This prospective hospital-based study was done at Tikur Anbessa Teaching Hospital (TATH) between July 1991 and July 1992 in order to obtain base-line data on the rates of caesarean section, pregnancy out-come, major indications for caesarean and caesarean section complications. Out of a total of 3237 deliveries conducted during the study period, 318(10%) were caesarean section. The leading or the major indication for abdominal deliveries were: repeat caesarean section, 103(32.4%), cephalo-pelvic disproportion, 93(29.2%), placenta previa and abruptio-placentae 40(12.6%). Age ranged from 15-40 years, 58(18.2%) were women under the age of 20, and 182(57%) were between 20 and 30 years of age which is the safest period to bear children. Eighty one (25%) of the mothers were primiparae, 158(50%) were between para one and para four, 79(25%) were grand multiparae. Seventy six (24%) of the cases were not registered for antenatal care in any health institutions. Two hundred and fifteen (67.6%) of the mothers had primary caesarean section. The chance of repeat caesarean section was statistically significant (P < 0.001). Fifty seven (18%) had elective caesarean section and 261 (82%) were emergency caesarean section. On eight (2.5%) of the patient, obstetric hysterectomy was performed. There were five maternal deaths among the caesarean section cases. The main cause of maternal death was failure to control bleeding during the caesarean section. These were preventable deaths in experienced hands. There were nine (2.8%) stillbirths and 15(4.7%) early neonatal deaths. A foetal wastage of 24(7.5%) was noted in the study. The mean birthweight among booked caesarean section were 3108 grams and unbooked caesarean section were 2991 grams. The major obstetric complications in the cases were antepartum haemorrhage, and 34(11%), pregnancy induced-hypertensive disorder, 32(10%). Sixty one (19%) had blood transfusion. The blood loss was significantly more in operations done by residents compared to consultants even using the most crude measurement as estimated by the operator at the time.