270 REPORTS ON THE STATE OF SCIENCE, ETC. 



such a core have been confirmed. A distortional wave incident on such a surface 

 from above would give rise to reflected compressional and distortional waves, and to 

 a transmitted compressional wave. The former are denoted by Gutenberg as SjP 

 and S4S. The latter, on emergence, is agam broken up into a compressional and a 

 distortional wave ; these are called S4P4P and S4P4S (in later work the suffix 4 is 

 replaced by c). These transmitted waves with change of type are found on the 

 seismograms at the predicted places and times. So is a wave that emerges after 

 havmg undergone a reflection at the mside of the region. The waves reflected at the 

 ©utside of the core are so spread out as to have only small amplitudes when they 

 emerge, but Prof. V. Conrad has found waves that are capable of this interpretation. 

 The velocity of P waves, as found from near earthquakes, is nearly 7'8 km. /sees., 

 which points to a rock of the nature of olivine within a few tens of kilometres of the 

 surface. Its density would be about 3'4. But at a depth of the order of 1,500 km. 

 such a rock would be so compressed by the weight of the overlying rock that its 

 density would be increased by about unity ; the density of the iron core would be 

 increased still more. Hence Wiechert's numerical results need revision, and it is found 

 that when the increase of density due to compression is taken into account the boundary 

 between the rocky shell and the iron core must be almost exactly where Gutenberg 

 finds the discontinuity of elastic properties to be situated. The failure of the core 

 to transmit S waves is most naturally attributed to its being fluid. Such a view has 

 been denied, because it was found by Kelvm that the earth's tidal properties showed 

 that as a whole it possessed about the rigidity of steel. But seismology now indicates 

 that the rocky shell has a mean rigidity about twice that of steel, and reopens the 

 possibility of a fluid core. In an investigation just published the present writer 

 finds that the tidal data definitely imply a rigidity of the core much lower than that 

 of the shell ; if both the core and the shell were homogeneous, and the core fluid, 

 the tidal yielding would be somewhat more than is observed, but allowance for varia- 

 tion of density witliin the laj^ers would reduce the discrepancy, and it seems very 

 probable that the core is truly fluid. 



Gutenberg's ScPcS. 



Of the waves or rays mentioned above by Dr. Jeffreys, the only one which has 

 received attention up to the present in the discussion of 1918-1922 is that denoted 

 first S4P4S and then ScPoS. By a curious coincidence this had attracted attention in a 

 particular earthquake, October 11, 1922, the night before a letter from Dr. Jeffreys 

 drawmg attention to Gutenberg's work was received. The accord with observation is 

 so close that the following extract from a note to that earthquake in the International 

 Summary may be reproduced here : — 



Note to 1922 Oct. lid. 14h. 49m. 453. 



The readings for S from near A =80° to about A = 110° probably refer to some- 

 thing preceding the true S. The residuals can be represented by the formula : — 



— (A— 80°)x4-6s. 



A 0. C, 0-C. A O. C. O-C. 



I 



