Chapter 9

THE DEFENCE OF THE ANGLO-NORMAN KINGDOM

March 1067–November 1085

In the summer of 1067, as William moved in triumph from Rouen to Fécamp, from the Dives to Jumièges, he was not only, as few of his predecessors had been, effectively master of Normandy, but he was also the consecrated and acknowledged king of the English. None the less his position was by no means secure. In France, Maine and Brittany were restive, and the French monarchy, whose heir was growing up to manhood, was ill-disposed towards its most powerful vassal. In England, only a part of the country was as yet under Norman control, and beyond the English frontiers, which were themselves ill-defined, stood apprehensive the Welsh princes and the Scottish king. Finally, there remained the longstanding opposition of Scandinavia to any control of England by Normandy. The defeat of Harold Hardraada at Stamford Bridge had prepared the way for William's victory, but other northern rulers would not lightly relinquish their ancient claims on a country which had recently formed a part of the Scandinavian political world.

Thus while the decision of 1066 was to mark an epoch, it was not in itself final, and it had to be confirmed. Its fulfilment in fact depended upon three main conditions. It was essential that Norman strength, as developed during the past fifty years, should be maintained, so that the duchy should retain its predominant place among the powers of northern Gaul. Secondly, the conquest of England had to be completed, and the surviving elements of opposition to the new order reduced to obedience. Thirdly, the continuing Scandinavian threat to the Anglo-Norman state had to be withstood. These three problems were, moreover, closely linked, and the manner in which one or other of them was dominant at any period of the reign can be roughly deduced from the movements of William himself. From the end of 1067 to 1072 he was primarily engaged in suppressing English rebellions, and establishing his power. From 1073 to 1085 he spent most of his time in Normandy. And, throughout, he had constantly to withstand attacks from Scandinavia. These took place, for instance, in 1069 and 1070; the threat appeared again in 1075; and it was the menace of another and very formidable attack which brought the Conqueror back to England towards the end of 1085. Between December 1085 and September 1087 he was to make some of his most lasting contributions to the future development of England, but, none the less, his last days were spent in defensive warfare on the Norman frontiers, and he was to die within sound of the church bells of Rouen.

In 1067, however, it was the English situation which was the most precarious. The task of the regents William fitz Osbern and Bishop Odo was not easy. They had substantial control over the south-east of the country, and the formal submission of the chief English magnates gave them a claim to obedience elsewhere. Nevertheless there were many who were ready to take advantage of the unsettled conditions, and prominent among them was a west-country magnate named Edric the Wild who raised a revolt in Herefordshire, and called to his assistance the Welsh princes, Bleddyn and Riwallon.¹ They did much damage but failed to obtain control of the shire, and retired with their booty back into Wales, where they prepared for further raiding. Meanwhile, a better organized revolt took place in Kent, where the insurgents called to their assistance Eustace, count of Boulogne.² It was a strange appeal, since Eustace in the previous year had fought at William's side at Hastings. He may now, however, have been influenced by the death on 1 September 1067 of his overlord Count Baldwin V of Flanders, who in 1066 had been friendly, or at least neutral, towards William.³ At all events, Eustace, doubtless recalling his earlier adventures in Kent in 1051, crossed over the Channel with a substantial contingent of knights. Both the regents were at that time north of the Thames, and Eustace was enabled to occupy the town of Dover. But he failed to take the newly erected castle, and a sortie from the garrison cut his force to pieces; whereupon he made an ignominious escape across the Channel. Neither of these risings had in fact seriously disturbed the new government. But a threat of more serious danger was already foreshadowed. According to William of Poitiers, overtures were now being made to the Danes,⁴and an invasion of England by Sweyn Estrithson was becoming an imminent possibility. It was in fact probably with this apprehension that on 6 December 1067 William himself came back from Normandy to England.⁵

On his return, however, his immediate attention was directed towards the south-west. The city of Exeter refused to accept the new régime, and sought to form a league of resistance among the neighbouring towns. William's reply was to march at once into Devonshire at the head of a force which included many English mercenaries. The thegns of Devonshire seem to have accepted the new king, but Exeter itself held out against him for eighteen days, and then at last only surrendered on the understanding that its ancient privileges should be confirmed.⁶ Thereupon the king built a castle within the city, and proceeded into Cornwall, where was soon to be established his half-brother, Robert, count of Mortain.⁷ Resistance in the south-west in fact was everywhere breaking down. To this period must be assigned the submission of Gloucester, and evidently Bristol had also by now accepted the new order.⁸ Indeed, in this same summer, when three of the illegitimate sons of Harold came over from Ireland, they were repelled by the citizens of Bristol, and on their retreat their force was routed by the thegns of northern Somerset.⁹ William's rapid campaign in the south-west had in fact been so successful that he could return at once towards his capital. He celebrated Easter 1068 at Winchester, and at Whitsuntide he held a great court at Westminster which was attended by many English notables. Thither too came Matilda, his wife, to be solemnly crowned as queen.¹⁰

These ceremonies were impressive, but the respite they reflected was short-lived. Very soon there were significant desertions from the new king's court. Edgar the atheling had already taken refuge with Malcolm, king of Scotland, and now Earls Edwin and Morcar departed to their earldoms. The Norman settlement had as yet barely affected the north, where despite the efforts of Archbishop Aldred a serious movement of resistance was rapidly growing. During the period immediately following King William's accession, Northumbria had been disputed between Osulf, a protégé of Earl Morcar, and Copsige a former adherent of Tosti. Both had perished in the struggle between them, but in 1068 local resistance was forming not only round Earl Morcar himself but also round Gospatric, a descendant of the more ancient Northumbrian house. Appeals were also being made both to Malcolm and to Sweyn Estrithson. In these circumstances, William thought it necessary to move northward at once. He went first to Warwick, where he placed Henry of Beaumont in charge of a newly built castle and then on to Nottingham. From there he moved up into Yorkshire and was able to enter York without a battle. Having received the formal submission of many of the local magnates, he negotiated a temporary truce with the Scottish king, and after erecting a castle on the spot now marked by Clifford's Tower, he turned southward to enforce the submission of Lincoln, Huntingdon, and Cambridge.¹¹

William's extraordinary activity during the first nine months of 1068 deserves note, for during this short period he had conducted an almost uninterrupted series of campaigns which had brought him in turn to Exeter, to Warwick, to York, and then through a large part of eastern England. Yet such respite as had thus been gained was to be short-lived, for the north was soon again to demand his presence. Late in 1068 a Norman, Robert de Commines, had been sent, with the title of earl, to restore order north of the Tees. But, on his arrival at Durham, he was, on 28 January 1069, set upon in the streets of that city, and subsequently burnt to death in the bishop's house.¹² The news quickly passed to York, where the Norman garrison was immediately attacked by local insurgents, and on hearing of these events Edgar Atheling at once made preparations to move down from Scotland. King William was thus forced to return to the north with all speed, and his march in this respect deserves some comparison with that of Harold to Yorkshire in 1066. The king was in York before his enemies expected him, and he dispersed the besiegers of the castle, thus taking possession of the city for the second time. On this occasion he took signal vengeance on the rebels, and set up a new castle near the town. Then, doubtless as a temporary measure, he placed Gospatric in charge of the earldom. But he could not himself afford to delay, and by 12 April 1069 he was back again at Winchester.¹³

Perhaps the most salient feature of the confused events that occurred in England between the beginning of 1067 and the spring of 1069 was the comparative ease with which the king, and his lieutenants, having at their disposal only a limited number of troops in a newly occupied land, were enabled to suppress each rising as it occurred. This may partly be ascribed to the severe losses which had assailed the warrior class in England during the great battles of 1066. Still more must it be attributed to the lack of any common purpose among the insurgents whose efforts were made in isolation and without any contact with one another. Again, there was, almost from the first, a substantial body of opinion which was favourable, or at least not actively hostile, to the new régime. Many of the ecclesiastics who had been appointed during the reign of Edward the Confessor such as Giso, bishop of Wells, William, bishop of London, or Baldwin, abbot of Bury St Edmunds were committed to the cause of the new king, and it is yet more significant that respected English prelates such as Wulfstan of Worcester and Aldred of York were likewise ready to support him. Nor could their example be wholly disregarded by lesser men. The ‘Englishmen’ who marched under King William to the assault of Exeter in 1068 may have been for the most part adventurers serving for pay, but there were many thegns and local officials who were ready to accept William as their king, and to help carry on the administration which he conducted. It was the thegns of Somerset who repelled the sons of Harold, and sheriffs of native stock found themselves with Bishops Giso and Wulfstan the recipients of the new king's writs.¹⁴

It was William's achievement to exploit these favourable conditions, by developing such advantages as he possessed in the techniques of war. The trained mounted troops who had stood him in such good stead at Hastings could be employed in effecting the swift reduction of scattered forces in a hostile countryside, but they were little use in operations against cities, and, by themselves, they were ill-adapted for holding down regions whose doubtful submission had only been hazardously acquired. For this purpose the establishment of fortified strong-points was essential, and contemporaries are unanimous in their opinion that much of William's success in these campaigns was due to his use of the castle¹⁵ which, as has been seen, had been consistently developed in Normandy, both by the dukes and by the new aristocracy.

