The risk factors for increasing cesarean section (CS) rates were studied in two population-based cohorts of singleton live births from families residing in the municipality of Ribeirão Preto, statè of São Paulo, southeast Brazil. The first comprises births from June 1978 to May 1979(6750 births - population survey) and the second births from May to August 1994 (2990 births - sample survey). The CS rate rose from 30.3% in 1978/79 to 51.1% in 1994. Multiple unconditional logistic regression modeling was used to control for confounding. In 1978/79 the following factors were associated with CS: four years of maternal schooling and over, delivery in specific hospitals, day of delivery, four prenatal visits and over, maternal age of 20 years and over, previous stillbirth, number of live births and birth weight of 3500 g. and over. In 1994, delivery in specific hospitals, day and time of delivery, attendance by the same physician in prenatal care and delivery, four prenatal visits and over, maternal age of 30 years and over, number of live births and birth weight from 3500 to 3999 g were found to be associated with CS. In 1978/79 socioeconomic, reproductive and demographic variables, and health service factors were associated with CS rate although in 1994 only reproductive, demographic and health services factors remained associated. This high rate indicates that cesarean section in Brazil is widely performed for non-medical reasons. There is an urgent need for public health interventions aimed at reducing the cesarean section rate in Brazil, mainly directed towards cultural beliefs and physician behavior.