Various techniques have been advocated for finishing retrofilling materials. This study evaluated marginal interfaces between tooth structure and the retrofilling material following three techniques of finishing IRM and EBA. This interface was evaluated for the presence of flash and for adaptation to the prepared canal walls. Thirty-one teeth were divided into six groups of five and one control. The teeth were instrumented to a minimum of a 35 K-file, obturated with gutta-percha, resected perpendicular to their long axes and ultrasonically prepared to receive a retrofilling. The teeth were filled with either IRM or EBA cement and finished by ball burnishing, burnishing with a moistened cotton pellet, or with a carbide finishing bur in a high-speed handpiece with air/water spray. Photomicrographs were made and examined by four evaluators. Retrofillings finished with a finishing bur displayed significantly better marginal adaptation and little evidence of flash when compared to the other techniques (p < 0.05). There were no significant differences between ball burnished or moistened cotton pellet techniques or between retrofilling materials used in this study (p < 0.05).