414 Bulletin American Museum of Natural History, [Vol. XIX, 



between two prongs the hinder end of the vomer. On the 

 latter there appear to have been four rows of teeth, but many 

 of these teeth are broken away. 



The counterpart of specimen No. 45i6{; (3701) presents the 

 parasphenoid bone. Posteriorly this seems to join a basi- 

 sphenoid. On the lower surface of the latter, in the midline, 

 there is a short, pointed, downwardly directed process. 



All the specimens were collected at Hajula. 



ELOPID^. 

 Holcolepis attenuatus {Davis). 



Clupea attenuata Davis (J. W.) Trans., Roy. Dublin Soc. (2), III, 

 1887, p. 580, pi. xxxiii, fig. 4. 



Osmeroides attenuatus Woodward (A. S.), Ann. and Mag. Nat. Hist. 

 (7) II, 1898, p. 409; Cat. Foss. Fishes, IV, 1901, p. 19. 



No. 4526 (3781), from Hakel, is a specimen which the writer 

 is unable to distinguish from Davis's Clupea attenuata, a 

 species hitherto known only from Sahel Alma. In the present 

 specimen the total length is 95 mm.; to the base of the 

 caudal fin, 77 mm. The head, including the opercular appara- 

 tus, is about 23 mm., but the extremity of the snout is miss- 

 ing. The depth is only 15 mm. There are a few more than 

 50 vertebrae, of which not more than 20 belong to the caudal 

 region. There are 20 interneurals supporting the dorsal fin. 

 Davis reports the presence of only 10 rays in the dorsal, but 

 this is doubtless an error. This fin, in our specimen, is equally 

 distant from the occiput and the base of the caudal fin. The 

 ventrals are inserted below the front of the dorsal. The anal 

 is disturbed, but there are 7 supporting interhaemals present. 

 It is entirely behind the dorsal. The bones of the head are 

 smooth. Nothing can be determined regarding the size or 

 the form of the scales. 



There are various reasons for not identifying this specimen 

 as H. sardinioides (Pictet). It is entirely too slender to be 

 H, lewisi (Davis). The dorsal and ventral fins of the latter 

 are farther backward, and there are said to be 35 vertebrae in 

 the abdominal region behind the operculum. 



