[Back] [Forward] [Contents] [Journal Homepage]


Period 2

Period 1 | Period 2 | Period 3 | Periods 4-5

Period 2A | Period 2B | Period 2C

Period 2A: The rear of the bank | The front of the bank | The historical context

The excavations in 1975 confirmed earlier conclusions that the stone wall fronting the bank had been inserted into it at a later stage, replacing part or (probably in many places) most of the turf revetment. It was typically about 1.4m (5ft) wide at the base, in places up to 1.8m (6ft) (not including the front plinth). A sample of the stones from the inner ditch filled during the destruction of the wall in Period 3A shows that the wall was composed of approximately equal proportions (by estimated volume) of stone from Coral Rag deposits, and slabs of Oolitic Limestone. The Coral Rag is from Blunsdon Hill, 5km to the south-east along Ermin Street, and the Oolitic Limestone is probably from the Middle Jurassic deposits which outcrop about 6km to the north-west along Ermin Street. It is clear from this evidence that Ermin Street was a major highway for the transport of goods in the late 9th and early 10th centuries.

The wall must have stood originally to a height of at least 2m to have been effective (as shown in the reconstruction, Figure 8). The total volume of stonework in the defensive wall, which was 2051m in length (see above) must have been at least 5620 cubic metres, and probably in the order of 7400 cu.m. On the assumption that one man could have built about 1 cubic metre of wall in a day, this would have required around 7500 man-days to build, or 75 men working for around 4 months. The extra labour of quarrying some 1500 tonnes of stone, carting it some 6km from the nearest quarries, squaring and facing some of the larger stones, procuring and mixing mortar, sand and gravel, as well as digging out the front of the bank, spreading the earth on the top of the bank and building the rear revetment wall, would have required an army of at least 20 times this number. The strengthening of the defences with the stone walls must therefore have occupied, at a minimum estimate, around 1500 people over a summer season - a figure perhaps significantly of similar magnitude to the number of hides given to Cricklade in the Burghal Hidage. The building of this wall must therefore have made as great a call on the local manpower, as well as on the regional organisation and administration, as was probably required to construct the original defences - an aspect commented on by Brooks (1979, 18-19) and by Haslam (forthcoming). See further comments on this in Part 1.

The historical context for this large-scale refortification requires further discussion. As has long been realised, stone walls of similar type to that added to the bank at Cricklade are found in a number of other fortresses in both Mercia and Wessex (Biddle 1976a, 135-7; Radford 1978, with a recent discussion in Alcock 1995). Of those in Wessex, besides Cricklade, there is archaeological evidence for added walls at Wareham, Wallingford, Lydford and Christchurch. In addition, there is evidence of refortification from a number of hilltop fortresses which belong to the early 11th century, of which South Cadbury is the most fully excavated (Alcock 1995, esp. discussion pp.154-7). In the Mercian examples - in particular Hereford and Oxford (archaeological evidence) and Towcester (documentary evidence) - the added stone walls are likely to date from the early 10th century. In western Mercia these can be attributed to the activities of Ealdorman Aethelred and/or Aethelflaed, and in the case of Towcester to Edward the Elder. In Wessex, on the other hand, the added walls are generally regarded, in particular by Radford, as being the work of Ethelred the Unready, built in response to the new Danish threats of the later 10th and early 11th centuries (RCHM 1959, 137; Radford 1970, 103; 1972, 106; 1978, 139, 141), and included with the example at South Cadbury. This conclusion is broadly accepted by Hinton (1977, 69-71) and for instance Wood (1981, 195) and Hodges (1982, 178-9), and in his discussion of the subject by Haslam (1984a, 109).

