A mixture of varicella-zoster virus-infected human embryonic lung fibroblasts and hamster embryo cells produced foci of morphologically transformed cells after several weeks of incubation. These transformed cells exhibited virus-specific antigens by immunofluorescence and developed surface Fc receptors. They induced aggressive fibrosarcomas when injected back into inbred hamsters. Cells derived from hamster tumor tissue exhibited similar properties. The tumor-bearing hamsters develop antibodies specific for varicella-zoster virus (VZV) antigens. Cell lines derived from both the original transformants and explanted hamster tumor tissue have varying growth properties. All maintain indefinite growth. All eventually lose varicella-zoster virus-specific immunofluorescence. None retain VZV-specific DNA sequences as determined by dot blot and Southern blot hybridization using radiolabeled whole VZV DNA and cloned VZV DNA fragments as a probe. A few transformed and tumor cell lines frozen relatively soon after isolation and stored in liquid nitrogen were also thawed, replicated to mass culture, and analyzed for VZV-specific DNA sequences. Hybridization with radiolabeled whole VZV DNA initially suggested that some of these cell lines did contain virus-specific DNA sequences. However, hybridization with cloned VZV DNA fragments radiolabeled in vitro and representing greater than 95% of the virus genome was negative. Karyotyping of these "positive" transformed cells indicated that they are of human and not hamster origin. The positive hybridization with whole VZV DNA therefore most likely represented contamination of the probe by host DNA sequences. The cells that survived freezing then were predominantly transformants of human origin. Attempts to repeat the transformation of hamster embryo and baby hamster kidney cells with laboratory-passaged VZV strains have been unsuccessful. Similarly, we have been unable to transform cells with whole VZV DNA or cloned VZV DNA fragments, although whole VZV DNA is demonstrably infectious. Apparently, transformation of cells by the varicella-zoster virus is a very rare event and one that may require a recent clinical isolate. Fresh clinical isolates of varicella-zoster virus are seemingly able to transform mammalian cells in vitro. The transformed cells have malignant properties and are capable of indefinite growth. Although VZV gene function can be detected early after transformation, there is no evidence that a VZV-specific protein product is required. Transformation of mammalian cells by varicella-zoster virus apparently occurs through a "hit-and-run" mechanism.