Elastic Waves, with Seismological Applications. 89 



(2) Oondensational-rarefactional Wave incident in the Rock. 



e. 



A. 



A,. 



e\ 



A'. 



<p. B r 







1 



•021 



0° 



•979 





14° 2' 



1 



•007 



10° 



•937 



8° -055 



26° 34' 



1 



•003 



19° 



•839 



15° -158 



45° 



1 



•104 



32° 



•G46 



24° -256 



59° 2' 



1 



•240 



40° 



•503 



30° -256 



73° 18' 



1 



•268 



46° 



•471 



34° -261 



84° 17' 



1 



•039 



48° 



•618 



35° -344 



90° 



1 



1 



48°-2 





35° 3 



The chief peculiarity in the first of these hypothetical cases 

 is, perhaps, the vanishing at two incidences of the refracted 

 wave in the fluid. It vanishes at the critical angle (35° 13') 

 at which the reflected condensational wave disappears ; and 

 then it has its own critical angle (50° 44 r ). 



Between these limits its energy rises to a pronounced 

 maximum. In these respects there is a broad similarity between 

 this case and the cases of rock and water, and rock and air. 

 The differences are only differences of degree, depending on 

 the different relations among the densities. 



The bearing of this investigation upon the question of 

 Earthquake sounds has been discussed above. There is, how- 

 ever, another point on which some light seems to be thrown 

 by these calculations. I refer to the Preliminary Tremors and 

 (comparatively) Large Waves, which were first observed in 

 1^89 and are now recorded on many delicate seismographs in 

 countries which are not, in the ordinary sense, subject to 

 earthquakes. The discussion of these forms the bulk of the last 

 British Association Report of the Seismological Committee: 

 no one doubts that they have come from a distant earthquake 

 origin, the preliminary tremors outrunning the big waves 

 as they pass through the earth. The origin being known, 

 it is an eas}^ matter to calculate approximately the average 

 velocity of the swiftest of these tremors. Following out one of 

 Milne's suggestions, 1 have found * that the square of this 

 average speed may be represented by the formula 

 w 2 = 2-9 + -026d, 



where d is the average depth of the chord joining the earth- 

 quake origin with the station where the tremors are recorded, 

 the units being miles and seconds. 



This involves an increase of about 1*2 per cent, per mile 



* See " The New Seismology " in the ' Scottish Geographical Magazine/ 

 January 1899. 



