BENTHAM ON PAPILION ACE.E AND C^SALI'I NI F-.T,. 127 



above, to prove the affinity of to Stylosanthes amonu; Hedy- 

 sarece, an affinity recognised by Bennett,* and by Torrey 

 and Grayf who have further confirmed it by the addition of 

 their new genus Chapmannia, This affinity appears to me 

 to consist not only in the " corolla? structiira Stylosanthis 

 simili," but in the remarkable structure and physiological 

 development of the sterile and fertile flowers in all their 

 parts, and in the pod as well as in habit. Vogel says indeed, 

 "qu£E vero similitudo Hedysareuni characterem non attinet, 

 sed in quavis tribu occurrere potest, ita ut banc causam non 

 agnoscam," but he does not point out any instance, nor has 

 it been my lot to observe a single example of similar flowers 

 in any other tribe of Leguminosa. 



2. Brongniartia (including Peraltea) and Harpohjce, which 

 to my eyes bear a much closer affinity to several Galegece^ 

 than to any genus of CcBsalpinieiP, excepting in the single 

 character of the embryo. 



3. Geoffroya and Andira, Dipteryx and Pterodon, Cyclolo- 

 oium, and perhaps some others among my Dalbergiece, where 

 't appears to me that their nearest allies are to be met with, 



4. A considerable number of genera with stamens free or 

 nearly so, the flowers papilionaceous in aestivation, but some- 

 times rosaceous in expansion, the habit and inflorescence 

 generally that of Dolbergiets, or of some Galegetr, and not 

 unlike that of a few C<ssalpiniea:, which I had collected under 

 the name o( Sophorece, and placed at the end of Papilionacece, 

 3s forming an approach to Casalpiniece. As my greatest 

 doubts have always been in relation to some of these genera, 

 It is to them I have more especially directed my attention on 

 this occasion. 



In order to ascertain what practical advantage may be 

 gained by the examination of the ovulum rather than of the 

 ^'pe embryo, I selected for comparison five species, of which 

 I happened to have abundance of flowers in various stages, 

 and in a good state for dissection, and also ripe seeds, viz., 



* Piantae Javanicae Rariores, p. 132. 

 f Flora of North America, v. i. p. 334. 



