The incidence and current management of red cell aplasia in children was determined from a retrospective survey of haematologists and paediatricians in the northern health region of England over a 7-year period. Thirty-three children were diagnosed: 4 had Diamond Blackfan anaemia, 22 transient erythroblastopenia of childhood, and 7 parvovirus B19 aplasia, with annual incidences of 1, 5 and 2 per 1,000,000 children respectively. Many were over-investigated. Three with Diamond Blackfan anaemia were steroid responsive. One with transient erythroblastopenia was retrospectively diagnosed because anaemia did not recur after steroids were stopped. Transient erythroblastopenia is the most common single cause of red cell aplasia in immunocompetent children. Time, observation and bone marrow examination before steroid therapy are the ways to distinguish transient erythroblastopenia from Diamond Blackfan anaemia or leukaemia. Interpretation of red cell indices using age-related percentiles may reduce the number of inappropriate investigations of the anaemia, but is often unhelpful in distinguishing the various causes of red cell aplasia.