PERFORMANCE AFTER IRRADIATION TO THE HEAD 709 



bends in the wire. The patterns are constructed in pairs and grasped in the 

 center by a vise. A reward, a candy Life Saver or a paper poker chip 

 punched through the center, is threaded onto the pattern, and the subject 

 solves the problem by removing the reward. 



Preliminary training 



Subjects were trained for 16 days to remove poker chips and Life Savers 

 from simple one segment patterns and for 19 additional days on variations 

 of the one segment problem (Davis, 1958b, experiment III). 



Radiation schedule trainiiig 



Each subject was trained with five bent wire problems during each of the 

 66 days of training before, between, and after irradiation. Pairs of two seg- 

 ment bent-wire problems were employed in each trial, with two candy Life 

 Savers as rewards, one on either side of the jaws of the vise. Patterns with 

 two segments, one paralleling the bars of the restraining cage, the other 

 extending toward the subject, were used during the 12 days of training prior 

 to irradiation. During the period between the first and second irradiation 

 and following the second irradiation, the second segment extended away 

 from the subject rather than toward it (Davis, 1958a; Davis et al., 1958). 



Subjects in the four groups were less well matched on detour problems 

 than on any other problems. During the 12 day preirradiation period, the P 

 group failed significantly more often than the control group, p < .05 using 

 the U-test. For this reason, we regarded comparisons between groups as 

 invalid and only compared performance within particular groups between 

 periods of training. 



The change in difficulty of patterns after the first irradiation produced 

 significantly more failures in the control group. This trend was not signifi- 

 cant in the irradiated groups. Performance of the control group became 

 significantly worse after the second irradiation than during the intrairradia- 

 tion period. The irradiated subjects followed a similar, but not significant, 

 trend. 



Patterned String Tests 



Patterned string tests are among the oldest and most frequently employed 

 tests of "animal intelligence" (Harlow, 1934). They consist of two or more 

 strings or chains arranged in patterns. Subjects must select the arbitrarily 

 correct string and draw it in to obtain the reward attached to the end. 



Preliminary training was not given prior to the 12 days preceding irradia- 

 tion. 



