A case-control study, carried out in the Mount Gambier region of South Australia, investigated the relationship between mothers' antenatal drinking water source and malformations in offspring. It was prompted by earlier descriptive findings of a statistically significant, and localized, increase in the perinatal mortality rate in Mount Gambier, due principally to congenital malformations affecting the central nervous system and multiple organ systems. Available for statistical analysis were 218 case-control pairs, from the period 1951-1979, individually matched by hospital, maternal age (+/- 2 years), parity and date of birth (+/- 1 month). Compared with women who drank only rainwater during their pregnancy (relative risk (RR) = 1.0), women who consumed principally groundwater had a statistically significant increase in risk of bearing a malformed child (RR = 2.8). Statistically significant risk increases occurred specifically for malformations of the central nervous system and musculoskeletal system. Reanalysis of the data by estimated water nitrate concentration demonstrated a nearly threefold increase in risk for women who drank water containing 5-15 ppm of nitrate, and a fourfold increase in risk for those consuming greater than 15 ppm of nitrate. A seasonal gradient in risk was evident among groundwater consumers, ranging from 0.9 for babies conceived in winter, 3.0 in autumn, to 7.0 and 6.3 for spring and summer conceptions, respectively. Linear logistic regression analysis, controlling for risk factors not accounted for in the study design, showed that maternal water supply, infant's sex, and mother's area of residence all contributed significantly to the risk of malformation. These results are discussed in relation to previous experimental and human descriptive studies, suggesting a plausible mechanism for nitrate-induced teratogenesis.