitudinally furrowed (ripe grains often visible in the open
groove), gradually narrowed upwards passing into three long, flattened, scaberulous
awns, margins thinly membranous, involute ; callus subacute, 1 • 3 mm long, densely
bearded; awns subequal, up to 45 mm long, spreading, when mature collectively
laxly spirally contorted. Palea 2-nerved, cuneate, hyaline, apex erose, 2 mm long
and 1 mm wide. Lodicules 2, membranous, glabrous, many-nerved, up to 2 mm long.
Anthers 3, linear to about 6 mm long. Stigmas densely plumose, pale. Caryopsis
linear, somewhat compressed, glabrous, brown, ventrally grooved, about 10 mm long;
hilum linear grooved; starch grains compound, consisting of numerous granules.
Embryo 3-^ the length of the grain.
Anatomy (Fig. 152, A-D)
Shoot: Young leaves rolled in the bud (Fig. 152, E).
Leaf-blade: Stereome strands projecting into the first order bundles, partly or completely encircling
the phloem and interrupting the outer bundle sheath. First order bundles 3-5; adaxial surface with
relatively few hairs; abaxial surface with very few hairs. Otherwise as described for the genus.
Awns as described for the genus (Fig. 8 G-H, p. 235).
Embryo as described for the genus (Fig. 152 H-I).
Transvaal. — Pietersburg' Blaauwberg above Leipzig Mission Station, Codd 8686;
v.d. Schi iff 537S', Strey & Schlieben 8502; Blaauwberg, Malaboch’s kraal, Schweickerdt
1820; 2019; 1807 (type).
This species has been collected only on the Blaauwberg in the Pietersburg district
of the Transvaal. It is possible that it also occurs in the Soutpansberg and other
mountains of the northern Transvaal since suitable habitats are undoubtedly present
in these ranges. It occurs on stony slopes at a altitude of 1,300-2,000 metres.
Though undoubtedly closely related to the other two species of the genus, S.
jucunda is distinguished by much smaller spikelets and scabrid lemmas.
2. S. angolensis (C. E. Hubb.) de Winter in Kirkia 3: 137 (1963)
Aristida angolensis C. E. Hubbard in Kew Bull. 1949: 359 (1949). Type: Angola:
Benguella, Gossweiler 4099a (holo, K). (Fig. 153).
385
Fig. 153. — Sartidia angolemis: A, lemma X B, callus X 6; C, contorted basal parts of the awns
X 2^; D, lower glume X 2V; E, upper glume X 2^-; F, palea X 6; G, lodicules X 60;
H, anthers X 6; 1, gynoecium X 6.
Perennial, densely caespitose up to 1-2 m high with numerous intravaginal inno-
vations. Culms erect about 2 ram in diameter, simple, very firm, solid, terete, glabrous,
finely striate; nodes mostly included; ultimate internode long exserted. Leaf-sheaths
glabrous, estriate and often with a dull sheen in the lower part, lax, finely puberulous
and striate upwards, the basal imbricate and persistent, pallid or pale brown, margins
free, firmly membranous. Ligule a dense fringe of short silky cilia; auricles bearded;
collar except for midrib with a line of long to short bristles. Leaf-blade when dry
a deep brown, hnear, up to 35 cm long, narrowed and convolute below, 2-4 mm wide,
tapering to a fine apex, pilose above the ligule and densely scaberulous on the upper
surface, finely scaberulous below. Panicle stiff and erect, narrow, laxly contracted,
up to 40 cm long including the awns, and with relatively few spikelets; rhachis striate,
compressed, sub-glabrous; branches solitary or binate, erect simple or little divided,
compressed, sub-glabrous or finely scaberulous upwards, bearing 1 to few spikelets
in the upper part, the lower most about 10 cm long, shorter upwards; pedicels smooth,
the lateral 3-8 mm long, the terminal elongated. Spikelets pale brown to pallid.
Glumes narrowly linear-lanceolate, firmly chartaceous below, tapering to a membranous
acute apex, subequal or slightly unequal; the lower 2 -5-3 -3 cm long, 5-nerved; the
upper about 2-5 cm long, 3-nerved. Lemma linear, coriaceous, 15-23 mm long
excluding the callus, glabrous (very finely scaberulous under a strong lens), without
a column or articulation, gradually passing into three long bristles, margins membranous
and involute in the lower part; callus elongate, acute 3-5-4 mm long, densely and
shortly bearded; bristles 3, scaberulous, flattened and shallowly grooved on the inner
surface, subequal, 7-5-10 cm long, spreading, when mature collectively laxly spirally
contorted at the base. Palea oblong, about 3 mm long, coriaceous below, membranous
upwards, ventrally grooved. Lodicules about 3 mm long, lanceolate, membranous,
many-nerved. Caryopsis linear, 9 mm long, very dark brown; hilum linear, nearly
as long as the grain, forming a narrow shallow groove; starch grains compound,
composed of numerous 3-4 granules; embryo \ the length of the grain.
386
Fig. 154. — Sartidia angolensis: A, cross section of vascular bundles of the leaf-blade; B, abaxial
epidermis of the leaf-blade; C, diagram of leaf-blade in cross section {de Winter 2779).
Anatomy (Fig. 154, A-C)
Leaf-blade: stereome strands in contact with the bundles but not or only partly interrupting
the outer sheath and not projecting into the bundle nor enveloping the phloem; first order bundles
usually 5; adaxial surface with very few hairs; abaxial surface usually densely covered with hairs
opposite the bundles. Otherwise as for the genus.
Awns as described for the genus.
Embryo as described for the genus.
387
Northern Rhodesia. — Kafue National Park, Ngoma, Mitchell 116.
South West Africa. — Otjiwarongo : top of the Waterberg Plateau, de Winter 2779.
Grootfontein : 30 m N. of Gautscha Pan, Story 6473.
A species found on Kalahari sand in north western South West Africa. It also
occurs in N. Rhodesia and Angola.
Henrard in his Monograph included specimens of this species in Aristidci vanderijstii
de Wild. This error was recognized by Hubbard who pointed out the differences,
including the characteristic absence of a column in S. angolensis. The South West
African specimens, collected many hundreds of miles from the type locality, agree
very well with the type and constantly lack a column. This confirms Hubbard’s
opinion that Henrard was incorrect in stating that the lack of the column in the
Angolan specimens was due to the immaturity of the material. As pointed out by
Hubbard, in Aristida species the awns and column develop befoie the body of the
lemma and this probably also holds good for Sartidia species.
S. angolensis is very closely allied to S. vanderijstii. The differences by which
the species may be distinguished are fully set out in the key to the species.
3. S. vanderijstii {De Wild.) de Winter in Kirkia 3: 137 (1963)
Aristida vanderijstii De Wild, in Bull. Jard. Bot. Brux. 6: 40 (1919) (Fig. 155). Types:
Congo: Nyungu, Vanderijst 3216; Mukulu, Vanderijst 3182 (syn., BR.).
Fig. 155. — Sartidia vanderijstii: A, lower glume, X 2f; B, upper glume, X 2); C, lemma, X 2)-;
D, callus, X 6; E, spirally twisted basal part of the awns, 2i; F, palea, X 6; G, lodicules,
X 6; H, anthers, X 6; I, gynoecium, x 6.
Perennial, densely caespitose, up to 80 cm high including the panicle. Culms
erect, usually simple, 1-3-noded, finely striate, glabrous or slightly scaberulous, terete,
solid, nodes included or rarely exserted, glabrous; upper internode long exserted.
388
Leaf-sheaths striate, rather dull, scaberulous, lax, pallid to pale reddish brown, margins
membranous, ciliate, free. Ligule a dense fringe of cilia; auricles glabrous or bearded
with long cilia; collar usually glabrous. Leaf-blade when dry a deep reddish brown.
Fig 156.— Sartidia vanderijstii : A, cross section of vascular bundles of the leaf-blade; B, diagram
of leaf-blade in cross section ; C, abaxial epidermis of the leaf-blade {Callens s.n.).
389
linear, up to 20 cm long, narrowed towards the base, 2-3 mm wide, tapering to a fine
apex, scabrous on the upper surface, finely scaberulous below, striate. Panicle strongly
exserted, erect, narrow, laxly contracted, much interrupted, up to 20 cm long, with
relatively few spikelets, rhachis scabrid, branches erect, flattened, solitary or binate,
scabrous, bearing 1-6 spikelets in the upper part. Spikelets deep reddish brown,
rarely pallid. Glumes linear-lanceolate, fiimly chartaceous below, tapering to a sub-
membranous acute apex, subequal or unequal, glabrous (finely scaberulous under a
strong lens), not awned; the lower 30-40 mm long, 5-nerved; the upper 25-30 mm
long, 3-nerved. Lemma linear, coriaceous convolute, 3-nerved, 40-60 mm long
including the slightly twisted column but not the callus, glabrous (very finely scaberulous
under a strong lens), margins membranous and involute in the lower part, column
loosely twisted, up to 40 mm long, without an articulation ; body of the lemma gradually
passing into the column which is terminated by 3 thick, coarse bristles; callus elongate,
acute, 3-3-5 mm long, densely but shortly bearded; bristles 3, scaberulous, flattened
and somewhat grooved on the inner surface, subequal, 8-9-5 cm long, spreading,
collectively laxly spirally contorted at the base when mature. Palea oblong, about
2-5 mm long, ventrally furrowed, coriaceous below with a membranous obtuse apex,
nerves two, short, evanescent. Lodicules about 2 - 5 mm long, membranous, lanceolate,
many-nerved. Caryopsis linear, 8-9 mm long, dark brown; hilum linear, nearly as
long as the grain, grooved; starch grains compound, composed of 3-4 granules;
embryo ^ the length of the grain.
Anatomy (Fig. 156 A-C)
Leaf-blade: stereome strands in contact with the bundles but not interrupting the outer bundle
sheaths nor projecting into the bundle; first order bundles 7-9; adaxial surface with a few short rather
abruptly acute hairs; abaxial surface with numerous rather short, blunt retrorse barbs. Otherwise
as described for the genus.
Awns as described for the genus.
Embryo as described for the genus.
Congo. — Leopoldville: Gombe ya Turn ba, Co////z5 3411 ; Katanga: Risopaulos
358.
S. Yanderijstii has so far been recorded only from the Congo Republic where it
occurs in grassland and bush-savannah on sandy soils. The differences by which it
may be distinguished from the closely allied S. angolensis are fully set out in the key
to the species.
PART 3
GENERAL CONSIDERATIONS
3.1. General Discussion of the Leaf Anatomy of the Aristideae
The anatomical details included in the taxonomic descriptions are based only on
the study of scrapes of the abaxial epidermis and cross-sections of the leaf-blade. For
this reason a general account giving further details of the anatomical structure has
been drawn up. In order to arrive at a better understanding of the anatomy, a thorough
study was made ot a few representative species where, in addition to the epidermal
scrapes and cross sections prepared for all species, the leaf in longitudinal section,
and the leaf-tissues in a macerated state, were included.
Stipagrostis (Fig. 157)
The Epidermis. The long elements of the abaxial epidermis seen in surface view
(Fig. 157, H) usually have strongly undulate, anticlinal walls, and cross sections show
that the outer (tangential) walls are strongly thickened. In longitudinal section under
5752152-7
390
high magnification the strong thickening of the outer walls is very evident and in this
thickened portion faint concentric complete or interrupted circles can be seen. These
concentric rings are the undulations of the anticlinal (radial) walls, which so clearly
show up in suiface view, seen end on (Fig. 157 F). From, this it may be inferred that
the faces of the anticlinal walls (those parallel to the longitudinal axis of the leaf) are
strongly undulate m.ainly in their upper part, i.e. the area approaching the outer
tangential wall of the cell concerned; the undulations, however, gradually become
weaker and finally disappear as the anticlinal walls “ approach " the inner tangential
wall of the cell. The long ripple-walled cells of the adaxial epidermis are thin-walled
and the anticlinal walls usually much less undulate. In some cases, however, their
inner tangential walls are thickened whereas their outer walls are thin. Two types
of epidermal hair occur; (1) Unicellular retrorse barbs more or less appressed to the,
surface of the epidermis occur in most species both ad- and abaxially. There is, however,
a marked tendency for a more copious production of hairs adaxially and these hairs
are usually less appressed to the surface thus forming a dense bristly layer. Appressed
unicellular hairs ate also present adaxially. In surface view these hairs are broadly
oblong in outline when very short, and the apex can often be distinguished only with
difficulty. They vary greatly in shape and length, viz. abruptly produced into an
acute apex, to attenuate with an acuminate, very fine, sharp apex. In longitudinal
section it is clearly seen that the base is often narrowed and inserted between the
adjoining narrow ends of the long ripple-walled elements. In the species with long
woolly hairs, the bases of these hairs are also naiTOwed, occupy the same position,
and should therefore be regarded as homologous with the papilla-like hairs found
Fio. 157. — Anatomical details of Stipagrostis: S. iiniplnmis var. neesii: A, one-celled retrorse barb
in longitudinal section in the stomatal area of abaxial epidermis; B, l.s. of motor cells;
C, adaxial epidermis, adjacent to the stereome, in longitudinal section showing one-celled
hair (OH), silicified cells (SC) and suberized cells (CC); D, l.s. of adaxial epidermis showing
retrorse hairs subtended by parenchyma cells; E, l.s. of bicellular hair in the stomatal
area of abaxial epidermis; F, l.s. of abaxial epidermis in silicified cell zone showing
silicified cells (SC) accompanied by suberized cells (CC) as well as long ripple-walled cells
(LC) and subtended by fibres; G, l.s. of bundle sheaths and chlorenchyma cells (de Winter
3204). S. dregeana: H, surface view of stoma and guard cells in abaxial epidermis, showing
large papillae, formed by projections of the walls of the adjacent long ripple-walled cells,
which project over the guard cells (Kinges 2572). S. iiniplnmis var. neesii: J, cell types:
1. chlorenchyma, 2. parenchyma cell supporting the chlorenchyma, 3. thick-walled pitted
parenchyma cell from the xylem; 4. small mx vessel; 5. ring vessel from px; 6. large
pitted vessel from mx; 7. transition fibre; 8. fibre; 9. inner bundle sheath cells.
391
in other species. A complete gradient between these extreme types is found (Fig.
157, A). (2) Hairs of the hiceUular linear type: The bases of these hairs are likewise
inserted between the adjoining narrow ends of long ripple-walled cells. In contrast
to the unicellular hairs they are relatively thin-walled and terete; they lie closely
appresscd to the surface of the epidermis (Fig. 157, E). Tht’sHicified cells are strongly
impregnated with silica. In longitudinal section the silicihed cells are usually about as
wide as deep, the anticlinal walls parallel or sloping inwards towards the inner walls so
that the cells are narrowed towards the base (Fig. 1 57, C, F). The inner tangential walls
as well as the anticlinal walls are much thickened, decreasing in thickness towards the
outer thin tangential wall which is dom.eshaped (Fig. 157, F).
In a few annual species the silicihed cells are dumb-bell-shaped in surface view
and thus agree with the silicihed cells of Aristida in general structure (see p. 392).
The silicihed cells in both the adaxial and abaxial epidermis are similar in shape and
size. They are usually more numerous in the abaxial epiderm.is. The suherized cells
can be recognized by the usually deeper staining of the contents of these cells which
in the mature epidermis often have retained their nuclei. They are thin-walled
throughout. In the m.ajority of cases they are found in association with the silicihed
cells. Where they are small and associated with silicihed cells which are narrowed
towards the inner walls they are partially hidden and often appear kidney-shaped in
surface view (Fig. 157, C, F). In the stomatal apparatus the cells subsidiary to the
guard cells are triangular (Metcalfe 1960, p. 667). In a few species the thick outer
tangential walls of the adjacent ripple-walled long cells are produced into large papilla-
like struetui'es projecting over the stomatal apparatus partially hiding the component
cells (Fig. 157, H). Examples of this are found in the abaxial epidermis of S. obtnsa,
S. lanipes and S. dregeana. This phenomenon has also been reported by other authors
for species such as ciliata and S. pangens.
The motor cells are large parenchymatous cells and do not exhibit any striking
peculiarities (Fig. 157, B).
The chlorenchyma, however, presents an entirely different picture in longitudinal
section compared with that in transverse section. The cells are elongated with the
tangential walls deeply undulated thus creating a laige number of intercellular spaces.
The circular structures observed in these cells, when studied in cross section, represent
the constricted portions of the cells seen in end view. This may be characteristic for
all the tribes showing a radial arrangement of the chlorenchyma (Fig. 157, G). The
outer sheath of the vascular bundles, which lies adjacent to the chlorenchyma, is
composed of cells which are either thin-walled or the walls of these cells may be
somewhat thickened. These cells are well supplied with chloroplasts and are larger
or smaller than the cells of the inner sheath. Study of longitudinal sections shows
that they are fairly short and occasionally approximately isodiametric. The length
of the longitudinal walls, however, usually exceeds their width (Fig. 157, G). The
cells of the inner sheath are usually rather elongated, thin-walled and devoid of
chloroplasts. They are parenchymatous cells with slightly, or oceasionally fairly
strongly thickened walls. These are traversed by simple or funnel-shaped pits; if
funnel-shaped, the openings are slit-like and lie obliquely.
The Xylem. The protoxylem vessels are usually either ring or spiral vessels and
are partially or wholly surrounded by thin-walled elongate proto-parenchyma cells.
The large metaxylem vessels are somewhat angular with transverse elongate bordered
pits on the longitudinal walls; the transverse walls of the original component cells
of the vessels are usually only slightly oblique, often with only a single large perforation,
and thus almost absent. The smaller cells between and adjoining the large metaxylem
vessels in longitudinal section, are vessels of various sizes with the remams of the
transverse walls usually more oblique than in the large metaxylem vessels, the ends
with one very large or a few fairly large perforations (Fig. 157, J, 3, 4, 5, 6). The cells
surrounding the phloem are fibres or occasionally parenchyma' cells with somewhat
thickened walls. Where pai'enchyma cells adjoin tire xylem vessels the walls ai'e pitted
complemeirtarily.
The adaxial stereome consists mostly of long, very thick-walled fibi'es, likewise
the abaxial stereome but, in a few species the abaxial stereome consists of large cells
which in longitudinal section are seen to be elongate somewhat thick-walled paren-
chymatous cells. Much elongated narrow cells, intermediate in shape between fibres
and parenchyma cells, frequently occur among the thick-walled parenchymatous cells
just mentioned (Fig. 157, J, 7).
Aristida (Fig. 158)
Aristida largely agrees with Siipagrostis in its general anatomical features, and the
description of the latter thus also applies to the former except for the bundle sheaths
and silicified cells. Both the outer and the inner bundle sheaths of Aristida contain
chloroplasts in contrast to Stipagrostis, in which only the outer sheath contains these
bodies. The silicified cells of Aristida are almost invariably dumb-bell-shaped varying
from short and broad to much longer than wide with a slender elongated central
constriction. In a few species, however, such as A. sciurus and A. dasydesmis the
silicified cells frequently are sub-circular. In species where the silicified cells are well
developed they often protrude strongly above the surface of the epidermis (Fig. 158, H).
Fig. \5%.— Anatomical Details of Aristida: A. mollissima: longitudinal section of: A, motor cells;
F, abaxial epidermis cells and stoma; C, adaxial epidermis showing curved unicellular
hairs sheathed at the base by enlarged suberized cells; D, bundle sheath cells (OS and
ICS), fibres (S) and xylem vessels; E, bicellular appressed hair, and outer walls of epidermal
cells showing circular patterns caused by undulate walls; F, A. hordeacea, inner and outer
siieath cells (ICS and OS) and chlorenchyma cells (C) seen in longitudinal view (macerated
material); G, A. barhicollis, longitudinal section through a third order vascular bundle:
AB, abaxial epidermis; AD, adaxial epidermis; BH, bicellular hair; CC, suberized cell;
C, chlorenchyma; ICS, inner bundle sheath; LC, long ripple-walled cell; OH, one-cel led
hair; OS, outer bundle sheath; P, ohloem; S, stereome; X, xylem. A. hordeacea,
macerated material: H, longitudinal section of silicified cell; I, silicified cell seen from below
showing surfaces of attachment to other cells; J, epidermis in surface view showing ripple-
walled and silicified cells.
393
The suberized cells on the adaxial epidermis are often very strongly developed adjacent
to unicellular hairs, forming a sheath-like structure enclosing the base of the hair
(Fig. 158, C, G). These cells are also described and depicted by Caceres (1961, p. 9).
In Fig. 158 various tissues and individual cells as seen in longitudinal section are
shown.
3.2. SiLICIFICATlON OF THE LEAF-EpIDERMIS IN THE ArISTIDEAE
It is well known that grass leaves are often difficult to section due to silica deposited
in the tissues. In order to obtain more information on the distribution of silica in
the leaves of the Aristideae, material of several species was treated with acid to remove
all organic material (see p. 204). The residue remaining after this treatment can readily
be dissolved in hydrofluoric acid, and therefore represents the silicified parts of the
leaf.
In some cases practically the whole of the epidermis was found to be silicified
and fairly large pieces of recognizable epidermal “skeletons” were present in the
residue. Two types of silicification were observed, viz. the impregnation of cell walls
with silica, and the formation of solid or hollow “ silica bodies ” within cell cavities.
The former was fairly general for all types of epidermal cells: even the cell walls of
stomata were occasionally observed to be silicified, but the silica skeletons are extremely
fine and fragile. The second type of silicification is found in “ silicified cells ” and some
long ripple-walled cells. Here the cells contain silica bodies, which may be hollow
or solid, in appeal ance and which fit tightly into the cell cavity. These bodies can be
dissected from the cells leaving the cavity lined by the cell walls. In epidermal scrapes
glass-like fractures are occasionally visible in the silica bodies. These fractures are
probably caused by pressure exserted during scraping and do not extend to the cell
walls as would be expected if the whole of the cell was silicified. This investigation
has also indicated that the tissues of species growing in very dry areas are often more
highly silicified than those of species growing under more temperate conditions.
SUMMARY
1. Anatomical, cytological and organographic data indicate that the Stipeae is a
tribe of uncertain affinity.
2. The Stipeae should be limited to the genera Stipa, Oryzopsis, Piptochaetium
and Trikeraia, of which only Stipa occurs in South Africa.
3. Of the four species of Stipa previously recorded from South Africa, only two
are true species of Stipa, the other two being one-awned representatives of the Aris-
tideae.
4. Stipa parvula Nees, is transferred to Aristida under the designation of Aristida
parvula (Nees) de Winter, and Stipa riamaquensis Pilger to Stipagrostis under the
designation of Stipagrostis anomala de Winter.
5. The differences between the Aristideae and Stipeae are discussed. The Aristideae
was justifiably separated from the Stipeae by Hubbard.
6. The phylogenetic position of the Aristideae is uncertain, but there are strong
indications that this tribe is distantly related to the Eragrosteae.
1. The inclusion of Amphipogon and Diplopogon in the Aristideae (Pilger,1956)
cannot be supported. Both these genera could probably be better accommodated in
the Danthonieae.
5752152-8
394
8. The genus Aristida sensu lato is divided into three genera viz. Aristida (awns
glabrous), Stipagrostis (awns plumose) and the new genus Sartidia.
9. The genus Sartidia differs from Aristida and Stipagrostis mainly in the structure
and size of the mature embryo, the anatomical structure of the leaf-blade and the
rolled position the leaf occupies in the bud. The genus at present consists of three
species, viz. 5. (Schweick.) de Winter, S. angolensis (C. E. Hubb.) de Winter,
and S. vanderijstii (De Wild.) de Winter, all previously included in section Aristida
{Chaetaria) of the genus Aristida L.
10. The leaf-anatomy and organography of all the South African species of the
Stipeae, Aristideae and the three species of the new genus Sartidia are described. In
addition the embryos of representative species were studied anatomically.
11. The chromosome number of several species of Aristida and Stipagrostis, and
one species each of Stipa and Sartidia is recorded. The majority of Aristida species
are diploids (2n = 22), while most Stipagrostis species are tetraploids (2n = 44).
Sartidia jucunda (Schweick.) de Winter has 2n = 22, and Stipa dregeana Nees has
2n = 48.
12. Keys based mainly on floral characters and keys based on a combination of
anatomical and vegetative characters, are given for the South African species of the
genera dealt with.
13. Natural groups of allied species can be recognized in Aristida and Stipagrostis
by correlating anatomical and organographic characters. These groups do not in all
cases agree with the generally recognized sections of Aristida sensu lato.
14. In both Aristida and Stipagrostis the absence or presence, and relative position
of the articulation of the lemma, when used for the delimitation of sections, occasionally
results in unnatural separation of closely allied species. For practical reasons the
division of genera into sections based on the absence or type of articulation ot the
lemma and awn, has, however, been retained.
15. Three sections are recognized in the genus Stipagrostis, namely Schistachne,
Stipagrostis and Anomala, the latter being a monotypic section described by the present
author to accommodate the single-awned Stipagrostis anomala de Winter {= Stipa
namaquensis Pilger).
16. Six sections are recognized in the genus Aristida namely Schizaclme, Pseudo-
chaetaria, Aristida, Arthratherum, Pseudarthratherwn and Streptachne. Schizaclme
described by Trin. & Rupr. as a section of Stipa to accommodate the single species
Aristida parvula (Nees) de Winter {=Stipa parvula Nees), is transferred to Aristida
as a new section Schizaclme (Trin. & Rupr.) de Winter.
17. Theron, in his classification of the South African species of Aristida sensu
lato based on anatomical features of the leaf-blade, did not achieve a natural grouping
of species mainly because he overlooked the characteristic differences in the bundle
sheaths of species with plumose awns (Stipagrostis) and species with glabrous awns
(Aristida).
18. Silica is deposited in the epidermis of the Aristideae in two ways, viz. as bodies
formed within the cell cavities, and secondly as an impregnation of the cell walls.
395
Ffg. 159. — Diagram of leaf-blades in cross section: 1, /I. parvM/a (Schweickerdt 2220); 2, A. hordeacea
(de Winter 3097); 3, A. hubbardiana iSch'WQicV.erdi 2111); 4, A. rhiniochloa (Codd 5977);
5, A. adscensionis var. guineensis (de Winter and Codd 574); 6. A. effusa (de Winter 3526);
7, A. citrvata (de Winter 2437a); 8, A, scabrivahis (de Winter 775); 9, A. bipartita
(Acocks 9563).
10. A. canescens (de Winter 795 A); 1 1, A. aequiglumis (de Winter 7515); 12, A. transvaalensis (de
Winter 7563); 13, A. monticola (Mogg PRE 20634); 14, A. junciformis subsp. Junciformis
(Baker 9); 14a, A. junciformis subsp. galpinii (Galpin 6900); 15, A. recta (Schweickerdt
1677); 16, A. sciurus (Pentz and Acocks 10314); 17, A. dasydesmis (Pearson 3487);
18, A. diffusa var. diffusa (Schlechter 9138); 18a, A. diffusa var. diffusa (var. pseudohystrix)
(Zeyher 447); 18b, A. diffusa var. burkei (de Winter 795),
19. A. vestita (de Winter 3584); 20, A. engleri var. engleri (Pole Evans 80); 21, A. spectabiUs (de Winter
773); 22, A. meridionalis Pole Evans 3307 (3)1; 23, A. stipoides (Schweickerdt 2129);
24, A. mollissima (Murray 2); 25, A. argentea (&hweickerdt 1806); 26, A. stipitata (de
Winter 2458); 26a, A. stipitata var. graciliflora (de Winter 752); 27, A. pilgeri (de Winter
2730); 28, A. congesta (de Winter 5986); 28a, Subsp. barbicollis (de Winter 791); 28b
Subsp. congesta (A. longicauda) (Moss and Rogers 608).
575215?
397
Fkj. 160. — Diagram of leaf-blades in cross section: 1, S. hochstetteriana var. secalina (de Winter
3408); la, S. hochstetteriana var. hochstetteriana (de Winter 2660); 2, S. ciliata var. capensis
(Keet 1697); 2a, S. ciliata var, capensis (weak form) (de Winter 6163); 2b, S. ciliata x S.
zeyheri (Acocks 14817); 3, S. schaeferi (Volk 26); 4, S. dinteri (Galpin & Pearson 7416);
5, S. proxima (Flanagan 1157); 6, S. amabilis (de Winter 3398); 7, S. nanuiquensis (de
Wintei" 3883); 8, S. damarensis (Moss Herb 17837); 9, 5. sabulicola (Keet 1612).
10, S. lutescens var. marlothii (Strey 2445); 11, 5. brevifolia (Adamson 1538); 11a, S. brevifolia x S.
namaquensis (de Winter 3266) ; 1 2, 5. fastigiata (de Winter 3573) ; 1 3, S. geminifolia (Marloth
12415); 14, S. garubensis (Kinges 2289); 15, S. dregeana (Kinges 2572); 16, S. zeyheri
subsp. zeyheri (van Breda 3); 16a, S. zeyheri subsp. barbata (Codd 9240); 16b, S. zeyheri
subsp. macropus (Bolus 12910); 16c, S. zeyheri subsp. sericans (Story N.H, 24574); 16d,
S. zeyheri subsp. sericans (CilUers 11).
17. S. gonatostachys (Dinter 1022); 18, S. lanipes (de Winter 6133); 19, S. obtusa (Krapohl H. 21826).
20, S. uniplumis var. uniplumis (Acocks 14330); 20a, S. uniplumis var. uniplumis (var
pearsonii) (de Winter 3410); 20b, S. uniplumis var. neesii (de Winter 793); 20c, S.^iplumis
var. neesii x S. hirtigluma var. patula (de Winter 2778); 21, S. hirtigluma var. hirt^luma
(Trott 134); 21a, S. hirtigluma var. patula (de Winter 3527); 22, S. hermannii (Kmges
2634); 22a, S. namibensis (Keet 1698); 23, S. subacaulis (Kinges 2633); 24, S. anomala
(Scott s.n.); 25, S. ranmlosa (de Winter & Hardy 8197),
75215>
I
i
I
i
399
REFERENCES
ACOCKS, J. P. H., 1953. Veld types of South Africa. Dept, of Agric. S. Afr., Bot. Survey Mem.
No. 28: 1-192.
AVDULOV, N. P., 1931. Karyo-systematische Untersuchung der Familie der Gramineen Bull. Appl.
Bot. Genet, and Plant Breed. Suppl. 44: 354-425 (German summary).
BENTHAM, G., 1882. Notes on Gramineae. J. Linn. Soc. Bot. 19: 14-134.
BENTHAM, G. & HOOKER, J. D., 1883. Genera Plantarum 3, 2. London: Reeve & Co.
BOR, N. L., 1954. Notes on Asiatic Grasses. Kew Bull. 1954: 22.
BOR, N. L., 1960. The Grasses of Burma, Ceylon, India and Pakistan. Oxford: Pergamon Press.
BROWN, W. V„ 1951. Chromosome Numbers of Some Texas Grasses. Bull. Torrey Bot. Club 78,
4: 292-299.
BROWN, W. V., 1958. Leaf Anatomy in Grass Systematics. Bot. Gazette 119, 3: 170-178.
BROWN, W. V. & EMERY, W. H. P., 1957. Persistent Nucleoli and Grass Systematics. Amer.
J. Bot. 44: 585-589.
BUGNON, P., 1921. La feuille chez les Graminees. Thesis, Caen.
BURBIDGE, N. T., 1946. Foliar Anatomy and the Delimitation of the Genus Triodia R. Br.
Blumea, suppl. 3: 83-89.
BURBIDGE, N. T., 1953. The Genus Triodia. Australian J. Bot. 1, 1: 121-184.
CACERES, M. R., 1961. Consideraciones relatavas a la anatomia foliar de las especies de Aristida
de Mendoza. Rev. Argent. Agron. 28, 1 : 1 (1961).
CHEADLE, V. L, 1955. The Taxonomic use of Specialization of Vessels in the Metaxylem of
Gramineae, Cyperaceae, Juncaceae and Restionaceae. J. Arnold Arb. 36: 141.
CHIPPINDALL, L. K. A., 1955. Stipeae in Meredith ef a/. Grasses and Pastures of South Africa
287-317. Central News Agency.
CONN, H. J. & DARROW, M. A., 1946. Staining Procedures Used by the Biological Stain Com-
mission. Biotech. Publications, Geneva, N.Y., U.S.A.
COVAS, G. & BOCKLET, M., 1945. Numero de Cromosomas de algunas Gramineae — Stipinae de
la Flora Argentina. Rev. Argent. Agron. 12: 261-265.
DARLINGTON, C. D. & WYLIE, A. P., 1955. Chromosome Atlas of Flowering Plants, Second
Ed. London: Allen & Unwin.
DE WET, J. M. J., 1954. Chromosome Numbers of a few South African Grasses. Cytologia, 19:
97-103.
DE WET, J. M. J., 1958. Additional Chromosome Numbers in Transvaal Grasses. Cytologia, 23:
113-118.
DE WET, J. M. J., 1960. Chromosome Numbers and some Morphological Attributes of various
South African Grasses. Amer. J. Bot. 47: 44^9.
DE WET, J. M. J. & ANDERSON, L. J., 1956. Chromosome Numbers in Transvaal Grasses.
Cytologia, 21 : 1-10.
DUVAL-JOUVE, M. J., 1875. Histotaxie des Feuilles de Graminees. Ann. Sc. Nat. Bot. Ser. 6,
14: 294-371.
ELIAS, M. K., 1942. Tertiary Prairie Grasses and other Herbs from the High Plains. Geol. Soc.
Amer., Special Papers 41: 9 122.
ENGLER, A., 1921. Die Pflanzenwelt Afrikas, 3, 2: 851-852.
EVANS, R. L, 1949. Delafields Hematoxylin Combined with Safranin for Anatomical Preparations
of Plant Parts. Stain Tech. 3: 191.
FIGARI & DENOTARIS, 1851. Mem. Acad. Sci. Torino 2, 12.
GARDNER, C. A., 1952. Flora of Western Australia 1, 1, Gramineae, Perth: Government Printer.
GOULD, F. W., 1958. Chromosome Numbers in South Western Grasses. Amer. J. Bot. 45 : 757-767.
GUENZEL, F., 1913. Blattanatomie Suedwestafrikanischer Graeser. Bot. Jahrb. 48, Beibl. 108:
1-55.
GUENZEL, F., 1921. Weitere Beitraege zur Kenntnis der Blattanatomie afrikanischer Graeser.
Bot. Jahrb. 57, Beibl. 126: 1-26.
HACKEL, E., 1896. The True Grasses. Westminster: Archibald Constable & Co.
400
HARTLEY, W., 1958 (a). Studies on the Origin, Evolution and Distribution of the Gramineae.
1. The Tribe Andropogoneae. Austr. J. Bot. 6: 116-128.
HARTLEY, W., 1958 (b). Studies on the Origin, Evolution and Distribution of the Gramineae.
2. The Tribe Paniceae. Austr. J. Bot. 6: 343-357.
HARTLEY, W. & SLATER, C., 1960. Studies on the Origin, Evolution and Distribution of the
Gramineae. 3. The Tribes of the Subfamily Eragrostoideae. Austr. J. Bot. 8: 256-276.
HENRARD, J. TH., 1926-1933. A Critical Revision of the Genus Aristida. Med. Rijksherbarium
Leiden No. 54 (1926), 54a (1927), 54b (1928), 54c (1933).
HENRARD, J. TH., 1929-1933. A Monograph of the Genus Aristida. Med. Rijksherbarium
Leiden No. 58 (1929), 58a (1932), 58b (1933).
HITCHCOCK, A. S., 1950. Manual of the Grasses of the United States. Ed. 2. Washington:
Government Printing Office.
HOLM, T., 1901. Some New Anatomical Characters for Certain Gramineae. Bot. Centralbl. 11,
Beih 2: 101-133.
HUBBARD, C. E., 1934. Gramineae, in Hutchinson, Earn. El. Plants 2; 199-229. London:
MacMillan & Co.
HUBBARD, C. E., 1948. The Genera of British Grasses. Reprint from Hutchinson, J., British
Flowering Plants.
HUBBARD, C. E., 1959. Gramineae in Hutchinson, Earn. FI. Plants Ed. 2, 2: 710-741. Oxford:
Clarendon Press.
HUBBARD, C. E. & VAUGHAN, R. E., 1940. The Grasses of Mauritius & Rodrigues. London:
Crown Agents for Colonies.
HUGHES, D. K., 1921. A Revision of the Australian Species of Stipa. Kew Bull. 1921: 1-30.
HUNTER, A. W. S., 1934. A Karyosystematic Investigation in the Gramineae. Canadian J. Res.
11: 213-241.
JACQUES-FELIX, H., 1958. J. Agric. Trop. Bot. appl.
JELENC, F., 1950. Anatomic de la Tige et de la Feuille des Aristida de L’Afrique du Nord. Bull.
Soc. Sci. Nat. Maroc. 30: 3-21.
KENNEDY, P. B., 1899. The Structure of the Caryopsis of Grasses with Reference to their Mor-
phology and Classification. U.S. Dept. Agric. Div. Agrost. Bull. 19: 1^4.
KINGES, H., 1961. Merkmale des Gramineenembryos. Bot. Jahrb. 81, 1, 2: 50-93.
KUNTH, C. S., 1813. Enumeratio Plantarum 1: 175-198.
KUNTH, C. S., 1833. Enumeratio Plantarum 1: 175-198.
KUNTH, C. S., 1835. Enumeratio Plantarum Suppl. 131-154.
LOMMASSON, R. C., 1957. Vascular Bundle Sheaths in the Genus Aristida. Phytomorphology 7:
364-369.
METCALFE, C. R., 1960. Anatomy of the Monocotyledons. 1. Gramineae, 1-731. Oxford:
Clarendon Press.
MOFFETT, A. A. & HURCOMBE, R., 1949. Chromosome Numbers of South African Grasses-
Heredity 3, 3: 369-373.
NEES AB ESENBECK, C. G., 1829. Agrostologia Brasiliensis 371-390. Stuttgart.
NEES AB ESENBECK, C. G., 1832. Plantae Ecklonianae. Linnaea 7: 290.
NEES AB ESENBECK, C. G., 1841. Flora Africae Australioris.
PARODI, L. R., 1944. Revision de las Gramineae Australes Americanas del Genero Piptochaetium.
Revista Mus. de la Plata, Bot. (Nueva serie) 6: 213-310.
PARODI, L. R. & FREIER, F., 1945. Observaciones Taxonimicas sobre las Gramineas Estipeas.
Ciencia e Investigacion 1, 3: 144-146.
PHILIPSON, W. R., 1934. The Morphology of the Lemma in Grasses. The New Phytologist 33,
5: 359-371.
PILGER, R., 1954. Das System der Gramineae. Bot. Jahrb. 76, 3: 281-384.
PILGER, R., 1956. Gramineae 2. Natuerliche Pfianzenfam. 14d.
PRAT, H., 1936. La Systematique des Graminees. Ann. Sci. Nat. Bot. 18: 165-258.
REEDER, J. R., 1957. The Embryo in Grass Systematics. Amer. J. Bot. 44, 9: 756-768.
REESE, G., 1957. Ueber die Polyploidiespektren in der Nordsaharischen Wuestenflora. Flora
144: 598-634.
ROSHEVITZ, P. J., 1937. Graser. Ref. in Pilger, Gramineae 2. Natuerliche Pfianzenfam. 14d: 120.
401
ROSHEVITZ, P. J., 1951. Stipeae. Bot. Mater. Herb., Komarov Bot. Inst., U.S.S.R. Acad. Sci. 14.
SABNIS, T. S., 1921. The Physiological Anatomy of the Plants of the Indian Desert. J. Indian
Bot. 2: I57-I67; 217-227; 271-299.
SCHWABE, H., 1949. Contribucion a la Anatomia Foliar de algunas Agrostideas. Lilloa 16: 142-160.
SCHWEICKERDT, H. G., 1941. An Account of the South African Material of Aristida L. in certain
European and South African Herbaria. Bothalia 4, 1: 91-175.
ST APE, O., 1899. Gramineae-Stipeae. This. Dyer FI. Cap. 7: 551-575.
STEBBINS, G. L., 1950. Variation and Evolution in Plants. New York: Columbia University
Press.
STEBBINS, G. L., 1956. Cytogenetics and Evolution of the Grass Family. Amer. J. Bot. 43:
890-905.
STEBBINS, G. L. & LOVE, R. M., 1941. A Cytological Study of California Forage Grasses. Amer.
J. Bot. 28: 371-382.
STOCKWELL, P., 1934. A Stain for Difficult Plant Material. Science 80: 121-122.
STURGEON, K. E., 1954. A Revised List of the Grasses of Southern Rhodesia, Part 9. Rhod.
Agric. J. 51: 495-508.
TATEOKA, T., INOUE, S. & KAWANO, S., 1959. Systematic Significance of Bicellular Microhairs
of Leaf Epidermis. Bot. Gazette. 121, 2: 80-91.
THERON, J. J., 1936. Anatomisch-systematische Untersuchungen der Laubblatter Suedafrikanischer
Aristida-Arten. Potsdam: A. W. Hayn’s Erben.
TRINIUS, C. B. & RUPRECHT, F. J., 1842. Graminum Stipaceorum. Actis Acad. Imp. Sci.
Petrop. Ser. 4, 5.
VAN TIEGHEM, Ph., 1872. Sur le Cotyledons des Graminees. Ann. Sci. Nat. 5, 15: 236.
VAUTIER, S., 1949. La vascularisation florale ches les Polygonacies. Candollea 12: 219-343.
VOLKENS, 1887. Die Flora der Aegyptisch-Arabischen Wueste auf Grundlage anatomisch-physio-
logischer Forschung. Berlin.
VON MOHL, H., 1845. Ueber die Bedeutung der unteren Blumenspelze der Graeser. Bot. Zeitung,
3, 3: 33.
ZEMKE, E., 1938. Anatomische Untersuchungen an Pfianzen der Namibwueste (Deutsch-Suedwest-
afrika). Flora 133: 365-416.
