VARIATION AND HEREDITY 33 



original mixed race. This is one of the methods of the 

 horse-breeder and the horticulturist, and their success 

 is a standing proof of the inheritance of the variations 

 found in a race, originally made more or less homo- 

 geneous by random breeding, but derived from a mixed 

 ancestry and containing a large number of germinal 

 differences. 



To a certain extent selective mating in mankind 

 does occur even without conscious choice. Families 

 with similar occupations, ideals, and characters tend to 

 associate, and this tendency increases the opportunities 

 for individuals with similar qualities to realize their 

 innate preference for each other. Instances will be 

 given later to show how this process results in the 

 segregation of distinct types of ability in various 

 sections of the community. Were this tendency to 

 selective mating not in action, the coefficients of cor- 

 relation between one parent and the offspring would be 

 less than they are, though it is difficult to estimate 

 what would be the decrease in their value. 



Another method of studying the correlation of 

 characters between parents and their descendants was 

 followed by Galton. He measured the average frac- 

 tional contribution of each ancestor to the total heritage 

 of the offspring in a mixed race. From an examination 

 of the records relating to the basset hounds of the 

 late Sir John Millais, he found that the contribution of 

 the parents was about one-half, that of the grand- 

 parents one-quarter, that of the great-grandparents 

 one-eighth, and so on. 



Now, if this law of ancestral inheritance be under- 

 stood to mean that the character of each individual is a 



D 



