Post-classical history



With the landing of Matilda at Arundel, the civil war may be thought to have begun, though as we have seen, the ‘rebellions’ of 1138 and 1139 seem to have been a concerted, if abortive, attempt to back Robert of Gloucester’s defiance. Indeed, at the time, Matilda’s arrival did not seem as noteworthy an event as it did afterwards. Stephen probably did not anticipate that it would have any great effect on the military situation, and he may have been right. It was the arrival of Robert of Gloucester which had the greater impact from this point of view. One doubts that Stephen would have arranged Robert’s journey with an escort to Bristol or anywhere else in England. Events which seem puzzling to us, often make better sense if we try to block out knowledge of what happened afterwards, and try to see with the eyes of someone there.

In order to understand the civil war, we need some knowledge of the state of warfare by 1139, the developments in technology, and thinking which affected it. We need to know something of the tactics of dismounted knights, of the kind of arms and armour now favoured, and of the current state of fortification. So, before we tackle the events of 1139–40, let us turn aside to consider these matters briefly.

The experience of both Stephen and Robert of Gloucester, was of warfare as members of the household of Henry I. Both had fought for him. Both were therefore fully aware of the kind of methods used by the King of France and his allies. Both had some experience of the kind of war waged by Geoffrey of Anjou, and perhaps felt some resentment against him for his conflict with Henry in 1134–5. Both were grandsons of William the Conqueror, and were aware of the significance of the fighting methods of the Normans. No one in this age was ignorant of warfare in the East and of the crusades. Indeed, Stephen’s father had fought in the First Crusade, where he is generally said to have disgraced himself by his flight from Antioch. I have argued elsewhere that this has been greatly exaggerated.1 Count Stephen-Henry had been a highly respected leader of the crusade, indeed elected as its chief by the leading crusaders themselves. The problem had arisen during his own illness, which only his enemies sought to question, and had happened because of a particularly good relationship which he had built up with the Byzantine emperor. Some letters of the count back to his wife, Adela, give his own account of affairs. He did not flee in a cowardly manner. He did decide to return home, falsely believing that all was lost. It was a realistic assessment of the situation as he then knew it, and largely dependent on the fact that the Byzantine emperor chose not to assist the crusaders at Antioch. Stephen-Henry had gone to seek aid for a desperate situation. When the emperor refused aid, having no army behind him, Stephen-Henry had no chance of breaking through to what seemed in any case like a lost stronghold. The comments by Anna Comnena on the incident need to be treated with caution, since she was chiefly concerned to protect her father’s reputation. Only afterwards did it look like cowardice because the crusaders won a great, if unlikely, victory. Nor should we be too ready to believe waspish comments by unfriendly chroniclers, in particular by Orderic Vitalis, who pretended to knowledge of conversation between Stephen-Henry and Adela in their bedchamber, ‘between conjugal caresses’. Was the chronicler hiding under the bed to obtain this information? What we truly know is that Stephen-Henry did return to the Holy Land, and then bravely fought in the continuing conflict, and lost his life in doing so. We must not forget that in the twelfth century most men were very provincial in their outlook; virtually all outsiders had to be ridiculed. Thus a Norman, perhaps especially if he himself were an outsider from England, like Orderic, was contemptuous of people like Geoffrey of Anjou, or Stephen-Henry of Blois. The English, even if they were of Norman origin, were equally contemptuous of the Scots, the Welsh and the Irish. The nearest neighbours were usually the most despised. Stephen-Henry was almost certainly much maligned over this episode, and undeservedly so, but his son Stephen knew of the sneers and the criticism, and it affected his view of warfare. He would not want to be seen as a coward, or give any excuse for such comment against himself.2

At the same time, like almost everyone else in the West, Stephen admired the achievements of the great crusaders. To what extent battle tactics and fortification methods of the East influenced development in the West remains controversial, but it is impossible not to believe that those crusaders who returned had not been affected by their experience, which included contact with Byzantine military methods, as well as those of the Saracens.

By the 1130s, the association between England and Normandy was a long-standing one. Only for relatively brief periods since 1066 had the two regions been separated. Both had been ruled together for some time, at least during the reigns of William I, William Rufus and Henry I. The troops which the kings used and the military system they depended upon, was largely Norman in origin, but had also been influenced by Anglo-Saxon traditions. The idea of agreed quotas for military service had been gradually introduced to England. The ruler, as king or duke, could call on military service in defence of the realm or duchy. In England the ruler had been able to hammer out individual deals with the magnates, to gain military service from the men of the magnates for national purposes.

Fundamentally, Stephen’s plan of military organization seems to have been to treat earls as military governors acting on his behalf; thus he selected men of military ability and aristocratic standing, rather than officials. A large number of new earls was appointed during the crucial four years from 1138 to 1142, the only period of his reign when Stephen made a large number of appointments. The strategic placing of many of these earldoms is significant, including border regions such as Yorkshire (1138), Pembroke (1138), Worcester (1138/9), Northumbria (1139), and Herefordshire (1140), together with vital areas bordering on Angevin territories, such as Wiltshire (1140) and Cornwall (1140). Not all of these men were able to establish themselves, but the intention is clear: to have loyal men of ability in vital military regions, apart from the dubiously loyal Henry of Scots in Northumbria. They were also to be men of sufficient social standing to command respect in their regions and act with a degree of independence. William of Ypres was also given an equivalent position to that of an earl in Kent, an essential area to protect, because of its cross-Channel significance. And even the interior earldoms must not be left out of the military equation; no area was safe to neglect in a situation of civil war.

Royal control was important, and this was especially noticeable in the system of fortifications that had developed since 1066. Stephen was prepared to delegate a degree of power to his earls, but he expected to have rights over the royal castles, usually the key castles in a time of war. By the time of the Conquest, the Normans possessed castles in the duchy, used both as military centres and as residences for families, dependents and garrisons. They had developed both stone towers and the earth and timber fortifications, which we call motte and bailey castles. Excellent illustrations of this type of fortification, developing throughout northern France at the time, are found on the Bayeux Tapestry, both in Normandy and in Brittany. Naturally, when the Normans invaded and conquered England, they developed such castles there too. In fact, because Edward the Confessor had lived for some time in Normandy, and had a number of Normans whom he favoured in England, a few of these castles had been built in England even before 1066. But after 1066 they mushroomed, until the whole country seemed to be smothered with mounds and moats and baileys. England was now governed from, and by, castles.

