As a matter of fact, the Saxons trusted little to walls: their strength, after they had settled in the country, lay neither in earthwork, nor in stonework, but in the boundary of wood or marsh that extended round their settlements. Consequently, during the six centuries and a half between the final departure of the Roman legions and the Norman conquest, the history of military construction in England is very obscure. Work in stone, which can be distinguished as Saxon, is practically confined to churches. Such fortifications of this long period as can be identified are entirely in earthwork. In only a few cases we hear of a stone wall of _enceinte_ being built, or an old Roman town wall being repaired. Further, it may safely be said that these fortifications, at any rate until the end of the period, whether their builders were Saxons or Danes, were intended to protect, not private individuals, but a community. Of the private citadel or castle we hear nothing until the period immediately before the Conquest, and then it is heard of only as a foreign importation. The most formidable earthworks of the Saxon period are the great dykes known as Wansdyke and Offa’s dyke, with the subsidiary works of the Bokerley dyke and Wat’s dyke. Offa’s dyke, which ran from the Dee in the north to the Wye in the south, with a ditch along its western side, and the parallel line of Wat’s dyke,[24] are generally acknowledged to have formed the boundary line between the Mercian kingdom of Offa (757-96) and the territory of the conquered Britons. The object and date of Wansdyke and the Bokerley dyke is not so clear. The Wansdyke ran from the Bristol Channel near Portishead, across north Somerset and along the downs south and south-west of Bath, passed through Wiltshire, north of Devizes and south of Marlborough, and, leaving Wiltshire east of Savernake park, turned southwards in the direction of Andover. The Bokerley dyke, in its present state, is only some four miles long, and forms the boundary between Wilts and Dorset, on the road from Salisbury to Cranborne. In both cases the ditch is upon the north or north-east side of the bank or dyke, which is clear proof that the defence was provided against attack from that quarter. The Wansdyke is obviously a late Roman or post-Roman work, for it encroaches in places upon the adjacent Roman road. The system on which it is planned resembles that of the Roman wall, in that its course includes a series of forts, presumably of earlier date. The conclusion which seems irresistible is that propounded by the late General Pitt-Rivers, that the Wansdyke was raised by the Roman Britons, to defend their last refuge in the south-west against the invading Saxons. If this is really the case, one can only wonder at the energy of despair which constructed this huge rampart, and at the uselessness of its builders’ attempt to ward off an invasion from the inland country alone. Ceawlin’s victory at