The United States Geological Survey (USGS) and the Colorado School of Mines (CSM) have initiated a project to locate more accurately all earthquakes recorded by instruments during the period 1900 to 1963. Seismicity for this period (hereafter referred to as historical seismicity, following Lee et al. ) is still poorly understood, even for basic parameters such as earthquake locations (see Figure 1). In some cases this is the result of inherent limitations in the distribution, response characteristics, and timing of the instruments. However, locations for most of the pre-1964 earthquakes are poorly determined simply because modern data analysis techniques have yet to be applied to the available arrival-time observations, which are mainly preserved as printed bulletins and not in a computer-ready digital format. The arduous task of hand-entering these data has prevented the systematic analysis and relocation of historical seismicity.