Fear of uterine rupture has led to the widespread practice of 'Once a section, always a section'. Between 1972 and 1982, 1498 patients with one or more previous cesarean sections were delivered at University College Hospital, Galway. Trial of labor was undertaken in 844 patients, while the remaining 654 patients underwent repeat elective section because they had two or more prior sections. Eight true ruptures and 22 scar dehiscences were found. Regional analgesia and oxytocin did not significantly affect the rate of true rupture. The mean parity with uterine rupture was five, and it occurred most frequently in the initial trial of labor. There were four perinatal deaths associated with true rupture. Failure to detect the already compromised fetus before labor and delivery, rather than the method of delivery, was responsible for fetal demise in some instances. Five true ruptures were found in the trial of labor group (i.e. a ratio of 1:169), with the loss of three babies. A further baby was stillborn in a mother who ruptured a classical scar before labor. There were no maternal deaths in trial-of-labor patients and one in the elective section group. Two patients with true rupture had their uterus repaired, and were subsequently delivered by section. Another two patients with bloodless dehiscence and no repair, had two subsequent elective repeat sections each, and the unrepaired scar dehiscence was not evident.