782 Bulletin 47, United States National Museum. 



HAPLODOCI: Ventrals jugular, with 2 or 3 soft rays; post-temporal undivided; gills 3, a slit 

 behind the last; no suborbital stay; spinous dorsal very short; no pseudobranchiae; tail diphy- 

 cercal. 



XENOPTEBYGII: Ventrals wide apart, I, 4 or I, 5, a broad sucking disk formed of folds of skin 

 between them; no spinous dorsal; no suborbital ring; no scales; tail diphycercal. 



BLENNIOIDEI: Ventrals jugular, usually with 1 spine; less than 5 soft rays, often wanting; dor- 

 sal fin very long, its anterior portion composed of numerous spines; hypercoracoid perforate; 

 shoulder girdle, jaws, nostrils, and suborbital normal; tail diphycercal; pseudobranchiae present; 

 scales usually small and smooth, often wanting; vertebra numerous. 



OPHIDIOIDEI: Ventrals without spines; no spines in the anterior portion of dorsal fin. Other- 

 wise essentially as in the Blennioidei, the tail diphycercal, the last vertebras sometimes much 

 reduced. 



ANACANTHINI: Ventrals jugular, of soft rays only, the number usually more or Ie?s than 5; uo 

 spines in any of the fins; hypercoracoid imperforate; tail isocercal; no pseudobranchiae; gills, 

 nostrils, pharyngeals, suborbital, and shoulder girdle normal; vertebrae numerous. 



TJENIOSOMI: Body ribbon-shaped; the ventrals thoracic, the.rays usually less than I, 5; poet- 

 temporal undivided; skin smooth or prickly; caudal fin wanting or else divided and peculiar. 

 Skin naked or prickly; vertebrae very numerous. 



HETEROSOMATA : Cranium twisted so that both eyes in the adult are on the same side of the head; 

 dorsal and anal fins very long; no spines in the fins; ventrals thoracic, of more than 5 soft rays; 

 coracoids normally developed, the hypercoracoid perforate; tail diphycercal; pseudobranchiae 

 present; vertebras in increased number. 



Suborder SALMOPERC^E. * 



(THE TROUT PERCHES.) 



We place provisionally as a suborder of the Acanthopteri, a singular 

 group of archaic 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. 

 The group shows the remarkable combination of true fin spines, ctenoid 

 scales, and a percoid mouth, with the adipose fin, abdominal ventrals, and 

 naked head of the Isospondyli. The relations of the Percopsidce with 

 such archaic spiny-rayed fishes as Aphredoderus and Elassoma are cer- 

 tainly not remote and the close resemblance of the head of Percopsis to 

 that of Gymnocephalus (Acerina) may be more than accidental. The sub- 

 order may be provisionally defined as follows : 



Ventrals abdominal, each with a short simple ray ; dorsal with 2 sim- 

 ple rays or spines ; anal with 1 or 2 ; mouth formed as in Percoid fishes, 

 the simple toothless maxillary not forming part of its border. Adipose 

 fin present. Scales ctenoid ; head naked ; pseudobranchise present. Air 

 bladder apparently with a rudimentary duct. Stomach siphonal, with a 

 few coeca. Shoulder girdle without mesocoracoid, apparently of the 

 normal percoid type ; vertebrae about 35. A single family. (Salmo, trout ; 

 Perca, perch.) 



* In describing Percopsis, Agassi/ refers to it as a generalized type and relic of an older fauna. 

 He says: "Now, the genus Percopsis is as important to the understanding of modern types as 

 Lepidosteus and Ceslracion are to the understanding of the ancient ones, as it combines charac- 

 ters which in our day are never found together in the same family of fishes, but which, in 

 more recent geological ages, constitute a striking peculiarity of the whole class. My Percopsis 

 is really such an old-fashioned fish, as it shows peculiarities which occur simultaneously in the 

 fossil fishes of the Chalk epoch, which, however, soon diverge into distinct families in the Ter- 

 tiary period never to be combined again. Now my new genus Percopsis is a just intermediate 

 between Gtenoids and Cycloids; it is what an ichthyologist at present would scarcely think 

 possible, a true intermediate type between Percoids and Salmonidae." (Agassiz, Lake Superior, 



