The incidence and height of antibody titers to colon, assayed by indirect hemagglutination with a heat stable colon extract from germ free rats, is significantly higher in sera from patients with ulcerative colitis than in those from healthy controls or from patients with amebic liver abscess or dysentery. While sera from ulcerative colitis patients and controls are indistinguishable in regard to incidence and height of antibody titers to Forsman antigen, Staphylococcus aureus S 209, Clostridium difficile, and several common strains of E. coli, they have elevated titers and increased incidence of antibodies to a heat stable antigen of E. coli O14. Patients with amebic dysentery have normal titers of such antibodies. Absorption of patients' sera with E. coli O14 antigen inhibits the colon directed hemagglutination reaction in approximately 30% of the cases tested. Likewise, the anti-E. coli O14 reaction can sometimes be inhibited with the colon extract. Other E. coli strains and other bacteria are inactive or have only weak inhibitory activity. Hemagglutination inhibition experiments show that germ free rat colon and E. coli O14 contain common structures, depicted by antibodies in the patients' sera. This pattern of reactivity closely resembles that seen in rats made autoimmune to colon by injection of newborn rabbit colon. E. coli O14 is known to carry a heterogenetic antigen present in lower concentration (or activity) in most Enterobacteriaceae. Hemagglutination inhibition experiments with rabbit antisera to E. coli O14 suggest that the antigen common for E. coli O14 and colon is related to this heterogenetic antigen. The findings imply that this antigen, which is constantly present in low concentrations in the human colon, may give rise to anticolon antibody formation in ulcerative colitis through breakage of tolerance. Since this antigen is present in healthy individuals as well, additional factors are required to explain the induction of anti-colon autoimmunity in ulcerative colitis.