Over the past decade, randomized controlled trials have established that, after eating either a low-carbohydrate, high-fat or a high-carbohydrate, low-fat diet for 4-6 wk, trained athletes performed equally well during a maximum oxygen consumption (VO2max) test; during 5 and 1.6 km laboratory treadmill time trials; during a 6 × 800 m interval repetition session; and during a prolonged cycling test to exhaustion at 70%VO2max. Indeed, during the 6 × 800 m interval repetition session, some subjects achieved the highest rates of fat oxidation (2 g/min) ever reported in humans; whereas ingestion of 10 g carbohydrate/h improved prolonged cycling test performance by 22%, equally following either diet. These data establish that muscle glycogen is not an obligatory fuel for exercise. Rather, exercise-induced hypoglycemia due to depletion of glucose in the small glucose pool in the liver and bloodstream, prevented by minimal carbohydrate ingestion during exercise, is the main metabolic contributor to premature fatigue during more prolonged submaximal exercise.