124 OLIVER AND DORMAN [CHAP. 8 



thick overlying the higher-velocity mantle while the lower velocity observations 

 indicate about twice this thickness with the upper part having a considerably 

 lower shear velocity than basalt. It is, perhaps, significant that the observations 

 made at the oceanic island stations, except for the Hawaiian data from the 

 direction of the Easter Island rise, give consistently higher group velocities 

 than those made at Berkeley. The effect of the portion of continental path near 

 Berkeley, although its length is short, would be to reduce the group velocity of 

 the waves observed there. Real differences in the data, however, undoubtedly 

 exist, and probably represent real differences in sub-bottom structure. How- 

 ever, because of the large dispersion in the case of continental paths for Love 

 waves in the crustal period range as compared with small dispersion for waves 

 of the same period range for oceanic paths, there is always a danger that the 

 effects of a small portion of continental or near-continental structure will 

 obscure the effect of the oceanic portion of the path. Furthermore, mantle 

 heterogeneities are of considerable importance in this f)roblem. 



Probably the most striking feature of seismic wave trains which have 

 traversed paths which are at least partly oceanic is the long, nearly sinusoidal 

 train of Rayleigh waves with periods slightly greater than about 15 sec (Fig. 1). 

 This appearance results largely from the effect of the water layer on Rayleigh 

 waves propagating in the crust-upper-mantle rock system. Because much of 

 the earth is covered by deep water and because this water layer is remarkably 

 uniform in depth over much of the earth, the predominant period in this wave 

 train varies little with different oceanic propagation paths. 



The theory of Rayleigh -wave propagation in rock overlain by a compressible 

 ocean was considered by Stoneley (1926). He derived the pertinent period 

 equation but did not obtain numerical solutions in the period range of significant 

 dispersion. In 1950, Ewing and Press showed that such a theory could account 

 for the main features of the observed Rayleigh -wave train (Fig. 9). The first 

 theoretical model consisted simply of a water laj^er overlying a solid half-space 

 with the average properties of the crust and mantle. Later, Jardetzky and 

 Press (1953) showed that the effect on the dispersion curves of considering the 

 crust and mantle separately in such a problem was small, as is also shown by 

 the theoretical curves in Fig. 13. The sedimentary layer in both cases was in- 

 cluded as part of the liquid layer. The properties of the Rayleigh waves of short 

 period, which travel at a velocity less than the sound velocity in water, were 

 clarified by Biot (1952). 



Once the propagation of oceanic Rayleigh waves was understood, they were 

 used for exploration of the various ocean basins. Ewing and Press (1952) in a 

 study of waves propagating directly, and also through the antipodal path to 

 Honolulu, Berkeley, Tuscon and Palisades from a shock in the Solomon Islands, 

 found no measurable difference in the basement structure of the Pacific, the 

 Indian and the Atlantic Oceans. This result brought Rayleigh-wave evidence 

 into agreement with the conclusions Wilson had made on the basis of Love- 

 wave data. 



Oliver, Ewing and Press (1955) confirmed these broad conclusions for much 



