24 ROBERT M. YERKES 



since it can be relied upon to place and hold the doors in a 

 constant position. 



For the large apparatus, it is extremely desirable to devise 

 some type of lever mechanism which shall be easily manipu- 

 lated, reliable, and inexpensive. All of the mechanisms thus far 

 proposed are either too cumbersome or too expensive to be 

 feasible, but it is hoped that shortly a method may be discovered 

 by which the experimenter may conveniently and accurately 

 control the various doors by means of levers, the maximum 

 excursion of which shall not exceed eighteen inches. Since the 

 various doors must be raised a maximum of seventy-two inches, 

 it will probably be necessary to introduce one or more forms of 

 multiplying device. Already an automatic locking device, to be 

 operated in connection with the proposed system of levers, has 

 been designed. 



In the absence of a satisfactory scheme for the use of levers, 



weighted cords and locks, which are operated independently, 



may be employed. But this system of control mechanism, as 



has been stated above, is both unreliable and troublesome to 



operate because of the numerousness of the parts. There must 



be a separate weighted cord for each of the twenty-six doors 



and a separate lock mechanism for each of the twelve boxes, 



entrance and exit door in each case being controlled by the 



same lock. 



USE OF APPARATUS 



The use of the convertible reactive tendency apparatus in con- 

 nection with each of the three methods in question will now be 

 described. For all of the methods alike, rewards and punishments 

 may be used as inducements to effort. As rewards, food pre- 

 sented in the food cups, or for children small presents similarly 

 presented, serve well. In certain exceptional instances, it may 

 prove desirable to present the reward for a successful choice, not 

 in the food cup of the correct box, but instead at the entrance to 

 the release box. As punishment, it has proved feasible to use 

 confinement in incorrect boxes. It seems probable that for cer- 

 tain organisms the electric shock may prove useful. 



Hamilton Method 



For use with the Hamilton method of quadruple choices, the 

 following procedure is suggested. This method involves the use 

 of only four reaction mechanisms. Boxes 5, 6, 7 and 8 may 



