326 K. S. LAS 1 1 LEY 



not attend to differences in form. Further data, obtained from 

 rats 4 and 6 when offered a choice of two new forms of unequal 

 size, support this view. 



EXPERIMENT 11 



Since the discrimination box places the animals under some- 

 what unnatural conditions a few tests were undertaken under 

 an environment more nearly approaching the normal. 



(a) Two cardboard food boxes were constructed, the one 

 rectangular, the other pyramidal in form and having small 

 openings in the sides to admit the rats. The exposed faces of 

 the boxes were of equal area (128 square centimetres) but, 

 owing to the sloping surface of the pyramid, they differed some- 

 what in brightness. Food was given in only one of the boxes 

 and they were interchanged irregularly to avoid position asso- 

 ciations. 



Of the three rats required to choose between the boxes not 

 one showed any evidence of discrimination after 150 trials. All 

 formed position and rhythm associations but did not distinguish 

 between the boxes in any way. 



(b) Rat No. 1 was trained to go to a large square of flashed 

 glass supported in an upright position upon the floor of the 

 laboratory. After a few trials she learned to run directly to 

 the square from a distance of two or more feet, passing behind 

 the glass to get a bit of food. When this association was estab- 

 lished a small piece of glass, one centimetre square, was placed 

 beside the larger (12 by 12 centimetres). The rat sniffed at 

 the small glass only once, then gave it no further attention. 

 When a plate of 25 square centimetres area was substituted 

 for the smaller square, the rat became confused and failed to 

 choose the larger. After 40 trials, however, she began to avoid 

 the small square and made a record of 17 correct choices in 

 25 trials. A rectangle of half the area of the larger square was 

 then used as the negative stimulus. In the first 20 trials the 

 rat made only five per cent error, then became frightened and 

 in 200 trials since has shown no evidence of discrimination. 



From this experiment it appears that the rat can distinguish 

 great differences in size without difficulty. With slighter differ- 

 ences, however, the dependence upon kinaesthetic sensations 

 appears, just as in the discrimination box, and learning does 

 not seem to be any more rapid. The objection to the unnatural- 