As Alcock has pointed out (1995, 155), however, there is no archaeological dating evidence for the construction of these walls which supports this conclusion. Furthermore, this historical model is contradicted by the detailed archaeological evidence from the 1975 excavations at Cricklade. It is suggested that this evidence is more consistent with a date for the construction of the wall at Cricklade (and by inference at the other four places mentioned above) in the early 10th rather than the early 11th century. This must be examined briefly before the alternative historical model can be discussed in general terms. The relevant evidence from Cricklade is provided by the deposits both on the back and the front of the bank. It has already been pointed out that many aspects of the structure of the defensive arrangements in period 2 (when the stone wall was added to the front of the bank) are inferred from the character of the destruction deposits of period 3, when the stone defences were arguably systematically slighted.

The rear of the bank

The existence of a rear revetment wall built part-way up the back of the bank (which nowhere survives intact) is inferred (sections 1 and 2) from the existence of a tumbled layer of large stones spread over and behind the back of the bank and over the laid stones of an intra-mural walkway or 'wall street'. These stones lie both on the walkway and within a thick earthy clay layer overlying it. These relationships are by no means unequivocal, but can best be interpreted by the hypothesis that the rear revetment wall was built soon after the construction of the intra-mural walkway (itself a primary feature), and that it began to collapse down the back of the bank onto the walkway when the latter (and the defensive system as a whole) fell into disuse, before being completely destroyed in period 3. In view of the inference that this rear wall was of unmortared dry-stone construction, this would be expected. Such a period of disuse of the defences of this and other fortresses (as for instance Hereford) can be most reasonably argued as occurring in the more stable conditions of the mid and later 10th century.

The front of the bank

The situation here is rather clearer. There are several considerations from which it can be inferred a) that the wall was inserted into the bank to replace the front turf revetment relatively soon after the latter was built; and b) that the destruction of the wall in period 3 (to which a date of c. 1016 can be assigned on historical grounds, argued below) followed a relatively long period in which the wall remained standing while weathering products accumulated both on the berm and in erosion hollows on the sides of the ditches. The basis for the former assertion lies in the fact of the survival of the structure of the front of the bank. The evidence from the excavations of both 1975 and earlier have combined to demonstrate that the wall was cut into the front of a turf-revetted bank which at the time was largely intact. Various writers have remarked on the impermanent nature of turf revetments (e.g. Biddle 1976a, 137), a situation which is well demonstrated from the archaeological evidence from Hereford. It is unlikely that the turf revetment at Cricklade - even if it were supported by a wooden palisade - would have survived the period of general neglect of the defences in the mid-later 10th century. The impermanent nature of such a revetment is demonstrated by the fact that it was necessary to replace at least once the considerably more substantial palisade newly erected around the whole town in the mid 12th century (period 4).

This conclusion is strengthened by the evidence obtained from a number of trenches excavated in 1975 of the existence of deposits on the inner berm (between the wall and the inner ditch) of material which by the time of the destruction of the wall in c. 1016 must have accumulated over an appreciable period as a result of weathering of surrounding features, which included the wall. The presence and nature of these deposits, which had a significant stone and mortar component, are not consistent with a process of weathering, decay and perhaps collapse of a timber palisade and turf revetment over a period from their probable construction in the late 9th century to the period of the assumed construction of the wall in the early 11th. Furthermore, in every instance where it was observed the inner lip of the inner ditch had been considerably eroded over what must have been an appreciable period after its excavation as part of the original defensive system. These erosion hollows were again filled with the same mortary deposits (as must have occurred with the ditch, before re-excavation in period 2C), which could only have come from weathering and erosion of the wall over a long period. That some of these lower deposits on the berm could well have come from the cleaning out of the ditches does not invalidate these arguments.

These observations are also possibly reflected in Wainwright's records of earlier trenches (Part 2). The two layers which are occasionally distinguishable on the berm could also be taken as reflecting a period of accumulation of weathering products from the wall, which was succeeded by the deposition of products of its destruction. This evidence is not very clear, but offers some support for the conclusions drawn from the 1975 excavations.

The stratigraphical evidence from the Cricklade excavations allows, therefore, a number of inferences, which are reflected in the three divisions of period 2.

