Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration 9781407310565, 9781407322506

One of the most tenacious and long-running controversies regarding the origin and development of the late Anglo-Saxon to

168 98 65MB

English Pages [151] Year 2012

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration
 9781407310565, 9781407322506

Table of contents :
Front Cover
Title Page
Copyright
TABLE OF CONTENTS
PREFACE
CHAPTER 1. HETEROGENEOUS TENURE IN ANGLO-SAXON TOWNS
CHAPTER 2. WILTSHIRE
CHAPTER 3. THE BURGHAL TERRITORIES IN WILTSHIRE
CHAPTER 4. HAMPSHIRE
CHAPTER 5. WARWICKSHIRE
CHAPTER 6. GLOUCESTERSHIRE
CHAPTER 7. WORCESTERSHIRE
CHAPTER 8. DISCUSSION – URBAN-RURAL CONNECTIONS AND THE DEVELOPMENT OF TERRITORIES
CHAPTER 9. URBAN-RURAL CONNECTIONS – FUNCTION AND ORIGINS: A MODEL
CHAPTER 10. OXFORDSHIRE, BUCKINGHAMSHIRE AND BERKSHIRE
CHAPTER 11. THE BURGHAL TERRITORIES OF OXFORD, WALLINGFORD, SASHES AND BUCKINGHAM
CHAPTER 12. DISCUSSION – THE DEVELOPMENT OF THE UPPER THAMES AREA, 9TH TO 11TH CENTURIES
CHAPTER 13. BEFORE AND AFTER THE BURHS OF THE BURGHAL HIDAGE
CHAPTER 14. THE FUNCTION OF THE BURGHAL SYSTEM OF THE BURGHAL HIDAGE
ACKNOWLEDGEMENTS
BIBLIOGRAPHY

Citation preview

BAR 571 2012 HASLAM

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

URBAN-RURAL CONNECTIONS IN DOMESDAY BOOK

B A R

Jeremy Haslam

BAR British Series 571 2012

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration Jeremy Haslam

BAR British Series 571 2012

Published in 2016 by BAR Publishing, Oxford BAR British Series 571 Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration © J Haslam and the Publisher 2012 The author's moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781407310565 paperback ISBN 9781407322506 e-format DOI https://doi.org/10.30861/9781407310565 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 2012. This present volume is published by BAR Publishing, 2016.

BAR PUBLISHING BAR titles are available from:

E MAIL P HONE F AX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Table of Contents Preface iii Chapter 1 Heterogeneous Tenure in Anglo-Saxon Towns 1 Chapter 2 Wiltshire 9 Chapter 3 The Burghal Territories in Wiltshire 19 Chapter 4 Hampshire 28 Chapter 5 Warwickshire 38 Chapter 6 Gloucestershire 43 Chapter 7 Worcestershire 51 Chapter 8 Discussion – Urban-rural Connections and the Development of Territories 60 Chapter 9 Urban-rural Connections – Function and Origins: a Model 70 Chapter 10 Oxfordshire, Buckinghamshire and Berkshire 82 Chapter 11 The Burghal Territories of Oxford, Wallingford, Sashes and Buckingham. 96 Chapter 12 Discussion – the Development of the Upper Thames Area, 9th to 11th Centuries 117 Chapter 13 Before and After the Burhs of the Burghal Hidage 123 Chapter 14 The Function of the Burghal System of the Burghal Hidage 133 Acknowledgements 138 Bibliography 139

i

Preface The terminology used in this study to describe rural manors and their tenements within the borough is that proposed by A Ballard (1904). Although possibly eccentric, this is nevertheless logical. The manors are ‘contributory’ or ‘contributing’ to the borough (they “give or furnish ... along with others ... to a common purpose” – Oxford English Dictionary) – even though it is the urban tenements which contribute their dues to the holders of the manors. The tenements at the centre are described as ‘appurtenant’ to the manors (“belonging to a property as of right” - OED). A property within a borough, whether attached to a manor or not, is referred to in this study by the neutral term ‘tenement’, rather than by any one of a number of possible contemporary Latin names such as haga or domus. This also includes the house occupied by a burgess. I have avoided the use of the term ‘messuage’, used in the Phillimore edition of Domesday, which is not a word in everyday use. The Latin term haga (plural hagae) is however used to refer to a comparatively large urban property held by a single person or religious institution which was often sub-divided into a number of tenements, sometimes with a church. I have also used, as a least-worst option, the contemporary term ‘burh’ to refer to a middle- or late-Saxon defended place (which may or may not have been ‘urban’ by the standard of the times), and a ‘borough’ (translating the Latin ‘burgus’) to refer to its Norman successor, notwithstanding the various caveats voiced by Susan Reynolds (Reynolds 1987, 295-300). For general discussions of the usage of the term ‘burh’, see Draper 2008, Draper 2011, Bassett 2011, 1-3, and Hall 2011, 600-2. Estates which were contributory to urban tenements are signified in the following tables as numbers in square brackets – [12] – which refer to their positions in the maps of each of the shires. References to Domesday are to the county, sections and sub-sections given in the various Phillimore editions; folio references, if required, can be worked out from these. The mention of shires by name in this study refers to those which existed at the time of Domesday.

