DISCRIMINATION OF SIZE AND FORM IN THE RAT 329 



hand are confused by the illness of the rat (No. 6) but it seems 

 probable that she could distinguish between the '20 by 30 milli- 

 metre rectangles used in experiment 6. The difference between 

 these forms is scarcely greater than that of the stimuli used in 

 experiments 1 and 4, so this difference must lie somewhere near 

 the threshold for the conditions given. 



In the discrimination of size the records of experiments 9 

 and 10 indicate that the rat can distinguish a difference of 

 one-fourth the area of the larger circle (2000 square millimetres). 

 Reduction below this limit caused uncertainty of reaction and 

 loss of attention. In addition to their ability to compare and 

 choose between the stimuli, the rats are able to fix upon a rough 

 standard of size. In choice by affirmation and negation they 

 tended to go toward a circle of more than 40 millimetres diam- 

 eter and to avoid one of less than 30 millimetres. But when 

 both circles were made larger than this standard the rats still 

 chose the larger of the two by comparison. 



The long failure to react to the interrupted light in experi- 

 ment 2 suggests that the factor of movement may play a less 

 important part in the activities of the animals than is generally 

 supposed; that an accustomed moving object is no more effec- 

 tive in catching and holding the attention than is a stationary one. 



THE NATURE OF THE RAT'S IMAGE PERCEPTION 



The rat perceives brightness differences readily, but ordinarily 

 without attention to the nature of the image. Animals which 

 have been trained to go to a white food dish will run to any 

 white object. Absolute size is not so easily recognized, but 

 greatly different areas are distinguished without difficulty. The 

 perception and recognition of form seem to be most difficult 

 for the animal and to require an unusually close attention. In 

 these respects the rat's visual perceptions resemble closely 

 human perception in the extreme peripheral field. Brightness 

 is most evident and size is more readily perceived than form. 



TRAINING METHODS 



The work of Hoge and Stocking, (191 2) indicates that pun- 

 ishment is more effective than reward as a motive for visual 

 discrimination. In the present experiments it was found that, 

 with the more difficult problems, punishment as a motive had 



