30 HARRY MILES JOHNSON 



is evidently present here, and the experimenter was unwilhng 

 to accept these negative results as conclusive until the tests 

 had been made under such conditions as would compel the 

 animal to make her choice very soon after the stimulus is 

 presented. 



From these preliminary experiments it becomes evident that 

 in order to reach reliable results in the field of pitch-discrimina- 

 tion new apparatus must be constructed and such conditions 

 fulfilled as will enable the experimenter to exclude certain 

 vitiating factors inherent in tests like my first ones, and of 

 the Kalischer and Rothmann type. An enumeration of these 

 defects, with apparatus proposed for their elimination, is set 

 forth in the section next followins:. 



THE CONDITIONS OF A DECISIVE EXPERIMENT ON PITCH-DISCRIMIN- 

 ATION IN ANIMALS 



It is probably evident that neither in the results of my own 

 preliminary experiments nor in the work of the other investi- 

 gators which has been discussed, can we be at all sure that the 

 dogs were reacting only to auditory stimuli. Some obvious 

 secondary cues ha\-e to be eliminated from the experimental 

 conditions if results of future experiments are to be reliable. 

 It seems worth while to mention these disturbing factors, some 

 of which others have previously pointed out in describing dis- 

 crimination-experiments of other kinds. 



I. "Unconscious helps.'' By this term is meant any sort of 

 body-movements, which the animal can learn to associate with 

 a definite reaction. Those likely to be made by the experi- 

 menter under such conditions as Kalischer 's and Swift's, or by 

 the assistant in Rothmann 's experiment — namely, such nascent 

 movements of the arms and body as would accompany readiness 

 to strike or to step back to allow the dog to obtain food, are 

 particularly vicious, as appears in the "Clever Hans" report. 

 They can be detected visually by many animals. But visually 

 sensed aids by no means exhaust the list. Suppose the operator 

 has been in the habit of recalling the dog when a wrong choice 

 is made, or of scolding him if he fails to inhibit, and encourag- 

 ing him by a word if he is timid in reacting, when the problem 

 is like Kalischer 's and Rothmann 's; a very slight change in 



