Key points are not available for this paper at this time.
By using a pulse method, attenuation and velocity measurements have been made for aluminum and glass rods in the frequency range from 2 to 15 megacycles. The sound pulses are generated by crystals waxed to the surface of the rod. This wax joint limits the band width of the transmitted pulse and measurements are made using long pulses which approach steady state conditions. The reflected pulses show evidence of several normal modes which can be minimized by using specially shaped electrodes. Longitudinal waves show delayed pulses of smaller magnitude that are caused by the longitudinal wave breaking up into reflected longitudinal and shear waves at the boundary. This effect is small if the diameter of the rod is 20 wave-lengths or more. The measured losses for aluminum rods show a component proportional to the frequency and another component proportional to the fourth power of the frequency. The first component is the hysteresis loss found for most solid materials. The component proportional to the fourth power of the frequency is caused by Rayleigh scattering losses which are the result of differences in the elastic constants between adjacent grains caused by changes in orientation. Calculated scattering losses agree quite well with the measured values. The fourth-power scattering law holds quite well until the grain size is equal to one-third of a wave-length. For higher frequencies the scattering loss increases more nearly with the square of the frequency. Glasses and fused quartz have a loss directly proportional to the frequency, showing that any irregularities must be of very small size.
Mason et al. (Thu,) studied this question.