62 REPORT — 1900, 



our earth the elasticity governiiig the transmission of condensational 

 waves could be determined. The following notes show that, although the 

 first conclusion and the consequent hypothesis do not require modification, 

 constants necessary in farther calculations require to be modified. 



With regard to the large waves my own assumption was that their 

 apparent increase in velocity with distance might be due to the fact that 

 it was only large waves which, travelling faster than small waves, reached 

 great distances. 



The observations brought together in this paper show that this idea 

 has to be abandoned, and in its place we are to accept either the hypothesis 

 of a surface wave which increases its velocity in regions 90° from the focus, 

 or of a distortional wave passing through the earth the outcrop of which 

 "■ives rise to similar surface undulations. 



3. Sources of Error. 



The phases of earthquake motion here considered are the first pre- 

 liminary tremors and the first group of large waves, which latter in a 

 seismogram representing an earthquake which lias originated at a great 

 distance uoually correspond to the maximum movement. 



Altliough near to the origin of an earthquake there is a varying 

 interval of several seconds between the first movements and the shock or 

 shocks, it is the time of occurrence of this latter phase which is taken as 

 the datum to which observations made at great distances from origins are 

 referred. The initial time for all large earthquakes has been a matter of 

 inference. It may be deduced from the times at which clocks have been 

 stopped, or which have been noted with varying degrees of accui-acy by 

 survivors in an epifocal district, but more generally it has been deduced 

 from automatic time determinations outside such an area, and subtracting 

 from the same an interval which the shock is assumed to have taken to 

 travel from its origin to the point or points where these chronographic 

 records have been made. The determination of this interval is based 

 upon repeated observations of earthquake velocities made between stations 

 well removed from an epicentre and well outside a meizoseismal area. 

 These figures are important, not only for this particular purpose, but also 

 for completing velocity curves which may represent transmission over the 

 surface and through the material of the whole globe. They have been 

 arrived at by many observers, the last being those given by Dr. F. Omori, 

 who for paths commencing 100 kms. from an origin and extending to 

 distances of 1,000 kms. gives the velocities of 2-2 km. for preliminary 

 tremors and 1*7 km. for large waves, and within these limits the former 

 outrace the latter at the constant rate of 15 seconds per 100 kms. 



When we remember that large earthquakes may sometimes originate 

 as practically simultaneous displacements over very lai'ge areas, it is seen 

 that the application of the method here considered might easily result in 

 determinations of initial times from a few to some sixty seconds earlier 

 than had really been the case. Errors of this nature would result in a 

 general lowering of the determinations for true velocity of transmission 

 of earthquake motion to distant stations, the deviation from the truth 

 being most marked for the preliminaiy tremors, and in records referring 

 to transmission to stations comparatively near to an origin. 



Another serious error afli"ecting the determination of initial time arises 

 from the difiiculty in accurately locating the position of a focus, especially 

 when this is sub-oceanic. 