402
INDEX
Page
Achnatherum pekenense (Hance) Ohwi, 217
Aciachne Bth., 205, 206, 207
AGROSTEAE, 205, 207
Amphipogon R. Br., 221, 222, 380
Anisopogon R. Br., 205
Aristella Bertol, 205
Aristida L., 229, 234, 300, 302, 392
§ARISTIDA, 238, 300, 302
§ARTHRATHERUM (P. Bcauv.) Reichcnb., 227,
233, 238, 240, 300, 301,302, 303
§CHAETARiA (P. Beauv.) Trin., 220, 227, 238
§PSEUDARTHRATHERUM Chiov., 233, 238, 249,
299, 300, 301, 302, 303
§PSEUDOCHAETARiA Henr., 226, 233, 236, 238,
300, 302
§SCHISTACHNE (Fig. & De Not.) Henr., 220,
226, 227, 228, 234, 307, 308, 309
§SCHIZACHNE (Trill. & Rupr.) de Winter 238,
299, 300, 302
§STiPAGROSTis Trin. & Rupr., 220, 226, 227,
228, 234, 307, 309
§STREPTACHNE, 220, 226, 227, 231, 243, 299
Aristida acutiflora Trin. & Rupr., 220
adscensionis Z.., 221, 225, 236, 238, 254, 301,
302
subsp. guineensis (Trin. & Rupr.) Henr., 239,
240, 250, 251, 254
var. pygmaea (Trin. & Rupr.) D. & S., 253
var. strictiflora (Trin. & Rupr.) D. & S., 253
aequiglumis Hack., 233, 239, 241, 261, 262,
265, 301, 302
alopecuroides Hack., 297
amahilis Schweick., 324
andoniensis Henr., 249
angolensi.s Hiibh., 226, 228, 384
angustata Stapf, 268
argentea Schweick., 240, 241, 288, 289
astroclada Chiov., 238
atroviolacea Hack., 270
harhicoUis Trin. & Rupr., 221, 233, 298, 392
var. conglomerala Henr., 298
bifida Karl, 355, 357
bipartita (Nees) Trin. & Rupr., 239, 240, 257,
260, 301, 302, 343
brachyathera Coss. & Ba/airsa, 220
brevifolia (Nees) Steud., 338
burkei Stapf, 211, 299, 346
caeride.scens Dest’., 251
var. brevi.seta Hack., 251
canescens Henr., 233, 239, 241, 259, 260, 265,
295, 301, 302
subsp. canescens, 258, 259, 260
subsp. ramosa de Winter, 260
capensis Thunb.. 348, 351
var. barbata Stapf, 347
var. canescens Trin. & Rupr., 347
var. dieterleniana Schweick., 229, 349
var. fuhiberbis Trin. & Rupr., 348
var. genuina Henr., 348
var. macropus Nees, 348
var. zeyheri (Nees) Walp., 347
ciUata Desf., 220, 221, 304, 316, 351
var. capensis Trin. & Rupr., 316
var. villosa Hack., 316
var. pectinala Henr., 316
Page
var. tricholaena Hack., 316
coma-ardeae Mez, 320
confusa Trin. & Rupr., 253
congesta R. & S., 232, 233, 237, 240, 241, 296,
302
subsp. congesta, 295, 297
subsp. barbicollis (Trin. & Rupr.) de Win-
ter, 298
var. genuina Chiov., 297
var. megalostachya Henr., 297
corythroides Karl, 336
curvata (Nees) Trin. & Rupr., 233, 239, 240,
253, 302
var. nana (Nees) Henr., 253
damarensis (Mez) de Winter, 329
dasydesmis Mez, 233, 235, 240, 241, 273, 275,
282, 283, 301, 302, 392
diffusa Trin., 235, 240, 241, 111, 275, 281,
282, 299, 301, 302, 346
var. diffusa, 275, 277
var. burkei (Stapf) Schweickerdt, 233, 111,
281, 283, 284, 301
var. brevistipitata (Trin. & Rupr.) Henr., 277
var. densa (Trin. & Rupr.) Henr., 277
var. ecklomiana (Trin. & Rupr.) Henr., 277
var. schraderiana (Trin. & Rupr.) Henr., 277
var. genuina Henr., 277
dinteri Hack., 320
dregeana (Nees) Trin. & Rupr., 344
effusa Henr., 233, 239, 240, 244, 251, 257, 260,
299, 302
elliptica Kiinth, 299
elymoides Mez, 287
engleri Mez, 231, 240, 241, 281, 302
var. engleri, 281
var. ramosissima de Winter, 282
fastigata Hack., 338
flocciculmis Mez, 279
fontismagni Schweickerdt, 286
funiculata Trin. & Rupr., 221
galpinii Stapf, 270
garubensis Pilger, 343
geminifoUa (Nees) Trin. & Rupr., 341
gonatostachys Pilger, 353
gossweileri Pilger, 270
graciliflora Pilger, 292
var. robusta Stent & Rattray, 291
gracilior Pilger, 361, 363, 365, 367
var. intermedia Schweickerdt, 361, 365, 306,
307
var. pear.s'onii Henr., 365
gracillima Oliv., 286
guineensis Trin. & Rupr., 250
hermannii Mez, 369
hirtigluma Steud., 221, 363
var. patula Hack., 365
hochstetteriana Beck., 315
hockii de Wild., 270
hordeacea Kunth, 233, 236, 238, 240, 244, 249,
250, 299, 300, 302, 392
var. longiaristata Henr., 244
hubbardiana Schweickerdt, 239, 240, 245, 250,
299, 300, 301, 302
hygrometrica R. Br., 238
hystrix sensu Thunb'., 277
403
Page
jiicuiula Schweickerclt, 226, 111, 228, 384 I
junciformis Triu. cS Riipr., 231, 232, 233, 239, '
241, 262, 266 |
subsp. junciformis, 268, 301, 302 I
subsp. galpinii (Stapf) de Winter, 231, 268,
270 '
lamarkii (R. & S.) Steud., 286
ianipes (Mez) de Winter, 355
lanuginosa Burch., 279
longicauda Hack., 297, 298
hitescens (Nees) Trin. & Rupr., 336
var. marlothii (Hack.) Stapf, 336
mariothii Hack., 336
meridionalis Henr., 233, 240, 241, 284, 287, 302,
303
mollissima Pilger, 240, 241, 284, 287, 290, 302, [
303, 392
monticola Henr., 239, 241, 265, 273, 301, 302
namaquensis (Nees) Trin. & Rupr., 221, 325
var. vagans (Nees) Walp., 325
obtusa Del., 221, 304, 306, 307, 355
parvula (Nees) de Winter, 231, 236, 238, 240,
242, 299, 301, 302
pennata Trin., 220
pilgeri Henr., 240, 241, 293, 296, 303
plumosa L., 220
proxirna Steud., 322
pseudo-hystrix (Trin. & Rupr.) Steud., 277
p ungens Desk, 220, 221, 306
var. pennata, 220
pusilla Trin. & Rupr., 253
pygmaea Trin. & Rupr., 253
rangei Pilger, 297
recta Franch., 231, 239, 241, 269, 270, 301, 302
rhiniochloa Hochst., 233, 239, 240, 249, 301, 302
rigidiseta Pilger, 249
riparia Trin., 301
subacaulis (Nees) Steud., 273
sabidicola Pilger, 331
scabrivalvis Hack., 233, 239, 240, 253, 254, 260,
302
var. scabrivalvis, 257
var. contracta de Winter, 257
Schaefer i Mez, 318
var. biseriata Henr., 318
schlechteri Henr., 351, 352
sciurus Stapf, 239, 241, 273, 275, 300, 302, 392
secalina Henr., 315
sericans Hack., 201, 234, 299, 304, 349
serrulata Chiov., 249
spectabilis Hack., 233, 240, 241, 283, 284, 285,
287, 288, 290, 302, 303, 341, 343
steudeliana Trin. & Rupr., 244
stipitata Hack., 233, 240, 241, 290, 302
var. stipitata, 233, 290, 291
var. robusta (Stent & Rattray) de Winter, 291
var. graciliflora (Pilger) de Winter, 292
var. spicata de Winter, 292
stipoides Lam., 240, 241, 285, 286, 302, 303
var. meridionalis Stapf 284
strictifiora Trin. & Rupr., 253
subacaulis (Nees) Steud., 373
submucronata Schiim., 254
transvaalensis Henr., 233, 239, 241, 263, 299,
301, 302, 303
imiplumis Licht., 221, 306, 359
var. pearsonii Henr., 360
var. neesii Trin. & Rupr., 360
Page
vanderijstii De Wild., 226, 227, 228, 328, 387
veslita Thunb., 240, 241, 275, 279, 282, 283, 302
var. parviflora Trin. & Rupr., 275
var. brevistipitata Trin. & Rupr., 277
var. densa Trin. & Rupr., 277
var. pseudohystrix Trin. & Rupr., 277
var. Schrader iana Trin. & Rupr., 277
var. diffusa (Trin.) Walp., 277
forma amplior Hack., 279
waibeliana Henr., 251
walteri Suessenguth, 338
welwitschii Rendle, 268
zevheri (Nees) Steud., 347
ARISTIDEAE Hubbard, 202, 205, 212, 219, 220,
221, 222, 223, 224, 225, 226, 228, 234, 379,
380, 393
Arthratherum (Beauv.) Reichenb., 238, 304
brevifolium Nees, 338
hitescens Nees, 336
namaquense Nees, 325
obtusum (Del.) Nees, 355
pungens Beauv., 220
subacaulis Nees, 373
uniplume (Licht.) Nees, 360
vestitum (Thunb.) Nees, 279
zeyheri Nees, 347
ARUNDINEAE Dumort., 224, 225
AVENEAE Dumort., 207
BAMBUSEAE, 209
Brachyelytrum Beauv., 207
Chaetaria Beauv., 205, 238
biparita Nees, 257
capensls (Thunb.) Beauv., 348
congesta (R. & S.) Nees, 297
ciirvata Nees, 253
lamarkii R. & S., 286
var. nana Nees, 353
vestita (Thunb.) Beauv., 279
CHLORIDEAE Agardh, 223, 224, 234
Danthonia DC., 222, 223, 380
DANTHONIEAE Nevski, 222, 223, 380
Dichelachne, 205, 206, 207
Diplachne Beauv., 380
Diplopogon R. Br., 220, 221, 222, 380
Ehrharta Thunb., 205
Eragrosteae Stapf, 207, 219, 223, 224, 230
Eragrostis Beauv., 380
ERAGROSTOIDEAE Pilger, 225
FESTUCEAE Dumort., 207, 208, 222
FESTUCOIDEAE Hitchc., 207, 209, 224
Lasiagrostis Link., 205, 206, 216
capensis Nees., 215, 216, 217
var. elongata (Nees) Trin. & Rupr., 216
elongata Nees, 216
Lepturus R. Br., 380
Lycurus Kunth., 207
Macrochloa Kunth., 205, 206
MAYDEAE Dumort., 223
Milium Adans, 205, 206, 207
Monerma Beauv., 380
Muehlenbergia Schreb., 207
NARDEAE Reichenb., 207
Nasella Desv., 205, 206, 208, 209, 210
Orthoraphium Nees, 205, 206
Oryza L., 205, 207
ORYZEAE Dumort., 209, 223
Oryzopsis Michx., 205, 206, 209, 210, 216, 217
aequiglumis Duthie, 217
404
I
Page
kememis Pilger, 210, 216
PANICEAE R. Brown, 207, 230
PANICOIDEAE, 209
PAPPOPHOREAE Kunth, 207, 223, 224, 234
Perieleima Presl., 207
Phaenosperma Munro ex Benth., 205
PiPTATHERUM Beauv., 205, 210, 216
coerulescens Desf., 210, 216
keuiense (Pilg.) Roshev., 216
PiPTOCHAETiUM J. & C., 205, 206, 208, 209
Plectrachne Henr., 222, 380
PoA Linn., 380
Ptilagrostis Griseb., 206
Relchella Steud., 206
Sartidia de Winter, 225, 228, 302, 303, 379, 380,
381
angolensis (Hubb.) de Winter, 111, 381, 382,
384
jucunda (Schweickerdt) de Winter, 227, 382, 384
vanderijstii {De Wild.) de Winter, 227, 382, 387
Schistachne Fig. & De Not., 307, 308, 309
SPOROBOLEAE Stapf, 207, 219, 223, 224, 234
Sporobolus R. Br., 217, 219
Stipa Linn., 201, 205, 206, 207, 208, 209, 210,
211, 212, 216, 217, 219, 222, 244, 299, 377
§ACHNATHERUM, 216
§PTILAGROSTlS, 210, 216
§SCHIZACHNE Trin. & Rupr., 236
§TORTiLiA Elias, 210
capensis Thiinb., 210, 211, 213, 217
dregeana Steud., 210, 211, 213, 216, 217, 232,
235
var. dregeana, 215
var. elongata (Nees) Stapf, 210, 215, 216
elongata (Nees) Steud., 216
extremi-orientalis, 217
hookeri Stapf, 206
leucotricha Trin. & Rupr., 209
namaquensis Pilger, 211, 212, 309, 375
parvula Nees, 211, 212, 236, 242, 243, 299
pennata Linn., 212
sibirica Lam., 217
tortilis Desf., 210, 217, 219
vaseyii Scribn., 217
Stipagrostis Nees, 212, 225, 226, 227, 228, 230,
231, 233, 236, 299, 300, 303, 304, 307, 308,
342, 351, 357, 359, 377, 379, 380, 389, 390,
392
§ANOMALA de Winter, 309, 311, 377, 378
§SCHISTACHNE (Fig. & De Not) de Winter, 228,
300, 307, 308, 309, 322, 378, 379
§STiPAGROSTis, 228, 300, 309, 322, 378, 379
amabilis (Schweick.) de Winter, 237, 306, 309,
311, 324, 328, 330, 332, 378, 379
anomala de Winter, 211, 237, 309, 311, 312,
375, 377, 378
brevifolia (Nees) de Winter, 305, 310, 312, 322,
328, 338, 378, 379
capensis Nees, 304, 355
ciliata (Desf.) de Winter, 235, 237, 305, 308,
309, 312, 315, 316,320,352, 378,379,391
var. capensis (Trin. & Rupr.) de Winter, 316
damarensis (Mez) de Winter, 310, 312, 324,
329, 378, 379
dinteri (Hack.) de Winter, 309, 312, 320, 357,
378 379
dregeana Nees, 304, 310, 312, 336, 344, 357, [
378, 379, 390, 391 j
Page
fastigiata (Hack.) de Winter, 305, 310, 312,
338, 378
garubensis (Pilger) de Winter, 310, 312, 336,
343, 346, 378, 379
geminifolia Nees, 304, 305, 310, 312, 341, 378
gonatostachys (Pilger) de Winter, 310, 311,
312, 322, 353, 357, 378, 379
hermannii (Mez) de Winter, 305, 308, 311,
313, 353, 369, 372, 373, 374, 378
hirtigluma (Steud.) de Winter, 305, 311, 313,
359, 360, 361, 367, 378
var. hirtigluma, 363
var. patula (Hack.) de Winter, 305, 365
var. pearsonii (Henr.) de Winter, 305, 306,
365
hochstetteriana (Boeck. ex Hack.) de Winter,
232, 237, 306, 309, 312, 313, 378, 379
var. hochstetteriana, 315
var. secalina (Henrard) de Winter, 315
lanipes (Mez) de Winter, 306, 310, 312, 322,
355, 357, 373, 378, 379, 391
lutescens (Nees) de Winter, 310, 312, 324,
334, 378, 379
var. lutescens, 336
var. marlothii (Hack.) de Winter, 336
namaquensis (Nees) de Winter, 232, 306, 309,
311, 324, 325, 330, 332, 338, 375, 378,
379
namibensis de Winter, 311, 313, 370, 378
obtusa (Del.) Nees, 232, 237, 305, 306, 308,
311, 312, 322, 334, 353, 355, 377, 378,
379, 391
proxima (Steud.) de Winter, 309, 312, 322,
314, 378, 379
pungens (Desf.) de Winter, 220, 306, 391
ramulosa de Winter, 310, 312, 333, 378
sabulicola (Pilger) de Winter, 306, 310, 311,
324, 328, 331, 334, 337, 378, 379
schaeferi (Mez) de Winter, 309, 312, 318, 377,
378, 379
subacaulis (Nees) de Winter, 305, 308, 311,
313, 370, 371, 373, 378
uniplumis (Licht.) de Winter, 305, 306, 311,
312, 313, 342, 359, 362, 363, 367, 370,
378
var. uniplumis, 306, 358, 360, 365
var. neesii (Trin. & Rupr.) de Winter, 305,
306, 360, 367
var. intermedia (Schweick.) de Winter, 306,
361
zeyheri (Nees) de Winter, 310, 346, 378, 379
subsp. barbata (Stapf) de Winter, 347
subsp. macropus (Nees) de Winter, 347,
348
subsp. sericans (Hack, apiid Schinz) de
Winter 299, 347, 349
subsp. zeyheri, 347
STIPEAE Nees, 201, 205, 206, 207, 208, 209, 210,
212, 219, 222, 223, 224
Streptachne R. Br. 205, 206, 207, 210
Timouria Roshev., 205, 206
TIMOURllNAE Roshev., 216
Trikeraia Bor, 206, 209, 210
Triodia R. Br., 222, 380
TRIODIINAE Pilger, 222
UNIOLEAE, 209
Urachne Trin., 205
Frontispiece: Encephalartos transvenosus (Modjadji Cycad).
THE CYCADS OF SOUTHERN AFRICA
(with one map and 97 illustrations)
by
R. A. Dyer
CONTENTS
Page
Frontispiece
Introduction 405
Literature 406
Acknowledgements 407
Family Characters 409
Fossil Records 410
Survival 411
Reproduction 412
Coning 412
Pollination and Fertilization 414
Germination 415
Vegetative propogation 420
Hybridization 421
Age and Rate of Growth 422
Conservation 425
Miscellaneous —
Toxicity 427
Edibility (Starch) 427
Yeast 428
Root tubercles or nodules 428
Decoration 428
Insects 428
Stangeria 429
S. eriopus 429
ii
' Page
Encephalartos 432
Key to species 433
1. E. gheUinckii .• 437
2. E. cycadifolius 439
3. E. friderici-guilielmi 442
4. E. lanatus 445
5. E. hiimilis 448
6. E. laevifolius 451
7. E. mopinus 453
8. E. eiigene-maraisii 455
9. E. pr bleeps 458
10. E. lehmannii 460
11. E. trispinosiis 463
12. E.Jiorridus 467
13. E. areiiarius 470
14. E. latifroiis 471
15. E. longifoliiis 474
16. E. altensteinii All
17. E. woodii 482
18. E. natalemis 486
19. E. lebomboensis 491
20. E. tramvenosiis 494
21. E. paucideiitatus 498
22. E.fevox 499
23. E. viUosus 502
24. E. iimbeluziensis 505
25. E. iigoyanus 508
26. E. caffer 509
Conclusion 514
Index to Species 515
405
THE CYCADS OF SOUTHERN AFRICA
Introduction
The object of this account of the Cycads of Southern Africa is to summarize the
salient information which will interest the botanically minded’ members of the general
public. It has been prepared in conjunction with a taxonomic treatment of the species
for the Flora of Southern Africa and for this reason the descriptions of individual
species have been abbridged, literature references minimized and herbarium specimens
not cited. The area concerned lies south of the Limpopo River.
Records show that keen interest in the Cycadaceae has not flagged much during the
past 200 years. Sixty years ago Worsdell, writing in the Annals of Botany Vol. 20: 129
(1906), said that “ no Family of Plants is exciting more interest at the present day than
the Cycadaceae, and for the prime reason that they represent one of those forms of
plant life which link the types of the far past with those of the present modern world ”.
The family Cycadaceae, as generally accepted in the past, is represented widely in
the southen hemisphere and in limited areas of the northern hemisphere. The greatest
Cycad concentrations are in Mexico, central America, and Queensland in Australia.
Queensland has three genera, Lepidozamia Macrozamia and Boweuia; Mexico three,
Dioon, Ceratozamia and Zamia. No other region can clafm more than two genera.
In Southern Africa we have Stangeria and Encephalartos.
When the Cycadaceae of Southern Africa were monographed in the Flora Capensis
by Hutchinson and Rattray in 1933, one species of Stangeria and 13 of Encephalartos
were recognized. When Henderson contributed his results towards a “ Revision of the
South African species of Encephalartos ” in the Journal of South African Botany in
1945, he listed 18 described species and suggested that it would be necessary to name and
describe several more, once it were possible to do some intensive field work. Most of
Henderson’s forecasts have proved well founded and only one completely new diseovery
has been recorded since his summary of 1945, and that is Encephalartos inopiniis.
The other newly established species since Henderson’s work are: E. arenarius
(distinguished from E. latifrons); E. hnmilis (segregated from E. lanatiis); E. lebom-
boensis (distinguished from E. altensteinii); E. natalensis (segregated from E. alten-
steinii); E. ngoyanus (segregated from E. caffer)\ E. princeps and E. trispinosus segre-
gated from E. horridus and E. lehmannii; and E. wm/WMc/ciwA (distinguished from £.
villosus). E.friderici-guilielmi has been reinstated and distinguished from E. cycadifolius,
whereas E. striatus (possibly equal to E. villosus or E. umbeluziensis), has been excluded
because of inadequate information for its certain identification with any known species.
Thus 26 species of Encephalartos and an indeterminate number of natural hybrids
are dealt with in the present work. The distribution of these begins in the north with
E. transvenosus on the Soutpansberg, and E. eugene-maraisii on the Waterberg, and
extends mainly in mountainous areas of the eastern Transvaal into Natal and the south-
eastern Cape as far as the district of Uniondale. Here E. longifolius is the species which
has penetrated furthest towards, or persisted nearest to, the south-western Cape flora.
No record of the genus has come from the Orange Free State or South West Africa.
Records of species in tropical and subtropical Africa are widespread but the number
is unlikely to be as many as half that for southern Africa.
In the case of the monotypic genus Stangeria, it occurs in coastal districts from
Bathurst in the eastern Cape to northern Zululand near Kosi Bay. It may well be
found to occur in southern Portuguese East Africa.
The accompanying map gives a guide to the limits of distribution of the family in
our area.
406
Map : Showing distribution of family, Cycadaceae, in Southern Africa.
Literature
A vast amount of literature about Cycads has accumulated over the years and no
attempt is made to present a list of references here. Hutchinson and Rattray give most
of the references to earlier taxonomic literature in Flora Capensis 5, 2 (Suppl.) 1933.
An article on economic aspects of Cycads published in Economic Botany 12, No. 3 (1958)
by John W. Thieret is supported by 296 references to literature. The literature on the
toxicity of Cycads is summarized by Marjorie Whiting in Economic Botany 17, No. 4
(1963) with 254 references, which gives some idea of what the preparation of a full
bibliography would entail. Henderson’s work has been referred to above and other
works are mentioned in the pages which follow.
The present account is based very largely on field observations but there are still
many problems which call for further attention and this by no means pretends to be the
last word.
407
Acknowledgements
As this project has gone on the number of those to whom I am indebted in one way
or another, has grown beyond expectations. I thank them all.
Material and information has been received from the Directors of the following
Institutions: Royal Botanic Gardens, Kew; Berlin-Dahlem' Herbarium; Botanisch
Museum en Herbarium, University of Utrecht; Rijksherbarium, Leiden; Jardin
Bontanique de I’Etat, Bruxelles; Institutum Botanicum, Szeged; Naturhistorisches
Museum, Vienna; National Botanic Gardens, Kirstenbosch; Bolus Herbarium; Nature
Conservation Dept., Transvaal Province; Penge Asbestos Mine; Parks Departments of
Port Elizabeth, East London and Queenstown.
Among private people my special thanks are recorded to Mr. G. G. Smith, who has
travelled several thousand miles in quest of field knowledge of our Cycads and has built
up a valuable study collection in the grounds of the East London Museum, and to the
late Basil Christian who accompanied him on several excursions; to the late Dr. N. R.
Smuts; to Mr. V. Pringle, Mrs. N. H. James, Mr. J. A. Smit, Mr. C. Els, Mr. Louis
Naude, Mr. Jimmie Hall, Mr. A. P. du Plessis, Mr. W. P. Diepraam, Mrs. Olive Hilliard,
Prof. R. H. Compton, Dr. E. E. A. Gledhill, Dr. Edna Plumstead, Prof. A. W. Bayer,
Miss Courtenay-Latimer, Mr. L. Renton, Mr. B. J. Huntley and Miss G. Blackboard,
also Mr. L. N. Prosser, Mr. T. Linley and Mr. A. Paton.
Fig. 1. — Cycad hunting: Miss Courtenay-Latimer, Miss Johns, Prof. R. H. Compton and Mr. G. G.
Smith investigating possible hybrids between Encephalartos altensteinii and E. trispinosus on
the west bank of the Bushmans River in the Alexandria district.
Most of my colleagues in the Botanical Research Institute have been involved to a
greater or lesser degree and with apologies to those I omit by name, I must mention
Dr. L. E. Codd (Chief), Miss M. Gunn (Librarian), Mrs. E. van Hoepen, Mr. M. J.
Wells, Mr. R. G. Strey, Messrs H. King and J. Reyburn (photographers), Mrs. M. B.
Coetzee (typist) and Miss I. C. Verdoorn, who has a fundamental knowledge of the
Cycads, who has contributed several articles on the group and who is joint author with
me of the taxonomic account being prepared for the Flora of Southern Africa.
408
Fig. 2. — Portion of Cycad collection at the Kirstenbosch Botanic Garden, started by Pearson about
1913: by 1916 fourteen to sixteen species were represented there by approximately 500 speci-
mens; photo taken 1948.
I 9g. 3. — Portion of Cycad collection of Botanical Research Institute, Pretoria, started in 1950; photo
taken 1965.
409
FAMILY CHARACTERS OF CYCADACEAE IN SOUTHERN AFRICA
The plants are either male or female and it is not possible to determine the sex in
the absence of cones. They have subterranean or aerial stems which are commonly
branched from the base but rarely above. The stems produce alternating tufts or
whorls of leaves and bracts from the apex (often termed palm-like although there is no
relationship between the two groups). The leaves have a strong central stalk or rhachis
from which are produced many leaflets, neither strictly alternate nor opposite. The
pollen of the male, and seeds of the female, are produced on scales which are densely
aggregated into cones, the female cone being larger than the male in their respective
species. The male scales produce pollen on their under surface in densely arranged cells,
the tip of the scale being sterile and produced into a beak of varying length (Fig. 4).
Each female scale produces two comparatively large exposed seeds, directed inwards
from the scale head or bulla and they lie above the scale stalk (Fig. 5); the seeds are red,
yellow or amber in colour when ripe and have a fleshy covering over a hard inner jacket
which encloses the endosperm (food supply of embryo).
Fig. 4. — Male cone scales, under-surface and face of: (a) Encephalartos cajfer (Henderson 1505);
(b) E. caffer from East London (G. G. Smith 6901); (c) E. ngovanus (Verdoorn and Christian
716b).
In Cycads, which belong to the group of plants known as Gymnosperms, the seeds
are exposed and are not enclosed in special containers such as pods and fleshy fruits as
they are in the flowering plants termed Angiosperms.
The two South African genera, Stangeria and Encephalartos may be distinguished as
set out in the following ‘ key ’ :
Stems subterranean, tuber-like; leaves somewhat fern-like, falling entire from the
stem; leaflets with a prominent mid-rib and with branched, spreading lateral
veins Stangeria.
410
Stems subterranean or up to 30 ft or more tall, protected by densely packed leaf-
bases which appear as leaf-scars ; leaves with a central stalk from which spread
at varying angles numerous leaflets along its length ; the leaflets have parallel
venation. Encephalartos.
l| |S| |8| |4| I si |6j ,7| |8| |9| , l|6| l|l
Fig. 5. — Female cone scales with 2 seeds each: left, upper and right, lower view of Encephalartos
lebomboensis.
FOSSIL RECORDS
The fossil history of the fern-like parent stock of the Cycadales from Mesozoic
times could be anything from 200 to 300 million years, according to Professor Dr.
P. Greguss, of Szeged University, who has done extensive research on the anatomical
features of the group.
There is general agreement that the Cycads, as we know them, had their origin in
the Carboniferous or Permian period, when some of the world-wide coal measures
were being formed and the Dinosaurs roamed our countryside, estimated at over 50
million years ago.
In his account of the Geology of South Africa, Alec du Toit, 1926, when dealing
with the fossil wood beds in the Uitenhage Series of the Cretaceous System, states that
some of the beds are crowded with the broad fronds of Zamites. These are fossil
Cycads. In setting a geological age for these, du Toit agrees with Seward’s conclusion
that the flora is related to that of the Upper Jurassic. He lists nine genera of fossil
Cycads with a total of 12 species. These were found from Pondoland in the east to the
Gamtoos River Valley in the west, with most of them from the Sundays River Valley.
Zamites recta, which du Toit figures, could, with good reason, be compared with the
modern species, Encephalartos longifolhis, which is found today in the mountains in the
catchment areas of the two rivers mentioned.
411
In the Annals of the S.A. Museum of 1927, dealing with The Fossil Flora of the
Upper Karoo Beds, du Toit described an incomplete specimen from Maclear which he
considered was another species of Zamites, nearest in affinity to the Indian fossil, Z.
rajmahalensis (Morris) du Toit.
Dr. E. Plumstead of the Witwatersrand University, informs me that a few remnants
of Cycads (Otozamites) of lower mid Jurassic age have been found interbedded with the
Drakensberg lavas in Portuguese East Africa and northern Zululand and that other
fragments have been recorded in formations of Upper Jurassic age on the coast of
Zululand.
Incidentally Dr. Plumstead has recently described three new species of Zamites in
fossil material from Antarctica.
Because the Cycads have survived for such a very long period of time with relatively
little modification, they are often referred to as living fossils.
SURVIVAL
For Cycads to have propogated themselves for over 50 million years, with little
change in basic character, is something to be marvelled at. One may speculate on the
reasons for this phenomenon. On the basis that the strength of a chain is the strength
of the weakest link, it is obvious that the ancestors of modern Cycads had no really
weak link, or, if they did, it was compensated for in another direction. This does not
imply that there has been no change in the species from one era to another, but that new
forms or species have evolved as and when environmental changes created a sufficiently
strong stimulus. The stock possessed the genetic factors required for adaptation to the
changing environment throughout the long period of survival.
Temperature fluctuation was probably the most testing factor in the earlier phase of
survival and species today suffer less permanent damage than most other plants under
wide fluctuations of temperature. After the evolution and spread of the grasses and the
advent of generally drier conditions within the environment, Cycads were put to the
critical test of fire. To meet this new hazard, the stems of Encephalartos species already
possessed the protection of persistent leaf-bases and scales to a thickness of 1-3 inches.
This protective covering has withstood the test of fire periodically for many centuries.
As with many perennial herbs and bulbs in grassland, fire may now act as a stimulus
to fresh growth. This is illustrated by an account by Mr. V. L. Pringel of Bedford in
the Cape Province; — “ I have been very interested this season in the behaviour of the
Cycads {Encephalartos cycadifoUus) which grow in these mountains. In August last
year I burnt down some stretches in the mountain which had become overgrown with
coarse grass and all the Cycads, of which there are hundreds, had their leaves burnt off.
They soon came out in full leaf and are now looking better than ever before — and there
is scarcely one in hundreds which has not fruited. I notice that where others grow,
which were not burnt, there is no sign of any fruit as yet. The good rains at the end of
July might have helped them but it has been dry through September-October-November.
The only explanation which I can see is that the burning has stimulated the plants ”.
Here it may be added that Stangeria, which is no doubt similarly responsive to the
stimulus of fire, but whose stems do not have the protection of leaf-bases, is restricted
to one species with a subterranean stem.
The kernel or endosperm surrounding the embryo has proved poisonous in all
species of Encephalartos which have been reliably tested. Encephalartos is not an
exception in the Family in this respect, which suggests that the toxicity of the seed has
also had a significant bearing on the persistence of the family throughout the ages.
412
What may be regarded as a weak link in the life history of the Cycads, at least
under South African conditions, is the lack of a real resting period in the seed and its
slow germination. If fertilization takes place, the embryo develops slowly and steadily
so long as favourable conditions prevail, but it takes a period of six to nine months
before the first root appears and two or three months longer before the growth of the
first leaf. If, however, any unfavourable circumstance should arise during this period to
check the development of the embryo, it is very likely to die.
The succulent tap root of the seedlings seems to be attractive as food for certain
animals, such as rodents, for relatively few of the seedlings which do germinate in the
wild state, ever reach maturity. On the other hand the possible life-span of individual
plants can be reckoned in hundreds or thousands of years, which period should be long
enough to ensure some natural regeneration in the absence of interference by man.
REPRODUCTION
Coning
It has been pointed out that the male and female cones of both Stangeria and
Encephalartos are produced on separate plants. It happens occasionally that the
apical scales of a female cone develop into miniature leaves, thus recalling the fact that
the cone itself is made up of modified leaves which have become seed-bearing scales,
closely aggregated into compact fruiting bodies. At least two cases have been recorded
for Encephalartos longifolius in South Africa (Fig. 6) and it is known to have occurred
in an unidentified species in tropical Africa.
The age at which plants of Encephalartos produce a cone for the first time is known
to be extremely variable, and while some plants cone fairly regularly, others are erratic.
One specimen at the Botanical Research Institute has not coned in 50 years. If a
plant has more than one stem it is most unlikely that all stems will cone at the same time.
It would be more in keeping if only one of several stems coned in any one year. A
sucker from an old plant may cone at ground level as was the case with a specimen of
E. laevifolius at the Botanical Research Institute. Cones usually alternate with one or
more leaf- whorls, but sometimes cones and leaves arise almost simultaneously, possibly
due to some physiological disturbance.
The effect of fire on the coning of specimens of E. cycadifolius on the farm of V. L.
Pringle is referred to in the paragraphs on ‘ survival ’.
There can be few accurate records of the time taken from the seedling stage to first
coning of an Encephalartos plant, so the following record by Schelpe in S.A. Journal of
Science 47 : 16 (1950), is of interest. Seeds of E. ferox (= E. kosiensis) from Kosi Bay,
germinated in Durban in 1937. One of these produced a small female cone 16 cm long
in 1 949. In the 1 2 years the stem, mostly subterranean, had reached a diameter of 25 cm.
The phenomenon of a rise in temperature within the dehiscing male cone of E.
altensteinii was recorded for the first time in Encephalartos by Amy Jacot Guillarmod in
the August issue of Nature 1958. A record of this observation merits repetition here.
A thermometer registering 21°C in the laboratory, was inserted between the sporophylls
(cone scales) of a dehiscing male cone of E. altensteinii, taken from the Botanic Garden
in Grahamstown. Within a matter of seconds the thermometer registered 35°C and
in the course of the afternoon rose to 37°C. Tests on other cones of the same species
in a similar stage revealed that there was a diurnal rhythm, the temperature within the
cones rising from about 11 a.m., to a maximum about 5 p.m., and then falling in the
evening to room temperature.
413
Fig. 6. — Female cone of Encephalartos longifolius with a few abnormal apical scales indicating that all
the cone scales represent modified leaves.
414
Pollination and Fertilization
There is no difficulty in the way of pollen dispersal. The male cone scales are
well spaced ; pollen is produced in excessive quantity and is carried away in the slightest
breeze as it drops from the microsporangia on the under-surface of the scales. The
cones of several of the species with woolly cones are observed to bend to one side at
maturity and this has been suggested as an aid to pollen dispersal. It seems to be
generally accepted that pollination in Stangeria is effected by wind dispersal of pollen
and, judging by fertility counts, it is very effective. In Encephalartos both insect and
wind agencies are thought to play a significant part. The claim for insect pollination
is largely on circumstantial evidence provided by Alice Pegler, a very keen naturalist
who lived at Kentani at the beginning of the twentieth century and contributed much to
South African botany. Certain insects are present on and in the developing cones and
the possibility of insect pollination cannot be refuted. But it is still an open question
whether it is fortuitous or by the design of nature.
The scales of the female cones appear so tightly packed that one is surprised that
pollen can be transported by wind or insect to the micropyle, especially in the case of
the thickly woolly cones of certain species, such as E. lanatus. However, close examina-
tion of female cones shows that the female scales open slightly to allow the entry of
pollen at the receptive stage, more obviously so in Stangeria than in Encephalartos.
One can also effect pollination by pumping fertile pollen into the female cone at the
receptive stage. Chamberlain in his booklet on The Living Cycads (1919) says that the
pollen grains, happening to reach receptive female cones, sift in among the scales while
the cells at the tip of the ovule are breaking down and secreting a clear mucilaginous
substance which on drying draws down into the pollen chamber whatever pollen has
fallen on it. Within the pollen chamber the male body develops slowly over a period of
months before fertilization takes place. The long interval between initial pollination and
subsequent fertilization is a peculiar feature of Cycads and very different from the
relatively rapid process in flowering plants.
Marloth, in his Flora of South Africa Vol. 1, 1913, gives the following summary:
“ The ovule is fully developed only when it has reached almost the size of the ripe seed;
it then secretes a drop of fluid at the micropyle, by means of which the pollen grains,
brought there by the wind or through the agency of insects, are captured and transported
into the pollen chamber; here they attach themselves to the walls of the chamber by
means of a short tube {haustorium). Their further development has not been observed
in any South African species as yet ”. That was written in 1913 and, regretably, very
little more is known 50 years later. This would be a wonderful field for University
research.
That fertilization is not effected with regularity in the wild state is certain. Some
cones are completely sterile while others are almost 100 per cent fertile. The following
extract from a letter in 1959 by an American nurseryman to a South African colleague
is full of interest:
“ The seeds of E. ahensteinii arrived in good shape and we were very glad to
get them. However, we are not going to plant them. After dissecting a repre-
sentative sample we have decided that none of them has an embryo. We enclose
a few for your examination. A longitudinal halving will disclose a small hollow
where the succulent embryo should be. The embryo should roundly fill the round
exit hole in the endosperm and should extend to about the centre of the seed,
gradually narrowing to a slender point ”.
415
“ Please be assured that I am a patient philosopher when it comes to getting
viable seed of African Cycads. I have been at it twenty years. Over this period I
have planted thousands of seeds that were without embryos, not realizing the
hopelessness of the effort. But think how rewarding it can be to get finally some
measure of success. I hope that you will not be discouraged in getting seeds for us.
The very plant from which these seeds came, may next year bear fertile seed, be-
cause a male plant in the vicinity may be coning at the right time ”.
Fig. 7. — Seed germination:
(a), (6) and (c), longitudi-
nal sections of fertile seeds
at different stages of de-
velopment; (d), (e), de-
velopment outside seed;
(a) to (e) denote growth
period of 6 to 9 months:
seeds without a thread-
like structure are infertile.
Germination
It is not difficult to observe whether the seed is fertile or not because the embryo
begins to develop even before the seed is shed from the cone and continues to grow slowly
while conditions remain favourable. In the initial stage it looks like a small head with
a thin coiled thread attachment (Fig. 7). Young stages of growth in Stangeria eriopus
are seen in Fig. 8.
6821035-2
416
Fig. 8. — Stangeria eriopiis: germ'nati'on and stages in development of seedlings.
Unfavourable circumstances such as exposure to high temperature and low humidity
at the same time, are liable to cause mortality of the developing embryos. This could
be confused with infertility or non-viability as the following occurrence proves; In
answer to a request, Mr. V. L. Pringle forwarded a pocket of fresh seed from Bedford
in the Cape in May, 1960, to the Botanical Research Institute, Pretoria, where knowledge
of the best method of germination of this species was wanting. It was decided to
carry out tests by immediate planting.
417
Five mixtures with different proportions of sand and compost were made up and
placed in open frames. Into each mixture 100 seeds were pressed just under the surface
and given regular watering but no special overhead protection. The highest germination
was 5 per cent by October, 6 months later and no plot improved on this.
The balance, of approximately 1,000 seeds, had been left in a jute bag and thrown
into a dark corner of the store room on a concrete floor. In November a request was
received from overseas for more seed. On opening the bag it was discovered that
approximately 90 per cent of the seed was germinating with the radicle well exposed or
showing signs of emerging. Half of this seed was then placed in a ^ sand, f compost
mixture in 2 frames. Protection from the summer heat was given by means of palm
leaves. The first seedling-leaves appeared in February 1961 and by May, i.e. a year
after collection, over 75 per cent were in leaf. About 5 stragglers produced leaves
after 18 months. Figure 9 shows the condition of the seedlings in their third year of
growth after a number had been distributed.
Fig. 9. — Encephalartos cycadifoliiis: phenomenal germination after seed had remained for several
months in a jute bag on the concrete floor of a store room.
This is not an isolated case of high percentage germination. A year after three
female cones of E. lebomboensis disintegrated from the parent plant growing in a lawn
at Capt. D. R. Keith’s residence near Stegi, Swaziland, seedlings had germinated from
practically every seed (Fig. 10). Some of these seedlings are now in the Pretoria National
Botanic Garden.
During ‘ Operation Wild Flower ’, before the completion of the Jozini Dam across
the Pongola Poort, seedlings of E. lebomboensis could be pulled up in handfulls where
fertile female cones had fallen the previous season. R. H. Compton reported the
germination of practically every seed of a fallen cone of E. umbelnziensis in the wild state.
A successful Australian correspondent writes: “ I am 3 miles from the coast in a
subtropical climate with no frost and a 40—50 in. summer and winter rainfall. I have
an unheated glasshouse in which I grow bromeliads and orchids. Cycad seeds are
418
placed in a mixture of silver-sand and leaf-mould. I press the seed approximately ^ their
depth into the sand leaving the germinating point (top, micropyle end, opposite base)
just visible. They are placed under the bench and receive plenty of water and no direct
sunlight until they germinate. They require high humidity and shade until the first
leaf shows ”.
Not all species require such moist conditions for, in the course of evolution, some
species, such as E. horridiis and E. lehmannii, have become adapted to a relatively dry
environment with an erratic rainfall often less than 40 cm (15 in.) per annum. In the
case of E. lehmanuU seed was observed to be germinating in the veld in the Jansenville
district on the surface of dry shale. M. J. Wells observed that year-old seed of E.
longifoliiis had germinated in full sun on open ground and on rocks.
Fig. \0.—Enceplialartos lebomhocnsis: germinating on top of soil in short grass under parent female
plant in Capt. D. R. Keith’s garden near Stegi, Swaziland, 1947; several now in Pretoria
National Botanical Garden of Botanical Research Institute.
419
It is found that in nature seed of Cycads normally germinates on the surface of the
soil. Some cases have been observed of seed germinating under the side of a boulder
and these were probably transported there by rodents. The seed first produces from
the micropile end a thick cotyledonary tube which curves directly to the soil surface,
where the root emerges from the tube into the soil and becomes tuberous [Fig. 7 {d) and
(e)]. The major part of the cotyledons remains inside the seed, absorbing the endosperm
food supply and passing it to the developing seedling. Once rooted, the first leaf is then
produced from within the top portion of the cotyledonary tube and is followed by
successively produced leaves for the first year or so, after which leaves are produced
two or more together in whorls.
Fig. 1 1 . — Encephalartos hor-
n'diis: plantlets produced
from the inner wall of a
rotted stem found in
nature near Uitenhage.
The first leaves of different species of Encephalartos show differences in the texture,
surface-colour and degree of pubescence of the leaflets but are otherwise similar. Apart
from size, in most cases the leaflets of the first leaf differ makedly from the adult form.
The main features of the juvenile leaflets are that they are fairly uniform in size and shape
and are all toothed on the margin, more consistently so on the upper half and round the
420
obtuse apex. There is no differentiation of the terminal spine which is characteristic
of most adult leaflets; there is no reduction to prickles of the lower leaflets which is a
feature in adult plants of many species, for example, E. villosus. The first leaf of the
subterranean-stemmed E. villosus could be confused with that of the robust-stemmed E.
altensteinii or even that of the giant E. transvenosus.
Vegetative Propogation
The main method is to remove basal suckers which may or may not have their own
root-system. If the off-shoot has no roots it will usually produce them within a fairly
short period and there is a fairly high rate of success by this means. Off-shoots which
are sometimes produced anywhere along the trunk above ground are also fairly readily
rooted. Trunks up to several feet long, with or without roots, may also be re-established
with success in most species, but the time taken to re-root and produce new leaves varies
very considerably between individual stems and between species. Marloth reports a
case of astonishing vitality, that trunks of E. altensteinii, which lay in a shed for four
years, produced new leaves the summer after planting. On the other hand E. lanatus
may show a mortality rate of 50 per cent in transplanted stems.
Fig. 12. — Encephalartos friderici-giiilielnii-. leaves breaking through the surface after the plant had been
buried during the construction of a national road near Cathcart. Photo: J. Hall.
421
A few growers of Cycads have had success in propogation by means of small pieces
of stem attached to leaf-bases. These were placed half-way into a very sandy mixture
and kept slightly moist and cool for several months. Something similar has been ob-
served in the veld, where the centre of the stem rotted away completely and a young
plant developed from the inner wall in the region of the leaf-bases (Fig. 11). For proof
of vitality refer also to Fig. 12 of a plant breaking through the surface of a national
road near Cathcart.
Hybridization
C. J. Chamberlain, writing about “ The Living Cycads ” in 1919, after a journey
through Australia and South Africa, was of the opinion that hybridization does take
place in nature in present circumstances. Henderson writing on Encephalartos in
South Africa in the Journal of S. Afr. Botany, 1945, found no evidence for the assumption.
Nevertheless, Henderson was unable to classify specifically quite a number of specimens
in the herbarium and in cultivation. This apparently did not suggest to him that the
problem could be involved with natural hybridization. Possibly he would have changed
his opinion had he been able to do field work.
Fig. 13. — Hybridization: On left is Encephalartos trispinosus, right hybrid E. trispinosiis x E. altens-
teinii, the latter parent just excluded from picture; on bank of Kap River in the Bathurst
district, where other suspected hybrids were observed at the same time.
With complete information about specimens of Encephalartos from the Transvaal,
one is able to name most of them specifically with little trouble. Each species has its
range of variation but hybridization nowhere appears to have been involved in causing
variability. The same can be said of specimens from Natal, even if one does feel the
distinction between E. ahensteinii and E. natalensis, and between E. ingoyanus and
E. cajfer, is not very strong. Yet when one studies the populations of Encephalartos
in the lower reaches of the Buffalo, Great Fish and Bushmans River Valleys in the
422
eastern Cape, two or more species may share the same habitat and in such circumstances
one may be confronted with specimens which cannot readily be classified under any of
the generally accepted species.
In the Buffalo River Valley specimens have been found which suggest that hybridi-
zation between E. altensteinii and E. villosus has taken place; in Bathurst a specimen
in a garden suggested hybridization between E. altensteinii and E. latifrons. In Bathurst
and Alexandria there are specimens which strongly suggest hybridization between
E. altensteinii and E. trispinosus (Fig. 1 and 13), while E. arenarim may also be involved
in hybridization with E. altensteinii in the Bushmans River area.
Several botanists have studied these populations and none has found it possible to
unravel all the complicated relationships which appear to exist. The first botanist to
comment on it was Charles Zeyher whose notes were read in 1852 before a scientific
gathering and published in the Phytologist. After mentioning that any person with
“ tact ” should have no difficulty in detecting the “ play of nature ”, he went on to say
that he intended to examine all the populations once more in their natural state. The
outcome was not recorded. The problems, being what they are, with slow growing and
slow maturing plants, are unlikely to be solved in the near future. It is not feasible to
give a specific identification for all living specimens of Encephalartos and the reason for
this is assumed to be largely due to the complicating effects of natural hybridization.
Age and Rate of Growth
Because Cycads have survived with relatively little modification in habit over the
past 50 million years or so, they are commonly referred to as living fossils. This may
give rise to misconceptions of very great age for living plants. Few individual trunks
standing 25 ft today, and there are not very many, can be as much as 500 years old
according to measurements and photographs of individual plants of known history.
On the other hand the rootstocks of species which produce suckers from the base, in
sequence as the old stems mature, may attain far greater age, running into thousands of
years. No means of calculating the age of rootstocks in the wild state has yet been
devised.
The length of the interval between the production of new whorls of leaves from a
stem varies between species and between different specimens of the same species. The
time-lag can be as short as six months under exceptional circumstances (when a whorl
is destroyed on emergence) though it is rarely less than one year, and can be as long as
several years. In the case of cones, these may, at a maximum rate, alternate with the
production of whorls of leaves, but it is usually far less frequent and some specimens of
fifty years or more old are not known ever to have produced cones. Although the over-
all rate of growth of Cycads is always slow, once young leaves or cones appear at the
apex of a stem their rate of growth to maturity is relatively fast. Leaves elongate
visibly from day to day in the early stages.
The rate of growth of the stems also varies considerably between species, moreover
it has been found that stems from the same root-stock do not have a uniform rate of
growth. No accurate information is available on the rate of growth of species such as
E. villosus with a subterranean stem, but observations have been recorded for several
species with aerial stems.
E. altensteinii: A plant of E. altensteinii (Fig. 14) was photographed in 1915 near
the verandah at the Botanical Research Institute (as its title now is). A tree of Euclea
undulata grew up beside and over the Encephalartos so it has had competition as it
would have done in the wild state. The estimated height of the trunk in 1915 was 34 in.
p. Height Increase Period
in. in. Years
1915 34 — —
1947 52 18 32
1964 60 26 49
8 in last 17
423
This gives a rate of slightly more than 0-5 inches per year and there is an indication
of a slight decrease in rate during the last 17 years. During the whole period the stem
produced no cone or sucker. Growth is in rhythms and an increase in length is only
apparent after the emergence of a crown of leaves which is completed in a relatively short
space of time. There is usually a period of consolidation after the production of cones
or another crown of leaves.
Fig. 14. — Rate of Growth; left, Encephalavtos altensteinii at Botanical Research Institute 1915; right,
same plant 1947, showing an increase in height of about 18 in. in 32 years.
E. natalensis'. In 1951 measurements were made of eleven of the stems of the female
type specimen at Monteseel near Inchanga, Natal. At the time the clump within which
there were a number of undeveloped suckers, had a total circumference of 26 ft 1 1 in. and
it must have a huge root-system. By 1963 the largest stem of 1951, then 19 ft 9 in., had
been snapped in two places by a hurricane which struck the area in 1961. Other stems
had become more curved and difficult to measure accurately. Hence the heights for
only 3 of the upright stems are given for comparison.
This gives an average increase in length of about 19 in. for the 12 year period or a
rate of 1 - 6 in. per year, which is higher than expected and possibly resulting from the
favourable local growing conditions, relatively small girth, and the extensive rootstock
and rootsystem of the arent.
424
A young unbranched female plant a few yards distance gave the following result :
This gives a rate of elongation of the young plant of 0-9 in. per annum, and it will
be noted that the girth was at least 6 in. greater than in any of the stems measured from
the old rootstock.
E. longifoHus: At the Royal Botanic Gardens, Kew, a young plant of this species
was introduced by Masson in 1775.
Fig. 15. — The Old Mill, Trappes Valley, eastern Cape, with two Enccphalartos latifrons on left and three
E. altensleinii on right (one obscured) ; established about 1870 according to Chamberlain : this
photograph taken 1945.
425
The first period of growth was slow but the young plant had to re-root, adapt itself
to changed conditions and possibly rise above ground level. If one calculates the rate
of growth from 1822 when the stem was 12 in. tall, one arrives at the figure of 120 in. in
138 years which is within the range limits recorded for E. altenstemii and E. natalensis.
Even E. latifrom, which gives the impression of much slower growth, is estimated
from a photograph taken in 1915 by the celebrated C. J. Chamberlain of two specimens
in front of the old mill in Trappes Valley in the eastern Cape, to have increased in height
by about 25 in. in 50 years. This is at some variance with his reference to the specimens ;
There are five plants in the row, three of them Encephalartos altensteinii and the other
two E. latifrons. A pleasant grey-haird lady told me that they had been set out when she
came to that house as a bride forty-six years before. She said that the E. altenstemii
may have grown a foot in the forty-six years, but that the E. latifrons did not seem to
have grown any, although they always had green leaves ”. One’s memory often plays
one false as the photographic records prove in this case (Fig. 15 and 16).
Fig. 16. — Encephalartos latifrons: in Old Mill garden shown in previous figure; the trunk of the
specimen on the left was approximately the height of the window-sill when photographed by
Chaimberlain in 1915 and in 1964 was about 5 ft tall, giving a growth increment of very
roughly 25 in. in 50 years.
Conservation
The density of the Cycad population in southern Africa was far greater before
European settlement and greater still before Hottentot and Bantu occupation. Several
of the early travellers testify to the chopping and use of the starch-laden pith by the
indigenous races of man for the making of crude bread. The evidence that the Encepha-
lartos population in the wild has been depleted since the arrival of the European is
before our eyes, in the thousands of specimens which presently adorn the streets of
municipalities, give character to public gardens and are to be found in private gardens
throughout the country. In addition hundreds, possibly thousands more have been
despatched to botanic gardens and nurserymen throughout the world. And there is
no knowing how many more have perished by the wayside.
426
The history of exportation starts with Thunberg and Masson shortly after 1770 who
were followed by Scholl, Ecklon, Zeyher and others. The following extracts from a
Register kept by McKen, curator of the Durban Botanic Garden 100 years ago, is
enlightening in this regard.
This is over 230 in one year and the number mounted to over 670 in about 7 years
and specimens of Stangeria accompanied many of the consignments.
It may be mentioned here that over 4,000 specimens of E. lebomboensis were rescued
by Operation Wild Flower from the area to be flooded by the construction of the Jozini
Dam across Pongola Poort between the Transvaal and Natal (Fig. 78).
For scientific reasons the protection of Cycads in the wild state is preferable to all
other means; protection in Botanic Gardens, where the life history of each individual
specimen is followed and recorded, is a valuable contribution to knowledge; the transfer
to private gardens of plants otherwise doomed to perish in the face of modern develop-
ments, is also justified ; but sooner or later the ownership of a private garden changes and
the history, and with it the scientific value of specimens, is lost to posterity.
In the Middelburg district of the Transvaal, E. eugene-maraisii is represented on a
large farm by less than a dozen old clumps, where it is impossible to find any seedling
regeneration whatsoever. The most recently discovered species, E. inopinus, from the
Lydenburg district of the Transvaal, appears to be one of the rarest. The reasons for
this are mere speculation but, until one has proof to the contrary, one may conjecture
that it is facing natural extinction unless strict conservation measures are followed.
E. woodii from Zululand, has apparently become extinct in the wild state within living
memory.
