Lateral hemisection lesions separated by 1 to 3 spinal segments were made on opposite sides of the mithoracic spinal cord in 1-month-old (N = 15; weanling operates) and newborn albino rats (N = 16; neonatal operates). Hindlimb behavior was assessed between 1 and 6 months p.o. for both groups of operates using a protocol and rating system that have previously proved effective in differentiating behavioral recovery of the hindlimbs as a function of age of spinal transection. In addition, at the conclusion of behavioral testing, operates received spinal injections of [3H]proline and HRP caudal to the spinal lesions to determine if lesions were complete and if neurons within the region between the two lesions (interlesion zone) projected into the caudal spinal cord. In both groups of operates, neurons were retrogradely labeled within the interlesion zone bilaterally, primarily in laminae VII-VIII. When both lesions were complete lateral hemisections in weanling operates, little behavioral recovery was observed, similar to complete spinal cord transection (N = 3). However, much greater behavioral recovery was seen, including supporting reactions and locomotor responses, when one or both lesions spared axons along the ventrolateral rim of the white matter. Neurons were retrogradely labeled in the brain stem reticular formation (N = 12) in these cases. All lesions were complete lateral hemisections in neonatal operates but much greater behavioral recovery was seen than in weanling operates with the same lesions, including supporting, placing, and locomotor responses. In an additional group of eight neonatal operates, the spinal cord rostral to the spinal hemisections was transected at 1 month of age. Supportive, placing, and locomotor responses were seen immediately after recovery from anesthesia and responses returned to pretransection levels in six of eight operates over the 10-day survival period. Fink-Heimer impregnation showed that degeneration argyrophilia from the transection bilaterally filled the interlesion zone but little argyrophilia was seen caudal to this region. Our results indicate that an intact propriospinal circuit remains in both neonatal and weanling operates but does not appear to contribute to hindlimb response development or recovery. The greater behavioral recovery in neonatal operates appears due to intrinsic connections (doral root, interneuronal) continuing to be able to drive the spinal circuitry underlying the spared behaviors.