44 



REPORTS ON THE STATE OP SCIENCE. — 1916. 



The ' final ' corrections could be improved by a slight change in the 

 distance from Pulkovo. In the next example (May 8): — 



Table VIII. 

 1914 May 8^ ISi" 2'° 0« : 37°-7 N. 15°-0 E. {Shide Determination). 



[Ekaterinburg is probably PRi, which arrives 72' after P.] 



the difference between Konigsberg and Baku is only partly com- 

 pensated by om* corrections, which may be fairly set against the 

 apparent over-compensation of the example preceding. A change of 

 epicentre 1°"2 in the azimuth SIC^ (which is the best that a rough 

 investigation suggests) cannot even now bring Konigsberg and Beku 

 quite together. 



These examples (out of a number which have been already examined) 

 will suffice to show how elaborate an investigation will probably be 

 required to decide the point fully ; moreover, it must be remembered 



(a) That the precise form of the curve of correction is still to be 

 determined, that above given being purely tentative. 



(b) That the observations of S must also be taken into account. 

 If the SP curve has an oscillation of the kind indicated, the cause must 

 be sought in the arrangement of density layers as we descend into the 

 earth; and this will affect S also. The chord of an arc of 30° lies 

 within 150 miles of the surface of the Earth, and of an arc of 15° 

 within 40 miles, so that the anomalies lie at no gi-eat depth, and may 

 reasonably be placed at the limit of the Earth's ' crust.' 



Without claiming more than that a case has been made out for 

 further inquiry (which will be conducted as opportunity offers), let us 

 now return to the phenomenon which suggested the hypothesis and 

 see how the figures given provisionally will fit the facts. We adopt 

 for time of P up to A = 45° the New Values of Table V., and from 

 A =45° onwards the figures of the table printed in the Shide Bulletins. 

 Let us now add together the times for arcs of 20°, 21°, 22°, &c., to 

 arcs of 120°, 119°, 118°, &c. : — 



