OBJECTIVE--To measure the prevalence of gonorrhoea in Ethiopian women attending gynaecologic, obstetric and family planning clinics: to determine the reliability of patient self history of sexually transmitted disease (STD); to correlate the serological diagnosis of gonorrhoea with clinical evidence of pelvic infection in order to define a reliable clinical diagnosis of gonorrhoea in a country where pelvic inflammatory disease is very common but where routine laboratory culture and serological tests for gonorrhoea are unavailable. SUBJECTS--1851 Ethiopian women: 50% symptomatic, 50% asymptomatic. SETTING--Gynaecological outpatient department, antenatal, postnatal and family planning clinics (Ethiopian Family Guidance Association (EFGA)), in two teaching hospitals and a mother and child health centre in Addis Ababa, Ethiopia. METHODS--The indirect haemagglutination test with gonococcal pilus antigen as an epidemiological tool was used in a cross-section study to screen 1851 sera for evidence of past or current gonococcal infection. The gonococcal antibody test (GAT) seropositivity was correlated with patient's history of STD, age, clinic attended and the clinical evidence of infection in "gonococcal target organs" urethra, salpinges or Bartholin glands. RESULTS--Fifty nine per cent of the study group were seropositive for the gonococcal antibody test, 22% with titres greater than or equal to 1/320, indicative of current, recent or recurrent infection. Seropositivity indicating past or present gonococcal infection was highest in those who gave a history of having had treated syphilis (85%), in women aged 40-49 (72%), and family planning attenders (EFGA) (66%) of whom 31% had titres greater than or equal to 1/320. Fifty per cent had clinical evidence of past or present infection in the urethra, salpinges or Bartholin glands. Gonococcal antibodies were present in 54% of women with no evidence of clinical infection, compared with 91% of those with pyosalpinx and 86% of those with triple infection of urethra, salpinges and Bartholin glands. CONCLUSION--The high prevalence of gonococcal antibodies in Ethiopian women, especially in asymptomatic clinic attenders must be of concern for all health workers especially those in gynaecology and obstetrics and the related disciplines of family planning and neonatal paediatrics. While seropositivity was highest in those giving a past history of syphilis, the patient's history of STD was unreliable, as of those who denied having any history of STD, fifty per cent were GAT seropositive. Despite a high correlation between GAT seropositivity with pyosalpinx and clinical evidence of infection in urethra, salpinges and bartholin glands, gonococcal antibodies were present in 54% of women with no clinical evidence of infection. Thus we were unable to define a diagnostic clinical picture of gonorrhoea in Ethiopian women.