Multiple lines of evidence support a central role of hormones in the etiology of breast cancer. In epidemiologic studies, considerable effort has focused on delineating the role of endogenous hormones in risk of breast cancer among postmenopausal women. Recently, substantial additional data has accrued from prospective studies where endogenous hormones are measured in study subjects prior to disease diagnosis. In this review, the epidemiologic evidence linking sex steroids--estrogens, testosterone, and progesterone, specifically--with subsequent risk of breast cancer in both premenopausal and postmenopausal women is summarized. Overall, a strong positive association between breast cancer risk and circulating levels of both estrogens and testosterone has now been well confirmed among postmenopausal women; women with hormone levels in the top 20% of the distribution (versus bottom 20%) have a two- to three-fold higher risk of breast cancer. Evidence among premenopausal women is more limited, though increased risk associated with higher levels of testosterone is consistent. However, both positive and null associations have been observed with estrogens and progesterone and clearly more evaluation is needed.