This is one reason why siege warfare was so far and away the most common kind of war in this period. Battles were very rare: only two major ones in a reign of constant border invasions and civil war, plus a few lesser skirmishes which we might just consider as battles, on the Welsh border, or at Wilton. There were no battles at all fought in Normandy, though from 1135 there was nearly constant war for a decade. Battles were very risky, their outcome very uncertain. No one risked battle unless fairly certain of victory, and to find two commanders in that frame of mind was not very common. Most good commanders avoided battle so far as possible. It has been said before, but it cannot be emphasized too greatly, since readers coming from other periods of history are often hard put to understand twelfth-century warfare unless they appreciate this basic fact. There was no disgrace in avoiding battle in most circumstances. Of course, war could have many aims: conquest, survival, expansion, defence, power, punishment, booty, order – and many others. The achievements and successes of war depended on the aims, as did the method of fighting. For example, in a war of conquest, one needed to establish firm authority over a region, and castles were extremely useful for that, defending against a hostile local population, and giving a secure base to conquering troops. In the Welsh border wars, where the Welsh aim was often to raid and collect booty, castles had a different use, more for containing mobile troops who could emerge to cope with raiders, often when on their way home.

With the dominance of siege war, the significance of the castle is only too obvious. The normal type of castle was the earthwork motte and bailey castle of the Norman period. This consisted typically of a large Christmas pudding shaped mound of earth, the motte, plus a larger and flatter mound which was the bailey. The motte was the chief defensive section of the castle, and could usually only be entered by a wooden bridge crossing a ditch of moat round the foot of the motte. The flattened top of the motte was protected by a palisade, and within that, built on the top of the motte, was a tower keep. Within the keep lived the lord, his family, some retainers, and an armed garrison to defend it. The flattened mound of the bailey was also usually surrounded by a ditch or a wet moat, and a palisade, and any entrance would be defended. It would contain various buildings of use in the castle, probably some dwellings for lesser folk, store places, stables and so on. If a motte and bailey castle were attacked, the bailey would be defended, but usually the last ditch defence would be from the keep on the motte.

Archaeology has shown that there was in fact a variety of motte and bailey designs. That described above is typical but by no means universal. There were earthwork castles without mottes, castles with two mottes, castles with two or more baileys and so on. It all reminds us that although there was a model within men’s minds, castles were generally built for individuals, and men have always fancied making improvements to their homes, or changes to suit themselves, or changes based on new fashions of which they had become aware.

Castles were certainly affected by fashions and by changes in ideas about defence. They were symbols of power and status as well as homes and fortifications. Therefore, lords liked their symbols to be up to date and impressive: they might improve the motte, add another bailey, or especially change the wooden palisades and keeps for stone ones. The Normans had been perfectly capable of building in stone, and there were stone towers in Normandy before the conquest, and early on in England. The two earliest in the kingdom, begun in the reign of William the Conqueror, were the White Tower in the Tower of London, and Colchester Castle. At first stone castles in England were scarce. Motte and baileys were cheaper to build, and quicker, and suited a situation of conquest and emergency. But as the disturbed condition eased, more castles were built or rebuilt in stone, both by the king and by the lords. By 1139, it was becoming rare to undertake any new building which was not mainly in stone, though the new disturbances of the civil war meant a temporary return to some motte and bailey construction.

Because castles had become stronger, with improved defences such as better gatehouses, towers built into the walls, battlement walks and so on, sieges tended to take longer. It was still possible, but less easy, to take a castle immediately with a storm attack. Therefore, castle besiegers had also to develop their techniques. Improvements were made in throwing engines for hurling stones and javelins. We do not know when the trebuchet was first introduced, but it may just have been at this time; certainly already men were interested in being able to throw heavier stones, further and with more force. The ballista and the mangonel were commonly used, as were great wooden towers on wheels for approaching walls. Mining was a common method of bringing down defences, and the throwing of fire often used to destroy gates or internal buildings. It was during Stephen’s reign that Greek fire was first used in a siege in western Europe.3 Improved archery was also important in sieges, both for attack and defence. And besiegers, having often to spend a lengthy period reducing a castle by blockade, made life more comfortable and more safe for themselves by constructing siege huts to live in, and counter castles to give protection. Speed, surprise, and originality of method by a commander, remained important, and we shall see a good deal of improvised and clever tactics in the sieges of the civil war.

Although battles were rare, they did occur, and the Battle of Lincoln in particular is a key moment in our story. It is no longer commonly believed by historians that medieval battle tactics were primitive or non-existent. But the Battle of Lincoln has not been analysed very closely with a view to studying the tactics.

As fortification provoked changes in siege war, so improvements in arms and armour led to changes in battle tactics. Armour was gradually becoming more elaborate. It was still, by 1139, not unlike the armour used in 1066, but there had been some changes. More men were usually now well armoured, and included more mercenary troops than had previously been the case, on horse as well as on foot. The ring-mail hauberk, tunic shaped, was the fundamental garment; men now tended to add additional pieces to protect legs or arms or chest. The helmet might still be in the conical shape of the familiar Norman style, but helmets covering more of the head and face were developing, and may have encouraged the need for more clothing to distinguish the knight from his fellows.

Our period seems to be the one when heraldry was beginning, though its rules were not yet developed, as later they would be. This development began in France but was soon taken up throughout Europe, with social as well as military connotations. The enamel in the museum at Le Mans of Matilda’s husband, Geoffrey of Anjou, is one of the most famous early examples of heraldic dress, and as I have argued elsewhere, may have derived from the coat of arms of the English kings as used by Henry I. Both his cap and his shield have a device with lioncels which are obviously similar to the lions of England.4

There were also developments in the major arms used. Swords were now made in larger numbers by experts, well balanced, sharp; some of the best swords ever made. Lances were used, and seem to have been growing in length, with great emphasis on balance, as the cavalry charge was perfected. Longbows were being used more than previously, and crossbows were becoming familiar weapons.

