During the past few years there have been numerous reports of changes in coda wave attenuation occurring before major earthquakes. These observations are important because they may provide insight into stress-related structural changes taking place in the focal region prior to the occurrence of large earthquakes. The results of these studies led us to suspect that temporal changes in coda wave attenuation might also accompany volcanic eruptions. By measuring power decay envelopes for earthquakes at Mount St. Helens recorded before, during, and after an eruption that took place during September 3--6, 1981, we found that coda Q-1 for frequencies between 6 and 30 Hz was 20--30% higher before the eruption than after. The change is attributed to an increase in the density of open microcracks in the rock associated with inflation of the volcano prior to the eruption. Q-1 was found to be only weakly dependent on frequency and displayed a slight peak near 10 Hz. The weak frequency dependence is attributed to the dominance of intrinsic attenuation over scattering attenuation, since it is generally accepted that intrinsic attenuation is constant with frequency, whereas scattering attenuation decreases strongly at higher frequencies. The weak frequency dependence of Q-1 at Mount St. Helens contrasts with results reported for studies in nonvolcanic regions. The peak in Q-1 near 10 Hz at Mount St. Helens is attributed to the scale length of heterogeneity responsible for generating backscattered waves. Results for nonvolcanic regions have shown this peak to occur near 0.5 Hz. Thus a smaller scale length of heterogeneity is required to explain the 10-Hz peak at Mount St. Helens. ¿ American Geophysical Union 1988