The association between family size and age at menarche was evaluated in 370 university athletes representing seven sports, 291 white and 79 black. Age at menarche, number of children in the family (family size), and birth order were collected retrospectively by questionnaire with a follow-up interview if necessary. The effect of family size on menarche was estimated with linear regression and multiple regression. For each additional sibling in the family-age at menarche was later by 0.17 yr in white athletes, 0.21 yr in black athletes, and 0.16 yr in the total sample (about 2.0-2.5 months). The slopes for white (beta = 0.167) and black (beta = 0.208) athletes did not differ significantly. Controlling for the effects of birth order significantly increased the family size effect in white athletes (beta = 0.221), but not in black athletes (beta = 0.203) or in the total sample (beta = 0.164). The slopes for white and black athletes, after controlling for birth order, also did not significantly differ. Thus, after controlling for birth order, menarche was later by 0.16-0.22 yr (about 2.0-2.6 months) for each additional sibling in the family, which was within the range of family size effects observed in samples of nonathletes and athletes. The family size effect also persists after controlling for maternal age at menarche in a subsample of 134 athletes, thus implying an independent effect of family size. Later ages at menarche observed in athletes may thus be due in part to larger family sizes.