PSYCHOLOGICAL MEASUREMENTS. 565 



or slightly below. Gar also shows 77 per cent at 1 vd. His record, 

 however, is not so consistent, since he showed 77 per cent also for 2 vd. 

 Other average records which are particularly good are with subjects 

 Pea, Pec, and Bro, all of whom have thresholds below 2 vd, Kon, 

 Tom, and Vea show almost exactly 75 per cent on 2 vd. The values 

 for Mon, since they frequently show less than 50 per cent correct 

 judgments, were in the case of pitch discrimination not included in the 

 averages which are for the other 9 regular subjects and appear at the 

 right of the table. The total average for all records, excluding Sep- 

 tember 29 (normal), shows 95, 88, 77, and 67 per cent correct judg- 

 ment for pitch-difference intervals 5, 3, 2, and 1 vd., respectively. 

 The average threshold is evidently very close to 2 vd., which corre- 

 sponds absolutely with the modal threshold found by Seashore from 

 records on about 800 undergraduate college students, including both 

 men and women. 



The data in table 150 for Squad B show no marked peculiarities 

 which differentiate them from those for Squad A. There was con- 

 siderable improvement between the first and second dates, October 

 6 and November 3, although these were separated by one month. 

 Individual averages for the 5 normal periods are on the whole not quite 

 so high as those shown for the 9 periods for Squad A. As many of 

 the values for Van are below 50 per cent, they have not been included 

 with the others in the group averages. The averages for the re- 

 maining 9 regular subjects show a total for the normal period of 90, 

 82, 72, and 62 per cent correct judgment for intervals 5, 3, 2, and 1 vd., 

 respectively. Again, 2 vd. appears to be very near the normal thres- 

 hold for this group. 



The total comparable averages for intervals 5, 3, 2, and 1 vd. for 

 the two squads on the different dates are most conveniently compared 

 in diagrammatic form, as in figure 103. The 8 curves give a some- 

 what confusing appearance to the figure. It will, however, be seen 

 that the results for the different increments are for a particular squad 

 at different levels. The four curves for Squad B do not touch or cross 

 each other at any point, showing consistently the increasing difficulty 

 of increments 5 to 1 vd., since on each date the percentage of correct 

 judgments regularly decreases. The same is true for the four curves 

 of Squad A, with the single exception of the curves for 3 and 2 vd. 

 on September 29, which begin at almost the same point. The confused 

 appearance is therefore due to the fact that the two squads in general 

 have so nearly the same level for the different pitch-discrimination 

 increments. 



In the case of 5 vd. the two curves ascend about equally between the 

 first and second experiments and thereafter maintain approximately 

 the same level, which is very near 95 per cent. Squad B shows some 

 depression on December 15, due to the unaccountably poor records for 



