118 Dr. Francis HaurrTOox's Commentary 
as numerous as in any Capparis. The appearance of the Cra- 
tæva is however so different from that of the Capparis, that they 
cannot enter one natural genus; and I think the character 
by which they may be best distinguished is, that the Capparis 
has petala sessilia, and the Crateva, petala unguiculata. Linnæus 
could not of course make use of this distinction, because the 
petala of the Marmelos are sessile. The flowers of the Crateva 
also are polygamous, a large proportion being entirely male, with 
only a rudiment of the pistillum. Whether or not they are all 
dicecious I cannot say, but some are certainly so. 
In the course of travelling, I observed that the specimens of 
the Crateva, which I examined, differed considerably from each 
other, so as to lead me to suspect that in India there may be 
several species: but as I never afterwards was long enough 
stationary in one place to observe the same tree in all its stages, 
or to try the effect of different situations on the seed of the same 
plant, I am by no means certain that my suspicions are well 
founded. I shall however mention the circumstances by which 
I was induced to suppose that there are at least four species of 
Cratæva in the Gangetic provinces. "This will at any rate throw 
light on the true generic character. Specimens of the whole 
have been deposited in the library at the India House. 
I shall first describe a species which I met with in Behar, and 
which, I believe, is Dr. Roxburgh's Capparis trifoliata, because 
he does not quote the Nürvala as synonymous, and because he 
thought it his plant when on my return from Ava I showed him 
specimens. If it were not for the long point at the end of the 
leaflets, their breadth would fully equal their length ; including 
these points, the breadth is about half the length. 
1. Crateva odora. : 
Capparis trifoliata. Hort. Beng. 41? 
Varuna 
