MODERN SEISMOLOGY HODGSON 



353 



again a fair agreement. We thus have a measure of the velocity of 

 propagation of the two types of waves in the uppermost layers of 

 the earth's crust, based on observations of five separate explosions, 

 registered at three or more stations in each case. Obviously we wish 

 to add to our observational data of this nature. 



To quote from our legal friends, " time is to be the essence " of 

 our experiments. The nearest of the La Courtine stations was 31/2 

 miles from the blast. The first tremor, marking the time of arrival 

 of the P wave, registered about one second after the explosion. The 

 S wave required about two seconds to travel the same distance. The 

 difference in time of arrival was thus about one second in the case 



N Coftingm 



rtJmlMf H-S 



AAMt^^ 



Figure 1. — Upper group : Seismograms registered at Goltingen and Jena, respectively, 

 of an earthquake in Kansu, China, 1020, December 16. Lower group : Seismograms 

 registered at Jena, of eai-thquakes in Kamchatka, on the dates indicated. (This 

 illustration has been copied from Gutenberg's discussion of " Ferubeben," in Uorn- 

 traeger's " Handbuch der Geophysik." See footnote 13) 



of the nearest station and about four seconds in the case of the 

 farthest. To render the onset of the S wave legible it was necessary 

 to record the tremors at a higher speed than that generally used for 

 earthquakes. Obviously, if the distance is increased to several hun- 

 dreds or thousands of miles, the difference in time of arrival will 

 be increased and, moreover, slight inaccuracies in the determination 

 of the exact instant of arrival of either phase will not greatly affect 

 the velocity determination. 



In order to convey some idea of the manner in which the two 

 types of waves make their appearance on a seismogram, the two 

 groups of records of Figure 1 are shown. The arrival of P and S 

 is indicated for the two seismograms of the first group. These are 



