The prevalence of amenorrhea was studied among 113 professional and student ballet dancers in the Netherlands (mean age 23.3 yr, SD 4.8). Sixty-one dancers not on oral contraceptives were included in the subsequent analyses. Six cases (prevalence 9.8 percent, 95 percent confidence interval: 2.4-17.2) with secondary amenorrhea (< or = 4 cycles per year; previous menstruation > or = 3 months prior to the study; menarche > or = 1 yr prior to the study) were found. Two dancers had primary amenorrhea (no menarche at the age of > or = 16). There was a negative correlation between the age of menarche and the number of menstrual cycles during the 12 months preceding the study (r = -0.46, P = 0.001). Body composition (four-compartment model), amount of dancing (recorded), resting energy expenditure (ventilated hood), dietary intake (recorded), and indices of eating disorders (Eating Disorders Inventory, EDI) were studied in 15 of the dancers, 5 amenorrheic and 10 eumenorrheic. No significant differences were found between the amenorrheic and eumenorrheic dancers. An explanation for the lower prevalence of amenorrhea in ballet dancers in the Netherlands, compared with U.S. dancers, was not obvious. Relatively low EDI scores (25.8, SD 14.5) in a subsample of 24 dancers could indicate less rigid emphasis on leanness and dieting.