2. The rapid development of resistance to 1000 units of streptomycin per cc. during treatment argues against metabolic adaptation of the bacteria.

3. Careful study of large samples of initial bacterial populations shows resistant members in all ten strains examined. The incidence varies from 1 in 1.1 billion to 1 in 13.8 billion organisms.

4. The proportion of resistant members is not significantly higher, before treatment, in patients who do poorly than in those who respond promptly.

5. The results indicate that emergence of resistance is the result of a selective process; after elimination of sensitive members the very few resistant ones, apparently present in all strains, make up most or all of the population in the cultures taken after commencement of treatment.

6. The survival of organisms which can grow in high concentrations of streptomycin, either in patients or in vitro, is influenced more by the size of the bacterial population than by any other known factor.

7. Since the bacterial population is relatively small in those patients with mild or moderately severe infections the likelihood of a significant number of very resistant members being present is remote.

8. The traits responsible for resistance of the organisms are apparently inherited: (a) The resistant state of one strain has been transmitted in vitro without change in degree through over one hundred subcultures in the absence of streptomycin, (b) Persistence of resistant organisms in the nasopharynx of one patient during a one year period raises an important public health problem.