THE San Andreas fault near Parkfield, California has experienced moderate to large earthquakes about every 20 years since 1881, the most recent of which occurred in 1966(1-3). Statistical recurrence and precursory quiescence models(1,4,5) led to a prediction of the next event for 1992 or before, but this event has not yet occurred, Here I present a model(6) that attributes the ’non-occurrence’ of the expected earthquake to the influence of adjacent thrusting on the rate of fault weakening from increasing pore pressures within the fault zone, Reductions in fault-normal strain rates(7-9), and the onset of seismic quiescence along the San Andreas fault, both coincide with adjacent blind thrust faulting in the 1982-85 earthquake sequence of New Idria, Coalinga and Kettleman Hills, These events ’turned off’ a compaction-induced weakening mechanism and extended the (proposed) seismic cycle at Parkfield, A model simulation of this effect shows a strong correlation in space and time of calculated and observed earthquake hypocentres, and shows that observed seismicity patterns from 1969-95 can be explained by reduced compaction rates brought on by adjacent thrusting, The model may have applications to other linked thrust/strike-slip fault systems, such as are found in southern California and other transpressional regions.