106 
Seleciions. 
[no. 3, NEW SERIES, 
and Fpica)yuriis to consist in the proportion of the cotyledons, I 
fear that they must "be again united : in Epicarpariis the cotyledons 
are often thick, but they are constantly folded and crumpled. Of 
the Tro2)his Americana,^' whicli is the type of the genus Trophis, 
I have not examined the seed, nor does M. Trecul (Ann. Sc. Nat. 
ser. 3. vol. viii. p, 147) describe it; but tVrere are abundant marks 
of distiaciion in the spicate inflorescence and tubular perianth. 
Trophis and Epicarpurus both belong to the MorecB as charac- 
terized by Trecul, the stamens being indexed during aestivation. 
In Trecul's memoir alluded to, an error occurs as to Trophis 
spmcsa, Roxb. This I have said is one of the thorny species of 
Ep}icarpuruSy and almost identical with E. Timorensis of Decaisne, 
as every one must acknowledge who reads attentively Roxburgh's 
description (Fior. Ind. vol. iii. p. 762) ; but Trecul refers it 
(p. 123) to Cadrania Javanensis, a plant belonging tohis Artocarpece, 
having the female flowers in dense capitula, arranged in umbels, 
and with a simple style. In this he has been, perhaps, misled by 
Blume, who, in his ' Bijdragen,' p. 489, appears to have described 
a species of Ciidrania (probably C. olovaia, Tree) under the name 
of Trophis spiaosci, Roxb. Indeed Roxburgh himself may have led 
others astray, the plant to which he gave the manuscript name of 
Trophis spinosa at an early period of his botanical career, and under 
which he deposited a drawing in the E. I. C. Museum (tab. 119), 
and which name was adopted by Willdenow, never having been 
published by him as such : in fact, his manuscript T. spinosa, and 
consequently the T. spinosa of Willd. Sp. PI. vol. iv. p. 735, is the 
Batis spinosa of the * Flora Indica' (vol. iii. p. 762) ; nor is T. acu- 
leata. Roth, Sp. Nov. p. 36S, at all distinct : this, although belong- 
ing to the MorecBy has the habit of Ciidrania, and is the Plecosper- 
mwn spiiiosu7n, Tree. (1. c. p. 124). So far as I can ascertain, the 
Cadrania Javanensis, Tree, was unknown to Roxburgh, although 
t. 
* Tropins Ramon from Mexico (Linnaea, vol. vi. p. 357) is scarcely distinct. 
The specific name, too, is unfortunate, being obviously the same as liamoon, by 
Tvliich jT. A77icricana is known in Jamaica. JRa?non is a Spanish expression for 
small boughs or tvigs, which, when broken ofl, ate suitable as fodder for cattle, 
and it dots not inoicate the species of plant. 
