MARY
GUNN LIBRARY
5000017916
View Nation1
,k.. Tosiiiuie
Volume
BOTHALI A
9 Jaargang 9
Published by the
Botanical Research Institute
Department of Agricultural Technical Services
Gepubliseer deur die
Navorsingsinstituut vir Plantkunde
Departement van Landbou-Tegniese Dienste
> m. VAN PMTf-
\ MMN- hitllTIlUT VIR PLANTKUNDE
Edited bv — Onder redaksie van
L. E. Codd. D.Sc.,
EDG/SOX 994
\t -1
prRpE-T CRIA-
REo/ INSTITUTE
Chief : Botanical Research Institute, Pretoria
Hoof : Navorsingsinstituut vir Plantkunde, Pretoria
THE GOVERNMENT PRINTER-DIE STAATSDRUKKER, PRETORIA
G.P.-S. 36919-1968-69-1,000
Digitized by the Internet Archive
in 2016
https://archive.org/details/bothaliavolume909unse
CONTENTS - INHOUD
Page / Bladsy Date/ Datum
Part 1 1-250 30 November 1966
Part 2 251-362 4 August/Augustus 1967
Parts 3 & 4 363-513 7 May/Mei 1969
Page
Bladsy
Anderson, J. G. The genus Andropogon in Southern Africa 5
A new combination in Hyparrhenia 130
A new species of Panicum 341
A new variety of Festuca costata 341
Badenhorst, H. Melastomaceae: a note on Memecylon australe Gilg & Schltr.
and M. grandiflorum R. & A. Fernandes 345
Codd, L. E. The Cassine complex 123
A new species of Canthium 345
New species of Kniphofia 139
Notes on Boerhavia in Southern Africa 113
The South African species of Kniphofia 363
The status of the genus Lydenburgia 124
De Winter, B. A new species of Eragrostis 137
Polexansia De Winter, gen. nov 130
Styppeiochloa De Winter, gen. nov 134
Gunn, M. D. Ethel Mary Doidge (1887-1965) 251
Hilliard, O. M. A new species of Streptocarpus ( Gesneriaceae ) 335
Jessop, J. P. The genus Asparagus in Southern Africa 31
Killick, D. J. B. Two new Erica species from the Drakensberg 126
Klesser, Patricia J. A virus disease of Datura spp 173
Leistner, O. A. The typification of Podocarpus elongatus 142
Marais W. Notes on South African Cruciferae 97
Marasas, W. F. O. New species of Ascomycetes and a new genus of Sphaeropsi-
daceae from Transvaal 203
Marasas, W. F. O. & Rabie, C. J. New and interesting records of South African
Fungi. Part IV 217
Marasas, W. F. O. & Schumann, I. H. Two species of Erysiphaceae from Pre-
toria 245
Page
Blaclsy
Marasas, W. F. O., Van der Westhuizen, G. C. A., Van Warmelo, K. T. &
Papendorf, M. C. New and interesting records of South African Fungi.
Part V ~. 229
Marsh, J. A. Notes on Widdringtonia 124
Obermeyer, A. A. The identity of Bidbinella carnosa 345
The identity of Frit Maria nana 345
Mrs. Reino Pott-Leendertz (1869-1965) and the Transvaal Museum Herbarium 1
A new combination in Ornithogalum 344
A new combination in Trachyandra 344
Two new species of Bulbine 342
Raven, P. H. A revision of African species of Epilobium ( Onagraceae ) 309
Simons, R. H. The genus Ceramium in South Africa 153
The genus Microcladia (red algae) in South Africa 351
A new species of brown alga from South Africa 357
A new species of Dictyotales from South Africa 169
Tolken, H. R. The species of Arthrocnemum and Salicornia ( Chenopodiaceae )
in Southern Africa 255
Van Warmelo, K. T. The cytology of Mycosphaerel/a pinodes 195
The identity of a pea blight fungus in South Africa 183
Verdoorn, I. C. A new species of Dombeya and a new variety of Dombeya
rotundifolia 143
Plate 1. — Mrs. Reino Pott-Leendertz (1869-1965); photograph taken about 1910
7230691
Botha Ha 9, 1 : 1-3.
Mrs. Reino Pott-Leendertz (1869-1965)
and the
Transvaal Museum Herbarium
by
A. A. Obermeyer
Reino Leendertz, the founder of the Transvaal Museum Herbarium, was born
in Medemblik, Holland, on the 2nd January, 1869. She qualified as a pharmacist
in 1891 but then turned her attention to Botany. At Amsterdam she studied under
Professor Hugo de Vries and afterwards became one of his assistants. In July, 1898,
she arrived in Pretoria and was appointed Botanist to the State Museum on the 1st
September, 1898. In her letter of appointment the Director, Dr. J. W. B. Gunning,
mentioned that, although it was contrary to the policy of the Committee to appoint
women on the scientific staff, her exceptional qualifications decided them to waive
this objection. Thus she became the first official botanist of the Transvaal as well as
its first woman civil servant. A large collection of plants had been purchased from
Rudolf Schlechter and these formed the nucleus of the Transvaal Museum Herbarium.
The State Museum was then housed in a building on the Market Square in Pretoria,
the part that is presently used as the covered market. This building, which was rented
for £10 per month, was officially opened by President Paul Kruger in May, 1894. It
soon proved to be too small and plans were made to put up a new building in Boom
Street, where the Zoological Garden was being constructed. A photograph of the
stone-laying ceremony for this new building, on the 22nd July, 1899. shows Miss
Leendertz as the only woman among the museum officials.
War broke out and Pretoria came under the English in 1901. The British com-
mander asked her to stay, but Miss Leendertz decided to return to Holland. At a
girls’ high school in Amsterdam she taught Natural History for a year but she did not
like teaching (in after years she confided to a niece that she had difficulty in keeping
order). So she re-applied for her former post at the museum, now called the Pretoria
Museum, and was accepted immediately, since the Director had a high opinion of
her capabilities. She returned to Pretoria in 1904.
She threw herself with heart and soul into the work and started a special Transvaal
herbarium. She collected extensively in the Transvaal, her numbers eventually reaching
nearly 6,000. Her first excursions were done by bicycle. In the beginning a Mr.
B. H. Rademacher and his sister accompanied her; later Mr. A. T. J. Janse (now
Professor Janse), who was on the staff of the Normal College, became her enthusiastic
companion during week-end excursions. At one stage they decided to make an exhibit
of the Transvaal timber trees. The trees were photographed; wood, bark, flowering
branches and fruits were all loaded onto the bicycles and taken home. Mr. Haak Snr.
who was the owner of the bicycle shop (which afterwards became Haak’s Garage),
once happened to see such a fully loaded bicycle and remarked that they were not
meant to be used as wagons.
Quite often she used her holidays to go on collecting expeditions further afield.
Gatherings from Ermelo (with a beautiful kloof, Mavieriestad, nearby), Zeerust, Bethal,
Potchefstroom, Standerton, Vereeniging, Moorddrift (near Potgietersrus), Heidelberg
and some spots nearer Pretoria (Daspoort. Kameeldrift, Roodeplaat. Irene) are localities
regularly met with in the herbarium.
2
The collection now expanded by leaps and bounds as can be seen from the
following extract from the Annual Report of the Transvaal Museum, 1905-1906:
“ I collected in the neighbourhood of Pretoria 84 species of plants in many duplicates,
nine species of wood, many fruits and seeds to complete the collections. On a botanical
excursion to the Zoutpansberg District, where I stayed at Marabastad, Waterval,
Rietfontein and Eersteling — only a fortnight — I collected 105 species of plants in many
duplicates. Most of them were new to the collection “ Transvaal Plants ”. On a five
days’ excursion to Crocodile River and Skeerpoort, I collected 68 species of plants
of which 50 species were new for the Transvaal collection: and on one to Rooiplaat
(Roodeplaat), Derdepoort, also of five days, 62 species of plants of which 40 species
were new to the collection ”.
In about 1906 the Reverend F. A. Rogers of the South African Church Railway
Mission is also mentioned by her as an assiduous collector, who regularly sent her
large parcels of plants from different localities. The plants were just put between
sheets of paper by him and Miss Leendertz did the rest, the pressing, naming, etc.
She also made duplicate sets for the collector. Other collections, which were purchased
by the Museum, were a collection of Ferns and Fern Allies from Mr. T. R. Sim. Miss
Leendertz pointed out that it was a standard collection, Mr. Sim having based his
book “ The Ferns of South Africa ” on it. Further additions were collections purchased
from Mr. G. N. Thorncroft, of Barberton, and the missionary, Henri Junod of Shilouvane
(near Tzaneen).
No wonder that soon the universal complaint was heard; the boxes housing the
collections “ were filled to the top ” and that she had no assistant to help her with the
pressing, mounting and labelling. The neat sheets with her copy-book handwriting
are a joy to behold. She was an extremely good taxonomist and hardly ever made
mistakes in her identifications although the literature at her disposal in those days
was limited.
In her next report she informs us that a junior assistant was appointed in September,
1907, and that new boxes had arrived. Incorporated in the herbarium that year were
2,436 species of plants and she remarked that “ many families are now fairly well
represented and form useful material for study and reference ”. She then again refers
to the Ferns and Fern Allies collection which through additions from Junod, Burtt
Davy and others had become one of the best in South Africa. She also thanks Dr.
Harry Bolus for his co-operation. In conclusion she mentions that the “ Preliminary
List of Transvaal Vascular Plants which Mr. J. Burtt Davy and myself have prepared ”,
had been read as a paper at the Grahamstown Meeting of the South African Asso-
ciation for Advancement of Science.
She also spent a good deal of time on the preparation of a botanical exhibition,
filling 8 show cases with examples of flowering timber trees of the Transvaal with their
Latin and vernacular names, and 16 other cases “ illustrating the biology of plants,
e.g. insectivorous plants, parasites, poisonous plants, etc., all provided with descriptive
labels in both languages ”,
In April, 1908, the Director, in a report presented in the first issue of the “Annals
of the Transvaal Museum ” informed us that the Herbarium contained 4,000 specimens.
In 1915 the number of sheets catalogued had risen to 14,400 and when she left to go on
pension on the 2nd January, 1925, the last number entered by her was 23,255.
Before that, in 1912, the Natural History and Geological collections of the Transvaal
Museum had been moved to a new building in Paul Kruger Street.
Staff members of the Museum, when on field trips, added many plants to the
herbarium. We frequently meet the names Dr. H. G. Breyer (at one time a Director
of the Museum) and his son, when the latter was stationed at Fort Namutoni in South
3
West Africa; C. Wilde, T. J. Jenkins, C. J. Swierstra, V. Fitzsimons, G. van Dam.
J. Hewitt, M. Collins, Austin Roberts, were industrious collectors. The Maputaland
Expedition, which visited Mozambique in 1914, also yielded a number of plants.
Names of well known South African collectors are also represented in the herbarium,
such as Wilms (apparently his early gatherings were not numbered), F. A. Rogers,
G. Thorncroft, H. Junod, W. Nelson, W. A. H. Harbor, M. Stewart, R. D. Bradfield,
R. E. Symons, J. Medley Wood, Alice Pegler, Florence Paterson, Daly & Cherry,
H. G. Flanagan, P. Krook, A. Penther, W. Tyson, Barbara Smart, Theo Kassner,
A. Stolz, etc.
Together with other well known scientists such as Sir Arnold Theiler, Prof. A. T. J.
Janse and Dr. I. B. Pole Evans, Miss Leendertz was a foundation member of the
Biological Society which was started in Pretoria in 1907. In 1909 she married Mr.
N. W. Pott.
From the time of her retirement in 1925, her interest in botany appears to have
ceased but to the last she remained active and alert and a real lady, much loved by all
around her. After a short illness she passed away on the 14th June, 1965, at the age
of 96.
A few plants were named after her, the best known being Stapelia leendertziae.
There is also a Blepharis leendertziae Oberm., Delosperma leendertziae N. E. Br.,
Indigofera leendertziae N. E. Br., Lotononis pottiae Burtt Davy and Thesium pottiae
N. E. Br. A list of her publications is appended.
Upon her retirement. Miss D. Kaplan succeeded her in 1925. In 1927 Mr. R. G. N.
Young took over and in 1929 Miss A. A. Obermeyer, who remained there until her
marriage in 1938. In 1953 the Board of Trustees of the Transvaal Museum donated
the herbarium, now consisting of about 40,000 numbers, to the National Herbarium,
Pretoria.
Acknowledgements
It was a pleasure to hear Professor A. T. J. Janse reminisce on the time that he
and Miss Leendertz went out collecting on bicycles and I thank him heartily for the
information so freely given. Grateful thanks are also due to Mrs. E. Patterson, a
niece, who lent me the few but important letters and documents Mrs. Pott had
preserved. These were very helpful indeed in building up a picture of her early years
at the Museum. Dr. V. Fitzsimons, the present Director, kindly lent us the photograph
reproduced here.
List of Publications by Mrs. R. Pott-Leendertz
1. The Amaryllidaceae of the Transvaal in Ann. Transv. Mus. 1: 46 (1911). Enumerating 11 genera
and 36 species.
2. A first checklist of the flowering plants and ferns of the Transvaal and Swaziland. J. Burtt Davy
and R. Pott-Leendertz in Ann. Transv. Mus. 3 : 119-182 (May, 1912). 3,240 species are enumerated.
3. A new species of Stapelia from the Transvaal, Stapelia gettleffii Pott in Ann. Transv. Mus. 3: 226,
1. 1 3 (1912).
4. A new record of plant species for the Transvaal, Duvalia polita N. E. Br. in Ann. Transv. Mus.
3: 226 (1912).
5. New species of Alepidea in Ann. Transv. Mus. 4: 206 (1914). A. jenkinsii Pott ; A. basinuda Pott.
6. A new species of Warburgia from the Transvaal, in Ann. Transv. Mus. 6: 60, fig. 1-2 (August 1918).
7. Addendum to the first checklist of the flowering plants and ferns of the Transvaal and Swaziland,
in Ann. Transv. Mus. 6: 119-135 (1918). Another 700 species were added.
'
5
Botlialia 9, 1 : 5-30.
The Genus Andropogon in Southern Africa
by
J. G. Anderson
Abstract
This study is limited to an investigation of the Southern African representatives of the genus
Andropogon, consisting of 14 species, two varieties and one record of doubtful identity. The history,
karyology, leaf anatomy and anatomy of the shoot and embryo are discussed briefly. A key based on
floral and vegetative characters is provided and a detailed description of the genus and of each species
and variety is given. The nomenclature and synonomy for each taxon is included as well as brief notes
on the distribution and habitat. Specimens deposited in the National Herbarium, Pretoria, and several
other South African Institutions are cited in full.
Introduction
No thorough revision of the genus Andropogon, as it occurs in South Africa and
including its nomenclature, has been undertaken since Stapf’s excellent work in
Volume 7 of the Flora Capensis (1898). An unpublished thesis by W. J. Louw (1943)
is available in which special attention is given to the leaf anatomy, but he dealt with
only nine species. A summary of the then known South African species, with short
descriptions and keys, was published by Chippindall in “ The Grasses and Pastures
of South Africa” (1955), but this work did not include nomenclatural information
nor the citation of specimens. Since the publication of this work, three new species
have been described by the present author, and one additional species has been
recorded for South Africa. The genus Hypogynium which included the South African
species H. festucifonnis has been referred back to Andropogon by Clayton in the Kew
Bull. 17, 3: 469 (1964). The present revision was, therefore, undertaken in order to
include the new information, as well as to investigate questions of synonomy and
nomenclature.
Most of the information regarding the anatomy of this genus was obtained from
Louw’s unpublished data but, as he did not deal with A. absyssinicus, A. lacunosus,
A. ravus, A. laxatus and A. festucifonnis, these species were investigated by the present
author.
The material in the National Herbarium, Pretoria, was studied as well as specimens
kindly sent on loan from the following herbaria: Bolus, Durban, Grahamstown,
Kimberley, Natal University and the South African Museum. Type material was
gratefully received on loan from Kew, Lisbon, Paris, Stockholm, Vienna and Zurich.
Special thanks are due to Dr. B. de Winter of the Botanical Research Institute, Pretoria,
for assistance and guidance.
All specimens cited are represented in the National Herbarium, Pretoria (PRE),
unless indicated otherwise.
History
Twelve species of Andropogon were included by Linnaeus in the first edition of
his Species Plantarum (1753) and, of these, A. distachyos L. has generally been accepted
as the type species. Most of the remaining eleven species, owing to their divergent
characteristics, have been assigned to other genera by subsequent authors. After
Linnaeus, the most important contributions to the knowledge of the genus for Southern
Africa came from Nees, Steudel, Hackel and Stapf.
6
In 1841 Nees dealt with 16 species of Andropogon and separated them into two
subgenera viz. Gymnandropogon and Cymbopogon. The subgenus Gymnandropogon
contained six species of which three, A. eucomus, A. amplectens and A. appendiculatus
were newly described from South Africa and are still regarded as true members of
Andropogon. The three remaining species are now placed in the genus Bothriochloa.
Under the subgenus Cymbopogon, Nees included ten species, eight of which were newly
described. These species have since been placed in the genera Monocymbium, Hyparr-
henia and Cymbopogon. Nees also described Heteropogon filifolius in this work, a
species subsequently transferred to Andropogon by Steudel.
Steudel, in 1855, recognised 459 species and divided the genus into five sections
and 29 subsections. Twenty species were cited from South Africa. Whilst several
of these were transferred to other genera by later workers, most of them were retained
as distinct taxa.
Hackel revised the tribe Andropogoneae in A. & C. de Candolle's “ Monographiae
Phanerogamarum ” in 1889 and divided the genus Andropogon into two series viz.
Isozygi and Heterozygi, consisting of thirteen subgenera with 193 species. Nine of
these were mentioned as coming from South Africa. Two new South African species
were described in this work, viz. A. schinzii and A. pilosissimus. During the present
study the holotype of A. pilosissimus was seen and it compared well with specimens
of Hyparrhenia aucta (Stapf) Stent deposited in the National Herbarium, Pretoria.
The necessary combination is made on p. 130 of this issue.
In 1898 the genus as it occurs in South Africa was fully treated by Stapf in the
Flora Capensis, Volume 7. Stapf followed Hackel in separating the genus into the
two series Isozygi and Heterozygi, dividing the South African representatives into
nine subgenera and 32 species. Stapf also states that this genus “ consists of several
heterogeneous elements ” and that “ several of the old genera of which it is made up,
will have to be restored, as for instance Chrysopogon, Cymbopogon and Heteropogon ”.
Only the seven species placed into the subgenus Arthrolophis are at present regarded
as true members of Andropogon.
In the Flora of Tropical Africa, 1919, Stapf treated the species of the genus occur-
ring in tropical Africa. In this work Stapf reinstated a number of genera formerly
placed in synonymy under Andropogon and gave a narrower circumscription to the
latter. The Andropogoneae were separated into sixteen subtribes with the genus
Andropogon placed in the subtribe Andropogonastrae. Stapf recognised 46 species
(and four sections). Of these, four species and one variety were known from South
Africa. Another four species and one variety subsequently proved also to occur in
South Africa.
The most recent work on the genus in this country is by Chippindall in “ The
Grasses and Pastures of South Africa”, (1955), in which ten species and one variety
are included. In this work brief descriptions of the species are supplied but synonymy
and citation of specimens are omitted.
Economic Importance
Economically this genus is not regarded as important and the species are usually
classified as “ sour grasses ”. They are usually grazed only in the young stages becoming
unpalatable when mature. Ecologically, however, certain species are important consti-
tuents in several of the major veld types of South Africa.
Karyology
In their study of A. distachyos, the type species of the genus, Gould (1956) and
Celarier (1956) found that in this species 2n = 36, indicating a basic chromosome
7
number of 9. For A. abyssinicus, Gould reported a count of 16. Subsequent studies
by various authors dealing with several other species of Andropogon, however, have
shown that 10 appears to be the commonest basic number in the genus.
The following table is a summary of the chromosome numbers so far recorded
for the South African species, with the authority and year of publication.
ANDROPOGON
Andropogon L., Sp. PI. 1045 (1753); Gen. PL ed. 5: 468 (1754); Nees. FI. Afr. Austr.
103 (1841), emend.; Steud., Syn. PI. Glum. 1: 363 (1855); Benth. & Flook. f., Gen.
PI. 3, 2: 1133 (1883); Hack, in DC., Monoar. Phan. 6: 359 (1889); Stapf in FI. Cap.
7: 334(1898); FI. Trop. Afr. 9: 208 (1919)f Phill., Gen. ed. 2: 84 (1951); Chippindall
in Gr. & Pastures S. Afr. 1: 495 (1955). Type species: A. distachyos L. (as “ dis-
tachyon ”, 1753).
Spikelets 2-nate, the one sessile, the other pedicelled, those of each pair differing
in sex and also more or less in shape and size or those of the lowest pair of the lowest
raceme male or imperfect or the pedicelled suppressed and the spikelets thus solitary.
Sessile spikelets 2-flowered, dorsally or laterally compressed, rarely pubescent, nearly
always with a bent and twisted awn; callus obtuse, subacute or acute, rarely broad
and obtusely rounded, shortly bearded. Glumes equal to subequal, subcoriaceous or
membranous; lower flat, concave or channelled on the back with sharply indexed
margins at least from the middle upwards, rarely with subcircular or elongated shallow
pits, 2-keeled; upper more or less boat-shaped, keeled upwards, 1-3-nerved, sometimes
aristulate or awned. Lemmas ciliate or ciliolate, rarely glabrous, lemma of lower
floret hyaline, 2-nerved; lemma of upper floret 2-fid or 2-dentate, usually awned from
the sinus, hyaline or firmer and sometimes substipiform below the insertion of the
awn; awn usually 2-8 times the length of the lemma. Paleas hyaline nerveless scales
or absent. Lodicules 2, minute, glabrous. Stamens 3, rarely reduced to staminodes.
Ovary glabrous, styles terminal; stigmas laterally exserted. Grains narrowly lanceolate
to oblong in outline, subterete to plano-convex; embryo about half the length of
the grain; starch grains simple or compound, granules mostly numerous. Pedicelled
spikelets often larger and more conspicuous than the sessile, sometimes small and
suppressed and represented by the pedicel only, more or less dorsally compressed,
rarely hairy. Glumes herbaceous, chaitaceous or membranous; lower usually flat
on the back, rarely grooved or with subcircular or elongated pits, muticous or very
rarely aristulate, usually many-nerved; upper membranous to subcoriaceous, 3-5,
rarely 7-nerved, ciliate or ciliolate on the margins, apex acute, acuminate, mucronate
or shortly awned. Florets: lower reduced to the lemma; upper male or sterile, rarely
bisexual; lemma, if present, hyaline, muticous or very rarely aristulate.
8
Perennials of various habit, rarely annuals, usually compactly caespitose, sometimes
with an oblique or short rhizome, very rarely fascicled or stoloniferous. Culms erect,
rarely geniculate, terete or sometimes compressed below, glabrous, smooth, rarely
hairy upwards, often channelled on the one side, usually simple, rarely branched either
from the base or from all the nodes resulting in a suffrutescent habit, or only from the
upper nodes. Leaves with the sheaths terete or the lower compressed and keeled,
glabrous or sparingly hairy, rarely densely pubescent, smooth, often tinged with purple;
ligules membranous with a ciliate or ciliolate rim, often scarious, truncate, rounded
to oblong or oblong-lanceolate; blades linear, expanded or longitudinally folded
(or revolute in dried specimens), more rarely filiform. Inflorescence of 1-many spike-
like racemes which are either in pairs, with the lower almost sessile, upper shortly
peduncled, or arranged subdigitately, or on a short central axis at the end of the culm
and, if present, its flowering branches ; raceme pairs or corymbs supported by a spathe
or spatheole and frequently collected into a false panicle; racemes glabrous or often
plumose with white or cream-coloured hairs; joints and pedicels linear, filiform,
cuneate or cuneate-clavate, often inflated and hollow or thickened upwards, tips more
or less cupular, auricled or dentate, rarely truncate.
Anatomy
Shoots rounded, compressed or strongly compressed in cross section, rarely cylin-
drical; leaf-blades rolled or folded in the bud; sheaths slightly to prominently keeled.
Leaf-blades flat, V-shaped or rarely terete in cross section, keels acute and prominent
or obtuse, rarely rounded, formed by an adaxial tissue of colourless parenchyma cells,
which are roughly triangular in transverse section. Silicified cells dumb-bell-shaped;
lateral walls smooth, exceptionally with a small central bulge. Suberized cells more or
less rectangular often with undulating lateral walls. Ripple-walled cells more or less
rectangular and arranged with their long axis parallel to the vascular bundles.
Stomata mainly located on the abaxial surface, usually in 1-2 longitudinal rows between
the nerves. Motor cells large and rectangular or hexagonal in shape, occupying £
the thickness of the leaf. Hairs variable and rare, of three kinds: (1) unicellular
usually antrorse sharp-pointed bulbous-based hairs; (2) unicellular tubercle-based
hairs, mostly on the abaxial surface and confined to motor cell and ripple-walled cell
areas; (3) bicellular hairs of which the basal and upper cells are approximately equal
in length; upper cell delicate and with a subobtuse apex. Bundle sheaths: outer sheath
thick-walled, parenchymatous, usually containing chloroplasts; inner sheath absent.
Chlorenchyma consisting of a usually well defined single layer arranged in the form
of an arc and surrounding each vascular bundle, often supported by irregulaily shaped
chlorophyll-bearing cells, mainly adaxially; the two adjacent arcs are either adjoining
or separated by a single row of larger rounded cells. Bulliform cells mainly consisting
of enlarged epidermal cells. Embryo about half the length of the grain; epiblast absent;
deeply cleft between the lower part of the scutellum and the coleorrhiza; the vascular
tissue with a distinct elongation between the point of divergence of the scutellum bundle
and the base of the coleoptile; margins of the primary leaf of the embryo (sheathed
by the coleoptile) overlapping and with several vascular bundles; the coleoptile
sheath with two lateral bundles and the scutellum with one median bundle.
Distribution
A large genus occurring mainly in the tropical and subtropical regions of both
hemispheres. Fourteen species, two varieties and one record of doubtful identity
(from the Natal Drakensberg) have been recorded from South Africa.
Most of the species are widely distributed throughout the eastern grassveld regions
of South Africa. Two species, A. eucomus and A. appendiculatus extend, in addition,
into mainly the coastal districts of the eastern and southern Cape, penetrating into
9
the south-western regions of the Cape. Three species and one variety, viz. A. eucomus,
A. amplectens , A. schinzii and A. schirensis var. angustifolius also occur in the central
regions while A. schinzii, A. eucomus, A. brazzae, A. huillensis, A. schirensis var.
angustifolius and A. gayanus var. squamulatus have been recorded from northern South
West Africa; all occur north of the tropic of Capricorn.
The name “Andropogon” is derived from the Greek words, aner, a man, and
pogon, a beard.
1 Basal parts usually distinctly flattened ; lower and sometimes the upper leaf-sheaths compressed
and keeled, usually strongly so; leaf-blades folded or sharply keeled, rarely expanded;
racemes 1-15:
2 Racemes plumose with silvery-white or cream-coloured hairs; joints and pedicels filiform
and slightly curved with hairs more than 5 mm long:
3 Sessile spikelets 2-3 mm long, the pedicelled supressed or represented by a minute rudiment ;
hairs on the joints and pedicels several times longer than the spikelets. ... LA. eucomus
3a Sessile spikelets 4—6 mm long, the pedicelled usually suppressed or rudimentary, more rarely
well developed; hairs on the joints and pedicels about as long as the sessile spikelets:
4 Racemes 5-12 (rarely 3 or 4), loose and usually flexuous; plant stout, 90-180 cm tall;
leaf-blades up to 30 cm long; lemma of the upper floret of the sessile spikelets
3-4 mm long, usually obscurely 3-nerved, rarely 1-nerved; awn 8-14 mm long
2. A. huillensis
4a Racemes 2-3, straight or slightly curved; plant up to 70 cm tall; leaf-blades up to
15 cm long; lemma of the upper floret of the sessile spikelets 2-3 mm long, usually
1-nerved; awn 15-20 mm long 3. A. laxatus
2a Racemes not plumose; joints and pedicels linear and straight, glabrous, or with hairs up to
5 mm long:
5 Racemes 1-3, 2^4 cm long; plant 20-100 cm tall; leaf-blades up to 30 cm long:
6 Racemes 1; sessile spikelets awnless 4. A. festuciformis
6a Racemes 2-3; sessile spikelets awned with awns up to 12 mm long 5. A. platybasis
5a Racemes 3-20 (if only 2, then racemes hairy), 4—16 cm long; plant 40-250 cm tall; leaf-
blades up to 50 cm long:
7 Sessile spikelets awnless; joints and pedicels glabrous or almost so; leaves expanded
or often revolute, margins scabrid or spinously ciliate 6. A. brazzae
7a Sessile spikelets awned, awns 10-15 mm long; joints and pedicels hairy with hairs up
to 5 mm long; leaves usually folded, rarely expanded, margins scabrid 7. A. appen-
diculatus
la Basal parts not distinctly flattened; lower leaf-sheaths rounded; leaf-blades filiform, convolute
or expanded; racemes 2, exceptionally 3, 4 or 5:
8 Glumes densely or sparingly hairy over the whole surface with hairs up to 4 mm long or,
if glabrous, then with 5-12 subcircular to elongated shallow pits (depressions) on either
side of a deep longitudinal median groove; joints and pedicels linear, straight:
9 Lower glume of the sessile spikelet dorsally flat or only slightly concave; glumes densely
or sparingly hairy with hairs about 4 mm long:
10 Annual; glumes densely hairy; racemes whitish-villous; joints and pedicels densely
hairy on the back and sides 8. A. abyssinicus
10a Perennial; glumes sparingly hairy; racemes appearing glabrous; joints and pedicels
hairy only on the sides 9. A. pilosellus
9a Lower glume of the sessile spikelet dorsally with a deep longitudinal median groove;
glumes glabrous and with 5-12 subcircular to elongated shallow pits on either side
of the groove 10. A. lacunosus
8a Glumes glabrous or ciliate only on the keels and margins, without depressions on either
side of the longitudinal median groove; joints and pedicels cuneate, cuneate-clavate
or cuneate-linear, usually inflated and hollow:
11 Lower glume of the sessile spikelet dorsally with a deep, narrow or wide, longitudinal
groove, groove often so deep that the wingless keels almost meet, especially in spikelets
borne on the upper part of the raceme:
10
12 Leaf-blades broad and rounded to cordate and more or less amplexicaul at the base;
callus acute or subacute 11. A. amplectens
12a Leaf-blades not or only slightly widened, often narrowed towards the base; callus
subacute, obtuse or obtusely rounded (if acute, then leaves filiform or narrowed
towards the base):
13 Upper glume of the sessile spikelet and both glumes of the pedicelled spikelet awned,
awn up to 10 mm long; culms branched, the branches either barren or bearing
inflorescences 12. A. schinzii
1 3a Glumes of both the sessile and pedicelled spikelets awnless, usually acute or mucronate ;
culms simple, rarely with one flowering branch:
14 Sessile spikelets 7-9 mm long; callus broad and obtusely rounded: plants often
with a branched system of knotty rhizomes; inflorescence usually glaucous-
grey or green 13. A. ravus
14a Sessile spikelets up to 7 mm long; callus obtuse, subacute or acute; plants densely
tufted; inflorescence and often the older leaves drying brown or reddish-brown:
15 Sessile and pedicelled spikelets very unequal in length; awn of sessile spikelets
3-5 cm long, stout; callus acute or subacute; leaves usually filiform or
when expanded then up to 3 mm wide, leaf-sheath usually breaking up
into fibres when old 14. A. filifolius
15a Sessile and pedicelled spikelets more or less equal in length; awn of sessile
spikelets up to 3 cm long, slender; callus obtuse to subacute; leaf-blades
usually expanded and 3-7 mm wide; lower leaf-sheaths not breaking up
into fibres 15. A. schirensis var. angustifolius
11a Lower glume of the sessile spikelet dorsally flattened or only very shallowly grooved:
16 Keels of the glumes of the sessile spikelets wingless; culms branched, robust, up to
3-5 m tall, usually glaucous, plants tufted 16. A. gayamts var. squamidatus
16a Keels of the glumes of the sessile spikelets distinctly winged; culms simple, up to
1-2 m tall; plant with a short oblique rhizome 17. A. distachyos
1. Andropogon eucomus Nees, FI. Afr. Austr. 1: 104 (1841); Steud., Syn. PI.
Glum. 1: 390 (1855); Hack, in DC., Monogr. Phan. 6: 421 (1889); Stapf in FI.
Cap. 7: 338 (1898); FI. Trop. Afr. 9: 230 (1919); Chippindall in Gr. & Pastures
S. Afr. 1; 501 (1955). Type: Clanwilliam, Olifantsriver, Eck/on & Zeyher s.n. (S, lecto.!).
Eriopodium krausii Hochst. ex Krauss in Flora 115 (1846). Type: Natal, between
Umlazi River and Durban Bay, Krauss 163 (K!; PRE, fragment!).
Perennial, densely tufted, occasionally with a short creeping rhizome. Culms up
to 90 cm long, erect or slightly geniculate, glabrous, 4-6 noded, compressed below,
simple to about the middle and with 2-6 or more flowering branches from the upper
nodes. Leaves up to 30 cm long and 4 mm broad, folded or rarely expanded, linear,
acute, glabrous or pubscent especially near the ligule, smooth or slightly scaberulous
on the margin, midrib slender, prominent below, lateral nerves 4-5 on each side, very
fine; ligules 0-5 mm long, truncate, membranous, ciliolate; sheaths glabrous or
beaided near the mouth, rarely lower down, firm, basal compressed and keeled, the
remainder shorter than the internodes, terete, distant, the uppermost subtumid.
Inflorescence of 2-5 subdigitate racemes, conspicuously plumose from the silvery-
white hairs surrounding the spikelets and on the rachis and pedicels; spathes linear-
lanceolate, acuminate or acute, usually without a blade, glabrous, pale-green or tinged
with purple or red, at length tightly inrolled. Racemes 3-5 cm long, plumose, often
nodding; joints finely filiform, 2 mm long, silky with soft 1-0-1 -2 cm long hairs;
pedicels similar, 4 mm long, quite barren or occasionally with the minute rudiment
of a spikelet at the apex. Spikelets all sessile, hermaphrodite, often purplish; callus
minute, truncate, long bearded. Glumes equal; lower 2-3 mm long, subcoriaceous,
narrowly lanceolate, concave dorsally, keeled, with 2-3 obscure nerves in each keel;
keels scabrid in the upper half, rarely long bearded near the apex, obtuse and glabrous
in the lower half; apex minutely bifid; upper boat-shaped, 2-3 mm long, membranous,
11
1- nerved, sparingly pilose on the margins in the upper half, dorsally scabrid on the
nerve; apex acuminate or shortly mucronate. Lower floret reduced to a lemma
2- 2-5 mm long, narrowly lanceolate, membranous, hyaline, nerveless, sparingly ciliate.
Upper floret hermaphrodite, lemma 1-5-2 mm long, lanceolate-oblong, 1-nerved,
ciliolate, bifid to about one third, lobes finely subulate, awned from between the lobes;
awn a fine bristle, 12-20 mm long, slender, slightly twisted near the base; palea very
minute, hyaline, nerveless; anthers 0-75 mm long; grain 1 mm long or longer, linear-
lanceolate in outline, subcylindric. Pedicelled spikelets absent or reduced to a minute
rudiment.
This species is widely distributed in southern Africa but has not been recorded
from the drier areas of the central Cape, south-western Orange Free State and southern
South West Africa. It also occurs in Mozambique, Rhodesia, Angola, the Congo
and Uganda. It usually grows in damp places, often in vleis and occasionally in open
grassland or on rocky outcrops.
Cape. — Ceres: Marloth 12643. Clanwilliam: Olifants River at Villa Brakfontein, Zeyher 47. East
London: Convict Station, Galpin 5580. Grahamstown: Hutton 249. Hay: Griquatown, Burchett
2102 (GRA; PRE, fragment); Witsand, Acocks 4474 (KMG); Leistner 1171. Humansdorp: Karee-
douw, Fowcade 2533. Komga: near Kei Mouth, Flanagan 2362. Paarl: Klein Drakenstein Mts.
on Berg River, Drege s.n. (S; PRE, fragment); Dal Josaphat, Tyson 6050 (SAM). Peninsula: Kloof-
neck, Marloth 3717; Steenberg near Lakeside, Bolus 14928 (BOL). Port Elizabeth: Schoemakers
Kop, Paterson 548 (GRA). Riversdale: Kortefontein, N. slopes of Langeberge, Acocks & Hafstrom
2039. Robertson: Omklaar, Stokoe 7196 (BOL). Somerset West: Steenbras River Mouth, Acocks
3936 (KMG). Taung: Thoming, Brueckner 1235; Rodin 3653. Tulbagh: Nieuwkloof, MacOwan
1691 (GRA, SAM). Vryburg: Armoedsvlakte, Mogg 8123. Worcester: Bainskloof Pass, Bolus
11743 (BOL).
Basutoland. — Tejatejaneng: Dieterlen 816.
Orange Free State. — Heilbron: Makouvlei, Brandmuller 44; Viljoensdrift, Rogers 13101.
Natal. — Bergville: Cathedral Peak Forest Research Station, Killick 1435; 1448. Camperdown:
Nagle Dam, Wells 1171 (NU). Dundee: Vants Drift, Codd 185. Durban: Bay End, Strey 4543;
Medley Wood 1654 (NH); 6004; Rehmann 8550 (GRA). Eshowe: Kotze 417. Estcourt: Mountain
hostel, West 253; Champagne Castle, Bayer 1246a. Newcastle: 20 miles S.E., Godfrey S.H. 1887.
New Hanover: Blinkwater, Sim 19361. Nkandla: Melmoth, Tinley 6; Qudeni, Leisegang 57 (NU).
Nongoma: Gerstner 3875 (NH). Pietermaritzburg: Table Mtn., Killick 282 (NU). Signal Hill,
Bews 37 (NU). Port Shepstone: Mogg 13201. Umlazi: Umlaas River, Drege s.n. (S, SAM); Isipingo
Beach, Ward 1275 (NU). Utrecht: Donkerhoek, Devenish 1124.
Swaziland. — Mbabane: Ukutula, Compton 25514; Forbes Reef, Compton 22391.
Transvaal. — Belfast: Schoemanskloof, Smuts 336. Brits: Jacksonstuin, Van Vuuren 400. Ermelo:
Athole Pasture Research Station, Norval 137. Johannesburg: Hutton 249 (GRA). Krugersdorp:
Jack Scott Private Nature Reserve, Wells 2367. Letaba: Duiwelskloof, Westfalia Estate, Scheepers
838. Lichtenburg: Malmanie Oog, Burn Davy 80. Middelburg: Pan Station, Burtt Davy 137. Nel-
spruit: Kaapsehoop, Devils Kantoor, Pole Evans 1008. Petersburg: farm Vaalwater, Skead 31;
Houtbosch, Rehmann 5673 (BOL). Pilgrims Rest: Mariepskop, Van der Schijff 4473. Potchefstroom :
Welverdiend, Louw 870. Potgietersrus : Crecy, Barenbrug M. Pretoria: Rayton vlei, Schweickerdt
1583; Waterkloof, Verdoorn 832; Wonderboom, Louw 864. Vereeniging: Brandmuller 44. Waterberg:
Nooitgedacht 699, Acocks & Naude 42.
Bechuanaland. — Lobatsi: Kanye, Bathosu Dam, Miller B/873.
South West Africa. — Caprivi: Katima Mulilo, Killick & Leistner 3082. Grootfontein : Omuramba,
Hohensee, Boss s.n.; Fontein Omuramba, De Winter & Marais 5041. Okahandja: Wiss & Kinges
994. Otjiwarongo: Waterberg, Liebenberg 4785; Waterberg plateau, Boss s.n.
The Ecklon & Zeyher specimen in Stockholm proved to be the most authentic
material seen. This specimen is written up in Nees’s own handwriting and was for this
reason chosen as the lectotype.
This species is characterized by the absence of pedicelled spikelets. The pedicels,
although present, are either barren or bear only the minute rudiments of a spikelet.
The hairs on the joints and pedicels are dense and several times longer than the spikelets.
These hairs, together with the long slender awns of the lemma of the upper floret, give
the inflorescence an attractive white, plumose appearance.
12
2. A. huillensis Rendle, Cat. Afr. PI. Welw. 2: 146 (1899); Stapf in FI. Trop.
Afr. 9: 231 (1919); Chippindall in Gr. & Pastures S. Afr. 1: 500 (1955). Type:
Angola, Huila, Humpata, Welwitsch 2670 (BM, holo.; PRE, fragment!, photo.!).
Perennial, tufted from a short oblique rhizome. Culms 90-180 cm long, erect,
stout, terete, 5 or more noded, glabrous, very smooth, shiny, simple up to the third
or fourth node. Leaves up to 60 cm long and 4 mm wide, linear, long tapering to a
fine apex, flat or folded, firm and hard, the upper almost reduced to the midrib, smooth
and glabrous or sparingly pilose towards the ligule, scaberulous on the margin; ligules
2 mm long, firm, scarious, ciliolate; sheaths firm, tight, glabrous or bearded near
the ligule, the lower up to 25 cm long, compressed and keeled or sometimes terete,
often shiny or tinged with purple. Inflorescence of 5-12 racemes arranged subdigi-
tately, plumose from the long white hairs on the jonts and pedicels; spathes glabrous,
the lowest resembling the preceding leaves; spatheoles narrow, linear-lanceolate to
linear, tapering to an acute apex, up to 8 cm long, reddish, at length tightly inrolled.
Racemes 4-10 cm long, slender, loose, flexuous and often nodding, plumose, sessile
or with peduncles 1-3 cm long; peduncles glabrous; joints 3-6 mm long, filiform,
often curved, bearded with fine, approximately 7 mm long hairs, apex shallowly cupular;
pedicels similar but approximately 5 mm long. Sessile spikelets 4-6 mm long, linear-
lanceolate, more or less tinged with purple; callus minute, rounded, bearded. Glumes
firmly membranous; lower 4-6 mm long, dorsally deeply or shallowly concave, keeled,
obscurely 3-nerved in each keel; keels scaberulous in the upper half, apex minutely
bifid; upper 5 mm long, narrowly boat-shaped, 1 -nerved, pilose on the margins in
the upper half; apex acuminate or mucronulate. Lower floret reduced to a linear-
oblong lemma, 3-5 mm long, membranous, hyaline, nerveless or obscurely 2-nerved,
pilose on the margins in the upper half, apex acute. Upper floret hermaphrodite;
lemma 3-4 mm long, narrowly linear-lanceolate, membranous, hyaline, obscurely
3-nerved, rarely 1 -nerved, pilose on the margins in the upper half, 2-lobed, lobes acute,
awned from between the lobes, awn a fine bristle 8-14 mm long, usually bent and
twisted below the middle, scaberulous; palea 2-2-5 mm long, oblong, hyaline, nerveless,
bearded near the obtuse apex; anthers 2 mm long; grain 4 mm long, linear-lanceolate.
Pedicelled spikelets variously reduced, sometimes to a glume or entirely suppressed,
occasionally well developed and male and then as long as or slightly longer than the
sessile.
A. huillensis is common in the Transvaal and is also found in Zululand, Swaziland,
South West Africa, Angola, Rhodesia and Mozambique. It usually grows in damp
places, along streambanks and in vleis, very often in sandy soils.
Transvaal. — Barberton: White River, Rogers 22965. Brits: Jacksonstuin, Van Vauren 392.
Johannesburg: Hutton 254. Krugersdorp: Lake, Louw 922. Lydenburg: between Lydenburg and
Dullstroom, Pole Evans 3811. Middelburg: Burtt Davy 13264; Hutchinson 2709. Nelspruit:
Pretoriuskop, Van der Schijff 2648. Pilgrims Rest: Graskop, Driekop Gorge, De Winter & Codd 210.
Potchefstroom: Elandsfontein, Louw 1522. Potgietersrus: Pyramid Estate, Gatpin 8884. Pretoria:
Rayton vlei, Schweickerdt 1593; Premier Mine, Repton 4058. Rustenburg: Rainhill, Story 991;
Codd 1074. Soutpansberg: Phiphiti Waterfalls, Obermeyer s.n. Warmbad: Bell s.n. Waterberg:
Twenty-four Rivers, 13 miles E. of Vaalwater, Codd 994; Rogers 25104.
Natal. — Ubombo: Bazwanu, Strey 5135.
Swaziland. — Hlatikulu: Mhlambanyati Pan, Michelmore 50. Mbabane: Black Mbuluzi Valley,
Compton 31269.
South West Africa. — Caprivi: Andara, Okavango River, Volk 2154.
A. huillensis can be confused with A. eucomus and A. laxatus but is generally a
taller, more robust plant with larger spikelets and with the hairs on the joints, and
pedicels shorter than in A. eucomus. In A. laxatus there are only 2-3 racemes in the
inflorescence whereas there are 5-12 in A. huillensis. In A. huillensis the pedicelled
spikelets are usually suppressed, these being represented by much reduced glumes.
These spikelets are occasionally well developed and male.
13
Welwitsch 2644, collected in Angola, was described by Rendle as A. huillensis
var. minor in Cat. Afr. PI. Welw. 2: 147 (1899). In the Flora of Tropical Africa,
Stapf states that the plant has smaller inflorescences and smaller spikelets and adds
that it is probably only a weaker state of A. huillensis. I have seen an isotype as well
as a fragment and photograph of the holotype and in my opinion this variety should
be included in the synonomy of A. laxatus.
3. A. laxatus Stapf in FI. Trop. Afr. 9: 237 (1919); Chippindall in Gr. & Pastures
S. Afr. 1: 501 (1955). Syntypes: Angola, Welwitsch 2693; Rhodesia, Inyanga , Mundy
s.n. ; Salisbury, Mundy s.n. (K!).
A. ternatus Nees var. africanus Rendle, Cat. Afr. PI. Welw. 2: 147 (1899). Type:
Angola, Huila, dry wooded pastures near Humpata, Welwitsch 2643 (LISU!; PRE,
photo.!). A. huillensis Rendle var. minor Rendle in Cat. Afr. PI. Welw. 2: 147 (1899).
Type: Angola, Huila, rather damp rich pastures on the river Monino, Welwitsch
2644 (LISU!; PRE, photo.!, fragment!).
Perennial, densely tufted. Culms up to 70 cm long; erect, terete, somewhat
compressed below, glabrous, smooth, shiny, sometimes pruinose below the lower
nodes, branched in the upper half, 1-3 noded. Leaves up to 15 cm long and 4 mm
wide, linear, expanded or folded, subacute, firm, flexuous or straight, glabrous or
sparingly hairy near the ligule, margins scaberulous; ligules 1 mm long, scarious,
rounded or truncate, ciliolate; sheaths firm, glabrous, smooth, the lowest up to 6 cm
long, compressed, keeled, the upper looser, much longer and more or less terete.
Inflorescence of 2, rarely 3, racemes on a short common axis, silkily plumose from the
hairs on the joints and pedicels; spathes resembling the preceding leaves and with
short blades; spatheoles very narrow, acute, 6-8 cm long, at length tightly inrolled,
tinged with purple. Racemes 4-6 cm long, slightly flexuous, plumose; peduncles short
or the upper raceme sessile; joints 4-6 mm long, filiform, flattened, bearded with
spreading hairs 5-8 mm long, apex cupular; pedicels similar, 4-5 mm long, apex
obtuse. Sessile spikelets 5-6 mm long, linear-lanceolate; callus small, obtuse, bearded.
Glumes more or less equal, membranous; lower 5-6 mm long, dorsally concave, keeled,
1-2-nerved in each keel; keels scaberulous in the upper half, apex shortly bifid; upper
5-6 mm long, narrowly boat-shaped, 1-3-nerved, pilose on the margins in the upper
half, dorsally scabrid on the nerves, apex mucronate, mucro approximately 1 mm
long. Lower floret reduced to a lemma 4-5 mm long, narrowly oblong, hyaline,
nerveless, pilose on the margins in the upper half, apex acute. Upper floret herma-
phrodite; lemma 2-3 mm long, narrowly oblong, hyaline, usually 1 -nerved, often
pilose on the margins, apex 2-lobed, awned from between the lobes; awn 1- 5-2-0
cm long, slender, kneed and twisted below, scaberulous ; palea 2 • 5-3 mm long, oblong,
hyaline, nerveless, ciliate upwards. Pedicelled spikelets usually reduced, occasionally
well developed and male and then of the same shape and size as the sessile. Glumes
unequal; lower 5-nerved, acuminate; upper 3-nerved, acuminate; lemmas of both
florets, if present, narrowly oblong, acute, 1 -nerved, ciliate.
A. laxatus is a rare species occurring in Angola, Rhodesia, Zambia and the Congo
Republic. It has been recorded only once from South Africa, namely at Graskop
in the Pilgrims Rest district of the Transvaal. It usually grows in wet places, near
streams or in vleis.
Transvaal. — Pilgrim’s Rest: Graskop, Driekop Gorge, De Winter & Codd 218.
Angola. — Huila: Humpata, Texeira 3411. Moxico: Miltie-Redhead 4217.
Zambia. — Mwinilunga: Sinkabolo Dambo, Milne-Redhead 4380. Barotseland: 10 miles N. of
Senanga, Codd 7281.
Congo. — Leopoldville: Kwango, Devred 1495.
14
This species is very closely related to A. huillensis but is a shorter, less robust
plant with 2, rarely 3, racemes to the inflorescence. The lemma of the upper floret
is shorter, 2-3 mm long, but with a much longer awn, 15-20 mm long, than is found
in A. huillensis.
Of the syntypes only Mundy s.n. from Salisbury was seen and consequently no
lectotype was selected.
4. A. festuciformis Rendle in Cat. Afr. PI. Welw. 2: 145 (1899); Clayton in Kew
Bull. 17: 470 (1964). Type: Angola, Huila, Humpata, Welwitsch 7505 (BM, holo. ;
PRE, photo. !).
A. schlechteri Hack, in Bull. Herb. Boiss. ser. 2, 6: 703 (1906). Type: Natal,
Clairmont, Schlechter 3143 (Z, holo.; BOL!).
Hypogynium schlechteri (Hack.) Pilger in Pflanzenfam. ed. 2, 14e: 156 (1940);
Chippindall in Gr. & Pastures S. Afr. 1 : 516 (1955).
Perennial, densely tufted, subflabellate. Culms up to 100 cm long, erect, simple
below but branched upwards, compressed, smooth and often glossy, usually purplish,
2-3-noded. Leaves up to 30 cm long and 3 mm wide, rigid, folded, linear, passing
gradually into the sheath, pubescent on the upper surface near the base, apex acute;
ligules 1 mm long, membranous, acute; sheaths striate, persistent, firm, often tinged
with purple, lower compressed and keeled, glossy. Inflorescence of solitary, peduncled
racemes, crowded to form a narrow panicle 5-15 cm long or scattered along the culm;
spathes scarious, narrow, the lower 2-3 cm long, decreasing in size upwards, acute,
usually tinged with purple. Racemes 2-4 cm long, straight, glabrous; joints and
pedicels unequal, 4-5 mm long, linear, straight, often with a shallow median groove,
scaberulous, apex truncate. Sessile spikelets female, 4-4 • 5 mm long, linear-lanceolate ;
callus short, obtuse, shortly bearded. Glumes subequal; lower 3 5-4 mm long, linear-
lanceolate, bicarinate, subcoriaceous, scaberulous on the keels in the upper half, dorsally
shallowly grooved, 1-nerved in each keel, apex minutely bifid; upper 3-3-5 mm long,
boat-shaped, sharply keeled, subcoriaceous, 1-nerved, scaberulous on the upper half
of the keel, apex acute. Lower floret reduced to the lemma, 2-5-3 mm long, narrowly
oblong, membranous, nerveless, apex obtuse or truncate. Upper floret with the lemma
lanceolate, 2-2-5 mm long, membranous, 1-nerved, apex shortly bilobed or subacute,
awnless; palea absent; anthers reduced to staminodes. Pedicelled spikelets male,
5 • 6-6 mm long, lanceolate. Glumes subequal ; lower 5 • 6-6 mm long, narrowly ovate,
subcoriaceous, 5-7-nerved, glabrous, apex acuminate; upper 4-7-5 mm long, subcori-
aceous, narrowly ovate, 5-7-nerved, apex acuminate. Lower floret with the lemma
3 • 5-4 mm long, lanceolate, membranous, 3-nerved, apex obtuse. Upper floret with
the lemma 3-2-4 mm long, narrowly lanceolate, membranous, 1-nerved, apex acute;
palea absent; anthers 2-5 mm long.
Andropogon festuciformis has been recorded from Natal, Angola, Zambia, the
Congo Republic and the Guinea Republic. It usually grows in damp places, often
in vleis.
Natal. — Durban: Clairmont, Medley Wood 8543. Hlabisa: W. of St. Lucia Estuary, Feely &
Ward9\ Umfolozi flats near Mtubatuba, Liebenberg 591 1. Pietermaritzburg: Table Mtn., Killick 284.
This species differs from all the other South African species in having solitary
racemes. Until recently it was placed in the genus Hypogynium Nees and was separated
from Andropogon mainly on account of the awnless lemmas of the sessile spikelets
and the solitary racemes. W. D. Clayton in the Kew Bulletin (1964) points out that
a few species of Andropogon in the section Piestum (e.g. A. textilis) also have solitary
racemes and that these characters are not sufficient to justify a separate genus. In
A. brazzae (see p. 15) the sessile spikelets are awnless as in A. festuciformis.
15
5. A. platybasis J. G. Anderson in Kirkia 1 : 102 (1961). Type: Transvaal, Belfast,
5^ miles from Dullstroom on road to Lydenburg, De Winter & Codd 180 (PRE, holo.!).
Perennial, tufted with a short rhizome. Culms up to 75 cm long, erect, simple,
2-3-noded, glabrous, striate, compressed; nodes glabrous, internodes glabrous and
smooth except for the upper which are pilose for approximately 3 cm below the
inflorescence. Leaves up to 15 cm long, 3-4 mm wide, folded or expanded in the upper
half, linear with the apex acute, glabrous, occasionally with a few long hairs near the
ligules, margins scaberulous; sheaths strongly compressed and keeled, glabrous,
striate, pallid or suffused with purple, basal persistent; ligules fimbriate, membranous,
1 mm long. Inflorescence of 3, occasionally 2, spike-like racemes arranged digitately
or subdigitately. Racemes 2-3 cm long, straight, greyish-violet; joints and pedicels
2-3 mm long, linear, pilose on the margins. Sessile spikelets female with the anthers
leduced to staminodes, 5-6 mm long, lanceolate. Glumes subequal, membranous,
glabrous; lower 5-6 mm long, narrowly lanceolate, minutely 2-dentate, concave
between the keels with 2 nerves in each keel, scaberulous from the middle upwards;
upper 3-6-4 mm long, narrowly boat-shaped, 3-nerved with the faint lateral nerves
often anastomosing with the distinct midnerve, midnerve excurrent into a 1 mm long
awn. Lower floret reduced to the lemma, 4 mm long, margins inflexed. hyaline, nerveless,
margins ciliate in the upper half. Upper floret with the lemma 2 mm long, lobed to
about the middle, awned from between the lobes; awn 9-12 mm long, lobes ciliate;
palea 0-75 mm long, truncate hyaline, nerveless. Pedicelled spikelets male, 6-7 mm
long, lanceolate. Glumes membranous; lower with a distinct midnerve and 3-4 nerves
on either side, keeled with the keels scaberulous in the upper half, entire at the apex
or with 2 membranous lobes on each side of the slightly excurrent nerve; upper narrowly
boat-shaped, 7 mm long, 3-nerved, membranous, apex mucronulate. Lower floret
reduced to the lemma, 5-5-6 mm long, lanceolate, 3-nerved with a distinct midnerve
and a faint lateral nerve on either side, apex obtuse. Upper floret male; lemma and
palea absent; anthers 3 mm long.
A. platybasis is a rare species recoided from the Belfast, Lydenburg and Wakker-
stroom districts of the eastern Transvaal and from Mbabane in Swaziland. It is a
water-loving grass growing in marshy soils, on streambanks and at high altitudes.
Transvaal. — Belfast: 6 miles N.E. of Dullstroom, Codd & De Winter 3225. Lydenburg: 13-6
miles from Lydenburg on road to Sabie, De Winter & Codd 199. Wakkerstroom: Oshoek, Devenish
910.
Swaziland. — Mbabane, Compton 30984.
A. platybasis is characterised by the upper floret of the sessile spikelet being female
with its anthers reduced to very tiny staminodes. It is also readily distinguished from
other South African species in being generally a smaller plant with 2-3 short, greyish-
violet, racemes.
6. A. brazzae Franch. in Bull. Soc. Hist. Nat. Autun 8; 326 (1895); Stapf in FI.
Trop. Afr. 9: 233 (1919); Robyns, FI. Agrost. Congo Beige 1: 126 (1929). Type:
Congo Republic, between Mokeno and Mongo, Brazza 234 (P, holo., iso.!; PRE,
photo.!).
A. appendiculatus Nees var. genuinus sensu Dur. & Schinz, Etudes FI. Congo 314
(1896), non Hack. (1889), fide Stapf.
Perennial, densely tufted with a short rhizome. Culms up to 250 cm long, erect,
stout, slightly compressed and simple below with a few long, erect, flowering branches
higher up, often channelled, glabrous and smooth, 3-many noded. Leaves up to 50 cm
long and 5 mm wide, linear, long attenuate to a fine apex, expanded or revolute in
dried specimens, puberulous on both surfaces, margins scabrid to spinously ciliate;
ligules membranous, ciliolate, 2 mm long, rounded; sheaths up to 20 cm long and
16
1 cm broad, lowest and intermediate compressed and strongly keeled, firm, striate,
glabrous, often tinged with purple. Inflorescence of 3-15 racemes, simple or divided;
spathes like the preceding leaves but with reduced blades. Racemes 3-10 cm long,
slender, loose and often flexuous, glabrous, pale or purple-brown, joints 4-7 mm long,
linear, slender, slightly thickened upwards, glabrous, scaberulous or shortly ciliate on
the margins, transparent towards the middle line, tips minutely cupular and denticulate;
pedicels similar, 3-5-5 mm long. Sessile spikelets hermaphrodite, 4-6 mm long,
linear-lanceolate, loosely wedged in between joint and pedicel; callus short, obtuse,
glabrous. Glumes subequal; lower 4-6 mm long, subcoriaceous, keeled, shallowly
concave and hyaline between the keels; keels smooth in the lower half, scabrid upwards,
2-nerved in each keel, apex bifid or acuminate; upper 4-5-5 mm long, membranous,
boat-shaped, acuminate, 3-nerved with the midnerve distinct, the lateral nerves shorter
and finer, glabrous or dorsally scabrid in the upper part of the midnerve. Lower floret
reduced to a lemma 3-5-5 mm long, membranous, hyaline, lanceolate, acute or
acuminate, indistinctly 3-nerved, winged in the upper half, glabrous or ciliolate on the
margins. Upper floret with the lemma 4-5-5 mm long, membranous, hyaline, 1 -nerved
or indistinctly 3-nerved, glabrous, acute, mucronate or rarely with rudiments of an
awn up to 1-5 mm long; palea 1-1-5 mm long, ovate-oblong, hyaline, nerveless;
anthers 2 mm long. Pedicelled spikelets male, 4-6 mm long. Glumes subequal; lower
4-6 mm long, linear-lanceolate, subcoriaceous, keeled, 5-nerved with a distinct mid-
nerve and 2 nerves close together in each keel, keels scabrid in the upper half, apex
acute or acuminate; upper 4-5 mm long, lanceolate, often winged in the upper half,
membranous, glabrous or ciliolate on the margins, 3-nerved, acuminate or often
mucronate. Lower floret reduced to a lemma 3-4 mm long, membranous, hyaline,
winged in the upper half, glabrous or ciliolate on the margins, indistinctly 3-nerved,
apex acute. Upper floret with the lemma 3-3 • 5 mm long, lanceolate, glabrous, hyaline,
obscurely 1 -3-nerved, mucronulate; palea 1 mm long, ovate-oblong, hyaline, nerveless;
anthers 2 mm long.
A. brazzae has been recorded from the Okavango Native Territory in South West
Africa, Angola and the Congo Republic. It is a rare species found on streambanks
or in marshy soils.
South West Africa. — Okavango: 16-8 miles W. of Nyangana Mission, De Winter & Marais 4775.
Angola. — Moxico: Lusavo Ferry, Milne-Redhead 4125.
Congo Republic. — Equateur: Conteaux 503; 564; Coquihatville, Germain 4717; Bondo Village,
Germain 1934. Leopoldville: Mongobele, Flamigny 5046. Orientate: Ango, Germain 4311.
This species is allied to A. appendiculatus but differs from that species in that the
spikelets are either awnless or do not produce fully developed awns. Investigation
of the material deposited in the National Herbarium, Pretoria, revealed that in the
sessile spikelets the apex of the lemma of the upper floret is usually acute or mucronate.
In some spikelets the midnerve of this lemma is thickened at the apex and frequently
produced into an awn up to 1-5 mm long. Other interesting features of this species
are that the leaf-blades are frequently revolute in dried specimens and that the leaf
margins are often very distinctly spinously ciliate.
7. A. appendiculatus Nees, FI. Afr. Austr. 105 (1841); Steud., Syn. PI. Glum.
1 : 379 (1855); Hack, in DC., Monogr. Phan. 6: 436 (1899); Stapf in FI. Cap. 7: 339
(1898); Chippindall in Gr. & Pastures S. Afr. 1: 500 (1955). Type: Cape, Klein
Buffelvallei, near Gaatjie, Drege s.n. (S, lecto.!, SAM!).
A. ischaemum Thunb., Prodr. 20 (1794), non Linn. Type: Cape, Thunberg s.n.
(S, iso.!). A. appendiculatus Nees var. genuinus Hack, in DC. Monogr. Phan. 6: 437
(1889). — var. ischaemum Nees, FI. Afr. Austr. 105 (1841). Syntypes: Uitenhage,
Van Stadensriver Mtns., Drege 696!; Zwartkei River, Ecklon s.n.!; Klein Buffelvallei,
near Gaatjie, Drege s.n.! — forma purpurea Nees, l.c., 106 (1841). Type: Winterberg,
17
Ecklon s.n. — var. serrulatus Nees, l.c., 105 (1841), excl. syn. ; Hack, in DC., Monogr.
Phan. 6: 437 (1889). Syntypes: Waterfall near Tulbagh, Ecklon s.n.!; Dutoitskloof,
Drege s.n.!.
Perennial, densely tufted with a creeping, sometimes branched, rhizome. Culms
up to 120 cm long, glabrous, compressed or terete, 4-6 noded, simple below, remotely
branched from the upper 2-3 nodes with 1-3 flowering branches; branches solitary,
long, glabrous. Leaves mostly basal; blades up to 30 cm long and 5 mm broad,
usually folded or sometimes expanded, upper very short or obselete, rigid, glabrous
or bearded near the ligule, acutely keeled, margins scabrid; ligules 1-5-2 mm long,
truncate, membranous, ciliolate; sheaths glabrous or bearded at the mouth, lower very
firm, strongly compressed, keeled, persistent, upper shorter than the internodes, tight,
often tinged with purple. Inflorescence of 2-20 racemes; peduncles generally long
exserted, glabrous. Racemes 4—16 cm long, fascicled, unequal, often slender and loose,
usually strongly flushed or wholly dark purple; joints linear, 4-6 mm long, hairy on the
convex back, hairs up to 5 mm long, white or yellowish, tips subcupular with irregularly
toothed margins; pedicels similar but more slender, produced into a subulate appendage
facing the upper glume. Sessile spikelets hermaphrodite, 6-7 mm long, lanceolate;
callus short, obtuse, scantily bearded. Glumes subequal, subcoriaceous to chartaceous;
lower 6-7 mm long, lanceolate, dorsally concave, keels acute, scabrid or rigidly ciliate
in the upper half, obscurely 3-nerved in each keel, apex bifid, rarely deeply bifid; upper
5-5-6 mm long, boat-shaped, 1-nerved, glabrous or scabrid on the keel. Lower floret
reduced to the lemma, 5-5-6 mm long, oblong-lanceolate, membranous, obscurely
2-nerved or nerveless, pilose on the margins in the upper half, apex acute. Upper
floret with the lemma 4—4-5 mm long, membranous, 1-3-nerved, pilose on the margin
in the upper half, 2-lobed, lobes acuminate, awned from between the lobes; awn
10-15 mm long, slender, geniculate below the middle; palea 1-2-3 mm long, ciliate,
hyaline, nerveless; anthers 4 mm long. Pedicelled spikelets male. Glumes subequal,
subcoriaceous; lower 5-4-6 mm long, oblong-lanceolate, flat or slightly rounded
on the back, scabrid or ciliate on the keels in the upper half, keels acute, midnerve
prominent and with 3 nerves in each keel, nerves often anastomosing near the apex,
minutely awned from a bifid or acute apex; awn 1-7-2 mm long; upper 4-5-5 mm
long, oblong-lanceolate, pilose on the margins in the upper half, 1-3-nerved, apex
acute or shortly mucronate. Lower floret with the lemma 4-5-5 mm long, lanceolate,
membranous, 3-nerved or nerveless, reversedly pilose on the margins in the upper half,
apex acute. Upper floret with the apex of the lemma minutely bifid, otherwise similar
to that of the lower floret; palea 3-7-4 mm long, elliptic, membranous, nerveless,
ciliate on the margins in the upper half; anthers 3-4 mm long.
This is a common and widespread species in South Africa, in which it is endemic,
its distribution ranging from the south-western, eastern and north-eastern Cape to
Natal, Swaziland, Transvaal, Basutoland and the Orange Free State. It is absent
from the drier areas of the central Cape and South West Africa. A. appendiculatus
occurs from sea level up to altitudes of 10,000 ft. above sea level in the Natal Drakens-
berg. It is usually found growing in damp situations, frequently in swampy soils and
very often in the shade.
Cape. — Albany: Grahamstown Nature Reserve, Britten 5961. Aliwal North: Elands Hoek, Bolus
10505. Barkly East: Ben McDhui, Galpin 6896. Caledon: Rivierzondereind, Schlechter 9895 (BOL,
GRA). Ceres: Koude Bokkeveld, Skurfdebergen, Schlechter 10175. Clanwilliam: Sneeuberg, south
Cedarberg, Edwards 162; Juriesberg, Compton 6263 (NBG). Herschel: Majuba Nek, Sterkspruit,
Hepburn 284 (GRA). Humansdorp: Witte Els Bosch Peak, Esterhuysen 6116 (BOL). Keiskamma
Hoek: Cata Forest Reserve, Story 3279. Kentani: Pegler 82. King Williams Town: Amatola Mtns.,
Dyer 268. Knysna: Pirie, Sim 2822. Komga: Flanagan 1020. Lady Grey: 4 miles W., Godfrey
1542. Molteno: Bews 80 (NU). Mount Ayliff: Amalfi, Godfrey & Scott SH. 1503. Paarl: Du
Toits Kloof Mtn., Tyson 6040 (SAM). Port Elizabeth: Liebenberg 5387. Queenstown: Mtaba
Magwele Mtn., Galpin 2634. Stellenbosch: Jonkershoek, Taylor 3253; 4508. Steynsburg: Molteno
Road, Acocks 8668. Stutterheim: Fort Cunynghame, Sim 2731 ; Rogers 12701. Uniondale: Niekerks-
berg Forest Reserve, near Misgund, Taylor 3259. Worcester: Hex River Valley, Tyson s.n. (GRA).
18
Basutoland. — Leribe: Dieterlen 349; 6323 (SAM). Maseru: Maluti Mtns., Staples 158; Roma,
Schutte s.n. (BOL). Quachas Nek: valley of the Legowa River, Pole Evans 84. Quthing: above
Mphaki, Archibald 534 (GRA).
Orange Free State. — Bethlehem: ravine near village, Phillips 3099. Ficksburg: Riverhill Farm,
Potts 3687. Harrismith: Van Reenen, Bews 81 ; 82; 83 (NU). Heilbron: Maccauvlei, Brandmuller 3.
Kroonstad: Pont 152. Senekal: Doornkop, Goossens 885. Winburg: 10 miles N., Godfrey SH. 1561.
Zastron: “ Mooifontein ”, Liebenberg 5247.
Natal. — Bergville: Mont-aux-Sources, Bayer & McClean 317; Cathedral Peak Forest Res. Stn.,
Killick 1153. Dundee: Vants Drift, Codd 1 86. Durban: Wood 6049 (GRA); Umgeni, Rehmann
8652 (GRA). Estcourt: Cathkin Peak Hostel, West 13; Research Stn., Acocks 9958. Klipriver:
Acton Homes, Doidge s.n. Mapumulo: Buchanan 191 (GRA). Mooi River: Meteor Ridge, Mogg
3388. Mpendle: Huntley 456. Nkahdla: Qudeni Forest Reserve, Fisher & Schweickerdt 40 (NH).
Mtunzini: “ Homewith ”, Mogg s.n. Newcastle: road to Utrecht, Edwards 628. New Hanover:
Balgowan, Mogg 3546. Nqutu: Codd 239. Utrecht: Doringkop, Edwards 1037; Boshoffs Vlei,
Edwards 1046.
Swaziland. — Mbabane, Compton 23763; 23840 (NBG); Rogers 11609 (BOL).
Transvaal. — Belfast: Dullstroom, Galpin 13030. Benoni: Bradfield 372. Bethal: Leendertz
10532. Bronkhorstspruit: on road to Pretoria, De Winter & Codd 167. Carolina: 19 miles S.E.
of Machadodorp, Codd 8278. Ermelo: Maviristad, Pott 5208; Athole Pasture Research Stn., Norval
52; Nooitgedacht, Henrici 1284; 1299. Heidelberg: Henley on Klip, Stent H. 21 584. Johannesburg:
Leendertz 6062; Burtt Davy 832; Canada Junction, Cohen 871. Lydenburg: Zwagershoek, Obermeyer
s.n. Petersburg: Blaauwberg, on Mohlakeng Plateau, Codd & Dyer 8989. Pilgrim’s Rest: Mt.
Anderson, Pole Evans 3794. Potchefstroom : Welverdiend, Louw 874 (GRA). Pretoria: Doornkloof,
Irene, Pole Evans 382; Wolwekloof, Mogg 15909: Rayton, Schweickerdt 1698; Wonderboom South,
Louw 863. Schweizer-Reneke : Burtt Davy 1697 . Soutpansberg: Tzaneen, Sampson 141. Standerton:
Beginsel, Burtt Davy 926. Swartruggens: Sutton 933. Vereeniging: Deneysville, Chippindall 343;
Volksrust: Burtt Davy 150. Wakkerstroom: Oshoek, Devenish 1118; Jantjies Hoek pass, Galpin
70217. Wolmaransstad: Oersonskraal 16, Sutton 77.
This species is variable as may be expected when its wide distribution and its
occurrence in a wide range of habitats is taken into consideration yet it is, on the whole,
easily recognizable and distinct.
When describing the species, Nees (1841) divided it into var. serrulatus, var.
ischaemum and the forma purpurea, on grounds of the size of the spikelets and the
difference in pubescence of the lemma of the sessile spikelets. Hackel (1889) retained
the var. serrulatus but placed the var. ischaemum and forma purpurea in synonomy
under his variety genuinus.
When studying the herbarium material, I found that a small form with conspicuously
hairy racemes, as well as forms with more numerous, but less hairy racemes, occur
in the south-western Cape. From this area to the northern provinces the species shows
a gradual decrease in pubescence as well as an increase in size of plants and of the
number of racemes to the inflorescence. A taller plant with several, almost glabrous
racemes, is found in the Transvaal, Natal, and Orange Fiee State. The size of spikelets
and degree of pubescence of the lemma of the sessile spikelets is not correlated with the
different forms.
Study of a large range of material has thus shown that, in spite of the extreme
variability, there is no valid reason for subdividing this species into subordinate taxa.
8. A. abyssinicus R. Br. in Fres. Mus. Senkenb. 2: 146 (1837); Steud., Syn. PI.
Glum. 1: 371 (1855); Hack, in DC., Monogr. Phan. 6: 465 (1889); Stapf in FI. Trop.
Afr. 9: 222 (1919); Chippindall in Gr. & Pastures of S. Afr. 1: 496 (1955). Type;
Abyssinia, Salt s.n. (BM, holo. ; PRE, photo.!).
Annual, loosely tufted. Culms up to 70 cm long, geniculate, 2-4-noded, simple
or sparingly branched from below or at the middle, terete, glabrous, smooth, the
internodes channelled downwards. Leaves up to 25 cm long and 4 mm wide, expanded,
narrowly linear, long tapering to an acute apex, glabrous or sparingly hairy from
minute tubercles, midrib prominent below, broader and whitish above; ligules 0-5 mm
19
long, membranous, fimbriate, truncate; sheaths terete, glabrous or sparingly hairy
from minute tubercles, smooth, shorter than the internodes. Inflorescence of 2 racemes,
one sessile, the other with a short hairy base. Racemes 8-10 cm long, densely whitish-
villous; joints and pedicels 3-5-4 mm long, narrowly cuneate-linear, transparent
between the thickened edges, densely hairy on the back and sides with the hairs
approximately 3 mm long, tips on one side produced into a short, membranous,
crenulate auricle. Sessile spikelets 6-8 mm long, oblong-lanceolate. Glumes subequal,
membranous; lower dorsally flattened, densely hairy, hairs approximately 4 mm
long, keeled, keels with very narrow subhyaline wings, intracarinal nerves 6-7, evenly
distributed, apex minutely bifid; upper narrowly boat-shaped, sparingly hairy, 3-nerved,
apex with a bristle 5-7-6 mm long. Lower floret reduced to a 5 mm long lemma,
linear-oblong, hyaline, nerveless, ciliate. Upper floret hermaphrodite; lemma 4-7-5
mm long, cuneate-oblong, 3-nerved below, 2-lobed to about the middle, awned;
lobes acute, ciliate; awn about 12 mm long, fine, bent at and twisted below the middle;
palea absent; anthers imperfect, 1-2 mm long. Pedicelled spikelets 6-8 mm long,
male or hermaphrodite. Glumes subequal; lower long hairy on the back, keels
narrowly winged, many nerved, nerves very fine and close with a prominent midnerve,
apex 2-mucronate with a fine bristle 4-5 mm long from between the mucros; upper
subhyaline, long hairy on the back, 3-nerved. apex acute or aristulate. Lower floret
with the lemma 4-5-5 mm long, oblong, faintly 2-nerved, ciliate, apex obtuse. Upper
floret with the lemma lanceolate-oblong, bifid, ciliate, often with a mucro from the
sinus or more or less reduced and entire with a male or hermaphrodite flower; anthers
2- 3 mm long.
This species has been recorded from Abyssinia, Uganda, Malawi, the Cape and
Natal. Tn the Republic it is restricted to the Drakensberg, where it grows on mountain
slopes at altitudes ranging from 5,000-9,000 feet.
Natal. — Estcourt: Ntabamhlope Mtn., West 41. Newcastle: Normandien Pass, Edwards 2819.
Cape. — Maclear: 23 miles N., Theron 2203.
Uganda. — Kigezi: Greenway & Eggeling 7144; Kashenji, Johnston 1375.
The holotype of this species was not seen but the South African material was
compared with the description and with a photograph of the holotype in the National
Herbarium, Pretoria. It is also identical with Schimper 223b, collected in Abyssinia
on Mt. Sholoda. The latter specimen is quoted by Stapf (1919) and is deposited
in Kew.
9. A. pilosellus Stapf in FI. Trop. Afr. 9: 221 (1919). Type; Eritrea, Amasen,
At-Zien, Pappi 5285.
Perennial, loosely tufted with a short oblique rhizome. Culms up to 20 cm long,
geniculate, slender, simple, 2-3-noded, glabrous, smooth. Leaves up to 10 cm long
and 3 mm broad, expanded, linear from a narrow base, sparingly hirsute on both
surfaces with bulbous based hairs, margins scabrid, apex acute; ligules 1-5 mm long,
membranous, obtuse; sheaths terete, sparingly hirsute with bulbous based hairs,
persistent. Inflorescence of 2 racemes, one sessile, the other with a short hairy base.
Racemes 3-5 cm long, straight, yellow-green to purplish; joints and pedicels linear,
3- 3 • 5 mm long, subtransparent between the thickened edges, ciliate on both margins
with hairs 1 • 5-2 mm long, tips cupular, truncate. Sessile spikelets hermaphrodite,
7 mm long, lanceolate-oblong; callus short, obtuse, densely bearded. Glumes subequal;
lower 7 mm long, lanceolate-oblong, 7-nerved, subchartaceous, dorsally shallowly
concave, long hirsute, unequally winged from the keels, one wing very narrow, scabrid,
apex unequally 2-dentate, each tooth with a fine mucro; upper 6 mm long, narrowly
boat-shaped, membranous, 3-nerved, shortly pilose on the margins, keel scabrid,
apex minutely bifid, awned, awn 3 mm long. Lower floret reduced to the lemma, 5 mm
20
long, oblong-lanceolate, hyaline, 2-nerved, sparingly pilose on the margins, apex acute.
Upper floret with the lemma 4 mm long, hyaline, 1-nerved, lobed to about the middle,
awned, awn 8 mm long, bent at and twisted below the middle; palea absent. Pedicelled
spikelets neuter, reduced to the glumes; lower 4-5 mm long, lanceolate-oblong, hirsute,
7- nerved, apex minutely bifid, shortly mucronate; upper 2 mm long, oblong, hyaline,
nerveless, apex truncate.
Natal.— Bergville: Drakensberg, low area pass between Saddle and Twins, Edwards 2133.
The plant, from which the above description was drawn up, was collected on the
Drakensberg in Natal in July, 1957, at an altitude of 9,200 feet. To my knowledge
this is the only record for South Africa. It compares very well with Stapf’s description
of A. pilosellus in the Flora of Tropical Africa (1919) and could possibly be a depauperate
specimen of that species. I have not seen the type of A. pilosellus, Pappi 5285, collected
at At-Zien in Eritrea, but Edwards 2133 was sent to Kew for comparison. Mr. W. D.
Clayton kindly examined it and reported that it was probably A. pilosellus but that
a definite identification would not be possible until a detailed study of the East African
montane species of Andropogon had been made.
10. A. lacunosus J. G. Anderson in Bothalia 8: 113 (1962). Type: Transvaal,
Ermelo, Athole Pasture Research Station, Gericke A. 49 (PRE, holo. !).
Perennial, caespitose with a short rhizome. Culms up to 80 cm long, erect, slender,
glabrous, striate, often grooved, terete or the basal internodes slightly compressed,
2-4-noded; nodes glabrous. Leaves 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 between the nerves, often densely pilose near the ligule; ligules 1 mm long,
membranous, ciliolate, truncate; basal sheaths often tinged with purple. Inflorescence
of 2-3 racemes, 5-5-7 cm long; joints and pedicels linear, 3-4-5 mm long, margins
pubescent. Sessile spikelets hermaphrodite, 6-8 mm long, linear. Glumes unequal;
lower 6-8 mm long, narrowly-lanceolate, deeply grooved on the back, the groove
densey pilose in the lower and with 5-8 variably shaped shallow pits on either side
of the groove, chartaceous, 7-9-nerved, keeled with the keels narrowly winged, ciliate
in the upper half, apex obscurely bifid; upper 5-6 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, lanceolate, hyaline, nerveless, margins inrolled and pilose in the
upper half, apex obscurely bifid. Upper floret bisexual with the stamens reduced to
staminodes; lemma 4-7-5 mm long, linear, hyaline, nerveless, lobed to about the
middle, lobes acute, pilose along the margins in the upper half, awned from between
the lobes; awn slender, 10-15 mm long; palea absent. Pedicelled spikelets male,
8- 9 mm long. Glumes unequal; lower 8-9 mm long, lanceolate, chartaceous, keeled,
keels and margins pilose, 11 -nerved with the nerves grouped together, one group of
3 in the middle and 2 marginal groups of 4 each, with 2 longitudinal grooves and 5—12
variably shaped, shallow pits on the back, midnerve excurrent into a very short bristle
or apex acute; upper 7-5-8 mm long, lanceolate, membranous, 3-5-nerved, obscurely
winged, pilose on the margins, apex acuminate. Lower floret reduced to the lemma,
6-5-7 mm long, hyaline, obscurely 2-nerved, pilose on 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, 4 mm long.
This species is rare and has been found only in the Ermelo, Groblersdal, Belfast,
Lydenburg, Pilgrims Rest and Wakkerstroom districts of the eastern and north-eastern
Transvaal. It occurs in vleis and near streams, usually in sandy soils, at altitudes
ranging from 4,700-6,800 feet.
21
Transvaal. — Belfast: Dullstroom, Hector s.n. Ermelo: Athole Pasture Research Station, Letole
River, Malherbe 54; Norval 8; 122. Groblersdal: 28-j miles N.E. by E., Acocks 20908. Lydenburg:
15 miles W. of Sabie, Codd 6441. Pilgrim’s Rest: Mariepskop, Meeuse 9940; near Reitz’s grave,
Van der Schijff 5992; 6 miles N. of Graskop, Codd 6450. Wakkerstroom : Oshoek, Devenish 1090.
A. lacunosus is a distinct species and is unique in the genus Andropogon because
of the shallow, variably shaped pits on the lower glumes of both the sessile and
pedicelled spikelets.
11. A. amplectens Nees, FI. Afr. Austr. 104 (1841); Steud., Syn. PI. Glum. 1 : 372
(1855); Hack, in DC., Monogr. Phan. 6: 453 (1889); Stapf in FI. Cap. 7: 342 (1898);
FI. Trop. Afr. 9: 243 (1919); Chippindall in Gr. & Pastures S. Afr. 1: 498 (1955).
Type: Cape, “ between Omtata and Omgaziana ”, Drege s.n.
Perennial, densely tufted. Culms up to 150 cm long, slender, few-9 noded, erect
or rarely subdecumbent, glabrous, smooth. Leaves glabrous, or rarely hairy to densely
pubescent; blades linear from a wider rounded base or those of the culms cordate
or subcordate and amplexicaul, tapering to a long setaceous apex, up to 30 cm long
and 10 mm wide at the base, those of the innovations up to 15 cm long and 2-5 mm
wide, subglaucous or suffused with purple, usually glabrous, rarely with a few scattered
hairs or densely pubscent, margins somewhat scabrid; ligules 1 mm long, membranous,
fimbriate, truncate or rounded; sheaths terete, tight, glabrous, the lowest and those
of the innovations short, sometimes purplish, subpersistent. Inflorescence of 2, rarely
3 racemes; spathes linear-lanceolate, long tapering upwards, glabrous, pale green, at
length tightly inrolled; spatheoles similar to the spathes, about 8 cm long, tapering
to a setaceous point; peduncles exerted. Racemes 5-8 cm long, straight or somewhat
flexuous; joints and pedicels very similar, 5-6 mm long, clavate or cuneate, distinctly
swollen upwards, densely ciliate on the margins, tips produced into very oblique, slightly
denticulate cupules. Sessile spikelets hermaphrodite, 7-9 mm long, laterally compressed,
wedged in between joint and pedicel, linear-oblong, glabrous, green, often reddish
towards the tips; callus acute to subacute, 2 mm long, shortly bearded. Glumes equal;
lower 7-5-8 mm long, lanceolate, keeled with the keels rounded, broad and smooth
below, acute and scabrid near the tips, narrowly grooved between the keels and with
2 fine lateral furrows in the upper third, apex bifid; upper 7-5-8 mm long, boat-shaped,
subcoriaceous, 1 -nerved, glabrous or the nerve dorsally scabrid, apex mucronate or
aristulate. Lower floret reduced to the lemma, 5-6 mm long, lanceolate, nerveless
or faintly 2-3-nerved, sparingly ciliate, hyaline, apex acute. Upper floret hermaphrodite;
lemma 5-7-6 mm long, linear-lanceolate, sparingly ciliate, 3-nerved, deeply lobed,
lobes acute, awned from between the lobes, awn stout, 3-5-5 cm long, kneed and
pubescent below the middle; palea 1-6-2 mm long, linear-oblong, nerveless, ciliate;
anthers 4 mm long. Pedicelled spikelets male, 9-12 mm long, lanceolate-acuminate,
dorsally compressed. Glumes subequal; lower 9-12 mm long, subherbaceous, linear-
lanceolate, many-nerved between the scabrid keels, usually produced into an awn
2-8 mm long; upper 9-11 mm long, linear-lanceolate, 3-5-nerved, sparingly ciliate
on the keels, apex acuminate. Lower floret reduced to the lemma, 6-9 mm long,
membranous, linear-lanceolate, faintly 3-nerved, apex acute. Upper floret with the
lemma 6-8 mm long, membranous, 1 -nerved, apex bifid; palea 2-3-5 mm long, mem-
branous, nerveless; anthers 4-4-5 mm long.
This species occurs from the East London district of the eastern Cape to Natal
and Swaziland and is widely distributed and common in the Transvaal. From here
it extends into the Orange Free State and northern Cape. It has also been recorded
from Bechuanaland, Mozambique, Rhodesia, Malawi and the Congo Republic. It
usually occurs in dry open veld and on hillsides, but is often found also in marshy
areas and other moist situations.
22
Cape. — Barkly West: Daniels Kuil, Gets 2999 (KMG). East London: Kidds Beach, Comins 1484
(GRA). Hay: Jaspar Hill at Clifton, Acocks 2142. Kentani: near Kei Mouth, Flanagan 981 ; Pegler
1056. Kuruman: Pole Evans 2414; Marloth 1114 (GRA). Mount Currie: Clydesdale, Tyson 931
(SAM). Tsolo: Umtwynkwa Cutting, Galpin 6893. Umtata: “ Morely ”, Drege s.n. (K, S). Vryburg:
Armoedsvlakte, Henrici 44: Bietjiesvlakte, Henrici 73.
Orange Free State. — Heilbron: Maccauvlei, Brandmuller 42. Kroonstad: Vais River, Pont 27;
143. Ventersburg: 12 miles S., Godfrey SH. 1563.
Natal. — Bergville: Ladysmith road, Edwards 661 . Camperdown: Nagle Dam, Wells 1190 (NU);
1865. Estcourt: Acocks 11454. Hlabisa: Hluhluwe Game Reserve, Ward 1977. Lions River: Moll
617. Nkandla; Qudeni, Edwards 1297. Pietermaritzburg; Albert Falls, Comins 265 (NU). Port
Shepstone: Port Edward, Huntley 720; Oribi Flats, McClean 528. Tugela: Bulwer Farm, Edwards
1920. Ubombo: Mpangazi Lake, Strey & Huntley 5017. Umzinto; Dumisa Station, Rudatis 884.
Upumulo: Buchanan 194 (BOL, NH).
Swaziland. — Mbabane: Ngwena Mtn. Range near Forbes Reef, Meeuse 10136; Ukutula, Compton
25787; 26698.
Transvaal. — Brits: Jacksonstuin, Van Vuuren 379. Delmas: Schmc/z 16 (GRA). Ermelo: Experi-
mental Farm, Burtt Davy 7720. Heidelberg: Henley-on-Klip, Stent s.n.; Uitgevallen 197, Burn Davy
13667. Johannesburg: 8 miles S., Lintner 45; Houghton, Cohen 650. Letaba: The Downs, Junod
4096; Duivvelskloof, Scheepers 822. Lichtenburg: Grasfontein, Sutton 392. Marico: Liebenberg
5321. Nelspruit: Nohpe, Van der Schijff 1512; Shabin Kop, De Winter & Codd 458. Pietersburg:
Iron Crown Peak, Swanepoel s.n. Potchefstroom: Burtt Davy 2700. Pretoria: Doornkloof, Pole
Evans 351; Hoorns Nek, Codd 770; Onderstepoort Station, Smith 6169; Premier Mine, Rogers
25048. Ventersdorp: Palmietfontein, Botha 26. Vereeniging: Leeukuil, Story 91. Warmbad: Pole
Evans & Smuts 690. Waterberg: Nooitgedacht 699, Acocks & Naude 31.
Bechuanaland. — Lobatsi: Kanye, Miller B/304.
Rhodesia. — Bubi : Gwampa Forest Reserve, Goldsmith 37/56. Matobo: Matopas, Rattray 300.
Que Que; Gokwe, Bingham 481A.
Mozambique. — Beira: Dondo, Torre 5899; Santaca, Maputo, Gomes <fc Sousa 3955.
I have not seen the type specimen of A. amplectens but the South African material
agrees with the description by Nees (1841) and with a specimen collected by Drege
simultaneously with the type. The latter specimen was recorded from “ Morely ” in
the Umtata district and as this is the type locality there is no doubt as to the identity
of the species.
A. amplectens is a distinct species with the culm leaves characteristically rounded
or subcordate and subamplexicaul at the base, tapering sharply to a long, fine, apex.
The leaves and culms may be dark green or frequently tinged with red, purple or brown.
In its early and flowering stages the plant is usually conspicuously glaucous.
The width of the leaf-bases and the pubescence of the vegetative parts of the plant
vary considerably. Examination of the South African material revealed that the leaf-
bases are from 2-5 mm to 1 - 3 cm wide. This variation occurs in plants from the same
locality and even in leaves on the same plant. For this reason 1 am reluctant to include
any of the broad-leaved South African plants in the var. diversifolius (Rendle) Stapf,
described originally from Angola.
It is of interest to note that plants collected along the coast in the Port Shepstone
district of Natal often have a subdecumbent or sprawling habit, with leaves 5-12 cm
long and 6-10 mm wide.
12. A. schinzii Hack, in DC., Monogr. Phan. 6: 458 (1889); Stapf in FI. Cap.
7: 341 (1898); FI. Trop. Afr. 9; 245 (1919); Chippindall in Gr. & Pastures S. Afr.
1: 499 (1955). Type: South West Africa, Amboland, Oshando, Schinz 670 (Z, holo. ! ;
BOL, iso.!; PRE, photo.!, fragment!).
A. schinzii Hack, ex Schinz in Verh. Bot. Ver. Brandenb. 30: 139 (1888), nom. nud.
Perennial, densely tufted with a slightly bulbous base. Culms up to 120 cm tall,
erect, slender, glaucous, glabrous, terete or often channelled, simple and sometimes
woody below with barren solitary or fascicled branches from the lower nodes and few
23
solitary (rarely 2-3-nate) leafy long and erect flowering branches higher up. Leaves
up to 30 cm long and 8 mm broad, linear from a narrow base, long tapering to a fine
apex, slightly rough or puberulous on both surfaces, rarely with some additional very
long and spreading hairs near the ligule; ligules 1-2 mm long, membranous, truncate,
laterally produced and adnate to the auricles of the sheath; sheaths terete, tight,
glabrous, firm, the lower and intermediate mostly longer than the internodes, shortly
auricled at the mouth, the basal long, persistent. Inflorescence of 2 racemes, conspi-
cuously hairy from the long soft hairs on the joints and pedicels; spathes subherbaceous,
gieen, sometimes faintly tinged with purple, at length tightly inrolled with a more
or less reduced blade. Racemes 5-8 cm long, one sessile, the other with a short base,
straight or flexuous, villous, dull or yellow-green, sometimes purplish; joints and
pedicels equal or subequal, cuneate, stout above, much attenuated towards the base,
4-5 mm long, rigidly ciliate on both margins, cilia 3 -5-4 -5 mm long, tips obliquely
cupular, minutely dentate. Sessile spikelets hermaphrodite (except the lowest 1 or 2
of the sessile raceme), 6-7 mm long, narrowly linear, laterally compressed, loosely
wedged in between joint and pedicel; callus short, obtuse, shortly bearded. Glumes
subequal; lower 6-7 mm long, subcoriaceous, keeled with a deep narrow hyaline
groove between the keels, nerves 4 on each side, one close to the groove, the other
in the keels; keels obtuse and smooth in the lower three-quarters, acute and scaberulous
upwards, apex minutely bifid; upper boat-shaped, 6-7 mm long, membranous, 3-nerved,
pilose on the margins, awned, awn 7-10 mm long, fine, straight. Lower floret reduced
to a lemma, 5-5-7 mm long, linear-oblong, 2-nerved. pilose on the margins, hyaline,
apex acute or obtuse. Upper floret hermaphrodite, lemma 5-6 mm long, oblong,
3-nerved below, pilose on the margins, conspicously 2-lobed, lobes acute, awned from
between the lobes, awn slender, 25-35 mm long, geniculate, scaberulous; palea 3-5 mm
long, linear-oblong, hyaline, nerveless; anthers 3-4 mm long. Pedicelled spikelets
male, dorsally compressed, oblong or lanceolate-oblong, acute, purplish or reddish.
Glumes subequal; lower flat or rounded on the back, 5-6 mm long, membranous,
keeled, keels ciliate, awned from a bifid apex, awn 5-7 mm long, fine, straight; upper
submembranous, 6-7 mm long, 3-nerved, pilose on the margins in the upper half,
awn 1-2 mm long. Lower floret reduced to the lemma, 6-7 mm long, hyaline, 2-nerved,
narrowly keeled, pilose on the margins. Upper floret with the lemma 6-7 mm long,
hyaline, 3-nerved, pilose on the margins, apex mucronulate; palea 4-5-5 mm long,
hyaline, nerveless, apex obtuse or acute; anthers 3-4 mm long.
A. schinzii is found in several districts of the Transvaal, the Hay, Postmasburg
and Kuruman districts of the northern Cape and in the central and northern part of
South West Africa. It has also been recorded from Swaziland. Bechuanaland, Rhodesia
and the Congo Republic. It usually grows on rocky slopes or ledges on hillsides.
Cape. — Hay: Floradale, Ferrar 47 ; 6098 (KMG); Esterhuysen 2368; Bingap, Acucks 5730 (KMG,
NH, BOL). Griqualand West: Klipfontein, Burchett 2164-2 (SAM). Kuruman: Broncote, Ester-
huysen 1152 (BOL, KMG). Postmasburg: Kaapstewel, Cooke s.n. (GRA); 6335 (KMG).
Swaziland. — Mbabane: Komati Bridge, Compton 26818.
Transvaal. — Krugersdorp: Gladysvale, Rodin 3930. Letaba: 174 miles W. of Letaba Camp,
De Winter & Codd 692; Duiwelskloof, Westfalia Estate, Scheepers 666. Messina: Govt. Ranching
Station, Du Plessis s.n.; Dongola, De Winter & Codd 286. Nelspruit: Skipberg, Van der Schijff 3397 .
Potgietersrus : Pyramid Estate, Galpin 8885; Swerwerskraal, Acocks 8869; Doornhoek, Walker s.n.
Pretoria: Klapperkop, Mogg 16231; Wonderboom Reserve, Repton 2070; Rust der Winter Dam,
De Winter & Codd 247. Rustenburg: ! lexkrantz, Codd 1066. Foster River, Bunt Davy 82. Waterberg:
104 miles N.E. of Oslo, Codd 3998.
Bechuanaland. — Francistown: between Totume and Bushman Mine, Pole Evans 3266; Gordon
87; 102.
South West Africa. — Grootfontein : Auros, Volk 594b; Farm Kumkauas, Kinges 2818; 2883;
20 miles W., Schoenfelder 26; 5 miles N., Schweickerdt 2063. Kaokoveld: Omberera, De Winter
& Leistner 5487. Otjiwarongo: Ozondjache, Volk 517. Rehoboth: Paulinenhof, Strey 2575. Wind-
hoek: Sittmann 3; between Haris and Aub, Pearson 9601 (BOL).
24
Rhodesia. — Makoni: Rusape, Davies 1075. Matobo: Matopos, Brain 3874. Shangani: Gwampa
Forest Reserve, Goldsmith 151/55.
A. schinzii is easily distinguished by its slightly bulbous base, the usually branched
culms and the glaucous-grey colour of the culms and leaves. Because of the long
soft hairs on the margins of the joints and pedicels, the racemes are villous in appearance.
A characteristic feature of this species is that the upper glumes of the sessile spikelets
as well as the lower glumes of the pedicelled spikelets are each provided with a straight
awn 5-10 mm long.
13. A. ravus/. G. Anderson in Bothalia 7: 417 (1960). Type: Natal, Bergville,
Cathedral Peak, Killick 1261 (PRE, holo. !).
Perennial, often with a branched system of knotty rhizomes. Culms up to 90 cm
long, usually 45-60 cm, erect, simple, 2-3-noded, glabrous, faintly striate or smooth.
Leaves up to 25 cm long, expanded, rarely almost filiform, 2-6 mm wide, linear, tapering
to a setaceous apex, narrowed towards the base or base slightly rounded, glabrous,
glaucous-grey or green, margins often scabrid; ligule 1-5-2 mm long, membranous,
scarious; sheaths glabrous, striate, pallid or upper suffused with purple, shorter than
the internodes, leaf-blade of uppermost sheath sometimes reduced to a short awn,
mouth often with a few long hairs. Inflorescence of 2 (very occasionally 3) spike-like
racemes, shortly pedicelled or lowermost sessile. Racemes 7-12 cm long, glaucous-grey
or green and usually suffused with dull purple; joints and pedicels 6-7 mm long,
cuneate-clavate, ciliate on the margins, cilia approximately 4 mm long. Sessile spikelets
hermaphrodite, linear-oblong; callus broad and obtusely rounded, bearded. Glumes
subequal; lower 6-9 mm long, dorsally with a deep, narrow or wide median groove,
thinly coriaceous; 3-nerved on each side of the groove, sharply keeled upwards, keels
narrowly winged, scabrid to ciliate in the upper half, apex mucronate; upper boat-
shaped, keel scabrid or ciliate in the upper third, finely ciliate on the margins, 1-3-
nerved, apex subacute. Lower floret with the lemma 6-8 mm long, 2-keeled with
sharply indexed margins, margins ciliate in the upper half, hyaline, 3-nerved, apex
obtuse ; palea a nerveless hyaline scale 2-3 mm long. Upper floret with the lemma
5-6 mm long, deeply lobed, lobes acute, awned from between the lobes, awn 1 • 5-2 cm
long, twisted below the knee; palea absent. Pedicelled spikelets male, dorsally
compressed. Glumes unequal; lower 7-12 mm long, conspicuously broad, keeled,
keels and margins ciliate or with stiff hyaline hairs, apex obtuse or mucronate; upper
shallowly boat-shaped, 6-9 mm long, 3 or faintly 5-nerved, margins ciliate, apex acute.
Lower floret with the lemma 7-8 mm long, margins indexed, sparsely ciliate in the upper
half, hyaline, 3-nerved, apex obtuse; palea a nerveless hyaline scale. Upper floret
with the lemma 5-6 mm long, margins indexed, apex obtuse; palea absent; anthers
linear, 4-5 mm long.
A. ravus is recorded from mountain slopes near Mbabane in Swaziland and in the
foothills of the Drakensberg from Bethlehem in the Orange Free State to Maclear in
the eastern Cape Province. It usually occurs in open grassveld at altitudes ranging
from 4,700-8,000 feet.
Cape. — Maclear: Lower Pitsing, Acocks 12191.
Orange Free State. — Bethlehem: Golden Gate, Story 1957.
Natal. — Bergville: Cathedral Peak, Edwards 1935; 1954 (NU); Mont-aux-Sources, Bayer &
McClean 179; Edwards 639; Cleft Peak, towards Indumeni, Gomel! s.n. Klip River: Van Reenen,
Phillips 8728.
Swaziland. — Mbabane: Ukutula, Compton 25593; Forbes Reef, Codd <£ Midler 305.
25
This species is usually glaucous-grey in colour. It can be confused with A.
amplectens but is distinguished from that species by the shape of the callus. In A.
ravus the callus is conspicuously broad and rounded, whereas it is acute or subacute
in A. amplectens.
A depauperate form of A. ravus has been collected near Cleft Peak in the Cathedral
Peak area. Except for being smaller in stature with narrow reddish leaves, I can
find no difference in spikelet characters. This form is represented by Gomel l s.n. and
Edwards 1935 and 1954.
14. A. filifolius ( Nees ) Steud., Syn. PI. Glum. 1: 374 (1855); Hack, in DC.,
Monogr. Phan. 6: 453 (1889); Stapf in FI. Cap. 7: 342 (1898); Chippindall in Gr.
& Pastures S. Afr. 1: 497 (1955). Type: Uitenhage, Zuurebergen, northern slopes,
Drege s.n. (S, lecto.!).
Heteropogon filifolius Nees, FI. Afr. Austr. 102 (1841). Syntypes: Olifantshoek,
Boschesman River, Ecklon s.n. (SAM!); Uitenhage, Tigerberg, Ecklon s.n.; Zuur-
bergen, Drege s.n. (S!).
Perennial, densely tufted. Culms up to 70 cm tall, simple, erect, glabrous, terete,
3-4-noded. Leaves mostly basal; blades up to 30 cm long, filiform, rarely expanded
and then 2-5 mm wide, firm, flexuous, glabrous or pubescent on the inside near the
ligule, apex acute; ligules coriaceous, 0-5-2 mm long, truncate; sheaths shorter than
the internodes, terete, tight, striate, glabrous, lowest persistent, at length breaking up
into fibres. Inflorescence of 2 spike-like racemes, one sessile, the other with a glabrous
base 7-11 mm long. Racemes 3-10 cm long, straight; joints and pedicels similar,
6-8 mm long, cuneate-linear, attenuated towards the base, rigidly ciliate on the margins,
tips crenulate, cupular. Sessile spikelets hermaphrodite, laterally compressed, wedged
in between joint and pedicel; callus acute to subacute, shortly bearded. Glumes
subequal, subcoriaceous, glabrous; lower 6-8 mm long, linear, keeled, keels rounded,
narrow, almost contiguous, narrowly grooved between the keels, apex obtuse, hyaline;
upper 6-8 mm long, boat-shaped, glabrous, 1-nerved, apex acute with a bristle 1 -3-1 -5
mm long. Lower floret reduced to the lemma, 5-5-6 mm long, lanceolate, membranous,
hyaline, 1-nerved or nerveless, ciliate on the margins, apex acute. Upper floret with
the lemma 4-5-5 mm long, linear, hyaline, 3-nerved, bilobed, lobes oblong, awned
from between the lobes, awn 3-5 cm long, stout, kneed at about the middle, pubescent
below; palea 2-7-3 mm long, ovate, hyaline, glabrous, nerveless, apex acute; anthers
3—4 - 5 mm long. Pedicelled spikelets male, lanceolate, dorsally compressed or subterete,
glabrous. Glumes unequal; lower 9-20 mm long, flat or rounded on the back,
herbaceous, acutely 2-keeled, keels narrowly winged, intercarinal nerves many, apex
lobed with a bristle from between the lobes, bristle about 2 mm long ; upper 8-14 mm long
lanceolate, 3-nerved, ciliate in the upper half or glabrous, apex finely acuminate. Lower
floret reduced to the lemma, 8-13 mm long, linear-oblong, 3-4-nerved, ciliate on the
margins, apex acuminate. Upper floret with the lemma 7-12 mm long, linear, 1-nerved,
glabrous or ciliate on the margins; palea 2-4 mm long, linear, hyaline, nerveless;
anthers 5-7 mm long.
A. filifolius is indigenous to Southern Africa and has been recorded from the
southern and eastern Cape, Basutoland, Natal, Swaziland and the northern and eastern
Transvaal. It grows in sour, open grassland on mountains and hillsides, often in rocky
soil.
Cape. — Albany: Cradock road, Wells 2758 (GRA); Grahamstown, Daly & Sole 109 (GRA).
Komga: near Kei Mouth, Flanagan 2348. Matatiele: Melkfontein, Story 537. Mount Ayliff: Kok-
stad, Goossens 268. Port Elizabeth: Van Stadensberg, Gillett 2384 (BOL).
Basutoland. — Butha Buthe: Moteng, Guillarmod 3842. Leribe: Dieterlen 397a.
26
Natal.- Bergville: Natal National Park, Galpin 10360; Cathedral Peak, Killick 949. Durban:
Clairmont, Schlechter 3092 (GRA). Estcourt: Giants Castle, Wood 10546; Griffins Hill, Acocks
10633; Ntabamhlope, West 470; 917; Mooi River, Mogg 3104. Lions River: Balgowan, Moll 1205.
Mpendle: Tillietudlem, Huntley 455. Nkandla: Codd 235; Qudeni, Fisher & Schweickerdt 30 (NH,
NU). Pietermaritzburg: Town Hill, Fisher 285 (NU); Bews 77 (NU). Utrecht: Edwards 630.
Swaziland. — Mbabane: Ukutula, Compton 24721; 24958.
Transvaal. — Barberton: Burtt Davy 87; Kaapsche Hoop, Pole Evans 1009. Belfast: Burtt Davy
88. Carolina: near Bossis, Burtt Davy 86. Ermelo: Maviristad, Pott 81 . Heidelberg: Vrisgewaagd,
Strey 3068. Letaba: The Downs, Junod 4092; Duiwelskloof, Scheepers 966. Pilgrim’s Rest: 12)-
miles from Graskop on road to Vaalbank, De Winter & Codd 221 ; Graskop, Kowyns Pass, De Winter
<& Codd 153. Waterberg: Krantzberg, Dyer & Verdoorn 4242.
This is a variable species in which several “ forms ” can be distinguished. These
forms differ mainly in size, leaf shape and length of the pedicelled spikelets. The most
common form occurs in the Drakensberg in the Cape, Basutoland and Natal and is
usually a smaller plant of which the leaves are all filiform, the ligules about 0-5 mm long,
the racemes 3-6-5 cm long and the pedicelled spikelets 9-12 mm long. A second
form is represented by plants with the leaves all filiform or filiform mixed with narrowly
expanded leaves, the ligules 0-5-2 mm long, the racemes 6-7 cm long and the pedicelled
spikelets up to about 17 mm long. An extremely robust form, represented by Dyer
& Verdoorn 4242, Pole Evans 1009, and Compton 24758, collected in the Waterberg,
Barberton and Mbabane districts respectively, has leaves which are filiform mixed
with leaves up to 5 mm wide, racemes 7-10 cm long and pedicelled spikelets 17-20
mm long.
Except for the variation in size of the various organs, no other noteworthy differences
could be found. In contrast to the variability in the length of the pedicelled spikelets,
the size of the sessile spikelets remains more or less the same. These forms intergrade
and should probably be regarded as ecotypes.
15. A. schirensis Hochst ex A. Rich. var. angustifolius Stapf in FI. Cap. 7: 340
(1898), sphalm “ angustifolia FI. Trop. Afr. 9: 247 (1919); Chippindall in Gr.
& Pastures S. Afr. 1: 497 (1955). Syntypes: Several. Lectotype: Orange Free State,
Thaba Nchu, Burke 434 (K, lecto.!; SAM!; PRE, fragment).
A. amplectens Nees var. natalensis Hack, in Mem. Herb. Boiss. 20: 9 (1900). Type:
Natal, near De Beers, Drakensberg, Wood 6002 (NH, iso.!).
Perennial, densely tufted. Culms up to 120 cm long, erect, slender, glabrous
and smooth, terete, simple, rarely branched, 3-4-noded. Leaves with the blades up
to 35 cm long and 7 mm wide, expanded, linear from a narrow base, long tapering
to a fine apex, glaucous or very often reddish, glabrous or often pubescent or scabrid
at the base, margins scabrid; ligules 1-5 mm long, membranous, truncate; sheaths
terete, smooth, rarely pubescent near the ligule, shorter than the internodes. Inflo-
rescence of 2 racemes (rarely 2-5), one sessile, the other with a base about 6 mm long.
Racemes 5-8 cm long, straight or flexuous, conspicuously dorsiventral exhibiting
on the one side the sessile spikelets with the joints and pedicels and on the other side
the broad backs of the pedicelled spikelets with the dense cilia of the pedicels between
them; joints and pedicels very similar, clavate-cuneate, 4-5 mm long, densely ciliate
on both margins, cilia usually shorter in the lower portion, 4-5 mm long towards the
apex, tips obliquely cupular or the pedicels produced into a broad tooth. Sessile
spikelets hermaphrodite, narrowly linear, laterally compressed, glabrous, gieen to
reddish-brown, tightly wedged in between joint and pedicel; callus 1-1.5 mm long,
obtuse to subacute, shortly bearded. Glumes equal; lower 5-6 mm long, linear-
lanceolate, thinly coriaceous, keeled with a narrow, deep hyaline groove between the
keels, keels obtuse, smooth below and scaberulous above the middle, 1 -nerved in the
keels, apex obscurely bifid ; upper 5-6 mm long, rounded on the back up to beyond
the middle then slightly compressed and keeled, subcoriaceous, 3-nerved with the side
27
nerves indistinct, margins very narrow, ciliate, apex acute to mucronulate. Lower
floret reduced to the lemma, 5-5-5 mm long, linear-oblong, membranous, hyaline,
2-nerved, winged, shortly ciliate on the wings, apex subobtuse. Upper floret with
the lemma 4-5 mm long, linear-oblong, membranous, hyaline, 3-nerved, deeply 2-lobed,
lobes acute, lanceolate, ciliolate, awned from between the lobes, awn 2-5-3 cm long,
bent and twisted, slender, scaberulous; palea 3-6-4 mm long, ovate-oblong, hyaline,
nerveless, apex obtuse; anthers 3-5 mm long. Pedicelled spikelets male, 7-8 mm
long and 1-2 mm wide, dorsally much compressed, reddish or purplish-brown. Glumes
more or less equal; lower 7-8 mm long, linear-lanceolate, subherbaceous, 7-15-nerved,
keeled, keels acute, scabrid or rigidly ciliolate, apex acute or acuminate; upper 7-8
mm long, membranous, 3-nerved, ciliolate on the margins, apex acuminate. Lower
floret reduced to the lemma, 7-7-5 mm long, narrowly-oblong, membranous, hyaline,
2-nerved, ciliolate on the margins, apex acute. Upper floret with the lemma 6-7 mm
long, linear, hyaline, 1-nerved; anthers 5 mm long.
This variety is common in Natal and the northern, eastern and central Transvaal.
It also occurs in the Orange Free State, Swaziland, the Barkly West, Kuruman,
Mafeking and Vryburg districts of the northern Cape and in South West Africa. Further
north it has been recorded from Angola, Rhodesia, Zambia and Malawi.
It grows mainly in open grassland or on hillsides, usually in stony or sandy soil,
occasionally in loamy soil on the edge of pans or along riverbanks.
Cape. — Barkly West: Daniels Kuil, Esterhuysen 2044; Pillans 2089. Kuruman: 11 miles N.W.,
Leistner 598. Mafeking: near Mosita, Brueckner 371. Vryburg: between Vryburg and Tigerkloof,
Brueckner 1087; Jakkalskop, Scheepers 2757.
Basutoland. — Leribe: Dieterlen 397b.
Orange Free State. — Ficksburg: farm "Aletta ”, Mat thy sen 11. Kroonstad: Vais River, Pont
131. Parys: Cohen 101. Winburg: Pretorius Game Reserve, Van Zinderen Bakker 1076; 1170.
Natal. — Bergville: Sentinel Park to Mont-aux-Sources, Edwards 639 (NU). Dundee: near Wasch-
bank, Pentz & Acocks 10282; near Rorkes Drift, Edwards 673. Durban: Berea, Broad 104
(NU). Estcourt: Giants Castle, West 1618 (NH); Symons 15700. Hlabisa: St. Lucia Estuary,
Feelv, Tinley & Ward 3; W. of Charters Creek, Feeiy, Tinlev & Ward 7. Kliprivier: Van Reenen,
Schlechter 6983 (BOL, GRA). Lions River: Mooi River, Rietvlei, Buchanan 193; 196; Nottingham
Road, McClean 948 (NH). Mpendle: Clarkson 156 (NU). Mtunzini: Ngoya Forest Reserve, Huntley
641; 644 ; 815. Newcastle: Road to Utrecht, Edwards 624; Buchanan 195 (NH). Nkandla: Qudeni
Forest Reserve, Fisher & Schweickerdt 28 (NH, NU). Pietermaritzburg: Table Mtn., McClean 195;
Bews 75 (NU); Inchanga, Most 92 (NU). Utrecht: Amajuba, Burtt Davy 154. Vryheid: Hlobani,
above Mkusi River, Johnstone 387 (NU). Weenen: Farm Schurfdepoort, West 1799 (NH).
Swaziland. — Mbabane: Ukutula, Compton 26696; Forbes Reef road, Compton 25742; Forbes
Reef, Meeuse 10134.
Transvaal. — Barberton: Thorncroft 73a. Belfast: Machadodorp, Williams 6315; Pole Evans
H. 16962. Brits: Jacksonstuin, Mogg 14277; Van Vuuren 457. Carolina: Van der Schijff 2605.
Ermelo: Billys Vlei, Burtt Davy 78. Heidelberg: Henley-on-Klip, Stent s.n. Johannesburg: Hougton,
Cohen 655. Letaba: Westfalia Estate, Scheepers 519; 1236. Lichtenburg: road to Zeerust, Kinges
1415. Lydenburg: Mt. Anderson, Strey 3036. Marico: Liebenberg 5323. Middelburg: Burtt Davy
152. Nelspruit: Shabin, Van der Schijff 1470. Pietersburg: 12 miles E., Van Vuuren 1455. Pilgrim's
Rest: Mariepskop, Van der Schijff 6151; 6 miles N. of Graskop, Codd 6450; Iron Crown Mtn.,
Meeuse 9849. Potgietersrus : Moord Drift, Norlindh & Weimarck 5241. Pretoria: Meintjies Kop,
Pole Evans 364; Hoorns Nek, Codd 769; Rietvlei Reserve, Repton 3227. Vereeniging: Burtt Davy
17194 (BOL). Waterberg: Sandrivierspoort, Codd 921 ; Towoomba, Acocks 8892. Witrivier: Numbi,
Van der Schijff 1605.
South West Africa. — Grootfontein : 30 miles N. of Gautscha Pan, Story 6472; Aha Mts.. Story
6387; between Cigarette and Tsammagaigai, Maguire 2509; Otavi, Dinter 5745. Otjiwarongo: Water-
berg, Volk 1077; Waterberg Plateau, De Winter 2791; 2805. Okavango: 8 miles E. of Tamso, De
Winter & Marais 4682; 4 miles S.W. of Nzinzi, De Winter & Wiss 4008; 5-8 miles E. of Schamvura
Camp, De Winter & Wiss 4435.
Rhodesia. — Bubi: Gwampa Forest Reserve, Goldsmith 19/56. Inyanga: Norlindh & Weimarck
4804. Salisbury: Eyies 1493; Brain 2443.
28
A. schirensis var. angustifolius is characterized by its conspicuously dorsiventral
racemes. On the one side the sessile spikelets with the joints and pedicels can be seen,
whereas the broad backs of the pedicelled spikelets with the dense cilia of the pedicels
between them are evident when viewed from the other side. The whole plant, especially
the racemes, is usually flushed with a reddish or purplish-brown colour.
Typical A. schirensis apparently does not occur in South Africa although several
specimens approach it very closely, in particular those from South West Africa. These
include: De Winter & Marais 4682, De Winter & Wiss 4008, Maguire 2509, Volk 1077,
Scheepers 1236, Van Vuuren 457 and Aeocks 8892. These specimens were compared
with tropical African material and it was found that the plants are smaller and less
robust than in typical A. schirensis and that the leaves, although comparatively wide,
are shorter.
A definite conclusion can be reached only after a study of a more extensive range
of African material has been undertaken.
16. A. gayanus Kunth var. squamulatus ( Hochst . ex A. Rich.) Stapf in Fl.Trop. Afr.
9: 263 (1919); Chippindall in Gr. & Pastures S. Afr. 1: 498 (1955). Type: Abyssinia,
Ferrfera Valley, Schimper 715 (K, iso.!; PRE, photo.!).
A. squamulatus Hochst. ex A. Rich., Tent. FI. Abyss. 2: 454 (1851); Hochst. in
Flora 27: 244 (1844), nom. nud. ; Steud., Syn. PI. Glum. 1: 374 (1855). A. gayanus
Kunth var. cordofanus Hack, in DC., Monogr. Phan. 6: 448 (1889), in part, excl.
Kotschy 54. A. helophilus Schum. in Pflanzenw. Ost-Afr. 5 : 98 (1895). Type: Usam-
bara, Hosigo, Holst 2535 (K, iso.!). A. hylophilus Schum. in Abhandl. Preuss. Akad.
Wiss. Berlin 38 (1894), nom. nud.
Perennial, tufted, glaucous. Culms up to 3-5 m long, stout, terete, glabrous,
smooth, shiny, many-noded with few to many flowering branches. Leaves glabrous
or softly pubescent, rarely villous or tomentose; blades up to 40 cm long and 18 mm
wide, glaucescent or ultimately reddish, expanded, linear to lanceolate-linear, the
lower portion usually much attenuated and often reduced to the midrib, widening
upwards and tapering to a fine apex, midrib prominent, margins scabrid; ligules
2-5-3 mm long, rounded or truncate, scarious, ciliolate; sheaths terete, striate, tight,
if wider than the base of the blade then forming more or less rounded shoulders or
only very shortly produced into auricles adnate to the ligule. Inflorescence of 2 racemes
appearing hairy from the long soft hairs on the joints and pedicels, one sessile, the
other with a glabrous base 7-10 mm long; spathes usually with well developed blades
except in the uppermost, glabrous. Racemes 6-10 cm long, more or less erect and
straight; joints 5-6 mm long, stout, cuneate-clavate, swollen upwards, densely ciliate
on both margins, cilia about 4 mm long, apex cupular, dentate; pedicels similar but
the cupular apex produced into a tooth on each side. Sessile spikelets hermaphrodite,
narrowly-oblong, pale green; callus obtuse, bearded at the base. Glumes equal;
lower 7-9 mm long, chartaceous, narrowly-oblong, scabrid or shortly ciliate on the
upper half of the keels, dorsally only shallowly grooved, 5-7-nerved on each side of
the groove, apex obtuse or minutely bifid; upper 7-9 mm long, boat-shaped, subchar-
taceous, scabrid on the upper half of the keel, ciliate on the margins, 3-nerved with the
midnerve distinct, the lateral shorter and finer, apex obtuse oi minutely 3-dentate.
Lower floret reduced to a linear-oblong lemma, 7-8 mm long, membranous, hyaline,
2-nerved, margins pilose, apex acute or truncate and ciliolate. Upper floret with the
lemma 6-7 mm long, membranous, hyaline, 1 -nerved, oblong, 2-lobed, lobes acute,
awned from between the lobes, awn 2-2-5 cm long, kneed at about the middle,
scaberulous; palea 3-7-4 mm long, oblong, membranous, hyaline, 2-nerved, obtuse
or minutely 2-dentate; anthers 4 mm long. Pedicelled spikelets male, linear-oblong,
29
scaberulous, pale-green. Glumes equal; lower 7-9 mm long, chartaceous, linear-
oblong, dorsally flattened, keeled, scabrid on the keels, many-nerved, the midnerve
distinct and produced into a straight, scaberulous awn 4-6 mm long, apex minutely
bifid or acute; upper 7-9 mm long, narrowly-oblong, membranous, 3-5-nerved, margins
shortly pilose, apex acute, acuminate or mucronulate. Lower floret reduced to the
lemma, 7-8 mm long, membranous, hyaline, 2-3-nerved, linear-oblong, margins shortly
pilose, apex obtuse or acuminate. Upper floret with the lemma 6-7 mm long, mem-
branous, hyaline, 1-nerved, glabrous, apex minutely 2-dentate or mucronate; palea
2-7-3 mm long, more or less oblong, membranous, hyaline, ciliolate on the margins,
apex obtuse; anthers 4 mm long.
A. gayanus var. squamulatus has been recorded from Abyssinia, the Sudan, Nigeria,
Ghana, the Congo Republic, Uganda, Tanzania, Zambia, Rhodesia and Mozambique.
In South Africa it is confined to the northern parts of Natal and Transvaal, Swaziland,
Bechuanaland and northern South West Africa. It usually grows in open grassland
in rocky situations and almost invariably in sandy soil.
Natal. — Hlabisa: W. of Charters Creek, Ward 2859; False Bay Park, Ward 4151. Ubombo:
Strey & Huntley 5011; Mkuzi Game Reserve, Ward 3996.
Swaziland. — Stegi: Malinda Hills, Compton 29790.
Transvaal. — Letaba: Duiwelskloof, Scheepers 639; between Komatipoort and Letaba River,
Rogers 2657. Pietersburg: Leipzig, Schweickerdt 1822. Pilgrim’s Rest: Orpen Dam, Van der Schijff
3473. Sibasa: Punda Milia, Van der Schijff 1826; Acocks 16765; Codd 4212; 5374; Godfrey H. 1715.
Soutpansberg: 5 miles N. of Louis Trichardt, Van Vuuren 1683.
Bechuanaland. — Francistown, Makarikari Pan, Pole Evans 3269; 3303; between Totume and
Bushman Mine, Pole Evans 3267 ; Lower Ngwezimba River, Miller s.n.; Ngamiland, Tsodito Hills,
Banks 45.
South West Africa. — Grootfontein : Aha Mtns., Story 6337; Gemsboklaagte, Schweickerdt 2064;
Otjihenena, Volk V. 68. Otjiwarongo: Bradfield 376; Otjiwarongo-Grootfontein Road, Liebenberg
4859; Waterberg, Volk 423; 999; 1078; Liebenberg 4802. Okavango: Andara Mission Stn., De
Winter & Marais 4827.
Angola. — Huila: Teixaira 55 3; 3459; 3575.
Rhodesia. — Hartley: Makwiro, Mainwaring s.n. (Eyles No. 2624). Marandellas: Eyles 4015.
Matobo: West 2158. Salisbury: Sturgeon 57588. Shangani: Goldsmith 48928.
Mozambique. — Cabo Delgado: 5 km from Mocamboa da Praia, Gomes e Sousa 4693. Lourenco
Marques: Manica, De Lernos & Macuacua 92. Niassa: Maudimba, Hornby 3340.
Zambia. — Chilinga: Sandwith 70. Fort Jameson: Senegallia farm, Grout 124.
Andropogon gayanus var. squamulatus differs from the other South African species
of Andropogon in being more robust with broader leaves. Like A. schinzii the culms
are usually branched, producing few to many flowering branches. Three well defined
varieties of A. gayanus are upheld by Stapf in the Flora of Tropical Africa (1919), of
which only the var. squamulatus is known to occur in South Africa. This variety differs
from typical A. gayanus in that the pedicelled spikelets are scaberulous with the joints
and pedicels ciliate on both margins.
17. A. distachyos L., Sp. PI. 1046 (1753); Steud., Syn. PI. Glum. 1: 372 (1855);
Hack, in DC., Monogr. Phan. 6: 461 (1889); Stapf in FI. Cap. 7: 343 (1898); FI.
Trop. Afr. 9: 218 (1919); Chippindall in Gr. & Pastures S. Afr. 1 : 496 (1955). Type:
Europe, Helvetia, Linn. Herb. 1211.10 (PRE, photo.!).
Perennial, caespitose with a short oblique rhizome. Culms up to 120 cm long,
erect, slender, terete, simple or branched towards the base, glabrous, 3-7-noded, lower
nodes often hairy, basal internodes usually naked. Leaves 5-30 cm long, 1-4 mm,
rarely 5 mm wide, expanded, linear, tapering to a long and sometimes setaceous apex,
rigid to flaccid, hairy or glabrous but usually with pubescent basal shoots, often glabre-
scent, more or less rough above and scabrid on the margins; ligules 2 mm long, mem-
branous with a fimbriate rim, truncate, sometimes pilose from behind; sheaths glabrous
30
or the lower hairy. Inflorescence of 2, rarely 3 racemes, the one sessile, the other shortly
peduncled, rather robust, straight or curved, greenish, often tinged with a dull or livid
purple, 5-10 cm long; joints cuneate-linear, hollow and very slightly inflated, 3-5-4 mm
long, semi-transparent along the middle, glabrous to subpubescent on the back, ciliate
on the outer margin, cilia 1-2 mm long, tips slightly cupular; pedicels similar, slightly
longer and more translucent. Sessile spikelets hermaphrodite, 9-12 mm long, broad-
lanceolate to acuminate, pale-green, often purplish towards the apex; callus short,
obtuse, shortly bearded. Glumes unequal; lower subherbaceous or chartaceous,
dorsally flattened, glabrous or puberulous, keels broadly winged, wings membranous,
whitish or purplish, intracarinal nerves 7-11, partly evancescent below; upper distinctly
shorter than the lower, boat-shaped, rigidly membranous, 3-nerved, softly ciliate,
tips minutely 2-dentate with a bristle 4-6 mm long. Lower floret reduced to the lemma,
5-7 mm long, oblong, acute or subacute, 2-nerved. Upper floret with the lemma 5-7
mm long, slightly gibbous on the back, 2-lobed to below the middle, chartaceous and
3-nerved below, lobes lanceolate, glabrous, awned from between the lobes, awn slender,
about 2-5 cm long, kneed at and twisted below the middle; palea minute or absent;
anthers 3 mm long; grain oblong, 2 mm long. Pedicelled spikelets similar to the
sessile but narrower. Glumes unequal, lower 9-1 1 mm long, subherbaceous to char-
taceous, very narrowly winged and terminated by a bristle 4-6 mm long; upper mem-
branous, 3-nerved, shortly aristulate. Lower floret with the lemma 5-7 mm long,
delicately hyaline, ciliolate, faintly 2-nerved. Upper floret with the lemma 4-5 mm
long, delicately hyaline, glabrous, 1 -nerved, apex slightly notched.
This species occurs in southern Europe, throughout the Mediterranean region,
tropical Africa and southern Arabia. In South Africa it has been recorded from
mountain slopes in the Drakensberg areas of Natal, the Orange Free State and the
northern Transvaal. It usually grows in rocky situations in open grassland, often
preferring wet places.
Orange Free State. — Harrismith: Wittebergen, Buchanan 263 (BOL). Vrede: Quagga Kop,
Acocks 12757.
Natal. — Bergville: Cathedral Peak, Chippindall 355. Estcourt: Cathkin Peak, Edwards 2006;
Mountain hostel, West 145; Ntabamhlope, A cocks 1 1460; 11480. Kliprivier: Codings Pass, Acocks
13487. Underberg: McClean 601.
Transvaal. — Pilgrim's Rest: Mariepskop, Van der Schijff 5583.
Linnaeus originally described this species as A. distachyon. The generic name
Andropogon is, however, masculine and, if Recommendation 75A of the International
Code of Botanical Nomenclature (1961) is taken into consideration, the specific name
should be changed to distachyos.
A. distachyos is readily distinguished from the other South African species by
the broadly winged lower glumes of the sessile spikelets. I have not seen type material
of this species but the South African material was carefully compared with the
description of the species. Furthermore, Chippindall 355 was compared with European
specimens at Kew by Dr. B. de Winter who states that it compares well except that
the South African specimen is slightly larger. I have no doubt that the South African
material is indeed A. distachyos L.
Bews s.n., collected at Himeville in the Underberg district, proved to be of interest.
This specimen, deposited in the herbarium of the Natal University (NU 3539), is
vegetatively similar to A. schirensis var. angustifolius. Even the colouring of the leaves
and racemes is typically reddish-brown. Examination of the spikelets revealed, however,
that the lower glumes of the sessile spikelets are dorsally flattened and broadly winged
as in A. distachyos. Except for minor details the spikelets are typical of this species.
As both A. schirensis var. angustifolius and A. distachyos occur in this area, this specimen
is most probably a hybrid between these two species.
31
Bothalia 9, 1: 31-96.
The Genus Asparagus in Southern Africa*
by
J. P. Jessop
Abstract
A revision of the genus Asparagus in South Africa, South West Africa, Bechuanaland, Basutoland,
and Swaziland has been undertaken.
Notes are given on the value of most of the characters which have been used in the separation
of species, recent literature on the nature of Asparagus assimilatory organs, and a technique for the
examination of chromosomes. Chromosome counts are listed for ten taxa.
There is a key to the forty species and four varieties. In the main part of the work these taxa
are described, and their synonymy, taxonomy, distribution and habitats dealt with. Six species and
one variety are new.
The following are the new species and combinations: A. setaceus (Kunth) (Asparagopsis setacea
Kunth), A. mucronatus, A. macowanii Bak. var. zuluensis (N. E. Br.) (A. zuluensis N. E. Br.), A. rigidus,
A. densiflorus (Kunth) (Asparagopsis densiflora Kunth), A. aethiopicus L. var. angusticladus, A. falcatus
L. var. ternifolius (Bak.) (A. aethiopicus L. var. ternifolius Bak.), A. aspergillus, A. obermeyerae, A.
krebsianus (Kunth) (Asparagopsis krebsiana Kunth), A. acocksii, A. crassicladus.
Several plants of horticultural importance occur in South Africa. The three best known are
A. plumosus, which is reduced here to synonymy under A. setaceus (Kunth) Jessop, and A. sprengeri
and A. myersii. A. sprengeri is being reduced to synonymy under A. densiflorus (Kunth) Jessop,. A.
myersii, which is a nomen nudum, is also regarded as belonging to A. densiflorus.
Introduction
The earliest record of a South African Asparagus, which can be authenticated,
is that of an artist, probably Hendrik Claudius, who accompanied Governor Simon
van der Stel on an expedition to Namaqualand during the second half of 1685 and
the beginning of 1686. One of the illustrations which he made was of A. capensis L.
Subsequently several other species were described, and some of them illustrated, but
by 1753 Linnaeus could include only five South African species of Asparagus in his
Species Plantarum. Thunberg, as a result of his own collecting, brought this number
up to 13 in his Prodromus. The next major contribution was that of Roemer and
Schultes, Systema Vegetabilium (1829), in which nearly thirty species were mentioned.
Many new names, many of them superfluous, were added by Kunth in 1850 when he
divided the genus as it is defined at the present into three genera. He referred seven
of his species to Myrsiphyllum, which consisted of plants with solitary, flattened cladodes
and axillary flowers. The remaining species were divided into Asparagus, with
supposedly dioecious flowers and eight South African species, and Asparagopsis, with
23 South African species.
Baker, whose monograph was published in 1875, provided the most recent complete
work on the genus. In this monograph. Baker recognised 40 species from South Africa
and, 21 years later, in the Flora Capensis, he included an additional four.
Since Baker, several authors have added an occasional name, based especially
on South West African material. The most important work has been that of Schinz
on South West African plants in 1896, Salter, who revised the genus for the Flora
of the Cape Peninsula, and Kies, who revised part of the section Africani in 1951.
Phillips (1951) estimated that the total number of South African species was about 70.
* Based on an M.Sc. thesis, submitted to the University of Cape Town in 1964
7230691-2
32
Partly as a result of these additional species, and in the absence of a revision of
the whole genus, or even the South African species, the available classification was
found to be in considerable need of revision. Salter stressed this need in the Flora
of the Cape Peninsula in 1950, and an examination of such works as Wilman’s Preli-
minary Check List of the Flowering Plants and Ferns of Griqualand West shows a
large number of unidentified specimens.
It was, therefore, decided to undertake a complete revision of the genus in South
Africa, accompanied where possible by field work. Most of this work was undertaken
while the author was a student at Cape Town University. The assistance of many
members of the staff of both the Bolus herbarium and the University of Cape Town
Botany Department is gratefully acknowledged. A special debt is owed to Dr. E. A.
Schelpe who supervised the work and provided much valuable advice. The revision
was completed at the National Flerbarium, Pretoria, where several members of the
staff provided very useful advice. In particular Mrs. A. A. Mauve’s knowledge of
the Transvaal species has been a great help. Acknowledgements are due to the
University of Cape Town Botany Department for the photographs of Asparagus
mucronatus and A. crassicladus. The remaining illustrations were prepared at the
Botanical Research Institute, Pretoria.
Specimens have been loaned, or other assistance has been provided, by the directors
and curators of the following herbaria : The Albany Museum Herbarium; the Herbarium,
Berlin-Dahlem ; The British Museum; the Herbarium of Trinity College, Dublin; the
Royal Botanic Gardens, Kew; the Herbarium of the University of Kiel; the National
Botanic Gardens and South African Museum herbaria at Kirstenbosch ; Rijksherbarium,
Leyden; the Botanische Staatssammlung, Munich; the Natal Herbarium; the National
Museum of Natural History, Paris; the Stellenbosch Herbarium; the Botanic Gardens
and Museum, Uppsala; the Windhoek Herbarium; and the Botanische Garten und
Museum der Universitat, Zurich.
In the distribution lists all specimens cited have been examined by the author.
In the larger species not all specimens seen have been cited, but a specimen from each
magisterial district has been selected for citing. When no herbarium abbreviation
follows the citation of a specimen, a specimen of that collecting is in the National
Herbarium, Pretoria. When there is no specimen in the National Herbarium only
one herbarium is cited for each collecting although duplicates may have been seen
from other herbaria.
Taxonomic Characters
Up to the present, there has been no satisfactory delimitation of the species. This
is partly because no-one has undertaken a major revision of the group since the Flora
Capensis in 1896, and partly because, in the absence of diagnostic floral characters,
most of the species must be defined on vegetative characters. These are often extremely
susceptible to environmental factors, and this is aggravated by the wide distribution
of many of the species.
A list of the characters of use in defining taxa below generic level is given here,
with notes on the general value of the different characters.
(a) Perianth segments
All have six, fairly small (1-6 mm long), white or near white tepals. The members
of the inner whorl are sometimes slightly broader than the outer segments, and the
distal margins are sometimes slightly ciliate. These characters are frequently variable
within a single species. A few species are characterised by having the distal half of
the tepals reflexed.
33
{b) Stamens
The six stamens sometimes have a pair of minute lateral spurs at the bases of their
filaments. This character was first used in Salter’s treatment in the Flora of the Cape
Peninsula, and is probably of value. The anthers vary from greenish through yellow
to orange and black. Work on fresh material suggests that colour is useful, but on
dried specimens the colour can usually not be determined.
(c) Fruit
The fruit is generally a wrinkled, fleshy berry with two or three seeds, and globose
in shape, but dry, smooth fruits as in A. exuvialis, and single-seeded berries as in
A. scandens, occur. One species, A. crispus, has an ovoid berry, and A. glaueus has
a persistent perianth enclosing the fruit. The fruit characters are usually reliable at
specific level.
( d ) Peduncles
All species have flowers borne solitarily on the ultimate flowering stalks. In
A. stipulaceus and A. capensis these stalks are reduced to discs. As all leaves are reduced
to scales, it is not strictly correct to refer to umbels or racemes. By treating the cladodes
as leaves, the use of the conventional terminology will be of value. The term peduncle
will be used for the flower stalk except in the section Racemosi, where the term pedicel
will be used for the final flower stalks and peduncle for the main axis of the inflorescence.
Inflorescences (termed racemes here) consisting of branches bearing only a series of
bracts and pedicels occur. However, in species producing racemes, cladodes some-
times appear at the bases of the pedicels. In two species, A. minutiflorus and A. rigidus,
the inflorescence, which appears to be a reduced raceme, consists of a peduncle, one to
three terminal pedicels, and sometimes one lateral pedicel. In some species both
terminal and axillary flowers occur as in A. africanus, but in others terminal, as in
A. setaceus, or axillary peduncles, as in A. scandens, are diagnostic. The umbels,
referred to by Baker, are clusters of terminal flowers, and are rarely found as the only
type of inflorescence on a plant. However, exceptions to this occur in the sections
Striati and Capenses. The peduncles may be solitary, paired or numerous, and these
characters are sometimes useful at specific level. In all species the peduncles are
articulated and in A. nelsii sometimes twice.
(e) Roots
The roots may be tuberous or fibrous. Classification cannot usually be based
on further subdivision of tuber type as the tubers are particularly susceptible to
environmental conditions. This is discussed further under A. asparagoides. In many
species there is a single root-stock bearing lateral tubers, but in most species of the
section Racemosi there are several roots each bearing lateral tubers. This condition
is preceded in young plants by the direct attachment of the tubers to the root-stock.
(/) Stems
Roots and aerial stems are produced from perennial rhizomes in all species. Annual
and perennial aerial stems occur, but the duration of the life of the aerial stem may
vary under different conditions. In A. exuvialis the aerial stem is perennial, but the
cladodes are deciduous, and after a month or two are replaced. Some species are
predominantly climbers and others erect. In each of these groups exceptions may be
produced by the effect of the environment, particularly the presence or absence of
support and the amount of light reaching the plant. The aerial stem may have a smooth,
wrinkled (Fig. lb), grooved (Fig. la), ridged (Fig. lc) or hispid surface. These characters
are sometimes useful. In A. exuvialis the outer layers of the aerial stems and branches
peel. (Fig. 2d). In several species, including A. compactus and A. retrofractus, the
grooved white outer layers disintegrate, leaving the stem smooth and brown. Another
character of occasional value is aerial stem colour, as in A. denudatus, which has a green
aerial stem.
34
c, A. aethiopicus var. aethiopicus, ridged.
fiteiPR vt.MWwe,
35
( g ) Branchlets
Branchlets, which are ultimate, specialised branches bearing cladodes and flowers,
occur in some species, especially in the section Capenses, where they are fascicled.
(Fig. 2b).
(h) Cladodes
Up to the beginning of the nineteenth century, the photosynthetic organs of
Asparagus were regarded as leaves. By the middle of the twentieth century, however,
it had come to be accepted by most botanists that they are modified branches. Several
conflicting theories exist, and these need to be considered. The three most important
theories propounded this century are those of Arber (1924 & 1925), Schlittler (1953
& 1959) and Kaussmann (1955).
Arber divided the genus Asparagus (sensu lato) into Myrsiphyllum (containing
A. asparagoides and A. undulatus ), in which she considered that the assimilatory organs
were leaves, and Asparagus in which she considered them to be branches. She cut
serial sections through the plumules of A. asparagoides and found what she believed
to be the axillary shoot of which the assimilatory organ was the prophyll. In Asparagus
she, and later Kaussmann, found what they took to be reduced leaves borne on the
organs, which they therefore regarded as cladodes.
While Kaussmann agreed with Arber about the nature of the organs of Asparagus
sensu Arber, and supported the view with developmental investigations, he disagreed
with her about the nature of the organs of A. asparagoides. His ontogenetic work
suggested that these were also cladodes, and not true leaves.
Schlittler took the opposite view to that of Kaussmann and, basing his evidence
on a criticism of other workers’ conclusions, and by comparing the assimilatory organs
of what he believed to be related plants, arrived at the conclusion that they were leaves
in all species.
While no satisfactory conclusion to this controversy has yet been reached, it has
been decided that the term cladode will be used throughout this revision. The present
author feels that the weight of evidence is on the side of this view, and believes this to
be the more generally accepted point of view.
The cladodes may be single, 3-nate or variable. These three characters are
generally reliable. In many species the cladodes are flattened. In most, however,
they are terete or angled, but several species are variable in this character, such as
A. juniperoides. All species have a single vein except the section Myrsiphyllum and
A. striatus. The cladodes are sometimes scabrid, but this is rarely a reliable feature.
(i) Leaves
Leaves are always reduced. Sometimes they are firm, as in A. exuvialis, but usually
with a more or less free, papery and scarious part. (Figs. 2c & d).
(j) Spines
Spines are formed from branches in the section Capenses, one in A. glaucus, three
or more in the others. (Fig. 2b). In all other species where spines are formed they are
formed from reduced leaves. Many species, which usually have spines, have individuals
without them. The spines may be straight or recurved, spreading or reflexed, but
these characters are generally rather variable. In species with spines, the presence
or absence of spines on the ultimate branches is often significant.
36
F.c 2— Diagnostic characters: a, spines formed from modified branches showing aborted nodes;
F,°' 2' “ Sded spines formed ftom modified branches lacking .aborted nodes: “f no!
Hifferentiated and fascicled; c, spine formed from a modified leaf with the non spino
portion chartaceous; d, spines formed from a modified leaf with the non-spinous por 10
hard, deltoid and entire; outer layers of the stem peeling in strips.
37
Subgeneric Groups
Other than Kunth (1850), the only botanists before Baker (1875) who subdivided
the genus, did so on the presence or absence of spines — in many species one of the
most variable characters. Kunth recognised three genera: Asparagus , Asparagopsis
and Myrsiphyllum. Myrsiphyllum included the herbaceous species with flattened
cladodes. Asparagus was separated from Asparagopsis in, supposedly, being dioecious.
Although in A. officinalis L. this is undoubtedly the case, for most of the species he
placed in Asparagus his information was faulty. The species within each genus
recognised by Kunth do not show a particularly large number of characters in common
while, in several cases, undoubtedly similar species were placed in different genera.
As none of the South African species which Kunth placed in his construction of
Asparagus is dioecious, his use of this character to separate Asparagus from Asparagopsis
was incorrectly applied to the South African species. His genus Myrsiphyllum was
retained by Baker (1875 and 1896) as a subgeneric group, and will be treated as a
section in this work.
Baker (1896) included all species in one genus, which he divided into nine sections.
The characters he used for separating these sections were: the presence or absence
of spines; the shape of the cladodes, and the number of these in a cluster; the number
of flowers in a cluster; whether or not the branchlets were fascicled; and the type of
inflorescence.
In each of his sections there are closely similar species, but also many with little
in common. His section Myrsiphyllum has been retained here with a slight change,
and the section Capenses includes all the species he included. While it would be unwise
to propose a revision of the sections of Asparagus with only a knowledge of the genus
in Southern Africa, the following sections appear to reflect the natural grouping of
the species in South Africa. In each, the more important characters are printed in
italics. The species which would be included in each are indicated by their numbers
in this work.
1. Capenses
Roots not tuberous. Aerial stems perennial, woody. Branchlets fascicled.
Cladodes fascicled, terete. Spines axial in origin. Peduncles usually terminal. Anther
filaments not spurred. Berries globose, fleshy or dry.
This section contains the same plants as Baker's section Capenses.
Species: 1-4.
2. African i
Roots not tuberous. Aerial stems usually perennial. Branchlets, when present,
usually solitary. Cladodes fascicled, terete. Spines frequent, foliar in origin. Peduncles
axillary or terminal. Anther filaments not spurred except in A. fasciculatus. Berries
globose, fleshy.
This section has no outstanding characters, but is identifiable by a combination
of root, spine, cladode and inflorescence characters. It contains species from Baker’s
sections Declinati, Umbellati, Thunbergi and Africani.
Species: 5-16.
3. Exuviali
Roots tuberous. Aerial stems perennial , woody. Cladodes fascicled, terete. Spines
foliar in origin, usually small and blunt. Peduncles axillary. Anther filaments not
spurred. Fruit dry and smooth.
A monotypic section containing A. exuvialis, which Baker included in his section
Declinati.
Species: 17.
38
4. Race mo si
Roots tuberous. Aerial stems woody or softly woody. Branchlets often present.
Cladodes fascicled or occasionally solitary, flattened or subulate, with one vein. Spines
often present, foliar in origin. Flowers generally borne in racemes. Anther filaments
not spurred. Fruit a berry.
This section includes Baker’s sections Racemosi, Falcati and A. juniperoides from
his section Myrsiphyllum.
Species: 18-33.
5. Striati
Roots tuberous. Aerial stems perennial. Branchlets absent. Cladodes sometimes
solitary, flattened to terete, firm, with one or several veins. Spines absent or poorly
developed. Peduncles terminal. Anther filaments not spurred.
This section contains part of Baker's sections Striati and Umbellati.
Species: 34 and 35.
6. Sc an dentes
Tubers present. Aerial stems somewhat herbaceous, perennial. Branchlets absent.
Cladodes ternate, flattened, with one vein. Spines absent. Peduncles axillary. Anther
filaments not spurred. Fruit globose, fleshy, with one seed.
This section contains two species which Baker included in his section Striati.
Species: 36 and 37.
7. Crisp i
Tubers present. Aerial stems annual. Branchlets absent. Cladodes ternate,
terete or flattened. Spines absent. Peduncles axillary. Anther filaments spurred.
Fruit ovoid, fleshy, many-seeded.
This section is monotypic, containing A. crispus, which Baker included in his
section Declinati.
Species: 38.
8. Myrsiphyllum
Tubers present. Aerial stems annual. Cladodes flattened, with more than one
vein, solitary. Spines absent. Peduncles axillary. Anther filaments spurred. Fruit
fleshy, globose.
This section is the same as Baker’s, except that A. juniper oides has been removed.
Species: 39 and 40.
Cytological Investigations
While a thorough investigation of chromosome numbers could not be undertaken,
a technique has been used which has provided counts for ten plants.
It was found that young shoots were as satisfactory as either germinating seed
radicles or roots, and that they are more readily obtainable. Pollen Mother Cells
are extremely difficult to get at the right stage as they develop when the buds cannot
easily be recognised.
The following procedure was developed:
39
Shoot tips were placed in 0-01 per cent aqueous colchicine solution for three to
four hours, before being fixed in Acetic Alcohol (1:3) for twenty to twenty four hours.
They were then hydrolised in 20 per cent HC1 at 63 degrees centigrade for three to
eight minutes, depending on the thickness of the material. Two changes of distilled
water for at least five minutes each were used before the shoot tips were squashed
in 2 per cent Aceto-Orcein.
The following numbers were obtained: —
Species No.
A . capensis 40
A . stipulaceus 80 ?
A. fasciculatus 20
A . thunbergianus 20
A . retrofr actus 20
A. densiflorus (sprengeri form) 60
A . krebsianus 20
A. asparagoides (broad cladode form) 20
A. asparagoides (narrow cladode form) 20
A. asparagoides (pubescent form) 20
ASPARAGUS
Asparagus L. , Sp. PI. 313 (1753); Thunb., Prodr. 65 (1794): FI. Cap. ed. Schult. 332
(1823); Bresler, Diss. (1823); R. & S., Syst. Veg. 7: 317 (1829); Kunth, Enum. PI.
5: 57 (1850); Bak. in Journ. Linn. Soc. Bot. 14: 594 (1875); Benth. & Hook, f.,
Gen. PI. 3: 765 (1883); Bak. in FI. Cap. 6: 256(1896); FI. Trop. Afr. 7: 425 (1898);
Marloth, FI. S. Afr. 4: 83 (1915); Krause in Pflanzenfam. Auf. 2, 15a: 362 (1930);
Salter in FI. Cape Penins. 172 (1950); Hutch., Fam. Flow. PI. 2: 608 (1959); Solch,
Beitr. FI. Siidwest-Afr. 31 (1961). Type species: A. officinalis L.
Medeola L., Sp. PI. 339 (1753), partly, as to M. asparagoides L.
Myrsiphyllum Willd. in Ges. Naturf. Fr. Berl. Mag. 2: 25 (1808); Kunth, Enum.
PI. 5: 105 (1850). Type species: M. asparagoides (L.) Willd.
Asparagopsis Kunth in Abh. Akad. Berl. 35 (1842); Enum. PI. 5: 76 (1850). Type
species: A. albus (L.) Kunth.
Hecatris Salisb., Gen. PI. Fragm. 66 (1866). Type species: H. asparagoides (L.)
Salisb.
Roots often tuberous, but frequently fibrous. Stems erect or climbing, annual
or more often perennial and woody, often much-branched; final branches sometimes
fascicled and morphologically distinct from the main branches. Cladodes solitary or
fascicled, with one or more veins. Leaves greatly reduced. Flowers borne singly on
the final flowering stalks, which may be arranged terminally, in the axils of cladodes,
or on specialised branches lacking cladodes; the final flowering stalk always articulated.
Perianth segments six, free or nearly so, more or less similar, ascending or spreading,
up to 6 mm long, white or nearly so. Stamens six, fused to the perianth segments
towards their base. Ovary of three fused carpels containing three chambers; ovules
several, axile; style with three short branches or divided nearly to the base. Berry
1 — several-seeded, globose, or rarely ovoid, usually red, less often black or brown.
Distributed throughout Africa and in most of Europe, Asia and Australia. 40
species recognised in Southern Africa.
40
1 Cladodes conspicuously flattened:
2 Cladodes more than one-veined:
3 Flowers axillary:
4 Stem smooth, usually climbing; tubers sessile or distant 39. A. asparagoides
4a Stem seabird and often with longitudinal wings, erect; tubers distant ... . 40. A. undulatus
3a Flowers terminal 34. A. striatus
2a Cladodes one-veined:
5 Plants herbaceous, twining; flowers axillary or terminal:
6 Cladodes all in one plane 36. A. scandens
6a Cladodes not all in one plane 37. A. ramosissimus
5a Plants woody or, if rather softly woody, then erect; flowers usually racemose:
7 Cladodes solitary:
8 Stems pubescent 18. A. wcyacanthus
8a Stems glabrous:
9 Spur of the leaves spine-shaped; tepals and cladodes without a ciliate margin;
stems usually well lignified; from the southern Cape and the eastern parts
of South Africa 22. A. densiflorus
9a Spur of the leaves not spine-shaped; tepals and usually the cladodes with a ciliate
margin; stems softly woody ; from the western Cape and S.W. A. 19. A. juniperoides
7a Cladodes several at a node:
10 Stems erect 22. A. densiflorus
10a Stems climbing:
1 1 Cladodes more than 35 mm long 26. A. falcatus var. falcatus
11a Cladodes less than 35 mm long:
12 Racemes branched; cladodes usually very dark after drying 26. A. falcatus var.
ternifolius
12a Racemes simple; cladodes not usually darkening on drying:
13 Branches grooved; from the southern and south-west Cape 25. A. aethiopicus
var. aethiopicus
13a Branches smooth; from the eastern Cape and the eastern and northern parts
of South Africa 25. A. aethiopicus var. angusticladus
la Cladodes terete, angled, or absent during the flowering period:
14 Flowers in racemes:
15 Cladodes solitary:
16 Spur of the leaves spine-shaped; tepals and cladodes without a ciliate margin; stems
usually well-lignified; from the southern Cape and the eastern parts of South
Africa 22. A. densiflorus
1 6a Spur of the leaves not spine-shaped ; tepals and usually the cladodes with a ciliate margin ;
stems softly woody; from the southern Cape and S.W. A 19. A. juniperoides
15a Cladodes several at a node, or absent during the flowering period:
17 Cladodes terete or oval in section, 0-8 mm or more thick 32. A. crassicladus
17a Cladodes angled or, if terete, then less than 0-8 mm thick, or absent during the flowering
period:
18 Branches, at least the younger, densely pubescent, scabrid or tubercled; peduncles
grey to brown; flowers white:
19 Branches ridged or grooved; cladodes often absent during the flowering period
27. A. asperigillus
31. A. acock sii
19a Branches smooth
41
18a Branches glabrous or, if slightly pubescent, then the stems erect and the flowers
yellowish or greenish yellow:
20 Stems erect :
21 Final branches grooved 22. A. densiflorus
21a Final branches smooth:
22 Flowers white; from the eastern parts of South Africa.... 23. A. saundersiae
22a Flowers yellowish or greenish yellow; mainly from the western and northern
Cape, Bechuanaland and S.W.A., but recorded from the northern Trans-
vaal 33. A. nelsii
20a Stems twining:
23 Spines present on all branches; branchlets never well differentiated 30. A. krebsianus
23a Spines absent from ultimate branches; branchlets sometimes well differentiated:
24 Stems usually grooved; branchlets differentiated; cladodes more than \ mm
thick, angled; perianth segments 2-5-3 -0 mm long 28. A. racemosus
24a Stems smooth: branchlets not generally differentiated; cladodes very narrow,
less than \ mm thick; perianth segments usually 1 -5-2-0 mm long
24. A. buchananii
14a Flowers not in racemes:
25 Branches spine-tipped; spines formed from modified branches, usually at least 3-nate:
26 Spines solitary; perianth enclosing the fruit 4. A. glaucus
26a Spines rarely solitary; perianth not enclosing the fruit:
27 Flowers sessile; spines completely smooth:
28 Average cladode length less than 5-5 mm; plant widely branched:
29 Cladodes at least 1-5 mm long; branches and stems generally fairly slender;
from the southern and western Cape, and S.W.A. away from the coast
1. A. capensis var. capensis
29a Cladodes less than 1 mm long; branches and stems more robust; from the west
coast 1 . A. capensis var. lit oralis
28a Average cladode length more than 5-5 mm; plants of a “fox-tail” form
2. A. stipulaceus
27a Flowers pedunculate; spines bearing nodes, which are usually reduced to minute
protuberances (Fig. 2a) 3. A. suaveolens
25a Branches not spine-tipped; spines formed from modified leaves, solitary, or absent:
30 Stem rigid; flowers mainly terminally 1-3-nate on peduncles, but sometimes also laterally
solitary on the peduncles:
31 Cladodes more than 0-4 mm broad; spines 4 mm or less long; stems glabrous
21. A. rigidus
31a Cladodes less than 0-3 nm broad; the longest spines at least 5 nm long; stems
usually more or less scabrid or pubescent 20. A. minutiflorus
30a Stems variable; flowers axillary or terminal on normal branches:
32 Cladodes three at a node except at nodes where there is a branch and then only two:
33 Cladodes 1 mm or more broad; fruit globose.... 37. A. ramosissimus
33a Cladodes less than 1 mm broad; fruit ovoid 38. A. crispus
32a Cladodes solitary, paired or numerous at a node, or absent during the flowering
period:
34 Flowers all terminal:
35 Stems smooth, climbing, repeatedly branched; cladodes and ultimate branches
all in one plane 6. A. setaceus
35a Stems grooved, erect or climbing to 2 m or rarely more; final branches and
cladodes not all in one plane.
36 Branches usually reflexed; cladodes laterally usually several; plant more or
less herbaceous, and often climbing or scrambling 35. A. subulatus
42
36a Branches usually ascending; cladodes laterally usually solitary; plant more
woody and robust, always erect 34. A. striatus
34a Flowers all, or at least some, axillary:
37 Cladodes solitary 19. A. juniperoides
37a Cladodes not solitary:
38 Young branches never both white and grooved:
39 Usually climbers or scramblers; stems tortuous:
40 Aerial stems annual, softly woody; peduncles articulated at the base of
the flower 9. A. fascicidatus
40a Aerial stems perennial, rarely annual; peduncle not articulated at the
base of the flowers:
41 Stems smooth or regularly grooved, sometimes lacking cladodes towards
the ends of the branches; spines often well-developed; colour of
stems variable 5. A. africanus
41a Stems irregularly, longitudinally wrinkled; cladodes always evenly
distributed along the entire length of the branches, or completely
absent; spines absent or poorly developed; stems green to greyish
green 7. A. denudatus
39a Stems erect, straight or zigzagging with straight internodes:
42 Outer layers of stems and branches peeling; leaf firm and entire (Fig. 2d)
17. A. exuvialis.
42a Outer layers of stems and branches not peeling; leaf usually with a char-
taceous free portion (Fig. 2c):
43 Plant completely glabrous:
44 Spines absent or poorly developed:
45 Branches green :
46 Stem zigzagging: branches more or less spreading. . 7 . A. denudatus
46a Stem straight; branches virgate 8. A. virgatus
45a Branches not green:
47 Plant up to 1 m high; from the southern Cape 12. A. macowanii
var. macowanii
47a Plant over 1^ nr high: widespread, but not in the southern Cape
12. A. macowanii var. zuluensis
44a Spines present and pungent, at least on the stems and main branches
10. A. thunbergianus
43a Ultimate branches pubescent 11. A. mucronatus
38a Young branches and often all the aerial stems and branches white and grooved:
48 Stems bearing fascicled, simple branches or less often secondary branches
present 29. A. obermeyerae
48a Axes repeatedly branched; branches solitary or less often 2- or 3-nate:
49 Spines present on all ultimate branches:
50 Cladodes ascending, straight; plant over 1 nr high, bushy 15. A. laricinus
50a Cladodes spreading, often arcuate; stems more or less erect, less than
1 m high 14. A. compactus
49a Spines absent from ultimate branches:
51 Cladodes 4 mm or less long 16. A. microrhaphis
51a Cladodes 5 mm or more long:
52 Branches zigzagging; stems often scrambling or climbing; from the
western half of South Africa 13. A. retrofractus
52a Branches straight; stems e ect; from the wetter parts of the southern
Cape and the eastern half of South Africa:
53 Stems up to 1 m high; from the Cape 12. A. macowanii var. macowanii
53a Stems over 1J2 m high; from Natal and the Transvaal 12. A. macowanii
var. ziduensis
43
1. A. capensis L., Sp. PL 314 (1753). Type: Cape of Good Hope, without precise
locality, collector unknown, in Herb. Plukenet (BM, holo.; PRE, photo.!).
Roots fibrous. Stems erect, usually with several wide-spreading branches, up to
1 m high, brown or green, zigzagging, terete, pubescent when young, terminating in
a spine. Branches similar to the stems. Branchlets fascicled, not replacing spines,
spineless except occasionally for a terminal spine, pubescent. Lateral spines formed
from modified branches, usually 3-nate, rarely more numerous, smooth, spreading at
right angles to one another, straight or recurved, the central one longest, brown or
green with a brown tip; some of the central spines on the older parts of the plant
replaced by normal branches. Cladodes terete, usually slightly recurved, 1-5-6 mm
or more long, glabrous or very minutely pubescent. Leaves scarious, with marginal
hairs, sometimes very greatly reduced. Flowers sessile, 1- or 2-nate, terminal. Perianth
segments oblong-obovate, 2-4-5 mm long, white with a green or purple median
streak; margin serrate. Stamens nearly as long as the perianth segments; anthers
about 0-5 mm long, yellow. Ovary about 1 mm long; style about 1 mm or less long.
Berry globose, 4 mm diameter, 1-seeded, red; perianth persistent.
South-west Cape, northwards to the southern parts of South West Africa, and
eastwards to Port Elizabeth.
For key to varieties, see key to species.
(a) var. capensis.
A. capensis L., Sp. PI. 314 (1753); Jacq., Hort. Schoen. t.266 (1798); Thunb., Prodr.
66 (1794); Bresler, Diss. 21 (1826); Bak. in Journ. Linn. Soc. Bot. 14; 617 (1875);
FI. Cap. 6: 263 (1896); Salter in FI, Cape Penins. 173 (1950); Solch, Beitr. FI. Siidwest-
Afr. 36 (1961). A. triacanthus Burm. f., Prod. FI. Cap. 10 (1768), nom. illegit. Type:
as for A. capensis. A. nelsonii Bak. in Journ. Linn. Soc. Bot. 14: 617 (1875); FI.
Cap. 6: 264 (1896). Type: South Africa, Nelson s.n. (BM, holo.; PRE, photo.!).
Asparagopsis passerinoides Kunth, Enum. PI. 5: 90 (1850), nom. illegit. Type: as
for A. capensis.
The typical variety is characterised by the generally smaller plants, narrower
stems and branches, and by the larger cladoles, which are always linear and more than
1 mm long.
This variety occurs over the entire range of the species except the coast of Nama-
kwaland and South West Africa. It occurs from sea-level to about 1,000 feet, especially
in sandy soils. Among shore vegetation it may form dense bushes about 1 m high,
with thick branches and short, thicker cladodes, while at higher altitudes it is generally
more sparsely branched and smaller, with narrower branches and cladodes. It is one
of the commoner species of Asparagus in the winter rainfall area. Flowering is affected
by, but not dependant on, fires. Flowers have been recorded in February and from
April to August. The flowers are sweetly scented.
Cape. — Bellville: Melkbosch, Wasserfall 159 (NBG). Bredasdorp: near Elim, Schlechter 10476
(BOL). Caledon: Grietjesgat, between Sir Lowrys Pass and Palmiet River, Ecklon & Zeyher s.n.
Calvinia: Calvinia. Compton 19484 (NBG). Cape: Lion’s Head, MacOwan 1977 (SAM). Carnarvon:
10 miles N.E. of Carnarvon, Hafstrom & Acocks 194. Clanwilliam: Langekloof, Schlechter 8029
(BOL). Hay: 16 miles E. of Griquatown, Acocks 2191. Hopefield: Hopefield, Jessop 455 (BOL).
Knysna: Keurboom’s River at Postpad, Fourcade 207 (BOL). Laingsburg: Ngaapkop, Compton
14591 (NBG). Malmesbury: Malmesbury, Marloth s.n. (STE). Middelburg: Witkransnek, Acocks
21415. Montagu: Fonteinkloof, Lewis s.n. (NBG). Murraysburg: Murraysburg, Tyson 38 (BOL).
Namakwaland: Garies, Thorne s.n. (SAM). Paarl: Tigerberg, Pillans 8656 (BOL). Piketberg: The
Rest, Barker 3590 (NBG). Port Elizabeth: Red House, Paterson 1109 (BOL). Somerset West:
Strand, Parker 3805 (BOL). Stellenbosch: Stellenbosch, Bond s.n. (NBG). Steynsburg: Steynsburg,
Thode 5636 (STE). Tulbagh: Saron, Schlechter 7886. Uitenhage: Uitenhage, Cooper 1573 (BOL).
Uniondale: Joubertina, Horn s.n. Vanrhynsdorp: near Vanrhynsdorp, Marloth 7794. Victoria
West: above Victoria West reservoir, Smith 2404. Worcester: Nature Reserve, Olivier 8. Wynberg:
Rondebosch, Wolley Dod 1143 (BOL).
44
South West Africa. — Luderitz: Zebrafontein, Merxmuller & Giess 3325 (M); 6 miles N. of Witputz
Police Station, Merxmuller & Giess 3446 (M).
Linnaeus, in his Species Plantarum, wrote: “Asparagus aculeatus, triplici spina,
surrectus. Pink. aim. 54. t. 15./. 4 But this figure does not show a plant with ternate
spines as described. On the page of text, page 54, to which Linnaeus referred, Plukenet
wrote: “Asparagus aculeatus, triplici spina surrectus, Prom. Bon. Spei. Phytogr. Tab.
78. fig. 3 While the polynomial is the same as that used by Linnaeus, the figure
is therefore different. This figure does show a plant with ternate spines. It would,
therefore, s;cm likely that Tab. 78, fig. 3 should accompany Linnaeus’ citation of
“ Pluk. aim. 54 ”, and the reference to Tab. 15, fig. 4 was an error. As the specimens,
labelled A. capensis , in the Linnaean Herbarium do not fit Linnaeus’ description,
they cannot be regarded as types. Both the Plukenet figure and the specimen on which
it was based in the Plukenet Herbarium match the polynomial in having ternate spines.
The specimen is therefore regarded as the holotype.
The binomial Asparagus triacanthus was accompanied, in Burmann’s work,
only by a reference to Plukenet’s Phytographia t. 78, f. 3, which is the figure regarded
by the present author as the illustration on which Linnaeus based his A. capensis.
These names are therefore synonymous.
Asparagopsis passerinoides Kunth is also a superfluous name for A. capensis of
Linnaeus. Kunth cited both Linnaeus’ description and the correct Plukenet figure.
Occasionally an atypical form occurs under trees. This form has longer
cladodes — up to 8 mm long — and a more lax habit. Also, it often has more numerous
spines at a node than is usual in this species. A specimen of this form, collected by
Nelson and now in the British Museum, was named A. nelsonii by Baker. He separated
A. nelsonii from A. capensis on the grounds that its branches were terminated by a
spine. There are few, if any, specimens of A. capensis in which the branches are not
terminated by a spine, however. No difference was found between these forms after
they had been cultivated near one another for two years. Nor is it possible to separate
these forms in the field because of specimens showing intermediate characters.
(b) var. litoralis Suess. & J. Karl, in Mitt. Bot. Staatssam. Miinchen 2: 50 (1950).
Type: South West Africa, Pomona, Dinter 6367 (B, holo.!; BOL!, PRE!, SAM!, Z!).
This variety is characterised by the larger plant, more robust stems and branches,
and small size of the cladodes. The cladodes are so reduced as to give a granular
appearance to the branchlets.
Retricted to the coast from Lambert's Bay in the Clanwilliam district to Luderitz
Bay in South West Africa. Flowers have been recorded in May.
Cape. — Clanwilliam: Lambert's Bay, Muir s.n. (SAM); 5 miles N. of Lambert's Bay, Taylor 3951;
Elandsbaai, Strey 427 (M). Namakwaland: south of Hondeklip Bay, Pi/lans s.n. (BOL).
South West Africa. — Luderitz: Luderitz Bay, Peyer 5 (Z); Pomona, Dinter 6367; Oranjemund,
Merxmuller 2335; Sturmvogelbucht, Merxmuller 2241; Buchubergen, Dinter 6463a (B).
There are specimens which show several intermediate characters between the two
varieties of this species. These occur in littoral vegetation but, although possessing
the more robust form of the second variety, do not have the extremely reduced cladodes.
A specimen of this form was grown near a specimen of the typical form of variety
capensis for about five years. During this time the differences were maintained. This
may indicate that a genetic difference is involved, but this is not of sufficient magnitude
either to create a third variety or to raise variety litoralis to specific status.
2. A. stipulaceus Lam., Encyc. 1: 297 (1783); Bresler, Diss. 5 (1826); R. & S.,
Syst. Veg. 7: 333 (1829); Salter in FI. Cape Penins. 173 (1950). Type: Cap e, Sonnerat
s.n. (P, holo.; PRE, photo.!).
45
Asparagopsis stipulacea (Lam.) Kunth, Enum. PI. 5: 91 (1850).
Asparagus densus Soland. ex Bak. in Journ. Linn. Soc. Bot. 14: 616 (1875). Type:
Cape, without precise locality, Auge s.n. (BM, holo.; PRE, photo.!).
Roots fibrous. Stems erect to 50 cm high, rarely more, not usually producing
branches, pubescent, grey-green to brown, ending in spines, angled or terete, ridged
when young, slightly zigzagging. Branches short, few or absent, not forming secondary
branches, spine-tipped. Lateral spines ternate, the longest occasionally replaced by
a branch, spreading at right angles to one another, straight or slightly curved, green
with a brown-red tip, up to 2 cm long. Branchlets fascicled, spineless or sometimes
terminating in a spine, grooved. Cladodes firm, terete or angled, slightly mucronate,
sometimes scabrid, 5-10 mm long, almost always 3-nate. Leaf scarious, fringed with
hairs. Peduncles represented by a minute disc, terminal, 1- or 2-nate. Perianth segments
oblong-obovate, white with a green or purple median streak, 5-6 mm long, spreading,
the margin serrate. Stamens almost as long as the perianth; anthers 0-7-0 -8 mm long,
yellow-green. Ovary 1-1-5 mm long; style 1 mm long. Berry fleshy, globose, red,
1- or 2-seeded.
This species grows in sandy soils, often in vegetation with a high proportion of
Restionaceae, and usually within sight of the sea, from the west coast of the Cape
Peninsula to the eastern parts of the Bredasdorp district. It is particularly affected
by fires and rarely flowers except after fires. Flowers have been recorded in April,
May and July. Fruits are rarely produced. Even in areas where large numbers of
plants have flowered, it is usually impossible to find fruits. This is possibly associated
with polypoidy.
Cape. — Bredasdorp: Mierkraal, Schlechter 10536; De Hoop, Barker 8731 (NBG); Baardskeerders
bos, Oliver s.n. (BOL); near Gansbaai, Oliver s.n. (BOL). Caledon: Rooi Els, Jessop s.n. (BOL);
Betty’s Bay, Oliver s.n. (BOL). Wynberg: Karbonkelberg, Oliver s.n. (BOL). Simonstown: Vasco
da Gama, Salter 8046A (BOL); .Jessop 135 (BOL); Buffels Bay, Guthrie s.n. (BOL); Oliver s.n. (BOL);
Miller’s Point, Pillans 8532 (BOL); Clovelly, Sailer 8210 (BOL); Kalk Bay, Jessop 47 (BOL); Oliver
s.n. (BOL); St. James, Michell s.n. (BOL); hills near Fish Hoek, Peers s.n.
The type specimen was collected by Sonnerat at the Cape. Two specimens have
been mounted on the same sheet. That on the left is of an Asparagus, and matches
Lamarck’s description. The other is of a Psoralea. The specimen on the left has,
therefore, been selected as the type.
A. capensis and A. stipulaceus are very similar, and their separation is sometimes
difficult. A. capensis is larger and considerably more branched, while A. stipulaceus
is usually more pubescent and has larger claodes and flowers. At Rooi Els where the
two species grow together on a slope above the river, they retain their distinguishing
characters.
Chromosome counts were made on specimens of A. capensis from about four
miles inland from Ysterfontein, in the Malmesbury district, and of A. stipulaceus from
about half way up the mountain behind Kalk Bay in the Simonstown district. The
diploid number was found to be 40 for the former species and very near to — probably
exactly — 80 for the latter. This difference in chromosome number suggests that the
small qualitative differences between the two species is not only the result of inadequate
sampling of the populations of each form.
3. A. suaveolens Burch., Trav. 2: 226 (1822); Pole Evans in Flow. PI. S. Afr.
11: t. 409 (1931). Type: Griquatown, Burchell 1956 (K, holo.!; PRE, photo.!).
A. spinescens Steud. ex R. & S., Syst. Veg. 7: 334 (1829). Type: Cape, without
precise locality, Ludwig s.n. A. triacanthus Willd. ex R. & S., Syst. Veg. 7: 334 (1829).
Type: Cape, Lichtenstein in Herb. Willd. no. 6693. A. burchellii Bak. in Journ. Linn.
Soc. Bot. 14: 618 (1875); FI. Cap. 6: 264 (1896). Syntypes: “Cap. Bon. Spei in
46
aridis ”, Burchell 2962 (K ! ) ; Zeyher ( Asparagus 10); Cooper 1574. A. stipulaceus
sensu Bak. in Journ. Linn. Soc. Bot. 14: 617 (1875); FI. Cap. 6: 264 (1896). A.
spinosissimus Kuntze, Rev. Gen. PI. 3: 315 (1893); Solch, Beitr. FI. Siidwest-Afr. 39
(1961). Type: Cathcart, Kuntze s.n. (NY, holo.; BOL, photo.!). A. omahekensis
Krause in Bot. Jahrb. 51 : 447 (1914); Solch, Beitr. FI. Siidwest-Afr. 38 (1961). Type:
“ Zwischen Gobabis und Oas ”, Dinter 2711 (B, holo.!; SAM !). A. intangibilis Dinter
in Fedde, Rep. 29: 269 (1931); Solch, Beitr. FI. SUdwest-Afr. 37 (1961). Type:
“ Grosse Karasberge ”, Dinter 5168 (B, holo.!).
Asparcgopsis triacantha (Willd.) Kunth, Enum. PI. 5: 91 (1850). A. zeyheri Kunth,
Enum. PI. 5; 92 (1850). Type: Without locality, Zeyher s.n. A. spinescens (R. & S.)
Kunth, Enum. PI. 5: 93 (1850).
Roots fibrous. Stems erect, to 1 m, or occasionally climbing or scrambling to
0- 5 m, zigzagging, spine-tipped, glabrous or pubescent, pale cream coloured to brown.
Branches similar to the stems, usually spreading or ascending, but occasionally reflexed.
Lateral spines 3-, 5- or 7-nate, at least the lateral ones of each group usually bearing
sterile nodes (Fig. 2a), pungent, well-developed, straight or arcuate, spreading or
reflexed. Branchlets fascicled, paired to numerous, herbaceous, arcuate or straight,
up to 3 cm long. Cladocles borne on branchlets, fascicled, subulate, spreading or
ascending, 3-6-nate, 1-10 mm or more long. Peduncles solitary, terminal on branchlets,
1- 4 mm long, articulated near the base. Perianth segments obovate, 1-5-3 -5 mm
long, the outer whorl frequently, perhaps always, with a purple streak. Stamens nearly
as long as the perianth segments, narrow; anthers about 1/3 mm long, orange to yellow.
Style divided for less than half its length; ovary sessile, 12-14-ovuled. Berry one-seeded,
globose, 3—3 - 5 mm diameter, red or black; perianth persistent.
This is one of the commoner species of Asparagus in the summer rainfall area,
occurring in Rhodesia, large parts of South West Africa, Bechuanaland and most of
the areas to the south, and extending into the winter rainfall area as far as the Clan-
william district in the north and Caledon in the east. It has not been recorded from
Basutoland and there are few records from the south-eastern parts of South Africa.
It occurs in coastal vegetation, low bush, open forest and Karroid vegetation. Flowers
have been recorded in most months, but particularly in autumn and early summer.
There does not seem to be any geographical variation in flowering period.
Cape. — Albany: Botha's Hill, Rogers 3417 (BOL). Alexandria: Debega Valley, Galpin s.n. Aliwal
North: Doctor’s Drift, Gerstner 245. Barkly West: near the Vaal River, Nelson 1 76. Bathurst: Port
Alfred, Salisbury s.n. (GRA). Bedford: Witmos, Galpin 3081. Bredasdorp: Mierkraal, Schlechter
10527. Caledon: Caledon, Esterhuysen s.n. (BOL). Cathcart: 1 mile from Henderson on Cathcart
road. Story 2805. Clanwilliam: Langkloof, Schlechter 8050. Cradock: 10 miles E. of Cradock,
Jessop 656. De Aar: De Aar, Acocks 632. East London: East London, Thocle 6653 (STE). Fraser-
burg: Fraserburg, Net s.n. (STE). Gordonia: Gordonia district, Kotze 852. Graaff-Reinet : Graaff-
Reinet, Bolus 411 (BOL). Hankey: Hankey, Barker 7853 (NBG). Hay: Elmadale, Acocks 2231.
Heidelberg: 5 miles W. of Heidelberg, Jessop 591. Kentani: Columbia Mission Station, Pegler 1451
(BOL). Kimberley: 7 miles N.E. of Kimberley, Leistner <& Joynt 2650. King Williams Town: King
Williams Town, Sim 1060 (BOL). Komga: Komga, Flanagan 1793. Kuruman: 10 miles S. of
Kuruman, Louw 842. Ladismith: Prinspoort, Bond 306 (NBG). Laingsburg: Tweedside, Marloth
12076. Middelburg: Grootfontein, Theron 94. Montagu: Cogman’s Kloof, Barker 8625 (NBG).
Murraysburg: Roads Poort, Tyson 212 (GRA). Namakwaland: Doornpoort, Pillans 5466 (BOL).
Pearston: by the Blyde River, Burchell 2962 (K). Phillipstown : Potfontein Station, Schweickerdt
1196. Oudtshoorn: near dam, Britten 1624. Port Elizabeth: Cradock Place, Galpin 6439. Post-
masburg: west entrance to Langkloof, Leistner 1164. Prieska: near Prieska, Bryant J368. Prince
Albert: Farm Damascus, Marloth 1 1276. Queenstown: Long Hill, Galpin 2095. Riversdale: Lange-
berg, Muir 2529 (BOL). Robertson: Klaasvoogds, Van Niekerk 372 (BOL). Somerset East : Somerset
East, Van Niekerk 435 (BOL). Swellendam: National Bontebok Park, Liebenberg 6536. Taungs:
Kankatjies, Acocks 538. Uitenhage: near Uitenhage, Schlechter 2573. Uniondale: 8 miles from
Uniondale on Willowmore road, Fourcade 6256. Vryburg: Tiger Kloof, Brueckner 487. Worcester:
Veld Reserve, Van Breda 21.
Natal. — Camperdown: Nagle Dam, Wells 1361. Ubombo: Ubombo, Gerstner 3400 (NH).
Weenen: Mielietuin, West 1204.
47
Orange Free State. — Bloemfontein: south of Glen, Mostert 326. BoshofF: between Boshoff
and Windsorton Road, Burtt Davy s.n. Fauresmith: Fauresmith commonage, Leistner 1099. Kroon-
stad: Kroonstad, Pont 196. Parys: Parys, Rogers 5175. Trompsburg: Lucerne, Henrici s.n. Winburg:
Willem Pretorius Game Reserve, Bakker 1004.
Swaziland.- — Hlatikulu, Ingwavuma Poort, Ben Dlamini s.n.
Trasvaal. — Barberton: Barberton, Thorncroft 844 (NH). Brits: Jacksonstuin, Van Vuuren 292.
Groblersdal: 4 miles S. of Bethal Mission Station, Monnig 48. Fleidelberg: The Kloof, Thode 3435
(STE). Johannesburg: Milner Park, Moss 17583. Lichtenburg: Grasfontein, Sutton 461. Lydenburg:
Farm Parys, Barnard & Mogg 702. Marico: Zeerust, Thode 1508. Messina: Messina, Rogers 1-9297
(BOL). Middelburg: Tautesberg, Young A 163. Pietersburg: Reebok Vlei, Mauve 4295. Potchef-
stroom: Losberg, Theron 693 (NF1). Potgietersrus: Mosdene, Galpin 342. Pretoria: near Pretoria,
Schlechter 3584. Roodepoort: Spitskop, Archibald 2829 (GRA). Sibasa: 7 miles N.W. of Punda
Milia, Codd & De Winter 5556. Ventersdorp: Goedgedacht, Sutton 612. Vereeniging: Klipriviersberg,
Gilliland s.n. (BOL). Waterberg: Towoomba Pasture Research Station, Codd 6116. Zwartruggens:
Zwartruggens, Story 5638.
South West Africa. — Gobabis: Okasondana, Walter 4145 (M). Grootfontein : Otaviberge,
Rehm s.n. (M). Rehoboth: Floachanas, Dinter 1948 (SAM). Warmbad: Narudas-Kanus, Dinter
5168.
Bechuanaland. — Molepolole, Story 4869.
In the south-western Karroo, Little Karroo and as far east as the Albany district
there are two forms. One of these is erect and has spreading to ascending branches,
while the other, to which Baker gave the name A. burc/iellii, varies from erect to
climbing and usually has reflexed branches and spines. Evidence, based both on
herbarium material and on field observations, has shown that these forms are linked
by many intermediates and cannot be given taxonomic recognition. These intermediates
occur in the eastern regions of the distribution of the two forms but are absent in the
west. In the Transvaal there are also two forms which exist side by side. The larger
form is glabrous, and has black berries, while the smaller is pubescent, has a particularly
neat appearance, smaller cladodes, spreading branches and red berries. It was again
not found possible to separate these two forms when the variability of the plants of
this species from other areas of South Africa is taken into consideration. Two “ fox-
tail ” forms have been recorded. One of these, from the Laingsburg district, is only
about 25 cm high. It has been collected by Rogers (BOL 16322, Tvl. Museum 21 167).
The other, which is often over 1 m high, has been collected several times, for example
Mogg 22407 (PRE), in the Loskop Dam area in the Transvaal. Little-branched forms
occur in most areas where this species is found, and it is not felt that this character
should be regarded as taxonomically significant in this species.
4. A. glaucus Kies in Bothalia 6: 229, fig. 5 (1954); Solch, Beitr. FI. SLidwest-Afr.
36 (1961). Type: Fauresmith, Groenvlei, Kies 340 (PRE, holo.!).
Roots fibrous. Stems erect, up to 50 cm high, smooth, glabrous, woody, slightly
zigzagging. Branches spreading to ascending, solitary; stems and branches white at
first, becoming brown, often spine-tipped, but frequently damaged. Branchlets 1-3 cm
long, solitary or paired, straight, grooved, greenish, borne at the bases of spines or
branches, terminating in a spine. Lateral spines solitary, mostly less than 1 cm long,
occasionally up to 2 cm long, straight or recurved, slightly ascending or spreading,
pungent, bearing sterile nodes and often bearing lateral spines. Cladodes 2-10 mm
long, glaucous, fascicled, frequently 3-nate, terete, frequently up to 0-5 mm thick.
Leaves very reduced; lateral buds present at the bases of branches and spines.
Peduncles axillary, 2-nate, 2 mm long, borne on branchlets or spines, articulated near
the centre. Perianth segments similar, broadly oblong-obovate, spreading, 3 '5-4-0 mm
long, margins smooth, rather tough, green in bud, becoming greenish-white with a
green keel, with a bloom. Stamens about 3 mm long; filaments white, flattened
laterally; anthers oblong, about 1 mm long. Ovary shortly stalked; the ovules about
9 in each chamber; style branches shortly fused at the base. Fruit globose, 4 mm
diameter, containing a single seed, surrounded by a persistent perianth.
48
Found in the dry areas from the Oudtshoorn district in the south-west, northwards
to Keetmashoop and eastwards to Bloemfontein. It appears to have a strong preference
for brakish soils. Flowers have been recorded from October to December.
Cape. — Cradock: Jakkalsfontein, Acocks 17516. Hanover: Hanover, Sim s.n. Kimberley: Mollers
Pan, Acocks & Hafstrom H874; Leeuwpoort, Burtt Davy 9553. Oudtshoorn: Kruispad, Compton
21779 (NBG). Uniondale: bank of the Olifant’s River, Fourcade 6265 (BOL).
Orange Free State. — Bloemfontein: De Brug, Acocks 8464. Fauresmith: Groenvlei, Kies 340;
Heuningberg, Marais 124; Veld Reserve, Henrici 3981. Luckhoff: Rose Marie, Verdoorn 2148;
Henrici 4535. Trompsburg: Ventersvlei, Verdoorn 1153.
