134 



F. OMORI. 



earthquakes. The hours at v^'hich these occur seem to be ditterent for 

 different localities and theref«^re these mny not each he shewn, 

 ■\vhen we mix up earth()uake records from distant places of the world 

 too-ether. AVhether there are more earth([uakes durinij;' the nig'ht than 

 diu'ing the day is not certain, as may be inferi-ed from the mean curves 

 drawn in red. 



The 4-hourly and (S-(or 1>) hourly periods indicated in the curves 

 of hourly earthquake frecpiency (Figs. 10, 11, and 1'2) are evidently 

 due to the nhove diurnal fluctuation. We h:ive not at present suf- 

 ficient data to determine whether other three constant periods of 4^, 

 12, and 3o days occur likewise in the frequenc}' of ordinary earth- 

 quakes. 



With reg;u'd to tlie annual fluctuation, it is to b:' noted that :dl 

 the maxima and minima of the monthly after-shock frequency for 

 Kumamoto (§ 2(J), with the tAvo exceptions next mentioned, occurred 

 in exactly, or nearly, the same months as the maxima and minima 

 of fre(|uencv for nil fTapan (§ 27. ('Aj ). Only the second and third 

 minima at Kum;imoto (§ 20) took ])lace res])ectively in February and 

 ]Sroveml)er, (LSilO), which are generally montjis of maximum earth- 

 (juake frequency. These mny denote fluctuations due to the " proper" 

 period of the Kumamoto after-shocks. 



Again, the four maxima (^f the mi^nthlv fre(juency i\)V Gifii 

 (§21) occurred in A|)ril and September, which are gener:dly uKMiths 

 of minimum e'irth((uake frequencv. These ni;iy denote the fluctua- 

 tions due to the "proper" period of the Mi no-Owari after-shocks. 

 Of the four corresponding minima, three occurred in months of 

 minimum frequency, and one in a month of maximum. 



As already remarked, strong shocks seem to have a tendency to 

 happen when the frecjuencv is going to rc-u-h a maxinuuu, and it 

 is interestino- to find thnt of the four severest ai'lcr-slincl.-s fthree of the 



