be suggested that it is a curtain wall of Saladin, I hasten to add that it lacks the narrow headers, so characteristic of his work.

It will be observed that a triangular space is left between this curtain wall and that of Muḥammad 'Aly. On entering the latter by the door M we find that its internal gallery extends about a metre and a half under the area we have just examined. The rest of the triangle has presumably been filled with rubble. The old curtain wall can only extend as far as it is actually visible, i. e. to the end of the vaulted gallery at the bottom of staircase g and the wall m; after that its track is occupied by the open court between Muḥammad 'Aly's wall and the Hospital, a court of which the level is nearly as low as the floor of the vaulted gallery.

DATE OF SQUARE TOWER. By its internal construction this great tower is closely related to Burg Kerkyalân, but its masonry has suffered and been refaced to such an extent that it is difficult to find a representative piece; however, a careful examination near the ground level reveals several blocks with a dressing, still intact, resembling the rusticated work we have already observed in the two square towers and the two great round ones. The arrow-slits, also, are similar, being covered, not by a lintel as in Saladin's work, but by a tapering tunnel vault like half a cone laid on its side. The vaults of the corner rooms on the upper and lower floor are pointed in section, as also are the vaults covering the arms of the cross on the lower floor. On the upper level, although the vaults of the corner rooms are pointed in section, the arms of the cross are covered by semi-circular tunnel-vaults, strengthened, in the case of the north-western arm, by semi-circular arches, the whole of very new appearance. The centre part instead of being cross-vaulted is covered by a shallow dome on spherical-triangle pendentives of the same curvature, also of very modern appearance. I therefore attribute this tower to al-'Âdil, and conclude that Muḥammad 'Aly vaulted parts of the upper storey, then no doubt ruined, before building on the top of it.

If we retrace our steps, we observe that next to this tower is a piece of wall measuring 25 m. go in length, the openings of which - a walled-up door and several windows - have a thoroughly xixth century appearance, the mouldings of the door in fact are obviously the work of Muhammad 'Aly. On examination,

it turns out to be part of Muhammad 'Aly's Palace, now occupied by the Stores Section of the Hospital. I shall not describe it in detail as the plan (Fig. 10) shows all that is necessary. It will be sufficient for our purpose to point out that a direct measurement, taken through the window n, shows that the south-eastern side of the great tower must have been lined, as indicated on the plan, by a wall which forms one side of the main hall of the Hospital (west wing). On passing through this hall into the yard already mentioned, the eastern corner of the great tower is seen to be chamfered off like the northern and western (1).

On passing along towards the doorway E, another length of wall, remarkable for the size of its masonry, is observed (see Plates XXII and XXIII). This masonry has suffered terribly, and, from the amount of plaster adhering to its upper portion, it appears probable that it once formed the side of a large hall or store-room, and a row of square holes high up suggests a wooden roof, Nevertheless a second glance suffices to reveal the existence of quite a number of large rusticated blocks similar to al-'Âdil's work. The openings seen lead to nothing, as this wall is backed by another forming the south-western side of the Quadrangle which serves the west wing of the Hospital. I climbed from the top of the Bab al-Mudarrag, along the wall built above the doorway E (see Plate XXII) and then on to the top of this massive wall. It does not rise quite so high as the Quadrangle wall, as may be seen in the Plate. These two walls meet at an acute angle as shown (Fig. 9), and there is a small space between them at the southern end, which allows the buttresses of the Quadrangle wall to be seen. This strip of wall is 25 m. 10 in length, and its southern end is clearly marked (see plan), but the opposite end is ragged, so that probably it once extended a little farther in that direction. Can it be the remains of a second great square tower? Seeking confirmation, I looked once more at Napoleon's map (Fig. 10), and there, clearly marked in this corner of the Citadel, are two heavy black squares, with a number 69 against them. On turning to the text, the explanatory Index reads: 69.- Tours en partie ruinées.

The tower shown next the corner clearly corresponds with the one we have already discussed, this therefore must be the other, which in Napoleon's day

(1) I doubt whether the corners were always so chamfered.

Fig. 10.

THE CITADEL (from NAPOLEON'S Description de l'Égypte).

was evidently foursquare, so I have dotted it in accordingly (1). I conclude that it was much more ruined than its fellow, so that when Muḥammad 'Aly built his Palace, he preferred to enfold the latter rather than face the cost of its destruction, but that the former was so far gone that he cleared the site for his Quadrangle, leaving one side of it only.

There have therefore been four successive stages at this corner of the Citadel, as follows:

(1) The Bab al-Mudarrag and a curtain wall running up to a round corner tower, perhaps on the site of the great square tower.

(2) The insertion of the two mighty towers by al-'Âdil.

(3) The abandon of these two towers and the construction of the wall F F F, together with the round corner tower, and its continuation beyond, parallel to the north face of the square tower. The space between the new and the old alignment must have been filled up with a great mass of material to a level 4 or 5 metres above the base of the corner tower, builders' refuse, no doubt, being the chief source of supply.

(4) Muḥammad 'Aly's extension, which advanced the north façade of the enclosure, enveloping part of the curtain wall of the third period, which ran east from the corner tower, but destroying the rest.

DATE OF ROUND TOWER AND CURTAIN WALL. The round tower is solid, and so is the curtain wall, at least no openings are visible on the exterior below the parapet. This suggests that it was built in the days of artillery, i. e. not earlier than Janbalât. Its parapet, which is of exactly the same masonry as the lower part, is arranged for musketry and cannon. This masonry, however, is quite different from that of the casing near the water tower which I have attributed to this Sultan. On the other hand, in the size of the stones and in the absence of the narrow headers characteristic of Saladin's work, it resembles the masonry of the Muqaṭṭam tower, except that it has been heavily pointed with cement, and the stones have not weathered so well. On the whole, I am inclined to place it in the first century of the Turkish period (i. e. 1517-1617).

(1) Burg Kerkyalán, Burg at-Turfa and the corner tower described above are all almost square, so it is a fair deduction thus to restore this tower; it is moreover shown square on Napoleon's

map.