iii

Chapter 1 Heterogeneous Tenure in Anglo-Saxon Towns

One of the most tenacious and long-running controversies regarding the origin and development of the late Anglo-Saxon town has been the nature and function of ‘heterogeneous tenure’, one of the defining characteristics of the Domesday borough. This refers to the basic division of the larger boroughs as described in Domesday Book into the customary burgesses or tenements which owed dues and obligations to the king alone, and the non-customary burgesses or tenements which were appurtenant to the various manors of tenants-in-chief of the shire (and sometimes neighbouring shires) to whom they paid rent and owed other dues and services (Maitland 1897, 17882; Roffe 2007, 110-1, 120). As David Roffe has made clear, although the occupants of these tenements were held in demesne by the various lords of the shire or district, they were of similar status to the king’s burgesses in 1086, holding their land in burgage tenure, and were as free as the customary burgesses of the king. This implies that they are not to be considered as merely part of the real estate of their lords (Roffe 2007, 123). Some of these tenements, indeed, comprised considerable sokes with distinct franchises which would have set them apart from the average burgage (Tait 1936, 43; Roffe 2007, 121-2). Heterogeneous tenure (though of perhaps a slightly different kind) is also characteristic of many of the larger pre-Conquest manors or multiple estate centres (Stephenson 1930, 184 & n.6). This is of some importance in the formulation of the explanatory hypothesis explored in this study.

The establishment of a complete list of Domesday boroughs which possessed heterogeneous tenure and which show evidence of having tenements appurtenant to rural manors is however somewhat problematical. In many cases such connections can only be inferred, in ways usefully discussed by Ballard (Ballard 1904, 24-31). Of the 62 places noted by Ballard with contributory lords or tenements/burgesses appurtenant to rural estates, all were probably burhs of late ninth or early tenth century origin (or earlier), with only 10 exceptions (Ballard 1904, 39-40; Roffe 2007, 114-5 table 4.1 includes rather fewer). These exceptions are Wimborne (Dorset), Sandwich (Kent), Bruton (Som), Milborne Port (Som), Milverton (Som), Calne (Wilts), Droitwich (Worcs), Tewkesbury (Gloucs), Gt Yarmouth (Norfolk), Dunwich (Suffolk) and Hythe (Kent). These were all either early royal trading places or the centres of early royal ‘multiple’ estates, or both. (The case of Hythe might well, however, be the exception which nevertheless proves the rule). In some cases the status of some of the places with such connections as late Saxon burhs is also a matter of inference, or has yet to be argued in detail – such as Bristol, Reading, Arundel, and possibly Dunwich - while the connections of tenements at Arundel with rural estates, for instance, are themselves a matter of inference rather than direct statement (Ballard 1904, 22 & map). These patterns of distribution, or sets of relationships, between urban tenements and rural manors have exercised the attention of every historian who has anything to say about the pre-Conquest and Domesday town for more than a century. As will be shown, a particular paradigm of interpretation has held the stage for almost as long. This study, however, presents a new interpretation which arguably provides a new set of explanations for this widelyobserved phenomenon. In order to do this it is necessary, in the words of Simon Keynes in relation to the study of the ‘Mercian Supremacy’ under Offa,

These arrangements had, however, virtually ceased to have any force by the later twelfth century (Turner 1990; Blair 1994, 157-8), but are evidenced in various charters from the late eighth century (the earliest in relation to Canterbury and Rochester – Tait 1936, 8-14), and in other places from the early tenth century. It is generally recognised therefore that the situation as it emerges in the folios of Domesday Book represents a system which was already in the process of fragmentation and decay. It is the purpose of this study to put forward evidence and arguments as to the ways in which these connections of urban tenements and non-customary burgesses with rural manors were important in the growth and development of Anglo-Saxon towns and burhs, possibly from the late eighth century onwards, and how this leads to a rather different perspective on some aspects of royal administration in the late Saxon period.

