HEEPETOLOGY OF JAPAN. . 401 



elusion that the males differ in having fewer ventrals, fewer scale 

 rows around the neck, more anals, longer tail and smaller circum- 

 ference of body relative to total length, and much more developed 

 spines on the ventrals. In all three species examined by him he found 

 a chfference of about 20 ventrals between the maximum of the males 

 and the minimum of the females. Thus in Lapemis hardvncMi 

 {Hydrophis loreatus) he found in the males 135-168 ventrals (average 

 153) and in the females 186-237 (average 202), the difference between 

 the averages being 49 ventrals. His result was calculated to carry 

 conviction as it was based upon a series of no less than 31 specimens. 

 Nevertheless, if we inquire into the data furnished by Boulenger 

 regarding the specimens of the same species in British Museum "■ 

 we find in a much smaller series (14 specimens, 5 males and 9 females) 

 proportions essentially different. In his list the ventrals of the males 

 range between 130 and 155 (average 140), those of the females 136-200 

 (average 156), consequently an overlap of 19 ventrals instead of a gap 

 of 18. The British Museum specimens are not from one locality, 

 however, but cover the whole area from India to the Philippines, and 

 it might be supposed that this cUscrepancy had a geographical signifi- 

 cance. Such is not the case, however, for three females from Negros, 

 Philippine Archipelago, consequently from practically the same 

 locality as Boettger's specimens, have 140, 156, and 180 ventrals, 

 respectively, thus nearh^ covering the whole range of Boettger's 18 

 males (135-168) without reaching the minimum of his 13 females 

 (186). 



These apparent discrepancies seem at present inexplicable, and 

 indicate one of the many difficulties with which the student of these 

 snakes has to deal. 



My own observations bear out Boettger's conclusions only partly. 

 In our four specimens of Disteira godejfroyi the female has more scale 

 rows, more ventrals, and fewer subcaudals. On the other hand, in 20 

 specimens of Disteira melanocephala there is apparei,itly no difference 

 among the sexes except that the males average a few more subcaudals 

 than the females. 



KEY TO THE GENERA OF HYDRIN.E IN JAPANESE AND CHINESE WATERS. 



a' Ventrals wide, at least four times as wide as long. 



6' Nasals widely separated by internasals; supralal)ials nornuil . . .L«/(Va(/f/o, p. 402. 

 6^ Nasals broadly in contact on median line of snout; middle supralabials fused into 



a large plate on each side Eiiii/florcphahis. p. 413. 



or Ventrals rudimentary, less than twice as wide as long. 



6' Two pairs of chin-shields, anterior always, posterior nearly always in contact. 



Distrira. p. 418. 

 b'- Chin-shields scarcely differentiated, not in contact. 



"Cat. Sn. Brit. Mus., Ill, p. 301. 

 26485— No. 58—07 26 



