EVOLUTION 405 



In other words, we can form a correlation table of which 

 the pairs of organs shall be the statures of the same 

 individual boy at 1 1 and at 13, and we may calculate 

 the coefficient of correlation, and this will be a measure 

 of individuality in growth due to age. Thus growth with 

 age is again rendered precise by the knowledge of the 

 same three quantities : (a) change in type ; (d) change in 

 variability ; and (<r) correlation in the values of the same 

 character in the same individual at different ages. 



This quantitative method of describing growth may be 

 applied to any organ or character in any form of life, 

 which proceeds by stages to an " adult " condition, i.e. a 

 condition in which at any rate for a relatively long period 

 there is no sensible change in type or variability. 



If I have ventured to give the reader a probable law 

 for the growth of like characters at the same time, he 

 may well ask me for a law of the growth of the same 

 character at unlike times. Changes of type and of vari- 

 ability with age have been frequently determined, but 

 what is the law of growth- correlation for the same char- 

 acter ? Frankly I must confess that no statement on this 

 important point is yet possible,^ and that until it is possible 

 no quantitative measure of natural selection during growth 

 can be considered above suspicion. We must confine 

 ourselves to forms of life — unfortunately not very many — 

 in which adult stages are clearly marked, and deal only 

 with selection during these stages. 



% 9. — Selection. Discovery of the Fittest 



Let us suppose some form of life chosen of which the 



types, variabilities, and correlations have been determined 



for a certain number of organs, and suppose a certain 



group of individuals in the " adult " stage are left for a 



given time to the influence of given organic and inorganic 



surroundings. Now let us imagine an ideal state of affairs 



in which no migration and no reproduction take place, 



1 I may on certain theoretical grounds have a very shrewd suspicion of what 

 its value will turn out to be, but the paucity of the data I have at present 

 collected on the point do not warrant even its suggestion here. 