In a war against the infidel, there was every encouragement to kill the enemy. In a civil war, the fighting could often be bloody, but there was every reason to capture rather than kill enemy knights, for ransom, and as a sign of tit for tat mercy. There was also good reason to treat the native population with care, since one might hope to govern the same people afterwards. In any case war between Christians was viewed as a different matter to war between different faiths. The church condemned war between Christians except on very good grounds, and could normally only approve of one side’s actions, for example, if fighting in self defence or in defence of the church. The arguments for what was or was not a just war were more than simply theories; men cared about seeming to have justice on their side, and war propaganda formed an important part of the whole picture. Knights saw themselves as Christian warriors, and other knights, even those fighting on the opposite side, as members of a class, a brotherhood; therefore mercy was normally to be praised, and death in war was much less common in this period in the West than in most other circumstances.5


There can be little doubt that Robert and Matilda planned a war to win the succession, but Stephen and his friends may not have realized this at first. Robert’s defiance had been taken as a declaration of war, and Stephen had already started the process of taking over his forfeited lands. Stephen was also aware of the threat posed by Robert, and had set up a guard on the coast, ‘closing the harbours by a very close watch’.6 He did not mean to be caught unawares by the landing of an invasion force.

But Matilda’s plans were not yet clear. She held virtually no land either in England or in Normandy, since Henry I had neglected to give her this vital base for taking over power, assuming that she would take over the royal lands. She held only those few castles in Normandy mentioned above. Her appearance in England was neither illegal nor necessarily a threat. The fact that she arrived at Arundel made it look even more normal, in that she received an invitation to the place from her step-mother Adeliza, widow of Henry I and now married to the lord of Arundel, William of Albini. William of Albini was with Stephen and had shown no signs of disloyalty; indeed, he was to remain at least nominally a royalist. Then again Matilda was a woman, and although this had been a disadvantage with regard to the succession, it gave her certain privileges in polite society. Stephen had no grounds for arresting her or otherwise ill treating her. Again, when we look through the eyes of contemporaries, Stephen was presented with a much more difficult and delicate situation than normally seems to be recognized, and it is not at all clear that he acted badly, mistakenly or stupidly. In some ways, the less he provoked Matilda’s hostility, the less likely she was to cause trouble. If she could be persuaded to accept the new regime, he would have won a major battle.

Of course, if we look through Matilda’s eyes, we see a very different picture. Two facts allow us to believe this. Firstly, she came in the company of Robert of Gloucester, who had already begun a war. But secondly, and it would take time for this point to be clear to her enemies, she was behind the sending of Baldwin de Redvers to Dorset, and Baldwin came in very hostile fashion, seeking to establish a maritime base for the Angevin cause.

Baldwin de Redvers had good cause to oppose Stephen. His offer to come in peace to the king, albeit late in the day, had been rejected. He had twice rebelled in 1136 and twice been defeated, losing both Exeter and the Isle of Wight. Since then he had joined up with Geoffrey of Anjou and been encouraged to cause trouble for Stephen in Normandy. Now he came to reclaim his lands in England, and to help the cause of Matilda and Robert. He landed first at Wareham, which had belonged to Robert of Gloucester but had been captured by Stephen. Baldwin arrived at about the beginning of August ‘with a fine and strong body of troops’.7 He does not seem to have been able to use Wareham except for landing, and moved on to nearby Corfe Castle, where the ruins still dominate the surrounding countryside from their imposing hilltop site.8 Here he was admitted. Stephen, with his usual promptness, learned of the arrival and came to besiege Corfe. But no sooner was he there than news came of the imminent arrival of Robert of Gloucester and Matilda, and Stephen realized he must give priority to the Arundel landing.

Again, Stephen took precautions against a new invasion. He also had forces at sea, and ordered ‘a careful watch, night and day, on all the approaches to the harbours’.9 The same source says he was watching ‘over the pacification of the kingdom with the greatest soldierly skill’. But he did not prevent the landing. To be fair, Arundel is not the most obvious of landing points for an invasion, and the force that arrived was more of an escort than anything else. Presumably none of the major ports was open to Robert and Matilda, so they preferred to slip in to Arundel. William of Malmesbury says Robert had come ‘with a far smaller military force than that with which anyone else would have ventured on so hazardous a war’. They were accompanied by just 140 knights, one of whom was Guy of Sablé.10

They slipped into Arundel on 30 September 1139, opened to them by Adeliza of Louvain; ‘welcomed as if merely guests’.11 Arundel is hardly a port; it stands near the sea and is approached by the River Arun. Its position has never allowed large ships to come close, but the nature of its approach is excellent for a secretive entry by a smallish group. It is also a very strong castle on a good defensive site, not easy to approach by land. It is situated on the edge of a rise of land which stands above the very flat, watery meadows, crossed by streams, which run down to the sea. The motte, which existed in 1139, is still impressive, though practically all of the structures around have been rebuilt. In the old entrance building on the approach to the motte is a chamber over a gateway, and this retains the name of ‘Matilda’s chamber’, though whether she ever used it is impossible to say. The Worcester chronicler suggests that Stephen was angry that his watch had failed and Robert and Matilda had managed to land at all.12

The arrival of Robert, as the Gesta Stephani says, was the signal for his supporters; ‘all who secretly or openly favoured the earl were keener than usual and more eager to trouble the king’.13 But Stephen’s readiness was soon apparent, in that he used the troops with him and advanced straight on Arundel, while Matilda was still within. Robert of Gloucester, however, recognizing that he could not delay, set off with a small escort of about a dozen troops, to reach safer territory. Stephen’s subsequent actions are often criticized, but Robert’s have never even been questioned. If he expected Stephen to come and attack Arundel, as his rapid move away suggests, what was he doing leaving his half sister behind? Perhaps neither Stephen nor Robert believed that she was the key to the conflict ahead. William of Malmesbury seems aware of some criticism which might be made of Robert on this count, and says he had left her to the protection of Adeliza, who then broke her oath and did a deal with Stephen, but it is a rather feeble defence. It seems as if Adeliza had offered some promise to Matilda, but she told Stephen that ‘none of his enemies had reached England by her means’, and she was merely being hospitable.14