In South Africa the separate Provincial authorities have their own Plant Protection
Ordinances, under which all Cycads are protected. It is pleasing to record that an
area has been reserved on the Department of Forestry ground in the Lydenburg district
427
for the protection of E. humilis, one of the smallest species in the genus. The largest
species, E. transvenosus, the Modjadji Cycad near Duiwelskloof shown on the frontis-
piece, has been protected by the Bantu Rain Queen and her people from prehistoric
times, with the result that the forest of large trunks is a unique sight in the Republic
today. The Ida Doyer Nature Reserve has been established near Barberton where
E. paucidentatiis is indigenous. E. longifolius is similary protected by the Port Elizabeth
municipality and a reserved area was proclaimed at Monteseel to include the type speci-
men of E. natalensis. No doubt further good examples could be cited and this should
be sufficient encouragement for other public bodies and private persons to take similar
steps to preserve local species in their natural surroundings.
The desirability of abiding strictly by measures of conservation for all our Cycads
must be abundantly clear to everyone.
MISCELLANEOUS INFORMATION
Toxicity (Poisonous Properties)
There are many references to the edibility of the fleshy outer covering of the seeds of
some species of Encephalartos. Baboons, monkeys, rodents and some birds (Louries)
are said to be particularly fond of them. Louries swallow some seed whole. There is
no doubt that the bright red and amber-coloured seeds are most attractive to look at.
It is not advisable, however, for humans to sample them and children should certainly
be warned not to do so. The flesh of some species is toxic and it is not possible for
children to discriminate between the poisonous and non-poisonous flesh.
Tests have shown that almost invariably the inner white kernels are poisonous.
Those actually proved poisonous include the following: E. cycac/ifoliiis, E. friderici-
guilielmi, E. horridus, E. lehmannii, E. longifolius, E. villosiis and E. eugene-maraisii.
It is a virulent liver poison, the chemical formula of which has not been established.
The species most likely to have been the cause of illness in General Smuts and some of
his men, when on commando during the Anglo-Boer war and recounted by Reitz in his
book ‘ On Commando ’ is E. longifolius.
On February 3rd, 1965, M. J. Wells reported the death from Encephalartos poison-
ing of cattle on Mr. G. Pohl’s farm near Riebeek-East (nort-west of Grahamstown).
Two died and the post mortem showed that the stomachs of the animals contained large
quantities of the partially digested seed of E. longifolius.
In parts of Australia, species of Macrozamia have caused extensive mortality among
sheep and cattle, the former through eating the seed and the latter through the grazing
of leaves. It is apparent that the extent and density of the Cycad growth in Australia
is far greater than is the case in South Africa. The question of eradication has received
serious consideration there in contrast to the strict protective laws passed in the Republic.
Edibility (Starch)
As mentioned, the fleshy covering of the seeds is always suspect for human con-
sumption and eating any part of a kernel might prove fatal. On the other hand many
species of Cycad in different parts of the world are exploited by the aborigines lor the
428
starch content of the pith within the trunks. Early travellers in southern Africa (1772-
1779) Thunberg, Masson, Swellengrebel, Paterson, refer to the use of the pith from the
stems of Encephalartos species for making crude bread. This gave rise to the common
names Hottentot- or Kaffir-bread (Kaffirbroodboom) or simply bread-tree (broodboom).
One may assume from the reports that all species of the eastern Cape Province were used
for this purpose. The method employed amounted in essence to the splitting open of
the trunks and the removal of the pith. This was placed in an animal skin and buried
in the ground for a period of a month or six weeks to induce it to ferment partially.
The fibrous mass was then ground between stones and mixed with water into a paste.
The paste was divided into the desired sizes and roasted under glowing embers.
Yeast
A side-light on the economic aspects of the Cycads is the presence of primitive
yeasts in the gut and frass of some of the inhabiting insects. A sample from a cone
produced in the garden of Mr. G. G. Smith of East London, infected with a member of
the Cossidae and possibly other insects, was supplied to Dr. J. P. van der Walt on the
staff of the C.S.I.R. He found several strains of an unrecorded yeast which he described
as Endomycopsis wickerhamii in Antonie van Leeuwenhoek Vol. 25 : 1959.
Root Tubercles or Nodules
In addition to thick tuberous and thin fibrous roots, Cycads, including both Stangeria
and Encephalartos produce roots with nodules which grow upwards to or near the soil
surface (Fig. 17). These appear in the early seedling stage and may develop into a
dense mass for a short distance round the base of old stems. They are somewhat
spongy and harbour a green alga, and possibly a bacterium also, without doing any harm
to the Cycad. The association is regarded as a symbiotic relationship from which all
the organisms derive benefit. The Cycad is probably assisted in nitrogen assimilation
and fixation. Little research has been done in this field in South Africa and the specific
identity of the alga and bacterium in the different species has not been determined.
Decoration
Cycads are cultivated the world over as ornamental plants and many of the species
have the advantage of enduring wide temperature fluctuations. There is an extensive
trade from the East to the United States of America and parts of Europe in the leaves of
Cycas revoluta, mainly for wreaths. About 3 million are shipped to the United States
annually.
Insects on Cycads
No systematic collection of insects has been made on Cycads throughout South
Africa and there is little doubt that if this were done a greater range would be found.
Nearly 200 years ago Thunberg described a weevil with a long slender snout as Curculo
zamiae which was later transferred as Anthiarhinus zamiae. This insect consumes com-
pletely the endosperm from Cycad seeds before the adult insects emerge. Another type
of weevil with a curved snout is Phloeophagus hispidus and its relative P. ebeninus and
yet another is Derelomus languidus, all of which occur in the eastern Cape Province and
429
have been considered as agents of pollination. A cone of E. eugene-maraisii from the
Transvaal yielded a new genus and species of insect, Apinotropis verdoornae, and in 1963
lavae, taken from cones of E. laevifolius at the Botanical Research Institute, proved to
belong to a moth of the genus Hercidia and is probably undescribed.
Fig. 17. — Root tubercles or nodules: here these are on Stangeria eriopiis and also occur on species of
Encephalartos: they contain a species of Alga and possibly also a species of Bacterium. Photo
R. G. Strey.
The moth Zeronopsis leopardina lays its eggs on the young leaves of Cycads and the
social worms which hatch soon eat the young foliage and make the plants look most
unsightly. The moth is particularly active in Natal.
STANGERIA
There is only one species in this genus which occurs in the eastern Cape and Natal.
The description of the species serves for the genus also. The genus is named in honour
of Dr. William Stander who sent specimens overseas and who became surveyor general
of Natal.
Stangeria eriopus (Kunze) Nash in New York Bot. Gard. 10 : 164, pi. 62 (1909).
Lomaria eriopus Kunze in Linnaea 13 : 152 (1839).
Fig. 8, 18, 19, 20.
430
Fig. 18. — Sla/igeria en'opiis: Prof. R. H. Compton and Miss Courtenay-Latimer investigate the tall
forest form in the Kentani district.
431
Fig. 19. — Stangeria eriopiis: a grassland form near Ngoye in Zululand, with male cone.
6821035-3
432
A perennial herb with male and female cones on separate plants; plants have a
large variously shaped tuberous main root up to 10 cm diam. continuous with the main
stem. Stem subterranean, branched or unbranched, with 1-4 leaves from the top of
each stem or branch. Leaves somewhat fern-like, very variable in size, 25 cm-2 m tall,
divided into separate pinnae or leaflets ; leaflets in 5-20 opposite or subopposite pairs,
the lowermost shortly stalked, the uppermost somewhat united, entire on the margin or
serrulate or lobed, acute or rounded at the tip, the median ones up to 40 cm long and 6 cm
broad. Cones solitary from each stem, stalked, silvery pubescent at first becoming
brownish with age: male cone more or less cylindric, narrowed to the apex, 10-15 cm
long and up to 3 cm in diam. with numerous spirals of scales overlapping upwards and
toothed on the margin: female cone ovoid-elliptic, up to 18 cm long, 8 cm diam. ; scales
stalked and shield-like, overlapping tightly upwards, irregularly toothed on margin:
seeds red, up to 3-5 cm long, 2-5 cm diam. with fleshy covering and hard thin inner
shell enclosing the endosperm.
Stangeria eriopus occurs in coastal grassveld and forest from Bathurst district in
the eastern Cape to northern Zululand. No doubt vast numbers of plants were eradi-
cated along the Natal coast as sugar cane fields replaced the natural vegetation, and the
expansion of the pineapple industry in the eastern Cape is causing a similar diminution
in the distribution of Stangeria eriopus in that area.
Controversy has existed for over a hundred years whether to recognize one or more
species of Stangeria because of the very obvious difference between some plants found in
open grassveld and those found in high forest. It seems that the different forms are
adaptations to the varying habitat conditions and there is no clear line of distinction
between the different growth forms. In open grassveld the leaves are often not more than
1 ft tall, the leaflets are hard and rounded at the tip and the rootstock may have up to
10 or 12 short underground branches. Within forests the leaves may attain a height of
over 6 ft with correspondingly luxuriant leaflets and single stems. But the cones, which
are most important in classification, are relatively uniform throughout, for which reason
no subdivision seems justified.
ENCEPHALARTOS
Plants either male or female with unbranched or branched trunks growing above or
below ground level and covered by alternating series of bracts and persistent leaf-bases.
Leaves produced in whorls, few to many together, 40-250 cm long, with numerous
leaflets in two rows along the leaf-stalk; the leafstalk with a swollen base or pulvinus
which persists and protects the stem ; leaflets of seedlings toothed round the apical part
of margin, becoming modified as plant matures, with entire, toothed or lobed margin,
light green, dark green or with bluish bloom for several months, with or without hairs,
reduced in size towards the base, sometimes into a long series of prickels. Pollen- and
seed-scales aggregated into dense spirals, forming cones. Cones on stout peduncles
either male or female on a plant, 1-several together from the apex of the stem (rarely
produced at the same time as leaves); male scales with numerous pollen-bearing bodies
on under surface of basal half and produced into a sterile apical beak: female scales
producing two large seeds lying above the stalk of the scale and directed towards the
central axis, apical expanded portion or bulla with an upper, lower and a variously
prominnent terminal facet and incurved lateral lobes which protect the seed. Seeds
oblong to somewhat rounded with a variably fleshy beak and a hard inner shell surroun-
ding the large endosperm (food reservoir).
Encephalartos is restricted in its distribution to the African Continent south of the
Sahara but does not occur in S.W. Africa.
The name is derived from the common name Bread-tree or Broodboom referring to
reports by early travellers of the making of crude bread by the Hottentots and Bantu
from the starchy pith within the trunks.
433
Key to Species of Encephalartos
For those who have not had experience with ‘ keys to species ’ it may be well to
say that it is at best a guide or aid to the identification of the species. It has not been
thought practical to present this one in any more simplified form. An identification
made by way of the key must be checked against the full description, illustrations,
distribution records and any other means available.
An effort has been made to arrange and number the species in the key in a sequence
to indicate the assumed relationships. But a key necessitates a linear arrangement,
whereas the relationships of the species of Encephalartos are two or three or more
dimensional and one cannot guess how far back in the history of evolution of the various
groups of species one would have to go to find the common ancestor.
Fig. 20.— Stangeria eriopiis: left, male cone; right, female cone showing the scales tightly overlapping
those above.
434
Thus in the 26 species listed, we find that Encephalartos inopinus is somewhat of an
enigma at the moment as regards its closest affinity because female cones are as yet
unknown: E.ferox is without any close relatives among the other South African species,
and the small group of species including E. villosus and E. caffer, mainly with subterranean
stems, seems rather isolated from all others. Yet when E. villosus meets E. altensteinii
under natural conditions, hybridization seems as good as proved. The key does not
provide for the identification of hybrids or aberrant forms. A natural hybrid between
E. altensteinii and E. villosus is unlikely to ‘ run down ’ to either parent in the key and
would be more likely to lead one to E. natalensis or E. lebomboensis, because of the
combination of characters. Similar confusion may arise with other natural hybrids
such as between E. altensteinii and E. trispinosus. Of all species E. trispinosus is the
most difficult to define.
Leaflets 1-A mm broad, 8-14 cm long, linear, entire with revolute margins, dark green;
crown of stem brown woolly; cones 2-5 together, densely woolly; male cone
scales on stipe 1-1-5 cm long 1. ghellinckii.
Leaflets 4 mm or more broad, margins usually thickened and sometimes recurved but
not revolute :
Leaflets towards the base of the petiole deflexed in relation to the rhachis, the basal
leaflets reduced to a short or long series of prickles, those about the middle of the
rhachis spreading at right angles to the axis and recurving, 14-21 cm long, 8-13
mm broad, linear, attenuate, 13-21 -veined on the under surface and glaucous
(silvery or bluish) on under surface when young, stems branched from the base,
up to 2 m long, rarely longer; face of male cone scales moderately greyish
pubescent 7. inopinus.
Leaflets spreading at right angles or ascending in relation to the axis of the rhachis,
none deflexed, glaucous or green:
Leaflets 4-8 mm broad, linear, entire, glaucous or somewhat silvery when young,
sometimes yellowish with age, rarely reduced to more than one prickle at
base, venation conspicuous on under surface; cones densely woolly or only
moderately so in E. laevifolius, seeds light brown or brown-madder to
ochre-yellow :
Cones with dense woolly covering:
Stems well developed above ground:
Stem crown conspicuously and rather persistently brown woolly;
trunk 35-60 cm diam. up to 3 m or more tall; cones 3-several
together; median leaflets 10-17 cm long and about 7 mm broad,
7-9-nerved on under surface 3. friderici-guilielmi.
Stem crown not conspicuously brown-woolly, trunk rarely up to 30 cm
diam. and not often up to 2 m tall:
Median leaflets 9-13 cm long, 4-5 mm broad, 5-6-nerved on under
surface; margin somewhat recurved; cones often single
2. cycadifolius.
Median leaflets 10-13 cm long, 6-8 mm broad, finely 10-14-nerved
on under surface, flat; cones 1-4 together 4. lanatus.
Stems rarely up to 20 cm above ground, forming dwarf branched clumps;
cones single; median leaflets 9-13 cm long, 4-6 mm broad, finely
about 9-nerved on under surface 5. humilis.
435
Cones with short greyish tomentum, not conspicuously wooly; median
leaflets up to about 12 cm long, 5-7 mm broad, 10-1 2- nerved on under
surface ; trunk 1-3 m tall, rarely more than 30 cm diam . . 6. laevifoUus.
Leaflets 1 cm or more broad, if less than 1 cm broad -then bright green not glau-
cous; _ margins entire or toothed; cones glabrous or apical part of scale
face with reddish or brownish hairs:
jStems usually well developed above ground if subterranean then leaves
glaucous as in E. horridus:
Leaves glaucous with bluish bloom when young and persisting so for
several months at least, contrasting with the old leaves:
Leaves positively glaucous when young, leaflets rarely reduced to more
than one prickle at base, median leaflets usually less than 2 • 5 cm
broad :
Median leaflets mainly entire or occasionally with 1 or 2 teeth on
lower margin :
Leaves slightly or markedly incurved near the tip; female
cone-scales with slightly rugose face.. 8. eugene-maraisii.
Leaves slightly or markedly recurved towards tip:
Female cone-scales with irregularly warty face, the pulvinus
of petiole often partly hidden by stem bracts, leaves
often pilose when young, leaflets usually overlapping
above middle of rhachis ................ 9. princeps.
Female cone scales with smooth or slightly rugose face,
pulvinus of petiole exposed and prominent, with brown
collar, leaves glabrous, leaflets often not overlapping.
10. lehmannii.
Median leaflets, some or most, with 1-3 spine-tipped lobes on the
lower margin, glabrous:
Stems mainly above ground ; median leaflets 1 • 5-2 ■ 5 cm
broad, some or all with 1 or 2 spine-tipped lobes on lower
margin, and lobes sometimes twisted out of the plane of the
leaflet ; female cone-scales yellowish-green, rugose on face
with the lateral incurved lobes 2-3 cm long
1 1 . trispinosus.
Stems mainly subterranean, occasionally up to 45 cm tall;
rarely taller, median leaflets rarely less than 2-5 cm broad,
deeply 1- or 2-lobed on lower margin and lobes twisted out
of plane of the leaflets, female cone-scales blackish or
reddish-brown, ridged towards terminal facet but otherwise
with a fairly smooth face, with the lateral incurved lobes
about 1 cm long... 12. horridus.
Leaves light or medium green, faintly glaucous; median leaflets usually
more than 2-5 cm broad; lower margin with 3 or occasionally 4
spine-tipped lobes: 13. arenarius.
Leaves light or dark green, with or without prickles or lobes along one or
both margins, not glaucous except sometimes in young stage of E.
longifolius and E. arenarius:
436
Lower margin of median leaflets with 2-4 triangular pungent lobes,
upper margin without lobes or teeth:
Leaves medium green, faintly glaucous, glabrous; leaflets not
prominently veined on under surface 13. arenarius.
Leaves dark green, finely pubescent when young, rigid, leaflets
prominently veined on under surface 14. latifrons.
Lower and upper margin of median leaflets entire or with one or more
prickles or teeth :
Leaflets reduced in size towards base of rhachis but not to more
than one or exceptionally 2 prickles :
Leaves 1-1-5 m long, suberect-spreading and recurved or
arcuate in the upper half ; leaflets dark green often with
bluish lustre and sometimes glaucous when young, finely
pubescent when young, median ones mainly oblong-
lanceolate, up to 20 cm long and 4 cm broad, generally
overlapping on upper half of rhachis, margins not or
sparingly toothed 15. longifoUus.
Leaves 1-1-5 m long, nearly straight; leaflets bright green,
median ones oblong-lanceolate to lanceolate up to 15 cm
long and 2-5 cm broad, rarely overlapping, variably
toothed on margins; incurved lateral lobes of cone scales
lacerate-winged, about 3 cm long and extending nearly
to axis 16. altensteinU.
Leaflets reduced to several or many prickles towards base of
rhachis :
Leaflets with a single terminal spine or point, broadest below
or occasionally near middle:
Leaflets often showing veins on lower surface but not in
regular raised formation:
Leaflets spreading and straight from the rhachis, stems
often branched from base, rarely up to 7 m tall:
Leaves curved, bow-shaped giving an umbrella-like
canopy; median leaflets dark green, more or less
ovate-lanceolate, up to 20 cm long and 5 cm
broad 17. woodii.
Leaves erect, spreading straight or slightly curved in
the upper third :
Median leaflets lanceolate to ovate-lanceolate,
16-23 cm long, 2 -5-4 -5 cm broad; face of
female cone-scales with prominent blunt
rugosities; incurved lateral lobes toothed or
laciniate, 2-3 cm long.... 18. natalensis.
Median leaflets mainly narrowly lanceolate, 12-17
cm long, 1-2-2 -2 cm broad; face of female
cone scales without prominent excrescences;
incurved lateral lobes, laciniate, about 1 - 8
cm long 19. lebomboensis.
437
Leaflets spreading-recurved from rhachis, lanceolate to
ovate-lanceolate, sometimes somewhat falcate, 10-20
cm long, 2-3-5 cm broad; stems rarely branched,
many over 6 m tall 20. transvenosus.
Leaflets with about 20 regular conspicuously raised longitu-
dinal veins on lower surface, median leaflets linear-
lanceolate, straight or somewhat falcate, 15-25 cm long,
2-3 cm broad, rarely more 21. paucidentatus.
Leaflets with 3-5 almost equal pungent lobes at the end, usually
3-4 cm broad, broadest above or sometimes near middle,
slightly narrowed to the base; cones salmon or shrimp-pink
to scarlet 22. ferox.
ttStems dwarf, subterranean or rarely up to 20 cm exposed above ground,
unbranched or sometimes branched; leaflets light or dark green, not
glaucous, linear or linear-lanceolate, with or without marginal teeth:
Leaflets reduced to many prickles towards base of petiole; median ones
averaging 15-20 cm long, usually with one or two prickles on both
margins : female cone scales with the terminal facet extending almost
to the basal margin of the bulla face or overlapping slightly the scale
below; male cones 6-5-12 cm diam 23. villosus.
Leaflets reduced to one or two prickels at the base or sometimes none :
Leaflets, some or nearly all with one or two prickles on the margin,
average length 12-14 cm, occasionally up to 30 cm, 8-15 mm
broad, female cone scales with terminal facet generally slightly
below central, rarely overlapping scale below, male cones 6-8
cm diam 24. umbeluziensis.
Leaflets (except on juvenile plants) mostly entire, average length 6-8 cm
long, 8-11 mm broad:
Female cone scales with terminal facet on the face indistinct and
with the lower margin as low as or slightly overlapping the
scale below; male cones less than 8 cm diam. ; median leaflets
usually toothed on lower margin (Natal) 25. ngoyanus.
Female cone scales with terminal facet on the face distinct, not
extending to basal margin; male cones 6-11 cm diam.;
median leaflets usually without prickles on lower margin
(Cape) 26. coffer.
1. Encephalartos ghellinckii Lem., 111. Hort. 14, Misc. 80 (1867).
Fig. 21, 22.
Plant unbranched or branched from the base; stems up to about 3 m tall, often
leaning with age, 30-40 cm diam. with open brown woolly crown. Leaves somewhat
yellowish-green up to about 1 m long, straight or spreading-incurved and sometirnes
twisted, at first densely greyish woolly but hairs fall with age, leaflets dense, spreading
with revolute margin; the median ones 8-14 cm long and 2-4 cm broad, pungent.
Cones 2-5 together on short stout storks and densely brown woolly: male cones more or
438
less cylindric becoming curved, about 25 cm long and 6-7 cm diam., narrowed slightly
to both ends; the nose of the scales projecting only slightly: female cones up to about 22
cm long and 14 cm diam., rounded at the top; the scale face projecting only slightly.
Seeds golden-yellow, about 2 • 5 cm long and of similar width.
Fig. 21. — Enceplialartos gbellinckii: a female plant in cone on the Drakensberg, Natal. Photo: Dr.
A. O. D. Mogg.
Enceplialartos ghellinckii occurs on the southern foothills of the Drakensberg in the
eastern Cape Province near Flagstaff, extending to higher altitudes northwards along
the mountain range in Natal to Mont-aux-Sources.
Because of its narrow leaves with revolute margins there is little chance of mis-
naming E. ghellinckii among South African species. It is most nearly related to E.
cycadifoliiis and E. friderici-gidliehni, both of which are eastern Cape species with
similar woolly cones but broader leaflets. It could be confused, however, with the
sometimes cultivated species Cycas revoluta, which occurs naturally in Japan, but this
has its seeds in a loose cluster, which is quite different from the compact woolly cones of
Enceplialartos ghellinckii and its allies.
439
Fig. 22. — Encephalartos ghellinckii: habit view on the Drakensberg, Natal; trunks usually hidden by
lower vegetation.
2. Encephalartos cycadifolius (Jacq.) Lelim., Pugill. 6 : 13 (1834). Zamia cycadi-
folia Jacq., Fragm. 27, t. 26 (1801). Encephalartos eximius Verdoorn in Bothalia 6:
426 (1954).
Fig. 9, 23, 24.
Plants usually branching from the base; stems up to about 1-5 m tall and 25 cm
diam. Leaves 60-90 cm long, often somewhat twisted, greyish or slightly silvery and
somewhat woolly when young; median leaflets 9-13 cm long, 4-5 cm broad, entire, with
a thickened, slightly recurved margin and with 5-6 prominent veins on under surface.
Cones 1 or 2 from a stem, closely off-white or tawny woolly; nude cone more or less
cylindric, narrowed to apex, 13-22 cm long, 4-7-7 cm diam., with the nose of the scales
440
Fig. 23. — Encephalartos cycadifoHus: a relatively large plant for this species, which is little more than
three feet tall and freely suckered from the base.
only slightly projecting: female cones more or less barrel-shaped with rounded top,
20-30 cm long, 16-18 cm diam., with the face of the scales much flattened. Seed orange-
yellow to amber-brown, about 3 cm long and 2 cm broad.
Encephalartos cycadifoHus is rather local on mountains of the Cradock and Bedford
districts of the eastern Cape.
One should observe that in this species, as compared with E. friderici-guilielmii
with which it has been confused, the trunk shows practically no wool on top, further the
middle leaflets are 9-13 cm long, 4-5 mm broad, 5-6-nerved on the under surface, the
margins are slightly recurved and the woolly cones are usually produced singly and not
more than two at a time on one stem.
Because of the similarity of the leaflets and other features, plants of E. friderici-
guilielmi from Queenstown and Cathcart have been identified as E. cycadifoHus, for
example in Flora Capensis, 1933, although they had been recognized as distinct and
described under the name E. friderici-guilielmi as early as 1834. It is necessary to
reinstate the name E. friderici-guilielmi for the Queenstown-Cathcart communities as
explained more fully in the Journal of South African Botany 31 : 116 (1965). Because
of the confusion in the application of these names, the new name E. eximius was given to
plants which have since proved to agree with the original E. cycadifoHus. The name
E. eximius is, therefore, redundant and must be dropped.
441
Fig. 24. — Encephalartos cycaclifoliiis: female cones, one or two are produced from a stem, contrasting
with the 3 to 5 or occasionally more per trunk produced by the closely allied species E.friderici-
guilielmi. Photo: V. L. Pringle.
Fig. 25. — Encephalartos friderici-gidlielmi: a group near Tsolo, Transkei, with D. Collett lost in admira-
tion.
442
It is certainly curious that true E. cycadifoUus remained so little known for so long
after its first description in 1 801 . There appear to be two main reasons for this. Firstly,
the areas in which it grows naturally are off the modern main roads between settlements,
and secondly, the plants are not particularly robust and are not readily re-established in
cultivation, that is according to experience in the garden of the Botanical Research
Institute. This makes it all the more remarkable that a small plant should have survived
the journey from its habitat before 1800, firstly to the Cape of Good Hope and from
there by slow sea transport to Vienna. The plant was taken by the Viennese collector-
gardener, Scholl, together with specimens of E. horridus and E. longifolius for cultivation
in the Royal Garden at Schoenbrunn. Here they were described and illustrated by the
famous botanist Jacquin.
Interesting observations by V. L. Pringle on the effects of fire on E. cycadifoUus have
been recounted in the introductory chapter of this account and facts about the germina-
tion of the seeds have also been mentioned.
3. Encephalartos friderid-guilielmi Lehm., Pugill. 6 : 8, t.l, 2 & 3 (1834).
Fig. 12, 25, 26, 27, 28, 29.
Fig. 26.~Enceplialartos friderici-guilielnii: not an abnormally large plant with a trunk of about 12 ft
tall in Cathcart district, eastern Cape.
443
Fig. 27. — Encephalartos friderici-guiliehni: female plant in cone near Tsolo, Transkei.
444
Fig. 28. — Encephalartos friderici-guilielmi: male cones, a normal head for this species.
Plant unbranched or branched from the base; trunks stout, up to about 4 m tall on
rare occasions, 35-60 cm diam., eventually leaning or procumbent with age, with open
brown woolly crown. Leaves fairly stiff, straight or curved, often spreading lower than
crown when stem has cones, 1-1 - 2 m long; the leaflets somewhat bluish or silvery when
young, and yellowish with age, closely spaced and overlapping in the upper half of the leaf
and reduced in size towards apex and base; median leaflets 10-17 cm long and about
7 mm broad, with 7-9 prominent veins on lower surface. Cones 3-several together,
up to 10 on large male trunks, covered with yellowish-grey to brownish wool: male cones
somewhat cylindric narrowed slightly to the apex, curved with age, 20-30 cm long, 6-7
cm diam., with the beak of the scales projecting about 5 mm: female cones barrel-
shaped, with rounded apex, 25-30 cm long and 15-17 cm diam., with the face of the
scales much flattened. Seeds pale yellow to pale orange-yellow, 3 cm long and about
2 cm diam.
Encephalartos friderici-guilielmi occurs on mountains and rocky hill-slopes of the
Queenstown and Cathcart districts of the eastern Cape, extending eastwards at intervals
to the neighbourhood of Kokstad. The more westerly records found in literature are
almost certainly incorrect and due to some confusion.
One should look for the following features. The crown of the stem is usually open
and conspicuously brown-woolly, the trunks are robust and 35-60 cm in diam., the
middle leaflets are 10-17 cm long, about 7 mm broad, 7-9-nerved on the under surface
and a few to several (up to 10) cones may be produced at one time.
445
Fig. 29. — Encephalartosfriderici-giiilielmi: female cones breaking up on maturity to shed seeds. Photo:
W. Everitt.
It is unfortunate that E. frkkrici-guilielmi has been confused with E. cycadifoHus
during the past half century and has been widely referred to under that name during
this period. The name E. cycadifoHus belongs correctly to plants on the mountains
further west in the Bedford and Cradock districts. The distinguishing characters of
E. friderici-guilielmi, particularly the very woolly crown of the stem, were remarked
upon by Lehmann, when he described it in 1834. E. friderici-guilielmi is remarkable
among its allies in the genus for its free coning; large male specimens bearing up to 10
cones at a time. It is far more robust and appears to be more adaptable to a change of
environment and cultivation than E. cycadifoHus, which is more exacting in its cultural
requirements, judging by experience in the garden of the Botanical Research Institute
One is diffident about saying that the species is locally frequent for fear that some-
one will take it as an invitation to remove specimens from their natural habitat. In its
lower altitude range, near the junction of the White and Black Kei Rivers it was reported
by Galpin to occur with E. princeps. In this connection Chamberlain (1919) remarked:
“ His warning prepared me for the striking variation which this species displays in
different localities, and guarded me against confusing it with a nearly related species which
closely resembles it and is associated with it in some places ”.
4. Encephalartos lanatus Slapf & Burtt Davy in Burtt Davy FI. Transvaal & Sw-
1 : 40, 90, Fig 4 D (1926).
Fig. 30, 31, 32, 33.
446
Fig. 30. — Encephalartos lanatus: an exposed site near the Wilge River in the Bronkhorstspruit district,
Transvaal.
Fig. 3 1 . — Encephalartos lanatiis: a dense stand in a sheltered kloof, in the Middelburg district, Transvaal
447
Fig. 32. — Encephalartos lanatus\ Male cones from a plant in the Middelburg district.
Fig. 33. — Encephalartos lanatus: female cones on a plant in the Middelburg districts
6821035-4
448
Plants unbranched or sometimes branched from the base; stems up to about 1 m
tall and only very rarely up to 3 m long and then reclining, up to 25 cm diam. Leaves
60-95 cm long, white or greyish woolly when young except under surface of leaflets;
recurved and often slightly twisted; leaflets with a light-silvery bloom when newly
developed, entire, pungent, set in V formation above middle of leaf and usually over-
lapping; median leaflets 10-13 cm long, 6-8 mm broad, finely 10-14-nerved on under
surface. Cones 1-4 together, covered by a dense light cream to greyish tomentum:
male cones more or less cylindric, narrowed to both ends, 25-30 cm long, 5-6 cm diam;
median scales with the face flattened and only slightly projecting: female cones barrel-
shaped, 25-35 cm long, 12-15 cm diam. with the face of the scales only slightly protruding.
Seeds yellow, 2-5-3 cm long, 2 -25-2 -5 cm broad.
Encephalartos lanatus is found in sheltered valleys of the upper reaches of the
Olifants River catchment in the Bronkhorstspruit, Witbank and Middelburg districts of
the Transvaal: sometimes it is common locally.
This species has been confused with both Encephalartos laevifolhis and E. humilis
and falls between the two in stature. Its woolly cones are not unlike those of E. humilis,
but are readily distinguished from those of E. laevifolius which are far less woolly. The
young leaves are more woolly than those of either of the other two species and it is
distinguished by the median leaflets being 10-13 cm long, 6-8 mm broad and having
10-14 veins visible on the under surface. Both the other two species occur in the upper
catchment area of the adjacent Crocodile River.
In some areas of its native habitat from Bronkhorstspruit to Middelburg district,
this species has to withstand grass fires annually. It is, however, one of the most
sensitive species to removal from the veld and attempts to re-establish it in gardens often
end in failure.
5. Encephalartos humilis Verdoorn in Bothalia 6 : 220, pi. 3, 241 (1951).
Fig. 34, 35, 36.
F.g. 34. — Encephalartos humilis: in the small reserved area of the Uitsoek Forest Station, Lydenburg
district, Transvaal.
449
Fig. 35. — Encephalartos hiimilis: single mature female cone on a plant in the Uitsoek Forest Station
Cycad Reserve. The early woolly covering has largely disappeared with age.
A dwarf plant branching from the parent root-stock and forming small clumps;
stems subterranean or occasionally up to about 30 cm above ground and of a total
length not exceeding about 50 cm and 13-18 cm diam. Leaves 30-50 cm long, loosely
greyish-woolly when young, losing hairs with age, recurved and sometimes somewhat
twisted in the upper part, with a greyish or silvery bloom when young; leaflets in V
disposition above middle and sometimes slightly overlapping, entire; the median ones
9-13 cm long, 4-6 mm broad and finely about 9-nerved on the under surface. Cones
single on a stem, densely greyish woolly: male cone 15-20 cm long, 4-5 cm broad, nar-
rowed slightly to both ends on stalks about 10 cm long; the face of the cone scales only
slightly projecting: female cone barrel-shaped, 25-30 cm long, 8-10 cm diam. on short
stalk, the face of the scales very slightly projecting. Seeds orange-yellow, 2-5-3 cm
long, 2-2 • 5 cm broad.
Encephalartos humilis is occasional on the mountains of the eastern Transvaal from
the Carolina district to Lydenburg and Nelspruit, usually associated with grassveld and
often wedged between sandstone rocks.
450
Prior to the recognition of E. humilis as a distinct species, specimens of it were
usually classified as E. lanatus. Apart from its dwarf habit and a difference in hairyness
(being less woolly and with straighter hairs on the leaflets) the median leaflets have only
about 9 nerves on the lower surface as compared with 10-14 in E. lanatus. It is of
interest also that it appears to be restricted in distribution to the catchment area of the
Crocodile River, in contrast to the Olifants River for E. lanatus. Its nearest associate in
the veld is E. laevifolius on the Kaapsehoop Mountain.
Fig. 36. — Encephalartos hi/mi!is: (left) an old male plant with cone, having survived the effects of grass
fires over a long period; the protective leaf bases may be up to about 5 cm or 2 in. long;
(right) a relatively young plant, mainly underground with fairly large tuberous roots.
451
1 40 &^9"9 0926)*^**** laevifolius Stapf & Burtt Davy in
Burtt Davy FI.
Transvaal & Sw.
Fig. 37, 38, 39.
Fig. 11 —Encephalartos laevifolius: specimens on Kaapsehoop Mountain in the Nelspruit district the
trunks attaining a height of 8-12 ft.
452
Fig. 38. — Eucephalartos laevifolius: specimens on Kaapsehoop Mountain; the very spreading old
leaves indicate that a new whorl of leaves or a cluster of cones is likely to be produced shortly.
Plants unbranched or branched from the base; stems 1-3 m tall, occasionally up to
3-5 m and 25-35 cm diam. Leaves up to about 1 m long, nearly straight or slightly
recurving and sometimes twisted in the upper half, with a greyish or silvery bloom when
young and at first with a short dense woolly covering except on the under surface of the
leaflets; leaflets in V disposition above middle of leaf, dense but not or only slightly
overlapping, pungent, entire, up to 12 cm long and 5-7 mm broad, rarely up to 1 cm, with
10-12 nerves on under surface. Cones 2-4 together, thinly covered by short whitish
woolly hairs : male cones subcylindric, tapering slightly to apex and base, 30-40 cm long
and 10 cm in diam., curved with age; the beak of the scales projecting about 1 cm:
female cone barrel-shaped with rounded apex, 20-30 cm long 10-13 cm diam.; the scale-
face progressively less hairy towards its base, flattened and not very prominent. Seeds
orange-yellow, 2 -5-2 -7 cm long, 2-2-3 cm broad, with short fleshy apex.
Encephalartos laevifolius is fairly local on the Kaapsehoop Mountains in the Nel-
spruit district, and within the catchment area of the Crocodile River. It also occurs
near Havelock Mine in Swaziland.
Because of the similarity in the general appearance of the leaves, this species has been
largely confused with Encephalartos lanatus. It is normally far more robust than the
latter and is readily distinguished when in cone, the cones being far less woolly. The
number of nerves (10-12) on the under surface of the median leaflets, also helps to dis-
tinguish it from E. lanatus and E. humilis. In the absence of cones the latter might be
mistaken for a young stage in its life history. The plants are hardy and attractive and
their numbers have been depleted considerably for cultivation and also to make way for
afforestation. The nearness of the Kaapsehoop Mountains to Barberton may have given
rise to the statement that the species occurs in the Barberton district. On the other
453
hand it has recently been recorded near Havelock Mine in Swaziland near the Barberton
district border. E. paucidentatus occurs occasionally on the mountains and in valleys
above Barberton but this is very different in character.
Fig. 39. — Encephalartos laevifolius: A cluster of female cones with only a thin covering of woolly hairs
on the face of the scales.
7. Encephalartos inopinus R. A. Dyer in Bothalia 8 : 169 (1964).
Fig. 40, 41.
Plant freely branched from base with the stems up to about 2 m long, rarely up to
3 m and then spreading or reclining, 17-25 cm diam. Leaves 0-8-1 -5 m long with the
rhachis nearly straight or slightly upcurved and twisted near the apex; leaflets with a
silvery or bluish bloom on the under surface when young, mainly entire or occasionally
with a minute prickle on lower margin; those above the middle directed sightly upwards
and curving outwards, those about the middle spreading and recurving gracefully more
or less at right angles from the rhachis, those below descending and curving slightly
downwards and gradually reduced in size to 2-6 prickles on either side of the rhachis;
the middle leaflets 14-21 cm long and 8-13 mm broad in the lower half and tapering gra-
dually to the slender pungent apex, with 13-21 veins on under surface. Cones: male,
immature, 2 or more together, on peduncles about 5 cm long and 1 • 5 cm thick, 8-10 cm
long, 4-4-5 cm diam.; scales dense (few showing some dehiscence) about 2 cm long
(tips all damaged) 1-7-2 cm broad, with acute lateral margin, maximum vertical thick-
ness 5 mm ; microsporangia dense on lower basal half ; beak projecting about 1 cm, with
upper facet humped in middle; lower facet nearly flat; glabrous round base of beak,
shortly greyish scurfy-pubescent towards apex: female not yet discovered.
Encephalartos inopinus is known only from the Lydenburg district of the Transvaal,
the positive records being from the Olifants and Steelpoort River Valleys between
Penge and Kromellenboog Asbestos Mines.
454
Fig. 40. — Encephalartos iiiopi-
iiiis: young male cones;
the only fresh cones seen
so far, from plant in the
Lydenburg district, Trans-
vaal.
Fig. A\.— Encephalartos inopinus: the type plant from near Kromellenboog in the Lydenburg district.
455
The finding of this species is one of the major botanical discoveries in South Africa
in recent times. Encephalartos inopinus differs in several details from all other local
species and at first glance reminds one of a Mexican Dioon. It is the deflection of the
lower leaflets and the graceful spreading of the long tapering upper leaflets which makes
this immediately distinctive among South African species of Encephalartos.
So far female cones of this plant have not been recorded and until this has been
done one cannot be sure of its correct classification. It may sound strange that, with a
plant of this size, only a few have been localized in the wild state. The first descovery
to be registered was made casually by a hiking party in 1955, when Dr. H. van Hoepen
fortunately brought home a small sucker for cultivation in Johannesburg. This was
not examined botanically until 1964. This led to the finding of a second one by Mr. Els,
another member of the original hiking party. He found it on a krans near Kromellen-
boog but could not locate the first one in the wild and it may be that it has died, been
removed or destroyed in the past few years. However, since then several more have
been investigated not far distant. The disappearance in living memory of a species
from the wild state is recorded in the case of E. wooclii and it is trusted that no other is
destined to suffer the same fate through wanton destruction by man.
8. Encephalartos eugene-maraisii Verdoorn in Journ. S. Afr. Bot. 11:1 (1945).
Fig. 42, 43, 44.
Fig. 42.~EncephaIartos eugene-maraisii: a speciman from near the type locality of the species in the
Waterberg, Transvaal, showing the typical up-curved tips of the leaves; older and longer
trunks often recline.
456
Fig. 43. — Encephalartos eiigene-maraisii: a large clump in the Middelburg district, Transvaal, where
the trunks are generally more robust and are known to have reached a reclining length of
nearly 30 ft ; the leaves are generally straighter than the typical form.
Plants unbranched or more often branched from the base with stems 2-4 m tall, rarely
more, but up to about 10 m long in some procumbent stems, 30-45 cm diam. Leaves
■ 7-1 • 8 m long, nearly straight with the apical portion of the rhachis slightly or markedly
up or inward-curved, with bluish or silvery bloom for some considerable time; leaflets
reduced in size towards base but not to more than one prickle: median leaflets 15-20 cm
long, 1 ■ 3-1 • 5 cm broad, cuneate at base, shortly narrowed to the pungent apex, entire or
occasionally with a single tooth on the lower margin. Cones 1-4 together; male cone
subcylindric, 22-42 cm long, 6-8 cm broad, median scales with the face projected into a
beak 7 mm long: female cone more or less egg-shaped, 30-50 cm long, 16-20 cm broad,
median scales with face slightly wrinkled and pimply, minutely hairy. Seed light
brown or date-coloured, 3-5-4 cm long and 2-3-3 cm diam.
Encephalartos eugene-maraisii occurs in rather remote mountainous areas in the
Waterberg, Pietersburg, Witbank and Middelburg districts of the Transvaal.
In some instances a critical eye is needed to detect the slight upcurve of the tip of
the leaves (rhachis) which is the characteristic feature of this species and distinguishes it
from Encephalartos lehmannii and its allies. Encephalartos eugene-maraisii is the only
one of its group in the Transvaal and it is already rare in parts of its distribution area.
In parts of the Middelburg district young plants are very scarce, which cannot be due
entirely to the depredations of collectors, and suggests that the strictest protective
457
measures will have to be taken to ensure its survival in those areas. Natural causes seem
to have diminished its numbers in the Pietersburg district even more than near Middel-
burg, while in the Waterberg the species has been observed to regenerate tolerably well
under favourable conditions. It is in the Witbank and Middelburg districts that the
largest specimens are found, and here also that the leaves are on the average straighter
than they are in the Waterberg district. Mr. Jimmie Hall records one procumbent
trunk to have reached the extraordinary length of 28 ft. A terminal 18 ft piece of trunk,
placed on top of the ground in Johannesburg, rooted and produced leaves for the first
time after four years.
In parts of the Middelburg district E. eugene-maraisii is found only a few hundred
yards distance from groups of Encephalartos lanatus but no evidence of hybridization
between the two species has been recorded, nor is suggested.
Fig. 44. — Encephalartos eugene-maraisii'. female cones from a plant in the Botanical Research Institute
Garden and originally from the Waterberg, Transvaal.
458
9. Encephalartos princeps R. A. Dyer in Journ. S.Afr. Bot. 31 : 111 (1965).
Fig. 45, 46, 47.
Fig. 45. — Encephalartos princeps: few miles north of Komghain the eastern Cape overlooking Kabousie
River Valley; main trunks nearly reaching their maximum upright height.
Plant unbranched or more often branched from the base with stem up to 3 m tall or
occasionally up to 5 m long with age and reclining, up to about 30 cm diam. Leaves
1-1-3 m long, fairly straight or recurved in the top third with often bluish or silvery
bloom for considerable period and pilose in early stage, leaflets reduced in size towards
base, possibly to one prickle, mainly overlapping in the top third in V disposition;
median leaflets about 15 cm long and 1-3 cm broad, pungent, entire or occasionally
toothed on lower margin. Cones 2 or more together, dull green: male cone subcylindric,
narrowing to both ends, 16-26 cm tall and 8-10 cm diam. with the face of median
scales projected into a beak about 1 -2-1 -5 cm long: female cone more or less barrel-
shaped, 30 cm long and 20 cm diam. with rounded apex; median scales with face
protruding about 2 cm, indistinctly ridged but coarsely and irregularly pimpled and with
sparse whitish or brownish hairs. Seeds red, about 4 cm long and 2 cm diam.
Encephalartos princeps is found on the doleritic cliflfs and outcrops in the Kei River
catchment area including parts of the districts of Queenstown, Cathcart, Komga, Tsomo
and Butterworth. The exact range of the species up the main Kei River Valley and the
459
Kabousie River Valley has not been determined nor how far it extends towards the
river mouth below the railway bridge, but it is quite frequent in parts. There is a record
as far east as Mquanduli in the Transkei but it is based on a cultivated plant and con-
firmation of its presence there is necessary.
Specimens now classified as E. princeps were formerly regarded by most workers as
belonging to E. lehmannii, but the case seems strong enough for specific separation. In
addition to the usually greater size of the pulvinus and leaf scars of E. lehmannii, E.
princeps becomes appreciably taller or longer when reclining, the leaves are often pilose
when young, the leaflets are denser and the face of the female cone scales is more warty.
There is an interesting difference in soil preference, E. princeps being found on doleritic
formations as opposed to the sedimentary sandstone for E. lehmannii and the largely
shale preference of E. trispinosus which is another closely related species. E. princeps
may well have been earlier in the evolutionary series than any of its eastern Cape relatives
but this is a speculative idea.
Encephalartos friderici-guilielmi is recorded by Galpin as an associate of E. princeps
in the south-eastern part of the Queenstown district. Confusion between the two is
unlikely in any circumstances and out of the question when cones are present, those of
the former being very woolly and of the latter inconspicuously scurfy-hairy only on the
face of the cone scale.
Fig. 46— Encephalartos princeps: on the west bank of the Great Kei River below the railway line in the
Komgha district, possibly all trunks originate from the same rootstock, whose ultimate roots
have probably reached the water-table.
460
Fig. 47. — Encephalartos princeps : female
cone from Komgha district, show-
ing warty female cone scales: Photo:
G. G. Smith.
10. Encephalartos lehmannii Lehm., Pugill. 6 : 14 (1834).
Fig. 48, 49, 50, 54 on right.
Plant unbranched or more frequently branched from the base with stems 1-2 m tall,
rarely more, 25-45 cm diam. Leaves 1-1 • 5 m long, stiff and nearly straight, occasionally
somewhat recurved in the upper half, with bluish or silvery bloom persisting for con-
siderable period, the base of the stalk swollen and with a red-brown collar; leaflets
spaced and reduced in size towards the base but to only one prickle, close in the upper
half but rarely much overlapping, in V disposition; median leaflets 12-18 cm long
1 -7-1 -9 cm broad, pungent, entire or occasionally with one or two small teeth on the
lower margin. Cones solitary blackish-red over green because of short scurfy hairs on
the face of the scales: male cone subcylindric, narrowed to apex and base, 25-35 cm long
and 8-10 cm diam., the median scales with the face projected into a beak about 1 - 5 cm
long: female cone somewhat egg-shaped, 45-50 cm long and about 24 cm diam., the
median scales with the face protruding about 2 cm and with little surface irregularity
except for central upper ridge and terminal facet. Seeds red, 4-7-5 cm long and about
2 cm diam.
Encephalartos lehmannii occurs mainly on sandstone hills and mountains in the
Cape from near Willowmore and Steytlerville and extending eastwards to Uitenhage and
Pearston.
461
Fig. 49. — Encephalartos lehmannii'. Pearston district, in the Cape, infrequent on hills. Photo: L. E.
Codd.
462
Fig. 50. — Encephalartos lehnmmii : Pearston district, the female cone scales with angeled relatively
smooth face.
Most previous workers on the genus have taken a broad view of this species and
various relationships have been suggested between it and E. honidus and the recently
described species E. princeps and E. trispinosus. The differences in the face of the female
cone scales was the guiding principle in distinguishing between these species: that of
E. lehmannii being nearly smooth, of E. hoiridus ridged, of E. trispinosus wrinkled, of
E. princeps warty. The spinous lobes on the lower margin of some, if not all, median
leaflets of E. horridus and E. trispinosus also readily distinguish them from the other two,
while E. lehmannii is distinguished from E. princeps by its larger pulvinus and leaf-scars,
shorter trunks and more spaced leaflets. On occasions E. lehmannii and E. princeps
show some toothing of the lower margin of median leaflets but comparable variation is
found in all related species.
The distribution areas of the four species mentioned above do not appear to overlap.
Both E. horridus and E. lehmannii are found in the Uitenhage district, but the former is
mainly in the middle undulating part of the Zwartskops River Valley, whereas the latter
463
is in the mountainous area of the Klein Winterberg. For the rest the distribution range
of E. lehmannii will- be found in the drainage areas of the Groot and Sundays River
Valleys. E. trispinosus is found further east in the Bushmans, Great Fish River and
intervening river valleys, while E. princeps is even further east in the catchment area of
the Great Kei River.
