Abstract Most in vitro testing of bonding systems is performed using specimens made in a mold with a low configuration ( C) factor (ratio of bonded/unbonded surfaces) whereas clinically the C-factor is usually much greater. This study compared the effect of thermal cycling on the measured shear bond strength of 3M Single Bond dental adhesive bonded to dentin using molds with two different C-factors. The hypothesis was that neither C-factor nor thermal cycling would affect measured bond strengths. Resin composite was bonded to human dentin in cylindrical molds with an internal diameter of 3.2 mm and either 1 mm or 2.5 mm deep. The 1 mm deep molds had a C-factor of 2.2 and the 2.5 mm deep molds had a C-factor of 4.1. Specimens were debonded either 10 min after they had been bonded to dentin, or after they had been stored for 7 days in water at 37±1°C, or after thermal cycling 5000 times for 7 days. Two-way ANOVA showed that overall both the C-factor and the storage condition had a significant effect on bond strength ( p<0.001). There was a significant interaction ( p<0.001) between the C-factor and how the specimens had been stored. The GLM/LSMEANS procedure with Sidak's adjustment for multiple comparisons showed that overall the specimens made in the mold with a high C-factor (4.1) had a lower bond strength than those that had been made in the mold with a lower (2.2) C-factor ( p<0.001). Thermal cycling had a negative effect on the bond strength only for specimens made in molds with a C-factor of 4.1 ( p<0.001).