Several lines of evidence attest to the existence of alternative ligand binding sites on the oestrogen receptor (ER), including non-competitive inhibition by trilostane or tamoxifen. It is possible that the inhibitory action of conventional oestrogen agonists at high concentrations may indicate that they too interact at alternative ER sites, albeit at low affinity. To test this possibility an oestrogen reporter assay was used to compare the activity of different oestrogens and antagonists in breast cancer and prostate cell lines. All four cell lines tested contained different amounts of oestrogen receptor α (ERα), ERβ, progesterone receptor and coregulator mRNA. Though differences were observed in response to stimulation and inhibition, these correlated only with the presence or absence of ERα, and not with the other components. Thus stimulation of the reporter by oestradiol and oestrone was biphasic in the breast cancer cells, while prostate cells were unable to respond. Only T47D cells were stimulated by oestriol or diethylstilboestrol, however reporter activity of all the cell lines was repressed by 10mM diethylstilboestrol. Reporter activity of MCF-7 cells was inhibited by tamoxifen, raloxifene and ICI 182,780, but stimulated by trilostane, yet all these antioestrogens inhibited agonist-stimulated activity. Trilostane also inhibited the agonism seen in cells co-treated with E2 and tamoxifen. It is clear that several of the compounds tested may have either agonist or antagonist effects under different conditions and at different concentrations, acting through ERα alone. Though biphasic dose response curves, or hormesis, have been attributed to various mechanisms, we here provide evidence that alternative ligand binding sites may contribute to this phenomenon.