In the second place, we must remember that the walls must have run on the highest available ground, in order to give advantage to the defenders over their enemies outside. This is invariably the arrangement of a fortified town in any age, and it is impossible to suppose that ancient engineers--any more than those of our own time--would build walls in valleys, leaving high ground immediately outside, where the towers and engines of the besiegers could be placed in positions commanding the town within. At Jerusalem the wall had to be carried across the head of a narrow valley on the north side, down which valley ran the street west of the Temple leading to Siloam. At Siloam also the wall had to be carried across a wider valley--the south part of the Tyropœon; but, unless it was desired to enclose that pool, it would here be kept as high as possible on the ground above the pool to the north. It is certain--at all events in the time of Josephus--that the Pool of Siloam was outside the wall; but, since it was flanked by scarps within easy bowshot, it would be sufficiently defended if the wall was built on these scarps. The same consideration makes it certain that the wall defending Jerusalem on the north-west and west must have stood on the higher ground which defends the lower city on two sides. It would therefore join the wall of the upper city near the north-west corner of the latter, and would run on the narrow saddle, or neck of high land, which separates the heads of the Tyropœon and Hinnom Valleys. No other position can be conceived, since if it began east of this saddle, it would have stretched through the Tyropœon, leaving the saddle outside with a command of at least 50 feet above the base of the wall. Farther to the north-west the wall must also have enclosed, or run over, the high knoll of rock which was shown later as the site of Calvary--a rock which is nearly as high as the level ground of the upper city, and which formed the natural defence of the lower city which lay in the Tyropœon Valley. The wall might have run farther to the west and north, where the rock is close to the present surface, but it could not run farther east or south without leaving high ground immediately outside the fortress; for the north slope of the broad Tyropœon hollow sinks very rapidly south of the knoll now shown as Calvary, while not far to the east of this knoll it also fell about 50 feet to the confluent valley coming from the present Damascus Gate. [Sidenote: THE GATES]