358 Mr. CoreBRookE on Boswellia 
pulpy, going to pieces when dry, but not opening by deter- 
minate valves: they contain, in one or more cells, solitary 
seeds; or, if all the ovules ripen, two, or possibly sometimes 
three in each cell. In fact, throughout this family the fruit is a 
berry, in which the natural number of cells, containing two, or 
sometimes three ovules, agrees with that of the petals and caly- 
cine divisions (sepala), and corresponds to twice as many stamina. 
Exceptions are to be admitted, if the plants have been rightly 
classed and described ; for A. acuminata and A. simplicifolia of 
Roxburgh, in which the germ exhibits but two cells, and A. Zey- 
. lanica of Retzius, in which osseous seeds coalesce and present a 
trilocular nut. But all three should perhaps, on account of 
these deviations from the natural structure, be removed from 
their present place to other genera. 
The staminiferous ring around the germ, which was — 
for a distinctive mark of the Icica as a separate genus, is not 
more characteristic according to Willdenow * : for it is found, 
more or less conspicuous, in divers species of Amyris, as it like- 
wise is in those of Bursera : and the insertion of stamina, as well 
as petals, in it, is an important character pervading the whole 
family. 
In truth, as long ago remarked*, the three genera Amyris, 
Icica, and Bursera, require to be thrown together and re-cast. 
The whole group comprises nearly forty species, including seve- 
ral yet unpublished ; and is likely to receive further accessions. 
It may be expected to become unwieldy for a single genus ; and 
it actually comprehends plants which do not assort well together. 
It should therefore be subdivided, and moulded anew into distinct 
genera. But for this purpose much the greater part of the spe- 
cies requires re-examination, with a view to the distribution of 
them by truly discriminative marks. 
* Sp. Plant. ii. $38. + Juss. Gen. Pl.371. Lam. Enc. ii. 768. 
If 
