GO TO:      Introduction    |    part II - the Burghal Hidage    |    bibliography

PART I - The defeat of the Vikings in Mercia and London

Introduction

The decade 876-886AD saw a total reversal of the position of King Alfred of Wessex in his relations with the Viking armies, from one of near subjugation, with the serious possibility of the annihilation of Wessex as an independent kingdom, to one in which Alfred had become the undisputed lord of the whole of central and southern England. The understanding of these developments has been significantly advanced by recent studies on Alfred's relationship with Mercia (Keynes 1998), the economic history and the coinage (Maddicott 1989; Blackburn 1998), the military background (Abels 1988; Peddie 1999), and by biographies of Alfred, from rather different viewpoints (Smyth 1995; Abels 1998). In a number of recent publications, Simon Keynes has explored the concept of the 'kingdom of the Anglo-Saxons' as constituting a new political order in the early 880s which obtained over much of what had before been the separate kingdoms of the West Saxons and Mercia (cf Keynes 2001: 44-5, 57-62).

However, a combination of the physical evidence of the fortifications which Alfred deployed in this process, and the relevant documentary evidence that may be regarded as being contemporary - in partiular the Treaty between Alfred and Guthrum and the Burghal Hidage document - casts the strategies which Alfred had in mind, and the tactics he employed to achieve his military and political goals, in a rather different light than has hitherto been accepted. In doing so, it provides an explanation as to how the new kingdom of the Anglo-Saxons came about.(1).

In summary, it will be argued that as a result of the Partition of Mercia by the Vikings in 877, king Alfred was deprived of his interests and influence in the London area, and probably the whole of the rest of Mercia, which he had enjoyed since his alliance with Burgred. In the period immediately following his victory over Guthrum's army at Edington in early 878, Alfred implemented military and political strategies whose goal was the removal of the Viking armies from Mercia and from London, over which they still exercised control from bases at both Cirencester and Fulham, and from which they still directly threatened Wessex. This strategy involved two complementary aspects. The first of these was the construction and garrisoning of a system of fortresses in Wessex and central Mercia, which both defended Wessex in depth and acted as offensive instruments against the Vikings in Mercia. The second involved the assumption by Alfred of overlordship of Mercia in a way which could be interpreted either as a coup d'etat within Mercia itself, or as a more gradual process which involved Alfred taking control of Mercia after Ceolwulf's demise. It is suggested that these tactics enabled Alfred to confront, and to dictate terms to, the two potentially hostile Viking armies stationed at Cirencester and Fulham, with the consequence that they were forced to leave Mercia. Guthrum's army in Cirencester retreated to East Anglia, and the Fulham army left for more rewarding prospects on the Continent - both moves recorded in the Anglo-Saxon Chronicle as occurring in late 879. It is furthermore argued that the document known as Alfred and Guthrum's Treaty, in which the limit of Guthrum's kingdom of East Anglia is defined by a boundary drawn to the east of London, was a contemporary record of the agreement reached between Alfred and Guthrum at this time.

The second part of this paper presents arguments for the view that the original Burghal Hidage document, which records some details of this rapidly executed system of fortresses in Wessex and central Mercia, was more-or-less contemporary with the creation of this system in 878-9.

Alfred and Guthrum's Treaty

A unique window into the elucidation of Alfred's strategies, which in the decade in question hinged around his relationship with the Viking leader Guthrum, is provided by the undated document known as Alfred and Guthrum's Treaty (Keynes & Lapidge 1983: 171-2). To appreciate the importance of this for the present discussion it is necessary to determine the events and processes which can be reasonably inferred to have led to the situation it records. David Dumville has recently argued (1992a) that the terms of the Treaty - and in particular its boundary clause, which is drawn around the eastern and northern sides of London - are those appropriate to the context of the conference between Alfred and Guthrum at Wedmore and Aller in May 878, soon after the former's victory over the Viking forces at Edington. He thus sees the boundary as defining Guthrum's territory in Mercia to the west, with Alfred's territory to the east.

See map

Some arguments against Dumville's thesis have been voiced more recently by Simon Keynes (1998a: 31-4), but its strategic implications have not been analysed. It is argued here that the two principal premises underlying Dumville's interpretation are ultimately untenable. The first is that the terms of the Treaty must be associated with a known treaty between both parties (i.e., one recorded in documentary sources) which might reasonably be inferred to have included these terms. Since the circumstances of which the Treaty was the result are not indicated by the terms of the document itself, there can be no demonstrable connection between these and any occasion in which a treaty - known or unknown - was entered into by the two parties. This leads Dumville to examine all the known occasions of such meetings and treaties, leading him by a process of elimination to the conclusion that the Treaty must belong to the occasion of the agreement between the two parties at Aller / Wedmore in May 878. This assumes however that the Anglo-Saxon Chronicle does in fact contain a record of all those occasions on which treaties were agreed between Alfred and the Vikings, with the inference that there were no others. In view of the known gaps and omissions in the Anglo-Saxon Chronicle it seems unsafe, to say the least, to infer that only those events recorded in it were of any significance. There is therefore every likelihood that there were other treaties which the Chronicle did not record, to which the terms of the extant Treaty might be equally or more appropriate.

Dumville's second premise, which logically follows his acceptance of the first, is that since the Vikings still held Mercia in 878 the Treaty gives the area to the east and north of the boundary to Alfred, while giving Guthrum territory to the west, i.e. all of Mercia. There are however a number of considerations of a strategic nature which Dumville has barely mentioned, and which make his thesis ultimately untenable. The following arguments will be clear from the map. Fig 000.

If, as Dumville argues, the terms of the Treaty implied that Guthrum ceded the whole of Essex to Alfred because it had always been part of the latter's kingdom, then it would be expected that the boundary in the Treaty would have been more-or-less coterminous with the land boundaries of Essex. Instead, however, the agreed boundary progresses from the Thames along the common border of Essex and Middlesex along (up) the river Lea, but departs from this line to swing westwards, following the Lea to its source, then northwards to Bedford on the Ouse and then westwards again to Watling Street. It thus inexplicably leaves to Alfred not only Essex but also the whole of north-east Hertfordshire as well as eastern and northern Bedfordshire, including Bedford itself. Without making any assumptions about the courses of the boundaries of the E. Midland 'shires' (or even the existence of discrete Viking armies based on shire centres) in the later 9th century, this would have left to Alfred an irregular and narrow strip of land between the agreed boundary on the south and (it has to be presumed) the southern borders of Northamptonshire, Cambridgeshire and Suffolk on its northern edge, which areas Alfred can never have held at any time. If Alfred obtained this territory under the terms of the Treaty, it also raises the question of why the agreed boundary did not define its western and northern edges - a difficulty pointed out, but not resolved, by Dumville.

This would have given Alfred control of a narrow strip of territory to which Wessex had no historic claim, which would have offered him no strategic advantage, to which he would have had very limited access, and which would have severely taxed his resources to 'defend' against Viking forces ranged along nine tenths of its boundaries - even given the unlikely assumption that he was able to gain the formal allegiance of the population. From Guthrum's point of view, the fact that Alfred would have had access to this wedge of territory would have considerably weakened his own hold on areas to the north (East Anglia), west (Mercia), and south (London and its region) - particularly since Alfred's territory impinged on the strategic corridor of Watling Street. This arrangement would therefore have made strategic nonsense from the points of view of both Alfred and Guthrum.

It may be doubted, for similar reasons, whether king Alfred had any meaningful control in Essex in the later 870's. It is argued below that from the partition of Mercia in August 877, when Guthrum's Vikings took effective control of Mercia (ASC sa 878), it is probable that London (as part of Mercia) was taken over by the Vikings. From this time passage for Alfred's ships up and down the Thames would have been severely restricted. This position was aggravated - or from the Vikings' point of view consolidated - by the arrival in the Thames in probably the summer of 878 of a large Viking army which settled at Fulham, which according to Asser (ch.58) "made contact with the army further upstream", by which it may be reasonable inferred that it came to support Guthrum's attempted take-over of Wessex, just as the army led by 'the brother of Ivar and Healfdene' had attacked Devon earlier in the year in support of Guthrum (Smyth 1995: 55-7). As will be discussed below, it thus effectively consolidated the annexation of London, and must have prevented any access along or across the Thames by forces in Wessex. Essex was therefore surrounded by Viking-held territory to the north and west, and by the Viking-dominated Thames estuary and the sea to the south and east. The proposition that Alfred controlled Essex after the partition of Mercia in 877, even before the arrival of the army at Fulham, in the sense of either receiving military allegiance from its people or being able to exact tolls and taxes and other revenues, cannot really be credited. In view of these considerations, Dumville's assertion that the Chronicle's record of the Vikings leaving Alfred's kingdom in 878 would have meant that they would have felt obliged to vacate Essex must by called in question.

