A silicone rubber serosal tunnel jejunostomy for postoperative nutritional therapy was placed in 38 patients who had major operations. A dilute, chemically defined diet was begun within 48 hours of surgery and an attempt made to advance it slowly to full strength by the fifth day. No catheter-related complications occurred. Intravenous fluid therapy was shortened to less than three days in 11 (29%) patients. Only 17 (45%) patients tolerated full-strength feedings within the protocol time. Three (8%) patients depended on their tube feedings for over 30 days and the need for parenteral nutrition was avoided. Of six septic patients, four had complete intolerance of the diet; two of these developed massive gastric hypersecretion. Serum albumin was a statistically significant indicator of whether a patient could tolerate tube feedings. No patient with an albumin less than 3 g/dl tolerated full-strength feedings and the pded. Of six septic patients, four had complete intolerance of the diet; two of these developed massive gastric hypersecretion. Serum albumin was a statistically significant indicator of whether a patient could tolerate tube feedings. No patient with an albumin less than 3 g/dl tolerated full-strength feedings and the patients with albumin greater than 4 g/dl had no problems. Postoperative feeding utilizing the serosal tunnel jejunostomy is technically safe, but feeding difficulties can be anticipated in those patients who are septic or severely malnourished; these are the patients whose nutritional needs are the greatest. The greatest benefits accrue to those patients who are in need of long-term nutritional support. We recommend routine placement of these catheters in major operations.