304 THE VARIATION OF ANIMALS IN NATURE 



conditioned to the new food. It appears that in most of the 

 experiments the transference has not been made at a suffi- 

 ciently early stage. When a certain food has been tasted it 

 may well be understood that a transfer will be more difficult. 

 It may be argued that the few survivors on the new substratum 

 form a specially adapted strain which has been selected out, 

 but this appears improbable when we find that a race once 

 accustomed to a new substratum may be almost as difficult 

 to retransfer to its original food as it was, in the first experiment, 

 to rear on the new. The whole question, however, is in need 

 of more numerous experiments on a larger scale. 



If we turn to other typical instances of habit difference, 

 we usually find our knowledge equally small and the difficulties 

 of a straightforward adaptational explanation just as great. 

 A large number of the minor specific differences in habits 

 appear, as far as we can see, to be non-adaptive. For instance, 

 in various leaf-mining insects the mine may be made on either 

 the under or the upper surface of the leaf, or it may begin at 

 the base, centre or margin of the leaf ; it may be of various 

 shapes (a loosely or tightly coiled spiral, blotch, etc.) ; and the 

 pupa (or puparium) may remain in the mine or the larvae 

 may pupate in the ground ; the frass of the larva in the mine 

 may be arranged in one or more continuous rows of pellets, 

 in discontinuous heaps, or in a single mass, or may be ejected 

 from the mine altogether. None of these habits has any 

 known adaptive significance. 



In another large class of examples the habits appear to 

 be adaptive in a general way without being specially adapted 

 to the particular case under consideration. We may instance 

 here numerous specific differences in nesting habits. Generally 

 speaking, each method of nesting appears to be reasonably 

 adapted to the needs of the animal, but we can rarely, if ever, 

 indicate how one method is more adapted to the need of the 

 particular species which employs it. It will perhaps be 

 retorted that it is too much to expect that we should be able 

 to demonstrate such adaptation ; but until we can (at least in 

 a fair proportion of cases) it is not very logical to assume that 

 all such habit differences must have some important reference 

 to the survival of the animal. 



Another difficult problem is raised by the consideration 

 of how far the habitat differences between species are likely to be 



