We have previously shown that action tremor persists during maximal wrist extension in patients with Parkinson's disease, and that this contributes to weakness at this joint by preventing the fully fused contraction of the forearm extensor muscles. Antiparkinsonian medication reduces the action tremor in torque and electromyographic (EMG) records, thereby improving strength at the wrist. In the present experiments, peak torque and action tremor were recorded during maximal extension of the elbow in nine patients with Parkinson's disease while they were on and off antiparkinsonian medication, and in eight age- and sex-matched healthy subjects. Peak torque and mean rectified EMG levels were reduced by 34% and 36%, respectively, during maximal elbow extension when patients off medication were compared to those in the treated state. Action tremor was visible in torque records and had a frequency of approximately 10 Hz both in parkinsonian patients and in normal controls. Activity of a similar frequency was often detected in EMG records, especially in patients off therapy. The absolute amplitude of action tremor in torque (A[torque]) and EMG (A[EMG]) records from the elbow was unaffected by therapy and was little different from that recorded in healthy controls. The relative action tremor in torque ([A(torque)/peak torque] x 100) and EMG ([A(EMG)/mean rectified EMG] x 100) was reduced by treatment, but this was the result of the increase in peak torque and mean rectified EMG. Thus, in contrast to the results at the wrist, antiparkinsonian medication has little effect on the 10-Hz action tremor at the elbow. Increased strength is produced by a different mechanism which allows recruitment of triceps motor units outside of this synchronizing influence.