Many chemicals in surface waters and sediments have recently been discovered to have estrogenic/antiestrogenic activity. Among these compounds, known as 'endocrine disrupters', are natural and synthetic hormones, phytoestrogenes and a variety of industrial chemicals, such as certain detergents and pesticides. These substances are supposed to affect the development and reproduction in wildlife and humans and may also be involved in the induction of cancer. In order to assess the estrogenic/antiestrogenic potential of pure compounds and complex environmental samples we compared an array of in vitro test systems, (i) two luciferase reporter gene assays using transgenic human MVLN cells (derived from MCF-7 cells) and HGELN cells (derived from HeLa cells); (ii) a competitive binding assay with recombinant human estrogen receptors (ER) alpha and beta; and (iii) a proliferation assay with MCF7-cells (E-Screen). The sensitivity of the assays for 17-beta-estradiol decreased in the order: MVLN-cells=E-Screen>HGELN-cells>binding to ER-alpha>binding to ER-beta. A good correlation was obtained between the estrogenic potencies of 11 compounds (17-beta-estradiol (E(2)), estrone (E(1)), estriol (E(3)), ethinylestradiol (EE(2)), diethylstilbestrol (DES), coumestrol, beta-sitosterol, genistein, 4-nonylphenol, 4-octylphenol, bisphenol A) in the three tissue culture assays. The relative potencies of the compounds obtained by the cell free binding assays were one to two orders of magnitude higher compared with the cell culture assays. The phytoestrogens showed a preference to bind to ER-beta, but only genistein showed a much lower activity in the E-Screen (growth induction in breast cancer cells) compared with the luciferase induction in MVLN and HGELN-cells.