The prevailing hypothesis of thyroid carcinogenesis is that thyroid-stimulating hormone (TSH) is a prime factor in the growth and possibly the genesis of thyroid carcinoma. On the other hand, Graves' disease is a hyperthyroid state caused presumably by human thyroid-stimulating immunoglobulin (HTSI), not TSH. However, the literature does contain scattered reports of an association between Graves' disease and thyroid carcinoma. With this background, we retrospectively reviewed our series of 117 consecutive patients operated on for Graves' disease from 1961 through 1984. The series contains 23 male and 94 female patients, with a mean age of 26.3 years at operation (median 26 years; range 8 to 58 years). Of these patients, six (two males and four females) were found to have carcinoma, four papillary and two follicular, for an incidence of 5.1%. Four of the 117 patients had a history of head and neck irradiation; two (50%) subsequently developed carcinoma. For the six patients with carcinoma, mean follow-up has been 8.1 years (median 8 years; range 1 to 14 years). To date there have been no recurrences of carcinoma, and all six are alive and well. Thus our series shows a carcinoma rate of 5.1% in patients with Graves' disease, which is markedly higher than the 0.1% to 0.2% incidence in random autopsy series of all patients. This raises the question of a possible carcinogenic role for HTSI, similar to that hypothesized for TSH, in patients with presumably suppressed TSH.