Least-squares adjustments of observations of waves of the P groups at central and southern California stations are used to obtain the speeds of various waves. Only observations made to tenths of a second are used. It is assumed that the waves have a common velocity for all earthquakes. But the time intercepts of the travel-time curves are allowed to be different for different shocks. The speed of P̄ is found to be 5.61 km/sec.±0.05. The speed for S̄ (founded on fewer data) is 3.26 km/sec. ± 0.09. There are slight differences in the epicenters located by the use of P̄ and S̄ which may or may not be significant. It is suggested that P̄ and S̄ may be released from different foci.
The speed of Pn, the wave in the top of the mantle, is 8.02 km/sec. ± 0.05. Intermediate P waves of speeds 6.72 km/sec. ± 0.02 and 7.24 km/sec. ± 0.04 are observed. Only the former has a time intercept which allows a consistent computation of structure when considered a layer wave. For the Berkeley earthquake of March 8, 1937, the accurate determination of depth of focus was possible. This enabled a determination of layering of the earth's crust. The result was about 9 km. of granite over 23 km. of a medium of speed 6.72 km/sec. Underneath these two layers is the mantle of speed 8.02 km/sec. The data from other shocks centering south of Berkeley would not fit this structure, but an assumption of the thickening of the granite southerly brought all into agreement.
The earthquakes discussed show a lag of Pn as it passes under the Sierra Nevada. This has been observed before. A reconsideration of the Pn data of the Nevada earthquake of December 20, 1932, together with the data mentioned above, leads to the conclusion that the root of the mountain mass projects into the mantle beneath the surface layers by an amount between 6 and 41 km.