Compilation
Brachystegia oliveri

VARIATION. Figures 36 and 37 show part only of the remarkable variation in size, shape and spacing of pairs of leaflets. Glabrous and pubescent forms seem to be mixed in all populations; thus, e.g. B. D. Burtt 5294 & 5296 (Tanganyika, Singida–Kondoa, Rift scarp, fl. Oct. 1935) have hairy and glabrous leaves respectively; Hoyle 1093 & 1092 (Tanganyika, Njombe–Songea, 40–50 km., fr. 24 July 1949) are a similar pair from adjacent trees. Extreme hairiness is, however, usually associated with high altitude and exposure. Racemes are usually hairy but often combined with glabrous leaves. Early, widely scattered gatherings, showing various combinations of characters, were made the basis of separate species now relegated to synonomy, but known combinations of characters are now becoming so numerous that even the definition of subspecies and varieties seems increasingly futile. On the whole, fewer ovate leaflets predominate in W. and SW. Tanganyika while rather more leaflets, with the distal pair tending to be obovate, occur mainly in NE. Tanganyika and Kenya; in the same direction the length of the stamina! tube and of the petiole tend to decrease. There are, however, numerous exceptions and almost identical forms occur hundreds of miles apart. Depauperate, apparently distinct local forms such as “ B. itoliensia ” near Lake Victoria and “ B. oliveri ” in Kenya and Tanga Province are associated with exceptional soil and climate at the limits of range but intergrade with the general population. In many places, some of the variation (especially narrow, pointed leaflets) may well be accounted for by introgressive hybridization, notably with sp. 5, B. microphylla (see 5 × 1 under the latter). Naturally enough, hybrids with B. microphylla are relatively easy to detect because of wide differences in obvious characters like the number, shape and size of leaflets and the nature of the inflorescence and flowers; this may partly account for their relative frequency among available specimens because they puzzle collectors. Rare putative hybrids with spp. 2, B. bussei and 3, B. utilis are cited under these species.Hybrids with species in Group B, although theoretically improbable, are suspected (see under spp. 12, B. × longifolia and 14, B. wangermeeana).