The development of immunologic approaches to the treatment of cancer has progressed through major changes in the past five years. In the 1970s considerable enthusiasm existed for immunotherapeutic approaches to cancer based on attempts to provide nonspecific stimulation of the host immune system in the hope that this general increase in immune reactiv­ ity would lead to an increased reactivity to putative tumor antigens on growing cancers. This approach was based on minimal experimental data from animal tumor models. A large number of clinical trials in humans utilized attempts to stimulate the immune system nonspecifically. Using substances such as Bacillus Calmette-Guerin (BCG), methenol-extracted residue (MER), Corynebacterium parvum, levamisole, and other immune stimulators, researchers treated patients with extensive tumor, as well as with minimal tumor burdens. This experience was almost universally unsuccessful and, both in animal tumor model systems and in the human, has largely been abandoned (I). An alternate approach to cancer immunotherapy can be categorized as "passive" or "adoptive" immunotherapy, in which the tumor-bearing host receives the systemic transfer of immunologic reagents such as antibodies or reactive cells already possessing antitumor reactivity (2-5). Because the