Nine subjects did repeated maximal voluntary contractions (MVCs) of the right knee extensors to induce fatigue. At regular intervals during this process, they matched the force of a 50% MVC in the fatiguing right muscle with a contraction of the unfatigued left muscle that felt like the same force. Perturbations in force sensation were achieved by applying high-frequency vibration (160 Hz) to the patellar tendon of the fatigued muscle during one of a pair of 50% MVCs. The neural drive to the muscle was monitored from the surface electromyogram. Cycles of 10 fatiguing contractions followed by 2 matching contractions were repeated until the subjects could no longer produce 50% of the initial MVC or until they chose to stop. On the average, MVC force fell to 68% of its initial value by the end of 50 fatiguing contractions; almost all of the force loss occurred after completing the first 20 contractions. Maximal electromyograms declined proportionally, and percutaneous shocks delivered during fatiguing contractions did not produce a measurable twitch. This indicated that there was no central fatigue and that force loss was due to failure of the contractile apparatus. In the non-vibrated trials, force sensation increased as a mirror image of force loss. Vibration increased force sensation in the fresh muscle but the effect of vibration disappeared in fatiguing contractions. These data suggest that the sensory analog of muscular fatigue is not necessarily a consequence of alterations in peripheral receptors but may also reside in the central nervous system.