Four experiments investigating the antitumor activity of bacterial lipopolysaccharide (LPS) against the autochthonous methylnitrosourea-induced mammary carcinoma are summarized. Administration of LPS alone i.v. caused distinct regression of small tumors following its first injection. This therapeutic effect, however, was short-lived and could not be maintained by administering a second dose. The observed antineoplastic activity of LPS was dose-related, whereas no dose-response relationship was observed with respect to its toxicity. A series of experiments in which LPS was combined with other compounds to possibly exploit its activity while reducing the toxicity were performed. Neither the combination with cytotoxic drugs such as 4'-(9-acridinylamino)-methansulfone-m-aniside or cyclophosphamide nor that with 1-octadecyl-2-methoxy-rac-glycero-3-phosphocholine or hexadecylphosphocholine showed sufficient anticancer activity at acceptable toxicity. In all experiments promising efficacy was observed at high dosages but also high toxicity. When the dosages were reduced, diminished antineoplastic activity was found together with overproportionally high mortality. It might therefore be concluded that the active dose range of LPS cannot be reached clinically because of its inherent toxicity.