Electrophysiological studies were used to evaluate neurological recovery in 14 rhesus monkeys with different nerve lesion lengths and graft lengths. After exposure of both sciatic nerves in each animal, baseline evoked nerve action potentials, muscle action potentials, and muscle strength values were determined for the posterior tibial nerves. Each nerve was then crushed over a measured distance. Three weeks later, the crushed segments were resected and the defects repaired with sural nerve grafts. In seven animals, 20-mm resection sites were repaired by 4 x 20-mm grafts in one leg and by 4 x 40-mm grafts contralaterally. In the other seven animals, the lengths of resection sites were 10 mm in one leg and 30 mm contralaterally; both nerve defects in these animals were repaired by 4 x 30-mm grafts. Electrophysiological studies were repeated at one interval of either 4, 7, or 12 months after repair. Postoperative electrophysiological values were compared to baseline values and described by the mean values and by percent recovery. Muscle strength recovery was significantly better in limbs with short lesions. In animals with identical lesion lengths, lesions repaired with shorter grafts (the same length as the defect) did significantly worse than did lesions repaired with longer grafts. This may suggest that any degree of tension at the graft repair site has a deleterious effect on functional nerve regeneration. Nevertheless, it was generally found that nerve lesion length had the greatest negative effect on functional nerve regeneration.