Lastly, it is clear from the events recorded in the Chronicle that Alfred was negotiating with Guthrum at Aller and Wedmore from a position of strength. Alfred was the undisputed victor at the battle of Edington, and had received Guthrum's submission in a solemn ceremony of Christian baptism. It is therefore quite unreasonable to suppose that in the negotiations leading to the Treaty Alfred would have accepted control of an indefensible rump of territory in Bedfordshire and Hertfordshire in which he had no historic interest, and which gave him no strategic advantage, while allowing Guthrum to take the cream: all of Mercia, as well as London, in which Alfred only a little while previously had had considerable economic and perhaps strategic interests. Alfred would certainly not have allowed himself to have been bullied into accepting such a one-sided, disadvantageous and precarious state of affairs by a vanquished Viking leader who had just accepted a public submission to himself as overlord. Furthermore, if Alfred himself had agreed a treaty with Guthrum's Vikings to allow them control of Mercia and London, this raises the question as to why these same Vikings abandoned these newly acquired lands in 879 to return to East Anglia, and why the army encamped at Fulham also chose to abandon this easy ride at the same time.

Since the logic of these arguments makes it difficult to accept Dumville's interpretation, it is necessary to follow Keynes (1998a: 33; 2001: 45-8) and return to the 'orthodox' interpretation of the Treaty - that its terms gave Alfred control of lands to the west of the boundary and Guthrum control of the area to its east. The Treaty cannot therefore have arisen from the circumstances of the negotiations at Aller and Wedmore in mid 878, since at this time the Vikings still held London and Mercia, and Guthrum was to spend the next 12-13 months or so in Cirencester (2), a position which (as will be discussed below) left him poised threateningly on the borders of Wessex. A context for the Treaty must therefore be sought in a situation in which the Vikings were about to leave or had just left Mercia altogether.

This is indeed what happened in late 879, when the Chronicle records (sa 880) that Guthrum's army "went from Cirencester into East Anglia and settled there and shared out the land", and that "the army which had encamped at Fulham went overseas…". There are therefore grounds for arguing (as have both Simon Keynes and Richard Abels - Keynes 1998a: 31-4; 2001:45; Abels 1998: 163-4) that the origin of the Treaty belongs to this time. However, the suggestion of Keynes that "it is conceivable that the Treaty was drawn up as early as 880 (when Guthrum first settled in East Anglia)" (1998a: 33) appears to be based on the premise that at the time it was drawn up Guthrum had already left Mercia and was settled in East Anglia. There is therefore no explanation on offer as to why it should have been drawn up at all, except to regularise an already existing situation, and does not answer the problem of how or when this situation had come about. The alternative is to view the terms of the Treaty, and the boundary clause which is its first provision, as being appropriate to the context of a meeting or conference held in the latter half of 879 at which Alfred and Guthrum agreed the terms of the Vikings' departure from Mercia and London. It was at this point - and not before - that the Vikings were forced to accept that they did indeed have no choice other than to leave Mercia for good. This implies that the conference in which all these agreements were thrashed out was held somewhere in Mercia, or even possibly at a place on or near the border defined by the Treaty, sometime in July or August 879 (3).

It might be argued that Guthrum and his army agreed to the terms of the Treaty as a natural extension of their alliance with and submission to Alfred established at Aller and Wedmore earlier in the previous year, or even that the substance of the terms had been agreed at this stage (eg Whitelock 1979: 31; Smyth 1995: 92). But Guthrum's army, still formidable and largely intact (Loyn 1977: 59; Smyth 1995: 85-92), displayed its still hostile military and strategic intentions by staying at Cirencester for a further year. As a royal tun, the site of a large minster church, and the collecting place of produce from a wide area (Gerrard 1994: 90-1), it would have had a natural attraction to a Viking army requiring sustenance over the winter. And most importantly, as a probably still adequately walled Roman town Cirencester represented the most readily defensible site within Hwiccian territory which was nearest to the borders of Wessex. It occupied a nodal position on the Roman road system which, in addition to being connected directly to Gloucester to the west and to the Mercian heartland to the north-east, commanded the easiest and most direct access into Wessex - to the south-west along the Fosse Way, and to the south-east along Ermin Street. It is quite clear from the map (Fig. 000) that no other position outside the borders of Wessex could have been chosen by Guthrum to have posed a greater threat to its continued security. Although by moving with his army to Cirencester Guthrum fulfilled the condition he had accepted that he would move away out of Alfred's kingdom, these developments show Guthrum to have been essentially two-faced - accepting an alliance with Alfred for his own convenience, if not also his political aggrandisement and the validation it gave him of his own importance and acceptability, - yet on the ready, in some sort of alliance with the army at Fulham, to attack Wessex again. The terms of the Treaty a year later therefore represent a new understanding and a new response to a new situation.

The extraordinary nature of the retreat of the two armies from Mercia in late 879 is brought into focus even more sharply when it is considered how the political and military relationship between Alfred and Guthrum must have been drastically altered by the arrival in the summer of 878 of the second Viking army which settled at Fulham, within a fortified enclosure of Roman origin on the river Thames upstream of London (Vince 1990: 83-4 & fig.42). It is likely that this move, as Asser pointedly remarks (ch 58), was intended to reinforce Guthrum's army in Cirencester. The army at Fulham occupied a position which secured and reinforced Viking domination of both London itself and its territory, as well as the whole of the Thames estuary and the river upstream of London. Equally importantly, it commanded Akeman Street, which led westwards into the heart of Wessex over a crossing of the Thames at Staines at the very south-west corner of Middlesex, the ancient kingdom of the Middle Saxons, and the westward extent of the territory dependent on London. It therefore placed itself in the most advantageous position from which it could both control the London area and directly threaten Wessex. From this it can be inferred that it was intent on further conquest in Wessex - especially when these tactics are seen in the light of its subsequent damaging exploits on the continent. Since the two Viking armies together comprised a significantly greater threat to Alfred's control of Wessex than that posed by Guthrum's army on its own, the circumstances which governed the dealings between Alfred and Guthrum immediately after the latter's defeat in early 878 were therefore radically changed. The Fulham army had no treaty agreement with king Alfred, were bound by no alliance, and had no memory of defeat in battle by Alfred. The balance of power had therefore shifted decisively in the Vikings' favour. The potential attacking force of the combination of Guthrum's army at Cirencester and the other army at Fulham, poised on the borders of Wessex, must have been quite formidable. As Professor Whitelock has observed (with a degree of perspicacity lacking in most other commentators) "With Danish armies so uncomfortably close as Cirencester and Fulham, the year from the autumn of 878 to that of 879 ... must have been an anxious one for the West Saxons" (1977:13). The threat to Wessex posed by these armies must have had a powerful and galvanising effect on Alfred - and indeed on the whole population of Wessex. The fact that both armies were persuaded to leave Mercia in 879 demonstrates that by this time there was a new set of factors that put Alfred once again on the offensive, and which gave him a degree of power and leverage which was more radical and far-reaching than that which had allowed him to dictate terms to Guthrum alone in mid 878.

Whitelock also observed that had the information in the Chronicle been more complete, "we might have known by what means the two Danish armies were persuaded to leave Mercia". (1977:9). But the significance of this event lies in the fact that this retreat to East Anglia must have represented for the Vikings a painful abrogation not only of their military, political and economic domination of the whole of Mercia, which they had begun to regularise in the partition of August of 877 (see below). It also deprived them of any hope of further conquest of Wessex, a territory they had fought so hard to subdue for the previous few years, and which they had in part begun to settle. It was, with hindsight, a crucial turning point in Alfred's dealings with the Vikings, which was comparable in significance to his victory at Edington at the beginning of the previous year, and an event which clearly set the stage for Alfred's subsequent political domination of Mercia and the emergence of Keynes' 'kingdom of the Anglo-Saxons'.