‘to determine the historical tradition transmitted to modern times, since in this way we can see how received tradition has influenced our own presumptions and preconceptions. Only then are we in a position to release ourselves from the accumulated weight of tradition, and from all that it

1

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

entails, and better able to work matters out afresh, as if from first principles.’ (Keynes 2005, 1).

of Maitland’s original hypothesis with a considerable (and valuable) range of detail – in particular using inferences about connections which derive from a comparison of the owners of properties in boroughs with those of rural estates, and in his tabulation of instances and numbers. He could well have elabourated on one of Maitland’s reservations about Tait’s own position by pointing out instances of the ‘distribution among divers rural manors of the burgages and burgesses that belong to one and the same lord’ (Maitland 1898, 210). He paid little attention, however, to some of the difficulties in Maitland’s original hypothesis or to any alternative interpretations of the evidence adduced in its support. In effect, he analysed this material as a creationist would look at order and design in nature to find evidence of the special creation of species – in other words, selecting material to illustrate and demonstrate an unquestioned and indeed self-evident hypothesis and ignoring evidence which contradicted it. Ballard subsequently discussed the thirteenth-century evidence of the wall-work performed by the burgesses of Malmesbury, which he adduced in support of the ‘garrison theory’ (Ballard 1906a). Although this paper was criticised in considerable detail by Mary Bateson, who argued that many of the connections proposed by Ballard originated in the twelfth century (Bateson 1906), the premise that these connections represent the survival of pre-Conquest arrangements is undoubtedly correct. Bateson’s interpretation cannot, as she points out herself, be applied to the ‘mural mansions’ of Oxford in Domesday. This is discussed further below (chapters 10 & 11). She, in her turn, introduced a red herring in considering that the burgesses attached to rural estates were living not in the boroughs but in the named estates (Bateson 1905; Bateson 1906).

The development of a paradigm In 1897 F W Maitland developed the so-called ‘garrison theory’ to explain these patterns of relationship of urban tenements appurtenant to the rural manors shown in Domesday Book. He argued that these connections were the vestiges of arrangements whereby the land-holders of the shire discharged their obligations for the defence of the borough for which the shire or burghal territory was responsible, a situation demonstrated in the attachment of territories to burhs in Wessex listed in the Burghal Hidage (Maitland 1897, 186-92). Already in 1896, however, he was beginning to explain the origins of tenurial heterogeneity in the older boroughs in terms of a military origin (Maitland 1896, 16-18), encapsulated in a passage which is worth quoting in full: What did the Anglo-Saxon thegn want with a town house ? He was not going to spend ‘the season’ there in order that he might take his wife and daughters to the county balls. Then, again, your ceorl who was ‘thriving to thegn right’ was expected to have a burh-geat-setl, and what is this but a house [with] in the gate of the burh ? Is it not a duty of burgward which obliges the thegns of the shire to have houses and dependents in the burh of the shire? (ibid., 17). This hypothesis, however, was soon roundly criticised by James Tait, who argued on the contrary that these connections originated as favourable ‘proprietary or jurisdictional grants of a profitable nature’ which were made as a result of the growth and trading activity of the borough as a town (Tait 1897). The importance he attached to this whole question can be gauged from the fact that he devoted half of his review article to Maitland’s hypothesis, which only comprised a very small part of Maitland’s book of 1897. Tait caricatures Maitland’s views on tenurial heterogeneity as ‘the shell of a dead military system’, and refers (somewhat anachronistically) to the burgesses attached to rural manors as ‘no peaceful traders, but warriors whose wants were supplied by the manors to which they belonged’. It is doubtful, however, whether Maitland would have agreed to either of these propositions, at least in the form expressed by Tait. Tait indeed cited Calne (Wilts) – discussed below - as an instance whose contributory manors could not be explained by the ‘garrison theory’.

