146 THE VARIATION OF ANIMALS IN NATURE 



occurs on alluvium inland, where it is associated with C. 

 campestris. The two moths Lasiocampa quercus and L. quercus 

 callunae, already noted as differing in emergence period, also 

 differ in habitat, the former being a lowland species, the latter 

 inhabiting moors and mountains. The two habitats in this 

 case are subject to considerable overlap. Fulton (1925) has 

 described two races of the common N. American cricket, 

 Oecanthus niveus. The races differ in song and habits of ovi- 

 position and also in habitat, one living on trees, the other on 

 bushes. Myers (1929, p. 50) records that the various New 

 Zealand species of Cicada are strictly confined to different 

 plant-associations. It is probably not usual, however, in 

 England for a species to be strictly confined to a plant-associa- 

 tion. Many species have a single food-plant, but few plants 

 are rigidly confined to a single association. Again, allied insect 

 species not rarely feed on the same plant, e.g. many Chryso- 

 melid beetles and weevils. It is not easy to find numerous 

 genera in which both taxonomic and ecological studies are so 

 advanced that we can say with certainty which species are 

 closely allied and what range of habitat is occupied. It is 

 certainly quite impossible to give a numerical estimate of the 

 frequency with which allied forms are found together or in 

 different habitats. We only know that both conditions may 

 be encountered. 



Amongst the vertebrates, closely allied forms tend to be 

 geographically isolated, so that this method of separation can 

 hardly arise. Amongst more distinct species, of course, 

 habitat-differences are common, but are probably not very 

 important in preventing interbreeding. 



I (d) . Differences in breeding habitats : minor geographical units. 



Differences of this type are best known in forms with a 

 definite breeding season. In migratory birds, for instance, 

 there is a well-known tendency for individuals to return to 

 breed in the locality where they were reared, and this tendency 

 makes possible the formation of geographical races, since races 

 which may mix in the winter, sort out and return to their own 

 areas in the spring. Though this phenomenon is largely part 

 of geographical differentiation, it must also lead to the forma- 

 tion of smaller units. Thus Schmidt (1931) has shown that 

 species of eel which breed in a single restricted area are 