Since this analysis of cause and effect differs from the conclusions of a number of historians, it is worth examining these in more detail. Stenton for instance merely repeats what the chronicle says (1971: 257) without further comment, and provides no explanation as to why the Fulham army left in 879. Whitelock, in spite of her reservations noted above, assumes that the retreat to E Anglia by Guthrum's army was a direct result of Alfred's victory at Edington (1979:31). Charles-Edwards suggests that the move of both the Fulham army to Ghent and Guthrum's army to E Anglia was the result of the 'alliance' with Alfred forged with Guthrum in 878, and, furthermore, that the Fulham army moved to the Continent to take advantage of the succession struggle of the Franks from the spring of 879 (1998: 49). But the fact that the second army arrived at Fulham only after this alliance had been forged between the Saxons and Guthrum shows that it saw some military advantage in doing so, in spite of this alliance. This alliance could hardly therefore have been the cause of its retreat. It may well have been drawn to the Continent in 879 by an eye for the main chance, but then this is precisely why it must have been attracted to the London area and why it had chosen a position poised on the borders of the West Saxon kingdom in 878. It is argued here that the only explanation for its movements is that, seeing Alfred's defensive arrangements taking shape throughout the autumn, winter and spring of 878-9 (described further below), it saw its opportunities of further conquest in Wessex evapourate, and decided as a result to grab the 'main chance' elsewhere. Smyth has also argued (1995: 87) that the removal of the Fulham army to the Continent was in some way the consequence of the strength of the peace agreed between Alfred and Guthrum (also bypassing the problem of why this army came to Fulham in the first place) and, on the premise that Asser is not a contemporary witness, dismisses the strength of his evidence for cooperation between the two armies. Lastly, Abels, who also dismisses Asser's evidence, suggests that the arrival of the Vikings at Fulham "may have persuaded Guthrum to look to his interests in East Anglia", and that Guthrum's departure drove off the Fulham Vikings to Ghent soon after (1998: 163, 175). In other words, Abels suggests that these events have an inbuilt momentum of their own and had nothing to do with Alfred's strategies.

None of these suggestions, however, explains why it was that the second Viking army - which to judge from its subsequent exploits recounted in the Chronicle and in Frankish annals (ASC sa 880-85; Smyth 1995: 101-116; Charles-Edwards 1998) was large, powerful, well-equipped and ferocious - came to winter at Fulham in a strategically threatening position in the first place. Most commentators seek to diminish Asser's near-contemporary witness for the Fulham army's collusion with, if not active support for, Guthrum. And none take into account the inference from the strategically advantageous positions of these two armies on Roman roads leading into Wessex that both were intent on further conquest. There seems to be no adequate explanation for the abrogation by both armies of these positions dominating Mercia and their inferred hostile intentions towards the West Saxons, apart from their recognition of some decisive military advantage on Alfred's part. This being so, the agreement of the terms of the Treaty requiring Guthrum to move back to E. Anglia would have deprived the Fulham Vikings from their expected support in their designs on Wessex, which in turn would have encouraged them to seek new conquests elsewhere.

These arguments so far can be summarised thus: the Anglo-Saxon Chronicle records the retreat of the Viking armies from Mercia in late 879, which can be associated with the (undated) record of an agreement (the Treaty), whose terms and provisions would be the appropriate and expected outcome of a negotiated agreement between Alfred and Guthrum at this juncture. There is therefore a reasonable basis for the inference that the Treaty was the contemporary record of the position agreed by both parties at negotiations which took place in the summer of 879, at which the Vikings agreed to leave Mercia and London. It can also be inferred that Alfred was able to dictate the terms implied by the existence of the Treaty from a position of military superiority which was recognised and acknowledged by the Vikings. It is therefore necessary to provide an explanation for the extraordinary nature of these events and outcomes in terms of processes that can be anchored within a short period of time - effectively between May 878 and the late summer of 879. It is suggested that the underlying factors that precipitated this remarkable turn of events can only be appreciated by examining the strategies of both Alfred and Guthrum up to this point.

The strategies of Alfred and Guthrum

The basic premise underlying this scenario is that Alfred negotiated the Treaty in late 879 from a position of such strength that Guthrum and his army, who it can be inferred still harboured hostile intentions towards Wessex, had no choice but to accept his terms. This position was of course a total reversal of that in January 878, when the West Saxon kingdom was in very real danger of becoming a Viking client state. In the previous year, the Viking army had moved to Gloucester, in the territory of the Hwicce, a manoeuvre which is shown by the Anglo-Saxon Chronicle to have been made in the context of a partition of Mercia imposed on the client king Ceolwulf II. In this process the Roman fortress of Gloucester would have formed a natural focus as a port within Mercia for the Vikings, as a bridging point of the Severn connecting Mercia and S.Wales, and as a secure base from which the Viking armies could threaten central Wessex from the north, to which they had easy access along the Roman road system.

The partition of Mercia in August 877 was clearly part of this strategy to take over the kingdom. Reasons are given below for suggesting that in this process the Vikings were able to assume direct control of London and its area, central Mercia (present-day Oxfordshire and Buckinghamshire), and western Mercia (essentially the old kingdom of the Hwicce) dependent on a 'capital' at Gloucester. Furthermore, this process of Partition can be seen as interrupting a long process, demonstrated by both the coinage and charters, which had begun in the 860s if not earlier, of involvement of the West Saxon kings in the affairs of Mercia as a whole and of London in particular - an alliance in which king Alfred had become the dominant partner (Keynes 1998a: 7-13). That the Vikings had taken control of London at this point provides the necessary context for the arrival of the other large army which sailed up the Thames past London in the summer of 878 to encamp at Fulham, the strategic importance of which is pointed out above. Further support for this hypothesis is provided by the existence of the Treaty itself, which demonstrates that the Vikings were obliged to give up control of an area they had held until this time (4).

This situation is also somewhat at variance with the hypothesis argued by Cyril Hart (1992: 115-25) that Essex "remained under West Saxon administration throughout the period of Danish autonomy in East Anglia" (ibid: 125). The inference from the strategic considerations are however clear - that for a time the Danes held control both of Essex and of the London area, which arguably also comprised Middlesex and Hertfordshire. This must be seen as part of an wider super-kingdom, which there is every reason to suppose they were in the active process of extending over Mercia, and in which they wished to forcibly include Wessex. It may well be that Essex kept its 'English' (West Saxon) ealdorman, as Hart suggests, but this must have been for a short period under Danish control - just as the Vikings seem to have pursuaded Ealdorman Wulfhere of Wiltshire to serve their interests after the rout of Alfred at Chippenham and their assumption of control in part of Wessex from early 878 (Abels 1998: 152). That the West Saxons had any meaningful control in Essex as a whole from 879 to 917 seems to be gainsaid by the efforts of Edward the Elder recorded in the Chronicle to gain control of the area from 911 onwards by a sustained programme of military campaigning and fortress-building, which the writer has argued (1997) was designed to ensure the permenant submission of the population to the West Saxon king. Furthermore, if this were not enough, the very existence of the Treaty between Alfred and Guthrum shows that Alfred had to concede any sole rights he may have thought he had in Essex to the control of the Vikings at the time it was drawn up - albeit in return for his regaining control of London and its region.

The consolidation of the Scandinavian interests in the London area after 877 has suggested to the present writer the existence of a boundary between their territory around London to the east and the rest of Mercia to the west (Haslam 1997: 118-23) (6). This however would not have been a defensive frontier, in the manner originally suggested, since the Vikings effectively held Mercia to its west as well, but can be suggested as being the delimitation of a particular area of administration. At the time of the Partition, Guthrum's Vikings must have felt that they had begun to realise their goal of creating an enlarged Scandinavian 'kingdom' of East Anglia. The course of this boundary between the central and eastern areas of (former) Mercia can be most reasonably suggested to have run along the western side of the ancient shire of Middlesex - the territory of the Middle Saxons - i.e. along the Colne river northwards from Staines. This must represent the western limit of the area formerly dependent on London. The boundary is likely to have continued directly northwards to the source of the Colne and across the Chiltern Hills to the river Ouzel. This river was certainly an important boundary between the Vikings and the West Saxons in the early 10th century (Haslam 1997: 119). Its logical course would have been to follow the Ouzel to where the latter crossed Watling Street at Fenny Stratford, and thence northwards along Watling Street. Its course to the north of this has been suggested by Gelling (1992: 128) as following Watling Street along the eastern border of Warwickshire, branching northwards off Watling Street to follow the eastern borders of Staffordshire and Cheshire.

From a strategic point of view the establishment and consolidation of Viking domination within the territory defined within this boundary would have given the Vikings control over both a wide area around London, as well as a considerable length of the Roman roads leading into London from the north, the northwest (Watling Street) and the west (Akeman Street). The existence of this boundary of 877 supplies the explanation for a fact which has puzzled many historians in the past - that the boundary recorded in Alfred and Guthrum's Treaty stops short at Stoney Stratford (where the Ouse is crossed by Watling Street), when it would be expected to have continued northwards. The Treaty would not be expected to have mentioned any such continuation if a pre-existing boundary beyond this point was not altered. From this it can be inferred that the area to the east of this - in effect, most of present-day Northamptonshire, Bedfordshire and areas to the east - had been part of Viking-held East Anglia at least from the partition of 877. The boundary in the Treaty can therefore be best interpreted as the modification of an earlier boundary to the west and north-west of London (which therefore had included London within Viking-held territory), a change forced on the Vikings by Alfred to give him strategic control of London and Watling Street. The nature of this Treaty boundary as a modification of an already existing situation is one of the clearest indications from which the existence of the Scandinavian control of London, which included the surrounding area to its north and west, can be inferred before this time (5).

