38 RUTLEDGE T. WILTBANK 



dence for this fact being furnished by the large saving 

 in the group's time of learning the second maze, as com- 

 pared with the time of the control-group. In addition 

 to this, the association between running the maze and the 

 procurement of food is firmly fixed at the beginning of 

 the maze-learning in the case of the rat that has learned 

 a preceding maze, while this connection has to be formed 

 by the inexperienced group in the course of the learning. 

 Still again, and perhaps most important of all, is the prac- 

 tise in error-elimination. This does not secure the animal 

 against entering blind alleys nor forming habits of doing 

 so, neither does it assist the animal in finding the true 

 pathway; but it enables the animal that has had this 

 practise to eliminate more promptly its entrances into 

 blind alleys. 



The Correlations among the Records of Transfer through a 

 Series of Mazes 



It was found that a positive correlation existed among 

 the savings in trials and errors, in trials and time, and in 

 errors and time, effected by the various groups as they 

 learned the various series of mazes; the savings being 

 estimated on the basis of the average number of trials, 

 the average total errors and the average total time per rat. 

 A positive correlation existed among the savings when 

 they were estimated on the basis of the average errors 

 and the average time per trial, except in the case of that 

 between trials and errors in the third group and that 

 between trials and time, and errors and time, in the fifth 

 group, in which cases it was negative. On this basis of 

 estimation there were also two .cases of zero correlation, 

 those of errors and time in the second and in the third 

 groups. The method of rank-differences was used in de- 

 termining these correlations, and the two sets of figures are 

 given in tables 31 and 32. 



