Photoreceptor loss causes irreversible blindness in many retinal diseases. Repair of such damage by cell transplantation is one of the most feasible types of central nervous system repair; photoreceptor degeneration initially leaves the inner retinal circuitry intact and new photoreceptors need only make single, short synaptic connections to contribute to the retinotopic map. So far, brain- and retina-derived stem cells transplanted into adult retina have shown little evidence of being able to integrate into the outer nuclear layer and differentiate into new photoreceptors. Furthermore, there has been no demonstration that transplanted cells form functional synaptic connections with other neurons in the recipient retina or restore visual function. This might be because the mature mammalian retina lacks the ability to accept and incorporate stem cells or to promote photoreceptor differentiation. We hypothesized that committed progenitor or precursor cells at later ontogenetic stages might have a higher probability of success upon transplantation. Here we show that donor cells can integrate into the adult or degenerating retina if they are taken from the developing retina at a time coincident with the peak of rod genesis. These transplanted cells integrate, differentiate into rod photoreceptors, form synaptic connections and improve visual function. Furthermore, we use genetically tagged post-mitotic rod precursors expressing the transcription factor Nrl (ref. 6) (neural retina leucine zipper) to show that successfully integrated rod photoreceptors are derived only from immature post-mitotic rod precursors and not from proliferating progenitor or stem cells. These findings define the ontogenetic stage of donor cells for successful rod photoreceptor transplantation.