It is however the coinage evidence, discussed most recently by both Mark Blackburn (1998) and Simon Keynes (1998a: 15-18), which clarifies the position around this time most exactly. Keynes has concluded that "Alfred was recognised, at some time before Ceolwulf's demise ..., as one of the two kings who had superior authority" in Mercia (ibid:16). Blackburn has demonstrated that Alfred's interests in both the London mint and another in southern Mercia in the mid 870s, which can be taken back to Burgred's reign, were superceded by a coin series issued by Ceolwulf alone (ibid:117-20). He concludes on the internal evidence of the coinage that Ceolwulf's London coin issues start in 877/8, and finish with the end of the reign in probably 879 (ibid: 119-20). From this he also draws the inference that during the duration of the issue of the Cross-and-Lozenge type from 875 to 879 - ie for both the Alfred-and-Ceolwulf and the Ceolwulf-alone phases - "London was in Anglo-Saxon rather than Viking hands" (ibid: 120). This conclusion is also accepted by Keynes, who appears to favour, amongst several alternative explanations, the idea of the reform of the coinage at this time being planned by Alfred, a process in which Alfred then persuaded Ceolwulf to participate (1998a: 17-18).

It is clear that this conclusion is at variance with the inferences from the strategic considerations discussed above, that London must have been in Viking hands in the two years from the Partition of Mercia in the summer of 877 to the summer of 879, when the Vikings were driven out of the London area to the east of the boundary set out in the Treaty. However, if viewed from a slightly different perspective, the evidence from the coinage offers remarkable support for this interpretation. It is significant that the Ceolwulf-alone phase of the coinage begins at around the time of the Partition of Mercia in 877. This could be taken as indicating that at this juncture the Vikings succeeded in denying to Alfred the political power and the economic interests he had exercised in Mercia and London up to that time, while assuming control of the management of the Mercian government and economy through their client king Ceolwulf alone. In this way the Vikings would seem to have acted as a 9th century mafia, creaming off the profits of the day-to-day economy for their own use, in this case without the awkward interference of Alfred. This scenario provides the essential background to an understanding of the motives behind Alfred's ultimately successful efforts to oust the Vikings from their position of parasitic control in Mercia and London in the next two years or so.

The liberation of Mercia and London

It is the purpose in what follows to give reasons for suggesting that there were two distinct but interconnected factors which between them provided both the means (the first factor) and the opportunity (the second factor) by which Alfred was able to liberate Mercia and London by causing the retreat of these armies away from Mercia in 879. The first is that in the period between Alfred's victory at Edington and the retreat of the Vikings in September 879 from Cirencester and Fulham (a period of some 15 months) the West Saxons constructed a system of fortresses around Wessex (including Oxford and Buckingham), which system is listed in the Burghal Hidage document. It is argued that these fortified places would have been part of a broadly two-pronged strategy: as an in-depth defence of Wessex, and as an offensive system of secure, garrisoned fortresses built in positions which acted as a counter-threat to the Viking armies poised at both Cirencester and Fulham. It can therefore be inferred that this strategy had as its aim the removal of the Vikings from both western Mercia and the London area. As well as posing a threat to the Viking armies' hold on Mercia, the rapid completion of this fortress system would have shown the Vikings quite clearly that they had little hope of achieving their goal of the conquest of Wessex.

The second factor, the success of which must have to a large extent been dependent on the effectiveness of the first, is the possibility that king Alfred, taking advantage of his former standing and close connections with Mercia, either initiated or sanctioned what could be interpreted as a coup d'etat in Mercia shortly before the Treaty was signed. This ousted or led to the assassination of king Ceolwulf of Mercia, who disappears from the historical record at this time (Dumville 1992: 7 n.37; Keynes 1998a: 13-14 & n.54). The alternative, that Ceolwulf could merely have died at this point without any drastic intervention, was a situation which in the circumstances decisively favoured Alfred's political and strategic intentions, and correspondingly disfavoured those of the Vikings. At this point Alfred must have swiftly filled the political vacuum (7). This can be argued as leading to the submission of the Mercians to Alfred, and the rise to power of Aethelred as Alfred's sub-regulus within Mercia (8).

The burghal system

It has been suggested by Patrick Wormald that the immediate result of Alfred's victory at Edington was the implementation of the "most sustained programme of military [and] administrative change in the West since Charlemagne" (1981:149) . Nicholas Brooks has discussed the "crash building programme" of fortresses and the organisation of their garrisons (1979: 17-20). Martin Biddle has argued for some time that the construction of the burghal system in Wessex can best be placed in the period between 880 and the 'capture' of London in 886 (Biddle 1971; 1973:251; 1975:273 & n.7, 310), a conclusion broadly accepted by Keynes and Lapidge (1983:23-5, 212 n.16, 340 n.6). It could however be observed that the placing of the construction of the fortresses in this period, after the Vikings had left Mercia and London altogether in late 879, would have been the military equivalent of bolting the stable door after the horse has left. It is argued here, however, that it was the successful implementation of this programme of fortress-building in the short period of 15 months or so between May 878 and August 879 which was the single most important and decisive factor which led to the Viking withdrawal from Mercia and London. It was the cause, rather than the consequence, of this development. While this suggested coup d'etat in Mercia may have been an adroit and perhaps opportunistic political move, the construction of the system of garrisoned fortresses, together with the reorganisation of the army, was the result of a well thought-out strategic plan.

The fortresses described in the Burghal Hidage show characteristics which have an important bearing on the question of Alfred's strategic thinking at this time. As will be argued in part 2 of this paper, it can be inferred from the logical order in which they were described (Hill 1969) that at the time of the composition of the Burghal Hidage these fortresses were regarded as a system. As such, not only would the military functions of each of the separate fortresses have complemented those of their neighbours, but all the fortresses in the system would have been the built to achieve a single goal - i.e. to implement a single overall strategic concept and to ensure a single outcome. It follows that this system was at the time of the composition of the document considered as being complete - i.e. that there were no other original components of the system, nor are any of the fortresses mentioned in the Burghal Hidage later additions. As a system it contrasts with the fortresses built by Edward the Elder in the E. Midlands in the early decades of the 10th century, and by Aethelred and Aethelflaed and latterly by king Edward in the West Midlands from the late 880s onwards. All of these comprise several piecemeal and non-contemporary series, the construction of which reflected quite different sets of military strategies. While the idea of such fortresses and the arrangements for their construction, garrisoning and upkeep were not new in either Mercia or Wessex (Biddle 1983:20-4; Brooks 1996b: 129), and had clear precedents in Francia in the 860s, there is no indication that the pre-Alfredian fortresses, at least in Wessex, had formed part of such a system. Since the creation of this system was thus entirely new (at least for Wessex), it must be inferred that the military objectives which brought the system into being were also new. It is therefore necessary to search for a suitable historical context for the implementation of this new strategy that was appropriate to the use of all the elements in the system.

Some insight into these strategies is given by an analysis of the positions of the northern line of the West Saxon fortresses in relation to the Viking forces. Just as the choice of both Cirencester and Fulham by the two Viking armies in the summer of 878 demonstrates their hostile intentions towards Wessex (discussed above), so does the siting of the fortresses on at least the northern side of Wessex demonstrate the implementation of a strategy to counter precisely this combined threat (Fig.000). Bath defended the common border of Wessex with the southern and south-western part of the Hwicce, and commanded the Fosse Way which led directly from Cirencester south-westwards into western Wessex. It also blocked access up the Avon to Viking ships (before the existence of Bristol) approaching from the Bristol Channel. The hilltop fortress at Malmesbury was also sited very near the border of Wessex with the Hwicce and the Fosse Way itself, and on the direct route (which the Viking army must have taken) between the battle site at Chippenham and Cirencester (Haslam 1984b: 111-117). Cricklade was sited on the Wessex side of the common border with the Hwicce at the crossing of the Thames by Roman Ermin Street, which led south-eastwards in a straight line from Gloucester through Cirencester into Wessex (Wainwright 1960; Haslam 1984b:106-110; Haslam forthcoming 2004). These three fortresses therefore effectively blocked access to the Viking army in both Cirencester and Gloucester by all the direct Roman roads into northern Wessex, and by sea from the Viking-dominated Bristol channel.

The positions of fortresses at Wallingford, Sashes and Eashing are equally significant in blocking access into eastern and central Wessex by the newly arrived army encamped at Fulham. The island site at Sashes with its bridge would have effectively controlled both the crossing of the Thames by the Roman road leading from St Albans (and Watling Street) to Silchester (Brooks 1964: 74-81; Astill 1984: 63), as well as passage up the Thames itself by the Viking army at Fulham. Wallingford was placed on or near the crossing of the Thames by the Icknield Way which leads from the north-east and Watling Street into central Wessex (Astill 1984:61; Peddie 1999: **), and its bridge would have complemented that at Sashes in functioning as a defence of the upper reaches of the Thames against passage by Viking ships. Similarly, the fortress at Eashing blocked navigation up the river Wey, a tributary of the Thames, which represented another good route into Wessex from the London area (Aldsworth & Hill 1971; O'Connell & Poulton 1984:43, 46). The context for the construction of this series of fortresses in these particular positions around northern Wessex becomes rather less appropriate for the time when Mercia was no longer a Viking client state after late 879.

