SNAKES COLLECTED AT HAKGALA, CEYLON. 145 



August. The adult, however, showed nothing unusual upon 

 anatomical investigation. The navel involved 2 shields in 

 both cases. In the 6 10 ventrals intervened between the navel 

 and the anal shield, in the ? 7. 



Sexes. — Of the 48 specimens collected, 21 were males and 27 

 females. Without actually measuring them, I think there is 

 no doubt that the females were usually larger. Adult males had 

 all the scales about the anal region keeled. In the median rows 

 these were tuberculate, and often bi, tri, or pluridentate, and 

 in the lowest lateral rows markedly spinose, with the points 

 directed backwards. The spines were hooked, reminding one 

 forcibly of rose-thorns, and were attached by a long base to the 

 anterior three-fifths of each scale. 



In young males, or at least in some, the keels were obtusely 

 tuberculate as in females. The tails of the males were longer, and 

 the subcaudals varied from 21-26, whereas in females these 

 shields numbered from 11-16. The ventrals in males were 

 however 135-148, against 139-151 in females. 



Scale peculiarities. — The scales anteriorly (two head-lengths 

 behind the head) number 15, mid-body 15, posteriorly (two head- 

 lengths before vent) 15, keels were absent in all these situations. 

 The vertebral row was the narrowest of the series, and the 

 ultimate row very slightly largest. The supracaudals were in odd 

 rows, as is the rule where the subcaudals are entire. The first 

 subcaudal was as often divided as entire. 



Abnormalities. — One specimen had no prseocular. The upper 

 postocular was confluent with the parietal in one. The labials 

 were 7, of which the 4th only touched the eye on one side in one 

 specimen ; 7 with the 5th only touching the eye on the left side 

 in two specimens. 



One very aberrant specimen requires special remark, viz., a 

 male in which the subcaudals numbered only 16. In addition 

 the scales anteriorly (two head-lengths behind the head were 

 15 or 14, mid-body 13, posteriorly (two head-lengths before the 

 vent) 13. Investigation showed that the 2nd and 3rd rows above 

 the ventrals blended on both sides, one rather before the other, 

 at the site I select to count the scales anteriorly. The large row 

 occasioned by this coalescence divided and blended two or three 

 times before the rows were finally established as 13. One sees 

 the same vacillation frequently exhibited in snakes whose scales 

 normally reduce in number, as for instance in Trojjidonotus 

 piscator, where the scales number 19 in the anterior and mid- 

 body, and 17 posteriorly. In all other respects this specimen was 

 normal, and so probably does not deserve specific isolation. 



