| || || |
424 THE ENDERE RUINS [Chap. XII
a mass of sun-dried brick and plaster, the curious shape of which the detailed plan will explain. The statues once occupying each of its four main faces had completely crumbled away, except on the west and north, where the feet with parts of the draped robe below the knees survived. As the feet measured in each case i ft. the statues must be supposed to have been about life-size. The arrangement of the drapery remains and the shape of the stucco backing suggested that the statues were seated. There was nothing to indicate what particular Buddhas or •Bodhisattvas they represented. The work was in relief much lower than that of the corner images. Remains of oval halos reaching down to the level of the feet could be distinguished on the plaster background. It appears probable that, for purposes of lighting, the cella was provided with a raised roof in the centre. With the decay of this the central group of stucco sculptures would necessarily have been much exposed to atmospheric influences, which may, together with their raised position, account for the scantiness of the remains.
The east and south-east facets of the base were found to have been completely destroyed by an old excavation, undoubtedly the work of treasure-seekers, which had been carried towards the centre, but could not be wholly cleared of sand from risk to the extant structure. Of the way in which this burrowing affected the MS. finds made in the shrine I shall have to speak presently. It also caused the almost complete loss of the fresco paintings with which the portion of the base facing the entrance appears to have been decorated. On the north-east facet the fresco remains showed two rows of seated male figures, seven in each, apparently Buddhas or Bodhisattvas. The colours had badly faded, and only the outlines could be made out.; even these were effaced in many places. In spite of this poor preservation the work seemed distinctly superior in drawing and execution to the corresponding type of wall-decoration in the DandänUiliq shrines. Of the fresco decoration of the eastern facet I fortunately recovered an interesting fragment in the small piece of painted plaster (E. i. oz 2), which was found lying in
the loose sand filling the old excavation made in front of the base. Its careful reproduction
in Plate LXXIX shows the delicate and harmonious colouring which the little fresco piece retains in spite of its faded condition. In it the head and shoulders of a figure surrounded by a green nimbus and probably seated can be clearly made out, with portions of the robes of other figures. Mr. Andrews' description in the list below indicates interesting technical details as to the preparation of the stucco ground.
The hopes of MS. remains which the first afternoon's experimental digging had raised were amply realized by the numerous finds made in the course of the subsequent excavation. These finds derive particular interest, not only from the variety of languages and texts represented among them, but also from the curious conditions in which the numerous MS. portions were recovered. That the three leaves (E. i. 2) of a Buddhist text in Sanskrit, discovered on the first day about i ft. above the floor of the east side of the central base, had belonged to a larger Pôthi was at once clear to me, from the pagination numbers on their margins and the string holes which showed them to be left halves. I was encouraged in my hope for the recovery of more of this text when on the following morning there turned up towards the north-east, and also not much above the floor, two more packets of broken folia (E. i. 4, 5)• containing between them 12 left halves and 4 right halves of the identical MS. Finally, when after excavating the rest of the cella the portion between the east side of the central base and the entrance could be completely cleared of sand, there were discovered in the cutting about 2 ft. deep which treasure-seekers had made into the floor of that portion two more packets of the same text (E. i. 39, 4o), containing 8 left and 19 right halves of folia, besides 2 detached half-leaves (E. i. 41, 43). From a comparison of the total number of left and right half folia,