Hospitals in New York and Illinois have wide variations in their primary and repeat cesarean section rates. A number of factors account for these differences. To investigate whether hospitals with higher or lower rates tend to continue these patterns over time, their rates in 1988 were compared with those in 1983. It was found that a hospital's cesarean section rate was consistent, but some regression to the mean process did occur. By 1988, teaching hospitals had lower rates than nonteaching hospitals; this difference is likely due to the greater response to calls for increasing trials of vaginal birth after a previous cesarean section by teaching hospitals. Over time this should contribute to further moderating of the rates. Data were from the Illinois and New York State Departments of Health.