CONTRIBUTIONS TO BOTANY. 
points of approach to the Olacacea, but it must be remembered 
that such a structure is not peculiar to that family, for it is 
found to exist equally in the Santalacece, StyracecB (excluding of 
course Symplocacece), Ebenacea, Myrsinacea, and Theophrastacece. 
We must therefore look to this general character of an unilocular 
ovarium, with a central placenta wholly distinct from the style, 
and more or less free or combined with spurious dissepiments, as 
belonging to a class composed of several orders, just as we unite 
into groups or classes, numerous other families, possessed of a 
bilocular or plurilocular ovarium, and others again that are uni- 
locular with parietal placentations ; and it does not follow, that 
we must associate other plants in Olacacea, merely because they 
have an ovarium constructed in a somewhat similar manner. 
The existence of an inner whorl of petals, the union of the 
stamens into a thick columnar tube, the anthers imbedded ex- 
trorsely in an annular and almost globular fleshy connective, the 
peculiar mode of the dehiscence of the anther-valves, the curious 
sti’ucture of the pollen, the absence of the deep hypogynous disk 
and of the thick epigynous gland, are points quite at variance 
with all we And in the Olacaceoe, where we meet with nothing in 
the smallest degree analogous to the very peculiar features that 
mark Aptandra. However striking its points of approach, it is 
evident that this genus cannot be referred to that family, although 
its position in the system may be proximate. 
There is yet another group of plants ofiering some featui-es of 
resemblance, to which it is worth while to direct our attention ; 
I mean the Canellacea of Von Martins, the characters and real 
affinities of which are yet too imperfectly understood. It con- 
sists of three genera, all with their stamens united into a tube, 
as in Aptandra, and with extrorse anthers, although the cells art- 
said to open longitudinally, but they vary greatly in their other 
characters, and evidently belong to three several families. Pla- 
tonia is clearly referable to the Guttifera, with w'hich it agi’ees 
in having opposite leaves ; Canella probably has a considerable 
affinity with the Humiriacere; and Cinnamodendron (the Canella 
axillaris, Mart. Nov. Act. 12. tab. 3) may perhaps be foimd to be 
allied to Aptandra, for besides its synantherous stamens, it has a 
similar whorl of petaloid scales intervening between the staminal 
tube and the petals. 
Hornshuckia has also a small truncated calyx, a corolla of six 
petals in two series, the inner smaller and carinated, extrorse 
stamens, and a 3-locular (?) ovarium, with a single ovule in each 
cell. 
Much will depend upon the structure of the fruit and seed be- 
fore any final decision can be made in regard to the nearest affi- 
nities of Aptandra, but taking the above-mentioned facts into con- 
