Sec. i] THE RUINED FORT. ON MAZAR-TAGH
more than about 300 feet above the absolutely barren waste of high dunes on either side, its isolation in the midst of the Taklamakan desert makes it a very striking and geographically interesting feature. Its bearing and the similarity of its geological structure to that of the chain of rugged isolated hills rising near Maral-bashi and Tumshuk on both sides of the Yarkand River justify the belief that the Mazar-tagh of the Khotan-darya forms the last remnant of an ancient mountain system, which jutted out south-eastwards into the Tarim Basin from the outermost chain of the Tien-shan subsequently visited on my way from Ak-su to Maral-bâshi. That it is the never-ceasing wind-erosion of countless ages which *has reduced this remnant to its present insignificant dimensions appears to me more than ever probable since the observations I was able to gather in 1913 on my attempt to reach the Mazar-tagh from the side of Maral-bashi.1
So striking a natural feature as this hill range rising abruptly in the midst of the desert was bound to attract local worship at all times, and this explains its modern designation as the ` Hill of the sacred shrine'.? I was, therefore, prepared for the conspicuous collection of poles bedecked with rags and other ex-votos which crowns the end of the ridge where it falls off with precipitous cliffs towards the left river-bank (Fig. 335 ; see site-plan, Plate 59). But my satisfaction was greater when, ascending to the crest of the ridge about a hundred feet higher, I found it occupied by the ruins of a small and relatively well-preserved fort. Its high walls standing out clearly defined against the sky had, when seen from below at a distance, looked almost like some rock formation ; for the total want of covering detritus or sand on the steep slopes has given the appearance of natural walls, terraces, or steps to the denuded sandstone strata which crop out at a sharp tilt.3
The crest of the ridge on which the ruins stand is very narrow where it approaches the river to the east, and even where it culminates in the small plateau bearing the detached watch-tower seen in Fig. 329 its width is only about 3o yards. To the south the ridge presents a very precipitous face, practically unscalable near the fort. The slopes to the north are easier ; but these too are completely commanded by the fort and the tower. A second ridge shown in the site plan, which runs parallel to the main one for a considerable distance, is both lower and beyond the range of bow-shot. The position thus occupied by the fort was naturally very strong, altogether inaccessible to attack from the south and east, and protected from the west by the tower already referred to. Apart from this outlying and very massive tower, the fort was defended on the north-west by a wall stretching across the rocky crest and by two bastions 16 feet square, projecting at either end and built solid (Fig. 329). These bastions and the curtain between them were constructed of flat pieces
of clay set in mortar, with layers of tamarisk branches between the masonry courses at intervals of to inches. In addition, posts and beams of Toghrak wood inserted within served to strengthen the masonry, both here and in the walls, fully io feet thick, which enclosed the fort proper on the south-
east and north-east. The latter walls, as well as those of the inner structures and the court v, were built with coarse bricks, measuring 15" x 8" x 3â~~•
By filling up the space within the walls a level area was created over the naturally steep north Interior of slope to serve for the interior of the fort. This comprised the inner court iv, measuring 5o feet fort. square, which was found covered with the débris of less substantial structures likely to have served. as quarters. Abundance of charred timber suggested that these may have been constructed mainly
Fort crowning end of ridge.
Position of fort.
Construction of walls and towers.
' See Third Journey of Exploration, Geogr. Journal, xlviii. pp. 113 sq. The application of the name Mazâr-High also to the isolated rocky massif rising island-like to the south-east of Maral-bâshi (Map No. 15. c. 7) has, of course, no bearing whatsoever on the orographical nexus above indicated; see the next note.
2 Regarding the folk-lore beliefs which cause isolated hills
to be looked upon as sacred and chosen for the location of shrines, or Mazârs', like the Svayambhû Tirthas ' of India, cf. Ancient Khotan, i. pp. 313 sq. ; also below, pp. i 3o2, 1310.
The photograph in Fig. 332, taken far away at a desert hill rising from the-plain near Marâl-bâshi, curiously illustrates also the formation of the rock strata as they appear on the north slopes immediately below the Mazâr-tagh ruins.