VOL.XVI,"] 

 1893. J 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



739 



They show, ainoug other things, that the relative number of com- 

 ])k'te and incomplete '' circular folds" relied upon by Boulenger in con- 

 structing his key to the species of this genus (Cat. Bat. Grad. Br. Mus., 

 1882, p. 96) is of no value. It is plain from the appended table that 

 Avhile in some of the specimens " nearly all the circular folds* com- 

 ]»letely surround the body," in others the majority of these folds are 

 widely separated on the anterior portion of the back, a few nearest to 

 the head being complete, however, in most cases. On the ventral sur- 

 face all the rings counted are continuous, the lateral impressions on 

 the posterior portion, which were not counted, alternating with the 

 comi)lete rings. It seems, therefore, better to rely upon the smaller 

 number of rings and their in(;ompleteness on the anterior portion of 

 the ventral surface in separating H. guentheri^ from H. rostratus. 



List of specimens. 



Hypogeophis alternaus, sp. no v. 



Diagnosis. — About 163 to 175 folds, the posterior 40 to oO complete 

 on the ventral and dorsal lines; the posterior 70 to 86 complete across 

 the dorsal surface as well, while anterior to these, above and below, 

 the complete primary rings alternate with secondary rings broadly 

 interrupted on the dorsal and ventral lines; snout shorter than width 

 of head across the eyes; tentacle halfway between and below eye and 

 nostril. 



Habitat. — Seychelle Islands. 



Type.—U. S. National Museum, No. 2()4i8; Mahe, Seychelles; Dr. 

 W. L. Abbott coll. 



* As "circular folds" I have only counted those which are visible on the 

 tipper and lower surfaces, whether interrupted on the middle of the hack and Itelly 

 or not. I have consequently left out those short impressions on the posterior half 

 of the body which are only visible if counting along the sides and which can not by 

 any stretch of language be termed "circular." As a result I count 105 to 111 circular 

 'folds against Boulenger's "about 125." 



t. Fudging from the number of rings and their completeness (»n the back of the 

 full-length figure on plate vii (Cat. Bat. Grad. Br. Mus., 1S82) it represents H. ros- 

 tratus rather than H. guenthcrl, although so designated. 