That Stephen saw Robert rather than Matilda as the threat, is shown by the explanation in the Gesta Stephani of Stephen’s immediate reaction. He came to Arundel, but learning that Robert had escaped, he left a force to blockade the castle, and himself set off in pursuit of Robert, concentrating ‘all his efforts on the capture of the earl’.15 But Robert travelled on quiet side roads and got safely through to Bristol. Stephen then turned back to the trapped Matilda. The same source gives another strange tale, also usually ignored. It says that Stephen’s brother, Henry, Bishop of Winchester, ‘had all the by-roads blocked by guards, and at length met the earl [Robert], it was rumoured, and after a compact of peace and friendship had been firmly ratified between them, let him go unharmed’. The author asks leave to doubt that such a story of treachery could be true, but gives the reader the impression that he believes it to be so, the bishop then returning to his brother without revealing the secret meeting. If the story is true, then the bishop must bear a great burden of responsibility for the war which followed, for the capture of Robert at this juncture would have given Stephen an unassailable advantage.

The same author then suggests that it was on the bishop’s advice that Stephen also made his next surprising move. The writer says that Stephen was keen to press on with the siege of Arundel, but Henry said it was ‘wiser for the king himself and more beneficial to the kingdom to let her go to her brother unharmed’, and concentrate on attacking Robert. We need not believe that Stephen was a gullible idiot who followed his brother’s advice automatically. His preference of Theobald of Blois for Canterbury is sufficient to dispel that idea, apart from his resistance to Henry’s efforts in the church council to condemn him for the arrest of the bishops. Neither brother was a particularly straightforward character, and both probably had a healthy distrust of the other’s actions. The pact of Bishop Henry with Robert is only a rumour rather than a known fact, but not to be utterly discarded; the advice to let the empress go seems far more certainly known to the anonymous author, who frequently shows himself to possess good information on Bishop Henry. Henry is often held up as the shrewder, tougher, more sensible of the brothers, yet on several occasions he gave advice of dubious benefit to his brother. It is unlikely that Stephen would have accepted the advice if he did not see its merit. As suggested above, he was faced with the dilemma of how to deal with Matilda, and the bishop’s advice offered a way out. So then ‘an agreement was made, and a truce accepted under sanction of an oath’.16 It is a pity we are not given the details of this agreement, and no reference is ever made to it afterwards. How wise Stephen was in this episode depends a lot upon what was agreed. Certainly Matilda was granted a safe conduct.17 William of Malmesbury’s reference to a boundary where Matilda was handed over to Robert, suggests some sort of frontier already recognized, somewhere west of Calne. It seems very likely that Matilda had promised not to war against him and broke her promise. That, after all, would seem to be Stephen’s main purpose in the move. Even if the agreement were not so specific, that must have been its intention. The Gesta Stephani suggests that the point was to put all his enemies in one region, but Matilda was no threat at Arundel, blockaded and isolated. That Stephen had gained something from this agreement is also suggested by William of Malmesbury, no doubt repeating the view of Earl Robert, that Adeliza had somehow been at fault here. William says that she had often pledged her faith, even sending envoys to Normandy, but had now reneged on it ‘with a woman’s fickleness’. But what ‘faith’ he refers to is uncertain; whether she had made some new secret pledge in support of Matilda in opposition to her husband’s stance, or whether it is the old promise extracted from Henry I.18

Stephen was able to pose as a merciful and chivalrous king, as even William of Malmesbury admits, providing an escort ‘which it is not the custom of honourable knights to refuse to anyone’. Another gain for Stephen was the removal of the problem of keeping forces at Arundel while Robert was his main target. Arundel was a strong castle, and Henry of Huntingdon points out the king was partly motivated by his knowledge that it would be hard to take. Orderic thought Stephen simply ‘foolish’, and that letting Matilda go was the cause of the war to come; but he was writing from a distance, from his monastery in Normandy, and ignored the fact that Robert had already escaped. Orderic suggests that the king’s enemies should have been cut down with the sword; the mind boggles at what would have been the church’s reaction if Stephen had taken that literally, with regard to Matilda!19

Stephen knew he must fight Robert, but there were still hopes that Matilda might be excluded from the conflict, hopes that were of course to be dashed, though she never actually took up arms. That she meant to join her half brother was no secret. After the agreement, she was escorted by Bishop Henry himself, and by Count Waleran of Meulan.20 Count Waleran went as far as Calne, and Henry continued to the arranged boundary agreed by the king, presumably as part of the agreement made at Arundel, where Robert met her and took her to Bristol. The west country base for Stephen’s opponents was already established.


Robert seems to have been disappointed in the number of men who came out in open support. William of Malmesbury, his apologist, says that ‘the nobles were either hostile or gave no help, apart from a very few’.21 One immediate ally was Brian fitz Count, who came from Wallingford for a brief meeting with Robert as he made his way from Arundel to Bristol. Another prompt response to the call came from Miles of Gloucester, who gave Matilda his homage, and accepted her into Gloucester, where she chiefly resided during the following years. John the Marshal, who held Marlborough Castle, was another early recruit. It seems that John’s rebellion had come just before the arrival of Robert, probably intended to coincide with it. Stephen was besieging Marlborough when he heard of their coming and set off for Arundel.22


The main war zone, 1139–41.

The position which had emerged in 1138 was now hardened into a stronger reality. Robert of Gloucester was in defiance of the king, and had arrived in England in person to defend his truncated estates. With the arrival of Matilda he could present his defiance as a defence of her claim to the throne, thus giving him a legality which would in her absence have been lacking or at least unconvincing.

Robert had been uncertain of how much support to expect. Now it was clear: Brian fitz Count, Lord of Wallingford; Miles, the castellan of Gloucester; a number of near neighbours in the west country; relatives; and feudal vassals. Not one of the greatest magnates in the land had joined Matilda, apart from Robert himself. The few advantages of the Angevin party were a compact west country base, considerable holdings in Wales which perhaps doubled their territorial base and from which troops could be drawn, though of questionable value for fighting elsewhere than in the border country, a moral argument over the oath and Matilda’s right, and the ability of the trio of leaders. They knew they had the support of David, King of Scots; but for the time being that was an inclination of the heart rather than a solid assistance. Since his defeat in the Battle of the Standard, King David was in no position to give much more than moral support. They could also hope for aid from Geoffrey of Anjou. His part in the end, in carrying on the war in Normandy, proved absolutely invaluable, even decisive, but his contribution to the English war, except indirectly, was negligible.

