ANNUAL MIGRATIONS 123 



condition as recited above, and thereafter (from 

 January 9) kept in an aviary in which it was possible 

 to reduce the lighting in stages. By February 13 

 their gonads had returned to the winter condition. 

 They were then turned into another aviary where 

 they received normal daylight then increasing at 

 about 4 minutes per day. The last samples, killed 

 May 30, were again in full breeding condition. 

 Assuming that these birds had bred in the summer 

 of 1927 (they were adults) before being trapped in 

 September, their gonads thus dropped from the 

 maximum in June to the minimum at the beginning 

 of November, were brought back to the maximum 

 by the beginning of January, again reduced to the 

 minimum by mid-February, returning once more to 

 the maximum in May. 



Ordinary electric light bulbs emit no ultra-violet 

 rays. The results obtained could therefore not be 

 attributed to ultra-violet radiation. The question 

 as to exactly what might account for them was 

 therefore of particular interest. All the evidence 

 suggested that the solution might be found in the 

 amount of exercise the birds were getting. By 

 curtailing the day-length, exercise would be re- 

 duced; by extending it, it would be increased, pro- 

 viding the birds kept awake during the available 

 period. This idea could obviously be put to a 



