Tanned red cell hemagglutination has been widely accepted for use in the detection of anti-thyroglobulin antibodies. The interpretation, however, especially when the hemagglutination was observed at a low titer of serum dilution, has remained equivocal. In order to separate specific and non-specific hemagglutination, we tried to concentrate serum IgG to compare the hemagglutination titer at various concentrations of IgG. If the hemagglutination is specific for the presence of antithyroglobulin antibodies, the hemagglutination titer would increase as the IgG concentration rises. On the contrary, when the hemagglutination is non-specific and irrelevant to the presence of anti-thyroglobulin antibodies, the hemagglutination titer is expected not to increase while the IgG concentration rises. Based on this hypothesis, sera which showed various hemagglutination titer for anti-thyroglobulin antibodies were studied to examine the parallelism between hemagglutination titer and IgG concentration. IgG was concentrated with the method of ammonium sulfate precipitation. The grade of concentrate was examined by the IgG determination with the method of immunodiffusion technique employing the kit distributed by Behringwerke. The hemagglutination test was performed with the Boyden's method using the kit distributed by the Wellcome Company. The results showed that 4 out of 15 sera whose original hemagglutination titer was 1:3 failed to show increase in hemagglutination titer even at the 2 to 4 fold concentration of IgG. On the contrary, only 2 out of 26 sera whose hemagglutination titer was more than 1:6, failed to increase the titer at the IgG concentrate to the same degree. Thus, it is possible to conclude that the hemagglutination observed at 1:3 dilution of the serum is highly non-specific, while that observed at a serum dilution of more than 1:6 is thought to be specific.