ON SEISMOLOGICAL INVESTIGATIONS. 



39 



book on Growth and Form, p. 122 (Camb. Univ. Press, 1917). The 

 actual length is apparently shorter than 156 years — nearer 150 years, 

 but the material is scarcely sufficient to warrant a very precise estimate. 

 The Chinese earthquakes show this periodicity of 156 years very 

 clearly. Dividing the period into 18 equal parts the totals are as in 

 the columns 0, the columns C being calculated from the formula : — 



C = 45 + 15-6 Cos (0 - 225°). 



Period of 156 years in Chinese Earthquakes, exhibited in 18 groufs 



of 7 years. 



The differences 0—0 show a variation in the half -cycle of 78 years, 

 to which attention has already been drawn in the paper on the 15-month 

 period {loc. cit.), and, moreover, the phases of the 15-month term 

 vary in this half-cycle of 78 years and in the quarter-cycle of 39 years. 

 This quarter-cycle appears in numerous meteorological phenomena, 

 and should probably i-eplace the supposed ' Briickner Cycle ' of 35 

 years (see Q.J.R. Met. Sac. xh., p. 322). 



But 39 years is no longer a multiple of 26 years, though related 

 to it. The submultiple of both — viz., 13 years — has, however, been 

 shown to affect a large number of meteorological phenomena, being a 

 double ' chapter ' of the kind indicated in the paper just cited, and 

 others which have followed it. 



As yet it cannot be said that we have anything really tangible in 

 the shape of a physical hypothesis, but these numerical relations are 

 certainly suggestive, and are sufficient to guide further inquiry. 

 Naturally in tentative exploration of this kind much time is spent 

 in unproductive essays, but this need not be grudged. 



B 2 



