Data on 56,012 current and 48,608 previous pregnancies were obtained by interview in 11 Montreal hospitals, 1982 to 1984, after delivery or spontaneous abortion. In 17,632 pregnancies in occupations with substantial use of visual display units (VDUs), users and nonusers had similar rates of congenital defects in both current and previous pregnancies and of abortions in previous pregnancies. In current pregnancies there was an excess of abortions in users which could have been due to biased recall. In a further analysis of all 42 occupational groups ranked according to percentage use of VDUs, the risk of spontaneous abortion in both current and previous pregnancies was the same irrespective of the amount of VDU use. Thus, the study does not support the suggestion that work with a VDU in pregnancy increases the risk of congenital defect or spontaneous abortion.