NATURAL SELECTION 315 



instances structures apparently useless may in the future be 

 found to play an important part in the life of the species ; 

 further, many ' useless ' characters may be correlated with less 

 obvious features which are of real use, but, even allowing for 

 this, the number of apparently useless specific characters is so 

 large that any theory which merely assumes that they are 

 indirectly adaptive is bound to be more a matter of predi- 

 lection than of scientific reasoning. 



A survey of secondary sexual characters (in which specific 

 differences are often displayed) shows that in any one group 

 they tend to occur very sporadically. They are often present 

 in one species and absent in another which is otherwise very 

 similar both in habits and structure. The explanation of the 

 evolution of such structure by some modified form of Darwin's 

 sexual selection theory still requires much more direct verifica- 

 tion. We hardly feel as yet that we have enough evidence to 

 estimate the value of the theory. The special case of specific 

 differences in the male or female genitalia is considered at some 

 length, and we conclude that there is very little evidence that 

 these structures play an important part in isolating species. The 

 evolution of such structures, where there must be some degree of 

 co-adaptation between the sexes, is very difficult to understand, 

 particularly if it is assumed to have resulted from the establish- 

 ment of a number of small variants, each one of which was 

 separately adaptive. 



Most of the so-called ' useful ' characters are regarded 

 as adaptive because they fulfil some role in the normal life- 

 cycle of the animal rather than because they have been proved 

 to have survival value. This tacitly assumes that any differ- 

 ence in habits must be adaptive. An analysis of a number of 

 particular examples shows that the problem of habit-differences 

 between species is by no means so simple. Quite a number of 

 differences in habit appear to be just as useless as the bulk of 

 structural specific characters. Where habit-differences appear 

 superficially to be more definitely adaptive, as in differences 

 in food- or habitat-range, each example still needs to be 

 studied on its merits. Increase of range may be beneficial to the 

 whole complex which forms the species, but is not often of such 

 obvious advantage to the individuals breaking new ground. 

 Specialisation in a more restricted range might be at least 

 temporarily advantageous for the pioneering individuals, but 



