This study confirms the findings of Douglas (1975) that single hospital admissions of children for up to a week carry no increased risk of later emotional or behavioural disturbance. The study also confirms Douglas' finding that repeated hospital admissions are significantly associated with disturbance in later childhood. The association probably applies to both emotional and conduct disorders, is most marked in the case of children from disadvantaged homes, and it may well reflect a causal influence. The validity of the finding is strengthened by the fact that detailed psychiatric assessments gave rise to much the same findings as did teacher questionnaire scores. However, in both cases, repeated hospital admissions are associated with only a small minority of disorers and account for little of the variance in children's behaviour.