Compilation
Embelia nyassana

This very widespread tropical African species is variable in leaf characters, and the many synonyms were based on particular leaf–forms, these often being paralleled in distant regions. The type specimen of E. schimperi (Ethiopia) has oblong–elliptic, rather coriaceous leaves rounded–acute at both ends, with prominent reticulate venation and apparently no resin–dots. The type specimens of E. guineensis (Sierra Leone), E. pellucida (Angola) and E. nyassana (Malawi) have obovate leaves of thin chartaceous texture (when dry), rounded at apex, tapering at base, with venation very inconspicuous or impressed and then the lamina puckered along each nerve; there are numerous conspicuous resin–dots. In F.W.T.A., Hepper recognised these two extreme forms and a third, species A, with the leaf–shape and resin–dots of “E. guineense” but the texture and strong venation of “E. schimperi”.Throughout tropical Africa, there is some degree of correlation between leaf–shape, leaf–texture and geography, in that the elliptic, coriaceous, strongly veined leaf is predominant in Ethiopia and countries of the F.T.E.A. area while the obovate, chartaceous, weak–veined, punctate leaf is the usual form in the F.Z. area. In Angola, Zaire and W. tropical Africa, both forms are equally represented. Despite this partial geographical separation, I am not convinced that if the specimens of each extreme form were grouped together they would represent two genuine genetically distinct taxa, and in any case many intermediates would remain. More work needs to be done to discover whether the differences are taxonomically significant or merely environmental. Brenan (in Mem. N.Y. Bot. Gard. 8: 490, 1954) took a similar view, though on a more limited scale, when he proposed that E. schimperi was the correct name for E. ahyssinica, E. kilimandscharica, E. nyassana and (probably) E. mujenja and E. pellucida.A few specimens from the extreme W. of the F.Z. area have rather small leaves, probably reflecting the relatively dry environment in these localities. They include Killick & Leistner 3387 (K) from the Caprivi Strip, which Dyer in Fl. Southern Afr.26: 5 (1963) mentions as possibly belonging to E. ruminata, and Codd 7316 (BM; K) from Zambia (B) which P. Halliday in Kew Bull. 32: 294 (1978) cites under E. xylocarpa. Drummond & Cookson 6587 (K; LISC; PRE; SRGH) from the same area is a similar form. A single gathering from Zimbabwe (E: Inyanga Distr., Stapleford, Chisanza R., 1220 m., fr. 4.iv.1962, Wild 5696 (FHO; PRE)) is very unusual in having dentate leaves. E. ruminata (E. Mey. ex. A. DC.) Mez, from S. Africa, is closely related to E. schimperi. It differs in having generally much smaller leaves (1–7 x 0·7–3 cm.), rather contorted branches with much closer nodes than in E. schimperi, and shorter, fewer–flowered racemes (up to 1 cm. long with up to 10 flowers).