Function of the knee and patellofemoral symptoms were correlated with the position of the implant in 101 consecutive patients with 116 posterior stabilized condylar knee prostheses. All of the patients were followed for a minimum of two and a half years with sequential physical examinations, radiographs, and functional evaluation of the knee. In sixteen knees (14 per cent), clicking or catching of the patella in terminal extension or painless crepitation throughout the arc of flexion developed without lowering the functional knee-evaluation score. Pain or mechanical problems, or both, that lowered the functional knee-evaluation score occurred in another fourteen knees (12 per cent), within the first postoperative year. Of these fourteen, eight required revision solely for patellofemoral complaints. Critical analysis of the tibial-patellofemoral mechanical axis identified three surgical variables that were found to markedly affect the functional result of the prosthesis: the distance from the center line of the tibial prosthesis to the center line of the tibial plateau, a change in the position of the joint line of the prosthesis relative to the hip and ankle, and the patellar height, measured as the perpendicular distance from the inferior pole of the patellar implant to the joint line of the prosthesis. Functional knee scores, range of motion, patellofemoral pain or mechanical symptoms, the need for revision, and the necessity of manipulation could all be statistically significantly correlated with the three independent variables. In addition, a range of neutral alignment was developed.(ABSTRACT TRUNCATED AT 250 WORDS)