The inter-relationship of these species is further discussed under E. trispinosus where
hybridization seems to have played a prominent part in the production of aberrant forms
11. Encephalartos trispinosus {Hook.) R. A. Dyer in Journ. S. Afr. Bot. 31 : 112
(1965). Encephalartos horridus var. trispinosus Hook, in Bot. Mag. t. 5371 (1863).
Fig. 13, 51, 52, 53, 54.
Plant unbranched or more often branched from the base with stems up to about 1 m
tall and 25-30 cm diam. Leaves O-l-l -25 m long, spreading-recurved and often twisted
towards apex with a bluish bloom when young and without hairs: leaflets reduced in
size and simple towards base, occasionally overlapping in the upper half ; median leaflets
often slightly curved to one side, 10-18 cm long, 1 •5-2-5 cm broad, pungent, with 1-2
pungent sometimes twisted lobes 1-3 cm long from the lower margin. Cones solitary,
bluish- or yellowish-green: male cone subcylindric, narrowed to both ends, 25-35 cm
long, 6 -5-7 -5 cm diam.: median scales with the face projected into a beak about 7 mm
long: female cone 45-50 cm long and 16-18 cm diam.; median scales with the face
protruding 2-5-3 cm, wrinkled and pimpled, rarely nearly smooth. Seeds red to yellow-
ish-orange, about 5 cm long and 2 cm diam.
Fig. 51. — Encephalartos trispinosus'. Kap River Valley, Bathurst district, eastern Cape.
6821035-5
464
Fig. 52. — Encephalartos trispinosus: near Plutos Vale, Albany district, eastern Cape.
Encephalartos trispinosus is located in the eastern Cape in the lower reaches of the
Bushmans, Great Fish and intervening river valleys in karroid scrub. There is appreci-
able variation between specimens in the lobing and width of leaflets.
Specimens classified under this species were first regarded by Hooker at Kew as a
variety of E. horridus. Later authors, for example Hutchinson, Rattray and Henderson
included the concept under E. lehmannii. Of the two ideas it is felt that the relationship
is decidedly stronger with E. horridus than E. lehmannii.
The decision to recognize E. trispinosus as a distinct species from these two others is
discussed in the Journal of S. Afr. Botany, 1965, and the decision was taken only after a
considerable amount of field work. Field work showed that in the lower reaches of the
Great Fish and Bushmans River Valleys there are isolated populations which are suffi-
ciently distinct and numerous to justify specific status.
Some of the specimens show a strong likeness to E. horridus in the lobing of the
median leaflets, but the leaflets from plants of comparable age of the two species show
that those of E. trispinosus are generally narrower and the twisting of the lobes on the
lower margin is rarely as uniform. The face of the female cone scale of E. trispinosus is
much wrinkled and pimpled, blue-green or yellowish-green in colour, whereas that of
E. horridus has a relatively smooth surface with prominent ridges directed towards the
terminal facet, and is usually dark reddish-brown in colour before maturity. M. J.
Wells records that cones of E. horridus, in the Botanical Garden at Rhodes University,
465
Grahamstown, turned a dull orange-yellow beneath their greenish-grey exterior at
maturity but that they did not get anywhere near as deep or definite an orange colour as
an E. trispinosm from Kap River. In addition E. trispinosus has longer incurved side
lobes to the female scales.
Fig. Sl.—Encephalartos trispinosus: female cone with the face of the scales angled and conspicuously
wrinkled.
466
Fig. 54.— Male Cones: ]c{t, Encephalartos trispinosus; right E. lehmmnii; former from Albany district,
latter from Steytlerville district: Photo: G. G. Smith.
467
Encephalartos lehmannii, as seen here, is more robust than either E. trispinosus or
E. horridus, has, for the most part, straight or nearly straight leaves with a large pulvinus
or swollen base and the leaflets are only rarely toothed on the lower margin. The
female cone scales are relatively smooth on the face by comparison with those of E.
trispinosus.
The distribution range of neither E. horridus nor E. lehmannii overlaps that of E.
trispinosus. However, in parts of its distribution range E. trispinosus comes into direct
contact with E. altensteinii and in such localities there are plants which strongly suggest
that natural hybridization has taken place. Some specimens are probably 1st generation
hybrids, whereas others give the impression of 2nd generation segregates, because of
their near approach in appearance to one or other of the parents.
Other species which come into direct or indirect contact, that is within a matter of
one to several miles of E. trispinosus and E. altensteinii are E. arenarius, E. latifrons and
E. caffer. Of these only the dwarf E. cajfer seems uninvolved in hybridization. Some
specimens, including several seen in the Bathurst district, seem to have an involved
parentage which eludes a reasoned interpretation.
12. Encephalartos horridus (Jacq.) Lehm., Pugill. 6 : 14 (1834). Zamia horrida
Jacq., Fragm. 27 : tt. 27 & 28 (1801).
Fig. 11, 55, 56, 57.
Fig. 55. — Encephalartos horridus: dwarf female plant few miles west of Port Elizabeth, eastern Cape, on
quartzite ridge.
Plant unbranched or more usually branched from the base with stems subterranean
or exposed about 30 cm, exceptionally up to 1 m tall, with the crown sparsely woolly.
Leaves rigid 0-5-1 m long, recurved in the upper third, with an intense bluish or silvery
468
bloom ; leaflets reduced in size and simple towards base but not reduced to more than
one prickle at base; median leaflets up to about 10 cm long and 2-5 cm broad with
pungent apex and 1-3 pungent lobes up to 4 cm prominent on the lower margin and
twisted out of the plane of the leaflet. Cones solitary with sparce blackish- or reddish-
brown scurfy hairs over dark green surface: male cone subcylindric narrowed to both
ends, 20-40 cm tall, 6-12 cm diam.; face of median scales projected into a beak up to
1 cm long: female cone more or less egg-shaped, 20-40 cm long, 15-20 cm diam.,
median scales with the face protruding 1 ■ 5-2 cm, ridged towards the tip but not much
wrinkled or pimpled. Seeds pale red or carmine, 3-3 • 5 cm long and about 2 ■ 5 cm diam.
Encephalartos hoiridus occurs in the Port Elizabeth and Uitenhage districts of the
eastern Cape.
It has been generally accepted that the typical form of this species originated from
the Uitenhage district where it grows in association with karroid scrub in which succulent
plants abound. The specimens must have been fairly abundant locally at one time
judging by the number now in cultivation and plants may still be seen scattered in parts
along the roadside when travelling from Uitenhage to Steytlerville. In the majority of
specimens the major part of the stem is below ground level, but one exceptional record
was made by R. Story of a plant with a stem over one meter (4 ft) tall. This indicates a
close relationship with the concept classified under the name E. trispinosus, under which a
fuller discussion of the problem of interrelationship will be found.
Fig. 56. — Encephalartos hoiridiis: unusual form with trunk above ground few miles from Uitenhage,
eastern Cape, on shale formation.
469
Fig. 57. — Encephalartos horridus: female and male cones and leaf from dwarf form in Port Elizabeth
district as shown in Fig. 55.
In the Port Elizabeth district a few miles west of the city, the record of Encephalartos
horridus is on a quartzite hill with a sparse sour vegetation (Fig. 55 and 57). The
specimens are much-branched and dwarf by comparison with those in the Uitenhage
district, to such an extent that they seem to merit varietal rank. There appears to be no
important difference in the cone formation, except a smaller size. It has not been
possible to compare growth reactions under cultivation of the two forms, for which
reason the status quo is maintained.
Except for some specimens of E. trispinosus, E. horridus is readily distinguished
from other related species, by its spreading recurved leaves with an intense bluish bloom
on the inner whorls, by the deeply lobed lower margin of the median leaflets and the
470
twisting of the lobes. It is best distinguished from E. trispinosus by the cone characters,
which unfortunately are not always available and one may have to rely on general
characteristics and distribution records. If one knows that the specimen was collected
near Uitenhage or Port Elizabeth, it is a rule of thumb identification because no allied
species grows within its distribution area. It is difficult to appreciate why some authors
have confused E. horridus with E. latifrons, which has a stout trunk developing over 3 m
tall, broad dark green leaves, hairy at first and very different cones. The only excuse for
confusion is the similarity in the lobing of the lower margin of median leaflets.
13. Encephalartos arenarius R. A. Dyer in Journ. S. Afr. Bot. 22 : 1 (1956).
Fig. 58, 59.
Plants unbranched or more often branched from the base with stems up to about I m
tall and at this stage often somewhat reclinate, 20-30 cm diam. Leaves 1-1 -5 m long,
recurved in the upper half, without hair: leaflets light green when young with a faint
bloom, becoming darker with age and losing the bloom, overlapping in the upper half
and in V disposition, reduced in size towards base occasionally to one prickle; median
leaflets 12-16 cm long, 2-5-4 cm broad with the attachment 1-1 -5 cm broad, pungent
with upper margin entire or occasionally with a single tooth, and lower margin usually
with 3, rarely 4 flat or slightly twisted pungent lobes. Cones single: male cones with
median scales projected into a beak about 1 cm long: female cone mainly barrel-shaped
with rounded base and apex, 35-45 cm long and 20 cm diam. green with a slight bloom;
median scales with the face projected about 2 cm, wrinkled or pimply. Seeds coral-red,
up to about 5 cm long, 2-2-5 cm broad.
Fig. 58. — Encephalartos arenarius'. plant of medium size, few with trunks up to 4 ft tall, from Alexandria
district, eastern Cape.
471
Fig. 59. — Encephalartos arenarius: female plant with single cone, produced contemporary with whor
of leaves, probably unusual, Alexandria district.
Encephalartos arenarius is local in the Alexandria district of the eastern Cape^
associated with coastal sand-dune scrub bush; at one time it was locally plentiful, but its
numbers have been reduced to make way for farming activities.
This species could well be confused with Encephalartos latifrons, yet it may readily
be distinguished by its smaller stature its less rigid, lighter green foliage with a thin
bloom, and, in the wild state, by its distribution pattern in sand-dune bush, as opposed
to bush groups on rocky outcrops.
It is reasonably certain that natural hybridization has taken place on the banks of
the Bushmans River between Encephalartos altensteinii and E. trispinosus. There are
also grounds for a suspicion that E. arenarius may also be involved in certain cases of
doubtful parentage. The elucidation of the taxonomic problems which present them-
selves in the identification of some specimens of Encephalartos in the lower reaches of the
Bushmans and adjacent river valleys of the Alexandria and Bathurst districts, would be a
long-term undertaking. At this stage it is not possible to be anything but vague on
many issues.
14. Encephalartos latifrons Lehm. in Tydschr. Nat. Gesch. 4 : 424, t. 9 A & B
(1837-38).
Fig. 15, 16, 60, 61.
472
Fig. 60. — Encephalarlos lalifrons: a large male plant about 9 ft in a bush group at Trappes Valley in the
Bathurst district, eastern Cape.
473
Fig. 61. — Encephalartos latifrons: female cone on right hand plant in garden of the Old Mill, Trappes
Valley, see Fig. 15 and 16.
Plants unbranched or more often branched from the base with stems 2-5-3 m tall.
Leaves 1-1-5 m long, recurved or curled in the upper half, hard and rigid, finely hairy
when young and gradually losing the hairs with age; leaflets overlapping in the upper
half of the leaf and in V disposition, reduced in size near base but only the lowest some-
times prickle-like; median leaflets 10-15 cm long, 4-6 cm broad with the attachment
474
1 • 5-2 cm broad, pungent with upper margin entire or sometimes toothed and with 2-4
deep triangular pungent, sometimes slightly twisted, lobes on the lower margin. Cones
1- 3, dark green or dark blue-green: male cones subcylindric, narrowed to both ends,
30-50 cm long, 8-17 cm diam.; median scales with the face projecting into a slightly
decurved beak about 2 cm long. Female cones barrel-shaped up to about 60 cm long and
25 cm diam. ; median scales with the face wrinkled and pimply and protruding 2-2 • 5 cm.
Seeds red, about 5 cm long and 2-2-5 cm diam.
Encephalartos latifrons is found in association with scrub-bush on rocky outcrops
in the Bathurst and Albany districts of the eastern Cape; it is now rare and shows little
sign of active regeneration from seed. Distribution records from outside this area are
suspect and require verification.
As mentioned earlier this is one of, if not the slowest growing arborescent species in
South Africa. It seems that two or more years elapse between the emergence of new
whorls of leaves, and cone production is irregular.
Some earlier workers seem to have experienced difficulty in distinguishing the robust
E. latifrons from the dwarf E. horridus, and as pointed out in the notes under the latter,
there is little justification for this. If, however, one introduces E. arenarius into the
discussion, one could visualize it as an intermediate species, which bridges the gap in
certain characters. It is certainly far more understandable that E. latifrons should have
been thought to include E. arenarius than that it should be confused with E. horridus.
Up to now, no direct overlap has been discovered in the distribution records of E.
latifrons, in the slightly inland broken scrub of the Albany and Bathurst districts, and of-
E. arenarius in the coastal dune bush of the adjacent Alexandria district.
Encephalartos latifrons has not been found in close association with E. altensteinii
in the Bathurst district, yet they are certainly not many miles apart in some places.
A specimen in a local farm garden strongly suggested that it had its origin in hybridi-
zation between these two species. Another, rather different plant, cultivated in the
Albany Museum grounds in Grahamstown, suggests the same possiblility.
15. Encephalartos longifolius (Jacq.) Lehm., Pugill. 6 : 14 (1834). Zamia longifolia
Jacq., Fragm. Bot. 1 : 28, t. 29 (1801).
Fig. 6, 62, 63, 64, 97.
Plants unbranched or branched from the base with stems up to about 3 m tall,
rarely 4-5 m, 30-45 cm diam. Leaves 1-1-75 m long sometimes nearly straight but
usually spreading and recurved towards the apex or occasionally arcuate, finely hairy
for some time but most hairs eventually falling; leaflets usually dark green, sometimes
with a bluish lustre or bloom, overlapping in the upper half and reduced in size towards
base but not to a series of prickles, at most 2 prickles; median leaflets up to about 20
cm long and 4 cm broad, pungent or blunt, entire for the most part or with the lower
margin with 1-3 short teeth. Cones 1 or 2 together, greenish-brown with reddish
adpressed hairs: male cones 40-60 cm long and 14-20 cm broad, narrowed to the ends;
median scales with the beak projecting about 2 cm: female cones more or less egg-
shaped, up to about 60 cm long and 40 cm diam. ; median scales with the face projecting
2- 2-5 cm, coarsely wrinkled on the surface. Seeds red, about 5 cm long and 2-5 cm
diam.
Encephalartos longifolius makes an imposing sight on the sourveld mountains from
the Uniondale to Albany district, and is sometimes locally frequent.
475
Fig. 62. — Encephalartos longifolius". in Van Stadens River Valley, eastern Cape; a robust specimen,
with Frank Stayner.
476
Fig. 63. Encephalartos longifoUus: female plant with single cone in Paardepoort, between Steytlerville
and Uitenhage, Cape.
Fig. 64. — Encephalartos longifoUus: a specimen repeatedly burned in fynbos fires in the Langkloof,
Uniondale district, Cape, showing unusual branching, with M. Wells.
477
This would be the first species of Encephalartos to be seen by early travellers ap-
proaching the eastern Cape via the Langkloof Valley. It appears first on the mountains
some miles west of Joubertina and from there it is seen at irregular intervals on the hills
of the Zuurberg approaching Grahamstown. It was the Swedish traveller Thunberg
who first collected this arborescent species and also a dwarf species on his way to the
eastern Cape with Masson, of Kew, in 1772. He mistook thern for forms of one species
and gave them the name E. coffer. The error was not discovered until much later,
when it was found necessary to use the name E. longifolius for the arborescent species,
while the dwarf retained the name E. coffer.
The stems of E. longifolius are often unbranched from the base and develop a
massive trunk. Occasionally specimens are seen to branch above ground, but this must
be regarded as abnormal and possibly caused by some injury (Fig. 64). The trunks
withstand fierce fynbos fires over a considerable period of time but there must be a
limit to their endurance of this hazard. One wonders whether the present western limit
of distribution was set by the influence of fire. Fynbos or macchia fires are very much
more intense than those in open grassveld. Grassveld becomes ecologically more
important as one goes eastwards from Port Elizabeth.
The stature of the plant, the metalic green colour and the scurfy hairs of the leaflets
are generally good guides to the identification of E. longifolius, but the degree of bloom
on the surface of the leaflets varies, sometimes to the extent of a silvery shine, and one
must be prepared for this. Normally there should be little difliculty in distinguishing
E. longifolius from its near ally E. oltensteinii, but specimens of the latter grown in full
sunlight could cause some uncertainty to arise. There can be no doubt about specimens
in the field, because the distribution ranges of the two do not overlap; the former keeps
to inland hills of Table Mountain Sandstone and the latter mainly to coastal bush and
on kranses, and for the most part is further east.
Although the relationsip to E. woodii in Natal must be remote, the leaf-canopy of
E. longifolius sometimes takes on the characteristic ‘ umbrella ’ shape of the former. In
such cases one must confirm the identification by noting the absence of prickles towards
the base of the leaves.
E. longifolius is associated on occasions, for example in Paardepoort on the Klein
Winterberg, with E. lehmonnii, but there is no close relationship between them, and no
question of interspecific hybridization has been raised so far.
16. Encephalartos altensteinii Lehm., Pugill. 6 : 11, t. 4 & 5 (1934).
Fig. 14, 65, 66, 67, 68.
Plants unbranched or branched from the base with trunks up to about 4 m tall,
rarely up to 7 m and then generally reclining, 25-35 cm diam. with a small amount of
wool on the crown. Leoves 1-5-3 m long, rarely 4 m long, nearly straight in exposed
situations and recurved in shade, hairy at first except the surface of the leaflets, hairs soon
falling; leaflets reduced in size at base but not to a series of prickles; median leaflets
up to about 15 cm long and 2-5 cm broad, fairly rigid in open situation, not usually
markedly tapering, with 1 or 2 or, less frequently, 3-5 teeth on upper and lower mergin,
sometimes on one margin and not on the other, mucronate. Cones 2-5 together, rarely
only one, yellowish-green: mole cones subcylindric, rounded or narrowed to apex, up to
about 50 cm long and 12 cm in diam., the median scales with the face extended into a
decurved beak about 1 • 5 cm prominent: femole cones more or less egg-shaped, 40-55 cm
long, 20-28 cm diam., with the median scales with the face wrinkled and lumpy and with
some hairs, projecting 2-5-3 cm. Seeds scarlet, 3-5—4 cm long, 2-2-5 cm diam.
478
Fig. 65. — Encephalartos altensteinii: in forest near East London, eastern Cape, trunk about 12 ft tall,
with Miss Courtenay-Latimer.
479
Fig. 66. — Horse-shoe Bend in Kowie River, Bathurst district, Cape, where Encephalartos altensteinii
occurs frequently in coastal forest, photo taken from overlooking krans.
Encephalartos altensteinii has its western limit of distribution in the coastal bush on
the west bank of the Bushmans River in the Alexandria district of the eastern Cape: it is
furthest from the sea in the King William’s Town district and extends eastward to the
southern border of Natal.
Specimens from the midlands of Natal previously classified as E. altensteinii will be
found classified under the name E. natalensis in this account. The absence of a series of
prickles at the base of the leaf stalk in the former and their presence in the latter, is a
fairly constant distinguishing feature between the two.
As in most species with prickles on the margins of the leaflets, those of E. altensteinii
vary considerably in number, size and position. The type figure of the species shows the
prickels to be conspicuous and fairly regular on both upper and lower margins. This
indicates that the type specimen was probably collected on the west bank of the Bush-
mans River towards the coast, where very similar forms have been seen in recent times,
Geographically too, this would be the most likely locality to be reached first by early
botanical collectors. Specimens of E. altensteinii on the east bank of the river have
fewer prickles, and, travelling eastwards, it is not long before some specimens are found
which are almost devoid of marginal prickles. One wonders how long the Cycads on
the Bushmans River banks, many reclining up to 25 ft long on the east bank, will survive
the encroachment of the beach cottage.
In the Bushmans, Great Fish, Buffalo and intervening river valleys, Encephalartos
altensteinii is found in association with several other species. In the case of E. trispinosus
and E. villosus, they are found side by side with E. altensteinii, while E. arenarius and
E. latifrons are not far distant from it in certain other areas. In such circumstances one
6821035=6
480
finds occasionally specimens which appear to combine the characteristics of the associated
species and these strongly suggest that natural hybridization has taken place. No other
species seems so much involved in natural hybridization as E. altensteinii and the sup-
posed hybrid forms have added greatly to the problem of classification.
Fig. 67. — Encephalartos altensteinii-. female cone from Bathurst district, near the area from which the
type specimen was probably collected, specimen from Dr. E. E. A. Gledhill.
48]
Fig. 68. — Encephalartos altensteinii: exposed in low scrub bush in the King William’s Town district;
main trunk about 15 ft tall, with G. G. Smith.
482
17. Encephalartos woodii Sander in Card. Chron. 1908, 257 (1908).
Fig. 69, 70, 71.
Plants (only male known) unbranched or branched from the base, sometimes
branched near the crown in cultivation, up to about 6 m tall, 40-60 cm diam., sometimes
expanding to 90 cm at base, narrowing to 30-40 cm above, with densely leafy umbrella-
like canopy and somewhat woolly crown. Leaves 1 •75-2-5 m long, gracefully curved
and spreading or broadly arched and bow-like, densely pale brown-woolly at first but
hair falling with age; leaflets dark green and leathery, gradually reduced to prickles at
the base, those from juvenile suckers variably 2-5-toothed in the lower half on one or
both margins, those on old stems often entire; median leaflets up to about 20 cm long
and 5 cm broad; the apex pungent or sometimes slightly hooked and blunt. Cones',
male 1-several, bright orange-yellow, subcylindric, 40-90 cm long, rarely up to 120 cm,
15-20 cm diam., median scales up to about 2-5-3 • 5 cm prominent.
There is only one locality record for this species near Ngoye in Zululand. It is now
probably extinct in the wild state.
Because of its unique character the story of its discovery and subsequent disappear-
ance from nature is summarized below.
The first record was made by Medley Wood in 1895 near Ngoye forest on a botanical
expedition by oxwagon. He noted a single male plant consisting of four large stems up
to 18 ft tall, with several off-shoots from the base. Some years later, probably in 1903,
Medley Wood sent John Wylie to collect living specimens for cultivation. He brought
back some of the smaller off-shoots, three of which were planted in the Durban Botanic
Garden and according to an unpublished report by Medley Wood other specimens were
sent to Kew and Messrs. Sander & Co. of St. Albans.
In Medley Woods’ Annual Report of 1906-7, which included a photograph of one of
the plants, he referred to it under the name E. altensteinii var. bispinosa. On 25th April,
1908, however, Sander gave a short account of his specimen in the Gardeners’ Chronicle
under the name E. woodii. This prompted Medley Wood to give further information in
the same journal of 2nd May, 1908. “ Having reached a spot where the country was
very rough, I stayed for several days botanising in the vicinity, and in so doing found a
solitary clump of Encephalartos . . . not another plant in the vicinity . . . but a number of
E. brachyphyllus (i.e. E. ngoyanus Verdoorn) ... we have in the garden seven species . . .
numerous specimens, but these three (E. woodii) as far as the foliage is concerned, are in
my opinion, not only the handsomest of all, but are strikingly different from any of the
others ”.
Wylie returned in 1907 to collect two of the larger trunks and found that one of the
original four had been mutilated by Bantu in the interim. A photograph of the parent
plant was taken in situ on that occasion and is preserved in the Natal Herbarium,
Durban (Fig. 69). The two trunks are still in good health on the lawn near the Old
Conservatory in the Durban Botanic Garden. These specimens prompted Medley
Wood to say in 1912 that “ the leaves are not at all like those of the other species known
to me, but are beautifully curved, and make the plant a very decorative one ”.
The next record of importance is the visit to the site by C. J. Chamberlain from
Chichago, about 1912, and reported in his booklet of 1919 under the title of The Living
Cycad. He refers to a single specimen more than 10 ft high about 20 miles from
483
Mtunzini in the midst of Stangeria and Encephalartos hnichyphyUus (i.e. = E. ngoyanus
Verdoorn) “ My Zulu guide, the son of the Zulu Chief, . . . had been well coached by
Mr. Wylie, otherwise there would have been little likelihood of finding such an isolated
specimen in a hilly country, with numerous stretches of forest and bush ” . . . “ They say
that it is the only Cycad, with a trunk, within a distance of fifty miles
Fig 69 —Encephalartos woodir. the original male plant found near Ngoye, Zululand; from a photo-
graph in the Natal Herbarium, Durban, taken on the expedition by Wylie in 1907.
484
Fig. 70. — Encephalartos woodii: plant established in Durban Botanic Garden, with Medley Wood, from
photo in Kew Bulletin 1914.
485
Fig. 71. — Encephalartos woodii: same plant as in previous figure with Miss H. Forbes, 1945, trunk
abnormally branched above, other trunks in the garden have not branched, base enlarged
probably due to compression through weight of canopy.
A further interesting record was revealed in 1964 by Mr. B. J. Huntley, who traced a
copy of a letter from Forester Prior, dated 22nd May, 1916. It reads:
“ Only two plants (presumably trunks from the old rootstock) of this species
{E. woodii) are known ”. “ I intend cutting both, one for him (Wylie) and one for
Pretoria ”. “ The vegetation has been burned in the vicinity during the last year
or two and the stems are clipped by natives, so if they are removed now they will
486
serve a better purpose to the botanical world than if they remained there A letter
dated 29th May, 1916 from the Acting District Forester states : “ Forester Prior has
now consigned the only known plant of this species in the Ngoye to the Chief:
Division of Botany, Pretoria ” . . . “ the weight of this plant is between 13 and 15
hundred pounds and is about 12 feet long and 18 inches in diameter
What may be the plant referred to above was traced in the Union Buildings rockery
and by kind permission of the Superintendent was replanted in the National Botanic
Garden of the Botanical Research Institute in 1964.
Huntley went on to say that in a three week’s stay he had not discovered a specimen
of E. woodii at Ngoye. He had, however, located the spot where the plants had been
removed, a steep, south facing slope on the margin of a forest outlier. This he had done
with the assistance of an old wizened Zulu of about 85 years, the only one who really
knew E. woodii. On some of the adjacent slopes specimens of E. ngoyanus were frequent
and were regarded by the old man as seelings of E. woodii. The other old Zulus, whom
Huntly consulted, confused it with Cyathia dregei, Stangeria eriopus and Encephalartos
ngoyanus. He concluded by saying that he doubted very much if the species was still
to be found there, an opinion with which I reluctantly agree.
18. Encephalartos natalensis Dyer & Verdoorn in Bothalia 6 : 205 (1951).
Fig. 72, 73, 74, 75.
Plants unbranched or more often branched from the base with stems 3-4 m tall and
occasionally up to 6-5 m, 30-40 cm diam., sometimes with wool on crown. Leaves
1 • 3-3 • 2 m long, nearly straight or somewhat recurved and slightly twisted towards
apex; leaflets reduced in size towards the base to several prickles on either side of the
stalk, dark green; median leaflets 16-23 cm long and 2- 5-4- 5 cm broad, pungent, entire
or with 1-5 sharp prickles on one or both margins, more often on the lower leaflets and
sometimes almost lobate on young trunks. Cones 2-3 together, dark yellowish-green
with thin covering of foxy wool: male cones up to about 45 cm long and 9-10 cm diam. ;
median scales extended into a beak about 1 • 5 cm long : female cones somewhat egg-
shaped, up to 50 cm long and 25 cm diam. ; median scales with the face wrinkled into
blunt papillae and projecting about 2 cm. Seeds scarlet, about 5 cm long and 2 cm diam.
Encephalartos natalensis is to be seen dn inland kranses and in kloofs from the
eastern Cape-southern Natal border to northern Natal in the upper catchment area of
the Umfolozi River.
This species was segregated from the broad concept held by some earlier workers of
Encephalartos altensteinii Lehm. In typical E. altensteinii the basal leaflets are not
reduced to a series of prickles as they are in E. natalensis. When E. natalensis was first
described it was suggested that there was a wide break between its distribution areas and
that of E. altensteinii. Since then the gap has been very much bridged by records of
E. altensteinii from the eastern Cape border and E. natalensis is at no great distance
from it. There is still a difference though, in that E. altensteinii is frequent in coastal
bush and is at its furthest from the coast in the district of King Williams Town, whereas
E. natalensis is restricted to inland areas well out of the coastal vegetation association.
487
Fig. 72. — Encephalartos natalensis: type female plant in Monteseel Reserve near Inchanga, Natal, with
Dr. Herd.
488
E. mtalensis used to be called the giant Natal Cycad even as early as 1867 when a
particularly large specimen from Kranskloof was painted by J. Sanderson. Charles
Millar, writing at the time to Hooker at Kew said that the huge Encephalartos was found
in a secluded kloof about 30 miles from the sea. It had a girth of 9 ft, was 16 ft tall
before branching and was 25 ft to the crown formed of 5 branches. Millar added that
one trunk filling a large wagon, was transferred to the Durban Botanic Garden. It is not
known whether or not this is the same specimen which Medley Wood commented on in
a note he sent to the Gardeners’ Chronile in 1908, about a plant in the Durban Botanic
Garden: “ This plant is a male and this year has no less than 18 cones, arranged in more
or fewer complete concentric circles. I think that this is a record number; the tree has
been partly buried in the side of an embankment, and it is intended to remove it to a
position where its full length can be seen to better advantage ”. There is a large male
specimen near the steps leading to the old Conservatory today, which seems to answer
this description.
Further information about E. nalalensis is included in the notes on the rate of
growth of trunks. The appearance of the old female type-plant in question has altered
considerably in recent times. The tallest trunk was snapped in two places by a tornado
which devastated the area in 1961. Several of the younger trunks have since madr
appreciable growth but the tallest one in 1963 was then only 1 1 ft 7 ins in height.
Fig. 73. — Encephahirtos nalalensis: cones on a branch of the type specimen shown in previous figure.
489
Fig. 74. — Encephalartos natalensis: habitat view at Howick Falls, near Pietermaritzburg, Natal.
490
Fig. 75. — Eiicephalartos iiatalensis: habitat view on Inhlazatje Mt. south of Vryheid, Natal.
491
19. Encephalartos lebomboensis Verdoorn in Flow. PI. Afr. 27 : P!. 1078, 1079
(1949).
Fig. 5, 10, 76, 77, 78.
Fig. 76. — Encephalartos lebomboensis: on the mountains in the vicinity of Stegi, Swaziland, with J.
Erens.
492
Fig. 77. — Encephalartos leboniboensis: female cone from Paulpietersburg district. Natal, showing
relatively flat unwrinkled face of the scales.
Plant unbranched or sometimes branched from the base with stems occasionally up
to about 4 m tall and 30 cm diam., somewhat woolly on crown. Leaves in a dense
crown, 1-2 m long, fairly straight or somewhat recurved towards apex, hairy at first but
hairs soon falling; leaflets reduced in size towards base and ending in several prickles
on either side of the leaf stalk ; median leaflets 1 2-1 7 cm long and 1 • 2-2 • 2 cm broad, with
493
1-4 prickles on both margins but more regularly on the lower margin and rarely without
prickles. Cones 1-3 together, apricot-yellow, or dull salmon-pink; male cones sub-
cylindric, up to about 45 cm long and 13 cm diam., narrowed gradually to both ends,
median scales projected into a beak 1-1 • 5 cm long : female cones more or less egg-shaped,
up to about 45 cm long and 22 cm diam., median scales fairly smooth and the face
protruding only about 1 cm, somewhat hairy at first. Seeds scarlet, 4 cm long and
1 •8-2-2 cm diam.
Encephalartos lebomboensis features prominently on cliffs and in kloofs of the
Lebombo Mountain range in Swaziland, southern Transvaal and northern Natal and
also extends into Portuguese East Africa.
There is a fairly close relationship between this and E. natalensis which occurs
further south. There does not seem to be an overlap in distribution, E. lebomboensis
not occurring south of the Pongola River catchment and E. natalensis not further north
than the Umfolozi River catchment. The face of the female cone scales of E. lebom-
boensis is relatively smooth and flattish compared with the more prominent wrinkled one
of E. natalensis. In addition the leaves of E. lebomboensis are straighter, the base of
the leaf stalk more densely prickled and the leaflets narrower.
Fig. li. —Encephalartos lebomboensis-. rescue by ‘ Operation Wild Flower ’ at the site of the Jozini Dam
across Pongola Poort. Photo; J. Hall.
494
The distribution of E. lebomboensis overlaps in places, such as near Pongola Poort
in southern Transvaal, that of E. villosus, and the appearance of certain young plants
in the area suggests that natural hybridization between the two species has taken place.
E. ngoyanus also occurs at the Jozini Dam site.
Several thousands specimens of E. lebomboensis doomed to be swamped by the
waters of the Jozini Dam across the Pongola Poort were removed beforehand, largely by
“ Operation Wild Flower ” for distribution to public gardens and interested growers.
(Fig. 78). In order to remove specimens there was no option but to float or take
them by raft from the north to the south bank and this photograph shows how some
of the larger trunks were handled. The aims of Operation Wild Flower are most praise-
worthy and the organization has met a real need.
It is unusual for plants of Encephalartos in cultivation to set fertile seed but a
striking example was observed in the garden of Capt. D. R. Keith near Stegi, Swaziland
in 1947 (Fig. 10). He had a short avenue of Encephalartos including both male and
female specimens of E. lebomboensis, which occurs naturally in the neighbourhood.
Female cones had developed and fallen to pieces the previous year and at the time of our
visit one parent plant was surrounded in the short grass cover by a dense stand of seed-
lings. A quantity of them was collected and a group of young plants is now established
in the Pretoria National Botanic Garden of the Botanical Research Institute. In
nearly 20 years the stems under dry conditions have not developed above ground.
20. Encephalartos transvenosus Stapf & Burlt Davy in Burtt Davy FI. Transvaal &
Sw. 1 : 40 & 99 Fig. 4b (1926).
Fig. Frontispiece, 79, 80, 81, 82.
Plants unbranched or occasionally branched, reaching a height of about 13 m
(40 ft) in exceptional cases but many are between 5 and 8 m and 40-65 cm diam., with
brown woolly hairs on the crown. Leaves 1-2-5 m long, fairly straight or slightly
recurving towards the tip, hairy at first but the hairs soon falling: leaflets reduced to
several prickles on either side at base; median leaflets spreading more or less at right
angles from the stalk and then recurving gradually and sometimes slightly sickle-curved,
10-20 cm long, 2-3-5 cm broad, generally toothed on both upper and lower margin
towards base. Cones 2-4 together: male cones subcylindric, 30-40 cm, 13-15 cm diam.,
rounded at apex; median scales with the face projecting into a beak 1-1-5 cm long:
female cone more or less egg-shaped, 50-80 cm long and 20-30 cm diam., the median
scales with the face somewhat wrinkled or nearly smooth and projecting about 2-5 cm.
Seeds red, 4-5-5 cm long and about 2-5 cm diam.
Encephalartos transvenosus, the giant Modjadji Cycad, is prominent in the mist belt
on the mountains of the Soutpansberg and Letaba districts of the Transvaal.
The Modjadji Cycad, sometimes referred to as Modjadji Palm, which enjoys royal
protection in the territory of the Rain Queen near Duiwelskloof, is certainly the most
majestic of all species in southern Africa and at its full height of about 40 ft has few rivals
in any other part of the world. In addition to its size, a distinguishing feature of the
species is the recurving of the leaflets from the leaf-stalk. The leaflets are somewhat
broader in proportion to their length than those of its relative E. paucidentatus with
which it might be confused in its middle height. The leaflets do not have the same
prominent venation on their under surface as do those of E. paucidentatus. The female
cones weigh up to about 75 lb.
495
Fig. 79. — Encephalartos transvenosus: the Modjadji Cycad, about 30 ft tall, with younger ones below,
in Modjadji Location, Letaba, Transvaal, and under the protection of the Rain Queen, with
Prof. R. H. Compton.
6821035-7.
4%
Another feature of E. tramvenosus at Modjadji, and unrivalled in southern Africa,
is its presence in forest formation. It is to be numbered in its thousands, on the hills
near the royal kraal. For how long protection has been enforced by the Bantu Queens
it is not possible to say, but the decree must have had some bearing on the present
density of the local Cycad population. Damaged trunks are the exception. In other
areas and in the case of most other species, the trunks frequently show signs of mutilation,
often the work of the Bantu medicine men. E. woodii, now probably extinct in nature,
certainly suffered at their hands.
Fig. 80.— Eiicephalartos lransveno.ms\ female cones weighing up to 75 lb on maturity.
497
F[G 82. — The Modjadji Rain Queen in 1947, with her lady in waiting, advisors and son seated on the
right.
Fig. 81. — Encephalartos transvenosiis: general view of remarkable forest of Cycads in protected area at
Modjadji Location.
498
21. Encephalartos paucidentatus Stapf & Burtt Davy in Burtt Davy FI. Transvaal
& Sw. 1 : 40 & 99 Fig. 4a (1926).
Fig. 83, 84.
Plants unbranched or occasionally branched from the base with stems up to about
6 m tall and 40-70 cm diam. and brown woolly on crown. Leaves 1-2 m long, fairly
straight or slightly recurved, hairy at first and gradually losing the hairs; leaflets gra-
dually reduced in size towards the base, becoming lobate and the lowest prickle-like;
median ones straight or somewhat curved, 15-25 cm long and 2-3-2 cm broad towards
the base, with up to 30 conspicuous raised veins on the lower surface where hairs persist
longer than elsewhere. Cones probably 2-3 together: male cones up to about 60 cm
long and 15 cm broad with the scales projecting into a decurved beak 1 • 5-2 cm long with
toothed or crenate margin: female cones large but exact measurements of a complete
one are not recorded ; the individual scales are 5-6 cm broad and about 3 • 5 cm thick
vertically ; the face is wrinkled and the beak protrudes about 3 cm. Seeds red, up to
about 4 cm long and 2-5 cm diam.
Encephalartos paucidentatus is found rather sparsely in low forest and mountain
bush in the eastern Transvaal near Barberton and over the border of Swaziland.
One may claim with good reason that this is one of the most handsome species for
cultivation, because of its graceful widely spreading leaves with long slender, often
slightly recurved, 'eaflets. The numerous (up to about 30) conspicuous longitudinal
raised \eins on the lower surface of the leaflets are distinctive among related species,
such as the giant Modjadji Cycad, E. transvenosus, from near Duiwelskloof and the
Fig. 83. — Encephalartos paucidentatus'. in mountain valleys near Barberton, Transvaal, with typical
whorl of leaves.
499
Fig. 84. — Encephalartos paucidentatus: on mountain near Barberton; male plant with trunk up to
about 12 ft tall and three or more specimens in the background.
Soutpansberg. It is comparatively rare, unfortunately, but in co-operation with the
Department of Nature Conservation of the Transvaal wise inhabitants of Barberton
have established the Ida Doyer Nature Reserve where fine specimens are included.
There has been some confusion about the distribution of E. paucidentatus in nature
because it was first recorded as having originated from the farm Breslau, north-west of
the Soutpansberg. All efforts to substantiate the record failed and there is little doubt
that the species is, in fact, restricted to the area mentioned above.
22. Encephalartos ferox Bertol. f. in Mem. Accad. Sci. Bologn. 3 : 264 (1851)
E. kosiensis Hutch, in Kew Bull. 1932 : 512.
Fig. 85, 86, 87.
Plants unbranched or occasionally branched from the base, usually less than 1 m
tall but occasionally up to about 2 m and 30-35 cm diam. Leaves 1-2 m long, nearly
straight and gradually spreading, hairy at first but soon without hairs; leaflets gradually
reduced to a few prickles near the base; median ones broadest above the middle, up to
about 15 cm long and 3-5 cm broad, rarely up to 5 cm broad, with 2-4 small teeth on
the upper and lower margins and terminating in 3-5 pungent lobes. Cones 1-3 together,
salmon- or shrimp-pink to scarlet : male cones subcylindric, narrowed to both ends, up to
about 40 cm long and 7-10 cm diam.; the median scales with a decurved beak about
500
1 cm long: female cones egg-shaped, 25-50 cm long 20-40 cm diam; the median scales
with the face somewhat wrinkled and 2-2-5 cm prominent. Seeds red, 4-5-5 cm long,
1 - 5-2 cm diam.
Fig. 85. — Eiicepha/arlos ferox: young male and female plants in Durban Botanic Garden, Natal.
Fig. 86. — Encephalartos ferox: vigorous young specimen in garden of Mr, Harry Butcher in Durban,
Natal; E. hiimilis in left foreground, 1947.
501
Fig. il —Encephalartos ferox: female cone from Xai Xai (Chai Chai) in Portuguese East Africa.
502
Encephalartos ferox occurs in coastal scrub bush and grassveld from about 400
miles north of Lourenco Marques to Sordwana Bay in Zululand.
After this species was first discovered in Zululand at Kosi Bay by Aitken and Gale,
the new name Encephalartos kosiensis was given to it and it was only some years later
that it was proved to be the same as the plant named earlier as E. ferox from the coastal
belt of Portuguese East Africa. This habitat along the eastern coast is far removed
from that of any other species in southern Africa. Its furthest inland record in southern
Africa is about 25 miles from Sordwana Bay on the road to Pongola. Here on the
Makatini Flats a single specimen was reported in 1964 by a Natal Parks Board ranger,
Mr. I. Steytler. But this does not appear to be within the range of natural distribution.
E. ferox is distinctive in the shape of the broad leaflets which are broadest above the
middle and terminate in 3-5 spinescent lobes. The salmon-pink or red colour of the
cones is also peculiar to it in southern Africa. It does not appear to be closely allied to
any local species. Its numbers have been diminished by the encroachment of habitation
on the Mozambique coast and by afforestation in Zululand.
As mentioned in the chapter on propogation, a record was kept of seed which germi-
nated in Durban in 1937 and it was noted that one of the plants, with the stem still
mostly subterranean, produced its first cone after twelve years’ growth.
23. Encephalartos villosus Lem., 111. Hort. 14, Misc. 79 (1867); 15 : t. 557 (1868).
Fig. 88, 89,
Plants unbranched or occasionally branched possibly due to injury, with stems
subterranean or sometimes with the crown slightly exposed, up to about 30 cm long and
20 cm diam. Leaves 1 •25-2- 5 m long, suberect to gracefully curved-spreading, with the
stalk white-woolly at first; leaflets glossy green, widely spreading to recurved, reduced
in size towards the base into a series of prickles; median leaflets linear to linear-lanceo-
late, acuminate, 15-25 cm long, rarely up to 30, 1-5-2 -5 cm broad, usually with both
upper and lower margins with 1-3 forward-directed sharp prickles and the apex with a
single point, or occasionally with 2 subequal prickles. Cones 1-several together,
greenish-yellow to orange- yellow : male cone up to about 65 cm tall, tapered to the apex
and 6-5-12 cm diam. above the base (varying considerably in size and partly dependent
on the number of cones from the head) ; median scales with face very little projecting
and irregularly toothed or rough on lower margin: female cone subcylindric, 30-40
cm long and 12-13 cm diam.; median scales with face flattened and somewhat over-
lapping the scale below. Seeds scarlet, 2-5-3 - 3 cm long, 1 - 8-2 cm diam.
Encephalartos villosus is a relatively common and widely spread species from the
vicinity of East London into the coastal parts of Natal, the south-eastern tip of the
Transvaal and into Swaziland: it is usually associated with low forest in temperate
situations. There is considerable variation in size of plants from different areas between
the eastern Cape and Swaziland and those from the latter area are sometimes so much
more robust than others that it has been suggested that they belong to a distinct variety.
This view is not supported here. The subterranean habit of the stem added to the
reduction of the lower leaflets to a series of prickles is sufficient to distinguish it from
most related species in southern Africa. It has its nearest relative in E. umbeluziensis,
which also has a subterranean stem but lacks the series of prickles on the base of the
leaf stalk.
503
Fig.
-EncephalartosviHosus: near Shongweni Waterfall north-west of Durban, Natal- thismavbe
blle’^S'the^eS species; the prickles can be seen extending ’almost to^he
6821035-8
504
It was mentioned in an earlier chapter that E. villosus was collected and exported in
large numbers from Natal in the last century and in recent times many plants have been
removed from the wild state near East London in the preparation of land for pine-apple
research. Yet, because the species regenerates well naturally, no fear need be felt that
its survival in nature is in any danger.
Fig. 89. — Encephalartos villosus: fleft, female cone from near Shongweni, north-west of Durban; right,
extreme forms of the male cone on plants in cultivation at the Botanical Research Institute,
Pretoria.
505
Near East London E. villosus sometimes grows in close proximity to E. altensteinii.
In the south-eastern tip of the Transvaal near Pongola Poort it is found in association
with E. lebomboensis. In both areas specimens with intermediate characters between
E. villosus and the second species have been found, strongly suggesting that natural
hybridization has taken place. Further examples of suspected natural hybridization
within the genus are mentioned under E. altensteinii, E. trispinosus and elsewhere.
24. Encephalartos umbduziensis R. A. Dyer in Flow. PL Afr. 28 : PL 1100 (1951).
Fig. 90, 91.
Plants unbranched or occasionally branched possibly due to injury; stem sub-
terranean or sometimes with crown slightly exposed, up to about 30 cm long and 20-25
cm diam. Leaves 1-2 m long, suberect in open situations and only recurving in shade,
woolly at first except on upper surface of leaflets, hairs falling with age; leaflets, reduced
in size towards base but not to more than 1 or 2 prickles; median leaflets 10-20 cm long,
rarely up to 30 cm, 8-15 mm broad tapering gradually to the pungent apex, and margins
usually with 1-3 prickles. Cones 1-3 together, sulphur-yellow on maturity: male cones
subcylindric up to about 30 cm long, 6-8 cm diam. tapering slightly to apex, median
scales flattened and only slightly projecting: female cones more or less cylindrical, up to
about 30 cm long and 12 cm diam. ; median scales with face flattened slightly, ridged and
not much projecting. Seeds scarlet, about 3 • 5 cm long and 2 cm thick.
Encephalartos umbeluziensis appears to be restricted in distribution to dry forest
without much undergrowth in southern Portuguese East Africa, and extends into
Swaziland up the Umbeluzi River Valley: it may be common locally.
This species is distinguished from Encephalartos villosus, its nearest relative, by
the absence of a series of prickles on the base of the leaf-stalk. Sometimes there may
be one or two prickles but not more. The leaf-stalk or rhachis is usually fairly straight,
round and smooth and the dark green upper leaflets are set like a half open Venetian
blind. The cones are generally smaller than those of Encephalartos villosus and the
face of the scales also differs somewhat. Anatomical differences also have been observed
by Prof. Dr. P. Greguss of the Botanical Institute, Szeged, Hungary. Both species
occur in Swaziland but under different habitat conditions. E. villosus favours relatively
high, cool and moist areas as opposed to the hot drier habitat of E. umbeluziensis in
the Umbeluzi River Valley. And in spite of this relatively near approach of the two
species, no actual overlapping of the distribution areas has been recorded. In addition
no specimen with intermediate characters between the two species has been reported
from either distribution area.
The evidence taken as a whole, is sufficiently strong to refute the suggestion that
E. umbeluziensis should be regarded as a variety of E. villosus.
506
Fig. 90. — Encephalartos umbehniensis: top, seedling and young plants from Swaziland; bottom left,
female plant in garden of Botanical Research Institute, Pretoria; right, closer view of female
cone.
507
Fig. 91. — Encephalartos umbeluziensis: male plant in cone; leaf rhachis smooth and round; leaflets not
reduced to a series of prickles towards base as in E. villosus.
508
25. Encephalartos ngoyanus Verdoorn in Flow. PI. Afr. 27: Plates 1053, 1054 (1949).
Fig. 4, 92.
Fig. 92. — Encephalartos ngoyanus: left, type female plant of the species near Ngoye Forest in the
Umtunzini district of Zululand; right, male cone from plant collected in the type locality.
