Fifteen male subjects performed a repetitive elbow flexion/extension task with a 7-kg mass until exhaustion. Average joint angle, angular velocity, and biceps brachii surface electromyographic (EMG) amplitude (aEMG) and mean power frequency (MPF) were calculated with each consecutive 250-ms segment of data during the entire trial. Data were separated into concentric or eccentric phases and into seven 20 degree-ranges from 0 to 140 degrees of elbow flexion. A regression analysis was used to estimate the rested and fatigued aEMG and MPF values. aEMG values were expressed as a percentage of amplitudes from maximum voluntary contractions (MVC). Under rested dynamic conditions, the average concentric aEMG amplitude was 10% MVC higher than average eccentric values. Rested MPF values were similar for concentric and eccentric phases, although values increased approximately 20 Hz from the most extended to flexed joint angles. Fatigue resulted in an average increase in concentric and eccentric aEMG of 35 and 10% MVC, respectively. The largest concentric aEMG increases (up to 58% MVC) were observed at higher joint velocities, whereas eccentric increases appeared to be related to decreases in velocity. Fatigue had a similar effect on MPF during both concentric and eccentric phases. Larger MPF decreases were observed at shorter muscle lengths such that values within each angle range were very similar by the end of the trial. It was hypothesized that this finding may reflect a biological minimum in conduction velocity before propagation failure occurs.