In recent years there has been increased interest in the possibility of reinnervating paralyzed muscles, particularly the muscles of the face and of the larynx. There has been dispute over the relative effectiveness of different approaches. This study was undertaken to compare direct nerve implantation with the transfer of a neuromuscular pedicle using denervated strap muscles in a rabbit model. Reinnervation was carried out both immediately after denervation and after a delay period. Evidence of return of function was obtained by a method of long-term electromyographic monitoring and was confirmed electrically by evoked electromyography and muscle tension measurement and histologically by brain stem uptake of horseradish peroxidase. Both methods of nerve transfer produced consistently functional neuromuscular units, with physiological activity and muscle strength comparable with those in normally innervated controls. Return of function was apparent within 6 weeks of both nerve implantation and neuromuscular transfer; with this model, neither method showed a clear advantage over the other.