358 W. B. Scott — Variations and Mutations. 



variation. " Those who reject the particular inferences, posi- 

 tive or negative, here drawn from that study, must not in 

 haste reject the method, for that is right beyond all question." 

 (p. XI.) " It is submitted that the study of variation gives us 

 a chance and perhaps the only one, of arriving at this knowl- 

 edge" [i. e. of the facts of evolution] (p. 13). "In introduc- 

 ing the method of the study of variation, I have said that it 

 alone can supply a solid foundation for inquiry into the man- 

 ner by which one species arises from another. The facts of 

 variation must therefore be the test of phylogenetic possibility. 

 Looking at organs instead of species, we shall now see that the 

 facts of variation must also be the test of the way in which 

 organ arises from organ and that thus variation must be the 

 test of homology " (p. 30). 



Bateson's entire argument is founded upon the assumption 

 that individual variations form the material, out of which new 

 species are constructed, an assumption which has almost passed 

 into an axiom. " It is upon the fact of the existence of this 

 phenomenon of variation that all inductive theories of evolu- 

 tion have been based " (p. 3). He has entirely omitted any 

 consideration of the palaeontological facts and appears not to be 

 familiar with the literature of this subject. Had he been so, 

 he would hardly have said : " It is really a strange thing that 

 so much enterprise and research should have been given to the 

 task of reconstructing particular pedigrees — a work in which 

 at best the facts must be eked out largely with speculation — 

 while no one has ever seriously tried to determine the general 

 characters of such a series " (p. 14). Palaeontologists have 

 repeatedly endeavored to do this and it is much to be regretted 

 that workers in other lines of morphology should pay so little 

 attention to their results. Progress cannot be hoped for except 

 by combining all classes of facts. The object of this paper is 

 to compare Bateson's methods and results with those drawn 

 from a study of fossil mammals, with special reference, of 

 course, to those phyla which may be regarded as fairly well 

 established. The outcome of such a comparison is an 

 unpleasant alternative : either the study of variation offers us 

 little hope of learning the real facts of evolution and the 

 assumption upon which it is founded is radically faulty, or all 

 the palseontological phyla have been erroneously arranged and 

 must be thoroughly reconstructed, though no reconstruction at 

 present possible would bring about any greater harmony of 

 results. Such rearrangement is necessarily limited by the 

 geological succession. 



Bateson's definition of variation may be accepted as an exact 

 statement of the significance to be attached to the term. " To 

 this phenomenon, namely, the occurrence of differences 



