Carbohydrate (CHO) ingestion during short-term recovery from prolonged running has been shown to increase the capacity for subsequent exercise in a warm environment. The aim of this study was to examine the effects of the amount of glucose given during recovery on substrate storage and utilisation during recovery and subsequent exercise in a warm environment. A group of 11 healthy male volunteers took part in two experiments in a controlled warm environment (35degreesC, 40% relative humidity), 1 week apart. On each occasion the subjects completed two treadmill runs (T1 and T2) at a speed equivalent to 60% of maximal oxygen uptake, for 90 min, until they were fatigued, or until aural temperature (T-aur) reached 39degreesC. The two runs were separated by a 4 h recovery period (REC), during which subjects consumed 55 g of naturally enriched [U-C-13]-glucose in the form of a 7.5% carbohydrate-electrolyte solution (CES, mass of solution 667 g) immediately after T1. The subjects then consumed either: the same quantity of CES, or an equivalent volume of an electrolyte placebo, at 60, 120 and 180 min during REC, providing a total of 220 g (C220) or 55 g (C55) of [U-13C]-glucose, respectively. Expired gases were collected at 15 min intervals during exercise and 60 min intervals during REC, for determination of total CHO and fat oxidation by indirect respiratory calorimetry, and orally ingested [U-13C]-glucose oxidation, estimated from the C-13:C-12 ratio of expired CO2. Substrate metabolism did not differ between conditions during T1. Despite the fact that total CHO (P < 0.05) and ingested glucose oxidation (P < 0.01) were greater during REC of the C220 condition, glycogen synthesis was estimated to be approximately fivefold greater (P < 0.01) than in the C55 condition. During T2 the rate of total CHO oxidation was higher (P < 0.01) and total fat oxidation lower (P < 0.01) at all times during the C220 compared to the C55 condition. The greater CHO oxidation during C220 appeared to be met from ingested sources, as the rate of [U-13C]-glucose oxidation was greater (P < 0.01) at all times during T2, compared to C55. Whilst more of the ingested substrate remained unoxidised on completion of T2 during C220, exercise duration was similar in the two experimental conditions, and was limited by thermoregulatory incapacity (T-aur > 39degreesC) rather than substrate availability per se.