In spite of Ballard’s work, Tait’s explanation has constituted an enduring paradigm which has been repeated, with variations, by all who have concerned themselves with the development of the Domesday borough and the AngloSaxon town – historians, geographers and archaeologists alike. This early emphasis on the commercial origin and function of these connections is shown in Mary Bateson’s comment on the references in the Domesday folios for Suffolk to Blythborough and Dunwich in her review of Ballard’s book: ‘Blythburgh made a large render of herring, and its river no doubt was navigable. A share in the herring trade had been obtained by some tenants of St. Etheldreda; to the Ely manor of Alneterne, in the hundred of Blithing, there appertained eighty burgesses in Dunwich. Mr. Ballard sees these eighty burgesses as primarily responsible for the ditch or wall of a borough at Dunwich, the manor of Alneterne being abnormally heavily taxed to this burden. We see them rather as enjoying some measure of commercial franchise, for which, according to its measure, they paid in service or in money. These particular burgesses were of advantage to the convent, because they provided

In spite of Tait’s contrary viewpoint, which does to some extent compromise the simple model put forward by Maitland, and notwithstanding Maitland’s own admission that he had said ‘too little of the borough as a tun and as the market and moot-stow of a shire’ (Maitland 1898, 210), Adolphus Ballard subsequently elabourated on the subject of the urban properties contributed by rural estates (Ballard 1904; Ballard 1906a). He came down decisively in support

2

Heterogeneous Tenure in Anglo-Saxon Towns

herring and had opportunities of enjoying borough market rights that made them valuable traders. For similar reasons Droitwich salt burgesses were a desirable acquisition to lands which fetched their salt from Droitwich.’ (Bateson 1905, 151)

were considered as appurtenant to the manor which they represented. This theory has the merit of coordinating the burh-bot of Anglo-Saxon law with the ‘tenurial heterogeneity’ displayed by the county boroughs in Domesday Book, but the objections to it are serious. In the first place it is only in regard to a small number of county towns that this definite connexion of urban and rural tenure is revealed by Domesday. Leicester itself is the one borough north of Welland in which this phenomenon occurs. In addition there is the fact that in 1086 these urban houses are clearly sources of profit to their lords, and an equally plausible explanation of the attribution of town houses to rural manors may be found in the assumption that the former were the abode of men whose place it was to supply the manors to which they were assigned with such articles of commerce as could only lawfully be bought and sold under the stringent conditions of witness and warranty which obtained within the burghal area.

In several of his introductions to the county Domesday folios, J H Round also came to share Tait’s views. On the evidence from Hereford, he remarks that ‘it is here [in a garrison town par excellence] that we should expect to find the garrison system in operation. And here the champions of that system [i.e. Ballard] emphatically fail to find it.’ With only three instances of urban-rural connections in the Herefordshire Domesday, there is ‘just enough to establish the absence of any general system of the kind. The evidence of this county is a reductio ad absurdam of the theory that the lords of the manors had to keep burgesses in the county town to discharge the duty of repairing the walls incumbent on their rural properties’ (Round 1908, 279). He has clearly forgotten momentarily that Domesday Book is not a complete log of practices and customs originating in earlier centuries, many of which may not have survived in the evidence recorded in a later era. In similar vein he has seen the lack of references to manors contributing to Guildford in Surrey as evidence which tells against Maitland’s garrison theory, and on the other hand seeing references to the connection of manors to Southwark as arising from the need for the tenants-in-chief to have houses near London (Round 1902, 286).

‘There is every reason to suppose that both AngloSaxon thegns and Anglo-Norman lords were sensible of the profit which would accrue to themselves if their men were to obtain the freedom of the borough market, and that the connexion here and there, as at Leicester, manifested between town and country property in Domesday has its origin rather in a desire for commercial advantage than in any rule of public law. In this connexion it is very significant that the borough was the seat of the county mint, and was, therefore, the centre of monetary exchange for the district; nor should we forget that in days when the county town was periodically thronged with visitors to the shire court, to which all freemen in theory owed suit and service, it was no small advantage to a lord to possess houses at which he himself and the men from the various manors of his fief might receive entertainment during the sessions of the assembly.’

