346 



J. F. DASHIELL 



mazes 9 and 10, the same alleys (in the same mazes) do not show 

 anything like a similar number of re-entrances. 



Remembering the different sorts of preliminary training or 

 " practice series," we may now consider the formal experiments 

 or "test series." All four groups of rats were run for this crucial 

 experiment hi the maze Q (fig. 6). In the first section of the 

 present paper the results of the test series with the first supply of 



ric v RE. 



/o 



3 to k io is 



m 



IIS 



rats used were given only hi part, the learning of the rats being 

 disturbed after the 8th trial by an extraneous factor unforeseen 

 (fig. 7). As amplification of that record the new series with the 

 second set of rats has great value. In graphic form the record 

 appears in figure 10. Here we have gradations between all the 

 four curves more definitely shown than in figure 7. Group I in 

 this case shows substantial inferiority to group II in its initial 

 attack upon Q; though in later trials the differences are insignifi- 



