Whereas intensive and regular physical training is known to alter female reproductive function, its potential role in growth is still controversial. At the beginning of a longitudinal growth study of young elite female gymnasts (n = 34, 15-25 h/wk training) and moderately trained swimmers (n = 19, 5-15 h/wk), patterns of recalled parental growth and pubertal maturation were compared with those of parents of 25 sedentary school girls. These data were also correlated to the height, weight, pubertal development as well as adult height prognosis of their daughters. Bone age was estimated using the methods of Greulich-Pyle and Tanner (RUS score) and adult height prognosis using the methods of Bayley-Pinneau (BP), Roche-Wainer-Thissen (RWT), and Tanner et al. (TW2). Parents of gymnasts were significantly lighter (fathers: P = 0.027; mothers: P = 0.038) and shorter (fathers: P = 0.034; mothers: P less than 0.001) than those of swimmers and controls. Consequently, target heights of gymnasts were also significantly shorter (P less than 0.001). Recalled menarche occurred significantly later (P = 0.030) in mothers of gymnasts who, in turn, grow much alike their mothers. At the first visit, the gymnasts were shorter and lighter for age than swimmers and controls. Their bone age (11.0 +/- 1.3 years, mean +/- SD) was retarded (P less than 0.001) when compared with chronological age (12.6 +/- 1.2 years). Adult height prognosis was lower for gymnasts than for other girls (BP: P less than 0.001; RWT: P = 0.023, TW2: P less than 0.001), but adequate for target height range.(ABSTRACT TRUNCATED AT 250 WORDS)