F M Stenton (as ever) expressed a consensus viewpoint in his introduction to the Leicestershire Domesday in the VCH (Stenton 1907, 303), which may be usefully quoted in full: ‘The attribution of urban houses to rural manors, which has just been mentioned, is noteworthy because of its bearing on what has been called the ‘garrison theory’ of the borough. According to this theory every normal borough had originally been a place of defence for the county in which it was situated, and it is further assumed that the burden of manning these strong places, and of keeping them in repair, was laid upon the landowners of the shire. In the discharge of this duty the theory goes on to assert that each landowner was required to keep up in his county-town a number of houses, inhabited by men-at-arms, roughly proportional to the amount of land which he held in the shire, and that each house was considered as fulfilling this obligation with regard to some particular portion of his rural estate. Thus, when in the description of Leicester we read that Hugh de Grentemaisnil has nine houses in the borough which belong to Stockerston, we are by this theory required to understand that Hugh de Grentemaisnil and his predecessor Earl Ralf of Hereford, in virtue of their possession of this important vill, have been expected to maintain a definite number of men-at-arms in the countytown, and that the houses which they occupied

Further comments in 1908 by Charles Petit-Dutaillis on the subject of heterogeneous tenure in early boroughs appeared to add another voice in support of the general paradigm articulated by Tait, Round, Stenton and Bateson (Petit-Dutaillis 1908, 79-82). His views again supported the general view that the connections between rural estates and urban tenements were determined by the need for a place in the market town to facilitate trading, citing Mary Bateson’s example of the 80 burgesses from Ely’s manor of Alneterne having houses in Dunwich (above). His objections to Maitland’s views – that these connections crossed shire boundaries, that Domesday book does not record a complete list of the holders of rural manors with appurtenant tenements, that there is no proportionality between the sizes of manors and the number of burgesses contributed by them, and that these burgesses were anyway resident on the rural manors (echoing Bateson’s views) - are easily met. In the first place (to anticipate the discussions

3

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

below), the Domesday shires were not the same as the burghal territories of earlier periods; the Domesday record cannot be held up as a complete and systematic record of anything, let alone contributory manors; and the numbers of burgesses attached to a manor in a particular borough, and occupying a single haga, can be shown in may instances to have grown by a process of internal expansion in the long period before Domesday in ways which have nothing to do with the size of the parent manor.

country by the attachment of town houses to rural manors’ (ibid., 528), which harks back to Tait’s original view of the origins of these connections as appendages granted to estate-holders to facilitate trade. The more systematic descriptive input of H C Darby and his colleagues in their analysis of the Domesday geography of the different regions of England seems only to have strengthened this received view. Although the individual authors of the chapters in his Domesday Geographies do not in general comment on the questions of the origins of the connections between tenements in boroughs and rural estates which they so carefully document (and often map, though in isolation), nevertheless Darby himself inclines to an essentially Taitian viewpoint (Darby 1977, 309-13). He sees, for instance, the connections of manors in southern Oxfordshire with Wallingford as representing a commercial rather than a military function (ibid. 312). These same connections are analysed in quite different terms in chapters 10 and 11 below. He describes the urban-rural connections somewhat misleadingly as ‘urban fields’, or more appropriately as ‘urban spheres of influence’, which are seen to give some indication of the commercial hinterlands of towns in the eleventh century (ibid. 312). This view has been elabourated more recently for instance by John Blair, Terry Slater and in particular Grenville Astill (Blair 1990, 17; Blair 1994, 152; Slater 2000, 594; Astill 2006, 250-2). A variation on this explanatory theme is introduced by A H J Baines, who has suggested that both the non-customary as well as the customary burgesses at Buckingham were introduced by Aethelmaer – the ‘sheriff of Buckinghamshire in all but name’ - after 949 as an exercise in promoting the old burh of Buckingham as the centre of a kind of ‘enterprise zone’, which of course included his own lands (Baines 1985, 61). Needless to say, this particular and ingenious explanation will not have much relevance to the understanding of a phenomenon which is common to so many other places besides Buckingham itself. John Blair has, for instance, suggested that ‘there are strong grounds for thinking they [links between urban tenements and rural estates] have more to do with the economic circumstances of the mid-eleventh century than the defensive ones of the ninth’ (Blair 1994, 152) – a proposition which, I would argue by the evidence adduced in this study, is untenable.

