420 BULLETIN 58, UNITED STATES NATIONAL MUSEUM. 



The 27 specimens of D. melanocephala examined by me are practi- 

 cally from the same locality, and so is the Miyakoshima specimen 

 recorded by Boett^er. Even the type of the species collected by 

 Sir. E. Belcher may well have been picked up at the same island. In 

 these 29 specimens the rano;e of the ventrals only amounts to 56, while 

 in the 13 specimens enumerated by Boulenger as H. fasciatus it 

 amounts to 147. His 13 specimens, however, cover a range from 

 Malabar to China. If we eliminate the two types of Guenther's 

 Ilydrophis atriceps, with 364 and 353 ventrals, respectively, which 

 I am inclined to refer to D. melanocephala, the variation in the 

 ventrals is reduced to only 98. 



The rational explanation seems to be that the indivi(kial variation 

 in a fairly large series from the same locality is within reasonable 

 bounds, and that the so-called "species" must be broken up on 

 different lines from those which have been established by Boulenger. 



This question is particularly important to us in the present connec- 

 tion, as some of the species which are reported from the waters adja- 

 cent to the region here treated of are very closely related, their 

 differences consisting mostly in slight gaps between their scale 

 formulas. The three species alluded to agree in having a very small 

 head and a long body very slender anteriorly. As a rule they have a 

 single large anterior temporal, and two pairs of small chin-shields in 

 contact. 



Carefully noting what is said in the following under the (Uscussion 

 of each species the various forms treated of may be distinguished by 

 the following: 



KEY TO THE SPECIES OP DISTEIRA IN CHINESE « AND JAPANESE WATERS. 



a' Frontal shield longer than broad. 

 6' Ventrals more than 295. 

 c^ A single anterior temporal (normally). 



d} Scale rows around the neck more than 23 (25-.33). 



<;' Scale rows around body 38, or less (38-32) ; ventrals less than 371 (.301-370). 



D. melanocephala. p. 421. 

 e'^ Scale rows around body 40, or more (40-5(5); ventrals more than 399 



(400-500) [D.fasciataj p. 427. 



fP Scale rows around neck less tlian 23 ( 19-21) [D. gracilis] p. 427. 



c^ Two anterior temporals D. cyanocincta. p. 428. 



b^ Ventrals less than 295 D. godcffroyi, p. 4.30. 



a^ Frontal shield as broad as long [D. viperina] p. 434. 



a The six Snakes mentioned by Peters (Mon. Ber. Berlin Akad. Wiss., 1872, p. 859) 

 under the name of Ilydrophis diadevia Guenther, from the coast of China imd Siam 

 (Berlin Mus. Nos. 4428, 7428, and 4747) can scarcely behmg to that species, if it is 

 identical with D. obsoira. for which Boulenger gives the following scale formula: 

 Scales around neck, 3.3-40; around body. 40-50; ventrals, 310-438. Peters's speci- 

 mens have 31-33 neck scales, 37-40 body scales, and the one specimen of which the 

 ventrals were given has 296. This scale formula does not agree with any of the 

 species as diagnc>sed by Boulenger, 



