MODERN SEISMOLOGY SCRASE 197 



The other is called the distortional or transverse wave since the dis- 

 placement is at right angles to the direction of propagation, as is the 

 case with light waves. For each type of wave the speed of propaga- 

 tion depends on the elastic constants and the density of the material. 

 Seismologists are now agreed that on the seismogram of an earth- 

 quake at a fairly large distance the primary pulse, usually denoted 

 by P, marks the arrival of the compressional wave, whilst the com- 

 mencement of the second phase, denoted by S, indicates the arrival of 

 the distortional wave. Professor Turner very appropriately de- 

 scribed these waves as the " push " and the " shake " waves. In 

 addition to these body waves which travel through the medium, 

 there is another type of wave which can travel along the surface 

 in much the same Avay as water waves are propagated. These sur- 

 face waves were discovered by Lord Rayleigh and are usually named 

 after him. They play an important part in the third portion of the 

 seismogram; the commencement of this portion is usually called the 

 long wave or L phase. 



With the improvement of instrumental records it was soon found 

 that the times of travel of the P and S waves gave a smooth curve 

 when plotted against the distance from the epicenter. In 1907 

 Wiechert and Zoeppritz constructed travel-time graphs from a care- 

 ful examination of the records of three well-defined earthquakes, 

 and the tables derived by Turner from these graphs have served 

 as a basis for all important seismological work until quite recently. 

 These tables have been used, for example, in the determination of 

 epicenters and times of origin of earthquakes which are published 

 regularly in the International Seismological Summary. From the 

 time interval between the P and the S phases at any station both 

 the epicentral distance and the time of origin can be estimated by 

 means of the tables, and if the time intervals for three well- 

 distributed stations are known the position of the epicenter can 

 be located. A rough estimate of the position of the epicenter can 

 often be obtained at a single station provided that good records 

 of all three components are available, for a comparison of the 

 amplitudes of the components of the initial longitudinal pulse 

 indicates the direction from which the wave has arrived; this 

 information together with an estimate of the distance obtained from 

 the S-P interval gives the locality of the epicenter. 



It has been realized for some years now that the Zoeppritz-Turner 

 tables have appreciable errors, amounting at some distances to 20 

 seconds, and much work has been done recently on the construc- 

 tion of more accurate tables. H. Jeffreys and K. E. Bullen have 

 obtained revised times of travel of P and S from a statistical 

 analysis of data taken from a large number of the most fully observed 

 earthquakes treated in the International Seismological Summary; 



