To investigate daily cortisol production and clearance rates in a group (n = 18) of normal unstressed pubertal males, we applied deconvolution analysis to serum cortisol concentrations obtained every 20 min for 24 h. Subject-specific characterization of adrenocortical secretory episodes, cortisol production rate, and serum hormone half-life for nine early pubertal (Tanner I or II; early) and nine late pubertal (Tanner IV or V; late) subjects was undertaken to assess potential roles of sexual maturation and changing gonadal steroid hormone concentrations on glucocorticoid physiology. The estimated cortisol production rate for the early group [16.8 +/- 1.3 mumol/m2 x day (6.1 +/- 0.4 mg/m2 x day)] was indistinguishable from that of the late subjects [14.8 +/- 1.4 mumol/m2 x day (5.3 +/- 0.5 mg/m2 x day)]. No differences were observed between the two pubertal groups in the secretory burst frequency and half-duration, mass of cortisol released per secretory episode, average maximal rate of hormone secretion, and serum cortisol half-life. A significant diurnal pattern of cortisol secretion was observed for all subjects manifest by nyctohemeral variations in the frequency of adrenocortical secretory bursts, the amplitude (maximal rate of cortisol secretion) and the mass of cortisol released per secretory episode. Maximum serum hormone concentrations occurred between 0706 and 1114 h. We conclude that in normal pubertal males: 1) cortisol production rates as estimated by deconvolution analysis are in agreement with other recent independent isotopic estimates, but are lower than many previous estimates; 2) the rise in serum gonadal steroid hormone levels is unassociated with alterations in the production rate or metabolic clearance of cortisol; and 3) increased secretory burst frequency, increased amplitude (maximal rate of cortisol secretion attained within each secretory event), and increased mass of cortisol released per adrenocortical secretory episode give rise to the normal diurnal rhythm of circulating cortisol.