20 



STELLA BURNHAM VINCENT 



observed differences in the general numerical results of time 

 and errors, as shown in the summary. (See fig. 2.) 



Summary of the results of Experiment IL 

 Average number of errors in the first twenty-five trials. .35. 



Average number of errors in the first trial 9.5 



Average number of trials in learning 16.5 



Average time of the last ten trials in learning 11. 8 sec. 



Average time of the last trial 8.8 



Average slips and falls in the first twenty-five trials 2.5 



/ 23^5678'^ /O 

 ' Fig. 2 — Learning curves, time and error, of normal rats on maze without sides. 



If the figures for this experiment are compared with those for 

 Ex. I, it will be seen that the average number of errors for 

 the first twenty-five trials is somewhat greater in the maze with 

 sides than in the maze without, and markedly greater in the first 

 trial. In spite of this fact the time of learning is the same and 

 the average speed is practically the same. That something is 

 influencing the different distribution of errors in this maze is 