The employment of the castle, not only as a fortified centre of administration but also as a means for conducting a campaign, had already before 1066 become a normal feature of Norman military life. In England, on the contrary, except in Normanized Herefordshire, its use had hardly been adopted, and it was regarded as a continental innovation of doubtful value. To the absence of castles, indeed, Ordericus Vitalis, who is here following William of Poitiers, attributes the lack of success of the opponents of King William in England during this warfare.¹⁶ Such was the English situation. By contrast, William in his English campaigns evidently employed precisely the same device which he had earlier used in France, and the Bayeux Tapestry shows no substantial difference between the castles at Dol, Rennes, or Dinant, and that which was erected at Hastings in 1066.¹⁷ Here, again, the castle is of the typical motte and bailey pattern which was already familiar in the duchy – that is to say, an earthen mound surrounded by a fosse, and surmounted by a palisaded rampart, crowned with a wooden tower. It could be speedily constructed, and it proved highly effective. Castles of this type were thus at once set up at Pevensey, at Hastings, and even at London, where an erection of this type, built about the time of the coronation, preceded the stone fortification which was to be the Tower of London.¹⁸ It was, however, during the campaigns of 1067, and more particularly in 1068, that the Norman castle as an instrument of war was fully used for the first time in England. The surrender of Exeter was marked by the beginnings of ‘Rougemont Castle’, and as William proceeded northwards the same plan was continued at Warwick and Nottingham. York at this time received its first castle, and on William's return march castles were erected at Lincoln, Huntingdon, and Cambridge.¹⁹

Some idea of the importance attached to these castles can be obtained by a contemplation of the men to whom they were entrusted. For these were chosen from among the most important of the Norman magnates. Thus Dover, within the special province of Odo of Bayeux, was put under Hugh of Montfort-sur-Risle.²⁰ The castle at Hastings, first given in charge to Humphrey of Tilleul, was soon to be the responsibility of Robert, count of Eu.²¹ The castle at Exeter was given to Baldwin of Meules, the brother of Richard fitz Gilbert, later of Clare, and the son of Gilbert of Brionne, the count. Warwick was assigned to Henry of Beaumont, brother of Robert, and son of the veteran Roger of Beaumont, who had in 1066 been left in Normandy as one of the regents.²² The first castle at York was entrusted to William Malet of Graville-Sainte-Honorine, near Le Havre,²³ and the second castle at York was to be given to William fitz Osbern himself.²⁴ Soon the Norman castles in England were to multiply, and what in 1068 was essentially a device of war came to be a permanent feature of the new feudal administrative order which the Norman here established. Indeed, before the end of the eleventh century there had been erected in England at least eighty-four castles, and a few of these were even then being reconstructed in stone.²⁵ Already, however, by the beginning of 1069, the motte and bailey castle, in charge of a trusted lieutenant of the king, was proving itself an essential, and a highly effective, instrument by which the conquest of England might be completed.

Yet, when all is said, much of William's success between January 1068 and the summer of 1069 must be attributed to the phenomenal energy which he personally displayed at this time. To appreciate the quality of his astonishing activity it is, however, necessary also to note the possibility that during these months, possibly towards the end of 1068,²⁶ but more probably in the early summer of 1069,²⁷ he found it necessary to return to Normandy. Perhaps he deemed it necessary at this critical juncture in his affairs to display his authority south of the Channel, and it is at all events certain that about this time Matilda returned to Normandy so that she might be received with royal honours in the duchy.²⁸ But his sojourn in Normandy, if in fact it occurred, must in any case have been brief,²⁹ for the English situation demanded his constant vigilance. By the summer of 1069 William might be said to have established himself in effective control over most of England south of the Humber. Now, however, the whole Norman position in England was to be tested more drastically than ever before.

In the summer of 1069 Sweyn Estrithson launched his long anticipated attack upon England. It was planned on a scale comparable with that of the invasion of Harold Hardraada of Norway three years previously. A fleet of 240 ships sailed under the leadership of King Sweyn's sons, Harold and Cnut, and of Osbern his brother. It brought to England an army of trained warriors which included many men of high rank in Denmark, and the threat of the expedition was enhanced by the fact that it could count on considerable support within those regions of England which had Scandinavian affinities. The ships first appeared off Kent, and then proceeded up the east coast. Raiding parties were thrown off, but were repelled, and at length the fleet reached the safe anchorage of the Humber. Its arrival was the signal for a general rising in Yorkshire. Edgar the atheling, Gospatric, and Waltheof collected a considerable force, and forthwith joined the Danes. Then the whole body marched upon York. The Norman garrisons were unable to hold the castles, and on 19 September they sallied out and started to fire the city. They perished after prolonged fighting, and on 20 September York fell. The Danes thereupon repaired to their ships, and, having moved across to the southern shore of the Humber, they fortified the isle of Axholme. Many of their troops dispersed over the countryside of North Lincolnshire, where they were welcomed by the peasantry, and entertained by them at the village feasts.³⁰

The whole Norman venture in England had thus been placed in peril, for at last the resistance to William was assuming a coherence which it had hitherto lacked. A strong Scandinavian force was at large in England, and it was supported by a considerable army led by powerful Saxon magnates. Nor is it surprising that the news of these events spread rapidly through England, and gave occasion for revolts elsewhere: in Dorset and Somerset, for example, in Staffordshire and South Cheshire.³¹ But the centre of the crisis was in the north. Yorkshire had been lost, and beyond Yorkshire in ‘Saint Cuthbert's land’, north of the Tees, there was whirling chaos in which was emerging the authority of Malcolm, king of Scotland. The Scottish king had in fact now thrown in his lot with William's opponents in England, and it was probably about this time that, by one of the most influential marriages in English history, he allied himself to Margaret, the sister of Edgar Atheling.³² The possibilities latent in the developing situation were in fact incalculable, and in the autumn of 1069 it must have seemed possible that a Scandinavian kingdom might once more be established in northern England, or even a realm created for Edgar Atheling, buttressed by the support of Malcolm and Sweyn, and perhaps even to be sanctioned with a separate coronation by a metropolitan archbishop of the distinct ecclesiastical province of York.

The magnitude of this crisis indicates the importance of the ensuing campaign, and explains (though it does not excuse) its terrible sequel. Never did William act with more vigour or at greater risk. He immediately moved up towards Axholme, where his approach caused the Danes to move back again over the Humber into Yorkshire. Then leaving the counts of Mortain and Eu to watch the situation in Lindsey, the Conqueror struck westward to cope with the rebellion which had broken out under Edric the Wild and the Welsh princes. This he suppressed, apparently without much difficulty, and then at once advanced towards Lincolnshire, leaving Geoffrey, bishop of Coutances, to cope with the Dorset rising that was threatening the newly constructed castle of Montacute. When the king reached Nottingham, however, he learnt that the Danes were preparing to reoccupy York, and so he turned northward. He found the Aire gap defended against him, but after some delay he managed to effect a crossing, and advanced directly on the northern capital which the Danes evacuated once more. On his route he savagely devastated the land through which he passed, sparing no male and leaving nothing behind him which could support life. Just before Christmas he reached York, and there in a burnt city, surrounded by a desolated countryside, he celebrated the Nativity of Christ.³³

The devastation which the king carried out on his march had been part of a rapid and critical campaign. That which he now ordered was inspired by a more cold-blooded design. The Norman troops split up into smaller bands and carried out a systematic harrying of Yorkshire. So terrible was the visitation that its results were still apparent twenty years later. But the king himself could not tarry in Yorkshire. In appalling weather he moved up to the Tees on a rapid and hazardous raid, and then without pause he set out upon what was probably the most difficult and arduous march of his career. Realizing that the western rebellion was barely suppressed, and that Chester still remained as the one outstanding centre of resistance, the king struck directly right across the Pennine Chain. It was the depth of winter, and the hardships of the route caused even his seasoned troops to threaten mutiny. But he pushed on, harrying as he went, and reached Chester before his enemies were ready to meet him. He occupied the city without difficulty, and placed a castle there, and also at Stafford. The resistance which had so nearly overwhelmed him was finally broken, and the Danish fleet, seeing its English allies defeated, accepted a bribe to depart from the Humber. The king himself moved back to the south. He reached Winchester before Easter 1070.³⁴

King William's campaign of 1069–1070 must rank as one of the outstanding military achievements of the age, and it was to prove decisive in ensuring that the Norman domination of England would endure. None the less, the cost of that achievement and its consequences deserve note in any estimate of the Norman impact upon England, and of the character of William the Conqueror. An eleventh-century campaign was inevitably brutal, but the methods here displayed were widely regarded as exceptional and beyond excuse,³⁵ even by those who were otherwise fervent admirers of the Norman king.

On many occasions [writes one of these] I have been free to extol William according to his merits, but I dare not commend him for an act which levelled both the bad and the good in one common ruin by a consuming famine … I am more disposed to pity the sorrows and sufferings of the wretched people than to undertake the hopeless task of screening one who was guilty of such wholesale massacre by lying flatteries. I assert moreover that such barbarous homicide should not pass unpunished.³⁶

Such was the view of a monk in Normandy. A writer from northern England supplies more precise details of the horrible incidents of the destruction, and recalls the rotting and putrefying corpses which littered the highways of the afflicted province. Pestilence inevitably ensued, and an annalist of Evesham tells how refugees in the last state of destitution poured into the little town. Nor is it possible to dismiss these accounts as rhetorical exaggerations, for twenty years later Domesday Book shows the persisting effects of the terrible visitation, and there is evidence that these endured until the reign of Stephen. Yorkshire, whose prosperity was thus destroyed for more than a generation, was, moreover, not the only region to suffer, for the devastation, though to a diminishing degree, spread as far west as Merseyside, and as far south as Derby.³⁷

Never again did William in England have to face such perils as those which menaced his rule in 1069–1070, and such subsequent opposition as he encountered was in the nature of an aftermath to the storm which had recently ravaged the country. The continuing centre of disturbance was the Danish fleet which had returned to the Humber. To it in the spring of 1070 came King Sweyn himself, and under his leadership the Danish force sailed southward towards the Wash.³⁸ The soldiers entered East Anglia, and entrenched themselves in the Isle of Ely, where they were joined by men of the countryside, and in particular by a Lincolnshire thegn named Hereward.³⁹ The first objective of the composite force was the abbey of Peterborough.⁴⁰ Brand the abbot, who had supported the cause of Harold, had recently died, and the abbey had been given to a certain Turold, who took possession with a considerable body of troops with which he hoped to hold down the countryside. On 2 June, however, a miscellaneous force composed of Danes and English, called vaguely by the chronicler ‘Hereward and his company’, entered Peterborough and gave the abbey over to fire and loot. It was an outrage which challenged constituted order in the region, but at first William did not go further than to negotiate with the Danes. Doubtless by means of another bribe, King Sweyn was persuaded to a truce, and at last the great Danish fleet set sail for home laden with booty. It had been off the coasts of England for nearly two years, and its departure marked a definite stage in the final settlement of England.⁴¹

