unless certain conditions prevail. The organism must be introduced into the uterus, dead tissue must be present at the time, and the injured tissue must remain in the uterus for long enough to permit incubation. In our patient probably the organism was introduced into the uterus from bowel during the unsuccessful attempt at amiocentesis, because the high vaginal swab taken before the abortion was negative for clostridia. Cl welchii septicaemia, especially when haemolysis is present, is regarded as an indication for total abdominal hysterectomy. The improved survival rates of the early 1970s were attributed to this aggressive surgical approach.2 In the absence of haemolysis, however, prompt and adequate antibiotic treatment with immediate curettage will often effect a cure if coupled with good supportive care.3 We are grateful to Mr M E Pawson and Dr J Curtis for permission to publish this case report. Douglas GW. Toxic effects of Welch bacillus in post abortal infection. NY StateJa Med 1956;56:3673-7. 2 Decker WH, Hall W. Treatment of abortion infected with Clostridium welchii. AmJ Obstet Gynecol 1966;95:394-9. 3Pritchard JA, Whalley PJ. Abortion complicated by Clostridium perfringens infection.

Facilitated and improved by advances in molecular biology, techniques for the immunodiagnosis of schistosomiasis, including assays based on the detection of antigens circulating in the serum and/or excreted in the urine, have now reached the stage of multi-centre trials. There is a need to complement parasitological techniques as some national programmes are becoming increasingly successful in establishing control of the disease and the classical approach frequently fails to reveal low-intensity infection. Epidemiological survey teams in some areas have tentatively started to use serology and their experience indicates that antibody detection suffices in eradicated or controlled areas with low expected prevalence but that detection of circulating antigens is needed for assessment of the incidence of infection or reinfection in areas recently brought under control. Before reagents and procedures can be recommended for routine use of national control programmes, the assays must be standardized with sera from clinically well-characterized patients in geographically defined regions, hence emphasizing the need for a reference serum bank. Implementation of serological testing, carried out by national public health laboratories using standardized testing systems, would permit valid comparisons between different areas providing support for decisions regarding national health policies.