The present work investigates the extent to which mature central neurons acutely or chronically axotomized by a spinal lesion still maintained the potential to regenerate an axon following post-traumatic nerve grafting within supra-lesional spinal structures. In adult rats, a C3 cervical hemisection (injury) was made and an autologous segment of the peroneal nerve was implanted 2 mm rostrally into the ventrolateral part of the ipsilateral C2 spinal cord. Nerve graft implantations were carried out acutely at the time of injury (group I, acute conditions) or chronically, three weeks post-injury (group II, chronic conditions). Central neurons axotomized by the spinal lesion were labeled by True Blue injected at the lesion site at the time of trauma. Central neurons regenerating axons within the nerve grafts were labeled with either horseradish peroxidase (only in group I, n = 4) or Nuclear Yellow (group I, n = 3 and group II, n = 6) applied two to four months post-grafting to the distal cut end of the nerve grafts. Neurons with dual staining (True Blue/Nuclear Yellow) represented central regenerating neurons which were previously axotomized by the spinal lesion and which had retained the capacity for axonal regeneration for a delayed period after injury. In group I (acute injury conditions), all types of labeled cells were found to be scattered with a clear bimodal distribution within the spinal cord and the brainstem. No labeled cells were found within the motor cortex. There was no statistically significant difference between horseradish peroxidase and all cells containing Nuclear Yellow (Nuclear Yellow and True Blue/Nuclear Yellow). In group II (chronic injury conditions), Nuclear Yellow- and True Blue/Nuclear Yellow-labeled cells had a similar dual distribution to that of group I, but were found to be significantly less represented (P = 0.019). These differences are discussed in terms of capacity for cell survival and axonal regrowth after acute and chronic injury. The main conclusion is based on the evidence of dual staining of central neurons in both groups, which demonstrates that brainstem and spinal neurons involved in acute and chronic axotomy after spinal C3 lesion can survive the trauma and still maintain the capacity to regenerate lesioned axons within nerve grafts inserted rostrally (C2 spinal cord) to the primary site of injury. Although exhibited to a lesser extent in chronic than in acute conditions, this capacity was found to occur for as long as three weeks post-injury. These results indicate that supra-lesional post-traumatic nerve grafts may constitute an efficient delayed strategy for inducing axonal regrowth of chronically axotomized adult central neurons. We suggest that surgical intervention; which is not always possible immediately after a spinal cord injury, may be satisfactorily carried out after an appropriate delay. (C) 2000 IBRO. Published by Elsevier Science Ltd. All rights reserved.