EDWARD C. ROSENOW 577 



to 1914 in which mutations of pneumococci into streptococci, and vice versa, were 

 induced.' It was noted then that coincidental with changes in morphological and 

 cultural characters marked changes occurred in the localizing power which was often 

 characteristic of the type. The changes induced experimentally were shown to be 

 complete in the case of a number of representative strains of the group. The new 

 strains corresponded to those obtained from the usual sources, morphologically and 

 culturally, in the Marmorek test, in dextrose-serum broth, in fermentative powers, 

 in bile solubility, in autolysis in sodium chloride solution, in agglutinating properties, 

 and in specific antigenic power. The production of marked changes in cultural char- 

 acter as well as in localizing or infecting power of members of this group has recently 

 been reported by Morgenroth, Schnitzer, and Berger.^ I fulfilled the pure-line re- 

 quirement by obtaining single-cell cultures of different members of the group. 



In the work of Morgenroth, Schnitzer, and Berger, as in mine, not all strains 

 yielded to attempts to produce mutations or "dissociation";^ a premutational stage, 

 therefore, seemed necessary. The changes which I induced often had the character- 

 istics of true mutations because they appeared suddenly under conditions more or 

 less obscure and the newly acquired properties persisted unless the organisms were 

 again placed under special conditions. In my in vitro experiments, the underlying con- 

 ditions which tended most to produce changes were, first, favorable media for luxuri- 

 ant growth and then unfavorable conditions, under stress or strain. Mutations in the 

 animal body have been observed almost exclusively in closed cavities such as joints, 

 peritoneum, pleura, or pericardium, and here, as in vitro, under conditions of stress or 

 when the tissues of the host were gradually getting the upper hand and the organisms 

 were being destroyed. The following observations made during my studies in trans- 

 mutation had particular bearing on the fundamental questions of elective locahzation, 

 which have been studied extensively since. 



It was found that as non-virulent strains became virulent by successive passage 

 through animals, and highly virulent strains became less virulent by cultivation on 

 artificial media, the site of localization with production of lesions changed markedly. 

 When the virulence was lowest the localization was almost wholly in relatively avas- 

 cular tissue, such as heart valves, joint structures, and tendinous ends of muscles; 

 when the virulence was moderate, iritis, myositis, ulcer of the stomach, cholecystitis, 

 and focal lesions of the kidneys were more prone to develop; and when it was high 

 from successive passage through animals or on isolation, lesions of the lung and death 

 from bacteremia occurred commonly. Striking as these results were, it was not until 

 the unusual localization in the mucous membrane of the stomach of strains from sev- 

 eral sources occurred, producing hemorrhage and ulcer, that the idea of elective local- 

 ization took definite form. 



Since variations in oxygen tension and salt concentration, growth in symbiosis with 

 other bacteria, and injection into closed cavities in animals commonly called forth 

 mutational forms in pneumococci and streptococci, it seemed highly probable that 



• Rosenow, E. C: ./. Inferl. Dis., 14, i. 1914. 



^Morgenroth, J., Schnitzer, R., and Berger, E.: Ztsc/ir. f. Immiinitdlsforsch. 11. exper. Therap., 

 43i 169, 209. 1925. 



5 Hadley, P.: /. Infect. Dis., 40, i. 1927. 