South West Africa. — Keetmanshoop : Great Karasberg, Noachabeb-Witmond, Ortendahl 425.
This species has similarities with A. suaveolens, but differs in having solitary spines,
fewer branches at each node, axillary peduncles, and larger, thicker perianth segments
which envelope the fruit.
5. A. africanus Lam., Encyc. 1: 295 (1783); Bresler, Diss. 9 (1826); R. & S.,
Syst. Veg. 7: 331 (1829); Bak. in Journ. Linn. Soc. Bot. 14: 619 (1875); FI. Cap. 6:
265 (1896); Marloth, FI. S. Afr. 4: pi. 20 (1915); Salter in FI. Cape Penins. 175
(1950); Kies in Bothalia 6: 177 (1951). Type: Cape without precise locality, Sonnerat
s.n. (P, holo.; PRE, photo.!).
Asparagus dependens Thunb., Prodr. 66 (1794); Bresler, Diss. 17 (1826); Kunth,
Enum. PI. 5: 74 (1850). Type: Cape of Good Hope, without precise locality, Thunberg
s.n. (UPS, holo.!; BOL!, PRE! photos). A. cooperi Bak. in Gard. Chron. 1: 818
(1874); Journ. Linn. Soc. Bot. 14: 613 (1875); FI. Cap. 6: 263 (1896); Flow. PI.
S. Afr. 13: t.503 (1933). Type: Cape without precise locality, Cooper s.n. (K, holo.!;
PRE, photo.!). A. multiflorus Bak. in Journ. Linn. Soc. Bot. 14: 610 (1875); FI.
Cap. 6: 261 (1896); Kies in Bothalia 6: 177 (1951). Type: “ Somerset division ”,
Bowker s.n. (K, holo.!; PRE, photo.!). A. asiaticus sensu Bak. in Journ. Linn. Soc.
Bot. 14: 618 (1875); FI. Cap. 6: 265 (1896). A. burkei Bak. in Journ. Linn. Soc.
Bot. 14: 607 (1875); FI. Cap. 6: 261 (1896). Type: South Africa, Crocodile River,
Burke s.n. (K, holo. ! ; PRE, photo. !). A. pilosus Bak. in Journ. Linn. Soc. Bot. 14: 610
(1875). Type: Bechuanaland, near Lake Ngami, McCabe 15 (K, holo.!; PRE, photo. !).
A. fleckii Schinz in Bull. Herb. Boiss. 1, 4, app. Ill: 44 (1896); Solch, Beitr. El.
Siidwest-Afr. 37 (1961). Type: S.W.A., Hoachanas, Fleck 901 (Z, holo.!; PRE,
photo.!). A.judtii Schinz in Bull. Herb. Boiss. 1, 4, app. Ill: 44 (1896); Solch, Beitr.
FI. Siidwest-Afr. 37 (1961). Type: S.W.A., Hoachanas, Fleck 901 (Z, holo.!). A.
conglomeratus Bak. in FI. Trop. Afr. 7: 428 (1898); in Vierteljahrsschr. Naturf. Ges.
Zurich 49: 176 (1904). Type: Ngamiland, Kwebe, Lugard 52 (K, holo.!; PRE,
photo.!). A. bechuanicus Bak. in FI. Trop. Afr. 7: 429 (1898). Type: Bechuanaland,
Nakalechwe, Lugard 25 (K, holo.!; PRE, photo.!). A. lugardii Bak. in FI. Trop.
Afr. 7: 431 (1898). Type: Bechuanaland, near Lake Ngami, Lugard 31 (K, holo.!;
PRE, photo.!). A. dinteri Engl. & Krause in Bot. Jahrb. 45: 155 (1910); Solch, Beitr.
FI. Siidwest-Afr. 36 (1961). Type: “ Bei Neitsas ”, Dinter 676 (B!, SAM!). A. patens
Krause in Bot. Jahrb. 51: 448 (1914). Type: S.W.A., Haribes, Engler 6570 (B!).
A. francisci Krause in Bot. Jahrb. 51: 449 (1914). Type: S.W.A., “ Zwischen
Omuramba und Omatako ”, Seiner 689a (B!). A. rivalis Burch, ex Kies in Bothalia
6: 179 (1951); Solch, Beitr. FI. Siidwest-Afr. 38 (1961). Type: Kosiefontein, Burchell
2587 (K, holo.!; PRE, photo.!).
Asparagopsis lamarckii Kunth, Enum. PI. 5: 87 (1850), nom. illegit. Type: as for
A. africanus. A. juniperina Kunth, Enum. PI. 5: 85 (1850). Type: Swartkops River,
Drege 8576. A. schlechtendalii Kunth, Enum. PI. 5: 90 (1850). Type: Olifants River,
Krebs s.n.
Roots fibrous. Stems woody, twining, either climbing to about 3 m or forming
low bushes up to about 1 m high, glabrous to pubescent, smooth or slightly grooved,
slightly zigzagging, grey-brown to greenish. Branches solitary, much-branched, similar
49
to the stems, but with a greater tendency to pubescence and grooving, usually spreading,
but in some forms — particularly the pubescent specimens — strongly reflexed. Cladodes
fascicled, up to 20-nate, but often about 8-nate, straight or slightly arcuate, subulate,
generally 5-10 mm long, ascending to spreading. Spines usually present on at least
the stems and larger branches, spreading to reflexed, usually straight, pale in many
specimens but frequently reddish brown, up to 1 cm or more long, usually with the
rest of the leaf forming well-developed spurs. Peduncles axillary and a few sometimes
terminal, up to 20-nate, but frequently 2- or 3-nate, up to 5-8 mm long, articulated
below the centre. Perianth segments similar, entire, oblong-obovate, 2 -5-3 -5 mm
long, white. Stamens nearly as long as the perianth segments. Style and ovary each
about 1 mm long. Fruit globose, fleshy, red, about 6 mm diameter, with a single seed.
This species usually occurs in bushy or wooded areas, where it climbs, but also
occurs in more open vegetation where it forms bushes or scrambles among rocks.
It grows mainly in the damper parts of South Africa, but also occurs in drier areas
and in South West Africa, Bechuanaland and further north. Flowers have been
recorded from August to May.
Cape. — Albany: Grahamstown, Galpin 284. Alexandria: Kowie Road from Southwell, Britten
2596. Aliwal North: Orange River, Gerstner 160. Barkly West: Vaal River at Sydney, Acocks
H830. Bathurst: Port Alfred, Rogers 16623 (BOL). Bedford: Bedford, Gane 118 (GRA). Bellville:
Tigerberg, Pi/Ians 8659 (BOL). Caledon: Baviaan’s Kloof, Genadendal, Gillett 855 (BOL). Cape:
Porcupine Buttress, Jessop 12 (BOL). Cathcart: 5 miles N. of Cathcart, Jessop 660. Cradock: 17
miles along Wapadberg road from Graaff-Reinet-Middelburg road, Jessop 653. East London: East
London, Bokelman 5-28 (NBG). Fort Beaufort : Beaufort, Cooper 523 (BOL). George: near George,
Barker 6856 (NBG). Gordonia: Kakoup, 50 miles S.W. of Olifantshoek, Leistner 2046. Graaff-
Reinet: near Graaff-Reinet, Bolus 132 (BOL). Heidelberg: 5 miles W. of Heidelberg, Jessop 589.
Herbert: St. Clair, without collector 237 (GRA). Humansdorp: Ratelsbosch, Fourcade 552. Kentani:
Kobongaba Hills, Pegler 1463. Kimberley: Christian’s Drift on Riet River, Leistner 1995. King
Williams Town: King Williams Town, Sim 1063 (BOL). Knysna: Leisure Island, Ta_ vlor 1019 (NBG).
Komga: Komga, Flanagan 299. Mafeking: Mafeking, Burtt Davy 11087. Middelburg: Groot
Vlei, Van Wyk’s Vlei, Archibald 3042 (GRA). Mossel Bay: Gouritz Valley, Horn s.n. Mount Currie:
Kokstad, Mogg 5177. Oudtshoorn: near Cango, Bolus 12369 (BOL). Paarl: Klein Drakenstein
at Salem, Bolus s.n. (BOL). Peddie: Fish River Valley below Breakfast Vlei, Acocks 1190. Port
Elizabeth: Cradock Place, Galpin 6408. Queenstown: Queenstown, Galpin 2152. Somerset East:
Zuurberg, Compton 20284 (NBG). Stellenbosch: Franschhoek, Schlechter 9218. Stockenstrom :
Hogs Back, Rattray s.n. (GRA). Stutterheim: Ripplemead, Kabousie, Hutton 482 (GRA). Tarkastad:
near Hofmeyer turning on Cradock-Tarkastad road, Jessop 655. Uitenhage: Enon, Thode A2769
(NH). Uniondale: between Krakeel and Joubertina, Jessop 595. Vryburg: Stella, 30 miles N. of
Vryburg, Hardy & Bayliss 1244. Wynberg: Orange Kloof, Wolley Dod 2470 (BOL).
Basutoland.— Leribe, Dieterlen 346.
Natal. — Bergville: Natal National Park, Edwards 962. Camperdown: Bothas Hill, Hutchinson,
Forbes & McClean 55 (NH). Durban: Berea, Forbes 383 (NH). Eshowe: Eshowe, Gerstner 2405.
Estcourt: Dalton Bridge, Acocks 10626. Himeville: 5 miles N. of Himeville, Jessop 664. Inanda:
Inanda, Wood 989 (NH). Klip River: near Colenso, Wood 4089 (NH). Lions River: 5 miles along
Merrivale-Boston road, Moll 937. Mtunzini: Umlalazi Nature Reserve, Ward 4275. New Hanover:
Albert Falls, Comins 84 (NH). Nongoma: Bangonoma, Gerstner s.n. (NH). Pietermaritzburg:
Pietermaritzburg, Schlechter 3297 (BOL). Pinetown: Gillitts, Lansdell s.n. (NH). Richmond:
Richmond, Schlechter 6717 (BOL). Umlazi: Isipingo, Wood 12480 (NH). Umvoti: Greytown, Wylie
s.n. Utrecht: Tweekloof, Altemooi, Thode A212.
Orange Free State. — Bethlehem: Kloof, Potgieter 81. Bloemfontein: Bloemfontein, Potts s.n.
(BOL). Boshoff: Smitskraal, Burtt Davy 10305. Fauresmith: Sleutelpoort, Henrici 2593. Harrismith:
Maccauvlei, Brandmuller 1 19. Kroonstad: Kroonstad, Powt 195. Winburg: Willem Pretorius Game
Reserve, Leistner 2975.
Swaziland. — Near Gollel, Compton 29105; Forbes Reef, Compton 30974; Stegi, Compton 26021.
Transvaal. — Barberton: Berea Hill, Galpin 622. Christiana: Bloemhof, Kameelpan, Theron
5633. Ermelo: Mavieriestad, Pott s.n. Heidelberg: Suikerbosrand, Mogg 18129. Johannesburg:
Zoo Kopjes, Moss 10498. Krugersdorp: Jack Scott Private Nature Reserve, Weds 2413. Letaba:
Tzaneen, Rogers 12506 (GRA). Lydenburg: Lydenburg, Thode A1648 (NH). Middelburg: Olifant’s
River Gorge, Mogg 22445. Nelspruit: 10 miles S. of Lower Sabie, Codd 6134. Pietersburg: Haenerts-
burg, Rogers 19041 (BOL). Pilgrim’s Rest: Sabie, Rogers 23090. Potchefstroom : Potchefstroom,
50
Burt/ Davy 48. Potgietersrus: Mosdene, Galpin 17346. Pretoria: Pretoria, Schlechter 3603. Sout-
pansberg: Spelonken, Jenkins s.n. Standerton: Standerton, Leendertz s.n. Ventersdorp: Goed-
gedacht, Sutton 522. Vereeniging: Lagerspoort, Mogg 19532. Wakkerstroom: Oshoek, Devenish
97. Warmbad: Warmbad, Leendertz s.n. Waterberg: banks of the Limpopo, Leipoldt 71 (BOL).
South West Africa. — Grootfontein: West Runtu “ in Okavongotal ”, Volk 1860a (M). Luderitz:
Aus, Peyer 15 (Z). Mariental: 60 miles from Mariental, Liebenberg 5154. Okahandja: Okahandja,
Hopfner c-65 (Z). Omaruru: Brandberg, Rodin 2713. Otjiwarongo: Okosongomingo, Volk 2568
(M). Rehoboth: Gravenstein, Volk 1 1 502 (M). Warmbad: Naruchas, Dinter 7214 ( M). Windhoek:
Voigtland, Walter 75 (M).
Bechuanaland. — Toteng, N.E. tip of Lake Ngami, Story 4690; 103 miles N.N.W. of Kan, De
Winter 7350.
Lamarck described A. africanus from a specimen which was sent from the Cape
by Sonnerat. The description mentions the angled stems and spine and cladode
characters which are typical of the form growing in the south-west Cape. This is
the area where Sonnerat is likely to have been able to have collected it. The holotype
is in the Lamarck Herbarium in Paris. Lamarck also cited Plukenet’s Phytographia,
t. 374 f. 4. This figure shows a slightly tortuose stem without spines, but Plukenet
mentioned the presence of spines in his description. The locality was given as Africa.
If this figure is intended to represent the Cape species, the shape of the cladodes and
absence of spines makes it a very poor likeness.
Kunth cited Drege 8576, which was collected at the Swartkops River, as the type
of Asparctgopsis juniperina. All the characters described are those of this species. In
particular, the spines, described as “ ferruginea ”, are very characteristic. Asparctgopsis
schlechtendalii Kunth is also likely to be a synonym for this species. The type was
collected along the Olifants River, but Kunth does not indicate which Olifants River
this is. The plant, is described as woody with flexuose branches. The cladodes and
axillary flowers are typical for A. africanus, but there is not enough information to
be absolutely certain.
Because various authors have used the name A. asiaticus L. for the species treated
here as A. africanus, it is desirable to discuss the typication of A. asiaticus here. Linnaeus
cited his Hortus Cliffortianus, page 122 (1737) in his description of A. asiaticus. In this
work, there is a reference to Plukenet’s Phytographia (1699) t. 15 f. 4. As there is
no specimen with this name in the Linnaean Herbarium, the Plukenet figure would
be regarded as the iconotype if there is no specimen in Herb. Sloane in the British
Museum upon which this illustration is based. Plukenet cited his species as “ e Maderas-
patans”; that is from the eastern coast of Peninsular India. Whether or not the
Indian species is synonymous with the South African plants called A. africanus in this
treatment would require further work on the material from India and Africa. There
are, however, several characters of Linnaeus’ description and of the figure which make
it very unlikely that these plants are conspecific. Linnaeus described the stems as
erect, and the figure also shows a rigidly erect stem. The figure also does not match
the South African species as it appears to have spines on the final branches. It therefore
seems better to regard this name as not applicable to the South African species until
work is done on the genus in a wider geographical context. This decision is in agree-
ment with that of Kies (1951).
Those South African Asparagi with fascicled, terete cladodes, stems tortuous
and not white and grooved, final branches and cladodes not horizontal as in A. setaceus
have been included in this species. The variation within these limits, as can be seen
from the description, is very great. It is possible that further field work would reveal
the limits of significant groupings within this definition, but, judging from the available
material and data, there seems to be no way of correlating any of the variable characters
in order to subdivide the group.
51
The form in the south-west Cape, A. africanus of Baker, has well-developed,
reddish spines, and tends to be less tortuous. This is replaced to the east by a more
pubescent form with straighter, reflexed branches and cladodes, referred to by Baker
as A. africanus var. dependens. More generally over most of the summer rainfall area
there is a less spiny, more tortuous form with narrower stems, branches and cladodes,
which was referred to A. asiaticus by Baker. This form becomes slightly herbaceous,
and may even produce annual stems.
In hot, dry areas specimens occur in which cladodes do not form, or are reduced,
towards the tips of the branches. This may be associated with the environment. Several
species have been based on specimens of this form, among them A. burkei Bak. and
A. judtii Schinz.
It is difficult to distinguish A. africanus from A. setaceus. The form of branching
and cladode arrangement which are so well known and characteristic of the horti-
cultural plants occurs in wild populations over a large part of South Africa. Inter-
mediate forms do occur, however, in which the branching shows some similarity with
A. africanus. When there is any doubt, these species may be separated by the terminal
flowers of A. setaceus, and axillary flowers of A. africanus.
6. A. setaceus ( Kunth ) Jessop, comb. nov. Asparagopsis setacea Kunth. Enum.
PI. 5: 82 (1850). Type: “ Drege, Herb. no. 8584. c.v.s. in Herb. reg. Berol. et Luc.”
(KIEL, lecto.!).
Asparagus declinatus sensu Bak. in Journ. Linn. Soc. Bot. 14: 609 (1875); LI. Cap.
6: 261 (1896). A. plumosus Bak. in Journ. Linn. Soc. Bot. 14: 613 (1875); Planchon
in Llor. des Serres ser. If, 23: t. 2413-2414 (1880); Bak. in FI. Cap. 6: 260 (1896).
Syntypes: Port Natal, Drege 4482 (K!); British Kaffraria, Cooper 202, partly, ( K ! ) ;
without locality, Gerrard & Me Ken 754.
Roots fibrous. Stems twining, smooth or grooved, glabrous, greenish to reddish-
brown. Branches similar to the stems, solitary. Cladodes linear, fine, arcuate, terete,
up to 10-nate, up to 10 mm long; final branches and cladodes all in the same plane,
horizontal (homalotropic). Spines not usually present, reflexed. Peduncles terminal,
usually solitary, rarely up to 3-nate, 2-5 mm long, articulated near the middle, pendent.
Perianth segments cream-white, oblong-obovate, entire, similar, about 3 mm long.
Stamens slightly shorter than the perianth segments; anthers ^-3 mm long. Styles
shortly divided, about 1 mm long; ovary about 1 mm long. Berry red, fleshy, up to
3-seeded, about 6 mm diameter.
This species occurs mainly in forests, but also in open country provided there is
sufficient support, from the Riversdale and Swellendam districts in the south-west,
through the eastern Cape and Natal into the Transvaal. To the north its range
continues into Rhodesia, and in the east into Mozambique. Llowers have been recorded
from September to April.
Cape. — Albany: Sandy Drift near Grahamstown, Daly 6 (GRA). Alexandria: Olifantshoek
between mouths of Bushman’s and Sunday’s Rivers, Ecklon & Zeyher s.n. Bathurst: Horseshoe on
Kowie River, Barker 9267 (NBG). George: Karatara, Drege 8584 (KIEL). Humansdorp: Ratelsbos,
Fourcade 377 (BOL). Kentani: woods near Ken tani, Pegler 345. King Williams Town : King Williams
Town, Sim 1043 (BOL). Knysna: Knysna, Marloth 7504. Komga: Komga woods, Flanagan 1711.
Queenstown: Umbombola Mountains, Galpin 2091. Riversdale: above Novo, Muir 2837 (GRA).
Somerset East: Boschberg, MacOwan 1917 (SAM). Stockenstrom: Katberg near Sanatorium,
Schonland 4307 (GRA). Swellendam: Grootvadersbos, Zeyher 4167 (BOL). Umtata: Convent of
the Holy Cross, without collector. Willowvale: Qora River mouth, Hilner 443.
Natal. — Durban: woods near Durban, Wood 437 (BOL). Eshowe: Eshowe, Lawn 730 (NH).
Estcourt: Giant’s Castle, Symons 40. Hlabisa: Hluhluwe Game Reserve, Ward 2159. Inanda:
Wood 1649 (BOL). Nkandla: Nkandla Forest, Nixon s.n. (NH). Ubombo: near turning to Mkuzi
Game Reserve, Jessop 668.
Swaziland.— Mbabane, Malagwane Hills, Ben Dlamini s.n. (NBG).
52
Transvaal. — Barberton, Thorncroft 2116. Belfast: Belfast, Leendertz 7953. Brits: Jacksonstuin,
Van Vuuren 65. Krugersdorp: 1 mile N.W. of Krugersdorp, Mogg 20305. Letaba: Grootbos
Government Forest Reserve, Sc/ieepers 896. Nelspruit: Numbi, Van der Schijff 1614. Pietersburg:
near Woodbush, Gillett 3192 (BOL). Pilgrim’s Rest: Mariepskop, Van der Schijff 5006. Pretoria:
Fountains Valley, Repton 717. Soutpansberg: Downs, Junod 4133. Ventersdorp: Goedgedacht,
Sutton 715.
In his description of Asparagopsis setacea, Kunth cited only one collecting, namely
Dregs 8584. However, he referred to two specimens, of which one was in Berlin,
and has now been destroyed, and one in “ Herb. Luc.”. This specimen is now in
Kiel. It is labelled Drege 8584b and was collected at Karatara in the George district.
Baker cited several specimens in his description of A. plumosus. Of these, Drege
4482 (K) from Port Natal, and Cooper 202, partly, (K) from British Kaffraria, have
been seen by the present author. A third specimen, Gerrard & McKen 754 without
precise locality, could not be found at Kew. Baker cited Asparagopsis setacea Kunth
under the synonymy of Asparagus declinatus L. He separated A. declinatus from
A. plumosus on the cladodes being filiform in A. plumosus but more robust in A.
declinatus. Asparagus declinatus was described by Linnaeus in his Species Plantarum
(1753) as being spineless, with declinate branches, and setaceous cladodes, and as
coming from Africa. Baker upheld this species in both the Journal of the Linnean
Society, and in the Flora Capensis, and cited MacOwan 1917 and Thunberg s.n. These
gatherings are included here in A. setaceus. It is possible that A. declinatus L. and
A. setaceus are conspecific, but the true identity of A. declinatus is obscure and it is
best discarded as a nomen dubium.
7. A. denudatus (Kunth) Bak. in Journ. Linn. Soc. Bot. 14: 606 (1875); FI. Cap.
6: 258 (1896). Type: Queenstown, Stormberg, Drege 3533 (BOL, lecto.!).
Asparagopsis denudata Kunth, Enum. PI. 5: 82 (1850).
Asparagus namaensis Schinz in Bull. Herb. Boiss. 1, 4, app. Ill: 44 (1896); Solch,
Beitr. FI. Siidwest-Afr. 37 (1961). Type: Great Namaqualand, Fleck 307a (Z, holo.!).
A. pearsonii Kies in Bothalia 6: 179 (1951); Solch, Beitr. FI. Siidwest-Afr. 38 (1961).
Type: Raman’s Drift, Pearson 4003 (K, holo.).
Roots fibrous. Stems erect to about 1 m or scrambling to nearly 2 m, zigzagging,
glabrous, smooth or wrinkled but never regularly grooved, green; the outer layers
easily separable from the aerial stems and branches, but not peeling. Branches usually
solitary but sometimes 2- or 3-nate, frequently — especially the more distal ones —
spreading or reflexed, not bearing branchlets, similar to the stems. Cladodes terete,
straight or slightly arcuate, up to 30 mm long, up to 5-nate laterally but usually 2- or
3-nate terminally (rarely up to 8-nate terminally), the outer layers often lost from the
distal parts of the cladodes after maturity. Spines developed only on the larger branches
and the stems, small, reflexed, straight, up to 3 mm long, pale coloured. Peduncles often
terminal and up to 4-nate or rarely more, axillarily only 2- or 4-nate, 2-6 mm long,
articulated near or below the centre. Perianth segments entire, similar, oblong to
oblong-obovate, greenish white to yellowish white, 2 -5-3-0 mm long. Stamens slightly
shorter than the perianth segments; anthers orange. Style about 1 mm long, shortly
divided; ovary about 1 mm long. Berry red, up to 3-seeded, about 4 mm diameter.
This species is found in two areas with rather different climates. In the north-west
it occurs over most of the drier areas of South West Africa, Bechuanaland and the
north-west Cape. In the south-east it is confined to Basutoland and the neighbouring
areas of the Cape and Orange Free State. Flowers have been recorded from October
to April in the south-east, but are apparently mainly restricted to the second half of
this period in the north-west.
53
Cape. — Aliwal North: Elandshoek, Bolus 151. Cathcart: Cathcart, Comins 1741. Gordonia:
Aughrabies Falls, Leipoldt 4437 (BOL). Herschel : Majuba Nek near Sterkspruit, Hepburn s.n. (GRA).
Queenstown: Long Hill, Galpin 2092 (BOL). Sterkstroom: Stormberg, Drege 3533 (BOL). Wode-
house: Buffelsfontein, Stretton 66.
Basutoland. — Leribe, mountain slopes, Dieterlen 828; Mafeteng, Tatai River, Dieterlen s.n.;
Mamathes, Jacot Guillarmod s.n.
Orange Free State. — Ficksburg: Westburg, Galpin 13881. Paul Roux: Rexford, Acocks 13174.
South West Africa. — Bethanie: Kuibis, Range 854 (SAM). Gibeon: Haribes, Volk 12168 (M).
Karibib: road to Roten Kuppe, Kinges 3184 (M). Keetmanshoop : 3-6 miles from Aroab on road
to Koes, De Winter 3457. Kaokoveld: Otjitundua, Barnard s.n. (SAM). Lude itz: Weissenborn,
Kinges 2404. Maltahohe: Duwisib, Volk 12482 (M). Okahandja: Osondu, Walter 1364 (M). Ovambo-
land: Odonga-Nukuambi, Rautanen 105 (Z). Swakopmund: Walvis Bay, Hardy & De Winter 1442.
Kunth based his description of this species on Drege 3533. The specimen in the
Berlin Herbarium has been destroyed and, as the specimen bearing this number in
the Bolus Herbarium matches Kunth’s description, it has been selected as a lectotype.
Asparagus denudatus was characterised mainly by the absence of cladodes, but
several specimens have been found in which a few cladodes are present. In all collec-
tions which include both flowering and sterile material, the sterile specimens invariably
have cladodes. This shows that the absence of cladodes is only a seasonal phenomenon.
The surface characters of the stems and the appearance of the spines are extremely
characteristic both of the typical material of A. denudatus , which was originally described
from the south-eastern Cape, and of typical A. namaensis, which was described from
South West Africa. These two species were separated from one another on the presence
or absence of cladodes. As it has been shown that this is not as significant a character
as previously thought, it is more satisfactory to combine these species.
8. A. virgatus Bak. in Saund. Ref. Bot. t. 214 (1870); Journ. Linn. Soc. Bot.
14: 606 (1875); FI. Cap. 6: 259 (1896). Type: Somerset East, Boschberg, Burchell
3197 (K, lecto.!).
Roots fibrous. Stems erect to 1 m, straight, glabrous, smooth, slightly herbaceous.
Branches similar to the stems, ascending, solitary, not bearing branchlets. Cladodes
up to 7-nate, but frequently 3-nate, up to 25 mm long, usually straight. Spines only
represented by a blunt spur, never pungent. Peduncles terminal and axillary, up to
6-nate, 5-9 mm long, usually articulated below the centre. Perianth segments oblong-
obovate, similar, entire, 3-4 mm long, pale yellowish green. Stamens slightly shorter
than the perianth segments. Style shortly divided, 1 mm long; ovary sometimes slightly
more than 1 mm long; cells 3- or 4-ovuled. Berry globose, 1-seeded, smooth or
wrinkled, about 5 mm diameter, red.
This species occurs in most parts of the wetter areas of the eastern half of South
Africa. It is particularly frequent on the margins of forests, but also grows in other
shady places. It also occurs in Rhodesia and Mozambique. Flowers have been
recorded mainly from September to January, but also in April.
Cape. — Aliwal North: Elands Hoek, Bolus 205 (BOL). Barkly East: Barkly East, Gerstner 180.
Cathcart: Inverthorn, Barker 3513 (NBG). East London: East London, Rattray s.n. (GRA). Kentani:
Kobongaba River, Pegler 1462. King Williams Town : King Williams Town, Sim 1050 (BOL). Komga:
Komga, Flanagan 1292. Mount Currie: Kokstad, Mogg 1805. Somerset East: Zuurberg, Holland
199. Stockenstrom: Katberg, Ecklon & Zeyher s.n. (BOL). Stutterheim: Donga Range, Acocks
9336. Uitenhage: Uitenhage, Schmutz s.n. (BOL).
Basutoland. — Maseru, Compton 22534 (NBG).
Natal.- — Bergville: Cathedral Peak, Esterhuysen 12878. Camperdown: Cato Ridge, Fairall 147
(NBG). Durban: Isipingo, Schlechter 2993 (BOL). Eshowe: near Reservoir, Lawn 554 (NH).
Estcourt: Estcourt, West 1334. Hlabisa: Hluhluwe Game Reserve, Ward 2283. Inanda: Inanda,
Wood 1373 (NH). Klip River: Van Reenen, Franks s.n. (NH). Lions River: Howick Falls, Barker
3567 (NBG). Ngwavuma: Lebombo Mountains, Gwalaweni Forest, Strey 4692. Nongoma: 11
miles N. of Nongoma, Jessop 672. Pietermaritzburg: Oribi, Lawson 641 (NH). Ubombo: near
turning to Mkuze Game Reserve, Jessop 667. Utrecht: Utrecht, Thode A1297.
54
Swaziland. — Mbabane, Fonteyn, Compton 25980 (NBG); Isateki Beacon, Compton 27313.
Transvaal. — Barberton: Barberton, Galpin 1155. Belfast: Belfast, Bruce 480. Brits: Jacksons-
tuin. Van Vuuren 211. Ermelo: Breyten, Burtt Davy s.n. Letaba: Duivelskloof, Scheepers 328.
Lydenburg: 4^ miles N. of Branddraai, Codd 6713. Marico: Zeerust, Leendertz 11324. Middelburg:
near Botsabelo, Marloth 1 1751. Pietersburg: Magoebaskloof, Papendorf 230. Pilgrim’s Rest: Nooit-
gedacht. Young A656. Potgietersrus: Naboomspruit, Galpin 343. Pretoria: Wonderboom Reserve,
Repton 2015. Rustenburg: Kloof, Lanham s.n. Soutpansberg: Soutpansberg, Burtt Davy 2654.
Wakkerstroom : Oshoek, Devenish 407. Warmbad: near Warmbad, Acocks & Hafstrom 196. Water-
berg: Zandspruit near Sandrivierspoort, Forssman s.n.
An illustration accompanies the type description. There is also a reference to
the manuscript name, A. sylxaticus Burch. There is a specimen in the Kew herbarium,
collected by Burchell on the Boschberg, bearing a label with the name A. sylvaticus
on it. As this specimen matches both Baker’s description and the figure, and it is
extremely likely that this is the source of the manuscript name to which Baker referred,
it is necessary to choose, in the absence of any other specimen, between this specimen
and the illustration for the lectotype of this species. The specimen is preferred.
De Winter 3909 from Okavongo (M, PRE) has been regarded as belonging to
this species, but has rather large thorns, and rather woody stems and branches. Its
true identity is obscure.
9. A. fasciculatus Thunb., FI. Cap. ed. 2: 329 (1818). Type: Cape without precise
locality, Masson s.n. (UPS, holo.!; BOL!, photos).
Asparagopsis consanguinea Kunth, Enum. PI. 5: 76 (1850). Type: Cape without
precise locality, Drege s.n. (7KIEL!, ?K!).
Asparagus consanguineus (Kunth) Bak. in Journ. Linn. Soc. Bot. 14: 615 (1875);
FI. Cap. 6: 260 (1896).
Roots slightly thicker than usual for this section, but not tuberous. Stems prostrate
or weakly climbing to about 1 m, glabrous, softly woody, usually light brown or less
often green, especially when young, to dark brown, zigzagging at the nodes. Branches
solitary, usually slightly reflexed; final branches not well enough defined to be regarded
as branchlets, very narrow, often consisting of a single internode with a fascicle of
cladodes at the base, or rarely with more than two or three internodes. Cladodes
often extremely fine, linear, arcuate, numerous in a fascicle, 15-30-nate or more, up
to 7 mm long, but often only about 4-5 mm long; the cladodes in each fascicle usually
not overlapping those of adjacent nodes. Spines may be present on the lower parts
of the main stems, but rarely exceeding 2-3 mm in length. Peduncles axillary or
terminal, 1-3-nate, variable in length, up to 5 mm long, articulated at the base of the
flower. Perianth segments oblong, reflexed, about 4 mm long, with a broad brownish
streak. Stamens as long as the perianth segments, with a spur at the base; anthers
yellow, about 0-5 mm long. Style about 2 mm long. Berry yellow, fleshy, globose,
wrinkled, normally 3-seeded.
This species occurs in the south-western Karroo, as well as parts of the western
Cape. It is frequent in the coastal vegetation of the Sand Veld, but has not been
recorded further south than Milnerton. Flowers have been recorded from March
to June.
Cape. — Calvinia: 5-) miles S. of Loeriesfontein, Acocks 14414. Cape: near Blaauwberg, Oliver s.n,
(BOL). Ceres: Muishond River, Esterhuysen 1805. Clanwilliam: Clanwilliam, Zeyher 1675 (STE).
Hopefield: Langebaan, Jessop 70 (BOL). Laingsburg: Whitehill, Compton 14645 (NBG). Malmes-
bury: S. side of Paardeberg, Van Niekerk 358 (BOL). Namakwaland: Witbank, Pi/tans 5113 (BOL).
Tulbagh: Saron, Schlechter 7890. Vanrhynsdorp: Bitterfontein, Ecklon & Zeyher 1675.
There appears io be no doubt from the protologue that this was the species which
Thunberg was describing. There are three specimens labelled A. fasciculatus in the
Thunberg Herbarium. Of these two do not match Thunberg’s description as they
55
have well developed spines. The third, however, matches in every character. This
specimen, on which is written: “ e. Cap. b. Spei. Mason”, belongs to A. fasiculatm
as contrued here. The collector was probably Francis Masson, who was in South
Africa at the same time as Thunberg.
Kunth based his species on a Drege specimen, but he did not give either the locality
or collecting number. In his description, he mentioned the apical articulation of the
peduncle, which is the best character of this species, the climbing habit, the lack of
spines, and the numerous, fascicled cladodes. Two specimens collected at the Cape
by Drege, and showing the above characters, have been examined by the author. The
specimen at Kiel lacks a number. As Kunth mentioned in his Enumeratio Plantarum
other specimens which are now in the Kiel Herbarium, it seems possible that this was
the specimen on which Kunth based his species. This specimen also lacks a locality.
A specimen in the Kew Herbarium was collected between the Kroin River and Petrus-
fontein on the Piketberg. It also lacks a number and may be an isotype.
In several respects A. fasciculatus does not seem to belong to the section Africani
but rather to the section Crispi. It has numerous cladodes, fibrous roots and few-
seeded fruits, characteristic of the section Africani, but the spurred anther filaments
and herbaceous habit would refer it to the section Crispi. On balance, it seems to be
best placed in the section Africani.
10. A. thunbergianus Schult. in R. & S., Syst. Veg. 7: 333 (1829); Bak. in Journ.
Linn. Soc. Bot. 14: 615 (1875); FI. Cap. 6: 262 (1896); Salter in FI. Cape Penins.
175 (1950). Type: Cape Peninsula, Thunberg s.n. (UPS, holo. ! ; BOL!, PRE!, photos).
Asparagus albus sensu Thunb., Prodr. 66 (1794). A. capensis sensu Marloth, FI.
S. Afr. 4: pi. 20 (1915).
Asparagopsis dregii Kunth, Enum. PI. 5: 84 (1850). Type: Cape of Good Hope.
Drege s.n. (?K, iso.!). A. thunbergii Kunth, Enum. PI. 5: 85 (1850), nom. illegit,
Type: as for A. thunbergianus. A. niveniana sensu Kunth, Enum. PI. 5: 88 (1850).
Roots fibrous. Stems erect to 1 m, or forming a dense bush up to 150 cm high
and 200 cm broad, in dense shade sometimes climbing to 150 cm, pale to dark brown,
glabrous, zigzagging. Branches glabrous, spreading or ascending, solitary, similar to
the stems. Cladodes fascicled, terete, slightly arcuate, 3-8 mm long, rarely up to
20 mm, ascending or spreading. Peduncles axillary or terminal, usually solitary, 3-6
mm long, articulated below the centre. Perianth segments spreading, white, obovate,
2-3 mm long. Stamens nearly as long as the perianth segments; anthers brown,
0-5 mm long. Styles about 1 mm long; ovary about 1 mm long. Berry globose,
up to 8 mm diameter, with a single seed, red.
This species occurs, usually in low bush, from the Cape Peninsula to Clanwilliam
in the north and Uitenhage in the east. Flowers have been recorded from January
to June.
Cape. — Bellville: Tigerberg, Pillans 8658 (BOL). Bredasdorp: Nachtwacht, Smith 3022A. Caledon:
Hermanus, Rogers 26572. Cape: above Camps Bay, Marloth 9492. Clanwilliam: Clanwilliam,
Leipoldt 408 (SAM). George: Montagu Pass near Herold, Jessop 592. Hopefield: between Hopefield
and Paternoster, Leipoldt 3895 (BOL). Humansdorp: along river, Britten 1 162. Knysna: Belvidere,
Duthie 763 (BOL). Malmesbury: Darling Flora Reserve, Rycroft \9[ 1 (NBG). Paarl: Dal Josaphat,
Tyson 858 (BOL). Piketberg: Het Kruis, Stephens & Glover s.n. (SAM). Riversdale: Albertinia
commonage, Muir 931. Simonstown: W. of Simonstown, Wolley Dod 933 (BOL). Somerset West:
Knorhoek, Jessop 77 (BOL). Stellenbosch: Stellenbosch, De Jongh s.n. Swellendam: Grootvadersbos,
Taylor 1042 (NBG). Tulbagh: Saron, Schiechter 7887. Uitenhage: Theescomb, Paterson 853 (BOL).
Wellington: Wellington, Thomson 153. Wynberg: Kirstenbosch, Middlemost 2008 (NBG).
Thunberg included Asparagus albus L. in his Prodromus plantarum Capensium.
In his herbarium, there is a sheet labelled “ Asparagus albus , p. Withering ", On this
sheet there is a specimen of A. thunbergianus Schult., but there is also a specimen of
56
another species — perhaps A. albus L. from Europe. The specimen of A. thunbergianus
is of a young shoot, but it shows the reflexed, solitary spines, smooth (somewhat
wrinkled during drying) stem and general appearance of a young shoot of A. thun-
bergianus. The description, however, applies equally well to either specimen. It
seems probable that Thunberg did not realise that there were specimens from different
countries on this sheet, but recognised the European specimen as A. albus L.
Schultes described A. thunbergianus as being from the Cape Peninsula. He
described it as a shrub up to two feet or more high with grooved stems and with solitary,
recurved spines and fascicled cladodes. He cited Thunberg’s A. albus L. as a synonym,
and as no other specimen on which he could have based his species is known to have
existed the Thunberg specimen may be regarded as the holotype.
Asparagopsis dregii was described by Kunth as having been collected at the Cape
of Good Hope by Drege, but this specimen was destroyed in Berlin. The branches
were described as smooth, glabrous and spinous. He regarded it as very similar to
Asparagopsis thunbergii (based on Asparagus thunbergianus Schult.) but the description
makes it very likely that these names were based on specimens of the same species.
A specimen in Kew was collected by Drege but, like the specimen which Kunth had,
had no number. It is labelled “ Table Mountain and Devils Mountain. Asparagopsis
dregei Kth.”. It seems very likely that this is an isotype.
Asparagus nivenianus Schult. was based on a specimen collected by Niven at the
Cape. This specimen has not been located and it is not possible to tell from the
description whether this was a specimen of A. thunbergianus or of A. africanus. This
name must therefore be regarded as a nomen dubium. Kunth transferred the epithet
to Asparagopsis and cited Ecklon 95b from the Cape as the only specimen he had seen
of Asparagopsis niveniana. A specimen of this collecting in the Munich Herbarium is
of A. thunbergianus.
11. A. mucronatus Jessop, sp. nov., A. thunbergiano Schult. afltnis, sed caulibus
ramisque maioribus, bracteis manifesto evolutis et ad basim ramulorum persistentibus
differt. Plate 1.
Radices non tuberosi. Caules recti ad 2 m, plerumque c. 1 m, basi bracteis
manifestissime evolutis ad 2 cm longis, glabri, costis magnis, basi recti, supra tortuosi,
cani vel fusci. Rami caulium partibus distalibus similes, rami minores nonnumquam
pubescentes vel scabridi, Ramuli fasciculati, ad 6-nati, pubescentes vel scabridi, recti
ad 6 cm longi; ramuli juvenes saepe bracteis manifesto evolutis et basi persistentibus
ad 5 mm longis. Cladodia recta vel arcuata, teretes, mucrone non-spinoso et non-
colorato, 3-8 mm longa, in ramulis orta, ad 10-nata. Spinae magnae, recurvae.
Pedunculi in ramulis axillares, 1-vel 2-nati, 4-7 mm longi, submedio articulati. Perianthii
segmenta oblongo-obovata, 3-4 mm longa, margine integro, similia. Stamina quam
perianthii segmenta paulum breviora; antherae £ mm longae, subglobosae. Ovarium
fere 1 mm longum; stylus breviter ramosus, c. 1 mm longus. Bacca globosa, succulenta,
deinde rugosa, rubra, fere 5 mm diam., seminibus 1 vel saepe 2.
Type: Cape: Middelburg district, Grootfontein, Theron 551 (PRE, holo.).
Roots fibrous. Stems erect to 2 m, usually about 1 m high, bearing very pronounced,
acute bracts up to 2 cm long at the base, glabrous, smooth, with large ridges, straight
at the base, becoming slightly zigzag, grey to brown. Branches similar to the distal
parts of the stems, the younger ones sometimes pubescent or scabrid. Branchlets
fascicled, up to 6-nate, pubescent or scabrid, straight, up to 6 cm long, often bearing
well-developed bracts up to 5 mm long when young and persisting towards the base.
As the stems are at least biennial, at times new and old branchlets occur simultaneously,
in which case the younger ones are paler, more pubescent, and bearing larger bracts,
and possibly peduncles. Cladodes straight or arcuate, terete, with a non-pungent,
57
colourless mucro, 3-8 mm long, borne on the branchlets, up to 10-nate. Peduncles
axillary on the branchlets, 1- or 2-nate, 4-7 mm long, articulated below the middle.
Perianth segments oblong-obovate, 3-4 mm long, entire, similar. Stamens slightly
shorter than the perianth segments; anthers £ mm long, rather globose. Ovary
about 1 mm long; style shortly divided, slightly less than 1 mm long. Berry globose,
fleshy, wrinkled when ripe, about 5 mm diameter, with a single or occasionally two
seeds, red.
This species is fairly common in at least some parts of the Karroo and Little
Karroo, but also occurs as far east as the Cathcart district. Flowers have been recorded
from October to December.
Cape. — Albany: Queens Road, Britten 5159 (GRA); Botha Hill, Dyer 1124 (GRA); Plutos Vale,
Jessop 637. Calvinia: Calvinia, Schmidt 358. Cathcart: about 5 miles N. of Cathcart, Jessop 661.
Graaff-Reinet : Graaff-Reinet, Henrici 4943. Hay: Blaauwbosput, Acocks 2011. Herbert: Valsch-
fontein, Acocks 2598; Heuningneskloof, Moran s.n. (BOL). Kimberley: Mauretzfontein, Acocks
2580; Acocks 4957 (BOL). Laingsburg: Matjiesfontein, Foley 62. Middelburg: Grootfontein,
Theron 551 (BOL). Montagu: between Dobbelaarskloof and Montagu, Esterhuysen 1872; 10 miles
N.E. of Montagu, Jessop 678. Port Elizabeth : Redhouse, Denman 282; Paterson 960 ( BOL) ; Paterson
977 (BOL). Vanrhynsdorp: Langeberge, Marloth 12953. Worcester; Veld Reserve, Olivier 276.
The greatest similarity is with A. thunbergianus. It is, however, a much larger
and more robust plant. It also differs in having scale leaves, and pubescence on the
final branches.
12. A. macowanii Bak. in Journ. Linn. Soc. Bot. 14: 609 (1875). Type: Albany,
Zeyher 879 (K, lecto.!; BOL!).
Roots fibrous. Stems erect, softly woody, glabrous, smooth or slightly grooved,
usually white, up to about 2 m high. Branches similar to the stems, solitary. Cladodes
fascicled, numerous, terete, arcuate, up to about 12 mm long. Spines not well-
developed, absent from final branches. Peduncles terminal or axillary, sometimes
solitary but usually up to 1 2-nate, articulated proximally. Perianth segments oblong
to oblong-obovate, 2-5 mm long, pale. Stamens slightly shorter than the perianth
segments; anthers less than 0-25 mm long; ovary 1-5-2 mm long, ovoid. Fruit not
known.
Southern Cape, eastwards through Natal, and in the eastern districts of the
Transvaal. Also in Mozambique.
For key to varieties see key to species.
(a) var. macowanii.
A. macowanii Bak. in Journ. Linn. Soc. Bot. 14: 609 (1875); FI. Cap. 6: 261
(1896).
The typical variety is characterised by its smaller size. The stems are rarely over
75 cm high. The stems are sometimes grooved at first, but soon become smooth.
This variety occurs in the southern Cape from Knysna to Komga, and inland
as far as Glen Grey. Flowers have been recorded from September to November and
in April and May.
Cape. — Albany: Howieson’s Poort, Zeyher 879 (BOL, K). Bathurst: Horseshoe on the Kariega
River, Jessop 616. Bedford: Bedford, Gone 88 (GRA). Glen Grey: White Kei Falls, Galpin 2504.
Knysna: Plettenbergbaai, Rogers 26141. Komga: Prospect Farm, Flanagan 296. Somerset East:
Somerset East, MacOwan s.n. (BOL, SAM); Rogers 161; 164 (GRA); Boschberg. Tyson 1878.
Baker, in his type description, cited Zeyher 879 from Albany and MacOwan 1917
from Somerset East. In the Flora Capensis, however, he removed the MacOwan
specimen to A. declinatus. It therefore seems best to regard the Zeyher specimen as
58
the lectotype. There are specimens of a MacOwan collecting of this species in the
Bolus and South African Museum herbaria. This collecting is not numbered but was
also collected in the Somerset East district. It seems possible that these are of the
same collecting as the specimen seen by Baker when he drew up the original description.
It would otherwise be hard to explain the epithet.
(. b ) var. zuluensis (N. E. Br.) Jessop, stat. nov.
A. zuluensis N. E. Br. in Kew Bull. 1912: 283 (1912). Type: Zululand, Wylie sub
Medley Wood 11969 (K, holo.!; NH!, PRE!).
This variety differs from the typical variety in being a much larger plant. The
stems are much-branched and up to 2 m high. The stems and branches are always
smooth and whitish. Spines are developed, although not large or pungent, on the
main stems and branches only.
Occurs from the Camperdown district in the south, through Zululand and into the
eastern Transvaal as far north as the Pilgrim’s Rest district. Flowers have been recorded
from September to November.
Natal. — Camperdown: Cato Ridge, McC/ean & Ogilvie s.n. (NH). Eshowe: Umlalazi Bridge,
Lawn 750 (NH). Hlabisa: Hluhluwe Game Reserve, Ward 4334. Mahlabatini: Umfolozi Game
Reserve, Ward 4379. Mtonjaneni: Mtonjaneni, Gerstner 3665. Mtunzini: Ngoya Forest Reserve,
Huntley 268.
Transvaal. — Barberton: Barberton, Thorncroft 16. Piet Retief: 13 miles from Pongola to Gollel,
Dyer & Verdoorn 5832. Pilgrim’s Rest: Eileen Orpen Dam near Tshokwane, Codd 6138.
N. E. Brown based A. zuluensis on a single collecting from Zululand by Wylie.
Medley Wood made the herbarium specimens which he distributed from a specimen
growing in the Durban Botanic Garden. A plant which is very likely the original
Wylie specimen is still growing in these gardens near the herbarium.
The differences between this and the type variety are not usually distinguishable
from herbarium material unless a note on the habit accompanies the specimen. The
distribution of the two varieties, however, makes it possible to identify with certainty
material of which the origin is known.
13. A. retrofractus L., Sp. PI. 313 (1753); Bresler, Diss. 10 (1826); R. & S., Syst.
Yeg. 7: 330 (1829); Bak. in Journ. Linn. Soc. Bot. 14: 621 (1875); FI. Cap. 6: 266
(1896); Salter in FI. Cape Penins. 176 (1950). Type: without locality or collector
(LINN No. 434.4, holo.; NBG!, PRE!, photos).
Asparagopsis retrofracta (L.) Kunth, Enum. PI. 5: 78 (1850).
Asparagus africanus Lam. var. microphyllus Bak. in FI. Cap. 6: 266 (1896). Type:
Rumman, Kosi Fontein, Burchell 2572 (K, holo.; PRE, photo.!). A. microphyllus
(Bak.) Kies in Bothalia 6: 178 (1951).
Roots fibrous. Stems scrambling or weakly climbing to 2 m or more, or forming
dense bushes up to 2 m or more high, white and grooved at first, but becoming brown
and smooth, zigzagging; swollen lateral buds are present at the nodes. Branches
similar to the stems, spreading or reflexed, not bearing branchlets, pubescent when
young. Cladodes fascicled, 6-30 mm long, arcuate, terete, spreading or ascending,
not all the same length. Spines sometimes present, strongly reflexed, occasionally
pungent, up to 6 mm long, never borne on the final branches. Peduncles axillary or
terminal, usually 2-6-nate, 5-40 mm long, articulated below the centre. Perianth
segments white with a green streak, narrow obovate, spreading, 3 mm long, margins
entire, similar. Stamens almost as long as the perianth segments; anthers orange.
Style divided for about half its length, 1 mm long; ovary slightly longer than the style.
Berry globose, 1-3-seeded, 5 mm diameter, orange; perianth generally persistent.
59
This species occurs mainly in the drier areas of South West Africa and the Karroo*
but extends into the winter-rainfall area as far as the Cape Peninsula. Flowers have
been recorded mainly from April to June, but also at other times.
Cape. — Barkly West: Klipvlei, Esterhuysen 2083 (BOL). Beaufort West: Sunnyside, Esterhuysen
4357. Bellville: Tigerberg, Pil/ans 4758 (BOL). Caledon: 12 miles west of Caledon, Salter 7215
(BOL). Calvinia: Calvinia, Schmidt 556. Cape Town: Porcupine Buttress, Pil/ans 4327 (BOL).
Carnarvon: 12 miles from Carnarvon on Van Wyks Vlei road, Acocks 1741. Ceres: Karroopoort,
Marloth 9008. Clanwilliam: Modderfontein, Gillett 3695 (BOL). Gordonia: Aughrabies Falls,
Jessop 351 (BOL). Graaff-Reinet : 14 miles north of Graaff-Reinet, Jessop 651. Herbert: Campbell
Kloof, Wilman 4069 (BOL). Laingsburg: Whitehall, Compton 8523 (NBG). Malmesbury: near
Darling, Bolus 12859 (BOL). Montagu: Cogman’s Kloof, Barker 959 (NBG). Namakwaland: Honde-
klip Bay, Pil/ans s.n. (BOL). Oudtshoorn: road to Armoed via Baakens Kraal, Britten 1706. Prieska:
Prieska district, Bryant J224. Stellenbosch: Papagaaisberg, Duthie 452 (STE). Vanrhynsdorp: Heeren-
logementsberg, Zeyher 1674 (BOL). Vryburg: Vryburg, Mogg 8976. Worcester: Welgevonden,
Esterhuysen 1903.
South West Africa. — Gibeon: Harris, Fleck 296a (Z). Keetmanshoop: Swartkop, Ortendahl 83.
Ludertiz: Aus, Peyer 15 (Z). Maltahohe: Duwisib, Volk 126363 (M). Warmbad: Kanus, Range
907 (SAM). Windhoek: Windhoek, Rogers 30480 (Z).
Linnaeus, in his type description, cited Plukenet, Phyt. t.375 f.3. This figure
shows the typical characters of A. retrofractus L. as construed here. There is a specimen
in the Linnaean Herbarium labelled A. retrofractus and probably received before
1753 (see Jackson, 1912). This specimen is therefore likely to be the holotype.
This species is replaced to the east by A. laricinus which, particularly on herbarium
material, is sometimes difficult to distinguish. Usually a less regular arrangement of
the cladodes distinguishes A. retrofractus, but the typically ascending arrangement
of the cladodes is not always apparent in A. laricinus, which can then be identified
by having small spines on the final branches, and by its more eastern distribution.
In the field, A. retrofractus can often be recognised as a climber or scrambler, while
A. laricinus is a large bush.
14. A. compactus Salter in Journ. S. Afr. Bot. 6: 165 (1940); FI. Cape Penins.
175 (1950); Kies in Bothalia 6: 179 (1951). Type: Cape Peninsula, Orange Kloof,
Wolley Dod 2521 (BOL, holo.!).
A. africanus Lam. var. wrightii Bak. in Journ. Linn. Soc. Bot. 14: 619 (1875); FI.
Cap. 6: 266 (1896). Type: Simons Bay, Wright 454.
Roots fibrous. Stems erect, normally about 50 cm long, zigzagging, woody,
white, minutely pubescent, grooved; outer layers sometimes lost leaving the stems
brown and smooth. Branches spreading to ascending, similar to the stems. Cladodes
fascicled, commonly up to 10-nate terminally, and fewer laterally, more or less straight,
ascending to spreading, 3-angled, rarely up to 1 mm thick, rigid, 10-30 mm long.
Spines pungent, spreading, recurved or reflexed, present on the aerial stems and all
branches. Peduncles axillary or terminal, 1- or 2-nate laterally, and up to 4- or 5-nate
terminally, 5-7 mm long, articulated below the centre. Perianth segments obovate,
similar, entire, white, about 3 mm long. Stamens slightly shorter than the perianth
segments, not spurred; anthers about f mm long. Style divided for nearly half
its length, about 1 mm long. Berry globose, fleshy, red, about 6 mm diameter, 1-3-
seeded.
This species is confined to the south-western Cape, occurring from near sea-level
to 3,000 ft or more on well-drained mountain slopes. Flowers have been recorded
in October, December, January and from March to May.
Cape. — Bellville: south side of Tigerberg, Pil/ans 8655 (BOL). Caledon: Swartberg near Caledon,
Schlechter 10364. Cape: Paarden F.iland, Salter 8268. Ceres: Milner Peak, Esterhuysen 7804 (BOL).
Malmesbury: near Kalabas Kraal, Salter 7198 (BOL). Paarl: above Franschhoek, Jessop 102 (BOL).
Piketberg: Porterville, Loubser 810 (BOL). Simonstown: Clovelly, Salter 8209 (BOL). Somerset
60
West: Knorhoek, Jessop 89 (BOL). Stellenbosch: Swartboskloof, Van Rensburg 268 (STE). Suther-
land: near Sutherland, Du Toit s.n. (BOL). Tulbagh: Sneeuwgat Valley, Great Winterhoek, Phillips
1880 (SAM). Wellington: Bains Kloof, Salter 7307 (BOL). Wynberg: Orange Kloof, Wolley Dod
2521 (BOL).
The holotype is cited by Salter as Wolley Dod 2521 (BOL). Salter regarded this
species as very probably identical with A. africanus var. wrightii Bak., which was based
on Wright 454. Baker’s description mentioned grooved branches. The cladodes
were given as very stout, angled, with a groove down each face, and the berries as
being larger than for the typical variety. This is not adequate to identify the species
with certainty, but with the locality, given as Simonstown, it is almost certainly this
species. According to Salter (1940), Professor Compton searched for the type in
Europe, but was unable to find it.
15. A. laricinus Burch., Trav. 1: 537 (1822); Bresler, Diss. 40 (1826); R. & S.,
Syst. Veg. 7: 337 (1829); Kunth, Enum. PI. 5: 75 (1850); Bak. in Journ. Linn. Soc.
Bot. 14:^620 (1875); FI. Cap. 6: 267 (1896); Solch, Beitr. FI. SUdwest-Afr. 37 (1961).
Type: Hay, between Griquatown and Wittewater, Burchell 1971 (K, holo.!; PRE,
photo. !).
Asparagus africanus Lam. var. concinnus Bak. in Journ. Linn. Soc. Bot. 14: 619
(1875); FI. Cap. 6: 266 (1896). Type: Pearston, Bruintjes Hoogte, Burchell 3020
(K, holo.!; PRE, photo.!). A. concinnus (Bak.) Kies in Bothalia 6: 178 (1951).
Root'> fibrous. Stems erect, forming much-branched shrubs up to 2 m high, but
usually only about 150 cm or less, often zigzagging or tortuous, woody, white, pubescent
and usually grooved when young, but soon losing the outer layers and becoming smooth
and darker. Branches similar, but retaining the grooved outer layers longer. Cladodes
fascicled, 15-60-nate laterally, more numerous terminally, more or less straight,
ascending, terete, fine but rigid, all approximately the same length at each node, 8-30
mm long. Spines pungent, ascending to recurved, usually present on all branches.
Peduncles axillary or terminal, commonly up to 7 or 8 terminally and 3 or 4 laterally,
up to 7 mm long, articulated below the centre. Perianth segments obovate, similar,
entire, white, 2-5-3 mm long. Stamens slightly shorter than the perianth segments,
not spurred; anthers about f mm long. Style divided for nearly half its length,
about 1 mm long. Berry globose, about 6 mm diameter, fleshy, red, 1-3-seeded.
This species has a wide distribution from the south-eastern Cape to South West
Africa, Bechuanaland and the Transvaal. Flowers have been recorded from September
to December.
Cape. — Albany: Mountain Drive, Grahamstown, Britten 2805 (GRA). Aliwal North: Elandshoek,
Bolus 204 (BOL). BarklyWest: Daniels Kuil, Esterhuysen 1153 (BOL). Cathcart: Fairford, Cotterell
126 (GRA). Fort Beaufort: Fort Beaufort, Zey/ier s.n. Flay: between Wittewater and Griquatown,
Burchell 1971 (K). Herbert: Heuningneskloof Station, Moran s.n. (BOL). Kimberley: Kimberley,
Moran s.n. (BOL). Mafeking: Mafeking, Burtt-Davy s.n. Mount Currie: Kokstad, Mogg 1806.
Port Elizabeth: Swartkops Estuary, Archibald 4914 (GRA). Port St. Johns: Eagles Nest, Schonland
4077. Somerset East: Somerset East, Bowker s.n. (BOL). Sterkstroom: Andriesberg, Galpin 2168.
Vryburg: Armoedsvlakte, Mogg 7922.
Batutoland. — Leribe, Dieterlen 406.
Natal. — Mooi River: Warley Common, Mogg 7152. Mtonjaneni: Melmoth, Lawn 1828 (NH).
Mtunzini: Ginginhlovu, Wylie s.n. (NH). Umzinto: Dumisa, Rudatis 1735.
Orange Free State. — Bloemfontein: Leeuwkop, Potts s.n. (BOL). Fauresmith: Veld Reserve,
Henrici 395 1 . Harrismith: Rensburgskop, Jacobsz 103. Kroonstad: Kroonstad, Pont 485. Senekal :
Doornkop, Goossens 649. Thaba Nchu: The Willows, Roberts 2992. Vredefort: near Vredefort,
Pole Evans 19474H. Wepener: sandstone hillside, Acocks 1172. Winburg: W. Pretorius Game
Reserve, Bakker 1006.
Transvaal. — Barberton: Malelane, Murphy s.n. Belfast: Dullstroom, Codd 5638. Bethal: Steen-
kool River, Weintraub s.n. Ermelo: Nooitgedacht, Henrici 1128. Germiston: Wattles, Moss 10503.
Heidelberg: Suikerbosrand, Mogg 18133. Johannesburg: near Johannesburg, Watt cfc Brandwyk
2207. Kroonstad: Kroonstad, Pont 485. Krugersdorp: 30 miles W.N.W. of Johannesburg, Mogg
61
24235. Lichtenburg: Witstinkhoutboom, Liebenberg 44. Lydenburg: Steenkampsberg, Young
A428. Marico: near Zeerust, Van der Merwe 128 (BOL). Pilgrim’s Rest: between Pilgrim’s Rest and
Sabie, Rogers 20564. Potchefstroom: Potchefstroom, Burtt-Davy 48. Pretoria: Irene. Rogers 11406
(BOL). Vereeniging: Vereeniging, Leendertz s.n. Wakkerstroom : Oshoek, Devenish 1157. Warmbad:
Warmbad, Leendertz s.n. Wolmaransstad: Boskuil, Sutton 56.
South West Africa. — Gobabis: Okasondana, Walter 4163 (M). Windhoek: Voigtland, Walter
116 (S.W.A. Herb.).
Bechuanaland. — Gaberones, Rogers 6254; Pharing, Kanye, Hillary & Robertson 466.
Burchell cited only his C.G. 1971 from between Griquatown and Wittewater.
This specimen, in the Kew Herbarium, is therefore taken as the holotype.
In some areas, particularly to the south and east of its distribution, plants occur
with rather small cladcdes, but otherwise the normal characters for this species. A
specimen of this form. Burchell 3020 from Bruintjes Hoogte, was described by Baker
as A. africanus var. concinmis. Kies raised this variety to specific status, but in the
present author’s opinion there are no grounds for giving this form taxonomic recog-
nition.
This species is similar in many respects to both A. retrofractus L. and A. macouanii
Bak. var. zuluensis (N. E. Br.) Jessop. The characters by which it can be distinguished
from A. retrofractus are given under that species. The regular arrangement of the
cladodes, the usually grooved stems, the development of spines on the final branches,
and the more western distribution serve to distinguish it from A. macowanii var.
zuluensis.
16. A. microrhaphis (Kunth) Bak. in Journ. Linn. Soc. Bot. 14: 612 (1875); FI.
Cap. 6: 262 (1896). Type: Queenstown, Drege 3534 (BOL, lecto.!).
Asparagopsis microrhaphis Kunth, Enum. PI. 5: 83 (1850).
Asparagus stellatus Bak. in Journ. Linn. Soc. Bot. 14: 612 (1875); FI. Cap. 6: 262
(1896). Syntypes: “Albert Div.”, Cooper 622 (K !) ; Aliwal North, Witteberge, Drege
8589 (K!).
Roots fibrous. Stems erect to 3 m, glabrous or minutely pubescent, ashen grey,
grooved but becoming smooth and brown, softly woody, straight, much-branched.
Branches similar to the stems, aiso branched, not bearing branchlets. Cladodes fascicled,
glabrous or minutely pubescent, usually up to about 8-nate but occasionally up to
25-nate, 1-4 mm long. Spines usually well developed on aerial stems and main
branches. Peduncles axillary, 1- or 2-nate, 5-6 mm long, articulated below the centre.
Perianth segments oblong-obovate, entire, similar, spreading, white, 3-4 mm long.
Stamens almost as long as the perianth segments. Styles shortly divided, about 1 mm
long. Fruit not known.
This species occurs in bush or on grassy slopes in the eastern Cape, Basutoland
and Natal. Flowers have been recorded in October and December.
Cape. — Herschel: Herschel, Gerstner 18. Mount Ayliff: Kokstad, Mogg 1912. Queenstown:
Table Mountain, Drege 3534 (BOL). Sterkstroom: Andriesberg, Galpin 2169. Wodehouse: Buffels-
fontein, St ret ton 65.
Basutoland. — Leribe, Dieterlen 301; Mafateng, Thababa-Morena, Dieterlen 1374.
Natal. — Bergville: Ndedema River Valley, Killick 1861. Estcourt: Giants Castle, Symons 38.
Lions River: Inhleyane, Moll 1286. Wakkerstroom: 7 miles N.E. of Wakkerstroom, Acocks 11525
(NH).
Kunth, in his description of Asparagopsis microrhaphis, cited Drege 3634. The
holotype is likely to have been destroyed in Berlin, but there is a specimen with this
number in the Bolus Herbarium. As it matches Kunth's description in every detail
this specimen has been selected as a lectotype.
62
A number of specimens, such as Cooper 622 and Drege 8589, are representatives
of a group possibly deserving varietal status. They have extremely small and very
numerous cladodes. This group was described by Baker under the binomial A. stellatus.
No specimens of this vegetative form have ever been collected in flower, however,
and it seems likely that this is only a seasonal form.
17. A. exuvialis Burch., Trav. 1: 432 (1822); Bresler, Diss. 40 (1826); R. & S.,
Syst. Veg. 7: 337 (1829); Kunth, Enum. PI. 5: 74 (1850); Bak. in Journ. Linn. Soc-
Bot. 14:" 608 (1875); FI. Cap. 6: 259 (1896); Schinz in Bull. Herb. Boiss. 1, 4, app-
III: 42 (1896); Salter in FI. Cape Penins. 175 (1950); Solch, Beitr. FI. Siidwest-A.fr-
36(1961). Type: “Along the Vaal River, Griqualand West ”, Burchell 1768 (K, holo.!;
PRE, photo. !).
A. ecklonii Bak. in Journ. Linn. Soc. Bot. 14: 615 (1875); FI. Cap. 6: 258 (1896).
Type: South Africa, without locality, Ecklon & Zeyher s.n. (TCD, holo.!; PRE,
photo.!; PRE, iso.!). A. hereroensis Schinz in Bull. Herb. Boiss. 1, 4, app. Ill: 43
(1896). Syntypes: Rehoboth, Fleck 295a (Z!); Harris, Fleck 549 (Z !).
Roots tuberous; the tubers almost sessile, about 5-10 cm long. Stems erect or
twining, usually about 50 cm high, but occasionally up to about 2 m, woody, glabrous,
smooth; the outer layers of the young stems pale, and peeling to leave the stems
darker, sometimes almost black, straight or zigzagging. Branches similar to the stems,
often spreading. Cladodes fine, terete, fascicled, from 5-30 mm long, deciduous.
Spines absent or poorly developed, and reflexed; the leaf firm and deltoid, of a rather
characteristic shape. Peduncles axillary, generally paired, but up to about 5-nate,
articulated near or below the middle, 3-5 mm long; the flowers opening in a racemose
sequence along the branches. Perianth segments oblong-obovate, entire, similar,
about 3-5-4 mm long, with a brown or purplish streak, spreading. Stamens not
spurred, nearly as long as the perianth segments; anthers yellow, about f mm long.
Style shortly divided, about 1 mm long; ovary about 1 mm long. Fruit globose,
turning black, dry, smooth; perianth not persistent.
This species is generally found in the Karroo and Kalahari, but does occur in
moist areas as well. It is distributed from the south-western Cape northwards to the
northern Transvaal, Bechuanaland and South West Africa. Flowers have been recor-
ded from October to April. In the Cape Peninsula, and probably other areas, the flo-
wers open in the morning, close at mid-day, and the following morning the next flo-
wer on the branch opens.
Cape.- Barkly West: Daniels Kuil, Esterhuysen 1149 (BOL). Bell vi lie : Kanonberg, Compton
115622 (NBG). Cape: Paarden Eiland, Salter 8294 (BOL). Clanwilliam: Lange Kloof, Schlechter
8051. Cradock: W. of Rayner's Koppie, Acocks 11926. Kimberley: Dorstfontein, Acocks 1404.
Kuruman: Kaap Plateau, Marloth 14070. Malmesbury: Darling Flora Reserve, Barker 8670 (NBG).
Namakwaland: Doornpoort, Pillans 5467 (BOL). Port Elizabeth: Redhouse, Paterson 963 (BOL).
Postmasburg: Postmasburg, Esterhuysen 1153. Prieska: Draghoender, Bryant 1174. Queenstown:
Queenstown, Rogers 2321 (GRA). Tulbagh: Wolseley, Schlechter 9949. Uitenhage: Uitenhage,
Ecklon & Zeyher 607. Vanrhynsdorp: Klawer, Andreae 411. Williston: Williston Station, Smith
2445. Worcester: near De Dooms, Bolus 13201. Wynberg: Princess Vlei, Jessop 88 (BOL).
Orange Free State. — Bloemfontein: Bestersput, collector unknown, 13. Fauresmith: Voelfontein,
Kies 338. Jacobsdal: near Heuningneskloof Station, Phillips 3480. Luckhoff: Rosemarie, Kalkbult,
Henrici 4260.
Transvaal. — Pilgrim’s Rest: 20 miles W. of Satara, Codd 6197. Potgietersrus: Naboomsrpuit,
Galpin 347. Pretoria: 6 miles W. of Hammanskraal, Kies 371. Sibasa: Punda Milia, Lang s.n.
Soutpansberg: Soutpan, Obermeyer, Schweickerdt & Verdoorn 51.
South West Africa. — Grootfontein : 157 miles E. of Grootfontein, Story 6199 (M). Karibib:
Marmorwerk, Volk 2757 (M). Luderitz: Aus, Dinter 6136 (B). Okahandja: Otjihua, Dinter 329
(SAM). Otjiwarongo: Okosongomingo, Volk 2523 (M). Outjo: Kamanyab and Kaross, Thorne
s.n. (SAM). Rehoboth: Rehoboth, Fleck 295a (Z). Warmbad: Sandfontein, Wilman 1602 (SAM).
Windhoek: Voigtland, Walter 2778 (M).
Bechuanaland. — Ngami See, Schinz 900 (Z); Mochudi, Marloth 950.
63
There is a form from the north-eastern Transvaal in which the steins are more
tortuous than usual, the flowers are more than 2-nate, and the perianth segments are
yellowish (e.g. Code I 6197, Obermeyer, Schweickerdt & Verdoorn 97, Long s.n. and
Van der Schijjf 911 ; all in PRE). This form does not seem to be sufficiently different
to warrant separate taxonomic status.
18. A. oxyacanthus Bak. in Journ. Linn. Soc. Bot. 14: 625 (1875); FI. Cap. 6:
270 (1896). Type: Somerset East, Bowker s.n. (K, holo.!; PRE, photo.!).
Tubers distant, about 6 cm long. Stems often unbranched, erect, smooth, straight,
up to about 50 cm long, pubescent, with numerous narrow, softly spinous leaf spurs,
grey to brown. Branches solitary, lacking branchlets, similar to the stems. Cladodes
solitary, oblong, up to 4 cm, rarely to 8 cm, long, with a median vein and thickened
margin, mucronate. Peduncles 2 mm long, solitary in the axils of large bracts, articulated
distally. Perianth segments obovate, obtuse, white, 2-3 mm long, spreading to ascending,
rather variable in width, but the inner whorl broader. Anthers large, up to 0-7 mm
long. Style branches free, short; ovary globose. Fruit not known.
This species is restricted to the area from the Bathurst district to the East London
district. It occurs mainly on rather dry slopes, particularly the slopes above the Fish
River. Flowers have been recorded from October to January.
Cape. — Albany: Overlooking Kaffir Drift, Leach & Bayliss 12646; Fraser’s Camp, Maguire 632
(NBG). Bathurst: Fish River, Dyer 3375; Schlechter 6105 (BOL); Kleinemonde, White 952 (SAM).
East London: East London, Rattray s.n. (GRA); Green Point, Smith 3765; East London Park,
Wood s.n. Peddie: Bergha River Valley opposite Wooldridge, Acocks 11885.
This is one of the “ fox-tail ” species of the section Racemosi, but can easily be
distinguished by the broad, mucronate cladodes, and the pubescent stem.
19. A. juniperoides Engl, in Bot. Jahrb. 10: 3 (1889); Schinz in Bull. Herb.
Boiss. 1, 4, app. Ill: 44 (1896); Solch, Beitr. FI. Slidwest-Afr. 37 (1961). Type:
S.W.A., Aus, Marloth 1538 (B, holo.!).
Tubers sessile or shortly stalked, fusiform. Stems erect, glabrous, grooved, straight,
annual, up to 1 m high. Branchlets narrow, numerous, up to 3 cm long, borne directly
on the stems; branchlets and cladodes ascending, forming a dense cylinder round the
stem. Cladodes solitary, flattened, usually very narrow, linear, often with ciliate
margins, up to 1 cm long. Leaves not forming spines. Peduncles axillary, solitary,
about 2 mm long, articulated distally. Perianth segments similar, oblong, with ciliate
margins, ascending, 6 mm long, white with a green streak. Stamens nearly as long
as the perianth segments; filaments not spurred; anthers minute, yellow. Ovary
globose, 1-0-1 -5 mm long. Berry globose, with one to three seeds, about 5 mm
diameter, red.
This species occurs from Tulbagh, up the west coast into South West Africa.
Flowers have been recorded from April to July.
Cape. — Clanwilliam: Lange Kloof, Schlechter 8028; Zeyher 4167 (STE); Graafwater, Zinn s.n.
(SAM). Namakwaland: Spektakel, Esterhuysen 5867 (BOL); Kamieskroon, Lewis 2146 (SAM);
Vioolsdrif, Middlemost s.n. (NBG); Doornpoort, Pillans 5379 (BOL); Witbank, Pillans 5139 (BOL).
Tulbagh: Tulbaghkloof, Zeyher s.n. (BOL). Vanrhynsdorp: 5 miles N. of Nuwerus, Compton 20571
(NBG); near Klawer Station, Marloth 7794 ; Heerenlogementsberg, Zeyher s.n. (BOL).
South West Africa. — Luderitz: Kuckaus, Dinter 3723; Spitskop, Merxmiiller & Giess 3418;
Tsirub, Range 207 4 (B).
20. A. minutiflorus ( Kunth ) Bak. in Journ. Linn. Soc. Bot. 14: 616 (1875); FI.
Cap. 6: 263 (1896). Type: Delagoa Bay, Forbes s.n. (B, holo.!; PRE, photo.!).
Asparagopsis minutifiora Kunth, Enum. PI. 5: 85 (1850).
7230691-3
64
Roots tuberous; tubers distant or lateral on the roots. Stems erect, almost
straight, softly woody, ridged, usually scabrid along the ridges, pale, up to about 50 cm
long. Branches usually absent. Branchlets straight usually fascicled, ascending,
branched, very narrow, spineless, ridged, scabrid, up to about 5 cm long. Cladodes
3-angled, up to about 10 mm long and about 0 2 mm thick, up to about 8-nate laterally
but more numerous terminally, slightly arcuate. Spines usually present on the stems,
slender, terete, pungent, more or less straight and spreading, the longest from 5-12
mm long, brown. Peduncles borne on stems and less often on branches, solitary or
paired, about 3 mm long. Pedicels usually 2- or 3-nate and terminal, but occasionally
lateral and solitary, borne in the axils of prominent bracts, about 1 mm long, articulated
near the middle. Perianth segments oblong-obovate, 2-3 mm long; the apex of the
inner whorl dentate. Stamens nearly as long as the perianth segments; anthers minute.
Style shortly divided, about 1 mm long; ovary about 1 mm long. Fruit with a smooth
surface, 1 -seeded, up to about 7 mm diameter.
This species has been recorded from a range of habitats, from stony hillsides to
dense thorn-scrub on the flats, widely spread over most of Natal, the eastern Transvaal
and Mozambique.
Natal. — Estcourt: Rondedraai, Acocks 10817. Ingwavuma: 10 miles S.E. of Ingwavuma, Codd
10296. Ngotshe: “ Ngotsche ”, Gerstner 2516 (BOL). Nongoma: Bangonomo, Gerstener s.n.
(NH). Ubombo: Mkuze Game Reserve, Gerstener 4597.
Swaziland. — Stegi, Tambuti Ranch, Compton 30940; St. Phillips Mission Station, Gerstner s.n.
(NH).
Transvaal. — Barberton: Avoca, Thorncroft 3006. Letaba: 3^ miles W. of Letaba Camp, Codd
& Dyer 4694. Pilgrim's Rest: 6 miles N.E. of Skukuza, Codd & Verdoorn 5484.
This and the following species have a unique type of inflorescence by which they
can be distinguished from any other species. In these the raceme has apparently been
so reduced as to leave only 2-4 flowers which are borne terminally. In a few specimen
there may be one or two lateral flowers as well showing what is probably an inter-
mediate stage between these species and the rest of the Section Racemosi.
21. A. rigidus Jessop , sp. nov., A. minutifloro Bak. affinis, sed cladodiis maioribus,
spinis minoribus, caulibus glabris differt. Plate 2.
Caules erecti, leviter tortuosi, lignei, costati, glabri, saepe virides, ad 60 cm longi.
Rami solitarii, leviter tortuosi, quam caules costatiores, brevissimi, inultum ramosi,
plus minusve ascendentes, ad 8 cm longi; rami ultimati non ramulos eximios lateraliter
facti. Cladodia valde 3-angulata, lateraliter ad 3-nata et terminaliter fere 6-nata,
plerumque at 10 cm longa et 0-5-0 -9 mm crassa, plus minusve recta, mucrone tereti
fuscoque; cladodiorum margines saepe costa subtiliter serrata cellarum dilatarum.
Spinae absentes vel ad 4 mm longae, reflexae. Pedunculi in ramis orti, solitarii vel
gemini, c. 3 mm longi. Pedicelli plerumque 2-4-nati terminales sed saepe in pedunculis
laterales, prope medium articulati. Perianthii segmenta oblongo-obovata, 2-3 mm
longa, integra. Stamina fere perianthii segmenta aequilonga; antherae c. 0-5 mm
longae. Stylus breviter ramosus, c. 1 mm longus. Ovarium c. 1 mm longum. Bacca
semine 1, c. 5 mm diam.
Type: Transvaal: Pilgrim’s Rest district, Mariepskop, Killick & Strey 2380
(PRE, holo.).
Stems erect, slightly zigzag, woody, ridged, glabrous, often green, up to 60 cm
long. Branches solitary, slightly zigzagging, more strongly ridged than the stems,
very short, much branched, more or less ascending, up to about 8 cm long; the final
branches not forming specialised branchlets. Cladodes strongly 3-angled, up to 3-nate
laterally and about 6-nate terminally, usually up to about 10 cm long and 0-5-0 -9
mm thick, more or less straight, with a terete, brown mucro; the margins of the cladodes
65
frequently bearing a minutely serrated ridge of enlarged cells. Spines absent or up
to 4 mm long, reflexed. Peduncles borne on the branches, solitary or paired, about
3 mm long. Pedicels usually 2- to 4-nate and terminal but often lateral on the peduncles,
articulated near the middle. Perianth segments oblong-obovate, 2-3 mm long, entire.
Stamens nearly as long as the perianth segments; anthers about 0-5 mm long. Style
shortly divided, about 1 mm long; ovary about 1 mm long. Fruit 1 -seeded, about
5 mm diameter.
This species grows among scrub in rocky areas. Although it is mainly restricted
to the south-eastern parts of the Transvaal, there is a single record from the Petersburg
district. Flowers have been recorded only in November.
Transvaal. — Barberton: Barberton, Pott 5546; Saddleback Mountain, Thorncroft 790 (NH).
Pietersburg: The Downs, Rogers 21982. Pilgrim’s Rest: Mariepskop, Kitlick & Strey 2380; Van der
Schijff 5589; 8 miles N. of Sabie on Graskop road, Codd & de Winter 3313.
This species resembles the preceeding in the nature of the inflorescence, but can
be separated by the fewer, larger cladodes, smaller spines, and glabrous stems.
22. A. densiflorus ( Kunth ) Jessop, comb. nov.
Asparagopsis densiflora Kunth, Enum. PI. 5: 96 (1850). Type; Cape, Lalande s.n.
(P, lecto. ; PRE, phioto.!).
Asparagus sarmentosus sensu Bak. in Journ. Linn. Soc. Bot. 14: 625 (1875); FI.
Cap. 6: 270 (1896). A. myriocladus Bak. in Journ. Bot. Lond. 18: 43 (1889); FI.
Cap. 6: 271 (1896). Type: Inanda, Medley Wood 355 (K, holo. !; PRE, photo.!;
NH!). A. sprengeri Regel in Acta Horti Petrop. 11: 302 (1890); Bak. in FI. Cap.
6: 271 (1896). Type: described from a cultivated specimen from Durban. A. myersii
Hort., nomen nudum.
Roots tuberous; the tubers either borne directly on the rhizome or laterally on
the roots. Stems erect or weakly decumbent, straight or slightly tortuous, finely ridged,
glabrous, 30-60 cm long, green to brown. Branches similar to the stems, but not
usually formed. Branchlets numerous, grooved, occasionally over 10 cm long, spreading
or ascending. Cladodes flattened or rarely 3-angled, slightly arcuate, 5-15 mm long,
weakly to strongly mucronate, commonly solitary, but sometimes ternate or numerous,
with a distinct midrib when flattened. Spines usually well-developed, but not often
pungent, usually absent from the final branches, reflexed, the leaf attenuate. Peduncles
rather shorter than the branchlets, occasionally not differentiated. Pedicels solitary,
generally articulated near the centre. Perianth segments more or less similar, oblong-
obovate, obtuse, spreading, white or pale pink, about 2-5 mm long. Stamens three
quarters of the length of the perianth segments; anthers small. Styles divided for
about half their length. Berry globose, about 5 mm diameter, 1 -seeded, red.
This species grows in a range of habitats, mainly in the coastal areas of the south-
eastern Cape and Natal, from coastal dunes to open rocky places or woods. It has
also been recorded from Inhaca Island. Flowers have been recorded from October
to March and in July.
Cape. — Albany: Howieson’s Poort Hills, Galpin 3084. Alexandria: Alexandria Forest, Jessop
612. Bathurst: Port Alfred, Galpin 2948. Bedford: Bedford, Potts 1730 (BOL). East London:
East London beach, Bokelman 3-28 (NBG). Kentani: Kentani, Pegler 1482 (BOL). Keiskammahoek:
Keiskammahoek, Stayner 53 (GRA). Kirkwood: between Rautenbach’s Drift and Addo Drift,
Bitrchell 4208 (K). Komga: Komga, Flanagan 221 1 . Mount Ayliff: Kokstad, Tyson 31 58. Mqanduli:
Koffiebaai, Van der Schijff 5447. Peddie: Kaffir Drift, Compton 17829 (NBG). Port Elizabeth:
Redhouse, Paterson 531 (BOL). Queenstown: Umbombola Mountain, Galpin 2094 (GRA). Somerset
East: Kommadagga, Burchell 3302 (K). Uitenhage: Swartkops River, Ecklon & Zeyher 1061.
Natal. — Durban: near Durban, Medley Wood 7559 (NH). Eshowe: Eshowe, Lawn 781 (NH).
Hlabisa: Hluhluwe Game Reserve, Ward 2941. Inanda: Inanda, Medley Wood 1626 (BOL). Lower
Tugela: Tugela Beach, Johnson 408 (NBG). Mahlabatini: Mahlabatini, Gerstner 3525 (NH). Port
66
Shepstone: Beach Terminus, Thode 3429 (STE). Ubombo: Mkuze Game Reserve, Ward 3984.
Umzinto: Scottburgh, Thode 3424 (STE). Utrecht: Utrecht, Thode A1298. Weenen: Griffins Hill,
Acocks 11413 (NH).
Orange Free State. — Without precise locality, Cooper 3603 (K).
Transvaal. — Barberton: Sheba, Thorncroft 14.
Linnaeus described A. sarmentosus as having solitary, linear-lanceolate cladodes;
as being flexuose, and as having recurved spines. He cited Hermann’s Hortus Academici
Lugduno-Batavus t. 650, which shows the tuber characters. Another plate, t. 63,
shows the aerial parts. There is a specimen, LINN No. 434. 15, labelled A. sarmentosus
in the Linnaean Herbarium, but this does not fit the description as it has fascicled
cladodes. There is no specimen of this species in the Hermann Herbarium in Leyden,
and the figures must be regarded as iconotypes. Both Linnaeus and Hermann gave
the source of the material as Ceylon. According to Trimen (1898) there are two species
to which this name might apply in Ceylon, A.falcatus L. and A. gonocladus Bak. These
are both much-branched twiners, and it is unlikely that either could by synonymous
with A. sarmentosus sensu Bak. from South Africa. The name A. sarmentosus L.
can therefore not be applied to the South African species.
As Regel described A. sprengeri from Durban, and as he mentioned that it had
racemes and a striate stem, it is very likely that this name is synonymous with A.
densiflorus as construed here. Regel also mentioned that this species had been introduced
into Europe through horticulture. The plant generally known in horticulture as A.
sprengeri fits Regel’s description and plants of this form grow near Durban, making
it very probable that this is the correct application of the binomial. Another cultivated
form of A. denstflorus is known as A. myersii (often misspelt meyeri ), but this name has
never been validly published.
This is an extremely variable species, but no practical way of subdividing it has
been found. It can be identified by the fact that it never climbs to more than a little
over a metre, and then only weakly twining, not twisting to any extent round its support.
The racemes may be obscured by the development of cladodes at the nodes, but are
never reduced to the form found in A. minutiflorus and A. rigidus. The stems are
grooved or ridged, but always glabrous, and branches are usually absent, although
branchlets are well developed.
23. A. saundersiae Bak. in Journ. Bot. Lond. 18: 42 (1889); FI. Cap. 6: 269
(1896). Type: Natal, Saunders s.n. (K, holo.!; BOL!).
Stems woody, erect to climbing, up to 2 m high, hardly zigzagging, slender, glabrous,
smooth, yellowish or golden-brown. Branches similar to the stems, wide spreading
to ascending, usually solitary, lacking spines. Cladodes fascicled, about 6-8-nate,
less than 4- mm brod, angled, subulate, ascending or spreading, up to 1-5 cm long,
slightly arcuate. Spines usually well-developed, absent from the final branches, up
to 1 cm or more long, pungent, spreading, slightly recurved. Peduncles 1-4-nate,
usually about 4 cm long. Pedicels solitary, 5-6 mm long, articulation variable. Perianth
segments serrate, obovate, about 3 mm long, white. Stamens nearly as long as the
perianth segments; anthers minute, globose, yellow. Ovary stalked; style shortly
divided. Fruit not known.
This species occurs mainly at the margins of forests in the south-eastern and
eastern parts of South Africa, possibly as far north as Kenya. It occasionally grows
in the open, even on dry, exposed slopes. Flowers have been recorded from November
to May.
Cape. — Lusikisiki : between Port St. Johns and Lusikisiki, Hutchinson 1780. Mqanduli: Mqanduli,
Pegler 599. Port St. Johns: about 5 miles N.E. of Port St. Johns, Jessop 663.
67
Natal. — Camperdown: Camperdown, Carnegie s.n. (NH); Franks s.n. (NH); Botha’s Hill, Medley
Wood s.n. (NH). Eshowe: Eshowe, Gerstner 2841 (BOL). Hlabisa: near Hluhluwe, Wells 2039;
Hlabisa, Ward 2911. Mtonjaneni: Umhlatuzi Road, Lawn 331 (NH). Nongoma: Nongoma, Gerstner
4423 (NH); 11 miles N. of Nongoma, Jessop 670. Pinetown: Marionhill, Kolbe s.n. (BOL). Rich-
mond: Richmond Road, Schlechter 6740.
Swaziland. — Isateki Beacon, Compton 27312.
Transvaal. — Letaba: Duiwelskloof, Scheepers 1059.
This species is rather similar to A. buchananii Bak. The main difference is that
A. buchananii is a climber with very large spines which are characteristically flattened
at the base. The yellowish colour of the stems and branches of A. saundersiae often
distinguishes it from A. buchananii which usually has rather browner stems and
branches.
24. A. buchananii Bak. in Kew Bull. 1893: 211 (1893); FI. Trop. Afr. 7: 434
(1898). Type: Nyasaland, Buchanan 1503 (K, lecto. !; PRE, photo.!); Buchanan 757.
Stems woody, climbing to at least 9 metres, tortuous and often strongly zigzagging,
glabrous, very variable in colour, smooth. Branches similar to the main stems.
Branchlets often well-defined, solitary, lacking spines, spreading to ascending. Cladodes
fascicled, about 6- to 8-nate, less than mm broad, angled, terete or grooved, subulate,
ascending or spreading, up to 1 • 5 cm long, slightly arcuate. Spines well-developed,
absent from final branches, up to 4 cm or more long on the main stems, dorsally
flattened towards the base, usually reflexed. Peduncles 1- to 4-nate, 2-4 cm long,
usually solitary. Pedicels solitary, 3-5 mm long, articulation variable. Perianth
segments serrate, obovate to oblong-obovate. 1 • 5-3 mm long, white. Stamens nearly
as long as the perianth segments; anthers minute, globose, yellow to brown. Ovary
stalked, style slightly divided. Berries fleshy, about 5 mm diameter, 1- or 2-seeded, red.
This species is a forest climber from Natal in the south to at least as far north
as Nyasaland. Flowers have been recorded from April to June, and in September
and November.
Natal. — Mapumulo: near Thringspost, Mauve 4207. EJbombo: Makatini Flats, Vahrmeyer 432.
Transvaal. — Barberton: Sheba Valley, Thorncroft 3006. Groblersdal: Loskopdam, Mauve 4278.
Lydenburg: Bothas Hoelc, Repton 5947; Sekhukhuneland, Barnard 448. Nelspruit: Numbi, Van der
Schijff 29. Petersburg: Thabina, Schweickerdt 1019; Molopo Reserve, Gerstner 5322. Potgietersrus :
Ihlenfeldt 2067. Sibasa: Punda Milia, Codd 5292. Soutpansberg : Elim, Obermeyer s.n. Wakker-
stroom: Slangapies Mountains, Burtt Davy 1937. Warmbad: 5 miles N. of Warmbad, Vahrmeyer
& Tolken s.n. Waterberg: near Elmeston, Meeuse 10537.
South West Africa. — Grootfontein : 8 miles E. of Runtu, De Winter 3176a. Ovamboland: Oshi-
kango, De Winter & Giess 7054.
Baker cited Buchanan 1503 and Buchanan 757, both from the Shire highlands in
Nyasaland, in his description of A. buchananii. It is proposed that Buchanan 1503
(K) be the lectotype.
This species, particularly if the characters of the main stems are not known, may
be difficult to distinguish from A. saundersiae. The characters of use in distinguishing
these species are given under A. saundersiae. It may also be difficult to distinguish
from the northern forms of A. aethiopicus, which have flattened cladodes and generally
lack the large spines. These spines are the most characteristic feature of the plant,
and may reach more than 4 cm and be extremely pungent and strongly reflexed. The
flattened base is conspicuous even in fairly small spines, but may occur to a small
degree in A. aethiopicus.
25. A. aethiopicus L., Mant. 32 (1767). Type: Cape without precise locality
(LINN No. 434.6, neo.; NBG!, PRE!, photos).
68
Tubers stalked, borne laterally on the main roots, about 30 cm long, fusiform
to ovoid. Stems climbing to about 7 m or erect to 1 m, woody, terete, glabrous.
Branches grooved, spreading or recurved, similar to the stems. Branchlets sometimes
well-defined, grooved, straight. CJadodes flattened, linear, mucronate in many speci-
mens, 3-6-nate or more numerous, 10-40 mm long, 1-2 mm broad, with a single vein.
Spines usually present at all nodes, but often absent on final branches, spreading or
reflexed, often pungent, up to 1 cm long or occasionally more. Leaves and bracts
sometimes ciliate. Peduncles usually 3-15 cm long, paired, unbranched. Peduncles
usually single or paired, 2-4 mm long, usually articulated near the centre, but variable.
Perianth segments obovate, about 3 mm long, often with a ciliate margin, spreading,
white. Stamens almost as long as the perianth segments; anthers orange, 0-5 mm
long. Style short, 0-6 mm long; ovary about 1 mm long. Berry globose, 5-7 mm
diameter, with 1-3 seeds, red; perianth not persistent.
Mainly a forest climber, this species also occurs in coastal vegetation and in parts
of the Little Karroo and Namakwaland. Distributed over a large area from the
south-western Cape to central Africa.
For key to the varieties see key to species.
(a) var. aethiopicus.
A. aethiopicus L., Mant. 32 (1767); R. & S., Syst. Veg. 7: 335 (1829); Bak. in
Journ. Linn. Soc. Bot. 14: 626 (1875); FI. Cap. 6: 271 (1896); Salter in FI. Cape
Penins. 175 (1950). A. lanceus Thunb., Prodr. 66 (1794). Type: without precise
locality, Thunberg s.n. (UPS, holo. !; BOL!, PRE!, photo.).
Asparagopsis lancea (Thunb.) Kunth, Enum. PI. 5: 106 (1850). A. aethiopica (L.)
Kunth, Enum. PI. 5: 95 (1850).
In this variety the stems are grooved. The cladodes appear to be more opaque
and tend to be a little broader and shorter than average.
This variety is confined to the Cape Province, occuring mainly in the coastal
forests from the Cape Peninsula to Humansdorp. Flowers have been recorded from
January to June and in October.
Cape. — Bredasdorp: Uilenkraal Forest, Taylor 286 (NBG). Caledon: Danger Point, Pil/ans 9541
(BOL). Cape: Camps Bay, Pappe s.n. (SAM). Clanwilliam: near Lambert’s Bay, Leipoldt 4244
(BOL). George: Wilderness, Van Niekerk 211 (BOL). Humansdorp: Ratelsbosch, Fourcade 245
(BOL). Knysna: The Point, Taylor 1331 (SAM). Ladismith: Anysberg Dam, Stokoe 8520 (BOL).
Namakwaland: Witbank, Pidans 5121 (BOL). Riversdale: Bottebersfontein, Muir 1421. Robertson:
Robertson, Schmidt 24. Simonstown: Clovelly, Walgate s.n. (BOL). Swellendam: Bonnievale,
Mar loth 12006. Wynberg: Houtbaai, Compton 15598 (NBG).
Linnaeus described A. aethiopicus as having solitary, recurved spines, angled
branches and linear-lanceolate cladodes. He also mentioned that it came from the
Cape, and was similar to A. falcatus, but differed in having fewer cladodes, reflexed
spines, and flexuose, angled branches. The specimen in the Linnaean Herbarium
bears the letters “ H.U.”, indicating a cultivated specimen from the Uppsala Gardens,
as does the description in the Mantissa, and shows the typical cladodes and stem
characters of this variety. The Linnaean specimen is possibly from the same plant as
that described, but is unlikely to have been the actual specimen described, as Jackson
(1912) does not indicate that this specimen was in the Enumeration of 1767 of the
specimens in the Linnaean Herbarium. This specimen is therefore selected as the
neotype.
In the south-western Karroo and in the Little Karroo there are plants which
combine some of the characters of this species and A. racemosus. Both have grooved
branches and the form of branching characteristic of the south-western Cape specimens
of A. aethiopicus. Where the cladodes are flattened, specimens are placed in this
69
species. In flat-cladode forms from the Swellendam area (e.g. Stokoe 8520 from the
Anysberg Dam) there are exceptionally well-developed spines present on the final
branches. In the Laingsburg area (e.g. Jessop 355 from near Matjesfontein) the same
development of spines occurs, but the cladodes are three-angled, and these plants
must therefore be placed in A. racemosus.
(, b ) var. angusticladus Jessop, var. nov.
A. ternifolius sensu Hooker fil. in Bot. Mag. t. 7728 (1900). A. sprengeri sensu
Wright in Bot. Mag. t. 8052 (1906).
Haec varietas caulibus levibus, cladodiis longioribus angustioribus et statu sicco
leviter perlucidis distinguitur. Plate 3.
Type: Transvaal: Middelburg district, Olifants River Gorge, Mogg 22470 (PRE,
holo.).
This variety is distinguished by the smooth stems, and the longer and narrower
cladodes which appear to be slightly translucent when dry.
It is found over most of the Transvaal, but occurs also in Natal, Swaziland,
the extreme eastern parts of the Cape Province, and in a large part of east Africa.
Flowers have been recorded in September and October.
Cape. — Kentani: Kentani, Pegler 840. King William’s Town: King William’s Town, Sim 1061
(BOL).
Natal. — Lower Tugela: Darnall, Schmidt 11 (NH). Mtunzini: Tugela Beach, Johnson 408 (NBG).
Utrecht: Kaffir Drift, Thode A250.
Swaziland. — Mbabane, Dalriach, Compton 30739 (NBG); Hlatikulu, Compton 28903 (NBG);
Mpisi, Compton 30112 (NBG); Stegi, Compton 28020 NBG).
Transvaal. — Barberton: Barberton, Thorncroft 843 (NH). Brits: Weltevreden 61, Mogg 20391.
Carolina: Arnhemburg, Roberts s.n. Heidelberg: Kuilfontein, Mogg 20121. Johannesburg: 9
miles N.W. of Johannesburg, Mogg 19188. Krugersdorp: Jack Scott Private Nature Reserve, Wells
2412. Middelburg: Olifants River Gorge, Mogg 22470. Nelspruit: Amajuba Mountain, Liebenberg
3118. Piet Retief: 27 miles S.E. of Piet Retief, Codd 10268. Pretoria: Pienaar’s River area, Repton
3727. Soutpansberg: Elim, Obermeyer 592.
Baker described A. aethiopicus var. ternifolius in 1871. This variety has been
transferred to A. falcatus. J. D. Hooker, in Bot. Mag. t. 7728, raised this variety to
specific rank, but it is clear that the plant on which the illustration was based is referable
to the variety described above. Wright, in 1906, described a plant which was also
illustrated in the Botanical Magazine, and considered it to be A. sprengeri, but referred
to A. ternifolius as being a later synonym. Again, the figure is of a climber and must
be referred to this variety.
26. A. falcatus L., Sp. PI. 313 (1753). Type: Burmann, Flora Zeylanica t. 13, f. 2.
(1737)
Stems climbing or scrambling, tortuous, smooth, glabrous, pale, much-branched.
Branches usually solitary, spreading, similar to the stems. Cladodes flattened, straight
or falcate, with a prominent vein, up to 9 cm long and 7 mm broad, dark green. Spines
well-developed, recurved, pungent, strong, often borne on vegetative branchlets; the
rest of the leaf not usually developed. Peduncles up to 4-nate laterally, more numerous
terminally, usually about 3-5 cm long, unbranched or branched. Pedicels 4-8 mm
long, articulated proximally, 1- or 2-nate. Perianth segments oblong to obovate, the
inner whorl slightly narrower than the outer, 2-5^1 mm long, cream to white. Stamens
almost as long as the perianth segments; anthers small and globose. Style short;
ovary about 1 mm long. Berry fleshy, about 5 mm diameter, 1- or 2-seeded, red.
This species occurs in the eastern Cape and northwards through east Africa to
Asia. Flowers have been recorded in South Africa from September to February.
70
For key to the varieties see key to species.
(a) var. falcatus.
A. falcatus L., Sp. PI. 313 (1753); Bresler, Diss. 2 (1826); Kunth, Enum. PI. 5:
71 (1850); Bak. in Journ. Linn. Soc. Bot. 14: 626 (1875); FI. Cap. 6: 271 (1896).
The typical variety has larger cladodes, which are usually 3-5-9 cm long and
more or less falcate. The inflorescence is simple.
This variety occurs over the entire range of the species, mainly in forest. Flowers
have been recorded from December to February in South Africa.
Cape. — East London: East London, MacOwan 2413 (BOL). Kentani: Kentani, Pegler 1335.
Port St. Johns: Port St. Johns, Galpin 3491.
Natal. — Durban: Berea , Medley Wood 150. Eshowe: Eshowe, Lawn 73 (NH). Ingwavuma:
Kosi Bay, Vahrmeyer 494. Lower Tugela: Tugela Beach, Johnson 607 (NBG). Mtunzini: Ngoya
Forest, Wells & Edwards (A (Nil). Nkandla: Qudeni Forest, Davis 119 (NH). Pinetown: Umgababa,
Bayer 862. Ubombo: Makatini Flats, Vahrmeyer 429.
Linnaeus described A. falcatus as having solitary spines, and ensiform, falcate
leaves. He reported its origin as being from Ceylon. He also cited Burmann’s Flora
Zeylanica t. 13, f. 2 (1737). This figure must be regarded as an iconotype in the
absence of a specimen from which it was drawn. This specimen should be in Paris
according to Lanjouw and Stafleu (1954), but is apparently missing from the collection.
According to Jackson (1912), a specimen did not appear in the Enumerations of the
Linnaean Herbarium. Moreover, the specimen in the Linnaean Herbarium, LINN
No. 434.3, has short cladodes, and is in fact A. aethiopicus var. aethiopicus. The
specimen in the Linnaean Herbarium cannot, therefore, be regarded as the type. Four
specimens from Ceylon, the type locality for A. falcatus , were loaned to the author
by the Director of the Kew Herbarium. These specimens match several specimens,
such as Thode A1563 (PRE) from Zululand and Goldsmith 123/62 (PRE) from Melsetter,
exactly. In a few respects most Ceylon specimens are not a perfect match of the South
African plants, however. The racemes are shorter, being only from 1-2 cm long,
and the cladodes are longer, from 4-12 cm, and narrower, often only about 2 mm broad.
These characters cannot, however, be regarded as being of significance taxonomically.
( b ) var. ternifolius {Bak.) Jessop, comb. nov.
A. aethiopicus L. var. ternifolius Bak. in Saund. Ref. Bot. t. 261 (1871). Type: Natal,
Cooper sub Saunders 1448 (K, holo. !; PRE, photo.!). A. aethiopicus L. var . natalensis
Bak. in FI. Cap. 6: 272 (1896). Type: Natal, Inanda, Medley Wood 1351 (K, holo.!;
PRE, photo.!; NH!).
This variety has shorter cladodes than the typical variety, being up to about
35 mm long. The inflorescence is branched.
This variety does not occur in the Cape Province and probably does not extend
as far north as the typical variety. It habit is similar, but it does not appear to climb
to such a height, perhaps only two or three metres. It tends to occur in drier areas
and the height may be associated with this. Flowers have been recorded from September
to December.
Natal. — Camperdown: Drummond, Gerstner 2421. Hlabisa: Hluhluwe Game Reserve, Ward
1535. Inanda: Umzinyati Falls, Medley Wood 1357 (SAM). Lower Tugela: Tugela Valley below
Sans Souci, Edwards 1685. Mtonjaneni: Mkubaan Valley, Acocks 12936. Mtunzini: Mandini,
Gerstner 2817 (NH). Nongoma: 1 1 miles N. of Nongoma, Jessop 671. Umfolozi: about 3 miles N.W.
of Empangeni, Jessop 665. Ubombo: Mkuze, Galpin 13649. Umvoti: between Greytown and Tugela
Ferry, Edwards 922.
Swaziland. — Ingwavuma Poort, Compton 28903; Hlatikulu, Sipopaneni, Murdoch 125 (NBG);
Blue Jay Ranch, Compton 29570 (NBG).
71
Transvaal. — Barberton: Komatipoort, Van der Schijff 1130. Waterberg: near Elmeston, Mecuse
£c Strey 10430.
The climbing species of Asparagus with racemes and flattened cladodes can be
readily divided into four groups, but the status which should be alotted to these is
not easily decided. It has been decided that the most natural grouping is to place
the two groups in which the cladodes dry very dark, often almost black, together under
A. falcatus. These two can be separated with certainty only because of the difference
in inflorescence. Although plants with branched racemes tend to have shorter cladodes
than those with simple racemes, there is a complete intergradation in the cladode length
of these varieties. The other two groups, in which the cladodes dry a lighter colour,
are placed as varieties of A. aethiopicus and are differentiated from each other on the
basis of the stems being grooved or smooth.
27. A. aspergillus Jsssop, sp. nov., A. racemoso Willd. affinis, sed caulibus scabridis,
rectioribus cinereisque, antheris fuscis differt. Plate 4.
A. racemosus sensu Solch, Beitr. FI. Siidwest-Afr. 38 (1961).
Caules ascendentes vel erecti ad 2 m, lignosi, sulcati vel costati, scabridi, cani.
Rami caulibus similes sed saepe sulcatiores et scabridiores, solitarii, divaricati vel
ascendentes. Ramuli 1-3-nati, plus minusve divaricati, recti, sulcatissimi. Cladodia
fasciculata, multa, angulata, recta vel arcutata, ad 4 cm longa et mm crassa, sed
plerumque c. 1-5-2 cm longa, plerumque flore absentes. Spinae in caulibus praecipue
ramis et plerumque ramulis, sed non in pedunculis, pungentes, divaricatae vel leviter
reflexae, leviter recurvae, ad 1 cm longae, pallide fuscae. Pedunculi gemini, nonnunquam
cladodiis instructi. Peclicelli solitarii vel interdum gemini, 1-5-2 -5 mm longi, prope
florem articulati. Perianthii segmenta oblongo-obovata, integra, 2-4 mm longa, alba.
Stamina quam perianthii segmenta paulo breviora; antherae fuscissimae. Stylus
1-1-2- mm longus; rami c. mm longi. Baccus globosus, seminibus 1 vel 2. c. 6 mm
diam., ruber, pericarpio rugoso.
Type: Transvaal: Letaba district, near Birthday Road at Mamaranga, Breyer
in TRV 19063 (PRE, holo.).
Stems twining or erect to 2 m, woody, grooved or ridged, scabrid, pale grey.
Branches similar to the stems, but often more prominently grooved and scabrid.
solitary, spreading to ascending. Branchlets 1-3-nate, more or less spreading, straight,
strongly grooved. Cladodes fascicled, numerous, angled, straight or arcuate, up to
i mm thick and 4 cm long, but usually about 1-5-2 cm long, usually absent during
the flowering period. Spines present on the main stems, branches and usually on
the branchlets, but not on the peduncles, pungent, spreading or slightly reflexed, slightly
recurved, up to 1 cm long, pale brown. Peduncles paired, sometimes bearing cladodes,
usually 2-5 cm long, but up to 15 cm, not usually bearing spines. Pedicels solitary
or less often paired, 1-5-2 -5 mm long, articulated near the flower. Perianth segments
oblong-obovate, entire, 2-4 mm long, white. Stamens a little shorter than the perianth
segments; anthers very dark coloured. Style 1-1-} mm long; branchlets about -} mm
long. Berry globose, 1- or 2-seeded, about 6 mm diameter, red, with a wrinkled
pericarp.
This species occurs in the northern Transvaal, Bechuanaland, northern South West
Africa and Rhodesia. Flowers have been recorded in October and June.
Transvaal. — Letaba: near Birthday Road at Mamaranga, Breyer'm TRV 19063. Messina: Zeerust,
Tokwe Farm, Breyer s.n.; Messina, Rogers 19418. Pilgrim’s Rest: 20 miles N. of Satara, Van der
Schijff 518.
South West Africa. — Caprivi: east of Cuando River, Curson 900. Grootfontein: Oshikango,
De Winter & Giess 7053. Kaokoveld: Orumane, Abner A23 (M); 15 miles E. of Ohopuho, Merxmuller
1517.
72
This species was confused with A. racemosus Willd. by Solch (1961), but is clearly dis-
tinct. The stems differ in being grey, scabrid, and straighter. The spines tend to be rather
more slender. The anthers are very dark, almost black, giving the inflorescence the
appearance of being covered by the fruiting bodies of an Aspergillus.
28. A. racemosus Willd., Sp. PI. 2: 152 (1799). Type: India, in herb. Willdenow
(B, holo.; PRE, photo.!).
A. tetragonus Bresler, Diss. 27 (1826); Wright in Bot. Mag. t. 8288 (1909). Type:
South Africa, Mundt & Maire s.n. A. racemosus Willd. var. tetragonus (Bresler) Bak.
in Journ. Linn. Soc. Bot. 14: 624 (1875); FI. Cap. 6: 269 (1896).
Asparagopsis floribunda Kunth, Enum. PI. 5: 98 (1850), nom. illegit. Type: as for
A. racemosus. A. subquadrangularis Kunth, Enum. PI. 5: 94 (1850). Type: “ Eckl.
et Zeyh. Herb. Cap. n. 171. (v.s. in Herb. Luc.)”.
Stems twining, climbing, woody, usually grooved, glabrous. Branches similar to
the stems, solitary, spreading to ascending. Branchlets usually 1- to 3-nate, spreading
or ascending, straight, grooved, spineless. Cladodes fascicled, numerous, angled,
straight, up to 1 mm broad, generally 1-2 cm long, green at first but turning greyish.
Spines pungent, usually about 3-4 mm long, brownish to greyish. Peduncles paired,
up to 10 cm long. Pedicels 1- or 2-nate laterally, more numerous terminally, 2-3 mm
long, articulated near or below the centre. Perianth segments similar, obovate-oblong,
2 -5-3-0 mm long. Stamens not appreciably shorter than the perianth segments;
anthers about ^ mm long. Style short, shortly divided. Berry globose, 1 -seeded,
about 6 mm diameter, with a wrinkled pericarp, red.
This species is distributed from the southern Cape through east Africa and into
Asia. It climbs or scrambles to four metres, but usually only about two metres, in
bush or forest. Flowers have most frequently been recorded from March to July in
South Africa.
Cape. — Albany: Queensroad, Schouland 4357 (GRA). Alexandria: De Bega, Archibald 5908.
Bathurst: Port Alfred, Schouland s.n. (GRA). Beaufort West: Sunnyside, Esterhuysen 4359. Cathcart:
5 miles N. of Cathcart, Jessop 658. Cradock: 10 miles east of Cradock, Jessop 654. East London:
near Nahoon Mouth, Galpin 5737. Heidelberg: 5 miles W. of Heidelberg, Jessop 590. Humansdorp:
Gamtoos Gorge, Compton 23448 (NBG). King William’s Town: King William’s Town, Sim 1052.
Knysna: Blaauwkrantz, Dyer 2029 (GRA). Komga: Prospect Farm, Flanagan 299 (SAM). Ladismith:
Sandberg W. of Calitzdorp, Wurts 1371 (NBG). Middelburg: Doornberghoek, Acocks 8665. Mossel-
baai: Gouritz Valley, Horn s.n. Port Elizabeth: near Port Elizabeth, Rodin 1228 (BOL). Riversdale:
near Kweekkraal, Muir 3140. Somerset East: Cookhouse, Schlechter 2750. Uitenhage: Redhouse,
Rogers 2793 (SAM). Umtata: near Umtata, Acocks 10910. Uniondale: Kouga River, Compton
5214 (BOL). Willowmore: Willowmore, Marloth 14129.
Natal. — Estcourt: Pasture Research Station, West 364. Lions River: 5 miles along Merrivale-
Boston road. Mol! 934. Ubombo: near the turning to Mkuze Game Reserve, Jessop 669. Utrecht:
4 km from Utrecht on Knight’s Pass road, Story 1766. Weenen: Mielietuin, West 1209.
Swaziland. — Ingwavuma Poort, Compton 28579 (NBG).
Transvaal. — Lydenburg: 16 miles beyond Malipsdrif, Obermeyer & Verdoorn 22; Erasmus Pass,
Strey 3719.
Willdenow described A. racemosus from India. The South African material
differs from most specimens examined from India by the author in having well-defined
branchlets, grooved stems and straighter cladodes, but all these characters are so
variable as not to warrant even varietal status when the whole range of variation over
the range of the species is considered. For example, one specimen from Madras,
Jacob 17549 (K), differs from South African specimens only in having spines on the
final branches. Walker 56 (K) from Ceylon matches the South African plants even
more closely, as it has grooved stems, and lacks the spines on the final branches.
Bresler considered that the South African plants were specifically different, and named
them A. tetragonus. Baker reduced this species to varietal status.
73
Asparagopsis floribunda Kunth is a superfluous name for A. racemosus Willd.
Asparagopsis subquadrangularis was described as having setaceous cladodes, grey stems
and pedicels 2 mm long. The inflorescence is a raceme. Kunth cited the specimen
as “ Eckl. et Zeyh. Herb. Cap. no. 171. (v.s. in Herb. Luc.)”. According to De Candolle
(1880), the Luca specimens were at Kiel, but this specimen appears to have been
destroyed. It is very likely that this name is referable to A. racemosus, although no
authentic material has been seen.
29. A. obermeyerae Jessop, sp. nov., A. racemoso Willd. affinis, sed caulibus erectis
et canis, ramis canis, floribus axillaribus differt. Plate 5 and 6.
Caules erecti, lignosi, glabri, plus minusve recti; caules juvenes cani vel luteo-
cani sulcati, maturi fusci leves; ad fere 1 m alti. Rami caulibus similes, fasciculati,
constanter cani et sulcati, simplices vel ramis paucis brevibus, ad 15 cm longi. C/adodia
fasciculata, multa, angulata, plus minusve recta, 12-15 mm longa, c. 0-2 mm crassa,
statu sicco fuscescentes. Spinae pungentes, fere rectae, patentes, caulibus ramisque
ortae, ad 15 mm longae, cauli concolores. Pedicelli solitarii vel gemini, ramis orti,
fere vel supra medio articulati, 2-3 mm longi. Perianthii segmenta obovata vel oblongo-
obovata, 2-5-3 mm longa, margine integro, similia, margine pallido. Stamina quam
perianthii segmenta breviora; antherae c. £ mm longae. Stylus c. {- mm longus, semi-
divisus. Baccus globosus, seniine 1, c. 8 mm diam., levis, perianthio persistenti.
Type: Transvaal: Pietersburg district, near Boyne on road to Wolkberg, Mauve
4442 (PRE, holo.).
Stems erect, woody, glabrous, more or less straight, pale grey or yellowish-grey
and grooved when young, but becoming brown and smooth, up to almost 1 m high.
Branches similar to the stems, fascicled, persistently greyish and grooved, simple or
with a few short branches, up to 15 cm long. Cladodes fascicled, numerous, angled,
more or less straight, 12-15 mm long, about 0-2 mm thick, darkening with drying.
Spines pungent, almost straight, spreading, borne on branches, articulated near or
above the middle, 2-3 mm long. Perianth segments obovate or oblong-obovate, 2 • 5-3
mm long, entire, similar, with a pale margin. Stamens shorter than the perianth
segments; anthers about y mm long. Style about § mm long, divided for about half
its length. Berry globose, 1 -seeded, about 8 mm diameter, smooth, with a persistent
perianth.
This species has been recorded only from dry, bushveld slopes in the Pietersburg
district. Flowers have been collected in October.
Transvaal. — Pietersburg: Subiaco Mission, Molepo Reserve, Gerstner 5359; Naauwpoort, Breme-
kamp & Schweickerdt s.n.; Chunes Poort, Hafstrom & Acocks 197; 10 miles E. of Boyne, Mauve 4294;
Mauve 4442.
While this species most nearly resembles A. racemosus, it differs in having axillary
flowers, in its erect habit, and in the pale grey, grooved stems and branches. This
last character is shared with several species, such as A. compactus, in the section Africani.
It also resembles the Africani in having axillary flowers, but the fascicled branches
place it rather in the Racemosi.
It is named after Mrs. Mauve who has taken a particular interest in the genus,
and collected the type material of this species.
30. A. krebsianus {Kunth) Jessop, comb. nov.
Asparagopsis krebsiana Kunth, Enum. PI. 5: 93 (1850). Type: Cape, without
precise locality, Krebs s.n.
74
Tubers stalked, borne laterally on the main roots, about 30 cm Jong, fusiform.
Stems erect or scrambling, slightly zigzagging, woody, smooth or slightly ridged, grey-
green, glabrous. Branches solitary, almost always grooved, much-branched, similar
to the stems. Cladodes fascicled, three-angled, rigid, mucronate, 15-20 mm long or
sometimes only 1 cm or less, linear. Spines present on all stems and branches, pungent,
slightly recurved, browner than the stems and branches; the rest of the leaf not con-
spicuous. Peduncles 1-4-nate, with or without spines, up to 10 cm long, articulated
near the centre; the proximal half greatly thickened. Perianth segments oblong to
obovate, the outer whorl narrower, 2-5-4 mm long, white, reflexed near the centre
of each segment. Stamens nearly as long as the perianth segments. Style slightly
less than 1 mm long, divided for about half its length; ovary about 1 mm long. Berry
smooth, 1-seeded, red; perianth not persistent.
A. krebsianus may scramble among rocks or on bushes in wetter areas, but is
generally more or less erect and usually occurs on fairly dry slopes. It has been recorded
from the eastern Cape and parts of the Transvaal. Specimens, which may be referrable
to this species, have been collected in South West Africa. Flowers have been collected
in most of the summer months from October to April.
Cape. — Albany: Grahamstown, Daly & Sole s.n. ; Penther 1462 (BOL). Alexandria: Alexandria,
Archibald 5488. Cathcart: Inverthorn, Barker 3469 (NBG). Komga: Komga, Flanagan Till. Peddie:
Peddie, Barker 7821 (NBG). Victoria East: Tukulu, Acocks 11145.
Transvaal. — Brits: Jacksonstuin, Codd 6805. Krugersdorp: Baviaanspoort, Smith 351. Lyden-
burg: Driekop, Mogg & Barnard s.n. Pretoria: Pretoria Koppies, Leendertz 531; Derdepoort,
Robertson 123.
Asparagopsis krebsiana is based on a specimen collected by Krebs at the Cape.
The Krebs specimens were mainly at Kiel according to De Candolle (1820). A large
part of the Kiel Herbarium has been destroyed, and it does not now have the specimen.
Most of the plants on which Kunth worked were in Berlin, but there is no specimen
there either. The plant was described as having three-angled cladodes, spinous final
branches, racemes, and pedicels 3 mm long. As Krebs did collect in the eastern Cape
where the species described above occurs, and as A. krebsianus as construed here does
match Kunth’s description in every detail, there seems little doubt as to the correct
application of the epithet. It is proposed, should an isotype of Krebs’ specimen not
be found, that Flanagan 2377 (PRE) from Komgha be taken as the neotype. Duplicates
of this collecting have been seen by the author in the Bolus and Albany Museum
Herbaria.
The type of A. confertus Krause was Engler 6764 from Schakalskuppe in the
Luderitz district. A specimen from the Berlin-Dahlem Herbarium of this collecting
has been examined by the author, and appears to be identical with A. krebsianus
although the cladodes are a little shorter than the average, being about 10-15 mm
long. It seems likely that despite the considerable gap between this locality and the
nearest known for A. krebsianus it should be referred to this species. Other specimens
of this form are Range 1715 (B, SAM) from the type locality and Dinter 6155 (B)
from Aus.
31. A. acocksii Jessop, sp. nov., A. racemoso Willd. affinis, sed caulibus erectioribus,
levibus et pubescentibus differt. Plate 7.
Caules lignei, erecti vel scandentes, tortuosi, teretes, pubescentes, leves, cani.
Rami divaricati vel ascendentes, recti vel leviter tortuosi, plerumque solitarii sed aliquando
ad 3-nati. Cladodia subulata, plerumque ad 6 mm longa, arcuata vel recta, acuminata,
multi-nata, glabra. Spinae manifeste evolutae, leviter recurvae, omnino in caulibus
et ramulis vegetativis. Pedunculi in paribus, plerumque fere 3 cm longi, aliquando
spinosi. Pedicelli 1— 3-nati, 3-4 mm longi, fere 1 mm e tamo articulati. Perianthii
75
segmenta similia, oblongo-obovata, fere 3 mm longa, margine integra. Stamina quam
segmenta perianthii leviter breviora; antherae fere \ mm longae. Stylus brevis, semi-
divisus; ovarium fere sessile, ad fere 2 mm longum. Bacca fere 6 mm longa, semine
solitario, laevigata.
Type: Cape: Cradock district, 12 miles N.N.E. of Fish River Station, Acocks
16296 (PRE, holo.).
Stems erect to scrambling, zigzagging, woody, terete, smooth, grey, densely
pubescent. Branches spreading to ascending, straight or slightly zigzagging, usually
solitary but sometimes up to 3-nate. Cladodes subulate, usually up to 6 mm long,
arcuate or straight, acuminate, many in a bundle, glabrous. Spines well-developed,
slightly recurved, present on all stems and vegetative branches. Racemes paired,
usually about 3 cm long, sometimes bearing spines. Pedicels 1 -3-nate, 3-4 mm long,
articulated about 1 mm from the branch. Perianth segments similar, oblong-obovate,
about 3 mm long, entire. Stamens slightly shorter than the perianth segments; anthers
about mm long. Style short, divided for about half its length; ovary almost sessile,
up to about 2 mm long. Fruit about 6 mm diameter, 1-seeded, smooth.
This species does not appear to be common anywhere, but is spread over a large
part of the drier southern areas oi the Cape. In the west it is known from Namakwaland
and in the east there are single records from both Zululand and the Transvaal. Despite
this disjunct distribution, the characters are sufficiently distinct to make it very likely
that the specimens are all correctly included in this species. Flowers have been collected
from October to January and in March.
Cape. — Beaufort West: Sunnyside, Esterhuysen 5036 (BOL). Bedford: 13 miles N.N.E. of Cook-
house, Acocks 11908. Cradock: 12 miles N.N.E. of Fish River Station, Acocks 16296. Graaff-Reinet:
Wimbledon, Henrici 4964. Namakwaland: Holgat, Pillans 5178 (BOL). Port Elizabeth: Redhouse,
Paterson 978 (BOL). Uitenhage: Aloes, Drege s.n.
Natal. — Ingwavuma: Ndumu Game Reserve, Ward 2034.
Transvaal. — Lydenburg: 2 miles N. of Ohrigstad, Young A616.
This is one of the most densely pubescent species in South Africa. The normally
strongly arcuate cladodes and recurved spines are also characteristic. The nearest
species is A. racemosus from which it can be distinguished by the pubescence, the smooth
stems, and more erect habit. Another species with many similarities is A. nelsii, which
has a whitish stem, well-defined branchlets and yellowish flowers.
32. A. crassicladus Jessop, sp. nov., A. racemoso Willd. affinis, sed ratnulis minus
manifestis, cladodiis crassioribus, racemis pedicellisque brevioribus differt. Plate 8.
Caules lignei, tortuosi, teretes, plerumque pubescentes, cani vel cano-fusci, ali-
quando cortice cano exuviales. Rami divaricati vel ascendentes, leviter tortuosi, 1-
vel 3-nati. Cladodia transversa sectioni ovalia, acuminata, 12-25 mm longa, interdum
scabrida, lateraliter 1- ad 3-nata, terminaliter ad 12-nata. Spinae manifeste evolutae,
reflexae, omnino in caulibus et ramulis vegetativis. Racemi in paribus dispositi, ad
6 cm longi. Pedicel/i 2-6-nati, 3-4 mm longi, c. 1 mm e ramo articulati. Segmenta
perianthii similia, oblonga, aliquando leviter obovata, 3 mm longa, lacticolorata.
Stamina f perianthii segmentorum aequantia; antherae minutae. Ovarium 1 mm
longum, breviter stipitatum: stylus fere 0-5 mm longus, semi-divisus.
Type: Cape: Uniondale district, Kouga, Esterhuysen 4670 (BOL, holo.).
Stems woody, tortuous, terete, usually slightly pubescent, grey or brownish grey,
sometimes with the outer layers peeling. Branches spreading or ascending, slightly
tortuous, 1- or 3-nate. Cladodes oval in cross section, acuminate, 1-3-nate laterally,
up to 12-nate terminally, sometimes scabrid, 12-25 mm long. Spines well-developed,
reflexed, sometimes persistent on all vegetative stems and branches. Racemes paired,
76
up to 6 cm long. Pedicels 2-6-nate, 3-4 mm long, articulated about 1 mm from the
peduncle. Perianth segments similar, oblong, sometimes slightly obovate, 3 mm long,
cream coloured. Stamens three quarters of the length of the perianth segments; anthers
minute. Ovary shortly stalked, 1 mm long; style about half as long, divided for about
half its length. Fruit not known.
This species is a climber to about 2 m on bushes in the southern Cape from tne
Uniondale district in the west to at least as far east as the Somerset East district.
Flowers have been recorded from December to February.
Cape. — Alban> : Sushington Valley, Britten 2724 (GRA); Penrock Farm, Dyer 1193 (GRA).
Alexandria: 9 miles from Salem towards Alexandria, Jessop 609. Bathurst: Horseshoe on the Kariega
River, Jessop 617. Kirkwood: Addo Reserve, Repton 5674. Somerset East: 1 mile W. of Bruintjes-
hoogte, Jessop 645. Uniondale: Kouga, Esterhuysen 4670 (BOL); Ecklon & Zeyher s.n. (BOL).
This species is very similar in general habit to A. racemosus , but sometimes has
spines on the final branches and diffeis in the shape of the cladodes and in the length
of the peduncles and pedicels. There is little differentiation of branchlets.
33. A. nelsii Schinz in Bull. Herb. Boiss. 1, 4, app. Ill: 44 (1896); Solch, Beitr.
FI. Siidwest-Afr. 38 (1961). Type: S.W.A., ‘ Hereroland ”, Nets 20 (Z, holo.!).
Stems erect, somewhat zigzagging, smooth, sometimes slightly pubescent, grey-
white becoming brown as the outer layers disintegrate, generally up to about 50 cm
high. Branches spreading to ascending, solitary except on the older parts of the plant
where they may be paired, similar to the stems. Branchlets usually a pale yellow
to pale brown, solitary except on the older parts where they may be paired. In young
plants solitary herbaceous branchlets occur on solitary woody branches, but in the
older parts, branches and branchlets become more numerous and may become indis-
tinguishable from one another. Cladodes narrow, terete, fascicled, 10-20 mm long,
appearing after the flowers. Spines well-developed on all stems, branches, branchlets
and peduncles, spreading to recurved; the larges ones grey with brown apices; the
smaller brown. Peduncles 3-5 cm long, unbranched, or up to about 10 cm long, and
branched, pale grey to yellowish, solitary or up to about 10-nate. Pedicels articulated
once just below the centre, or sometimes twice, 2-5 mm long, up to 4-nate. Perianth
segments similar, free or fused for up to about one fifth of their length, entire, 3-4 mm
long, oblong-obovate, yellowish. Stamens slightly shorter than the perianth segments.
Style and ovary each about 1 mm long; ovules 8 or 9 per cell. Berry globose, about
5 mm diameter, 1- or rarely 2-seeded, red.
This species is widespread and fairly frequent in the dry Karriod and Kalahari
areas of the northern Cape, Bechuanaland and South West Africa. Flowers have been
recorded from August to October.
Cape. — Gordonia: 11 miles N.E. of Twee Rivieren, Leistner 2899. Hay: 8 miles E. of Postmasburg,
Acocks s.n. Kenhardt: near Pofadder, Jessop 352 (BOL). Kuruman: Tellery Pan near Kuruman
River, Leistner 2167.
Transvaal. — Soutpansberg: Hackthorne Farm, Code I 4336.
South West Africa. — Gibeon: Haribes, Volk 12188 (M). Grootfontein: Onguma, Walter 416
(M). Keetmanshoop : Gariganus, Acocks 15656. Maltahohe: Duwisib, Volk 12685 (M). Okahandja:
Okahandja, Dinter 282 (SAM). Otjiwarongo: Quickborn, Bradfield 89. Outjo: Outjo, Thorne s.n.
(SAM). Ovamboland: Oshikango, Loeb 447. Rehoboth: Rehoboth, Fleck 212 (Z). Warmbad:
Warmbad, Wandres s.n. (Z). Windhoek: Windhoek, Rogers 29773 (Z).
Bechuanaland. — Toteng, Story 4661; Ngamiland, Curson 556.
34. A. striatus (Linn, f) Thumb., Prodr. 65 (1794); Bak. in Journ. Linn. Soc. Bot.
14: 621 (1975): FI. Cap. 6: 267 (1896); Solch, Beitr. FI. Siidwest-Afr. 39 (1961).
Type: Cape of Good Hope without precise locality, Thunberg s.n. (UPS, holo.!).
77
Dracaena striata Linn, f., Suppl. 204 (1781). D. erecta Linn, f., Suppl. 204 (1781);
R. & S., Syst. Veg. 7: 346 (1829). Syntypes: Cape of Good Hope without precise
locality, Thunberg s.n. (UPS, a!, b!). D. stricta Schult. in R. & S., Syst. Veg. 7: 346
(1829), in errore D. striata.
Asparagus erectus (Linn, f.) Thunb., Prodr. 65 (1794); Bak. in Journ. Linn. Soc.
Bot. 14: 621 (1875); FI. Cap. 6: 267 (1896). A. engleri Krause in Bot. Jahrb. 51:
450 (1914). Type: S.W.A., Great Karasberg, Englsr 6651 (B, holo.!).
Myrsiphyllum striatum (Linn, f.) Kunth, Enum. PI. 5: 110 (1850). M. erectum
(Linn, f.) Kunth, Enum. PI. 5: 109 (1850).
Tubers borne laterally on the roots. Stems erect, glabrous or minutely pubescent,
woody, grooved, almost straight, greenish. Branches ascending, similar to the stems,
but often more zigzagging, not bearing branchlets. Cladodes terminal and 2- or 3-nate
and lateral and solitary in most specimens, very variable in shape from linear to almost
lanceolate, 10-40 mm long, 1-5 mm broad, rarely almost terete, very firm and rigid.
Spines often not developed, small, reflexed, hardly ever pungent: the remaining part
of the leaf with an entire margin. Peduncles in terminal fascicles varying from 1-2
or less often up to 15 or more, about 5 mm long, articulated below the centre. Perianth
segments 3-4 mm long, oblong-obovate with a serrate apex or not, similar. Stamens
not spurred, almost as long as the perianth segments; anthers about £ mm long. Style
branches very short; style about 1 mm long. Berry fleshy, about 5 mm diameter,
generally 1- or 2-seeded, red.
This species occurs in dry areas, particularly on rocky hills, over much of the Cape
Province east of Laingsburg. There are also a few records from the Orange Free
State and South West Africa. Flowers have been recorded from August to January
and in March.
Cape. — Adelaide: 3 miles from Adelaide on Bedford road, Jessop 644. Albany: Grahamstown,
Schlechter 2633 (GRA). Alexandria: Bushman’s River Poort, Galpin 2976. Aliwal North: Orange
River, 3 miles down from Aliwal North, Gerstner 62. Bathurst: Hopewell, Acocks 11128. Beaufort
West: Sunnyside, Esterhuysen 4356 (BOL). Calvinia: Calvinia, Schmidt 232. Clanwilliam: Brak-
fontein, Ecklon & Zeyher s.n. (BOL). Cradock: Cradock, Killick 796. De Aar: near De Aar,
Schweickerdt 1206. Graaff-Reinet : near Graaff-Reinet, Bolus 516 (BOL). Hay: Floradale, Esterhuysen
s.n. Humansdorp: Kabeljaauw, Burtt-Davy s.n. Kirkwood: Addo Reserve, Repton 5703. Kuruman:
Lohatlha, Esterhuysen 2414 (BOL). Laingsburg: Ngaap Kop, Compton 9242 (NBG). Middelburg:
Middelburg, Theron 454. Murraysburg: The Cave, Tyson 242 (GRA). Oudtshoorn: near Cango,
Britten 1771 (GRA). Port Elizabeth: Swartkops River, Zeyher 480. Prieska: Prieska, Bryant J330.
Queenstown: Oxton Manor, Galpin 7373. Riversdale: Riversdale, Bolus s.n. (BOL). Somerset
East: Boschberg, MacOwan 1113 (SAM). Uniondale: Kouga near Misgund, Compton 7483 (NBG).
Uitenhage: The Aloes, Drege 3164 (GRA). Victoria West: Slypfontein, Rehm s.n. Willowmore:
Willowmore, Watt & Brandwyk 2396.
Orange Free State. — Fauresmith: Botanical Reserve, Pole Evans 1794. Thaba Nchu: Thaba
’Nchu, Roberts 2945.
South West Africa. — Warmbad: Kanus, Dinter 3072 (SAM). “ Grosse Namaland ”, Fleck
306a (Z).
Thunberg did not cite a Linnaean synonym for A. striatus but there is no reasonable
doubt that the omission of Dracaena striata was accidental, as all other species collected
at the Cape by Thunberg and regarded as Dracaena species by Linnaeus are cited.
D. striata and D. erecta were described in the same work, and the epithet striatus is
preferred by the present author as it is in more common use.
Linnaeus separated these two species on D. erecta being herbaceous and D. striata
being woody, with striate cladodes. Neither of these characters has any significance
as both are woody, although often green, and have striate cladodes. Baker (1896),
who realised the worthlessness of these characters, separated his concept of these two
species on whether the flowers were solitarW (in A. erectus) or not. This character
does not seem to have any taxonomic significance.
78
Schultes cited both A. striatus and D. striata as synonyms for his D. stricta, the
spelling of which is, therefore, very likely to have been an orthographic error.
35. A. subulatus Thunb., Prodr. 66 (1794); R. & S., Syst. Veg. 7: 326 (1829);
Kunth, Enurn. PI. 5: 74 (1850); Bak. in Journ. Linn. Soc. Bot. 14: 612 (1875); FI.
Cap. 6: 262 (1896). Type: Cape of Good Hope without precise locality, Thunberg
s.n. (UPS, holo.!; BQLb PRE!, photos).
Tubers borne laterally on the roots. Stems erect, slightly zigzagging, glabrous
or minutely pubescent, woody, grooved, greenish. Branches reflexed or at least
spreading, similar to the stems, but more zigzagging. Cladodes often paired terminally
and single laterally, but sometimes up to 10 or more terminally and paired laterally,
terete, grooved, up to about 2 ■ 5 cm long and 1 mm broad, rather firm. Spines rudi-
mentary or absent, the remainder of the leaf not scarious. Peduncles in terminal
fascicles varying from 1-15, 3-5 mm long, articulated below the centre. Perianth
segments 3-4 mm long, oblong-obovate, sometimes much narrowed towards the apex,
entire, similar. Stamens not spurred, almost as long as the perianth segments; anthers
about \ mm long. Style branches rather short; style about 1 mm long. Fruit about
5 mm diameter, fleshy, 1-3-seeded; perianth persistent and conspicuous.
This species occurs mainly in dry areas, especially on rocky slopes, mainly in
the south-eastern parts of the Cape, but also in parts of the Transvaal, Natal and
Mozambique; flowers have been recorded from September to December.
Cape. — Albany: Sandy Drift near Grahamstown, Daly 70 (SAM). Alexandria: valley south of
Signal Hill, Gatpin 2923. Bathurst: West bank of Kowie, Britten 1875. Butterworth: overlooking
the Butterworth River, Pegler 2078 (BOL). Cathcart : near the Swart Kei River, 15 miles N. of Cathcart,
Roberts 1827A. East London: Chalumna causeway, Barker 6966 (NBG). Fort Beaufort : Fort Beau-
fort, Taylor s.n. (GRA). Humansdorp: Zeekoe River mouth, Fourcade 5802 (NBG). Kentani:
Kentanij Pegler 2078. King William’s Town: King William’s Town, Sim 1047 (BOL). Kirkwood:
Addo National Park, Archibald 3781. Komga: Kabousie River near Komga, Flanagan 2584 (BOL).
Peddie: 15 miles from Peddie towards East London, Barker 3969 (NBG). Port Elizabeth: Swartkops
River, Ecklon & Zeyher 250. Queenstown: Newstead below Otterburn, Acocks 12158. Uitenhage:
Aloes, Drege 3162. Umtata: near Umtata, Acocks 10971. Victoria East: Breakfast Vlei, Barker
2834 (NBG).
Natal. — Camperdown: Umgeni Valley, Forbes 1257 (NH). Lower Umfolosi: Umfolosi Game
Reserve, Ward 1665 (NH). Mahlabatini: Mahlabatini, Gerstner 2825. Msinga: 3 miles S. of Tugela
Ferry, Codd 6336. Ngotche: about -V mile from Pongola Bridge on Magudu road, Edwards 3193.
Nongoma: 11 miles N. of Nongoma, Jessop 673. Ubombo: Mkuze Game Reserve, Ward 3991.
Umzinto: Umpambinyoni River, T/iode 3430 (STE).
Transvaal. — Barberton: Budd’s Farm near Kaap Muiden, Thorncroft 3065 (NH). Lydenburg:
Lulu Mountains, Mogg 16890. Pilgrim’s Rest: 1 mile N.E. of Skukuza, Codd 5718. Soutpansberg:
Ingwe Motel, Jacobsen 1876.
The only significant variation in this species is in the number of cladodes in a
bundle. Although there are few species of Asparagus in South Africa in which both
solitary and numerous cladodes occur on different specimens, this character has been
found so variable in this species as not to be of any value taxonomically.
36. A. scandens Thunb., Prodr. 66 (1784); Bresler, Diss. 5 (1826); R. & S., Syst.
Veg. 7: 325 (1829); Bak. in Saund., Ref. Bot. t. 21 (1869); Journ. Linn. Soc. Bot.
hT 622 (1875); FI. Cap. 6: 268 (1896); Salter in FI. Cape Penins. 174 (1950).
Syntypes: Cape of Good Hope without precise locality, Thunberg s.n. (UPS, a!, b ! ;
BOL!, PRE!, photos).
Asparagus pectinatus Delile in Red., Lil. t. 407 (1813). Type: “ l’orangerie du
jardin la Malmaison ”.
Asparagopsis scandens (Thunb.) Kunth, Enum. PI. 5: 78 (1850).
79
Roots tuberous; the tubers 2-3 cm long, about 0-5 cm broad, borne on long
stalks. Stems weakly scrambling or climbing to 2 m, perennial, glabrous, green,
tortuous, not zigzagging at the nodes. Branches wide-spreading or ascending, angled.
Cladodes ternate, one conspicuously longer than the other two, spreading in the same
plane on all ultimate branches on the same branch, linear-lanceolate, falcate, with a
single prominent vein, bright green, 0-5-1 -5 cm long, 0-8-1 -6 mm broad. Leaves
much-divided, scarious, not spine-forming. Peduncles axillary, usually solitary, occas-
ionally 2-3-nate, 8-12 mm long, articulated in the distal half. Perianth segments
obovate, the outer whorl slightly smaller, the margins entire or minutely toothed,
spreading, 3-4 mm long, white or rarely pinkish. Stamens nearly as long as the perianth
segments; filaments not spurred; anthers yellow, \ of the length of the filaments.
Style 1 mm long; ovary 1 mm long, with 3 ovules per chamber. Berry globose,
1 -seeded, red.
This species grows in shady places where there is perennial moisture, from the
Cape Peninsula in the west to Zitzikama in the east. Generally it is found on coastal
mountains below about 2,000 feet. Flowers are produced irregularly, but have been
recorded in January, September and October.
Cape. — Caledon: Betty’s Bay, Topper 88 (NBG). George: Forest between Wilderness and George,
Story 3561. Humansdorp: Holman's Bos, Britten 1325 (GRA). Knysna: Blaaukrantz Pass, Galpin
4738. Paarl: Franschhoek, Phillips 1345 (SAM). Riversdale: Grootvadersbos, Zeyher 8570b (STE).
Stellenbosch: Jonkershoek, Rodin 3233. Swellendam: Strawberry Hill, Van der Merwe s.n. (STE).
Tulbagh: Tulbaghkloof, Ecklon & Zeyher s.n. Wynberg: Devils Peak, Wolley Dod 1739 (BOL).
37. A. ramosissimus Bak. in Gard. Chron. n. s., 2: 6 (1874); Journ. Linn. Soc.
Bot. 14: 622 (1875); FI. Cap. 6: 268 (1896). Type: “Cooper in Hort. Saunders
5102”.
A. scandens Thunb. var. deflexus Bak. in Journ. Linn. Soc. Bot. 14: 622 (1875);
FI. Cap. 6: 268 (1896). Type: Somerset East, Boschberg, MacOwan 1771 (BOL!,
SAM!).
Stems weak and scrambling or climbing, up to 1 m long, probably usually perennial,
herbaceous. Stems and branches grooved, glabrous, slightly zigzagging, greenish.
Ultimate branches reflexed at the base, similar to the other branches. Cladodes angled
or flattened, linear, 3-nate, 5-15 mm long. Leaves not spine-forming. Peduncles
solitary, 2-10 mm long, articulated distally. Perianth segments similar, entire, obovate,
about 3 mm long, spreading, white. Stamens nearly as long as the perianth segments;
filaments not spurred; anthers up to 1 mm long. Style sometimes at least, unbranched
about 1 mm long; ovary about 1-5 mm long. Berry globose, fleshy, up to nearly,
1 cm diameter, 1- or 2-seeded, red.
This species occurs in moist shady places from the Worcester district through the
southern Cape and Natal up to the eastern Transvaal. Flowers have been recorded
from September to January and in June.
Cape. — Heidelberg: between Port Beaufort and Riversdale, Lewis 5941 (NBG). George: Cradock-
berg, Esterhuysen 30168. Keiskammahoek : Cala Forest Reserve, Story 3269 (GRA). King William’s
Town: Pirie, Flanagan 2235. Mount Ayliff: Insiswa Mountains, Schlechter 6439. Mount Currie:
Mount Currie, Stephany 33 (GRA). Riversdale: Kampscheberg, Muir 2977 (BOL). Somerset East:
Boschberg, MacOwan 1771 (SAM). Stockenstrom: Hogsback, Barker 1901. Stutterheim: Fort
Cunynghame, Galpin 2439. Swellendam: Tradouw Pass, Walgate 914 (NBG). Victoria East: Victoria
East, Rattray 338. Worcester: Welgevonden, Esterhuysen 1902 (BOL).
Basutoland. — Leribe, Pitzeng, Dieterlen 707.
Natal. — Bergville: Ndedena, Esterhuysen 28502 (BOL). Estcourt: Cathkin Park, Howlett 47.
Orange Free State. — Bethlehem: Golden Gate National Park, Roberts 3229.
Swaziland. — N.W. of Mbabane, Compton 30552.
80
Transvaal. — Barberton: Ivy Range, Moodies, Thorncroft 338. Ermelo: Spitskop, Pott 15118.
Pilgrim’s Rest: Graskop, Ga/pin 14531. Pretoria: Wonderboomspruit, Mogg 14754. Wakkerstroom:
Oshoek, Devenish 206.
Baker described A. ramosissimus as “ Wide climbing, leaves obscurely spurred at
the base, cladodes 3- to 8-nate, flattened, linear, falcate, acute, -j- to f inches long,
spreading. Flowers solitary at the tips of the branchlets There are two characters
in this description which are rare in the specimens examined by the present author.
These are the 3- to 8 -nate cladodes and the terminal flowers. Both of these characters
do occur, however, and there is no other species which fits this description. In 1875,
Baker cited “ Cooper in Flort. Saunders 5102 ”, but added that the description was
made from a figure by Fitch. This drawing does not appear to have been published,
and Moss (Ms. in National Herbarium, Pretoria) could not find a specimen of the
species among the specimens in Cooper’s complete set at Kew.
In 1896, Baker wrote that his description of A. ramosissimus was from a living
plant in Wilson’s garden at Reigate. It is likely that no specimen has been preserved,
and no other specimen was cited in any of Baker’s works.
38. A. crispus Lam., Encyc. 1: 295 (1783); Bresler Diss. 22 (1826); R. & S.
Syst. Veg. 7: 326 (1829); Kunth, Enum. PI. 5: 73 (1850); Bak. in Journ. Linn. Soc.,
Bot. 1 4 : 607 (1875); FI. Cap. 6: 259 (1896); Marloth, FI. S. Afr. pi. 20(1915); Salter
in FI. Cape Penins. 174 (1950). Type: “ L’lsle de France”, collector unknown, (P,
holo.; PRE, photo.!).
A. flexuosus Thunb., Prodr. 66 (1794); Kunth, Enum. PI. 5: 74 (1850). Type:
Cape of Good Hope without precise locality, Thunberg s.n. (UPS, holo.!; BOL!,
PRE!, photos). A. decumbens Jacq., Hort. Schoenbr. 1: 51, t. 97 (1797); Bresler,
Diss. 21 (1826). Type: Cape of Good Hope without precise locality, collector unknown.
Asparagopsis decumbens (Jacq.) Kunth, Enum. PI. 5: 77 (1850).
Roots tuberous; the tubers borne on short stalks, 3-6 cm long, 0-5-1 -5 cm
broad. Stems annual, weak, straggling or climbing, usually up to 1 m long, much-
branched, angled, zigzagging at the nodes, glabrous, green. Branches reflexed at the
base, curving upwards, similar to the stems. Cladodes ascending, ternate, except
where a normal branch is formed when they are paired, borne at nodes towards the
ends of branches, soft or slightly firm, slightly arcuate, flattened or angled, 3-9 mm long,
less than 1 mm broad and usually less than \ mm broad. Leaves frequently forming
soft spine-shaped processes, grey. Peduncles solitary, axillary, 6-1 1 mm long, articulated
distally. Buds cylindrical. Perianth segments reflexed from their centres, but forming
a cylinder proximally; the base of the flower truncate; segments similar, entire, linear-
oblong, whitish with a pale brown or green streak, 4-5 mm long. Stamens about
3 mm long, orange; filaments broadened towards their bases, and with minute spurs.
Style 1-5 mm long; ovary 1-5 mm long, with 2-4 ovules in each chamber. Berry
ovoid, 3-9-seeded, 8-15 mm long, pale coloured; style persistent and conspicuous;
perianth persistent.
This species grows in a wide range of habitats from coastal sands to rich mountain
soils, but has a rather restricted distribution. It occurs in the south-western Cape from
St. Helena in the north, round the coast as far as Swellendam. It also grows as far
inland as Worcester. Flowers have been recorded from June to October.
Cape. — Caledon: Van Niekerk 739 (BOL). Cape: Robben Island, Walgate 494 (NBG); Signal
Hill, Dimmer 1600 (SAM). Clanwilliam: between the Olifant’s River and Brakfontein, Ecklon &
Zeyher s.n. Humansdorp: mouth of Klein River, Zeyher 41 55 (BOL). Malmesbury: Darling Flora
Reserve, Rycroft 1931 (NBG). Paarl: Tigerberg, Pillans 8657 (BOL). Robertson: Robertson, Van
Niekerk 383. Simonstown: Lakeside Plateau, Jessop 62 (BOL). Somerset West: Knorhoek, Jessop
61 (BOL). Stellenbosch: Jonkershoek, Duthie 527 (STE). Tulbagh: Saron, Schlechter 7892 (GRA).
Worcester: Karroo Garden, Leighton 1037 (NBG). Wynberg: Kirstenbosch, Salter 7356 (BOL).
81
Lamarck described this species from a specimen in the Lamarck Herbarium. This
specimen is described as having been collected on Mauritius. The only Mauritian
species is A. umbellu/atus Bresler, which differs in having fascicled cladodes, flowers
not solitary, peduncles articulated near the centre and a globose fruit. The locality
must, therefore, either be wrongly given or the plant must have been cultivated.
The illustration by Jacquin includes an accurate drawing of the flower, which
shows the truncate base and cylindrical proximal part characterising A. crispus Lam.
39. A. asparagoides (L.) Wight in Century Dictionary II: 845 (1909); Salter in
FI. Cape Penins. 173 (1950). Type: Tilli, Cat. Plant Horti Pisani t. 12, f. 1,2 (1723).
Medeola asparagoides L., Sp. PI. 339 (1753); Delile in Red., Lil. t. 442 (1816). M.
angustifolia Mill., Gard. Diet. ed. 8 (1768). Type: Tilli, Cat. Plant. Horti Pisani t. 12
“ myrti folio angustiore ” (1723).
Dracaena medeoloides Linn, f., Suppl. 203 (1781). Type: Cape of Good Hope
without precise locality, Thunberg s.n. (UPS, holo.!; BOL!, PRE!, photos). D. vo/u-
bilis Linn, f., Suppl. 204 (1781). Type: Cape of Good Hope without precise locality,
Thunberg s.n. (UPS, holo.!; BOL!, PRE!, photos).
Asparagus volubilis (Linn, f.) Thunb., Prodr. 66 (1794); Bak. in Journ. Linn. Soc.
Bot. 14: 628 (1875); FI. Cap. 6: 272 (1896). A. medeoloides (Linn, f.) Thunb., Prodr.
66 (1794); Marloth, FI. S. Afr. 4: pi. 20 (1915). A. krausii Bak. in Journ. Linn. Soc.
Bot. 14: 628 (1875); FI. Cap. 6: 272 (1896), nom illegit. Type: as for Myrsiphyllum
krausianum Kunth. A. krausianus (Kunth) MacBride in Contrib. Gray Herb. 56:
17 (1794); Salter in FI. Cape Penins. 174 (1950), sub A. kraussianus. A. kuisibensis
Dinter in Fedde, Rep. 29: 270 (1931). Type: S.W.A., Kuisib, Tjuezu in Herb. Dinter
4698 (B, holo.!). A. ovatus Salter in Journ. S. Afr. Bot. 6: 167 (1940); FI. Cape
Penins. 174 (1951); Dyer in Flow. PI. Afr. 29: t. 1146 (1943). Type: Cape, Rugby,
Salter 8214 (BOL, holo!.; NBG!, PRE!). A. multitub erosus Dyer in Bothalia 6: 442
(1954). Type: Ceres, Karoopoort, Marloth 9006 (PRE, holo.!; STE!).
Ruscus volubilis Thunb., Prodr. 13 (1794); Kunth, Enum. PL 5: 276 (1950). Type:
Cape of Good Hope without precise locality, Thunberg s.n. (UPS, holo.!; BOL!,
PRE!, photos).
Myrsiphyllum asparagoides (L.) Willd. in Ges. Naturf. Fr. Berl. Mag. 2: 25 (1808);
Kunth, Enum. PI. 5: 105 (1850); Hook. f. in Bot. Mag. t. 5584 (1866). M. angusti-
folium (Mill.) Willd. in Ges. Naturf. Fr. Berl. Mag. 2: 25 (1808); Kunth, Enum. PI.
5: 106 (1850). M. falciforme Kunth, l.c., 107 (1850). Type: Cape without precise
locality, Drege 2704a in “Herb. Luc.” (K, iso.!). M. gramineum Kunth, l.c., 108
(1850). Type: “ M. striatum Schlechtend. in Herb. reg. Berol., excl. synon. (v..s\) etl
Hort. Berol. 1847 ”. M. krausianum Kunth, l.c., 107 (1850). Type: Wynberg, Con-
stants, Krauss 1333.
Hecatris asparagoides (L.) Salisb., Gen. Plant. 66 (1866).
Tubers varying from almost sessile to 7 cm distant, usually 5-15 cm long. Stems
climbing to 3 m, scrambling or under exceptional circumstances erect, tortuous, glabrous,
terete or angled, annual. Branches similar to the stems, spreading or ascending.
Cladodes broadly ovate to lanceolate, with an obtuse base, sessile, 1-7 cm long, up
to about 7 prominent veins and numerous smaller ones. Leaves scarious, not forming
spines. Peduncles 1- or 2-nate, 3-8 mm long, articulated near the flower. Perianth
segments similar, entire, linear-oblong, 5-6 mm long, spreading at first, later reflexed
from the centre of each segment, white. Stamens nearly as long as the perianth
segments; anthers 0-5 mm long; filaments with two basal spurs. O vary 2 mm long;
style 2-3 mm long, the branches sometimes only loosely fused; ovary chambers with
4-9 ovules. Berry globose, 6-10 mm diameter, up to 8-seeded, red; perianth persistent.
82
It is distributed from the south-western Cape through the eastern parts of South
Africa and north to Kenya and the Congo. The flowering period is mainly June to
September, but varies according to the time of the rainfall. In the Transvaal, flowering
usually occurs in January and February.
Cape. — Albany: Frasers Camp, Maguire 636 (NBG). Alexandria: near Sandflats, Burtt Davy
14249 (BOL). Barkly East: Barkly Pass, Rattray 7324. Bathurst: Kowie River, MacOwan 1920
(SAM). Bellville: De Grendel, Esterhuysen 23057 (BOL). Bredasdorp: Strandkloof, Compton 14778
(NBG). Caledon: Uilenkraal, Taylor 1 584 (SAM). Calvinia: near Middelpost, Salter 3484 (BOL).
Cape: near Rugby, Salter 8214. Ceres: Karroopoort, Mar/oth 9006. Clanwilliam: Brakfontein,
Schlechter 7974. Colesberg: Arundel, Acocks 17992. East London: near Cove Rock, Galpin 3102,
Fort Beaufort : near Fort Beaufort, Story 2226. George: Wilderness, Van Niekerk 251 (BOL). Graaff-
Reinet: Graaff-Reinet, Bolus 463 (BOL). Humansdorp: Ratelsbos, Fourcade 53 (BOL). Kentani.
Kentani, Pegler 1240 (BOL). King William’s Town: Amabele, De Vries 23. Kirkwood: Addo
Bush, Mar loth 6439. Knysna: Knysna Village, Fourcade 5885 (BOL). Komga: Prospect Farm,
Flanagan 315. Laingsburg: White Hill, Compton 11200 (NBG). Malmesbury: Darling Flora Reserve,
Rv croft 1969 (NBG). Montagu: Montagu, Levyns s.n. (BOL). Namakwaland: Silverfontein near
O’okiep, Drege 2704b (K). Peddie: near Wooldridge, Acocks 1 1886. Piketberg: Avontuur Mountain,
Pillans 7571 (BOL). Port Elizabeth: Perseverence, Paterson 67 (GRA). Queenstown: Invana,
Cooper 328 (BOL). Riversdale: Vet Rivier, Muir 353. Simonstown: Smitswinkel, Wolley Dod 3009
(BOL). Somerset East : Boschberg, MacOwan s.n. (SAM). Somerset West: Fir Grove, Parker 4485
(BOL). Stellenbosch: near Mill Stream, Duthie 1479 (STE). Stockenstrom : Andriesberg, Galpin
2536 (GRA). Tulbagh: Saron, Schlechter 4876 (GRA). Uitenhage: Enon, Thocle A2111. Uniondale:
near Laudina Store, Acocks 14643. Vanrhynsdorp: Giftberg, Phillips 7619 (SAM). Wellington:
Wellington, Thomson s.n. Worcester: De Dooms, Geriche s.n. Wynberg: Kirstenbosch, Salter
7596 (BOL).
Basutoland. — Leribe, Dieterlen 102.
Natal. — Bergville: Cathedral Peak Forest Reserve, Killick 1723. Camperdown: Mid Illovo:
Thode 3432 (STE). Eshowe: Eshowe, Gerstner 3754 (NH). Inanda: Inanda, Medley Wood 623
(NH). Kranskop: Kranskop, Thode 3419 (STE). Lions River: Dargle Forest, Mol! 1233. Ngotche,
Ngome Forest, Gerstner 4396. Pietermaritzburg: Allerton, Mogg 6620.
Orange Free State. — Bethlehem: Clarens, Van Hoepen s.n. Bloemfontein: Glen, Mostert 363.
Lindley: near Lindley, Phillips s.n. (BOL).
Swaziland. — Near Hlatikulu, Compton 26393 (NBG).
Transvaal. — Barberton: Umvoti Creek, Galpin 857. Belfast: near Machadadorp, Bruce 481.
Ermelo: Ermelo, Rogers 14782 (BOL). Johannesburg: Jeppe, Gilmore 520. Krugersdorp: near
Krugersdorp, Mogg 23103. Letaba: Duiwelskloof, Scheepers s.n. Lichtenburg: Grasfontein, Sutton
581. Middelburg: Boschhoek, western Steenkampsberg, Young A384 (NH). Petersburg: Blaauw-
berg, Dyer 9152. Pilgrim’s Rest: Mariepskop, Van der Schijjf 4784. Potchefstroom : Nooitgedacht,
Cohen s.n. Wakkerstroom : Wakkerstroom, Beeton 124 (SAM).
South West Africa. — Grootfontein: Otavibirge, Dinter 743 (SAM); Otavi, D inter 5547 (B).
Linnaeus, in his Species Plantarum, cited M. A. Tilli’s “ Catologus Plantarum
Horti Pisani ” (1723) t.12, f.l & 2. There is no specimen labelled Medeola asparagoides
in the Linnaean Herbarium so that Tilli’s figure, which shows specimens with solitary,
many-veined cladodes and tortuous stems should be regarded as the iconotype, as
this was the only illustration cited.
Miller regarded the specimen in Tilli’s plate with larger cladodes as the type of
Linnaeus’ species, and coined the name Medeola angustifolia for the form with smaller
cladodes shown in the other figure on the same plate.
Dracaena volubilis Linn. f. was based on a Thunberg specimen from the Cape.
The specimen in the Linnaean Herbarium, a photograph of which has been examined
by the author, has ternate cladodes as opposed to the alternate cladodes mentioned
in the type description and according to Jackson (1912), it does not occur in the
enumerations of the Linnaean Herbarium. It can, therefore, not be considered as
a type. The sheet in the Thunberg Herbarium labelled Asparagus volubilis has three
specimens which agree with the younger Linnaeus’ description of Dracaena volubilis
and with the present author’s concept of Asparagus asparagoides. Thunberg's
A. volubilis was described in the same words as Dracaena volubilis Linn, f., and
83
reference was made to “ Linnaeus Syst. 334 The only reference to this species earlier
than Thunberg’s Prodromus appears to be the Supplementum of the younger Linneaus.
Stearn (1961) mentions that both Linnaeus the younger and Thunberg played a part
in adding to the elder Linnaeus’ manuscript of this work. It is therefore very likely
that “ Linnaeus Syst. 334 ” refers to this manuscript.
15 20 50 40 50 70 Km
CLADO&E LENGTH
Fig. 3. — Scatter diagram and histograms showing cladode length and breadth, and the relation of
these characters to one another, in Asparagus asparagoides
84
The holotype of Myrsiphyllum falciforme was cited by Kunth as being in “ Herb.
Luc.”. This specimen was, therefore, likely to have been in Kiel, but is not there now,
and has probably been destroyed. The specimens cited by Kunth under his M.
gramineum are all likely to have been destroyed in Berlin. He described this species
as a climber with flattened, solitary cladodes, and there can, therefore, be no reasonable
doubt as to the correct application of the name. Krauss 1333, which is the type of
M. kraussianum, has not been traced. Kunth’s description, however, clearly places
it in the synonymy of this species. He spelt the collector’s name Krause, but the correct
spelling is Krauss, and the correct spelling of the binomial should have been M.
kraussianum.
There is considerable variation in the size and shape of the cladodes of A. aspara-
goides as construed by the present author. Both A. krausianus and A. volubilis were
described from specimens with lanceolate cladodes, while A. ovatus and A. multi-
tuberosus were described from specimens with ovate cladodes. A. asparagoides appears
to have been described from an illustration showing cladodes of an ovate-lanceolate
foim in the left hand figure and lanceolate cladodes in the right hand figure. As A.
kraussianus sensu Salter and A. asparagoides sensu Salter (1950) are frequently found
in similar environments and sometimes near one another, it seems likely that their
difference is genotypic, and not merely the effect of different habitats. In order to
determine the taxonomic value of cladode shape in this group in South Africa, the
scatter diagram and histograms (Fig. 3) were drawn. These show that there is no
justification for subdivision of A. asparagoides , as construed here, on the basis of cladode
shape or size. In the eastern Cape there is a narrow-cladode form with a strong garlic-
like smell. Without any other subsidiary character, however, it was felt that this form
could not be given taxonomic recognition.
The other important character which has been used for the separation of species
in this complex is the tuber arrangement. The tubers of A. ovatus were described as
distant as opposed to those of A. asparagoides sensu Salter (1950), which were sessile.
The tuber character of A. ovatus may be associated with the loose sand in which the
plants occur. The typical form of A. asparagoides with sessile tubers generally occurs
in soils of a heavier texture. Parker collected specimens which he claimed had both
sessile and distant tubers and, in the experience of the present author, it is, often
impossible to define the difference between sessile and distant tubers as there is a complete
gradation between these two forms. There is also a gradation between the rooting
system in A. multitub erosus, where the tubers are very numerous and only about 5 cm
long, and that in A. asparagoides sensu Salter in which there are about twelve tubers,
each about 10-15 cm long. The specimens which have been placed in A. multitub erosus
come from dry areas with very hard soils such as occur on the Giftberg near Van-
rhynsdorp.
40. A. undulatus {Linn. /.) Thunb., Prodr. 66 (1794); Bak. in Journ. Linn. Soc.
Bot. 14: 628 (1875); FI. Cap. 6: 273 (1896); Marloth, FI. S. Afr. 4: pi. 20 (1915);
Salter in FI. Cape Penins. 173 (1950); Solch, Beitr. FI. Siidwest-Afr. 39 (1961). Type:
Cape of Good Hope without precise locality, Thunberg s.n. (UPS, holo.!; BOL!,
PRE!, photos).
Dracaena undulata Linn, f., Suppl. 203 (1781); R. & S., Syst. Veg. 7: 346 (1829).
Myrsiphyllum undulatum (Linn, f.) Kunth, Enum. PI. 5: 109 (1850).
Asparagus klinghardtianus Dinter in Fedde, Rep. 29: 270 (1931). Type: S.W.A.,
Klinghardtgebirge, Dinter 3998 (B, holo.!; PRE!).
Tubers distant, 0-8-15 mm broad, 25-50 mm long. Stems erect, frequently with
minute obtuse processes giving them a rough surface, annual, up to 60 cm high, but
usually less, angled, the edges often produced into wings in the upper parts, green.
85
Branches similar to the stems, ascending. Cladodes solitary, sessile, ovate to lanceolate,
undulate and often folded towards the tips, acutely attenuate, with prominent veins
which are sometimes produced into longitudinal wings or ridges, scabrid, 8-40 mm
long. Leaves deltoid, not forming spines. Peduncles 1-3-nate, 5-10 mm long, articu-
lated in the distal third. Perianth segments similar, entire, linear-oblong, the distal
half spreading or reflexed, 6 mm long, purplish or greenish. Stamens almost as long
as the perianth segments; filaments spurred; anthers 0-5 mm long. Ovary 2 mm long;
styles 2-5 mm long. Berry 5-6 mm diameter, 1-3-seeded, red; perianth persistent.
This species is particularly abundant on the flat sandy area stretching from the
Cape Peninsula north to near Citrusdal. It also occurs in the drier areas of Nama-
kwaland and S.W.A Flowers have been recorded from July to October.
Cape. — Bellville: Tigerberg, Pillans 8660 (BOL). Calvinia: Glen Lyon, Lewis 5726 (NBG). Cape:
Signal Hill, Thode 6079 (STE). Clanwilliam: Clanwilliam, Schlechter 8416. Hopefield: near Hopefield,
Bolus 12858 (BOL). Malmesbury: near Groenkloof, Bolus 3690 (Z). Namakwaland: Witbank,
Pillans 5222 (BOL). Piketberg: south of Piketberg, Barker 5775 (NBG). Stellenbosch: near Libertas,
Duthie 542 (STE). Tulbagh: Tulbagh New Kloof, Compton 11708 (NBG).
South West Africa. — Grootfontein: Otavi, Dinter 5547 (Z). Luderitz: Klinghardtgebirge,
Dinter 3998. “Ex reg. : Hereroland ”, Dinter 623 (Z).
There is a certain amount of variation in cladode breath of specimens from different
areas. For example, the specimens from north of Citrusdal have particularly narrow
cladodes. There is, however, a complete gradation between the different cladode
shapes.
Doubtful Species
A. confertus Krause in Bot. Jahrb. 51: 449 (1914); Solch, Beitr. FI. Siidwest-Afr.
40 (1961). Type: Luderitz, Schakalskuppe, Engler 6764 (B, holo.!).
Material matching the type shows many similarities with A. krebsianus, but differs
mainly in having shorter cladodes, which are about 10-15 mm long. These specimens
have been collected only in the Luderitz district of South West Africa. A. krebsianus
has not been recorded closer to this than the central districts of the Transvaal. In
view of the paucity of material, the smaller cladodes and the considerable gap between
this population and A. krebsianus, it is considered best not to refer these specimens
to that species. Similarly it does not seem justified to uphold A. confertus without
further material.
South West Africa. — Luderitz: Aus, Dinter 6155 (B); Schakalskuppe, Engler 6764 (B); Range
1715 (B, SAM).
A. declinatus L., Sp. PI. 313 (1753). Type: No specimen or earlier description
cited.
As there is no specimen in the Linnaean Herbarium, as Linnaeus did not cite any
earlier work on which this binomial could have been based, and as the description,
which does not even give the geographical area from which the plant came, is com-
pletely inadequate for identifying the species, the name must be regarded as a nomen
dubium. Baker (1875) used the name for a form which is treated here as A. setaceus,
and A. declinatus sensu Baker is therefore discussed under that species.
A. nivenianus Schult. in R. & S., Syst. Veg. 7: 331 (1829). Type: Cape without
precise locality, Niven s.n.
The type specimen of this binomial has not been located. It may have been
destroyed in Berlin. From the description it is not possible to tell whether the Niven
specimen should be referred to A. africanus or A. thunbergianus, and the name must
therefore be regarded as a nomen dubium. Kunth, when he made the combination
in Asparagopsis, cited Ecklon 95b from the Cape. This specimen is referable to
A. thunbergianus.
86
References
ARBER, A., 1924. Myrsiphyllum and Asparagus: a morphological study. Ann. Bot. 38: 635-659
ARBER, A., 1925. Monocotyledons: a morphological study. London: Cambridge University
Press.
BAKER, J. G., 1875. A Revision of the Genera and Species of Asparagaceae. Journ. Linn. Soc.
Bot. 14: 508-632.
BAKER, J. G., 1896. Asparagus in: FI. Cap. 6: 256-274. London: L. Reeve & Co.
BRESLER, M., 1826. Generis Asparagi historia naturalis atque medici. Berlin.
DE CANDOLLE, A., 1880. La Phytographie ou l’art de decrire les Vegetaux. Paris: G. Masson,
Editeur Libraire de FAcademie de Medecine.
JACKSON, B. D., 1912. Index to the Linnaean Flerbarium. Suppl. to Proc. Linn. Soc. Lond.,
124th session.
KAUSSMANN, B., 1955. Histogentische Untersuchungen zum Flachsprossproblem. Bot. Studien 3,
KIES, P., 1951. Notes on Asparagus africanus and related Species. Bothalia 6: 177-181.
KUNTH, K. S., 1850. Enumeratio Plantarum 5. Stuttgart and Tubingen: J. G. Cotta.
LANJOUW, J. & STAFLEU, F. A., 1954. Index Herbariorum, pt. II, Collectors, 1st instalment.
LINNAEUS, C., 1753. Species Plantarum, ed. 1. Stockholm: Laurentius Salvius.
PHILLIPS, E. P., 1951. Genera of South African Flowering Plants. Pretoria: Government Printer.
SALTER, T. M., 1950. Asparagus in: FI. Cape Penins. CaDe Town and Johannesburg: Juta &
Co., Ltd.
SCHLITTLER, J., 1953. Bliitenartikulation und Phyllokladien der Liliaceae organophylogenetische
betrachtet. Fedde, Rep. 55: 154-258.
SCHLITTLER, J., 1959. Die Asparageenphylokladien erweisen sich auch ontogenetisch als Blatter.
Bot. Jahrb. 79: 428-446.
SOLCH, A., 1961. Beitrage zu einer Flora Stidwest-Afrikas. Munich: A. Solch.
STEARN, W. T., 1961. In: Cat. Books R. M. M. Hunt 2, 1: civ.
ROEMER, J. J. & SCHULTES, J. A., 1829. C. Linne, Systema Vegetabilium 7. Stuttgart: J. G.
Cotta.
TRIMEN, H., 1898. Handbook of the Flora of Ceylon. London: Dulau & Co.
Index
Asparagopsis aethiopica (L.) Kunth, 68
albus (L.) Kunth, 39
consanguinea Kunth, 54
decumbens (Jacq.) Kunth, 80
densiflora Kunth, 65
denudata Kunth, 52
dregii Kunth, 55
floribimda Kunth, 72
juniperina Kunth, 48
krebsiana Kunth, 73
lamarckii Kunth, 48
lancea (Thunb.) Kunth, 68
microrhaphis Kunth, 61
minutiflora Kunth, 63
niveniana sensu Kunth, 55, 85
passerinoides Kunth, 43
retrofracta (L.) Kunth, 58
scandens (Thunb.) Kunth, 78
schlechtendalii Kunth, 48
setacea Kunth, 51
spinescens (Schult.) Kunth, 46
stipulacea (Lam.) Kunth, 45
subquadrangularis Kunth, 72
thunbergii Kunth, 55
triancantha (Willd.) Kunth, 46
zeyheri Kunth, 46
Asparagus acocksii Jessop, 74, pi. 7
aethiopicus L., 67, 71
TlllUpiLLlb U/, / I
var. aethiopicus, fig. 1, 68, 70
var. angusticladus Jessop, 69, pi. 3
var. natalensis Bak., 70
var. ternifolius Bak., 69, 70
africanus Lam., 33, 48, 85
var. concinnus Bak., 60
var. deoendens (thunb.) Bak., 51
var. microphylhis Bak., 58
var. wright ii Bak., 59
albus sensu Thunb., 55
albus L., 55
asiaticus sensu Bak., 48
asiaticus L., 50
asparagoides ( L .) Wight, 33, 35, 39, 81
aspergillus Jessop, 71 pi. 4
bechuanicus Bak. 48
buchananii Bak. 67
burchellii Bak., 45
burkei Bak., 48
capensis L., 31, 33, 39, 43, 45
capensis sensu Marloth, 55
var. capensis, 43
var. litoralis Suess. & J. Karl., 44
compactus Salter, 33, fig. 1, 59, 73
87
concinnus Kies, 60
confertus Krause, 74, 85
conglomerate Bak., 48
consanguineus (Kunth) Bak., 54
cooper i Bak., 48
crassicladus Jessop, 75, pi. 8
crispus Lam., 33, 80
declinatus sensu Bak., 51, 57
declinatus L., 85
decumbens Jacq., 80
densiflorus (Kunth) Jessop, 39, 65
densus Bak., 45
denudatus (Kunth) Bak., 33, fig. 1, 52
dependens Thunb., 48
dinteri Engler & Krause, 48
ecklonii Bak., 62
engleri Krause, 77
erectus (Linn, f.) Thunb., 77
exuvialis Burch., 33, 35, 37, 62
falcatus L., 66, 69
var. falcatus, 70
var. ternifolius (Bak.) Jessop, 70
fasciculatus Thunb., 37, 39, 54
fleckii Schinz, 48
flexuosus Thunb., 80
francisci Krause, 48
glaucus Kies, 35, 47
gonocladus Bak., 66
hereroensis Schinz, 62
intangibilis Dinter, 46
judtii Schinz, 48
juniperoides Engler, 35, 38, 63
klinghardtianus Dinter, 84
krausii Bak., 81
krausianus (Kunth) MacBride, 81
krebsianus (Kunth) Jessop, 39, 73, 85
kuisibensis Dinter, 81
lanceus Thinb., 68
laricinus Burch., 59, 60
lugardii, Bak., 48
macowanii Bak., 57
var. macowanii, 57
var. zuluensis (N. E. Br.) Jessop, 58, 61
medeoloides (Linn, f.) Thunb., 81
meyeri nom. nud., 66
microphyllus (Bak.) Kies, 58
microrhaphis (Kunth) Bak., 61
minutiflorus (Kunth) Bak., 33, 63, 64, 66
mucronatus Jessop, 56, pi. 1
multiflorus Bak., 48
multituber osus Dyer, 81
myersii nom. nud., 65
myriocladus Bak., 65
namaensis Schinz, 52
nelsii Schinz, 33, 75, 76
nelsonii Bak., 43
nivenianus Schult., 85
obermeyerae Jessop, 73, pi. 5, 6
officinalis L., 37, 39
omahekensis Krause, 46
ovatus Salter, 81
oxyacanthus Bak., 63
patens Krause, 48
pearsonii Kies, 52
pectinatus Redoute, 78
pi /osus Bak., 48
plumose Bak., 51
racemose sensu Solch, 71
racemosus Willd., 68, 71, 72, 73, 74, 75
var. tetragonus (Bresler) Bak., 72
ramosissimus Bak., 79
retrofractus, L., 33, 39, 58, 61
rigidus Jessop, 33, 64, 66, pi. 2
rivalis Kies, 48
sarmentosus sensu Bak., 65
sarmentosus L., 66
saundersiae Bak., 66, 67
scandens Thunb., 33, 78
var. de fie xus Bak., 79
setaceus (Kunth), Jessop, 33, 51, 85
spinescens Schult., 45
spinosissimus O. Kuntze, 46
sprengeri Regel, 65
sprengeri sensu Wright, 69
stellate Bak., 61
stipulaceus Lam., 33, 39, 44
stipulaceus sensu Bak., 46
striatus (Linn, f.) Thunb., 35,^76
suaveolens Burch., 45, 48
subulatus Thunb., 78
sylvaticus Burch. Ms., 54
ternifolius sensu Hooker f., 69
tetragonus Bresler, 72
thunbergianus Schult., 39, 55, 57, 85
triacanthus Burm. f., 43
triacanthus Willd., 45
umbellulatus Bresler, 81
undulatus (Linn.f.) Thunb. , 35, 84
virgatus Bale., 53
volubilis (Linn, f.) Thunb., 81
zuluensis N. E. Br., 58
Dracaena erecta Linn, f., 77
medeoloides Linn, f., 81
striata Linn, f., 77
stricta Schult., 77
undulata Linn, f., 84
volubilis Linn, f., 81
Hecatris asparagoides (L.) Salisb., 39, 81
Medeola angustifolia Miller, 81
asparagoides L., 39, 81
Myrsiphyllum angustifolium (Miller) Willd., 81
asparagoides (L.) Willd., 39, 81
erectum (Linn, f.) Kunth, 77
falciforme Kunth, 81
gramineum Kunth, 81
krausianum Kunth, 81
striatum (Linn, f.) Kunth, 77
undulatum (Linn, f.) Kunth, 84
Ruscus volubilis Thunb., 81
x
89
Plate I. — Asparagus mucronatus Jessop (Theron 551, holotype).
7230691
90
Plate 2. — Asparagus rigidus Jessop (Kil/ick & Strey 2380, holotype).
91
/5 /.
'f/'C ‘
r*fort 4
Plate 3. — Asparagus aethiopieus L. var. angusticladus Jessop (Mogg 22470, holotvpe).
92
Plate 4. — Asparagus aspergillus Jessop ( Breyer in TRV 19063, holotype).
93
Plate 5. — Asparagus obermeyerae Jessop ( Mauve 4442, holotype).
94
Plate 6.
Asparagus- obermeyerae Jessop ( Hafstrom & Acocks\91).
95
N A noNAI. (IEKBARILrM,
PRETORIA.
DISTRICT <\
CM,- ' Hf;€fee ■'/■ Vo.
Ycm. Name
Locality
Habitat
Dcscriyjion etc.
SSi)V
NAS ION M
N \TION M
III- iUlMUl M, I*Rl TORI
> j-' '
•Ki -C rt H j
. i, v IbAlo
Plate 7. — Asparagus acocksii Jessop ( Acocks 16296, holotype).
96
UNIVERS'TY OF C A°F TOWN
. , (. U \ •< u • M Arm '«M:M P
l).u-niMi. ,.i il* ii ssor i" '■
Plate 8. — Asparagus crassicladus Jessop ( Esterhuysen 4670, holotype).
97
Bothalia 9, 1: 97—112
Notes on South African Cruciferae
by
W. Marais
Abstract
A new genus in the section Arabideae, sixteen new species and four new varieties are described,
and eleven new combinations made: — Heliophila arenaria Sond. var. acocksii — var. glabrescens (Schulz)
(H. sabulosa Schltr. var. glabrescens Schulz),' — var. agtertuinensis (Schulz) (H. maximilianii Schltr.
var. agtertuinensis Schulz), H. diffusa (Thunb.) DC. var. flacca (Sond.) (H. flaeca Sond.), H. elata
Sond. var. pillansii, H. latisiliqua E. Mey. ex Sond. var. macrostylis (E. Mey. ex Sond.) (H. macrostylis
E. Mey. ex Sond.), H. linearis (Thunb.) DC. var. linearifolia (Burch, ex DC.) (H. linearifolia Burch,
ex DC.), — var. reticulata (Eckl. & Zeyh.) ( H . reticulata Eckl. & Zeyh.), H. meyeri Sond. var. minor,
H. pusilla L.f. var. maerosperma, — var. setacea (Schltr.) (H. setacea Schltr.), — var. lanceolata (Adamson)
(H. lanceolata Adamson), H. alpina, H. cedarbergensis, H. cuneata, H. esterhuyseniae, H. eximia, H.
filieaulis, H. katbergensis, H. obibensis, H. promontorii, H. rimicola, H. schulzii, Lepidium divarication
Ait. subsp. trifurcum (Sond.) ( L . trifurcum Sond.), L. basuticum, L. suluense, L. transvaalense, Thlas-
peocarpa nama uensis, APLANODES, type A. doidgeana, A. sisymbrioides (Schltr.) ( Heliophila sisym-
brioides Schltr.) Brachycarpaea juncea (Berg.) ( Cleome juncea Berg.).
In continuation of my studies on the South African Cruciferae for the Flora of
Southern Africa (see Bothalia 8: 166-169, 1964), further combinations and new
descriptions are required.
In the lists of specimens cited, no herbarium reference is given for collections
represented in the National Herbarium, Pretoria.
HELIOPHILA
Heliophila arenaria Sond. in Abh. Ges. Naturw. Hamb. 1: 262 (1846). Type:
Drege 7568 (K!).
(а) var. arenaria.
H. arenaria Sond., l.c. (1846); FI. Cap. 1: 53 (1860). H. maximilianii Schltr. in
Bot. Jahrb. 27: 136 (1899). Type: Schleehter 8229 ( B ! ). H. sabulosa Schltr., l.c. 138
(1899). Type: Schleehter 8339 ( B ! ). H. lightfootii Phillips in Ann. S. Afr. Mus. 9:
111 (1913). Type: Phillips 7683 (SAM!).
(б) var. acocksii Marais , var. nov., a typo siliquis crassioribus tnarginibus fere
rectis et stylis in fructu longioribus differt.
Type: Calvinia, Lokenburg, Acocks 19750 (PRE, holo. ; K).
Cape. — Vanrhynsdorp: Tigerberg, Lewis 4959 (SAM); Urionskraal, Barker 8554 (NBG). Calvinia:
Lokenburg, Acocks 19750; Leistner 317; between Nieuwoudtville and Oorlogskloof, Leipoldt 3960;
between Nieuwoudtville and Graskop, Lewis in BOL 19823 (BOL); Botterkloof Pass, Maguire 196,
p.p. (NBG); Lewis 3124 (SAM); 3125(SAM); Compton 20898 (BOL, NBG). Clanwilliam: Brandewyn
River, Maguire 245 (NBG); Pakhuis, Compton 19972 (NBG); between Pakhuis and Wupperthal,
Bolus s.n. (BOL); Leipoldt 3968; between Pakhuis and Bidouw, Bolus 8922; Wupperthal, Thode
1960; between Nardouws Pass and Brakvlei, Marais 1436.
(c) var. glabrescens ( Schulz ) Marais, comb. nov.
H. sabulosa var. glabrescens Schulz in Bot. Arch. 31: 541 (1931). Type: Stephens
& Glover 8705 (K!).
( d ) var. agtertuinensis (Schulz) Marais, comb. nov.
H. maximilianii var. agtertuinensis Schulz, l.c., 31: 540 (1931). Type: Schlechter
10858 (B!).
Heliophila diffusa (Thunb.) DC., Syst. 2: 685 (1821). Type: Thunberg Herb.
No. 14956.
(a) var. diffusa.
H. diffusa (Thunb.) DC., l.c. (1821); Sond. in Abh. Ges. Naturw. Hamb. 1: 220
(1846); FI. Cap. 1: 42 (1860). H. peltaria DC., l.c. (1821); Sond. in Abh. Ges.
Naturw. Hamb. 1: 222 (1846); FI. Cap. 1: 43 (1860). Type: Thunberg Herb. No.
14958. H. capensis (L.f.) C. A. Smith in Kew Bull. 1931 : 155 (1931), non (L.f.) Kuntze
(1898).
Lunaria diffusa Thunb., Prodr. 107 (1800). L. pinnata Thunb., l.c. (1800). Type:
Thunberg Herb. No. 14958.
Peltaria capensis L.f., Suppl. 296 (1781). Type: LINN 829.3 or 829.4.
(b) var. flacca (Sond.) Marais, stat. nov.
H. flacca Sond. in Abh. Ges. Naturw. Hamb. 1: 223, t.22, f.l (1846); FI. Cap.
1: 43 (1860). Type: Ecklon & Zeyher 69 (K!).
Heliophila elata Sond. in Abh. Ges. Naturw. Hamb. 1: 247 (1846). Type: Drege
7566a (K!).
(o) var. elata.
H. elata Sond., l.c. (1846); FI. Cap. 1: 49 (1860).
(b) var. pillansii Marais, var. nov., a typo petalis et filamentis lateralibus exappen-
diculatis, sepalis interioribus basi non saccatis, foliis saepe lobatis, indumento plerumque
densiore et longiore differt.
Type: Cape: Piketberg District, between Avontuur and Zebra Kop, Pillans 7603
(BOL, holo. ; GRA, K).
Cape. — Piketberg: between Avontuur and Zebra Kop, Pillans 7603 (BOL, GRA, K); PHlans 7506
(BOL); Kapteinskloof, Pillans 8017. Clanwilliam: Edwards in BOL 14450; 14451 (BOL); Elands-
kloof, Lewis in BOL 27109 (BOL); Pakhuis Pass, Compton 6950 (BOL, NBG); Lewis in BOL 27125
(BOL); Middelburg, Esterhuysen 1119 (BOL, K); Compton 12693 (NBG); Esterhuysen 2468; Sneeu-
berg, Esterhuysen 13099 (BOL).
