THE ALTERNATION PROBLEM 377 



by letters. The value for each point of the curve represents 

 the percentage of correct choices for the group out of 224 

 trials. 



The following conclusions may be derived from the results 

 of these control tests. 1. The introduction of the novel con- 

 ditions decreased the number of correct choices for the group 

 by 10%. 2. The alterations did not disturb two of the eight 

 animals. The percentage of correct choices of the rat manifest- 

 ing the greatest disturbance was lowered from 91 to 75%. No 

 animal fell below a record of 75%. 3. The most disturbing 

 conditions were those in which the animals were handled by 

 strange hands and in which they were subjected to a new body 

 orientation in starting. 4. The rats quickly adapt to these 

 novel conditions. This fact is evident from an inspection of 

 the graph. 5. The interpolation of these novel conditions in- 

 terfered little, if any, with the progressive perfection of the 

 two habits. At the beginning of the tests the animals had just 

 attained a consistent group average of 85% of correct choices. 

 At the end of the tests a record of 95% was secured. An im- 

 provement of 10% was thus attained during the period in which 

 the tests were given. The perfection of the two habits during 

 this period relative to the progress attained during the previous 

 learning period is represented by the solid line graph of fig. 2. 

 The curve up to the point A represents the progress attained 

 during the learning period. The part of the curve between A 

 and B represents the records secured from the tests for normal 

 conditions which were interpolated among the various control 

 experiments. The rate of progress during the control period 

 is somewhat less than that obtaining for the period of learning. 

 It is impossible to assert, however, that this decreased rate of 

 learning is due to the introduction of the controls. 6. As pre- 

 viously noted the animals experienced difficulty in mastering 

 the initial choice for each day's trials. This fact indicates that 

 the animals were not relying exclusively upon sensory data 

 derived from the mode of handling or the position in which 

 they were placed in the apparatus. If such stimuli were effica- 

 cious, the first choice should have been no more difficult than 

 the subsequent ones. 



The above results prove rather conclusively that the animals 

 did not rely exclusively upon the second class of stimuli. Neither 



