26 

 norm taken foiv the learning of the maze was an average 



of six seconds in three successive days , and f in addition 

 to this criterion, the runs had to he relatively free of 

 error. No definite standard was taken for an error, but 

 one was counted when a rat turned in the direction of a 

 cul de sac or retraced its steps. Mere hesitation at 

 any point was not considered an error. V/hen error is 

 adopted as an additional criterion it tends to become, 

 in later trials, the sole standard, for the runs in the 

 maze were made at one time without error in sis seconds, 

 and again with several errors in the same time. This is 

 not because the standard of time is too high, for some rats 

 can barely make an error-free run in six seconds. An active 

 rat, on the other hand, may make the run in less than six 

 seconds and have several errors to its credit. The norm 

 established in this experiment was a run attained in six 

 seconds and relatively free from error. 



That it becomes necessary to take into account errors 



