INTEGRATION OF MOVEMENTS IN LEARNING IN THE RAT 387 



ing of the nature or of the possibilities of any problem rests upon 

 the same factors. Accordingly, care must be taken in selecting 

 the problem as well as the animals used. It is necessary in this 

 work to select animals that can be had in sufficient numbers, so 

 that a given observation of any or of all movements can be re- 

 peated in order to establish their relationship, their constancy, 

 and their importance. Furthermore, to have animals of a con- 

 venient size is an essential condition, for in larger animals are pro- 

 duced many lesser movements unfavorable for a given observation. 

 No animal seems to fulfill these conditions, and appear better 

 fitted for experimentation than the white rat. 



The selection of the first problem revealing in a small space 

 all the movements made by the rat and offering the greatest op- 

 portunity for the investigation of integrated movements would 

 be of one producing movements of simplicity and of definiteness. 

 No other problem seems to be as suitable in revealing these quali- 

 ties as the latch-box problem. Accordingly, it is the first one 

 selected for the investigation of learning of integrated movements. 

 The only movements necessary for the solution of the problem are 

 progression to the latch, and the simple act of lifting the latch 

 out of the notch by the raising of the head. These integrated 

 movements involved are sufficiently definite and constant, and 

 are inherently conjoined or coordinated. Preliminary work with 

 several rats established these important points. 



The latch-box (fig. 1) was made entirely of metal. The box 

 frame was constructed of 1.5 cm. square rod of aluminum. The 

 box was 25 cm. square with a door 6 by 10 cm. in the lower cen- 

 ter of one side. Within the sides, the top and the door were 

 fitted lightly constructed frames covered with 25 cm. wire mesh. 

 The door was closed by a latch, which was snugly held by the 

 pressure of a close spiral spring attached to the inside of the 

 door and to the opposite side of the box. When the latch was 

 released, this same spring gently opened the door. The problem 

 box was covered with a wire hood with plate glass covering part 

 of one side, so that the behavior of the rats in any position under 

 the latch could be observed. For the same reason, one side of 

 the entrance box to the hood was covered with glass for observa- 



