134 HELEN B. HUBBERT AND K. S. LASHLEY 



TABLE 1 

 The Average Number of Trials Required by the 56 Rats for the Elimina- 

 tion of Each of the 11 Errors Possible in the Watson Circular Maze. 

 The Numerals Indicate the Position of the Errors in Figure 1. 



TABLE 2 



The Number of Cases in Which Each of the Blind Alleys Was Not Ex- 

 plored Once Before the Maze-habit Was Thoroughly Established 



Errors of passing a door Errors of turning wrongly 

 Number Number of cases Number Number of cases 



The most striking fact brought out by this table is that errors 

 of type I are eliminated in less than two-thirds as many trials 

 as are those of type II. Many rats never made certain errors 

 and the distinction between the types is shown here also. Table 

 2 gives the number of cases in which each of the errors was not 

 made once during the course of training. Blind alleys, entrance 

 to which constitutes an error of type II remained unexplored 

 in only two cases. In contrast to this, blind alleys beyond the 

 doorways remained unexplored in 31 cases. 



What is the explanation of this fact? The first suggestion is 

 that there is a transfer of the reaction learned for one doorway 

 to other situations, but it may be also that a break in the smooth 

 walls of the alleys excites some instinctive mechanism which car- 

 ries the animals through the doorways. To test this latter pos- 

 sibility the records were examined to discover whether the 

 animals more frequently went through the first door that they 

 came to in their first trial in the maze or passed beyond it into 

 the blind alley, that is, whether or not error 2 was made at the 

 first possible opportunity. It was found that, of the 56 records 



