Much confusion still exists concerning the pathological definitions and clinical significance of double spinal cord malformations. Traditional terms used to describe the two main forms of these rare malformations, diastematomyelia and diplomyelia, add to the confusion by their inconsistent usage, ambiguities, and implications of their dissimilar embryogenesis. Based on the detailed radiographic and surgical findings of 39 cases of double cord malformations and the autopsy data on two other cases, this study endorses a new classification for double cord malformations and proposes a unified theory of embryogenesis for all their variant forms and features. The new classification recommends the term split cord malformation (SCM) for all double spinal cords. A Type I SCM consists of two hemicords, each contained within its own dural tube and separated by a dura-sheathed rigid osseocartilaginous median septum. A Type II SCM consists of two hemicords housed in a single dural tube separated by a nonrigid, fibrous median septum. These two essential features necessary for typing, the state of the dural tube and the nature of the median septum, do not ever overlap between the two main forms and can always be demonstrated by imaging studies so that accurate preoperative typing is always possible. All other associated structures in SCM such as paramedian nerve roots, myelomeningoceles manqué, and centromedian vascular structures frequently do overlap between types and are not reliable typing criteria. The unified theory of embryogenesis proposes that all variant types of SCMs have a common embryogenetic mechanism. Basic to this mechanism is the formation of adhesions between ecto- and endoderm, leading to an accessory neurenteric canal around which condenses an endomesenchymal tract that bisects the developing notochord and causes formation of two hemineural plates. The altered state of the emerging split neural tube and the subsequent ontogenetic fates of the constituent components of the endomesenchymal tract ultimately determine the configuration and orientation of the hemicords, the nature of the median septum, the coexistence of various vascular, lipomatous, neural, and fibrous oddities within the median cleft, the high association with open myelodysplastic and cutaneous lesions, and the seemingly unlikely relationship with fore and midgut anomalies. The multiple facets of this theory are presented in increasing complexity against the background of known embryological facts and theories; the validity of each facet is tested by comparing structures and phenomena predicted by the facet with actual radiographic, surgical, and histopathological findings of these 41 cases of SCM.