Few ideas have gained such strong acceptance in the scientific community as the monoclonal origin of tumors; the idea that tumors start with a single mutated cell (or a single clone of cells) that go on to accumulate additional mutations as a tumor develops. The certainty with which this concept is held by the scientific community reflects the length of time it has been unchallenged and the experimental difficulty in obtaining direct evidence to the contrary. Yet, recent findings regarding X chromosome inactivation patch size indicate that the X-linked marker data previously interpreted as evidence of monoclonal tumor origin is actually more consistent with polyclonal tumor origin, a situation where two or more cells or clones of cells interact to initiate a tumor. Although most tumors show homotypy for X-linked markers (as expected given the bias conferred by X chromosome inactivation patch size), the literature contains numerous examples of tumors with X-linked marker heterotypy, examples of which encompass 24 different tumor types. Chimeric models have yielded direct unequivocal demonstrations of polyclonality in rodent and human tumors. Also, mutational data are consistent with polyclonal tumor origin. Methods that analyze levels of tumor-associated oncogene and tumor suppressor gene mutations demonstrate that initiated cells are much more common in normal tissues than previously realized. Also, while tumors have higher levels of mutation than normal tissues, oncogenic mutations frequently are present as subpopulations within tumors, rather than as the pure mutant populations expected to develop from a single initiated cell. Understanding the mutational basis of tumor etiology has important practical significance for assessing cancer risk, as well as in modeling and treating cancer. Therefore, the scientific community needs to re-examine this issue and consider the implications of polyclonal origin for, perhaps, a majority of tumors, encompassing many different tumor types.