Migration Flights 33 



Daily records have been made each year from early in Feb- 

 ruary until late in May for the past fourteen years and each 

 record includes a list of all the species seen, with approximate 

 numbers of each, tog-ether with memoranda concerning nests 

 and songs. During the season of migration activity these 

 records include results of early morning trips by the writer 

 and others ; trips of sections of a class in Ornithology at 8-10 

 a. m. or 3-5 p. m. taken with an instructor; and individual 

 work of a few experienced students takings more advanced 

 work who are assigned special territory for regular observa- 

 tions. In addition to these records there are nearly always 

 some members of the University Staff and other citizens, often 

 former students, who contribute results of frequent trips and 

 help materially in the adequacy of the data. The great super- 

 iority of such composite data over that possible by a single ob- 

 server has already been discussed by Cooke (Auk. 1907: 346) 

 and Stone (Proc. Acad. Sci. Phila. 1908: 128). 



The combined results of a number of observers makes a list 

 of dates of " first seen " conform much more closely with a list 

 of dates of first arrivals than is possible with a single observer. 

 A comparison of the records of the writer with those of the 

 other observers during the years of 1912-1916 shows the fol- 

 lowing relations between his list of dates of " first seen " and 

 those of the other observers. 43.5% of the "firsts" were 

 recorded on earlier dates by others; 26.5% were recorded 

 earlier by the writer; and 30% were recorded on the same 

 dates by both. This is a decidedly better showing- than that 

 of Mr. Cooke and his collaborators at Washington who only 

 recorded about 8.5% of the " firsts " on the same days. 



In the problem under discussion we must determine what 

 measurable phenomenon we shall use as an evidence of migra- 

 tion having taken place. We may use the appearance of species 

 which have not previously been recorded for the season ; or 

 we may use obvious increases of numbers of individuals or 

 bulk movements ; or we may use the non-appearance of birds 

 which have been recorded on the next preceding lists. For 

 the purposes of this paper I shall deal with records containing 



