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

   Sec. viii]   REMAINING RUINS OF MIRAN   537

and a half inches, which is some two and a half inches in excess of the standard width of silk for later Han and Chin times, as ascertained above from the specimens of silk found at the Lou-lan Site and at the station T. xv. a of the Tun-huang Limes.I" The plain-weave texture of the piece, in the present state of our knowledge, does not allow even an approximate determination of its age. But from its general condition and the position in which it was found, well above the floor, I am inclined to conclude that this much-worn ` kamarband ' is a relic left behind by some later traveller who may have sought refuge in the structure while its roof was still intact.

The antiquity of the ruin, apart from the architectural features already discussed, was sufficiently demonstrated by the depth of the wind-erosion around the ` witness ' on which it stood (Fig. 132). Deep-cut erosion trenches separated it from three or four other Yârdangs lying close by to the south and east. These evidently marked the original extent of the ground once occupied by a group of buildings. One of these Yârdangs, about 8o feet to the east of M. x, bore the remnant of a structure built of sun-dried bricks and apparently 16 feet square. The bricks were of the same size as at M. x, and the north wall, still clearly traceable, had a thickness of three feet. The wind-eroded trough running close to it with almost vertical banks had its bottom fully 17 feet below the original ground level as marked by the wall.

This excessive effect of erosion, clearly seen in Fig. 132, at first caused me surprise, considering that the ` Sai ' around had almost everywhere a surface of fine gravel. But a closer examination of the soil laid bare on the slopes of these Yârdangs showed that this surface layer of gravel was very thin, and that the soil beneath was composed of strata of soft alluvial clay, interspersed only at rare intervals with a scant admixture of fine pebbles. The surface layer which had so far formed over the ground from this gravel was not sufficiently thick effectively to protect the soil beneath from deflation. On the contrary, this gravel seemed to act as a corrasive agent, facilitating erosion wherever the configuration of the ground particularly exposed it to being lifted and driven along by the violent north-east winds. The resistance offered by structures occupying exposed positions would create wind eddies and thus intensify the local erosive action of the winds.

The condition of the soil on the surface was very peculiar also in the wide area of tamarisk-covered cones which, as the site-plan, Plate 29, shows, spread away to the north of the ruins so far described. In that area, too, all the ground that was left bare between the thick-set tamarisk-cones was covered with a thin crust of light gravel. But immediately beneath this there came a thick layer of fine dust, apparently disintegrated alluvial clay, into which the feet of man and beast sank deeply. The effect curiously resembled that of moving over treacherous ground where sand is held in suspension by water. Here, about a mile to the north-north-west of M. II, I was shown by Tokhta Akhün, my Loplik guide, the last ruin we were then able to trace at the site. It was that of a

Pao-t`ai-like tower, M. xII, about 18 feet square at its base, and solidly built of sun-dried bricks measuring 18-19" x 9-10" x 4". The masonry, slanting slightly inwards, still rose to about 12 feet

above the ground, which had here suffered scarcely any wind-erosion. The tower must, of course, have been constructed before the maze of sand cones covered with tamarisk scrub grew up around it, and their height, reaching here up to 16 feet or more, sufficed to attest its antiquity. Like a similar tower, discovered in 1914 some two-thirds of a mile further north, it may have served the purpose of a look-out station.

It was obvious that on such very deceptive ground as this area of tamarisk jungle other remains might escape attention, and, in fact, when revisiting the site in 1914, I found two more small ruins

1R See above, pp. 373 sq. with note 13 a. The width of the Mirân piece agrees with that of the silk used in the painting Ch. 0067 from the cave-temple of the Thousand Buddhas,

   1174   3

which was walled up at the beginning of the eleventh century ; see below, chap. xxv. sec. ii.

z

Wind- eroded terraces at M. x.

Soil eroded under gravel surface.

Peculiar surface of ground.

Tower M. XII.

Chronology of site.