Skeletal muscle fatigue is most often studied in vitro at room temperature and is classically defined as a decline in maximum force production or power output, exclusively linked to repeated isometric contractions. However, most muscles shorten during normal use, and we propose that both the functional correlate of fatigue, as well as the fatigue mechanism, will be different during dynamic contractions compared with static contractions. Under isoflurane anesthesia, fatigue was induced in rat soleus muscles in situ by isotonic shortening contractions at 37 degrees C. Muscles were stimulated repeatedly for 1 s at 30 Hz every 2 s for a total of 15 min. The muscles were allowed to shorten isotonically against a load corresponding to one-third of maximal isometric force. Maximal unloaded shortening velocity (V(0)), maximum force production (F(max)), and isometric relaxation rate (-dF/dt) was reduced after 100 s but returned to almost initial values at the end of the stimulation protocol. Likewise, ATP and creatine phosphate (CrP) were reduced after 100 s, but the level of CrP was partially restored to initial values after 15 min. The rate of isometric force development, the velocity of shortening, and isotonic shortening were also reduced at 100 s, but in striking contrast, did not recover during the remainder of the stimulation protocol. The regulatory myosin light chain (MLC2s) was dephosphorylated after 100 s and did not recover. Although metabolic changes may account for the changes of F(max), -dF/dt, and V(0), dephosphorylation of MLC2s may be involved in the fatigue seen as sustained slower contraction velocities and decreased muscle shortening.