part 1] iisr the eeequejstct of earthquakes. S"' 



[The distribution of shocks by hours is not tabulated here, but 

 was extracted and submitted to the usual process of harmonic 

 analysis. The solar record gave for the diurnal and semidiurnal 

 periods the formula 



F = l-f30 sin {t-\-58° 30') +-10 sin (t-\-12°), 



which is very close to the formula (1) on p. 103 (079. cit.), and, 

 as in that case, almost completely covers the diurnal variation in 

 frequency. The lunar record, treated similarly, gives the formula 



p = l + -05 sin (t + 151^ 30')4--015 sin (2t-f 168°). 



Here the epochs of maximum frequency are very different from 

 those of the solar record, but the coefficients, which do not, even. 

 in the case^ of the solar record, exceed the expectanc}^ are also 

 reduced to an insignificant magnitude ; the real meaning, therefore, 

 of the formula is that there is no indication of either diurnal or 

 semidiurnal variation of frequency which can be correlated with 

 the hour-angle of the moon. Hence we may conclude that the 

 variation in frequency of earthquakes at different times of the day, 

 which is conspicuous and consistent, so far as Italy and the last 

 quarter-century are concerned, cannot be attributed in any appre- 

 ciable degree to the gravitational stresses set up by the sun. 



This leaves the third term of the formula (5) as the only one 

 which can be attributed to this cause.] 



b2 



