The measurement of hepatic blood flow based upon the rate of disappearance of colloidal particles from peripheral blood by the Kupffer cells depends upon two basic assumptions: 1) that the colloid is completely removed from the blood in one passage through the liver, and 2) that the liver is its sole source of removal. Dobson and Jones (1) showed in animals that radioactive colloidal chromic phosphate approximated satisfactorily to these theoretical postulates, although the efficiency of hepatic extraction of this colloid has not been determined in man. Radioactive colloidal gold has also been used as an index of liver blood flow, and in the original work of Vetter, Falkner and Neumayr (2) the hepatic extraction efficiency averaged 80 per cent in three subjects. In a later paper (3), however, Vetter and co-workers reported a lower average hepatic blood flow in normal subjects, and subsequently it has been shown that commercial preparations of radiogold are not suitable for measurement of hepatic blood flow, since their particle sizes are not homogeneous (4, 5).Halpern and colleagues (6) described a heatdenatured albumin colloidal complex labeled with I131 (CAI131). The particle size could be controlled by the extent of heating and checked by biological standardization. Preliminary results in animals and in man suggested that extrahepatic removal was minimal and hepatic extraction was over 90 per cent (7-9). This colloid might therefore be suitable for measuring liver blood flow and our communication deals with its use for this purpose in control subjects and patients with cirrhosis. The results suggest that the method is valid for