Classification of Tel eos lean Fishes. 299 



pterygials, two or three are in contact with the coracoid and 

 one or two with the scapuLa, this being the reverse of what 

 obtains in the Berycidte and most Acanthopterygians. 

 3. Absence of pseudobranchiaj. 



If we now compare the Pleuronectidaj to the Gadoids, we 

 find that in these three cliaracters they differ from them and 

 agree with the majority of the Acanthopterygians, especially 

 with those which, geologically and morphologically speaking, 

 may be termed the oldest. The tail, whenever a caudal fin 

 is well developed, belongs to the homocercal type (hetero- 

 cercal in the embryo), with comparatively few rays (20 or 

 less) . The pectoral fin, in its fullest development, is supported 

 by four pterygials, of which three are attached to the scapula 

 and one to the coracoid, and the fenestra is in the scapular 

 bone. Pseudobranchige are present. 



From a consideration of these characters alone, the Pleuro- 

 nectidas canriot be held to have been derived from the Gadoids, 

 but their ancestors must be sought for among more primitive 

 Acanthopterygians. Bearing in mind Holt's suggestion 

 quoted above, 1 have proceeded to make a search among the 

 deep-bodied, strongly compressed types, such as the so-called 

 Squamipinnes and some of the Scombiiformes. I may 

 mention that in seeking for extinct allies of the Zeidaj I had 

 already arrived at the conclusion that a form placed among 

 the Carangidte by Woodward *, Amphistium, agreed very 

 closely with them in the structure of the vertebral column, 

 notwithstanding the lower number of vertebrse (L0-i-l4) ; its 

 caudal fin, quite similar to that of the Zeida?, precluded its 

 reference to the Carangidye, and a careful examination of one 

 of the specimens preserved in the British Museum, kindly 

 placed at my disposal for stud}^ by Dr. A. S. Woodward, 

 convinced me that as manj^ as eight branched rays in addition 

 to a spine are borne by the pelvic bone — this, of course, 

 affording a confirmation of the supposed affinity with the 

 Zeidaj. The latter differ from all other Acanthopterygians, 

 except the Berycidre, J\lacrurida3, and various Gadidaj, in 

 having more than five (6 to 8) articulated rays to the ventral 

 fin — an important character, indicative of descent from a 

 lowly type, for I am not disposed to admit that the number 

 of rays on one bony support having once been reduced can 

 again become multiplied. We know, in the evolution of 

 the dorsal and anal fins, that the exoskeletal rays, having first 

 been in excess of their endoskeletal supports (interspinous 

 bones), becaaie reduced so as to correspond with them in 



* Cat. Foss. Fish. iv. p. 434 (1001). 



22* 



