Abstract: Two experiments were designed to determine whether gonadal steroids during the estrous cycle may modify the characteristics of the plasma melatonin rhythm. In the first experiment, 12 ovariectomized estradiol-treated ewes were used and exposed to constant short days. The experimental design was a latin square to distinguish between steroid treatments and individual effects on melatonin secretion. Twenty four hours before the bleeding period (hourly during 20 hr) and with a 1 week interval, animals were treated with a) additional subcutaneous estradiol implants, b) progesterone devices, or c) control. In the second experiment, nine ewes received a treatment combining fluorogestone acetate devices and pregnant mare serum gonadotrophin to induce synchronous ovulations. Samples for melatonin determination were obtained hourly for 13 hr at three stages of estrous cycle: follicular phase, early luteal phase, and late luteal phase. Ovarian activity was monitored by taking daily samples for progesterone analysis and ovulation rate was determined by laparoscopy. Duration and mean melatonin plasma concentrations of the elevation were calculated for each ewe and each night and analysed by latin square test (experiment 1) or ANOVA (experiment 2). Melatonin concentrations of elevation and duration of elevation were not significantly affected by hormonal treatments or by phase of estrous cycle. A strong individual effect was detected (P<0.01) for both parameters in both experiments. It was concluded that melatonin secretion is unaffected by steroid administration or by phase of estrous cycle. The existence of very high inter-individual variation suggest that both parameters are individual characteristics of each animal which may have a strong genetic basis.