Data of sound attenuation constant in sand is nesessary for designing detection apparatus for underground buried pipes by the sound echo method. The measurement of the sound absorption constant in sand with a continuous or pulsed sinusoidal wave would be difficult to perform due to obstruction by electrical induction directly from the transmitting system to the receiving system. The authors devised the method to measure the frequency property of attenuation constant in sand with an impulsive sound wave. In this method, the impulsive sound is radiated from the electromagnetic induction type sound source placed on the ground surface and the sound pressure propagated in sand is received by, a piezoelectric microphone buried at a depth of 1 m just below the sound source, and the received output signals are analyzed with 1/3-Oct. band pass filters. Similar measurements are performed varing the distance from the sound source to the microphone by digging down into the sand. The distance was varied from 1 m to 20 cm. By the variety of each frequency spectrum component of recieved output signals at the varied distance, the sound attenuation constant in sand is determined. In this study, the experiment to obtain the sound attenuation constant in the model sand bath (11 m × 8 m × 2 m) filled with wet mountain sand was performed by the method mentioned above. Soil tests to define the fundamental property of wet mountain sand were carried out. The results of the soil tests are shown in Fig. 1 and Table 1. Fig. 2 shows the block diagram of the measuring apparatus of radiated sound pressure. The received signals were recorded by a magnetic tape recorder. In this experiment, the driving current frequency f_c was increased from 200 Hz to 950 Hz by decreasing the condenser capacitance from 180 μF to 10 μF, and the input energy was kept at 250 joule by increasing the charging voltage as the capacitance is decreased. Fig. 3 shows the block diagram of the apparatus for frequency analysis. The received output signals recorded by a magnetic tape recorder were processed using 1/3-Oct. band pass filter during reproduction. In order to obtain an accurate sound attenuation constant, it is necessary to use received output signals with large S/N ratios. Accordingly, we adopt the result obtained by frequency analysis in case of three kinds of f_e in each frequency band, as shown in Table 2. Fig. 4 shows, as an example, the attenuation constant in frequency band at 355 to 450 Hz. Because measurement points are fairly scattered in the figure, the attenuation constant was obtained by the method of least squares. Besides, each measurememt value was normalized with the value in the case where the distance from the sound source to the microphone was 20 cm, or the sound attenuation constant in sand was the smallest. Attenuation constants of other frequency bands were obtained by a similar method. The results are shown in Fig. 5. We also found by the experiment in a model sand bath that the attenuation constant in wet mountain sand (Particle size: 0. 4 mm, Specific gravity of particle: 2. 45, Porosity: 34. 45%, Water content in percentage of dry weight: 11. 6%) is expressed as α&sime;3. 1×10^<-5>×f^2 (dB/m) where f is the frequency in Hz.