Sputtered quartz crystals of known natural frequency were used as sources of high frequency sound in a Pierce acoustic interferometer. The interfering sound waves react on the crystal driver in such a way as to change periodically the plate current, ${I}_{p}$, when the mirror is displaced a half wave-length or more. The galvanometerdeflection, mirror-displacement curves show that multiple reflections take place at the movable mirror and at the crystal surface. The general agreement of these curves and those which represent the resultant pressure amplitude in the gas at the crystal surface as ordinates and the mirror displacement as abscissa shows that the changes in galvanometer deflection are a direct function of the changes in pressure amplitude. Check runs made with a torsion vane detector indicate that this function expresses a proportionality between the deflection change and the square of the change in the pressure amplitude, provided the latter change remains small.Observed facts indicating a variation of wave velocity with intensity are: 1st, the asymmetry of the deflection peaks; 2nd, the change in the type of this asymmetry with increasing mirror displacement; 3rd, non uniform spacing of deflection maxima; 4th, sudden changes in the logarithmic decrement of the ${I}_{p}$-maxima. The precision in the value of the wave velocity is diminished considerably by this velocity variation.Measurements are made on air and C${\mathrm{O}}_{2}$ at frequencies ranging from ${3(10)}^{5}$ to ${14(10)}^{5}$ cycles per sec. The apparent velocity obtained from the average spacing of ${I}_{p}$-maxima is slightly higher than the commonly accepted value for audible frequencies.The absorption by air and by C${\mathrm{O}}_{2}$ increases with frequency through this range, C${\mathrm{O}}_{2}$ being nearly opaque at ${14(10)}^{5}$ vps. Lebedew's constant, $A$, has a value of 0.00037 for air at 20\ifmmode^\circ\else\textdegree\fi{}C. The observed value obtained with the interferometer at ${12(10)}^{5}$ vps. was 0.00039 which agrees well with a torsion vane determination. At the lower frequencies the observed value is higher, partly because of the multiple reflections. The torsion vane arrangement used with C${\mathrm{O}}_{2}$ and with the highest frequency in air was not suitable for the lower frequencies in air. The observed value of $A$ for C${\mathrm{O}}_{2}$ was 0.0073 at ${12(10)}^{5}$ and 0.012 at ${3(10)}^{5}$ vps. A careful torsion vane determination at ${6.555(10)}^{5}$ gave 0.0096. Abello's data at ${6.12(10)}^{5}$ extrapolated to 100 percent C${\mathrm{O}}_{2}$ give 0.0063. The humidity has a marked effect on the absorption in C${\mathrm{O}}_{2}$.