Four hundred and seventy height Culdocenteses were carried out in cases of possible ectopic pregnancy between the 20th September 1986 and 31st December 1987. Culdocentesis was considered to be positive if 2 cm3 or more of dark non-coagulated blood was removed, and negative if only a yellow liquid or blood stained serum was removed. It was not conclusive if nothing could be aspirated or if the blood was coagulated. Of the 94 cases where culdocentesis was positive, 74 were found to have an ectopic pregnancy. There were 20 false positive cases (due to 5 haemorrhagic ruptures of follicles, 3 refluxes of menstrual blood, 2 with other aetiology, and 10 without known cause). There were 21 cases of ectopic pregnancy in the 293 cases where culdocentesis was non-conclusive. Of the 91 cases where culdocentesis was negative, a second culdocentesis showed an ectopic pregnancy. It was positive 11 days after the first. In our series this diagnostic test was reliable in 77.1% of cases. Laparotomy was carried out in 22.3% of cases and only 18.6% had to have laparoscopy thanks to the use of culdocentesis.