Heliophila latisiliqua E. Mey. ex Sond. in Abh. Ges. Naturw. Hamb. 1 : 224
(1846). Type: Drege 585 (B!).
(a) var. latisiliqua.
El. latisiliqua E. Mey. ex Sond., l.c. (1846); FI. Cap. 1: 43 (1860).
Several-stemmed herbs with erect or procumbent stems up to 35 cm high. Leaves
1-5 cm long, the lobes 0-2-2 cm long. The flowers are bigger than in var. macrostylis
but, though the filaments are shorter, the anthers are longer. Distributed from the
Hex River Mts. to Calvinia.
(b) var. macrostylis (E. Mey. ex Sond.) Marais stat. nov.
Heliophila macrostylis E. Mey. ex Sond. in Abh. Ges. Naturw. Hamb. 1 : 225 (1846);
FI. Cap. 1: 43 (1860). Type: Drege s.n. (K!).
Stout erect herbs up to 90 cm high, usually with a single leafy stem. Leaves up
to 10 cm long, the lobes 2-5-4 cm long. The flowers are small, but the filaments are
proportionately much longer than those of var. latisiliqua. Confined to Namaqualand.
99
Heliophila linearis ( Thunb .) DC., Syst. 2: 697 (1821). Type: Thunberg Herb.
No. 15167 (UPS!).
(a) var. linearis.
H. linearis (Thunb.) DC., !.c. (1821); Sond. in Abh. Ges. Naturw. 1 : 238 (1846);
FI. Cap. 1: 47 (1860). H. linearifolia var. hirsuta Burch, ex DC., l.c. 692 (1821).
Type: Burchell 5518 (K.!). H. falcata Eckl. & Zeyh., Enura. No. 77 (1834-5). Type:
Ecklon & Zeyher 77 (K!). H. linearifolia var. pilosiuscula Sond. in Abh. Ges. Naturw.
Hamb. 1: 253 (1846); FI. Cap. 1: 51 (1860), partly, as to Burchell 5518. H. tenuis
N. E. Br. in Kew Bull. 1894: 99 (1894). Type: Pappe s.n. (K!).
Cheiranthus linearis Thunb., Prodr. 108 (1800). Type: Thunberg Herb. No. 15167.
(b) var. linearifolia {Burch, ex DC.) Marais, comb. nov.
H. linearifolia Burch, ex DC., Syst. 2: 692 (1821), excluding var. hirsuta-, Sond.
in Abh. Ges. Naturw. Hamb. 1: 252 (1846); FI. Cap. 1: 51 (1860), partly, excluding
var. pilosiuscula, and var. filifolia as to its basionym. H. linearifolia var. dolichocarpa
Schulz in Bot. Arch. 31 : 540 (1931). Type: Rogers s.n. (K!). H. elongcita Adamson
in FI. Cape Penins. 417 (1950), partly, excluding basionym of var. filifolia, non
(Thunb.) DC.
(c) var. reticulata {Eckl. & Zeyh.) Marais, comb. nov.
H. reticulata Eckl. & Zeyh., Enum. No. 90 (1834-5); Sond. in Abh. Ges. Naturw.
Hamb. 1 : 243 (1846); FI. Cap. 1 : 48 (1860). Type: Ecklon & Zeyher No. 90 (SAM!).
El. christieana Schulz in Bot. Arch. 31: 530 (1931). Type: Christie 49 (B!). H.
cornigerci Fourcade in Trans. Roy. Soc. S. Afr. 21 : 90 (1932). Type: Eourcade 1815a
(BOL!).
Heliophila meyeri Sond. in Abh. Ges. Naturw. Hamb. 1: 226 (1846). Type:
Drege s.n. (K!).
{a) var. meyeri.
H. meyeri Sond., l.c. (1846); FI. Cap. 1: 44 (1860). H. dentifera Sond. in Abh.
Ges. Naturw. Hamb. 1: 219 (1846); FI. Cap. 1: 42 (1860); Adamson in FI. Cape
Penins. 415 (1950). Type: Ecklon & Zeyher s.n. (SAM!).
{b) var. minor Marais, var. nov., a typo floribus et fructibus in partibus omnibus
minoribus, filamentis omnibus basi dilatatis papillatis sine appendice papillata differt.
Type: Cape: Swellendam District, Bonnievale, Leighton in BOL 21146 (BOL,
holo.).
Cape. — Swellendam: Bonnievale, Leighton in BOL 21146 (BOL). Worcester: Audensberg, Compton
9762 (NBG); Esterhuysen 3428 (K); Louwshoek Mts., Stokoe in SAM 59125 (SAM).
These plants differ from var. meyeri mainly in size. Since the appearance of the
plants is reminscent of some of the other species with smaller white flowers like H.
pendula and H. diffusa, rather than of var. meyeri, it is desirable to give it a name.
Heliophila pusilla L.f., Suppl. 297 (1781). Type: Thunberg Herb. Linn. 840.4.
{a) var. pusilla.
El. pusilla L.f., l.c. (1871); Sond. in Abh. Ges. Naturw. Hamb. 1: 210 (1846); FI.
Cap. 1: 40 (1860); Adamson in FI. Cape Penins. 414 (1950). H. divaricata Banks
ex DC., Syst. 2: 687 (1821); Sond. in Abh. Ges. Naturw. Hamb. 1 : 236 (1846); FI.
Cap. 1: 46 (1860). Type: Masson s.n. (BM!).
100
( b ) var. macrosperma Marais , var. nov., a typo floribus majoribus, filamentis
lateralibus papillatis, ovarii stylo longiore, ovulis paucioribus, siliquis latioribus et
seminibus majoribus differt.
Type: Cape: Stellenbosch District, Prior s.n. (K., holo.).
Cape. — Stellenbosch: Prior s.n. (K); Golf Course, Strey 665; Stellenbosch flats, Garside 158;
410 (K); Vanderstel, Smith 3207; Faure, Parker 3973 (BOL, K); Firgrove, Compton 13453 (BOL,
NBG); Elsenburg, Marioth 9281, partly ; 9281b; Simondium, Morrison in STE 1 1 109 (STE). Tulbagh:
De Hoek, Saron, Barker 5838 (BOL, NBG). Wellington: Compton 11624 (NBG). Without exact
locality: between Paarl and Pont, Drege 7555.
(c) var. setacea ( Schltr .) Marais, comb. nov.
H. setacea Schltr. in Bot. Jahrb. 49: 417 (1913). Type: Schlechter 4992 (B!).
(d) var. lanceolata ( Adamson ) Marais, comb. nov.
H. lanceolata Adamson in J. S. Afr. Bot. 7: 197 (1941). Type: Adamson 2973
(BOL!).
Heliophila alpina Marais, sp. nov. distincta.
Herba perennis parva, 10-15 cm alta, caulibus pluribus procumbentibus, partibus
junioribus puberulis. Folia 5-9 mm longa, pinnatim 5-7 lobata, stipulis minutis.
Inflorescentia pseudoterminalis, pauciflora; pedicelli breves, ± interpetiolares, bracteis
2 minutis subtendentes, saepe rami floriferi laterales pedicellos superantes. Sepala
4-5-5 mm longa, oblonga, puberula; exteriora 2 apice cucullata, breviter corniculata,
interiora 2 basi saccata. Pet ala 6-7 mm longa, 2-5-3 mm lata, anguste obovata.
FUamenta 2-4-3 mm et 3-4 mm longa. Antherae 1-1 -2 mm longae. Ovarium anguste
oblongo-ellipticum, replo minute puberulo; stylus longus; stigma breviter bilobatum;
ovula 8-12. SUiquae 1-1-5 cm longae, 3-4 mm latae, oblongae vel lineari-oblongae,
brevissime stipitatae; valvae 1-nervae; stylus 2-3 mm longus, robustus, apicem versus
attenuatus. Semina 2 -3-2 -8 mm longa, 1-5-1 -7 mm lata, ambitu oblongo-ovata,
papillosa.
Type: Basutoland: Thaba Ntlenyana, Guillarmod 2358 (PRE, holo.).
Procumbent, several-stemmed perennial herbs. Stems 10-15 cm long, the young
parts puberulous, glabrescent. Leaves 5-9 mm long, pinnately 5-7-lobed with linear,
entire lobes, probably = L fleshy, minutely puberulous, glabrescent, minutely stipulate.
Inflorescence pseudo-terminal, few-flowered, the pedicels ± interpetiolar, subtended by
2 minute bracts, often overtopped by lateral flowering branches. Sepals 4-5-5 mm
long, ovate-oblong to oblong-elliptic, puberulous; the outer 2 cucullate, shortly horn-
tipped; the inner 2 saccate and with broad membranous margins. Petals 6-7 mm
long, 2-5-3 mm broad, narrowly obovate, cuneate, very minutely puberulous at the
base outside. Filaments 2-4-3 mm and 3-4 mm long, the 2 short ones slightly thickened
at the base and with a glandular area. Anthers 1-1-2 mm long Ovary narrowly oblong-
elliptic, subsessile, finely puberulous on the replunr; style long; stigma shallowly
2-lobed; ovules 8-12. Fruits 1—1-5 cm long, 3-4 mm broad, oblong or linear-oblong,
very shortly stipitate; valves 1 -nerved, net-veined, bulged by the seeds; stipe and
replum puberulous; style 2-3 mm long, stout, tapering. Seeds 2 -3-2 -8 X 1-5-1 -7
mm, oblong-oval, subcompressed; seed-coat minutely papillose.
A very distinctive plant found on screes and steep grassy slopes at altitudes between
9.800 and 11,000 ft. Collected in fruit with a few late flowers in January; in fruit
only in March.
Cape. — Barkly East: Ben McDhui, Galpin 6573.
Basutoland. — Thabana Ntlenyana, Guillarmod 2358.
101
Heliophila cedarbergensis Marais, sp. nov., a H. scoparia Burch, habitu patenti
vel patulo, inflorescentiis intercalaribus, fructibus brevioribus seminibus paucioribus
differt.
Fruticulus patens, dense virgatus, glaber. Folia 0-5-3 cm longa, 1-2-5 mm lata,
oblanceolata vel anguste elliptico-lanceolata, levia vel cartilagineo-papillata, stipulis
minutis. Injiorescentia intercalaris. Sepala 4-5-5 3 mm longa, elliptica vel elliptico-
obovata; exteriora 2 apice cucullata, breviter corniculata, interiora 2 basi saccata.
Petala 10-11-5 mm longa, 4-5-5 mm lata, obovato-spathulata, longe ungulata. Fila-
menta exappendiculata. Antherae 1-5-1 -8 mm longae. Ovarium lanceolatum, stipi-
tatum, ovulis 10-12. Siliquae 2-4 cm longae, 2- 5-3 -5 mm latae, lanceolatae, breviter
stipitatae; valvae 1-nervatae; stylus 2-5-5 mm longus, subulatus. Semina 2-5 mm
longa, 1-6-1 -8 mm lata, ovalia, compressa, papillata.
Type: Cape: Clanwilliam District, Cedarberg, Wolfberg, Esterhuysen 20566
(BOL, holo.; K, PRE).
Trailing or densely twiggy, woody shrublets; branches glabrous with red-brown
bark, smooth or finely cartilaginous-papillate. Leaves 0-5-3 cm long, 1-2-5 mm
broad, oblanceolate to narrowly elliptic-lanceolate, acute or acuminate, mucronate,
smooth or finely cartilaginous-papillate, minutely stipulate. Inflorescence few- to
several-flowered, terminal or apparently so but soon over-topped by new leafy growth
of the central axis, i.e., intercalary. Pedicels 0-9-1 -6 cm long, erect or ascending,
subtended by 2 minute filiform bracts. Sepals 4 -5-5 -3 mm long, 1 -8-2-3 mm broad,
elliptic to elliptic-obovate; outer 2 cucullate and shortly horn-tipped, with mem-
branous margins, inner 2 membranous, saccate. Petals 10-11-5 mm long, 4-5-5 mm
broad, obovate-spathulate, rounded, long-clawed. Filaments 3-5 mm and 4 -5-5 -5
mm long. Anthers 1-5-1 -8 mm long. Ovary lanceolate, stipitate; style long, stigma
capitate; ovules 10-12. Fruits 2-4 cm long, 2 -5-3 -5 mm broad, lanceolate, shortly
stipitate; valves flat, 1 -nerved, slightly bulged by the seeds; style 2-5-5 mm long,
stoutly subulate. Seeds 2-5 mm long, 1 -6-1 -8 mm broad, oval, compressed, papillate.
Growing among rocks and boulders or on shady ledges at an altitude of 4,000-
5,500 ft. Collected in flower in October; flowers described as white or pink.
Cape. — Clanwilliam: Cedarberg, Wolfberg, Esterhuysen 20566; Esterhuysen 22448 (BOL, K);
slopes below Wolfberg Cracks, Esterhuysen 29974 (BOL); Tafelberg, Esterhuysen 30000 (BOL).
Heliophila cuneata Marais, sp. nov., a H. integrifolia L. planta suffruticosa perenni,
floribus multo majoribus, fructibus latioribus et seminibus majoribus; a H. lineari
(Thunb.) DC. forma foliorum, fructibus multo latioribus stylis longioribus, et ovulis
paucioribus differt.
Frutex diffusus usqae 1-3 m altus, ramis junioribus sparse pilosis. Folia 3-6-5
mm longa, 1 • 5-3 cm lata, cuneata vel obovato-cuneata, in parte superiore 5-9 dentata,
superne sparse vel dense pilosa vel hirsuta, subtus minus dense pilosa vel glabra.
Racemus terminalis, multiflorus, initio contractus, denique elongatus et laxus. Sepala
7-9-5 mm longa, lineari-oblonga, dense pilosa; exteriora 2 apice cucullata; interiora
2 basi saccata, marginibus membranaceis. Petala 10-16-5 mm longa, 6-5-11-5 mm
lata, obovata, rotundata, breviter unguiculata, appendicibus papillosis. Filamenta
4-5-5 mm, 5-5-7 mm longa, breviora 2 appendicibus magnis instructa. Antherae
3-3-5 mm longae. Ovarium oblongum, stylo longo gracili, ovulis 6-8. Siliquae 6-8
cm longae, 7-8 mm latae, lineari-oblongae; valvae 3-nervae; stylus 1 • 1-1 -4 cm longus,
robuste subulatus. Semina 6 mm longa, 5 mm lata, subrotundata, compressa, alata.
Type: Cape: Stellenbosch District, Swartboskloof, Taylor 5839 (PRE, holo.).
Straggling shrub up to 1-3 m high with long weak branches; young branches
sparsely pilose. Leaves 3-6-5 cm long, 1-5-3 cm broad, cuneate or obovate-cuneate,
gradually narrowed below, 5-9 toothed or lobulate-dentate in the upper half; sparsely
102
to densely pilose or hirsute on the upper surface, less densely hairy to glabrous
underneath. Racemes terminal, many-flowered, contracted in flower; elongate and
very lax in fruit; rhachis glabrous. Sepals 7-9-5 mm long, linear-oblong, densely
pilose; outer 2 cucullate; inner 2 saccate with broad membranous margins. Petals
10-16-5 mm long, 6-5-11-5 mm broad, obovate, rounded, shortly clawed, each with
a papillate appendage. Filaments 4-5 • 5 mm and 5 • 5-7 mm long, the short ones each
with a large appendage. Anthers 3-3-5 mm long. Ovary oblong with a long, slender
style, as long as or longer than the ovary; ovules 6-8. Fruits 6-8 cm long (including
the style), 7-8 mm broad, linear-oblong with straight margins; valves flat, strongly
3-nerved (sometimes with 2 or 4 additional weaker veins); style 1-1-1 -4 cm long,
stoutly subulate. Seeds 6 mm long, 5 mm broad, subcircular, winged.
The flowers are blue with a white centre, produced in spring. Known only from
the one locality where it was first recorded in 1960, growing among boulders in scrub
vegetation.
Cape. — Stellenbosch: Jonkershoek, Swartboskloof, Taylor 5839; Marais 1451 (K); Van der Merwe
22-58; Van der Merwe s.n.
Heliophila esterhuyseniae Marais, sp. nov., a H. dregeana Sond. floribus majoribus,
fructibus non moniliformibus differt.
Fruticulus lignosus confertus, erectus vel patens, glaber, 0-6-1 m altus. Caules
glabri, angulati vel anguste alati. Folia 1-4 cm longa, 1 • 5-4 mm lata, lineari-lanceolata
vel elliptico-lanceolata, sessilia, levia vel marginibus costaque subtiliter cartiligineo-
papillatis, stipulis minutis. Inflorescentia intercalaris, floribus pluribus, pedicellis bracteis
minutis 2 subtendentibus. Sepala 5 • 3-6 mm longa, ovata vel obovata, membranacea,
3-nervata; interiora 2 basi saccata. Petata 11-14 mm longa, 6- 5-7 -5 mm lata,
rotundato-obovata, unguiculata. Filamenta exappendiculata. Antherae 1-5-2 mm
longae. Ovarium lanceolatum, ovulis 4-8. Siliquae 0-8-2 -8 cm longae, 2-5-4 mm
latae, lanceolatae, breviter stipitatae, erectae; valvae 1-nervatae, denique asperae et
naviculares; stylus 1 - 5—5 mm longus, robuste subulatus. Semina 3-5-4 mm longa,
3-3-2 mm lata, citriformia, papillosa.
Type: Cape: Paarl District, Wemmershoek Pk., Esterhuysen 11290 (BOL, holo.).
Dense, erect or trailing, woody shrublets 0-6-1 m high. Stems angled or narrowly
winged by the decurrent leaf-margins. Leaves 1-4 cm long, 1-5-4 mm broad, linear-
lanceolate to elliptic-lanceolate, sessile, smooth or the margins and the midrib, which
is prominent on the lower surface, finely cartilaginous-papillate; subtended by 2 minute
stipules. Inflorescence intercalary, few- or several-flowered, each pedicel subtended by
2 minute bracts; the old pedicels from previous seasons’ flowers persistent. Sepals
5-3-6 mm long, 2-3-3 mm broad, membranous, 3-nerved; outer 2 obovate, cucullate,
shortly horn-tipped ; inner 2 ovate, strongly saccate. Petals 1 1-14 mm long, 6 • 5-7 • 5 mm
broad, subcircular-obovate, clawed. Filaments 4-4-3 mm and 4 -5-4 -8 mm long.
Anthers 1-5-2 mm long. Ovary lanceolate, shortly stipitate; style long, slender;
ovules 4-8. Fruits 0-8-2 -8 cm long, 2-5-4 mm broad, lanceolate, shortly stipitate,
erect; valves 1 -nerved, at first smooth, later rough, i ridged and warted, bulged
by the seeds, eventually boat-shaped; style 1-5-5 mm long, stoutly subulate. Seeds
3-5-4 mm long, 3-3-2 mm broad, citriform, papillose.
Growing on rock ledges in partial shade, old screes, at foot of cliff's or on stream-
banks. The flowers are pale blue, mauve or white, produced in spring.
Cape. — Ceres: Witelskloof, Esterhuysen 26396 (BOL); Skurfdeberg, Esterhuysen 14724 (BOL).
Worcester: Slanghoek Mts., Esterhuysen 22306 (BOL, K); Stettynskloof, Esterhuysen 17629 (BOL);
Hex River Mts., Morainekloof, Esterhuysen 28073 (BOL). Paarl: Wemmershoek Pk., Esterhuysen
11290 (BOL). Caledon: Fransch Hoek Mts., Roesbos Pk., Esterhuysen 29422 (BOL). Paarl-
Worcester: Du Toitskloof, Esterhuysen 30359 (BOL).
103
Heliophila eximia Marais, sp. nov. distincta.
Herba perennis, scandens?, glabra. Folia 4-5-6- 5 cm longa, 1-2 cm lata, elongato-
cuneata prope apicem 3-7 dentata, in petiolum attenuata. Racemus terminalis,
contractus, pauci- ad multi-florus. Sepala 3-3 • 8 mm longa, oblonga; interiora 2 leviter
saccata. Petala 6 mm longa, 4 • 5-4 • 8 mm lata, rotundato-obovata, breviter unguiculata.
Filamenta subulata. Antherae 1 -8-2 mm longae. Ovarium anguste oblongum, breviter
stipitatum; stylus brevis; stigma capitatum, magnum; ovula 8-12. Fructus ignotus.
Type: Cape: Namakwaland district, Hellskloof, Hall 789 (NBG, holo.).
Perennial scandent herbs?, glabrous. Leaves 4 -5-6 -5 cm long, 1-2 cm broad,
elongate-cuneate, 3-7-toothed near the apex, narrowed into a petiole. Racemes terminal,
few- to many-flowered, contracted. Sepals 3-3-8 mm long, 1-5 nun broad, oblong;
outer 2 cucullate, shortly horn-tipped; inner 2 slightly saccate with broad membranous
margins. Petals 6 mm long, 4 -5-4 -8 mm broad, obovate-circular, shortly clawed.
Filaments 2-2-4 mm and 2- 5-2- 8 mm long, subulate. Anthers 1 • 8-2 mm long. Ovary
narrowly oblong, shortly stipitate; style short; stigma large, capitate. Fruits not
known.
Known from only one gathering mentioned above which was collected on 19th.
September, 1953, in flower; the flowers are white. The affunity of this plant is obscure,
all the more so in the absence of ripe fruits.
Heliophila filicaulis Marais, sp. nov., a H. tulbaghensi Schinz floribus majoribus,
fructibus compressis differt.
Fruticulus multo ramosus tenuis; rami glabri, striis minutis cartilagineis, plerumque
valde moniliformibus. Folia 1-3 cm longa, lineari-filiformia, acuta, mucronata, stipu-
lata. Racemus terminalis, pauci ad pluri-florus. Sepala 5-5 • 5 mm longa, anguste
oblonga, membranacea; interiora 2 basi saccata. Petala 10-11-5 mm longa, 3 -8-4 -3
mm lata, oblongo-obovata, unguiculata. Filamenta exappendiculata. Antherae 1-7-
1-9 mm longae. Ovarium lanceolatum, stipitatum; stylus longus; ovula 4-8. Siliquae
1-5-3 cm longae, 1 -6-1-8 mm latae, lineares, compressae, moniliform.es, stipite 0-5-2
mm longo; stylus 2-5-3 mm longus, linearis, robustus. Semina 2 -2-2 -3 mm longa,
1-3-1 -4 mm lata ovata, compressa, papillosa.
Type: Cape: Worcester District, Hex River Mts., Esterhuvsen 15936 (BOL,
holo.; K. NBG. PRE).
Much-branched slender shrublets with wiry branches; branches glabrous with
minute cartilaginous lines which are mostly strongly beaded, these disappearing with
age. Leaves 1-3 cm long, linear-filiform, acute, mucronate, subtended by 2 minute,
filiform stipules. Racemes terminal, few- to several-flowered. Pedicels 0-6-1 cm
long, erect or ascending, usually subtended by 2 minute bracts. Sepals 5-5 • 5 mm long,
1-4-2 mm broad, narrowly oblong, membranous; outer 2 cucullate; inner 2 saccate.
Petals 10-11-5 mm long, 3 -8-4 -3 mm broad, oblong-obovate, narrowly clawed.
Filaments 4-5 mm and 5 • 5-6- 5 mm long. Anthers 1 • 7-1 • 9 mm long. Ovary stipitate,
lanceolate; style long; ovules 4-8. Fruits 1-5-3 cm long, 1-6-1 -8 mm broad, linear,
compressed, moniliform with narrow waists between the seeds; stipe 0-5-2 mm long;
style 2-5-3 mm long, linear, stout. Seeds 2 -2-2 -3 mm long, 1-3-1 -4 mm broad,
ovate, flat, papillose.
Growing on southern, south-eastern and eastern aspects in shelter of rocks.
Collected in flower in October; flowers blue to pinky-blue.
Cape. — Worcester: Hex River Mts., Esterhuvsen 15936; Outer Sanddrift Peak, Esterhuysen 29777
(BOL).
7230691-4
104
Heliophila katbergensis Marais, sp. nov., a H. rigidiuscula Sond. habitu, petalis
exappendiculatis et seminibus paucioribus differt.
Herba perennis. Caules plures, procumbentes, striis minutis cuticularibus. Folia
1-1 • 8 cm longa, 1 • 8-2 • 3 mm lata, oblanceolata vel oblanceolato-elliptica, integra,
acuta, stipulis 2 minutis. Racemus terminalis, laxus. Sepala 3 • 5-4 mm longa, oblonga;
interiora 2 basi saccata. Petala 6 • 5-7 • 5 mm longa, 3 • 5-4 • 5 mm lata, late obovata. Fila-
menta subulata, breviora 2 appendiculata. Antherae 1-8-2 mm longae. Ovarium
ellipticum, stipitatum, ovulis 4-6. Fructus immaturus 1-6-1 -7 cm longus, 4-5-5 mm
latus, oblongus, pedicellis recurvis 1 -5-2 cm longis; valvae planae, 1-nervae, reticulatae;
stylus 3 -5-6 -5 mm longus, robustus, c. 1 mm crassus.
Type: Cape: Stockenstroom District, Katberg, Hutchinson 1614 (PRE, holo.;
BOL, K).
Procumbent, many-stemmed, glabrous herbs. Stems with minute cuticular ridges.
Leaves 1-1-8 cm long, 1-8-2 -3 mm broad, oblanceolate or oblanceolate-elliptic, entire,
acute, mucronate, subtended by 2 minute stipules. Racemes terminal, lax, many-
flowered. Sepals 3-5-4 mm long, oblong, membranous-margined; outer 2 thickened
at the apex; inner 2 saccate. Petals 6 -5-7 -5 mm long, 3 -5-4 -5 mm broad, broadly
obovate. Filaments subulate, the short ones each with an appendage. Anthers 1 • 8-2
mm long. Ovary elliptic, stipitate; ovules 4-6. Immature fruit 1-6-1 -7 cm long,
4-5-5 mm broad, oblong, on recurved pedicels 1-5-2 cm long; valves flat, 1 -nerved,
reticulately veined; style 3 -5-6 -5 mm long, stout, c. 1 mm thick.
Known from only the one gathering from the top of Katberg; collected in flower
on 27th November, 1928. The flowers are described as deep mauve. Like so many
other of Hutchinson’s specimens, this one was misidentified and in “A Botanist in
Southern Africa ” he refers to it as Heliophila glauca Burch.
Heliophila obibensis Marais, sp. nov., a H. amplexicauli L.f. foliis angustis linearibus,
floribus multo minoribus, petalis et filamentis exappendiculatis; a H. pectinata Burch,
ex DC. foliis simplicibus, siliquis latioribus lineari-oblongis differt.
Herba annualis, erecta, parva, glabra. Folia linearia, integra. Racemus terminalis,
pluriflorus. Sepala 1-7-2 mm longa, anguste oblonga. Petala 2 -5-3 -5 mm longa,
1 • 2 mm lata, anguste oblongo-obovata. Filamenta 1 • 8-2 mm et 2-2 ■ 3 mm longa,
filiformia. Antherae 0 ■ 5-0 • 6 mm longa. Ovarium stylo brevi robusto, stigmate capitato
magno, ovulis 8-12. Siliquae 1-2-1 -7 cm longae, 2-2-5 mm latae, lineari-oblongae,
submoniliformes; valvae planae, 1-nervae, reticulatae; stylus c. 0-5 mm longus.
Semina ignota.
Type: South West Africa: Luderitz District, Obib, Merxmuller & Giess 3340
(M, holo.).
Erect, branched, glabrous, annual herbs up to 15 cm high. Leaves up to 2 • 5 cm
long, 0-7-0 -8 mm wide, linear, simple, each subtended by 2 small filiform stipules.
Pedicels each subtended by 2 small filiform bracts; in flower slender, erect, 2-4 mm
long; in fruit spreading or ± recurved, 5-6 mm long. Sepals 1 • 7-2 mm long, 0- 6-0- 8
mm broad, narrowly oblong; the outer 2 ± cucullate; the inner 2 slightly saccate.
Petals 2 -5-3 -5 mm long, 1-2 mm broad, narrowly oblong-obovate, clawed. Filaments
1-8-2 mm and 2-2-3 mm long, filiform. Anthers 0-5-0 -6 mm long. Ovary linear-
oblong; style short, stout; stigma large, capitate; ovules 8-12. Siliquae 1-2-1 -7 cm
long, 2-2 -5 mm broad, linear-oblong, submoniliform; valves flat, 1 -nerved, reticulately
veined; style c. 0-5 mm long. Seeds not seen.
This distinctive species is known from only the one gathering made on the 1st
of September, 1963, in quartzite screes on the slopes north of Obib. The small white
flowers turn pinkish with age.
105
Heliophila promontorii Marais, sp. nov., a H. macowaniana Schltr. planta fere
glabra, floribus majoribus caeruleis vel lilacinis siccitate roseis, siliquis sursum versus
curvatis stylis inconspicuis parvis differt.
Herba annualis, tenuis, sterea; caules basin versus fusci; nodi inferiores minute
puberuli, reliqui glabri. Folia integra vel bipinnatifida, linearia vel lineari-filiformia.
Sepala 2-8-4 mm longa, oblonga, interiora 2 basi saccata. Petala 5 -5-8 -5 mm longa,
2-5-6 mm lata, obovata, exappendiculata vel appendicibus minutis, caerulea vel lilacina,
siccitate rosea. Filamenta lateralia plerumque appendicibus minutis. Ovarium lineare,
ovulis 16-25. Siliquae 2-3 cm longae, 0-8-1 -2 mm latae, lineares, tenues moniliformes
sursum versus curvatae; stylus 1-1 -6 mm longus, parvus, subfusiformis vel subclavatus.
Semina 1 mm longa, 0-7 mm lata, ovalia anguste marginata.
Type: Cape Peninsula, Kenilworth, Bolus 7922 (PRE, holo.; BOL, K).
Slender wiry annual herbs 6-25 cm high. Stems dark in the lower part only or
all over; lower part of stem, sometimes only a few nodes, minutely puberulous, rest
of the plant glabrous. Leaves 1-5-5 -5 cm long, simple or 3- to 7-lobed, rarely
subbipinnately lobed, subtended by 2 minute stipules; lobes linear or linear-filiform.
Racemes terminal, lax, few-flowered. Pedicels 0-7-1 -2 cm long in flower, slender;
0- 9-1 -3 cm long in fruit, spreading; subtended by 2 minute bracts. Sepals 2-8-4 mm
long, 0-9-1 -5 mm broad, oblong; outer 2 cucullate, thickened at the tip; inner 2
saccate with broad membranous margins. Petals 5 -5-8 -5 mm long, 2-5-6 mm broad,
obovate to circular-obovate, clawed, without or with a very small appendage. Fila-
ments 1 -8-2-8 mm and 2-3-1 mm long, the short ones usually with a small appendage.
Anthers 1-1 ■ 5 mm long. Ovary linear; stigma capitate; ovules 16-25. Fruits 2-3 cm
long, 0-8-1 -2 mm broad, linear, shallowly moniliform, curving upwards; valves
1- nerved; style 1-1-6 mm long, subfusiform or subclavate. Seeds broadly oval,
narrowly margined.
This is Ft. trifida sensu Adamson in FI. Cape Peninsula and sensu Schulz in
Botanisches Archiv 31: 542 (1931), a species which has not yet been given a name.
True Ft. trifida Thunb., an illegitimate name, is a synonym of H. pinnata L.f. (see
Bothalia 8, 2: 167, 1965). Some of the specimens have been identified as H. tenuifolia
Sond. which is, however, a synonym of H. digitata L.f.
The flowers are described as blue, blue with a white eye, or mauve with a white
centre. They all dry pink. The flowering period is from September to November.
The species is confined to the Peninsula, usually found on sand dunes and flats.
Cape. — Peninsula: Kenilworth, Bolus 7922; Salter 5654 (BM, BOL, K); Bolus in PRE 28953;
Levyns 5200; Rondebosch, Bolus 3965 (BOL); Vlakkenburg, Compton 20177 (BOL, NBG); Smith’s
Farm, Salter 7044 (BOL); Smitswinkel, Salter 234/3 (BM); Rooihoogte, Leighton in BOL 27350
(BOL); Compton 13781 (NBG). Without exact locality; Zeyher 41 (BOL, K, SAM); Wright s.n.
(K); Grey s.n. (K).
Heliophila rimicola Marais, sp. nov., a Ft. esterhuyseniae Marais fructibus com-
pressis et seminibus compressis levibus differt.
Fruticulus perennis, ramis rigidis. Folia 0-8-4 cm longa, 1-2-8 mm lata, lineari-
oblonga vel oblanceolata, acuta, crassa, stipulis 2 minutis. Inflorescentia intercalaris
vel ramulis lateralibus abbreviatis terminalis, pauciflora, rare floribus axillaribus.
Pedicelli 1-2-3 cm longi, erecti, apice in fructu valde incrassato, bracteis 2 filiformibus
minutis subtendentibus. Sepala 5-7 mm longa, ovata vel obovato-oblonga, exteriora
2 cucullata. Petala 1-1-1 -5 cm longa, 5-7 mm lata, obovato-spathulata. Filamenta
exappendiculata. Antherae 1-8-2 -7 mm longae. Ovarium lanceolatum, stipitatum;
stylus tenuis; ovula 4-8. Siliquae 1-5-3 cm longae, 2-7-4 mm latae, lanceolatae,
longe acuminatae, compressae, stipite 0-5-2 mm longo; valvae 1-nervae; stylus 2-3-5
mm longus, tenuis, apicem versus attenuatus. Semina 2-1-3 mm longa, 1-8-2 -5 mm
lata, subrotunda vel oblongo-rotunda, compressa, levia.
106
Type: Cape: Ladismith District, Swartberg, Towerkop, Esterhuysen 26800 (BOL,
holo.).
Glabrous, perennial shrublets with stiff branches; young branches striate with
ridges from the thickened leaf-base; older branches smooth, round, with a light brown
bark. Leaves 0-8-4 cm long, 1-2-8 mm broad, linear-oblong, linear-oblanceolate or
oblanceolate, acute or shortly acuminate, fleshy, subtended by 2 minute stipules.
Inflorescence intercalary or terminal on abbreviated side-branches, rarely with some
axillary flowers. Pedicels 1-2-3 cm long, erect, the apex thickened and becoming
strongly swollen in fruit, subtended by 2 minute filiform stipules. Sepals 5-7 mm
long, 1 • 5-3 • 5 mm broad ; outer 2 ovate or obovate-oblong, cucullate, thickened at
the apex; inner 2 ovate to elliptic, membranous. Petals 1-1-1 -5 cm long, 5-7 mm.
broad, obovate-spathulate, clawed. Filaments 4-4-8 mm and 5 -3-6 -5 mm long.
Anthers 1 -8-2-7 mm long. Ovary lanceolate, stipitate; style slender; ovules 4-8.
Fruits 1-5-3 cm long, 2-7-4 mm broad, lanceolate, long acuminate, flattened; stipe,
0-5-2 mm long; valves 1 -nerved; style 2-3-5 mm long, slender, tapering. Seeds
2-1-3 mm long, 1-8-2 -5 mm broad, subcircular or oblong-subcircular, flattened,
smooth.
Apparently always found in moist crevices on rock faces at an altitude of 6-7,000
ft. The flowers are described as white, light purplish or tinged with purple. Flowering
specimens have been collected during December; specimens with ripe fruits during
March and April.
Cape. — Ladismith: Swartberg, Towerkop, Esterhuysen 18516 (BOL); Esterhuysen 26800 (BOL);
Esterhuysen 30672 (K). Prince Albert: Swartberg, Ridge Peak, Andreae 1223; Primos 55; Stokoe
1752; 1796.
Heliophila schulzii Marais, sp. nov., a H. lactea Schltr. foliis minute stipulatis,
pedicellis bracteis 2 similibus subtendentibus, sepalis puberulis differt.
Herba annualis, 15-45 cm alta. Folia 5-12 cm longa, integra vel 3-lobata, lobis
linearibus vel lineari-filiformibus, supeme minute puberula, stipulis minutis caducis.
Sepala 4-5-6 mm longa, oblonga, breviter puberula; interiora 2 basi saccata. Petala
0-8-1 -2 cm longa, 3 -5-7 -8 mm lata, obovata, appendiculata. Filamenta 3 -5-4 -5 mm
et 4-5 mm longa. brevioribus 2 appendiculatis. Antherae 1-5-2 mm longae. Ovula
(36-) 44-70 (-80). Siliquae 3 • 5-6 cm longae, 1-1 ■ 8 mm latae, lineares, margine recto,
denique pendulae; valvae 1-nervae, glabrae vel puberulae; stylus 1 -8-3-5 mm longus.
Type: Cape: Naniakwaland district, Bowesdorp, Stokoe in SAM 59126 (SAM,
holo.).
Slender or stout annual herbs 15-45 cm high, simple or with many ascending
branches; basal part of stem and sometimes the axils of the branches shortly puberulous.
Leaves 5-12 cm long, entire or 3-lobed, the lobes linear or linear-filiform, the upper
leaves entire, shorter; upper surface minutely puberulous; stipules minute, caducous.
Racemes lax. Pedicels shortly puberulous at the base, with 2 minute bracts; in flower
slender, erect, in fruit 1-1-5 cm long, widely spreading but finally sharply reflexed.
Sepals 4-5-6 mm long, 1-2-1 -9 mm broad, narrowly oblong or oblong-elliptic, with
membranous margins, shortly puberulous; outer 2 cucullate, thickened or very shortly
horn-tipped; inner 2 saccate. Petals 0-8-1 -2 cm long, 3 -5-7 -8 mm broad, obovate,
rounded, cuneate or clawed, each with an appendage. Filaments 3 -5-4 -5 mm and
4-5 mm long, the short ones appendaged. Anthers 1-5-2 mm long. Ovary linear;
stigma large, capitate; ovules (36-) 44-70 (-80). Fruits 3-5-6 cm long, 1-1-8 mm
broad, linear with straight margins; valves 1-nerved, strongly bulged by the seeds,
glabrous or puberulous all over or just near the base; style 1 -8-3-5 mm long, straight
or very slightly swollen near its base, or subclavate. Seeds 0-9-1 -4 mm long, 0-7-1
mm broad, subcircular to oblong, emarginate.
107
The flowering time is from August to October and the flowers are blue or purple.
Apparently confined to the Kamiesberg.
Cape. — Namakwaland: Brakdam, Schlechter 11147; Maguire 958 (NBG); Kamiesberg, De Kom,
Leipoldt 3914; Leipolclt 3915 (BOL); between Kamieskroon and Leliefontein, Salter 1506 IK); Kamies-
kroon, Herre in STE 11786 (STE); Macdonald 62 (BM); Bowesdorp, Stokoe in SAM 59126 (SAM);
Garies, Compton 11565 (NBG); Henrici 2145; Spektakel, Morris in BOL 5580 (BOL); between
Swartdoring River and Groen River, Drege (BM); Grootvlei, Compton 17285 (BOL, NBG).
LEPIDIUM
Lepidium africanum (Burm.f) DC., Syst. 2: 552 (1821), quoad synon. tandem,
excl. specim. et descr. ; non auct.
Thlaspi africanus Burm.f., Prodr. FI. Cap. 17 (1768). Type: Burmann in Herb.
Delessert (Gl).
Lepidium divaricatum sensu DC., l.c., 543 (1821). quoad specim. Burm.; sensu auct.
pro parte.
In Burmann’s herbarium there are three specimens with the name Thlaspi africanus.
Burmann’s diagnosis: “ Siliculis subovatis, foliis lanceolatis serrulatis glaberrimus ”,
agrees very well with one of them. This specimen, however, is the one which De
Candolle excluded from Lepidium africanum when he made the combination. On
p. 543 under L. divaricatum he writes: “ Haec species est in herb. Burmanni confusa
cum Lepidio suo africano, ad quod revera accedit ”.
Thus, the name L. africanum must now be applied to plants which, since De
Candolle’s treatment, have been included in L. divaricatum Ait. The plants which
I include in L. africanum have mostly been put into L. divaricatum subsp. linoides.
In my opinion the type of L. linoides Thunb. does not belong here, but should go into
L. divaricatum sens, strict. The plants that we used to know as L. africanum must
now be called L. capense Thunb.
Lepidium capense Thunb., Prodr. 107 (1800). Type: Thunberg Herb. No. 14761
(UPS!).
L. decumbens Desv. in Journ. Bot. Appl. 3: 165, 176 (1814). Type: Herb. Desvaux
(ex cult.) (P, holo.!; BM!). L. africanum (Burm. f.) DC., Syst. 2: 552 (1821), pro
parte, quoad specim. et descr., excl. synon. Burm.f. — var. typicum Thell. in Viertelj.
Naturf. Ges. Ziir. 51: 187 (1906), excl. synon. Burm.f. — var. capense (Thunb.) Thell.,
l.c. — var. serratum Thell., l.c., 188 (1906). Type: Thunberg Herb. No. 14792
(UPS!). — var. serratum f. glabratum Thell., l.c., 189 (1906). Type: Ecklon & Zeyher
No. 43, p.p., not seen. — var. burchellii Thell., l.c. (1906). Type: Burchell 137 (K!).
L. africanum auct.
My concept of this species does not differ from that of Thellung except that I
do not uphold all his varieties. Only the name, and thus the synonymy, has changed.
Lepidium divaricatum Ait., Hort. Kew. ed. 1, 2: 375 (1789).
(a) subsp. divaricatum.
L. divaricatum Ait., l.c. Type: “Lepidium divaricatum Banks’ ex Hort. Kew”
in Herb. Delessert (G ! ). — subsp. eu-divaricatum Thell. in Viertelj. Naturf. Ges. Ziir.
51: 167(1906). Type as above. L. linoides Thunb., Prodr. 107 (1800). Type: Thunberg
Herb. No. 14776 (UPS!). L. divaricatum subsp. linoides (Thunb.) Thell., l.c. (1906).
— subsp. linoides var. typicum Thell., l.c., 168 (1906). Type as above. L. subdentatum
108
Burch, ex DC., Syst. 2: 545 (1821). Type: Burchell 1299 (K!). L. linoides var. subden-
tatum (DC.) Sond. in FI. Cap. 1: 28 (1860). L. divaricatum subsp. linoides var.
subdentatum (DC.) Thell., l.c. (1906). L. pinnatum sensu Sond. l.c., 30 (1860), pro
parte, quoad syn. L. divaricatum.
The type specimen is a cultivated plant, preserved in Herb. Delessert in Geneva,
with large, broad, pinnately lobed leaves like those of Bolus 25 from Graafif-Reinet.
They do not represent the usual form of the species; the majority are fairly slender
plants with small, narrower leaves.
(. b ) subsp. trifurcum (Sond.) Marais , comb. nov.
L. trifurcum Sond. in Linnaea 23: 4 (1850); FI. Cap. 1: 30 (1860). Type: Zeyher
213 (K!, PRE!).
Thellung, in Vierteljahreschrift der Naturforschenden Gesellschaft, Zurich 51 : 171
(1906), interpreted L. trifurcum incorrectly, probably because Sonder had, in Flora
Capensis, misidentified Drege 7542a with his species. Rehmann 3033, 3034 and 3219
and Drege 7542a from Vienna, cited by Thellung, all belong to L. schinzii Thell. 1
have not seen the sheet of Drege 7542a from Herb. Delessert which is also cited by
Thellung.
Lepidium ecklonii Schrad., Ind. Sem. Hort. Goetting. 3 (1830). Type: ex Hort.
Goettingen (LE!).
L. divaricatum subsp. ecklonii (Schrad.) Thell. in Viertelj. Naturf. Ges. Ziir. 51:
169 (1906). L. sylvaticum Eckl. & Zeyh., Enum. 6, No. 37 (1834-5). Type: Ecklon
& Zeyher 37 (LE!). L. pinnatum Eckl. & Zeyh., l.c. 7, No. 45 (1834-5), non Thunb.
L. hirtellum Sond. in FI. Cap. 1: 30 (1860). Type: Ecklon & Zeyher 45 (LE!). L.
capense sensu Sond., l.c. 29 (1860), pro parte, quoad syn. L. ecklonii Schrad.
Lepidium basuticum Marais , sp. nov., a L. desertorum Eckl. & Zeyh. siliculis magnis
ellipticis stylis in emarginaturis inclusis differt.
Herba perennis. Caules pilis ± falcatis dense induti. Folia basilaria rosulata,
petiolata, 8-10-jugata, pinnata, pinnis lobatis vel dentatis. Folia caulina lobis pauci-
oribus angustioribus et minus lobulatis; folia superiora minora, oblanceolata, 3-fida
vel integra. Racemi fructiferi 1-5 cm longi, densi. Sepala 0-8-0 -9 mm longa, oblonga.
Petala 0. Stamina 2. Nectaria 0-4-0 -5 mm longa, filiformia. Siliculae 2-8-3 mm
longae, 1 -7-1 -9 mm latae, ellipticae, leviter emarginatae, stylo in emarginatura incluso.
Semina 1 — 1-4 mm longa, 0-5-0-6 mm lata.
Type: Basutoland: Butha Buthe, Pela Tseou Streams, Guillarmod 2073 (Hb.
Guillarmod, holo.).
Perennial herb; stems prostrate or erect up to 30 cm high; with a long taproot.
Stems densely ± retrorsely set with flattened ± sickle-shaped hairs. Basal leaves in
a rosette, petiolate, up to 12 cm long, pinnate, 8-10-jugate, the pinnae variously lobed
or dentate; both surfaces of petiole and lower surface of the rhachis and midribs set
with short, flattened hairs. Cauline leaves pinnately lobed with fewer, narrower and
less lobulate lobes, upwards smaller, the uppermost oblanceolate, 3-fid to entire.
Racemes terminal, in fruit 1-5 cm long, dense; rhachis puberulous. Pedicels 1-5-3 -5
cm long, wide-spreading, puberulous all round. Sepals 0-8-0 -9 mm long, oblong
with membranous margins, subpersistent. Petals 0. Stamens 2; filaments subulate,
thickened at the base. Nectaries 0-4-0 -5 mm long, filiform. Siliculae 2-8-3 mm long,
1-7-1 -9 mm broad, elliptical, shallowly emarginate with the style shorter than or as
long as the sinus. Seeds 1—1-4 mm long, 0-5-0-6 mm broad, brown.
109
A distinctive plant. The leaves are reminiscent of those of L. desertorum, and
so are the short fruiting racemes. The long filiform nectaries are found in both species.
Differs from L. desertorum in the large elliptical siliculae.
A high mountain species found between 9,200 and 10,300 feet.
Basutoland. — Butha Buthe, Pela T eou St earns, Guillarmod 2073; Lehaha-la Sekhonyana, Guil-
larmod 348; Maluti Mts., Dieterlen 1260.
Cape. — Barkly East: Doodmanskrans, Galpin 6574, p.p.
Lepidium suluense Marais, sp. nov. distincta.
Herba annua vel perennis, 30-60 cm alta. Caules pilis d= clavatis brevibus puberuli.
Folia caulina inferiora sub-bipinnatipartita, lobis patulis vel plusminusve recurvis
angustis, acutis, lineari-lanceolatis. Folia superiora pinnatipartita, lobis brevibus, folia
omnia glabra vel aliquando petiolis sparse ciliolatis. Racemus fructiferus elongatus,
laxus. Sepala 0- 7-0- 8 mm longa, oblonga. Petala 0 -2 m longa. Stamina 2. Nectaria
parva, triangularia. Siliculae 2-2-3 mm longae, 1-7-1 -8 mm latae, ovato-oblongae
vel ovatae, emarginatura satis profunda, lobis saepe plusminusve convergentibus;
stylus quam emarginatura brevior. Semina 1-3-1 -4 mm longa, 0-6-0 -7 mm lata.
Type: Natal: Hlabisa District, Hluhluwe, Ward 4577 (PRE, holo.).
Annual? or biennial herbs 30-60 cm high with spreading branches in the upper
part. Stems puberulous with short dr clavate hairs. Lower stem leaves sub-bipinnati-
partite with spreading or dr recurved, narrow, acute, linear-lanceolate lobes, or some
few leaves oblanceolate, pinnatifid. Upper leaves pinnatipartite with short lobes; all
leaves completely glabrous or occasionally the petioles sparsely ciliolate. Racemes
terminal, in fruit elongate, lax; rhachis puberulous. Pedicels puberulous on the upper
surface; in fruit 1-5-3 -5 mm long, slender, ascending or arcuate. Sepals 0-7-0 -8
mm long, oblong with membranous margins. Petals 0-2 mm long. Stamens 2; fila-
ments subulate. Nectaries small, triangular. Siliculae 2-2-3 mm long, 1-7-1 -8 mm
broad, ovate-oblong or ovate, fairly deeply emarginate with the lobes frequently d:
converging; style shorter than the sinus. Seeds 1 -3-1 -4 mm long, 0-6-0- 7 mm broad,
red-brown.
A distinctive species, apparently not a weed, growing in sandy soil in light shade
in open savanna.
Natal. — Hlabisa: Hluhluwe, Ward 4577. Ntambanana, Salberg sub Curson 15.
Mozambique. — Lourenco Marques: Gomes e Sousa 3654 (K); Khocene, Junod 314 (Z).
Synonymy: L. divaricatum subsp. eu-divaricatum var. dissectum Thell. in Viertelj.
Naturf. Ges. Ziir. 51: 166, 167 (1906). Type: Junod 314 (Z, holo.!).
Lepidium transvaalense Marais, sp. nov. distincta.
Herba perennis parva; caules plures, 15-25 cm alti, sparse vel dense retrorso-
pubescentes. Folia basilaria et caulina inferiora 2-5-5 cm longa, angusta, 7-9 pinnati-
lobata, lobis brevibus oblanceolatis, vel folia oblanceolata breviter lobata. Folia
caulina media 0-5-3 cm longa, 3-7 pinnatilobata, lobis brevibus acutis puberulis cilio-
latis. Folia superiora acute dentata, lineari-oblanceolata vel cuneato-oblanceolata.
Racemus fructiferus densus. Sepala 0 • 6-0 • 8 longa, oblonga. Petala 0 • 3-0 • 4 mm longa.
Stamina 2. Nectaria parva, triangularia. Siliculae 1-8-2 -5 mm longae, 1-4-1 -7 mm
latae, ovato-oblongae vel subrotundo-oblongae, leviter emarginatae, stylo in emargina-
tura incluso. Semina 0-9-1 mm longa, 0-4-0 -6 mm lata.
Type: Transvaal: Bethal District, Leendertz 3574 (PRE, holo.).
110
Small perennial herbs, several-stemmed from a woody crown, or sometimes with
branched woody stems, 15-25 cm high, sparsely to densely retrorse pubescent. Basal
and lower stem-leaves 2 • 5-5 cm long, narrow, pinnately 7-9-lobed with short (0 • 5 cm),
lanceolate lobes, or leaves broadly oblanceolate, shortly lobed. Middle stem leaves
0- 5-3 cm long, pinnately 3-7-lobed with short sharp lobes, puberulous (especially
on the lower surface), ciliolate. Upper leaves sharply toothed, linear-oblanceolate to
cuneate-oblanceolate. Racemes terminal, in fruit dense, 4-8 (-12) cm long. Pedicels
puberulous on the upper surface, ascending or arcuate. Sepals 0-6-0 -8 mm long,
oblong, membranous-margined. Petals 0-3-0 -4 mm long, white. Stamens 2; fila-
ments subulate. Nectaries small, triangular. Siliculae 1 -8-2-5 mm long, 1 -4-1 -7 mm
broad, ovate-oblong to subcircular-oblong, shallowly emarginate, the style included
in the sinus. Seeds 0-9-1 mm long, 0-4-0 -6 mm broad, brown.
Transvaal. — Bethal: Leendertz 3574 (PRE); Machadodorp, Young in TRV 27231. Ermelo: Burtt
Davy 7721; Henrici 1278. Heidelberg: Leendertz 2639; Burtt Davy 9145. Standerton: Jenkins
in TRV 10043. Volksrust: Mogg in T.D.A. Herb. 9687 (K). Potchefstroom : Liebenberg 1019.
Natal. — Estcourt: Edwards 432 (NU). Pietermaritzburg: Wilms 1855 (K).
Orange Free State. — Harrismith: Witzieshoek, Junod in TRV 17369.
Basutoland. — Leribe: Dieterlen 137b.
Cape. — ? Prieska: Bryant B 7.
The habit of this species is distinctive, but its relationship is not clear. Included
by Thellung in his wide concept of L. divarication , which is here treated as several
distinct species.
The majority of the specimens are remarkably uniform. Junod in TRV 17369,
Mogg in T.D.A. Herb. 9687, Wilms 1855 and Dieterlen 137b differ in the broader,
cuneate-oblanceolate leaves. Edwards 432, collected in a very advanced state, obviously
belongs here but is much more robust than the other specimens.
THLASPEOCARPA
Thlaspeocarpa namaquensis Marais, sp. nov., a T. capensi C.A. Sin. floribus fruc-
tibus et seminibus minoribus, racemis fructiferis laxis differt.
Herba annua; caules et rami minute puberuli. Folia 3-10 cm longa, pectinate
3-9-lobata, lobis linearibus. Racemi fructiferi laxi. Sepala 1-5-2 mm longa, 0-8-1
mm lata, puberula vel glabra, interiora 2 basi leviter saccata. Petala 2-3 ■ 5 mm longa,
1- 3-1 -8 mm lata, obovata. Filamenta breviora minute appendiculata. Antherae
0- 3-0 -5 mm longae. Pedicelli fructiferi 1-1-5 cm longi, patuli, gracillimi. Siliculae
1 — 1-3 cm longae, 0-8-1 cm latae, subrotundae vel subrotundo-obovatae, compressae,
pendulae; valvae 1-nervae, membranaceae, maturitate reticulatae. Semina 4-7-5 mm
diametro, rotundata, compressa, late alata.
Type: Cape: Namakwaland District, Steinkopf, Acocks 19536 (PRE, holo.).
Annual herbs 15-35 cm high, branched in the upper part; stems and branches
minutely puberulous. Feaves 3-10 cm long, pectinately 3-9-lobed, the lobes broadly
linear, acute, puberulous. Racemes soon lengthening, lax in fruit. Sepals 1 • 5-2 mm
long, 0-8-1 mm broad, puberulous or glabrous, membranous-margined; the inner
2 slightly saccate. Petals 2-3-5 mm long, 1-3-1 -8 mm broad, obovate, rounded,
cuneate. Filaments of the short stamens with minute appendages. Anthers 0-3-0 -5
mm long. Fruiting pedicels 1-1-5 cm long, wide-spreading, very slender. Siliculae
1- 1-3 cm long, 0-8-1 cm broad, subcircular or subcircular-obovate, compressed,
pendulous; valves 1 -nerved, thin, finely but conspicuously reticulately veined at
maturity. Seeds 4-7-5 mm in diameter, circular, broadly winged with a narrow sinus.
The flowering time is from August to September. Apparently confined to northern
Namaqualand. The plants are more strongly pigmented than in T. capensis , the whole
plant usually having a purplish tinge. Differs from T. capensis in the small fruits and
seeds, the lax fruiting racemes, and in the flowers being smaller in all their parts.
Cape. — Namakwaland: Kasteelpoort, near Klipfontein, Bolus 6502; Bolus in Herb. Norm. Austro-
Afric. No. 1102; Steinkopf, Acocks 19536. Without exact locality: Stokoe in PRE 29579.
APLANODES, gen. nov.
Aplanodes Marais, gen. nov. (Cruciferae-Arabideae), Cardamini L. affinis, sed
foliis simplicibus vel pinnatifidis, indumento pilorum ramosorum, fructuum valvis
costatis manifestis differt.
Herbae perennes, pilis irregulariter ramificatis. Folia ovata, petiolata, inferiora
basi cordata vel pinnatifida lobis brevibus patentibus. Racemi terminales. Sepala
patentia, interiora 2 leviter saccata. Petala obovata, unguicularia, alba usque ad
purpurata, leviter emarginata. Filamenta plana vel basi crassa. Nectaria lateralia
circum bases filamentorum brevium annulatiformia, in annulo nectariis 2 medianis
loratis extra filamenta longa conjunctiva. Ovarium sessile; ovula in quoque loculo
6-18; stigma capitatum, brevissime 2-lobatum. Siliquae lineares, valvis planis 1-nervatis
reticulatis acutis. Semina uniseriata, late ovata, alata, cotyledonibus accumbentibus.
Type species: A. doidgeana Marais.
Perennial herbs with irregularly branched hairs. Leaves ovate, petiolate, the lower
ones cordate at the base or pinnatifid with short spreading lobes. Racemes terminal.
Sepals spreading, the inner 2 slightly saccate. Petals obovate, clawed, slightly emar-
ginate, white to purple. Filaments flattened or thickened at the base. Lateral nectaries
ring-shaped, closed round the bases of the short filaments; united into a ring with
the ribbon-shaped median nectaries which are outside the long filaments. Ovary
sessile; ovules 6-18 per locule; stigma capitate, very shallowly 2-lobed. Siliquae
linear, the valves flat, 1 -nerved, net-veined, acute. Seeds in 1 row, broadly oval,
winged; cotyledons accumbent.
Species 2, natives of South Africa.
The name is derived from the negative of the Greek word P/anodes, “ to wander
about ”, hence “ not to wander about ”, i.e. not an introduced plant as I first thought
it might be. Also, the type species has never been described before (cf. Planodes Greene
ex Schulz in Pflanzenfam. 17b: 541, 1936).
Aplanodes doidgeana Marais, sp. nov., species typica.
Herba perennis, ad 75 cm alta. Caules sparse vel dense puberuli, pilis irregulariter
ramificatis. Folia simplicia, pilis ramificatis dense puberula, subtus pilis plerumque
longioribus minus ramosis, petiolis basi leviter dilatatis; basalia ovato-rotundata vel
reniformi-rotundata, basi cordata, obtusa, crenata, conferta, rosulata, mox caduca;
folia caulina inferiora basi leviter cordata; superiora basi cuneata, apicibus obtusis
vel acutis, dentata vel acute serrata, ovato-oblonga vel ovata. Pedice/li fructiferi 1-1*7
cm longi, graciles, puberuli, late recurvato-patentes. Sepala 3-5-5 mm longa, puberula;
exteriora 2 cucullata, corniculata; interiora 2 basi leviter saccata. Petala 6-9-5 mm
longa, 2 -5-4 -5 mm lata, obovato-cuneata, unguiculata, leviter emarginata. Filamenta
2 -5-3 -5 mm et 3-5-5 mm longa, longioribus 4 basi applanatis, brevioribus 2 basi
incrassatis. Antherae 1-5-1 -7 mm longae. Ovarium puberulum vel glabrum; ovula
in quoque loculo 6-10. Siliquae 3-5-5 cm longae, 2-5 mm latae, lineares; valvae
1-nervae; stylus 4-6 mm longus, robustus. Semina 3-5 mm longa, 2-3 mm lata, late
ovalia, alata.
112
Type: Natal: Bergville District, Cathkin Peak Hostel, West 15 (PRE, holo.).
Perennial branched herbs up to 75 cm high. Stems thinly to densely puberulous
with irregularly branched hairs; some hairs longer, less branched to simple. Leaves
simple, densely puberulous with branched hairs, lower surface usually with longer,
less branched hairs, petiolate, the petiole slightly dilated at the base; basal leaves
petiolate, the petiole slightly dilated at the base; basal leaves ovate-circular or reniform-
circular, cordate at the base, obtuse, crenate, crowded, rosulate but dying off early;
lower stem leaves shallowly cordate, the upper cuneate at the base, obtuse to acute,
dentate to sharply and irregularly or doubly serrate, ovate-oblong to ovate. Racemes
dense in flower, lax in fruit; rhachis puberulous. Fruiting pedicels 1-1-7 cm long,
slender, widely recurved-spreading, puberulous. Sepals 3-5-5 mm long, puberulous,
membranous-margined; outer 2 narrowly oblong, ± cucullate, horn-tipped; inner
2 ovate, ± saccate. Petals 6-9-5 mm long, 2 -5-4 -5 mm broad, obovate-cuneate,
clawed, shallowly emarginate. Filaments 2 -5-3 -5 mm and 3-5-5 mm long; long
ones flattened at the base; short ones thickened at the base. Anthers 1-5-1 -7 mm
long. Ovary linear-oblong, glabrous or with coarse hairs on the valves; ovules 6-10
per locule. Siliquae 3-5-5 cm long, 2-5 mm broad, linear; valves 1-nerved, net-veined;
style 4-6 mm long, stout; stigma capitate, shallowly 2-lobed. Seeds 3-5 mm long,
2-3 mm broad, broadly oval winged.
Named after Dr. Doidge who first collected the species in 1920. The flowers
are described as white, but some of the dried specimens show a definite pinkish tinge.
Collected in flower from October to May.
Natal. — Bergville: National Park, Galpin 10317 (K); Cathedral Peak Forestry Stn., Killick 1862;
Cathkin Peak Hostel, West 15; Tugela Valley, Doidge s.n.; Acocks & Hafstrom 524.
Since the description was drawn up, I have received from the Government
Herbarium, Stellenbosch, a specimen of Thode 7778, collected in the Injasuti Valley,
altitude 6-7,000 ft., in November 1890. This specimen differs from all the others in
having only simple hairs, and somewhat narrower, ovate-lanceolate upper stem leaves.
Its flowers are described as dull lilac. This specimen consists of only the upper part
of the plant, and appears rather more robust than the other specimens. All other
characters are the same.
Aplanodes sisymbrioides ( Schltr .) Marais, comb. nov.
Heliophila sisymbrioides Schltr. in Journ. Bot. 35: 340 (1897). Type: Queenstown,
Galpin 2260 (B !).
An insufficiently known species, differing from A. doidgeana in the pinnatifid
leaves.
BRACHYCARPAEA
Brachycarpaea juncea {Berg.) Marais, comb. nov.
Cleome juncea Berg., Descr. PI. Cap. 164 (1767); Thunb., Prodr. 109 (1800). Type:
Grubb s.n. (SBT!). C. laxa Thunb., l.c. (1800). Type: Thunberg, Herb. no. 15423!
Fleliophila flava L.f., Suppl. 297 (1781). Type: Thunberg in Herb. Linn. no. 8407!
Brachycarpaea varians DC., Syst. 2: 699 (1821); Sond. in FI. Cap. 1: 33 (1860).
B. varians var. flava (L.f.) DC., l.c. (1821); Sond., l.c. (1860). B. varians var. pur-
purascens DC., l.c. (1821); Sond., l.c. (1860). Type: Herb. Burmann sub. Polygala
bracteolata. B. polygaloides Eckl. & Zeyh., Enum. no. 51 (1834-5). B. linifolia Eckl.
& Zeyh., l.c. no. 53 (1834-5). B. laxa (Thunb.) Sond., l.c. (1860). B. laxa var.
stricta Sond., l.c. (1860). B. flava (L.f.) Druce in Rep. Bot. Exch. Cl. Br. 1. 1916:
610(1917). B. capensis Fourc. in Bot. Surv. Mem. 20: 36 (1941), excl. synon. Linn.
Coronipus anomalus Spreng., Syst. Veg. 2: 853 (1825), nom. illeg.
113
Bothalia 9, 1: 113— 121.
Notes on Boerhavia in Southern Africa
by
L. E. Codd
Abstract
A key is provided to the six species recognized in South Africa, of which B. pterocarpaS. Wats,
and B. erecta L. are introduced weeds from America. B. deserticola (South West Africa) is described
as new.
During the preparation of a handbook on the more common weeds of South
Africa, it became necessary to review the names in use for the species of Boerhavia
present in this country. Many of the species are cosmopolitan weeds and, as is common
experience in such cases, their nomenclature presents some difficulty. Thus, although
the available herbarium material may be sorted into groups of specimens, it is not
always easy to assign a name to each group. For assistance in this respect I am
grateful to our Liaison Officer at Kew, Dr. O. A. Leistner, who, in turn, received
valuable advice from Mr. R. D. Meikle, the author of the treatment of Boerhavia in
the second edition of the Flora of West Tropical Africa, Vol. 1, p. 177 (1954).
In the Flora Malesiana 6, 3: 452 (1964), Stemmerik takes a broad view of the
genus and reduces Commicarpus Standi, to synonymy. This view seems unnecessarily
drastic as the two genera can be distinguished chiefly by the character of the fruits,
although florally they are very close. In South Africa, six species of Boerhavia and
four species of Commicarpus are recognized. The present study is concerned with
Boerhavia sens, strict.
Boerhavia fruits (anthocarps) are very small but, under magnification, often provide
useful diagnostic features. When a clear-cut fruit character is associated with a
relatively constant facies, it is considered that such a combination provides grounds
for separate taxonomic status, especially when the type of pubescence is also relatively
constant within the group of specimens.
BOERHAVIA
Boerhavia I., Sp. PI. 3 (1753); Gen. PI. ed. 5: 9 (1754); Willd., Sp. PI. 1: 19 (1797) (as
“ Boerhaavia ”) ; Choisy in DC., Prodr. 13, 2: 449(1852); Heimerl in Pffanzenfam. 3,
lb: 26 (1889); 16c: 117 (1934); Bak. & Wright in FI. Trop. Afr. 6, 1 : 2 (1909);
T. Cooke in FI. Cap. 5, 1: 393 (1910); Standley in N. Amer. FI. 21: 204 (1918);
Burtt Davy, FI. Transv. 1 : 209 (1926); Cufodontis in Bull. Jard. Bot. Brux. 23 (Suppl.):
76: (1953) (as “Boerhavia”); Meikle in FI. W. Trop. Afr. ed. 2, 1: 177 (1954);
Stemmerik in FI. Males. 6, 3: 452 (1964).
A small genus containing several cosmopolitan weeds and other species of more
restricted distribution. Six species are recognized in South Africa, one of which is
subdivided into three varieties. Probably three or four species are indigenous and
there is a possibility that, in certain cases, there may be indigenous and introduced
populations.
Synonymy was investigated only to the extent necessary to ascertain, as far as
possible, the correct name for each taxonomic unit recognised in South Africa.
114
B. diffusa var. diffusa is reported to be a good fodder plant, but the sticky fruits
x>f this, and certain other species, are generally regarded as being objectionable,
especially in wool-producing areas.
Fruit eglandular:
Fruit 3-4-winged 1. B. pterocarpa
Fruit 5-ribbed:
Fruit truncate at the apex, ribs obscurely undulate; perianth purple, up to 2-5 mm long;
leaves broadly ovate; stems puberulous 2. B. erecta
Fruit rounded at the apex, ribs smooth: perianth white to pale pink, 3-5 mm long; leaves
ovate or ovate-elliptic to ovate-lanceolate; stems puberulous to glandular-hispid
5. B. deserticoia
Fruit glandular:
Flowers in ultimately lax, much-branched cymes; leaves usually more than 2-5 cm long:
Leaves broadly ovate to subrotund; stems glabrous to densely glandular-hirsute:
Stems usually glabrous; young leaves fringed with long septate hairs, otherwise glabrous,
usually not wrinkled below; flowers deep purple 3a. B. diffusa var. diffusa
Stems puberulous to glandular hirsute; leaves thick-textured, often wrinkled and whitish
below, usually puberulous to hispid; flowers white to deep pink:
Stems puberulous to hispid often with scattered multicellular hairs; leaves usually thinly
hispid below, glabrescent 3b. B. diffusa var. hirsuta
Stems and leaves densely hirsute, viscid, with stipitate glands often present 3c. B. diffusa
var. viscosa
Leaves ovate-elliptic to lanceolate; stems densely glandular-hirsute 4. B. hereroensis
Flowers in axillary clusters or short cymes up to 2-5 cm long; leaves small, usually less than
2-5 cm long 6. B. repens
1. B. pterocarpa S. Wats, in Proc. Am Acad. 17: 376 (1882); Standley in N.
Amer. FI. 21: 209 (1918). Type: from Arizona, U.S.A.
B. cordobensis Kuntze, Rev. Gen. 3, 2: 264 (1898). Type: from Argentinia, S.
America.
Annual herb. Stems prostrate, puberulous to hispid with scattered long septate
hairs. Leaves broadly ovate to subrotund; blade 2-5-4 cm long and 1-5-2 -5 cm
broad, sparsely hispid on the margins and nerves below, glabrous above, apex acute
to obtuse, base truncate; petiole 0-5-2 -5 cm long. Inflorescence cymose, usually
freely branched: peduncles glandular-puberulous. Flowers small, in many-flowered
glomerate capitula, colour not recorded on specimens collected in South Africa.
Anthocarp eglandular, obconical-clavate, 3-4-winged, broadly truncate at the apex
(Fig. 1:2).
A weed, widespread in the warmer parts of America. First recorded in South
Africa from the Kimberley District in 1919 and now known to occur in several districts
in central and northern Cape Province and western Orange Free State. Common
and evidently spreading along roads and in gardens.
Cape. — Beaufort West: Adamson D162; Theron 1615; Nelspoort, Lewis 4015 (SAM); [Fraserburg
Road, C. A. Smith 2497. Graafif Reinet: Swart River, Maguire 725 (NBG). Kimberley: Leistner
938; Schmidt’s Drift, Jan. 1919, Wilman in BOL 15854 (BOL), in SAM 13307 (SAM); Kalk Drift,
Brueckner 780. Prieska: Bryant J198; Wilman in BOL 3047 (BOL). Prince Albert: Prince Albert
Road, Compton 23903 (NBG). Steytlerville: Gill 55. Taung: Parker s.n.
Orange Free State. — Fauresmith: Leistner 1084. District unknown: Pocock in SAM 52266
(SAM); Mostert 1217.
Superficially similar to B. diffusa var. diffusa but is a smaller plant and may readily
be separated by the eglandular fruits, truncate at the apex, with 3 or 4 wing-like ridges
(see Fig. 1). The margins of the ridges are smooth, not undulate as in the next species.
115
The South African specimens cited above are an excellent match of the type of B.
cordobensis Kuntze, which comes from the Argentine. In addition, a specimen of
Leistner 1084 was sent to the Gray Herbarium of Harvard University where it was
reported to be an almost perfect match of the type of B. pterocarpa S. Wats.
2. B. erecta L., Sp. PI. 3 (1753); Standley in N. Amer. FI. 21 : 210 (1918); Meikle
in FI. W. Trop. Afr. ed. 2, 1: 178 (1954); Woodson et al. in Ann. Miss. Bot. Gard.
48: 51 (1961). Type: America, LINN 9-1 (PRE, photo.).
Annual herb. Stems decumbent to prostrate, puberulous, with scattered long
septate or gland-tipped hairs, sometimes viscid. Leaves ovate to broadly ovate, sparsely
hispid to subglabrous on the margins and on the nerves below, glabrous above, apex
acute to obtuse, base rounded to truncate; blade 2-4-5 cm long and 1 -5-3 cm broad;
petiole 1-2 cm long. Inflorescence cymose, usually freely branched; peduncles glandular-
puberulous. Flowers minute, magenta, in glomerate, 4-10-flowered capitula. Anthocarp
eglandular, clavate, broadly obtuse at the apex, tapering to the base, 5-ribbed, the ribs
slightly undulate along the margins (Fig. 1:1, la, b, c, d).
A weed, widespread in the warmer parts of America. First recorded in South
Africa from the Bloemfontein District in 1934. Now known, in addition, from Pretoria
and several localities in the eastern Cape Province. It also occurs in tropical Africa.
Cape. — Bedford: 5 miles S.E. of Daggaboers Nek, Acocks 15802; Baviaansrivier, Theron 407.
Somerset East: Marais s.n. Uitenhage: Weed Inspector s.n.
Orange Free State. — Bloemfontein: Tempe Farm, 1934, Motsamoi 4958.
Transvaal. — Pretoria: Repton 4856; Walters in PRE 29073; Codd 10514.
May be recognised by the eglandular, 5-ribbed fruits, truncate at the apex, and
with obscurely undulate margins.
3. B. diffusa L., Sp. PI. 3 (1753). Type: India, LINN 9-3 (PRE, photo.).
Stems decumbent to prostrate, diffusely branched, glabrous or variously pubescent
from minutely puberulous, with or without long brown septate hairs, to hispid or
glandular-hispid, sometimes viscid. Leaves broadly ovate to subrotund, glabrous,
except for a fringe of septate hairs on the margins, to sparingly or densely hispid, apex
obtuse to rounded, base truncate to subcordate; blade 3-5-5 cm long and 2 -5-4 -5 cm
broad; petiole 1-3 cm long. Inflorescence cymose, ultimately diffusely branched;
peduncles glabrous to puberulous. Flowers small, whitish to red or purple, in 2-5-
flowered capitula. Anthocarp glandular, clavate, 5-ribbed, rounded at the apex, tapering
to the base; ribs entire, not undulate.
Three varieties recognized in South Africa, differing mainly in pubescence of stems
and leaves. For key to varieties, see key to species.
(a) var. diffusa.
B. diffusa L., Sp. PI. 3 (1753); Choisy in DC., Prodr. 13, 2: 452 (1852); Cufodontis
in Bull. Jard. Bot. Brux. 23 (Suppl.): 76 (1953), pro parte, excl. syn. B. repens I..;
Meikle in FI. W. Trop. Afr. ed. 2, 1: 178 (1954). B. repens var. diffusa (L.) Hook.f.
in FI. Brit. Ind. 4: 709 (1885); Bak. & Wright in FI. Trop. Afr. 6, 1 : 5 (1909); T.
Cooke in FI. Cap. 5, 1: 394 (1910). B. adscendens Willd., Sp. PI. 1: 19 (1797); Bak.
& Wright in FI. Trop. Afr. 6, 1 : 4 (1909); T. Cooke in FI. Cap. 5, 1: 395 (1910).
Type: from Guinea (Herb. Willd.).
Stems glabrous or minutely puberulous with a few long brown septate hairs at
the nodes; leaves glabrous except for scattered long septate hairs on the margins;
underside not markedly wrinkled; flowers mauve, reddish or purple (Fig. 24).
116
117
A weed, widespread in the tropics of both hemispheres, possibly introduced into
South Africa; recorded from the coastal areas of Natal and from eastern and northern
Transvaal.
Natal. — Durban: Medley Wood 7199; 9345; 12220; Isipingo, Ward 586. Pietermaritzburg:
Umgeni Dam, Barker 5221 (NBG). Umlazi: Amanzimtoti, Medley Wood 13206.
Transvaal. — Barberton: Pott 5475; Kaapmuiden, Rogers 30050. Nelspruit: Liebenberg 2499.
Letaba: Modjadji's Reserve, Krige 126 ; Duiwelskloof, Scheepers s.n. Pretoria: Wonderboom,
C. A. Smith 2292; 2298; Wasserfal & Van Niekerk s.n. Sibasa: Punda Milia, Lang in TRV 32168;
13 miles N.E. of Sibasa, Codd 6883.
( b ) var. hirsuta Heim, in Bot. Jahrb. 10: 9 (1889). Type: “ Kuruman ”, Mar/oth
1102 (on the PRE specimen the locality of this number is givep as “ Namaqualand,
Ga Mhana Peak ”).
B. schinzii Heim, in Ver. Bot. Verein. Brandenburg 31 : 222 (1889). Type: Ovambo-
land, Schinz 741 (Z, holo.; BOL!; PRE, photo.). — — forma cmgustcita Heim., l.c.
223 (1889). Type: Hereroland, Liideritz 174. — — forma vi/losa Heim., l.c. 223 (1889).
Syntypes: Hereroland, Omaruru, Schinz 742 (Z; PRE, photo.); Liideritz s.n. B.
marlothii Heim, in Bot. Jahrb. 10: 9 (1889). Type: Otjimbingwe, Marloth “ 1372 ”
(this is the number cited with the type description on p. 9 though, later in the same
publication, Heimerl changed the number to 1342; on the PRE specimen, the original
number 1342 has been crossed out and replaced by 1372). B. diffusa forma subhirsuta
Heim, in Bull. Herb. Boiss. 5, Appendix 3: 67 (1897). Type: Amboland, Hopfner
66. B. bracteata T. Cooke in Kew Bull. 1909: 421 (1910); FI. Cap. 5, 1: 394 (1910);
Burtt Davy, FI. Transv. 1: 209 (1926). Syntypes: Transvaal, Barberton, Galpin 1240
(PRE); Bechuanaland, Chue Valley, Burchell 2381 (PRE); Natal, Tugela, Gerrard
1787. B. diffusa sensu Burtt Davy, l.c. 209 (1926). B. coccinea Mill. var. a Meikle
in FI. W Trop. Afr. ed. 2, 1 : 177 (1954).
Stems hispid, often with numerous long brown septate hairs, sometimes glabrescent
with age; leaves with scattered hispid hairs on the nerves below, with septate or hispid
hairs on the margins, usually glabrous above, thick-textured; lower surface usually
whitish and wrinkled; peduncles glabrous; flowers white, pink or mauve, rarely
pinkish-purple.
Widespread in the drier parts of tropical Africa and America. This is the com-
monest member of the genus in South Africa, occurring in northern Natal, Swaziland,
Transvaal, northern Cape Province and South West Africa. It is probably indigenous
but behaves as a weed, spreading on bare areas, the sticky fruits being dispersed by
grazing animals.
Natal. — Ngotshe: Pongola Bridge, Gerstner 2460.
Swaziland. — Hlatikulu: Big Bend, Compton 30913.
Transvaal. — Barberton: Avoca, Galpin 1240; Kaapmuiden, Rogers 24286: Komatipoort ,Schlechter
11874. Lydenburg: Sekukuniland, Barnard & Mogg 981. Witrivier: Kruger National Park, near
Skukuza, Letty 31 ; Codd & De Winter 5010; Lower Sabie Rest Camp, Van der Schijffl99. Pretoria:
Pienaars River, Schlechter 4221; Leendertz 774. Sibasa: 32 miles N.E. of Punda Milia, Codd 5981.
Soutpansberg: Wylliespoort, Breyer in TRV 22025; Rodin 4234; Waterpoort, Bremekamp &
Schweickerdt 298; Soutpan, Schweickerdt & Verdoorn 551; Messina: Messina, Rogers 19273; 19433;
19434; 21191; 21226; Dongola area, Codd 4385.
Cape. — Kimberley: Windsorton, Marloth 4823; Leeuwpoort, Acocks 8578. Namakwaland: Ga
Mhana Peak, Marloth 1102. Postmasburg: Klapin, Leistner 2065. Kuruman: Silk 1341; Chue
Valley, Burchell 2381.
South West Africa. — Bethanien: Kinges 2074; Buchholzbrunn, Pearson 3664 (BOL). Groot-
fontein: Schoenfelder 551. Kaokoveld: Namutoni, Barnard 550 (SAM); Kunene Banks, Barnard
551 (SAM). Karibib: Seydel9\l\ 1335; Otjimbingwe, Marloth 1404 (SAM); 1372. Keetmanshoop:
Great Karasberg, Pearson 8473 (BOL). Rehoboth: Bullspoort, Tdlken & Hardy 662. Okahandja:
Dinter 357 p.p.; Merxmiiller & Giess 1020; near Okahandja, Tdlken & Hardy 723; 22 miles from
Okahandja on road to Kalkfield, Tdlken & Hardy 730. Okavango Native Territory: Cigarette,
118
Maguire 2317 ; 6 miles E. of Tondoro Mission, De Winter 3968. Outjo: Thorne in SAM 31701 (SAM);
near Kamanjab, De Winter & Leistner 5118; near turn-off to Welwitschia on Fransfontein road to
Outjo, Tolken & Hardy 846; 20 miles from Fransfontein on road to Outjo, Tdlken & Hardy 859.
Ovamboland: Schinz 741 (BOL). Windhoek: Gil/man 152 (SAM); Wilman in BOL 15227 (BOL).
Bechuanaland. — Ngamiland, Van Son in TRV 28953.
Intermediates between this variety and the previous one occur in the northern
and eastern Transvaal, but the great majority of specimens can be allocated without
difficulty.
Pearson 9176 (BOL) from Ababes, Rehoboth District, probably is a form of
B. diffusa var. hirsuta, although it is described as being a “ 3 ft. bush The leaves
(immature?) are only 1-1 -4 cm long and 0-6-1 cm broad; the inflorescence is diffusely
branched, but there are only 1 or 2 flowers to each ultimate peduncle. However, the
anthocarps and the pubescence on stems and leaves agree well with B. diffusa var.
hirsuta.
(i c ) var. viscosa (Lag. & Rodr.) Heim, in Beitr. Syst. Nyct. 27 (1897). Type: from
South America.
B. viscosa Lag. & Rodr. in Anal. Cienc. Nat. 4: 256 (1801). B. coccinea Mill. var.
viscosa (Lag. & Rodr.) R.M. Mosc. in Cat. FI. Doming. 1: 180 (1943); Meikle in
FI. W. Trop. Afr. ed. 2, 1: 177 (1954).
Stems densely glandular-hispid, viscid; leaves densely hispid on both surfaces;
peduncles glandular-puberulous; flowers pink to reddish-purple.
Indigenous in dry, rocky places in northern Cape Province and South West Africa.
Described originally from South America and also present in tropical West Africa.
Cape. — Gordonia: near Nakop Station, Acocks 18813. Hay: Magoras, Acocks 1985. Prieska:
between Prieska and Marydale, Acocks 1757; 16 miles E. of Draghoender, Codd 1237.
South West Africa. — Keetmanshoop : Great Karasberg, Pearson 8174 (BOL); near Narubis,
Wilman 328 (BOL). Maltahohe: Dinter 2002 (2 sheets in SAM dated 16/3/1911 and 31/3/1911).
4. B. hereroensis Heim, in Bot. Jahrb. 10: 9 (1889); Bak. & Wright in FI. Trop.
Afr. 6, 1 : 3 (1909). Type: Otjimbingwe, Marloth 1403 (PRE, SAM).
Stems suberect to erect, 30-40 cm long,g landular-hirsute, viscid. Leaves ovate-
elliptic or ovate-lanceolate to lanceolate; blade 2-3 cm long and 0-5-1 -5 cm broad,
glandular-hispid on both surfaces, apex acute, base rounded; petiole 1-2 cm long,
glandular-hispid. Inflorescence cymose, freely branched; peduncles glandular-puber-
ulous. Flowers solitary or in 2-4-flowered capitula; perianth 2-2-5 mm long, purple.
Anthocarp glandular, elliptic-clavate, 5-ribbed, rounded at the apex, margin of ribs
entire (Fig. 2: 5).
Indigenous in the northern Cape Province and South West Africa, usually in
stony places.
Cape. — Gordonia: near Riemvasmaak, Acocks 16391.
South West Africa. — Bethanien: Kuibis, Range 922 (SAM). Kaokoveld: Kapupa Valley, Story
5889. Karibib: Kinges 3136; Seydel 1436; Otjimbingwe, Marloth 1403; Small Spitzkop, Tolken
& Hardy 810. Keetmanshoop: Great Karasberg, Pearson 8474 (BOL). Okahandja: Dinter 357
(SAM), p.p. Omaruru: Brandberg, Merxmiiller & Giess 1663.
The pubescence is similar to that found in B. diffusa var. viscosa, but the leaves
are constantly narrower in relation to their length, and the stems tend to be more
ascending.
119
5. B. deserticola Codd, sp. nov., B. hereroensi Heim, affinis, sed fructibus eglan-
dulosis et perianthio majore differt.
Caules decumbentes 40-80 cm longi; rami puberuli vel glanduloso-hirsuti. Folia
opposita, ovata vel ovato-lanceolata, subglabra vel glanduloso-hispidula, 3-5-5 cm
longa, 1-3-2 -8 cm lata, apice acuta, basi cuneata vel obtusa. Inflorescentia 15-30 cm
longa; bracteae lineari-lanceolatae, acuminatae. Flores in cymis multi-floris dispositi;
perianthium album 3-5 mm longum. Fructus 5-costatus, eglandulosus, elliptico-clavatus,
3 • 5-4 mm longus.
Type: South West Africa: Kaokoveld Reserve, Orupembe Waterhole near Sani-
tatas, 5th May, 1957, De Winter & Leistner 5722 (PRE, holotype; K, M).
Perennial herb. Stems woody at the base, decumbent, 40-80 cm long, divaricately
branched, puberulous to glandular-hirsute and viscid. Leaves opposite, ovate to
ovate-lanceolate, subglabrous to glandular-hispidulous and viscid; petiole 0-5-2 -5 cm
long; blade 3-5-5 cm long and 1-3-2 -8 cm broad, apex acute, minutely apiculate,
base cuneate to rounded; lower surface usually paler and somewhat wrinkled. Inflore-
scence elongate, divaricately branched, 15-30 cm long; rhachis glandular-hispidulous
or puberulous: bracts linear-lanceolate, gradually acuminate, 3-5 mm long. Flowers
glomerate in dense, many-flowered cymes; bracteoles ovate-lanceolate, acuminate,
2 mm long; perianth white to pale pink, 3-5 mm long, 5-lobed; stamens 4, anthers
just exserted at anthesis; style filiform, 6 mm long; stigma peltate. Anthocarp 5-ribbed,
eglandular, 3 • 5^1 mm long, elliptic-clavate, rounded at the apex, ribs entire, conspicuous
(Fig. 2: 1, la, b, 2, 2a).
Known as yet from two restricted areas in South West Africa, one in the Kaokoveld
near Sanitatas, and the other in the Omaruru District, inhabiting dry watercourses
and rocky places.
South West Africa. — Omaruru: Brandberg, Boss in TRV 36251; Merxmiiller & Giess 1645; 8
miles from Uis on road to Welwitschia, Tdlken & Hardy 831 ; 28 miles from Uis on road to Welwitschia,
Tdlken & Hardy 836. Kaokoveld Reserve: Orupembe Waterhole, near Sanitatas, De Winter & Leistner
5722; Anabib (Orupembe), Story 5733.
In general facies this species comes nearest to B. hereroensis Heim, but differs in
the eglandular anthocarps and longer perianth, which is white (rarely pale pink) instead
of deep pink to purple. In the field, the two may also be distinguished by the difference
in habit, B. hereroensis having semi-erect to erect stems, forming a bushy, viscid, herb
about 30 cm tall, while the stems of B. deserticola are decumbent, up to 80 cm long,
ascending at the ends, and are less conspicuously viscid.
The specimens now included in B. deserticola vary considerably in pubescence.
Merxmiiller & Giess 1645 comes nearest to B. hereroensis in this respect, having glandular-
hirsute stems with leaves glandular-hispidulous (glandular-hispid in B. hereroensis).
At the other extreme is the specimen selected as the type, namely De Winter & Leistner
5722, with minutely puberulous stems, glandular-puberulous inflorescence-branches
and peduncles, and subglabrous to minutely puberulous leaves. This gathering was
made on “ outcrops of volcanic origin (amygdaloidal basalt and a lava-like rock)
with overlying limestone ”, and the stems and leaves of the dried specimen have a
greyish-green appearance.
At first, consideration was given to maintaining these two gatherings as distinct
varieties, but the two recent collections Tdlken & Hardy 831 and 836 are somewhat
intermediate in the character of the pubescence. These two gatherings and Merxmiiller
& Giess 1645 are found on granite formation, in dry watercourses and among rocks.
120
Fig. 2.— 1, la, Boerhavia deserticoIa,xl (De Winter & Leistner 5722), lb, fruit, x 6. 2, B. deserticola',
inflorescence, X 6 ( Merxmiiller & Giess 1645); 2a, fruit, x 6. 3, B. repens, fruit, x 6
( Bryant 198a). 4, B. diffusa var. diffusa, fruit, x 6 (Pott 5475); 5, B. hereroensis, fruit, x 6
( Tolken & Hardy 810).
121
6. B. repens L., Sp. PI. 3 (1753); Meikle in FI. W. Trop. Afr. ed. 2, 1 : 177 (1954).
Type: Egypt, LINN 9-8 (PRE, photo.).
B. repens var. minor Del., FI. Egypte Icon, t.3, 1 (1813). Type: It is not known
if a specimen exists, but the illustration will suffice. B. diffusa var. minor (Del.) Cuf.
in Bull. Jard. Bot. Brux. 23 (Suppl.): 77 (1953).
Stems prostrate, pallid, glabrous. Leaves small, broadly ovate or subrotund to
ovate-lanceolate; blade 1-2-5 cm long and 0-6-1 -5 cm broad; petiole 0-5-1 -4 cm
long. Flowers in axillary clusters, in short cymes up to 2-5 cm long, rarely solitary,
often in 2-4-flowered capitula on axillary peduncles, white, pink or mauve. Anthocarp
glandular, elliptic-clavate, rounded at the apex, 5-ribbed, margin of ribs entire (Fig. 2:3).
A species of tropical Africa and Asia, probably indigenous in southern Africa,
where it occurs in dry water-courses and at the sides of pans in northern Cape Province
and South West Africa. Although locally common, it does not appear to be a potential
weed.
Cape. — Gordonia: Obobogorap, Leistner 1783; Gemsbok National Park, Leistner 1048; 1889.
Hay: Acocks 1978. Kenhardt: near Kakamas, Acocks 16367. Namakwaland: near Springbok,
Krapohl in PRE 23295. Prieska: Bryant 198a.
Sooth West Africa. — Gibeon: Mariental, Dinter 1978 (SAM). Karibib: Farm Otjimbojo, Kinges
3380; Farm Ameib, Tolken & Hardy 767. Keetmanshoop : De Winter 3348; Liebenberg 5193;
Seeheim, Dinter 2957 (SAM); Pearson 3724 (BOL); near Asab Tolken & Hardy 637. Windhoek:
Avisdam, Liebenberg 4494.
The type of B. repens L. appears, from the photograph in PRE, to be an immature
specimen, with the result that the interpretation of it has varied considerably. It has
been united with B. diffusa L., both as B. diffusa var. repens and as B. repens var.
diffusa.
The specimens cited above are an excellent match of the plant described by Delile
as B. repens var. minor Del., and it is clear that Meikle interpreted B. repens in the
Flora of West Tropical Africa ed. 2, 1 : 177 (1954) on the basis of var. minor Del. It
is convenient to follow this interpretation here. The South African specimens, whether
they should be correctly known as B. repens L. or B. repens var. minor Del., are certainly
distinct, at the specific level, from those cited under B. diffusa and its varieties, and
from the other South African species.
.
123
Boihalia 9, 1: 123 — 151.
New and Interesting Records of African
Flowering Plants
by
Various Authors
CELASTRACEAE
The Cassine Complex
Dr. N. K. B. Robson, in Bol. Soc. Brot. 39: 5-55 (1965), published an analysis
of the genera allied to Cassine L. occurring in the Flora Zambesiaca area and included
those found in the neighbouring territory of Southern Africa. He deserves credit for
his meticulous study, which has brought together a great deal of data. However,
many of his conclusions do not commend themselves to the present author, who prefers
the broader concept presented by Ding Hou in Flora Malesiana 6, 2: 284 (1962).
An important contribution by Robson concerns the typification of Cassine L.
On the evidence produced, it appears that the genus should be typified by the Cape
species, C. peragua L., Sp. PI. 268 (1753) (= C. capensis L.), an opinion that will
probably be generally accepted. Differences of opinion may arise, however, regarding
his conclusion to restrict the genus to the two Cape species, C. peragua L. and C.
parvifolia Sond. This conclusion, he argues, follows from an attempt to classify the
species into “ natural ” groups, leading him to uphold the following five genera in
addition to Cassine sens. str. : Elaeodendron Jacq.f. ex Jacq. (in which he includes
Lauridia Eckl. & Zeyh.), Crocoxylon Eckl. & Zeyh. (in which he includes Pseudocassine
Bredell), Hartogia L.f., Mystroxylon Eckl. & Zeyh. and Allocassine N. Robson.
Certain of these genera, such as Mystroxylon and Allocassine sens, str., may well
be “ natural ” ones in the sense that they appear to be relatively discrete units in an
intricate and reticulate pattern. The question arises whether any useful purpose is
served by attempting to split up a heterogeneous but easily recognized group into a
number of units based on characters which appear to be matters of degree. The result,
as applied to the South African species, is unsatisfactory and, until more information
of a fundamental nature is forthcoming, which supports this treatment throughout
the entire range of the group, it is preferred to adopt a broad concept of Cassine and
to include within it the genera Elaeodendron , Lauridia , Crocoxylon , Pseudocassine and,
probably, Mystroxylon.
The genus Pseudocassine, with its 3-merous flowers, has previously been regarded
as an aberrant member of the Celastraceae, in which family flowers with three stamens
are exceptional. When seen in perspective, it may be interpreted as the product of
an evolutionary trend within the Cassine complex towards reduction in numbers of
flower parts.
This broad concept of Cassine calls for certain combinations, which are now
effected.
Cassine reticulata {Eckl. & Zeyh.) Codd, comb. nov.
Lauridia reticulata Eckl. & Zeyh., Enum. PI. Afr. Austr. 124 (1834-35).
Elaeodendron reticulatum (Eckl. & Zeyh.) Ettingshausen in Denkschr. K. Akad.
Wiss. Wien 13: 58 (1857); Robson in Bol. Soc. Brot. 39: 39 (1965).
124
Cassine transvaalensis (Bunt Davy) Codd , comb. nov.
Salacia( ?) transvaalensis Burtt Davy in Kew Bull. 1921: 51 (1921).
Pseudocassine transvaalensis (Burtt Davy) Bredell in S. Afr. J. Sci. 33: 330 (1937).
Crocoxylon transvaalsense (Burtt Davy) Robson in Bol. Soc. Brot. 16: 41 (1965).
Hartogia L.f. (1781) is an illegitimate name, being antedated by Hartogia L.
(1759). Consideration may be given to including it in Cassine , but it shows certain
anomalous features which first require thorough investigation. It differs from Cassine
mainly in the endosperm being scanty or absent(?), while the fruit is relatively thin-
walled like that of typical C. aethiopica. Although the fruit usually contains one seed,
a second seed may partially or fully develop. It may be noted that a similar situation
arises in C. tetragona , included by Robson as a second species in his genus Allocassine.
C. tetragona appears to be out of place in Allocassine. The similarities which it shows
to the typical species, A. laurifolia (Harv.) Robson, are suggestive of convergent evolution
rather than a natural relationship. Allocassine sens. str. may well be a good genus,
but it is preferred to return A. tetragona to Cassine as it is clearly related to C. aethiopica.
The Status of the Genus Lydenburgia
Another point on which I must differ from Robson is in connection with his
decision to erect the genus Lydenburgia Robson, based on L. cassinoides Robson, a
species found in Sekukuniland, Lydenburg District. It so happens that I recently
studied a range of material of this species and was in the process of describing it in
the genus Catha. As stated by Robson, it differs from typical Catha in its longer seeds
which are trigonous and possess a small aril, lacking the wing-like aril of typical Catha.
In floral characters and general facies it closely resembles Catha edulis and I have no
hesitation in reducing Lydenburgia to synonymy.
Catha Forsk. ex Scop., Introd. Hist. 228 (1777).
Lydenburgia Robson in Bol. Soc. Brot. 39: 34 (1965).
Catha cassinoides (Robson) Codd, comb. nov.
Lydenburgia cassinoides Robson in Bol. Soc. Brot. 39: 35 (1965).
L. E. Codd
CUPRESSACEAE
Notes on Widdringtonia
In the course of a revision of the genus Widdringtonia Endl. for the Flora of
Southern Africa, attention was given to the typification of W. juniperoides (L.) Endl.,
the name generally applied to the South African “ Cedar ” occurring on the Cedarberg
in the Clanwilliam District of the Cape Province. As a result, the conclusion was
reached that the Linnaean basionym, Cupressus juniperoides L., should be rejected
as a nomen dubium, a view already expressed by Stapf in Flora Capensis 5, 2: 23 (1933).
Unfortunately, Stapf did not carry his opinion to the conclusion required by the Inter-
national Code of Botanical Nomenclature, and he upheld the name W. juniperoides
Endl., Syn. Conif. 34 (1847), non (L.) Endl. Endlicher validly effected the combination
W. juniperoides (L.) Endl., taking up the Linnaean epithet, and enlarging the concept
of the species to include certain specimens from the Cedarberg.
Linnaeus described Cupressus juniperoides in his Species Plantarum ed. 2, 1422
(1763), basing his description on two-year-old plants, originating, according to him,
from “ Caput Bonae Spei ”. There is no evidence that any subsequent author saw
a specimen of the plant described by Linnaeus, and no specimen exists in the present
Linnaean Herbarium. Judging from historical records, it is very unlikely that any
collector visited the Cedarberg at such an early date. If the seed did come from the
125
Cape it is more probable that the species concerned would be the one readily available
at the Cape, which was adequately described a few years later as Thuia cupressoides
L., Mantissa Plantarum 125 (1767). The name W. cupressoides (L.) Endl. has generally
been accepted as the correct one for the S.W. Cape plant. Up to the time of Endlicher,
botanists had occasionally applied the epithet “juniperoides ” to the S.W. Cape plant
e.g. specimens collected by Oldenland (BM) and Burniann (G). The epithet “juni-
peroides ” thus, as the oldest name, should probably be applied to the S.W. Cape plant
known as W. cupressoides. However, this might cause confusion, so it has been decided
to reject the epithet altogether, on the grounds that the epithet is inadequately described
and typified, and might even belong to a genus other than Widdringtonia. If this
course is adopted, W. cupressoides (L.) Endl. remains the correct name for the S.W.
Cape plant, while a name other than W. juniperoides becomes necessary for the Clan-
william “ Cedar ”. Although numerous synonyms with copious ecological and economic
notes exist, none of them appears to be supplied with a botanical description. Thus
the species is now redescribed and typified.
Widdringtonia cedarbergensis Marsh, sp. nov., W. schwarzii (Marloth) Mast,
affinis, sed seminibus minoribus angustissimis alatis et cicatriculis minoribus differt.
Arbor vel frutex 5-20 m altus. Folia adulta squamiformia, appressa, decussata,
ovata. Coni masculi parvi, 4 mm longi; squamae c.12, peltatae, late ovatae, decussatae.
Coni feminei globosi; valvae 4, lignae, rugosae, verrucis ad margines regulariter dispo-
sitis. Semina ovoidea, obscure alata, trigona; seminis cicatricula c. 4-5 mm lata,
6 mm longa. Cotyledones post germinationem 35 mm longae, 5 mm latae.
Callitris arborea Schrad. ex Drege, Zwei Pflzgeogr. Doc. 73 (1843), nom. nud.;
Dallimore in Kew Bull. 35: 222 (1913), nom. nud.; Stapf in FI. Cap. 5, 2: 24 (1933),
in synon. C. ecklonii Schrad. ex Pappe, FI. Cap. Med. Prodr. ed 1 : 25 (1850), nom.
nud. C. juniperoides sensu Durand and Schinz, Consp. FI. Afr. 5: 951 (1894); Engl.,
Pflzw. Afr. 2: 89 (1908); Marloth, Kapland 167 (1908).
Widdringtonia wallichii Endl., Syn. Conif. 34 (1847), nom. nud.; Stapf in FI.
Cap. 5, 2: 23 (1933), in synon. W. juniperoides sensu Endl., Syn. Conif. 32 (1847);
sensu Stapf in FI. Cap. 5, 2: 23 (1933); sensu Hubbard in S. Afr. Journ. Sc. 33: 572
(1937); sensu Smith in Journ. S. Afr. For. Add. 25 (1955); sensu Chapman in Kirkia
1: 138 (1960-61). W. wallichiana Gordon, Pinet. suppl. 107 (1875), nom. nud.
Type: Clanwilliam, Cedarberg Mts., near Middelberg West Peak, 17.6.1965,
Liickhojf in PRE 29824 (PRE, holo.).
Tree normally 5-7 m high, occasionally attaining a height of about 20 m with stem
diameter up to 2 m; crown pyramidal when young, spreading with age; bark on
young trees reddish grey, thin, fibrous, flaking off annually. Leaves of two types;
juvenile leaves on seedlings and young trees needle-like, spreading, 1-2 cm long and
up to 2 mm broad; adult leaves scale-like, appressed, ovate, 2-4 mm long, usually
strictly decussate, semi-circular in transverse section, adnate at the base, free portion
often much shorter than adnate portion. Male cones 2 mm long, mostly on short,
lateral branchlets; scales usually 6 pairs, decussate, coriaceous, peltate, broadly ovate,
acuminate, with 4 pollen sacs at the base of each scale. Mature female cones subglobose,
diameter about 2-5 cm, usually consisting of 4 (rarely 5 or 6) woody valves; valves
rough, with regular verrucae along the margin. Seeds ovoid, obscurely winged,
trigonous; seed scar about 4-5 by 6 mm. Cotyledonary leaves about 35 mm long
by about 5 mm broad. Plate 1.
Found on the Cedarberg Mts. near Clanwilliam in the Cape Province; occurs
singly or in groups over some 30 miles on rocky outcrops at an altitude of between
3,000 and 5,000 feet.
126
Cape. — Clanwilliam: Cedarberg Mts., Coetzee s.n. (PRE); Compton 12776 (NBG); Drege s.n,
(BM, G, L, P, S); Forester Bath 5535 (FD Herb., S); Forester s.n. (L, PRE); Galpin s.n. (PRE);
Kappler 9626 (FD Herb.); Leipoldt 1649 (BM, G, K, P, Z); Liickhojf s.n. (PRE): MacOwan in BOL
27689 (BOL); Marloth 3086 (PRE); Mogg 2010 (PRE); Pocock 522 (STE); Wallich s.n. (BM, K);
Ecklon & Zeyher 74-3 (GRA, L, MO, W, Z; this gathering is attributed to Drege in E and G).
A detailed study of the genus revealed that the species can be divided into two
main groups on the basis of the valve structure of the female cone and the shape of the
adult leaves. One of these groups consists of the two species W. schwarzii and W.
cedavbergensis , which differ only in their seed characteristics. The other group consists
of the species W. cupvessoides in the broad sense, which is now interpreted as including
the following species which were kept separate by Stapf, l.c. : W. stipitata Stapf, W.
dracomontana Stapf and W. whytei Rendle. Although local geographical forms in
this group can sometimes be recognised, the degree of variation is such that no constant
differences could be found on which separate specific or even subspecific status could
be justified.
J. A. Marsh
ERICACEAE
Two New Erica Species from the Drakensberg
Erica dominans Killick, sp. nov., E. glaphyrae Killick affinis, sed foliis adpressis,
inflorescentiis plerumque terminalibus, sepalis basi manifeste imbricatis, ovario ovoideo-
globoso distinguitur.
Fruticulus erectus vel suberectus, 15-45 cm altus; ramuli dense virgati, juvenes
albo-tomentosi, maturi glabri. Folia constanter 3-nata, conferta, adpressa vel interdum
nonnihil suberecta, recta, elliptica, 1-9-3 -3 mm longa, 0-7-0 -9 mm lata, supra plana
vel concava, infra convexo-triangularia, sulcata, praeter margines petiolosque glabra,
olivaceo-viridia, nitida. Flores in acervis cernuis terminalibus vel interdum in axillis
foliorum superiorum dispositi; pedicelli recurvati, 0-7-1 -8 mm longi, albo-tomentosi;
bracteae 2, 1 mm longae, marginibus ciliatis. Sepala 4, cinnamomea erubescentia
vel rubra, basi manifeste imbricata, ovata, 1-1-1 -2 mm longa, 0-5-0 -7 mm lata
(interiora saepe exterioribus angustiora et crassiora), concava, dorse carinata et in
parte superiore sulcata, cartilaginea, marginibus hyalinis ciliatis. Corolla erubescens
vel purpurata accedens (ex Lubkeo brunnea, sed flores speciminis ejus probabiliter
veteres et sicci), cyathiformis, basi nonnihil angusta, 2 -0-2 -5 mm longa; tubus inter
sepala ventricosus; lobi obtusi, 0-5-0 -9 mm longi. Stamina 8, filamentis capillaribus
1 mm longis; antherae 1 mm longae, basi setis (haud caudis) 0-6 minutis instructae.
Ovarium ovoideo-globosum, nonnunquam apice subumbonatum, glabrum, stylo tenui
1-1-1 -2 mm longo, stigmate cyathiformi-peltato incluso vel raro exserto (in Schwei-
ckerdtio s.n. et Lubkeo 275 longe exserto).
Type: Basutoland, Quachas, summit of Drakensberg in Organ Pipes Pass area,
c. 9,800 feet, 2 October 1964, Killick 3519 (PRE, holo.). Fig. 1: 2.
Erica dominans is probably the most common species of Erica on the summit of
the Drakensberg between 9,800-10,500 feet and is often dominant over large areas,
hence the epithet dominans. It is either purely dominant or co-dominant with Heli-
chrysum trilineatum var. tomentosum in what has been described as alpine heath
(Killick, 1963). E. dominans commences flowering in July, reaching a peak in October.
The range of this species is along the Drakensberg from Doodmans Krans Mountain
(not traceable on maps, but probably in the Mount Fletcher District) in the south to
Mont aux Sources in the north with several records from the northern part of Basutoland.
Basutoland. — Leribe : Butha Buthe area, Nchoatle/Khotjoane Pass, Jacot Guillarmod 4004; Khatibe
Camp, above Tlolohatze River Valley, Lubke 275; Maluti Mountains, Staples 19; Ntibokho Valley,
Jacot Guillarmod 296. Quachas: Tsanatalana Valley, Killick 1728; Organ Pipes Pass area, Killick
3519; 3538; 3539; 3540.
127
Fig. 1. — 1, Erica alticola, flowering branch (Thode A1620). 2, E. dominans, flowering branch ( Killick
3519). 3, E. glaphyra, flowering branch ( Killick 3541). a, portion of branchlet with leaf,
x 4; b, leaf, x 6; c, flower, X 10; d, pistil, x 12; e, stamen, anterior, lateral and dorsal
views, x 10.
128
Natal. — Bergville: Mont aux Sources, Edwards 360; Ga/pin 9497; Hutchinson, Verdoorn & Forbes
117; Schweickerdt s.n.; Sidey 2016; Castle Buttress, Esterhuysen 10185; Cleft Peak, Esterhuysen
12892; Organ Pipes Pass, Schelpe 847. Estcourt: Giants Castle, Ndnni s.n. Mount Fletcher?: Dood-
mans Krans Mountain, Ga/pin 6764.
Below the summit of the Drakensberg, between 6,500-8,500 feet, there occurs
what may prove to be a form of E. dominans. It is larger in all parts than typical
E. dominans. This taxon is represented in the National Herbarium by Killick 3525,
Esterhuysen 7939, 8828, 10187, 10197, 15471, 15556, 17360, 17364 and 17365. However,
careful field observations with particular reference to altitudinal variation are necessary
before its status can be definitely established.
Erica glaphyra Killick, sp. nov., E. dominanti Killick affinis, sed foliis patentibus’
inflorescentiis pseudo-spicatis vel interdum terminalibus, sepalis basi plerumque liberis,
ovario depresso-globoso distinguitur.
Fruticulus erectus vel suberectus, 15-50 cm altus; ramuli subvirgati, juvenes
puberuli, maturi glabri. Folia constanter 3-nata, patentia vel interdum suberecta
(sed petioli semper erecti), recta vel leviter incurvata, elliptica, 2- 1-3-7 mm longa,
0-5-0 -9 mm lata, supra plusminusve plana, infra convexa, sulcata, praeter margines
petiolosque glabra, olivaceo-viridia vel atro-viridia, nitida. Flores in pseudo-spicis ad
1 cm longis vel interdum in acervis cernuis terminalibus dispositi; pedicelli recurvati,
0-8-2 -5 mm longi; bracteae 2, 1 mm longae, puberulae. Sepala 4, cinnamomea vel
erubescentia vel rubra, basi libera vel interdum leviter imbricata, ovata, 11-1-2 mm
longa, 0-4-0 -6 mm lata (interiora saepe exterioribus angustiora et crassiora), concava,
dorsaliter carinata et in parte superiore sulcata, nonnihil cartilaginea, marginibus
hyalinis nonnunquam ciliatis. Corolla plerumque vinosa, 1- 5-2-0 mm longa, cyathi-
formis; tubus inter sepala ventricosus; lobi obtusi, 0-7-0 -8 mm longi. Stamina 8,
filamentis 0-8 mm longis, antheris 0-6-0 -8 mm longis basi setis (haud caudis) 0-6
minutis instructis. Ovarium depresso-globosum, glabrum; stylus crassus 0-7-1 -2 mm
longus, stigmate cyathiformi-peltato incluso vel leviter exserto. Fig. 1 : 3.
Type: Basutoland, Quachas, summit of Drakensberg in vicinity of Castle Buttress,
10,000 feet, 11 July 1952, Killick 1729 (PRE, holo.).
Erica glaphyra is less frequent than E. dominans and is rarely dominant in alpine
heath. In the Organ Pipes Pass area it occurs on south-facing slopes often in the shelter
of boulders at the edge of the escarpment. Miss E. Esterhuysen of the Bolus Herbarium,
who has made several collectings of this specie s in different parts of the Drakensberg
records E. glaphyra chiefly from south- and south-west-facing slopes. It would appear
to prefer the more mesic slopes — unlike E. dominans, which occurs on the drier north-
facing slopes and on horizontal portions of the summit. The flowering periods of
the two species are practically the same with E. glaphyra starting to flower a little later
than E. dominans. The epithet glaphyra refers to the polished or glossy appearance
of the plant. E. glaphyra has been recorded along the Drakensberg from Cathkin
Peak in the south to Mont aux Sources in the north between 8,000-10,000 feet.
Basutoland. — Quachas: Summit of Drakensberg, Castle Buttress, Killick 1729; 3520; 3541;
3542; 3543.
Natal. — Bergville: Cathedral Peak, Esterhuysen 15466; Cleft Peak, Esterhuysen 12891; Mitre,
Esterhuysen 14538; Mont aux Sources, Esterhuysen 15554; The Bell, Esterhuysen 14467. Estcourt:
Cathkin Peak, Esterhuysen 17359.
The two Erica species described above were first collected by the author in 1952,
while engaged on a botanical survey of the Cathedral Peak area. In the National
Herbarium the specimens collected were found to match a number of specimens
determined as forms of E. alticola Guth. & Bol., a Transvaal species described from
“ near Devil’s Kantoor ”. It was considered, however, that the Drakensberg material
could be divided into two distinct species differing from E. alticola. In order to reach
Table 1. — Main differences between Erica dominans , E. glaphyra and E. alticola
129
■ ■ E
* E
o oo
U c
<D
U c
2 ^
« c
8 ou-S
C X
-o'4
O £
G X) 03
X; a >,
-a a. nTx
- W 03 -G OX)
cr c > g io c* ^
d) t OX) w o-2
n 51 ,1 C o
a.r-r30scCc3_S
-a t iz ■ o o ■£ w
od r+~t c/o ^ O vX
c X
<L>
« CL
* o °*
O s— < +_•
G x ad
^ x: a
_r .opj2 = >. ,
X cd — • —
3 ^ <D .S '
(Jo ; Sx: «
0> G O -O TO u ■
i G O
° E
E u
O „ E >
— Eu ^ °
.r ' _C _d0* ^ • O
C -Si 00 ■£ £ <7 - £
O (3X-5, J oo o
- c S.SPgT §13
.Enxx
Q-m
<L> 1>
■G ^
on >
^ MO
i 2
M O > ^
^jx X o
s *
o 03
-5 cd
— £ <L>
g-g o5 >»
0^ — <L> ^
C Q 03 V o
w 00
, a =
o 3 J2
12 o c
3-* G
- 0>
a «<n
xj ."3 £
w „ <L>
( — ^-o ^
c5 5 ■ oo 5 o
-G G O ^ C — ' —
>, T
— O
c -o
O r,
£ £
Q.£
130
a definite conclusion, the author in October 1964 collected 100 random samples of
the two species in the Organ Pipes Pass area with the object of applying some statistical
test of resemblance. The specimens, together with those available of E. alticola, were
examined and the resulting data discussed with a statistician. After careful consideration
it was decided that there was no point in applying statistical tests, because the data showed
so clearly the presence of three distinct entities. Some of the Drakensberg specimens
examined showed evidence of hybridization. A summary of the main differences
between E. dominans, E. glaphyra and E. alticola as revealed by the investigation is
given in Table 1. See also Fig. 1.
The possibility of two such closely related species occurring in the same area may,
at first sight, seem questionable: but, as already pointed out, they do occupy different
ecological habitats. This is probably another example of ecological isolation, where
“ closely related species are kept apart by the spatial pattern which results from their
differing ecological requirements ” (Cain, 1944). Hybridization will occur at the
ecological boundaries.
The precise position of E. dominans and E. glaphyra in the subgenus Euerica is
difficult to establish. Although obviously closely related, they possess inflorescences
which could refer them to two different sections. E. dominans with its usually terminal
inflorescence could be placed in the section Arsace, while E. glaphyra which usually
has a short pseudo-spike could be placed in Chlorocodon. The fact that both species
sometimes have the opposite inflorescence suggests that they occupy a position inter-
mediate between the two sections.
References
CAIN, S. A., 1944. Foundations of Plant Geography. Harper and Brothers: New York.
KILLICK, D. J. B., 1963. An Account of the Plant Ecology of the Cathedral Peak Area of the Natal
Drakensberg. Mem. Bot. Surv. S. Afr. 34.
D. J. B. Killick
GRAMINEAE
A New Combination in Hyparrhenia
Hyparrhenia pilosissima {Hack.) J. G. Anderson , comb. nov.
Andropogon pilosissimus Hack, in DC., Monogr. Phan. 6: 690 (1889). Type: Natal,
Drakensberg, Biggarsberg, Rehmann 7109 (W, holo.!). A. audits Stapf in FI. Cap.
7: 357 (1898). Syntypes: U itenhage, Ecklon & Zeyher 626 ; Queenstown, near Shiloh,
Baur 897; Thaba Nchu, Burke 427!; Zeyher 1799!; Griqualand East, grassy slopes
near Clydesdale, Tyson 3105; Natal, Rietvlei, Buchanan 222.
Hyparrhenia aucta (Stapf) Stent in Both. 1: 249 (1924); Chippindall in Gr. &
Pastures S. Afr. 1 : 514 (1955).
J. G. Anderson
POLEVANSIA De WINTER, GEN. NOV.
Polevansia De Winter, gen. nov. (Gramineae-Chlorideae) Willkommiae Hack, affinis,
sed lemmate fertili exaristato, lemmate paleaque subcoriaceis (haud membranaceis),
anthoecii callo brevi obtuso differt.
Spiculae a latere visae anguste oblongae, exaristatae, dorsaliter compressae, adaxi-
ales, solitariae. distichae, alternae, contiguae, oblique vel demum horizontaliter patentes;
rhachilla inter glumam superiorem et anthoecium disarticulans, supra anthoecium haud
producta. Anthoecium 1, hermaphroditum; callus brevissimus, obtusus, minute
pilosus. Glumae persistentes, inaequales; inferior circiter \ superiorem aequans,
131
explanata, lanceolata vel anguste ovata, obtusa, enervis, membranacea; superior
lemma aequans, explanata oblongo-lanceolata, obtusa, tenuiter coriacea, dorsaliter
rotundata, 1-nervis. Lemma spiculam aequans, dorsaliter plus minusve rotundatum,
marginibus incurvis, explanatum, lanceolatum vel anguste ovatum, obtusum, tenuiter
coriaceum, 3-nerve, nervibus lateralibus in triente superiore evanescentibus, nervo
medio in mucrone brevissimo excurrente, prope margines infra medium breviter
pubescens, ceterum glabrum. Palea lemma fere aequans, a dorso visa lanceolata,
membranacea, 2-nervis, nervibus ciliatis. Lodiculae 2, oblique cuneatae truncatae,
glabrae. Stamina 3. Ovarium glabrum; styli terminates, liberi; stigmata plumosa,
albida, e anthoeciis lateraliter exserta.
Gramen perenne humile, mattam rhizomatibus et stolonibus ramosis brevibus
formans; culmi graciles, erect i vel geniculati, simplices, paucinodosi; ligulae brevis-
simae, truncatae, inembranaceae, glabrae; laminae lineares, in apicem obtusum gradatim
attenuatae, breves, planae, rigidiusculae, marginibus cartilagineis; racemi plures, in
axi primario inflorescentae contractae orti.
Type species: P. rigida De Winter.
A monotypic genus known only from Basutoland.
This genus is named in honour of Dr. I. B. Pole Evans well known for his valuable
contributions to South African botany. The spelling Polevansia is preferred to the more
unwieldy alternative Pole-Evansia.
The position of Polevansia as a member of the tribe Chlorideae can be well
substantiated. Certain characters, however, such as the single floret which usually
is devoid of a prolongation of the rhachilla, and the adaxially positioned lower glume,
can be regarded as somewhat unusual for this tribe. These characters it has in common
with several other genera of the Chlorideae. Willkommia Hack., Microcldoa R. Br.
and Catalepis Stapf et Stent exhibit both one-flowered spikelets and adaxially placed
lower glumes, while Craspedorhachis is one-flowered but has the glumes laterally
appressed to the rhachis. The other characters of Polevansia are typical of the Chlori-
deae. A study of the leaf anatomy has supplied additional evidence confirming its
position in the tribe. The silicified cells of the epidermis are suborbicular to squarish
in shape and the alternate cells are deeply ripple-walled. The epidermal hairs are
two-celled and club-shaped and typical especially of the Chlorideae and Zoiseae (Fig. 3).
The cross-section of the leaf shows a radially arranged chlorenchyma, one layer of
cells thick. The motor cells, which are characteristic in shape, occupy at least half
the thickness of the leaf (Fig. 3). For comparison of its anatomy with that of other
genera in the Chlorideae, the figures given by Prat, in Ann. Sc. Nat. bot. Ser. 10, 18:
165-258 (1936) can be used.
Polevansia rigida De Winter, sp. nov. Gramen perenne, humile, infra 10 cm alta,
mattam rhizomatibus et stolonibus ramosis brevibus formans; culmi graciles, erecti
vel geniculati, simplices, paucinodosi; innovationes intra-vaginales, breves, dense
foliosi; vaginae leves. glabrae, marginibus tenuiter membranaceae; ligulae brevissimae,
truncatae, membranaceae; laminae lineares in apicem obtusam gradatim attenuatae,
breves, planae. lnflorescentia contracta, racemis spiciformis secundis breviter pedicel-
latis. Spiculae a latere visae angustae oblongae, exaristatae dorsaliter compressae
adaxiales, solitariae, alternatim distichae, contiguae. Anthoecium 1, hermaphroditum
callus obtusus, minute pilosus. Glumae persistentes, inaequales; inferior enervis;
superior l-nervis. Lemma spiculam aequans, 3-nerve, prope margines infra medium,
breviter pubescens, ceterum glabrum. Palea lemma fere aequans, a dorso visa lanceolata,
membranacea, inter carinas dorsaliter ciliata. Lodiculae 2. Stamina 3. Ovarium
glabrum. (Fig. 2).
132
Type: Basutoland, near top of Lekhalabatesi Valley, Pole Evans 12 (PRE,
holo.).
Perennial, about 10-0 cm high, matforming with a branched system of relatively
slender rhizomes; culms ascending or prostrate, densely sheathed, rooting at the
nodes; innovations arising from the underground rhizomes and the prostrate culms,
intravaginal, short, densely leafy, ascending to erect. Leaf-sheaths chartaceous, pallid,
rather loose, 1 -0-1 -5 cm long, striate, glabrous; margins thinly membranous, glabrous.
Auricles rounded, brownish, glabrous or scabrid. Collar glabrous. Ligule a short
membranous minutely fimbriate rim. Leaf-blade linear, 1 • 0-3 • 0 cm long, 1 • 5-2 • 0 mm
broad, flat, tapering to an obtuse apex, with a few scattered long hairs above, glabrous
below; nerves numerous, arranged closely together, raised below, slightly raised above,
midnerve more or less prominent below; margin narrow, cartilaginous, scabrid.
Inflorescence rigid, oblong, spikelike, consisting of a number of short, dense, appressed,
secund racemes, arranged alternately on the axis, 2 -0-3-0 cm long and 0-8-1 -5 mm
wide ; rhachis and branches triangular in cross-section, scabrid on the angles, otherwise
glabrous. Spikelets 3 • 5^4 - 5 mm long, solitary, dorsally compressed, pedunculate,
alternate, lower glume always adaxial to the rhachis; peduncles 0-5-1 -5 mm long,
triangular in cross-section, scabrid on the angles; rhachilla disarticulating above the
glumes, not produced beyond the base of the floret. Floret solitary, hermaphrodite;
callus short, obtuse, shortly hairy. Glumes persistent, unequal; lower 1-5-2 -5 mm
long, lanceolate to narrowly ovate, subhyaline, subcoriaceous, nerveless, apex obtuse
truncate or emarginate ; upper 3 • 5-4 • 5 mm long, lanceolate (when flattened), greenish,
subcoriaceous, one-nerved, slightly keeled towards the apex, scaberulous on the keels
Fig. 2. — Polevansia rigida. A, Side view of spikelet showing adaxial lower glume, X 10; B, lower
glume, X 10; C, upper glume, x 10; D, lemma, X 10; E, palea, X 10; F, lodicules,
X 10; G, stamens and gynoecium, X 10 ( Pole Evans 12).
133
Fig. 3. — Polevansia rigida: A, cross section of the leaf; B, abaxial epidermis of the leaf: C, diagram
of the cross section of a leaf. AB, abaxial epidermis; AD, adaxial epidermis; BH, two-
celled hair; C, chlorenchyma : LC, long ripple- walled cell; MC, motor cell; SC, silicified
cell; S, mechanical tissue or stereome; ST, stoma: 1 VB, first order vascular bundle; 3VB,
third order vascular bundle ( Pole Evans 12).
134
and upper part of midnerve, apex obtuse. Lemma 3 -5-4 -5 mm long, narrowly
lanceolate-oblong (when flattened), 3-nerved, midnerve excurrent into a very small
mucro, lateral nerves evanescent (occasionally a few short inconspicuous nerves present
at the base between the midnerve and lateral nerves) ; lateral nerves hairy in the lower
half. Palea 3 -5-4-0 mm long, narrowly oblong-elliptic, two-keeled, subhyaline, sub-
coriaceous; apex truncate; margins hairy in the lower half. Lodicules 2, free, small,
truncate, 0-25 mm long. Stamens 3, anthers cultrate. Ovary glabrous, oblong, styles
distinct, plumose, Caryopsis unknown. Fig. 2.
This species has been collected only in two different valleys in the Quachas District
of Basutoland. It is reported to be a pioneer on old lands by Pole Evans. After the
discovery of the new genus by Pole Evans, a special search was undertaken by Killick,
who found it only once, growing on a rocky outcrop. Killick found Catalepis gracilis
in large quantities as a pioneer of abandoned lands. Since Polevansia resembles Catalepis
in some respects the latter may have been mistaken for Polevansia. The ecology of
the species should therefore be more thoroughly investigated.
Basutoland. — Near top of Lekhalabatesi Valley, colonizing worn out native lands, 17.2.1950,
alt. 8,500 ft, Pole Evans 12; Mokhotlong River Valley, abundant on rock outcrop, 1.5.1953, alt.
8,000 ft, Killick 1983.
B. de Winter
Styppeiochloa De Winter, gen. nov.
Styppeiochloa De Winter, gen. nov. Danthoniae DC. affinis, sed lemmatibus
breviter trifidus nervis 3 prominentibus etiam plerumque basi nervis 1-4 brevibus,
hilo lineari f- caryopsin subaequanti differt.
Spiculae breviter aristatae anguste elliptico-oblongae vel elliptico-oblongae, latera-
liter compressae, solitariae, ramis paniculae sparse ramosis, appressis. Rhachilla
supra glumas et inter anthoecia disarticulans. Anthoecia 2-5, hermaphrodita vel
supremum sterile vel ad lemma redactum, callus brevis truncatus, dense pilosus.
Glumae persistentes, subaequales vel inaequales, lanceolatae, 1-3-nervatae (raro 5-ner-
vatae nervis lateralibus tenuissimis), arista breve vel nulla, margine interdum mucrone
singulari vel duobus. Lemmata apice trifida, basin versus marginibus et dorso pilosa,
lanceolata 3-5-nervata (vel 7-nervata, nervis lateralibus tenuissimis) nervo medio in
aristam scabram rigidam brevem excurrenti, utrinque nervo singulari in mucronem
vel in aristam brevem excurrenti. Paleae anguste lanceolatae, lemmatibus longiores
vel breviores, inter Carinas duas basin versus paucipilosae, spica bifida, carinis scabris.
Lodiculae 2, connatae vel liberae, parvae truncatae interdum mucronatae, 2-5-nervatae.
Stamina 3, flavida vel purpurea. Ovarium ellipitico-oblongum glabrum, appendice
nodosa inter bases stylorum, stylis distinctis, stigmatibus plumosis. Caryopsis elliptica,
embryone quarta parte grani longitudinis, hilo lineari tribus quartibus grani longi-
tidinis.
Gramen perenne dense caespitosum, basi dura fibrosa pilosa, innovationibus
intravaginalibus, culmis erectis gracilibus glabris paucinodosis, foliis convolutolinearibus,
paniculis sparse ramosis, ramis singulis vel binis laxis vel contractis.
Type species: Styppeiochloa gynoglossa (Goossens) De Winter.
A monotypic genus from Southern Africa.
135
C
Fig. 4. — Styppeiochloa gynoglossa: A, cross section of the leaf; B, abaxial epidermis; C, diagram
of the cross section of a leaf. AB, abaxial epidermis; AD, adaxial epidermis; BH, two-
celled hair; CC, suberized cell; C, chlorenchyma; LC, long ripple-walled cell; MC, motor
cells: OU, one-celled retrorse hair; SC, mechanical tissue or stereome; S, silicified cell;
1 VB, first order vascular bundle: 3 VB, third order vascular bundle (Killick 1013).
7230691-5
136
Spikelets shortly awned, narrowly elliptic-oblong to elliptic-oblong, laterally
compressed, appressed to the contracted branches of the panicle. Rhachilla disarti-
culating above the glumes and between the florets. Florets 2-5, hermaphrodite,
uppermost floret sterile or reduced to the lemma. Callus short, truncate, densely
pilose. Glumes persistent, subequal to unequal, lanceolate, 1-3-nerved (rarely 5-nerved
with the lateral nerves short) apex trilobed or lobes absent and apex acute. Lemma
with a tri-lobed apex, lanceolate, 3-5-nerved (rarely 7-nerved with two lateral nerves
very short), median nerve and one lateral nerve on each side of the median excurrent
into a mucro or short awn. Palea narrowly lanceolate, subequal to the lemma.
Lodicules 2, small, truncate, 2-5-nerved. Stamens 3, yellow or purplish. Ovary elliptic-
oblong, glabrous, with the style bases knob-like. Caryopsis elliptic; embryo \ the
length of the grain; hilum linear, J the length of the grain.
Dense sclerophyllous tufts with hard fibrous basal sheaths forming a thick dense
mat.
The name of the genus is a combination of the Greek words “ stuppeion ” (fibre)
and “ chloa ” (grass), an allusion to the tough fibrous base of the type species.
Styppeiocliloa gynoglossa, the only species of the new genus, was first described
as Crinipes gynoglossa by Goossens in Kew Bull. 1934: 200 (1934), although at the time
Goossens thought that it might represent an undescribed genus. Specimens were sent
to Kew for an opinion on the generic affinities and, since it was similar in several
respects to the genus Crinipes Hochst., which was then only known from a few collec-
tions, and because the material was inadequate for a complete study. Professor Goossens
was advised to include it in Crinipes. Since then much more material of true Crinipes
and of species which are related to it, but could hardly be included in Crinipes, has
come to light. For this reason a re-assesment of generic delimitations in this group
was undertaken by Hubbard with the result that the South African species was excluded
from Crinipes and is here described as a distinct genus.
An investigation of the leaf anatomy of Styppeiocliloa gynoglossa has shown that
the genus does not belong in the Eragrosteae (in which, as a member of the genus
Crinipes , it had been placed) but is more closely allied to Danthonia and its allies in the
Danthoniae. The silicified cells are dumb-bell-shaped (Fig. 4) and quite different
from the typical silicified cells of Eragrostis and its allies, which are more or less round
or square in shape and usually have accompanying cells associated with them. Moreover
the chlorenchyma is not radially arranged round the vascular bundles in a single row
of cells, as in the Eragrosteae. The leaf is of the “ Danthonoid ” types where the
chlorenchyma is more or less continuously distributed between the bundles, the motor
cells being comparatively small and not forming divisions between the vascular bundles
as in the “ Eragrostoid ” type (Fig. 4). Bicellular hairs similar to those of Danthonia
are also present (Fig. 4). It has even proved somewhat difficult to find good characters
to distinguish this species from members of Danthonia. The very reduced awns and
the rather long linear hilum together with the different facies distinguishes this genus
from typical Danthonia. The hilum in all the members of Danthonia investigated
proved to be rather short, usually narrowly elliptic in outline, or not very well defined.
None of the species investigated had a distinctly linear hilum such as possessed by
Styppeiocliloa.
Styppeiochloa gynoglossa ( Goossens ) De Winter, comb. nov.
Crinipes gynoglossa Goossens in Kew Bull. 1934: 200(1934); Chippindall in Grasses
and Pastures of S. Afr. 1 : 123 (1955). Type: Transvaal, Barberton District, Kaapsche
Hoop, Pole Evans 1022 (PRE, holo.!; K).
S. gynoglossa occurs all along the Drakensberg escarpment from the Impendhle
district in Natal northwards through Swaziland to the Eastern Transvaal and the
eastern districts of Rhodesia. At Namaacha on the border of Swaziland it is found
within the territory of Mozambique. It prefers seepage areas where the rock is overlain
137
by a thin layer ot soil. The extremely tough fibrous bases of these plants form a thick
mat so resistant that a pickhandle is easily broken in trying to lift it. Some of the stands
are several acres in extent. It is found mainly at high altitudes usually between
5,000-6,000 ft. above sea level. At Millers Falls in Swaziland and the nearby Namaacha
in Mozambique the altitudes are 4,500 and 2,000 ft. above sea level respectively. The
rainfall in the greater part of its range exceeds 880 mm p.a.
SOUTH AFRICA. Transvaal.— Ermelo: Pont s.n. (K). Barbe ton: Kaapsche Hoop, Pole
Evans 1022. Lydenburg: Steenkampsberg, 29 miles W. of Kydenburg, shallow moist seepage area
over quartzite slabs, Codd 8219. Pietersburg: Blaauberg on way to Trig. Beacon, Codd & Dyer 9071.
Pilgrim's Rest: 7 miles from Graskop on road to Vaalbank, De Winter & Codd 220 \ Mariepskop,
plateau near beacon. Van der Schijff 6344; Driekop Gorge near Graskop, De Winter & Codd 211.
Natal.— Bergville: Cathedral Peak Forest Research Station, Killick 1013; 1237; 1248; 1548;
1565; 1789. Impendhle: along streambank, Huntlev 481 ; 798. Utrecht: farm Naauwhoek, Devenish
935; 172.
SWAZILAND. Mbabane: Millers Falls, Compton 26188.
MOZAMBIQUE. Sul do Save. — Lourenco Marques: Namaacha, Myre & Carvalho 300; 1312.
RHODESIA. Melsetter: Chimanimani, rocks west of airfield, Goodier & Phipps 20.
S. gynoglossa is extremely closely allied to the Madagascan Redfieldia hitchcockii
A. Camus and the latter may on further investigation prove to belong to the genus
Styppeiochloa. It certainly does not belong to the monotypic American genus Redfieldia
Vasey, from which it differs in several important points.
B. de Winter
A New Species of Eragrostis
Eragrostis moggii De Winter , sp. nov., E. glandulosipedatae De Winter affinis, sed
spiculis minoribus et angustioribus lemmatium nervis conspicuis differt.
Gramen perenne, decumbens vel suberectum (rare erectum). Culmi geniculati,
3-multinodosi, multiramosi, saepe fasciculati. ad 90 cm longi, glabri, saepe purpurascenti.
Vaginae laxae, glabrae, marginibus pilosis, carinis costisque glanduloso-punctatis.
Ligula dense ciliata. Laminae planae vel involutae, infra glabrae supra minute scabridae.
Panicula angusta, plus minusve contracta vel patens, ramis solitariis vel binis. Spiculae
plumbeae, distincte pedicellatae, 4-11-florae, 3-7 mm longae, 1-1 -5 mm latae; pedicelli
glande annulari instructi. Glumae subequales vel nonnihil inaequales purpurascentes,
ovales, apice obtusae vel subtruncatae, 1-nervatae, carinis scabridiusculis. Lemmata
subcoriacea, nitida. late ovata, obtusa, ca 1 - 5 mm longa, 1 0 mm lata, nervis 3 con-
spicuis. Paleae anguste ovato-ellipticae, 1-3-1 -4 mm longae, bicarinatae, truncatae,
membranaceae. Lodiculae 2, carnosae, cuneiformes. Stamina 3, antheris anguste
oblongis ca 0-7 mm longis. Ovarium oblongum, stylis 2. Caryopsis nonnihil lateraliter
compressa, apice et basi subtruncata, opaca, nonnihil rugosa, ca 6 cm longa.
Type: Mocambique, Inhaca Island, Mogg 26816 (PRE, holo.; K, SRGH).
Perennial up to 90 cm high (usually 30-40 cm), decumbent, geniculate-ascending,
or erect. Culms somewhat branched to much-branched and occasionally fascicled
from the nodes, rarely prostrate and rooting from the nodes which bear dense fascicles
of culms, few- to many-noded, internodes glabrous and smooth, flushed with purple
or pallid, easily compressed. Leaf-sheaths usually shorter than the internodes, lax
and slipping from the culms leaving them bare, glabrous except for the ciliate margins,
somewhat striate, with glands on the ribs upwards, often purplish. Collar conspicuous,
yellowish, glabrous or hairy on the back, the auricles conspicuously bearded. Ligule
long ciliate. Lecf-hlades expanded or rolled, up to 3 mm wide, up to 20 cm long,
138
tapering to a fine point, glabrous and smooth below, finely ribbed above with the ribs
minutely scaberulous. Panicle narrow and contracted to effuse and open, branches
single or paired, stiff, glabrous, finely ribbed. Pedicels of the spikelets slender, stiff,
furnished with a distinct annular gland. Spikelets lead-coloured (rarely pallid), narrowly
oblong, 4-7 mm long and 1 -0-1 • 5 mm wide, 4-13-flowered, rhachilla not disarticulating
the lemmas deciduous, the paleas remaining attached to the rhachilla. Florets herma-
phrodite. Glumes 1 -nerved, subequal or somewhat unequal, broadly ovate-oblong,
obtuse, finely scaberulous on the keels, subcoriaceous, usually distinctly purplish.
Lemmas about 1 -5 mm long and about 1 -0 mm wide, broadly elliptic-ovate in outline,
very finely scabridulous, 3-nerved, nerves prominent, apex obtuse to subtruncate.
Palea equalling the lemma or slightly shorter, membranous, subhyaline, bicarinate,
margins inflexed, keels scaberulous. Lodicules truncate, cuneate, somewhat fleshy,
small. Stamens 3; anthers about 0- 7 mm long, cultrate, pale yellow. Ovary glabrous;
styles free; stigmas 2, plumose. Caryopsis somewhat laterally compressed, apex and
base subtruncate giving it a nearly square outline seen in side view, smooth, whitish-
opaque; embryo half to slightly less than half the length of the grain, brownish; hilum
punctiform, basal.
Two varieties can be distinguished.
Inflorescence effuse and open, the branches and branchlets spreading var. moggii
Inflorescence narrow and contracted the branches appressed to the rhachis var contracta
var. moggii.
Inflorescence effuse and open, the branches and branchlets spreading.
Occurs in Mozambique particularly in the vicinity of Lourenco Marques and
islands off the coast. It has also been collected in northern Zululand. It prefers sandy
soils and often occurs in open forest.
SOUTH AFRICA. Natal, — Ubombo: Mtenzi Game Reserve, Tinley 596.
MOZAMBIQUE. Sul do Save. — Lourenco Marques: Costa da Sol, Balsinhas 261; Myre 864; Inhaca
Island, Mogg 27205; 26816; 29354; 26870 (pro parte); 26746 (J); Gillard J. 34664 (J); Weintraub
J. 20469 (J); Mauve & Verdoorn 103; Bazaruto Island, Mogg 28512 (J); Masinga, Schlechter 12127.
Maputo. — Catuane, Myre & Balsinhas 585.
Inhambane. — Inharrime, De Lomos & Balsinhas 1 74 ; Velho, Gomes e Sousa 2136.
var. contracta De Winter, var. nov., inflorescentia angusta contracta, rami ramulique
rhachi adpressi.
Type: Mozambique, Inhaca Island, Mogg 26947 (PRE, holo.; K).
This variety has been recorded from the vicinity of Lourenco Marques and has
the same ecological preferences as the typical variety.
MOZAMBIQUE. Sul do Save. — Lourenco Marques: between Lourenco Marques and Costa da Sol,
Schweickerdt 1910; Inhaca Island, Mogg 26870 (pro parte); 26947; 26870; 29362; 27571; 30069
(J); 30119 (J).
The two varieties often occur together and one collection, Mogg 26870, consisted
of specimens bearing either narrow and contracted or divaricate open inflorescences.
So far the species has been collected mainly in the surroundings of Lourenco Marques.
The Zululand specimen is the only record from South Africa and shows that the species
is probably much more widely spread than the present records indicate.
This species seems to be allied to E. glandulosipedata de Winter, with which it
agrees in the annular glands on the pedicels, the glands on the leaf-sheaths and the
general configuration of the spikelets. It is, however, smaller in all the parts of the
139
spikelet and very easily distinguished. E. moggii could be confused with E. lehmanniana
which it resembles generally in habit but the latter has no glands on the pedicels and
the nerves of the lemmas are not as prominent nor are the glumes purplish.
LILIACEAE
B. de Winter
New Species of Kniphofia
Names are required for three undescribed species of Kniphofia which are included
in check-lists being compiled for the territories of Swaziland by Prof. R. H. Compton
and of Basutoland by Mrs. A. Jacot Guillarmod, respectively. The three species are
described here, together with a fourth species which was recently collected in the
Mountain Zebra National Park, Cradock District, Cape Province.
K. ritualis Code! , sp. nov., K. sarmentosae (Ait.) Kunth affinis, sed foliis marginibus
scabridulis, floribus longioribus differt.
K. sarmentosa sensu Phillips in Ann. S. Afr. Mus. 16: 294 (1917).
Planta 40-80 cm alta. Folia carinata, glauca, 40-90 cm longa, 1 • 2-2 • 4 cm lata,
margine conspicue serrulata. Inflorescentia densa, rhomboideo-cylindrica vel ovoidea,
9-14 cm longa, 4-5 cm lata, gemmis pendulis corallinis ve! rubris, floribus maturis
flavo-virentibus. Bracteae lineari-lanceolatae, acutae vel acuminatae, 10-12 mm
longae, 2 mm latae, minute denticulatae. Pedicelli 1-2 mm longi. Perianthium
subcylindricum, 2-5-3 cm longurn. Stamina inclusa vel vix exserta. Stylus 3 mm
exsertus. Capsu/a subglobosa, 7-8 mm longa. Plate 2.
Type: Basutoland, Berea District, Bitsolebe area, alt. 6,500 ft, 22/1/1957, Marais
1327 (PRE, holo.).
Plants solitary or in small groups. Leaves in 3 or 4 ranks, 8-12 per peduncle,
at first erect, later bending at the middle, soft in texture, glaucous, 40-90 cm long and
1-2-2 -4 cm broad near the base, deeply keeled; margin conspicuously and regularly
serrulate. Peduncle subequal or shorter than the leaves, 40-80 cm long. Inflorescence
rhomboid-cylindrical to ovoid, dense, 9-14 cm long and 4-5 cm in diameter; buds
and flowers pendulous; buds coral or salmon to orange-red, becoming greenish-
yellow as the flowers open. Bracts lanceolate to linear-lanceolate, tapering to the
acute or acuminate apex, 10-12 mm long and 2 mm broad; margin usually minutely
denticulate. Pedicels 1-2 mm long at flowering stage, elongating to 5 mm in the
fruiting stage. Perianth subcylindrical, 2-5-3 cm long, very slightly constricted above
the ovary, 3 mm wide at the base, increasing gradually to 5 mm wide at the mouth;
lobes ovate, rounded, 3-3-5 mm long, not spreading. Stamens 6, of two lengths;
anthers not or scarcely exserted at anthesis, later withdrawn into the perianth tube.
Style equal to the stamens at anthesis, later included or exserted by up to 3 mm.
Capsule subglobose, not markedly triquetrous, 7-8 mm long.
Common in Basutoland and the adjoining areas of the eastern Orange Free State,
mainly on sandstone formation, in shallow soil in rock crevices, in dry water courses
and on grassy slopes, at altitudes of 5,500 to 9,000 ft. The flowering time is from
January to March.
Basutoland. — Mafeteng: Malealea, Munro s.n. (cultivated in Pretoria). Maseru: Thaba Bosigo,
Van der Merwe 1156; Jordan Valley, Guillarmod 1671; Senqunyane Valley, Guillarmod 2256. Berea:
below Molimo-Nthuse, Marais 1326; Bitsolebe area, grassy slope among shrubs, Marais 1327 (PRE,
holo.); at foot of Pulane Mt., Bruce 360; Pulane area, Bruce 361 ; 362; near Mateka, Bruce 364;
Thaba Chitja, Bruce 365; Mamathes, Bruce 390. Leribe : Dieterlen 174; 174a.
Orange Free State. — Ficksburg: farm Wintershoek, near Kirklington, Codd 10540; 10541; farm
Franschhoek, 20 miles N.E. of Ficksburg, Codd 10549. Fouriesburg: 23 miles N.E. of Ficksburg,
Bruce 343; 12 miles S. of Clarens, Bruce 393; near Golden Gate, 11 miles E. of Clarens, Codd 10537.
Senekal: Goossens 838. Thabanchu: Thabanchu Mt., Roberts s.n.
140
Phillips, l.c., misidentified this species as K. sarmentosa (Ait.) Kunth, a species
largely restricted to the Roggeveld area of the Cape Province. He records the Sotho
common names as “ Leloele or Lelole ” and states that: “A decoction prepared from
this species is drunk to cure pains in the shoulders. Women, during the times 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 information could be obtained
K. sarmentosa and K. ritua/is are similar in having conspicuously long, narrow
bracts. The latter species can, however, readily be distinguished by the markedly
serrulate leaves and the longer perianth, 2-5-3 cm long, as against 2-2-5 cm long
in K. sarmentosa.
K. ritualis also shows an affinity to K. caulescens, but there are several characters
distinguishing them. K. caulescens is one of the few caulescent species in the genus
and is gregarious, forming large groups. With its markedly glaucous leaves and
attractive inflorescences grading from coral or red buds to lemon-coloured floweis, a
group in flower is an impressive sight. The perianth (2 -2-2 -4 cm) is distinctly shorter
than in K. ritualis , while the stamens are exserted by 8-13 mm at anthesis. It is
interesting to note that the natives regard K. caulescens as a charm against lightning
and frequently plant the species near their huts. Phillips, l.c., who misidentified this
species as K. tysonii Bak., also records that the natives bathe themselves with a decoction
from the plant in order to protect themselves against lightning.
K. hirsuta Codd , sp. nov., K. ritualis Codd affinis, sed foliis hirsutis differt.
Planta 45-60 cm alta. Folia carinata, erecta, 40-60 cm longa, 1-5-2 -5 cm lata,
hirsuta, margine conspicue serrulata. lnflorescentia densa, rhomboideo-cylindrica vel
ovoidea, 8-12 cm longa, 4-4-5 cm lata, gemmis pendulis corallinis vel rubris, floribus
maturis flavo-virentibus. Bracteae lineari-lanceolatae, acuminatae, 10-14 mm longae,
2 mm latae, minute denticulatae vel integrae. Pedicelli 2-3 mm longi. Perianthium
subcylindricum, 2 -2-2 -7 cm longum. Stamina inclusa vel vix exserta. Stylus 3 mm
exsertus. Capsula subglobosa, 7-9 mm longa. Plate 3.
Type: Basutoland. Berea District, Mamalapi, alt. 8,500 ft, 20/1/1957, Guillarmod
& Marais 1307 (PRE, holo.).
Plants usually solitary. Leaves in 3 or, rarely, 4 ranks, 8-12 per peduncle, eiect,
soft in texture, dull green, 40-60 cm long and 1 -5-2-5 cm broad near the base, deeply
keeled., hirsute along the nerves on both surfaces; margin conspicuously and regularly
serrulate. Peduncle equal to or longer than the leaves, 45-60 cm long. Inflorescence
shortly rhomboid-cylindrical to ovoid, dense, 8-12 cm long and 4-4-5 cm in diameter;
buds and flowers pendulous; buds orange or dull coral to salmon-pink tipped green,
becoming dull greenish-yellow as the flowers open. Bracts linear-lanceolate, tapering
to the acuminate apex, 10-14 mm long and 2 mm broad; margin minutely denticulate
to entiie. Pedicels 2-3 mm long at flowering stage, elongating to 4 mm in fruit. Perianth
subcylindrical, 2 -2-2 -7 cm long, very slightly constricted above the ovary, 3-5 mm wide
at the base, increasing gradually to 5 mm wide at the mouth; lobes ovate, rounded,
2-5 mm long, not spreading. Stamens 6, of two lengths, not or scarcely exserted at
anthesis, later withdrawn into the perianth tube. Style equal to the stamens at anthesis,
later exserted by up to 3 mm. Capsule subglobose, not markedly triquetrous, 7-9
mm long.
Found in the mountains of Basutoland on grassy slopes and stream banks, at
altitudes of 8,000 to 9,000 ft. The flowering takes place in December and January.
Basutoland. — Berea District, Mamalapi, in grassveld near rocks, alt. 8,500 ft, 20/1/1957, Guillarmod
& Marais 1307 (PRE, holo.); Guillarmod & Marais 1308; Mamalapi, December 1948, Compton 21266.
141
K. hirsuta is closely allied to K. ritua/is, having similar inflorescence, perianth
and bract characters, but the former species is unique in the genus in having pubescent
leaves. The only other species which approaches it in this respect is K. stricta Codd,
which sometimes has a few scabrid hairs on the nerves. K. hirsuta was apparently
discovered in December 1948 by Prof. R. H. Compton at Mamalapi at an altitude
of about 8,500 ft. Since then, good flowering and fruiting material was collected at
the same locality by Mrs. A. Jacot Guillarmod and Mr. W. Marais, who report that
it tends to flower earlier than K. ritua/is and to grow at higher altitudes (up to 9,000 ft),
where it grows in mountain grassland, associated with Danthonia-Erica-Cliffortia
scrub of rocky slopes or streams.
K. umbrina Codd, sp. nov., K. typhoidi Codd affinis, sed foliis conspicue carinatis,
floribus et bracteis longioribus diflert.
Planta 70-90 cm alta. Folia carinata, erecta, apice reflexa, flaccida, 45-70 cm
longa, 1-2 cm lata, margine carinaque levi. Inflorescentia densa, subcylindrica, 7-15
cm longa, 1-5-2 cm lata, gemmis patulis, floribus maturis primum patulis demum
pendulis, umbrinis vix aromaticis. Bracteae lineari-lanceolatae, acuminatae, 9-1 1 mm
longae, 1-5 mm latae, integrae. PediceUi 1-5 mm longi. Perianthium cylindricum,
8-9 mm longum. Stamina 2 mm exserta. Stylus 3 mm exsertus. Capsula ovoidea-
triquetra, 5-6 mm longa. Plate 4.
Type: Swaziland, 4 miles S. of Forbes Reef, 20/2/1951, Bruce 272 (PRE, holo.).
Plants usually in small groups. Leaves not distichous, 6-8 per peduncle, at first
erect, later reflexed, mid-green, somewhat flaccid, 45-70 cm long and 1-2 cm broad,
tapering towards a rather blunt apex, V-shaped in cross section, breaking up into a
few persistent fibres at the base; margin and keel smooth. Peduncle overtopping
the leaves, 70-90 cm tall. Inflorescence subcylindrical, tapering slightly towaids the
apex, dense, 7-15 cm long and 1 -5-2 cm in diameter; buds spreading, flowers eventually
pendulous; buds purplish-brown, flowers reddish-brown to dull red, slightly scented,
drying blackish. Bracts linear-lanceolate, acuminate, 9-11 mm long and 1-5 mm
broad; margin entire. Pedicels 0-5 mm long. Perianth cylindrical, not constricted
above the ovary, 8-9 mm long, 1 -5 mm wide at the base, increasing to 2-5 mm at the
mouth; lobes ovate, 1 mm long, not spreading. Stamens 6, of two lengths, exserted
by up to 2 mm at anthesis. Style exserted by 1-2 mm at anthesis and eventually by
3 mm. Capsule ovoid-triquetrous, 5-6 mm long.
Known from only a small area a few miles south of Forbes Reef in Swaziland,
at an altitude of about 4,500 ft, where it grows in or near grassy vleis. The flowering
time is Februaiy to early March.
Swaziland. — Four miles S. of Forbes Reef. 20/2/1951, Bruce 272 (PRE, holo.); 6/3/56, Codd 9521 ;
12 miles N. of Mbabane on Forbes Reef road, 19/2/51, Reynolds 5866; near Forbes Reef, 21/2/58,
Compton 27552; 8/2/62, Compton 31290.
With its reddish-brown, slightly scented flowers, this species shows an obvious
relationship to K. tvphoides Codd (Notosceptrum natalense Bak.), but there are several
clear distinguishing characters. The leaves are not distichous or glaucous, as in K.
typhoides, and they are distinctly keeled. In addition, the perianth is longer (4 -5-6 -5
mm long in K. typhoides ) and more pendulous, while the floral bracts are longer (4-6
mm long in K. typhoides) and are acuminate at the apex, not rounded as in K. typhoides.
In view of the restricted distribution of this species and the growing pressure of
agricultural activity, it is desirable that adequate steps be taken to ensure that it is
protected from eradication.
142
K. acraea Codd, sp. nov., K. breviflorae Bak. affinis, sed inflorescentiis densioribus:
bracteis latis ovatis differt.
Planta 50-70 cm alta. Folia nervosa, carinata, 50-70 cm longa, 1-1-5 cm lata,
margine minute denticulata. Inflorescentia densa, ovoidea vel subcylindrica, 6-12 cm
longa, 2 -2-2 -5 cm lata, gemmis patulis viridi-luteis, floribus maturis primum patulis
demum pendulis luteis. Bracteae latae ovatae, obtusae, 4-5 mm longae, 2-2-5 mm
latae, minute eroso-denticulatae. Pedicelli 1-5-2 mm longi. Perianthium breviter
cylindricum, 8-9 mm longum, 2 mm latum. Stamina 2-3 mm exserta. Stylus 3-4 mm
exsertus. Fructus non visi. Plate 5.
Type: Cape Province, Cradock District, National Mountain Zebra Park, on top
of Bankberg, alt. about 6,000 ft, 6/3/1964, Liebenberg 7120 (PRE, holo.).
Plants in small groups. Leaves about 12 per peduncle, 50-70 cm long and 1-1 -5
cm broad, folded along the midrib, nerves conspicuous; margin minutely and somewhat
distantly denticulate, keel smooth. Peduncle subequal to the leaves, 50-60 cm long.
Inflorescence dense, ovoid to subcylindrical, tapering to the apex in the young stage,
6-12 cm long and 2-2-2- 5 cm in diameter; buds greenish-yellow, spreading; mature
flowers yellow, at first spreading, later deflexed. Bracts broadly ovate, obtuse, 4-5 mm
long and 2-2-5 mm broad, chartaceous, erect and overlapping the buds in the young
stage, later reflexed; margin minutely erose-denticulate. Pedicels 1-5-2 mm long.
Perianth shortly cylindrical, not constricted above the ovary, 8-9 mm long, about 2 mm
wide at the base, widening slightly to 2-5-3 mm wide at the mouth; lobes bioadly
ovate to oblong, rounded to acute, 1 • 5-2 mm long. Stamens exserted by 2-3 mm at
anthesis. Style equal to the stamens at anthesis, eventually exserted by 3-4 mm.
Fruit not seen.
Known only from the above gathering, consisting of 3 specimens, collected on
top of the Bankberg, in a seepage area among rocks. It is a very distinctive species,
resembling K. breviflora Harv. ex Bak. in the length and colour of the flowers, but
its real relationship is not clear.
L. E. Codd
PODOCARPACEAE
The Typification of Podocarpus elongatus
The taxon was first described by Solander in Aiton’s Hortus Kewensis ed. 1, 3:
415 (1789) as Taxus elongata. It is mentioned that the plant is a native of the Cape
of Good Hope and that it was introduced in 1774 by Thomas Lucas.
Two sheets at the British Museum (Natural History) were annotated by Solander
as Taxus elongata. The first was collected by Masson at the Cape but no exact locality
or date aie indicated. As Masson undertook his last collecting trip at the Cape during
1774 the specimen could not have been gathered later than that year. Mounted on
the sheet are two branches with male cones and both represent the Cape plant.
On the second sheet three specimens are mounted, all bearing male cones. One
of them is annotated by Solander: “ Hort. Dni. Lucas Lee in Kent 1777 ”. (“ Dni.”
stands for “ Domini ” and Lee is the village in Kent in which the Lucas garden was
situated). The two other specimens are labelled in a different hand.: “ Hort. Dni.
Salisbury 1785 ” and “Hort. Dni. Vere in Kensington Gore 1788 ”. These three
specimens are referable to Podocarpus macrophyllus (Thunb.) Don subsp. maki Sieb.,
an Asiatic species which closely resembles P. elongatus.
Stapf in Flora Capensis 5, 2 (Suppl.): 8 (1933) cites (among others) the Masson
specimen and the specimens from the gardens of Lucas, Salisbury and Yere as P.
elongatus. In a typewritten report on the South African species of Widdringtonia
and Podocarpus (27/5/1921), he expresses the opinion that the specimen from the garden
of Lucas in the BM is the one that Solander described in the Hortus Kewensis.
143
It appears, however, that Solander included a mixture in his concept and there
is no conclusive evidence that his description of Taxus elongata is based solely or even
mainly on the Lucas specimen. The plant mentioned in the Hortus Kewensis was
introduced in 1774 whilst the “ Hort. Lucas ” specimen was collected in his garden
in 1777. As most plants were brought to England as seed or cuttings it is unlikely
that a specimen of Podocarpus which was introduced in 1774 could have borne male
cones in 1777.
The Hortus Kewensis refers to a plant in the Lucas garden which is a native of
the Cape of Good Hope, while the “ Hort. Lucas ” specimen is undoubtedly of Asiatic
origin. It thus appears that Lucas had both species in his garden. The source of the
Asiatic species is not known but the date 1774 given for the introduction of the Cape
species points to Masson who sent large botanical collections from the Cape to England
during that year. Salisbury mentions in a note in Burchell's Memoranda Botanica
(MS. in Kew Library) that he had a specimen of “ Taxus elongata ” growing in his
garden which Masson had collected at the Cape.
Gray (Journ. Arn. Arb. 34: 166, 1953) pointed out that the specimens from the
gardens of Lucas, Salisbury and Vere in the BM do not represent the Cape species
and argued that the Lucas specimen can thus not be taken as the type of Podocarpus
elongatus. She proposed two “ lectotypes ” : as male type she chose the Thunberg
specimen in UPS and as female type the Sonnerat specimen in P. Obviously this
treatment cannot be accepted. Apart from the fact that a name cannot be typified
by more than one specimen, neither of the specimens chosen by Gray beat s Solander’s
handwriting and it is unlikely that he saw either of them. Instead, the Masson specimen
in the BM is proposed as lectotype of the name Podocarpus elongatus. The specimen
was identified by Solander and although no specimens are explicitly mentioned in
the Hortus Kewensis there is every reason to assume that the author saw it before
publishing the description. The only alternative lectotype is one of the specimens
from the gardens of Lucas, Salisbury or Vere and this would involve transference of
the name P. elongatus to the Asiatic species presently known as P. macrophyllus subsp.
maki whilst the Cape plant would require a new name.
I am grateful to Dr. L. Codd of the Botanical Research Institute, Pretoria, and to
Mr. A. A. Bullock of Kew who provided valuable advice.
O.A. LE STNER
STERCULIACEAE
A New Species of Dombeya and a New Variety of Dombeya rotundifolia
Dombeya autumnalis Verdoorn, sp. nov. D. rotundifolia (Hochst.) Planch, affinis
sed plantis plerumque fruticosis caulibus gracilioribus, cortice levioie, floribus et foliis
plerumque in auctumno coaetaneis, ovario breve stellato-pubescente nunquam setosi
diffeit.
Frutex vel arbuscula, 1-6-5 m altus, caulibus pluribus gracilibus ad 5 cm diam.,
cortice leviore. Ramuli graciles, foliati, breve stellato-pubescentes. Folia suborbiculata,
1-5-5 cm longa, 1-5 cm lata, utrinque dense vel sparse stellato-pubescentia, subtiliter
cienato-dentata, subtus obscure reticulato-nervosa petiolis gracilibus 2-12 cm longis
stellato-pubescentibus. Cymae in axillis foliorum super iorum, quam folia subtendentia
longiores vel breviores, pedunculis gracilibus c 2-5 cm longis stellato-pubescentibus,
pedicellis giacilissimis 1-1-5 cm longis stellato-pubescentibus. Bracteolcie lineari-
naviculares, c. 2-5 mm longae, utrinque stellato-pubescentes. Calyx basi rotundatus
obtusus, 0-5 mm connatus, lobis reflexis c. 4 mm longis 2 mm latis, extus stellato-
pubescentibus. Petala alba, c 7 mm longa, 5 mm lata, persistentia. Stamina basi
144
connata in tubum c 1 mm longum, filamentis ad 2-5 mm longis, antheris c 1 mm longis
stami aodiis c 5 mm longis. Ovarium globosum, c 3 mm diam., breve stellato-tomentosum
(non setosum). Stylus glabrescens vel minute stellato-pubescens, c 2 mm longus in
ramos 3 c 2 mm longos divisus. Capsula c 5 mm diam., stellato-tomentosa.
Type: Lydenburg, Penge Mine, Verdoorn 2470 (PRE, holo.).
Shrub or small tree 1-5-5 m tall; stems several, slender, up to about 5 cm diam.;
bark compaiatively smooth, stringy. Branchlets slender, leafy, new growth shortly
stellate pubescent, the pubescence formed of short spreading hairs from a scaly base.
Stipules caducous, deltoid or linear-subulate from a deltoid base, densely pubescent.
Leaves more or less orbicular, on flowering branches 1-5-5 cm long, 1-5 cm bioad,
densely to sparesely stellate-pubescent on both surfaces, finely crenate-dentate on the
margins, reticulate veins rather obscure beneath; petiole slender, 2-12 cm long, stellate-
pubescent. Cymes in the axils of the upper leaves, often overtopping them, sometimes
shorter; peduncles slender about 2-5 cm long, pubescent; pedicels very slender 1-1-5
cm long, pubescent. Bracts 7 or 2 at the calyx-base, 1 further down or all 3 scattered
on the pedicel, linear-navicular, about 2-5 mm long pubescent on both surfaces. Calyx
rounded at the base, united for 0- 5 mm; lobes reflexed about 4 mm long, 2 mm broad,
dorsally stellate-pubescent. Petals persistent, white turning cinnamon rufous with
age, about 7 mm long, 5 mm broad. Stamens united at the base for more oi less 1 mm;
filaments with the longest about 2-5 mm long; anthers up to 1 mm long; staminodes
about 5 mm long. Ovary globose about 3 mm diam. shortly stellate-tomentose (not
setose); style glabrous or minutely stellate pubescent, about 2 mm long, branches
about 2 mm long; ovules 2 in a cell. Capsule about 5 mm diam., stellate-tomentose.
Recorded from the eastern Transvaal in mountainous country on mesophytic,
well-wooded slopes among rocks and in riverine bush.
Transvaal. — Lydenburg: Abel Erasmus Pass, Schlieben & Strey 8387; Codd 10027; near Penge
Mine, Codd & Dyer 7737; Codd & Verdoorn 10488; Repton 5936 (partly); Verdoorn 2470; near
Weltevreden Asbestos Mine, Verdoorn 2471. Letaba: Dublin Mine, Miller 4271.
The only other species in southern Africa, with suborbicular leaves is D. rotundifolia
which has a wide distribution stretching from Natal and the Tiansvaal northwards
into east tropical Africa and westwards through Bechuanaland to South West Africa.
D. autumnalis differs from D. rotundifolia (as it occurs in the eastern and central areas
of southern Africa) in that it flowers in late summer and autumn (with its leaves) while
D. rotundifolia flowers in spring on more or less leafles branches. From all forms
including those in South West Africa, our species differs mainly in that its slender
stems do not develope a lough bark. If collector’s notes about bark are lacking on
a specimen this feature can be determined by the examination of a cut bianchlet,
because the cortex in D. autumnalis appears as a thin, solid outer layer while in D.
rotundifolia it is a thicker, porous layer which evidently gives rise to the rough bark
on the old stems. A further diagnostic feature is that the pubescence on the ovary
of D. rotundifolia, throughout the length and breadth of its distribution, is setose as
well as stellate-tomentose, the bristle-like hairs standing more or less erect, and reaching
a length of up to 1 mm while in D. autumnalis the ovary is stellate-tomentose and not
setose as well. Other small differences can be found such as the distinctly reticulate
veining on the under surface of the leaf in D. rotundifolia while on our species the
reticular veining, if any, is obscure.
Dombeya rotundifolia ( Hochst .) Planch, var. velutina Verdoorn, var. nov., a typica
ovario breve stellato-tomentoso non setoso foliis utrinque velutino-tomentosis differt
Frutex elatus ad 5 m altus caulibus pluribus virgatis, cortice rugosa, ramulis
molliter et breve tomentosis Folia suborbiculata vel oblongo-orbiculata, 3-9 cm
longa, 2 -5-9 -5 cm lata, utrinque velutino-tomentosa, crenato-dentata, cordata, 5-7
145
palmato-nervosa, petiolis 1-2 cm longis molliter tomentosis. Cymae ramulis lateialibus
et terminalibus confertae, pedunculis 10-15 mm longis, pedicellis 7-10 mm longis.
Calyx c. 6 mm longus, extus dense et breve tomentosus. Petala c. 7 mm longa. Stamina
basi in tubum brevissimum connata, 0-5 mm longum; filamenta c. 2-5 mm longa;
staminodia c.5 mm longa. Ovarium dense et breve stellato-tomentosum (non setosum).
Stylus stellato-pubescens.
Type: South West Afiica. Rehoboth, Buellsport, Aub Schlucht, Strey 2010 (PRE,
holo., BOL, NBG).
Tall shrubs with several virgate stems, up to 5 m tall; bark rough, new growth
softly and shortly tomentose. Leaves suborbicular oi broadly oblong-orbicular,
broadest in the upper half 3-9 cm long, 2 -5-9 -5 cm broad, velvetty tomentose on
both surfaces (the tomentum made up of minute stellate-pubescent scales, the hairs
short and silky), crenate dentate, cordate at the base, palmately 5-7-nerved ; petiole 1-2 cm
long, softly tomentose. Cymes crowded on lateral and terminal branchlets; peduncles
10-15 mm long; pedicels 7—10 mm long. Calyx about 6 mm long, densely and shortly
tomentose without. Petals about 7 mm long. Stamens united at the base for about
0-5 mm. filaments unequal lengths, about 2-5 mm long, staminod.es about 5 mm long.
Ovary densely and shortly stellate-tomentose (not setose); style 2-3 mm long, stellate-
pubescent.
Occurs along the banks of the permanent stream at Aub Schlucht in the Naukluft
Mountains, South West Africa.
S.W.A. — Rehoboth: Naukluft Mountains, farm Buellsport, Aub Schlucht, Strey 2010; 2328;
Tolken & Hardy 666.
In the common and widespread species, D. rotundifolia, three features, in the
gross morphology, have been found to be constant and of diagnostic value. They
are the suborbicular leaves, the rough bark and the pubscence on the ovary which
is stellate-tomentose and setose. Microscopically this pubescence consists of stellate
and tufted hairs, the tufted being in the upper part of the ovary giving the bristly or
setose appearance. D. rotundifolia thus characterised occurs widespread in South
West Africa although not in the optimum form found so commonly in the eastern
and central regions of tropical and subtropical Africa. In South West Africa it generally
occurs as a group of low bushy shrubs and only where conditions are favourable does
it grow up into the characteristic, rough-barked tree.
In a valley in the Naukluft Mountains, Rehoboth district, along a permanent
stream the variety here described, grows. It has the suborbicular leaves and rough
bark of typical D. rotundifolia but the ovary is constantly shortly stellate tomentose,
not setose. Investigation has shown that the habit of these plants also differs from
that of typical D. rotundifolia as found in the rest of South West Africa. The plants
along the stream are all tall, several-stemmed shrubs. These two differences together
with a third, the velvetty texture of the pubescence seem to justify at least the varietal
rank here given this form of D. rotundifolia.
I. C. Verdoorn
147
Plate \.—Widdringtonia cedarbergensis Marsh (Luckhoff in PRE 29824, holotvpe)
7230691
148
Plate 2. — Kniphofia ritualis Codd ( Marais 1327, holotype, in PRE)
149
?P-gJ3363 — JM7-&— 144-Hon *. ^
NATIONAL HERBARIUM,
PRETORIA.
DISTRICT /,
Uotlrrlnr
A Uitude
Dau, . //. y
Same
fs - . ,7
So. /$c /
Vem Same
Isocalittf
Desert jtt i
Plate 3 .—Kniphofia hirsute Codd ( Guillarmod
& Marais 1307, holotype, in PRE)
150
Plate 4. — Kniphofia umbrina Codd ( Bruce 272, holotype, in PRE)
151
Plate 5 .—Kniphofia acraea Codd ( Liebenberg 7120, holotype, in PRE)
7230691
153
Bothalia. 9, 1: 153-156.
The Genus Ceramium in South Africa
by
R. H. Simons
Abstract
A revision of the genus Ceramium in South Africa has been undertaken. The morphology of
the genus is discussed, and a key to the ten species which are recognised is provided. The following
new taxa and new combinations are published: Ceramium arenarium, C. papenfussianum, C. diaphanum
(Lightf.) Roth var. capense, — var. pulchellum (Kuetz.) (Hormoceras pulchellum Kuetz.), C. centro-
ceratiforme.
Ceramium capense Kuetz., which was transferred to Microcladia by Papenfuss, is regarded as
belonging to Ceramium.
Introduction
The Delimitation of Ceramium and Allied Genera
The genus Ceramium (family: Ceramiaceae), although used in a restricted sense
for over one hundred years, is still not fully understood. Dixon (1960) studied the
British forms recently and his observations together with those of Feldmann-Mazoyer
(1940) are detailed and informative. But certain problems, some highlighted by South
African forms, remain and, although the present author has attempted to resolve some
of them, it cannot yet be claimed that the genus Ceramium is satisfactorily delimited.
One of these problems concerns the attachment of tetrasporangia. According
to Dixon (1960), and some other authors, they arise from unspecified cells of the cortex.
The inference is that any cortical cell may act as the supporting cell of a tetrasporangium.
The present author was able to trace the origin of many tetrasporangia in five out of
the ten species listed in this paper. In every case they were attached to pericentral cells.
No other supporting cell was indicated. Where the actual derivation was not directly
observable the tetrasporangia always occupied positions such that attachment to the
pericentrals was not only feasible but highly likely. If this derivation is as constant as
the present author suggests it might be then it must be regarded as of fundamental
importance and any interpretation of Ceramium should take it into account. Obviously
more information is required on this subject before its value can be properly assessed.
The occurrence in South Africa of Ceramium centroceratiforme Simons, sp. nov.,
raises the second problem dealt with here, namely, that of the status of the genus
Centroceras Kuetz. From time to time species such as this one have been described
which link Ceramium and Centroceras, and opinions have been divided as to their
autonomy. Most authors have accepted them as different although some have done
so without much conviction. The tendency of Ceramium centroceratiforme Simons
to have an orderly arrangement of cortical cells but at the same time to have distinct
genicules somewhere along its length puzzled the author because the species appeared
to belong to both genera equally well. A solution seemed to be to regard Centroceras
as synonymous with Ceramium as some other authors have done. It was decided
instead to make a closer study of the South African form of Centroceras clavulatum
(Ag.) Mont., the type species of Centroceras, to discover whether there was any justifi-
cation for its separation from Ceramium. Certain facts emerged from this investigation
which the writer regards as notable.
154
First, in C. clavulatum the regular arrangement of basipetal secondary cortical
cells in a single layer is the result of their exclusively transverse segmentation. The
resultant cortical files of any single articulation develop equally by what appears to be
simultaneous segmentation of their apical cells. In species of Ceramium, on the other
hand, there is no notable regularity of segmentation of cortical cells either in time or
space: apical cells of the laterals of limited growth divide transversely, longitudinally
and obliquely. Secondly, in C. clavulatum acropetal development of cortex is virtually
absent; instead there is almost radial development to form a girdle of tissue raised
slightly from the surface of the thallus at every node. As a result of this, all genicules
retain their individuality and there is no confluence of cortical tissue from adjacent
axial cells. In contrast to this, completely corticated species of Ceramium have a tissue
which is continuous between successive nodes. Thirdly, in male plants of C. clavulatum,
spermatangia were found confined to the terminal part of each fertile articulation,
i.e. on the radially developed girdle of tissue at each node.
South African species of Ceramium bearing spermatangia have not been found
by the writer, but Dixon (1960) and others have reported on various other species
in which they found spermatangia scattered haphazardly over the surface of the cortex.
Of the characteristics of Centroceras listed above, the most significant seems to be
that relating to the development of the basipetal cortex of each articulation. The
contrast of the latter with what occurs in species of Ceramium is sufficient, in the present
author’s opinion, to justify the separate generic status of Centroceras Kuetz., as typified
by C. clavulatum (Ag.) Mont. The other differences seem to corroborate this view.
As to the affinities of C. centroceratiforme Simons: its cortical development is like
that of species of Ceramium, to which genus it, therefore, undoubtedly belongs.
A third problem has resulted from the recognition by Papenfuss (1940) of South
African material as a species of Microcladia. The main reason for doing this seems
to have been a well developed multi-layered cortex. Feldmann-Mazoyer (1940),
writing of Microcladia , claimed that the genus was little understood and contained
unrelated species. She suggested that the only justification for removing M. glandulosa
(Soland.) Grev. to a separate genus was its unique arrangement of cells of different
sizes in the innermost layer of the cortex. The latter characteristic leads to a flattened
bilateral thallus. The present author’s study of Microcladia capensis (Kuetz.) Papenf.,
the South African species referred to above, revealed that the thallus is not bilateral
nor does it possess any other characteristic to exclude it from Ceramium. This species
is dealt with in this paper as C. capense Kuetz.*
Some Morphological Features of South African Species of Ceramium
Commenting on the glandular cells so often found in species of Ceramium, Dixon
(1960) stated that these cells are rather short-lived. Tt is interesting to note, therefore,
that similar cells in one or two South African species are sometimes very persistent.
In C. glanduliferum Kylin, for example, they are often obtrusive and so long-lived
that the specific epithet was suggested by their presence. Glandular cells in C. papen-
fussianum Simons, sp. nov., were found to be almost as persistent but not nearly so
obvious as those of the previous species.
* Since writing this article the author has seen the paper by Max Hommersand on the ’'Morphology
and Classification of some Ceramiaceae and Rhodomelaceae" (University of California Publications in
Botany Vol. 35, No. 2, pp. 165-366, published in 1963). In this paper Hommersand discusses the
genera Ceramium , Microcladia and Centroceras ; he upholds the separate status of these three.
Microcladia produces upwardly and outwardly directed cortical filaments only which entirely
enclose the axial cells. This interpretation makes no mention of bilateral symmetry but does not
affect my placing of C. capense in the genus Ceramium.
Hommersand’s observations on the derivation of cortical filaments in Centroceras are essentially
similar to mine and are apparently sufficient reason for a generic separation.
Ceramium poeppigiana Grunow, removed to the genus Rheinboldiella and not considered in my
paper, is returned by Hommersand to the genus Ceramium.
155
Several South African species have extremely thick-walled axial cells. The mature
axial cells of C. papenfussianum Simons (Fig. 2: lc), for instance, have such thick
lamellate walls that the central lacunae are very narrow. The thickening appears to
develop centripetally in clearly demarcated layers on the inner faces of the cell walls,
and the result is a likeness to the annual rings of higher plants. This characterictic
no doubt leads to the comparative rigidity of these diaphanous plants.
Intercellular hyphae ramifying within the cortex of C. planum Kuetz. have been
seen in no other South African species. Their presence adds greatly to the girth of
an already thick cortex in the mature parts and renders the thallus very cartilaginous
and difficult to dissect when fresh. The origin of these hyphal filaments was not traced,
but they appeared sometimes to arise from a cingula-like thread of similar material
encircling a median constriction of an axial cell. Attachment of this cingula to the
axial cell was at times evident. Perhaps this species belongs to Campylaephora.
Intracellular hyphae ( Thylles of Feldmann-Mazoyer, 1940), such as those described
by Dixon (1960), have possibly been seen in only one South African species. However,
they were not entirely convincing.
Terminology of Morphological Features
The joints between adjacent axial cells are called nodes. The cortex consists of,
primarily, the pericentral cells differentiated at the distal end of every axial cell, and,
secondarily, the laterals of limited growth remaining closely appressed to the axial
cells. The laterals of limited growth arise by segmentation of apical cells cut off acro-
and basi-petally from the pericentrals. A cortical zone of common origin is referred to
as a genicule and uncorticated zones are inter genicules. Unless otherwise mentioned,
length is the dimension in the plane of the long axis of the plant, and width is measuied
at right angles to this.
CERAMIUM
Ceramium Roth in Cat. Bot. 1: 148 (1797).
Dictiderma Bonnemaison (1822).
Hormoceras Kuetz. (1841).
Gongroceras Kuetz. (1841).
Echinoceras Kuetz. (1841).
Acanthoceras Kuetz. (1841).
Chaetoceras Kuetz. ( 1 847), non Chaetoceros Ehr. ( 1 844), nee Chaetoceras Kuetz* (1849).
Trichoceras Kuetz. (1849).
Celeceras Kuetz. (1849).
Pteroceras Kuetz. (1849).
Ceramothamnion Richards (1901).
Plants of pseudodichotomously or irregularly branched uniaxial filaments, partly
or completely corticated by closely appressed cells in one or more layers, radially
symmetrical in transverse section, cells of inner layers larger than those of outer, arising
as laterals of limited growth from pericentrals, and cut off aero- and basi-petally by
random segmentation. Pericentrals attached to, and forming a ring around the distal
end of every axial cell. Tetrasporangia immersed in or exserted from cortex in vicinity
of nodes. Cystocarps naked, subtended by one or more involucral branches. Sper-
matangia scattered over surface of cortex (not seen in South African species).
Without any implication as to primitiveness or derivation, the species listed here
are arranged in a series decided, primarly, by the extent of tetrasporangial immersion
in the cortex and, secondarily, by the degree of cortication.
156
Genicules distinct in lower part of plant:
Intergenicules below equal to or longer than genicules:
Tetrasporangia unilateral or whorled, exserted 1. C. glanduliferum
Tetrasporangia whorled, partly or entirely immersed:
Tetrasporangia immersed partly:
Pericentrals at bases of genicules 3. C. tenerrimum
Pericentrals median in genicules; genicules contiguous in upper part of filaments
4. C. arenarium
Tetrasporangia immersed entirely:
Tetrasporangia patent through outer cortex; pericentrals frequently exposed and cortical
cells often transversely elongated 5. C. papenfussianum
Tetrasporangia not visible through outer cortex; pericentrals also occluded:
Plants flaccid, branches ultimately incurved strongly 7 (a) C. diaphanum var. capense
Plants setaceous, ultimately branches almost straight; apical cells patent 7 (, b ) C. diaphanum
var. pulchellum
Intergenicules all shorter than genicules:
Cortical cells randomly disposed 6. C. atrorubescens
Cortical cells somewhat ordered into longitudinal and transverse rows in lower parts of
genicules 9. C. centroceratiforme
Genicules not distinct in lower parts of plant:
Genicules distinct above; branch apices digitate 2. C. planum
Genicules nowhere distinct:
Branching more or less alternate:
Branching pinnate, racemo-corymbose 2. C. planum
Branching pseudodichotomous, alternate; adventitious branches frequently secund
10. C. obsoletum
Branching on all sides, profuse 8. C. capense
1. C. glanduliferum Kylin in Lunds Univ. Arsskr. N.F. Avd. 2, 34, 8: 14, fig.
7A-B (1938).
Holdfast : prostrate filaments with distinct but weakly developed genicules of one
half or one third the length of intergenicules; genicules often with single whorl of
large terminal, exserted glandular cells; rhizoids numerous at nodes. Erect system :
filaments up to 1-5 cm high, pseudodichotomously branched, adventitious branches
rare, 180 p wide at base tapering gradually to about 60 p terminally with slightly
incurved apices; axial cells up to one-and-a-half times longer than broad; genicules
shortly [separated above but widely so below, often up to twice as wide as long, some-
times even shorter and then pericentrals often at bases, cells angular minute (about
10 p X 5 p) most often with much larger and very conspicuous ovate (up to 30 p X 18/*),
exserted glandular cells in an apical whorl; number of pericentrals variable, up to nine,
often superficially visible through the cortex, more or less spherical; tetrasporangia
totally exserted. mostly unilateral on inner faces of upper dichotomies; cystocarps
not seen. Plate 1: 1.
Cape. — Mossel Bay: Tergniet, Simons 634. Knysna: Beacon Isle, Isaac B.942. East London:
Ecol. Sttrv* L66. Willowvale: Dwessa, Isaac B. 1013. Elliotdale: Bomvanaland, Flanagan s.n.
(Tyson s.n. sub BOL 27476).
Natal. — Durban: Ecol. Surv. D. 100.
Typically, this species is readily identifiable by its conspicuous glandular cells
but, when the latter are absent, it is not easily recognisable. The variable position
of the pericentrals in the genicules. however, is a fairly reliable characteristic.
* The abbreviation “ Ecol. Surv .” used in citations throughout this paper refers to material in
the University of Cape Town collection from the ecological survey of marine, littoral fauna and flora
of South Africa. This survey was directed during the period 1931-1940 by Prof. T. A. Stephenson and
is now being continued under the guidance of Prof. J. H. Day, both Professors acting in their capacities
as Head of the Department of Zoology, Cape Town University.
157
2. C. planum Kuetz. in Sp. Alg. 687 (1849); Tab. Phyc. 13: 5, pi. 1 1, fig. c-d (1863);
Papenf. in J. S. Afr. Bot. 17: 177 (1952).
C. cancellation Ag. in Syst. Alg. 136 (1824); Sp. Alg. 2: 145 (1828), non C. cancel-
latum (L.) DC. in Bull. Sci. Soc. Philom. Paris 3, 94: 264 (1804-1805). C. flexuosum
(Kuetz.) Grunow in Alg. Nov. 64 (1867), non C. flexuosum Ag. in Syst. Alg. 141 (1824).
Pteroceras cancellation (Ag.) Kuetz. in Sp. Alg. 690 (1849); Tab. Phyc. 13: 8,
pi. 22, fig. a-b (1863). P. flexuosum Kuetz. in Sp. Alg. 690 (1849); Tab. Phyc. 13:
8, pi. 22, fig. c-e (1863).
Holdfast: short haustorium-like rhizoids produced at base of erect filaments
attach plants to others. Erect system: up to 15 cm high, terete, about \ mm wide;
branching almost alternate and more or less distichous, racemo-corymbose; cortication
either complete or interrupted in upper parts, multi-layered below, large-celled inner
layer often with intercellular ramifying hyphae; pericentrals six, almost spherical;
tetrasporangia exserted, frequently bilateral, or hemicyclic, attached to pericentrals;
cystocarps surrounded by five or six involucral branches. Plate 1 : 2.
Cape. — Namaqualand: Port Nolloth, Simons 273; 331; Ecol. Surv. N. 19; Strong s.n. (Tyson
s.n. sub BOL 27435); Tyson s.n. sub BOL 27431; Buffels River, Ecol. Surv. BR2H. Clanwilliam:
Lamberts Bay, Ecol. Surv. B50. Malmesbury: Steenberg’s Cove, Ecol. Surv. C4E; Saldanha Bay,
Simons 594; Dassen Island, Tyson s.n. sub BOL 27429. Peninsula: Table Bay, Tyson s.n. sub BOL
21 All to 27428 incl., 27433, 27434, 27436, to 27448 incl.; Oudekraal, Ecol. Surv. A13; Kommetjie,
Simons 25; False Bay, Ecol. Surve. F. 115 in part sub Simons 655.
