In June 1973, seismic noise measurements were made in Long Valley, California, as part of the U.S. Geological Survey's geothermal investigations. Spatial variation of the average noise power shows high levels of noise extending over most of the eastern half of the Long Valley caldera. Since the noise high is almost similar in extent to the soft sedimentary Owens River basin, it is possible that ground amplification of seismic waves is at least partially responsible for the noise anomaly. Two lines of evidence indicate that geothermal noise may be present in Long Valley. (1) Relative amplification of teleseismic waves over soft ground, with respect to a reference station on hard rock, is about 12 dB. The noise anomaly, however, is at least 12 dB higher than this value. It is therefore difficult to explain the anomaly by postulating ground amplification of regional noise, thus indicating that a noise source may be present in the area of the anomaly. At wave frequencies below 2 Hz, river and cattle noise do not contribute much to the anomaly. (2) Group velocities of seismic noise, measured by using arrays, are in general quite low except at a few stations along the southern edge of the noise anomaly. The wave azimuths in the low-velocity areas show random propagation, whereas azimuths associated with the high-velocity waves point to the area where surface geothermal phenomena are found. The high-velocity waves also have frequencies below 2 Hz. If a noise source is present under the southern edge of the sedimentary basin, it could excite the basin much more than it does the hard ground directly above it and thus produce the observed noise anomaly.