Systolic pulmonary arterial pressure was determined serially over the first 10 days of life in 33 babies with hyaline membrane disease by measuring the peak velocity of pansystolic tricuspid valve regurgitation, using Doppler ultrasound, and applying the Bernoulli equation. Results are presented in age groups 0-12, 13-36, 37-72, and 73-96 hours respectively. The incidence of tricuspid valve regurgitation was 92, 97, 80, and 64% (falling to 35% by day 10) compared with 53, 50, 31, and 0% in 17 healthy premature infants. In comparing healthy babies with those with hyaline membrane disease, no allowance was made for right atrial pressure. The derived 'right ventricle to right atrial (RV-RA) pressure difference', was expressed as a ratio of systemic arterial (systolic) pressure. Over the first three days, this ratio fell much faster in the healthy babies. Values were 0.78:1, 0.77:1, and 0.72:1 in babies with hyaline membrane disease and 0.87:1, 0.53:1, and 0.44:1 in healthy babies. Ductal patency was prolonged in babies with hyaline membrane disease (75% on day 4 compared with 6% in healthy babies). The incidence of bidirectional ductal flow, indicating balanced pulmonary and systemic arterial pressures, was 79, 53, 30, and 20%, and in healthy babies was 41% at 0-12 hours and zero thereafter. Pulmonary arterial pressure was then calculated by adding a right atrial pressure estimate of 5 mm Hg to the RV-RA difference when the babies were ventilated. Babies of lower gestation had lower values. The pulmonary: systemic arterial pressure ratio showed considerable temporal variability, but fell with age and was raised by high mean airway pressure and pneumothorax (through a reduction in systemic pressure), and less noticeably by carbon dioxide tension. It did not correlate significantly with other indices of disease severity. Hyaline membrane disease is associated with delayed postnatal circulatory adaptation characterized by pulmonary hypertension, systemic hypotension, and prolonged ductal patency.