The author performed both in vitro and in vivo experiments of double resistant tubercle bacilli with the purpose to determine whether SM or INAH should be used at first to obtain a better result. As the result the following findings were obtained. 1) With the use of H(37)Rv parent strain, SM-resistant H(37)Rv-strain, resistance increment experiments were performed in Kirchner's serum liquid media. In the successive cultures by the tenth successive culture both H(37)Rv parent strain and SM-resistant strain attained the resistance against INAH up to 100 γ/cc., while the parent strain and INAH-resistant strain acquired the resistance against SM up to 10,000 γ/cc. by the tenth successive culture. In other words, in the acquiring of resistance there is a significant difference between SM and INAH. 2) In comparing the resistance increment during ten successive cultures, SM-resistant strain acquired the resistance against INAH later than the parent strain. INAH-resstant strain likewise was slower in acquiring the resistance against SM than the parent strain. Because no difference can be found between the two in vitro experiments, it can not be said that clinically INAH should be better used prior to the use of SM. 3) In the experimental tuberculosis of mice, a fairly high resistance has been acquired by SM injection or INAH injection. 4) In the experiments with mice, the acquiring of resistance against SM in the cases inoculated with INAH-resistant strain is extremely slower than the acquiring of resistanec against INAH in the cases inoculated with SM-resistant strain. Therefore, judging from the experiments with mice, it may be said that clinically INAH would be better used first.