A MONOGRAPH OF THE SHEA-SNAKES (HYDROPHIIN2),. 173 
cens, and torquata, and very small in Thalassophis annandalei. With rare exceptions 
the costals are of equal or sub-equal size in the whole body length. In the genus 
Enhydris, however, there is a notable enlargement of the three or four rows near the 
ventral median line. This point is not very well brought out in figure 63, which is 
intended to show the imperfect ventrals. In Emydocephalus tjime@ some vertebrals 
are much enlarged, but even in this species this is not a constant feature throughout 
the same individual, nor in different individuals. 
Shape.—As a general rule the costals are rhomboidal when imbricate, more or 
less hexagonal where juxtaposed. The edges of the scales are peculiar in one species 
especially, viz., Astrotia stokest, where in the lowest rows they are irregularly dentate, 
and the apices emarginate (see fig. 66 D). 
Carination.—In many genera the costals are quite smooth as in Platurus, Emydo- 
cephalus and Arpysurus, but in other genera they are furnished with short, median 
keels or tubercles to which many authors have attached considerable importance. 
Personally I do not share their views. Much attention to this character leads me to 
think little if any weight can be attached to it, either in the separation of genera or 
species. I find the degree to which these tubercles are developed varies very much in 
individuals from birth to maturity, and in individuals of the same species of similar 
growth. It is not unusual to see young specimens with these tubercles so little in 
evidence that the scales feel smooth, or almost smooth to the touch, and to meet with 
old examples which are very rough to the touch. Some authors are inclined to think 
the degree of development dependent upon the sexes, the males especially showing 
more pronounced tuberculation. Inthe case of Enhydris curtus, I have seen speci- 
mens in which the lowest and enlarged costal rows have the tubercles modified, so as to 
form spines resembling in size and shape the teats of some small mammals like the 
guinea-pig. Mr. Boulenger believes this occurs in males (Catalogue, Vol. iii, page 
300), but my notes on this point are too imperfect to criticise this,or the belief enter- 
tained by others that males are more strongly tuberculate than others. In many 
species the scales are bi-, tri- or pluri-tuberculate or spinose, but I cannot see in this 
condition any means of assisting the classification of genera or species. 
HEAD SHIELDS.—The actual presence in their entirety of many of these shields 
is of great importance in the separation of genera and species, but the relationship of 
these shields is of far less importance, and very secondary to most characters which 
affect the ventrals and costals. I find that the relationship of many of those shields 
which preserve the greatest degree of constancy in individuals, and which one must 
employ in the separation of species, is open*to some variation in certain of these 
species, and it is therefore impossible to lay down hard-and-fast rules regarding head 
shields for distinguishing the various forms. A very open mind must be kept to prevent 
creating new species on insufficient grounds. 
I find these shields in most, if not in all, species very prone to become rough and 
granular with age. In the young they are usually quite smooth, and often glossy, but 
in very old specimens the asperities are sometimes very pronounced. A good example 
