In 17 female recreational athletes, ovarian function was monitored using daily hormone measurements and serial ultrasound determinations. Whereas 11 out of 13 women of a control group showed estradiol (E2) maxima beyond 470 pmol/l, progesterone (P4) maxima of 19 nmol/l or more, and a luteal phase length of 9 days or more, only 10 out of 17 athletes satisfied these criteria. Six athletes showed disturbed follicular development, and one athlete showed luteal phase disturbance. Both athletes with disturbed menstrual function (n = 7) and athletes fulfilling the above-mentioned minimal criteria (n = 10) had lower E2 concentrations in all phases of the menstrual cycle (P less than 0.05). P4 concentrations were significantly decreased in the group with disturbed menstrual function (P less than 0.05). Maximal aerobic capacity in the two athlete groups was similar. Neither athlete group showed the expected increase in caloric intake compared with the sedentary controls. It is concluded that recreational running is associated with altered ovarian function. Inadequate nutritional adaptation may be a contributing factor.