Continuous records from a borehole strainmeter and a long baseline tiltmeter in the Long Valley caldera provide critical insights into the origin of at least one episode of minor seismicity in volcanic regions triggered by the 28 June 1992, ML 7.3 Landers, California, earthquake. A strain transient reaching a peak of 0.25 microstrain occurred in the few days following the Landers event and decayed over the next 20 days. A tilt perturbation during the same time reached a peak amplitude of 0.2 microradians. These signals correspond approximately in time to the primary seismic moment release across a 50 km2 region of the south part of the caldera at depths between 2 and 10 km. Corresponding strain transients in 5-km geodetic lines across the south caldera are not apparent above the 95% confidence limits of about 0.4 microstrain in daily sampled data during this same period. These data rule out models involving single localized inflation sources within the upper crust beneath the caldera, including that responsible for the current rapid inflation of the resurgent dome. They also preclude models involving aseismic slip on single strike-slip or normal faults in the caldera. A single source in the form of a relaxing magma body at a depth of 50 km beneath the caldera can account for the deformation data, but whether the small stress changes are sufficient to drive the triggered seismicity is not clear. An alternate possibility involves distributed deformational sources triggered by the passage of the 10 microstrain peak amplitude surface waves from the earthquake. This distributed deformational source could result either from rupturing of overpressured fluid or gas chambers commonly encountered in volcanic regions or from advective gas overpressure during release of gas bubbles in hydrothermal or magmatic fluids.