I915.] F. H. GRAVELY : Oviental Tarantulidae. 435 
The finger may be armed at the base with 0, I or 2 dorsal 
spines, which remain throughout life. It is always unarmed 
ventrally. 
Subfamily CHARONTINAE. 
The structure of the second visible abdominal sternum and of 
the margin of the carapace opposite the lateral eyes, the relative 
lengths of the two dorsal spines on the hand and of those on the 
end of the arm, and the jointing of the finger and of the tibia of 
the fourth leg, appear to be the principal characters that have been 
used in the definition of genera. 
The structure of the posterior margin of the second visible 
abdominal sternum seems to be very variable, and I am unable to 
attach any importance to it. 
The segmentation of the hind tibiae is often less marked in 
the young of species in which it occurs, than in adults; it is 
sometimes variable within the limits of a single well-marked 
species, and it reaches its highest development in more than one 
genus, among them the specialized cavernicolous genus Stygo- 
phrynus. There can, I think, be little doubt that the extent of 
this segmentation is a mark of the degree of specialization in the 
species in which it occurs. Probably increased segmentation 
facilitates in some way the activities of the animal exhibiting it, 
and may appear independently in different branches of the 
subfamily. It is also found in the genus Damon of the subfamily 
Phrynichinae. 
In species in which the hind tibiae are normally not more 
than 3-jointed, the tarsi (excluding the metatarsi) appear to be 
invariably 4-jointed. In most species in which the hind tibiae are 
4-jointed the tarsi are 5-jointed. Sarax javensts is the only species 
known to me which appears to have both tibiae and tarsi of the hind 
legs 4-jointed, and as I have only one specimen before me the tibiae 
may be abnormal. The structure of the tarsi appears to be con- 
stant within the limits of each species, whereas in Phrynichosarax 
cochinensts and singapurae, and perhaps therefore in other species 
also, the structure of the tibiae is variable. Although, therefore, 
the structure of the hind tibiae is usually much easier to distin- 
guish than is that of the tarsi, it seems best to use the latter 
rather than the former for the separation of genera. 
The structure of the margin of the carapace appears to be of 
more fundamental importance from a taxonomic point of view 
than is the structure of the legs. By its means the subfamily may 
split into two distinct groups. One of these, which may be termed 
the Sarax group, includes only small species whose distribution 
extends from India through Malaysia as far as the Solomon Islands. 
The other, which may be termed the Charon group, includes 
the small species found on the outskirts of and beyond this area 
from the Seychelles to the Galapagos Islands, together with the 
large and highly specialized species belonging to the genera Stygo- 
phrynus and Charon. 
