OBJECTIVE To assess the relationship between mid-pregnancy maternal serum zinc and copper concentrations and neural tube defects. METHODS A prospective case-control study during 1978-1988 within a statewide hospital-based neural tube defect screening program measuring maternal serum alpha-fetoprotein levels at mid-pregnancy. METHODS Cases were 69 women with fetuses with confirmed neural tube defects. Controls were 592 women with fetuses without neural tube defects who were individually matched to cases for hospital, calendar date of screening, age and parity; there was a variable control-to-case ratio. RESULTS For both unmatched and adjusted matched analyses, mean maternal serum zinc concentration was higher in cases than controls (P = 0.02 and P = 0.03, respectively). There were no case-control differences for serum copper concentrations. Conditional logistic regression analysis showed a (statistically non-significant) 50% increase in risk of neural tube defects in women whose serum zinc concentration was more than two standard deviations above the population mean. CONCLUSIONS Within the normal range of maternal serum zinc and copper concentrations there is no variation in risk of neural tube defects. However, women with very high serum zinc levels may have an increased risk of neural tube defects. This could reflect deficient maternal-to-fetal transfer of zinc in some of those individuals. Any such phenomenon would be manifest in observational, but not experimental, studies.