We measured forces in the quadriceps tendon and ligamentum patellae in six human cadaver knee joints loaded through a range of flexion angles from 30 to 120°. Using standardized loads based on one-sixth of maximum isometric quadriceps moments reported in human volunteers, we measured maximum average forces at 60° flexion of 532 N in the quadriceps tendon (F_Q) and 470 N in the ligamentum patellae (F_L). Linear extrapolation to full maximum extensor moments results in estimates of 3,200 N for F_Q and 2,800 N for F_L. The force ration (F_L/F_Q) reached a maximum value of 1.27 at 30° and minimums of 0.7 at 90 and 120° knee flexion. Contrary to prevailing opinion, our results indicate that the patella is not a simple pulley that serves only to change the direction of equal forces in the quadriceps tendon and ligamentum patellae. Instead, the ratio of forces in these structures varies significantly with flexion angle. The ratio appears to be determined by the changing location of the patellofemoral contact area relative to the insertions of the tendon and ligament. These findings emphasize the biomechanical importance of patellar length and of the vertical dimensions and locations of the patellofemoral contact area. Attempts at surgical intervention for the treatment of disorders of the extensor apparatus should recognize these variations. Procedures that tend to move the contact area more proximally (at a particular flexion angle) will also tend to decrease the F_L/F_Q ratio.