GREGORY: FISH SKULLS 



239 



posterior border of the orbit and the preopercular is relatively rnuch greater than it is in 

 the primitive berycid (Fig. 110), while in Luciolates (Fig. 117) th>s lengthening is further 



^""^Throtolith (sagitta) of Micropterus salmonoides,-accordmg to Frost {\927b, p. 301), 

 while frail and delicate, in general resembles that of Pwa. 



Mechanism and Evolution of the Protrusile Upper Jaw.-\n the percoids the protrusile 

 upper jaw varies from incipient to extreme stages. The phenomenon of protrus.hty of the 

 mouth has been partly described by Thilo (1920) and by Delsman (192.) from the mechan- 

 istic and the morphological viewpoints respectively. There is need, however, of a recon- 

 sideration of the subject from the phylogenetic viewpoint, as protrusd.ty has ar.sen inde 

 pendentlv in different groups of fishes. In the incipient stage (e.g., m Lates, tig. 114) the 



£phoh 



Lates Tiilotlcus 



Fig. 114. Lates niloticus. 



premaxill^ have relatively short ascending processes, which are more or less closely con- 

 joined in the mid-line. Beneath them (Fig. 115) is a median piece of cartilage, the rostral, 

 which slides on the upper surface of the vomer; the latter bears a low keel for the groove 

 on the under side of the rostral cartilage. The conjoined ascending processes and their 

 supporting cartilage form an inverted V, the top of which is Aanked by a ^^^h formed by 

 the nasals and upper part of the mesethmoid. Immediately ^^hmd and latera to the as- 

 cending process of each premaxilla is a shorter broader process called by Allis (1909 p 24) 

 the articular process of the premaxiUary. This fits posteriorly into the anterior fork of the 

 . So named by Allis (1909, p. 28) in the mail-cheeked fishes (Fig. 116). No: to be confused with the surface plate of the 

 same name in Palaeozoic ganoids. 



