Mechanomyography (MMG) provides a measure of muscle mechanical changes during contractions. The purpose of this study was to quantify alterations in MMG signals during fatigue at two muscle lengths. Comparisons with electromyographic (EMG) recordings were made. A group of 13 subjects performed isometric dorsiflexions (50% of maximum for 60 s) at 40° of plantarflexion (long, ll) and 5° of dorsiflexion (short, ls). The mean power frequency of the EMG (f¯EMG) and MMG (f¯MMG) signals and the mean rectified MMG ( ) and EMG ( ) were determined over each 1-s period, normalized to the respective maximal value, regressed against time, and the resulting slopes (units = %max · s−1) were analyzed. The slopes were larger (P = 0.007) at ll compared to ls [mean ll 0.50 (SD 0.26), mean ls 0.27 (SD 0.16)], however there were no differences (P = 0.24) between mean f¯MMG slopes [ll−0.10 (SD 0.16), ls−0.16 (SD 0.11)]. Similarly, slopes were larger (P = 0.001) at ll versus ls [ll 0.26 (SD 0.13), ls 0.08 (SD 0.15)] and there were no differences (P=0.89) between mean f¯EMG slopes [ll−0.15 (SD 0.14), ls−0.14 (SD 0.12)]. At 5 s following the exercise to fatigue mean MVC (units = %max) were not significantly different between ll and ls [P = 0.08; ll 78.8 (SD 9.1), ls 85.2 (SD 6.0)]. These results showed that during fatiguing contractions, MMG and EMG amplitudes increased while frequency characteristics decreased at both muscle lengths. The change in and was greater at ll but no differences in f¯MMG or f¯EMG slopes occurred between lengths. These results would suggest that larger increases in motor unit recruitment occur with time during fatigue at ll compared to ls.