The numerous pinnate branches, arranged more or less distichously on the main
filaments, make this plant unique amongst the South African Ceramium species,
separating it clearly from C. capense Kuetz. especially. These two species have cortices
which at times are almost indistinguishable superficially. In longitudinal section their
thalli can be distinguished because the inner cells of C. capense are all more or less
elongated longitudinally whereas those of C. planum are more polygonal, those in the
middle of an articulation being more or less ovate and orientated so that their long
axes are radial to the adjacent axial cell. The ramifying hyphae present in C. planum
have not been seen in C. capense.
3. C. tenerrimum {Mart.) Okam. in Icon. Jap. Alg. 4: 112, pi. 179 fig. 1-7 (1921);
Feldmann-Mazoyer in Ceramiacees de la Mediterranee 289 (1940).
Hormoceras tenerrimum Mart, in Preuss. Exped. ost-Asien, Bot. Teil: Tange: 146,
pi. 13, fig. 2 (1866).
Holdfast: prostrate lower parts of erect filaments with distinct very short genicules;
numerous rhizoids attach plant to rock or other algae. Erect system: up to 2 cm high
and 200 p thick, branching pseudodichotomous, ultimate branches inrolled ; adventitious
branches occasional; genicules distinct throughout, short, up to three times wider
than long; pericentrals up to seven at bases of genicules; glandular cells present,
inconspicuous; tetrasporangia in single whorls on genicules of penultimate branches,
apically exserted, attached to pericentrals; cystocarps of three or four gonimolobes
subtended by about five branches up to five times longer than height of gonimolobes.
Fig. 1: 1; Plate 1: 3.
Cape. — Namaqualand: Port Nolloth, Simons 303; Hondeklip Bay, Ecol. Surv. F1B5B; Groen
River, Ecol. Surv. GR4F. Clanwilliam: Lamberts Bay, Ecol. Surv. B70. Malmesbury: Ysterfontein,
Ecol. Surv. YZ1F. Caledon: Flermanus, Simons 539; 540; F.ansk aal, Simons 647. Knysna:
Robberg, Isaac B. 788.
Although present among the older collections, this species was overlooked by
other workers. The specimens cited here compare well with one from the Mediter-
ranean supplied to the present author by Mme Feldmann. The most striking thing
about this species, compared with other South African species, is the basal position
of the pericentrals in almost every genicule. Cell rows initiated basipetally seldom
project below the bases of their pericentrals and are generally only one cell long.
158
Fig. 1. — 1, Ceramium tenerrimum, surface view of a genicule; la, longitudinal section through a
genicule; lb, transverse section through the pericentrals of a tetrasporiferous genicule 2, C.
arenarium, longitudinal section through thallus near the apex of a branch; 2a, longitu-
dinal section through a genicule of the mature portion of a thallus; 2b, transverse sec-
tion through the pericentrals of a genicule.
a.c. — axial cell; c. — cortical cell; g.c. — glandular cell; p. — pit connection; p.c. — pericentral
cell; t. — tetrasporangium.
159
Besides Japan, the Mediterranean and the Cape, there appears to be only one
other recorded locality for this species, viz. N. Pacific (Dawson, 1962, p. 49). Such
disjunctions seem strange and it is probable that C. tenerrimum will be found elsewhere.
4. C. arenarium Simons, sp. nov.
Thallus usque ad 3 cm altus, pileum formans, inferne circa 150 p latus pro parte
repens cum rhizoidibus multis ex geniculis excurrentibus, sursum quoque ramorum
gradatim gracilior; rami ultimi circinati; genicula superne contigua, inferne multo
distincta intergeniculis quam genicula triplo longioribus; cellulae pericentrales c. 7
cellulis corticalibus occlusae, plus minusve in medio geniculorum dispositae; tetras-
porangia in ramulis superioribus monoverticillata dimidio superiore exserto; cysto-
carpia subapicalia, geminata ramulis nunc paucis nunc saepius quinque vel sex incurvatis
involucrata. Fig. 1: 2; Plate 1: 4.
Type: Namaqualand, Port Nolloth, Isaac B.676 sub Simons 632 (PRE, hole., spirit
material only).
Holdfast : prostrate basal filaments with many rhizoids at each genicule entwining
neighbouring filaments and attaching plant to substratum; genicules distinct, half as
long as wide; intergenicules up to four times longer but of same width as genicules.
Erect system : up to 3 cm high, filaments dichotomously branched, frequently producing
adventitious branches in the axils in a plane at right angles to dichotomies, 150 y thick
or more, narrowing slightly at each dichotomy to about 100 y subapically, ultimately
circinately inrolled; pericentrals about seven completely obscured in the mature
parts, more or less median in genicule; genicules contiguous in upper half of filaments,
separated below and then about one-and-a-half times wider than long, intergenicules
becoming progressively longer abapically until about three times longer than genicules;
tetrasporangia in a whorl at each of the upper genicules, upper half exserted when
mature; cystocarps subapical, two gonimolobes, involucral branches up to six, some-
times absent.
South West Africa. — Luderitz, Simons 432.
Cape. — Namaqualand: Port Nolloth, Isaac B. 662; B. 676; Simons 632. Clanwilliam: Lamberts
Bay, Ecol. Surv. B. 13. Penin ula: Table Bay, Simons 117.
The characteristic circinate apices, the contiguous genicules above and the more
or less median pericentrals separate this species from C. tenerrimum (Mart.) Okam.
with which it shares the habit of partially exserted tetrasporangia. The epithet chosen
for this species is suggested by its sandy habitat amongst other low-growing turf-
forming algae.
Transverse sections of axial cells frequently showed thick-walled cellular inclusions.
Their true shape could not be deduced nor could their origin because longitudinal
sections provided no evidence of their presence. There appeared to be seven such
bodies circular in outline and closely appressed to the inner wall of the axial cells in
which they were seen. Their diameter was approximately equal to half the greatest
dimension of the pericentrals. It is possible that these structures may be similar to
the intracellular filaments referred to by Dixon (1960, p. 338) and Feldmann-Mazoyer
(1940, p. 161).
5. C. papenfussianum Simons , sp. nov.
C. pulchellum sensu Kylin in Lunds. Univ. Arsskr. N.F.Avd.2, 34, 8: 14 fig. 7C-E
(1938); non Hormoceras pulchellum Kuetz. (=C. diaphanum war. pulchellum (Kuetz.)
Simons, comb, et stat. nov., see p. 162).
160
Plantae parvae usque ad 2-5 cm altae, in aliis algis habitae, ramosissimae paene
fastigiatae, ramis supra hamatis; genicula usque ad 150 p. lata, omnibus praeter
terminalia distinctis; cellulae pericentrales plerumque 6 patentes; cellulae corticales
angulatae, basi-petalibus praecipue in geniculis superioribus saepe transverse elongatis;
tetrasporangia monoverticillata, immersa, patenti; cystocarpia ramulis 5 vel 6 involu-
crata. Fig. 2: Plate 2: 1.
Fig. 2. — 1, Cevamium papenfussianum, surface view of two genicules; la, transverse section through
the pericentrals of a mature genicule; lb, longitudinal section through a tetrasporiferous
genicule showing attachment of tetrasporangium to pericentral cell; lc, transverse section
through intergenicule showing curious wall thickening.
a.c. — axial cell; c. — cortical cell; p.c. — pericentral cell; t. — tetrasporangium; t.p.c. — tetras-
poriferous pericentral cell.
Type: Peninsula, Sea Point, Simons 694 (PRE, holo.; iso. in spirit collection).
Holdfast : prostrate lower filaments with widely spaced genicules from which
numerous rhizoids arise to attach plant to others. Erect system', filaments up to 3 cm
high, branching pseudodichotomous, almost fastigiate, terminal branches more or less
straight but hooked inwards at the tips; adventitious branches absent; genicules distinct
except terminally, widely separated below, about as wide as long; intergenicules up
to four times longer than wide; pericentrals usually six, median in genicules, often
clearly visible between outer smaller cortical cells; cortical cells angled, basipetal
cells frequently transversely elongated especially in upper genicules; glandular cells
sometimes conspicuous as an apical whorl on genicules, distinctly angled, and of same
size as other cortical cells; tetrasporangia whorled, subapical, immersed but patent,
as high as parent genicules, attached to pericentrals; cystocarps surrounded by five
or six involucral branches up to twice as high as gonimolobes.
Cape. — Namaqualand: Port Nolloth, Simons 275; 302. Malmesbury: Paternoster, Simons 482;
Ecol. Surv. P10A; Saldanha Bay, Simons 571; 572; 576; 581; 586; 587; 597; 609; 611; 619.
Peninsula: Sea Point, Simons 694; Levyns 0170 ( Tyson s.n. sub BOL 27470); St. James (False Bay),
Ecol. Surv. F115. Caledon: Hermanus, Simons 538; 541; 542; 545. Mossel Bay: Tergniet, Isaac
B. 949; B. 950. Knysna: Robberg, Isaac B. 806. Port Elizabeth, Ecol. Surv. E75. East London,
Ecol. Surv. L66. Komga: Cape Morgan, Flanagan 111 ( Tyson s.n. sub BOL 27471). Port St. Johns,
Isaac B. 834. Locality not stated: Flanagan s.n. ( Tyson s.n. sub BOL 27460-27463 inch). Collector
unknown ( Tyson s.n. sub BOL 27467; 27468).
161
C. papenfussianum is common on both the west and south coasts of the Cape.
The occurrence of the transversely elongated cortical cells caused earlier workers to
assign this species to C. gracillimum Grift', et Harv., but these two species differ in the
positions of their tetrasporangia, those of C. papenfussianum being immersed while
those of C. gracillimum are exserted. A more recent confusion has been that of Kylin
(1938) who misidentifiea this species as Hormoceras pulchellum Kuetz. The following
characteristics of H. pulchellum Kuetz. can be deduced from Kutzing’s (1849, 1862)
description and/or figures: first, the patent and somewhat diverging apical cells;
secondly, the completely obscured pericentral cells; thirdly, the presence of many
and often opposite adventitious branches. None of these characteristics is possessed
by C. papenfussianum.
J. G. Agardh (1851) assigned H. pulchellum Kuetz. to Ceramium diaphanum (Lightf.)
Roth because he could find no essential difference between the two. The present author
agrees with Agardh to the point of including H. pulchellum Kuetz. in the species C.
diaphanum (Lightf.) Roth, but considers it sufficiently different from the typical form
to be treated as a separate variety. In this paper H. pulchellum Kuetz. is placed as
C. diaphanum var. pulchellum (Kuetz.) Simons, comb. et. stat. nov. (p. 162).
The presence of the ring of glandular cells sometimes seen girdling the upper
margin of the genicules of C. papenfussianum might confuse the observer but, unlike
the relatively large ovate glandular cells of C. glanduliferum Kylin, those of this species
are angled and no bigger than normal cortical cells.
The specific epithet for this species was chosen in honour of Prof. G. F. Papenfuss,
who has contributed so much to the knowledge of South African marine algae.
6. C. atrorubescens Kylin in Lunds Univ. Arsskr. N.F. Avd. 2, 34, 8: 15, fig. 7F
(1938).
Holdfast : prostrate lower parts of filaments with numerous rhizoids arising from
narrowly separated genicules. Erect system : up to about 3 cm high and 500 p thick
below, narrowing upwards to about 100 p: branching pseudodichotomous, adventitious
branches at right angles in dichotomies, often opposite; branches ultimately somewhat
incurved; genicules variable in length, somewhat wider than long and scarcely separated
above, up to one-and-a-half times longer than wide with intergenicules varying greatly
in length below, acropetal development of cortex greater than basipetal, upper margin
of genicule undefined; pericentrals more than eight, superficially scarcely visible;
tetrasporangia immersed, patent, in a single whorl on every genicule throughout almost
entire length of erect filaments; cystocarps subapical, subtended by four or five
involucral branches somewhat longer than height of gonimolobes. Plate 1 : 5.
South West Africa. — Swakopmund, Simons 641; “ Walfisch Bay”, Cleverly s.n. ( Tyson s.n. sub
BOL 27473; 27474); Luderitz, Simons 417; Isaac B. 754.
Cape. — Namaqualand: Port Nolloth, Ecol. Surv. N42.
This species has a superficial resemblance to C. circinatum (Kuetz.) J. Ag., but
its genicules are, as it were, inverted. It is also very like C. arenarium Simons, sp.
nov. The chief difference between C. arenarium and C. atrorubescens is that the
first has genicules in which the pericentrals are always median, whereas those of C.
atrorubescens seldom are. Other differences between them are the circinate tips, the
continuous cortex above, the often half-exserted tetrasporangia and the production
of rhizoids for the greater length of the filaments of C. arenarium , all of which charac-
teristics contrast with their counterparts in C. atrorubescens.
7. C. diaphanum (Lightf.) Roth in Cat. Bot. 3: 154 (1806); Feldman-Mazoyer
in Ceramiacees de la Mediterranee 306 (1940).
1 62
Holdfast4, prostrate filaments with distinct genicules from which many rhizoids
arise; litho- or epi-phyte. Erect system: genicules distinct throughout, often wider
than intergenicules; seven pericentrals completely obscured by outer cortical tissue of
large and small cells; tetrasporangia immersed, often inconspicuous; cystocarps
subapical, subtended by involucral branches.
(a) var. capense Simons, var. nov. Plate 1 : 6.
A var. typico Feldmann-Mazoyer (1940), tetrasporangiis plus minusve geniculorum
maturorum dimidio superiori circumscriptis, ramis ultimis valde incurvatis differt.
Type: Hopefield, Schaapen Island (Saldanha Bay), Simons 614 (PRE, holo.,
spirit collection only).
Differs from var. typicum Feldmann-Mazoyer (1940) by its tetrasporangia being
more or less confined to upper half of mature genicules and its strongly incurved branch
apices.
Holdfast: prostrate filaments with distinct genicules; numerous rhizoids from
genicules attach plant to substratum. Erect system: up to 5 cm high, intergenicules
about 250 p thick in lower parts; genicules tumescent in lower parts, as much as 450 p
wide medianly and equally high, becoming relatively less swollen distally and ultimately
scarcely wider than the intergenicules at about 50 p; branching mainly pseudodichoto-
mous but adventitious branches common, branches ultimately strongly incurved;
tetrasporangia immersed and obscured in upper halves of mature genicules; cystocarps
not seen.
Known only from the single record cited above. This plant, like var. pulchellum
(Kuetz.) Simons, stat. nov., shows an arrangement of cortical cells essentially similar
to that described and figured for C. diaphanum (Lightf.) Roth by Feldmann-Mazoyer
(1940). The other shared characteristics of seven pericentrals and immersed tetras-
porangia have convinced the present author that it should be assigned to C. diaphanum.
Many varieties of this species have been recognised and these two (var. capense and
var. pulchellum) are sufficiently different from other varieties and from each other to
deserve separate varietal rank. These two differ from each other mainly in texture
and in the shapes of their ultimate branches.
( b ) var. pulchellum {Kuetz) Simons, comb, et stat. nov.
Hormoceras pulchellum Kuetz. in Sp. Alg. 676 (1849); Tab. Phyc. 12: 23, pi. 75,
fig. d-i (1862).
Ceramium pulchellum (Kuetz.) Kylin in Lunds Univ. Arsskr. N.F. Avd. 2, 34,8: 14,
fig. 7C-E (1938).
Holdfast: plant epiphytic with basal haustorial attachment and prostrate filaments
provided with rhizoids at genicules which are distinct and somewhat swollen. Erect
system: filaments up to 7 cm high, about 400 p below tapering ultimately to patent
relatively large apical cells about 10 p wide, much branched both pseudodichotomously
and adventitiously; adventitious branches arising from genicules singly, oppositely
or in occasional whorls, strongly divergent at first but soon becoming parallel with the
parent axis, branched, all tapering upwards and downwards; ultimate dichotomies
slightly incurved but apical cells remaining patent; genicules distinct throughout
with intergenicules about as long as genicules; tetrasporangia immersed in whorls in
penultimate and older branches; cystocarps surrounded by between three and six
involucral branches somewhat mnger than the height of the gonimolobes. Fig. 3: 1;
Plate 2: 2.
Cape. — Hopefield: Schaapen Island (Saldanha Bay), Simons 633. Peninsula: Table Bay, Tyson
s.n. sub BOL 27475, 27477 to 27482 inch
163
164
This variety differs from var. diaphanum by its relatively large, patent, divergent
apical cells.
All the Tyson specimens cited under this variety were designated “ C. kylinii sp.
nov. ined. G. F. Papenfuss”; a further two specimens similarly designated are of
C. atrorubescens Kylin.
8. C. capense Kuetz. in Linnaea 15: 740 (1841); Sp. Alg. 686 (1849); Tab. Phyc.
13: 3, pi. 5, fig. c-e (1863).
Microcladia capensis (Kuetz.) Papenf. in Bot. Not. 223 (1940).
Holdfast: epiphytic plant attached by basal haustorial rhizoids. Erect system-
plant up to 20 cm high and 1 mm thick near base, tapering downwards slightly, and
upwards gradually to about 50 p subapically; apically strongly incurved; main
branching pseudodichotomous, adventitious branches frequent, often in whorls, at
almost every articulation; cortication complete, multilayered when mature, outer
layer of very small cells, inner layer of large elongated or isodiametric cells; seven
pericentrals; articulations half as long as wide, cylindrical, translucent in upper parts,
lower down one-and-a-half times longer, medianly constricted; tetrasporangia whorled,
completely immersed, attached to pericentrals, often in two transverse ranks; cystocarps
clasped by five or six involucral branches. Fig. 3: 2.
South West Africa. — Luderitz: Elizabeth Bay, Simons 257.
Cape. — Namaqualand: Port Nolloth, Simons 278; 313; Isaac B. 658; Ecol. Surv. N16; N20A ;
N40; Groen River, Ecol. Surv. GR4E. Clanwilliam: Lamberts Bay, Ecol. Surv. B12; B13; B39;
B51; B52. Malmesbury: Paternoster, Ecol. Surv. P6A; Saldanha Bay, Simons 570; Isaac B. 31.
Peninsula: Table Bay, Tyson s.n. sub BOL 27485-27491 inch; Kommetjie, Levyns 0269 ( Tyson s.n.
sub BOL 27492).
This species differs from C. rubrum (Huds.) Ag. in having seven pericentrals
instead of eight (Feldmann-Mazoyer, 1940). Very similar to C. vimineum J. Ag.,
C. capense differs by its more extensive branching and rather shorter mature articu-
lations. True nodes, indicated by the point of origin of lateral branches, alternate
with false ones marked by transverse constrictions of the frond especially in the older
parts.
The description given above applies to the plant designated by Papenfuss as
Microcladia capensis, but it does not seem to apply well to Ceramium capense Kuetz.
which Papenfuss has designated as the basionym of this species. For instance, the
articulations of C. capense are all half as long as wide and the cortical cells are larger
(than those of C. rubrum). The present author is of the opinion that C. capense Kuetz.
is a form of C. obsoletum Ag., but on the available evidence hesitates to make the change
because it would necessitate the erection of a new name for M. capensis Papenf. The
missing type of C. capense must be found and re-examined before this matter can
be satisfactorily resolved.
9. C. centroceratiforme Simons, sp. nov.
Fila prostrata, intricata, pulvinum parvum formantia; genicula distincta, pellucida,
rhizoidibus numerosis: cellulae pericentrales in dimidio superiore geniculorum; corticales
sub pericentralibus plus minusve longitudinaliter vel transverse seriatae, omnibus sub-
pellucidis; fila erecta, pseudo-dichotomoso-ramosa, usque ad 2 cm alta, 300 p lata,
interdum ramis adventitiis instructa; rami ultimi suberecti, breves, divaricati et fere
triangulares; genicula superiora contigua vel fere contigua, inferiora plus minusve
distincta ; pericentrales 6 vel 7, in partibus superioribus medianae, in partibus inferioribus
in dimidio superiore geniculorum dispositae; cellulae corticales plus minusve in stratis
transverse et longitudinaliter Centrocerati similes dispositae, praecipue in partibus
165
geniculorum infra pericentrales; tetrasporangia verticillata subterminalia omnino
immersa vix manifesta, pericentralibus orta; cystocarpia bilobata uno ramulo subtensa.
Plantae subflavae vel galbanae. Fig. 3: 3; Plate 2: 3.
Type: Caledon, Hermanus, Simons 555 (PRE, holo., spirit collection only).
Holdfast : prostrate filaments matted together forming a small cushion, rhizoids
abundant at distinct genicules; pericentrals in upper half of genicules; cortical cells
below pericentrals arranged more or less in longitudinal and transverse rows, all more
or less colourless. Erect system : up to 2 cm high and 300 n thick, branching dichoto-
mous, adventitious branches occasional, ultimate branches short, more or less straight
and somewhat triangular in contour; upper genicules contiguous or almost so, lower
genicules the same or distinct and scarcely different from those of prostrate filaments,
longer than intergenicules; cortex, in vicinity of articulations and sometimes elsewhere,
two-layered; pericentrals six or seven, median in upper parts, in lower parts occur
in upper half of genicules; cortical cells arranged more or less in transverse and longi-
tudinal rows similar to those of the genus Centroceras, especially in parts of genicule
below pericentrals; tetrasporangia immersed, whorled, subterminal, scarcely discer-
nible, attached to pericentrals; cystocarps two-lobed, subtended by only one ramule.
Plant yellowish or greenish-yellow.
Cape. — Caledon: Hermanus, Simons 489; 555; 557. Knysna: Noetzie, Isaac B. 755; Robberg,
Isaac B. 939; B. 941. Willowvale: Dwessa, Isaac B. 1014.
The superficial appearance of this species is so like that of a Centroceras that
sometimes only the distinct genicules of the prostrate filaments disclose its identity.
Mostly, however, the genicules in other parts of the plant are also distinct. The author’s
reasons for assigning this species to Ceramium and not to Centroceras have been given
in the introduction to this paper.
10. C. obsoletum Ag. in Sp. Alg. 2: 145 (1828); Kuetz. in Sp. Alg. 687 (1849);
Tab. Phyc. 13: 4, pi. 10, fig. a-d (1863); Papenf. in J. S. Afr. Bot. 17: 178 (1952).
C. fnrcellatum Kuetz. in Sp. Alg. 687 (1849); Tab. Phyc. 13: 4, pi. 11, fig. a-b
(1863).
Holdfast : epiphytic on other algae, haustorial rhizoids arising from the base.
Erect system : somewhat rigid and setaceous, up to 20 cm high, 700 /u thick below,
tapering slightly to base and upwards gradually; branching pseudodichotomous,
often diverging strongly; adventitious branches sparse, but numerous and secund
when fertile; cortex continuous throughout, one or more layered, inner layer of large
cells of almost hexagonal outline with small angular cells seeming to cover only the
interstices between the larger cells; pericentrals nine or ten, more or less spherical
and largest of the cortical cells; plant translucent throughout, articulations as long
as cr shorter than width; tetrasporangia immersed in adventitious stichidia-like branches
and in penultimate branches; cystocarps on short adventitious branches subtended
by about five involucral branches. Plate 2: 4.
Cape. — Peninsula: Kommetjie, Simons 692. Knysna: Robberg, Ecol. Surv. RR1C. Bathurst:
Kowie (Shark’s Bay), Tyson s.n. sub BOL 27483. East London: Ecol. Surv. LI 14. Komga: Cape
Morgan, Flanagan 56 ( Tyson s.n. sub BOL 27484).
There is a confusing amount of variation in the habit of this species: in well-
developed plants the pseudodichotomies diverge at right angles to one another, whilst
in other plants they converge somewhat. Ultimately, however, they are never inrolled.
The most consistent features of this plant are the large cortical cells outlined by very
much smaller cells and the translucent thallus. When dry the plant is more or less
opaque.
166
References
AGARDH, C. A., 1824. Systema Algarum. Lund.
AGARDH, C. A., 1828. Species Algarum. 2. Greifswald.
AGARDH, J. G., 1852. Species, Genera et Ordines Algarum 2, 2, Lund.
AGARDH, J. G., 1894. Analecta Algologica. Continuatio 2. Lunds Univ. Arsskr. 30.
BONNEMAISON, T., 1822. Essai d’un Classification des hydrophites Loculees, ou Plantes marines
articulees qui croissent en France. J. Phys. Chem. Hist. Nat. 94.
DAWSON, E. Y., 1962. Marine Red Algae of Pacific Mexico. Part 7. Ceramiales: Ceramiaceae,
Delesseriaceae. Allan Hancock Pacific Expeditions 26 (1).
DE CANDOLLE, A. P., 1804-5. Note sur la Mousse de Corse. Bull. Sci. Soc. Philom. 3, 94.
DIXON, P. S., 1960. Studies on marine Algae of the British Isles: the Genus Ceramium. J. Mar.
Biol. Ass. U.K. 39: 331-374.
FELDMANN-MAZOYER, G., 1940. Ceramiacees de la Mediterranee Occidentale. Alger.
GRUNOW, A., 1867. Die Algen der Reise der Fregatte Novara urn die Erde. Wien.
KUTZING, FR. T., 1841. Uber Ceramium Ag. Linnaea 16: 727-746.
KUTZ1NG, FR. T., 1847. Diagnosen und Bemerkungen zu neuen oder kritischen Algen. Bot.
Zeit. 5.
KUTZING, FR. T., 1849. Species Algarum. Leipzig.
KUTZING, FR. T., 1862. Tabulae Phycologicae 12. Nordhausen.
KUTZING, FR. T., 1863. Tabulae Phycologicae 13. Nordhausen.
KYLIN, H., 1938. Verzeichnis einiger Rhodophyceen von Siidafrika. Lunds Univ. Arsskr. N.F.
Avd. 2, vol. 34, No. 8.
MARTENS, G., 1866. Die preussische Expedition nach ost-Asien. Botanischen Teil. Die Tange.
Berlin.
OKAMURA, K., 1921. leones of Japanese Algae 4. Tokyo.
PAPENFUSS, G. F., 1940. Notes on South African marine Algae. 1. Bot. Not. Lund.
PAPENFUSS, G. F., 1952. Notes on South African marine Algae. 3. J. S. Afr. Bot. 17, 4: 167-188.
RICHARDS, H. M., 1901. Ceramothamnion codii, a new rhodophyceous Alga. Bull. Torrey Bot.
Club 28: 257.
ROTH, A. G., 1797. Catalecta Botanica 1. Leipzig.
ROTH, A. G., 1806. Catalecta Botanica 3. Leipzig.
167
5 6
Plate 1.— 1, Ceramium glanduliferum, terminal fertile dichotomies, X 60 ( Isaac B 942)' la part of
mature filament (refractive cells are the glandular cells), x 60; 2, 2a, C. planum , completely
and partially corticated terminal portions of racemes, X 10 (2a, Simons 655); 3, C tener-
rimum, terminal and part of prostrate filament with rhizoids, X 30 ( Simons 647)- 4 C
arenarium, terminal portion of erect system (tetrasporiferous), X 30 ( Simons 632); 5’ C
atrorubescens, portion of a mature filament with two fertile laterals, each indicating a node
x 30 (Ecol. Sin v. N, 42); 6. C. diaphanum var. capense, terminal portions of laterals one
attached to a mature genicule, X 30 (Simons 614).
7230691
168
Plate 2. — 1, Ceramium papenfussianum, terminal tetrasporiferous portion of erect system, x 30
( Simons 609); la, transverse section through distal end of an axial cell showing several
pit-connections to the ring of pericentrals, X 500 (Ecol. Surv. F. 115); 2, C. diaphanum
var. pulchellum, general habit, X 2; 2a, detail of tetrasporiferous plant (swollen genicules
indicate presence of tetrasporangia), X 30 (Simons 633); 3, C. centroceratiforme, terminal
portions of filaments and portion of more mature thallus, X 30 ( Isaac B. 939); 4, C. obso-
letum, terminal portions of filaments (slight swelling of some genicules indicate presence of
tetrasporangia), X 15 (Ecol. Surv. L. 114),
169
Botha Ha 9, 1:169-171
A New Species of the Dictyotales from
South Africa
by
R. H. Simons
Whilst examining some seaweeds from Natal, I came across a few specimens
collected by Mr. M. Meyer, Lecturer at the Botany Department of the University
of Natal, which had the appearance of a dichotomising Zonaria. The specimens were
mounted on a single sheet and came from Kosi Bay in the District of Ingwavuma.
The characteristics of these plants were unlike any Zonaria previously described from
South Africa and it was therefore decided to make a detailed study of them.
Externally, the plant looks like a small Dilophus suhrii (Kutz.) Papenf. with its
dichotomising, strap-shaped branches with distinctly darker margins. Near the base
of the plant, multicellular brown rhizoids are produced, forming a stupose mat. The
resemblance to Dilophus ends here because the Kosi Bay plant has a terminal meristem
instead of a single apical cell. The fronds develop zonately, a dark transverse band
marking the termination of a period of growth. Scattered over both surfaces,
particularly towards the distal parts of the fronds, were raised sori of two kinds: one
containing nothing but multicellular uniseriate filaments (Fig. 1, C) and the other
containing large oval bodies, which were empty or had undivided contents, together
with occasional marginally disposed uniseriate filaments as in the other type of sorus.
The large oval bodies were sessile. Neither type of mature sorus was found with a
complete indusium but the remnants of the ruptured cuticle were always visible at
the margins of the sectioned sori (Fig. 1, C). The multicellular uniseriate filaments
are interpreted as hairs (paranemata) because they show no signs of being fertile.
The oval bodies appear to be fertile since they were often observed to be without
contents. Because their contents, when present, were undivided they are almost
certainly oogonia. No sporangia or male sori were found.
Closely appressed tubular skeletons of a Bryozoan forming, as it were, an additional
tissue were often seen. The Bryozoan, so far as could be seen, was confined
to the ventral surface of the fronds. Although the presence of the Bryozoan
did not seem to prevent sori from developing, the sori on the affected surface were
not as frequent as on the other surface.
The tissue is differentiated into four : two epidermal layers, a cortex and a
medulla (Fig. 1, A, B, and C). The dorsal epidermis consists of a single layer of
rather small sub-isodiametric cells, whereas the ventral epidermis has somewhat
longer cells. A central medulla of a single layer of elongated and relatively high cells
is contained within a dorsal and ventral multi-layered cortex. The medullary cells
differ from the cortical cells only by their greater height. The cells of the internal
tissue are almost always twice as wide as the epidermal cells (Fig. 1, B) except that the
subepidermal layer, on the dorsal side only, sometimes has cells of the same width
as the epidermis. On the other hand cells of the ventral epidermis are occasionally
also twice as wide as those of the dorsal epidermis. Thus, the differentiation into
dorsal and ventral surfaces is quite clear. In longitudinal section (Fig. 1, E) the origin
of new tissue in the zonate bands is seen to be from the medulla and inner cortex, the
superficial remnant layers forming flaps of truncated tissue heavily pigmented on the
upper and lower surfaces.
170
A
Fig. 1. — Pocockiella dichotoma Simons: A, L.S. of mature frond; B, T.S. of mature frond; C, T.S.
of frond showing hairs; D, L.S. through an oogonium; E, L.S. of frond through zone of
regeneration.
c: cortex; de: dorsal epidermis; e: epidermis; h: hairs; m: medulla; nf: new
frond; o: oogonium; of: old frond; ve: ventral epidermis.
As to the relationships of this plant, its zonate development and terminal meristem
indicate that it is allied to species of Zonaria. Besides, new tissue originates from the
middle layers of the frond (Fig. 1, E) as in Zonaria (Simons, 1964). it does not agree
in all features, however with the typical Zonaria construction. The absence of a stalk
cell at the base of an oogonium is found only in Zonaria tournefortii and Pocockiella
variegata ( =Gymnosorus variegatns) amongst the Dictyotales (fide Papenfuss, 1943).
171
Secondly, the presence of the medulla as distinct from the cortex is not found in species
of Zonaria. Thirdly, there is a tendency towards dorsiventrality of the thallus as
evidenced by the longer ventral epidermal cells. Finally, the arrangement of hairs in
discrete sorus-like groups instead of transverse bands is not found in species of Zonaria.
A plant which differs from typical Zonaria in like manner to the above is Pocockiella
variegata on which the genus Pocockiella Papenfuss is based. The present plant could
therefore be placed in Pocockiella. The characteristic separating Pocockiella from
Zonaria is the absence of paraphyses in its sporangial sori. These structures, unfor-
tunately, are not present on the Kosi Bay specimens. Papenfuss in his discussion
(l.c., p. 465) expressed the opinion that the peculiar internal differentiation into
cortex and medulla found in P. variegata was diagnostic and on that basis transferred
Gynmosorus nigreseens (Sond.) J. Ag. to Pocockiella. On the same basis, therefore,
the present plant should be regarded as a Pocockiella and it is described in this
genus below. The specific epithet refers to its distinctive dichotomous habit.
Pocockiella dichotoma Simons sp. nov.
Thallus dorsiventralis, ramosus, ramis linearibus, pseudodichotomis, margine
integerrimis, zonatis, inferne rhizoides gerens, ex quattuor stratis, stroma dorsali
cellularum subisodiametricarum, strato pleiostromatico corticali, stroma cen-
trali medullari et stroma ventrali cellularum sub-elongatarum, constans; cellulis
corticalibus medullaribusque aeque elongatis sed medullaribus in altitudine
majoribus, utriusque quam cellulis superficialibus duplo latioribus et ad usque octem
longioribus; paranemata et oogonia in soris discretis nudis; oogonia is sessilia; sporangia
et antheridia ignota.
Type: Natal: Ingwavuma, Kosi Bay, Meyer sub Simons 625 (PRE, holo.).
References
PAPENFUSS, G. F., 1943. Notes on Algal nomenclature. IF Gynmosorus J. Agardh. Amer. J.
Bot., 30 (7): 463-468.
SIMONS, R. H., 1964. Notes on the species of Zonaria in South Africa. Bothalia 8: 195-200.
173
bothalia 9, 1 : 173-1 82
A Virus Disease of Datura Spp.
by
Patricia J. Klesser
Introduction
Many malformed Datura plants were noticed on a land lying fallow for the summer.
The first impression was that this was damage due to hormone sprays e.g. 2-4D, as
these symptoms were similar to those caused by weedkiller preparations.
The plants were not noticeably stunted; but the leaves were severely distorted
with exaggerated indentations resulting in sharply serrated margins. They had irregular
chlorotic areas and very prominent raised veins. There was an almost complete
suppression of the spines on the seed capsules.
After severe frosts, only the skeletons of the plants remained and a count of the
spineless pods was possible. Twelve to 15 per cent of the plants were affected.
Preliminary sap transmissions to a few Solanaceous test plants resulted in unmis-
takeable virus symptoms on Datura, Nicotiana glutinosa, tobacco and petunia, and the
causal virus was therefore investigated.
Procedure
Malformed Datura plants were collected at intervals throughout the year from
different parts of the land. Seed from affected plants was harvested for seed trans-
mission tests; and seed from healthy plants was used for experimental test plants in
the glasshouse.
These Daturas were probably mixed D. stramonium x D. tatula hybrids.
Standard virus technique was used throughout this investigation.
Results
1. Physical Properties
As no local lesion host was found, these values are based on the systemic symptoms
induced on Datura, Nicotiana glutinosa, N. langsdorffii and N. tabacum. The source
of inoculum was the first three named above.
Thermal inactivation point 65°-70 C.
Dilution end point 5,000-10,000.
Longevity 3-4 days.
2. Transmission
(a) Mechanical sap inoculation. — The virus is readily sap transmissible even without
the addition of carborundum powder.
( b ) Insect vectors. — In glasshouse experiments the following were used: —
(i) Aphids. At least two species are efficient vectors : Myzus persicae (Sulz.) —
small pale green aphid found naturally on Daturas.
Aphis craccivora Koch — black aphid colonising groundnuts.
(These two were used only as a criterion of whether or not the virus
was aphid transmitted).
174
(ii) Whitefly. These are often found in the field, under the Datura leaves.
All results were negative.
Details of aphid transmission experiments: Datura was used as the source of
inoculum. Five to ten aphids were fed on excised leaves in Petri dishes for 24 hours,
and then transferred to the test plants for a further 24 hours. Symptoms developed
on the following plants: Datura sp., Nicotiana glutinosa , N. langsdorjfii, N. rustica,
N. sanderae , N. sylvestris, N. tabacum and Physalis peruviana.
A feature of the aphid-infected Daturas was the extreme malformation and dis-
tortion of their leaves. Often they were reduced to only a thickened midrib or possibly
a half lamina only 2 mm wide along the entire main vein (Plate 2). There was a complete
suppression of spines on the capsules. This increase in intensity of symptoms was not
apparent on other host plants.
( c ) Seed transmission. — As the virus distorts the flowers and suppresses the spines
on the capsules of Daturas, the possibility of the virus being transmitted through the
seed was investigated. (The Datura Quercina virus was transmitted to 79 per cent
of the progeny. Blakeslee 1921).
Seed harvested from D. stramonium plants individually infected with six of the
nine isolates was sown in sterilized soil. Normal Datura seed is black and firm, and
the capsules are fully packed. In contrast, the seeds from infected pods are small,
white or gray, and shrivelled.
Both strains of the virus reduced the percentage of viable seeds per pod ; and
from plants infected with the severe strain less than half of the seeds germinated. Not
a single seedling showed virus symptoms.
The numbers of seeds sown, germinated and infected were recorded:
Severe strain : 402 seeds sown, 171 germinated i.e. 42-5 per cent (3 isolates).
Mild strain : 90 seeds sown, 77 germinated i.e. 85-5 per cent (3 isolates).
3. Host Range Experiments
GROUP A
The following plants developed obvious virus symptoms and the virus was readily
transmitted back to Datura :
Solanaceae: Datura ferox L., D. stramonium L., Nicotiana clevelandii A. Gray, N.
glutinosa L., N. langsdorjfii Weinm., N. rustica L., N. sanderae Sander., N. sylvestris
Speg. & Comes., N. tabacum L. vars. Delcrest, Hicks and White Burley, Petunia hybrida
Hort. ex Vilm. var. Rosy Morn and Physalis peruviana L.
GROUP B
Clear virus symptoms developed on 3 hosts but few transmissions back to Datura
were positive:
Solanaceae: Capsicum cmnuum L., C. frutescens L. and Lycopersicon esculentum
Mill. var. Little Marvel.
GROUP C
Apparent virus symptoms developed on several hosts but all sub-inoculations to
Datura were negative.
Aizoaceae: Tetragonia expansa Thunb. — diffuse chlorotic blotches on the young
leaves.
Amaranthaceae: Gomphrena globosa L. — local necrotic rings.
175
Chenopodiaceae : Chenopodium amar anticolor Coste & Reyn. — local white lesions
which become necrotic.
Compositae: Callistephus chinensis Nees — raised darker green blisters on some leaves.
Cucurbitaceae : Cucumis sativus L. — local white necrotic lesions.
Leguminosae: Crotalaria juncea L. — local necrotic specks.
Phaseolus lathy roides L. — systemic chlorotic spots. Vicia faba L. — local necrotic
rings.
GROUP D
Latent carriers i.e. the test plants were symptomless but sub-inoculations to Datura
were positive.
Compositae: Ageratum houstonianum Mill., Dahlia pinnata Cav., Zinnia elegans Jacq.
Malvaceae: Althaea rosea Cav.
Primulaceae: Primula malacoides Franch.
Scrophulariaceae: Antirrhinum majus L.
GROUP E
The following plants were not susceptible to this virus:
Boraginaceae: Anchusa capensis Thunb.
Caryophyllaceae: Dianthus barbatus L.
Chenopodiaceae: Beta vulgaris L.
Compositae: Arctotis breviscapa Thunb., Sonchus oleraceus L.
Cruciferae: Alyssum maritimum Lam.
Graminae: Zea mays L.
Labiatae: Salvia splendens Ker-Gawl.
Leguminosae: Lupinus albus L. var. Sweet, L. luteus L. var. Sweet, Medicago sativa
L., P. vulgaris L. var. Canadian Wonder, Pisum sativum L. var. Greenfeast, Trifolium
subterraneum L. Vigna unguiculata (L.) Walp.
Onagraceae: Gaura lindheimeri Engelm. & Gray.
Polemoniaceae : Phlox drummondii Hook.
Ranunculaceae : Aquilegia Hort. sp.
Violaceae: Viola Hort. sp.
4. Reaction of Susceptible Species
After several experiments it was evident that there were differences in the symptoms
induced on the test plants by the virus isolates from the many naturally infected Daturas
used. This was a difference in degree of severity and not a different type of symptom.
Therefore it was assumed that two strains of the same virus occurred in that area — mild
and severe. This variation was not apparent on all host plants, so the 2 strains are
described separately only where the difference was significant.
GROUP A
SOLANACEAE
Datura stramonium. There is no local reaction. The severe strain causes a vein-
clearing at the base of the young leaves in 14 days. As these leaves mature the main
vein becomes arched downwards and the leaf margins are rolled. Later formed leaves
have distinct linear chlorotic areas which fuse and cause distortion; and there are
also dark green blisters. The teeth of these leaves are sharper than normal.
These symptoms increase in severity on all new growth until the laminae are reduced
to narrow bands of tissue along the midrib from which long “ teeth ” extend. The
last leaves formed before flowering may be stringlike. The flowers are twisted and
may also be distorted.
7230691-6
176
The seedpods or capsules are relatively small and completely smooth or have
only a few vestigial spines. Plate 1 . There is a marked dark green network on these
bare pods. Fewer seeds are formed and a high percentage are infertile.
The mild strain induces similar early symptoms. However, on later growth the
indentations are not as deep, there is less malformation of the laminae and the leaves
show an irregular chlorotic mottle. The capsules have isolated short spines, or the
spines may be confined to one valve only. Plate 2.
(The symptoms which develop after insect transmission are described in that
section).
Datura ferox. With both strains the reaction is similar to, but not as severe as that
of D. stramonium. The capsules have few short stubby spines.
Nicotiana clevelandii. With both isolates there is no local reaction; nor is there a
preliminary veinclearing or chlorotic spotting. After a month the young developing
leaves are misshapen and fluted, and this malformation increases on subsequent leaves.
With the severe strains the laminae are reduced to occasional dark green blisters
protruding from the side of the distorted stringlike midrib. The whole plant is extremely
stunted.
With the mild strain the leaves are broader but more distorted with frequent dark
green blisters and the plant is stunted. Plate 4.
Nicotiana glutinosa. There is no local reaction. In 6-10 days there is a chlorotic
network followed by a mottle. As the leaves mature the chlorosis intensifies to form
large blotches on an uneven leaf surface.
In addition to the mottle, the severe strain induces short dark green veinbands,
and the leaves are elongated.
Nicotiana langsclorffii. The local reaction is slow in appearing. With the severe strain,
large chlorotic blotches develop after a month, or the leaves just become flaccid and
later collapse and drop. Leaves inoculated with the mild strain develop chlorotic
blotches and become flaccid but do not collapse.
The first systemic symptoms of the severe strain are seen in 10-14 days as a vein-
clearing and interveinal chlorotic spotting on the young leaves. On later formed leaves
the veins are severely puckered, resulting in dark green blisters. The leaves are small
and the plant is stunted.
With the mild strain there is a veinclearing of the young leaves, followed by a
chlorotic mottle and slight crinkle on the next formed leaves. After several months,
only the base of the new leaves is a lighter green than normal with dark green along
the veins. The plant is only slightly stunted. Plate 3.
Nicotiana rustica. There is no local reaction. Both isolates induce a veinclearing
on the young leaves followed by a chlorotic mottle with short dark green veinbands
and irregular islands. The leaf surface is rugose with both convex and concave blisters.
The leaves are elongated.
Nicotiana sanderae. No local reaction develops. With both strains the first symptoms
are a slight puckering of the veins and rolling of the leaf margins. Later a chlorotic
mottle with irregular dark green islands develops.
After a month, the severe strain causes a necrosis of the veins, which results in a
distortion of the lamina.
Nicotiana sylvestris. There is no local reaction nor preliminary veinclearing or chlorotic
spotting. The young leaves are a lighter green with dark green rectangles forming
an angular mosaic. They are slightly puckered.
177
Nicotiana tabacum var. White Burley. There is no local reaction. In 8-10 days a
chlorotic network appears on the young leaves, followed by a chlorotic mottle with
short dark green veinbands. On later formed leaves there are large chlorotic blotches
and irregular dark green islands. The laminae are slightly crinkled and the margins
are waved.
With the severe strain the chlorotic and dark green symptoms are accentuated,
and there is a distortion of one half of the lamina causing the leaf to curve towards
that side. These leaves are also elongated and the plant is stunted.
Petunia hybrida. No local reaction develops. The first systemic symptom with the
severe strain is a clearing of the veins of the young leaves in 7 days, followed by a
chlorotic speck mottle in 2 weeks and then speckled ring patterns. These patterns
become almost white as the leaves mature. On new secondary shoots there is a shock
reaction of necrotic areas which soon turn white. These areas may fuse and the leaves
become flaccid. Later formed leaves on these shoots are elongated and malformed.
There is a distinct colour break on most flowers.
With the mild strain there are only a few chlorotic blotches and later chlorotic
speck patterns.
Physalis peruviana. There is no local reaction. After 18-20 days areas of chlorotic
network appear on the young leaves, but these soon fade. After several months there
is an indistinct dark green mottling of some leaves.
GROUP B
SOLANACEAE
Capsicum spp. There is no local reaction. After several weeks the young leaves show
isolated chlorotic flecks and dark green islands. The flecks enlarge and become vivid
bright yellow. If they are adjacent to the midrib, this bends down at right angles at
that point.
Lycopersicon esculentum. No local reaction. Isolated irregular vivid yellow areas
develop after 6-8 weeks, usually along the veins. Most leaflets are severely curled
back and there may be a diffuse mottle.
Discussion
A survey of the literature shows that several virus diseases of Datura spp. have
been reported — some found primarily on Daturas, others more often on other host
plants in the Solanaceae.
1. Datura virus 1 (Smith & d’Oliviera in Smith 1937) does not cause any malfor-
mation on Datura spp. and is able to infect Phaseolus and Vigna whereas this virus is
apparently confined to the solanaceous plants.
2. Datura virus 2 (Thomas & Krishnaswami 1939) is not sap transmissible whereas
this virus does not even need the addition of an abrasive.
3. Datura virus 3 (Datura distortion mosaic virus) (Capoor & Yarma 1948,
1952) does not infect D. stramonium systemically, although the symptoms on D. inoxia
do include malformation. Capoor’s virus caused a systemic necrosis of White Burley
but most other Nicotianas showed only a local necrotic reaction.
4. Datura virus 3a (Garga 1958) is confined to Datura spp. and petunia while
this virus can infect several other host plants.
5. Datura rugose leaf curl virus (Grylls 1954) is transmitted solely by leafhoppers.
178
6. Datura necrosis virus (Badami & Kassanis 1959) as its name implies induces
a lethal local and systemic necrosis on D. tatula and severe mosaic and stunt on
tobacco.
7. Pepper veinbanding virus v
>(Dale 1954, 1956). Both infect a large number of
8. Eggplant mosaic virus J
solanaceous plants but Datura spp. were not included in the host range. Nevertheless,
symptoms on other plants, hosts and insect vectors differ from those described for
this virus.
9. Solanum phureja virus (Bagnall 1963) does not infect Datura spp., tobacco,
N. glutinosa or Capsicum.
= (Blakeslee 1921). Both caused severe malformation
10. Datura Quercina virus'
11. Datura Z-virus S
of Datura leaves and the suppression of the spines on the capsules. With the methods
used at that time Blakeslee was unable to sap-transmit the Quercina virus, although
he did graft-transmit it to other Daturas, but not to tobacco, tomato, petunia, etc.
Furthermore, this Q-virus was seed transmitted to 79 per cent of the progeny of Datura
spp., whereas the virus described in this paper is definitely not seed-borne.
The Z-virus was easily sap-inoculable and not seed transmitted; but tobacco was
not a host. Nevertheless, Smith (1957) places this Z-virus in the tobacco etch virus
group.
12. Tobacco etch virus group (Johnson 1930). A comparison of this Datura virus
with those in the tobacco etch virus group shows some similarities with several, but
no complete correlation with any one of the present strains.
Bawden (1964) cites the suppression of spines on the seed pod as a feature of
tobacco etch virus on Datura. However, the necrotic etch symptom on tobacco which
is a characteristic of the viruses in this group does not occur with the virus described
here. The V-strain of Stover (1951) is the only one which, under greenhouse conditions,
does not induce systemic necrotic etch on tobacco, only a mild mottle and veinbranding;
but on Daturas the Y-strain causes only a transient chlorotic spotting and no malfor-
mation. Even the mild strain of Bawden & Kassanis (1941) causes an initial etching
on tobacco which later fades.
Doolittle & Alexander (1951) working on tobacco etch in greenhouse tomatoes
which showed a mild mottle and curl similar to that caused by this Datura virus, state
only that all isolates induce “ severe malformation ” of D. stramonium leaves.
The pepper strain of Greenleaf (1953) resulted in a very severe wilt of Capsicum
spp. and is thereby excluded from this comparison. However, the bell pepper strain
of McLean (1962) does cause distortion of D. stramonium leaves but the typical necrotic
etch occurs on tobacco.
A comparison with the host range of tobacco etch virus as given by Holmes (1946)
shows many similarities in the susceptible species, but not in the symptoms induced
on these species.
Hollings (1959) describes the symptoms of tobacco severe etch virus on a new
indicator test plant, Nicotiana clevelandii, as a yellow network followed by a chlorotic
mottle, dwarfing and rosetting. In contrast, this Datura virus causes such severe
malformation that the leaves are reduced to strings with occasional dark green blisters.
179
Conclusions
The Datura virus described above appears to be a new virus or strain belonging
to the tobacco etch virus group, even though necrotic etching does not occur on tobacco.
Suppression of the spines on Daturas is a feature of both this virus and tobacco etch.
According to Rossouw (private communication) tobacco etch has never been isolated
from tobacco in South Africa, and the occurrence of a virus in this group is thus a
new record for this country.
The malformation of Datura leaves is extremely severe with this new virus, and
a similar malformation occurs on N. clevelandii leaves. Therefore, the name DATURA
MALFORMATION VIRUS is suggested.
References
BADAMI, R. S. & KASSANIS, B., 1959. Some properties of three viruses isolated from a diseased
plant of Solarium jasminoides Paxt. from India. Ann. Appl. Biol. 47: 90-97.
BAGNALL, R. H., 1963. A virus from Solatium phureja. Phytopath. 53: 546-547.
BAWDEN, F. C., 1964. Plant viruses and virus diseases. 4th Ed. Ronald Press Co., New York.
BAWDEN, F. C. & KASSANIS, B., 1941. Some properties of tobacco etch viruses. Ann. Appl.
Biol. 28: 107-118.
BLAKESLEE, A. F., 1921. A graft-infectious disease of Datura resembling a vegetative mutation.
J. Genet. 11: 17-36.
CAPOOR, S. P. & VARMA, P. M., 1948. A mosaic disease of Datura alba Nees. Curr. Sci. 17:
151-152.
CAPOOR, S. P. & VARMA, P. M., 1952. Studies on the “ distortion mosaic ” virus of Datura inoxia
Mill. Ind. Jour. Agric. Sci. 22: 303-314.
DALE, W. T., 1954. Sap-transmissible mosaic disease of Solanaceous crops in Trinidad. Ann. Appl.
Biol. 41: 240-247.
DALE, W. T., 1956. Virus diseases of Solanaceous crops in Trinidad. Trop. Agric. Trin. 33: 35-50.
DOOLITTLE, S. P. & ALEXANDER, L. J., 1951. Occurrence of tobacco etch virus on greenhouse
tomatoes in Ohio. PI. Dis. Rept. 35: 351-353.
GARGA, R. P., 1958. Studies on virus diseases of plants in Madhya Pradesh. II. A new strain of
Datura virus 3. Ind. Phytopath. 11: 133-136.
GREENLEAF, W. H., 1953. Effects of tobacco-etch virus on peppers ( Capsicum sp.). Phytopath.
43: 564-570.
GRYLLS, N. E., 1954. Rugose leaf curl — a new virus disease transovarially transmitted by the
leafhopper, Austrogallia torrida. Austr. J. Biol. Sci. 7: 47-58.
HOLLINGS, M., 1959. Nicotiaua clevelandii Gray as a test plant for plant viruses. PI. Path. 8:
133-137.
HOLMES, F. O., 1946. A comparision of the experimental host ranges of tobacco-etch and tobacco-
mosaic viruses. Phytopath. 36: 643-659.
JOHNSON, E. M., 1930. Virus diseases of tobacco in Kentucky. Kentucky Agric. Exp. Sta. Bull.
306.
KASSANIS, B., 1941. Transmission of tobacco etch viruses by aphids. Ann. App. Biol. 28: 238-243.
MCLEAN, D. M„ 1962. Tobacco-etch virus on peppers and tomatoes in the Lower Rio Grande
Valley of Texas. PI. Dis. Rept. 46: 565-568.
SMITH, K. M., 1937. A textbook of plant virus diseases. Churchill, London.
SMITH, K. M., 1957. A textbook of plant virus diseases. 2nd Edition. Churchill, London.
STOVER, R. H., 1951. Tobacco etch virus in Ontario. Canad. Jour. Bot. 29: 235-245.
THOMAS, K. M. & KRISHNASWAMI, C. S., 1939. “Little leaf"— a transmissible disease of
brinjal. Proc. Ind. Acad. Sci. Sect. B. 10: 201-212.
VALLEAU, W. D., 1940. Classification and nomenclature of tobacco viruses. Phytopath. 30:
820-830.
VALLEAU, W. D. & JOHNSON, E. M., 1928. Some virus diseases of tobacco in Kentucky. Phyto-
path. 18: 132-133.
.
• ,
181
Plate 1. — Left: Healthy Datura plant. Centre and right: Severe strain showing malformation of
leaves, distortion of flower and spineless pod. Sap transmission.
Plate 2. — Left: Severe strain on Datura showing string-like leaves. Aphid transmission. Right:
Mild strain showing distorted chlorotic areas. Sap transmission.
7230691
182
Plate 3. — Top: Nicotiana iangsdorffii. Sap transmission. Left: Mild strain. Right: Severe strain.
Bottom: Daturas, showing almost complete suppression of spines on the pods.
Plate A.— Nicotiana clevelandii. Sap transmission. Left: Severe strain. Right: Mild strain.
Healthy plant in the centre.
183
Bothalia 9, 1:183-194.
The Identity of a Pea Blight Fungus in South Africa*
by
K. T. van Warmelo
Plant Protection Research Institute, Pretoria
Abstract
The perfect stage of Ascochyta pinodes (Berk. & Blox.) Jones, a cause of pea blight in Natal, was
compared with type material of Sphaeria pinodes Berk, and Blox., Mycosphaerel/a pinodes (Berk. &
Blox.) Stone, and Didvmella pinodes (Berk. & Blox.) Petrak and the development of its ascocarps
studied.
Two types of ascocarp were found on the material of Didymella pinodes , one perithecial and the
other ascolocular in structure. The ascocarp of the South African fungus was typically ascolocular
in development and construction and similar to that of other species of Mycosphaerella. These ascocarps
were identical to those of Sphaeria pinodes and Mycosphaerella pinodes and the ascolocular ascocarps
of the Didymella pinodes material.
In development and morphology this fungus agrees more closely with the original generic concepts
of the genus Mycosphaerella Joh. than with Didymella Sacc. and should thus be named Mycosphaerella
pinodes (Berk. & Blox.) Stone.
Introduction
While investigating pea blight in Natal, it was noticed that an anomaly existed
concerning the identity of Mycosphaerella pinocles (Berk. & Blox.) Stone, one of the
three causal organisms listed by Doidge, Bottomley, van der Plank and Pauer (1952).
No paraphyses could be seen in the ascocarps found on overwintered straw but Muller
and von Arx (1950, 1962) followed Petrak (1924) who transferred this fungus to
Didymella Sacc. This latter genus differs from Mycosphaerella Joh. by the presence
of interascal paraphyses (Saccardo, 1882b).
Although Muller and von Arx (1950, 1962) do not regard the presence or absence
of paraphyses as an important character in the taxonomy of this group of Ascomycetes,
Miller (1949) and Luttrell (1951, 1955) showed that in the Dothidea-type ascocarp,
as found in Mycosphaerella, the centrum is composed of pseudoparenchyma which is
squashed as the asci develop. This results in the formation of a locule occupied by
a cluster of aparaphysate asci. In the Pleospora-type ascocarp, as found in Didymella
(Gaumann, 1952), the centrum consists of pseudoparaphyses which, according to
Luttrell (1955), originate as separate paraphysis-like hyphae prior to formation of the
asci. The asci later develop among them. The pseudoparaphyses are thus not
remnants of interthecial tissue. Luttrell (1951, 1955) believed that these differences
afford an adequate basis for the separation of orders in the Ascomycetes.
Muller and von Arx (1950) considered that the presence or absence of pseudo-
paraphyses can not be used as a criterion for the separation of genera. They maintain
that pseudoparaphyses are remnants of interthecial tissue which may be eventually
completely absorbed. The presence or absence of pseudoparaphyses is thus a variable
condition depending on the stage of maturity of the ascocarp. These two authors
(1962) do, however, employ the presence or absence of pseudoparaphyses in their key
to separate Mycosphaerella and Didymella.
* Based on a thesis submitted to the University of Natal, Pietermaritzburg, in partial fulfilment
of the requirements for the degree of M.Sc. Agric.
184
Petrak (1924) considered that all species of Mycosphaerella which possess Ascochyta
imperfect stages should be placed in Didymella. This was accepted by Muller and
von Arx (1950, 1962) who stated that the most important distinction between Myco-
sphaerella and Didymella was the shape of the ascospores. In Mycosphaerella species
the ascospores have a length : breadth ratio of 3 : 1 or more, whereas this ratio is lower
in the ascospores of Didymella species.
Berkeley and Bloxam in 1861 first described Sphaeria pinodes, a fungus producing
hyaline, two-celled ascospores in single perithecia (Muller and von Arx, 1962). This
name was changed to Sphaerella pinodes by Niessl (Saccardo, 1882a) who accepted the
genus Sphaerella proposed by Cesati and de Notaris in 1861 (von Arx, 1949). Johanson
(Grove, 1912) showed in 1884 that Sphaerella was not a valid name and substituted
Mycosphaerella which, according to Saccardo (1891) did not possess paraphyses (pseudo-
paraphyses sensu Luttrell).
Stone (1912) transferred Sphaerella pinodes (Berk. & Blox.) Niessl to Myco-
sphaerella pinodes (Berk. & Blox.) Stone. In his detailed description of M. pinodes.
Stone did not mention paraphyses which must, therefore, be assumed to be absent. He
also considered the imperfect form of Mycosphaerella pinodes to be Ascochyta pisi Libert.
Linford and Sprague (1927) showed that the imperfect stage of M. pinodes, named
Ascochyta pinodes by Jones (1927), differed from Ascochyta pisi Lib., thus disproving
Stone’s assumption. Jones (1927) was unable to find the perfect stage of Ascochyta
pisi Lib.
Von Hohnel (1918) transferred Mycosphaerella pinodes (Berk. & Blox.) Stone to
Didymellina, a genus which Petrak (1923) later considered to be invalid.
Petrak (1924) examined fresh material of a fungus he identified as Didymellina
pinodes (Berk. & Blox.) v. Hohn. and published a detailed description of the ascocarp.
He observed that the centrum was filled with paraphysoids (pseudoparaphyses) which
consisted of pseudoparenchymatous cells. Eventually the pseudoparaphyses formed
an indefinite fibrous matrix in which the asci were embedded and difficult to remove.
Petrak then reclassified this fungus as Didymella pinodes (Berk. & Blox.) Petr, on the
basis of the morphology of the ascocarp and asci and the presence of the Ascochyta
imperfect stage.
As a result of these conflicting views about the morphology and taxonomy of
this ascocarpic pea blight fungus it was decided to investigate the morphology of the
ascocarp and to establish the identity of the South African fungus by examination
of type and authenticated specimens of the fungi concerned.
Materials and Methods
The South African fungus studied here was found on pea straw from Pietermaritz-
burg, Natal. Type material of Sphaeria pinodes Berk. & Blox., Mycosphaerella pinodes
(Berk. & Blox.) Stone and Didymella pinodes (Berk. & Blox.) Petr, was obtained for
comparison.
Ascocarps from herbarium material were softened in a 1:1:2 mixture of glacial
acetic acid, lactic acid and water. This allowed satisfactory squash preparations
showing the internal structures to be made. Squash preparations were mounted in
lactophenol.
The fungus was cultured on 3 per cent malt agar to which barley kernels, sterilized
by dry heat at 160°C for 2 hours, were added. Ascospore or pycnospore suspensions
were used as inoculum for subcultures.
185
The development of the ascocarp was studied from microtome sections cut at
16//. Ascocarps were fixed in formal-acetic-alcohol (Riker and Riker, 1936) for 24
hours, dehydrated in Cellosolve (2-Ethoxy-ethanol) (Gurr, 1956) and embedded in
paraffin wax of 55 °C melting point.
The ribbon was mounted with Haupt’s adhesive (Gurr, 1956). Sections were
stained in Pianeze Illb (Conn and Darrow, 1946) for one hour, differentiated in acid
alcohol, washed in alcohol and mounted in Canada Balsam.
Descriptions
Sphaeria pinodes Berkeley and Bloxam, in Ann. Mag. Nat. Hist., ser. 3, 7: No.
981 (1861). (Muller and von Arx, 1962. p.363).
Material examined: Berkeley Herb. No. 1879 in Herb. Royal Botanic Gardens,
Kew, England.
Figs. 1, 11, 15.
Ascocarps caulicolcus, scattered over the surface, not in distinct lesions, sub-
epidermal becoming partially erumpent, opening by central ostiole, brown, 100-1 80, //
in diameter; wall of ascocarp stromatic; centrum composed of thin-walled, hyaline,
pseudoparenchymatous cells; asci cylindrical to cylindrical-clavate, bitunicate,
45- — 65 X 11 — 15//; spores eight, irregularly biseriate, hyaline, ellipsoid, two-celled,
slightly constricted at the septum, upper cell somewhat larger, 12-5- — 16 X 5 — 7//.
No indication as to the host is given on the specimen mount.
Mycosphaerella pinodes (Berk. & Blox.) Stone, in Ann. Myc. 10: 564-592 (1912).
Material examined: From Plant Pathology Herbarium, Cornell University, United.
States of America. (CUP-A).
Figs. 2, 12, 16.
Ascocarps caulicolous, scattered over the surface, not in distinct lesions, sub-
epidermal, becoming partially erumpent, opening by central ostiole, brown, 100-200//
in diameter; wall of ascocarp stromatic, one to three cell layers thick; centrum com-
posed of thin-walled, hyaline, pseudoparenchymatous cells; asci cylindrical-clavate,
bitunicate, not all maturing at the same time, 60-80 X 12-16//; spores eight, irregularly
biseriate, hyaline, elliptical-obovate, two-celled, slightly constricted at the septum,
upper cell usually somewhat larger, 12-15-5 X 4-6//.
In this material a large number of ruptured asci were found in each ascocarp.
The walls of these asci were not always easily visible while the central lumina often
appeared as dark lines (Fig. 16) which could be mistaken for pseudoparaphyses.
On Pisum sativum L.
Didymella pinodes {Berk. & Blox.) Petrak, in Ann. Myc. 22: 16-18 (1924).
Material examined: Mycotheca Generalis No. 428. Two distinct types of ascocarp
were present on this material.
Type A
Figs. 3, 13.
Ascocarps caulicolous, scattered over the surface, not in distinct lesions, sub-
epidermal becoming partially erumpent, opening by central ostiole, brown, 90-180//
in diameter; wall of ascocarp stromatic, one to three cell layers thick; centrum composed
186
of thin-walled, hyaline, pseudoparenchymatous cells which are compressed by the
developing asci; asci cylindrical to cylindrical-clavate, bitunicate, not all maturing
at the same time, 60-75 x 12-16//; spores eight, irregularly biseriate, hyaline, ellipsoid
to elliptical-obovate, two-celled, slightly constricted at the septum, upper cell usually
somewhat larger, 12-5-16 X 4 -5-5 -5//.
On Pisum sativum L.
Type B
Fig. 17.
Ascocarps caulicolous, scattered over the surface, not in distinct lesions, sub-
epidermal becoming erumpent, opening by central ostiole, brown to black, 130-250//
in diameter; wall of ascocarp hard; asci cylindrical-clavate, 70-95 X 18-24//,
immature, interspersed with slender septate paraphyses with free apices, 65-100 X 3-4//.
Spores not found.
On Pisum sativum L.
Mycosphaerella pinodes {Berk. & Blox.) Stone.
Material examined: Mycological Herbarium, Nat. Herb., Pretoria. PRE 42549.
Fig. 4, 5, 6, 7, 8, 9, 10, 14.
Ascocarps caulicolous and fructicolous, scattered over the surface, not in distinct
lesions, subepidermal becoming partially erumpent, opening by central lysigenous
ostiole, brown, 80-180// in diameter; wall of ascocarp stromatic, one to three cell
layers thick; centrum composed of thin-walled, hyaline, pseudoparenchymatous cells
which are compressed by the developing asci, leaving strands of tissue which may
disappear completely; asci cylindrical to cylindrical-clavate, bitunicate, not all maturing
at the same time, 65-80 X 12-15//; spores eight, obliquely uniseriate to irregularly
biseriate, hyaline, elliptical-obovate, two-celled, slightly constricted at the septum,
upper cell somewhat larger, 13-15-5 x 4-5-6//.
Pycnidial stage: Ascochyta pinodes (Berk. & Blox.) Jones.
Fig. 18.
Pycnidia caulicolous and foliicolous, scattered in irregular purplish lesions, sub-
epidermal becoming partially erumpent, opening by central ostiole, brown appearing
black under reflected light, globose with papillate rostrum, 70-180// in diameter;
pycnidial wall pseudoparenchymatous; spores hyaline, ellipsoid, straight or sometimes
somewhat curved, two-celled, slightly constricted at the septum, 11-20 x 2 -5-6-0//.
On Pisum sativum L.
Development of the ascolocule. The ascostromata were produced singly, each
stroma forming one locule only.
Locule development was initiated in young growing ascostromata. Stromal cells
in the centre of the globose ascocarp became thin-walled and larger while the outer
two to three layers of cells became thick-walled and dark in colour. No ascogonia
were observed but it is probable that the ascogenous hyphae arose in this less dense,
pseudoparenchymatous, stromal tissue (van Warmelo, 1966) (Fig. 5). The asci
developing from the ascogenous hyphae grew into this pseudoparenchyma, crushing
it and so forming the locule (Fig. 6). Lateral growth of the asci resulted in strands
of compressed cells (Fig. 7) which finally disappeared leaving a locule completely
187
filled with aparaphysate asci in various stages of maturity (Fig. 8). A lysigenous,
aperiphysate ostiole was formed (Fig. 9). As the asci matured they elongated through
the ostiole, ruptured to release their spores and collapsed completely. Old ascocarps
could, therefore, be found only partially filled with asci or completely empty (Fig. 10).
The development described above is similar to that observed in several other species
of Mycosphaere/la (Barr, 1958; Higgins, 1920, 1929, 1936; Jenkins, 1930, 1938, 1939).
Discussion
From the above descriptions it is evident that the fungi collected by Berkeley and
Bloxam, Stone, Petrak (type A) and the author are markedly similar in location, size,
shape and morphology of the ascocarps. Their asci are also similar in shape as can
be seen from figures 11, 12, 13 and 14. The small variation in their size was ascribed
to the desiccation caused by long storage. Only slight differences in the sizes of the
ascospores were observed, but none in their shape. These four fungi are thus considered
identical. The specific epithet “ pinocles ” can thus be applied to the South African
species.
The two types of ascocarp found on the material obtained from Dr. Petrak showed
marked differences in centrum structure. In type A ascocarps the centrum consisted
of hyaline, isodiametric, thin-walled, pseudoparenchymatous cells, strands of which
sometimes remained between the asci (Fig. 13). There was no evidence of the presence
of pseudoparaphyses. The centrum was thus typically ascolocular and identical with
that of Mycosphaere/la pinodes (Berk. & Blox.) Stone. In the type B ascocarps slender
and septate true paraphyses with free apices could be distinguished (Fig. 17). The
asci were similar in shape to those in type A but were larger. In the absence of
ascospores the genus could not be determined. The presence of true paraphyses does,
however, show that this fungus possesses a true perithecium and is thus not a species
of Didymella Sacc. which has ascolocular ascocarps. In view of the differences shown
to exist between the ascocarps in Petrak’s material, it would appear that the description
of Didymella pinodes (Berk. & Blox.) Petr, was based on a mixed collection.
The average length: breadth ratio of the ascospores of the four fungi examined
is 2-7 which is less than the ratio for Mycosphaere/la quoted by Muller and von Arx
(1950). It must, however, be noted that in the descriptions of Mycosphaerel/a and
Didymella species by Muller and von Arx (1962) the length: breadth ratios are not
constant for all species of these two genera.
Luttrell (1951) and Muller and von Arx (1950) agree that the development and
morphology of the ascocarp and of the asci and ascospores are the criteria on which
taxonomic divisions should be based. Muller and von Arx (1950) stated that charac-
teristics of the vegetative phase may only be considered to the extent of their direct
relation to the sexual stage.
Petrak (1924) and Muller and von Arx (1962) stated that species with Ascochyta
imperfect stages should be classified in Didymella. The perfect stage of Ascochyta
pinodes has been shown here to lack certain morphological structures found in the type
and other species of Didymella. These structures are regarded as most important
in the classification of Ascomycetes by Luttrell (1951, 1955) and Muller and von Arx
(1950, 1962). Furthermore, species of both the genera Phoma and Phyllosticta are
known to have ascocarpic forms in both the genera Didymella and Mycosphaerella
(Muller and von Arx, 1962; Wolf, 1940; Brooks, 1953). The restriction of the perfect
stages of Ascochyta species to the genus Didymella thus appears unjustified.
The development of the ascolocule in the South African fungus has been shown
to follow that of a typical Dothidea-type, sensu Luttrell (1951). Pseudoparaphyses
were not observed at any stage of development. In view of the similarity of this fungus
188
with those of Berkeley and Bloxam, Stone and Petrak (type A) it must be assumed
that pseudoparaphyses were not present at any stage of development of these older
collections, of which the development could not be examined.
It is thus evident that the perfect stage of the South African pea blight fungus,
Ascochyta pinocles, agrees much more closely with the concept of Mycosphaerella
Joh. than with Didymella Sacc. and should, therefore, be known as Mycosphaerella
pinocles (Berk. & Blox.) Stone.
Acknowledgements
The author is indebted to The Director, Royal Botanic Gardens, Kew, England;
The Curator, Plant Pathology Herbarium, Cornell University, Ithaca, United States
of America and Dr. F. Petrak, Naturhistorisches Museum, Vienna, Austria for the
loan of the type and authentic material of Sphaeria pinodes Berk, and Blox.; Myco-
sphaerella pinodes (Berk. & Blox.) Stone and Didymella pinodes (Berk. & Blox.) Petr,
respectively.
References
ARX, J. A. VON, 1949. Beitrage zur Kenntnis der Gattung Mycosphaerella. Syd. Ann. Myc. 3:
28-100.
BARR, MARGARET, E., 1958. Life history studies of Mycosphaerella tassiana and M. typhae.
Mycologia 50: 501-513.
BROOKS, F. T., 1953. Plant diseases. Oxford Univ. Press. London, 457p.
CONN, H. J. & DARROW, MARY A., 1946. Staining procedures used by the Biological Stain
Commission. Biotech Publications, Geneva, N.Y.
DOIDGE, E. M„ BOTTOMLEY, A. M., VAN DER PLANK, J. E. & PAUER, G. D., 1952. A
revised list of plant diseases occurring in South Africa. Science Bull. No. 346. Botany and Plant
Pathology Series No. 16, 122p.
GAUMANN, E. A., 1952. The fungi. A description of their morphological features and evolutionary
development. Elafner Publ. Co. New York. 420p.
GROVE, W. B., 1912. Sphaerella v. Mycosphaerella. Journ. of Bot. 50: 89-92.
GURR, E„ 1956. A practical manual of medical and biological staining techniques. Leonard Hill
Ltd. 320p.
HIGGINS, B. B., 1920. Morphology and life history of some Ascomycetes with special reference
to the presence and function of spermatia. Amer. Journ. Bot. 7: 435-444.
HIGGINS, B. B., 1929. Idem. II. Ibid. 16: 287-296.
HIGGINS, B. B., 1936. Idem III. Ibid. 23: 598-602.
HOHNEL, F. VON, 1918. Mycologische Fragmente CCVIII. Ober Sphaeria ( Depazea ) phaseolicola
Roberge. Ann. Myc. 16: 66-67.
JENKINS, W. A., 1930. The cherry leafspot fungus, Mycosphaerella cerasella Aderh., its morphology
and life history. Phytopathology 20: 329-337.
JENKINS, W. A., 1938. Two fungi causing leaf spot of peanut. Journ. Agr. Res. 56: 317-332.
JENKINS, W. A., 1939. The development of Mycosphaerella berkeleyii. Ibid. 58: 617-620.
JONES, L. K., 1927. Studies of the nature and control of blight, leaf and pod spot, and footrot of
peas caused by species of Ascochyta. New York State Agr. Exp. Sta. Bull. 547. 46p.
LINFORD, M. B. & SPRAGUE, R., 1927. Species of Ascochyta parasitic on the pea. Phytopathology
17: 381-397.
LUTTRELL, E. S., 1951. Taxonomy of the Pyrenomycetes. Univ. Missouri Studies 24 (3): 1-120.
LUTTRELL, E. S., 1955. The ascostromatic Ascomycetes. Mycologia 47: 511-532.
MILLER, J. H., 1949. A revision of the classification of the Ascomycetes with special emphasis on
the Pyrenomycetes. Mycologia 41: 99-127.
MOLLER, E. & VON ARX, J. A., 1950. Einige Aspekte zur Systematik pseudospharialer Ascomy-
ceten. Ber. Schweiz. Bot. Ges. 60: 329-397.
MULLER, E. & VON ARX, J. A., 1962. Die Gattungen der didymosporen Pyrenomyceten. Beitr.
Kryptogamenflora Schweiz 11: 1-922.
189
PETRAK, F., 1923. Mykologische Notizen. V. No. 199. Ober Didymella, Mycosphaerellopsis und
verwandte Gattungen. Ann. Myc. 21: 19-30.
PETRAK, F., 1924. Mykologische Notizen. VII. No. 305. Ober Didymellina pinodes (Berk. & Blox.)
v. Hohn. Ibid. 22: 16-18.
RIKER, A. J. & RIKER, REGINA S., 1936. Introduction to research on plant diseases. Publ.
by authors. Univ. Wisconsin. 117p.
SACCARDO, P. A., 1882a. Spltaerella pinodes (B. et Blox.) Niessl. Syll. Fung. 1: 514.
SACCARDO, P. A., 1882b. Didymella Sacc. Ibid. 1: 545.
SACCARDO, P. A., 1891. Mycosphaerella Johans. Ibid. 9: 659.
STONE, R. E., 1912. The life history of Ascochyta on some leguminous plants. Ann. Myc. 10:
564-592.
WARMELO, K. T. VAN, 1966. The cytology of Mycosphaerella pinodes. Bothalia (in press).
WOLF, F. A., 1940. A leafspot fungus on Nyssa. Mycologia 32: 331-335.
EXPLANATION OF FIGURES
Figs. 1-4. — Host material showing ascocarps. 4.3x.
Fig. 1. — Sphaeria pinodes (Berk. Herb. 1879).
Fig. 2. — Mycosphaerella pinodes (CUP-A).
Fig. 3. — Didymella pinodes (Myc. gen. 428).
Fig. 4. — Mycosphaerella pinodes (PRE 42549).
Figs. 5-10. — Development of an ascolocule in Mycosphaerella pinodes (PRE 42549). 500x.
Fig. 5.- — Young ascostroma showing less dense central tissue.
Fig. 6.— Stroma showing developing asci growing into the pseudoparenchyma.
Fig. 7. — Stroma showing advanced locule formation. Pseudoparenchyma still noticeable.
Fig. 8.— Mature locule with open ostiole.
Fig. 9.— Mature locule with asci emerging from the ostiole.
Fig. 10. — Mature locule with copious ascospore liberation.
Figs. 11-16. — Preparations showing asci and ascospores. 500x.
Fig. 11. — Sphaeria pinodes (Berk. Herb. 1879).
Fig. 12. — Mycosphaerella pinodes (CUP-A).
Fig. 13. — Didymella pinodes (Type A) (Myc. gen. 428). Note the remaining pseudoparenchyma.
Fig. 14. — Mycosphaerella pinodes (PRE 42549).
Fig. 15. — Sphaeria pinodes (Berk. Herb. 1879). Note the remaining pseudoparenchyma.
Fig. 16. — Mycosphaerella pinodes (CUP-A). Note the ruptured asci giving the appearance of pseudo-
paraphyses.
Fig. 17. — Preparation of Didymella pinodes (Type B) (Myc. gen. 428) showing immature asci and true
paraphyses with free apices. 500x.
Fig. 18. — Section through pycnidium of Ascochyta pinodes (PRE 42549) showing pycnospores. 500x.
.
.
.
l '■ 11
, ...
.
.
. ',(} • .
.
.
.
.
191
7230691
192
193
194
195
Bothalia 9, 1 : 195-202.
The Cytology of Mycosphaerella pinocles*
by
K. T. van Warmelo
Plant Protection Research Institute, Pretoria
Abstract
The behaviour of nuclei during ascus and spore formation and maturation in Mycosphaerella
pinodes (Berk. & Blox.) Stone was studied by means of the HCl-Giemsa technique. The mechanisms
of dikaryotization could not be determined but typical croziers were formed. Nuclear fusion occurred
in the young ascus and was followed by four conventional successive divisions. At pachytene the
chromosomes could be counted and measured. The haploid chromosome number was found to be
15. Ascospore delimitation could be seen at metaphase IV so that each spore is homokaryotic. Two
nuclear divisions occurred in the ascospores so that each cell became quadrinucleate.
Introduction
In an earlier paper (van Warmelo, 1966) the development of the- ascocarp of
Mycosphaerella pinodes (Berk. & Blox.) Stone was described. As it is parasitic on
pea plants it was thought that knowledge of the possible variation potential could be
of importance in the study and control of this fungus. A review of the literature
showed that there was considerable variation in the nuclear behaviour of different
species. The cytology was, therefore, investigated to obtain an indication of the
variation potential and to obtain data which may be of value in taxonomic studies
of this large group.
The development of the ascocarp and nuclear behaviour preceding ascus formation
in many species of Mycosphaerella is well known from the work of Barr (1958),
Higgins (1920, 1929, 1936), Jenkins (1930, 1938), Latham (1934) and Wolf (1940a, b).
The formation of archicarps and spermatia was reported in several species of
Mycosphaerella, for example M. tulipiferae (Higgins, 1936) and M. arachidicola and
M. berkeleyii (Jenkins, 1938). Each archicarp consisted of a basal carpogonium and
tricbogyne, embedded in the ascostroma. The haploid spermatial nucleus migrated
down the trichogyne to initiate the dikaryophase by association with the single carpo-
gonial nucleus. After association of the nuclei the carpogonium became the ascogonium.
Conjugate division of the ascogonial nuclei led to the formation of ascogenous hyphae.
In M. tassiana (Barr, 1958) the ascogenous hyphae developed from ascogenous cells.
No archicarps, spermatia or ascogonia were found and the origin of the ascogenous
cells was not determined.
The presence of croziers has been reported in several species of Mycosphaerella
(Higgins, 1929, 1936; Jenkins, 1938) whereas no croziers were found in other species
(Higgins, 1920; Jenkins, 1930; Latham, 1934). Barr (1958) reported that croziers
were not always formed in M. tassiana.
* Based on a thesis submitted to the University of Natal, Pietermaritzburg, in partial fulfilment
of the requirements for the degree of M.Sc. Agric.
196
The number of asci which may develop from one ascogenous hypha was found
to vary in different species of Mycosphaerella. In M. bolleana only one ascus arose
from each ascogenous hypha due to the lack of branching ascogenous hyphae and the
absence of croziers (Higgings, 1920). After conventional ascus formation in M. tulipi-
ferae (Higgins, 1936) and M. brassicicola (Dring, 1961) the terminal and basal cells of
the crozier fused. The nuclei passed into a branch developing from the basal cell, thus
forming a second ascus. In M. per sonata (Higgings, 1929) multiple croziers were formed
by repetition of the process outlined above, resulting in the development of several asci
from one ascogenous hypha.
Barr (1958) investigated nuclear behaviour after ascus formation in Mycosphaerella
tassiana and M. typhae. In M. tassiana karyogamy took place after ascus formation.
Growth of the ascus and nuclear divisions following karyogamy were conventional.
Spore formation occurred coincidentally with the production of eight nuclei in the ascus.
The spore initials became binucleate and, after septum formation, two further divisions
followed resulting in ascospores with quadrinucleate cells. The mycelial cells in this
fungus were uninucleate except for the multinucleate tips. Barr (l.c.) reported the
haploid chromosome number in M. tassiana to be four.
Barr (1958) did not observe meiotic and mitotic stages in Mycosphaerella typhae.
The nuclear sequence, however, appeared to be similar to that in M. tassiana. A
noticeable difference between these two species was the uninucleate ascospore cells
in M. typhae. The chromosome number was not determined.
Considerable differences in nuclear behaviour and ascus development are thus
evident in different species of Mycosphaerella.
Materials and Methods
Culture. — The fungus studied was isolated from ascocarps on naturally infected
pea straw. Cultures were maintained as described earlier (van Warmelo, 1966).
Fixation. — Ascocarp material was fixed overnight in a 3:1 mixture of absolute
ethyl alcohol and glacial acetic acid. The fixed material was then stored in 70 per cent
ethyl alcohol at room temperature after washing for ten minutes in 90 per cent alcohol
followed by ten minutes in 80 per cent alcohol.
Staining. — Mounts for examination were prepared by dissecting out centra from
fixed ascocarps in distilled water. The centra were then transferred to 1 N HC1 at
room temperature for ten minutes and hydrolyzed for 6^ minutes in 1 N HC1 at 60 °C.
After hydrolysis the centra were washed in distilled water and taken through a phosphate
buffer (pH 6-9) to the staining solution.
The Giemsa staining technique as used in several other investigations (Hrushovetz,
1956; Knox-Davies and Dickson, 1960; Ward and Ciurysek, 1962) was employed.
Two drops of stain per ml of buffer solution were used. Staining was done overnight
in tightly sealed glass vials. After staining the centra were mounted in a drop of buffer
on a slide and squashed. The preparations were not permanent so that nuclear stages
were photographed immediately after preparation.
Results
No ascogonia, trichogynes or spermatia were observed so that the mechanism
by which dikaryotization occurs is still unknown.
Centrum. — Within individual centra a wide variety of stages could be found.
Most of the early stages were found in 10-day old cultures whereas centra consisting
mainly of mature asci occurred in 14-day old cultures.
197
Crozier formation. — Croziers, which developed in the conventional manner, were
common in centra from 10-day old cultures (Figs. 1, 2, 3, 4). Nuclear fusion did not
occur in the subterminal cell and the two nuclei passed into the young ascus formed
by the enlargement of the subterminal cell (Fig. 5). The origin of the ascogenous
hyphae could, however, not be traced.
Croziers with anucleate terminal cells were sometimes observed and it is assumed
that the nucleus in the terminal cell degenerated after septum formation. At no time
was fusion of the terminal cell with the basal cell observed.
Ascus formation and development. — The young binucleate asci could be readily
distinguished from the invariably uninucleate cells of the basal tissue (Fig. 29).
Karyogamy took place in the young ascus (Fig. 6) resulting in the diploid nucleus
(Fig. 7), after which the ascus enlarged considerably. Occasionally fusion failed to
occur as two pachytene stages were sometimes seen in one ascus.
Nuclear divisions in the ascus.
Division I. — The young asci grew rapidly during the period from nuclear fusion
to pachytene, after which stage the rate of growth diminished. The diplophase was
of short duration as the reductional division was initiated soon after fusion.
Meiotic prophase was typical consisting of five stages. At leptotene (Fig. 8) the
nuclear membrane was still present. The chromosomes were long and somewhat
indistinct and a large nucleolus could usually be distinguished.
At zygotene (Fig. 9) pairing of chromosomes could be observed. The chromosomes
were long and thin and a nucleolus was usually visible.
At pachytene the paired chromosomes had become much more definite in
appearance. Centromeres could, however, not be distinguished. It was possible to
count and measure individual chromosomes at this stage (Fig. 10) and they were
numbered (Fig. 11) following the system employed by Singleton (1953). The approxi-
mate lengths of the chromosomes were: 1 = 9-5/t; 2 = 8-5^; 3 = 8-5^; 4 = 7-01«;
5 = 6-5/<; 6 = 6-0/r; 7 = 5 • 5/< ; 8 = 5-5/4; 9 = 5-5/4; 10 = 5-0/t; 11 = 4-0/t;
12 = 4-0/t; 13 = 3-5/4; 14 = 3-5/4; 15 = 3.0/x. Chromosome 15 was associated
with the nucleolus.
At diplotene the chromosomes were much contracted and fuzzy in appearance
(Fig. 12), as has been reported by Carr and Olive (1958) and Knox-Davies and Dickson
(1960).
Diakinesis was not observed.
At metaphase I (Fig. 13) individual chromosomes could not be distinguished
except where movement separated one or more chromosomes from the rest. Rod-like
centrioles appeared and the spindles were always parallel to the long axis of the ascus.
At anaphase I the chromosomes were spread irregularly in a band between the two
centrioles (Fig. 14). The distance between the centrioles at mid-anaphase was larger
than at metaphase and it appeared that movement of the centrioles occurred during
anaphase separation. As this stage was seldom seen it was inferred to be of short
duration.
At telophase I the chromosomes were closely grouped at the spindle poles (Fig. 15).
Lagging chromosomes, which were frequently found in Trichometasphaeria turcica
(Knox-Davies and Dickson, 1960), were not observed. One case was seen where the.
two telophase I nuclei were connected by strands of aggregated spindle fibres (Fig. 16).
198
Division II. — This, and all succeeding divisions were mitotic. During prophase
of the second division recondensation of chromosomes took place. At metaphase II
the spindles were again parallel to the long axis of the ascus (Fig. 17). Anaphase II
figures were not observed. The structure of the telophase II (Fig. 18) and interphase II
(Fig. 19) nuclei was identical with that of telophase I and interphase I nuclei respec-
tively.
Division III. — -Prophase III nuclei (Fig. 20) were similar to those found at
prophase II. At prometaphase III (Fig. 21) the arrangement of the chromosomes
was still random. No metaphase or anaphase figures were seen, but the arrangement
of nuclei suggested that the spindles were either parallel with or oblique to the long
axis of the ascus. Telophase III nuclei were closely aggregated, as before (Fig. 22).
Division IV. — Prophase IV (Fig. 23) was similar to previous prophase stages.
At metaphase IV the spindles were not visible but individual chromosomes could be
clearly seen (Figs. 24, 25). At this stage chromosomes could again be counted. The
average number of fifteen found here agreed with the number determined at pachytene.
Occasionally cleavage planes could be detected in the cytoplasm at metaphase IV,
but at telophase IV (Fig. 26) eight one-celled binucleate ascospores were usually clearly
delimited. Transverse septa started to form in the young ascospores soon after
telophase IV.
Nuclear division in the ascospore. In the immature ascospores two further
divisions of the nuclei took place during development. In the mature state the spores
contained four nuclei per cell (Figs. 27, 28). Ascospores were released from the asci
only after this stage had been reached.
Somatic mycelium. — -The cells of young hyphae which developed from germinating
spores were uninucleate although multinucleate terminal cells were occasionally seen.
Cells in the older mycelium were, however, usually multinucleate.
Discussion
In Mycosphaerella pinodes no trichogynes, spermatia or ascogonia were observed,
the earliest stage seen being the early binucleate crozier. It thus appears that ascus
development in this fungus is similar to that of M. tassiana as described by Barr (1958).
The chromosomes of M. pinodes appear to be longer than those of M. tassiana
(Barr, 1958) but are shorter than the chromosomes of Neurospora crassa (Singleton,
1953), Sordaria fimicola (Carr and Olive, 1958) and Trichometasphaeria turcica (Knox-
Davies and Dickson, 1960).
Mycosphaerella tassiana, with chromosome number of 4 (Barr, 1958), is the only
other species of Mycosphaerella of which this character is known. This number is
appreciably lower than the high number of 15 found in M. pinodes. The regularity
and stability of the divisions in the latter species and the lack of lagging and aberrant
chromosomes indicate that the chromosome complement is not an incomplete tetraploid.
The haploid chromosome number of Mycosphaerella pinodes is exceeded only by
the sixteen chromosomes observed in Aleuria ( Peziza ) rut Harts by Wilson (1937). Hirsch
reported twelve chromosomes in Pyronema confluens but the highest number counted
previously among the perithecial and ascolocular forms was ten in Phyllactinia corylea
(Colson, 1938). M. pinodes thus has the highest number of chromosomes known
among the ascolocular and perithecial forms.
The large difference in haploid number between M. tassiana and M. pinodes
suggests that there may be different groups within this genus. The available evidence
indicates a possible relationship between a moniliaceous imperfect stage ( Cladosporium
herbarum) and a low chromosome number (in M. tassiana ) and a sphaeropsidaceous
199
imperfect stage ( Ascochyta pinodes ) and a high chromosome number in (M. pinodes).
A large number of form genera of Fungi Imperfecti are associated with species of
Mycosphaerella, for example Ramularia (Plakidas, 1941), Cercosporella (Doidge, 1950),
Scolecotrichum (Wehmeyer, 1946), Cercospora (Jones, 1944), Stigmina (Smith and
Smith, 1941), Phyllosticta (Wolf, 1940b), Phoma (Brooks, 1953), Piggotia (Wolf and
Davidson, 1941), Ascochyta (Wiant, 1945), Phleospora (Brooks, 1953), Septoria (von
Arx, 1949), Marssonia (Wolf and Davidson, 1941), Septogloeum (Doidge, 1950) and
Cylindrosporium (Doidge, 1950). It may thus be possible to correlate chromosome
number with the imperfect stages of various species of Mycosphaerella. Such correlation
may allow division of the genus into a number of smaller taxa. These relationships,
however, require verification by similar studies of other species in this genus.
Since cytoplasmic cleavage planes occur in the ascus before the 16-nucleate stage,
each ascospore contains two identical nuclei at telophase IV. The mature ascospore
thus contains eight identical nuclei. This confirms the results of Hare and Walker
(1944) who demonstrated by cultural studies that the two cells of the ascospore of
M. pinodes are genetically identical. This mode of ascospore formation is important
in fungi in which the ascospores are not obligately uniseriate. If metaphase IV spindles
are irregularly arranged and spore delimitation occurs at telephase IV, a two-celled
ascospore may incorporate two nuclei of different compatibility reaction. In hetero-
thallic fungi this could give rise to pseudohomothallism (Raper, 1960). But, in M.
pinodes, which is homothallic (Hare and Walker, 1944), it would have no effect.
The significance of the quadrinucleate condition of the mature ascospore cells
of M. pinodes is not clear at present but the similarity with M. tassiana is noteworthy.
The multinucleate condition may be a mechanism of survival as has been suggested
for other fungi with multinucleate spores (Knox-Davies. Pers. comm.).
References
ARX, J. A. VON, 1949. Beitrage zur Kenntnis der Gattung Mycosphaerella. Syd. Ann. Myc. 3:
28-100.
BARR, MARGARET E., 1958. Life history studies of Mycosphaerella tassiana and M. typhae.
Mycologia 50: 501-513.
BROOKS, F. T., 1953. Plant Diseases. Oxford EJniv. Press, London. 457p.
CARR, A. J. H. & OLIVE, L. S., 1958. Genetics of Sordaria fimicola. II. Cytology. Amer. Journ.
Bot. 45: 142-150.
COLSON BARBARA, 1938. The cytology and development of Phyllactinia corylea Lev. Ann.
Bot. (N.S.) 2: 381-102.
DRING, D. M., 1961. Studies on Mycosphaerella brassicicola (Duby) Oudem. Trans. Brit. Mycol.
Soc.’ 44: 253-264.
DOIDGE, ETHEL M., 1950. The South African fungi and lichens. Bothalia 5: 1-1094.
HARE, W. W. & WALKER, J. C., 1944. Ascochyta diseases of canning pea. Wise. Agr. Exp. Sta.
Res. Bull. 150, 31p.
HIGGINS, B. B., 1920. Morphology and life history of some Ascomycetes with special reference
to the’ presence and function of spermatia. Amer. Journ. Bot. 7: 435-444.
HIGGINS, B. B., 1929. Idem II. Ibid. 16: 287-296.
HIGGINS, B. B., 1936. Idem III. Ibid. 23: 598-602.
HIRSCH, HILDE E., 1950. No brachymeiosis in Pyronema confluens. Mycologia 42: 301-305.
HRUSHOVETZ, S. B., 1956. Cytological studies of ascus development in Cochliobolus sativus. Can.
Journ. Bot. 34: 641-651.
JENKINS, W. A., 1930. The cherry leafspot fungus, Mycosphaerella cerasella Aderh., its morphology
and life history. Phytopathology 20: 329-337.
JENKINS, W. A., 1938. Two fungi causing leaf spot of peanut. Journ. Agr. Res. 56: 317-332.
JONES, F. R., 1944. Life history of Cercospora on sweet-clover. Mycologia 36: 518-525.
KNOX-DAVIES, P. S. & DICKSON, J. G., 1960. Cytology of Helminthosporium turcicum and
its ascigerous stage, Trichometasphaeria turcica. Amer. Journ. Bot. 47: 328-339.
200
LATHAM, D. H., 1934. Life history of a Cercospora leaf spot fungus of cowpea. Mycologia 26:
516-527.
PLAKIDAS, A. G., 1941. Purple leaf spot of strawberry. Phytopathology 31: 225-240.
RAPER, J. R., 1960. The control of sex in fungi. Amer. Journ. Bot. 47: 794-808.
SINGLETON, J. R., 1953. Chromosome morphology and the chromosome cycle in the ascus of
Neurospora crassa. Amer. Journ. Bot. 40: 124-144.
SMITH, D. J. & SMITH, C. O., 1941. Species of Stigmina and Stigmella occurring on Platanus.
Hilgardia 14: 205-231.
WARD, E. W. B. & CIURYSEK, K. W., 1962. Somatic mitosis in Neurospora crassa. Amer. Journ.
Bot. 49: 393-399.
WARMELO, K. T. VAN, 1966. The identity of a pea blight fungus in South Africa. Bothalia 9 (1)
(in press).
WEHMEYER, L. E., 1946. Studies on some fungi from North-Western Wyoming. I. Pyrenomycetes.
Mycologia 38: 144-170.
WIANT, J. S., 1945. Mycosphaerella black rot of cucurbits. Journ. Agr. Res. 71: 193-213.
WILSON, IRENE M., 1937. A contribution to the study of the nuclei of Peziza rutilans Fries. Ann.
Bot. (N.S.) 1: 655-671.
WOLF, F. A., 1940a. Cercospora leafspot of red bud. Mycologia 32: 129-136.
WOLF, F. A., 1940b. A leafspot fungus on Nyssa. Mycologia 32: 331-335.
WOLF, F. A. & DAVIDSON, R. W., 1941. Life cycle of Piggotia fraxini, causing leaf disease of
Ash. Mycologia 33: 526-539.
EXPLANATION OF FIGURES
All figures magnified 2000x.
Fig. 1 . — Early binucleate crozier.
Fig. 2. — Binucleate crozier.
Fig. 3. — Conjugate division in a crozier.
Fig. 4. — Quadrinucleate crozier.
Fig. 5. — Binucleate ascus.
Fig. 6. — Ascus with fusing nuclei.
Fig. 7. — Diploid ascus.
Fig. 8. — Leptotene.
Fig. 9. — Zygotene.
Fig. 10 & 11 . — Pachytene.
Fig. 12. — Diplotene.
Fig. 13. — Metaphase I.
Fig. 14. — Anaphase I.
Fig. 15. — Telophase I.
Fig. 16. — Telophase I with strands of aggregated spindle fibres.
Fig. 17. — Metaphase II.
Fig. 18. — Telophase II.
Fig. 19. — Interphase II.
Fig. 20.— Prophase III.
Fig. 21. — Prometaphase III.
Fig. 22. — Telophase III.
Fig. 23. — Prophase IV.
Fig. 24 & 25. — Metaphase IV.
Fig. 26. — Telophase IV.
Fig. 27. — Ascospores with binucleate cells.
Fig. 28. — Ascospores with quadrinucleate cells.
Fig. 29. — Uninucleate pseudoparenchyma.
201
fif
7230691
202
^ f ♦ #
i » ** * *
* •
#>
* «
*
• «
28
203
Botha lia 9, 1 :203-215.
New Species of Ascomycetes and a New Genus of
Sphaeropsidaceae from Transvaal*
by
W. F. O. Marasas
Plant Protection Research Institute, Department of Agricultural Technical Services,
Pretoria
Abstract
A new genus of Sphaeropsidaceae, Oncostroma Batista & Marasas, and its type species O. toddaliae
Batista & Marasas, are described. Five new species of Ascomycetes, Brasiliomyces entadae Marasas
& Rabie, Thyriopsis sphaerospora Marasas, Microcyclus halleriae Marasas & Rabie, Phyllachora
ehrhartae Marasas and Trichopeltum af icanum Batista & Marasas, are described. Two new com-
binations, Trichopeltum carissae (Doidge) Marasas and Trichopeltum kentaniensis (Doidge) Marasas,
are made.
In the course of an investigation of the foliicolous Ascomycetes and Deuteromycetes
occurring in the indigenous forests of the northern Transvaal, the fungi described
below were found to be new species. With the exception of Thyriopsis sphaerospora,
all these fungi were collected on indigenous host plants growing in an ecological habitat
described as Inland Tropical Forest by Acocks (1953). T. sphaerospora was collected
on a Eucalyptus tree planted in a garden.
The type collections of the new species described below are deposited in the
Mycological Herbarium, Plant Protection Research Institute, Pretoria and in the
herbarium of the Commonwealth Mycological Institute, Kew, England.
Brasiliomyces entadae Marasas & Rabie, sp. nov.
Figs. 1, 5.
Colonies epiphyllous; mycelium superficial, white, very thinly effused over the
leaf surface. Cleistothecia very numerous, gregarious, globose-depressed, attached to
the mycelium at the base, at first hyaline, becoming orange, astomous, semitranslucent,
without appendages, 47-74/t in diameter, 32-5-42-5/t high; wall composed of one
layer of large, irregularly lobed cells. Asci four or five per ascocarp, ovate, thin-walled,
briefly stipitate, five- to eight-spored, 45-0-52-5 x25-0-32-5/<. Ascospores elliptical,
hyaline, continuous, 15-0-17-5 X 7 • 5—9 - 5/«.
Conidiophores and conidia not seen.
On living leaves of Entada spicata (E. Mey.) Druce, Entabeni, Louis Trichardt,
Transvaal, Apr. 1964, Marasas & Rabie in PRE 42626 (PRE, holotype); Mariepskop,
Pilgrims Rest Dist., 1962, PRE 42627.
* Part of a thesis submitted in partial fulfilment of the requirements for the M.Sc. (Agric.) degree,
University of Pretoria, Pretoria, 1965.
204
The asci were measured after squashing the cleistothecia in lactophenol. The
cleistothecia were measured in microtome sections. Their dimensions were smaller
than the correct values because of a certain amount of shrinkage which occurred during
preparation.*
Two genera of Erysiphaceae with one-celled ascospores and cleistothecia lacking
appendages are known. Viegas (1944) described the genus Brasiliomyces to accom-
modate a single species, B. malvastri. Blumer & Muller (1964) made a new genus
Salmonia for Erysiphe malachrae Seaver. The distinction between Brasiliomyces and
Salmonia is not at all clear. The author’s collection could be placed in either of these
two genera according to the original generic descriptions. Prof. Emil Muller (personal
communication) is at present trying to obtain the type material of B. malvastri to
undertake a critical study.
Brasiliomyces malvastri , Salmonia malachrae and B. entadae are compared in
Table 1.
Table 1. — Comparison of characters of Brasiliomyces malvastri, Salmonia malachrae
and Brasiliomyces entadae
From this comparison it is evident that three closely related yet distinct species
are involved. The fungus on Entada is tentatively proposed as a new species of
Brasiliomyces until the matter is settled by Prof. Muller’s researches.
Brasiliomyces entadae Marasas & Rabie, sp. nov.
Plagulae epiphyllae; mycelium superficiarium, albidum, sparse effusum; conidio-
phora et conidia non visa. Cleistothecia numerosa, gregaria, globose-depressa, prima
hyalina, deinde aurea, non ostiolata sine appendicibus, 47-74// diam., 32-5-42-5// alt.;
* The microtome sections were pepared according to the following technique suggested by van
Warmelo (personal communication):
A small piece of leaf bearing cleistothecia was embedded in 2 per cent water agar. A small block
of agar with the material embedded in it was then fixed in formal-acetic-alcohol (Johansen, 1940) for
24 hours and dehydrated for at least two hours in each of four changes of Cellosolve (2-ethoxy-ethanol)
(Gurr, 1956). The dehydrated material was placed in xylene for two hours and embedded in molten
wax (B.D.H. Paraffin Wax with Ceresin, congealing point 55°C). Sections of 8 p and 12/t were cut on
a Jung rotary microtome and mounted on slides with Haupt’s adhesive (Johansen, 1940). The wax
was removed by placing the slides in xylene for two hours. The sections were then dehydrated in a
mixture of equal parts of absolute alcohol and xylene and stained in Pianeze Illb (Conn & Darrow,
1946) for 45 minutes. The sections were then washed with 70 per cent ethyl alcohol, briefly differentiated
in acid alcohol, washed in 70 per cent alcohol, dehydrated in a mixture of equal parts of alcohol and
xylene, cleared in xylene and mounted in Canada Balsam
205
parietis ordinis unius cellularum magnarum irregulariter lobatum consistus. Asci in
quoque cleistothecio 5-8, ovati, breviter pedicellati, 5-8-spori, 45 • 0-52 • 5 x 25 • 0-32 • 5 //.
Ascosporae ellipsoideae, hyalinae, continuae, 15 -0-17 -5 x 7 • 5—9 • 5/^.
Hab. in foliis vivis Entadae spicalae (E. Mey) Druce, Entabeni, Louis Trichardt,
Transvaal, Apr. 1964, Marasas & Rabie in PRE 42626 (PRE, holotypus).
Thyriopsis sphaerospora Marasas, sp. nov.
Figs. 9, 10, 11, 12.
Ascostromata scutellate to conical, suborbicular, irregularly tuberculate, amphi-
genous, subcuticular, black, 280-884// in diameter, 65— 120/< high, producing coralloid,
hyaline haustoria in the epidermal cells, guard cells, cells surrounding the stomatal
cavity and the first layer of mesophyll cells; unilocular or multilocular, opening by
irregular apical fissures exposing the ascospores; upper wall stromatic, composed of
several layers of dark brown flattened, very thick-walled cells; basal wall well developed,
composed of two layers of cells, the outer layer composed of light brown, relatively
thin-walled cells, 5-7 -5// high, which extend laterally for some distance beyond the
base of the upper wall as an intracuticular, unicellular layer of radially elongated,
hyaline cells, and an inner layer, 5-10// high, of dark, thick-walled, stromatic cells.
Asci parallel on the basal wall between hyaline, deliquescent pseudoparaphyses, broadly
clavate, wall thickened, especially at the apex, bitunicate, deliquescing before spore
maturation, 28-42-5 x 15-23/t. Ascospores hyaline when immature, becoming dark
brown, thick-walled, subglobose to globose, finely verrucose, uni-septate, 10-15 x
10-12.5//; mature spores collecting in the ascolocular cavity.
On leaves of Eucalyptus camaldulensis Dehn., Nylstroom, Transvaal, Aug. 1963,
Marasas in PRE 42659 (PRE, holotype); Nylstroom, Apr. 1964, PRE 42660.
This fungus agrees with the genus Phaeothyriolum ( = Micromycothelia sensu
Muller & v. Arx, 1962) of the Stigmateaceae as defined by Hansford (1946) in respect
of the subcuticular stromata and the coralloid haustoria. However, the structure of
the upper walls of its stromata is not radial and the stromata open by irregular fissures
and not by definite ostioles as in Phaeothyriolum.
With the key presented by Muller & v. Arx (1962) this fungus can be placed in
the Order Dothiorales because of the bitunicate asci and the presence of ascostromata
opening by fissures to expose the spores. Within that order it can be placed in the
family Leptopeltaceae because of the shield-shaped, subcuticular stromata. Although
most members of this family are saprophytes on ferns and conifer needles, some occur
on phanerogams (Leptopeltella) and some penetrate the epidermis to form cell com-
plexes or hyphae (Muller & v. Arx, 1962). Within the Leptopeltaceae it fits best into
the genus Thyriopsis, which is however assumed to be limited to conifer needles by
Muller & v. Arx (1962).
The fungus on Eucalyptus agrees with the genus Thyriopsis in the subcuticular,
scutellate ascostromata opening by irregular fissures; the basal layer of parallel,
broadly clavate, bitunicate asci developing between pseudoparaphyses and the uni-
septate ascospores which are dark brown at maturity. The edge cells of the stromata
of Thyriopsis halepensis (Cooke) Theiss. & Syd., the only known species of the genus,
are radially elongated. The basal wall of the stromata of T. sphaerospora is extended
radially as an intracuticular, unicellular layer of radially elongated cells. Numerous
haustoria are produced by these radial extensions of the stromata. This can possibly
be interpreted as a more highly developed adaptation to the parasitic habit than that
found in T. halepensis.
206
T. sphaerospora differs from the type species of Thyriopsis by having the upper
walls of the stromata composed of several layers of stromatic cells and the basal walls
well developed and composed of two layers of cells. Furthermore it differs from
T. halepensis in that it occurs on a phanerogam ( Eucalyptus ) and not on conifer needles,
in the presence of well developed haustoria and in the characteristically subglobose
to globose instead of oblong ascospores.
The author does not consider these differences to be of generic rank and this
species on Eucalyptus is thus described as the second species of the genus Thyriopsis
Thyriopsis sphaerospora Marasas, sp. nov.
Ascostromata scutata, orbicularia, irregulariter tuberculata amphigena, subcuti-
cularia, atra, 280-884// diam , 65—1 20/t alta, haustoriis hyalinis, coralloidibus; unilo-
cularia vel multilocularia, rimis irregularibus dehiscentia, strato tegente ordinum plurum
cellularum complanetarum parietibus crassis composito; membrana basali duorum
ordinum cellularum composita. Asci paralleli in membrana basali inter pseudo-
paraphyses hyalinos deliquescentos; lata clavati, bitunicati, octospori, deliquescenti,
28 • 0-42 -5 x 1 5-23 //. Ascosporae primum hyalinae, deinde atro-brunneae, subglobosae
vel globosae, parietibus crassis, minute verrucolosae, uniseptatae, 10-15 x 10-12-5//.
Hab. in foliis Eucalypti camaldulensis Dehn., Nylstroom distr., Transvaal, Marasas
in PRE 42659 (PRE, holotypus).
Microcyclus halleriae Marasas & Rabie, sp. nov.
Figs. 3, 15, 16.
Stromata hypophyllous, surrounded by reddish-brown haloes on yellowish dis-
coloured areas of leaf tissue, erumpent, gregarious, irregularly tuberculate, 65-150//
in diameter, 64-78// high; stroma not very well developed, causing a yellowish
discolouration of the epidermal cells and the subepidermal layers of mesophyll cells;
outer stromatic wall composed of several layers of dark-brown, thick-walled cells,
9-13 X 4-7//; attached to the hypostroma over its entire width or becoming narrower
to form an indistinct foot, 38-80// wide; locules one to six in each stroma, subglobose
to ellipsoidal, 32-80// wide, 42-58// high, opening by an apical, flat, rounded ostiole;
filled with hyaline, thin-walled cells which are replaced by the asci. Asci few, ovate
or obclavate, ventricose, stipitate, apical wall prominently thickened, aparaphysate,
8-spored, 30-38 x 12-5-17-5//. Ascospores polystichous, oblong-elliptical, hyaline,
immature spores uniseptate in the middle, mature spores three-septate, slightly con-
stricted, rounded at both ends, 16-19 x 4-5//.
On leaves of Halleria lucida L., Entabeni, Louis Trichard, Transvaal, Oct. 1963,
Marasas & Rabie in PRE 42687 (PRE, holotype); Entabeni, Louis Trichardt, Apr.
1964, PRE 42688.
This fungus is not a typical representative of the genus Microcyclus and differs
from the species of this genus described by Doidge (1948), Miiller & Sanwal (1954)
Vital (1956), Tilak (1959) and Muller & v. Arx (1962) in a number of characters.
The stroma is not connected with the hypostroma by a definite foot; the asci are
ventricose and not clavate to cylindrical as in most of the other species; the spores
become three-septate when mature.
It resembles other species of Microcyclus in the presence of erumpent, tuberculate,
uni- or multi-locular stromata, the absence of paraphysoids, the immature one-septate
spores of which each cell is later divided by a septum so that the mature spores are
four-celled (see Fig. 3). This process appears to be similar to the division of contents
207
(Inhaltsteilung) described by Sydow (1939) as taking place in the ascospores of Gilletiella
chusqueae. Gilletiella was however considered a didymosporous genus of the Pleos-
poraceae by Muller & v. Arx (1962) although some spores may have more than one
septum.
This fungus is somewhat intermediate between Mycosphaerella and Microcyclus
but because definite loculate stromata are present, it is placed in the latter genus rather
than in Mycosphaerella. As this species differs in all respects from those described
previously, it is proposed as a new species of Microcyclus'.
Microcyclus halleriae Marasas & Rabie, sp. nov.
Ascostromata hypophylla, sine maculis, decoloratione flavo textus folii et corona
rufo-brunneola circumdata, gregaria, irregulariter tuberculata, 65—1 50// diam., 64-78/*
alta; paries exterioris ordinum plurum cellularum brunnearum parietibus crassis,
9-13 X 4-7/* consistus; basi tota ad hypostromata fixa vel in pedem 38-80/* latum
attenuata. Loculi immersi, usque 6 in quoque stromate, subglobosi vel ellipsoidei,
32-80/* lati, 42-58/* alti; ostiolo complaneto rotundo. Asci ovati vel obclavati,
ventricosi, stipitati, aparaphysati, octospori, 30-38 x 12 -5-1 7 -5/*. Ascosporae poly-
stichae, oblongae, hyalinae, immaturae medio uniseptatae, maturae triseptatae, vix
constrictae, 16-19 x 4-5/*.
Hab. in foliis vivis Halleriae lucidae L., Entabeni, Louis Trichardt, Transvaal,
Marasas & Rabie in PRE 42687 (PRE, holotypus).
Phyllachora ehrhartae Marasas, sp. nov.
Figs. 2, 13, 14.
Ascostromata amphigenous in straw-coloured discoloured areas, the whole leaf
dying later and becoming straw-coloured, shiny-black, scattered, well-defined, convex,
lengthened, oblong to ellipsoidal or irregular in outline, developing parallel to the veins,
0 • 5-2 -5x0- 2-0 • 5 mm. Clypeus well developed on both sides of the leaf but extending
through more cells of the abaxial epidermis, completely filling the infected epidermal
cells, very dense, opaque, black, 6-5-19/* thick; stroma in the mesophyll rather poorly
developed. Perithecia one to nine in each stroma, ellipsoid, occupying the entire
thickness of the leaf, 109-330/* in diameter, 109-182/* high; ostiole, completely
immersed in the clypeus, periphysate; perithecial wall well defined, merging above
and below with the clypeus, brown, concentric fibrose, 6- 5-12-5/* thick. Asci cylin-
drical, rounded at the apex, attenuated below to a short, straight or geniculate pedicel,
paraphysate, eight-spored, 64-80 x 9-5/*.; paraphyses numerous, filiform, branched,
hyaline, 1-2 p. in diameter. Ascospores obliquely monostichous, subglobose to broadly
oval, hyaline, continuous, 6-5-13 X 5-0-6-5/x.
On living leaves of Ehrharta erecta Lam., Mariepskop, Pilgrims Rest distr., Dec.
1963, Marasas in PRE 42628 (PRE, holotype); Entabeni, Louis Trichardt, Apr. 1964,
PRE 42629.
According to Dr. D. G. Parberry (personal communication), Phyllachora species
are usually confined to hosts in one or a few closely related tribes of grasses. At
present the genera Ehrharta, Tetrarhena, Microlaena and Petriella are generally included
in the tribe Ehrharteae (Tateoka, 1963). Only one record of a Phyllachora sp. on any
one of these genera could be traced. Doidge (1950) cited: “ ? Phyllachora sp. (Syn.
Sphaeria graminis Pers. var. ehrhartae Berk.) on Ehrharta sp., Uitenhage Drege 83 ”.
This fungus was actually described by Berkeley (1843) from Zeyher 83 and not Drege 83
as cited by Doidge (1950). The asci were described as clavate, swollen in the middle,
apiculate and with a double row of oblong-ovate spores. No material of this fungus
is available in the Mycological Herbarium, Pretoria and it was not included in the
208
Sylloge Fungorum or in the treatise by Theissen & Sydow (1915). According to
Berkeley’s description, it also differs from the present collection in the characters of
the asci and ascospores. As no other species of Phyllachora are known on the tribe
Ehrharteae, it was considered necessary to propose a new species:
Phyllachora ehrhartae Marasas, sp. nov.
Ascostromata amphigena, sine maculis definitis decoloratione flavobrunnea cincta,
atra, nitida, sparsa, convexa, oblonga vel ellipsoidea aut irregularia, venis parallela
evoluta, 0-5-2 -5 X 0-2-0 -5 mm; clypeo amphigeno bene evoluto sed in epidermide
hypophyllo per plus cellulas extento, 6-5-19-0// crasso. Perithecia in quoque stromate
1-9, ellipsoidea, 10-330// diam., 109-182// alta; ostiolo piano, papilliformi, in clypeo
omnino immerso; pariete perithecii 6-5-12-5// crasso, brunneo, concentrice fibroso.
Asci cylindracei, superne rotundati, inferne pedicello breve recto vel geniculato,
octospori, 64-80 X 9-5 //; paraphyses numerosae, filiformes, ramosae, hyalinae, 1-2//
crassae. Ascosporae oblique monostichae, subglobosae vel late ovatae, hyalinae,
continuae, 6-5-13 x 5 -0-6 -5/t.
Hab. in foliis vivis Ehrhartae erectae, Mariepskop, Pilgrims Rest distr., Transvaal,
Marasas in PRE 42628 (PRE, holotypus).
Trichopeltum africanum Batista & Marasas, sp. nov.
Figs. 4, 6.
Free mycelium forming an epiphyllous, thalloid mycelial membrane; membrane
radiate, prosenchymatous, linear, branched, brown-black, single or coalescing, 75—3 10/*
broad, composed of narrow rectangular cells, 5-5-11-0 x 3 -0-3 -5/t. Ascostromata
formed under the mycelial membrance by pycnosis, applanately hemispherical, dimidiate,
darker than the membrane, 115-192// in diameter; upper wall composed of two layers
of cells, 9-11// thick; basal wall inconspicuous. Asci ellipsoid to subclavate, broadly
rounded at the apex, apical wall thickened, briefly stipitate, eight-spored, aparaphysate,
28-33 x 10-16-5//. Ascospores polystichous, oblong-clavate, straight, hyaline, one- to
four-septate, slightly contricted, 10-15 X 4//.
On leaves of Clivia sp., Mariepskop, Pilgrims Rest distr., Transvaal, 1962, Marasas
& Rabie in PRE 42707 (PRE, holotype).
Also present: Vitalia ekmanii (Petr. & Cif.) Bat. & Cif. (PRE 42708).
Batista, Costa & Ciferri (1957) proposed the new genus Trichopeltum as a substitute
for Trichopeltis Speg., which is a synonym of Trichothyrium according to Hughes
(1953). The genus Trichopeltum accommodates fungi belonging to the Trichopeltinaceae
sensu Batista, Costa & Ciferri and having a linear mycelial membrane, hyalophrag-
mospores and lacking paraphysoids. In the genus Trichopeltum the ascostromata
develop beneath the mycelial membrane by pycnosis and in this character it differs
from Trichothyrium which has superficial ascostromata.
Two species of Trichopeltula were described by Doidge (1922). The genus Tricho-
peltula Theissen was considered to be a synonym of Trichopeltis by Clements & Shear
(1931). Trichopeltis is however a synonym of Trichothyrium (Hughes, 1953). The
type of the genus Trichopeltula (T. hedycaryae Theiss.) will have to be examined to
determine whether it belongs in Trichothyrium or in the Trichopeltinaceae. Trichopeltula
Theissen (Zentr. Bakteriol. 39: 636, 1914), is not a valid name because it is a later
homonym of Trichopeltulum Speg. (Fung. Puigg. No. 342, 1889). If Trichopeltula
hedycaryae is found to belong in the Trichopeltinaceae it may probably be placed in
the genus Trichopeltum.
209
The types of Trichopeltula carissae Doidge and T. kentaniensis Doidge in the
Mycological Herbarium, Pretoria were examined. The ascostromata of both these
species develop underneath the mycelial membrane. Accordingly the following new
combinations are proposed for these two species: Trichopeltum carissae (Doidge)
Marasas, comb. nov. ( Trichopeltula carissae Doidge in Bothalia 1 : 78. 1922) and
Trichopeltum kentaniensis (Doidge) Marasas, comb. nov. ( Trichopeltula kentaniensis
Doidge in Bothalia 1 : 79. 1922).
The fungus on Clivia is closely related to both these species and appears to be
somewhat intermediate between them. These two species and PRE 42707 differ from
the type of the genus ( Trichopeltum hawaiiensis Bat. & Costa) in that their mycelial
membranes are branched, the ascostromata are ostiolate and the spores are four-septate.
Because the author’s collection cannot be satisfatorily placed in any one of these species
and was also collected on a hitherto unreported host, it is described as a new species:
Trichopeltum africanum Batista & Marasas, sp. nov.
Thallus epiphyllus, ramosus, vittaeformis; vittae brunneo-atra, radiato-prosen-
chymaticae, 75—3 1 0/^ latae, cellulis anguste rectangularibus, 5-5-11 X 3-3-5//. Asco-
stromata sub thallo formata, applanato-hemisphaerica, thallo ateriora, 115-192//
diametro; ostiolo rotundato, 7-5-11// diametro; paries superus ordinum duorum
cellularum consistus. Asci ellipsoidei vel subclavati, aparaphysati, supra late rotundati,
breviter pedicellati, octospori, 28-33 X 10-16-5//. Ascosporae polystichae, oblongo-
clavatae, rectae, hyalinae, 1-4-septatae, leviter constrictae, 10-15 X 4//.
Hab. in foliis Cliviae sp., Mariepskop, Pilgrims Rest distr., Transvaal, 1962,
Marasas & Rabie in PRE 42707 (PRE, holotypus).
Oncostroma Batista & Marasas, gen. nov.
Mycelium superficial, brown, with hyphae radiating from the pycnostromata to
form a loose mycelial network. Pycnostromata superficial, developing underneath
the mycelial network, brown-blackish, membranous, uni- or bi-loculate, ostiole not
defined. Conidiophores subulate, hyaline, simple or branched, at first continuous
later becoming septate. Pycnidiospores ellipsoid to fusoid, continuous, smooth, hyaline.
Type: Oncostroma toddaliae Batista & Marasas.
Etym. : Gr. ’o/dcos = bulk, mass and stroma.
Oncostroma Batista & Marasas, Sphaeropsidaecearum.
Mycelium superficiale, brunneum, hyphis ab pycnostromatibus radiatis; pycno-
stromata superficialia, sub reticulo hypharum evoluta, atro-brunnea, membranosa,
uni- vel biloculata, sine ostiolo definiti; conidiophorae subulatae, hyalinae, simplices
vel ramosae, primum continuae deinde septatae; pycnidiosporae ellipsoideae vel
fusoideae, continuae, hyalinae, leves.
Typus: Oncostroma toddaliae Batista & Marasas.
Oncostroma toddaliae Batista & Marasas, sp. nov.
Fig. 7, 8.
Mycelium hypophyllous, superficial, composed of sparingly branched, few septate,
brown hyphae, 2-5-4// wide, radiating from the pycnostromata to form a loose mycelial
network; haustoria coralloid, hyaline to olivaceous, developing in the leaf glands
only. Pycnostromata formed underneath the mycelial network, brown-blackish, uni-
or bi-loculate, without a defined ostiole, 144-384// in diameter; walls stromatic, pseudo-
parenchymatic, soft, composed of subglobose cells, 2-4-5// in diameter. Conidiophores
210
subulate, hyaline, simple or branched, at first continuous but becoming septate,
4 -0-7 -5 X 1 • 5—3 • 0//. Pycnidiospores abundant, ellipsoid or fusoid, continuous,
smooth, hyaline, 7-5-12-5 x 2-0-2- 5//.
On living leaves of Toddalia asiatica (L.) Lam., Mariepskop, Pilgrim’s Rest Dist.,
Transvaal, 1962, Marasas & Rabie in PRE 42662 (PRE, holotype); F. C. Erasmus
Reserve, Bosbokrand, Jan. 1963, PRE 42663.
The stromata develop over the leaf glands and coralloid haustoria are present
in the cells of these glands. For these reasons it is assumed that Oncostroma toddaliae
is parasitic on leaf glands.
The new genus Oncostroma does not appear to be closely related to any other
genus of stromatic Sphaeropsidaceae. The closest genera according to the key presented
by Clements & Shear (1931) are Phellostroma and Ascochytopsis. The pycnostromata
of Phellostroma are however, suberose and the conidia are falcate. According to the
key presented by Barnett (1960), the new genus approximates Creothyriella, but the
pycnostromata of the latter genus contain several well defined, globose loculi and
catenate conidia.
The genus Oncostroma is characterised by the one- or two-loculate pycnostromata
developing beneath the mycelial network, the haustoria in the leaf glands, the subulate
conidiophores and the one-celled, hyaline, ellipsoid or fusoid pycnidiospores.
Oncostroma toddaliae Batista & Marasas, sp. nov.
Mycelium hypophyllum, superficiale, hyphis ab pycnostromatibus radiatis, paulo
ramosis, paulo septatis, brunneis, 2 -5-4- (fit latis, haustoriis in glandis foliorum, coral-
loidibus, hyalinis vel olivaceis. Pycnostromata superficialia, sub reticulo hypharum,
brunneo-atra, uni- vel biloculata, sine ostiolo definito, pseudoparenchymatica, mollia,
144-384, u diam., ex cellulis subglobosis, 2-4 -5/t diam. composita. Conidiophorae
subulatae, hyaline, simplices vel ramosae, primum continuae deinde septatae, 4-7 • 5 x
1 • 5-3 /n. Pycnidiosporae copiosae, ellipsoideae vel fusoideae, continuae, hyalinae,
leves, 7-5-12-5 X 2-0-2- 5^.
Hab. in foliis vivis Toddaliae asiaticae (L.) Lam., Mariepskop, Pilgrim’s Rest
distr., Transvaal, Marasas & Rabie in PRE 42662 (PRE, holotypus).
Acknowledgements
The author wishes to express his sincere thanks and appreciation to Dr. C. J. Rabie
of the Department of Plant Pathology, University of Pretoria, for his interest and
invaluable assistance during the course of the work; Prof. Dr. E. Muller of the Institut
fur spezielle Botanik, Eidgenossische Technische Hochschule, Ziirich, Switzerland, for
his generous advice on the taxonomic position of certain new species and for material
of Thyriopsis halepensis; Prof. Dr. A. C. Batista of the Instituto de Micologia,
Universidade do Recife, for his kind advice and criticism of certain descriptions;
Mr. K. T. van Warmelo and Mr. G. C. A. van der Westhuizen of the Plant Protection
Research Institute, Pretoria for their unfailing interest, advice and encouragement
and invaluable assistance with the microtechnique and photomicrography; Dr. D. J. B.
Killick of the Botanical Research Institute, Pretoria for checking the Latin diagnoses;
the staff of the National Herbarium, Pretoria for the identifications of the host plants
and the Department of Forestry and the Department of Nature Conservation of the
Transvaal Provincial Administration for permission to collect fungi at the various
forestry stations and nature reserves.
211
References
ACOCKS, J. P. H., 1953. Veld types of South Africa. Mem. Bot. Surv. S. Afr. 28: 1-192.
BARNETT, H. L., 1960. Illustrated genera of imperfect fungi. Second Ed. Burgess Publ. Co.,
Minn. 225 p.
BATISTA, A. C., COSTA C. A. A. & CIFERRI, R., 1957. Organogenese e sistemdtica dos fungos
Tnchopeltinaceae (Theiss.) emend nobis. Inst. Mic. Univ. Recife Publ. 90: 1-24.
BERKELEY, M. J., 1843. Enumeration of fungi collected by Elerr Zeyher in Uitenhage. Hooker’s
London J. Botany II: 507-524.
BLUMER, S. & MULLER, E., 1964. Uber zwei Mehltauarten auf Baumwolle aus Peru. Phytopathol.
Z. 50: 379-385.
CLEMENTS, F. E. & SHEAR, C. L., 1931. The genera of fungi. H. W. Wilson Co., New York.
496 p.
CONN, H. J. & DARROW, MARY A., 1946. Staining procedures used by the Biological Stain
Commission. Biotech Publications. Geneva, New York.
DOIDGE, ETHEL M., 1922. South African Ascomycetes in the National Herbarium. II. Bothalia
1 : 65-82.
DOIDGE, ETHEL M., 1948. South African Ascomycetes in the National Herbarium. VI. Bothalia
4: 837-880.
DOIDGE, ETHEL M., 1950. The South African fungi and lichens. Bothalia 5: 1-1094.
GURR, E., 1956. A practical manual of medical and biological staining techniques. Leonard Hill
Ltd., London. 320 p.
HANSFORD, C. G., 1946. The foliicolous Ascomycetes, their parasites and associated fungi. Mycol.
Papers, Commonwealth Mycol. Inst. 15: 1-240.
HUGHES, S. J., 1953. Fungi from the Gold Coast. II. Mycol. Papers, Commonwealth Mycol.
Inst. 50: 1-104.
JOHANSEN, D. A., 1940. Plant Microtechnique. McGraw-Hill Book Co., New York. 523 p.
MULLER, E. & SANWAL, B. D., 1954. Uber die Gattungen Microcyclus Sacc., Coccoidella v.
H5hn., Coccodothis Theiss. & Sydow und Coccodothella Theiss. & Sydow. Sydowia 8: 231-244.
MULLER, E. & VON ARX, J. A., 1962. Die Gattungen der didymosporen Pyrenomyceten. Beitr.
Krypt. FI. Schweiz 11 (2): 1-922.
SYDOW, H., 1939. Fungi Aequatoriensis. Ann. Mycol. 37: 275-438.
TATEOKA, T., 1963. Notes on some grasses XIII. Relationships between Oryzeae and Ehrharteae
with special reference to leaf anatomy and histology. Botan. Gaz. 124: 264-270.
THEISSEN, F. & SYDOW, H., 1915. Die Dothideales. Kritisch-systematische Originaluntersuch-
ungen. Ann. Mycol. 13: 149-746.
TILAK, S. T., 1959. A new species of Microcyclus from India. Sydowia 12: 197-199.
VIEGAS, A. P., 1944. Alguns fungos do Brasil. II. Ascomisetos. Bragantia 4: 1-392.
VITAL, A. F., 1956. Una nova especie de Microcyclus associado a Stromatopycnis. Inst. Mic.
Univ. Recife Publ. 15: 1-6.
EXPLANATION OF FIGURES
Fig. 1-4. — Fig. 1. Brasiliomyces entadae, ( a ) ascocarp; ( b ) ascus; (c) ascospores; (d) five asci being
released from a squashed ascocarp. Fig. 2. Phyllachora ehrhartae, (a) asci; ( b ) ascospores.
Fig. 3. Microcyclus halleriae, (a) asci: (b) immature and mature spores. Fig. 4. Trichopeltum
africanum, (a) part of mycelial membrane showing a portion of an ascostroma; ( b ) asco-
spores; (c) asci.
Fig. 5-8. — Fig. 5. Brasiliomyces entadae, cleistothecia, X 500. Fig. 6. Trichopeltum africanum, mycelial
membrane with ascostroma, X 125. Fig. 7, 8. Oncostroma toddaliae. Fig. 7. Pycnidia
under hyphal network, X 125. Fig. 8. Pycnidiospores, X 500.
Fig. 9-12. — Thyriopsis sphaerospora. Fig. 9. Cross section of biloculate ascostroma, X 156. Fig. 10.
Cross section of stroma showing covering and basal membranes and ascospores in deli-
quescent asci, x 625. Fig. 11. Ascostromata on leaf surface, X 20. Fig. 12. Ascospores,
x 1250.
Fig. 13-16. — Fig. 13, 14. Phyllachora ehrhartae. Fig. 13. Longitudinal section of stroma showing five
perthecia, X 156. Fig. 14. Cross section of stroma, X 400. Fig. 15, 16. Microcyclus halleriae.
Fig. 15. Cross section of uniloculate stroma, x 500. Fig. 16. Cross section of stroma
showing three loculi, x 625.
212
213
7230691
214
215
217
Bothalia 9, 1 : 217-227.
New and Interesting Records of South African
Fungi, Part IV*
by
W. F. O. Marasasf and C. J. Rabie 1
During 1962, 1963 and 1964 a large number of foliicolous fungi were collected
at the following forestry stations in the Transvaal Province of the Republic of South
Africa: De Hoek Forestry Station, Tzaneen, Letaba Distr. ; Entabeni Forestry Station,
Louis Trichardt, Soutpansberg Distr.; Mariepskop Forestry Station, Pilgrims Rest
Distr.
The fungi described here were found to be new records for the Republic of South
Africa. All the host plants on which these fungi were collected also represent new host
records for the various fungi.
The specimens are deposited in the Mycological Herbarium, Plant Protection
Research Institute, Department of Agricultural Technical Services, Pretoria.
1. Parapeltella maitlandii (Hansf.) Batista, in Inst. Mic. Univ. Recife Publ. 56:
292(1959).
Micropeltella maitlandii Hansford, Proc. Lin. Soc. London 153: 49(1941).
Fig. 1.
Free mycelium absent. Ascostromata hypophyllous, superficial, olivaceous black
with a hyaline margin, glabrous, composed of reticulate hyphae, 280-400/1 in diameter,
opening with an irregular pore. Asci bitunicate, wall more or less uniformly thickened,
obclavate, sessile or briefly stipitate, six to eight-spored, 60-85 x 17-24//. Ascospores
at first obliquely monostichous, becoming polystichous, clavate, hyaline, two- to six-
septate, constricted at the septa, cells not equal in size, 29-44 x 5-7//.
Specimen examined". PRE 42621, on leaves of Rawsonia lucida Harv. & Sond.,
De Hoek, Tzaneen, Letaba Distr., Transvaal, Oct. 1963.
Rawsonia lucida is a new host record for P. maitlandii which had been collected
previously on Apocynaceae and Rubiaceae in Uganda (Batista, loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
2. Micropeltidium obscurum Batista & Bezerra, in Inst. Mic. Univ. Recife Publ.
391: 18(1963.)
* Part of a thesis submitted by the senior author in partial fulfilment of the requirements for the
M.Sc. (Agric.) degree. University of Pretoria, Pretoria, 1965.
f Plant Protection Research Institute, Department of Agricultural Technical Services, Pretoria.
J Department of Plant Pathology, University of Pretoria, Pretoria.
218
Fig. 2.
Free mycelium absent. Ascostromata hypophyllous, superficial, olivaceous black
with hyaline to subhyaline margins, glabrous, rounded, composed of reticulate hyphae,
opening with an irregular, stellate fissure, 372-390/* in diameter. Asci bitunicate, wall
more or less uniformly thickened, clavate, attenuated towards the base or briefly
stipitate, two- to eight-spored, 54-83 X 12-22/*; paraphysoids present, filiform,
branched, 2-5/* in diameter. Ascospores polystichous, clavate, hyaline, with two to
twelve transverse septa, constricted at the septa, cells not equal in size, becoming
separated, 27-42/* long, the basal cell 2- 5-5-0/* and the apical cell 4 -5-7 -5/* wide.
Specimen examined : PRE 42620, on leaves of Ilex mitis (L.) Radi., De Hoek,
Tzaneen, Letaba Distr., Transvaal, Jan. 1963.
This is the first record of the occurrence of this fungus in South Africa.
3. Scolecopeltidium racemosae Batista & Lima apud Batista, in Inst. Mic. Univ.
Recife Publ. 56: 222 (1959).
Fig. 3.
Free mycelium absent. Ascostromata hypophyllous, superficial, scattered, rounded,
olivaceous black with a hyaline margin, upper membrane meandriform, 490-500/* in
diameter, ostiole central, round, 65-75/* in diameter. Asci bitunicate, cylindrical to
obclavate, briefly stipitate, two- to six-spored, 81-110 X 17-23/*; paraphysoids filiform,
branched, hyaline, up to 2/< wide. Ascospores distichous to polystichous, clavate,
hyaline, with five to twelve transverse septa, constricted at the septa, 37-104 X 5-8/*,
cells becoming separated.
Specimen examined : PRE 42622, on leaves of Syzygium gerrardii (Harv.) Hochst.,
De Hoek, Tzaneen, Letaba Distr., Transvaal, Oct. 1963.
The asci mostly contained immature spores or no spores at all and only a few
mature phragmospores were seen.
Syzygium gerrardii is a new host record for S. racemosae which is known on
Paullinia from Brazil (Batista, loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
4. Chaetothyrina musarum ( Speg .) Theissen, in Ann. Mycol. 11: 495 (1913);
Batista, Inst. Mic. Univ. Recife Publ. 56: 448 (1959); Muller & v. Arx, Beitr. Krypt.
FI. Schweiz 11 (2): 548 (1962).
Fig. 8.
Colonies hypophyllous, black, confluent; mycelium superficial, pelliculose; hyphae
septate, branched, reticulate, olive-brown, 2 -5-6-0/* in diameter; mycelial setae
scattered, rigid, septate, simple, attenuated, obtuse, straight or curved, dark brown,
165-213/* long, 5/* wide at the base and 2-5/* at the apex. Ascostromata gregarious,
globose-depressed, ostiolate, olive-brown, 126-165/* in diameter, composed of mean-
driform hyphae, setose; setae like the mycelial setae. Asci bitunicate, obclavate,
briefly stipitate, eight-spored, 42-49 X 12-14/*; paraphysoids present, clavate, septate,
hyaline, up to 2-5/* wide at the apex. Ascospores oblong-clavate, septate, hyaline,
12-15 X 3-5/*.
Specimen examined : PRE 42619, on leaves of Euclea crispa (Thunb.) Guerk. var.
crispa, Mariepskop, Pilgrims Rest Distr., Transvaal, Dec. 1963.
219
Euclea crispa var. crispa is a new host record for C. musarum which is known to
occur on Musa, Smi/ax, and various other unrelated host plants in tropical America,
Africa and Asia (MUller & v. Arx, loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
5. Ainsworthia xanthoxylii Batista & Costa apud Batista & Ciferri, in Inst. Micol.
Univ. Recife Publ. 158: 7 (1962).
Fig. 4.
Mycelium epiphyllous, superficial, pelliculose, not setose, composed of hyaline,
septate hyphae with cylindrical cells. Ascostromata globose-depressed, membranous,
brown, semitranslucent, 164-190// in diameter. Asci numerous, subglobose to clavate,
sessile or briefly stipitate, 37-44 x 17-22 //; paraphysoids present, filiform, coiled at
the apex, hyaline, 1// thick. Ascospores polystichous, cylindraceous, straight to slightly
curved, hyaline but becoming subhyaline when mature, seven to eight transverse and
two to four longitudinal septa, constricted at the septa, 29-42 x 5-5-10//.
Specimen examined : PRE 42614, on living leaves of Ilex mitis (L.) Radik., De Hoek
Forestry Station, Tzaneen, Letaba District, Transvaal, Jan. 1963.
This is a new host record for A. xanthoxylii which had been collected previously
on Ilex glabra and various other unrelated host plants in the United States and South
America. (Batista & Ciferri, loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
6. Ainsworthia roraimensis Batista & Cavacanti, in Inst. Micol. Univ. Recife Publ.
431 (1964).
Fig. 5.
Mycelium hypophyllous, superficial, pelliculose, composed of hyaline, reticulate,
septate hyphae with cylindrical cells, 5 -0-7 -5// wide. Ascostromata single, scattered,
globose-flattened, not setose, brown, semitranslucent, 152-210// in diameter. Asci
subglobose to broadly clavate, sessile, eight-spored, aparaphysate, 54 x 34//. Ascos-
pores conglobate, polystichous, cylindraceous to clavate, straight or curved, hyaline,
with seven to fifteen transverse and none to two longitudinal or oblique septa, slightly
constricted at the septa, 39-69 x 7-10//.
Specimen examined : PRE 42613, on living leaves of Rawsonia lucida Harv. & Sond.,
De Hoek Forestry Station, Tzaneen, Letaba District, Transvaal, Oct. 1963.
This is the first record of the occurrence of this fungus in South Africa.
7. Deslandesia javanica ( Zimm .) Batista & Ciferri var. harana ( Trotter ) Batista &
Ciferri, in Inst. Micol. Univ. Recife Publ. 158: 44 (1962).
Fig. 6.
Mycelium epiphyllous, superficial, pelliculose, not setose; hyphae brownish,
branched, septate, constricted, composed of cylindrical cells, 12-5-20 X 4—5//. Asco-
stromata single, scattered, globose-depressed, membranous, glabrous, brown, semi-
translucent, pseudo-ostiolate, 154-260// in diameter. Asci ovoid to ellipsoid, rarely
clavate, sessile or briefly stipitate, wall prominently thickened apically when immature
7230691-7
220
but becoming uniformly thickened when mature, six- to eight-spored, 50-73 x 23-33/*.
Ascospores polystichous, cylindraceous to clavate, straight or slightly curved, hyaline
to subhyaline, six to nine transverse septa, apical cells becoming muriform, constricted
at the septa, 35-68 x 6-5-10/*.
Specimen examined : PRE 42616, on living leaves of Syzygium guineense (Willd.) D.C.,
Entabeni, Louis Trichardt, Soutpansberg District, Transvaal, Oct. 1963.
This is a new host record for D. javanica var. harana which is known to occur
on Citrus sp. in Japan and on Mangifera indica on the island of St. Thomas (Batista
& Ciferri, loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
8. Vitalia ekmanii ( Petrak & Ciferri ) Batista & Ciferri , in Inst. Micol. Univ.
Recife Publ. 158: 111 (1962).
Figs. 7, 10, 11.
Mycelium epiphyllous, superficial, pelliculose, blackish-brown; hyphae septate,
slightly constricted, branched, dull-brown, setose; mycelial setae arising from a multi-
cellular base, scattered, numerous, simple, erect, straight, acute, continuous, blackish-
brown, 80-120/* long, 9-10/* wide at the base and 2-5/* at the apex, covered by mycelial
hyphae. Ascostromata globose-depressed, scattered, membranous, setose, 144-175/*
in diameter; setae numerous, erect, continuous, simple, straight or slightly curved,
dark brown but becoming lighter brown towards the apex, 75-700/* long, 7-5/* wide
at the rounded base, tapering to the acute apex, 1-5-2 -5/* wide. Asci ellipsoid to
sub-clavate, very briefly stipitate, apical wall thickened, six- to eight-spored, apara-
physate, 45-58 x 10-15/*. Ascospores mono- or di-stichous, fusiform-elliptical,
hyaline, continuous when immature, becoming three-septate when mature, not con-
stricted, 12-15 X 5/*.
Specimen examined : PRE 42708, on living leaves of Clivia sp., Mariepskop Forestry
Station, Pilgrims Rest District, Transvaal, 1962.
This is a new host record for C. ekmanii which had previously been collected on
various unrelated host plants in South Africa (Batista & Ciferri, loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
9. Johansonia amadelpha (Syd.) v. Arx apud Miiller & v. Arx, in Beitr. Krypt.
FI. Schweiz 11 (2): 212 (1962).
Microcallis amadelpha Syd., Ann. Mycol. 24: 342 (1926).
Chaetothyrina amadelpha (Syd.) Syd., Ann. Mycol. 32: 10 (1934).
Fig. 9.
Mycelium hypophyllous, hyaline, not hyphopodiate, setose. Setae brown, septate,
straight or slightly curved, obtuse, concentrated round the ascostromata, 122-160 X
5-7 • 5/*. Ascostromata superficial, gregarious, globose-depressed, brown, membranous,
glabrous, 135-400/* in diameter; superior wall epithecioid. Asci bitunicate, clavate,
sessile or briefly stipitate, eight-spored, 36-42 x 12-15/*; paraphysoids dichotomously
branched, anastomosed, hyaline, filiform, 1/*. in diameter. Ascospores oblong-clavate,
hyaline, one-septate, not constricted at the septum, 7-12 x 3-5/*.
Specimen examined: PRE 42617, on leaves of Syzygium cor datum Hochst., Mariepskop,
Pilgrims Rest Distr., Transvaal, 1962.
221
The ascostromata occur in groups of two definite sizes: small, ca. 135-185/* in
diameter and large, ca. 335-400/* in diameter. Sydow (Ann. Mycol. 24: 342, 1926)
described Microcallis amadelpha (= Johansonia amadelpha (Syd.) v. Arx) and M.
consociata (— J. consociata (Syd.) v. Arx) as occurring together on Roupala veraguensis
and on Phoebe tonduzii in Central America. The only difference noted between these
two species were that M. amadelpha had larger ascostromata and longer setae than
M. consociata. Accordingly he doubted whether these two fungi were in fact distinct
species or only forms of the same species. In the authors’ collection the smaller
stromata contain no differentiated asci and are considered to be immature stromata
of J. amadelpha and not a separate species.
Syzygium cordatum is a new host record for J. amadelpha which is known to occur
on Phoebe (Lauraceae) and Roupala (Proteaceae) in Central America (Muller & v. Arx,
loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
10. Asterina bukobensis Hansford, in Proc. Linn. Soc. London 157: 201 (1945).
Fig. 12.
Colonies amphigenous, black. Mycelium reticulate, composed of radiating,
septate, branching, brown hyphae, 3—6/* in diameter. Hyphopodia at first terminal
on the hyphae, later becoming lateral by growth of the cell immediately behind them,
solitary, darker brown than the hyphae, composed of two cells of very variable shape,
from sublobate to irregular, 1 6 • 0—22 - 5/< long, basal cell 6-5-11/* long and 10-13/*
broad, terminal cell 8-11/* long and 6-5-10/* broad. Ascostromata numerous, scattered,
rounded, covered by dark mycelial hyphae, up to 550/* in diameter; basal membrane
subhyaline, indistinct; covering membrane convex, subopaque, dark brown, composed
of radiating hyphae; margin fimbriate, fringing hyphae olive-brown, tortuous; dehiscing
by means of radiating fissures. Asci basal, cylindraceous-clavate, nodose-stipitate,
rounded at the apex, eight-spored, 70-5-105 x 25-30/*; paraphysoids very numerous,
hyaline, clavate, septate, up to 2-5/* in diameter. Ascospores obliquely monostichous,
ovate to ellipsoid, dark brown, smooth, one-septate, constricted, 32-38 x 15-19/*,
basal cell narrower than the upper cell and attenuated towards the base.
Specimen examined : PRE 24678, on leaves of Eugenia natalitia Sond., Mariepskop,
Pilgrims Rest Distr., Transvaal, Jan. 1963. Also present: Asteridiella atra (Doidge)
Hansf. (PRE 42679), parasitised by Helminthosporium dorycarpum Mont. (PRE 42680).
The ascostromata are mostly immature and only a few asci and ascospores were
seen.
A. bukobensis is characterised by the typical solitary, two-celled hyphopodia which
are at first terminal on the hyphae and later become lateral (Fig. 12). Eugenia natalitia
is a new host record for this species of Asterina which was dscribed on E. bukobensis
from Uganda by Hansford (loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
11. Dimerina mindanaense (P. Henn.) Hansford, in Mycol. Papers, Commonwealth
Mycol. Inst. 15: 56 (1946).
Fig. 13.
Mycelium densely reticulate between the hyphae of the Meliola- host, branched,
very indistinctly septate, subhyaline, not hypopodiate, 3-3 • 5/* wide. Ascocarps super-
ficial, ostiolate, dark-brown, smooth, 75-112/* in diameter; wall composed of dark-
brown, polygonal cells. Asci very numerous, cylindrical to clavate, nodose-stipitate,
222
wall somewhat thickened at the apex, eight-spored, 32-48 x 6—8/*. Paraphysoids
filiform, very indistinct, deliquescing before the asci mature. Ascospores obliquely
monostichous or distichous, oblong-fusoid or subclavate, straight or slightly curved,
rounded at both ends, smooth, hyaline, one-septate but becoming three-septate when
mature, 9-5-11 X 3-5/*.
Specimen examined : PRE 42632, parasitic on Meliola carissae Doidge on Carissa
bispinosa (L.) Des. f. ex Bren. var. acuminata (L.) L. E. Codd, Tzaneen, Letaba District,
Transvaal, Oct. 1963.
Under a dissecting microscope the mature spores may be seen oozing out of the
ostioles and forming white droplets on the ascocarps. These spores are three-septate
(Fig. 13).
Hansford (loc. cit.) tentatively placed Dimerina eutricha (Sacc. & Berl.) Theiss.
(= Dimerosporium eutrichim Sacc. & Berl.) in synonymy with D. mindanaense. He
expressed the opinion that Dimerina eutricha would be the correct name for this species
if the suggested synonymy could be established by comparison of specimens. In their
original description of Dimerosporium eutrichum, Saccardo and Berlese (Rev. Mycol.
7: 156. 1885) stated that three-septate ascospores had been seen. The author’s
collection fits very well into this description. The fact that three-septate ascospores
were slso observed, further substantiates Hansford’s suggestion that Dimerosporium
eutrichum and Dimerina mindanaense are identical.
This is the first record of the occurrence of this fungus in South Africa.
12. Diplodia longipedicellata T. S. & K. Ramakrishnan, in Proc. Indian Acad.
Sci., B, 32: 78 (1950).
Figs. 14, 15.
Pycnidia amphigenous, not on definite leaf spots but causing purplish-brown
protuberances of the leaf surrounded by a yellowish discolouration, gregarious, black,
completely immersed in the leaf, 1 12—385/* in diameter, 85-225/* high; ostiole distinct,
protruding, rounded; contents exuding in long black tendrils which spread over the
leaf surface to form black conidial mats; cavity globose to ellipsoid or irregular in
form, sometimes occupying the entire thickness of the leaf, convoluted, uniloculate
but occasionally two to four pycnidia coalesce; wall poorly developed so that the
delimitation of the pycnidium is indistinct, pseudoparenchymatous, hyaline to brownish,
composed of a few layers of thin-walled cells, becoming thicker and more stromatic
round the ostiole, 6-22 /< in diameter. Conidiophores arising from the cells of the
pycnidial wall, erect, cylindrical, hyaline. Pycnidiospores acrogenous, single, elliptic
to broadly obpyriform, smooth-walled, dark olive-brown, 12-5-16-0/* long, one-septate
below the middle; basal cell much shorter and narrower than the upper cell, base
truncate or sometimes rounded, often with the conidiophore attached, 4 -5-6-5/* long
and 4-5-6-5/t wide at the septum; upper cell broadly elliptical with a broadly rounded
apex, brown but with a lighter coloured area directly above the septum, 8 -0-9 -5/*
long and 6-5-12-5/* in the widest part.
Specimen examined : PRE 42714, on living leaves of Syzygium guineense (Willd) D.C.,
Entabeni, Louis Trichardt, Soutpansberg District, Transvaal, Apr. 1964.
As noted by Ramakrishnan and Ramakrishnan (loc. cit.), the majority of the
spores have cells of unequal size. These are not typical Diplodia spores. This fungus
is very similar to Mycohypal/age congesta (Berk. ; Br.) Sutton, another fungus occurring
on Syzygium, in a number of respects: the structure of the pycnidium;
the exudation of the spores in long black tendrils which spread over the leaf surface
223
and the two-celled, brown spores. The only character that excludes D. longipedicellatci
from the genus Mycohypallage Sutton, is the absence of apical setae on the pycnidiospores
of the former. As the suthor is unable to suggest a more suitable genus, he follows
Ramakrishnan and Ramakrishnan (loc. cit.) in placing this fungus in Diplodia.
Syzygium guineense is a new host record for D. longipedicellatci which was described
on S. montanum from India by Ramakrishnan and Ramakrishnan (loc. cit.).
This is the first record of the occurrence of this fungus in South Africa.
Acknowledgements
The authors wish to express their sincere thanks and appreciation to: Prof. Dr.
A. Chaves Batista of the Instituto de Micologia, Universidade do Recife, Brazil for
the identification of certain specimens, kind advice and criticism of certain descriptions.
The staff of the Commonwealth Mycological Institute, Kew, England for identifying
certain collections. Mr. K. T. van Warmelo and Mr. G. C. A. van der Westhuizen
of the Plant Protection Research Institute, Pretoria, for invaluable advice and assistance
with the microtechnique and photomicrography. The staff of the Botanical Research
Institute, Pretoria for the identifications of the host plants.
EXPLANATION OF FIGURES
Fig. 1-3. — Fig. 1. Parapeltella maitlandii (a) asci; (b) ascospores. Fig. 2. Scolecopeltidium racemosae,
(a) asci; ( b ) ascospores. Fig. 3. Micropellidium obscurum, asci with ascospores and a para-
physoid.
Fig. 4-9. — Fig. 4. Ainsworthia xanthoxylii, ( a ) asci; (b) ascospores. Fig. 5. Ainsworthia roraimensis
ascospores. Fig. 6. Deslandesia javanica var. harana, (a) asci; ( b ) ascospores. Fig. 7. Vitalia
ekmanii, (a) mycelial seta; (b) ascostromatal seta; (c) asci: (d) ascospores. Fig. 8 . Chaetothyrina
musarum, (a) asci; (b) ascospores; (c) ascostromatal seta; (d) base of mycelial seta. Fig. 9.
Johansonia amade/pha, (a) asci; ( b ) ascospores; (c) part of mycelial seta.
Fig. 10-15. — Fig. 10-11. Vitalia ekmanii. Fig. 10. Mycelial setae showing multicellular base, x 400.
Fig. 11. Ascostromatal setae, x 400. Fig. 12. Asterina bitkobensis, hyphopodia, X 500. Fig. 13.
Dimerina mindanaense, ascospores, X 1250. Fig. 14-15. Diplodia longipedicellata. Fig. 14.
Cross section of pycnidium, x 200. Fig. 15. Pycnidiospore, x 1250
224
(5>cp0Cb^ i
C^Ooooq
^OQcoa^
3jOOCo03^
Xi
•D
O
c^Xl33ac> 1
cOQQOod
©
8 9 9 8
227
7230691
229
Bothalia 9, 1 :229-243
New and Interesting Records of South African
Fungi, Part V*
by
W. F. O. Marasasf, G. C. A. van der Westhuizenf,
K. T. van Warmelof, and M. C. Papendorft
Fifteen species are described below. Dried down cultures or specimens on natural
substrata of all these species have been deposited in the mycological collection of the
National Herbarium (PRE), at 590 Vermeulen Street, Pretoria.
1. Tritirachium roseum Beyma, in Antonie van Leeuwenhoek 8: 118-120 (1942);
MacLeod, D.M., Can. Journ. Bot. 32: 818-890 (1954).
Figs. 1, 2.
Colony on 3 per cent malt agar slow-growing, raised, cottony, “ Rhodonite pink ”
in colour, reverse “Clay colour”; mycelium subhyaline, reddish orange in mass
under the microscope, septate, 1-5-2 -2// in diameter, sparingly branched, usually
near septa; conidiophores either arising singly from the mycelium or compound, erect
or recumbent, branched, either single or in whorls of 2-4; fertile branches short,
forming a 60°-90° angle with the hypha or conidiophore, sterile portion with septum
at base, cylindrical to subulate, hyaline, 8 -4-19 • 5x1 -4-2- 1/t, across widest part, tapering
to rachis-like zigzag, hyaline fertile region, 5-6-14-0 x 0-9-1 -2 //; conidia hyaline,
subglobose to ovoid, 1-9-2 -8 x 1 -4-2-1//.
Specimen examined : PRE 43023, dried culture on 3 per cent malt agar, isolated from
maize silage, Mafeking. Feb. 1965.
This specimen is very similar to T. album described by Limber (Mycologia 32:
23-30. 1940), except for the colour. It agrees very well with the brief description
of T. roseum given by MacLeod (loc. cit.).
This fungus was isolated from decomposing maize silage together with other
saprophytic fungi, and is the first record of the occurrence of this genus and species
in South Africa. — K.T. v. W.
2. Ascobolus immersus Per soon, in Obs. Myc. 1: 35 (1796); Seaver, F. J. The
North American Cup Fungi (Operculates) 83 (1942).
Fig. 3.
Apothecia gregarious, partially immersed, subturbinate, up to 1-5 mm in diameter,
yellow becoming brown on drying, hymenium convex; asci few, very large, clavate,
projecting above the hymenium at maturity, eight-spored, operculate, interspersed
with paraphyses, 550-720 x 85-100// when extended; spores irregularly biseriate.
* Parts I-IV appeared in Bothalia 6: 183-204. 1951; ibid. 6: 489-500. 1956; ibid. 7: 109-116.
1958 and ibid. 9: respectively.
t Plant Protection Research Institute, Private Bag 134, Pretoria.
J Department of Botany, Potchefstroom University, Potchefstroom.
230
ellipsoid, hyaline darkening to deep violet, unicellular, surrounded by a gelatinous
sheath, surface sculptured by anastomosing bands, 55-64 x 25-34/*; paraphyses
slender, hyaline, sparsely septate, of varying length, 4- 5-5-5/* in diameter.
Specimen examined : PRE 42927, dried apothecia from horse dung.
The author has found this species several times on horse dung from several
localities, e.g. Pietermaritzburg, Natal and Pretoria, Transvaal, as well as on Hippo-
potamus ( Hippopotamus amphibius L.) dung from the Kruger National Park, Transvaal.
The occurrence of this species in South Africa has not been reported before. — K.T. v. W.
3. Podospora anserina ( Rabh .) NiessI, in Hedwigia 22: 156 (1883).
Sordaria anserina (Ces.) Winter in Abh. Nat. Ges. Halle 13: 99 (1873); Cain, R. F.
Univ. Toronto Studies No. 38: 46 (1934).
Figs. 4, 5.
Perithecia superficial or semi-immersed, dark brown appearing black under reflected
light, pyriform, lacking appendages, 360-540 x 240-330/*; neck conical, setose, setae
long, thin, straight, brown, aggregated, 80-110 x 3-4/<; asci hyaline, cylindrical-
clavate, on long slender stipes, thin-walled, four-spored, 210-240 x 25-32/* (pars
sporif. 144-154 x 25-32/*); spores obliquely uni-seriate at maturity, tadpole-shaped
when young, appendiculate, hyaline, becoming uniseptate ; upper cell swollen, darkening
to black at maturity, ellipsoid, 32-35 X 16-23/*; basal cell hyaline, cylindrical, straight
or curved, 25-29 x 8/*; gelatinous appendages long and coiled, attached eccentrically
at spore apex; paraphyses slender, septate, hyaline, 200-250 x 3-5/*.
Specimen examined: PRE 42925, dried perithecia from horse dung.
This species was found several times on dung of various herbivorous animals,
e.g. horse, sheep, hippopotamus, from various localities, e.g. Pietermaritzburg, Natal,
Pretoria, Transvaal and the Kruger National Park, Transvaal. Its occurrence in
South Africa has not been recorded before. — K.T. v. W.
4. Sporormia minima Auersw., in Hedwigia 7: 66 (1868).
Figs. 6, 7, 8, 9.
Ascocarps on horse dung numerous, single, very small, totally immersed with
only the short papillate ostiole exposed, brown to black, smooth, 100-112 /* in diameter;
asci eight-spored, cylindrical-saccate, widest below the middle, briefly stipitate, bituni-
cate, 67-83 x 16—18/* ; pseudoparaphyses few, deliquescing at maturity, spores obliquely
arranged in the ascus, cylindrical, three-septate, deeply constricted at the septa, hyaline
at first but darkening rapidly to olivaceous, 32-35 x 4 -5-6 -5/*, surrounded by a
gelatinous sheath about 4/* thick. After liberation the spores separate along the
middle septa.
Specimen examined: PRE 42926, dried ascocarps from horse dung, Onderstepoort
Vet. Res. Inst., July 1964.
This is the first collection of this species in South Africa. — K.T. v. W.
5. Tripterospora longicaudata Cain, in Can. Journ. Bot. 34: 699-710 (1956).
Figs. 10, 11, 12.
Ascocarps on dung scattered, brown appearing black by reflected light, globose,
superficial, inostiolate, without appendages, 160-190/* in diameter; surface mycelium
sparse; peridium brown, pseudoparenchymatous, about 8/* thick, composed of
231
irregularly shaped interlocking cells; asci eight-spored, clavate, with thin basal stipe,
thin-walled, deliquescing at maturity, 65-80 x 14-18//, spore mass held together by
transparent membrane not composed of cells ; ascospores irregularly biseriate, tadpole-
shaped when young, becoming uniseptate at base of swollen portion, upper cell at
first hyaline, darkening through brown to black at maturity, containing a single oil
globule, ellipsoid, 12-14 x 8-9//; basal cell remaining hyaline, cylindrical, straight
or slightly curved, 11-13 x 3//, disintegrating at maturity resulting in unicellular
ellipsoid spores with truncate bases.
Specimen examined : PRE 42928, dried ascocarps from horse dung, Onderstepoort
Vet. Res. Inst., July, 1964.
This is the first record of the occurrence of this species in South Africa. — K.T. v. W.
6. Tripterospora erostrata (Griff.) Cain, in Can. Journ. Bot. 34: 699-710 (1956).
Pleurage erostrata Griff-, in Mem. Torrey Bot. Club 11: 71 (1901).
Figs. 13, 14.
Ascocarps on dung scattered, dark brown appearing black by reflected light,
globose, superficial, inostiolate, 200-240// in diameter, covered with very long, flexuous,
septate, brown hairs up to 2 mm in length and 3-5// in diameter at the base; surface
mycelium sparse; peridium brown, about 10// thick, composed of somewhat angular
pseudoparenchymatous cells; asci eight-spored, clavate, with thin basal stipe, thin-
walled, deliquescing at maturity, 48-65 X 13-18//, spore mass held together by trans-
parent membrane not composed of cells; ascospores usually biseriate, tadpole-shaped
when young, becoming uniseptate at base of swollen portion, upper cell at first hyaline,
darkening through brown to black at maturity, ellipsoid with apical germ pore, 9-12
X 6-7-5//; basal cell remaining hyaline, cylindrical, straight, 7-8 x 3//, disintegrating
at maturity resulting in unicellular ellipsoid spores with truncate bases.
Specimen examined : PRE 42942, dried ascocarps from horse dung, Onderstepoort
Vet. Res. Inst., August, 1964.
This is the first record of the occurrence of this species in South Africa.
The two species described here agree very closely with Cain’s (loc. cit.) descriptions
of the type specimens. These fungi were found on horse dung kept in a moist chamber
and are, apparently, the first collections of these two species of this rare genus outside
Canada. — K.T. v. W.
7. Auxarthron umbrinum ( Boudier ) Orr & Plunkett, in Can. Journ. Bot. 33: 1439-
1456 (1963).
Figs. 15, 16.
Colonies on potato-carrot agar slow-growing, cottony, white at first, turning
buff later and becoming orange-brown in the centre as ascocarps mature, reverse of
colony orange-red with pigment insoluble in water and not diffusing into the medium.
Hyphae buff, 0- 5-2-0// in diameter, septate, branched. Ascocarps superficial, spherical,
reddish-orange to orange-brown, diameter 80-0-300-0//; peridial hyphae yellow-orange
to orange-brown, asperulate, sometimes more or less smooth, thick-walled, 1-5-2 -5//
in diameter, septate and usually swollen at the septa forming knuckle-joints measuring
3 -0-4 -5// in diameter, branching more or less dichotomously and anastomosed and
interwoven to form a reticulate peridium; peripheral peridial elements forming inverted
Y-shaped arches from which short and elongate appendages arise, free apices of arches
often truncate suggesting that appendages were either not formed or broken off; short
232
appendages spine-like with apices acute or rounded, aseptate or with 1 or 2 septa,
6- 0-60-0// long; elongate appendages not abundant, smooth, yellow-orange to orange-
brown paling towards a delicate apex, simple, straight, bent or hooked apically, non-
septate or often with 1, 2 or rarely 3 knuckle-joints above the junction of appendage
and peridium, more or less 3 - 0// in diameter and up to 650// long; asci hyaline,
subglobose to ovoid, thin-walled, evanescent, 8-spored, 5 -0-6-0 x 7 -0-8 -5//; asco-
spores pale yellow-green, echinulate at all stages, globose to subglobose with diameter
2 -5-3 -5// or ovoid and measuring 2 -0-2 -5 x 2 -5-3 -5//. No asexual stage was
observed.
Specimen examined : Potchefstroom Pure Culture Collection, No. 99; PRE 43068,
dried culture on potato-dextrose agar.
Isolated from leaf-litter of Acacia karroo , Potchefstroom, Transvaal, Jan./Febr. 1964.
This isolate agrees very closely with Auxarthron umbrinum as described by Orr
and Plunkett (1963). It differs from it mainly in the maximum diameter of the ascocarps,
the maximum length of the long appendages and especially in the diameter of the
knuckle-joints. The maximum measurements given for these structures are 600//,
1080// and 12-0// respectively as compared with 300//, 650// and 4-5// for the Potchef-
stroom isolate.
The poorer medium on which this isolate was grown could account for the
difference in measurements but does not seem to explain the difference in ratio between
the maximum diameter of appendage and knuckle-joint. Orr and Plunkett’s ratio is
1 : 4 while mine is only 1 : 1-8. According to this the knuckle-joints of the Potchefstroom
isolate must be far less prominent than those of some other specimens.
A culture of this specimen was submitted to Dr. C. F. Orr for observation and
diagnosis and he kindly confirmed the indentification.
This is the first record of the occurrence of this genus in South Africa. — M.C.P
8. Beltrania rhombica O. Penzig, in Nuovo G. bot. ital., 14: 72 (1884). Michelia.
2: 474; F. ital., tab. 1204; Pirozynski, Mycol. Papers, Commonwealth Mycol. Inst.
90: 7 (1963).
Beltrania indica C. V. Subramanian, Proc. Indian Acad. Sci., B. 36: 45 (1952).
Beltrania multispora H. J. Swart, Leeuwenhoek ned. Tijdschr. 24: 221 (1958).
Fig. 17.
Colonies growing rapidly on potato-carrot agar, white, spreading, with little
aerial mycelium, becoming grey-brown as fertile conidiophores appear, often zonate
when young. Mycelium scanty, partly superficial and partly immersed in the medium;
hyphae hyaline or light greenish-yellow to faintly straw-coloured, becoming brownish
later, thin- walled, septate, branched, up to 6-0// in diameter; setae rarely produced,
growing vertically from prostrate hyphae or from the upper part of vertical hyphae
bearing conidiophores lower down, simple, erect, straight, thick-walled, darker than
conidiophores, septate, smooth, 140 0-170-0 X 3 -0-3 • 5//, tapering to an acute apex;
conidiophores vertical, arising from prostrate hyphae, often from lobate cells, sometimes
from the lower cells of hyphae that become setiform towards the apex, conidiophores
flexuous, straw-coloured to yellow-brown, thin-walled, septate, mostly simple, occasi-
onally branched, often proliferating and geniculate, apex hyaline to subhyaline, often
inflated, denticulate, bearing varying numbers of conidia and/or one to several separating
cells, 25-0-130-0 X 2 -5-4-0//; separating cells numerous, arising as blown-out ends
of conidiophores, subhyaline to faintly coloured, thin-walled, ovoid to obovoid, 1 -denti-
culate at the base, 1-2-denticulate above, secondary separating cells may develop,
7- 0-15-0 X 4 -0-7-0//; conidia aoundant, borne directly on conidiophores or on
233
MARY GUNN LIBRARY
NATIONAL BOTANICAL INSTITUTE
PRIVATE BAG X 101
PRETORIA 0001
REPUBLIC OF SOUTH AFRICA
separating cells, 1 -celled, rhombic or equally to unequally biconic, smooth, olive-brown
spore wall two-layered with the outer layer of uniform thickness and the inner layer
thinner immediately above the widest part of the conidium where consequently a pale
transverse band resembling a broad septum is formed, 22-5-28-0 x 7 • 5-10-0/u., basal
end 1-denticulate or rounded, apex furnished with a rigid spike-like, hyaline appendage,
3-0-6-5u long.
Specimen examined : Potchefstroom Pure Culture Collection, No. 67; PRE 43066,
dried culture on potato-dextrose agar.
Isolated from leaf-litter of Acacia karroo, Potchefstroom, Transvaal, Jan./Febr. 1964.
The identity of this specimen was confirmed by the Director of the Centraalbureau
voor Schimmelcultures in the Netherlands.
This is the first record of the occurrence of this species in South Africa.— M.C.P
9. Myrothecium verrucaria {Alb. & Schwein .) Ditmar ex Fries, in Syst. Myc. 3:
216-218 (1829); Preston, Trans. Brit. Mycol. Soc. 26: 167 (1943).
Figs. 18, 19, 20, 25, 27.
Colonies develop rapidly on potato-carrot agar, appressed with slightly raised
aerial mycelium, cottony to floccose, white. Hyphae smooth, occasionally verrucose,
hyaline, septate, sparingly branched, diameter 1 -5-4-0//. Sporodochia minute and
confluent into larger masses up to 2 mm in diameter, discoid or irregular in outline,
sessile and raised above the surface of the medium, composed of loosely arranged
hyphae arising from prostrate mycelium or from parenchymatous aggregations of
hyphae, terminating in a dense disc-like layer surrounded by small numbers of long,
sterile, simple or sparingly branched, verrucose or tuberculate hyphae which occasionally
protrude from and extend above the surface of the disc; conidiophores composed
of fertile hyphae and phialides, fertile hyphae erect, branched, hyaline or faintly
greenish-yellow, 3-4-celled with cells up to 40// long and about 2-5 // wide, branches
up to 38// long, 1-2-celled, arising singly or in pairs or whorls directly from the apical
cell of the main axis or immediately below a septum of an intermediate cell, each branch
bearing a terminal whorl of phialides, fertile hyphae and phialides closely intertwined
and forming a compact palisade-like layer; phialides narrowly clavate or sub-cylindrical,
straight or bent near the base, hyaline to faintly yellow-green, arranged in whorls
at the apices of the fertile hyphae and their branches, occasionally arising singly or
in small whorls below a distal septum of an intermediate cell, 10-0-18-0 X 1 -0-2-0//;
conidia elliptical with the base truncate and the apex usually slightly pointed or rounded,
showing a germ pore at each end, smooth, one-celled, subhyaline to pale olive-green,
7 -0-8-0 x 8 -0-3-0//, usually bearing an extremely delicate membranous, funnel-
shaped or bicornute appendage at the pointed end, 2-5-4-0// long and 3-0-4-0/t wide
distally; spore-mass globular or conical, diameter 10-0// and more and up to 2-0
mm, viscid, green at first, becoming black later.
Specimen examined : Potchefstroom Pure Culture Collection, No. 65: PRE 43064,
dried culture on potato-dextrose agar.
Isolated from leaf-litter of Acacia karroo, Potchefstroom, Transvaal, Jan. /Feb. 1964.
This isolate agrees very closely with Myrothecium verrucaria as described by Preston
(loc. cit.) but differs from it mainly in the width of the conidia which was, in this case,
never found to be as much as 4-5//. An interesting feature of the conidia which, as
far as I know, has never been reported previously, is the extremely delicate, mem-
branous appendage borne on the pointed end of most of the spores. It is generally
funnel-shaped or triangular in outline, often apparently with a wide, more or less
234
Y-shaped notch making it appear bicornute. It is 2-5-4-0/t long and 3 -0-4 • 0/* wide
at its broad, distal end. It was not possible to establish with absolute certainty whether
the appendage is a flat, two-dimensional or a hollow, three-dimensional structure.
It appears to be the latter because it is often seen with a recurved edge.
Spores with similar appendages characterize the genera Starkeyomyces and Loma-
chashaka. Species of both genera differ from the fungus described here in having
hyaline or bright-coloured sporodochia and irregularly branched conidiophores in
Starkeyomyces and simple conidiophores in Lomachashaka. The difference between
the fungus described here and these two genera appears to be greater than between
this fungus and Myrothecium. Because the appendages on the spores apparently do
not persist in old cultures and because the type specimen of Myrothecium verrucaria
was not available for examination and comparision, this fungus is assigned to M.
verrucaria with which it agrees very closely in all other respects.
The described appendages are not visible with ordinary illumination but become
quite distinct when phase contrast equipment is used (Figs. 18, 19).
A culture of this isolate was submitted to the Director of the Centraalbureau voor
Schimmelcultures in the Netherlands who confirmed the identification.
This is the first record of the occurrence of this species in South Africa. — M.C.P.
10. Robillarda sessilis Saccardo, in Michelia 2: 8; Syll. Fung. 3: 408 (1884).
Figs. 21, 22, 24, 28.
Colonies on potato-carrot agar restricted, appressed, aerial mycelium sparse,
cottony, white. Hyphae hyaline, septate, branched, up to 3 • 5 /< in diameter. Pycnidia
abundant, scattered, partly immersed in the medium, at first minute, globose, faintly
yellow-brown and becoming darker, when mature black in incident and reddish-brown
in transmitted light, globose to sub-globose or lacrymoid, ostiolate, often beaked,
up to 450,w in diameter; wall composed of radially arranged rows of yellow brown,
thick-walled cells, 5-0~8-0/< in diameter and lined internally with large, hyaline, thin-
walled more or less isodiametric or flaskshaped sporogenous cells 5- 0-8-0^ or
7-0-10-0 X 3 0-6-0^ respectively with highly vacuolated contents, often forming
short 2-3-celled filaments projecting into the pycnidial cavity; spores cylindrical to
cylindric-elliptical with ends more or less tapering or rounded, 10-0-16-0 x 2-5-3-5/i,
faintly olivaceous, often slightly constricted at the single median septum, apical cell
rounded or truncate, bearing a short, clavate, hyaline appendage branching into 2-4
simple or branched, filiform, straight or bent, hyaline, widely divergent setulae
10-0-25-0 X 0-5//, basal cell rounded or more or less truncate where it was attached
to the bearer cell.
Specimen examined : Potchefstroom Pure Culture Collection, No. 74; PRE 43065,
dried culture on potato-dextrose agar.
Isolate from leaf-litter of Acacia karroo , Potchefstroom, Transvaal, Jan./Febr. 1964.
In his paper on Robillarda phragmitis, Cunnell (Trans. Brit. Mycol. Soc. 41 : 405.
1958) refers to the diverse opinions which exist as to whether the appendages of the
spores of the various Robillarda species are situated basally or apically and stresses
the necessity for detailed studies on the development of the spores in order to reach
agreement on this point. From his paper it appears that the appendages of R. phrag-
mitis are borne basally.
The development of the spores of R. sessilis differs completely from that of R.
phragmitis as recorded by Cunnell (loc. cit.).
235
The young spore appears as a rounded or oblong projection from a bearer-cell.
It soon assumes a clavate or oval shape and becomes separated from the bearer-cell
by a cross-wall (Fig. 28: 1, 2). The young spore now enlarges until it reaches a size
of approximately 10-0 x 2-0/* (Fig. 28: 2). At this stage the spore grows out apically
and forms an oval-shaped protuberance or extension which is attached to the main
body by means of a slender neck (Fig. 28: 3). The appendages are then initiated
when 2 — 4 papillate outgrowths appear apically and/or laterally on the apical pro-
tuberance (Fig. 24, 28: 4). These initials elongate and develop into the characteristic
filiform setulae of the mature spore (Fig. 22, 28 : 5). Because the developing spore
is highly vacuolate at all stages it is extremely difficult to observe the exact stage at
which the median and apical septum at the junction between the neck and the main
body are formed. It probably happens before the spore becomes detached because
septate spores are often found still attached to their bearer-cells.
The identity of this isolate was confirmed by the Director of the Centraalbureau
voor Schimmelcultures in the Netherlands. This is the first record of the occurrence
of this species in South Africa. — M.C.P.
11. Hyalotia viridis ( Torrend ) Guba, in Monograph of Monochaetia and Pestalotia
310 (1961).
Figs. 23, 26.
Colonies develop rapidly on potato-carrot agar, white, floccose to granulose.
Hyphae hyaline, profusely branched, septate, 1 • 0—6 - 0/« in diameter. Fruiting pustule
suggestive of a pycnidium, abundant, scattered, lightly covered with a loose reticulum
of dark hyphae often aggregating to form ribbon-like strands, shape variable, subglobose
to oblong or subcylindrical with definite or indefinite or irregular outline, single or
coalescing, up to 400^ in diameter, dark-brown to red-brown in transmitted and black
in incident light, distinctly ostiolate when young or completely closed and opening
irregularly, carbonaceous shell when crushed straw-coloured to light-brown; spores
cylindrical to cylindric-fusiform, erect, sometimes slightly curved, hyaline to light
greenish-yellow, all excepting apical cells with finely granular, iridescent contents,
3-4-septate, 22-0-28-0 x 2 -5-3 -5 p, apical cell empty, narrow-conical, acute, rather
long and crested with 2 or usually 3 simple, widely divergent, hyaline, setulae, 0-5-1 -Cffi
wide at the base, 15-0-20-5/z long, basal cell obconical with a rounded or truncate
base supported by a slender, filiform, straight or curved pedicel which is either basal
or attached eccentrically or sub-laterally, 8 - 0—12 * 5/* long.
Specimen examined : Potchefstroom Pure Culture Collection, No. 73; PRE 43067,
dried culture on potato-dextrose agar.
Isolated from leaf-litter of Acacia karroo, Potchefstroom, Transvaal, Jan./Febr. 1964.
This specimen was kindly examined by Dr. Emil F. Guba and Dr. J. A. von Arx
who supplied valuable comments and also verified the identification.
Guba (1961) described the fruiting body of Hyalotia as an acervulus but stated
that he found the pustule of H. Iateripes to be a globose pycnidium with ostiole. From
his description of H. viridis it appears that the pustule of this species is considered to
be an acervulus. According to my observations the structure is a pycnidium which is
distinctly ostiolate, especially when still young, but sometimes it is completely closed
and opening irregularly (Fig. 23).
In a private communication Dr. Guba also pointed out that he found the conidia
of H. viridis to be 1-4-septate.
This is the first record of the occurrence of this genus in South Africa. — M.C.P.
236
12. Helminthosporium pedicellatum Henry, in Univ. Minn. Agr. Exp. Sta. Bull.
22: 42 (1924).
Bipolaris pedicel/ata (Henry) Shoemaker in Can. Journ. Bot. 37: 884 (1959).
Figs. 29, 30.
Colonies on potato dextrose agar “ Olivaceous Black ”, darkening as sporulation
increases, fast-growing, covering the entire plate in five days at 25 °C. Conidiophores
sparingly branched, 5-6/< in diameter, brownish olive at the apices, lighter towards
the base, relatively straight up to the position of the first spore then geniculate and
bearing one to four conidia acropleurogenously, leaving prominent scars upon detach-
ment; conidia broadly fusiform, straight, brownish-olive, widest near the middle and
decidedly attenuated towards both ends, apical cell apiculate or broadly rounded in
some spores, basal cell narrowed to a short, cylindrical pedicel, epispore strongly
thickened with the hilum included within the wall of the basal cell, germinating from
both ends but usually from the base, 3-7 septate (mostly 6-septatej, 58-68 x 23-28/t.
Specimen examined : PRE 42940, dried culture on potato dextrose agar, isolated from
corn (Zea mays) roots by Mr. J. J. du Toit, Plant Protection Research Institute, Pretoria,
Dec. 1963.
H. pedicellatum was first isolated by Henry (loc. cit.) from diseased wheat roots
in Minnesota, U.S.A. He found this species to be only weakly parasitic. Tveit (Rev.
Appl. Mycol. 33: 345. 1954) isolated this fungus from oat seed from Brazil. Hassan
(Plant Dis. Reptr. 40: 890-897. 1956) isolated H. pedicellatum from the basal parts
of oat plants. In pathogenicity tests he found this fungus to be non-pathogenic to
certain varieties of oats, barley and wheat. Shepherd, Hall & Pendry (Phytopathology
52: 752. 1962) reported H. pedicellatum as the causal organism of a severe root necrosis
of corn in California. They also isolated this fungus from diseased sorghum roots.
H. pedicellatum has been found to be in constant association with root rot of corn
in the Transvaal province (du Toit, private communication).
This is the first record of the occurrence of this fungus in South Africa.— W.F.O.M.
13. Curvularia trifolii ( Kaufm .) Boed. f.sp. gladioli Parmelee & Luttrell apud
Parmelee, in Mycologia 48: 558-567 (1956).
Fig. 31.
Culture on potato dextrose agar dark olive grey, woolly, zonate with margin
pale or hyaline and darkening to olivaceous — Black in the centre. Hyphae prostrate
and submerged, branching, septate, sub-hyaline at first, darkening to pale-brown or
fuscous 1-0-7-5/t in diameter, and forming intercalary or terminal sclerotia consisting
of inflated, globose or subglobose cells, hyaline and thin-walled at first, later with
thickened dark-brown walls, forming irregular masses of about 17-0-50 x 37-0-80-0/<
mostly embedded in the agar; conidiophores erect, sub-hyaline to fuscous, arising
from trailing hyphae and the sclerotia, septate, straight, unbranched and narrow,
2 -0-2 -5 n, in the lower portion up to the first conidium, or with one or two branches
below the first conidium then geniculate and widening to 5-0-7-0/t and more frequently
septate, bearing one to several conidia acropleurogenously, length very variable 30-250, « ;
conidia unequally ventricose — fusiform, rounded at the apex, tapering towards the
prominent, protruding, basal hilum, strongly curved, often sharply bent almost at
right angles, usually three-septate occasionally with one or two additional septa, the
terminal cells hyaline or sub-hyaline, the middle cells larger fuscous or dark brown
thicker-walled, the penultimate cell largest, swollen and usually unequally distended
imparting the curvature to the conidia, 20-5-42-5 X 8-5-15-O/t.
237
Specimen examined-. PRE 42934, dried culture on potato-dextrose agar, isolated from
gladiolus corm. Springs District, Tvl., Feb. 1964.
This specimen agrees very well with the descriptions and data given by Parmelee
(loc. cit.) who emphasized the protruding hilum which distinguishes this species from
C. lunata (Wakker) Boed. Parmelee (Plant Dis. Reptr. 28: 515-517. 1954) also
mentioned the “ undifferentiated masses of mycelium made up of globose pigmented
cells ” which he found on the infected host tissues and in culture. On moist filter
paper, these bodies, which he regarded as structures which ensure survival of the fungus,
produced clumps of conidiophores.
The South African collection was found intermingled with a Botrytis sp. on a
single corm. Although no lesions similar to those illustrated by Parmelee were present,
there is no doubt that this isolate is identical to the fungus described by him.
This is the first record of the occurrence of this fungus in South Africa. — G. C. A.
v.d. W.
14. Chalara terrestris Agnihothrudu & Barna, in Lloydia 25: 172-175 (1962).
Figs. 32, 33.
On potato dextrose agar growth is slow, restricted, colony up to 50 mm in dia-
meter after 4 weeks at 26°C.
