The search for antibodies that will reliably diagnose, predict and monitor the autoimmune thyroid diseases is a quest that is beset with difficulties. This review will describe the various antigens involved in autoimmune thyroid disease, the development of a humoral response to these antigens and some of the technical difficulties involved in trying to detect and then measure their concentrations. Multiple antigen configurations of thyroglobulin (TG) are produced when it is iodinated resulting in functionally active but immunologically distinct molecules. Thyroid peroxidase (TPO), originally described as thyroid microsomal antigen, is present on the apical surface of thyroid follicular cells and is the antigen most closely involved in cell-mediated cytotoxicity. Multiple B cell reactive epitopes exist each giving rise to different antibodies. The third antigen is the TSH receptor (TSHR) which is a two subunit glycoprotein. The extracellular A subunit is recognised by thyroid stimulating antibodies whilst those antibodies recognising the B subunit, located much nearer the cell surface, appear to function as blocking antibodies. The aetiology and mechanics of the autoimmune cellular and antibody responses involves a combination of HLA linkage, genetics and environmental factors to determine the initial and subsequent stages of the development of autoimmune thyroid disease. Starting off with immunofluorescence or passive tanned red cell agglutination assays, we now favour more sensitive TPO antibody quantitative immunoassays which are standardised using MRC 66/387. The apparent incidence of these antibodies is influenced by the techniques used to detect them and the significance of those positive antibody concentrations detected using assays that report very high incidence rates in the "normal" population remains to be proven using longitudinal studies. Older immunofluorescence and passive tanned red cell agglutination methods have been improved upon to give the current, competitive and non-competitive immunoassays that are much more sensitive and specific for anti-TG antibodies. However, because a wide-range of methods are still being used in clinical laboratories, the sensitivity and specificity of available methods will vary depending on the method used. There are still assays that are calibrated with purified or crude preparations of TG antibody by pooling patient sera or blood donor material. These various secondary standards are often, but not always, calibrated against the primary standard (MRC 65/93). All requests for TG estimation in thyroid carcinoma patients should have TG antibody estimated at the same time because of the possibility of interference in the tumour marker assay by the antibody. Two methods are widely used for the estimation of TSHR antibodies. The first involves bioassays based on cultured cells to measure the stimulating class of antibodies and the second involves receptor assays based on (125)I-labeled TSH. The most widely used assay uses detergent-solubilized porcine TSHR with TSHR antibodies to inhibit the TSHR-(125)I-TSH interaction, and polyethylene glycol to separate receptor-bound and free labelled-TSH by precipitation. TSHR antibody heterogeneity can coexist within an individual patient and change over time is one reason why it has been difficult to develop diagnostically accurate TSHR antibody tests.