It was recently reported that the red wine phytoestrogen resveratrol (RES) acts as a superagonist to oestrogen-responsive MCF-7 cells. This activity of RES was speculated to be relevant to the 'French paradox' in which moderate red wine consumption is reported to yield cardiovascular health benefits to humans. We report here that RES binds to oestrogen receptors (ER) isolated from rat uterus with an affinity approximately 5 orders of magnitude lower than does either the reference synthetic oestrogen diethylstilboestrol (DES) or oestradiol (E2). In comparison with E2 or DES, RES is only a weak and partial agonist in a yeast hER-alpha transcription assay and in cos-1 cell assays employing transient transfections of ER-alpha or ER-beta associated with two different ER-response elements. Resveratrol was also concluded to be inactive in immature rat uterotrophic assays conducted using three daily administrations of 0.03-120 mgkg(-1)/day(-1) RES (administered by either oral gavage or subcutaneous injection). These data weaken the suggestion that the oestrogenicity of RES may account for the reported cardiovascular protective effects of red wine consumption, and they raise questions regarding the extent to which oestrogenicity data derived for a chemical using MCF-7 cells (or any other single in vitro assay) can be used to predict the hormonal effects likely to occur in animals or humans.