20 HISTORY OF THE OCEANS 



multiplicity of arrivals, and an increase in the apparent surface 

 velocities to about 9 and 5 km/sec for P and S, respectively. 

 Lehmann (1934, 1952, 1955) concluded that "there is likely to be 

 a layer in which there is a decrease in the velocity of 5 waves." 

 Gutenberg (1948) concluded from studies of more detailed data, 

 that in Southern California there is a low-velocity layer at a depth 

 of 80-100 km affecting both P and S. Gutenberg (1953) introduced 

 a new method which made use of the fact that the velocities at 

 the depth of an earthquake focus may be determined by the slope 

 of the tangent of the travel time curve at the inflection point. Data 

 from small epiccntral distances were available for intermediate 

 and deep earthquakes, but mainly for Japan alone. He reported 

 that elastic wave velocities "show a clear decrease with depth 

 with a minimum at a depth of roughly 100 km for longitudinal 

 waves and 150 km for tranverse waves" and presented evidence 

 for a strong increase of compressional and shear velocities below 

 200 km (see Fig. 7). This method was even more restricted in its 

 areas of applicability than the earlier one, but it was the first 

 capable of giving direct measures of velocity variation within a 

 low-velocity zone. The previously used methods could only 

 indicate such a layer by a "shadow zone" in which the data 

 obtained seemed very poor, and by a delay between the travel-time 

 curve for large distances, compared to that extrapolated from 

 small distances. 



Lehmann (1955) reported that in northeastern North America 

 the near branch of the 5 travel-time curve extended to 14°. 

 Contrasting this to the 5° limit which she had found in Europe, 

 she inferred regional differences in upper mantle structure, and 

 postulated that a "soft" layer, having reduced shear wave velocity, 

 was somewhat deeper in northeastern North America than in 

 Europe. However, Lehmann believed that no low-velocity channel 

 exists for P since there is no delay between the near and far 

 boundaries of the time curve, the small amplitude of P at distances 

 less than 15° being explained by the straightness of the time 

 curves. (Compare Fig. 7 for Gutenberg's results.) 



Lehmann (1954) discussed the relation of Gutenberg's low- 

 velocity zone to the strain rebound studies of Benioff (1954), 



