Serum magnesium and potassium concentrations were determined in 188 patients undergoing diuretic treatment for mild to moderate hypertension in four controlled clinical trials (eleven treatment groups). Measurements were made before and after 12 weeks of treatment with piretanide monosubstance (including the retarded form) in the range 3 to 12 mg daily, with triamterene as monosubstance and in fixed dose combination with piretanide, and with amiloride in combination with hydrochlorothiazide. Eumagnesaemia and eupotassaemia were preserved at all dosage of the piretanide monosubstance. Mean serum potassium concentrations were significantly increased following treatment with the potassium-sparing agents triamterene (with and without piretanide) and amiloride in combination with hydrochlorothiazide. These changes were not accompanied by changes in mean serum magnesium levels, that is, a magnesium-sparing effect was not demonstrated. Individual values of changes in serum magnesium levels before and after diuretic therapy did not correlate with the corresponding changes in serum potassium levels in any of the eleven groups of patients. This result is in contrast to the findings of other authors, which suggested that serum magnesium and potassium levels were correlated. Long-term treatment with diuretics can lead to hypomagnesaemia and hypokalaemia, and, if in combination with a potassium-sparer, to hyperkalaemia. The results of this analysis show that piretanide, in addition to its smooth and effective antihypertensive action, is also able to preserve serum magnesium and potassium homeostasis when administered in the range of 3 to 12 mg for 12 weeks.