As part of a community health programme, a population-based case-control study was undertaken to determine whether crowding and alcohol are risk factors for tuberculosis in Mamre. Forty notified cases of tuberculosis were compared with 84 randomly selected controls. Cases and controls were group-matched for age and sex. Seventy-six per cent of dwellings had a housing density ratio greater than 100%. There was no association between crowding and tuberculosis, but cases lived in houses that were significantly older than those of controls (odds ratio, 4,8, 95% confidence interval 1,5-15,6). There was an association between alcohol problems in the household and tuberculosis (odds ratio adjusted for employment status 2,2, 95% confidence interval 1,3-3,8). Further studies using refined methods are needed to evaluate the role of crowding in tuberculosis. The need for alcohol intervention campaigns in conjunction with tuberculosis control programmes directed at high-risk groups should be considered.