Magnetic resonance spectroscopy is a non-invasive and repeatable method of studying muscle metabolism. Magnetic resonance spectroscopy uses specific radiofrequency pulses in a strong magnetic field to determine the relative concentrations of chemical compounds in the sample. 31P Magnetic resonance spectroscopy provides indirect measures of phosphate compounds such as adenosine triphosphate (ATP), phosphocreatine and inorganic phosphate. Muscle intracellular pH can also be determined. Exercise tests can be performed in the magnet such that the metabolic response to steady-state exercise can be measured. The ratio of inorganic phosphate to phosphocreatine reflects the relative metabolic rate of mitochondrial respiration (V) and the extrapolated maximum capacity of oxidative metabolism (Vm). Normal humans vary considerably in their metabolic response to exercise. These differences are reflected in their Vms and the degree of acidosis during exercise. Active muscles in endurance trained athletes have higher Vms and faster recovery rates than normal controls. Preliminary studies have been done to assess muscle glycolytic capacity by measuring the degree of acidosis during ischaemic exercise. Exercise-induced muscle injury can be detected as an increased inorganic phosphate to phosphocreatine ratio in resting muscle. The increase in the inorganic phosphate to phosphocreatine ratio with injury reaches a peak 1 to 2 days after the injury and lasts for up to a week. Similar increases in the inorganic phosphate to phosphocreatine ratio occur in patients with destructive neuromuscular diseases. Thus changes in the resting inorganic phosphate to phosphocreatine ratio may be used to detect the degree of muscle injury following exercise. Levels of H2PO4- in muscle are thought to be important in causing muscle fatigue during exercise. As 31P magnetic resonance spectroscopy can measure H2PO4-, magnetic resonance spectroscopy has become a useful technique in the study of the metabolic causes of muscle fatigue. It may also be possible to identify the relative populations of fast twitch and slow twitch fibres in a skeletal muscle using pH changes measured with 31P magnetic resonance spectroscopy. Magnetic resonance spectroscopy using other nuclei, such as 1H, 13C and 23Na, have the potential to provide information on other metabolic changes which occur with exercise. Magnetic resonance spectroscopy has shown promise as a technique to monitor the effects of training, including overtraining, in specific muscle groups in athletes.