ANNUAL MIGRATIONS 125 



returning to the floor and thence along it and so on. 

 Nowhere in the cage could a bird sit still for more 

 than 20 seconds at a time as it was moved on by the 

 travelling bar. When the juncos had been trained 

 to the movement in full daylight, the room was 

 shuttered from 6 in the evening till 9 the following 

 morning. The room was then completely dark but 

 for a small 5 watt electric-light bulb suspended from 

 the ceiling. This provided just enough glow for 

 the birds to see the advancing bar. It was so feeble 

 that the control birds in the other cage, getting the 

 same illumination fell asleep in spite of it. The first 

 night the mechanism ran for 7| minutes, the next 

 for 15 and so on, the 7^-minute increase being 

 adhered to till the termination of the experiment. 

 In brief, the rate of development of the gonads 

 exactly paralleled that of the birds getting similar 

 periods of extended illumination outside, while the 

 gonads of the caged controls remained stationary. 

 It was impossible to run the machinery in complete 

 darkness as it would have meant certain death for 

 the birds. The illumination, however, was such as 

 to preclude any question of radiation. The length 

 of time spent in activity, rather than the amount of 

 exercise obtained (which must have been slight 

 owing to space limitations) seems to be the crucial 

 factor. Here then, apparently, is the missing link 



