To investigate the effects of alterations in GnRH pulse frequency on gonadotrophin secretion, we administered low dose GnRH pulses (25 ng/kg) at hourly or 2-hourly frequencies to eight normal men. All subjects received GnRH pulses i.v. every 2 h for 88 h. Following this, exogenous GnRH was discontinued in four normal men (Group A, GnRH withdrawal), and the frequency of GnRH injections was increased to one pulse every hour for 24 h in the other four normal men (Group B, hourly GnRH). Blood samples were obtained every 20 min for LH and FSH and every 12 h for testosterone (T) and oestradiol (E2). Plasma LH increased in all subjects during injection of GnRH pulses every 2h. Withdrawal of GnRH pulses in Group A men was accompanied by a fall in mean LH, reductions in LH pulse amplitude (means +/- SEM: control 6.5 +/- 1.0; GnRH withdrawal 4.0 +/- 0.5 mIU/ml) and pulse frequency (control 5.5 +/- 0.2; GnRH withdrawal 3.5 +/- 0.7 pulses/12 h), and an increase in plasma E2 (control 122 +/- 15; GnRH withdrawal 340 +/- 37 pmol/l). Gonadotrophin responses to GnRH (25 ng/kg) were normal when tested 32 h after GnRH withdrawal. Injection of hourly GnRH pulses in Group B men was accompanied by a time-dependent change in mean LH, which transiently rose, then fell, and subsequently rose to a plateau during the second 12 h period of hourly GnRH. The final rise in LH was accompanied by an increase in LH frequency to 11.8 +/- 0.3 pulses/12 h. These data suggest that: (1) increases in gonadal steroids decrease LH secretion by reducing the amplitude and frequency of endogenous GnRH pulses; and (2) the normal adult male pituitary requires approximately 12 h to initiate a sustained increase in LH secretion in response to a doubling in GnRH pulse frequency.