NO. 3 TELEOSTEAN FISHES — GOSLINE 33 



Macruri formes, Gasterostei formes, Lampridi formes, most Cyprino- 

 donti formes, Phallostethi formes, Stephanoberyci formes, Beryci- 

 formes, Mugiliformes, and most higher teleostean orders. 



Of these, the very pecuHar mouth structure of the isospondylous 

 Phractolaemus has been taken up briefly by Ridewood (1905, p. 279). 

 It seems to be constructed on a very different plan from that of the 

 percoids. In Bathyclupea the protrusile jaw is typically percoid, as 

 indeed is the fish itself. The protrusile, somewhat tubular mouth of 

 ateleopids seems never to have been adequately described ; no speci- 

 mens are available to this author for dismemberment of the mouth 

 parts. The protrusile upper jaw of cyprinoid fishes has been dealt 

 with a number of times (cf. Fiebiger, 193 1 ; Gregory, 1933); its 

 construction is very different from that of the percoid fishes and 

 certainly represents an independent development. 



In most of the rest of the fishes with a protrusile premaxillary there 

 is a basic structure essentially similar to that of the percoids. So far 

 as the structural elements mentioned in the description of the percoid 

 jaw are concerned, there would seem to be no basic difference between 

 the protrusile upper jaws of the cods, sticklebacks, holocentrids, and 

 phallostethids and those of the percoids. Whether, however, the 

 peculiar type of jaw protrusion found in the Lampridi formes (Regan, 

 1907) is of basically percoid type or has been derived independently 

 is a question about which the author has no first-hand information. 



The cyprinodonts, some of which have a protrusile and some a fixed 

 upper jaw, form a rather special category. Eaton (1935, pp. 166-167) 

 has stressed the similarities in jaw structure between Fundulus and 

 the Percesoces. There are certainly superficial resemblances between 

 Mugil and Fundulus, but I believe these to be secondary. The peculi- 

 arities of the mugilid jaw structure can be traced, via the atherinids 

 and sphyraenids, directly back to the percoid type. The protrusile 

 upper jaw of cyprinodonts operates on a rather different system. In 

 Fundulus, Belonesox, Rivulus, Goodca, Orestias, and others the pre- 

 maxillary ends laterally in a strong downward hook which is mem- 

 branously attached to the coronoid portion of the lower jaw forward, 

 more or less independently of the maxillary. Lowering of the mandible 

 thus forces the premaxillary forward. The maxillary apparently has 

 very little to do with premaxillary protrusion. 



Among cyprinodont genera with a nonprotrusile premaxillary, 

 Oryzias has the same downward process at the tip of the premaxillary 

 as Fundtdus and the others mentioned. Loss of protractile premaxil- 

 laries is here undoubtedly a secondary character. In Chologaster and 



