HABIT FORMATION IN THE DOG 75 



boxes, which the animals had learned and were then opening in 

 one second or less, were provided with contacts of brass. When 

 the door of the box was closed, a circuit was formed, the cur- 

 rent running from the positive pole of a dry battery through 

 the magnet of a* Schallhammer in the experimentor's room, to 

 the metal contacts on the door and jamb of the box, to the 

 negative pole of the battery. When the latch of the door was 

 released this circuit was broken and the hammer being drawn 

 from the magnet by a spring gave a signal-click. The animal- 

 room was darkened by heavy black tar-paper nailed o^'er the 

 windows and openings about the door. The tests were made 

 at night and all lights in the animal-room and the room ad- 

 joining were put out before the animal was admitted to the 

 experiment-room. Time was taken from the animal's passing 

 between the glass door and the food-box, and the click of the 

 Schallhammer when the box was opened. This is accurate to 

 within one-half second. The blind dogs were slightly disturbed, 

 as is shown by the fact that the time of their first ten reactions 

 varied between four and seven seconds ; the next ten averaged 

 about two. Dog 5, normal, was much more disturbed. His 

 first reaction under the changed conditions required thirteen 

 seconds and there were many useless movements. Several 

 times in the first six trials he climbed on top of the box and 

 scratched there. The reactions after the sixth were more defi- 

 nite, and after the twelfth there were no more errors, and the 

 time for each thereafter averaged 1.3 seconds for thirty trials. 

 These results indicate that the light had some stimulating effect, 

 on both the blind and the normal dogs. 



The second control was a test on Dogs i, 2 and 5, of disturb- 

 ance when a box previously learned by the dog was oriented 

 in another direction. As the boxes stood when the animals 

 w^ere learning them, the latch was always at the northwest 

 corner; when the box was turned 90°, then the latch was on 

 the northeast corner, etc. The following table summarizes the 

 disturbance which follows : 



