According to popular belief the graves, which are cut in the volcanic tufa, fill with water, and the dead dissolve into a sort of soup. But this is not true; the rock is dry and porous. It discharges its drainage by a little spring on the north-east that in process of ages has worked itself from stage to stage lower down the hill. The Dewerstone Camp consists of two stone walls drawn across the headland. No walls were needed for the sides that were precipitous. Cranbrook Castle is in very good preservation, except on the side towards the Teign, where it has been removed by road-menders, but not within recent years. It richly deserves to be investigated, and the owners have recently granted permission to do so to the Dartmoor Exploration Committee. We come next to the earthen-banked camps. Of these there is a very fine example at Hembury, near Buckfastleigh. But the finest of all is in Burleigh Wood, in the parish of Bridestowe. Here the side accessible from Galford Down has been cut through, with a trench and a bank thrown up on the camp side, and this is carried right across the neck. The earthen banks were almost certainly crested with palisades. Hard by this early camp, where a bronze palstave has been found, is another of a different character, occupying the extreme point of the hill. This consists of a tump or mound, with an earthwork round it as a ring. In this are remains of iron-smelting. There can be little doubt as to the period of this latter. It was the _burrh_ of the Anglo-Saxon, and was in every point similar to the _mottes_ of the Merovingians in France. On the Bayeux tapestry three fortified places are represented--Dinan, Dol, and Rennes--and all are of the same type. A mound of earth was either thrown up, or a hilltop was artificially shaped like a tumulus. On the top of this the _thegn_ erected his fortress of wood. In the Bayeux representations the superstructures at Dol and Rennes are of timber, and that of Dinan is partly of timber and partly of stone. A flying bridge of wood led from the gate in the palisading of the outer ring, supported on posts, and conducted by an incline to the gate of the citadel. An example of one of these camps at Bishopston in Gower has been explored recently.[11] The stumps of the pales were there found embedded in the clay of the bank, in tolerable preservation. In the valley below Burleigh Camp, commanding the ancient road from Exeter by Okehampton to Launceston, was a third camp, that has been for the most part obliterated; it occupied a rising knoll of limestone, and this latter has been quarried, so that the camp earthworks have been either destroyed or buried under the accumulations from the quarry. The locality is of great interest. The ridge goes by the name of Galford, and there is reason to think that this was the Gavulford of the Anglo-Saxon Chronicle, where, in 823, the Britons made their last stand against Egbert and the Saxons of Devon.