Statement of problem. When implants are placed interforaminally in mandibular implant-supported screw-retained prostheses, the most anterior implant is usually positioned lingual to the incisors of the prosthesis. This creates an anterior cantilever with the entire prosthesis acting as a class I lever and possibly placing the anterior implant under alternating tension and compression during function. Purpose. The purpose of this study was to measure the anterior cantilever of randomly chosen patients with restored mandibular implant-supported fixed prostheses, and to establish the proportions of anterior to posterior cantilever lengths relative to the anteroposterior spread. Material and methods. Thirteen edentulous patients were recruited from the University of Minnesota's dental implant program. Each patient had been restored with a maxillary complete denture and a mandibular implant-supported screw-retained prosthesis supported by 5 endosseous implants. Each patient had 1 mandibular impression made with irreversible hydrocolloid, which was poured in type III gypsum. A FaroArm precision 3D measuring stylus was used to make multiple-axis (X-Y-Z) measurements (mm) on the casts of the dental implants, of anteroposterior spread, and anterior and posterior cantilevers. Presence or absence of screw loosening was noted using a screwdriver with finger pressure. Data were analyzed using a 1-way analysis of variance to compare prostheses with loose screws to prostheses without loose screws, for each of 3 outcome measures: length of anterior cantilever, length of posterior cantilever, and anteroposterior spread (P<.05). Results. Mandibular anterior cantilever lengths ranged from 5.5 to 14.4 mm with a mean of 8.78 mm. Combined posterior cantilever lengths ranged from 9.2 to 20.9 mm with a mean of 16.2 mm. Anteroposterior spread ranged from 5.2 to 12.3 mm with a mean of 7.9 mm. From a total of 65 retaining screws, 7 were found to be completely loose. No apparent correlation was found between length of mandibular anterior cantilever and screw loosening (P=.45), although the ratio of posterior cantilever to anteroposterior spread (2:4) was significantly associated with screw loosening (P=.006). Conclusion. Within the limitations of this study, anterior cantilevers in mandibular implant-supported screw-retained prostheses were frequent and appeared to depend on implant placement and prosthesis design. The ratio of anterior to posterior cantilever lengths was approximately 1:2.