Immunological endeavor in recent years calls for a reappraisal of the concept of immunosurveillance against neoplasia. This concept proposes an immunological policing system capable of aborting tumor growth by the recognition of "nonself" tumor associated antigens on neoplastic cells. The model is supported by evidence of tumor induction in the immunosuppressed host and the demonstration of an immune response to tumors in animals. The occurrence of tumor, regarded as a failure of immunosurveillance, is attributed to selection of neoplastic cells for immunological or other reasons or abnormal humoral or cellular antitumor immune responses. However protagonists of the postulate are faced with mounting evidence that fails to support the surveillance hypothesis. These observations include, inter alia, the monoclonality of certain tumors, the low incidence of spontaneous tumors in genetically immunodeficient mice and immunological privileged sites, and new ideas about the pathogenesis of lymphoproliferative neoplasms. However, contradictory arguments are not sufficiently substantiated to prosecute the case against surveillance conclusively. In citing highlights of the evolving quandary, both the pros and cons of immunological surveillance are presented here.