There are also grounds for viewing the construction of the fortress at Southwark, together possibly with a defensive bridge over the Thames, as a key element in this strategy (Keene 2000: 144). Various writers have suggested that this fortress was built in 886 or soon after to complement the defensive function of the City and bridge, which was 'captured' in that year (e.g. Biddle and Hudson 1973: 23; Keynes and Lapidge 1983: 266 n.200; Keynes 1998a: 24, 26-7 n.119) (9). However, it could be argued that since it would have been a direct challenge to the Vikings both on the north bank of the Thames in London itself, as well as at Fulham, the period 878-9 would be a more appropriate context for its construction. It would have curtailed the freedom of movement of the Fulham Vikings along the Thames, and would have given Alfred strategic control over the lowest crossing-point of the river and the most important entry point into eastern Wessex from Viking-held territory in the north.

Southwark is the only fortress in the Burghal Hidage list which was used as an element in a system of two fortresses linked with a bridge over a river, a military device used frequently in the decades both before and after 879, in England and on the continent (Hassall & Hill 1970; Brooks 1979: 17-20; Haslam 1983; 1984a, 1987a, 1987b). In some cases in England the second fortress in the fort-bridge-fort system was only built after the primary fortress had been captured. Such was the situation in Hertford in 912, Bedford in 914, and Nottingham, captured in 918 with the second fortress built in 920 (ASC sa) (Haslam 1987b:). However, fortresses were also used offensively, in that their construction resulted on several occasions in the submission of the Viking armies in entrenched positions nearby. This process is seen in the construction of two fortress built in 914 during Edward the Elder's stay at Buckingham, which led directly to the submission of part of the army at Bedford and some of the people in the territory of the Northampton army. The whole army of Northampton submitted to Edward the Elder when Towcester was 'walled' in 917, and the whole of East Anglia and Cambridgeshire submitted to Edward when he occupied and restored Colchester in 917. And lastly, the submission of the army in the fortress at Stamford in 918 was apparently achieved by the building of an opposing fortress on its southern side. Although all these examples are later in date, in would be appropriate to the military realities of the time to see the construction and garrisoning of the fortress at Southwark in 878 or early 879, together with the rebuilding of London Bridge, as one of the proximate causes of the capitulation of the Viking military presence in the London area and the retreat of the Fulham army back to the continent in the summer of 879.

There are also reasons for regarding both Oxford and Buckingham, the two fortresses in the Burghal Hidage not in Wessex proper, as part of this initial system. Oxford, although on the Mercian side of the Thames, was strategically placed to control access along the important north-south routeway between Wessex and Mercia, which connected middle Saxon Hamwic on the south coast with Northampton and northern Mercia (Hassall 1987; Blair 1994: 87-92), as well as reinforcing Wallingford and Eashing in controlling access up the Thames. The evidence of the coins from Oxford, discussed by Blackburn (1998:111), shows that the mint at Oxford was producing coins at the same time as the celebratory issues produced by Alfred in both London and Gloucester soon after the Vikings left Mercia in late 879 (see below). This being so, Oxford is likely to have been 'refounded' by Alfred as a defended urban place in probably the spring or early summer of 879 as part of the fortress system described in the Burghal Hidage.

A fortress at Buckingham would also have played an important if not crucial role at this time. This would have been ideally sited to command the northern stretch of the Watling Street corridor (10). Its construction can therefore be seen as an important factor both in forcing the Vikings back from a boundary on the west side of Watling Street to the line defined in the Treaty well to its east, and in maintaining Alfred's hold over this crucial area once it had been gained. Since the presence of a West Saxon garrison in Buckingham in 914 was enough to cause the submission of part of the army in both Bedford and Northampton (ASC, sa) it could be inferred that the construction of a new fortress at Buckingham in 879 would also have acted as part of an offensive strategy to ensure the West Saxon domination of the northern sector of Watling Street which was taken into Alfred's hands in the Treaty, and against a possible Viking presence in Bedford itself, the only place mentioned in Alfred and Guthrum's Treaty. The fortresses at Southwark and Buckingham can therefore best be interpreted as essential elements in an offensive strategy that culminated in the retreat of the Vikings from Mercia and London reported in the Chronicle for 879. Their inclusion as elements in the system of fortresses described in the Burghal Hidage gives further support for the hypothesis of the origin of this system as a whole in the period 878-9.

It is furthermore of some significance that the combined hidage figures of the 8 fortresses placed around the northern border of Wessex in the Burghal Hidage - Bath, Malmesbury, Cricklade, Oxford, Wallingford, Buckingham, Sashes and Southwark - (totalling 12,500) is nearly as much as the combined hidage figures (totalling 15,671) of all the other 22 fortresses put together. It must be inferred that the border of Wessex with Mercia was, at the time of the construction of this system, the frontier line to which most of the manpower resources of the relevant shires were channelled. There would have been little point in committing these huge resources along a border between two areas owing allegiance after 879 to the same king, and against an adversary which had by this time effectively disappeared.

It can be inferred both from this evidence, and from the unequivocal evidence of the outcomes of these tactics (the retreat of the Viking armies away from Mercia and their abrogation of control of London), that Alfred set out to implement an overall strategy in which the individual fortresses were acting both on their own and as part of a coherent system which was designed from the outset to put the West Saxons on the offensive against the Vikings who controlled Mercia and the London area. The role of both Buckingham and Southwark in particular focuses attention on the offensive function of many of the other Wessex fortresses, especially those around its northern borders already discussed. A key element in Alfred's strategy lay in how the construction of these fortified and garrisoned strongholds in Wessex would have appeared to the Vikings. If Guthrum's army perceived their own occupation of the fortress of Cirencester as being an offensive act, as must be inferred from its position on the Roman road system in close proximity to the Wessex border, then the construction of the fortress at Cricklade, backed up by that at Malmesbury, must have appeared to Guthrum's army for what it clearly was - part of an offensive strategy directed against themselves. They cannot have been comfortable with the distinct possibility that they might be attacked at any time by the West Saxon forces garrisoned in a securely defended fortress - which their scouts cannot have failed to have been aware - only a few hours' march down the Roman road, and backed up by similar garrisons in neighbouring fortresses. In this way Sashes, Wallingford and Oxford would have complemented the offensive functions of Buckingham and Southwark.

An important aspect of this offensive role would have been to provide the means to guarantee the permanence of the intended military and strategic outcomes. Seen in this light, the context of the construction of the fortress at Buckingham is more appropriate to a time immediately before Guthrum's army retreated behind the boundary to the east of London and Watling Street, rather than after this had taken place. Similarly, the construction of the fortress at Southwark makes more sense in a military context before the Viking army at Fulham retreated to the continent in late 879, directly threatening its position upstream at Fulham and making sure that it could not reoccupy London. And the most appropriate context for the construction of the fortress at Cricklade would have been at a time when Guthrum's army was in occupation of Cirencester, just a little way up the Roman road, destabilising its presence there, preventing a direct foray into Wessex along Ermin Street, and creating a situation which would have made its return to either Cirencester or Gloucester impossible.

By constructing the system of garrisoned fortresses at this time, Alfred would therefore have achieved four distinct but interconnected strategic goals: a) the provision of an adequate protection for most of the population and territory of Wessex, and to counter Viking foraging parties or raids; b) the prevention of forseeable adverse military outcomes from taking place; c) the provision of offensive instruments designed to bring about required military objectives, and d) the establishment of the means to make certain that these outcomes, when achieved, would continue in force. The hypothesis of the construction of Alfred's burghal system in the short period before August 879, which was demonstrably designed to fulfil this four-part strategy, is quite distinct from the prevailing view which sees the initial system of fortresses described in the Burghal Hidage as being built in the 880s and perhaps into the 890s as a purely preventative measure "to protect the kingdom from the threat of further Viking attack" (Keynes and Lapidge 1983: 25, 212 n.16). As an explanation for their construction this is a necessary but not a sufficient condition, leaving the reasons which drove the Vikings from Mercia in the first place both unacknowledged and unexplained.

