COXTRIBUTIONS TO BOTANY. 
27 
EbenacecE, as before suggested, appear to belong to the neigh- 
bourhood of the Anomcece rather than of the Aquifoliacece, with 
which family they are strangely consociated by Dr. Lindley 
(Veg. Kingd. p. 594') in the same alliance with the Gentianacea , 
Apudjnaceee, &c. The affinity of the Styracea with the Hu- 
iniriacece has been already indicated. The Primulacece, together 
with the LentibulariacecE, appear to have more relation with the 
Plautaginacefe and HydrojAiyllacecE, an alliance that differs little 
from that shown by Dr. Lindley (Veg. Kingd. p. 637). The 
farther prosecution of these considerations would be foreign to 
the present purpose, and they are now only indicated with the 
view of assisting us in the determination of the true affinities of 
the OlacacecE. 
There is yet another family, to which the Olacacece, compre- 
hending all the genera included in it by Mr. Bentham, will 
be found to offer many points of approximation, — I mean the 
Aquifoliacece of DeCandolle, the Ilicinece of Brongniart, Endlicher 
and others ; but I am not aware that this affinity has been before 
noticed. INIany species of Ilex bear much the habit of the Ola- 
cacece and differ little in the structure of the flower from the 
tribe Icacinece, except in the aestivation of the corolla and the 
unilocular apex of the ovarium. Leretia, indeed, bears a re- 
markable resemblance in its habit and inflorescence, and in the 
structure of its flowers, to a Brazilian species of Villaresia, dif- 
fering principally in the aestivation of the corolla, and in the 
want of an inner carinated midrib in the petals ; but in other 
points of arrangement there is very little variance, agreeing even 
in its unilocular ovarium, with two collateral ovules suspended 
almost parietally from near the apex of the cell. The stimcture 
of the fruit of Villaresia corresponds so far with that of the Ola- 
cacece, in having a single seed, with copious albumen, containing 
a small embryo near its summit, with a superior radicle, and 
small cotyledons. It may be well here to mention a fact, ap- 
parently yet unknown, which may serve to throw some better 
light upon the real affinities of the Aquifoliacece. I have found 
that the suspension of the ovules in the ovarium of Villaresia 
is not really parietal, as generally stated, for it is sometimes 
completely bilocular, with two ovules in each cell, collaterally 
suspended from each side of the dissepiment by a cupshaped 
strophiole, like that seen in the ovules of the Celastracece ; but in 
ordinary cases the ovarium is unilocular, only by the suppres- 
sion of one of the cells, and the confluence of the dissepiment 
with the pericarpial covering, for it is then always somewhat 
gibbous, and its wall much thicker on the side of the abortive 
cell, towards which the style is then constantly somewhat lateral : 
this fact serves to bring the genus completely within the pale of 
E 2 
