The objective of this study was to investigate the relation between violence victimization and levels of ovarian hormones during women's late reproductive years, as measured by serum levels of follicle-stimulating hormone and estradiol, which respectively rise and fall with onset of menopause. In March 1999, 732 women 36-45 years of age from the Harvard Study of Moods and Cycles cohort completed a survey of lifetime experiences of physical and sexual harm. Follicle-stimulating hormone and estradiol levels were measured during the menstrual period after entry into the cohort. Associations for violence and follicle-stimulating hormone and estradiol levels were estimated using crude and adjusted risk differences. Overall, women who experienced abuse during childhood or adolescence relative to never-abused women had a slight positive association of violence with high follicle-stimulating hormone. However, a positive association with high follicle-stimulating hormone was not observed among women whose first abuse occurred during adulthood. Age stratification indicated modification of the association between violence and low estradiol. Women 36-40 years of age had no evidence of a positive association between violence before adulthood and low estradiol, whereas first violence in adulthood was associated with an 11% increase in the estimate of risk difference [95% confidence limits (CL) = -0.14, 0.36]. Among women 41-45 years, there was a 17-23% increase in the estimate of risk difference for low estradiol, regardless of life stage at first experience of abuse (before adulthood, 95% CL = 0.06, 0.28; during adulthood, 95% CL = 0.01, 0.46). This investigation supports the credibility of a hypothesis that physical and sexual abuse may lead to neuroendocrine disruption, thereby affecting ovarian function and potentially leading to altered age at perimenopausal transition.