The logistics of the construction of the individual fortresses in the system give some support to the idea that all the fortresses in Wessex (including Oxford and Buckingham) could have been built within a relatively short timespan. In the case of Cricklade, where excavations on the defences over the last 50 years have produced some significant data (Radford 1973; Haslam 1984:106-10; forthcoming 2003), it can be estimated that all the elements in the defensive circuit could have been constructed by 1000 men working for about eight months. The calculations are as follows: given a wall length of 2280 yards [2083 metres] (the longest estimate), an average width of 6m, and an average (estimated) height of 2.5m, the bank would have comprised around 34,200 cubic metres of material. Excavation has determined that the bank was of simple dump construction with turf revetments, built from material derived from the three ditches and with turves stripped from the area of the bank and berms (Haslam 1992 & forthcoming 2003). On the premise that a team of 4 men could have built 2 cubic metres of bank in a day's work, the whole defensive circuit would have been completed in 68,400 working days, or by 1000 men working for 68.4 working days, or 13.6 weeks of 5 working days - say around 4 months. The defences also comprised an intra-mural walkway made of laid stones which would have had to have been quarried and brought some distance, and a possible simple timber revetment on top of the bank (Haslam 1992 & forthcoming 2003). If a similar amount of manpower was needed to have constructed these elements sequentially with the bank (in the order: 1-walkway, 2 - bank and ditches, 3 - palisade), this gives a total of eight months for a team of 1000 men to have constructed the basic defensive system (11).

The period available was some 15 months - from May 878 until August 879 - which would have given a good 11 or 12 months with reasonable weather. Even allowing for other works to have used the available human resources (building a bridge and causeway, laying out streets, constructing gateways and watchtowers, and general supportive provision for the workers), and allowing extra time for surveying and laying out the site, bad weather, general inefficiency and human wastage, the whole enterprise, if reasonably well organised, would therefore have been comfortably completed with the manpower resources available from the 1,400 hides appurtenant to Cricklade in the Burghal Hidage, at the rate of one man being conscripted from each hide. Cricklade was perhaps a special case in being the nearest fortress in Wessex to the Vikings in Cirencester. But, given the pressures on Wessex from the two Viking armies on its borders, it can reasonably be assumed that a similar degree of political will and local organisation would have been applied to the construction of all the other fortresses in the system. There is therefore good reason to accept the proposition that all the fortresses in Wessex as set out in the Burghal Hidage could have been in place as a functioning military system by the summer of 879.

The conclusion that the fortresses were able to be built in a comparatively short time-span is given support by the records of fortress-building in the Anglo-Saxon Chronicle, in particular those built by Edward the Elder in the period 912-20. While not all instances of fortress construction in the Chronicle give an indication of a time scale, it is clear that some of these were built in a matter of weeks, or at most a few months. In two instances (Towcester and Wigingamere in 917), the fortresses were completed in time to withstand sieges not more than three months later. In a third, the Danish fortress at Tempsford was completed in a similar time-span (917), sufficiently quickly for it to be considered a safe refuge for a Danish king and an earl with his two sons, even though the fortress was over-run by king Edward's forces. Maldon was also complete enough to withstand a siege in the autumn of 917 only a little more than a year after the start of its construction in the summer of 916. Two further examples of the speed of fortress construction are given in the Chronicle for the year 895. The Danes built a fortress on the river Lea in probably the autumn of 894, which was able to withstand a siege in the summer of 895. And in the later summer of that year Alfred started work on a double fortress linked with a bridge to prevent the Vikings from escaping with their boats down the Lea. This appeared to have had such an immediate impact that the Vikings abandoned their ships and fled to Bridgenorth before the winter. Lastly, Alfred's fortress at Athelney, completed by a "small force" in Easter 878, was in use as an apparently effective military stronghold only a few weeks later.

In other instances (Witham in 912, the two at and near Buckingham in late 913, Towcester [stone wall] in 917 and Bakewell in 920) the construction of these fortresses was clearly sufficiently far advanced after only a matter of six or possibly even four weeks to compel the submission to the king of earls, armies and/or the neighbouring populations - and in the case of Bakewell the Scots, the Northumbrians and the Strathclyde Welsh. The writer has argued elsewhere (Haslam 1997: 114-8) that the annals in the period 911-17 were written to establish a pattern - which there is no reason to doubt reflected a real strategy - of the submission of the Danes in the various conquered areas to king Edward in person as a result of the building of a fortress in that area. This does not necessarily mean that these fortresses were completed in all aspects before this happened. It does however show that the construction of their defences and the establishment of garrisons was sufficiently far advanced in the 4-6 week period in which Edward was recorded as having stayed in the area to have forced the desired outcome.

It can furthermore be argued that the period immediately after the battle of Edington was for political reasons the most appropriate time for the system of fortresses to have been constructed. It has been emphasised by several writers that the process of the creation of this system in the late 9th century, together with the establishment of the burghal territories and the arrangements for their garrisoning and settlement, would have been an essential component of the exercise of royal power, and therefore of the consolidation of the state. Nicholas Brooks has pointed out that "the development of royal authority in England was directly connected with the successful enforcement of public works and general military obligations so that an adequate defence against the Vikings was provided" (1971:84), a conclusion amplified in his later discussion of the physical and social effects which this military programme would have had on the population (1979: 17-20). Richard Abels has pointed out the processes by which royal interests in minting, marketing and the exercise of justice, as well as of defence, were concentrated within the fortresses as an essential means whereby royal power was centralised and underpinned (1988: 79-80 & passim; 1998: 208-9). After his defeat of the Vikings at Edington in May 878, the victorious Alfred was able perhaps for the first time to count on the undivided loyalty of the ealdormen and thegns in every shire in Wessex. The two fortresses in Wessex which had already proved their worth in action against the Vikings, namely Athelney (constructed by Alfred in early 878) and Countisbury, N.Devon (unsuccessfully stormed by the Vikings also in early 878), must have demonstrated to Alfred how effective such fortifications could be - just as he doubtless appreciated from his own experience how their absence in Wessex had facilitated the successes of the Vikings, occupying their fortress at Reading, in the earlier part of the decade. It would have been at this time that the population, faced with a common enemy on its borders and knowing how easily the Vikings had nearly overrun Wessex in the very recent past, would have been most prepared to submit to the enormous scale of conscription and coercion which such a programme would have entailed (Brooks 1979:17-20). Conversely, it would have been immeasurably harder at a local level to raise the motivation to undertake this massive project if there were no standing Viking armies anywhere near Wessex, which was the situation after late 879 (12).

This programme of fortress-building can also be seen as an essential prerequisite for the maintenance of Alfred's prestige and the consolidation of his power within Wessex after his victory at Edington, and, most importantly, for underpinning the power and standing which allowed him to assume the overlordship of Mercia, including London, soon after. Richard Abels has observed how the construction of the fortresses enhanced "the institutional power of the West Saxon Monarchy over its subjects", and how they reinforced and regulated "the traditional connection between landholding and the military obligation to the Crown" (1998:208). The hypothesis of the creation of the system at this time is therefore supported both by its appropriateness in relation to the internal political dynamics of the state, and by its necessity as a response to the perceived and real threat from the two Viking armies poised on the borders of Wessex. While the fortress-building programme was clearly a complex process rather than a single event, and whether or not the construction of all the fortresses had been entirely completed, these considerations suggest that by the middle or late summer of 879 its implementation was sufficiently far advanced to have had the effect of persuading two relatively large and potentially dangerous Viking armies to accept that they had no future in prosecuting any further hostilities against Wessex, and of forcing them to abandon London and leave Mercia altogether.

Alfred and Mercia

By the time the agreement between Alfred and Guthrum's Vikings was ratified and recorded in the extant Treaty, arguably in the summer of 879, Alfred must have been in a position of domination throughout Mercia. Ceolwulf had gone the way that ineffectual parties to unequal power struggles have tended to depart throughout history (13). Alfred could now - unlike any ruler before him when confronted with the Vikings - virtually dictate his own terms. This juncture marked the end of one era, which was dominated by the lack of any strategy to combat the unexpected Viking raids and massive exactions of wealth, and characterised by political uncertainty if not actual chaos. It saw the beginnings of a new era which saw both the Vikings essentially contained, when Alfred had achieved his strategic and political goals of the overlordship of the whole of Mercia and Wessex and the repossession of London - this time without having to accommodate the interests of a Viking client king in Mercia. It could be said therefore that by the end of 879 the 'kingdom of the Anglo-Saxons' was well on the way to becoming a reality.

