is supposed to be much older than the rest of the group. A central circular chamber, with walls of upright stones, has a roof in which an attempt has been made to form a kind of dome, the stones of which project and overlap each other, marking, clumsy as is the construction, a considerable advance on anything previously accomplished, and adding considerably to the solidity of the monument. An examination of the megalithic monuments still standing enables us to judge of the difficulties with which their builders had to contend, bearing in mind the primitive nature of their tools. We have already given the dimensions of the stones forming the alignments at Carnac. Those at Avebury vary in height from about fourteen to sixteen feet, and in the Deccan is a tumulus surrounded by fifty-six blocks of granite of an even greater size. One of the slabs of the Pedra-dos-Muros (Portugal) is remarkable for its size; and the length of the table of a dolmen on the road from Loudun to Fontevrault is more than seventy-two feet long; that of the dolmen of Tiaret (Algeria) is some seventy-five feet long by a width of nearly twenty-six feet and a thickness of nine and a half feet. This extremely heavy block rests on supports rising more than thirty-nine feet from the ground.27 Stone as well as wood can be much more easily cut in one direction than in any other. Men early learnt to recognize this peculiarity, and to take advantage of it in attacking rock. With their stone hammers they struck in straight lines, always aiming at the same points, and then, probably with the help of a fierce file, they succeeded in breaking off fragments. They also employed wedges of wood, which they drove into natural or artificial fissures, pouring water on to this wedge again and again. The wood became swollen with the damp, and in course of time a block of stone would be detached. Neither time nor sinewy arms were wanting, and Fergusson has remarked that any one who has seen the ease with which Chinese coolies transport the largest monoliths for considerable distances, will not look upon the difficulties of transport as insurmountable. A more serious difficulty would be the placing of the table of the dolmen on the supports, which are often raised to a great height above the ground. It is supposed that earth was piled up against the jambs so as to form an inclined plane, up which the table was slid into place with levers and rollers of the most primitive form, such as were in use in the most remote antiquity. Sometimes the way in which these stones are balanced is perfectly marvellous. The Martine stone, near Livernon (Lot), for instance, is the shape of a boat, and the slightest touch is enough to make it rock on its two supports. That of Castle Wellan (Fig. 55) rests on three stones pointed at the top, and some of the trilithons of India are of even more remarkable construction.