134 PROCEEDINGS OF THE ACADEMY OF [April, 



Or, tabulating the averages obtained as above for eleven species, 



we have: No. of days 



I. II. III. difference. 



Chimney Sivift April 21 April 21 April 21 



Phoebe Mar. 27 Mar. 15 Mar. 20 12 



Chipping Sparroiv " 29 April 1 "29 3 



Scarlet Tanager Mav 8 May 4 May 7 4 



Barn Swallow April 19 April 19 April 22 3 



Black-throated Blue Warbler. May 5 Mav 2 May 3 3 



Ovenbird " 1 April 29 April 30 2 



Maryland Yellow-throat April 25 " 26 Mav 2 7 



Catbird " 28 '•' 25 April 30 5 



Brown Thrasher " 19 " 21 " 18 3 



Wood Thrush " 30 " 27 '' 29 3 



This demonstrates conclusively that the average date of arrival 

 for a number of years, based upon the observations of a single individual, 

 varies materially from the average date obtained by another equally 

 accurate observer stationed but a few miles distant. The amount of 

 difference in the case of individual observers is even greater than that 

 shown above, as in these cases the record given for each of the three 

 stations is the result of the combined work of several observers. 



I called attention to the percentage of error in the records of indi- 

 vidual observers in a paper read before the American Ornithologists' 

 Union at New York in November, 1905, and during the Spring of 1907 

 Prof. W. W. Cooke made some experiments along the same line, and 

 his results showed that, compared with the combined work of twenty- 

 three other observers, in the immediate vicinity of Washington, D. C, 

 in this single season his dates of arrival averaged one and three- 

 tenths of a day late, and this in spite of the fact that he spent more 

 time in the field and covered a greater variet}: of country. In my 

 summary given above a single station averages one and nine-tenths 

 of a day later than the earliest average date recorded for the species. 



This information, however, does not help us in using the record of a 

 single observer for comparative study, either as between different 

 years or different stations, and we are forced to the conclusion that 

 results based upon such individual records are really of but little 

 value for comparative work, so great is the possibility of error. 



For instance, quoting from Prof. W. W. Cooke's papers on the 

 Migration of Warblers and Thrushes, as recorded in the schedules of 

 the U. S. Department of Agriculture,^ we have the average dates of the 

 arrival of the following species at Germantown, Pa., a suburb of 

 Philadelphia, and at Washington, D. C. : 



1 Bird Lore, 1905-1907. 



