The outcome of new cases of psychiatric illness identified by general practitioners in a prevalence survey of 46 practices in London, was studied over a period of three years. Information on 422 patients was obtained by examination of the N.H.S medical records, interview with general practitioners and postal questionnaire. Conclusions about outcome based on these data ware validated by applying a number of screening tests and by clinical interview of a one-in-four sample of patients. Three years after index consultation 73 per cent of all new cases were regarded as free from psychiatric symptoms and the progress made by these patients at one year had indicated probable outcome at three years. Duration of illness was demonstrated to be an important prognostic factor when comparisons were made with chronic patients, and there is evidence that psychiatric referral should be considered when symptoms have persisted for one year. Long-standing social difficulties, evident when patients were interviewed, suggest that social agencies could be used to greater advantage in the management of chronic psychiatric patients in general practice. © 1969 Springer-Verlag.