Although much of cancer research relies on Nowell's clonal evolution hypothesis as a conceptual framework, large gaps remain in understanding how tumors develop. The multistage skin cancer model in mice provides continuing insight on fundamental aspects of tumor evolution. In this model, mutation of the oncogene Hras is frequently the initiating event while mutation of the tumor suppressor p53 is a late event, associated with malignant progression. Recent evidence demonstrates that intracellular signaling from the initial Hras mutation leads directly to the activation of p53, creating selective pressure in favor of cells with mutant p53. Thus, selection for subsequent mutations is mechanistically linked to the initial mutation, explaining the preferred order of mutational events observed. Analysis of this model also reveals that a diverse array of signals can selectively impair or enhance clonal expansion of Ras mutant cells into a visible neoplasm. These modifiers can be genetic, physiological, or environmental and are often highly specific to tumor cells. This indicates that tumor cells have an inherent reduced capacity to buffer against perturbations. Reduced buffering may play an important role in both tumor evolution and therapy response and may be a hallmark of cancer cells.