160 The Science of Life. 



though the change that resulted was not, of course, one 

 that was not in a sense implicit in the reproductive cells, 

 it was none the less an alteration of the natural bias. 

 Similarly, it is possible that very decisive functional and 

 environmental modifications may saturate deeply into 

 the organism, and affect the reproductive cells in such a 

 definite manner that a tendency to change in the same 

 direction may be transmitted to the offspring. But, as 

 we have said, it is not justifiable at present to admit 

 more than a possibility, and science does not deal with 

 possibilities. 



In his work entitled Natural Inheritance Gal ton was 

 Filial led by statistical methods to a very impor- 



Regression. tant generalization, which from one of its 

 aspects may be called the law of filial regression. 



A strange regularity is observable in the peculiarities 

 of large populations throughout a series of generations. 

 "The large do not always beget the large, nor the 

 small the small; but yet the observed proportion 

 between the large and the small, in each degree of size 

 and in every quality, hardly varies from one generation 

 to another.'* A specific average is sustained. And 

 this is not because each individual leaves his like behind 

 him, for this is obviously not the case. It is rather due 

 to the fact of a regular regression which brings the off- 

 spring of extraordinary parents in a definite ratio nearer 

 the average of the stock. 



"However paradoxical it may appear at first sight, 

 it is theoretically a necessary fact, and one that is 

 clearly confirmed by observation, that the stature of 

 the adult offspring must on the whole be more mediocre 

 than the stature of their parents that is to say, more 

 near to the median stature of the general population. 

 Each peculiarity of a man is shared by his kinsmen, 

 but on an average in a less degree. It is reduced to a 

 definite fraction of its amount, quite independently of 

 what its amount might be. The fraction differs in 

 different orders of kinship, becoming smaller as they 

 are more remote." 



As it is easy to misunderstand this important gener- 

 alization, let us give some further illustration. It does 



