GREGORY: FISH SKULLS 



289 



was derived from very small-mouthed nibbling forms at least closely related to the balistids, 

 requires then that, pari passu with the acquisition of a cutting beak in the ancestors of the 

 puffers, the jaws increased greatly in breadth and to a less extent in the length. But so 

 large a cutting and crushing mechanism could not advantageously be operated at the distal 

 end of so prolonged and narrow a rostrum as that in the trunk-fishes. Hence, according to 

 our hypothesis, there was a rapid period of transverse growth of the skull-top accompanied 

 by a marked secondary shortening of the rostrum and by the enlargement of all the elements 



Spheroides stictonotus 



Fig. 167. Stheroides stidonotus. 



that brace the jaws posteriorly. That the rostrum has been shortened is indicated, among 

 other reasons, by the marked difference in the length of the mesethmoid in different species 

 of puffers. In Lagocephalus (Fig. 168) it is relatively narrow and much resembles that of 

 the narrow-nosed balistids. In Spheroides (Fig. 167) it has become very wide and short, 

 due partly to the dorsal encroachment of the posterior processes of the premaxillae. That 

 the palatine has received a marked secondary increase in size is evident by comparison with 

 the minute palatine of the balistids, teuthids, chsetodonts and other nibbling types. 



In Lactophrys, of the family Ostraciontids, the entopterygoid and metapterygoid 

 share this enlargement and still retain much of the arrangement shown in Triacanthui 

 (Fig. 160) where they were placed near the lower end of the downwardly elongate face. 

 But in Spheroides (Fig. 167), according to our hypothesis, a marked relative shortening of 

 the face from the postorbital process to the quadrate-articular joint has shifted the entire 

 palato-metapterygoid tract backward beneath the orbit. However, even with this second- 

 ary backward shift, the upper part of the metapterygoid has not regained its primitive 

 percoid connection with the upper part of the hyomandibular, from which it is still sepa- 

 rated by a deep indenture. This hypothesis of a secondary shortening of the face and 

 its bony scaffolding, in the line leading to the puffers, does not in any true sense require an 



