CATOSTEOMI 629 



Atlantic and Mediterranean and from distant points in the 

 Pacific ; specimens are occasionally captured on our coasts. It 

 feeds on other fish, but little is known of its habits and nothing 

 of its development. 



The affinities of the Zamprididae are very doubtful. Zampris 

 has usually been placed with the Acanthopterygians, a view which 

 is still upheld by Gill.^ 1 now agree with this high authority in 

 regarding the bone which I took for an infraclavicle as a much 

 developed coracoid, and the bone termed by me the coracoid as 

 a pterygial. But it has also been shown, by Starks, that such 

 a thing as an infraclaAdcle does not exist in the Stickleback, the 

 bone so-called being only a part of the coracoid ; and as in most 

 of the Sticklebacks the pelvic bones join the latter, the resemb- 

 lance between them and Zampris remains. As I have previously 

 pointed out, the absence of spines in the fins, and the position of 

 the ventral fins, together with the great number of rays in the 

 latter, which is only met with in the lower Teleosteans, are 

 characters which necessitate the removal of Zampris from the 

 Acanthopterygians, and I cannot find a better place for them 

 than near the Gastrosteidae. 



The whole question of the arrangement of the Physoclists 

 with abdominal ventrals (Catosteomi and Percesoces) is, I feel, 

 much in need of revision, and it may be found advisable to 

 break up this group into a greater number of sub-orders, in 

 which case the Selenichthyes would stand by themselves; the 

 Hemibranchii and Lophobranchii would be united under the former 

 name, as proposed by Woodward, or under that of Thoracostei 

 (Swinnerton) or Phthinobranchii (Hay). The position in the 

 system of the Pegasidae is still somewhat doubtful. This family 

 is regarded by some authors as related to the mail-cheeked 

 Acanthopterygians. 



Fam. 2. Gastrosteidae. — Body more or less elongate, naked 

 or protected by bony shields, tapering to a slender caudal 

 peduncle. Head moderate, with short or elongate and tubiform 

 snout; mouth small, terminal, toothed; opercular bones well 

 developed ; suborbitals in contact with praeoperculum, protecting 

 the cheek. Gills four, pectinated. Praecaudal vertebrae with 

 strong transverse processes and slender, free ribs. Spinous dorsal 

 represented by isolated spines. Pectoral fins with short ptery- 



1 Froc. U.S. Nat. Mus. xxvi. 1903, ]>. 915. 



