THE HERPETOLOGY OF HISPANIOLA 



323 



A second specimen in our collection (No. 66713) has a very similar 

 scale formula. In coloration, however, this snake is markedly differ- 

 ent, as shedding of the skin was about to take place when the serpent 

 was captured. The whole body (in alcohol) is a beautiful pale bluish 

 gray, with the saddlelike blotches a very little darker than the body 

 color. Only a trace of the lateral series of spots is visible. The 

 ventral surface is very pale bluish, the chin and throat cream-colored. 

 In both specimens the alternating large and small scales in the first 

 scale row are very noticeable (see the accompanying figure), while 

 from the second row onward a great diminution in length with a 

 corresponding increase in number of body scales seems to have 

 taken place. Though in most nonconstricting snakes and boas the 



Figure 93.— Dorsal pattern of Epicrates gracilis gracilis, same specimen as shown in fig. 92. 



dorsal scales are equal in length and therefore equal numerically to 

 the ventral plates, in this species there are twice as many dorsals 

 in a given length as there are ventrals. A similar but much less 

 frequent diminution of dorsal scales may be noted in E. striatus 

 although the scalation in that species is not so symmetrical and 

 there are no alternating small scales in the first scale row. In the 

 drawing of the type of E. gracilis this unusual scalation was indicated. 



Dimensions: Total length of No. 55025, 700 mm.; tail (defective), 

 100 mm. Total length of No. 66713, 890 mm.; tail, 150 mm. 



Variations.— The only other known specimens of this exceedingly 

 rare snake are the types described by Fischer, one of which was 

 figured in his original description, and three in the Museum of 

 Comparative Zoology. The specimens in the National Museum 

 agree very well with Fischer's figure, although some small differences 

 may be noted. 



In the figure of the type, and also in No. 66713, three scales in a 

 transverse row separate the frontal from the second pair of prefrontals, 

 while in No. 55025 there are five scales making the separation. The 

 second pair of prefrontals is separated from the loreal and preocular 



