followed discovery of a leak of toxic chemicals in November 1981. The point source was an underground waste solvent storage tank at a semiconductor plant, 0.7 kilometers from a well that supplied drinking water to nearby industrial and residential areas (Swan et al., 1989). The well was removed from service in December 1981 after contamination with solvents (predominantly 1,1,1-trichloroethane or methyl chloroform) was discovered. Because of the concern of residents, the California Department of Health Services conducted two studies. In one, there was an increased prevalence of major cardiac anomalies in children born to persons who lived in the service area of the water company that operated the contaminated well (Swan et al., 1989). A relative risk of 2.2 (95 percent CI 1.2-4.0) was determined for the contaminated area in comparison with the remainder of the county. No excess was observed for the period between September 1982 and December 1983. However, on the assumption that the greatest degree of chemical contamination of drinking water occurred in the three-month period before the leak was discovered, it was expected that the excess cases would be found in children born between May and August 1982. Cases were not found in excess of the expected rate. Therefore, it seemed that the chemical contamination of the drinking water could not explain the excess of cardiac anomalies in children. This conclusion was reinforced by a follow-up telephone interview of 145 mothers of children born with severe cardiac anomalies and 176 mothers of children without such anomalies (Shaw et al., 1990). However, longer-term analyses were not conducted. Although there was a positive association between a mother's consumption of tap water during the first trimester of pregnancy and cardiac anomalies in her infant for the year 1981 that was not present for 1982 or 1983, this appeared unrelated to the source of water and to the incident of contamination. Chlorination by-products should be considered further, given recent studies of TCE and cardiac anomalies discussed below. Consumption of bottled water appeared protective, however. The data could not be used to distinguish between a potential causal agent in the water and differential reporting of exposure by study subjects.
In the second study, a cluster of adverse pregnancy outcomes, including spontaneous abortions, low birth weight, and congenital malformations, was identified (Deane et al., 1989). An odds ratio of 2.3 for spontaneous abortion was found in comparison to a census tract free of contaminated water, while the odds ratio for congenital malformations was 3.1. In a subsequent study, Wrensch et al. (1990a) also investigated adverse pregnancy outcomes during 1980-1985 in two communities where solvents had leaked from the underground