The United States Public Health Service has undertaken a series of controlled trials in different situations to explore the possibilities of prophylactic isoniazid. A trial among household contacts of newly reported cases of active tuberculosis was conducted in cooperation with 39 health departments. A total of 25,512 contacts in 6219 households were examined soon after diagnosis of an index case for each household. On the first examination, 479 contacts had active tuberculosis 343 children with primary tuberculosis and 136 adults with pulmonary disease, a rate of 19 new cases per 1000 contacts examined. The remaining 25,033 contacts were enrolled in the trial. Half of the households were randomly assigned isoniazid and half, placebo. They were asked to take the unidentified pills every day for a year, isoniazid in an average dose of 5 mg/kg. The composition of the 2 groups as to race, sex, tuberculin sensitivity, and other characteristics remained similar throughout the year of medication. Indirect measures of pill taking indicate that approximately two-thirds of both groups took their pills quite regularly for the entire year, while less than 10% took pills for only a short time or very irregularly. Isoniazid markedly reduced the incidence of tuberculosis among contacts while they were taking pills. It was successful in preventing primary tuberculosis among initially uninfected children, with 16 cases in the placebo group and 5 in the isoniazid group and in preventing extra-pulmonary forms of tuberculosis, with 16 cases in the olacebo group and 4 in the isoniazid group. Adult pulmonary tuberculosis appeared in 62 persons in the placebo group and in only 14 in the isoniazid group. The hope had been that a course of isoniazid would also reduce the risk of disease in subsequent years but after the medication year, in the placebo group there were 6 cases of primary disease, 2 extra-pulmonary, and 17 pulmonary, compared with 6 primary, 1 extra-pulmonary, and 10 pulmonary cases in the isoniazid group. Further observation is needed. The trial has shown that, in the year following the diagnosis of a case, other members of the family are at a very high risk of developing tuberculosis. The importance of prompt examination of all household associates is emphasized by the large number of active cases found on the initial examination. Isoniazid prophylaxis is an effective and practical measure to add to existing programs of contact supervision.