We present results from over 30 yr of precise surveys of creep along the Hayward fault. Along most of the fault, spatial variability in long-term creep rates is well determined by these data and can help constrain 3D-models of the depth of the creeping zone. However, creep at the south end of the fault stopped completely for more than 6 years after the M7 1989 Loma Prieta Earthquake (LPEQ), perhaps delayed by stress drop imposed by this event. With a decade of detailed data before LPEQ and a decade after it, we report that creep response to that event does indeed indicate the expected deficit in creep. ¿ 2001 American Geophysical Union