In South Carolina over a nine-year period, the adjusted odds that infants who died of injuries had been of low birth weight were 2.65:1 as compared to a live-born control group. The odds that injury victims were black were 1.84:1 as compared to control subjects. For victims of accidents only, the odds ratio for low birth weight was 2.37, and for black race, 1.61. For victims of homicide, the odds ratio for low birth weight was 6.15, and for black race, 4.76. Male sex, high birth order, low Apgar score, and the presence of a congenital anomaly were not significant infant risk factors by multifactorial analyses.