Stephen weighed up the possibilities and made his decision. He gave his priority to ending the English war, to finishing off the efforts he had already begun against the territories of Robert of Gloucester. His immediate problem was that the return of Robert had reinvigorated the efforts of his friends, and many of the gains which Stephen had made before September 1139 were soon to be lost again. Therefore, Stephen, as usual with undiminished vigour, though he might be forgiven a sigh of resignation at having to start all over again, set about reducing the minor fortresses around Bristol to allow an attack on the final target.

Before that, however, there was another closer target. Brian fitz Count had made his own position clear; he was even ‘delighted’ by Matilda’s coming.23 The caput of his honor was the great castle at Wallingford. The Angevin party had no more easterly stronghold and, indeed, Wallingford could either be viewed as their salient into enemy territory, or as an isolated stronghold. Its isolation made it vulnerable, and strategically Stephen had made a good choice of target. Unfortunately for him, it was a strongly built and strongly defended castle, with a double moat round three of its sides. The surviving foundations are massive, and suggest that this was a powerful structure. It was also under the resolute command of Brian fitz Count, and for all his efforts throughout the whole civil war, Stephen was never able to capture it. Wallingford alone is a perfect illustration of the value of castles in the warfare of this age.

We do not in fact know a great deal about Wallingford Castle. When an opportunity for excavation in the modern period offered, it was refused on the grounds that archaeological ability in castle work was inadequate to tackling the site, which was referred to as ‘a unique, uncut archaeological jewel’, and the decision was made to delay excavation to a later period. As was claimed, ‘there is not another Wallingford Castle’.24 This came about from plans to build flats and a house on the site in 1976. An earlier dig had established that it had been an Anglo-Saxon settlement, and Domesday Book shows that there had been some destruction of dwellings for the building of the castle.

The castle was built in the north-east corner of the Saxon burh, itself placed on the banks of the Thames at an important crossing place.25 It is mentioned as early as 1071, and was one of the first Norman castles in England, suggesting its vital strategic position. Basically, it was a typical Norman motte and bailey castle, and was strengthened in stone. But without excavation, and with the knowledge of so many periods of later work on it, including under Henry II, John and Henry III, it is not possible to describe precisely how it appeared in 1139.

At any rate, Stephen’s decision was to begin by attacking Wallingford. Brian had a sizeable garrison and an ‘impregnable castle’ with strong walls. Stephen planned at first to blockade the castle, but soon changed his mind on advice from his barons and decided, that as with Bristol, he was not yet ready for the major challenge.26 The strength of the place was clear on sight, and it was known that Brian had brought in plentiful supplies, ‘enough to last for a great number of years’. The king decided to build two counter castles and leave a besieging force in them while he went to wage war in the west. One of these counter castles either used the church of Saint Peter or was built alongside it.

His first target in the enemy’s base region was Trowbridge in the west of Wiltshire and on a route to the far west. It was held by Humphrey de Bohun III, who was married to Miles of Gloucester’s daughter, Margaret. The town around the castle was probably the largest in the county. Humphrey was in the conspiracy against the king, and had been advised by his father-in-law to make his castle ready for defence. But en route to Trowbridge, Stephen had two other successes. First, passing by the castle of South Cerney, on the Gloucestershire–Wiltshire border, built by Miles of Gloucester himself, Stephen decided on a sudden attack. He took the garrison by surprise and the castle fell by storm.

Then he approached Malmesbury, a prosperous little town then, with both an abbey and a castle. The castle was well built and held by a mercenary captain, Robert fitz Hubert: ‘a man of great cruelty and unequalled in wickedness and crime’.27 Robert had been employed by the Earl of Gloucester to hold the castle, but soon chose to surrender to the royal army, apparently on the advice of his relative, William of Ypres. Even William of Malmesbury sees merit in Stephen’s dealing firmly with this character.28

So Stephen approached Trowbridge, with two easy successes to boost his confidence. But his chief enemies now showed their hand for the first time, and their mettle. Miles of Gloucester was to prove the most effective of all the commanders engaged in the early stages of the war, and his death in 1143 was a boon to the king. Now Miles cut across Stephen’s rear, and attacked the besieging force he had left behind at Wallingford. He took them by surprise, wounding and killing those who fought him. Others were taken prisoner and chained up. Stephen’s initial minor successes had been nullified. It soon became clear that his enemies were too powerful for him to win an easy or quick victory. They would not be lured into a decisive battle, and their major fortifications were too strong to be easily captured. It was likely to be a long war of sieges.

In November 1139 an army from Gloucester attacked the city of Worcester. The Worcester chronicler vividly describes the fears and activities of the citizens in the months preceding this attack, which they had anticipated. Citizens took their furniture into churches for safe-keeping so that the church seemed like a warehouse. The cathedral became like an inn or council chamber, filled with chests and sacks, and the monks’ services were drowned by the howling of infants brought into the church by their mothers, as well as the crying of the women. The church seemed stripped bare, since valuable items were pulled down and hidden for safety, including the cross and a statue of the Virgin Mary, along with such things as curtains and vestments.

On Tuesday 7 November, what the chronicler calls the first day of winter, the enemy arrived while the monks were chanting the service of prime. The monks then dressed themselves in splendid clothes, thinking it better to wear them than leave them behind. They made a procession through the city, carrying the relics of Saint Oswald, while the church bells rang. But the attack proceeded, and the garrison of the castle, together with the citizens, resisted. The first attack was beaten off, and then a second was made against the north walls. This time the Angevins broke in and set fire to the houses in their path, so that the flames spread through much of the town. The attackers took away furniture, valuables, oxen, horses, sheep and cattle. Captives were bound in pairs like hounds and carried off so that they could later be ransomed. Then the attackers departed, ‘maddened and drunken’.

