The enzyme aromatase converts testosterone (T) into 17 beta-estradiol and plays a pivotal role in the control of reproduction. In particular, the aromatase activity (AA) located in the preoptic area (POA) of male Japanese quail is a limiting step in the activation by T of copulatory behavior. Aromatase-immunoreactive (ARO-ir) cells of the POA are specifically localized within the cytoarchitectonic boundaries of the medial preoptic nucleus(POM), a sexually dimorphic and steroid-sensitive structure that is a necessary and sufficient site of steroid action in the activation of behavior. Stereotaxic implantation of aromatase inhibitors in but not around the POM strongly decreases the behavioral effects of a systemic treatment with T of castrated males. AA is decreased by castration and increased by aromatizable androgens and by estrogens. These changes have been independently documented at three levels of analysis: the enzymatic activity measured by radioenzymatic assays in vitro, the enzyme concentration evaluated semi-quantitatively by immunocytochemistry and the concentration of its messenger RNA quantified by reverse transcription-polymerase chain reaction (RT-PCR). These studies demonstrate that T acting mostly through its estrogenic metabolites regulates brain aromatase by acting essentially at the transcriptional level. Estrogens produced by central aromatization of T therefore have two independent roles: they activate male copulatory behavior and they regulate the synthesis of aromatase. Double label immunocytochemical studies demonstrate that estrogen receptors(ER) are found in all brain areas containing ARO-ir cells but the extent to which these markers are colocalized varies from one brain region to the other. More than 70% of ARO-ir cells contain detectable ER in the tuberal hypothalamus but less than 20% of the cells display this colocalization in the POA. This absence of ER in ARO-ir cells is also observed in the POA of the rat brain. This suggests that locally formed estrogens cannot control the behavior and the aromatase synthesis in an autocrine fashion in the cells where they were formed. Multi-neuronal networks need therefore to be considered. The behavioral activation could result from the action of estrogens in ER-positive cells located in the vicinity of the ARO-ir cells where they were produced (paracrine action). Alternatively, actions that do not involve the nuclear ER could be important. Immunocytochemical studies at the electron microscope level and biochemical assays of AA in purified synaptosomes indicate the presence of aromatase in presynaptic boutons. Estrogens formed at this level could directly affect the pre-and post-synaptic membrane or could directly modulate neurotransmission namely through their metabolization into catecholestrogens (CE) which are known to be powerful inhibitors of the catechol- omicron - methyl transferase (COMT). The inhibition of COMT should increase the catecholaminergic transmission. It is significant to note, in this respect, that high levels of 2-hydroxylase activity, the enzyme that catalyzes the transformation of estrogens in CE, are found in all brain areas that contain aromatase. On the other hand, the synthesis of aromatase should also be controlled by estrogens in an indirect, transynaptic manner very reminiscent of the way in which steroids indirectly control the production of LHRH. Fibers that are immunoreactive for tyrosine hydroxylase (synthesis of dopamine), dopamine beta-hydroxylase (synthesis of norepinephrine) or vasotocine have been identified in the close vicinity of ARO-ir cells in the POM and retrograde tracing has identified the origin of the dopaminergic and noradrenergic innervation of these areas. A few preliminary physiological experiments suggest that these catecholaminergic inputs regulate AA and presumably synthesis.