We have conducted 8 precise seismic surveys near Hollister, CA, over a period of 1 y in an attempt to detect the traveltime variation caused by the solid-earth tidal stress. The surveys were conducted along a 600 m baseline located in quartz monzonite hills 2 km west of the San Andreas fault. A 656 cm3 air gun fired in a mud-filled pit 2 m deep provided a repeatable seismic source. The signals from two 2.3 Hz vertical-component geophones 600 m apart were digitized at a nominal rate of 600 samples/s by two cassette recorders modified for precise synchronization of data sampling against a master clock. Each survey consists of ~100 traveltime measurements over the 12 h period between 6 p.m. and 6 a.m. local time; the time frame of each experiment was limited by daytime cultural noises. Analysis of traveltime variation is done either by timing of amplitude extrema or by cross-correlation of a waveform contructed from the digital data by a cubic-spline interpolation. Fractional error of the repeatablility of traveltime measurement is typically ¿3.3¿10-4 for the first high-frequency, large amplitude arrival following the direct body waves. The first survey, conducted at a spring tide in August, 1981, showed a variation of Δt/t -2¿10-3 and correlated in time with the extensional tidal strain component along the baseline direction. The next two surveys, conducted at two neap tides, showed variation of Δt/t ~6¿10-4 and also correlated with the same tidal strain component. However, the other 5 surveys conducted after the onset of 1981 rainy season and into the 1982 dry season, 4 at spring tides and 1 between a spring and a neap tide, showed traveltime constant to within 1 standard deviation. These results corroborate only partially the previously reported tidal stress variation of traveltimes in the shallow crust.