2l6 yournal of Comparative Neurology and Psychology. 



into a hole in the door frame. The box had entirely closed sides. The raccoons were taught to go into 

 this box to get food. At first both levers had to be pushed up but later they were arranged so that 

 considerable force would push them downward. The plug was very difficult to draw. I have com- 

 pared this with Thorndike's Box K. 



Observations on Reactions to Fastenings. — The raccoons learned 

 very readily to perform a certain act in a particular situation. 

 This learning is doubtless of the trial and error type, yet when a 

 latch has been operated a few times there is probably present in 

 the animal's mind a distinct memory image of the act, includmg 

 a memory of its difficulty. Experiments with colored cards, to be 

 described later, gave evidence in support of this opinion. At first 

 I supposed that, as was true in Thorndike's work with cats, the 

 raccoons would be found to learn chiefly from the stimulus of 

 hunger. As already stated, however, they soon showed a ten- 

 dency to unfasten latches and set themselves free from the mere 

 pleasure of performing the act. This motive was not strong 

 enough to overcome the discouraging difficulties of a box of six 

 or seven fastenings, but the tables show so-called "play trials" 

 for all boxes of fewer fastenings. The term "play trials" means, 

 then, that though the animal unfastened the latches and escaped 

 from the box, he refused to eat or drink milk on coming out or at 

 best merely tasted the milk and turned away from it. Generally 

 this work was deliberately done, but often rapidly. It seemed, 

 therefore, that this tendency to be occupied was the motive for some 

 of the raccoon's normal learning and careful records were kept 

 of all play trials. In Table I, I have indicated the cases that were 

 certainly play trials with italics, but the cases which were certain 

 to the observer were fewer than the actual number, for as the 

 raccoon's hunger was gradually allayed he worked partly for the 

 mere pleasure of doing the work and partly from hunger. This 

 is shown in the longer times taken to escape toward the close of 

 each day's work. When the animal showed any eagerness for 

 food, the reaction was recorded as a hunger trial even though play 

 trials had preceded it. The tables show that this was unusual, 

 the rule being that play trials began only when hunger began to 

 be satisfied. Even when using the most complicated fastenings 

 I did not employ "utter hunger." I usually gave the raccoons 

 considerable food after I had finished the day's experiments. In 

 several trials with the raccoons, when they were young, I was able 

 to get one to work, which otherwise would not do so, by bringing 



