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

Sec. iii]   THE WALL BETWEEN YO-MÊN AND YANG KUAN   697

to dig even near it. I could not feel altogether sorry for this, as continued veneration of the spot seemed in itself an additional archaeological proof for the identification that I have now been able to record in detail.

SECTION IV.—T. xiv. A AND THE REFUSE-HEAPS OF T. xv. A

From the site of the Jade Gate now determined we may return to the line of wall and watch- Limes wall towers and resume its survey from the point where we left it by the side of the great marshy T:W` of depression to the north-west of T. xiv. Beyond this, I found the wall reappearing on the north-

eastern edge of the marsh and continued in the direction from south-west to north-east across a pro-

jecting tongue of the gravel-covered plateau to the edge of a smaller depression which contained a winding lakelet, seen in Fig. 189. Along this section, about three-quarters of a mile in length,

I could trace no remains of a tower or station, but the wall, built with reed fascines in the usual

fashion, still rose in places to 5 or 6 feet in height. Where it descended towards the edge of the marsh on either side of the plateau, it had decayed into a low mound of earth, which could be distinguished

only by its straightness amidst the scrub and reeds fringing the open sheets of water. On the north-

east the traces of the wall stopped short at about 4o yards' distance from the small salt lake and

about 15 feet above its level at the time. To the south-west, in the wide depression, I found

the earth mound extending for about a quarter of a mile from the foot of the plateau to where the

soil became actually boggy.

The salt lakelet just referred to is separated from a much longer depression eastwards by Clay ridge

another narrow tongue or ridge of clay which further north breaks up into a large cluster of eroded bearing

terraces. Fig. 187 shows these as seen from the south, together with the tower T. xiv. a perched T. xtv.a.

on the ridge to the right. The wall crosses this ridge where it narrows to a neck only about 40o yards wide. On the south-west the wall, with its layers of reeds still quite distinct though

curiously contorted in places, as Fig. 189 shows, descends the steep slope of the ridge to its base,

some 6o feet lower, and could be traced to within about loo yards of the shore of the lake. The level of the lake was then about zo feet lower, but a further stretch of wall, decayed here into

a mere low mound, may well be covered by the thick scrub. To the north-east the wall runs down to the gravel-covered edge of the winding basin already mentioned, and was traceable over this to within about 150 yards of the limit of the nearest marsh as marked by a recent crust of salt-cake fringing open water. The traces of the wall ended about 1 2 feet above the water's level. The view in Fig. 188, taken from near the tower T. xiv. a, shows a portion of this marsh-basin where it extends to the north. It is a characteristic glimpse of the desolate landscape past which the Su-lo Ho runs its terminal course westwards. The bed of the river forms a deep-cut fosse, quite distinct from the marshes that I have described from T. x onwards. It is so well hidden from view that on my first reconnaissance from T. xiv. a I completely failed to realize its existence, though approaching it to within a mile or so.

The tower T. xiv. a was situated about 5o yards from the western slope of the ridge and, Remains of

though much decayed, still rose to about 15 feet in height. Its base (Plate 39) measured close Watch-tower

T. my. a.

on 24 feet square. The brickwork showed close resemblance to that in the towers T. xi', T. xi'. a, T. xlIi, with layers of reeds after every fourth course. On the top of the tower were remains of a small conning room, about 8 feet square, in which the curious wooden hoe, T. xiv. a. ii. ooi (Plate LII), was found. From a small apartment adjoining the east foot of the tower came the interesting clay sealing, T. xwv. a. i. 001 (Plate LIII), with four Chinese characters in angular

seal writing, which have been read by Mr. Hopkins as Tun, ftao yin-lsin" qi   seal of the

1974   4 u