126 BENTHAM ON PAPII.IONACE^ AND C^SALPIN] EiE. 



some errors, especially in regard to Cercis, which I consid- 

 ered as papilionaceous, which it certainly is not. I have since 

 had occasion to examine some species of above sixty Csesalpi- 

 nieous genera, more especially with reference to the struc- 

 ture of the flower, and to the diversity of aestivation adverted 

 to by Vogel in the Linncea {v. xi. p. 381), and the conclusions 

 I have been induced to come to are stated in my paper on 

 Schomburgk's LeguminoscB [Vol. ii. p. 69, et seq. of the Jour- 

 nal of Botany), I have there given a primary importance to 

 the aestivation of the corolla, and considered the form of the 

 embryo as a more secondary character ; an opinion which 

 appears once to have been that of Dr Vogel also, but he now 

 thinks that the most absolute value should be given to the 

 character derived from the curved or straight embryo, to be 

 determined in cases of doubt by the curvature or straightness 

 of the ovule (i. e. of the nucleus ;) an opinion to which I con- 

 fess I see less reason to subscribe, the more I examine into it. 



It will, I believe, be generally agreed, that the essential 

 character of the great mass of PapilionacecB, is to have a co- 

 rolla papilionaceous in its aestivation (that is to say, the poste- 

 rior petal overlaps the two lateral ones, and these in their 

 turn overlap the two anterior or carinal petals), combined 

 with a decidedly curved embryo, the radicle being usually 

 conspicuously curved down on the edge of the cotyledons 

 and directed towards the hilum; and that the greater num- 

 ber of CfEsalpinie(s have an apparently straight embryo, with 

 a corolla either carinal [i. e. with the anterior petals outside,) 

 in its aestivation, or more or less irregularly imbricate. The 

 difficulty lies in those cases where these characters do not go 

 together, and especially in those genera, now rather nume- 

 rous, where the papilionaceous aestivation is combined with 

 a straight embryo. 



These genera, such of them at least as have come under 

 my observation, may be considered as forming four different 

 groups, corresponding to four of the recognised tribes ot 

 Papilionace(e as follows : — 



1. Arachis, which I have endeavoured in a paper quoted 



