Acoustic Emission (AE) event recordings were made with 30 and 110 kHz resonant piezoelectric transducers and a 1 Hz to 20 kHz accelerometer at the San Francisco Presidio during the times of the Lake Elsman earthquake and its largest aftershock on August 8, 1989, the Loma Prieta earthquake aftershocks of October 17, 1989, and a small earthquake on the Hayward fault in November 4, 1989. Each earthquake generated abrupt increases in AE production at the time of arrival of the seismic waves that were clearly above the background rates. Corresponding strain changes recorded on strainmeters in the same thermally insulated test bunker are from a few tens to a few hundreds of nanostrain, with strain rates from ^SxicrV1 to slorV1. Sensitivity of AE to tidal strains was checked by comparison of the 30 kHz background AE and the strain recorded by one of the USGS strainmeters during a 30-day interval in April and May, 1991. At this site tidal strains are greater than thermoelastic strains at semi-diurnal periods. Power spectral densities and coherence were computed for the strain and AE data. Dominant spectral peaks at approximately 24 hours (Oi, klt Si, MI, etc., tidal periods) and approximately 12 hours (O2, k2, S2, M2 etc., tidal periods) are clear in the strain spectrum and suggested in the AE spectrum. A weak correlation between the rate of AE production and the observed strain is observed at these neardiurnal and semidiurnal periods. This suggests that AE production occurs at strain levels and strain rates even lower than those estimated to be caused by passage of seismic waves from the 1989 earthquakes. We believe these to be the lowest strain levels and strain rates for which acoustic emission has been reported.