476 SCIENCE PROGRESS 



parts the basis of comparison, so long the sole criterion of 

 classification must be found in likeness of structure. Ques- 

 tions of function must be rigidly excluded. This conclusion 

 ris expressed clearly by Geoffroy in the words : " II est 

 evident que la seule generalite a appliquer dans l'espece est 

 donnee par la position, les relations et les dependances des 

 parties." 



Up to this point, whether from sufficient or insufficient 

 premises, the position had been logically developed. But not 

 unnaturally zoologists came in course of time to forget the 

 exact steps of the argument, and to extend its conclusion 

 beyond the proper limits. What had been established was 

 that, so long as classification proceeded by the tracing of 



! homologies, structural characters alone could serve as the basis 

 of classification. Beyond this the assumption was tacitly made 

 that the method of homology was the only possible one in 

 classification. Accordingly, classification became completely 

 identified with the comparison of structure, while any possible 

 use of functional characters was tabooed. 



It is difficult to see, if one disregards the results which 

 the two methods achieved, why the system of Geoffroy should 

 have been accepted rather than that of Cuvier. Geoffroy found 

 likeness of structure transcending difference of function, while 

 Cuvier treated likeness of function as though it were more funda- 

 mental than difference of structure. It is not inconceivable that 

 biology might have developed along the lines of a common 

 plan of function. Complexity of functional arrangement might 

 have been made the basis of classification, the structure which 

 served each function being regarded as of subsidiary import- 

 ance. But the doctrine of a common plan of organisation had 

 a very strong hold on zoologists, and it was essentially a 

 doctrine of structural likeness : as developed by the school of 

 " Naturphilosophie " it amounted almost to an ideal geometry 

 of animals. We have to reckon also with the overwhelming 

 evidence which soon accumulated, showing the fertility of the 

 method of structural homology. Under its influence classifi- 

 cation grew rapidly into a consistent science. Finally, the 

 method seemed to receive complete confirmation when the 

 embryonic development of animals came to be studied. For 

 it was then found that many structures which had previously 

 been recognised as homologous were actually developed from 



