336 On Random Occurrences in Space and Time 
Fig. V. Number of intervals. 
-2 - 1 0 1 
0- 
/ 
/ 
/ 
/ 
/ i 
I ; REGRESSION CURVES 
/ / TIME ON INTERVALS 
/ ,' N° OF INTERVALS ON TIME 
Below are given the values of the constants /S and ?/i found by the different 
methods; it should be remembered that the last two are not found from precisely 
the same material as the others, only a portion of the whole being used. 
Vario us methods of find ing /3 att d m. 
A. From duration of intervals in an indefinitely long time. 
/3 m 
(i) First Group /3 to -5 -370,552, 2 488,007, 
(ii) First Group /3 to 1 -362,918, 2-495,527. 
B. From frequency of occurrences in periods T. 
(iii) ^„ =('^-/^)" method '353,862, 2-580,641, 
\ /o / 
(iv) = '\) method... -393,772, 2-495,831. 
These values agree moderately well with each other and we may laii-ly safely 
predict that although their probable errors are at present unknown, they will be 
found small enough to make the methods sufficiently reliable for practical purposes. 
The tests of goodness of tit are reasonably satisfied and the experimental data may 
