National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

398   THE LOU-LAN SITE   [Chap. XI

In the second main group of wood-carvings may be placed the numerous fragments of openwork panels which, as already explained, must have occupied a considerable portion of the spaces in

the wall and served for the admittance of light and air.9a Simplest in design are L.B. I1. 0028 (Plate XxxiII), showing heavy trellis-work with diamond openings, and the pieces, L.B. Ii. 0023-24, in which an open trellis-work of diagonal bars is combined with wheels about each point of junction. More characteristic in style of design is the large open-work panel, L.B. II. 0016-17 (Plate xxxI), in which a four-petalled lotus with sepals between fills a rectangular plain-bordered frame. Fragments of similar panels are L.B. Ir. oo5, 0030 (Plate xxxII). Other panels of this class must have contained representations of grotesque beasts carved in relief as shown by the interesting fragments L.B. 11. 0011-13, 0053 (Plate xxxiii) and II. 0021 (Plate xxxIv).

Among wood-carvings distinct from the above groups the most interesting piece is L.B. u. 0014

(Plate xxxIi), the fragment of a richly-carved capital of Indo-Corinthian type with acanthus leaves and volutes. Both the way in which the octofoil plan of the top is combined with a circular lower plan and the method of carving are characteristics which, as pointed out by Mr. Andrews in his description in the list below, are found again in Byzantine capitals. In view of what recent research has proved as to th'e dependence of Byzantine art on the Eastern development of Hellenistic art, the early date of this Lou-lan capital, which can scarcely be later than the third century A. v., and its discovery on the very high road to the Far East, may claim special interest. Among other wood-carvings may be mentioned the large block, L.B. II. 0014 (Plate XXXII), showing a circular eight-petalled lotus or rosette which, as comparison with the panel M. v. 0012

from a Mirân temple shows,' °b may have formed part of a lintel ; the triple wooden bud, L.B. n. 004 (Plate xxxiv), and the fragment, L.B. n. oo6, of a lotus-wreath carved in the round.

Wooden   The curious wooden stave, L.B. iI. 002 (Plate xxxv), slightly tapering at one end and with

pins for ectural a large ball at the other, like similar wooden staves subsequently found at a neighbouring ruin, decoration. L.B. iv. i. 2 ; iv. 001-002 (Plate xxxv) ; v. ooio, was for a long time rather puzzling. That they

had all served as pins for fastening beams or panels to the timber framework, and that their knob-like ends, whether shaped as balls or cones, or to resemble inverted thistle-heads, were added for the sake of architectural decoration, was first made clear to me by my architect friend Mr. J. H. Lyon. The use still made in timber construction of similar large-headed metal pins for ornamental purposes supplies an exact parallel. This explanation is fully confirmed by the measureinents of the pin portions proper in all these pieces. Their length varies from nine to ten inches, and thus corresponds exactly to the usual thickness of ten inches which I measured in the foundation beams of L.B. I-III and neighbouring ruins, and which is likely to have been that of the walls also. They all show a section of seven-eighths to one inch square, and with this they fitted exactly into the holes which I actually found at the ends of two beams excavated in the hall L.B. iv. iv.10

The whisk-broom of grass, L.B. II.0054 (Plate LII), which came to light from the sand protecting

these scanty architectural remains of the little temple, presents a quasi-pathetic interest ; for, just like the brooms found in 1900 in one of the temple cellas of Dandân-oilik," it must have been in the hands of the very last attendants who endeavoured here to keep the objects of worship clear of dust and sand.- It is, perhaps, significant of the prevalent Chinese influence on the ancient route through Lou-lan. that the make of this humble household implement agrees in its constructive principle with that of the broom T. xili. iii. 001, subsequently discovered at one of the watch-

Open-work panels.

IndoCorinthian capital.

Ancient
whisk-
broom.