218 MIGRATION 



the evidence against the latter, the supposed regularity no longer 

 seems important, but it is worth while to examine it for itself, as 

 well as in relation to length of day and to temperature. 



J. C. Phillips (1913) expresses his view of regularity as follows: 

 "This brings us back to our inquiry into the mechanism by which birds 

 are enabled to arrive each year at a given locality at almost exactly 

 the same time. From a physico-chemical standpoint, the accuracy 

 of time sense in certain species is little short of marvelous." Wetmore 

 (1926, 1930) considers that "the regularity of travel when birds are on 

 migration constitutes one of the most interesting facts in connection 

 with this phenomenon and is one familiar to all ornithologists. 

 Through years of observation average dates of spring arrival and 

 autumn departure have been established for many localities, and birds 

 come and go with surprising regularity on their appointed dates. Ar- 

 rival in spring is particularly punctual with the majority, and unusual 

 is the season when the first of the travellers fail to put in their appear- 

 ance within a few days of the average dates. At Washington, D. C, 

 the barn swallow, on the average, arrives April 12, the least flycatcher 

 ]May 2, the chipping sparrow March 22, and the house wren April 18. 

 On or near these dates one is always sure to find them. Individuals 

 which breed about our homes come with particular promptness on 

 their appointed days. Severe indeed is the weather that delays them 

 for any length of time. Departure for the south in autumn is prompt, 

 but has greater range of variation, particularly in middle latitudes, as 

 prolonged mild weather may induce birds to remain beyond their 

 custom." It is manifest that "regularity," "punctual," and "prompt" 

 are here to be interpreted in terms of averages, and hence it is inter- 

 esting to examine the records to secure an objective evaluation. 



The belief in regularity has been fostered by averages over a short 

 period, which convey fictitious values, but it springs chiefly from 

 prepossession. The most striking example of this has been cited by 

 Phillips {loc cit., page 195) in the following statement: "Cooke's 

 method of averaging migration arrivals consists in throwing out dates 

 which are more than six days out of the way, his experience teaching 

 him that 'birds seldom vary on account of the season more than six 

 days either way from the average date of their arrival.' The method 

 may seem to some, as to the writer, rather arbitrary." Just how 

 arbitrary is revealed in the following paragraphs, where it may be 

 noted that few species possess the degree of regularity demanded by 

 the method. 



A brief record is misleading. Of the six utilized by Phillips, five 

 range from 6 to 11 years; the other is 23 years in length, during which 



