34 PROCEEDINGS OF THE 



uals and species, and its relations to variation in temperature 

 and meteorological conditions. 2n(l. The variation in the time 

 of arrival of the various species at a number of stations of differ- 

 ent geographical position both with respect to latitude, altitude 

 and proximity to water courses. 



The first line of investigation requires a daily estimate of the 

 number of individuals of each species present in a given district, 

 and necessitates extreme care and accurateness on the part of 

 the observer as well as uniformitj' in the length of time spent in 

 the field and in the extent of country covered each day — re- 

 quirements almost impossible on the part of the majority of 

 observers. In fact, previous experience has shown us that in 

 the majority of records of this kind, the increased opportunity 

 for observation on Sundays or holidays always makes an appar- 

 ent increase in the migratory movement on these dates, and 

 while this work was carried on with success by a few observers 

 in 1890 and 1891 it is impractical for general investigations. 



Our efforts for the present year have therefore been devoted 

 to recording the " first arrivals" at as many stations as possible. 

 Occasional early arrivals in advance of the bulk of the species 

 are often noted, but no attempt has been made to have the ob- 

 server note the "arrival of the bulk" since experience has 

 shown such a variation in the interpretation of this term that 

 the records are much less comparable than those of "first 

 arrival." 



In tabulating the arrivals of a species at a number of nearby 

 stations a considerable variation in the dates is noticed, due: 

 (1) to the fact that some observers are more regularly afield and 

 detect the presence of a species as soon as it arrives, while others 

 are not out so often or are not as keen observers; and (2), to 

 the fact that each observer covers only a limited area and often- 

 times the first individuals of a species may drop into his neigh- 

 bor's wood several days before any of them find their way into 

 his own ground. These causes of error are found to be largely 

 eliminated by having a number of observers clustered in one 

 vicinity and by combining their records; selecting the earliest 

 date for each species. The resultant record compared with a 

 similar composite record for some other district will give a more 



