trough—at least ioo miles in length by 3o in width—which appears to have received either the whole of the present water-supply [of Sistan] or the overflow of the old and greater flood ; otherwise it is impossible to account for its vast area. The Shila runs in a briny stream when there is a large accession to the lake." Yate, writing of a journey made about 1894, speaks of the Shila (p. 98) as being from 150 to 200 yards wide, with precipitous banks 3o to 4o feet high, where he first saw it. There was no sign of a running stream, but merely pools of bitterly salt water. Thirty miles farther downstream he found (a, p. 99) the Shila " 150 yards in width, with sloping banks some zo feet high and full of sand, there having been no flood-waters down for the last five years." Below this point (p. 102) the banks of the Shila gradually decrease in height and finally die out on approaching the God-i-Zirrah. From these two descriptions the outlet appears to be through a broad, deep channel, which could only be cut by a strong, vigorous stream far different from the trickling overflow of the rare floods which now traverse it. Smith, however (a, p. 254), describes the Shila in very different terms as "a low, shallow sort of ditch or canal, about thirty yards wide, and quite dry." As he crossed the channel near the places seen by Sykes and Yate, there seems to be no way to reconcile the opposing statements except by supposing that the outlet, like the tributaries, is terraced, and preserves the record of two periods of overflow corresponding to the two elevated shorelines which will presently be described.
The lowest portion of the basin of Sistan and the final resting-place of the waters which escape via the Shila is the God-i-Zirrah, or Hollow of Zirrah. McMahon (a, p. 19) describes it as "a large lake of clear, deep-blue water, some 25 miles long and 5 miles wide, standing in the midst of a wide margin of solid salt. It used to be fed by flood-water from the Helmund, but it now seldom receives any replenishment. The last time it received any water from the Helmund is, so far as I can ascertain, as long ago as 1880, i. e., seventeen years ago. [Yate, p. 105, says 1885.] All the drainage which, in the natural course of events, should flow into it from the mountain ranges south of it is intercepted and swallowed up by the wide barriers of sand lying in the way and thus never reaches it. Its water is now so salt that even water-fowl avoid it."
A glance at the map shows that the Shila and the God-i-Zirrah form an arc parallel to the arc of the Helmund and sloping in the opposite direction. In ancient times, according to Ishtakhri, who wrote in the tenth century (quoted by Sykes, a, p. 365), the Helmund was diverted so as to flow across the neck between the two arcs and discharged directly into the Shila. The lake of Zirrah was too miles long and covered an area ten times as great as to-day. It was filled to overflowing, apparently, and the direction of the Shila was reversed, for the people of Sistan have a legend that in former times the supply of the Hamun-i-Sistan entered at the southern end of the lake instead of the northern. Part of the water of the Helmund, however, still flowed to Sistan along the present course and watered the numerous villages the ruins of which crowd the upper part of the Helmund delta. In course of time this main channel regained all the water, perhaps, as McMahon (a, p. 20) suggests, because the Hehnund deepened its bed below the mouth of the canal which leads to the Shila. To-day that canal still exists, and until a hundred years ago