Planar impact experiments were employed to induce dynamic tensile failure in Bedford limestone. Rock discs were impacted with aluminum and polymethyl methacralate flyer plates at velocities of 10 to 25 m/s. This resulted in tensile stresses in the range of ~11 to 160 MPa. Tensile stress durations of 0.5 and 1.3 μs induced microcrack growth which in many experiments were insufficient to cause complete spalling of the samples. Ultrasonic P and S wave velocities of recovered targets were compared to the velocities prior to impact. Velocity reduction, and by inference microcrack production, occurred in samples subjected to stresses above 35 MPa in the 1.3-μs PMMA experiments and 60 MPa in the 0.5-μs aluminum experiments. Apparent fracture toughness of 2.4 and 2.5 MPa m1/2 are computed for the 1.3- and 0.5-μs experiments. These are a factor of ~2 to 6 greater than quasi-static determinations. Three-dimensional impact experiments were conducted on 20 cm-sized blocks of Bedford limestone and San Marcos gabbro. Compressional wave velocity deficits up to 50--60% were observed in the vicinity of the crater. These damage levels correspond to O'Connell and Budiansky damage parameters of 0.4 as compared to the unshocked rock. The damage decreases as ~r-1.5 from the crater indicating a dependence on the magnitude and duration of the tensile pulse. Using the observed variation in damage with tensile stress from the one-dimensional experiments, and estimates of the variation of peak dynamic tensile stress and tensile stress duration with distance from an impact on an elastic half-space, the observed dependence of damage with radius in the three-dimensional experiments are theoretically predicted and compare favorably to experimental data. ¿ American Geophysical Union 1993