In the Museum at Devizes there are several hard pieces of perforated and fashioned chalk which offer more conclusive evidence. Of these Mrs. M. E. Cunnington, the Curator, writes me: "All the weights here have holes bored right through. Two large ones stand easily on the floor. Others are more irregular in form and will not stand upright. This latter type is, as far as I am aware, the more usual in this part of the country. They are commonly cut out of the hard chalk, and weigh about 3 or 4 lbs. (1·5-2 Kilos). We think these weights are loom weights because we find them with Romano-British remains, as at Westbury, and late Celtic remains on our chalk uplands, far from water where fishing could have been carried on. With the same remains we find weaving combs, numerous spindle whorls and other tools of bone that were also probably used in weaving operations." The Westbury, in Wiltshire, referred to, is some thirty miles in a straight line from the mouth of the Severn, and about forty miles from the English Channel. These pieces of chalk cannot therefore have been used as net-sinkers, leaving out of consideration their composition; they were found with weaving tools and they fit the position. So far the ingenuity of our ablest archæologists at home and abroad has not succeeded in ascribing the use of these objects to anything else than net-sinking or warp tension. The adaptability of the articles for use as warp weights, the small groups in which they are found, the discovery of weaving implements in the closest proximity, our knowledge of the Greek representations of warp-weighted looms, the Olafsson illustration, and the loom in the Copenhagen Museum all tend to prove that these articles are really warp weights. As regards the practical possibility or impossibility of working a "Greek" loom, I had a simple frame made in the Museum and showed Mr. J. Smith, a mill "Overlooker" at Messrs. Wayman and Sons, Ld., Halifax, the illustration in Montelius' book already referred to, and asked him to weave me a small piece of cloth on it. In the course of a few hours he did the warping, beaming and weaving, making the pick with his fingers and using a ball of weft thread instead of a spool or shuttle. The result is shown in the accompanying illustration, Fig. 36, conclusively proving that weaving on such a frame is quite feasible, and practically proving that Olafsson's and the Copenhagen warp weighted looms are properly constructed workable looms. [Illustration: Fig. 36.--A warp weighted loom made at Bankfield Museum, to show the possibility of weaving by this method. There is no heddle nor shuttle used. The weaver made the "shed" and pushed the weft through with his fingers. He naturally worked _down_wards.] [Illustration: Fig. 37.--Diagram to show how the warp is kept taut on a Syrian loom.]