Sec. v] THE DECORATIVE DESIGNS OF THE L.C. FABRICS 239
foliage. Mr. Andrews interprets this rather confused design as a clumsy adaptation of fragments of three or possibly four different cloud scroll patterns, treated by the turn-over method whereby exact bi-symmetrical schemes are obtained producing new and adventitious forms '.11 The point here raised as to the possible origin of certain peculiar motifs, such as the feature recalling the so-called ` tree of life ' of ` Sasanian ' and other Near Eastern textiles, fully deserves to be kept in view by future investigators.
Before we turn to patterns composed mainly of floral motifs or scrolls, we should notice a small but important group of designs, also showing animal figures but strikingly different in style from the fabrics so far examined. This group is represented by two fragments, L.C. ii. of (Pl. XXXVI) and L.C. x. 04 (Pl. XXXV), the designs of which agree closely in the figured subject, treatment and colour scheme, but differ in scale. The most striking feature in both is the introduction of that characteristically Chinese ogre, the ` T`ao-t`ieh', and the extremely stiff, distinctly archaic execution of all details, which is strangely reminiscent in some ways of Pacific work. In L.C. ii. of (Pl. XXXVI, XL), which has been illustrated and fully described already in Mr. Andrews' paper,12 we have on the right a frontal view of the T`ao-t`ieh, in very angular drawing, showing the monster's disproportionately large head with grinning jaws and huge hexagonal eyes. Bandy legs project below, with spines bristling from elbow to ear. To the left of the ogre and separated from it by a tree indicated in symmetrical outline, a winged lion, passant, is drawn in a style which in simple silhouette treatment and vivacity closely recalls the figurative work of Han bas-reliefs. A second tree, far less stiffly drawn, with curving stem and branches, stands between the lion and the hind part of a stealthily moving dragon. The pattern is repeated vertically in close-set rows. The colours are rich, but confined to a dark yellow-brown for the ground and a dark bronze-green for the figures.
The design of the fabric L.C. x. 04 (Pl. XXXV, XXXVII), preserved in numerous pieces in a ragged state and difficult to open, consists of an extremely close but much larger rendering of the T`ao-t`ieh as seen in L.C. ii. of, and of a winged lion, similar in type to that of the latter fragment, but facing the monster in the manner of a heraldic supporter. Here too the pattern repeats vertically quite close, and the colours are restricted to two—rich blue and golden yellow.
By their style and treatment alone these two fabrics could clearly be recognized as the oldest among our figured Lou-lan silks. But definite archaeological evidence of this has been provided by the fortunate discovery of a patchwork of figured silk, T. xxii. c. oofo. a, closely allied in style and treatment of pattern, at one of the watch-towers of the Tun-huang Limes explored by me in 1907. The early date of this fabric is fixed with fair chronological exactness by a Chinese wooden document, dated 98 B.C., which was found in the same refuse heap. As the patchwork has already been fully described and illustrated by Mr. Andrews in Serindia,13 there is no need to refer here in detail to the elements of the pattern. It will suffice to state that it is an ` all-over' diaper of diagonally placed squares, with groups of four highly stylized ogre heads forming symmetrical bosses at the crossings of the straight cloud scrolls that enclose the squares. Within the squares two different patterns alternate diagonally. In one a pair of dragons confronts a pair of phoenixes ; the other shows two uniform pairs of birds, placed feet to feet 14 The close connexion between this design and that of the last two L.C. fabrics is clearly recognizable in the stylized ogre heads ; in the silhouette treatment of the figures, which conforms to the archaic convention of the ancient bronzes ; and also in the stiff angular shapes, probably meant for trees, that form the centres of the squares. There can be no doubt that we have in all three fabrics specimens of a decorative
11 See Andrews, ibid., p. 9.
12 Cf. ibid., p. 12, Fig. 8.
13 See Serindia, ii. pp. 72o, 785 sq. ; iv. Pl. LV, CXVIIi;
also Andrews, Chin. Figured Silks, pp. 12 sq., Fig. 9.
14 M. Goloubew, B.E.F.E.O. xx (1920). p. 173, suggests that these birds may be meant for hua chung (` flowery bird ').