HISTORICAL ACCOUNT OF EVOLUTION THEORY 39 



resulting from his computations of parental generations with those of 

 offspring, he arrived at two laws of heredity: the law of filial regres- 

 sion, and that of ancestral shares of inheritance. The essence of the 

 first was that the offspring of exceptional parents tend to regress 

 toward mediocrity in proportion to the degree of parental excep- 

 tionalness. The second law was really explanatory of the first, for it 

 was found that the offspring inherit not only from parents, but from 

 the various grades of ancestors, and it was the pulldown of a miscel- 

 laneous ancestry that made for regression toward mediocrity. It 

 appeared that half of the hereditary influence could be assigned to 

 parents, half of the remainder to grandparents, half of the remaining 

 remainder to great-grandparents, and so on down the line. 



Karl Pearson, a pupil and follower of Galton, has carried the study 

 of biometry to a more highly refined state. His attempt has been to 

 apply to the study of evolution the precise quantitative methods which 

 are used in physics and in chemistry. While much of Pearson's work is 

 far beyond the range of the average professional biologist today, it 

 is extremely useful as a tool in handling data in which great accuracy 

 is demanded. Frequently, however, the methods are far too refined 

 for the material, and much time is wasted in handling crude data 

 by means of highly refined instruments of measurement and ultra- 

 accurate mathematical methods. 



On the whole the contributions of biometry to our understanding 

 of the causes of evolution are rather disappointing. About the only 

 clean-cut finding has been the discovery that some variations are 

 continuous and others discontinuous. The former are capable of being 

 expressed in a single curve with a single mode, while the latter are 

 expressed in bimodal or polymodal curves. If material is homo- 

 geneous to start with it is likely to give monomodal curves, but if it is 

 heterogeneous, its heterogeneity will be revealed by the plural modes. 

 In a subsequent connection (chapter xliii) some further account of the 

 details of biometry will be presented. We must for the present be 

 content with having placed biometry in its setting as one step in the 

 advance of the evolution idea. 



EXPERIMENTAL BREEDING 



"While De Vries," says Castle, 1 "was engaged in his studies of the 

 evening primrose he hit upon an idea far more important, as most 

 biologists now believe, than the idea of mutation, though De Vries 



1 W. E. Castle, Genetics atui Eugenics (Harvard University Press, 1920), p. 82 



