210 KANSAS UNIVERSITY SCIENCE BULLETIN. 



from the totals of the individual daily records. The two sets 

 of results do not, therefore, agree perfectly, but whenever there 

 is any divergency, Table III6 should be considered the more 

 exact. In Table III6 the per cent of increase or decrease in the 

 second test is determined by dividing the per cents of omissions 

 and errors of the second test by the corresponding number of 

 the first test. 



Two of the subjects walked between tests for two to three 

 hours' traveling distances ranging from nine to fifteen miles, 

 as recorded by the pedometer. As these walks occurred but 

 once each week, and as neither subject was accustomed to 

 much walking, they were very fatiguing. The other six sub- 

 jects used the big spring ergograph already described. How- 

 ever, it had been found that the subjects suffered so much 

 from sore hands after the first day that they were unable to 

 lift sufficiently hard to render themselves very fatigued. Ac- 

 cordingly a harness was introduced, which consisted of a 

 padded strap over each shoulder and attached to tugs in front 

 and behind. The subject stood astride the ergograph and the 

 tugs were attached to the handles of the ergograph and short- 

 ened sufficiently that the subject could stand erect only when 

 he lifted his maximum amount. This meant that he was 

 always under strain, for when he was not lifting, he was 

 standing in an awkward position that was in itself fatiguing. 



ASSOCIATION TEST. 



In the association tests nonsense syllables were used, and 

 the method was the usual method of successes. The appara- 

 tus used to expose the syllables was the Wirth form made by 

 Zimmermann, slightly modified. Ten pairs of syllables usually 

 constituted a series. As there are fourteen spaces in the 

 cylinder of this apparatus, this left four blanks. The syllables 

 were always arranged so these blanks came together at the 

 end of the series. In the case of one subject, R., these blanks 

 were filled, making fourteen pairs of syllables. Four seconds, 

 counted by the metronome, were allowed for each space, in- 

 cluding the blanks. The number of repetitions of the series 

 was determined for each subject by preliminary tests. An 

 approximate average of seventy-five per cent for the normal 

 score was chosen, but in some cases it was found necessary to 

 accept a lower score, as some subjects required more time 

 than was available to reach the higher score. The number of 



