1. The fatigue of force that occurs during the first 60 s of a maximum voluntary contraction of the human quadriceps has been examined by comparing the voluntary force with that obtained by brief tetanic stimulation at 50 Hz in nine healthy subjects. In three subjects the voluntary force declined in parallel with the tetanic force whereas in the remainder it fell more rapidly, suggesting that central fatigue was present. 2. For those subjects who showed little or no central fatigue, surface electromyograph (EMG) activity remained approximately constant while the force declined by about 60%. In the others, EMG activity and force declined in parallel but when an extra effort was made the subjects could briefly increase their force and this was accompanied by a proportionately greater increase in EMG activity (generally up to the original value). 3. It is concluded that in sustained maximum voluntary contractions of the quadriceps (a) central fatigue may account for an appreciable proportion of the force loss, (b) surface EMG recordings provide no evidence that neuromuscular junction failure is the limiting factor determining the loss of force in this muscle.