on the Hortus Malabaricus, Part III. 117 
tarum called the plant Crateva Tapia, denoting thereby that he 
meant the American plant, continued to quote the Nürvala as 
synonymous. Besides this error, both authors fell into one 
infinitely worse, by placing the Tapia and Nürvala in the same 
genus with the Covalam already mentioned ; the former belong- 
ing to the order of Capparides, while the latter is one of the 
Aurantie. The circumstance which seems to have misled Lin- 
nsus was, that the Covalam was called Belou by the Brahmans 
of Malabar, while the Nürvala is their Rana Belou. When 
Linnæus published the second edition of the Species Plantarum, 
he separated the plant of Jamaica, calling it Crateva gynandra; 
and Willdenow (Sp. Pl. ii. 853.) separated the Niirvala from the 
Tapia of Brazil, and joined it to the Cratæva religiosa of Forster 
and Vahl, of which a figure is given by M. Lamarck (IUl. Gen. 
t. 395.). Although he is followed in this by M. Poiret (Enc. 
Meth. vii. 582.), yet I suppose that this is an erroneous opinion, 
as the berry of the Crateva religiosa is described and represented 
as globular, and no larger than a small plum, and the leaves as 
still smaller and shorter than those figured by Plukenet. 
I have already observed what an unnatural genus the Crateva, 
as left by Linnæus, must be considered; and Dr. Roxburgh 
wished to abolish it altogether. He therefore not only removed 
the species ( Marmelos) belonging to the order of Aurantie, but 
the remaining plant, that he knew, he considered as a mere 
Capparis, which he called trifoliata (Hort. Beng. 41.) Who- 
ever, in fact, endeavours to point out an essential character 
common to all the species of Capparis, will find it a difficult 
matter to exclude the Crateva. Linnæus seems to have at- 
tempted it, by removing the Crateva to the class Dodecandria 
from the class Polyandria, where he at first placed it close by 
the Capparis; but this is trifling: one undoubted Capparis has 
only nine stamina, and in several Cratævas the stamina are fully 
; as 
