A clinical trial of applied, slowly oscillating, weak electric fields was performed in dogs with naturally occurring spinal cord injuries due to intervertebral disc herniation. Criteria for admission to the study were: complete paraplegia, defined by neurological examination and electrophysiological testing; intact segmental reflexes; radiologic and myelographic evidence of spinal cord compression due to disc herniation and a focal lesion, without appreciable rostrocaudal spread of necrosis; weight less than 16 kg; onset of paralysis less than 1 month before surgery. The injured cord was exposed by laminectomy, and decompressed by aspiration of disc material. Active (n = 13) or sham (n = 11) stimulators were implanted subcutaneously, with platinum/iridium electrodes sutured to muscle several millimeters above the cord surface, at either end of the laminectomy site. Active stimulators delivered 200 μA of direct current, switching polarity every 15 min, for 3, 6, or 15 weeks. Neurological and electrophysiological examinations were repeated approximately 6 weeks and 6 months after implantation. Few complications were noted. None were attributable to current application. The group of dogs with active stimulators showed greater improvement in function than the group with sham implants, with a trend towards greater recovery in all neurological measures, and evoked potentials. The combined neurological score derived from these separate tests was significantly different from controls at 6 weeks and 6 months (P < 0.05). Application of slowly oscillating electric fields thus appears to be beneficial in the treatment of paraplegia in dogs.