COLLECTED IN HONG-KONG. 317 
In the economy of the ovary again the most striking similitude ob- 
tains between the two Orders. In both it adheres more or less to the 
: inner wall of the calyx, and in both is usually, but not always, 2-celled. 
In Bruniacee the cells are either 1, 2, or 3 in number. The genera 
which have only one cell are Berzelia and Thamnia. In the former 
instance this is owing to the abortion of a carpel, and in the other to 
the suppression of the dissepiments of an ovary consisting of five car- 
pels. In Ardouinia there are three cells, while all the other genera 
have two. Hitherto the Hamamelidacee have always been considered 
to be a strictly bilocular family, but the discovery of Tefracrypta, in 
which there are four cells, destroys this uniformity. 
In both tribes the number of ovules in each cell is variable. In 
Bruniacee they are always pendulous, and either one or two in number, 
in the latter case being always collateral. In Hamamelidacee they are 
always solitary and pendulous in the normal genera; but in the aberrant 
ones they are numerous, and attached to placentze which adhere to the 
dissepiment. In both tribes the styles are always equal in number to 
the carpels. In Bruniacee they are either distinct or more or less 
united, while in Hamamelidacee they are always distinct. The stigmas 
in Bruniacee are always small and papilliform, and such is al8o the 
case in several of the genera of the other Order; but in others, Eustig- 
ma for instance, they are much enlarged, and covered with carunculate 
papille, which often extend down the inner face of the styles, as in 
Eustigma and Bucklan ; 
In neither of the tribes is there an uniform structure of fruit. In 
both it is more or less capsular, the capsule being of a more woody 
texture in Hamamelidacee than in Bruniacee. In both the seeds are 
pendulous, and furnished with an abundant fleshy albumen. In both 
the embryo is orthotropal, with the radicle pointing to the hilum ; but 
while in Bruniacee the embryo is placed at the apex of the albumen 
next the hilum, and has a short conical radicle and. small fleshy coty- 
ledons, it occupies about the centre of the albumen in amamelidacee, 
has a cylindrical radicle, and the cotyledons are somewhat foliaceous. 
Grubbia and Ophiria, two Cape genera, formerly referred to Santa- 
lacee, but now to Bruniacee, I have not seen. By Endlicher they are 
formed into a distinct tribe under the name of Grubdiacee, which in his 
* Enchiridion Botanicum " he places subordinate to Bruniacee. 
careful examination of them by Arnott, leads him also to consider that 