In particular, the sailing of the Danish fleet was to prove fatal to the cause of Hereward. William, who was at this time preoccupied with the much more formidable threat to his power that was developing on the Continent, did not at first think it necessary to proceed against the outlaw, and his delay made the situation in the Fens more serious than it need otherwise have been, for in the absence of the king, Hereward was now joined by other more prominent men, including Earl Morcar himself. But without Scandinavian support the rising in the Fens was doomed to failure and when William advanced against Ely in person, the rebels surrendered unconditionally. Earl Morcar was taken prisoner, and Hereward, having escaped with difficulty, passed out of history into legend.⁴²

The Norman régime in England had thus survived the first disturbances which followed its establishment. The chief English cities had submitted; the north had been subdued; the Fenland rebellion suppressed; Earl Morcar was a prisoner and soon to die; whilst Earl Edwin about this time was killed by his own followers while flying to Scotland. Nevertheless, the problems facing King William had only been partially solved, and indeed they were now in some measure to be increased. To treat the resistance which his government met in England as if it was something to be regarded in isolation is to misconceive the nature of the dominion which he had established. This was a conjoint dominion stretching across the Channel, and politically united under a single rule. Any attack upon any of its parts from any quarter was equally a menace to its survival, and both its enemies and its defenders were to show themselves well aware of the fact. Indeed, it was to become a cardinal feature of the policy of King Philip I of France to exploit this situation. The intimate connexion in this respect of English and continental politics between 1067 and 1085 deserves, therefore, more emphasis than it usually receives. The suppression of risings in England was always connected with the imminence of attacks from Scandinavia or Scotland, from Anjou or Maine; and throughout all this period the maintenance of the northern frontier beyond Yorkshire could never be dissociated from concurrent threats from France, from Flanders, or from the Baltic lands. It was only by means of a far-flung and integrated defence that the Anglo-Norman kingdom was to survive under the rule of William the Conqueror.

The freedom from attack enjoyed by Normandy during the period of the English conquest had been a prime factor in its success, but this continued immunity could not be presumed, and it was now to be disrupted. In 1069, even while the king was conducting his desperate campaign in the north of England, the city of Le Mans revolted against Norman rule.⁴³ The event might doubtless have been expected but it was none the less sudden. Arnold, the bishop who succeeded Vougrin at Le Mans in 1065, was a nominee and partisan of King William, and, to judge from a charter alleged to have been given to the abbey of La Couture in 1068, Norman administration was still operative in Maine in that year.⁴⁴ Now, however, a powerful party in Maine supported the citizens of Le Mans in the interests of Azzo, lord of Este in Liguria, who was the husband of Gersendis, the sister of Count Hugh IV. Azzo, who arrived in Maine before 2 April 1069, managed to collect a large body of adherents, and in particular attracted to his cause Geoffrey of Mayenne, the powerful border lord whose influence had so often been decisive in the politics of the comté. The confederation thus formed was too strong for the Norman rulers of Maine to resist. A certain Humphrey, described as the senescallus of King William was killed, and the Norman knights were expelled, among them being William of La Ferté-Macé, a brother-in-law of Odo, bishop of Bayeux. Azzo, having achieved thus much, now retired to his Italian lordship, leaving in charge of Maine, Gersendis and their young son Hugh, who was recognized as count. Geoffrey of Mayenne, who forthwith took Gersendis for his mistress, remained the dominant figure in the partnership.⁴⁵

From the first it was an unstable government. In March 1070 the citizens of Le Mans revolted once more, this time against Geoffrey, and formed themselves into what was described as a commune.⁴⁶ They forced Geoffrey to recognize their claims to special privileges, but were unable to maintain their position. Setting out in company with their bishop to reduce the castle of Sillé, which was held against them, they were betrayed by Geoffrey and routed. Geoffrey himself, however, did not think it safe as yet to re-enter the city to rescue the countess. He took refuge therefore in Château-du-Loir, and the young Hugh was sent for safety to his father in Italy. None the less, before the end of the year, the revolt was finally crushed, and the countess and Geoffrey were together re-established in Le Mans.⁴⁷

These events must have caused the greatest concern to King William in England. Within a few months the Norman rule in Maine, which had been operative since 1063, had collapsed, and the disturbance of Maine invited the intervention of some stronger power which as in the past might use the comté as the base for an attack upon the duchy. Moreover, before the end of 1070 there had occurred another revolution on the Continent which entailed peril to Normandy. On 16 July 1070, within five weeks of the departure of Sweyn's fleet from East Anglia, there died King William's brother-in-law, Baldwin VI, count of Flanders. A succession question was immediately opened with which Normandy was to be at once concerned. The two young sons of Baldwin VI, namely Arnulf and Baldwin, received respectively Flanders and Hainault, and in view of their youth the government was conducted by their mother, Richildis. Her rule was strongly resisted, particularly in Flanders, and the opposition was headed by Robert, ‘le Frison’, a son of Baldwin V. Richildis immediately sought the aid of King Philip I, and, looking about for further assistance, turned to William fitz Osbern, the closest personal associate at this time of King William. Early in 1071 the king had sent the earl to Normandy, doubtless to watch over the developing danger from Maine, and now Richildis offered herself in marriage to William fitz Osbern, and placed her son Arnulf in his wardship. He accepted the offer, and, in support of his ward and of his designated wife, he hastened to Flanders, ‘as if to a game’, accompanied, as it was said, by only ten knights. The decisive battle was fought at Cassel on 22 February 1071, and it resulted in the overthrow of Richildis, the establishment of Robert le Frison as count of Flanders, and the death of William fitz Osbern.⁴⁸ King William thus lost his most powerful secular supporter in England, and, within a few months of Maine slipping from his control, he saw established a hostile power in the Low Countries.

During 1070–1071 the events in Maine and Flanders must have disturbed William far more than the continuing resistance of Hereward in the Fens, and even in Britain there were other matters which were beginning to press more urgently on his attention. The Norman régime which had been established in England confronted both Celtic Wales and Celtic Scotland with a challenge. On the Welsh border the great Norman palatine lordships were about this time being established,⁴⁹ and the defeat of Bleddyn and Riwallon set the stage for a new Norman movement westward. But in 1070 it was Scotland that was most immediately affected. Ever since 1066 Scotland had served as a refuge for disinherited English magnates. Edgar Atheling remained an honoured guest at Malcolm's court; and Malcolm had married Edgar's sister Margaret. In these circumstances, the victorious campaigns of William through the north of England during the winter of 1069–1070 was bound to provoke an immediate reaction from Scotland, which now could provide a new and most formidable threat to the Anglo-Norman kingdom.

The situation was, indeed, of wider significance then even these facts would by themselves suggest. Here was a question of frontiers. The centre of Malcolm's dominion was the kingdom of Alban based upon Perthshire, and flanked to the north by the Scandinavian settlements, and to the south by the provinces of Cumbria and Lothian: Cumbria stretching from the Clyde to the Westmorland fells, and Lothian from the Forth southward; and both with their southern boundaries still undefined. Both these provinces were of vital interest alike to Malcolm and to William, and on both of them the impact of the campaigns of 1069–1070 were immediately felt. In this manner there was at once posed the question which was to dominate Anglo-Scottish relations for the next quarter of a century. What henceforward were to be the political filiations of Cumbria and Lothian, or (in other words) what was to be the northern frontier of the newly established Anglo-Norman state?⁵⁰ That question was not in fact to be even partially resolved before 1095, but in 1070 it was already urgent. The devastation of the north had created, so to speak, a vacuum of political authority in the debatable region, and already in the spring of 1070, almost before William had reached Winchester, Malcolm was carrying out a terrible devastation of Durham and Cleveland, whilst Gospatric, as earl, was taking reprisals against his former Scottish ally in Cumbria.⁵¹ It was a measure of the menace to William from the north.

The problem before the Conqueror at the end of 1071 was in its essence not wholly dissimilar from that which faced Harold Godwineson in the autumn of 1066. The Anglo-Norman state was being threatened on two of its extreme frontiers, and it was a question which danger should first be met. How pressing was the problem, and how closely connected were its two parts, can be judged by the astonishing rapidity of William's movements during the next fifteen months. In the winter of 1071–1072, with the northern menace behind him, he departed for his duchy.⁵² His actions in Normandy at this juncture are but sparsely recorded, but it may be assumed that the court he at once held was concerned with the situation in Maine, and it is significant that his half-brother, Odo of Bayeux, was also in the duchy at this time.⁵³ The king, however, could not tarry long, and before Easter 1072 he was back in England,⁵⁴ and there his first action was to begin the extensive preparation essential for coping with the threat from Scotland. The measures he took were in fact to have a considerable influence upon the establishment of the Norman nobility régime in England, and they imposed burdens not only on the Norman magnates but on the bishoprics and abbeys of England.⁵⁵ They were also executed with the utmost speed, and during the summer they were completed. In the early autumn therefore he was ready to act, and he forthwith ‘led a land force and a naval force to Scotland’.⁵⁶

So began one of the most remarkable of the many military ventures of the Conqueror. His plan was to make a two-pronged thrust by sea and land into the heart of Malcolm's kingdom. The army, consisting chiefly of horsemen, moved up by the eastern route through Durham, and then on through Lothian, crossing the Forth by the ford near Stirling to turn eastward towards Perth and the upper reaches of the Tay. The fleet, acting in conjunction, sailed up the east coast of Britain and entered the estuary of the Tay to make contact with the land forces. It was a bold plan, and it achieved a hazardous success. Doubtless, the Conqueror hoped for an engagement somewhere in Lothian where his horsemen might be able to show to advantage, but Malcolm was not disposed to afford him this opportunity. He was, however, so daunted by the invasion that he consented to negotiate, and the two kings met at Abernethy within a few miles of the Norman ships. As a result, Malcolm gave hostages to William and became his man.⁵⁷ Whether such homage was held to involve the kingdom of Alban itself, or merely lands in Cumbria and Lothian is uncertain. Nor was it of great significance. What was important was that the Scottish king had been brought formally to recognize the new régime in England, and as a token of this, Edgar Atheling was expelled from the Scottish court.

