The gastric emptying of six infant feedings (20 kcal/oz; whey to casein ratio, 60:40) with varying fat and carbohydrate composition was studied. Feedings contained either predominantly long-chain triglycerides (94%) or predominantly medium-chain triglycerides (94%) as the fat and lactose, glucose, or glucose polymers (Polycose) as the carbohydrate. Eleven premature infants were fed 22 ml/kg body wt of all six feedings over a 3-4-day period, and the volume of gastric contents was measured every 20 min using polyethylene glycol 4000 as the marker. Analysis of variance demonstrated that the use of medium-chain triglycerides resulted in faster gastric emptying than long-chain triglycerides (p less than 0.001). Analysis of variance and Tukey's test showed that use of glucose polymers instead of glucose resulted in less volume of gastric contents at 40 min (p less than 0.05). Use of glucose polymers instead of lactose resulted in less volume of gastric contents at 60 and 80 min (p less than 0.05). Gastric emptying can be altered by changes in nutrient composition. The difference between medium-chain and long-chain triglycerides was more pronounced than the differences between the carbohydrates studied. Feedings with medium-chain triglycerides may be more suitable than long-chain triglycerides in patients with delayed gastric emptying.