Vol. 6 3.] ORIGIN OF THE OCEANS. 347 



we may conclude that the difference, in whatever it consists, does 

 not extend to a depth of one-third, but that it does extend to some- 

 thing like a quarter of the radius from the surface. 



We may conclude, therefore, that there is good ground for 

 supposing that the oceanic basins and continental elevations are 

 not merely irregularities of surface-form, but are accompanied by, 

 and probably related to, differences in the constitution of the sub- 

 stance beneath them which extend inwards to a distance of about 

 one-quarter of the radius. It is not possible to state exactly the 

 nature of this difference, beyond that it is such as to cause a slowing- 

 down of the rate of propagation of the second-phase waves under 

 the oceans. 



More data are required before these conclusions can be posi- 

 tively accepted ; but, if not amounting to proof, the evidence is too 

 strong to be ignored, and its confirmation or refutation must depend 

 on the accident of the occurrence of great earthquakes originating 

 at suitably situated localities. 



Appendix. 



Data utilized in the foregoing Paper. 



The following table (II) embodies the data utilized in the paper. 



Each line refers to the record of a distinct instrument or 

 pendulum. 



The first column gives the name of the station. 



The second column indicates the nature of the instrument. 

 These are classified into types, irrespective of variations in detail, 

 of construction, represented by letters as under : — 



M. = Light horizontal pendulum, with photographic registration, of 



the Milne pattern. 

 R. = Light horizontal pendulum, with photographic registration, of the 

 Rebeur-Paschwitz type. 

 H.P. = Heavy horizontal pendulum, with mechanical registration. 

 V.P. = Vertical pendulum, with mechanical registration. 

 Vert. = Instrument designed for recording the vertical component of the 

 motion. 

 W. = Inverted astatic pendulum of the Wiechert type. 



In the two groups of three columns referring to the Calif ornian 

 and Colombian earthquakes the first column gives the approximate 

 distance, in whole degrees, from the origin, while the other two 

 give the time of arrival of the first- and second-phase waves. In 

 the case of the Californian earthquake they represent minutes after 

 13 h 12 m of April 18th, 1906 ; in the case of the Colombian earth- 

 quake minutes after 15 h 36 m of January 31st, 1906 : these times 

 being those of origin in each case, as near as they are known. 



Figures underlined indicate that the times were determined in 

 the observatory on the original record ; the figures not underlined 

 represent times determined by myself on copies and, therefore, 

 necessarily, of somewhat less accuracy. 