This campaign must be regarded as one of the most bizarre exploits of Norman arms in the eleventh century, and the risk which attended the enterprise of bringing knights from the Risle and the Seine, from the Bessin and the Hiémois up to the gates of the Highlands were very considerable. William, late in a campaigning season, was operating perilously far from his base, and even the ships which might serve in an emergency as a means of retreat could hardly diminish the menace of his isolation. In these circumstances, by securing from Malcolm a pact so far favourable to himself, he made at great risk a notable contribution to the defence of the Anglo-Norman state. Its existence had been formally recognized in the north; a centre for the assembly of its enemies had been neutralized; and its northern frontier had been asserted if not defined.

William's expedition to Scotland must, moreover, be viewed as part of a larger political strategy, and here time was of the essence of his task. The safety of the state he had created now involved a unified defence which stretched from Abernethy to Nonancourt, from Flanders to the boundaries of Brittany, and in the autumn of 1072 events on the Continent had made it imperative not only that a settlement should be obtained in the north, but that this should be accomplished with the minimum of delay. Maine was in revolt; Flanders had become hostile; and William's presence was urgently needed in Normandy. The preservation of the Anglo-Norman state had in fact become a single problem: it was essentially the same whether it was conducted in Britain or in France; and events in either country had immediate repercussions on the other. Immediately after his pact with Malcolm, William therefore moved south with speed. By 1 November he was at Durham.⁵⁸ Where he heard the Christmas Mass is not known – but early in 1073 he was back again in Normandy at the head of a large force which he had transhipped from England.

His return was not too soon, for even during the months of his Scottish campaign his position in France had again deteriorated. The instability of the government of Geoffrey of Mayenne invited intervention in Maine by a stronger power, and some time in 1072 a new turn was given to the situation when the citizens of Le Mans invited the help of Fulk le Rechin, count of Anjou. Anjou was no longer the force it had been in the time of Geoffrey Martel, but by 1072 the repulsive Fulk le Rechin had so far overcome the anarchy he had done much to create, that he could no longer be disregarded as an opponent. The invitation to the count was thus of considerable significance in reconstructing an earlier pattern of politics. Nor did Fulk hesitate to accept it. He immediately entered Maine, and advanced towards Le Mans. The citizens rose in his favour, and with the count's aid, Geoffrey of Mayenne was expelled. Once again a count of Anjou had gone some way towards establishing himself on the frontiers of Normandy.⁵⁹

Such was the developing situation which had brought the Conqueror with such haste from Scotland, and the unified character of the defensive campaigns on which he was engaged can be seen in the composition of the force which accompanied him. Even as Norman knights had been concerned in the expedition to Scotland, so now did English troops take part in the campaign in France, and their participation in the warfare which ensued was both noteworthy and noted.⁶⁰ Once again, the chief feature of William's plan was speed. Fulk himself seems to have left Anjou at this time, and William did not wait for his return. Although the campaigning season had barely begun he entered Maine by way of the valley of the Sarthe and attacked Fresnay. This stronghold together with the neighbouring fortress of Beaumont surrendered without much resistance, and William then attacked Sillé, which likewise submitted.⁶¹ The way was now open to Le Mans, which William invested and captured. The seizure of the capital entailed the general submission of the comté.⁶² By 30 March (as it would seem) it was all over, and the Conqueror was back at Bonneville-sur-Touques, having re-established the Norman ascendancy in Maine.⁶³

The rapidity with which the successful campaign in Maine followed the expedition to Scotland undoubtedly enabled the Conqueror to escape from a dangerous crisis in his affairs, and by the summer of 1073 his position was much stronger than it had been twelve months before. Nevertheless, the defence of his composite realm could not be relaxed, the more especially as a new factor was beginning to complicate the problem. The French monarchy was emerging from the eclipse into which it had entered with the minority of Philip I. In 1067 the young king left the tutelage of the count of Flanders,⁶⁴ and he was now developing an active policy of his own, against Normandy. For that purpose he sought allies, and these were available. He seems in the first instance to have turned towards Flanders. He had, it is true, in 1071 opposed the advancement of Robert le Frison, but when the latter was established, it was apparent that the interests of the Flemish count were here attuned to those of the French king. Robert, with justice, regarded the Norman duke as his most dangerous rival; he had not forgotten the intervention of William fitz Osbern in Flemish affairs; and it is significant that very soon after the pact between William and Malcolm he gave Edgar Atheling asylum in Flanders.⁶⁵Philip, on his side, in dealing with his over-mighty Norman vassal, had obviously much to gain from friendship with the count of Flanders. A rapprochement between them therefore rapidly took place, and it was symbolized by the marriage, in or before 1072, of Philip with Bertha of Hainault, who was Robert's half-sister.⁶⁶

Flanders was, moreover, not the only quarter to which the French king could turn for help against Normandy. There was also Anjou. Already by 1068, profiting by the war between Fulk and his brother Geoffrey for the Angevin inheritance, he had made a pact with the former by which he obtained the Gâtinais, and there is documentary evidence that in 1069 the relations between Philip and the count were friendly.⁶⁷ Fulk's incursion into Maine in 1072 must therefore have been welcome to the French king, and the situation could be developed. The fundamental characteristic of the policy of Philip I during the next twenty years was thus disclosed. It was to be a ceaseless opposition to Normandy, conducted in alliance with Flanders and Anjou. So consistently in fact was this policy pursued that William was compelled henceforth to spend most of his time in Normandy, and the defence of his kingdom for the remainder of his reign was to be concentrated in France, though (as in 1075 and 1085) it had constant repercussions upon England.

It was in fact during 1074 that William's opponents on both sides of the Channel began noticeably to act in concert. Thus in the course of that year Edgar Atheling returned from Flanders to Scotland where he was received with honour,⁶⁸ and the French king at once saw how he might be used as a centre for an alliance against Normandy. He therefore offered him the important castle of Montreuil-sur-Mer which would have placed him in a position of great strategic advantage.⁶⁹ For Montreuil was the chief Capetian outlet to the English Channel; it was within easy access of Flanders; and it was at the same time a base from the east. It seemed indeed as if Edgar might once again serve as a rallying point for all the enemies of the Anglo-Norman kingdom, and so seriously did William take the threat that he treated with the atheling, and consented to receive him back at his own court.⁷⁰ The French king had, therefore, to seek another centre of opposition to William, and he was to find this in Brittany where a situation was arising that could be made highly dangerous to Normandy. It was, indeed, through Brittany that between 1075 and 1077 there was now to develop a movement in which all William's opponents – English, French, and Scandinavian – became once more for a time associated.

As has been seen, William's Breton campaign of 1064 had prevented the establishment of a hostile state on the border of Normandy during the time of the English expedition, and the death of Conan in December 1066 had still further reduced the power of the Breton ruling house. He was succeeded by his son-in-law Hoel, count of Cornouailles. The new ruler inherited all Conan's difficulties in coping with the feudal magnates of Brittany. Among these a most important group was established in the northern and eastern part of Brittany facing Normandy. Prominent among them were the members of the cadet branch of the ruling house represented in Eudo of Penthièvre and his sons. Of these, most had already begun their careers in England under King William,⁷¹ but there remained the eldest, Geoffrey Boterel I, whose lordship stretched all along the northern coast, and included wide lands in the dioceses of Dol, Saint-Malo and Saint-Brieuc.⁷² Again, there was Geoffrey ‘Granon’, a bastard son of Alan III, whose lands were likewise concentrated in the diocese of Dol.⁷³ Finally to the south of these honours was the great lordship of Gael, a compact barony immediately to the west and north-west of Rennes, which at this time included both Montfort and Montauban, and stretched westward as far as Tremorel and Penpont to comprise no less than forty parishes.⁷⁴ This lordship of Gael was in 1074 held by Ralph ‘de Gael’, who together with many other Breton lords had established himself in England, and who, about 1069, had become earl of Norfolk.

Ralph de Gael⁷⁵ was in fact to be the central figure in the crisis that now ensued. He had inherited both his Breton lands and his English earldom from his father, also named Ralph, who had served as staller at the court of Edward the Confessor, and who had later assisted the Conqueror in the settlement of England. Ralph de Gael, the son, had therefore both Breton and English connexions, and he was particularly strong in being a natural leader of all those Bretons of middle rank who had followed in the wake of the Conqueror to receive lands in England. It was in fact to them in the first instance that Ralph appealed when in 1075 he conceived the plan of rebelling against the Conqueror in England, and he managed to associate with his rising no less a personage than Roger ‘of Breteuil’, earl of Hereford, the second son of William fitz Osbern. The precise causes of the rebellion are obscure, but its occasion was the marriage of Ralph with Roger's daughter, and the plot was hatched at the wedding feast held at Exning near Newmarket.⁷⁶Further importance was moreover given to the revolt by the fact that Waltheof, son of Earl Siward of Northumbria, and now earl of Huntingdon, and high in the Conqueror's favour, allowed himself to be associated with it.⁷⁷ Thus Breton and English opposition to the Conqueror was combined and, as if to make the movement more logically complete, Ralph appealed to Denmark for help.⁷⁸ Meanwhile, Ralph's fellow magnates in Brittany were ready to revolt against Hoel, or to raid into Normandy, whilst Count Robert of Flanders, Fulk le Rechin of Anjou, and King Philip were alertly watching the developing situation.

