Electrolyte excretion was observed during 24 oestrous cycles in housed sheep, together with mixed salivary Na/K ratio during 10 additional cycles. 1. The sharp fall in food and fluid intake at oestrus accompanied a peak of sodium excretion which changed to peak retention 3 days later, both in faeces and urine. 2. Potassium excretion declined with food intake at oestrus but subsequently failed to recover to pre-oestrous levels dispite full recovery of dietary intake. 3. Curiously, water intake also recovered completely whereas urinary and faecal water retention continued; faecal loss actually exceeded renal excretion on these liberal water intakes. 4. Changes in salivary, urinary and faecal Na/K indicated an aldosterone peak neither during the luteal phase nor at oestrus but three days later. The data rarse questions concerning the regulation of water and electrolyte balance within the normal cycle. They also provide a baseline for the investigation of renal effects of gonadal steroids. Possible roles for aldosterone, ADH and progesterone in maintaining fluid and electrolyte balance are discussed, emphasising problems confronting species which have evolved with heavy obligatory potassium excretion but undependable supplies of sodium and water.Sheep vary their sodium intake during the oestrous cycle, taking least at oestrus and most at the climax of the luteal phase [Michell, 1975]. This seemed consistent with the idea that oestrogens stimulate salt retention whereas progesterone promotes natriuresis [Landau, 1973], but specific data in sheep were lacking.Salt retaining effects of oestrogens were suspected almost as soon as those of adrenal steroids [Thorn, Nelson and Thorn, 1938] but considerable conflict emerges from the literature on gonadal steroids and electrolyte excretion [Christy and Shaver, 1974;Lindheimer and Katz, 1977]. The factors influencing aldosterone levels during the human cycle remain uncertain [Schwartz, 1975]. The luteal rise can be attributed to sodium loss induced by progesterone but the rise near ovulation is less easily explained [Sundsfjord and Askvaag, 1973]. Such questions have practical significance since the aetiology of both premenstrual tension and contraceptive hypertension remain uncertain