Comparing the drawings of Grelot and the plan given by Du Cange, both published in 1680, with the present remains, it would appear that there were formerly ten of these buttresses; two being merged in the central belfry, and the two outside ones incorporated in the minarets, on the sides of which traces of them may still be seen. Two others have either been destroyed by the Turks, or Grelot’s drawings are wrong to this extent, as no trace seems to remain of more than eight. Of these eight which now in part remain, Salzenberg only reserves the four at the centre, on which he places the horses. Our Figs. 26 and 29 represent the original west front and the altered façade of the ninth century; see also Plan, Fig. 24. _Cisterns._--On the south side of the right-hand pier is a small arch which gives access to a little recessed chamber in the buttress. From this and from a similar recess north of the central entrance, water from the cisterns beneath the church was probably obtained: a cross on the wall of the little chamber would seem to show that it was a “holy well.”[326] Clavijo says the cisterns beneath the church would float ten galleys, and C. Lebrun (1714) speaks of ten cisterns and forty columns standing in the water. The only real description of the cistern we have been able to find is in Dr. Covel’s MS. diary in the British Museum. In 1676 he writes, “We went to see the vaults under S. Sophia; they were full of water, then 17 feet deep, and overhead, from the water up to the top of the arch was about 2 yards and 6 inches. Every pillar is square (4½ feet), and distant from another just 12 feet. The bricks are very broad, thin, and well baked; [it is] not plastered within, the mortar very hard. They say it goes under [the] At-Meidan, but we could not enter it. The waste water of the Aqueduct enters into it, and [going] out of it passing through the Seraglio, goes into the sea by the dunghill. [There is] severe punishment to [those who] have houses with offices [draining] into it; or [for those who] throw any filth into it: the well of S. Sophia [opens] into it and many wells in the Seraglio.” He gives a diagram plan, showing two rows of eight piers and a third row of three, although, as no boundary is shown, it is impossible to say if this is the whole extent (see below).[327]