Drug resistance has long been identified as a major reason for therapy failure in cancer patients. Concurrently, work from many laboratories in the past 10 years has established tumor heterogeneity as a phenomenon of critical importance in the natural history of individual neoplasms. The two most sinister aspects of intraneoplastic diversity in human solid tumors are the genesis of clones with metastatic potential, and the existence of drug-resistant variants in primary cancers and their metastases. Thus, recent investigations on drug resistance and on tumor heterogeneity have converged to focus attention on the clonal organization of primary tumors and their metastases as the underlying basis for anticancer drug resistance. This review examines the degree of heterogeneity observed within tumors and the relationship of this diversity to resistance that might be anticipated for any given agent. A question critical to our discussion is "How many subpopulations are there?" The impact of multiple tumor clones on therapy is next discussed in relationship to normal tissue tolerance, the barrier clinicians face regardless of the specific agent used in treatment. Finally, laboratory and clinical approaches are presented for addressing a drug resistance problem that is seemingly overwhelming because of its complex biological roots.