70 SCHOMBURGK'S GUIANA PLANTS, 
more the species are known, the more difficult it becomes to 
draw the precise line of demarcation between the two, most 
the characters, hitherto considered as absolute, having ul 
mately broken down when better examined. All that remains 
to be done, is to combine the characters common to the great 
mass of each Suborder, without giving to any one a value so 
definite as to remove particular species from others with which — 
they are, in other respects, intimately connected. The most | 
important, and that which it now appears may be the most 
safely trusted to, in almost every instance, is the sestivation 
of the corolla, which in the Papilionacee is always regularly 
papilionaceous, the vexillum overlapping the two ale and | 
these in their turn enclosing the carina. In Cesalpiniee, it is 
most commonly the very reverse, being what Vogel* has ap- 
propriately denominated carinal, that is to say, the lower oF - 
carinal petals enclose the alee, and the vexillum is inside of 
all. In some genera, as for instance, Exostyles, the sestiva- 
tion is regularly twisted, each petal overlapping the adjoining : 
one on one side, and sometimes the ale are outside instead 
of the carina, the vexillum remaining inside of all. It is only | 
in the genus Leptolobium that some species, having a truly 
papilionaceous estivation, appear however better ime 
mda Cesalpiniee. ; 
The character next in value, derived from the ibo of the 
mdsyo and direction of the radicle, is the one to which the 
. most importance has been attached, and has induced the abso- 
lute division of Leguminose into Curvembriee and Rectembriee; 
but it is now ascertained that if rigidly followed up, this divi- 
sion would be most unnatural. Taking it however merely 
as a general character, it appears that the embryo is, with 
very few exceptions, curved in Podalyrice, Lotee, Hedysarees — 
Viciee, and JPAaseolee, in all of which the papilionaceous : 
corolla is also the most decided; more or less curved or quite 
—— in the seer oes and Sophoree, which vem an 
been first pointed out € Vogel. | 
xd Dainida XI. p. 381, iere thi station appen to have 
