328 MR. F. N. WILLIAMS : REVISION 
more workable than one based on a combination of theoretic 
characters, which would either fail to include some species, or 
result in the overlapping of groups. 
Most of the other proposed genera frequently included under 
Arenaria, in its comprehensive sense, are best considered as 
primary subdivisions of <Alsine. The genus Arenaria, in this 
paper, is defined more in accordance with Fenzl's circumscription 
of the admitted species, and as propounded by him in Endlieher's 
* Genera Plantarum. 
It is difficult, in the species of a genus like Arenaria, to select 
constant charaeters which may seem to define satisfactorily and 
in a natural manner the primary subdivisions of the genus. 
After examination of a large number of specimens, both in the 
herbarium and in the living state, I have decided to base the 
primary subdivisions of the genus on the association of two 
characters which seem fairly constant, viz., the structure of the 
disk, and the number of teeth formed by dehiscence of the ripe 
capsule. This latter character is generally dependent on and 
associated with a definite number of styles, but not invariably. 
For, though one of the important generic characters is that the 
capsular teeth are twice the number of the styles, it is found 
that the number of styles sometimes varies in specimens of the 
same species, or even in flowers of the same specimen, while 
the capsular teeth are not so liable to casual variation. In such 
cases, the furrows on the surface of the ovary, which are present 
in many species, and which correspond with the usual number of 
styles, are a more trustworthy character, though these furrows 
usually disappear with the distention of the ovary after fertiliza- 
tion. The structure of the disk is a character which few botanists 
besides Edmond Boissier have noticed in their descriptions of 
species, and the verification of this character has entailed a 
considerable examination of material. Plants belonging 
Arenaria are of little horticultural interest, and it is difficult to 
obtain for examination an extensive series of living oF fresh 
specimens, Several species grown and cultivated in the Her 
ceous Ground at Kew, which have flowered in the course of the 
past few seasons, have afforded advantages for examining Spec 
mens of different groups side by side under favourable conditions. 
After carefully studying the delimitation of other gene" © 
Caryophyllacez formed on the basis of analogous fruit-characte™ 
and their association into groups, and taking into accoun 
