5-toothed cup, nearly closed at the orifice; and on the 
outside of this cup are placed the anthere, three to each 
tooth ; the latter are two-celled, with divaricating cells, 
which open longitudinally, and are so entangled with each 
other, that the whole surface of the cup appears, when 
the anthere have burst, to consist of a single, many-celled 
anthera. The pollen is spherical and smooth. The ovarium 
is seated within the cup of stamens, and is so entirely con- 
cealed that it cannot be discovered till some part of the cup 
is removed by violence; it is ovate, smooth, and formed of 
five inseparable cells, each of which has two ovula placed 
one above the other, and attached to their placenta by 
their inner margin ; the stigma is sessile, with five radiating 
lobes. From the. Chinese drawing, the half-ripe fruit 
appears to be fleshy, with five deep angles, and five cells, 
without any remains of calyx, and with a slight appearance 
of separation between the lobes. The ripe fruit is an 
obovate, 5-angled, 5-celled, 5-valved, retuse, woody cap- 
sule, with a loculicidal dehiscence, and no separable axis. 
The seeds are attached one to each side of the valves, and 
are expanded at their lower end into a wing. 
* From this description it is obvious, that, with the 
single exception of the contents of the seed, we are in pos- 
session of all that it is essential to know of the structure of 
this plant. The next subject of consideration is its affinity. 
“The stellate pubescence, the thickening of the petiole 
at the point where it expands into the lamina, the station 
of the stamens upon a long filiform torus, the external 
position of the antherz, and the union of the filaments by 
threes into a cup surrounding the ovarium, are all cha- 
racters that forcibly call to recollection the genus Sterculia. 
The calyx, indeed, in that genus is generally divided much 
more deeply than in the plant now under aggre 
and the anthere are usually seated at the base of the 
ovarium ; but, on the other hand, in Sterculia colorata of 
Roxburgh, which, if a distinct genus (Erythropsis), as I am 
inclined to believe, is nevertheless next of kin to Sterculia, 
the calyx is of the same figure, and divided in the same 
degree, and the anther are also combined in a capitate 
cup enclosing the ovarium. If, however, we pursue this 
comparison further, we find that, with the characters now 
adverted to, the similarity ceases; 1n Sterculia there are 
no petals, the calyx has a valvular, not imbricate, «estiva- 