There had been a great deal of destruction, but the castle had held out. At the end of the month the new earl, Waleran, came to Worcester to view the damage. He was enraged, and made a revenge attack on Sudeley, held by John fitz Harold for Robert of Gloucester. Stephen also came to Worcester and was upset by the injuries the town had sustained. While there he appointed William Beauchamp, the sheriff of Worcester, as his constable to replace Miles of Gloucester.29

Stephen made a reconnaissance expedition to the west, but made no attack on either Hereford or Bristol, around which he raided. He did the same near Dunster, ‘leaving nothing at all, as far as it lay in his power, that could serve his enemies for food or any purpose’. This was good military thinking in the period, and common practice, but it very much helps to explain the poverty and hardship which accompanied the civil war in England, and especially those areas which bore the brunt of the fighting. William of Malmesbury sums up the situation by the end of 1139: ‘so the whole district around Gloucester far into Wales, partly through force and partly from good will, gradually went over to the lady empress in the remaining months of that year’. Of the following year, the chronicler says:

there were many castles all over England, each defending its own district, or rather plundering it. The knights from the castles carried off both herds and flocks, sparing neither churches nor graveyards … pillaging the dwellings of the wretched countrymen to the very straw.30

Also of 1140, Henry of Huntingdon, breaking into verse, exclaimed:

  Gaunt famine, following, wastes away

  Whom murder spares, with slow decay.31

Stephen had been worried by Miles’ activities, but not diverted from his initial target of Trowbridge, which he proceeded to besiege. The king built throwing engines and set up a blockade. But the garrison held out with defiance. The barons who had accompanied him began to murmur, clearly not relishing a long blockade. This was to prove one of Stephen’s greatest problems throughout the war. There were more lords who gave him half-hearted support than actually went over to the enemy, but it meant that he rarely had troops he could rely on 100 per cent. It meant also that the morale of his army was often less than that of an enemy which was more unified in its attitude, even if inferior in numbers. Both sides used mercenary troops, but Stephen had to rely on them more, and again this did not improve the morale of the barons with him or his army as a whole. In the end Stephen abandoned the close siege of Trowbridge, though he did not leave it altogether in peace. He put a reinforced garrison into nearby Devizes Castle with the intention of hampering any activity out of Trowbridge. These troops were ‘very prompt in battle’; in other words, probably light horse which were mobile.32 As Leland put it, Stephen could say of Trowbridge: I came, I saw, but he could not add, I conquered.33

At the end of 1139 Roger, Bishop of Salisbury, died. He perished of a quartan fever, one which recurred every fourth day. William Malmesbury suggests his health had declined because of the arrest, which had depressed him and led to ‘mental suffering’. Comments the chronicler: ‘fortune, which had favoured him greatly for so long, finally stung him cruelly with a scorpion’s tale’.34 His death underlines Stephen’s achievement in removing the bishops without serious repercussions. Roger left behind at Salisbury considerable wealth in money and vessels of precious metal, and the king used it, not we are told for his armies, at any rate not all of it, but to improve the churches at Abbotsbury and Malmesbury.35

Whatever the motive for allowing Matilda to join her friends in the west, Stephen must certainly have hoped to confine conflict to that region, but he was to be disappointed. The west remained the one solid base for Robert of Gloucester. David of Scots remained a brooding threat to the king in the north. But the region which proved most troublesome to Stephen, throughout the war, second only to the west country, was East Anglia and neighbouring districts. The nobles of the region who opposed him were less dedicated to Matilda’s cause than the Gloucester coterie, and hence their threat was more intermittent and less coherent, but they caused enough difficulty to divert Stephen from his main target at several vital moments in the war.

Already Nigel, Bishop of Ely, had apparently been involved in at least discussion with men conspiring against the king, and may even himself have acted against him. Stephen had taken no chance with regard to Nigel when moving against Roger of Salisbury in 1139. It was Nigel who had given Stephen the most problems then, preparing to defend himself in arms against the king at Devizes, and only surrendering because Matilda of Ramsbury wished to give way to the king’s threats against her lover and her son. Nigel had yielded with a bad grace. With Matilda present and war against Stephen opening up, Nigel came back into the limelight. He now joined the revolt against Stephen, hiring knights and attacking royalist neighbours.

Ely was an ideal base for rebellion. It possessed a castle at Cherry Hill, now in the park close to the cathedral. It was in essence a motte and bailey castle, but improved by Bishop Nigel with stone and cement.36 There was a second castle at the main entrance to the island, Aldreth. Even better, the isle was isolated by geography. In the twelfth century Ely genuinely was an island, cut off from the mainland by a swathe of rivers and marshes. It was ‘impenetrably surrounded on all sides by meres and fens, accessible only in one place, where a very narrow track affords the scantiest of entries’.37 It was very difficult country to cross. Normal entry to the island was by boat, and easy approach was hard. The nature of the country, and its protection to rebels, had been demonstrated by the rebellion of Hereward the Wake against William the Conqueror. Hereward had finally been flushed out, but only with great difficulty, and he had survived into an obscure end. During the civil war, Ely would twice harbour enemies of the king.

Stephen was probably inspired by a knowledge of the Conqueror’s campaign against Hereward, and as usual showed no decline in his vigour against new enemies. Now he set about building a bridge of boats, placed broadside on against each other, supported where necessary on hurdles laid across the marsh. So a timber bridge was constructed over the boats, and his army was transported to the island. Even so, there was difficult ground to cross, and the king had to rely on a monk who knew the area well, showing him a ford through the fens. This was the monk, Daniel, who was later rewarded by being made Abbot of nearby Ramsey Abbey.

Now Stephen’s men erupted into the Isle of Ely, capturing Nigel’s knights and seizing booty. They also captured the ‘small castle’ of Aldreth at the entry to the island, with its garrison, where some of Nigel’s knights had gone for safety. This castle has not been found, and is not obvious in any present remains.38 Bishop Nigel, like Hereward before him, managed to elude his enemies by fleeing through the marshes. He fled to Gloucester, ‘a poor and humbled man’. Stephen had conducted a rapid and brilliantly successful campaign in very difficult conditions, and had closed down the Angevins’ second front, at least for the time being.

