Peak and average moment of concentric elbow flexion was assessed in 30 males without previous elbow or shoulder injuries. Isokinetic assessment was performed with the forearm in supination and in the neutral position, at isokinetic velocities of 30, 60 and 120degrees.s(-1) for both dominant and nondominant sides, Eight volunteers were examined twice to determine test-retest reliability and estimate the variability of isokinetic moment data associated with normal fluctuations of strength performance over repeated measurements (inter-session variability) using the intraclass correlation coefficient (ICC 3,1) and the 95% standard error of measurement (95% SEM). The ICC values and 95% SEM ranged between 0.76-0.99 and 3.3-16.2 Nm, respectively. Our findings revealed higher reliability and lower inter-session variability for average compared to peak moment data, the dominant compared to the non-dominant side and at the angular velocity of 120 compared to 30 or 60degrees.s(-1). Isokinetic moment of elbow Hex ors was significantly greater with the forearm in supination compared to the neutral position (p < 0.001) regardless of the side or the isokinetic velocity used. Isokinetic strength of elbow flexion was significantly decreased, as the velocity increased (p < 0.001). The dominant side was also stronger compared to the non-dominant side (p < 0.05). The results of our study suggest that elbow flexors can be reliably assessed with an isokinetic dynamometer, Consistent positioning of the forearm is necessary when isokinetic elbow flexion is monitored in longitudinal studies. A bilateral difference of less than 5% should be taken into account when the contra-lateral is used for comparison in pre and post-injury or rehabilitation situations.