Over a 5 year period 232 ectopic pregnancies were recorded at Ullevaal Hospital in Oslo, Norway. There had been 10,294 births during this July 1976 to June 1981 period. In 3 cases tubal sterilization had been performed prior to the development of the extrauterine pregnancy. A total of 1047 female tubal sterilizations were performed during these 5 years. Almost all the sterilizations were done by laparoscopy. Different methods of sterilization were used: unipolar diathermy; spring clips according to Hulka; silicone rings; and endotherm coagulation. Each case of the 3 ectopic pregnancies, observed following tubal sterilization, is reviewed. A 36 year old patient became pregnant 3 years after diathermy sterilization. The right tube was found normal, and the pregnancy located in the lateral part of the divided left tube. 14 months following silicone ring sterilization a 26 year old patient had an ectopic pregnancy in the lateral right tube. The silicone ring was in perfect position on the left side. The medial right tube showed fibrous scarring after the ring application, but the ring was located in the mesosalpinx. A 37 year old patient was admitted to the hospital after a tubal pregnancy 10 months after diathermy sterilization. The pregnancy was in the lateral part of the tube. Both tubes had been transected, and there was a diastase of about 2 centimeters. The etiology of ectopic pregnancies is complex. It is only recently that previous sterilization has been recognized as a factor in this condition. Luteal phase pregnancies are because of a failure in the timing of the procedure and are unrelated to the procedure itself. Pregnancies resulting from operative failure range from 0-2.4/1000 sterilizations in different series. Technical failure may be caused by recanalization, fistula formation, and product failure. An important cause of ectopic pregnancy after laparoscopic sterilization is probably fistula formation that allows sperm to pass out of the uterus. Such fistulas have frequently been documented. Thus, if it is considered necessary to confirm tubal occlusion by salpingogram, this should be delayed for at least 12 weeks. As fistula formation probably is a major cause in this entity, division of the tube at the time of sterilization is neither necessary nor desirable. It is important to damage the tube in the isthmic segment and to minimize the involvement of the parametrium in the destructive process.