224 MIGRATION 



ccntages are 42, 17, 24, and 22 per cent. For the combined lists, the 

 number of earliest arrivals on the year of .sunspot maximum or mini- 

 mum is 145 by comparison with 542 at other years of the cycle, or a 

 percentage of 27. 



When a period of 3 years centering on M or m is employed, thus 

 giving a relation comparable with the smoothed curves of De Lury 

 and of Wing, the number of earliest arrivals at these times is greatly 

 increased, in accordance with the expectation. For Oberlin they are 

 nearly trebled, amounting to 88 out of 159 and yielding a ratio of 

 88:71 and a percentage of 55, while for Raleigh they are 87 of 140, 

 with a ratio of 87:53 and 62 per cent. If the two successive Ann 

 Arbor records are combined and the earliest rate of arrival taken for 

 each species, the respective values are 104 species out of 169, 64 mak- 

 ing their earliest appearance about the maximum, and 40 about the 

 minimum. 



In the case of 25 species with the most complete records at Ann 

 Arbor, the 3 earliest and 3 latest arrivals were taken for each, some 

 dates being duplicated in the first, and the number that fell at or 

 within 1 year of the maximum or minimum determined. For earliest 

 arrival, the ratio was 93:163 or 57 per cent; for the latest, 75:149 or 

 50 per cent. For arrivals on the precise year of an extreme, the num- 

 ber was the same for both, namely 31. The number of M and m years 

 in which there was no early arrival ranged from 4 in the kingbird, 

 with 2 such years lacking in the record, to 6-9 in a possible total of 

 9 for 19 of the 25 species. By combining the records of Atkins at 

 Locke (1883) and Wood at Ann Arbor, made less than 50 miles apart, 

 a record of 76 years is obtained for the redstart, the spring return 

 ranging from April 5 to June 5, or 62 days. This yielded but 1 very 

 early return, namely, April 5, 1903, 2 years after the minimum, the 

 next 3 in order being 20-23 days later, 1 on the minimum of 1889. 

 Of the 13 April dates, all of which might be considered early, 3 oc- 

 curred at the minimum and none at the maximum. 



The above summary affords little support to the hypothesis that 

 spring arrivals are earlier at sunspot maxima and minima, only a few 

 species exhibiting a limited amount of agreement. However, the de- 

 gree of general accord may prove somewhat greater than now appears, 

 when the sunspot extremes are correlated with temperature and rain- 

 fall and resultant food, and these with the time of migration. Such 

 a procedure might serve to explain the wide divergence in date of 

 arrival for the same year between two localities. Thus, the day of 

 return of the piedbill grebe at Ann Arbor and Oberlin may differ as 

 much as 4-6 weeks for a particular year and hence partly reflect 



