Active and passive surveillance methods were utilized in an attempt to identify all neural tube defect (NTD)-affected pregnancies in South Carolina, a state previously identified to have a high prevalence of these defects (Greenberg et al., 1983). Sources of case ascertainment included maternal serum alpha-fetoprotein (MSAFP) programmes, prenatal diagnosis (amniocentesis) programmes, physician offices, perinatal centres, hospital medical record departments, and vital records. One hundred and nine NTD cases were identified over 24 months, indicating a prevalence of 1.60 cases per 1000 recorded births and fetal deaths (surveillance 1 October 1992-30 September 1994). Fifty-three (49 per cent) of the 109 NTD-affected pregnancies were either spontaneously or electively aborted prior to 26 weeks' gestation. Only three (6 per cent) of these early termination NTD cases were recorded in vital records, while medical records recorded 40 cases (75 per cent). By monitoring MSAFP programmes and maintaining frequent contact with physician offices and perinatal centres, 85 per cent of these early termination NTD-affected pregnancies were identified. However, for complete ascertainment of NTD-affected pregnancies, the utilization of all of these ascertainment methods was necessary, as each ascertainment method identified NTD cases missed by the other methods. Consideration of the significant impact of prenatal diagnosis on NTD surveillance and the use of multiple ascertainment methods in an attempt at complete ascertainment is particularly important now because of the recent recommendation for the periconceptional use of folic acid to prevent neural tube defects and the need for epidemiological studies to monitor the effectiveness of this prevention.