  1. The massive stone wall was inserted into the front of the bank after an interval of no more than a few decades from the time of the initial construction of the bank (period 2A).
  2. A wall of dry stone construction was built on the upper part of the back slope of the bank, in all likelihood at the same time as the construction of the fronting wall.
  3. The front wall stood in position, and its weathering products accumulated on the inner berm and in the ditches, for some considerable time (period 2B) before it was possibly refurbished and reused as the main element in a new defensive system (period 2C), until totally destroyed in one operation in c. 1016 (period 3A).
  4. The ditches were exposed to weathering and erosion for some considerable time before being cleaned out - or perhaps rather re-excavated - (period 2C) only a very short time before being filled with stones from the destruction of the wall in period 3A.
  5. The rear wall began to collapse down the back of the bank during the period of neglect, until it was dismantled and thrown down the back of the bank in one operation (period 3A).
  6. At the time of the construction of front and rear walls the intra-mural walkway was still open and, by inference, being used for its initial purpose of access to all parts of the bank - even though it may be inferred that the rear revetment wall supported a wall walk at the top of the bank.

The systematic programme of refortification of the defences at Cricklade with a stone wall can therefore be inferred from the archaeological evidence as having taken place relatively soon after the initial construction of the turf-revetted bank, which there is no reason to doubt was constructed as part of the Alfredian scheme for the defence of Wessex in c. 880. The later the date assumed for the insertion of the walls, the less easily are the archaeological observations and inferences accommodated. It is suggested that similar inferences can be made with regard to the addition of stone walls to the earth banks of the other fortresses discussed in detail below.

The historical context

This conclusion calls into question both the accepted date for this phase of refortification in the early 11th century and the aptness of its supposed historical context, alluded to above. The historical 'evidence' for this later date is by analogy only. It relies mainly on the dating of new mints which were set up in newly fortified hilltop 'emergency burhs'. These were located at Castle Gotha, Cornwall (c. 997). Old Sarum, Wilts (1003), Beochore, Glos (c. 1007), South Cadbury, Somerset, and Cissbury, Sussex (both the latter two c. 1009-1010). The evidence for these fortresses has been discussed fully by Alcock (1995, 154-7). The new fortress at South Cadbury with its uncompleted church, its minting operations and its substantial (but uncompleted) defences may well have been envisaged as more than an emergency set-up (Alcock 1995, 168), but these examples do not, as has been all too easily supposed, add up to a systematic programme for national defence. They suggest, on the contrary, a piecemeal provision of locally secure centres in response to specific threats. This view is perhaps tempered by the evidence from Cricklade that the two inner ditches were cleaned out probably around this time (period 2C), which implies that its defences were put in order. It may also be put in perspective when it is realised that the bank and stone wall defences of South Cadbury were constructed on a comparatively large scale (two-fifths of the area of the largest in the Burghal Hidage - Alcock 1995, 168), the walls of which consisted of about 2000 cubic metres of stone which had to be brought from quarries some distance away. It is, however, difficult on either archaeological or historical grounds to see the royal policy for the provision of fortifications for the redefence of the realm in the late 10th-early 11th centuries as anything more than a succession of acts of expediency. It was not a unified policy or systematic strategy offering a successful and lasting solution, as was the case in Alfred's time.

