Purpose: To test the null hypothesis that three self-adhesive luting materials had the same microtensile bond strength when used to lute ceramic IPS Empress II disks to the dentin of perfused teeth. Materials and Methods: Occlusal enamel and the roots of 9 human third molars were removed and crown segments connected to a perfusion system (30 cm H2O). Nine ceramic disks (IPS Empress II) were prepared, conditioned with 5% HF (20 s), rinsed with water, and air dried. A primer silane agent was applied (Monobond-S) for 60 s and air dried. Teeth were bonded to disks using one of three materials: Multilink System, RelyX Unicem, or Panavia F light. Specimens were vertically sectioned to obtain square bars. Each bar was fixed to a rigid custom-made tensile device and submitted to tensile force until debonding. Microtensile bond strength (µTBS) was given in MPa. Because the means of bonded areas (BA) were different between groups, it was impossible to compare µTBS results directly. Thus, the regression line TBS(MPa) = -2.15 + [19.92 / BA (mm2)] was calculated to correlate µTBS results and BA. Residuals (difference between the value estimated by the regression line and the actual value) were stored as "distances". ANOVA was performed to examine the statistical significance of differences between means distances of the groups', and the Tamhane's post hoc test was used to locate eventual differences. Results: The number of bars used in the analysis was n = 147. Residuals should not be statistically different if the null hypothesis (represented by the regression line) was credible. ANOVA showed statistically significant differences between the distances of the groups' means (p 0.00001). Tamhane's post hoc test (p 0.05) showed that different group means were RelyX Multilink Panavia F. Conclusion: The null hypothesis was rejected: materials showed different tensile bond strengths. Panavia F obtained highest bonding values, followed by Multilink System. RelyX Unicem, the only material that does not require pretreatment of dentin, achieved the lowest TBS values under these study conditions. Keywords: self-adhesive luting, ceramic, IPS Empress II, microtensile