In the study of skeletal muscle bioenergetics, 31P magnetic resonance spectroscopy (MRS) allows frequent measurement of the cytosolic pH and the concentrations of phosphocreatine, inorganic phosphate, and adenosine diphosphate (ADP) during exercise and recovery, which can be supplemented by 1H MRS (or biopsy) measurements of muscle lactate content and 13C MRS (or biopsy) measurements of muscle glycogen. We review the many methods now described by which 31P MRS measurements can be made to yield quantitative estimates of adenosine triphosphate (ATP) turnover, oxidative capacity, and proton handling in skeletal muscle. In particular, we describe how to estimate the rates of glycogenolytic and aerobic ATP synthesis during exercise and oxidative ATP synthesis and proton efflux during recovery from exercise and how to assess oxidative capacity using data from steady-state exercise, work jumps, or recovery. We discuss the metabolic relationships that make these methods possible and the assumptions (eg, about cytosolic buffer capacity and mitochondrial control mechanisms) on which they depend. We show how these methods, although sometimes based on apparently conflicting metabolic models, can be analysed in a common framework. Finally, we discuss some examples of the current and potential applications of these methods in clinical and experimental studies of skeletal muscle.