468 THE GRAMMAR OF SCIENCE 



of genetic selection,^ — a survival of the most fertile, when 

 the most fertile are not the socially fittest. 



I have diverged somewhat from my theme in this 

 paragraph, because I wished to illustrate a point insisted 

 on in my first chapter (p. 25), i.e. the direct bearing 

 science has on moral conduct and on statecraft. The 

 theory of evolution is not merely a passive intellectual 

 view of nature ; it applies to man in his communities, as 

 it applies to all forms of life. It teaches us the art of 

 living, of building up stable and dominant nations, and it 

 is as important for statesmen and philanthropists in 

 council as for the scientist in his laboratory or the 

 naturalist in the field. 



\ I o.^On Bi-parental Inheritance 



Hitherto we have considered the correlation between a 

 pair of relatives only ; we have ascertained what change 

 will be made in the offspring if we selected one parent 

 only. We have now to consider how the offspring will be 

 influenced if we take into account both parents at the 

 same time. We have then three organs to measure, 

 which may or may not be the same for three individuals, 

 according as we are dealing with direct or cross heredity. 

 Let in^ be the mean value or type of the organ in fathers, 

 fi^ the deviation from the type in a particular father, a^ 

 the standard -deviation or variability (p. 387) of this 

 organ in fathers in general. Let w., and cr., give the type 

 and variability of the organ for mothers, and /^., the 

 deviation from type of the mother mated to the particular 

 father. Let the mean and variability of the offspring of 

 one sex, say sons, be given by w.^ and o-.^, and the devia- 

 tion from type of a particular son of the union by x.^. 

 We know that the sons of parents of given organs will 

 form an array, and our object is to find the type and 

 variability of this array ; these wiil fix the regression and 

 the correlation which determines the intensity of the inherit- 

 ance. Let the type of the array of sons due to parents 

 whose deviations are //^ and h.^ be h.,, and the variability of 



