SummaryThe problem of menstrual dysfunction in athletes was recognised at about the same time as a theory was developed that a critical fat level was necessary for the onset and maintenance of menstrual function (17% and 22% of bodyweight, respectively). This theory was acceptable because of the frequency of leanness in athletes experiencing menstrual dysfunction and because of the role of adipose tissue in the intraconversion of hormones which could affect hypothalamic and pituitary regulatory centres. Research on this topic has been hampered by the extensive use of surveys, confining sampling to specific sports, use of inaccurate methods of body composition assessment, and lack of data on hormonal changes. Studies using appropriate body composition measures do not support the critical fat theory, but they do not exclude a role for body composition changes in the regulation of menstrual function. The probability of finding menstrual dysfunction in very lean athletes is high, but not absolute, and there is no assurance that dysfunction will improve merely by increasing body weight. Perhaps of more concern, however, is recent research on very lean, long term amenorrhoeic athletes who train intensely and show a loss of bone mineral, apparently related to low oestrogen levels. This mineral deficiency is apparently a factor in stress fractures. The bone mineral content of these athletes is similar to that of post-menopausal women susceptible to osteoporosis. Although the loss of bone due to short term amenorrhoea may be reversible, the prognosis for a long term deficiency is not presently known.