Sixty-two consecutive cemented modular unicompartmental knee arthroplasties in 51 patients were studied prospectively. At surgery, the other compartments had at most Grade 2 chondromalacia. The average age of the patients at arthroplasty was 68 years (range, 51-84 years). One patient was lost to followup and 10 died with less than 6 years followup. The average followup of the remaining 51 knees was 7.5 years (range, 6-10 years). The preoperative Hospital for Special Surgery knee score of 55 points (range, 30-79 points) improved to 92 points (range, 60-100 points) at followup; 78% (40 knees) had excellent and 20% (10 knees) had good results. The mean range of motion at followup was 120 degrees with 26 knees (51%) having range of motion greater than 120 degrees. One patient underwent revision surgery for retained cement, one patient underwent knee manipulation, and one patient underwent revision surgery at 7 years for opposite compartment degeneration and pain. Radiographically, 26 knees (51%) had at least one partial radiolucency. There were no complete femoral radiolucencies, but there were three complete tibial radiolucencies, all less than 2 mm. No component was loose as seen on radiographs. At final followup, five of the opposite compartments (10%) and three of the patellofemoral joints (6%) had some progressive radiographic joint space loss; this was less than a 25% loss in all but one knee component that was revised. At 6- to 10-years followup, cemented unicompartmental knee arthroplasty yielded excellent clinical and radiographic results. The 10-year survival using radiographic loosening or revision as the end point was 98%. Using stringent selection criteria, unicompartmental knee replacement can yield excellent results and represents a superb alternative to total knee replacement.