An earthquake of magnitude ML=6.0 oocurred within the U.S. Geological Survey Yellowstone Park, Wyoming, seismic network on June 30, 1975. The two stations closest to the epicenter were Norris Junction (NJ) (8 km) and Madison Junction (MJ) (20 km). In the 2 years preceding the earthquake, 12 nuclear devices detonated at the Nevada Test Site and Novaya Zemlya were recorded by NJ, MJ, and other stations of the network. The proximity of the two stations to the epicenter, together with their recording of the nuclear shots, gives us an exceptional opportunity to test for P wave travel time and spectral changes precursive to the earthquake. The travel times at NJ and MJ were measured in relation to two stations of the Yellowstone net that were 50--60 km from the epicenter. To estimate the noise in the data reduction scheme, measurements similar to these were made at Hanford. Washington. In general, the relative arrival times at Yellowstone were as stable as those at Hanford, a suggestion that no anomaly larger than fluctuations caused by data reduction (0.09 s) was associated with the impending earthquake. The P wave forms of the shots recorded at NJ also were stable as a function of time, and in 1975 no anomaly was found in the P wave amplitudes.