510 MR. G. BENTHAM—-REVISION OF THE GENUS CASSIA. 
gence from the Oncolobium type originating in Australia or Asia*, and C. ligustrina a 
corresponding one of American origin, the two being connected by C. occidentalis, this 
would only be another instance of the diversity of divergence in distant countries, which 
has frequently been brought to my notice in working up the Australian flora. Of this 
I need only notice two examples, Pelargonium and „Nicotiana. Pelargonium is essen- 
tially a South-African genus; but one species (only one according to F. Mueller), P. aus- 
trale, has, probably at some far distant epoch, found its way into Australia, become there 
very abundant, and diverged into a great variety of forms, of which one only, and that 
not a common one, appears to be identical with one of the rarer varieties of the S. African 
P. grossularioides, a species which in S. Africa is also very variable, but there diverges from 
the form eommon to the two countries in a totally different direction, and with different 
combinations of characters from what is observed in Australia. So also in the case of the 
Australian representative of the S. American genus Nicotiana; N. suaveolens, Lehm., one 
of its numerous forms, has remained undistinguishable from the Chilian N. acuminata, 
Grah. (apparently one of the several varieties of JN. angustifolia, Ruiz & Pav.); whilst 
the great mass of them has diverged in one direction in Australia, in another in S. Ame- 
rica, so as to give a generally distinct character to each geographical type. 
There is one group, the Bicapsulares, or Chamefistulas with several pairs of leaflets 
and a cylindrical scarcely dehiscent or indehiscent pod, in which, as above mentioned, I 
have included six American and three African species ; but the group is perhaps, like the 
whole section Chamefistula, too artificial; and the American species may be more nearly 
allied to the strictly American Chamesenne Platycarpe, and the African ones on the 
one hand to some Asiatic or African Fistule, and on the other to the Asiatic Chame- 
senne Floride. 
I have said that two very natural groups, the Floride and the Picte (both Chame- 
senne), are, like the Fistule, really common to the two hemispheres. Like the Fistule 
also, the species are all perfectly distinct, but do not form geographical groups. The 
Floride have four American species, three Asiatic ones (of which one extends to Aus- 
tralia), and one confined to Australia. In America their nearest allies appear to be the 
strictly American Chamefistule Speciose, whilst in the Old World they are more con- 
nected with the African Chamefistule and, to a certain degree, with the Asiatic and 
African Plane. The Picte are very remarkable in their distribution: there are ten 
American, two African, and two Australian species, most of them very limited in their 
areas ; and although not distributable into groups, the two which are perhaps the nearest 
allied to each other are very widely dissevered, the tropical African C. didymobotrya and 
the tropical Australian C. venusta, a disseverance which calls to mind a similar one in 
such marked genera as “Adansonia and Erythrophleum, each with one tropical African 
and one Australian species. The common parent of the Picte must have been a con- 
temporary of the progenitor of the Fistule, although his descendants, more dainty 
* More probably in Australia than in Asia, if we consider the peculiar varieties it has formed in the one whilst 
it remains comparatively constant in the other, and if we also take into account other evidences of a more direct 
early connexion between the vegetation of tropical América and Australia than between that of tropical America 
and Eee cone 1871. 
- 
