Titres of E. coli O6 antibody obtained with the direct bacterial agglutination (DBA) technique, the indirect haemagglutination (IHA) technique and the ammonium sulphate precipitation (ASP) technique were compared using rabbit immune sera and IgG and IgM antibody fractions thereof, untreated as well as reduced with β-mercaptoethanol (ME). Poor rank correlation was obtained between the different assays on sera whether reduced or not. However, statistically significant correlation was obtained for the DBA and IHA titres on the IgM as well as IgG antibody fractions. There was also good correlation between the IHA and ASP titres on IgG antibody fractions. The DBA and IHA techniques both favoured IgM antibodies compared to the ASP method. This effect, however, was most accentuated by the IHA technique. In contrast, the IHA technique underestimated IgG antibodies more than the DBA technique. This could not be explained by influence of different antibody avidities, although the methods previously have been shown variously affected by such antibody properties. The study thus indicates an explanation for discrepancies in E. coli·antibody titres in patients with urinary tract infections obtained by different investigators using direct and indirect agglutination techniques or radioimmunoassay.