When the rebellion began in England, William was still in Normandy and it is significant that Archbishop Lanfranc, in whose hands he had left the administration of England, wrote at once to the king in order to persuade him not to return.⁷⁹ The king's place at this juncture was in Normandy, and it would be a disgrace if his loyal vassals could not by themselves deal with the revolt in England. In the event, they were able to do so without great difficulty. Wulfstan, bishop of Worcester, and Æthelwig, abbot of Evesham, both native prelates, combined with the Norman lords already established in the western Midlands to prevent Earl Roger from advancing out of Herefordshire to join his fellow rebel, whilst Odo, bishop of Bayeux, and Geoffrey of Coutances, Richard, son of Count Gilbert, and William of Warenne in like manner barred the westward progress of Earl Ralph from Norfolk.⁸⁰ Ralph thereupon retreated to Norwich, and then, leaving the castle at that place to be defended by his wife, he departed overseas. He may have gone to Denmark to incite his Scandinavian allies to action, and he eventually reached Brittany.⁸¹ Norwich was forthwith besieged and after some resistance surrendered upon terms which allowed the countess and many members of the garrison to depart for Brittany. Meanwhile the expedition had sailed from Denmark. Sweyn Estrithson had recently died, but his son Cnut, accompanied by many Danish magnates, led a great fleet of over two hundred warships to England. It arrived too late. Norwich had already fallen to the king's men, and the Danes contented themselves with sailing northward from Norfolk, pillaging the coast-lands, and York. After this, they departed home with their booty by way of Flanders so that Lanfranc could report to his king in Normandy that the south was now in greater tranquillity than at any time since the king's last departure.⁸² Thus at Christmas 1075 William could return to a pacified country in order to deal out punishment to the rebels.⁸³ The Bretons were savagely dealt with, Earl Roger was thrown into captivity, and Earl Waltheof was immediately cast into prison, there to languish for several months until on 31 May 1076 he was beheaded on Saint Giles Hill outside Winchester.⁸⁴ By that time, however, William was already back again across the Channel ready for the war in Brittany.

The revolt of the earls is of great interest as illustrating the problems arising from the establishment of Norman feudalism in England, whilst the participation of Earl Waltheof – and his execution – has a considerable bearing upon the policy adopted by the Conqueror towards his English subjects. But in respect of the defence of the Anglo-Norman kingdom, the chief importance of the rebellion of 1075 lies in its connexion with continental politics. The appeal to Scandinavia immediately gave to the English rising a wider significance, and it was in fact this aspect of the situation which most concerned King William, who, while still in Normandy, wrote urgently to Lanfranc, ordering that the east coast should be put in a state of defence.⁸⁵ Nor is it only with Scandinavia that this rising must be related. This was a Breton as well as an English revolt; it was the Bretons in England who were appropriately singled out for William's vengeance; and it was in this connexion also that the movement most concerned William's enemies in France. Earl Ralph at once continued the war from his Breton possessions, and if his rebellion should be successful it was clear that William would find a hostile power established on his western frontier. Such a situation,, providing constant opportunities for further attack, would be to the advantage of King Philip and his associates. William, on his part, was equally concerned to prevent it.

The war that followed in 1076 had thus wide general implications, and it is not surprising that it attracted the attention not only of Norman and Breton writers, but also of English and Angevin annalists. Their testimony is not wholly consistent and is difficult to interpret, but the general course of events can be discerned with a reasonable degree of probability.⁸⁶ At the time of Earl Ralph's return to Brittany, Count Hoel was already engaged in war with Geoffrey Granon. Earl Ralph, it would seem, joined with Geoffrey Granon, and together the two magnates established themselves in the castle of Dol.⁸⁷ It was a situation that King William could not ignore. Dol was near the Norman frontier; its chief defender was the rebel earl of Norfolk, and the whole operation could be regarded as directed as much against William as against Hoel. For this reason, too, the defenders of Dol were early reinforced by a contingent of troops from Anjou.⁸⁸ William, therefore, in September advanced against Dol, and, according to Breton writers, he acted in the campaign which followed in close association with Hoel. But despite all the resources of siege-craft Dol held out.⁸⁹ And it was then that the French king seized his opportunity. Philip was at Poitou early in October urgently seeking the assistance of Geoffrey, count of Aquitaine,⁹⁰ and towards the end of the month he moved at last at the head of a large force to the relief of Dol. His intervention was well timed, and was completely successful. Dol was relieved, and William, having suffered heavy losses in men and material, was forced to retire.⁹¹

William's defeat at Dol was the first serious military check that he had suffered in France for more than twenty years, and its importance has been unduly minimized.⁹² Indeed, his failure in Brittany in 1076 went some way to counterbalance the successful suppression of the rebellion of the earls in England during the previous year. Ralph remained a great lord in Brittany, strongly entrenched in power, and it is significant that during 1077 King Philip was able to consolidate his position in the Vexin without serious opposition from William. William's losses at Dol had been severe,⁹³ the damage to his prestige considerable, and his opponents were given an opportunity to follow up their success. Thus it was that, probably in the late autumn of 1076, or during the early months of 1077, Fulk le Rechin, with the assistance (as it seems) of Breton as well as Angevin troops took the offensive and attacked John of Le Flèche, one of William's strongest supporters in Maine.⁹⁴ John, however, managed to hold out in his castle until William came to his assistance, and Fulk, who appears to have been wounded in the course of the siege, was compelled to withdraw.⁹⁵ An uneasy truce was thereupon made. It was marked by a pact between William and Philip (which was certainly ratified in 1077),⁹⁶ and secondly by a pact between William and Fulk which may have been made at the same time and which at all events can hardly be placed elsewhere than in 1077 or 1078.⁹⁷

The real victor in the French campaigns of these years was in effect the French king. His policy had attained its first objective. The reverse suffered by William at Dol had been due to Philip's diplomacy and intervention, and William had now been brought to negotiate at a disadvantage. Philip was quick to turn the occasion to his profit, and his opportunity came in the Vexin. There Ralph of Crépi had been succeeded by his son Simon, who had albeit with difficulty maintained his position against the French king. But now he was seized by one of those violent impulses which were so characteristic of the eleventh century. Having obtained in marriage Judith, daughter of Robert II, count of Auvergne, he chose the occasion of his wedding night to vow himself and his bride to continence, and forthwith renouncing the world, he entered the monastery of Saint-Claude in the Jura. The event, which not unnaturally inspired widespread comment, gave King Philip the opportunity he had so long desired. He immediately occupied the Vexin, and thus extended his demesne up to the Norman frontier on the Epte.⁹⁸ The act constituted a new menace to Normandy, but Duke William in his present circumstances could do nothing to prevent it. He was perforce to acquiesce in the change, though in the sequel he was to meet his death in trying to reverse it.

Duke William's position in France had in fact been impaired. It is true that he was not himself forced to cede territory, and although the Angevin overlordship over Maine was once more recognized, the Norman administration there was continued under Robert, the duke's son. But the settlement of 1077–1078 was clearly based upon a compromise which could not be expected to endure, and it marked a reverse in William's fortunes. For the first time since 1054 a limit had been placed on the growth of Norman power in France, and it is impossible to escape the impression that from now on until the death of William the Conqueror the initiative of the struggle in France lay with the French king.

King Philip could certainly at once give a new turn to his offensive against William, and he did so by fostering divisions within Normandy, and by exploiting the difficulties latent in the relations between William and his eldest son. Robert had for long been used as a factor in the Conqueror's policy. In 1063 he had been given the title of count of Maine, and this had been confirmed in the arrangements of 1077–1078. Moreover, on more than one occasion he had been formally recognized as his father's heir, and when William returned to England in December 1067 he became permanently associated with the government of the duchy.⁹⁹ Thus it was that after Matilda's own departure for her crowning in England in 1068, Robert was made, as it appears, primarily responsible for the Norman administration.¹⁰⁰ His influence steadily grew, and during the ensuing decade he may even on occasion have been recognized as duke of Normandy under his father the king, for in two charters issued in 1096 he seems to reckon his tenure of the duchy from 1077 or 1078.¹⁰¹

Such arrangements might well entail a threat to the unity of the Anglo-Norman kingdom unless they were handled with discretion. Much here depended upon Robert himself, and unfortunately the young man's character was ill-suited to the delicate situation in which he was placed. Ordericus Vitalis was in this matter perhaps a harsh critic, but his vivid description of the young count carries conviction.¹⁰² Robert (he says) was personally brave, and very adventurous, a witty talker, and an attractive companion. But his acts were frequently ill-considered. He was extravagant in word and deed, so that he wasted his substance, and was lavish with promises upon which little reliance could be placed. ‘Wishing to please everybody, he was too ready to accede light-heartedly to any request.’ The portrait is of a young man possessed of the cruder feudal virtues, and it is not surprising that he was very popular among those of like mind to himself. Nor is it without interest that he speedily came to be a firm friend of Edgar Atheling whom Orderic describes in somewhat similar terms. But, equally certainly, Robert was devoid both of statesmanship and sagacity. Impetuous and vain, he was suitably fashioned to be a tool in the hands of men less frank, and more astute, than himself.

