This study investigated the marginal adaptation of mineral trioxide aggregate (MTA) as a root-end filling material, compared with commonly used rootend filling materials by scanning electron microscopy (SEM). Eighty-eight single-rooted freshly extracted human teeth were cleaned, shaped, and obturated with gutta-percha and root canal sealer. Following root-end resection and cavity preparation, the root-end cavities were filled with amalgam, Super-EBA, Intermediate Restorative Material (IRM), or MTA. Using a slow-speed diamond saw, 40 roots were longitudinally sectioned into two halves. Resin replicas of resected root ends of the remaining nonsectioned roots were also prepared. After mounting longitudinal sections of roots and resin replicas of resected roots on aluminum stubs, the distance between the test rootend filling materials and surrounding dentin was measured at four points under SEM. Examination of the original samples showed numerous artifacts in the longitudinal sections of the specimens. In contrast, the resin replicas of resected and filled root ends had no artifacts. Statistical analysis of data comparing gap sizes between the root-end filling materials and their surrounding dentin shows that MTA had better adaptation compared with amalgam, Super-EBA, and IRM.