ON SEISMOLOGICAL INVESTIGATIONS. 



41 



arcs. In order that this may be possible the angles of incidence and 

 reflection must be equal, and these angles depend essentially on SP, 

 the difference of time for (say) 1°; so that SP must have the same 

 value for a large arc as for a small one. "With the adopted tables this 

 does not occur. The values of 5P steadily diminish, as may be seen 

 by the following figures: — 



A= 0° 10° 



s. s. 



5P = 15-5 15 



20° 



s. 

 12 



30° 



s. 

 10 



40° 



50° 



s. 



7 



60° 70° 80^ 90° 

 s. s. s. s. 



7 6 6 5-5 



100° 



s. 

 5 



If these figures are correct we cannot explain PX in the way 

 suggested. It is now proposed to challenge the correctness of the 

 figures between 0° and 45°, leaving those > 45° practically unaltered. 

 The nature of the proposed change is best seen in diagrammatic form 



