Fetal rat spinal cord tissue was obtained on gestational day 14 (E14) and transplanted into 2-4-mm-long intraspinal cavities produced by partial spinal cord lesions in adult and neonatal rats. At regular post-transplantation intervals, light and electron microscopy, autoradiographic demonstration of tritiated thymidine labelling, and immunocytochemical localization of glial fibrillary acidic protein (GFAP) were used to identify surviving donor tissues and to study their differentiation and extent of fusion with recipient spinal cords. In some experiments, wheat germ agglutinin-conjugated horseradish peroxidase (WGA-HRP) was also employed to examine whether neurons within the grafts projected axons into the host spinal cord and vice versa. Lastly, immunocytochemistry was used to determine whether any supraspinal serotoninergic (5-HT) axons from the host extended into the transplants. Over 80% of the grafts survived in lesions of both the neonatal and adult rat spinal cord for periods of 1-16 months (duration of experiment), and considerable maturation of donor tissue was evidenced, which even included the appearance of some topographical features of the normal spinal cord. Many of the transplants extended the entire length of the lesion, and were often closely apposed to the injured surfaces of the recipient spinal cords without an intervening dense glial scar. At post-transplantation intervals of 2-4 months, injection of WGA-HRP into the host spinal cord (5 mm from the transplant in adult animals or as much as 20 mm in neonatal recipients) demonstrated retrogradely labelled neurons and anterogradely labelled axons in the grafts. Likewise, injecting WGA-HRP into transplants in adult recipients resulted in labelling of neurons in adjacent segments of the host spinal cord; some labelled axons, derived from donor neurons, were also present in neighboring spinal gray matter. Finally, immunocytochemistry revealed 5-HT-like immunoreactive fibers in transplants that had been prelabelled with tritiated thymidine. These observations demonstrate the potential of embryonic spinal cord transplants to replace damaged intraspinal neuronal populations and to restore some degree of anatomical continuity between the isolated rostral and caudal stumps of the injured mammalian spinal cord.