The validity and reliability of hand-held dynamometry and Cybex dynamometry were investigated using maximal isometric contractions of the external rotators of the shoulder in 14 healthy male subjects. Three maximum voluntary contractions were recorded from each subject by a hand-held dynamometer and a Cybex isokinetic dynamometer at two testing sessions approximately 1 week apart. Analysis of variance did not reveal any significant differences between the mean peak torques obtained with either instrument or between days. The intrarater reliability was clearly established for both the hand-held dynamometer (r = 0.986) and Cybex dynamometer (r = 0.993). Within-day correlations between the two instruments accounted for 27% (day 1) and 60% (day 2) of the explainable variance. This suggests that although both techniques produced identical peak torque values and measured the same element of performance-strength, they did so in a slightly different manner. The possible nature of these differences is discussed.J Orthop Sports Phys Ther 1988;10(6):213-217.