Abstract Cylinders of small-particle light-cured composite resin were bonded to the superficial dentin layer of young bovine incisors that had previously been subjected to 4 different treatments: 1) immersed in 35% hydrogen peroxide for 60 min, then etched with 37% phosphoric acid gel for 60 s; 2) immersed in saline for 60 min, then etched with 37% phosphoric acid gel; 3) etched with 37% phosphoric acid gel for 60 s, then immersed in 35% hydrogen peroxide for 60 min; and 4) etched with 37% phosphoric acid gel, then immersed in saline for 60 min. Specimens were stored in water at 37°C for 1 and 7 d prior to tension and shear testing. A total of 128 teeth were tested: 8 teeth in each group, for each day and for each test. No measurable adhesion of resin to peroxide-treated dentin was observed and statistical analysis of the test results indicated that this represented a highly significant reduction in the bond strength as compared with values obtained when resin bonded to saline-treated dentin was tested. SEM examination of randomly selected fracture test specimens indicated that the bond failure was both cohesive and adhesive in nature, suggesting that peroxide treatment adversely affects the setting of bonding resin and its adhesion to a filled resin.