Neonatal mortality is examined by place and circumstances of delivery in North Carolina during 1974 through 1976 with attention given to home delivery. Planned home deliveries by lay-midwives resulted in three neonatal deaths per 1,000 live births; planned home deliveries without a lay-midwife, 30 neonatal deaths per 1,000 live births; and unplanned home deliveries, 120 neonatal deaths per 1,000 live births. The women whose babies were delivered by lay-midwives were screened in county health departments and found to be medically at low risk of complication, despite having demographic characteristics associated with high-risk of neonatal mortality. Conversely, the women delivered at home without known prenatal screening or a trained attendant had low-risk demographic characteristics but experienced a high rate of neonatal mortality. Planning, prenatal screening, and attendanttraining were important in differentiating the risk of neonatal mortality in this uncontrolled, observational study.