The use of a devitalized skeletal muscle graft and conventional nerve graft to repair a 5mm long segmentary sciatic nerve lesion was studied in rats by means of functional, morphometric and spinal cord motor neuron cell response evaluation. Thirty-four rats were used and divided into four groups according to the procedure: (1) sham operation; (2) conventional nerve grafting; (3) muscle grafting; (4) unrepaired lesion. The sciatic functional index (SFI) was evaluated every fortnight up to the 105th postoperative day by measuring three parameters in the rats' footprint. The animals of Groups 2 and 3 presented initial complete functional loss, followed by slow but steady recovery, with final similar SFIs. The histologic and morphometric studies showed an increased small diameter/thin myelin sheath nerve fiber density distally to the lesion site for both types of graft. An increased population of motor neurons was observed in the anterior horn of the lumbar spinal cord segment with both types of grafts, but not in the control groups. The SFI, histologic and morphometric data did not differ significantly between the two types of graft, thus indicating a similar behavior. The authors conclude that a 5mm long skeletal muscle graft works as well as a conventional nerve graft.