Doppler ultrasound was used to assess the pressure drop between the ventricles in 109 infants and children (61 less than two years old) with a ventricular septal defect who underwent cardiac catheterisation. The pressure in both ventricles was measured at catheterisation in 103 patients either simultaneously through two catheters (41) or with a single catheter withdrawn across the septum or removed from one ventricle to the other (62). When pressure was measured simultaneously with two catheters (41 patients) the peak to peak and instantaneous gradients showed a maximum difference of 20 mm Hg with levels within 10 mm Hg of each other in 36. Comparison of the difference in the gradients with the average of the measurements demonstrated a tendency for Doppler to underestimate the difference when it was high (greater than 50 mm Hg) and overestimate it when it was low. A Doppler estimate of a low pressure difference between the ventricles indicates pulmonary arterial hypertension and a high one low pulmonary artery pressure, but in the intermediate group Doppler is as yet not sufficiently sensitive to allow selection of those patients who require further investigation and possible operation. Doppler ultrasound was found to be a sensitive method of detecting a very small ventricular septal defect. Thus although Doppler is a very useful means of assessing and following patients with a ventricular septal defect, further studies are required to determine its exact place in clinical practice.