Beginning in October 1976, a program of water-level monitoring of abandoned water wells was initiated in the Palmdale area with the purpose of identifying possible water-level changes premonitory to a major earthquake on the San Andreas fault. In October 1977, the program was extended southeastward along the rift zone to the Valyermo area. In November 1977, the monitoring of water wells along the San Jacinto fault was initiated with the expectation of experiencing a moderate size earthquake while monitoring was in progress. Currently over thirty wells are being monitored. Eleven wells are monitored continuously with Stevens Type F recorders, two of which have been modified to operate with the Caltech Remote Observatory Support Systems (TIMS). The remaining wells are probed weekly, or in some cases semi-weekly or daily, by volunteers. We are endeavoring to improve the volunteer program by increasing the frequency of measurements and simplifying the procedure to minimize measurement errors. Weekly water-level data are displayed on computer-generated hydrographs for each well. Rainfall and earthquakes are plotted on the graphs for direct comparison with water levels. The hydrographs are updated and reviewed weekly. Weekly hydrographs are also prepared from recorder charts on two wells maintained by W. R. Moyle, Jr., of the U.S. Geological Survey, Water Resources Division Office, Laguna Niguel, California. A M5.5 earthquake occurred on 25 February 1980, between the Buck Ridge and San Jacinto strands of the San Jacinto fault zone, about 8 miles east-southeast of Anza. This is the largest earthquake which has occurred on the faults within our monitoring network. The Stevens recorder chart for well number 11S/6E-3N4 located in Borrego Valley about 20 miles southeast of the epicenter indicates that, during a period of about four hours on 21 February, the water level rose 1.5 to 1.6 feet and returned to its prior level. This is one of the most remarkable short-term water-level fluctuations observed during our monitoring program; it occurred about 88 hours prior to the earthquake. This well has been monitored since October 1977 and has had a Stevens recorder since October 1978. Long-term water levels have been remarkably steady compared to those in other wells we are monitoring. This well shows a strong response to earth tides. The Stevens record on a second well in Borrego Valley (11S/6E-1C1) showed a much smaller (0.1 foot) spike in water level at the same time as the spike in well number 11S/6E-3N4. Other continuously recording wells in Anza and Ocotillo Wells showed no identifiable water-level anomalies. The two wells in Borrego Valley which showed possible strain-induced water-level spikes appear to be more sensitive to earth tides than wells in the same area which did not show spikes in water level. We cannot conceive of any nontectonic cause for the spike on the Stevens record of well 11S/6E3N4. A creep event on the Coyote Creek fault about four miles from the well is one possibility. Because the spikes in the two wells in Borrego Valley are unique for the long-term record of these wells, they may represent precursors to the 25 February earthquake. Local seismic activity was also recorded in water levels of the two wells in Borrego Springs with continuous water-level recorders. One well showed peculiar high frequency water-level fluctuations from 22 July to 21 August 1980; since then the record has appeared normal. Several wells in the Palmdale-Valyermo area have shown peculiar changes in water level within the past year or so. The long-term hydrograph of well number 5N/12W-4H1 shows the most unusual behavior. The well showed no response to the 1977 and very heavy 1978 rainstorms, yet began to rise in early 1979 and continued to rise through the dry season of 1979. By comparison, most wells show a more normal seasonal response to rainfall. Eight wells in the Palmdale-Valyermo area have shown water-level changes which are different than would have been predicted from the previous history of water-level changes and seasonal rainfall. A longer period of observation is required to determine whether the changes are anomalous. Changes in the strain pattern and other geophysical phenomena have also been observed in southern California during about the same period as the water-level changes. If the water-level changes are the result of tectonic strain, the mechanism and significance are unknown. However, it is interesting that five of the eight wells with possibly anomalous water-level changes have been identified as good strain meters based on their response to earth tides. Six wells showing an unexpected rise in water level are located west of the earthquake swarm which occurred in 1976-1977 (McNally et al, 1978), whereas the two which show water levels lower than would have been predicted are located east of the earthquake swarm.