Abstract: The objective of this study is to evaluate the effect of estrogens on fish endocrine disruption in river water and treated wastewater. Endogenous estrogen estrone (E1), 17β-estradiol (E2), river water, and treated wastewater were used for exposure tests on male Japanese medaka. Vitellogenin induction in male medaka was regarded as the endpoint of endocrine disruption. The effective concentrations of E1 and E2 on vitellogenin induction in medaka were evaluated by breeding medaka for 14 days in tanks with various concentrations of E1 or E2. Vitellogenin induction increased with elapse of time during exposure, with higher estrogen concentrations causing greater vitellogenin induction. According to the test results, the lowest observed effect concentrations (LOECs) of E1 and E2 were estimated to be 31.6 and 5.0 ng/l, respectively; and the E2 equivalent (EEQ) LOEC value of E1 was 5.2 ng-E2/l, derived by multiplying the relative potency of E1 by that of E2 as estimated in this study. In continuous exposure tests using river water or treated wastewater, the estrogenic activity, an index of total estrogenic potential measured by yeast screen assay and expressed as EEQ, varied widely during the exposure tests, and significant vitellogenin induction was observed after several days of high levels of estrogenic activity in water. Vitellogenin concentration tended to significantly increase if EEQ exceeded the level of 5 ng-E2/l. The threshold value was substantially the same as the results for E1 and E2 exposure tests of medaka. Consequently, EEQ over 5 ng-E2/l was revealed to have the potential to cause endocrine disruption of male medaka. As estrogenic activity exceeding 5 ng-E2/l was observed in some rivers in Japan, total estrogenic potential needs to be further decreased at wastewater treatment plants to prevent fish endocrine disruption.