1. Twelve patients with the nephrotic syndrome were prescribed for 4 week periods a normal protein diet (NPD) containing 1 g of protein/kg ideal body weight. They were then prescribed for further 4 week periods in random order diets with high (HPD) and low (LPD) protein contents, respectively 2.0 and 0.5 g/kg ideal body weight. 2. Compliance was confirmed by dietary history and measurement of urinary excretion. 3. Serum albumin was the same on all diets. Twenty-four hour urinary protein excretion increased progressively with increasing dietary protein (LPD 6.1 g. NPD 8.2 g. HPD 9.2 g). Recumbent plasma renin activity and serum phosphate were significantly increased on HPD (plasma renin activity: LPD 5.7, NPD 4.6, HPD 8.2 pmol of angiotensin I min-1 1(-1); serum phosphate: LPD 1.27, NPD 1.26, HPD 1.41 mmol/l). 4. There was no evidence of protein-induced hyperfiltration or hyperperfusion: 51Cr-ethylenediaminetetra-acetate and [125I]iodohippurate clearances were similar on all three diets. 5. Since proteinuria, increased plasma renin levels and hyperphosphataemia may contribute to progression of renal failure and because HPD did not improve hypoalbuminaemia, the use of HPD in the nephrotic syndrome should be abandoned. 6. Until it can be established that LPD, which is accompanied by the least proteinuria, does not, with long-term feeding, lead to malnutrition, NPD should be used in the treatment of the nephrotic syndrome.