Sec. iv] ANCIENT DOCUMENTS ON WOOD AND LEATHER 355
left proper. The relief characters composing it are, however, so faint that Prof. Percy Gardner, who was good enough to examine them in different lights and under the microscope, could assert of them only that they were not Greek. Prof. Rapson came to the conclusion that the inscription was Kharosthi, but he could only distinguish traces of a few detached characters, among them ma at the bottom of the impression. Traces of characters appear also on the flat part of the seal-impression on N. ii. 2 (see Plate LXXI, while on the seals of N. i. 9, iv. 8o, xv. 137, the only other ones showing square settings, the clay surface corresponding to the latter is too poorly preserved to indicate whether there was a legend or not. It is thus possible to assume that the square setting, wherever provided for these seals, might have been specially intended for engraving the name of the owner. Its impression would have sufficed for the distinct indication of the sender where the latter was not named in the document.
In this connexion it is curious to note that the impressions of seals provided with settings appear without exception on wedge-shaped tablets (see N. i. 9, 17, 67 ; ii. 2 ; iv. 8o ; N. xv. 24, 137, 307, in Plate LXXI, i. e. on a class of documents where, judging from the specimens already translated or examined, the senders' names would not be stated in the text. On the other hand, none of the nine rectangular documents which have preserved their seal-impressions" show any settings. Is it possible to assume that the seals with settings and legends served as a kind of office stamp, specially designed for use on documents where the issuing official's name would not otherwise appear ? Such an explanation would render it easy to understand the combination of seals of classical design and make with Kharosthi legends. While the former would be imported from the West, the settings engraved with the latter had to be prepared locally to show the names of their users. That these were officials can be safely asserted of all seals provided with settings ; for the wedge-shaped tablets on which their impressions occur were used solely for the record of brief official orders.
A Pallas Athene of the same type as above described, but from a different and less carefully made intaglio, is seen in the seal-impression of the wedge-tablet N. xv. 137 (see Plate LX XI). N. ii. 2 (ibid.) shows a well-modelled male figure standing to the right proper, which from its muscular development appears to be meant for a Heracles or possibly an Eros. The latter god may be intended also in the seal of N. i. 9, if the male figure standing to the right proper is to be taken as winged. But the impression of this seal is too poorly preserved to permit of a certain opinion. Also of the seal on N. iv. 8o (see Plate LXXI) impressed in very coarse clay it is impossible to say more than that it represents a male figure of classical design.
Turning to the series of seals preserved on rectangular tablets we find two which plainly display western, i.e. Roman, workmanship. N. xv. 166 (Plate LX X I I) shows a Pallas standing to the left proper, but with a figure far less carefully modelled and engraved than in the seal found impressed on N. xv. 24, 307, &c. More curious and also far better preserved is the seal of the small covering-tablet N. xv. 33o, which Plate LXXI reproduces in its original size, together with an enlargement of the seal only. The figure shown by the impression within an elliptical ribbed border of mouldings is that of a youth apparently nude, seated to the left proper on a stool. The appearance of what look like folded wings at the back of the figure suggests a representation of Eros, and the strongly-curved bow seen in front of the figure and above its knees supports this identification. The good preservation of this seal is probably due to the superior quality of the clay used for it. Of seals which are rather Oriental than Roman, yet show too much of classical influence to be taken for the work of Khotan artists,
18 N. xv. 38, 155, 163, z66, 167, 196, 330, 344; xxiii. I (see Plate LXXII).