ON THE AFTER-SHOCKS OF EARTHQUAKES. 13][ 



The curve of the cliurnal earthquake fluctuation for Gifu (Fi.ir. 11), 

 (1)) indicates distinctly tlirce ma.rima, which occur respectively 

 between 4 and 5 a.m., between 11 a.m. and noon, and between (> and 7 

 p.m.. and tliree iiiiiiiiiui which occur respecti\'ely ])etween 1) and 10 

 a.m., between o and 4 p.m., and at 11 p.m. The intervals between 

 successive maxima are 7, 8, and *J lioiu's, and those between successive 

 minima 6, S, and 10 hours. 



The corresponding curve for Xag'oya (Fig. 11*. ('2) ) indicates six 

 more or less distinct oscillations, giving a mean period of 4 hours. 



It is evident (see § 21) that both the 4-liour and the 8-hour 

 periods existed together, but that the longer period predominated in 

 the diurnal eartlujuake frequency for Gifu and the shorter one in that 

 for Nagoya. 



The curves drawn in red (Fig. 11). (1) and (2)) seem to indicate a 

 very slight diurnal variation of the mean frequencv, the minimum 

 occurring between 1 and '2 p.m., and the maximum at about 1 a.m. 



The hourly distril)utions of 148 shocks at Kumamoto durinir 

 fourteen days, from July olst to Aus^ust lotli, 1881), and of 233 

 shocks as Chiran during an equal time interval, from 8th to 21st, 

 September, 1893, (see Tables II [ and X\'I), are, as shewn in Fig. 20, 

 very similar respectively to those for Gifu and Xagoya. 



§ 2d. The Annual Fliicinalion of tlte After-shock Freiaencij. 



Curves (1). (2), and (3), Fig. 21, shew the annual fluctuation of 

 after-shock frequency at Kumamoto averaged respecti^'eIy f )r four 

 years (1890-181)3), three years (1891-1893), and two years (1892- 

 1893). The first curve indicates ilivee inaxima which occur respec- 

 tively in M:u*cli, May, and October; and tliree minima which occur 

 respectively iu April, between August and September, and in Decem- 

 ber. The curves drawn in red indicate the annual variation of the 

 mean frequency. 



