It is essential to have regular, accurate blood pressure (BP) monitoring of sick preterm infants. Invasive direct arterial BP measurements are often recommended, but it was our clinical experience that such measurements may not be possible in all infants. We therefore assessed the proportion of infants receiving neonatal intensive care in whom reliable arterial BP measurements could be made, the accuracy of a non-invasive method (Doppler technique) and established a reference range of BP results related to postnatal age using this non-invasive technique. Arterial catheters were established in 44 infants (study population) of the 45 in whom access was attempted. Median gestational age was 27 weeks and birth weight 949 g. Over the 1st week, however, arterial catheters had to be removed because of complications or improving respiratory status. From those catheters that remained in situ, accurate measurements were not always possible because of damping; by day 7 only 22 catheters remained in situ and 14% of the arterial waveforms were damped. Doppler systolic BP measurements correlated well with the accurate (non-damped) arterial results (r = 0.96, P less than 0.01). Systolic blood pressure was measured on all 44 infants daily for the 1st week using the Doppler technique and increased linearly with increasing postnatal age (r = 0.92, P less than 0.01) from a mean of 41.7 mmHg on day 1 to 49.3 on day 7. We conclude that Doppler non-invasive BP monitoring is a useful method for regular monitoring of sick preterm neonates as it can be applied accurately, unlike direct arterial monitoring, to all patients.