The frequency of tumors and other conditions of the thyroid gland were examined in 686 female radium dial workers first employed before 1930, who had a radium body-burden measurement while living (1958-76). If one assumed that the two thyroid cancers ascertained were radiation-induced and that a linear dose-response relationship existed, the estimated thyroid cancer risk was 69 (4-124, 95% confidence range) per 10(6) person-rem thyroid dose equivalent from internal and external radiation. Using data from the Connecticut tumor registry to obtain expected numbers of thyroid cancer, the estimated risk (2 observed vs. 0.67 expected cases) was 46 (95% confidence interval = -19 to 101) excess cases per 10(6) person-rem. Risk estimates were based on crude estimates of external radiation exposure and uncertain quality factors for internal radiation from alpha particles ingested. The frequencies of benign tumor (adenoma), nodules, and goiters were not significantly higher in the higher thyroid-dose groups (5-19, greater than or equal to 20 or greater than or equal to 50 rem) than in the lowest dose group (less than 5 rem). In 1237 female dial workers first employed before 1930, with or without a radium body-burden measurement, no deaths due to thyroid cancer (underlying cause of death on death certificates) were observed during 1950-76, when 0.4 deaths were expected. In a subgroup of 84 Illinois female dial workers who were long-term survivors, means for thyroid function test (T3 resin uptake and free thyroxine index) results did not differ among the thyroid-dose groups.