THE QUESTION OF FATIGUE PROM THE ECONOMIC STANDPOINT. 311 



the accidents of one year in Illinois (Table XIII. Col. 7), of which 246 

 happened to occur in the fourth hour of the morning spell, with 257 

 in the fifth ; while 227 occurred in the third and 260 in the fourth (and 

 probably last) hour of the afternoon spell ; Imbert because he was too 

 intent on fatigue and failed to notice considerable drops at 11 a.m., at 

 any rate, in some of the very curves he gives in his diagrams (Revue 

 Scientifique, September 24, 1904, page 388) and Table XV.d. 



In the very last half-hour of the morning spell, it is interesting to 

 note a rise of the rate of accidents over that of the preceding hour in 

 almost every case where this period is separately recorded. In the 

 English cotton industry this rise is from 623 to 651 in the morning 

 and 512 to 615 in the afternoon (Table IX. a), and in the morning spell 

 in Massachusetts (Table XI. a and b) while the pronounced rise for 

 ' Total Industry ' may be due to some industries diverging from the 

 normal lunch hour, yet there are rises of accident in the last half- 

 hour of the morning in the Building, AVool, Eubber, Paper, Printing, 

 and Autocar Industries, whose midday hours are fairly normal. 



Alluding to their accident statistics of metal-work, cotton (in 

 U.S.A.), and Indiana and Wisconsin manufactures reproduced here 

 respectively in Diagrams I., III., and IV., the U.S. Federal Eeport 

 (page 97) speaks of the ' remarkable accord from group to group ' : 

 ' Here are four sets of figures collected by different agencies in differ- 

 ent parts of the Union at different times and covering different in- 

 dustries, each agency v/orking independently of the others. Yet the 

 figures thus gathered show striking similarity.' Yet these particular 

 figures, as can be seen in the tables, are among the more irregular 

 that we have recorded; and compared, for instance, with the Massa- 

 chusetts, Illinois, Ohio, and individual factory accidents, differ among 

 themselves quite considerably. In fact our figures agree with one 

 another to such an extent, particularly those of accidents, that we are 

 justified in speaking of a ' normal ' time-distribution of output and of 

 accidents, or considered inversely, accident immunity. The shape of 

 the output and accident-immunity curves for a five-hour spell may for 

 purposes of illustration be summarised as follows : 



Hour of Spell Output Accident Immunity 



1st small very great 



2nd ..... very great great 



3rd great fair 



4th fair *small 



5th *small fair 



* Where there are only four hours in the spell, strike out the last output, but the 

 fourth accident hour. 



In seeking on explanation of this ' normal ' time-distribution of the 

 accident rate and the output in a spell of manufacturing work, let us 

 concentrate on the illustrative table. Here we find the four same 

 degrees : very great, great, fair and small, succeeding one another in 

 both the output and the accident-immunity column, though earlier in 

 the spell with accidents than with output. Now both output and 

 accident immunity vary inversely to fatigue ; these four decreasing 

 degrees, therefore, may well be measuring an increase in fatigue. 



