EARTHQUAKE MOTION TO GREAT DISTANCES. l(j 



As the density of iron is about 7 '5, and as its density in the interior of the earth 

 is, on the one hand, lessened by increased temperature, and, on the other, increased 

 by pressure, we may take it that the central metallic core extends to about 0'55 of the 

 radius from the centre, or to about the same depth as is reached by the wave paths 

 which emerge at a distance of 90 of arc from the origin. 



It will be interesting to see if observations which may be obtained hereafter at 

 distances of more than 90 of arc from the origin bear this out ; at present we have 

 only the few observations of the Argentine earthquake of 1894. If the Tokio record 

 of this can be accepted as representing the arrival of the condensational waves, there 

 is a marked fall in the time curve between 90 and 155, indicating an earlier 

 emergence of the waves at the greater as compared with the lesser distance. 



The European records of the same shock do not indicate a drop in the time curve ; 

 those of the first phase lie very close to the continuation of the curve for the 

 condensational wave, but to bring the second phase near the continuation of this 

 curve it has to be bent downwards into greater parallelism with the time curve of the 

 first phase, thus indicating a change in the ratio of the elastic constants. The 

 observations are, however, too few for any dependence to be placed upon them. 



7. The records of the third phase show no such variation of velocity with 

 distance from the origin as was noticed in the case of the first and second phases. 

 The velocities of propagation as deduced from the times of commencement show 

 a good deal of variation among themselves, but there is no indication of an increase of 

 apparent rate of propagation with distance. 



I have already referred to the difficulty there is in determining with certainty the 

 time of commencement of this phase of wave motion, and if we turn to the time of 

 maximum of this phase, usually determinable with greater certainty, we find a close 

 agreement in the rate of propagation at all distances, except in the case of the 

 Turkestan earthquakes and the shortest arcs dealt with. These are slightly but 

 distinctly less than those deduced from observations over longer arcs, but with this 

 exception the observations point to the conclusion that the apparent rate of 

 propagation is uniform at all distances from the origin. 



This conclusion is not in concordance with the latest results published by Professor 

 MILNE,* who has deduced as average rates of propagation of the large waves the 

 following values : 



Distance from origin 20 60 80 110 



Velocity of propagation, kiloms. per second . . . 2'1 2'8 2'9 3'3 



These values if plotted do not fall into a smooth curve, but such as it is the 

 curve would point to propagation through the earth along brachistochronic paths, 

 slightly concave towards the centre of the earth. If this be the case, the form of 



* ' Brit. Assoc. Rep.,' 1898, p. 220. 

 VOL. CXCIV. A. 2 



