ON SEISMOLOGICAL INVESTIGATIONS. 



55 



a rest period of 20 days. The mean line through these various deter- 

 minations indicates that activity and rest are directly proportional. 

 After marked efforts to bring about adjustments in the crust of our earth 

 there are long periods of quiescence and vice versa. A definite time- 

 interval is required to bring about a condition for hypogenic activity. 



VIII. A Catalogue of Large Earthquakes. 



In the British Association Eeport for 1908 I drew attention to the 

 fact that existing catalogues of earthquakes consisted of materials ex- 

 tremely heterogeneous in their character. Earthquakes which had only 

 shaken a few square miles were included with those which might have 

 shaken the whole world. Further than this the heterogeneity varied 



10 20 30 40 



Average No. of Days of Rest. 

 Fig. 2. 



in different historical periods. Ancient records only referred to large 

 earthquakes, while, as we approach modern times, this type of dis- 

 turbance was eclipsed by numerous entries relating to tremors which 

 had only a local significance. If we take this as a fact we see in it an 

 explanation why the numerous analyses of earthquake statistics have 

 failed to reveal any striking results respecting the distribution of earth- 

 quakes either in regard to space or time. 



To obtain materials which might throw light upon seismic frequency 

 and periodicity, it would be necessary to draw up lists for districts and 

 one for the world from which seismic trivialities were so far as possible 

 excluded. With this object in view, I have made certain progress with 

 a catalogue which only refers to earthquakes which have been accom- 

 panied by destruction, or by changes of the earth's surface, or which 

 have extended over large areas. In many instances these disturbances 



