The effects of nutrition on the hypothalamo-pituitary-gonadal axis were studied in three groups of six mature Merino rams that were fed for 56 days with a ration that maintained their initial live mass (intermediate diet: 675 g chaff plus 175 g lupins), the same ration with a lupin supplement (high diet: 675 g chaff plus 825 g lupins), or about half of the intermediate ration (low diet: 475 g chaff plus 125 g lupins). Lupin seed provides a highly (95%) digestible source of energy and protein. Plasma concentrations of LH, FSH, testosterone and inhibin were measured in blood samples collected over 24 h on the day before dietary treatments began (day -1), then on days 0, 1, 5, 14, 28 and 56. Compared with the intermediate diet, the high diet significantly increased live mass within 14 days and testicular size within 28 days, and these differences increased steadily throughout the experiment. Plasma FSH concentrations and LH pulse frequency increased within 5 days, but these effects were maintained for only 14 days. Decreasing the nutritional status reduced live mass and testicular size within 7 days, led to a low LH pulse frequency that persisted throughout the experiment, but did not affect FSH concentrations. Significantly less testosterone was secreted over 24 h in the low dietary group than in the intermediate or high group until day 28. The high group tended to secrete more than the intermediate group, but only at the beginning of the experiment when LH pulse frequencies differed between these groups. The testosterone response to each endogenous LH pulse, or following an injection of ovine LH i.v. (200 ng kg-1 live mass), was not related to testicular size or dietary treatment at any stage of the experiment. Similarly, plasma inhibin concentrations were not related to change of diet, despite large differences in testicular. size. We concluded that the effects of nutritional status on testicular size in mature rams are at least partly mediated through changes in gonadotrophin secretion. Both increases and decreases in food supply affected LH pulse frequency, suggesting the involvement of hypothalamic mechanisms. However, the lack of an effect of a decrease in nutritional status on the secretion of FSH and inhibin and the inconsistent long-term relationship between LH pulse frequency and testicular size suggest that the effects of diet on testicular growth also involve mechanisms that are independent of changes in gonadotrophin secretion.