The results of mapping locked and unlocked patches on the Hayward fault using the local recurrence time method are compared with models derived from surface creep and geodetic measurements. The data used were the M>1.0 earthquakes in the USGS catalog since 1981.0, located within 5 km of the fault trace. Along the central and southern segments, agreement with the other methods is satisfactory. The 1868 rupture, except for its two ends, is classified as a locked segment by our method. In the northern part of the Hayward fault, the data are poor and the results of the three methods differ. Short local recurrence times in the northern part suggest that an asperity exists there, which is capable of an M6.5 class event. ¿ 2001 American Geophysical Union