Colony somewhat sodden at first, wrinkled around the inoculum, isabelline, slowly
developing fine, compact, dow'ny white mycelium in patches which later turn grey as
conidiophores develop; hyphae hyaline, very narrow, branching, septate 1 • 5—2 - 5/«
in diameter; conidiophores borne singly or in groups of 2-3 on the narrow hyaline
hyphae, dark yellowish brown, erect or reclining, long subulate or narrowly obclavate
to almost cylindrical, widest in lower or middle part of the sporogenous cell tapering
gradually upwards towards the tube or narrowing abruptly above the middle, thick-
walled, with 1-20 septa in lower part, tube sub-hyaline to hyaline, slightly flared,
thin-walled and with a prominent septum above the sporogenous cell, forming conidia
endogenously (80)— 90— 1 20~(223) x 6 -0-8 •()/<, tube 2 -4-3 -6 /« in diameter; conidia
cylindrical, ends truncate or slightly rounded, hyaline, thin-walled, two-celled, 11-16 x
3-0-3-6/n
Specimen examined-. PRE 42936, isolated on potato dextrose agar from decaying
roots of young Eucalyptus saligna, Tzaneen, Transvaal, April 1964.
The fungus described here, agrees very well with the description by Agnihothrudu
and Barna (loc. cit.) of a fungus isolated from the roots of Camellia sinensis suffering
from Hittiali disease in upper Assam. Although the South African fungus was isolated
from a different host, there is little doubt that this fungus is identical with the species
found on the roots of tea bushes in Assam. The conditions under which these isolations
were made, however, suggested this species to be a soil saprophyte rather than a root
parasite (v.d. Westhuizen, S. A. For. Journ. 54: 12-16. 1965).
One other species Chalara kriegeriana Bres. had been reported from soil in Swit-
zerland (Gilman, J. C., A Manual of Soil Fungi, p. 331, Iowa, 1957) but this species
differs from C. terrestris by having smaller conidia.
Saccardo (Syll. Fung. IV pp. 333-336) listed C. fusidioides Corda, C. heterospora
Sacc. and C. montellica Sacc. on species of Eucalyptus. These three species had been
found on dead or dying branches or wood of the host and differ from C. terrestris
in the characters of the conidiophores and conidia.
7230691-8
238
This is the first record of the occurrence of this genus in South Africa. — G. C. A.
v.d. W.
15. Melanospora episphaeria Phillips & Plowright, in Grevillea 10: 71 (1881).
Sphaeroderma episphaeria (Phill. & Plowr.) Sacc. in Syll. Fung. 2: 560 (1893);
Petch, Trans. Brit. Myc. Soc. 21: 254 (1938); Martin, Mycologia 47: 606-608 (1955).
Figs. 34, 35.
On malt agar colony thin, fast growing, reaching a diameter of about 70 mm in
one week; mycelium thin cottony, sub-hyaline to “pale ochraceous buff” raised at
first, later collapsing and often with woolly overgrowth of mycelium in patches; hyphae
narrow, branched, hyaline, thin-walled, septate 2 -0-6-0//; perithecia appearing at
first as small mucilaginous droplets on the mycelium, darkening gradually to amber,
at maturity, scattered, superficial, gelatinous and without subiculum, subglobose,
glabrous, 120-500/t in diameter, the dark spores visible through the wall; ostiole
slightly papillate, surmounted by a collar of straight, erect, hyaline setae 40-150//
long; wall thick, consisting of translucent, pale amber, parenchymatous cells; asci
hyaline, clavate, thin-walled 8-spored, deliquescing early, 50-70 x 12-18//, paraphyses
lacking; ascospores dark brown, limoniform, somewhat flattened with epispore covered
by reticulate ridges and with minute light-coloured germ pores at the apicula, 16-20 x
10-15//.
Specimen examined : PRE 42937 on malt agar, isolated from decaying roots of Eucalyptus
sa/igna, Tzaneen, Transvaal, March 1964.
The fungus described here, agrees very well with Martin’s (loc. cit.) description
of Sphaeroderma episphaeria (Phill. & Plowr.) Sacc. which he isolated from surface
litter under beeches in Surrey, England. Although Martin’s isolate had smaller spores
than those mentioned by Petch (loc. cit.) in his description of the species, Martin
regarded his isolate as representative of the fungus described by Phillips and Plowright
(loc. cit.) and transferred to Sphaeroderma by Saccardo (loc. cit.). Martin further
regarded early gelatinization of the asci, maturation of the ascospores in the gelatinous
cavity of the ascocarp and their eventual discharge in a gelatinous tendril as alien to
the process of spore formation in typical members of the Hypocreales. He suggested
that Sphaeroderma should be grouped with other species having the same type of ascus
dehiscence.
According to Petch (loc. cit.) species of Sphaeroderma Fuckel differ from those
of Melanospora Corda by the absence of perithecial necks. Von Arx and MUller
(Beit. Krypt. FI. Schweiz II (I): 139. 1954) regard perithecial necks as variable and
their absence or presence subject to influence by environmental conditions. Further-
more, perithecia of the two genera are very similar in their internal structure. The
soft gelatinous perithecial wall and asci deliquescing early, leaving the ascospores to
mature inside the perithecia, are important characters of Melanospora Corda. For
these reasons, Von Arx and Muller (loc. cit.) regarded Sphaeroderma as synonymous
with Melanospora Corda.
Since the fungus described above agrees with these characters, it is considered
to be best placed in the genus Melanospora in which it was described originally.
This is the first record of the occurrence of this species in South Africa. — G. C. A.
v.d. W.
239
EXPLANATION OF FIGURES
Fig. 1. — Tritirachium ro.se it m, fertile branch, X 2200.
Fig. 2.—T. roseitm, branching, x 2200.
Fig. 3. — Ascobolus immersits, mature ascus, X 130.
Fig. 4. — Podospora anserina, mature ascospore, x 520.
Fig. 5. — P. anserina, immature ascospore, x 520.
Fig. 6. — Sporormia minima, mature ascospores, x 520.
Fig. 7. — S. minima, extended ascus, x 520.
Fig. 8. — S. minima, unextended ascus, x 520.
Fig. 9. — S. minima, ascospores showing mucus sheath, x 520.
Fig. 10. — Tripterospora iongicaudata, ascospores with appendage, X 520.
Fig. 11. — T. Iongicaudata, mature ascospores, x 520.
Fig. 12. — T. Iongicaudata, asci and ascospores, x 520.
Fig. 13. — T. erostrata, ascocarp, x 130.
Fig. 14. — T. erostrata, mature ascospores, x 520.
Fig. 15. — Auxarthron umbrinum, reticulate peridium showing inverted Y-shaped arches, short anne
dages and elongate appendage, X 250.
Fig. 16. — A. umbrinum, mature ascocarps, X 100.
Fig. 17. — Beltrania rhombica, conidiophores, separating cells and conidia, x 330.
Fig. 18 & 19. — Myrothecium verrucaria, conidia with membranous appendages, X 830.
Fig. 20. — Myrothecium verrucaria, conidiophores, X 715.
Fig. 21. — Robillarda sessi/is, flask-shaped spre-bearing cells with young conidia, X 715.
Fig. 22. — R. sessiiis, spre-bearing cells with young conidia developing appendages, x 715.
Fig. 23. — Hyaiotia viridis, pycnidia, X 110.
Fig. 24. — Robillarda sessiiis vr>ung conidia with developing appendages, X 715.
Fig. 9 5. — Myrothecium vei •/, via, phialides. x 715.
Fig. 6. — Hyaiotia viridis, mature conidic "15.
Fig. 27. — Myrothecium verrucaria, conidic t res, and conidia, x 2200.
Fig. 28. — Robillarda sessiiis, five stages in c idial development, x 1500.
Fig. 29. — Helminthospori.im pedicellatum, conidiophore and conidia, x 500.
Fig. 30. — H. pedicellatum conidia, x 500.
Fig. 31. — Curvidaria trifolii f.sp. gladioli, conidia, x 500.
Fig. 32. — Chalara terrestris, conidiophore with emerging conidium, X 1000.
Fig. 33. — C. terrestris, two-celled conidia, X 1000.
Fig. 34. — Melanospora episphaeria, three ascospores showing form, surface reticulation and germ
pore, x 1000.
Fig. 35. — M. episphaeria perithecium showing ostiole, hyaline setae and extruded ascospores, X 250.
240
©
241
7230691
Vi\
242
243
245
Bothalia 9, 1 :245-249.
Two Species of Erysiphaceae from Pretoria
by
W. F. O. Marasas and I. H. Schumann
Plant Protection Research Institute, Department of Agricultural Technical Services,
Pretoria
During the early winter of 1965 the junior author collected two interesting species
of Erysiphaceae on trees growing in the gardens of the Botanical Research Institute,
Pretoria. One of these, collected on Ehretia rigic/a (Thunb.) Druce, appeared to be
a new species of Uncinula and is described here as Uncinula praeterita Marasas &
Schumann sp. nov. The other, Phyllactinia acaciae Sydow, had been collected only
once before and was then described from immature material. A more complete
description of this species is presented.
Uncinula praeterita Marasas & Schumann, sp. nov.
Figs. 1, 3, 5.
Colonies amphigenous; mycelium superficial, white, persistent, inconspicuous,
thinly effused; hyphae hyaline, branched, sparingly septate, 3 -0-6 -5// wide; apressoria
lobate, single or opposite, 6 ■ 5— 1 3 • 0/t x 3-0-6-5//. Perithecia scattered, globose-
depressed, at first yellowish, becoming brown, 95— 1 40// in diameter; wall composed
or irregularly angular cells; appendages on the upper half of the perithecium, very
numerous, 25-80 or more, usually more than 40, hyaline, straight or curved with
uncinate to helicoid or occasionally obtuse tips, simple, continuous but often one- or
two-septate in the lower half, thin-walled throughout, smooth, 5 * 0—6 - 5// in diameter,
uniform in width or tapering slightly towards the apex, 65—290,/# long, usually about
equal in length to the diameter of the perithecium, occasionally up to twice as long.
Asci 4-7 per perithecium, broadly ovate, pedicellate, 3-spored, 50-60 x 30—35//.
Ascospores hyaline, continuous, ellipsoid, 20-30 x 13-16, /< (mostly 25 x 15//). Coni-
diophores one- or two-celled, straight, hyaline, 25-50 x 8-10//. Conidia hyaline
solitary, oblong, continuous, broadly rounded at the apex and usually truncate at the
base, 33-45 X 12-15//.
On leaves of Ehretia rigida (Thunb.) Druce (Boraginaceae), Gardens of the
Botanical Research Institute, Pretoria, June 1965, Schumann in PRE 43035 (PRE,
holotype).
Type collection deposited in the Mycological Herbarium, Plant Protection Research
Institute, Pretoria and in the Herbarium of the Commonwealth Mycological Institute,
Kew, England.
Plagulae amphigenae; mycelium superficiarium, albidum, sparse effusum, non
conspicuum, persistens; hyphae hyalinae, ramosae, paulo septatae, 3 -0-6 -5// diam.;
appressoria lobata, singularia vel opposita, 6-5-13-0 x 3 -0-6 -5//. Perithecia sparsa,
globoso-depressa, prima flavida, deinde brunnea, 95-140// diam.; cellulae parietis
exterioris irregulariter angulatae; appendices numerosae, in quoque perithecio 25-80,
hyalinae, rectae vel curvulae, simplices, aseptatae sed saepe uni- vel bi-septatae, tenue
tunicatae, leves, in latitudine aequales vel ad apicem sensim attenuatae, 5 -0-6 -5//
diam., perithecii diametrum subaequantes sed nonnunquam usque ad duplo longiores.
246
65-290 // longae. Asci in quoque perithecio 4-7, late ovati, pedicellati, 3-spori, 50-60 x
30-35//. Ascosporae hyalinae, ellipsoideae, 20-30 x 13-16// . Conidiophorae uni- vel
bi-cel lulatae, hyalinae, rectae, 25-50 x 8-10//. Conidia sola, oblonga, continua,
hyalina, antice rotundata postice ad basim truncata, 33-45 x 12—15//.
Hab. in foliis vivis Ehretiae rigidae (Thunb.) Druce, Pretoria, Schumann in PRE
43035 (PRE, holotypus).
U. praeterita appears to be related to U. incrasscita Salmon (Ann. Mycol. 6: 524,
1908) (= U. pterocarpi Doidge, Bothalia 4: 844. 1948) and U. combreticola Doidge
(Bothalia 4: 844. 1948), both of which occur in South Africa. U. praeterita, U. incras-
sata and U. combreticola all have about the same number of appendages (40-80), but
can easily be distinguished by other characteristics of the appendages viz.: —
U. incrassata : Appendages thick-walled, aseptate, length seldom exceeding the
diameter of the perithecium.
U. combreticokr. Appendages thin-walled, aseptate, widening towards the tip,
length usually equal to the diameter of the perithecium.
U. praeterita : Appendages thin-walled, aseptate or one- to two-septate, of
uniform width or tapering slightly towards the tip, length usually equal
to the diameter of the perithecium but occasionally up to two times greater.
Furthermore, both U. incrassata and U. combreticola have a well developed,
dense mycelial mat covering the greater part of the leaf while U. praeterita forms
inconspicuous colonies of thinly-effused mycelium. All the asci seen of U. praeterita
were three-spored, while the asci of U. incrassata (= U. pterocarpi, PRE 1805) contain
four spores. The ascospores of U. incrassata are also smaller than those of U. praeterita,
measuring 13-18 X 12-13// (Doidge, Bothalia 4: 845. 1948). No mature asci were
present in the type material of U. combreticola so that the number of ascospores per
ascus and the size of the ascospores could not be determined.
The only other species of Uncinula known to occur on the Boraginaceae is U.
ehretiae Keissler (Osterr. Botan. Zeitschr. 1924 p. 123), which was described on
Ehretia sp. from China. The perithecia of this species, however, have only 4-8
appendages of which the length is less than the diameter of the perithecium (Tai, Bull.
Torrey Botan. Club 73: 125. 1946).
Phyllactinia acaciae Sydow, in Ann Mycol. 33: 233 (1935); Doidge, Bothalia 4:
840 (1948.)
Fig. 2, 4, 6.
Colonies amphigenous, mycelium persistent, very well developed and covering
the entire leaf surface with a thick, white mat; hyphae hyaline, branched, septate,
1-4// in diameter. Perithecia scattered, fairly numerous, at first orange, becoming
brown, 120-275// in diameter, wall composed of irregularly angular cells. Appendages
7-18 per perithecium, equatorial, hyaline, rigid, continuous, 88-185// long, with a
bulbous base, 20-35// in diameter, above which the appendages are 6-5-10-0// in
diameter and the walls 3// thick, tapering to the apices. Asci 12-26 per perithecium,
ovate, pedicellate, two-spored, 50-68 x 25-30//. Ascospores hyaline, continuous,
ellipsoid, vacuolate, 23-30 X 13—16// (mostly 25 x 15//). Conidiophores three-celled,
straight or curved, hyaline, 50-62 x 3-5//. Conidia solitary, oblong-clavate or cylin-
drical, obtusely rounded at both ends with the lateral walls concave, 40-80// long
(mostly 62//), 10-15// broad at the ends and 8 - 0— 12-5// in the centre.
On leaves of Acacia karroo FLayne (Mimosaceae), Gardens of the Botanical Research
Institute, Pretoria, June 1965, PRE 43036.
247
Only the type material of P. acaciae had been collected previously (PRE 23428,
Klapperkop, near Pretoria, 1928). Sydow (loc. cit.) and Doidge (loc. cit.) observed
that very few, apparently not quite mature perithecia, were present in the type collection.
The present authors could not find any perithecia on the type material in the Mycological
Herbarium, Plant Protection Research Institute, Pretoria.
Abundant material of a Phyllactinia with a well developed conidial as well as
perithecial stage was collected on Acacia karroo in the gardens of the Botanical Research
institute, Pretoria during June, 1965. This fungus was considered identical with P.
acaciae because of the typical “ dumb-bell-shaped ” conidia and the size and shape
of the asci and ascospores.
Acacia karroo is a new host record for P. acaciae which was described on A. robust a
by Sydow (loc. cit.). A more complete and somewhat emended description of P. acaciae,
made from more mature material, is given above.
Acknowledgements
The authors wish to express their sincere thanks and appreciation to the stall
of the Botanical Research Institute, Pretoria for the identifications of the host plants.
Dr. D. J. B. Killick of the same Institute for checking the Latin diagnosis and Mr.
G. C. A. van der Westhuizen of the Plant Protection Research Institute, Pretoria for
advice and criticism during the preparation of the manuscript.
EXPLANATION OF FIGURES
Fig. 1-4. — Fig. F Uncinula praeterita, ascocarp, X 125.
Fig. 2. Phyllactinia acaciae, ascocarp, X 125.
Fig. 3. Uncinula praeterita, appendages, X 500.
Fig. 4. Phyllactinia acaciae, appendages, X 500.
Fig. 5-6. — Fig. 5. Uncinula praeterita, (a) appendages; (b) conidiophores; (c) conidia; (cl) asci;
(e) ascospores; (/) appressoria.
Fig. 6. Phyllactinia acaciae, (a) appendages; (b) conidiophores; (c) conidia; (d) asci;
( e ) ascospores.
mm
249
7230691
*
*
Plate 1. — Ethel Mary Doidge (photograph taken about 1930)
10862
Bothalia 9, 2: 251-253.
Ethel Mary Doidge (1887-1965)
by
M. D. Gunn
Ethel Mary Doidge, who died on 22nd September, 1965, at Anerley, South Coast,
Natal, occupied a leading place for almost forty years as a research worker in the field
of plant bacteriology, pathology, and systematic mycology.
She was born of Methodist stock at Nottingham, England, on 31st May, 1887.
Her father, Henry Doidge, an analytical chemist, who died early, and his wife, Elizabeth
Craven, came to Natal in 1897, with their young family and settled in Pietermaritzburg.
Ethel Doidge was educated at Epworth High School, Pietermaritzburg, and
Huguenot College, Wellington, Cape Province, where she received her early botanical
training under Dr. Bertha Stoneman. Graduating in 1907 with distinctions in
botany, she was awarded a scholarship at the University of the Cape of Good Hope,
which she did not take up.
In 1908 her long association with the Public Service commenced when she was
appointed as assistant to the Plant Pathologist and Mycologist in the Division of
Botany, Transvaal Department of Agriculture, Pretoria, where pioneer work in plant
pathology had begun three years earlier under Dr. I. B. Pole Evans. In 1909 she
obtained the M.A. degree in botany, with special reference to mycology, from the
University of the Cape of Good Hope. In the same year, the University awarded
her the Cornwall and York prize for a research paper entitled “ The flora of certain
kaffir beers ”. In 1912 she was appointed professional assistant in the Division of
Botany and Mycology (the Union of South Africa Public Service) and in 1914 was
awarded the degree of D.Sc. by the University of the Cape of Good Hope. The thesis
dealt with “A bacterial disease of mango, Bacillus mangiferae, n.sp.”, a previously
undescribed disease which was then unknown outside South Africa and had for some
years been causing considerable loss to mango growers in this Country.
Dr. Doidge was elected an F.L.S. in 1912 and supported the main South African
scientific associations, in which she played an active part. She was a foundation member
of the S.A. Biological Society, served on its council and was granted the Society’s
major award, the Senior Capt. Scott Memorial Medal in 1922 for her researches in
plant pathology in South Africa. She joined the Royal Society of South Africa, of
which she was elected a Fellow in 1915, and the S.A. Association for the Advancement
of Science, becoming President of Section C in 1918. Her address on this occasion
was entitled “ The role of bacteria in plant diseases ”. She was also appointed as a
member of the first council of the University of South Africa. Dr. Doidge is com-
memorated in the name Meliola doidgeae Syd.
Her accomplishments were not confined to science, however, for she was interested
in music and held two licentiate diplomas, the L.T.C.L. and U.P.L.M., for singing.
For many years she was a member of the Pretoria Musical Society.
In 1929 she was appointed as Principal Plant Pathologist, a post which she held
until her retirement in 1942 at the age of 55 years. Her services were retained for
a further four years, during which time she rounded off her official career with a
monumental work of 1094 pages: “ The South African fungi and lichens ”, published
as Bothalia Vol. 5 (1950).
252
Commenting in the preface to this publication, Dr. I. B. Pole Evans, her former
Chief, writes: “ Dr. Doidge took up this study more than 35 years ago. She has made
her impress on it and has left a record of it for which younger workers in years to come
will never cease to thank her. The field for research in Mycology in this country is
thus immense, but this Science rarely appeals to the average Administrator under
whom much of the work usually falls. Its progress, therefore, must naturally be slow
and far from encouraging. However, there are still some labourers in the field of
Science whose only thought and ambition is to advance the knowledge of their particular
subject for the general benefit of mankind, and in this Dr. Doidge has certainly
succeeded. She has tirelessly paved the way and truly laid the foundation on which
Mycology in Southern Africa will be built. Great things no doubt will come from
the further study of African fungi and more that is new from Africa will still arrive,
but come what may, future workers in this field will rightly appraise the value of Dr.
Doidge’s work ”.
Scientific Publications
A list of her main scientific publications is given below, in addition to which she
wrote over a hundred semi-popular papers relating to plant diseases.
1909. The flora of certain kaffir beers. Transv. Dept. Agric. Sci. Bull. No. 8.
1915a. A bacterial disease of the mango, Bacillus mangifera, n.sp. Ann. Appl. Biol. 2: 1-45.
1915b. Some notes on the South African Erysiphaceae. Trans. Roy. Soc. S. Afr. 5: 237-245.
1915c. The South African mulberry blight, Bacterium mori. Ann. Appl. Biol. 2: 113-124.
1916a. On the occurrence of Bacterium campestre in South Africa. S. Afr. J. Sci. 11: 401^409.
1916b. The origin and cause of citrus canker in South Africa. Union Dept. Agric. Sci. Bull. No. 8.
1917a. The bacterial blight of pear blossoms occurring in South Africa. Ann. Appl. Biol. 4 : 50-74.
1917b. A bacterial spot of citrus. Ann. Appl. Biol. 3: 52-80.
1917c. The Perisporiaceae. Trans. Roy. Soc. S. Afr. 5: 713-750.
1919a. The role of plant diseases. S. Afr. J. Sci. 16: 65-92.
1919b. The bacterial blight of beans. Bacterium phaseoli. S. Afr. J. Sci. 15: 503-505.
1919c. Walnut bacteriosis. S. Afr. J. Sci. 15: 407-412.
1919d. Revisional notes. Trans. Roy. Soc. S. Afr. 7: 193-198.
1920a. Notes on four species of Meliola hitherto unrecorded from South Africa. Trans. Row Soc.
S. Afr. 8: 107-110.
1920b. Notes on an interesting collection from Natal. Trans. Roy. Soc. S. Afr. 2: 137-144.
1920c. Melio/aster. A new genus of the Microthyriaceae. Trans. Roy. Soc. S. Afr. 8: 121-124.
1920d. Mycological notes. I. Trans. Roy. Soc. S. Afr. 8: 117-120.
1920e. Some changes in the nomenclature of South African Ascomycetes. S. Afr. J. Nat. Hist.
2: 39^11.
1920f. South African Microthyriaceae. Trans. Roy. Soc. S. Afr. 8: 235-282.
1921a. A tomato canker. Ann. Appl. Biol. 7: 407-430.
1921b. The haustoria of the genera Meliola and Irene. Trans. Roy. Soc. S. Afr. 9: 117-127.
1921c. South African Ascomycetes in the National Herbarium. I. Bothalia 1 : 5-32.
1922a. South African Ascomycetes in the National Herbarium. II. Bothalia 1: 65-82.
1922b. A fungus of economic importance on the avocado. Bothalia 1: 179-186.
1924a. South African Ascomycetes in the National Herbarium. III. Bothalia 1: 195-221.
1924b. A preliminary check list of plant diseases occurring in South Africa. Mem. Bot. Surv. S. Afr.
No. 6, pp. 56.
1924c. The cause of citrus scab. Trans. Brit. Mycol. Soc. 10: 119-121. (With E. J. Butler).
1927a. South African Ascomycetes in the National Herbarium. IV. Bothalia 2: 229-241.
1927b. A preliminary study of the South African rust fungi. Bothalia 2: 1-228.
1928a. The South African species of the Meliolineae. Bothalia 2: 372-424. (With H. Sydow).
253
1928b. South African rust fungi. II. Bothalia 2: 473-474.
1929. A study of some Alternarias affecting citrus in South Africa. Union Dept. Agric. Sci. Bull.
No. 69, pp. 27.
1931. A revised list of plant diseases occurring in South Africa. Mem. Bot. Surv. S. Afr. No. 11,
pp. 78. (With A. M. Bottomley).
1936. The fungi which causes rots in stored citrus fruits in South Africa. Union Dept. Agric. & For.
Sci. Bull. No. 162, pp. 23. (With J. E. van der Plank).
1938. Some South African Fusaria. Bothalia 3: 331-483.
1939. South African rust fungi. III. Bothalia 3: 487-512.
1941a. South African rust fungi. IV. Bothalia 4: 229-236.
1941b. Some South African Valsaceae. Bothalia 4: 47-74.
1941c. South African Ascomycetes in the National Herbarium. V. Bothalia 4: 193-217.
1942a. A revision of the South African Microthyriaceae. Bothalia 4: 273—420.
1942b. Revised descriptions of Phyllachora and related genera. Bothalia 4: 421-463.
1948a. South African rust fungi. V. Bothalia 4: 895-918.
1948b. South African rust fungi. VI. The species of Uromyces on Iridaceae. Bothalia 4: 919-937.
1948c. Cercospora species recorded from Southern Africa. Bothalia 4: 881-937. (With C. Chupp).
1950. The South African fungi and lichens to the end of 1945. Bothalia 5: 1-1094.
.
255
Bothalia 9, 2: 255-307.
The Species of Arthrocnemum and Salicornia
(Chenopodiaceae) in Southern Africa
by
H. R. Tolken
Abstract
In the genus Arthrocnemum, ten species and three varieties are recognized. The new species and
combinations are: A. mossianum, A. terminate, A. xerophilum, A. decumbens, A. natalense var. affine
(Moss) (A. affine Moss), A. pillansii var. dunense (Moss) (A. dunense Moss). A. perenne var. lignosum
is a new record for South Africa. In Salicornia, of the three species dealt with, one is new: S. uniflora.
Two keys are provided for the identification of fertile and sterile material respectively. The anatomy
is discussed. The presence or absence, and the type and shape of sclereids were found to have taxonomic
significance. Twenty-four naturally occurring hybrids in the genus Arthrocnemum and one in the
genus Salicornia are recognized on the basis of sterility and pollen irregularities.
Introduction
Throughout the world this taxonomically critical group has provided difficulties
at both specific and generic levels. Moquin (1840) separated the genus Arthrocnemum
from the genus Salicornia L. on the basis of only one character, viz. the perennial
habit of the former as against the annual habit of the latter. Although he later (Moquin,
1849) attempted another circumscription of the two genera, based mainly on endosperm
characters which proved to be unreliable, it was Ungern Sternberg (1866) who formulated
Arthrocnemum to include species with endospermous seeds while those with ex-endo-
spermous seeds were placed in Salicornia.
This definition was adopted by Bentham & Hooker (1880), Volkens (1894), Paulsen
(1918), Chevalier (1922) and Ulbrich (1934). However, doubts soon arose as to whether
the genera should be separated on a single character such as this. In addition, the
absence of seeds in many specimens makes it difficult to ascribe them to either genus
and, even recently, species have been somewhat indiscriminately transferred from one
genus to the other.
Wright (1912) in Flora Capensis probably followed Ungern Sternberg’s circum-
scription of the two genera when he upheld three species and three varieties of Salicornia.
His delimitation of the species soon proved to be unsatisfactory, while the names of
two European species, S. herbacea and S. fruticosa, were incorrectly used when applied
to South African species.
Since then the main student of the Salicorniae in South Africa has been C. E.
Moss, who had previously worked on this group in Europe (Moss, 1911; 1912; 1912b;
Moss & Salisbury, 1914), before he came to South Africa. Unfortunately he did not
bring his notes on the South African species to finality before his death in 1930.
Adamson eventually undertook the task of editing and publishing the notes (see Moss
1948 and 1954). In the former, Adamson clearly accredits the work to Moss when
he states that “ the following descriptions are practically as they were drawn up by
their author ”. I have not seen Moss’s manuscript dealing with the Salicorniae and
have based my interpretations entirely on Adamson’s publications. Unfortunately
Moss’s notes, possibly due to their being edited and published posthumously, proved
256
to be insufficiently critical. In addition, the keys to the species soon proved to be
unsatisfactory, while in some cases the nomenclature and typification may not have
been presented as Moss would have desired. The need thus became apparent for a
further taxonomic revision of the two genera in Southern Africa and the present review
became urgently required as a result of recent ecological studies of estuaries.
Moss (1948; 1954) accepted Moquin’s original circumscription of the two genera
and added that the flowers of Arthrocnemum are protogynous and have bifid stigmas
while those of Salicornia are protandrous and have tufted stigmas. This separation
proved to be effective as applied to the South African species and is accepted as a basis
for the present study, although it may be mentioned that Dalby (1962), Ball (1964)
and Ferguson (1964) have queried whether some of the floral characters can be used
for distinguishing the genera.
In the revision of the two genera (Moss, 1954), two species of Salicornia and
fifteen species of Arthrocnemum were recognized. The genus Arthrocnemum was sub-
divided into two subgenera and two sections. In my opinion, however, the affinities
are rather different from the concept presented by Moss (1954), and the subdivisions
he proposed were not adhered to in the present study. His species order was also
changed following the affinities revealed by morphological and anatomical information
but, unfortunately, technical difficulties did not permit the application of cytological
evidence to these problems.
The subgenus Angianthemum, for instance, was separated by Moss from the sub-
genus Gymnanthemum on the basis of its hidden flowers and the presence of endosperm.
However, no endosperm was observed in A. africanum and A. affine, both now placed
as synonyms of A. natalense, and their inclusion together with A. indicum in the subgenus
Angianthemum, on the basis of the hidden flowers only, is unfortunate. If affinities
are evaluated, it becomes obvious that A. indicum is different, both anatomically and
also in its seed structure, from the other species of Arthrocnemum and Salicornia in
Southern Africa. Experiments have shown that the seeds do not germinate as readily
as those where a fully developed embryo with chlorophyll and no endosperm is found.
It seems, therefore, that the character, endosperm present or absent, is not an isolated
character, as indicated earlier, and that its significance in the dissemination biology
seems to be much more complex than was realized by Moss (1954). However, further
investigations on a world-wide basis are needed before the generic delimitations can
be reliably evaluated.
Anatomical and seed characters appear to be of taxonomic value and show a
remarkable coincidence when used for the grouping of the species. Although A.
natalense, for instance, is superficially very similar to A. indicum, it is different in the
seed structure and the anatomy, which are more like those of all the other species of
Arthrocnemum and Salicornia investigated. Yet the complete absence of sclereids
in the palisade tissue, as well as passage cells, which, at least in some cases, are
undeveloped sclereids (see p. 261), is peculiar. Sclereids, in particular the spirally
thickened sclereids, are very characteristic of the species of Salicornia and Arthroc-
nemum.
Also, the two very similar species A. mossianum and A. terminate differ from the
rest in the shape of their uniformly thickened sclereids, the almost black, glabrous
seeds and odd stigma behaviour. The absence of chloroplasts in the secondary cortex
of the old branches separates these two species from other shrubby species (A. pil/ansii,
A. littoreum, A. xerophilum) which, again, are similar to one another. On the other
hand A. decumbens and A. capense, the other two South African species with uniformly
thickened sclereids, exhibit greater affinities to A. perenne, a species with spirally
257
thickened sclereids, than to A. mossianum and A. terminate. The four species with
spirally thickened sclereids are very similar to one another although each tends to
grow in a different habitat. An evaluation of the affinities of the South African species
of the two genera with those from other countries is extremely difficult as the descriptions
of these plants in many floras are inadequate.
The revision presented here is based mainly on fresh or preserved specimens
collected by the author, and on extensive habit and habitat observations. Although
most of the material was collected in the salt marshes at Milnerton near Cape Town,
I have investigated various populations of Arthrocnemum and Salicornia species at
many different localities from Liideritz Bay around the coast to Durban. Preserved
specimens from Lourenco Marques and Swakopmund were also at my disposal.
In this treatment, Southern Africa is delimited by the Kunene River on the west
coast and the Limpopo River on the north-east coast. This delimitation seems to
cover the main centre of distribution of this group of plants in Southern Africa.
Immediately to the north of this area only S', pachystachya and A. indicum are recorded
on the east coast, and A. indicum and possibly also A. natalense var. affine on the west
coast.
The author wishes to express his grateful appreciation in particular to Dr. E. A.
Schelpe for his guidance and encouragement. He also wishes to thank the staff of the
Bolus Herbarium and the Botany Department, University of Cape Town, for their
helpful advice and for providing the facilities that made this study possible. Prof.
H. B. Rycroft kindly looked up some type specimens while visiting the Leningrad
Herbarium. The author is also indebted to the Chief of the Botanical Research Institute,
Department of Agricultural Technical Services, for support in connection with the
project and for travelling facilities, and to Dr. W. Macnae (University of Witwatersrand)
for valuable information and specimens from Mozambique. Specimens have been
loaned, or other assistance has been provided, by the directors and curators of the
following herbaria: State Herbarium of South Australia, Adelaide; Albany Museum
Herbarium; Herbarium, Berlin-Dahlem; British Museum; Herbarium, Royal Botanic
Gardens, Kew; National Botanic Gardens and South African Museum herbaria at
Kirstenbosch; Moss Herbarium, University of the Witwatersrand; Natal Herbarium;
National Museum of Natural History, Paris; Stellenbosch Herbarium; Windhoek
Herbarium.
Taxonomic Value of some Celaracters
The species of Arthrocnemum and Salicornia are extremely variable in habit and
in the shape and size of the different organs. In addition, a description from which
living plants as well as herbarium material can be identified, is a requirement not easily
met in succulent plants. Therefore measurements given in the descriptions are rarely
of taxonomic value and usually a range is given which includes the dimensions of a
particular character in both the living as well as dried plants. Clear-cut diagnostic
characters are rare in this group and consequently it is important to know the range
of variability of the characters used.
1. Adventitious Roots
Adventitious roots are commonly found in the species of Arthrocnemum with
prostrate growth, but they have also been observed in all salt marsh species of this
genus, if the stems are buried. Normally, adventitous roots are absent in the erect and
shrubby species, especially in herbarium material, because specimens are usually collected
from the top of the plants. Konig (1960) records several cases of rooting at the nodes
in the genus Salicornia in Europe, if the branches are buried by dune sand, but this
behaviour has not been observed in South African species.
258
2. Succulent Leaf Sheath
If the anatomy, and in particular the vascular traces of the succulent stems of
the Salicorniae, is investigated it is found that “ the succulent cortex which covers
the internodes of a Salicornian shoot is foliar in origin” (De Fraine, 1911). In fact
the ‘ cortical envelope ’ of the segments of the stems seems to represent inverse dorsi-
ventral (in comparison with the general use of the term), opposite leaves which have
enlarged and fused laterally and of which the abaxial sides have expanded vertically
to cover the whole internode below the respective node. According to the recurved
upper leaf traces, the whole abaxial surface of the leaf has experienced a downward
movement. A difference between the tissues of the leaves and those of the stem cannot
be ascertained.
In herbarium material the dried leaves often develop characteristic shapes and
colours. The characteristic spinescent appearance of A. mossianum, which distinguishes
this species from all the other shrubby species of Arthrocnemum, is due to the markedly
spreading leaf apex. It must also be mentioned at this stage that a distinction must
be made between dried leaves of herbarium material and dead leaves which are normally
adhering to the branches of living plants. For example, in A. capense, the dried leaves
are cylindrical to obconical with a spreading leaf apex, whereas the dead leaves are
barrel-shaped and cork-like.
The lateral ridges of a segment are the vertical colourless ridges along the central
axis on the upper half of the leaf rudiments. Sometimes they are pronounced ridges
but often only a colourless line below the leaf apex.
3. Colour of the Segments
In Salicornia, Moss and Salisbury (1914), Ball and Tutin (1959) and Aellen (1961)
have used the colour of the segments for distinguishing some of the species. However,
Konig (1960) considered segment colour useless in the delimitation of the species as
he found no clear-cut distinctions.
Colours of the fleshy segments were found to be useful field characters in separating
species of Arthrocnemum , but it is very difficult to describe these colours. If plants
are observed from a distance of about two metres a specific tinge can often be recognized.
However, if the plant is investigated from nearby, a full range of colours is found.
Colour differences are most noticeable in the red range. Here it is immediately realized
that A. littoreum and A. perenne var. lignosum produce no red colouration, or only a
light tinge. Their segments fade to yellow, whereas in the other species a brownish-red
or a bluish-red coloration can be observed.
These colours were found to be a useful hint in the identification of the parents
of putative hybrids. Good examples of this are the two hybrids between A. natcilense
and A. pillansii.
4. Anatomy of the Branches
In this study further work has been done on the anatomy of the different taxa
in an attempt to discover additional characters on which they might be separated.
For detailed accounts of the anatomy of this group of plants, one may refer to De
Fraine (1911), Baumgartel (1917) and Metcalf and Chalk (1950). For purposes of
identification, I found transverse sections of the succulent stems sufficient. All South
African species of Arthrocnemum and Salicornia are very similar in the basic arrange-
ment of the tissues except A. indicum. This species has, in contrast to all the other
species, only one layer of palisade cells, equal in size and each one about twice as long
as wide. A single layer of isodiametric cells, which normally become lignified with
age, is found to the inside of the palisade tissue. This layer was not observed in other
species investigated.
259
Two characters were found to be of taxonomic value in the other species of
Arthrocnemum and Salicornia. They are the sclereids in the palisade tissue, and the
nature of the secondary cortex.
(i) Sclereids and Passage Cells in the Palisade Tissue
Although the sclereids in the palisade tissue were mentioned by Moss (1954), no
use was made of them in the delimitation of species, as is done here. Basically two
types of sclereids can be distinguished and both types, even in young stages, do not
contain chlorophyll.
Firstly, there are uniformly thickened sclereids, i.e. a sclerenchyma cell, varied
in form, and having thick lignified walls with many pits. They are often long and slender
or even branched in A. mossianum. Into this group fall A. capense, A. decumbens,
A. mossianum and A. terminale.
Secondly, there are the spirally thickened sclereids which are sclerenchymatous
cells normally longer and wider than the palisade cells, becoming lignified in spiral
striations. Sclereids of this type are found in A. pillansii, A. xerophilum, A. littoreum,
A. perenne and the South African species of Salicornia.
De Fraine (1911) reported finding intermediate forms between the two types in
A. perenne and some species of Salicornia. In the case of plants with spirally thickened
sclereids, uniformly thickened sclereids are also found around the stele. However,
the sclereids around the stele are always of the uniformly thickened type, except in
A. xerophilum. In this species spirally thickened sclereids or uniformly thickened
sclereids with additional trabecula-like bars of lignin occur frequently around the stele
especially at the nodes.
It was observed that in South African species either the one or the other type of
sclereids is found, if the sclereids in the palisade tissue only are considered. No interme-
diates were observed. Young stages where the spiral striations are still incomplete should
not lead to confusion as the walls are either not lignified or not lignified uniformly.
This is often found in A. perenne var. lignosum where spirally thickened sclereids rarely
develop fully.
Also, sclereids are often found in flowering spikes in the water tissue surrounding
the flowers. Their form and development suggest that they are merely lignified paren-
chyma cells. However, sometimes, especially in A. decumbens, the pit canals are
narrowly widened at their mouth in such a way as to give the appearance of incomplete
striations in surface view.
The faint spiral striations usually found in A. pillansii are almost perpendicular
to the lateral walls. This character, which is used to separate this species from other
species with spirally thickened sclereids, cannot be applied to putative hybrids involving
A. pillansii. In such hybrids the spiral striations are always oblique to the lateral
walls.
Sclereids in the palisade tissues of putative hybrids were studied extensively and
three main groups were distinguished among such hybrids.
(a) Hybrids in which A. natalense is one parent usually develop no sclereids,
irrespective of the sclereids exhibited by the other parent. Only in hybrids of A. natalense
with A. mossianum were uniformly thickened sclereids present.
(b) Hybrids between species having the same type of sclereid also exhibit that
type of sclereid. In such cases, if there is a difference in shape between the sclereids
of the parents, usually an intermediate shape of sclereid is found in the hybrid.
260
(c) Hybrids from parents having different types of sclereids, show intermediate
types in the progeny, i.e. in such instances slcereids with more or less complete striations
overlaying otherwise uniformly thickened walls are produced (Fig. 2: 6). Again,
A. mossianum is the exception as, in the hybrid A. mossianum x A. pillansii, sclereids
similar to those of A. mossianum are found.
Fig. 1. — 1, Diagram of a transverse section through an internode of Arthrocnemum and Salicornia.
2, Arthrocnemum capense, paradermal section of a segment illustrating the distribution of the
palisade tissue, passage cells and sclereids. The stomata are at right angles to the axis.
airch._air chamber below stomata; camb. — cambium; e. — endodermis; epi. — epidermis; l.t.—
vascular leaf traces; p. — pith; pal.t. — palisade tissue; pas.c. — passage cells; p.can.— pith canal;
pr.p. — primary phloem; pr.x. — primary xylem; sec.c. — secondary cortex; sec.p. — secondary
phloem; sec.x. — secondary xylem; u.scl.— uniformly thickened sclereids; w.t.— water tissue.
261
In the genus Salicornia only spirally thickened sclereids have so far been found.
The European species S. dolichostachya and S. strictissima, have no sclereids at all
(Nannfeldt, 1955).
Large non-chlorophyllose cells often occur around the sclereids in the palisade
tissue (Fig. 2). They are referred to as passage cells as this appears to be their function.
They are twice or three times as wide as the surrounding palisade cells. Usually a
row of two or three such cells extends across the palisade tissue from the epidermis
to the water tissue. It seems significant that these passage cells are most commonly
found in species with uniformly thickened sclereids, whereas in species with spirally
thickened sclereids they are rare or completely absent. In species with the latter type
of sclereid, the initial stages of the sclereids and the passage cells are very alike. In
fact, the passage cells are only undeveloped spirally thickened sclereids, as found in
A. perenne var. lignosum. Normally, however, the passage cells become about twice
as wide as the sclereids.
Baumgartel (1917) found that no sclereids are formed if the seedlings are cultivated
under very moist conditions. The present author observed that usually in very old
segments of such plants growing in more moist regions, e.g. A. decumbens at Lourenco
Marques as opposed to Mossel Bay, eventually at least some sclereids are formed.
However, dry conditions and apparently also high salt consentrations seem to promote
the formation of sclereids.
Paradermal sections reveal that the sclereids in general are arranged singly or in
groups in the palisade tissue between the stomata. They occur only occasionally
almost next to the air chambers below the guard cells, but are never in contact with
them. In A capense the sclereids together with the passage cells produce an almost
complete reticulation around a clump of palisade cells below the respective stomata
(Fig. 1:2).
(ii) Secondary Cortex
Towards the outside the cambium produces a parenchymatous secondary cortex.
In species with a shrubby habit and spirally thickened sclereids, such as A. pillansii,
A. littoreum and A. xerophilum, the cells of the secondary cortex of old stems contain
chlorophyll, but this does not occur in those species with uniformly thickened sclereids,
such as A. mossianum and A. terminale.
In Salicornia the thickness of the secondary cortex appears to be a useful taxonomic
character. As it is applicable only to older specimens, it might be partly due to the
much stouter nature of S. pachystachya, which exhibits a thicker secondary cortex
than do the other two species, S. meyerana and S. uniflora.
5. Inflorescence
The inflorescence or the “ spike ” of Arthrocnemum and Salicornia is a compound
inflorescence consisting of three- or more-flowered, sessile cymules in the axils of the
opposite decussate bracts. Perhaps the term spike-like thyrse is more appropriate.
A recent tendency has been to substitute “ spike ” by “ pseudo-spike ” (Scheinahre
in Konig, 1960; Aellen, 1961). This might point out the wrong use, but it is still not
the correct morphological term. As the term “ spike ” has been so widely used in
taxonomic treatments of this group, this terminology is used in the present study.
262
pIG. 2. — Sclereids of the palisade tissue in relation to the stomata and passage cells. 1, Arthrocnemum
decumbens, sclereids usually rectangular, not reaching the epidermis. 2, A. capense, sclereids
usually narrower in the middle, reaching the epidermis. 3, A. mossianum, branched sclereids
reaching the epidermis and into the water tissue. 4, A. terminate, branched sclereids not reaching
the epidermis but into the water tissue; 4a, sclereids not branched. 5, A. pillansii, spirally thickened
sclereids; 5a, detail of the cell wall showing lignin deposited in striations. 6. A. decumbens X A.
pillansii, lignified walls plus spiral striations.
pal.t. — palisade tissue; pas.c. — passage cells; w.t. — water tissue.
The bracts are anatomically similar to the leaf with the exception that, in the former,
more sclereids are often found both in the palisade tissue and around the stele. The
internodes do not elongate as much as the barren branches, which results in a condensed
inflorescence. Characters such as tapering spikes, fruiting segments thicker than the
263
barren segments or bulging segments are often good indications of species concerned.
Spike length on the other hand must be used with care, as it is often very variable,
especially in species of Arthrocnemum.
In A. natalense a continuation of vegetative growth above the spike has often
been observed. Alternating vegetative and fruiting growth may be repeated up to three
times. This terminal proliferation was also recorded, but much less frequently, in
A. pillansii and A. capense.
Another phenomenon which is thought to be lateral proliferation is the presence
of lateral branches of spikes, frequently observed in a form of S. meyerana. Lateral
branches may develop in this form in the position normally occupied by one or both
lateral flowers of the cymule. Another branch below the central flower can also occur
with all three flowers developed. The central flower is never absent. These lateral
branches always bear flowers.
A similar arrangement of the branches has been observed in the axils of non-
fruiting segments of the same form of S. meyerana. Here the maximum of eight
branches per node is possible, yet seldom found, but up to six branches in various
combinations are not uncommen at a node. It is remarkable that a whole population
of this odd form was found at Saldanha Bay.
6. Cymules
In all the species of Arthrocnemum investigated, the flowers are completely sessile,
but in A. natalense a bulge indicates the pedicel of the flower. However, in all the
species a vascular strand, one or two tracheids thick, is visible at the base of each
flower.
In Arthrocnemum all the flowers are approximately equal in height, whereas in
most of the species of Salicornia the central flower is bigger and does not completely
separate the two lateral flowers. Only in the Australian species S. blackiana (which
should be placed in Arthrocnemum ) are two to three flowers often found below the central
flower. Cymules are usually three-flowered but up to sixteen flowers per cymule have
been recorded for S. australis Sol. ex Benth. (Cooke, 1912). The number of flowers
may vary greatly and Moss (1954) grossly overrated the usefulnes of this character
in delimiting A. variiflorum and, even more so, in separating the section Trianthemum
and Polyanthemum on the number of flowers per cymule only. Moss (1954) must
have been aware of this as he admits that “ A . variiflorum forms a connecting link
between the sections . . However, there are certain species, such as A. indicum
and A. capense, with consistently three flowers per cymule.
7. Flowers
In all species of Arthrocnemum and Salicornia the perianth tube exhibits a number
of segments at the apex. The number of segments at the apex of the perianth tube
is constantly three or four in the genera Arthrocnemum and Salicornia in South Africa.
Black (1919) does not regard these segments as perianth lobes in S. australis, but
considered them to be the result of bursting open of the perianth. If this view is correct,
one would expect irregular segments as the result of fragmentation. In South African
species these perianth lobes are already visible before anthesis. In the two species
with laterally compressed perianth tube, namely A. indicum and A. natalense, usually
only three lobes are seen. In all the other species of Arthrocnemum and Salicornia
where the truncate perianth tube is found, usually four segments were observed, rarely
only three in lateral flowers.
264
Furthermore, in all the species except A. indicum, an abscission layer is formed
at maturity of the seeds well above the tissues of the bracts, so that a ridge outlining
the flower remains when the nodes have fallen out.
A useful additional character for separating some species, e.g. S', meyerana from
S. pachystachya, was found to be whether the perianth becomes cork-like when mature
or shrivels up. However this character must be used with care, since in certain species,
such as A. perenne, a cork-like perianth is often an indication of a specimen grown
under unfavourable conditions.
8. Stamens
The number of stamens was never found to exceed two. The record of four or
five stamens by McCann (1952) in the Australasian species of Salicornia is open to
doubt since Allan (1961) failed to find more than two in the New Zealand species also
investigated by McCann. However, sometimes one or no stamens are produced in
functionally female plants as recorded for A. pillansii and A. mossianum.
The size of the anthers as a taxonomic character has been investigated in Salicornia
in Europe (Ball, 1964), but does not give workable results in South Africa.
9. Ovary and Stigma
The ovary in all species is very similar except in A. mossianum, where two lateral
grooves adjoining the anthers are found. This is a very useful character and is clearly
visible in hybrids derived from this species.
In A. natalense and A. indicum, the bifid stigma is papillose and spreading. In
A. pillansii, A. littoreum and A. xerophilum, the bifid stigma is plumose and erect, the
lobes adhering to one another. In A. perenne, A. capense and A. decumbens, the bifid
or trifid stigma is plumose with reflexed lobes, sometimes even adhering to the fleshy
segments. In Arthrocnemum the stigma was found to be a very reliable character, but
unfortunately the stigmata are extremely brittle and are not visible in herbarium
material.
10. Seeds and Seed Dispersal
The present author found seed characters, especially the papillae and hairs on the
seed testa, very reliable in the delimitation of species. Among the South African
species of Arthrocnemum, the seeds of A. indicum are unique, because they produce
a crustaceous pericarp, while endosperm is present. In all the other South African
species of Arthrocnemum and Salicornia, the embryo is conduplicate and without
endosperm. At the end of the seed the radicle and the cotyledons are separated by
part of the testa which is invaginated. In most of the species this double membrane,
resulting from invagination, is not well fused and the radicle can be split from the
cotyledons without damaging the seed testa. It appears that embryological studies
might reveal interesting data pertaining to the relations of the taxa above the species
rank.
In most of the species the perianth tube remains attached to the seeds. Wind
and especially rain apparently aid the release of these propagules, which drop onto
the ground below the parent plant. If the soil is moist enough, the seeds germinate
within two days and the seedlings establish themselves within a week. In the lower
zones of salt marshes, the seeds are washed away as the result of tidal movement or
river flooding. It is, however, interesting to note that the dried perianth enables the
seeds to float. But the question arises how these species can establish themselves if
the seeds are washed away. In temporary vleis the seeds have just enough time to
265
establish themselves before the lower levels of the vleis are submerged after rain. It
was noted that seedlings of many species of Arthrocnemum and Salicornia were able
to endure a long period of submergence, during which time the seedlings were not
observed to grow.
In contrast with this, endospermous seeds, such as those of A. indicum, are released
only by the eventual decay of the hard cork-like spike. Also, germination is erratic
and takes from a week to three weeks.
11. Flowering Period
In the field it has been observed that all the plants of a species at a given locality
usually flower within a very definite period, and that only the hybrids in each population
flower at irregular times. Most of the species flower from January to May. Judging
from herbarium records, however, there appears to be some variation in the times
of flowering at the different localities, which is not unexpected as the species of Arthroc-
nemum and Salicornia occur in a wide range of climates. For instance, A. natalense
seems to flower twice in Natal, namely in December to April, and flowers begin to
appear again in June or July. On the other hand A. natalense var. affine seems to
flower throughout the year. Only A. capense and A. xerophilum are exceptions in
the genus Arthrocnemum and flower in October and November.
As specimens in the flowering and fruiting stage look so alike, the term flowering
period, used in the descriptions, includes both.
12. Habitat and Zonation
Two of the species of Arthrocnemum and Salicornia in South Africa do not occur
in salt marshes. A. xerophilum is found on dry slopes in the Knersvlakte (Vanrhynsdorp
district) among succulents. The other species is A. littoreum, which grows in rock
crevices along the sea shore but, in contrast to the previous species, it grows under
moist conditions.
All the other species were observed to occur at characteristic levels in salt marshes.
Although the limits of the various zones are not always clear-cut, a fairly clear zonation
is always noticeable. In fact, the species of Arthrocnemum occur at definite levels of
the salt marshes and, although this is not of great taxonomic value, it is interesting
for the delimitation of the species.
Populations of Arthrocnemum and Salicornia at lower levels may be subjected to
either flooding by fresh water in the rainy season or may be inundated by tidal sea
water. According to which is applicable, either A. natalense or A. perenne , respectively,
occur in the lowest zone. This is indicated by the distinction between “ inland salt
marshes ” and “ maritime salt marshes ”. However, in estuarine salt marshes the
two types may eventually merge where the effects of the fresh w'ater overlap those of
the tidal movement. Inland salt marshes include the temporary vleis in the Cape and
also those estuaries which Day (1952) has referred to as “blind estuaries”, because
their access to the seas has been barred by a sand bank. In these localities, A. natalense
occupies the lowest zone because of its ability to endure submergence for up to three
months. In contrast with this, A. perenne can endure only short periods of inundation.
At Rugby, near Cape Town, this species occurs together with A. natalense in seasonal
vleis and it occupies the zone at the highest water level and above. However, this
population of A. perenne is most unusual because this species normally grows in
maritime salt marshes. It is thought that its occurence at this locality must be attributed
to acclimatization, as Day (1952) demonstrated for animals, after these vleis were cut
off from the main lagoon.
266
S. meyerana is generally found in the lower zones of salt marshes. It is a pioneer
plant and does not necessarily occur in a definite zone of its own. Plants of this annual
species die before enough rain water has accumulated to submerge them in inland
salt marshes, but they also occur in maritime salt marshes.
The next zone higher up in the salt marshes is occupied by A. decumbens, immediately
followed by the A. piUansii zone. It appears that there is a strong competition between
the two species and the zonation is effected by seasonal flooding, because A. decumbens
seems to be more tolerant to occasional inundation than A. piUansii. Both the varieties
of A. piUansii always occur outside the reach of the high water level.
Higher up still in the zonation occurs A. mossianum and at similar levels, but at
different localities A. terminate is found. Neither of these species was ever found
in localities near the sea. At the higher level the salinity factor seems to be operative.
A. capense also occurs in a zone above A. piUansii, and is especially abundant
in sandy localities. It is thought that it requires well-aerated soils.
Putative Hybrids
In taxonomic investigations certain intermediate specimens have been recorded.
Moss often identified them as hybrids. In the present study, when pollen was investi-
gated, it was found that the grains of these intermediates are often not round as is
normal, but are dented and in addition there are much smaller grains without an
intermediate range being observed. In the immature pollen it was found that the
“ tetrads ” were abnormal in being composed of more than four grains of varying
sizes. Goodspeed (1954) found that a similar phenomenon in hybrids in the genus
Nicotiana was due to irregularities in meiosis: “ The smaller nuclei contain chromosomes
which did not reach the poles at the end of anaphase I . . . At the sporal stage two
or more lagging chromosomes may become united and are surrounded by a membrane ”.
In addition, in putative hybrids no seeds develop or these are aborted at an early
stage, and in no case was a fully developed embryo found, although often the brown
seed membrane can be recognized. The two characters, the presence of abnormal
pollen and the absence of seeds, were taken as a criterion for hybrids in this work.
Unfortunately it was not possible to substantiate this view by the experimental production
of artificial hybrids, as Turesson (1925) could prove in the genus Atriplex, also a member
of the Chenopodiaceae.
In South Africa it was observed that salt marshes with gently sloping banks and
ill-defined zonation often have a higher frequency of hybrids. In a few localities, such as
at Milnerton, hybrids appeared to form about a quarter of the population. The large
quantity of hybrid plants in this locality is primarily due to the occurrence of the two
very vigorous hybrids, A. piUansii X A. natalense and A. perenne x A .pillansii. These
hybrids produce extensive mats up to three metres in diameter and apparently compete
very successfully with their parents.
Since the species of Arthrocnemum and Salicornia are wind-pollinated and all
species flower at about the same time, hybridization is highly likely when they grow
together. In species like A. piUansii , in which male-sterile plants may occur, the possi-
bility of hybridization is enhanced by the occurrence of purely female plants which
must be fertilized with pollen from plants nearby. This may perhaps account for the
many hybrids of which A. pillansii is one of the parents.
In the genus Salicornia in Southern Africa only one putative hybrid has been
recorded. This can probably be ascribed to the fact that the different species tend
to grow in different localities. In Europe, where more than one species of Salicornia
may occur at the same locality, hybridization is more likely. However, recent authors
such as Konig (1960), Aellen (1961) and Dalby (1962) could not find any evidence
for such hybrids.
267
Taxonomically the hybrids are not as important as might seem from the treatment
now presented, and the number of such specimens seen in herbaria did not constitute
more than 1 per cent. When it was realized that these hybrids are sterile, a special
search was made for them in order to describe them and avoid later confusion. As
a result of this close study, it was found that apparently two types of hybrids may occur
between each pair of parents but, nevertheless, I did not succeed in finding all the
possible combinations.
In the field, hybrids can be recognized according to several peculiarities which
they may exhibit. With some knowledge of the species occurring in an area, hybrids
can easily be recognized by their abnormal habit. The branches are often excessively
tortuous and lateral branches arise from the main branches at varying angles. Once
struck by the odd habit, one immediately recognizes the intermediate character of the
plant, while colours are often especially striking. For example, in a population at
Bredasdorp containing A. natalense and A. decumbens, a hybrid showed the habit
and colour of A. natalense and could be mistaken for this species at first glance. The
multi-flowered cymules and the type of flowers of this hybrid indicate that A. decumbens
is the other parent. Dried material could, however, be mistaken for A. perenne or
A. decumbens.
The stigmata of hybrids were found to be larger and to persist for a longer time
even after the pollen is shed, than in the parent plants. Even more characteristic is
the irregular flowering of hybrids with odd spikes being produced throughout the year.
This presumably has resulted in many collections of hybrids when the species them-
selves are not flowering. However, the main flowering period for all the hybrids is
from February to May, even in putative hybrids of A. capense, a species which
normally flowers in October and November.
Furthermore the fruiting spikes of hybrids usually dry from the apex downward.
The spikes of the parental plants become mature from the lower cymules upwards,
and the whole spike dries at more or less the same time. Also the perianth tubes shrink
considerably when drying and are, as a rule, not cork-like.
The hybrids are classified in three groups according to the sclereids in the palisade
tissue. Hybrids between species with uniformly thickened sclereids and spirally thickened
sclereids usually produce an intermediate type of sclereid (Fig. 2: 6). The walls of
these sclereids are at first uniformly thickened but produce additional, often incomplete,
spiral striations. A separation of specimens with this type of sclereid from those with
spirally thickened sclereids would be impractical as sections of herbarium material
are often insufficiently clear to distinguish between the two.
A key to the hybrids recognized by the present author, and which probably includes
all the commoner combinations, is given below, although this does not exclude the
possibility that additional hybrid combinations could be found.
In accordance with Article 1, Appendix 1, of the International Code of Botanical
Nomenclature (1961) the names of the putative parents in the formula of the hybrids
were placed in alphabetical order. However, in view of the fact that two different
types of progeny seem to be possible from the hybridization of a pair of putative parents
a type A and a type B were designated, indicating whether the resemblance of the
hybrid is considered to be greater to the parent placed first or second in the formula.
268
Key to the Putative Hybrids
Group A. Hybrids without sclereids in the palisade tissue; flowers usually with ill-defined lateral
shield; (usually hybrids with A. natalense as one of the parents).
la. Cymules 3-flowered:
2a. Branches woody, ascending or decumbent:
3a. Branches ascending: segments green 1. A. natalense X A. terminate
3b. Branches decumbent, almost procumbent; segments glaucous 2. A. natalense X A. pillansii
2b. Branches herbaceous, succulent, prostrate or decumbent:
4a. Barren segments succulent, 3-5 mm thick:
5a. Dried segments light coloured 5. A. capense X A. natalense
5b. Dried segments dark coloured 3. A. natalense X A. perenne
4b. Barren segments scarcely succulent, 2-3 mm thick 4. A. capense x A. natalense
lb. Cymules (3-) 4— 7-flowered:
6a. Cymules 3 or 4 (-5)-flowered ; flowers with ill-defined lateral shield 6. A. decumbens x A. natalense
6b. Cymules 4-7-flowered: flowers with clearly defined lateral shield 7. A. decumbens X A. natalense
Group B. Hybrids with uniformly thickened sclereids.
la. Barren segments 8-15 mm long, 3-5 mm thick; branches often rooting at the nodes 11. A. decum-
bens X A. capense
lb. Barren segments 3-8 mm long, if longer then narrower than 3 mm; branches not rooting
at the nodes:
2a. Plants prostrate or procumbent; barren segments often longer than 5 mm:
3a. Terminal spikes 15-20 mm long 8. A. mossianum X A. natalense
3b. Terminal spikes 20-30 (-70) mm long 9. A. mossianum X A. natalense
2b. Plants ascending rarely decumbent; segments 3-5 mm long. . 10. A. mossianum X A. pillansii
Group C. Hybrids with spirally thickened sclereids or uniformly thickened sclereids with spiral
strations superimposed.
la. Cymules 3-7-flowered:
2a. Plants ascending, woody 21. A. decumbens X A. pillansii
2b. Plants prostrate, scarcely woody:
3a. Branches straight: cymules 3-5 (-7)-flowered 23. A. decumbens X A. perenne
3b. Branches tortuous: cymules 3- or 4-flowered 24. A. decumbens X A. perenne
lb. Cymules 3-flowered:
4a. Flowers scarcely exposed, somewhat laterally flattened, with ill-defined lateral shield:
5a. Barren segments 8-12 mm long 12. A. natalense X A. pillansii
5b. Barren segments 12-18 mm long 13. A. natalense var. affine X A. pillansii
4b. Flowers exposed, obpyramidal, with lateral shield:
6a. Branches becoming woody:
7a. Dead segments barrel-shaped, cork-like:
8a. Plants with ascending branches: spikes cylindrical when fruiting 18. A. capense x A.
pillansii
8b. Plants with decumbent branches, sometimes creeping; spikes tapering when fruiting
17. A. capense X A. pillansii
7b. Dead segments not barrel-shaped and not cork-like:
9a. Sclereids uniformly thickened with usually incomplete spiral striations 14. A. terminate
X A. pillansii
9b. Sclereids spirally thickened 16. A. pillansii x A. perenne
6b. Branches scarcely woody:
10a. Dead segments barrel-shaped, cork-like:
11a. Fruiting segments bulging:
12a. Spikes often clustered at the end of branches, thicker than barren segments
15. A. perenne X A. pillansii
12b. Spikes not clustered at the end of branches, not thicker than barren segments 20. A.
capense X A. perenne
269
lib. Fruiting segments not bulging:
13a. Barren segments 3-4 mm thick; spikes 20-30 mm long 24. A. decumbens X A. perenne
13b. Barren segments 2-3 mm thick; spikes 10-20 mm long, if longer than segments
not succulent:
14a. Plants much branched; segments green fading to brown or brownish-red
19. A. cape use X A. perenne
14b. Plants little branched; segments glaucous fading to bluish-red 17. A. capense X
A. pillansii
10b. Dead segments cylindrical or obconical, not cork-like:
15a. Spikes usually cylindrical when flowering; fruiting segments not bulging 21. A. decum-
bens X A. pillansii
15b. Spikes tapering when flowering; fruiting segments bulging:
16a. Plants much branched; spikes terminal 24. A. decumbens X A. perenne
16b. Plants little branched; spikes terminal and lateral.. 15. A. perenne X A. pillansii
Group A
1. A. natalense X A. terminate
Cape.— Vanrhynsdorp : Salt River, Tolken 557.
This hybrid resembles A. terminate in producing dull green and brownish-red
segment colours and in its shrubby habit, but the spikes are terminal and lateral, the
lateral shield ill-defined, the stigmas are spreading and the sclereids in the palisade
tissue are absent; all these characters are typical of A. natalense.
2. A. natalense X A. pillansii (Type B)
Cape. — Peninsula: Milnerton, Tolken 542. Bredasdorp: Seekoeivlei, Tolken 514. Port Elizabeth:
Swartkopsrivier, Archibald 4900/3 (GRA).
This hybrid resembles A. pillansii in its woody branches and the glaucous green
and bluish-red segment colours, but it also exhibits a procumbent habit and spreading
stigmas characteristic of A. natalense.
3. A. natalense x A. perenne
Cape. — Peninsula: Salt River, Moss 8754.
The specimen of this hybrid appears to be one involving A. natalense, especially
since it exhibits spreading leaf apices and lack of sclereids. However, it has exposed
flowers with a definite lateral shield typical of A. perenne, another prostrate species.
A very similar specimen ( Moss 8757) is probably a plant of A. perenne grown under
very moist conditions, so that the sclereids are not formed. Moss on the sheet described
it as A. capense X A. perenne which is highly unlikely because of the absence of sclereids.
4. A. capense X A. natalense (Type A)
Cape. — Peninsula: Kommetjie, Moss 8883b (J).
Most of the specimens on the sheet must be regarded as normal A. capense, but
one detached branch is certainly a hybrid, although it looks very much like the rest
of the collection. The flowers of this specimen contain aborted anthers and no seeds
are developed. No sclereids but only passage cells were observed, whereas the rest
of the collection has sclereids typical of A. capense.
5. A. capense X A. natalense (Type B)
Cape. — Peninsula: Kommetjie, Moss 11681 (J).
The specimen of this hybrid resembles A. natalense in its fleshy, obconical segments
and the absence of the sclereids, but the exposed flowers with a lateral shield resembles
those of A. capense, the other species present at the locality.
270
6. A. decumbens x A. natalense (Type A)
Mozambique. — Lourenco Marques: Matola Road, Tolken 1091; Matola Bridge, Tolken 1094.
This hybrid differs from A. decumbens X A. natalense (Type B) by its shorter
and barrel-shaped, rarely cylindrical segments which are glaucous green. Also 3 or
4 (-5)-flowered cymules are produced. This hybrid resembles A. decumbens in its
glaucous segment colours, fading to bluish-red and rarely individual segments becoming
yellow. It is similar to A. natalense in its long tapering spikes with flowers with ill-
defined lateral shield and the absence of sclereids and passage cells in the palisade
tissue.
7. A. decumbens x A. natalense (Type B)
Cape. — Bredasdorp: De Hoopvlei, Tolken 517.
Mozambique. — Lourenco Marques: Matola Bridge, Tolken 1093.
When fruiting this putative hybrid superficially resembles A. perenne, but the
fertile segments are hardly wider than the barren segments, and 4 to 7 flowers per
cymule are found. This hybrid is different from A. decumbens x A. natalense (Type A)
in that it produces 4-7-flowered cymules and the flowers have a clearly defined lateral
shield. This hybrid is very similar to A. natelense in its green, often yellow or yellowish-
brown segments colours and the absence of sclereids, but occasionally passage cells
occur.
Group B
8. A. mossianum X A. natalense (Type A)
Cape. — Caledon: 10 miles west of Riviersonderend, Tolken 528.
This hybrid resembles A. mossianum in its woody branches and short spikes with
truncate flowers, as well as having the typical sclereids in the palisade tissue. However,
the uniformly thickened sclereids vary in form, 2-3 times as long as wide, angular
and not branched. The stomata are sunken. The pink colour of the segments which
are often shiny, the prostrate habit and the more succulent segments are characteristic
of A. natalense.
9. A. mossianum X A. natalense (Type B)
Cape. — Bredasdorp: Spitzkop, Tolken 522.
This hybrid is similar to A. natalense in producing long tapering spikes with flowers
with ill-defined lateral shield and branches tough in texture. The dead leaves do not
adhere to the branches, the segments are glaucous in colour and the type of sclereids
in the palisade tissues is typical of A. mossianum. The uniformly thickened sclereids
are angular and variable in form but never branched. The stomata are sunken.
10. A. mossianum x A. pillansii
Cape.- — Caledon: 10 miles west of Riviersonderend, Tolken 554.
The uniformly thickened sclereids present in the palisade tissue are not branched.
This hybrid is very similar to A. mossianum, but the branches are flexuous and flexible,
instead of producing straight erect, brittle branches. This and the many spikes clustered
at the end of branches are reminiscent of A. pillansii. As this hybrid does not produce
any stamens, it could be confused with functionally female plants of A. mossianum.
271
11. A. decumbens X A. capense
Cape. — Caledon: Uilenkraalrivier, Tolken 502.
Although A. decumbens was not found in the immediate vicinity, it occurred about
half a mile further up the river. The second putative parent can, however, only be
A. decumbens because of the segment colours and the type of sclereids, which do not
reach the epidermis. The frequent rooting at the nodes and the sunken stomata are
characteristic of A. capense.
Group C
12. A. natalense X A. pillansii (Type A)
Cape. — Peninsula: Rugby, Tolken 547.
The pink colour of the segments, the long tapering spikes with flowers with ill-
defined lateral shield and spreading stigmas found in this hybrid are characteristic
of those of A. natalense , but the shrubby habit and spirally thickened sclereids in the
palisade tissue are reminiscent of A. pillansii.
13. A. natalense var. affine X A. pillansii
Cape. — Hopefield : 8 miles west of Koperfontein, Tolken 474b.
The long prostrate main branches and long segments of this hybrid resemble
those of A. natalense var. affine. However, characters such as the woody branches
and spirally thickened sclereids in the palisade tissue are characteristic of A. pillansii.
14. A. pillansii X A. terminate
Cape. — Namaqualand: Swart Lintjiesrivier, Pillans in BOL 18215b (BOL, STE).
On both sheets only one sprig of this hybrid is found together with the typical
A. terminate . The spreading branches and the incomplete spiral striations of the
sclereids indicate hybridity with A. pillansii, the other species recorded from the locality.
Although no fertile material of this hybrid was available, the specimen was immediately
suspected of hybridity.
15. A. perenne X A. pillansii (Type A)
Cape. — Peninsula: Rugby, Tolken 549.
This hybrid is very similar to A. capense x A. perenne (Type A) but differs in
having bulging fruiting segments as in A. perenne. It differs from A. perenne x A.
pillansii (Type B) in having prostrate main branches, which are tough in texture and
are rarely woody in old branches. The bulging fruiting segments and the brownish-red
segments colour is similar to A. perenne , but the clustered spikes and the non-succulent
segments are rather like A. pillansii.
16. A. perenne x A. pillansii (Type B)
Cape. — Peninsula: Rugby, Tolken 550.
This hybrid differs from A. perenne x A. pillansii (Type A) and A. perenne in
having brittle hardly ascending, secondary branches which become woody with age.
However, the almost prostrate main branches with ascending lateral branches are very
unlike the habit of A. pillansii.
17. A. capanse x A. pillansii (Type A)
Cape. — Peninsula: Milnerton, Adamson s.n. (CT); Rugby, Tolken 548; Paardeneiland, Tolken 147.
This hybrid resembles A. capense in its scarcely woody but tough branches with
old leaves adhering to them and its prostrate growth, but the clustered spikes and
spirally thickened sclereids are characteristic of A. pillansii. This hybrid is very similar
to A. perenne x A. pillansii (Type A), but differs in producing barrel-shaped and
cork-like dead leaves, and not bulging fruiting segments.
272
18. A. capense X A. pillansii (Type B)
Cape. — Peninsula: Paardeneiland, Tolken 546.
This hybrid is very similar to A. pillansii, but can be recognized by its persistent
stigmas and by the fact that the flowers are equal in size and instead of bulging, they
are dented. These flowers and the sunken stomata are typical of A. capense. In
herbarium material the cork-like segments are characteristic. It differs from A. capense
X A. pillansii (Type A) in having a shrubby habit.
19. A. capense X A. perenne (Type A)
Cape. — Peninsula: Paardeneiland, Tolken 551; Salt River, Moss 8761 (J).
This hybrid differs from A. perenne X A. pillansii (Type A) in having a different
colour and smooth spikes. This hybrid resembles A. capense in its cylindrical spikes
of which the segments do not bulge and in the cork-like dead leaves adhering to the
branches, however the spirally thickened sclereids are typical of A. perenne.
20. A. capense X A. perenne (Type B)
Cape. — Peninsula: Rugby, Tolken 549.
This hybrid is in all its characters very similar to A. perenne when grown under
dry conditions, but the thick walls of the sclereids suggest hybridity with A. capense.
21. A. decumbens X A. pillansii (Type A)
Cape. — Caledon: Mierkraal, Tolken 532. Port Elizabeth: Swartkopsrivier, Archibald 5821.
This hybrid differs from A. decumbens x A. pillansii (Type B) in its decumbent
habit and its consistently 3-flowered cymules. This hybrid resembles A. decumbens
in its decumbent habit, whereas the clustered spikes and spirally thickened sclereids
in the palisade tissue indicate hybridity with A. pillansii.
22. A. decumbens X A. pillansii (Type B)
Cape. — Caledon: 10 miles west of Riviersonderend, Tolken 531. Port Elizabeth: Swartkopsrivier,
Archibald 5820 (BOL, GRA).
This hybrid differs from A. decumbens X A. pillansii and A. decumbens in its
ascending habit similar to that of A. pillansii, but it often produces more than three
flowers per cymule, a character typical of A. decumbens.
23. A. decumbens x A. perenne (Type A)
Cape. — Bathurst: Kowie River, L. Britten 5200 (GRA).
Mozambique. — Lourenco Marques: Lourenco Marques, Macnae 1.
This hybrid has straight not flexuous branches, tapering spikes when flowering
and usually more than three flowers per spike in contrast with A. decumbens x A.
perenne (Type B). In herbarium material it is almost indistinguishable from A. decum-
bens. In fresh material it resembles A. perenne in producing green segments, not
glaucous as characteristic in A. decumbens.
24. A. decumbens X A. perenne (Type B)
Cape.— “Albany ”, Phillips s.n.
This hybrid differs from A. decumbens X A. perenne (Type A) in having cork-like
segments and the branches thin and flexuous. This hybrid resembles A. perenne in
its prostrate habit and flexuous branching and three flowers per cymule, but the much
thickened walls of the sclereids suggest hybridity with A. decumbens.
273
Key to the Genera in Southern Africa
Perennial, prostrate to ascending; not all branches ending in fruiting spikes;
stigma bifid or trifid; normally protogynous 1. Arthrocnemum
Annual shrublets; all branches ending in fruiting segments; stigma tufted; nor-
mally protandrous 2. Salicornia
1. ARTHROCNEMUM
Arthrocnemum Moq., Chenop. Monogr. Enum. Ill (1840); in DC., Prodr. 13,
2: 114 (1849); Ung.-Sternb., Vers. Syst. Salic. 36 (1866); in Atti Congr. Int. Bot.
Firenze 1874: 268 (1876); Benth. & Hook.f., Gen. PI. 3, 1 : 65 (1880); Moss in Adamson
in J. S. Afr. Bot. 20: 4 (1954); Brenan in FI. Trop. E. Afr. Chenopodiaceae : 18
(1954). Type species: A. fruticosum (L.) Moq.
Perennial plants, prostrate, decumbent or ascending. Branches often rooting at
the nodes and not all branches terminating in flowering spikes. Cymules often three
flowered but also up to nine-flowered. Flowers equal in size, or nearly so, normally
distinctly protogynous or protandrous (A. mossianum), rarely flowering simultaneously.
Stigma bifid or trifid, minutely papillose to plumose, up to 4 mm long, distinctly visible
in fresh material. Seeds with membranous testa, glabrous, papillose or covered with
hairs, with no endosperm and a conduplicate embryo or with hard pericarp, endosperm
present and a straight embryo.
The genus has a world wide distribution and occurs mainly in saline marshes.
Key to the Species of Arthrocnemum based on Flowering Material
Hybrids are not included in this key: they are dealt with separately (see pp. 268).
Hybrids can be recognized by no embryo developing in the seeds; the pollen grains
not round but often dented; and much smaller grains occuring together with normal
ones. Indications of hybrid material are: —
(1) Branches often tortuous.
(2) Prominent stigmas persisting for a long time.
(3) Flowering spikes produced outside the normal flowering period.
la. Flowers covered by the bract below, or nearly so; perianth tube dorsiventrally flattened,
opening at the apex of the flower:
2a. Flowers deeply embedded in and fused to upper segment; fruit with glabrous, hard pericarp;
endosperm present; fruit released only with the decay of the spike l.A. indicum
2b. Flowers scarcely embedded in and not fused to the segments; seeds with papillose, mem-
branous testa; endosperm absent; fruit falling out easily 2. A. natalense
lb. Flowers scarcely covered by the bract below; perianth tube obpyramidal, opening in the centre
of the lateral shield:
4a. Branches (not including the succulent leaf sheath) 5-30 mm in diam., not swollen at the
nodes, often laterally compressed: plants ascending:
5a. Spikes terminal, solitary, rarely more than one per branch; secondary branches erect,
fastigiate; fleshy perianth becoming cork-like, not shrivelling when fruiting:
6a. Growing in inland salt marshes; spikes less than 25 mm long; vegetative segments
8-15 mm long, 2-3 mm wide 4. A. terminate
6b. Growing on gravelly beaches; spikes 20-30 (-100) mm long; vegetative segments
5-8 mm long, 3-5 mm 7. A. littoreum
5b. Spikes clustered at the end of branches; secondary branches spreading; fleshy perianth
shrivelling when fruiting (if not, then segments shorter than 5 mm):
7a. Segments rarely up to 4 mm in length, distinctly keeled; leaf apex very much spreading,
at right angles to stem when dried; uniformly thickened sclereids present in the palisade
tissue 3. A. mossianum
274
7b. Segments usually longer than 4 mm, faintly keeled, leaf apex hardly spreading when
dried; spirally thickened sclereids present in the palisade tissue:
8a. Succulent leaf sheaths shrinking considerably when dried, exposing the bare stem
between them 6. A. xerophilum
8b. Succulent leaf sheaths hardly shrinking when dried 5. A. pillansii
4b. Branches (not including the succulent leaf sheath, if present) up to 5 mm in diam., rarely
thicker, but then conspicuously swollen at the nodes, not laterally compressed; plants
prostrate, decumbent:
9a. Seeds glabrous or irregularly papillose; spikes clustered at the end of the branches;
branches not rooting at the nodes 5. A. pillansii
9b. Seeds uniformly papillose or covered with hairs; spikes usually terminal, solitary on
the branches; branches often rooting at the nodes:
10a. Seeds papillose; 5-7 (-10) flowers per cymule; stems decumbent, rarely rooting at
the nodes 10. A. decumbens
10b. Seeds covered with hairs; 3 rarely 4 or 5 flowers per cymule; stems prostrate, usually
rooting at the nodes:
11a. Flowers always 3 per cymule; barren segments 10-16 mm long, 2-3 mm in diam.;
uniformly thickened sclereids present in the palisade tissue 9. A. capense
lib. Flowers 3-5 per cymule: barren segments 5-15 mm long, 3-5 mm in diam.; spirally
thickened sclereids present in the palisade tissue 8. A. perenne
Key to the South African Species and Varieties of Arthrocnemum and Salicornia
BASED ON ANATOMICAL AND VEGETATIVE CHARACTERS
As there are numerous specimens without flowers in herbaria, a second key based
on anatomical and vegetative characters has been constructed. In the identification
of herbarium material, even flowering specimens of doubtful nature can be checked
by means of their anatomical characters. Hybrids are not included in the key, but are
dealt with separately (see p. 268).
All the anatomical characters used are based on transverse sections of the segments,
but it is advisable to cut longitudinal sections in case branched sclereids are present.
When dealing with herbarium material, a whole old segment or the top half of
an old segment, preferably one that is not completely flattened by the drying process,
is detached. It is soaked in tap water for one or two days. (The addition of a little
salt was found helpful). Only in the case of very cork-like segments is it advisable
to heat the material to boiling point in order to hasten the process of softening. Other-
wise, boiling normally softens the tissue too much and, apart from the treated material
being difficult to cut, it never regains its shape satisfactorily. Ordinary hand-sections
are cut just below the membranous leaf apex, preferably several sections through the
lateral ridges and several sections below that.
In assessing the anatomical features, it should be noted that the cells of the water
tissue touching the epidermis in the lateral ridges are not regarded here as passage
cells.
Furthermore the author is well aware that sunken stomata do not constitute a
very constant character (Esau, 1960, Baumgartel, 1917). However, for the identification
of herbarium material which is normally collected in the field and not cultivated, the
position of the stomata was found to be consistent.
la. Palisade tissue a single cell thick; palisade cells of equal size, surrounding a single layer of
isodiametric, lignified cells, in contact with the vascular bundles 1. A. indicum
lb. Palisade tissue 2-3 (-4) cells thick; palisade cells not of equal size, continuing into the paren-
chymatous water tissue, lignified cell layer absent; vascular bundles not in contact with the
palisade tissue:
2a. Sclereids absent in the palisade tissue:
3a. Passage cells absent; palisade tissue with all cells containing chloroplasts 2. A. natalense
4a. Segments 5-15 mm long, 2-4 mm in diam.; main branches herbaceous, with lateral
branches often longer than half their length var. natalense
275
4b. Segments 15-35 mm long, 4-6 mm in diam.; main branches hard, with lateral branches
never longer than half their length var. affine
3b. Passage cells or other non-chlorophyllose cells present in the palisade tissue:
5a. Palisade tissue with many passage cells present; main branches usually tortuous, thin
(2 -4 mm in diam.) 10. A. decumbens
5b. Palisade tissue with very few passage cells present; main branches usually straight,
thick (4-10 mm in diam.) 8(b). A. perenne \ ar. lignosum
2b. Sclereids present in the palisade tissue:
6a. Uniformly thickened sclereids present:
7a. Sclereids branched and narrower than the palisade cells:
8a. Sclereids longer than the layer of the palisade tissue, reaching the epidermis
3. A. mossianum
8b. Sclereids shorter than the layer of palisade tissue, not reaching the epidermis
4. A. terminate
7b. Sclereids long, rectangular, as wide as, or wider than, the palisade cells:
9a. Sclereids rectangular but normally narrower in the middle, reaching the epidermis
9. A. capense
9b. Sclereids rectangular, not reaching the epidermis 10. ^4. decumbens
6b. Spirally thickened sclereids present:
10a. Sclereids about 8 times as long as wide 6. A. xerophilum
lib. Sclereids 2-4 (-6) times as long as wide:
11a. Old segments not becoming cork-like; erect annual shrublets; (usually the whole
plant with tap root is collected):
12a. Occurring in Nothern Natal and Mozambique, secondary cortex 5-8 cells thick
S. pachystachya
12b. Occurring in South Africa except northern Natal and Mozambique; secondary
cortex 2-3 (-4) cells thick:
13a. Lateral branches spreading; shrublets up to 30 cm high; stomata protruding
S. meyerana
13b. Lateral branches erect, fastigiate; shrublets up to 15 cm high; stomata sunken
S. uniflora
lib. Old segments often becoming cork-like; perennials, erect bushes or prostrate herbs;
(normally only a branch, often with adventitious roots, is collected):
14a. Branches usually without succulent leaf sheath, 5-30 mm in diam. (excluding
leaf sheath), often laterally compressed; segments normally becoming light brown,
seldom greyish when dried; plants frutescent or shubby:
15a. Chloropasts absent in the secondary cortex of branches that have lost their
leaf sheath 8 (b). A. perenne var. lignosum
15b. Chloroplasts present in the secondary cortex of branches that have lost their
succulent leaf sheath:
16a. Spiral striations of the spirally thickened sclereids normally faint, perpendicular
or nearly so to the lateral walls; stomata at the same level as the epidermis:
17a. Branches erect, fastigiate; growing on gravelly beaches.... 7. A. littoreum
17b. Branches spreading; growing in inland salt marshes 5. A. pillansii
16b. Spiral striations of spirally thickened sclereids normally strongly lignified,
oblique to the lateral walls; stomata sunken:
18a. Segments 2 (3) mm in diam.; western Cape Province 5. A. pillansii
18b. Segments (3-) 4-6 mm in diam.; Luderitz Bay, S.W.A. 5 (b) A. pillansii var.
dunense
14b. Branches usually with succulent leaf sheath but, if not, then up to 5 mm in diam.,
not laterally compressed; segments normally becoming dark coloured when
dried ; plants prostrate :
19a. Adventitious roots absent; spiral striations of spirally thickened sclereids
perpendicular, or nearly so, to the lateral walls 5. A. pillansii
19b. Adventitious roots often present; spiral striations of spirally thickened sclereids
oblique to the lateral walls:
20a. Branches without succulent leaf sheath, rarely exceeding 2 mm in diam.
8 (a) A. perenne var. perenne
20b. Branches without succulent leaf sheath 2-5 mm in diam. 8 (b) A. perenne var.
lignosum
276
1. A. indicum ( Willd .) Moq., Chenop. Monogr. Enum. 113 (1840); in DC., Prodr.
13, 2: 151 (1849); Ung.-Sternb., Vers. Syst. Salic. 69 (1866); in Atti Congr. Int. Bot.
Firenze 1874:274 (1876); Hook.f., FI. Brit. India 5, 1: 12 (1886); Bak. & C.B.C1. in
FI. Trop. Afr. 6, 1 : 86 (1909), quoad descr., excl. specim. cit. ; Moss in Adamson in
J. S. Afr. Bot. 20: 5 (1954); Brenan in FI. Trop. E. Afr. Chenopodiaceae: 18 (1954).
Type: India, Madras Presidency, Tranquebar, Klein in Herb Willdenow 54 (B, holo.;
PRE, photo.!).
Salicornia indica Willd. in Ges. Naturf. Fr. Neue Schriften 2: 111-112, pi. 4, fig. 2
(1799); Vahl, enum. 1: 10 (1804); Wight, Icon. 3, 1, fig. 737 (1844).
Woody or scarcely woody perennial, prostrate or decumbent up to 25 cm high,
forming mats about 60 cm in diameter. Barren segments 5-9 mm long, 4-6 mm in
diameter, obconial to barrel-shaped, with keeled leaf apex and lateral ridge, glaucous
green fading to bluish-red; leaf apex of young leaves spreading when dried, cylindrical
and cork-like in older leaves; dead leaves adhering to the branches. Sclereids uniformly
thickened and in shape similar to the palisade cells; palisade tissue in a single layer,
each cell about twice as long as wide, surrounding a single layer of isodiametric cells
which become lignified. Spikes terminal, 10-30 mm long, rarely longer, tapering
when flowering, often cylindrical when fruiting; flowering segments wider than the
barren segments below; fruiting spikes becoming hard and cork-like, and fruit released
by its decay. Cymules 3-flowered; flowers completely embedded in lower and upper
segments. Perianth tube 3-dentate, not truncate but dorsiventrally flattened, opening
at the apex, fused to the upper segments. Anther 1 or 0. Stigmas bifid, papillose,
spreading. Fruits triangular, much flattened laterally with hard pericarp, glabrous,
light yellow to brown; endosperm starchy, anterior; embryo straight, posterior.
Fig. 3:3.
A. indicum occurs in salt marshes in all tropical regions around the Indian Ocean
and occasionally on the west coast of Africa. No specimen of this species from South
Africa has been seen by the present author, and the Galpin collection reputedly from
Eshowe, cited by Moss as the southernmost locality, could not be traced. Flowering
and fruiting period from July to March.
Mozambique. — Lourenco Marques: Maputo, Hornby 2591; Lourenco Marques, Moss 3135 (GRA,
J); St. Georges Island, Gomes e Sousa 4166; Matola Road, Moss & Ottley 11740 (J); Porto Amelia,
Macnae in J 37543 (J).
Brenan (1954) mentions that he was unable to find stamens in the material investi-
gated. However, in fresh material from Lourenco Marques an anterior stamen was
observed, but it does not develop to its full size in all flowers and, in addition, the
staminode is sometimes fused to the anterior wall of the perianth tube. A second
posterior stamen was never seen.
Although Moquin effected the combination for this species under Arthrocnemum,
there is some doubt whether he referred to the same species. “ Semen . . . hispidulum.
Albumen parcissimum. Radicula exerta . . .” suggests seeds very unlike those of
A. indicum.
According to Moss (1954) the greater part of A. macrostachyum sensu Hiern,
Cat. Afr. PI. Welw. 1,4: 899 (1900), non Moris et Delport, must be referred to A.
indicum. This can also be deduced from the descriptions of the specimens by Hiern.
The same specimens are cited by Baker & C. B. Clark (1909) under A. macrostachyum
and A. fruticosum and not under A. indicum, though mentioned as a different species.
When Chevalier (1922) transferred Salicornia pachystachya to the genus Arthroc-
nemum, he based his decision on the collection, Perville 661, a specimen of A. indicum
from Madagascar (see S. pachystachya, p. 298).
277
2. A. natalense ( Bunge ex Urtg.-Sternb.) Moss in Adamson in J. S. Afr. Bot. 20: 15
(1954). Type: “Port Natal”, Drege s.n. (LE, lecto. ; PRE, photo.!).
Scarcely woody perennial, prostrate or decumbent, forming dense mats or solitary
in irregular rosettes 20-80 cm in diameter, not higher than 30 cm; main branches
flat on the ground, often rooting at the nodes. Barren segments 5-35 mm long, 2-6
mm in diameter, obconical or cylindrical, slightly keeled, with or without a lateral
ridge, shining to dull grey-green or light yellowish-green, fading to brownish-yellow
or pinkish-red ; leaf apex spreading when dried; dead leaves adhering to the branches.
Sclereids absent in the palisade tissue. Spikes terminal, rarely lateral, (10-) 20-30
(-80) mm long, tapering when flowering and fruiting; flowering segments not wider
than barren segments below. Cymules 3-flowered; flowers hidden by the bract below
in fresh material, in dried material flowers often exposed, protogynous. Perianth
tube with three dentate lobes, seldom with a fourth anterior lobe, not truncate but
dorsiventrally flattened, adhering to the seeds. Anthers 1 or 2. Stigmas bifid, papillose,
spreading. Seeds compressed, oblong to ovate with membranous testa, rough or
papillose; endosperm absent; embryo conduplicate.
A. natalense is the most common of all the species occurring in South Africa and
is found from Angola to Mozambique, mainly along the coast, but also as far inland
as Calvinia, Worcester and Riversdale. It has also been recorded from Madagascar
( Bosser 15431). It does not thrive when submerged by sea water, even temporarily.
Moss and others remarked on the similarity of A. natalense and A. indicum but the
resemblance is merely superficial. If the flowers, seeds and the anatomy are investigated,
significant differences are found.
Two varieties are recognized in South Africa:
Barren segments 5-15 mm long, 1-4 mm in diameter; main branches softly herbaceous, with
lateral branches often longer than half their length (a) var. natalense
Barren segments 15-35 mm long, 4-6 mm in diameter; main branches hard, with lateral branches
never longer than half their length (b) var. affine
(i a ) var. natalense
Salicornia natalensis Bunge ex Ung.-Sternb., Vers. Syst. Salic. 62 (1866); in Atti
Congr. Int. Firenze 1874:292, fig. 11 (1876); Wright in FI. Cap. 5, 1:450 (1911).
Salicornia fruticosa var. capensis Ung.-Sternb., Vers. Syst. Salic. 59 (1866), non Wright.
Type: “Cap. (‘Salicornia indica, W. Ecklon und Zeyher pi. Cap. 26n. oder ’62,
11 ’?)”.
Arthrocnemum natalense (Bunge ex Ung.-Sternb.) Moss in Adamson in J. S. Afr.
Bot. 20: 15 (1954). A. africanum Moss ex Fourcade, Mem. Bot. Surv. S. Afr. 20:19
(1941), nomen nudum. A. africanum Moss in Adamson in J. S. Afr. Bot. 14: 37 (1948);
J. S. Afr. Bot. 20: 7 (1954); Adamson in FI. Cape Penins. 357 (1950). Type: Peninsula,
Fishhoek, Moss in BOF 14243 (BOL, holo.!).
Herbaceous perennial, prostrate, forming dense mats, seldom higher than 15 cm;
much branched with inconspicuous main branches and secondary branches irregularly
branched and longer than half the main branches, often tortuous, soft and flexible.
Barren segments 5-15 mm long, 2-4 mm in diameter, usually obconical, shining to
glaucous green. Seeds compressed, oblong to obovate. Fig. 3:1.
278
Fig. 3. — 1, Arthrocnemum natalense var. natalense, flowering spike; la. fruiting spike; lb, gynoecium;
lc, seed in surface view; Id, seed in longitudinal section. 2, A. natalense var. affine, seed in surface
view. 3, A. indicum, flowering spike; 3a fruit in longitudinal section,
end. — endosperm; per. — pericarp.
This variety is recorded from Lambert’s Bay and the Vanrhynsdorp district along
the South African coast to Mozambique. It grows normally in soils of lesser salinity
and not along the sea unless it has a supply of fresh water. It occupies the lowest
zone in inland salt marshes. In the Cape it is often submerged in temporary vleis
for up to three months. The old leaves rot off and only small terminal and lateral
buds on the apparently dead branches remain dormant until the branches re-appear
above the surface of the water. This has led to the frequently expressed view that this
species is an annual. The flowering and fruiting period is usually from January to
May. In Natal it begins to flower again in June-July indicating that it has two main
flowering seasons there.
279
Cape. — Bathurst: Kowie River, L. Britten 2118 (GRA, J, K); Great Fish River, Tolken 396.
Bredasdorp: Elim, Tolken 505; De Hoop Vlei, Barker 8729 (NBG); Spitzkop, Tolken 519. Caledon:
Botrivier, Tolken 488; Hermanus, L. Guthrie in BOL 1700 (BOL); Riviersonderend, Tolken 530.
Calvinia: Leighton 2440. Clanwilliam: Lambert’s Bay, Pole Evans 27; Tolken 142 (BOL). East
London: Buffalo River, Ottley 2568 (J); Nahoon River, Galpin 5673. Elliotdale: Bashee River,
Macnae in J 33177 (J). George: Wilderness, Stephens in BOL 18839 (BOL). Hopefield: 2 miles
west of Hopefield, Tolken 417. Humansdorp: Gamtoos River, Tolken 376. Knysna: Swartvlei,
Martin 4521; Keurboomsrivier, Tolken 409. Malmesbury: Ysterfontein, Tolken 484a. Mossel Bay:
Great Brak River, Moss 4342 (J). Peddie: Great Fish River, Tolken 396; Keiskama River, Tolken
401. Peninsula: Rugby, Tolken 347; 348; Salt River, Burchell 667 (K); Black River, H. Bolus 3708
(BOL); Uitvlugt, Wolley Dod 2691 (BOL); Mowbray Flats, Marloth 8145; Noordhoek, Barnard
in SAM 44148 (SAM); Fishhoek, Moss in BOL 14243 (BOL). Port Elizabeth: Swartkopsrivier,
Archibald 5814 (BOL, GRA); Mow 17041 (J). Port St. Johns: Second Beach, Tolken 405. Riversdale:
Muir 4645 (J); Gouritz River, Tolken 350. Stellenbosch: Kuilsrivier, Tolken 300. Somerset West:
Strand, Parker 4732 (BOL, K, NBG). Vanrhynsdorp : Salt River, Acocks 14187. Vredendal: near
Heerenlogement, Tolken 539. Worcester: Over Hex Station, Tolken 491.
Natal. — Durban: Isipingo, Forbes & Obermeyer 32; Strey 4566; Durban Bay, Pillans 6277.
Hlabisa: St. Lucia Estuary, Ward 4375; False Bay Park, Ward 4216. Ingwavuma: Kosi Estuary,
Tinley & Ward 3. Lower Umfolozi: Richard’s Bay, Lawn 1608 (NH). Mtunzini: Umlalazi, Ward
4228. Umlazi: Umgababa River, Tolken 407.
Mozambique. — Lourenco Marques: Maputoland, Wager s.n. (GRA); Costa da Sol, Macnae s.n.
(BOL, J); Inhaca Island, M. Moss 27423; Mogg 27145.
The type of S. natalensis is cited as being in the Bunge Herbarium and in the type
description both flowers and adventitous roots are mentioned. Also the type is given
as being a collection made by Drege at Port Natal (Durban).
The only sheet from Herb. Bunge containing a specimen collected by Drege at
Port Natal, that could be found by Prof. Rycroft during his visit to Leningrad, was
one on which were three specimens bearing characteristic Drege collecting labels.
These specimens are annotated as “ Salicornia indica Willd. a Cap. b. sp. Drege;
Salicornia herbcicea procumbens Cap. b. sp. Drege; 80. 20 Cap. b. sp. Drege ”, Judging
from the citation in Drege, Zwei pflanzengeographische Documente, the only collection
made by Drege at Port Natal was that distributed under the name S. indica. This
specimen bears both adventitous roots and flowers and appears to be the main element
from which Ungern Sternberg prepared his description, as it is the only specimen
with flowering spikes. It seems that this specimen on the bottom left hand corner
must be taken as the lectotype despite the fact that the specimen at the bottom right
hand corner, without flowers, is accompanied by a label, in an unknown hand, reading
“ S', natalensis Bunge sp. nov. teste Bunge It is possible that this label was intended
to refer to all three specimens.
The ovate seeds and the mention of the similarity of the flowers to A. indicum
by Ungern Sternberg (1866) can apply only to A. natalense as defined here and not
to A. perenne, the only other species occurring in the vicinity of Durban. In the Cape
the seeds of this variety are oblong in contrast to the ovate seeds of the var. affine.
In Natal, however, they are more ovate. It is curious to find that Moss regarded the
rooting at the nodes, mentioned in the type description, as a major obstacle in accepting
that A. natalense is the same species as what he called A. africanum. This character
can be seen in numerous specimens. Moss’s suggestion that the Drege specimen is
a hybrid does not seem likely, as Ungern Sternberg gave such detailed descriptions
of the seeds. The present author has never found seeds in specimens regarded as being
hybrids. Eventually Moss distinguished between A. africanum and A. natalense on
the presence or absence of endosperm, respectively. In specimens identified and cited
by Moss as A. africanum no endosperm could be found and this material could not be
separated from A. natalense. The type specimen of A. africanum is in the flowering
stage and does not bear any seeds. Therefore, A. africanum is treated as a synonym
of A. natalense.
280
It is not possible to decide finally whether S. fruticosa var. capensis is also syno-
nymous with A. natalense, but the obconical segments combined with prostrate growth
strongly suggest that this is the case.
Two specimens of A. natalense ( Tolken 142a, 417) were found which show peculi-
arities in their anatomy, in exhibiting dumb-bell shaped sclereids, in longitudinal
sections of the secondary cortex. Superficially this form is indistinguishable from the
typical form of A. natalense, which lacks these sclereids.
In A. natalense, terminal proliferation, i.e. vegetative growth above the spike, can
often be observed (cf. Tolken 147b). This might be repeated up to three times.
The following putative hybrids are recorded (for key see p. 268) :—
A. mossianum X A. natalense A. capense x A. natalense
A. natalense x A. terminate A. decumbens x A. natalense
A. natalense X A. pillansii
(b) var. affine (Moss) Toelken, stat. nov.
A. affine Moss in Adamson in J. S. Afr. Bot. 20: 8 (1954). Type: Swakopmund,
Moss 18122a (PRE, lecto.!; J!).
Scarcely woody perennial, prostrate or decumbent, forming rosettes or mats up
to 30 cm high; main branches long, straight; lateral branches arising in pairs at each
node, shorter than half the length of the main branches, usually not branched. Barren
segments 15-35 mm long, 4-6 mm in diameter, usually cylindrical or nearly so, light
greyish-green or bluish-green. Seeds ovate. Fig. 3 : 2.
The var. affine is recorded from Angola, southwards along the coast to Ysterfontein.
It occurs in similar habitats to the var. natalense, except that it is also found around
lagoons with sea water. The main flowering and fruiting period is from November
to April, but it appears that in the Namib, if water is plentiful, continuous flowering
throughout the year is possible.
Cape. — Hopefield : 8 miles east of Koperfontein, Tolken 475b. Malmesbury: Ysterfontein, Tolken
484. Namaqualand: Orange River, Pillans 5577 (BOL); Wallekraal, Pillans in BOL 18218; 18219
(BOL, STE). Vanrhynsdorp: Salt River, Tolken 538.
South West Africa. — Kaokofeld: Unjab River, Giess 3927. Swakopmund: 11 miles north of
Swakopmund, Hardy & De Winter 1454; Swakopmund, Moss 18122; Merxmiiller & Giess 1740;
Bradfield 590; Dinter in SAM 71408 (SAM); Galpin & Pearson 7586; Sandwich Harbour, Giess 5049
(Windhoek).
In its typical form A. affine has stouter and longer segments, but intermediates
occur. The collection Merxmiiller & Giess 1740 consists of branches of the typical
form and of a much less vigorous plant, both found at the same locality. These two
forms are similar in habit. In typical A. affine a few conspicuous main branches with
much shorter lateral branches are found, arising regularly in pairs at each node. In
contrast, the branching in the typical A. natalense is irregular and if long main branches
are produced, such as in shade forms of this species, these do not branch as regularly
as in the typical A. affine. This difference in branching is the only distinguishing
character. The two forms are not geographically isolated and consequently A. affine
is regarded as a variety of A. natalense.
The type is a mixed collection. On the sheet of the type in PRE there is a specimen
of A. perenne var. lignosum probably from Liideritz Bay. A. perenne has never been
recorded to the north of Liideritz Bay. As recommended by the International Code
of Nomenclature (1961), Article 9, a lectotype is designated.
The following putative hybrid is recorded (for key see p. 268): —
A. natalense var. affine X A. pillansii
281
3. A. mossianum Toelken, sp. nov., ab A. pillansii segmentis obonicis ad 4 mm
longis differt.
A. hottentotieum sensu Moss in Adamson in J. S. Afr. Bot. 20: 10 (1954), partly,
excl. lectotype.
Arbuscula erecta, lignea ad 50 cm alta, ad 80 cm lata; rami primarii 8-30 cm lati,
saepe compressi, asperi, raro leves; rami secundarii, ramis lateralibus multis plerumque
1- 2 cm longis sed nonnunquam gracilioribus. Articuli obconici, clare carinati, costis
lateralibus, 2-3 mm longi, raro 4 mm longiores, 2-4 mm lati, obscure virides raro
glauci vel rubescentes; foliorum siccorum apices extensissimi; folia mortua ramis
non inhaerentia. Spicae terminales et laterales, 8-20 mm longae, ad apicem fastigiatae
flore, cylindricae fructu, ad fines ramorum dispositae. Cymae triflorae, flore medio
floribus lateralibus paulo longiori, floribus protandris. Periartthii tubus lobis 4, truncatus,
scutulo laterali, semini inhaerens. Antherae 2 vel 0. Stigmata 2 vel 3, undique papillosa.
Semina oblonga, atra, c.2 mm longa, testa coriacea, glabra, saepe nitida; endospermum
nullum; embryo conduplicatus. Plate 1. F G. 4: 1.
Type: Cape, Caledon District, 10 miles west of Riviersonderend, Tolken 552
(PRE, holo.).
Erect woody shrubs forming bushes up to 50 cm high, 80 cm in diameter; main
branches 8-30 mm thick often laterally flattened, with a rough bark, seldom smooth;
secondary branches with many short and thin lateral branches. Barren segments
obconical, distinctly keeled with a lateral ridge, 2-3 mm long (rarely up to 4 mm),
2- 4 mm in diameter, dull green or seldom glaucous, fading to brownish-red; leaf
apex much spreading when dried; dead leaves not adhering to the branches. Sclereids
in the palisade tissue branched, uniformly thickened, reaching from the water tissue
to the epidermis. Spikes terminal and lateral, 8-20 mm long, tapering when young,
cylindrical when fruiting, clustered at the end of branches; flowering segments scarcely
wider than barren segments below. Cymules 3-flowered; central flower scarcely longer
than the lateral ones; flowers protandrous. Perianth tube with 4 crenate segments,
truncate with lateral shield, cork-like when mature and adhering to the seed. Anthers
2 or 0. Stigmas bifid or trifid, plumose, erect. Seeds compressed oblong, very dark,
about 2 mm long with membranous testa, glabrous, often shining; endosperm absent;
embryo conduplicate.
A. mossianum is recorded only from the area between Caledon and Riversdale,
always away from the coast. It occupies the highest zone in salt marshes or often
on river banks. Floweiing and fruiting period from March to May.
Cape. — Bredasdorp: Vogelvlei, Galpin 11345a; Spitskop, Tolken 520. Caledon: Rietkuil, Henrici
3676 (BOL); 10 miles west of Riviersonderend, Tolken 525; 526; 552; 553. Montagu: Montagu,
Michell in BOL 15659 (BOL); Rogers in BOL 15498 (BOL); Tolken 555. Riversdale: Marloth 3587;
Docrnkraal, Muir 3079 (BOL). Swellendam: Tolken 412b.
The type collection of A. hottentotieum Moss consists of a mixture; the portion
which answers to the description must be referred to A. pillansii , while the remaining
portion represents an undescribed species (see A. pillansii, p. 2x5). It is now described
as A. mossianum in honour of the late Prof. C. E. Moss for his pioneer work done on
the South African species of Arthrocnemum and Salicornia.
Two barren specimens, Rogers in BOL 15498 and Michell in BOL 15659, from
Montagu, are in all respects bigger than the typical A. mossianum. However, from
fruiting specimens collected at the same locality ( Tolken 555) it is concluded that these
specimens represent a slightly different but localized population.
10862-2
282
Fig. 4. — 1, Arthrocnemum mossianum, flowering spike: la, fruiting spike; lb, seed in surface view;
lc, seed in longitudinal section. 2, A. terminale, flowering spike: 2a, seed in surface view.
Although the flowers of A. mossianum are usually hermaphrodite, some plants
have been found in which only female flowers are produced (cf. Tolken 525 and 526
in flowering stage; Tolken 552 and 553 in fruiting stage). In these plants the anthers
are represented only as staminodes (cf. A. pil/ansii). As A. mossianum is protandrous,
these female plants might be mistaken as protogynous, but in fact the anthers never
develop. Since only a proportion of the flowers in these female plants set seed, it is
thought that they are fertilized by pollen from hermaphrodite plants in the vicinity.
The following hybrids are recorded (for key, see p. 268): —
A. mossianum x A. natalense
A. mossianum x A. pillansii
4. A. terminale Toelken , sp. nov., ab A. mossiano articulis cylindris longioribus
5-12 mm longis, spicis terminalibus differt.
Arbuscula lignea ad 1 m alta; rami primarii ad 3 cm lati, leves, ascendentes vel
decumbentes, ramis lateralibus erectis, fastigiatis. Articuli cylindrici, obscure carinati,
costis lateralibus, 5-12 mm longi, 2-3 mm lati, obscure virides raro glauci, vel flavo-
virentes et testacei; folia sicca ut videtur sulcata et foliorum siccarum apices extensi;
folia mortua ramis non inhaerentia. Spicae solum terminales, 8-20 mm longae, non
longiores, cylindricae flore et fructu. Cymae triflorae, flore medio floribus lateralibus
paulo longiori, floribus protogynis. Perianthii tubus lobis 4, truncatus, scutulo laterali,
semini inhaerens. Antherae 2. Stigmata 2 vel plures, undique papillosa, saepe ut
videtur cristata. Semina oblonga, atra, c. 2 mm longa, testa coriacea, glabra, saepe
nitida; endospermum nullum; embryo conduplicatus. Plate 2. Fig. 4:2.
Type: Cape, Vanrhynsdorp District, Salt River, Tolken 536 (PRE, holo.).
283
Woody shrub forming bushes up to 1 m high with main branches smooth, ascending
or decumbent, with erect, fastigiate lateral branches. Barren segments cylindrical,
with faintly keeled leaf apex and lateral ridge, 5-12 mm long, 2-3 mm in diameter,
dull green to glaucous, fading to yellowish-green and brownish-red; the segments
shrinking regularly so as to appear grooved; leaf apex spreading when dried; dead
leaves not adhering to the branches. Sclereids in the palisade tissue uniformly thickened,
branched or often star-shaped, next to the water tissue, not reaching the epidermis. Spikes
terminal, 8-20 mm long, not longer, cylindrical when flowering and fruiting; flowering
segments not wider than barren segments below. Cymules 3-flowered; flowers about
equal in size, protogynous. Perianth tube with four crenate segments, truncate with
lateral shield, adhering to seeds. Anthers 2. Stigmas with 2 or more branches, plumose,
often appearing tufted. Seeds compressed oblong, very dark, about 2 mm long with
membranous testa, glabrous, often shining; endosperm absent; embryo conduplicate.
This species is found along rivers in Namaqualand, growing at a higher level on
the banks than A. pillansii. The type plant was flowering and fruiting from February to
April.
Cape. — Namaqualand: Swart Lintjiesrivier, Pillans in BOL 18215a (BOL, STE). Vanrhynsdorp :
Salt River, Tolken 536; 8 miles north of Vanrhynsdorp, Moss 18043 (J).
A. terminate is very similar to A. mossianum but differs from that species in its
long narrow segments and consistently terminal spikes, hence the choice of the name.
Flowering in A. terminate is protogynous, but the anthers soon appear in rapid
succession while the stigmas are maintained in apparently receptive condition even
after the pollen is shed in the individual flowers. Furthermore, the stigma is often
not only bifid or trifid, but with many branches, which gives it an appearance similar
to that of the tufted stigma of the species of Salicornia.
The following hybrids are recorded (for key, see p. 268): —
A. natalense X A. terminate
A. pillansii X A. terminate
5. A. pillansii Moss in Adamson in J. S. Afr. Bot. 14: 38 (1948). Type: Cape,
Milnerton, Moss 8764 (PRE, holo.!).
Woody shrub, either ascending or decumbent, up to 70 cm high; main branches
up to 30 mm thick, often laterally compressed. Barren segments cylindrical or slightly
obconical, faintly keeled with lateral ridge, 4-20 mm long, 2-6 mm in diameter, green
to glaucous fading to bluish- red ; leaf apex scarcely spreading when dried ; dead leaves
adhering to the branches, but older branches bare. Sclereids in the palisade tissue
spirally thickened, with faint spiral striations almost perpendicular to the lateral walls.
Spikes terminal and lateral, (15-) 20-30 (-70) mm long, tapering when flowering and
fruiting, rarely cylindrical, clustered at the end of branches ; flowering segments hardly
wider than barren segments below. Cymules 3-flowered ; central flower often one-third
larger than the lateral ones, or more or less equal in size; flowers protogynous.
Perianth tube with three or four crenate teeth, truncate with lateral shield, shrivelling
when mature, adhering to the seeds. Anthers 1 or 2, rarely 0. Stigmas bifid or rarely
trifid, plumose, usually erect, branches often tangled with one another, sometimes
reflexed. Seeds compressed oblong, light brown, about 1-1 • 5 mm long, with mem-
branous testa, glabrous or slightly rough; endosperm absent; embryo conduplicate.
A. pillansii occurs along the South African coast from Liideritz Bay to Mozambique.
Usually it is found in inland marshes and occassionally also near the sea, but then
always at a higher level than the flooded zone. The flowering and fruiting period is
from January to June.
10862—3
284
This species is not only very variable in form and habit but has apparently also
hybridized freely with a wide range of other species, thus forming the most diverse
complex encountered in this genus in Southern Africa. The species can vary in habit
from ascending to decumbent. Sometimes it has short spreading, almost spinescent
lateral branches with spikes up to 25 mm long, or it produces long slender branches
with long spikes up to 70 mm long.
In typical A. pillansii, specimens with very prominent stigmas, which persist for a
long time, have been observed ( Levyns 3535; Archibald 4900/4b; Tolken 543). This
form does not produce any anthers and not even staminodes were found (cf. A. mos-
sianum). When the spikes were covered with polythene bags, no seeds had been set
after a month, which leads one to conclude that these plants are functionally female.
This behaviour was observed in one individual plant in two successive years. It should
be noted that functionally female plants are also found in A. pillansii var. dunense
( Tolken 461).
Two varieties are recognized in South Africa:
Old barren segments 8-20 mm long, 2-^4 mm in diameter, rarely cork-like; stigmas 1-2 mm long;
southern Africa except Luderitz Bay (a) var. pillansii
Old barren segments 4-8 mm long, 4-6 mm in diameter, becoming cork-like; stigmas 3 or 4 mm
long; Luderitz Bay (b) var. dunense
(a) var. pillansii
Salicornia fruticosa var. densiflora Ung.-Sternb., Vers. Syst. Salic. 59 (1866); in
Atti Congr. Int. Bot. Firenze (1874) 301 (1876); Wright in FI. Cap. 5, 1:449 (1912).
Type: Cape, Paardeneiland, “ Drege Sammlung (‘ III, B), b ’ ” (LE, holo.).
Arthrocnemum pillansii Moss ex Fourcade, Mem. Bot. Surv. S. Afr. 20:20 (1940);
nomen nudum; Moss in Adamson in J. S. Afr. Bot. 14: 38 (1948); J. S. Afr. Bot.
20: 13 (1954); Adamson in FI. Cape Penins. 358 (1950). Type: Cape, Milnerton,
Moss 8764 (PRE, holo.!). A. hottentoticum Moss in Adamson in J. S. Afr. Bot. 20: 10
(1954), partly, as to lectotype & descr. Type: Cape, Milnerton, Moss 8874 (sheet
in PRE, lecto.!). A. namaquense Moss in Adamson in J. S. Afr. Bot. 20: 15 (1954).
Type: Vanrhynsdorp, Salt River, Moss 17908 (J, holo!; BOL!).
Woody shrubs, decumbent or ascending with main branches up to 30 mm in
diameter, often laterally compressed. Barren segments cylindrical, 8-20 mm long,
2-4 mm in diameter, seldom becoming cork-like when old. Sclereids in the palisade
tissue usually only a few, spirally thickened with faint spiral striations perpendicular
to the lateral walls. Stigmas short, erect, branches often tangled with one another,
hardly visible in dried material. Fig. 5: 1.
Var. pillansii is recorded from the Orange River along the South African coast
to Mozambique. It forms a zone well above the high water level, lying between A.
decumbens and A. mossianum.
Cape.- — Alexandria: Bushmans River, Archibald 5010 (BOL); Kariega River, Tolken 384. Bathurst:
Kowie River, Acocks 13276; L. Britten 5198 (GRA, J). Bredasdorp: Vogelvlei, Galpin 1 1345b; near
Struisbaai, Levyns 3535 (CT). Caledon: 10 miles west of Riviersc nderend, Tolken 529. Clanwilliam:
Lambert’s Bay, Tolken 342; Leipoldtville, Tolken 332. Hcpefield: Langebaan, Tolken 478; 8 miles
west of Koperfontein, Tolken 473; 474. Humansdorp: Gamtoos River, Tolken 375. Kirkwood:
Addo Bush, Cohen in J 37948 (J). Knysna: Duthie 885a (BOL, STE); Duthie 5198 (J). Malmesbury:
Ysterfontein, Tolken 482; Skrywershoek, Walgate 1065 (BOL). Mossel Bay: Great Brak River,
Bolus in BOL 18835; 18836 (BOL). Namaqualand: Orange River, Pillans 5585 (BOL); Okiep,
Pillans 6247 (BOL); Konfontein Gorge, Giffen 34 (CT); Hondeklipbaai, Pillans in BOL 18221 (BOL,
STE); Wallekraal, Pillans s.n. Peddie: Keiskama River, Tolken 400. Peninsula: Milnerton, Moss
8874; Rugby, Levyns s.n. (CT); Salt River, Moss 3141 (J); 17908 (J, BOL): Lakeside, Weintraub
in J 18367 (J). Piketberg: Elandsbaai, Tolken 540. Port Elizabeth: Swartkops River, Archibald
4900/4a & b (GRA); Rodin 1212; Troughton (J 36913). Riversdale: Stilbaai, Tolken 358; Gouritz
River, Tolken 351. Swellendam: Tolken 524. Vanrhynsdorp: Salt River, Moss 17863 (J); 17908
(J, BOL); Acocks 14186. Vredenburg: St. Helenabaai, Tolken 431; Danger Bay, Tolken 423.
Worcester: Chervonnes, Van Breda 219; Over Hex Station, Tolken 492.
Natal.- — Mtunzini: Umlalazi River, Macnae s.n.
285
Fig. 5.— 1, Arthrocnemum pillansii var. pillansii, flowering spike; la, seed in surface view; lb, seed
in longitudinal section. 2, A. pillansii var. dunense, flowering spike. 3, A. xerophilum, seed in
surface view; 3a, seed in longitudinal section.
The typification of A. hottentoticum Moss is referred to briefly under A. mossianum
(p. 281). A. hottentoticum is described as a plant about 70 cm high, 1*3 m in diameter,
the segments 5-6 mm long, the seeds covered with curved hairs and with spirally
thickened sclereids recorded from the palisade tissue. These characters do not apply
to A. mossianum. The type of A. hottentoticum is cited as Moss 8874 from Milnerton.
Three sheets with this number have been seen (in PRE, J and K) and it is apparent
that a mixture is involved. Of these three, the PRE sheet agrees with the description
and is, therefore, nominated as the lectotype. However, this specimen proves to be
conspecific with the type of A. pillansii ( Moss 8764) and so A. hottentoticum is relegated
to synonymy. The remaining two sheets in J and K represent a previously undescribed
species, now described as A. mossianum (p. 281), typified by Tolken 552. Another point
of interest is that I have not been able to find A. mossianum at Milnerton, nor at any
locality west of Caledon. This throws doubt on the locality cited on the J and K
sheets.
A. namaquense, typified by Moss 17908, cannot be distinguished from A. pillansii.
There are, however, indications that Moss intended the name A. namaquense for the
species now described as A. xerophilum (see p. 287).
A. fruticosa var. densiflora Ung.-Sternb. based on a Drege specimen from Paarden-
eiland can only refer to A. pillansii, because Ungern Sternberg described it as a woody,
frutescent plant.
A form of var. pillansii exhibiting shorter spikes and more cylinderical segments
occurs in the drier areas in the Cape and can easily be confused with A. mossianum
(see Pillans 5585, Levyns 4102, Giffen 34). Although this form, unlike typical A. pillansii,
exhibits many heavily lignified spirally thickened sclereids with spiral striations, strongly
oblique to the lateral walls, no other distinction could be found.
286
The following putative hybrids are recorded (for key, see p. 268) : —
A. natalense x A. pillansii, A. natalense var. affine x A. pillansii,
A. mossianum x A. pillansii, A. perenne x A. pillansii,
A. capense x A. pillansii, A. debumbens x A. pillansii,
A. pillansii X A. terminale.
(b) var. dunense (Moss) Toelken, stat. nov.
A. dunense Moss in Adamson in J. S. Afr. Bot. 20: 14 (1954). Type: S.W. Africa,
Liideritz Bay, Moss 18048 (PRE, holo.!; J.!).
Scarcely woody or woody perennial, normally decumbent, rarely with ascending
branches. Barren segments barrel-shaped, rarely cylindrical, normally 4-8 mm long,
4-6 mm in diameter (rarely 5-15 mm long, 3-4 mm in diameter), becoming cork-like
when old. Selereids in the palisade tissue numerous, spirally thickened with spiral
striations oblique to the lateral walls. Stigmas up to 5 mm long, spreading or reflexed,
usually visible in herbarium material. F:G. 5:2.
Var. dunense is recorded from Liideritz Bay and, possibly, from the Orange River.
It grows on very brackish soil near the sea and is often almost completely covered
by drifting sand. The main flowering period seems to be in January and February,
but flowering specimens are also recorded at various times throughout the year.
Cape.- — Namaqualand: Orange River, Pillans 5587 (BOL, K).
South West Africa. — Liideritz Bay: Martoth 1163; Moss 18048; 11707 (J); Radford Bay,
Dinter 6012 (BOL, STE); Lagoon, Range 1668 (SAM); Tolken 461; 466; Grosse Bucht, Tolken 464;
Diaz Point, Tolken 469; Halifax Bay, Tolken 471.
Typical var. dunense differs from var. pillansii in its shorter and stouter segments,
becoming cork-like with age. However, the specimens Dinter 6012, Tolken 461, 466,
469, 47 1 show a complete range merging into typical var. pillansii. On a few branches
of Tolken 466 the segments are short and thick below, then suddenly become narrower
and much longer on the same axis. This could be due to the unusually heavy rainfall
for the area recorded the previous year. The long stigmas of var. dunense, visible
even in herbarium material, provide the only constant character separating the two
varieties.
6. A. xerophilum Toelken, sp. nov., ab A. pillansii spicis brevioribus, seminibus
denticulatis, habitatis aridis differt.
Arbuscula erecta, lignea ad 50 cm alta, ramis primariis ad 1 ■ 5 cm latis, raro com-
pressis, asperis. Articuli succulenti, ad fines constricti vel cylindrici, obscure carinati,
costis lateralibus, 5-7 mm longi, 4-6 mm lati, glauci vel purpurascentes ; folia sicca
atra et magnopere contrahaerentia ita adeo caulus nudus inter articulos apparens;
foliorum apices non extensi: folia mortua ramis non inhaerentia. Spicae terminales,
et laterales, 10-25 mm longae, non longiores, cylindricae, nodis, ad fines ramorum
dispositae. Cymae triflorae, flore medio floribus lateralibus paulo longiori. Perianthii
tubus lobis 3, truncatus, scutulo laterali, fructu rugescens, semini inhaerens. Antherae 2.
Stigmata 2, papillosa. Semina obovata, fusca, testa coriacea, denticulata praecique
in radicula; endospermum nullum; embryo conduplicatus. Fig. 5: 3. Plate 3.
Type: Cape, Vanrhynsdorp District, Compton 20874 (NBG, holo.; BOL).
Stunted woody shrubs up to 50 cm high with several woody branches 5-15 mm
thick, rarely laterally compressed, rough. Barren segments barrel-shaped or cylindrical,
faintly keeled with a lateral ridge, 5-7 mm long, 4-6 mm in diameter glaucous or
greyish-green fading to bluish-red, becoming dark coloured and shrinking considerably
when dried, so that the bare stem between them becomes visible; leaf apex not spreading;
dead leaves not adhering to the blanches. Selereids in the palisade tissue spirally
287
thickened, about eight times as long as wide. Spikes terminal and lateral, 10-25 mm
long, not longer, cylindrical, bulging, often clustered at the end of branches ; flowering
segments hardly wider than barren segments below. Cymules 3-flowered; flowers
about equal in height, protogynous. Perianth tube with three crenate segments, truncate
with lateral shield, shrivelling when mature, adhering to the seeds. Anthers 2. Stigmas
bifid, spreading, plumose. Seeds compressed obovate, light brown with membranous
testa, denticulate mainly on the radicle; endosperm absent; embryo conduplicate.
A. xerophilum is known only from the dry mountain slopes in the Knersvlakte,
Namaqualand. It is the only species in the whole genus in South Africa which grows
away from the moister marshy areas. The flowering and fruiting period is from
September to December.
Cape. — Namaqualand: Garies, Marloth 5620. Vanrhynsdorp: Pillans in BOL 18786 (BOL);
Knersvlakte, Compton 20874 (BOL, NBG); Acocks 14788; Hutchinson 812.
From a personal letter (in Bolus Herbarium) from Professor Moss to N. S. Pillans
it is known that Moss originally recognized this new species, as he gave a recognizable
description of it in his letter when referring to the specimens Pillans in BOL 18786
and Marloth 5620. However, A. namaquense, as published, falls, both as regards
type and description, within the range of variability of A. pillansii. The two specimens
mentioned above were not cited by Moss (1954).
A remarkable characteristic of this species is the early flowering in September
and October. Even under cultivation in Worcester Karoo Gardens, the plants adhered
to their early flowering period.
7. A. littoreum Moss in Adamson in J. S. Afr. Bot. 14: 38 (1948); J. S. Afr. Bot.
20:9 (1954); Adamson in FI. Cape Penins. 358 (1950). Type: Cape, Camps Bay,
Moss 8775 (PRE, holo.!; J!).
Salicornia fruticosa sensu Harv., Gen. PI. ed. 1:285 (1838), nomen nudum, non
L.; H. Bol. & Wolley-Dod in Trans. S. Afr. Phil. Soc. 14: 311 (1903), nomen nudum;
Wright in FI. Cap. 5, 1:449 (1911), pro majore parte.
Arthrocnemum fruticosum sensu Harv., Gen. PI. ed. 2: 314 (1868), non Moq.
Erect woody shrub up to 1 m high with several erect stems 3-8 cm thick at the
base, seldom laterally compressed; secondary branches erect, fastigiate. Barren
segments cylindrical, with slightly keeled leaf apex, with lateral ridge, 5-8 mm long,
3-5 mm in diameter, glaucous or green fading to yellow, rarely to reddish-brown;
leaf apex spreading when dried; dead leaves not adhering to the branches. Sclereids
in the palisade tissue spirally thickened. Spikes terminal, 20-30 (-100) mm long,
tapering when flowering, cylindrical when fruiting; flowering segments wider than
barren segments below. Cymules 3-flowered ; flowers about equal in size, protogynous.
Perianth tube with three crenate segments, truncate with lateral shield, becoming cork-
like when mature, adhering to the seeds. Anthers 2. Stigmas bifid, plumose, erect,
branches normally tangled with one another. Seeds compressed oblong, light brown,
with membranous testa with recurved hairs, seldom coiled at the apex; endosperm
absent; embryo conduplicate. Fig. 6.
This species occurs in the Cape on rocky beaches below or at spring-tide mark,
often much battered by the sea. The flowering and fruiting period is from September
to May, and often spikes are found releasing ripe seeds below and continuing to flower
above.
288
Fig. 6. — 1 , Arthrocnemum littoreum, flowering spike ; a, fruiting spike; b, seed in surface view; c, seed
in longitudinal section.
Cape. — Alexandria: Kariega River, L. Britten 5237 (GRA). Bredasdorp: Cape Agulhas, Salter
4128 (BOL); Pillans 8101 (BOL). Caledon: Pringle Bay, Tolken 487; Hermanus, L. Guthrie s.n.
(BOL); Danger Point, Pillans 3539 (BOL); Compton 10212 (NBG). Hopefield: Langebaan, Tolken
477. Knysna: Moss 8763 (J); Plettenberg Bay, Smart s.n. Malmesbury: Ysterfontein, Tolken 481.
Mossel Bay: Mom 4341; Burtt Davy s.n. (BOL). Namaqualand: Hondeklipbaai, Pillans s.n. (BOL,
STE). Peninsula: Robben Island, Walgate 489 (NBG); Camps Bay, W alley -Dad 3056 (BOL); Moss
8775; Kommetjie, Simons 88; Schusters Bay, Tolken 205; Olifantsbosch, Tolken 165 (BOL); Buffels
Bay, Tolken 159 (BOL); below Paulsberg, Wolley- Dod 3012 (BOL); Kalkbaai, Pillans 3242; Guthrie
1395 (CT). Port Elizabeth: Lurie in J 18181 (J); Skoenmakerskop, Smith D.107; Bird Island
Schonland 5128 (BOL). Riversdale: Stilbaai, Muir 3149; 166. Vredenburg: Danger Bay, Tolken 428.
A. littoreum is the only species occurring along unsheltered, rocky beaches. Thus
the woody specimen referred to by Harvey can only be this species. Only once, as
Langebaan, were plants of A. littoreum found growing next to those of A. pillansii.
8. A. perenne (Mill.) Moss ex Fourcade, Mem. Bot. Surv. S. Afr. 20: 20 (1941).
Type: from England.
289
Woody and scarcely woody perennial, prostrate or decumbent, forming mats up
to 25 cm high; main branches prostrate and often rooting at the nodes. Barren
segments barrel-shaped to cylindrical, with faintly keeled leaf apex and lateral ridge,
6-12 mm long, 3-5 mm in diameter, dull green to shiny green, fading to yellow or
reddish-brown; leaf apex seldom spreading and the whole specimen becoming dark
grey when dried; dead leaves adhering to the branches but disintegrating on old bushes.
Sclereids in the palisade tissue spirally thickened with striations oblique to the lateral
walls. Spikes terminal, 15-35 (-50) mm long, tapering when flowering, cylindrical
when fruiting; fruiting segments wider than barren segments below, and bulging
about the middle. Cymules 3-flowered, rarely 4- or 5-flowered; central flower slightly
higher than the lateral ones; flowers protogynous. Perianth tube with three crenate
segments, truncate with lateral shield, often becoming cork-like when mature, adhering
to the seeds. Seeds compressed, oblong or obovate, brown, with membranous testa
with hairs recurved or coiled at the apex; endosperm absent; embryo conduplicate,
A. perenne, as construed here, occurs along the coast of South Africa, western
north Africa and south-western Europe. It usually forms a separate zone from below
the high-water mark and just above it. In South Africa the flowering and fruiting
period is from January to June.
Two varieties are recognized in South Africa :
Branches (not including the succulent leaf sheath) rarely exceeding 3 mm in diameter, often rooting
at the nodes; succulent segments green fading to reddish-brown (a) var. perenne
Branches (not including the succulent leaf sheath) up to 8 mm in diameter, rarely rooting at the
nodes; succulent segments green, fading to yellow (b) var. lignositm
(a) var. perenne.
Salicornia perennis Mill., Gard. Diet. ed. 8 (1768); Moss in J. Bot. Lond. 49: 179
(1911). S. radicans Smith, Eng. Bot. 1. 1691 (1807). S. fruticosa var. radicans (Smith)
Gr. & Godr., FI. Fr. 3:28 (1855). var. paardeneilandica Ung.-Sternb., Vers.
Syst. Salic. 59 (1866); Wright in FI. Cap. 5, 1:449 (1912). Type: Cape, Paarden-
eiland, “ Drege Sammlung, n. 221 ” (LE, holo.). 5. perennis var. radicans (Smith)
Moss in New Phytol. 11:408 (1912); Moss & Salisbury in Camb. Brit. FI. 2: 188,
fig. 195 (1914). V. arabica var. paardeneilandica (Ung.-Sternb.) Chev. in Rev. Bot.
appl. 2: 33, (1922).
Arthrocnemum perenne (Mill.) Moss ex Fourcade Mem. Bot. Surv. S. Afr. 20 : 20
(1941). Type: from England. —var. radicans (Smith) Moss ex Fourcade, Mem.
Bot. Surv. S. Afr. 20 : 20 (1941); Moss in Adamson in J. S. Afr. Bot. 14 : 40 (1948);
J. S. Afr. Bot. 20 : 11 (1954); Adamson in FI. Cape Penins. 358 (1950). A. fruticosum
var. radicans (Smith) Moq., Chenop. Monogr. Enum. 112 (1840); in DC., Prodr. 13,
2 : 151 (1849). A. variiflorum Moss in Adamson in J. S. Afr. Bot. 14 : 39 (1948);
J. S. Afr. Bot. 20 : 17 (1954); Adamson in FI. Cape Penins. 358 (1958). Type: Cape,
Milnerton, Moss 8765 (PRE, holo.!).
Scarcely woody perennial, with bare main branches seldom thicker than 3 mm,
not conspicuously swollen at the nodes, often rooting at the nodes. Barren segments
dull to shiny green often with a pinkish tinge between the segments, fading to reddish-
brown. Seeds oblong (1-2 mm long) densely hairy; hairs usually coiled at the apex
Fig. 7 : 1.
This variety grows in the lowest zone of muddy maritime salt marshes, and has
been mainly recorded from the eastern coast of South Africa.
290
Cape. — Alexandria: Bushmans River, Dyer 3368; Archibald 5009 (BOL). Bathurst: Kowie River,
L. Britten 5202 (GRA, J, K); 5206 (GRA, J). East London: Nahoon River, Tolken 403. Elliotdale:
Bashee River, Tolken 404. Hopefield: Langebaan, Isaac in BOL 23472 (BOL, PRE). Humansdorp:
Gamtoos River, Tdlken 377. Kentani: Pegler 648. Knysna: Moss 9937 (J); 8770 (J); Duthie 894
(BOL, J); 893 (BOL, STE); Fourcade 1997b (BOL); Bitou River, Tdlken 368. Komga: Great Kei
River, Flanagan 1119; M. Moss in J 27425 (J). Mossel Bay: Great Brak River, Bolus in BOL 18837
(BOL); Tdlken 361. Peddie: Keiskama River, Comins 1612; Wager in PRE 29115. Peninsula:
Milnerton, Moss 8756 (J); 11330 (J); Rugby, Moss 8776 (GRA, J); Salt River, Moss 3130 (GRA);
5259 (J). Port Elizabeth: Swartkopsrivier, Rodin 1213 (BOL, K); Archibald 4927 (BOL); 5815
(BOL, GRA). Riversdale: Stilbaai, Muir 3209 (J); 46456 (J); Gouritz River, Tdlken 352. Umtata:
Umtata River, West 3171.
Natal. — Durban: Medley Wood 418 (SAM); Umbilo, Moss 3139; 4335 (J); Salisbury Island,
Forbes 62 (NH); 324; Moss 18397 (J); Galpin 12132. Mtunzini: Umlalazi Estuary, Macnae s.n.
(PRE).
South West Africa. — Luderitz Bay: Orange River, Merxmuller & Giess 2290.
According to the International Code of Botanical Nomenclature (1961), Article 24,
the var. radicans, having been described as the typical form, has been changed to var.
perenne.
Fig. 7.— 1, Arthrocnemum perenne var. perenne, flowering spike; la, fruiting spike; lb, seed in surface
view. 2, var. Iignosum, seed in surface view: 2a, seed in longitudinal section.
291
From field observation it appears that A. variiflorum is merely a vigorous form of
A. perenne. Furthermore, the type seems to be an unfortunate choice, as it exhibits
spikes hardly longer than 2 cm instead of 3-3-5 cm long, and not a single cymule with
more than three flowers could be found. The growth and pollen suggest that the
specimen is a hybrid. A. variiflorum is regarded as a synonym of A. perenne var.
perenne.
The following putative hybrids are recorded (for key, see p. 268):—
A. nata/ense x A. perenne, A. capense x A. perenne,
A. perenne X A. pillansii, A. decumbens X A. perenne.
(b) var. lignosum (Woods) Moss in New Phytol. 11 : 409 (1912); Moss & Salisbury
in Camb. Brit. FI. 2 : 189 (1914); Moss in Adamson in J. S. Afr. Bot. 14 : 40 (1948).
Type: from England.
Salicornia lignosa Woods in Bot. Gaz. (London) 3 : 31 (1851). S. natalensis sensu
Marloth, FI. S. Afr. 1 : 184, fig. 84b (1913), non Bunge ex Ung.-Sternb.
Woody perennial, prostrate or decumbent, rarely rooting at the nodes, with bare
main branches 3-5 (-10) mm thick, often conspicuously swollen at the nodes. Barren
segments dull or shiny green, fading to yellow, never reddish-brown. Seeds compressed
obovate, 2-3 mm long, up to twice the size of those of var. perenne, sparcely hairy
with hairs not coiled at the apex. Fig. 7 : 2.
The var. lignosum has so far been found only at Knysna and at Liideritz Bay in
Southern Africa. It grows in the lowest zone in gravelly maritime salt marshes,
preferably in sheltered localities.
Cape. — Knysna: Moss 8875 (J); Tolken 365.
South West Africa. — Liideritz: Moss 1175 (J); 18123 (J); De Winter & Hardy 7936; Lagoon,
Kinges 2034; Range 1750 (SAM); Diaz Point, Tolken 470; Halifax Bay, Tolken All.
This variety is very similar to A. littoreum in its colour and habitat, but differs
from that species in its decumbent or prostrate habit.
Seeds collected at Liideritz Bay indicated that the seed characters used by Moss
to differentiate the two varieties are not completely reliable. The two entities are
kept separate because the var. lignosum never produces the red segment colour. All
plants of A. perenne at Liideritz Bay must be referred to the var. lignosum, although
most of the specimens grow on a sandy substrate and not on gravel. Also specimens
of both varieties were grown under similar conditions at the Botanical Research Institute,
Pretoria, but they never changed their respective colours.
Marloth’s S. natalensis is referred to this variety in view of its recorded occurrence
at lower levels at Liideritz Bay.
9. A. capense Moss in Adamson in J. S. Afr. Bot. 14 : 38 (1948); J. S. Afr. Bot.
20 : 13 (1954); Adamson in FI. Cape Penins. 358 (1950). Type: Cape, Milnerton,
Moss 11621 (PRE, holo.!).
Scarcely woody or woody perennial, decumbent or prostrate, forming dense mats
or irregular rosettes 40-80 cm in diameter and up to 25 cm high; main branches
prostrate, often rooting at the nodes, with terminal and lateral branches erect, but
young branches often prostrate. Barren segments cylindrical, with keeled leaf apex
and pronounced lateral ridge, 10-16 mm long, 2 (-3) mm in diameter, glaucous or
dull green fading to reddish-brown; leaf apex of only young leaves spreading when
dried ; dead leaves adhering to the branches. Sclereids in the palisade tissue uniformly
thickened, reaching from the water tissue to the epidermis. Spikes terminal and lateral.
922
15-35 mm long, seldom longer, tapering when flowering, cylindrical when fruiting,
usually solitary but also clustered at the end of branches ; flowering segments not wider
than the barren segments below. Cymules 3-flowered; central flower hardly larger
than the lateral ones ; flowers protogynous. Perianth tube with three crenate segments
truncate with lateral shield, shrinking when dried, adhering to the spike and the seeds.
Anthers 2. Stigmas bifid or trifid, plumose, reflexed. Seeds compressed oblong,
light brown, with membranous testa with few hairs recurved not coiled at the apex;
endosperm absent; embryo conduplicate. Fig. 8.
Fig. 8. — 1, Arthrocnemum capense, flowering spike: a, fruiting spike; b, seed in surface view; c, seed
in longitudinal section.
This species was previously recorded from the vicinity of Cape Town only, but
has since also been found at isolated localities at Hermanus and Port Elizabeth. It
grows usually in sandy, hardly saline soils, forming a zone above the A. pillansii zone.
The flowering and fruiting period is from October to April.
Cape.— Caledon: Mosselrivier, L. Guthrie s.n. (BOL); Uilenkraalrivier, Tolken 500; 503. Malmes-
bury: Mamre Road, Compton 9812 (NBG). Peninsula: Milnerton, Moss 11621; Rugby, Tolken 326;
Uitvlugt, Wolley-Dod 1480 (BOL); Raapenburgvlei, Wolley-Dod 2690 (BOL); Isoetesvlei, Tolken
304; Rondevlei, Andraea 237 (STE); Noordhoek, Moss 3144; Pillans 3222 (BOL); Kommetjie,
Moss 8883a. Port Elizabeth: Blue Hole, Archibald 5818 (GRA). Stellenbosch: Kuilsrivier, Tolken 301.
Although this species produces few or no flowering spikes in dry seasons, it can be
easily recognized by its thin, but tough branches with marked lateral ridge on the
segments.
293
The following putative hybrids are recorded (for key, see p. 268) : —
A. capense X A. natalense, A. capense X A. pillansii,
A. capense X A. decumbens, A. capense x A. perenne.
10. A. decumbens Toelken, sp. nov., ab Salicornia blackiana floribus cymae omnibus
in linea transversa dispositis, spicis saepe longioribus 3 cm; ab Salicornia australis
seminibus obovatis, denticulatis, differt.
A. heptiflorum sensu Moss in Adamson in J. S. Afr. Bot. 20 : 18 (1954), partly, excl.
type. A. australasicum sensu Moss in Adamson in J. S. Afr. Bot. 20 : 19 (1954), non
(Moq.) Moss.
Suffrutex decumbens, vix ligneus, ad 40 cm altus; rami primarii ad 15 mm lati,
non compressi; rami secundarii erecti, ramis lateralibus paulo. Articuli plerumque
cylindrici, carinati, costis lateralibus, 5-15 mm longi, 3-6 mm lati, glauci vel purpura-
scentes; foliorum siccorum apices, novellorum extensi; folia mortua saepe cortica,
ramis inhaerentia. Spicae 15-40 (-50) mm longae raro ad apices fastigiatae, plerumque
cylindricae, sub articulis latissimae. Cymae (4-) 5-7 (-9)-floribus flore medio prope
instar florum lateralium. Perianthii tubus lobis, 4, 3, 2, saepe rima, truncatus scutulo
laterali, semini non inhaerens. Antherae 2. Stigmata 2, raro 3, undique papillosa, reflexa.
Semina obovata, raro oblonga, fusca testa coriacea, papillosa; endospermum nullum;
embryo conduplicatus. Plate 4. Fig. 9.
Type: Cape: Mossel Bay District, Great Brak River, Tolken 362 (PRE, holo.).
Scarcely woody to sometimes woody perennial, decumbent, up to 50 cm high
with main branches up to 15 mm thick not laterally compressed; secondary branches
usually straight erect and with few lateral branches. Barren segments cylindrical to
barrel-shaped, keeled with lateral ridge, 5-15 mm long, 3-6 mm in diameter, glaucous
fading bluish-red, leaf apex of only young leaves spreading when dried; dead leaves
often becoming cork-like, adhering to the branches. Sclereids in the palisade tissue
uniformly thickened next to the water tissue, not reaching to the epidermis. Spikes
15-30 (-50) mm long, rarely tapering, normally cylindrical; flowering segments usually
much wider than barren segments below. Cymules (4-) 5-7 (-9)-flowered ; flowers
about equal in size, protogynous. Perianth tube with 4, 3 or 2 crenate segments, often
only a slit, truncate with lateral shield, becoming cork-like when mature, not adhering
to the seeds. Anthers 2. Stigmas bifid, seldom trifid, plumose, reflexed. Seeds com-
pressed obovate or ovate, light brown with membranous testa with pointed or blunt
papillae; endosperm absent; embryo conduplicate.
A. decumbens occurs along the east coast of South Africa from Bredasdorp to
Lourenco Marques. The gap in the distribution from the Kei River to Lourenco
Marques is significant, and is attributed to the higher rainfall which does not provide
extensive, seldom flooded salt marches. The species usually occupies a separate zone
below the zone of A. pillansii, with which it appears to be in strong competition. The
flowering and fruiting period is from January to June.
Cape. — Bathurst: Kowie River, L. Britten 5199 (GRA); 5025 (GRA, J); M. Moss in J 20844 (J)-
Bredasdorp: De Mond, Taylor 105 (BOL); Seekoeivlei, Tolken 511; De Hoopvlei, Tolken 518
Caledon: 10 miles west of Riviersonderend, Tolken 416; 527. East London: Gulu River, Galpin
9590. Humansdorp: Gamtoos River, Tolken 378. Komga: Great Kei River, M. Moss in J 27425
(J). Mossel Bay: Great Brak River, Moss 4344; 4345 (J): Bolus in BOL 18805 (BOL); Tolken 362.
Peddie: Keiskama River, Tolken 399. Port Elizabeth: Swartkopsrivier, Archibald 4891 (GRA, BOL);
5819 (BOL); Zeyher 3615 (GRA); Paterson 497 (GRA); Cape Recief II, Zeyher (BOL, GRA, K,
SAM, STE). Swellendam: Tolken 413; 523.
Mozambique.- — Lourenco Marques: Moss 3136 (J); 4335 (J); Inhaca Island Moss 16996 (J); M.
Moss in J 27630; in J 27632 (BOL).
294
Fig. 9.— 1, Arthrocnemum decumbens, flowering spike; a, fruiting spike; b, seed in surface view;
c, seed in longitudinal section.
Although the new species here described was linked by Moss with various Australian
species, it seems to be distinct. Moss applied the following two names to the South
African plants which I now regard as belonging to one species, A. decumbens.
1. A. heptiflorum Moss, based on one of the syntypes of Salicornia australis Soland.
ex Benth. (see discussion below) and under which he included Salicornia pachystachya
Black (1921), non Ung.-Sternb.
2. A. australasicum (Moq.) Moss, based on Halocnemum australasicum, an Aus-
tralian species.
Solander’s S. australis was a manuscript name, and Forster, Florulae Insularum
Australium Prodromus (1786), also published just the name. The first description
of this species appears to be that of Bentham in 1870. This name is antedated by
S. quinqueflora Ung.-Sternb. (1866). Moss (1954) seems to have been unaware of
Bentham’s description as he treated S. australis as a nomen nudum, and described
A. heptiflorum based on one of the syntypes of S. australis ( Forster , New Zealand).
Thus Moss incorrectly included the South African specimens under A. heptiflorum, a
synonym of S. quinqueflora.
295
The following table, drawn up from herbarium and preserved material investigated,
indicates the characterstics of the taxa, probably species, as seen by the present author.
Black (1921) described S. pachystachya, non Ung.-Sternb. (1866), for which Ulbrich
(1934) then proposed the new name S. blackiana. This species is very similar to A.
decumbens, but differs from the latter in often producing additional flowers in a second
row below the central flower, and having spirally thickened sclereids in the palisade
tissue.
From Moquin’s (1840) description of Halocnemum australasicum it is not clear
to which species he referred. In his more detailed description (Moquin, 1849), where
he still referes to the reputed R. Brown specimen only, it can be recognized that he
refers to the more recent A. lylei as already pointed out by Black (1921). The phrase
“ sepala ovata, tenuia, membranacea ” gives the impression of a three-lobed perianth,
membranous at the apex, as it is found in A. lylei (Black, 1919; plate 34). It also
296
reflects Moquin’s motive for putting his species into the genus Halocnemum, the flowers
of which he described as “calyx 3-sepalus; sepalis squamaeformibus, adpressis ”,
in contrast with the perianth of Arthrocnemum described as “ calyx ventricosu^ . . .
apice truncatus vel 3, 4, 5 dentatus The subsequent misunderstanding by Moquin
and Betham regarding the identity of H. australasicum and the R. Brown specimen
seems to be based on different specimens seen by the different authors. According
to Stearn (1960) it is unlikely that Moquin had seen a Brown specimen in Herb. Paris,
as duplicates of Brown’s collections were distributed only after his death in 1851. This
is much later than Moquin’s publications. Perhaps this might explain the query after
Moquin’s (1849) citation of the Brown specimen.
A. australasicum sensu Moss seems to refer to a mixture consisting of specimens
(a) of a form of A. decumbens and (b) the putative hybrid A. decumbens X A. perenne.
In Moss’s description of A. australasicum, the mention of “ green shoots ” and the lesser
number of flowers per cymule applies to both these elements, whereas “ few large
spirally thickened cells ” indicate the hybrid, which in herbarium material is almost
indistinguishable from this form of A. decumbens. This form of A. decumbens, where
the sclereids do not normally develop but in which the typical uniformly thickened
sclereids are sometimes formed, combined with the stomata being in line with the
epidermis, exhibits all transitional forms between it and the typical A. decumbens.
It is thus regarded as an ecological form of A. decumbens, as it exhibits all the symptoms
of specimens grown under humid conditions.
In A. decumbens, papillose seeds are usually found, though occasionally seeds with
stout, straight hairs may occur, but these hairs are never hooked at the apex.
It is remarkable that the several flowers below the central flower of a cymule of
S. blackiana have never been described before. This is unique in the genus Arthroc-
nemum.
The following putative hybrids are recorded (for key, see p. 268) : — -
A. decumbens x A. natalense, A. decumbens x A. pillansii,
A. decumbens X A. perenne, A. capense X A. decumbens.
2. SALICORNIA
Salicornia L. Sp. PI. 3 (1753); Moq., Chenop. Monogr. Enum. 113 (1840);
Adamson in FI. Cape Penins. 358 (1950); Moss in Adamson in J. S. Afr. Bot. 20 : 20
(1954). Type species: S. europaea L. S. subgen. Eusalicornia Gr. & Godr., FI. Fr.
3 : 27 (1855); S. subgen. Vulgata Ung.-Sternb., Vers. Syst. Salic. 43 (1866); S. sect.
Annuae Duval-Jouve in Bull. Soc. Bot. Fr. 15 : 170 (1866); Moss in J. Bot. Lond.
49 : 178 (1911); Ulbrich in Pflanzenfam. ed. 2, 16c : 550 (1934). S. sect. Salicorniella
Moss & Salisbury in Camb. Brit. FI. 2 : 187 (1914).
Annual shrublets, erect, rarely decumbent; branches not rooting at the nodes
and all branches ending in spikes. Barren segments usually drying out at anthesis or
just before maturity of the seeds. Cymules 3-flowered; lateral flowers not completely
separated by the central flower, or only a central flower with rarely 1 or 2 much smaller
lateral flowers present. Flowers protandrous, sometimes overlapping. Stigma tufted,
hardly visible. Seeds with membranous testa covered with slender hairs, coiled at the
tip; endosperm absent; embryo conduplicate.
Species of this genus occur in salt marches throughout the world, except in Australia.
297
Key to the Species of Salicornia
Flowers 1, rarely 2 or 3 per cymule 3. S. uniflora
Flowers consistently 3 per cymule:
Spikes becoming thicker, up to 8 mm wide when fruiting: perianth cork-like when mature;
apex of central flower strongly obtuse 1. S. pachystachya
Spikes not becoming thicker, 2-3 mm wide when fruiting: perianth shrivelling when mature;
apex of central flower acute, rarely becoming obtuse when fruiting 2. S. meyerana
1. S. pachystachya Bunge ex Ung.-Sternb., Vers. Syst. Salic. 51 (1866); in Atti
Congr. Int. Bot. Firenze 1874 : 278, fig. 18A, B (1876); Brenan in FI. Trop. E. Afr.
Chenopodiaceae : 21 (1954). Type: Madagascar, Bay of Suarez, Boivin 2370 (LE,
holo., P!).
S. perrieri Chev. in Rev. Bot. appl. 2 : 748 (1922); Moss in Adamson in J. S. Afr.
Bot. 20 : 22 (1954); Cavaco in FI. Madagascar, Chenopodiaceae : 9 (1954). Type:
Madagascar, Ambongo Province: Soalala, De la Bathie 1580 (P, holo.!).
Fig. 10. — 1, Salicornia meyerana, fruiting spike; la, fruiting spike with lateral proliferation; lb, seed
in surface view; lc, seed in longitudinal section showing a clearly differentiated stele. 2, S. uniflora,
fruiting spike; 2a, seed in surface view. 3, S. pachystachya, fruiting spike; 3a, seed in surface
view.
298
Erect woody shrublets up to 40 cm high; main branches up to 10 mm in diameter,
erect with shorter lateral branches; apices of lateral branches not higher than the
apex of the main stem. Barren segments cylindrical, faintly keeled with lateral ridge,
5-10 mm long, 2-4 mm in diameter, glaucous or green fading to brownish-red; leaf
apex hardly spreading when dried. Sclereids in the palisade tissue spirally thickened.
Spikes terminal and lateral, 10-25 (-80) mm long, becoming longer and thicker when
fruiting, up to 8 mm thick, tapering when flowering and fruiting or sometimes cylindrical
when fruiting. Cymules 3-flowered; apex of the central flower strongly obtuse; flowers
protandrous. Perianth tube with three crenate segments, truncate with lateral shield,
becoming cork-like when fruiting. Anthers 2, 0-5-1 mm long. Stigma tufted often
bifid or trifid. Seeds compressed oblong, brown with membranous testa with few
long hairs coiled at the apex; endosperm absent; embryo conduplicate. Fig. 10 : 3.
This species occurs in tropical East Africa and Madagascar. In South Africa it
is found only in northern Natal. The flowering period is insufficiently known, but the
few specimens seen indicate a flowering time from June to October and fruiting from
December to January.
Natal. — Hlabisa: False Bay Park. Ward 4220. Ingwavuma: Nyamiti Pan, Tinley 521, Sheet I
and II.
Mozambique. — Lourenco Marques: Roger & Moss in J 4340 (J); Moss 3137 (J); Inhaca Island,
M. Moss in J 27633; J 28245; Matola, Macnae 2; Moss 11742 (J); Costa da Sol, Macnae 3.
Two forms of this species were observed in the field and these are clearly shown
in Sheet I and 11 of Tinley 521. The one form which grows apparently always on drier
soils, has an erect growth with the apex of the lateral branches not higher than the apex
of the main branches. All branches end in spikes, which are 20-30 (-50) mm long.
The second form is much more branched, almost decumbent and the apex of the lateral
branches are often higher than the apex of the main branch. A few branches at the
end of the main branch begin flowering in June while the lower branches still grow
vegetatively. Flowers are born on the latter type of branches later or not at all. The
length of the spikes of this second form are 20-50 (-80) mm long.
The specimen Ward 3267 from Lake St. Lucia, which shows the typical growth
of the second form, is thought to be a hybrid because no seeds develop and anthers
are often aborted. Unfortunately no anthers with mature pollen were found in order
to establish whether pollen of two sizes is found in putative hybrids of Salicornia,
as is the case with Arthrocnemum. The problem is, however, that no other species of
Salicornia has been recorded from that area and no indications of any species of
Arthrocnemum being involved, could be found.
The typical spikes of S. pachystachya can be recognized from the illustrations in
Ungern Sternberg (1876), 1. 18, fig. A, B. S. brachiata Roxb., a species from India
and the East Indies, is very similar to S. pachystachya, the differences in the shape of
the perianth, as described by Ungern Sternberg, seem to be rather insignificant.
Chevalier’s (1922) claim that the type specimen of S. pachystachya has seeds with
endosperm is thought to be based on a specimen of A. indicum ( Perville 661 !). I therefore
agree with Moss (1954) that A. pachystachyum sensu Chevalier is conspecific with
A. indicum, while S. perrier i Chev., being described as the second species occurring
in Madagascar, is a synonym of S. pachystachya.
2. S. meyerana Moss in Adamson in J. S. Afr. Bot. 14 : 36 (1948); J. S. Afr.
Bot. 20 : 20 (1954); Adamson in FI. Cape Penins. 357 (1950). Type: Cape, Paarden-
eiland, Levyns in J 16969 (PRE, holo.l; J!).
299
Scarcely woody shrublet, erect rarely decumbent, up to 40 cm high; erect main
branches up to 10 mm in diameter at the base, often twisted; secondary branches
spreading and longer than the central stem. Barren segments cylindrical, faintly keeled
or not keeled at the leaf apex, with lateral ridge, 5-15 mm long 2-3 mm in diameter,
glaucous or green fading to deep bluish-red, then often shiny; leaf apex of young
leaves slightly spreading when dried. Sclereids in the palisade tissue spirally thickened.
Spikes terminal and lateral, (5-) 10-30 (-120) mm long, tapering when flowering,
cylindrical when fruiting. Cymules 3-flowered; central flower acute or nearly so at
its posterior apex; flowers protandrous. Perianth tube with three crenate segments,
truncate with lateral shield. Anthers 2, 0-5-1 mm long. Stigmas tufted. Seeds com-
pressed, oblong, 1-2 mm long, brown, with membranous testa covered with long seed
hairs coiled at the apex; endosperm absent; embryo conduplicate. Fig. 10 : 1.
This species is recorded from all along the South African coast from Vanrhynsdorp
to Durban. It is a pioneer plant mainly in maritime salt marshes, rarely in inland salt
marshes, and rarely occurs in the lowest zone. The flowering and fruiting period is
from February to May.
Cape. — Bathurst: Kowie River, L. Britten 2096 (GRA); 2862 (GRA); Great Fish River, Tolken
397. Bredasdorp: Vogelvlei, Tolken 507; near Struisbaai, Levyns 3536 (CT). Caledon: Hermanus
Lagoon, Jordaan s.n. (STE); Tolken 499; Uilenkraalrivier, Tolken 504. Clanwilliam: Lambert’s
Bay, Tolken 340. Humansdorp: Kabeljourivier, Tolken 273a & b. Knysna: Duthie 884 (BOL);
902 (BOL, STE); Duthie sub Moss 8771 (J). Malmesbury: Ysterfontein, Tolken 485a & b. Mossel
Bay: Little Brak River, Moss 4346 (J); Great Brak River, Bolus in BOL 18807 (BOL). Peninsula:
Milnerton, Moss 3131 (J); 8826 (J); Rugby, Tolken 329; 330; Paardeneiland, Levyns in J 16969;
Salt River, Schlechter 254 (GRA); Moss 8755 (GRA, J); 8758 (J). Port Elizabeth: Swartkopsrivier,
L. Britten 3016 (GRA); Troughton in J 36909 (J). Vanrhynsdorp: Salt River, Tolken 535. Vredenburg:
Danger Bay, Tolken 424; 480.
Natal. — Durban: Congella, Moss 4336 (J); 5252.
This complex species is insufficiently known and some specimens differ very little
from the European species but never entirely agree with the descriptions of Ball (1964).
A form which is different from the typical S. meyerana has been observed in the field,
often near disturbed areas. This form has very long spikes, 4-10 cm long, with wider
flowering segments, more than two branches at the node and the central flower more
acute and bigger in comparison with the lateral flowers. Usually only single specimens
are seen, but at Danger Bay (Vredenburg district) a whole population was found.
The following specimens are referred to this form: Moss 5252; Tolken 273; 424;
480; 485a. Tolken 485a shows somewhat intermediate characters. When plants of
this form were grown from seeds collected at Danger Bay, together with those of the
typical form under uniform conditions, they were indistinguishable from one another.
This form is very similar to the European species, S. dolichostachya, but differs in
always producing spirally thickened sclereids in the palisade tissue.
3. S. uniflora Toelken, sp. nov., ab 5. meyerana cymis unifloris, raro bifloris, seminis
ad 2 mm longis differt.
Suffrutex erecta ad 20 cm alta, ad 15 cm lata. Articuli cylindrici carinati, costis
lateralibus, 4-8 mm longi, 2-3 (-4) mm lati, obscure virides vel purpurascentes ; folia
sicca vix extensi: folia mortua ramis inhaerentia. Spicae terminates et laterales, (8-)
10-20 (-30) mm longae, ad apices fastigiatae flore et fructu. Cymae uniflorae, raro
biflorae, flore medio laterali longissimo. Perianthii tubus lobis 3, truncatus scutulo
laterali, semini nonnunquam inhaerens. Antherae 2. Stigma cristata. Semina oblonga,
fusca 2 mm longa, testa coriacea, pilosa, pilis adpressis apice circinatus; endospermum
nullum; embryo conduplicatus. Plate 5. Fig. 10 : 2.
Type: South West Africa, Luderitz Bay, Tolken 460 (PRE, holo.).
300
Herbaceous annual, erect, up to 20 cm high and 15 cm in diameter; lateral branches
fastigiate but not arising at an acute angle to the main branch. Barren segments
cylindrical to barrel-shaped with keeled leaf apices and lateral ridge, 4-8 mm long,
2-3 (4) mm in diameter, glaucous green fading to bluish-red; leaf apices scarcely
spreading when dried. Sclereids in the palisade tissue spirally thickened. Spikes
terminal and lateral, (8-) 10-20 (-30) mm long, tapering when flowering and fruiting.
Cymules 1-flowered, rarely 2, 3-flowered; central flower much larger than the lateral
flowers and rounded at the apex ; flowers protandrous. Perianth tube with three crenate
segements, truncate with lateral shield, sometimes adhering to seeds. Anthers 2,
0-5-0 -8 mm long. Stigma tufted. Seeds compressed oblong, brown, 2-3 mm long,
with membranous testa covered with long hairs, adpressed, coiled at the apex; endos-
perm absent: embryo conduplicate.
This species is so far recorded only from Luderitz Bay and Malmesbury district.
It grows at Luderitz Bay in maritime salt marshes in open spaces above the A. perenne
zone. S. uniflora, judging from observations in situ at Luderitz Bay, flowers in December
and January and fruits in February and March.
Cape.- — Malmesbury: near Darling, Walgate 1070 (BOL).
South West Africa. — Luderitz Bay: Dinter 6016 (BOL, STE); Lagoon, Tolken 460; 467.
References
Aellen, P. (1961). In Hegi, Illustrierte Flora von Mitteleuropa 3, 2 : 723-734. Miin-
chen: C. Hauser Verlag.
Allan, H. (1961). Flora of New Zealand 1 : 232. Wellington: R. E. Owen.
Ball, P. W. (1964). A taxonomic review of Salicornia in Europe. Repert. nov. Spec.
Regn. veg. 69, 1 : 1-8.
Ball, P. W. & Tutin, T. G. (1959). Notes on the annual species of Salicornia in
Britain. Watsonia 4 : 193-205.
Baumgartel, O. (1917). Die Anatomie der Gattung Arthrocnemum Moq. S.B.
Akad. Wiss. Wien Abt. 1, 126 : 41-73.
Bentham, G. (1870). Flora Australiensis 5 : 205. London: L. Reeve.
Bentham, G. & Hooker, J. D. (1880). Genera Plantarum 3, 1 : 65, 66. London:
L. Reeve.
Black, J. M. (1919). A revision of the Salicorniae. Trans. Roy. Soc. S. Austr. 43 :
355-367.
Black, J. M. (1921). Additions to the Flora of South Australia No. 19. Trans. Roy.
Soc. S. Austr. 45 : 8-10.
Chevalier, A. (1922). Le Salicornes, leur Biologie et leur Distribution geographic.
Rev. Bot. appl. 2 : 748-750.
Cooke, F. W. (1912). Observations on Salicornia australis. Trans. N.Z. Inst. 44 :
349-362.
Dalby, D. H. (1962). Chromosome number, morphology and breeding behaviour
in the British Salicorniae. Watsonia 5 : 150-162.
Day, J. H. (1952). The ecology of South African estuaries I. Trans. Roy. Soc. S.
Afr. 33 : 53-91.
301
De Fraine, E. (1911). The anatomy of the genus Salicornia. J. Linn. Soc. ( Bot .)
41 : 317-339.
Esau, K. (1960). Anatomy of Seed Plants. New York: J. Wiley.
Ferguson, I. K. (1964). Notes on the stigma morphology and flowering behaviour
in British Salicorniae. Watsonia 6 : 15-27.
Goodspeed, T. H. (1954). The Genus Nicotiana. 209-211. Waltham, Mass: Chronica
Botanica Co.
Hiern, W. P. (1900). Catalogue of the African Plants collected by F. Welwitsch 1,
4 : 899. London: Longmans.
Konig, D. (1960). Beitrage zur Kenntnis der deutschen Salicornien. Sonderdruck
aus Mitteilungen der floristisch soziologischen Arbeitsgemeinschaft, 8 : 5-58.
McCann, C. (1952). Notes on the genus Salicornia L. (Chenopodiaceae). Bombay
Nat. Hist. Soc. 50 : 872-877.
Metcalf, C. R. & Chalk, L. (1950). The Anatomy of Dicotyledons 2 : 1074-1084.
Oxford: Clarendon Press.
Moquin-Tandon, C. H. B. A. (1840). Chenopodiarum Monographica Enumeratio.
Paris.
Moquin-Tandon, C. H. B. A. (1849). In De Candolle, Prodromus Systematis Naturalis
Regni Vegetabilis 13, 2 : 149-152. Paris.
Moss, C. E. (1911). Some species of Salicornia. J. Bot., Loud. 49 : 177-185.
Moss, C. E. (1912). The genus Salicornia in Denmark. J. Bot., Lond. 50 : 94-95.
Moss, C. E. (1912b). International phytogeographical excursion in the British Isles.
New Phytol. 9 : 409-411.
Moss, C. E. (1948). Plantae novae Africanae. /. S. Afr. Bot. 14 : 36-40.
Moss, C. E. (1954). The species of Arthrocnemum and Salicornia in Southern Africa.
J. S. Afr. Bot. 20 : 1-22.
Moss, C. E. & Salisbury, E. (1914). The Cambridge British Flora 2 : 187-196.
London: Cambridge University Press.
Nannfeldt, J. A. (1955). Nagot om Slakett Salicornia: Sverige. Svensk. Bot. Tidskr.
49:97-110.
Paulsen, O. (1918). In Ostenfeld, Contributions to West Australian Botany II.
Dansk. Bot. Ark. 2, 8 : 61-65.
Stearn, W. T. (1960). Introduction to Facsimile Reprint of Brown, Prodromus Florae
Novae Hollandiae.
Turesson, G. (1925). Studies in the genus Atriplex L. Lunds Universitets Arsskrift.
N.F. 21, 4 : 1-15.
Ulbrich, E. (1934). In Engler & Prantl, Naturlichen Pflanzenfamilien (2nd edn.).
16c : 548-552. Leipzig: W. Engelman Yerlag.
Ungern Sternberg, F. (1866). Versuch einer Systematik der Salicornien. Dorpat.
Volkens, G. (1894). In Engler & Prantl, Naturlichen Pflanzenfamilien 3, la : 76-77 .
Leipzig: W. Engelman Verlag.
Wright, C. H. (1912). In Thiselton-Dyer, Flora Capensis 5, 1 : 449-450. London:
L. Reeve.
302
Index
Arthrocnemum Moq., 273
affine Moss, 280
africanum Moss, 277
anstralasicuni (Moq.) Moss, 293
capense Moss, 291
decumbens Toelken, 293
dunense Moss, 286
fruticosum (L.) Moq., 273
var. radicans (Smith) Moq., 289
heptiflorum Moss, 293
hottentoticum Moss, 28 1 , 284
indicum ( Willd .) Moq., 276
littoreum Moss, 287
lylei ( Ewart & White ) Black, 295
macrostachyum sensu Hiern, non Moris & Del-
port, 276
mossianum Tolken, 28 1
namaquense Moss, 284, 287
natalense ( Ung.-Sternb .) Moss, 277
var. affine (Moss) Toelken, 280
var. natalense, 279
pachystaehyum (Ung.-Sternb.) Chev., 276, 298
perenne (Mill.) Moss ex Fourcade, 288
var. radicans (Smith) Moss, 289
var. lignosum ( Woods) Moss, 29 1
var. perenne, 289
pillansii Moss, 283
var. dunense (Moss) Toelken, 286
var. pillansii, 284
terminale Toelken, 282
variiflorum Moss, 289
xerophilum Toelken, 286
Halocnemum australasicum Moq., 294
Salicornia L., 296
arabica L. var. paardeneilandica (Ung.-Sternb.)
Chev., 289
australis Sol. ex Benth, 294
blackiana Ulbrich, 295
brachiata Roxb., 298
dolichostachya Moss, 261, 299
europaea L., 296
fruticosa L., 255
var. capensis Ung.-Sternb., 280
var. densiflora Ung.-Sternb., 284
var. paardeneilandica Ung.-Sternb., 289
var. radicans (Smith) Gr. & Godr., 289
fruticosa sensu Harv., 287
indica Willd., 276
lignosa Woods, 291
meyerana Moss, 298
pachystachya Ung.-Sternb., 297
pachystachya Black, non Ung.-Sternb., 295
pachystachya sensu Chev., non Ung.-Sternb.,
276, 298
perennis Mill, 289
var. radicans (Smith) Moss, 289
var. lignosa (Woods) Moss, 291
per rieri Chev., 297
quinqueflora Ung.-Sternb., 294
radicans Smith, 289
strictissima K. Gram., 261
uniflora Toelken, 299
303
Plate
l™.. Ll .fi. iUfea-0.
to,. iT ^ jjjjp s„. 5CJ
1. — Arthrocnemum mossianum Toelken (Tolken 552, holotype, PRE).
10862
304
Lfsl» H -JoAksyru
Anno ,ef 3 No. S36
Plate 2. — Arthrocnemuin terminate Toelken ( Tulken 536, holotype, PRE).
305
Plate 3. — Arthrocnemum xerophilum Toelken (Compton 20874, holotype, NBG).
306
Plate 4. — Arthrocnemum
deeumhens Toelken ( Tolken 362, holotype, PRE).
307
Plate 5. — Salicornia uniflora Toelken ( Tolketi 460, holotype, PRE),
10862
309
Bothalia 9, 2: 309-333.
A Revision of the African Species of Epilobium
(Onagraceae)
by-
Veter H. Raven*
Abstract
There are in Africa 10 native species of Epilobium, one of Circaea, and between 8 and 12 of Ludwigia
(including Jussiaea), depending upon how many species are considered native. These three genera
include all native African species of Onagraceae. Four sections of Ludwigia, comprising a total of
five species, and one fairly distinctive group of Epilobium are the most characteristic African groups.
Despite the fact that H. Perrier de la Bathie (in Humb., FI. Madagasc., Oenotheracees : 1-26, 1950)
considered all three of the species of Epilobium that he recognized and four of the 1 1 species of Ludwigia
(including Jussiaea), to be endemic to Madagascar, there are in my opinion no Mascarene endemics
in the family Onagraceae and nothing to suggest any great antiquity for the family on that island.
This observation agrees with the general pattern for the family in Africa in suggesting that plants of
this family are relatively recent arrivals on that continent. Indeed this generality could be extended
to apply to the Old World as a whole, where there is no endemic genus of the family (as opposed to
16 endemic genera and 2 endemic tribes — Lopezieae and Onagreae — in the New World, principally
in Mexico and the western United States) and relatively few endemic sections or subgenera.
The ten African species and one subspecies of Epilobium recognized are described, and a key
provided.
Introduction
The genus Epilobium (Onagraceae) probably comprises more than 200 species,
but is best represented at relatively high latitudes. Thus only a few members of the
genus occur in Africa. Nevertheless, the ten species and one additional subspecies
here recognized for the continent provide an interesting pattern phytogeographically,
and it has been more than 80 years since the last attempt to treat them as a group.
Representatives of the specimens examined are cited under each species.
In fact, the only comprehensive treatment that attempts to cover all African species
of Epilobium is that of Carl Hussknecht who, in his Monographic der Gattung
Epilobium (1884), included a section on the species found in Africa. Relatively little
material was available from Africa at the time Haussknecht wrote, and he adopted a
narrow species concept for the characteristically African group including E. capense,
E. salignum and E. stereophyllum. Haussknecht considered material, that is here
referred to one of these three species, as belonging to a total of 13 species, five of which
are here referred to E. capense and four each to the other two. He recognized eight
other species from Africa, among which E. angustifolium and E. lamyi (=E. tetragonum
subsp. lamyi) are considered as doubtfully African in the present work, E. madeirense
a synonym of E. obscurum, and E. tournefortii a subspecies of E. tetragonum (Hauss-
knecht’s E. adnatum).
Following Haussknecht’s revision, new species, subspecies and varieties were
described from tropical and southern Africa by various European authors, including
Hector Leveille, Th. Loesener, Adolph Engler, W. P. Hiern and E. G. Baker. Not
a single one of these is considered worthy of taxonomic recognition in the present
* Division of Systematic Biology, Stanford University, Stanford, California, U.S.A.
310
treatment. By the early part of this century, there were some 26 names that had been
applied to various forms of the three species that ranged into tropical Africa: E.
capense, E. salignum and E. stereophyllum. Various authors working on regional
treatments or works of more limited scope, for example H. Leveille, Ic. Gen. Epil. 68
(1910), and W. Robyns, FI. Sperm. Parc Nat. Albert 1: 684 (1948), clearly realized
that not all of these entities were distinct. The first comprehensive attempt to compare
them, however, was undertaken by J. P. M. Brenan of the Royal Botanic Gardens,
Kew, in preparing his treatments of Onagraceae for the floras of East and West
Tripical Africa. For the first time, Brenan brought together these segregates into
the same three species that are recognized here, and the present treatment is largely
based on an extension of his concepts beyond his geographical limits.
Several entities were likewise described from North Africa, of which E. atlanticum
and E. psilotum are strikingly distinct endemics of the Haut Atlas, the former also
occurring in the Sierra Nevada of Spain. Several other entities recently described from
North Africa, namely E. mirei Quezel, E. numidicum (Batt.) Batt., E. x caballeroi
Pau and E. tetragonum var. ampelusii Maire, are here reduced to -synonymy.
Relationships of the African Species
To summarize the preceding discussion, Haussknecht recognized 22 species of
Epilobium from Africa, of which two are tentatively excluded from the flora, 1 1 reduced
to synonymy and one regarded as a subspecies. Two more recently described species
are added, bringing the continental total to 10 species and one additional subspecies.
Of these, five ( E . hirsutum, E. parviflorum, E. lanceolatum, E. tetragonum subsp. tetra-
gonum and E. obscurum) range widely in Europe; two others ( E . tetragonum subsp.
tournefortii and E. atlanticum) have Mediterranean distributions and are found in
Europe; and one {E. psilotum , endemic to the Haut Atlas of Morocco) is clearly related
to European and north temperate species. Thus these eight species and subspecies
are identical with or closely related to European species, and may have reached Africa
relatively recently. Epilobium hirsutum extends down the eastern half of the continent
to South Africa, where it is common, and E. tetragonum subsp. tetragonum is likewise
common in South Africa where it was probably introduced from Europe. The other
six species and subspecies are restricted to the Atlantic islands and to a narrow band
within 150 miles of the Mediterranean, extending from Morocco to Tunisia.
The remaining three species, E. capense, E. salignum and E. stereophyllum, are a
characteristically African group without any clear affinities elsewhere in the genus.
Epilobium capense, found from southern Tanzania to the Cape, is an extremely variable
species with a deeply 4-lobed stigma. The other two species have more or less clavate,
undivided stigmas, and occur in the mountains of east Africa. Epilobium stereophyllum
ranges from equatorial east Africa north to the mountains of Ethiopia, where it is
common and variable, whereas E. salignum which overlaps with E. stereophyllum in
the northern end of its range, occurs from the equatorial mountains south to the eastern
half of South Africa. Both E. capense and E. salignum are found on Madagascar,
where they are the only two species of the genus present.
In the light of the data presented in this revision, it is interesting to consider the
representation of the family Onagraceae as a whole in Africa. Three native genera
occur in this continent: Circaea, Epilobium and Ludwigia (including Jussiaea ). The
only other genus of the family which has native species in the Old World is Fuchsia ,
with an endemic subgenus in New Zealand and Tahiti. Circaea is a characteristic
North Temperate genus not represented at all in the Southern Hemisphere; a single
widespread European species, C. lutetiana L. subsp. lutetiana, reaches North Africa,
where it occurs only in the coastal mountains of Algeria and Tunisia. In its pattern
of distribution it is obviously comparable to the primarily European taxa of Epilobium
311
discussed above and, like them, it may have reached Africa relatively recently. If
the European species of Epilobium did not reach Africa over water, it is difficult to
account for the absence in Africa of such characteristic European taxa as E. angusti-
folium L., E. montanum L., E. collinum Gmel., E. tetragonum subsp. lamyi (F. W.
Schultz) H. Lev., E. roseum Schreb. and E. palustre L. The ability of the genus to
disperse, by means of its plumed seeds, across considerable water barriers, is clearly
illustrated by the presence of three species on Madeira, more than 400 miles from
the mainland (in an area where no species of Epilobium grows); two on the Azores,
more than 800 miles from the mainland; and four on the Canary Islands, far from the
nearest stations for the species involved. There seems no compelling reason to think
the eight primarily European taxa of Epilobium and Circaea lutetiana may not have
reached Africa relatively recently, as the result of chance dispersal.
In Ludwigia, for which I recently revised the Old World species (in Reinwardtia
6 : 327-427, 1963), the pattern is very different. The genus is primarily tropical in
distribution, although it does include characteristically North Temperate groups. One
species of clear North Temperate affinities, L. palustris (L.) Ell., occurs in coastal
Morocco, Algeria and Tunisia (in the same area as the characteristically European
species of Epilobium and Circaea ), as well as on Socotra and in temperate southern
Africa, in both of which areas it may well have been introduced. South America
is clearly the centre of distribution and evolution of the tropical species of Ludwigia,
and five of the 13 remaining African species* are shared with South America, from
which two were certainly, and all may possibly have been, introduced. One other
L. perennis L., is found nearly throughout the tropics of the Old World. The seven
remaining species are essentially endemic to Africa, and represent six sections, four
of them endemic to Africa. The four endemic sections have no evident relationship
to one another, and it is therefore likely that several lines of the genus reached tropical
Africa at a relatively early date, with others coming more recently, up to the present.
The characteristic African group of Epilobium (comprising E. capense, E. salignum
and E. stereophyllum ) probably likewise reached Africa at a fairly early date and evolved
their characteristic combinations of characteristics in isolation there.
Acknowledgments
The present work was supported by U.S. National Science Foundation Grants
GB-141 and GB-4161. I am most grateful to Mr. J. P. M. Brenan, of the Royal
Botanic Gardens, Kew, for his helpful comments during the course of this study.
Dr. Elena Paunero, of the Instituto “Antonio Jose Cavanilles ”, Madrid, kindly supplied
me with useful information about the type of Epilobium caballeroi. Prof. Paul Quezel,
of the Faculte des Sciences de Marseille, made helpful comments on North African
species and allowed me to examine several critical specimens in his private herbarium.
I am likewise grateful to those in charge of the herbaria of the following institutions
who very kindly made all or part of their African material of Epilobium available for
this study: University of California, Berkeley: Jardin Botanique de l’Etat, Bruxelles;
Botany School, University of Cambridge; Bolus Herbarium, Cape Town; Compton
Herbarium and South African Museum Herbarium, Cape Town; Botanical Institute
of the University of Coimbra; Natal Herbarium, Durban; Royal Botanic Garden,
Edinburgh; Herbarium Universitatis Florentinae, Instituto Botanico, Firenze; For-
schungsinstitut und Naturmuseum Senckenberg, Frankfort; Herbarium of the Albany
Museum, Grahamstown; Institut fur Spezielle Botanik and Herbarium Haussknecht,
Jena; The Moss Herbarium, University of the Witwatersrand, Johannesburg; Royal
Botanic Gardens, Kew; Centro de Botanica de Junta de Investigacoes do Ultramar,
Lisboa; Jardin e Museu Agricola do Ultramar, Lisboa; Institute of Botany, Faculty
* I now regard L. pulvinaris Gilg as a synonym of L. senegalensis (DC.) Troch.
312
of Science, Lisboa; British Museum (Natural History), London; Institut de Botanique,
Universite de Montpellier; Botanische Staatssammlung, Miinchen; The East African
Herbarium, Nairobi; Museum National d’Histoire Naturelle, Laboratorie de Phanero-
gamic, Paris; National Herbarium, Botanical Research Institute, Pretoria; Federal
Herbarium, Salisbury; Dudley Herbarium, Stanford University; Botanical Depart-
ment, Naturhistoriska Riksmuseum, Stockholm; and U.S. National Museum,
Washington.
I am also grateful to the Plant Biology Division, Carnegie Institution of Washington,
Stanford, California, for generously affording me the facilities in which strains of
two of the endemic African species were grown, and to Prof. R. H. Compton, Mbabane,
Swaziland, and to the East African Herbarium, Nairobi, Kenya, for kindly collecting
these seeds for me. I am indebted to Miss Joan C. Burger for the illustrations.
EPILOBIUM
Key to the African Species
Stigma deeply 4-cleft:
Stems clothed with spreading pubescence; leaves subsessile:
Leaves distinctly clasping at base; petals 6-16 mm long 1. E. hirsutum
Leaves not clasping at base; petals (4 — )6— 9 mm long 2. E. parviflorum
Stems clothed with appressed pubescence or subglabrous; leaves petiolate:
Leaves narrowly cuneate at base, the petiole 3-10 mm long; petals white, fading rose; inflore-
scence drooping in bud 3. E. lanceolatum
Leaves not narrowly cuneate at base, the petiole 1-2-5 mm long; inflorescence not drooping
in bud 8. E. capense
Stigma entire, clavate or capitate:
Plants pubescent:
Plants less than 3 dm tall, with conspicuous leafy stolons creeping through cushions of moss;
Haut Atlas of Morocco 6. E. atlaniicum
Plants usually taller and lacking leafy stolons:
Pubescence entirely strigose except for a few glandular trichomes on the calyx; plants with
leafy stolons at the base; Madeira and Canary Islands 5. E. obscunim
Pubescence entirely strigose or with a copious admixture of glandular trichomes; plants
lacking leafy stolons:
Leaves narrow, subsessile and strongly decurrent, evenly and strongly denticulate; flowers
erect at anthesis:
Petals 2 - 5— 5(— 7) mm long; stigma surrounded by the anthers at anthesis 4a. E. tetragonum
subsp. tetragomim
Petals 7-12 mm long; stigma held above the anthers at anthesis 4b. E. tetragomim subsp.
tournefortii
Leaves often broader, distinctly petiolate (the petiole often short), not decurrent ; flowers
usually strongly nodding at anthesis:
Cauline leaves normally broadly rounded to subcordate at base; petals rose-
purple, 6-16 mm long, 4-11 mm wide 9. E. stereophylhmi
Cauline leaves cuneate to narrowly rounded at base; petals white or cream, then
rose following pollination, 5-15 mm long, 2-7 mm wide 10. E. saligmim
Plants entirely glabrous, to 2-5 dm tall 7. E. psilotum
1. Epilobium hirsutum L., Sp. PI. 347 (1753); Harv. in FI. Cap. 2 : 506 (1862);
Oliv. in FI. Trop. Afr. 2 : 487 (1871); Hausskn., Monogr. Epil. 53, t. I, fig. 20 (1884);
Batt. in Batt. & Trab., FI. Alger. 316 (1889); H. Lev., lc. Gen. Epil. tt.230, 231 (1911);
Muschler, Man. FI. Egypt 679 (1912); Jehandiez & Maire, Cat. PI. Maroc. 2 : 516
(1932); Adamson in FI. Cape Penins. 605 (1950); Andrews, FI. PI. Anglo-Egypt.
Sudan 1 : 143 (1950); Tackholm, Stud. FI. Egypt 208 (1956); Ross-Craig, Drawings
Brit. PI. 9 ; 1. 18 (1958); Quezel & Santa, FI. Nouv. Alger 2 : 640 (1963). Type: from
Europe.
