364 



H. M. JOHNSON 



ference once in a small number of repetitions of the experiment. 

 Two subjects showed slight differences in favor of the darkness, 

 the differences being statistically attributable to chance; while 

 one subject, the astronomer, showed a difference in favor of 

 darkness which would justify odds of 100 to 1 that repetition 

 of the experiment would not result in a reversal of the direction 

 of the difference. 



TABLE 2 



Summary of differences between the performance of the several subjects in light and 



in darkness 



Explanation: The difference in performance in light and in darkness for any 

 particular day is given in per cent performance in darkness, and indicated by 

 the symbol A. In other words: 



100(P L -P D ) . 



A = , in which 



"D 



P = number cards correctly sorted per 1000 seconds; 

 P L = Performance in light, and P D = performance in darkness, as thus 

 measured; while M indicates the mean, and PE the probable 

 error, of the determinations indicated by the subscripts. 



