Speculations are made regarding the significance of the well-known observation that seismic reflection energy is usually in the frequency range of from 20 to 100 cycles per second. The general absence of reflected energy below 20 cps is attributed to the fact that the wavelengths of seismic waves in this frequency range are becoming large compared to the thicknesses of reflecting beds; accordingly, the reflection coefficients are low with the results that the geologic section appears more or less homogeneous, the low frequency energy is unweakened by reflections, is transmitted efficiently, and can only return to the surface by refraction.As the frequency is increased the wavelengths become comparable to the vertical discontinuities represented by stratification and more efficient reflection takes place with the result that reflected energy is returned and detected at the surface. At still higher frequencies the wavelengths become comparable to small inhomogeneities distributed at random throughout the geologic section and the energy is therefore diffused and scattered to such an extent that transmission into the earth is limited. This weakening of the main wave front by scattering, plus the weakening by absorption processes involving viscous and solid friction, constitute an effective cutting off of high frequency transmission. The high frequency scattered energy diffuses back to the surface and appears on the seismogram as 'hash,' unless eliminated by filters, or is absorbed before it reaches the surface.Such a speculative picture of seismic energy propagation accounts qualitatively for (1) the continuous reception of random energy that is always superimposed upon the reflection energy, (2) the tendency for deep reflections to be of lower frequency than shallow reflections, and (3) the fact that theoretical considerations of absorption do not always account for known attenuation of high frequency seismic energy.