This context provides a particularly appropriate explanation for the issue of the London Monogram coinage of Alfred (Blackburn 1998: 120-2) . This was a "celebratory issue struck intensively over a short period" (Archibald 1991: 286), and although traditionally associated with the reoccupation of London in 886, must now be placed rather earlier for numismatic reasons. The issue can now be dated soon after Ceolwulf's demise, when the Mercian mints reverted to Alfred, and marked an Alfredian monetary reform (Blackburn 1998: 120). Blackburn concludes that this type "does not celebrate the 'occupation' of London and the submission of all the English people to Alfred, but Alfred's assumption - or rather resumption - of control over the city from the Mercians after Ceolwulf's demise". The purpose of this issue "appears to have been to give publicity to Alfred's assumption of authority over Ceolwulf's former kingdom" (ibid.: 121,122). In view of the strategic considerations put forward above, this is certainly likely to have been so, and dates this issue to the last few months of 879 and the first few of 880. But the issues of similar coins from both Gloucester and Oxford at the same time shows that this reform had a deeper meaning. It suggests that together they marked and truly celebrated Alfred's defeat of the Vikings in Mercia and London in late 879 and his full assumption of power within Mercia, represented symbolically by the three 'capitals' of eastern, central and western Mercia. It must have been intended to be a clear and unequivocal signal to his subjects of his power and newly-won prestige - a piece of true propaganda. Perhaps it is at this point that Alfred did in fact receive the submission of all the English people, for he had at this juncture achieved his long-held strategic and political goal, and the people would certainly have had something to celebrate. It could be said that the issue of these series of coins was the most potent signal to the population as a whole for the inauguration of Keynes' 'kingdom of the Anglo-Saxons' - literally the beginning of a new era.

It is in the context of the extension of Alfred's lordship over the whole of Mercia that the rise of Aethelred must be viewed. His probable rise to power at this time (Keynes 1998a: 19ff) suggest that he was already a natural successor to Ceolwulf and the Mercian 'Crown', waiting in the wings for the removal of the Vikings. Keynes has demonstrated (ibid.) how he functioned as virtually a king of Mercia in his own right, yet almost always with acknowledgement of Alfred as his overlord. This could be interpreted as the behaviour of a legitimate aspirant to kingship with perhaps his own independent power base and/or family connections, who nevertheless had to accept that his own position and power could only be maintained under the overlordship of a more powerful ruler - especially as this powerful ruler was instrumental in bringing about the conditions which allowed him the exercise of power in the first place. Whether Aethelred was perceived as being a political rival to Alfred within Mercia, or whether he was entirely Alfred's protégé without any realistic claims of his own, his marriage in the mid 880s to Alfred's daughter (Keynes 1998a: 27) certainly cemented this subordinate relationship. This must have had the effect of containing the ambitions of Aethelred's circle (whether legitimate or not), and of strengthening Alfred's power and his hold over Mercia.

The events of 886, as recorded in the Anglo-Saxon Chronicle, can also be viewed afresh in the light of the strategic intentions of Alfred discussed above. It is possible that Aethelred may have assisted Alfred in what could be interpreted as a long see-saw struggle against the Vikings for possession of London in the period 880-886 (Keynes 1998a), so that Alfred's gift of London to Aethelred, recorded in 886, could be seen as a just reward for services rendered. However, many indications - in particular Alfred's economic and political interests in London prior to the Viking partition of Mercia in 877, his perception of London as his ultimate prize as shown by the strategies to regain the city from Viking domination, and the issue of the London Monogram coinage to celebrate its recapture - all indicate that possession of London was Alfred's long-held strategic goal, the potent symbol of his prowess, power, stature and triumphs. This being so, it is both too altruistic and too sentimental to assert, as many historians have tended to do (eg Stenton 1971: 259), that Alfred handed London back to a true Mercian in the spirit of benign generosity. Alfred had gained his prize and consolidated his power; if was giving anything away, there must have been a substantial 'quid' as part of the 'quid pro quo'.

It is suggested that some insight into this process is indicated by the circumstances in which Mercia became divided after Alfred's death. Aethelred and Aethelflaed maintained control of Mercia until the former's death in 911, when control of only London and Oxford and their dependent territories passed to Edward the Elder. The evidence already discussed above suggests that soon before late 879 - and indeed possibly rather earlier - Alfred took over direct control of the eastern part of Mercia (probably as a result of his deposition of Ceolwulf), thus giving himself the ability to negotiate directly with Guthrum, while allowing Aethelred, with his base in Gloucester, some control over western Mercia at probably a slightly later stage. It seems very likely therefore that in or before 886 - before his marriage to Aethelflaed - some of the economic benefits from London and Oxford and their dependent territories (ie present-day Oxfordshire, Buckinghamshire, Middlesex and much of Hertfordshire) would have been granted by Alfred to form Aethelred's personal estate for his lifetime. Similarly, some arrangement must have been made later by Alfred whereby Aethelflaed would maintain control of the rest of Mercia for her own lifetime after her husband's death. It can be suggested that in return - the 'quid' in the 'quid pro quo' - king Alfred required Aethelred to take direct responsibility for the defence of the eastern borders of Mercia with Viking East Anglia. That Aethelred had sole responsibility for these matters until his death seems to be indicated, firstly, by the role that he played in 893 in supplying prince Edward with reinforcements from London in the latter's engagement with the Vikings at Farnham, as reported by Aethelweard (Keynes & Lapidge 1983:190); and secondly, by the fact that it was only after Aethelred's death that Edward the Elder was able to take the offensive against the East Anglian Vikings by the building of a series of fortresses to the north and north-west of London, which was to culminate in the submission of the whole of Viking East Anglia to Edward in 917 (Haslam 1997) (14). The record of the events of 886, in which Alfred "entrusted the borough [of London] to the control of ealdorman Aethelred" (ASC sa), aptly describes the culmination of this process of transference of responsibility (though not of power) in the context of a general reaffirmation of allegiance of both the Mercians and the West Saxons to Alfred. This is perhaps underscored by the possible connection between this event and Aethelred's marriage (or at least betrothal) to his daughter, which would in practice have represented a contractual as well as a ceremonial consolidation of Aethelred's allegiance to Alfred.

Conclusion

This paper has sought to bring out explanatory connections between the different classes of evidence which provide the most credible historical context for all the pieces of the evidence as a whole, and to do so in a way which reflects the strategic realities of the time. It is salutary to reflect on the relative abundance and complementary character of these various pieces of evidence, and on the fact that if any of the parts of this jigsaw of evidence had been missing it would not have been possible to put forward the overall hypothesis set out above. It has been argued in this paper that the goal of Alfred's strategic planning was the removal of the Vikings from Mercia and London; that the Treaty between Alfred and Guthrum is the formal record of the agreement required for this outcome; that the Burghal Hidage and the evidence of the fortresses themselves is tangible evidence as to the means by which this outcome was achieved; and that the Anglo-Saxon Chronicle and the coinage together provide the evidence for the historical context which connects the means and the outcome.

The first fundamental conclusion of this paper is that Alfred was deprived of the considerable control he exercised in both Mercia and London prior to 877 as a result of the Partition of Mercia by the Vikings. The second is that the lifting of the continuing threat to Wessex by the retreat of the Vikings from Mercia and London in late 879, described in the Anglo-Saxon Chronicle, was neither a historical accident nor the inevitable outcome of earlier events or agreements. It was, rather, the direct result of military strategies thought out and implemented by King Alfred and his circle, which involved the construction and garrisoning of a system of fortresses over Wessex and eastern Mercia in the period between early 878 and late 879. This system is recorded in the arguably contemporary Burghal Hidage (discussed further in part 2). It is also argued that the agreement which regulated this retreat was documented in the contemporary Treaty between Alfred and Guthrum. This event - the recapture of London and the liberation of Mercia from Viking domination - also appears to have been marked by the celebratory issue of a new coinage from London, Oxford and Gloucester, the three 'capitals' of southern Mercia. Other processes and events, such as the displacement and/or death of Ceolwulf of Mercia, Alfred's assumption of the mantle of political power within Mercia and his relationship with Aethelred, were clearly important - perhaps vital - to the unfolding of events and the success of the overall strategy. From all this Alfred emerges (or rather has his character confirmed) as a determined, single-minded, energetic and ruthless operator, a clever political opportunist, a far-sighted strategist, and a king in the true Saxon mould in his concern to maintain and augment his economic resources, his military power and his political control over the territories he had come to rule.

APPENDIX

The following is a developmental model showing the suggested temporal and functional interconnections of significant events and processes in the years 876-886, as set out in the two parts of this paper:

· 877 (Aug) - Partition of Mercia -establishment of Viking-Mercian boundary to west of London, with loss of London and its region to Alfred. Ceolwulf functions in London as the Viking client king. Guthrum's army occupies Gloucester. Alfred's position and influence within Mercia is diminished if not entirely extinguished.

· 878 (Jan) - Guthrum attacks Alfred at Chippenham; Alfred retreats to Athelney marshes; some of N. Wessex dominated by Vikings.

· 878 - Alfred rallies and defeats Guthrum's army at Edington; submission of Guthrum to Alfred in the context of religious ceremonies at Aller and Wedmore (which end in early July); Guthrum's army stays in Chippenham; moves to Cirencester in mid-late Aug (before the harvest).

