White’s type of this subspecies is ‘not typical’ of the afromontane taxon as it occurs in the upland evergreen forests; in fact it seems to me nearer the type of S. guineense itself. It would be foolish to tamper with the names in common use until the whole complex has been examined throughout Africa by modern methods so I have left matters as they are. Certainly the afromontane populations are different from W African coastal populations. Purseglove 866 (Uganda, Ankole, Ijara at 1650 m) growing in short grassland (derived from forest?) is stated to be only 1.2 m tall.Subsp. afromontanum merges with subsp. guineense but there are specimens from lowland areas which have narrowly acuminate leaves e.g. Vaughan 2734 (Tanzania, Uzaramo District, km 16 Utete road (from Dar es Salaam), 21 Jan. 1939 at under 50 m which has slender petioles 1.7 cm long and slender pseudopedicels). Battiscombe 120 seems indistinguishable from subsp. afromontanum and is stated to be ‘common tree on grassland in Coast District’ but no locality is given. Battiscombe, T.S.K. ed. 1: 20 (1926) states ‘also a tall tree of the coast district’.Subsp. afromontanum also merges or crosses with S. masukuense and in the Mbizi Forest, 2180 m, Ufipa District, Tanzania the leaves seem to be rather more coriaceous and the pseudopedicels scarcely developed (e.g. Whellan 1359); despite the ± narrowly cuneate leaf-bases one suspects some influence from S. masukuense .