Differences in hormonal and constitutional parameters between women with at least one first and/or second degree relative with breast cancer (RBC) and women without such affected relatives were studied in a group of healthy, nulligravid women aged 19-25 years. Present oral contraceptive (OC) users were analysed separately. In women not presently exposed to OCs we found significant correlations between RBC and reduced concentrations of testosterone during both the follicular (P < 0.001) and luteal menstrual cycle phases (P = 0.016). 17 beta-oestradiol was also significantly negatively correlated with RBC in the follicular (P = 0.044) and in the luteal phase (P = 0.027). RBC was significantly correlated with a lower waist/hip ratio (P = 0.044) compared with women without such a history. In multivariate analyses, the results for testosterone but not 17 beta-oestradiol remained significant. In these analyses high IGF-1 (P = 0.05) in the follicular phase and low sexual hormone-binding globulin (SHBG) (P = 0.04) in the luteal phase were also related to RBC. Including all 66 women in a multivariate model that analysed the specific effects from OCs and RBC on plasma testosterone showed that plasma testosterone was significantly lower among present OC users (P = 0.004) and in women with RBC (P = 0.005) during cycle days 5-10, with a significant positive two-way interaction between present OC use and RBC (P = 0.007). During cycle days 18-23 plasma testosterone showed a significant negative relationship with present OC use (P < 0.001) and RBC (P = 0.016) no significant interaction was seen during cycle days 18-23. Factors not significantly related to RBC were height, weight, breast size, age at menarche, p-progesterone and p-prolactin. It is concluded that a family history of breast cancer significantly lowered plasma testosterone concentrations in both cycle phases among healthy, nulligravid women compared with women without such history.