316 BULLETIN 17 7, UNITED STATES NATIONAL MUSEUM 



very irregular, owing to the dividing of small portions from it; again 

 there may be a suture lengthwise down the middle, dividing the 

 frontal into two even halves. The only scales that seem to have 

 some permanence in shape and size in this series of specimens are the 

 anterior prefrontals and the supraoculars; the anterior prefrontals 

 are constant in the prolongation backward of their outer posterior 

 borders, and the supraoculars are fairly constant in having a small 

 anterior and a large posterior shield, although occasionally there is a 

 single large supraocular or three small ones. The ring of ocular 

 scales is incomplete in every specimen except No. 60604, where it is 

 complete on the right side only. The number of upper labials entering 

 the eye is usually two, although the entrance of but one is not in- 

 frequent. There is usually one scale, but occasionally two, inter- 

 calated between the loreal and the upper labials; in two specimens 

 (Nos. 60603 and 60604) there is no such shield on the left side of the 

 head. The very elongate loreal is sometimes vertically divided or 

 semidivided. The upper labials vary in number from 14 to 16. The 

 Tortuga Island specimen has 17 labials on both sides, the ninth and 

 tenth entering the eye; on the right side of the head the eleventh, 

 twelfth, and thirteenth are fused as they approach the mouth. 

 Another slight abnormality in this same specimen seems to be the 

 horizontal division on one side of the head of the anterior part of the 

 loreal, which has separated completely from the posterior part. 



The number of scales around the body ranges from 48 to 60 in the 

 specimens before me. Both Fischer and Boulenger record specimens 

 having 63 rows as a maximum. The ventral scales are 272 to 295 in 

 my series, while Boulenger reports one with 302 from Santo Domingo. 

 The subcaudals vary from 85 to 90. A large specimen, U.S.N. M. 

 No. 84260, has a head and body length of 1,723 mm. ; tail, 230 mm. 



In this species, as in gracilis, the dorsal scales compared to the 

 ventral plates seem to have lessened in length and consequently 

 increased in number, although not to so great an extent as in gracilis. 

 In No. 55044, for example, at the center of the body I find 25 ventrals 

 occupying the same amount of space as 35 to 40 dorsals, while in 

 gracilis the proportion of ventrals to dorsals is quite regularly 1 to 2. 

 The increase takes place in the fourth to the eighth dorsal row in 

 striatus. 



The variation in color is not great, the pattern on the whole series 

 remaining nearly as I have described it, with slight differences in 

 intensities on different individuals. 



Relationships. — Epicrates striatus seems to be readily distinguishable 

 from others of the genus by the large numbers of scale rows around the 

 body. The Cuban species, E. angulijer, is the only other known form 

 approximating it in this respect. The presence of a complete ocular 



