242 



more marginal ones icrminating above and l)clo\v. But few, i. e., the ccnlral, 

 grooves are truly circular. No abdominal carina. 



Tlic two pelvic bones are together truncate heart-shaped, the acuminate 

 apex presented forward. Tlieir posterior portion is a strung transverse rib; 

 anteriorly, each is a thin plate, with thickened outer edge, uuiling with its 

 fellow on the median line. The median portion is so thin as to be readily 

 broken away. The ventral fins are short and wide, with numerous rays. 

 The coracoid bone is a broad lamina, and the pectoral fin evidently had the 

 su])port of rod-like humeral bones of no great length, after the type of most 

 physostomous fishes; but their form cannot be made out. Pectorals not 

 elongate. The anal fin originates but a short distance behind the ventrals, 

 and was not armed with an anterior spine; its length cannot be made out. 

 Immediately above it, a dorsal fin, with slender rays, is represented by the 

 bases of these rays. From above the ventrals to above the distal portion 

 of Ihc pectorals, a line of projecting points appears in the specimen, which 

 I am disposed to ascribe to the articular portions of the interneural spines 

 and attached fin-rays of a first dorsal; but of this I cannot be entirely sure. 



The vertebrae are longer than deep, and present the two deep lateral 

 grooves frequently seen. The number of the cervico-abdominal series is 

 twenty-si.\. The ribs are delicate, and supernumerary ribs are present. 



In comparing this genus with forms already known, points of distinction 

 from 0,11 of them may be detected. Thus, the lack of pectoral sjiine will 

 distinguish it from the known genera of Saurodon/idcc at least. The character 

 of the dorsal fin distinguishes it from Characinidcc, SalmonidcB, etc., which, 

 with the scales, point toward Clupcidoi and Elojndce. From these, the form 

 of the pelvic bones distinguishes it. 



The end of both muzzle and caudal region are destroyed. The latter 

 evidently contracts from the anal fin, and was not probably very elongate, 

 but more as in Elops or Sanrus. 



ApSOPELIX SATJRIFORMIS, Cope. 



Scales large; ten longitudinal series to be counted across the obliquely- 

 depressed body. No lateral line visible. About seventeen transverse series 

 between ])ectorals and ventrals. Ventral broad ; when laid backward nearly 

 leaching anal, but far behind the pectoral. Anal probably rather short; but 

 ' thirj is not entirely certain. Radii, D. ? — 12, P. 16, V. 12. 



