354 THE ANCIENT SITE BEYOND THE NIYA RIVER [Chap. XI
or use. About the first possibility it is, of course, impossible to form any opinion at present ; but the careful, neat way in which these tablets are made up and sealed, seems to speak against it. In support of the second alternative the double seals found on two out of these three tablets (N. xv. 151, 155, see Plate XCIV) may be mentioned ; for these would suggest an agreement or contract which two parties attested and which might well be kept unopened until the time for adducing its evidence came, or until the period of its use had passed. In the case of N. xv. 196, which bears only a single seal, the assumption of its being a bond or similar record seems also possible.
I am the more inclined to favour such a conjecture because, as far as our available materials show, seals in clay were the only means used to authenticate any of these ancient Kharosthi documents. In none of those which are translated, or which I have been otherwise able to examine, do we meet with anything that could be taken for a signature. In the wedge-shaped tablets we do not even find the senders' names mentioned, their seal-impressions being evidently held sufficient to indicate the origin of the missives and attest their contents. If this assumption as to the use of seals as the sole means of authentication is true, it would seem to follow that documents which were intended eventually to serve as legal evidence had, when written on wood, to be kept sealed and unopened until produced in court ; for only so long as the fastening remained intact could the seal-impressions on the covering-tablets be appealed to as proofs of the genuineness of the contents within. Inconvenient as such an arrangement might have been at times, it would at least help us to understand why documents which had ceased to possess value found their way to the rubbish-heap without ever being opened.
Of the clay seal-impressions which originally must have been borne by the obverses of all covering-tablets, whether of wedge or rectangular shape, not less than twelve were recovered more or less intact from this antique rubbish-heap, besides two detached clay seals. In Plates LXXI and LXXII reproductions, in some cases slightly enlarged, are presented of them, as well as of the few equally curious specimens which are preserved on tablets found in the ruined buildings N. 1 and Ix. Apart from their use as means of authentication, exceptional interest may be claimed for these seals ; for they have furnished direct evidence of the influence of classical art in distant Khotan. It was a delightful surprise when the first intact seal-impression that turned up here, N. xv. 24 (see Plate LXXI), presented me with the unmistakable figure of Pallas Athene carrying aegis and thunderbolt, deeply impressed from a well-engraved intaglio. The identical seal appears to have been used also on the wedge-tablets N. xv. 307 (see Plate LXXI) and N. i. 17, 67, though the impressions there are not quite so well-preserved. It shows the goddess in profile standing to the left proper, helmeted, with the right arm drawn back in the act of raising the thunderbolt. The outstretched left supports an object which must be meant for the aegis. Passing over both arms and drawn behind the shoulders there appears a scarf-like chlamys, with its loose ends hanging down from the elbows. The pose of the figure, rigid but yet alert, suggests imitation of an archaic type of Athene Promachos, according to information kindly supplied by Mr. Cecil Smith. Considering the smallness of the figure, measuring only about j! in., and the relative coarseness of the material in which the impression was taken, the fact of these details being still distinguishable shows that the original seal must have been finely cut. A restored but faithfully conceived enlargement of it, by the artist hand of Mr. F. H. Andrews, is presented in the vignette of the title-page.
The seal, with its bold elliptical mouldings, must have been mounted on a square setting. In N. xv. 24, where the impression of this setting appears more clearly (see Plate LXXI), it is still possible to make out the traces of a legend running round the outermost moulding on the