The present study was designed to investigate the fixation strength of a quadruple semitendinosus-gracilis graft compared with a middle-third bone-patellar tendon-bone graft using a new interference screw developed to fix hamstrings grafts for ACL reconstructions (RCL Smith & Nephew Donjoy). Five pairs of human cadaveric knees from donors with a mean age of 43 (range 33-52) years were used. One knee of each pair was randomly allocated to be reconstructed on the femoral side with a semitendinosus-gracilis graft from the same donor using RCL screw fixation. As the control, the contralateral knee was correspondingly reconstructed with a bone-patellar tendon-bone graft using the same interference screw. The grafts were pulled out at a velocity of 30 mm/s by an axially applied load using a MTS machine. The mean (SD) failure load for the bone-patellar tendon-bone graft fixations was 505 (25) N, 110% stronger than the mean failure load for the semitendinosus-gracilis graft fixations, which was 240 (47) N (P = 0.003). The stiffness for the patellar tendon-bone graft fixations was 46 (11) N/mm, 120% stiffer than the semitendinosus-gracilis graft fixations, which was 22 (11) N/mm (P = 0.01). This study shows that the interference screw principle used for ACL reconstructions with hamstrings tendons is inferior to that for bone-patellar tendon-bone reconstructions although the screw was developed especially for soft-tissue fixation in bone tunnels.