We have used over 16,000 compressional wave travel times recorded on 11 ocean bottom seismometers to determine the magnitude of P-wave heterogeneity and anisotropy in the upper 2 km of the oceanic crust in an ~150 km2 area around Hole 504B, the deepest hole drilled into the oceanic crust. Our best fitting one-dimensional, isotropic upper crustal velocity model for the 150 km2 area around the drill site is similar to velocity models at Hole 504B determined using downhole sonic logs, vertical seismic profiles, oblique seismic experiments and conventional refraction profiling. There is no evidence seismic layer 2 is anomalously thin at Hole 504B, suggesting that the occurrence of the seismic layer 2/3 boundary within the upper sheeted dike section, which has been documented in Hole 504B, is a general feature of this 5.9 Ma crust. On the scale of a few kilometers and averaged over seismic wavelengths of a few hundred meters the heterogeneity in upper crustal P wave velocity around Hole 504B is comparatively small. Within the 571-m thick extrusive section maximum P wave velocity differences are ~300--400 m s-1 (a 6%--8% velocity anomaly). These velocity differences decrease to 1 km into the crust. The highest upper crustal velocities are found beneath basement ridges west and south of Hole 504B, while the lowest upper crustal velocities occur beneath the flanking basement troughs. There is an azimuthal dependence to the travel time residuals (slow direction N-S) which is consistent with 2%--4% crack-induced anisotropy in the upper crust. The magnitude of upper crustal heterogeneity observed near Hole 504B is significantly less than that found in zero-age crust at the Mid-Atlantic Ridge but comparable to that observed off-axis in crust formed at faster spreading ridges. ¿ 1998 American Geophysical Union