AbstractTo make a direct biomechanical comparison between the sandblasted and acid‐etched surface (SLA) and the machined and acid‐etched surface (MA), a well‐established animal model for implant removal torque testing was employed, using a split‐mouth experimental design. All implants had an identical cylindrical solid‐screw shape with the standard ITI thread configuration, without any macroscopic retentive structures. After 4, 8, and 12 weeks of bone healing, removal torque testing was performed to evaluate the interfacial shear strength of each surface type. Results showed that the SLA surface was more powerful in enhancing the interfacial shear strength of implants in comparison with the MA surface. Removal torque values of the SLA‐surfaced implants were about 30% higher than those of the MA‐surfaced implants (p = 0.002) except at 4 weeks, when the difference was at the threshold of statistical significance (p = 0.0519). The mean removal torque values for the SLA implants were 1.5074 Nm at 4 weeks, 1.8022 Nm at 8 weeks, and 1.7130 Nm at 12 weeks; and correspondingly, 1.1924 Nm, 1.3092 Nm, and 1.3226 Nm for the MA implants. It can be concluded that the SLA surface achieves a better bone anchorage than the MA surface, and that sandblasting before acid etching has a beneficial effect on the interfacial shear strength. As regards the bone–implant interfacial stiffness calculated from the torque‐rotation curve, the SLA implants showed an overall more than 5% higher stiffness compared with the MA implants, although the difference did not reach the statistical significance level. © 2002 John Wiley & Sons, Inc. J Biomed Mater Res 60: 325–332, 2002; DOI 10.1002/jbm.10063