Because preliminary data suggested a relation between risk of spontaneous abortion and tapwater consumption during pregnancy, the California Department of Health Services included questions on prenatal water consumption in all reproductive studies conducted between 1982 and 1988. Results from four of these five retrospective data bases suggest that women abstaining from tapwater or drinking bottled water during the first trimester of pregnancy may be at reduced risk of spontaneous abortion. Fetal resorption frequencies seen in an accompanying toxicology study were consistent with these epidemiologic findings, although not conclusive. Tap and bottled water samples from these study areas were analyzed for agents that might account for these findings. Differences in trace element composition and biological activity were observed, but the reproductive significance of these differences is unknown. This paper presents an overview of these studies, which are presented in detail separately. Three alternative explanations for these findings--bias, chance, and causality--are reviewed.