A preventive effect of dietary marine n-3 fatty acids on early delivery and toxaemia has recently been hypothesized. In only one published controlled trial fish oil has been given to pregnant women, namely in that conducted during 1938-9 in London by the People's League of Health with a dietary supplement containing vitamins, minerals, and halibut liver oil. Although it was of high quality and its findings are hitherto unexplained, neglect and misinterpretation of the trial seem to occur commonly in reviews. Of the 5644 women who were enrolled the 622 withdrawals were independent of treatment. Alternate allocation to treatment was used, producing two groups that were well balanced as to age and parity. The supplement was given from about week 20. The control group did not receive any supplement. Reductions of 20.4% (95% confidence interval 9-30%, P = 0.00083) and 31.5% (95% confidence interval 11-47%, P = 0.0047) were seen in odds of delivering before 40 weeks of gestation and pre-eclampsia respectively. No significant effects were seen on perinatal mortality, average birth weight, deliveries after 40 weeks, hypertension in the absence of oedema and proteinuria, duration of labour, sepsis or breast-feeding occurrence. Later controlled trials with vitamins or minerals given in the same amounts as in this trial have largely failed to show convincing effects as seen here. A controlled trial assessing the isolated effects of fish oil in pregnancy is warranted.