31P magnetic resonance spectroscopy (31P MRS) can yield much information about bioenergetics in skeletal muscle. During mixed aerobic/glycolytic exercise, changes in phosphocreatine (PCr) concentration and pH may be abnormal because of reduced muscle mass or reduced efficiency (which the authors combine here as "effective muscle mass") or because of reduced oxidative capacity. The authors show how these can be distinguished by calculating the nonoxidative and oxidative costs of mechanical work, and also of work per unit of effective muscle mass (measured using the initial rate of ATP turnover). These quantities are substantially time-independent during incremental exercise, and so can be used to compare exercise studies of differing duration. The authors illustrate this analysis by showing that in dialyzed patients with chronic renal failure, the substantial exercise abnormalities seen by 31P MRS are due mainly to a decrease in effective muscle mass, which outweighs the oxidative defect implied by the abnormal PCr recovery kinetics.