38 THE ROYAL SOCIETY OF CANADA 



stated the longitudinal and transverse waves suffer variable velocity, 

 dependent upon the depth of path traversed, while the long or un- 

 dulatory waves confined- to the crust, which although not homo- 

 geneous, are practically constant, and not a function of the distance. 

 These last waves have a velocity of about 3-8 km. per second or say 

 230 km. per minute, a determination made from several hundred 

 seismograms obtained at Ottawa. It may be observed that when the 

 quake is particularly violent and deep-seated we obtain not only a 

 record of the long waves on the shorter arc to the observing station, 

 but also by the supplementary arc completing the great circle or 

 circumference of the Earth. Even a third wave has been recorded. 

 These successive waves give us a measure of the co-efîfiicient of absorp- 

 tion which is of the order -0003, which means that for the distance of 

 some 2,500 km. the energy has decreased one-half. (e--ooo28A = i^). 

 It is obvious that were there no absorption we would have an earth- 

 quake at the anti-epicentre of equal intensity as the original one. 

 Perhaps it may be interesting to remind one of this expansion and 

 concentration of energy passing over the surface of the earth in con- 

 centric waves, the case of the violent explosion in the Sunda straits 

 between Sumatra and Java of Krakatoa in August, 1883, when on the 

 self-recording mercurial barometer at the meteorological office in 

 Toronto several impulses of the atmospheric waves, direct and re- 

 bound, were recorded. 



The surface or long waves of an earthquake have an average 

 period of about 20 seconds; at the beginning of their appearance they 

 are frequently longer and gradually decrease to a period of about 12 

 seconds, so that we may put their length as lying approximately be- 

 tween 40 and 80 kilometres, and with a fairly uniform velocity of 

 propagation of 230 km. per minute. After the long waves are well 

 launched on the seismogram there are practically no complications in 

 the record, differing in this respect from the longitudinal and trans- 

 verse waves which we shall presently note. The rate of propaga- 

 tion of the longitudinal waves is conditioned by the expression 



A 4- 2/^ II 



Vp = / — , and that of the transverse waves by Vg = ^~> where 



[1 and p are the usual symbols for rigidity and density, and 

 X = '^ — f Z^, in which k- = modulus of cubical compression. 



These two waves are propagated as spherical waves, so that at a 

 station from a single impulse from the epicentral region we would 

 have the P wave arrive first, then a quiescence followed by the S wave. 

 However, following Huygen's principle each spherical wave as it 

 reaches the surface becomes again the centre of a disturbance from 

 which issue P and S waves, so that our seismogram does not show this 



