Sec. i] RUIN N. i AND FIRST FINDS OF INSCRIBED TABLETS 317
interment of many centuries beneath the sand. As it was, the sun of one year and its rain
and snow, however slight in quantity, had sufficed to bleach and partly efface the fully-exposed
writing of the topmost tablets.
After placing a guard over the room where Ibrahim's scattered finds lay, so as to prevent Clearing of
further injury or abstraction, I set the men to work to clear the room N. i. (see Plate VI), where room N. i.
he had first come upon them. This was an easy task, as the room measured only i6 by 14 ft.,
and the sand filling it was not deep. On the north-west side, near the edge of the eroded slope,
it lay only to a depth of about 2 ft., which increased to about 4 ft. further in, where the better
preserved south-east wall retained the drift-sand. While the clearing proceeded cautiously, I had
time to examine the system of construction followed in the ruined building. As the same was
subsequently found in almost all other structures of the site, it will be convenient to describe
it in this place. In its general features it showed close resemblance to that noticed in the
Dandan-Uiliq structures, being based on the use of timber ; but there were characteristic minor
differences. Massive squared beams of White Poplar or Terek wood, usually extending below Construe-
several rooms, and in some instances exceeding 4o ft. in length, formed a kind of foundation ; houses.
their thickness, which varied from 6 to to ins. according to the size and importance of the
walls they supported, and their perfect finish and fitting always caused my workmen to wonder
at the skill of those ancient carpenters. On this foundation were set wooden posts from
4 to 6 ins. square, which supported the roof and at the same time served as a frame for the
walls. These, and smaller but equally well-finished intermediary posts, fixed at regular intervals
usually of about one foot, were joined by heavy crossbeams on the top and light ones between.
In the ruin N. i. the upper portions of the walls had decayed too far to show the crossbeams
in position ; but they can be seen clearly, or else the dowels that once held them, in the photograph
(Fig. 41) of the better preserved large dwelling N. iii. The arrangement of the upright posts
is, however, fully visible in Plate VI, showing the room N. i. after excavation. To this
framework, and usually on the outside of the small intermediary posts, was fixed a strong kind
of matting of thin tamarisk branches woven diagonally. This again was covered on each side
with layers of hard, white plaster, giving a total wall-thickness varying from 6 to 8 ins. in
different structures. In the photograph (Plate VI), showing the north-east and south-east
walls of room i. in N. i. as seen from outside before excavation, the diagonal matting, which
had become denuded of its plaster covering owing to exposure above the sand, is distinctly
visible. This diagonal tamarisk matting undoubtedly supplied a stronger core to the walls than
the horizontal layers of reeds used for the same purpose at Dandân-Uiliq. But that the latter
system was known also at the earlier period to which the ruins of the Niya Site belong is
proved by its employment in several instances in the ruins N. irr., N. iv., and N. v., to be
A third, far rougher, method of wall-construction, which closely resembles that still in vogue Walls of
in the Khotan region, was applied apparently only to cattle-sheds, stables, and similar outhouses. Plaster and
It consisted of vertically-placed and closely-packed rushes covered with layers of mud plaster.
Walls of this kind, in which rough wooden corner-posts supported the roof, while saplings or
roughly cut branches of trees were inserted at intervals to strengthen the rushes, were found
near N. i. and most of the ruined buildings of the site, and have been distinctively shown in
the plans. In N. i. also the more solidly built walls had completely decayed, where not actually
covered by sand, but many of the stronger posts originally holding them still rose high above
the surface, in some instances to feet and more. Their splintered and shrivelled appearance
showed the long periods for which they had been exposed to the destructive forces of the desert.