One major factor hindering progress of pediatric cancers of the nervous system has been the lack of satisfactory model systems for testing novel therapies. A mouse strain, mutant for the Rb1 gene was generated 12 years ago in the hope of producing a model in which to study retinoblastoma. Surprisingly, the Rb(+/-) mice never developed retinoblastoma. Now, Zhang, Schweers and Dyer produce triply deficient Rb, p107 and p53 mutant retinal progenitor cells. All such mice develop intraocular retinoblastoma with invasion of the tumor into the anterior chamber of the eye. This dramatic finding represents the first description of a heritable mouse model of retinoblastoma, which has eluded investigators for the last 12 years. Such models provide an unprecedented opportunity to advance knowledge of tumorigenesis and to develop non-toxic intervention strategies which eradicate disease.