DEVONIC FISHES OF THE NEW YORK FORMATIONS 2/ 



granules. Still more far-reaching in its consequences is Traquair's dis- 

 covery of Ateleaspis [text fig. 3], a form considered by him to indicate a 

 genetic connection between the Heterostraci and Osteostraci. Assuming 

 the correctness of the Scottish author's conclusions/ not only Pteraspis, 

 Palaeaspis, Cyathaspis, and the like, but also Cephalaspis, are traceable to 

 the primitive Coelolepidae, in which the dermal covering consists of small 

 shagreen granules, or of minute hollow spines. Only in the case of the 

 Antiarcha, the order to which Pterichthys and Bothriolepis belong, does 

 their immediate origin remain uncertain, owing to the absence of inter- 

 mediate forms linking them either with Pteraspids or Cephalaspids. Not- 

 withstanding their nearer approach to the latter, as indicated by the presence 

 of bone lacunae and dorsal fins, it is practically certain that they could have 

 had no direct derivation from that group ; and an inverse hypothesis, by which 

 the Cephalaspids are derived from the Asterolepids, appears vastly improb- 

 able. As to the theory, recently revived in some quarters, that Asterolepids 

 are transitional between vertebrates and invertebrates, and are more or less 



' A dissenting opinion has lately been expressed by Mr C. Tate Regan, in his article 

 on the " Phylogeny of the Teleostomi " [Ann. Mag. Nat. Hist. ser. 7, 1904. 13 : 329-49]. 

 Admitting, as he does, the resemblance between Ateleaspis and Cephalaspis, yet he cannot 

 see that there is "the least reason for regarding it [Ateleaspis] as allied to anything else," 

 and accordingly denies that it is annectant between the Heterostraci and Cephalaspidae. 

 The latter family is considered by Mr Regan to have been derived from the Asterolepidae, 

 being in fact specialized and degenerate members of that group, a view which lacks the 

 support of morphological evidence, and is at variance with the known facts of geological 

 succession. 



The extreme position taken by Mr Regan is apparent from the following extract : 

 " The Placodermi (Coccosteidae, Asterolepidae, Cephalaspidae) are a natural group, not 

 related to the Heterostraci, which are Chondropterygii. They may probably be regarded 

 as armored primitive Crossopterygii, this view being in accord with (a) the arrangement of 

 the cranial roof bones in Coccosteus ; (/') the structure of the ventral fin in Coccosteus ; 

 (<:) the structure of the pectoral limb of the Asterolepidae. . . The Teleostomi probably 

 originated from the Pleuropterygian Elasmobranchii in the Lower Silurian, and the Cross- 

 opterygii, with their specialized offshoots, the Dipneusti and Placodermi, must have rapidly 

 evolved, since all are well represented in the Lower Devonian, and the highly specialized 

 Cephalaspidae are found in the Upper Silurian." [p. 347] Some further discussion of the 

 problems here involved will be found in an article on " Asterolepid Appendages," published 

 jn the American Journal of Science for August 1904. 



