Previous studies have shown that voluntary motor activity decreases the ability to detect near-threshold electrical stimuli applied to the skin, but has no effect on the perception of either suprathreshold electrical stimuli or natural thermal stimuli (warmth, heat pain). The present study was undertaken to determine if the perception of natural tactile stimuli (vibrotactile) is diminished by motor activity (rhythmical isometric flexions and extensions about the elbow). The stimuli were applied at three different sites on the operant arm--ventral forearm, thenar eminence and distal digit--to examine also the influence of the proximity of the stimulated site to the active muscles on perception. The ability to detect near-threshold stimuli at the two more proximal stimulation sites was significantly reduced during the motor task, and these effects were more pronounced and widespread with higher levels of target force (20 N vs 50 N). Discrimination of small differences in the intensity of suprathreshold stimuli, at all three sites, was unchanged during the motor task. Finally, the subjective intensity of suprathreshold vibrotactile stimuli was reduced, in a nonlinear fashion, during the motor task; proximity again influenced the degree of modulation. In contrast a previous study showed no change in magnitude estimates of suprathreshold electrical stimuli during isotonic flexion and extension. Some possible reasons for the discrepancy are discussed. In addition, our previous suggestion that movement produces a simple reduction in the signal-to-noise ratio (i.e. the gating signal modelled as a masking stimulus) cannot explain the present results, so more complex models are required.