Sec. ii] EXCAVATION OF ANCIENT RESIDENCES, N. II, N. nt, N. iv. 329
The great majority of the tablets belonged to the class of oblongs, counting sixteen specimens among the twenty that were capable of transport. The various modifications of shape as described with regard to the finds of N. iv. are represented here also, as a reference to the inventory list will show. Writing in columns, or else in short detached items suggesting lists of names or accounts, is frequently seen in the still legible pieces. The large proportion of `oblongs' of considerable length was here particularly striking. Among the above specimens there are seven measuring over 14 in. in length, N. v. I and N. v. 6 attaining the respectable dimensions of 302 and 3I in. respectively. N. v. 6, reproduced in Plate CI I, is curious also for showing portions of its columnar text both on Obv. and Rev., written originally with a better ink than the rest, and hence presumably later additions. Owing to the dark colour assumed by the partially-perished wood the writing, though clear enough in reality, is barely traceable in the reproduction. The fact of some bark still remaining attached to the narrow sides of this tablet shows the rough methods with which this wooden ` stationery ' was prepared on occasion. Among pieces that retained no trace of writing, and could not be carried away owing to their complete decay, I counted eight having a length of over 16 in., one of them showing the imposing but inconvenient dimensions of 42 feet in length, with a width of 42 in. Such dimensions, and the prevalence of oblong tablets with columnar writing and detached notes, point to records kept in some office. The very small number of wedge-shaped and rectangular documents 1 found might be taken to indicate that the office once housed in N. v. had little to do with correspondence. Fragments of a large pottery jar and small pieces of felt and coarse sacking (N. v. 17) were also found here.
In view of the conditions already described it could not surprise me that the clearing of half a dozen more ruined dwellings yielded no further results. Everywhere erosion had done its work thoroughly, sparing neither walls nor any objects that might have been left behind within them. By the rows of posts it was just possible to distinguish the division of apartments, and the sketch-plan of Plate X X I X shows that the latter were numerous in several of these ancient houses. But the decay was far too advanced to permit the character, communication, &c., of the rooms to be ascertained. The only discovery of interest in the course of a second day's excavations was an ancient ice-pit in the outhouse of a modest dwelling-place on the western edge of the area. Here, in a small room measuring about I2 by 9 ft., the labourers came upon two unhewn trunks of Toghrak half-embedded in the floor and lying parallel close together. Abdullah, my guide from Keriya, at once suggested that we had found a muz-khâna or ice store-room, trunks of trees being ordinarily used now in exactly the same way for keeping the ice from touching the ground. Abdullah's conjecture was soon confirmed by the discovery of thick layers of poplar leaves filling the space of about 2 feet between the trunks, heaps of such leaves being still the usual covering for the ice which well-to-do villagers are accustomed to store for use in the summer.
The sand covering the dwelling just referred to was fully 4-5 ft. deep, yet owing to ancient erosion nothing was left of the walls but the bleached and withered posts that had guided me to it. The absence of walls, of course, greatly increased the difficulty of excavation, since the drift-sand was ever flowing back into the space cleared. Seeing how slow progress would be under these conditions, and how little prospect there was of adequate return for further labours, I decided, after the conclusion of the second day's work, to abstain from clearing the remaining structures of N. II traceable on the northern side of the area, and to turn to more