The Eriosema squarrosum complex ( Papilionoideae , Fabaceae ) in southern Africa

Eriosema squarrosum (Thunb.) Walp. has traditionally been the dumping ground for all densely pubescent Eriosema species in southern Africa. This study clarifies the identity of E. squarrosum; recognizes three new taxa: E. luteopetalum C. H. Stirton, E. rossii C. H. Stirton and E. umtamvunense C. H. Stirton; effects the combina­ tions E. latifolium (Benth. ex Harv.) C. H. Stirton and £. acuminatum (Eckl. & Zeyh.) C. H. Stirton; and re­ instates E. dregei E. Mey. The species E. preptum C. H. Stirton, described earlier, also belongs to this complex. Rhynchosia barbertonensis C. H. Stirton is given as a new name for E. rogersii Schinz.


INTRODUCTION
The Eriosema squarrosum (Thunb.)Walp.com plex remains the only unresolved complex among the Eriosema species of southern Africa.As in the E. cordatum E. Mey.complex (Stirton 1978(Stirton , 1981a) ) it is partly a nomenclatural muddle and partly a taxo nomic problem.Once again hybridization has played a prominent role in the development of the complex (Stirton 1981b).
The complex comprises the majority of the densely pubescent Eriosema species in southern Africa.Most of these plants have in the past been referred to either E. zeyheri E. Mey. or E. squarro sum.
Central to the complex is E. squarrosum.

This species has been and is easily confused with E. acuminatum (below) and E. preptum (no. 8). For differences see under the latter species.
Eriosema squarrosum is the smallest of all the Cape, Transkeian and Natal species.It flowers from September through to March.Eriosema acuminatum occurs mainly in grasslands in the Transkei (Fig. 4).It has not often been col lected and is probably more common within its known overall distribution area than its representa tion in herbaria suggests.According to herbarium labels this species flowers between October and November, but also occasionally in December and January.No ecological data were found on herbar ium labels.This species is endemic to southern Natal (Fig. 6) but may yet be discovered in the Transkei.It is com mon along the coast and seems to grow best on sandy soils in previously burnt grassland.(Fig. 7).The range of this species appears to be extending as a result of roadbuilding activities.Eriosema luteopetalum is a very showy shrub worthy of consideration as a garden plant.It pro duces masses of inflorescences in spring providing a flash of yellow colour, soon to be followed by colourful brown fruits that persist on t"he plant long after the fruits have explosively scattered their seeds.

Eriosema acuminatum (Eckl
The specific epithet luteopetalum, was chosen to draw attention to the massed yellow flowers.It seems remarkable that this distinctive and locally abundant species has, until recently, been so rarely collected.This species appears to have been missed by most of the early collectors.Its nearest allies are E. latifolium (no.7) and E. dregei (no.5) from which it differs in its very much larger flowers and distinctive pubescence.Eriosema umtamvunense is endemic to the rolling grasslands decking the plateaux on either side of the Umtamvuna Gorge (Fig. 8c).So far it has been re corded only above 300 m.It is restricted, with E. latifolium, to Acocks's Pondoland Coastal Plateau Sourveld (Fig. 9).Flowering takes place in Novem ber and December.

E. umtamvunense is a very distinctive, locally abundant
Eriosema, yet like E. latifolium it has been collected rather infrequently.Strey 7242, collected as recently as 1967, is the first record of the species.
It seems to have been missed by all the early collec tors.This is not surprising as it is distributed on top of the escarpment and generally grows in grassland that would have been largely inaccessible to early ex plorers.One wonders what other treasures are still to be discovered in the Umtamvuna Gorge and its escarpment.
This species has been consistently lumped with E. squarrosum (no.1), E. dregei (below), E. latifolium (no.7) and E. luteopetalum (no.3).It differs from all of these species in its red and yellow flowers and golden-haired calyces; occasional yellow morphs can occur.The latter species all have yellow flowers and silver-haired calyces.From E. preptum (no.8) it can be separated by its much larger flowers and fruits and by the wing petals exceeding the keel petals.

FIG. 12. -Isotype of Eriosema rossii (Stirton 1205).
Eriosema rossii is restricted to the higher-lying Ngongoni Veld (Acocks's Veldtype 5) of southern Natal and the eastern Transkei (Fig. 13).The Umkomaas River Valley bisects Acocks's Veldtype 5 and it may be significant that many legumes found in the southern portion of this veld type have not been recorded from the area north of the Umkomaas River.Another example in Eriosema is E. populifo lium Harv.Although this species is fairly common within its distribution range, it was until recently poorly repre sented in herbaria.With its large stipules and small compact inflorescences it is generally easily separ ated from the other species in the E. squarrosum complex.
There is, however, a group of plants which may form part of this species.I have previously anno tated this group in various herbaria as E. superpositum mss.It is allopatric with E. rossii and is separ ated from it by its long-peduncled, few-flowered box-like inflorescences held high above the sparsely pubescent subtending leaves.But since the interven ing area between the two ranges has not been col lected I do not wish to recognize it formally until more is known about the variation present in both groups.
The following brief description may be useful to collectors who might be fortunate to find plants of this unnamed group which grows in small scattered colonies.This taxon has disappeared rapidly from the area between Durban and Key Ridge and is now found only in a few isolated patches of the natural grass land that has not yet been built on or been destroyed through overgrazing.

7.
Eriosema  Although described by Harvey as long ago as 1862 this species has been collected only rarely.It occu pies Acocks's Veld Type 3, his Pondoland Coastal Plateau Sourveld.This veld type has until recently been little explored and I am certain that once a full fenumeration has been made of its constituents it will receive the recognition it deserves as an area of endemism.
Eriosema latifolium is closely allied to E. dregei (no.5) and E. luteopetalum (no.3), two other yel low-flowered suffrutices from the same general re gion.It differs from these two species in its much smaller flowers, the appendages on the standard be ing well-free of the auricles, its shorter stipules and narrowly oblong racemes.
The presence of variably sized pollen grains strongly suggests that this species may be of hybrid origin.It is perhaps significant that this species oc curs as scattered individuals or small colonies.The few fruits that have been found contained shrivelled seeds only.Eriosema preptum is endemic to Natal and ex tends some 100 km inland from the coastal belt (Fig. 17).It occurs in Acocks's Coastal Forest and Thornveld (V T l), Ngongoni Veld (VT5) and his Zululand Thornveld (VT6).It favours sandy, well drained sites along roadsides and ditches but is also com monly found in regularly burned grassland.Flow ering extends from September to February but oc curs mainly in October.

(Lamiaceae).
There is a note written by the late Dr J. Raynal in the Paris Herbarium (27-6-1963) to suggest that 'this is probably an incorrectly labelled plant'.The protologue of this species is too vague and as it could be applied to any of a number of species it cannot be appUed until a specimen so named is found.