For 5 days, eight well-trained cyclists consumed a random order of a high-carbohydrate (CHO) diet (9.6 g. kg(-1). day(-1) CHO, 0.7 g. kg(-1). day(-1) fat; HCHO) or an isoenergetic high-fat diet (2.4 g. kg(-1). day(-1) CHO, 4 g. kg(-1). day(-1) fat; Fat-adapt) while undertaking supervised training. On day 6, subjects ingested high CHO and rested before performance testing on day 7 [2 h cycling at 70% maximal O(2) consumption (SS) + 7 kJ/kg time trial (TT)]. With Fat-adapt, 5 days of high-fat diet reduced respiratory exchange ratio (RER) during cycling at 70% maximal O(2) consumption; this was partially restored by 1 day of high CHO [0.90 +/- 0.01 vs. 0.82 +/- 0.01 (P < 0.05) vs. 0.87 +/- 0.01 (P < 0.05), for day 1, day 6, and day 7, respectively]. Corresponding RER values on HCHO trial were [0. 91 +/- 0.01 vs. 0.88 +/- 0.01 (P < 0.05) vs. 0.93 +/- 0.01 (P < 0.05)]. During SS, estimated fat oxidation increased [94 +/- 6 vs. 61 +/- 5 g (P < 0.05)], whereas CHO oxidation decreased [271 +/- 16 vs. 342 +/- 14 g (P < 0.05)] for Fat-adapt compared with HCHO. Tracer-derived estimates of plasma glucose uptake revealed no differences between treatments, suggesting muscle glycogen sparing accounted for reduced CHO oxidation. Direct assessment of muscle glycogen utilization showed a similar order of sparing (260 +/- 26 vs. 360 +/- 43 mmol/kg dry wt; P = 0.06). TT performance was 30.73 +/- 1.12 vs. 34.17 +/- 2.48 min for Fat-adapt and HCHO (P = 0.21). These data show significant metabolic adaptations with a brief period of high-fat intake, which persist even after restoration of CHO availability. However, there was no evidence of a clear benefit of fat adaptation to cycling performance.