The effects of intrahypothalamic and subcutaneous implants of testosterone (T) and those of hypothalamic lesions on resting levels of circulating LH and pituitary responsiveness to exogenous LHRH were studied in castrated male rats to elucidate hypothalamic and pituitary regulation of LH secretion. Two hundred mug implants of testosterone propionate (TP) in the median eminence region suppressed plasma LH titers before evidence of direct inhibition of pituitary function (as indicated by testing with LHRH) was found. Such implants release appreciable amounts of T into the peripheral circulation in the immediate post-operative period, and SC Silastic (constant release) capsules containing T have similar effects. The findings suggest that, regardless of the site of implant, the initial negative feedback inhibition of LH by T is not dependent on direct action at the pituitary levels but rather appears to be a hypothalamic effect. In the days following exposure to hypothalamic or peripheral implantation of T, however, a progressively developing decline in the response to exogenous LHRH was observed. In order to determine whether this effect results from suppression of endogenous LHRH release, the median eminence-arcuate region was destroyed to remove the source of LHRH. In these animals, the suppression of plasma LH was evident on the first day after the lesion, but pituitary responsiveness to LHRH was unaffected until after one week. When Sialastic capsules were implanted SC into lesioned animals, a more rapid (less than 1 week) inhibition of pituitary responsivity ensued. Suprachiasmatic lesions did not affect basal LH secretion or pituitary responses to LHRH. The data provide evidence for a dual feedback action of T on LH in castrated male rats: an initial inhibitory effect presumably due to hypothalamic inhibition (commencing at around 6h after hypothalamic of SC implantation of T), and a subsequent suppression of pituitary responisveness (after one day) presumably due to direct action of T on the pituitary. In addition to these phenomena, findings in rats bearing median eminence-arcurate lesions suggest that the removal of endogenous LHRH by itself leads to an eventual decline in pituitary responsiveness (greater than one week postoperatively).