Mr.J.MiersonGoupia, - = 291 
convex when there are two, or angular when there are three 
seeds in each cell: it is covered by a smooth and very fleshy 
tunic, marked on its ventral face by a prominent line ; this cannot 
be an arillus, because it contains the simple chord of the raphe, 
which extends from a somewhat lateral and basal hilum to a 
small hollow in its summit, which it penetrates to reach the 
second tunic; it is analogous to the same kind of covering as 
that I have described in Magnolia and Clusia, and designated as 
an arilline*, being a fleshy development of the primine of the 
ovule: the next is a hard testaceous shell, regularly oval, densely 
pitted by minute hollows arranged longitudinally in parallel 
lines; it cannot be the testa, because it has no raphe, being a 
growth of the secundine of the ovule; in its summit there is a 
hollow, with a small diapylar hole, through which the chord of 
the raphe passes to reach the chalaza of the inner membranaceous 
integument or ¢egmen, which closely invests the fleshy albumen. 
The embryo is somewhat shorter than the albumen, in which it 
is imbedded, but is much narrower ; the cotyledons are oblong- 
ovate and foliaceous, equal'in length to the terete radicle, the 
extremity of which reaches the base, and is therefore close to the 
hilum of the seed. 
We have here a structure which does not correspond in all its 
characters with any particular family, but which evidently be- 
longs to the Cedastral alliance. It agrees with the Hippocratacee 
in the position of its stamens within a cupular disk, but differs 
in the imbricate estivation of its petals and in the position of 
its ovules. It accords with the Celastracee in its erect ovules 
originating in the base of the cells of the ovary, the external 
tunic of its seeds being an arilline, that is to say, a fleshy coat- 
ing containing the vessels of the raphe, and covering an internal 
testaceous integument; but it differs in the position of its sta- 
mens in regard to the disk, and in the estivation of its petals. 
It agrees with the Jcacinacee in the estivation of its sepals and 
petals, in the long apical inflection of the petals (as in Mappia, 
Stemonurus, &c.), in the hairy appendages of its stamens (as in 
Stemonurus), and in the number of cells in its ovary (as in Em- 
motum) ; but it differs in the position of its stamens with respect 
to the disk, and in its erect ovules and seeds. 
If we maintain any consistency in our demarcations of the 
limits of these several families, Goupia cannot be admitted into 
any of them, and it must stand as the type of a distinct group, 
which I propose to call Goupiacee, distinguished by the follow- 
ing characters :— 
* Linn. Trans. xxii. 80; Ann, Nat. Hist. 3 ser. iv. 26; Contributions to 
Botany, i. 211. 
