320 ANNALS NEW YORK ACADEMY OF SCIENCES 



the Amphibia. With regard to the Arthrodira I can only say that I 

 have long pondered over the conflicting views and arguments of Smith 

 Woodward, Eastman, Dean, Hussakof, Tate Eegan, Jaekel and Patten; 

 that I have endeavored to consider the question without bias of any kind 

 and have returned again and again to readjust the conflicting claims of 

 diverse evidence for the relationship of the Arthrodira: either with the 

 Dipnoi (Eastman, Smith Woodward), or with the Antiarchi (McCoy, 

 Dean, Hussakof), or with the Devonian Crossopter3^gii (Tate Eegan) or 

 finally with the stem of the Dipnoi and Amphibia on the one side and 

 the Antiarchi, Ostracodermi and even the Arthropoda on the other 

 (Jaekel, Patten). Out of this maze of conflicting resemblances I gain 

 the impression that the resemblances which tend to brigade the Arthrodira 

 with the Antiarchi are quite specific and of evidential value,^ while the 

 resemblances in the jaws and head-shield to those of Dipnoi and still 

 more to those of Crossopterygii, are, with certain exceptions, of a vague 

 and inconclusive character, suggestive of convergent evolution between 

 widely removed groups. And in this connection I would call attention 

 to the arthrodiran appearance of the skull-top of the modern armored 

 catfish Clarias magur (figured by Goodrich, 1909, p. 381), as illustrating 

 the deceptive effects of convergent resemblances ; a similar example is the 

 Ceplialaspis-like appearance of Loricaria. 



Until it shall be learned whether M aero petaUcli thy s is really related 

 to the Arthrodira and what its affinities are with other groups, it is 

 hardly likely that the Arthrodira problem can be definitely settled. In 

 the meantime I must hold provisionally : first, that the arthrodiran jaws, 

 skull and thorax represent progressive, carnivorous adaptations^ from a 

 generalized antiarchian stem ; secondly, that the resemblances in the skull- 

 roof to those of Dipnoi and Crossopterygii have been independently as- 

 sumed; thirdly, that the Arthrodira are not true gnathostomes and con- 

 sequently not anywhere near the line leading toward the Tetrapoda, but 

 tliat they are, on the contrary, the last and highest of a pre-gnathostome 

 adaptive radiation. In brief this pregnathostome radiation was char- 

 acterized by an accelerated evolution of the exoskeleton of the head and 

 thorax and b}' a retarded evolution not only of the endoskeleton but of 

 the whole locomotive apparatus. 



ELASMOBRA^TCHII 



In the Acanthodii, which are the earliest elasmobranchs, the gnatho- 

 stome series is first made known by its relatively advanced adaptations 

 to predatory, free SAvimming habits : first, the whole pharyngeal region 



Cf. the list of common characters given by Hussakof. lOOG. pp. 123-136, 