313
E. tomentosum Vent., Descr. PL Nouv. Jard. Cels, t.90 (1802). Type: from Persia,
cultivated in France. E. hirsutum var. villosissimum (“ villosissima ”) Koch, Syn.
FI. Germ. Helv., ed. 1 : 240 (1835); Batt. in Batt. & Trab., FI. Alger. 316 (1889);
Brenan in FI. Trop. E. Afr., Onagraceae: 2, fig. 1 (1953); Fernandes & Fernandes
in Garcia de Orta 5 : 473 (1957); 7 : 493 (1959). Type: presumably from central
Europe. var. tomentosum (Vent.) Boiss., FI. Or. 2 : 746 (1872). var. villosum
Hausskn., Monogr. Epil. 55 (1884). No type indicated. var. africanum H. Lev.
in Bull. Herb. Boiss., ser. 2, 7 : 589 (1907). E. mirei Quezel in Bull. Soc. Hist. Nat.
Afr. Nord. 48 : 90 (1957).
Robust herb, 0-2-2 -5 m tall, the subligneous stems sometimes persistent; rhizome
stout, producing thick white underground runners with very scattered cataphylls;
plants more or less white-pubescent, densely covered, especially in the inflorescence,
with long spreading trichomes. Leavers mostly opposite, alternate above, oblong-
lanceolate, densely hairy, acute, sessile, clasping at the base, coarsely toothed, 2-12 x
0-5-4 cm; rosette leaves longer and less hairy than later leaves; young leaves more
glabrous, to 20 cm long, petiole to 5 cm. Inflorescence with an admixture of glandular
trichomes, erect in bud; flowers erect in bud. Hypanthium 2-5-3 m across c. 1-1 -5 m
deep. Sepals 6-10 mm long, 2-2-5 mm wide, apiculate. Petals 6-16 mm long, 6-15
mm wide, deeply notched, bright purplish-rose. Anthers 1 • 5-2 mm long, the filaments
of the longer 3-5-6 mm long, of the shorter 2-3 mm long. Styles 6-10 mm long;
stigma deeply 4-lobed, held above the anthers at anthesis, the lobes 1 -5-2-5 mm long.
Capsules densely villous, 3-8 cm long, on a pedicel 2-12 mm long; seeds dark brown
or even coppery, oblong-obovoid, acute at the base, coarsely papillate, 0 • 9-1 ■ 1 5 mm
long, the coma 5-7 mm long, dull white. Gametic chromosome number, n = 18.
Reportedly protandrous.
Fig. 1. — Range of Epilobium hirsutum in Africa.
314
Geographical range (Fig. 1): Wet places as in swamps and marshes and beside
rivers and lakes, Canary Islands, Cape Verde Islands, and Morocco along the coast
of North Africa to Egypt, whence south in the eastern part of the continent to the
Cape; sea level to 2,600 m elevation, but mostly above 1,200 m elevation in equatorial
regions. The species ranges throughout Europe except for the extreme north and
across Asia to China and Japan, reaching south to about 27° N in Nepal. It has
become naturalized in North America.
Canary Islands. — La Palma, Gran Caldera, in 1892, Murray s.n. (K).
Cape Verde Islands. — Santo Antao, Cova, Bordeiros, Chevalier 45606 (P). San Tiago, Ponta da
Cruz, Chevalier 44610 (COI, K, P). Fogo, at the Fonte up Monte Whuco, Lowe s.n. (BM).
Morocco. — Bou Gamez Taorirt near Ait Milch, c.2000 m, Spence SI 89 (E). Arrond, Haut Atlas,
c.2000 m, Polunin 2172 (BM).
Spanish Morocco. — near Xauen, 600 m, Font Quer 289 (BM, G).
Algeria. — Sidi-bel-Abbis, prov. d’Oran, in 1863, Lefrane s.n. (P). Environs d’Alger, in 1856,
Durando s.n. (E). Kroubs, Dep. Constantine, in 1954, Doppelbaur s.n. (M).
Tunisia. — Au Nord d'Ai'n Draham, Kroumirie, in 1883, Cosson et al. s.n. (P).
Libya. — Cyrenaica, Derna, U. Derna, Pampanini & Pichi-Sermolli 5367 (FI).
Egypt. — Marais entre le canal Makmoudieh et le Ch. du fer du Caire, sous Alexandrie, in 1871,
Du Parquet s.n. (P). Entre le Jardin du Khedive et 1’ extremite du petit lac du Bamleh, in 1872, Du
Parquet s.n. (P).
Chad. — Mare d’Arigui, Mire (herb. Quezel). Tigui, in 1956, Quezel s.n. (herb. Quezel; holotype
of E. mirei). Tibesti, Nema-Nemasso, 1400 m, Quezel s.n. (herb. Quezel).
Sudan. — Darfur Prov., Nyuringya, on west side of Marra Mountains, c. 2000 m, Dandy 169 (BM).
Darfur Prov., Suni, 1700 m, Robertson 117 (BM).
Ethiopia. — Eritrea, Hamasen, road from Asmara to Keren, Buscaloni 400 (FI). Begemdir, Gondar
and vicinity, Massey 8 (K). Shoa, near Wofasha, 39° 48' E, 9° 44' N, 2500 m, Mooney 7886 (BR,
K). Harar, Galla Pass, 2250 m, 9° 28' N, 42° 19' E, Gillett 5168 (F, K, P).
Republic of the Congo. — Orientale, entre Aru et Mahagi, Yibali-Ytieri, 1650 m, Lebrun 3731
(BR). Kivu, Terr. Lubero, S.A.L. Luhotu, route Lubero-Butembo, 2100 m, Christiaensen 2127 (BR).
Rwanda. — Lubona, Astrida, 1700 m, Becquet 701 (BR).
Burundi. — Muruketchu, Ngozi, 1700 m, Becquet 873 (BR).
Uganda. — West of Lake Bunyoni, 35 miles east of Rutshuru, 220 m, Chapin 268 (BR, US). N. E.
Elgon, 8,000 feet, Tweedie 711 (BR, K).
Kenya. — Endebiss, 7,500 feet, Webster (EA). South side of Lake Naivasha, 1870 m, Mearns 701
(BM, BR, US).
Tanzania. — Muleha, Makuyuni Dist., Koritschoner 1360 (EA, K). Kilimanjaro, southeast side,
1800 m, Schlieben 4383 (BM, BR, G, LISC, M). Near Mondo, Kondoa Dist., 1340 m, Burtt 1022
(BM, EA, K). Uluguru, Uhehe, Usangu, V. Prittwitz & Gaffron 104 (BM). Mbeya Forest Reserve,
Mbeya, Southern Highlands, 2,400 m, Myembe 53 (EA, K).
Angola. — Huila, Mendes 1408 (LISC). Huila, ad rivulem de Lopolo, 1500 m, Welwitsch 4457
(BM, COI, K, G, LISU, P).
Mozambique. — Manica e Sofala, Himalaya, roadside in indigenous forest down P.E.A. side, 1400 m,
Lady Drewe 26 (BM, SRGH).
Rhodesia. — Mazoe, 1350 m, Eyles 524 (BM, BOL, K, PRE, SAM, SRGH). Mann’s Farm, Salisbury
Dist., Drummond 6781 (COI, EA, K, M, SRGH). Inyanga Dist., near Cheshire, c.1300 m, Norlindh
& Weimarck 4795 (BM, BR, K, PRE).
South West Africa. — Gross Waterberg Mt., near Waterberg, Rodin 2586 (BOL, K, PRE, US).
South Africa. — Cape.— Aliwal North: Eland’s Hoek, 1350 m, Bolus 29 (BOL). Bedford: Upper
Cowie, 750 m, Bennie 264 (PRE). Bredasdorp: Papiesvlei, near Elim, 100 m, Schlechter 10446 (BM,
BOL, BR, E, GRA, K, P, PRE, US). Peninsula: Camps Bay, Marloth 6044 (PRE). Clanwilliam:
Clanwilliam, Leipoldt 138 (PRE). Kuruman: Batlharo, Silk 269 (BOL, K). Matatiele: Mvenyani,
P.O. Cedarville, Bandirt 166 (PRE). Riversdale: Plattebosch, near Still Bay, 30 m, Muir 103 (J).
Orange Free State. — Harrismith: Harrismith, in 1928, Kulling s.n. (PRE). V redefort: Parys, Vaal
River, Obermeijer s.n. (BOL). Natal. — Estcourt: Winterton, King 357 (NH). Umzinto: 750 m,
Rudatis 1649 (BM, E, G, K, P, PRE, US). Transvaal. — Barberton: near Barberton, Thorncroft
1070 (K). Letaba: 1050 m, Scheepers 625 (K). Lichtenburg: Lichtenburg, Jenkins 11203 (BOL).
Lydenburg: Lydenburg, 1600 m, Schlechter 3939 (BM, BOL, BR, COI, GRA, K, P, PRE). Pretoria:
Pretoria, Rehmann 4752 (GRA). Soutpansberg: Louis Trichardt, in 1919, Breyer s.n. (PRE).
Lesotho.— Leribe, Dieterlen 764 (K, NH, P, PRE, SAM). Roma 5500 ft., Ruch 1581 (PRE).
315
Epilobium hirsutum is, as pointed out by Brenan (in Turrill and Milne-Redhead,
FI. Trop. E. Afr., Onagraceae : 3, 1953), relatively uniform in appearance in Africa,
in contrast to its great variability in the European and Asiatic portions of its range.
Brenan (l.c.) suggested that this might indicate the possibility of a “ relatively recent
and rapid migration southwards of a single race capable of withstanding African
conditions On the other hand, it is interesting to note that Michaelis (in Zeitschr.
Pflanzenzuchtung 30 : 250-275, 1951) found that cytoplasmically, South African
strains of E. hirsutum were distinct from all European ones he tested. The present
species is relatively rare between lat. 10° S and lat 20° S, but again abundant in South
Africa; but there is apparently no strong reason to think that it reached South Africa
in historical times.
Epilobium hirsutum grows with E. capense, E. salignum and E. tetragonum subsp.
tetragonum in South Africa. I have seen one specimen of what is surely a hybrid
between E. capense and E. hirsutum, from South Africa: Mt. Currie Dist., near
Kokstad, 1500 m, Feb. 1883, Tyson 1423 (SAM). This sheet consists of two branches,
both evidently from a single vigorous plant, and is perfectly intermediate between
the parental species. The pubescence is somewhat appressed but longer than in E.
capense, the leaves almost sessile, with teeth intermediate between the forward-directed
ones of E. capense and the spreading ones of E. hirsutum and the plants have large,
purlish-rose petals. The capsules are obviously failing to set, and only 20-6 per cent
of the pollen was stainable in cotton blue in lactophenol, based on an examination
of 1,500 grains. Generally only one grain of each tetrad was filled. Hybrids with
E. salignum and E. tetragonum have not been observed, although the hybrid with the
last-mentioned species is known to be rather frequent in Europe.
Epilobium caballeroi, supposed to be the hybrid between E. hirsutum and E. tetra-
gonum subsp. tournefortii, is here considered a synonym of E. parviflorum. Epilobium
mirei is clearly a form of E. hirsutum with relatively small flowers and small, rounded
leaves. Such plants are often found among European material, particularly among
those flowering late in the season or in relatively dry habitats. I have no records of
E. hirsutum growing with any other species in North Africa or on the Atlantic islands,
although it must occasionally do so. Hybrids with E. lanceolatum, E. obscurum, E.
parviflorum and E. tetragonum subsp. tetragonum are known from Europe, and the
hybrid with E. tetragonum subsp. tournefortii has been described from Syria (Hausskn.,
Monogr. Epil. 74, 1884).
2. Epilobium parviflorum Schreb., Spicil. FI. Lips. 146, 155 (1771); Lowe, Man.
FI. Madeira 1 : 271 (1868); Hausskn., Monogr. Epil. 66, t.I fig. 21 (1884); Webb
& Berthol., Hist. Nat. lies Canar. 3, 2 : 7 (1888); Batt. in Batt. & Trab., FI. Alger.
316 (1889); Pitard & Proust, FI. Arch, ties Canar. 183 (1909); H. Lev., Ic. Gen. Epil.
t. 232 (1911); Maire, Cat. PI. Maroc 2:516 (1932); Ross-Craig, Drawings Brit.
PI. 11 : 1. 1 9 (1958); Quezel & Santa, FI. Nouv. Alger 3 : 640 (1963). Type: from
Europe.
E. parviflorum var. numidicum Batt. in Bull. Soc. Bot. Fr. 40 : 478 (1896). E.
numidicum (Batt.) Batt. in Batt. & Trab., Synop. Alger. 127 (1902); Quezel & Santa,
FI. Nouv. Alger 2 : 640 (1963). Epilobium X caballeroi Pau (“F. hirsutum x tourne-
fortii ”) in Bol. Soc. Esp. Hist. Nat. 22 : 59 (1922); Jehandiez & Maire, Cat. PI. Maroc.
2 : 516 (1932). E. parviflorum “var. mollissimum (Welw.) H. Lev.”; Jehandiez &
Maire, Cat. PI. Maroc 2 : 516 (1932).
Perennial herb, 0-2-1 -4 m tall; in autumn producing short-stalked leafy rosettes
near the base of the plant; plants covered with short soft spreading hairs throughout,
the stems with raised lines decurrent from the margins of the petioles, glandular-
pubescent in the inflorescence or rarely subglabrous. Leaves narrowly lanceolate to
narrowly ovate, 3-12 cm long, 1-3-5 cm wide, acute, subsessile, denticulate, softly
316
hairy, opposite below, alternate in the upper half. Inflorescence erect in bud. Hypan-
thium 1—1 - 9 mm deep, 1-3-2 mm across at the summit. Sepals (2-) 3-6 mm long,
(1-) 1 -4—2 mm wide at the base, lanceolate, acute. Petals (4-) 6-9 mm long, 3 -5-4 -8
mm wide, deeply notched, rose-purple. Anthers 1-1 -2 mm long, the filaments of the
longer 1 - 8—3 - 5 mm long, of the shorter 1-2 mm long. Style 2 -2-4 -7 mm long; stigma
4-lobed, the lobes 1-1 ■ 8 mm long, surrounded by the shedding anthers at anthesis.
Capsule 3 • 5-7 cm long, glabrous to densely pubescent with erect hairs, on a pedicel
0-8-1 -7 cm long; seeds brown, obovoid, rounded at the base, coarsely papillose,
0-9-1 mm long, the coma c. 5-6 mm long, dull white. Gametic chromosome number,
n — 18.
Geographical range (Fig. 2): In moist places from sea level to 1700 m elevation,
Madeira, Canary Islands, and from Morocco to Tunisia. Also known from the Azores
(e.g. July, 1868, Drouet, BM). The general range of this species is from Europe, where
it occurs throughout except for the extreme north, to western China. In the Himalaya
it ranges east to Nepal at about 27° N lat.
CR. t
Fig. 2. — Range of Epibolium parviflorum in Africa (triangles), and total range of E. stereophyllum
(circles) and of E. capense (dots).
Madeira. — Along the Levada in the Ribeira d’Ametade, in 1849, Lowe s.n. (BM). Queimadas,
bank of Levada in woodland ( Erica scoparia), 1000 m, Coleridge 8 (BM).
Canary Islands. — La Palma, Barranco del Rio, Sprague & Hutchinson 211 (K). Gomera, San
Sebastian, in 1905, Pitard s.n. (P). Tenerife, San Andreas, Knoche 116 (DS).
Tangier. — Fide Haussknecht, Monogr. Epil. 67 (1884).
Spanish Morocco. — Xauen, in 1921, Vidal s.n. (MA, holotype of E. caballeroi, not seen; photo-
graph, DS).
317
Morocco. — Great Atlas, in valley of the Aid Messane, c. 1300 m, Lindberg 3664 (K). Middle
Atlas, Ras-el-Ma, Wooddell 157 (BM). Moyen Atlas, Dai'et Achlef, bords de la Daya, 1700 m,
Jehandiez 583 (P).
Algeria. — Murs de Tlemcen, Jourdan 111 (P). Kerrata, Kabylie, 800 m, in 1897, Reverchon s.n.
(E). La Calle, Battandier (MPU, lectotype of E. numidicuni). Djebel Megris, Battcmdier (MPU).
Tunisia. — Burgous, Tunis, Letourneux (P).
Epilobium caballeroi was described as a hybrid between E. hirsutum and E. tetra-
gonum subsp. tournefortii but, in a photograph of the type (DS), it seems clearly to
be a luxuriant form of E. parviflorum. The petals were described as being 10 mm
long, but no flowers this large are visible on the photograph I have examined, the
largest being about 8 mm long. Epilobium numidicuni is likewise, although relatively
slender in habit, indistinguishable from many European plants of E. parviflorum.
Haussknecht (Monogr. Epil. 67, 1884) also recorded E. parviflorum from Santa
Antao in the Cape Verde Islands. I have seen no material, and A. & R. Fernandes
do not mention it in their treatment of the Onagraceae of these islands (in Garcia de
Orta 5 : 473-474, 1957). Its presence there requires confirmation.
I have no records of E. parviflorum growing with other species in North Africa
or on the Atlantic islands, but hybrids with E. hirsutum, E. lanceolatum, E. tetragonum
subsp. tetragonum and E. obscurum are known from Europe.
3. Epilobium lanceolatum Seb. & Mauri, FI. Rom. Prod. 138 (1818); Hausskn.,
Monogr. Epil. 90, t. II fig. 22 (1884); Batt. in Batt. & Trab., FI. Alger. 316 (1889);
Ross-Craig, Drawings Brit. PI. 11: t. 21 (1958); Quezel & Santa, FI. Nouv. Alger
2 : 639 (1963). Type: from Europe.
Perennial herb, 0 - 1-0-6 m tall, producing leafy rosettes at ground level in autumn;
stems strigulose usually with a strong admixture of glandular hairs above and in the
inflorescence, or more rarely subglabrous below, with obscurely marked raised lines
running down a short distance from the margins of the petioles. Leaves elliptical
to elliptical-lanceolate, 3-12 cm long, 1-3-5 cm wide, acute, narrowly cuneate
at the base, on a petiole 3-10 mm long, sharply serrulate, opposite near the base,
alternate above. Inflorescence drooping in bud, the pubescence strigulose with
many glandular hairs intermixed; flowers erect in bud. Hypanthium 1 -1-2 mm deep,
1-1-3 mm across at the summit. Sepals 3-4-5 mm long, 0-8-1 -3 mm wide at base,
lanceolate, acute. Petals 4-5-8 mm long, 2-6-4 mm wide, shallowly notched, white
or very pale purplish-rose at first, becoming darker after pollination. Anthers 0-8-1 • 1
mm long, the filaments of the longer 3-5 mm long, of the shorter 2 -5-3 -2 mm long.
Styles 4-5-5 mm long; stigma 4-lobed, the lobes c. 1 mm long, erect or spreading,
surrounded by the shedding anthers at anthesis. Capsules strigulose, usually with an
admixture of glandular hairs, 3-7 cm long, on a pedicel 1-2-5 cm long; seeds reddish-
brown, oblong-obovoid, acute at the base, finely papillose, c. 1 mm long, the coma
c. 5 mm long, dull white. Gametic chromosome number, n — 18.
Geographical range (Fig. 3): Petite Kablyie and Djuradjura Mountains of northern
Algeria (cf. Hausskn., Monogr. Epil. 92, 1884). The general range of this distinctive
species is in Europe, where it is common in the central and southern parts and rarer
northwards, eastward to western and central Anatolia and the Caucasus.
Algeria. — Djebel Tababor, Petite Kabylie, prov. de Constantine, 22 July, 1861, Cosson (P); 25
June, 1880, Cosson (P).
Epilobium lanceolatum is not known to grow with other species in Africa, although
it may do so. Hybrids with E. hirsutum, E. obscurum, E. parviflorum, and E. tetragonum
subsp. tetragonum have been described from European material.
318
Fig. 3. — Ranges of Epilobium obscurum (crosses) and E. lanceolatum (solid triangles) in Africa, and
total ranges of E. atlanticum (circles), E. psilotum (hollow triangle), and E. salignum (dots).
The related E. montanum L., a Eurasian species not otherwise known from Africa,
was collected once as an introduced plant on Madeira: “A weed in the Rectory Garden,
Lea ”, August, 1863, Lowe (G). It can easily be distinguished from E. lanceolatum
by its leaves, which are rounded at the base, and the colour of its petals, which are
bright purplish pink from the time they first open.
4. Epilobium tetragonum £., Sp. PI. 1 : 348 (1753); Harv. in FI. Cap. 2 : 597
(7862); Lowe, Man. FI. Madeira 1 : 273 (1868); Batt. in Batt. & Trab., FI. Alger.
315 (1889); Pitard & Proust, FI. Arch, lies Canar. 183 (1909); Jehandiez & Maire,
Cat. PI. Maroc 2 : 515 (1932); Adamson in FI. Cape Penins. 606 (1950); Quezel
& Santa, FI. Nouv. Alger. 2 : 639 (1963). Type: from near Montpellier, France.
Perennial herb, often flowering the first year, 0-1-1 -8 m tall; in autumn producing
leafy rosettes from short stolons near the base; plants strigulose, especially above,
and with raised lines running down from the bases of the petioles, these pubescent
above. Leaves narrowly lanceolate to nearly elliptical, 1-5-7 -5 cm long, 0-3-1 -5
(-2-1) cm wide, bluish-green, mostly opposite, alternate above, acute, subsessile and
usually decurrent into the raised lines on the stem, evenly and strongly denticulate.
Inflorescence erect in bud; pubescence of inflorescence entirely strigulose; flowers
erect in bud. Hypanthium 0-8-2 -5 mm deep, 1-3-5 mm across at summit. Sepals
2-8-7 -5 mm long, 0-8-2 -2 mm wide at base, lanceolate, apiculate. Petals 2-5-12
mm long, 1-8-7 mm wide, shallowly notched, pale lilac to rich rose-purple. Anthers
0-3-3 mm long, the filaments of the longer 1-3-4 mm long, of the shorter 0-3-3 mm
long, the filaments of the longer 1-3-4 mm long, of the shorter 0-3-2 -2 mm long.
319
Styles 1 -5-9 mm long; stigma clavate, 1 -8-3-6 mm long, surrounded by or held well
above the shedding anthers at anthesis. Capsules densely strigulose, (3 • 5-) 5-1 1 cm
long, on a pedicel 1-2-3 cm long; seeds brown, oblong-obovoid, acute at the base,
coarsely papillose, 1-1-3 mm long, the coma c. 8-10 mm long, dull white. Gametic
chromosome number, n = 18.
(a) subsp. tetragonum
E. adnatum Griseb. in Bot. Zeit. 10 : 851 (1852); Hausskn., Monogr. Epil. 97, t.II
fig. 31 (1884); Ross-Craig, Drawings Brit. PI. 11 : t.24 (1958). Type: from Europe.
E. tetragonum subsp. adnatum (Griseb.) Maire in Jehandiez & Maire, Cat. PI. Maroc
2 : 515 (1932); Quezel & Santa, FI. Nouv. Alger 2 : 639 (1963). var. ampelusii
Maire, Mem. Soc. Sci. Nat. Maroc 15 : 33 (1927); Jehandiez & Maire, Cat. PI. Maroc
2 : 515 (1932). var. rodriguezii “(Hausskn.)” Maire in Mem. Soc. Sci. Nat. Maroc
15 : 33 (1926); Jehandiez & Maire, Cat. PL Maroc 2 : 515 (1932). E. virgatum sensu
Batt. in Batt. & Trab., FI. Alger. 315 (1889); non Lam., Diet. Encycl. 2 : 375 (1786).
E. palustre sensu Webb & Berthelot, Hist. Nat. ties Canar. 3, 2 : 8 (1888); Pitard &
Proust, FI. Arch, ties Canar. 184 (1909), probably; non L., Sp. PI. 348 (1753). E.
capense sensu H. Lev., Ic. Epil. t. 40 (1910); non Buch. ex Hochst. in Krauss in Flora
27 : 425 (1844).
Plants mostly 0-1-1 m tall, autogamous. Hypanthium c. 1 mm across and deep.
Sepals 2 - 8—4 - 2 mm long, 0-8-1 -8 mm wide at base. Petals 2-5-5(-7) mm long,
2-3 -5(^1 -5) mm wide, pale lilac. Anthers 0-3-1 -3 mm long, the filaments of the longer
1- 3-2-1 mm long, of the shorter 0-3-1 -2 mm long. Styles 1 -5-3-2 mm long; stigma
surrounded by the shedding anthers at anthesis. Capsules (3 ■ 5— )5— 1 1 cm long. Gametic
chromosome number, n = 18.
Geographical range (Fig. 4): Generally in moist places, Madeira, Canary Islands,
and from Tangier to Tunisia in the coastal hills and mountains at low elevations.
The general range is throughout Europe, except for the extreme north, east to Iran
and the Caucasus.
Madeira. — Cliffs to the east of Funchal, in 1862, Lowe s.n. (BM).
Canary Islands. — Gomera, in humidis convalle Sancti Sebastiani, in 1845, Bourgeon s.n. (BM,
K, P). Tenerife, above Cruz de Carmen, Cook 1076 (US). In humidis alterioribus insulae Teneriffeae,
“ Epilobium palustre”, Barker-Webb (P).
Tangier. — Djebel Kebir, in 1921, Pitard s.n. (P); Point Ampelusius near Tangier, Maire (type of
var. ampelusii, AL, RAB, not seen).
Algiers. — Pres Bone, in 1841, Durieu s.n. (P). Philippeville, in 1840, Durieu s.n. (P). Mila, prov.
de Constantine, in 1840, Durieu s.n. (P).
Tunisia. — Sidi Athman El-Ahdded, Kroumirie orientale, in 1888, Cosson et al. s.n. (P).
South Africa. — Cape. — Albany: Coldspring, Glass 224 (GRA). Aliwal North: Eland’s Hoek,
near Aliwal North, c. 4700 ft.. Bolus 215 (BOL, PRE). Barkley West: Danielskuil, Esterhuysen 2028
(BOL, NBG). Graaff-Reinet: Graaff-Reinet, 2500 ft, Bolus 415 (BOL, PRE). King William’s Town:
Kei Road, 2000 ft, Schlechter 6129 (BOL, GRA). Middelburg: Van Wyk’s River, 4500 ft, Archibald
3037 (GRA). Murraysburg: Mt. Randveld near Murraysburg, Tyson 419 (GRA). Oudtshoorn:
Huis River Pass, Compton 20330 (NBG). Peninsula: Wolley Dod 672 (BM). Queenstown: Berry
Reservoir, east of Queenstown, Hilner 303 (GRA). Riversdale: Oakdale, Muir 3028 (PRE). Somerset
West: Somerset West, Parker 3638 (BOL, K, NBG). Tarka: Great Winterberg south of Tarkastad,
Fairfield Farm, c. 5000 ft, Comins 198 (K, PRE). Uitenhage: Zwartkopsrivier, Zeyher (SAM). Victoria
East: Woburn, 1800 ft, Acocks 11099 (K, PRE). Orange Free State. — Bethlehem: Golden Gate,
Compton 22476 (BOL. NBG). Harrismith: Harrismith, Sankey 71 (K). Natal. — Estcourt: Mooi
River, Johnston 369 (E). Utrecht: Kafir Drift, Tweekloof, Thode A320 (K, NH, PRE). Transvaal. —
Johannesburg: Van Wyksrust, West Witwatersrand, Moss 18312 (BM, J). Middelburg: Klein Olifants
River near Middelburg, Young A109 (PRE). Pretoria: Rietvallei 221, north portion, c. 5000 ft,
Acocks 11334 (PRE).
Lesotho.— Leribe, 5000-6000 ft, Dieterlen 936 (K, P, PRE, SAM). Mokhotlong Dist., 9300 ft,
Guillarmod 1101 (PRE).
320
There have been several reports of Epilobium tetragonum subsp. lamyi (F. W.
Schultz) H. Lev. in Africa, for example: Madeira. — Mandon (Hausskn., Monogr.
Epil. 107, 1884); Algeria.— Mitidja, Pomel (Batt. in Batt. & Trab., FI. Alger. 315,
1889); Tenerife. — Cabrera (Pitard & Proust, FI. Arch, lies Canar. 184, 1909);
Algeria. — Djebel Kebir, Lindberg (Jehandiez & Maire, Cat. PI. Moroc 515, 1932);
Quezel & Santa, FI. Nouv. Alger 2 : 639, 1963). 1 have not seen any of the specimens
concerned and consider it probable that they are either subsp. tetragonum or perhaps
E. obscurum.
The two populations of E. tetragonum subsp. tetragonum in Africa are separated
by more than 4,000 miles and the entire width of the tropics; furthermore, they are
very distinct from one another. The North African population is extremely variable,
intergrading with the very large-flowered subsp. tournefortii , whereas the South African
population is extraordinarily constant morphologically. These South African plants
are exactly similar to the common small-flowered race of central and northern Europe
morphologically, and it is likely that they have been derived from a single early
introduction from Europe into South Africa. The earliest South African collection
I have seen is in the Burman collection (G), and is labelled “ Caput bonae Spei ”,
As pointed out to me by Dr. B. de Winter, however, weeds were introduced into South
Africa very soon after the settling of the Cape in 1652, and it would not be surprising
to find one well established some 80 years later. This species is commonly represented
in 19th Century collections but almost always in settled regions. As the South African
plants have very little in common with most North African plants it is very unlikely
that they were introduced from North Africa. In South African material, the pale
petals are 2 -5-3 -5 mm long, with other flower parts correspondingly small. This
race, like the common European one, is very highly autogamous.
Although it may do so, E. tetragonum subsp. tetragonum is not known to grow
with other species of the genus in Africa. In Europe, hybrids with E. hirsutum, E„
lanceolatum, E. obscurum and E. parviflorum are known.
(b) subsp. tournefortii (Michal.) H. Lev., Monde des PI. 6 : 22 (1896); Jehandiez
& Maire, Cat. PI. Maroc 2 : 515 (1932); Quezel & Santa, FI. Nouv. Alger 3 : 639
(1963).
E. tournefortii Michal., Bull. Soc. Bot. Fr. 2 : 731 (1855); Hausskn., Monogr.
Epil. 112, t. I fig. 11, t. II fig. 29a, b (1884); Batt. in Batt. & Trab., FI. Alger. 315
(1889). (Based on plants from Corsica, Sardinia, Portugal and Tangier, herb. Gay.).
E. tetragonum var. ? tingitanum Ball, J. Linn. Soc. Bot. 16 : 459 (1878).
Plants mostly 1-1-8 m tall, allogamous. Hypanthium 1-2-2 -5 mm deep, 2-3-5
mm across. Sepals 5-7-5 mm long, 1 -2-2-2 mm wide at base. Petals 7-12 mm long,
4-5-7 mm wide, rich rose-purple. Anthers 1-3 mm long, the filaments of the longer
3-4 mm long, of the shorter 1-5-2 -2 mm long. Styles 4-9 mm long; stigma usually
held well above the shedding anthers at anthesis. Capsules mostly 5-6 cm long.
Geographical range (Fig. 4): Moist places near the Mediterranean, from Tangier
to Tunisia: from sea level to 1,000 m elevation. Outside of Africa, this subspecies
is known from southern Portugal, Spain and France; Sicily, Corsica, Sardinia and
Malta; Turkey, where it ranges to eastern Anatolia, Syria and Lebanon.
Tangier. — Tangier, “ E . tingitanum Salzman ”, in 1835, Salzman s.n. (P, type of var. ? tingitanum,
G). Near Tangier, Lindberg 1331 (K).
Spanish Morocco. — In argillosis humidis montis Djebel Zem-Zem, 200 m. Font Quer 459 (G, P).
Algeria. — Maison Carree, near Algiers, Gandoger 586 (P). Kabylie, 1000 m Reverchon 237 (BM,
DS, E, G).
Tunisia. — Tabarque, Kroumirie orientale, in 1888, Cosson et at. s.n. (P). Between Bizerte and Sidi
Hassoun, in 1887, Letourneux s.n. (P).
321
Fig. 4.- — Range of Epilobium tetragonum in Africa. Dots indicate stations for subsp. tetragonum
and triangles localities where subsp. tournefortii occurs with it. Subsp. tournefortii is not known
to occur without subsp. tetragonum except locally.
The modally outcrossing plants of this handsome and distinct subspecies differ
from those of subsp. tetragonum principally in their larger, darker-coloured flowers,
doubtless associated with an increased frequency of visits by insects. They likewise
appear to have broader and more evenly and coarsely denticulate, somewhat bluish
green leaves. The peculiar, disjunct pattern of distribution of E. tetragonum subsp.
tournefortii around the Mediterranean, coupled with the fact that it is one of the very
few outcrossing entities in the genus, suggests strongly that it may be a relic similar
to the populations from which the autogamous subsp. tetragronum and subsp. lamyi
were derived. It is always found in the same areas as subsp. tetragonum , and here
intermediates between the two are frequent and apparently completely interfertile.
In my opinion, the prevalent autogamy of subsp. tetragonum contributes to the main-
tenance of its distinctness where it occurs with subsp. tournefortii, and microgeographical
and perhaps ecological isolation doubtless also plays a role in their separation.
Epilobium tetragonum subsp. tournefortii is not known to grow with other species
in Africa, although it probably does so. Hybrids with E. parviflorum have been described
from Syria (Hausskn., Monogr. Epil. 74, 1884).
5. Epilobium obscurum Schreb., Spic. FI. Lips. 147, 155 (1771); Loew, Man. FI.
Madeira 1 : 274 (1868); Hausskn., Monogr. Epil. 114, t. II fig. 33a, b (1884); Pitard
& Proust, FI. Arch, lies Canar. 184 (1909). Type: from Europe.
Epilobium madeirense Hausskn. in Oester. Bo . Zeit. 29 : 90 (1879); Hausskn.,
Monogr. Epil. 232, t. XI fig. 59a, b, c (1889). E. lanceolatum sensu Lowe, Man. FI.
Madeira 1 : 274 (1868); non Seb. & Mauri, FI. Rom. Prod. 138 (1818). E. miguelense
10862.— 4
322
H. Lev. in Monde des PI. 3 : 91, t.3 (1894). Type: Azores, Ilh. de S. Miguel, Cande-
laria, Aug. 1891, Carreiro. E. tetragonum L. subsp. gilloti H. Lev. in Monde des PI.
6 : 21 (1896). Based on E. obscurum and E. virgatum Lam., Diet. Encycl. 2 : 375 (1786).
Perennial herb, 0-2-0 -8 m tall; in late summer producing elongated leafy stolons
above or on the surface of the ground, these not terminating in distinct rosettes; plants
rather sparsely strigulose above, glabrous below, with raised lines running down from
the margins of the petioles. Leaves lanceolate to narrowly ovate, 1 • 5-7 cm long,
0-5-1 -8 cm wide, acute to obtuse, subsessile and often somewhat decurrent, sparsely
denticulate, mostly opposite, alternate above. Inflorescence erect in bud, the pube-
scence strigulose but with a few glandular hairs on the hypanthium; flowers erect
in bud. Hypanthium c. 1 mm deep, c. 1 • 5 mm across at summit. Sepals 2 • 5^1 mm
long, 1-1-3 mm wide at base, lanceolate, apiculate. Petals 3-5-6 mm long, 1-8-3
mm wide, shallowly notched, rose-purple. Anthers 0-5-1 mm long, the filaments of
the longer 2-2-2 mm long, of the shorter 1-1-3 mm long. Styles 2-5-3-3 mm long;
stigma clavate, 1 • 5-2 mm long, surrounded by the shedding anthers at anthesis.
Capsules strigulose, 4-6 cm long, on a pedicel 0-4-1 -6 cm long; seeds brown, oblong-
obovoid, acute at the base, coarsely papillose, c. 1 mm long, the coma 4-5 mm long,
dull white. Gametic chromosome number, n = 18.
Geographical range (Fig. 3); In moist places from sea level to 1,300 m elevation,
Madeira and Canary Islands. Also in the Azores (e.g. Trelease 341, BM). The general
range is throughout Europe, except for the extreme north. E. obscurum reaches Asia
only in the Amanus Mountains of southern Turkey.
Maderta. — Mason (W, holotype of E. madeirense; isotype, BM). Madeira, in 1856, Mason 304
(CAM, JE). Pico Grande, Caminho Central, 1300 m, Bommiiller 611 (BM, JE). Fajao da Corte
(Curral das Freiras), in 1851, Lowe s.n. (BM, K). Ribeira d’Ametade, in 1849, Lowe s.n. (BM).
Canary Islands. — Rabacal, Moniz 743 (BM). La Caldera, in 1858, Lowe s.n. (BM).
It is problematical whether Epilobium obscurum occurs on the African mainland
or not. I have seen no material, but Haussknecht (Monogr. Epil. 117, 1884) cites a
specimen from near Bufarick, Algeria, Gandoger, and several other authors have given
references which might indicate the presence of Ihis common European species in
Algeria (for example: Batt. in Batt. & Trab., FI. Alger. 315, 1889; Letourneux, Etude
Bot. Kabylie du JurJura 42, 1871; Quezel & Santa, FI. Nouv. Alger 2 : 639, 1963).
Epibolium madeirense (“ maderense ” of Hausskn., Monogr. Epil. 232, 1884) seems
clearly to be a maritime form of this species with larger leaves and a somewhat lusher
growth habit. It is curious that Haussknecht (op. cit.) referred it to his group Capenses,
which otherwise comprises species here referred to the very different E. capense. There
appears to be no reason to continue to recognize this form taxonomically.
Epilobium obscurum is not known to grow with other species in Africa, although
it may do so. Hybrids with E. hirsutum, E. lanceolatum, E. parviflorum and E. tetra-
gonum subsp. tetragonum have been described from European material.
6. Epilobium atlanticum Litard. & Maire, Contrib. fitude FI. Grand Atlas (Arch.
Sci. Maroc) 2 : 3 (1930); Litard. & Maire, Mem. Soc. Sci. Nat. Maroc 26 : 15 (1930);
Jehandiez & Maire, Cat. PI. Maroc 2 : 515 (1932).
Epilobium samuelssonii Raven, Svensk Bot. Tidskr. 56 : 61, fig. 1 (1962). Type:
Spain, Prov. de Granada, Sierra Nevada, Rio Monachil, 2,700 m.s.m., 24 July 1950,
Roivainen (S).
323
Slender subglabrous herb, 3-5-28 cm tall; rhizomes slender, the lower part of the
stems abundantly provided with elongate (to 10 cm) leafy stolons, typically creeping
through cushions of moss, from which new plants arise; stems lightly strigulose above,
especially along the lines decurrent from the margins of the petioles. Leaves opposite,
alternate near the inflorescence, oblong-lanceolate, to narrowly ovate, 0-8-2 -4 x 0-3-1
cm, subentire or slightly denticulate, at times ciliate with microscopic pubescence,
subsessile or on a petiole up to 1-5 mm long, the apex broadly cuneate, the base obtuse.
Inflorescence densely white-hairy, few-flowered, nodding before anthesis but erect in
fruit; flowers erect in bud. Hypanthium broadly funnel-form, strigulose without,
c. 1 mm deep, c. 2 mm wide. Sepals 3 -8-4 -5 mm long, 1-2-1 -3 mm wide, very
narrowly deltoid, narrowly acute, more or less densely strigulose, especially near the
base. Petals 4-4-8 mm long, c. 4-5 mm wide, bright rose-purple, deeply notched.
Anthers c. 1-1-2 mm long, the filaments of the longer c. 15 mm long, of the shorter
c. 0-5 mm long. Style 4-7-6 mm long; stigma capitate, c. 1-8 mm thick, held above
the anthers at anthesis; ovaries densely white-strigulose, the pedicels shorter than the
subtending leaves. Capsules densely strigulose, 4^4-5 cm long, on a pedicel 3-4 cm
long; seeds brown, narrowly obovoid, attentuate, very minutely papillose, c. 1 - 5 mm
long, with a conspicuous pellucid beak; coma white, c. 5 mm long.
Geographical range (Fig. 3): Along streamlets and in moist cushions of moss,
2,700-2,900 m elevation, in the Haut Atlas of Morocco and the Sierra Nevada of
Spain.
Morocco. — Haut Atlas, in Valle Reraya, in scaturiginosis alpinis, in clivo septentrionali montis
Toubkal, secus rivulos convallis Immouzer, solo porphyrico, 2800-2900 m, in 1924, Emberger, Litardiere,
& Maire (MPU, holotype).
Epilobium atlanticum is an attractive and very distinctive species. In Africa, it
is known only from the type collection, but it is also found in similar habitats in the
Sierra Nevada of Spain. In 1962, unaware that E. atlanticum was the same, I described
the European populations as a new species, E. samuelseonii. As presently understood,
E. atlanticum adds yet another floristic link between the two mountain ranges where
it occurs (e.g. Quezel in Mem. Soc. Brot. 9 : 1-77, 1. 1—6, 1953). The stigmas of E.
atlanticum were said to be clavate by Litardiere and Maire (Mem. Soc. Nat. Moroc
26 : 16, 1930), but have been capitate in the material I have seen, which however includes
only a very small portion of the type collection. They are certainly capitate in the
collections from the Sierra Nevada, which are identical in all other respects.
Epilobium atlanticum is not known to grow together with any other species of
the genus.
7. Epilobium psilotum Maire & G. Samuels., Ark. Bot. Stockh. 29A, no. 11, 26,
t.3 (1939); Emberger & Maire, Cat. PI. Maroc 4 : 1081 (1941).
Entirely glabrous, slender, delicate, clumped perennial herb 6 • 5-25 cm tall, arising
from a slender creeping rhizome from the underground portions of which arise leafy
shoots, which are at first short and fleshy; stems with prominently raised lines decurrent
from the margins of the petioles. Leaves almost entirely opposite, alternate only in
the inflorescence, lanceolate to narrowly ovate, 0-9-2 -6 x 0-30-6 mm, obscurely and
sparsely serrulate, paler below, on a short but distinct petiole 1-2 mm long, the apex
acute, the base obtuse. Inflorescence short, nodding before anthesis but erect in fruit;
flowers erect in bud. Hypanthium 1-2-1 -8 mm deep, 1-2-1 -8 mm across. Sepals
3 -1-3 -8 mm long, 1-1-1 mm wide, very narrowly deltoid. Petals 5-8 mm long,
2 -5-3 -6 mm wide, light rose-purple, the notch prominent, 1-6-3 mm deep. Anthers
0-6-0 -9 mm long, the filaments of the longer 3-5 mm long, those of the shorter 1 -3-3-2
324
mm long. Style 3 -5-4 -3 mm long; stigma clavate, 2-2-1 mm long, 0-6-0 -8 mm
thick, surrounded by both sets of anthers at anthesis. Capsules 1-8-4 -5 cm long, on
a pedicel 1-2 cm long; seeds light brown, very finely papillose, 1-2-1 -4 mm long,
c. 0-4 mm thick, acute at both ends, not beaked; coma dull white, 4-5-5 mm long.
Geographical range (Fig. 3): Endemic to the Haut Atlas of Morocco, where it
occurs in wet places from 2,100-3,100 m elevation.
Morocco. — Tashdirt, 2100 m. Balls B3117 (S, holotype; isotypes, BM, E, K). North face of Djebel
Augour, c. 3100 m, Polunin 2200 (BM). G. Tistouit, 2850 m, Newbould 211 (BM).
As was pointed out in the protologue of the species, Epilobium psilotum is clearly
related to the North Temperate group of the species which Haussknecht assigned to
his group Alpina, but entirely distinct within this group. It is not known to grow
together with any other species of the genus.
8. Epilobium capense Buck, ex Hochst. in Krauss in Flora 27 : 425 (1844) ; Hausskn.,
Mon. Epil. 229 (1884).
Epilobium flavescens E. Mey. ex Harv. in FI. Cap. 2 : 506 (1862); Hausskn., Monogr.
Epil. 230 (1884); H. Lev., Ic. Gen. Epil. tt. 42, 43 (1910). E. bojeri Hausskn., Oester.
Bot. Zeit. 29 : 90 (1879); Hausskn., Mon. Epil. 231, t. XII fig. 60a, b (1884); H. Lev.
Ic. Gen. Epil. t. 44 (1910); H. Perr. in FI. Madagasc., Oenotheracees : 3, fig. 1 (1950).
H. biforme Hausskn., Monogr. Epil. 230 (1884); H. Lev., Ic. Epil. t. 41 (1910). E.
jonathum Hausskn., Monogr. Epil. 231, t. X fig. 56a (1884); H. Lev., Ic. Gen. Epil.
t.45 (1910). E. species No. 2, Brenan in FI. Trop. E. Afr., Onagraceae: 4, fig. 1
(1953).
Perennial herb, 0-1-1 -2 m tall; underground stems vertical or nearly so, densely
invested with thick white, fleshy, rounded scales c.4 mm long and 2-10 mm wide after
first year of growth, the new shoots scaly and arising from this region; plants strigulose
with some glandular trichomes in the more densely pubescent inflorescence; stems with
weakly marked elevated lines decurrent from the margins of the petioles. Leaves
opposite near the base, alternate above, the margins and veins and sometimes the
entire surface finely strigulose, narrowly ovate to narrowly lanceolate, mucronate to
long acuminate at the apex, rounded to obtuse or more rarely subcordate at the base,
weakly or more often coarsely serrate with prominent forward-directed teeth, especially
in the upper half, 2-5 cm long, 0-4-2 -5 cm wide; petioles short but distinct, 1-2-5
mm long. Inflorescence densely strigulose with an admixture of glandular trichomes,
erect in bud; flowers erect or somewhat drooping in bud. Hypanthium 2-2-5 mm
across, 1-1-1 -5 mm deep, usually long-ciliate at the mouth with trichomes mostly
0-6-0 -8 mm long. Sepals 4-2-10 X 1-2-2 -5 mm, narrowly oblong, acuminate or
apiculate. Petals obovate, 6-16 mm long, 3-10-5 mm wide, bright rose-purple, paler
purplish, creamy, or white, especially in smaller-flowered forms, the notch about 1/5 the
length of the petal. Anthers white, 1-3-2 -5 mm long, c. 0-8-1 mm thick; filaments
pale rose to white, those of the longer stamens 2-5-8 mm long, of the shorter ones
1 -5-5 mm long; pollen yellow. Styles pale rose to white, glabrous, 5-15 mm long;
stigma white, 4-lobed, usually deeply so, the lobes 0-7-2 mm long, papillate and
receptive within, held far above the anthers in larger-flowered plants but reached by
them in smaller-flowered ones. Capsules densely strigulose with admixture of glandular
trichomes, erect, 3-9 cm long, on a pedicel 1-6 cm long; seeds brown, oblong-obovoid,
densely papillose, 1 -3-1 -6 mm long, the coma c. 5-7 mm long, dingy white. Gametic
chromosome number, n = 18.
325
Fig. 5. — Epilobium capense. 1, from Madagascar, corresponding to E. bojeri ( Humbert 11821, in P).
2, cultivated from seed from Swaziland, corresponding to E. jonantkum, E. capense sensu stricto
( Compton 30558, in DS). 3 and 3a, from the Cape Province, corresponding to E. flavescens (Tvson
1813, in SAM).
326
Geographical range (Fig. 2): In moist places and mountain meadows, from
southern Tanzania south and west to the vicinity of Cape Town; mountains of central
Madagascar. 900-2,000 m elevation.
Tanzania. — Southern Highlands District, Mbeya to Tukuyu Road, where it descends steeply to
Mwakaleli, Napper 1182 (K). Elton Plateau, Ngjombe Dist., 2700 m, Procter 1604 (EA; K). Kigoro
Forest Reserve, Mufindi, Iringa Dist., Michelmore 935 (K).
Zambia. — Between Mts. Kongula and Kangampanda and the Malawi border near Mt. M’Wanda,
high moorland on granitic soil, in 1932, Temperley s.n. (BM). Nyika Plateau, 2100 m, Lees 102 (K).
Malawi.- — Damp ground near streams, Nyika Plateau, 1350 m, Benson 1382 (BM). Dedza Moun-
tain slopes, Banda 285 (BM). Ncheu Dist., Lower Kirk Range, Chipusiri, 1460 m, Excell, Mendonca
& Wild 965 (BM; LISC, SRGH).
Rhodesia. — About 3 km west of Mt. Inyangani, Inyanga Dist., c. 2000 m. Fries, Norlindh & Weimarck
3479 (BR). Umtali Dist., South Patrol, Stapleford Forest Reserve, 1500 m, Chase 5883 (BM, K,
SRGH). Chimanimani-Stonhenge, Melsetter Dist., 1700 m, Phipps 376a (BR).
South Africa. — Cape. — Ceres: Koude Bokkeveld, in monte Tafelberg, 1500 m, Schlechter 10096
(BOL, GRA), Engcobo: near Engcobo, in 1896, Bolus s.n. (BOL). George: Knysna River, George,
in 1840, Krauss s.n. (LZ, holotype of E. capense, destroyed; G). Herschel: Majuba Nek, near Sterk-
spruit, Hepburn 242 (GRA). Mount Currie: near Kokstad, 1500 m, Tyson 1813 (E, SAM). Murraysburg:
Murraysburg, Tyson 1 14 (BOL). Somerset : Boschberg near Somerset East, 1050 m, MacOwan 729 (type
collection of E. biforme; BM, BOL, G, GRA, K, NH, P, SAM). Stellenbosch: Biesie Vlei, Sonkang-
hoek, Hamson 3124 (BM). Stockenstrom : Hogsback, Barker 914 (NBG). Tulbagh: Sneeuwgat
Valley, Great Winterhoek, Phillips 1740 (SAM). Umzimkulu: near Clydesdale, 2500 ft., Tyson 2012
(BOL). Orange Free State.— Ficksburg: Strathcona, Isaac 118 (NBG). Harrismith: Bester’s Vlei,
Harrismith, 5400-6200 ft, Bolus 8161 (BOL). Orange Free State, Cooper (K, holotype of E. jonanthum;
isotype, BOL). Natal. — Alfred: Rooi Vaal, Harding, Jacques 5233 (NBG). Bergville: Oliviershoek
Pass, Gillett 1163 (NH). Camperdown: Umlaas, Krauss 154 (lectotype of E. flavescens, BM, G, K).
Inanda: Medley Wood 868 (BM, BOL, GRA, K, NH). Newcastle: near Newcastle, 900-1200 m,
Medley Wood 6752 (E). Swaziland. — Forbes Reef Road, Mbabane Dist., 1200 m, Compton 25662
(NBG, PRE). Mankaiana, Mankaiana Dist., 1200 m, Compton 27493 (NBG). Transvaal. — Johan-
nesburg: Craighall, Moss 4027 (J). Lydenburg: near Lydenburg, Atherstone 174 (K). Potgietersrus :
Palala River, in 1918, Breyer s.n. (PRE). Soutpansberg: Louis Trichardt, Breyer in 1919 (PRE).
Ventersdorp: Klerkskraal, 1400 m, Sutton 730 (PRE). Wakkerstroom : Oshoek, 2000 m, Devenish
375 (BM).
Malagasy Republic. — An karat ra Mts., Hildebrandt 3869 in part (BM, JE, K). Mt. Iarambao,
north of Inanatona, basin of the Andrantsay, Perrier 6623 (P). Manambolo, vicinity of Isomono,
Mt. Morahariva, Humbert 13213 (P). Mt. Itrafanaomby (Ankazondrano), haut Mandrare, Humbert
13477 (P). Near Ambositra, Perrier 14622 (P). Near Bezong zong, Feb.-Mar., Bojer (JE, lectotype
of E. bojeri ).
The 4-lobed stigma of E. capense , taken together with its distinctively serrate
leaves and dense-set fleshy scales on its underground stems make it unmistakable.
Haussknecht (Monogr. Epil. 229-232, 1884) considered the material he saw to belong
to five species, differing primarily in leaf-width and flower colour. Together with
E. madeirense, which is here considered a synonym of E. obscurum, these five species
comprised his section Capenses. With the more abundant material available at the
present day, it is impossible to find any subdivisions within his species recognizable
on the basis of leaf-shape. Larger-flowered and presumably more highly outcrossed
populations of this species normally have more or less erect bright rose-purple flowers
and often, particularly in South Africa, relatively narrow leaves (Fig. 5: 2). Smaller-
flowered, presumably more highly autogamous populations, normally have cream
or white flowers, which are often nodding and relatively broad leaves (Fig. 5: 3 and
3a). But these correlations do not hold consistently, and specimens representing
all possible combinations of these characteristics have been seen. Epilobium bojeri
(Fig. 5: 1), considered by H. Perrier de la Bathie (FI. Madagasc., Oenotheracees :
4, 1950) to be endemic to Madagascar, is matched perfectly by numerous collections
from the mainland of Africa; such plants are relatively short, with crowded leaves.
327
Fig. 6. — Epilobium capense. 1, detail of underground parts, showing imbricated fleshy scales, from
Njombe District, Tanzania ( Proctor 1809, in EA). 2, inflorescence, from Mbeya Range. Tanzania
{Procter 1618, in EA). These specimens correspond to Brenan’s (1953) Sp. No. 2, and are from
the north end of the range of the species.
328
Some plants of this species from the northern end of its range, especially those
from Tanzania, have nodding flowers with purple rose-purple petals; relatively broad,
rounded, more weakly serrulate leaves; and sometimes less deeply divided stigmas
(Fig. 5:1). These plants were treated by Brenan (FI. Trop. E. Afr., Onagraceae: 4,
1953) as an unnamed species, No. 2. It appears possible that some of the charac-
teristics of these populations may have resulted from historical introgression between
E. stereophyllum and E. capense. At present, however, the ranges of the two species
are separated by a gap of more than 300 miles. Taken as a whole, the northern popula-
tions of E. capense appear to have no distinctive combination of characteristics that
would allow their segregation as a geographical entity.
In summary, there appear to be no units within E. capense the formal recognition
of which would shed greater light upon the complexities of the group. Indeed, treating
the complex as a single, variable species seems much more informative.
The gametic chromosome number, n = 18, was determined from a single population
grown at Stanford, collected June, 1964, the seeds from Compton 30558 (DS), Mbabane
Dist., Swaziland. These plants had relatively erect, large, purple flowers (Fig. 5 : 2).
Epilobium capense has been collected growing together with E. hirsutum, E. salignum
and E. tetragonum subsp. tetragonum in South Africa; a hybrid between it and the
first-mentioned species is discussed under E. hirsutum. Hybrids with the other two
species are not known.
9. Epilobium stereophyllum Fres. in Mus. Senckenberg 2 : 151 (1837); Oliv. in
FI. Trop. Afr. 1 : 487 (1871); Hausskn., Monogr. Epil. 233 (1884); H. Lev. Ic. Gen.
Epil. t.47 (1910); Brenan in FI. Trop. E. Afr. Onagraceae: 4, fig. 1 (1953); Hedberg
in Symb. Bot. Upsal. 15, 1 : 133 (1957).
Epilobium fssipetalum Steud. ex A. Rich., Tent FI. Abyss. 1 : 273 (1848); Hausskn.,
Monogr. Epil. 234 (1884); H. Lev., Ic. Gen. Epil. t.50 (1910). E. cordifolium A. Rich.,
Tent. FI. Abyss. 1 : 274, t.50 (1848); Oliv. in FI. Trop. Afr. 2 : 487 (1871); Hausskn.,
Monogr. Epil. 233 (1884); H. Lev., Ic. Gen. Epil. t.48 (1910). E. schim-
perianum Hochst. ex A. Rich., Tent. FI. Abyss. 1 : 272 (1848); Oliv. in FI. Trop. Afr.
2 : 487 (1871); Hausskn., Monogr. Epil. 232 (1884); H. Lev., Ic. Gen. Epil. t.46
(1910). E. kilimandscharensis H. Lev. in Bull. Herb. Boiss., ser. 2, 7 : 589 (1907);
H. Lev., Ic. Gen. Epil. t.49 (1910). E. kiwuense Loes. in Mildbr., Wiss. Ergebn. Deutsch.
Zentr.-Afr.-Exped. 1907-8, 2 : 588 (1913); Robyns, FI. Sperm. Parc. Nat. Albert
1 : 682, t.lxxii (1948). E. neriophyllum subsp. ellenbeckii Engl, in Veg. Erde lx. Pflanzenw.
Afrikas 3, 2 : 773 (1921). E. stereophyllum var. kiwuense (Loes.) Brenan in Kew Bull.
8 : 163 (1953); Brenan in FI. Trop. E. Afr., Onagraceae: 5, fig. 1 (1953).
Perennial herb, 0-25-1 -5 m tall, from a long rhizomatous base from which leafy
shoots arise; plants puberulous or short-pubescent with spreading hairs, with elevated
pubescent lines running down from the sides of the petioles, the stems often thick
and hollow. Leaves mostly opposite, alternate in and near the inflorescence, the margins
and veins and sometimes the entire surface finely strigulose, narrowly ovate, acute, cordate
or more rarely to obtuse at the base, sparsely serrulate, 1-5-6 -5 cm long, 0-3-2 -7
cm wide; petiole short but distinct, up to 2-5 mm long. Inflorescence densely strigulose
or spreading pubescent, sometimes with a strong admixture of glandular trichomes,
erect in bud; flowers somewhat nodding in bud and when they first open, later erect.
Hypanthium 1-8-3 -5 mm across, 1-2-5 mm deep, its mouth glabrous or with a few
hairs. Sepals 5-5-10 mm long, 1 -8-2-8 mm wide, apiculate. Petals obovate, 6-16
mm long, 4-1 1 mm wide, rose-purple, the notch deep, about 1/6 of the length. Anthers
1 -2-2-3 mm long; filaments of the longer stamens 3-5-7 mm long, those of the shorter
2-4 mm long. Styles 5-9-2 mm long; stigma obovoid-clavate, 1 -5-4 mm long, 1 -2-3
329
mm thick, the longer anthers surrounding it at the base. Capsules densely strigulose
or spreading-pubescent, erect, 3 -5-8 -5 cm long, on a pedicel 1-9 cm long; seeds 1-1 -2
mm long, 0-4-0 -6 mm thick, oblong-obovoid, obutse at the base, finely papillose,
brown to black, the coma 4-8 mm long, white. Fig. 7:1.
330
Geographical range (Fig. 2): Common in the mountains of Ethiopia, from the
vicinity of Lake Tana southward, and south in the principal mountain masses of
equatorial Africa: eastern Republic of the Congo and Rwanda; Ruwenzori, Yirunga
Volcanoes, Elgon, Cherangani Hills, Aberdares and Kinagop Plateau, Mt. Kenya,
Mt. Kilimanjaro and the Pare Mountains; 1750-3660 m elevation. Usually in moist
places in open mountain grassland.
Ethiopia. — Daugila, Lake Tana region, 2000 m, in 1926, Cheesman s.n. (BM). Mt. Buahit, Schimper
1 348 (P, holotype of E. fissipetalum; isotypes, BM, G, K, M). Northeast side of Mt. Buahit, 4300 m,
Pichi-Sermolli 2703 (BR, FI, K). Gara mulata Mt., 3000 m, 9° 15' N, 41° 45' E, Gillett 5374 (K).
Mt. Silke, Schimper 972 (type of E. schimperianum , BM, K, P). Choke Mts., Gojjam, vicinity of the
upper Ghiedeb Valley, 10° 40' N, 37° 50' E, at Arat Makere near Mt. Talo, c. 3600 m, Evans 500
(BM, K). Eastern slopes of Mt. Delo, Amaro Mountains, 5° 48' N, 37° 54' E, 2500-3000 m, Gillett
14909 (BR, EA, FI, K). Between Temben and Semen, Ruppell (FR, holotype of E. stereophyllum).
Ethiopia, Quartin Dillon & Petit (P, holotype of E. cor di folium).
Republic of the Congo. — Orientale, Nioka (Route Mawa), 1800 m, Taton 1100 (BR). Kivu, Kishi,
Terr. Lubero, 2000 m, Christiaensen 1079 (BR). Kivu, Numbi, Terr. Kalehe, 2200 m, Leonard 4570
(BR).
Rwanda .—Mildbraed 722, 1555, 1646 (B, destroyed; syntypes of E. kiwuense). Terr. Shangugu,
Route Bukavu-Astrida, environ d’Uwinka, 2400 m, Troup in 9782 (BR, EA). Dorwa, Biumba
Becquet 190 (BR).
Uganda. — Ruwenzori, 2850 m, Purseglove P238 (K). Toro Dist., Ruwenzori, Namwanba Valley,
3150 m, Taylor 2911 (BM). Virunga Volcanoes, Kigezi Dist., saddle between Muhavura and Mgahinga,
c. 3000 m, Stauffer 653 (K, PRE). Elgon, edge of stream just above Bulambuli, 2700 m, Liebenberg
1634 (K).
Kenya. — Elgon, eastern slope above Japata estate, along the Saum River, 3500 m, Hedberg 1000
(K). Cherangani Hills, 2400 m, McLoughlin (PRE). East Aberdares, Kimakia Forest Reserve, 2500 m,
Kerfoot 617 (EA). Kinagop Plateau, Loreko, 3000 m, Napier 1248 (K). Mt. Kenya, northwest slopes,
3150 m, Synge S1762 (BM).
Tanzania. — Kilimanjaro, southeast side, 3200 m, Schlieben 4879 (BM, BR, LISC, M, P). Kili-
manjaro, 2700 m, Volkens 1847 (G, holotype of E. kilimanscharense; isotypes, BR; E; K, photograph).
South Pare Mts., between Tona and Muhasi swamp, Peter K.661 (O.III.142) (K).
Epilobium stereophyllum is extraordinarily variable in pubescence, leaf shape and
size, and flower size, particularly in Ethiopia. In the mountains of equatorial Africa,
plants with longer, spreading pubescence (“ var. kiwuense ”) are generally found below
2,500 m elevation, those with shorter, appressed pubescence above this level. As this
is only one of the several obvious parameters of variation, and is not useful in classifying
populations found in Ethiopia, I prefer not to accord formal taxonomic recognition
to these two entities. Some of the specimens referred here are difficult to distinguish
from E. salignum, although the two species are normally totally distinct. It is likely
that hybridization between them contributes to the local blurring of the distinctions
between them. Between the southernmost stations for E. stereophyllum in the South
Pare Mountains of Tanzania and the northernmost for E. capense in the Porotos is
a lowland gap of more than 300 miles; yet the northern populations of the latter are
more similar to E. stereophyllum than are the southern populations of the same species.
Possibly past introgression has occurred between the two species, which are obviously
relatively closely related.
Epilobium neriophyllum subsp. ellenbeckii is referred doubtfully here. It was
based on material from “Arussi-Galla-Land bei Jidah um 2,600 m This collection,
which was doubtless in the herbarium at Berlin, was presumably destroyed in World
War II, and no duplicates have come to my attention. The sub-species was contrasted
with E. neriophyllum (= E. salignum) by its longer, cylindrical style; short petioles;
and slightly toothed leaves 2-4 cm long and 4-5 mm wide. It is doubtful that it is
E. salignum , as Brenan (in Turrill & Milne-Readhead, FI. Trop E. Afr., Onagraceae:
6, 1953) considered it, for that species is not known closer to Jidda than the vicinity
of Mt. Elgon, more than 600 miles to the south.
331
10. Epilobium saligmim Hausskn. in Oester. Bot. Zeitschr. 29 : 90(1879); Hausskn.,
Monogr. Epil. 236, t. XII fig. 62a, b, c (1884); H. Lev., Ic. Gen. Epil. t.54 (1910);
Perr. in Not. Syst. 13 : 138 (1947); H. Perr. in FI. Madagasc., Oenotheracees: 4
(1950); Brenan in FI. Trop. E. Afr., Onagraceae: 5, fig. 1 (1953); Brenan in Hutch.
& Dalz., FI. W. Trop. Afr., ed 2, 1, 1 : 166 (1954); Fernandes & Fernandes in Garcia
de Orta 5:111 (1957); 7 : 493 (1959).
Epilobium neriophyllum Hausskn. in Abh. Naturw. Verein Bremen 7 : 19 (1880);
Hausskn., Monogr. Epil. 236, t. XII, fig. 61a (1884); H. Lev., Ic. Gen. Epil. t.55 (1910);
Robyns, FI. Sperm. Parc Nat. Albert 1 : 682 (1948). subsp. benguellense (Welw.
ex Hiern) Engl, in Veg. Erde IX. Pflanzenw. Afrikas 3, 2 : 773 (1921). subsp.
benguellense var. welwitschii Engl, in Veg. Erde IX. Pflanzenw. Afrikas 3, 2 : 773
(1921). subsp. cylindro stigma Engl, in Veg. Erde IX. Pflanzenw. Afrikas 3,
2 : 773 (1921). E. mundtii Hausskn., Monogr. Epil. 235, t. XI fig. 58a (1884); H.
Lev., Ic. Gen. Epil. t.52 (1910). E. natalense Hausskn., Monogr. Epil. 235, t. X fig.
57a (1884); H. Lev., Ic. Gen. Epil. t.51 (1910), probably. E. oliganthum Bak. in J.
Linn. Soc. Bot. 21 : 345 (1886); non Michaux, FI. Bor. Amer. I : 223 (1803). E.
benguellense Welw. ex Hiern, Cat. PI. Welwitsch, Dicot. 1 : 378 (1898). E. madagas-
cariense H. Lev. in Fedde, Rep. Nov. Sp. 4 : 225 (1907); H. Lev., Ic. Gen. Epil. t.56
(1910). E. schinzii H. Lev. in Fedde, Rep. Nov. Sp. 4 : 225 (1907); H. Lev., Ic. Gen.
Epil. t.53 (1910). E. perrieri H. Lev. in Rev. Geogr. Bot. 27 : 3 (1917); H. Perr. in
Not. Syst. 13 : 139 (1947); H. Perr. in FI. Madagasc., Oenotheracees: 4, fig. 1 (1950).
E. palustre sensu Ridley, The Dispersal of Plants throughout the World 152 (1930);
non L., Sp. PI.: 348 (1753).
Perennial herb, 0-2-1 -6 m tall, the stems sometimes persistent and subligneous;
strongly rhizomatous, the rhizomes lacking scales, long-spreading and giving rise to
new leafy shoots; plants evenly strigulose, sometimes sparsely so with faintly marked
lines running down from the margins of the pedicels. Leaves mostly opposite or sub-
opposite, alternate in the inflorescence, the margins and veins and sometimes the entire
surface finely strigulose, very narrowly to narrowly elliptic, rarely almost lanceolate,
acute, rarely acuminate, narrowly cuneate to attenuate at the base, very rarely rounded,
weakly serrulate or rarely serrate to near the base, 2-8 x 0-3-2 cm; petiole 1-8 mm
long, distinct. Inflorescence densely strigulose, erect in bud, to 3 dm long, the leaves
usually not much reduced; flowers nodding in bud and when they first open, later
erect. Hypanthium 0 • 8-2 mm across, 1 • 5-2 • 3 mm deep, its mouth glabrous or long-
ciliate. Sepals 3 • 5-8 • 5 mm long, 1 • 2-2 • 2 mm wide, apiculate. Petals narrowly
obovate, 5-15 mm long, 2-7 mm wide, at first white or cream, then rose following
pollination, the notch 0-5-2 mm deep. Anthers brownish, fading rose, 0-8-2 mm
long; filaments white, those of the longer stamens 2-5-8 mm long, those of the shorter
1-5-5 mm long. Styles white, 4-2-10 mm long; stigma white, usually clavate, rarely
subcapitate, 1 -8-4 mm long, 1 -2-1 -8 mm thick, the longer anthers usually just reaching
the base of the stigma. Capsules densely strigose, erect, 3-7 cm long, on a pedicel
0-8-4 -5 cm long; seeds light brown or tan, oblong-obovoid, obtuse at the base, very
minutely pitted, 1-1-35 mm long, 0-35-0-6 mm thick, the coma c.5-9 mm long,
copious, white. Gametic chromosome number, n — 18. Fig. 7 : 2.
Geographical range (Fig. 3): Moist places, 500-3,000 m elevation, from the eastern
Republic to the Congo and the vicinity of Mt. Elgon in Uganda south in the mountains
of east Africa to the eastern Cape Province; also in Cameroons and Benguela, Huambo,
Huila and Bie in Angola; throughout Madagascar.
Nigeria. — Cameroons, in marsh on the bank of River Bui, Banso, Bameda, Tamajong FHI 23474
(K).
Cameroun. — Dschang Meletan, 2500 m, Saxer 164 (K). Bamboutos, 1900 m, Lepesme, Pautian
& Villiers 36 (P).
332
Republic of the Congo. — Orientale, Nioka, marais affluent Korda, Taton 1074 (BR, K, P). Kivu,
Kahuzi saddle on the road from Bukavu to Walikale, c. 2300 m, Stauffer 1110 (K). Kivu, Bwit, Kikuyu,
Katwe, 1600 m, Deru 239 (BR). Katanga, envir. Elisabethville, in 1923, DeGiorgi s.n. (BR). Katanga,
Parc National de l’Upemba, De Witte 3540 (BR). Katanga, Kipiri, Marungu Mts., 2200 m, Dubois
L.1367 (BR).
Rwanda. — Kamiranjovu, Terr. Shangugu, 2000 m, Christiaensen 1412 (BR).
Burundi. — Dans le marais des bords du lac Kanzigiri, Elskens 169 (BR).
Uganda. — Near Bwamba Pass, E. Ruwenzori, Toro Dist., 2250 m, Eggeling 4031 (K). Kigezi
Dist., 1 mile north of Mpalo, 1800 m, Norman 79 (EA). Bugishu Prov., Mt. Elgon, Butandiga, 2100 m,
Synge S817 (BM).
Kenya. — N.E. Elgon, swamp on Kisano River, 2400 m, Tweedie 1194 (K). Cherangani Mts.,
1° 5' N, 32° 26' E, 3000 m, Maas Geesteranus 6334 (BR, COI, G, K, PRE, more or less intermediate
to E. stereophyllum). Aberdare Range, Ol Bolossat, Meinertzhagen AH9343 (EA). Limoru, Nairobi,
2100 m, Diimmer 1643 (BM, K, PRE).
Tanzania. — Ujamba, Kigoma Dist., Mahinda HSM1184 (EA, K). Foothills at Mt. Mmemya,
Ufipa Dist., Bullock 3701 (BR, K). Kikondo Camp, Poroto Mts., 1960 m, Richards 13993 (K). Kigogo
Forest Reserve, Mufindi, Iringa Dist., Eggeling 6814 (BM, BR, EA, FI, K, LISC, P, PRE, SRGH);
cultivated material from this collection, grown at Kew, collected 29 June, 1955, (K). West side of
Uluguru Mts., Lukwangule Highland, vicinity of Morogoro, 2400 m, Schlieben 3515 (BM, BR, G,
LISC, M, P).
Angola. — Benguela, between Ganda and Caconda, 1700 m, Hundt 840 (BM, DS, P). Huambo,
Dist. de Nova Lisboa, Perimetro Florestral de Sacaala, junto as rio Curimaala, Murta 107 (COI).
Huila, Huila cataract, 1700 m, Exell & Mendonga 2852 (BM, COI, LISJC). Huila, margins of the
river Monino, Wehvitsch 4458 (LISU, lectotype of E. benguellense, A. & R. Fernandes in Garcia de
Orta 7 : 12, 1959; isolectotypes, BM, COI, G, K). Huila, margins of the rio Caculovar, Quilengues,
Wehvitsch 4459 (type of E. neriophyllum subsp. benguellense var. welwitschii; BM, COI, G, LISU,
P). Bie, Menongue, margins of the rio Cutato-Cubango, near Sa Macaca, Gossweiler 3818 (BM,
COI, LIJSC).
Zambia. — Abercorn Dist., Saisi River Marsh, 1500 m, Richards 8381 (K). Lundazi Dist., Nyika
Plateau, 2100 m, Richards 10416 (K). Solwezi, 1350 m, Robinson 3461 (BR, K, M, SRGH); cultivated
at Kew, Aug. -Sept., 1961, K.5208 (K). Mochipapa, 5 miles east of Choma, 1300 m, Robinson 1279
(BR, K, SRGH).
Malawi. — Kondowe to Konongu, 600-1800 m, in 1896, Whyte s.n. (K). Ncheu Dist., Lower Kirk
Range, Chipusiri, 1460 m, Exell, Mendonca & Wild 967 (BM, LISC, SRGH). Zomba Plateau, Zomba
Dist., 1500 m. Brass 16314 (BM, BR, K, SRGH, US).
Mozambique. — Niassa, Vila Cabral, Metonia, Torre 97, 98a (COI, LISC).
Rhodesia. — Prince Edward Dam, Salisbury, 1400 m, Eyles 8943 (SRGH). Iona Farm, Melsetter
Dist., Chase 4819 (BM, PRE, SRGH). Matopos, Rogers 5257 (SRGH). “ Southern Rhodesia, Epi-
lobium palustre, L. (Introduced from Europe) ”, Hislop 147 (K).
South Africa. — Cape. — Bizana: 3 miles northwest of Fort Donald, between Emagushine and
Bizana, Lewis 4475 (SAM). Komga: Komga, 500 m, Flanagan 665 (E, K, PRE). Queenstown;
Katberg, Galpin 2075 (GRA). Somerset East: Boschberg, 900 m, MacOwan 1487 (K, chosen lectotype
of E. neriophyllum by Brenan in FI. Trop. E. Afr., Onagraceae: 6, 1953; GRA, NH). Stutterheim:
3 miles from Amabele, De Vries 112 (PRE). Umtata: Umtata, 1040 m, Schlechter 6537 (BOL).
Umzimkulu: Clydesdale, 750 m, Tyson 2012 (BOL, K, PRE). Xalanga: Cala Kebelo Kloof, Pegler
1694 (PRE). Without definite locality, ad promont. bonae spei, in 1820, Mundt & Maire s.n. (JE,
lectotype of E. mundtii ). Orange Free State. — Bethlehem: Golden Gate, Compton 22488 (NGB).
Natal. — Kliprivier: Van Reenen, 900 m, Medley Wood 13101 (PRE). New Hanover: Noodsberg,
Noodsberg, o00-900 m, Medley Wood 5282 (Z, holotype of E. schinzii, not seen; isotypes, BM, NH,
SAM). Umzinto: Dumisa, 850 m, Rudatis 1346 (BM, E, K). Underberg: Underberg, McClean
662 (NH, PRE). Natal Bay, Gueinzius (LZ, holotype of E. natalense, destroyed; isotype, JE).
Transvaal. — Barberton: Reimers Creek near Barberton, Thorncroft 1069 Heidelberg: Vandeleur
in 1901 (BM). Letaba: 1050 m, Scheepers 625 (K, M, PRE, SRGH). Rustenburg: 3 miles south
of Breedts Nek, 1200 m, Story 957 (PRE).
Swaziland. — Mbabane: Fyfes Swamp, 1050 m, Compton 24826 (NBG). Mankaiana: Mankaiana
Swamp, 1200 m, Compton 27491 (NBG).
Malagasy Republic.— Mantagne d’Ambre, Perrier (P). Bassin superieur du Sambirano, foret de
Besantribe, c. 1200 m, Humbert 18707 (K, P). Ankaratra, 200 m, Perrier 6630 (P, lectotype of E. perieri).
Ankaratra Mts., Hildebrandt 3869 in part (BM, JE, K, M, P, US). Sambiana, north of Antsirabe,
Waterlot 754 (P). Bassin de P Itomampy, col d’Andriteny, Humbert 6857 (P). Antananarivo and
Be’zongzong, Bojer (P, lectotype of E. salignum, Brenan in FI. Trop. E. Afr., Onagraceae: 5, 1953;
K, photograph). Tanarivou, prov. Emirina, in 1830, Bojer s.n. (JE). Central Madagascar, Barron
2269 (K, type of E. oligodontum). Ranobe, near Sirabe, in 1895, Forsyth-Major s.n. (G, holotype of
E. madagscariense).
333
Epilobium salignum is somewhat variable, and a number of its variants have been
given names. Some specimens from Madagascar, for example, are relatively low in
habit and have somewhat narrower leaves than usual ; these have been called E. perrieri.
In his treatment of Onagraceae for the Flore de Madagascar, H. Perrier de la Bathie
(op. cit. : 2, 1950) stated that such plants had “Stolons tres greles et tres longs”,
whereas plants of E. salignum are said to have “ Stolons nuls, mais base des tiges
radicante ”. I have been unable to substantiate this distinction with the available
material; all plants that I have seen have long, somewhat woody, rhizomatous bases,
from which arise leafy shoots after the flowering and fruiting cycle is completed.
Some of the plants of this species from South Africa (e.g., Flanagan 665 BOL,
GRA, PRE; McClean 913, PRE; Mogg 6830, PRE; Tyson 2012 E, PRE) have rather
distinctly serrulate and often somewhat hairy leaves; they appear to constitute a rather
distinctive local variant. The type collections of E. mundtii, E. natalense and E. schinzii
represent plants of this sort.
As herein constituted, therefore, Epilobium salignum comprises Haussknecht’s
entire group Anomalifoliae. A few collections made where the range of this species
overlaps with that of E. stereophyllum are assignable with difficulty to one species
or the other, and it is likely that hybridization between them may contribute to this
local blurring of the boundaries between what are otherwise very distinct species. In
South Africa, E. salignum grows sympatrically with E. capense, E. hirsutum and E.
tetragonum subsp. tetragonum, but hybrids have not been observed.
The gametic chromosome number of this species was determined in two collections
grown at Stanford: Raven 19737 (DS), June, 1963, the seeds from Procter 2078 (K),
Little Ruaha River, Sao Hill, Iringa, Tanzania, Jan., 1962; Raven 20339 (DS), June,
1964, the seeds from Compton 30554 (DS), Forbes Reef, Mbabane Dist., Swaziland,
3 March, 1961.
Doubtful and Excluded Species
Epilobium angustifolium L., Sp. PI. 347 (1753). This widespread circumboreal
species was reported by Haussknecht (Monogr. Epil. 39, 1884) from Africa “ in ins.
Madera leg. Forster, et in ins. Canar. ad Pic de Teneriffa alt. 1 1 000’ s. m. repertum ”.
I have not seen any material from these islands and have been unable to discover any
further record of the presence of this species on them in the literature. Unless additional
or authentic collections be discovered, the occurrence of this species in Africa must
be regarded as extremely doubtful.
'
335
Bothalia 9, 2 : 335-339.
A New Species of Streptocarpus (Gesneriaceae)
by
O. M. Hilliard
Bews Botanical Laboratories, University of Natal
Streptocarpus trabeculatus Hilliard, sp. nov., ex affinitate S. denticulati Turrill et
S. cooksonii B. L. Burtt, ab ambobus folio crassiore subtus insigniter viscoso, ab hac
corollae tubo angustiore distinguitur.
Herba acaulis, monocarpica. Folium solitarium, oblongum, 400 mm x 160 mm
usque, basi obtusum, apice marcidum, pagina supra laete viridi scabridula venis impressis
rugosa, subtus pallidiore pilis glandulosis et eglandulosis dense obsita venis valde promi-
nentibus, costa percrassa ex hypocotylo brevi crasso abeunte. Inflorescentiae cymosae,
plures in serie prorsum ex apice hypocotyli et basi costae orientes; flores multi simul
aperti. Pedunculi 20-100 mm longi, validi, rubro-brunnei vel virides, pilis glandulosis
et eglandulosis pubescentes. Bracteae ellipticae, 5x3-5 mm usque, virides, pube-
scentes. Pedicellus c. 6 mm longus, sub fructu fere duplo longior, rubro-brunneus
vel viridis, pubescens, pilis parcis glandulosis. Flos 25 mm usque longus. Calyx
quinquepartitus, segmentis lanceolatis 3x1 mm viridibus pubescentibus. Corolla
bilabiata; tubus anguste subcylindricus, c. 8 mm longus, basi 4 mm diametro, medio
abrupte deflexus et angustatus, fauce ampliatus 4-5 mm diametro, ex albo ad pallide
violaceum, pilis glandulosis et eglandulosis instructus ; limbus obliquus, lobis superior-
ibus divergentibus suborbicularibus 4x3-5 mm, labio inferiore 10-13 mm longo
lobis suborbicularibus 6x4-5 mm; lobi omnes pallide violacei, intus glandulosi,
extra pallidiores parce pubescentes; labium inferius basi et tubo ventraliter albo striis
intense violaceis. Stamina fertilia 2, filamentis c. 5 mm supra corollae basin orientia,
3 mm longa, medio curvata et ampliata, alba intense violaceo suffusa, parce glandulosa;
antherae applanatae, 1 mm longae, violaceae, connectivo albo parce glanduloso.
Staminodia 3, posterius minutum, lateralia 1 mm longa, alba. Discus annularis.
Ovarium 5 mm longum, viride, pubescens; stylus teres, 5 mm longus, glanduloso-
pubescens; stigma stomatomorphum, papillosum, album. Capsula bivalva, spiraliter
torta, 16-36 (25 ± 4) mm longa, 3 mm diam., styli basi persistente prominente.
Type: Natal, Port Shepstone District, Plains, Izotsha Falls, c. 450 m., Hilliard
2750 (NU, holo. ; NH, iso.).
Stemless, monocarpic herb. Leaf 1, oblong, up to 400 x 160 mm, base obtuse,
apex withered, upper surface rugose, light green, scabridulous, veins impressed, lower
paler densely pilose with glandular and eglandular hairs, veins very prominent, midrib
very thick and merging into the short thick hypocotyl. Inflorescences cymose, several
in acropetal succession from base of midrib and apex of hypocotyl, crowded, many
flowers expanded simultaneously. Peduncle 20-100 mm long, stout, reddish-brown or
green, pubescent with glandular and eglandular hairs. Bracts elliptic, up to 5 x 3-5
mm, green, pubescent. Pedicel c. 6 mm long, almost doubling in length in fruit, reddish-
brown to green, pubescent, with occasional glandular hairs. Flower up to 25 mm long.
Calyx 5-partite, lobes lanceolate, 3 x 1 mm, green, pubescent. Corolla bilabiate,
tube narrowly subcylindric, sharply deflexed about the middle and narrowed, expanded
in throat, c. 8 mm long, base 4 mm diam., throat 4-5 mm diam., white to pale violet,
pilose with glandular and eglandular hairs; limb oblique, upper lobes divergent, sub-
orbicular, 4x3-5 mm, lower lip 10-13 mm long, lobes suborbicular, 6x4-5 mm,
336
all lobes pale violet, glandular inside, paler, sparsely pubescent outside, base of lower
lip and part of floor of tube white with deep violet streaks. Stamens: fertile 2, filaments
inserted two-thirds up corolla tube, 3 mm long, curved, curve dilated, white suffused
deep violet, sparsely glandular; anthers applanate, 1 mm long, medium violet, connec-
tive white, sparsely glandular; staminodes 3, posterior minute, laterals 1 mm long,
white. Disc annular. Ovary 5 mm long, green, pubescent; style terete, 5 mm long,
glandular-pubescent; stigma stomatomorphic, papillose, white. Capsule of 2 spirally
twisted valves, 16-36 (25 ± 4) mm long, 3 mm diam., persistent style-base prominent.
Fig. 1. Plate 1.
S. trabeculatus has a known distribution range from Dumisa in the Umzinto
district of Natal to near Lusikisiki in the Transkei. Plants grow wedged in rock crevices
along the usually precipitous faces of some of the numerous escarpments and river
gorges in this region, often in broken shade at the upper limits of forest, sometimes
sheltered by aspect alone. The flowering period is November to February.
Natal. — Alfred: Escarpment above Otterburn, c. 600 m, Hilliard 1153 (E, K, NH, NU); Burtt
2998(E). Port Shepstone : Plains, Izotsha Falls, c. 450 m, Hilliard 2750 (NU, holo. ; NE1, iso.); Prosser
1417 (J, K); Burtt 2996 (E); Oribi Gorge, “ Fairacres ”, c. 400 m, Hilliard 917 (E, NU). Umzinto:
Dumisa, farm “Ellesmere”, 650 m, Rudatis 1251 (E, K); Dumisa, Rudatis 1237 (BM).
Cape.- — Lusikisiki: Mkambati Leper Institute, on way to Umtentu Falls, Marais 977 (PRE).
The solitary leaf is long in relation to its breadth, and very thick owing to the
extreme prominence of the veins on the lower surface. The architecture of this surface
resembles a closely and intricately beamed ceiling, and suggested the specific epithet.
A shaggy indumentum of both glandular and eglandular hairs clothes the prominent veins,
making the leaf extremely viscid to the touch. The indumentum combined with the
shape and texture of the leaf are so distinctive that the species can be recognized by
these characters alone, which is useful, because it is a peculiarity of S. trabeculatus
that it seems to flower only sporadically even in nature, and seldom does so in cultivation.
This handsome plant has been largely overlooked since it was first collected by Rudatis
near Dumisa some fifty-five years ago. Its reluctance to flower, and the general inacces-
sibility of most of its haunts, no doubt account for the scanty records.
There appears to be considerable variation between populations both in the size
of the flowers and in corolla colour, markings and the shape of the lobes. For example,
corollas in populations at Izotsha Falls and Oribi Gorge are pale violet with a white
blotch at the base of the lower lip and darker violet lines in the throat extending out
onto the anterior lobe. The flowers are up to 25 mm long, the lower lip about 13 mm
long with suborbicular lobes. Flowers in a population at Otterburn, about five miles
as the crow flies from each of the other two localities, are up to 17 mm long, the
lower lip 9 mm long, the lobes obiong-elliptic, medium violet in colour with white
wedges on the lower lip and an inverted V of deep violet in the throat. The descrip-
tion on the sheet from the Umtentu Falls records “ flowers deep purplish-blue ”.
Notwithstanding such variation, the narrow corolla tube and conspicuous limb,
the densely massed, short, stout peduncles (which can, however, elongate if water is
plentiful, as in cultivation) and stout, apiculate capsules, together with the remarkable
leaf, combine to make this a most distinctive species. Its closest affinity is probably
with S. denticulatus from the south-eastern Transvaal. Both species favour the same
type of habitat and have similar capsules and corolla form, but they differ in the colour
and patterning of the limb. In S. denticulatus this may be “ rose-purple with lines of
deeper carmine on the lower lip, and dark-purple blotches on the lower part of the
upper lip ” (Curtis’ Bot. Mag. t. 8632, describing a plant from Barberton) or “ light
magenta, green on floor of throat and small purple-red-spotted there, larger spots
on side and roof of throat ” (B. L. Burtt, personal communication describing a plant
from the Belfast-Stoffberg road). In contrast, the limb of S. trabeculatus is in varying
337
tones of violet, white at the base of the lower lip with streaks of dark violet. The leaf
of S. denticulatus, too, is rather thin, cordate, almost glabrous, and sprinkled with
reddish-orange pigment granules, altogether quite unlike that of S. trabeculatus.
Of Natal species, only S. cooksonii is in the affinity of S. trabeculatus, from which
it is distinguished by its cordate, non-viscid leaf and broadly cylindric corolla tube.
Fig. 1. — 1, Streptocarpus trabeculatus, from Izotsha Falls, Natal, habit, x a, inflorescence about
natural size; b, half-flower, X If; c, stamens, X 5; d, gynaecium, x 2.
339
Plate 1. — Streptocarpus trabeculatus Hilliard, lower surface of leaf, about life size. The strongly
raised veins, which suggested the specific epithet, as well as the long glandular hairs, are clearly
seen.
10862
341
Bothalia 9, 2 : 341-350.
Notes and New Records of African Plants
by
Various Authors
GRAMINEAE
A new Variety of Festuca costata
Festuca costata Nees var. longiligulata J. G. Anders., var. nov.; planta miiiore
tenuiore; ligula prominenti, 7-12 mm longa, vaginae marginibus contigua, oblonga,
apice acuta vel acuminata, plerumque lacerata vel diffisa distinguitur. Plate 1.
Natal. — Bergville: Cathedral Peak Forest Research Station, occasional in moist gully in Indumeni
Valley, Killick 3509; 3510 (PRE, holo.).
This variety has been recorded only from the Indumeni Valley in the Cathedral
Peak area of the Natal Drakensberg. There it is found at an altitude of about 5,200
ft in moist gullies below the Cave Sandstone cliffs. According to Dr. D. J. B. Killick
(personal communication) local Bantu residents claim that this grass also grows on
the summit of the Drakensberg at approximately 10,000 ft.
In most respects it compares well with typical F. costata, but is generally a smaller
and less robust plant with culms up to 75 cm high, the leaves consistently narrow and
with the sheaths only slightly wider than the base of the leaf. It is characterized by
the prominent ligules which are 7-12 mm long, oblong, with the apex acute or acuminate
and usually lacerated or split. In typical F. costata the ligules are ovate, obtuse and
3-4 mm long.
The grass, which is known as uVinde by the local Bantu, is much used in the
making of domestic articles such as basins and beer sieves. The leaves have a very
high tensile strength when dry.
J. G. Anderson
A new Species of Panicum
Panicum dewinteri J. G. Anders., sp. nov. distincta nulla affinitate manifesta obvia.
Gramen perenne, caespitosum, ad 90 cm altum. Culmi erecti, tenues, duri et
ligni, a nodis multo ramosi, ramo quoque inflorescentiam ferenti, internodis inferioribus
et superioribus inaequalibus. Foliorum vaginae culmis veteris latae constrictae, firmae,
laminis caducis, culmis juvenalioribus laxae, marginibus versus ligulam praesertim
pilosae; ligula ciliata; collum hispidulum; laminae planae, rigidae, lanceolatae,
2- 8 cm longae, 1-6 mm latae. Inflorescentia 4-15 cm longa, 2-10 cm lata, in axillis
ramorum omnium pilosa. Spiculae 3-5-4 mm longae, lanceolatae, acutae. Glumae
inaequales; inferiores 2-5 mm longae, anguste ellipticae, 3-nervatae; superiores
3- 5-4 mm longae, lanceolatae, 7-nervatae. Flosculus inferior sterilis; lemma 3-5-4
mm longum, lanceolatum, 5-7 nervatum; palea 1-0-1 -5 mm longa, anguste ovata.
Flosculus superior bisexualis; lemma 2-5 mm longum, sub-oblongum; palea 2 -3-2 -5
mm longa, suboblonga. Antherae 1 mm longae. Caryopsis 1-5 mm longa, elliptica;
embryo 1/3 caryopsidis aequans. Plate 2.
Type: Transvaal, Soutpansberg District, Lejuma, near Louis Trichardt, De Winter
6006 (PRE, holo.).
342
Transvaal. — Petersburg: Blouberg, in kloof leading to trigonometrical beacon, Codd 8766; Codd
& Dyer 9080; at edge of Mohlakeng plateau, Codd & Dyer 9122. Soutpansberg : south end of Wyllies-
poort, Codd 8366; Lejuma, near Louis Trichardt, De Winter 6006.
A tufted perennial up to 90 cm tall. Culms erect, terete, many-noded, slender,
hard, wiry or woody, branching profusely from the rather thickened nodes, particularly
in the upper portion, each branch bearing an inflorescence, glabrous, striate; lower
internodes 4-8 cm long, those of the secondary culms 1-3 cm long. Leaves variable
in size; sheaths finely striate, longer than the internodes, those of the main culm and
older branches tightly embracing the culm, glabrous or sparingly pilose, soon becoming
dry and brittle, those of the branches looser, glabrous on the back, pilose on the
margins particularly towards the ligule; ligule reduced to a shortly ciliate rim; collar
hispidulous ; blades on the main culm and older branches falling off early, those of
the younger branches rigid, expanded, 2-8 cm long, 1-6 mm broad, lanceolate, tapering
to an acute apex, pubescent on the upper surface, lower surface glabrous, margins
scabrous. Panicle 4—15 cm long and 3-10 cm broad, narrowly ovate to ovate in outline,
bearing relatively few spikelets, pilose in the axils of all branches; rhachis more or
less terete, finely striate, glabrous; primary branches solitary, 3-7 cm long, sub-terete,
finely striate; secondary branches 2-3 cm long, slightly angular, straight or flexuous,
scabrous; pedicels 1-6 mm long, filiform, abruptly widened at the apex. Spikelets
3 • 5-4 mm long, lanceolate, acute, yellowish or pale green flushed with purple. Glumes
unequal, membranous; lower 2-5 mm long, narrowly elliptic, acute, 3-nerved, glab-
rous, pale; upper 3-5-4 mm long, lanceolate, acute, 7-nerved, glabrous, pale green
flushed with purple. Lower floret sterile; lemma 3-5-4 mm long, lanceolate, acute,
membranous, 5-7-nerved, glabrous, pale green flushed with purple; palea 1-1 -5mm
long, narrowly ovate, obtuse to acute, minutely scabrid on the keels, translucent.
Upper floret bisexual; lemma 2-5 mm long, sub-oblong, obtuse to sub-acute, light
brown, smooth and glossy, indistinctly 5-nerved; palea 2 -3-2 -5 mm long, sub-oblong,
obtuse to sub-acute, indistinctly 2-nerved. Anthers 1 mm long. Grain 1-5 mm long,
elliptic, brownish-purple; embryo 1/3 the length of the grain.
This grass has been recorded from the Pietersburg and Soutpansberg districts of
the northern Transvaal. Here it was found growing on rocky outcrops and in rock
crevices at altitudes ranging from 3,500-6,000 ft.
Panicum dewinteri is a distinct species with no close affinities amongst the South
African members of the genus.
It is an erect, tufted plant, branching repeatedly and profusely from the many-
noded culms, particularly in the upper parts, with each branch bearing an inflorescence.
The culms are slender and are hard and woody, or wiry, in appearance. The internodes
are variable in length : those of the primary culm 4-8 cm long and those of the branches
1-3 cm long. The main culm and stronger branches are very often naked because
the leaf-blades are cast off at maturity and only the hard, dry, dark-coloured sheaths,
or relics of the sheaths, firmly embracing the culm, remain. Like the internodes, the
leaf-blades and panicles vary considerably in size.
J. G. Anderson
LILIACEAE
Two new Species of Bulbine
Bulbine inflata Oberm., sp. nov., B. asphodeloidi (L.) Roem. & Schult. afflnis, sed
capsula tumida, pedicello firmo longiore differt.
Planta rhizomata glabra. Folia 10-15, erecta subteretia subulata c. 50 cm longa
1 cm diam. sulcata succulenta. Racemus multiflorus; scapus c. 35 cm longus fistulosus;
pedicelli teretes in fructu 2-5 cm longi; perianthii segmenta reflexa. Capsula inflata,
globosa, c. 13 mm diam. viridis nitida. Plate 3.
343
Type: Transvaal, Carolina District, 6 miles S.E. of The Brook, Codd 9503 (PRE,
holo.).
Perennial rhizomatous glabrous herbs. Roots long, spreading, numerous, from
a small compact rhizome. Leaves rosulate, 10-15, erect, terete, c. 50 cm long, 1 cm
wide, the vaginate bases forming a short tube below, attenuated to a subulate point,
dorsally convex, grooved, light green, succulent. Raceme 60-125 cm tall, bearing over
100 closely packed flowers which open consecutively, about 3 opening each day during
anthesis; scape cylinderical, hollow, c. 35-75 cm long, rhachis elongating during
anthesis, 1 5—35 cm long; bracts small, aristate, exserted in bud-stage; pedicels spreading-
reflexed, afterwards patent with erect apices, up to 2-5 cm. Flowers sweet smelling,
with a yellow perianth; segments reflexed, linear-lanceolate, 10 x 3 mm, outer with
an orange spot; stamens erect, filaments 7 mm long, with patent erect hairs up to
3 mm long, attached to the outside of the filament except at the apex where they cover
all sides; locules introrse, versatile, attached near the base dorsally; ovary turbinate,
3 mm, truncate at the apex, grooved: ovules about 8, biseriate, style filiform, 7 mm
long, stigma apical, minute. Capsule inflated, globose, about 13 mm in diam., green,
glossy; seeds angled through pressure, c. 2 mm in diam., black, sparsely tuberculate
and ridged.
Transvaal. — Carolina; 6 miles S.E. of The Brook at foot of rocky koppie, Codd 9503.
Swaziland. — Malkerns: near Manzini, farm Marula, Adelaar in PRE 30027. Mbabane: Hill
west of Mbabane, Compton 25721; Usutu Forests, hills, Compton 25640.
Natal.— Ixopo : Kununata, Strey 6127.
As far as is known one other species, B. namaensis Schinz from South West Africa,
also bears inflated fruits but it differs from the above species in being shorter and
caespitose, with the many shiny brown leaf-bases persisting as in B. stenophylla
Verdoorn.
Bulbine coetzeei Oberm., sp. nov., B. inflata Oberm. affinis, sed planta minore
solitaria, capsula minore non inflata differt.
Planta solitaria glabra basi bulbosa, radicibus multis tumidis. Folia filiformia
c. 50 cm longa et 1 mm diam. sulcata. Racemus foliis aequens vel plus minusve
superans pedicelli steriles recurvati; loculi 4-ovulati. Capsula minor dura.
Type: Transvaal, Barberton District, 8 miles from Nelspruit on road to Kaapse-
hoop, Coetzee in PRE 30026 (PRE, holo.).
Solitary, slender, glabrous herbs forming a small bulbous base, c. 2 cm long and
1 -5 cm broad from which a circle of swollen spreading roots are given off (the younger
roots situated above those of the previous year). Leaves erect, in a rosette, sheathing
below to form a very short neck about 2 cm long; about 7 leaves per flowering shoot,
lamina filiform, c. 50 cm long and 1 mm in diam. deeply grooved, acuminate, somewhat
succulent, applegreen. Raceme about as long as, or just overtopping the leaves; scape
firm, terete, 1-5 mm in diam. Flowers congested at first, rhachis lengthening during
anthesis, up to 15 cm long; bracts aristate, 1 cm long, brown keeled, membranous,
much exserted in bud stage; pedicels 15 mm long, recurving if infertile flower drops
off. Perianth yellow, segments 6 mm long, 1 mm broad; stamens typical, as long as
the perianth; ovary ovoid, 3-lobed, membranous, ovules 4 in each locule. Capsule
small, obovoid, 4 mm, woody; seed (somewhat immature) 2 mm in diam. angled
through pressure, ridged, obscurely and sparsely tuberculate.
Transvaal. — Barberton: 8 miles from Nelspruit on road to Kaapsehoop, on mountain slopes in
moist crevices rich in humus, Coetzee in PRE 30026.
Swaziland. — Piggs Peak, Emlembe, moist slope, Compton 24979, 30519.
Natal. — Vryheid: Zungwini Mountain, 84 miles N.E. of Vryheid, steep moist hillside below sand-
stone krantzes, solitary plants, Codd 6959. Eltrecht: Pongola Mountains, to Kaffir Drift, Thode
A.345. Nkandla: Nkandla Mountain, summit above forest, Wells 2504A.
344
The type plant was collected by Mr. J. Coetzee, who investigated the anatomy
in a research project for his M.Sc. thesis at the Pretoria University in 1966. The species
is undoubtedly related to B. inflata but it is smaller in all respects and does not bear
inflated fruits.
A. A. Obermeyer
A new Combination in Trachyandra
Trachyandra filiformis (Ait.) Oberm., comb. nov.
Anthericum filiforme Ait., Hort. Kew. 1 : 451 (1789); ed. 2, 2 : 267 (1811); Willd.,
Sp. Plant. 2 : 135 (1799); Roem. & Schult., Syst. Veg. 7 : 456 (1829). Type: Cape,
Masson', flowered at Kew Gardens in 1778 (BM, holo.). A. longepedunculatum Steud.
in Roem. & Schult., Syst. Veg. 7 : 457 (1829). Type: Cape, Ludwig (K, holo.). A.
aitonii Bak. in J. Linn. Soc. 15 : 294 (1876).
Phalangium filiforme (Ait.) Poir. in Lam., Encycl. 5 : 247 (1804).
Bulbinella filiformis (Ait.) Kunth, Enum. PI. 4 : 572 (1843); Bak. in FI. Cap. 6 : 356
(1896).
Trachyandra longepedunculata (Steud. ex Roem. & Schult.) Kunth, Enum. PI. 4 : 584
(1843). Oberm. in Bothalia 7 : 736 (1962).
The basionym, Anthericum filiforme, was unfortunately overlooked in the revision
of the South African species of Anthericum, Chlorophytum and Trachyandra (Bothalia 7,
1962). Baker described the flower as bright yellow, but in the original description
Aiton describes the perianth segments as white and green; he mentions also that the
leaf-bases are faintly maculate as well as minutely scabrid. Mr. John Lewis of the
British Museum (Natural History) kindly compared a specimen of T. longepedunculata
with Aiton’s type of Anthericum filiforme and found them to be cospecific.
A. A. Obermeyer
A New Combination in Ornithogalum
Ornithogalum ornithogaloides (Kunth) Oberm., comb. nov.
Bulbinellal ornithogaloides Kunth, Enum. 4 : 693 (1843). Type: Cape, Aliwal
North, banks of Orange River near Aliwal North, Drege 8695 (G!; PRE, photo.!).
Ornithogalum ? bulbinelloides Bak. in FI. Cap. 6 : 509 (1897), nom. illegit. Type:
as for Bulbinella ? ornithogaloides.
Anthericum ornithogaloides (Kunth) Bak. in J. Linn. Soc. (Bot.) 15 : 294 (1877).
A. pauper Poelln. in Bol. Soc. Broteria. 16, 2 : 47 (1942). Type: Transvaal, Heidelberg,
Wilms 1542 (B, holo.!; PRE, photo.!). A. longipedicellatim Poelln. in Repert. nov.
Spec. Regn. veg. 53 : 128 (1944). Type: Riversdale, Rust 14 (B, holo.!; PRE, photo.!).
Ornithogalum zeyheri Bak. in J. Linn. Soc. (Bot.) 13 : 281 (1873); FI. Cap. 6 : 502
(1897); Leighton in J. S. Afr. Bot. 11 : 163 (1945). Type: Cape, Uitenhage, near
Uitenhage in uliginosis ad Zwartkops River, Zeyher 1686 (K, holo.!; PRE, photo.!;
PRE!).
This hygrophyte is a common species in the summer-rainfall area, extending
from the south-eastern Cape to Rhodesia and Malawi. It is a slender glabrous plant
with the flowers borne on long pedicels. Of interest is the paucity of ovules, from
1-3 per locule, which is unusual for this genus.
A. A. Obermeyer
345
The Identity of Bulbinella carnosa
The South African species of Bulbinella are confined to the south-western Cape.
An exception appeared to be B. carnosa (Bak.) Bak. which was originally described
as Anthericum carnosum Bak. (Journ. Linn. Soc. (Bot.) 15 : 296, 1876) but which he
subsequently transferred to Bulbinella (FI. Cap. 6 : 358, 1896). It was based on a
specimen collected by Gerrard & McKen 1890 (K, holo., NH) from Natal, from damp
rocks between the rivers Umzinto and Ifafa.
I regard these binomials as synonyms of Ornithogalum paludosum Bak. (J. Bot.,
Lond. 12 : 366, 1874), a species which at times forms an atypical rootsystem. Possibly
due to a marshy habitat no firm bulb is formed, only a scaly swollen base.
Other synonyms include O. gracilentum Bak. (1897), O. ebulbe Schltr. (1907) and
Anthericum quadrifidum Poelln. (1942).
Bulbinella burkei Benth., which is also recorded from the summer rainfall area
in the Flora Capensis, has been transferred to Trachyandra burkei (Benth.) Oberm.
(Bothalia 7 : 721, 1962).
A. A. Obermeyer
The Identity of Fritillaria nana
In Baileya 13 : 7 (1965) Ingram made the combination Whiteheadia nana based
on Fritillaria nana Burm.f. (Prodr. FI. Cap. 9, 1768). This name has priority over
Whiteheadia bifolia (Jacq.) Bak. (J. Linn. Soc. (Bot.) 13 : 226, 1873) since the latter
is based on Eucomis bifolia Jacq. (Collect. 4 : 215, 1790). The type of the Burmann
species (Plate 4) was kindly sent on loan to me by the Director of the Delessert Her-
barium, Geneva. I was thus able to establish that it is a species of Eucomis and not
Whiteheadia. Therefore the solitary Whiteheadia species retains the binomial W.
bifolia (Jacq.) Bak., while Fritillaria nana was correctly placed under Eucomis by
L’Heritier as E. nana (Burm.f.) L’Herit.
A. A. Obermeyer
MELASTOMACEAE
While naming a specimen from Port Shepstone, Natal (Strey 5888), it was noted
that Memecylon australe Gilg & Schltr. in Bot. Jahrb. 40: 94 (1908), from Natal,
appears to be the same species as M. grandiflorum R. & A. Fernandes in Bol. Soc.
Brot. 29 : 63 (1955), based on Galpin 9568 from the Lusikisiki district. It so happens
that the former name is an illegitimate homonym, being antedated by M. australe
F. Muell. ex Triana (1871) and M. australe C. Moore (1893), thus M. grandiflorum
R. & A. Fernandes is the correct name for the species when the two entities are brought
together in one circumscription.
In this species it is found that the leaves vary from rotund (as described by R.
& A. Fernandes) to broadly elliptical, acute at apex and cuneate at the base (as described
by Gilg and Schlechter).
H. Badenhorst
RUBIACEAE
A New Species of Canthium
Canthium suberosum Codd, sp. nov., a C. ventoso (L.) S. Moore foliis ellipticis
vel oblanceolatis, inflorescentiis subsessilibus differt.
Frutex vel arbuscula 3-6 m alta omnino glaber, inermis; ramuli ad nodos incras-
sati; cortex demum suberosus. Folia opposita, 2-4 mm longe petiolata, glauca,
coriacea, elliptica vel oblanceolata, 3 • 5-7 cm longa, 1 • 4-3 • 5 cm lata, integra, basi
346
cuneata, apice obtusa vel rotundata. Cymae axillares, subfasciculatae, 1-4-florae,
pedunculo 0-2 mm longo, pedicellis 2-4 mm longis. Calyx breviter 5-denticulatus.
Corolla parvula, tubo 2 mm longo, lobis 5, oblongo-ovatis, 3 mm longis, reflexis,
margine papilloso. Antherae 5, breviter exsertae. Ovarium ovoideum, 3-loculare,
loculis uniovulatis; stylus exsertus, 4 mm longus; stigma capitatum, 3-lobatum.
Fructus drupaceus, circa 1 cm longus, plerumque lobatus, nigrescens, seminibus 1-3.
Type: Transvaal, Heidelberg Kloof, Mogg 20473 (PRE, holo.).
Shrub or small tree 3-6 m high, entirely glabrous, unarmed; branchlets conspi-
cuously swollen at the nodes; bark grey, powdery on young branchlets, becoming
corky on older trunks. Leaves opposite, with petioles 2^1 mm long, glaucous, leathery,
elliptic to oblanceolate, 3-5-7 cm long, 1-4-3 -5 cm broad, entire, cuneate at the base,
apex obtuse to rounded. Inflorescence axillary, subfasciculate, consisting of 1-4-
flowered cymes; peduncle 0-2 mm long, hidden by the stipules; pedicels 2-4 mm long.
Calyx shortly 5-toothed, teeth about 0-5 mm long. Corolla small; tube 2 mm long;
lobes 5, oblong-ovate, 3 mm long, margin papillate, especially at the tip which is
thickened and curved inwards, ultimately reflexed. Anthers 5, shortly exserted, oblong,
apex acute; filaments short and broad, inserted just below the throat of the corolla.
Ovary ovoid, 3-locular, with one descending ovule in each locule; style exserted, 4 mm
long, glabrous, broadest at the base; stigma capitate, 3-lobed; disc slightly raised,
sticky. Fruit drupaceous, about 1 cm long, more or less oblong, usually lobed when
2 or 3 seeds develop, black when mature, 1-3-seeded.
Found on wooded, rocky hillsides, usually in kloofs or along dry water-courses,
associated with quartzite or granite formations on the central Transvaal plateau region,
with outliers in western Swaziland.
Transvaal. — Brits: Jacksonstuin, Code! 6804; Van Wyk 267 ; Mogg 23538; Van Vuuren 275.
Heidelberg: Kloof, Mogg 20473. Krugersdurp: Hekpoort, Phillips 456. Pietersburg: Molepo Reserve,
30 miles E. of Pietersburg, Gerstner 5346; 5565; Wolkberg, 12 miles S.E. of Boyne, Codd 10399.
Rustenburg: Castle Gorge, Meeuse 9257. Waterberg: Geelhoutkop, Mogg 24495.
Swaziland. — Hlatikulu, Compton 28157.
C. suberosum is allied to C. ventosum (L.) S. Moore, of which C. inerme (L.f.)
Kuntze is regarded as a synonym, but differs in several respects. C. ventosum shows
a tendency to be scandent with longish internodes and spine-like branchlets, and the
bark on the older twigs is smooth and longitudinally wrinkled, somewhat waxy and
not powdery or corky, while the bark on the main trunk never becomes thickly corky;
the leaves are more or less ovate, in outline, obtuse at the base and acute to acuminate
at the apex and are relatively thin in texture; and the inflorescence is usually branched
and many-flowered.
C. suberosum , on the other hand, is never scandent, but has thickish, brittle branches
with shorter internodes and more conspicuously thickened nodes; the bark on the
older twigs is scurfy, cracking into a more or less rectangular network while on the
main trunk the bark often becomes conspicuously corky with age; the leaves are
elliptic to oblanceolate-obovate in shape, cuneate at the base and obtuse to rounded
at the apex, greyish and relatively thick in texture; and the inflorescences are few-
flowered and subsessile.
There is also a difference in distribution with C. ventosum scattered from the
Cape Peninsula, through the eastern Cape Province, Natal, Swaziland and the eastern
and northern Transvaal to Rhodesia and Mozambique, while C. suberosum is essentially
a species of the wooded quartzite and granite formations of the central Transvaal
plateau, with outliers in western Swaziland.
L. E. Codd
347
Plate 1.— Festuca costata Nees var. longiligulata J. G. Anders. (Killick 3510, holotype, PRE).
10862
348
Sit Hiil' I’reioru
PKE Ncg- \m .
i m ' i ' i 1 i n
lejur*. near LomTncUdt Erect broad
kdwd.fxrernal&W outcrops.
Aml /<T /i/'?SS V tOOl
Plate 2.— Panicum dewinteri J. G. Anders. (De 6006, holotype, PRE).
349
Plate 3. — Bulbine inflata Oberm. ( Adelaar in PRE 30027).
350
Plate 4. — Eucomis nana (Bunn, f.) L’Herit. (Without collector, holotype, in
Herb. Delessert, G).
351
Bothalia 9, 2 : 351-355.
The Genus Microcladia (Red Algae) in South Africa
by
R. H. Simons
The genus Microcladia is typified by M. glandulosa (Soland. ex Turn.) Grev.,
and has been regarded as distinct from the genus Ceramium Roth for over a century.
M. glandulosa was distinguished by its flattened and unarticulated thallus but, since
its establishment, other species have been associated with it which do not conform
with this circumscription. As a result, Feldmann-Mazoyer (1940, p. 345) was persuaded
that the genus should be revised. In fact, Mme Feldmann’s (l.c.) comments about
the genus suggested a lack of conviction about the separate status of Microcladia.
She seemed to think that, if this status were to be maintained, it should be based on the
bilateral nature of the thallus of M. glandulosa. This bilateral structure is the result
of the disposition of the unequal pericentral cells, the largest of which are arranged
in two opposite longitudinal ranks and thus increase the width of the thallus in that
plane. The presence or absence of articulation in the thallus was, for her, scarcely
relevant since she observed that articulations are visible in freshly collected plants
(Feldmann-Mazoyer, l.c., p. 343). Be that as it may, the diagnostic characters of
Microcladia seemed to be obscure and Kylin (1956, p. 380) offered no light on
the problem. The present author became concerned with the status of Microcladia
when, during his study of the South African species of Ceramium , the position of
Microcladia capensis sensu Papenf. (1940, p. 233), had to be considered.
The plant known in South Africa as Microcladia capensis (Kuetz.) Papenf., appeared
to be closely related to, if not identical to, Ceramium rubrum (Huds.) Ag., the type
species of the genus Ceramium. When compared with the accounts of Microcladia
glandulosa, the structure of M. capensis showed no reasonable likeness to them, and
the criteria for identifying Microcladia after Mazoyer (l.c.) did not apply. When
writing up the revision of the genus Ceramium in South Africa, Simons (1966), therefore,
included M. capensis sensu Papenf. under the name of its alleged basionym, Ceramium
capense Kuetz.
Some time after the paper (Simons, l.c.) was accepted for publication, Hommer-
sand’s (1963) account of Microcladia appeared. Here was published for the first time,
what seemed to be a valid reason for upholding the genus Microcladia: species, related
to Ceramium, having secondary cortical cells produced in an upward direction only,
and, at the same time, being completely corticated. This combination of charac-
teristics differentiates species of Microcladia from completely corticated species of
Ceramium, since the latter produce secondary cortical cells in both an upward and
downward direction. Hommersand (l.c.), recognized ten species of Microcladia and
suggested that M. capensis sensu Papenfuss might make a possible eleventh. But
the specimens of this species examined here have secondary cortical cells produced
both upwards and downwards and they must, therefore, belong to Ceramium.
Microcladia capensis sensu Papenf. was the only plant from South Africa which
had any pretensions to the genus Microcladia, and, after its removal to Ceramium,
it was thought that there were probably no South African representatives of Microcladia.
352
But since the revision of the genus Ceramium (Simons, l.c.), a plant bearing a remarkable
resemblance to Microcladia glandulosa (Soland. ex Turn.) Grev. was collected at Port
St. Johns on the eastern Cape coast. The plant was epiphytic on a specimen of Gelidium
cartilagineum (L.) Gaill., which was growing on a rocky substratum immediately below
the sublittoral fringe (sensu Stephenson, 1948, p. 212). It appears to be comparatively
rare, because further search in the area yielded only two more specimens. There
are no other records of this plant known to the author. Microscopic examination
of these plants confirmed their structural similarity to Microcladia coulteri Harv.,
(fide Hommersand, l.c., pp. 214-223), and thereby, their identity. The tetrasporic
plant in the collection, revealed by its exserted sporangia, that it was unique amongst
the known species of Microcladia.
In regard to its tetrasporangia, it is interesting to note that they were observed
to be attached to both pericentral and other cortical cells (Fig. 1, a-f). This contrasts
with the situation of tetrasporangia in five out of the ten species of Ceramium discussed
by Simons (l.c.), where the attachment was to pericentrals only; although the actual
attachment was not discerned in the other five species, the position of the tetrasporangia
as revealed in sections was such as to indicate a high probability of a pericentral
attachment. Simons (l.c.) suggested that this sporangial attachment in Ceramium
might be a distinguishing character and this idea seems to be enhanced by Hommer-
sand’s (l.c.) observations: of the four species of Ceramium he described in detail, all
had sporangia attached only to pericentrals, whereas Microcladia coulteri Harv., (Hom-
mersand, l.c., p. 223), had tetrasporangia attached to both pericentrals and other cortical
cells, which compares with the situation of the tetrasporangia in the plant described
below.
This newly discovered South African plant is named Microcladia gelidii, because
it was found as an epiphyte of Gelidium cartilagineum (L.) Gaill. It is at present the
only species of Microcladia known in South Africa.
Microcladia gelidii Simons, sp. nov., M. glandulosae (Soland. ex Turn) Grev.
affinis, sed ab ea thallo sursum et deorsum parum attenuato, cellulis axialibus brevi-
oribus et tetrasporangiis exsertis differt.
Planta epiphytica atrosanguinea subfastigiata subsetacea. Hapteron intra hospitem
ramificans et nodulum, c. 2 mm diam., in superficie hospitis formans. Axes plures
ex nodulo excurrentes cauliformes complanati bilaterales ramosi ad 8 cm alti ad 1 mm
lati sursum deorsum parum attenuati. Rami distichi laterales pinnatim in uno piano
decompositi, in planta tetrasporifera inferiores subrecurvati superiores forcipati, in
planta feminea omnes forcipati. Rami adventitii curti frequentes. Tetrasporangia
exserta verticillata in ramis subultimis vel in ramis adventitiis ad cellulas corticales
et, saepius, ad cellulas pericentrales affixa. Cystocarpia subterminalia involucrata.
Planta mascula ignota. Thallus intra ex cellulis axialibus uniseriatis omnino corticatis
quam altis parum latioribus constans. Cortex ex pericentralibus et secondariis cellulis
corticalibus constans. Pericentrales ad usque 9, sed 6 vel 7 saepius, circum axiales
cellulas parentales, inaequimagnae, maximae oppositae biseriatae margines thalli
complanati congruentes, omnes quam cellulis axialibus angustiores sed altitudinem
earum aequantes et cum eis alternantes. Secondariae cellulae corticales quam aliis
parviores stratum ad duas cellulas latas extrorsus formantes ex pericentralibus solum
sursum exorientes. Plates 1 and 2. Fig. 1.
Type: Cape, Port St. Johns, epiphytic on Gelidium cartilagineum, Simons 698
(PRE, holo.).
353
Fig. 1. — Microcladia gelidii; a-f, various transverse sections through tetrasporangial thalli indicating
the attachment of the tetrasporangia; a and c, tetrasporangia connected directly to a pericentral
cell; d, a constriction in the lower portion of a maturing tetrasporangium; b, e and f, a cortical
cell intermediate between tetrasporangia and pericentral cells; g, an oblique longitudinal section
through a mature portion of the thallus.
a. — axial cell; p. — pericentral cell; s.c.- — secondary cortical cell; t. — tetrasporangium.
354
Plant epithytic, dark blood-red, subfastigiate, subsetaceous. Holdfast ramifying
within host and forming an external nodule, about 2 mm in diam., on the surface of
the host. Axes several, arising from nodule, stem-like flattened bilateral branched up
to 8 cm high and 1 mm thick tapering slightly upwards and downwards. Branches
distichous lateral pinnately decompound in one plane, in tetrasporic plant the lower
ones somewhat recurved the upper forcipate, in female plant all forcipate. Adven-
titious branches short lateral frequent. Tetrasporangia exserted whorled on subultimate
branches or on adventitious branches; attached to cortical cells and, more often, to
pericentral cells. Cystocarps subterminal involucrate. Male plants unknown. Thallus
internally composed of entirely corticated uniseriate axial cells a little wider than high.
Cortex consisting of pericentrals and secondary cortical cells. Pericentrals as many
as 9, but more often 6 or 7, forming a ring around the parent axial cells, unequal in
size, the largest in two opposite series corresponding to the edges of the flattened
thallus, all narrower than but equalling the height of the axial cells and alternating
with them. Secondary cortical cells smaller than the rest forming up to two layers
outwards arising from the pericentrals in an upward direction only.
Cape.— Port St. Johns: Port St. Johns, Simons 698 (PRE); 699 (BOL).
References
Feldmann-Mazoyer, G. (1940). Ceramiacees de la Mediterranee Occidentale. Alger:
Imprimerie Minerva.
Hommersand, Max H. (1963). Morphology and classification of Ceramiaceae and
Rhodomelaceae. Univ. Calif. Publ. Bot. 35, 2 : 165-366.
Kylin, H. (1956). Die Gattungen der Rhodophyceen. Lund: CWK Gleerups Forlag.
Papenfuss, G. F. (1940). Notes on South African marine algae. I. Bot. Notiser.
Simons, R. H. (1966). The genus Ceramium in South Africa. Bothalia 9, 1 : 153-168.
Stephenson, T. S. (1948). The constitution of the intertidal fauna and flora of South
Africa. III. Ann. Natal Mus. 11,2: 207-324.
355
10862
Plate 1.— Microcladia gelidii Simons, tetrasporic plant (Simons 698, holotype, in PRE).
Plate 2. — Microcladia gelidii, fertile female branches adventitious on a mature axis.
357
Bothalia 9, 2 : 357-361.
A New Species of Brown Alga from South Africa
by
R. H. Simons
Carpomitra longicarpa Simons sp. nov., a speciebus Carpomitrae aliis fronde
ecostata tereteque, soris elongatis et filamento uniseriato fili sporangiferi paucis cellulis
dilatatis terminato differt.
Frons obscure olivaceo-brunnea cornea ecostata, base callum stuposum deinde
unum vel plures primarios axes cauliformes formans. Axes filiformes ad 1 mm lati
et ad 50 cm alti teretes articulati ramosi penicillis coloratis terminati. Rami alterni
vel oppositi quam axibus graciliores sed ceterum conformantes. Sori pedicellati in
ramis vel axibus laterales cylindro-clavati ad 6 mm alti et ad 0-5 mm lati, quisque
ab annulo terminali constrictione patenti separati, ex duobus stratis, interiore cellulis
elongatis sterilibus, exteriore filamentis fertilibus radiantibus, constantes. Filamenta
fertilia ex filis sporangiferis constantia circa 200 p longa. Filum sporangiferum ex
sporangiis in filamento uniseriato lateraliter portatis constans. Filamentum uniseriatum
ex circa decern cellulis, superioribus quattuor vel quinque coloratis dilatatis suglobosis
c. 16 p diam., cellula basali subconica, cellulis intermediis cylindraceis c. 8 p latis
constans. Sporangia unilocularia cylindro-clavata ad 60 p longa et 15 p lata in cellulis
inferis filamentorum uniseriatorum portata. Plate 1. Fig. 1.
Type: Cape: Bizana district, Mzamba, sublittoral on rocks, Simons 865 (PRE,
holo.).
Frond dark olive-brown horny unribbed, forming at the base a stupose callus
from which one or more primary stem-like axes arise. Axes filiform up to 50 cm high
and 1 mm broad terete articulated branched terminating in a pigmented penicillus.
Branches alternate or opposite more slender than the axes but otherwise conforming.
Sori lateral on branches (or axes) cylindro-clavate up to 6 mm high and 0-5 mm broad,
each separated from a terminal annulus by a patent constriction; consisting of two
tissues, an inner of elongated sterile cells, an outer of radiating fertile filaments. Fertile
filaments consisting of sporangiferous threads about 200 p long. Sporangiferous threads
each consisting of sporangia carried laterally on a uniseriate filament. Uniseriate
filament consists of about 10 cells, the upper 4 or 5 pigmented dilated subglobose about
16 p diam., the basal cell subconical, the intermediate cells cylindrical about 8 p broad.
Sporangia unilocular cylindro-clavate up to 60 p long and 15 p broad pedicellate on
the lower cells of the uniseriate filaments.
Cape. — British Kaffraria, Flanagan 45 (BOL). Bomvanaland Coast, Filmer s.n. (BOL). Bizana:
Mzamba, Simons 865 (PRE, BOL); Simons 866 (BOL).
All the cited specimens except Simons 865-866 bear the inscription Carpomitra
chytraphora Kuetz., and one bears the further name “ Carpomitra filiformis (Suhr)
Papenf. (1950)”. Chytraphora filiformus Suhr, the basionym of Carpomitra filiformus
(Suhr) Papenf. and Carpomitra chytraphora Kuetz. are both described as flattened and
with a mid-rib whereas the present plant has terete axes and there is no sign of a mid-rib
anywhere. Then, too, the sori of C. chytraphora and C. filiformis are cupule-shaped
which contrasts with the cylindro-clavate shape of the sori of C. longicarpa. There
is a further difference between the present plant and the other(s) and indeed it is one
358
which, as far as can be ascertained from the literature, occurs in no other member of
the Sporochnales to which order these plants belong: terminating its sporangiferous
threads is a series of four or five dilated pigmented cells; in other species there is
apparently only one dilated cell at the apex of such a thread.
In general, this plant has the characteristics of members of the order Sporochnales
of the Phaeophyceae and, more particularly, of the family Sporochnaceae sensu Decaisne
(1842). Because of the all but terminal position of its sori, this species belongs either
in the genus Sporochnus or in the genus Carpomitra. Of these two genera, only
Carpomitra has a sterile annulus differentiated above its sori (Fritsch, 1945, p. 173).
Thus the present plant would appear to be a member of this genus. But there is a slight
difficulty, because typically the species of Carpomitra have sori crowned with a penicillus
of fine hairs. The present plant showed no evidence of such a penicillus. However,
in the author’s opinion this seems an insufficient reason for excluding this plant from
the genus Carpomitra.
Fig. 1. — Carpomitra longicarpa; a, a single sorus supported on its pedicel; b, a single sporangiferous
thread from a sorus.
p. — pedicel; p.c. — pigmented cell; sp. — sporangium; s.s. — evacuated sporangium.
359
References
De Toni, J. B. (1895). Syl/oge Algarum, Vol. 3. Patavii.
Fritsch, F. E. (1945). Structure and Reproduction of the Algae , Vol. 2. Cambridge
University Press.
Kuetzing, F. Tr. (1859). Tabulae Phycologicae. 9: 37, t. 89,\fig. 2. Nordhausen.
Papenfuss, G. F. (1943). Notes on South African Marine Algae. II. J. S. Afr.
Bot. 9: 79-92.
Suhr, J. N. (1834). Algae von Ecklon. Flora, Jena 17 (46) : 721, t. 1.
361
iji
j
i
Plate 1 Carpomitra longicarpa Simons ( Simons 865, holotye, in PRE).
10862
v»3C»
CONTENTS
Vol. 9, Parts 3 and 4
The South African Species of Kniphofia (Liliaceae). L. E. Codd
Page
363
363
The South African Species of Kniphofia (Liliaceae)
by
L. E. Codd
CONTENTS
Page Plate
Introduction 367 —
The Generic Name 367 —
The K. uvaria and K. praecox Complexes 369
Pre-Linnaean History 373 —
Post-Linnaean History 376 —
Taxonomic Value of Characters 388 —
Table of Flowering Times 391 —
Natural Hybrids 392 —
Acknowledgements 393 —
Taxonomic Treatment 394 —
Generic Citation 394
Key to Species 397
1 . K. multiflora Wood & Evans 400 1
2. K. parviflora Kunth 403
3. K. acraea Codd 405 2b
4. K. umbrina Codd 406
5. K. typhoides Codd 407 3
6. K. brachystachya (A. Zahlbr.) Codd 410 —
7. K. buchananii Bak 411 4a
8. K. breviflora Harv. ex Bak 413 —
9. K. albescens Codd 417 5
10. K. flammula Codd 418 6
11. K. evansii Bak 419
12. K. fibrosa Bak 422
13. K. crassifolia Bak 424
14. K. gracilis Harv. ex Bak 424 7
15. K. laxiflora Kunth 428
16. K. ichopensis Bak. ex Schinz 433 8
17. K. rufa Bak 434 2a
364
Page Plate
18. K. pauciftora Bak 438
19. K. tabular is Marl 438
20. K. ensifolia Bak 440 —
a. subsp. ensifolia 440 9
b. subsp. autumnalis Codd 442
21. K. splendida E. A. Bruce 442 10
22. K. praecox Bak 444
a. subsp. praecox 447 —
b. subsp. bruceae Codd 449 —
23. K. sarmentosa (Andr.) Kunth 453 1 1
24. K. coralligemma E. A. Bruce 453 12
25. K. caulescens Bak 455 —
26. K. ritualis Codd 458 —
27. K. hirsuta Codd 460
28. K. stricta Codd 461 13
29. K. northiae Bak 464 —
30. K. galpinii Bak 466 14
31. K. triangularis Kunth 468
a. subsp. triangularis 469 15
b. subsp. obtusiloba (Diels ex Berger) Codd 472 16
32. K. thodei Bak 472 4b
33. K. porphyrantha Bak 474 17
34. K. fluviatilis Codd 476 18
35. K. littoralis Codd 478 19
36. K. elegans Codd 480
37. K. urepanophylla Bak 481 20
38. K. baurii Bak 482 21
39. K. latifolia Codd 484
40. K. rigidifolia E. A. Bruce 487 22
41. K. rooperi (Moore) Lem 488 23
42. K. maria (L.) Hook 492 24
43. K. citrina Bak 497 25
44. K. linearifolia Bak 498 26
45. K. tysonii Bak 504
a. subsp. tysonii 506 27
b. subsp. lebomboensis Codd 506
Excluded Species 508
Opsomming 509
Index to Collectors’ Numbers 509
Index 512
Fig. 1. Johannes Hieronymus Kniphof, 1704-63, Professor of Medicine at Erfurt University, after
whom the genus Kniphofia Moench is named.
367
THE SOUTH AFRICAN SPECIES OF KNIPHOFIA
(LILIACEAE)
Abstract
A revision of the South African species of Kniphofia (Liliaceae) is presented in which a key is
provided to the 45 species and 4 subspecies recognized. The new species, subspecies and combinations
are: K. ensifolia Bak. subsp. autumnalis, K. praecox Bak. subsp. bruceae, K. triangularis Kunth subsp.
obtusiloba (Diels ex Berger) ( K . obtusiloba Diels ex Berger), K. littoralis, K. elegans, K. latifolia, K.
tysonii Bak. subsp. lebomboensis. Freely illustrated, with 27 species depicted life-size in colour.
INTRODUCTION
The genus Kniphofia Moench, which includes the plants commonly known as
Redhot Pokers, was named in honour of Johannes Hieronymus Kniphof, 1704-63
(Fig. 1), who was Professor of Medicine at Erfurt University, Germany. He was
interested in botany and one of his best known works was entitled Botanica in Originali
or Herbarium Vivum. It comprised 1,200 botanical illustrations which were produced
by a somewhat unique process, whereby dried plant specimens were coated with
printer’s ink and pressed on paper, resulting in a silhouette effect. According to Kew
Bulletin 1895: 157 (1895): “ It is interesting historically both on account of its being
a record of plants cultivated at that date in Germany and also on account of its being
one of the earliest, if not actually the first, work of considerable extent, in which the
process of nature-printing was employed to illustrate plants ”. This was an adaptation
from an earlier method in which the plant specimens were blackened by holding them
over the smoke of a candle or oil lamp; by placing these between sheets of paper and
rubbing them down with a smoothing bone, the lamp-black was transferred to the paper
(see Blunt, “The art of botanical illustration”, Collins, London, 1950, pp. 138-9;
and Nissen, “Die botanische Buchillustration ”, Hiersemann, Stuttgart, 1951, pp.
246-7).
The 1,200 illustrations were published in twelve folio volumes of 100 plates each,
and each centurion is preceded by a title page. One of the illustrations depicts the
species then known as Aloe uvaria L. which later became the type species of the genus
Kniphofia Moench. The dates on which the various illustrations were published cannot
be determined with certainty. In the copy at the British Museum (Natural History),
the plate of Aloe uvaria is No. 805 in centurion IX, of which the title page, dated 1762,
is reproduced in Fig. 2. This date may be accepted as authentic, although, in Index
Londinensis, which is based on the Kew copy, the plate of Aloe uvaria is cited as Herb.
Viv. 1 : t.46 (1758). As will be seen from Fig. 2, reproduced from the British Museum
copy, the characters of the plants are depicted with remarkable fidelity.
368
D. 10. HIERON. KNIPHOFII
CAOEM. EUFV'RTKNS. K t RKCTOR. MAC N IF.
? A T H O t. n HAS. PROP. PVBL. OBDIS.
PACYlT. MEO, SENIOR, it AfiSESJ. PElMAB-
ACAl>, CAESAR. NAT. CVRtOSOR,
ACVNCTf ET aiflMOTBf'CAlW,
BOTANICA IN ORIGINAL!
HERBAR1VM VIVVM
in qvo
PLANTAR YM TAM JND1GENARVM
QY.\M EXOT ICARVM.
PECVLIARI <*VAD.AM OPKKOSAC^VE ENCHTRESJ
AIRaMENTO JMPRESSORtO OPOVCTaRVM
AD MCTROOVW
VLLVSTRIVM NOSTJU AEVI BOTANICORVM
LINNAEI et LVDWHill
JOANNIS GODOFR EDI TRAMPE
CENTVK. IX.
Fig. 2.— Title page of Centurion IX (1762) of Kniphof’s publication Herbarium Vivum.
369
It may be noted that Kniphofia Moench (1794) was antedated by Kniphofia Scop.
(1777). The latter is generally regarded as being a synonym of Terminalia L. (Com-
bretaceae) and has never been taken up. In order to preserve the name for the Redhot
Pokers, the recommendation to conserve Kniphofia Moench, in accordance with the
provisions of the International Code of Botanical Nomenclature, was approved (Taxon
8: 140, 1959).
Linnaeus described the only species of the genus then known to him as Aloe uvaria
in his Species Plantarum (1753). In 1771 he transferred it to the genus Aletris while
Willdenow, in 1799, placed it in Veltheimia, but it clearly does not belong to either
of these genera. In 1794 Moench created the genus Kniphofia to accommodate the
species, but took the unnecessary step of changing the epithet to K. alooides. The
generic name Kniphofia was not taken up until 1843, when Kunth reviewed the species
then known to him. In the meantime Ker Gawler described the genus Tritoma in
1804 and made the combination Tritoma uvaria (L.) Ker Gawl. The generic name
Tritoma became widely used in Britain and America for the Redhot Pokers and still
persists in nursery catalogues to this day, even though the name Kniphofia has been
in general use in scientific literature for the last 100 years. Subsequently Link, in
1821, proposed the generic name Tritomanthe Link in place of Tritoma , pointing out
that the latter name had previously been used for a genus of insects and, in 1829, used
the name Tritomium Link, without giving reasons for the change. In 1866 a further
change to Triclissa Salisb. was published, attention again being drawn to the prior
use of the name Tritoma for a genus of insects. Further notes on the history of the genus
will be found on pp. 376-88).
The genus is almost entirely confined to the continent of Africa, with only one
species in Madagascar and one in southern Arabia. In Yol. 6 of Flora Capensis (1896),
Baker recognized 32 species in South Africa and, in Vol. 7 of Flora of Tropical Africa
1898), he dealt with 13 tropical species. Since then the genus was monographed as
a whole by Berger (1908), who deals with 67 species, 44 of which are recorded from
South Africa. The genus Notosceptrum, separated by Bentham in 1883, was upheld
by both Baker and Berger (3 species in South Africa and 2 species in Angola). In
the present revision, Notosceptrum is included in synonymy under Kniphofia (see pp.
392, 394, in which genus 45 South African species are maintained.
THE K. UVARIA AND K. PRAECOX COMPLEXES
Before proceeding to a historical outline of the genus, it is appropriate to comment
on members of the K. uvaria and K. praecox complexes which have contributed largely
to the commonly cultivated Redhot Pokers. Much of the early history of the genus
is concerned with the application and misinterpretation of species names within these
groups, thus illustrating some of the difficulties encountered in studying a genus
favoured for horticultural purposes.
Fortunately, until about 1800, apparently only one member from these two groups
of species was known to botanists in Europe, namely K. uvaria (L.) Hook., the typifi-
cation of which is not in doubt. This is the species which occurs in the south-western
Cape Province and was readily available to early visitors to the Cape. However, as
botanical exploration spread further to the east, certain more robust Kniphofia plants,
with larger inflorescences, found their way to European gardens. These were not
immediately recognized as distinct and were grown and illustrated under the epithet
“ uvaria ” or its synonym, “ alooides
370
Early examples which may be quoted are the plants illustrated on Plate 4 of Jacquin's
Fragmenta (1800 or 1801) under the name Veltheimia maria (reproduced here as
Fig. 9), and Plate 291 of Redoute’s Liliacees (1810 or 1811) under the name Tritoma
maria (see Fig. 11). Not only can these two be distinguished from true K. maria,
but they are clearly distinct from one another. Yet 70 years or more elapsed before
the species represented in these illustrations received names of their own.
K. praecox Bak. The identity of the Jacquin illustration (see Fig. 9), which has
an elongate inflorescence, perianth tube constricted above the ovary and well-exserted
stamens, provided a puzzle for some time. Only recently has it been satisfactorily
matched with naturally occurring plants which are restricted to a few localities in the
Knysna and Komga areas. Another characteristic then became apparent, namely
that it possesses narrow, long-acuminate bracts which are very different from the ovate-
oblong, obtuse to rounded bracts of the K. maria complex.
The bracts are of particular importance in the diagnosis of this group in the
herbarium with the result that it is now possible for the first time to account satis-
factorily for certain old herbarium specimens in various European herbaria (discussed
in some detail on p. 448). These specimens, which were confused under various names,
have narrow, long-acuminate bracts, and are thus obviously allied to the plant illustrated
by Jacquin. They were widely grown in Europe in the early 1800’s and superficially
resemble certain members of the maria complex, but can readily be identified by their
characteristic bracts. Derivatives of these plants are grown in gardens to this day
and, over the years, some of the floral characters have become modified in cultivation,
as may be expected.
In 1870 Baker somewhat diffidently described K. praecox in Saunders’ Ref. Bot.
1. 1 69, stating that it differed from “ uvaria ” mainly in its earlier flowering time. In
subsequent revisions Baker himself and, later, Berger (1908) included the species in
their concept of K. uvaria. The illustration (see Fig. 46, p. 446) cannot readily be
distinguished from the maria complex, but fortunately a specimen is preserved in Kew
Herbarium and it may be seen at a glance that the bracts are narrow and long-acuminate.
The type of K. praecox is, therefore, clearly allied to the plant illustrated by Jacquin.
It is not a perfect match of the known wild plants now associated with the Jacquin
plate (flowers of the type plant are longer and not distinctly constricted above the
ovary, with the stamens only slightly exserted), but the wild plants may be included
in a broad concept of K. praecox. It may be mentioned that another cultivated member
of this complex with pale yellow flowers was later illustrated in the Bot. Mag. t.7623
(1898) where it was erroneously called K. longicollis Bak.
Unless, therefore, adequate botanical details are shown (or described), the identity
of an illustration (especially in the K. maria or K. praecox complexes) cannot be
satisfactorily established unless a corresponding herbarium specimen is preserved.
For this reason one may conveniently ignore many semi-horticultural names, such
as T. saundersii Carr. (1882) and T. nobilis Guill. (1882), of which no authentic type
specimens have been traced.
The K. uvaria Complex. This complex is widespread and variable, extending
from the western Cape Province, through eastern Cape Province, eastern Free State,
Natal, Swaziland and eastern Transvaal to eastern Rhodesia. The copious modern
material now available appears to fall into five main groups according mainly to the
shape of the inflorescence, which varies from globose (large, medium or small) to
cylindrical, and the perianth length, which varies from 20 mm to 45 mm. The groups
vary in size and degree of variability, and are by no means clear-cut, being linked by
intermediates. However, each group has a fairly well-defined geographical distribution,
as indicated in Maps 13 and 14.
371
Opinions may differ as to the rank which should be accorded to each of these
groups. From what has been said above it may seem logical to designate each group
as a subspecies of K. uvaria. Apart from the cumbersome names that would result
from such a treatment, it is preferred, at the present state of our knowledge, to give
each group the rank of species because more field work is required than has been possible
during the present study to assess the status of the so-called intermediate specimens.
Some are undoubtedly true intermediates in genetic composition, while others, as
already suspected, will prove to be abnormal variants, either depauperate or unusually
large specimens of recognized groups, resulting from local or seasonal abnormalities,
cultivation, etc.
Partly for these reasons, it is impossible to construct a key which will necessarily
lead to the correct identification of every specimen, regardless of its condition. The
key provided on pp. 397-400 is based on average specimens and is, therefore, intended
mainly as a guide. The position is further complicated by the fact that Kniphofias,
like so many other Monocotyledons, are difficult to transform into good herbarium
specimens. When identifying a member of the K. uvaria complex, it is suggested that
the key should be used in conjunction with distribution maps Nos. 13 and 14 (p. 497)
and in consultation with the group characters summarised briefly below.
In determining the correct epithet at species level for each group, it is necessary
to review 11 epithets which, in chronological order of priority, are as follows: K.
uvaria (L.) Hook. (1753), K. burchellii (Herb.) Kunth (1836), K. rooperi (Moore) Lem.
(1852), K. tysonii (Bak. (1889), K. longicollis Bak. (1893), K. citrina Bak. (1893), K.
linearifolia Bak. (1896), K. longiflora Bak. (1901), K. bachmannii Bak. (1901), K.
occidenlalis Berger (1908) and K. rhodesiana Rendle (1911). These are allocated to
the five groups as shown below. Full descriptions and discussion of synonymy will
be found in the taxonomic section, as indicated by the page numbers quoted.
1. K. uvaria (L.) Hook. Plants of small to medium stature, with medium-sized
oblong to globose inflorescences, perianth tube 28-38 mm long, usually conspicuous
pedicels (2-)3-5 mm long and slightly exserted stamens. Distributed from south-
western Cape Province northwards to Khamiesberg in Namaqualand, eastwards to
King William's Town District, and from there inland to Murraysburg, Steynsburg
and Xalanga Districts. In Humansdorp and Albany Districts it grades into Group 4,
while in the north-eastern Cape Province it may be difficult to distinguish satisfactorily
from Group 5. Synonyms: K. burchellii (Herb.) Kunth, K. bachmannii Bak. and K.
occidentalis Berger (see p. 492).
2. K. rooperi (Moore) Lem. Relatively robust plants (but varying according to
soil conditions) with broad, somewhat arcuate leaves, large globose inflorescences,
perianth tube 32-44 mm long, short pedicels and slightly exserted stamens. Described
in 1852 from a cultivated plant originally from “ Kaffraria ”, and is distributed near
the coast from King William’s Town to Port Shepstone Districts. Synonym: K.
longicollis Bak. (see p. 488).
3. K. tysonii Bak. Robust plants with erect leaves, oblong to cylindrical inflore-
scences, perianth tube 22-30 mm, short pedicels and well-exserted stamens. Based
on a wild plant collected by Tyson in East Griqualand and distributed from there to
the Natal Midlands, with a subspecies extending to the Lebombo Mts. in Swaziland
(see p. 504).
372
373
4. K. citrina Bak. Plants of small stature with narrow, mainly arcuate-spreading
leaves, small globose inflorescences, perianth tube 20-28 mm long, short pedicels and
well-exserted stamens. The type is a cultivated plant said to have its origin in the
Albany District and it is in this and neighbouring districts that matching plants are
found (see p. 497).
5. K. linearifolia Bak. Robust to medium-robust plants with medium to large
rhomboid to elongate inflorescences, perianth tube 28-40 mm long (sometimes shorter
in cultivation), short pedicels and slightly exserted stamens. This is the most wide-
spread group of the K. uvaria complex and, probably, in the whole genus. Its main
area of distribution is from about Somerset East in the Cape, through eastern Cape,
eastern Free State, Natal, Swaziland and eastern Transvaal to eastern Rhodesia. It
may come as a surprise, therefore, that it was the last group in the complex to receive
a species name. The reason is, of course, that the plants of this group were grown in
Europe under the name “ uvaria ” or “ alooides ”. Examples which are well known
are the plants illustrated in Redoute's Plate 291 (see Fig. 11, p. 383) and Bot. Mag.
t.48 16, 1854 (see Fig. 12, p. 385). The type of K. linearifolia was a Bachmann specimen
in Berlin which was destroyed during World War II. It was collected in Pondoland,
not far from Lusikisiki, and a study of Kniphofias in this area indicated that the present
concept answers to the description. The name K. linearifolia is, therefore, interpreted
in the sense of the specimens cited on p. 499, and a neotype has been designated. There
are two synonyms at species level, K. longiflora Bak. and K. rhodesiana Rendle.
PRE-LINNAEAN HISTORY
The first known reference to the plant now called Kniphofia uvaria appeared in
1644 under the name Iris uvaria promont. bonae spei (“ uvaria ” indicates a resemblance
to a bunch of grapes) in the edition of Theophrasti Eresii de Historia P/antarum by
Johannes Bodaeus a Stapel (or Van Stapel, latinized as Stapelius), published in
Amsterdam. In this work five Cape plants are illustrated and shortly described in
Latin including, on p. 335, the drawing reproduced here (Fig. 4). It is probable that
the drawings and descriptions were done by Justus Heurnius, a missionary from Leiden,
Holland, when he called at the Cape in the year 1624. The description of the plant
may be roughly translated as follows : “ Flowers purple-red, inflorescence wedge-shaped,
stem about 40 inches tall, round, brown, leaves long, f inch broad, both sides acute
glandular (i.e. serrulate?), flowers foetid, grows in wet places ”. Actually the flowers
of this plant have no odour at all and it may be concluded that the reference to foetid
flowers became transferred from the description of Stapelia variegata, which is illustrated
on the same page.
Another illustration of K. uvaria of considerable historical interest is in a collection
of early paintings of Cape plants acquired by the Africana Museum, Johannesburg.
The paintings are attributed to Hendrik Claudius, a German apothecary, who was sent
to the Cape from Batavia in 1682 to collect and draw botanical specimens. They are
reproduced (much reduced and not in colour) in a “ Catalogue of Pictures in the
Africana Museum” compiled by the former Director of the Museum, R. F. Kennedy,
and K. uvaria appears as illustration C 314 in Vol. 2 (1967) of the Catalogue.
It was customary, up to the time of Linnaeus, to designate plants by means of a
descriptive phrase. The various phrase names used for the species are summarised
below, followed in each case by authors who used them.
Aloe africana folio triangulo longissimo etc.: Paul Herman, Hort. Acad. Lugd.-Bat,
p. 18 (1687); Commelin, Hort. Amst. 2, p. 29, 1. 1 5 (1701); Boerhaave, Index, 2, p. 131.
No. 43 (1727).
374
Fig.
H1STOR. PLANT. LIB. IV.
P lanta fine nomine. Hanc placet verbenam lanugino-
fam Indicam vocarc, aue his verbis abHeurniodefcri-
bitur. Folia habet obionga, ambitu incifa , inextremo
obrufb mucronata, lanueinofa, inodora, prop& radicem
crcbriora ; caulem rotundum, femicubitalem ; cauli or-
biculatiin aflurgunt flofculi parvuli , colons obfcure
phcenicei , quingue foliorum , racematim ex caliculis
fubluteisprotruh, inodori; radix fibrofa. Floret Aprili,
id eft, auturani in iliis regionibus initio.
Sedum arborefcens promontorii
bona: fpei.'
Scium arborefcens promontorii bom fpei. Scclum arbo*
rcfcens folia digitum fcrb crafli, ex viridi albicantia ha-
bet, volx minus fermfc magnicudine.
Frutillaria craffa promontorii
boose fpei.
33?
FritillarU craffa (dentillaria: fpccies videtur)promon-
torii bona: fpei vel Soldanica a loco , foliorum loco ex-
crefcentias habet digiti longicudine & craifnic luccu-
lenta, angulis rupium inftar praeruptis, circumquaquc c-
rainentibits , & in aculc'um dcfinentibus. Folia hie in
fummo ad livorcm purpureumvcrgunt, inferius viridia;
quibus flos adnafcitur quinq; conftans foliis ccnacibus.iu
extremitatibus acuminatis ; quae corollam in medio po-
fuam cingunt , quaque ilium attingunt lata funt, Sc or-
biculato marginc in acumen vergunt. Corolla protu-
berat, intufque concavum habcc , cui alrcr quad nofeu-
lusiniidct, conftans quinquemucronatis foliolis in cxrre-
miratibus fufeis , cetera lutcis , ftamimbus aliquot in
centro phoeniccis. Color ftoris exterius fordidus.purpu-
reus eft ; intus fundus luteus maculis fufeis undique
confpergirur , ram in foliis quamin corolla. Caulisquo
flos foliis nedfitur, ftriis purpureis diftinguitur, vis femi-
digitilongitudinem habens, folia pnrva appcndicc radi-
cituberolx*, Candida: coherent. Plants plane inodora
eft , nafeitur locis inontanis arenofis.
Iris maria premont.
bonx fpei.
Jrisuvarlt Florcm habet coloris phcrnicci, quiuvu-
!is deorfnn propendenribus , capitelli vel cunei formam
habet. caulis bicubitalis, rotundus fufei coloris; folia
Jonga, dieiruin lata , ycroque latere acuta glandioli for-
ma; flos rortldi odoris eft. uafeitur locis paludofs.
i The first known illustration of a Kniphofia (here entitled
heophrastii Eresii de Historic Plan, arum. Amsterdam, 1644.
“ Iris uvaria ”) in Van Stapel's
375
Fig. 5. — An illustration of Kniphofia
1. 1 3, fig. 5, Berlin, 1696.
maria in Mentzel's Index Nominum Plantarum Multilinguis,
24379—2
376
Iris uvaria flore iuteo etc.: Jacob Breyne, Prodr. Fasc. Rar. PI. ed. 1, p. 74 (1688).
Aloe africana foetida , folio triangulari longissimo et angustissimo, radice lutea:
Kiggelaar, Horti Beaumont iani, p. 5 (1690).
Aloe folds linearibus triangularis etc.: Linnaeus, Hort. Cliff, p. 133 (1737).
Aloe africana foetida folio triangulari etc.: J. Burman, Cat. Plant. Afric. p. 2
(1737).
Aloe uvaria: Weinmann, Phyt. Icon, t.45 (1737).
POST-LINNAEAN HISTORY
1753. Aloe uvaria L., Sp. PI. 1 : 323 (1753). This name was also used by Linnaeus
in his Flora Capensis p. 4 (1759) and by other authors, such as Kniphof, Herb. Viv.
9: t.805 ( 1 762) ; N. L. Burman, Prodr. FI. Cap. p. 10 (1768); P. Miller, Diet, n.23
(1768); John Hill, Hort. Kew. p. 333, No. 11 (1769) etc. For a full bibliography,
see p. 493.
1771. Linnaeus transferred the species to the genus Aletris in his Mantissa Plan-
tarum pp. 367-368 (1771).
1783. Lamarck, Encycl. Meth. Bot. 1: 90 (1783), described the same species
under the name Aloe longifolia Lam.
1789. William Aiton, Hortus Kewensis 1: 464 (1789), listed Aletris uvaria and
described a second species, A. pumi/a Ait. This is a distinctive species with short,
funnel-shaped flowers, credited by Aiton as having been introduced to England from
the Cape by Francis Masson in 1774. However as pointed out in J. S. Afr. Bot. 29:
145 (1963), K. pumila is an Abyssinian species, and its long association with the South
African flora is based on an error. There is a specimen in the Herbarium of the British
Museum (Natural History) annotated “ Hort. Kew 1781, Aletris pumila ” (in Dryander’s
handwriting), which may be accepted as the type of K. pumila. It is not matched by
any material known to occur in South Africa, but is conspecific with Schimper specimens
collected in Ethiopia and described as K. comosa Hochst. (1844).
Francis Masson did, in fact, figure in the early history of the genus in South Africa
because there is another specimen in BM labelled “Aletris uvaria? Hort. Kew 1786
(e Cap b. Spei per Masson 1780)”. Although not recognized as such at the time, this
specimen represented a distinct species, the second to be found in South Africa. The
flowers are considerably shorter than in K. uvaria, but are longer than those of K.
pumila. The 1786 specimen clearly belongs to the species described many years later
as K. ensifolia Bak. (1885), and again as K. tuckii Bak. (1893) and K. rivularis Berger
(1908).
Masson apparently did not himself reach the areas where this species is known
to occur. One can only speculate how it came into his possession. Possibly the well-
known explorer and friend of Masson, Colonel Gordon, found the species during his
travels to the Orange River.
1794. Thunberg, who accompanied Masson on some of his travels, records
collecting only one species, Aletris uvaria, in his Prodr. Cap 1 : 60 (1794). It is evident,
however, that he and Masson penetrated eastwards into regions where other species
are found. Thus it was left to later collectors, probably before and about the turn of
the century, to introduce the more robust plants into Europe which came to alter the
concept of K. uvaria. One of these collectors was Scholl, who collected extensively
as far east as the Great Fish River, and who sent quantities of plants back to the Royal
Garden at Schoenbrunn, Vienna, towards the end of the eighteenth century.
377
Fig 6 Kniphofia uvaria illustrated by Commelin in Horti Medici Amstelaedamensis , Vol. 2, 1. 1 5
378
* *■* ' /£*>** r* u
Fig. 7. — Kniphofia uvaria, from a folio of early eighteenth-century water-colour illustrations in the
library of the Botanical Research Institute, Pretoria. The artist is not known, but the illustrations
are annotated by Johannes Burman, 1706-79 (see Fig. 8).
379
1794. The genus Kniphofia Moench, Meth. 631 (1794), was created to accommodate
Aloe uvaria L. Unfortunately, Moench renamed the species Kniphofia alodides Moench,
a procedure not unusual at the time, but not acceptable according to the present
International Code of Botanical Nomenclature. This was the first generic name to
be created specially for the Redhot Pokers.
1797. The third species to be described was illustrated in 1797 on Plate 54 of
Andrews’ Botanical Repository under the name A/etris sannentosa Andr. The figured
plant was cultivated in Lee’s Nursery at Hammersmith and was recorded as being
“introduced . . . from the Cape of Good Hope about 1789 ”. We now know
that the species occurs mainly in the Roggeveld, an area visited by Masson, although
his name is not linked with the introduction. The younger Aiton in Hort. Kew ed.
2, 2: 290 (1811) says it was “introduced about 1789 by Mr. Richard Williams”.
The species is distinct from K. uvaria and was never confused with it.
1799. Willdenow, Sp. PI. 2: 182 (1799), evidently unaware of the creation of
Kniphofia Moench, transferred two of the three known species to the genus Veltheimia,
making the combinations V. uvaria (L.) Willd. and V. pumila (Ait.) Willd.
1800 or 1801. The first illustration which calls for close scrutiny with regard to
its identity was published by Jacquin as Plate 4 of his Fragmenta, a series which appeared
in six fascicles from about 1801 to 1809. Plate 4, which was entitled Veltheimia uvaria
Willd., probably appeared in 1801 (see Konig and Sims in Ann. Bot. 1: 64, 1806).
Although its identity was not questioned by subsequent workers, it has several clear-cut
characters which distinguish it from true K. uvaria. A glance at the illustration (see
Fig. 9) immediately reveals its distinguishing features, namely: the cylindrical inflore-
scence; relatively short perianth tube, constricted above the ovary and expanding
Fig. 8. — Annotation in 1755 by Johannes Burman to the folio of early eighteenth-century drawings
in the library of the Botanical Research Institute, Pretoria (see Fig. 7).
380
rather abruptly, with long-exserted anthers; and the lanceolate, acuminate bracts.
The identity of the figured plant remained uncertain until recently, when it was matched
with a few gatherings from the Uniondale-Knysna Districts and specimens collected
first by Flanagan and later by Miss Bruce in the Komga District (see Figs. 49 and
50). The species is evidently comparatively rare at a few scattered localities. Specimens
corresponding to Jacquin’s plate are not present in the Herbarium of the Natural
History Museum, Vienna, but it was interesting to find matching specimens in the
herbaria of the British Museum (Natural History) and Geneva, made from specimens
cultivated in Europe during the early part of the Nineteenth Century (see p. 448). They may
immediately be recognized by the long, narrow, acuminate bracts, a characteristic
shown by K. praecox Bak. (1870), in which species the Jacquin plate is now included.
Jacquin does not explain the origin of his plants, but the most probable explanation
is that they were sent by the collector Scholl.
1804. Ker Gawler, editor of Curtis’s Botanical Magazine, no doubt also unaware
of the publication of Kniphofia Moench, described the genus Tritoma Ker Gawl.,
illustrating it with plates of the three species known to him: T. media Ker Gawl.,
which name he introduced as being more appropriate than “ sarmentosa ”, in Bot.
Mag. t.744 (1804); T. uvaria (L.) Ker Gawl., l.c., t.758; and T. pumi/a Ker Gawl.,
l.c., t.764.
1810 or 1811. Redoute published a plate in his Liliacees 5: t.291 (about 1810)
under the name Tritoma uvaria. The illustration does not, however, depict typical
K. uvaria, but rather one of the more robust plants from the eastern Cape Province.
Support for this view may be found in the more robust inflorescence with densely
overlapping buds and flowers (see Fig. 11). However, subsequent authors did not
question its identity and it is clear that, from about this time, the concept of K. uvaria
became altered to apply to one or more of the more robust species (see, for example,
Fig. 12 on page 385, showing the illustration which appeared under the name K. uvaria
(L.) Hook, in Bot. Mag. t.48 1 6, 1854). It is thus an interesting exercise to examine
the numerous literature references that have appeared under the specific epithets uvaria
and alodides in order to sort out the concepts involved. These two illustrations are
now included in K. linearifolia Bak., which was described only in 1896 (see pp. 498-504).
1811. William Townsend Aiton, in the second edition of Hortus Kewensis 2: 29
(1811), listed the three species Tritoma uvaria, T. media and T. pumi/a.
1821. A. G. Roth described Veltheimia speciosa in his Nov. PI. Sp. 190 (1821),
a publication dealing with plants collected by Dr. Benj. Heyne in India. He relates
it to V. uvaria and the description could apply to one of the large, showy Redhot Pokers
of the eastern Cape Province, for example, K. praecox Bak., K. linearifolia Bak. or
even K. rooperi (Moore) Lem. No type specimen has been traced and thus the
possibility cannot be exluded that the plant is not a Kniphofia at all. Heyne is not
known to have collected in South Africa. Therefore, although the name “ speciosa ”
would be an appropriate one for the robust species, and subsequent authors have
treated it as a synonym of “ uvaria ”, it is considered advisable to treat it as a name
of doubtful application.
1821. H. F. Link in Enum. PI. Hort. Reg. Bot. Berol. 1: 333 (1821) proposed
the name Tritomanthe Link in place of Tritoma Ker Gawl., pointing out that the name
Tritoma had previously been used for a genus of insects. He published the combina-
tions Tritomanthe uvaria (L.) Link, T. media (Ker Gawl.) Link and T. pumi/a (Ait.)
Link.
381
Fig. 9. — A robust Kniphofia illustrated by Jacquin in his Fragnrenta, t.4 (1800 or 1801). Although
named “ Veltheimia uvaria ”, it is not that species, but represents K. praecox subsp. bniceae Codd
(see p. 450, fig. 48).
382
Fig 10 The illustration in Botanical Magazine t.758 (1804) accompanying the description ol the
genus Tritoma Ker Gawl. Although reputed to be “ Tritoma uvana , it looks somewhat like
Jacquin’s plant (see Fig. 9).
383
///,•/
Fig. 11.— Although referred to as “ Tritoma uvaria ” in Redoute's Liliacees t.29 1 (1810 or 181 1), this
is probably the first illustration of the more robust plants now included in K. linearifolia (see
p. 498).
384
1823. Thunberg’s Flora Capensis edited by Schultes appeared in this year. Only
the one species Veltheimia uvaria is listed, with a more extensive description than is
given in his Prodromus (1794).
1829. Link, Handbuch 1: 170 (1829), published the generic name Tritomium
and made the combination Tritomium uvaria (L.) Link. No reasons are given for the
change.
1829. Schultes in Roem. & Schult., Syst. Yeg. 7: 629-631 (1829) follows Link
(1821) and lists the three species as Tritomanthe. He gives an extensive bibliography
under each species, but the descriptions are compilations from earlier authors. Under
T. uvaria he includes, amongst others, the Jacquin illustration (1800/01), Redoute’s
plate (1810/111 and V. speciosa Roth (1821), all of which are now excluded from K.
uvaria proper, though admittedly he expresses doubt about the identity of the Redoute
plate.
1836. It is evident that by this time the epithet “ uvaria ” was generally applied
to the more robust plants commonly cultivated in gardens, and which probably originated
in the eastern Cape Province. Hybridization and selection had also no doubt played
a part in modifying and improving the cultivated stock. Thus when true K. uvaria
was again introduced into England by Burchell, it was looked upon as a distinct species,
and was illustrated and described as Tritoma burchellii Herb, ex Lindl. in Bot. Reg.
t . 1 745 (1836). The error was perpetuated, even by monographers of the genus, so
that members of the large eastern species became consistently known as “ uvaria ”
or “ alooides ”, while the epithet “ burchellii ” came to be accepted as the correct
name for the true K. uvaria of the south-western Cape Province. See, for example,
Hooker (1854) and Baker (1870) below.
1841. Steudel, Nom. Bot. ed. 2: 718 (1841), repeats the three species listed by
Link (1821) and adds Tritomanthe burchellii (Herb, ex Lindl.) Steud. and T. odorata
Steud., a nomen nudum.
1843. The generic name Kniphofia Moench was taken up by Kunth, Enum. 4:
550-553 (1843) and the genus was reviewed. In addition to listing K. alooides Moench,
he made the combinations K. burchellii (Herb, ex Lindl.) Kunth, K. pumila (Ait.) Kunth
and K. sarmentosa (Andr.) Kunth, thus correctly restoring Andrews’ epithet. He had
before him the extensive collections made by Drege, from which he described three
new species, K. laxiflora, K. parviflora and K. triangularis.
1852. Tritoma rooperi Moore was described and figured in Gard. Comp. 1: 116
(1852). It is a robust plant with a large, globose inflorescence, obviously a component
of the K. uvaria complex, but relatively uniform throughout its distribution range
from about East London, along the coast to just north of Port Shepstone. The original
plant was cultivated in England from a plant sent by Capt. Edward Rooper, who was
stationed in East London from time to time between 1848 and 1850. In the library
of the Botanical Research Institute, Pretoria, is a collection of original water-colour
paintings by Capt. Rooper, among which Plate 99 is an illustration which is a good
match of the figure in the Gardener’s Companion (1852). The illustration is annotated
by the artist: “ Common in Albany and Kaffirland. Flowers nearly all year.
March ’49 ”.
1854. In the Bot. Mag. t.48 16 (1854), the combination Kniphofia uvaria (L.) Hook,
was published. This is the correct name for the species from the S.W. Cape, but the
illustration published by Hooker (see Fig. 12) shows the robust eastern species which
was, by that time, commonly known as “ uvaria ”. In fact. Hooker refers to the ealier
Bot. Mag. Plate 758 (1804) in the following terms: “ The figure of Mr. Gawler is so
indifferent that we can well excuse its not being easily recognised and have no hesitation
in offering a better one ”.
385
Fig 12.— Probably a specimen
in Bot. Mag. t.4816 (1854)
of the robust species now known as
, when he effected the combination
K. linearifolia , figured by Hooker
K. uvaria (L.) Hook.
386
1866. In R. A. Salisbury’s Gen. PI. (p. 75), published nearly 40 years after his
death, the generic name Triclissa Salisb. is proposed in place of Tritoma Ker Gawl.,
mention again being made of the prior use of the name Tritoma for a genus of insects.
1870. J. G. Baker, subsequently Keeper of the Herbarium, Kew, described K.
praecox Bak. with an accompanying illustration (see Fig. 46) in Saunders’ Ref. Bot. 3:
1. 1 69 (1870). Baker advanced the name with some diffidence, stating that it differed
from “ K. alooides ” only in its earlier flowering time. Later he reduced it to a synonym
of “ K. alooides ”. It appears to be the first valid specific name for the robust eastern
species illustrated by Jacquin (see Fig. 9, p. 381). Some years before this, plants had
been introduced into horticulture under such names as T. recurva and T. recurvata
(1863), and T. grandis (1865). Although some of these are accompanied by illustrations
and resemble K. praecox or the species described later as K. iinearifolia, their origin
is uncertain and they may even be garden hybrids. No corresponding herbarium
specimens have been traced and these horticultural names are best disregarded as
being of uncertain application.
K. praecox is based on a plant cultivated in England and said to have been received
from Thomas Cooper (see also p. 445), who came to South Africa to collect plants
for Mr. W. W. Saunders of Reigate. Although the original locality is not known,
there is a specimen in the Herbarium of the Royal Botanic Gardens, Kew, which may
be regarded as the type. Cooper spent from 1859 to 1862 in South Africa and
proceeded from Port Elizabeth to the north-eastern Cape Province and thence through
the eastern Orange Free State to the Basutoland border. Continuing round the
northern boundary of Basutoland, he crossed the Drakensberg and eventually returned
via Port Natal. Unfortunately his localities are usually vague. Two further species
were based on his gatherings, namely K. breviflora Harv. ex Bak. (1871) and K. porphy-
rantha Bak. (1874), from the Orange Free State.
1871. Baker published a review of the genus in J. Linn. Soc. 11: 360 (1871) in
which he recognized 14 species, nine of which were recorded as originating in South
Africa.
1883. Bentham & Hook, f., Genera Plantarum 3: 775 (1883), considered that
there were 16 species of Kniphojia. In this work the segregate genus Notosceptrum
was described, based on two species from Angola.
1885. A synopsis of the South African species, including the first key, was
published by Baker in J. Bot. Lond. 23: 275 (1885). Eighteen species and three varieties
were upheld, several of which were newly described.
1888. Engler reviewed the genus briefly in Pflanzenfamilien 2: 5 (1888) but made
no useful contribution to our knowledge of the group. Only 16 species were recognized.
1896. Baker, in Flora Capensis 6: 275 (1896) and 6: 533 (1897), revised the South
African species. A further six species were described in this work, bringing the total
to 35 for South Africa. Two South African species of Notosceptrum were upheld.
About this time. Max Leichtlin, a nurseryman in Baden Baden, Germany, began
to take a particular interest in the genus Kniphojia and went to considerable trouble
to import seeds and plants from South Africa. One of his most active helpers was
Medley Wood in Durban. He regularly supplied specimens to Kew, on which Baker
based several new species, including K. leichtlinii (from tropical Africa), K. citrina,
K. longicollis, K. rufa and K. tuckii. Unfortunately, precise information regarding
original localities of these introductions is usually lacking and they cannot always
be matched exactly with naturally occurring plants. This applies particularly to the
type specimens of K. longicollis and K. rufa.
387
1898. The species of tropical Africa were dealt with by Baker in FI. Trop. Afr.
7: 450 (1898), in which 13 species of Kniphofia and 2 species of Notosceptrum were
dealt with.
1908. The genera Kniphofia and Notosceptrum were monographed in their entirety
by Alwin Berger, the German horticulturalist and botanist, in his monumental work
on the Aloineae published in Pflanzenreich 4, 38: 31-72 (1908). Berger was Curator
of the famous La Mortola Gardens in Italy from 1899 to 1918 and was particularly
interested in succulent plants. Of 67 species and 13 varieties of Kniphofia recognized,
44 species and 13 varieties were recorded from South Africa, including three newly
described species. Five species of Notosceptrum were upheld, of which three were
South African.
Berger’s work served an important purpose in bringing together the greater part
of the information available at the time, including a comprehensive list of hybrids and
garden cultivars. It is evident that he consulted the types represented in Berlin and
Kew, covering most species, several of which he illustrated by means of black-and-
white drawings. The descriptions of the species are carefully prepared and the full
transcriptions of notes appearing on collectors’ labels are often helpful. In those cases
where no type material remains after the extensive damage to the Berlin Herbarium
during World War II (e.g. K. bachmannii, K. linearifolia and K. decaphlebia), the
information in Berger’s monograph is of particular value in interpreting the species.
Unfortunately, Berger’s work suffers from two main weaknesses. Firstly, his
nomenclature is, in several cases, not acceptable according to our present International
Code. Examples maybe seen in discussions under K. buchananii (p.416), K. breviflora
(p. 416), K. parviflora (p. 405) and K. gracilis (p. 427).
Secondly, he had no opportunity to study the plants in the field, while very scanty
material apart from the type specimens was available to him, which makes his treatment
unrealistic in several respects. Thus he shows no grasp of geographical distribution
of species and little concept of the variation which occurs within a taxonomic group,
resulting in too many “species” (among those known to him) being upheld on
unreliable “ differences In practice, therefore, it is often difficult to name specimens
according to Berger’s treatment and it would seem as if his monograph had a stultifying
effect on the progress of our knowledge of the genus. Since 1908, a considerable
quantity of herbarium material has accumulated in South African herbaria, but little
constructive taxonomic work was undertaken until after World War II.
From 1947 to 1952, Miss Eileen A. Bruce (see Fig. 51, p. 452), previously of the
Herbarium of the Royal Botanic Gardens, Kew, was employed at the National
Herbarium, Pretoria, and started a revision of the South African Kniphofia species.
She came to South Africa with a sound background of taxonomic research and set
about her studies with typical enthusiasm and energy, being determined to see as
many species as possible in their natural habitats. After her return to Kew in 1952,
she found it necessary to relegate her Kniphofia studies to periods of spare time between
work on other groups for the Flora of East Tropical Africa and, unfortunately, her
career was cut short by her untimely death in 1955. Three species which she recognized
as new, and for which she had drawn up descriptions, were published posthumously
in Flowering Plants of Africa, Yol. 30 (1955). They are: K. splendida E. A. Bruce,
K. coralligemma E. A. Bruce and K. rigidifolia E. A. Bruce.
As I had been associated with Miss Bruce in many of her field excursions in my
capacity as Officer in Charge of the Botanical Survey Section and had continued to
co-operate in her Kniphofia studies after her return to Kew, it was fitting that I should
continue with the revision of the genus. Several additional species have been illustrated
in Flowering Plants of Africa since 1955, including some new to science, while a review
388
of the Transvaal species appeared in Wild Flowers of the Transvaal by Cythna Letty
(1962). After the criticisms directed at Berger, it must be confessed that, in the last-
mentioned publication, the plant with the robust inflorescence illustrated under the
name K. praecox Bak. should, in fact, be called K. linearifolia Bak.
While it is felt that the identity of K. praecox has been established (p. 370), it
should be realised that there are still many problems which have not been satisfactorily
dealt with, and for which much more field work is necessary than could be allocated
during the present revision. The eastern Cape Province, in particular, is of interest
because here a complicated interaction is found between complexes of K. uvaria, K.
rooperi, K. linearifolia and their allies. Intensive studies are required of variation within
populations and between neighbouring populations of this group of species before
a satisfactory classification can be achieved. Also, more extensive and thorough
collecting is required in order to ensure that a good representation of the distribution
of each species is obtained. Similarly, in Natal, an attempt should be made to analyse
the variation now included within K. gracilis, while closer study is required of the
interactions within the K. rufa — K. ichopensis and K. linearifolia— K. tysonii complexes.
The present revision should, therefore, be regarded essentially as a framework for
further study.
TAXONOMIC VALUE OF CHARACTERS
Root System. The subterranean part of the plant consists of a thick rhizome
from which arise numerous somewhat fleshy roots. In most species the rhizome
divides, forming groups of stems, while in others the stems are more or less solitary.
In K. sarmentosa, slender rhizomes 4-10 cm long arise from the main, thick rhizome,
so that new shoots may arise at a short distance from the parent plant.
Stem. The great majority of species do not produce an aerial stem, but exceptions
occur and stems are usually to be seen on old plants of K. caulescens and K. northiae,
reaching to a height of 30 cm in the former and 60 cm or more in the latter.
Leaves. All species of Kniphofia have non-succulent leaves and this is a useful
character for separating them from the closely related genus Aloe. Leaf characters
are also useful in recognising species, and variation may be seen in (a) arrangement
of leaves and whether erect or spreading; ( b ) shape, e.g. length, breadth, cross section;
(c) texture and colour; and (d) character of leaf margin. These characters are elaborated
below.
{a) Arrangement. The leaves may or may not be arranged in distinct ranks. A
distichous arrangement is to be seen in K. typhoides , K. brachystachya and K. ichopensis
but, when distinct ranks are visible in the broader-leaved species, there are usually
4 or 5, as in K. rooperi and allied species, and in K. multiflora. In some species the
leaves are spreading arcuate or recurved, e.g. K. drepanophylla and K. citrina, but in
the great majority of species the leaves are erect at first, later sharply bent (kneed)
at or below the middle due to their relatively flaccid texture. In K. stricta the leaves
remain rigidly erect.
(, b ) Shape. Length and breadth (especially the latter) are useful diagnostic
characters, but are liable to be affected by habitat factors, for example cultivation,
and so allowance must be made for this when dealing with cultivated plants. The
leaves die down annually (except in species such as K. northiae ) and are produced
again in spring in clusters at the end of each growing point, from 6 to 16 leaves,
appearing in such a cluster. The extreme outermost of these leaves are often broad
scale-like structures and may be ignored; measurements should be taken from those
outer leaves (of which there may be 3 to 8) which develop more or less to their maximum
389
length. In the centre of the cluster, the leaves become shorter and narrower to more
or less filiform, and these should also be ignored for diagnostic purposes. The peduncle
arises in the centre of the leaf-cluster and, in early-flowering varieties, usually rapidly
overtops the leaves. The cross section of the leaf may also be diagnostic. In most
species the leaves have a distinct keel to the leaf, in which cases the keel forms the base
of a broad or shallow V in cross section. Exceptions are K. stricta (U-shaped) and
K. northiae (broadly U-shaped to almost flat). The depth to which the leaf is channelled
on the upper surface also varies in different species. In some with very narrow leaves
there may be practically no channel above, so that the cross section is an inverted
triangle.
(c) Texture and colour. These characters are not easy to define and are usually
omitted from collectors’ notes. Texture varies from the soft, flaccid leaves of K.
sarmentosa and K. porphyrantha to the fibrous, rigid leaves of K. stricta and K. uvaria.
When the leaves die down annually, the bases of the leaves may break into separate
fibres which persist at the base of the plant in several species. Leaf colour varies from
yellow-green to milky-glaucous. Texture can usually be recognized in the Herbarium
but colour, although conspicuous in living plants, is not easily detected in dried
specimens.
( d ) Leaf Margin. The leaf margin, and often the keel also, may bear minute
teeth (serrulations), which are characteristic for certain species. Other species have
consistently smooth-margined leaves. There are also several species in which the
character is not constant, especially when the teeth are few and scattered (sometimes
only towards the apex of the leaf, or present on some leaves and absent on others).
The leaf surface is glabrous in all except one species, namely, K. hirsuta, although
K. stricta may occasionally have small papilla-like outgrowths on the main nerves
as well as on the margin.
Peduncle. Length and stoutness of peduncle are reasonably characteristic for
particular species but have little diagnostic value except, possibly, in the case of K.
multiflora, which has a long scape 1 -5-2 m in length and a markedly elongate inflore-
scence, making it the tallest member of the genus in South Africa (2-3 m tall). The
shortest species is probably K. brachystachya with a peduncle about 30 cm long. The
peduncle projects above the leaves in the great majority of species (in many of which
the leaves are equally long, but are bent downwards), one of the exceptions being
K. latifolia. In this species the broad, erect leaves overtop the inflorescence at the
flowering stage, but the peduncle elongates exceeding the leaves at the fruiting stage.
The peduncle is normally simple but branched peduncles are occasionally found in
K. tabularis.
Inflorescence, (a) Shape and density. The shape of the inflorescence is a useful
species characteristic, best observed in living plants, where an impression of average
shape can be obtained. In the herbarium this is not always easy, and odd or abnormal
inflorescences, collected deliberately or through being the only one available, may
cause difficulty. The inflorescence varies from narrowly cylindrical (markedly elongate
in K. multiflora ) through narrowly ovoid and ovoid to globose or even subcapitate
(for example in some specimens of K. drepanophylla). As flowering progresses, an
inflorescence may change in shape from ovoid (conical at the apex), through rhomboid
to obovoid. In the descriptions of each species the shape is indicated at approximately
the mid-flowering stage, when about half the flowers in the inflorescence have opened.
Length and diameter (in herbarium specimens this is measured as width) should be
measured at this stage. The reliability of these figures obviously depends on the material
available and the efficiency with which the specimens have been prepared. While in
most species the flowers are densely placed, often densely overlapping so that the
rhachis is not visible, in other species the flowers are fewer per unit length and may be
390
widely spaced (as in K. pauciflora, K. rufa and K. tabularis ). Variation in density
may be found within certain species, such as K. gracilis , K. laxiflora and K. ichopensis.
Normally the buds are more densely placed than the open flowers, due to the elongation
of the rhachis as flowering progresses but, in certain species, the apex of the inflore-
scence may be more lax than the central or lower portion.
( b ) Colour. In general the buds and unopened flowers tend to be more intensely
pigmented than the open flowers. This is to be seen in the commonly cultivated
Redhot Pokers in which the apex of the inflorescence (buds and young flowers) is
brilliant scarlet in colour, grading down to the yellow or orange-yellow flowers towards
the base of the inflorescence. This pattern may be observed in many species, although
the red pigment may vary in intensity and colour, or be lacking, while the flowers
may be yellowish to yellow-green. In several species, on the other hand, the inflore-
scence may be more or less concolorous. In one group of species, consisting of K.
evansii, K. triangularis (together with its subspecies) and a colour form of K. rufa,
the inflorescence is more or less uniformly coral-red in colour, a fact which may be
used in separating these from closely allied plants. Some species have white or yellow
flowers, which are slightly more pigmented or sometimes tinged with red at the apex
of the inflorescence. Among these may be listed K. albescens, K. buchananii , K. fibrosa
and K. crassifolia (always white); K. brevifolia, K. ensifolia, K. gracilis and K. multiflora
(white or yellow-flowered forms); K. acraea , K. brachystachya and K. pauciflora
(pale yellow); K. parviflora (greenish brown); while in two species the flowers are
chocolate-brown, namely K. typhoides and K. umbrina.
(c) Scent. Although in the description of K. uvaria , the first species known in
Europe, Linnaeus repeats references by earlier authors to the “ foetid ” flowers of this
species, it must be accepted (see p.373) that this reference became accidentally transferred
from another plant ( Stapelia variegata ) to the Kniphofia. The great majority of
Kniphofia species, including K. uvaria , may be described as having scentless flowers.
However, a faintly sweet scent is found in certain species, all of which have short,
yellow or brownish flowers, and which form a closely related group, namely K. brachy-
stachya, K. parviflora, K. typhoides and K. umbrina.
(d) Flowering sequence. In all South African species, flowering progresses from
the base of the inflorescence upwards (acropetal), with the exception of an occasional
freak inflorescence which may start at the apex (basipetal). N. E. Brown, in Gardener’s
Chronicle of January 25th (1910), p. 35, records an inflorescence of this kind. However,
in Ethiopia, at least two species, namely K. pumila (Ait.) Kunth ( =K . comosa Hochst.)
and the closely related K. isoetifolia Engl., may flower in either basipetal or acropetal
sequence. An investigation of the underlying causes for this should make an interesting
study.
(e) Flowering Time. Each species has a relatively constant flowering season.
Abnormal weather conditions may cause occasional flowers to appear at unusual
times but, in general, flowering of a particular species may be predicted to within a
few weeks. Exceptions are K. rooperi, which is one of the few winter-flowering species
but which has an extended flowering period from about May to November, and K.
uvaria, which may flower at almost any time of the year depending on veld fires.
Cultivated plants, especially during the first year they are brought into cultivation,
may deviate from their normal flowering times but, once established, they usually
settle down to a regular cycle corresponding approximately to the species in the wild
state. In Table 1 the main flowering times are indicated for each species (x = main
flowering season; o = occasional flowers recorded). If used in conjunction with the
key based on morphological characters, this table may assist with the identification
of specimens collected in the wild state. As nearly all species are dormant in winter,
the months are listed from July to June.
391
Table 1. — Flowering times of Kniphofia species
Bracts. Sterile bracts are often found as a small coma at the apex of the inflore-
scence. Also, below the inflorescence, occasional sterile bracts are produced which
may be broadly deltoid to long-caudate, almost leaf-like structures. Although accorded
some importance by certain authors, sterile bracts do not appear to have much
taxonomic significance.
24379—3
392
Fertile bracts, each of which subtends a flower in the inflorescence are, on the
other hand, useful diagnostic characters in many cases. Although small and requiring
a lens for adequate study, they show characteristics of shape which are relatively
constant for a particular species. They are thin-textured, chartaceous structures,
somewhat concave on the upper surface, varying in colour from white to brownish.
The brown coloration occurs mainly at the apex and along the median nerve. Usually
only one nerve is present but, in a few species (chiefly in the K. uvaria complex), the
bracts towards the base of the inflorescence may have 3 or 5 nerves. In shape, the
bracts vary from subrotund or ovate to linear-lanceolate, with the apex varying from
rounded to long-acuminate. Cultivation may result in an increase in the length of the
bracts, but the shape (and particularly the shape of the apex) remains fairly constant.
As in the leaf margin, the margins of the bracts vary from entire to serrulate or irregular-
ly toothed (eroso-denticulate).
Pedicels. Pedicel length, which should be measured at anthesis, does not vary
a great deal, but can be useful in separating species such as K. uvaria (L.) Hook, (pedicels
3-6 mm) and K. linear folia Bak. (pedicels 0-5-3 mm). The longest pedicels are found
in K. praecox Bak. (up to 10 mm). The bracts and pedicels are persistent. If the
ovary is not fertilized, the ovary and perianth are soon shed, in which case the pedicel
does not elongate. If the ovary develops into a fruit, the pedicel elongates slightly
and carries the fruit in a horizontal or ascending position.
Perianth. Length. The perianth consists of a tubular portion which varies in
length from about 1 - 5 to 50 mm, with six equal lobes which vary scarcely at all (1 -5-
3 mm in length) between species. The width of the tube at the base is also fairly constant
(1-5-3 mm). Perianth length can often be used as a diagnostic character though,
in cultivated plants, it is less reliable than in naturally occurring plants.
