Ingesting carbohydrate beverages during exercise in cool temperatures can improve endurance performance. However, because hyperosmotic solutions leave the stomach more slowly than water, carbohydrate beverages could be less effective in minimizing the dehydration and hyperthermia that accompany exercise in the heat. To determine the effect of osmotically different beverages on prolonged (2 h) treadmill exercise (65% VO2max) in the heat (T db = 35 degrees C), five male runners (age 24 to 41 yr) performed three separate runs drinking 200 ml every 20 min of either 10% glucose polymer (GP), 10% glucose (G), or saccharin-sweetened water (WH). A fourth run was performed in a cool (T db = 25 degrees C) room and included drinking saccharin-sweetened water (WC). Drink osmolalities (Osm) for runs GP, G, WH, and WC were 194, 586, 94, and 71 mmol . kg-1, respectively. No significant differences were observed between runs in the heat (GP, G, and WH) for heart rate, rectal and mean skin temperatures, sweat rate, percent change in plasma volume, and gastric residue volume. When compared to the WC run, both the GP and G runs yielded greater (P less than 0.05) declines in percent change in plasma volume, but only the G run had a greater (P less than 0.05) gastric residue volume. Neither plasma osmolality, total protein, nor [Na+] varied between runs. Plasma glucose, insulin, and respiratory exchange ratios were similar between the GP and G runs. However, the GP run yielded the lowest (P less than 0.05) plasma glycerol values. Although gastric residue volume (r = 0.68) and final percent change in plasma volume (r = 0.69) were significantly correlated with drink osmolality, thermoregulation was similar between runs in the heat despite the beverage consumed.