NO. 2034. BATS OF THE GENUS GL08S0PHAGA— MILLER. 421 



Diagnosis. — Very similar to Glossophaga soridna mutica but prob- 

 ably averaging larger (the two adults examined are about equal in 

 size to the largest among 24 mutica) ; molars both above and below 

 with crown area sensibly increased. 



Measurements. — For detailed measurements, see table, page 428. 



Specimens examined. — Five (three young), all from Port Antonio, 

 Jamaica. 



GLOSSOPHAGA LONGIROSTRIS Miller. 

 (Synonymy under subspecies.) 



GeograpJiic distribution. — Coast region of Colombia and Venezuela; 

 southern Lesser Antilles (Grenada, Grenadines, Dominica). Limits 

 of range very imperfectly known. 



Diagnosis. — Like Glossophaga soridna, but skull (fig. 16) longer, 

 its condylobasal length usually more than 21 ram. (21 to 22.4 mm.), 

 with rostral portion so much more developed as to appear nearly as 

 long as braincase; outer upper incisor about equal to inner in bulk; 

 pm^ and pm* not noticeably unlike in crown outline; braificase 

 broad and high, its dorsal profile usually forming evident angle 

 with that of rostrum m interorbital region. 



Color. — The color does not differ appreciably from that of Glosso- 

 pTiaga soridna. Among the skuis examined the extremes of colora- 

 tion, blackish, reddish, and pallid, are less frequent than in the 

 smaller animal. 



Shull. — The skull (fig. 16) is distinguishable from that of Glosso- 

 pTmga soridna by its larger size and by the greater development of 

 the rostrum, the region Ijdng to the front of the postorbital con- 

 striction appearing about equal to the braincase in length, while in 

 the smaller animal it is obviously shorter. Dorsal profile not con- 

 spicuously different from that in G. soridna, but braincase rising less 

 abruptly in front, this reduction of abruptness not sufficient to do 

 away with the noticeable angle over anterior part of orbit. In the 

 features not directly affected by the essential proportions of rostrum 

 and braincase the skulls of the two species present no tangible points 

 of difference. Individual variation appears to be about as exten- 

 sive as in Glossophaga soridna, though the material examined is not 

 sufficient to give so satisfactory results as in the case of the smaller 

 animal. In 12 specimens of true longirostris the variation in condy- 

 lobasal length is 1.4 mm., or 6.4 per cent of the mean, while in 17 

 skulls of rostrata it is only 1 mm., or 4.6 per cent. The breadth of 

 braincase in the two races shows a variation of 0.6 mm., or 6.7 per 

 cent, and 0.4 or 4.5 per cent, respectively. Referred to mean con- 

 dylobasal length of skuU the percentages for the two forms are 2.7 

 and 1.8. While the percentages of variation in length and breadth 

 are about the same as in G. soridna the appearance of the skulls is 

 more uniform, probably because the anterior portion of braincase is 

 less subject to change. 



