Water hardness can no longer be considered as the most reliable "water factor' with regard to the cardiovascular risk observed in epidemiologic studies. Only two out of three studies have shown a reverse correlation between cardiovascular mortality and water hardness. But studies carried out on the water Mg level alone, as opposed to those on water hardness (Ca + Mg) have all shown a reverse correlation between cardiovascular mortality and the Mg level. In developed countries, the Mg intake is often marginal and the Mg intake coming from drinking water represents the critical factor through which the Mg intake is deficient or satisfactory. Thus, Mg deficiency, either experimental or in man facilitates cardiovascular pathology. The importance of the Mg intake in drinking water is both quantitative and qualitative. Water containing Mg is better and more quickly absorbed than dietary Mg. This particular availability might help to understand why an adequate water Mg level may determine a better state of health, even without any Mg deficiency. Epidemiological data in man and experimental data in rats have demonstrated that the intake of water containing a sufficient amount of Mg may prevent arterial hypertension and correlated ionic and nervous disturbances. Indirectly the water Mg level also interferes in the leakage of food-borne Mg during cooking. There is an inverse correlation between the Mg loss in the cooked food and the Mg level of the cooking water itself. Mg appears to be an antagonist of noxious polluting agents (e.g. in the human amnion, Mg is a competitive inhibitor of Pb and Cd). It is not advisable to enrich water in Mg in the course of the processing since its corrosivity index would also increase. The best pathway is probably to neutralize corrosive water by filtration on calibrated grains of earth-alkaline metals (Neutralite or Magno or Akdolit) to ensure the highest possible Mg/Ca ratio, with the best anticorrosive power.