128 
CONTRIBUTIONS TO BOTANY, 
which extends nearly to the centre of the cell — a fleshy albumen, 
enclosing in its summit an embryo with a superior terete radicle 
—are all characters surprisingly in accordance with Villaresia, 
the principal point of its disagreement being its inferior fruit. 
With this single exception, Bursinopetalum is quite conformable 
with the Aquifoliacea ; and even this exception, making due 
allowance for its aberrance, in great measure may be reconciled. 
The ovary, with its fleshy summit assuming the appearance of 
an epigynous disk, as in Styracea, is half superior, that is to 
say, its cell rises above the level of the insertion of the stamens ; 
and we And a parallel of this instance in Halesia. I have shown 
how' it happens in that genus*, contrary to what occurs in Styrax 
and its congeners, that the superior moiety of the ovary remains 
almost unchanged, while the principal growth takes place in the 
lower moiety, from wdiich it results that a half-superior ovary 
becomes converted into an inferior fruit. Now, precisely the 
same occurrence takes place in Bursinopetalum ; and if on this 
account we were to deny its right to rank in Aquifoliacece, then, 
for the same reason, w'e ought to exclude Halesia from the Sty- 
racece, which few would venture to propose. For the same I’eason 
that the Halesieee have been made a tribe of the Styraceoe, Bursi- 
nopetalece may be considered a second tribe of the Aquifoliacece, 
the Ilicinece being the first. 
I will add, in justice to Mr. Thwaites, that three years sub- 
sequently to his note before mentioned, in his 'Enumeration of 
Ceylon Plants’ (1858), he abandoned his former conclusion; for 
he there classes Bursinopetalum among the Olacacea, meaning, 
I presume, Mr. Bentham’s tribe Icacinece. Mr. Thwaites has 
thus ignored the unquestionable grounds upon which the Icaci- 
nece must remain separated from Olacacece. The Icacinacece, I 
have shown, must be a distinct order, contiguous to the Aqui- 
foliacece, the one only dififering from the other in. the sestivation 
of the corolla. This last ai’rangement of Mr. Thwaites brings 
Bursinopetalum close to the place I have assigned it ; and if the 
aestivation of its corolla had been truly valvate, his determination 
would have been perfectly correct, according to the Candollean 
arrangement ; but as the case is otherwise, the genus falls into 
Aquifoliacece, under the condition above proposed. 
Mr. Thwaites is perfectly correct in his statement that the 
raphe is on the face of the seed opposite to the pi’ojection in the 
cell round which the seed is folded, but is wrong in his infer- 
ence that, because it is so, such a projection can have nothing to 
do with the placenta : the fact is that not only in this instance, 
but in others of this family, as also in different orders of the 
* Contributions to Botan}’, i. 168; Ann. Nat. Hist. 3 ser. hi. 137. 
