the very low state of Welsh architecture at this time; it may be paralleled only by the artless fabric of a cattle-house." This certainly is a hall and not a castle.

2

The so-called Dimetian Code indeed tells us that the king is to have a man and a horse from every hamlet, with hatchets for constructing his castles (gestyll) at the king's cost; but the Venedotian Code, which is the older MS., says that these hatchet-men are to form encampments (uuesten); that is, they are to cut down trees and form either stockades on banks or rude zerebas for the protection of the host. It is clearly laid down in the Codes what buildings the king's villeins are to erect for him at his residences: a hall, buttery, kitchen, dormitory, stable, dog-house, and little house. In none of these lists is anything mentioned which has the smallest resemblance to a castle, not even a tower. We can imagine that these buildings were enclosed in an earthwork or stockade, but it is not mentioned.*

3

Wales was never one state, except for very short periods. Normally it was divided into three states, Gwynedd or North Wales, Powys or Mid-Wales, and Deheubarth, all almost incessantly at war with each

WALES AND THE SAXONS

253

as

other.1 Other subdivisions asserted themselves opportunity offered, so that the above rough division into provinces must not be regarded as always accurate. A Wales thus divided, and perpetually rent by internal conflicts, invited the aggression of the Saxons, and it is probable that the complete subjugation of Britain would have been accomplished by the descendants of Alfred, if it had not been for the Danish invasions. The position of the Welsh kings after the time of Athelstan seems to have been that of tributaries, who threw off their allegiance whenever it was possible to do SO. But still the Anglo-Saxon frontier continued to advance. Professor Lloyd has shown, from a careful examination of Domesday Book, that even before the Norman Conquest the English held the greater part of what is now Flintshire and East Denbighshire, and were advancing into the vale of Montgomery and the Radnor district. The victories of Griffith ap Llywelyn, an able prince who succeeded in bringing all Wales under his sway, devastated these English colonies; but his defeat by Earl Harold in 1063 restored the English ascendancy over these regions. The unimpeachable evidence of Domesday Book shows that a considerable district in North Wales and a portion of Radnor were held respectively by Earl Edwin and Earl Harold before the Norman Conquest. Moreover, the fact mentioned by the Anglo-Saxon Chronicle in 1065, that Harold was building a hunting-seat for King Edward at Portskewet, after he had subdued it, suggests that the land between Wye and Usk, which Domesday Book reckons under Gloucestershire, was a conquest of Harold's."

2

The Norman Conqueror was not the man to slacken his hold on any territory which had been won by the Saxons. But there is no succinct history of his conquests in Wales; we have to make it out, in most cases, from notices that are scarcely more than allusive, and from the surer, though scanty, ground of documents. It is noteworthy that the Anglo-Saxon Chronicle is so hostile to the Norman kings that it discounts their successes in Wales. Thus we have only the briefest notice of William I.'s invasion of South Wales, which was very probably the beginning of the conquest of that region; and several expeditions of William II. are spoken of as entirely futile, though as we are told that the existing castles were still held by the Normans, or new ones were built, it is clear that this summing-up is not strictly correct. Our Welsh authorities, the Annales Cambria and the Brut y Tywysogion, seem to give a fairly candid account of the period, although the dates in the Brut are for the most part wrong (sometimes by three years), and they hardly ever give us a view of the situation as a whole. They tell us when the Welsh rushed down and burnt the castles built by the Normans

WALES AND THE NORMANS

255

in the conquered districts, but do not always tell when the Norman recovered and rebuilt them.

1

Fortunately we are not called upon here to trace the history of the cruel and barbarous warfare between Normans and Welsh. No one can turn that bloodstained page without wishing that the final conquest had come two hundred years earlier, to put an end to the tragedy of suffering which must have been so largely the portion of the dwellers in Wales and the Marches after the coming of the Normans. Our business with both Welsh and Normans is purely archæological. We hold no brief for the Normans, nor does it matter to us whether they kept their hold on Wales or were driven out by the Welsh; our concern is with facts, and the solid facts with which we have to deal are the castles whose remains still exist in Wales, and whose significance we have to interpret.

"Wales was under his sway, and he built castles therein," says the Anglo-Saxon Chronicle, in summing up the reign of the Conqueror; a passage which is scarcely consistent with its previous almost complete silence about events in Wales. There can be little doubt that William aimed at a complete conquest of Wales, and that the policy he adopted was the creation of great earldoms along the Welsh border, endowed with special privileges, one of which was the right of conquering whatever they could from the Welsh. To these earldoms he appointed some of his strongest men, men

2

little troubled by scruples of justice or mercy, but capable leaders in war or diplomacy. It was an essential part of the plan that every conquest should be secured by the building of castles, just as had been done in England. And we have now to trace very briefly the outline of Norman conquest in Wales by the castles which they have left behind them.

We shall confine ourselves to those castles which are mentioned in the Brut y Tywysogion, the Pipe Rolls, or other trustworthy documents between 1066 and 1216, the end of King John's reign. Of many of these castles only the earthworks remain; of many others the original plan, exactly similar to that of the early castles of Normandy and France, is still to be traced, though masked by the masonry of a later age. Grose remarked but could not explain the fact that we continually read of the castles of the Marches being burnt and utterly destroyed, and a few months later we find them again standing and in working order. This can only, but easily, be explained when we understand that they were wooden castles built on mottes, quickly restored after a complete destruction of the wooden buildings.

North Wales appears to have been the earliest conquest of the Normans, though not the most lasting. North Wales comprised the Welsh kingdoms of Gwynedd and Powys. Gwynedd covered the present shires of Anglesea, Carnarvon, and Merioneth, and the mountainous districts round Snowdon.1 Powys stretched from the estuary of the Dee to the upper course of the Wye, and roughly included Flint, Denbigh, Montgomery, and Radnor shires. Hugh of Avranches, Earl of Chester, was the great instrument of Norman