We measured bone mineral density (BMD) in 25 premenopausal ultramarathon (56 km) runners aged 29-39 yr and related risk factors for decreased BMD with actual BMD. Fifteen runners who had never had oligo/amenorrhea (R) were compared with 10 runners (OA): 4 oligomenorrheic, 2 amenorrheic, and 4 with prior oligo/amenorrhea. Menstrual, dietary and training data were obtained. BMD of the lumbar spine (LS) and proximal femur (F) were measured by dual energy x-ray densitometry. Both groups had similar body mass (58 +/- 8 vs 57 +/- 8 kg), running and dietary histories. F BMD was not different (P = 0.07) and correlated only with BMI (P < 0.05; r = 0.43). LS BMD was lower in OA (0.946 +/- 0.098 g.cm-2) than R (1.088 +/- 0.069 g.cm-2; P < 0.001). Menstrual History Index (MHI), (estimated periods.yr-1 since age 13), was higher in R (11.6 +/- 0.6) than OA (9.4 +/- 2.1; P < 0.01). LS BMD correlated with MHI (P < 0.0005; r = 0.67) and years oligomenorrheic (P < 0.01; r = -0.58) but not years amenorrheic, parity, breastfeeding, diet, or training. In conclusion, in mature women distance runners low LS BMD is related to a history of oligo/amenorrhea regardless of resumption of regular menstrual cycles in some subjects. Not only amenorrhea, but also prolonged oligomenorrhea may negatively influence peak adult bone mass.