348 THE ANCIENT SITE BEYOND THE NIYA RIVER [Chap. XI
with those ascertained in the case of the first find, the length of the wedges varying from 78 in. to 15 in., while their width at the square end proportionately varies from 11 in. to 21 in.
The complete `double wedges' (N. xv. 12, 24, 66, 71, 137, 14o) still retaining the string with which they had been originally fastened by their senders, afford the clearest possible illustration of the manner in which documents of this type were written and dispatched. They consisted invariably of pairs of pieces, fitted exactly to match each other in size, as seen in the complete specimens N. xv. 24, 71, 137, reproduced in Plates XCVIII-C. The thickness of the two pieces, except where the raised seal-socket is placed on the covering-tablet, is also fairly uniform in each double wedge. Otherwise it varies, in some relation, perhaps, to the size of the tablets and the quality of the wood, from T:7.6 in. to 1: in. It is very probable that the carpenter engaged in preparing this kind of stationery first shaped the ' double wedge' as a whole, and then
separated its two pieces by sawing or splitting. This was certainly the method adopted by the maker of N. xv. 24 (Plate IC), the inner surfaces of both covering- and under-tablet showing identical faults in the grain. One end of the double tablet thus formed was cut off square ; the other runs out into a point, near which a string-hole is drilled through both pieces. The text written in lines parallel to the upper long side of the wedge, and nowhere exceeding four in number, occupies the smoothed obverse of the under-tablet, and is protected by the upper- or covering-tablet, which rests on it and serves as a kind of envelope. Whenever the length of the communication required it, the writing was continued on the reverse of the covering-tablet in such a way that, with the latter turned back, the whole of the text could be seen simultaneously and with the writing placed the right way in both tablets. The position of the covering-tablet of N. xv. 24, as reproduced in Plate IC, will explain this.