Context: Investigators have observed electromyographic (EMG) activity of the supraspinatus muscle and reported conflicting results. Objective: To quantify EMG activity of the supraspinatus, middle deltoid, and posterior deltoid muscles during exercises commonly used in rehabilitation. Design: One-factor, repeated-measures design. Setting: Controlled laboratory. Patients or Other Participants: Twenty-two asymptomatic subjects (15 men, 7 women) with no history of shoulder injury participated. Main Outcomes Measure(s): The dominant shoulder was tested. Fine-wire EMG electrodes were inserted into the supraspinatus, middle deltoid, and posterior deltoid muscles. The EMG data were collected at 960 Hz for analysis during maximal voluntary isometric contraction (MVIC) and 5 repetitions of 3 exercises: standing elevation in the scapular plane ("full can"), standing elevation in the scapular plane with glenohumeral internal rotation ("empty can"), and prone horizontal abduction at 100° with glenohumeral external rotation ("prone full can"). We calculated 1-way repeated-measures analysis of variance (P < .05) and post hoc 2-tailed, paired t tests to detect significant differences in muscle activity among exercises. Results: No statistical difference existed among the exercises for the supraspinatus. The middle deltoid showed significantly greater activity during the empty-can exercise (77 ± 44% MVIC) and prone full-can exercise (63 ± 31% MVIC) than during the full-can exercise (52 ± 27% MVIC) (P = .001 and .017, respectively). The posterior deltoid showed significantly greater activity during the prone full-can exercise (87 ± 53% MVIC) than during the full-can (P = .001) and the empty-can (P = .005) exercises and significantly greater activity during the empty-can exercise (54 ± 24% MVIC) than during the full-can exercise (38 ± 32% MVIC) (P = .012). Conclusions: While all 3 exercises produced similar amounts of supraspinatus activity, the full-can exercise produced significantly less activity of the deltoid muscles and may be the optimal position to recruit the supraspinatus muscle for rehabilitation and testing. The empty-can exercise may be a good exercise to recruit the middle deltoid muscle, and the prone full-can exercise may be a good exercise to recruit the posterior deltoid muscle. © by the National Athletic Trainers' Association, Inc.