Clinical trials in autologous BMT to date have indicated that significant salvage and potential cures can be obtained in patients with high grade non-Hodgkin's lymphoma (NHL) who have failed primary therapy and are treated with high dose chemoradiotherapy and autologous marrow rescue. The major need in NHL is to better define those patients who might benefit by autologous BMT and to reduce the relapse rate by improved pre- or post-transplant therapy. Similar results to those in NHL could be obtained in acute leukemia if occult tumor cells could be eliminated from autologous marrow. Animal model experiments have shown that it is feasible to eliminate low level contamination with tumor cells by in vitro immunologic or pharmacologic treatment. While it is too early to accurately assess the efficacy of ongoing clinical trials using those marrow purging techniques, a few patients have exhibited encouraging durations of CCR. Should these approaches prove to be effective in only a fraction of cases, combination in vitro treatment or the use of more efficient effector mechanisms for cell killing (e.g., ricin conjugated antibody) may very well clear occult tumor from the marrow of most patients. The encouraging results with autologous BMT in leukemia and lymphoma stand in sharp contrast to the disappointing results so far achieved with the non-hematologic solid tumors. It is, however, possible that those cancers have not been subjected to the most rigorous test for successful autologous BMT and that the search for newer agents which can produce operationally irreversible aplasia may provide a fairer test of this approach. It is to be hoped that future research will settle this point.