164 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.39. 



These contrasted characteristics are merely the most saUent or 

 obvious of the many that confront the observer;" they are assuredly 

 enough to certify to the generic distinction between AnarrTiichas and 

 Lyciclithys. In fact, in most respects the cranial and scapular differ- 

 ences between the genera in question are not only more obvious but 

 of greater taxonomic importance than those between AnarrJiicJias and 

 Anarrhichthys. Further, the distinction between the last two genera, 

 based on the relative length and size of the tail has been exaggerated. 

 The caudal fin of Anarrhichthys is of the same type as that of 

 Anarrhichas and merely reduced in size. 



The relations and most important, or rather most easily ascertained 

 diagnostic peculiarities of the three genera of the family are now 

 indicated. 



ANALYTICAL CHARACTERS. 



Key to genera of Anarrhichadids . 



a^. Body moderately elongated and robust backward and vertebrse moderately numer- 

 ous (74-81, e. g. 26+50 pm.); dorsal and anal with a moderate number of rays; 



caudal distinct Anarrhichadin^. 



6'. Teeth mostly subacute and not crowded; vomerine patch rather short (not 

 extending as far back as palatine patches) ; cranium flattened behind inter- 

 orbital area; cranial axis almost rectilinear; branchiostegal rays 7 . .Lyciclithys. 

 h^. Teeth chiefly blunt, or molarand crowded ; vomerine patch elongated (extending 

 backward beyond palatine patches); cranium pinched behind interorbital 

 area; cranial axis highly angulated by the extension downward of the vomer; 



branchiostegal rays 8 Anarrhichas. 



a?. Body greatly elongated and attenuated backward and vertebrae extremely numer- 

 ous (e. g. 350=38+212 pm.) ; dorsal and anal with corresponding number of rays 



(D. 250; A. 230 pm.); caudal connected with dorsal and anal 



Anarrhichthyin^. 



Teeth blunt, or molar and crowded, vomerine patch elongated; cranial axis 

 moderately angulated Anarrhichthys. 



The primary groups of the family — the subfamilies — are sufficiently 

 defined here, and clearness of conception of their differences would be 

 marred rather than enlarged by the few coordinated characters, so 

 relatively unimportant are they. The genera, however, are so distinct 

 that their characteristics are numerous, as will appear by a com- 

 parison of those now to be given. 



The genera are considered in the order of their assumed develop- 

 ment, the most generalized first, the most specialized last. 



o Among other notable differences between the genera are those manifested between 

 the dorsal rays signalized by Steenstrup and Collett. 