(a) Shape. In the majority of species the perianth tube is subcylindrical. Excep-
tions are to be found in those species (a) with very short perianths (1-5-3 mm long and
1 -5-2-5 mm in diameter) in which the tube may be described as sub-campanulate;
and ( b ) in which the tube expands noticeably from about the middle, resulting in a
relatively funnel-shaped tube.
( b ) Lobes. The perianth lobes are relatively constant in size throughout the genus,
rarely exceeding 3 mm in length so that, in species with very short perianth tubes, the
lobes may be as long as the perianth tube, or even exceed it. It was on such species
that the genus Notosceptrum Benth. was based, but there is a complete gradation
from long to extremely short perianth tubes so, in the present revision, Notosceptrum
is not upheld as a distinct genus. The lobes tend to spread more in some species than
in others and this led Berger (1908) to create a section Obtusilobae to accommodate
a group of species (see K. triangularis, p. 468) with this characteristic.
Fruit. The 3-locular, capsular fruit is a generic characteristic which varies a
little in size and, in outline, from subglobose to 3-angled (triquetrous). In general,
the fruit has little diagnostic value and, in fact, is rarely available to the herbarium
worker. K. littoralis Codd (p. 478) has relatively large, acute capsules up to 10 mm
long, which assist in separating this species from its allies, K. baurii Bak. (p. 482) and
K. elegans Codd (p. 480). Seeds have not been thoroughly studied and are seldom
available in the herbarium. Those that have been examined have not shown distinctive
characteristics of taxonomic value.
NATURAL AND GARDEN HYBRIDS
All species investigated have 2n = 12 chromosomes and it is evident that hybrids
are readily formed when two or more species flower together in a garden. N. E.
Brown makes a timely reference to this in the Gardeners’ Chronicle 56: 410 (1914)
393
when he says: “ Indeed, from the investigations I have made, I am of the opinion
that in a very great majority if cases plants of this genus raised from seed produced
in any garden where more than one kind is grown, whether that garden be in South
Africa or Europe, will not be true to name, but will be more or less affected by
hybridization, unless precaution has been taken to fertilize the plant with pollen of the
same species and to exclude all insect visitors
Unfortunately, species were frequently described in Europe from cultivated plants
and there is evidence that the types of the following species are of hybrid origin: K.
longicollis Bak., K. primulina Bak. and K. rufa Bak. Many new cultivars have been
produced by horticulturalists, and Berger (1908) provides an extensive list of those
found in European gardens, with an indication of their parentage, when known.
Two hybrid Redhot Pokers are commonly found in South African gardens. One
of these flowers in summer and has relatively narrow, dull-green, erect leaves. It is
near to K. praecox and is illustrated on p. 447. The other flowers in winter and has
broad, yellow-green, flaccid leaves. It is possibly the var. serotina listed by Baker in
Flora Capensis under “ K. alooides ” but, as no type specimen has been traced, its
identity is uncertain.
In nature, where it is found that one species grades into another, it may be assumed
that hybridization has contributed to the presence of intermediates and will continue
to do so while neighbouring populations flower at approximately the same time.
Hybrids between widely different species are rare, probably because when more than
one species occurs in an area they either do not flower together or are separated by
different habitat requirements. The following few specimens may be regarded with
a fair degree of certainty as being hybrids between the species indicated, the supposed
parents being listed in alphabetical sequence.
E. evansii X K. porphyrantha: Trciuseld 741 from Giants Castle Game Reserve,
Estcourt District, Natal.
K. ensifolia subsp. autumnalis X K. linearifolia: Jacobsz s.n. from Rensburgskop
and Codd 10529 from Kerkenberg, both localities being in Harrismith District, O.F.S.
Acknowledgements
The Directors of the following institutions and herbaria kindly sent specimens on
loan, including types, during the course of this study and their assistance is gratefully
acknowledged: — -
Albany Museum Herbarium, Grahamstown.
Bolus Herbarium, Cape Town.
Royal Botanic Gardens, Kew.
National Botanic Gardens, Kirstenbosch.
Natal Herbarium, Durban.
State Herbarium, Stellenbosch.
Natural History Museum, Stockholm.
Instituut fur Systematische Botanik, Zurich.
The internationally accepted abbreviations for herbaria (Index Herbariorum) are
given after all type specimens which have been seen but, in citing specimens in the
text, a herbarium abbreviation is added only in those cases where the gathering is not
represented in the National Herbarium, Pretoria.
I am particularly grateful to Miss M. D. Gunn, Librarian of the Institute, for much
historical information regarding early botanical collectors and literature.
394
KNIPHOFIA
Kniphofia Moench, Meth. 631 (1794); Kunth, Enum. 4: 551 (1843); Hook, in Bot.
Mag. t.48 1 6 (1854); Bak. in J. Linn. Soc. 11: 360 (1871); Benth. & Hook, f., Gen.
PI. 3: 775 (1883); Bak. in J. Bot. Lond. 23: 275 (1885); Engl., Pflanzenfam. 2: 5
(1888): Dur. & Schinz, Consp. FI. Afr. 5: 291 (1895); Bak. in FI. Cap. 6: 275 (1896);
FI. Trop. Afr. 7: 450 (1898); Berger in Pflanzenr. 4, 38: 31 (1908); Marloth, FI.
S. Afr. 4: 102(1915); Krause in Pflanzenfam. 15a: 299 (1930); Phillips, Gen. S. Afr.
Flow. PI. ed. 2: 185 (1951); Letty, Wild Flows. Tvaal. 32 (1962). Type species: K.
uvaria (L.) Hook.
Aloe L., Sp. PI. 1: 323 (1753), partly; Mill., Diet. ed. 8: n.23 (1768), partly.
Aletris L., Mant. PI. 367 (1771), partly; Ait., Hort. Kew. 1: 464 (1789), partly.
Veltheima sensu Willed., Sp. PI. 2: 182 (1799), partly; Thunb., FI. Cap. ed. Schult. 309 (1823).
Tritoma Ker Gawl. in Bot. Mag. t.758 (1804); Ait.f., Hort. Kew. ed. 2, 2: 290 (1811).
Tritomanthe Link., Enum. PI. 1: 333 (1821); Schult. in Roem. & Schult., Syst. Veg. 7: 631 (1829).
Tritomium Link, Handb. 1: 170 (1829).
Rudolphoroemeria Steud. ex Hochst. in Flora 27: 30 (1844).
Triclissa Salisb., Gen. PI. 75 (1866).
Notosceptrum Benth. in Benth. & Hook. f., Gen. PI. 3: 775 (1883); Engl, in Pflanzenfam. 2: 5 (1888);
Dur. & Schinz, Consp. FI. Afr. 5: 291 (1895); Bak. in FI. Cap. 6: 285 (1896); FI. Trop. Afr. 7: 454
(1898); Berger in Pflanzenr. 4, 38: 69 (1908); Krause in Pflanzenfam. 15a: 300 (1930); Phillips,
Gen. S. Afr. Flow. PI. ed. 2: 185 (1951).
Plants perennial, herbaceous, caespitose or solitary from a thick rhizome, rarely
with a well-developed stem; rhizome simple or branched. Leaves rosulate, usually
in 4 or 5 ranks, rarely distichous, linear, tapering gradually to the apex, usually
keeled; margin smooth to minutely serrulate. Peduncle terminal, stout, erect, subequal
to the leaves, simple or very rarely branched, naked except for occasional sterile bracts
below the inflorescence. Inflorescence a subspicate raceme of, usually, numerous
flowers, dense or somewhat lax; bracts scarious, persistent, longer than the pedicels;
pedicels short to almost obsolete, articulated at the apex; flowers spreading or
pendulous, white, yellow, brownish or various shades of red, the red pigment often
more conspicuous at the apex producing a bicolorous appearance. Perianth-tube
campanulate to cylindrical or somewhat funnel-shaped, varying in length from 1-5-
45 mm, lobed; lobes short, subequal, 1-5-5 mm. Stamens 6, hypogynous, usually
as long as or longer than the perianth at anthesis, the 3 stamens opposite the inner
segments longer than the others; anthers dorsifixed, versatile, dehiscing introrsely.
Ovary sessile, ovoid, 3-chambered, with many ovules in each chamber; style filiform,
subequal to the stamens at anthesis, usually finally exserted; stigma minute, capitate.
Fruit globose to ovoid, often triquetrous with loculicidal dehiscence. Seed somewhat
flattened, acutely 3-angled or winged; endosperm fleshy.
A genus of about 70 species distributed essentially in eastern and southern Africa,
with one species in Madagascar and one in southern Arabia. In South Africa, 45
species are recognized (four of which are subdivided into subspecies), and are found
mainly along the mountain ranges and in moist places. No species have been recorded
from South West Africa or Botswana.
The status of the genus Notosceptrum Benth. was discussed in Bot. Notiser 120:
53 (1967), and it was concluded that insufficient grounds existed for maintaining it
as distinct from Kniphofia L. The South African species K. multiflora (see p. 402)
provides some basis for this conclusion. Although it has a markedly elongate inflore-
scence with ascending flowers, as in typical Notosceptrum species, the perianth is
cylindrical, 7-12 mm long (as in Kniphofia ), not campanulate nor deeply lobed as in
Notosceptrum. Some grounds may be adduced for upholding Notosceptrum as a
subgenus based on its distinctive facies, in which case K. multiflora would be accom-
395
modated in this subgenus. However, no worthwhile objective can be seen in such a
step. Nor can any great advantage be seen in separating Kniphofia into sections as
was done by Berger in his monograph (l.c.). Many of the affinities between species
as indicated by Berger appear to be sound, though a good deal of recasting of sections
would be necessary in order to produce reasonably satisfactory subdivisions.
In the outline given below, the South African species are grouped into sections
according to their probable affinities which provide the basis for the sequence of species
in the present revision.
Sect. 1. Plants of robust stature; leaves broad; inflorescence markedly elongate',
flowers very numerous, short, ascending. (Sect. Multiflorae Berger, l.c. p. 44).
1. K. multiflora Wood & Evans, p. 400.
Sect. 2. Plants of small to medium stature; inflorescence very dense to medium-
lax, greenish-yellow to dark brown', flowers very short, often scented. (Sect. Parviflorae
Berger, l.c. p. 38, partly, and genus Notosceptrum sensu Berger, l.c. p. 69, partly).
2. K. parviflora Kunth, p. 403.
3. K. acraea Codd, p. 405.
4. K. umbrina Codd, p. 406.
5. K. typhoides Codd, p. 407.
6. K. brachystachya (A. Zahlbr.) Codd, p. 410.
Sect. 3. Plants of fairly small stature with narrow, grass-tike leaves', inflorescence
small, dense, white to yellow (red at the apex in K. flammula)', flowers very short to
medium. (Sect. Parviflorae Berger, l.c. p. 38, partly, and Sect. Pauciflorae Berger, l.c
p. 41, partly).
7. K. buchananii Bak., p. 411.
8. K. breviflora Harv. ex Bak., p. 414.
9. K. albescens Codd, p. 417.
10. K. flammula Codd, p. 418.
12. K. fibrosa Bak., p. 422.
13. K. crassifolia Bak., p. 424.
Sect. 4. Plants of small to medium stature; leaves fairly narrow; inflorescence
lax; flowers medium length to longish. (Sect. Pauciflorae Berger, l.c. p. 41, partly;
Sect. Modestae Berger, l.c. p. 42; and Sect. Laxiflorae Berger, l.c. p. 50).
14. K. gracilis Harv. ex Bak., p. 424.
15. K. laxiflora Kunth, p. 428.
16. K. ichopensis Bak. ex Schinz, p. 433.
17. K. rufa Bak., p. 434.
18. K. pauciflora Bak., p. 438.
19. K. tabularis Marl., p. 438.
Sect. 5. Plants of medium to robust stature; leaves medium to broad; inflorescence
very dense, often large, oblong; bracts lanceolate, acuminate ; flowers medium-length
to long, anthers conspicuously exserted in some species. (Sect. Comosae Berger, l.c.
p. 44; Sect. Densiflorae Berger, l.c. p. 55, partly; Sect. Uvaria Berger, partly; and
Sect. Caulescentes Berger, l.c. p. 65).
20. K. ensifolia Bak., p. 440.
21. K. splendida E. A. Bruce, p. 442.
396
22. K. praecox Bak., p. 444.
23. K. sarmentosa (Andr.) Kunth., p. 453.
24. K. coralligemma E. A. Bruce, p. 453.
25. K. caulescens Bak., p. 455.
26. K. ritualis Codd, p. 458.
27. K. hirsuta Codd, p. 460.
Sect. 6. Plants of medium to robust stature; leaves medium to very broad, not
distinctly keeled ; inflorescence very dense; bracts oblong to ovate-lancolate; flowers
fairly long. (Sect. Aioifoiiae Berger, l.c. p. 65).
28. K. stricta Codd, p. 461.
29. K. northiae Bak., p. 464.
Sect. 7. Plants of small to medium stature with narrow-leaves', inflorescence
dense, small, concoiorous, orange-red in colour', bracts lanceolate, acuminate', flowers
medium to longish; perianth lobes spreading. (Sect. Pauciflorae Berger, l.c. p. 41,
partly, and Sect. Obtusilobae Berger, l.c. p. 53).
11. K. evansii Bak., p. 419.
31. K. triangularis Kunth, p. 468.
Sect. 8. Plants of medium to smallish stature with narrow-leaves; inflorescence
dense, small, bicolorous; bracts lanceolate, acuminate; flowers long; perianth lobes
not spreading. (Sect. Uvaria Berger, l.c. p. 56, partly).
30. K. galpinii Bak., p. 466.
32. K. thodei Bak., p. 472.
33. K. porphyrantha Bak., p. 473.
34. K. fluviatilis Codd, p. 476.
Sect. 9. Plants of medium to smallish stature; leaves of medium breadth;
inflorescence dense, subglobose, mainly spring-flowering; bracts ovate, obtuse; flowers
long. (Sect. Isoetifolia Berger, l.c. p. 46).
35. K. littoralis Codd, p. 478.
36. K. elegans Codd, p. 480.
37. K. drepanophylla Bak., p. 481.
38. K. baurii Bak., p. 482.
Sect. 10. Plants of medium to robust stature; leaves medium to broad; inflore-
scence dense, often large, globose to oblong-cylindrical, spring or autumn-flowering;
bracts ovate to ovate-oblong, obtuse; flowers medium to long. (Sect. Densiflorae
Berger, l.c. p. 55, partly, and Sect. Uvaria Berger, l.c. p. 56, partly).
39. K. latifolia Codd, p. 484.
40. K. rigidifolia E. A. Bruce, p. 487.
41. K. rooperi (Moore) Lem., p. 488.
42. K. uvaria (L.) Hook, p. 492.
43. K. citrina Bak., p. 497.
44. K. linearifolia Bak., p. 498.
45. K. tysonii Bak., p. 504.
397
Key to Species
Perianth less than 2 cm long:
Inflorescence elongate, 30-80 cm long; flowers 7-12 mm long, ascending.... 1. K. multiflora
Inflorescence less than 30 cm long; flowers spreading or pendulous:
Perianth 4-10 mm long:
Inflorescence secund 2. K. parviflura
Inflorescence not secund:
Flowers yellow-brown to fuscous, drying dark purplish-brown:
Bracts linear-lanceolate, acuminate; perianth 8-9 mm long 4. K. umbrina
Bracts ovate-deltoid to subrotund; perianth 4-6-5 mm long:
Inflorescence 15-30 cm long; leaves 35-65 cm long 5. K. typhoides
Inflorescence 4-9 cm long; leaves 15-28 cm long 6. K. brachystachya
Flowers yellow, cream or white, not drying dark purplish-brown:
Bracts broadly ovate, obtuse; leaves 10-15 mm broad, margin minutely denticuiate
3. K. acraea
Bracts ovate-lanceolate, acute; leaves 2-5-6 mm broad, margin smooth:
Perianth 4-5 -5 mm long 7. K. buchananii
Perianth 7-11 mm long 8 . K. breviflora
Perianth 11-19 mm long:
f Inflorescence dense (occasionally medium-dense in K. breviflora):
Inflorescence white, cream or yellow (buds may be tinged with red):
Perianth not exceeding 15 mm in length:
Bracts ovate-lanceolate to linear-lanceolate, acute to acuminate:
Leaves 3-6 mm broad:
Perianth 7-11 mm long 8. K. breviflora
Perianth more than 11 mm long:
Leaf margin scaberulous to finely denticulate, often smooth towards the base;
perianth 12-15 mm long 12. K. fibrosa
Leaf margin conspicuously serrulate; perianth 15-18 mm long 13. K. crassifolia
Leaves 6-12 mm broad; perianth 11-15 mm long 9. K. albescens
Bracts ovate to ovate-oblong, obtuse to rounded 14. K. gracilis
Perianth exceeding 15 mm in length:
Bracts lanceolate to linear-lanceolate, acute to acuminate:
Leaves narrow, grasslike, 3-6 mm broad; plants of small stature with inflore-
scences 5-8 cm long 13. AT. crassifolia
Leaves 10-35 mm broad; robust plants with inflorescences 10-30 cm long
20. K. ensifoiia
Bracts ovate to ovate-oblong, obtuse to rounded:
Inflorescence oblong to elongate, 2 8-3-8 cm in diarn 14. AT. gracilis
Inflorescence globose to subglobose, 5-5-5 cm in diam 43. K. citrina
Inflorescence orange, red or scarlet (buds distinctly suffused with red):
Inflorescence 3-7 cm long (rarely to 10 cm and then concolorous, coral red); plants of
small stature with narrow leaves 2-10 mm broad:
Stamens reaching the throat of the perianth tube at anthesis or slightly exserted;
inflorescence flame-red at the apex grading to yellow at the base. ... 10. K. flamnmla
Stamens not exceeding half the length of the perianth tube at anthesis; inflorescence
concolorous, coral-red 11. K. evansii
Inflorescence elongate, 8-30 cm long; leaves 8-35 mm broad:
Leaves relatively rigid, usually exceeding 60 cm in length; leaf margin usually finely
serrulate; bracts 6-10 mm long:
Inflorescence white (flowers rarely pale yellow), often suffused with red at the apex;
perianth 15-20 mm long 20. K. ensifoiia
Inflorescence yellow to orange-red, deeper red at the apex; perianth 19-25 mm
long 21. K. splendida
Leaves milky-glaucous, flaccid, 30-65 cm long; leaf margin smooth; bracts 11-15 mm
long 23. K. sarmentosa
398
fflnflorescence lax:
Bracts ovate to oblong, obtuse to rounded 14. K. gracilis
Bracts lanceolate to linear-lanceolate, acute to acuminate:
Inflorescence 10-30 cm long; perianth 19-30 mm long, in various colours (including
yellow), not markedly expanded about the middle 17. K. nifa
Inflorescence 3-10 cm long; perianth 14-18 mm long, yellow, distinctly expanded
from about the middle to the throat 18. K. pauciflora
* Perianth exceeding 2 cm long:
Inflorescence lax, particularly in the lower part:
Bracts ovate to ovate-oblong, obtuse:
Perianth up to 21 mm long 14. K. gracilis
Perianth 25-35 mm long 15. AT. laxiflora
Bracts ovate-lanceolate to linear-lanceolate, acute to acuminate:
Perianth exceeding 30 mm in length 16. K. ichopensis
Perianth up to 30 mm long:
Leaves 2-5 (rarely -8) mm broad (Natal) 17. K. rufa
Leaves 8-30 mm broad (S.W. Cape):
Leaves 60-150 cm long, pendulous from cliff faces; pedicels 5-7 mm long 19. K. tabularis
Leaves 30-60 cm long, erect; pedicels 1-3 mm long 23. K. sarmentosa
Inflorescence dense, especially in the lower part (sometimes ± lax towards the apex):
Leaves not keeled or V-shaped in cross-section:
Leaves rigid, erect, U-shaped in cross-section, markedly ribbed, 0 - 8—1 -4 cm broad 28. K. stricta
Leaves spreading-arcuate, broad and relatively flat, 3-12 cm broad; plants often caul-
escent 29. K. northiae
Leaves keeled or V-shaped in cross-section:
iBracts lanceolate to linear-oblong, acute to gradually acuminate at the apex:
Stamens well exserted at anthesis and remaining exserted:
Plants usually caulescent; leaves and bracts usually drying with a purplish tinge;
inhabiting high mountain areas at 6,000 to 10,000 ft 25. K. caulescens
Plants acaulescent; leaves not drying with a purplish tinge; at low to medium elevation,
usually below 6,000 ft:
Pedicels 1-3 (rarely -4) mm long; perianth up to 24 (rarely -25) mm long:
Plants spring-flowering; flowers whitish when open, often red in the bud stage
20a. K. ensifolia subsp. ensifolia
Plants autumn-flowering; flowers whitish to yellow when open, often red in the
bud stage:
Leaves semi-glaucous, 2-3-5 cm broad; flowers whitish to lemon-yellow when
open (eastern O.F.S.) 20b. K. ensifolia subsp. autumnalis
Leaves mid-green, 1-2 cm broad; flowers greenish-yellow to orange-yellow
when open (eastern Transvaal and Swaziland) 21. K. splendida
Pedicels 4-10 mm long; perianth 24-30 mm long (Cape: Uniondale to Komga;
hybrid derivatives often cultivated) 22. K. praecox
Stamens included or shortly exserted at anthesis, usually withdrawn later:
Plants with peduncles 1-2 m tall 22. K. praecox
Plants of small to medium stature rarely exceeding 80 cm tall:
Leaves with margins serrulate for entire length:
Leaves less than 2-5 cm broad:
Leaves 8-20 mm broad; inflorescence dull red at the apex, flowers usually
greenish-yellow:
Leaves glabrous 26. K. ritualis
Leaves pubescent 27. K. Iiirsuta
Leaves up to 6 mm broad; inflorescence concolorous, coral-red 31a. K. triangularis
subsp. triangularis
Leaves more than 2-5 cm broad 29. K. ncrthiae
Leaves with margins completely smooth or with a few scattered teeth:
Perianth 20-25 mm long; inflorescence not concolorous coral-red in colour:
Leaves 30-60 cm long and 8-30 mm broad; peduncle 25-45 cm long; inflore-
scence lax to medium-dense (western Cape) 23. K. sarmentosa
399
Leaves 50-90 cm long, rarely exceeding 10 mm broad; peduncle 50-90 cm
long; inflorescence dense (Transvaal) 24. AT. coralligemma
Perianth 26-50 mm long or, if shorter, then inflorescence concolorous coral-red:
Leaves 2-6 mm broad:
Leaves grass-like, fibrous in texture:
Inflorescence concolorous coral-red in colour.... 31a. K. triangularis subsp.
triangularis
Inflorescence mainly scarlet, often yellowish at the base 30. K. galpinii
Leaves soft-textured:
Buds brownish-red at the apex, flowers white 32. K. thodei
Buds red at the apex, flowers lemon-yellow; leaf margins completely
smooth 33. K. porphyrantha
Leaves 7-25 mm broad:
Leaves soft-textured, rare'y less than 8 mm broad :
Perianth 28-40 (rarely -42) mm long:
Inflorescence concolorous, coral-red.. 31b. K. triangularis subsp. obtusiloba
Inflorescence bicolorous with a few red buds at apex, remainder of
inflorescence lemon-yellow 33. K. porphyrantha
Perianth 42-50 mm long 34. K. fluviatilis
Leaves fibrous in texture, up to 8 mm broad; inflorescence scarlet in the
upper half, yellow below 30. K. galpinii
Bracts ovate to oblong or, rarely, lanceolate; apex rounded to acute:
Flowering during the period July to November:
Pedicels 2-5-6 mm long:
Leaves more than 3 cm broad 29. K. northiae
Leaves usually less than 3 cm broad:
Leaves 1-5-3 cm broad, yellow-green, usually recurved-falcate with serrulate
margins; inflorescence short, subglobose-capitate (Pondoland) 37. K. drepano-
phylla
Leaves 0-6-1 -5 cm broad, erect, dull to glaucous-green, erect with margins smooth
or nearly so; inflorescence subglobose to oblong (W. Cape).... 42. K. uvaria
Pedicels 1-2 mm long:
Leaf margins smooth or with a few scattered teeth:
Inflorescence more than 6 cm in diameter, globose; plants usually robust with
peduncles 60-140 cm long, leaves 50-120 cm long and 15^40 mm broad
(depauperate plants apt to be confused with K. littoralis. but then bracts
not white and fruits small, subglobose) 41. K. rooperi
Inflorescence 4-6 cm in diameter, globose to oblong; plants of small to medium
stature with peduncles 25-65 cm long, leaves 15-70 cm long and 8-25 mm
broad :
Buds flame-red. flowers yellow; inflorescence globose; leaves relatively narrow,
erect 36. K. elegans
Buds dull red, flowers greenish-yellow; inflorescence subglobose to oblong;
leaves medium-broad, erect or falcate:
Bracts entire, white; leaves dark green, not glaucous; fruits large, ovoid,
8-10 mm long 35. K. littoralis
Bracts eroso-denticulate; leaves glaucous to semi-glaucous: fruits subglobose,
5-8 mm long 38. K. baurii
Leaf margins distinctly serrulate:
Leaves erect:
Leaves glaucous to dull green (eastern Cape and Natai):
Leaves up to 2-5 cm broad 38. A", baurii
Leaves usually more than 2-5 cm broad .39. AT. latifolia
Leaves yellow-green (mainly eastern Transvaal) 40. A-, rigidifolia
Leaves arcuate-spreading to falcate:
Inflorescence usually less than 6 cm in diameter, globose to oblong; plants
of small to medium stature with peduncles up to 60 cm long and leaves
up to 50 cm long 38. K. baurii
400
Inflorescence usually more than 6 cm in diameter, globose; plants usually
robust with peduncles 60-140 cm long and leaves 50-120 cm long 41. K. rooperi
Flowering during the period December to June:
Pedicels 3-6 mm long:
Leaves more than 2-5 cm broad 29. K. northiae
Leaves less than 2-5 cm broad 42. K. uvaria
Pedicels less than 3 mm long:
Inflorescence globose or subglobose:
Perianth 19-28 mm long; plants of small stature with small, globose inflorescences
and well-exserted stamens 43. K. citrina
Perianth exceeding 28 mm; plants of medium to robust stature with stamens
scarcely exserted:
Leaves distinctly tough and fibrous, usually glaucous, 6-15 (rarely -20) mm
broad; eastern Cape, inland, usually rocky situations:
Leaf margin distinctly serrulate 38. K. batirii
Leaf margin smooth or with a few teeth towards the apex 42. AT. uvaria
Leaves not markedly tough, usually not distinctly glaucous, 15-40 mm broad:
Leaves arcuate-spreading; mainly coastal, eastern Cape to southern Natal,
usually in marshy situations 41. K. rooperi
Leaves erect; inland, eastern Cape to Transvaal 44. K. liitearifolia
Inflorescence oblong, rhomboid or subcylindrical:
Leaves usually arcuate-spreading, tough and fibrous, glaucous; plants of medium
stature not exceeding 80 cm tall; inflorescence shortly oblong to subglobose:
Leaf margin distinctly serrulate 38. K. baurii
Leaf margin smooth or with a few teeth towards the apex 42. K. uvaria
Leaves at first erect, later reflexed (kneed) about the middle, not arcuate, more
or less flaccid; plants of robust stature often exceeding 90 cm tall:
Perianth exceeding 28 mm in length (rarely less); stamens scarcely exserted;
inflorescence large, oblong to rhomboid 44. K. linearifolia
Perianth 20-28 (rarely -30) mm long; stamens usually markedly exserted;
inflorescence oblong to subcylindrical 45. K. tysonii
1. Rniphofia multiflora Wood & Evans in J. Bot. Lond. 35: 353 (1897); Wood
in Gard. Chron. 29: 56 (1901); Natal PI. 3: t.206 (1900); Hook. f. in Bot. Mag.
t.7832 (1902); Berger in Pflanzenr. 4, 38: 44, fig. 154 H-J (1908); Codd in Flow.
PI. Afr. 37: 1. 1 445 (1965). Type: Natal, between Van Reenen and Nelson’s Kop,
Medley Wood 5972 (NH!, holo.; K!).
Plants robust, usually in groups. Leaves broad, linear, more or less in four ranks,
8-12 per peduncle, at first ascending, later reflexed, mid-green, 80-180 cm long and
2-4 cm broad, channelled above and deeply keeled; margin and keel serrulate with
minute forward-pointing teeth. Peduncle over-topping the reflexed leaves, 80-160 cm
long (excluding the inflorescence), usually with a few sterile deltoid bracts below the
inflorescence. Inflorescence narrowly cylindrical, elongate, tapering gradually towards
the apex, lax to medium-dense at the base, denser towards the apex, 30-80 cm long
and 2 -5-4-0 cm in diameter; buds and flowers ascending; buds greenish-white to
orange-yellow tipped with red, becoming whitish to yellow as the flowers open. Bracts
oblong-spathulate, 3 -5-5 -5 mm long, 1 -5-2 mm broad near the base, acute to rounded
at the apex, scariose with a median brown nerve, overlapping the buds when young,
soon withering; margin minutely eroso-denticulate. Pedicels 1-5-4 mm long at
flowering stage, elongating to 4-5 mm in fruit. Perianth shortly subcylindrical, slightly
constricted above the ovary, 7-12 mm long and 2-5 mm wide at the base, increasing
gradually to 3 mm wide near the apex; lobes ovate, 1-5 mm long, rounded, not
spreading. Stamens 6, of different lengths, exserted at anthesis by 4-8 mm; anthers
broadly oblong, 1 -5 mm long. Style subequal to the anthers at anthesis, finally exserted
by 10-12 mm. Fruit ovoid-triquetrous, 5-6 mm long. Figs. 13, 14. Plate 1. Map 1.
Plate 1. — Kniphofia multiflora Wood & Evans
401
Distributed along the Drakensberg escarpment from Van Reenen and Harrismith
in the south to northern Natal, Swaziland and eastern Transvaal, reaching its northern-
most limit in the Soutpansberg. It grows in mountain marshes, vleis and stream-
banks, usually at altitudes of 4,000 to 6,000 ft. but, in the eastern Transvaal, it is found
in vleis at altitudes as low as 2,500 ft. Flowering occurs from February to April.
Fig. 13. K. multiflora, on Nelshoogte Forestry Station (Bruce 313).
402
Fig. 14. — K. multiflora, inflorescence, X | ( Bruce 313).
O.F.S.— Harrismith: Kerkenberg, Jacobsz 299.
Natal.— Klip River: near Van Reenen, Medley Wood 5972 (K, NH); 6684. Utrecht: Altemooi,
Thode s.n. (NH, STE); 3? miles S.E. of Groenvlei, Codd 6952. Vryheid: Kambula Mission Station,
Gerstner 4326; 17 miles N. of Vryheid, Bruce 292; Pevaan Store, Shirley s.n. (NU).
Swaziland. — Goedgegun, Net s.n.
Transvaal. — Barberton: Nelsberg, 28 miles W. of Barberton, Bruce 313; Codd 10334; Looke
s.n.; Van der Merwe 1671; Van Balen s.n.; Botha's Nek, near Nelshoogte Forestry Station, Meeuse
10089. Carolina: 6 miles N. of Carolina, Adelaar s.n.; 7 miles E. of Zevenfontein, Codd 10340.
Letaba: near Tzaneen, Pigeon Hole Farm, McCallum in PRE 12466; near Duiwelskloof, Scheepers
s.n. Nelspruit: Townlands, Unkles s.n.; Mtimba, Cholmondeley s.n.; Crocodilepoort Mt., 24 miles
403
E. of Nelspruit, Code l 7771. Pietersburg: Haenertsburg, Thompson s.n.; 2 miles S. of Houtbosch,
Bruce 410. Pilgrims Rest : 3 miles W. of Acornhoek, Codd 7851 ; 10 miles E. of Graskop, Codd 10051.
Soutpansberg: near Hangklip, Meeuse 10152. Wakkerstroom : Thode s.n.; Oshoek, Devenish 608.
Readily distinguished from other South African species by the exceptionally
long, narrow infloresence, 30-80 cm in length, and its tall stature, frequently reaching
a height of 2 m or more. It exhibits two colour forms, which are associated with
differences in distribution. In the south, from Van Reenen to Wakkerstroom, the
typical form occurs with greenish-white buds and whitish to cream flowers. From the
Carolina and Barberton Districts, northwards, the buds are orange-yellow, sometimes
tinged with red, and the flowers on opening are pale to deep yellow. Otherwise the
species is relatively constant throughout its range and therefore separate taxonomic
status for the colour variant is not considered justified.
The flowers, unlike other species of Kniphofia in South Africa, do not become
pendulous with age, but remain ascending. Its relationship to certain tropical species
previously placed in a separate genus, Notosceptrum, is discussed on p. 394.
2. Kniphofia parviflora Kunth, Enum. PI. 4: 553 (1843); Bak. in J. Linn. Soc.
11: 361 (1871); J. Bot. Lond. 23: 277 (1885); FI. Cap. 6: 277 (1896); Berger in
Pflanzenr. 4, 38: 40 (1908), excluding var. albiflora Berger. Type: Pondoland, hills
between Umzimvubu and Umsikaba Rivers, Drege 4528 (K!, lecto.; G!).
K. modesta Bak. in J. Bot. Lond. 27: 43 (1889); FI. Cap. 6: 277 (1896), partly,
as to type, non Bot. Mag. t .7293 . Type: East Griqualand, near Kokstad, upper slopes
of Mt. Currie, Feb. 1884, Tyson 1418 (K!, holo.; PRE!, BOL!, GRA!, G!). K. krookii
A. Zahlbr. in Ann. Hof-Mus. Wien 15: 14 (1900). Type: East Griqualand, Mt.
Insizwa, Krook sub Penther 547 (W, holo.f; BOL!).
Plants usually solitary with a number of spreading, fleshy, cylindrical roots.
Leaves 5-8 per peduncle, narrowly linear, at first ascending and later recurving, 20-75
cm long, 3-7 mm broad, keeled, nerves fairly distinct, breaking into numerous persistent
fibres at the base; margin smooth or, rarely, sparingly serrulate; outer two leaves
usually shorter and broader, clasping the base of the plant. Peduncle overtopping
the recurved leaves, 25-80 cm long, with a few to several sterile bracts below the
inflorescence. Inflorescence elongate oblong, secund, fairly lax, 6-28 cm long, 1 -2-1 -7
cm broad; buds ascending, flowers at first spreading and later pendulous; buds
greenish-brown to maroon, flowers creamy-yellow, yellow, yellow-green to greenish,
rarely suffused with dull red. Bracts ovate, usually broadest about the middle then
narrowing abruptly to an acute to acuminate apex, 4-6-5 mm long, 2-2-5 mm broad,
concave, pale brown with a dark brown central region; margin entire. Pedicels 1-2
mm long at flowering stage, elongating to 2-5-3 mm in fruit. Perianth cylindrical,
sometimes slightly constricted above the ovary, 9-10 mm long, 2 mm broad at the
base, scarcely increasing towards the apex; lobes ovate, 1 mm long, not spreading.
Stamens of two lengths, the longest exserted by 1-2 mm at anthesis and later withdrawn
into the perianth tube. Style equalling the anthers at anthesis and eventually exserted
by 2-3-5 mm. Fruit subglobose, 5-6 mm long. Figs. 15, 16. Map 1.
Distributed from King William’s Town District, through the Transkei, Pondoland,
Tembuland and East Griqualand to the Polela District of southern Natal. Found
in dense sourveld, usually on hillsides or, occasionally, in marshy places, from near
sea level to 6,000 ft. altitude. The main flowering time is January to March, but
occasional specimens are recorded in flower in August, September, November and
December.
Cape. — Without locality: Mrs. Barber (GRA). Bizana: 2 miles W. of Bizana, Acocks 13363.
Elliot: 15f miles N.E. of Engcobo, Marais 1382. Engcobo: near Engcobo, Reynolds 2503 ; Engcobo
Mt., Bolus 10335 (BOL); Flanagan 2712; 3f miles from Engcobo on road to Cala, Bruce 559;
404
miles from Engcobo on road to Cala, Bruce 560; 4 miles from Engcobo on road to Elliot, Marais
1388. Keiskammahoek : Cata Forestry Station, Bruce 550; Hogsback, N. R. Smuts s.n. Kentani:
Sept. 1902, Pegler 851; 1906, Pegler 851 (GRA); Nov. 1911, Pegler 851. King William’s Town:
slopes of Mt. Kemp, B. H. Dodd sub Galpin 8032; near Evelyn Valley Forestry Station, Bruce 547:
Komga: near Kei Mouth, Flanagan 1805. Lusikisiki: between Umzimbuvu and Umsikaba Rivers,
Drege 4528 (G, K); near Lusikisiki, Bruce 597; near Magwa Falls, Codd 9317. Maclear: Glen Airy,
Galpin 6864; Ugie, Britten 4582; 6 miles S.W. of Ugie, Acocks 21984; near Naudes Nek, Bruce 577.
Mount Ayliff: Mt. Insizwa, Krook sub Penther 547 (BOL); Schlechter 6461 (Z). Mount Currie:
near Kokstad, upper slopes of Mt. Currie, Tyson 1418; 6 miles S. of Kokstad Bruce 588. Stutterheim:
summit of mountain above Toise River Railway Station, Flanagan 2261. Umtata: Baur 617 (K).
Umzimkulu: Ensikeni, Haygarth sub Wood 12068 (NH). Xalanga: near Cala, McNeil s.n.
Natal. — Polela: 10 miles from Bulwer on road to Elimeville, Marais 1446.
Fig. 15.— K. parviflora, from Naudes Nek Fig. 16. — K. parviflora , inflorescence, life size
(Bruce 577). ( Bruce 577).
K. parviflora is unique in the genus in having a secund inflorescence. In the bud
stage the inflorescence is narrowly cylindrical but, as the flowers open, the inflorescence
usually leans to one side and the flowers turn in that direction. Considering the
relatively wide distribution, variation is comparatively slight and is noted chiefly in
stature and flower colour. According to collectors’ notes, the latter varies from
creamy-yellow or whitish-green to yellow or yellow-green. The flower colour of dried
specimens is usually yellow-brown to greenish-brown, and the bracts are pale brown
with a dark brown central region. Berger, l.c., mentions that the collector Baur records
the flowers as being scented, which suggests a relationship with K. typhoides and K.
Plate 2 a.—Kniphofia rufa Bak
Plate 2b. — Kniphofia acraea Codd
405
umbrina. In the early flowering stage K. parviflora may sometimes be confused with
K. breviflora, but can usually be distinguished by the longer inflorescence, darker
flower colour and broader more ovate bracts.
K. modesta Bak. (1889) is based on Tyson 1418, which is conspecific with K.
parviflora Kunth. Subsequently Baker misinterpreted his species because the plant
figured in Bot. Mag. t.7293 (1893) as K. modesta is not that species, being clearly distinct
from Tyson 1418. It is mainly in this wrong sense, also, that Baker treats K. modesta
in Flora Capensis (1896). This erroneous interpretation was followed by subsequent
authors, such as Mallett in the Gardener’s Chronicle (1906) and Berger (1908). Berger
correctly cites Tyson 1418 (type of K. modesta Bak.) under K. parviflora Kunth, but he
then upholds K. modesta in the sense of the Bot. Mag. t.7293 as a distinct species,
whereas the correct procedure should have been to give the plant illustrated in the
Bot. Mag. plate a new name. N. E. Brown made the necessary correction when he
renamed the Bot. Mag. plate K. sparsa N. E. Br. The latter is now placed as a synonym
of K. gracilis, though the possibility is not excluded that further study may prove that
it is worthy of separate status of some sort.
3. Kniphofla acraea Codd in Bothalia 9: 142 (1966). Type: Cape, Cradock
District, National Mountain Zebra Park, on top of Bankberg, alt. about 6,800 ft.,
flowering 6.3.1964, Liebenberg 7120 (PRE!, holo.).
Plants in small groups. Leaves about 12 per peduncle, 50-70 cm long and 1-1 -5
cm broad, V-shaped in cross section, tending to fold along the midrib, nerves con-
spicuous in dried specimens; margin minutely and somewhat distantly denticulate,
keel smooth. Peduncle subequal to the leaves, 50-60 cm long. Inflorescence dense,
ovoid to subcylindrical, tapering to the apex in the young stage, 6-12 cm long and
2 -2-2 -5 cm in diameter; buds greenish-yellow, spreading; mature flowers yellow,
at first spreading, later deflexed. Bracts broadly ovate, obtuse, 4-5 mm long and
2-2-5 mm broad, chartaceous, at first erect and overlapping the buds in the young stage,
Map 1. — Distribution of Kniphofia multi flora,
K. parviflora and K. acraea.
Map 2. — Distribution of Kniphofia umbrina,
K. typhoides and K. brachystachya.
406
later reflexed; margin minutely eroso-denticulate. Pedicels 1 • 5-2 mm long. Perianth
shortly cylindrical, not constricted above the ovary, 8-9 mm long, about 2 mm wide
at the base, widening to 2-5-3 mm at the mouth; lobes broadly ovate to oblong,
rounded to acute, 1 • 5-2 mm long. Stamens exserted by 2-3 mm at anthesis. Style
equal to the stamens at antheris, eventually exserted by 3-4 mm. Fruit not seen.
Plate 2b. Map 1.
Known from only one gathering (3 specimens) from the top of the Bankberg in
the Cradock District, where it was found growing in small seepage areas among rocks.
It was collected in flower in March 1964, and has been successfully cultivated in
Pretoria.
Cape. — Cradock: National Mountain Zebra Park, on top of Bankberg, Liebenberg 7120; from
same locality, cultivated in Pretoria, Liebenberg 7120a.
A distinctive species, resembling K. breviflora Harv. ex Bak. in the length and
colour of the flowers but from which it may be distinguished by the ovate, obtuse
bracts and the denser inflorescence. Its true relationship is not altogether clear, but
probably is closer to K. brachystachya (A. Zahlbr.) Codd than to any other species.
4. Kniphofia umbrina Codd in Bothalia 9: 141 (1966). Type: Swaziland, 4 miles
south of Forbes Reef, flowering 20.2.1951, Bruce 272 (PRE!, holo.).
Plants usually in small groups. Leaves not in clear ranks, 6-8 per peduncle, at
first erect, later reflexed, mid-green, somewhat flaccid, 45-70 cm long and 1-2 cm
broad, tapering towards a rather blunt apex, V-shaped in cross section, breaking up
into a few persistent fibres at the base; margin and keel smooth. Peduncle overtopping
the leaves, 70-90 cm tall, with several sterile bracts below the inflorescence. Inflorescence
subcylindrical, tapering slightly towards the apex, very dense, 7-15 cm long and 1-5-2
cm in diameter; buds spreading, flowers eventually pendulous; buds purplish-brown,
flowers reddish-brown to burnt umber, drying blackish, slightly scented. Bracts linear-
lanceolate, acuminate, 9-11 mm long and 1-5 mm broad; margin entire. Pedicels
0-5 mm long. Perianth cylindrical, not constricted above the ovary, 8-9 mm long,
1 -5 mm wide at the base, increasing to 2-5 mm at the mouth; lobes ovate, 1 mm long,
not spreading. Stamens of two lengths, exserted by up to 2 mm at anthesis. Style
exserted by 1-2 mm at anthesis and eventually up to 3 mm. Fruit ovoid-triquetrous,
5-6 mm long. Fig. 17.
The species is known only from a small area a few miles south of Forbes Reef,
in Swaziland, at an altitude of about 4,500 ft., where it grows in or near grassy vleis.
The flowering time is February to early March.
Swaziland. — 4 miles S. of Forbes Reef, Bruce 272; Codd 9521 ; 12 miles N. of Mbabane on Forbes
Reef road, Reynolds 5866; near Forbes Reef, Compton 27552; 31290.
With its brown, slightly scented flowers, K. umbrina shows an obvious relationship
to K. typhoides Codd, but differs in several important characters: the leaves are not
distichous or glaucous, as in K. typhoides , and they are V-shaped in cross section;
the perianth is longer and more pendulous; and the floral bracts are longer and
acuminate, not rounded.
The species was first collected in 1951: on 19.2.1951, by Dr. G. W. Reynolds
and, independantly, on 20.2.1951 by Miss E. A. Bruce, accompanied by the author,
who collected it again in March, 1956. All the above gatherings are from one small
locality adjoining the road about 4 miles south of Forbes Reef. Further information
is desired as to the possibility of its occurring at other localities as well because, with
the growing pressure of agricultural activity, steps should be taken to preserve it in
its natural state if it is indeed as rare as would appear.
407
Fig. 17. — K. umbrina, 4 miles S. of Forbes Reef ( Bruce 272).
5. Kniphofia typhoides Codd in Flow. PI. Afr. 36: 1. 1 424 (1964). Type: Natal,
“Klip River County”, Mrs. K. Saunders in Herb. Medley Wood 3895 (K!, holo.;
BM!, NH!).
Notosceptrum nataJense Bak. in FI. Cap. 6: 285 (1896); in Hook, f., Ic. PI. t.2523
(1897); Berger in Pflanzenr. 4, 38: 71 (1908); non Kniphofia natalensis Bak.
Plants solitary or two or three stems together from a short thickened rhizome
from which a number of cylindrical fleshy roots arise. Leaves mainly distichous, 6-8
per peduncle, glaucous, erect, eventually bending over at the apex, shorter than the
peduncle, 35-65 cm long, 0-8-2 -5 cm broad, with a distinct midrib, shallowly chan-
24379—4
408
nelled above, slightly keeled, tending to be spirally twisted, firm and semi-fleshy in
texture, not conspicuously nerved but outer leaves breaking into persistent fibres at
the base; margin smooth. Peduncle 50-75 cm long with several sterile bracts below
the inflorescence. Inflorescence cylindrical, very dense, 15-30 cm long and 1 -5-2-5 cm
in diameter, rounded at the apex; buds erect, dark brown; flowers erect or spreading,
brown to purplish-brown, faintly scented. Bracts obovate or subrotund to ovate,
rounded at the apex, concave, 4-6 mm long, 3 mm broad, scariose with a central brown
nerve; margin entire to minutely eroso-denticulate. Pedicels 0. Perianth subcam-
panulate or shortly cylindrical, not constricted above the ovary, 4. 5-6 -5 mm long,
2-5-3 mm broad; lobes ovate to rounded, 1-5 mm long, not spreading. Stamens
of two lengths, exserted by 3-4 mm at anthesis; anthers brown. Style exserted by
1-2 mm at anthesis and finally by up to 3 mm. Fruit ovoid, more or less triquetrous,
5-6 mm long. Fig. 18. Plate 3. Map 2.
Fig. 18. — K. typhoides, Heidelberg District, Transvaal ( Codd 8509).
Plate 3. — Kniphofia typhoides Codd
409
Recorded from the northern districts of Natal, southern, central and western
Transvaal and the north-eastern part of the Free State, at medium altitudes of 4,000
to 5,500 ft. It is almost invariably found on black clay soil and shows a preference
for low-lying pans or vleis. The type locality is given by Baker as Klip River County,
but on the sheet in the Natal Herbarium the locality is given as “ near Newcastle
The main flowering season is from mid-February to near the end of March.
Natal. — Newcastle: near Newcastle, Mrs. Saunders sub Wood 3895 (BM, K, NH); Utrecht: farm
Klipspruit, Sprengel s.n. Vryheid: 13f miles N. of Vryheid, Bruce 293.
O.F.S. — Vredefort: between Leeuwspruit and Vredefort, Barrett-Hamilton s.n. (BM); near Parys,
Hutchinson 2979; between Parys and Greenland Station, Wagener sub NBG 264/32 (BOL).
Transvaal. — Bethal: H miles W. of Bethal, Bruce 259; 9 miles E. of Bethal, Bruce 260; 1 mile
E. of Kinross, Codd 7693. Brakpan: near Witpoort, Murray 202. Heidelberg: 10 miles N. of Villiers,
Bruce 339; 8 miles N. of Villiers, Codd 8509. Lydenburg: Van der Merwe s.n.; between Lydenburg
and Orighstad, Strey 3758; 10 miles N. of Lydenburg, Dyer 3953; 14 miles N. of Lydenburg, Bruce
322. Potchefstroom: De Wet 1230; Haaskraal, Lomv 1689. Rustenburg: between Swartruggens
and Groot Marico, Gilliland in Hb. Moss 27050. Springs: vlei near Delmas, Pole Evans s.n.; Codd
2683; Bruce 258. Vereeniging: Burtt Davy s.n. (K).
K. typhoides is characterised by the erect, glaucous leaves arranged mainly in a
distichous manner, and the dense inflorescence of small brown flowers, carried well
above the leaves and bearing some resemblance to a bullrush ( Typha sp.), from which
the species name is derived. This name was given when the species Notosceptrum
natalense was transferred to Kniphofia, because the name K. natalensis has already
been used for a different species. The reasons for the transfer are discussed on p. 392.
Notes on the relationship of K. typhoides are given in the discussion on K. umbrina
(p. 406) and K. brachystachya (p. 411). These three species form a closely related
group with relatively small, closely placed, brown or yellow-brown, slightly scented
flowers. They may, however, readily be separated from one another on perianth
length and bract shape.
6. Kniphofia brachystachya (A. Zahlbr.) Codd in Flow. PI. Afr. 36: sub. 1. 1 424
(1964). Type: Griqualand East, Mt. Insizwa, 26.1.1895, Krook sub Penther 536
(W, holo.f); neotype: Cape Province, Umzimkulu District, Ensikeni, December,
1912, Haygarth sub Wood 12062 ( PRE !).
Notosceptrum brachystachyum A. Zahlbr. in Ann. Hof-Mus. Wien 15: 15, t. 3 (1900);
Berger in Pflanzenr. 4, 38: 71 (1908).
Plants solitary or in small clusters. Leaves 6-8 per peduncle, not noticeably
distichous, linear-lanceolate, acute, erect or somewhat falcate, rigid, 15-28 cm long,
4-12 mm broad, shallowly keeled, breaking up into persistent fibres at the base;
margin thickened, smooth to minutely denticulate; bases of outer leaves broader and
clasping base of plant. Peduncle overtopping the leaves, 25-65 cm long, with several
sterile bracts below the inflorescence. Inflorescence cylindrical, dense, 4-9 cm long,
1-1-5 cm in diameter; buds erect and flowers erect to spreading; buds brownish,
flowers dull yellow to brownish, drying purplish-brown. Bracts ovate to ovate-deltoid,
acute, 4-5 mm long and 2 mm broad at the base, brown to purplish-brown; margin
eroso-denticulate. Pedicels 0. Perianth campanulate to subturbinate, not constricted
above the ovary, 4-5 mm long, 1-5-2 mm broad at the base, increasing gradually
to 2-5 mm at the apex; lobes ovate, 1-1 -5 mm long, not spreading. Stamens subequal,
the longest exserted by 2 mm at anthesis. Style subequal to the anthers at anthesis
and finally exserted by 3 mm. Fruit globose, mature fruits not seen. Fig. 19. Map 2.
Recorded from the mountains of East Griqualand and the foothills of the Drakens-
berg in the midlands of Natal, where it grows at altitudes of 4,000 to 6,500 ft. in
mountain grassland. The flowering time is from November to January.
410
Fig. 19. — K. brachystachya, from Ann. Naturhist. Hofrnus. Wien 15: t.3 (1900).
411
Cape. — Mount AylifF: Insizwa Mt., Schlechter 6462 (BOL, Z). Umzimkulu: Ensikeni, December
1912, Haygarth sub Wood 12062.
Natal. — Estcourt: Little Berg above Champagne Castle Hotel, Killick 1867; Giants Castle Game
Reserve, Trauseld 582; Highmoor Forest Station, Killick & Vahrmeijer 3630. Lions River: Nottingham
Road, Medley Wood 5238 (NH); near Lidgetton, Medley Wood 6391 (K, NH); Shawlands, near
Nottingham Road, December 1941, Coetzee 4 (NH).
A species of small stature with short, dense inflorescences of very short flowers,
4-5 mm long. The only other South African species with such short flowers is K.
buchananii Bak. Collectors report on the flower colour in the living state as “ dull
yellow ”, “ brown and yellow ”, “ brown and white ” or “ buds yellow-green, flowers
brown ”, but in dried specimens the buds and flowers are uniformly dark purplish-
brown, as in K. typhoides. It may be separated from the latter species by the smaller
stature, the shorter, narrower leaves, the usually shorter perianth and the bracts, which
dry a purplish-brown colour and are ovate, acute. In K. typhoides the bracts dry a
pale brown colour and are rounded at the apex. K. buchananii, on the other hand,
has creamy-white flowers and lanceolate, acuminate bracts.
There is a specimen in PRE, Nicholson s.n., reputed to come from Karkloof,
Natal, which is somewhat intermediate between K. brachystachya and K. buchananii.
The bracts are ovate, acute, with lacinate margins, and are pale brown in colour. The
flowers have dried a brownish colour. It is not possible to place this specimen with
certainty until more material is forthcoming. It is considered to be a form of K.
brachystachya.
The type of K. brachystachya, Krook sub Penther 536 in W, appears to have been
destroyed during World War II and no further material of this gathering has been
traced. The specimen in PRE of Haygarth sub Wood 12062 from Ensikeni has been
nominated as the neotype.
7. kniphofia buchananii Bak. in J. Bot. Lond. 23: 276(1885); FI. Cap. 6: 276(1896);
Berger in Pflanzenr. 4, 38: 39 (1908), partly, excl. Cooper 1029, 3294; Engl. & Drude,
Yeg. Erde 9, 2: 313 (1908). Type: Natal, Rev. J. Buchanan (K!, holo.).
K. buchananii var. flavescens Berger, l.c. 40 (1908), partly, as to Wood 1972.
Plants usually solitary with a number of spreading, fleshy, cylindrical roots.
Leaves about 8 per peduncle, narrowly linear and grass-like, at first ascending and later
recurving, 40-60 cm long, 2-5-4 mm broad, keeled, triangular in cross section,
mid-green, strongly nerved, breaking into numerous persistent fibres at the base;
margin smooth; outer two leaves often shorter and broader, clasping the base of
the plant. Peduncle overtopping the leaves, 40-85 cm long, with several sterile
bracts below the inflorescence. Inflorescence oblong to cylindrical, dense at the apex,
often laxer below, 3-5-9 cm long and 1-1-4 cm in diameter, rounded at the apex;
buds erect, flowers erect to spreading, sometimes eventually deflexed; buds greenish-
cream or tinged with red, flowers white. Bracts ovate-lanceolate to lanceolate, acute
to gradually acuminate, 4-5-5 mm long, 1-1 -5 mm broad; margin entire or minutely
eroso-denticulate. Pedicels 0-5-1 mm long at flowering stage, elongating to 2 mm
in fruit. Perianth subcampanulate to oblong, not constricted above the ovary, 4-5-5
mm long, 1-5 mm broad at the base increasing gradually to 2 mm at the apex; lobes
broadly ovate, 1 mm long, not spreading. Stamens subequal, the longest exserted
at anthesis by 1-1 -5 mm. Style equalling the stamens at anthesis, eventually exserted
by 2-5-3 mm. Fruit globose to ovoid, 2-5 mm long. Fig. 20. Plate 4a. Map 3.
Found in the Natal Midlands and semi-coastal grassland, mainly at medium
altitudes from about 1,000 to 4,000 ft., on grassy slopes, often among dolerite rocks.
The main flowering period is February-March with an occasional record as early as
October and as late as April.
412
Fig. 20. — K. buchananii, Pinetown District, Natal {Lansdell s.n,).
Natal. — District not known: Liddesdale, Medley Wood 3753 (NH); Medley Wood 4798 (NH);
near Boston, Medley Wood 4572 (K, NH); Mowbray, Handley 87 (NH). Camperdown: Drummond,
Van Nouhuys s.n.; between Alverstone and Drummond, Forbes 397 (NH); Alverstone, Strey 7374.
Helpmekaar: Reynolds sub NBG 418/41 (robust plants with flowers 6-7 mm long). Lions River:
Howick, Junod 202 (Z); Shafton, near Howick, Evans s.n. ; Cedara, Wylie sub Wood 9977 ; Lidgetton,
Medley Wood 7431; near Balgowan, Dyer 4867; Caversham, Mogg 2455; Tweedie, Mogg 6819;
10 miles S.E. of Nottingham Road, Acocks 13879; Merrivale, Moll 1339; Karkloof, Moll 3457.
Nkandla: Wylie sub Wood 8958 (NH); 9 miles S. of Nkandla, Codd 6986. Pietermaritzburg: near
Sweetwaters, C. A. Smith 8890. Pinetown: Forest Hill above Kranskloof, Lansdell s.n. ; Fields Hill
Medley Wood 1972 (BM, K, NH, Z). Richmond: Schleehter 6723; Hele Hele, Gerstner 7025; 71-
miles N.E. of Richmond, Acocks 22014. Umvoti: Greytown, Galpin 14835; Van der Merwe 2495
Umzinto: Ellismere, Rudatis 1469 (BM, G, K, Z).
In length of perianth K. buchananii resembles K. brachystachya, but may be
distinguished by its taller stature, its creamy-white flowers and the bracts which are
straw-coloured, ovate-lanceolate to lanceolate and usually acuminate.
Its nearest relative is K. breviflora Harv. ex Bak., from which it is distinguished
mainly on perianth length, which is usually 4-5-5 (rarely -7) mm in K. buchananii
and 7-11 mm long in K. breviflora. There are one or two specimens which are inter-
mediate and difficult to place with certainty, but the great majority of specimens present
no difficulty. K. breviflora occurs more inland and at higher elevations than K.
buchananii. Thus K. parviflora var. albiflora Berger, based on 1 Tow/ 4826 from Glencoe
with flowers 6-7 mm long, is placed in synonymy under K. breviflora.
413
There is an element of confusion in the type material of K. breviflora which is
discussed more fully under that species. K. breviflora is based on Cooper 1029, of which
two sheets are in Kew Herbarium. One sheet has a specimen with an inflorescence
devoid of flowers but with three loose flowers, measuring about 5 mm long, in a capsule
(these, of course, may or may not belong to the specimen); the other sheet has an
inflorescence with a few apical flowers 7-7-5 mm long attached, and is now designated
as the lectotype of K. breviflora. The presence of the short flowers in the capsule no
doubt led Berger to place K. breviflora (1871) as a synonym of K. buchananii (1885),
though he then upheld K. breviflora sensu Bak. in Bot. Mag. t.7570 (1897) as distinct
(see p. 416).
8. Kniphofia breviflora Harv. ex Bak. in J. Linn. Soc. 11: 361 (1871); Bak. in
FI. Cap. 6: 277 (1896); Bot. Mag. t.7570 (1897); Berger in Pflanzenr. 4, 38: 40
(1908). Type: Orange Free State, Drakensberg, Cooper 1029 (there are two specimens
of this gathering in K and the specimen with flowers attached, mounted on the same
sheet as Cooper 3924, is selected as the lectotype).
K. sclilechteri Schinz in Mem. Herb. Boiss. No. 20: 10 (1900). Type: Natal, Van
Reenen’s Pass, Schlechter 6940 (Z !, holo.; PRE!; BOL1, GRA!, K!, G!, BM!).
K. concinna Bak. in Bull. Herb. Boiss. 2 ser., 1: 784 (1901). Type: Natal, near
Newcastle, Rehmann 7024 (Z!, holo.). K. fibrosa sensu Wood, Natal PI. t.322 (1903).
K. buchananii var. flavescens Berger, l.c. 40 (1908), partly, excl. Wood 1972. Type:
Based on K. breviflora Bak. K. parviflora var. albiflora Berger, l.c. 40 (1908). Type:
Natal, near Glencoe, Medley Wood 4826 (K !, holo.; NH!). K. breviflora var. concinna
(Bak.) Berger, l.c. 40 (1908).
Plants usually solitary with a number of spreading, fleshy, cylindrical roots.
Leaves 6-10 per peduncle, narrowly linear, at first erect then bending over about the
middle, grass-like, 40-70 cm long, 2-5-6 mm broad, keeled, distinctly nerved, breaking
into persistent fibres at the base; margin smooth; outer few leaves broad-based,
clasping base of plant. Peduncle overtopping the recurved leaves, 40-80 cm long
with a few sterile bracts below the inflorescence. Inflorescence ovoid to oblong, usually
dense at the apex, medium lax below, 5-8 cm long and 2-2-4 cm in diameter, narrower
towards the apex; buds erect to spreading, flowers ascending to spreading eventually
becoming pendulous; buds yellow-green, often tinged with red, flowers whitish to
yellow. Bracts ovate-lanceolate to lanceolate, acute to gradually acuminate, 4 -5-7 -5
mm long and 1-5-2 mm broad; margin usually entire, occasionally very minutely
denticulate. Pedicels 1-2 mm long at flowering stage and in fruit. Perianth oblong to
shortly subcylindrical, not constricted above the ovary, 7-11 mm long, 1-5-2 mm
broad at the base, widening gradually to 3 mm at the apex, in occasional specimens
broadest above the middle (3-5 mm) and then narrowing again to 3 mm at the apex;
lobes broadly ovate, 2-2-5 mm long, not spreading. Stamens of two lengths, the
longest exserted by 1-1-5 mm at anthesis, later withdrawn into the perianth tube.
Style equalling the longest stamens at anthesis, eventually exserted by 3^4-5 mm.
Fruit globose to ovoid, 4-5-5 mm long. Figs. 21, 22, 23. Map 3.
Recorded from the foothills and slopes of the Drakensberg from Underberg
District to Newcastle and extending into the Free State in the Harrismith District,
occurring in dense mountain grassland on hillsides or in vleis. The flowering season
is mainly from January to March, but it has been collected in flower as early as
October.
Natal. — Cultivated by Max Leichtlin, Baden-Baden, October, 1896, from seed collected in Natal
by Medley Wood, type of Bot. Mag. t.7570 (K). Bergville: Tugela Valley, Hafstrom & Acocks 237;
Royal Natal National Park, Schetpe 1246 (NH); Trauseld 165; Mont-aux-Sources, Bayer & McClean
155; Oliviershoek, Thode 4502 (STE); Vogts s.n.; Code! 10530. Dundee: Glencoe. Medley Wood
4826 (K, NH). Estcourt: Kamberg, Pentz 426; 11 miles from White Mt. Inn on road to Hlatikulu,
414
t
I;
jmwmM mmen.
J >r. A Hi .MM ANN:
Kxsktata Airaw a»sli ex ai;n. l$<i — 1S80,
,v. /. //^./ / ,/,
Z. /
K$WC: tfe
Fig. 21. -Holotype of K. concinna ( Rehmann 7024 in Z), now placed as a synonym of K. breviflora.
415
Killick & Marais 2157; Ntabamhlope, Downing s.n. Klip River: Van Reenen’s Pass, Krook sub
Penther 538 (S); Schlechter 6940; Medley Wood 5606; 8922 (NH); Bews 543; Germishuizen s.n.
Newcastle: near Newcastle, Rehmann 7024 (Z). Underberg: near Underberg, McClean 622 (NH),
O.F.S. — Without locality, Cooper 1029 (K); 3924 (K). Harrismith: near Swinburne, farm Rens-
burgskop, Jacobsz 182; Codd 10522.
Fig. 23. — K. breviflora, inflorescence,
life size ( Killick & Marais 2157).
{ Fig. 22. — K. breviflora , habit (Estcourt
District, Killick & Marais 2157).
It appears that two colour forms are found in this species. The plants seen in
eastern O.F.S. and the neighbouring Oliviershoek Pass had clear yellow flowers, while
those in the Estcourt area had white flowers. Medley Wood describes the flower
colour of his plant No. 8922 (figured in Natal Plants t. 322, 1903, under the name K.
fibrosa and represented by a specimen in the Natal Herbarium) as pale yellow, so it
is possible that there are intermediate colours as well. In the herbarium all flowers
dry a dirty white colour and so it is not possible to form an opinion of flower colour
in the absence of collector’s notes. Cooper, who records the locality of the type
plant as O.F.S., does not mention colour of flowers but the locality indicates that
the typical flower colour for the species is probably yellow.
416
K. breviflora shows affinities with the three species K. buchananii Bak., K. parviflora
Kunth and K. albescens Codd. From the first, K. buchananii, it differs in the longer
flowers, which are usually 9-10 mm long, although the range extends from (6-) 7-11
mm, as against 4-5-5 rarely -7 mm in K. buchananii, and the inflorescence is often
more lax than in K. buchananii. The distributions of the two species do not appear
to overlap, K. breviflora occurring more inland and at higher altitudes than K. buchananii.
The type of K. parviflora var. alb flora Berger, Medley Wood 4826 from near Glencoe,
has flowers 6-7 mm long and is thus somewhat intermediate between the two species.
From the second species, K. parviflora, it differs in the inflorescence not being
secund. In living material, the difference is usually clear cut but, in the herbarium,
it sometimes happens that specimens of K. parviflora are pressed in such a way as to
obscure the secund nature of the inflorescence. There are, however, additional small
differences which assist one to decide on the plant’s identity; for example, in K. parviflora
the flowers tend to dry a greenish-brown colour (pale brown to yellow-brown in K.
breviflora ), the inflorescence is usually longer, the leaves are shorter and the bracts
are usually broader and more abruptly acute. There is also a difference in distribution,
as K. parviflora occurs mainly in the eastern Cape Province, reaching its northernmost
limit in southern Natal.
The third affinity mentioned earlier, namely K. albescens, is closely related to
K. breviflora, but is altogether a more robust plant, usually growing in clusters, with
longer flowers (11-15 mm) and longer bracts (7-9 mm). It has, in general, a more
northerly distribution, but where the two species overlap in nothern Natal it is some-
times difficult to distinguish between them with certainty.
The type material of K. breviflora has been mentioned under K. buchananii (p. 413).
K. breviflora is based on Cooper 1029, of which two specimens, mounted on different
sheets, are in Kew Flerbarium. One of these specimens is mounted on the same sheet
as Cooper 3924 and has an inflorescence with a few apical flowers 7-7*5 mm long
attached. This specimen agrees with Baker’s concept of K. breviflora and is now
selected as the lectotype. The other specimen is alone on a sheet and has an inflorescence
devoid of flowers, but there are three flowers, measuring about 5 mm long, in a capsule.
These flowers resemble those of K. buchananii but, of course, they may or may not
belong to the inflorescence. If the former specimen is accepted as the lectotype, the
two names K. buchananii and K. breviflora can be maintained for the two species
separated by Baker on the basis of perianth length.
Berger, l.c., adopted a course which is in conflict with the present rules of nomen-
clature. He placed K. breviflora Harv. ex Bak. (1871) in synonymy under K. buchananii
Bak. (1885), apparently basing his view on a drawing in Kew Herbarium made by
Harvey of K. breviflora. I have not seen this drawing, so I cannot say if it conflicts
with the concept of Baker, who published the species without citing the drawing.
Berger went on to describe K. buchananii var. flavescens Berger, basing it on “ K.
breviflora Bak.”, and citing Wood 1972 (which is, in my opinion, K. buchananii). Then,
on p. 40, he upheld K. breviflora Bak. in Bot. Mag. t.7570 (1897) as a valid species.
A specimen corresponding to this Bot. Mag. plate has been seen in Kew Herbarium
and, although more robust than Cooper 1029 (lectotype of K. breviflora), is considered
to fall within the range of this species.
Regarding the synonyms K. schlechteri Schinz and K. concinna Bak., the type
specimens have been examined and both have flowers about 8 mm long. They are,
therefore, not distinguishable from K. breviflora Bak. The plant figured in Medley
Wood’s Natal Plants t.322 (1903) as K. fibrosa should also be included in K. breviflora.
The figured plant is Medley Wood 8922 from Van Reenen’s Pass and is represented
in the Natal Herbarium.
Plate 5. — Kniphofia albescens Codd
417
9. Knipholia albescens C odd in Flow. PI. Afr. 34: 1. 1 325 (1960). Type: Natal,
mountains north of Utrecht, Code/ 8248 (PRE!, holo.).
Plants usually in groups. Leaves about 8 to each peduncle, spreading-recurved,
dull green to glaucous, tough and fibrous in texture, 60-100 cm long, 8-15 mm broad,
keeled, breaking up into numerous persistent fibres at the base; margin and keel
smooth or sometimes distantly and minutely scabrid. Peduncle overtopping the
recurved leaves, 40-75 cm long, with several sterile, ovate-lanceolate bracts below the
inflorescence. Inflorescence long-ovoid to subcylindrical, tapering to the base and apex,
dense, often with a coma of bracts at the apex, 6-10 cm long, elongating to 30 cm
long in fruit, 2 -7-3 -5 cm in diameter; buds greenish-white or tinged with pink, erect;
flowers becoming white or cream, spreading at anthesis, later pendulous. Bracts
lanceolate to linear-lanceolate, acuminate, 7-9 mm long and 2 mm broad, chartaceous,
1 -nerved, erect in the bud stage, longer than the buds; margin smooth to minutely
denticulate. Pedicels 1 mm long. Perianth cylindrical, not constricted above the
ovary, 11-15 mm long, 2-5-3 mm broad; lobes ovate, rounded, 2 mm long, slightly
spreading. Stamens exserted by 2-2-5 mm at anthesis; anthers yellow, oblong, 1-1-5
mm long, drying brown. Style equal to the stamens at anthesis, eventually exserted
by 5-6 mm. Fruit ovoid-triquetrous, 7-8 mm long. Fig. 24. Plate 5. Map 3.
Distributed mainly from Utrecht District in Natal to Piet Retief and Ermelo
Districts in southern Transvaal, with one record from Mont-aux-Sources. It occurs
in dense grassveld on mountain slopes and, occasionally, on the flats and in marshy
places, at altitudes from 4,500 to 6,500 ft. The flowering season is from mid-January
to early March.
Natal. — Bergville: Mont-aux-Sources, Pardoe s.n. Utrecht: mountains north of Utrecht and
cultivated in Pretoria, Codd 8248; between Utrecht and Vryheid, Bruce 289; Altemooi, Thode s.n.;
Thode 3333 (STE); Wonderhoogte Farm, Parkhouse s.n.; 4 miles W. of Viljoenspos, Codd 6956;
Fig. 24.— K. albescens, Utrecht District ( Marais 1472). Photo by W. Marais.
418
12^ miles S.E. of Groenvlei, Marais 1472; 4\ miles S.E. of Groenvlei, Codd 6953; 7 miles N.W. of
Groenvlei, Codd 2532; 5 miles S. of Wakkerstroom, Bruce 282; near Wakkerstroom, Reynolds 1135
(BOL).
Transvaal. — Ermelo: near Athole Pasture Research Station, Bruce 266; Spitskop, Scheepers in
TRV 15228; 14 miles S. of Ermelo, Codd 10267; Vlakfontein, Du Plessis 100. Piet Retief: near
Iswepe, Sidey 1564; 4 miles N. of Iswepe, Bruce 275; 4 miles S. of Idalia, Bruce 277; near Moolman
Station, Brace 274. Wakkerstroom; in vlei, Van der Merwe 28; farm Oshoek, Devenish 559; Goed-
gegun, Du Plessis 108; Driehoek, Du Plessis 109; 5 miles S. of Wakkerstroom, Mauve & Tolken 4518.
For differences between K. albescens and the related species K. breviflora, see
p. 416. In perianth length, K. albescens approaches K. fibrosa (p. 422) and K. crassifolia
(p. 424), but these two species are much less robust, with narrow, grasslike leaves
which have distinctly serrulate margins.
10. Kniphofia Hammula Codd in Flow. PI. Afr. 34: 1. 1 326 (1960). Type: Natal,
3+ miles N.W. of Glencoe, Codd 6779 (PRE!, holo.).
Plants growing in groups or occasionally singly. Leaves 6-8 per peduncle, at
first erect then curving over towards the apex or bending at the middle, dull green
to semi-glaucous, 60-90 cm long, 6-12 mm broad, narrowly keeled below, shallowly
channelled above, not prominently nerved, breaking into some persistent fibres at the
base; margin and keel smooth. Peduncle subequal to the leaves, 50-85 cm long with
several sterile bracts below the inflorescence. Inflorescence ovoid to shortly rhomboid,
3-7 cm long and 2-3-5 cm in diameter, dense, often with a coma of sterile bracts at
the apex; buds pendulous, flame scarlet; flowers at first subspreading, later becoming
pendulous, pale orange-yellow tinged with salmon-orange. Bracts ovate-lanceolate,
lanceolate or linear-lanceolate, acuminate, 6 -5-7 -5 mm long, 2 mm broad; margin
almost entire or minutely serrulate. Pedicels 0-5-1 mm long at flowering stage,
elongating to 2 mm long in fruit. Perianth subcylindrical, not constricted above the
ovary. 1-3-1 -5 cm long, 2-2-5 mm broad at the base, increasing gradually to 3-5-4
mm at the apex; lobes ovate, 1-5 mm long, not spreading. Stamens of different
lengths, the longest exserted by 3-4 mm at anthesis, later withdrawn into the mouth
of the perianth tube. Style subequal to the stamens at anthesis, eventually exserted
by 4-5 mm. Fruit subglobose to ovoid 6-7 mm long. Fig. 25. Plate 6. Map 4.
Map 3. — Distribution of Kniphofia buchananii,
K. breviflora and K. albescens.
Map 4. — Distribution of Kniphofia flammula,
K. evansii, K. fibrosa and K. crassifolia.
Plate 6. — Kniphofia flammula Codd
419
Fjg. 25. — K. flammula, from Glencoe, Natal ( Codd 6779).
Known only from a restricted area near Glencoe in the Dundee District of Natal
where it grows with dense tall grasses and sedges in and at the margins of vleis, at
an altitude of about 4,500 ft. The main flowering season is from mid-November to
the end of January.
Natal. — -Dundee: Kelvin Grove, Medley Wood 5188 (NH, Z); 31- miles N.W. of Glencoe, Codd
6779: Codd 8247; near Glencoe, Killick 2219.
If the pressed material is scrappy or the collector’s notes incomplete, this species
might be confused with K. albescens or K. breviflora. From the latter it differs in the
longer perianth, which is 13-15 mm as against 7-11 mm in K. breviflora. From both
species it differs in the buds being distinctly pendulous, instead of erect, and in the
flower colour. The bracts of K. flammula are also not as long acuminate as in the other
two species, though the specimen Medley Wood 5188 is intermediate in this respect.
11. Kniphofia evansii Bak. in FI. Cap. 6: 278 (1896); Berger in Pflanzenr. 4, 38:
42 (1908). Type: Natal, Drakensberg, Tiger Valley, Jan. 1895, Evans 353 (K!, holo.;
NFL!).
Plants solitary or in small groups. Leaves up to 30 per peduncle, grass-like, fibrous,
somewhat spreading, green to slightly glaucous, 60-75 cm long, 2-3 mm broad, keeled,
distinctly nerved, breaking into persistent fibres at the base; margin and keel smooth;
420
outer few leaves broad-based, clasping base of plant. Peduncle subequal to or shorter
than the leaves, 55-65 cm long, with a few sterile bracts below the inflorescence.
Inflorescence oblong to subcylindrical, dense, 5-10 cm long and 3-4 cm in diameter;
buds ascending, flowers spreading, eventually pendulous; buds and flowers yellow-
orange to scarlet, becoming dark purple on withering. Bracts ovate-lanceolate
to lanceolate, long acuminate, 4-6 mm long, 1 -5 mm broad, concave, scariose; margin
entire. Pedicels 1 mm long, elongating to 2 mm in fruit. Perianth subcylindrical,
not or scarcely constricted above the ovary, 1 -3-1 -6 cm long, 2 mm broad at the base,
expanding rather abruptly at the apex to 4 mm; lobes ovate, rounded, 2-5 mm long,
distinctly spreading. Stamens in the basal part of the perianth tube, of two lengths,
reaching approximately 4 mm and 6 mm from the base of the tube respectively. Style
1-1-5 mm long at anthesis, not elongating further. Fruit subglobose, 5-6 mm long.
Figs. 26, 27.
I
Fig. 26.— K. evansii, from Cathedral Peak
( Kill irk 1640).
Fig. 27. — K. evansii , inflorescence, life
size (Killick 1640).
421
'*4^4
^sdd.
■/*- r-ifi t+'/s tl /,'
. & **&*>,
—>A
lU^a^LM
/* ,?« 7 , ,,
Fig. 28— K. fibrosa, isotype in Natal Herbarium (Evans 649).
422
Known at present from only a small area on the upper slopes of the Drakensberg
in the Bergville District at altitudes of 6,000 to 7,000 ft., where it grows among hygro-
philous grasses and herbs on stream-banks and moist ledges. Its flowering time is
January to early February.
Natal. — Bergville: Drakensberg, Tiger Cave Valley, Evans 353 (K, NH); Cathedral Peak Forest
Research Station, Killick 1640; Killick 1642; Killick 1653; Germishuizen 42; Killick & Vahrmeijer
3552. Estcourt: Giants Castle Game Reserve, Trauseld 341.
The concolorous inflorescence of orange to scarlet flowers with spreading lobes
indicates a close relationship to the species in Berger’s Section Obtusilobae, in which
he includes K. triangularis Kunth, K. nelsonii Mast., K. macowanii Bak. and K. obtusiloba
Diels ex Berger (see p. 470). From all these entities K. evansii may be recognised by
its shorter flowers. In addition, K. evansii shows a characteristic which is unique in
the genus, namely, the fact that the stamens are included in the lower half or third
of the perianth tube, while the style does not exceed 1-5 mm in length. This results
in the stigma being placed below the anthers at anthesis and about 3 mm from the
base of the tube. After anthesis, the style does not elongate as is the case in other
species. It is a rare species, first collected by Maurice Evans in January, 1895.
An odd specimen, Trauseld 741, was recently collected on Giants Castle Game
Reserve which combines characters of K. evansii and K. porphyrantha, and is assumed
to be a natural hybrid between the two. In perianth length, which is 3-2 cm, and
general appearance the specimen resembles K. porphyrantha but the coral-red perianth
colour and the stamens extending to only half the length of the perianth tube suggest
hybridization with K. evansii.
12. Kniphofia fibrosa Bak. in FI. Cap. 6: 533 (1896); Berger in Pflanzenr. 4, 38:
41 (1908). Type: Natal, “ Mahwaqua Mt.”, Feb. 1896, Evans 649 (K!, holo.; NH!).
Plants apparently solitary. Leaves 6-10 per peduncle, grass-like, at first erect
then bending over about the middle, 35-60 cm long, 3-4 mm broad, keeled, breaking
into persistent fibres at the base; margin scaberulous to finely denticulate especially
towards the apex, often smooth towards the base. Peduncle overtopping the recurved
leaves, 30-60 cm long, with a few sterile bracts below the inflorescence. Inflorescence
subglobose to oblong, dense, 2-5-7 cm long and 2-5-3 cm in diameter; buds spreading,
flowers pendulous, pale yellow. Bracts ovate-lanceolate, acuminate, 5-7 mm long,
1 -5-2 mm broad; margin entire to minutely denticulate towards the apex. Pedicels
0-5-1 mm long. Perianth cylindrical, not constricted above the ovary, 1-2-1 -5 cm
long, 2-5-3 mm broad; lobes broadly ovate, 1 mm long, not spreading. Stamens
of two lengths, the longest exserted by up to 3 mm at anthesis. Style equalling the
anthers at anthesis, eventually exserted by 4-5 mm. Fruit not seen. Fig. 28. Map 4.
A little known species recorded from scattered localities in the eastern Cape Province
and the mountains of the Natal midlands, occurring usually in dense, sour grassveld.
The flowering season is February to March.
Cape. — King William's Town: Pirie, Taylor sub NBG 268/36 (BOL); Evelyn Valley, S.E. of Keis-
kammahoek, Bruce 549. Lusikisiki: near Port Grosvenor, Long s.n. Stutterheim: Dohne Hill, Sim
1143 (BOL).
Natal. — Drakensberg, Mrs. Prescott Decie s.n. (BM). Polela: Marwaga Mt., Evans 649 (K, NH).
Underberg: McClean 622.
Related to K. breviflora Bak. but differs in the longer flowers and minutely denti-
culate leaves. For differences between K. fibrosa and K. crassifolia , see p. 424.
A specimen from Ceza in the Mahlabatini District, Gerstner 4316, appears to be
closely allied to K. fibrosa. The bract and perianth characters agree very well, but
the leaf margins are smooth. More material is required from this locality before a
definite decision on its identity can be reached.
423
Fl°' from fhisTathenr^g h°,0type in Z (Rehmann 5796 from Houtbosch, Transvaal). Known only
24379—5
424
13. Kniphofia crassifolia Bak. in Bull. Herb. Boiss. 2 ser., 1: 784 (1901); Berger
in Pfianzenr. 4, 38: 42 (1908). Type: Transvaal, Houtbosch, Rehmann 5796 (Z!,
holo.).
Plants probably solitary. Leaves about 8 per peduncle, narrow, grasslike, 40-50
cm long, 3-6 mm broad, keeled, breaking into persistent fibres at the base; margin
conspicuously serrulate; outer leaves broad-based. Peduncle about 30 cm long.
Inflorescence subglobose to oblong, dense, about 5 cm long; flowers probably whitish
or pale yellow. Bracts lanceolate, acuminate, 5-7 mm long, entire. Pedicels 1-1-5
mm long. Perianth cylindrical, not constricted above the ovary, 1-5-1 -8 cm long,
1-5 mm broad at the base, increasing to 2-5 mm broad at the throat; lobes ovate,
1-5-2 mm long, slightly spreading. Stamens of two lengths, the longest exserted by
3 mm at anthesis. Style equal to the anthers at anthesis, eventually exserted by 4 mm.
Fruit not seen. Fig. 29. Map 4.
Known only from the type gathering in the Houtbosch area of the Transvaal.
Transvaal. — Petersburg: Houtbosch, Rehmann 5796 (Z).
Related to K. fibrosa Bak., but differs in the slightly longer flowers and the
conspicuous and characteristic serrulation of the leaf margin. More material of both
species is, however, required in order that the relationship between the two may be
more accurately evaluated.
14. Kniphofia gracilis Harv. ex Bak. in J. Linn. Soc. 11: 362 (1871); FI. Cap.
6: 278 (1896); Berger in Pfianzenr. 4, 38: 41 (1908). Type: Zululand, Gerrard &
McKen (Kl, holo.).
K. modesta sensu Bak. in Bot. Mag. t.7293 (1893); FI. Cap. 6: 277 (1896) partly
as to Bot. Mag. 7293 and Wood 4409; Mallett in Gard. Chron. 39: 83 (1906); Berger,
in Pfianzenr. 4, 38: 42 (1908); non K. modesta Bak. (1889). K. woodii Wats, in The
Garden 48: 292 (1895) (as “ woodi ”). Type: Hort. Kew, Oct. 14, 1895 (K! holo).
K. natalensis var. condensata sensu Bak. in FI. Cap. 6: 533 (1897), partly, as to syn.
K. woodii. K. modesta var. woodii (Wats.) Berger, l.c. (1908). var. lutescens
Berger, l.c. 43 (1908). Type: Manderston, Wood 5463 (K!, holo. ; SAM!). K. sparsa
N. E. Br. in Gard. Chron. ser. 3, 56: 410 (1910). Type: the plant on which Bot.
Mag. t.7293 is based, hort. Kew, Oct. 1893 (K!, holo.!). K. wyliei N. E. Br. in Kew
Bull. 1910: 129 (1910). Type: Zululand, Wylie sub Wood 8996 (K!, holo.; NH!).
Plants solitary or in small clusters. Leaves 8-10 per peduncle, recurving about
the middle, 40-120 cm long, 4-14 mm broad, channelled above, keeled below; margin
smooth to sparingly serrulate, often only towards the apex. Peduncle overtopping
the recurved leaves, 25-80 cm long, usually with a few sterile bracts below the inflore-
scence. Inflorescence oblong to narrowly cylindrical, usually tapering towards the
apex, 9-35 cm long and 2 -8-3 -8 cm in diameter dense or lax; buds ascending, over-
lapped by the bracts, whitish tinged with pink, to yellow or orange; flowers at first
spreading, later deflexed, whitish, creamy-yellow or yellow. Bracts broadly ovate to
oblong, 4-6-5 mm long, 2-5-3 mm broad, concave, obtuse to rounded, rarely acute,
scariose with a median brown nerve; margin minutely serrulate Pedicels 1 - 5—2 mm
long. Perianth subcylindrical to narrowly funnel-shaped, 1- 1-2-0 cm long, slightly
to markedly constricted above the ova y, 1-2 mm in diameter near the base, widening
gradually to 4-5 mm at the throat, or widening more or less abruptly about the middle;
lobes ovate to oblong, 2-2-5 mm long, rounded, often spreading. Stamens in the
throat of the perianth or exserted by up to 2 mm at anthesis. Style subequal to the
stamens at anthesis, eventually exserted by 5 mm. Figs. 30, 31, 32. Plate 7. Map 5.
425
AW,
■v/:Ayy
<//. < A^~ *•*♦*;> $ *
<■ s^^f< -
J/~X^ Post XaTal, B. Afkxca
Oerr&rd k M'Ke.
Fig. 30. K. gracilis, holotype in K ( Gerrard & McKen 2140 from Port Natal).
426
Distributed from Mt. Ayliff District in the eastern Cape Province to Hlabisa
District in Zululand, occurring on grassy slopes from near sea level to mountain grass-
land at altitudes of 2,000 to 5,000 ft. The flowering season is from December to April.
Cape. — Mt. Ayliff: 154 miles from Kokstad on road to Port St. Johns, flowered in cultivation,
Pretoria, Marais 850.
Natal. — Without locality: Hort. Kew, Oct. 1895, type of K. woodii (K). District unknown: Man-
derston, Medley Wood 4827 (K, NH); 5463 (K, SAM); Zululand, Medley Wood in NH 11316 (NH)
Camperdown: Near Botha’s Medley Wood s.n. (NH, this sheet bears an original label with the number
978 which is crossed out); 978 (BM); 4409 (K); 6652; Inchanga, McClean s.n. (NH); Bothas Hill,
Strey 7376; Drummond, Strey 7377. Durban: near Durban, Medley Wood 11591 (NH); Jenkins
in TRV 7047; Brighton Beach, Forbes 503 (NH); Bruce 148; Bluff, near Wentworth, Strey 7380.
Eshowe: Medley Wood 1 1098 (NH): Lawn 279 (NH); Ngoye, Medley Wood 10367 (NH); Entumeni,
Medley Wood in NH 3726 (NH); Wylie sub Wood 8996 (K, NH); Umhlatuzi Valley, Gerstner in
NH 22565 (NH); Melmoth, Schelpe 5187 (BM). Hlabisa: Hluhluwe Game Reserve, Ward 3376.
Ixopo: Highflats, Gerstener 7005. Kranskop: 11 miles N.W. of Mapumulo, Codd 10191. Nkandla:
Wylie sub Wood 8824 (NH); near Qudeni, Codd 6995; 8 miles S. of Nkandla, Reynolds 5851. Port
Shepstone: Uvongo Beach, Liebenberg 8047. Richmond: Inhlasuka, Gerstner 7006. Umzinto:
Campbelltown, Rudatis 1881 (G, Z).
A good deal of variation is included among the above citations, particularly with
regard to length and density of the inflorescence and degree of constriction of the
perianth above the ovary (sometimes difficult to interpret in herbarium specimens).
Flower colour may also vary from almost white to cream or yellow, usually more deeply
pigmented, or tinged with red, towards the apex. In typical K. gracilis the inflorescence
is short and relatively dense with the perianth 14-20 mm long, more or less parallel-
sided, expanding at the mouth. On the other hand, the plants described as K. woodii
and K. sparsa (i.e. “ K. modesta ” of Bot. Mag. t.7293) have lax, elongate inflorescences
with the perianth 11-16 mm long, constricted above the ovary and expanding about
the middle, varying in colour from white (K. sparsa) to yellow (K. woodii and K. modesta
var. lutescens). There are, however, several herbarium specimens (including the type
of K. wyliei ) which are intermediate in inflorescence and perianth characters, making
it impossible to separate the material satisfactorily into infraspecific groups in the
herbarium. The possibility cannot be excluded that further investigation in the field
may well reveal good grounds for separating the lax-flowered form as having distinct
status.
The somewhat involved citations included in the synonymy of the species have
resulted from several extraordinary misconceptions and nomenclatural errors perpet-
uated by Baker and Berger, and may be clarified by the following explanatory notes.
A plant with lax inflorescences and short, white flowers was figured in Bot. Mag.
t.7293 (1893), to which the name “ K. modesta” was erroneously applied by Baker,
the true K. modesta Bak. (1889) being a synonym of K. parviflora Kunth. The mistake
was recognized by N. E. Brown, who gave the name K. sparsa N. E. Br. (1910) to the
Bot. Mag. plate, based on a plant which flowered in Kew Gardens in October, 1893,
and is preserved in Kew Herbarium.
Shortly after the above Bot. Mag. plate appeared, W. Watson, in The Garden
(October, 1895), described a lax-flowered plant under the name K. woodi which had
yellow flowers and was related to “ K. modesta ”. One feels that this was a provisional
description and scarcely qualified as valid. Two specimens (flowered September and
October, 1895, respectively) are preserved in Kew Herbarium, annotated “ K. woodii ”
in N. E. Brown’s handwriting. Although the authorship of K. woodii is attributed
to “ Bak.” or “ Bak. ex Wats.”, Baker did not uphold the name and, in Flora Capensis
6: 533 (1897), he included “ K. woodii Bak.” in synonymy under K. natalensis var.
condensata Bak., a variety described in FI. Cap. 6: 281 (1896) from a specimen received
from Max Leichtlin in January, 1895. This latter specimen is somewhat intermediate,
but is nearer to K. laxiflora {=K. natalensis ) than it is to K. gracilis.
<•
427
Fig. 31. — K. gracilis, typical form with relatively Fig. 32. — K. gracilis, form with elongate
dense inflorescences, at Qudeni ( Codd 6995). inflorescence from near Durban (Bruce 148).
There followed an inexplicable step by Berger in his monograph, p. 42 (1908).
Having correctly cited the type specimen of true K. modesta Bak. (1889) under K.
parviflora Kunth, he upheld the name K. modesta Bak. in the sense of the Bot. Mag.
plate 7293 (1893). This he distinguished from K. gracilis mainly on the basis of the
perianth being distinctly constricted near the base. Berger also separated two additional
varieties within his “ K. modesta One of these, var. lutescens Berger, was based on
Medley Wood 5463 from Manderston, Natal. The other, var. woodii Berger (which
he validly published), is based on the same plant as K. woodii Wats. Both varieties
have long, lax inflorescences of yellow flowers and are indistinguishable from one
another.
428
Finally N. E. Brown described K. wyliei in 1910. The type, Wylie sub Wood
8996 in Kew Herbarium, has a somewhat laxer inflorescence than the type of K. gracilis,
but there is no justification for separating it from that species.
A few specimens have been seen with perianths about 2 cm long, which are thus
intermediate between K. gracilis and K. laxiflora (perianth usually 2-4-3 -2 cm long).
The distinction between K. gracilis and K. laxiflora becomes, therefore, a bit arbitrary,
but no advantage can be seen in merging the two and making an even more heterogeneous
aggregation of material. Thus the present view is that K. gracilis varies in perianth
length from 1 -1 to 2-0 cm, and in colour from white to yellow (often tinged with red
at the apex). In K. laxiflora, plants with yellow flowers may be found but there are,
in addition, striking shades of red, orange and salmon colours which are not recorded
in K. gracilis. Both species have oblong, obtuse to rounded bracts as against the species
K. rufa and K. ichopensis in which the bracts are lanceolate and acute to acuminate
at the apex.
15. Kniphofia laxiflora Kunth, Enum. PI. 4: 552 (1843); Bak. in FI. Cap. 6: 281
(1896); Berger in Pflanzenr. 4, 38: 50 (1908). Type: Pondoland, between Umtentu
and Umsikaba Rivers, Drege 4527 (K!, lecto.).
K. natalensis Bak. in J. Bot. Lond. 23: 278 (1885); FI. Cap. 6: 281 (1896), partly,
excl. var. angustifolia Bak.; Wood, Natal PI. 4, 1: t.305 (1903); Berger, l.c. 50 (1908).
Type: Inanda, Medley Wood 636 (K!, holo.; BM). var. condensata Bak. in
FI. Cap. 6: 281 (1896). Type: Hort. Leichtlin, received Jan. 1895 (K!, holo.).
var. robustior Berger, l.c. 52 (1908). Type: Zululand, Tugela River, Medley Wood
3871 (K!, holo.). K. decaphlebia Bak. in Bot. Jahrb. Beibl. 35: 6 (1892); FI. Cap.
6: 282 (1896); Berger, l.c. 57 (1908). Type: Pondoland, Feb. 1888, Bachmann 281
(Bf).
Plants solitary or in groups. Leaves 8-12 per peduncle, recurving about the
middle, 50-120 cm long, 6-10 (rarely-15) mm broad, channelled above, keeled below,
semi-glaucous; margin smooth or serrulate. Peduncle overtopping the recurved
leaves, 40-90 cm long, usually with few to many long sterile bracts below the inflore-
scence. Inflorescence oblong and fairly dense to markedly elongate and lax, tapering
at the apex, 10-45 cm long and 4 - 5—5 - 5 cm in diameter; buds at first ascending
overlapped by the broad bracts; flowers at first spreading, later deflexed, varying
greatly in colour from pale yellow or yellow-green to orange, salmon-pink, coral-red
or red-orange, but not white or cream, buds usually darker in colour, but not producing
a distinctly bicolorous effect. Bracts broadly ovate to ovate-oblong, 6-9 mm long,
3 mm broad, concave, obtuse or rounded at the apex, scariose, with a median brown
nerve; margin minutely eroso-serrulate. Pedicels 1-5-2 mm long. Perianth sub-
cylindrical, 2 -4-3 -5 cm long, slightly to markedly constricted above the ovary, often
somewhat curved, 2-2-5 mm in diameter above the ovary, widening gradually to 4-5
mm at the throat; lobes ovate to oblong, 2-2-5 mm long, rounded, often spreading.
Stamens of different lengths, scarcely exserted at anthesis, later withdrawn. Style
subequal to the stamens at anthesis, eventually exserted by 4-5 mm. Fruit ovoid-
triquetrous, large, up to 1 cm long. Figs. 33, 34, 35, 36, 37. Map 6.
A variable species, found on grassy slopes and among rocks at altitudes from
near sea level in the neighbourhood of Port St. Johns, through Pondoland, Natal
coastal districts and midlands, to over 5,000 feet above sea level in northern Natal,
and just entering southern Transvaal on the northern slopes of the Drakensberg Range.
The normal flowering time is from February to May.
429
Fig. 33. — K. laxiflora, typical form growing near Lusikisiki ( Codd 9307).
Cape.- Lusikisiki: between Umtentu and Umsikaba Rivers, Drege 4527 (K); near Lusikisiki,
Bruce 598; Codd 9307; 3+ miles N.W. of Mtafufu Mouth, Codd 9305; near Dindini Store, Codd
9319. Mt. Currie: Tyson 1768 (BOL, GRA, SAM); 2 miles W. of Ingeli Forest, Codd 8560. Port
St. Johns: near Port St. Johns. Flanagan 2547; Bruce 599. Tabankulu: 5 miles S. of Tabankulu,
Bruce 593; 14 miles S. of Tabankulu, Bruce 595. Umzimkulu: 6 miles S. of Umzimkulu, Marais 840;
2\ miles S. of Kingscote, Marais 1421.
Natal. — Alfred: near Staffords Post, Marais 844. Dundee: Shirley s.n. (NU). Durban: Berea
Medley Wood 1 1452. Estcourt: Griffins Hill, Acocks 1 1217; 11218; 11219; 10 miles S.E. of Estcourt
Codd 8518; near Mooi River, O'Connor s.n.; Germishuizen 51. Hlabisa: St. Lucia Bay, Van der
Merwe 2390\ Rodin 4736; Dukuduku Forest, Harrison s.n. ; Codd 7004: Killick 3835; near Hluhluwe,
Harrison s.n. Mthole Hill, Letley H. 18. Impendhle: Gordon-Gray s.n.; 10 miles S.E. of Impendhle
Marais 822. Ixopo: 3 miles N.W. of St. Faiths, Codd 9365. Lions River: Reynolds 3440; Codd
6796; Balgowan, Bruce 159; 160; 163; Tweedie, Mogg 1357; Lidgetton Medley Wood 9762 (NH);
Mogg 3568; near Nottingham Road, Gerstner s.n.; near Howick, Medley Wood in NH 3583 (NH);
Karkloof, Wylie s.n. (NH). Mtunzini: Ngoya, Medley Wood 10367 (NH); Huntley 264; near
Mtunzini, Codd 9646; Gerstner s.n. (NH). Ndwedwe: Inanda, Medley Wood 636 (BM, K); 4762
(NH). Ngotshe: near Ngome, Bruce 294; 296; 299. Paulpietersburg: between Kempslust and
Paulpietersburg, Bruce 290. Pietermaritzburg: Zwartkop, Medley Wood 11424 (NH). Polela: near
Bui wer, Codd 8541. Port Shepstone: 8 miles N. of Port Shepstone, Codd 9354. Richmond: Gerstner
430
7007; 7008. Stanger: Tugela River, Medley Wood 3871 (K). Umzinto : Dumisa, Rudatis 617 (BM>
K, Z). Underberg; McClean 620 (NH); 621 (NH). Utrecht: 4 miles W. of Viljoenspos, Codd 7955,
9 miles N.E. of Utrecht, Bruce 288.
Transvaal.— Amersfoort: 12 miles S.W. of Dirkiesdorp, Bruce 279.
FIG. 34. X. laxiflora, typical form ( Codd 9307 from Lusikisiki).
431
Fig. 35. — K. laxiflora, form with fairly elongate
inflorescences, near Nottingham Road
(N. R. Smuts 1021).
Fig. 36. — K. laxiflora, form with markedly I
elongate inflorescences, from near Est- ' -3 ►
court, x 3- ( Codd 8518).
432
The decision to reduce K. natalensis to synonymy results from the difficulty
experienced in attempting to subdivide the specimens cited above into distinct groups.
K. decaphlebia has also been included although the type in Berlin has been destroyed
and no further material of the type gathering has been seen. Berger, who saw the
type, included it in the K. uvaria complex but, from the description, flowering time
and locality, it could equally well be a form of K. laxiflora. The separation of these
as distinct species on the basis of the keys provided in Flora Capensis and Berger’s
monograph does not work in practice when a range of material is available.
However, it is possible to recognise three main forms among the above citations
and further field work is necessary to determine what status should be given to them.
At this stage in our knowledge it is preferred not to give them separate taxonomic
rank. The characteristics and distribution of the three forms is set out below.
Form A. This is the typical form represented by Drege 4527. The plants tend
to be solitary or in small groups; the inflorescence is relatively short, 8-20 cm long,
fairly dense about the middle, tapering and more lax towards the apex; flower colour
varies mainly from orange-red to yellow. The distribution of this form covers a wider
area than the other two, extending from Port St. Johns, through Pondoland and along
the coastal areas of Natal to Lake St. Lucia in the north. In the northern part of the
range, for example in Mtunzini and Hlabisa Districts, only the yellow flower colour
is recorded, and the plants tend to flower later (April-May) than is the case further
south (February-April).
Form B. This form is commonly found in the Natal Midlands and extends from
East Griqualand to the Estcourt District and may overlap with Form A between
Pietermaritzburg and Durban. The two forms cannot always be distinguished in the
herbarium. The plants of Form B tend to be in clusters, on grassy slopes or among
rocks; the inflorescence is distinctly elongate, 18-45 cm long, and lax-flowered, tapering
gradually to the apex; flower colour shows an extensive range from pale yellow or
yellow-green to orange, salmon-pink, coral-red or orange-red, the colour varying
among adjacent plants. Flowering time is normally February-March, and the groups
of graceful inflorescences in various shades of colour make this one of the most
attractive members of the genus. Although the name K. natalenis Bak. has frequently
been applied to this form, it is not at all certain that the type, Wood 636 from Inanda
(in Herb. Kew), represents this form with the markedly elongate inflorescences. As
regards the varietal names included in synonymy, var. condensata Bak. is based on a
cultivated plant in the garden of Max Leichtlin, Baden-Baden. The type material
at Kew consists of several inflorescences in the bud stage with densely imbricate bracts;
one inflorescence, which is partly open, has flowers (possibly abnormal) about 2 cm
long, and a leaf about 2 cm broad. This may be an abnormal specimen of Form B,
but var. condensata Bak. is best treated as a name of dubious application. The other
variety, robustior Berger, is based on Wood 3871 (in Herb. Kew) from Zululand and
also appears to represent Form A rather than Form B.
Form C. This form occurs among dolerite rocks in the Drakensberg range north-
east of Utrecht. The plants grow in groups with subglaucous, arcuate-spreading
leaves, and the inflorescence is relatively short, 12-18 cm long, dense at the middle
and somewhat laxer towards the tapering apex; the flower colour is generally yellow-
green with the buds often tinged with dull red. Not many gatherings of this form are
known and it requires further investigation.
K. laxiflora may be distinguished from other lax-flowered species, such as K.
ichopensis and K. tabularis, by its broader bracts which are rounded or obtuse at the
apex. A few specimens have been seen, however, with bracts of intermediate shape,
which may be hybrids.
433
Fig. 37. — K. laxiflora, form growing in northern Natal, always among rocks ( Codd 6955).
16. Knophofia ichopensis Bak. ex Schinz in Bull. Herb. Boiss. ser. 2, 2: 1066
(1902); Berger in Pflanzenr. 4, 38 : 52, t.20E (1908); Codd in Flow. PI. Afr. 33 : 1. 1 292
(1959). Type: Natal, Ixopo, Schlechter 6662 (K!, holo.; BOL!, GRA!, PRE!, Z!).
K. schlechteri Bak. in Bull. Herb. Boiss. ser. 2, 1: 784 (1901), non Schinz (1900).
Type: as for K. ichopensis.
Plants solitary or in small groups. Leaves 6-8 per peduncle, more or less distichous,
soft, erect or curving over near the apex, 50-80 cm long and 5-10 mm broad, dull green
to subglaucous, shallowly channelled above and keeled below; margin and keel smooth
or remotely and minutely denticulate towards the apex. Peduncle 50-90 cm long,
overtopping the leaves, with 2 or 3 sterile deltoid bracts below the inflorescence.
Inflorescence lax, oblong-cylindrical, 10-30 cm long and 6-7 cm in diameter; buds
dull yellow, often tinged with red, to coral or dull red; flowers at first spreading, soon
deflexed, varying in colour from cream to yellow-green or, rarely, salmon-pink. Bracts
lanceolate to linear-lanceolate, acute to acuminate, 7-11 mm long, 2-5 mm broad
434
near the base; margin entire. Pedicels 2-3-5 mm long. Perianth subcylindric, often
slightly arcuate, not constricted above the ovary, 3 -0-4 -2 cm long, 2-5-3 -5 mm broad
at the base, widening gradually to 4-5 mm at the throat; lobes ovate, 4 mm long,
not spreading. Stamens included or slightly exserted at anthesis, later withdrawn;
anthers usually visible through the perianth tube. Style subequal to the stamens at
anthesis, finally exserted by 3-5 mm. Fruit subglobose to ovoid, about 6 mm long,
not markedly triquetrous. Plate 8. Map 5.
Found in grassy vleis and marshy places, occurring mainly in the Natal Midlands,
from Pietermaritzburg and Greytown to the foothills of the Drakensberg in the Estcourt
District, at altitudes of 3,000 to 6,500 ft. Flowering is mainly from December to
March, but may be as early as November and as late as April.
Map 5. — Distribution of Kniphofia gracilis and Map 6. — Distribution of Kniphofia laxiflora
K. ichopensis. and K. rufa.
Natal. — Estcourt: Kamberg, Smuts 1066; Tabamhlope, Pent: 67; West 595; Killiek & Marais
2167; Highmoor Forest Station, Killiek & Vahrmeijer 3590; 3648; Giants Castle Game Reserve,
Trauseld 372a. Ixopo: Schleehter 6662. Lions River: Fort Nottingham, Buchanan 142 (K); Wylie
sub Wood 7154; Lidgetton, Medley Wood 6334 (K, NH); Nottingham Road, Shirley s.n. (NU);
15 miles S.W. of Nottingham Road, Codd 10651. New Hanover: between Pietermaritzburg and
Greytown, Killiek & Marais 2112; Hardy s.n. Pietermaritzburg: Zwartkop, Medley Wood in NH
11341 (NH); Reynolds 3721. Umvoti: near Greytown, Wylie in NH 21686 (NH); Killiek & Marais
2114.
In its lax, usually concolorous inflorescence, K. ichopensis resembles some forms
of K. laxiflora, but it may be recognised by the narrowly lanceolate, acute to acuminate
bracts, and by the perianth showing no constriction above the ovary. Its nearest
affinity is K. rufa Bak. a variable species discussed under the next heading, and there
is some indication of intergrading between the two. K. ichopensis is the more robust
of the two with longer, broader (5-10 mm) leaves and taller, stouter peduncle with
longer flowers (3 -0-4 -2 cm long).
17. Kniphofia rufa Bak. in Bot. Mag. t.7706 (1900); Mallett in Gard. Chron.
39: 101 (1906); Berger in Pflanzenr. 4, 38: 56 (1908). Type: Hort. Max Leichtlin,
Baden-Baden (K!, holo.).
K. natalensis var. angustifolia Bak. in FI. Cap. 6: 281 (1896). Type: Natal,
Tabamhlope Mt., 6-7,000 ft., February, 1895, Evans 411 (K!, holo.; NH!, PRE!).
435
Plants solitary or in small groups. Leaves 6-10 per peduncle, not distichous,
soft, erect at first, later arcuate, 40-70 cm long and 2-5 (-8) mm broad, dull green to
subglaucous, more or less triangular in cross section with a narrow channel above,
slightly keeled below; margin and keel smooth or remotely denticulate towards the
apex. Peduncle 40-65 cm long, slender, overtopping the leaves, with an occasional
sterile deltoid bract below the inflorescence. Inflorescence usually lax, rarely subdense,
oblong-cylindrical, 8-25 cm long and 5-6 cm in diameter; buds cream or dull yellow
to orange-red; flowers pendulous, often somewhat arcuate, varying in colour from
white to cream, yellow or coral-red. Bracts lanceolate to linear-lanceolate, acute to
acuminate, 6-10 mm long, 2-2-5 mm broad near the base, margin entire. Pedicels
2-3 mm long. Perianth subcylindric, not constricted above the ovary, 1 -9-3-0 cm
long, 2-5 mm broad at the base, widening gradually to about 4 mm broad at the throat;
lobes ovate, 2-5 mm long, not or slightly spreading. Stamens included or slightly
exserted at anthesis, later withdrawn; anthers usually visible through the perianth
tube. Style subequal to the stamens at anthesis, finally exserted by 3-5 mm. Fruit
subglobose to ovoid, about 6 mm long, not markedly triquetrous. Figs. 38, 39, 40.
Plate 2a. Map 6.
Found beside mountain streams, on grassy slopes and in marshy places in mountain
grassland, at altitudes of 4,500 to 7,500 ft., chiefly along the eastern slopes of the
Drakensberg range from Polela to Estcourt Districts. The flowering season extends
from the end of November to early April, the peak period being January-March.
Natal. — Estcourt: Tabamhlope Mt., Evans 411 ; West s.n.; Bushmans Pass, West s.n.; Cathkin
Park, Galpin 11347; 11759; 11774; Cathedral Peak Forest Station, Killick 1405; 1431; 1442; 1466;
1467; Bruce 407; Germishuizen 47; 48; near White Mt. Inn, Killick & Marais 2161; Kamberg,
Smuts 1064; 1065; 1072, Wright 2. Lions River: near Nottingham Road, Smuts 1028/1; 1028/2;
1038; 1039; 1414; Codd 8533. Polela: near Bulwer, Lansdell in NH 15802 (NH); Killick & Marais
2099; Marwagwa Mt., Pole Evans 4869; upper Umkomaas, Van der Merwe 2830.
This species is characterised by its narrow, grass-like leaves and slender inflore-
scences which are medium-lax to very lax, with flowers about 2-3 cm long, varying
in colour from white to yellow or coral-red. The latter colour is particularly attractive
and is reminiscent of the flower colour found in the K. triangularis complex. The
variation is reflected in the multiple gatherings made by several modern collectors
(usually with notes on flower colour) but, unfortunately, the colours cannot be discerned
in most herbarium specimens. Largely for this reason, no attempt has been made to
subdivide the species into varieties, though further study may well reveal that sub-
divison is possible.
There is, for instance, some indication that colour may be associated with perianth
length. At one end of the range of variation is the type of K. natalensis var. angustifolia
(Evans 411) with white flowers about 1-9 cm long, and medium-lax inflorescences.
At the other extreme are specimens with coral-red flowers about 3 cm long, borne on
elongate, often arching inflorescences. These latter plants are among the most graceful
members of the genus. Some specimens of this form have inflorescences which are
more compact than usual, and are then difficult to separate with certainty from the
Natal form of K. triangularis subsp. triangularis (see p. 000). Thus it is necessary
to consider whether the plants with coral-red flowers should not rather be placed as
a lax-flowered variety of K. triangularis than as a colour form of K. rufa. It seems
probable that K. triangularis has played a part in the evolution of the lax, coral-red
inflorescences but, as mentioned earlier, colour is not easily discerned in herbarium
specimens, whereas a separation can be made on the basis of relative density. It is,
therefore, preferred to include these plants as a colour form of K. rufa.
The type of K. rufa is a plant of doubtful background and is not matched exactly
by any plant collected in the wild state. It was grown by Max Leichtlin in his nursery
at Baden-Baden and forwarded to Kew in June, 1899, where it was figured for the
436
Fig. 38 —K. rufa, holotype in K (cultivated plant).
437
Botanical Magazine t.7706, published in 1900. The specimen is preserved in Kew
Herbarium, together with a specimen cultivated subsequently at Kew. The illustration
shows a medium-lax inflorescence of yellow flowers about 2-4 cm long, tinged with
red in the bud stage. The floral characters are not unlike some of the wild specimens
cited above, but the leaves are somewhat broader, namely, 8 mm broad as against
2-5 mm for the wild specimens.
Fig. 39. — K. rufa , form with coral-red flowers Fig. 40. — K. rufa, form with white flowers
(N. R. Smuts 1064, Kamberg, Natal). (N. R. Smuts 1028/1 from Nottingham
Photo by Dr. N. R. Smuts. Road, Natal). Photo by Dr. N. R. Smuts.
Regarding the origin of the type plant. Max Leichtlin reports in a letter to Kew,
which is filed with the type specimen: “The Kniphofia came with original seeds of
K. nelsonii and seeing some difference in leafage I kept it separate ”. Unfortunately
he does not record where the seeds of K. nelsonii (now placed as a synonym of K.
triangularis, p. 469) came from. It may also be significant to quote from Mallett in
Gard. Chron. 39: 101 (1906): “ It appears incapable of producing seeds, and the
pollen is often sterile ”.
There is thus some indication that the type of K. rufa is of hybrid origin, but the
evidence is inconclusive. Although the most desirable procedure would be to discard
the name K. rufa, it seems necessary to accept, though with reluctance, the nomen-
clatural legacy resulting from the repeated descriptions of species based on cultivated
plants, and to apply the name to the taxonomic group which it resembles most closely
i^see also K. praecox, p. 445 and K. citrina, p. 497).
438
a I Fig. 42. — K. pauciflora, inflorescence,
life-size ( Reyburn s.n.).
Fig. 41. — K. pauciflora, from near
Pinetown ( Reyburn s.n.).
18. Kniphofia pauciflora Bak. in J. Bot. Lond. 23: 280 (1885); Gard. Chron.
ser. 3, 12: 65, t.10 (1892); Bot. Mag. t.7269 (1892); FI. Cap. 6: 279 (1896); Mallett
in Gard. Chron. 39: 101, t.43 (1906); Berger in Pflanzenr. 4, 38 : 41 (1908). Syntypes:
Natal, Sanderson 416 (K!); Clairmont Flats, Wood 1096 (K!, NH!).
K. pedicellata Bak. in Bull. Herb. Boiss. ser. 2, 4: 998 (1904). Type: Clairmont,
Schlechter 3160 (Z!).
Plants with short stolons forming groups of stems. Leaves 4-6 per peduncle,
soft, erect, yellow-green, 20-35 cm long, 2-8 mm broad, shallowly channelled above
and keeled below, more or less triangular in cross section; keel and margin smooth.
Peduncle overtopping or subequal to the leaves, 30-50 cm long. Inflorescence ovoid,
lax, few-flowered, 3-10 cm long and 3-3-5 cm in diameter; buds ascending, deep
yellow; flowers at first ascending, eventually deflexed, yellow to pale yellow. Pedicels
3-4 mm long. Bracts narrowly lanceolate, gradually acuminate, 4-6 mm long; margin
entire. Perianth narrowly funnel-shaped, 1-4-1 -8 cm long, not constricted above the
ovary, 2-5 mm in diameter near the base, increasing to 5-6 mm at the throat; lobes
439
ovate, obtuse, 3 mm long, spreading. Stamens exserted by 2-3 mm at anthesis, later
withdrawn. Style subequal to the stamens at anthesis, eventually exserted by 5 mm.
Fruit small, subglobose. Figs. 41, 42.
Known only from a restricted area near Durban, where it grows in marshy grassland.
Flowering is mainly from September to November.
Natal. — District unknown: Sanderson 416 (K); between Durban and Pietermaritzburg, Wahtberg
s.n. (S). Durban: Clairmont, Medley Wood 1096 (BOL, K, NH); 4662 (BOL, NH, Z); 7293; 10495
(NH, Z); Schlechter 3160 (Z). Pinetown: on road to Marionhill, Reyburn s.n.
With its short stature and small lax inflorescences, this is one of the least conspicuous
members of the genus. It shows some relationship to K. rufa (p. 434), but its nearest
affinity is clearly to the solitary species in Madagascar, K. pallidiflora Bak., which has
short, funnel-shaped flowers and long pedicels.
The species is threatened as a result of urban development around Durban and can
no longer be found at the type locality, Clairmont. The late Mr. J. W. Reyburn, who
collected specimens near Marionhill in 1951 and 1954, informed me that the area has
since been drained and the plants which he knew can no longer be found.
19. Kniphofia tabularis Marl, in Trans. S.A. Phil. Soc. 18: 49 (1907); FI. S. Afr.
4: 102, t.24c (1915); Berger in Pflanzenr. 4, 38: 329 (1908). Type: Cape, Table
Mountain, Marloth 4366 (PRE!, holo.; BOL!).
Plants in groups in crevices on moist cliff faces. Leaves 12-18 per peduncle, soft,
bright green, pendulous, 60-1 50 cm long, 6-20 mm broad, channelled above and keeled
below, breaking into numerous fibres at the base, the outer short, broad and sheathing
at the base; keel and margin smooth. Peduncle curving-erect, 60-120 cm long, with
an occasional sterile bract below the inflorescence, occasionally branching. Inflorescence
very lax, oblong to cylindrical, 10-25 cm long, 5-5-7 cm in diameter; buds ascending;
flowers at first ascending, later spreading and eventually deflexed, red or scarlet to
pale orange, blackish at the tips of the lobes. Pedicels conspicuous, 5-7 mm long,
elongating to 10 mm in fruit. Bracts linear-lanceolate, long-acuminate to aristate,
7-11 mm long, 2-2-5 mm broad; margin entire. Perianth subcylindrical, 2-4-3 cm
long, not constricted above the ovary, 2-5 mm wide at the base, widening to 6-7 mm
at the throat; lobes ovate, 3 mm long, not spreading. Stamens in the throat or slightly
exserted at anthesis, yellow. Style finally exserted by 3 mm. Fruit subglobose to ovoid,
7 mm long. Map 8.
Grows on moist cliffs on the southern aspect of Table Mountain and in similar
situations on mountains in neighbouring districts at altitudes of 2,500 to 5,500 ft.
Flowers appear in December and January. Marloth states: “A very handsome plant
which seems to grow in a few, not easily accessible spots and, as the flowering season
hardly lasts one month, the plant would not be noticed during the remainder of the
year. The flowers were visited by sugarbirds ( Anthobaphes violacea ) and a butterfly
( Meneris tulhaghia ) ”.
Cape. — Caledon: Hottentot Holland Mts., Stokoe 201; Landdrost Kop, Esterhuysen 2638 (BOL).
Peninsula: Table Mt., Bolus in BOL 24692 (BOL); Marloth 4366; Bruce 245; Esterhuysen 24694;
Compton 12936 (NBG). Stellenbosch: Banhoek Mts., Marloth 10028; Esterhuysen 14369; Victoria
Peak, Esterhuysen 9733; Drakenstein Mts., Stokoe s.n. (BOL, SAM). Tulbagh: Groot Winterhoek,
Andreae 903; 1132. Worcester: Milner Peak, Esterhuysen 14236 (BOL); Waaihoek Peak, Ester-
huysen 18341.
A distinct species with long, pendulous leaves, markedly lax inflorescences and long
pedicels, first collected on Table Mt. by Bolus in December, 1879. The peduncle
occasionally branches sparingly.
24379—6
440
20. Kniphofia ensifolia Bak. in J. Bot. Lond. 23: 278 (1885). Type: Transvaal,
Matebe River, Holub 1530 (K!, holo.).
Plants in groups. Leaves 8-12 per peduncle, erect, bending over towards the
apex, glaucous, 50-120 cm long, 1 -5-3-5 cm broad, V-shaped in cross section, broad
and chartaceous at the base, not breaking up into fibres; margin and keel usually
conspicuously serrulate, rarely smooth. Peduncle 60-180 cm long with usually a few
sterile bracts below the inflorescence. Inflorescence very dense, cylindrical, tapering
slightly towards the apex, 9-20 cm long and 3-5-4 cm in diameter; buds spreading,
greenish-white, red-tinged, or dull to flame-red; flowers pendulous, greenish white to
cream or yellow (subsp. autumnalis) with a green line down each perianth lobe. Pedicels
2-3 mm long, increasing to 5 mm in fruit. Bracts lanceolate to linear-lanceolate,
gradually acuminate, 7-10 mm long, 2 mm broad at the base; margin smooth or,
rarely, minutely eroso-serrulate. Perianth narrowly funnel-shaped, 1-5-2 cm long,
2-2-5 mm in diameter at the base, expanding about the middle to 4-5 mm and eventually
5-6 mm at the throat; lobes ovate, 2-5 mm long, not spreading. Stamens exserted
by 6-8 mm or more at anthesis. Style subequal to the anthers at anthesis, eventually
exserted by 8-10 mm. Fruit subglobose, 6 mm long.
A robust species with an essentially inland distribution west of the Drakensberg
escarpment. Two subspecies are recognized as indicated below.
Flowering time October-December; flowers on opening white to greenish white (a) subsp. ensifolia
Flowering time February-March; flowers on opening usually yellowish (b) subsp. autumnalis
(a) subsp. ensifolia.
K. ensifolia Bak. in J. Bot. Lond. 23: 278 (1885); FI. Cap. 6: 280 (1896); Berger
in Pflanzenr. 4, 38: 45 (1908); Letty, Wild Flows. Tvaal. 33, 1. 19, 2 (1962). Type:
Transvaal, Matebe River, Holub 1530 (K.!, holo.). K. tuckii Bak. in Gard. Chron.
13: 68 (1893); FI. Cap. 6: 278 (1896); Bot. Mag. t.7644 (1899); Bean in The Gard.
58: 232(1900); Mallett in Gard. Chron. 39: 101(1906); Berger, l.c. 56 (1908). Type:
Hort. Leichtlin, from plants collected originally at Colesburg by Mr. Tuck (K!, holo.).
K. foliosa sensu Bak. in FI. Cap. 6: 279 (1896), non Hochst. K. pumila sensu Bak.,
l.c. 279 (1896), non (Ait.) Kunth. K. rivularis Berger, l.c. 55, t.21D, E (1908); Dyer
in FI. PI. S. Afr. 22: t.866 (1942). Type: Transvaal, Modderfontein, Conrath 643
(W, holo.f; K!). K. ensifolia var. albiflora E. A. Bruce in Bothalia 7: 231 (1951).
Type: Transvaal, 13 miles W. of Witbank, Codd & de Winter 3156 (PRE!, holo.).
In the typical subspecies the open flowers are white to greenish white while the
buds may or may not be pigmented with red to dull red. Flowering time is during
early summer from October to December and the plants are locally frequent in marshy
places and on stream banks, usually on black clay soil, from the relatively dry areas
of the northern Cape and western O.F.S., across the western and central Transvaal
to Lydenburg District in the east and northwards to the Waterberg District. Fig. 43.
Plate 9. Map 7.
Without Locality: ex Hort. Kew 1786 (BM).
Cape. — Barkly West: near Koopmansfontein, Leistner 948. Burghersdorp: Rev. T. W. Pocoek
86 (GRA); Colesberg: near Colesberg, Shaw s.n. (K); Bruce 220; between Colesberg and Steynsburg,
Marais 276. Kuruman: Kosifontein, Burchell 2554 (K). Phillipstown: near Petrusville, Van der
Walt s.n.; 16 miles S.W. of Phillipstown, Bruce 219a. Steynsburg: between Steynsburg and Venterstad,
Theron 889.
O.F.S. — Bloemfontein: Dewetsdorp road, Potts 1018 (UOFS); Fauresmith: near Groenvlei, Ver-
doorn 1035; 2176; near Veld Reserve, Bruce 217a, 217b; 16 miles S.E. of Fauresmith, Bruce 218;
24 miles S.W. of Fauresmith, Bruce 221.
Transvaal. — Belfast: near Draaikraal, Codd 8058. Bronkhorstspruit : 54 miles E. of Bronkhorst-
spruit, Codd 10077; farm Wag-’n-bietjiekop, Francis 1. Johannesburg: Bezuidenhout’s Valley, Rand
1091 (BM); near Zoo Lake, Moss 17750 (WU); Parkwood, Moss 17815 (WU); Wattles, Moss 13700
(WU); Florida, Moss 10020 (WU); Frankenwald, Gilliland in WU 26173; Jackson's Drift, Gilliland
in WU 26225; Modderfontein, Conrath 643 (K). Krugersdorp: King’s Kloof, Moss 10506 (WU).
Lydenburg: Sekukuniland ; farm Schoonoord, Barnard 163. Marico: Matebe River, Holub 1530
Plate 9 .—Kniphojia ensifolia Bak. subsp. ensifolia
441
(K); Linokana, Bruce 23 1 ; 462; Eersteling, Leendertz 4043. Middelburg: 40 miles N. of Middelburg:
Rogers 24829; Olifants River, Reynolds 2142: Slaghoek, Prosser 1878; 2 miles E. of Middelburg,
Codd 5161. Potgietersrus: Klein Magalakwin, near Koenap, Van der Merwe 1743. Pretoria: Rehmann
4769 (Z); hort. Kew ex Pretoria, Loveday s.n. (K); Aapies River, Leendertz 5710; near Boys’ High
School, Goossens 33; Forssman s.n.; Rietvlei, Dyer & Verdoorn s.n.; 14 miles S.E. of Pretoria, Codd
<6 Bruce 3519; Bruce 24; 25; 26; 27; 28: 35; 229; near Irene, C. A. Smith 5164; 25 miles N.E.
of Pretoria, Codd 8025. Rustenburg: Swartruggens, Sutton 937; Bruce 230; Kroondal, Von Wolff
in TRV 34695. Ventersdorp: Louw 775; 1615. Waterberg: Warmbaths, Leendertz 5695; Krantzberg,
farm Groothoek, Van der Merwe 2022; Codd 8020; Nylstroom, Van der Merwe 2003 ; 10 miles N.W.
of Nylstroom, Erens 2249; 16 miles N.W. of Nylstroom, Codd 6153; Sandrivierspoort, Codd 8833;
Palala Heights, Erens 1941. Witbank: Gilfillan sub Galpin 7259; 13 miles W. of Witbank, Codd
& De Winter 3156; Codd All A.
Fig. 43. — K. ensifolia subsp. ensifolia, from Middelburg, Transvaal ( Reynolds 5765).
442
The earliest specimen seen is in BM herbarium, “ex Hort. Kew 1786 ”, said to
be introduced by Francis Masson. Subsequently the species was collected by Burchell
at Kosifontein and Dr. Shaw at Colesberg, but Baker wrongly identified these gatherings
as K. pumila , which has been shown to be an Ethiopian species (Codd in J. S. Afr.
Bot. 29: 145, 1963). This led to a search for K. pumila at Colesberg by Mr. Tuck,
at the instigation of Max Leichtlin, and the resulting plant was, in fact, the same species
as the Shaw plant, but was described as a new species, K. tuckii. In the meantime
K. ensifolia had been described, based on a plant collected by Dr. Holub on the Matebe
River, some miles west of Zeerust, and this is the oldest name for the species.
Fig. 44. — K. ensifolia subsp. autumnalis, from Harrismith District (Jacobs: s.n.).
443
Consideration was given to separating K. tuckii as a distinct variety. In its typical
form around Colesberg and western O.F.S., the leaves are relatively short and markedly
glaucous, while the inflorescences are distinctly bright red in the upper half (bud stage).
However, there are many intermediates linking this form with typical K. ensifolia.
K. rivularis Berger is described as having smooth-margined leaves. Such specimens
are found in K. ensifolia , but are rare. It has even been found that on a single plant
the margins of the outer leaves are smooth and the inner leaves serrulate. In all other
characters the type of K. rivularis does not deviate from typical K. ensifoPa.
K. ensifolia var. albiflora E. A. Bruce is merely a colour form in which the buds
completely lack red pigment and, as this is the only difference, it is not worth separate
taxonomic rank. This form is commonly encountered in the eastern part of the distri-
bution range, for example in the Witbank and Middelburg Districts.
(b) subsp. autumnalis Cockl, subsp. nov., a subsp. ensifolia plantis autumno floren-
tibus, floribus luteis differt.
Type: O.F.S., Harrismith District, farm Rensburgskop, February, 1967, Jacobsz
s.n. (PRE!, holo.).
In this subspecies there is usually (but not always) yellow pigment in the buds
and flowers which, when present, produces more colourful inflorescences than in the
typical subspecies, while the flowering time is in autumn, during February and March.
Recorded only from the Harrismith District in the eastern Free State, where it occurs
on black clay soil on stream banks and low-lying areas. Fig. 44. Map 7.
O.F.S. — Harris/with: 15 miles S.E. of Harrismith, Codd 8510; near Swinburne, Codd 10520; 10521;
farm Rensburgskop, Jacobs z s.n.; 306: Codd 10523.
This subspecies flowers at the same time as K. linearifolia near Swinburne in the
Harrismith District and, when they occur in close proximity to one another, occasional
hybrids between the two are found (see p. 393).
21. Kniphofia splendida E. A. Bruce in Flow. PI. Afr. 30: 1. 1 183 (1955); Letty,
Wild Flows. Tvaal. 33, t.20, 1 (1962). Type: Transvaal, Magoebaskloof (cultivated
in Pretoria), Groeneveld in PRE 28420 (PRE!, holo.).
Plants in groups. Leaves 12-20 per peduncle, fairly rigid, distinctly nervose when
dry, generally more or less recurved, dull green to slightly glaucous, 80 cm to 2 m
long and usually 1-2 cm broad (rarely to 3 cm); leaf surface more or less flat, keeled
below and channelled above, or folded along the midrib, breaking into numerous
fibres at the base; keel and margin finely and somewhat irregularly serrulate, rarely
almost smooth. Peduncle 1 -2-2-5 m long, overtopping the recurved leaves, usually
with several triangular to linear-caudate sterile bracts below the inflorescence. Inflore-
scence very dense, subcylindrical, 10-22 cm long, 5-6 cm in diameter, usually tapering
towards the apex; buds ascending to spreading, yellow to yellow-green, usually tipped
with orange-red to scarlet; flowers pendulous, lemon-yellow to yellow. Pedicels
1-5-4 mm long. Bracts ovate-lanceolate to lanceolate, 5-9 mm long, 2 mm broad
at the base, acute; margin slightly to distinctly serrulate. Perianth narrowly funnel-
shaped, 1 -9-2-5 cm long, slightly contricted and 1-5-2 mm in diameter above the
ovary, expanding about the middle to 4-6 mm and about 5-7 mm near the throat.
Stamens well exserted by 5-10 mm at anthesis. Style subequal to the stamens at
anthesis, eventually exserted by 10-12 mm. Fruit subglobose, 5-6-5 mm long. Fig. 45,
Plate 10. Map 7.
444
Fig. 45. — K. splendida, near Lochiel, eastern Transvaal ( Bruce 303).
Distributed from Swaziland, through eastern Transvaal to the mountains of eastern
Rhodesia and Malawi, growing among rocks or on grassy slopes, often adjacent to
forest. The main flowering season is February-March.
Swaziland. — 4 miles N. of Forbes Reef, Reynolds 3483; Bruce 273; Hlatikulu, Compton 29258.
Transvaal. — Carolina: 4 miles E. of Lochiel, Bruce 303; Nelshoogte, near the Carolina/Barberton
boundary, Codd 10336. Letaba: Westfalia Estate, Scheepers 626; Grootbos Forest Reserve, Scheepers
925; Magoebaskloof, Van der Schijff 4410; cultivated in Pretoria, Groeneveld in PRE 28420. Peters-
burg: 2 miles S. of Houtbosch. Bruce 413; Haenertsburg, Thompson s.n. Pilgrims Rest: near Sabie
Van der Merwe 1290; Bruce 315. Soutpansberg: near Louis Trichardt, Meeuse 10151.
Rhodesia. — Melsetter: Martin Forest Reserve, Crook 546; Farm Rocklands, Crook 547; between
Melsetter and Cashel, Pole Evans 5700; Chimanimani Mts., Plowes 2171.
Malawi. — Zomba Plateau, Brass 16323.
A robust species, allied to K. ensifolia but with somewhat narrower and more
fibrous leaves, longer and more colourful flowers. It flowers in autumn, whereas
K. ensifolia flowers mainly in early summer, and it occurs along the eastern escarpment,
so that the two do not overlap in distribution. The first known gathering of the species
was by Dr. F. Z. van der Merwe near Sabie in March 1937.
22. Kniphofia praecox Bak. in Saund., Ref. Bot. 1. 1 69 (1870). Type: Hort. Saunders
(K!, holo.).
Plants robust, in groups. Leaves more or less in four ranks, 12 or more per
peduncle, fairly rigid, distinctly nervose when dry, erect to more or less spreading,
dark green, 90 cm-2 m long and 2-4 cm broad, deeply keeled below and channelled
above; keel and margin serrulate to obscurely serrulate, rarely almost smooth.
Peduncle 1-2-2 m long, well overtopping the recurved leaves, with several triangular
CylhhCL. UITy
Plate 10 .—Kniphofia splendida E. A. Bruce
445
sterile bracts below the inflorescence. Inflorescence dense to very dense, subcylindrical,
12-30 cm long, 6-7 cm in diameter, tapering somewhat towards the apex; buds
spreading, orange to reddish orange or scarlet; flowers spreading to pendulous, pale
yellow, often brown-tipped especially when dry. Pedicels distinct, 4-5 (-8) mm long,
elongating to 8-10 mm in fruit. Bracts lanceolate to linear-lanceolate, 8-12 mm long,
1-5-2 mm broad at the base, tapering gradually to the small rounded apex; margin
finely eroso-denticulate. Perianth subcylindrical to narrowly funnel-shaped, 2-4-3 -4
cm long. Stamens exserted by 4-15 mm at anthesis. Style subequal to the stamens
at anthesis, eventually exserted by about 16 mm. Fruit subglobose, erect, 5-7 mm
long.
K. praecox is a robust species with a long history of cultivation in Europe. It
has been the subject of considerable misinterpretation due to confusion with K. uvaria
and K. linearifolia, which some forms of K. praecox superficially resemble in the large,
showy inflorescences. There is evidence that it has hybridized with other species in
cultivation to produce many of the spectacular Redhot Pokers of gardens; in fact,
as indicated below, the possibility cannot be excluded that the type specimen itself
may be of hybrid origin.
It is clear that the chief monographers of the genus, Baker (1896) and Berger
(1908), included within one circumscription (under “ K. alodides ” by Baker and under
“ K. uvaria ” by Berger) elements which are now separated into two species: K. praecox
Bak. and K. linearifolia Bak. The separation is made on the basis of the long, acuminate
bracts and longish pedicels of K. praecox and the oblong, acute to obtuse bracts of
K. linearifolia and allied species. In addition, two subspecies within K. praecox are
recognized, as indicated in the key below which is based on plants occurring naturally
in two restricted, but separate, areas in the Cape Province. Of these, subsp. hruceae,
known from Komga and King William’s Town Districts, shows a close affinity to
K. splendida E. A. Bruce (p. 442) of the eastern Transvaal, and flowers in April-May,
but it cannot always be readily separated in the herbarium from some specimens of
subsp. praecox from the Knysna-Uniondale area, which flower in December-January.
The latter are clearly related to the type of K. praecox, a specimen preserved in
Kew Herbarium, cultivated by Mr. W. W. Saunders of Reigate and reputed to have
been collected by Mr. Thomas Cooper in South Africa. Unfortunately, this cannot
be verified because, although Cooper made some herbarium specimens in the field,
an original Kniphofia specimen matching the cultivated type of K. praecox in its narrow,
acuminate bracts has not been seen. The possibility exists, therefore, that the type
plant of K. praecox was not introduced by Cooper, but was a garden plant which
Saunders had obtained from some other source. In fact, the type specimen does not
match exactly any plants now known from the wild state, but it resembles plants
known in Europe for many years (see p. 448) and which were present in England at
least by 1852 (see cited specimen ex Herb. Thomas Moore).
Baker’s description of K. praecox was somewhat tentative and he subsequently
did not uphold the species stating, in Flora Capensis 6: 283 (1896): “ K. praecox . . .
is a plant imported by Mr. T. Cooper, that on one occasion flowered in summer,
and at other times in autumn ”, The question may be asked whether some confusion
did not perhaps take place in the records maintained by Mr. Saunders and, although
one may wish that adequate grounds could be advanced for discarding the name
K. praecox altogether, such a course does not seem feasible.
Key to Subspecies
Perianth tube 25-35 mm long, expanding more or less gradually from base to apex; stamens
exserted by up to 6 mm; buds red, flowers red to yellow: flowering time December-January
(a) subsp. praecox
Perianth tube 23-26 mm long, expanding abruptly about 1 cm from the base; stamens exserted
by up to 15 mm; buds orange, flowers pale yellow; flowering time April-May (b) subsp . bruceae
446
pIG. 46. K. praecox subsp. praecox, illustration of type plant in Saunders’ Ref. Bot. t.169 (1870.)
447
(a) subsp. praecox
K. praecox Bak. in Saund., Ref. Bot. 1. 1 69 (1870). Type: Hort. Saunders (K, holo.!).
K. alodides sensu Bak. in FI. Cap. 6: 283 (1896), partly. -var. nobilis Bak., l.c.
283 (1896). Type: a specimen in K (!) labelled: “ Hort. Kew as K. nobilis, Sept. 14,
1883 ” (see discussion on p. 449). K. uvaria var. praecox (Bak.) Berger in Pflanzenr. 4,
38: 62 (1908). — — var. nobilis (Bak.) Berger, l.c. 62 (1908).
The naturally occurring plants now included in this subspecies consist of a few
gatherings from the Uniondale, George and Knysna Districts of the Cape Province,
where they occur on stream banks and low-lying places. The flowering time is
November-January. A few cultivated specimens are listed together with the type on
the basis of the characteristic narrow, acute to acuminate bracts, while the flowering
time is given where it is known. A further discussion of the plants in cultivation follows
after the citations. Fig. 46, 47. Map 8.
Fig. 47. — K. praecox subsp. praecox, garden escape near Pretoria (Code! 9965).
448
Cape. — George: Wilderness, Mogg s.n. Knysna: 6 miles E. of Knysna, Codd 9922; 16 miles E,
of Plettenberg Bay, Marais 577. Uniondale: Saptou, Bo-Kouga Valley, Van Breda & Admiraal 2174.
Cultivated. — Pretoria: Union Buildings, Nov. 1936, Dyer s.n.; Division of Botany Jan. 1950,
Bruce 148A; Botanic Garden, Nov. 1951, Van der Elide s.n.; garden escape, 10 miles S.E. of Pretoria,
on stream bank, Dec. 1959, Codd 9965. Great Britain: Hort. Saunders , type of Ref. Bot.t.169 (K);
Hort. Chelsea, Jul. 1852, Herb. Thomas Moore dO: Hort. T. Cooper. Sept. 1877, Cooper 3629 (K).
Cultivated specimens cited above as subsp. praecox may superficially resemble
some forms of K. linearifolia, suggesting that grounds for separating two species on the
shape of the bracts may seem inadequate. It must be admitted that this is one of the
most difficult problems in the genus. The decision to regard bract shape as being
important in this case is consistent with the separation, for example, of K. ichopensis
and K. laxiflora and is prompted largely by the obvious differences between some of
the wild specimens now included in subsp. praecox (e.g. Van Breda & Admiraal 2174)
and those placed in K. linearifolia (very short pedicels and acute to obtuse bracts).
Although K. linearifolia is very much more widespread in its natural distribution,
K. praecox appears to have been more widely cultivated in Europe, as indicated by the
following specimens, which are of considerable historical interest, and which are
included in K. praecox on grounds of the longish pedicels and narrow, acuminate
bracts. Unfortunately, the older specimens lack information on date and place of
cultivation, but the species must have been introduced to Europe before 1800. As
these specimens are not referred to by the chief monographers of the genus. Baker
and Berger, several of them are listed in some detail below.
Herbarium of the Conservatoire et Jardin Botaniques, Geneva (G). Four sheets
seen, as follows:
1. Herb. Ventenat, named Veltheimia uvaria , with no collector or date.
2. Herb. Delessert, in Herb. DC., labelled Aletris uvaria L. with no collector or
date.
3. A specimen labelled Aletris sarmentosa, without any information on origin or
date.
4. Herb. Daniel De la Roche, “ fait au Jard. de Leyde sous van Royen. acq. en
1829 ”. Labelled Aloe uvaria , without collector or date.
Herbarium of the British Museum ( Natural History) (BM). Three sheets seen;
1. A sheet with two inflorescences and two leaves, with the name Aletris uvaria
on the front of the sheet and, on the back, (i) “ Russia, Dr. Thornton ” and (ii) “ Herb.
Pallas ”.
Mai* 7. — Distribution of Kniphofia ensifolia
subsp. ensifolia, subsp. autumnalis and
K. spelndida.
Map 8. — Distribution of Kniphofia tabularis,
K. praecox subsp. praecox, subsp. hruceae
and K. sarmentosa.
449
2. A sheet with an inflorescence and leaf, with no label on the front but, on the
back: “ Hort. 1782 ”.
3. A sheet with an inflorescence only, labelled on the front Aletris uvaria and, on
the back: “Herb. Pallas”.
It may be noted that Peter Simon Pallas (1741-1811) lived in Leningrad and his
collection was purchased by the British Museum in 1842.
Rijksherbarium, Leiden (L). One sheet seen with the name Tritomanthe uvaria
Link, without collector or date.
Royal Botanic Gardens, Kew (K). Several specimens seen, extending over the
period from 1852 to 1952. Some of these are not quite typical, suggesting that they
may be of hybrid origin. The following are worth noting:
1. A specimen with a label in N. E. Brown’s handwriting: “ Kniphofia alodides
Moench. Tritoma nobilis Guillon in Rev. Hort. 1882, p. 24: Hort. Kew (as K. nobilis)
Sept. 14, 1883. The flower stem of this specimen was seven feet high ”. This specimen
has flowers 3-5 cm long, but is otherwise quite a good match of K. praecox.
2. A specimen with two labels, one in N. E. Brown’s handwriting: “ From the garden
of Mr. Gumbleton . . . Ireland, 1884 ”. The other label, no doubt supplied with
the specimen, states: “ Kniphofia nobilis ex Horto Aureliae Aquensis. M.L. Baden-
Baden ”, i.e. from Max Leichtlin’s Nursery.
3. Three specimens collected in Kew Gardens during July, 1952, Bruce C, D and 1
(flowers 2-5 cm, 3-2 cm and 3 cm long, respectively), bear some resemblance to K.
praecox. Specimen D is labelled: “ Kniphofia sp. (hybrid), named K. nobilis in
gardens”; and specimen I: “Kniphofia sp., named K. uvaria var. nobilis” .
The repeated use of the epithet “ nobilis ” requires that it should be carefully
scrutinised. Tritoma nobilis Guillon is published in Rev. Hort. p. 24 (1882). The
description is a very general one and makes no reference to diagnostic characters, such
as length and shape of the perianth and floral bracts, nor does Guillon contrast it with
a known species. In fact, he appears doubtful whether it is a good species. A few
years later an illustration of Tritoma nobilis was published by Carriere in Rev. Hort.
p. 252 (1885), accompanied by a description similar to Guillon’s, which adds nothing
of diagnostic value.
Herbarium specimens of the actual plants dealt with by Guillon and Carriere
would assist in establishing the identity of their “ T. nobilis ”, but no such specimens
have been found in the Herbarium of the Museum National d’ Histoire Naturelle,
Paris, nor in Herb, de Vilmorin. There is thus justification for discarding the species
name used by Guillon (1882) and Carriere (1885) as being of uncertain application.
A clue to its identity is provided by the two contemporary specimens in Kew Herbarium
listed as 1 and 2 above. An annotation on the former specimen in Baker’s writing
may be accepted as typifying K. alodides var. nobilis Bak. in FI. Cap. 6: 283 (1896).
It is now included in K. praecox subsp. praecox.
( b ) subsp. bruceae Codd, subsp. nov., a subsp. praecoci floribus brevioribus,
staminibus circa 15 mm exsertis differt.
Veltheima uvaria sensu Jacq., Fragm. t.4 (1800/01).
Type: Cape, plant collected on Bedeford Farm, near Komga and flowered in
Pretoria, 7.4.1955, Bruce 604 (PRE, holo.!).
450
Fig. 48. K. praecox subsp. bmceae, inflorescence, reduced (Bruce 604).
illustration in Fig. 9.
Compare this with the Jacquin
451
A I Fig. 49. — K. praecox subsp.
T ; bnteeae, individual flowers,
life size (Bruce 604).
Fig. 50. — K. praecox subsp.
bruceae, from Komga
District (Bruce 605)
452
As indicated in the key on p. 445, the perianth tube in this subspecies is 23-26
mm long and it expands abruptly about 1 cm from the base, giving it a narrowly funnel-
shaped appearance, while the anthers are exserted by up to 15 mm at anthesis Fig*
48, 49, 50. Map 8.
its distribution appears to be restricted to small areas in the Komga and King
William’s Town Districts, where it occurs in grassy valleys. The flowers appear in
April and May.
Cape.— King William’s Town: near Kei Road, Comins 1552. Komga: near Komga, May 1894,
Flanagan 2323: Bedeford Farm, Bruce 604.
The illustration in Jacquin’s Fragmenta t.4, published in 1800 or 1801, was
confused in the past with K. uvaria, from which it differs strikingly in the well-exserted
stamens, the shape of the perianth and the narrow, acuminate bracts. It is a perfect
match of the modern gatherings cited above. Although the origin of Jacquin’s plant
is not recorded, there is a strong possibility that it was sent to Vienna by the collectors
Boos and Scholl, who penetrated well into the eastern Cape Province in search of
interesting plants for the Emperor’s garden at Schonbrunn. The subspecies was first
recorded in the wild state by Flanagan, who collected it near Komga in May, 1894.
Miss E. A. Bruce (see Fig. 51), who recognised it as distinct, made a special effort
to locate plants and eventually succeeded in 1954, with the help of Miss Courtenay-
Latimer and Mr. G. G. Smith of East London.
Plate 11. — Kniphofia sarmentosa (Andr.) Kunth
453
23. Kniphofia sarmentosa ( Andr .) Kunth, Enum. PI. 4: 552 (1843); Bak. in J.
Linn. Soc. 11: 362 (1871); J. Bot. Lond. 23: 279 (1885); FI. Cap. 6: 282 (1896);
Berger in Pflanzenr. 4, 38: 62 (1908), partly, excl. Rehmann 4769; Codd in Flow.
PI. Afr. 34: 1. 1 327 (1960). Type: the illustration in Andrews’ Bot. Rep. t.54 (1797),
of a plant said to be introduced from the Cape about 1789; collector not recorded.
Aletris sarmentosa Andr., Bot. Rep. t.54 (1799).
Veltheima media Donn, Hort. Cantab, ed. 2: 131 (1800), nom. illegit. Type: as
for A. sarmentosa. V. repens Ker, Recens. PI. 18 (1801), nom. illegit. Type: as for
A. sarmentosa. V. sarmentosa (Andr.) Willd., Enum. Hort. Berol. 380 (1809).
Tritoma media (Donn) Ker-Gawl. in Bot. Mag. t.744 (1804); Red., Liliac. 1. 1 6 1
(1807); Ait. f., Hort. Kew. ed. 2, 2: 290 (1810), nom. illegit.
Tritomanthe media (Donn) Link, Enum. Hort. Berol. 1: 333 (1821); Roem. &
Schult., Syst. Veg. 7: 630 (1829), nom. illegit.
Plants with short rhizomes, forming small groups of stems. Leaves 6-8 per
peduncle, erect to recurved, glaucous, soft and flaccid in texture, linear-ensiform.
30-65 cm long, 0-8-3 cm broad, V-shaped in cross section; keel and margin smooth.
Peduncle stout, 30-60 cm long, overtopping the recurved leaves, with several oblong-
lanceolate sterile bracts below the inflorescence. Inflorescence pyramidal, ovoid or
subcylindrical, tapering to the apex, at first dense but often elongating considerably
and becoming laxer, 8-30 cm long, 5-6-5 cm in diameter; buds at first erect, soon
spreading, coral to coral-scarlet with a greyish bloom; flowers pendulous, salmon to
creamy-buff. Bracts long, erect in the bud stage and covering the buds, later deflexed,
lanceolate to oblong-lanceolate, acuminate, 11-15 mm long and 2-2-5 mm broad at
the base, chartaceous; margin entire. Pedicels 1-3 mm long. Perianth subcylindrical,
not constricted above the ovary, 2-2-5 cm long, 2-3 mm broad near the base, widening
gradually to 5 mm broad at the middle and at the mouth: lobes ovate, rounded, 2 mm
long, not spreading. Stamens exserted by 2-5 mm at anthesis, later withdrawn. Style
subequal to the stamens at anthesis, eventually exserted by 6-8 mm. Fruit ovoid-
triquetrous, 7-8 mm long. Plate 11. Map 8.
Found beside mountain streams in the Sutherland (Roggeveld), Calvinia, Worcester
and Caledon Districts. Flowering has been recorded in nature from June to October.
Cultivated: without locality or date (G, L); cult. France, 1815, Herb. J. Gay (K); cult. Luxemburg,
1818, Herb. J. Gay (K); cult. Cobham, Surrey, 1833, Herb. Forbes Young (K); Hort. Kew, 1882 (K);
Hort. Kew, 1883 (K); cult. Ledbury by W. J. Grant, 1887 (K).
Cape. — Calvinia: Augustfontein Mt., Acocks 18982. Sutherland: Klein Roggeveld, Marloth 9609;
Compton 22239 (NBG); between Matjiesfontein and Sutherland, Comins 1100: 1101; Acocks 16843.
Worcester: top of Hex River Pass, Hardy 47; Matroosberg, Van Breda s.n.; between Osplaats and
Tunnel Siding, Rogers 30226 (Z).
A fairly distinctive species whose closest affinity is with the Transvaal species,
K. coralligemma E. A. Bruce. It differs from that species (see below) in the longer
bracts (which are probably the longest in the genus), the somewhat shorter and more
glaucous leaves which are never serrulate, in the flowering time, and in flower colour.
When Andrews described the species in 1799, he recorded that it was introduced
from the Cape of Good Hope in 1789, but made no mention of the original collector.
One is inclined to attribute the introduction to Francis Masson, who was collecting
in South Africa at the time. The species was not again collected until over 130 years
later, when Marloth rediscovered it in the Roggeveld in 1920. This gathering was
the first indication of the natural habitat of the species.
24. Kniphofia coralligemma E. A. Bruce in Flow. PI. Afr. 30: 1. 1 1 86 (1955);
Letty, Wild Flows. Tvaal. 36, t.20, 2 (1962). Type: Waterberg District, Kransberg,
farm Groothoek, Codd & Erens 2090 (PRE! holo.).
454
Plants solitary or in small groups. Leaves 8-12 per peduncle, at first erect, later
bending over above the middle, dull mid-green to slightly glaucous, 55-110 cm long,
1-2 cm broad, soft in texture, channelled above and keeled below, not conspicuously
fibrous at the base; margin distantly serrulate towards the apex, generally smooth
in the lower part and on the keel. Peduncle 60-120 cm long, overtopping the reflexed
leaves, with several lanceolate sterile bracts below the inflorescence, inflorescence
ovoid-cylindrical, usually tapering slightly towards the apex, 8-15 cm long, 4-5 cm
in diameter, dense, or sometimes rather lax towards the apex; buds at first erect, later
spreading, coral to pale orange or orange-scarlet; flowers spreading, varying in colour
according to colour form (see notes) from cream or ivory to creamy-yellow or orange-
yellow. Pedicels 2-5 mm long. Bracts erect and covering the buds in the early bud
stage, later deflexed, oblong-lanceolate to narrowly lanceolate, acute to acuminate,
7-12 mm long, 1-5-2 mm broad at the base; margins generally minutely serrulate
especially towards the apex or, occasionally, entire. Perianth subcylindrical, not
constricted above the ovary, 2 -2-2 -5 cm long, 2 mm in diameter at the base, widening
slightly to 3 mm near the mouth and with a slight constriction at the base of the lobes;
lobes oblong-ovate, about 3 mm long, slightly spreading. Stamens exserted by about
2 mm at anthesis, later withdrawn. Style subequal to the stamens at anthesis and
eventually exserted by 5 mm. Fruit ovoid-triquetrous, 5 mm long. Fig. 52. Plate 12.
Map 1.
Found in marshy places, forest margins and grassy, south-facing slopes at altitudes
of 5,000 to 6,000 ft. in the following mountain areas of the Transvaal: Kransberg
(Waterberg), Blouberg, Soutpansberg, Wolkberg and Haenertsburg. Flowering has
been recorded from the end of February to early April.
Transvaal. — Letaba: The Downs, Thompson s.n. Pietersburg: Iron Crown Mt., Meeuse 9856;
Haenertsburg, Bruce 195; Thompson s.n.; Magoebaskloof, Cunliffe s.n.; Mogg 20301; Hodge s.n.;
near Houtbosch, Bruce 414; Wolkberg, Van Biljon s.n.; Blouberg, Codd 8764 (cult, in Pretoria);
Esterhuysen 21474. Soutpansberg: near Louis Trichardt, Koker 13; Hangklip, Govt. Forester s.n.;
Meeuse 10177; near Hangklip, Meeuse 10192; Pisangkop, Meeuse 10235. Waterberg: Kransberg,
farm Groothoek, Codd & Erens 2090; 3966; Bruce 408; Story 6518; Meeuse & Strey s.n.
Three forms, differing mainly in colour of the inflorescence, are included in the
above citations. In the typical form, known only from a small area on the western
extremity of the Waterberg (Kransberg), the inflorescence is coral-pink at the apex,
shading to creamy-white or ivory as the flowers open. The second form, represented
by plants from the Soutpansberg and Haenertsburg areas, on the other hand, have
orange-scarlet buds which change to orange-yellow as the flowers open, while the
leaves tend to be somewhat longer and broader. Miss Bruce referred tentatively to
this form as K. pulchra, but the differences are inadequate for separating it as a distinct
species and this name was never validly published. The difference in inflorescence
colour, though striking, is not of great significance but, combined as it is with a difference
in distribution, there may be support for an argument in favour of giving this form
the status of a variety.
The third form, from the Blouberg, requires further study in the field. It is known at
present mainly from cultivated plants of one gathering, from which its taxonomic status
cannot easily be assessed. These cultivated plants show the following deviations
from the typical form: the buds are orange-yellow, becoming pale yellow as the flowers
open; the leaves are narrower (6-10 mm); the peduncle is shorter (20-50 cm); and
the perianth is shorter (1 -9-2-3 cm). Experience has shown that cultivated plants,
especially those from mountain areas, are frequently abnormal when grown in the
hotter and drier climate of Pretoria; stature and perianth length, in particular, tend
to be below normal. For this reason, the measurements given above for this form
are not included in the description of the species. Flowering specimens collected in
the natural habitat are required before a decision can be reached regarding the status
of this form.
Plate 12. — Kniphofia coralligemma E. A. Bruce
\
455
Fig. 52. — K. coralligemma at the type locality, Kransberg, Waterberg ( Bruce 408).
The Soutpansberg-Haenertsburg form is the most common of the three and was
collected several times before the Waterberg plant was discovered and described, the
earliest known herbarium specimen having been sent in by Mrs. Cunliffe from Magoe-
baskloof in April, 1928. Although the flame-coloured inflorescence of this form suggests
an affinity with the K. uvaria complex (and this no doubt explains why the species was
not described earlier), the long, narrow bracts and the soft-textured leaves indicate
the relationship to K. sarmentosa. For the differences between K. coralligemma and
K. sarmentosa , see the discussion under the latter species (p. 453).
25. Kniphofia caulescens Bak. in Bot. Mag. t. 5946 (1872); J. Bot. Lond. 23: 281
(1885); FI. Cap 6: 284 (1896); Mallett in Gard. Chron. 39: 82 (1906); Berger in
Pflanzenr. 4, 38: 65 (1908). Type: the Bot. Mag. plate, made from a plant cultivated
by Mr. W. W. Saunders of Reigate, originally collected by Thomas Cooper in the
Stormberg.
Tritoma caulescens (Bak.) Carr, in Rev. Hort. 132, t28 (1887).
K. tysonii sensu Phillips in Ann. S.A. Mus. 16: 294 (1917), as to Dieterlen 642.
24379—7
456
Fig. 53. — K. caulescens, inflorescence, slightly reduced, from Basutoland (Bruce 344),
457
Plants in groups, usually caulescent, gregarious, often forming large colonies.
Stems simple or sparingly branched, up to 60 cm tall, crowned with a large rosette of
leaves. Leaves many per peduncle, erect or bending over towards the apex, very
glaucous, somewhat fleshy in texture, outer leaves 45-70 cm long, 2-5-5 cm broad,
V-shaped in cross section, tapering rapidly (inner leaves much narrower), drying with
a purplish tinge towards the base, not breaking up into fibres; margin and keel distinctly
and finely serrulate. Peduncle stout, 30-60 cm long, overtopping or subequal to the
leaves, with several sterile bracts below the inflorescence. Inflorescence oblong to
subcylindrical, 9-30 cm long, 5-6 cm in diameter, tapering slightly towards the apex,
very dense; buds coral-pink to flame-coloured; flowers pale greenish-yellow to creamy-
yellow. Pedicels 3-5 mm long. Bracts linear-oblong to linear, acute to acuminate,
7-11 mm long, 1-5-2 mm broad, chartaceous, brownish; margin entire or sometimes
finely denticulate at the apex. Perianth subcylindrical, 2 -2-2 -4 cm long, slightly
constricted and about 1-5 mm in diameter above the ovary, widening gradually to
about 3-5 mm above the middle and to about 4 mm at the mouth; lobes ovate-oblong,
2 mm long, not or slightly spreading. Stamens markedly exserted by 8-13 mm at
anthesis. Style subequal to the stamens at anthesis, eventually exserted by about
1-5 cm. Fruit ovoid-triquetrous, erect, 5 mm long. Figs. 53, 54. Map 9.
Occurs in high mountain areas at altitudes of 6,000 to 10,000 ft. on peaty soil
overlying rock formations, marshy places and seepage areas in mountain grassland,
in north-eastern Cape Province, eastern O.F.S., Lesotho and the adjoining crest of
the Drakensberg in Natal. The main flowering time is January to March, though
flowering specimens have been collected in Natal in late November and December.
Cape. — Barkly East: Ben McDhui, Galpin 6865; Witteberg, Bruce 570; Lundeans Nek, Reynolds
578 (BOL); 3451: near Rhodes, Naude 6. Herschel: Majuba Nek, Hepburn s.n. (GRA); near
Lundeans Nek, Marais 1347. Lady Grey: Bruce 569. Sterkstroom: Stormberg, Andriesberg, Galpin
1929. Tarkastad: Great Winterberg, Story 3754; Bruce 525.
Lesotho. — Without locality. Staples 244. Likolobeng, Compton 21261 (NBG); Ntibokho Valley
Jacot Guillarmod 272; Butha Buthe, Jacot GuHlarmod 4350; Pulane, Bruce 366; near Mateka, Bruce
344; 379; Berea, Mamalapi, Marais 1309; Leribe, Dieterlen 642; near Qachas Nek, Galpin 14254;
Machaba Peak, Galpin 13667; Mokhotlong, Liebenberg 5760; summit of Drakensberg, Cleft Peak
area, Killick 1871; Killick & Marais 2179.
O.L.S. — Bethlehem: 24 miles N.E. of Clarens, Bruce 395; Golden Gate National Park, Scheepers
s.n. Harrismith: Drakensberg, Nelson's Kop, Medley Wood 6071 (NH); Rensburgskop, Jacobsz
286; 314.
Natal. — Bergville: adjoining Basutoland, Cuthbert sub Edwards 2292; Mont-aux-Sources, Schwei-
ckerdt 696; Cathedral Peak area, Schelpe 488 (NU). Underberg: Sani Pass, Killick & Vahrmeijer
3771.
The species is described in Bot. Mag. t.5946 (1872) and this plate may be regarded
as the type as no specimen of the figured plant appears to have been preserved. It
was originally collected by Thomas Cooper and cultivated by Mr. W. W. Saunders
of Reigate.
It is a very distinctive species and shows relatively little variation. One of the few
caulescent species, the stem may be simple or may branch dichotomously, with each
branch ending in a dense rosette of glaucous leaves which are markedly serrulate on
the margins. The inflorescence is very dense and is bicolorous, the upper half varying
from coral-pink to flame and the lower half from pale greenish-yellow to cream. The
plants tend to be gregarious and a colony in full flower is an impressive sight. In
herbarium specimens the leaf-bases dry with a characteristic purplish tinge not found
in any other species, while the densely placed, medium-length flowers with markedly
exserted stamens assist in the easy identification of this species.
K. caulescens is related to K. sarmentosa and K. ritualis but differs from both
in its caulescent habit, the denser and less tapering inflorescence, and the markedly
exserted stamens.
458
Fig. 54. — K. caulescens, Castle Buttress, Drakensberg. Photo by Dr. D. J. B. Killick.
The Basutos regard it as a charm against lightning and so, for this reason, it is
frequently cultivated near their huts.
26. Kniphofia ritualis Code! in Bothalia 9: 139 (1966). Type: Lesotho, Berea
District, Bitsolebe area, flowering 22.1.1957, Marais 1327 (PRE!, holo.).
K. sarmentosa sensu Phillips in Ann. S. Afr. Mus. 16: 294 (1917), as to Dieter/en
174; 174a.
Plants solitary or in small groups. Leaves in 3 or 4 ranks, 8-12 per peduncle,
at first erect, later bending at the middle or semi-spreading, soft in texture, glaucous,
40-90 cm long and 1 -2-2-4 cm broad near the base, deeply keeled; margin conspi-
cuously and regularly serrulate. Peduncle subequal to or shorter than the leaves,
40-80 cm long. Inflorescence rhomboid-cylindrical to ovoid, dense in the centre,
laxer at the apex, 9-14 cm long and 4-5 cm in diameter; buds and flowers pendulous;
buds coral or salmon to orange-red, becoming greenish-yellow as the flowers open.
Brads lanceolate to linear-lanceolate, tapering to the acute or acuminate apex, 1 0—12
mm long and 2 mm broad; margin usually minutely denticulate. Pedicels 1-2 mm long
at flowering stage, elongating to 5 mm in the fruiting stage. Perianth subcylindrical,
2-5-3 cm long, very slightly constricted above the ovary, 3 mm wide at the base,
increasing gradually to 5 mm wide at the mouth; lobes ovate, rounded, 3-3-5 mm
long, not spreading. Stamens 6, of two lengths; anthers not or scarcely exserted at
anthesis, later withdrawn into the perianth tube. Style equal to the stamens at anthesis,
later included or exserted by up to 3 mm. Fruit subglobose, not markedly triquetrous,
7-8 mm long. Fig. 55. Map 9.
Common in Lesotho and in the adjoining areas of the eastern Orange Free State,
mainly on sandstone formation, in shallow soil in rock crevices, in dry water-courses
and on grassy slopes, at altitudes of 5,500 to 9,000 ft., extending to the Bergville District
in Natal and with recent records from the Wolkberg in Transvaal. Some additional
specimens from the Natal Drakensberg region may also belong to this species. The
flowering time is from January to March.
459
Map 9. — Distribution of Kniphofia coralli-
gemma, K. caulescens and K. ritiialis.
Map 10. — Distribution of Kniphofia stricta and
K. northiae.
O.F.S. — Bethlehem: near Generaalskop, Liebenberg 6979; near Clarens, cult, in Pretoria, Marais
1287; Golden Gate National Park, Roberts 3430; 3452. Ficksburg: Farm Wintershoek, near
Kirklington, Codd 10540; 10541 ; Farm Franschoek, 20 miles N.E. of Ficksburg, Codd 10549. Fouries-
burg: 23 miles N.E. of Ficksburg, Bruce 343; 12 miles S.E. of Clarens, Bruce 393 ; near Golden Gate,
Codd 10537; near Fouriesburg, cult, in Pretoria, Marais 1288. Harrismith: Witzieshoek, Thode
5677 (STE); Rensburgskop, Jacobsz 322: 323. Senekal: Goossens 838. Thabanchu: Thabanchu
Mt., Roberts s.n.
Lesotho. — Berea: Below Molimo-Nthuse, Marais 1326; Bitsolebe area, Marais 1327; at foot of
Pulane Mt., Bruce 360; Pulane area, Bruce 364; Thaba Chitja, Bruce 365; Mamathes, Bruce 390.
Leribe: Dieterlen 174; 174a; Phillips 574 (SAM). Mafeteng: Malealea, and cult, in Pretoria, Munro
s.n. Maseru: Thaba Bosigo, Van dev Merwe 1 156; Jordan Valley, Jacot Guillarmod 1671 ; Senqunyane
Valley, Jacot Guillarmod 2256; Roma, Ruch 1722. Summit of Drakensberg in Cathedral Peak area,
Killick 1857; Killick & Marais 2174.
Natal. — Bergville: Cathedral Peak, Kromhout & Knuffel s.n.; Killick & Vahrmeijer 3547: 3550;
Cleft Peak, Edwards 1186; Champagne Castle, slopes of Grey’s Pass, Cuthbert sub Edwards 2290;
Organ Pipes Pass, Cuthbert sub Edwards 2297; Mont-aux-Sources, Trauseld 204. Underberg: Sani
Pass, Killick & Vahrmeijer 3759.
Transvaal. — Petersburg: Wolkberg, Van Biljon s.n.; Hardy 2330.
Phillips, l.c., misidentified this species as K. sarmentosa (Ait.) Kunth, a species
largely restricted to the Roggeveld area of the north-western Cape Province. He
records the Sotho common names as “ Leloele ” or “ Lelole ” and states that: “A
decoction prepared from this species is drunk to cure pains in the shoulders. Women,
during the times 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 information
could be obtained ”. It is on the basis of this information that the name K. ritualis
was chosen for this species in preference to a manuscript name “ K. basutica ” given
tentatively to the species by Miss E. A. Bruce.
K. sarmentosa and K. ritualis are similar in having conspicuously long narrow
bracts. The latter species can, however, readily be distinguished by its markedly
serrulate leaves and the longer perianth which is 2-5-3 cm long, as against 2-2-5 cm
long in K. sarmentosa.
The long bracts and serrulate leaves of K. ritualis are also reminiscent of K.
caulescens but, among the many characters that distinguish the two, are the caulescent
habit of the latter and its shorter perianth with markedly exserted stamens.
460
Fig. 55. — K. ritualis, Lesotho ( Marais 1327).
27. Kniphofia hirsuta Cocld in Bothalia 9: 140 (1966). Type: Lesotho, Berea
District, Mamalapi, alt. 8,500 ft., in flower 20.1.1957, Jacot Guillarmod & Marais
1307 (PRE!, holo.).
Plants usually solitary. Leaves in 3 or, rarely, 4 ranks, 8-12 per peduncle, erect,
soft in texture, dull green, 40-60 cm long and 1 -5-2-5 cm broad near the base, deeply
keeled, hirsute along the nerves on both surfaces; margin conspicuously and regularly
serrulate. Peduncle equal to or longer than the leaves, 45-60 cm long. Inflorescence
shortly rhomboid-cylindrical to ovoid, dense, 8-12 cm long and 4-4-5 cm in diameter;
buds and flowers pendulous; buds orange or dull coral to salmon-pink tipped with
green, becoming dull greenish-yellow as the flowers open. Bracts linear-lanceolate,
tapering to the acuminate apex, 10-14 mm long and 2 mm broad; margin minutely
denticulate to entire. Pedicels 2-3 mm long at flowering stage, elongating to 4 mm
in fruit. Perianth subcylindrical, 2 -2-2 -7 cm long, very slightly constricted above the
461
ovary, 3-5 mm wide at the base, increasing gradually to 5 mm wide at the mouth;
lobes ovate, rounded, 2-5 mm long, not spreading. Stamens not or scarcely exserted
at anthesis, later withdrawn into the perianth tube. Style equal to the stamens at
anthesis, later exserted by up to 3 mm. Fruit subglobose, not markedly triquetrous,
7-9 mm long.
Found in the mountains of Lesotho on grassy slopes and stream banks, at altitudes
of 8,000 to 9,000 ft. Flowering takes place in December and January.
Lesotho. — Berea: Mamalapi, December 1948, Compton 21266; Jacot Guillarmod 1238; Jacot
Guillarmod & Marais 1307; 1308.
K. liirsuta is closely allied to K. ritualis, its inflorescence, perianth and bract
characters being similar, but the former species is unique in the genus in having
pubescent leaves. The only other species which approaches it in this respect is K.
stricta Codd, which sometimes has a few scabrid hairs on the nerves. K. hirsuta was
first collected by Prof. R. H. Compton at Mamalapi at an altitude of about 8,500 ft.
Since then, good flowering and fruiting material has been collected by Mrs. A. Jacot
Guillarmod and Mr. W. Marais, who report that it tends to flower earlier than K.
ritualis and to grow at higher altitudes (though this is not strictly correct) where it is
associated with Danthonia-Erica-Cliffortia scrub of rocky slopes or streams.
28. K. stricta Codd in Flow. PI. Afr. 34: 1. 1 328 (1960). Type: Cape, Graaff
Reinet, Loskop near New Bethesda, Kitching 54 (PRE!, holo.).
Plants in groups. Leaves not clearly in ranks, 8-12 per peduncle, erect, rigid,
strongly nerved, dark green, 30-70 cm long and 0-8-1 -4 cm broad near the base,
U-shaped in cross section (not keeled), channelled above; margin serrulate; nerves
prominent on lower surface with occasional scabrid papillae on the nerves. Peduncle
subequal to the leaves, 30-60 cm long, bearing usually several large deltoid sterile
bracts below the inflorescence. Inflorescence subcylindrical to ovoid, dense, 8-14 cm
long and 5-6 cm broad; buds pendulous to spreading, flowers at first spreading but
soon becoming pendulous; buds orange or pale orange-red to coral, becoming yellow
to greenish-yellow as the flowers open. Bracts ovate to ovate-lanceolate, rounded
to acute at the apex, 7-10 mm long and 2-5-3 mm broad at the base; margin eroso-
denticulate. Pedicels 1-5-3 mm long at flowering stage, elongating to 6 mm long in
fruit. Perianth subcylindrical, slightly constricted above the ovary, 2-7-3 -3 cm long
and 3 mm wide at the base increasing gradually to 5-6 mm wide at the mouth; lobes
ovate, rounded, 3 mm long, slightly spreading. Stamens exserted at anthesis by 0-8
mm, usually withdrawn eventually into the perianth tube though sometimes remaining
exserted. Style subequal to the anthers at anthesis, finally exserted by 8-12 mm.
Fruit subglobose, not markedly triquetrous, 7-9 mm long. Fig. 56. Plate 13.
Map 10.
Found on grassy slopes, often among dolerite rocks, at altitudes of 5,500 to 8,000
ft., in the mountains of the north-eastern Cape Province and the adjoining area of
south-western Lesotho. Its flowering season is from mid-January to early March.
Cape. — Barkly East: Kraalberg, near Barkly Pass, Rattray sub Galpin 7326; 1 mile S. of Lundean's
Nek Post Office, Marais 1349; Naude’s Nek, Marais 1361; 1364; Godfrey VH 1526; Buttermead,
near Rhodes, Naude 2; Maartenshoek, Lavranos 1091. Graaff Reinet: Loskop, on Farm De Toom,
6 miles from New Bethesda on road to Graaff Reinet, cult, in Pretoria, Kitching 45 ; 54. Herschel :
4 miles N. of Lundean's Nek, Marais 1346. Lady Grey: 6 miles from Lady Grey on Joubert Pass
road, Bruce 567. Middelburg: Sneeuwberg, between Compassberg and Rhenosterberg, Drege s.n.
(K, partly). Sterkstroom: Halseton, Farm Buffelsfontein, Stretton 5. Wodehouse: Stormberg Mts.,
west of Dordrecht on road to Boshoff's Kraal, Bruce 563; near Dordrecht, Bruce 565.
Lesotho. — Berea: Jacot Guillarmod 1716; 1744; 1753. Mokhotlong: Magapung Valley, Coetzee
556. Qachas Nek: On top of Rafase Kop near Qachas Nek, Galpin 14083.
462
Fig. 56. — K. stricta, from New Bethesda ( Kitching 54).
In the herbarium, K. stricta might be confused with K. ritualis but the differences
are clear-cut. The leaves of K. stricta are U-shaped in cross section, not V-shaped
as in all other species in South Africa with the exception of K. northiae; when
herbarium specimens are prepared, however, the leaves are often folded in such a
manner that the characteristic shape of the cross section cannot be seen. When living
plants are available, no confusion will arise; the leaves of K. stricta are stiff and erect
with the margin tending to be inrolled, so that they appear to be terete at the apex.
While at first sight there does not appear to be a distinct mid-vein, there is always
one nerve more strongly developed than the others, but this nerve is not centrally
placed, being always nearer one margin than the other. Along the leaf margins and
on the nerves are minute papillae, making the leaf surface slightly scabrid to the touch.
In occasional specimens, for example Godfrey VH 1526 and Marais 1364, both from
Naude’s Nek, there are minute protuberances on the nerves which vary from scarcely
raised, discoloured spots to bristle-like papillae, similar to the serrulations on the
leaf margins.
Plate 13. — Kniphofia stricta Codd
)
463
In floral characters, K. stricta resembles K. ritualis but there is a constant difference
in the bracts, which are shorter and ovate to ovate-lanceolate in K. stricta , while in
K. ritualis they are lanceolate and long-acuminate. The two species were, at first,
confused in the herbarium until living plants were sent in by Mr. S. Kitching, when
it became obvious that two distinct species were involved.
Fig. 57. — K. northiae, from Cathcart,
\ (Mrs. Bowker s.n.).
464
The species was first discovered by Drege on the Sneeuwberg, but was not recog-
nized as distinct. A sheet at Kew shows a mixed gathering, consisting of an inflorescence
and leaf of K. stricta together with an inflorescence and leaf of K. linearifolia.
The leaves of typical K. northiae also lack a keel, but they are much broader
(3-12 cm) and are recurved, while the perianth is shorter (2-2-5 cm) and the anthers
are more conspicuously exserted than in K. stricta.
29. Kniphofia northiae Bak. in J. Bot. Lond. 27: 43 (1889); Gard. Chron. 10:
67 (1891); Bot. Mag. t.7412 (1895); FI. Cap. 6: 284 (1896); Mallett in Gard. Chron.
39: 100, t.42 (1906); Berger in Pflanzenr. 4, 38: 65 (1908). Type: A drawing by
Miss North of a plant from the mountains north of Grahamstown (in K); a plant
from this locality was also sent to Kew, where it was cultivated and figured in Bot.
Mag. l.c.
Fig. 58. K. northiae, from Cathcart (Mrs. Banker s.n.).
Plants solitary, caulescent; stems simple or, rarely, branching from the base,
20 cm to 1-7 m tall. Leaves in a rosette at the apex of the stem, broad, persistent,
broadly Y-shaped in cross section to non-keeled, leathery, recurved, shallowly and
465
broadly channelled above, 0-5 — 1-5 m long and (3) 3-5 — 12 cm broad, nervose; margin
strongly serrulate. Peduncle overtopping the recurved leaves, 20-30 cm long, 1-2-5
cm in diameter, bearing several to many large, oblong, empty bracts below the
inflorescence. Inflorescence cylindrical to ovoid, very dense, 10-22 cm long and 5-6 cm
in diameter; buds and flowers subspreading; buds pinkish red opening to whitish
flowers, or orange-red opening to yellow flowers. Bracts oblong-lanceolate to linear-
oblong, 7-12 mm long and 1-5-2 mm broad, acute to obtuse at the apex; margin
entire or minutely eroso-denticulate at the apex. Pedicels 2-5-3 mm long at flowering
stage, elongating to 7-10 mm in fruit. Perianth subcylindrical, 2-2-3 -2 cm long,
3 mm broad near the base, slightly constricted above the ovary, increasing to 5 mm
broad below the middle and then approximately parallel-sided to the mouth; lobes
ovate, 2 mm long, not spreading. Stamens markedly exserted by 1-1 -3 cm at anthesis.
Style subequal to the anthers at anthesis, eventually exserted by about 1 -5 cm. Fruit
subglobose, obscurely triquetrous, erect, 8 mm long. Figs. 57, 58, 59. Map 10.
Occurs in mountain areas at altitudes of 5,000 to 10,000 ft. in grass or on sparsely
wooded, stony slopes in the eastern Cape Province, the Midlands and Drakensberg
region of Natal, and eastern Lesotho, with outliers at lower altitudes near Mqanduli
in the Transkei and Pietermaritzburg in Natal. The main flowering time is December
to February, though cultivated plants from the two last-mentioned localities flowered
in September and October (see notes on variation below).
Cape. — Barkly East: Ben McDhui, Galpin 6866; near Rhodes Bruce 572. Cathcart: Hart in NBG
169/29 (BOL, NBG); Gaikas Kop, Bowker s.n. Mqanduli: 6 miles N. of Mqanduli, Marais 1014.
Queenstown: Katberg, Bruce 530; King s.n.; near Fairford, Bruce 539; Cotterell s.n. (GRA).
Stockenstroom : Top of Katberg Pass, Dyer 2307. Umzimkulu: near Kokstad, Mogg 1843.
Lesotho. — Summit of Drakensberg, Cleft Peak area, Killick & Marais 2175; Castle Buttress area
Killick 1872.
Natal. — Bergville: Cathedral Peak, Killick 3223; Mont-aux-Sources, Trauseld 174. Estcourt:
Cathkin Park, Galpin 11733; Tabamhlope, Pentz 34; West 709; Marais 1469; Giants Castle Game
Reserve, Trauseld 498: Highmoor Forest Station, Killick & Vahrmeijer 3640. Lions River: Spioen
Kop, 18 miles S. of Nottingham Road, N. R. Smuts 1022. Pietermaritzburg: near Pietermaritzburg
(locality uncertain), Beard s.n.
The original description of the species is based on a painting by Miss North
which may be accepted as the type. A living plant was also presented to Kew by Miss
North and this plant flowered and was illustrated in the Bot. Mag. t.74 1 2 (1895). This
shows the colour form found commonly in the eastern Cape Province, with pale red
buds changing to creamy-white flowers. There is another colour form, in which the
buds are orange-red to flame-red, changing to yellow as the flowers open, occurring
in Natal ( Beard s.n. and Trauseld 498) and near Mqanduli in the Transkei ( Marais
1014). Although the usual flowering time for the species is December to February
in the high moutain areas, two of these gatherings from lower altitudes ( Marais 1014
and Beard s.n.) flowered in cultivation in September and October.
In its typical form, K. northiae may be separated from all other species by its
broad, shallowly channelled, arched leaves, which lack a distinct keel. However,
specimens collected in Lesotho and the Bergville and Estcourt Districts ( Schelpe s.n.,
Killick 3223, West 709) may have narrower leaves, 3-4 cm wide, which are broadly
V-shaped in cross section with a distinct midrib. Further study is required to determine
whether the species should be subdivided into groups with separate taxonomic status
but, at the present state of our knowledge, it is considered best to leave the specimens
in one variable species.
The large, dense inflorescence resembles that of K. cualescens and. like that species,
K. northiae may also develop a distinct stem.
466
Fig. 59. — K. northiae, near Nottingham Road (N. R. Smuts 1022). Photo by Dr. N. R. Smuts.
30. Kniphofia galpinii Bak. in FI. Cap. 6: 281 (1896); Berger in Pflanzenreich 4,
38: 56 (1908); Phillips in Flow. PI. S. Afr. 20: t.783 (1940); Letty, Wild Flow. Tvaal.
36, t.20, 3 (1962). Type: Barberton District, Upper Moodies, Galpin 1208 (K!, holo,;
BOL!, PRE!).
Plants solitary or in small groups. Leaves not clearly in ranks, 8-12 per peduncle,
narrow, grass-like, 40-65 (-100) cm long, 3-6 (-8) mm broad, dull green, markedly
nervose, triangular in cross section, recurved, breaking into numerous fibres at the
base; margin entire or occasionally minutely denticulate towards the apex. Peduncle
standing above the recurved leaves, 30-60 cm long, bearing several sterile bracts below
the inflorescence. Inflorescence subglobose to ovoid, pyramidal, with a coma of sterile
bracts at the apex, dense, 5-8 cm long and 4 cm broad, buds and flowers pendulous;
buds flame-red to orange-red, flowers yellow to orange. Bracts linear-lanceolate,
Plate 14. — Kniphofia galpinii Bak.
467
Fig. 60. — K. galpinii, Barberton District (Bruce 304).
9-13 mm long, 1-5-2 mm broad, acuminate to subcaudate, especially the lowermost
which may be up to 2 cm long; margin subentire to minutely denticulate towards
the apex. Pedicels 1-5-3 mm long at flowering stage, elongating to 4 mm in fruit.
Perianth subcylindrical, 2 -7-3 -5 cm long, 3 mm broad at the base, not constricted
above the ovary, expanding gradually to 4 mm broad near the throat; lobes ovate,
obtuse, 2 mm long, not spreading. Stamens not or scarcely exserted at anthesis, later
withdrawn. Style subequal to the stamens at anthesis, finally included or shortly
exserted by up to 5 mm. Fruit not seen. Fig. 60. Plate 14. Map 11.
Grows in dense grass on hillsides and grassy vleis at altitudes of 3,000 to 6.000 ft.
Recorded from the south-eastern Transvaal, northern Swaziland, northern and central
468
Natal. The flowering period is mainly from January to March, though flowering
specimens have been collected as early as December.
Transvaal. — Barberton: Pott-Leendertz 5544; Thorncroft in TRV 5137; Upper Moodies, Galpin
1208; 6 miles S. of Barberton on road to Havelock Mine, Reynolds 4241; 20 miles S. of Barberton,
Reynolds 4138; near Angle Station, Bruce 311; 2 miles S.E. of Havelock Mine, Reynolds 5874; near
Transvaal border, Bayliss 2749; Sheya-lo-ngubo Dam, 10 miles S.W. of Louws Creek, Reynolds 5885;
Bruce 314. Nelspruit: near Kaapschehoop, Van der Merwe 1648; Reynolds 3462 (BOL, K); near
Montrose Falls, Key s.n.; Codd 4878. Pilgrim’s Rest: Mt. Anderson, Galpin 13717; Reynolds 5895;
Meeuse 10082; Peach Hill, Galpin 14501; Jubilee Peak, Galpin 14510; Mauchsberg, Van der Merwe
333; 1665; in PRE 2105; Brent 109; Long Tom Pass, Werdermann & Oberdieck 2145.
Swaziland. — 6 miles W. of Piggs Peak, Codd 9524; Piggs Peak, Emlembe, Comton 30528.
■Natal. — Dundee: Medley Wood 11628 (NH). Helpmekaar: between Helpmekaar and Vants
Drift, Codd 6799 ; 14 miles S.E. of Dundee on road to Rorkes Drift, Codd 6782. Vryheid: Hlobane
Mt., Johnstone 609.
