CONTRIBUTIONS TO BOTANY. 
the hj'pogynous Polyjietaleie, not far from the Anonacea, rather 
than in the monopetalous group, where it is placed in the ' Pro- 
dromus ’ of DeCandolle, and in the arrangements of other modern 
botanists. 
j\Ir. Bcntham in his memoir before quoted gives his opinion, 
that among dichlamydeous plants, the family of the Hwniriacete 
approaches most to that of the Olacacea ; but in this inference 
he had probably in view bis tribe Icacinece, which I propose to 
remove altogether from the order : 1 cannot indeed perceive any 
close approximation between the two families. In the Humi- 
riacea, the {estivation of the corolla is imbricated or contorsive, 
the stamens are many-seried, and numerous in respect to the 
petals, generally united into a monadelphous tube, or combined 
in jihalanges, and they have a singular expansion of their fleshy 
connective ; the ovarium is surrounded at its base by a thin, and 
somewhat membranaceous dentate ring ; it has four or five com- 
plete cells, which by the thickening of the axile placenta are 
often again divided by a transverse partition. The fruit is a 
berry, having a 5-cellcd osseous nut, the cells being often 2- 
locellate, and the seeds are provided with the usual integumental 
coverings. This is in no way analogous to what is seen in 
Olacacete ; but the Humiriacece present a more manifest affinity 
with the Styracece. 
A considerable degree of analogy between the Myrsinacea and 
Olacacetv is shown in the position of its stamens opposite the 
petals, which present an sestivation so little imbricated as to be 
sometimes mistaken for being valvate ; they agree also much in 
habit and inflorescence. In Icacorea the ovarium is unilocular, 
with four ovules attaehed to a central free jplacenta, of which 
sometimes only one becomes matured, as in T)lacacece; but here 
the analogy ceases, as the {estivation of the corolla is contorsivelv 
imbricate and the seed presents all the characters of the Myrsi- 
nacece. This family has been arranged by most authors among 
the Monopetalca-, but for the reasons before urged in regard to 
the Ebenacete and Styracea, it should be transferred to the Pleio- 
petalece. In Masa, Samara [Choripetalum , A. DC.), and Embelia, 
the corolla is decidedly pleiopetalous, and in the other genera of 
the order the petals are only slightly coherent at base, the 
ovarium being in all cases superior, except in Masa, where it is 
said to be partly inferior, but probably not so at an early period. 
The disposition to produce red dots in all parts of the plant in 
Liriosma, as in the Myrsinacece, is common to several families of 
the ThalamiflorcB of DeCandolle’s arrangement. Some degree 
of analogy may also be perceived between the Mxyrsinacea and 
the Anonacea, Lardizabalaceae, and Menispermacece, in the de- 
velopment of the ovTile, in the arilliform growth of the placentary 
VOL. I. E 
