The aim of this investigation was to evaluate the urinary profile of androgen metabolites during menstrual cycle in both young-trained female athletes, and young sedentary women, not presenting any pathological signs. Urines were collected for 24 hours (08 : 00 a. m. the first day to 08 : 00 a. m. the second day) from all sportive and sedentary subjects. All steroids were measured by specific radioimmunological analysis, and the implications of these results in terms of concentrations and modifications by exercise will be discussed. During follicular phase, control values were respectively, testosterone glucuronide (TG): 1.67 +/- 0.70 nmol x mmol C -1; epitestosterone glucuronide (ETG): 2.51 +/- 0.88 nmol x mmol C -1; TG/ETG ratio: 0.72 +/- 0.26 and cortisol (FLU): 10.02 +/- 0.79 nmol x mmol C -1. No significant modifications were observed during luteal phase (respectively: TG: 1.48 +/- 0.50 nmol x mmol C -1; ETG: 2.65 +/- 0.93 nmol x mmol C -1; TG/ETG ratio: 0.67 +/- 0.31 and FLU: 9.29 +/- 3.37 nmol x mmol C -1. Similarly, no significant effect of physical training was observed on studied parameters between these two groups during either follicular phase (TG: 1.96 +/- 1.00 nmol x mmol C -1; ETG: 1.97 +/- 0.70 nmol x mmol C -1; TG/ETG ratio: 0.66 +/- 0.05 and FLU: 11.31 +/- 3.73 nmol x mmol C -1) or luteal phase (TG: 1.93 +/- 0.86 nmol x mmol C -1; ETG: 3.19 +/- 1.23 nmol x mmol C -1; TG/ETG ratio: 0.69 +/- 0.33 and FLU: 9.52 +/- 3.86 nmol x mmol C -1). It is concluded that although physical training could play a role in androgen metabolism, it has no significant incidence on urinary TG/ETG ratio. This study thus confirms that sportswomen can also be considered as normal subjects when they do not present any obvious endocrine disorder induced by physical activity.