136 PROCEEDINGS OF THE ACADEMY OF [April, 



ally suggest some method of recording migration by which results 

 sufficiently accurate for comparative work are to be obtained. This, 

 I think, is to be found by securing a large number of observers in a 

 limited area and by combining their results, as has been done by the 

 Delaware Valley Ornitliological Club in the vicinity of Philadelphia. 

 If we had seven-year records kept by thirty-five individuals within 

 ten miles of Washington, and a similar series within ten miles of 

 Boston for comparison with the Philadelphia series, then I think we 

 should be able to estimate with some degree of accuracy the progress 

 of migration between these points. 



In a composite record of this kind it is especially worthy of note that 

 more or less fragmentary records are of great value, as an observer 

 who only records a hmited number of species may note some of them 

 earlier than any other observer, while species which he fails to record 

 are noted by others. 



The way in which a number of indvidual records from one vicinity 

 are to be combined in order to get the most reliable results is quite a 

 problem. 



Take, for example, the Ovenbird, Seiurus aurocapillus, for the years 

 1905, 1906 and 1907, as recorded within ten miles of Philadelphia by 

 respectively thirty, thirty-two and thirty-four observers — the number 

 of the observation corps varying somewhat from year to year. 



We find that in 1905 it arrived at one station on April 25; at another 

 on April 28; at eight stations on the 29th, ten on the 30th, etc., i.e. : 



1905— April 25, 28, 29 (8), 30 (10), May 2, 3, 4, 6, 7 (2), 8, 12 (2). 

 1906— April 28 (2), 29 (7), 30 (4), May 1 (5), 2 (3), 3 (4), 4 (2). 5 (3), 



8 12 

 1907— April 26 (2), 27, 28 (4), 29 (5), 30 (2), May 1 (5), 2 (2), 5 (4), 6, 8, 



11 (4), 12, 13, 15. 



If we select the earliest date for each year as the basis of our com- 

 parison, we shall say that 1905 was the earliest season and 1906 the 

 latest. The objection to this is that it considers only the earliest 

 stragglers, whose movements may or may not reflect those of the bulk 

 of the species. 



If we select the average of all the dates for each year we shall have 

 for 1905 May 2, 1906 May 2, 1907 May 3, or 1905 earhest and 1907 

 latest. The objection in this case is that some at least of the late 

 dates of arrival represent errors of observation — i.e., failures to detect 

 the species until it had been present for some days — while others are for 

 stations which are not congenial haunts of the species under considera- 

 tion and at which it is only occasionally seen, and by including these 

 in our computation we obviously make the resultant date too late. 



