STATEMENT OF PROBLEM Ceramic restorations have been known to cause wear of opposing enamel. PURPOSE The purpose of this study was to evaluate enamel wear caused by 3 ceramic substrates in the glazed and polished conditions. MATERIAL AND METHODS Sixty ceramic discs (10 x 2 mm)-20 each of Finesse, All-Ceram, and IPS-Empress-were prepared and glazed. Each group of 20 was divided into 2 groups of 10. The surfaces of one group were ground and polished using a porcelain polishing kit (Dialite). The remaining 10 were left as glazed. Ten specimens of a type III gold alloy were cast into rectangular shapes of 10 x 12 x 2 mm and polished. Seventy human cusps were prepared from sound, caries-free, extracted teeth and abraded against the substrates in a wear machine for a total of 10,000 cycles. The cusp height loss was traced before and after the wear test using a profile projector. Mean surface roughness (R(a)) values for the substrates were also recorded with a profilometer before testing. Differences in R(a) were evaluated using 1- and 2-way ANOVA and the Scheffe post hoc test (alpha = .05). RESULTS One-way ANOVA indicated that enamel height loss was significantly different by material (P < .001) and surface condition (glazed and polished or glazed; P < .05). Gold, polished Finesse, and polished All-Ceram were the least abrasive, whereas glazed IPS-Empress was the most abrasive. There was no significant interaction effect between substrate type and surface condition. Significant differences were found when R(a) of the substrate condition was compared with enamel wear (P < .01). CONCLUSIONS Gold, polished Finesse, and polished All-Ceram caused the least enamel wear, whereas IPS-Empress caused the most wear. Cast gold was significantly different than glazed IPS-Empress (P < .05), whereas other groups overlapped. There was significant correlation between R(a) and enamel wear (P < .01).