358 K. S. LASHLEY 



The second group (B) received similar training except that the 

 animals were allowed to explore the entire maze, without food, 

 for 20 minutes on the day before training was begun. 



The third group (C) was treated like the first except that the 

 animals were not allowed to correct their errors but were returned 

 immediately to the starting compartment if they entered the cul 

 de sac or turned back along the correct path. Each time that 

 they were returned to the starting compartment a trial was re- 

 corded, whether or not they had reached the food. 



The fourth group (D) was trained like the first except that a 

 dish of food, covered with fine wire netting so that the food could 

 not be reached, was placed in the end of the cul de sac. 



Each of these groups was subdivided into two, one of which 

 received ten, the other two trials per day. A summary of the 

 groups is given in table 1. 



TABLE 1 

 Conditions of training for animals tested in the simple maze 



Training was continued until the animals made ten successive 

 trials without error. In some cases training was continued be- 

 yond this point for as much as 200 trials. None of the animals 

 so trained made more than 5 per cent of errors in these later 

 trials, so that the criterion may be looked upon as indicating 

 practically perfect learning. 



The average numbers of trials per day required by the different 

 groups are given in table 2. Under all the various conditions of 

 training the animals which were trained with only two trials per 

 day learned more rapidly, per unit of practice, than those which 

 were trained with ten trials per day. The average difference for 

 all the groups resulting from the different distributions of practice 



