Introduction. Patellar tracking in total knee replacements has been extensively studied, but little is known about patellar tracking in isolated patellofemoral replacements. We compared patellar tracking and the position of the patellar groove in the natural knee, followed by implantation of the femoral component of a PFR (patella unresurfaced) and after implantation of the femoral & patellar component of the PFR. Methods. Computer navigation was used to track the patella in eight whole lower extremities of four cadavers in the natural knee, in the same knee with the femoral component of the PFR (PFR-P) and with the femoral and patellar component of the PFR (PFR+P, patella resurfaced) (Depuy Sigma PFR). The form and position of the trochlea in the natural knee and the patellar groove of the femoral component was also analysed. Values are means+/−SD, two tailed Student's t-test for paired samples. Results. With a PFR-P the patella had a slightly more lateral tilt (0.8+/−0.8° to 2.8+/−2.5° at 40–100° of flexion, p<0.05 vs. Nat), this was more pronounced with the PFR+P (2.0+/−0.7° to 4.9+/−1.8° at 20–90° flexion, p<0.05 vs. Nat., p<0.05 vs. PFR-P at 20–80° flexion). No differences in patella rotation were seen between the three groups. In the PFR-P group the patella tracked a little more medially compared to the natural knee (0.6+/− 0.7mm to 1.3+/−2.6mm, p<0.05 at 20°,80°,90° flexion). The difference was more pronounced after patella resurfacing (PFR+P) (2.1+/−2.0mm to 3.0+/−2.2mm, p<0.05 vs. Nat. at 10°–100°, p<0.05 vs. PFR-P from 10–100°). When analysed relative to the patellar groove of the trochlea/femoral component the patella in the natural knee tracked slightly lateral to the groove (2.0+/−1.7mm to 2.9+/−2.0mm at 50–100° p<0.05), so did the patella of PFR-P (2.0+/−2.3mm to 2.3+/−2.3mm at 60–90° flexion, p<0.05), whilst the PFR+P tracked right on the groove (0.6+/−3.7mm medially to 0.6+/−2.9mm laterally, p<0.05 vs Nat at 10–30° & 70–100°, p<0.05 vs. PFR-P at 10–100°). Distance from the patellatot the epicondylar axis was slightly larger in the PFR-P group (0.6+/− 0.7mm to 1.3+/−1.4mm, p<0.05 vs. Natu at 20,80 & 90°. This was more pronounced with patellar resurfacing (2.1+/−2.0 to 3.0+/−2.2mm, p<0.05 vs. Nat at 10–100°, p<0.05 vs. PFR-P at 20–100°) The patella groove on the natural knee and the implanted femoral component of the implanted PFR had the same radius, inclination relative to the femoral mechanical axis, antero-posterior position and medio-lateral orientation. As intended by the designers the groove of the patellar component extended about 13mm further superiorly and 0.5mm more inferiorly. Discussion. The patella groove on the femoral component of the PFR reproduces the natural trochlear anatomy well. Patella tracking in the PFR-P shows only minor differences compared to the natural knee. Resurfacing of the patella in the PFR+P group causes the patella to tilt a little more laterally and track a little more medially, the distance to the epicondylar axis is slightly larger but this allows the patella to engage better in the patellar groove of the femoral component