Oxidation tests on spent-fuel fragments and rod segments were conducted between 250 and 360°C to relate temperature and defect size to fuel oxidation rate and time-to-cladding-splitting. Defect sizes from an equivalent circular diameter of 8 µm (the approximate size of a stress-corrosion-cracking-type breach) to 760 µm were used. Samples, held at temperature in a flowing air atmosphere, were frequently weighed and visually observed to determine the oxidation rate and effects of oxidation. Both the size and shape of the defect appear to influence the time-to-cladding-splitting. Above 283 °C, time-to-cladding-splitting was longer for the sharp small defect than for the large circular defect, an effect that diminished as the temperature decreased. By 250°C the sharp small defects split open before the large circular defects, indicating that, at lower temperatures, the defect’s shape and not its size may be more important when determining time-to-cladding-splitting. At both 283 and 295°C, the defects in fuel rod segments with lower burnups propagated sooner than those in rod segments with higher burnup from the same parent rod. The cumulative damage fraction approach, using a reasonable decreasing time/temperature profile, was applied to determine time-to-cladding-splitting for pressurized water reactor (PWR) fuel with a burnup >640 MWh/kg of uranium. Breached PWR fuel rods will not split open from fuel oxidation during 100 yr of storage if the rod is not exposed to air until the temperature drops below 230°C. Lower burnup fuel apparently requires lower temperature limits. The temperature limits appear to depend more on the time/temperature profile in the storage container than on oxidation rates.