Epidemiological evidence indicates that estrogens are one of the risk factors of breast cancer. However, there have been reports that pre-pubertal estrogen exposure is related to reduced breast cancer risk. These discrepancies made us investigate the time-point and duration of estrogen exposure. Our studies focus on the effect of estradiol-3-benzoate (EB) on the mammary gland that was exposed to carcinogens. Ninety-six female Sprague-Dawley rats were randomly divided into 6 groups. Animals at 7 weeks of age were injected with 7,12-dimethylbenz[a]anthracene (DMBA) in groups 1, 2 and 3 or N-methyl-N-nitrosourea (MNU) in groups 4, 5 and 6. One week later, the animals were subjected to sustained treatment with 0 micro g (groups 1 and 4), 30 micro g (groups 2 and 5) or 300 micro g (groups 3 and 6) of EB containing pellets for 4 weeks. All animals were sacrificed at 5 weeks or 21 weeks after carcinogen treatment, for the examination of mammary gland differentiation or mammary gland tumors, respectively. At 21 weeks after carcinogen treatment, the incidence of mammary tumors in group 2 was significantly decreased (P<0.05). EB treatment decreased the multiplicity of DMBA- or MNU-induced mammary gland tumors. At 5 weeks after carcinogen treatment, there were increased branchings of the mammary gland, and there was also a decrease of ERalpha and ERbeta in EB treatment groups. Taken together with these results, we conclude that EB has an inhibitory effect on mammary carcinogenesis, and it suggests that this inhibition may be associated with the differentiation of mammary gland and modulation of ERalpha and ERbeta.