Some thought was given to placing K. galpinii as another subspecies of K. triangu-
laris, but it is felt that the many small differences in a number of features justify retaining
it as a separate species. The narrow, grass-like leaves are very similar to those of K.
triangularis but are, in general, more nervose and fibrous. The floral characters are
also similar, though the inflorescence of K. galpinii is not concolorous (except in the
early flowering stage), the flame-red buds changing eventually to orange-yellow as the
flowers open; the perianth tends to be longer and the lobes are not spreading, while
the bracts are longer and more narrowly acuminate.
Four specimens (not included in the citations above) from Carolina and Ermelo
Districts need further investigation. Three of these are from Jessievale Plantation,
Carolina District, collected on 18th November, 1961, by Germishuizen and do not
entirely agree with K. galpinii. They have narrow, fibrous leaves 5-8 mm broad with
a few scattered teeth on the margins near the apex, and the bracts vary from narrowly
acuminate in the sheet designated “ c ”, to oblong-lanceolate, acute to obtuse in the
sheets labelled “ a ” and “ b The last two cannot be reconciled with K. galpinii,
which consistently has narrowly acuminate bracts. The fourth specimen of doubtful
identity was collected by Miss O’Connor on the Athole Pasture Research Station,
Ermelo District, in November, 1948, and has ovate-lanceolate, acute bracts. This
matches sheets “ a ” and “ b ” above. It may also be significant that all four specimens
flowered in November, earlier than is normal for K. galpinii. Although they are now
excluded from K. galpinii, these specimens are obviously closely related to that species,
but further information is required before their taxonomic status can be determined.
The specimens cited from Natal ( Medley Wood 11628, Codd 6782, 6799 and
Johnstone 609) also differ slightly from typical K. galpinii in their longer (70-100 cm),
broader (4-8 mm) and more glaucous leaves, but the differences do not appear sufficient
to warrant separate status. They represent a form which also requires further investi-
gation.
31. Kniphofia triangularis Kunth, Enum. PI. 4: 551 (1843). Type: “Aliwal.
District”, Wittebergen, Drege 3524 (K! lecto.; G!).
Plants solitary or in small groups. Leaves not clearly in ranks, recurved, soft
or somewhat fibrous, 28-60 cm long, 1-5-8 mm broad, triangular in cross section,
or more or less flat with a narrow channel above, keeled below, breaking up into
fibres at the base; margin entire to distinctly serrulate. Peduncle subequal to the
leaves, 30-60 cm long, bearing several sterile bracts below the inflorescence. Inflore-
scence subcylindrical to ovoid, pyramidal at the apex, dense, 5-8 cm long and 4-5 cm
broad; buds and flowers pendulous, more or less concolorous, coral-red, red-orange,
salmon-red to orange-yellow. Bracts ovate-lanceolate to lanceolate, acute to long-
acuminate, 6-8 mm long, 1-5-2 mm broad, entire or nearly so. Pedicels 1-2 mm long
at flowering stage, elongating to 3 mm long in fruit. Perianth cylindrical, 2-4-3 -5 cm
469
long, 3 mm broad, not constricted above the ovary, expanding abruptly at the mouth
to 5-8 mm broad; lobes ovate, 1 -5-2 mm long, usually somewhat spreading. Stamens
exserted by 2-3 mm at anthesis, later withdrawn. Style eventually exserted by 5-10
mm. Fruit ovoid, obscurely triquetrous, 6-8 mm long, erect.
Two subspecies are recognized.
Key to Subspecies
Leaves 1-5-3 mm broad or, if broader, leaf-margins sparingly to distinctly serrulate
(a) subsp. triangularis
Leaves 3-9 mm broad; leaf margins smooth (b) subsp. obtusiloba
(a) subsp. triangularis.
K. triangularis Kunth, Enum. PI. 4: 551 (1843); Bak. in J. Linn. Soc. 11: 362
(1871); J. Bot. Lond. 23 ; 278 (1885); FI. Cap. 6: 280 (1896); Berger in Pflanzenreich
4,38: 53 (1908); Bruce in Flow. PI. Afr. 30: 1. 1 184 (1955). Type: “Aliwal District ”,
Wittebergen, Drege 3524 (K, lecto!; G!). K. macowanii Bak. in J. Bot. Lond. 12: 3
(1874); Bot. Mag. t.6167 (1875); J. Bot. Lond. 23: 278 (1885); FI. Cap. 6: 280
(1896); Mallett in Gard. Chron. 39: 83, t . 38 (1906); Berger in Pflanzenreich 4, 38:
54, t.21 C (1908). Type: Somerset East, Boschberg, MacOwan 1536 (Kl, holo. ;
BOL!, G!, GRA!, PRE!, SAMI, Z !). K. nelsonii Mast, in Gard. Chron. 11: 554,
t.83 (1892); Bak. in FI. Cap. 6: 280 (1896); Mallett in Gard. Chron. 39: 82, t.37
(1906); Berger in Pflanzenreich 4. 38: 53 (1908); Phillips in Ann. S. Afr. Mus. 1:6
(294 (1917). Type: O.F.S., Nelson s.n. (Kl, holo.).
Tritoma macowanii (Bak.) Carr, in Rev. Hort. 390 (1879).
Included in subsp. triangularis are plants with leaves very narrow and grasslike,
1-5-3 mm in diameter, with smooth margins (typical), and those with leaves 2-5-6
(-8) mm broad and margins sparingly to distinctly serrulate, which were previously
separated as K. macowanii (see notes below). Fig. 61. Plate 15. Map 11.
Found in mountain grassland, often in peaty soil, on grassy slopes, stream-banks
and moist places among sandstone rocks at altitudes of 3,000 to 6,500 ft. in the eastern
Cape Province, eastern O.F.S., Lesotho and the southern and central Drakensberg
area of Natal. The flowering time is from January to April.
Cape. — Without exact locality: Winterhoek, Zeyher (SAM). Barkly East: Stretton 120; Witteberg-
Drege 3524 (K, G); Thode A. 520; Bruce 571; Barkly Pass, Reynolds 2506; Bittermead, Naude 1;
2 miles S. of Lundeans Nek, Marais 1350. Cathcart: Mrs. Bowker s.n.; near Hogsback, Rattray 32;
Story 3743; near Toise River Station, Flanagan 2250; near Hogsback, Reynolds 1176 (BOL). Keis-
kammahoek: Ghulu Kop, Dyer 397 (GRA); Cata Forest Station, Bruce 551. King William’s Town:
Mt. Kemp, B. H. Dodd sub Galpin 8033. Maclear: Naudes Nek, Bruce 573; Marais 1371. Matatiele:
Galpin 14084. Queenstown: Hangklip Mt., Galpin 1778; near Queenstown, Bruce 528. Somerset
East: Boschberg, MacOwan 1536; Bruce 520. Stockenstroom: Katberg Pass, Elfers 2; Bayliss
s.n.; 12 miles S.W. of Fairford, Reynolds 3794. Stutterheim: Dohne Hill, Sim 540. Tarkastad: 9
miles S. of Spring Valley, Bruce 523; Winterberg, Reynolds 1772. Umzimkulu: Ingeli Mt., Tyson
1380 (SAM).
Lesotho. — Leribe, Dieterlen 474 (K, BM); Phillips s.n. (SAM). Qachas Nek, Brooke 33 (BM).
O.F.S. — Bethlehem: Harding s.n.; Potgieter in TRV 21936; near Retiefsnek, 1 Vium 25; Golden
Gate National Park, Roberts 3387; 20 miles S. of Kestell, Codd 10536. Ficksburg: Gum Tree,
Martley in BOL 23830 (BOL). Fouriesburg: Brierley s.n. (BM); Key in PRE 28303: Wyndford,
Gemmell 7556. Harrismith: Witzieshoek, Junod s.n. (G); Besters Vlei, Bolus 8260 (BOL); 8261
(BOL); Flanagan 1838; 1841; Thode s.n. (STE); Junod s.n. (G); Zaaihoek, Thode s.n. (BOL, Z);
Van Reenens Pass, Pole Evans s.n.; Oliviershoek Pass, Bruce 401; Kerkenberg: Jacobs: 290; Codd
10528. Herschel: Hepburn s.n. (GRA).
Natal. — Kranskop: near Olifantshoek store, Edwards 3334. Polela: Marwaga Hill, near Bulwer,
Marais 830; near Xumeni Forest, Gerstner 7012. Underberg: McC/ean 672; Umkomazana River
Valley, Marais 1429; 9 miles N. of Kingscote, Codd 8547.
470
Fig. 61. — K. triangularis subsp.
triangularis illustrating the
form with serrulate leaves
described as K. macowanii
Bak.
Berger in his monograph, l.c. (1908), upholds the following four species in his
section Obtusilobae: K. triangularis, K. nelsonii , K. macowanii and K. obtusiloba.
These are now included within the compass of one species as K. triangularis subsp.
triangularis and K. triangularis subsp. obtusiloba. Berger distinguished K. triangularis
and K. nelsonii on the basis of the perianth length in the former being 2-0-2- 5 cm
and, in the latter, 3 -0-3 -5 cm. However, in the present investigation, a complete
range of perianth length was found from 2 -4-3 -5 cm, no specimens having flowers
as short as 2-0 cm and only some specimens of subsp. obtusiloba being as long as
3-5 cm. The type of K. triangularis ( Drege 3524) has flowers about 2-5 cm long and
that of K. nelsonii (Nelson s.n.) flowers about 3-0 cm long. Both have narrow, grass-like
leaves 1 -5-3 mm broad with smooth margins, and no reliable basis can be found for
separating them even at varietal level.
In view of the markedly serrulate leaf margins of the type of K. macowanii,
consideration was given to separating it taxonomically, possibly as a variety. However,
a complete range of intermediates is found linking it with the smooth, grass-like leaves
(1 -5-3 mm broad) of typical K. triangularis. Oddly enough, the intermediate specimens
occupy a different geographical zonation to the east of the other two groups. Thus
typical K. triangularis is distributed mainly to the west of the Drakensberg escarpment
from Barkly East District in the extreme eastern Cape Province to the eastern O.F.S.
Plate 15. — Kniphofia triangularis Kunth subsp. triangularis
471
and the adjoining parts of Lesotho; specimens matching the type of K. macowanii
with leaves 2-5-6 mm broad and markedly serrulate margins occur to the south-west
of this area in the districts of Cathcart, Keiskammahoek, King William’s Town,
Queenstown, Somerset East, Stockenstroom and Stutterheim; while the intermediate
specimens with leaves 1 - 5—6 mm broad, sparingly serrulate to almost smooth, are
found on the eastern foothills of the Drakensberg, from the Umzimkulu District (Cape
Province) to Polela and Underberg Districts (Natal). This unusual situation cannot
satisfactorily be ascribed to hybridization between typical K. triangularis and typical
“ K. macowanii
Berger’s Section Obtusilobae, now reduced to one variable species, is characterised
by the small, dense, more or less concolorous inflorescences, coral-red to orange-red
in colour, with the perianth lobes somewhat spreading at anthesis. On p. 435 reference
is made to an attractive colour form of K. rufa which has lax, concolorous, coral-red
inflorescences with flowers 2-3 cm long, and narrow leaves with smooth leaf margins.
The distinction between lax and dense inflorescences is not always clear-cut and
occasional specimens may be difficult to classify.
24379—8
472
In the herbarium, K. thoclei, with its narrow leaves and sparingly toothed to entire
margins, may be confused with some forms of K. triangularis. However, its inflorescence
is bicolorous, the buds being dull red and the flowers whitish, with the perianth lobes
not spreading, so that it is clearly closer to K. porphyrantha in its affinity than to
K. triangularis.
(b) subsp. obtusiloba (Berger) Codd, stat. nov.
K. obtusiloba Diels ex Berger in Pflanzenreich 4, 38: 54, t . 2 1 (1908); Letty in Wild
Flow. Tvaal 33 (1962); Codd in Flow. PI. Afr. 36: 1. 1 422 (1964). Type: Transvaal,
Lydenburg District, Wilms 1559 (B !, holo.; BM!, K!).
This subspecies is characterised by the leaves being relatively soft and (2-5-)
4-10 mm broad with entire margins. The flowers tend to dry a purplish-brown colour.
Fig. 62. Plate 16. Map 11.
Occurs among quartzite rocks in mountain grassland at altitudes of 4,000 to 7,000
ft in the eastern Transvaal, with one record from Ngotshe District, Natal. The flowering
season is mainly from mid-January to the end of April.
Natal. — Ngotshe: near Ngome, Bruce 297.
Transvaal. — Belfast: Elandskop, near Izaak Siding, Galpin 13275; near Machadodorp, Rogers
18198 (BM); 16 miles S.E. of Machadodorp, Bruce 477; 10 miles W. of Slaaihoek, Bruce 484; 16 miles
N.E. of Machadodorp on road to Mareskop, Bruce 514; Mareskop, Bruce 516. Carolina: near
Slaaihoek, Pole Evans 3923; 4724; Codd 8274; 10306. Lydenburg: Wilms 1559 (B, BM, K); farm
Zwagershoek, Obermeyer 352; 10 miles S.E. of Lydenburg, Prosser 1790; Kemps Heights, 15) miles
S.E. of Lydenburg, Codd 8308; 17 miles S.E. of Lydenburg, Marais 51. Pilgrims Rest : Mariepskop,
Meeuse 9956; Van der Schijff 4518; 6524.
Subsp. obtusiloba is separated geographically from subsp. triangularis and was,
at first, considered to be specifically distinct. Recent gatherings from Mariepskop
in eastern Transvaal are very reminiscent of K. triangularis in their narrow leaves
which are, however, softer in texture and thus form a link between the two subspecies.
In subsp. obtusiloba the leaves are consistently smooth-margined and herbarium
specimens are apt to be confused with K. porphyrantha. The latter may be separated
by its bicolorous inflorescence with flowers becoming lemon-yellow as they mature.
The Mariepskop gatherings referred to above are not clearly distinct from K. galpinii
which also occurs in the eastern Transvaal, but the latter species is again distinguished
by its bicolorous, red and yellow, inflorescence and the more rigid, fibrous, grass-like
leaves.
32. Kniphofia thodei Bak. in FI. Cap. 6: 533 (1897); Berger in Pflanzenreich 4,
38; 57 (1908); Codd in Flow. PI. Afr. 37: 1. 1 446B (1966). Type: O.F.S., Caledon
Range, Roodeberg, Moperis Peak, Thode 62 (BOL!, holo.; K!, NH!, PRE!, STE!).
Plants usually solitary. Leaves 6-8 per peduncle, narrow, at first erect, later
recurved or bent downwards, 25-50 cm long, 2-5 mm broad, broadly triangular in
cross section, narrowly channelled above and slightly keeled below, pale glaucous-
green, soft in texture, breaking into fibres at the base; margin finely serrulate, parti-
cularly towards the apex, to smooth. Peduncle 30-50 cm long, usually overtopping
the leaves, with an occasional sterile bract below the inflorescence. Inflorescence
subglobose to ovoid, dense, 4-6 cm long and 4-5 cm in diameter; buds and flowers
pendulous; buds dull red to reddish-brown often tipped with white; flowers white
or tinged with reddish-brown towards the base. Bracts lanceolate, long-acuminate,
6-7 mm long, 2 mm broad near the base; margin entire. Pedicels 1-5-2 mm long.
Perianth subcylindrical, somewhat curved, not constricted above the ovary, 2 -8-3 -5 cm
long, 3 mm broad near the base, scarcely widening to 3-5-4 mm at the throat; lobes
473
broadly ovate, 1-5 mm long, not spreading. Stamens reaching the throat of the
perianth at anthesis, later withdrawn. Style eventually exserted by 4 mm. Fruit not
seen. Fig. 63. Plate 4b. Map 11.
Found in moist places in mountain grassland at altitudes of 7,500 to 9,000 ft in
Lesotho and eastern O.F.S., and at altitudes of 5,000 to 6,000 ft. in the Estcourt District,
Natal. The flowering season is mainly in January, but flowering specimens have been
collected as early as November and as late as March.
Fig. 63. — K. thodei Bak., Kamberg, Natal (N. R. Smuts 1420). Photo by Dr. N. R. Smuts.
474
O.F.S.- Harrismith: Witzieshoek, Moperis Peak, Thode 62 (BOL, K, PRE, STE).
Lesotho. — Without locality. Staples 245; Pitseng, Malutis, Dieterlen 1286; Leribe, Dieterlen 872;
between Tsienyane Valley and Bokong River, Coetzee 505 (UOFS); Little Bokong Valley, Jacot
Guillarmod 334; Jordan Valley, Jacot GuiUarmod 1670; Butha Buthe, Tsehlanyane Valley, Jacot
Guillarmocl 4011.
Natal. — Estcourt: near White Mt. Inn, Killick & Marais 2155; 2163; Kamberg Nature Reserve,
Wright 3; TV. 7?. Smuts 1420. Lions River: Nottingham Road, Shirley s.n. (NU).
Herbarium specimens of K. thodei may be confused with either K. porphyrantha
or K. triangularis but, when fresh specimens are available, the species may be recognized
by the unusual colour combination in the inflorescence: the buds are coral-red to
dull red, often tipped with white, while the open flowers are whitish in colour. Its
nearest relation is K. porphyrantha (see p. 476), which has lemon-yellow flowers, tinged
with red at the apex of the inflorescence, and leaves yellow-green in colour with smooth
margins. The leaves of K. thodei are normally narrower than those of K. porphyrantha ,
pale glaucous-green in colour, while the margins are usually sparingly denticulate.
It is the minute teeth on the leaf margins and the narrow leaves which suggest an affinity
to some forms of K. triangularis (see p. 472), but the inflorescences of the latter are
uniformly coral-red, while the perianth lobes spread when the flowers are open.
Baker, in Flora Capensis, states that there is no specimen of the species at Kew.
However, a sheet of the type, Thode 62, is at present in Kew Herbarium. The specimen
in Bolus Herbarium is annotated in Baker’s handwriting and may thus be regarded
as the holotype.
Justus Thode, after whom the species is named, came originally from Europe
(probably Germany) and collected plant specimens in all four Provinces. In a paper
on the Botanical Regions of Natal, presented to the Durban Field Naturalists Society
in May, 1901, and published in the Natal Mercury, he mentions in passing the presence
of the “ showy, scarlet or orange Kniphofias ( K . thodei , a brilliant crimson and
white) . . .’’at high altitudes in the Drakensberg.
33. Kniphofia porphyrantha Bak. in J. Bot. Lond. 12; 4 (1874); ibid. 23: 279
(1885); FI. Cap. 6: 281 (1896); Berger in Pflanzenreich 4, 38: 59 (1908); Codd in
Flow. PI. Afr. 33: 1. 129 1 (1959); Letty, Wild Flow. Tvaal. 32, 1. 1 9, 1 (1962). Type:
O.F.S., Cooper 3208 (K!, lecto.); 3207 (K!).
K. conratliii Bak. in Bull. Herb. Boiss. 2 ser., 4: 998 (1904); Berger, l.c. 59 (1908).
Type: Transvaal, Modderfontein, Conrath 644 (K!, holo.; Z!).
Plants usually in dense groups, rarely growing singly. Leaves 10-12 per peduncle,
yellow-green, flaccid, 30-45 cm long, 6—14 mm broad, more or less flat above, shallowly
channelled, narrowly keeled below, at first erect, later reflexed; margins and keel
smooth. Peduncle overtopping the leaves, 40-60 cm long. Inflorescence subglobose
to shortly cylindrical, dense, 4-8 cm long and about 5 cm broad, with a coma of sterile
bracts at the apex; buds spreading, orange-flame tipped with yellow, flowers pendulous,
lemon-yellow. Pedicels 1*5 mm long. Bracts ovate-lanceolate, 6-9 mm long, 2-3 mm
broad, acute to acuminate, sometimes minutely toothed near the apex. Perianth
narrowly subcylindrical, not constricted above the ovary, somewhat arcuate, 3-4 -2 cm
long, 3 mm broad at the base, widening gradually to 4-5 mm at the throat; lobes
ovate, obtuse, 2-5 mm long, at first erect, sometimes spreading at a later stage. Stamens
exserted by 1-3 mm at anthesis, finally withdrawn. Style subequal to the stamens
at anthesis, finally exserted by about 6 mm. Fruit erect, subglobose, obscurely trique-
trous, 8-10 mm long. Fig. 64. Plate 17. Map 12.
Widespread in grassy vleis and mountain grassland at altitudes of 4,500 to 7,000 ft
in eastern O.F.S., north-western Natal, southern and south-eastern Transvaal, extending
into the western part of Swaziland. The flowering season is mainly October to December
especially in the Transvaal, but extends to January and February in Natal.
Plate 17. — Kniphofia porphyrantha Bak.
475
O.F.S. -Without locality: Cooper 3207 (1C) ; 3208 (K). Bethlehem: Golden Gate National Park'
Roberts 345 1 . Harrismith: Van Reenen, Bews 541 ; near Witzieshoek, Bolus 8259 (BOL); Van Reenens
Pass, Jacobs z s.n.; Renburgskop, Jacobsz 293.
Natal.— Bergville: Mont-aux-Sources, Schweickerdt 697; Oliviershoek, Vogts s.n.; Royal Natal
National Park, Killick & Marais 2212; Cathedral Peak Forest Station, Killick 1265; 1305; Germis-
huizen 45 ; 46. Estcourt: Cathkin Peak, Galpin 11740; Tambamhlope, West 400. Newcastle: near
Normandien Pass, Codd 9978. Utrecht: Tweekloof, Altemooi, Thode A213; near Groenvlei, Bruce
283: Codd & Dyer 6285; 13 miles S.E. of Groenvlei, Codd 6954; Naauwhoek, Devenish 1002.
Fig. 64. — K. porphyrantha, Utrecht District ( Codd & Dyer 6285). Photo by Dr. R. A. Dyer.
Transvaal. — Belfast: near Belfast, Leendertz 9208; Hutchinson 2782; Codd 7592; near Dull-
stroom, Galpin 13175. Bronkhorstspruit : farm Wag-'n-bietjie-kop, Francis 2. Carolina: Rademacher
in TRV 8204; Witpoort, Van der Merwe in PRE 24080; 8 miles N. of Carolina, Reynolds 3551; 11^
miles S.E. of Carolina, Bruce 302; 3 miles W. of Oshoek, Codd 47 60. Ermelo: near Ermelo, Burtt
Davy 917 (K); Van der Merwe s.n.: Spitskop, Pott-Leendertz in TRV 15116; Nooitgedacht, Potter
sub Henrici 1736; near The Gem, Walker 101; Codd 4768; Gunn s.n.; near Warburton, Bruce 267 ;
6-j miles N. of Ermelo, Codd 8069. Heidelberg: Leendertz 7711; near Heidelberg, Schweickerdt 983;
Prosser 1677. Johannesburg: Diepkloof Farm, Watt 4684; Modderfontein, Conrath 644 (K, Z).
Lydenburg: 14 miles S.W. of Spitskop on Lydenburg-Dullstroom road, Bruce 325. Middelburg:
Van der Merwe s.n. PietRetief: Galpin 9613; Iswepe, Sidey 2056. Potchet’stroom : near Kraalkop,
Codd 3434. Pretoria: Fairy Glen, Leendertz 3689; Mogg 18991; Derdepoort, Leendertz 4042;
Baviaanspoort, C. A. Smith 3511; 19 miles E. of Pretoria, Codd 3188; near Tygerpoort, Bruce 30;
14 miles S.E. of Pretoria, Bruce 29; 228; Bapsfontein, Codd & Bruce 3382; Bruce 31; 32; 33; 34.
Volksrust: near Volksrust, Galpin 11338. Wakkerstroom: Van der Merwe 3; 2056; Austin Roberts
in TRV 12382; farm Oshoek, Devenish 225; 515; 793; farm Damascus, Devenish 298. Witbank:
15 miles W. of Witbank, Reynolds 2650.
Swaziland. — Mbabane: Forbes Reef road, Compton 27186.
K. porphrantha is a fairly widespread species, flowering from early to mid-summer,
relatively small in stature with attractive small inflorescences, tinged with orange-flame
at the apex, the flowers becoming pale to lemon-yellow. The leaves have smooth
margins and are flaccid and yellow-green in colour, often with a waxy bloom. In
the herbarium it is not always easy to distinguish between specimens of this and certain
allied species with acute to acuminate bracts but, with fresh material, the above
476
characters provide a basis for separating K. porphyrantha from the following: (a) K.
triangularis subsp. obtusiloba , which has a concolorous, coral-red to orange-red in-
florescence and spreading perianth lobes; (b) K.galpinii, with its predominantly scarlet-red
inflorescence and narrower, fibrous and somewhat grass-like leaves; ( c ) K. thodei ,
with the unusual colour combination in the inflorescence consisting of dull red buds
and whitish flowers, while the leaves are narrower with sparingly denticulate margins;
and ( d ) K. fluviatilis , in which the inflorescence has a similar colour pattern to that
of K. porphyrantha , though the flowers tend to be longer (4-5 cm), and which differs
mainly in the broader, more glaucous leaves which are V-shaped in cross section (see
also p. 478).
Map 11. — Distribution of Kniphofia galpinii ,
K. triangularis, subsp. obtusiloba and K.
thodei.
Map 12. — Distribution of Kniphofia porphy
rantha and K. fluviatilis.
When he described K. conrathii, Baker related it to K. citrina Bak., an autumn-
flowering species from the eastern Cape Province with obtuse bracts, shorter flowers
and markedly exserted stamens. Berger, in his monograph, correctly aligned it to
K. porphyrantha , but separated the two on the basis of the style eventually being exserted
in K. conrathii and remaining within the perianth tube in K. porphyrantha. An exami-
nation of the respective type specimens shows little difference in this respect and the
degree of exsertion of the style depends largely on the age of the flower.
34. Kniphofia fluviatilis Codd in Flow. PI. Afr. 36: 1. 1 42 1 (1964). Type: Trans-
vaal, Belfast District, 6 miles N.E. of Dullstroom, Codd 7625 (PRE!, holo.; K!).
Plants in groups. Leaves 8-12 per peduncle, erect, firm in texture, dull to glaucous
green, tapering rapidly, 35-70 cm long, 0-8-2 -5 cm broad, V-shaped in cross section
with a rather thick keel, not strongly nerved and not breaking into fibres at the base;
margin and keel quite smooth. Peduncle moderately stout, 45-60 cm long, sometimes
with 1 or 2 sterile, lanceolate, acuminate bracts below the inflorescence. Inflorescence
ovoid or subglobose, often pyramidal at the apex, dense, 6-9 cm long and 6-8 cm broad,
with a coma of sterile bracts at the apex; buds and flowers pendulous; buds flame- to
orange-red, flowers apricot-yellow to greenish-yellow. Bracts ovate-lanceolate to
linear-lanceolate, acute to long-acuminate, 7-13 mm long, 3-4 mm broad; margin
entire. Pedicels 1-5-2 mm long, elongating to 2-5-3 mm at fruiting stage. Perianth
477
subcylindrical, not constricted above the ovary, often somewhat arcuate, 4-2-5 0 cm
long, 4 mm broad at the base, increasing gradually to 6 mm at the apex; lobes ovate,
2 mm long, obtuse, not spreading. Stamens just included or slightly exserted at anthesis,
later withdrawn. Style subequal to the stamens at anthesis, finally exserted by 2-4 mm.
Fruit erect, globose to ovoid, 8-10 mm long. Fig. 65. Plate 18. Map 12.
On the banks of streams in mountain grassland, usually growing partly in running
water, at altitudes of 4,000 to 7,000 ft, recorded from the eastern Transvaal, and the
foothills of the Drakensberg Range in Natal and East Griqualand. The main flowering
time is during November and December, but specimens in flower have been collected
as late as February.
Transvaal. Belfast: Doidge & Bottomley s.n.; O'Connor 6; 6 miles N.E. ol Dullstroom, Codd
5639; 6768; 7625; Codd & De [Vinter 3234; 8 miles N. of Dullstroom, Codd 6655; Velorevlei,
Steenkampsberg, Du Plessis s.n. Nelspruit: Schagen, Amajuba Mt., Liebenberg 3116. Piet Retief:
10 miles N. of Iswepe, Sidey 1553; Bruce 276. Pilgrim's Rest: Mt. Anderson, Smuts & Gillett 2379,
Wakkerstroom : farm Oshoek, Devenish 153; 168; 721.
478
O.F.S. — Harrismith: Jacobsz 321.
Natal. — Estcourt: near White Mt. Inn, Killick & Marais 2154: 11 miles from White Mt. Inn on
road to Hlatikulu, Killick & Marais 2156; Kamberg Nature Reserve, Wright 1. Lions River: near
Rosetta, Rogers 28148 (Z); Nottingham Road, Galpin 9485; 3 miles N. of Nottingham Road, Marais
929; 3j miles W. of Nottingham Road, Marais 932; Umgeni Poort, Moll 1367. Underberg: near
Himeville, Killick & Marais 2088. Weenen: Culvers, Rogers 28274 (K, Z).
Cape. — Matatiele: 8-)- miles S.W. of Cedarville, Bruce 581. Umzimkulu: 28 miles from Swartberg
on road to Underberg, Killick & Marais 2085.
The species was first collected by Miss Doidge and Miss Bottomley in 1924
near Belfast and later, in 1928, by Galpin near Nottingham Road. It is closely related
to K. porphyrantha (see also p. 476), but has longer flowers, 4-2-5 cm long (among
the longest recorded in the genus) and broader leaves which are dull green to semi-
glaucous in colour, and are distinctly V-shaped in cross section K. fluviatilis also
shows an ecological distinction in its occurrence almost invariably along mountain
streams, usually partly in the water, while K. porphyrantha inhabits grassy vleis and
moist hillsides. The two species are rarely found growing together at the same locality,
although their general distribution overlaps to some extent.
35. Kniphofia littoralis Codd, sp. nov., a K. baurii Bak. foliis integris vel subintegris,
bracteis obtusis vel acutis, fructibus majoribus ovoideis 1-2-1 -5 cm longis differt.
Planta 20-65 cm alta. Folia erecta vel recurvata, carinata, subglauca, flaccida,
15-70 cm longa, 1-2-5 cm lata, margine integra vel parce serrulata. Inflorescentia
densa, subglobosa vel breviter cylindrica, 5-9 cm longa, 5-5-5 cm lata, gemmis pendulis
rubris vel lateritiis, floribus maturis flavo-virentibus. Bracteae ovatae vel oblongo-
ovatae, acutae vel obtusae, 5-8 mm longae, 2-5-3 mm latae, integrae. Pedicelli 1-1 -5
mm longi. Perianthium subcylindricum, 2-8-3 -4 cm longum. Stamina inclusa vel
vix exserta. Stylus 3-5 mm exsertus. Capsula ovodia, 1-2-1 -5 cm longa, 8-10 mm
lata.
Type: Natal, Umzinto District, near Park Rynie, Codd 9764 (PRE!, holo.).
Plants in groups. Leaves 6-10 per peduncle, erect (when growing in dense grass)
to strongly recurved (when in an exposed place, or after the grass has been burnt in
winter), dull to glaucous-green, soft in texture, 15-70 cm long, 1-2-5 cm broad,
chartaceous at the base, not or scarcely breaking into persistent fibres at the base,
V-shaped in cross section, usually folded along the midrib (especially in dried specimens
with recurved leaves); margin and keel smooth or minutely serrulate towards the
apex. Peduncle overtopping the spreading leaves, 20-65 cm long, with an occasional
sterile bract below the inflorescence. Inflorescence globose to shortly cylindrical, dense,
5-9 cm long, 5-5-5 cm broad, usually with a few sterile bracts at the apex, buds and
flowers pendulous; buds red to dull red (dragon’s blood red, Ridgeway XIII), turning
to pale yellow-green (primrose yellow, Ridgeway XXX to sea foam green, Ridgeway
XXXI) as the flowers open. Bracts ovate to ovate-oblong, 5-8 mm long, 2-5-3 mm
broad, acute to obtuse, white, chartaceous, with a median brownish nerve; margin
entire. Pedicels 1-1 -5 mm long at flowering stage, elongating to 5 mm long in fruit,
arcuate. Perianth subcylindrical, slightly arcuate, 2 -8-3 -4 cm long, 3 mm broad at
the base, very slightly constricted above the ovary, expanding to 5-6 mm broad about
the middle and then parallel-sided to the throat; lobes broadly ovate, obtuse, not
spreading. Stamens in the throat or slightly exserted at anthesis, later withdrawn.
Style more or less equal to the stamens at anthesis, later exserted by 3-5 mm. Fruit
ovoid, erect, I -2-1 -5 cm long, 8-10 mm broad, acute at the apex. Plate 19.
Found in Natal in moist, grassy, low-lying places near the coast from just above
sea level to about 600 ft. in altitude, from Umzinto District in the south to Hlabisa
District in the north. The flowering time is from August to October.
Plate 19. — Kniphofia littoralis Codd
479
Natal. — Durban: near Durban, Medley Wood 2634 (NH); 7162 (NH); Umgeni, Rehmann 8563
(Z). Eshowe: near Eshowe, Lawn 1074a. Hlabisa: Palm Ridge Farm, 8 miles N. of Mtubatuba,
Harrison I. Lower Tugela: Sheffield Beach, 37 miles N. of Durban, Bruce 155. Umlazi: Isipingo,
Miss Thompson s.n. Mtunzini: Ngoye, Venter 316 (UCZ); Umhlatuzi River, Venter 2695; Hamewith,
Mogg 4493. Umzinto: Rudatis 485; Park Rynie, Gerstner 6814 (K); 6815; Bruce 423; Codd 9699;
9764.
The large fruits of this species are characteristic, being ovoid in shape and acute
to almost beaked at the apex, 1 -2-1 -5 cm long, carried erect on stout, arcuate pedicels
up to 5 mm long. Thus when fruits are present, K. littoralis can be distinguished from
all other members of the genus. In floral characters and general appearance, however,
it bears a resemblance to K. baurii (p. 482), K. drepanophylla (p. 481 ), and to depauperate
forms of K. rooperi (p. 487).
K. bauri is probably its closest affinity, but has leaves which are markedly serrulate
while the bracts are broader and rounded at the apex. K. drepanophylla is similar to
K. baurii in these respects and can be differentiated from K. littoralis in the same way.
These two species are not known to occur near the coast in Natal. K. rooperi , on
the other hand, is essentially a coastal species, extending to a few miles north of Port
Shepstone in Natal. There is no clear evidence, as yet, that its distribution overlaps
with that of K. littoralis, which reaches its southernmost known limit a few miles south
of Scottburgh and extends northwards to beyond Mtubatuba.
K. rooperi is a robust species with broad, recurved leaves in four ranks, which
flowers from about May to October. It reaches its maximum development in coastal
marshes, and the plants are noticeably smaller towards the drier soil at the edges of
the marsh. These smaller plants, which are referred to as depauperate specimens,
may easily be confused with K. littoralis in the herbarium if fruits are lacking. However,
there are small differences, which are not easy to put into words, but which enable
specimens to be identified with a high degree of certainty. The bracts of K. littoralis
are usually white and papery, acute to obtuse and not erect in bud while, in K. rooperi ,
they are usually discoloured and brownish, rounded at the apex and erect in the bud
stage. K. littoralis also has slightly longer pedicels and the flowers are not so densely
placed that the rhachis of the inflorescence cannot be seen; in K. rooperi, on the other
hand, the almost sessile flowers are so densely placed as to obscure the rhachis. Although
the leaves of K. rooperi are normally broader, with serrulate margins, in many of the
depauperate specimens the leaves are narrower and with margins almost completely
entire, or with a few teeth near the apex as in K. littoralis. Although such intermediate
specimens may give the impression of being hybrids, it is considered that they are rather
variants of K. rooperi (see p. 491).
On page 491 the name K. longicollis Bak., based on a cultivated plant said to have
been introduced from Natal by Max Leichtlin of Baden-Baden, Germany, is discussed.
The type specimen, which is in K.ew Herbarium, superficially resembles the specimens
now placed in K. littoralis, but has bracts which are oblong-lanceolate, tapering to a
somewhat rounded apex, with three distinct nerves and minutely serrulate margins.
These are so unlike the bracts of K. littoralis that the two cannot be regarded as
conspecific. In addition, the flowers of K. longicollis are described as being “ bright
yellow without a tinge of red ”, while no mention is made of its fruits, which are such
a characteristic feature of K. littoralis. In many respects K. longicollis resembles the
depauperate forms of K. rooperi mentioned above; the bracts do not match perfectly,
but K. rooperi (and occasionally K. baurii) sometimes has 3-nerved bracts (see also
p. 487). The plant illustrated in the Botanical Magazine t .7623 (1898) as K. longicollis
is, of course, quite different from the type of that species.
480
Fig. 66. — K. elegans, at
Mkambati Leper Insti-
tution, Pondoland
(Marais 952). Photo
by W. Marais.
36. Kniphofia elegans Coild , sp. nov., a K. Jittorali Codd bracteis subrotundis,
capsulis globosis 6-8 mm longis differt.
Planta 40-65 cm alta. Folia erecta, angusta, subglauca, 30-60 cm longa, 0-8-1 -5
cm lata, carinata, margine integra vel apicem versus parce serrulata. Inflorescentia
densa, globosa, 5-7 cm longa, 5-6 cm lata, gemrnis patentibus rubris, floribus maturis
demum pendulis luteis vel luteo-aurantiacis. Bracteae late ovatae vel subrotundae,
obtusae vel rotundatae, 6-8 mm longae, 3-4 mm latae, integrae vel minute erosodenti-
culatae. Pedicelli 1-2 mm longi. Perianthium subcylindricum, 2-8-3 -2 cm longum.
Stamina inclusa vel vix exserta. Stylus 4-5 mm exsertus. Capsula glbosa, 6-8 mm
longa.
Type: Cape Province, Lusikisiki District, Mkambati Leper Institution, Codd 9720
(PRE!, holo.).
Plate 18. — Kniphofia Jluviatilis Codd
iJiii"
481
Plants solitary or in small groups. Leaves 6-8 per peduncle, erect or rarely recurved,
narrow, dull glaucous-green, 30-60 cm long (the outermost much shorter), 0-8-1 -5 cm
broad, chartaceous at the base, not breaking into fibres, V-shaped in cross section;
margin and keel smooth or sparingly denticulate towards the apex. Peduncle over-
topping or subequal to the leaves, 40-65 cm long with an occasional sterile bract
below the inflorescence. Inflorescence globose, dense, 5-7 cm long and 5-6 cm broad;
buds spreading, flame-red; flowers pendulous, yellow to orange-yellow. Bracts broadly
ovate to subrotund, obtuse to rounded, 6-8 mm long, 3-4 mm broad, chartaceous;
margin almost entire to minutely eroso-denticulate. Pedicels 1-2 mm long. Perianth
subcylindircal, not or scarcely constricted above the ovary, somewhat arcuate 2 -8-3 -2
cm long, 2-5 mm broad at the base, increasing to about 5 mm broad about the middle
and then parallel-sided to the throat; lobes broadly ovate, 2 mm long, not spreading.
Stamens in the throat or just exserted at anthesis, finally withdrawn. Style subequal
to the anthers at anthesis, finally exserted by 4-5 mm. Fruit globose, 6-8 mm long
and 6-9 mm in diameter. Fig. 66.
Grows in grass associated with rocks of Table Mountain Sandstone at altitudes
of 200 to 1,000 ft. in the Pondoland area of the eastern Cape Province. The flowering
time is in spring, from August to October.
Cape.— Bizana: 4 miles inland from Umtamvuna mouth, Cockt 9713. Lusikisiki: Mkambati Leper
Institution, Codd 9720; 9729; Marais 952; 970; Hone 26; 13 miles N.W. of Port Grosvenor, Codd
9737.
K. elegans is a spring-flowering species with small, globose, brightly-coloured
inflorescences, flame-red at the apex and turning yellow as the flowers open. Related
to K. littoralis, it differs in the globose fruits, borne in dense clusters on short pedicels,
the globose inflorescence and the more rounded bracts. The species was first collected
by Mr. W. Marais in October, 1955, in the grounds of the Mkam Leper Institu-
tion, where it is fairly common.
37. Kniphofia drepanophylla Bak. in Bot. Jahrb. 15, Beibl. 35: 5 (1892); FI. Cap.
6: 282 (1896); Berger in Pflanzenr. 4, 38 ; 48 (1908); Codd in Flow. PI. Afr. 37: 1. 1 447
(1965). Type: Pondoland, Bachmann 281 (K, holo.!; Bt).
Plants forming small groups. Leaves 6-8 per peduncle, broad, falcate, strongly
recurved, flaccid, more or less in 4 ranks, pale yellow-green, 15-30 cm long at flowering
stage, eventually elongating to 60 cm long, 1-5-3 cm broad, V-shaped in cross section,
tending to fold along the midrib, often strongly nerved, not breaking up into persistent
fibres at the base; margin and keel finely and distinctly serrulate. Peduncle overtopping
the recurved leaves, 25-50 cm long. Inflorescence subglobose, 5-7 cm long and 5-6 cm
broad, often lax at the apex, denser below; buds spreading, greenish-yellow, often
suffused with dull red; flowers ultimately pendulous, lemon-yellow. Bracts broadly
oblong, rounded to obtuse, 5-6 mm long, 3 mm broad, scariose; margin subentire
to somewhat eroded. Pedicels 3- 5-5 mm long. Perianth subcylindrical, not constricted
above the ovary, 3.5-4 cm long, 3 mm in diameter at the base, increasing gradually
to 5-6 mm at the throat; lobes spreading, broadly ovate to rotund, 3 mm long.
Stamens exserted by up to 5 mm at anthesis, later withdrawn. Style subequal to the
anthers at anthesis, finally exserted by 6 mm. Fruit ovoid-triquetrous, 7-9 mm long.
Plate 20.
Grows in marshy places at altitudes of 200 to 500 ft. near the coast in Pondoland,
where it is locally common, with one record from southern Natal at an altitude of
3,500 ft. The flowering time is mainly August to October.
Cape.- Without precise locality: Pondoland, Bachmann 281 (K). Lusikisiki: near Port Grosvenor,
Codd 9738; Mkambati Leper Institute, Story 4233; Codd 9717; Marais 953; Hone 35.
Natal.- Alfred: near Weza, Killick & Marais 2014.
This is a spring-flowering species, small in stature with relatively few-flowered inflore-
scences of lemon-yellow flowers, tinged with red at the apex. The leaves are broad,
482
yellow-green, strongly recurved, with finely serrulate margins. In the herbarium it
may be confused with K. baurii which has somewhat narrower, glaucous leaves, slightly
shorter flowers and a more inland distribution. Until a more detailed study of the
two in the field has been made, it is preferred to keep them distinct. The specimen
Killick & Marais 2014 from Weza falls within the distribution range of K. baurii but
its characteristics render it almost indistinguishable from K. drepanophylla.
Until a few years ago the species was known only from the type material gathered
in Pondoland by Bachmann in August 1888. The second collection was made by
Dr. Story in August, 1953, also in Pondoland, where it is locally common in marshy
grassland, protected from grazing, on the Mkambati Leper Institution.
38. Kniphofia baurii Bak. in FI. Cap. 6: 281 (1896); Berger in Pflanzenr. 4, 38: 48
(1908); Codd in Flow. PI. Afr. 36: 1. 1423 (1964). Type: Cape, Tembuland, Baziya,
Baur 755 (K !, holo.).
Plants in small groups. Leaves 8-10 per peduncle, erect or somewhat falcate or
occasionally recurved, soft in texture, glaucous, 20-50 cm long, 1-2-2- 5 cm broad,
V-shaped in cross section, often distinctly nerved, not breaking into persistent fibres
at the base; margin and keel markedly to sparingly serrulate or almost entire. Peduncle
Plate 21. — Kniphofia baurii Bak.
483
overtopping the leaves, 25-60 cm long. Inflorescence globose to oblong, sometimes
broader than long, dense, 4-10 cm long, 4-6 cm broad; buds spreading, tinged with
dull red; flowers becoming pendulous, greenish to greenish-yellow. Bracts ovate to
ovate-oblong, obtuse to rounded, 4-5-10 mm long, 2-5-3 mm broad, scariose, with
a brownish, central nerve; margin eroso-denticulate, especially towards the apex.
Pedicels 1-2 mm long. Perianth cylindrical, 2 -8-3 -8 cm long, slightly constricted
and 2 mm in diameter above the ovary, widening to 4-5 mm broad at the throat; lobes
ovate, obtuse, 2-5 mm long. Stamens exserted by up to 2 mm at anthesis, later
withdrawn. Style subequal to the anthers at anthesis, finally exserted by 3-5 mm.
Fruit subglobose-triquetrous, 7 mm long. Figs. 67, 68. Plate 21.
Found in eastern Cape Province, southern and northern Natal on moist, grassy
slopes or depressions and on stream banks, at altitudes of 2,000 to 4,000 ft. and
flowers from about mid-September to early November. (See notes below for reference
to a late-flowering form).
Cape. — Engcobo: near Engcobo, Marais 1028 ; 1032; KiUiek & Marais 2051. Libode: near Libode,
Marais 1003; Killick & Marais 2034. Mount Currie: 17 miles E. of Kokstad, Killick & Marais 2024.
Nqamakwe: near Nqamakwe, Killick & Marais 2042. Unitata- Baziya, Baur 755 (K, 2 sheets);
Baziya Forest Station, Marais 1020; 1021; 6 miles W. of Umtata, Marais 1018.
Natal. — Alfred: Weza Forest Reserve, Killick & Marais 2016. Dundee: near Vants Drift, Baker
s.n. Newcastle: 9 miles S. of Newcastle, Codd 5609; 6161; Marais 958; 7 miles S. of Newcastle,
Bruce 461: Ingagane, Reynolds in NBG 624/40 (NBG); One Tree Hill, Shirley s.n. (NU). Nqutu:
east bank of Buffalo River near Vants Drift, Codd 6797 . Richmond : 8y miles S. of Richmond, Killick
& Marais 2000. Utrecht: near junction of Blood and Buffalo rivers, Codd 6788.
484
The typical form, on which the description is largely based, is found in the eastern
Cape Province and southern Natal. With its relatively small stature and often recurved
leaves which are markedly serrulate, it may be confused in the herbarium with K.
drepanophylla (p. 482), but a study of living plants leads to the conclusion that the
two are distinct. K. baurii has dark green to glaucous leaves, while the inflorescence
consists of dull red buds grading to greenish-yellow flowers, which are not well represented
in Plate 21 reproduced herewith.
In northern Natal, separated geographically from the typical form, several gatherings
have been recorded which are more robust and tend to grow in larger groups, with
somewhat softer and less markedly serrulate leaves than in typical K. baurii. In essential
characters they agree with K. buarii and are not given separate status. The citations
referred to are from two main localities, namely: (a) several miles south of Newcastle
on the road to Ladysmith; and (b) on the banks of the Buffalo and Blood Rivers,
where the Districts of Dundee, Utrecht and Nqutu meet.
A further group of specimens which deviate from typical K. baurii and which have
been collected mainly in East Griqualand, must be recorded. These were in flower
during January to March, i.e. a good deal later than is normal for the species, which
is essentially spring-flowering. Apart from this difference the plants compare well
with K. baurii, having greenish flowers, often tinged with red at the apex, dull green to
glaucous leaves which are hard and fibrous when dry and have serrulate margins.
The leaf characters assist in separating this form of K. baurii from certain plants
classified as K. linearifolia (see p. 502), which have similar inflorescence colour and
approximately the same flowering time, but in which the leaves, although glaucous,
are soft in texture with no tendency to recurve, as is the case with K. baurii. Difficulty
may also be experienced in distinguishing between this form of K. baurii and the form
of K. uvaria found in the eastern Cape Province, which has arcuate, glaucous and
tough, fibrous leaves, but in which the leaf-margin is smooth (see p. 496).
The significance of these late-flowering specimens of K. baurii is not understood
and, as they are not distinguishable morphologically, they are not accorded separate
status at present. For convenience they are listed separately below.
Cape.- -Barkly East: 12 miles N.W. of Elands Heights, flrd. in cult. 15. 1 .57, Marais 1046. Maclear:
6 miles E. of Naudes Nek, 24.2.52, Bruce 575; 124 miles E. of Naudes Nek, 24.2.52, Bruce 576.
Matatiele: 64 miles N.E. of Paspalum, flrd. in cult. March, 1956, and January, 1957, Marais 856.
Mount Ayliff: 9 miles S. of Kokstad, 25.2.52, Bruce 590. Mount Currie: 5 miles W. of Kokstad,
25.2.52, Bruce 584; 6 miles N. of Kokstad, 25.2.52, Bruce 585; 6 miles S. of Kokstad, 25.2.52,
Bruce 586; 3 miles W. of Kokstad, 25 . 2 . 52, Bruce 587; 10 miles E. of Kokstad, flrd. in cult. 1.12.56,
18.12.56, Marais 942; 44 miles S.E. of Kokstad, 29.1.57, Marais 1414; 3 miles N. of Kokstad, 29.1. 57,
Marais 1415; 124 miles S.E. of Swartberg, 24.3.54, Codd 8555.
It is desirable that a study be made in the field to determine the extent to which
individual plants vary in their flowering time from year to year. Obviously, weather
conditions will exert an influence, but the impression gained from cultivated and wild
plants is that flowering of a particular clump may extend over several weeks, but is
relatively constant, within a few weeks, from year to year. An occasional late inflore-
scence may appear in a clump which has flowered several months earlier, but this
does not account for the consistent later production of inflorescences in this late-
flowering form of K. baurii. There is also no real evidence that these plants have been
derived as a result of hybridisation between K. baurii and K. linearifolia.
39. Kniphofia latifolia Codd . sp. nov., a K. baurii Bak. foliis latioribus, per anthesin
pedunculos aequantibus vel superantibus, inflorescentia versus apicem gradatim
decrescenti differt.
Planta 60-100 cm alta. Folia carinata, erecta, ensiformis, 60-90 cm longa, 2-4
cm lata, margine carinaque serrulata. Inflorescentia densa, anguste ovoidea, versus
apicem gradatim decrescens, 7-11 cm longa, 4.5-5 cm lata, gemmis pendulis rubris
485
vel haematinis, floribus maturis flavo-virentibus. Bracteae latae oblongae, obtusae
vel rotundatae, 7-9 mm longae, 3-4 mm latae, minutae denticulatae vel integrae.
Pedicelli 1 mm longi. Perianthium subcylindricum, 3-3.5 cm longum. Stamina vix
exserta. Stylus 6-8 mm exsertus. Capsula ovoidea, 5 mm longa.
Type: Natal, plant collected 3 miles S.W. of New Hanover and cultivated in
Pretoria, flowered 28.10.1953, Codd 6791 (PRE, holo.).
Fig. 69. — K. tatifolia, from
New Hanover (Codd 6791).
Plants in groups. Leaves 8-12 per peduncle, broad, erect, yellow-green, over-
topping the inflorescences at flowering stage, 60-90 cm long (the outermost much
shorter), 2-4 cm broad, V-shaped in cross section, tapering rapidly to the apex, not
breaking into fibres at the base; margin and keel finely and regularly serrulate.
Peduncle shorter than the leaves at flowering stage, 40-80 cm long, eventually elongating
to 100 cm. Inflorescence narrowly ovoid, conical to tapering at the apex, dense, 7-1 1
cm long, 4-5-5 cm broad; buds pendulous, red to dull blood-red; flowers pendulous,
imbricate, greenish to yellow-green. Bracts broadly oblong, 7-9 mm long, 3-4 mm
broad, obtuse to rounded and somewhat discoloured at the apex, very minutely serrulate
to almost entire. Pedicels 1 mm long, elongating in fruit to 3 mm. Perianth sub-
cylindrical, 3-3-5 cm long, scarcely constricted above the ovary, 2-5-3 mm broad
near the base, widening gradually to 4-5 mm broad at the throat; lobes ovate, 3 mm
486
long, not spreading. Stamens exserted by 4-6 mm at anthesis, usually remaining
exserted by 2-3 mm. Style subequal to the stamens at anthesis, finally exserted by
6-8 mm. Fruit erect, broadly ovoid, 5 mm long. Figs. 69, 70.
Found on grassy slopes and river banks, usually in moist depressions with dense
grass and sedge, in a restricted area in the Natal Midlands at altitudes of 2,500 to
3,500 ft. The flowering season is October-November.
Fig. 70. — K. latifolia ( Codd 6791), inflorescence and portion of leaf, life size.
Plate 22. — Kniphojia rigidifolia E. A. Bruce
<
487
Natal. — New Hanover: 3 miles S.W. of New Hanover, cult, in Pretoria, Code I 6791. Nkandla:
5 miles W. of Nkandla, cult, in Pretoria, Codd 6989. Pietermaritzburg: Albert Falls, Bruce 460.
Umvoti: 7 miles S. of Greytown, Killick & Marais 2113.
Related to K. baurii Bak. (p. 482), this species differs in its broader leaves and
its inflorescnce tapering at the apex. The tapering inflorescence and the broader,
rounded bracts distinguish the species from K. rigidifolia E. A. Bruce (below), which
is restricted to the eastern Transvaal. K. lat [folia is a spring-flowering species in which
the broad, erect leaves usually overtop the inflorescnces at the flowering stage, but
later the scapes elongate, eventually becoming longer than the leaves in the fruiting
stage. It appears to have been first collected by Miss Bruce at Albert Falls in October,
1951.
40. Kniphofia rigidifolia E. A. Bruce in Flow. PI. Afr. 30: 1. 1 185 (1955); Fetty,
Wild Flows. Tvaal. 32 (1962). Type: Transvaal, Belfast District, 7 miles S. of
Dullstroom, Codd 3224 (PRE!, holo.).
Plants in groups. Leaves 12-16 per peduncle, erect, ensiform, firm in texture,
yellow-green, 50-80 cm long (1-5-) 2-3-5 cm broad, V-shaped in cross section, not
breaking into fibres at the base; margin and keel finely serrulate. Peduncle subequal
to the leaves or slightly longer, 50-90 cm long with an occasional deltoid sterile bract
24379—9
488
below the inflorescence. Inflorescence subglobose to broadly ovoid, very dense, 5-8
cm long, 5-6 cm broad, with a small coma of bracts at the apex; buds and flowers
pendulous; buds flame or orange-red to coral-red, flowers yellow-green to greenish.
Bracts oblong-ovate to oblong-lanceolate, subacute to obtuse, 5-7 mm long, 2-2-5
mm broad; margin finely eroso-denticulate to almost entire. Pedicels about 1 mm long,
increasing to 3 mm long at fruiting stage. Perianth subcylindrical, not constricted
above the ovary, 3-4 cm long, 3-3.5 mm broad near the base, widening gradually
to 5-5-5 mm at the throat; lobes broadly ovate, 2-5 mm long, slightly spreading at
anthesis. Stamens in the throat or slightly exserted at antheis, later withdrawn.
Style subequal to the stamens at anthesis, finally exserted by 2-5 mm. Fruit erect to
spreading, subglobose to broadly ovoid, about 7 mm long. Fig. 71. Plate 22.
Grows in dense grass among dolerite rocks and on fertile soil beside streams in
the eastern Transvaal, at altitudes of 4,500 to 6,500 ft. The main flowering season
is from mid-October to the end of November.
Transvaal. — Belfast: Van der Merwe 1095: Harding 1; 2; 3; 4; 5; near Dullstroom, Gatpin
13127; 7 miles S. of Dullstroom, Codd 3224; 5636; 11 miles N. of Belfast, Codd 6166: 6654; 5i
miles N.E. of Dullstroom, Codd 6770; 3 miles S. of Dullstroom, Reynolds 5771 ; 7 miles E. of Belfast,
Prosser 1506; 14 miles N.E. of Machadodorp, Codd 8066. Lydenburg: Van der Merwe 1613; 21
miles S.W. of Lydenburg, Codd 5641 ; P.O. Skaapkraal, Codd 6661 : near Kemps Heights, Codd 7596;
15 miles W. of Lydenburg, Codd 8047; Boschhoek, Young 440. Pilgrim's Rest : 6} miles N. of Graskop,
Codd 6740.
A spring-flowering species with globose inflorescences reminiscent of K. rooperi
(below), but K. rigidifolia is readily distinguishable by its erect, yellow-green leaves,
while the buds are not erect and covered by broad, imbricate bracts as is the case in
K. rooperi. The two species are completely separated geographically.
41. Kniphofla rooperi (Moore) Lem., Jard. Fleur, t.362 (1854); Bak. in J. Linn.
Soc. 11: 363 (1871); J. Bot. Loud. 23: 280 (1885), partly, excl. Bot. Mag. t.6 1 16;
FI. Cap. 6: 283 (1896), partly, excl. Bot. Mag. t .6 1 16; Berger in Pflanzenr. 4, 38: 62
(1908), partly, as to type only; Codd in Flow. PI. Afr. 33: 1. 1 290 (1959). Type: the
specimen ex Herb. T. Moore on which the plate of the species in Gard. Comp., l.c.,
is based (K!, holo.).
Tritoma rooperi Moore in Gard. Comp. 1: 113 cum tab. (1852).
K. longicollis Bak. in Gard. Chron. 13: 682 (1893); FI. Cap. 6: 284 (1896); Berger’
l.c. 60 (1908). Type: “Natal, Hort. Leichtlin, May 30, 1893 ” (K!, holo.).
Plants usually in groups. Leaves 10-16 per peduncle, usually in four ranks,
arcuate-spreading to strongly recurved, dull green, somewhat fibrous, 50-110 cm long,
1 -5-3-5 (-4) cm broad, deeply keeled and broadly V-shaped in cross section; margin
and keel distinctly to sparingly serrulate, rarely almost smooth. Peduncle overtopping
the spreading leaves, stout, 60-140 cm long, usually with a few large, sterile, deltoid
bracts below the inflorescence. Inflorescence large, very dense, ovoid in the early
flowering stage, becoming globose and later obovoid, 8-11 cm long and 5-8 cm broad,
apex rounded to truncate when in full flower, conical in the young stage with the buds
erect, enclosed in the overlapping, imbricate bracts; buds brilliant flame-red, orange-
yellow or greenish with scarcely a tinge of red; open flowers pendulous, orange-red
to yellow-green. Bracts oblong to obovate, 7-11 mm long, 2- 5-3- 5 mm broad, obtuse
to rounded, usually brownish, minutely serrulate to almost entire. Pedicels 1 mm long,
elongating in the fruiting stage to 3-4 mm long. Perianth subcylindrical, straight
or slightly arcuate, 3 -5-4 -2 cm long, scarcely constricted above the ovary, 2-5-3 mm
broad near the base, widening gradually to 5-6 mm at jhe throat; lobes ovate, 2.5
mm long, slightly spreading at anthesis. Stamens slightly exserted at anthesis, later
withdrawn. Style subequal to the stamens at anthesis, finally exserted by 4-6 mm.
Fruit ovoid-triquetrous, erect, 8-10 mm long. Figs. 72, 73, 74. Plate 23. Map 13.
489
Fig. 72. — K. rooperi, holotype in Kew Herbarium.
490
Found in marshy places near the coast from East London district in the eastern
Cape Province to several miles north of Port Shepstone in southern Natal. The main
flowering season is during the winter and early spring, from June to September, but
specimens have been known to flower, especially in cultivation, as early as March
and as late as November.
Cape. — Without locality: Cult. Brighton, ex Herb. T. Moore (K). Bizana: near Umtamvuna
mouth, Story 4134. East London: Malcomess in TRV 17129; Gonubie Springs, Compton 17036
(NBG); Bruce 605; 605A; near East London, Galpin 1862. Kentani: Pegler 838, partly, as to PRE
specimen; near Mazeppa Bay, Dyer 5335. Komga: near Komga, Bruce 601; 603; Flanagan 761;
Kei Mouth, Codd 6346. Lusikisiki: Mkambati Leper Institution, Codd 9728.
Natal. — Without locality: Hort Leichtlin, 1878 (K): Hort. Leichtlin, May 30, 1893 (K). Port
Shepstone: St. Michael-on-Sea, Van der Merwe 1940; near Margate, Bruce 417; 418; 419; 421;
422; 2 miles S. of Hibberdene, Bruce 424; \\ miles S. of Melville, Bruce 425; 428; Sunwich Port,
Codd 6793; Port Edward beach. Strey 5840; near Umtamvuna Bridge, Strey 5860.
Plate 23. — Kniphofia rooperi (Moore) Lem.
491
The original plant of K. rooperi which was described and figured by Thomas Moore
in 1852 came from the garden of the Rev. Thomas Rooper, of Wick Hill, Brighton,
who received it from his son, Capt. Edward Rooper of the Rifle Brigade, from
“ Kaffraria Moore’s herbarium, which included a specimen of the illustrated plant,
was later acquired by Kew, and this specimen may be regarded as the holotype. The
illustration and specimen agree in showing a plant with a large globose inflorescence,
similar to those recorded above from near East London to southern Natal.
Capt. Rooper was stationed in the Border area from time to time between 1848
and 1850, and it so happens that a collection of his original water-colour paintings
is in the library of the Botanical Research Institute, Pretoria. Plate 99 of this set is
an illustration of a large, globose Redhot Poker, annotated by the artist: “ Common
in Albany and Kaffirland. Flowers nearly all year. March ’49 ”. On a list accom-
panying the paintings is the record: “99. Tuberous rooted, in marshy places, varies
to deep red His references to “Albany ” and “ flowers nearly all year ” indicate
that his observations may be based on more than one species, and it must be borne in
mind that other members of the difficult K. uvaria-K. linearifolia complex may have
been encountered by Capt. Rooper.
K. rooperi can usually be recognized by the large, globose inflorescences which
appear mainly during the winter months, carried well above the arcuate-spreading
leaves which are broad, dull green and are normally arranged in 4 or 5 ranks. Flower
colour is not an important characteristic and may vary in wild populations, depending
on the amount of red pigment present, from dull greenish-yellow to brilliant scarlet.
The species reaches its maximum development of over \\ m in height in boggy, peaty
marshes near the coast. It is noticeable that the most robust plants are to be found
in the wettest, central part of the marsh while, toward the drier edges, the plants tend
to be smaller. Having noted this, it becomes possible to account for several gatherings
from the Natal South Coast which, because of their smaller stature, shorter leaves
and peduncles, were at first considered to be possible hybrids with other species, such
as K. littoralis and K. drepanophylla. The following are, therefore, considered to be
merely depauperate specimens of K. rooperi : Port Shepstone: near Melville, Bruce
426; 427; near Southport, Bruce 436; near Sunwich Port, Bruce 434; 435.
The type of K. longicollis Bak. is discussed under K. littoralis (p. 479). Apart
from the bracts, which are more tapering, it closely resembles the smaller specimens
of K. rooperi listed above. The type specimen was sent to Kew from Baden-Baden
in May, 1893, with a note from Max Leichtlin which reads: “ I beg to send you specimen
of a new Kniphofia from Natal which is now flowering freely here. Perhaps longicollis
or longitubulosa would be an appropriate name. In living specimens the unopened
flowers have a tinge of orange but become bright yellow when opening ”. It may be
significant that at Kew there is also a specimen labelled “ Natal, Hort. Leichtlin, 1878 ”
which is recognisable as K. rooperi and from which the 1893 specimen may possibly
be derived.
The plant illustrated in the Botanical Magazine t.7623 (1898) as K. longicollis
is, as already stated, quite a different plant (see also p. 387). It is probably a form
of K. praecox Bak. It is also doubtful if the plant figured as K. rooperi in the Botanical
Magazine t.6116 (1874) is correctly identified. The inflorescence is too elongate for
K. rooperi and it appears to be a form of K. linearifolia Bak.
From the neighbourhood of East London two specimens have been collected
which are of doubtful status, namely, Bruce 607 and Comins 1284. They are small
in stature, as in the case of the “ depauperate ” specimens listed above, and have
markedly curved, almost coiled leaves. These characters have remained constant in
cultivation, suggesting that a distinct variety is involved, but obviously more field
study is necessary before separate taxonomic rank can be allocated to these specimens.
24379—10
492
There is also some doubt about the identity of certain specimens listed below which
occur where K. rooperi and the eastern forms of K. uvaria find a common meeting
ground. They are somewhat intermediate between the two species but, because of
their subglobose inflorescences with broad, rounded bracts, they are regarded as being
nearest to K. rooperi.
Cape. — Humansdorp: between Storms River and Humansdorp, Bruce 233. Knysna: near Noetzie,
Bruce 238. King William's Town: near Berlin, Bruce 608; Marais 444; 729; Van Breda 890; near
Kei Road, Bruce 557. Komga: 8 to 10 miles W. of Komga, Bruce 601; 602: Codd 6350; 6351.
Stutterheim: 6 miles N. of Amabele Hotel, Bruce 558A; 558B.
42. Kniphofia uvaria (L.) Hook, in Bot. Mag. t.4816 (1854), partly, excl. fig. and
descr. ; Berger in Pflanzenr. 4, 38: 61 (1908), partly, excl. spec, cited; Adamson in
FI. Cape Penins. 170 (1950); Codd in Flow. PI. Afr. 33: 1. 1 289 (1959). Type: the
specimen in Herb. Hort. Cliff. (BM).
493
Aloe uvaria L., Sp. PI. 323 (1753); Kniphof, Herb. Viv. Cent. 9: t.805 (1762); L.,
FI. Cap. 4 (1759); Burm. f., Prodr. FI. Cap. 10 (1768); Mill., Diet. ed. 8: 23 (1768);
Hill, Hort. Kew. ed. 2: 333 (1769); Knorrs, Thes. rei Herb. Hort. t. A. 13 (1770);
Buchoz, Hist. Univ. Reg. Veg. 8: 1 (1775). A. longifo/ia Lam., Encycl. 1: 90 (1783),
nom. illegit. Type: as for K. uvaria.
Aletris uvaria (L.) L., Mant. 367 (1771); Ait., Hort. Kew. ed. 1 : 464 (1789); Thunb.,
Prodr. Cap. 60 (1794); L., Syst. Yeg. ed. 10: 359 (1797).
Kniphofia alodides Moench, Meth. 631 (1794); Kunth, Enum. PI. 4: 551 (1843);
Heynh., Nom. Bot. 2: 335 (1846); Bak. in J. Linn. Soc. 11: 364 (1871); J. Bot. Lond.
23: 279 (1885); FI. Cap 6: 283 (1896), partly; nom. illegit. Type as for K. uvaria.
K. burchelUi (Lindl.) Kunth, l.c. 552 (1843); Heynh., l.c. 335 (1846); Bak. in J. Linn.
Soc. 11: 363 (1871); J. Bot. Lond. 23: 280 (1885); FI. Cap. 6: 284 (1896); Berger
in Pflanzenr. 4, 38: 63 (1908). K. odorata Heynh., Nom. Bot. 2: 335 (1846), nom.
illegit. Type: as for K. uvaria. K. bachmannii Bak. in Bull. Herb. Boiss. 2 ser. 1:
784 (1901); Berger, l.c., 58 (1908). Type: Cape. Malmesbury District, Bachmann
1229 (Bf, holo.; Z!). K. occidentalis Berger, l.c., 57 (1908). Type: Cape Peninsula,
Smithwinkelbay near Simonstown, Schlechter 817 (Z!, holo).
Veltheimia uvaria (L.) Willd., Sp. PI. 2: 182 (1799); Pers., Syn. PI. Enchir. 1: 377
(1805); Thunb., FI. Cap ed. Schult. 309 (1823); Willd., Ann. Blumist. 2: 291, t.17
(1827).
Tritoma uvaria (L.) Ker.-Gawl. in Bot. Mag. t.758 (1804); Ait. f., Hort. Kew. ed. 2,
2: 290 (1811). T. burchellii Herb, ex Lindl. in Bot. Reg. 1. 1 745 (1836); Herb, ex
Sweet, Hort. Brit. ed. 2: 517 (1836), nom. nud; Nees & Sinning, Samml. Schonbl.
Gew. 19, t.7 (1831). Type: Hort. Kew, originally from the Cape, Burchell s.n. (K!,
holo.).
Tritomanthe uvaria (L.) Link, Enum. PI. Berol. 333 (1821). T. burchellii (Lindl.)
Steud., Nom. ed. 2: 718 (1840-41).
Tritomium uvaria (L.) Link, Handb. 170 (1829).
Triclissa uvaria (L.) Salisb., Gen. PI. 75 (1866).
Plants usually in small groups, 50-120 cm high. Leaves 10-20 per peduncle,
stiffly erect to arcuate-spreading, shorter than the peduncle, dull green to glaucous,
35-80 cm long and about 0-6-1 -8 cm broad, keeled, V-shaped in cross section, usually
tough and often drying with a hard, fibrous texture; margin and keel smooth to sparingly
serrulate with few scattered teeth mainly towards the apex. Peduncle overtopping
or subequal to the leaves, up to 100 cm long. Inflorescence oblong to globose, dense
to subdense or sometimes lax at the apex, 4-5-11 cm long and 5-8 cm broad; buds
spreading, brilliant scarlet to greenish tinged with red; flowers becoming pendulous,
orange-yellow to greenish-yellow. Bracts broadly ovate to oblong-ovate, rounded
or obtuse to subacute, 3-9 mm long; margin almost entire to eroso-denticulate.
Pedicels (1 -5-) 3-5 mm long, elongating to 8 mm long in the fruiting stage. Perianth
subcylindrical, almost straight, 2 -8-4-0 cm long, slightly constricted and about 3 mm
broad above the ovary, widening to 4-5 mm broad at the throat; lobes ovate, obtuse,
2 mm long, slightly spreading. Stamens included or just exserted at anthesis, later
withdrawn. Style subequal to the anthers at anthesis, finally exserted by 3-5 mm.
Fruit ovoid-triquetrous, 7-14 mm long. Figs. 75, 76. Plate 24. Map 13.
Plants now included in this species are distributed from Cape Peninsula, where the
typical form occurs, northwards to Khamiesberg and eastwards to about Barkly East.
It is found in seepage areas, marshy places and beside streams in fynbos of the south-
western Cape from near sea-level to 4,000 ft. flowering profusely after veldflres, and
in moist grassland in the eastern Cape Province from about 1,500 ft. in Albany District,
to over 6,000 ft. from Queenstown to Xalanga Districts. Flowering occurs in the
494
eastern Cape Province in autumn, from January to December, but in the south-western
Cape it may occur at almost any time from September to June, depending largely on
veld-fires, with two peaks, namely from October to December, and in April-May.
Fgi. 75. — K. uvaria, form with
large fruits, Malmesbury
District ( Bruce 227A).
Photo by E. A. Bruce.
Cape.— Without locality: Thunberg s.n. (BM, UPS); Oldenburg s,n. (BM); Burchell s.n. (K).
Albany: near Grahamstown, Hill s.n.; Comins 1020; Waainek, Jacot-Guillarmod s.n. ; near Atherstone
Station, Bruce 232. Alexandria: Ncanaha, Comins 1010. Bedford: Winterberg Mts., Story 3858;
Bruce 526 ; 527; MaphamlO. Bredasdorp: Potteberg, Barker 4548 (NBG). Caledon: near Grabouw,
Bolus 4216 (BOL); Swartberg, Schlechter 10367; Elgin, Marloth 11099; between Palmiet R. and
Hangklip, Pillans 8333 (BOL); Rivier Sonder End, Wilman 516 (BOL); Stokoe 7483 (BOL); Babylons
Tower, Esterhuysen 5008 (BOL); Genadendal, Hodge s.n.; Middlemost 1685; near Napier, Admiraal
2092. Cathcart: near Toise River, Codd 6354; near Thomas River, Bruce 545; near Elogsback,
Johnson 1211; Taylor 2494. Ceres: Elandskloof stream, Compton 16789 (NBG). Clanwilliam:
Paleisheuwel, Compton 20041 (NBG); Elandskloof, Lewis in SAM 58072 (SAM); between Citrusdal
and Elandskloof, Stokoe 7420 (BOL); 7133 (BOL); Cedarberg, Stokoe in SAM 58194 (SAM); Ester-
huysen 14340 (BOL); 18123; 18170; Langberg, Bond 1379 (NBG); Greys Pass, Godfrey SH 1253;
Lamberts Bay, Van Breda 3; near Paleisheuwel, Wedermann & Oberdieck 482. Graaff Reinet:
farm Konda, Bayliss 3917. Humansdorp: between Cambria and Combrink, Lewis in SAM 65735,
9 miles W. of Storms River Bridge, Codd 991 1 : Tsitsikama National Park, Nichol 1 ; 2. Keiskamma-
hoek: Ndola Mt., Story 4564. King William’s Town: Pirie, Godfrey s.n. (GRA). Knysna: Dieprivier,
Admiraal s.n. Laingsburg: Matjiesfontein, Pearson in NBG 2533/14 (BOL). Malmesbury: Bachmann
Plate 24. — Kniphofia uvaria (L.) Hook.
495
1229 (Z); 6 miles S. of Velddrift, Bruce 227 ; 221 A. Montagu: Montagu Pass, Hutchinson 1184 (K,
BM). Murraysburg: Englishmans Kraal, Tyson 84 (GRA). Namaqualand: Khamiesberg, Pearson
6249 (BM, BOL, K): 6320 (K); Leliefontein, Thorns s.n. (NBG). Oudtshoorn: Taylor s.n. (BOL,
GRA); Kamanassie Mt., Zinn s.n. (SAM). Paarl: French Hoek Pass, Barker 4914 (NBG). Peninsula:
Hout Bay, Thocle A59; Kirstenbosch, Esterhuysen 7620 (BOL); 21340; 24685 (BOL); Barker 1088
(NBG); 2948 (NBG); Klaassenbosch, Marloth 5896; near Kommetje, Compton 14017 (NBG); Bruce
223; between Millers Point and Smitswinkel, Worsdell s.n. (K); near Smitswinkel Bay, Galpin 12832;
Pillans 7609 (BOL); Isaac in BOL 24680 (BOL); Orange Kloof, Wolley Dod 2444 (K); Schusters
Kraal, Bruce 222; Steenberg Plateau, Salter in BOL 24682 (BOL); near Rhodes Memorial, Isaac
s.n. Piquetberg: Veloren Vlei, Barker 2618 (NBG); Leighton 198 (BOL); Pillans 8067 (BOL); 8652
(BOL); Kapiteins Kloof Mt., Pillans 8076 (BOL). Port Elizabeth: Bethelsdorp, Parmw; 669 (GRA),
Queenstown: Hangklip Mt., Galpin 1810; 17^ miles S.W. of Whittlesea, Marais 880. Riversdale:
near Albertinia, Muir 917; Garcias Pass, Thorne in SAM 38914 (SAM); Admiraal 1001. Somerset
West: Viljoens Pass, Strey 2897. Stellenbosch: Velorekloof, Van der Merwe 23/65: Van Rensburg
271. Steynsburg: Zuurberg, Schonland 3246a. Stockenstroom: 41 miles S. of Queenstown, Erens
2187; between Fort Beaufort and Post Retief, Story 4027: Katberg Pass, Elfers 3. Swellendam:
Swartberg, MacOwan s.n. (K); Warmbad, Little Karoo, Hutchinson 1136 (BOL, K); Leeuw River
Mts., Stokoe 8519 (BOL); Lemoenshoek Peak, Esterhuysen 14419 (BOL); Zuurbrak Peak, Thorne
in SAM 50163 (SAM); National Bontebok Park, Liebenberg 7148. Tulbagh: Bokkeveld Kloof,
Saron, Compton in NBG 1931/36 (BOL). Uitenhage: near Swartkops River, Pappe in SAM 22597
(SAM); Hilwacht, Admiraal 466. Vanrhynsdorp: Ebenezer, Compton in NBG 2082/36 (BOL). Wor-
cester: Bosjesveld Mts., Stokoe in SAM 58195 (SAM); 20 miles S. of Worcester, Andreae 330; Bothas
Halt, Van Breda 202 ; Brandwagt, Van Breda B; Dutoitskloof, Van Breda s.n. : Bruce 226; Stormsvlei,
Van Breda 200. Xalanga: Cala Pass \Marais 878.
Fig. 76. — K. uvaria, the form found in Albany and King William’s Town Districts (Comins 1010).
496
Four forms are recognisable, but they grade into one another and it has not been
possible to define them adequately for separate taxonomic status, even at infraspecific
level. They are outlined below.
(a) Typical K. uvaria occurs in the Peninsula and surrounding districts. The
leaves are erect to somewhat arcuate and relatively soft. The inflorescences are medium-
sized, ovoid to oblong, and brightly coloured (buds brilliant scarlet, grading down to
orange yellow as the flowers open). The pedicels are fairly conspicuous (2-4 mm long).
Towards the east, specimens are recorded which dry with hard, fibrous leaves (e.g.
Worcester: Botha’s Halt, Van Breda 202; Brandwagt, Van Breda B; Humansdorp:
Baviaanskloof Mts., Lewis in SAM 65735) and these form a link with forms (c) and
{d) below. Occasional specimens have abnormally short flowers (Oudtshoorn: Taylor
s.n.; Zinn s.n.).
(b) From Malmesbury District K. bachmannii Bak. was described and from this
district northwards the plants have somewhat longer and laxer inflorescences, but
still brightly coloured, with conspicuous pedicels (3-5 mm). The leaves are similar
to those of the typical form, but the fruits are exceptionally large (12-14 mm) in the
few fruiting specimens seen.
(c) From Humansdorp to Albany Districts the plants are medium to small in
stature with leaves often narrower than usual, erect to arcuate or recurved and fairly
tough in texture. The inflorescences are oblong to globose, varying a good deal in
colour from brilliant red to greenish yellow, while the pedicels are not very conspicuous
(2-3 mm long). A representative of the smaller plants ( Paterson 669) was listed without
description as K. patersoniae by Schonland in Bot. Surv. S. Afr. Mem. 1: 38 (1919).
This form, now included in K. uvaria, grades to some extent into K. citrina (p. 497),
which is closely related and may even be considered hardly worthy of being placed
as a separate species. It is, however, kept distinct on the basis of the short perianth
(2. 2-2. 7 cm long) and the markedly exserted stamens.
(d) Further to the north and north-east, and more inland at higher altitudes’
usually on rocky (especially sandstone) formations, a form of medium stature occurs
with arcuate, often very glaucous, tough leaves. The inflorescences are oblong or
subglobose, fairly large, somewhat reminiscent of K. rooperi and, like that species,
may vary from brilliant red to greenish-yellow tinged with red, or sometimes pure
yellow ( Johnson 1211; Taylor 2494). The bracts are oblong-ovate, longer than in
typical K. uvaria, and the pedicels are shorter (1-5-3 mm). Morphologically, this
form occupies a position between K. uvaria and K. rooperi, but it is separated from both
geographically. This form may also be confused in the herbarium with specimens
of K. linearifolia, but it can usually be recognized by the hard, fibrous and somewhat
narrower leaves. Separate taxonomic status was considered for this form but insufficient
support for such a step could be found.
As indicated on p. 370, K. uvaria is closely allied to plants now included in K.
linearifolia (p. 498) and, where the two species overlap in the eastern Cape Province,
some difficulty may be experienced in separating the two with certainty. The latter
species is generally more robust, with broader, softer leaves that do not tend to fold
along the midrib on drying, larger inflorescences and short pedicels.
The types of K. burchellii and K. occidentalis fall within the range of the typical
form of K. uvaria. As stated on p. 370, Baker in Flora Capensis (1896) and Berger
(1908) applied the epithets alooides or uvaria to the more robust species now dealt
with under the names K. praecox and K. linearifolia. The varieties upheld by the two
monographers are also allocated to these two species.
497
Map 13.— Distribution of Kniphofia rooperi,
K. uvaria and K. citrina.
Map 14. — Distribution of Kniphofia linearifolia-
K. tysonii subsp. tysonii and subsp. lehom,
boensis.
43. Kniphofia citrina Bak. in Gard. Chron. 14: 552 (1893); FI. Cap. 6: 278
(1896); Mallett in Gard. Chron. 39: 82 (1906); Berger in Pflanzenr. 4, 38 : 57 (1908);
Codd in Flow. PI. Afr. 37: 1. 1 448 (1965). Type: Hort. Leichtlin, Oct. 1893, “ from
mountains north of Grahamstown ” (K!, holo.).
K. linearifolia sensu Phillips in Flow. PI. S. Afr. 9: t.340 (1929).
Plants in small groups, 40-60 cm tall. Leaves 8-12 per peduncle, narrow, suberect
to arcuate, 40-70 cm long, 6-10 (-12) mm broad, keeled, V-shaped in cross section,
tending to fold along the midrib, somewhat glaucous, strongly nerved, breaking into
fibres at the base; margin minutely and sparsely denticulate to smooth. Peduncle
standing above the leaves, 40-60 cm long. Inflorescence subglobose to globose, dense,
4-5-5 cm long and 5-5-5 cm in diameter; buds erect to spreading, red, red-tipped or
yellow; flowers pendulous, yellow to yellow-green. Bracts ovate-oblong, 4-5-5 mm
long, 2-2.5 mm broad, subacute to obtuse; margin finely eroso-denticulate. Pedicels
1-5-2 -5 mm long. Perianth subcylindrical, 2 -2-2 -7 cm long, scarcely constricted
above the ovary and 2 mm broad for about 1 cm, then widening gradually to 4-5 mm
in diameter at the throat; lobes broadly ovate, 2-5 mm long, spreading. Stamens
exserted by 4-6 mm at anthesis and remaining exserted. Style subequal to the stamens
at anthesis, finally exserted by 8 mm. Fruit globose-triquetrous, about 8 mm long.
Plate 25. Map 13.
Found in dense grass from Flumansdorp to Bathurst districts and, inland, to
Albany and King William’s Town Districts, from near sea level to 2,000 ft. in altitude.
The flowering time is mainly March to May.
Cape. — Albany: Mountains north of Grahamstown, Hort. Max Leichtlin (K); Highlands Station,
Story 4542. Alexandria: near Bushman R. Mouth, Galpin 11809; Archibald 4222. Bathurst: farm
“ Brooklands ”, Comins 1018: 1019; farm “ Cowley ”, Comins 2032. Humansdorp: near Humans-
dorp, Marais 380. King William's Town: Evelyn Valley, Storv 3814. Port Elizabeth: Parsons Vlei,
Long 1241.
The essential characters of K. citrina are its relatively small stature, the arcuate,
somewhat fibrous leaves, the small globose inflorescences which appear in autumn,
with short flowers and markedly exserted stamens. As stated on p. 496 it is closely
related to plants from much the same area which are now included in K. uvaria and
intermediate specimens may be found, for example: Comins 1015 from between
Grahamstown and Port Elizabeth (perianth 3 .0 cm), Urton s.n. from near Port Elizabeth
(perianth 2.8 cm) and Story 3848 from Keiskammahoek (perianth also 2.8 cm).
Further study in the field is required in order to determine the status of K. citrina;
in the meantime it appears worthy of separate rank.
498
Inflorescences may vary in colour from clear yellow to bright red at the apex,
often with a varnished appearance, grading to yellow flowers, or dull red at the apex
grading to greenish-yellow flowers.
44. Kniphofia linearifolia Bak. in Bot. Jahrb. 15, Beibl. 35: 5 (1892); FI. Cap.
6: 282 (1896); Berger in Pflanzenr. 4, 38: 58 (1908). Type: Pondoland, Bachmann
279 (Bf, holo.); neotype: Cape, Idutywa, Cocid 9262 (PRE!).
Tritoma uvaria sensu Redoute, Lil. t.291 (1810).
Kniphofia uvaria sensu Hook, in Bot. Mag. t.48 1 6 (1854), non (L.) Hook; sensu
Berger, l.c. 61 (1908), partly. -var. maxima Bak. in Bot. Mag. t.6553 (1881);
Berger, l.c. 61 (1908). Type: Hort. Kew Nov. 1879, originally from Max Leichtlin
(K!, holo.). K. alodides sensu Van Houtte in FI. d. Serres 13: 187, 1. 1 393 (1858);
sensu Bak. in FI. Cap. 6: 283 (1896), partly. — - — var. maxima (Bak.) Bak. in FI.
Cap. 6: 283 (1896). K. rooperi sensu Hook, in Bot. Mag. t.61 16 (1874); sensu Berger,
l.c. 62 (1908), partly, as to Bot. Mag. t.6 1 16 and Scldechter 6518; sensu Wood, Natal
PI. 6: t.540 (1912). K. longiflora Bak. in Kew Bull. 1901: 134 (1901); Berger, l.c.
60 (1908). Type: “ Natal, probably ”, Hort. Kew Nov. 1898 from seeds communicated
by Mrs. Watkins, Euston Square, London (Kl, holo.). K. linearifolia var. montana
Berger, l.c. 58 (1908). Syntypes: East Griqualand, Insizwa Mt., Krook 548 (W |);
Tembuland, Umtata, Krook 368 (W f)- — - — var. kuntzei Berger, l.c. 58 (1908). Type:
Natal, near Colenso, Kuntze (K!, Z !). K. rhodesiana Rendle in J. Linn. Soc. Bot.
40: 214 (1911). Type: Rhodesia, Nyahodi River, Swynnerton 723 (BM!, holo.).
K. praecox sensu Letty, Wild Flows. Tvaal. 1. 1 8 (1962).
Fig. 77. — K. linearifolia from Nqeleni District, eastern Cape Province (Code/ 9280).
Plate 25. — Kniphofia citrina Bak.
499
Plants robust, usually in groups, up to 1.5 m tall. Leaves 8-16 per peduncle,
fairly broad, at first erect, usually folding back or bending with their weight towards
maturity, 60-140 cm long, 1 - 2—2 - 8 cm broad, strongly keeled, fairly rigid to flaccid
intexture, yellow-green to dull green, rarely subglaucous; margin serrulate to sparingly
serrulate or, occasionally, smooth. Peduncle standing above the reflexed leaves, 80-150
cm long. Inflorescence ovoid or oblong to rhomboid, dense, 6-16 cm long, 5- 5-6- 5
cm broad, usually tapering towards the apex; buds spreading to pendulous, pinkish-red
to green tinged with dull red; flowers pendulous, yellow to yellow-green or greenish.
Bracts broadly ovate to ovate-oblong, 4-10 mm long, 3-4 mm broad, subacute or obtuse
to rounded; margin usually finely serrulate. Pedicels very short, 1-1-5 mm long.
Perianth subcylindrical, 2 -5-3 -5 cm long, scarcely constricted above the ovary and
2-5 mm broad near the base, widening gradually to 5 mm broad at the throat; lobes
broadly ovate, 2-3 mm long, sometimes spreading. Stamens exserted by 4-5 mm at
anthesis, later almost or completely withdrawn. Style subequal to the stamens at
anthesis, finally exserted by 8 mm. Fruit ovoid-triquetrous, about 8 mm long. Figs.
77, 78, 79, 80. Plate 26. Map 14.
Found at altitudes of 1,000 to about 6,500 ft. in eastern and north-eastern Cape
Province (with an outlier at Worcester), eastern Orange Free State, Natal, the moun-
tainous regions of Swaziland, eastern Transvaal and probably eastern Rhodesia, in
mountain grassland, marshy places and on stream banks. The main flowering period
is January to March, but specimens may be found in flower in the eastern Transvaal
until April.
Without Locality: Hort. Saunders Nov. 1872, type of Bot. Mag. t.6116 (K); Hort. Kew Nov.
1879, originally from Max Leichtlin, type of Bot. Mag. t.6533 (K); Hort. Kew Oct. 12, 1886 (K).
Cape. — Aliwal North: Elandshoek, F. Bolus 199. Bedford: “ Smalldeel ”, Burke s.n. (K). Bizana:
Zuurberg, Codd 9342. Cathcart: near Thomas River, Codd 6355; Bruce 546; near Hogsback, Bruce
536: 537; 8f miles N. of Hogsback, Bruce 538; Kingsnek, near Hogsback, Erens 2216; Windvogel-
berg, Marais 515. Elliotdale: Marais 1009. Flagstaff: 10f miles N.W. of Holy Cross Mission, Codd
9329; 6 miles N.E. of Flagstaff, Codd 9332; 13 miles N. of Bizana, Codd 9334. Graaff Reinet: 26 miles
S.W. of New Bethesda, Kitching 55. Idutywa: near Idutywa, Codd 9262; 5 miles N. of Idutywa,
Codd 9263. Keiskammahoek: Bruce 553; Cata Forest Station, Story 3794; Bruce 552. Kentani:
Abernethy A; 2 miles N.W. of Nqunduza Store, Codd 9256. Libode: 9 miles W. of Libode, Marais
1006. Lusikisiki: Bruce 596; near Magwa Falls, Codd 9316; 9322; Mkambati, Marais 974; 7 miles
S. of Mkambati, Marais 994. Maclear: Between Naudes Nek and Maclear, Bruce 579; 2\ miles
S.W. of Maclear, Marais 871; 5 miles S.W. of Ugie, Marais 876; Gatberg, Marais 877; Naudes Nek,
Marais 1370; 28 j miles from Rhodes to Maclear, Marais 1379. Matatiele: near Paspalum, Marais
858. Middelburg: Potts 1704 (UOFS); between Compassberg and Rhenosterberg, Drege s.n., partly
(K); Grootfontein, Theron 160; s.n. Mqanduli: near Mqanduli, Codd 9264; 9265; Marais 1012.
Mt. Ayliff: Mt. Insizwa, Schlechter 6518 (Z); 9 miles W. of Mt. Ayliff, Codd 8558; 5f miles W. of
Old Goxe Store, Marais 851; 9 miles S. of Kokstad, Marais 852; 3 miles S. of Connomore, Marais
854. Mt. Fletcher: Bruce 580. Nqeleni: Codd 9280; Killick 2036. Port St. Johns: 5 miles N. of
Port St. Johns, Codd 9304. Queenstown: Galpin 1786. Somerset East: Boschberg, MacOwan 1908
(GRA, K, Z); Galpin 9287; Bruce 519; 522; Oudekraal, Admiraal s.n. Sterkstroom: Stormberg
Mts., Bruce 564; Halseton, Stretton 6. Stockenstroom : 12 miles N. of Deelkraal, Marais 882. Tarka-
stad: 9 miles S.E. of Spring Valley, Bruce 524. Umtata: Baziya, Bruce 600 ; Killick 2064. Umzimkulu:
10 miles S. of Umzimkulu, Codd 8564. Wodehouse: 10 miles N.W. of Dordrecht, Bruce 561 ; 13 miles
N.W. of Dordrecht, Bruce 562; 6 miles S. of Jamestown, Reynolds s.n. Worcester: Bothas Halt,
Van Breda 210.
O.F.S. — Without locality: Cooper 3599 (K, Z). Harrismith: Oliviershoek Pass, Bruce 400; Rens-
burgskop, Jacobsz 305; Codd 10524; Walton, Jacobs: 304; Kerkenberg, Jacobs: 307 ; Van Reenen's
Pass, Codd 8515.
“ Basutoland ”. — Cooper 3234 (K, Z).
Natal. — Without locality: Hort. Kew Nov. 1898, from Mrs. Watkins (K). Alfred: 2\ miles W.
of Harding, Marais 846. Bergville: Royal Natal National Park, Trauseld 251; Cathedral Peak area,
Bruce 416; Killick 1333; 1407; 1465; Codd & Dyer 6252; 6289; Germishuizen 43; 44. Eshowe:
near Mbongolwane, Strey 4189. Estcourt: Colenso, Kuntze s.n. (K, Z.); Cathkin Park, Galpin 11831 ;
near White Mt. Inn, Killick 2162; Giants Castle Game Reserve, Trauseld 372; Highmoor Forest
Station, Killick & Vahrmeijer 3669. Impendhle: 27 miles S.W. of Nottingham Road, N. R. Smuts
1037; 20 miles N. of Bulwer, Codd 8534. Ixopo: 3£ miles N. of Highflats, Codd 9367. Klip River:
24379-11
500
20 miles from Ladysmith on road to Van Reenen, Germishuizen 2. Kranskop : 2 miles N. of P.O.
The Ranch, Cocld 9668. Lions River: 10 miles W. of Nottingham Road, Dyer 5318; near Caversham,
Edwards 2303. Ngotshe: Ngome, Bruce 298. Polela: Bulwer, Codd 8539 ; 3 miles from Donnybrook
on Polela Road, Edwards 3108. Underberg: 12 miles S. of Underberg, Codd 8546. Utrecht: Groenvlei
area, Bruce 284; 285; 286; Codd & Dyer 6258; 6266; Codd 6951; 5 miles N.E. of Utrecht, Bruce
287; Maraisl41l; 2 miles S.E. of Viljoenspos, Codd 6957. Vryheid: 17 miles N. of Vryheid, Bruce 291.
Fig. 78. — K. linearifolia from Utrecht District, X f ( Codd & Dyer 6258).
501
Swaziland. — Near Mbabane, Nicholson s.n.; Bruce 305; Dlamini s.n.; 5| miles S. of Forbes Reef,
Bruce'iOl', 7 miles N.E. of Forbes Reef, Bruce 309; Piggs Peak, Compton 30626; 5 V miles E. of Havelock
Mine, Bruce 310.
Transvaal. — Amersfoort: 5 miles S.W. of Dirkiesdorp, Bruce 278; 10 miles E. of Wakkerstroom,
Bruce 280. Barberton; near Barberton, Repton 1000; Saddleback Range, Galpin 1230; Nelsberg,
Van der Merwe 1672; Bruce 312; Meeuse 10090; Berlin Forest Station, Codd 5593; 2 miles N.W,
of Havelock Mine, Codd 7845. Belfast: 2 miles E. of Airlie Station, Vermeulen s.n.; near Machado-
dorp, Van der Merwe 1249; Bruce 330; 338; Schoemanskloof, Bruce 337. Carolina: Waterval Boven,
Van der Merwe 901; 4y miles E. of Lochiel, Bruce 304; Slaaihoek, Bruce 504; 5 miles S.E. of Grootkop,
Codd 9559. Ermelo: 25 miles E. of Ermelo, Bruce 265; O'Connor s.n. ; Iswepe, Sidey s.n. Lydenburg:
near Lydenburg, Wilms 1558; Galpin 13718; 13 miles S.E. of Lydenburg, Bruce 323; 8^ miles S.W.
of Lydenburg, Bruce 324; 12 miles W. of Lydenburg, Bruce 326. Nelspruit; near Kaapschehoop,
Rogers 21053; Bruce 331. Pilgrims Rest: Rogers 18717 (K, Z); Sabie Valley, Galpin 13607; farm
London, Galpin 14516; near Sabie, Van der Merwe 1666; Munro s.n.; Brent 41; near Graskop,
Van der Merwe 1291; Codd 4307; Bruce 318; 319; Brent 33; 4V miles W. of Pilgrims Rest, Bruce
320; Bushbuckridge, Verdoorn 2453; Mariepskop, Codd 7878; 7922; Van der Schijff 5 111; Verdoorn
2456. Piet Retief: 16 miles E. of Moolman, Codd 7037. Wakkerstroom: Van der Merwe 29; Galpin
10014; Oshoek, Devenish 592; 1010.
Rhodesia. — Melsetter District: Nyahodi River, Swynnerton 723 (BM); farm Albany, Crook 544;
Pasture Research Station, Crook 545.
Fig. 79. — K. linearifolia, near Mqanduli, eastern Cape Province (Codd 9265).
502
The somewhat lengthy synonymy is partly due to the fact that the name K. uvaria
came to be applied mainly to the more robust plants of the eastern Cape Province
now included in K. linearifolia. It is also apparent that previous monographers of the
genus were not clear in their concepts of species such as K. uvaria, K. rooperi, K. praecox,
K. linearifolia etc. with the result that too many species names were upheld.
K. praecox (p. 447), which may superficially resemble the large inflorescences of
K. linearifolia, is distinguished from the latter on the grounds of the lanceolate,
acuminate bracts. K. uvaria, K. rooperi and K. linearifolia have ovate-oblong, rounded
to subacute bracts, and form a closely related complex. Although intermediates are
found between the three species, it is considered that over ninety per cent of the ample
modern gatherings can be allocated with a high degree of certainty, so no justification
is seen for a broader view which would include the entire range within one species.
As here defined, K. linearifolia is the most widespread species in South Africa.
The plants are robust with large, showy inflorescences which are ovoid or oblong,
but usually not subcylindrical (see K. tysonii, p. 504). The leaves are fairly broad and
usually soft in texture so that they tend to bend over about the middle, or lower, often
at a sharp angle. Variation is mainly in stature (due largely to growing conditions),
size and colour of inflorescences, and width and colour of leaves. Although the inflore-
scences are usually brightly coloured with pinkish-red at the apex, a colour form is
found in Natal ( Bruce 406, Code! & Dyer 6252, 6289 etc.) in which the buds are green
(sometimes tinged with dull red) and the flowers are greenish-yellow. This colour
form is associated with leaves that are more glaucous than is generally the case, and
is not easily distinguishable from the late-flowering specimens of K. baurii listed on
p. 482.
The species appears to have been introduced into Europe in the early part of the
19th Century, judging by the illustration by Redoute (1810 or 1811). The illustrations
in Bot. Mag. t.48 1 6 (1854) and Flore des Serres 13: 1. 1 393 (1858) also probably belong
here, but it is necessary to refer back to the statement made on p. 3/0 when dealing
with K. praecox, that it is often not possible to distinguish between illustrations of
K. praecox and K. linearifolia without supporting herbarium specimens. N. E. Brown’s
apt comment on hybridization in gardens (see p. 393) should also be remembered.
Due to the absence of corresponding herbarium specimens, it is possible to ignore
such semi-horticultural epithets as Tritoma grandis (1865), T. saundersii (1882) and
T. nobilis (1882, 1885), even if it is claimed that the descriptions are sufficiently adequate
for the names to be taken into account. They could equally well apply to K. praecox
as to K. linearifolia, and no corresponding specimens have been traced.
The plant illustrated as K. rooperi in Bot. Mag. t.6 1 1 6 (1874) belongs rather in
K. linearifolia and was grown in the garden of Mr. W. W. Saunders of Reigate, where
it flowered in November, 1872. Thomas Cooper, who collected for Mr. Saunders,
made two herbarium specimens of the species while in South Africa, Nos. 3234 and
3599. The former is recorded from “ Basutoland ”, though no other gathering is known
from the territory as defined at present.
The name K. uvaria var. maxima Bak. was applied to a plant of this species which
was illustrated in Bot. Mag. t.6553 (1881). It flowered at Kew in November, 1879,
and came from Max Leichtlin. Apart from this, its origin is obscure.
K. linearifolia Bak. (1892) appears to be the earliest available specific name for the
widespread species now under discussion. There appears to be no type material extant
and thus some explanation of its interpretation and the reasons for selecting Codd
9262 as the neotype would be appropriate.
nO~~
Plate 26. — Kniphofia linear if olia Bak.
503
Fig. 80. K. Unearifolia,
Ermelo District. Photo
by Miss O’Connor.
Baker described three species in 1892, based on Bachmann gatherings in Pondoland,
namely: K. Unearifolia ( Bachmann 279, no date recorded); K. decaphlebia ( Bachmann
280, February, 1888); and K. drepanophylla (Bachmann 281, August, 1888). A specimen
of the last mentioned exists in Kew Herbarium, so the identity of K. drepanophylla
is not in doubt. No material has been traced, however, of the types of K. Unearifolia
and K. decaphlebia.
Bachmann published an account of his travels (1901), from which his route in
Pondoland can be roughly traced, but he makes no mention of collecting Kniphofia
specimens. A fairly thorough study of the area was, therefore, made during 1955
and 1956 in order to investigate the species that Bachmann is likely to have encountered.
Two autumn-flowering species were found to be widespread in Pondoland. One of
these is K. laxiflora Kunth which may have a relatively dense inflorescence and to
which the description of K. decaphlebia Bak. agrees quite well. The other is the species
to which the name K. Unearifolia Bak. is now applied.
Reference to the description of K. Unearifolia indicates that Bachmann 279 had a
smaller inflorescence (5-7 cm) and shorter flowers (25-30 mm) than is normal for the
specimens now cited in K. Unearifolia. This reduction in size is found on plants growing
under relatively dry conditions. A good example of this is Codd 9262. The original
504
plant found growing a little distance from a stream near Idutywa (in the Transkei)
had inflorescences about 6 cm long, flowers 27 mm long and leaves 11 mm broad.
In cultivation in Pretoria, the same plant developed inflorescences about 10 cm long,
flowers 30 mm long and leaves 2 cm broad. The wild specimen is, therefore, selected
as the neotype.
The type of K. longiflora , cultivated at Kew of unknown origin, as well as the types
of K. linearifolia var. kuntzei Berger and of K. rhodesiana Rendle are all very scrappy
specimens but are considered to fall within the range of variation now included in
K. linearifolia.
45. Kniphofia tysonii Bak. in J. Bot. Lond. 27 : 43 (1889). Type: East Griqualand,
Zuurberg, Tyson 1709 (K!, holo.; BOL!, SAM!).
Plants in groups, robust, up to 2 m tall. Leaves 8-16 per peduncle, at first erect
but later folding over at about the middle, 80-150 cm long, 12-40 mm broad, yellow-
green to dull green or subglaucous, strongly keeled, relatively flaccid in texture; margin
serrulate to smooth. Peduncle overtopping the reflexed leaves, 1-2 m tall. Inflorescence
oblong to subcylindrical, dense, 7—1 5 cm long, 4-5-6 cm broad; buds spreading to
Fig. 81. — K. tysonii subsp,
tysonii , from The Dargle
{Mrs. Forssman s.n.).
mlnn 1 1 1 1 1 1 ITT 1 1 TTrrt i r \ timihi n trrrrtn rrmtittifi inn I iimmroxa
505
Fig. 82. — K. tysonii subsp. tysonii, form from Zululand (Cudd 7000).
506
pendulous, orange-red to dull blood-red; flowers spreading to pendulous, yellow to
greenish. Bracts oblong-ovate to oblong, obtuse to rounded 4-8 mm long, 2-5-3 mm
broad; margin finely serrulate. Pedicels almost obsolete, about 1 mm long. Perianth
subcylindrical, scarcely constricted above the ovary, 2 -0-2 -8 cm long, 2-5 mm broad
near the base, increasing to 4 mm at the throat; lobes broadly ovate, 2-5 mm long,
scarcely spreading. Stamens exserted by 6-9 mm at anthesis and remaining exserted.
Style subequal to the anthers at anthesis, finally exserted by 1 cm or more. Fruit
subglobose-triquetrous, 4-5 mm long.
Two subspecies are recognized.
Leaves relatively broad (up to 3-5 cm), usually serrulate; buds usually spreading, orange-red;
flowers spreading to pendulous, yellow to yellow-green subsp. tysonii
Leaves T2-2 cm broad, usually smooth or almost so; buds imbricate, dull blood-red; flowers
pendulous, greenish subsp. lebomboensis
(a) subsp. tysonii.
K. tysonii Bak. in J. Bot. Lond. 27: 43 (1889); FI. Cap. 6: 279 (1896); Berger in
Pflanzenr. 4, 38: 56 (1908).
Characterised by the broader (1-5-3- 5 cm), usually yellow-green leaves, larger
and more oblong inflorescences with buds spreading, orange-red to pinkish-red. Figs.
81, 82. Plate 27. Map 14.
Recorded from East Griqualand in the Cape to Nkandla District in Natal, in dense
tall grass on hillsides and at forest margins, from near sea level in the Port Shepstone
District to over 4,000 ft in the Natal midlands.
Cape. — District unknown: Zuurberg, Tyson 1709 (BOL, K, SAM). Umzimkulu: Kromhoek
Marais 839.
Natal. — Alfred: near Staffords Post, Marais 842. Camperdown: Inchanga, Eshuis s.n. Enton-
janeni: 2 miles S. of Melmoth, Codd 7003. Ixopo: 8 miles S.E. of Ixopo, Codd 8570. Lions River
near Cedara, Bruce 158; The Dargle, Letty 272; Marais 816; Lidgetton, Mogg 3596. Ndwedwe
Inanda, Medley Wood 1330 (K). Nkandla: 10 miles S. of Nkandla, Codd 6977 ; Nkandla Forest,
Codd 6971; 9693; Qudeni, Codd 6990; 6998; 7000. Nqutu: 2} miles E. of Nqutu, Codd 6790.
Pietermaritzburg: Wilms 2309 (K); Town Hill, Dyer 5325; near Pietermaritzburg, Codd 6792; 14
miles S. of Pietermaritzburg, Marais 838. Pinetown: Kranskloof Falls, Bruce 153. Polela: 3 miles
W. of Donnybrook, Gerstner 7011. Port Shepstone: 5 miles S. of Port Shepstone, Codd 9353;
Mehlomyama, Codd 9364. Richmond: Nels Rust, Killick 2106. Umvoti: 9 miles S. of Rietvlei,
Codd 8589; between Sevenoaks and Mt. Alida, Marais 805. Umzinto: 4 miles E. of Dumisa,
Codd 8572.
K. tysonii subsp. tysonii is closely related to K. linearifolia and intermediate
specimens are encountered which may be difficult to allocate with certainty. The
cultivated specimens Codd 9277 from near Umtata and Bruce 445 are examples of
this, but here it must be remembered that plant organs, such as leaves and flowers,
may be amplified or reduced by the effects of cultivation. In general, subsp. tysonii
has more elongate inflorescences with shorter flowers and long-exserted stamens.
The general colour is orange rather than bright red and, in this connection, a problem
is posed by the specimens recorded from the Zululand districts of Melmoth, Nkandla
and Nqutu. In these, the buds are pinkish-red, resembling K. linearifolia, while the
individual flowers and pedicels are longer than in typical subsp. tysonii. Closer investi-
gation may prove that they should be accorded separate taxonomic rank, and possibly
be included in K. linearifolia rather than in K. tysonii. The other alternative, to merge
K. linearifolia (1892) in K. tysonii (1889), does not commend itself at the present stage
of our knowledge.
(b) subsp. lebomboensis Codd, subsp. nov., a subsp. tysonii foliis angustioribus,
marginibus levibus, gemmis pendulis differt.
Plate 27. — Krtiphofia tysonii Bak. subsp. tysonii
507
Type: Swaziland, Lebombo Mts., about 2 miles S.E. of Stegi, 6.3.51, Bruce 306
(PRE!, holo.).
Characterised by its narrower (1-2-2 cm) leaves which are distinctly flaccid, and
mid-green to subglaucous with smooth margins, and the inflorescence which tapers
to the apex with buds pendulous, imbricate, dull blood-red in colour, while the mature
flowers are greenish. Fig. 83. Map 14.
Fig. 83. — K. tysonii subsp. lebomboensis, from near Stegi, Swaziland (Erens & Dyer 2025).
508
Found in moist, grassy places or seepage areas on hillsides at altitudes of 1,500
to 2,500 ft. in Swaziland, where it occurs on the Lebombo Range, and northern Natal.
The flowering period is February to early April with the peak in March.
Natal. — Hlabisa: 3 miles S. of Hlabisa, Code/ 7009; Hluhluwe Game Reserve, Ward 2189. Non-
goma: 16 miles E. of Nongoma, Codd 7010.
Swaziland. — 3 miles from Ravelston on road to Stegi, Dyer & Erens 2025; 2 miles S.E. of Stegi,
Bruce 306; 3 miles S. of Stegi, Nel s.n.
Although the inflorescence shape and colour is distinctive and the leaves are on
the average narrower, with usually smooth margins, subsp. lebomboensis is not easy to
separate from subsp. tysonii in the herbarium. It is probable, however, that it was
derived independantly from K. linearifolia, rather than from subsp. tysonii , because
an occasional specimen intermediate in character with K. linearifolia is found in northern
Natal in the Ngotshe District (near Louwsburg, Bruce 300; Codd 7022; near Ngome,
Codd 7012), at altitudes of 4,000 to 4,500 ft.
Subsp. lebomboensis does not overlap in distribution with subsp. tysonii or with
K. linearifolia but, because of its short flowers with well-exserted stamens, it is convenient
to include this group in a broad circumscription of K. tysonii.
EXCLUDED SPECIES
The following names have been excluded from the treatment of South African
species for the reasons stated after each.
Knipliofia carinata C. H. Wr. in Bot. Mag. t.8545 (1914). Type: Hort. Kew (K!,
holo.). Although said to be a South African species of which seed was supposedly
sent to Kew in 1892 by Miss Ayliflf of Rose Cottage, Grahamstown, it is not matched
by any South African material and is considered to be a synonym of the Ethiopian
species, K. leichtlinii Bak. ex Hook.f. See Codd in J. S. Afr. Bot. 29: 149 (1963).
K. infundibularis Bak. in J. Bot. Lond. 23: 277 (1885); FI. Cap. 6: 277 (1896);
Berger in Pflanzenr. 4, 38 : 44(1908). Type: Hort. Kew about 1780, from the herbarium
of Bishop Goodenough (K!, holo.). This is a synonym of K. pumila (Ait.) Kunth
listed below, which is also considered to be an Ethiopian species.
K. primulina Bak. in FI. Cap. 6: 533 (1897). Type: “ Natal, Hort. Leichtlin,
flowered in the Temperate House at Kew, January, 1897 ”. No corresponding herbarium
specimen with this date, or annotated by Baker as K. primulina, has been seen. The
description mentions smooth-margined leaves; dense, oblong raceme, 3-4 in long;
pedicels very short; bracts small, ovate; flowers all pale yellow; perianth an inch long.
Although Baker aligns it to K. natalensis Bak. (which is a synonym of K. laxiflora
Kunth), the description differs a good deal from the latter complex and there appears
little possibility of ascertaining the correct identity of K. primulina.
It may be noted that the name K. primulina Bak. was linked, probably wrongly,
by N. E. Brown with the plant illustrated in Bot. Mag. t.7623 (1898) as “ K. longicollis
Bak ”. The figured plant (from the garden of W. E. Gumbleton of Ireland), of which
a herbarium specimen is preserved in Kew annotated in N. E. Brown’s handwriting
“ this is K. primulina Baker ”, is quite distinct from true K. longicollis Bak. and also
does not agree with the description of K. primulina Bak. except in the colour of the
flowers. The specimen annotated by N. E. Brown may be accepted as the type of the
Bot. Mag. plate 7623, having leaves with serrulate margins; subcylindrical inflorescence
6 inches long; long pedicels; long-acuminate bracts; and perianth an inch and a
quarter long. This can now be interpreted as a colour form of K. praecox Bak. (see
also p. 449, where a specimen ex hort. Gumbleton, originally from Max Leichtlin
under the name “ K. nobilis ”, was included in K. preacox). Berger, in Pflanzenr. 4,
509
38: 60 (1908), apparently accepted N. E. Brown’s annotation because his description
of K. primulina corresponds with Bot. Mag. t.7623 and the Gumbleton specimen, and
thus differs from the type description of K. primulina Bak.
K. pumila (Ait.) Kunth, Enum. PI. 4: 552 (1843). Type: a specimen in BM
labelled “ Hort. Kew 1781, Aletris pumila” (BM!, holo.). See Codd in J. S. Afr.
Bot. 29: 145 (1963) for synonymy where reasons are given for concluding that this
is an Ethiopian species. Regarding synonymy, it should be noted that Ve/theima
abyssinica Redoute should be included.
K. saundersii (Carr.) Nicholson, Diet. Gard. Suppl. 459 (1901). See Tritoma
saundersii Carr.
Notosceptrum aloides (Bol.) Benth. This proves to be a species of Aloe and the
combination A. alodides (Bol.) Van Druten was published in Bothalia 6: 544 (1956).
Tritoma. On pp. 387, 449, and 502 reference is made to certain semi-horticultural
names applied to various robust Redhot Pokers that have been introduced to gardens
from time to time. The majority of them are inadequately described and, even when
an illustration is presented, the botanical details necessary for correct identification
are lacking. The following names, taken up mainly from Index Kewensis may,
therefore, be excluded: T. glauca Hort., T. recurva Hort., T. recurvata Hort. (1863),
T. grandis Hort. (1865), T. saundersii Carr. (1882), T. nobilis Guill. (1882, 1885).
Veltheimia speciosa Roth: see p. 380 for a discussion of this name and reasons
for discarding it.
OPSOMMING
Die Suid-Afrikaanse Kniphofia soorte is hersien en ’n sleutel is opgestel vir die
45 erkende spesies en 4 subspesies. Die nuwe spesies, subspesies en kombinasies is:
K. ensifolia Bak. subsp. autumnalis, K. praecox Bak. subsp. bruceae, K. triangularis
Kunth subsp. obtusiloba (Diels ex Berger) ( K . obtusiloba Diels ex Berger), K. littoralis,
K. elegans, K. latifolia, K. tysonii Bak. subsp. lebomboensis.
INDEX TO COLLECTORS’ NUMBERS
Included in the list below are collectors whose specimens are represented normally
in several herbaria. No attempt has been made to list every collector cited in the
text. The species number is given in a bracket after the collector’s number and may
readily be identified by reference to the table of contents in the Introduction (p. 000).
ACOCKS, J. P. H., 11217; 11218; 11219 (15); 13362 (2); 13879 (7); 16843; 18982 (23);
21984 (2); 22014 (7).
BAYLISS, R. D„ 2749 (30); 3917 (42); s.n. (31a). BOLUS, H.. 4216 (42); 8259 (33); 8260
Ola); 8261 (31a); 10335 (2); 24692 (19). BRUCE, E. A., 24; 25; 26; 27; 28 (20a); 29; 30; 31;
32; 33; 34 (33); 35 (20a); 148A (22a); 153 (45a); 155 (35); 158 (45a); 159; 160; 163(15); 195
(24); 217; 218; 219; 220; 221 (20a); 222; 223; 226: 227 (42); 228 (33); 229; 230; 231 (20a);
232 (42); 245 (19); 258; 259; 260 (5); 265 (44); 266 (9); 267 (33); 272 (4); 273 (21); 274; 275
(9); 276 (34); 277 (9); 278 (44); 279 (15); 280 (44); 282 (9); 283 (33); 284; 285; 286; 287 (44);
288 (15); 289 (9); 290 (15); 291 (44); 292 (1); 293 (5); 294(15); 296 (15); 297 (31b); 298(44);
299 (15); 302 (33); 303 (21); 304; 305 (44); 306 (45b); 307; 309; 310(44); 311(30); 312(44);
313(1); 314(30); 315(21); 318; 319; 320 (44); 322 (5); 323; 324(44); 325 (33); 326; 330; 331;
337; 338 (44); 339 (5); 340 (20a); 343 (26); 344 (25); 360; 364; 365 (26); 366; 379 (25); 390;
393 (26); 395 (25); 400(44); 401 (31a); 406 (44); 407 (17); 409(24); 410(1); 413(21); 414(24);
417; 418; 419; 421; 422 (41); 423 (35); 424; 425; 428(41); 460 (39); 461 (38); 477; 484 (31b);
504 (44); 514; 516 (31b); 519(44); 520 (31a); 522(44); 523 (31a); 524 (44); 525 (25); 526; 527
(42); 528 (31a); 530 (29); 536; 537; 538 (44); 539 (29); 545 (42); 546 (44); 547 (2); 549 (12);
510
550 (2); 551 (31a); 552; 553 (44); 559; 560 (2); 561; 562 (44); 563 (28); 564 (44); 565; 567 (28);
569; 570 (25); 571 (31a); 572 (29); 573 (31a); 577 (2); 579; 580 (44); 581 (34); 588 (2); 593; 595
(15); 596 (44); 597 (2); 598; 599 (15); 600 (44); 601; 603 (41); 604 (22b); 605 (41).
CODD, L. E., 2532 (9); 2683 (5); 3188 (33); 3224(40); 3434 (33); 4307 (44); 4760 (33); 4774
(20a); 4878 (30); 5593(44); 5609 (38); 5636 (40); 5639 (34); 5641(40); 5767; 6153 (20a); 6166
(40); 6346 (41); 6355 (44); 6654 (42); 6655 (34); 6661; 6740(40); 6768 (34); 6770 (40); 6779 (10);
6782; 6785 (30); 6788 (38); 6790 (45a); 6791 (39); 6792 (45a); 6793 (41); 6796 (15); 6797; 6799
(30); 6951 (44); 6952 (1); 6953 (9); 6954 (33); 6955 (15); 6956 (9); 6957 (44); 6971; 6977 (45a);
6986 (7); 6989 (39); 6990 (45a); 6995 (14); 6998; 7000; 7003 (45a); 7004 (15); 7009; 7010 (45b);
7037 (44); 7592 (33); 7596 (40); 7625 (34); 7686 (10); 7693 (5); 7771 (1); 7845 (44); 7851 (1);
7878; 7922 (44); 8020; 8025 (20a); 8047(40); 8058 (20a); 8066(40); 8069 (33); 8247 (10); 8248
(9); 8274; 8308 (31b); 8509 (5); 8515 (44); 8518 (15); 8533 (17); 8534; 8539 (44); 8541 (15);
8546 (44); 8547 (31a); 8510 (20b); 8558 (44); 8560 (15); 8564 (44); 8570; 8572; 8589 (45a); 8764
(24) ; 8833 (20a); 9256; 9262; 9263; 9264; 9265; 9280; 9304 (44); 9305; 9307 (15); 9316 (44);
9317(2); 9319(15); 9322; 9329; 9332; 9334; 9342 (44); 9353 (45a); 9354 (15); 9364 (45a); 9365
(15); 9367 (44); 9521 (4); 9524(30); 9646 (15); 9668(44); 9693 (45a); 9699 (35); 9713 (36); 9717
(37); 9720 (36); 9728 (41); 9729; 9737 (36); 9738(37); 9674 (35); 9911 (42); 9978 (33); 10051
(1); 10077 (20a); 10191(14); 10267 (9); 10306 (31b); 10334 (1); 10336 (21); 10340 (1); 10520;
10521 (20b); 10522 (8); 10523 (20b); 10524(44); 10528 (31a); 10530(8); 10536 (31a); 10537; 10540;
10541; 10549 (26); 10651 (16). CODD, L. E. & BRUCE, E. A., 3381 (33); 3519 (20a). CODD,
L. E. & DE WINTER, B„ 3156 (20a); 3234 (34); CODD, L. E. & DYER, R. A., 6252; 6258; 6266
(44); 6285 (33); 6289 (44). CODD, L. E. & ERENS, J„ 2090; 3966 (24). COMINS, D., 1010 (42);
1018; 1019(43); 1020 (42); 1100; 1101(23); 1552 (22b); 2032 (43). COMPTON, R. H„ 12936
(19); 14017; 16789(42); 17036 (41); 20041(42); 21261 (25); 21266 (27); 22239 (23); 27186 (33);
27552 (4); 30528 (30); 30626 (44); 31290 (4); CONRATH, P„ 643 (20a); 644 (33). COOPER,
T„ 1029 (8); 3207; 3208 (33); 3234; 3599 (44).
DEVENISH, N. J., 153; 168 (34); 225; 298; 515(33); 559 (9); 592 (44); 608 (1); 721(34);
793; 1002 (33); 1010 (44). DIETERLEN, A., 174 (26); 474 (31a); 642 (25); 872; 1286 (32).
DYER, R. A., 397 (31a); 2307 (29); 3953 (5); 4867 (7); 5318 (44); 5325 (45a); 5335 (41). DYER,
R. A. & ERENS, J., 2025 (45a).
EDWARDS, D., 1186 (26); 2303; 3108 (44); 3334 (31a). ERENS, J„ 1941 (20a); 2187 (42);
2216 (44); 2249 (20a). ESTERHUYSEN, E., 2638 (19); 5008; 7620 (42); 9733; 14236 (19); 14340
(42); 14369 (19); 14419; 18123; 18170 (42); 18341(19); 21340; 24685 (42); 24694 (19). EVANS,
M. S., 353 (11); 411 (17); 649 (12); s.n. (7).
FLANAGAN, H. G„ 761 (41); 1805 (2); 1838; 1841 (31a); 2050 (33); 2250 (31a); 2261 (2);
2323 (22b); 2547 (15); 2712 (2).
GALPIN, E. E., 1208 (30); 1230(44); 1778 (31a); 1786 (44); 1810(42); 1862 (41); 1929 (25);
6864 (2); 6865 (25); 6866 (29); 9287(44); 9485 (34); 9613 (33); 10014(44); 11338 (33); 11347
(17); 11733 (29); 11740 (33); 11759; 11774 (17); 11809 (43); 11831 (44); 12832(42); 13127 (40);
13175 (33); 13275 (31b); 13607 (44); 13667 (25); 13717 (30); 13718 (44); 14083 (28); 14084 (31a);
14254 (25); 14501; 14510 (30); 14516 (44); 14835 (7). GERMISHUIZEN, P., 2 (44); 42 (11); 45;
46 (33); 47; 48(17); 51(15). GERSTNER, J„ 4326 (1); 6814; 6815 (35); 7005; 7006 (14); 7007;
7008 (15); 7011 (45a); 7012 (31a); 7025 (7).
HUTCHINSON, J., 1136 (42); 1137 (33); 1184 (42); 2782 (33).
JACOBSZ, M. L„ 182 (8); 286 (25); 290 (31a); 293 (33); 299 (1); 304: 305 (44); 306 (20b);
307 (44); 314 (25); 321 (34); ;322; 323 (26). JACOT-GUILLARMOD, A., 272 (25); 334 (32)
1238; 1307; 13 8 (27); 1673(32); 1671 (26); 1716; 1744; 1753 (28); 2256 (26); 4011 (32); 4350
(25) ; s.n. (42).
KILLICK, D. J. B„ 1265; 1305 (33); 1333(44); 1405 (17); 1407 (44); 1431; 1442 (17); 1465
(44); 1466; 1467 (17); 1640; 1642; 1653 (11); 1857 (26); 1867 (6); 1871 (25); 1872 (29); 2036;
2064(44); 2106 (45a); 2162 (44); 2219(10); 3223 (29); 3835 (15). KILLICK, D. J. B„ & MARAIS,
W„ 2000 (38); 2014 (37); 2016; 2024; 2034; 2042 (38); 2085; 2088 (34); 2099 (17); 2112(16);
2113(39); 2114(16); 2154 (34); 2155 (32); 2156 (34); zl57 (8); 2161 (17); 2163 (32); 2167(16);
2174 (26); 2175 (29); 2179 (25); 2212 (33). KILLICK, D. J. B. & VAHRMEIJER, .1.. 3547; 3550
(26) ; 3552 (11); 3690 (16); 3630 (6); 3640 (29); 3648 (16); 3669 (44); 3759 (26); 3771 (25).
LEENDERTZ, R., TRV 3689; TRV 4042 (33); TRV 4043 (20a); TRV 5544 (30); TRV 5695
(20a); TRV 5710 (20a); TRV 7711; TRV 9208 (33). LIEBENBERG, L. C. C., 3116 (34); 5760
(25); 6979 (26); 7120 (3); 7148 (42); 8047 (14). LOUW, W. J., 775; 1615 (20a); 1689 (5).
MACOWAN, P., 1536 (31a); 1908 (44). MARAIS, W., 51 (31b); 380 (43); 515 (44); 577
(22a); 805; 816 (45a); 822(15); 830 (31a); 838; 839 (45a); 840 (15); 842 (45a); 844 (15); 846;
851; 852; 854; 858; 871; 876; 877(44); 878; 880 (42); 882 (44); 929(34); 932 (34); 952 (36);
953 (37); 958 (38); 974; 994 (44); 1003 (38); 1006; 1009; 1012(44); 1014(29); 1018; 1020; 1021;
1028; 1032 (38); 1287: 1288 (26); 1309 (25); 1326; 1327 (26); 1346 (28); 1347 (25); 1349 (28);
1350 (31a); 1361 (28); 1370 (44); 1371 (31a); 1379 (44); 1382; 1388 (2); 1421 (15); 1429 (31a);
51 1
1446 (2); 1469 (29); 1471 (44); 1472 (9). MARLOTH, R„ 4366 (19); 5896 (42); 9609 (23); 10028
(19); 11099 (42). McCLEAN, A. P. D., 620; 621(15); 622 (8); s.n. (14). MEEUSE, A. D. J., 9856
(24); 9956 (31b); 10082 (30); 10089 (1); 10090(44); 10151(21); 10152(1); 10177; 10192; 10235
(24). MOGG, A. O. D„ 1357 (15); 1843 (29); 2455 (7); 3568 (15); 3596 (45a); 4493 (35); 6819
(7) ; 18991 (33); 20301 (24); s.n. (22a). MOLL, E. J., 1339 (7); 1367 (34); 3457 (7). MOSS, C. E„
10020; 10506; 13700; 17815 (20a).
PENTZ, J. A., 34 (29); 67(16); 426 (8). P1LLANS, N. S., 7609 ; 8067; 8076; 8333 (42). POLE
EVANS, 1. B., 3923; 4724 (31b); 4869 (17); 5700 (21). POTT-LEENDERTZ, R. (See Leendertz,
R.). PROSSER, S. W., 1506 (40); 1677 (33); 1790 (31b); 1878 (20a).
REHMANN, A., 4769 (20a); 5796 (13); 7024 (8); 8563 (35). REYNOLDS, G. W„ 578 (25);
1135(9); 1172; 1176 (31a); 2142 (20a); 2503 (2); 2506 (31a); 2650 (33); 3440 (15); 3462 (30);
3483 (21); 3551 (33); 3794 (31a); 4138: 4241 (30); 5771 (40); 5851 (14); 5866 (4); 5874; 5885 -
5895 (30). ROBERTS, B. R., 3387 (31a); 3430 (26); 3451 (33); 3452 (26). ROGERS, F. A., 18717
(44); 18198 (31b); 21053 (44); 24829 (20a); 27673 (42); 28148; 28274 (34); 30226 (23). RUDATIS,
H., 485 (35); 617 (15); 1469 (7): 1881 (14).
SCHEEPERS, J. C., 626; 925 (21); s.n. (1). SCHELPE, E. A., 488 (25); 1246 (8); 5187 (14).
SCHLECHTER, R., 817 (42); 3160(18); 6461 (2); 6462 (6); 6518 (44); 6662 (16); 6723 (7); 6970
(8) ; 10367 (42). SCHWEICKERDT, H. G., 696 (25); 697; 983 (33). SIDEY, J, L., 1553 (34);
1564 (9); 2056 (33). SMITH, C. A., 3511 (33); 5164 (20a); 8890 (7). SMUTS, N. R., 1022 (29);
1028/1; 1028/2 (17); 1037 (44); 1038; 1039; 1064; 1065 (17); 1066 (16): 1072 (17); s.n. (2); 1414
(17); 1420 (32). STOKOE, T. P., 201 (19); 7133; 7420; 7483; 8519 (42); s.n. (19). STORY, R.,
3743 (31a); 3754 (25); 3794(44); 3814 (43); 3858 (42); 4134(41); 4233 (37); 4542 (43); 4564(42);
6518 (24). STREY, R. G., 2897 (42); 4189 (44); 5840; 5860(41); 7374 (7); 7376; 7377; 7380 (14).
THODE, J., 59 (42); 62 (32); 213 (33); 520 (31a); 3333 (9); 4502 (8); 5677 (26); s.n. (31a).
TRAUSELD, W. R„ 165 (8); 174 (29); 251 (44); 341 (11); 372 (44); 372a (16); 498 (29); 582 (6).
TYSON, W., 84 (42); 1380 (31a); 1418 (2); 1709 (44); 1768 (15).
VAN BREDA, P. A. B., 3; 200 (42); 201 (44); 202; s.n. (42). VAN BREDA, P. A. B. &
ADMIRAAL, J., 2174 (22a). VAN DER MERWE, F. Z., 3 (33); 28 (9); 29 (44); 333 (30); 901
(44); 1095 (40); 1156 (26); 1249; 1276 (44); 1290 (21); 1291 (44); 1613 (40); 1648: 1665 (30);
1666 (44); 1671(1); 1672 (44); 1743 (20a); 1940(41); 2003; 2022 (20a) 2056 (33); 2390 (15); 2495
(7) ; 2830 (17). VAN DER SCHIJFF, H. P„ 4410 (21): 4518 (31b); 5111(44); 6524 (31b). VER-
DOORN, I. C., 1035; 2176 (20a); 2456 (44).
WARD, C. J., 2189 (45b); 3376 (14). WEST, O., 400 (33); 595 (16); 709 (29). WILMS, F.
1558(44); 1559 (31b); 2309 (45a). WOOD, J. MEDLEY, 636 (15); 978 (14); 1096 (18); 1330 (45a);
1972 (7); 2634 (35); 3753 (7); 3871 (15); 4409(14); 4572 (7); 4662 (18); 4762 (15); 4798 (7); 4826
(8) ; 4827 (14); 5188(10); 5238 (6); 5606 (8); 5972 (1); 6071 (25); 6334 (16); 6391 (6); 6652 (14);
7293 (18); 7431 (7); 8922 (8); 10367(14); 10495 (18); 11098 (14); 11424(15); 11591(14); 11628 (30).
512
INDEX
Names upheld for South African species are in bold type, while synonyms and
excluded species are in italics. Names of species from outside South Africa are in
ordinary type and personal names are in small capitals.
Page
Aiton, William 376
Aiton, William Townsend 380
A/etris L 394
pumila Ait 376,509
sarmentosa Andr 379, 453
uvaria (L.) L 376,493
Aloe L 394
uvaria L 367,376,493
longifolia Lam 3 76, 493
Baker, J. G 369,387
Berger, A 369,387
Bruce, E. A fig. 51; 387,452
Cooper, Thomas 386, 415, 445, 502
Excluded Species 508
Hybrids 392
Iris uvaria promont. bonae spei 374
Jacquin, N. J 379,381
Ker Gawler, J. B 380
Kniphof, J. H 367
Kniphofia Moench 367,379,394
acraea Code! pL 2 b; map 1 ; 405
albescens Codd . .fig. 24; pi. 5; map 3; 417
alodides Moench 369, 379, 493
var. maxima Bak 498
var. nobilis Bak 447
var. serolina Bak 393
bachmannii Bak 493
basutica ined 459
baurii Bak figs. 67, 68; pL 21; 482
brachystachya (A. Zahlbr.) Codd fig.
19; map 2; 410
breviflora Bak. figs. 21,22,23; map 3; 413
var. concinna (Bak.) Berger 413
buchananii Bak. fig. 20; pL 4a; map 3; 411
var . flavescens Berger 411,413
6iirc/ie///i (Herb.exLindl.)Kunth . ..394,493
carinata C. H. Wr 508
caulescens Bak .. .figs. 53, 54; map 9; 455
citrina Bak. fig. 3; pi. 25; map 13; 373, 497
comosa Hochst 376
concinna Bak 413
conrathii Bak 474
coralligemma E. A. Bruce fig. 52; pi. 12;
map 9; 387, 453
erassifolia Bak fig. 20; map 4; 424
Page
decaphlebia Bak 428
drepanophylla Bak pi. 20; 481
elegans Codd fig. 66; 480
ensifolia Bak 376,440
subsp. ensifolia fig. 43; pi. 9; map 7; 440
subsp. autumnalis Codd fig. 44; map 7 ; 442
var. albiflora E. A. Bruce 440
evansii Bak figs. 26, 27; map 4; 419
flammula Codd . .fig. 25; pi. 6; map 4; 418
fluviatilis Codd . .fig. 65; pi. 18; map 12; 476
fibrosa Bak fig. 28; map 4; 422
fibrosa sensu Wood 413
galpinii Bak p!. 14; map 11; 466
gracilis Harv. ex Bak. figs. 30, 31, 32;
pi. 7; map 5; 424
hirsuta Codd 460
ichopensis Bak. ex Schinz pi. 8; map 5; 433
infundibularis Bak 508
latifolia Codd figs. 69, 70; 484
laxiflora Kunth figs. 33, 34, 35, 36, 37;
map 6; 13, 428
linearifolia Bak. figs. 3, 11, 12, 77, 78, 79,
80; pi. 26; map 14; 5, 7, 373, 380, 498
var. kuntzei Berger 498
var. montana Berger 498
littoralis Codd pi. 19; 478
longicollis Bak 393,488
longiflora Bak 387,498
macowanii Bak 469
modest a Bak 403
var. lutescens Berger 424
var. woodii Berger 424
modesta sensu Bak 405,424
multiflora Wood & Evans figs. 13, 14;
pi. 1 ; map 1 ; 394, 400
natalensis Bak 428
var. angustifolia Bak 434
var. condensata Bak 428
var. robustior Berger 428
nelsonii Mast 437,466
northiae Bak. figs. 57, 58, 59; map 10; 464
obtusiloba Diels ex Berger 472
Occident alis Berger 387, 493
odor at a Heynh 493
pallidiflora Bak 438
513
Page
parvittora Kunth ..figs. 15, 16; map I ; 384,403
var. albifiora Berger 413
pater soniae Schonl 496
pauciflora Bak afigs. 41, 42; 438
pedicel lata Bak 438
porphyrantha Bak fig. 64; pi. 17;
map 12; 386,474
praecox Bak 370, 379, 386, 444
subsp. bruceae Code! figs. 48, 49, 50;
map 8 ; 449
subsp. praecox . .figs. 46, 47; map 8; 447
primulina Bak 508
pulchra ined 454
pumila (Ait.) Kunth 376,442,509
pumila sensu Bak 440
rhodesiana Rendle 498
rigidifolia E. A. Bruce fig. 71; pi. 22; 387,487
ritualis Codd fig. 55; map 9; 458
rivularis Berger 387,440
rooperi (Moore) Lem. figs. 3, 72, 73, 74;
pi 23; map 13; 371, 384, 488
rufa Bak. figs. 38, 39,40; pi. 2a; map 6; 434
sarmentosa (Andr.) Kunth pi. 11; map 8; 379,
384, 453
schlechteri Bak 433
schlechteri Schinz 413
sparsa N. E. Br 405,424
splendida E. A. Bruce fig. 45; pi. 10;
map 7; 387, 442
stricta Codd fig. 56; pi. 13; map 10; 461
tabularis Marl map 8 ; 438
thodei Bak ..fig. 63; pi. 46; map 11; 472
triangularis Kunth 384,435,468
subsp. obtusiloba (Berger) Codd fig. 62;
pi 16; map 1 1 ;
subsp. triangularis ..fig. 61; pi. 15;
map 1 1 ; 469
tuckii Bak 386,440
typhoides Codd fig. 18; pi. 3; map 2; 407
tysonii Bak 371,504
subsp. lebomboensis Codd . . .fig. 83;
map 14;
subsp. tysonii figs. 3, 81, 82; pi. 24;
map 14;
umbrina Codd fig. 17; map 2;
uvaria (L.) Hook. figs. 4, 5, 6, 7, 75, 76;
pi. 24, 13; 371, 384, 492
var. maxima (Bak.) Berger 498
var. nobilis (Bak.) Berger 444
var. praecox (Bak.) Berger 444
var. serotina (Bak.) Berger 393
woodii Wats 424
wyliei N. E. Br 424
Kunth, K. S 384
472
506
506
406
Page
Lamarck, J. B 376
Leichtlin, Max 386,437,449
Link, H. P 380, 384
Linnaeus, C 369, 376
Masson, Francis 376,453
Notosceptrum Benth 369, 386, 394, 403
abides (Bol.) Benth 509
brachystachyum A. Zahlbr 410
natalense Bak 407
Pallas, Peter Simon 449
Pre-Linnaean Epithets 373
Redoute, P 370, 380
Rooper, Capt. Edward 384, 491
Roth, A. G 380
Rudolphoroemeria Steud. ex Hochst . . . 394
Salisbury, R. A 386
Schultes, J 384
Thode, Justus 474
Thunberg, C. P 376, 384
Triclissa Salisb 369, 386, 394
uvaria (L.) Salisb 493
Tritoma Ker Gawl 369, 380, 394
burchellii Herb, ex Lindl 493
cau/escens (Bak.) Carr 455
glauca Hort 386, 509
grandis Hort 386, 509
macowanii (Bak.) Carr 469
media Ker Gawl 380, 453
nobilis Guill 386, 449, 509
pumila (Ait.) Ker Gawl 380
recurva Hort 386, 509
recurvata Hort 386, 509
rooperi Moore 384, 488
saundersii Carr 386, 509
uvaria Ker Gawl 380, 493
Tritomanthe Link 380, 493
burchellii (Herb, ex Lindl.) Steud. . . . 493
media (Donn) Link 453
odor at a Steud 384
uvaria (L.) Link 493
Tritomium Link 384,394
uvaria (L.) Link 493
Veltheimia sensu Willd 369, 379, 394
ab.vssinica Redoute 509
media Donn 453
pumila (Ait.) Willd 379
repens Ker. Gawl 453
speciosa Roth 380
uvaria (L.) Willd 379, 493
Willdenow, K. L 376
515
INDEX - INDEKS
Acacia karroo, 232, 246
Acanthoceras Kuetz., 155
Ageratum houstianum Mill., 175
Ainsworthia roraimensis Batista & Cavacanti,
219, 225
xanthoxylii Batista & Costa, 219, 225
Aletris L., 394
pumila Ait., 376, 509
sarmentosa Andr., 379, 453
uvaria (L.) L., 376, 493
Aleuria rutilans, 198
Allocassine N. Robson, 123
Aloe L., 394
longifolia Lam., 376, 493
uvaria L., 367, 376, 493
Alyssum maritimum Lam., 175
Althaea rosea Caw, 175
Anchusa capensis Thunb., 175
Anderson, J. G. The genus Andropogon in
Southern Africa, 5-30
Anderson, J. G. A new combination in Hyparr-
henia, 130
Anderson, J. G. A new species of Panicum,
341-342
Anderson, J. G. A new variety of Festuca cos-
tata, 341
Andropogon L., 5, 7
abyssinicus R. Br., 5, 18
amplectens Nees, 6, 21, 25
var. natalensis Hack., 26
appendiculatus Nees, 6, 16
forma purpurea Nees, 16
var. genuinus Hack., 16
var. genuinus sensu Dur. & Schinz, 15
var. ischaemum Nees, 16
var. serrulatus Nees, 17
auctus Stapf, 130
brazzae Franch., 15
distachyos L., 5, 29
eucomus Nees, 6, 10
festuciformis Rendle, 5, 14
filifolius (Nees) Steud., 7, 25
gayanus Kunth var. cordofanus Hack., 28
var. squamulatus (Hochst. ex A. Rich.) Stapf,
7, 28
helophilus Schum., 28
huillensis Rendle, 7, 12
var. minor Rendle, 13
hylophilus Schum., 28
ischaemum Thunb., 16
lacunosus G. Anderson, 5, 20
laxatus Stapf, 5, 13
pilosellus Stapf, 19
pilosissimus Hack., 6, 130
platybasis J. G. Anderson, 15
ravus T. G. Anderson, 24
schinzii Hack., 6, 22
schinzii Hack, ex Schinz, 22
Andropogon L. —
schirensis Hochst. ex A. Rich. var. angustifolius
Stapf, 26, 27
schlechteri Hack., 14
squamulatus Hochst. ex A. Rich., 28
ternatus Nees var. africanus Rendle, 1 3
Andropogonastrae, 6
Angianthemum, 256
Annuae, 296
Anthericum, 344
aitonii Bak., 344
carnosum Bak., 345
filiforme Ait., 344
longepedunculatum Steud., 344
longipedicellatim Poelln., 344
ornithogaloides (Kunth) Bak., 344
pauper Poelln., 344
quadrifidum Poelln., 345
Antirrhinum majus L., 175
Aphis craccivora Koch, 173
Aplanodes, 111
doidgeana Marais, 11 1
sisymbrioides (Schltr.) Marais, 112
Aquilegia Hort. sp., 175
Arctotis breviscapa Thunb., 175
Arthrocnemum Moq., 255, 273
affine Moss, 256, 280
africanum Moss ex Fourcade, 256, 277
australasicum (Moq.) Moss, 293
capense Moss, 260, 262, 291
capense x natalense, 269
capense X perenne, 272
capense x pillansii, 271
decumbens Toelken, 262, 293, 306
decumbens x capense, 271
decumbens x natalense, 270
decumbens x perenne, 272
decumbens x pillansii, 272
dunense Moss, 286
fruticosum (L.) Moq., 273
var. radicans (Smith) Moq., 289
fruticosum sensu Harv., 387
heptiflorum Moss, 293, 294
hottentoticum Moss, 281, 284
indicum (Willd.) Moq., 256, 276, 278
littoreum Moss , 287
lylei (Ewart & White) Black, 295
macrostachyum sensu Hiern, 276
mossianum Toelken, 281, 303
mossianum X natalense, 270
mossianum x pillansii, 270
namaquense Moss, 284, 287
natalense (Bunge ex Ung.-Sternb.) Moss, 277
var. affine (Moss) Toelken, 280
var. natalense, 277
natalense var. affine X pillansii, 271
natalense x perenne, 269
516
Arthrocnemum Moq —
natalense X pillansii, 269
natalense x terminate, 269
pachystachyum (Ung.-Sternb.) Chev., 276, 298
perenne {Mill.) Moss ex Fourcade, 288, 295
var. lignosum ( Woods ) Moss, 291
var. perenne, 289
var. radicans (Smith) Moss, 291
perenne x pillansii, 271
pillansii Moss, 262, 283
var. dunense {Moss) Toelken, 286
var. pillansii, 284
pillansii x terminate, 271
terminate Toelken, 262, 282, 304
variiflorum Moss, 289
xerophilum Toelken, 286, 305
Arthrolophis, 6
Ascochyta, 187, 199
pinodes {Berk. & Blox.) Jones, 184, 186, 194,
199
pisi Lib., 184
Ascochytopsis, 210
Ascolobus immersus Persoon, 229, 241
Asparagopsis aethiopica (L.) Kunth, 68
albus (L.) Kunth, 39
consanguinea Kunth. 54
decumbens (Jacq.) Kunth, 80
densiflora Kunth, 65
denudata Kunth, 52
dregei Kunth, 55
floribunda Kunth, 72
juniperina Kunth, 48
krebsiana Kunth, 73
lamarckii Kunth, 48
lancea (Thunb.) Kunth, 68
microrhaphis Kunth, 61
minutiflora Kunth, 63
niveniana sensu Kunth, 55, 85
passerinoides Kunth, 43
retrofracta (L.) Kunth, 43
scandens (Thunb.) Kunth, 58
schlechtendalii Kunth, 48
setacea Kunth, 51
spinescens (Schult.) Kunth, 46
stipulacea (Lam.) Kunth, 45
subquadrangularis Kunth, 72
thunbergii Kunth, 55
triacantha (Willd.) Kunth, 46
zeyheri Kunth, 46
Asparagus acocksii Jessop, 74, 89
aethiopicus L., 67, 71
var. aethiopicus, 34, 68, 70
var. angusticladus Jessop, 69, 91
var. natalensis Bak., 70
var. ternifolius Bak., 69, 70
africanus Lam., 33, 48, 85
var. cone inn us Bak., 60
var. dependens (Thunb.) Bak., 51
var. microphyllus Bak., 58
var. wrightii Bak., 59
albus sensu Thunb., 55
albus L., 55
asiaticus sensu Bak., 48
asiaticus L., 50
asparagoides {L.) Wight, 33, 35, 39,81
aspergillus Jessop, 71, 92
bechuanicus Bak., 48
buchananii Bak., 67
burchellii Bak., 45
Asparagus —
burkei Bak., 48
capensis L., 31, 33, 39, 43, 45
var. capensis, 43
var. litoralis Suess. & J. Karl., 44
capensis sensu Marloth, 55
compactus Salter, 33, 34, 59, 73
concinnus Kies, 60
confertus Krause, 74, 85
conglomeratus Bak., 48
consanguineus (Kunth) Bak., 54
cooper i Bak., 48
crassicladus Jessop, 75, 96
crispus Lam., 33, 80
declinatus sensu Bak., 51, 57
declinatus L., 85
decumbens Jacq., 80
densiflorus {Kunth) Jessop, 39, 65
densus Bak., 45
denudatus {Kunth) Bak., 33, 52
dependens Thunb., 48
dinteri Engl. & Krause, 48
ecklonii Bak., 62
engleri Krause, 77
erectus (L.f.) Thunb., 77
exuvialis Burch., 33, 35, 37, 62
falcatus L. 66, 69
var. falcatus, 70
var. ternifolius {Bak.) Jessop, 70
fasciculatus Thunb., 37, 39, 54
fleckii Schinz, 48
flexuosus Thunb., 80
francisci Krause, 48
glaucus Kies, 35, 47
gonocladus Bak., 66
hereroensis Schinz, 62
intangibilis Dinter, 46
judtii Schinz, 48
juniperoides Engl, 35, 38, 63
klinghardtianus Dinter, 84
krausii Bak., 81
krausianus (Kunth) MacBride, 81
krebsianus {Kunth) Jessop, 39, 73, 85
kuisibensis Dinter, 81
lanceus Thunb., 68
laricinus Burch., 60
lugardii Bak., 48
macowanii Bak., 57
var. macowanii, 57
var. zuluensis {N. E. Br.) Jessop, 58
medeoloides (L.f.) Thunb., 81
meyeri nom. nud., 66
microphyllus (Bak.) Kies, 58
microrhaphis {Kunth) Bak., 61
minutiflorus {Kunth) Bak., 33, 63, 66
mucronatus Jessop, 56, 89
multiflorus Bak., 48
multituberosus Dyer, 81
myriocladus Bak., 65
namaensis Schinz, 52
nelsii Schinz, 33, 76
nelsonii Bak., 43
nivenianus Schult., 85
obermeyerae Jessop, 73, 93, 94
officinalis L., 37, 39
omahekensis Krause, 46
ovatus Salter, 81
oxyacanthus Bak., 63
patens Krause, 48
517
Asparagus —
pearsonii Kies, 52
pectinatus Redoute, 78
pilosus Bak., 48
plumosus Bak., 51
racemosus sensu Solch, 71
racemosus Willd., 68, 72
var. tetragonus (Bresler) Bak., 72
ramosissimus Bak., 79
retrofractus L., 33, 39, 58, 61
rigidus Jessop, 33, 64, 90
rivalis Kies, 48
sarmentosus sensu Bak., 65
sarmentosus L., 66
saundersiae Bak., 66
scandens Thunb., 33, 78
var. deflexus Bak., 79
setaceus ( Kunth ) Jessop, 33, 51, 85
spinescens Schult., 45
spinosissimus O. Kuntze, 46
sprengeri Regel, 65
sprengeri sensu Wright, 69
stellatus Bak., 61
stipulaceus Lam., 33, 39, 44
stipulaceus sensu Bak., 46
striatus (L.f.) Thunb., 35, 76
suaveolens Burch., 45, 48
subulatus Thunb., 78
sylvaticus Burch. Ms., 54
ternifolius sensu Hook, f., 69
tetragonus Bresler, 72
thunbergianus Schult., 39, 55, 85
triacanthus Burm. f., 43
triacanthus Willd., 45
umbellulatus Bresler, 81
undulatus (L.f.) Thunb., 35, 84
virgatus Bak., 53
volubilis (L.f.) Thunb., 81
zuluensis N. E. Br., 58
Asteridiella atra ( Doidge ) Hansf, 221
Asterina bukobensis Hansf., 221, 227
Auxarthron umbrinum ( Boudier ) Orr & Plunkett,
231, 242
Bacillus mangiferae, 251
Badenhorst, H., Melastomaceae, 345
Beltrania indica C. V. Subramanian, 232
multispora H. J. Swart, 232
rhombica O. Penzig, 232, 242
Beta vulgaris L., 175
Blepharis leendertziae Oberm., 3
Boerhavia L., 113
adscendens Willd., 115
bract eata T. Cooke, 117
coccinea Mill, var., 117
var. viscosa (Lag. & Rodr.) R. M. Mosc., 118
cordobensis Kuntze, 114
deserticola Codd, 119
diffusa, L., 115
forma subhirsuta Heim., 117
var. diffusa, 115, 120
var. hirsuta Heim., 117
var. minor (Del.) Cuf., 121
var. viscosa {Lag. & Rodr.) Heim., 118
diffusa sensu Burtt Davy, 117
erecta L., 115, 116
hereroensis Heim., 118, 120
marlothii Heim., 117
pterocarpa S. Wats., 114, 116
Boerhavia L. —
repens L., 115, 121
var. diffusa (L.) Hook.f., 1 15
var. minor Del., 121
schinzii Heim., 117
forma angustata Heim., 117
forma villosa Heim., 117
viscosa Lag. & Rodr., 118
Bothriochloa, 6
Botrytis sp., 237
Brachycarpaea, 112
capensis Fourc., 112
flava (L.f.) Druce, 112
juncea (Berg.) Marais, 112
laxa (Thunb.) Sond., 112
var. stricta Sond., 112
lini folia Eckl. & Zeyh., 112
polygaloides Eckl. & Zeyh., 112
various DC., 112
var. flava (L.f.) DC., 112
var. purpurascens DC., 112
Brasiliomyces, 204
entadae Marasas & Rabie, 203, 212, 213
malvastri, 204
Bruce, E. A., 387, 452
Bulbine asphodeloides (L.) Roem. & Schult., 342
coetzeei Oberm., 343
inflata Oberm., 342, 349
namaensis Scinz, 343
Bulbinella, 345
burkei Benth., 345
carnosa (Bak.) Bak., 345
filiformis (Ait.) Kunth, 344
? ornithogaloides Kunth, 344
Callistephus chinensis Nees, 175
Callitris arborea Schrad. ex Drege, 1 25
ecklonii Schrad. ex Pappe, 125
juniperoides sensu Dur. & Schinz, 125
Camellia sinensis, 237
Campy laephora, 155
Canthium suberosum Codd, 345
ventosum (L.) S. Moore, 345
Capsicum annuum, L., 174
frutescens L., 174
spp., 177, 178
Carissa bispinosa (L.) Desf. ex Bren. var. acumi-
nata (L.) Codd, 222
Carpomitra, 358
chytraphora Kuetz., 357
filiformis ( Stthr ) Papenf, 357
longicarpa Simons, 357, 358, 361
Cassine L., 123
capensis L., 123
parvifolia Sond., 123
peragua L., 123
reticulata (Eckl. & Zeyh.) Codd, 123
transvaalensis ( Burtt Davy) Codd, 124
Catalepis Stapf & Stent, 131
gracilis, 134
Catha Forsk. ex Scop., 124
cassinoides ( Robson ) Codd, 124
edulis, 124
Celeceras Kuetz., 155
Centroceras Kuetz., 153, 165
clavulatum, 153, 154
Ceramiaceae, 153
Ceramium Roth, 153, 155, 351
arenarium Simons, 159, 167
518
Ceramium Roth —
atrorubescens Kylin , 161, 167
cancellation Ag., 157
capense Knetz., 154, 157, 163, 164, 351
centroceratiforme Simons, 153, 154, 163, 164, 168
diaphanum ( Light f.) Roth, 161
var. capense Simons, 162, 167
var. pulchellum ( Kuetz .) Simons, 159, 161, 162,
163, 168
var. typicum Feldman-Mazoyer, 162
flexuosum (Kuetz.) Grunow, 157
furcellatum Kuetz., 165
glanduliferum Kylin, 156, 161, 167
gracillimum Griff. & Harv., 161
obsoletum Ag., 164, 165, 168
papenfussianum Simons, 155, 159, 160, 161, 168
planum Kuetz., 155, 157, 167
poeppigiana Grunow, 154
pulchellum sensu Kylin, 159
pulchellum (Kuetz.) Kylin, 162
rubrum (Huds.) Ag., 164, 351
tenerrimum (Mart.) Okam., 157, 167
Ceramothamnion Richards, 155
Cercospora, 199
Cercosporella, 199
Chaetoceras Kuetz., 155
Chaetoceros Ehr., 155
Chaetothyrina amadelpha (Syd.) Syd., 220
musarum ( Speg .) Theiss., 218, 225
Chalara fusidioides Corda, 237
heterospora Sacc., 237
kriegeriana Bres., 237
montellica Sacc., 237
terrestris Ag. & Barna, 237, 243
Cheiranthus linearis Thunb., 99
Chenopodium amaranticolor Coste & Reyn., 175
Chlorophytum, 344
Chrysopogon, 6
Chytraphora filiformis Suhr, 357
Circaea, 310
lutetiana L. subsp. lutetiana,- 310
Cladosporium herbarum, 198
Cleome juncea Berg, 112
laxa Thunb., 112
Clivia sp., 208, 220
Codd, L. E. The Cassine complex, 123-124
Codd, L. E. A new species of Canthium, 345-346
Codd, L. E. New species of Kniphofia, 139-142
Codd, L. E. Notes on Boerhavia in Southern
Africa, 113-121
Codd, L. E. The South African species of Knip-
hofia
Codd, L. E. The status of the genus Lydenburgia,
124
Commicarpus Stand!., I i3
Cooper, Thomas, 386, 415, 445, 502
Coronipus anomalus Spreng., 112
Craspedorhachis, 131
Creothyriella, 210
Crinipes, 136
gynoglossa Goossens, 1 36
Crocoxylon Eckl. & Zeylt., 123
transvaalense (Burtt Davy) Robson, 124
Crotalaria juncea L., 175
Cucumis sativus L., 175
Cupressus juniperoides L., 124
Curvularia lunata ( Wakker ) Boed., 237
trifolii (Kaufm.) Boed. f.sp. gladioli Parmelee &
Luttrell, 236, 243
Cylindrosporium, 199
Cymbopogon, 6
Dahlia pinnata Cav., 175
Danthonia DC., 134, 136
Datura, 173
ferox, L., 174, 176
inoxia, 177
malformation virus, 179
necrosis virus, 178
Quercina virus, 174, 178
rugose leaf curl virus, 177
stramonium L., 174, 175, 177, 178
stramonium x tatula, 173
tatula, 178
virus 1, 2, 3, 3a., 177
Z-virus, 178
Delosperma leendertziae N. E. Br., 3
Dcslandcsia javanica (Zimm.) Batista & Ciferri var.
harana ( Trotter) Batista & Ciferri, 219, 225
De Winter, B. A new species of Eragrostis,
137-139
De Winter, B. Polevansia De Winter, gen. nov.,
130-134
De Winter, B. Styppeiochloa De Winter, gen.
nov., 134-137
Dianthus barbatus L., 175
Dictiderma Bonnemaison, 155
Didymella Sacc., 183
pinodes (Berk. & Blox .) Petrak, 184, 185, 187,
193, 194
Didymellina, 184
pinodes (Berk. & Blox.) v. Hohn., 184
Dilophus suhrii (Kutz.) Papenf, 169
Dimerina mindanaense (P. Henn.) Hansf, 221, 227
Dimerosporium eutricha (Sacc. & Berl.) Theiss.,
222
eutrichum Sacc. & Berk, 222
Diplodia, 223
longipedicellata T. S. & K. Ramakrishnan, 222,
227
Dombeya autumnalis Verdoorn, 143
rotundifolia ( Hochst .) Planch., 143, 144
var. velutina Verdoorn, 144
Dracaena erect a L.f., 77
medeoloides L.f., 81
striata L.f., 77
stricta Schult., 77
undulata L.f., 84
volubilis L.f., 81
Echinoceras Kuetz., 155
Eggplant mosaic virus, 178
Ehretia rigida (Thunb.) Druce, 245, 246
Ehrharta, 207
erecta Lam., 207
Elaeodendron Jacq.f. ex Jacq., 123
reticulatum (Eckl. & Zeyh.) Ettingshausen, 123
Entada spicata (E. Mey.) Druce, 203
Epilobium, 312
adnatum, 309
angustifolium L., 333
atlanticum Litard. & Maire, 318, 322
benguellense Welw. ex Hiern, 331
bi forme Elausskn., 324
bojeri Hausskn., 324
caballeroi, 315
capense Bitch, ex Hochst., 316, 324
capense sensu H. Lev., 319
cordifolium A. Rich., 328
519
Epilobium—
fissipetalum Steud., 328
flavescens E. Mey. ex Harv., 324
hirsutum L., 312
var. africanum H. Lev., 313
var. tomentosum (Vent.) Boiss., 313
var. villosissimum Koch, 313
var. villosum Hausskn., 313
jonathum Hausskn., 324
kilimandscharensis H. Lev., 328
kiwuense Loes., 328
Ianceolatum Seb. & Mauri, 317, 318
lanceolatum sensu Lowe, 321
madagascariense H. Lev., 331
madeirense Hauskn., 321
miguelense H. Lev., 321
mirei Quezel, 313
mundtii Hausskn., 331
natalense Hausskn., 331
neriophyllum Hausskn., 331
subsp. benguellense (Welw. ex Hiern) Engl.,
331
var. welwitschii Engl., 331
subsp. cylindrostigma Engl., 331
subsp. ellenbeckii Engl., 328
numidicum (Batt.) Batt., 315
obscurum Schreb., 318, 321
oligantluim Bak., 331
palustre sensu Ridley, 331
parviflorum Schreb., 315, 316
var. moUissimum (Welw.) H. Lev., 315
var. numidicum Batt., 315
perrier i H. Lev., 331
psilotum Maire & G. Samuels, 318, 323
salignum Hausskn., 318, 329, 330, 331
samuelssonii Raven, 322
schimperianum Hochst. ex A. Rich., 328
schinzii H. Lev., 331
steriophyllum Fres., 316, 328
var. kiwuense (Loes.) Brenan, 328
tomentosum Vent., 313
tetragonum L., 318, 321
subsp. adnatum (Griseb.) Maire, 319
subsp. gilloti H. Lev., 322
subsp. lamyi, 309
subsp. tetragonum, 319, 321
subsp. tournefortii ( Michal. ) H. Lev., 320
tetragonum L. var. ampelusii Maire, 319
var. rodriguezii Maire, 319
var. ? tinaitanum Ball, 320
tournefortii Michal., 320
virgatum sensu Batt., 319
Eragrostis, 136
glandulosipedata De Winter, 137
moggii De Winter, 137
var. contracta De Winter, 138
var. moggii, 138
lehmanniana, 139
Erica, 126
alticola Guth. & Bol, 127, 128, 129
dominans Killick, 126, 129
glaphyra Killick, 127, 128, 129
Eriopodium kraussii Hochst. ex Krauss, 10
Erysiphe malachrae Seaver, 204
Ethel Mary Doidge (1887-1965), by M. D. Gunn,
251-253
Eucalyptus, 203
camaldulensis Dehn., 205
saligna, 237, 238
Euclea crispa ( Thunb .) Guerke var. crispa, 218
Eucomis, 345
bifolia Jacq., 345
nana ( Burm.f. ) L'Herit., 345, 350
Eugenia natalitia Solid., 221
Eusalicornia Gr. & Godr., 296
Festuca costata Nees, 341
var. longiligulata ./. G. Anderson, 341, 347
Fritillaria nana Burm.f., 345
Gaura lindheimeri Engelm. & Gray, 175
Gelidium cartilagineum ( L .) Gaill., 352
Gilletiella, 207
chusqueae, 207
Gomphrena globosa L., 174
Gongroceras Kuetz., 155
Gunn, M. D. Ethel Mary Doidge (1887-1965),
251-253
Gymnandropogon, 6
Gymnanthemum, 256
Gymnosorus nigrescens (Sond.) J. Ag., 171
variegatus, 170
Halleria lucida L., 206
Halocnemum australasicum Moq., 294
Hartogia L.f, 123
Hecatris Salisb., 39
asparagoides (L.) Salisb., 39, 81
Helichrysum trilineatum var. tomentosum, 126
Heliophila, 97
alpina Marais, 100
amplexicaulis L.f., 104
arenaria Sond., 97
var. acocksii Marais, 97
var. agtertuinensis (Schulz) Marais, 98
var. arenaria, 97
var. glabrescens (Schulz) Marais, 97
capensis (L.f.) C. A. Smith, 98
cedarbergensis Marais, 101
christieana Schulz, 99
cornigera Fourcade, 99
cuneata Marais, 101
dent if era Sond., 99
diffusa (Thunb.) DC., 98
var. diffusa, 98
var. Dacca (Sond.) Marais, 98
divaricata Banks, 99
dregeana Sond., 102
elata Sond., 98
var. elata, 98
var. pillansii Marais, 98
elongata Adamson, 99
var. filifolia (Thunb.) DC., 99
esterhuyseniae Marais, 102, 105
eximia Marais, 103
falcata Eckl. & Zeyh., 99
filicaulis Marais, 103
flacca Sond., 98
ffava L.f., 1 12
integri folia L., 101
katbergensis Marais, 104
lactea Schltr., 106
lanceolata Adamson, 100
latisiliqua E. Mey. ex Sond., 98
var. latisiliqua, 98
var. macrostylis (E. Mey. ex Sond.) Marais, 98
lightfootii Phillips, 97
I inear i folia Burch, ex DC., 99
var. dolichocarpa Schulz, 99
var. hirsuta Burch, ex DC., 99
var. pilosiuscula Sond., 99
520
Heliophila—
linearis ( Thunb .) DC ., 99, 101
var. linearifolia (Burch, ex DC.) Marais, 99
var. linearis, 99
var. reticulata (Eckl. & Zeyh.) Marais, 99
macowaniana Schltr., 105
macrostylis E. Mey. ex Send., 98
maximilianii Schltr., 97
var. agtertuinensis Schultz, 98
meyeri Sond., 99
var. meyeri, 99
var. minor Marais, 99
obibensis Marais, 104
pectinata Burch, ex DC., 104
pell aria DC., 98
promontorii Marais, 105
pusilla L.f, 99
var. lanceolata (Adamson) Marais, 100
var. macrosperma Marais, 100
var. pusilla, 99
var. setacea (Schltr.) Marais, 100
reticulata Eckl. & Zeyh., 99
rigidiuscula Sond., 104
rimicola Marais, 105
sabulosa Schltr., 97
var. giabrescens Schultz, 97
schulzii Marais, 106
scoparia Burch., 101
setacea Schltr., 100
sisymbrioides Schltr., 1 12
tenuis N. E. Br., 99
tulbaghensis Schinz, 103
Helminthosporium dorycarpum Mont., 221
pedicellatum Henry, 236, 243
Heteropogon, 6
fHifoUus Nees, 25
Hilliard, O. M. New species of Streptocarpus
(Gesneriaceae), 335-339
Hormoceras Kuetz., 155
pulchellum Kuetz., 161, 162
tenerrimum Mart., 157
Hyalotia lateripes, 235
viridis (Torrend) Cuba, 235, 242
Hyparrhenia, 6
aucta (Stapf) Stent, 6, 130
pilosissima (Hack.) ./. G. Anderson, 130
Hypogynium, 5
festuciformis, 5
schlechteri (Hack.) Pilger, 14
Ilex glabra, 219
mitis (L.) Radi, 218, 219
Indigofera leendertziae N. E. Br., 3
Iris uvaria promont. bonae spei, 374
Jessop, J. P. The genus Asparagus in Southern
Africa, 31-96
Johansonia amadelpha (Syd.) v. Arx, 220, 225
consociata (Syd.) v. Arx, 221
Jussiaea, 310
Killick, D. J. B. Two new Erica species from the
Drakensberg, 126-130
K lesser, Patricia .1. A virus disease of Datura
spp., 173-182
Kniphof, J. H., 367
Kniphofia Moench, 139, 367, 379, 394
acraea Codd, 142, 151, 405
albescens Codd, 417
Kniphofia Moench —
alooides Moench, 369, 379, 493
var. maxima Bak., 498
var. nobilis Bak., 447
var. serotina Bak., 393
bachmannii Bak., 493
basutica ined., 459
baurii Bak., 482
brachystachya (A. Zahlbr.) Codd, 410
breviflora Harv. ex Bak., 142, 413
var. concinna (Bak.) Berger, 413
buchananii Bak., 411
var. flavescens Berger, 41 1, 413
burchellii (Herb, ex Lindl.) Kunth, 394, 493
carinata C. H. Wr., 508
caulescens Bak.. 455, 140
citrina Bak., 373, 497
comosa Hochst., 376
concinna Bak., 413
conrathii Bak., 474
coralligemma E. A. Bruce, 387, 453
crassifolia Bak., 424
decaphlebia Bak., 428
drepanophylla Bak., 481
elegans Codd, 480
ensifolia Bak., 376, 440
subsp. autumnalis Codd, 442
subsp. ensifolia, 440
var. albiflora E. A. Bruce, 440
evansii Bak., 419
flammula Codd , 418
fluviatilis Codd, 476
fibrosa Bak., 422
fibrosa sensu Wood, 413
galpinii Bak., 466
gracilis Harv. ex Bak., 424
hirsuta Codd, 140, 149, 460
ichopensis Bak. ex Schinz, 433
infundibular is Bak., 508
latifolia Codd, 484
laxiflora Kunth, 13, 428
linearifolia Bak., 373, 380, 498
var. kuntzei Berger, 498
var. montana Berger, 498
littoralis Codd , 478
longicollis Bak., 393, 488
longiflora Bak., 387, 498
macowani Bak., 469
modes ta Bak., 403
var. lutescens Berger, 424
var. woodii Berger, 424
modesta sensu Bak., 405, 424
multiflora Wood & Evans, 394, 400
natalensis Bak., 428
var. angustifolia Bak., 428
var. condensata Bak., 428
var. robustior Berger, 428
nelsonii Mast., 437, 466
northiae Bak., 464
obtusiloba Diels ex Berger, 472
occidentalis Berger, 387
odorata Heynh., 493
pallidiflora Bak., 438
parviflora Kunth, 403
var. albiflora Berger, 413
patersoniae Schonl., 496
pauciflora Bak., 438
pedicel lal a Bak., 438
521
Kniphofia Moench —
porphyrantha Bak., 386, 474
praecox Bak., 370, 379, 386, 444
subsp. bruceae Codd , 449
subsp. praecox, 447
primulina Bak., 508
pulchra ined., 454
pumila {Ait.) Kunth, 376, 442, 509
putnila sensu Bak., 440
rltodesiana Rendle, 498
rigidifolia E. A. Bruce, 387, 487
ritualis Codd, 139, 148, 458
rivularis Berger, 387, 440
rooperi (Moore) Lem., 371, 384, 488
rufa Bak., 434
sarmentosa ( Andr .) Kunth, 379, 384, 453
sarmentosa sensu Phillips, 139
schlechteri Bak., 433
schlechteri Schinz, 413
sparsa N. E. Br., 405, 424
splendida E. A. Bruce, 387, 442
stricta Codd, 461, 141
tabularis Marl., 438
thodei Bak., 472
triangularis Kunth, 384, 435, 468
subsp. obtusiloba (Berger) Codd, 472
subsp. triangularis, 469
tuckii Bak., 386, 440
typhoides Codd, 141, 407
tysonii Bak., 140, 371, 504
subsp. lebomboensis Codd, 506
subsp. tysonii, 506
umbrina Codd, 141, 150, 406
uvaria (L.) Hook., 371, 384, 492
var. maxima (Bak.) Berger, 498
var. nobilis (Bak.) Berger, 444
var. praecox (Bak.) Berger, 444
var. serotina (Bak.) Berger, 393
woodii Wats, 424
wyliei N. E. Br., 424
Laurida Eckl. & Zeyh., 123
reticulata Eckl. & Zeyh., 123
Leichtlin, Max, 386, 437, 449
Leistner, O. A. The typification of Podocarpus
elongatus, 142-143
Lepidium, 107
africanum ( Burnt. f '.) DC., 107
var. typicum Thell., 107
var. capense (Thunb.) Thell., 107
var. serration Thell., 107
forma glabratum Thell., 107
var. burchellii Thell., 107
basuticum Marais, 108
capense Thunb., 107
decumbens Desv., 107
desertorum Eckl. & Zeyh., 108
divaricatum Ait., 107
subsp. divaricatum Thell., 107
subsp. ecklonii (Schrad.) Thell., 108
subsp. eu-divaricatum Thell., 107
var. dissection Thell., 109
subsp. linoides (Thunb.) Thell., 107
var. subdentatum (DC.) Thell., 108
var. typicum Thell., 107
subsp. trifurcum (Sond.) Marais, 108
divaricatum sensu DC., 107
ecklonii Schrad., 108
hir tell ion Sond., 108
Lepidium —
pinnatum Eckl. & Zeyh., 108
suluense Marais, 109
sylvaticum Eckl. & Zeyh., 108
transvaalense Marais, 109
Leptopeltella, 205
Lotononis pottiae Burtt Davy, 3
Ludwigia, 310
palustris (L.) Eli, 311
perennis L., 311
Lunaria diffusa Thunb., 98
pinnata Thunb., 98
Lupinus albus L. var. Sweet, 175
luteus L. var. Sweet, 175
Lycopersicon esculentum Mill., 177
var. Little Marvel, 174
Lydenburgia Robson, 124
cassinoides Robson, 124
Mangifera indica, 220
Marais, W. Notes on South African Cruciferae,
97-112
Marasas, W. F. O. New species of Ascomycetes
and a new genus of Sphaeropsidaceae from
Transvaal, 203-215
Marasas, W. F. O. and Rabie, C. J. New and
interesting records of South African fungi,
Part IV, 217-227
Marsh, J. A. Notes on Widdringtonia, 124-126
Marssonia, 199
Masson, Francis, 376, 453
Medeola angustifolia Mill., 81
asparagoides L., 39, 81
Medicago sativa L., 175
Melanospora Corda, 238
episphaeria Phillips & Plowright, 238, 243
Meliola carissae Doidge, 222
doidgeae Syd., 251
Memecylon australe Gilg & Schltr., 345
grandiflorum R. & A. Fernandes, 345
Microcall is amadelpha Syd., 220
consociata, 221
Microchloa R. Br., 131
Microcladia, 154, 351
capensis (Kuetz.) Papenf., 164, 351
coulteri Harv., 352
gelidii Simons, 352, 355
glandulosa (Soland. ex Turn.) Grev., 154, 351
Microcyclus halleriae Marasas & Rabie, 206, 212,
215
Microlaena, 207
Micromycothelia sensu Muller & v. Arx, 205
Micropeltella maitlandii Hansf., 217
Micropeltidium obscurum Batista & Becerra, 217,
224
Monocymbium, 6
Musa, 219
Mycohypallage congesta (Berk.; Br.) Sutton, 222
Mycosphaerella Joh., 183, 195, 207
arachidicola, 195
berkeleyii, 195
bolleana, 196
brassicicola, 196
personata, 196
pinodes (Berk. & Blox.) Stone, 183, 185, 187,
191, 193, 195
tassiana, 195, 198
tulipiferae, 195
typhae, 196
522
Myrothecium verrucaria (Alb. & Schwein.) Dit mat-
ex Fries, 233, 242, 243
Myrsiphyllum angustifolium (Miller) Willd., 81
asparagoides (L.) Willd., 39, 81
erectum (L.f.) Kunth, 77
falciforme Kunth, 81
gramineum Kunth, 81
krausianum Kunth, 81
striatum (L.f.) Kunth, 77
undulatum (L.f.) Kunth, 84
Mystroxylon Eckl. & Zeyh., 123
Myzus persicae, 173
Neurospora crassa, 198
Nicotiana clevelandii A. Gray, 174, 176, 178, 182
glutinosa L., 173, 176, 178'
langsdorffii Weinm., 173, 176, 182
rustica L., 174, 176
sanderae Sander, 174, 176
sylvestris Speg. & Comes., 174, 176
tabacum, L., 173
var. Delcrest, 174
var. Hicks, 174
var. White Burley, 174, 177
Notosceptrum Benth, 369, 386, 394, 403
abides (Bol.) Benth., 509
brachystachyum A. Zahlbr., 410
natalense Bak., 141, 407
Obermeyer, A. A. Identity of Bulbinella carnosa,
345
Obermeyer, A. A. The identity of Fritillaria nana,
345
Obermeyer, A. A. Mrs. Reino Pott-Leendertz
(1869-1965) and the Transvaal Museum Her-
barium, 1-3
Obermeyer, A. A. A new combination in Orni-
thogalum, 344
Obermeyer, A. A. A new combination in Trachy-
andra, 344
Obermeyer, A. A. Two new species of Bulbine,
342-344
Onagraceae, 307
Oncostroma Batista & Marasas, 209
toddaliae Batista & Marasas, 209, 213
Ornithogalum? bidbinelloides Bak., 344
ebidbe Schltr., 345
gracilentum Bak., 345
ornithogaloides (Kunth) Oberm., 344
paludosum Bak., 345
zeyheri Bak., 344
Panicum dewinteri J. G. Anderson, 341, 348
Parapeltella maitlandii ( Hansf. ) Batista, 217, 224
Paullinia, 218
Pelt aria capensis L.f., 98
Pepper veinbanding virus, 178
Petriella, 207
Petunia hybrida Hort. ex Vilm., Ill
var. Rosy Morn, 174
Phaeothyriolum, 205
Phaseolus, 177
lathyroides L., 175
Phalangium filiforme (Ait.) Poir., 344
Phellostroma, 210
Phleospora, 199
Phlox drummondii Hook., 175
Phoebe, 221
tonduzii, 221
Phoma, 187, 199
Phyllachora ehrhartae Marasas, 207, 212, 215
Phyllactinia acaciae Sydow, 246, 248, 249
corylea, 198
Phyllosticta, 187, 199
Physalis peruviana L., 174, 177
Piggotia, 199
Pisum sativum L., 185, 186
var. Greenfeast, 175
vulgaris var. Canadian Wonder, 175
Pleurage erostrata Griff., 231
Pocockiella Papenf, 171
dichotoma Simons, 170, 171
variegata, 170
Podocarpus, 142
elongatus, 142, 143
macrophyllus (Thunb.) Don subsp. maki Sieb.,
142, 143
Podospora anserina (Rabh.) Niessl, 230, 241
Polevansia De Winter, 130
rigida De Winter, 131, 132, 133
Polyanthemum, 263
Poly gala bracteolata, 112
Primula malacoides Franch., 175
Pseudocassine Bredell, 123
transvaalensis (Burtt Davy) Bredell, 124
Pteroceras Kuetz., 155
cancellation (Ag.) Kuetz., 157
flexuosum Kuetz., 157
Pyronema confluens, 198
Ramularia, 199
Raven, Peter H. A revision of the African species
of Epilobium (Onagraceae), 309-333
Rawsonia lucida Harv. & Sond ., 211, 219
Rheinboldiella, 154
Robillarda phragmitis Cunnell, 234
sessilis Saccardo, 234, 242, 243
Rooper, Capt. Edward, 384, 491
Roupala, 221
veraguensis, 221
Rudolphoroemeria Steud. ex Hochst., 394
Ruscus volubilis Thunb., 81
Salacia (?) transvaalensis Burtt Davy, 124
Salicornia L., 255, 296
arabica L. var. paardeneilandica (Ung.-Sternb.)
Chev., 289
australis Sol. ex Benth., 294, 295
blackiana Ulbrich, 295
brachiata Roxb., 298
dolichostachya Moss, 261, 299
europaea L., 296
fruticosa L., 255
var. capensis Ung.-Sternb., 280
var. densi flora Ung.-Sternb., 284
var. paardeneilandica Ung.-Sternb., 289
var. radicans (Smith) Gr. & Godr., 289
fruticosa sensu Harv., 287
herbacea, 255
indica Willd., 276
lignosa Woods, 291
meyerana Moss, 291, 298
pachystachya Black, non Ung.-Sternb., 295
pachystachya sensu Chev., 276, 298
pachystachya Ung.-Sternb., 297
perennis Mill., 289
var. lignosa (Woods) Moss, 291
var. radicans (Smith) Moss, 289
523
Salicornia L. —
perrieri Chev., 297
quinqueflora Ung.-Sternb., 294
radicans Smith, 289
strictissima K. Gram., 261
uniflora Toelken, 297, 299, 307
Salicorniella, 296
Salmonia malachrae, 204
Salvia splendens Ker-Gawl., 175
Scolecopeltidium racemosae Batista & Lima, 218,
224
Scolecotrichum, 199
Septoglocum 199
Septoria, 199
Simons, R. H. The genus Ceramium in South
Africa, 153-168
Simons, R. H. The genus Microcladia (red algae)
in South Africa, 351-355
Simons, R. H. A new species of brown algae from
South Africa, 357-361
Simons, R. H. A new species of the Dictyotales
from South Africa, 169-171
Smilax, 219
Solanum phureja virus, 178
Sonchus oleraceus L., 175
Sordaria anserina (Ces.) Winter, 230
Sphaerella, 184
pinodes, 184
Sphaeria graminis Pers. var. ehrhartae Berk., 207
pinodes Berk. & Blox., 184, 185, 191, 193, 194
Sphaeroderma Fuckel, 238
episphaeria (Phill. & Plowr.) Sacc., 238
Sporochnus, 358
Sporormia minima Auersw., 230, 241
Stapelia leendertziae, 3
Stigmina, 199
Streptocarpus cooksonii B. L. Burtt, 335
denticulata Turrill, 335
trabeculatus Hilliard, 335, 337, 339
Styppeiochloa De Winter, 1 34
gynoglossa ( Goossens ) De Winter, 134, 135, 136
Syzygium cordatum Hochst., 220
gerrardii ( Harv .) Burtt Davy, 218
guineense ( Willd.) DC., 220, 222
Taxus elongata, 142
Tetragonia expansa Thunb., 174
Tetrarhena, 207
Thesium pottiae N. E. Br., 3
Thlaspeocarpa, 110
capensis C. A. Sm., 110
namaquensis Marais, 110
Thlaspi africanus Burm.f., 107
Thode, Justus, 474
Thuia cupressoides L., 124
Thyriopsis, 205
halepensis (Cooke) Theiss. & Syd., 205, 210
sphaerospora Marasas, 203, 205, 214
Tobacco etch virus group, 178
Toddalia asiatica (L.) Lam., 210
Tolken, H. R. The species of Arthrocnemum and
Salicornia (Chenopodiaceae) in Southern Africa,
255-307
Trachyandra, 344
burkei ( Benth .) Oberm., 345
filiformis (Ait.) Oberm., 344
longepedunculata (Steud. ex Roem. & Schult.),
Kunth, 344
Trianthemum, 263
Trichoceras Kuetz., 155
Trichometasphaeria turcica, 197, 198
Trie hope It is Speg., 208
Trichopeltula, 208
carissae Doidge, 209
hedycaryae Theiss., 208
kentaniensis Doidge, 209
Trichopeltum africanum Batista & Marasas, 208,
212, 213
carissae (Doidge) Marasas, 209
hawaiiensis Bat. & Costa, 209
kentaniensis (Doidge) Marasas, 209
Trichothyrium, 208
Triclissa Salisb., 369, 386, 394
uvaria (L.) Salisb., 493
Trifolium subterraneum L., 175
Tripterospora erostrata (Griff.) Cain, 231, 241
longicaudata Cain, 230, 241
Tritirachium album, 229
roseum Beyma, 229, 241
Tritoma Ker-Gawl., 369, 380, 394
burchellii Herb, ex LindL, 493
caulescens (Bak.) Carr., 455
glauca Hort., 386, 509
grandis Hort., 386, 509
macowanii (Bak.) Carr., 469
media (Donn) Ker-Gawl., 380, 453
nobilis Guill., 386, 449, 509
pumila (Ait.) Ker-Gawl., 380
recurva Hort., 386, 509
recurvata Hort., 386, 509
rooperi Moore, 384, 488
saundersii Carr., 386, 509
uvaria Ker-Gawl., 380, 493
Tritomanthe Link, 380, 493
burchellii (Herb, ex Lindl.) Steud., 493
media (Donn) Link, 453
odorata Steud., 384
uvaria (L.) Link, 493
Tritomium Link, 384, 394
uvaria (L.) Link, 493
Uncinula, 245
combreticola Doidge, 246
incrassata Salmon, 246
ehretiae Keissler, 246
praeterita Marasas & Schumann, 245, 248, 249
pterocarpi Doidge, 246
Van Warmelo, K. T. The cytology of Mycos-
phaerella pinodes, 195-202
Van Warmelo, K. T. The identity of a pea blight
fungus in South Africa, 183-194
Veltheimia sensu Willd., 369, 379, 394
abyssinica Redoute, 509
media Donn, 453
pumila (Ait.) Willd., 379
repens Ker-Gawl., 453
speciosa Roth, 380
uvaria (L.) Willd., 379, 493
Verdoom, I. C. A new species of Dombeya and
a new variety of Dombeya rotundifolia, 143-145
Vicia faba L., 175
Vigna, 177
unguiculata (L.) Walp., 175
Viola Hort. sp., 175
Vitalia ekmanii (Petr. & Cif.) Bat. & Cif, 208,
220, 227
524
Vulgata, 296
Whiteheadia, 345
bi folia (Jacq.) Bak., 345
nana, 345
Widdringtonia End!., 124, 142
cedarbergensis Marsh, 125
cupressoides (Z,.) Emil., 124
dracomontana Stapf, 126
juniperoides sensu Endl., 125
schwarzii, 126
Widdringtonia Endl.
stipitata Stapf, 126
wallichiana Gordon, 125
wallichii Endl., 125
whytei Rendle, 126
Wilkommia Hack., 131
Zea mays L., 175
Zinnia elegans Jacq., 175
Zonaria, 169
tournefortii, 170
‘OT- 0f M"-
CUl,*iX>N aiTK***,
rjw &»»»©»»•