EARTHQUAKE MEASUREMENT AT MIYAKO. 189 



38. The duration of an earthquahe. 



It is impossible to measure exactly the total duration of an 

 earthquake, which is of course much longer when instrumeutally 

 recorded than when simply felt without such an aid. Again, 

 the duration recorded by an instrument depends on its sen- 

 sibility. Thus, for instance, the duration of motion in Eqke. 

 No. 23 was, as given in the Table, only 120 seconds ; but the 

 duration of the same earthquake in Tokyo, recorded by Omori's 

 Horizontal Pendulum, was 2 hours. The fact is that the ordi- 

 nary seismographs of the Ewing and Milne and Gray types so 

 widely used in Japan, are capable of registering only the quicker 

 vibrations of an earthquake but insensitive to waves of long 

 period, the latter being very easily obscured by the friction of 

 the instruments. 



The total earthquake durations given in this paper, which 

 varied between 8.5 and 200 seconds (see the 3rd column of the 

 Table), are to be regarded as the durations of the sharper vibra- 

 tions in different earthquakes, namely of those waves causing 

 the greatest effects on our senses. 



Speaking generally, the duration of an earthquake will 

 depend both on its magnitude and on the distance of the 

 observing station from the origin. As a rough approximation, 



let us assume 



A 



D = 



kTT d 



in which -=3.1416; k is a constant; D, the duration in 

 seconds of an earthquake (less the duration of the preliminary 

 tremor) at a given station ; d the distance in kilometres between 

 the origin and the observing station ; and A the area of dis- 

 turbance, in which the intensity of motion is sufficiently strong 