Furthermore, the evidence linking the two systems - the refortification with stone walls of the Burghal Hidage fortresses of the late 9th century, and the hilltop fortresses of the early 11th century - is tenuous and circumstantial. The similarities in construction between the walls at Cricklade and South Cadbury, pointed out above, could as easily be explained by suggesting that that was the way walls were best built at any period, and cannot by themselves be taken as evidence of contemporaneity - a point also emphasised by Alcock (1995, 156). Nor can the rather circumstantial evidence from Shaftesbury, in the form of an inscription regarded as indicating a period of refortification in stone between 975 and 1050 (RCHM 1972, 57; Radford 1978, 150), be taken as evidence of a phase of the redefence of the fortress with stone walls in Ethelred's reign. The inscription, which is dated 880, is more likely be from a gateway built during the initial construction of the fortress (Keynes 1998, 38 and further refs, notes 41, 43; Haslam forthcoming). As with the other examples discussed above, Shaftesbury may well have been fortified with a stone wall in the early 10th century. But its defences were certainly considered to be inadequate in 1001, when the bones of Edward the Martyr, formerly resting in the abbey, were transferred to Bradford-on-Avon 'as a refuge from the inroads of the barbarians' (Gem 1978, 109-10; Haslam 1984a, 90-3). The example of Shaftesbury cannot therefore be taken as the evidence for the systematic programme of defence in the early 11th century it has been made out to be - though of course it may have been refortified, or its defences put in order, at some stage soon after 1001.

An alternative hypothesis is introduced by David Hinton, although he stresses at the same time the paucity of the dating evidence. He has suggested that the process of refortification could have been the result of a 'policy of a powerful ruler more concerned about his personal prestige', suggesting Athelstan as a possible instigator (1977, 69-71). I consider, however, that the needs of defence, and in the case of Cricklade the consolidation of the existing turf bank, would have been a far more pressing need than the expression of personal prestige. The enormous task of conscription of the local population to build the wall, argued above, is unlikely to have been suffered - let alone successfully organised - merely for royal aggrandisement, especially since similar walls are also found in other fortresses.

Ascribing the phase of the redefence of the lowland fortresses to the early 10th century does, however, provide a direct parallel to those in Mercia which were so fortified at this time (see above). The best parallel for the archaeological evidence from Cricklade is Hereford, where both front and rear stone walls have survived relatively intact, and where these features are dated to the early 10th century. The programme of the redefence of the larger Wessex fortresses with stone walls in the same period could also have formed the model for the replacement of timber palisades with stone walls by Edward the Elder at such places as Towcester (Anglo-Saxon Chronicle s.a. 917; Biddle 1976a, 137) and Oxford (Durham et al. 1983, 14-18).

The example of Towcester, however, is ambiguous and puzzling, but perhaps revealing in a rather different way than commentators tend to suggest. The Anglo-Saxon Chronicle for 917 states that '..before Easter King Edward ordered the borough of Towcester to be occupied and built'. Not more than three months later it was stormed unsuccessfully by several combined Viking armies. That autumn the king went to Passenham 'and stayed there while the borough of Towcester was provided with a stone wall'. But Towcester was a former Roman town whose stone wall defences, refurbished and provided with bastions in the 4th century (Woodfield 1992), are likely to have to have survived into the later Saxon period, and it must have been those which resisted the advances of the Viking armies in the summer of 917. It is not, therefore, appropriate to view this process as the replacement of an initial wooden palisade with stone, as does Radford (1978, 150), and cannot be taken as a direct comparison with the examples from Wessex. I would, however, suggest that the autumn works represent a repair of the Roman defences, which had been hastily requisitioned in the spring, one of the purposes of which was to bring to submission part of the army of Northampton - which it did. It must also be seen in the context of the political need to ensure the consolidation of the defence of Watling Street. In this case the statement of the Anglo-Saxon Chronicler that the borough was 'provided with a stone wall' is not so much a factual statement as a symbolic assertion of a situation which had a wider goal of the submission of the Viking armies (see further, Haslam 1997). In the same way the provision of the stone walls for the Wessex fortresses can be seen not only as practical replacements for decaying earth or possibly timber revetments, but also as the expression of the need to make both a strategic statement against the potentially still hostile Viking armies, and to reinforce and consolidate the power of the king.

Radford suggests (1978, 78) that the wall at Hereford was built as a reaction to the Danish raid of 914. However, the scale of the defences suggests that it was built as part of an overall policy for the defence of Mercia as a whole, rather than as a local reaction to one incident. The failure of the raid of 914 could anyway be taken as an indication of the strength of the defences, and suggests the probability that the phase of refurbishment in stone could rather belong to the period of Aethelred's active ealdormanship, between c. 880 and 900 (Shoesmith 1982, 73, 81-2).

