64 KNIGHT DUNLAP 



" wrong" act has been in every series. In that case, the animal 

 should eventually repeat the "right" act uniformly preceded by 

 the "wrong" one. 



3. If the apparatus is so disposed that satisfaction is not 

 given to the animal until several "wrong" acts have been done 

 after the act which really makes the satisfaction available, learn- 

 ing will be made especially difficult, unless the animal is able to 

 make a conceptual analysis of the problem. 



4. In a problem involving the necessity of a definite series 

 of actions for its solution, and allowing the performance of 

 "wrong" acts at various points, the elimination of wrong acts 

 will follow no law except that of probability, based on a num- 

 ber of repetitions. Of the wrong acts which have been done 

 up to a certain point in the learning, those which have been done 

 the fewest times will be eliminated soonest. 



5. In the human animal, in spite of the non-emotional fac- 

 tors which enter most learning problems, the hormone factor 

 should be discernible. 



