362 TRIANDRIA “TRIGYN iA Houttuynia. 
on the angles of their ovarium, a little above its base —Anthers erect, 
oblong, two-celled, bursting lengthways along their margins. —Qze- 
rium round, three-cornered, most slightly pubescent, one-celled, ma- 
ny-seeded; ovula attached to the sides of the cavity. — Styles three, | 
persistent, rising from the apex of the angles on-the ovarium, above 
the insertion of the filaments, leaving its vertex naked and somewhat 
depressed ; they are spreading awl-shaped, acute, marked along the 
upper part and the furrowed inside (the stigmas?) with numerous ` 
short papillz.— Capsule about the size of a mustard seed, brownish 
green, of the same shape as the óvarium, membranaceous and uni- 
locular, bursting at the top, with a triangular opening reaching half 
way up the styles.— Seeds about eight, brown, smooth, striated, ob- 
. long, acute at each end, somewhat darker coloured: at their bases, 
which are attached to three parietal, oblong, elevated placenta al~ 
 ternating with the angles of the cavity. — Integument single, crusta- 
ceous. Embryo minute, lodged in a copious milk-white, mealy 
perisperm, towards the umbilicus, centrifugal. 
- Obs. Fo the numerous $padices which I have examined I have 
with Father Loureiro invariably found three stamina and as masy 
styles attached to each ovarium, the former above the base, the lat 
ter at the apex of its angles ; I have therefore not hesitated continu- 
ing this most interesting plant in the very class and. order where it 
has been placed in the flora of Cochinchia. As there is no reason: 
for considering it at all different from the original Japan plant Lam 
at a loss to. account for the difficulty which the celebrated Chevalier 
Thunberg experienced in determining its station in the sexual sys 
tem ; nor can there at present be any doubt of its neither belonging. 
to Heptandria, Polyandria, or Monoecia, —— ; 
The seeds are so small and their embryo proportionably minute 
that Í have not been able to ascertain the structure of the latter; 
they appear to me however, i to bear great resemblance to those of . 
Tacca and Aristolochie, to which family the plant is still further ale 
lied by the epigynous insertion of the stamens. But its 
affinity seems to be to the Aroideæ, forming an additional cone 
necting link between the twe mentioned families. ‘The! leaves m 
