There is no trace of any _arx_ or citadel within these enclosures. Their defenders were the citizens. In France, for reasons sufficiently indicated, the art of fortification was more advanced. Roman traditions survived there without that abrupt break which the historical continuity of England had suffered. No fortress of stone and wood, such as that which Charles the Bald, in 869, built within the _enceinte_ of the abbey of St Denis, is heard of in Saxon England. The state of society in which, as early as 864, Charles found it necessary, by the edict of Pistes, to forbid his vassals to raise private fortresses without royal authority, did not exist in England, and was only beginning to exist there two centuries later. In both centuries we have to deal with the same invaders, but with defenders whose state of social development was quite different. Although the private fortress or castle was introduced into England by the Normans, and although the type of earthwork associated with it was developed to its highest extent by Northmen, not only in Normandy and England, but also in Denmark, it is nevertheless probable that the earliest development of that form of earthwork took place on Frankish soil. The Danish _geweorc_ or _burh_, where it can be traced with any certainty—and this is in very few cases—supplied accommodation for the force, and probably a harbour for its vessels, but no private stronghold belonging to a prominent leader. The earthwork called Gannock’s castle (32), close to the Ouse near Tempsford, is sometimes supposed to be the _geweorc_ wrought by the Danes in 921. In plan, it very closely resembles a rather small mount-and-bailey castle of the usual early Norman type, and could have accommodated only a very small body of defenders. But the point in which it differs from the ordinary mount-and-bailey plan—the smallness of the mount, which is a mere thickening of the earthen bank, and the absence of a moat round its base—may show that here the Danes anticipated their Norman successors with a plan with which some of them may have gained acquaintance during marauding expeditions in France. This, however, is mere conjecture, and the utility of such a fortress for the immediate purposes of the Danes may well be questioned.[37] Of the private dwellings of the Danish leaders, and how far they may have approximated to the later type of the castle, we know nothing. St Mary’s abbey at York is on the site of Galmanho, the residence of the Danish earls outside the western wall of the Roman city, but nothing of its earthworks remain. If they were at all considerable, like those of a mount-and-bailey castle of Norman times, it seems strange that no trace of them should be left.