A randomised double blind study was designed to evaluate haemodynamic response to dobutamine and dopamine in 20 hypotensive preterm infants of less than 32 weeks' gestation. Neonates initially received dopamine or dobutamine 5 micrograms/kg/min. If mean arterial pressure (MAP) remained below 31 mm Hg, the infusion rate was increased in increments of 5 micrograms/kg/min. If 20 micrograms/kg/min of the initial drug failed to achieve a MAP above 30 mm Hg, it was discontinued and the other drug was administered at the same infusion rate. Left ventricular output (LVO) was measured by pulsed Doppler echocardiography. Mean (SE) MAP increased significantly from 24.4 (1.0) to 32.0 (1.4) mm Hg at a median dobutamine dosage of 20 micrograms/kg/min and from 25.6 (1.2) to 37.7 (1.5) mm Hg at a median dopamine dosage of 12.5 micrograms/kg/min. The percentage LVO increase was +21 (7)% with dobutamine compared with -14 (8)% with dopamine. Dobutamine failed to increase MAP above 30 mm Hg in six infants out of 10, whereas dopamine succeeded in all 10 infants. Six switches from dobutamine to dopamine were thus performed, providing a rise in MAP (29.2 (0.5) to 41.2 (2.0) mm Hg) and drop in LVO (356 (40) to 263 (36) ml/kg/min). These data indicate that dopamine is more effective than dobutamine in raising and maintaining MAP above 30 mm Hg; however dopamine does not increase LVO.