The relation between irregular and inconvenient working hours and the outcome of pregnancy was studied among women employed at a hospital in Sweden some time between 1980 and 1984. A questionnaire was distributed to 807 women; 81% replied. The pregnancies were divided into six groups with respect to work schedules during pregnancy. A slightly, but not significantly, increased risk of miscarriage was found in women who worked irregular hours or rotating shifts compared with women who worked only during the day (RR = 1.44, 95% confidence interval 0.83-2.51). Infants of non-smoking mothers who worked irregular hours had significantly lower birth weights than infants of non-smoking women working day time only. This difference was largest at birth order 2+. Similar results were found for infants of this birth order whose non-smoking mothers worked evenings or rotating shift.