It has been hypothesized that young athletes, undertaking intensive training, have delayed menarche due to the effects of training at an early age, although it is known that other genetic and environmental factors contribute to this observed later menarche. As part of a longitudinal study of the effects of intensive training during puberty and adolescence we investigated age of menarche in 222 athletes and their mothers. All the sports studied (gymnastics, swimming and tennis) had later mean ages of menarche (14.3, 13.3 and 13.2 years respectively) than the previously reported UK reference value of 13.00 years. A positive correlation was found between menarcheal age in mothers and daughters (r = 0.27, p < 0.01). Analysis of covariance, using maternal menarcheal age, socioeconomic group, duration of training and type of sport, confirmed that maternal menarcheal age and type of sport were having a significant influence on subject's age of menarche. As maternal menarcheal age and sport were the best predictors of menarcheal age in the athletes we studied, it would appear that menarche was intrinsically late rather than delayed. This suggests that some form of sport-specific selection may have occurred. It therefore seems likely that late maturation of gymnasts contributes to a girl's decision to continue participating in the sport rather than intensive training delaying menarche.