Muscle oxidative function has been investigated in subjects with various training status (VO2 max, 41–72 mL O2 kg-1 body wt min-1, n=10). Mitochondria were isolated from biopsies taken from m. vastus lateralis. Maximal mitochondrial oxygen consumption (QO2) and ATP production (MAPR) were measured with polarographic and bioluminometric techniques, respectively. The yield of mitochondria, calculated from the fractional activity of citrate synthase (CS), averaged 26%. With pyruvate + malate, the respiratory control ratio was 5.7 ± 0.4 (X ± SE) and the P/O ratio was 2.83 ± 0.02, which demonstrates that the isolated mitochondria were functionally intact. QO2 was significantly correlated to aerobic training status expressed as muscle CS activity (r=0.86), VO2 max (r=0.84) and lactate threshold (r=0.83) but not to the fibre type composition. A highly significant correlation (r=0.93) was observed between ATP production calculated from QO2 and MAPR, but ATP production derived from QO2 was higher than MAPR both for pyruvate + malate (255%) and for a-ketoglutarate (23%). QO2 extrapolated to a temperature of 38 °C averaged 68 mL O2 min-1 kg-1 wet wt, which is similar to previous findings in vitro and in vivo during the post-exercise period. However, calculated muscle O2 utilization during exercise was three- to fivefold higher than QO2 measured on isolated mitochondria. It is suggested that additional factors exist for activation of mitochondrial respiration during exercise. It is concluded that muscle oxidative function can be quantitatively assessed from the respiration of mitochondria isolated from needle biopsy specimens and that QO2 is closely correlated to whole-body VO2 max.