8 D. B. CASTEEL 



according to the inclination of the animal. If the turtle was 

 eager for food it would usually swim rapidly into the runway, 

 crawl up the sandy "bank," making choice the while of the 

 box into which it would enter. If it entered the "right " box 

 it received no repulse but continued on to the back of the box, 

 found the meat hidden behind the cork blind, turned around, 

 scrambled out of the box and down the bank into the water. 

 Here it might stop to eat the meat or might continue on to the 

 entrance room, in which case the slide was immediately dropped 

 and the operator rearranged the boxes. Ifit remained in the runway 

 while it w^as eating the meat or afterward, it was urged to enter 

 the entrance room and the slide was dropped before another trial. 



If the turtle chose the " wrong " box it received an electric 

 shock at the entrance to the box immediately upon touching 

 the electrode upon the floor of the box with its fore feet or snout, 

 its hind feet being in contact \\ith the long electrode in front 

 of the boxes or with the wet sand. As a rule this caused it to 

 jump backward though it might or might not retire to the water. 

 If the turtle remained in the runway it was, as a rule, replaced 

 in the entrance room and the slide dropped before another 

 trial was given. However, at the beginning of its training, a 

 turtle would at times be allowed to remain in the runway until 

 it found the " right " box and obtained the food from it, though 

 the trial was counted as " wrong " if it at any time approached 

 the electric box sufficiently near to receive an electric shock. 



It was considered that a turtle had learned any particular 

 problem when it was able to make a perfect record of correct 

 choices in three successive tests of ten trials each. 



The relative position of food and electric boxes was changed 

 with irregular sequence during each series of trials so that learn- 

 ing by an appreciation of regular alternation might be avoided. 

 If no change of position was made between two trials the boxes 

 were lifted and replaced that lack of noise might not give a clue. 



Full records were kept during the progress of each experi- 

 ment showing the relative positions of the boxes at each trial, 

 the time consumed in making each choice, the success or failure 

 of the trial and miscellaneous data regarding the actions of 

 the individual throughout its experimental history. The descrip- 

 tions of the experiments and the appended tables are sum- 

 marized from these records. 



