324 H. M. JOHNSON 



It must be explicitly granted that discrimination based on 

 these characteristics can be detected as such in control tests, so 

 that no one need be deceived in drawing conclusions. My point 

 is merely that control of these factors, at some stage of experimen- 

 tation, is necessary to warrant conclusions; that both factors 

 cannot be controlled simultaneously; and that the method of 

 control sometimes proves troublesome. Breed 5 obtained positive 

 results with a chick, keeping these factors variable throughout 

 experimentation. Other students have been less successful. 

 Yerkes 6 recently recommended a different course : that one present 

 stimuli differing in more than one characteristic — e.g., in bright- 

 ness and form as well as in size, until discrimination on some 

 basis appears. This course simplifies the problem at the beginning 

 of training. Certainly some advantages are thus to be gained. 

 Perhaps the greatest drawback is the risk of a large waste of 

 time. It has happened that an experimenter discovers after 

 several weeks of training that the animal had learned to react 

 to difference of e.g. brightness, and had not been affected by the 

 other stimulus-differences. It would seem that a difference in 

 luminous intensity is especially objectionable since in dark 

 surroundings the animal can choose the brighter or darker alley 

 without attending specifically to the stimulus-forms. 



With the apparatus under discussion some experimenters 

 cause the animal to be fed near the positive form, or to pass 

 quite near it, at least, after each correct choice. Thus it is 

 possible for olfaction to become a factor. Lashley (op. cit. 

 supra) reports disturbance of this kind in his work on the rat, at a 

 stage when discrimination was not present. Substitution of a 

 new stimulus-plate for one which has been used for some time 

 may introduce a disturbance which might be unjustly ascribed 

 to the animal's inability to discriminate visually. It is difficult 

 to exclude olfaction, although its presence may be detected in 

 control tests. 



The foregoing remarks apply only to questions of convenience 

 in using the standard apparatus. It possesses also a fundamental 



6 Breed, F. S., " Development of Certain Instincts and Habits in Chicks," 

 Behavior Monographs, vol. I, No. 1. 



, " Reactions of Chicks to Optical Stimuli," Journal of Animal Behavior, 



1912, pp. 280 ff. 



• Yerkes, R. M., "The discrimination method." This Journal, vol. 2, 1912, pp. 

 142 ff. 