Miles, after his victory at Wallingford, returned to the security of Gloucester, where his castle was well stocked and well fortified, and where Stephen’s enemies from less secure areas mustered. Miles set about a war of raiding and pillaging against royal supporters in Gloucestershire and Herefordshire, with ‘the terrible burning of villages and towns, which he turned into desert land’.39 That the Angevins had to use such tactics in an area so close to home demonstrates how restricted was their power. Among the places Miles captured was the castle at Winchcombe on 31 January 1140, where some of the garrison afterwards joined him. He also recaptured South Cerney and took Hereford. Miles had little difficulty in recapturing the latter from the small garrison left there, but he failed to take Sudeley.


Lands mainly under Pro-Angevin control, 1140.

The primary base of the Angevins in the west was expanding. Stephen himself, held considerable land in the county of Cornwall. William fitz Richard held much from the king, but now William turned against the king. His reasons are unknown, but he clearly had a connection with another of Henry I’s illegitimate sons, Reginald. Reginald does not appear to have been one of the initial ringleaders of the conspiracy, but it may simply be that we lack knowledge of his actions. He was to prove as consistent in his loyalty to Matilda as his fellow bastard, Robert. Reginald and William fitz Richard now made a pact. Reginald married William’s daughter, and was given entry to the castles which William held under the king in Cornwall. So far as we know this was Reginald’s first important link with the county, but it was to become his base for the rest of his life. Reginald’s activities in attacking church property soon brought him into conflict with the local church, and the Bishop of Exeter excommunicated him. But at first his rebellion looked as if it would be short-lived. Stephen, with his amazingly consistent energy against new threats, brought an army to Cornwall in the summer of 1140, to regain his own lands there. He recovered all the castles which had fallen into Reginald’s hands, with the exception of the one Reginald himself inhabited. Stephen introduced his own loyal supporter, Alan of Penthièvre, into Cornwall, gave the recovered lands into his hands, and shortly afterwards created him Earl of Cornwall. Matilda gave Reginald the same title, and the two rivals were left to contest their claims.40


The 1139–41 war zone.

In 1140 William of Malmesbury claims that Robert of Gloucester recaptured South Cerney, as well as Harptree and Sudeley, and that he levelled the counter castle at Wallingford. This is almost certainly one of the numerous instances of the chronicler exaggerating the role of his own hero, since, as we have seen, a less biased writer had said that it was Miles who took South Cerney and destroyed the counter castle at Wallingford, though perhaps he could be construed as acting on behalf of Robert. Reginald was made Earl of Cornwall, and William of Malmesbury says that this was done by Robert.41 No doubt Robert’s influence and wishes were regarded, but he could not in any way be considered qualified to create earls, even by his own side. It was Matilda who was claiming royal power in order to make such creations. Reginald certainly proved an effective choice, and was able to push the royal nominee, Earl Alan, out of the county, winning effective authority for himself.

The devious mercenary Robert fitz Hubert, thrown out of Malmesbury by Stephen, and employed by Robert of Gloucester, on 26 March 1140 suddenly seized Devizes Castle from the king. He was said to be related to Stephen’s mercenary captain, William of Ypres.42 William of Malmesbury calls him a blasphemer, saying that he was always boasting that he had once roasted alive eighty monks in a church, and that he would do so again. He wrote that Robert fitz Hubert ‘used to smear his prisoners with honey and expose them naked in the open air under the sun, stirring up flies and similar insects to sting them’.43 The chronicler says Robert’s aim was to build a principality for himself, and that he was planning to send for mercenaries from Flanders to aid him. But his career was cut short by John the Marshal, who believed that Devizes should be his. When the mercenary sought to make John agree to submit to him, the Marshal captured Robert fitz Hubert and, when later he refused to give up Devizes, hanged him on the spot. William clearly approved of the execution, though it seems a blatant case of taking the law into his own hands by John the Marshal.

It was still by no means certain that a full-scale civil war would develop. Already in 1140 peace overtures were made, according to William of Malmesbury through ‘the legate’, presumably meaning Henry of Blois.44 There was a conference at Bath, which Robert of Gloucester attended on Matilda’s behalf, while Henry of Blois, Theobald of Canterbury and Queen Matilda represented Stephen. The stumbling block appears to have been the clerics’ demand that the settlement be dictated by the church. Matilda was willing for this to be done, but Stephen refused. The conference, therefore, broke up without a settlement.

Later in the year, Henry of Blois, seeking peace in England, travelled to meet Louis VII and Theobald, Count of Blois, but still without success. Again, according to William of Malmesbury, Matilda and Robert were prepared to accept whatever proposals Bishop Henry brought back (we do not know what they were), but Stephen refused. Virtually throughout the reign the views and aims of Bishop Henry seem to have diverged from those of his brother Stephen. Then, says the chronicler, Bishop Henry abandoned his attempts and sat back ‘to see how things would turn out’.45

In 1140 Stephen returned to Worcester, and used it as a base for further raids. He also made an attack on Robert’s property at Tewkesbury, where the earl’s ‘magnificent house’ was set on fire, along with all the property around it.46 There was some sort of fighting near Bath on 15 August 1140, but we have very little detail about it. Robert of Gloucester had set out against Bath, one of the few important strongholds held by Stephen in the west. But Stephen had already sent men who probably laid an ambush. Two knights called Roger and John led the royal forces. One of those in Robert of Gloucester’s army was Geoffrey Talbot, for whom Stephen had developed a healthy dislike since the previous episode at the same city, when he had escaped the bishop’s clutches. In the fighting Geoffrey was severely wounded, and died a week later on 22 August.47 He was one of the first major casualties of the war, and his death a success for the royalists. The chronicle also says that the clash itself was a royalist victory.

In September 1140 the Angevins made an attack on Nottingham, which like Worcester before it, was sacked. It is a sign of the respective power and aims of the two sides that the Angevins should pursue this tactic. The idea was clearly to try and show Stephen’s weakness, but it was not the action of a powerful or winning party. If they hoped to rule over such towns as Worcester and Nottingham, such destructive measures were not likely to aid their cause, so the actions speak more of desperation. At Nottingham, ‘the whole city was destroyed by the flames’.48 Even those in the churches died, as the buildings caught fire and collapsed upon them.

