582 CLASSIFICATION OF THE MAIL-CHELKED FISHES. 



PRINCIPLES OF CLASSIFICATION. 



The question now comes up, which of the views entertained respect- 

 ing the classification of the mail-cheeked fishes and promulgated are 

 the more correct ? Three of the systems adverted to may be specially 

 considered : (1) the Cuvierian, in which all the mail-cheeked fishes were 

 associated closely together; (2) the Guntherian, wherein those fishes 

 are segregated according to the relative proportions of the spinous 

 and soft parts of the dorsal fin; or (3) the Copean, in which the pri- , 

 niary distinction is based upon the development or nou development of 

 the so-called muscular tube, or, in other words, whether there is a 

 double or single " basis cranii," and whether there are two or four epi- 

 pharyngeal bones on each side. 



Considering the various forms with reference to the development of 

 the dorsal tin, we find that at least some forms (Caracanthi) that have 

 been referred to the family Scorpcenidce by Giinther actually have the 

 soft portion of the dorsal longer than the spinous portion, and not, as 

 the definition requires, the proportions reversed. We also find that it 

 is difficult in practice to fit the definitions to certain fishes, for there is 

 really a gradation, if we take into account all those which have been 

 thus distributed into the four Guntherian families, between those forms 

 with an elongated spinous portion and those with au abbreviated spinous 

 portion of the dorsal fin, as well as those having a short or loug anal fin. 

 It is fouud also that the groups of Giinther traverse those proposed by 

 Cope, and that the characters derived from the structure of the dorsal 

 fin are not coordinated with those of the skeleton as signalized by 

 Cope or by nature. Thus it appears that the Triglidce, which, by Giin- 

 ther, are associated with Cottidce, are by Cope separated from the latter 

 and approximated next to the Scorpcenidce, while, on the other hand, 

 the genus Hemitripterus, which, by Giinther, is referred to the family 

 Scorpcenidce, is by Cope considered to be one of the constituents of the 

 family Cottidce. 



Long ago the present author had considered the questious thus in- 

 volved and had been led to the conclusion that the various mail-cheeked 

 fishes had been, on the whole, naturally associated by Cuvier, although 

 of course, in accordance with modern views, the species constituting the 

 family of Cuvier required to be segregated into at least a number of fami- 

 lies. He had found that the development of the dorsal fin was of much 

 less value than had been claimed for it by Giinther, and that the defini- 

 tions of Cope referring to the double or single basis cranii were inappli- 

 cable in the classification of these fishes. 



All of the Cottidce naturally have a double "basis cranii" although 

 less developed than in the Scorpcenid.ce, nor would the term " rudimen 

 tary," even, be applicable to the condition exemplified in the Cottidce 

 whatever may be the sense in which that word has been used by Profes 

 sorCope with reference to the genus Pelor, which is said to have the 



