The ability to measure skeletal muscle motion with phase-contrast magnetic resonance (MR) imaging was tested with a motion phantom that simulated muscle activity. Quantitative analytic data on unidimensional, bidirectional skeletal muscle motion measured in vivo was obtained in four healthy volunteers. MR images of the subjects' forearms were obtained during flexion and extension of the fingers and of the anterior and posterior muscle compartments of the lower leg with various resistances to ankle dorsiflexion and plantar flexion. It was necessary to correct the data for the effects of eddy currents. In vitro evaluation of the technique was done by studying through-plane sinusoidal motion of solid objects. The largest error was underestimation of the peak excursion of 11.5 mm by 0.09 mm (the root mean square error for the cycle was 0.04 mm) In vivo experiments demonstrated the contraction of muscles in relation to each other. Data acquisition and analysis techniques must be refined, but measuring skeletal muscle motion with phase-contrast MR imaging should enhance the understanding of bioengineering fundamentals and muscular changes in disease and adaptation.