Objective: To test the retention of smooth-surface sealants bonded with different etching protocols against toothbrushing and, secondarily, to characterize the type and location of sealant loss. Materials and Methods: Eighty-nine extracted human teeth were randomly assigned one of two etching protocols: 37% phosphoric acid etch (ETCH) or self-etching primer (SEP). Six specimens at a time were placed in a toothbrushing machine to simulate 4, 8, 12, and 24 months of toothbrush abrasion. Using black-light photographs of each specimen taken before and after brushing, four blinded coinvestigators determined new sealant loss, loss along the edge of an initial defect, and the location of sealant loss. Results: Overall, there were significantly (P , .05) more teeth with sealant loss in the SEP group (38.6%) than in the ETCH group (15.5%). New loss of sealant was significantly (P , .05) more likely in the SEP group (27.2%) than in the ETCH group (2.2%). Of the teeth with new loss of sealant, all (100%) had loss at the edge, and 23% had progressive loss. There was no significant group difference in sealant loss from initial defects. Of the teeth that showed enlargement of initial defects, 91% had loss at the edge and 91% had progressive loss. Conclusions: Using SEP to apply facial sealants results in lower retention rates than using ETCH. The vast majority of sealant loss occurs at the edges. Loss of sealant due to enlargement of an initial defect is highly progressive over time. (Angle Orthod. 2015;85:750-756.)