1. It is well established that tactile acuity is greater over digits than over the proximal parts of the upper limb and that the corticospinal projection is especially dense for distal muscles. To determine whether the acuity for judgements of forces exerted by distal muscles differed from that for proximal muscles, a weight-matching task was used with first dorsal interosseous, flexor pollicus longus and elbow flexors. 2. Reference weights equivalent to approximately 3% (light) and 15% (heavy) of the maximal voluntary contraction were lifted by one muscle group on the left. They were matched with a variable weight, lifted by the same group, on the right. 3. In naive subjects, the coefficient of variation for repeated estimates of perceived heaviness was significantly lower for proximal than distal muscles. Measured in this way, 'accuracy' (i.e. reproducibility of the estimates) was not greater for the intrinsic muscles of the hand. This result could not be explained by the way in which the weights were supported by the index finger. When the data were expressed as the relative difference between the reference and the matched weight, each muscle group behaved similarly. 4. For a particular muscle, accuracy was greater when the heavy rather than the light weights were lifted. 5. Given that estimates of forces and heaviness are biased by signals of central motor command (McCloskey, 1981; Gandevia, 1987; Cafarelli, 1988; Jones, 1988). these signals could be graded no more finely for distal than proximal muscles. Furthermore, relative accuracy was greater for forces at the high rather than the low end of the comfortable 'matching' range of force for a particular muscle.