A reverse relationship was found between the magnesium content of drinking water, ie. an adequate supply of highly available dietary magnesium, and the occurrence of preeclampsia. Where various parts of a settlement receive their drinking water from different sources (for example, surface water or well water) there may be an important difference in the level of magnesium supply to different population groups. In the population studied here, the occurrence of preeclampsia varied pari passu with the magnesium content of the drinking water in different parts of the city. The results show that during pregnancy the magnesium content of serum and erythrocytes decreases. This decrease may be prevented by magnesium therapy, which lowers the incidence of preeclampsia.