Bilateral sanding is used by occupational therapists as a therapeutic exercise technique, however, to-date no data have been available identifying the musculature involved in the activity nor the extent of that involvement. The dual purpose of this pilot study was to; (a) delineate the activity of selected shoulder muscles during bilateral sanding, and (b) determine whether the method used would be feasible in the analysis of other therapeutic activities. For this investigation, a quadriplegic individual was matched for height and weight with a normal subject and both were tested under several controlled conditions of bilateral sanding. Right shoulder complex musculature was electromyographically (EMG) investigated and the muscles studied were: pectoralis major, biceps brachii, the three heads of deltoid, and the lateral and medial heads of t riceps brachii. EMG data for the two subjects were normalized and represented as percentages of the maximal voluntary contractions (MVC) relative for each muscle. This technique allowed intra-and inter-subject comparisons of muscle activity.Recognizing the limitations of the study with the major one being the small number of subjects, useful information was obtained from the results. The quadriplegic individual appeared to be more inefficient in his muscle usage than the normal subject. This may be related to trunk instability in the quadriplegic subject. All three heads of the deltoid appeared to be the most highly used of monitored muscles in both the normal and the quadriplegic with t riceps being only moderately active in the quadriplegic individual. Spring resisted bilateral sanding produced the most EMG activity in all muscles studied in both subjects. Further investigations employing a larger number of subjects may further elucidate EMG activity during bilateral sanding.It would appear that EMG investigations of therapeutic activities offer an informative method of physical activity analysis.