The purpose of this study was to compare the efficacy of sealing the coronal 2-mm of the root canals versus covering the entire pulpal floor with one of two dental-resin cements (Principle or C&B Metabond). Sixty-two molars with the occlusal half of the crowns and the apical half of the roots removed were used. Each canal was enlarged by using a #3 Gates Glidden bur and obturated with unsealed gutta-percha cones. The teeth were randomly assigned to four groups, each containing 15 teeth, plus a negative and a positive control. In group 1, 2 mm of Principle were placed over the entire pulpal floor. In group 2, Principle was placed 2 mm into each canal orifice. Groups 3 and 4 were the same as groups 1 and 2, except C&B Metabond cement was used. After the cement set, the gutta-percha was removed and the integrity of the seal was tested by fluid filtration at a pressure of 20 cm H2O at 1 h and at 1, 2, and 4 weeks. The data were analyzed by a three-way ANOVA and the Student-Newman-Keuls tests at alpha = 0.05. The controls behaved as expected. Results showed that there were no statistically significant differences among the materials used or the location (p > 0.05), but there was a significant difference with respect to time. Principle leaked significantly more than C&B Metabond at 1 h (p < 0.05), but the seal became tighter over time. C&B Metabond leaked less early (p < 0.05) but increased in leakage at 4 weeks. Both materials sealed well over the 4-week study. Principle was easier to use, and sealing the entire pulpal floor was easier than sealing only the canal orifice.