An experimental method to determine the stress–elongation relation for a thin adhesive layer loaded in peel is presented. The method is based on equilibrium of the energetic forces acting on a DCB-specimen. These energetic forces are identified to be associated with the geometrical positions of the acting loads and the start of the adhesive layer. The first energetic force is shown to be given by the product of the force and the rotation of the loading point. The second energetic force is shown to be given by the area under the stress–elongation curve for the adhesive layer. Using equilibrium of these energetic forces, the shape of the stress–elongation curve is determined. A test set-up is developed to facilitate the experiments. Special consideration is given to the accuracy of the measurement of the elongation of the adhesive. Results from two sets of experiments with slightly varying geometry are presented. The main result is that the stress–elongation relation can be described by a curve divided into three parts; initially the stress increases proportionally to the elongation. This corresponds to a linear elastic behaviour of the layer. The next part is given by a constant limiting stress. The curve ends with a parabolically softening part. After this point, a crack has been initiated in the adhesive. The experimental results are first compared to an asymptotic analysis using linear elastic fracture mechanics. This shows that the new method to evaluate the fracture energy gives consistent results. The experiments are also simulated using the measured stress–elongation law. Good agreement with the experiments is achieved which further validates the method. The fracture energy and the maximum peel stress are found to agree well within each set of experiments. Some variation is found between the two sets. This is accredited to differences in fracture initiation