ON SEISMOLOGICAL iNVE8TIGATI0N; 65 



When these observations are plotted on squared paper it is found 

 that they practically lie on the straight line referring to large waves in 

 fig. 1, indicating that this form of movement passes from its origin to 

 its antipodes with a constant arcual velocity of 3 km. per second. If, 

 however, the direction of propagation has been along a diameter, the 

 average velocity becomes 1'9 km. per second. The time taken for an 

 earthquake to travel from its origin to its antipodes, whether it does so as 

 a surface wave or as a mass wave, is about 110 minutes. 



One modification to this general statement respecting a constant 

 velocity rests on the fact that repeated observations made within ten 

 degrees of an earthquake origin have shown that the large wave velocity 

 within that region is about 1 S km. per second. Whatever the conditions 

 may be which give rise to this increase in velocity in a wave as it radiates 

 from its origin, it seems probable that the converse would take place as it 

 approached its antipodes, while the maximum velocity should be sought 

 for in the equatorial or quadrantal ' region of the earthquake's transit. 

 Inasmuch as curves drawn for the Alaskan and Ceram earthquakes show 

 that between 70° and 110° from their respective origins velocities may 

 reach 4 km. per second, and that many earthquakes indicate an increased 

 average velocity as their paths increase up to 110° in their lengths, there 

 are strong reasons for suspecting that the suggested phenomena may 

 exist. The comparatively small initial velocity and the slightly increased 

 quadrantal velocity above the average arcual velocity are indicated in 

 fig. 1 by dotted lines ; but whether this modification can be retained 

 remains to be determined by further observations. That the average 

 arcual velocity between O"* and 90^ is practically 3 km. per second finds 

 confirmation in the records for earthquakes Nos. 36, S3, 100, 119, and 

 193, originating in Japan, 133 and 134, originating near Borneo, and 

 105, from N.E. India, all of which were recorded by the same instrument 

 in the Isle of Wight. 



6. Interval betioeen the First Tremor and the Maximum Motion, 



In the British Association Reports for 1898, pp. 221-224, I dis- 

 cussed a table showing the duration of preliminary tremors or the interval 

 in time between the first tremor and the commencement of the large wave 

 phase of motion at difierent distances from a number of known origins. 

 One object of the discussion was to establish a working rule enabling an 

 observer to determine from the inspection of a single seismogram the 

 distance of an origin from the station at which such a record had been 

 obtained. Inasmuch as the table was to a great extent based upon 

 descriptions of records obtained from different types of instruments which 

 had difierent degrees of sensibility, the results obtained could not be 

 expected to be more than approximately correct. The following table, 

 which gives the time in minutes by which the first tremor has outraced 

 the maximum movement over paths of varying lengths, is based on 

 measurements made on seismograms obtained from similar instruments. 

 These intervals not only enable us to correct the working rule indicated 

 above, but, as it will be shown, they enable us to check the accuracy of 

 the curves relating to the arcual velocity of preliminary tremors and 

 large waves. 



■^o^ 



' Thia word means the district 90° distant from the earthquake origin. 

 1900. ff 



