i902.] R. b. Oldham — Tidal Periodicity in Earthquakes of Assam. J i9 



successive similar meridian passages of the moon as representing 24 lunar 

 hours, and convert the recorded times into lunar times, it is obvious 

 that, in a long series of observations, any irregularity of frequency, at 

 any particular hour of solar time, will get spread over the whole of the 

 lunar day, and in its place will be introduced any fresh irregularity 

 due to the position of the moon. Now as the moon has twice the effici- 

 ency of the sun, as a tide producer, any irregularities due to the tide 

 producing forces should be double as great as in the case of the sun. 



Unfortunately the test cannot be applied in this case as, on trial, 

 it was found that the series of observations was not sufficiently long 

 to eliminate the effect of the diurnal irregularities. 



This method of elimination failing, we must fall back on the 

 recorded times, to see whether there is no other method of eliminat- 

 ing the non-tidal diurnal variation, and a method appears which depends 

 on the fact that, taking the year as a whole, the tidal effect is on the 

 average the same all through, since the times of passage of the tidal 

 circles during the six hours on either side of midnight are the same for a 

 south declination as the times on either side of midday in the case 

 of the same amount of north declination. 



If, then, we take the recorded frequency of shocks for each hour, 

 write them down in two lines, placing those for the hour after midday 

 under those for the hour after midnight and so on, and then add the 

 two linas, we obtain a series of numbers representing the semi-diurnal 

 curve of frequency. In this curve any diurnal periodicity, which is of a 

 harmonic nature, is completely eliminated, and any non-harmonic 

 periodicity largely reduced in amount. On the other hand any semi- 

 diurnal periodicity which is harmonic in character, or which, if not 

 harmonic, has its irregularities similarly distributed with regard 

 to midnight and midday, will be exaggerated ; that is to say the effect 

 we are looking for will be increased, while that which we wish to elim- 

 inate will be reduced, in amount. 



In the next tabular statement the process is illustrated as regards 

 the total number of shocks, and four more lines given, showing the 

 results obtained in the case of certain combinations of shocks, which 

 will be referred to further on. 



