PSYCHOLOGICAL MEASUREMENTS. 567 



reduced-diet date, was at about the same level, and on the following 

 two dates the values are still lower; in fact, lower than for the first and 

 second experiments. 



With 1 vd. Squad A showed considerable improvement at the start. 

 Their first threshold was, however, undoubtedly too low, being only 48 

 per cent. Their maximum level was attained in the third, fourth, and 

 fifth sessions and maintained very evenly. Beginning with De- 

 cember 8 there was a marked depression, with some recovery on 

 January 26 and February 2, especially on the latter date, but it did 

 not reach the former level. Squad B shows a continuing low level for 

 the first three sessions. The expected improvement from practice 

 with this most difficult judgment of the series does not reach a maxi- 

 mum in the third test as do the other curves for this squad but, as 

 might naturally occur, is delayed until the fourth and fifth sessions. 

 The remaining points in the curve represent January 13, 19, and 27, the 

 food reduction period, when there was a considerable decrease in the 

 percentage of correct judgments with the 1 vd. interval. 



In this pitch discrimination test, the records of most interest are 

 those for the increments 3, 2, and 1 vd., which were near the threshold 

 value and thus required the greatest care and attention for correct 

 judgment. The average results for Squad B are fairly consistent in 

 demonstrating poorer discrimination during the low-diet period. This 

 is quite noticeable in the two smallest increments. For Squad A an 

 adequate base line is lacking but beginning with December 8, these 

 men on the average not only failed to improve over former results but 

 for the remainder of the research show a general decrement, which is 

 associated with the prolongation of the reduced diet regime. 



(3) DISCRIMINATION FOR SPECIFIED NUMBER GROUPS. 



As described on page 145, the subject was required to mark in each 

 line of 10 numbers every combination of 2 successive digits which added 

 together equaled 11. In correcting the records, three things were 

 noted: (1) the number of combinations found and correctly marked in 

 that part of the material which the subject was able to cover in 5 min- 

 utes; (2) the number passed over unmarked by the subject; and (3) the 

 number of incorrect combinations marked. It will be recalled that the 

 same amount of new material was provided the subjects at each test, 

 the quantity having been from the first experiment made large enough 

 so that no subject would be likely to complete the material efficiently 

 in the 5-minute period allowed. Each man began at the first of the 

 blank and proceeded as far as he could within the time Umit. In suc- 

 ceeding trials, as the men gained in proficiency and became thoroughly 

 familiar with the number combinations to be marked, they covered 

 more and more of the material provided. It is therefore clear that the 

 sum for combinations marked and combinations passed unmarked 

 (errors of omission) will not be the same from day to day but is a meas- 



