To find out if the use of steroids affected the incidence of infection in babies who were nursed in the neonatal intensive care unit for nine weeks or more, 24 preterm babies who had received a three weeks course of dexamethasone (0.6 mg/kg/day, reducing to 0.3 mg/kg/day after a week, and 0.15 mg/kg/day after two weeks) were compared with 18 preterm babies who had not been so treated. No differences were found in the incidence or pattern of septicaemia or other bacteriologically proved infections between the groups. Of 57 episodes of septicaemia, 44 (77%) were caused by coagulase negative staphylococci.