Due to the small dimensions of the adhesive layer, the high non uniformity of the strain field and the non linear elastic behaviour of the adhesive material, the strain distribution at an adhesive joint can be predicted by F.E.M., but can not be experimentally obtained with classical approaches; only non standard procedures like Moire interferometry, or special artefacts like KGR extensometers may afford some insights on the behaviour of the adhesive. Due to their small size, ensuring low perturbation of the strain field, and their innate ability to measure strain and strain gradient along the sensor, fiber Bragg gratings offer a good opportunity to solve this problem, and it is a good example of situations that may benefit from these new sensors. Fiber Bragg gratings may be placed or at the interface, within the adhesive layer, or embedded at the adherents, if these were made of composite material. Tests may be run at different temperatures, changing the adhesive characteristics from brittle to pseudoplastic without additional difficulties. When loading the joint, the strain field is obtained by analyzing the distorted spectrum of the reflected light pulse; the algorithm for doing it has already been published. A comparison with theoretical results is done, and the validity and utility of these sensors for this and similar applications is demonstrated.