52 THE ROYAL SOCIETY OF CANADA 



likely be in a different phase from that which it would have had, had 

 the simple sine wave motion continued its consequent effect on the 

 compound wave. 



For the L waves the outstanding periods are 12 and 18 seconds» 

 I>eriods that have often been directly measured. But it will be ob- 

 served that other waves ranging from about 5 to 30 seconds are mark- 

 edly conspicuous, more so than those we recognize in the later L waves 

 when they are free from the P and S waves, and show themselves in 

 fairly uniform sinuous waves. Now, it is a characteristic of L waves, 

 when one can see them directly on the seismogram, to begin with a 

 long period, often of 40 seconds, then rapidly to fall to about 20 

 seconds, later to drop by gradations to about 12 seconds, at which 

 period they remain frequently to the end of the quake or as long as 

 one can directly measure them. The writer some years ago offered as 

 an explanation of this decrease in wave length for the L waves, the 

 greater surface mass involved at the beginning of the quake, and with 

 the continuance of the quake effect in travelling over the earth for 

 half an hour, j.n hour, or even several hours, the depth of the earth's 

 crust involved became less and hence the period less also. 



Now, the graduation of period which we directly measure on the 

 seismogram does not, from the inherent conditions of the analysis, 

 always manifest itself. The theory upon which the analysis is based 

 presupposes a combination and superposition of regular, continuous 

 sine curves or waves forming a compound curve of definite periodicity. 

 As already stated, it is difficult to select the section for the P, S, or L 

 waves which appears to form a complete cycle. As a matter of fact, 

 such cycles do in reality not exist, for the waves affecting the earth 

 particle are supplemented continuously by new wave trains, reflected 

 by various paths. 



The analysis, therefore, is limited to that extent, and periods of 

 waves which are not sub-multiples of the assumed period of the 

 compound curve are excluded in the above analysis, even though 

 they exist. For a satisfactory study of seismograms by means of the 

 analyzer and synthesizer these instruments should be available for 

 instant use — on the one hand that consideration may be given to the 

 assumption of different compound curves as the periodic one, in 

 order that the really predominant waves, their period and amplitude 

 may emerge ; and on the other hand to study and compare the analyses 

 of different seismograms, that may possibly lead to the discovery 

 of selective waves, dependent upon the nature of the material and 

 conditions along their paths, modified by depth. The Observatory 

 is in hopes of securing an analyzer and synthesizer. 



