Carbohydrate containing drinks are commonly consumed as an ergogenic aid during endurance sports activities. The efficacy of a given drink is limited by the rate of absorption, which is in turn limited by gastric emptying. A myriad of factors influence gastric emptying. The influence of several of these factors (training status, exercise intensity, and carbohydrate composition) was investigated by repeated experiments using a nasogastric tube and a modification of the technique of George. A group of well-trained bicyclists and a group of untrained subjects performed similar experiments. Three different carbohydrate containing drinks (15 g.100 ml-1 glucose (G), 15 g.100 ml-1 maltodextrin plus 3 g.100 ml-1 fructose (MD), 7 g.100 ml-1 sucrose (I] and artificially sweetened water were compared during rest and 50 and 70% Wmax bicycling. Experimental design was crossover. There was a trend for the carbohydrate drinks to empty initially more slowly under the influence of exercise. Differences in drink volume remaining in the stomach were significant (P less than 0.05), with I at 10 min (70%, mean = 48.9%; rest, mean = 30.5%) and at 20 min (70%, mean = 28.9%; rest, mean = 23.8%) and with MD at 10 min (70%, mean = 71.1%; rest, mean = 55.9%). A similar trend was also seen with 50% Wmax exercise; however, this trend was only significant with MD at 10 min (50%, mean = 72.1%; rest, mean = 55.9%). Drink composition was a much stronger inhibitor of gastric emptying. However, all drinks emptied exponentially with fast-phase initial emptying rates. No differences in gastric emptying or secretion were observed between trained and untrained subjects.