In a pursuit-tracking task consisting of 100 positioning movements between targets at 5 fixed positions, target location (proximity to the boundary of the task) varied independently of movement amplitude. Eighty-seven subjects performed 12 trials of the task, with target width (at 3 levels) as a between-subject variable. A microprocessor system detected the location of the end-point of each primary movement. Movement accuracy (measured as end-point dispersion not constant error) varied with target location but not movement amplitude, while movement time varied with both factors. The effect of target width on movement accuracy was less consistent. The observed effects are discussed in terms of a mass-spring model of muscular action. It is concluded that apart from having important consequences for the design of positioning experiments, these results call into question information-processing and impulse-variability theories which implicate movement amplitude in determining movement accuracy, and support theories which emphasize target location.