This study was conducted to test the sensitivity to gonadal steroids of the systems regulating food intake in sea bass. Animals were treated with silastic implants containing 17-beta-estradiol or testosterone. Self-feeding was recorded for 31 days using computerized demand feeders and unfed-pellet recovery systems. Both steroids strongly decreased self-feeding levels, feed efficiency and specific growth rates. The linear growth of fish treated with testosterone was higher than in 17-beta-estradiol treated fish. In the second experiment, fish were treated with lower 17-beta-estradiol doses and 11-keto-androstenedione, a precursor of the main fish androgen (11-keto-testosterone). The results demonstrated a dose-response effect of estrogen and no effect of non-aromatizable androgens on food intake or growth performance. The inhibitory effect of testosterone on food intake seems to be mediated by its aromatization to estradiol, while linear growth promotion is mediated by the androgen per se. Data suggest that gonadal steroids may be involved in the seasonal feeding pattern of sea bass. The results demonstrate the sensitivity of the mechanisms regulating food intake to estrogenic compounds and point to the risk of including feed containing estrogenic substances in fish diets as well as the risk involved in exposure to "estrogenic environments".