320 Journal of Couiparativc Neurology and Psychology. 



In his experimental work Thorxdike emphasizes the 

 method of learning as that of a selection from a large number 

 of random mov^ements of certain movenients which are stamped 

 in by the pleasure of success. I believe, however, that a guinea 

 pig may be taught a trick without waiting for selection from 

 among random movements. What was done was to "control the 

 impulse," and by impulse we mean the amount and direction of 

 muscular innervation. 



On the previous Saturday the guinea pigs had remained 

 almost motionless for two hours after being put into the experi- 

 mental cage. On Thursday the problem had been learned. 



A typical series of reactions is given, after the problem has 

 been learned. Until that time the difficulties of fright, strange- 

 ness, etc. , already mentioned, rendered any time record or 

 other measure wholely meaningless. 



The apparatus used was a wire box, lO by lO by lO in., 

 with a wire door hung from the top so as to swing freely in and 

 out Care was taken that nothing should distinguish the door 

 from the rest of the box. In every case the guinea pig has had 

 no food for 24 hours. The food stimulus used is always carrots 

 freshly cut, which has a strong odor ; also it is in plain sight in 

 the wire box. The animal is adult, and in this example is of 

 the solid red Peruvian variety, though mixed with the solid red 

 English. 



' It will be observed that throughout the work more than one trial was given 

 during an experiment. At any one time only a little food was given. 



[:8] 



