BIRD MIGRATION. 1 7 



species which seem to be exceptions to my rule, really do migrate (that 

 is to say, the majority do) in advance of the young, but either by different 

 routes or without stopping at the same places. Indeed, no other explanation 

 seems possible, for it would be absurd to suppose that in the case of Oforor- 

 nts ag/'lis or any other species which lays but four or five eggs, there are in 

 autumn only five old birds to every hundred young. 



Hence the fact that the earlier flights of any given species are made up 

 largely or even almost entirely of young birds, furnishes no proof that the 

 young of that species migrate first, unless it can be further shown that the 

 later flights in the same locality consist chiefly of old birds. In short 

 the latter must be accounted for in some way, or they cannot be used as 

 a factor in the case. A flight chiefly or largely of old birds, succeeding one 

 similarly composed of young of the same species, has never come under my 

 observation, and I venture to assert that such a phenomenon never occurs 

 unless under purely fortuitous conditions. 



With Limicolas, as far as I have observed them, the earlier flights are 

 composed almost wholly of old birds, the later ones of young. During 

 many seasons' shooting on the salt marshes of the New England coast I 

 have invariably found this to be true of all the species which occur in suffi- 

 cient numbers to furnish conclusive evidence. On more than one occasion in 

 July shooting I have seen large flocks of Yellow-legs, Snipe (^Macro7-ham- 

 f/ms) and Ring-necks fyyEgialitis semifalmata) composed entirely of old 

 birds in worn breeding plumage. 



5. That migratory nocturnal flights often start simultaneously over 

 thousands of square miles of country, that they frequently converge and 

 unite, that the routes which they follow usually correspond more or less 

 closel}' with the trend of mountain chains, river valleys, the shores of lakes 

 or seas, or the wooded borders of extensive plains, and flnally that these 

 routes for considerable distances are not always exactly or even approxi- 

 mately in line with the general direction of the migration, are facts too firmly 

 established and widely known to require detailed consideration here. 



Thus far I have dealt only with diurnal-feeding land birds which migrate 

 by night. Turning now to land birds which migrate as well as feed by 

 day, we find that the majority of species fly in large, compact flocks, not 

 only when migrating, but at all other times except during the breeding 

 season. These flocks almost invariably contain more or less old birds who 

 seem to be recognized as leaders, and whose movements or signals are 

 closely followed and obeyed. That these old birds act as guides during 

 migration, and that the younger birds depend on them in this way, is un- 

 questionable. If the old birds are killed or otherwise lost, the young may 

 be thrown on their own resources for a while, but they rarely have to wait 

 long for an opportunity to join, or at least follow, another flock. Thus, as 



