Over the past several years, many investigators have argued that static stress changes caused by large earthquakes influence the spatial and temporal distributions of subsequent regional seismicity, with earthquakes occurring preferentially in areas of stress increase and reduced seismicity where stress decreases. Some workers have developed quantitative methods to test for the existence of such static stress triggering, but no firm consensus has yet been reached as to the significance of these effects. We have developed a new test for static stress triggering in which we compute the change in Coulomb stress on the focal mechanism nodal planes of a set of events spanning the occurrence of a large earthquake. We compare the statistical distributions of these stress changes for events before and after the mainshock to decide if we can reject the hypothesis that these distributions are the same. Such rejection would be evidence for stress triggering. We have applied this test to the November 24, 1987, Elmore Ranch/Superstition Hills earthquake sequence and find that those post-mainshock events that experienced stress increases of at least 0.01--0.03 MPa (0.1--0.3 bar) or that occurred from 1.4 to 2.8 years after the mainshocks are consistent with having been triggered by mainshock-generated static stress changes. ¿ 1999 American Geophysical Union