Six men were studied during exercise to exhaustion on a cycle ergometer at 73% of VO2max following ingestion of glycerol, glucose or placebo. Five of the subjects exercised for longer on the glucose trial compared to the placebo trial (p less than 0.1; 108.8 vs 95.9 min). Exercise time to exhaustion on the glucose trial was longer (p less than 0.01) than on the glycerol trial (86.0 min). No difference in performance was found between the glycerol and placebo trials. The ingestion of glucose (lg X kg-1 body weight) 45 min before exercise produced a 50% rise in blood glucose and a 3-fold rise in plasma insulin at zero min of exercise. Total carbohydrate oxidation was increased by 26% compared to placebo and none of the subjects exhibited a fall in blood glucose below 4 mmol X 1-1 during the exercise. The ingestion of glycerol (lg X kg-1 body weight) 45 min before exercise produced a 340-fold increase in blood glycerol concentration at zero min of exercise, but did not affect resting blood glucose or plasma insulin levels; blood glucose levels were up to 14% higher (p less than 0.05) in the later stages of exercise and at exhaustion compared to the placebo or glucose trials. Both glycerol and glucose feedings lowered the magnitude of the rise in plasma FFA during exercise compared to placebo. Levels of blood lactate and alanine during exercise were not different on the 3 dietary treatments.(ABSTRACT TRUNCATED AT 250 WORDS)