Examination of a linked birth and death certificate file from Missouri revealed a remarkable increase in the use of cesarean section for very low—birth weight infants (500 to 1499 g) from 24% to 44% in the years 1980 to 1984. In the same years the rate for 1500- to 2499-g infants went from 21% to 26% and the rate for 2500- to 7000-g infants went from 14% to 18%. We studied first-day death rates in very low—birth weight infants as an indicator of potential benefit from this increase in cesarean sections. During the 5-year period first-day deaths averaged 10% and 22% in the cesarean and vaginal birth groups, respectively. This difference was explained almost entirely by deaths in the 500- to 740-g birth weight group, where the death rates were 33% and 59%, respectively. Although this difference remained statistically significant after adjustment for gestational age and other factors that differed between the groups, it was nullified by an excess of deaths in the succeeding 6 days of life. Overall, the odds of death in the first week in these infants weighing 500 to 749 g was 0.85 (95% confidence interval, 0.52 to 1.39) in the cesarean vs vaginal deliveries. We conclude that there is little evidence that the use of cesarean section for the delivery of very low—birth weight infants, independent of maternal or fetal compromise, improves overall survival. We were unable to find reasons to justify the sharp increase in the use of cesarean sections for these small infants.