This study examined the relationship between the delivery of insulin into the peritoneal space and its absorption into the peripheral circulation. Studies were performed in conscious dogs receiving somatostatin (5.0 microgram/min) to suppress endogenous insulin secretion, and intravenous glucose (50 mg/min) to prevent hypoglycemia. The biologic effectiveness of the absorbed insulin was determined by its hypoglycemic effect. The possibility of direct absorption of insulin into the portal circulation from the peritoneal space in anesthetized, portal vein-catheterized dogs was examined with radiolabeled I125 insulin. Our results suggest that absorption of insulin from the peritoneal space is volume, concentration, and time-dependent. Maximal absorption of insulin was observed at 50 min when 1.92 U of insulin in a volume of 3 ml was infused intraperitoneally over 30 min. More rapid absorption was observed at 30 min when this quantity of insulin was given in a 1-min intraperitoneal bolus, compared to 30 min of intraperitoneal infusion. Least rapid absorption of insulin followed the delivery of 1.92 U of insulin in a volume of 15 ml. Intermediate absorption of insulin was observed at 40 min when the 1.92 U was delivered in a volume of 0.6 ml. Peripheral intravenous insulin delivery of 1.92 U reached a maximal plasma concentration at 20 min, which was more than three times the concentration observed with intraperitoneal insulin. Isotopic tracer studies, in which radioiodinated insulin was placed into the peritoneal space in anesthetized dogs, demonstrated greater radioactivity in the portal vein than in the aorta throughout a 30-min observation period. These studies demonstrate that intraperitoneal administration of insulin results in absorption of insulin which is volume, concentration, and time-dependent. Thus, the peritoneal space may be an appropriate site for insulin delivery through a transcutaneous catheter.