A supernatant of bouillon cultures of S. miami contains a kind of phage. This is the phage which S. miami liberates and it is temporarily called λ phage. This phage is weak and is mostly inactivated by heating of 60°C for 30 minutes. This has closely serological relation with palos phage (this is the phage which S. paratyphi A komatsubara liberates) among the phages which we have studied in our laboratory. Futhermore, in the other characteristics its morphology of plaque and host-range are the same with the palos phage and only one step growth curve is different from the palos phage, and therefore we cannot say the two to be equal.On the other hand S. miami retains O [I] antigen. However, λ phage which is liberated from S. miami in the above mentioned method, is consequently of no use to convert O [I] antigen.Then, what controls O [I] antigen of S. miami? In case of superinfection of S. miami with s2 phage which S. typhimurium 4066 liberates by the same method as was followed on S. sendai at our laboratory, although it is a very small amount, the rate of 50 phages can be expelled from bacteria of 3×108/ml. This expelled phage was quite the same with that of p3 phage which is liberated from S. pullorum. Thus, as can be seen from this study, in respect of the phage, S. miami is the double lysogenic strain. Namely, it might be considered that λ phage existed as prophage in it and p3 phage attached to the bacterial chromosome of S. miami at the shape of “defective prophage”, and it proves that this p3 defective prophage controls O [I] antigen of S. miami. In this case also we did not use the word “defective lysogeny” but used the word “abnormal lysogeny” as in the case of S. sendai.