In 1077 Robert was not more than twenty-five years of age, and so far he had shown himself a loyal son to his father. Late in that year, however, or possibly in the spring of 1078, the fatal weakness of his character was displayed with lamentable results. Yielding to the flattery of his companions, he now demanded from his father that he should henceforth have independent control of Normandy and Maine.¹⁰³ Such a schism within the Anglo-Norman realm would at this juncture have been fraught with peril. William, however, seems to have refrained from any premature action until he was forced to quell a brawl which broke out at Laigle between the followers of Robert and those of his other sons, William and Henry. This brought matters to an open rupture. Robert precipitately withdrew from his father's court, and, accompanied by a large following, he attempted with incredible folly to gain possession of Rouen itself. Roger d'lvry, the king's butler, who was then in charge of the castle at the capital, was able to withstand the assault, but William realized that the situation demanded prompt action. He ordered the immediate arrest of the insurgents, and he threatened to confiscate their lands. Robert with most of his associates thereupon fled from Normandy.¹⁰⁴

The full consequences of Robert's action were, however, still to be disclosed. Robert's position within Normandy and Maine had already received recognition, and his personal popularity among the young nobles was so great that he was able to attract to his support many of the cadet members of the greatest houses in Normandy.¹⁰⁵ Chief among these was Robert of Bellême, son of Roger of Montgomery, earl of Shrewsbury, and with him was associated his brother-in-law, Hugh of Châteauneuf-en-Thimerais, whose castles of Châteauneuf, Sorel, and Rémalard thus became ready to provide the fugitives with bases outside the Norman frontiers from which they could conduct operations against the duchy. At this time, too, or perhaps a little later, the rebels were joined by William of Breteuil, the eldest son of William fitz Osbern, whose brother Roger had forfeited his English earldom after the rebellion of 1075, and who still languished in prison. Yves and Aubrey, sons of Hugh of Grandmesnil, likewise formed part of this company, as did also at some time Roger the son of Richard fitz Gilbert, lord of Tonbridge and Clare. All these were young men, and their action was distasteful to the heads of the great families to which they belonged – men who had achieved the conquest of England and who were even now organizing the settlement of the conquered kingdom. None the less, a threat had developed which menaced the Norman unity which since 1060 had been the chief source of Norman strength.

Such a situation could not fail to attract the attention of William's enemies in Europe, and it is not surprising that Robert's action led to an immediate revival of the coalition which had previously menaced the Anglo-Norman kingdom. Thus, about this time Robert visited the count of Flanders, and perhaps went on to solicit the aid of the archbishop of Trèves.¹⁰⁶ But the most interested party was the king of France, who saw here an admirable opportunity of developing a policy against William which had already achieved much success. Robert evidently sought his aid, and not in vain, and an unnamed ‘steward’ of the king of France is known at this time to have kept in contact with the rebels. As a result, Robert was joined by contingents from France and Brittany, from Maine and Anjou.¹⁰⁷ The earlier co-operation of William's enemies had in fact been reproduced, but this time its centre was the Conqueror's heir.

William could not afford to delay. At the time of the outbreak, he was apparently engaged against Rotrou I of Mortagne, but he immediately broke off these hostilities and attacked the rebels who were gathered at Rémalard.¹⁰⁸ In this engagement at least one of the chief supporters of Robert was killed,¹⁰⁹ and the rebels forthwith moved from the south-western frontier of Normandy round to its eastern border, establishing themselves at Gerberoi, a castle near Beauvais which had been placed at their disposal by the French king. There Robert was joined not only by a flux of new adherents from Normandy but also by many knights from France.¹¹⁰ It was a new crisis, and William was constrained to advance at once against the stronghold. The siege of Gerberoi which began shortly after Christmas 1078¹¹¹ lasted some three weeks, and it was only terminated when the rebels sallied out to risk a pitched battle. They were unexpectedly successful. William himself was unhorsed, possibly by his son, and wounded in the arm. His life was indeed only saved by one of his English followers, Toki son of Wigot, who was himself killed.¹¹² Thereafter the king's forces were put to flight, and Robert was left master of the field. It was a reverse for the Conqueror comparable to that which he had sustained in 1076 at Dol, and it was even more damaging to his prestige. William of Malmesbury indeed speaks of it as the greatest humiliation suffered by the Conqueror in his whole career.¹¹³

After his defeat at Gerberoi, William returned to Rouen, and there he was forced to enter into negotiations with his opponents. An influential group of senior members of the Norman aristocracy including Roger of Montgomery (now earl of Shrewsbury), Hugh of Grandmesnil, and the veteran Roger of Beaumont¹¹⁴ at once strove to effect a pacification in the interests of Robert and his young associates, many of whom were the sons or younger brothers of the negotiating magnates. King Philip, who had made their cause to some extent his own, supported them,¹¹⁵ and William was at last forced to treat. Perhaps before the end of 1079, and certainly before 12 April 1080, Robert was reconciled to his father, and William solemnly renewed to him the grant of succession to the Norman duchy.¹¹⁶ His control over his son had in fact been substantially weakened, and King Philip in particular had good reason to be satisfied with the results of Robert's first rebellion. The separation of Normandy from England – always the fundamental objective of the French king's policy – had been brought a stage nearer. And it is not wholly surprising that it was for a short time to be actually achieved immediately after the Conqueror's death.

William's defence of the dominion he ruled had always to take account of the menace which threatened all its frontiers, and a reverse in one region was almost invariably followed by an attack from elsewhere. Consequently as soon as the news of the engagement at Gerberoi penetrated to the north, Malcolm, king of Scotland, took immediate advantage of William's defeat in France.¹¹⁷ Between 15 August and 8 September 1079 he ravaged the whole area from the Tweed to the Tees. It was a severe raid which brought the Scottish king much plunder, and the fact that it had for the time being to remain unpunished, stimulated all the latent opposition to Norman rule in Northumbria. All through the ensuing winter it would seem the unrest grew, and in the spring of 1080 a revolt broke out which threatened the whole Norman settlement in the north, and culminated in one of the most horrible crimes of violence that disgraced the age.

In 1071 the bishopric of Durham had been given to Walcher, a clerk from Lorraine. He was in some respects a notable man, and he played a significant part in the monastic revival which took place at this time in his diocese. Moreover, in all his policy he seems to have been disposed to collaborate to some extent with the native magnates, and he is known, for example, to have been on good terms with Earl Waltheof. Indeed, Walcher's prestige had become so high by 1075 that when Earl Waltheof rebelled, and forfeited his possessions for treason, the king was induced to confide to the bishop the Northumbrian earldom itself. Walcher's qualities, however, were not suited to maintain order in a savagely turbulent province. From the first he seems to have compromised, leaving much of the administration to Gilbert, one of his kinsmen, and at the same time seeking to placate local feeling by favouring a certain Ligulf who was a cadet member of the comital house of Siward. Such a policy could hardly hope to succeed, and the inevitable clash occurred. In the spring of 1080 Gilbert overthrew and killed Ligulf with the aid of the bishop's household knights, and with the approval of Leobwin, the bishop's chaplain. Walcher himself, who had probably been guilty of nothing worse than negligence, was thereupon accused of connivance in the crime, and with more candour than wisdom he offered to prove his innocence by oath before an assembly which he summoned to Gateshead. The folly of this proceeding was at once demonstrated when the supporters of Ligulf arrived armed at the meeting and drove Bishop Walcher with Gilbert, Leobwin, and some of his knights into the adjoining church. This was promptly fired, and as the bishop with his followers emerged from the burning building they were severally butchered. The insurgents then proceeded to Durham. They failed to take the castle, but the resulting chaos invited a fresh attack from the hostile king of Scotland.¹¹⁸

The massacre of Bishop Walcher and his retinue took place on 14 May 1080. William was then still in Normandy, and indeed he was not able to return to England until late in July.¹¹⁹ In the meantime, however, Odo, bishop of Bayeux, was sent to the north on a punitive expedition, and in the autumn Robert, now reconciled to his father, set out for Scotland with a large force.¹²⁰ He pushed north as far as Falkirk, ravaging Lothian on his way, and forced Malcolm to a pact which recalled most of the provisions of the former agreement at Abernethy.¹²¹ Then turning southward he set up a fortress at the place afterwards to be known as Newcastle-on-Tyne.¹²² It was a notable expedition, but it achieved no final solution to the problem of the frontier. The selection of Newcastle for Robert's fortress indicated that the country north of the Tyne had still to be regarded as a debatable land, whilst to the west the dominion of the Scottish king extended as far south as Stainmoor.¹²³ The future filiations of Lothian and Cumbria had been foreshadowed, but they were still undetermined at the Conqueror's death, and the northern boundary of his kingdom never ceased to cause him anxiety.

Indeed, the threat which William constantly experienced from the north can be contrasted with his comparative immunity from attack from Wales. During the Confessor's reign Wales had been dominated by Griffith ap Llewellyn, prince of North Wales, who had been able to take advantage of the rivalries of the English earls.¹²⁴ In August 1063, however, he had perished when campaigning against Harold Godwineson, and for the next twenty years there was no one in Wales to take his place. As a consequence ‘there was no more heed paid to the Welsh’,¹²⁵ and the Welsh border in the time of King William constituted for him not so much a problem of defence as an opportunity for Norman expansion. The establishment of the marcher earldoms of Chester, Shrewsbury, and Hereford, which was to be a feature of the Norman settlement of England,¹²⁶ marked the beginning of a persistent Norman penetration into Wales. William fitz Osbern, earl of Hereford, for instance, planted colonies at Chepstow and Monmouth before his death; Hugh of Avranches, earl of Chester, extended his power as far as the river Clwyd, and provided means for further encroachments by his cousin Robert of Rhuddlan; whilst Roger II of Montgomery, earl of Shrewsbury, had before 1086 acquired much of the county which was afterwards to bear his name. The effects upon Wales of the Norman conquest of England were not, however, to be fully developed until after the Conqueror's death.¹²⁷ And during his lifetime Wales added little to his difficulties in defending the Anglo-Norman kingdom.

