G. MORANT 
837 
Fig. VI, Lengths ot intervals in periods of 10 seeouds when there are 
3 intervals in each period. 
300n 
250- 
200- 
i '50 
100- 
50- 
/?=-354 
m = 2-581 
OBSERVED N=663 
CALCULATED N= 590-2 
CALCULATED N- 663 
4 5 6 
Lentjlli of iiitenyih 
8 
— I 
10 
be held to confirm the above theory of random occurrences, if it is supposed that 
the discrepancies observed in dealing with short intervals in the neighbourhood of 
^ are due to fluctuations throughout the experiment of the least interval between 
occurrences. More satisfactoiy evidence of the value of the theory might be given 
if experimental or observational data could be found for which /S was moi-e 
adequately constant. 
I am indebted to Professor Karl Pearson for suggestion and assistance especially 
in the algebraic portion of this paper and to Miss Ida McLearn for the preparation 
of the diaorams. 
