Periodicity of Earthquakes. 885 



It is worth while noticing that somewhat similar results 

 are obtained for the annual period of shocks of different in- 

 tensities (Table IV.). In strong shocks, the maximum-epoch 

 is nearly always in winter. There appear, however, to be two 

 distinct classes of slight shocks, one with the maximum-epoch 

 in winter and the other in summer. Of the two classes, the 

 former probably contains the stronger shocks, for the winter 

 maximum is manifested by the less complete ot two catalogues 

 for the same district ; while the summer maximum is only 

 visible in the results obtained from the fullest records or from 

 catalogues in which the very slightest shocks are included *. 



Thus, with one exception (Toronto), the amplitude of the 

 semi-diurnal period is greater in weak, than in strong, earth- 

 quakes. Except at Bombay, there is no evidence of any 

 inversion, or considerable change of epoch. 



Origin of the Diurnal Periodicity of Earthquakes. 



In Japan the diurnal periodicity of ordinary earthquakes 

 may be due to more than one cause ; that of after-shocks is 

 probably due to the diurnal variation of barometric pressure f 

 The effects of the latter variation will no doubt depend on 

 the mode of origin of the earthquakes. If the Ordinary 

 earthquakes were due to slips along a normal fault, the 

 maximum-epoch of seismic frequency should coincide 

 approximately with that of barometric pressure. If they 

 were due to slips along a reversed fault, the maximum-epoch 

 of seismic frequency should coincide approximately with the 

 minimum-epoch of barometric pressure. Lastly, if the 

 ordinary earthquakes were due to slips along a reversed 

 fault, the after-shocks would be caused by returns of the 

 displaced mass towards its former position of equilibrium, 

 and thus the diurnal variation in pressure (with a maximum 

 shortly after midnight) would cause the maximum-epoch of 

 ordinary earthquakes to occur about noon, and that of after- 

 shocks about midnight. In this way, we may account in 

 part perhaps, but not entirely, for the curious reversal of 

 maximum-epoch manifested in several of the above records. 



* Phil. Trans. 1893 A, pp. 1116-1120. 

 t Phil. Mag. vol. xli. pp. 915-916 (1921). 