It seems most likely therefore that the walls at Hereford, and possibly those at other places in Mercia, were the model for the Wessex examples. The independent dating of the stone walls at Cricklade to this period on stratigraphical grounds underlines the degree of co-operation, at least in military matters, between Edward the Elder and his sister and brother-in-law in western Mercia. It has been argued (Haslam forthcoming) that Aethelred was given responsibility by King Alfred in the 880s for the defence of London and eastern Mercia (which included Oxfordshire) against the Vikings, but that in the last decade of his life he did little to discharge these responsibilities. It was only after his death in 911 (Anglo-Saxon Chronicle, sa) that Edward the Elder was able to implement a coherent strategy to force the Danes in eastern England into submission, which he achieved in 918 (Haslam 1997). My view, therefore, is that the stone wall which was added to the front of the bank at Oxford is most likely to have been the work of Edward the Elder, soon after he gained control of the area on the death of Aethelred.

It can be suggested as a hypothesis that the immediate context for the refurbishment of the defences at Cricklade was the raid of the East Anglian Vikings into Mercia and Wessex in late 902, when they crossed the Thames at Cricklade (Anglo-Saxon Chronicle, sa 903). That the Viking army was able to raid into Mercia (presumably around north Oxfordshire) and Wessex without apparently much resistance from garrisons from Cricklade, Oxford or Buckingham, or elsewhere in Wessex or Mercia, implies that the arrangements for garrisoning the burhs put in place by King Alfred had possibly become somewhat disorganised by 902. It might indeed be wondered why an army from East Anglia should want to cross the Thames at Cricklade, when a more direct route into Wessex south of the Thames would have been along the Icknield Way to cross the Thames at Wallingford. It must be assumed that this army first raided in Oxfordshire and then moved southwards towards Cirencester and Cricklade. The fact of this raid, combined with the doubtless precarious state of the turf-revetted defences after 25 years of weathering, must have brought home to the king the necessity to put in order the defensive capability of the kingdom.

This fits with what is known about the more general policies of Edward the Elder who, in the first decade of his reign, was arguably concerned to consolidate the defensive provision of his kingdom (Fleming 1985; Dumville 1992). Although the previously stated view that Edward was responsible for the building of a whole new series of fortresses in the decade 900-910 (Haslam 1984b, 262-7) must now be superseded (Haslam forthcoming), it is nevertheless likely that this programme of consolidation included the refurbishment of the defences of existing fortresses in Wessex by the replacement of decaying turf and/or timber ramparts with stone walls, a general process which is attested by the archaeological evidence from Wallingford, Wareham, Christchurch and Lydford as well as Cricklade. As remarked above, by this process the king was also consolidating his own authority within his kingdom (Brooks 1971, 84; 1979, 17-20; Abels 1988, 79-80; 1998, 208-9; Haslam forthcoming).

Although I see the time marked by the accession of a new king as representing the most likely circumstance for the general refurbishment with stone walls of the perhaps decaying defences of the fortresses of Wessex, other possibilities cannot be ruled out. It has been argued (Haslam forthcoming) that the 880s were marked by a phase of consolidation of the defences of Wessex whereby small, probably ungarrisoned, forts of the Burghal Hidage scheme were replaced by new fortresses whose garrisons were provided, at least in part, by permanent inhabitants. For instance, Pilton was arguably replaced by Barnstaple, Halwell by Totnes, Eashing by Guildford, Chisbury by Marlborough, and the Roman site at Clausentum (Bitterne) by Southampton, amongst others. As with the construction of the Burghal Hidage system of fortresses itself, the scale of these changes must have involved a considerable commitment in time and labour on the part of the population as a whole to ensure the adequate defence of the kingdom. The addition of stone walls to existing defences could therefore have formed part of this ongoing and largely undocumented process. This could have happened before the renewed Viking attacks in the 890s (thus explaining why the Vikings largely avoided Wessex), or alternatively as a direct response to the dangers represented by this new phase of attacks. Just as the labour for the construction of the Burghal Hidage scheme in 878-9 was arguably motivated by the presence of the two Viking armies poised on the borders of Wessex in Cirencester and Fulham, so the effect of the Viking depradations in the 890s could well have galvanised the king, the reeves and the population to reorganise and strengthen the defences which had proved so effective in driving the Vikings from Mercia and the London area in late 879.

