The relationship between body fat percentage and menstrual patterns was examined in a longitudinal study of 25 collegiate athletes and 41 college-aged nonathletic control subjects. The athletes were training in gymnastics, figure skating, synchronized swimming, volleyball, or track. Data were collected each month concerning cycle, flow length, weight, and skinfold thicknesses (iliac crest, triceps, and subscapular). Height was determined once. Body fat percentage was estimated using the Sloan and Weir (1970) nomogram. Total energy expenditure was estimated for athletes according to sport. Mean body fat percentages of controls and athletes differed significantly,t= 4.4,p< .01, but no significant difference was found in cycle length,F= 3.3,p= .078. Multiple regression analysis revealed no clear relationship between length of menstrual cycle and height, weight, skinfold thickness, or total energy expenditure. No significant relationship existed between body fat percentage and menstrual cycle length as tested with the Pearson correlation coefficient. This study confirmed that menstrual phenomena cannot be defined as normal or pathologic by rhythmic occurrence alone. Rather, an individual's menstrual cycle length may be more suitably evaluated according to a multivariable continuum of menstrual patterns.