Muscle metabolism and force production were studied in sprint trained runners, endurance trained runners and in untrained subjects, using 31P-MRS. 31P-spectra were obtained at a time resolution of 5 s during four maximal isometric contractions of 30-sec duration, interspersed by 60-sec recovery intervals. Resting CrP/ATP ratio averaged 3.3 +/- 0.3, with no difference among the three groups. The sprint trained subjects showed about 20 % larger contraction forces in contraction bouts 1 and 2 (p < 0.05). The groups differed with respect to CrP breakdown (p < 0.05), with sprinters demonstrating about 75 % breakdown in each contraction compared to about 60 % and 40 % for untrained and endurance trained subjects, respectively (p < 0.05). The endurance trained runners showed almost twice as fast CrP recovery (t 1/2 = 12.5 +/- 1.5) compared to sprint trained (t 1/2 = 22.5 +/- 2.53) and untrained subjects (t 1/2 = 26.4 +/- 2.8). From the initial rate of CrP resynthesis the rate of maximal aerobic ATP synthesis was estimated to 0.74 +/- 0.07, 0.73 +/- 0.10 and 0.33 +/- 0.07 mmol ATP x kg -1 wet muscle x sec -1 for sprint trained, endurance trained and untrained subjects, respectively. Only the sprint trained and the untrained subjects displayed a significant drop in pH and only during the first of the four contractions, about 0.2 and 0.1 pH units, respectively, indicating that only under those contractions was the glycolytic proton production larger than the proton consumption by the CK reaction. Also, in the first contraction the energy cost of contraction was higher for the sprinters compared to the two other groups. The simple 31P-MRS protocol used in the present study demonstrates marked differences in force production, aerobic as well as anaerobic muscle metabolism, clearly allowing differentiation between endurance trained, sprint trained and untrained subjects.