It was otherwise with Scotland. A threat from the north persisted throughout the Conqueror's reign, and, as has been seen, it could never be dissociated from the attacks made against his dominions in France. It was fortunate, for instance, that Robert's expedition of 1080 permitted the Conqueror to turn his attention across the Channel, for early in the next year Fulk, count of Anjou, doubtless taking advantage of William's absence in England, began a fresh attack on Normandy through Maine.¹²⁸ Supported by Count Hoel of Brittany, he advanced again against La Flèche, and this time he took the stronghold and burnt it.¹²⁹ Once again, therefore, William was compelled to cross the Channel in haste, and shortly after the fall of La Flèche he was advancing across Maine with a large force composed both of Norman and English troops. According to one account, a great battle impended, but this was averted by certain ecclesiastics who were in the neighbourhood.¹³⁰ A new pact was then made between the king and the count. It reproduced the compromise of 1077, and it appears to have been ratified at a place vaguely described by a later chronicler as Blancalanda or Brueria.¹³¹ Even so, Maine continued to be a source of danger. It was only after controversy and with papal support that in 1081 Hoel, a bishop with Norman sympathies, was appointed to succeed Arnold at Le Mans, and when some years later Hubert, vicomte of Beaumont, revolted against the Norman administration it proved impossible even after a lengthy siege to dislodge him from his castle of Sainte-Suzanne except on his own terms.¹³²

At this time, moreover, a new threat to the Conqueror's position came from his own family. It was in 1082 – the year after Fulk's second attack on La Flèche – that there occurred the famous quarrel between William and his half-brother, Odo bishop of Bayeux and earl of Kent. Odo, as has been seen, had been active in support of the king during 1080, and he seems to have remained in England when William was in Normandy in 1081. In 1082, however, he was arrested by the king.¹³³ The circumstances surrounding this important event are obscure, but according to a later chronicle it would appear that Odo had aspired to the papacy, and that he had also sought to entice some of the more prominent vassals of the king to follow him on fresh adventures to Italy.¹³⁴ Certainly, no medieval monarch could afford to countenance this latter project, and it is stated that William forthwith crossed to England, and despite the reluctance of many of his counsellors seized the person of the earl of Kent, and brought him back captive to Normandy. Odo seems indeed to have been kept in prison from this time until the Conqueror's death in 1087, but his possessions were not forfeited, since in Domesday Book he appears, as, after the king, the largest single landowner in England.¹³⁵

The defection of Odo involved a serious menace to the Norman dynasty, and before the end of 1083 this danger was further increased. Shortly after 18 July in that year Robert, whose personal status in Normandy had been much advanced as a result of his first revolt, decided once more to rebel, and departed from the duchy. His movements during the next four years are uncertain, but he remained the ideal agent for the French king, who gave him full support, and he became the centre of all the opposition to the Conqueror in France.¹³⁶ As 1084 advanced therefore King William and his realm might be seen to be approaching a new period of peril, and with the two most powerful members of his family in open opposition, the king's sense of isolation must have been acute. Before the end of the year it was made yet more distressful by the death of his wife to whom, according to the widespread opinion of contemporaries, he had been exceptionally attached. Queen Matilda died on 2 November 1083 and was buried in her own nunnery at Caen, where one of the finest epitaphs of the eleventh century still testifies to her memory.¹³⁷

Thus at the beginning of 1085 the security of the Anglo-Norman realm was still not assured, and in fact the final crisis of the Conqueror's reign was now at hand.¹³⁸ None the less William's achievement in ensuring its survival between 1067 and 1084 (which it has been the purpose of this chapter to watch) is not to be minimized. It was only accomplished despite reverses by means of an unremitting effort which, through two decades, had perforce to be expended over wide areas and against many enemies. Indeed, on looking back on that defence, its chief feature may well seem to have been the close interrelation which throughout existed between all its parts. The defence of the Anglo-Norman kingdom throughout the English reign of the Conqueror must in fact always be viewed as a unity, and the related campaigns which were conducted by William and his lieutenants had a common purpose, whether they were waged in Northumbria or Maine, or were directed against Sweyn Estrithson or Fulk le Rechin, King Malcolm of Scotland, or King Philip of France. Finally the astonishing energy on the part of King William that was involved in this defensive endeavour also deserves emphasis, for only thus can the total achievement of the Conqueror be appraised. The vitally constructive work carried out by King William between 1067 and 1087 (which has now to be considered) was accomplished in the midst of incessant warfare, and in a realm whose ruler was never secure from attack.

1 AS. Chron., ‘D’, s.a. 1067. On Edric the Wild – ‘Silvaticus’, ‘Guilda’ – there has been much learning. The curious may be referred to Freeman (Norman Conquest, vol. IV, note I), and to Ritchie (Normans in Scotland, chap. I). See also Douglas, Feudal Documents, pp. xci–xciii. Ordericus (vol. II, p. 166) says he made his submission at Barking. Legends, moreover, gathered rapidly about him. Thus after dinner one night he came on the fairies dancing, and fell in love with one of them whom he married (Walter Map, De Nugis, vol. II, p. 12). The king heard of this and ordered her to be brought to court. A conversation between William the Conqueror and the Queen of the Fairies would have been worth hearing.

2 Will. Poit., p. 264; Will. Jum., p. 138.

3 Prou, Rec. Actes – Philippe I, p. xxxii. If this suggestion were correct, the raid of Eustace took place in the autumn of 1067.

4 Will. Poit., p. 264.

5 AS. Chron., ‘D’, ‘E’, s.a. 1067.

6 Ibid.; Ord. Vit., vol. II, p. 180; Round, Feudal England, p. 433. Gytha, the widow of Godwine, had taken refuge at Exeter, and on its surrender she went to the island of Flatholme in the Bristol Channel.

7 Ord. Vit., vol. II, p. 180; Complete Peerage, vol. III, p. 428.

8 Freeman, op. cit., vol. IV, p. 175.

9 AS. Chron., loc. cit.

10 Ibid.; Ord. Vit., vol. II, p. 181; Regesta, vol. I, no. 23.

11 Ord. Vit., vol. II, p. 185; Stenton, William the Conqueror, p. 265.

12 Simeon of Durham, Hist. Regum (Opera, vol. II, p. 187).

13 Regesta, vol. I, no. 23.

14 Regesta, vol. I, nos. 7, 9.

15 See generally, E. S. Armitage, Early Norman Castles in the British Isles (1912).

16 Ord. Vit., vol. II, p. 184.

17 E.H.D., vol. II, p. 250.

18 Will. Poit., p. 169; Ord. Vit., vol. II, p. 165; Armitage, op. cit., p. 229. The stone Tower of London was begun about a decade later under the supervision of Gundulf, bishop of Rochester (Hearne, Textus Roffensis, p. 212).

19 Ord. Vit., vol. II, pp. 181–185; Armitage, op. cit., pp. 151, 242.

20 Will. Poit., p. 267; Will. Jum., p. 138.

21 Ord. Vit., vol. II, p. 186; vol. III, p. 111; Mason, Eng. Hist. Rev., vol. LXVI (1956), p. 61.

22 Ord. Vit., vol. II, pp. 181, 184. Henry must have been very young at the time, for his elder brother Robert was a tiro at Hastings.

23 Ord. Vit., vol. II, 188. William Malet is a difficult personality. He may have been in England before the Conquest, and one report says that the Conqueror entrusted to him the burial of Harold after Hastings (Freeman, op. cit., vol. III, p. 514).

24 He is back at Winchester on 13 April 1069 (Regesta, vol. I, no. 26).

25 e.g. Lincoln, Colchester, Dover, Richmond.

26 It is usually stated that the Conqueror was continuously in England during 1068. This is doubtless correct, but the matter is by no means clear. Two Norman charters which passed in his presence are dated 1068. The one is for Troarn (Sauvage, L'abbaye de Troarn, Preuves, no. II; cf. Regesta, vol. II, p. 391); the other is for Saint-Pierre de la Couture (Cart. S. Pierre, no. XV). Neither of these texts is wholly satisfactory in its present form, but the main reason why suspicion has fallen upon them is that they conflict with the assumed chronology of William's movements at this time. Recently, however, L. Musset has printed the text of two charters for the abbey of Saint-Gabriel in Calvados. These are stated to have passed in William's presence at Valognes, and they are dated 1069 (Actes Inédits du Xī^e Siècle, pp. 21–23 – Bull. Soc. Antiq. Norm., 1954). If these, in fact, were given at Christmas 1068 they might have been thus dated by a clerk who began the year at Christmas. On this hypothesis, they could be made to supply some confirmation of the other two Norman charters. Too much reliance should certainly not be placed on testimony of this character. But the possibility that William may have visited Normandy during the winter of 1068–1069 is not to be wholly disregarded.

27 Ord. Vit. (vol. II, p. 189), without giving a precise date, places Matilda's visit to Normandy after William's second campaign in the north; that is to say, between the Easter court at Winchester (4–11 April 1069) and the beginning of the ‘Rising of the North’ in the early autumn of that year. And the Valognes charters noted above could with equal, or greater, propriety be cited in support of this date.

28 Ordericus (loc. cit.) lays special emphasis on this point, and it is significant that the charters given at Valognes were there ratified not only by the king but also by the queen.

29 The only possible times for such a visit (or visits) to Normandy by King William are: (1) very late in 1068, i.e. between the end of the first northern campaign and the opening the second northern campaign in (?) February 1069; or (2) between April 1069 (the Winchester court) and September 1069, by which time William was back in England preparing to withstand the invasion of Sweyn.

30 AS. Chron., ‘D’, ‘E’, s.a. 1069, 1070; Ord. Vit., vol. II, p. 191; Simeon of Durham, Hist. Regum (Opera, vol. II, pp. 187, 188).

31 Ord. Vit., vol. II, p. 194.

32 AS. Chron., ‘D’, s.a. 1067 (an interpolation). The marriage probably took place late in 1069. Cf. Ritchie, Normans in Scotland, pp. 25, 26.

33 AS. Chron., ‘D’, s.a. 1069; Ord. Vit., vol. II, pp. 192–195.

34 Ord. Vit., vol. II, p. 197; Simeon of Durham (op. cit., vol. II, p. 198).

35 Simeon of Durham shows, however, that the crimes committed by Malcolm in the north had been equally horrible (op. cit., vol. II, p. 191).

36 Ord. Vit., vol. II, p. 196.

37 Simeon of Durham (op. cit., vol. II, p. 188); Chronicon de Evesham (ed. W. D. Macray), pp. 90, 90; Will. Malms., Gesta Regum, p. 309; G. Creighton, History of Epidemics, vol. I, pp. 27, 29.

