The effects of ingesting a mixed-snack food (CB), fructose (FRU), or placebo (PBO) prior to exercise (70% peak V̇O2) on the metabolic response during and after cycle exercise were studied in eight normal healthy volunteers with a wide range of peak V̇O2 (30–70 cc·kg−1·min −1). The study was designed to minimize the impact of confounding factors by using various strategies. First, the volunteers were grouped in teams with stratification by peak V̇O2, and the tests were randomized by a Latin-square design. Second, subjects received two acclimation trials in the cycle ergometer to diminish the effect of learning experiences and allow them to get used to the room and equipment. In addition, financial incentives were offered for team and individual endurance times. The test meals were administered 30 min prior to the beginning of exercise, and the subjects exercised to exhaustion, which was defined with clear-cut endpoints. Gas and blood samples were taken at regular intervals before, during, and for 60 min after each exercise bout. CB and FRU induced higher pre-exercise glucose and insulin concentrations. Blood lactate increased 100% with FRU ingestion. Despite these differences; endurance time, substrate, and hormone concentrations as well as rates of substrate oxidation during exercise were identical among the three conditions. During the post-exercise recovery period, PBO was associated with a starvation-like pattern of substrate utilization in which lipid oxidation was 60% greater and carbohydrate oxidation 50% less than following either CB (75 ± 11, 248 ± 27 mg·min−1, P < 0.05) or F ingestion (93 ± 4, 221 ± 14 mg·min−1). Our results demonstrate that pre-exercise feeding does not affect endurance or the metabolic response during exercise, yet it may attenuate post-exercise starvation-like response.