For differences among materials to be easily detected, low variation in in vitro wear tests is desirable. The working hypothesis of this paper was that antagonists standardized for shape and size and according to materials would show mean values similar to those found in natural, non-standardized cusps, and that standardization would lead to a reduction in mean variation. First, the shapes and sizes of palatal cusps of non-erupted human upper third molars were measured. The cusp cupola was best described by the formula y = 0.001 x2 and was symmetrical around the axis of rotation. Up to 200 microm of the y-axis, this parabola corresponded best to a ball radius of 0.6 mm. Based on this information, standardized antagonists were fabricated from both human enamel and steatite. Wear in the occlusal contact area and the wear of opposing conventional ceramic and fine hybrid composite, respectively, were quantified in a computerized chewing simulator. As a control, natural human enamel cusps were used. Standardization of enamel cusps did not reduce the variation of the resulting wear compared with that of non-standardized enamel antagonists. Furthermore, standardization led to significantly different results both in the antagonists and in the opposing restorative materials. Thus, natural enamel antagonists are preferable for the simulation of wear in the occlusal contact area.