260 G. F. Eaton —Collection of Triassic Fishes at Yale. 



based on inadequate material. It is to be regretted that Dr. 

 Woodward had not an opportunity for a thorough revision of 

 the American species of this genus. His decision to unite 

 Tschypti rus with Semionotus is satisfactorily borne out by the 

 investigations of Dr. E. Schellwien, recorded in his valuable 

 monograph.* 



At the outset the question arises — what characters should be 

 chosen for the basis of classification ? Prof. Newbenw placed 

 much confidence in proportional length and depth. Dr. 

 Woodward, relying on these proportions and also upon scale 

 conditions, recognized, as stated above, only two American 

 species of Semionotus. Although he has thus taught a whole- 

 some and well-deserved lesson, he has probably been too severe 

 in his judgment upon the work of American authors. Even 

 granting that the cranial and facial bones originally possessed 

 specific differences, it would be a profitless task to base the 

 classification of the various species upon such characters, obscured 

 as they now are throughout nearly the whole series of speci- 

 mens. To the present writer the only safe course out of the 

 difficulty seems to be to arrange this group of fishes mainly 

 according to the character of the fins and scales, and when 

 these are indistinct not to attempt the specific identification. 

 Strict adherence to this method has resulted in the revision of 

 species offered in this paper. 



The search for specific characters has, however, led to some- 

 thing of greater interest. A specimen of S. Marshi from 

 Sunderland, Mass., has been prepared in which the head bones 

 are so clearly defined that an accurate drawing can be made of 

 them (Plate VI, fig. 1). Although the bones of the maxillary 

 arch and of the circumorbital series cannot be recognized, and 

 the post-temporals are obscure, in other respects the preserva- 

 tion is remarkabl}- good, and the forms of the bones which 

 have been drawn may be considered fairly characteristic of the 

 genus. The premaxillaries, which are lacking in this example, 

 are shown distinctly in an otherwise poor specimen of 

 Semionotus from Boonton, N. J. (Plate YI, fig. 4), whose 

 specific identification is uncertain. In the figure of this latter 

 specimen the left premaxillary is seen in advance of the right, 

 and the paradoxical overlapping of the ascending process of 

 the right premaxillary by that of the left probably resulted 

 from maceration. In order to expose the entire premaxillaries 

 it was necessary to remove part of the right frontal. Appar- 

 ently, during life, the frontals overlay the upper third of the 

 ascending processes of the premaxillaries. 



With the drawing of the head bones it may be well to record 



* " XJber Semionotus Ag.," Konigsberg in Pr., 1901. 