509
Plant dwarf, unbranched or possibly branched occasionally due to injury, stem
subterranean or with crown slightly exposed, up to 30 cm long and 20 cm diam. Leaves
0-5-1 -25 m long, erect-spreading or slightly recurved in the upper half, with silky and
woolly hairs which fall with age; leaflets reduced in size towards the base but rarely to a
prickle; median leaflets 7-8 cm long rarely up to 12 cm, 9-1 1 mm broad, rarely 1 • 5 cm,
tapering gradually from near the base to the pungent apex, usually with 1-3 teeth on
lower margin, rarely entire. Cones solitary, yellow at maturity: male cones subcylindric,
20-25 cm long, 4-5-6 cm broad, narrowed to apex; the median scales with the face
projected into a beak about 7-8 mm long; female cone more or less egg-shaped, about
23 cm long, 10 cm diam.; median scales with the face flattened and the lower margin
sometimes slightly overlapping the scale below. Seeds scarlet, 2 - 7-3 cm long, 2 cm diam.
Encephalartos ngoyanus is sometimes common in grassveld and on forest margins,
often near boulders, in Zululand and possibly extends over the southern border of
Swaziland and Transvaal at Pongola Poort.
This species seems distinguishable from Encephalartos cajfer in a number of incon-
spicuous ways. The root-system seems less tuberous, the stems less consistently sub-
terranean, the leaflets less crowded, not twisted and with the lower margin often toothed
and the female cone scales with a less pronounced terminal facet. While each of these
characters shows variability the combination is considered to justify specific separation.
The break in distribution between Encephalartos cajfer in the eastern Cape and
Encephalartos ngoyanus in Zululand is also noteworthy but not necessarily significant on
a species level.
26. Encephalartos caffer (Thimb.) Lehm., Pugill. 6 : 14 (1834). Zamia cajfra Thunb.
in Nov. Act. Soc. Sc. Uppsal. 2 : 285 (1775).
Fig. 4, 93, 94, 95, 96.
Fig. 91 —Encephalartos caffer: plant at Trappes Valley in the Bathurst district, Cape.
510
Plant dwarf, unbranched or occasionally branched possibly due to injury, with
tuberous roots and subterranean stem up to 30 cm long and 15-25 cm diam., woolly at
crown. Leaves 40-90 cm long, erect-spreading or curved or twisted, brown-woolly at
first, hair soon falling; leaflets crowded, sometimes irregularly twisted from the axis,
reduced in size towards base, sometimes to 1 or 2 prickles ; median leaflets up to about
10 cm long and 8-10 mm broad, gradually narrowed from near base, mostly entire on
mature plants. Cones solitary, greenish-yellow: male cone subcylindric, narrowed to
apex, 20-30 cm long, 6-11 cm diam.; median scales with the face slightly projected into
a beak 5-6 mm long, with the lower margin sometimes toothed: female cone sub-
cylindric, up to about 30 cm long and 15 cm broad, slightly narrowed to the rounded
apex; median scales flattened on face and not much projecting. Seeds scarlet, up to
3 • 8 cm long and 2 • 3 cm diam.
Encephalartos caffer is found only infrequently in sour grassveld of the eastern Cape
coastal belt from Uitenhage district to the neighbourhood of Willowvale in the Transkei.
At Kirstenbosch there is a record from the Steytlerville district but the environmental
conditions in that area are such as to cast a doubt as to its accuracy.
The species has not been rediscovered in the Uitenhage district for many years and
the record of its presence in the Steytlerville district could have been caused by plants
being received at Kirstenbosch from that postal address. One would need the co-
operation of all farmers and their herdsmen in the eastern Cape coastal districts in order
to establish the true position as regards the distribution of this dwarf species. It is
very readily overlooked in grassveld.
Fig. 94. — Encephalartos caffer (form): near Slippery Drift in the East London district, Cape.
511
Fjg. 95. — Encepkalartos coffer-, left, female cone; right, male cone, from plants in cultivation by Miss G.
Blackbeard, Grahamstown, Cape.
Notable features of the typical form of E. caffer are the density and twisting of some
of the median leaflets and the relatively large tuberous roots. A feature likely to cause
confusion is that the leaflets of young plants are freely toothed on the margin whereas
those on mature plants are generally entire and neddle-pointed.
512
Fig. 96. — Encephalartos cuffer:
Cape, showing large
Institute, Pretoria.
yp.cal old male plant in Miss G. Blackbeard’s ^rahamsto^
berous rootsystem; since transferred to garden of Botanical Research
513
Fig. 97. — Encephalartos longifolius: typical old male plant, thought originally by Thunberg to be the
adult form of the dwarf species E. coffer shown in Fig. 93-96.
514
G. G. Smith notes that specimens from the neighbourhood of East London have
less rigid leaves than those from Bathurst and the leaflets are rolled inwards in the young
stage, unlike the usually spreading ones in the Bathurst district. But M. J. Wells has
recently detected inrolling in a Bathurst plant. Lack of adequate study material has so
far prevented a full picture being obtained and for the time being no further segregation
from Encephalartos cajfer is contemplated other than E. ngoyanus from Zululand.
E. coffer was one of the first two species of Encephalartos to be discovered in southern
Africa. The Swedish botanist Thunberg and the English botanical collector Masson
travelled down the Langkioof valley in 1772 on their way to the neighbourhood of
Coega, north-east of Port Elizabeth. On the journey Thunberg collected material of an
arborescent and a dwarf species, but considered them to be forms of one species which he
named Zamia caffra. It was not until long after that the confusion was unravelled and
the name Encephalartos coffer restricted to the dwarf species while the name E. longifolius
was restored to the arborescent species (Fig. 97).
Conclusion
Whatever is recorded here on the subject of the Cycads of Southern Africa must be
viewed in the light of the old adage that the exception proves the rule. Very little about
Cycads can be said to be constant except their inconstancy. To detail all the variations,
and deviations from the presumed normal for growth form and growth rhythm of stems
leaves and cones, would be a major undertaking. This applies to plants both in the
wild and in cultivation but more particularly to the latter.
It is as well to remind students that the Cycads present many problems of interest,
fascinating to some, which call for further investigation. The field of research is still
wide open.
A measure of support exists for the views of Johnson in Proc. Linn. Soc. N.S.
Wales 84 : 64 (1960) that the genera should be segregated under several smaller famillies.
Tiie problems involved are still under consideration.
515
INDEX TO SPECIES Species
No. Page
Encephalartos altensteinii 16 477
arenarius ■. 13 470
caffer 26 509
cycadifolius 2 439
eugene-maraisii 8 455
eximiits — 439
ferox 22 499
friderici-guilielmi 3 442
ghellinckii 1 437
horridus 12 467
humilis. 5 448
inopinus 7 453
kosiensis — 499
laevifolius 6 451
lanatus 4 445
latifrons 14 471
lebomboensis 19 491
lehmannii 10 460
longifolius 15 474
natalensis 18 486
ngoyanus 25 508
paucidentatus 21 498
princeps 9 458
St Hat us — 405
transvenosus 20 494
trispinosus 11 463
umbeluziensis 24 505
villosus 23 502
woodii 17 482
Stangeria eriopus 1 429
Zamia cycadifolia 2 439
caffra 26 509
honida 12 467
longifolia 15 474
516
!
t
1
517
INDEX TO VOLUME VIII
Page
(S = Supplement)
Achnatherum pekenense (Hance) OhwL. . . 217
Acanthosicyos Welw. ex Hook f. 52
horrida Welw. ex Hook. f. 3, 53
Aciachne Benth 205, 207
Acocks, J. P. H S5
Acokanthera sp 174
Adenopus Benth 96
Admiraal, J S5
Agrosteae 205, 207
Albuca namaquensis Bak 136
Aloe prinslooi S21
Amaryllidaceae, Notes on the 163
Amphipogon, R. Br 221, 222, 380
Anderson, J. G SI
Anderson, J. G. Notes on the Gramineae 113, 170
Andropogon lacunosus J. G. Anderson. . . 113
pratensis Hochst. ex Hack 113, 114
Anisopogon, R. Br 205
Anthericum 147
longituberoswn Poelln 147
vaUistrappii Poelln 147
Aristella Bertol 205
Aristida L 229, 234, 300, 302, 392
Sect. Aristida 238, 300, 302
Sect. Arthratherum (P. Beaiiv.) Reichenb. 266,
227, 233, 238, 240, 299, 300, 301, 302, 303
Sect. Chaetaria (P. Beauv.) Trin. 220, 226, 227,
238
Sect. Pseudarthratherum Cliiov. ex Henr. 233,
238, 240, 299, 300, 301, 302, 303
Sect. Pseudochaetaria Henr. 226, 233, 236, 238,
300, 302
Sect. Schistaclme (Fig. & De Not.) Henr. 220,
226, 227, 228, 234, 307, 308, 309
Sect. Schizachne {Trin. & Riipr.) De Wint. 236,
238, 299, 300, 302
Sect. Stipagrostis Trin. & Rupr. 220, 226, 227,
228, 234, 307, 309
Sect. Streptachne 220, 226, 227, 231, 243, 299
Aristida acutiflora Trin. & Rupr 220
adscensionis L. 221, 225, 236, 238, 254, 301,
302
subsp. guineensis {Trin. & Rupr.) Henr. 239,
240, 250, 254
var. pygmaea (Trin. & Rupr.) Dur.
& Schinz 253
var. strictiflora (Trin. & Rupr.) Dur.
& Schinz 253
aequiglumis Hack. 233, 239, 241, 261, 262,
265, 301, 302
alopecuroides Hack 297
amabilis Schweick 324
andoniensis Henr 249
angolensis C. E. Hubb 226, 228, 384
august at a Stapf 268
argentea Schweick 240, 241, 288, 289
astroclada Chiov 238
Page
atroviolacea Hack 270
barbicollis Trin. & Rupr. 173, 221, 233, 298,
392
var. conglomerata Henr 298
bifida Karl 355, 357
bipartita {Nees) Trin. & Rupr. 239, 240, 257,
260, 301, 302, 343
brachyathera Coss. & Balansa 220
brevifolia (Nees) Steud 338
burkei Stapf Til, 299, 346
caerulescens Desf 251
var. breviseta Hack 251
canescens Henr. 233, 239, 241, 259, 260, 265,
295, 301, 302
subsp. canescens 258, 259, 260
subsp. ramosa De Wint 260
capensis Thunb 348, 351
var. barbata Stapf 347
var. canescens Trin. & Rupr 347
var. dieterleniana Schweick 229, 349
var. fulviberbis Trin. & Rupr 348
var. genuina Henr 348
var. macropus Nees 348
var. zevheri (Nees) Walp 347
ciliata Desf. 220, 221, 304, 316, 351
var. capensis Trin. & Rupr 316
var. pectinata Htm 316
var. tricholaena Hack 316
var. villosa Hack 316
coma-ardeae Mez 320
confusa Trin. & Rupr 253
congesta Roem. & Schult. 173, 232, 233, 237,
240, 241, 296, 302
subsp. barbicollis {Trin. & Rupr.) De
Wint 173, 298
subsp. congesta 173, 295, 297
var. genuina Chiov 297
var. megalostachya Henr 297
corythroides Karl 336
curvata {Nees) Trin. & Rupr. 233, 239, 240,
253, 302
var. nana (Nees) Henr 253
damarensis (Mez) De Wint 329
dasydesmis Mez 233, 235, 240, 241, 273, 275
282, 283, 301, 302, 392
diffusa Trin. 235, 240, 241, 273, 275, 281, 282,
299, 301, 302, 346
var. brevistipitata (Trin. & Rupr.)
Henr 277
var. burkei {Stapf) Schweick. 233, 277, 281,
283, 284, 301
var. densa (Trin. & Rupr.) Henr... 277
var. diffusa 275, 277
var. eckloniana (Trin. & Rupr.)
Henr Ill
var. genuina Henr 211
var. schraderiana (Trin. & Rupr.)
Henr Ill
0140996—2
518
Page
dinteri Hack 320
dregeana (Nees) Trin. & Rupr 344
effusa Henr. 233, 239, 240, 244, 251, 257,
260, 299, 302
elliptica Kiinth 299
elvmoides Mez 287
engleri Mez 231,240,241,281,302
var. engleri 281
var. ramosissima De Wint 282
fastigiata Hack 338
floccicidmis Mez 279
fontismagni Schweick 286
funiculata Trin. & Rupr 221
gatpinii Stapf 270
garubensis Pilg 343
geminifolia (Nees) Trin. & Rupr 341
gonatostachys Pilg 353
gossweileri Pilg 270
gracili flora Pilg 292
var. robusta Stent & Rattray 291
gracilior Pilg 361, 363, 365, 367
var. intermedia Schweick. 173, 306, 307,
361, 365
var. pearsonii Henr 365
gracilUma Oliv 286
guineensis Trin. & Rupr 250
hermannii Mez 369
hirtigluma Steud 221, 363
var. panda Hack 365
hocbstetteriana Back 315
hockii De Wild 270
hordeacea Kiinth 233, 236, 238, 240, 244,
249, 250, 299, 300, 302, 392
var. longiaristata Henr 244
hubbardiana Schweick. 239, 240, 245, 250,
299, 300, 301, 302
hygrometrica R. Br 238
hystrix sensu Thunb 277
jucunda Schweick 226, 227, 228, 384
junciformis Trin. 8c Rupr. 231, 232, 233, 239,
241, 262, 266
subsp. galpinii {Stapf) De Wint. 231, 268,
270
subsp. junciformis 268, 301, 302
lamarckii (Roem. & Schult.) Steud... 286
lanipes (Mez) De Wint 355
lanuginosa Burch 279
longicauda Hack 297, 298
lutescens (Nees) Trin. & Rupr 336
var. marlothii (Hack.) Stapf 336
marlothii Hack 336
meridionalis Henr. 233, 240, 241, 284, 287,
302, 303
mollissima PUg. 240, 241, 284, 287, 290, 302,
303, 392
monticola Henr. 239, 241, 265, 273, 30l’, 302
namaquensis (Nees) Trin. & Rupr.. .221, 325
var. vagans (Nees) Walp 325
obtusa Del 221, 304, 306, 307, 355
parvula (Nees) De Wint. 231, 236, 238, 240,
242, 299, 301, 302
pennata Trin 220
pilgeri Henr 240, 241, 293, 296, 303
phunosa L 220
proxima Steud 322
pseudo-hystrix (Trin. & Rupr.) Steud. 277
AGE
pungens Desf 220, 221, 306
var. pennata 220
pusilla Trin. & Rupr 253
pygmaea Trin. & Rupr 253
rangei Pilg 297
recta Franch. 231, 239, 241, 269, 270, 301, 302
rhiniochloa Hochst. 233, 239, 240, 249, 301,
302
rigidiseta Pilg 249
riparia Trin 301
subacaulis (Nees) Steud 273
sabidicola Pilg 331
scabrivalvis Hack. 233, 239, 240, 253, 254,
260, 302
var. contracta De Wint 257
var. scabrivalvis 257
schaeferi Henr 351, 352
sciurus Stapf 239, 241, 273, 275, 300, 302,
392
secalina Henr 315
sericans Hack 201, 234, 299, 304, 349
serrulata Chiov 249
spectabilis Hack. 233, 240, 241, 283, 284, 285,
287, 288, 290, 302, 303, 341, 343
steudeliana Trin. & Rupr 244
stipitata Hack 233, 240, 241, 290, 302
var. graciliflora {PUg.) De Wint.... 292
var. robusta (Stent & Rattray) De
Wint 291
var. spicata De Wint 292
var. stipitata 233, 290, 291
stipoides Lam. 240, 241, 285, 286, 302, 303
var. meridionalis Stapf 284
strict iflora Trin. & Rupr 253
subacaulis (Nees) Steud 373
submucronata Schum 254
transvaalensis Henr. 233, 239, 241, 263, 299,
301, 302, 303
uniplumis Licht 211, 306, 359
var. neesii Trin. & Rupr 360
var. pearsonii Henr 360
vanderijstii De Wild. 226, 221, 228, 328, 387
vestita Thunb. 240, 241, 275, 279, 282, 283,
302
forma amplior Hack 279
var. brevistipitata Trin. & Rupr... 277
var. densa Trin. & Rupr 277
var. diffusa (Trin.) Walp 277
var. parviflora Trin. & Rupr 275
var. pseudo-hystrix Trin. & Rupr... 277
var. schraderiana Trin. & Rupr 277
waibeliana Henr 251
walteri Suesseng 338
welwitschii Rendle 268
zeyheri (Nees) Steud 347
ARISTIDEAE C. E. Hubb. 202, 205, 212, 219,
220, 221, 222, 223, 224, 225, 226, 228, 234, 279,
380, 393
Arthratherum (Beauv.) Reichb 238, 304
brevifolium Nees 338
lutescens Nees 336
namaquense Nees 325
obtusum (Del.) Nees 355
pungens Beauv 220
subacaulis Nees 373
uniplume (Licht.) Nees 360
519
Page
vestitum (Thunb.) Nees 279
zeyheri Nees 347
ARUNDINEAE Diimort 224, 225
Asclepiadaceae, Notes on the 164
Atheranthera Mast 8
AVENEAE Diimon 207
BAMBUSEAE 209
Bezuidenhout, W S3
Blastania Kotschy & Peyr 11
cerasiformis (Stocks) A. Meeiise 12
fimbristipula Kotschy & Peyr 12
garcinii (L.) Cogn 11
liiederilziana Cogn. ex Schinz 10
Botanic Gardens S. 18, 19, 20
Staff S5
Review of work SI 8
Botanical Survey Section
Staff S5
Review of work S21
Bothalia S37
Brachyelytrum Beaiiv 207
Breytenbach, Mrs. A. M. C S4
Briggs, Miss R. A S4
Brink, Mrs. E S4
Britten, Miss G. V S4
Bryonia africana L 24, 32
africana sensu Thunb 30, 32, 33
angulata Thunb 19, 30, 32
convohiiloides A. Rich 105
cordata Thunb 19
digitata Thunb 31, 32
dissecta Thunb 32
fimbristipula Fenzl 12
grossulariaefolia Drege, nom. niid. ... 32
jatrophaefolia A. Rich 105
laciniosa L 33
laevis Thunb 19, 20
lagenaria Drege, tiom. nud. 98
micropoda Drege, nom. nud 14
multifida Drege, nom. nud. 32
nana Lam 30
obtusiloba Drege, nom. nud 27
pinnatifida Burch 32
punctata Thunb 19
quinqueloba Thunb 99
re panda Blume 19
scabra Thunb 19
scabrelia L.f 14
triloba Thunb 30, 32
Bryoniopsis laciniosa (L.) Naud 33
Calitz, Miss A. S. C S3
Campanulaceae, Notes on the 164
Capparidaceae, Notes on the 165
Capparis calvescens Gilg & Bened 165
fianaganii Gilg & Bened 165
rudatisii Gilg & Bened 165
sc/dechteri Schinz 165
solanoides Gilg & Bened 166
transvaalensis Schinz 165
var. calvescens (Gilg. & Bened.)
Marais 165
var. transvaalensis 165
Cephalandra Schrad 95
rnackenii Naud 97
palmata Sond 97
pubescens Sond f05
Page
quinqueloba (Thunb.) Schrad. ex Eckl.
& Zeyh 99
sessilifolia Sond 98
Ceratiosicyos Nees 20
ecklonii Nees 20
laevis (Thunb.) A. Meeuse 20
Ceropegia ampliata E. Mey 164
carnosa E. Mey 164
estelleana R. A. Dyer 164
fimbriata E. Mey 164
Chamaegigas intrepidus S33
Chaetaria Beauv 205, 238
bipartita Nees 257
capensis (Thunb.) Beauv 348
congesta (Roem. & Schult.) Nees.... 297
curvata Nees 253
lamarckii Roem. & Schult 286
var. nana Nees 353
vestita (Thunb.) Beauv 279
Cheiranthus carnosus Thunb 168
elongatus Thunb 167
gramineus Thunb 168
CHLORIDEAE Agardh 223, 224, 234
Chlorophytum Ker-Gawl 147
bowkeri Bak 147
dregei Poelln 147
krookianum Zahibr 147
leipoldii Poelln 147
longituberosum (Poelln.) Oberm 147
namaquense Schitr. ex Poelln 147
pulchellum Kunth 147
rigidum Kunth 147
schlechterianum Poelln 147
triflorum (Ait.) Kunth 147
vallistrappii (Poelln.) Oberm 147
wilmsii Engl. & Krause 147
Citrullus Schrad. in EckL & Zeyh 54
subgen. Pseudocucumis A. Meeuse 55
subgen. Citrullus 55
amarus Schrad 57
battich Forsk 57
ecirrhosus Cogn 58
lanatus (Thunb.) Mansf. 3, 57, 59
naudinianus (Sond.) Hook. f. 3, 55
vulgaris Schrad. ex Eckl. & Zeyh. 54, 55, 57
Cleome minima Stephens 166
Coccinia Wight & Arn 95
adoensis (A. Rich.) Cogn 105
cordifoUa Cogn 96
grandis (L.) /. O. Voigt 96
hirtella Cogn 101
indica Wight & Arn 96
jatrophaefolia (A. Rich.) Cogn 105
var. australis Cogn 105
rnackenii (Naud.) Cogn 97, 100
ovifera Dinter & Gilg ex Dinter 102
palmata (Sond.) Cogn 96, 99, 100, 101
parvifolia Cogn 105
pubescens (Sond.) Cogn. ex Elarms.. 105
quinqueloba (Thunb.) Cogn 99, 101
rehmannii Cogn 15, 42, 102
var. littoralis A. Meeuse 104
var. rehmannii 102
roseifiora Suesseng 105
schinzii Cogn 98
sessilifolia (Sond.) Cogn 98, 100, 106
variifolia A. Meeu.se 100
520
Page
Codd, L. E SI, 37, 54, 67
Codd, L. E. Mr. L. C. Leach S67
Codd, L. E. Notes on the Guttiferae . . . 174
Codd, L. E. Notes on some Kniphofias
in the eastern Free State S37
Codd, L. E. Notes on the Scrophulariaceae 176
Codd, L. E. The South African Species of
Orthosiphon 149
Codd, L. E. and Voorendyk, S. Plants
which cause Gousiekte poisoning S54
Coetzee, Mrs. M. B SI
Colocynthis Mill 54
citruUus (L.) ex Kuntze 54, 55, 57
ecirrhosiis (Cogn.) Chakrav 58
naiidiniamis (Sond.) Kuntze 55
vulgaris Schrad 54, 58
Congresses S26
Coniandra Schrad 24
africana (L.) Sond 32
digitata (Thunb.) Sond 32
dissecta Schrad 32
glauca (Thunb.) Schrad 32
var. dissecta Sond 34, 35
grossulariaefolia E. Mey. ex Arn 32
moUe (Kunze) Sond 30
pinnatisecta Schrad 32
pimctidata Sond 32
thimbergii Sond 30
zeyheri Schrad 31
Contemporary Botanists and Collectors
1. Dr. F. Z. van der Merwe S65
2. Mr. L. C. Leach S67
Corallocarpus Welw. ex Benth. & Hook. f. 38
bainesii (Hook. /.) A. Meeuse. . . . 38, 40, 41
bequaertii De Wild 39
dinteri Cogn 41
dissectus Cogn 38
gilgianus Cogn 39
glaucicaidis Dinter 39
scaber Dinter 39
schinzii Cogn 39
var. lobatus Cogn 40
sphaerocarpus Cogn 40, 41
var. hastatus Cogn 41
var. scaberrimus Cogn 41
welwitschii (Naud.) Hook. f. ex Welw. 38, 39
var. siibintegrifolia Cogn 40
Creffield, Mrs. T. M SI
Cruciferae, Notes on the 166
Ctenolepis Hook, f H
cerasiformis (Stocks) Hook, f 12
Ctenopsis Hook. f. ex Naud 11
cerasiformis (Stocks) Naud 12
Cucumis L ; 59
abyssinicus A. Rich 77
abyssiniciis Schimp. ex Hook, f 81
africanus L.f. 3, 70, 72, 76, 79, 81
var. aciitilobus Cogn 79
var. myriocarpus (Naud.) Burtt Davy 74
var. zeyheri (Sond.) Burtt Davy... 79
angolensis Hook. f. ex Cogn 64
anguria L 77
var. anguria 77
var. longipes {Hook. /.) A. Meeuse 77, 80
arenarius Schrad 79
asper Cogn 66
chrysocomus Schum. & Thonn 79
Page
cogniauxianus Dinter & Cogn 66
cordifolius Drege, nom. nud 47
dinteri Cogn 65
dissectifolius Naud 3, 55, 74
ficifoUus sensu Cogn. var. echinophorus
(Naud.) Cogn 77
figarei Del. ex Naud 81
var. echinophorus Naud 77
figarei sensu Hook. f. 77
heptadactylus Naud 69, IQ
hirsutus Sond. 63
homblei De Wild 64
hookeri Naud 3, 70, 72, 80
humifructus Stent 62
kalahariensis A. Meeuse 70
leptodermis Schweick 76
longipes Hook, f 77
maderaspatensis L 14
melo L 3, 60, 61
membranifoUus Hook. f. 22
merxmuelieri Suesseng 74
metuliferus E. Mey. ex Schrad. 68
myriocarpus Naud. 3, 74, 76, 77, 80
naudinianus Sond 55
prophetarum Jacq 74, 81
pueschelii Dinter 63
pustulatus Hook 71, 74, 76, 81
rigidus E. Mey. ex Naud. 71
sativus L 60
seretii De Wild 64
seretioides Suesseng 64
sonderi Cogn 63
subsericeus Hook, f 22
umbrosus A. Meeuse 8c Strey 67
welwitschii Cogn 64
wildemanianus Cogn 64
zeyheri Sond. 3, 73, 74, 79
Cucurbita citrullus L 57
anguria Duch 57
lagenaria L 82, 83
leucantha Duch 83
siceraria Mol 83
CUCURBITACEAE 3
Cucurbitaceae of Southern Africa, The 1
Curatorial Services SI 2
Cycadaceae, Notes on the 169
Cycads of Southern Africa, The 405
Cycloptychis marlothii S35
Cynanchuni sarcostemmatoides K. Schum. 164
tetrapterum {Turcz.) R. A. Dyer 164
Cyrtanthus eucallus R. A. Dyer 163
sanguineus (Lindl.) Walp 163
Cyrtonema Schrad 24
molle Kunze 30
sphenoloba Schrad 31
triloba (Thunb.) Schrad 30
Dactyliandra Hook. f. 10
luederitziana (Cogn.) Cogn 10
welwitschii Hook. f. 10
Danthonia DC 222, 223, 380
aureocephala /. G. Anderson 170
davyii C. E. Hubb 170, 171, 172
draicensbergensis Schweick 171, 172
macowanii S/ap/. 171, 172
stereophylla J. G. Anderson 171, 172
521
Page
DANTHONIEAE Nevski 222, 223, 380
Decimal coins. Flowers on S43
De Groot, Miss C S2
De Winter, B SI, 26
De Winter, B. New combinations and new
species in the Aristideae 173
De Winter, B. The South African Stipeae
and Aristideae 201
Dichelachne 205, 206, 207
Dietes butcheriana S9
Dipcadi, The South African species of. 117
Dipcadi Medik 118
Sect. Tricharis 117
Sect. Uropetalum 117
adelae Beauv 122
anthericoides Engl. & Gilg 127
ausense Dinter 136
avasimontanum Dinter 136
bakerianum H. Bolus 118, 119, 120, 129, 131,
135, 136
bakerianum sensu Schinz 125
baumii Engl. & Gilg 134
brevifolium (Thunb.) Fourc 117, 120
brevipes Bak 129
ciliare (Zeyh. ex Harv.) Bak. 117, 120, 121
ciliatum Engl. & Krause 134
cinnabarinum Suesseng 131, 133
clarkeanum Schinz 135
comosum Welw. ex Bak 131, 133
conrathiiBak 131
crispociliatum Suesseng 128, 129
crispum.Bayt 118, 120, 128, 129, 135
dinteri BaA: 118, 135
durandianum Schinz 125
elatum Bak 131
ernesti-ruschii Dinter 125
filamentosum Medik 131
geniculatum Dinter & Suesseng 131, 133
glaucum (Ker-Gawl.) Bak. 118, 119, 120, 130,
131, 135, 136
gonocarpum Suesseng 129
gracilipes YjidiUSQ 130
gracillimum BaA: 117, 118, 120, 124
helenae Beauv 131
heterocuspe Bak 117
hyacinthiflora (Berg, ex Schlechtd.)
Hutch 121
hyacinthoides Bak 121
involutum Suesseng 124
juttae Engl. & Krause 135
lateritium Welw. ex Bak 131, 133
lividescens Engl. & Gilg 131
longibracteatum Sc\ur\z 130
Engl. & Krause 131
longifolium (Lindl.) Bak. 117, 118, 119, 120,
126, 130, 131
magnum Bak 130, 131
marlothii Engl. 117, 118, 119, 120, 121, 123,
125
megalanthum Zahlbr 122, 123
oligotrichum Bak 125
palustrel&ak 131
papillatum 119, 120, 123
platyphyllum BaA: 118, 120, 134, 135
polyphyllum Bak 118, 124, 126
rautanenii Bak 129
readii Bak 122, 123
Page
rhodesiacum Weim 133
rigidifolium BaA: 120, 133
serotinum (L.) Med/A: 117, 119
setosum Bak 128, 129
var. read/7 Bak 122
spirale Bak 121
stenopliyllum Dinter 131, 133
sulcatum Suesseng 131
tortile R. A. Dyer 128
umbonatum (Bak.) Bak 118, 131
undulatifolium Schinz 128, 129
vaginatum BaA: 120, 134, 135
ve«e/7ar//w Schinz 126, 128
viride (L.) Moench..m, 118, 119, 120, 131
volutum Bak 128, 129
zambesiacum Bak 126
Diplachne Beauv 380
Diplopogon B. Br 220, 221, 222, 380
Dombeya rotundifolia var. nov S33
Du Toil, Mrs. L. D S3
Du Toit, S. W SI
Dyer, R. A S2, 49, 51, 52
Dyer, R. A. Field work is not all fun.. S50
Dyer, R. A. Notes on the Amaryllidaceae 163
Dyer, R. A. Notes on the Asclepiadaceae 164
Dyer, R. A. Notes on the Cycadaceae . . 169
Dyer, R. A. The Cycads of Southern
Africa 405
Economic Botany SI 7
Edwards, D S5, 28
Edwards, D. Overseas study tour S28
Ehrharta Thunb 205
Elodea crispa Hort 145
Els, C. A S52
Encephalartos Lehm 432
altensteinii Lehm 4ni
arenarius R. A. Dyer 470
caffer (Thunb.) Lehm 509
cycadifolius (Jacq.) Lehm 439
eugene-maraisii Verdoorn 455, S51, 53
ex/w/'//^ Verdoorn 439
ferox Bertol. f. 499
friderici-guilielmi Lehm 442, S5]
ghellinckii Lem 437
horridus (Jacq.) Lehm 467
var. trispinosus Hook 463
humilis Verdoorn 448
inopinus R. A. Dyer 169, 453, S52
laevifolius Stapf & Burtt Davy 451
lanatus Stapf & Burtt Davy 445
latifrons Lehm 471
lebomboensis Verdoorn 491
lehmannii Lehm 460
longifolius (Jacq.) Lehm 474
natalensis Dyer & Verdoorn 486
ngoyanus Verdoorn 408
paucidentatus Stapf & Burtt Davy 498
princeps R. A. Dyer 458
transvenosus Stapf & Burtt Davy 494
trispinosus (Hook.) R. A. Dyer 463
umbeluziensis R. A. Dyer 505
villosus Lem 502
woodii Sander 482
ERAGROSTEAE Stapf.. 101, 219, 223, 224, 230
Eragrostis Beauv 380
ERAGROSTOIDEAE Pilg 225
0140996—3
522
Page
Eriochrysis brachypogon Stapf. 170
subsp. australis J. G. Anderson 170
Eulenburgia mirabilis Pax 47
Eureiandra Hook. f. 96
Exchanges S16
Expeditions S32
Fadogia monticola S63
Felicia echinata S12
Ferns SIO
FESTUCEAE Dumort 207, 208, 222
FESTUCOIDEAE Hitchc 207, 209, 224
Flora of Southern Africa S6
Flowering Plants of Africa S37
" Fluvialea” Drege 140
Forssman, Mrs. C. L. (See also Cythna
Letty) S3
Fiicus cirrhosis Turner 191
corallorhiza T urner 191
cornutus Turner 189
Garcinia gerrardii Harv. ex Sim 174
natalensis Schltr 174
Gerrardanthus Harv. ex Benth & Hook. f. 8
macrorhizus Harv. ex Benth. & Hook. f. 8
megarhiza Decne & Harv 8
tomentosus Hook. f. 9
transvaalensis Burtt Davy 174
volkensii Engl 174
Gerber, Miss M. H S3
Graderia iringensis Melch 176
fruticosa Ba//. /. 176
linearifolia L. E. Codd 176, 181
scabra (L./.) Benth 176
subintegrae Mast 176
Gramineae, Notes on the 113, 170
Grobler, P. J S5, 40
Gunn, Miss M.D SI
Guttiferae, Notes on the 174
Gymnopetalum sensu Baill 85
Hagedoorn, Miss E. F. E SI
“ Haloragea" Dregs 140
Hardy, D. S S6
Heliophila, Notes on the genus 166
abrotanifolia Banks ex DC 168
abrotanifolia Banks ex Sond 168
var. heterophylla Sond 168
var. tenuiloba Sond 168
var. tripartita (Thunb.) Sond 168
aspera Schltr 166
carnosa {Thunb.) Stead. 168
clavuligera Schulz 167
cornuta Sond. 167
var. cornuta 167
var. squamata (Schltr.) Marais 167
crithmifolia Willd. 166, 168, 169
dolichostyla Schltr 167
elongata (Thunb.) DC 167
exilis Schltr 167
graminea (Thunb.) DC 168
grandiflora Schltr 168
laciniata Marais 168
linearifolia Burch, ex DC. 168
longifolius Eckl. & Zeyh 167
marlothii Schulz 167
minima (Stephens) Marais 166
nigellifolia Schltr 167
oreophila Schltr 167
pearsonii Schulz 166
var. edentata Hainz 166
var. Schulz 166
Page
pinnata L.f. 167
platysiliqua R. Br 168
scoparia Burch, ex DC 166, 167
var. aspera (Schltr.) Marais 166
var. scoparia 166
seselifolia Burch, ex DC 166
var. marlothii (Schulz) Marais 167
var. nigellifolia (Schltr.) Marais 167
var. seselifolia 166
smithii Schulz 166
squamata Schltr 167
stylosa Burch, ex DC 167
subcornuta Beauv 168
sulcata Conrath 168
trifida Thunb 167
tripartita Thunb 168
virgata Burch, ex DC 167
Hemizygia bolusii (N. E. Br.) L. E. Codd 159
canescens (Guerke) Ashby 159
foliosa S. Moore 159
mossiana (R. Good) Ashby 159
petrensis (Hiern) Ashby 159
Herbert, Miss W. A S5
Heterosicyos Welw. ex Benth. & Hook. f. 85
Hippocrateaceae, Notes on the 114
Homoeostrichus harveyana Pappe ex Kuetz. 197
Hyacinthus brevifolius Thunb 118, 120
serotinus L 118
viridis L 131
Hydrilla dregeana Vresl 140
muscoides Planch 140
Hymenosicyos Chiov 21, 60
bequaertii (De Wild.) Harms 23
bryoniifolius Merxm 23
membranifolius (Hook, f.) Chiov 23
subsericeus (Hook, f.) Harms 23
triangularis (Cogn.) Harms 23
villosus (Cogn.) Harms 23
International Biological Programme S26
International Botanical Congress S26
International Union of Biological Sciences S26
Iridaceae, Notes on the 115
Jellinek, Mrs U S3
Jenkinson, T. R SI
Jessop, J. P S2, 40
Johansson, Miss E. Y S3
Kalahari, aerial view S22
Kat, Miss M S3
Kedrostis Medik 24, 38
subgen. Gilgina Cogn 25
subgen. Kedrostis 25
subgen. Toxanthera (Hook.f.) A. Meeuse 25
africana (L.) Cogn 24, 29, 32, 35
angulata (Berg.) Fourc 30
bainesii (Hook, f.) Cogn 38, 41
capensis (Sond.) A. Meeuse 28, 32
cinerea Cogn 17, 38
crassirostrata Bremek 35
digitata (Thunb.) Cogn 31
var. major Cogn 31
eminens Dinter & Gilg 36
foetidissima (Jacq.) Cogn 26
subsp. foetidissima 26
subsp. obtusiloba (Sond.) A. Meeuse 27
var. divergens (Hochst.) Cogn 26
var. genuina Cogn 26
var. microcarpa Cogn 27
var. perrottiana (Ser.) Cogn 26
523
gilgiana Cogn
glauca (Schrad.) Cogn
var. dissecta (Sond.) Cogn.
ledermannii Cogn
longipedunculata Cogn
minuti flora Cogn.
Page
36
32
35
37
36
27
mollis (Kunze) Cogn 30, 32
nana {Lam.) Cogn 19, 21, 25, 29, 33
var. nana 30, 31, 32
var. schlechteri {Cogn.) A. Meeuse 31
var. zeyheri {Schrad.) A. Meeuse.. i), 32
natalensis {Hook. /.) A. Meeuse 25, 36
ohtusiloba (Sond.) Cogn 27
otaviensis Dinter 36
punctidata (Sond.) Cogn 32, 35
var. tenuUoba (Sond.) Cogn 34
rau tone nil Cogn 36
rigidiuscula 37
rostrata (Rottl.) Cogn 26
schlechteri Cogn 31
velutina Cogn 30, 32
zeyAm (Schrad.) Cogn 25, 29, 31
var. angustiloba (Sond.) Cogn. 29, 31, 34
Keytel, C. L S6
Killick, D. J. B S5
Killick, D. J. B. Notes on the Campanu-
laceae 164
Killick, D. J. B. A note on the identity
of Myrica conifera Burm. f. 175
Kniphofia caulescens S39
ensifolia S38
multiflora S39
sp. nov S38
Lachenalia viridis (L.) Thunb 131
Lagarosiphon, The South African species of 1 39
Lagarosiphon Harv 139
cordofanus Casp 146
crispus Rendle 143
fischeri Guerke 146
ilicifolius 06erw 141, 145
major {Ridlev) Moss ex Wager 139, 140, 141,
145
muscoides //arv 139, 140, 144
var. major Ridley 145
schweinfurthii Casp 146
steudneri Casp 146
verticillifolius Oberm 141, 142
Lagenaria Ser 82
mascarena Naud 84, 86
sagittata Harv 87
siccrana (Molina) Standi 82, 83, 91
sphaerica Drege ex Naud 84
sphaerocarpa E. Mey. ex Arnott 84
vulgaris Ser 82, 83
Landman, J S2
Lasiagrostis Z,/>7^ 205, 206, 216
capensis 'iAtQS 215, 216, 217
var. elongata (Nees) Trin. & Rupr. 216
elongata Nees 216
Leach, Mr. L. C S67
Leendertz, Miss R S40
Leistner, O. A SI, 40
Lemmer, Miss C. M S3
Leptormus iongifolius Eckl. & Zeyh 167
Lepturus R. Br 380
Letty, Cythna S3, 45, 46, 47
Page
Letty, Cythna. The flowers on the new
decimal coins S43
Liebenberg, Miss C S3
Liebenberg, Miss E S3
Lomaria eriopus Kunze 429
Loubser, Miss E. J S4
Lufl'a Caledon ica Sond 94
sphaerica E. Mey. ex Sond 84
Lycurus Kunth 207
Macrochloa K;//;?/? 205, 206
Marais, W S2, 35
Marais, W. Notes on the Capparidaceae. . 165
Marais, W. Notes on the Genus Heliophila 1 66
Marine Algology SIO
Marsh, Miss J. A S2, 34
Mauve, Mrs. A. A. Obermeyer- S2, 33
Maydeae Dumort 223
Meeuse, A. D. J. The Cucurbitaceae of
Southern Africa 1
Melhania acuminata 177, 178, 179
forbesii 178, 179
Integra Verdoorn 177
polygama Verdoorn 178, 179
prostrata 177, 178
Melothria L 13, 67
acutifolia Cogn 15
breyeri Burtt Davy 17
cinerea (Cogn.) A. Meeuse 17, 38
cordata (Thunb.) Cogn 19, 32
foetida Lam 26
hederacea (Sond.) Cogn 21, 30
hispidida Burtt Davy 17
maderaspatana (L.) Cogn 14, 24
marlothii Cogn 15, 18
membranifolia Cogn 19
minutiflora Cogn 18
natalensis Cogn 21
obtusiloba (Sond.) Cogn 27
parvifolia Cogn 17
pendula L 13
punctata Rafin 20
punctata (Thmb.) Cogn 19, 32, 82
scabra Naud 20
tomentosa Cogn 19
tridactyla Hook. f. 21
velutina (Schrad.) Cogn 19
'MWmm Adans 205, 206, 207
Moll, E. J S5
Momordica L 45
sect. Raphanistrocarpus Baill 42
adoensis Hochst. ex A. Rich 105
angustisepala Harms 47
balsamina L 49, 51
var. huberi Naud 50
var. leucantha Ch. Huber 50
boivinii Baill 44
cardiospermoides liS.\oizsc\\ 46
charantia L 46, 51
var. abbreviata Ser 51
clematidea Sond 46
cordata Cogn 47
cordifoUa Sond 47
cucullata Hook. f. 47
foetida Schum. & Thonn 47
gariepensis Drege ex E. Mey 49
huberi ToA 50
524
Page
involucrata E. Mey. ex Sond. 50
lanata Thunb 55, 57
marlothii Harms 51
morkorra A. Rich 47
palmata E. Mey 96
repens Bremek 51
schimperiana Naud 47
schinzii Cogn 49
welwitschii Hook. f. 48
Monerma Beam 380
Muehlenbergia Schreb 207
Muller, P. J S5, 40
Mukia Arn 13
scabrella Arn 14
Myrica aethiopica L 175
cerifera L 175
conifer a Burm. f 175
Myrica conifera Burm. f., A note on the
identity of 175
Myrica serrata Lam 175
NARDEAE Reichenb 207
Nasella Desv 205, 206, 208, 209, 210
Natal Herbarium S3
National Herbarium SI, 3
Nautochilus Brem 149, 150
amabilis Brem 149, 157
breyeri Brem 149, 157
labiatus (N. E. Br.) Brem 149, 157
urticaefolius Brem 149, 157
Obermeyer, A. A. Notes on the Iridaceae 115
Obermeyer, A. A. The South African
species of Anthericum, Chlorophytum
and Trachyandra. Addenda et Corri-
genda 147
Obermeyer, A. A. The South African
species of Dipcadi 117
Obermeyer, A. A. The South African
species of Lagarosiphon 139
Obermeyer-Mauve, Mrs. A. A. (see Mauve,
Mrs. A. A.)
Oliver, E. G S4, 40
Olivier, J. du T S6
OreosycQ Hook. f. 21,60
africana Hook, f 21, 22, 23
aspera Cogn 22
bequaertii De Wild 22
parvifolia Cogn 23
subsericea (Hook, f.) A. Meeuse 22
triangularis Cogn 22
villosa Cogn 22
Ornithogalum dipcadioides Bak 124
Orthoraphium Nees 205, 206
Orthosiphon, The South African species of 149
Orthosiphon Benth 149
sect, serrati Ashby 149, 150
subgen. Nautochilus (Brem.) L. E. Codd 150
subgen. Orthosiphon 150
allenii (C. H. Wr.) L. E. Codd 160
amabilis (Brem.) L. E. Codd 158
ambiguus H. Bolus 159
australis Vatke 149, 150
bolusii N. E. Br 159
dissimilis N. E. Br 151
Perkins 159
fruticosus L. E. Codd 153, 162
glabratus var. africanus Benth ..... 150
holubii A. £■. Br 159
Page
inconcinnus Briq 150
labiatus N. E. Br 152, 157
neglectiis Bx'iC) 150
petrensis Hiern 159
pseudoserratus 149, 156
rubicundus (D. Don) Benth. 152, 153, 160
serratus Schltr 149, 155, 156
tubiformis R. Good 149, 154
varians N. E. Br 159
vernalisL. E. Codd 152, 161
wilmsii Guerke 150
var. komghensis N. E. Br 150
Oryza L 205, 207
ORYZEAE Diimort 209, 223
Oryzopsis Michx. 205, 206, 209, 210, 216, 217
aequiglumis Diithie 217
keniensis VWgQX 210, 216
Pachystigma pygmaeum S55, 57
thamnus S59, 60
PANICEAE R. Br 207, 230
PANICOIDEAE 209
PAPPOPHOREAE Kunth. . . .207, 223, 224, 234
Pavetta harborii S61
schumanniana S62
Peponia Naud 91
caledonica (Sond;) Cogn 94
Naud 91,93
Peponium Engl 91
caledonicum (Sond.) Engl...9\, 92, 94, 95
chirindense (Bak. /.) Cogn 95
kilimandscharicum (Cogn.) Engl 95
mackenii (Naud.) Engl 91, 92, 93
usambarense (Engl.) Engl 92, 93, 94
vogelii 92
Periboeal brevifolia (Thunb.) Kunth 121
Perielema Presl. 207
Phaenosperma Munro ex Benth 205
Phillips, E. P S39
Phormium viride (L.) Thunb 131
Physedra Hook. f. 95
Pilogyne Eckl. & Zeyh. ex Schrad 13
garcinii Harv 90
suavis Schrad 20
velutina Schrad 19
Piptatherum Bcomv 205, 210, 216
coerulescens Desf. 210, 216
keniense (Pilg.) Roshev 216
Piptochaetium J. & C 205, 206, 208, 209
Pisosperma Sond 24
capense Sond 25, 28, 32
Plant Collecting Expeditions S32
Plectrachne Henr 222, 380
Plectranthus allenii C. H. Wr 159, 160
ambiguus (H. Bolus) L. E. Codd 159
bolusii T . Cookt 150, 157
coloratus E. Mey. ex Benth 159
dregei L. E. Codd 159
Plocamium on the South African coast.
