Sec. vi] RECORDS FROM THE HU-KUO CONVENT 281
is intended 11. But it must be pointed out that, among the attributes mentioned in the descriptions of Maiïjugri forms accessible to me, the cup here seen in the right hand does not figure. On the other hand, we can well reconcile the natural pink colour of the flesh with the saffron colour which certain of the Sâdhana texts indicate for Manju§ri 12. I have already had occasion to state that this panel was found fully one foot above the floor, and the good preservation of the wood also shows that the damage suffered by the colours cannot be due to moisture. This makes me think that possibly the partial disappearance of the coloured surface may be the result of long exposure to which the panel was perhaps subjected, while the nail passed through the hole of its handle kept it fixed to the wall of the deserted dwelling, as above suggested.
SECTION VII.—OTHER RUINS OF DANDAN-UILIQ AND GENERAL OBSERVATIONS ON SITE
The remaining ruins explored by me at Dandan-Uiliq can be dealt with more briefly. Central They consist of what I may call the central group of remains immediately to the south of group of my camp, and of certain isolated structures elsewhere. To the former belonged D. I, the much
decayed small cella, the débris heaps of which were examined by me immediately after my
arrival, and which has already been described'. D. vIII, a small structure situated about 30 yards
to the north-east of it, proved to contain a single room measuring 17 by i 2 ft. inside. Though
covered by some 6 ft. of sand, the room on excavation yielded nothing but the fragment of
a paper document in cursive Brahmi (D. VIII. 2), measuring about 6 by 4 inches, which turned
up close to the floor in the north corner. Some débris of plaster and completely decayed
timber which lay half-covered by sand close to the north-east wall of D. VIII, probably originated
from earlier burrowings of ` treasure-seekers'. On clearing it, a small and brittle piece of crumpled
paper was found containing the fragmentary Chinese document D. viii. i (see Plate CXVI).
M. Chavannes' translation shows that it probably formed part of a letter written by, or addressed
to, the resident monk of a Buddhist shrine 2. It derives interest from the mention made of
the monk Ta p`i, whom we have already met in the document D. vii. 4. c, and who is here
designated by the title of Wei-na or Karmadana of a monastic establishment. The reference
to the contract about a female servant given as security suggests that this communication, too,
may have been connected with business arising from some loan or contract.
At a distance of about 85 yards to the north-east of D. vIII were the much-decayed remains Dwelling
of a large dwelling (D. xIII on plan), measuring about 6o feet from east to west. According to D. x"II.
Turdi's statement a find of silver equivalent to about Rs. 200 had been made here in the time
of Niâz Hakim Bag, and he, as well as others of his profession, had since tried their luck at
this ruin by burrowing into the walls or floor of exposed rooms in the hope of further finds.
This information was corroborated by the completely exposed and eroded state of what seemed
to have once formed the north flight of rooms in the building. The rooms to the south, being
apparently built on a lower level, had preserved their walls better, and were filled with sand
from 5 to 7 feet in height. I had the central one, measuring about 22 by 18 feet inside, cleared,