African-American women have a long-standing approximately 20% higher breast cancer incidence rate than USA White women under age 40 while rates among Latinas are lower than those of Whites. The reasons for this are not clear, however they may be due to ethnic differences in circulating oestradiol and progesterone levels. In a cross-sectional study, we investigated whether anovulation frequency and circulating serum oestradiol and/or progesterone levels vary among normally cycling nulliparous African-American (n=60), Latina (n=112) and non-Latina White (n=69) women. Blood and urine specimens were collected over two menstrual cycles among healthy 17-to 34-year-old women. Frequency of anovulation was greater among White women (nine out of 63, 14.3%) than African-American women (four out of 56, 7.1%) or Latina women (seven out of 102, 6.9%), although these differences were not statistically significant. African-American women had 9.9% (P=0.26) higher follicular phase oestradiol concentrations than Latina women and 17.4% (P=0.13) higher levels than White women. African-American women also had considerably higher levels of luteal phase oestradiol (vs Latinas, +9.4%, P=0.14; vs Whites, +25.3%, P=0.003) and progesterone (vs Latinas, +15.4%, P=0.07; vs Whites, +36.4%, P=0.002). Latina women were also observed to have higher follicular oestradiol, and luteal oestradiol and progesterone levels than White women (follicular oestradiol: +6.8%, P=0.48; luteal oestradiol: +14.6%, P=0.04; luteal progesterone: +18.2%, P=0.06). These results suggest that exposure to endogenous steroid hormones may be greater for young African-American and Latina women than for Whites. British Journal of Cancer (2002) (Gray et al, 1980;Parkin et al, 1997;SEER Program, 2001). The reasons for this are not clear, however they may be due to ethnic differences in circulating oestradiol (E 2 ) and progesterone (P 4 ) levels. Oestradiol is an established breast-cell mitogen, and considerable epidemiological and experimental evidence indicates that women with higher circulating levels are at greater risk of developing breast cancer (Bernstein and Ross, 1993; Pike et al, 1993;Thomas et al, 1997;Hankinson et al, 1998). Considerable evidence also exists indicating that progestins play an important role in breast carcinogenesis. Recent epidemiological studies have shown that postmenopausal hormone replacement therapy in the form of estrogen -progestin replacement therapy (EPRT) increases breast cancer risk to a much greater extent than estrogen (alone) replacement therapy (ERT; Magnusson et al, 1999;Ross et al, 2000;Schairer et al, 2000). These results are supported by the finding that mammographic densities are increased much more by EPRT than ERT (Greendale et al, 1999), and so is breast-cell proliferation (Hofseth et al, 1999).In the present study, we evaluated differences in ovarian function in African-American, Latina and non-Latina White young women in Los Angeles who were recruited as part of a crosssectional investigation of the frequency of anovulatio...