A system of pneumographic study of infants considered at increased risk from the sudden infant death syndrome is described. It is simple for nursing staff to use and for clinicians to interpret. A total of 166 'at risk' infants were studied and 85 were subsequently monitored at home. Seventeen of these infants, 16 of whom had had abnormal pneumograms, subsequently suffered significant apnoea. Four of 20 babies who had abnormal pneumograms did not have subsequent episodes of apnoea. Two babies died; the first was on a monitor but the second, despite having had an abnormal pneumogram, was not.