This description may be sufficient to account for the natural water-supply, which was always most abundant on the south-east, where the dolomite bed is nearest to the surface in the valleys. The principal spring is in the Kidron, below the steep eastern slope of the Ophel spur south of the Temple. It rises under the floor of a cave, where there must be an underground reservoir in the rock, resembling many in the Lebanon and in other limestone regions. Towards the end of winter, when the heavy rains have fallen, this reservoir overflows frequently through a fissure which acts as a natural syphon, sucking out all the water as soon as the reservoir is full. The sudden gush--like that of the Sabbatic River in Syria--occurs every few hours in early spring, but at the interval of several days in autumn. The stream originally flowed down the rocky bed of the Kidron, which is now filled in to a depth of 30 feet. But from early times it would seem that attempts were made to carry the water to the foot of the east slope of the upper city hill, in order to bring it nearer to the fortress. By the time of Hezekiah at least--as will be detailed later--a rock tunnel carried the waters of the spring to Siloam, or “westwards to the city of David.”[84] This statement--in consequence of the English mistranslation--has become the foundation of a literary theory according to which the city of David was a mere hamlet of 15 acres on Ophel, whereas in reality it appears to show that the stronghold of Jebus lay towards the west. It is not impossible that a yet earlier rock-cut channel existed, with the same object of conveying the waters of this intermittent spring towards the western citadel; and, as the point has some importance in connection with the history of the city, the reasons may be given more fully. [Sidenote: GIHON] Excavations were made in front of the cave in which the Kidron spring bursts forth, in the year 1902, and it was then discovered that a rock tunnel leads away towards the south outside the entrance to the cave.[85] The level of its floor is only 5 feet above the water-level at Siloam, and this aqueduct unfortunately has not been explored along its whole length, nor has it furnished any indications of the age in which it was made. It has been thought to be part of an old rock channel traced for 600 feet northwards from the old pool below the Siloam reservoir. This, however, is doubtful, as the channel in question rises rapidly, and the levels in consequence would oblige us to suppose that pipes must have been used, as water does not run uphill in an open channel.[86] This Siloam channel was still connected, in 1874, with a series of surface channels on the slopes of Ophel, which have been quarried away since, but which once carried the surface rain-water to the old pool.