TOWERS, WOOD-CARVINGS ETC. OF LOU-LAN. 645
on; some Chinese copper coins; and — the most important of all — a tablet of wood, 4 x 6 cm., crowded with lettering, the nature of which I leave it to the specialists to explain.
The two wings of the building were devoid of ornamentation. The answer
-
to the question, as to whether this little temple was thus tastefully and richly decorated on the inside or on the outside, must be that it was probably on the inside. The structure is not so big but that it is quite conceivable, that the whole of the interior may have been decorated in this manner. As I have stated, the carvings were also painted. In what way the various parts, the planks and squares, were fastened together cannot be determined, though the latter were evidently used to break the continuity of the former. The large plank shown on one of the plates in the archæological part and also on fig. 294 shows plainly that the pattern was not contained entirely upon it, but ran over on to the adjacent planks, which must consequently have touched it. Yet even with the help of all the fragments that survive, it would be too rash to attempt a reconstruction that would be likely to carry anything like conviction. The ground-plan of the edifice may be inferred from the four massive foundation beams, with the holes cut in them for the vertical posts. It may be assumed that the • doorway and door, or double doors, had the same appearance as those of the house M ("village at Camp CLIX). But here again we have no , sure means of deciding what the roof was like. The turned spiral also shown on the right of one of the plates in the archological part may possibly afford some slight clue. Although we found several of these, we only brought away one as a specimen. They were probably used to decorate the roof ridge, or its horns. I observed roof terminals of this kind on every temple in Kum-bum, the architecture of which is greatly influenced by Chinese models; though the temple buildings at Hemi are constructed in the Tibetan style. With regard to the temple and the larger houses of Lôu-lan, we may be quite certain that Chinese models had a far greater influence upon their builders than they had at Kum-bum. If then the small spirals of different sizes (the largest is 57 cm. high) of which I have been speaking were employed in the manner indicated, the roof can hardly have presented any other form except that of the usual high-pitched Chinese ridge with gently up-curving horns at the ends.
This neighbourhood too, in the extreme west of the villages we discovered, was virtually free from drift-sand. It was only on the south-west side of the jardangs and mounds that an occasional small dune had accumulated. In the vicinity of the temple mollusc-shells were extraordinarily abundant, and very often they lay in heaps in the little hollows, having been swept together by the wind like withered leaves. In consequence of the uninterrupted wind-erosion an increasing number of mollusc-shells continue to be shaken out of the clay deposit which formerly settled on the bottom of the lake. We often used to see them sticking like white dots all over the sides of the jardangs; these were of course only half exposed, but as the corrasion continues they will gradually become loosened and fall out.
This concludes my brief description of what was to be seen on the site of our two visits, unfortunately too short, to the place. The ruins, consisting of four villages and a solitary tower, stretch along a line that runs from N. 60° W. to
IIIFカラー高解像度
