The effect of pre-exercise meal composition on metabolism and performance in cycling were investigated in a crossover study. Twelve competitive cyclists ingested high-fat, high-carbohydrate, or high-protein meals 90 min before a weekly exercise test. The test consisted of a 1-hour pre-load at 55% peak power, five 10-min incremental loads from 55 to 82% peak power (to measure the peak fat-oxidation rate), and a 50-km time trial that included three 1-km and 4-km sprints. A carbohydrate supplement was ingested throughout the exercise. Relative to the high-protein and high-fat meals, the high-carbohydrate meal halved the peak fat-oxidation rate and reduced the fat oxidation across all workloads by a factor of 0.20 to 0.58 (p =.002-.0001). Reduced fat availability may have accounted for this reduction, as indicated by lower plasma fatty acid, lower glycerol, and higher pre-exercise insulin concentrations relative to the other meals (p =.04-.0001). In contrast, fat oxidation following the high-protein meal was similar to that following the high-fat meal. This similarity was linked to evidence suggesting greater lipolysis and plasma fat availability following high-protein relative to high-carbohydrate meals. Despite these substantial effects on metabolism, meal composition had no clear effect on sprint or 50-km performance.