More than 20 years later, in his introduction to the Oxfordshire Domesday, Stenton also alludes to the aspect of tenurial heterogeneity in which contributory properties from both Oxfordshire and Berkshire have appurtenant properties in Oxford and Wallingford (Stenton 1939, 3889), though he offers no explanation for what he must have recognised as an anomalous situation, with connections between tenements in both places and rural manors in both shires (a situation discussed in chapters 10 and 11 below). Stenton’s views were echoed by Carl Stephenson, who asserted that the urban-rural connections were ‘not the product of government ordinance, but usages that had grown up to suit the needs of seignurial management’ (Stephenson 1930, 184). By the early 1930s the subject seems to have been exhausted and its original impact forgotten. In 1930 J H Round, following an unpublished lecture given on the subject in 1912, observed that ‘...the [garrison] theory seemed . . . to have definitely lost favour’ (Round 1930, 252), and Stephenson in the same year concluded that ‘... if we re-examine that mooted question, perhaps it will be found not to deserve the prominence that it has hitherto enjoyed’ (Stephenson 1930, 181). In The Medieval English Borough of 1936 James Tait virtually dismisses the subject as of little importance, asserting that these connections were ‘rather a natural and very general, but not universal, result of burghal growth than the essential pre-requisite implied in the ‘garrison theory’ of Maitland and Ballard’ (Tait 1936, 64). The problem had been solved, and the tsunami generated by Maitland appeared to have subsided to mere ripples; the waters of the Domesday sea were now safe to navigate again. Subsequent comments by historians over the decades have only served to heighten this sense of monolithic calm. In the third edition of his Anglo-Saxon England F M Stenton once again expressed a consensus viewpoint in suggesting that, while the origin of the heterogeneous tenure characteristic of the Domesday borough might be able to be traced to the foundation of the borough by the king, the tenement in a borough attached to a ‘neighbouring’ rural manor had its origin in the need on the part of an estate holder for ‘a lodging when he came to the borough on business and with a place of refuge in time of trouble’, and that it was treated as ‘a profit-yielding appendage of the manor’ (Stenton 1971, 531). This merely repeats his views as stated in the first edition of 1943. He remarks that the pre-Conquest borough was ‘welded into the economy of the surrounding

Other authors have followed suit. The careful scrutiny of the evidence from Winchester has for instance led Martin Biddle to explain the notices of these connections in Domesday Book and charters as a way of providing the lord of the estate ‘with access for his produce to the largest market in the region, together with a town house that was necessary for the maintenance of his social position’, or, perhaps initially, as ‘the means by which the inhabitants of the county were guaranteed accommodation within the defences of Winchester in times of trouble’ (Biddle (ed.) 1976, 382-3) – somewhat sidestepping the issue as to who ‘guaranteed’ this facility and how this was achieved. Similarly, Geoffrey Martin has explained this heterogeneous tenure as arising from the need of manorial

4

Heterogeneous Tenure in Anglo-Saxon Towns

lords for ‘privileged access to markets, an additional stake in the community, and the simple convenience of accommodation near the shire court’ (Martin 1987, 60). David Roffe, while noting the importance of these connections in the composition of the Domesday borough, nevertheless inclines towards the received view, on the basis, firstly, that there were many different types of towns in the Anglo-Saxon period with different modes of origin; secondly, that the evidence has been altered perhaps out of all recognition by changes in tenurial structure both before and immediately after the Conquest; and thirdly, that many of the non-customary burgesses appear to have originated and developed in different ways and with different functions (Roffe 2007, 120-7). All these issues will be met in the discussions below.

over a long period before Domesday in response to what today would be loosely called ‘market forces’, to enable the holders of the rural manors to enjoy the economic and other benefits of a town house. The issues which this extended controversy have highlighted have not, however, disappeared – Brooks’ admonition is testimony to this – but go to the heart of any explanatory view about the origin of the late Saxon burh or borough, and raises issues which – as is hoped to show in these pages – cast a new light on the royal administrative mind-set and practices in the later Saxon period. The spatial patterns What has not been brought to bear on this issue in any systematic way, however, is any regard for the evidence of the spatial relationships of these connections either with each other, with the urban centre itself, or with shires and other historic land divisions and their boundaries in which they are set. As will be shown in this study, it is possible to make both temporal and therefore functional inferences from these spatial relationships which have an important bearing on any view of the controversy characterised above. Spatial language, let alone analysis, is seldom used by historians in the first half of the twentieth century; and where it is, with some sort of implied functional connotation, it is generally quite off-beam. F M Stenton, for instance, describes the relationships of contributory manors to the borough as ‘neighbouring’ (Stenton 1971, 531), without considering the ambiguities of this word in terms of the spatial patterning of these connections to particular centres rather than to others. Carl Stephenson remarks that ‘the basis of the connection was geographical rather than political, for it disregarded county lines whenever the borough lay close to one of them’ (Stephenson 1930, 183). Not only is this observation incorrect as a general statement – as is shown in many instances examined in this study; it also serves to highlight the fact that it is the distributions of these connections and their spatial relationships to the shire and other boundaries which form the evidential basis for coming to a diametrically opposite conclusion.

