Female athletes exhibit a higher prevalence of exercise-associated amenorrhea and oligomenorrhea compared with nonathletic women, and both conditions are related to reduced bone mineral density (BMD), particularly at the spine. This study investigated bone mass and oligomenorrhea and amenorrhea in two groups of competitive female athletes with different skeletal loading patterns: gymnasts and runners. Bone mineral density (g/cm2) of the femoral neck, lumbar spine (L2-4), and whole body was assessed by dual energy X-ray absorptiometry (QDR-1000/W, Hologic Inc., Waltham, MA) in collegiate gymnasts (n = 21) and runners (n = 20), and nonathletic college women (n = 19). The runners and gymnasts had similar values for percent body fat (14.7 +/- 2.2% and 15.6 +/- 2.9%, respectively), which were lower (p < 0.001) than controls (22.3 +/- 3.0%). Lean body mass (LBM) did not differ among the groups, but when adjusted for body surface area, gymnasts had a higher LBM/height2 (p = 0.0001) compared with runners and controls. Muscle strength was significantly greater (p < 0.05) in gymnasts for quadriceps, biceps, and hip adductor force, compared with runners and controls. Gymnasts had a significantly later menarche age (16.2 +/- 1.7 years) compared with runners (14.4 +/- 1.7 years) and controls (13.0 +/- 1.2 years). The prevalence of oligo- and amenorrhea was 47% for gymnasts (6 amenorrheic, 4 oligomenorrheic), 30% for runners (3 amenorrheic, 3 oligomenorrheic), and 0% for controls. Furthermore, athletic groups had similar menstrual histories given the higher proportion of gymnasts who had experienced primary amenorrhea. When evaluated since menarche, however, runners had somewhat longer histories due to an earlier age at menarche and slightly older ages. Dietary calcium intake did not differ among groups, although mean values were below the RDA of 1200 mg/day. By athletic group, BMD at any site did not differ among women with amenorrhea versus oligomenorrhea versus eumenorrhea, although there was a trend for the regularly menstruating athletes in both groups to have slightly higher values. Lumbar spine BMD was lower (p = 0.0001) in runners (0.98 +/- 0.11 g/cm2) compared with both gymnasts and controls (1.17 +/- 0.13 and 1.11 +/- 0.11 g/cm2, respectively). Femoral neck BMD differed among all groups (p = 0.0001): gymnasts = 1.09 +/- 0.12 g/cm2 > controls = 0.97 +/- 0.10 g/cm2 > runners = 0.88 +/- 0.11 g/cm2. Whole body BMD was lower (p < 0.01) in runners (1.04 +/- 0.06 g/cm2) compared with gymnasts and controls (1.11 +/- 0.08 and 1.09 +/- 0.06 g/cm2, respectively).(ABSTRACT TRUNCATED AT 400 WORDS)