The drain, which furnished us the clue, is hewn in the top of the rock for a long part of its course. It is 2 feet wide at the top, 1 at the bottom (where there is a groove), and 4 feet high. From about the point E' the floor falls both to the east and to the west. At various points it is fed by smaller drains from the city to the north. Beyond the gate it falls rapidly to the south-west, and comes to an end at right angles with the valley, at a point above a steep pitch, where it poured its filth into the valley of Hinnom. Immediately under the point where we later found the gate the sides were seen to be, not rock, but well-chiselled slabs of stone. At the time it occurred to us that this might be the point where it passed under the wall, but finding the scarp further out, we gave up the idea, only to find it the true one later on. The drain is roofed with slabs, which form part of the pavement of the street. One of these slabs was carved with a large Jerusalem cross, showing that it was used and repaired in Crusading times. It was completely choked up with rubbish, not ordinary earth, but actual sewage. At the point E' it is only 4 feet under the surface of the ground, which, however, I understand has been recently levelled down.
The pavement of the street always shows the sign of wear, and was clearly trodden by feet. The actual pavement as seen is not more than 10 feet wide, but at the two points measured the rock has been levelled down at its side, adding 8 or 10 feet more to the width of the road; the flat rock here also showing signs of wear. At other points the pavement may be wider. At the point C, in following the pavement, we had to break through the walls of a house, evidently of Byzantine times, which had been built over the disused pavement. This house had a mosaic pavement of its own, 212 feet higher than the street, and its walls were covered with plaster over 2 inches thick. We followed the street, all through the length of the house 1212 feet), broke through the second wall, and continued along the road—pavement. A few feet from the gate the flagstones disappeared, but I have shown that the road could have not turned anywhere else, and, indeed, it had been pointing towards the gate for a length of 230 feet, sure proof of its destination. I give the elevation of the wall from the gate at B' to the point where it joins the scarp of the fosse at A'. The dressed masonry does not rest on the rock, but on rough rubble built on the rock. The base of the wall may be seen to rise rapidly. The stones have smooth faces, are dressed with the comb-pick (without draft), and the point of jointure is so fine that sometimes it is difficult to find it. The top of each layer is perfectly horizontal. In other words, the workmanship is exquisitely careful. Between a and b (on the elevation) below the regular lower course another finely dressed course projects a few inches. Beyond the point a the dressed stones cease, but the course of the wall may be traced to the fosse along the line of rubble.
From the gate we also traced the wall along its inside face for some 25 feet. The stones at the corner were well dressed, but beyond, the masonry was rougher, as is natural on an inside face. The width is 9 feet.
