抄録
It is generally believed that most strains of the tubercle bacillus are a mixed population consisting of bacterial cells of various grade of pathogenicity. Increase of virulence by animal passage and its decrease by culture medium transfer are therefore explained by the selective operation of those environments upon the bacterial population in the manner that animal body acts to select virulent cells and culture medium transfer has a tendency to work in the reverse way. BCG and H37Ra strain have often been cited as an evidence to prove that an avirulent mutant can be selected on specially devised culture medium (Dubos, 1949) .
However, it is an understanding of the present authors that the above view requires more firm experimental prooves before it is accepted as an unanimous truth. Because, such discussion hitherto made on bacterial dissociation of the tubercle bacillus has mostly based upon empirical facts or upon speculations from other species of bacteria and not upon convincible data obtained from exactly planned experiments of the tubercle bacillus itself.
As reported in the preceding paper (Katsuyama and Kanai, 1958), the authors succeeded in single cell cultivation of tubercle bacilli using a newly devised method. Ever since, an attempt has been undertaken by them to analyse the virulence of a strain of human tubercle bacilli (H2) applying that method, principally aiming at isolation of an avirulent variant. The result will be presented in this paper. As far as the present paper is concerned, the definition of virulence and pathogenicity is that the latter is the qualitative and the former is the quantitative concept both concerning the tuberculous lesion-producing capacity of tubercle bacilli in guinea pigs. This way of thinking is basing upon the suggestion of Ushiba (1955) . In this sense, the virulence of tubercle bacilli can be also interpreted as the degree of their multiplying ability in guinea pigs, as formerly pointed out by Rich (1944) .
