During 1973/1978, more than 90% of the pregnant women in Malmö, Sweden, were examined by ultrasound to improve the early diagnosis of twin pregnancy. This screening programme detected 90% of the twin pregnancies delivered during this period, 105 were subjected to a preventive treatment with bed rest in hospital, usually from the 29th to the 37th week (mean: 55 days). This study evaluates the outcome of these pregnancies, compared with those not treated with bed rest and with all twin pairs born a decade earlier (1963/1965). All twins of this group born up to 1 May 1977 were also monitored by regular postpartum check-ups until at least 18 months of age. Perinatal mortality was reduced to 0.5% in the group with bed rest, but remained as high (8%) as ten years earlier in the group without bed rest. The bed rest group, compared with the group without bed rest and the group of twins born 1963/1965, also showed a significantly reduced frequency of deliveries before the 37th week; also, the incidence of twins born light-for-gestational age and with birth weight below 1500 g was reduced. The incidence of twins born with cerebral palsy, mental retardation, and late development was 3% in 1973/1977 and 8.1% in 1963/1965. The reduction of mortality and morbidity of twins paralleled the improved obstetric and neonatal care. The fact that perinatal mortality and preterm delivery were reduced only among twins subjected to special antenatal supervision suggests that large gains are to be made by early detection and antenatal hospitalization. The lower CNS morbidity provides evidence of the quality improvement among the twin survivers.