MR. G. BENTHAM ON EUPHOHBIACEJS. 219 



Brasiliensis,' differing much in habit and slightly in character, 

 but evidently a congener. 



6. Ckotoke^:. 



The Crotonese, extending them, as we should propose, to all 

 Euphorbiaceae with uniovulate cells to the ovary, without the 

 special characters assigned to any of the preceding tribes, com- 

 prise two thirds of the genera, though less than one third of the 

 species of the whole order. The genera are, with few small 

 exceptions, tropical ; they are mostly, but not all, shrubby or 

 arborescent ; the leaves are rarely opposite, the flowers in axillary 

 clusters only in a very few genera ; otherwise both foliage and 

 inflorescence are much varied. Digitate primary veins in the 

 leaves are frequent, and in some genera pass into digitately lobed 

 or compound leaves ; and racemose or spicate male infloresences, 

 axillary or terminal, are very prevalent. In the flowers, the calyx 

 is very frequently valvate ; the petals, when present, are usually 

 much more developed than in Phyllantheae ; the stamens are 

 frequently central and indefinite, sometimes exceedingly nume- 

 rous, and, when few and definite, those of the outer or single 

 series are always alternate with the sepals or opposite the petals. 

 The fruit is usually the normal Euphorbiaceous capsule, sepa- 

 rating elastically into 2-valved cocci ; but in a very few genera it 

 opens loculicidally in entire valves ; or the cocci remain closed ; 

 or the whole fruit is apparently indehiscent, forming in two or 

 three genera a several-celled drupe. The indehiscent one-seeded 

 drupe, frequent in Phyllanthese, has not been observed in Cro- 

 tonese ; and there is only one monotypic genus in which the 

 albumen is reduced to a membrane enclosing the large embryo 

 with thick fleshy cotyledons. Notwithstanding, however, these 

 general differences between the two great tropical tribes, there 

 remains only as a positive character the uni- or biovulate ovary - 

 cells, one which we overlook in Stenolobeae and Buxeae ; and it is 

 not impossible that a further acquaintance with the rather 

 numerous imperfectly known genera, and the discovery of new 

 ones, may induce a better distribution of the genera not always 

 governed by this one character. In the meantime a subordinate 

 grouping of the 120 genera of the Crotoneae has become necessary 

 for practical use ; the division proposed by Mueller, relying 

 sometimes absolutely on slight differences in the calyx, on the 

 presence or absence of the rudimentary pistil, &c, has frequently 



