448 THE GRAMMAR OF SCIENCE 



the old environment individuals with a character of value 

 a' were not only more frequent than those of character a, 

 but more fertile, being nearer to the centre of fertility c. 

 In the differentiated group they are, owing to natural 

 selection, less frequent than individuals of character a, but 

 will they still be more fertile ? Under the old environ- 

 ment they are nearer to the centre of fertility c, and 

 would be more fertile. Similarly the individuals of 

 character b' are more fertile than those of character b, 

 and thus as reproduction takes place (supposing no cross- 

 ing between the differentiated groups) the individuals 

 represented by the portions Ka(j and B(5H of the 

 differentiated groups would continually tend to increase 

 more rapidly than the individuals represented by the 

 portions haY and V>b] ; in other words, we should have 

 a continual tendency for the new modes a and b to regress 

 toward c. We have attempted to get over this diffi- 

 culty by supposing a differentiation of the character c of 

 maximum fertility, or that ultimately a and b become 

 centres of maximum fertility. But how is this compatible 

 zvith the theory of heredity? If fertility be inherited, the 

 offspring of the part population represented by ArtG will 

 be on the whole more fertile than the offspring of the 

 part represented by A^^F. Now there is scarcely any 

 doubt that fertility is inherited (p. 462), hence what possi- 

 bility can there be of the differentiation of the character 

 of maximum fertility ? Are we not thrown back on the 

 need for a continuously active periodic selection cutting 

 down the most fertile in each generation ? 



The answer, I think, should be of the following kind. 

 We know that all sorts of characters and organs are 

 inherited, but all our investigations turn on pairs of 

 relatives living under the same environment. What would 

 happen if we compared pairs of relatives, when the 

 environment for one relative was different from that for 

 the other ? Health, strength, and fertility are functions 

 we may suppose of the fitness under a given environment 

 of a complex of organs in the individual organism. If 

 they are such then it may be shown that they will be 



