This report is the last of a series of nine publications from the Tuberculosis Chemotherapy Centre, Madras, concerning various aspects of an investigation of the role of ambulatory chemotherapy for pulmonary tuberculosis. It presents the attack rates of tuberculosis over a 5-year period of follow-up of close family contacts of patients, all of whom were treated for one year with isoniazid plus PAS, half (selected at random) in sanatorium and half at home. The incidence of active tuberculosis and of tuberculous infections was no greater in the contacts of patients treated at home than in the contacts of patients treated in sanatorium, either in the first year or over the subsequent four years. The major risk to the contacts resulted from exposure to the patient before diagnosis. These findings reaffirm that close family contacts of patients treated at home were at no additional risk of developing tuberculosis, provided the patients received effective chemotherapy. Finally, this study has shown that it is possible in South India to obtain extremely good co-operation from a group of families over a period of several years.