Ninety-five premature newborns who had hyaline membrane disease and were struggling against the ventilator were randomised to one of three treatment groups: morphine (group M), pancuronium (group P) or morphine with pancuronium (group M+P). The dose of morphine was 50 micrograms/kg per h but was increased to 100 micrograms/kg per h in group M infants if they continued to struggle. The dosage of pancuronium was 100 micrograms/kg given as required to inhibit spontaneous respiration. Plasma catecholamine levels were measured on entry and at 24 h. Blood pressure and ventilatory requirements were determined on entry and at 6 h. The clinical outcome of the infants was documented. Group M infants (n = 29) showed a significant reduction in noradrenaline levels (median change -2.2 nmols/l (range -47.2 to +7.2 nmols/l), although seven were withdrawn from this group because of failure to settle. Group P (n = 28) and group M+P (n = 38) showed no significant change in noradrenaline levels. Comparison between the groups showed that group M infants had a significant reduction in noradrenaline levels compared with group P. The immediate effects of treatment on blood pressure and ventilatory requirements were similar in the three groups. The clinical outcome did not differ for any of the measured parameters. When adequate sedation is achieved, morphine may reduce the stress of newborn intensive care.