Shâh•tokhtaning•köli oot—oo3. Three pottery frs.; coarse hand-made ware of ill-levigated clay, hard-fired on an open hearth ; red to blackish grey. All c. 2â" x I x*" to i".
Shàh•tokhtaning•köli 004. Pottery fr.; coarse handmade ware of ill-levigated black clay, covered outside with (decayed) green glaze. a" x xt" x A".
Shfi.h•tokhtaning•köli cos. Pottery fr. ; hand-made like 001-003, but of less gritty paste. xi"x I "xc. i".
Shâh.tokhtaning•köli oo6. Terracotta spinning. whorl, orn. round outside with single line of incised dashes between two border lines, and on lower surface round edge of hole with small incised triangles and circles alternating ; nearly half broken away. Diam. iTle, thickness
SECTION II.—THE RUINED FORT OF PtIIRAN
In Chapter X I have already recorded the impressions of my first rapid survey of the Mirân Site and mentioned the interesting finds which rewarded the trial excavation commenced at the ruined fort. To this I moved out on the morning of January 23 and had our camp pitched under its walls, in order to be near our work and to secure what shelter they might give from the icy winds which rarely cease sweeping the bare desert glacis at this season. Then the men, nearly fifty in number, were promptly set to work to continue the systematic clearing of the interior where our first excavation had stopped a month and a half earlier. It did not take long to get proof that the ruined fort was likely to fulfil the promise held out by our former experimental digging. But before I proceed to describe the resumed operations and their abundant results, it will be convenient to give an account of the structural features of the ruin, as they presented themselves before excavation.
The ruined fort of Mirân, marked M. 1. in the site plan, Plate 29, rises on the stretch of bare pebble-covered plain Which extends, with a total width of about three miles, from the belt of vegetation watered by the present course of the Mirân stream eastwards to an ancient river-bed, now completely dry. This bed passes not more than half a mile to the east of the fort, and at the time of its occupation is likely to have carried water. Rather nearer on the north begins the belt of tamarisk-cones and scrub which edges the ancient site all along that side and stretches its sombre expanse for miles away. In Fig. 114, which shows the northern portion of the interior of the fort, it is seen filling the background.
Imposing as was the appearance of the ruin with its walls still rising to a considerable height above the flat ground, yet its irregular shape and construction had struck me from the first as features indicative of a non-Chinese and comparatively late origin. As Plate 3o shows, the ground-plan is that of an irregular quadrangle with the two longer sides facing approximately east-north-east and north-west respectively, and the other two facing roughly west-south-west and south-south-east. For the sake of convenience I shall designate these wall-faces hereafter as the east, north, west, and south sides of the fort enclosure. The two first wall-faces measure about 240 feet each, if we reckon from the outer foot of the adjoining wall-curtains. The west and south faces are considerably shorter, being respectively about 168 and 200 feet in length. Projecting oblong bastions of irregular shape and varying dimensions protected the corners, as seen in the plan. But these were too badly decayed for any approach to exact measurement. The south-west and south-east corner bastions, which are seen in Fig. 113, still stood. to heights of about 22 and 24 feet respectively. That at the north-east corner rose to 28-29 feet.
In addition, there was a bastion projecting from the central portion of each wall-face, but the exact position of these central bastions varied greatly, and their size still more. The one to the