The recent observation by Tyler Jacks' group that loss of ARF accelerates pituitary tumors in Rb(+/-):Arf(-/-) mice by inducing proliferation without affecting apoptosis allows one to speculate about an alternative model for retinoblastoma. Retinoblasts in human and pituitary gland melanotrophs in the mouse might be uniquely susceptible to bi-allelic inactivation of Rb because upon loss-of-heterozygosity, these tissues are able to slowly proliferate without effectively activating the cell death machinery, and respond to an additional oncogenic signal. But what is the nature of this proliferation-permissive oncogenic alteration?