We have tested sera from patients with autoimmune thyroid disorders with or without ophthalmopathy for immunoreactivity, in a dot blot assay, against a recombinant 98 amino acid fragment of a cloned 64 kDa protein, D1, which is expressed in human eye muscle and thyroid, in the form of a Lac Z fusion protein. Tests were positive in 19 out of 40 patients with established thyroid-associated ophthalmopathy (TAO), in 12 out of 21 patients with Graves' hyperthyroidism (GH) without clinically evident ophthalmopathy, in 5 out of 10 patients with thyroid autoimmunity and lid retraction but no other signs of ophthalmopathy, in 4 out of 23 patients with Hashimoto's thyroiditis (HT) without evident ophthalmopathy and in 2 out of 18 patients with benign adenoma or multinodular goitre, but in only 2 out of 37 normal subjects tested. SDS-polyacrylamide gel electrophoresis and Western blotting for an antibody reactive with a 64 kDa antigen in pig eye muscle membranes was also carried out on sera from patients with TAO and GH. While immunoblotting for antibodies reactive with a 64 kDa protein was more often positive in patients with TAO, in whom 58% had serum antibodies which reacted with a 64 kDa protein, this was not the case in patients with GH without eye signs in whom the prevalence of positive immunoblot tests was 35%. Overall there was a fairly close correlation between the two tests although there were many exceptions. While the apparent discrepancy between antibodies recognizing the human recombinant 98 amino acid peptide in dot blot and those binding to a 64 kDa protein in pig eye muscle membranes in immunoblot could be explained by the existence of more than one epitope on the native protein, it is possible that the peptide D1and the 64 kDa molecule are different. Although the relationship of these antibodies to the ophthalmopathy is uncertain, the high prevalences of antibodies to the D1fusion protein in patients with GH without evident ophthalmopathy (but who may have sub-clinical disease) and in patients with thyroid autoimmunity with lid retraction, which many believe to be an early stage of ophthalmopathy, raises the possibility that these antibodies may be an early marker of the eye disease in patients with thyroid autoimmunity. Further molecular studies will show whether D1and the 64 kDa protein are the same or different molecules while prospective clinical studies of these patient groups will be necessary to determine the significance of the corresponding antibodies.