T o define the pathophysiologic mechanisms of cold agglutinin disease, we investigated a human model of this syndrome in normal volunteers and in patients with diminished levels of serum complement. Subjects received intravenous injections of autologous, chromated (51Cr) erythrocytes which had been exposed in vitro to purified cold agglutinin preparations and to fresh autologous serum (as a source of complement). In vitro tests confirmed that such cells were coated with activated complement components (C3b), but not with immunoglobulin. Studies of erythrocyte clearance and simultaneous organ scanning showed that erythrocytes sensitized with low levels of cold agglutinin primarily undergo reticuloendothelial sequestration by the liver rather than intravascular hemolysis. After the initial sequestration of C3b-coated erythrocytes, a fraction of the cells are released back into the circulation and survive normally thereafter. Both phenomena are dose dependent and closely follow the sequestration and release pattern observed with IgM isoagglutinin sensitization. Experiments that used heated autologous serum as a source of B3 inactivator demonstrated that functionally intact C3b is required for hepatic sequestration. Erythrocytes coated with C3d were not cleared from the circulation. In vitro assays that sued human macrophage monolayers suggested that the intrahepatic conversion of C3b to C3d is responsible for the release of sensitized erythrocytes back into the circulation. The clearance of cold agglutinin-sensititzed erythrocytes was compared to the clearance mediated by IgM isoagglutinin. We found that the rate of complement fixation by an IgM antibody proceeds rapidly in vivo that the time for complement activation is not a factor in limiting the rate of hepatic sequestration. The major limiting factor appears to be the rate of liver blood flow. Maximal in vitro coating of erythrocytes with C3d conferred protection from further cold agglutinin sensitization but not from IgM isoagglutinin-mediated clearance. This suggests a mechanism for the resistance to lysis observed in cells obtained from patients with the cold agglutinin syndrome and confirms the marked dependence of the site of C3 attachment on the site of membrane localization of the sensitizing antibody.