Strain seismograms of the Montana shallow earthquake of August 17, 1959 recorded at Isabella, California have the wave pattern predicted in 1904 by Lamb for a surface pressure pulse. This is equivalent to a dipole source such as given by the vertical bilateral fault slip which was observed at the surface. Similar recordings made at Ñaña of the deep Peruvian earthquakes of August 19 and 30, 1961, (Δ = 600 ± km, h = 600 ± km) have the pattern calculated by Pekeris for a buried vertical downward force in the form of a step in time. It thus appears that these shocks were generated by a sudden volume contraction at the focus which could be the result of a sudden change of state.
Failure of conventional seismographs to record the patterns calculated by Lamb and by Pekeris is due to their low sensitivity to the very long period ground movement components involved, and their relatively high sensitivity to the short period components which are rendered oscillatory by the departure of the crustal characteristics from the homogeneous half space assumed in the theoretical computations.