346 MR. R. D. OLDHAM ON THE [Aug. I907, 



Prom the foregoing table it will be seen that the intervals are 

 distinctly greater in the case of the Colombian earthquake and 

 oceanic wave-paths, the mean difference being I/O minute in time 

 for an excess of only 2° in distance ; as the time-interval increases 

 by almost exactly 0*1 minute per degree at about 90° from the 

 origin, the observed difference is 5 times as great as that which 

 would be due to the increased distance if the rate of transmission 

 had been the same in each case, and the excess (being about 7 per 

 cent, of the whole interval) is greater than can be accounted for 

 by errors of record or interpretation of the seismograms. In other 

 words, the substance through which these waves of the Colombian 

 earthquake travelled transmitted the second phase at a slower rate 

 — as compared with the rate of transmission of the first phase — • 

 than that traversed by the corresponding waves of the San Francisco 

 earthquake. 



It may be a mere accident that the wave-paths in the one case 

 lie under a continental and, in the other case under an oceanic area, 

 but the coincidence is certainly noteworthy and suggestive, and the 

 notion that it is not merely accidental derives some support from 

 the Japanese records of the San Francisco earthquake. According 

 to Prof. Omori the mean interval derived from the records of four 

 distinct stations in Japan was 9*7 minutes for a mean distance of 

 75°, the wave-paths running under the Northern Pacific and being, 

 therefore, oceanic. The European records of the earthquakes of 

 Japan, August 9th, 1901, and Assam, June 12th, 1906, give mean 

 intervals of 10*6 minutes for a mean distance of 83-3° in the first 

 case, and 8-3 minutes for a mean distance of 64*2° in the second, 

 from which we may deduce an interval of about 9*4 minutes for a 

 distance of 75° when the wave-paths lie under the continents of 

 Asia and Europe. The interval as deduced from the average-curves 

 on p. 462, vol. lxii (1906) of this Journal, which are practically 

 based on continental wave-paths only, is 9*0 minutes ; so here 

 again we find that the material under the Pacific Ocean transmits 

 the second-phase waves at a slower rate, as compared with the first- 

 phase waves, than does the material under the Eurasian continent. 



The exact depth to which this difference extends is not, at present, 

 determinable with accuracy, but the figures in Table I (p. 345) 

 allow us to form an approximate estimate. It will be noticed 

 that the distances range from 75° to a little over 90°; now, the 

 maximum depth of the chord from the surface is 0*21 of the radius 

 for an arc of 75°, and 0*29 for an arc of 90°, and the wave-paths 

 descend somewhat below the direct line of the chords. We may, 

 therefore, take it that the wave-paths under consideration reached a 

 maximum depth of from one-quarter to one-third of the radius. A 

 consideration of the difference between the intervals for the two 

 shocks at the same station shows that it does not appear to increase 

 with an increase of distance, but rather tends to decrease in amount, 

 suggesting the conclusion that the deeper wave-paths entered a 

 region where there was no difference, or at any rate a smaller 

 difference, in the rate of propagation of the two phases. From this 



