How the non-Theodosian character of the walls to the north of the Kerko Porta is to be accounted for admits of more than one explanation. It may be due to changes in works of Theodosian origin, or to the fact that they are works of an earlier period,[462] or to the fact that they are works of a later age. On the supposition that these fortifications defended originally the Fourteenth Region, the second explanation is the most probable, for the division of the city into Regions was anterior to Theodosius II., and there is every reason to believe that the isolated Fourteenth Region was a fortified suburb from the earliest period of its history.[463] Accordingly, the second answer to the question how the north-western side of the city was defended before the erection of the Walls of Heraclius, Leo, and Manuel Comnenus, would have more in its favour if it maintained that the defence was effected by the junction of the Theodosian Walls with pre-existing fortifications around the western spur of the Sixth Hill.[464] The chief difficulty attending this view is that the _Notitia_ speaks of the Fourteenth Region as still an isolated suburb in the reign of Theodosius II.[465] As regards the opinion that the Theodosian Walls proceeded from the Kerko Porta to the Golden Horn in a north-eastern course and reached the water between the Church of St. Demetrius and Balat Kapoussi, it has upon its side the patent fact that those walls, if produced according to their trend at the Kerko Porta, would certainly follow the line indicated. On this view, the walls around the western spur of the Sixth Hill were either the fortifications of the Fourteenth Region (modified), or walls built expressly to defend the Palace of Blachernæ, after the fifth century. The trend of the walls at the Kerko Porta affords, unquestionably, a very strong argument for this view of the case. But the view is open to objections. The absence of all traces of the walls along the line indicated should, perhaps, not be pressed, as such works are apt to disappear when superseded. A more serious objection is that the Theodosian Moat does not follow the north-eastern course of the walls, but proceeds northwards, for a short distance, in the direction of Aivan Serai. Furthermore, if the western spur of the Sixth Hill was already fortified when the Theodosian Walls were built, it is reasonable to suppose that the land defences of the city were completed by the simple expedient of uniting the new works with the old. Any other proceeding appears cumbrous and superfluous. Still, after all is said, the information we have is so meagre, the changes made in the walls beside the Kerko Porta have manifestly been so numerous, that a decided judgment upon the point at issue does not seem warranted by the evidence at our command. Footnote 449: Paspates, p. 19. Footnote 450: Dr. Mordtmann was the first to prove this. See below, p. 122. Footnote 451: