This study was undertaken to compare the efficacy of four different types of perioperative intravenous nutritional support. Fifty-five patients undergoing routine major surgery were studied. They were prospectively assigned to one of four study groups. Group 1 received formal total parenteral nutrition (90 g amino acids, 3,000 calories as glucose, per day); Group 2, 100 g glucose per day; Group 3, 90 g amino acids per day; and Group 4, peripheral parenteral nutrition (90 g amino acids plus 1,600 calories, 60 percent as fat per day). Group 1 was maintained on therapy for 3 weeks and the other groups for 8 days. Nitrogen balance, maintenance of body cell mass, serum albumin levels, and maintenance of exercise capacity were measured. Patients receiving peripheral parenteral nutrition maintained their nutritional parameters, as did those receiving total parenteral nutrition. These infusions were both markedly superior to those receiving glucose alone or those receiving amino acids alone. Nitrogen balance was not correlated with maintenance of function, but maintenance of body cell mass was correlated with maintenance of exercise capacity (r = 0.66, p less than or equal to 0.01). We conclude that perioperative peripheral parenteral nutrition, in contradistinction to hypocaloric infusions of glucose or amino acids, is capable of maintaining postoperative muscle mass and function close to preoperative levels after major surgery, and in situations of relatively mild surgical stress, approaches the efficacy of total parenteral nutrition in this regard. A significant correlation exists between changes in body cell mass determined from isotope dilution and changes in the exercise capacity of large muscle masses.