The purpose of this study was to compare four measurements used as estimates of body fat that have been used in previous studies to determine whether the association between body fat and athletic menstrual irregularity (AMI) is measurement dependent. In a sample of 112 marathon runners, 94 responded to questions regarding their menstrual cycle. Of these, 30 (32%) reported irregular or absent menses, and 64 (68%) reported regular menses. Of the 30 subjects reporting menstrual irregularities or amenorrhea, 13 (43%) reported having had menstrual irregularity or amenorrhea prior to taking up running. Estimates of body fat were based on Mellits and Cheek's (1970) equation for estimating percentage of body water, Lutter and Cushman's (1982) height and weight categories, actual gross body weight, and weight loss of 10 lb (4.5 kg) or more since taking up the sport. In this sample of marathon runners, none of the four methods used to estimate body fat supported a relationship between menstrual irregularity and low body fat. A significant (p less than .001) relationship was found between prior menstrual irregularity and the development of AMI after starting to run.