The local detail of the geometry of the adhesive interface can have a significant effect on the measurement of dentin bond strengths and may be a contributory factor in the discrepancies among data in the published literature. The potential effect on the dentin bond strength due to modifications of the local stress distribution at the adhesive/dentin interface has been assessed. Tensile bond strength measurements for specimens with and without an adhesive flash were carried out and compared with the stress distribution at the adhesive interface determined by finite element stress analysis. The results showed that when the adhesive was constrained to the interface only, the tensile bond strength was 3.10 MPa, which increased to 6.90 MPa when a flash of adhesive was present. For a realistic measurement of dentin bond strength, the adhesive should be constrained to the interface only. Extension of the adhesive beyond the interface will result in an artificially high value for the dentin bond strength. A standardized method for the measurement of dentin bond strength is urgently needed, but must take these as well as all other known factors into account if results from different testing centers are to be directly comparable.