The hormonal activity of natural estrogens is influenced by the degree to which they bind to serum proteins. In the pregnant female and in the fetus, greater than 99% of estradiol may be bound by serum binding proteins. Therefore, even though total serum levels of estradiol appear very high in fetuses, we have found that in rodent fetuses, there is a very low free concentration of estradiol (0.2 pg/ml). Naturally occurring variation in fetal serum estradiol predicts differences in numerous postnatal traits, including prostate size. In addition, when this low level of free estradiol was experimentally increased from 0.2 to 0.3 pg/ml during the last third of fetal life, treated male mice showed an increase in adult prostate weight. Fetal exposure to low doses of xenobiotic estrogens by feeding to pregnant females, including the compounds methoxychlor (20 and 2000 micrograms/kg body weight), DES (0.02 to 2 micrograms/kg body weight) and bisphenol A (2 and 20 micrograms/kg body weight), also led to increased prostate weight in adulthood. In contrast, fetal doses of natural estradiol and DES above the physiological range of estrogenic activity, and within a toxicological dose range, led to the opposite outcome, a reduction in subsequent adult prostate weight. This indicates that it may be impossible to assess endocrine-disrupting activities in response to low doses within a physiological range of activity by using high, toxic doses of xenoestrogens in testing procedures. We have developed approaches in vitro to predict the potential estrogenic bioactivity of compounds in the physiologically relevant range in animals and humans. We address the following factors in predicting the final observed endocrine-disrupting effect in the animal: (1) the intrinsic estrogenic activity of a given molecule, (2) the effective free concentration determined by how the molecule is carried in serum, (3) partitioning between aqueous and lipid compartments in body and cell lipids, and (4) absorption and metabolism relative to the route of exposure. The studies and strategies we describe are important in developing criteria for a tiered testing system for the detection of estrogenic chemicals as well as endocrine-disrupting chemicals with different modes of action.