Purpose: To compare the torsional strengths and the effects of functional loading on five different implant surface structures. Materials and Methods: This in vivo longitudinal study of 9 months' duration examined osseointegration in 180 stepped cylindric implants placed into the mandibles of 18 healthy sheepdogs. Implants with five different surface structures were placed (n = 36 in each group): (1) smooth surface; (2) deep profile structure (DPS); (3) titanium plasma sprayed (TPS); (4) hydroxyapatite coated; and (5) sol-gel hydroxyapatite coated (SGHA). All implants were investigated under standardized conditions. At uncovering (3 months), half the implants were loaded with prefabricated crowns for 6 months, and the other half remained unloaded. Removal torque values were evaluated at 3, 6, and 9 months after implant placement. The data were analyzed using analysis of variance and Scheffé correction (a .05). Results: All smooth-surface and SGHA implants failed in succession during the fourth and ninth months of the study. After 3 months the removal torque resistance of DPS, TPS, and hydroxyapatite-coated implants was significantly higher (P .001) than that seen in the other two groups. Initially, the removal torque resistance of unloaded hydroxyapatite-coated implants was superior, and the decrease during the observation period was not significant. The increase in removal torque resistance under functional loading was not significant for the DPS implants. A significant increase was found in removal torque resistance for the loaded TPS implants (P .05). The unloaded DPS and TPS implants showed no change in removal torque levels after the closed healing period of 3 months. Conclusions: Successful osseointegration was achieved with DPS, TPS, and hydroxyapatite-coated implants, and smooth-surface and SGHA implants failed. Removal torque resistance was enhanced with controlled functional loading. Keywords: functional loading, implant surfaces, osseointegration, torsional strength