256 REPORTS ON THE STATE OF SCIENCE, ETC. 



deduction, instead of the observations of one or more stations of different dis- 

 turbances. This characteristic of the method offers one important advantage. 

 For when we attempt to obtain information as to the depth of earthquake foci 

 by using observations of various disturbances the number of unknowns (viz. 

 the actual depths) is equal to the number of different disturbances recorded in 

 the observations. When the observations are all associated with a single 

 disturbance only one focal depth is involved. 



Depth of Focus for 1920, December 16. 



As an example of the first method a note may be given on the great Chinese 

 earthquake, of which some particulars were furnished in the last report. Adopting 

 the epicentre 35°.5 N. 105°. 5 E. and To = 1920 Dec. 16d. 12h. 5m. 46s., as given 

 in the last report, the records from twenty-seven stations now available give a 

 mean correction to To of only —0.5 sec. The mean numerical residual is 

 + 7.1 sec, so that the probable error of the mean of 27 is about 1 sec, and 

 we may regard To as exceptionally well determined. If only we had as many 

 stations near the anticentre we could determine the focal depth with great 

 accuracy. But up to the present La Paz is the only station from which results 

 have been received. The time atA= 160°. 2 was 20m. 12s. after To, or +4s. 

 in excess of the adopted formula. So far as this evidence goes, then, the focus 

 was slightly above the normal depth d ; but news from other stations in South 

 America would be of the greatest interest. 



Earthquake Periodicity. 



Hitherto the periods considered in connection with earthquakes have been 

 long; recently attention has been directed to some periods near 300 years, and 

 even the fifteen months on which much work has been done is very long com- 

 pared with that now to be mentioned, which is of only twenty-one minutes. But 

 the results obtained are so startling, and yet so well supported, that, although 

 they are only a few weeks old, it seems desirable to give some account of them. 



It is mentioned in a preceding paragraph that the work of collation has recently 

 been extended to the smaller earthquakes, which are sometimes only scantily 

 recorded and therefore difficult to identify. It is a great help that they often 

 come from the same neighbourhood as a previous shock, and possibly, in many 

 cases, from the same actual focus. At any rate, it is a convenience to refer 

 them to the same focus, when discrepancies, if any, will be shown by the 

 residuals. 



On August 3 to 10, 1917, some twenty shocks were recorded by Mizusawa and 

 Osaka, which clearly came from the same focus, or nearly the same. It was 

 natural to inquire whether there was any regularity in the intervals between 

 shocks, and (without any special anticipation of finding it) a period of 

 21m. presented itself. The half-period of 10.5m. and the third of 7m. are 

 both liable to present themselves in one connection or another, but there seems 

 little doubt that 21m. is the master-period. 



In a preceding paragraph it is mentioned that 21m. is approximately the 

 time taken by a P wave to go right through the earth. In the investigation 

 there mentioned 20m. 17s. appears ; but this is the time from a focus of the 

 average depth to the opposite face, and must be shorter than the complete time 

 from face to face. The defect from 21m. may therefore, if the interpretation 

 suggested is correct, give us incidentally the depth of the average focus. But 

 it is too early to consider this as more than a possibility. 



it follows that 21m. is also the time taken by a P wave to travel to the 

 earth's centre and back again to the surface, i.e. it is the time for a possible 

 pulsation. It is again too early to put forward this view of the following facts 

 definitively, but it may help to co-ordinate the facts if this possibility is kept 

 in mind. 



The following figures will show how the 21m. period appeared and what 

 oart was played bv the half-period 10.5m. The first shock came at August 

 8d. 5h. 25.0m., and was followed by two others 2x21m. and 8x21m. later. 

 The actual intervals were not 43m. and 168m., but 45.9m. and 170.6m., so that 

 the .shocks do not come precisely but only approximately at the expected 



I 



