508 MR. G. BENTHAM—REVISION OF THE GENUS CASSIA. 
stamens; but in other species, such as Q. moschata and C. leiandra, with the same di- 
verging foliage, the pod retains nearly the ordinary Fistula shape. The Asiatic species 
perhaps the most distinct in foliage, C. javanica, has the smallest difference in the pod, 
consisting merely in the substance surrounding the seeds being dry and pithy instead of 
pulpy; and the most distinct pod among the African species, that of C. abbreviata, has 
nothing peculiar in its foliage. A remarkable thickening of the filament of the lower 
anthers is observable in one or two of the species of each of the four continents without 
any correlative character hitherto observed. | 
Of the subgenus Senna, we have above 160 species, of which about 120 are American 
and 40 in the Old World, besides a few, such as C. occidentalis and C. Tora, which are 
now ubiquitous weeds of tropical regions. They have been variously distributed, by 
De Candolle into four sections, by Vogel into five, of which one is again subdivided into 
four; but some of these are purely artificial, being founded on the fruit alone, and 
widely separating species which have evidently a near connexion. For the present 
purpose it has appeared better to arrange them into sixteen groups, believed to be more 
natural, although not always very well defined, and which, in the enumeration given 
below, 1 have collected into five more marked sections, adopting, with some modifica- 
tions the Chamefistula, Oncolobium, Prososperma, and Chamesenna of Vogel. 
Of these sixteen groups, ten, with about one hundred species, are American ; three, with 
twenty-eight species, belong exclusively to the Old World; one, the Bicapsulares, has 
six American and three African species, but the species of the two countries are sepa- 
rated by characters which may perhaps indicate a separate origin, and should form 
perhaps separate groups; whilst two only, the Floride with eight species and the Picte 
with fourteen, are common to the New and the Old World, without any common cha- 
_ racter separating the American from the Asiatic, African, or Australian species, although 
each species of each country is quite distinct, and further divergence into the nearest 
allied groups is in a different direction in the two hemispheres. 
There is but little of general interest to point out in the distribution of the American 
groups. Most of them and many of the species extend over a great part of tropical 
America, from Brazil to Mexico; many other species and some groups are more local: 
the Platycarpe are chiefly, but not exclusively, Chilian or Bolivian; the Molles mostly 
sub-Andine, from Mexico to Ecuador, or West-Indian; the two aphyllous species belong 
to extratropical South America, where there is a tendeney to the abortion of foliaceous 
expansions in a great variety of genera, analogous to that observable in the Australian 
scrub and in the Egypto-Arabian-deserts, the shrubs becoming more generally phyllo- 
dineous in Australia, spinescent in Africa, junceous in South America. In other respects 
these two aphyllous species appear to be an offset of the widely spread American group 
of Biflore. The group of Chamefistule Brachycarpe, a very natural one, characterized by 
the calyx and other points besides the pod, is Mexican and Texan, and therefore chiefly 
extratropical, with the addition of one species from extratropical South Brazil, without 
any intermediate Andine continuity—a circumstance frequently observed in other genera : a - 
and orders, where closely allied or even identical species are widely severed between 
California and the Argentine States, or still more so between the more temperate or 
