128 BLASTOGENIC VARIATIONS. 



gression, Pearson came to the conclusion that there was 

 not at that time sufficient ground for forming any defi- 

 nite conclusion as to the manner in which lineal and col- 

 lateral heredity were related. Thus it did not appear 

 necessary to him that the coefficient of the former 

 should be half that of the latter, as Galton had sup- 

 posed. On attacking the problem a second time,* 

 however, Pearson succeeded in proving that they were 

 connected, according to a mathematically ascertainable 

 relationship, so that, starting from Galton's law of 

 heredity, it was possible to calculate the coefficients of 

 regression or correlation between an individual and any 

 of his kinsmen, either direct or collateral. Thus Pear- 

 son calculated the coefficient of regression between mid- 

 parent and son to be .6, or somewhat less than that 

 found by Galton. Between a single parent and a son, 

 it would, therefore, be .3. Between grandparent and 

 grandson it was .15, between great-grandparent and 

 great-grandson .075, and so on. Between brothers it 

 was .4, or considerably less than the coefficient found by 

 Galton. Nevertheless this value confirms Galton's 

 conclusion that brothers are more closely related to 

 each other by blood than are fathers and sons. 



It may be pointed out that in a stable population the 

 coefficients of regression and of correlation between an 

 individual and an ancestor are one and the same thing. 

 If, however, the population is not stable, so that the 

 variability of the offspring differs from the variability 

 of the parents, then these coefficients also differ slightly. 



The importance of this extension of Galton's law can- 

 not be rated too highly, for by its means the whole 



*Proc. Roy. Soc., Ixii. p. 386. 



