section of our paper, we aimed to determine only alterations of day-time (early morning), but not nocturnal, plasma melatonin levels to investigate the role of melatonin in male hypogonadism and the effects of treatment on its plasma levels. Furthermore, early morning measurement of elevated melatonin could mean an increase in total melatonin production; however, they also could mean a delay in the early morning fall in melatonin secretion. We, as others (2,3), have been able to confirm the increased circulating melatonin in patients with IHH, and a tendency is also pointed to in primary hypogonadism, even using a single determination. Although subtle time-dependent changes in circadian melatonin secretion are best identified by sampling an individual subject at frequent intervals for a 24-h period, and interindividual variation in plasma melatonin levels is substantial, it has previously been shown that intraindividual variations tend to be relatively small (4-7). Moreover, Luboshitzky et al. (7) recently suggested that melatonin secretory patterns are specific for each individual. We have, therefore, evaluated melatonin levels in blood samples drawn at the same time (0730 h) before and after therapy to prevent the effect of intraindividual variation on melatonin levels.Short-term treatment with gonadotropins or testosterone, although inducing significant increases in circulating testosterone, did not achieve normal testosterone levels, either in primary or in secondary hypogonadal patients. Therefore, decreases in melatonin observed in both groups may not have reached significant differences, because of an insufficient testosterone milieu. Also, as demonstrated recently by Luboshitzky et al. (7) (after 4 months of testosterone treatment), it can not be ruled out that a much more prolonged time of treatment may be necessary to obtain significant changes in plasma melatonin.The recent demonstrations of melatonin receptors on human granulosa cell membranes (8) and prostate glandular epithelium (9), 'ZSI-labeled binding sites in the testes and ovaries of the duck (lo), and androgen receptors in rat pinealocytes (11) suggest a direct relation between the pineal and the gonad. In our study, observation of a tendency for melatonin to decrease after treatment in both groups suggests some degree of regulation of melatonin secretion by gonadal steroids or gonadotropins. However, we failed to demonstrate any correlation between early morning melatonin and circulating gonadotropin, PRL, and gonadal steroid levels. Although we did not measure integrated melatonin values and FSH, LH, testosterone, or estradiol profiles, it seems that the antigonadal action of melatonin is not related to circulating gonadotropin, PRL, or sex steroid levels. Similar observations were also demonstrated by some investigators (6, 12, 13). In contrast, others found a modulatory relationship between melatonin and sex steroids (2, 14, 15). Thus, the regulation of melatonin by sex steroids requires further investigation. Furthermore, the lack of an ...