506 MR. G. BENTHAM—REVISION OF THE GENUS CASSIA. 
These three great subgenera thus modified will, I believe, be found to be distinguished 
by positive characters in the fruit or in the flower, or in both, and so combined with 
habit and inflorescence that there can be no hesitation as to the subgenus into which 
any specimen, whether in flower only or in fruit only, should be placed ; nor have I met 
with a single species really intermediate between either, although in each subgenus there 
may be a subordinate group or species approaching to, without overstepping, the general 
boundary. | 
For the subdivision of these subgenera 1 have not in Fistula succeeded in tracing any 
natural groups intermediate between the subgenus and the species; in Senna 1 have 
adopted several sections and minor groups, either founded chiefly on the pod but not 
always recognizable in the flower, or characterized by the anthers without any distinctive 
pod, or by the seeds or glands, but none of them so clear and well defined as the sub- 
genera. Lasiorhegma, however, has three very good natural and well-marked sections, | 
Apoucouita, Absus, and Chamecrista, characterized chiefly by inflorescence. Such 
details on these and other groups as it may be worth entering into, will find their place 
in the following observations on their geographical distribution. 
On this subjeet of geographical distribution I am induced to enter into some detail, as 
Cassia seems to present, in this respect, a good example of analogous phenomena in the 
majority of large natural tropical genera common to the New and the Old World, such 
as Hibiscus, Crotalaria, Bauhinia, Mimosa (in the Linnean 'sense), Æugenioid Myriee, 
Hedyotis (in its broad sense), Psychotria, Solanum, &c.—phenomena which may be 
shortly summed up as a general diffusion of uniform primary types, with more or less of 
divergence into secondary types in different directions in different countries, the smaller 
groups becoming, generally speaking, more and more local, diverging in different direc- 
tions and with different combinations of characters in different countries, and the number 
of identical species over the whole range very few, beyond those which, as weeds of 
cultivation, are liable to frequent interchange by existing means of transport. 
I would premise, also, that in all considerations of geographical distribution of species 
or other groups, it seems necessary to consider, as an established fact, the great principle 
that natural affinity results from community of descent. The admission of a separate 
arbitrary creation of species, unmeaningly modified from an imaginary type, not only 
reduces the science to a dry statistical enumeration, but entails the preliminary exact 
determination of what is a species, which no one has as yet succeeded in giving ; whereas, 
in adopting the doctrine of descent, natural genera, sections, species, and varieties 
become superior and subordinate groups, similar in nature and differing only in degree. 
Under this aspect, when treating of the area occupied by a group, it matters little 
whether you call it a section, a species, or a variety, provided only you make yourself 
clearly intelligible as to what is the grade of subordination you have in view, and, follow- 
ing the terminology which you believe to be the most generally received, you keep to the 
same term for what you consider the same grade throughout the genus or order you 
are considering. Thus in speaking, for instance, of the Chilian Cassia Candolleana, Vog., 
_ Whether we follow Closs in breaking it up into three species, or - with Vogel regard it as 
a om E UES species, or, as suggested by some others, reduce it to a variety 