The first year of war had seen little more than the two sides settling down and testing out their opponents. The Angevins had to accept that their loyal territory was very restricted in area. Their main aim must be to seek new allies and increase the territories under Matilda’s power. Their main hopes in this respect lay in keeping Wallingford, while building more support in the lands around it, and in opening up another area of war, in which case East Anglia looked the best bet, while keeping touch with the sympathetic David of Scots. They clearly also hoped that Geoffrey of Anjou would succeed in Normandy, and ultimately lend a hand in England.

The royalists had to accept that it was going to be difficult to capture the main bases at Bristol and Gloucester, and even the Angevin outpost at Wallingford had proved impossible to take. It was unlikely that the Angevins would risk a pitched battle, so the probable prospect was for a prolonged and bitter siege warfare. Stephen’s aim in this first phase of the war proper, had been much the same as in his earlier operation: to subdue any additional threats to the peace in the country, and to take the lesser fortifications around the major strongholds in the west, with the ultimate aim of besieging Gloucester and Bristol. But as in many wars, fate had a surprise in store for all concerned.


  1.  J. Bradbury, The Medieval Siege, Woodbridge, 1992, pp. 113–14.

  2.  Orderic, v, pp. 106, 324; Anna Comnena, The Alexiad, ed. E.R.A. Sewter, Harmondsworth, 1969, pp. 348–9; Die Kreuzzugsbriefe aus den Jahren 1088–1100, ed. H. Hagenmeyer, Innsbruck, 1901, pp. 138–40, 149–52.

  3.  J. Bradbury, ‘Greek Fire’.

  4.  Bradbury, ‘Geoffrey V’, pp. 21–38, pp. 34–5; S. and M. Nikitine, L’Émail Plantagênet, Nancy, 1981.

  5.  See F.H. Russell, The Just War in the Middle Ages, Cambridge, 1975.

  6.  Gesta Stephani, pp. 86–7.

  7.  Ibid., pp. 84–5.

  8.  Davis, Stephen, p. 37.

  9.  Gesta Stephani, pp. 84–5.

10.  William of Malmesbury, p. 34.

11.  Gesta Stephani, pp. 86–7; John of Worcester, ed. Weaver, p. 55, says it was in October and the landing at Portsmouth, and n. 4, that the G manuscript gives August, but this seems muddled, and is usually considered to be incorrect; Worcester, in Stevenson, p. 361, following the error of Thorpe has an account with a landing both in August and October! Orderic, vi, pp. 534–5 merely says they arrived in the autumn.

12.  Worcester, in Stevenson, p. 361.

13.  Gesta Stephani, pp. 86–7.

14.  Worcester, in Stevenson, p. 361.

15.  Gesta Stephani, pp. 88–9.

16.  Ibid.: ‘Datis igitur dextris, et sub iureiurando acceptis induciis’. This is one of the examples of ‘right hands given’, probably an actual symbolic handshake to confirm an agreement.

17.  Henry of Huntingdon, ed. Forester, p. 272; Henry of Huntingdon, ed. Arnold, p. 266: ‘ire permisit ad Bristowe’.

18.  Gesta Stephani, pp. 88–91; William of Malmesbury, p. 35.

19.  William of Malmesbury, p. 35; Henry of Huntingdon, ed. Forester, p. 272; Henry of Huntingdon, ed. Arnold, p. 266: ‘quia castrum videbat inexpugnabile’; Orderic, vi, pp. 534–5.

20.  Gesta Stephani, pp. 88–9; and n. 1, where the editor suggests that the bishop may have been Matilda’s choice for an escort.

21.  William of Malmesbury, p. 35.

22.  John of Worcester, ed. Weaver, p. 55, n. 4, from the G manuscript of the Gloucester continuation: Worcester, in Stevenson, p. 361. Since this is in the uncertain account of the Worcester chronicle, the dating remains unclear, but must have been at about the time of the Arundel landing.

23.  Gesta Stephani, pp. 90–1: ‘qui de illorum aduentu eximie laetificatus’.

24.  Tom Hassall, ‘Wallingford Castle: life on the bailey’, The Times, 1976.

25.  D. Renn, Norman Castles in Britain, 2nd edn, London, 1973, pp. 337–8; R.A. Brown, Castles from the Air, Cambridge, 1989, pp. 219–20.

26.  Gesta Stephani, pp. 90–1: ‘firmato inexpugnabili’.

27.  Ibid., pp. 92–3.

28.  William of Malmesbury, p. 37; Worcester, in Stevenson, p. 366.

29.  John of Worcester, ed. Weaver, pp. 56–8; Worcester, in Stevenson, pp. 362–4.

30.  William of Malmesbury, pp. 36, 40–1.

31.  Henry of Huntingdon, ed. Forester, p. 273; Henry of Huntingdon, ed. Arnold, p. 267.

32.  Gesta Stephani, pp. pp. 96–7.

33.  W.H. Jones, ‘Early Annals of Trowbridge’, in VCH, Wiltshire, vii, 1953, p. 216.

34.  William of Malmesbury, pp. 37, 39.

35.  Gesta Stephani, pp. 96–9.

36.  Liber Eliensis, ed. E.O. Blake, Camden Society, 3rd ser., xcii, London, 1962, p. 314.

37.  Gesta Stephani, pp. 98–9.

38.  Ibid., pp. 100–1: ‘castellulum’.

39.  Ibid., p. 94.

40.  Ibid., pp. 102–3. The castle Reginald retained is not named.

41.  William of Malmesbury, p. 42.

42.  Worcester, in Stevenson, p. 366; Worcester, ed. Thorpe, ii, pp. 125–6.

43.  William of Malmesbury, pp. 43–4.

44.  Ibid., p. 44.

45.  Ibid., pp. 44–5.

46.  Worcester, in Stevenson, p. 365; Worcester, ed. Thorpe, p. 124: ‘magnificam domum’.

47.  Worcester, in Stevenson, p. 367; Worcester, ed. Thorpe, p. 128.

48.  Worcester, in Stevenson, p. 368; Worcester, ed. Thorpe, p. 128.

If you find an error please notify us in the comments. Thank you!