Screw-shaped implants were prepared with three different surface topographies: One was left as machined, ie, a truned surface, and two were blasted surfaces with differing degrees of surface roughness. The surface topography was measured with a confocal laser scanning profilometer and the surface roughness was characterized using height and spatial descriptive parameters. The turned surface had an average surface roughness of 0.96 um and an average peak spacing of 8.6 um. The two blasted surfaces had surface roughness values of 1.16 um and 1.94 um, respectively; the corresponding values for the peak spacin g parameter were 10.00 um and 13.22 um, respectively. After 1 year in rabbit bone, the bone response to the turned implants was compared with the response to the two blasted implant surfaces. Firmer bone fixation was found for the two blasted surfaces, with statistically significant increases in removal torque and percentage of bone-to-metal contact. Furthermore, about 2 mm from the implant surface, the titanium release was similar for the turned and the 25-um aluminumoxide-blasted implants.