232 TRANSACTIONS OF THE AMERICAN PHILOSOPHICAL SOCIETY 



late Dr. O. P. Hay that in order to "define" a varying group one need only state what 

 forms are referred to it. 



SALMOPERCiE 



In the rivers and streams of Canada and the northeastern United States lives a certain 

 species of small fish, and in the sandy or weedy lagoons along the Columbia lives another 

 and related species, which are of the greatest interest to the student of "missing links," 

 because in their anatomical characters they almost exactly divide the differences between 

 the lower, or soft-rayed, and the higher or spiny-finned teleosts. They have one to four 

 spines on the prominent dorsal and anal fins, the opercular and preopercular both bear a 

 posterior spine, the scales are ctenoid, the ventrals are beneath the pectoral fins, as in acan- 

 thopts. On the other hand, the air-bladder retains an open duct and they have an adipose 

 dorsal fin like the salmonids and other isospondyls. 



Authorities differ as to the proper place of the percopsids in the system of teleosts. 

 Jordan and Evermann (quoted by Boulenger, 1910, p. 620) concluded that they were "ar- 

 chaic fishes, relics of some earlier fauna, and apparently derived directly from the extinct 

 transitional forms through which the Haplomi and Acanthopteri have descended from allies 

 of the Isospondyli." To Boulenger, however, "an analysis of their characters shows them 

 to belong to the Haplomi, of which they may be regarded as highly specialized members, 

 having evolved in the direction of the Acanthopterygii." To Tate Regan (1929, p. 318) 

 they are "an isolated order, without evident relationships except to the Isospondyli or 

 primitive Iniomi." The branchiostegal rays of these peculiar fishes have afforded important 

 evidence on this question, which has been thus stated by Hubbs (1919, p. 68): "The Sal- 

 mopercae, long considered as intermediate between the soft-rayed and spiny-rayed fishes, 

 have six branchiostegals, arranged exactly as in the Acanthopteri. Both of the species 

 usually referred to this group, Percopsis omisco-maycus and Columbia transmontana, have 

 been examined. Aphredoderus sayanus, referred by Regan to the same group, has bran- 

 chiostegals in all essential respects similar to those of Percopsis and the following groups. . . ." 



According to Tate Regan (1929, p. 318) the head of the percopsids, like that of certain 

 berycoids, bears large muciferous channels; those of the frontals are continued forward on 

 the large thin concave nasal bones, which nearly or quite meet in the mid-line. In a skull 

 (No. 981) of Percopsis guttatus in the British Museum (Natural History) the orbits are 

 large with raised rings, the suborbitals small with reflected outer border. The olfactory 

 fossa is also relatively large. The mouth is very small, bordered by minute teeth; the rod- 

 like prcmsxillae have relatively long ascending processes; the maxillae are smaller than the 

 premaxillse and act as levers as do those of acanthopts; no supramaxillae are visible; the 

 mandible is small, with minute dentary and relatively large articular. Many of these 

 characters indicate that the mouth is protrusile. The opercular region suggests the gen- 

 eralized percoid type, although the point on the opercular is feeble. The large preopercular 

 has a reflected border which bears a moderate spine. The occipital condyles are triple as 

 in the berycoid and percoid groups. The general appearance of the skull suggests an ex- 

 tremely primitive percoid type, but with a minimum of crests and spikes, in accordance with 

 the small size, especially of the mouth. 



The otoliths of the Salmopercae, according to Frost (1926c, p. 470), "closely resemble 

 those of Ophichthys gomesii (Order Apodes); in his classification of this order Mr. Tate 



