To investigate the hypothesis that training-induced increases in muscle mitochondrial potential are not obligatory to metabolic adaptations observed during submaximal exercise, regardless of peak aerobic power (V) over dot o(2peak)) Of the subjects, a short-term training study was utilized. Two groups of untrained male subjects (n = 7/group), one with a high (HI) and the other with a low (LO) (V) over dot o(2peak) (means +/- SE; 51.4 +/- 0.90 vs. 41.0 +/- 1.3 ml . kg(-1) . min(-1); P < 0.05), cycled for 2 h/day at 66-69% of (V) over dot o(2peak) for 6 days. Muscle tissue was extracted from vastus lateralis at 0, 3, and 30 min of standardized cycle exercise before training (0 days) and after 3 and 6 days of training and analyzed for metabolic and enzymatic changes. During exercise after 3 days of training in the combined HI + LO group, higher (P < 0.05) concentrations (mmol/kg dry wt) of phosphocreatine (40.5 +/- 3.4 vs. 52.2 +/- 4.2) and lower (P < 0.05) concentrations of P-i (61.5 +/- 4.4 vs. 53.3 +/- 4.4), inosine monophosphate (0.520 +/- 0.19 vs. 0.151 +/- 0.05), and lactate (37.9 +/- 5.5 vs. 22.8 +/- 4.8) were observed. These changes were also accompanied by reduced levels of calculated free ADP, AMP, and P-i. All adaptations were fully expressed by 3 min of exercise and by 3 days of training and were independent of initial (V) over dot o(2peak) levels. Moreover, maximal activity of citrate synthase, a measure of mitochondrial capacity, was only increased with 6 days of training (5.71 +/- 0.29 vs. 7.18 +/- 0.37 mol . kg protein(-1) . h(-1); P < 0.05). These results demonstrate that metabolic adaptations to prolonged exercise occur within the first 3 days of training and during the non-steady-state period. Moreover, neither time course nor magnitude of metabolic adaptations appears to depend on increases in mitochondrial potential or on initial aerobic power.