An electrical resistivity monotoring experiment was conducted on a portion of the San Andreas fault near Hollister, California, in 1979 and earlier 1980. Receiving dipoles up to 20 km away from a 140-A 2-km transmitting dipole were uses to monitor apparent resistivities with accuracies as high as 0.1% for some of the sites. Several of the transmitter-receivers clearly show long-term changes in resistivity greater than the data errors; annual variations at a site outside the fault zone exceed 10%, and sites in the fault zone show month to month changes of a few percent. There are also significant variations in the vicinity of the transmitter which would be expected to produce correlated changes at other sites, but this was not observed in the data. There were no earthquakes of magnitudes greater than 3.0 during this period of high-accuracy monitoring so that no conclusions can be drawn about changes in resistivity associated with major strain events. Future experiments to relate resistivity changes to earthquakes must not only have high sensitivity but must also be designed to distinguish such an event from normal background fluctuations that we have measured. Unless the resistivity variations associated with the event are of the order of several percent their detection will be a truly formidable task.