It is axiomatic that a given dose of an antitumor agent will not produce the same effect in 100% of the treated subjects. Numerous explanations regarding the sources of this heterogeneous response to drugs have been offered; however, there is a scarcity of experimental data allowing critical evaluation of the sources of variance. It is possible to study heterogeneous antitumor drug response in experimental, inbred animals. One animal model system, the advanced Ridgway osteogenic sarcoma, exhibits marked variation in its response to maximally tolerated doses of a number of clinically active antitumor agents. To evaluate the role of the host in the variable drug response, the tumor was bilaterally implanted into the flank regions of recipient AKR male mice. Treatment of the advanced tumor (200 mg-1,500 mg) with maximally tolerated doses of vincristine or L-phenylalanine mustard produced marked, but variable antitumor responses. Evaluation of a number of quantal and graded parameters of the chemotherapeutic response suggested that host heterogeneity contributes to variability. The host contribution was more apparent in this experiment model when the agent was noncurative. The underlying biological basis for the host heterogeneity is not known; however, it appears likely that pharmacological, immunological or other differences between the inbred animals account for the heterogeneity. Identification of these factors may be experimentally feasible in this animal model and help in the design of future studies in humans.