GREGORY: FISH SKULLS 



171 



(Fig. 62), in which the middle part of the skull is already well elongated. The stout 

 hyopalatine arches articulate closely with the lower lateral borders of the neurocranium, 

 concealing the parasphenoid in the side view of the skull. The symplectic has been lost 

 and the broad preopercular articulates by a long irregular suture with the wide hyoman- 

 dibular. The stout opposite premaxillae are fused at the proximal end. They are placed 

 horizontally and bear small teeth. The stout curved maxilla is toothless and is displaced 

 to the outer angle of the gape. The dentary is rather short, thick, and must be capable of 

 giving a strong bite, since it is moved by powerful adductor mandibulse muscles, which 

 have an extensive area of origin on the elongated tract behind the mandible. Teeth are 

 wanting from the vomer, palatine and pterygoid bones. The subopercular is hidden 

 beneath the opercular and the interopercular is produced into a long narrow rod. The 

 large lateral cranial foramen (for the reception of a vesicle that penetrates the cranial 

 cavity) is covered laterally by a thin scale bone ("supratemporal," extrascapular). 



Gnaihonemus curvirostris 



mH 



dn 



Fig. 63. Gnaihonemus curvirostris. 



In Gnaihonemus curvirostris (Fig. 63) the typical mormyrid specializations are much 

 further emphasized. The skull could be derived from that of Mormyrops by the inflation 

 of the braincase and by the pulling out of the ethmoid, pterygoid and articular bones into 

 a long decurved trunk. The terminal mouth has become minute but is doubtless capable 

 of acting like powerful small pincers. 



In Petrocephalus (Fig. 64) it would seem that mormyrid specialization had reached its 

 present limit. As figured by Ridewood, the skull appears to me to have been derived 

 from a long-skulled mormyrid type by a relatively rapid shortening of an already down- 



