Adenovirus 2 does not produce tumors when injected into newborn hamsters, and this failure is not corrected by the acquisition of the portion of the SV40 genome present in any of the nondefective Ad2-SV40 hybrid viruses. However, nonhybrid Ad2 will transform hamster cells in vitro, and these transformed cells will produce tumors when injected into newborn hamsters. This finding, taken in conjunction with other studies (Graham et al., this volume; McAllister et al., 1969; McDougall et al., this volume; Williams 1973), suggests that the categorization of Ad2 and Ad5 as "nononcogenic" viruses is a reflection of host response rather than an intrinsic property of the viruses. However, transformation of hamster cells by Ad2 is inefficient, requiring 10(7)-10(8) PFU to produce a focus of transformed cells. Transformation of hamster kidney cells by the nondefective hybrids cannot be associated with portions of the SV40 genome. During the transformation process, complex interactions must occur between the hamster cell genome and the viral genome of Ad2 or the nondefective hybrids. Thus while it appears that there is very little Ad2 DNA in one nonhybrid Ad2-transformed cell line, there is a larger amount of Ad2 DNA in all of the hybrid virus-transformed cell lines. Moreover, while one Ad2+ND2-transformed line has apparently lost the SV40 portion of the viral genome, it is present and transcribed in another line transformed by the same virus. Finally, in the lines of nondefective hybrid-transformed cells examined thus far, there is extensive transcription from the Ad2 H strand. This pattern of transcription differs from transcription of the viral genome in most hybrid Ad2-transformed hamster and rat cells in which L-strand transcripion predominates.