On two occasions six competitive, male triathletes performed in random order, two experimental trials consisting of either a timed ride to exhaustion on a cycle-ergometer or a run to exhaustion on a motor-driven treadmill at 80% of their respective peak cycling and peak running oxygen uptakes (VO₂peak)- At the start of exercise, subjects drank 250 ml of a 15 g.100 ml⁻¹ w.v⁻¹ glucose solution with U-¹⁴C glucose added as tracer and, thereafter, 150 ml of the same solution every 15 min. Despite identical metabolic rates (VO₂ 3.51 ±0.06 vs. 3.51 ±0.10 l.min⁻¹; values are mean± SEM for the cycling and running trials, respectively), exercise times to exhaustion were significantly longer during cycling than running (96 ±14 vs. 63 ±11 min; P<0.05). The superior cycling than running endurance was not associated with any differences in either the rate of blood glucose oxidation (3.8 ±0.1 vs. 3.9 ±0.4 mmol.min⁻¹ ), nor the rate of ingested glucose oxidation (2.0 ± 0.1 vs. 1.7 ±0.2 mmol.min⁻¹) at the last common time point (40 min) before exhaustion, despite higher blood glucose concentrations at exhaustion during running than cycling (7.0 ±0.9 vs. 5.8 ±0.5 mmol.l⁻¹; P<0.05). However, the final rate of total CHO oxidation was significantly greater during cycling than running (24.0 ±0.8 vs. 21.7 ±1.4 mmol C6 .min⁻¹;P<0.01). At exhaustion, the estimated contribution to energy production from muscle glycogen had declined to similar extents in both cycling and running (68 ±3 vs. 65 ± 5%). These differences between the rates of total CHO oxidation and blood glucose oxidation suggested that the direct and/or indirect (via lactate) oxidation of muscle glycogen was greater in cycling than running.