PURPOSE: This study aimed at assessing the impact of physical training on psychological functioning at the onset of a prospective study of psychological and somatic maturation of adolescent female athletes. METHODS: Twenty-seven highly trained gymnasts aged 12.7 +/- 1.1 year (mean +/- SD, training load = 18-26 hr/week) and 16 age-matched but moderately trained swimmers (13.0 +/- 0.9 yr, training load 4-15 hr/wk) were submitted to standardized somatic and psychiatric examinations during training camps. RESULTS: Gymnasts were significantly shorter, lighter and thinner (p < 0.001) than swimmers. Their bone age was moderately but significantly retarded (-1.42 +/- 0.99 yr, p < 0.001) in contrast with swimmers in whom it was adequate for chronological age (+0.28 +/- 0.94 year, ns). Only 7.4% of gymnasts had already had menarche in contrast with 50% of age-matched swimmers (p = 0.003). Psychological functioning was considered as normal in all subjects. However, seven athletes including 3/27 gymnasts and 4/16 swimmers (p = 0.394) were considered as subjects "at risk" to develop a manifest mental disorder over time. Ten gymnasts (41.7%) presented with a global delay in psychological maturation, whereas no such case was observed among swimmers (p = 0.015). No correlation could be established between psychological delay and pubertal retardation (p = 0.210). CONCLUSION: Strenuous training in gymnastics for more than 1 yr has so far no detectable interference with the normal maturational events of adolescence. The outcome of athletes at risk to develop psychopathology as well as those with a global delay in psychological maturation who presented as if they were still in the latency period, remains uncertain.