It is generally assumed that passively fitting superstructures are a prerequisite for long-lasting implant success. In the study presented, the strain development of three-unit implant fixed partial dentures (FPDs) was evaluated at the bone surrounding the implant and on the superstructure using a strain gauge technique. Six groups of three-unit FPDs representing the commonly used techniques of bridge fabrication were investigated with 10 samples each, in order to quantify the influence of impression technique, mode of fabrication and retention mechanism on superstructure fit. Two ITI implants (Straumann, Waldenburg, Switzerland) were anchored in a measurement model according to a real-life patient situation and strain gauges were fixed mesially and distally adjacent to the implants and on the bridge pontics. The developing strains were recorded during cement setting and screw fixation. For statistical analysis, multivariate two sample tests were performed setting the level of significance at P=0.1. None of the investigated bridges revealed a truly passive fit without strains occurring. About 50% of the measured strains were found to be due to impression taking and model fabrication, whereas the remaining 50% were related to laboratory inaccuracies. The two impression techniques used did not reveal any significant differences in terms of precision. Both modes of fixation--i.e. cement and screw retention--provoked equally high stress levels. In the fabrication of screw-retained FPDs, similar results were obtained from the use of burn-out plastic copings and the technique of casting wax moulds to premachined components. Bonding bridge frames onto gold cylinders directly on the implants significantly reduces strain development.