NO. 2177. FOSSIL FISHES IN NATIONAL MUSEUM— EASTMAN. 293 



this or a closely related species, from the Madison Valley, Montana, 

 which is shown of slightly larger than the natural size on plate 18, 

 figure 4. 'Mi. Earl Douglass, who collected a small number of fish 

 remains from this locality, is of the opinion that the strata containing 

 them is of Ohgocene or Lower Miocene age. An undetermined spe- 

 cies of Osmerus is apparently indicated by the original of plate 18, 

 figure 5, from the same locahty as the last. Both are preserved in 

 the Carnegie Museum in Pittsburgh. In plate 19, figure 3, is shown 

 an example of an undetermined cyprinodont species from the Tertiary 



of Mexico. 



Family SILURIDAE. 



Fossil representatives of this family are. uncommon, and for the 

 most part imperfectly preserved. Fragmentary remains from the 

 Lower Eocene have been assigned to Arius, and others from 

 the Lower Miocene to Ameiurus. The undermentioned specimen is 

 the only nearly complete example of the latter genus that has been 

 discovered in the fossil state. 



Genus AMEIURUS Rafinesque. 



AMEIURUS PRIMAEVUS, new species. 



Founded upon a unique specimen without locality label, but as far 

 as may be judged from the appearance of the matrix it would seem 

 to have been derived from the Green River Eocene of Wyoming. 

 Clearly related to the more generalized and representative genus of 

 modern North American catfishes Ameiurus, it differs from all living 

 species in its shorter anal, oonsistmg of only 12 rays, and in this 

 respect resembles the more specialized genus Leptops, known by a 

 single species. It is shown in plate 20. 



The holotype of this species is a nearly complete skeleton having a 

 total length to the base of the caudal fin of 21.5 cm., in which the 

 length of the head including the supraocoipital process is contained 

 three times. Head broad behind, not much contracted foi-wards; 

 surface sculpture consistmg of anastomosing rugae and pittmgs as 

 in the recent A. catus; orbits small; fontanclle situated just before 

 the origin of the supraocoipital process. Maximum breadth of head 

 in its flattened condition equal to its length. Vertebrae 29 in number, 

 with strong neural spines. Dorsal fin between the pectorals and ver- 

 trals, with about six branched rays of moderate length (its spine 

 broken away). Pectoral spines strong, smooth, less than half as long 

 as the head. Ventrals with eight, anal with apparently not more than 

 twelve rays. Rays of caudal fin mostly broken away, but the curved 

 line formed by their articulation with hypural fin-supports indioatmg 

 that the fin was rounded. (Cat. No. 8122, U.S.N.M.) 



This is an extremely interesting and w^ell-preserved specimen, save 

 that the caudal fin is for the most part lackmg, and the dorsal has 

 been folded under the neural spmes and partly concealed hj them. 



