1905.] JAPAN AND THE LOO CHOO ISLANDS. 513 



in this area are peculiar to it, except Lachesis okinavensis, which 

 encroaches upon the Japanese area on Yaku, and Tropidonotus 

 pryeri, which extends further south into the next division, (3) The 

 South Loo Ohooan area, comprising the islands of Miyako, Irio- 

 mote, and Ishigaki. Only 15 specimens were collected in this 

 area, including 5 species. Two of these extend to Formosa, viz., 

 Lachesis mucrosqiuimatus and Dinodon rv,fozonat'iis ; one is common 

 to the whole Loo Chooan area, viz. Trojndonotus pryeri ; and the 

 remaining two hitherto have only been obtained from this area, 

 viz. Colaber schmackeri and Ahlahes hermince. 



Group 1. Japanese Snakes. 

 Family CoLUBRiDiE. Subfamily Colubkinje. 



1. Tropidonotus vibakabi. — The 14 specimens were all obtained 

 from Japan (Hondo)*. The supralabials were eight with the 

 fourth and fifth only touching the ej'e in three specimens, and 

 eight with the third, fourth, and fifth touching the eye on 

 one side in one specimen. The anterior chin-shields touched five 

 infralabials in one specimen. One example captured in July 

 contained five eggs, the largest measuring It^^q-" X i^". 



2. Tropidonotus tigkinus. — Of 56 specimens, 53 were collected 

 in Japan (Hondo), two in Yaku, and one in Tanega. The supra- 

 labials Avere eight with the foui'th and fifth touching the eye in 

 two, the postoculars two on one side in one. The loreal was con- 

 fluent with the postnasal on both sides in three, and on one side in 

 one. A toad had been ingested in two examples. 



3. DiNODON JAPONicus. — A single normal specimen was obtained 

 from Yamanashi (Hondo). The ventrals were 202 and the sub- 

 caudals 74. 



4. Coluber conspicillatus. — Of 19 specimens, 18 were from 

 Japan (Hondo) and one from Tanega Island. The temporals 

 were two in one Hondo specimen, and the scales 23 in another 

 Hondo specimen in the middle of the body. Yentrals and sub- 

 caudals were as follows in five specimens: — 209-^71, 203 + 64, 

 214 + 71, 217 + 66, 221 + 69 (the fourteenth and fifteenth sub- 

 caudals entire). The Hondo adult examples, with one exception, 

 were uniform brown with each scale outlined darker ; the belly 

 rose or cherry-coloured, with a double series of large, rectangular, 

 black, median spots frequently confluent across the belly ; two 



* Bouleuger, op. cit. vol. i. p. 221, mentions Formosa as a habitat for this species. 

 If this is correct, it is a very singular circumstance. There are no other instances 

 of Japanese species occurring in Formosa, except Dinodon rufozonatus and Ancls- 

 trodon hlomhoffii. The existence of both in Formosa is easily understood, since they 

 are common snakes on the Chinese mainland, and the Formosan Snake fauna is com- 

 posed almost entirely of Chinese species. In the case of Tropidonotus vibakari, 

 however, it is significant that it has not been recorded from China (except Manchuria) 

 and it does not occur in the Loo Choos. I am prompted to regard this record as an 

 eiTor. 



