We have tested the hypothesis that recurrent respiratory symptoms in the third year of life in patients born prematurely were more likely to reflect a family history of atopy rather than adverse neonatal events. Comparison of 28 symptomatic and 72 asymptomatic children revealed that a family history of atopy (p < 0.01), prolonged dependence on respiratory support in the neonatal period (p < 0.01) and extreme immaturity (p < 0.02) were significantly commoner in the symptomatic group. The relative risk of having symptoms was 2.27 for a family history of atopy, 2.48 for prolonged dependence on respiratory support and 1.7 for low gestational age. We conclude that respiratory morbidity in the third year of life following premature delivery has a multifactorial aetiology.