Jordan and Evermann. Fishes of North America. 861 



or bony plates, but rarely ctenoid. Lateral line various, usually undu- 

 late or with an anterior arch and a posterior straight part, at least not 

 regularly arched; sometimes wanting. Flesh in typical forms firm, oily, 

 and reddish in color, but in some cases pale and soft. Caudal peduncle 

 almost always slender and strong, the caudal fin if present more or less 

 deeply forked, except in certain deep-sea forms arid in aberrant families, 

 the structure typically adapted for swift propulsion. Dorsal fin usually 

 long, the spinous portion generally shorter than the soft part, sometimes 

 absent; the spines seldom very strong, sometimes not differentiated from 

 the soft rays; anal fin always long, more of less similar to soft dorsal. 

 Ventrals thoracic, subjugular, or subabdominal, usually with 1 slender 

 spine and 5 rays, sometimes many-rayed, sometimes rudimentary or wholly 

 wanting. Branchiostegals few , usually 7. Gills 4, a slit behind the fourth ; 

 gill rakers various ; gill membranes usually separate, sometimes joined 

 together, rarely attached to the isthmus. Mouth and dentition various. 

 Skeleton firm or variously soft, the structure as in spinous-rayed fishes 

 generally; the shoulder girdle attached to the cranium by a distinctly 

 forked post-temporal, which is not adnate to the cranium. Vertebra) 

 varying from 24 to more than 100, the high numbers found in pelagic spe- 

 cies. Intestinal canal generally short. This group of mackerel-like fishes 

 is not capable of exact definition, its deviations from the ordinary type of 

 spiny-rayed fishes being various and in various directions, so that no set 

 of diagnostic characters will cover them. The group is not a suborder as 

 the term is generally understood; it is incapable of simple definition, and 

 in its divergence some members approach to other groups more nearly 

 than to extreme or even to typical members of their own. The group is, 

 however, a somewhat natural one, as by the common consent of ichthy- 

 ologists its different types have always been kept near each other in the 

 system of classification. This naturalness, however, has been obscured by 

 the intrusion of aberrant forms, as Zeus, Oreosoma, Capros, Antigonia, Pem- 

 pheris, and Lampris. These forms have no near relationship to Scomber, 

 Caranx, or Coryphcena, the central groups of the Scombroidei, and it may be 

 that none of them are Scombroid fishes at all. Perhaps the earliest type 

 of existing Scombroids is that of the genus Lampris ; of the others, the 

 Scombridce proper form a center of divergence. The Zeidce and Caproidcv 

 are certainly not Scombroid fishes, nor are they at all closely related to 

 each other. While it is uncertain just where they should go in a natural 

 arrangement, we have thought best to remove them from the Sombroidei, 

 with which they have usually been associated by recent writers, and 

 reserve them for insertion elsewhere. From Scomber-like forms are 

 descended on the one hand the GempyUdce, Lepidopidm, and Trichiuridce. 

 Another line of descent leads to Acanthocybiince, Istiophoridcc, and Xiphiidce'^ 

 still another apparently runs through Carangidcv to Nomeidce and perhaps 

 to Coryplicenidm and Bramidce. Other forms, as Tctragonuridcc, are appar- 

 ently not Scombroid, but their line of descent is not apparent and at 

 present, knowing no better place for them, we leave them associated with 

 the Scombroidei. (cvcd^poj-, mackerel; 160$, likeness.) 