Species of 183
Plocamium abnorme Okam 192
beckeri Simons 185
coccineum (Huds.) Lyngb 187, 191
coxaWoxhxzdL (Turner) Harv 185, 190, 191
coxxxutnm (Turner) Harv 187, 188, 189
delicatulum Baardseth 187
fullerae Schmitz & Mazza 189
glomeratum J. Ag 183
525
Page
Page
ghmemtum sensu Kylin 185
latiusculum Kuetz 185
membranaceum 5'MAr 184, 185
membranaceum sensu Ky\m 189
nobile J. Ag 189, 192
procerum Suhr 189
rigidum 186, 187, 189
\ax. tenuior Gruno'M 189
robertiae Schmitz & Mazza 191
subfastigatum Kuetz 183
suhrii Kuetz 189
telfairiae Harv. ex Kuetz 180, 192
Poa L 380
Po/ewaw/j/fl Berg, ex Schlechtd 119
hyacinthiflora Berg, ex Schlechtd 121
Polemannia Eckl 118, 119
Pott, Mrs. R. Leendertz S39
Ptilagrostis Griseb 206
Publications S37
Quantitative ecology S24
Radersma, J S6
Raphanistrocarpus (Baill.) Pax 42
asperifoUus Cogn 44
boivinii (Baill.) Cogn 44
Raphanocarpus Hook. f. 42
asperifoUus (Cogn.) Chiov 44
boivinii {Baill.) Chiov 44
humilis Cogn 43
var. prostratus Suesseng 43
kirkii Hook. f. 42
tuberosus DinXet 44
welwitschii Hook. f. 43
Raphidiocystis 79
chrysocoma (Schum. & Thonn.) Killick
& A. Meeuse 79
caillei Hutch. & Dalz 79
Relchella Steud 206
Reybum, J S2
Rhynchocarpa Schrad 24
bainesii Hook. f. 38, 41
foetida Schrad 26
welwitschii Naud 39
Salacia gerrardii Harv. ex Sprague 114, 115
wardii Verdoorn 114
Sartidia De Wint. 225, 228, 302, 303, 379, 380,
381
angolensis (Hubb.) De Wint. 227, 381, 382,
384
jucunda (Schweick.) De Wint. 227, 382, 384
vanderijstii {De Wild.) De Wint. 227, 382,
387
Scheepers, J. C S5
Schistachne Fig. & De Not. 173, 307, 308, 309
Schlieben, H. J S3
Scilla brevifolia (Thunb.) Roem. & Schult. 121
Scrophulariaceae, Notes on the 176
Seaweeds SIO
Sesamothamnus guerichii S32
Sicyos angulata Berg 30
Simons, R. H S4
Simons, R. H. Notes on the species of
Zonaria in South Africa 195
Simons, R. H. Species of Plocamium on
the South African coast 183
Sphaerosicyos Hook. f. 82
meyeri Hook. f. 84
sphaericus (Drege. ex Naud.) Cogn 84
Sphaerosicyus Post. & Kuntze 82
SPOROBOLEAE Stapf 207, 219, 223, 224, 234
Sporobolus R. Br 217, 219
Staff 1964/65 S2
Stangeria eriopus {Kunze) Nash 429
Sterculiaceae, Notes on the 177
Stipa L. 201, 205, 206, 207, 208, 209, 210, 211,
212, 216, 217, 219, 222, 244, 299, 377
sect. Achnatherum 216
sect. Ptilagrostis 210, 216
sect. Schizachne Trin. & Riipr 236
sect. Tortilia Ellias 210
capensis Thunb 210, 211, 213, 217
dregeana Steud. 210, 211, 213, 216, 217, 232,
235
var. dregeana 215
var. elongata {Nees) Stapf . .2\0, 215, 216
elongata (Nees) Steud 216
extremi-orientalis 217
hookeri Stapf. 206
leucotricha Trin. & Rupr 209
namaquansis Pilg 211, 212, 309, 375
parvula Nees. . . .211, 212, 236, 242, 243, 299
pennata L 212
sibirica Law 217
rort/Vw Desf 210, 217, 219
vaseyii Scribn 217
Stipagrostis Nees 212, 225, 226, 227, 228, 230,
231, 233, 236, 299, 300,303,304,307, 308,
342, 351, 357, 359, 377, 379, 380, 389, 390, 392
sect. Anomala De Wint 309, 311, 377, 378
sect. Schistachne {Fig. & De Not.) De
Wint. 173, 228, 300, 307, 308, 309, 322, 378,
379
sect. Stipagrostis 228, 300, 309, 322, 378, 379
amabilis {Schweick.) De Wint. 237, 306, 309,
311, 324, 328, 330, 332, 378, 379
anomala De Wint. 211, 237, 309, 311, 312,
375, 377, 378
brevifolia {Nees) De Wint. 305, 310, 312, 322,
328, 338, 378, 379
capensis Nees 304, 355
ciliata {Desf.) De Wint. 235, 237, 305, 308,
309, 312, 315, 316, 320, 352, 378, 379, 391
var. capensis {Trin. & Rupr.) De
Wint 316
damarensis {Mez) De Wint. 310, 312, 324,
329, 378, 379
dinteri {Hack.) De Wint. 309, 312, 320, 357,
378, 379
dregeana Nees 304, 310, 312, 336, 344, 357,
378, 379, 390, 391
fastigiata {Hack.) De Wint. 305, 310, 312,
338, 378
garubensis {Pilg.) De Wint. 310, 312, 336,
343, 346, 378, 379
geminifolia Nees 304, 305, 310, 312, 341, 378
gonatostachys {Pilg.) De Wint. 310, 311, 312,
322, 353, 357, 378, 379
hermannii {Mez) De Wint. 173, 305, 308, 311,
313, 353, 369, 372, 373, 374, 378
hirtigluma {Steud.) De Wint. 305, 311, 313,
359, 360, 361, 367, 378
var. hirtigluma 363
var. patula {Hack.) De Wint 305, 365
var. pearsonii {Henr.) De Wint. 305, 306,
365
526
Page
hochstetteriana {Beck, ex Hack.) De
Wint. 232, 237, 306, 309, 312, 313, 378, 379
var. hochstetteriana 315
var. secalina (Henr.) De Wint 315
lanipes {Mez) De Wint. 306, 310, 312, 322,
355, 357, 373, 378, 379, 391
lutescens {Nees) De Wint. 310, 312, 324, 334,
378, 379
var. lutescens 336
var. marlothii {Hack.) De Wint 336
namaquensis {Nees) De Wint. 232, 306, 309,
311, 324, 325, 330, 332, 338, 375, 378, 379
namibensis De Wint. 173, 311, 313, 370, 378
obtusa {Del.) Nees 232, 237, 305, 306, 308,
311, 312, 322, 334, 353, 355, 377, 378, 379,
391
proxima {Stead.) De Wint. 309, 312, 322, 314,
378, 379
pungens {Desf.) De Wint 220, 306, 391
ramulosa De Wint... Hi, 310, 312, 333, 378
sabulicola (Pilg.) De Wint. 173, 306, 310,
311, 324, 328, 331, 334, 337, 378, 379
schaeferi {Mez) De Wint. 309, 312, 318, 377,
378, 379
subacaulis {Nees) De Wint. 305, 308, 311,
313, 370, 371, 373, 378
uniplumis {Licht.) De Wint. 173, 305, 306,
311, 312, 313, 342, 359, 362, 363, 367,
370, 378
var. intermedia {Schweick.) De Wint. 173,
306, 361
var. neesii {Trin. & Rupr.) De Wint. 305,
306, 360, 367
var. uniplumis 306, 358, 360, 365
zeyheri {Nees) De Wint. . .i\0, 346, 378, 379
subsp. barbata {Stapf) De Wint. . . 347
subsp. macropus (Nees) De Wint. . . 347, 348
subsp. sericans {Hack, apad Schinz)
De Wint 299, 347, 349
subsp. zeyheri 347
Stipeae and Aristideae, The South African 201
STIPEAE Nees 201, 205, 206, 207, 208, 209,
210, 212, 219, 222, 223, 224
Strachan, J. H S3
Streptachne R. Br 205, 206, 207, 210
Strey, R. G S3
Study Tours S26
Syncolostemon macranthus (Gaerke) Ashby 159
Systematic Botany Section
Staff SI
Review of work S6
Taylor, H. C S4, 5
Taylor, H. C. A Cycloptychis hunt S35
Thamnocarpus cornutus Kuetz 189
Thamnophora corallorhiza C. Ag 191
Thompson, Miss M. F S4
Timouria Roshev 205, 206
TIMOURIINAE Roshev 216
Tdlken, H. R S2
Toxanthera Hook, f 24
kwebensis N.E. Br 36
higardae N.E. Br 36
natalensis Hook, f 25, 36
Trachyandra 147
revoluta (L.) Kanth 148
thyrsoidea (Bak.) Oberm 148
Tricharis Sa\\%b 137
Page
Trichosanthes foetidissima Jacq 26
Trikeraia 206, 209, 210
Triodia R. Br 222, 380
TRIODIINAE Pilg 222
Tritonia watsonioides Bak 116
Trochomeria Hook. f. 85
sect. Eutrochomeria Cogn 85
sect. Heterosicyos {Welw. ex Benth. &
Hook. /.) Cogn 85
sect. Trochomeria 85
baumiana 86
brachypetala R.E. Fr 87
ddoWis, (^Sond.) Hook. f. 86, 87, 89
hookeri Harv 85, 90
macrocarpa (Sond.) Hook. f. 19, 86, 87, 88
nudiflora Burtt Davy 87, 88
pectinata (Sond.) Cogn 89, 91
var. subintegrifolia Cogn 90
rehmannii Cogn 91
rotundata Burtt Davy 90
sagittata {Harv. ex Sond.) Cogn. 21, 85, 86,
87
subintegrifolia (Cogn.) Burtt Davy. ... 90
vitifolia 86
wyleana (Sond.) Cogn 99
Troughton, Miss S. C S3
Typokedrostis Cogn 25
UNIOLEAE 209
Urachne Trin 205
Urginea I or at a Bak 130
Burch, ex Ker-Gawl 118
Uropetalum Burch, ex Ker-Gawl 122
ciliare Eckl. & Zeyh 122
crispum Burch 128
glaucum Burch, ex Ker-Gawl 130
hyacinthoides Spreng 121
longifolium Lindl 126
umbonatum Bak 132
viride (L.) Ker-Gawl 132
Vahrmeijer, G S6
Vahrmeijer, J S2
Van Breda, P. A. B S39
Van der Ende, A S6
Van der Helde, Mrs. P. W S3
Van der Merwe, Mrs. A S3
Van der Merwe, Dr. F. Z S65
Van der Merwe, P S39
Van der Walt, Miss A. J S3
Van der Westhuizen, Mrs. G. S. H S3
Van Hoepen, Mrs. E S2
Van Pletzen, J. S SI
Veld Reserve, Fauresmith S39
Veld Reserve, Worcester S39
Venter, Miss H. J S3
Verdoorn, Miss I. C S2, 27, 39, 65
Verdoorn, I. C. Dr. F. Z. van der Merwe. S65
Verdoorn, I. C. Notes on the Hippocratea-
ceae 114
Verdoorn, I. C. Notes on the Sterculiaceae 177
Visagie, A. F. J S39
Visitors S25
Visser, Miss C. H S2
Vogts, Mrs. M. M S39
Von Broembsen, H. H S5, 40
Voorendyk, Miss S S3, 54
Wahlenbergia monotropa Killick 164
Watsonia flavida H. Bolus 116
527
Page Page
latifolia N. E. Br. ex Oberm 115
meriana M/// 115
watsonioides (Bak.) Oberm 116
Wells, M. J S4, 5, 34
Wormser, Miss J. R S3
Zamia caffra Thunb 509
cycadifolia Jacq 439
horrida Jacq 467
longifolia (Jacq.) Lehm 474
Zehneria Endl 13
cerasiformis Stocks 12
debilis Sond 89
garcinii Sond 90
hederacea Sond 30, 32
macrocarpa Sond 88
obtusiloba Sond 27
pectinata Sond 90
scabra Sond 19
wyleana Sond... 89
Zonaria in South Africa, Notes on the
species of 195
Zonaria harveyana (Pappe ex Kuetz) Aresch
195, 196, 197, 199
subarticulata (Lamoiir.) Papenf. 196, 197
tournefortii (Lamoiir.) Mont 196, 197
Ziiccagnia Thunb 118
viridis (L.) Thunb 131
“ Zuccangia ” 118
Bothalia
Vol. VIII May, 1966 Supplement No.l
Contributions from the Botanical
Research Institute, South Africa,
including a review of its activities
during the period 1st July, 1964
to 30th June, 1965.
Edited by
L E. Codd, D.Sc.
NAVORSiNGSIKSTITL . TVIR P L A ^.TKUND E
BIBL 0 "EEK.
BOTANICAL RES
CATALOGUE NO. ...
L I B
botanical RE8
Headqu.\rters:
CeiJef
, T ruTE.
'ARCH INSTITUTE
Department of Agric iltural Technical Services
Republic of South Africa
P.O. Box 994, Pretoria
Assistant Chief
■Clerical Staff
Systematic Botany
Section
Economic Botany
Section
I
Botanical Survey
Section
Taxonomy of South
African flora, e.xcluding
Fungi
Applied botanical studies;
economic importance of
indigenous plants
Plant geographical and
ecological studies; quan-
titative ecology
National Herbarium,
P.O. Box 994,
Pretoria
Natal Herbarium,
Botanic Gardens Road,
Durban
Albany Museum Herbarium,
P.O. Box 101,
Grahamstovvn
McGregor Museum Herba-
rium, Kimberley
Stale Herbarium,
Private Bag 29,
Stellenbosch
Marine Algology Section, c/o
Department of Botany,
University of Cape Town,
P.O. Rondebosch
Pretoria National Botanic
Garden,
P.O. Box 994, Pretoria
Officer in Charge of Botanical
Survey, P.O. Box 994, Pre-
toria, with regional staff at:
Pietermaritzburg: c/o
Botany Department,
Natal University
Grahamstovvn: P.O. Box
101, Grahamstown
Middelburg, C.P. : Groot-
fontein College of Agri-
culture, Middelburg, C.P.
Stellenbosch: Botanic
Station, Private Bag 29
Stellenbosch
Bothalia
Vol. VIII May, 1966 Supplement No, 1
Contributions from the Botanical Research Institute, South Africa, including
a review of its activities during the period 1st July, 1964 to 30th June, 1965
Edited by L. E. Codd, D.Sc.
CONTENTS
Front Cover. — Ceropegia ampliata, from Steelpoort Valley, Transvaal.
Back Cover. — Stultitia liardyi from Wylliespoort, Transvaal.
Page
1. Botanical Research Institute, 1.7.64 to 30.6.65 1
Staff and Activities 1
Review of Work —
Systematic Botany Section 5
Economic Botany 10
Botanic Garden 11
Botanical Survey Section 15
Visitors 18
Congresses and Study Tours 18
Collecting Expeditions 22
Publications 31
Personalia 32
2. Elowers on the New Decimal Coins. Cythna Letty 35
3. Field Work is not All Fun. R. A. Dyer 43
4. Plants Which Cause “ Gousiekte ” Poisoning. L. E. Codd and S. Voorendyk 47
5. Contemporary Botanists and Botanical Collectors —
Dr. F. Z. van der Merwe. I. C. Verdoorn 59
Mr. L. C. Leach. L. E. Codd 61
6. Vegetative Propagation of Proteaceae and Ericaceae. J. Admiraal 65
Fig 1. — The National Herbarium, Pretoria.
BOTANICAL RESEARCH INSTITUTE
1st July, 1964 — 30th June, 1965
Staff and Activities
L. E. W. CODD, D.Sc., Chief of the Institute and Director of the Botanical Survey
of South Africa; taxonomic studies, particularly in Labiatae, Kniphofia and Boerhavia.
B. DE WINTER, D.Sc., Assistant Chief of the Institute and Curator of the National
Herbarium, Pretoria; taxonomic studies, particularly in Gramineae, Hermannia and trees
of Southern Africa.
CLERICAL SECTION
J. S. VAN Pletzen, Administrative Officer.
S. W. DU Ton, Senior Clerk.
T. R. Jenkinson, Temporary Clerk.
Mrs. T. M. Crefeield, Records Clerk.
Mrs. M. B. Coetzee, Typist.
LIBRARY
Staff supplied by the Department of
Education, Arts and Science
Miss M. D. Gunn, Librarian.
Miss E. F. E. Hagedorn, Librarian.
SYSTEMATIC BOTANY SECTION
Chief Professional Officer
Vacant
Principal Professional Officer
O. A. Leistner, D.Sc., Liaison Officer,
stationed at the Herbarium, Royal Botanic
Gardens, Kew; miscellaneous botanical
studies in response to enquiries from the
Institute; Podocarpus and Malvaceae.
National Herbarium, Pretoria
Senior Professional Officers
J. G. Anderson, B.Sc. ; Gramineae, parti-
cularly Andropogon.
R. A. Dyer, D.Sc. (temporary: formerly
Chief of the Institute); Encephalartos,
Stapelieae and Ceropegia.
W. Marais, M.Sc., stationed at the Royal
Botanic Garden, Kew; revision of the
Cruciferae for the Flora of Southern Africa.
Mrs. a. a. Obermeyer-Mauve, M.Sc.;
Helobiae, Liliiflorae, with particular reference
to Ornithogalum and Dietes.
Miss I. C. Verdoorn (temporary: formerly
Curator of the National Herbarium);
Melhania, Dombeya, Criniim and Encepha-
lartos.
Professional Officer
Mrs. E. van Hoepen, M. Sc. ; co-ordination
of general plant identifications and infor-
mation services; identification of Scrophula-
riaceae.
Assistant Professional Officers
Miss C. de Groot, B.Sc.; identification of
Compositae.
J. P. Jessop, M.Sc.; Asparagus; general
identifications, particularly Liliiflorae and
Restionaceae; part editor of “ Forum
Botanicum ”.
J. Landman, B.Sc.; seconded to Pretoria
University for Hons, degree.
Miss J. A. Marsh, B.Sc. (Hons.); revision
of Widdringtonia.
H. R. Tolken, B.Sc. (Hons.); revision of
Arthrocnemum and Salicornia; identifications
of Chenopodiaceae and related families;
identification of plants from South West
Africa.
J. Vahrmeijer, B.Sc.; general identifica-
tions, particularly woody plants; plant collec-
ting in northern Zululand.
Miss C. H. Visser, M.Sc.; anatomy of
plants; resigned January, 1965.
2
Senior Technician
Senior Technical Assistants
J. W. Reyburn; photographer.
Artists
Mrs. C. L. Forssman.
Mrs. a. van der Merwe, B.A. (Fine Arts).
H. J. Schlieben; preparation of specimens;
distribution of duplicates.
Miss A. J. van der Walt, assistant in
Library.
Technical Assistants (Plant identification)
Technicians
Mrs. G. S. H. van der Westhuizen,
B.Sc.; general Identifications, particularly
Labiatae and Verbenaceae.
Miss S. Voorendyk, B.Sc.; general identi-
fications, particularly Ericaceae and
poisonous plants; attends to visiting groups
of scholars and students.
W. S. Bezuidenhout, Learner Technician.
Miss A. S. C. Calitz; Gramineae.
Mrs. L. D. du Toit; cultivated plants.
Miss M. H. Gerber; Rubiaceae; organi-
sing supplies of seed and plant material.
Miss E. Y. Johansson; ferns, mosses and
liverworts.
Miss C. M. Lemmer; Leguminosae and
Geraniaceae; assisting with alkaloid survey.
Miss C. Liebenberg; Anacardiaceae,
Celastraceae and related families.
Fig. 2. — The National Herbarium, Pretoria, in its setting below the Union Buildings on Meintjies Kop.
3
Technical Assistants (Curatorial)
Mrs. U. Jellinek; Mtss M. Kat; Miss
E. Liebenberg; J. H. Strachan; Mrs. P.
W. VAN DER Helde; Miss H. J. Venter;
Miss J. R. Wormser.
Durban Botanic Station
Senior Technician: R. G. Strey, Officer in
Charge and Curator of the Natal Herbarium;
identification and collecting of Natal plants.
Technical Assistant: Miss R. A. Briggs.
Clerical Assistant: Mrs. A. M. C. Breyten-
BACH.
Grahamstown Botanic Station
Senior Professional Officer
M. J. Wells, M.Sc., Officer in Charge and
Curator of the Albany Museum Herbarium
(see also under Botanical Survey Section).
Senior Technical Assistant: Miss G. V. Brit-
ten.
Technical Assistant: Miss S. C. Troughton,
B.Sc.
Assistant Professional Officer (part-time):
Mrs. E. Brink, B.Sc.
Kimberley Botanic Station
Technical Assistant: Miss W. A. Herbert;
curatorial supervision of McGregor Museum
Herbarium.
Marine Algology Section, Cape Town
Senior Professional Officer
R. H. Simons, M.Sc.; collection and
taxonomy of marine algae; Galaxaura, Sar-
gassum, Ceramium.
Stellenbosch Botanic Station
Senior Professional Officer
H. C. Taylor, B.Sc. (For.), B.Sc. (Hons.),
Officer in Charge (see also under Botanical
Survey Section).
Assistant Professional Officer
E. G. Oliver, M.Sc., Curator of State
Herbarium, Stellenbosch; revision of
Acrostemon\ identification of S. W. Cape
plants. Part editor of “ Forum Botanicum. ”
Technician
Miss M. P. Thompson, B.Sc.; identification
of S.W. Cape plants; revision of Spiloxene.
Technical Assistant: Miss E. J. Loubser.
BOTANICAL SURVEY SECTION
Chief Professional Officer
D. J. B. Killick, Ph.D., Officer in Charge
of Botanical Survey Section; mountain
ecology; revision of Myrica.
Principal Professional Officers
J. P. H. Acocks, M.Sc., Middelburg, C.P. ;
revision of Veld Types of South Africa.
D. Edwards, Ph.D., study tour in Europe
and United States; survey of Tugela Basin.
H. H. VON Broembsen, M.Sc., Pretoria;
quantitative ecology.
Senior Professional Officers
H. C. Taylor, B.Sc. (For.), B.Sc. (Hons.),
Officer in Charge of Botanic Station, Stellen-
bosch; survey of Cape Peninsula.
M. J. Wells, M.Sc., Officer in Charge of
Botanic Station, Grahamstown; survey of
forests of the Eastern Cape; palaeo-ethno-
botanical studies.
Professional Officers
P. J. Grobler, M.Sc., Stellenbosch; survey
of Oudebos area, Kogelberg Forest Reserve.
J. C. Scheepers, B.Sc., Pretoria; final
stage of survey of Westfalia Estates, Duiwels-
kloof.
Assistant Professional Officer
E. J. Moll, B.Sc. (Hons.), Pietermaritz-
burg; survey of upper Umgeni River catch-
ment.
Technician
P. J. Muller, Pretoria; assistant to
Mr. von Broembsen in quantitative ecology
unit.
0015958—2
4
PRETORIA NATIONAL BOTANIC
GARDEN
Superintendents of Gardens Gr. I
J. Admiraal, gardener in charge; land-
scape planning of garden, etc. ; collection of
plants, particularly Erica spp.
A. VAN DER Ende, Vrede Huis Garden.
Technician
D. S. Hardy; collection and maintenance
of succulent plants.
Superintendent of Gardens Gr. II.
J. Radersma; maintenance of nursery.
Technical Assistants: J. du T. Oliver;
C. L. Keytel, orchid collection and nursery.
Learner Gardener: G. Vahrmeijer.
Fig. 3. — Staff photo, September, 1963. Standing at back: D. J. B. Killick, G. J. Anderson, J. C. Scheepers,
O. A. Leistner, B. de Winter, R. G. Strey, M. J. Wells, S. W. du Toit, A. F. J. Visagie,
Miss I. C. Verdoorn, H. H. von Broembsen, J. P. H. Acocks, P. J. Grobler, H. R. Tolken,
Mrs. C. L. Forssman, E. G. Oliver, L. E. Codd, Mrs. El. G. Codd, P. J. Muller, P. van der Merwe,
Mrs. M. M. Vogts. Kneeling and sitting: Miss C. Lemmer, Mrs. P. van Wyk, Miss A. van den
Berg, Mrs. E. van Eloepen, Miss Evan Vuuren, Miss H. Gilliland, Miss R. Venter, Miss E. van
Rensburg, Miss R. van Wyk, Miss E. Liebenberg, Mrs. T. Creffield, Miss H. Gerber, Mrs. A. A.
Mauve, Miss C. Liebenberg, Miss^M. Bennett, Mr. B. Vogts, Mrs. M. Coetzee, Miss J. Prinsen,
Miss C. de Groot, D. Edwards. (Photo by R. H. Simons).
5
Review of Work
SYSTEMATIC BOTANY SECTION
Flowering Plants
Flora of Southern Africa
Good progress was made with Vol. 1 of
the Flora. This will include: five families of
Gymnosperms, which are being revised by
Dr. Dyer and Miss Verdoorn {Encephalartos
and Stangeria), Miss Marsh {Widdringtouia),
Dr. (Podocarpus) and Miss Verdoorn
{Welwitschia); the Typhaceae by Mr.
Anderson, and the nine families of water-
plants belonging to the order Helobiae
(Zosteraceae, Potamogetonaceae, Ruppiaceae
Zannichelliaceae, Najadaceae, Aponogeto-
naceae, Juncaginaceae, Alismataceae and
Hydrocharitaceae) by Mrs. Mauve.
Encephalartos presented a series of pro-
blems. Typification of the species was
carefully investigated and led to changes of
accepted nomenclature in certain cases. The
determination of species limits called for
several field excursions, particularly in the
eastern Cape Province, when Dr. Dyer was
accompanied by Mr. Wells, and in the south-
eastern Transvaal and southern Lebombo
range, when he was accompanied by Miss
Verdoorn and Mr. Admiraal. A highlight
of the investigation was the discovery of an
undescribed species in the eastern Transvaal,
on the dolomite formation adjoining the
Olifants River. It was subsequently
described as E. inopinus R. A. Dyer.
Water-plants are not well represented in
South African herbaria and our collectors
were requested to keep a special look-out for
them. Fortunately the heavy early summer
rains led to good conditions for aquatics and
produced a number of new records, most of
which were investigated by Mrs. Mauve with
the help of the National Herbarium staff.
In October, 1964, while examining an
infestation of Lemna minor in the Crocodile
River near Brits, she collected the Indian
species Spirodela oligorrhiza, already present
in our Herbarium but unidentified. In
February 1965, in a farm dam near Brits, she
collected Heterantha callifolia, a species not
represented in our Herbarium, although it is
including in Phillips’ Genera of South African
Flowering Plants. A few days later, a study
of aquatic plants in a farm dam near
Petronella led to the conclusion that Apono-
geton junceus and A. rehmannii are forms of
one species; here Najas graminea was
collected for the first time in South Africa,
What makes these gatherings all the more
remarkable is that they were recorded within
25 miles of Pretoria. With Mr. Downing she
visited Ntabamhlope (Estcourt District) in
April, 1965, where she collected Utricularia
vulgaris, the first record for Natal. In
northern Zululand, Messrs. Tolken and
Vahrmeijer collected two minute water-
plants, Wolffiella denticulata and W. xvelwit-
schii. The former was orginally described
from a few individuals adhering to a specimen
of Lemna minor collected in Natal by Krauss
about 1839, and has been recorded once since
then, but is not present in South African
herbaria; the latter is a new record for South
Africa.
Vol. 21 of the Flora is nearing completion.
Miss Verdoorn worked intensively on
Melhania and Domheya and has cleared up
most of the problems in these two genera,
studying them, where possible, in the field.
Arising from these studies were the recog-
nition of the Rhodesian species, Melhania
randii, in the Barberton District and a new
species allied to it in the Wolkberg. The
Dombeya rotundifolia complex presented
difficulties in the herbarium, but yielded to
field investigation, resulting in the recogni-
tion of an undescribed species in the eastern
Transvaal near Penge Mine and Erasmus
6
Pass: although florally very close to D.
rotimdifolia, it differs markedly in habit and
ecology. Work on the largest genus in this
volume, namely Hermannia, was interrupted
by the death of Mr. Pillans. The revision is
being continued by Dr. de Winter.
Another volume in which progress can be
reported is Vol. 13, for which a revision of the
Cruciferae was completed by Mr. Marais.
In this family 35 genera and some 140 species
are recognized, the largest being Heliophila
with 70 species. Mr. Marais had the oppor-
tunity of studying some of the problems in
the field with Mr. Taylor while on holiday in
South Africa in November/December 1964.
However, more material is required of
several species which are comparatively
unknown. Included in the family are a
number of cosmopolitan weeds which were
in need of classification for the revision of
the Weed Book (see p. 10).
Fig. 4. — Ceropegia estelleana R. A. Dyer, a new
species collected in the Eastern Cape by
Mrs. Estelle Bayliss, X 1.
Other Revisions
Miscellaneous revisions and descriptions of
new species were published in Bolhalia Vol.
8, parts 2 and 3. Part 2 contained revisions
of Dipcadi and Lagarosiphon by Mrs. Mauve
and of Orthosiphon by Dr. Codd. Part 3
consisted of the thesis presented by Dr. de
Winter for his Doctorate and dealt with
anatomical, cytological and taxonomic studies
in the important grass genera belonging to the
tribes Stipeae and Aristideae.
Revisions completed and prepared for
publication during the year deal with Andro-
pogon (Mr. Anderson), Asparagus (Mr.
Jessop) and Boerhavia (Dr. Codd). Studies
on Kniphofia and Plectranthus (Dr. Codd),
Ornithogahim and Dietes (Mrs. Mauve), Scil-
leae (Mr. Jessop), Crinum (Miss Verdoorn),
Stapelieae and Ceropegia (Dr. Dyer), Myrica
(Dr. Killick), Arthrocnemum and Salicornia
(Mr. Tolken), Acrostemon (Mr. Oliver) and
Spiloxene (Miss Thompson) are continuing.
The appointment of a second artist,
Mrs. Aleida van der Merwe (nee De Bruyn),
has resulted in a satisfactory increase in the
production of paintings for the series
Flowering Plants of Africa. As the Govern-
ment Printer could no longer continue
printing this series, it became necessary to
resort to the unsatisfactory procedure of
calling for tenders for each section (two parts)
of 20 plates, as completed, with a resultant
variation in quality of printing. Vol. 36,
parts 3 and 4, were published, while Vol. 37,
parts 1 and 2, planned to appear in June,
1965, were unavoidably delayed; 20 plates
are available for parts 3 and 4 to complete
this volume. In the process of writing the
botanical descriptions accompanying these
plates, it is necessary to undertake a careful
review of the species depicted and its allies.
Our supply of plates was further augmented
by the purchase of 12 plates of Cape flora
by Fay Anderson.
The well-known and valuable Genera of
South African Flowering Plants by E. P.
Phillips has been out of print for some time.
Since 1951, when the second edition appeared,
many new generic and even family records
have been discovered within our boundaries.
7
These, together with changes in nomenclature
due to modern studies, necessitate the pre-
paration of a revised edition of the text.
The revision is to be initiated by Dr. Dyer
with the help of specialists in certain families.
Ferns, Mosses and Seaweeds.
No research on ferns, mosses or liverworts
has been undertaken by the Institute for many
years. Miss Johansson, curatorial assistant
in the National Herbarium, compiled a
popular article on the ferns of the Transvaal,
including 29 illustrations by Mr. Louis Steyn,
artist of the Transvaal Nature Conservation
Department. Salvinia auriculata, which has
caused such concern at Kariba in Rhodesia,
has been reported as being widespread in the
eastern part of the Caprivi Strip, and has
been recorded from the eastern and south
eastern parts of South Africa. It is often
cultivated in small aquaria and fish-ponds.
Two papers on marine algae were published
by Mr. Simons, dealing with the species of
Plocamium (Gigartinales) and Zomiria (Dic-
tyotales) along the South African coast.
A revision of Ceramium was submitted for
publication. The genera Galaxaura (Nema-
lionales) and Sargassiim (Fucales) were given
special attention and a new species of
Pocockiella of the Dictyotales was described.
A report on the distribution of South African
“ sea bamboos ” (Laminariales) was prepared
and recommendations were made to com-
mercial firms on the utilization of certain
seaweeds. A special report on the export of
seaweed was drawn up. A collection of
seaweeds from the Verna Seamount, about
500 miles west of the Cape, was received and
they appear to be most interesting; two of
those studied so far belong to species
previously considered to be endemic to our
south coast.
New and Interesting Records
Preliminary surveys in the northern part of
Zululand, in which Messrs. Vahrmeijer,
Tolken and Strey have taken part, have
produced a number of new species records,
some of which belong to genera recorded for
the first time within our boundaries. These
include Ballya zebrina, Coffea racemosa.
Neurotheca sclilechteri, Clerodendnun viola-
ceum, Solamim zanziharense var. vagans,
Ipomoea urbaniana, Caesalpinia bonduc. Ficus
mucosa and a species of Sophora. From other
parts of Natal came Eriosemopsis subaniso-
phylla, recorded only once before, and three
gatherings of Prunella vulgaris from a
restricted area in the Natal Midlands.
Although it appears to grow naturally in
a vlei, it is probably a recent introduction of
this widely distributed species. New records
of water-plants have been mentioned on p. 5.
Altogether, 22 new species of South African
plants were described during the year by the
staff of the Institute.
Curatorial Services
The curatorial work in connection with the
herbaria, as well as the identification of plant
specimens and the dissemination of infor-
mation on indigenous plants are routine
Fig. 5. — Dietes biitcheriana, a rare species from
Inanda, Natal (R. G. Strey, 4875), x|.
8
services undertaken by the Institute which
occupy a large proportion of the time of several
professional officers and other staff members.
The following table summarizes the more
important work of this nature carried out
in the National Herbarium and at the
Regional Herbaria.
Identifications
Specimens received at the National Her-
barium included the following contributions
from members of the staff of the Institute:
Drs. Codd (45, N.E. Free State and E.
Transvaal), De Winter (53, Pretoria and
Bulawayo), Dyer (71, mostly Encephalartos
from E. Cape and E. Transvaal), Killick (12,
Wolkberg, Transvaal), Messrs. Taylor (1703,
S.W. Cape), Strey (1008, Natal), Moll (826,
Natal Midlands), Tolken and Hardy (450,
S.W. Africa), Jessop (94, mostly Asparagus),
Vahrmeijer {1260, mainly Zululand),
Schlieben (400, South Africa and South West
Africa), Hardy (208, Namaqualand), Mrs.
Mauve (90, Transvaal and Natal), Miss Marsh
(about 200, N.W. Transvaal), while Miss
Verdoorn’s collections were mainly incor-
porated under the numbers of other
collectors.
Numerous specimens were identified for
regional officers of the Department, other
Institutes, students and members of the
public. The majority of these were not
incorporated in the Herbarium. Contribu-
tions which enriched the Herbarium were
received from the following persons and are
gratefully acknowledged : —
Col. R. D. Bayliss, who travels extensively
in South Africa on behalf of B.M.C.,
brought in a number of specimens (main-
ly from the Eastern Cape Province).
Prof. R. H. Compton, now residing at
Mbabane, continued with his survey of
Swaziland (specimens received from his
Swaziland collections).
Mr. R. B. Copley (N. Transvaal).
Mr. N. J. Devenish, farming in the moun-
tains near Wakkerstroom, continued
collecting specimens occurring on his
farms.
Miss M. D. Gemmel, acting Curator of the
University of O.F.S. Herbarium, sent
in some of her specimens from Bloem-
fontein.
Dr. W. B. G. Jacobsen, geologist, collected
in Rhodesia, the Transvaal and Natal.
Mrs. M. Jacobsz, whose husband farms at
Rensburgskop near Swinburne, O.F.S.,
collected specimens on her farm and in
the neighbourhood and has been the
main driving force behind the proposal
to start a Botanic Garden near
Harrismith.
Mrs. A. Jacot-Guillarmod, Lecturer in
Botany at Rhodes University, has con-
tributed further specimens from her
Basutoland collections.
Mr. J. Lavranos, Insurance Specialist in
Johannesburg, sent in some plants
collected in S.W. Arabia as well as
isotypes of certain new species he
described, and has started a collection
of plants growing near Johannesburg.
Mr. L. C. Leach (see p. 61) supplied
specimens collected at Nelspruit, in
Rhodesia and Mozambique, including
type material of his newly described
species of Euphorbia.
9
Dr. W. J. Louw of Potchefstroom Univer- Mr. H. J. T. Venter of the Botany Depart-
sity Botany Department, contributed ment. University College, Ngoye, is
specimens collected near Potchefstroom making a study of the surrounding vege-
and in the Waterberg District. tation and has contributed a few speci-
mens from his collection.
Dr. A. O. D. Mogg, Witwatersrand Univer-
sity Botany Department, continued his
survey of Inhaca Island and also ,
collected in the Waterberg District.
Natal Parks, Game and Fish Depart-
ment— specimens received mainly from
Hluhluwe Game Reserve.
Mr. A. Renny, farming in northern Trans-
vaal, sent in further specimens from the
Harmony Block.
Mr. J. Repton, Assistant Director of the
Pretoria Parks and Recreation Depart-
ment, brought in further collections of
indigenous and cultivated plants.
Mr. B. R. Roberts, Senior Lecturer in
Pasture Research at the University of
O.F.S., Bloemfontein, collected in eastern
O.F.S. and at Queenstown.
Mr. D. Shearing, sheep farmer in Fraser-
burg and Beaufort West Districts, is
trying out the grazing system advocated
by Mr. Acocks, whereby intensive
grazing is applied to the veld for short
periods, followed by relatively long rests;
he is getting to know the more important
plants on his farm and will extend his
study to an inventory of all plants.
Mr. J. L. Sidey, for some years on the staff
of the Natal Tanning Extract Co., now
resident in Pietermaritzburg, has con-
tinued his interest in general collecting.
Mr. W. R. Trauseld, Ranger with the Natal
Parks, Game and Fish Department, is
active in photographing and collecting
the more attractive plants of the Drakens-
berg.
Dr. B. van Ginkel collects in the neigh-
bourhood of Bethlehem, O.F.S.
Prof. W. J. van der Merwe and Mr. D. R. J.
van Vuuren of the Botany Department,
University College. Turfloop, have sent
in small collections for naming.
Almost 4,000 specimens were sent on loan,
the main recipients being as follows:—
Royal Botanic Gardens, Kew: 207 Podo-
carpus and 22 Melhania (Dr. Leistner);
170 Cruciferae (Mr. Marais); 749 Acacia
and Albizia (Mr. Brenan).
Government Herbarium, Salisbury: 164
Compositae (Dr. Wild).
British Museum (Natural History): 99
Xyris (Dr. Lewis).
Botanische Museum der Universitat,
Zurich: 18 Thesium.
Botanical Museum, Lund: 202 Penaeaceae
(Dr. Dahlgren).
Bolus Herbarium, Cape Town: 162 Saty-
rhim (Dr. A. Hall); 156 Proteaeceae,
minor genera (Mr. J. Rourke).
Forest Herbarium, Oxford: 603 Coin-
brelum (Dr. Exell).
Natal University, Pietermaritzburg: 168
Acacia (Mr. Ross); 78 Fimbristylis (Dr.
Gordon-Gray); 614 Streptocarpus (Mrs.
O. Hilliard).
Eaboratorium voor Plantensistematiek,
Wageningen: 81 Oligomeris, Reseda and
Liparia.
Stanford University, California: \22 Liid-
wigia and Gaura (Dr. Raven).
Komarov Academy of Science, Leningrad :
118 Osmunda regalis.
Botanical Museum, Utrecht : 94 Corrigiola
and Hernaria.
10
When revisions are carried out at the
National Herbarium it is the policy to see as
much of the material in South African
Herbaria as possible and also to obtain types
on loan or, where this is not permissible,
photographs of types from overseas Herbaria.
A record of these loans is not easily obtainable
for the year under review, but a system of
recording this information is being intro-
duced.
Exchanges
The exchanges indicated below appear to
be to the disadvantage of our Institute but it
should be noted that many of our duplicates
are sent to overseas Herbaria in exchange for
microfilm of rare literature and in return for
expenses involved in supplying photographs
of type specimens.
Exchanges of Duplicates
ECONOMIC BOTANY
A handbook dealing with the more
common weeds of South Africa was submitted
for publication. It is provided with 207
illustrations and deals mainly with weeds of
cultivation, gardens and waste places.
A new parasite on sugar-cane in Swaziland
was recorded by Prof. R. H. Compton. It
proved to be Thesium resedoides and at first
caused considerable concern, but tests proved
that it could be controlled by 2-4-D.
The identification of scraps of material
became an important matter in con-
nection with a number of investigations.
One of these, concerned with the presence of a
poisonous adulterant in a cattle-feed mixture,
was satisfactorily solved. Other cases are
concerned with material ingested during
grazing trials by livestock, while a few identi-
fications were of importance in court cases.
The resignation of Miss Visser, who had made
a particular study of plant anatomy, caused a
temporary set-back, but Mr. Tolken, with the
assistance of a Technician and a Technical
Assistant, is continuing with the project.
BOTANIC GARDEN
Land and Buildings
The acquisition from Pretoria University
last year of the area on which the Botanic
Garden was originally started has made it
possible to proceed with the planning of the
proposed new building to accommodate the
Botanical Research Institute. A plan was
prepared and received preliminary Treasury
approval. It is now with the architect for the
final details. Owing to the decision to place
a national road across the Garden, a new
site for the building was selected.
A further two morgen of land were
purchased, consisting of the two plots 3/F
and Rem/F, lying between the northern
boundary of the Garden and the Silverton
Road. This area, although small, is an
important acquisition because it lies in front
11
of the main succulent collection which is one
of the most attractive parts of the Garden.
A start was made with the removal of old
buildings and pine trees on these plots.
Cape Flora
The Protea and Erica collection recovered
satisfactorily after the exceptionally severe
winter and additional plantings were made in
this section. An experiment was initiated in
conjunction with Mrs. Vogts, involving the
sowing of over 50 samples of Proteaceae seeds
direct in the soil and in nursery tins.
Considerable success was achieved in the
vegetative propagation of Proteaceae and
Ericaceae, and the results are reported in this
publication. Through the efforts of Mr.
Admiraal, Superintendent of the Garden,
assisted by Dr. van Zyl of Pretoria, Dr. P. du
Toit of the Cape and Mr. P. A. B. van Breda,
Officer in Charge of the Worcester Veld
Reserve, the Erica collection was increased
to 127 species in cultivation.
Fig. 6. — Aerial view of the Pretoria National Botanic Garden, situated about six miles east of Pretoria.
Roughly triangular in shape, the southern boundary adjoins the east-west road to Brummeria and
is approximately a mile long. At its widest (eastern) boundary the Garden is three-quarters of
a mile wide. The third side of the triangle is formed by the Silverton road, along which there are
a number of privately owned properties. Some of these have been acquired for the garden.
At A, the warm north facing slope is planted with Aloes and other large succulent plants; a collect ion
of Cycads occupies the cool, southern slope at B; a section on the open Hats at C is devoted to the
Cape flora; while the nursery, glasshouses and garden offices are situated at D.
12
Garden Acquisitions
Further expeditions were made to collect
plants for the Garden. Areas visited by
Pretoria staff members included Namaqua-
land, south-western Cape, South West Africa,
north-eastern Free State and northern Zulu-
land. Contributions were also received from
Botanical Survey Officers and members of the
public, whose assistance is acknowledged
below.
In this way over 1,500 gatherings were
added to the Garden, including a large
number of miscellaneous trees and shrubs,
among which several interesting and unusual
species came from Dr. Ritchken of Salisbury,
Rhodesia. Attention was mainly concen-
trated on scientific collections of certain
groups of plants. The Ericas and Proteas
have already been mentioned. In addition,
the majority of South African species of the
following genera have been brought together
in cultivation : Coleus, Plectranthus, Encepha-
lartos, Zanledeschia, Crinum, Kniphofia and
Dietes. These collections are of particular
value to the herbarium workers undertaking
revisions of these groups.
As before, the collection of African, Ara-
bian and Madagascar succulent plants con-
tinued to thrive under the expert care of
Mr. D. S. Hardy. Additional plants from
Madagascar were received from Prof. Rauh
of Heidelberg University, Germany, and of
Arabian plants from Mr. John Lavranos of
Johannesburg. Mr. Hardy himself collected
about 140 contributions for the Garden,
about 30 of which were succulent plants,
including several interesting and possibly new
species. He investigated a new Aloe species
collected near Muden by Mr. G. J. Prinsloo
of Pretoria. It was figured and described as
Aloe prinslooi Verdoorn & Hardy. Al-
together, 41 gatherings of Aloe spp. were
added, including donations of A. hildebrandtii
and A. schelpei from Dr. G. W. Reynolds.
Succulent plants were also donated by the
following persons to whom our thanks are
recorded; Dr. van der Merwe of the C.S.I.R.
Fig. 7. — View of the Aloe garden at A in the aerial photo.
13
(mostly from S.W. Africa), Dr. W. J. Louw
of Potchefstroom, Col. R. D. Bayliss (E.
Cape), Mr. J. J. Minnaar (Cape), Mr. A. J.
Joubert of Stellenbosch (Little Karoo),
Mr. Harry Hall of Kirstenbosch and Mr. J.
van Zanten.
Plants were collected for the Garden by
the following members of the Institute’s
staff: Drs. Codd (22, including a new species
and four other species of Kuiphofia), Dyer
(24, including Encephalartos inopinus R. A.
Dyer, Crinum and Euphorbia spp.), De Winter
(5), Killick (5); Messrs. Admiraal (56, mainly
Ericas and Proteaceae from the Cape and
truncheons of a number of Convniphora
spp. from N. Transvaal), Scheepers (6),
Taylor (14), Oliver (17), Wells (13), Tolken
(15), Jessop (2), Moll (2), Vahrmeijer (121,
including many interesting trees and bulbs
from northern Zululand), Hardy (115, non-
SLieculents), Radersma (8), Keytel (260,
mainly forest herbs, orehids, etc.) and Strey
(24, of which several were Dietes spp.
including a new species from Jozini Dam,
and a new generic record for South Africa,
Ballya zebrina).
The following donations of plants or seeds
are gratefully acknowledged: Col. R. D.
Bayliss (32, mainly E. Cape), Dr. M. de
Villiers (3, Cape plants), Mr. F. During (12,
sueculents from E. Transvaal), Dr. P. du Eoit
(22, species of Erica and 3 Proteaceae),
Mr. Gilliland (15, Lydenburg), Mr. J. Hall
(9, Aloe spp., Crinum acaule), Mrs. H. J.
Herbst (5, Natal), Mrs. M. L. Jacobsz (7,
eastern O.F.S.), Mrs. E. Jenkins (6, Pearston),
Dr. J. D. M. Keet (2, ineluding a white
Agapanthus inapertus), Mr. J. Kloppers (6,
mainly Liliiflorae), Mr. Koch (6, Lydenburg),
Mr. J. Lavranos (40, including interesting
plants and seeds from Aden Protectorate),
Mr. L. C. Leach (21, including 5 gatherings of
Dietes from Mozambique), Mr. E. Lubbe (8,
orchids from Magoebaskloof) ,Mr. Van Oudts-
hoorn (19, mainly Ericas, also Ocluiaoconnori),
Mr. L. N. Schroeder (22, including seeds of low-
veldtrees) and Dr. Van Zyl (18, spp. of Erica).
Fig. 8. — View of the section at C in the aerial photo, devoted to members of the Cape flora, such as
Proteaceae and Ericas.
14
Fig. 1 1 — The Cycad group at B in the aerial photo (Fig. 6),
15
BOTANICAL SURVEY SECTION
Vegetation Surveys
Tugela Basin. — The final report of this
survey, carried out by Dr. Edwards for the
Natal Town and Regional Planning Com-
mission, is in the press and will appear as
Botanical Survey Memoir No. 36. The
report, which describes in considerable detail
the vegetation and environment of the
1 1 ,200 sq. mile catchment of the T ugela River,
consists of several hundred pages and
includes a 1 :250,000 vegetation map in
colour. It will assist greatly in the assessment
of land-use potential in the Tugela Basin.
Upper Umgeni Catchment. — This survey
represents the first phase of the “ Three
Rivers Survey ” being undertaken for the
Natal Town and Regional Planning Com-
mission on similar lines to and for the same
purpose as the Tugela Basin Survey. It will
eventually be expanded to include the catch-
ments of the Umgeni, Umlaas and Illovo
Rivers, and Mr. Moll, the responsible officer,
has completed the field work of the pilot
survey and is well advanced in writing up this
preliminary report.
Southern Kalahari. — Dr. Leistner com-
pleted the survey of 48,000 sq. miles of the
Southern Kalahari last year and the final
report is at present being edited with a view
to publication as a Botanical Survey Memoir.
Oudebos-Kogelberg Reserve. — An area of
85 morgen of fynbos vegetation in the
Kogelberg Reserve, Caledon District, was
investigated botanically by Mr. Grobler.
Sixteen stands comprising four plant associa-
tions were delimited and 286 species recorded.
The relation of the plant communities (and
of some individual species) to the factors of
the environment was studied. Besides con-
tributing much to our knowledge of a fairly
unknown flora, such studies yield basic
information about the ecological require-
ments of certain Protea species — information
which might be useful in the commercial
cultivation of Proteas.
Nature Reserves. — Minor surveys of two
nature reserves were completed during the
year. A report by Mr. Wells on the Jack
Scott Nature Reserve near Krugersdorp was
published in Fauna and Flora No. 15, 1964,
and a guide to the trees and shrubs of the
Willem Pretorius Game Reserve by Dr.
Leistner is to be published shortly by the
Orange Free State Administration.
Veld Types of South Africa. — Mr. Acocks
continued work on a revised edition of his
Veld Types of South Africa. The 1953
edition is now almost out of print, conse-
quently the need for a revised edition has
become urgent. Field work, intended to fill
gaps in our knowledge of the various veld
types, was concentrated in the Orange Free
State, the Transvaal, the Great Karoo and the
interior of the S.W. Cape. Eighty-six new
samples of veld were examined, yielding over
13,000 new distribution records to bring the
total up to over 329,000. The 3,046 samples
of veld so far examined were numbered,
classified, catalogued and indexed.
In the course of his travels Mr. Acocks
encountered some glaring examples of veld
deterioration. For example, he found that a
disastrous change has come over the veld
overlying the sandstone formation in parts
of the eastern Orange Free State in the past
20 years. Themeda triandra (Rooigras) has
disappeared as dominant: the valleys look
white with Eragrostis plana (Taaipol) and
other species, while the mountains look black
with Aster filifolius (Draaibossie) and other
weedy shrubs. In the north-eastern Orange
Free State, Aristida junciformis (Koperdraad,
Asegaaisteekgras) and Elyonurus argenteus
(Suurpol) as well as Eragrostis plana are
becoming the dominants. Only on black
turf soil has Themeda triandra survived,
patchily, as dominant. In the southern
Orange Free State the False Karoo has
continued to move north and east at an
average rate of about a mile a year and is
now almost twenty miles further north and
east than it was in 1953. In the Cape the
False Karoo now extends beyond Lehmanns-
drift in the Queenstown District. Themeda
triandra (Rooigras) disappeared as dominant
16
■^WSC-
• • ■.%;;'6*‘v;‘ , ••j
tV^r r •
Fig. 12. — Aerial view of typical sand dunes in the Molopo region of the Southern Kalahari,
17
from this area years ago, leaving Eragrostis-
veld, but in the past three years the Karoo
pioneers have established themselves to make
it False Karoo.
