The generic order is in general that of Bentham’s monumental “Revision of the Suborder Mimoseae” in Trans. Linn. Soc. 30: 336–664 (1875), although not necessarily the true and final one. Dnyansagar (in Journ. Indian Bot. Soc. 34: 362–374 (1955)) has suggested a new classification based on embryological and pollen characters, by which the Mimosoïdeae would be divided primarily by the presence of either simple or compound pollen-grains. He has, however, examined less than half the genera occurring in our area, although others have been accounted for by Erdtman (Pollen Morphology and Plant Taxonomy, Angiosperms: 225–6 (1952)). Any classification on this basis would cut across most of the main tribes recognized by Bentham.Existing classifications make it often difficult to identify the genera of this subfamily without complete material, including flower and fruit. As one or the other is so frequently absent, I have endeavoured to construct two alternative keys, one designed for flowering, the other for fruiting specimens, together with a conspectus of the various types of pod encountered in the subfamily. It must be clearly understood, however, that the keys and conspectus are artificial and apply to East Africa, and must not be relied upon in other areas or for identifying exotics.