This study was conducted to determine whether rigorous exercise training adversely affects ovarian hormone levels and bone health in cyclically menstruating trained runners. Ovarian hormones, bone mineral density (BMD), body composition, 3-d diet records, 3-d estimated energy expenditure, and menstrual histories were evaluated in 10 trained collegiate runners and 10 moderately active controls. The trained runners had lower total body calcium per kg of soft lean tissue measured by DEXA (P = 0.045). Half of the trained runners had experienced stress fractures compared with only one of the moderately active controls. The trained runners' lumbar (L2-L4) BMD (1.178 g.cm-2) was not significantly different from that of the active controls (1.283 g.cm-2) (P = 0.074) but, for all subjects combined, there wasa significant inverse relation between L2-L4 BMD and distance run per week (P = 0.036). Further, adding age, body weight, percent body fat, daily energy intake, and daily calcium intake to a stepwise multiple regression analysis did not significantly improve predictive precision. The trained runners consumed nearly twice the amount of calcium (1089 mg.d-1 vs 641 mg.d-1, respectively; P = 0.036), while intake of other nutrients did not differ significantly between groups. Urinary estrone conjugates (E1C) were lower in the trained runners during the early follicular phase (P = 0.028), while pregnanediol-3-glucuronide (PdG) was not significantly different between groups during the luteal phase (P = 0.213). Thus, it appears that lower estrogen production, especially during the early follicular phase, and not progesterone, is associated with lower whole body calcium per kg of soft lean tissue and, probably, L2-L4 BMD. Results of this study also suggest that regular menstrual cycles do not imply normal ovarian hormone function in young women who are engaged in either recreational or competitive running.