Forests of the Eastern Transvaal. — The de-
tailed survey by Mr. Scheepers of Westfalia
Estates in the Duiwelskloof area is nearing
completion.
Mountain Fynbos Vegetation. — This is a
major, long-term project which was started
this year by Mr. Taylor. Fynbos (macchia)
is a complex vegetation type which has
received very little attention from ecologists.
There are numerous problems connected with
its management, particularly relating to veld-
burning, consequently a detailed ecological
investigation of fynbos seems highly desirable.
The project will be tackled in several stages,
the first being a survey of the Cape Peninsula.
At Jonkershoek, Mr. Taylor is continuing
with the intensive study of regeneration of the
vegetation on a firebelt. The area is visited
monthly to collect information on phenology
and growth-rate of selected plants. A re-
survey of plant cover by an adaption of
Tidmarsh and Havenga’s Wheel-point
Method was undertaken and further infor-
mation on growth-rate of the chief species
was obtained from dry-weight determinations.
Quantitative Ecology
Two papers embodying the results of
research work in this very important field
were submitted by Mr. von Broembsen to
the Ninth International Grassland Congress
held in Sao Paulo, Brazil, during December
1964/January 1965. The first dealt with a
study of the point method of vegetation
measurement by means of an electronic
analogue apparatus and the second with a
wheel-point apparatus for the survey and
measurement of open and semi-open savanna
vegetation. In addition to research, much
time was spent on advisory work. Guidance
was given to several workers employing the
Williams and Lambert method of association
analysis. This also involved supervising the
programming of the data obtained.
Miscellaneous Services
The Botanical Survey Section continued to
work in close co-operation with the Toxi-
cology Department of the Veterinary Re-
search Institute at Onderstepoort and with
field veterinarians in connection with
problems of plant poisoning of stock.
Among the plant species involved were the
Mountain Cycad, Encephalartos longifolius.
Fig. 13. — Sandy flats in the Southern Kalahari with
scattered Camel-thorn trees (Acacia giraffae)
and a sparse grass cover; tufts of Stipag-
rostis iiniplumis in the foreground.
Cestrum laevigatum (Inkberry) and Dichape-
tahim cymosum (Gifblaar). In addition,
several plants, suspected of being poisonous,
were collected and sent to Onderstepoort for
testing.
Among the special reports prepared by the
Institute was one by Mr. Taylor on the
control of the alien Hakea for the Winter
Rainfall Region and one by Mr. Wells on the
soils of the 1820 Settlers Memorial Provin-
cial Nature Reserve for the technical com-
mittee responsible for the development of
the Reserve.
18
VISITORS
Visitors to the National Herbarium in-
cluded Dr. E. Lindner, Entomologist, of Bad
Godesburg, Germany, and Dr. Ernst Schiiz
of Stuttgart, Germany, who is compiling
biographical information on the collector
Krauss, who travelled from the Cape to Port
Natal in 1838-39. Mr. B. L. Burtt, from the
Royal Botanic Garden, Edinburgh, an
authority on the Gesneriaceae, visited the
Transvaal and Natal to collect plants and
seeds of Streptocarpus. He was taken to the
eastern Transvaal by Mr. J. C. Scheepers and
through Swaziland and Natal by Mrs. Olive
Hilliard, a member of the staff of Natal
University, Pietermaritzburg, who has been
engaged on a revision of the Natal members
of the genus.
Sir Oliver Leese, whose interests are
divided mainly between succulent plants and
cricket, was taken on a short collecting
expedition around Pretoria during the
M.C.C. tour of South Africa. Mr. J. Gomes
Pedro, an old friend of botanists in Pretoria,
now holding a high government post in
connection with economic planning of
Mozambique, visited the Institute with
Mr. J. M. Santareno, Director of Agriculture,
Luanda, Angola.
Prof. Liitjeharms of Cape Town and
Prof, van Zinderen Bakker of Bloemfontein
paid welcome, though short, visits to discuss
various botanical matters. Dr. David
Woods, recently appointed as lecturer in the
Botany Department of Natal University,
Pietermaritzburg, spent a few days at the
Institute, where he discussed his speciality,
viz. certain aspects of quantitative ecology,
with Dr. Killick and Mr. von Broembsen.
Prof. R. H. Compton and Mr. L. C. Leach
worked in the Herbarium from time to time,
while Col. R. D. Bayliss was a frequent visitor.
Mr. H. J. T. Venter, Lecturer in Botany at
the University College, Ngoye, brought some
of his problems to the National Herbarium.
The Botanic Station, Durban, receives
numerous calls from tourist ships stopping
over at the port. Among visitors whose
interests were of a more strictly botanical
nature the following may be mentioned:
Mr. Cannon of Michigan, U.S.A., Dr. E. A.
Schelpe of the Bolus Herbarium, Mr. R. H.
Simons of the Seaweed Laboratory of the
Institute; Mrs. J. Day and Miss P. Haygarth,
great-nieces of Medley Wood, Mr. L. H.
Wood and R. M. Wood, great-nephews of
Medley Wood, and Mr. Nolan Weaver
Wood, son of Mr. R. M. Wood.
A visitors book was instituted at the Albany
Museum Herbarium, Grahamstown, and 192
visitors (not including 120 students) were
recorded in 126 days.
Visitors to the Stellenbosch Botanic Station
included Prof. H. B. Rycroft, Prof. J. B.
Gouws, Dr. P. W. Roux (Middelburg),
Mr. W. Marais (on holiday from Kew),
Mr. D. Donald and Mr. O. Kerfoot (Depart-
ment of Forestry), Mrs. Vogts and Mrs. Irma
Kerr. Groups from the Cape Natural
History Club and the Botanical Society of
South Africa also visited the Station in
organized parties.
CONGRESSES AND STUDY TOURS
Dr. B. de Winter represented the Institute
and the Department of Agricultural Technical
Services at the following congresses: — •
International Union of Biological
Sciences, Prague, Czechoslovakia, from
18th to 22nd July, 1964.
International Biological Programme,
Paris, 23rd to 25th July, 1964.
International Botanical Congress, Edin-
burgh, 28th July to 1 1th August, 1964.
In the first two congresses the exchange
of ideas was fostered and international co-
operation in the various fields of biology
was promoted.
Some impressions which he gained at the
International Botanical Congress may be
summarized as follows: —
1. It is fruitless that some botanists, who
are not prepared to express their views
during sessions on nomenclature, are ready
19
to criticize the decisions which are arrived
at by the more articulate “ nomencla-
turists
2. Although the use of computors for
objective classifications was advocated in
several papers, the majority of botanists
are of the opinion that computors cannot
do taxonomy, but may play a role in
correlating the ever-increasing mass of
information available to taxonomists.
3. The experimental approach to taxo-
nomy, frequently referred to as the “ New
Systematics ” is gaining momentum and
any institute engaged on taxonomic studies
must take cognisance of this.
4. There are still large areas served by
only a few botanists where the primary
need is for the production of standard
“ Floras ”, based essentially on studies of
external morphology. This approach pro-
vides a basic framework for experimental
studies which may be undertaken later.
Before and after the International Con-
gress, Dr. de Winter spent some time
studying herbarium material, mainly of
Hermannia, at Kew, British Museum (Natural
History), Stockholm, Uppsala and Copen-
hagen, in connection with his revision of the
genus for The Flora of Southern Africa.
South African Association for the Advance-
ment of Science. — At the Kimberley meeting
of the Association held in July, 1964,
Miss Verdoorn, President of Section B,
delivered her presidential address entitled
“ Plant Taxonomy in South Africa”. Other
members of the Institute who attended the
Congress were Dr. Dyer (Recorder for
Section B), Miss Gunn and Dr. Codd.
Mr. M. J. Wells attended ten meetings of
the Technical Advisory Committee of the
1820 Settlers Memorial Nature Reserve and
gave advice and practical assistance in the
establishment of a garden to surround the
1820 Settlers Monument in Grahamstown.
Overseas Study Tour. By D. Edwards
For a botanist, native to the southern
hemisphere, a visit to Europe is like a
pilgrimage to Mecca : there our science has
developed and there is the vegetation of
which we read in the textbooks and scientific
journals. For aiding this pilgrimage, I wish
to record my grateful thanks to the Public
Service Commission for a bursary for a year’s
study of plant ecological trends in Europe and
the United States of America. It is also a
pleasure to acknowledge the help and warm
hospitality received from the many people
visited in the different countries.
August 1964, the first month of the year’s
study, was spent at the Centre d’Etudes
Phytosociologiques et Ecologiques (C.E.P.E.)
in Montpellier, France, under the direction of
Professor L. Emberger. Here, under the
guidance of M. Godron, I encountered for the
first time the quantitative group ecological
approach developed by this efficient and well-
organized Institute whose concepts and
methods are yet to make their full impact on
international ecological thought. The prac-
tical side of the C.E.P.E. approach at field
stations at Aumont Aubrac in the Central
Massif area and at Lamotte Beuvron in the
Sologne area were ably demonstrated by
M. Daget, M. Poisonnet and M. Grandjouan.
While in France, a short visit was also made
to Professor Rey in Toulouse at his well-
known air photo vegetation mapping
Institute.
During September, a week in Zurich,
Switzerland, allowed visits to Professor Ellen-
berg of the Riibel Institute, Dr. Markgraf of
the Botanic Gardens and the fine Forestry
Research Institute. In addition to the
stimulating discussions held with Professor
Ellenberg, he assisted us by providing a most
pleasant solution to the accommodation
shortage, resulting in our camping on the
shores of Lake Zurich.
Brief visits were paid to Professor
Merxmiiller and Dr. Roessler of the Munich
Botanic Gardens, to Drs. Kreeb and Leith of
Professor Walter’s Botany Department at the
0015958—4
20
Technische Hochschule in Stuttgart-Hohen-
heim, and to Professor Krausel and
Dr. Conert of the Naturmuseum in Frankfurt.
Nearly two weeks were then spent under the
expert guidance of Professor R. Tiixen in the
delightful countryside of Rinteln in the
Teutoburgerwald and at Haselune in north-
west Germany. This stay provided a prac-
tical insight into the Zurich-Montpellier or
Braun-Blanquet phytosociological system.
Later, during November, further contacts
were made with other well-known Ziirich-
Montpellier phytosociologists: Drs. West-
hoff, Barkmann and Segal in the Nether-
lands. It became apparent how little real
understanding there is of the Zurich-Mont-
pellier approach amongst Anglo-American
ecologists in general.
The autumn months of October and
November found us in the old world town of
Delft in the Netherlands. Here a basic
knowledge of photogrammetry and air photo
interpretation was acquired at the Inter-
national Air Training Centre for Aerial
Survey. Despite the floristic complexities of
African vegetation, air photos provide a most
useful tool in ecological survey work. The
conclusion was inescapable that the full
possibilities of air photos for ecological work
are as yet unrealized.
While in Holland it was a pleasure to renew
the 1963 contact made in South Africa with
Professor Lanjouw of the Botany Depart-
ment, University of Utrecht, and of Professor
A. Meeuse, formerly a member of the
Botanical Research Institute and now head of
the Hugo de Vries Laboratories of the
University of Amsterdam.
Our stay on the Continent concluded with
brief visits to Professor Robyns of the Royal
Botanic Gardens, Brussels, with whom con-
tact had also been made during his 1963 visit
to South Africa, and to Professor Duvig-
neaud of the Universite Libre in Brussels,
whose many fields of activity include
researches in phytosociology and production
ecology.
Fig. 14. — Releve in the Aumont Aubrac area of France by M. Daget and M. Poissonet of the Centre
d’Etudes Phytosociologiques et Ecologiques, August, 1964.
21
In gloomy December we arrived in Britain.
Despite the weather, visits were made to
Dr. A. Watt of the Botany School
Cambridge, to the Nature Conservancy in
London and at Monkswood, to the Land
Resources Division of the Directorate of
Overseas Surveys, to the Botany Departments
of London University College and the
University of Southampton, and to the
herbaria at Kew and the British Museum.
The second half of the study tour, which led
to the American continent, commenced with
a week’s visit to various branches of the
United States Department of Agriculture
and the Patuxent Wild Life Research Centre
in Washington, DC. Amidst continental
midwinter weather we proceeded to Madison,
Wisconsin, our home for the next four
months. Under Professor G. Cottam the
Plant Ecology Laboratories of the University
of Wisconsin became internationally recog-
nized for their quantitative ordination work
on vegetation under the aegis of Professor
J. T. Curtis. The months spent studying the
“ Wisconsin approach ” were amongst the
most rewarding of the year. While stationed
in Wisconsin, a journey to the south-eastern
United States with Professor H. litis and his
students provided a most valuable practical
understanding of the phytogeography and
vegetation of the United States. This back-
ground was particularly valuable during the
last month of June when a tour of the
western and southern parts of the country
was made, visiting Institutes in South Dakota,
Montana, Utah, Arizona, New Mexico,
Texas, Lousiana and Tennessee. Thus was
acquired a general appreciation of the
vegetation in areas which, in their ecological
aspects and practical problems of land
management, often bear a striking similarity
to certain parts of South Africa.
Fig. 15.— Professor R. Tiixen and student making a releve of a Jimipems community near Haselune,
north-west Germany, September, 1964.
22
The last month of the study tour was
again spent in Britain, now so different in its
lush greenery from the grey of December.
Visits were paid to the well-known quan-
titative ecological school under Mr. Greig
Smith at Bangor in Wales, to the Botany
Department in Aberystwyth, and to the
Nature Conservancy at Bangor. In Scotland
a series of instructive visits was made to the
Botany Department of the University of
Aberdeen, and to Dr. J. King of the Scottish
Hill Farming Research Organization, in-
cluding a visit to the experimental station in
the Cheviot Hills.
The end of July found us on our way back
to South Africa, experiencing once again the
sharp change of season. Many new friends
were made over the past year, experience was
enriched and much new knowledge was gained.
Fig. 16. — Sesamolhanmiis giierichii, Omaruru
District, South West Africa.
COLLECTING EXPEDITIONS
In addition to the routine intensive plant
collecting activities of the ecological survey
section (see p. 15) over the past year,
several expeditions were undertaken by
various staff members of the Institute to
explore botanically lesser known areas and
to collect plants for particular botanical or
other interests. Such collecting expeditions
tended to concentrate in four areas: South
West Africa, north-eastern Natal, the eastern
Transvaal and the south-western Cape Pro-
vince. Apart from the new distribution
records, these expeditions yielded a number of
new plant records for the area covered by
the Flora of Southern Africa.
South West Africa
During March and April 1964 a seven
weeks’ expedition in South West Africa was
carried out by Messrs. Tolken and Hardy
of the Institute, Dr. Jordaan of the Council
for Scientific and Industrial Research and,
later, Mr. Giess of the Windhoek Herbarium.
The primary objective was to collect plants
for the combined plant-alkaloid survey
project with the C.S.I.R. In this preliminary
sampling of plants of South West Africa, the
endeavour was to cover most of the different
types of vegetation in the country but,
because of the scattered rains of the season,
much of the vegetation was in poor condition
for collecting. The north-western parts,
however, provided many interesting plants
such as Moringa ovalifolia, Sesamothamnus
guerichii and many species of the interesting
genus Commiphora. Although usually
regarded as rare. Aloe dinteri was found in
many localities, while an Aloe closely allied
to A. mendesii, and possibly new, was found
in the Kaokoveld. Once again the botanists
were intrigued by the aquatic Chamaegigas
intrepidus (Scrophulariaceae) a perennial
showing amazing survival in the sporadically
water-filled pot-holes in the granite of the
Erongo Mountains. Many plants were also
collected in the areas bordering the Namib
Desert, which was in an exceptionally fine
condition. These included what appears to
be a new species of Citrulliis, while further
material of the Dombeya growing at Bulls-
poort enabled Miss Verdoorn to conclude
23
Map 1. — Route followed in South West Africa by
Messrs. H. R. Tolken and D. S. Hardy.
that it was worth separating as a variety of
the widely spread D. rotundifoUa. Altogether
over 900 herbarium specimens, over 300
specimens for chemical analysis and many
seeds and plants for the garden were obtained.
Northern Zululand.
Another area that has recently attracted
the plant collectors because of its rich and still
relatively unknown tropical flora, is that of
north-eastern Natal (i.e. the northern
extremity of Zululand near the Mozambique
border). Three short expeditions by
Mr. Vahrmeijer during December 1964,
March and May 1965, to the Makatini Flats,
Sordwana Bay, Lala Nek, Kosi Lakes and
Makane’s Pont areas yielded two new genus
and six new species records for South Africa
(see p. 7). One of the tragedies of the
increasing land use in the area is the
destruction of the Coast Forests, such as that
at Banga Nek, resulting in encroachment by
drift sands; the Sibayi Forests, being
destroyed by fires within the forest and the
extremely localized Raphia farinifera Swamp
Woodland being severely reduced because of
the use of this palm as a building material.
Further south the interesting and, in many
ways, unique Dukuduku Coast Forest has
been made available for agricultural exploi-
tation. This forest was visited and reported
upon by Mr. Strey of the Durban Botanic
Station. It is hoped that some remnants of
the forest can be set aside as small nature
reserves.
Shorter Expeditions
Short expeditions by various members of
the Institute also produced interesting results.
Mrs. Obermeyer-Mauve, in connection with
her revision of South African aquatic plants,
undertook several excursions around Pretoria,
to the eastern Transvaal and to Natal (see
p. 5). Three short excursions to the
Woodbush, Mariepskop, Lydenburg and
Fig. 17. — A new variety of Dombeya rotimdifolia at
Naukluft, Rehoboth District, South West
Africa.
24
Map 2. — Plant collecting expeditons to Northern
Zululand by Mr. J. Vahrmeijer.
Barberton areas of the eastern Transvaal were
undertaken by Mr. Scheepers. During one
of these, in December, 1964, he was accom-
panied by Mr. B. L. Burtt of the Royal
Botanic Garden, Edinburgh, to collect living
plants and seeds of Streptocarpus spp. These
excursions also provided much useful taxo-
nomic material, for example Podocarpus
specimens for Dr. Leistner’s revision of the
genus, enabling him to clear up several
doubtful points. Contrary to the treatment
in Flora Capensis (1933), it now appears that
only two species, P. falcatus and P. latifolius,
occur in the Transvaal. Mateiial was also
collected towards the elucidation of forest
species of the genera Eugenia, Ochna and
Trichilia, while the collection of Melhania
randii under extremely localized conditions
near Barberton provided the confirmation
required by Miss Verdoorn for inclusion of
this species in her revision of the genus.
Fig. 18. — Chamaegigas intrepidus in a deep pocket which remains dry for long periods, in granite formation
in South West Africa.
25
Dr. Dyer and Miss Verdoorn carried out
four field excursions to study Encephalartos
spp., verifying the south-eastern Transvaal
records of E. lebomboensis and E. villosus, and
investigating the newly-discovered species E.
inopinus at two localities in the eastern Trans-
vaal. Material for the description of a new
Dombeya species, D. autumnalis, was collected
near Penge Mine. Further study of a
Crinim, which occurs on black clay soil
adjoining the Nyl River in northern
Transvaal, confirmed that it is probably
C. forbesii; this is a markedly disjunct
distribution from its previously known
locality nearly 300 miles away in the Pongola
area of northern Natal.
In the south-western Cape Province the
Superintendent of the Botanic Garden,
Mr. Admiraal, made two short expeditions
with Mr. van Breda of the Worcester Veld
Reserve to the Jonkershoek, Dutoitskloof,
Ceres, Boschjesveldberg and Bredasdorp
areas. They collected over 350 plants of
Ericaceae, together with seeds of various
plants and nearly 100 herbarium specimens.
A considerable number of the plants collected
are now established in the Pretoria National
Botanic Garden. Mr. Oliver spent three
weeks collecting, mainly Ericaceae, in the
Outeniquas and Swartberge, while a 10-day
field study by Mr. Taylor and Mr. Marais is
described below.
Fig. 19. — Flower of Chamaegigas intrepidiis, showing
the mauve and white corolla above the
four green floating leaves, X8.
Map 3. — Plant collecting expeditions carried out in the
south-western Cape Province: by
Messrs. Taylor and Marais; by
Messrs. Admiraal and Van Breda.
A Cycloptychis Hunt. By H. C. Taylor
Cycloptychis marlothii, a rare shrubby
Crucifer, is known only from the outer fringe
of the fynbos on dry ridges separating the
Koue Bokkeveld plateau from the arid flats
of the Tanque Karoo. The only two
previous collections of this plant consisted of
meagre fruiting material and, as Wessel
Marais, revising the Cruciferae at Kew, could
not complete the description without flowers,
we set off in mid-September, 1964, to look
for the necessary material.
26
Marloth’s type specimen, collected sixty
years ago, bears only the cryptic legend
“ Swartruggens ” and the plant has not been
recorded since from that locality. As the
Swartruggens, more like an ill-defined
elevated rocky plateau than a ridge, stretches
for some 50 miles northward from Karoo-
poort (north of Ceres) it was with some mis-
giving that we reached the top of Katbakkies
Pass, which crosses the centre of the “ range ”,
and looked around for a “ broom-like shrub
up to 6 ft. high ” — the only description we
had to go on. For the rest of that day we
searched unsuccessfully, wandering many
miles examining every broom-like shrub in a
flora consisting mostly of broom-like shrubs.
We came at length to a place of broken
rocks where our flagging spirits were revived
by the discovery of Brachycarpaea and Sili-
cularia sigillata, the latter in itself rare enough
to make the trip worth while. This looked
promising Crucifer country. But the sun was
sinking and the sky looked threatening so we
pitched camp in the lee of a group of
columnar red rocks and battened ourselves
down in our tent. An icy wind blew and
rain fell in torrents during the night.
Fig. 20. — Wessel Marais collecting Cruciferae in the
Ccies District.
Next day dawned clear. Straight after
breakfast Wessel set off to scour a likely
looking rocky outcrop about a half-mile
distant. At noon he returned jubilant: he
had found our quarry! True, there was
only one ailing bush of it but it bore sufficient
flowering and fruiting material to make
specimens for the principal herbaria of
the world. And so at last, sixty years after
its discovery, Cycloptychis marlothii could be
described in detail.
Later in the day we found the shrub in
quantity close to the site of Wessel’s first
discovery; the following day, on the farm
Suurvlakte, 30 miles to the north, in the
same ridge of hills, we made another
collection of this rare plant. It would be
interesting to know how frequently it occurs
in the intervening area. These parts are
botanically not well explored but it is unlikely
that the places which we stopped at almost
by chance are the only two localities where
it grows.
For the rest, our route lay northward to
Nieuwoudtville in search of Heliophila species.
We skirted the north-easterly edge of the
Cedarberg along a Jeep-track which
tenuously followed a level shelf on the
mountain slope and then descended in a series
of breathtaking zigzags to the neat white-
walled Mission of Wupperthal. The Arid
Fynbos of the Swartruggens, different from
the vegetation of the Cedarberg proper,
extends in a narrow band along the north-
eastern slopes of the Cedarberg, following our
route as far as Wupperthal. From here to
the north. Karoo veld types occur, except
for an isolated patch of Arid Fynbos on the
plateau’s edge above Botterkloof.
From Nieuwoudtville, where Heliophilas
were disappointingly scarce, we wended our
way by slow degrees along the usual route
over Van Ryns Pass to Vanrynsdorp, making
an interesting detour from there through the
Nardouw Valley and over Pakhuis Pass to
Clanwilliam. From here we followed the
new National Road, tarred, straightened,
levelled and built for speed, but our progress
was interrupted, and our search rewarded, by
more Heliophilas in these well-worn tracks
than in all the special localities we had
visited to the north.
27
Notes on some Kniphofias in the eastern Orange
Free State. By L. E. Codd.
During March 1965 I undertook a short
tour from Pretoria to Harrismith and thence
to Ficksburg to investigate certain Kniphofia
problems. One of these arose out of a dried
specimen received from Mrs. M. L. Jacobszof
Rensburgskop, recorded as flowering in
March, which appeared to be K. ensifolia
Bak., a normally spring-flowering species.
More information seemed desirable in order
to ascertain the status of this late-flowering
form.
Another problem concerned what appeared
to be a new species, represented by several
herbarium speciments from the eastern Free
State and adjoining regions of Basutoland.
Efforts to cultivate the species in Pretoria
having failed, it was considered advisable to
study living plants in the field.
Mrs. Jacobsz, who had kindly invited
me to stay at their farm Rensburgskop, met
me at Harrismith and first showed me the
site which the Harrismith municipality has
set aside for development as a botanical
reserve, for cultivation of attractive plants of
the neighbourhood. This interesting project
owes a great deal for its inception to the
initiative and efforts of Mrs. Jacobsz. A
form of K. praecox Bak. was in flower on
the damp, stony slopes in the reserved area.
At Swinburne we turned off to Rensburgs-
kop and, in less than a mile from the national
road, encountered several flowering specimens
of the first-mentioned problem, i.e. the late-
flowering K. ensifolia. The plants were
growing in black clay soil on grassy flats
beside a stream. The inch-long flowers with
well exserted stamens are practically identical
with those of typical K. ensifolia, though
there was usually more red pigment in the
buds, and the flowers were more yellow, than
is normal for this species. The robust
peduncles 4 to 5 ft. tall carry the cylindrical
inflorescences well above the somewhat
spreading leaves. The later flowering thus
seems to be a fixed character in these plants
and, together with the slight differences in
flower colour and growth habit, would
justify placing them as a distinct variety of
K. ensifolia.
On the farm .Rensburgskop, situated near
the conspicuous mountain of the same name,
I was able to see three species of Kniphofia
in flower, each in its own ecological niche.
The late-flowering K. ensifolia grew on black
clay soil; K. praecox occurred in quantity on
rocky slopes above a stream, with its compact,
rhomboid inflorescences varying in colour
from those with bright red buds and yellow
flowers to some with dull red buds and
greenish flowers; then, at higher altitudes
along a gully, together with scattered bushes
Fig. 21.— Mrs. M. L. .lacobsz on the Harrismith
Botanical Reserve.
(Photo by Mr. W. Weber).
0015958—5
28
of Leucosidea sericea, were several plants of
K. breviflora Harv. ex Bak., usually solitary,
with graceful, slender inflorescences of clear
yellow flowers and narrow grass-like leaves.
The typical form of this species grows at the
foothills of the Drakensberg on the Natal side
and has white flowers.
Next day we saw K. breviflora, again with
yellow flowers, in Oliviershoek Pass. An
interesting visit followed to the historical site
known as Kerkenberg, a mountain massif
at the top of the Pass. Here, among great
boulders as large or larger than a house,
Piet Retief’s daughter. Debora, camped with
a group of Voortrekkers, waiting for news
of her father concerning his journey into
Natal to negotiate with the Zulu chief
Dingaan for a grant of land. On receipt of
his message that they could follow him, she
commemorated the occasion by painting his
name and the date (which was his 57th
birthday) in green paint on the side of a huge
overhanging rock mass. The inscription,
now a national monument and protected by
a frame with a sheet of glass, reads:
“ P. Relief, den 12 Novr. 1837 ”.
In the Kerkenberg sanctuary four species of
Kniphofla were in flower. K. triangularis
Kunth was scattered in dense grass, pro-
ducing a charming effect with its short,
coral-red inflorescences on slender stems
arising from tufts of narrow grass-like
leaves, while K. ensifolia, K. praecox and K.
multiflora Wood & Evans grew together in
quantity in a marshy place between some
large rocks. A marshy habitat is normal for
K. multiflora, with its elongate inflorescences
of numerous small, whitish flowers, carried on
5 ft. peduncles. K. ensifolia and K. praecox,
however, normally have different habitat
requirements, and these three have not been
seen growing side-by-side elsewhere. The
explanation is that the late Mr. F. A. Steytler,
during his term of office as school inspector
in Harrismith, and with the help of members
of the local Voortrekker movement, planted
indigenous flowers from the neighbouring
countryside in the reserve. The close proxi-
mity of K. ensifolia and K. praecox at
Kerkenberg has led to hybridization and the
production of a series of intermediates
between the two.
Fig. 22. — A view of the Botanical Reserve near Harrismith, Orange Free State.
29
From Kerkenberg I took the route through
Golden Gate to Ficksburg. In the Golden
Gate National Park, K. triangularis is common
along the shady banks of mountain streams
and, in the Park, I found the undescribed
species (my second problem) in flower. The
leaves are somewhat flaccid and markedly
serrulate along the edges. The inflorescence
is dense, usually subglobose to ovoid in shape
though sometimes more elongate, dull red
at the apex and grading down to greenish-
yellow or yellow flowers, borne on peduncles
2 to 3 ft. tall. The long narrow bracts help
to distinguish this species and are reminiscent
of the bracts in the Cape species K. sannen-
tosa which, among other dilferences, has
smooth-margined leaves. Mine Dieterlen,
who was stationed at a mission station in
Basutoland, was the first to collect this
undescribed species and Phillips (who mis-
identified it as K. sarmentosa) records, in
Annals of the S. African Museum (1917),
that: “A decoction prepared from this
species is drunk to cure pains in the shoulders.
Women, during the time girls are undergoing
initiation rites, make use of this plant for
some purpose, but this being one of the
sacred rites of the tribe, no further infor-
mation could be obtained ”.
Fig. 23. — -The inscription by Piet Retief’s daughter
on a large boulder at Kerkenberg. (Photo
by Mr. H. BrunzlaflF.)
Fig. 24. — Kerkenberg Sanctuary. (Photo by Mr. H.
Brunziaff.)
?0
Fig. 27. — Kniphofia caiilesceiis.
31
A species worth mentioning because of its
spectacular appearance is K. caulescens Bak.,
which is common in western Basutoland, with
outliers in the eastern Free State. It is one
of the few Kniphofia species with a distinct
stem, which may branch several times. It is,
in addition, gregarious and forms extensive
colonies of plants with distinctly glaucous
leaves and dense inflorescences, red at the
apex and grading to creamy-white flowers
with markedly exserted stamens. The
Basutos believe that it acts as a charm
against lightning and it is, therefore,
frequently planted near their huts. It has
repeatedly been introduced into cultivation,
but rarely survives for long away from its
natural habitat.
After leaving Ficksburg I called on Mr. and
Mrs. Bodham-Whetham who had, from time
to time, corresponded with our Institute.
They were kind enough to ask me to stay
with them at their farm, Kirklington, which
has been described by the present Mrs.
Bodham-Whetham’s aunt, Mrs. R. E.
Bodham-Whetham, in her book A Garden
in the Veld (1933). It was fascinating
wandering in the garden, with its extensive
terraces and unusual trees and plants, but the
greatest asset to me was having the assistance
of two people with a keen interest in the
local indigenous plants. David Bodham-
Whetham’s main interest is studying the
ecological requirements of birds, but he is also
a keen observer of the plants in the veld.
He was able to show me plants of the new
species growing on the mountain behind his
house and took me to a locality near Gumtree
where he had, by chance, seen a Kniphofia
flowering more than fifteen years before.
Unfortunately, the plants collected, like
previous gatherings of this species, failed to
grow in Pretoria. However, the specimens
seen and collected at Kirklington and again
at the farm Franschhoek, about 20 miles
north of Ficksburg, provided adequate
material to confirm that this was indeed a
clearly defined new species, which will be
described in Bothalia, Vol. 9, part 1 (in
press).
PUBLICATIONS
The following parts of the serial
publications issued by the Institute appeared
during the year under review: —
Bothalia: Vol. 8, part 2 (December 1964).
Bothalia: Vol. 8, part 3 (May 1965).
Flowering Plants of Africa: Vol. 36,
parts 3 and 4 (20 plates).
Publications by members of the staff are
listed below in alphabetical order of authors’
names: — ■
Acocks, J. P. H., 1965. Karoo vegetation
in relation to the development of deserts.
Ecological Studies in Southern Africa.
Junk, The Hague, Holland.
Anderson, J. G., 1964. Eriochrysis brachy-
pogon subsp. australis, subsp. nov.
Bothalia 8 (2), 170. Danthonia aitreoce-
phala, sp. nov. ibid 170-172.
Codd. L. E., 1964. Kniphofia fluviatilis, sp.
nov. Flow. PI. Afr. 36, t. 1421. Kniphofia
obtusiloba. ibid. t. 1422. Kniphofia baiirii.
ibid. t. 1423. Kniphofia typhoides, nom.
nov. ibid, t. 1424. Thorncroftia longifiora.
ibid. t. 1425.
Codd, L. E., 1964. The South African
species of Orthosiphon. Bothalia 8 (2),
149-162.
Codd, L. E., 1964. Garcinia gerrardii.
Bothalia 8 (2), 174. Graderia lineari-
folia, sp. nov. ibid. 176-177.
Codd, L. E., 1965. Progress with the Flora
of Southern Africa. Webbia 19, 901-903.
De Winter, B., 1964. New combinations
and new species in the Aristideae.
Bothalia 8 (2), 173.
De Winter, B., 1965. The South African
Stipeae and Aristideae (Giamineae). An
anatomical, cytological and taxonomic
study. Bothalia 8 (3), 201-494.
Dyer, R. A., 1964. Ceropegia. African
Wild Life 18 (3), 215-226.
Dyer, R. A., 1964. Stathmostelma specta-
bile. Flow. PI. Afr. 36, t. 1436.
Ceropegia dimorpha. ibid. t. 1437.
Caralhinia bredae, sp. nov. ibid. t.
1438a. Caralluma bredae var. thomallae,
var. nov. ibid. t. 1438b. Thwanthos
nocturnale, sp. nov. ibid. t. 1439,
Brunsvigia orientalis. ibid. t. 1440.
32
Dyer, R. A., 1964. Cyrtanthus eucallus, sp.
nov. Bothalia 8 (2), 163. Ceropegia
esteUeana, ?,p.nov. ibid.\6^. Encepha-
lartos inopinus, sp. nov. ibid. 169-170.
Dyer, R.A., 1965. The type specimen of
Encephalartos paucidentatus. S. Afr.
Journ. Sci. 61, 50.
Dyer, R. A., 1965. New species and notes
on type specimens of South African
Encephalartos. Journ. S. Afr. Bot. 31,
111-121.
Hardy, D. S. (with J. Lavranos), 1965.
Aloe audhalica, sp. nov. Journ. S. Afr.
Bot. 31, 65-68.
Johansson, E. Y., 1964. Ferns and fern
allies of Transvaal. Fauna and Flora
14, 54-87.
Killick, D. J. B., 1964. Wahlenbergia
monotropa, sp. nov. Bothalia 8 (2), 164-
165. A note on the identity of Myrica
conifera. ibid. 175.
Marais, W., 1964. A note on Capparis
transvaalensis Schinz. Bothalia 8 (2),
165-166.
Marais, W., 1964. Notes on the genus
Heliophila. Bothalia 8 (2), 166-169.
Obermeyer, A. A., 1964. The South
African species of Aponogeton. News-
letter of the Limnological Soc. 1 (3),
12-13.
Obermeyer, A. A., 1964. Barleria rotiin-
difolia. Flow. PI. Afr. 36, t. 1426.
Burmannia madagascarensis. ibid.
t. 1427. GethylUs britteniana. ibid.
t. 1428. Dipcadi papillatum. ibid. t.
1429.
Obermeyer, A. A., 1964. The South
African species of Dipcadi. Bothalia
8 (2), 117-137.
Obermeyer, A. A., 1964. The South
African species of Lagarosiphon.
Bothalia 8 (2), 139-146.
Obermeyer, A. A., 1964. The South
African species of Anthericum, Cldoro-
phytiim and Trachyandra. Addenda and
Corrigenda. Bothalia 8 (2), 147-148.
Simons, R. H., 1964. Species of
Plocamium on the South African coast.
Bothalia 8 (2), 183-193.
Simons, R. H., 1964. Notes on the species
of Zonaria in South Africa. Bothalia
8 (2), 195-197.
Verdoorn, L C., 1964. Orchids as I see
them. Orchid Soc. Bull. 10, 4-5.
Notes on an interesting South African
orchid, ibid. 11, 5-6.
Verdoorn, I. C., 1964. Crinum baumii.
Flow. PI. Afr. 36, t. 1432. Crinum
variablle. ibid. t. 1433. Peltophorum
africanum. ibid. t. 1434.
Verdoorn, F C., 1964. Melhania integra,
sp. nov. Bothalia 8 (2), 177-178. Mel-
hania polygama, sp. nov. ibid. 178-179.
Verdoorn, L C., 1965. Plant taxonomy in
South Africa. S. Afr. Journ. Sci. 62,
43^9.
Von Broembsen, H. H., 1964. An appa-
ratus using a radio-active source for
generating random electrical pulses.
National Conference on Nuclear Energy
1963: Application of Isotypes and
Radiation.
Wells, M. J., 1964. The vegetation of the
Jack Scott Nature Reserve. Fauna and
Flora 15, 17-25.
PERSONALIA
Obituary
We regret to record the death of Mrs. Reino
Leendertz-Pott on 14th June, 1965, after a
short illness, at the age of 96 years. A more
complete account of her career will appear in
Bothalia, Vol. 9, part 1. Miss Leendertz
took up an appointment as Botanical Assis-
tant in the Transvaal Museum in 1895, thus
becoming the first official botanist of this
province. In 1910 she married Mr. Pott, a
Pretoria business man, and in 1925 she went
on pension at the age of 55. Subsequently
Miss A. A. Obermeyer (Mrs. Mauve) became
botanist at the Transvaal Museum. The
accompanying photograph was taken in 1953,
when the Transvaal Museum Herbarium was
transferred to the National Herbarium.
33
Appointments, Transfers and Resignations
Mrs. Marie M. Vogts was transferred to
the Fruit and Food Technology Research
Institute, Stellenbosch, because the accent of
her Protea research programme was slightly
changed towards commercial production and
its allied problems; she is continuing with
her experiments at Betty’s Bay. Mr. Phillip
van der Merwe and Mr. P. A. B. van Breda,
together with the Worcester Veld Reserve,
were transferred to the Winter Rainfall
Region of the Department of Agricultural
Technical Services, while the Fauresmith
Veld Reserve and Mr. A. F. J. Visagie were
transferred to the Orange Free State Region.
In August 1964, Mr. Hugh Taylor assumed
duty as Officer in Charge of the Stellenbosch
Botanic Station. Miss Mary Thompson was
appointed to the National Herbarium in
January 1965 and was transferred to Stellen-
bosch in March 1965. The professional staff of
this Station thus consists at present of two
Botanical Survey Officers (Messrs. Taylor and
Grobler) and two Systematic Botanists (Mr.
Oliver and Miss Thompson).
Mr. Hellmut Tolken and Miss Judy
Marsh, both with B.Sc. (Hons.) degrees,
assumed duty in the National Herbarium in
July 1964 and January 1965, respectively.
Mr. John Jessop also joined the National
Herbarium after completing a revision of the
South African species of Asparagus for his
M.Sc., degree.
Mr. Johan Landman was appointed in
January 1965 and was seconded to Pretoria
University with a bursary to take the B.Sc.
(Hons.) degree.
Fig. 28. Mrs. A. A. Obermeyer-Mauve, Dr. E. P. Phillips, the late Mrs. R. Leendertz-Pott and
Miss i. C. Verdoorn, taken in 1953 when the Transvaal Museum Herbarium was donated to the
National Herbarium. (Photo by Dr. R. A. Dyer).
34
Miss Sarah Troughton was appointed to
the Botanic Station, Grahamstown, in
January 1965, while Miss Wendy Herbert
was appointed to the McGregor Museum
Herbarium, Kimberley, in August 1964.
At the National Herbarium, the remaining
post of Artist was filled by Mrs. Aleida
van der Merwe in May 1965. Other appoint-
ments were Mr. Willie Bezuidenhout as
Technician (June 1965) and Misses Janet
Wormser and Miep Kat as Technical Assis-
tants in January and May 1965, respectively.
Miss Carina Visser resigned during the
year to take up a post as lecturer at Cape
Town University. The following Technical
Assistants also resigned for various reasons:
Misses Willie Bouwmeister, Beryl Nutt, Joke
Prinsen, Holdi Hartig and Ursula Radmacher.
Degrees, Awards, etc.
Our congratulations go to Dr. O. A.
Leistner for being awarded the D.Sc. degree
by Stellenbosch University for a thesis
entitled “ The Plant Ecology of the Southern
Kalahari congratulations also to the
following members of the staff for the award
of the M.Sc. degree: Messrs. P. J. Grobler
(Stellenbosch), Mr. E. G. Oliver (Cape Town)
and Mr. J. P. Jessop (Cape Town).
Mr. von Broembsen was awarded a Public
Service Commission Bursary for two years
overseas study and will be departing for
Southampton University in September, 1965.
Mr. P. J. Muller received recognition for
meritorious service in connection with the
Quantitative Ecology programme and was
awarded special promotion of one notch on
his salary scale.
Marriages
Miss Ida van Vuuren and Mr. Sakkie van
der Westhuizen (previously of the Institute’s
staff and now with the Atomic Energy
Board) were married in October 1964. She
was re-appointed in a temporary capacity in
the National Herbarium.
Births
It is a pleasure to congratulate Mr. and
Mrs. Mike Wells of Grahamstown on the
birth of their second child, a son, in
February 1965.
General
The Institute was subjected to an inspection
by an officer of the Public Service Commission
during the year. His recommendations have
been submitted to the Public Service Com-
mission and include an overall increase in the
research staff.
A general meeting of the Institute was held
in Pretoria during March 1965 when members
of the staff reported briefly on their recent
activities, and discussions were held on
matters of mutual interest.
Representatives of the National Film Board
visited the National Herbarium to collect
information on “ General Smuts as a
Botanist ” for a school film on South
Africa’s Prime Ministers. Photographs
were taken of the General’s private collection,
which is kept as a separate unit, and of plants
named after him, such as Ficus smutsii and
Digitaria smutsii.
Mrs. Forssman (Cythna Letty) celebrated
her 70th birthday on 1st January 1965, when a
small presentation was made to her by the
staff of the Institute.
Miss Judy Marsh returned in July 1964
from a three months visit overseas during
which time she spent a month working at
Kew and also visited the British Museum,
Edinburgh and Paris Herbaria in connection
with her revision of Widdringtonia.
Mr. Wessel Marais spent a short holiday
in South Africa from Kew in August 1964
and took the opportunity to study some of
his Cruciferae problems when he and Mr.
Hugh Taylor made a 10-day excursion in the
south-western Cape Province (see p. 25).
35
e QJ lowers
on the DECIMAL COINS
by Cythna Letty
■]D EAL good luck sometimes comes to a
back-room boy. It was my luck to
be on hand, re-employed to continue painting
flowers at the Botanical Research Institute
after reaching retiring age, when the Com-
mission for the new decimal coinage
announced in 1963 that our indigenous birds,
flowers and animals were to be depicted on
the proposed new series of coins.
On previous occasions, when advice was
required on floral designs for our stamps and
coins, my former Chief, Dr. R. A. Dyer,
and I were consulted. So it was not
altogether unexpected that the Institute
should again be approached. But this time
the approach was more direct and I admit I
was thrilled and gratified when Mr. Malan,
Director of the Mint at the time, asked me
to submit designs for four of the new coins.
Fig. 30. — Some of the new decimal coins, very slightly larger than natural size. Photo by pei mission
of the Government Printer.
Fig. 29. — Glon'osa virescens drawn as a design for the
Rhodesian sixpence, much reduced.
36
At about the same time I also accepted an
undertaking to draw the Flame Lily
{Gloriosa virescens) for the Rhodesian silver
sixpence, which was minted in 1964. This
was a relatively straight-forward design but,
when I started on our own coins, I found
that there was more to putting plant-life
on a coin than I had ever imagined.
Plants, unless highly stylized, tend to be
unmanageable; they seem to have no shape
and, when reduced from life-size to about
three-quarters of an inch or less, are inclined
to “ clot ”. Air must come in from all sides
and long stems along the diameter must be
avoided as they hinder the working of the
pantograph when reducing the drawing in the
process of making the die.
StreUtzia reginae (The Crane Flower), with
its orange, blue and white colours corres-
ponding to those of the South African flag,
seemed an obvious choice for the 50 cent piece
which was to be about 27 mm (a little over
an inch) across. This idea was dropped
after several attemps, the results were always
too spikey and too full of points. As an
alternative, I suggested that we use three
separate well-known flowers having the
colours of our flag: the white Zantedeschia
aethiopica (Arum Lily or Varkoor) and a
blue Agapanihus species combined with the
StreUtzia reginae. After submitting two or
three unsatisfactory arrangements, I made
tracings of the flowers and leaves and pinned
All sketches much reduced in size.
37
them on to a 15 inch circle until all the
requirements for a coin (including a large
clear 50) were met.
A Protea was a must, but which one: the
handsome Giant {P. cynaroides) or the
homely, beloved Suikerbossie {P. repens).
At risk of “ clotting ”, they were both used;
a large “ 20 ” completed the design.
For the half-cent I was asked to draw a
wreath of typical South African leaves.
This proved more difficult than one might
think. What could be chosen as a typically
South African leaf? A thorn-tree leaf could
easily be mistaken for a Wattle or a
Jacaranda; the Cycad could be a palm.
I tried Celtis africana (Camdeboo- or White-
stinkwood) with its characteristically veined
leaves bearing small, stalked berries in the
leaf-axils. It was not recognized by anyone
on the Commission and so it was not
accepted. Then the Aloe came to my rescue
and I used Aloe acideata because of its
inwardly curving leaves and very simple
inflorescence. This design was accepted, but
was adapted for use on the 10 cent coin.
I then drew the Acacia torti/is (Umbrella
Thorn or Haak-en-steek) for the half-cent,
should such a coin be required.
Fig. 35. — Wreath of Celtis africana drawn for the
proposed -\c coin but not used.
Fig. 34. — Protea repens and Protea cynaroides drawn
for the 20c coin.
Fig. 36. — Second design for the i-c coin, using Aloe
acideata.
Sketches reproduced by permission of the Director of the Mint.
38
Much thought was given by the Com-
mission to the designs, sizes, weights and
alloys to be used in making the coins. The
needs of the blind were considered and it is
for the benefit of this handicapped group that
large, clear numbers are present on all the
coins, while milled edges are provided on the
bronze coins and the silver R1 piece. A new
feature is the sunken image, which will not
easily, rub off.
I have been asked by several people (some
have been quite incensed) why my initials do
not appear on the coins which I designed.
It seems to be traditional that the artist who
makes the plaster relief from a design has
this privelege. Thus the initials of Thomas
Sasseen, the Mint’s artist, appear on the
obverse of all the coins, while Jan van Zyl’s
initials appear on the reverse of the animal
designs because he was responsible for both
the illustrations and the plaster reliefs.
The great day, the 3rd December, 1964,
arrived, when the Minister of Finance,
Dr. Donges, struck a limited number of the
new coins at a select party at the Mint.
As they came out for inspection I thought
Jan van Zyl’s animals and Dick Findlay’s
birds were lovely, that the 10 cent with its
Aloe looked better than I had expected and
that the Proteas had not “clotted”. However,
when the 50 cent piece was taken out of the
bucket and lay shining and groomed on its bed
of cotton-wool in front of me, I suddenly felt
collapsed like a string of boiled macaroni.
Fig. 37. — The Aloe aculeata design was adapted for
the 10c coin.
Fig. 38. — Cythna Letty and Mr. Thomas Sasseen in
the latter’s work-room at the Mint.
Fig. 39. — Dies prepared by Mr. Sasseen for some of
the coins, much reduced.
39
Editor’s note. — When Mrs. Forssman
(Cythna Letty) refers to herself as a “ back-
room boy ” it is, perhaps, not without some
justification. Since her return to the Institute
in 1945, she has occupied a small south-
western corner of about 10 y 8 ft. in the
National Herbarium. Her back is so near to
a herbarium cupboard that her arm is liable
to be jolted each time a botanist consults
the Flacourtiaceae, and there is scarcely
room for her sloping drawing-table between
the herbarium cupboard and a steel filing
cabinet in front of her. On her left is a
south-facing window and to the right is a
thick green curtain which cuts out the
afternoon light from a west-facing window.
