Salivary progesterone profiles were studied in a group of recreational women runners (average 12.5 miles per week) and nonexercising control subjects. Although no differences were observed in the average cycle lengths of the two groups, luteal progesterone levels, both peak and average, were found to be significantly lower in the runners. A salivary progesterone level two standard deviations above the follicular phase average was used to discriminate luteal activity with 95% confidence. Runners averaged fewer days with sample values two standard deviations above this level and a shorter average interval between first and last samples observed above this level. However, no evidence was found of a delayed luteal progesterone rise among the runners. These results suggest that even moderate amounts of aerobic exercise may have effects on female reproductive function which would not be suspected from menstrual patterns alone.