Until recently, assessment of muscle metabolism was only possible by invasive sampling. 31P magnetic resonance spectroscopy ( 31P MRS) offers a way to study muscle metabolism non-invasively. The aim of the present study was to use spatially-resolved 31P MRS to assess the metabolism of the quadriceps muscle in sprint-trained, endurance-trained and untrained individuals during exercise and recovery. 5 sprint-trained (STA), 5 endurance-trained (ETA) and 7 untrained individuals (UTI) completed one unlocalized 31P MRS session to measure phosphocreatine (PCr) recovery, and a second session in which spatially-resolved 31P MR spectra were obtained. PCr recovery time constant (τ) was significantly longer in STA (50±17 s) and UTI (41±9 s) than in ETA (30±4 s), (P<0.05). PCr changes during exercise differed between the groups, but were uniform across the different components of the quadriceps within each group. pH during recovery was higher for the ETA than for the UTI (P<0.05) and also higher than for the STA (P<0.01). Muscle volume was greater in STA than in UTI (P<0.05) but not different from ETA. Dynamic 31P MRS revealed considerable differences among endurance and sprint athletes and untrained people. This non-invasive method offers a way to quantify differences between individual muscles and muscle components in athletes compared to untrained individuals. © 2013 Georg Thieme Verlag KG Stuttgart.New York.