MR DAVIB DON 
extensive monographs of Melastoma and Rhexia^ there does 
not exist almost a single species (as has been justly observed 
by Mr Brown) which really belongs to these two genera. 
The whole of MeIastomace(e agree in some important points 
of structure, the most remarkable of which, first observed 
by Mr Bkown *, is the singular position of the stamens in 
aestivation ; the filaments are inserted in the margin sur- 
rounding the mouth of the calyx ; the anthers hang down 
in a direct position, in the space between the calyx and 
ovarium. Almost immediately on the expansion of the 
flowers, the stamens ascend upwards, and hence they fre- 
quently become declinate, or curved. On the increase of 
the ovarium, the space between it and the calyx is gradually 
filled up ; and, in some cases, the capsule becomes closely 
united to the tube of the calyx, as in Miconia, Cmostegiay 
Tococa, &c. The anthers of all open with terminal pores ; 
and in the greater part of the genera, each anther opens with 
only a solitary pore, but in Bldkea^ and a few others, with 
two : they are all likewise appendiculated at the base. The 
stigma is constantly simple. The capsule varies, with from 
four to six cells, rarely eight, and very rarely but three , 
each cell opens in the middle with an oblong fissure, by 
which the seeds are shed, and the dissepiments are inserted 
into the centre of each valve. The receptacles correspond 
with the number of cells, except in Cmostegia, where they 
are contiguous with the alternate dissepiments, which are 
consequently eight in number. They are sometimes lunate, 
and attached by a short flat pedicel, to the central axis or 
columella, as in Rlieocia ; but in most genera, they are ob- 
long three-sided, and attached longitudinally to the central 
axis. Where they are thick and fleshy the capsule assumes 
*- Observations on thg Botany of the Congo, p. -435. 