In a wide-ranging examination of the relationships between the ‘rural elites’ and their urban holdings at the time of Domesday and earlier Robin Fleming also follows the accepted paradigm in seeing the acquisition of urban tenements by these thegns and tenants-in-chief as arising from steps which they themselves took to avail themselves of opportunities to meet their needs, whether of power or profit, and how urban properties were subject to being bought and sold (Fleming 1993). Fleming’s treatment has been criticised from a different perspective by Richard Holt (Holt 2009, 69), whose novel views offered in explanation of the abundant pre-Conquest evidence of these urban-rural connections at Worcester are discussed in chapter 7. John Blair’s thorough analysis of these connections with Oxford also follows the prevailing paradigm, arguing that the pattern of urban-rural connections represents ‘a commercial catchment area rather administrative planning’, and that the development of the pattern of these connections shown in Domesday was the result of accretion within the context of an ‘evolving system’ (Blair 1994, 158). These views are commented on in greater detail in chapters 10-12 below. In 1996, exactly a century after the controversy was first aired, Nicholas Brooks felt able to say that Maitland’s original hypothesis, ‘exploded long ago by James Tait, should not be revived today’ (Brooks 1996b, 142). In spite of one or two lone voices (e.g. that of Julian Munby – Sturdy and Munby 1985, 50), Maitland’s ‘garrison theory’ has been well and truly consigned to the recycle bin of outmoded historical interpretations. Tait’s paradigm still holds sway.

A number of studies have analysed and mapped these connections for individual towns or on a shire basis – amongst which are those relating to Stamford (Roffe 1977), Leicester (Phythian-Adams 1986, 11), Winchester and Hampshire (Biddle (ed.) 1976, 382-5), Wallingford and Berkshire (Roffe 2009, 42-5), Oxfordshire (Jope 1956; Blair 1994, 117-9) and Warwickshire (Slater 1981, 30). The evidence pertaining to Worcester, relatively plentiful from pre-Conquest charters as well as Domesday Book, has been described in detail by several commentators (Hooke 1980, 39-40, 44, 48-9; Baker & Holt 1996, 136-40; Baker & Holt 2004, 261-7, 368-9; Holt 2009, 67-9). Richard Holt’s novel take on this evidence is discussed in chapter 7 below. All of these connections existing at the time of Domesday have been noted, and some mapped, in the various contributions on individual shires to H C Darby’s regional Domesday Geographies, though none of these groups is shown in

In what follows it is hoped to show that it is not quite time to press the ‘delete permanently’ button on this controversy. The divergence of viewpoint between Maitland and the followers of Tait has brought out a dichotomy in respect of both the function and origin of these connections which has never really been addressed. The Maitland camp (one or two lost souls huddled in respectful silence round his grave) would view these connections as being formed at the beginning of the process by which burhs and towns were formed to facilitate the implementation of defensive and other social and organisational functions of the burh. The Tait camp (by contrast an army or self-assured protagonists) would see these same connections as originating by degrees

5

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

relation to any other, with the exception of Somerset (Finn & Wheatley 1967, fig. 47, 198). A few have sought to describe these connections on a regional basis, in depicting the orbits of connections to different centres together on the same map – as for instance in the pioneering study of the Oxfordshire region by E M Jope in 1956 (followed by John Blair in 1994), the West Midlands by Terry Slater (Slater 1982, 176 fig. 8.1), and in the depiction of these relationships over most of England, albeit on a somewhat diminutive scale, by H C Darby (Darby 1977, 309-13). All of these treatments, however, have in the main been descriptive of the evidence in map form, rather than attempts to interpret these spatial relationships as evidence of past processes and therefore functions. A notable development in the use of this evidence is, however, David Roffe’s analysis of the patterns of these connections to help define the burghal territories of the late ninth century in the upper Thames region (Roffe 2009, 42-5). Roffe’s thesis is developed further in chapters 10-12 below.

