BEHAVIOR OF VERTEBRATES 437 



Rouse that pigeons are very sensitive to sounds. From the 

 tests with the rotated maze Hunter draws the conclusion that 

 pigeons seem capable of changing their system of cues in response 

 to the varying demands of a new situation. It is unfortunate 

 that Hunter could not carry out these rotation experiments in 

 an apparatus which permitted of the simultaneous rotation of 

 the background. 



Cole (7) continued the work of Yerkes and Dodson on the 

 relation of the strength of the stimulus (punishment by electric 

 shock) to rapidity of habit formation when the discrimination is 

 (1) easy, (2) medium, (3) difficult. Chicks (barred Plymouth 

 Rocks) were used as the subjects. The relative brightness of the 

 stimulus plates in the three sets of tests is given as follows : 



(1 ) Easy discrimination o: 8.9 



(2) Medium " i:i 3-7 



(3) Difficult * 1: 5.1 



Punishment was administered by grills connected with a cali- 

 brated inductorium. It was found that the number of trials 

 necessary to form the habit of choosing the darker of the two 

 screens when the difference in illumination is great decreases 

 with the stimulus. This law holds, for discriminations of medium 

 difficulty, only for the lower intensities of the stimulus. It holds 

 likewise for difficult discriminations if the records of the chicks 

 which succeeded in making it are alone considered. " If, how- 

 ever, we consider only the chicks which fail, the optimal stim- 

 ulus recedes once more to a point nearer the threshold of stim- 

 ulation than in the case of medium discrimination." 



In the present state of the subject it is very doubtful if such 

 results can be wholly relied upon. Quantitative experiments 

 designed to isolate the effect of punishment are full of pitfalls. 

 The effect is masked by practice effects, individual variation in 

 animals, variations from day to day in the same animal, and 

 by the almost insuperable difficulty of keeping an inductorium 

 running for long periods of time without introducing serious 

 errors due to the oxidation of the contacts through sparking, 

 heating, etc. No condenser was connected across the terminals. 

 Apart from the difficulties inherent in the method. Cole certainly 

 worked with extreme care. The only criticism which can be 



