It was first raised to a height of fifteen feet, so as to cause it to enter the sand by virtue of its own gravity. It thus penetrated to a depth of about twenty inches. After this the workmen were ordered to man the windlasses and hoist out some of the sand. This caused the tubbing to descend about eight inches more, when it came to a standstill. It was now loaded with 17,000 pounds of pig iron, but in vain, for it refused to budge. Mr. Chavatte therefore had recourse to a dredge with vertical axis, constructed as follows: Upon a square axis, A B (Pl. 2, Figs. 1, 2, and 3), provided with double cross braces, C D, and strengthened by diagonals, were riveted, by their upper extremities, two cheeks, G H, whose lower extremities held the steel plates, I J I' J', which, in turn, were fastened to the axis, A B, by their other extremities. These plates were so inclined as to scrape the surface of the ground over which they were moved. They each carried two bags made of coarse canvas and strengthened by five strong leather straps (Figs. 2 and 4). To the steel plates were riveted two plates of iron containing numerous apertures, through which passed leather straps designed for fastening thereto the lower part of the mouth of the bags. That portion of the mouth of the latter that was to remain open was fastened in the same way to two other plates, X Y, X¹ Y¹ (Fig. 1), held between the lower cross-braces. When the apparatus was revolved, the plates scraped the earth to be removed, and descended in measure as the latter entered the bags. These bags, when full, were hooked, by means of the five rings which they carried, to the device shown in Fig. 8 (Pl. 2), and raised to the surface and emptied into cars. The dredge was set in motion by four oak levers (Figs. 5 and 6). Two of these were manned by workmen stationed upon the surface flooring, and the other two by workmen upon the flooring in the tubbing. The axis was elongated, in measure as the apparatus descended, by rods of the same dimensions fastened together by cast iron sleeves and bolts (Fig. 7). The steel plates were not capable of acting alone, even in cases where they operated in pure moving sand containing no pebbles, for the sand was too compact to be easily scraped up by the steel, and so it had to be previously divided. For this purpose Mr. Chavatte used rakes which were in form exactly like those of the extirpators, U and V, of Figs. 1, 2, and 3, of Pl. 2, except that the dividers carried teeth that were not so strong as those of the extirpators, and that were set closer together. These rakes were let down and drawn up at will. They were maneuvered as follows: