HERPETOLOGY OF JAPA]<r. 459 



The young have clearer, brighter colors, more contrasting pattern, 

 and the tail always light colored toward the tip both above and below. 



Variation. — The most obvious instability in the scale formula is 

 ■the varying number of temporals, but this variability is only confined 

 to the upper scales near the parietals. As a matter of fact the outer 

 edges of the later shields seem to be on the verge of breaking up into 

 small scales as evidenced by the many nicks and half-finished sutures 

 on their outer margins. The prefrontals also show sign of breaking 

 up, and there is a distinct regularity in this process. In nearly all the 

 specimens the inner anterior corner of these sliields is more or less 

 convex, and in a number of cases this slightly elevated region is cut 

 off by regular sutures so as to form an additional anterior pair of pre- 

 frontals adjoining the internasals. Such is the adult female, No. 

 34043, from Yokohama. Quite similar in this respect is the adult 

 female No. 15421, from Omachi, province of Shinshiu, and what is 

 even more interesting, no less than three of her five young just born 

 have these accessory prefrontals exactly like the mother, while only 

 two of them are normal. 



The number of supralabials, 7, is remarkably constant. Less so is 

 the number of scale-rows, namely, 21, but 23 is rather of rare occur- 

 rence. The ventrals range between 132 and 146; subcaudals between 

 44 and 56 pairs. 



The coloration is very variable. Wliile specimens lighter than 

 the one described above do occur, individuals much darker seem to 

 be more frequent, and it is not uncommon to find specimens with 

 the whole underside uniformly black or nearly so. "VVliile the tip 

 of the tail seems to be always light yellowish in the 3^oung I find 

 among the adults before me an even proportion of specimens with 

 dark and light tail ends, there being apparently no distinction as to 

 sex or locality. 



The specimen from Hachijo shima (U.S.N.M. No. 36540), the south- 

 ernmost island of the Idzu group, kindly submitted to me by Doctor 

 Ijima, difl^ers greatly in coloration from the normal specimens. The 

 ground color is of a rich reddish brown with nearly all traces of black- 

 ish gone. In all other respects the specimen seems to be normal. 

 Doctor Ijima writes me that he has six more specimens from the same 

 locality of a similar coloration, but adds that the Science College 

 Museum also has a specimen (No. 58) from the sarne island of a dark 

 color similar to those from Hondo. In view of this fact I refrain from 

 imposing any sA^stematic name on this color variety. 



Habitat. — As here understood A. hlomhoffii is restricted to the 

 islands of Japan, being represented on the mainland opposite by 

 closely allied forms, in Korea by A. hlomlioffii Irevicaudus, in the 

 Russian Coast Province bj^ A. hlomlioffii in term eel ius. 



