120 THE RIDDLE OF MIGRATION 



lows : Two aviaries were used , the one provided with 

 electric light, the other not. Birds housed in the 

 former will be referred to below as the experimen- 

 tals, those in the other (unlit) as the controls. The 

 lights consisted of ordinary frosted electric globes 

 totalling 1050 watts, distributed in such a way as to 

 ensure uniform lighting throughout the aviary. 

 The illumination was not intense but sufficiently 

 strong to enable minute details to be readily made 

 out on the ground without stooping. Lighting was 

 started at the beginning of November (1927)run- 

 ning, on the initial day, from sunset, 5.05 p.m., till 

 6 p.m. The following day the lights went on again 

 at sunset (as daily thereafter) but ran till 7 J minutes 

 after six; the next day till 6.15 and so on — an increase 

 of 7| minutes each night. Sunset was chosen as 

 lighting-up time as the birds were then still fully 

 active even on cloudy days. The interval of 7J 

 minutes, after deducting the 2 or 2| minutes of later 

 sunrise each morning, gave an effective daily in- 

 crease of about 5 minutes, roughly equal to the 

 spring increase of day-length experienced by north- 

 bound juncos in southern and central Alberta. 



From December 3 onwards the interval was 

 reduced to 5 minutes daily till December 15 when 

 the lights were burning from sunset till 11 p.m. 

 Thereafter increases were discontinued, the lights 



