XL] MY NEW IDEAS 393 



that with any species of migratory bird breeding can, as a 

 rule, be only safely accomplished in a certain area; and, further, 

 that during a large part of the rest of the year sufficient food 

 cannot be obtained in that area. It will follow that those 

 birds which do not leave the breeding area at the proper 

 season will suffer, and ultimately become extinct ; which will 

 also be the fate of those which do not leave the subsistence 

 area at the proper time. Now, if we suppose that the two 

 areas were (for some remote ancestor of the existing species) 

 coincident, but through geological and climated changes 

 gradually diverged from each other, we can easily understand 

 how the habit of incipient and partial migration at the proper 

 season would become hereditary (through the action of natural 

 selection), and so fixed as to appear to be what we term an 

 instinct. It will probably be found that every gradation still 

 exists in various parts of the world, from a complete coinci- 

 dence to a complete separation of the breeding and subsistence 

 areas ; and when the life-histories of a sufficient number of 

 species are thoroughly worked out, we shall find every link 

 between species which never leave a restricted area in which 

 they breed and live the whole year round, to those other 

 cases in which the two areas are very widely separated. The 

 actual causes that determine the exact time, year by year, at 

 which certain species migrate, will, of course, be difficult to 

 ascertain. I would suggest, however, that they will be found 

 to depend upon the climatal changes which most affect each 

 species. The change of colour, or the fall of certain leaves ; 

 the change to the pupa state of certain larvae ; prevalent 

 winds or rains, or even the decreased temperature of earth or 

 water, may all have their influence. Ample materials must 

 now exist, in the case of European birds, for an instructive 

 work on this subject The two areas should be carefully 

 determined for a number of species ; the times of their 

 movements should be compared with those of the natural 

 phenomena likely to influence them ; the past changes of 

 surface, of climate, and of vegetation should be taken account 

 of ; and there seems no reason to doubt that such a mode of 

 research would throw much light on the problem." 



