38 THE POPULAR SCIENCE MONTHLY 



genetic and the descriptive equivalent merely to the distinction between 

 past and present. It need hardly be said that genetic inquiries in sci- 

 ence are not necessarily purely historical or archeological inquiries, since 

 phenomena of genesis may be recurrent phenomena, taking place in ac- 

 cordance with the same laws in past or present. But, though he 

 blurred the idea somewhat, it remains true that, in his contrast be- 

 tween two types of scientific research, Kant exhibited his inclination to 

 what, in the vaguer sense, may properly be described as an evolutionary 

 habit of mind. It still remains, however, to determine just how far 

 this carried him, when he came to the consideration of definite prob- 

 lems. 



His problem of predilection, as I have said, was that of the nature 

 of a " race," the relations of different races, and the causes of their 

 diversity in physical characters. And this made necessary, at the very 

 outset, a consideration of the nature of a " species." Here, once more, 

 Kant follows Buffon : " Animals, however different they may be in form, 

 belong to the same physical species if, when mated with one another, 

 they produce fertile offspring." 



This Buffonian rule gives a definition of natural species as such {die Defi- 

 nition einer Naturgattung der Tiere uberhaupt) , in contrast with all artificial 

 species (Schulgattungen) . The artificial classification deals with classes, which 

 are grouped together upon the basis of similarity, the natural classification 

 deals with lines of descent, grouping animals according to blood-kinship. The 

 one provides an artificial scheme to aid the memory, the other a natural system 

 for the understanding. The purpose of the former is merely to bring animals 

 under labels, that of the latter is to bring them under laws. 



These references to Naturgattung en, determined by the criterion of 

 fertility of offspring, are themselves hardly in the language of trans- 

 formism. Yet one who employed such language might still regard 

 these " true species " as eventual results of divergent descent from com- 

 mon ancestors. But when we examine Kant's way of further defining 

 these species, we find that his notion of them expressly precludes the 

 possibility of any transformation of one into another through descent. 

 By the Buffonian test, he says : 



All human beings belong to one and the same natural species, since in 

 mating they always beget fertile offspring, however dissimilar the parents may 

 be in appearance. For this unity of natural species there can be but one nat- 

 ural cause, vis., that all men belong to a single stock (Stamm), from which they 

 have originated or at least could have originated. In the former case [i. e., of 

 actual descent from common ancestors], they belong not only to one and the 

 same species, but also to one family; in the latter case they would be similar 

 to one another but not related, and it would be necessary to assume a number 

 of separate local creations: an opinion which multiplies causes beyond 

 necessity. 26 



20 The same ideas are perhaps still more clearly expressed in the article " On 

 the Use of Teleological Principles in Philosophy," 1788: "There could be no 

 more certain test of diversity of stock (des urspriinglichen Stammes) than 

 the inability of two different hereditary branches of mankind to engender fertile 



