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

Sec. iii)   HUA-HAI-TZÛ AND ITS LIMES REMAINS   391

The fair state of the remains of the wall at this point seemed to indicate the prevalence of Relics at

climatic conditions more favourable to the preservation of ancient objects than those met with on T. XLIII. a.

the right bank of the Su-to-ho. The hope raised by this circumstance, that I should here find

relics of the men who had kept watch on the Limes, was not disappointed. The first watch-tower

traced, T. XLIII. a, had, indeed, decayed into a mere gravel-covered mound of small dimensions.

But around it ancient potsherds abounded, and from a refuse heap discovered under a thin layer

of gravel, some 4o yards to the south, we recovered three wooden slips inscribed in Chinese ` grass '

script, besides several fragments of blank slips and the miscellaneous small objects in wood, wool,

&c., described in the List below. They include a wooden seal-case, T. XLIII. a. oio, of a type

represented on the Tun-huang Limes, and pieces of a strongly woven rug, dyed green, T. XLIII. a.

o6. Close to the mound a Wu-chu coin was picked up on the surface. [Two of the Chinese records

refer to convicts exiled to the frontier for military service. T. XLIII. a. 013 also mentions the ` signal

post ' of Chên-chung   I$!, the name of which recurs in T. XLIV. b. 2, 024.]

After following the line for half a mile to the south-east we came upon unmistakable remains watch of a potter's kiln. The ground was reddened by fire and covered with slag and potsherds. The T. 'cm!. b. piece, T. XLIII. a. i. 01-2 (PI. XLVIII), from the lip and side of a large vessel is a good specimen of this local ware. After proceeding another half-mile we came to the tower T. XLIII. b (Fig. 223), still rising to about II feet. It was solidly built of bricks, 15 by 8 inches in size and 5 inches thick, and originally measured 16 feet square at the base. This had been subsequently enlarged to 20 feet, as was clearly seen through a gap where the outer brick casing had fallen on the east side. Here the white plastering of the original structure was still visible. The outer masonry in bricks of the same size had its courses separated by layers of reeds. The top of the tower had once probably carried a small guard-room, but nothing was found there. Among the pottery debris recovered close by the fragments one or more bowls in a fine glazed frit, T. XLIII. b. 01-7, deserve notice.

The next two ruined posts, T. XLIII. C, d, were marked only by mounds, completely decayed, Decayed

with ancient pottery debris around them. At the second I was able to ascertain that the bricks of posts

T.   c, d.

the tower which once stood here were of the same size as at T. XLIII. b, and this was observed also

at other ruins of this section of the Limes. A curious feature at T. XLIII, d was a row of eight very low mounds, just distinguishable over the gravel flat, stretching to the south and separated from each other by distances of 3o to 5o yards. At each of these little mounds, layers of brushwood, slag, or clay reddened by fire could be traced under the thin cover of gravel which the winds had blown over them. Is it possible that these remains mark the position of hutments of an ancient camp ?

About a mile farther on, the site of an ancient watch-station, T. XLIII. e, was indicated by Remains

a mound, about 15 feet high and 22 yards across, composed of layers of clay and brushwood. of posts

T. XLIII, e, f.

Besides numerous potsherds of Han type lying about, there was found here a curious hollow

pottery bar, T. XLIII. e. oi, made of very hard dark grey clay and described in the List below ; its use still remains to be determined. At T. XLIII. f, less than a mile farther on, the line of the wall was seen to take a slight turn due east. Here, too, only pottery debris survived to mark the position of a post. For a distance of nearly a mile from this point the wall could be followed with ease over the bare gravel. Then it passed into an area of wind-eroded clay with low scattered tamarisk- cones. Here its line remained traceable for a short distance only, in the shape of a wind-eroded clay bank, 3 or 4 feet high. This obviously was a ` witness ', due to the fact that the soil which the agger had once covered had resisted _wind-erosion longer than the adjoining unprotected ground.

Beyond this the line of the wall was lost completely, until, at a distance of about three miles Finds at

from T. XLIII. f and on the northern edge of a belt of tamarisk-cones, we came upon a low mound

T. XLIII. g.