In order to improve the sensitivity of our previously established thyroid carcinogenesis model and to clarify whether endocrine disrupting chemicals with weak estrogenic activity have any modifying effects on the development of thyroid proliferative lesions, 6-week-old female castrated F344 rats were first given a single subcutaneous injection of 2000 mg/kg body weight of N-bis(2-hydroxypropyl)nitrosamine. From 1 week later, they received diets with: no supplement (basal diet (BD) group); cholesterol pellets containing 0.5 mg 17 beta-estradiol 3-benzoate (EB); or diet admixed with 1000 ppm methoxychlor (MXC) or 10,000 ppm bisphenol A (BPA) for 20 weeks. Furthermore, additional groups were administered 200 ppm sulfadimethoxine (SDM) in the drinking water simultaneously with the BD, EB, MXC or BPA treatments. Thyroid follicular cell hyperplasias, adenomas and/or carcinomas were induced only in the EB+SDM group, the incidences of non-malignant lesions being significantly increased, as compared with the BD+SDM group values. Furthermore, the serum level of thyroid stimulating hormone (TSH) was significantly increased in this group. No significant variation in quantitative values for thyroid proliferative lesions or TSH levels were observed in the other treated groups. The results of the present study convincingly indicate that EB, with strong estrogenic activity, but not MXC and BPA, with weak estrogenic activities, exerts promoting effects on thyroid carcinogenesis in rats. The present modified rat two-stage thyroid carcinogenesis model appears to have advantages over our previous model for screening purposes.