With the exception of that citadel’s outer wall, erected by Leo the Armenian, the defences from the Kerko Porta to the Golden Horn have usually been ascribed to the Emperor Heraclius.[[449]] But this opinion is at variance both with history, and with the striking diversity in construction exhibited by the various portions of the works. As a matter of fact, the fortifications extending from the Kerko Porta to the Golden Horn comprise walls that belong to, at least, three periods: the Wall of Heraclius, the Wall of Leo, and the Wall of Manuel Comnenus.[[450]] Curiously enough, the Wall of Manuel Comnenus, though latest in time, stands first in order of position, for it intervenes between the Theodosian Walls, on the one hand, and the Heraclian and Leonine Walls, on the other.
Here, therefore, a question presents itself which must be answered before proceeding to the study of the walls just mentioned. If the various portions of the fortifications between the Kerko Porta and the Golden Horn did not come, respectively, into existence until the seventh, ninth, and eleventh centuries, how was the north-western side of the city defended previous to the erection of those walls?
Two answers have been given to this important and very difficult question. Both agree in maintaining that the city was defended on the north-west by the prolongation of the Theodosian Walls; but they differ as regards the precise direction in which the walls were carried down to the Golden Horn.
One view is that the Theodosian Walls upon leaving the Kerko Porta turned north-eastwards, to follow the eastern spur of the Sixth Hill,[[451]] along a line terminating somewhere in the vicinity of Balat Kapoussi.[[452]] According to this view, the quarter of Blachernæ, which until 627 lay outside the city limits,[[453]] was the territory situated between the spur just mentioned and the line occupied eventually by the Walls of Comnenus and Heraclius.
The second view on the subject is that the two Theodosian Walls were carried northwards along the western spur of the Sixth Hill, and enclosed it on every side. On this supposition, the suburb of Blachernæ, with its celebrated Church of the Theotokos, without the fortifications, was the plain extending from the foot of the western spur of the Sixth Hill to the Golden Horn, the plain occupied now by the quarter of Aivan Serai.[[454]]
In support of the first opinion, there is the undoubted fact that the Theodosian Walls, as they approach the Kerko Porta, bend north-eastwards, so that if continued in that direction they would reach the Golden Horn near the Greek Church of St. Demetrius, to the west of Balat Kapoussi.
The opinion that the Theodosian Walls were carried to the foot of the western spur of the Sixth Hill rests upon the fact that traces of old fortifications enclosing that spur are still distinctly visible; while the Theodosian Moat is, moreover, continued towards Aivan Serai, until it is stopped by the Wall of Manuel, which runs transversely to it.[[455]]
The fortifications referred to are found mostly to the rear of the Comnenian Wall, but portions of them are seen also to the north of it.
One line of the fortifications proceeded from the Kerko Porta along the western flank of the spur, and joined the city walls a little to the south of the “Tower of Isaac Angelus;” another line ran from that gate along the eastern side of the spur to the fountain Tsinar Tchesmè in the quarter of Londja, a short distance to the south-east of the Holy Well which marks the site of the Church of Blachernæ; while a third wall, facing the Golden Horn, defended the northern side of the spur, and abutted against the city walls, very near the southern end of the Wall of Heraclius.[[456]] Within the acropolis formed by these works of defence, the Palace of Blachernæ and the Palace of the Porphyrogenitus were in due time erected.
Both answers to the question before us have much in their favour, and possibly the truth on the subject is to be found in their combination. Their respective values as rival theories will, perhaps, be more easily estimated, if we begin with the consideration of the second answer.