The ability of transplanted Schwann cells to modify the sprouts formed by cut central axons, and in particular to induce branching and extension of axon sprouts, is an encouraging sign for their possible future use in repair. The accessibility of the Schwann cells in the culture stage before transplantation offers a practical opportunity for genetic engineering (e.g. to introduce genes directing the expression of specific growth factors) which might be useful in designing a future method for the repair of human spinal injury. It must be borne in mind, however, that even the most successful cases of peripheral nerve grafts have shown only a limited proportion of axons growing back from the grafts into the environment of the CNS (Carter et al., 1989). When we constructed Schwann cells transplanted into the thalamus (Brook et al., 1994), we did not observe axons leaving the artificial tracts. In our experiments with Schwann cells transplanted into the spinal cord (Li & Raisman, 1994), the axons have only been studied within the graft, and we have as yet not been able to assess the extent to which they re-enter the CNS. For effective regeneration to occur, regenerating axons must not only be able to re-enter their original pathways and elongate along them, but also leave them in a correct manner--i.e. by making appropriate choices from a wide range of destinations. Therefore the effectiveness of a Schwann cell "bridging" repair must depend upon the self-organising capacity of the adult CNS (e.g. Florence et al., 1996).