No attempt has previously been made to assess the morbidity of non-operative treatment of possible appendicitis. This paper uses the need for appendicectomy at a later date as the main criterion of morbidity. On the basis of a prospective study of 209 patients and a retrospective study of 1,284 patients, it appears that about 25 per cent of patients admitted to hospital with possible appendicitis but not operated on, require appendicectomy at a later date. In addition, only some 57 per cent of patients treated non-operatively do not again have pain. The presence of a past history of similar pain appears to prejudice unfavourably the results of non-operative treatment. Patients who return quickly to work and are not allowed to think that they have had appendicitis are the most likely to have a favourable prognosis.