EARTHQUAKE MOTION TO GREAT DISTANCES. 



171 



and may rise to over 4'0 Idioms, per second for the long low waves which outrun the 

 principal ones. 



The conclusion drawn from the records is greatly strengthened by the fact that, in 

 the case of the great earthquake of 1897, the only one where the rate of propagation 

 has been carefully determined within the seismic area as well as at a distance, it 

 was found that the rate of propagation of the sensible shock, from the origin to 

 distances up to 1 5 of arc, was practically the same as the rate of propagation of the 

 waves of the third phase to a distance of 65 of arc. The actual difference as 

 calculated is but O'l kilom. per second, or about one-thirtieth of the value, a 

 difference which is well within the inevitable limits of error of the observations. 



The nature of these waves lias yet to be elucidated. The elastic surface waves 

 investigated by Lord RAYLEIGH should travel, in material of the nature of the 

 rocks with which we are acquainted, at a rate of about 0'9 of the rate of propagation 

 of a distortional plane wave in an infinite solid. This for continuous rock of the 

 nature of that which forms the crust of the earth is about 2'6 kiloms. per second, so 

 that if we take the rate of propagation of the greatest surface waves at 2'9 kiloms. 

 per second, the excess is just about what the defect should lie. 



The form of the molecular movement in the waves investigated by Lord RAYLEIGH, 

 does not seem to be consonant with that recorded in the neighbourhood of the origin. 

 At great distances it may lie in closer accord, but apart from this, the rate of propa- 

 gation of the purely elastic surface waves is not a function of either their length or 

 amplitude, while that of the great surface undulations of an earthquake apj>ears 

 to be a function of one or both of these. This is intelligible if, as was suggested by 

 Lord KELVIN,* the propagation of these waves is accelerated by gravity, and the 

 fact that the rate of propagation seems to be in some way a function of their size is 

 a support to the suggestion. 



8. There remain now for notice only those cases where the record has commenced 

 earlier than the time at which the condensational waves, set up by the earthquake 

 to which the record is attributed, would l>e expected to emerge. In all, seven such 

 cases are included in the records noticed alx>ve, and are tabulated Ixjlow. 



'Seismol. Jour. Japan,' vol. 3, 1894, p. 87. 

 Z 2 



