Rupture tests on internally pressurized, thin-walled hollow cylinders of Westerly granite with impermeable inner membranes suggest that the conventional, or Terzaghi, effective stress law does not describe tensile failure at high internal pressurization rates near 6 MPa/s. Unjacketed and saturated samples with an initial pore pressure and for which the inner cavity pressure was increased rapidly with respect to the diffusivity, display substantially increased apparent tensile strengths and deformation moduli much higher than similarly configured but more slowly pressurized tests. Alternatively, the properties of completely dry test pieces with no pore pressure show little, if any, dependence on pressurization rate. Further, the behavior of the rapid unjacketed tests was similar to that for completely dry samples. These observations can not be explained by the predicted undrained response, but they provide indirect evidence for diminished pore pressure effects reminiscent of dilatant hardening observed in compressive failure experiments. Calculated pore pressure diffusion rates support this suggestion as pore pressure perturbations cannot be damped out on the time scale of the rapidly pressurized tests. It is not clear if they effects are produced by elastic microcrack dilatancy, of which the nonlinear stress-strain curve of granites is symptomatic, or the irreversible production of new porosity as in compressive shear failure tests. ¿ American Geophysical Union 1992