This space accommodates, in addition to her
high stool and drawing-table, two heaters
which come into use when the temperature
falls below about 60" F, and a table on which
stands a plant press and an assortment of
bottles and tins for holding the flowers she
is painting.
In this little corner she has produced her
incomparable water-colour paintings of
African plants, which have been the mainstay
of our serial publication Flowering Plants
of Africa for many years. Her ambition
to reach a total of 500 plates for this series
has been well exceeded and her total now
stands at 730, an achievement which will not
easily be surpassed.
Cythna Lindenberg Letty was born in
Standerton, Transvaal, on the 1st January
1895, the first of five children of the marriage
between Walter Edward Letty of Greenwich,
England, and Josina Christina Lindenberg
of Worcester, Cape Province, who had
already borne six children during her earlier
marriage to David Johannes de Vaal
Leibbrandt.
Cythna’s father emigrated from England
in the interests of his health and arrived in
Cape Town in 1889 at the age of 25. A few
years later he married Cythna’s mother and
they moved to the Witwatersrand in 1893.
Obviously an accomplished but possibly
somewhat unpredictable person, he preferred
to abandon a career as a naval architect and
engineer, for which he was trained, and to
turn his hand to any occupation which carried
Fig. 40.— Mr. Letty, Cythna and Mrs. Letty in Johannesburg.
40
Fig. 41. — Travelling between Standerton and Vrede, 1898. Cythna in white bonnet near back of wagon.
Fig. 42. — Estcourt, 1900. Cythna standing at extreme right, front row, with her family.
41
with it a hint of adventure. In 1899, with
the rinderpest sweeping from the north
through the cattle population of the country,
he accepted a post in Natal with the
Veterinary Services, who were engaged on
efforts to curb the disease. The family
stayed in and around Estcourt, where Cythna
spent her first two years at school, until they
moved to Standerton in 1904 and to Heidel-
berg in 1910. At the outbreak of World
War I he joined the South African Expedi-
tionary Force to France and stayed on in
France after the war until his death in 1948.
Fig. 43. — First study of a Transvaal plant by
Cythna, in 1915, with a book on “ Wild
Flowers of the Transvaal ” already in
mind.
Fig. 44. — Scientific drawing in colour for Dr.
Quinlan, Veterinary Research Institute,
Onderstepoort, about 1926. The caption
reads : “ Cervicitis followed by Pavilionitis.
Histological examination showed chronic
changes in the uterine mucosa ”.
Cythna’s mother was an able and talented
woman, interested in everything around her.
She named her daughter after the heroine of
Shelley’s “ Revolt of Islam ”, a name which
present-day type-setters and editors persist in
changing to Cynthia. Although largely self-
taught, her paintings show the delicate touch
and the talent for life-like portrayal which was
to develop in Cythna. Together they picked
and studied the wild flowers which faded so
quickly that there was scarcely time to paint
them. In the Botanical Research Institute
are two books of her mother’s paintings,
depicting many of the veld flowers they found
in the Transvaal from 1909 to 1911. These
volumes stimulated another ambition in
Cythna’s mind which was not achieved until
1963, when her own book on Wild
Flowers of the Transvaal was published.
In her mother’s books it is of considerable
interest to see species illustrated such as
Nerine krigei and Stapelia leendertziae, which
were “ discovered ” and brought to the notice
of botanists only many years later.
In 1914 Cythna spent her last year of
schooling at the Girls’ High School, Pretoria.
In that year she walked up the path to the
National Herbarium for the first time, but
42
Fig. 45. — Cythna Letty at a small function cele-
brating the publication of her book.
Fig. 46. — Mrs. Verwoerd, wife of the Prime Minister,
with Dr. Dyer and Cythna Letty in the
National Herbarium, 1962.
not with employment in mind. She and her
mother took her mother’s books of paintings
to the Herbarium to have the plants named;
considering the state of our knowledge at the
time, it is not surprising that few of them
could be identified. After working as a
governess for a short time, as a nurse for a
year (qualifiying as a mid-wife) and as a clerk
in the Railway Stores for a few years, she
went to Cape Town from 1920 to 1924 to
assist her brother-in-law. Dr. Bosenberg.
Returning to Pretoria in 1925, she received
her first appointment in which she could
exercise her artistic talent — ^at the Veterinary
Laboratories, Onderstepoort, where an artist
was required to record, in exact colour, the
appearance of diseased animal tissues and
the organisms responsible. We can be
grateful that, in 1927, it was decided that a
man would be more suitable for the Onderste-
poort post, with the result that Cythna was
transferred to the then Division of Plant
Industry, of which Dr. I. B. Pole Evans was
the Chief.
In 1938 she resigned to marry Oscar
William Alric Forssman, of Swedish extrac-
tion, whose grandfather had come to South
Africa in 1842. One son, Bruce, was born
in 1940. In 1945 she returned to her little
corner in the National Herbarium and took
up where she left off in 1938. Although her
work has long been a source of pride in the
National Herbarium, it received com-
paratively little recognition until her book
Wiki Flowers of the Transvaal appeared,
when her fellow-countrymen and even
overseas botanists suddenly became aware of
her genius. It was, in fact, the belated
recognition of our own “ back-room boy ”.
She is commemorated in the botanical
names Aloe lettyae Reynolds and Crassula
lettyae Phillips but what apparently has
pleased her more than anything else is the
creation of the “ Cythna Letty Nature
Reserve ” in the Barberton mountains. This
is an area of about 15 acres of natural veld,
set aside by the Twello Bosboukorporasie to
preserve the only locality in the Transvaal
where the rare species Aloe vryheidensis is
found. — L. E. C.
43
FIELD WORK
is not all fun
by R. A. Dyer
you tell a friend at Head Office
that you are going on a botanical field
excursion and need an advance of subsistence,
his usual reaction is to say; “ How nice! ”
And he leaves you with the impression that
he thinks you are going on a picnic at his,
the taxpayer’s, expense. This is rarely the
true picture and may be quite contrary to
fact.
It really is unnecessary to delve into the
distant past for an example of a hair-raising
experience, but I recall one involuntarily
which took place during my brief visit to
Tristan da Cunha in 1937. While on a
climb near the Settlement 1 was temporarily
marooned on a precipitous mountain ledge
in a sudden Antarctic squall. It seemed that
I would not survive to tell the tale and never
again sing the song: “ We are marching to
Pretoria ”.
In order not to be accused of selecting
exceptional cases, however, only occurrences
of the last four field trips, all within a year,
will be mentioned.
The first outing was to the eastern Cape in
September, 1964. Nothing untoward hap-
pened during the field work in the neigh-
bourhood of Queenstown, Cathcart, Komga,
East London and Grahamstown. Here
Encephalartos friderici-guilielmi, E. princeps,
E. latifrons, E. coffer, E. alteiisteinii and E.
trispiiwsiis were investigated and other
Fig. A1 .^Encephalartos eugene-maraisii, Waterberg District, with the author.
44
species were studied in cultivation. From
Grahamstown 1 travelled westwards with M.
J. Wells, Officer-in-Charge of the Botanical
Research Station. Soon after our discovery
of the first specimens of E. longifolius
towards Joubertina, on the first day out,
weather conditions deteriorated. By the next
morning the higher mountains were snow-
clad. Rain and sleet fell almost incessantly
while we trudged several miles into the
mountains to a small grove of E. longifolius
of special interest.
Conditions remained grim throughout the
day and by nightfall when we arrived at
Jansenville we were wet to the skin and
numb with cold. On the following day,
between spells of torrential rain, we investi-
gated E. longifolius and E. lehmannii on the
Klein Winterberg. Drifts became impassable;
the gravel road was turned into either a
skating rink or a quagmire and we slid from
one to the other and got stuck for a couple of
hours. Only with the help of a farm truck
and tow rope were we eventually extricated.
At night-fall, when it appeared that we were
safely at Uniondale, the unusual sway of the
van foretold a puncture!
The following day Mike Wells had to wade
through a wide drift of icy-cold water to test
its depth before we were able to proceed to the
habitat of E. horridus near Uitenhage.
Unknown to us or the attendant at the
wayside garage, the underground petrol
reservoir had been flooded by the overnight
Fig. 48. — Encephalartos friclen'ci-giiilielmi, Cathcart District, with Mrs. Dyer on the left and
Mrs. N. H. James on the right.
45
storm. Hardly were we out of sight of the
pump than the van spluttered to a jerky stop.
Don’t let me bore you with details. Our
petrol tank had to be drained but even after
this we stopped half a dozen times to empty
the water-trap below the carburetter. In
addition, we were caught in the middle of a
double convoy of military and commercial
vehicles stretching over three miles of
winding road to Grahamstown. What a
joy!
Excursion number two was in November,
1964, to the Pongola and Jozini Dam area,
where E. lebomboensis, E. villosus and E,
ngoyanus are known to occur. In addition,
a Cycad of interest had been reported south
of Sulphur Spring. Because of a misunder-
standing and a broken luggage carrier which
had to be spot-welded at a wayside garage.
Miss Verdoorn, Mr. Admiraal and I arrived
later than expected at Sulphur Spring.
From there we were guided by young
Mr. Andre du Plessis and friends into a
rugged area of Bantu reserve. The two cars
were parked where the rough track petered
out. We were told that the Cycads were not
over the next or the next, but over the next
mountain ridge. There was no alternative
but for some of us to go forward with
collecting impedimenta in spite of the lateness
of the hour which, by then, was after 3 p.m.
The further we tramped and the later it got,
the less important Cycads became to me.
All we found by the time the sun was setting
were a couple of seedlings in a steep kloof.
Night was drawing in fast, a storm was
brewing so it was agreed that our Bantu
guide would lead us back over the mountain
by a short cut.
Rain began to fall before we reached the
top of the ridge and soon the storm broke
in full force with vivid flashes of lightning
alternating with inky darkness and deafening
peals of thunder. Our guide soon slipped
out of sight, no doubt into a friendly kraal,
and left us. One can sympathize with him
now; the conditions were fearful — enough to
scare one stiff. We missed the mountain
track, missed our footing, tumbled down
slopes and over boulders, collecting a score of
bruises on various parts of our anatomy until
eventually a faint light was seen in the
distance. It proved to come from a Bantu
kraal where the occupants were preparing for
the night’s rest. The headman volunteered
to lead us to where he had seen our cars
parked earlier in the day. He did so through
another couple of miles of slippery slush.
By the time we reached the cars it was well
after nine o’clock, but our ordeal was not
quite over. On the return journey to the
main road, a dry watercourse was now a
torrent nearly 100 yards wide; how we
managed to cross the causeway with nothing
to guide us but the track emerging beyond
the flood is still a wonder. All hotel accom-
modation at Pongola had been taken and
it was only through the kindness of the town
ranger and his wife that we had a late
supper and a bed to sleep in. It is true that a
botanist and a gardener shared a double bed,
but there have been stranger bed-fellows and,
anyway, beggars can’t be choosers. The rest
of the trip was tolerably successful and
uneventful. One of the young men said
something to the effect that there is life in
the old dog yet.
Fig. 49. — Encephalartos inopiitiis, collected for the
Botanic Garden, Pretoria, with the author
on the right and Mr. C. A. Els, the dis-
coverer of the plant.
46
The third field job was in search of what
turned out to be a species new to science,
which I later described as Encephalartos
inopinus. This investigation entailed wading
through the Steelpoort River near Kromellen-
boog Asbestos Mine with rucksacks, rope,
crowbar and cameras, and scrambling up
and along a cliff where any foot-fault would
land one as a mutilated mess in the river
below. Not all the party was able to
negotiate the climb and some recoiled
from the dizzy height. The accompanying
picture shows the single specimen on the rock-
face at the top of a sheer 200 ft. cliff above
the Steelpoort River.
The last of the four field trips, which was
undertaken in March, 1965, was again in
search of E. inopinus in the Olifants-Steel-
poort complex of valleys. On this occasion
our troubles arose through the over-
estimation of our endurance by our athletic
guide Louis Naude. For once we took
neither food nor water as a precautionary
measure before leaving the car at Welte
vreden Asbestos Mine at 2 p.m. It was a hot
three-hour trek up hilt and down dale to our
first view of the Cycad on the heights over-
looking the Olifants River. On the way home
at nightfall, when climbing up a steep rocky
track about three miles from our base, the
strain began to tell. Miss Verdoorn collapsed
with cramp and nausea. If ever there was an
example of mind over matter, it was shown
by her that evening. It was comforting to
see a search party with torches as we limped
round the last bend to come in sight of
Weltevreden.
In connection with the hunt for E. inopinus
the kindness and assistance of the officials
and staff of the Penge, Kromellenboog and
Weltevreden Asbestos Mines is gratefully
acknowledged.
These experiences, which make botanical
fieldwork memorable, but not a picnic,
bring to mind the old ditty which runs : “ In
the morning, in the evening, don’t we
have fun ”.
Fig. 50. — Encephalartos inopinus on a cliff over the Steelpoort River, with the author on the right and
Bantu assistant. (Photo by L. E. Codd).
47
Plants which cause
^VOUSIEKTE^^ POISONING
by L. E. Codd and S. Voorendyk
Introduction
“ ^OUSIEKTE ”, which is caused by several
Rubiaceous plants, may be literally
translated as “ Rapid Sickness ” and is
characterized by the sudden death of the
animal without any warning symptoms.
Only ruminants, such as cattle, sheep and
goats, appear to be affected. No cases have
been recorded in horses and, as yet, it has
apparently not been possible to induce deaths
from gousiekte in small laboratory animals.
The latter factor has greatly increased the
difficulties involved in investigating plants
suspected of causing gousiekte; it is also a
handicap in attempting to isolate the toxic
principal, which still remains unknown.
Further complications are that the toxicity
of the plants apparently varies from year to
year and according to the season, while there
is evidence that individual animals vary in
susceptibility. Yet another complication is
that death does not occur shortly after the
animal has ingested the plant material, as is
the case in poisoning caused by “ Gifblaar ”
{Dichapetalum cymosum), but may take two
to six weeks, by which time the animals may
have moved to a different camp.
It has been established that death is due to
heart failure associated with lesions in the
muscle tissue of the heart (myocardium).
Under the microscope it is usually possible to
see the scar tissue in the heart muscle,
especially towards the apex of the heart, and
this evidence can be regarded as diagnostic
for gousiekte.
Stock losses due to gousiekte were known
from the early days of this century and were,
at first, confused with gifblaar poisoning in
which death takes place with equal suddemess.
However, the gifblaar plant was already well-
known to farmers before the turn of the
century and it soon became apparent that
gousiekte losses occurred in areas where the
gifblaar was not present, for example in the
Ermelo District. It was finally established
by Theiler, du Toit and Mitchell (1923) that
Pachystigma pygmaeum, already known in
the Pretoria District as the “ gousiekte-
bossie ”, was indeed the cause of gousiekte.
Two related species, also members of the
Rubiaceae, classified at the time as Van-
giieria zeyheri and V. chamaedendnim, were
tested and found to be non-toxic. The
history of these investigations is outlined by
Codd (1961).
Fig. 51. — Pachystigma pygmaeum, the “ Gousiekte-
bossie ”, flowering stage, November 1965,
Pretoria District.
48
ftLEiofi v.D. taeRwe,
Fig. 52. — Pachystigma pygmaeuni, the “ Gousiektebossie ”, {a) twig with buds; {b) flower and buds, both
natural size; (c) cross-section of ovary, X 10; {d) habit, reduced; — indicates ground
level.
49
The nomenclature of the dwarf Rubiaceous
shrublets was a source of considerable con-
fusion until Robyns (1928) produced a
satisfactory basis of classification. He places
the plants concerned in two genera as shown
below: —
1. Pachystigma — ovary 4-5-celled:
P. pygmaeiim (Schltr.) Robyns
P. thanmus Robyns
2. Pygmaeothamnus — ovary 2-3-celled:
P. charnaedendrum (Kuntze) Robyns
.... var. setuhsus Robyns
P. pilosus Robyns
P. longipes (Conrath) Robyns
P. zeyheri (Sond.) Robyns
.... var. oatesii (Rolfe) Robyns
.... var. livingstonianus Robyns
.... var. rogersii Robyns
The main criticism of the above classifi-
cation is that probably too many species and
varieties are upheld. It is now considered
that Pygmaeothamnus pilosus and P. longipes
should be reduced to synonyms of P.
charnaedendrum var. setulosus, while the P.
zeyheri complex should be reduced to two
varieties, a glabrous one, var. zeyheri, and a
pubescent one, var. rogersii. The genus Pyg-
maeothanmus is closely related to Canthium,
but it is felt that the former name could well
be retained for the dwarf, semi-submerged
shrublets listed above. It is of interest to
note that material of Pygmaeothamnus spp.
has repeatedly been subjected to investigation
but has not been found toxic.
From the early 1930’s evidence of stock
mortalities with typical gousiekte mani-
festations started coming in from areas where
close scrutiny of the veld failed to reveal the
the presence of Pachystigma pygmaeum.
Attention was extended to other Rubiaceous
plants but many years passed and much
frustration was experienced from negative
results before it was finally proved that
certain other genera were also involved.
One of the first examples is described by
Uys & Adelaar (1957) who proved that
Pavetta harborii was the cause of gousiekte
poisoning in the north-western Transvaal in
the neighbourhood of Ellisras and in sandy
areas to the north of the Soutpansberg.
Later, the plant responsible for mortalities of
the same nature in the central Waterberg and
Bronkhorstspruit Districts was shown to be
Fadogia monticola.
A description of these plants follows, but
the wider significance of these discoveries
should not be overlooked. Pavetta is a
genus of probably 300 species widely distri-
buted in Africa, Asia and Australasia. It
has since been shown that another species,
P. schumanniana, is also toxic. Fadogia
cienkowskii, a species closely related to F.
monticola, is distributed from the Transvaal
to Ethiopia. It seems highly probable,
therefore, that although gousiekte has not
been recorded from any other country but
South Africa, such mortalities have in fact
occurred in other countries, but have not been
diagnosed.
Descriptions of Species Causing
Gousiekte
Pachystigma pygmaeum {Schltr.) Robyns
Gousiektebossie.
(Synonyms: Vangueria pygmaea Schltr.;
V. setosa Conrath; V. nana K. Schum.).
Pachystigma pygmaeum is a perennial
shrublet with branches spreading under-
ground. During the winter nothing can be
seen of the plant but, about October or early
November, with the first good rains, short
leafy branchlets appear above ground from
the underground stems. These branchlets
are 8 to 15 cm high, bearing leaves which
are opposite, dull green to yellowish-green,
rough and pubescent (hairy), 4 to 10 cm long
and up to 3-5 cm broad, widest at or above
the middle, rounded, obtuse or abruptly
acuminate at the apex and tapering to a very
short petiole up to 3 mm long. The flowers
are small, pale yellow-green in colour, and
50
Map 4. — Distribution of Pachystigma pygmaeum, the
“ Gousiektebossie ”,
Fig. 53. — Pachystigma pygmaeum, fruiting stage.
are borne in clusters on shortly-branched
inflorescences (cymes) in the axils of the
lower leaves. The 5-lobed corolla lasts only
a few days, leaving the five narrow sepals
persisting at the apex of the developing fruit.
The calyx and corolla are pubescent on the
outside.
The ovary is normally 5-cel!ed with one
ovule in each cell but, as the fruit develops,
some of the cells may abort, often resulting
in only one, two, or three of the cells
producing seeds. The fruit is subglobose to
oblique, up to 2 cm in diameter, drupe-like,
with a thick layer of hard green flesh
surrounding a stony centre in which the seeds
are embedded. Because the number of cells
is a distinguishing feature between the genus
Pachystigma and the non-poisonous genus
Pygmaeothamnus (see further notes under the
next species on p. 53), it is important to
examine the ovary in the flowering or early
fruiting stage, before the cells abort, when
attempting to identify one of these Rubiaceous
shrublets.
P. pygmaeum is essentially a species of the
high-altitude, grassy plains, concentrated
mainly in the southern half of the Transvaal
from Lichtenburg and Ottoshoop in the west
to Nelspruit, Carolina and Piet Retief in the
east; there are outliers in the Waterberg
District and in the adjoining territories of
northern O.F.S., at Bergville in Natal, at
Mbabane in Swaziland, with a few records
near Salisbury in Rhodesia (see Map 4).
Even before it was proved to be poisonous,
P. pygmaeum was already known as the
Gousiektebossie (Rapid-sickness shrublet) by
farmers in the Pretoria District. Found in
dense grassland, often on red loam soils, it
is one of many perennial shrublets which
remain dormant and out of sight in the dry
winter months, sending up young green shoots
at about the time of the first good rains in
October and early November, usually before
the grasses have had time to put on growth.
It is, therefore, at this stage that livestock,
especially where green fodder has been in
short supply during the winter months, are
most liable to eat the first green material
available.
51
As explained earlier, there is a latent period
of two to six weeks between ingestion of the
Gousiektebossie and the death of the animal,
so that deaths from P. pygmaeum occur
mainly from the end of November to early
January. However, if a drought intervenes
in early summer, a second growth of P.
pygmaeum may occur in January or February,
causing much later deaths than usual. This
situation occurs periodically in the drier,
western Transvaal, for example in the summer
of 1957-58, when heavy gousiekte mortalities
were recorded in autumn in the Potchefstroom
District. The losses were more severe on
overgrazed veld where the grass cover had
been much reduced in relation to the P.
pygmaeum. During this outbreak it
appeared that animals which had received
supplementary feeding suffered a lower pro-
portion of losses than those which were
dependent entirely on the veld, while a sound
system of veld management, involving
rotation of grazing, was also instrumental in
reducing losses. It is also believed that the
toxicity of the plant may vary according to
the type of soil on which it is growing.
Pachystigma thamnus Robyns
Natal Gousiektebossie
Pachystigma thamnus is very similar to
P. pygmaeum. Like the latter species it is a
perennial shrublet with branches spreading
underground, dormant in winter and sending
up short leafy branchlets 3-5 to 10 cm high
in October to early November. The leaves
are soft or thick-textured, dark green and
glabrous, 5 to 12 cm long and 1-5 to 5 cm
broad, broadest at or above the middle, acute,
obtuse to rounded and abruptly apiculate at
the apex and tapering gradually or abruptly
into the short petiole about 3 mm long.
The flowers and fruits are almost identical
with those of P. pygmaeum, but the petals and
sepals are glabrous, not pubescent.
P. thamnus is the cause of Gousiekte in
northern Natal, the stock losses occurring
mainly in November and December, as for
P. pygmaeum. It is common in parts of the
Utrecht, Newcastle and Dundee Districts,
reaching its southernmost limit near Dann-
hauser. In these districts P. pygmaeum does
Map 5. — Distribution of Pachystigma thamnus.
Fig. 54. — Pachystigma thamnus, flowering twig,
reduced.
Fig. 55. — Pachystigma thamnus, flowering in the veld,
November 1965, Pretoria District.
52
«ue\OR v,T>. K\eRwe
Fig 56 — Puchystignia thauinus, (a) flowering twig, natural size; (b) cross-section through ovary,
(c) flower, X 21; (r/) habit, reduced; indicates ground level.
X 10;
53
not occur, but the two species may be found
together in the Transvaal from the Pretoria
District south-eastwards to Lydenburg,
Carolina and Piet Retief, and also in the
Waterberg area. P. thamnus is found in
dense grassland, usually on poor, sandy
soils in Natal, but may also occur on red
loam soils, as is the case with P. pygmaeiim.
Identification of Pachystigma pygmaewn
AND P. thamnus
Difficulty is often experienced in distin-
guishing with certainty between the two
Gousiektebossies on the one hand, and the
similar-looking species and varieties of Pyg-
maeothamnus, none of which are regarded
as being poisonous. The problem concerns
the separation of the following entities,
which may be placed initially in three groups
as shown below: —
1 . Pachystigma thamnus (leaves glabrous)
and P. pygmaeum (leaves pubescent).
2. Pygmaeothanmus chamaedendnim var.
chamaedendrum (leaves glabrous) and
P. chamaedendrum var. setulosus
(leaves pubescent).
3. Pygmaeothanmus zeyheri var. zeyheri
(leaves glabrous) and P. zeyheri var.
rogersii (leaves pubescent).
The third group usually presents no
difficulty because the two varieties of Pyg-
maeothamnus zeyheri (commonly known as
the “ goorappei ”) have distinctly larger
leaves (5 to 17 cm long) than the members of
the two other groups, while the leaves are
broadest well above the middle (oblanceolate
to obovate), and taper gradually to a petiole
6 to 10 mm long. Therefore this group
need not be considered further, although in
the past P. zeyheri var. rogersii, which occurs
in great quantities on the Waterberg, was
thought to ht Pachystigma pygmaeum because
of the frequent occurrence of gousiekte in this
area (see note under Fadogia monticola on
p. 57).
The first two groups can be readily
separated if flowers or young fruits are present
because Pachystigma has 4 to 5-celled
ovaries, while Pygmaeothanmus has 2 to
3-celled ovaries. When flowers or young
fruits are not present, however, considerable
difficulty may arise in deciding betVv'een the
following pairs:—
(a) Pachystigma pygmaeum and Pyg-
maeothanmus chamaedendrum var.
setulosus, both of which have
pubescent leaves.
(h) Pachystigma thamnus and Pyg-
maeothanmus chamaedendrum var.
chamaedendrum, both of which have
glabrous leaves.
The distinction between the members of
each pair is very slight and not easy to express
in words. The two pubescent entities can
usually be distinguished by the character of
the pubescence. In Pachystigma pygmaeum
the pubescence on the leaves is yellowish and
tends to be denser along the midrib and
around the edge of the leaf where it forms a
yellow “ halo ”, which can be seen when the
leaf is held up to the light; in Pygmaeo-
thanmus chamaedendrum var. setulosus the
pubescence is white to off-white and varies
from sparsely to densely pubescent.
The two glabrous entities are even more
difficult to separate satisfactorily and it is
often impossible to give a reasonably certain
decision when flowers or young fruits are
absent. Pachystigma thamnus has leaves
which are about 3 to 4 times as long as broad,
the leaves are somewhat roughish and the
tertiary veins on the underside are slightly
raised ; Pygmaeothanmus chamaedendrum
var. chamaedendrum, on the other hand, has
leaves which are 4 to 5 times as long as
broad and which are very smooth on the
surface, while the tertiary veins on the under-
side are relatively inconspicuous. In
northern Natal, where both occur, the Zulus
distinguish between them and state that the
fruits of P. chamaedendrum var. chamaeden-
drum are edible, without ill effects, while
those of Pachystigma thamnus cause deafness
if eaten.
Identifying a specimen on the basis of
locality is not always possible because the two
pubescent elements may occur together in the
54
PiG. 57. — Pavetta harborii, (a) fruiting stem; (/)) flowering branch, both natural size; (c) flower, X 2;
(r/) cross-section through ovary, X 10.
55
Johannesburg, Pretoria, Belfast, Ermelo and,
possibly, Bergville districts, while the distri-
butions of the two glabrous elements overlap
in the Pretoria, Lydenburg, Newcastle and
Dundee Districts.
Pavetta harborii S. Moore
Cases of Gousiekte in the neighbourhood of
Ellisras in the Waterberg District and north
of the Saltpan in the Soutpansberg District
were reported over a number of years and
yet a study of the veld revealed no trace
of Pachystigma. Suspicion was directed
towards Pavetta harborii, the commonest
Rubiaceous plant present and, as mentioned
earlier, this species was eventually proved
capable of causing Gousiekte by Uys &
Adelaar (1957). Mortalities caused by
Pavetta harborii usually occur from March
to May.
Pavetta harborii is a perennial shrub,
branching below ground level, producing
groups of stems 20 to 45 cm tall, often in
patches of 6 ft or more in diameter. The
stems have creamy-white bark, are woody,
and persist through the winter, although the
leaves are shed. The leaves are opposite,
sessile, narrowly elliptical, grey-green and
sparsely hairy on the upper surface and
densely white-felted below; like all Pavetta
species, there are several opaque, thickened
dots scattered on the surface of the leaf.
Opposite clusters of white, sweet-scented,
long-tubed, 5-lobed flowers, with long-
exserted styles, appear on the old wood in
early summer and are followed by groups of
small round berries, the size of a small pea,
green at first, with a black, wrinkled skin
when ripe.
As shown on the accompanying map,
P. harborii is distributed from eastern
Bechuanaland, through north-western Trans-
vaal to northern Transvaal, occurring always
on deep, sandy soil varying from whitish to
pale red in colour. The species is named
after the collector, Cyril Harbor, son of
William Harbor who had a hotel and store
at Mochudi Station (now Molotoane Siding)
some eight miles north of the present
Mochudi village. William Harbor also
collected plants, which were usually handed to
the Rev. Rogers’ A visit to the type locality
revealed that William Harbor is buried at
Molotoane Siding and his grave stands in a
small enclosure surrounded by numerous
large patches of Pavetta harborii.
Fig. 58. — Grave of William Harbor at Molotoane
Siding, Bechuanaland, the type locality of
Pavetta harborii, with a servant who
worked for him. Photo by L. E. Codd,
1958.
56
RLElDfl W.D l^eJiue
PiG. 59. — Fadogia monlicola, (a) habit, X -1^; (b) flowering stem; (c) fruiting twig, both natural size;
(r/) cross-section of ovary, X 2; (c) flower, x 2.
57
Pavetta schumanniana F. Hoffm. ex K. Schum.
It may be expected from the foregoing that
other species of Pavetta may be found capable
of causing Gousiekte poisoning. P. schuman-
niana is one that has proved to be poisonous
at the Veterinary Research Institute,
Onderstepoort. P. schumanniana is a deci-
duous much-branched shrub 2 to 8 ft tall
with thickish, woody branches. The leaves
are large, obovate, 8 to 13 cm long and 2-5
to 5 cm broad, leathery, conspicuously net-
veined on the underside, and with the usual
opaque dots. Small, white, sweet-scented
flowers are produced in clusters in early
summer usually on the previous season’s
wood, followed by small berries which are
black when ripe.
It has a wide distribution from tropical
central Africa to northern South West Africa,
Bechuanaland, northern and eastern Trans-
vaal, reaching its southernmost extent in
northern Zululand.
Fadogia monticola Robyns
WiLDEDADEL (WiLD DATE)
(Synonyms: F. oleoides Robyns; F.
fragrans Robyns)
For many years cases of Gousiekte have
been known from the Waterberg plateau and
from the Bronkhorstspruit and neighbouring
districts, associated with the so-called “ sour
busveld ”, i.e. tree species such as Burkea
africana and Ochna pidchra with a ground
cover of wiry (sour) grasses. It is common
practice to burn this veld annually to remove
the hard, wiry grasses with the result that
there is an upsurge of numerous dwarf
shrublets and herbs with the first rains.
Among these Pygmaeothamnus zeyheri var.
rogersii is usually a common constituent and
was originally misidentified as the Gousiekte-
bossie (Pachystigma pygmaeuni). As early
as 1944, attempts were made to eradicate this
plant in the neighbourhood of Twentyfour
Rivers by means of various chemicals, but
without success. Intensive feeding trials with
this and other species were carried out by the
Veterinary Research Institute and, eventually,
it was demonstrated that Fadogia monticola
was the cause of Gousiekte in these areas,
while Pygmaeothamnus zeyheri var. rogersii
gave consistently negative results.
Fadogia monticola has a perennial rootstock
with short branches below ground, producing
clusters of slender erect stems 15 to 20 inches
tall. The stems are yellowish and produce
whorls of 3 to 5 leaves at intervals up the
stem. The leaves are shiny, dark green and
glabrous on the upper surface and greyish-
white felted below. Small clusters of
yellowish-green flowers are produced in the
axils of the leaves and are followed by globose
berries, purplish-black when ripe, containing
a sappy pulp. It is from a superficial
resemblance which the fruit bears to a small
date that the common name “ wildedadel ”
(Wild Date) is derived.
Fig. 60. — Pavetta harborii, flowering in November at
Molotoane Siding, Bechuanaland.
58
Map /. — juistribution of Fadogia monticola.
It is a common species in parts of the
Transvaal, especially on white, grey or pate
red sands, extending from Pretoria eastwards
to Swaziland and northwards to Waterberg,
Potgietersrust, Soutpansberg and Sibasa
Districts, and into Rhodesia, Zambia and
Mozambique.
References
CoDD, L. E., 1961. Notes on poisonous
plants, with special reference to the
Gousiekte problem. Journ. S. Afr. Biol.
Soc. 2, 8-17.
Robyns, W., 1928. Monographiae Van-
gueriae. Bull. Jard. Bot. Brux. 11, 1-359.
Theiler, a., DU Toit, P. J. & Mitchell,
D. T., 1923. Gouwziekte in sheep.
Ninth and 10th Rep. Dir. Vet. Ed. and
Res. 9-108.
Uys, P. L. & Adelaar, T. F., 1957. A new
poisonous plant. Journ. S. Afr. Vet.
Med. Assoc. 28, 5-8.
Fig. 61. — Fadogia monticola, flowering in November 1965, Pretoria District.
59
Contemporary Botanists
and
Botanical Collectors
1. FREDERICK ZIERVOGEE VAN DER MERWE
by I. C. Verdoorn
‘4
^ \ . J
Fig. 62. — Dr. F. Z. van der Merwe, 1965.
pREDERICK ZIERVOGEE VAN DER
MERWE was born in Stellenbosch on
the 10th December, 1894. He went to school
at Beaufort West and later at Boshof in the
Orange Free State. Already at an early age
he took an interest in natural history and time
has shown that his approach to nature was
scientific as well as aesthetic. In 1913 he
obtained the B.A. (Classics) degree at Grey
College, Bloemfontein; qualified as a doctor,
M.B., B.Ch., B.A.O. at Trinity College,
Dublin, in 1920; in 1921 was awarded the
D.T.M. at Eiverpool University, England;
and in 1930 he got his D.P.H. at theUniversity
of the Witwatersrand, Johannesburg.
As a Medical Inspector of Schools,
Dr. van der Merwe travelled widely and
often to places off the beaten track. This
provided him with the opportunity to study
the varying forms of plant-life in their natural
habitat and their relationship to the environ-
ment, a subject in which he was particularly
interested. Wherever he went, no matter in
which Province, he found interest and beauty
in nature. In a letter, he once wrote: “ I
should have liked to rhapsodise on the
unspoiled natural beauties of the Transvaal,
which I was priveleged to see, before they
started to disappear ”.
60
Dr. van der Merwe’s contact with the
National Herbarium commenced in the late
1920’s when he brought or sent in Aloes to
Dr. Pole Evans for naming. At that time
Dr. Pole Evans was the authority on South
African Aloes and he had recently described
several new species from the Transvaal.
Later on, various species of Aloe, recognized
by Dr. van der Merwe as new, were described
by B. H. Groenewald; some in Tydskrif vir
Wetenskap en Kims and some in Flowering
Plants of Africa.
In the course of time Dr. van der Merwe,
under the protection (as he puts it) of the
Division of Botany, increased his knowledge
Fig. 63. — Dr. F. Z. van der Merwe with a specimen
of Euphorbia piilviuata, 1939.
of our indigenous plants and their classifi-
cation and he has contributed more than
2,500 specimens to the National Herbarium.
In particular, the genus Scilla and related
groups claimed his attention. In the process
of his studies he published an account of 34
species of Scilla and related genera of which
not less than sixteen species and two genera
were new to science, with Dr. van der Merwe
as author. Most of these were published in
Flowering Plants of Africa (in volumes 21,
22, 23 and 24), while in Tydskrif vir Weten-
skap en Kims (part 6) he published
“ Aantekeninge vir die hersiening van die
genus Scilla L. in Suid-Afrika ” and “ ’n
Nuwe Genus: Resnova His numerous
specimens, with valuable field notes and
personal observations, which are now in the
National Herbarium, have contributed greatly
to our knowledge of this genus. Dr. van der
Merwe’s wide knowledge of the Transvaal
veld was of material benefit to Dr. Dyer
during several field expeditions in 1939.
These were mainly concerned with the
gathering of information and photographs
for the two handsome volumes on the
Succulent Euphorbieae {Southern Africa) by
White, Dyer & Sloane, which appeared in
1941. He is commemorated in Euphorbia
vandermerwei R. A. Dyer, Aloe vandermerwei
Reynolds and Eucomis vandermerwei Ver-
doorn.
Dr. van der Merwe’s other interests include
Africana, especially printed South African
music and Afrikaans glossaries of terms,
particularly medical terms. Besides the many
botanical works mentioned above, his publi-
cations include Suid-Afrikaanse Musiek-
bibliografie 1887-1952, Mediese Woorde-
boelc (joint author, J. D. Louw), a trans-
lation of a book by Scoles and several
Afrikaans booklets on health. Dr. van der
Merwe is a faculty member of the South
African Biological Society and of the Suid-
Afrikaanse Akademie vir Wetenskap en
Kuns. In 1960 he retired on pension and is
now living in Durban, where he is continuing
with certain research work in connection with
iiis hobbies.
61
2. LESLIE CHAREES LEACH
by L. E. Codd
J ESLIE (“ Larry ”) CHARLES LEACH
^ was born at Southend, Essex, on 18th
November 1909 and was educated in Eondon,
after which he started on an engineering career,
concerning himself mainly with the electrical
equipment of motor vehicles and aircraft.
In 1938 he emigrated to Rhodesia and settled
in Salisbury.
During 1956-59, collecting was carried out
mainly in Rhodesia, South Africa and the
southern half of Mozambique, with one
longer expedition to Malawi, when Zomba
and the Mlange Mountains were visited.
By this time Mr. Eeach had acquired a fairly
clear grasp of the Euphorbia species of the
His interest in succulent plants began in
1950 after he had purchased a 10-acre plot,
which he named “ Farview ”, in the suburb
of Greendale. At first his enthusiasm led
him to cultivate a wide variety of plants, but
gradually his attention centred on certain
groups, including the Stapelieae, Euphorbieae
and Aloe.
It was not until 1956, after he had disposed
of his business in Salisbury, that his interest
in the taxonomy of the Euphorbieae began
to predominate; in that year he visited the
eastern Transvaal and southern Mozambique
in order to compare Euphorbia confinalis
R. A. Dyer with closely related plants which
he had studied at Chibi and Bikita in
Rhodesia. At the same time he acquired a
25-morgen plot of ground near Nelspruit.
This initial essay into the taxonomic field,
which had satisfied him that the Rhodesian
plants were distinct from E. confinalis, was
soon followed by a number of expeditions to
investigate and collect the species of the
Elora Zambesiaca area, and to see as
many as possible of the succulent Euphorbias
at their type localities. In the course of
these journeys collections were made of both
living plants and herbarium material, the
latter being presented mainly to the Govern-
ment Herbarium, Salisbury, as well as to the
Royal Botanic Gardens, Kew, the National
Herbarium, Pretoria, and the Herbarium of
the Centro de Botanica, Lisbon. Living
plants were cultivated and studied at
“ Farview ” and herbarium specimens were
prepared when the plants flowered.
Fig. 64. — Mr. L. C. Leach in the National Her-
barium, Pretoria, 1965. By permission of
the ,, Pretoria News ”.
1
62
Zambesi region and was convinced that
several of them were undescribed. He had
also developed a sound knowledge of the
Stapelieae and recognized several problems
that needed working out. In the course of the
next few years several expeditions with definite
objectives were carried out, as indicated below.
1960. Bechuanaland Protectorate was
first visited to study E. griseola and, later,
Mozambique to collect, study and photograph
the species later described as E. graniticola,
E. livicliflora and others. Similarly, on the
Zambesi escarpment, the new species E.
decidua and E. persisteiitifolia were studied in
their natural habitat. A tour was also made
to East Africa via Zambia to see certain
published species at their type localities,
visiting Abercorn, Kalambo Falls, Lake
Tanganyika, Mbezi Forest overlooking Lake
Rukwa, Morogoro, the Usambara and Pare
Mountains, Mt. Kilimanjaro and Lake Chala,
Mombasa, Voi, Kibwezi, etc. On the return
journey Malawi was traversed from north to
south.
1961. A special journey was undertaken
to study E. angidaris at its type locality on
Goa Island, off the cost of northern Mozam-
bique but, unfortunately, the plants found
were sterile at the time. Among the other
collections made en route were two new
species, subsequently described as Hiiernia
erectiloba and Euphorbia contorta. In the
same year the type locality of the lost E.
johnsofui was finally located on the Mozam-
bique coast, just north of the Save River.
This expedition also yielded the type material
of E. ambroseae, found growing in the shade
of bush groups situated on flat, swampy
country.
1962. An expedition was undertaken to
northern Mozambique, with the Namuli
Peaks as the main objective. Dr. E. A.
Schelpe, Curator of the Bolus Herbarium,
accompanied him in order to collect ferns and
orchids and thus fill in gaps in their distri-
bution records for the Flora Zambesiaca
area. The results were of considerable value
as the total number of ferns known from
Fig. 65. — Mr. Leach with his newly described species, Euphorbia persisteiitifolia.
63
northern Mozambique was increased from
55 to 80, including a new species in each of
the genera Cheikmthes and Notholaena.
Among the flowering plants, several
undescribed species were found, including
Dombeya leachii Wild and two new orchids
which remain to be described.
1963. The Sul do Save district of
Mozambique was explored and many
interesting collections were made, including
the new species E. baylissii Leach, named
after Col. R. D. Bayliss, formerly of Kenya,
who was with him on this occasion. Living
plants of Huernia kirkii contributed to the
solution of the problems concerning this and
allied species.
1964. In this year Goa Island was again
visited ; this time E. angidaris was found both
in flower and fruit, and adequate herbarium
material was prepared. Some additional
collections were made on the mainland which
are expected to yield further undescribed
plants. Visits were again paid to Sul do Save
and to Bechuanaland Protectorate where
living plants of Euphorbieae and Stapelieae
were made for further study in cultivation at
Nelspruit.
1965. A journey was made to South West
Africa as far north as Ovamboland, again
accompanied by Col. Bayliss, in order to
study E. monteiroi and the CaraUuma vaga
complex.
During the last few years Mr. Leach has
gradually been transferring his plants from
Salisbury to Nelspruit, where he proposes
to make his home. He is a member of a
number of scientific societies and associa-
tions, in several of which he plays an active
part. He was a committee member of the
Botanical Society of Southern Rhodesia for
a number of years and is a member of the
Royal Horticultural Society, the Botanical
Society of South Africa, Rhodesian Orchid
Society, the South African Association for the
Fig. 66. — An undescribed Euphorbia, related to E. confiualis, at Chibi, S. Rhodesia. (Photo by L. C. Leach).
64
Advancement of Science, the International
Association of Plant Taxonomy and the
Association pour I’Etude Taxonomique de la
Flore d’Afrique tropicale, to mention the
more important. For a short while he acted
as secretary of the “ Herbarium and Botanic
Garden Fund ”, aimed at the development of
the Alexandra Park Botanic Garden and
Herbarium, Salisbury, and served as a
member of the Liaison Committee.
He is commemorated in the species
Huernia leacliii Lavranos, Aloe leachii Rey-
nolds, Dombeya leachii Wild, Eulophia leachii
Greatrex ex Hall and Notholaena leachii
Schelpe. His published work is concerned
mainly with Euphorbia species of the Elora
Zambesiaca area, in which genus he has
described the following species: E.
lividijiora, E. graniticola, E. contorta, E.
ambroseae, E. bougheyi, E. baylissii, E. male-
vola, E. jubata and E. fortissima. With Mr.
Lavranos he is co-author of Huernia erectiloba
Leach & Lavranos.
Fig. 67. Edithcolea graiidis, X collected by Dr. P. J. Greenway (No. 4456) in Kenya and cultivated
in Pretoria.
65
Vegetative Propagation of Proteaceae and Ericaceae
by J. Admiraal
Interest in the South African Proteaceae
and Ericaceae is increasing daily, both here
and overseas. Their value as cut-flowers is
widely appreciated and, in this connection,
it has repeatedly been stressed that factors of
importance to the trade are uniformity of
colour and size of bloom, and length of stem.
At present, propagation is almost entirely
by means of seed, an unsatisfactory pro-
cedure as far as commercial production is
concerned because a high degree of unifor-
mity cannot be ensured, while seeds of many
kinds are difficult to obtain in sufficient
quantity. Furthermore, occasional hybrids
with desirable cut-flower potential cannot be
perpetuated by means of seed.
Authoritative publications usually state
that Erica cuttings may be rooted in 4 to 6
weeks, while members of the Proteaceae take
up to three months to root, when about
50 per cent success or less may be expected.
Experiments were, therefore, conducted on
the Pretoria National Botanic Garden to
improve on present methods of vegetative
propagation.
In view of the many factors involved, a
series of preliminary trials was carried out
using various combinations of {a) soil and
air temperatures, {h) sunlight, (c) mist-sprays
for various durations, and {d) pH and com-
position of rooting medium. Other factors,
such as time of year, ripeness of wood and
size of cuttings, were also taken into con-
sideration. At the higher temperatures and
humidity, at which rooting was most success-
ful, fungus diseases were at first troublesome,
but have now largely been overcome.
The preliminary results have proved so
promising that a short note at this stage
seems justified. Species such as Leuco-
spermmn reflexian, L. nutans, Mimetes lyrigera
and Serruria florida started rooting in 2 to
3 weeks and in some cases up to 100 per cent
of the cuttings had rooted in six weeks. The
Fig. 68. — Experiment with cuttings of Proteaceae.
66
Protea species tested did not do so well, but
this is attributed to the wood not being at the
right stage of ripeness. The best results so
far have been obtained with a rooting medium
of pH 3-9 held at 25° C, with the air tem-
perature at 30° C, and the application of a
mist-spray for one minute in every ten
during the hours of daylight. Under this
treatment, Erica cuttings usually started
rooting in 10 to 14 days with up to 100 per
cent rooting in four weeks.
Subsequent handling also requires further
experimentation. At first many plants were
lost after potting-up, but losses have been
greatly reduced by introducing a more
gradual transition from the high tempera-
tures of the rooting chamber to air tempera-
ture in the shade house.
Some of the results are illustrated in the
accompanying photographs. A general view
of part of the experiment is shown in Fig. 68.
Mimetes lyrigera and Serruria florida are
shown in Fig. 69. Both species were
put in the rooting medium on 16/7/65 and
after 14 days the first mentioned started pro-
ducing roots. The photographs were taken
after 26 days, on 12/8/65, when rooting had
made good progress in Mimetes lyrigera.
In this species 30 out of 50 cuttings had
rooted in 14 days and the remainder rooted
within four weeks. In the case of Serruria
florida, only four cuttings were available and
all four had rooted after 23 days, on 9/8/65.
Figs. 72 and 74 show plants of each species
grown from these cuttings, ready for planting
out on 10/2/66, approximately seven months
after putting in the cuttings.
Fig. 69. — Cuttings of Mimetes lyrigera (left) and Serruria florida, showing root
development after 26 days.
67
Fig. 70. — Erica tenella (left) and E. oatesii, cuttings
showing rooting medium sticking to the
roots after 7 weeks.
Fig. 71 shows a cutting of Leiicospennum
reflexmn starting to form roots at 15 days
old. Of 45 cuttings put in on 2/8/65, 15 had
started to root on 17/8/65 when this photo-
graph was taken. After 30 days. 100 per
cent had rooted. Subsequently, with
improved technique, 50 cuttings of Leucosper-
miiin nutans were put in on 1 5/10/65 and all 50
had started rooting in 14 days. After one
month, on 15/11/65, they were transplanted
into individual containers.
Equally good results were obtained with
species of Erica. Cuttings of E. oatesii, E.
tenella and E. hirtiflora were put in on
27/6/65 and root formation started in 14 days.
Fig. 70 was taken on 12/8/65, after nearly
seven weeks, when the cuttings were well
rooted and almost ready for transplanting.
Mature plants of E. tenella and E. hirtiflora
grown from these cuttings were also photo-
graphed on 10/2/66 and the former is shown
m Fig. 73.
Fig. 7 1 . — Leiicospennum refle.xiim, cutting starting to
root after 15 days in the rooting medium.
68
Fig. 72. — Sen iin'a florida.
Fig. 73. — Erica tenella..
Fig. 72 to 75. — Plants approximately seven months'old, grown from cuttings and ready for planting out.
Fig. 74. — Minie/es lyrigera.
Fig. IS.—Leucospermum nutans.
69
Fig. 76. — Felicia ecliinata, an attractive perennial blue-flowered daisy, cultivated on the Botanic Garden,
Pretoria.
T H B ( . C) V I- K N M B N T P R I N T B R . PRETORIA
(PRETORIA)
CALEDON STREET
OASPOORT