1957] 



THE BOTANY OF THE GUAYANA HIGHLAND — PART II 



337 



The relationship of this species is certainly with P. nitida Miq. , a Guianan 

 species. P. truncata differs by its alternate leaves, by its lack of a waxy bloom 

 on the lower leaflet surfaces, by truncate leaflets (hence the specific epithet), 

 and by the erect inflorescences of heads, most of which are smaller than most of 

 those P . nitida I have seen. 



The fruits of p . truncata are dark-brown-velutinous but the presence or absence 

 of pubescence in this group may not be very significant. In material assigned by 

 Dr. Amshoff to P. nitida, the hermaphrodite flowers at the apex of the heads have 

 glabrous or nearly glabrous pistils but the flowers near the base of the hermaphro- 

 dite part of the head are densely pubescent. In the flowering material of P. 

 truncata the ovaries are uniformly glabrous but the fruiting specimen has densely 

 velutinous pods. There is little doubt that the two collections are conspecific, 

 although it is possible that two subspecific taxa are involved. It appears better 

 to conclude that the presence or absence of pubescence is a character in this 

 group which varies independently and indiscriminately. 



LEGUMINOSAE-CAESALPINIOIDEAE 



Dicymbe Spruce ex Benth. & Hook. Gen. PI. 1(2): 564. 1865. 



In the winter of 1950, Drs. Maguire, Wurdack, and Cowan collected in Vene- 

 zuela a leguminous plant which at the time was of unknown affinity generically 

 and until this year it remained so. This winter, duplicates of the collections of 

 this plant were sent to Mr. N. Y. Sandv/ith, who kindly studied it and suggested 

 that it was to be associated with the genus Dicymbe. He also compared some of 

 the types in this genus; for this assistance I am very grateful. 



With. the recognition of the relationship of that plant, several other collections 

 proved to represent still other new species of the genus. In the light of these 

 additions, it seems wise to present a preliminary treatment of the genus. 



It is evident that in this group we are confronted with another of the reduction 

 series which seem almost characteristic of this part of the Caesalpinioideae; the 

 problem to resolve is the status to be accorded each of the stages of the series. 

 The corolla is more involved in this group than any of the other flower parts. In 

 sect. Dicymbe the corolla is composed of five, equal or subequal, petals; sect. 

 Triplopetala, as the name implies, is characterized by three complete petals and 

 two petalodia 13 ; in sect. Eremopetala four of the five petals are reduced to 

 petalodia but the fifth has a well-developed blade; and the species of sect. 

 Apoxypetah. have all the petals reduced to subulate-lanceolate petalodia. In 

 spite of the obvious differences between the extremes of this reduction series, I 

 feel that the phylogeny of the group is better expressed by treating it as one 

 genus, consisting of two subgenera, each with two sections. 



Of the four sections recognized, three require names and they have been 

 named with Greek compounds signifying the petal condition of each group. 

 Eremopetala refers to the single, "lonely" petal; Apoxypetala refers to the 

 absence of complete petals, that is, the petals are "reduced"; and Triplopetala 

 refers to the three complete petals of the one species included here. This is the 

 species formerly referred to Dicymbopsis by Ducke and ordinarily the generic 

 name would be used as a infrageneric epithet but here it is impossible, according 

 to paragraph four of Article 31 of the Rules; the sectional name must not repeat 

 the generic name with the ending "-opsis." 



13 For the distinction between petals and petalodia, see Mem. N. Y. Bot. Gard. 8:259 

 (1953). 



