During a 19-month period, 95% of all pregnant women in the greater Helsinki area, Finland, entered a study to compare one-stage ultrasonography screening with selective screening according to antenatal hospital use, obstetric procedures, and fetal outcomes. Of 9310 women who entered the trial, 4691 were randomly allocated to ultrasound screening between the 16th and 20th gestational weeks and 4619 to follow-up only. Screened and control groups otherwise had the same antenatal care, which included ultrasonography according to usual practice. Screened women made fewer visits to the antenatal outpatient clinic than did women in the control group (2.3 vs 2.6). There were no differences in the number of labour inductions or mean birthweights in the two groups. Perinatal mortality was significantly lower in the screened than in the control group (4.6/1000 vs 9.0/1000); this 49.2% reduction was mainly due to improved early detection of major malformations which led to induced abortion. All twin pregnancies were detected before the 21st gestational week in the screening group compared with 76.3% in the control group; perinatal mortality in the small series of twins was 27.8/1000 vs 65.8/1000, respectively.