THE EVOLUTION OF MIGRATIONS 109 



It will be seen, then, that the necessity to migrate 

 must have existed before the ice-age if birds had at 

 that time reached the northern sections of the north- 

 ern hemisphere. Towards the end of the Pliocene, 

 when modern birds were fully established and the 

 climate was much as it is now, a majority of species 

 must already have been fully migratory. The 

 southward march of the ice-fields may have crowded 

 most of them into the southern section of the North 

 American continent but the subsequent retreat of 

 the ice would have enabled the entire story to repeat 

 itself — the spreading north, the winter extermina- 

 tions and the final establishment of the migratory 

 swing. Whatever the climates at different periods, 

 day-lengths have undergone no change and the 

 species that now require a certain minimum amount 

 of daylight for feeding or of ultra-violet radiation 

 must have needed it then. Various species may 

 have continued to use the far North as summer 

 quarters even at the height of the various glaciations 

 and there is, indeed, indirect evidence of this in their 

 present migratory routes. But in either case the 

 ice-age has had no miraculous effect on birds and 

 there is not the least necessity to postulate that 

 birds have returned north after the various periods 

 of glaciation because they had for hundreds of 

 generations retained a memory of pre-glacial homes. 



