The incidence of thyroid carcinoma and of clinically palpable thyroid nodules is compared in three groups of individuals many years after exposure to ionizing radiation in childhood or infancy. The estimated mean cumulative doses to the thyroid gland ranged from 20 rad in the Ann Arbor series irradiated with x-rays as infants for thymic enlargement upward to 1225 rad of mixed radiation (including that from ingested radioiodine) in the Marshallese children. Intermediate between these extremes is the high-risk subgroup in the Rochester series of persons irradiated for thymic enlargement; they received an estimated mean thyroid dose of 335 rad. A plot of the incidence of thyroid nodularity against the total cumulative thyroid dose gives what could be a linear dose response with no threshold or, at least, a threshold below 20 rad. The risk of developing carcinoma in the three studies ranges from 0 to 5.5 cases and that for nodularity 38 to 52 cases per 106 persons exposed per rad (thyroid dose) per year. There is evidence dence from animal experimentation that the biologically effective dose in the case of the Marshallese would have been considerably lower than the cumulative physical dose; hence, the dose response might actually be curvilinear at least in the higher dose range. In view of the uncertainties regarding dose and the assumptions made in my study, the risk values may not be exact; however, these values reflect the trends regarding risks estimated from the best available data in man. Furthermore, the estimated risk values apply only if radiation exposure occurred in childhood.