SECT. 1] 



THE MANTLE ROCKS 



107 



near the M discontinuity would be difficult to detect. The three lowest values of 

 Vs in Table I were determined by refractions very near critical angle and thus 

 must have measured the uppermost part of the mantle. The other measurement 

 (4.6 km/sec) was based on arrivals appreciably beyond the critical range, and 

 this is in agreement with the velocity determined by the study of surface-wave 

 dispersion (Dorman, Ewing and Oliver, 1960). 



To test for a relationship between sub-crustal compressional-wave velocity 

 and depth, the graph shown in Fig. 3 was plotted. The data were taken from 

 Atlantic and Pacific profiles in which the water depth was between 4 and 6 km, 

 trenches being avoided. There appears to be a trend of the data in the direction 

 of increasing velocity with increasing depth. Some of this trend may be the 

 effect of lines having been drawn on time-distance graphs which, although 



20 



I 5 



H 10 



Q 5 



O PACIFIC 

 □ ATLANTIC 



7.5 8.0 8.5 



COMPRESSIONAL WAVE VELOCITY (KM/SEC) 



Fig. 3. Graph of mantle compressional wave velocities versus depth of the M discontinuity. 



passing through the centers of gravity of the points, had incorrect slopes. For 

 typical oceanic profiles errors produced by misalignment will cause the data 

 to trend approximately in the direction indicated by the line in Fig. 3. This line 

 fits the points about as well as any other, hence one could deduce that both the 

 high and low velocities are improperly determined values and that there is no 

 relationship of velocity with depth. The observed trend in Fig. 3 would, however, 

 be expected if there is a velocity gradient in the upper mantle. Evidence from 

 shear waves was cited previously as possibly indicative of a gradient. 



In Fig. 4 are plotted crustal thicknesses versus elevation of the sea floor 

 from about 90 of the better published profiles in the Pacific and Atlantic 

 Oceans, the Caribbean Sea and the Gulf of Mexico. For this graph the data were 

 selected if the sub-crustal velocity was 7.7 km/sec or higher, and, as before, 

 trench profiles were avoided. Although there is a great amount of scatter, a 

 trend is indicated which could extend smoothly to continental crustal thick- 

 nesses and appropriate depths of the M discontinuity (30-40 km crust for ele- 

 vations of 0-2 km). 



4. Summary 



Diagrams showing observed velocities higher than 7 km/sec have peaks at 

 8.1-8.2 km/sec indicating this as the average value of compressional waves in 

 the upper mantle in deep -ocean areas. Some deviations above and below the 



