Seismic waves in the earth's crust and mantle are known to be sensitive to density contrasts over large volumes of rock, which contrasts tend to cause focusing effects. The end results of such effects observed at seismograph stations are hard to detect unless sufficient arrivals are sampled. It is a common fault to confuse such effects with those of local structures and properties. In a study of teleseismic, short-period (1 s) P-wave travel-time residuals and variations of amplitudes in western North America. R. Butler of the Hawaii Institute of Geophysics has found a high level of correlation to which he attributes qualitatively the focusing and defocusing of seismic waves (Nature, December 15, 1983). The correlation indicating that slow travel times relate to higher, and fast travel times to lower, amplitudes of seismic waves measured in western North America. Conversely, faster travel times and higher amplitudes are generally observed in eastern North America (defined as stations located east of the Rocky Mountain front). Although there may be less attenuation of seismic waves in the upper mantle beneath eastern North America, indications are that the degree of attenuation is highly variable. According to Butler, “On the large scale, the variations between western and eastern North America are probably rooted in lateral differences in temperature. Higher temperatures beneath the tectonically active west produce higher attentuation of P-waves, lower velocities in the upper mantle, and high surface heat flow.” The focusing and defocusing effects of low and high velocity lenses, respectively, may be most effecitive if such lensese are located close to a seismic station. Butler noted that lenses, or anomalous regions, must have dimensions of one or more wavelengths of a P-wave, which translates to a minimum dimension in the earth of 6–8 km (for 1-s period waves). Positive correlations have been observed characteristically over large seismic arrays, suggeting the existence of lenses of several tens of square kilometers in cross section.