A MONOGRAPH OP THE SEA-SNAKES (HYDBOPEIINM). 223 



these on one side only.) Temporals, — usually broken up and replaced by two or 

 more superposed scales anteriorly. (A well developed single anterior shield occurs in 

 sixteen specimens, 1 and in four of these on one side only.) S u p r a 1 a b i a 1 s, —subject 

 to great variation, the third and succeeding shields subject to division; the third and 

 fourth , and usually the fifth, touch the eye. Infralabials, —four ; the last in con- 

 tact with three or four scales behind ; the suture between the first usually less than 

 that between the anterior sublinguals. Marginals, —usually a complete row after 

 the second supralabial ( sometimes one , or more after the third ). Sublinguals,— two 

 well-developed pairs, the fellows of each in contact. (In six examples the posterior 

 are quite separated.) Cost a Is,— anterior, 25 to 36 (usually 28 to 33) ; midbody 33 

 to 44 (usually 36 to 41) ; posteriorly 34 to 43, imbricate, or subimbricate throughout. 

 Ventrals, — 280 to 397, distinct everywhere; twice or nearly thrice the breadth of 

 the last costal row. 



Colour. — The many varieties have been summed up by Mr. Boulenger, and I 

 have little to add to his arrangement. 



A. Well-defined black bands, more or less connected ventrally. Analogous to 

 variety A. (4) of spiralis. 



(a) All the bands complete. A common form ranging from the Persian Gulf 



to the Philippines. With this I would place the semperi of Garman. 



(b) Some of the posterior bands interrupted costally or subcostally. Not un- 



common. In the British Museum, in Reeves' specimen from China, and 

 others, the bands are briefly interrupted costally. In a specimen of 

 Jayakar's, from Muscat, the interruption is subcostal, and more extensive. 

 Ventral spots occur corresponding to the dorsal bars. With this mac- 

 farlani (Boulenger) should be placed (see fig. 36 on plate viii). It 

 occurs between the Persian Gulf and Australia. 



(c) Some of the posterior bands deficient ventrally, and thus converted into 



bars. Xot uncommon. With this I would include kingi (Boulenger) 

 from Australia. Analogous to variety 6(a) of spiralis. 



B. Well-defined black bands not united ventrally. A common form occurring 

 between the Persian Gulf and China. With this I would place the tuber culata of 

 Anderson. Analogous to variety brugmansii of spiralis. 



C. Obscure bands or bars. A common form usually met with in adult specimens, 

 and occurring between the Persian Gulf and the Philippines. With this, I think, 

 should be included the crassicollis of Anderson, the grandis of Boulenger, aspera of 

 Gray, and pacificus of Boulenger. Analogous to specimens of variety A(6) of spiralis. 



D. Well-defined dorsal bars. 



{a) Costal, and subcostal spots. An uncommon form from Australia, viz., the 

 elegans of Gray (see fig. 3j on plate viii). Somewhat comparable to 

 variety A(5) of spiralis. 



1 Five of these are British Museum specimens labelled cyanocincta. 



2 None in one specimen, viz., belcheri, in the British Museum, 



