There are indications of two low-velocity layers in the earth's crust and mantle, one, in the continents, at depths between roughly 10and 20km, the other at depths from roughly 60 km to about 150 km for P and about 250 km for S. The first is called here the lithosphere channel, the second the asthenosphere channel. If an earthquake originates in a low-velocity layer, waves leaving the source downward cannot enter the layers with relatively high velocity above the channel and reach the surface unless they reach a layer below the source with a wave velocity at least equal to the maximum velocity above the source. If an earthquake occurs in the lithosphere channel, waves starting downward must consequently reach at least the depth where the velocity equals the maximum reached above the channel (which is about 6¾ km/sec in Southern California) in order to arrive at the surface and to be recorded. The travel-time curve of such waves with a deepest point above the Mohorovičić discontinuity1 may easily be misinterpreted as caused by waves refracted at a (nonexisting) discontinuity below the source where the velocity is believed to increase suddenly to about 7 km/sec.
A large fraction of the energy in earthquakes originating in a low-velocity layer remains in the channel. In earthquakes with a source above the lithosphere channel, much more energy is transmitted to the epicentral area than in those originating in this channel, and relatively great damage may result.
New evidence for the asthenosphere channel is discussed. It is based mainly on travel times and amplitudes of longitudinal and transverse waves on records of the largest Kern County, California, earthquakes in July, 1952, for which epicenters are established within about 4 km, and origin times within about a quarter of a second.