270 ADA W. YERKES 



The correct path is shown by the solid line in the figure.' There 

 are eleven blind alleys which may be entered either by going 

 straight ahead past an opening or by turning in the wrong 

 direction after passing through an opening. Entering any one 

 of these alleys was counted an error; also turning back so that 

 the animal was running in the wrong direction in any section 

 of the maze, i.e., between successive openings, was counted an 

 error for each section of the maze through which it ran. The 

 dotted line in figure 1 shows a path in which three such errors 

 occur. This definition of an error is in accordance with Mrs. 

 Hicks' two propositions, namely, that errors shall include all 

 total and partial returns, and that a runway, the distance be- 

 tween two corners, or its equivalent in the circular maze, the 

 distance between two openings, shall be taken as the unit of 

 error. 6 



After a number of animals had been tested in this form of 

 maze, it was decided to simplify the maze and shorten the path 

 to the centre by cutting out the outer circles. The inner wall 

 of the outer circle was shifted so that the opening faced the 

 entrance to the maze. Blocks were then placed on either side 

 of the opening so that a straight passageway was formed from 

 the entrance through to the third circle. The correct turns 

 were then L, R, L, L, R. This form is hereafter referred to as 

 the short maze. That this change made the problem much 

 easier is shown by the fewer trials required to learn it correctly, 

 as indicated in table 1, where the stock rats (average age about 

 66 days) in the long maze required an average of 55.25 trials; 

 the inbred, 65 trials; in the short maze (age 35 days), stock, 

 19.2 trials; inbred, 23.75. 



The attachment for recording the maze path consists, briefly, 

 of a mirror attached to the ceiling; this mirror reflects the image 

 of the maze through a lens to a drawing board, enveloped in 

 a hood of black cloth, and holding a pad of paper on each sheet 

 of which had been printed a diagram of the maze. This dia- 

 gram exactly fitted the image of the maze received from the 

 lens. The sheets of paper could be easily and quickly removed 

 with practically no interruption of the tracing and thus the 

 course of a rat in its devious wanderings could be readily fol- 



6 Hicks, V. C. The relative values of different curves of learning. Jour. An. 

 Behavior, 1911, 1, 138-156. 



