Seismic wave velocities (compressional and shear) are important parameters for estimating the seismic response characteristics of various geologic units when subjected to strong earthquake ground shaking. Seismic velocities of various units often show a strong correlation with the amounts of damage following large earthquakes and have been used as a basis for certain types of seismic zonation studies. In the current program seismic velocities have been measured at 59 locations 1n the San Francisco Bay Region. This report is the third in a series of Open-File Reports and describes the in-situ velocity measurements at locations 35-59. At each location seismic travel times are measured in drill holes, normally at 2.5-m intervals to a depth of 30 m. Geologic logs are determined from drill cuttings, undisturbed (cored) samples, and penetrometer samples. The data provide a detailed comparison of geologic and seismic characteristics and provide parameters for estimating strong earthquake ground motions quantitatively at each of the sites. A major emphasis of this program is to obtain a detailed comparison of geologic and seismic data on a regional scale for use in seismic zonation. There is a variety of geologic and seismic data available in the San Francisco Bay Region for use 1n developing the general zoning techniques which can then be applied to other areas. Shear wave velocities 1n near-surface geologic materials are of especial interest for engineering seismology and seismic zonation studies, yet in general, they are difficult to measure because of contamination by compressional waves. A comparison of various in-situ techniques by Warrick (1974) establishes the reliability of the method utilizing a "horizontal traction" source for sites underlain by bay mud and alluvium. Gibbs, and others (1975a) present data from 12 holes and establishes the reliability of the method for sites underlain by a variety of different rock units and suggest extending the measurements to a large number of sites. Data collected from the first 12 holes also provide an opportunity for developing a routine and efficient procedure for collection and reduction of the data. Gibbs and others (1975b) report preliminary comparisons of the data with the amplification data recorded from nuclear explosions, and the intensity data for the 1906 earthquake. These comparisons show that correlations exist between the three data sets. Average shear wave velocity in each geologic unit form the basis for the preliminary correlations, however, some of the geologic units (that is, granite and Franciscan Formation) exhibit a wide range of shear wave velocity. It is apparent that other factors must be considered in these units, perhaps near surface weathering or fracture spacing. Work is continuing in this area.