THE BREEDING SEASON 23 



being through the agency of natural selection, in consequence of the 

 necessity to most birds of daylight for the procuring of food. This 

 hypothesis explains both the northerly migration in spring and the 

 southerly migration in autumn, since at both times the birds are 

 travelling in the direction of increased light (or, if they start before 

 the equinox, towards regions where they will enjoy longer daylight 

 later in the season). The suggestion that the time of the spring 

 migration is determined in each individual by a stimulus set up by 

 the growing genital organs is in no way opposed to Schafer's theory, 

 which provides an explanation of the general fact of migration. 



It has been noted that the northerly spring migration is far more 

 hurried than the somewhat leisurely autumn migration in the reverse 

 direction. Furthermore, although the north-south migratory move- 

 ments are as a rule extraordinarily regular, it has been observed that 

 the birds do not all set out together, and that the times of departure 

 and arrival for each species may vary in any one year by several 

 weeks. Moreover, golden plover are found migrating across Britain 

 on their way northward (perhaps to Iceland) at a time when other 

 individuals of the same species are rearing young in Britain. (The 

 breeding season in Iceland is about a month or six weeks later than 

 in Britain.) In view of these facts it is evident that the occurrence 

 of the migratory movement is dependent not merely upon external 

 or environmental influences, but also upon internal or individual 

 ones, and, as already stated, it is not improbable that one of the 

 factors involved is the state of development of the organs of 

 generation. 



Many birds are double-brooded, having young ones not only in 

 spring, but also in autumn before the close of the mild weather (in 

 temperate climates). Swifts are stated to have a second brood in 

 Southern Europe after leaving Britain in August, and the same is 

 said to be the case with nightingales. Wiltshire l mentions that a 

 pair of swifts that stayed behind the others, had a brood in September, 

 which migrated with the parent birds in October. Whether birds 

 are single- or double-brooded probably depends to a large extent 

 upon the duration of the period of incubation. This period in wading 

 birds and sea-birds is approximately double that of passerine birds ; 

 but, within the limits of the group to which they belong, it is closely 

 related to the size of the birds, the size of the egg, and the temperature 

 of the bird. 2 Thus the incubation-period of the stormy petrel is 

 thirty days ; that of the starling is fifteen or sixteen days ; while 

 that of the raven (the largest passerine bird) is about nineteen days. 

 The starling is, as a rule, almost certainly double-brooded, while the 



1 Wiltshire, loc. cit. 



2 Bergtold, A Study of the Incubation Period of Birds, Denver, 1917. 



