Section III, 1920 [47] Trans. R.S.C. 



Analysis of Earthquake Waves 

 By Otto Klotz, LL.D., F.R.S.C. 



(Read May Meeting, 1920) 



This preliminary note is to give the results of the first atternpt 

 to apply the Henrici Harmonic Analyzer to the analysis of earthquake 

 waves. When Professor Dayton C. Miller, of the Case School of 

 Applied Science, Cleveland, gave his splendid address on the Science 

 of Musical Sounds before our Section, illustrated by many diagrams, 

 the writer was struck with the resemblance of some of the compound 

 waves to earthquake waves and saw, or believed to see, its applicability 

 to studies in seismology. The investigation determined upon was 

 made possible through the kind co-operation of Professor Miller, with 

 whose Henrici Harmonic Analyzer the numerical data was obtained. 



The analyzer and the 32-element harmonic synthesizer to which 

 reference will subsequently be made were described by Professor 

 Miller in the Journal of the Franklin Institute; the former in September, 

 1916, and the latter in January, 1916. 



The seismogram of September 6, 1915, giving a fine record of the 

 earthquake in the Pacific off the coast of Central America was chosen 

 for this first investigation. The epicentre of this quake was given 

 by the Dominion Observatory as 0=14° 14', A = 90° 30' W.; and the 

 energy released £=10x10^3 ergs, computed from the Observatory 

 seismogram. The exponent 23 is preserved, as it occurs in some other 

 investigations, so that comparisons are confined to the co-efficients. 



Whatever part of the seismogram is selected for analysis that 

 part which is chosen for a period, that is, for the compound wave, 

 after which repetition is supposed to take place, must be enlarged to 

 exactly 400 mm. in order to fit the machine — the analyzer. It may 

 be remarked that selecting a period, particularly for the P and S 

 waves, is rather difficult, and its position on the seismogram not too 

 obvious. Sections or periods were selected at the beginning of each 

 of the waves P, S and L, representing the preliminary, first or longi- 

 tudinal waves; the second or transverse waves; and the last, long or 

 surface waves. Only for the beginning of the P waves, the cataclysism 

 of the quake is the time absolutely certain. The mathematical 

 process of analysis by Fourier series is too time-consuming, as the 

 writer found out years ago, to apply the many analyses one wants, 

 so that one interprets or reads the periods or supposed periods directly 



