The Hayward fault, thought a likely source of large earthquakes in the next few decades, has generated two large historic earthquakes (about magnitude 7), one in 1836 and another in 1868. We know little about the 1836 event, but the 1868 event had a surface rupture extending 41 km along the southern Hayward fault. Right-lateral surface slip occurred in 1868, but was not well measured. Witness accounts suggest coseismic right slip and afterslip of under a meter. We measured the spatial variation of the historic creep rate along the Hayward fault, deriving rates mainly from surveys of offset cultural features, (curbs, fences, and buildings). Creep occurs along at least 69 km of the fault's 82-km length (13 km is underwater). Creep rate seems nearly constant over many decades with short-term variations. The creep rate mostly ranges from 3.5 to 6.5 mm/yr, varying systemically along strike. The fastest creep is along a 4-km section near the south end. Here creep has been about 9mm/yr since 1921, and possibly since the 1868 event as indicated by offset railroad track rebuilt in 1869. This 9mm/yr slip rate may approach the long-term or deep slip rate related to the strain buildup that produces large earthquakes, a hypothesis supported by geoloic studies (Lienkaemper and Borchardt, 1992). If so, the potential for slip in large earthquakes which originate below the surficial creeping zone, may now be 1/1m along the southern (1868) segment and ≥1.4m along the northern (1836?) segment. Substracting surface creep rates from a long-term slip rate of 9mm/yr gives present potential for surface slip in large earthquakes of up to 0.8m. Our earthquake potential model which accounts for historic creep rate, microseismicity distribution, and geodetic data, suggests that enough strain may now be available for large magnitude earthquakes (magnitude 6.8 in the northern (1836?) segment, 6.7 in the southern (1868) segment, and 7.0 for both). Thus despite surficial creep, the fault may be ready for the recurrence of large earthquakes today. However, the timing (Williams, 1992) and size of future events may vary greatly due to uncertainties in the tectonophysical model assumed for the Hayward fault within the greater San Andreas fault system (Lisowski and Savage, 1992).