ON THE AFTER-SHOCKS OF EARTHQUAKES. 



119 



Now n' = 493 repvcsent.s the linlf-day interval wliich is equally 

 distant from the heiiinning and end of tlie year 1892, and .t = 1225 

 that of the year 1898. The values of y/ corresponding to tliese x as 

 calculated from e(|uation (c) are respectively 0.9 and 0.86. In tlie 

 actaal case, there were (S67 and 808 shocks at Gifu daring' tlie alcove 

 two years, giving the average half-daily nnnihers of 1.2 and 0.42.^ 



It is interesting th;it equation (f), deduced from the ohservations of 

 only a few da^'s immediately succeeding the initial earthquake, thu.^ 

 represents with tî)lerable accuracy the frequency of eartlKpiakes a 

 year ov two later on. 



In the curve re[)resented 1)y equation (l>), the point of the maxi- 

 mum curvature occurs at .t — VÂ:— /'. Substituting the values of the 

 constants ?j and // adopted in equation (c), we find .t= 19, denoting 

 an epoch of time about ten days after the great earthquake (see § 9). 



Again, b}' taking the monthly activities for the successive eighteen 



1 Assuming equation (a) and using the same data as in deducing equation (c), we 

 olitain 



log y = 2.25-.rx0.10G, 



which gives satisfactory results when .c is small, (thus, when .r = 0, )/ = 175; when .r=l, j/ = 13S; 

 when .(' = 2, ij = 103, etc.), but fails when .v becomes great. 



