STUDY OBJECTIVE--To estimate the prevalence of active pulmonary tuberculosis in a homeless population in London and to assess whether those with suspected disease could be integrated into the existing health care system for further follow up and treatment. DESIGN--Voluntary screening programme based on a questionnaire survey and chest x ray. SETTING AND CASES--Screening programmes were set up over the Christmas period in 1992 and 1993 at a shelter for the homeless in London. An offer of screening was made to all individuals who visited the centre and an interviewer administered questionnaire was completed on those who volunteered for the screening. Chest x rays were carried out, developed, and read on site. Individuals with chest x rays features suggestive of tuberculosis or other medical problems were referred to a hospital of their choice. RESULTS AND OUTCOME--In 1992 nearly 1600 people visited the centre, of whom 372 volunteered for the screening and 342 were x rayed. Nineteen of the 342 (5.6%) had radiological features suggestive of active tuberculosis. In 1993 around 2000 homeless people visited the centre, of whom 270 volunteered for the screening and 253 were x rayed. Eleven (4.3%) had features consistent with active tuberculosis on the basis of the chest x rays and clinical examination by a chest physician. Overall, of 595 people x rayed in the two surveys, 30 (5%) had changes suggestive of active tuberculosis. Further investigations confirmed nine (1.5%) with active pulmonary disease and eight with no active tuberculosis. In 13, the diagnosis was not determined as four declined further investigation and nine did not attend their hospital appointment. CONCLUSION--Tuberculosis among the homeless remains a cause for concern. Follow up and treatment present unique difficulties. Services for the homeless need to include mechanisms for timely diagnosis and monitored treatment. Control programmes designed for the needs of the homeless are required.