A wide-band (0.03 to 10 cps) accelerometer system was used for recording seismic motions at a distance of 18 km from the underground nuclear explosion Benham. The records clearly show a unique difference in spectral structure between the radial and transverse components of horizontal motions. The source of the transverse component appears to be a more efficient radiator of long waves than the source of the radial component, strongly suggesting that the transverse component motion may not be due to transformation of P, SV or Rayleigh waves but due to an additional source which was triggered by the explosion. Combining the result from Love wave spectra observed at Weston, Massachusetts, we arrived at the following conclusions. (1) A faulting was triggered by the explosion in such a timing that the shear waves from the starting point of fault arrived at our station 3 seconds after the arrival of the explosion-generated shear waves. (2) The size of faulting estimated from long-period surface waves is such that the product of fault area and dislocation is about 1013 cm3. This value is consistent with the length, displacement and depth of faulting as suggested from field observations and aftershock studies.
An alternative theory which postulates a release of high pre-existing stress (400 to 3000 bars) by the formation of a crushed region can account for the seismic moment estimated from long-period Love waves observed at a long distance. This theory, however, does not harmonize with observations on surface faulting and aftershocks.