GENERAL ON A. rubifolius. There remain a large number of forms of which there may be large stable populations in the areas concerned. Glabrous- or almost glabrous-leaved forms have mostly been labelled as var. alnifolius but are very different from the small-leaved coastal form traditionally looked on as typical alnifolius (but not by any means identical with the type), e.g. Newbould 7398 (K 1, Northern Frontier Province, upper Turkwell R., Kaputir, 900 m.) has obovate ± glabrous leaflets, up to 9 × 5.5 cm. These are perhaps glabrous forms of var. dasystachys but also approach A. ferrugineus. A similar form is from Zanzibar, Faulkner 2388 (Mkokotoni, 26 Oct. 1959) with glabrescent leaflets 10 × 5 cm. and very densely flowered inflorescences 1–23 cm. long, and there is actually a name based on this form - A. stachyanthus Gilg forma calvus Radlk. in E.P. IV, 165: 542 (1932); type: Zanzibar, [Mkokotoni] Kokotoni, Stuhlmann 703 & 704 (B†, syn.). Other similar forms but with branched inflorescences have been referred to A. africanus, e.g. Paulo 684 (Masai District, Mosiro [Musilo], 19 Aug. 1960) and Verdcourt 1527, a liane to 7.5 m. (Masai District, 43 km. W. of Magadi, ‘Sweet Waters’, 29 July 1956).A form from K 2, 3 (Baringo) with glabrescent leaflets, coarsely crenate in upper half and cuneate and entire beneath has somewhat the facies of A. tanzaniensis F.G. Davies, e.g. Hemming 361 (Turkana District, 40 km. W. of Kangetet, 7 July 1954, in sandy riverbed); it has been considered a form of var. alnifolius .Around Nairobi (1700–1950 m.) there is a well known form which I have referred to var. alnifolius which has usually been called A. aff. repandus (Baker) Radlk. or A. kilimandscharicus Taub.; it has glabrous (save for domatia) coarsely toothed leaves and somewhat larger flowers. Populations have certainly been constant for over a century as evidenced by the numerous collections, e.g. Bally 10504, Dale 2884, Dummer 2007, Kibue 157, Linton 135, Mearns 179, Napier 294, Polhill 357, Rayner 18, Rogers 469, 547 & 548, and Whyte. There are, however, too many intermediates when a much larger area is considered but local botanists could call it var. alnifolius ‘Nairobi form’.I am not entirely certain of the identity of the fragment of the type of A. kilimandscharicus. It could be an abnormally small leaved form of A. ferrugineus .Something must also be said about forms occurring on Zanzibar. The Mkokotoni form has been mentioned above and var. alnifolius certainly occurs but other material is more difficult. Vaughan 1065 (Mbweni, 9 Jan. 1930) has been referred to A. rubifolius var. rubifolius but is only sparsely hairy and intermediate with var. alnifolius; the inflorescences are sparsely branched which has led to it being named A. africanus. Faulkner 3359 (Chukwani, 14 Mar. 1964) and Vaughan 1421 (Chukwani, 28 July 1930) with simple inflorescences are similar and the same kind of intermediates. A really detailed study of the populations on the island might help one to understand the reproductive biology of the genus. Gereau et al. 4997 (Ufipa District, 8–9 km. from Sumbawanga on Mbala road, Safu, in Brachystegia woodland at 1600 m.) with stiff branched inflorescences from leafless nodes, obovate leaflets 6.5× 4 cm., with rounded apices and yellow-grey indumentum may be better considered a form of A. africanus .