We determined the effects of varying daily carbohydrate intake by providing or withholding carbohydrate during daily training on endurance performance, whole body rates of substrate oxidation, and selected mitochondrial enzymes. Sixteen endurance-trained cyclists or triathletes were pair matched and randomly allocated to either a high-carbohydrate group (High group; n = 8) or an energy-matched low-carbohydrate group (Low group; n = 8) for 28 days. Immediately before study commencement and during the final 5 days, subjects undertook a 5-day test block in which they completed an exercise trial consisting of a 100 min of steady-state cycling (100SS) followed by a 7-kJ/kg time trial on two occasions separated by 72 h. In a counterbalanced design, subjects consumed either water (water trial) or a 10% glucose solution (glucose trial) throughout the exercise trial. A muscle biopsy was taken from the vastus lateralis muscle on day 1 of the first test block, and rates of substrate oxidation were determined throughout 100SS. Training induced a marked increase in maximal citrate synthase activity after the intervention in the High group (27 vs. 34 micromol x g(-1) x min(-1), P < 0.001). Tracer-derived estimates of exogenous glucose oxidation during 100SS in the glucose trial increased from 54.6 to 63.6 g (P < 0.01) in the High group with no change in the Low group. Cycling performance improved by approximately 6% after training. We conclude that altering total daily carbohydrate intake by providing or withholding carbohydrate during daily training in trained athletes results in differences in selected metabolic adaptations to exercise, including the oxidation of exogenous carbohydrate. However, these metabolic changes do not alter the training-induced magnitude of increase in exercise performance.