Procter collected his (sterile) specimens at two altitudes. At 2070 m., normal B. floribunda (Nos. 1773 & 1774, with non-flaking bark and 3–4 pairs of leaflets) was common on top of the scarp but one aberrant tree (No. 1775) had flaking bark and 4–5 pairs of narrow, very acute leaflets of a shape and venation unprecedented in this species. At 1935 m., a group of trees included normal B. floribunda (No. 1777 with 3 pairs of leaflets) and three apparent hybrids; No. 1778 with 6–8 pairs of leaflets narrower and smaller than those of B. floribunda; No. 1776 with 6–8 pairs still smaller but relatively broader; No. 1779 with 15–18 pairs of small oblong leaflets much broader than in normal B. microphylla, of which several small populations were d on the slope of the scarp. Michelmore 1073 (Ufipa), when collected at 1980 m., was thought to be a hybrid or form of B. microphylla [which occurs on the plateau], “one tree only, in company with Muputu [ B. spiciformis ] and Musompa [ B. floribunda ]”; glabrous except for vegetative buds and the apparent remnants of panicles, it has leaves with 9–11 pairs of very narrow, acute leaflets on a non-channelled rhachis.In addition to the above, Procter 1849 (Chimala scarp at 1525 m., fr. May 1961), with “bark smooth and grey like B. microphylla ”, pubescent vegetative buds like B. spiciformis but otherwise glabrous, and 7 pairs of medium-sized leaflets on a narrowly channelled rhachis with obscure stipellar wings, suggests by its characters a triple hybrid, (sp. 5 × sp. 1) × sp. 6; all three putative parents were present.