138 PROCEEDINGS OF THE ACADEMY OF [April, 



1906— March 9, April 10, 16, 17 (2), 19 (3), 20, 21 (4), 22 (3), 24 (2), 



25 (2), 26, 27, 28 (2). 30, Mav 1 (2), 5. 6. 

 1907— March 13, 17, April 20, 23. 25, 26 (6), 27 (7), 28 (2), 29 (2), 30 



(2), Mav 1 (3), 2, 3, 4'(2), 5, 8, 11 (2). 

 1905 April 22, 1906 April 21, 1907 April 27. 



The above plan gives us a definite date for all sorts of comparisons 

 and one which is independent of the personal equation. The term 

 "became common" may mean a different thing to each individual, but 

 the date upon which a species reached half of the stations at which it 

 was observed represents a definite 'point in the increase of its abundance, 

 and is a matter of record and not of opinion. 



As so little has been attempted in the way of combining local migra- 

 tion records, I find it difficult to discuss the comparative value of dif- 

 ferent methods. Some casual allusions by Prof. Cooke to the methods 

 employed by him form indeed the only contribution to the subject 

 with which I am familiar. He recognizes the clanger of including the 

 latest dates of arrival in computing averages and rejects them, just as 

 I have advocated above, but in deciding how many to reject his method 

 seems to lack clefiniteness and to involve the personal equation. He 

 says {Auk, 1907, p. 347), "When using migration records for the calcula- 

 tion of average dates of arrival, I usually discard dates that are more 

 than six days later than the probable normal date of arrival." This 

 would seem to imply an arbitrary selection of "the probable normal" 

 date before any averaging is done, which seems to be a dangerous 

 method. Again, in referring to the combination of the observations, 

 of twenty-three observers at Washington, D. C, in the Spring of 1907, 

 he says, "Many of the notes were duplicates or of no value, but after all 

 these had been eliminated," etc. [Italics mine]. This is exactly the 

 reverse of my method, instead of rejecting "duplicate" records, 

 these seem to me to be of the utmost value as pointing to the dates 

 upon which the greatest migration took place. It must, however, be 

 borne in mind that Prof. Cooke in this instance is ascertaining the 

 earliest date — not the date of bulk arrival which, as just explained, 

 seems to me a more reliable basis for comparison of migration between 

 two distant points, but one which, as I have also explained, is practically 

 impossible in the absence of a large corps of observers at each point. 



Graphic Representation of Migration. 



In the Auk for 1889 (p. 139) and 1891 (p. 194) I pubhshed some 

 papers on the Graphic Representation of Bird Migration, based in 

 part upon records of the Delaware Valley Ornithological Club for 1890. 



The attempt was made at this time to record tke actual number of 

 individuals or the relative abundance of certain species, as noted each 



