The tumulus was enlarged proportionately and usually completely covered the chamber. Its height averages ten to fifteen feet, and its diameter over ninety. The culvert-like entrance had to be lengthened accordingly. But one large chamber did not suffice for successive generations. It was often extended or additions were made so that quite complicated forms occur. In England we find frequently a row or cluster of small chambers. Here the roof is sometimes made of successive layers of stone approaching as they ascended until one slab covered the "false arch." In Brittany we find great diversity as well as complexity of form. In some parts of France the entrance continues the main line of the chamber instead of being at right angles to it. The French have well characterized these as "_Allées couvertes_." Some of these "gallery chambers" were very large and contained a large number of bodies; sometimes from 40 to 60, in one case 100. The tumulus at Mont St. Michel measures 115 by 58 metres, and forms a veritable hill. Thirty-five thousand cubic metres of stone were employed in the construction of the chamber. Other chambers are from 30 to 50 feet in length. The celebrated chamber at Bagneux, 25 feet long, is composed of fourteen great blocks, of which three form the roof. The great tumulus at _Fontenay-le-Marmion_ in Normandy covered eleven chambers in two parallel rows. All the material for these great structures could hardly have been found in the same vicinity. In one case it appears to have been brought from a quarry two miles away. Some large stones, weighing thousands of tons, seem to have been transported many miles. Some of the latest structures show a certain amount of degeneration. Certain galleries were apparently roofed with timber. We find "dry" masonry, of smaller stones laid in courses but without mortar, alternating with or replacing the great blocks, especially in structures of Æneolithic or Bronze Age. The custom was declining and soon after this disappeared.[90] The age of these stone monuments can generally be fairly closely determined by the contents, unless these have been removed or destroyed by treasure-hunters, as is often the case. In many cases the objects originally deposited seem to have been few and insignificant. Later, secondary interments were often made in tumuli, but these usually betray their later date by their position above the original chamber or near the side of the mound. We must keep in mind that chambers in the north containing only stone implements may be often of the same age as those farther south containing copper or even bronze, for metal made its way northward only gradually. The custom of building dolmens seems to have persisted later in England than in France. The English round tumuli or barrows belong to the Bronze period. It is not surprising that one country should be more conservative than another, especially if it is somewhat remote.