The occurrence of rebound hypoglycaemia may depend on the timing of carbohydrate intake. The aim of the present study was to investigate the metabolic and performance responses to the ingestion of carbohydrate at differing times before exercise. Eight subjects [mean (SEM)] [28 (3) years, 74.5 (2.6) kg, maximal oxygen uptake 63.1 (3.1) ml·kg–1·min–1] performed three experiments. They ingested 75 g glucose dissolved in 500 ml water, thereafter resting for either 15, 45 or 75 min (15-Pre, 45-Pre and 75-Pre) before exercising for 20 min at 65% maximal power output followed by a time trial [total work 685 (18) kJ]. There were no differences in performance between conditions [mean powers 268 (10), 269 (7) and 276 (12) W for 15-Pre, 45-Pre and 75-Pre, respectively]. There were significant differences in plasma glucose concentration between 15-Pre [6.6 (0.6) mmol·l–1; P<0.05] and both 45-Pre [4.5 (0.2) mmol·l–1] and 75-Pre [3.7 (0.2) mmol·l–1] immediately before exercise. Insulin concentrations immediately before exercise were higher (P<0.05) during 15-Pre [72.6 (10.4) µU·ml–1] than during 45-Pre [50.8 (9.9) µU·ml–1], which was higher (P<0.05) than during75-Pre [33.9 (5.5) µU·ml–1]. These differences disappeared within 10 min of exercise. Two subjects became hypoglycaemic (plasma glucose concentration of less than 3.5 mmol·l–1) in the 15-Pre while three and five subjects were transiently hypoglycaemic in the 45-Pre and 75-Pre, respectively. Performance and rating of perceived exertion did not seem to be related to hypoglycaemia. Altering the timing of the ingestion of carbohydrate before exercise resulted in differences in plasma glucose /insulin responses which disappeared within 10 min of exercise and which had no effect on performance. Hypoglycaemia was observed in some subjects during the first 10 min but this did not affect performance.