a palm leaf and a birch-bark manuscript. Insignificant as the written remnants are, they may yet help to determine the origin of manuscripts from Kuchâ which have been purchased at different times by the Kashgar Consulate General and otherwise for the Indian Government collection, and which were formerly in Dr. Hoernle's charge.4
Farther to the north, and some 5o feet lower down, lies the cave-shrine Jig. II
(Pl. 44). It comprises a cella, 17 feet wide and over to feet deep, with a passage running round the sides and behind the central portion of rock ; this once held a stucco image in a niche. The front of the cella has fallen in and with it most of the plastered ceiling. What remained of the latter was decorated with a painted lozenge diaper, now much injured ; like the traces of a seated Buddha painted in the image niche, it suggested old and good work. The clearing of the cella and passage yielded no finds. About 6o yards to the north and on a slightly lower level there is a large vaulted cave, about 31 feet wide and over 21 feet deep in its present broken condition. It looked like a hall, intended perhaps for monastic gatherings. Its front had fallen in, and masses of fallen rock covered the floor to a considerable depth. About 5o yards to the north there lie two caves now communicating with each other through an opening, apparently recent, in the thin rock partition which divides them. The roughly cut rock walls were blackened with smoke, and the layer of earth and straw refuse covering the floor showed signs of having been repeatedly burrowed into. The larger of these caves, about i8 feet wide and 12 feet deep, looked like a natural recess of which the walls had been smoothed. Through a hole in the floor, which had partly given way, a similar recess, narrowing tunnel-like westwards, had become accessible.
On the eastern side of the valley the caves were all small, and had suffered from the water and mud brought down by occasional rain from slopes of decomposed rock. The detritus that covered these slopes probably hid some excavations altogether, and made the ascent to some of the higher caves distinctly difficult. The best preserved of these was a small but carefully cut cella, only 5 feet by 6, situated about 8o feet above the brook. I found its floor completely cleared. A troublesome climb from this point towards the northern end of the spur, seen on the extreme left of the photograph, brought me past a completely collapsed cella to the narrow crest. On this, at an elevation of some 200 feet, I found two caves, for the most part filled with deposits of mud ; their plan proved that they had served as monastic quarters. One of them, Jig. iii, shown in Pl. 44, had its entrance at an inside corner, to which a rock-cut passage gave access. It was cleared without anything being found. On a small terrace below it, I noticed remains of charred wood, with the clayey soil below burned red, apparently indicating that a small shrine, built of timber, had once stood here and had been destroyed by fire. Pottery debris was to be found all the way up this eastern spur, a sign of prolonged occupation.
4 For an inventory of MS. fragments in Sanskrit and Kuchean, see Prof. Sten Konow's Appendix F ; also Mr. Pargiter's Appendix E.
Apart from the MS. remains, the following small objects were found here :
Jig. z. i. oi. a. Fr. of oblong wooden tablet, one end
rounded, with hole through it ; other end broken off. Obv. blank. Rev. somewhat sand-encrusted. Traces only of
one 1. of Brâhmi (?) script along one edge. Wood hard. 6" x 3I"xI„
Jig. I. i. ox. b. Fr. of wood, covered each side with layer of clay over which appears to have been a layer of fine plaster on which were written chars. ; but surface almost entirely destroyed, and chars. (if such) the merest fragments.
21-g" x i 6" x (max.) r'b".
Jig. z. i. 05. Fr. of iron or steel (?) knife-blade ; one-edged, straight, with tang, of type Kao. ni. 0172, Pl. LXXI. Blade broken off i}" from tang. Back, fiat. Fair condition. Length with tang 2", gr. width of blade i".
Jig. z. f. 06. Fr. of ornamental bone (?) plate,
apparently formerly riveted on to some flat surface. Long and narrow, with one edge straight, and cut down at right angles to face. Other edge (bevelled) at first straight, then cut in long tongue-like curve to join other in blunt point. Other end broken off, but may also have been origin. pointed. Two holes drilled through for attachment to surface below, a short black horn rivet still remaining in one. Length 2 i", gr. width I".