Two groups of patients suffering from extensive pulmonary tuberculosis treated with daily or twice-weekly regimens of isoniazid (INH) plus rifampicin (RMP) plus ethambutol (EMB) were formed by random selection. The effectiveness of treatment was expressed by the slope of regression line calculated for each patient in terms of quantitative time dependent decrease of mycobacteria isolated by culture per 1 ml of sputum. Chemically serum concentrations of drugs in five time intervals following simultaneous administration were established. It was shown that the rate of decrease of mycobacteria did not significantly vary in the compared groups. Negativity was reached in 48 day on the average. The mean regression line with double standard deviation allocated patients with rapid, normal or slow sputum conversion. Rapid convertors were the youngest, excreted largest amounts of mycobacteria before the start of treatment, and their x-ray showed predominantly changes of exudative character. Slow convertors excreted least amounts of mycobacteria. In the group of slow convetors with daily regimen was significantly higher number of rapid INH inactivators compared with groups of rapid and normal convertors. Analysis of multiple correlation and variance of the ratio of mycobacterial decrease rate and of the parameters of the biological availability of drugs as expressed by the area under the curve of drug serum levels reveal the rate of mycobacterial decrease to be dependent always on the drug out of the combination. In daily regimen the decrease rate was controlled by INH, in the intermittent regimen by RMP. The speed of negativization could therefore be increased in daily regimen by increased doses of INH, chiefly in rapid INH inactivators. The most important share of RMP in the mycobacterial decrease rate during intermittent administration can be judged from dependence of biological availability on the dose of RMP. Under the experimental circumstances of this study the pharmacological means for speeding up negativization were best utilized. Further shortening of time necessary for obtaining negativity would be practicable only when residual factors are involved presumably existing beyond the region of pathogen and drug relationship.