THE QUESTION OP FATIGUE FROM THE ECONOMIC STANDPOINT. 319 



tables are the stamp-pressing, where levers have to be worked as 

 well as material inserted and removed (Table VIII.), and complex to 

 a lesser extent is the machine sewing. In all these processes the 

 afternoon output compared with tJie morning is higher than the normal, 

 the average per afternoon hour compared with the average per morning 

 hour being 40,676 to 40,047 in the stamping-press work, and 111"72 

 to 97*1 in the machine-sewing at Kettering (Table IV.). In other 

 respects the curves differ, no doubt owing to the fact that stamping- 

 press work is highly uniform but not concentrative, while machine- 

 sewing is fauly variegated but tiring to the fingers (see above). In 

 explanation of increase of output during the day for both these 

 complex processes, it is interesting to note the following extract from 

 Weber : 



' Experiments show . . . that fatigue sets in more slowly in the 

 case of disturbed or complex work than where there is no disturbance ; 

 indeed, in the hours at the end of a spell the output is often very much 

 greater where work is disturbed than the normal increase which 

 results from practice would lead us to anticipate. These results, which 

 suggest that complex work is less fatiguing and more easily acquired 

 by practice, are only apparently paradoxical. . . . The paradox is 

 explained by the simple fact that the disturbed operation — that is to 

 say, in the case of complex work, each of the individual acts into which 

 it can be analysed — begins at a very much lower point in virtue of the 

 disturbance. Hence the output increases with continuous habitua- 

 tion and inner adaptation to the disturbance or to the other work very 

 much more quickly, until it reaches its maximum, than is the case with 

 uninterrupted operations. This is due to the fact that not only the 

 results of practice, but also the gradual adaptation to the disturbance, 

 make their influence felt and are only fully realised when the work is 

 practically at an end. ' 



(c) Of great uniformity or repetitiveness is the work with the metal 

 stamping-press, and of this particular process we are lucky in having 

 both accident and output time-distribution. Covering chocolate by 

 apparatus is also unifonn in that the same rack is always handled 

 instead of the individual chocolate creams, as in hand-covering. This 

 process, however, is not general enough to yield sufficient accidents for 

 useful statistical comparisons. 



Turning first to the output distribution in these uniform processes 

 given in Tables VIII. and II., we find that in both cases ' the curve ' 

 differs from the normal in increasing right throughout the spell, 

 except, in each case again, for the last hour but one of the afternoon 

 spell. In the last half-hour of the stamp-pressing spells the increase 

 is sudden enough to be attributable to ' spurt.' 



The two processes differ, however, in that the output is less in 

 afternoon than morning with the apparatus-covering, but greater with 

 the stamping-press. 



The explanation of the specially high aftei'noon output of stamping- 

 presses was given under complexity, but the increase of output during 

 the spell in both these uniform processes is due probably to the power 



