302 U. S. NATIONAL MUSEUM BULLETIN 217 



most of the species have met at any time on their wintering gromids. 

 We have no measure of the casualties which result from this exposure 

 to cold, but by their long existence it is evident that the birds regularly 

 endure these wide variations in temperature. The tolerance for cold 

 exhibited by such small birds as sandpipers is remarkable but it 

 must nevertheless be regarded as a normal ability, for it is exercised 

 every year. 



Nesting 



The social behavior and family relations of birds are elaborately 

 developed and their importance for the existence of races is more con- 

 spicuous than is generally the case among other animals, which can 

 detect others of their kind through scent and in darkness. Such ani- 

 mals may thus remain strange to us because of our lack of suitable 

 perception, whereas among birds social and family coherence is main- 

 tained by visible and audible actions. These are as apparent to us as 

 they are to the birds for which these signals are produced, and in many 

 cases their significance is in a general way understandable to us. 

 Without claiming full comprehension of the demonstrations by which 

 birds maintain coherent flocks and families we can often recognize the 

 general purport of their activity at a given stage near breeding. 



Duration of Migratory Behavior 



The latest date when the birds of a species nesting in the area have 

 arrived can be determined only by circumstantial evidence, but because 

 of their expressive behavior it is generally easy to distinguish in even 

 a single bird whether it is migrating or has settled down to nest. Mi- 

 gratory flocks disintegrate before the pairs nest. In any single season 

 the record of the first migrating bird observed may be a day or so after 

 the first arrival of the species on migration. To decide from journal 

 notes when the migration has ended is not so easy. While there is thus 

 uncertainty in setting the end of migration as a normal date, it appears 

 to be fairly regular, and in most years probably few birds of a species 

 arrive after the dates registered for the last migrants in table 13. I 

 believe, however, that favorable or unfavorable weather may shorten 

 or prolong migration in any year, whereas the first arrival does not 

 appear to vary in date. 



The duration of the migratory arrival period among 28 species 

 varies from 2 to 60 days. The 2 days attributed to the wagtails is 

 probably shorter than the actual period, which is nevertheless brief. 

 The long period of arrival shown for snow buntings is substantially 

 correct, for none remain to nest at Anaktuvuk. Since these early mi- 

 grants first appear at Barrow about April 15 they also have time there 



