On Lunar and Solar Periodicities of Earthquakes. 463 



is in the first case only 1 in 300,000, and in the second almost 

 infinitesimally small. The reality of the period would be thereby 

 established beyond reasonable doubt, unless the peculiarity of earth- 

 quakes occurring in groups, as discussed in the previous section, can- 

 be shown to raise the expectancy sufficiently. The fact, however, 

 that in each hemisphere the phase of the periodicity found is nearly 

 identical in a great number of cases disposes of all doubt which 

 might remain on that point. 



In some of the details Mr. Davison's results would seem to require 

 further confirmation. The evidence, for instance, that the strong 

 and weak shocks follow different laws is not very strong when 

 examined by means of the theory of probability. Mr. Davison 

 classifies the Japanese earthquakes into three groups, according to* 

 their intensity, and finds that the maxima of the annual period 

 agree in the two stronger groups and take place in winter, while the 

 maxima for the weakest group occur in summer. The magnitude of 

 the amplitude for the former is 0'17, while the expectancy is 0'074 

 and 0*133 respectively. Here the excess of amplitude over the 

 expectancy is no.t sufficiently marked to allow of any certain con- 

 clusions being drawn. A similar remark applies to the record of 

 Zante (see § 45 of Mr. Davison's paper) . 



7. The daily periods of earthquakes have been fully discussed by 

 the same author in a paper published in the ' Philosophical Magazine/ 

 The following table summarises some of the more important results, the 

 twenty-four hours period only being taken into account. The epoch 

 given is that of the maximum, and I have added the expectancy of 

 amplitude calculated on the principles of this paper. 





dumber of 



1 Expectancy, 













earthquakes, 

 n. 





Amplitude. 





Epoch. 



Japan (Tokio), Summer . 



543 



0-076 



0*176 



9 1 



58° 



1 A.M. 



Winter . . 



661 



0-069 



0-093 



10 



39 





Year .... 



1204 



0-051 



0-130 



10 



14 







597 



0-073 



061 







2 



P.M. 





578 



074 



0-239 



11 



50 



A.M. 



Year 



1175 



0-052 



0-147 



11 



53 







210 



0-122 



0*273 



10 



49 





Italy 



8177 



0-020 



0-324 







25 



P.M. 



The amplitudes are seen to exceed the expectancy considerably in 

 all cases but one. The reality of the daily period must be considered 

 established, unless the evaluation of the expectancy is faulty, owing 



