410 Dr. Puancuon on Melianthee, 
Until now, while speaking of Melianthus, I have purposely restricted my 
observations to Melianthus major and M. Himalayanus, because they deviate 
in several important points from the structure of the two other Cape species, 
viz. Melianthus minor, L.and Melianthus comosus, Vahl: in these, the constant 
abortion of the fifth anterior (apparently posterior) petal, the total absence of 
any depression in that part of the receptacle which is inclosed by the gland, 
-and of any lobes at either end of the capsule,—all these points concur with 
the striking peculiarities of facies, to point out the plants in question as types 
of an independent genus; the name of which, Diplerisma, will allude to the 
character of their free, lateral, subulate stipules, so very different from the 
wide, double, intrapetiolar stipule of the real Melianthus. 
Having thus brought under a comparative review the four genera which 
constitute the order of Melianthew, we may conclude with some general in- 
ferences on their common relations to other groups; or, rather, we may 
establish upon proofs what has been anticipated above of their being equally 
removed from Rutacee and Zygophyllec, while they are closely allied to Sa- 
pindacee on the one hand and to Geraniacee on the other. 
First, although the sagacious and profound A. Laur. de Jussieu had con- 
nected, under the common name of Rutacew, the genera which Mr. Robert 
Brown distributed afterwards into the independent orders of Zygophyllee 
and Diosmec, it strikes me that those orders belong to natural classes truly 
bordering on each other, but quite distinct. On one side, Diosmec (in- 
cluding Rutew, Diosmee proper, Xanthoxylew and Aurantiacee) form, with 
Simarubec and Meliacec, a vast and indivisible class; on the other side, Zygo- 
phyllecee, Oxalidew, Connaracew, Leguminose and Moringew are connected by 
so many points of structure and habit, that they offer, in my humble opinion, à 
rare example of a well-marked and at the same time complete natural groups . 
where the constant tendency of the folioles of the com pound leaf to periodical 
sleep, or sometimes to sudden motion under an irritating influence, is always 
connected with that important structural fact, the articulation of the foliole 
with the stipes on which it moves. Sapindacew do not seem to me to belong 
" the first class, any more than Geraniacee deserve to be united with Oza- 
— although this last opinion is generally prevalent. In fact, the true 
spirit of improvement in science is not to submit tamely and blindly to received 
