84 REMAINS OF KASHGAR, OASES OF YARKAND AND KARGHALIK [Chap. IV
receding square terraces of the Stûpa base ; the begging-bowl turned upside down the dome ; and the mendicant's staff the pinnacle, which, mostly of wood and fitted with Chattras or umbrellas, adorned all Stûpas.
It would be an interesting task to trace the regions and periods of Buddhist worship in which the orthodox system of Stûpa construction prescribed in this legend actually prevailed ; but the time and space available for the present work does not permit of such inquiries. Nor does it appear to me safe, in view of the very limited number of Turkestan Stûpas still sufficiently well preserved for accurate restoration and approximately datable, to draw any conclusion at present as to the development of their main architectural features. It may be assumed that in Turkestan, as in India, the strictly hemispherical dome, so aptly illustrated in the above legend by the begging-bowl turned upside down, represents the oldest shape adopted for relic memorials, and that bulb-shaped domes with exaggerated drums are later. But only a detailed study of all Stûpa remains and Stûpa models still extant on the Indian north-west frontier and beyond could furnish reliable indications as to the chronology of these changes, and even then the data deduced could not be applied with implicit assurance to monuments in widely separated regions.
About 200 feet behind the Stûpa, to the north-west, rises a great oblong mound (see Pl. I), consisting of solid masonry of sun-dried bricks, to a height of 34 feet above the same low gravelly ridge which bears the Stûpa. The whole exterior is far too decayed to permit of an exact reconstruction of the plan and dimensions ; nor can its original character and purpose be determined with certainty. Three stories, with traces of cells or large niches built into the brickwork, are roughly recognizable on the south-east face, which here, too, seems to have suffered least from exposure. The plan shows that the length of the structure from south-west to north-east was about 86 feet on the ground-level. Its original breadth seems to have been about 71 feet on the same level, but is more difficult to ascertain, even approximately, owing to the accumulation of débris in front of the south-east face. This is possibly due to the former existence here of a series of platforms or flights of stairs leading up to the higher stories.
It would be difficult to form even a conjectural opinion as to the original purpose of this structure, were it not for the indications furnished by some better preserved ruins of a similar type with which the investigations of Russian and German scholars in the vicinity of Turfan have recently made us acquainted. These present themselves as square structures, displaying on each face a series of vaulted niches, arranged in a succession of receding stories, which still contain remains of large stucco statues of Buddhas or Bodhisattvas. The brief description of these ruined buildings, which I take from Dr. D. Klementz's report on his preliminary survey of antiquarian remains at Turfan 14, together with the illustrations given there of two of them, situated at Astana and Syrkheb, show plainly that their outlines resemble those of a truncated pyramid, suggested by the present appearance of the Mauri-Tim mound. Their dimensions, too, seem in several cases to approach closely those recorded for the latter 15. Of the chamber or hall which two at least of the Turfan structures contain in their interior, I failed to trace any indication in the case of the Mauri-Tim ruin. Yet I have little doubt that it was built for identical purposes of worship. These quadrangular structures, with their rows of Buddhist