MYELOID Graffi leukaemia induced by a virus in mice contains both leukaemia virus and virus-induced antigens1,2. The presence of leukaemia virus can be demonstrated by cell-free transmission of the disease to new-borns and by the ability of the virus to induce the production of antibodies in adults (MAP test) against antigens appearing on the surface of leukaemia cells3,4. Specific antibodies have been shown by the cytcotoxic test in the presence of complement and by the indirect fluorescent–antibody technique using living cells as targets2. Virus-induced surface antigens of leukaemia cells can be detected by the same techniques. In vivo, syngeneic inocula of leukaemia cells are rejected in Graffi immune animals, which suggests the presence of specific surface antigens, too. For tumours induced by polyoma, SV 40, adenovirus, and Rous virus there is evidence that the specific surface antigens represent new cellular antigens unrelated to viral antigens5,6. This last problem, however, is still unsolved for virus-induced leukaemia of mice. Cells of primary and transplanted leukaemias are continuously producing virus by budding it from the cell surface. It is therefore conceivable that the virus-induced surface antigens consist of both viral and new cellular antigens. On the other hand, Landschutz sarcoma 1 cells, which naturally harbour the Graffi virus7, were found to be resistant to the specific rejection response in vivo and to cytotoxic antibodies in vitro8. These cells, however, reveal a weak but distinct punctiform membrane fluorescence in the indirect fluorescent antibody test with Graffi hyperimmune sera of mice8. Provided the positive fluorescence reaction with sarcoma cells is solely the result of the presence of virions or viral antigens on the cell surface, then a differentiation between viral and new cellular antigens of leukaemia cells must be possible. To substantiate this hypothesis the following experiments were carried out. Landschutz sarcoma 1 cells were used to absorb a Graffi immune serum pool containing virus-neutralizing antibodies and antibodies against specific surface antigens of leukaemia cells. Usually 0.1 ml. samples of the serum were added to cell pellets containing 80 million living cells. The mixture was then incubated for 60 min at 37° C. After centrifugation the supernatant serum was again absorbed with another pellet of sarcoma cells. After two absorptions the serum was tested against syngeneic L 414/2a Graffi chloroleukaemia cells using the fluorescent antibody technique. Its virus-neutralizing capacity was subjected to the MAP test.