Serial radioautographic data obtained in two patients with Sezary syndrome following intravenous tritiated thymidine administration were analysed using multicompartment kinetic models. In both patients, the grain count halving time in the cutaneous Sezary cell compartment was too long to account for the grain count halving rate observed in the peripheral blood. This implies that some proliferating cell compartment other than the skin is primarily responsible for producing the Sezary cells that circulate in the peripheral blood. In both patients, the cell compartment that served as the source of circulating Sezary cells consisted of 5-6 x 10" cells (500-600 g of tumour) with an average cell cycle time of 3 4 days. By comparison, the cutaneous Sezary cell compartment was estimated to contain 2 . 2 4 6 x 1OI2 cells (2.2-4.6 kg of tumour) with an average cell cycle time of 7-70 days. While this study does not permit direct anatomic localization 'of the primary site of Sezary cell production, the properties of this compartment that can be deduced from the available data suggest the lymph nodes as likely candidates. Thus, the Sezary syndrome may well be a true lymph node malignancy with prominent cutaneous manifestations.