ON THE GENUS VILLARESIA. 261 
xind of placentation, into one cohort, called Cionospermales. The 
chief distinctive character of this very natural group of families, is that 
owing to the want of connection between the style and placente there 
does not exist that direct line of feeundáting vessels, which in ordinary 
cases descends straight from the stigma to the ovules, reaching the 
point where they meet the nourishing vessels coming from an opposite 
direction ; here, on the contrary, the stigmatic channels for the con- 
veyance of the pollen-tubes must either terminate at the base of the 
style, or must descend along the parietal nerves before described, so as 
to reach the base of the cell, and thence ascending the central column 
in company with the nourishing vessels, in order to arrive at the 
placent. This fundamental difference in the organization of the 
carpels, and in the mode of fecundation of the ovules, assuredly places - 
the whole group in a distinct category.* It a pears, therefore, a vio- 
lation of the rules on which any natural system is founded, adopting 
as its basis the carpellary structure of plants, to unite in one family the 
“ Olaciniee”’ and “ Icacinee,” which in their entire organization stand 
so diametrically opposed to one another. 
But the whole of the case has not yet been stated, for there exists a 
still greater discrepancy iu the structure of the seed. In Zcacinacee, 
although only one seed is usually perfected, its development is pre- 
cisely the same as that in the Zgwifoliacee, and analogous to that of 
the Celastracea, where it differs only in its beiag erect, and often pro- 
vided with an arillus. In all, the albuminous seed is enveloped in two 
distinct integuments, testa and tegmen, furnished with a basal chalaza, 
and with the dorsal raphe, which is peculiar to nearly all the genera of 
the Celastral alliance, and one of its most distinguishing features; the 
embryo, usually large, has a terete radicle pointing to the hilum, aud 
large flat and foliaceous cotyledons. 
In Olacacee, on the contrary, when we break the putamen, we find 
in its single seed an albumen quite naked, and impressed on one side 
with a distinct furrow ; we find also adhering to the inner face of the 
putamen a very delicate membrane, without the smallest trace of either 
raphe or chalaza, but between it and the wall of the shell a free thread 
s structure must not be confounded with that observed in Caryophyllacee 
dum Pts oracee, where the ovary is normally, and at an early stage, completely 
plurilocular, but vere, Ai en the rupture and = of the dissepiment, it becomes 
wholly or od ially un 
