The surface roughnesses of pure titanium implants were compared with scaling in vitro with curettes of dissimilar composition. Each of 10 transmucosal implant extensions (TIEs) was divided into three experimental surfaces and an untreated control surface. The three experimental surfaces were instrumented with either a titanium-alloy tipped curette, a curette of stainless steel, or a plastic curette. All experimental surfaces received 30 strokes with the designated curette within a 2 mm wide area. Alteration of the surfaces due to instrumentation was evaluated by a helium neon (HeNe) laser and reported as relative specular reflectance (RSR). Scanning electron microscopy (SEM) conformed the quantitative HeNe laser results. A significant decrease in mean RSR (greater roughness) was observed for surfaces treated by metal curettes compared to either untreated control surfaces (P less than 0.01) or surfaces treated by the plastic curette (P less than 0.01). No statistically significant difference was noted between untreated surfaces and those treated by the plastic curette. The titanium-alloy curette produced a significantly lower mean RSR (greater roughness) compared to those surfaces treated by the stainless steel curette (P less than 0.05). In summary, plastic instruments produced an insignificant alteration of the titanium implant surface following instrumentation, while metal instruments significantly altered the titanium surface.