We have studied the cesarean section frequencies and changing spectrum of indications in Norway during six consecutive 3-year periods from 1967 to 1984. The data set consists of 1,046,162 births notified to the Medical birth registry of Norway, of which 52,426 were specified as cesarean sections, the frequency rising from 1.9% in 1967–69 to 9.7% in 1982–84. A check for completeness of cesarean section notification was made against the birth protocols of the largest obstetrical unit in Norway. The error rate was about 3%. Information on the indication for performing cesarean section was missing in 11.4% of the cases. We grouped thirty-one specified indications under seven headings. In 1967–69 the operation was most often performed for ‘maternal’ reasons, followed by ‘mechanical’, ‘abnormalities of fetal presentation’ and ‘acute placental’ conditions. Throughout the study period, ‘mechanical’ (which can also be labelled ‘dystocia’) accounted for about one third of the indications for cesarean section, while ‘maternal’ (including high maternal age and pre-eclampsia) and ‘acute placental’ (placenta praevia and placental abruption) decreased in relative importance. ‘Fetal presentation’ (including twins) doubled its relative share, while increasing from 2.4 per 1000 births in 1967–69 to 24.8 per 1000 births in 1982–84. A steep rise in the group ‘fetal asphyxia’ corresponded to the period when the mess of electronic monitoring devices was introduced, in the early 1970s. ‘Fetal asphyxia’ had a higher relative share of the indications among para 0 mothers than in the total birth population.