Neurons containing calcitonin gene-related peptide in the medial preoptic nucleus exhibit the largest neurochemically denned sex difference in the rat preoptic area with a 20-fold difference in cell numbers. The gonadal steroid hormones responsible for this sexual dimorphism have been investigated by examining calcitonin gene-related peptide immunoreactivity in the preoptic area of adult rats receiving a variety of perinatal and adult gonadal steroid manipulations. Cells immunoreactive for calcitonin gene-related peptide were examined in two populations within the preoptic area, one in its ventrolateral aspect and the other located in the lateral division of the medial preoptic nucleus. Cell profile counts estimate numbers of calcitonin gene-related peptide-containing cells in the medial preoptic nucleus of the female to be22.2 ± 3.0 cells/section compared with1.0 ± 0.2 in the male ( P < 0.01). No sex differences existed in the preoptic ventrolateral population of calcitonin gene-related peptide cells (males4.3 ± 0.2, females4.4 ± 0.6 cells/section). Gonadectomy of male rats on postnatal day 2 resulted in the appearance of a calcitonin gene-related peptide-containing cell population in the medial preoptic nucleus which was indistinguishable from intact female rats (19.3 ± 2.2 cells/section). Gonadectomy of adult male rats resulted in a modest increase in calcitonin gene-related peptide cell numbers within the medial preoptic nucleus (8.8 ± 0.4 cells/section) and this was fully reversed by replacement of testosterone (0.7 ± 0.2 cells/section). Replacement with oestradiol or dihydrotestosterone instead of testosterone resulted in only partial reversal of the gonadectomy-induced increase in calcitonin gene-related peptide cell numbers (4.4 ± 0.2 and5.4 ± 1.8 cells/section, respectively). In female rats, calcitonin gene-related peptide cell numbers in the medial preoptic nucleus were significantly reduced by perinatal testosterone treatment with a single administration of testosterone on postnatal day 5 (6.5 ± 2.1 cells/section) having the same effect as a 10-day period of perinatal testosterone administration (8.5 ± 1.4 cells/section). Female rats receiving a single testosterone injection on postnatal day 5 and then a testosterone implant as an adult displayed a male pattern of calcitonin gene-related peptide immunoreactivity (1.2 ± 0.3 cells/section). In all experimental groups, no significant differences in the numbers of calcitonin gene-related peptide cells were identified within the preoptic ventrolateral population. These results show that sex differences in the medial preoptic nucleus calcitonin gene-related peptide population are initiated by testosterone on or around postnatal day 5 but that full expression of this sexual dimorphism is dependent on testosterone exposure as an adult. The mechanisms by which testosterone inhibits calcitonin gene-related peptide immunoreactivity in the neonatal and adult male are unknown but may, in the latter case, result from a combination of androgen and oestrogen receptor activation. The calcitonin gene-related peptide cells of the rat medial preoptic nucleus are an example of a sexually dimorphic neural population which results from both “organizational” and “activational” effects of testosterone.