We report preliminary results of a prospective study conducted to prevent sudden death in asymptomatic infants. From 1977-1984, 3658 infants were studied polygraphically. There were 923 siblings of SIDS victims and 2735 infants with no personal of family history of SIDS. The infants were studied at 8 weeks of age. Polygraphic "risk factors" were defined by central apnoeas longer than 15 s; periodic breathing above 5% sleep time; or obstructive apnoeas above 3 s. In 937 infants "risk factors" were seen and a second study was requested 4 weeks later. Out of 891 infants re-studied at 12 weeks, 153 still presented some "risk factors" and were selected for a home monitoring programme; 150 families agreed to monitor their infants at home with a cardiorespiratory monitor with the alarms set at 20 s apnoea, and 50 beats per min bradycardia. Repeated alarms were reported for 97/150 (65%) infants; 48/150 (32%) infants were stimulated and 8/150 (5.3%) were resuscitated on at least one occasion. No death occurred during monitoring, which could be interrupted before the end of the first year of life in all infants. In the group of 3459 infants with normal results and not monitored, three siblings (0.35%) and one infant without history (0.04%) died of SIDS. Of the infants with abnormal polygraphic results, one sibling not returned for the second recording, and two out of three infants for whom the parents refused monitoring, died of SIDS. It is concluded that the programme, may prevent the death of some infants, but that the outcome of a child with normal results cannot be foreseen.