Gonadotropin secretion is pulsatile in prepubertal and early pubertal boys, and the onset of puberty is characterized by a sleep-associated rise in LH pulse amplitude. To determine whether an augmentation in LH pulse frequency as well as amplitude occurs at the onset of puberty, we studied gonadotropin secretion in 21 early to midpubertal boys. Blood samples were taken every 20 min (every 15 min in 4 boys) for LH determinations. A 2-fold increase in LH pulse frequency occurred during the nighttime sampling period (2200-0400 h) compared to that in the hours when the boys were awake (1000-2200 h). The maximum frequency (0.7 pulses/h) occurred between 2400 and 0200 h. The mean plasma LH concentration increased during the night from 2.3 +/- 0.2 (+/- SE) mIU/mL (2.3 +/- 0.2 IU/L) between 2000-2200 h to a maximum of 6.2 +/- 0.4 (6.2 +/- 0.4 IU/L) between 0200-0400 h. The mean plasma LH decreased to 5.5 +/- 0.4 mIU/mL (5.5 +/- 0.4 IU/L) between 0400-0600 h and to 4.2 +/- 0.5 (4.2 +/- 0.5 IU/L) between 0600-0800 h. Plasma testosterone rose during the night to a mean maximum value of 2.4 +/- 0.5 (+/- SE) ng/mL (8.3 +/- 1.7 nmol/L). This finding suggested that the rise in testosterone might play a role in decreasing LH secretion during the later hours of sleep (after 0400 h). To address this question and to study further the effects of testosterone in early puberty, we measured plasma LH concentrations every 10 min from 2000-0800 h in 8 early to mid-pubertal boys before and during short term testosterone administration. Saline or testosterone at a concentration of 9.33 micrograms/mL (32 mumol/L) was infused at a rate of 10 mL/h from 2100-1200 h to shift the nighttime testosterone rise 3 h earlier than would occur spontaneously. Blood samples were obtained every 10 min for LH and every 30 min for testosterone determinations from 2000-0800 h. Pituitary responsiveness was assessed by administering sequential doses of synthetic GnRH (25 and 250 ng/kg) at 1000 and 1200 h, respectively. The nighttime increase in LH pulse frequency and mean plasma LH concentration occurred between 2300 and 0200 h despite testosterone infusion. However, testosterone infusion was associated with significantly lower mean plasma LH concentrations from 0200-0800 h compared to those on the night of the saline infusion. Pituitary responsiveness to synthetic GnRH was unaltered by testosterone administration.(ABSTRACT TRUNCATED AT 400 WORDS)