Sec. ill] THE RAWAK STUPA 487
I found from the first that the excavation of this wealth of statuary was attended with serious difficulty, even in that limited portion of the quadrangle where the conditions of sand permitted operations. Owing probably to the moisture rising from the neighbourhood of subsoil water, the strong wooden framework which once supported internally the masses of stucco and fastened them to beams let into the wall behind, had completely rotted away. The cavities left by these beams, which were apparently 5 in. square, and fixed at a uniform height of about 8- ft. above the floor of the court, could be traced all round the excavated wall portions wherever the latter had retained a corresponding height (see e. g. Figs. 62, 65, 66, 68 ; P1. X III. b ; X I V. a, b, &c.). The beams had been set sufficiently deep into the wall to permit of their being covered on the outside with a layer of plaster about i to 2 in. thick, flush with the rest of the wall surface. For some distance on the cleared portion of the south-west wall this plaster covering had survived, and it is owing to its presence that in Figs. 61, 62 the cavity left by the perished beams does not appear. Here and at some other points I could still trace small remains of the beams in bits of decomposed wood. The wooden framework of the colossal statues must have been joined to the beam behind usually at the height of the heads or shoulders 4. The wood of the internal framework had everywhere completely perished, but the position once occupied by portions of it was often still indicated by the matrix this wooden core had left, as visible, e. g. in the arms of colossal images (see Figs. 61, 62, 68, 69 ; P1. XV. a, b, c ; Pl. XV I I. c).
Deprived of this support, the heavy stucco images, especially those still retaining much of their upper portions, threatened to collapse when the protecting sand was being removed. The strong winds blowing day after day greatly added to this risk. They carried away the fine dust of riverine loess which had filled the interstices between the edges of the relief work and the wall behind, and thus increased the danger of the friable masses of clay sliding down through their own weight to immediate destruction. Regrettable experience of this kind was gained in the case of some statues on the inner south-west and south-east walls. It soon showed me that these risks could be obviated only by extreme care in clearing the reliefs—a few hours of exposure to light winds often proved safer than the application of brushes for removing the layers of dust from drapery folds, &c.,—and by covering up again the lowest portions of the statues as soon as they had been examined and photographed. Even so damage could not altogether be prevented. In some instances it was necessary to secure the heads, &c., of statues still intact by means of ropes while they were being photographed. Fig. 69 illustrates this procedure followed in regard to some of the minor statues excavated on the inner side of the south-west wall ; it also helps to mark the true size of the colossal image seen to the extreme right by comparison with the labourers.
The conditions here briefly indicated, which rendered the work so difficult and risky, also explain why many of the colossal statues were found badly damaged in their upper portions and all of them without their heads. Their upper portions, just like the top segments of the great aureoles seen in Figs. 63, 64, had necessarily been left much longer without the protecting cover of sand, and had accordingly fallen away from the wall that formed their backing and support. Numerous fragments of colossal heads which had thus become detached long ago turned up at different depths of sand, especially near the south corner. Their extremely friable condition, which made them break at any attempt to lift or remove them, showed that they had long been exposed to atmospheric vicissitudes. The heads of the smaller images, which the dust accumulations had had time to cover up even while perhaps a roof above gave shelter, were
' See Fig. 62, and in Fig. 6r the series of small holes passed through the plaster into the beam behind.
above, marking the points where the joining pegs or tangs