Ever since Kobayashi,'in 1957, noted a parallel between the geographic distribution of the acidity of water in Japanese rivers and the distribution of what was then one of the major causes of mortality in Japan, apoplexy, an increasing number of investigators all over the world have attempted to elucidate and confirm a geographic relationship between quality of drinking water and mortality, particularly from cardiovascular causes. The now voluminous literature in this field has been subject to several comprehensive As remarkable as the geographic diversity of these studies is the great diversity of the hypotheses that have been favored by daerent investigators, both as regards the identity of the water-borne factor to which they impute a good or bad influence, and as regards the nature of the disease or pathologic process induced. Re~ ently,~ we expressed regret at the lack of any emerging consensus among those who have contributed to the literature over the last 15 years, and at the failure of most studies to yield evidence capable of discriminating among any of 64 major classes of explanatory hypotheses that need to be considered. Kobayashi did not refer to any category other than apoplexy, but ecological studies, from Schroeder on, have usually tried several cause-specific death rates as dependent variables.

Thus, even though most investigators continue to refer to the hazard of residing in soft-water areas as if it related specifically to cardiovascular disease, support for this view has been progressively weakened by the admission of a bronchitis effect lo and by the suggestion that there is also an infant mortality effect in England, Wales 11, l2 and possibly Canada, 13 though not in the United States. 14 Stocks l5 showed that whatever may be the environmental or genetic factor responsible for regional mortality patterns of cardiovascular disease in the United Kingdom, it would appear to have a pronounced effect also on the incidence of congenital malformation of the central nervous system, and to a lesser extent this would apply to nephritis and carcinoma of the stomach. In the United States, Sauer l6 found as strong a negative correlation of hardness with malignant neoplasm as with cardiovascular deaths, Canadian statistics l7 show that more than half the excess mortality in soft-water areas is certified to noncardiovascular causes of death.