The purpose of this study was to test the hypothesis that shoulder rhythm is affected by rotator cuff pathology during arm elevation. We divided 42 subjects into 3 groups: those with full-thickness rotator cuff tears (RCTs) (n = 14), those with tendinopathy (n = 13), and control subjects (n = 15). Shoulder kinematics was recorded while subjects performed elevation in the sagittal and scapular planes. Euler angles were computed for scapular and humeral elevation. Data were divided into 3 equal phases. Lines were fit and the slope determined for each phase. Significant differences in slopes ( P < .05) were found between experimental groups for both motions. The RCT group had higher slopes in the initial and middle phases for sagittal elevation and in the middle phase for scapular abduction. The scapula was elevated more in the RCT group in the initial two thirds of movement. This motion may change the length of remaining muscles so that they operate on a more effective part of their length-tension curve.