Conflicting results on tidal variation of seismic travel times exist in the literature. With improved methods, we have conducted a seismic survey at a Massachusetts granite quarry. The survey was conducted in the intervals (230d 23h, 231d11h) and (231d22h, 233d10h), 1983 (U.T.) along a 148 m baseline situated in nearly flat topography. The source for the present experiment was an air gun placed in a mud-filled pit. Travel times for the first five body wave extrema were analyzed. The results are: I. The 1st, 2nd, 4th, and 5th extrema show travel time variations; the 3rd extremum remains constant throughout the experiment. The magnitude of the fractional travel time variation, Δt/t, ranges from 0.5 to 0.9%. II. Changes in travel times of the 1st and 2nd extrema correspond to opposite changes in travel times of the 4th and 5th extrema. III. Two sets of nearly orthogonal joint systems are observed in the granite; the theoretical tidal strain in the direction perpendicular to the nearly vertical joints matches the travel time variations of the 1st and 2nd extrema whereas the tidal strain in the direction perpendicular to the nearly horizontal sheets matches the travel time variations of the 4th and 5th extrema, when a 4-hr delay is introduced for all the tidal strains. These results are interpreted in terms of the velocity changes of seismic rays as the two joint systems open and close due to the tidal stress.