adornment of the Stûpa base, and to prove that the wall of the circular passage was decorated with a frescoed frieze and dado.
The Stûpa (Fig. 129) was built with sun-dried bricks, of the same size as those composing the walls, and covered with a thick layer of hard white plaster. Its circular base, of which the elevation is shown in Plate 32, measured 124 feet in diameter on the level of the floor. Up to a height of close on 7 feet it had an elaborate series of well-designed mouldings, of which the two most prominent, together with the straight-edged plinth at the foot, may have been intended to represent the traditional three stories of a Stûpa base. The foot of the Stûpa had been dug into from the east in early times, no doubt with the hope of finding ` treasure ' ; for heavy débris from the vaulting blocked the approach to this cutting. The badly broken condition of the Stûpa prevents any estimate of its original height, and for the same reason we are left without any indication of the height of the dome which must have once risen above it and covered the cella. The dimensions of the Stttpa base and the circular passage around it, however, show that this dome must have had a span of 262 feet. It is much to be regretted that we have no means to judge of the method of vaulting likely to have been used for the dome. A comparison of it with the constructive methods employed in existing structures, of approximately the same period in Syria and other parts of Western Asia, would have offered considerable interest. From the fact that some small fragments of coloured plaster turned up in the débris quite close to the top of the Stûpa remains it may be concluded, with some probability, that the interior of the dome was also decorated with tempera paintings just like the cella walls.
The discovery of several fragments of fine wood-carving in the top layers of the débris filling the circular passage affords interesting evidence that the vaulting of the cella must have been high enough to allow space, not merely for the Stitpa itself, but also for a ` Tee ' or superstructure, obviously in wood, carrying that succession of Chattras, or umbrellas, which always surmounted the Stûpas of Gandhdra, just as we still find it over the actual Pagodas of Burma and other Buddhist lands.' It is to the staff-like support of such Chattras that I should ascribe a thick and badly broken piece of timber which was found in the eastern portion of the circular passage, at a height of about five feet above the floor. Two square holes passing through the middle, and at right angles to each other, were evidently intended for wooden cross-pieces likely to have carried an umbrella, perhaps modelled in plaster. It is to the decoration of the square pedestal or crown intervening between the Stûpa dome and the ` Tee ' proper, as seen in many of the small Stupas from Gandhâra,2 that I am inclined to assign the excellently carved fragment of a small wooden capital and shaft, probably belonging to a pilaster, M. v. ooi, which is reproduced in Plate XXXIV. The type of acanthus which decorates the front and sides of the capital is plainly Hellenistic and frequently met with in Gandhâra relievos.' It is also represented on the painted lintel of the gate seen on the south-eastern portion of the frieze of the circular passage (Fig. 134).
The comparison of this interesting piece with the fragments of decorative wood-carving from the L.B. site of Lou-lan, shown in the same plate, helps to demonstrate the close chronological connexion between the two groups of ruins. Both this piece and the small fragment of a similarly carved capital, M. v. 003, have remains of polychrome decoration. The carved lotus, M. v. oo6 (Plate XLVII), which was found close to the top of the surviving part of the Stûpa, is also likely to have adorned some member of the superstructure. Some of the petals still retain their gilding. The iron tang which passes through the centre of the carved'ornament in
' Cf. Foucher, L'an! du Gandhdra, i. pp. 74 sqq. ; also
above, p. 38.
s See Foucher, ibrd., i. Figs. zo, 70, 71 ; also above, p. 38,