National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

Nan-hu make it more difficult to ascertain with sufficient accuracy how far the great reduction Settlement
in the size of the settlement, as indicated by the ancient remains to be noticed presently, is reduced
due to that potent cause, desiccation, of which the general aspect of the ground would make us since earlier
think in the first place, and how far to circumstances connected with the human factor, i.e. the period:
available population, and local history. For the present I must content myself with recording possible
two conclusions of a general character. On the one hand, it appears to me certain that the water- causes.
supply at present available, over eighty cubic feet per second from all sources, would permit of
a far larger area being cultivated than is possible now with the labour of the thirty odd
homesteads. On the other hand, I feel inclined to doubt whether the agricultural resources thus
provided would by themselves suffice to account for such an extensive area of close occupation
in ancient times as the surviving 'Tati' remains indicate. The explanation may well be sought
in historical and geographical circumstances, which made Nan-hu a point of special quasi-strategic
importance during an early period, and which I shall have occasion to set forth presently.

SECTION III.—THE ANCIENT REMAINS OF NAN-HU

The only conspicuous remains of antiquity which survive within the basin of Nan-hu are The ruined
the ruins of a small walled town, known to the people merely by the name of Nan-hu-ch'êng, 'the town of
walled city of Nan-hu'. Its broken walls, reached within a mile eastwards from the edge of the Nan-hu.
present oasis and approached over scrub-covered ground once manifestly under cultivation, form an
irregular rectangle, as shown by the plan in Plate 35. Of the north face, measuring about 400 yards
in length, a considerable portion still survives, though half-buried under high dunes which have
helped to protect it (Figs. 159, 160). Of the somewhat closer east wall (on right in Fig. 159) and
of the west wall, too, portions are still extant to a fair height, though cut through and broken up by
wind erosion. This, with the abrading drift-sand close at hand as its instrument, can work here to
full effect. On the south, curiously enough, the wall has disappeared completely, though its position
was clearly traceable by the mound into which the clay rampart once bearing it had decayed. Yet
the wall proper was of very solid construction, being built with carefully stamped layers of clay
5 to 5½ inches thick, and seemed of early date. From fourteen to twenty feet thick at its base,
it still rises in places eighteen to twenty-one feet in height. Its foot rests on a broad clay rampart,
which seemed to raise it another twelve feet or so above the level of the ground in the centre. But
as the whole of the interior is covered with drift-sand bearing slight scrub, as seen in Figs. 159, 160,
neither the original level of the ground nor the real height of the rampart could be made out with
certainty. There was a much-decayed inner wall (Fig. 160, on left) on the north-west, marking
a small separate enclosure. Owing to the effects of erosion and the presence of drift-sand in the
gaps of the walls, the position of the gates could no longer be ascertained.
The interior contained no recognizable ruins, only some low mounds covered with drift-sand. Mounds
Being able to obtain a number of additional labourers from the neighbouring hamlets, I had trenches within
cut through these down to a depth of about five feet, where the men reached what seemed the walled town.
natural soil. But the only finds made here consisted of fragments of very hard burnt bricks, dark
grey in colour (for a specimen see Nan. Ft. 007), and two intact burnt bricks, also very hard, but of
a coarser clay and yellowish in colour; these measured 14 by 7½ inches, with a thickness of
4½ inches, being thus of the same size as that which prevails in the watch-towers of the Limes.
Like some large pieces of rubble, they all lay loose in the earth and not far from the top of the
mound, as if left over from some structure the materials of which had been quarried and removed.
Chiang Ssŭ-yeh. who, like other educated Chinese of antiquarian tastes, always showed a lively