BACKGROUND Testosterone inhibits gonadotrophin release in men either directly or after aromatization to oestradiol. We hypothesized that in males the androgen receptor-mediated effect of testosterone on LH release is negligible relative to that of oestradiol. OBJECTIVE To compare the effect of experimentally induced variations of plasma oestradiol levels on LH levels in normal (physiological testosterone levels) and castrated men (very low testosterone levels). METHODS Prospective, open label, intervention. METHODS We suppressed endogenous oestradiol in 10 young men with letrozole 2.5 mg once daily. In these men and in 10 young healthy castrated men, we restored plasma oestradiol levels with oestradiol patches (first week 100 mug/day, second week 50 mug/day, third week 25 mug/day and fourth week no oestradiol patch). METHODS The effect of the intervention on plasma levels of LH were monitored and compared between the groups. RESULTS With the intervention, the mean plasma oestradiol level in the two groups varied from supraphysiological to below the lower reference range. Levels of LH mirrored plasma oestradiol levels in both the groups, as did testosterone in the intact group. Despite similar oestradiol levels, mean levels of LH were significantly higher in the castrated group compared to the intact group for all doses of oestradiol, and supraphysiological levels of oestradiol were unable to suppress LH into the physiological range in the castrated group. CONCLUSIONS Physiological plasma oestradiol levels have a substantial suppressive effect on LH in men. However, low-normal testosterone levels are a prerequisite for suppression of LH into the normal range.