· 878 - from middle of year: Alfred plans and begins building a series of fortifications around Wessex; general mobilisation and conscription of population; organisation of fortress garrisons.

· 878 (summer) - New Viking army encamps at Fulham, directly supporting Guthrum's army in menacing Wessex.

· 879 (possibly early summer) - Alfred possibly stages coup d'etat in Mercia, leading to death of Ceolwulf (though possibly natural): Alfred now in control of central part of Mercia (later Oxfordshire & Bucks), and possibly also western Mercia (though Guthrum's army still in Cirencester). The emergence of the 'Kingdom of the West Saxons and other southern peoples' from the 'Kingom of the West saxons' (Keynes 2001: 61).

· 879 (summer) - system of fortresses in Wessex (with Oxford and Buckingham in eastern Mercia) completed, and garrisons installed; army system also reformed. Alfred is able to dictate terms to the Viking armies from a position of military and political advantage; stages conference with the Vikings - the terms as set out in Alfred and Guthrum's Treaty agreed.

· 879 (late summer) - Fulham Vikings retreat across the Channel; Guthrum's army retreats to East Anglia (early-mid September, after the harvest in Mercia).

· 879 (late) - Burghal Hidage document (which describes a functioning fortress system) is composed, and widely circulated to all the shires

· 879 (late summer) - Alfred gains control of whole of Mercia and regains London; issues celebratory coinage from London, Oxford and Gloucester; Aethelred (possibly waiting in the wings) is installed as sub-regulus under Alfred in the area of the Hwicce. Probable general submission of all Mercian people to Alfred. These events mark the emergence of the 'Kingdom of the Anglo-saxons'.

· 880-86 - London under military pressure from Vikings: possibly briefly regained by them in 883. Alfred (?and Aethelred) fight them off; probably pushing back A & G's boundary northwards and eastwards from London. Alfred gives the responsibility for the defence of E Mercia to Aethelred, in return for life-long control of London plus its 'territory' (modern Oxfordshire (including Oxford), Bucks, Middlesex and S Hertfordshire).

· 880s - Enlargement and extension of fortress system in Wessex; replacement of small temporary forts with larger urban fortified settlements, generally associated with defensible bridges, on new sites. The Appendix is added to the Burghal Hidage to facilitate the calculation of new hidage assessments and manpower requirements for this enlarged system in Wessex.

· 886 - general witnessing of Aethelred's responsibilities in context of new general submission of population of Mercia and Wessex to Alfred; further binding of Aethelred to Alfred by the marriage of the former to Alfred's daughter (with the betrothal possibly in the context of London celebrations, with the marriage slightly later).

ENDNOTES

1. The TS of this paper, in one or another of its various drafts, has been read by several people, including David Hinton, Nicholas Brooks, Simon Keynes and Richard Abels, for whose comments I am most grateful. In the two parts of this paper the writer uses the neutral term 'fortress', rather than 'burh', to describe the fortified sites of the period, following the comments on the matter by David Hill and Alex Rumble (1996: 3) - notwithstanding the pertinent comments of Martin Biddle that many of these places were not merely fortresses but fortified towns (1976: 124 & n.197).

return to text

2. The Viking army clearly remained in Chippenham for some weeks after the baptism ceremonies at Aller and Wedmore, which according to the timescale in the Chronicle's account had ended in early July. For a discussion of the chronology, see Whitelock 1977. See also in note 7 below.

return to text

3. See the detailed chronology suggested in the appendix below. The wording of the Treaty is somewhat equivocal, and the implications for this chronology not at all clear.

return to text

4. This view is however in contrast to that of Keynes, who has suggested that Alfred continued to hold London after 877, and did so continuously until 886 (Keynes 1993: 12-19, 21-4, 34-5), even though toying with the idea that the presence of the Viking army at Fulham might be taken to imply the hold over the London area by the Vikings. In direct contrast, Dyson argues that prior to 886 Alfred "had not previously been in occupation of the town in any sense", and that its 'occupation' "involved the replacement of the Danish presence by an English one" (1990:100-01). A variation of this scenario is espoused by Peddie (1999), who has argued that the Vikings held eastern Wessex and London until driven back behind the boundary set out in Alfred and Guthrum's Treaty in 886, which he dates to this time.

It is however one of the principle premises of this paper that after the 'partition' of Mercia in 877 Alfred was in no position to hold onto anything - let alone the major strategic and economic site in the south of England - but that by late 879 his position vis-à-vis the Vikings was completely reversed through the implementation of the tactics described in this paper.

return to text

5. In the discussion of the boundary of the Treaty the writer has pointed out (Ibid.: 121 & Fig. 5) the artificial nature of its northern half, which follows no natural features between the source of the Lea and the Ouse at Bedford. This implies that it was set out not so much to define a discrete area of land, but rather to give to Alfred control of the maximum length possible of the strategic corridor of Watling Street as it approached London. This should be clear from the map, fig. 000.

6. The existence of this boundary has been discussed more fully by the writer (Haslam 1997). The writer recognises today however that the inference of its existence as a political reality is rather more speculative than was perhaps originally thought.

return to text

return to text

7. The timing of this can be inferred from the likely sequence of subsequent events. Guthrum's army must have travelled to East Anglia at the latest by the end of October 879 (880 in the Anglo-Saxon Chronicle) in order to avoid too much autumnal bad weather, but after the harvest in Mercia. They must therefore have concluded the agreement recorded in the Treaty by the end of August. In order to finally pin them down, Alfred would have had to have put in place the construction of fortresses at Oxford and Buckingham (see further below). In view of the commanding presence of the Vikings in Mercia after the Partition, he is unlikely to have been able to do this before taking charge of at least central Mercia after Ceolwulf's death. The latter event must therefore have taken place in the spring of 879 - say March or April. This would have still given Alfred time to have substantially completed the rest of the system of fortresses in Wessex, which gave him the strategic leverage he needed to dictate terms to the Vikings.

return to text

8. Dumville regards it as a coincidence that Ceolwulf's reign "ended in the year in which Alfred's triumph had revolutionised the political situation" (ibid: 7, n.37). It must be said however that invoking coincidence is not an explanation. The alternative is, as here, to recognise cause and effect, and to acknowledge that it was as a direct result of Ceolwulf's death - whether assisted or not - that Alfred was able to turn the tide of events in Mercia to his advantage. Richard Abels has suggested (1998: 146) that "Alfred had taken advantage of Burgred's deposition to extend his control over London and parts of eastern Mercia, or even that Ceolwulf had allowed this as the price for Alfred's 'friendship'".

return to text

9. The existence of a fortress at Southwark and its associated bridge in the 9th century is however questioned by Dyson (1990: 110 n.57) on the premise that the composition of the Burghal Hidage document could well be earlier than the foundation of the fortresses it lists. He suggests that the uniqueness of the name in the document implies that it was a "paper expression ... earmarking a proposed burh [fortress] on a hitherto unnamed site whose future construction and maintenance was to be the responsibility of the men of Surrey." Reasons are given in the second part of this paper for rejecting the notion that the Burghal Hidage was merely a prescriptive planning brief; rather, it describes a system which was already in place and operational, and which therefore included Southwark.

return to text

10. In an earlier paper the writer pointed out that the construction by Edward the Elder of two fortresses when he stayed at Buckingham in 914 (one of them presumably at Buckingham itself - the other at Newport Pagnell) was one of the first actions of his campaign against the East Anglian Vikings after 911. It was argued that this was part of a strategy to consolidate control over the vital Watling Street corridor into London. Precisely the same strategy is shown by the initial construction of a fortress at Buckingham in 879, in consolidating Alfred's hold over the Watling Street corridor which he gained in the terms of his Treaty with Guthrum. An analysis of the Roman road network in the vicinity of Buckingham is given in Baines 1985.

return to text

11. This is a particularly generous estimate compared with Abels' estimate of 120,000 man-hours of labour - approximately 20,000 working days - for the construction of the rather larger fortress at Wallingford (Abels 1998: 206). On the assumption that 4 men could have built 4 cubic metres of bank in a day - more nearly the basis of Abels' estimate (personal communication) - the total time estimated above would have been halved.

return to text

12. The psychological, social and religious factors in the dynamics of the interplay of the perceived threat of Viking attack and the motivation for work on the public fortifications are emphasised in no uncertain terms by Asser himself (ch 91) - see Abels 1988: 75-8.

return to text

13. Keynes has remarked (1993a: 20) that the reference to Ceolwulf as a "foolish king's thegn" in the Anglo-Saxon Chronicle "could have arisen from a desire in Alfredian court circles to justify a turn of events which had ousted Ceolwulf from power". Ceolwulf was dismissed in death as he was in life.

return to text

14. This inference of Aethelred's responsibility for the defence of eastern Mercia from a base in London has also been made by Tony Dyson (1990: 102).

return to text