Alternate changes of specific surface antigen(s) (S antigen) were examined in transformed and tumor cells induced by human adenovirus type 12. All of the hamster and mouse cells transformed in vitro showed ring-form membrane fluorescence staining by anti-S antigen rabbit sera, whereas tumor cells, either induced by the virus in vivo or produced by inoculation with the S(+)-transformed cells, did not show any fluorescence. When the S(-) tumor cells were serially subcultured in vitro, all of them converted to S(+) cells, although more than ten subcultures were necessary. For the S(+) cells to form tumors in hamsters about ten times as many cells were necessary as the S( -) cells. This difference became greater when tumor formation was tested in preimmunized hamsters, while little, if any, when tested in X-irradiated hamsters. In addition, immunogenicity of the S(+) cells was suggested to be higher than that of the S(-) cells. These findings indicate that the S(+) cells are more immunosensitive and immunogenic than the S(-) cells, and that in vivo conversion from S(-1-) to S(-) may be due to selection of S(-)-mutant cells. In vitro conversion from S(-) to S(+) was also suggested to be due to the appearance of S (+) -mutant cells. [1,27,31,33] and transplantation rejection [5,24,26]. One of them which was demonstrable by immunofluorescence staining was designated as tumor-specific surface antigen (S antigen) and the other which was demonstrable by transplantation immunity is designated as tumor-specific transplantation antigen (TS-TA). However, the exact relationship between S antigen and TSTA is not clear yet.As pointed out in our preceding paper [6], we have noticed that S antigen was detectable by immunofluorescence staining on hamster and mouse cells transformed in vitro by human adenovirus type 12 and maintained by subcultures in vitro, while it was not detectable on tumor cells induced in vivo by the same virus. Since these observations were made with a small number of cell lines, extended and systematic experiments were carried out in order to determine 1) whether or not it is a general phenomenon, and, if general, 2) whether it is possible or not to convert S antigen-positive (S(+) ) cells to S antigen-negative (S( -) ) cells and vice versa, and when possible, 3) what the mechanism of the antigen conversion is.