Oncogenic viruses may cause human neoplasia; the failure to isolate human tumor viruses may be technical. Immunotherapy for cancer has produced remissions; recent work on host-tumor relations suggests that it may be a useful adjunct to surgery and a stimulant to the immune system after chemotherapy and radiation therapy. A new method of immunotherapy could result from a method useful in animals. Normal, nonimmune cells are made specifically immunoreactive by incubation with preparations rich in RNA extracted from lymphoid cells of immune animals, after incubation with tumor antigen. This method has been successful using lymphoid cells of animals different from the tumor's original host. Improved therapy of advanced cancer may develop from recent animal experiments using drugs to interrupt the single cell's life cycle. Prognosis for some cancers can often be determined by measurement of the patient's immunological competence, his specific immune response to tumor antigens, and a measurement of host-tumor interaction. Immunocompetence and a slow tumor doubling time offer optimum prognosis.