The present study was designed to determine whether fatigue alters the ability to estimate an index of individual muscle force from shear elastic modulus measurements ( experiment I), and to test the ability of this technique to highlight changes in load sharing within a redundant muscle group during an isometric fatiguing task ( experiment II). Twelve subjects participated in experiment I, which consisted of smooth linear torque ramps from 0 to 80% of maximal voluntary contraction (MVC) performed before and after an isometric fatigue protocol, beginning at 40% of MVC and stopped when the force production dropped below 30% of MVC. Although the relationships between modulus and torque were very similar for pre- and postfatigue [root mean square deviation (RMSdeviation) = 3.7 ± 2.6% of MVC], the relationships between electromyography activity level and torque were greatly altered by fatigue (RMSdeviation = 10.3 ± 2.6% of MVC). During the fatiguing contraction, shear elastic modulus provided a significantly lower RMSdeviation between measured torque and estimated torque than electromyography activity level (5.7 ± 0.9 vs. 15.3 ± 3.8% of MVC). Experiment II performed with eight participants consisted of an isometric knee extension at 25% of MVC sustained until exhaustion. Opposite changes in shear elastic modulus were observed between synergists (vastus medialis, vastus lateralis, and rectus femoris) of some participants, reflecting changes in load sharing. In conclusion, despite the fact that we did not directly estimate muscle force (in Newtons), this is the first demonstration of an experimental technique to accurately quantify relative changes in force in an individual human muscle during a fatiguing contraction.