DEVELOPMENT OF SEISMOLOGY IN JAPAN 261 



(d) Secular variation. At least with regard to destructive 

 earthquakes the materials are sufficiently rich for studying the secular 

 variation of frequency. Several attempts have been ma;'.e to establish 

 the ' period ', or more properly the cycle of earthquake occurrences^^'. 

 As for an instance, the sequence of destructive earthquakes in the 

 districts near the Province of Sagami has shown a remarkable succes- 

 sion of nearly equal intervals^-^ Unfortunately, no reliable criterion 

 has yet been established to discriminate the true pi3riodicity with a 

 definite physical significance from the apparent ones which may be 

 brought forth by more or less arbitrary choice of the points distributed 

 irregularly on the line of chronology^"\ The fact that the destructive 

 earthquakes seem to occur in groups is also a statistical truth, though 

 a similar grouping may happen also in the cases of pui-oly arbitrary 

 distribution. In this connection, Saem. Nakamura's method of invest- 

 igation may l)e cited''"'', though his paper refers to a much shorter 

 duration of time. He studied the frequencies of the days with 0, 1, 



2, 9 earthquakes recorded instrumentally, with reference 



to the data observed in Tokyo during four years. On com[)aring the 

 results with the theoretical expectations corresi)onding to the random 

 distribution, it was found that the number of the days with no shock 

 was less than the theoretical value, while those with one or two 

 shocks were above the expectation. On the other hand, T. Terada^'' 

 suggested another method of statistical investigation regarding the 

 time distribution of earthquakes which may in some case be utilized 

 for getting some insight into the physical cause governing the time 

 distribution. At any rate, it seems desirable to elaborate the method 

 of statistics in this direction, before we may arrive at any definite 

 conclusion on the much discussed periodicit}' of earthquakes. 



(e) Other periodic variations. Onioi-i'^''' remarlced tliat with 

 regard to the aftershocks of some destructive eartliquakes, there exist 



(1) Omori, Ho., 26 (1899); 57 (1906); 68E (1910); 88C (1920). Imamura, Ho., 

 77 (1913); 95 (1922). 



(2) Among others, M. Matsuyaina, Comptes Rendus des Seance de la 2'" Conference, 

 Madrid, 1924, Union geodesique et geophysique, Section de Ssismologie, 72. 



(3) See Suda's discussion on this point, Mem. Imp. Mar. Obs., 1 (1924), 215 — 

 216. 



(4) Journ. Met. Soc, 39 (1920), 79; T.S.B.K., [ill] 2 (1920), 82. 



(5) T.S.B.K., 9 (1917), 515. 



(6) Ho., 57 (1906). 



