Lower vertebrates, such as fish and urodele amphibians can regenerate complex body structures including significant portions of their central nervous system by recruiting progenitor cells to repair the damage. Significant ability to regenerate the nervous system is observed also during development in higher vertebrates, for example in the chick spinal cord, though it is not yet clear whether this involves de novo neurogenesis, in addition to axonal re-growth, also at the latest stages of development permissive for regeneration. The mechanisms underlying recruitment of progenitor cells in response to injury, particularly within the nervous system, are still poorly understood. Although it has been suggested that some neurogenesis can be induced even in regions of the adult mammalian brain, this potential is largely lost with evolution and development. Following tail amputation in urodeles, an ependymal tube, resembling a developing neural tube, forms from ependymal cells that migrate from the cord stump towards the terminal vesicle, and elongates by cell proliferation. The new cord might originate from stem cells, with possibly only a subset of ependymal cells displaying such properties, or via a process of dedifferentiation / transdifferentiation of these cells. Data currently available are more supportive of the latter hypothesis. Whereas dedifferentiation is a well demonstrated phenomenon in a broad range of urodele tissues, transdifferentiation seems to occur less widely and in extreme circumstances, and may contribute significantly to regeneration only in a few cases. In higher vertebrates it is even less clear how common and relevant to repair transdifferentiation is, as much work both in favour and against it has recently been published. However, the existence of multipotent neural progenitors in adult mammalian CNS and of a much higher neural cell plasticity, at least in vitro, than previously believed, encourages the view that if we were to better understand progenitor cell recruitment and plasticity in species where it does occur spontaneously, we might then find the way to make it happen effectively in mammals.