328 K. S. LASHLEY 



The responses to sudden changes in the form of the stimuli 

 are of importance as evidence of discriminating ability. Rat 

 No. 6 was choosing the larger of two circles accurately (100%). 

 Two squares of the same area, respectively, as the circles, were 

 substituted for the circles and during the next 10 trials No. 6 

 was greatly disturbed, running in and out of the starting box, 

 creeping a little way down the passages, then turning back. 

 She attended to the squares as to something new. There had 

 been no noise to account for this behavior and it is not probable 

 that the new forms had any peculiar odor, so it seems that the 

 rat must have been disturbed by the change of form. 



This is only one instance of a type of behavior which appeared, 

 to a greater or less extent, throughout all the experiments when- 

 ever the form of the stimuli was changed. It suggests that the 

 animals perceived the difference, although they never associated 

 it with the motive, and raises the question whether the discrim- 

 ination method offers a fair test of the animal's discriminative 

 ability. 



SUMMARY OF RESULTS 



In all the records there is a constant deviation in the per- 

 centage of error, and a series of 20 successive correct trials is 

 very rare. Sudden increases in the number of mistakes may 

 frequently be explained by the physical condition of the animals. 

 If they are fed too much or too little, their attention to the 

 stimuli is affected and the percentage of error increases. When 

 punishment is used, the rat's fear of the punishment grill after 

 too severe a shock causes irregularities in the records. 



Making allowance for such deviations, the selection of the 

 vertical line 164 times in 200 trials (experiment 5) proves that 

 rat No. 5 could distinguish between the vertical and horizontal 

 lines. In discrimination of size, the record of only 14 mistakes 

 in 150 trials shows that rat No. 4 could discriminate between 

 the 30 and 50 millimetre circles. 



The number of animals used in the experiments has been 

 small, necessarily, since the time required for each series is 

 considerable. The rats show marked individual differences and 

 all vary from day to day in their attention to the problem. The 

 quantitative results are therefore only approximate. The tests 

 for the threshold of form discrimination have been carried 

 scarcely far enough to give definite results, and the data at 



