Cretaceous Fishes from Mount Lebanon. 409 



preceding species, differs from all the known forms of T/irisso- 

 pteroides in its general proportions ; the length of the head 

 with opercular apparatus not much exceeding the maximum 

 depth of the trunk and contained nearly three times in the 

 length from the pectoral arch to the base of the caudal tin. 

 The axial skeleton and fins are essentially identical with 

 those of the last species, and there are no ridge-scales. 



4. C/upea curta, J. W. Davis, loc. cit. p. 579, pi. xxxiii. 



tig. 5. [= OsmeroideSj sp.j 



The crushed and distorted fragment thus described is 

 clearly excluded from the genus Clupea by the shortness of 

 its ribs and the absence of ventral ridge scutes. It is, 

 however, too imperfect for precise determination. The axial 

 skeleton is most suggestive of that of Osmeroides, and it is 

 quite possible that the fossil may belong to the species of this 

 genus to which Davis gave the name of Clupea Leioisi (loc. 

 cit. p. 571, pi. xxxiii. fig. 1). 



5. Clupea attenuate, J. W. Davis, loc. cit. p. 580, pi. xxxiii. 



fig. 4. [= Osmeroides attenuatus.~\ 



The cranium in this specimen is almost completely 

 destroyed, but the parasphenoid is shown to be straight and 

 comparatively stout. Below this are remains which may 

 perhaps be interpreted as a fragment of a stout maxilla. As 

 noted by Davis, the mandible exhibits traces of very minute 

 clustered teeth at its symphysial end, while the outer face of 

 the dentary is marked by two irregular longitudinal series of 

 pits, evidently connected with the sensory canal. Between 

 the mandibular rami there are fragments which might be 

 parts of a gular plate; but this is uncertain. The opercular 

 apparatus of the right side is imperfectly exposed from 

 within, and there seem to be traces of a much-expanded pre- 

 operculum marked with radiating ridges. The vertebral 

 axis is much obscured by the thick squamationj but the 

 centra in the anterior abdominal region are shown to be com- 

 paratively short and deep, while those in the caudal region 

 are a little elongated. The stout neural spines in the anterior 

 abdominal region seem to be separate from their supporting 

 arches, and the ribs clearly do not completely encircle the 

 abdominal cavity. The stout neural and haemal arches in the 

 caudal region are sharply inclined backwards. The total 

 number of vertebrae is approximately 35 in the abdominal, 

 20 in the caudal region. Remains of one of the pectoral fins 

 prove these to have been small and delicate, while the inser- 

 tion of the pelvic tins is shown to have been opposite the 



