Abstract Our data indicate that critical and postcritical reflections from crustal layers and the Moho produced increased shaking at discrete distances along the San Francisco Peninsula during the 1989 Loma Prieta earthquake. These reflections may have produced an increase in amplitude that is as much as 10 times greater than that of the direct arrival. Peak amplitude-distance patterns measured from explosive sources, synthetic seismograms, aftershocks, and the mainshock of the 1989 Loma Prieta earthquake indicate that (1) point sources, such as explosions, produce similar peak amplitude-distance relations as distributed, double-couple sources when the crustal structure is approximately a Poisson solid; (2) peak amplitudes from smaller point sources (explosions) may be scaled to those of larger sources (earthquakes); and (3) reflections caused a pattern of high amplitudes at specific distances along the San Francisco Peninsula that geographically correlates with areas of reported damage following from the Loma Prieta mainshock. Our study indicates that critical and postcritical reflections were stronger influences on the locations of strong shaking than local geology because a number of sites around the San Francisco Bay that are underlain by unconsolidated sediments experienced much less shaking than sites that were underlain by hard rock. Furthermore, some severely shaken hard-rock sites were farther from the epicenter than many of the less severely shaken “soft-sediment” sites. Models of the regional crustal structure and seismic-wave propagation paths may be useful in identifying sites around the San Francisco Bay region that are at risk due to reflected seismic energy, and this type of analysis may be useful in other seismically active regions.