Regeneration in the peripheral nervous system is often incomplete though it is uncertain which factors, such as the type and extent of the injury or the method or timing of repair, determine the degree of functional recovery. Serial electrophysiological techniques were used to follow recovery from median nerve lesions (n = 46) in nonhuman primates over 3 to 4 years, a time span comparable with such lesions in humans. Nerve gap distances of 5, 20, or 50mm were repaired with nerve grafts or collagen-based nerve guide tubes, and three electrophysiological outcome measures were followed: (1) compound muscle action potentials in the abductor pollicis brevis muscle, (2) the number and size of motor units in reinnervated muscle, and (3) compound sensory action potentials from digital nerve. A statistical model was used to assess the influence of three variables (repair type, nerve gap distance, and time to earliest muscle reinnervation) on the final recovery of the outcome measures. Nerve gap distance and the repair type, individually and concertedly, strongly influenced the time to earliest muscle reinnervation, and only time to reinnervation was significant when all three variables were included as outcome predictors. Thus, nerve gap distance and repair type exert their influence through time to muscle reinnervation. These findings emphasize that factors that control early axonal outgrowth influence the final level of recovery attained years later. They also highlight that a time window exists within which axons must grow through the distal nerve stump in order for recovery after nerve lesions to be optimal. Future work should focus on interventions that may accelerate the growth of axons from the lesion site into the distal nerve stump.