GREGORY: FISH SKULLS 



381 



As to the loss of spines in the fin and the loss of the homocercal caudal skeleton, these 

 characters have also been attained by some of the blennioids, while the presence of numerous 

 rays in the pelvic fins is not necessarily a sign of primitive descent, since soft rays in the 

 other fins have assuredly increased in number. 



The otoliths of some Macruridae, according to Frost (1926(f, p. 489) resemble those of 

 Umbra of the order Haplomi; but if the figures of these otoliths are correct the alleged 

 resemblance hardly carries definite evidence of relationship, especially in view of the im- 



sphot 



ep/ol 



(^u ^ pop 

 Melanonus gracilis 



Fig. 260. Melanonus fry. Sketch of skull from specimen prepared by Miss Gloria Hollister for Dr. William Beebe. 



mense differences between the skulls of macrurids and Umbra. Equally unconvincing is 

 the suggestion {op. cit., p. 483) of resemblance between some otoliths of the Macruridae 

 to the more specialized forms of the order Apodes, since the skulls of macrurids and all 

 Apodes are extremely unlike in basic features. 



For the present then, it seems that the evidence is insufficient to warrant setting aside 

 the many suggestions of relationships between the anacanths and the blennies and ophidiids. 



The typical anacanth skulls shown in Figs. 258, 259 call for but little comment. The 

 skulls are relatively long and low, the hyomandibular and preopercular being exceptionally 

 elongate anteroposteriorly, as in Ophiocephalus. The opercular is small and practically 

 excluded from the margin of the opercular flap by the enlarged subopercular. The oper- 

 cular has a more or less concave postero-ventral border lying between two sharp processes, 

 an arrangement that is seen also in some pediculates. Some of the branchiostegals are 

 also very long, as in pediculates. The lacrymal is extended nearly to the end of the snout; 

 it broadly overlaps the very weak maxilla. The bones are thin and tend to have mem- 

 branous borders. The'jaws are feeble and the ascending processes of the premaxillae are 

 weak or absent. The ethmoid of Gadus, however, still bears a median keel, as if for the sup- 

 port of protrusile premaxillae. The opposite frontals of Gadus are fused into a median plate, 

 which bears crests and creases probably caused by enlarged lateral line organs. The 

 parietals are enlarged, perhaps secondarily, but do not meet above the well developed 

 supraoccipital, which retains its percoid contact with the frontals. 



