HERPETOLOGY OF JAPAN. 473 



brownish band on top of head covering the two scale rows nearest to 

 the supraoculars and extending backward to the neck ; a well-defined 

 dark-brown band or line, one scale wide, starting from the posterior 

 margin of the orbit below the center of the eye and extending back- 

 ward beyond the angle of the mouth passing the latter on the scale 

 row above the last supralabial; a dark brown vertical line on the pos- 

 terior half of the second supralabial from the pit to the edge of the lip, 

 lower surface pale buff with ill-defined markings of pale brownish gray. 



Dimensions. 



mm. 



Total length 370 



Snout to vent 305 



Vent to tip of tail 65 



In the adult male from Ishigaki shima the coloration is essentially as 

 described above, but the lateral series of spots, one below each dorsal 

 blotch, is as well marked and dark colored as the latter; tip of tail is 

 without dark cross-bars but is not paler than the basal portion. This 

 specimen (Sci. Coll. Mus. Tokyo, No. 18) measures 655 mm. in total 

 length, of which the tail is 123 mm. This is considerable less than the 

 length reached by the species, as one of the types, according to Doctor 

 Boettger, measures 945 mm. 



The edge of the supraocular is turned up in all the specimens, but 

 according to my observations on the specimens in the Hamburg 

 Museum less so in the old specimens than in the young ones. 



Variation. — The present species is rather variable so far as the 

 scutellation of the head is concerned. In Boettger's t^'pes of T. 

 luteus, which he kindly allowed me to examine in 1898, the scales 

 covering the canthus rostralis are small and irregidar. Of four 

 specimens in the Hamburg Museum two have one large canthal scale 

 followed by three small irregular ones, while two have two large 

 scales followed by a small one, and of the three specimens before me 

 two have one and one two enlarged canthal scales. In his t3"pes, the 

 temporal scales are rather strongly keeled ; in the Hamburg specimens 

 three of the specimens have the upper temporals more or less strongly 

 keeled, in the fourth one the keels are only faintly indicated, and so 

 they are in my specimens. In one of the types of T. elegans the 

 temporals are slightly, but distinctly keeled, as already stated. 

 Usually there are two rows of scales between the subocular and the 

 labials, in one Hamburg specimen (No. 2582) only one such row. 

 The number of scales between the supraocidars varies between 10 and 

 13. Supralabials probably most commonly 7, but one of ni}^ speci- 

 mens has S on both sides, and two have 7 on one side, 8 on the other. 

 In the specimen in the Kumamato Fifth Iliglier Middle School the 

 long upper preocidar is divided vert icall}^ near the eye on both sides. 



