A cross-sectional survey of 701 Indian children aged 1-15 in five northern Canadian communities showed that 82% had a documented history of BCG vaccination while only 78% had a visible scar. The prevalence of 5 mm or more induration on Mantoux testing varied from 6 to 26%. Multiple logistic regression indicated that community, age, past use of isoniazid and time since the last Mantoux test were positive predictors of sensitivity. The presence of a BCG scar and the number of past vaccinations were not significantly associated but the elapsed time since the last BCG was negatively associated with a positive reaction. Overcrowding, nutritional status and past history of household and community contact with an active case of tuberculosis were also investigated but their independent effect could not be demonstrated. Despite the presence of mass BCG vaccination in this population, tuberculin testing is still a useful tool in assessing the risk of infection.