Doses equivalent to 18, 72 and 216 mg testosterone (T)/week were administered for 6 months to eugonadal men. Urinary excretions of androgen glucuronides (G) were quantitatively analyzed by gas chromatography-mass spectrometry with stable isotope dilution during periods of control (without hormone treatment), T administration and recovery. The lowest dosage T administration did not affect the androgen profile, while higher dosages generally increased urinary excretions of T metabolites (TG, T sulfate, glucuronides of androsterone, etiocholanolone, 5 alpha- and 5 beta-androstane-3 alpha,17 beta-diol) and decreased excretions of conjugates of epitestosterone (ET) and its precursor androgen 5-androstene-3 beta,17 alpha-diol. A dose-dependent decrease of urinary LH in response to T was also observed. The ratio (T/ET)G, which is the sole official criterium for assessment of T self-administration by athletes, increased above the threshold value of 6 in most of the subjects, but not all, after the two highest dosage T regimens, and returned to normal during the recovery period. False positive or negative testing emphasizes the need for improvement of testing procedures. In this regard, valuable complementary information may be gained from ratios such as TG/ET(Total), TG/LH, (T/5-androstene-3 beta,17 alpha-diol)G, (5 alpha/5 beta)androstane-3 alpha,17 alpha-diol and (5 alpha/5 beta)androstane-3 alpha,17 beta-diol.