CONTRIBUTIONS TO BOTANY. 
49 
bating long-established opinions, it becomes essential, at the 
risk of the charge of prolixity, to recapitulate here a few of the 
arguments before assigned, in order to guide us to a right deci- 
sion. Proceeding upon the principle, that the most scientific 
basis for the distribution of plants is that founded on the cha- 
racters which from their nature must be the most invariable, 
I mean those offered by the development of the organs of repro- 
duction, we perceive a feature of most general occurrence, where 
the margins of the carpellary leaves are supposed to become pla- 
centiferous : sometimes these are believed to unite either by their 
edges and thus to form parietal placentations, or by the variable 
degree of their inflexion, to constitute either loculigerous or 
axile placentations ; and from the evidence I shall be able to show, 
it will be evident that it is among the latter we must seek a 
place for the Icacinacece. But I have suggested the existence of 
several famihes, now distributed unsatisfactorily in different parts 
of the system, where we must imagine the placentae to have 
originated, not from the margins of carpellary leaves which in 
such cases may be considered as sterile, but where their o\Tili- 
gerous development is to be traced from the more basal or 
petiolar portions of the carpellaiy leaves : under this point of 
view, we may reconcile the idea of the original formation of an 
ovarium, which, though constituted of several carpels, will some- 
times be unilocular, and at other times often incompletely pluri- 
locular at the base, while in every instance they are all invax'iably 
1 -celled at the summit, the ovniles being always attached to an 
erect placenta arising from the base of the cell, and completely 
unconnected with the style. This extensive group I have pro- 
posed to associate together in a distinct class, the Cionosperma 
{huj. op. p- 8). 
It will be in vain to urge, that there exists only a slight dif- 
ference in the structure of the plurilocular ovarium with axile 
placentation, and the one-celled ovary with central placentation : 
this has been contended by several able botanists, who have 
argued that in such cases the dissepiments have been originally 
complete, but that by their attenuation they have broken away, 
until they have left the placentary column free. I do not deny 
that under certain circumstances this sometimes happens, but in 
these cases we can always trace the indications of such rupture, 
and we also invariably find, that the axile column, rendered thus 
free in the middle, is always attached by its summit to the style. 
On the other hand, in Myrsinacece, we cannot discern any indica- 
tion of parietal expansion, or the smallest involution of what we 
may conceive to have been the sterile margins of the carpellary 
leaves; we find there the ovules often crowded around a free 
globular placenta, rising but little above the base of the cell, and 
springing directly from the pedicel of the flower : here, at least, 
VOL. I. H 
