DOCKERAY : PHYSICAL FATIGUE AND MENTAL EFFICIENCY. 237 



The sounder test could not have been divided into smaller 

 equal periods, for the four sounders presented unequal diffi- 

 culties for discrimination. 



The chief characteristic that is brought out by the subdivi- 

 sion of the addition and multiplication tests is the almost total 

 lack of any constant tendency, either in the speed or accuracy 

 of the processes. In Table la one may find instances in the 

 first tests that indicate that Ba. decreases speed after the 

 first period and increases accuracy, but there are other in- 

 stances in which he shows no such tendency. This is illus- 

 trated by the tests before rest and before running eighteen 

 laps. In the second tests these series show the reverse ar- 

 rangement. Co. (Table Ha) manifests a constant speed in 

 all averages of tests preceding the interval, and a fairly con- 

 stant speed following the interval, whether it is after rest 

 or physical work; but there is no uniformity in accuracy in 

 either the first or second tests. Even the subjects that did 

 more poorly in the second tests on work days do not show 

 definite signs of fatigue in the first tests. In the results with 

 the sounder tests (Table Ilia) there is a more definite tend- 

 ency observable for some subjects. Bt. improves throughout 

 the first tests, though his average in the second quarter is 

 better than in the third in the series before rest. In the 

 second tests his accuracy decreases until the fourth period. 

 De. improves regularly in the first tests and in the first three 

 quarters after rest, but the fourth quarter is lower than the 

 first. After fatigue he decreases gradually in efficiency until 

 the fourth quarter and then manifests a decided improvement. 

 This might be due to a realization of the fact that he had 

 been doing poorly and that the end must be near. All the 

 subjects realized approximately the progress of the multiplica- 

 tion test. The other subjects, however, showed less definite 

 tendencies, except that W. always did better in the first quar- 

 ter of the second test and gradually decreased in the following 

 quarters. It seems probable, therefore, that if such phe- 

 nomena as the "antrieb" and the "anregung" do exist, they 

 are, as Thorndike (39) finds them, purely occasional occur- 

 rences which lack any great importance. In general, a single 

 test may be selected from any subject's series that illustrates 

 any theory desired, but the average of several tests represents 

 a blur. As was indicated earlier, the mean variation, if given, 

 would be very large. 



