north to south ([Fig. 135]). The four angle towers are larger in diameter than those which are set along the walls. The intermediate bastions are perfectly regular in size and shape except the two on either side of the southern gate, from which a segment is cut off by the door openings, and the bastion immediately to the west of the same gate which has a small addition to the western part of its curve, an addition which I do not believe to be later in date though the brickwork is of a slightly different character. The southern gate is a triple opening in the middle of the wall where it would be natural to look for the miḥrâb ([Fig. 138]). There are remains of mouldings round the inner face of the central opening ([Fig. 139]). The upper part of the south wall is pierced by twenty-four windows, two of them being placed over the smaller openings of the central gateway ([Fig. 122]). These windows, together with the trenches in the interior of the mosque which mark the line of the columns, determine the number of the colonnades; there must have been twenty-four, each one ending against the wall between the windows. The central aisle which terminated at the main gate and was wider than the rest, was not provided with a window. The space between the colonnades was undoubtedly roofed with beams; the holes into which the large cross-beams were fitted can still be seen on the inner side of the south wall. The windows, placed with regard to the aisles, bear no relation to the position of the round bastions on the exterior of the wall. They break into them at haphazard, frequently impinging upon their sides, while in one instance a window is cut straight through a tower ([Fig. 120]). On the inner face the windows are covered by a cusped arch ([Fig. 142]). The east and west walls are broken by numerous doors. Beginning from the southern end there is first a small entrance, 1·25 m. wide, close to the angle bastion ([Fig. 141]). A wall about a metre in length projects from the main wall to the south of the door opening and has been connected with the top of the main wall by a section of vaulting. Immediately beyond this postern there is a large gateway 4·55 m. wide, and then another which is still larger, being 4·75 m. wide. The next door is 3·85 m.; the fifth, which is only 2·62 m., is found in the west wall alone. Then follows another of the larger doors, about 4 metres wide, beyond which there is, in the west wall only, a door 2·62 m. wide; then on both sides a large door 4·05 m. wide and a small door 1·50 m. wide. The north wall is broken by five gates, the two at the outer ends averaging 1·50 m. and the other three 4 metres in width. All the smaller doors exhibit an exceedingly curious piece of construction ([Fig. 140]). The brickwork of the wall runs uninterruptedly over the door opening without the intermission of arch or lintel. It is as if the door had been cut out of the wall with a knife, and the bricks above it, so far as they keep their place, do so only by reason of the excellence of the mortar. The wall above the larger doors has in every case fallen away, but there is evidence of the former existence of some kind of lintel or arch strengthened by wooden beams, the round holes for the beams being visible in the existing masonry ([Fig. 143]). I incline to the theory of a lintel; the faced wall above the holes leaves no room for an arch. Above this lintel there would seem to have been a row of small arched windows two or three in number (cf. the two side openings of the south gate where there is a single window above the arch). Along the top of the east, west, and north walls runs a brickwork decoration consisting of a series of recessed squares, each of which contains the recessed segment of a sphere. The walls are seamed from top to bottom with narrow runnels, which were no doubt connected with the drainage system of the roof. There is no unanimity of opinion among those who have planned the mosque concerning the number of the colonnades in the interior. As I have already said, it seems to me evident that there were twenty-four rows of columns or piers, from east to west, at the northern and southern ends of the mosque. I made out the colonnades to be ten deep upon the south side and three deep upon the north, while upon the east and west sides I counted four rows of columns.[130] The supports of the arcades must have been either columns or small piers. From the absence of any structural remains, such as might have been expected if the supports had taken the form of brick piers, I incline, with Herzfeld, to the view that the roof must have been carried on columns. Their total disappearance may possibly be accounted for by the fact that they were of wood,[131] though Muḳaddasî, writing at the end of the tenth century, relates that the mosque of Sâmarrâ was built upon marble columns and his evidence cannot be wholly dismissed. In the centre of the open court was placed, in all probability, the famous stone basin called the Kâs i Fir’aun (Pharaoh’s Cup), which is described by Mustaufî.[132] The minaret, with its singular spiral path, stands to the north of the mosque. The summit, though somewhat ruined, still retains a decoration of niches. There can be little doubt that the mosque is that which was erected by Mutawakkil (A.D. 847-861) to replace Mu’tamid’s Friday mosque, but Yâḳût asserts that the minaret is a relic of Mu’tamid’s foundation. Yâḳût, however, wrote in 1225 when Sâmarrâ had long been in ruins.
Next in importance to the mosque is the castle or palace on the opposite bank of the Tigris, known as the ’Ashiḳ ([Fig. 145]).[133] The first time I visited it we crossed in a guffah from a point a little below the town where there is usually a bridge of boats. The bridge had been swept away by the floods and the guffah landing was very bad. It was a full hour’s
ride up the river to El ’Ashiḳ, but I was rewarded for my trouble by finding indubitable traces of a masonry bridge in the low ground almost exactly opposite a curious little building called Ṣlebîyeh. My attention was called to the bridge by seeing men digging out the brick piers and arches for building material. The peasants told me that when the river is low, piers can be seen in the bed of the stream and that the bridge ran in the direction of the Beit el Khalîfah. I give this information for what it is worth. Ya’ḳûbî mentions a bridge of boats (ed. de Goeje, p. 263); it is not impossible that pontoons may have been thrown across the deepest and swiftest part of the river and connected with the high ground on the west bank, which is at some distance from the stream, by a series of masonry arches of which I saw the remains. The piers and arches would therefore have stood on ground which was under water in time of high flood. This is exactly the arrangement of the modern bridge at Môṣul. The castle of the ’Ashiḳ consists of a great enclosure, 123 metres from north to south and 85 metres from east to west, surrounded by a wall with round bastions which are set upon a rectangular base ([Fig. 146]). All the buildings that may have stood within the wall have vanished, but adjoining the north wall there are remains of a gatehouse consisting of five parallel chambers opening on to a corridor or platform. The chambers and the corridor are built upon a substructure of vaults. Under the corridor the vaults run from east to west, except in the central part where the vault running from north to south is a continuation of the vault under the central chamber. Under the five chambers all the vaults run from north to south.[134] The vaults are built of flat tiles laid in slices against the head-wall without centering. They have the usual small set forward from the wall, but in one case, perhaps in more than one, there is a slight divergence from the customary arrangement. From the spring of the vault the tiles are laid horizontally for the first sixteen or seventeen courses, projecting forward so as to form a shallow curve;