NO. 3 TELEOSTEAN FISHES — GOSLINE II 



One of the most frequent variations and one that is apparently of 

 little systematic significance is the loss of postterminal centrum 2 ; 

 this occurs in the adult Synodus, in Lampanyctus, Omosudis, Lesti- 

 dium, and Alepisaurus. The loss of this structure in alepisauroids is 

 probably correlated with the rather sharp upturning of the last 

 vertebra as compared with myctophoids. There is also a repeated 

 fusion of hypurals, particularly the three lowermost. The three 

 epurals seem more constant; they are reduced to i or 2 only in the 

 synodontids among the material examined. The synodontids also are 

 unique in that the anterior uroneurals are broken up into a number 

 of platelets in the adult (Hollister, 1937b), 



Order Angiiilliformes. — The caudal skeletons of five eels belong- 

 ing to three families are illustrated by Whitehouse (1910, pi. 48, figs. 

 9-13). Though these are obviously specialized, it is tempting to iden- 

 tify the lower, double hypural as the fusion of the hypurals usually 

 numbered 2 and 3. If this is correct, then these eels have one complete 

 postterminal centrum, plus a second fused with the posteriormost 

 hypural plate. The elements labeled n. a. (equals neural arch) in 

 Whitehouse's plate would appear to be uroneurals. Under the ex- 

 planation offered here, the eel caudal skeleton is merely a specialized 

 offshoot of a basically primitive type. 



Miscellaneous orders. — The halosauroids and notacanthoids all have 

 a long, tapering tail with little or no possibility of a caudal skeleton. 

 The same is true of the macruroids. The caudal skeleton of the cods 

 has received a considerable amount of attention (e.g., Barrington, 

 1937). However, it seems to be far more aberrant than even that of 

 the eels, and its parts cannot easily be homologized with those of any 

 of the fishes dealt with here. 



In the Beloniformes, Syngnathi formes (c.f., Whitehouse, 1910, 

 pi. 48, fig. 15), and Gasterostei formes the caudal skeletons consist of 

 platelike hypurals that obviously represent secondary simplification, 

 leaving little trace of their derivation. They cannot profitably be dis- 

 cussed here. 



No caudal skeletons of the Phallostethi formes have been available. 



Order Percopsiformes. — The caudal skeleton of "Columbia" has 

 been figured by Regan (1911b, p. 295, fig. B) ; those of Percopsis and 

 Aphredodorus are illustrated here (figs. 2C and D). In a number of 

 features all are peculiar. The terminal vertebra and ist postterminal 

 centrum have fused, but in Percopsis (fig. 2C) there is a sharp up- 

 ward bend in this combined element, and in Aphredodorus (fig. 2D) 

 it has a peculiar diagonal line which seems to represent either an oddly 



