23 



cently modified by a cross as to allow the analysis and recombi- 

 nation of characters derived from different ancestral lines. 



The complete return of the offspring of an extreme parent, to 

 the mean condition of the "pure line" to which it belongs, or 

 in technical language the entire want of " regression " in the 

 "pure line," is presented by Johannsen as a fundamental ex- 

 ception to the conclusions of Galton. 



Weldon and Pearson have criticized * the work of Johannsen 

 in considerable detail and although the tone of their criticism is 

 adverse throughout, they grant that his main contention may well 

 be true, that small seeds and large seeds of the same plant do 

 not give rise to plants bearing small seeds and large seeds re- 

 spectively. If read aright, their criticism must be held to be 

 confirmatory in so far as Johannsen's data are capable of bio- 

 metric analysis. Certainly their conclusion that his results are 

 closely identical with those found for other plants and for animals 

 when we compare mean parental and mean filial characters, 

 agrees precisely with that reached by Johannsen, for these means 

 represent the condition in the population or mixture of several 

 "pure lines," and not in the individual "pure line." 



The relation between this work of Johannsen and that of Gal- 

 ton on sweet-peas may now be considered. In the first place, 

 the actual results were the same when the treatment of the ma- 

 terial was the same, and in so far the work of Galton was con- 

 firmed ; but when the "pure lines" were followed separately 

 they were found to offer an apparent exception in the complete 

 return of the offspring of extreme parents to the mean condition 

 of the " pure line." Instead of this being fundamentally opposed 

 to Galton's results, however, it is the condition which should have 

 been derived a priori from Galton's " Law of natural inheritance." 



Regression is lucidly explained by Galton f as due to the fact 

 that the child inherits partly from his parents, partly from his 

 more remote ancestry, and that if " traced far backwards his an- 

 cestry will be found to consist of such vyied elements that they 

 are indistinguishable from a sample taken at haphazard from the 



* Inheritance in Ph.iseolus vulgaris. Biometrika, 2 : 499-503. N 1903. 

 t Natur.il inheritance, 105. 



