THE PREVISION OF EARTHQUAKES 



By CHARLES DAVISON, Sc.D., F.G.S. 



Between foreseeing and foretelling an unexpected event, there 

 would seem to be little if any difference, beyond the fact that the 

 one may be conducted in private while the other implies publica- 

 tion of some kind. But, to the corresponding words " prevision " 

 and " prediction," somewhat different meanings seem to be attri- 

 buted, prevision being apparently considered as an approximate, 

 and prediction as an accurate, form of forecast. This, at any 

 rate, is a distinction that will be assumed in the present paper, 

 for the prediction of an earthquake, the accurate forecast of its 

 occurrence at a particular time and in a particular place, is a 

 problem far beyond our powers, and likely to remain so for many 

 a year to come. 



It must be admitted, moreover, that this is the only kind of 

 forecast that is of any practical value. It is futile to urge that a 

 great earthquake will occur somewhere upon the globe on a 

 certain specified day, if the particular region to be affected is 

 unknown, as such an earthquake actually does occur on an 

 average more than once a week. And, on the other hand, it is 

 of little advantage to proclaim that a certain area will be de- 

 vastated at some unknown time within the next few years, or 

 even within the next year, for no population will consent to 

 spend their nights out of doors for so long and so indefinite a 

 period. Until, therefore, we can fix the area in one case, or the 

 time in the other, within fairly close limits, the forecasting of 

 earthquakes can have little practical importance. But the 

 problem is so complex, and at the same time its solution would 

 possess such untold value for the dwellers in seismic countries, 

 that it does seem worth while to examine the progress that has 

 been made in the hope that further knowledge and greater 

 experience may in time to come lead us to the desired goal. 



It is clear that if, of the two elements, it be possible to deter- 

 mine only one with accuracy, it is more useful to know the site 

 with precision, and the epoch with somewhat less certainty. At 



639 



