The response of postural wrist tremors to brief mechanical displacements was compared in two groups of patients, one with classical hereditary essential tremor (n = 18) and another with typical Parkinson's disease (n = 13). These groups were compared with an additional group of normal subjects mimicking wrist tremor (n = 9). The degree to which brief mechanical displacements of the wrist produced by torque pulses of three different sizes could modulate the timing of rhythmic electromyographic bursts in the forearm flexor muscles was quantified by deriving a resetting index, which could range between 0 (no phase resetting) and 1 (complete phase resetting). In all three groups of subjects studied, the resetting index varied significantly with the size of the mechanical perturbation and, in an inverse fashion, with the ongoing tremor amplitude. When due allowance for these factors was made, the difference in mean resetting indexes between the three groups of patients and subjects was reduced to the extent that no definitive statement could be made as to whether brief mechanical perturbations had more effect on essential tremor than parkinsonian tremor. The method is therefore unlikely to be useful in differentiating the common causes of postural wrist tremors.