To help us identify appropriate techniques and laboratories for measuring hormones, we studied the variability and reproducibility of assay measurements of androstanediol glucuronide, androstenedione, dehydroepiandrosterone (DHEA), DHEA sulfate, dihydrotestosterone, testosterone, androstanediol, androsterone glucuronide, and androsterone sulfate for five men. Four sets of two aliquots from each sample were sent to participating laboratories, and one set was used for analyses monthly for four consecutive months. For each assay, estimates of components of variance were then used to estimate the coefficient of variation, the intraclass correlation between measurements on different days from a given individual, and the minimum detectable relative difference for a standard design. These data indicate that for at least one of the laboratories a single sample with two laboratory replicates per sample of androstanediol glucuronide, androstenedrone, DHEA, DHEA sulfate, and dihydrotestosterone yields an intraclass correlation coefficient exceeding 0.80 and can be used to discriminate reliably among men. The results for testosterone, androstanediol, androsterone, glucuronide, and androsterone sulfate do not meet this test. These data do not allow us to estimate the component of variation that corresponds to repeated blood samples taken over time from the same man. This reliability study design is, however, entirely appropriate for the typical case-control study which utilizes only one sample per subject.