The objective of this study was to measure and compare the precision of fit of implant-supported prostheses in the edentulous jaw by using both the master cast replicas and the intraoral implants as references. Seven maxillary and 10 mandibular prostheses were randomly selected and measured by means of a three-dimensional (3-D) photogrammetric technique. The results indicated that prostheses routinely connected to osseointegrated implants could demonstrate distortion between the framework and individual implants of up to several hundred microns. When master casts were used as a reference, the mean 3-D distortion of the center point of gold cylinders was 37 µm (SD 18) and 75 µm (SD 40) for mandibular and maxillary prostheses, respectively. The corresponding mean displacement was 90 µm (SD 51) and 111 µm (SD 59), respectively, when the intraoral implants were used as references. The mean 3-D distortion was significantly higher for the intraoral measurements in both arches (P .001 and P .05). Furthermore, the overall distortion was significantly higher for the maxillary prostheses when the master casts were used as the reference (P .05). However, for the intraoral measurements, no statistically significant difference of fit between the arches was possible to observe (P .05). This could possibly be explained by the finding that intraoral measurements of the mandibular prostheses indicated a deformation and rotation of the mandible that was not observed in the maxillary prostheses. A further factor in the lack of statistical significance could be the relatively small sample size. Keywords: framework, implant, precision