64 



REPORT — 1902. 



IV. Duration of the First rreliminary Tremors. 



In seismograms from the Milne horizontal pendulum the first 

 preliminary tremors, which usually appear as a thickening of the normal 

 trace, are only seen in connection with fairly large disturbances. The 

 reason for this is at least twofold : first, as a recorder of elastic vibrations 

 the multiplication of the instrument is low, with the result that when 

 these vibrations are minute they may be lost in the thickness of the 

 trace ; and second, because as the recording surface only moves at a rate of 

 1 mm. per minute it is difficult to measure very small intervals of time. 

 For ' near ' earthquakes, therefore, the seismograms usually show a dis- 

 turbance commencing suddenly, and the duration of the preliminary 

 tremors connected with the same can only be inferred by the continuation 

 of the curve of the durations of the movements as recorded at distant 

 stations backwards towards its origin. It is satisfactory to notice 

 that these inferred durations closely agree with actual measurements 

 of the same made by seismographs adapted to record ' near ' earthquakes. 



The following four tables give the durations of preliminary tremors 

 in minutes for earthquakes originating near Japan, Mexico, Alaska, or in 

 the East Indies as recorded at Shide, Kew, Toronto, Victoria (B.C.), 

 Bombay, Batavia, Mauritius, Madras, and the Cape of Good Hope. 



The number following a duration and placed in parenthesis is the 

 number of the earthquake as entered in the iShide register. For districts 

 see map, fig. 2 (^ Plate I.). 



Origins West of Alasl:a (District A). 



