The acromioclavicular (AC) joint enjoys the dubious distinction of being one of the few joints in the body whose total dislocation is routinely treated by simply leaving the joint dislocated. Adherents of both conservative and operative treatment have presented reasons for their viewpoints. Residual shoulder weakness has been offered as a sequela of untreated acromioclavicular injury and a reason for repairing the joint. An objective evaluation of shoulder strength would be valuable in determining the optimum treatment for this injury. The purpose of our study was to quantitate, using the Cybex II, the residual shoulder weakness following various modes of treatment. Seventeen patients with Grade III AC separations and eight patients with Grade II AC sprains were reviewed. Nine of the Grade III injuries were treated and eight nonoperatively. All Grade II injuries were treated nonsurgically. All patients were tested on the Cybex II isokinetic dynamometer at both slow and fast speeds through various ranges of motion. Grade III injuries treated nonoperatively showed no significant strength deficits. Surgically treated Grade III injuries had a significant strength deficit in vertical abduction at fast speeds (19.8%) when compared to the uninjured shoulder. Interestingly, the Grade II injuries led to a significant weakness in horizontal abduction (24.3%) at fast velocity. Evaluation of subjective results showed that Grade III injuries treated conservatively had the most pain and stiffness, despite their strong shoulders. Patients with Grade III injuries treated operatively rated their overall outcome below that of those treated conservatively.(ABSTRACT TRUNCATED AT 250 WORDS)