1888.] PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 587 



are imperfectly known, and it remains for future investigation to deter- 

 mine what are their exact relationships and characteristics. 



The Trigiids and Peristediids depart very widely from the other 

 groups, as will become hereafter mauifest, but, notwithstanding, their 

 relationships appear to be more intimate with the generalized mail- 

 cheeked fishes than with any other group. 



The Dactylopterids depart still more from all other fishes than do the 

 Trigloidea. We look in vain, however, for any nearer relation of those 

 fishes than the Trigloidea, and consequently it may be assumed that 

 they are the derivatives from a type from which the Trigiids have least 

 diverged. 



In fine, the relationships of the various families of mail-cheeked fishes, 

 in the present state of our knowledge, may be expressed in the follow- 

 ing genealogical tree iu which the left-hand branch in each case repre- 

 sents the more generalized type of each pair : 



Caracanthida? ? 



Cirri tidse. 



K 



o 



ft 



a 



eS 



to 





In this connection it seems advisable to refer to views enunciated by 

 Prof. W. Kitchen Parker. That eminent anatomist has proposed to 



