In a significant proportion of patients with thyroid disorders, thyroglobulin antibodies (TgAb) invalidate double-antibody radioimmunoassays (RIA) for thyroglobulin (Tg). A modified RIA procedure for determining serum Tg in the presence of TgAb, is described. Two measurements are made on each sample. a Ratio of immunoglobulin (Ig)-bound thyroglobulin to free thyroglobulin (bound:free ratio). This is measured by the distribution of a trace quantity of 125I-Tg between bound and free fractions. b Concentration of free, i.e. unbound thyroglobulin. This is measured in the supernatant after precipitation of IgG-bound Tg. Then, bound Tg=B:F x (free Tg) and total Tg=(bound Tg)+(free Tg). Critical factors examined in the validation of the modified method were: (1) use of immunologically intact 125I-Tg (specific activity less than or equal to 5 mCi/mg); (2) equilibration of tracer Tg with endogenous serum Tg, and (3) clean separation of IgG-bound and free fractions of Tg with goat anti-human IgG. Recovery of added Tg in the presence of TgAb was 105% and interassay precision of total Tg assay was less than or equal to 9.3%. In euthyroid subjects without TgAb, the correlation between total Tg levels by the modified and standard double-antibody method was 0.931. The correlation was much poorer in the presence of TgAb (r=0.752) reflecting their interference in the standard method. Reference ranges for serum total Tg levels by the modified method in euthyroid, hyper- and hypothyroid subjects with and without TgAb are detailed.