708 ROGER T. DAVIS AND ARNOLD A. McDOWELL 



Measurement of Performance 



Object-quality Discrimination 



Preliminary training 



Animals were trained 60 days on 4-trial object-quality discrimination 

 problems in the WGTA. Twelve problems were given to each animal on 

 each day. Discrimination stimuli consisted of pairs of commonly used, 

 randomly assorted dissimilar objects. Each problem required the selection of 

 one of the two objects which shifted in left and right positions in a prede- 

 termined random order. One object was arbitrarily correct and was placed 

 over a foodwell containing a raisin; the other object had no reward. A new 

 pair of objects was employed on every four trials. 



The learning during preliminary training was comparable to that previ- 

 ously reported for subjects given training on 4-trial object-quality dis- 

 criminations. 



Radiation schedule training 



Each subject received six 4-trial discrimination problems a day on each 

 of the 12 days prior to the first irradiation and during each of the 27 days 

 of training following the first and second periods of irradiation. The dis- 

 crimination problems given after irradiation differed in one important 

 respect from those presented prior to irradiation. Half of the problems given 

 each day consisted of planometric patterns composed of three pairs of pic- 

 tures selected at random from magazines and pasted on 2 X 2 in. fiberboard 

 squares. The other three pairs were similar to those used daily prior to 

 irradiation. 



A detailed analysis of the data, including a breakdown of the various 

 categories of errors outlined by Harlow (1950), indicated no significant 

 trends between groups. One significant trend within a group was a decrease 

 in diflferential-cue errors made by the P group on trial 4 of problems presented 

 following the first radiation. Three parallel analyses, each involving 36 com- 

 parisons, were made between the performance of subjects in a particular 

 group on trials 2, 3, and 4 in the preradiation and two postradiation periods, 

 and the occurrence of one significant difference is less than would be 

 expected by chance. 



Bent Wire Detour Problem 



Bent wire problems occur in various patterns, from a simple straight seg- 

 ment of 0.187 in. welding rod 3.0 in. long to four segments separated by 90° 



