THE ORDERS OF TELEOSTOMOUS FISHES 443 



The first step is to note the necessity for a larger number of 

 grades of taxonomic divisions between the subfamily and the class 

 than is found in the English system, which deals only with the 

 order (Teleostei), the suborder, the division, the family, and 

 the subfamily. As degrees of homological resemblances and of 

 phylogenetic affinities are infinite in number even a highly differ- 

 entiated system of classification must be more or less Procrustean 

 in nature, since in order to force all the different grades of assem- 

 blages into appropriate compartments some phylogenetic values 

 must be relatively compressed and others somewhat stretched. 

 But surely by using more grades of subdivision we may distort 

 the facts less than by using fewer grades ; although common sense 

 must soon impose a limit to the increase in the number of 

 grades, since each grade requires a corresponding set of terms 

 throughout the system. One advantage of a highly differentiated 

 system with many grades of divisions is that it permits us to 

 retain on different taxonomic levels many old useful and ex- 

 pressive names such as Malacopterygii, Isospondyli, which if 

 applied to divisions of the same taxonomic rank would compete 

 with each other as synonyms. 



The process of the differentiation of taxonomic grades has been 

 going on for a long time in ichthyology and elsewhere, and it has 

 usually been accompanied by the elevation in rank of certain 

 taxonomic grades and the lowering in rank of others. Thus the 

 rank of the grade called "species" by Linnaeus has really been 

 lowered, since many of his species are now called " genera " and 

 his genera " families " while in Gill's system many of Gunther's 

 " families " were elevated to the rank of the division called by 

 Gill "superfamily." From these inevitable shiftings many of 

 the differences between the American and English systems have 

 arisen. 



The desirability of a highly differentiated system has suggested 

 the use of the terms class, subclass, infraclass, cohort, superorder, 

 order, suborder, division, superfamily, family, subfamily in the 

 accompa.nying classification, which is offered as a tentative com- 

 promise between the American and English systems. 



The only ones of these terms requiring special comment are 

 the infraclass, superorder, cohort, and order. The infraclass is- 



