Six subjects exercised for 120;min on a cycle ergometer (65 (se 3) % V˙O2max) when ingesting a placebo or glucose, fructose or galactose (100g in 1000 ml water) labelled with 13C. The oxidation of energy substrates including exogenous hexoses was compared using indirect respiratory calorimetry and 13CO2 production at the mouth. Total carbohydrate progressively decreased and total fat oxidation increased over the 120min exercise period in the four experimental situations. During the 120min of exercise, the amount of fructose oxidized (38·8 (se 2·6) g; 9·0 (se 0·6)% energy yield) was not significantly (approximately 4%) lower than that of exogenous glucose (40·5 (se 3·4) g; 9·2 (se 0·8)% energy yield), while that of galactose (23·7 (se 3·5) g; 5·5 (se 0·9) % energy yield) was only 59% and 61% that of glucose and fructose, respectively. When compared with the placebo, the ingestion and oxidation of the three hexoses did not significantly modify fat oxidation or total carbohydrate oxidation, but it significantly reduced (9–13%) endogenous carbohydrate oxidation. The present data indicate that fructose and exogenous glucose ingested during exercise could be oxidized at a similar rate, but that the oxidation rate of galactose was only approximately 60% that of the exogenous glucose and fructose, presumably because of a preferential incorporation of galactose into liver glycogen (Leloir pathway). The reduction in endogenous carbohydrate oxidation was, however, similar with the three hexoses.