However, as Stenton remarks (1971, 269), the period from the first raids in 892 until 910 was one in which 'the Danish colonies in the north and east appear as the avowed enemies of the new state formed by the political association of Wessex and English Mercia'. Perhaps it would be fair, in the absence of any real evidence, to accept that any time in the 880s, 890s or early 900s would have been an appropriate context for the augmentation of the effectiveness of the original turf-revetted defences of Wessex by the addition of stone walls.

Period 2B

As has been argued above, one reason for placing the construction of the stone wall fronting the bank in the late 9th century is the presence on the berm of deposits which can only have accumulated over a relatively long period while the wall acted as a stabilising revetment to the bank. From the fact that the ditches were considerably eroded, and then also filled with these weathering products, it can be inferred that this period was one in which the defensive system as a whole was not actively maintained, and was indeed allowed to decay, for some considerable time.

This conclusion is reinforced by the evidence of the loss of the intra-mural walkway. In every place excavated a tumble of stones lay over the walkway and on the ground surface beyond it, as well as within a matrix of dirty clay (representing degradation products of the bank) overlying it. It can be inferred from these observations a), this is evidence of a rear revetment wall placed part-way up the back of the bank, as at Hereford; and b), that both the walkway and the suggested rear revetment wall had at some point ceased to function as elements in the defensive system, and were allowed to decay naturally over an appreciable period. On the premise that these processes of weathering and decay on the front and the back of the bank belong to the same period, it seems most reasonable to argue that this period of neglect can best be placed in the mid-late 10th century (from about AD 930-1000). The campaigns of Edward the Elder in the Midlands, and the success of his burghal policy, virtually removed the Viking threat to all England south of the Humber. During the resulting period of peace and economic consolidation the defences of the fortresses can have had little practical use - literally a casualty of their own success. David Hill has recently argued (Hill 2000) that King Athelstan, Edward the Elder's successor, was responsible for a phase of fortress construction in the 930s. But I have maintained (Haslam 1984b, 262-7) that Athelstan had little if anything to do with fortress-building. For later kings, especially Edgar, it could also be argued that there would have been little incentive to invest time and effort in maintaining the defences of what by then must have been regarded as an obsolete system - even though most of these same places were fast developing as minting, market and administrative centres. It is of interest that this period of neglect and decay can also be recognised from the archaeological evidence at Hereford (Shoesmith 1982, 82-3 and passim) (stage 4), and is attributable to the same causes.

Period 2C

In this period the existing defences were to some extent refurbished and otherwise put in order. It has been suggested above that in period 2B the ditches had become eroded and at least partly filled up, and must have been completely overgrown. The cleaning out - or perhaps more accurately the re-excavation - of the inner ditches can be placed in the years of renewed Viking threat in Ethelred's reign (in the years around or soon after AD 1000). The stone wall fronting the bank must have still been standing, and would probably have needed some repairs, although the stone wall at the rear of the bank had probably been in a state of partial collapse - to a degree that it no longer played much part in the functioning of the defensive system as a whole.

Period 1 | Period 2 | Period 3 | Periods 4-5

[Back] [Forward] [Contents] [Journal Homepage]

© Internet Archaeology URL: http://intarch.ac.uk/journal/issue14/1/period2.htm
Last updated: Mon Jul 7 2003