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

Southern main group of caves.

Varying tiers of excavations.

792   THE CAVES OF THE THOUSAND BUDDHAS   [Chap. XXI

the whole way passes over gravel fans of the same uncompromising sterility, I may refer to my Personal Narrative.3

After less than a mile from the point where the cart-track from Tun-huang turns into the silent valley, here still open, the first grottoes come into view, marking the northern end of the sacred site (see Plate 42). They are cut, like all the rest of its shrines, into the almost perpendicular conglomerate cliffs lining the western edge of the wide sandy bed where the stream descending the valley finally loses itself through evaporation, except on occasion of rare floods. The multitude of dark cavities, mostly small, belonging to this northernmost group of shrines honeycomb the sombre rock-faces in irregular tiers up to a level of about 5o to 6o feet above the bed. There the lowest and most precipitous step of the cliff gives way to easier sand-covered slopes. Much of the rock-face that once contained approaches to the various cave-chambers and the passages between them has fallen and crumbled away completely. This is partly due to the erosive action of the wind which sweeps up from the north, and the slow undercutting by the stream which washes here the very foot of the cliff. Along the length of close on 500 yards over which the grottoes of this group extend no trace has survived of wooden galleries and stairs such as must once have served to facilitate approach and communication. The same is the case at a second and smaller group of caves, which is found about 15o yards further up and is shown in its full length on the right of Fig. 193. At both these groups, I may state at once, want of time and difficulty about improvising means of approach prevented close examination on my part. But a variety of indications suggested later origin,3a and from the small size of the -majority of the recesses and the absence of any wall-paintings in most of them it seemed safe to conclude that they had served largely as quarters for Buddhist monks. Their sombre aspect and setting recalled pictures of troglodyte dwellings of anchorites in some western Thebais.

It is very different with the southern and main group of caves, which extends along the face of the gradually ,rising hill scarp for close upon a thousand yards. The panoramic view, Fig. 191, shows almost its whole length, as seen from the gravel plateau across the rubble bed of the stream to the south-east. But the groves of fine elms growing on the cultivated strip of fertile alluvium which stretches here between the foot of the cliff and the bed of the stream, as marked in Plate 42, help to hide in this photograph most of the cave-shrines occupying the lower face of the cliff. Even on closer approach it is difficult to obtain any clear view of the general character and arrangement presented by this wonderful agglomeration of cave-temples ; so bewildering is their multitude and the diversity of their disposition.' In the obvious absence of any systematic planning for the whole, and in view of the difficulty which any attempt at definite grouping must present, I think that it will best serve the purpose of this general introduction to the site if I reproduce here briefly the impressions received on my first visit.

Along the whole length of the scarp of the hill, from below the position marked on Plate 42 by Ch. 11, on the north, to above Ch. xvi on the south, the precipitous portion of the rock face shows an unbroken succession of grottoes. Some high, some low, they are all closely serried

Lower groups of grottoes.

s See Desert Cathay, ii. pp. 21 sq. The approach to the site and its general aspects are quite correctly described by the auThor of the Tun-huang lu, transi. Giles, J.R.A.S., 1914, pp. 707 sqq.

3a The Tun-huang lu, J.R.A.S., 1914, p. 709, estimates the distance of the cliff face occupied by cave-temples at z li. This corresponds very closely to the actual extent of the southern or main group of caves, taking the li meant at the value indicated by the same text's preceding estimate of the

distance from Sha-chou to the site at 25 li. Ih any case it may be inferred from this statement that the northern groups of caves did not exist when the text was composed about the close of the Tang dynasty.

' Rai Ram Singh in the course of his topographical survey of the site counted roughly 35o separate grottoes in the southern main group. To this number some 165 excavations, most of them quite small, have to be added in the two northern groups.