260 ON THE GENUS VILLARESIA. 
what resembling that of Liriosma. Of these varied conditions in the 
development of the calyx, we find no parallel in Jcacinacee. 
There is much to remark in regard to the development of the ovary 
and seed. In Jeacinacee, the ovarium in most of its genera is l- 
locular, but it is only so by the early abortion of its other cells, as its 
gibbous form sufficiently indicates. This is manifested in Pennantia, 
where the ovary, from this cause, is sometimes 1-locular, but at other 
times it is distinctly 2-locular, with a complete dissepiment, having a 
placentary line in the axis communicating with the style and with the 
base of the cell, and along which the nutrimentary and fecundating 
vessels pass in opposite directions, meeting at a point a little below 
the summit of the dissepiment, from which 2 nearly collateral ovules 
are suspended in each cell, precisely in the same position as when - 
it is unilocular; invariably only a single seed is perfected in each cell. 
When the ovary is l-loeular, we may easily trace the same line of 
"vessels, together with the vestiges of the abortive cells, along the wall 
of the ovary, which is consequently thickened in that direction. In 
Eimmotum, as Mr. Bentham himself showed many years ago, the ovary 
at the time of flowering is completely 5-locular, and sometimes has a 
perfect seed in each cell at maturity, though more generally 3 or 4 of 
the cells become abortive. This direct connection of the style with the 
placentze, through the medium of an axile line of vessels in Zcacinacee, 
is in conformity with the usual organization in Phanerogamous plants. 
But the Olacacee, together with Styracee, Santalacee; and a few . 
other families, form an exception to this general rule, the normal con- 
struction of their carpels being fundamentally of an opposite character. 
In Olacacee the ovary is constantly 1-locular in its upper portion, and 
is either eontinuously so to the base, or it is there divided into 2, 3, or 
4 short basal pouches, formed hy as many very short rudimentary dis- 
sepiments, whieh gradually vanish upwards into as many carinated or 
nerve-like projections along the wall of the cell; these pseudo-dissepi- 
ments are united at the base upon a placentary column, which rises in 
the centre in the form of a short free cylinder, from whose summit a 
corresponding number of ovules are suspended, so that a single ovule 
hangs in the cavity of each pocket: in Opilia, which has no pseudo- 
dissepiment, there is only a single ovule suspended from the apex of 
the central column. I proposed some years ago to call this placentary 
column a cionosperm, and to unite all the families provided with this 
