This study determined whether the strength with which resin composite bonds to dentin is influenced by variations in the curing rate of resin composites. Resin composites were bonded to the dentin of extracted human molars. Adhesive (AdheSE, Ivoclar Vivadent) was applied and cured (10 seconds @ 1000 mW/cm2) for all groups. A split Teflon mold was clamped to the treated dentin surface and filled with resin composite. The rate of cure was varied, using one of four LED-curing units of different power densities. The rate of cure was also varied using the continuous or pulse-delay mode. In continuous curing mode, in order to give an energy density totaling 16 J/cm2, the power densities (1000, 720, 550, 200 mW/cm2) emitted by the various curing units were compensated for by the light curing period (16, 22, 29 or 80 seconds). In the pulse-delay curing mode, two seconds of light curing at one of the four power densities was followed by a one-minute interval, after which light cure was completed (14, 29, 27 or 78 seconds), likewise, giving a total energy density of 16 J/cm2. The specimens produced for each of the eight curing protocols and two resin composites (Tetric EvoCeram, Ivoclar Vivadent; Filtek Supreme XT, 3M ESPE) were stored in water at 37 degrees C for seven days. The specimens were then either immediately subjected to shear bond strength testing or subjected to artificial aging (6,000 cycles between 5 degrees C and 55 degrees C baths) prior to testing. Failure modes were also assessed. The shear bond strengths were submitted to factorial analysis of variance, and the failure modes were submitted to a Chi-square test (alpha = 0.05). All but power density (curing mode, resin composite material and mode of aging) significantly affected shear bond strength. The curing mode and resin composite material also influenced the failure mode. At the selected constant energy density, pulse-delay curing reduced bonding of the resin composite to dentin.