We investigated the role of central activation in muscle-length dependent endurance. Central activation ratio (CAR) and rectified surface EMG (rsEMG) were studied during fatigue of isometric contractions of the quadriceps at 30° and 90° knee angles (full extension=0°). Subjects (n=8) were tested on a custom-built ergometer. Maximal voluntary isometric knee extension with supramaximal superimposed burst stimulation (three 100µs pulses; 300Hz), was performed to assess CAR and maximal torque capacity (MTC). Surface EMG signals were obtained from vastus lateralis (VL) and rectus femoris (RF) muscles. At each angle intermittent (15s on 6 off) isometric exercise at 50%MTC with superimposed stimulation was performed to exhaustion. During the fatigue task a sphygmomanometer cuff around the upper thigh ensured full occlusion (400mmHg) of the blood supply to the quadriceps. At least two days separated fatigue tests. MTC was not different between knee angles (30°:229.6±39.3 vs. 90°:215.7±13.2Nm). Endurance times however, were significantly longer (p<0.05) at 30° vs. 90° (87.8±18.7 vs. 54.9±12.1s, respectively) despite the CAR not differing between angles at torque failure (30°:0.95±0.05 vs. 90°:0.96±0.03) and full occlusion of blood supply to the quadriceps. Furthermore, rsEMG values of the VL (normalized to pre-fatigue maximum) were also similar at torque failure (30°:56.5±12.5 vs. 90°:58.3±15.2%), while RF EMG activity was lower at 30° (44.3±12.4%) vs. 90° (69.5±25.3%). We conclude that differences in endurance at different knee angles do not find their origin in differences in central activation and blood flow but may be a consequence of muscle-length related differences in metabolic cost.