In the quest to reduce cladding strain from power ramps and thereby reduce the incidence of defects, various aspects of pellet geometry have been examined. The influence of pellet end squareness was investigated in steady-power irradiation of UO2 with end tapers well in excess of the currently allowable deviation from end squareness, and no increase in postirradiation cladding strains was noticeable. Five ramp tests were performed to examine the effects of pellet length-to-diameter ratio, the width of the loadbearing annulus (shoulder) at the dished end, and the effect of adding a small 45-deg edge chamfer. Although cladding strains were significantly reduced by changes in pellet geometry, there appears little incentive to do so from the viewpoint of preventing Zircaloy cracking, since defects occurred in rods that had lower overall strains than those that survived.