4 The Wilson Bulletin — No. 82. 



The remaining" records show an average extreme variation 

 of 10.0 days between the earUest and the latest dates of ar- 

 rival for each species, with an average variation of 2.9 days 

 from the normal date of arrival. 



The records of these same thirty-six species for the same 

 six years at Grinnell, Iowa, after rejecting 13 per cent be- 

 cause too late or too early, show an average extreme varia- 

 tion of 9.4 days and an average variation of 3.4 days from 

 the normal date of arrival. The records from Lanesboro are 

 thus seen to be a trifle more regular than those from Grinnell. 

 As it is hardly supposable that the birds themselves are more 

 variable in one of these places than the other with such closely 

 similar physical surroundings, it must be that this difiference 

 is due to differences in ability or opportunity on the part of 

 ihe observer. In this particular case it is probable that the 

 two observers knew birds about equally well, but the Lanes- 

 boro observer was so situated that he could spend more hours 

 per day in the open air than the observer at Grinnell, with 

 the result that he averaged seeing the birds a little sooner 

 after their arrival. 



These final averages of 2.9 and 3.4 days represent the prob- 

 able variation in the observance of the arrival of the first, 

 or in other words a first-class observer, who is in the field 

 every day and is able to spend time enough each day to cover 

 his neighborhood satisfactorily will make records that show 

 an average variation from the normal date of arrival of about 

 three days, and unless extreme attention is paid to the birds 

 another half day will be added. This time — three days — is 

 the sum of two variables, first, the variations due to the birds 

 themselves as they vary their actual date of arrival from year 

 to year; and second, variations due to failures on the part of 

 the observer to note the species immediately on its arrival. 

 vSince the second of these causes must have some influence it 

 follows that the birds themselves must be remarkably uniform 

 in their spring movements. 



The above investigation was undertaken for the purpose 

 of obtainino- some standard that could be used in testing mi- 



