HABIT FORMATION IN THE DOG 21 



It seemed better also to adopt a different test of discrimina- 

 tion from that used by Kalischer, Rothmann and Swift. This, 

 as has been mentioned, was to allow the animal to take food at 

 one or more tones and to refrain from it at others. If an animal 

 is hungry and eager to obtain food, one would expect a ten- 

 dency to react to the mere striking of any tone ; while if the 

 animal is timid there may be a tendency to inhibit reaction, even 

 though the tone were "recognized" as the food-tone. So, to the 

 two tones used as stimuli, two reactions quite dift'erent were 

 chosen; to the deeper tone the animal was trained to react by 

 placing her forefeet on a chair at the operator's left, and waiting 

 there for food; to the higher tone, by mounting a low box at 

 the operator's right, and " sitting down " on it until fed. Between 

 stimuli the animal sat on the floor, at the experimenter's feet. 

 Food was given after a correct reaction had been chosen. In 

 case of incorrect choice she was recalled without being fed unless 

 she was unusually nervous, in which event occasionally she w^as 

 allowed to perform the correct reaction and take food, the reac- 

 tion being recorded as an error. The problem was considered 

 "learned" when the animal had perfonned three successive 

 daily series of reactions without error. 



The two stimulus-tones chosen were middle c (256 d.v.) and 

 the g above (384 d.v.). The tones were sounded at first on two 

 standard tuning forks, mounted on wooden resonators, and 

 struck by hand. The forks were placed close together on a 

 shelf in front of the experimenter, and rested on heavy cotton 

 felt pads. Their relative position was frequently changed. 

 Later in this experiment the tones were sounded also on several 

 Stern variators, large and small, blown from a Stem tank, and 

 carefully tuned each day to the tuning forks. The merits and 

 defects of this apparatus will be discussed later in this paper. 

 Each tone was sounded until the animal had reacted — usualh' 

 not more than one and one -half to two seconds. 



For about six days, the animals, both of which were seem- 

 ingly frightened when the forks were first struck in their pres- 

 ence, were "put through" the proper reactions — to one tone 

 several times and then to the other for the rest of the series. As 

 soon as they showed a tendency to react voluntarily they were 

 allowed to work without consciously given help, at least, from 

 the experimenter. Records were taken from this point on. 



