442 WILLIAM K. GREGORY 



that when intermediate forms between related groups are dis- 

 covered, these connections of form and of kinship should be 

 expressed by the assembling of the extreme forms and the middle 

 forms in one group, usually without any higher subdivisions than 

 families. Thus the Zeus-like fishes are thought to be related 

 to the Flatfishes through the Eocene Amphistiidae. Hence 

 Boulenger abandons the groups Zeoidea, Heterosomata, and by an 

 ingenious definition links the two in a new group called Zeorhombi. 



Whether related groups are now continuous or discontinuous 

 is partly an accident of time and of the degree of completeness 

 of our collections of fossil and recent forms. Surely such terms 

 as Nematognathi for the Catfishes, Squamipinnes for the Chaeto- 

 donts and their allies, and many other useful group-names stand 

 for perfectly clear types of structure, in forms clustered around 

 central types but grading into other groups at the peripheries. 

 The idea underlying the American method is that the best way 

 to map out the topography of this varied morphological expanse 

 is to assign a name to every conspicuous cluster of elevations, 

 •even if some lesser outlying elevations may connect with neigh- 

 boring systems. 



Thus the two classifications emphasize different sets of facts 

 about the same subject-matter, so that in a general way the Eng- 

 lish method emphasizes better both resemblances and phylogeneti- 

 gaps between different groups. Furthermore, as we have seen, the 

 results of the two methods are expressed in terms of a standard 

 the "order" which has a very different value in the two sys- 

 tems, in the English system covering the entire range of forms 

 from certain generalized Triassic physostomes (the Pholido- 

 phoridae) to the most advanced spiny-finned fishes and even to 

 such wonderfully metamorphosed beings as Mola and Malthe; 

 while in the American system the same term "order" implies 

 a much narrower range, as for example in the Haplomi. 



The American and English methods are fortunately not en- 

 tirely irreconcilable or contradictory, not like the two horns of a 

 dilemma between which only a bad choice is possible. Con- 

 ceivably the differences may be adjusted, and all the antitheses 

 and syntheses which the two systems seek to convey may be 

 harmoniously expressed. 



