On gaining the rock, Lieutenant Alexander decided to make use of the holes which had been driven into the granitic mass by Captain Swift to receive the piles of the previous structure. The twisted and broken pieces of iron were withdrawn and the holes cleaned out. Simultaneously the upper surface of the rock was pared and trimmed by the aid of chisels, which was no easy task, because at times the masons were compelled to manipulate their tools as best they could in two or three feet of water. This preparation of the rock to receive the base constituted one of the most notable features of the work. In the greater number of other outstanding achievements upon sea-rocks the surface of the latter has been above the waves at lowest spring-tides, whereas in this case a great part of the foundation work was continuously submerged. This preparation of the rock-face necessitated the final trimming and shaping upon the site of many of the masonry blocks forming the root of the tower. They could not possibly be prepared ashore to bring about the tight fit which was imperative. Accordingly, all but the bottom faces of the blocks were prepared in the depot on the mainland, and they were then shipped to the ledge for final paring and trimming. The attachment of the bottom courses to the rock-face was carried out very ingeniously. Bags of sand were brought on to the rock and laid around the spot upon which a particular block of stone was to be laid. The sacks, being filled with sand, were pliable, so that, when deposited, they adapted themselves to the contour of the ledge, and prevented the water making its way in under the rampart. The water within this small dam was then removed, sponges being used in the final emptying task, so as to suck out the salt sea from the cracks and crevices, leaving the surface on which the block of stone was to be laid quite dry. A film of cement was then trowelled upon the rock surface, and upon this was laid a sheet of muslin. The inclusion of the muslin was a wise precaution, because while the work was in progress a wandering wave was liable to curl over the rock, swamping the small dried space, when, but for the presence of the muslin, the cement would have been carried away. At the same time the cement was able to penetrate the meshes of the muslin when the stone was deposited, so as to grip the surface of the latter and to hold it tightly in position.