Because the effect of exercise on leptin was not established, we controlled energy intake (I) and exercise energy expenditure (E) to distinguish the independent effects of energy availability (A = I - E) and exercise stress (everything associated with exercise except its energy cost) on the diurnal leptin rhythm in healthy young women. In random order, we set A = 45 and 10 kcal. kg lean body mass(-1) (LBM) x day(-1) for 4 days during the early follicular phase of separate menstrual cycles in sedentary (S, n = 7) and exercising (X, n = 9: E = 30 kcal x kg LBM(-1) x day(-1)) women. Low energy availability suppressed the 24-h mean (P < 10(-6)) and amplitude (P < 10(-5)), whereas exercise stress did not (both P > 0.2). Suppressions of the 24-h mean (-72 +/- 3 vs. -53 +/- 3%, P < 0.001) and amplitude (-85 +/- 3 vs. -58 +/- 6%, P < 0.001) were more extreme in S vs. X than previously reported effects on luteinizing hormone pulsatility and carbohydrate availability. Thus the diurnal rhythm of leptin depends on energy, or carbohydrate, availability, not intake, and exercise has no suppressive effect on the diurnal rhythm of leptin beyond the impact of its energy cost on energy availability.