The effects of postulated Cascadia subduction zone earthquakes on inelastic structural response have been quantified. The earthquakes studied ranged in size from those previously recorded to the largest plausible event, a magnitude 9.5, 240 second duration earthquake. Artificial acceleration records attenuated to epicentral distances corresponding to coastal range sites and Puget Sound sites were generated. These records were used as input for inelastic response history analyses of single-degree-of-freedom systems with either bilinear or degrading stiffness hysteretic relationships. The results indicate that the maximum displacements are not significantly greater than those produced by previously recorded events or by records that are compatible with current design code response spectra. However, the inelastic energy dissipated and the numbers of displacement cycles are somewhat greater for the largest events, although the energy demands and cyclic demands are similar to those from previous events for magnitudes up to 8.5. Since, the maximum credible event is not well established at this time no changes to the current design procedures are recommended.