Plasma testosterone concentrations in men, first quantified nearly half a century ago, are now measured routinely as a primary index of androgen status. Most clinical laboratories employ a multichannel, fully automated analyzer. Current evidence suggests that these analyzers are capable of satisfactorily quantifying the concentration of total plasma testosterone in men. Newer technology, in the form of tandem mass spectrometry, may, in the future, replace these automated platforms, providing a more specific estimate of testosterone concentration. When concentrations of plasma testosterone are found to be around the lower limit of normal ( 9.0 nmol/l; microg/l=nmol/lx0.288), some measure of bioactive testosterone should be sought. This may be a free (non-protein-bound) or bioavailable (free plus albumin-bound) testosterone and may be measured (laborious and time-consuming and therefore unsuited to routine clinical laboratories) or calculated using any one of a variety of mathematical expressions. Sampling for the estimation of plasma testosterone should be carried out in the morning, before 11:00 h, to obviate the effect of the marked diurnal variation in testosterone production. In samples found to have an equivocal concentration (7.0-12.5 nmol/l) at least one more estimate should be obtained to account for the possible significant intra-individual variability. Although it is generally accepted that the concentrations of total, free and bioavailable testosterone decline as men age, the majority of elderly men have testosterone levels in the young adult range (9.0-35 nmol/l) and some maintain a diurnal rhythm. Salivary testosterone offers a non-invasive estimate of free testosterone but there does not appear to be an immediate demand for a routine salivary testosterone service.