Exercise is known to produce changes in the amount and distribution of water in skeletal muscle. Because MR imaging is highly sensitive to changes in water distribution, these changes should be detectable under appropriate imaging conditions. Imaging of the forearms and/or legs was performed in 16 volunteers at 0.35 T, before and after exercise. Exercises included finger flexion and extension, wrist flexion, ankle plantar flexion, and great toe extension. In the case of handgrip exercise, the level of exertion was quantitated. Individual muscles were frequently indistinguishable on preexercise scans. After exercise, active and inactive muscles could be clearly distinguished. For example, in the flexor digitorum profundus, finger flexion resulted in an increase in the image-derived estimate of T1 (T1 postexercise was 1037 +/- 162 msec vs T1 preexercise of 590 +/- 49 msec, p less than .001). T2 also increased (T2 postexercise was 35 +/- 2 msec vs T2 preexercise of 28 +/- 1 msec, p less than .001). Relative spin density also increased (p less than .001). T1, T2, and spin density subsequently decreased with time but were still increased above baseline at 10 min postexercise (p less than .005). Signal changes correlated moderately with the level of exertion (r = .63) and fatigue (r = .45). Vascular occlusion did not prevent intensity changes. Thus, changes in skeletal muscle MR signal intensity occur with exercise and appear to parallel known alterations in water distribution.