The hypothesis that the use of saccharin may increase the risk of spontaneous abortion was tested in a case-control study. Women who had had spontaneous abortions were compared with women who had been delivered after 28 weeks' gestation. Women with a history of diabetes were excluded from the analysis. The use of sugar substitutes during pregnancy was similar in cases and controls: 5.5 per cent of cases and 5.8 per cent of controls reported using sugar substitutes. The power of the analysis to detect a twofold increase in the frequency of use of sugar substitutes in cases as compared to controls was approximately 81 per cent. The analysis effectively controlled for all known potentially confounding variables: that is, age at last menstrual period, history of previous spontaneous abortions, smoking during pregnancy, and usual weight were controlled in the analysis.