Mechanical testing of the implant-tissue interface has been the focus of numerous investigations concerning the anchorage capacity of implants. The purpose of this study was to measure reverse-torque failure after 6 months of healing for three different biomaterials in the posterior jaws of four adult female baboons. The animals had all of their posterior teeth surgically extracted and, following 10 weeks of healing, 7 implants were placed in each quadrant. The biomaterials included titanium plasma-sprayed surfaces, titanium-aluminum-vanadium surfaces (both 3.8 mm x 10 mm), and a commerically pure titanium surface (3.75 mm x 10 mm). After 6 months, torque data were collected using a counterclockwise computerized torque driver and were analyzed by repeated measures analysis of variance for differences related to biomaterial, jaw, and biomaterial/jaw. Post-hoc Tukey Dramer analysis was also performed for within-group differences (a = .05 level). The biomaterial comparison reveal ed a significant difference between the titanium plasma-sprayed and the combined commerically pure titanium/titanium-aluminum-vanadium groups (analysis of variance, Tukey Kramer, P < .05). The jaw comparison showed no significant difference, although the data suggest that higher forces may be required for mandibular torsional failure. The biomaterial/jaw comparison revealed that jaw differences for the mean values of commercially pure titanium and titanium-aluminum-vanadium implants were gre ater than jaw differences for mean values of titanium plasma-sprayed implants, although these differences were not statistically significant. Because of the lack of correlation between single-cycle biomechanical tests and clinical performance, it is necessary to be selective in assigning usefulness to data of this type.