NATURAL SELECTION 301 



In the majority of the small characters which distinguish 

 families or genera, it appears legitimate to distinguish quite 

 sharply between habit and structure and to inquire which 

 changes first in the course of evolution. If we set aside the 

 numerous structural features which seem to be functionless 

 and the numerous habits in correlation with which no co- 

 adapted structures have been developed, we are left with 

 many examples of small structural characters clearly asso- 

 ciated with small habit differences which are not neces- 

 sarily adaptive. It is with regard to this residuum that the 

 inquiry as to the priority of structure or function has to be 

 made. First let us suppose that the structural changes precede 

 the co-adapted change in function. Then, at the time when 

 they occur, all such changes will be non-adaptive and they 

 can become adaptive only after the necessary habit changes 

 have been made and in so far as the new habit is more advan- 

 tageous than the old. 



Secondly, if the habits change first, then any structural 

 change making the new habit more easy of fulfilment will be 

 adaptive, at any rate in so far as the new habit is adaptive. 

 It is evidently much easier to imagine evolution happening in 

 this way, especially if Natural Selection has played a big part 

 in it. But if we want to decide which of the two alternatives 

 has actually been most usual, it is very difficult to find much 

 evidence ; most authors appear to attribute the major im- 

 portance to habit. It is probable that any change in habits 

 may provide a use for some hitherto trivial structure, while the 

 existence of so many useless structures might be regarded as 

 an incentive to a change in habits allowing some of them to be 

 used. The condition is one where ' pre-adaptation ' might 

 be expected to be rife. 



We will now consider how far certain typical specific 

 differences in habits can be considered adaptive. Perhaps one 

 of the commonest types of habit difference is in the nature 

 of the food. We usually know very little about the variety of 

 foods eaten by carnivores, and especially of the relative import- 

 ance of the various items, and it will be simpler to consider a 

 vegetarian animal feeding on a few allied species of one plant 

 genus. There are quite a number of examples, for instance, 

 where in one insect genus some species feed on poplars, 

 others on willows. There appear to be two ways in which 



