A national probability survey was conducted to evaluate changes since 1979 in cesarean delivery and trial of labor rates after a previous cesarean delivery. Questionnaires were mailed to 538 hospitals and 87% responded. In 1979, of women with a prior cesarean birth 2.1% (+/- 0.3%) were given a trial of labor. By 1984, the rate increased fourfold to 8.0% (+/- 0.7%). Trials of labor increased dramatically with the size of the delivery service. The rates ranged from 2% in smaller hospitals to 25% in larger hospitals. More than 50% of the trials of labor were successful. However, the fraction of hospitals with no trials of labor remains high (54%). Cesarean birth rates increased from 14.1% (+/- 0.1%) in 1979 to 19.0% (+/- 0.2) in 1984. Hospitals with fewer births had a larger variation in cesarean birth rates; the fifth and 95th percentiles were 5% and 33%, respectively, for hospitals with fewer than 500 deliveries per year, and 14% and 31% for hospitals with 1000 deliveries or more per year. Fetal distress accounted for a larger proportion of primary cesarean deliveries in 1984 (21%) compared with 1979 (14%). The observed increase in the rate of trial of labor does not seem to be large enough to stem the rising cesarean delivery rate. This survey and other national surveys provide no evidence that the cesarean delivery rates are leveling off or decreasing.