This genus, a member of the tribe Rondeletieae, Avas established by Ducke in 1935 based on D. spigelioides Ducke collected from the state of Amazonas, Brazil (ad ripas inundatas fluminis Curicuriary affluentis Rio Negro superioris). When he described this genus, Ducke believed that its closest allies were Ucriana and Sipanea, both of which genera have loculicidally dehiscent capsules. In his original description of the genus Ducke (loc. cit. pp. 69-70) states that the capsule "ab apice loculicide dehiscens," but an examination of mature fruiting capsules of an isotype specimen at NY (Ducke 24426) shows indubitably that the dehiscence is septicidal.
The septicidal type of dehiscence is further demonstrated by the collection of Maguire, Wurdack, d' Bunting 37592, the holotype of one of the new species of this genus described below. The small, angled, many-sided scrobiculate seeds of the Maguire, Wurdack, & Bunti)ig collection are obviously of the same type as and congeneric with those of D. spigelioides.
The relationship of Dendrosipanea, because of its septicidal type of dehiscence, is much closer to the genus Ghedepophyllum, discussed in another section of this paper, than with any other genus. The main differences between the two genera are to be found in the type of inflorescence, character of seeds, ornamentation of the lower surface of the leafblades, stipules, and amount of viscous exudation of the upper nodes of the branches.
The new collections of Dendrosipanea by Maguire and Wurdack and their associates, especially those assigned to the new species, D. revoluta, demonstrate the close affinities between Denelrosipanea and Chalepophyllum, but the two genera m a y be distinguished on the basis of the differences outlined above.
In view of the erroneous original description of the mode of dehiscence of the fruit by Ducke, it is necessary to emend the description