the disc, making scrambling movements with his feet. He 

 need not have leaned over so far, as the bridge dropped with 

 less pull on the string, and he could have omitted, for the 

 most part, the foot movements, so that the reaction was not 

 reduced to the greatest efficiency. 



These movements, however, were always made on the 

 side where he could reach the string, even though he some- 

 times wandered about the disc at random. After a lapse of 

 five weeks he did at first revert to his old attempts to climb 

 down anywhere, but this reaction was soon corrected. 



In the following table are stated both the time the mouse 

 took to drop the bridge after being placed on the disc and 

 the time he took to descend to the ground after the bridge 

 had been dropped. 



*Time given in seconds. 

 The time elapsing between the tenth and the eleventh 

 series was five weeks, whereas that between all other adjacent 

 series averaged twenty-two hours, with a maximum interval 

 of thirty-six hours. 



[TWENTY 



