In this research on motor short-term memory it was hypothesized that subjects making free-selected movements would evidence superior reproduction accuracy over subjects making preselected movements, whose reproduction performance would be superior to subjects making constrained movements. Subjects making free-selected movements were allowed to determine distance moved, terminal location, and speed of response. Subjects made preselected movements under similar conditons, except that the experimenter reminded them of the importance of preselections, as well as to move at a slow but continuous speed, and not to choose the same movement more than once during the sequence. When making constrained movements, subjects moved at a slow but continuous speed to a terminal location that was chosen at random by the experimenter. The methodology and design of this research is described in this report. Results indicated that the reproduction accuracy of free-selected movements were superior to that of preselected movements, which in turn exhibited less reproduction error than constrained movements. However, those performers making preselected movements were no less variable than those making constrained movements. A discussion of the possible bias provided by experimenter instruction is included. (JD)