446 THE GRAMMAR OF SCIENCE 



lives to be once fertile, or we deal only with seasonal 

 fertility. Here, I think, we must look upon fertility not 

 as merely associated with purely reproductive characters, 

 but consider, at any rate, net^ fertility as closely allied 

 to health, fitness, and strength in the whole complex of 

 organs and characters which form the individual. But 

 it is just this health, fitness, and strength for a given 

 environment which determine ultimately the type char- 

 acter for a given form of life. Thus it seems possible to 

 understand, even if the point still wants quantitative 

 demonstration, how the modal character tends to become 

 associated with the character of maximum net fertility. 

 Where the environment by natural selection produces a 

 given type, with that type it ultimately associates the 

 maximum fertility. Differentiation of type connotes 

 a differentiation of fertility. When two modes arise in 

 a species, then arise two maxima of fertility. If we admit 

 that fertility is not random, but that it is correlated with 

 other characters, then our hypotheses for the origin of 

 species would be : — 



{a) Differentiation of type owing to change of environ- 

 ment creating two centres of fitness, i.e. natural selection. 



{U) Maintenance of this difference of type owing to 

 some barrier to inter-breeding, i.e. owing to some form of 

 sexual selection. 



{c) Check to prevent reproductive selection destroying 

 the differentiation, either (i.) a stringent periodic selection, 

 or (ii.) a differentiation of the most fertile character 

 associated with the differentiation of type. 



id') Progressive change of types until by the prin- 

 ciple of correlation members of the two type groups are, 

 owing to mechanical or physiological causes, mutually 

 sterile. 



These hypotheses at any rate give a scheme for 

 quantitative inquiry ; there is no stage which cannot be 



^ The distinction between net and gross fertility is very important for the 

 problem of evolution. The working classes have a greater individual gross, 

 but a less individual net fertility than the professional classes. Owing to a 

 greater marriage rate, however, the total net fertility of the former is greater 

 than the total net fertility of the latter classes. 



