31P magnetic resonance spectroscopy (31P MRS) has been used to measure intramuscular magnesium concentrations and muscle metabolism. Abnormal intramuscular magnesium has been reported in several patient populations with suspected metabolic disorders. The purpose of this study was to evaluate our ability to measure intramuscular magnesium and muscle metabolism in the quadriceps muscles of healthy subjects, and to test whether these measurements were influenced by prior exercise. Twelve normal, healthy male volunteers were tested in a 3 Tesla magnet on four separate days. Resting [Mg2+] was calculated from the heights and frequency shifts of the phosphate, phosphocreatine and ATP peaks. Phosphocreatine (PCr) recovery kinetics were measured after 30-39 second bouts of isometric exercise. Thirty minutes prior to the 3rd test session the subjects completed a 2 hour treadmill walk at 40-60% of heart rate reserve. Resting [Mg2+] averaged 0.388 mM and had an interclass correlation coefficient between days (ICC) of 0.352. The mean end exercise PCr was 47.6% and the mean end exercise pH was 6.97. PCr recovery averaged 39 seconds (p = 0.892) and had an ICC of 0.819. Prior long duration exercise did not produce significant alterations in either PCr recovery kinetics or intracellular magnesium levels (p = 0.440). In conclusion, the reproducibility of Resting [Mg2+] was less than that of PCr recovery measurements, and may reflect the sensitivity of these measurements to phasing errors. In addition, prior exercise is unlikely to alter measurements of resting metabolites or muscle metabolism suggesting that rigorous control of physical activity prior to metabolic testing is unnecessary.