220 ME. G. BENTHAM ON EUPHOUBIACE^. 



the inconvenience of widely separating very closely allied genera, 

 or of collecting in one genus most dissimilar species ; and the 

 characters of Baillon's series are not sufficiently explicit. I have 

 endeavoured, however, to conciliate the views of both in pro- 

 posing eight subtribes, some still too artificial, but, as I am led 

 to believe, fairly recognizable. They are based much on inflores- 

 cence, partly on calyx and other floral characters. The first three 

 are usually but not always petaliferous, at least in the males, the 

 remaining five always apetalous. 



Subtribe 1. Jatbophe^;. Flowers usually in a terminal 2-3- 

 chotomous centrifugally cymose panicle, often androgynous 

 with the female flower central and solitary or several in the 

 primary forks, the lateral ones all male, but sometimes unisexual, 

 and occasionally reduced to a terminal cluster, or the female 

 flower very rarely alone and terminal. The floral characters are 

 various, the petals being wanting in several genera ; but the 

 inflorescence appears to be constant and readily observed. All 

 the genera, except two small ones, are in America, and most of 

 them chiefly or entirely from that hemisphere. The first five 

 are always apetalous, the other six usually petaliferous, at least 

 in the males. They are : — 1. Elateriospermum, a single species 

 from the Malayan archipelago, with very imbricate sepals and 

 indefinite stamens, without any rudimentary pistil. The large 

 seeds are quite exceptional in Crotonese, the albumen being 

 reduced to a chartaceous membrane enclosing the embryo, of 

 which the cotyledons are very thick and fleshy, enclosing at the 

 base the small radicle. 2. Cunuria, and 3. Micrandra, two well- 

 defined Brazilian genera of two or three species each, well 

 placed in juxtaposition by Baillon, although Mueller had asso- 

 ciated Micrandra by some mistake with Eucrotonese. 4. AveU 

 lanita, a single Chilian species, only known to me from Philippic 

 description, which would seem to place it in this group. The 

 calyx described is that of Aleurites ; but there are no petals, and 

 the capsule is a normal 3-eoccous one. 5. Hevea, nine tropical 

 American species, with digitately 3-foliolate leaves and a shortly 

 lobed or toothed calyx, but otherwise with very nearly the cha- 

 racters and, as stated by Spruce, the habit of Cunuria. 6. Joan- 

 nesia, the Anda of Jussieu, a single well-known Brazilian species. 

 7. Jatropha, the principal genus of the subtribe, of which the 

 species have been extended by the ' Flora Brasiliensis ' to 68, 



