302 REPORTS ON THE STATE OP SCIENCE. — 1915. 



which fatigue would act here especially would be a weakened activity in 

 judgment and memory. At any rate, it seems likely that the output 

 and the accident time distribution do not test fatigue at quite the 

 same points. It is interesting also to note that if it were possible to 

 record a time distribution of the quality of work (as measured inversely 

 by the number of errors each hour) it would follow the accident rather 

 than the quantity of work distribution in measuring fatigue indirectly 

 and especially through the channel of attention, muscular control, and 

 judgment and memory. The importance of this suggestion is that 

 since such time distributions of errors are very difficult to record, the 

 accident distribution may be taken as an equivalent test of the effect of 

 fatigue on quality. 



This whole discussion may be summarised, and here and there 

 expanded in the following table : 



, J . f decreases quantity of output ; 



Fatigue \ „°^^ j \ increases accident immunity 

 [ bpeea ^ ^^^ quality slightly. 



/_ ,,. .. r Decrease accident 



Less Attention, 



Less Attention immunity and 



Less Muscular Control | ^^^^.^^ ^^^^^^ 



■p, , . f Loss of j Less Sensitivity, f Decrease accident 



*tigu6 I Intensity I Longer Reaction 1 immunity mainly. 



„, , ,, f Decrease quality, 



Weaker Memory, ^^^.^^^^ immunity, and 



\Less Judgment [ output equally. 



Section IV. — Do the Tests Vary only with Fatigue? 



All the possible factors have been enumerated which besides previous 

 duration of work might be causing fatigue. It must now be pointed 

 out that besides fatigue there may be other factors causing variations in 

 the output and accident distribution. Though Imbert shows that out- 

 put and accidents may be expected to vary with the degree of fatigue, 

 he cannot show that either I. output and accidents, II. the output 

 only, or III. accidents only, do not also vary in amount with other 

 factors besides fatigue. 



I. All the factors that may influence both output and accident 

 distribution ' immediately ' without the interposition of fatigue may be 

 considered in two groups : 



A. Those factors which like fatigue itself are psycho-physical states in 



the worker. 



B. Those factors which are some outside condition or fact. 



A. For the full discussion of all the ' psycho -physical ' component 

 states which together with fatigue influence the ' work curve, ' 

 reference must be made to the translation of Weber's ' Psychophysik ' 

 (Index D8). Of these ' states ' the most important for short working 

 periods (the spell and the day) are ' Practice,' ' Spurt,' and 'Incite- 

 ment. ' 



In ' practice ' Weber sees the process which leaves ' the conscious 

 will 'and the power of attention ' free for application elsewhere, and 



