Two-hundred infants of <30 weeks gestational age were included in a randomized double-blind controlled trial to study the effect of thyroxine administration on neurodevelopmental outcome in very preterm children. The infants were given either a fixed dose of thyroxine (8 microg/kg birthweight/day) or placebo for the first 6 weeks of life. This paper evaluates the effect of thyroxine administration on behavioural outcome at the age of 2 years. More externalizing, especially destructive, behaviours were found in the group given thyroxine than in the placebo group. This difference was more pronounced in boys and in children born after 27 weeks' gestation. The thyroxine-treated children with behavioural problems had lower plasma-free thyroxine levels than the thyroxine-treated children without behavioural problems. This finding suggests that the presence of more behavioural problems in the group given thyroxine was not an immediate consequence of the treatment.