The San Andreas fault zone has been a very significant source of major California earthquakes. From 1812 to 1906 it generated four major earthquakes of M ∼7 or larger in two pairs on two major portions of the fault. A pair of major earthquakes occurred on the central to southern region, where the 1857 faulting overlapped the 1812 earthquake faulting. A pair of major earthquakes occurred on the northern region, where the 1906 faulting overlapped the 1838 earthquake faulting. Also, earthquakes of M ∼7 occurred in the San Francisco Bay area on the Hayward fault in 1868 and the Santa Cruz Mountains near Loma Prieta in 1989 and on the Imperial fault near the border with Mexico in 1940.

The 1838 earthquake's damage effects throughout the Bay area, from San Francisco to Santa Clara Valley and Monterey, were unequalled by any historical earthquake other than the 1906 event. This, and numerous strong possible aftershocks during the following 3 years in the San Juan Bautista vicinity, suggest 1838 faulting from San Francisco to San Juan Bautista.

Cycles of seismicity and quiescence were associated with the Bay area earthquakes of 1868, 1906, and 1989. The 1868 earthquake on the Hayward fault was preceded by 12 earthquakes of M ≥5.5 from 1855 to 1866, within 60 km of the Hayward fault, and was followed by 13 quiet years. The 1906 San Andreas fault event was preceded from 1881 to 1903 by 18 earthquakes of M ≥5.5 and was followed by quiescence, with only three earthquakes of M ≥5.5 until 1954. The Bay area has been seismically quiet at the M ≥5.5 level since the 1989 Loma Prieta earthquake and its 1990 aftershocks, which contrasts with the 10 years before 1989, when five M 5.5–6.2 events occurred. The Loma Prieta earthquake is of similar magnitude to the 1868 Hayward event and could be followed by a similarly short quiet period.

The 1857 earthquake had immediate foreshocks in the Lonoak–Bitterwater region ∼50 km northwest of Parkfield. In the northern end zone of the 1857 rupture, extending southeast from Bitterwater ∼70 km to Parkfield, the rate of seismic moment release has decreased with time since 1857. This may reflect the decay with time of the stress loading due to the ∼9 m 1857 fault displacements ∼80 km southeast of Parkfield and explain why the predicted earthquake, which was based on the assumption of regular recurrence of Parkfield earthquakes, has not yet occurred.

The extent of the 1812 earthquake fault rupture is not well defined. Jacoby et al. (1988) estimated that it extended ∼170 km from Cajon Pass to Tejon Pass. Based on this estimate, we present the hypothesis that the rupture occurred in two segments in December 1812. The eastern segment generated the 8 December earthquake that damaged San Juan Capistrano, San Gabriel, San Fernando, and San Buenaventura. Thirteen days later the western segment ruptured generating the earthquake that damaged San Fernando and San Buenaventura again, as well as Santa Barbara, Santa Ynez, and Purisima Concepcion.