GREGORY: FISH SKULLS 



267 



flattening down of the top of the head. Again, as in the pikes and eels, with a forwardly 

 inclined suspensorium, the jaws have been lengthened by growing forward, especially at the 

 front end. This in turn has involved a lengthening of the lacrymal and ethmoid region. 

 In the top view (Fig. 142) the ascending rami of the premaxillae, which are conjoined by a 

 growing cartilaginous-dermal symphysis are blocked posteriorly and braced by the meseth- 

 moid, much as in the Mugilidae. Detailed comparisons of the premaxillae and maxillae 



A.Sphyr^na B. Atherinopsis 



C. Mugil 



Fig. 143. Skulls of Sphymna argentea (A); Atherinopsis californiensis (B); Mugil cephalus (C). After Starks. Top view. 



show that they are plainly of true percomorph type, modified by the presence of the enlarged 

 laniary teeth. The enlargement of the jaws has reduced the protrusility of the premaxillae 

 almost to zero since the extreme posterior position of the maxillo-mandibular ligament 

 makes it impossible to pull the premaxilla forward. Hence the ascending processes of the 

 premaxillae are much reduced and the proximal forks of the maxilla firmly clamp the pre- 

 maxillae. The dentigerous layer is growing forward at the tip, like a predentary or prenasal 

 bone. I cannot agree with Allis that it is this intermaxillary tissue which serves to connect 

 the ascending process of the premaxillae with the shafts of these bones. In three small 

 specimens of Sphyreena there is not a sign of separation between the ascending process and 

 the body of the bone. 



The series of excellent comparative views (Fig. 143) of the skulls of Sphyrana, Atherin- 

 opsis and Mugil, which are to be found in the paper by Starks (1899i) on the osteology of 



