Hepatic clearance of a small dose of a medium chain fatty acid was used to estimate portasystemic shunting in patients with cirrhosis. Sodium octanoate dissolved in isotonic saline was administered by rectum (11.4 mg per kg of lean body mass) in three groups of 10 patients with cirrhosis and in 10 noncirrhotic controls. Two groups of patients were classified as mild and severe cirrhotics on the basis of a clinical evaluation. The third group of cirrhotic patients had a surgical portacaval shunt. Serum octanoate levels were measured in each subject for a 1-hr period following the administration of sodium octanoate by rectum. A highly significant difference (P less than 0.001) was demonstrated between all four groups of subjects with highest levels occurring in the shunted cirrhotics and the lowest levels in the control group. Serum levels obtained in 2 patients before and after a portacaval shunt increased by a factor of 2 following the operation. Two patients with proven occlusion of a surgical shunt fell well within the range of the nonoperated cirrhotic patients. The results suggest that the rectal octanoate tolerance test can probably be used for evaluation of collateral circulation in cirrhotic patients and for testing the patency of surgical shunts.