404 THE LOU-LAN SITE [Chap. XI
a bronze cup, L.B. iv. v. 0031 (Plate xxxvi), is of curious shape and uncertain use. Among miscellaneous finds in wood, I will only mention the knife sheath, L.B. iv. v. 007, and the half baluster, L.B. iv. v. 0026, which suggests that this room may have had a decorative panelling similar in style to that of the small neighbouring ruin L.B. v. A textile relic of a kind not found elsewhere is the collection of small pieces of yellow felt, L.B. iv. v. 0032-33, painted on a tempera surface with floral and geometrical designs in a variety of colours.
To the north-east of the rooms v, vi, and separated from them by a passage over eight feet wide, a larger apartment, vii, and a narrow closet, viii, could be traced. A portion of their timber and wattle walls still rose to a height of a few feet, owing to a thick protecting crust of dung which the photograph, Fig. 108, shows in course of removal. It is likely that these apartments represent all that remains of the private portion of the house which, lying to the north-east, has otherwise been demolished by the erosive action of the wind. The `finds' made here were scanty, but included the fine fragments of decorative open-work wood-carving, L.B. iv. vii. 003 ; viii. 001, 002 (all in Plate xxxiv), with classical design in berried laurel leaves and palmette, to which reference has already been made.i2
About thirty yards to the east there rose a small isolated terrace bearing on its top the previously mentioned remains, L.B. v, of what evidently had once been a shrine closely resembling L.B. it and of similar modest dimensions (Fig. 112). A great deal of timber débris strewed the eroded slopes, but it was all badly splintered, and only one foundation beam, about twenty-one feet long, lay still approximately in situ.13 The character of the other big pieces could not be determined with certainty. However, carved fragments picked up among the débris clearly show that the constructive features and the ornamentation of the walls in the building must have been of exactly the same type as in L.B. n. Among these wood-carvings we have a beam, L.B. v. 0013 (Plate xxxI), decorated in relief with the same floral scroll as found in L.B. H. 0036 ; an open-work panel, L.B. v. 009 (Plate xxxiI), showing the same four-petalled open lotus which is seen in L.B. ii. 0016-17 ; a piece of heavy diamond trellis-work, L.B. v. oot 2 (Plate XXXIII), of a type identical with that of L.B. ii. 0028 ; another open-work panel with diagonal bars and wheels, L.B. 0018 (Plate XXXIII), similar to L.B. iI. 0023-24 ; and also a series of lathe-turned balusters, L.B. v. oo8 (Plate xxxiii), ooio, 0014-18, which differ from those found at L.B. ii only by having seven instead of five ball mouldings, and in a few other unimportant details. A wooden rail from such a balustrade was also recovered, L.B. v. 0019."
Such close agreement in the construction and style makes it highly probable that the small isolated ruin was a Buddhist shrine of the same period as L.B. ii. This again raises the question whether the large and well-built residence L.B. Iv adjoining it, with its relatively spacious rooms, may not have been originally intended to shelter a monastic establishment.14a Unfortunately we have no definite evidence to decide it. It is worth. noticing, however, that this is the only place of the Lou-lan Site where clear surface proofs of ancient cultivation could be discovered. An extensive line of rush-fencing was traced to the north-east, half-hidden but also protected by the sand which had accumulated along it in a low, well-defined ridge. It first ran due north from L.B. iv and then turned to the east, enclosing what must have been a large garden or piece of cultivation. Within the line of the fence and near it we came upon fallen trunks of dead mulberry and Jigda trees. About eighty yards to the west of L.B. iv there lay in a row the withered trunks of eight big