38 AS. Chron., ‘E’, s.a. 1070.

39 All that is known, or could possibly be surmised, about Hereward is exhaustively discussed by Freeman (op. cit., vol. IV, pp. 454–487). On his alleged descendants, see Round, Peerage and Pedigree, vol. II, pp. 259–286.

40 The confused chronology of these events is discussed by E. O. Blake (Liber Eliensis, pp. lv, lvi).

41 AS. Chron., loc. cit.

42 Ibid., ‘E’, s.a. 1071.

43 Latouche, Comté du Maine, chap. V. The chief authority is the Actus pontificum Cenomannis in urbe degentium (ed. Busson and Ledru (1902)).

44 Cart. S. Pierre de la Couture, no. XV.

45 Actus, pp. 376–377; Latouche, op. cit., p. 37; Douglas, Domesday Monachorum, pp. 35, 36.

46 Whether this could with any propriety be called a ‘commune’ in the later sense of a collective feudal lordship is discussed by Latouche in Mélanges – Halphen, pp. 377–383.

47 Actus, pp. 278, 379.

48 AS. Chron., ‘E’, s.a. 1070; Ord. Vit., vol. II, p. 235; Fliche, Philippe I, pp. 252–261. For the date of the battle of Cassel, see J. Tait, in Essays – Lane Poole, pp. 151–167.

49 Below, pp. 295, 296.

50 See G. W. S. Barrow, The Border (Durham, 1962).

51 Simeon of Durham (op. cit., vol. II, p. 190).

52 Ord. Vit., vol. II, p. 237.

53 Ord. Vit., vol. II, p. 238.

54 Regesta, vol. I, no. 63.

55 Below, pp. 325, 326.

56 AS. Chron., ‘D’, ‘E’, s.a. 1072.

57 AS. Chron., ‘E’, s.a. 1072; Skene, Celtic Scotland, vol. I, p. 424.

58 Simeon of Durham, Hist. Dunelm. Eccl. (Opera, vol. I, p. 106).

59 Actus, p. 379; Latouche, op, cit., p. 38; K. Norgate, Angevin Kings, vol. I, pp. 219–220.

60 AS. Chron., ‘D’, s.a. 1074 (equals 1073).

61 Actus, pp. 380, 381.

62 Ord. Vit., vol. II, p. 255.

63 On 30 March 1073 William was at Bonneville-sur-Touques in the company of Arnold, bishop of Le Mans, and there he confirmed the rights in Maine of the abbey of Solesmes (Cart. S. Pierre de la Couture, no. IX). It would seem unlikely that he would have issued this charter before he had himself recovered his jurisdiction in Maine, and if this be so, the campaign must have taken place very early in 1073. The matter cannot, however, be taken as certain. A charter for Saint-Vincent du Mans (Cart. S. Vincent du Mans, no. 177) probably passed on the same occasion.

64 Prou, Rec. Actes – Philippe I, p. xxxii.

65 AS. Chron., ‘D’, s.a. 1075 (equals 1074).

66 Fliche, Philippe I, p. 36.

67 Prou, Rec. Actes – Philippe I, no. XLI; Fliche, op. cit., pp. 138, 142, 143.

68 AS. Chron., loc. cit.

69 Ibid.

70 Ibid., ‘E’, s.a. 1074.

71 e.g. Brian, Alan the Red, and Alan the Black.

72 La Borderie, Histoire de Bretagne, vol. III, p. 11.

73 Chron. S. Brieuc (Rec. Hist. Franc., vol. XII, p. 566).

74 La Borderie, op. cit., vol. III, pp. 68, 69. See also the admirable map in the same writer's Neuf Barons de Bretagne (1895).

75 Complete Peerage, vol. XI pp. 573 et sqq.

76 AS. Chron., ‘E’, s.a. 1075. It might have been thought that the marriage was made against the king's prohibition. A letter of Lanfranc (Ep. 39) shows this was not the case. The speeches put into the conspirators by Ord. Vit. (vol. II, pp. 258, 259), though interesting, are imaginative.

77 Ord. Vit. (vol. II, p. 260) makes Waltheof agree only reluctantly.

78 AS. Chron., loc. cit.

79 Ep. 34.

80 AS. Chron., loc. cit. Flor. Worc. (vol. II, p. 11) says the engagement took place near Cambridge. Ord. Vit. (vol. II, p. 262) puts it at a place he calls ‘Fageduna’.

81 Ord. Vit., vol. II, p. 263.

82 Lanfranc, Ep. 40.

83 AS. Chron., loc. cit.

84 Ord. Vit., vol. II, pp. 265–267. ‘It can only be left an open question whether his execution can be justified in morality as well as in law.’ (Stenton, op. cit., p. 603.)

85 Lanfranc, Ep. 35.

86 Below, Appendix E.

87 Chron. S. Brieuc; Chron. Britannicum (Rec. Hist. Franc., vol. XI, p. 413; vol. XII, p. 566), Flor. Worc., s.a. 1076.

88 Ann. ‘de Renaud’ (Halphen, Annales, p. 88).

89 Ibid.

90 Prou, Rec. Actes – Philippe I, nos. LXXXIII, LXXXIV.

91 AS. Chron., ‘E’, s.a. 1076.

92 Cf. Stenton (William the Conqueror, p. 341): ‘a small continental war’.

93 AS. Chron., loc. cit.; Ord. Vit., vol. II, p. 291; Prou, op. cit., nos. LXXXIX, XC.

94 Ann. Saint Aubin and Saint Florent (Halphen, op. cit., pp. 5, 129).

95 Halphen, Comté d'Anjou, p. 311, no. 233.

96 AS. Chron., ‘E’, s.a. 1077.

97 There may be a reference to this in a charter for Saint-Vincent du Mans (Cart. S. Vincent du Mans, no. 99). And see below, Appendix E.

98 Fliche, Philippe I, pp. 147, 149.

99 David, Robert Curthose, pp. 17–41.

100 Will. Jum., p. 139; Ord. Vit., vol. II, p. 188.

101 Gall. Christ., vol. XI; Instrumenta, col. 76; Haskins, Norman Institutions, p. 67.

102 Ord. Vit., vol. III, p. 262.

103 The revolt must have started after 13 September 1077 when Robert was still at his father's court (Gall. Christ., vol. XI; Instrumenta, col. 72) and some time before the opening of the siege of Gerberoi about Christmas 1078.

104 Ord. Vit., vol. II, pp. 297, 298.

105 Ord. Vit. gives two lists of the supporters of Robert (vol. II, pp. 296—298; vol. II, pp. 380, 381). All of these persons were certainly associated with Robert at one time or another in his revolts against his father.

106 AS. Chron., ‘D’, s.a. 1079; Ord. Vit., vol. II, p. 381.

107 Ord. Vit., vol. II, p. 296.

108 Ibid. The cause of the quarrel is unknown.

109 It was clearly a local revolt. Villerai is within a few miles of Rémalard, and Rémalard is some ten miles from Mortagne.

110 Ord. Vit., vol. II, p. 387.

111 Below, Appendix E.

112 AS. Chron., ‘D’, s.a. 1079. Toki has been presumed to be the son of Wigot of Wallingford who is frequently mentioned in Domesday Book.

113 Will. Malms., Gesta Regum, p. 317.

114 Ord. Vit., vol. II, p. 388.

115 There is a great difficulty here, for a charter of unimpeachable authenticity (Prou, op. cit., no. XCI) shows Philip and William in association outside the walls of Gerberoi. The only hypothesis which I can tentatively advance is that William after his defeat was forced to retreat, and Philip, in return for some concession, mediated on his own terms between Robert and his father, and took steps to enforce some temporary settlement to his own advantage.

116 Ord. Vit., vol. III, p. 390; vol. II, p. 242. For date, see Cart. S. Trin. Roth., no. LXXXII.

117 AS. Chron., ‘E’, s.a. 1079, implicitly connects Malcolm's invasion with the engagement at Gerberoi.

118 Simeon of Durham, Hist. Dunelm. Eccl. (Opera, vol. I, pp. 105, 106, 114); Hist. Regum (Opera, vol. II, p. 200).

119 William and Robert were still in Normandy on 14 July (Regesta, vol. I, no. 135). They must have returned to England very shortly afterwards.

120 Simeon of Durham, loc. cit.

121 Ibid. (Hist. Regum – Opera, vol. II, p. 211); Hist. Mon. Abingdon, vol. II, pp. 9, 10.

122 Ibid.

123 Ritchie, Normans in Scotland, p. 50.

124 J. E. Lloyd, History of Wales, vol. II, pp. 358–371.

125 Gaimar, v. 5084.

126 Below, pp. 294, 295.

127 The whole process (which falls outside the scope of this book) is admirably surveyed by Sir Goronwy Edwards in ‘The Normans and the Welsh March’ (Brit. Acad., Proceedings, vol. XLII (1956), pp. 155–177).

128 Annals of Saint-Aubin, and ‘de Renaud’ (Halphen, Annales, pp. 5, 88).

129 Ord. Vit., vol. II, p. 256.

130 Below, Appendix E.

131 Ord. Vit., vol. II, p. 356. All attempts to identify the place have hitherto proved more interesting than convincing.

132 Annals of Vendôme (Halphen, op. cit., p. 65); Ord. Vit., vol. III, pp. 194–201; Latouche, Comté du Maine, p. 39; David, Robert Curthose, p. 35.

133 AS. Chron., ‘E’, s.a. 1082.

134 Ord. Vit., vol. III, pp. 189–192, 247.

135 Ibid. But the matter is difficult. If Regesta, vol. I, no. 147, was issued in England, Odo was in this country in 1082, but he appears also to have witnessed a charter in Normandy in that year (ibid., no. 150). The Durham charter (ibid., no. 148) cited by the Complete Peerage (vol. VII, p. 128) is, I think, a forgery.

136 David, op. cit., p. 36.

137 Ord. Vit., vol. III, p. 192.

138 Below, chap. 14.