of the primary mechanisms designed to create sustainable communities. One of the new functions of this arrangement (but by no means the only one) would have been to ensure the defence of the central burh and the protection of the burghal territory and /or the shire. As will be shown below, this arose from the obligations due to the king from booked land, although similar obligations would appear to have been due from folkland and from royal leanland. This being so, it follows that the attachments of the non-customary tenements in the burh or centre to rural manors would have been established by the prerogative of the king, as originator and upholder of the burh or estate centre, to serve the king’s military, social and economic agendas, rather than by the holder of the manor to serve his own. In this sense the areas from which these services were owed, and which would have included the orbits of these attachments, can be characterised as ‘territories of obligation’. This process was not, however, confined to the initial stages of the formation of burhs, but can also be seen as developing through successive periods of the reorganisation of burghal territories and later shires – especially in the W Midlands – around their administrative centres. Some of the implications of this model, and the mechanisms by which these connections were arguably established, are discussed further in chapters 8, 9 and 14.

It is the purpose in this study, therefore, to outline a preliminary model for the development of these ruralurban connections, based primarily on a reassessment of the evidence in Domesday Book and in earlier charters, where available, and the spatial relationships of the manors enumerated in it to their central boroughs, their neighbours, and to shire and other early boundaries, as well as to other features of the physical and historic landscape. This will be developed and tested by the analysis of evidence from several adjoining areas in central England – 1) Wiltshire (chapters 2 and 3); 2) Hampshire (chapter 4); 3) Warwickshire and south Staffordshire (chapter 5); 4) Gloucestershire (including the former Winchcombeshire) (chapter 6); 5) Worcestershire (chapter 7); and 6) Oxfordshire, Berkshire and Buckinghamshire (chapters 10-12). The overall model which derives from an analysis of these areas is that the urban-rural connections given in Domesday Book and shown in earlier charters are a function of the dominant position of the centres within political, administrative and military territorial units, as places to which services due to the king from these territories were rendered. These may be shires, burghal territories or early ‘multiple’ estates. As such, they can be interpreted in one sense, as Maitland originally suggested, as vestiges of arrangements established in the formative stages of the development of the boroughs and important estate centres as fortified burhs, or relating to other central administrative foci, invariably royal tuns - of which a burh is after all a specialised instance. In other ways, however, these connections can also be seen as arising from a wider range of marketing, legal, administrative and political functions of these places within a territory, the inhabitants (or more particularly landholders) of which owed services to the centre.

A precautionary note must be made, however, concerning some of the basic premises behind the methodology adopted in this study. In the first place, it should be reasonably self-evident that the Domesday record of the urban-rural connections, even when supplemented by earlier charter evidence and later sources, is by no means a complete list of such connections as were originally formed. The operation of such factors as the incompleteness of the seignurial returns to the Domesday commissioners, the idiosyncratic practices of the commissioners themselves (Fleming 1993, 5-7; Roffe 2007, 116-9), and the demonstrable loss of such original connections in the long period before Domesday through market and other forces, has meant that the surviving evidence is only a sample of an original tally. (Clearly, this factor was not taken into account in the comments of J H Round on the significance of urban-rural connections at Hereford and Guildford, quoted above). However, since such losses and gaps in the information will, from a broad statistical point of view, be unbiased in a spatial sense, it is seen as a valid procedure to draw inferences from, and to attempt to provide explanations for, the spatial distribution of these connections as have survived. In other words, the surviving evidence from many shires (but by no means all of them) comprises a meaningful sample which can be taken as being representative of an initial more complete distribution. Whether it is representative in each and every case and in every detail is ultimately a meaningless concept, because it is unknowable. But clearly, the more examples which have survived (and which can be reasonably inferred) in any shire, the more meaningful will be the patterns of distribution. An important corollary of this is that it is possible to make valid inferences from the evidence as we have it, but not from perceived gaps in this evidence. It is

With this in mind, it is argued further that these connections were put in place in the ninth and tenth centuries by the king and earl working in partnership with the tenants-in-chief (or the thegns or other land-holders) of the shire as one

6

Heterogeneous Tenure in Anglo-Saxon Towns

also taken as being a valid procedure that conclusions which pertain to the patterns of distributions of these connections in one area or shire can be applied to other areas where the pattern is perceived to be somewhat less complete.

above (Blair 1994, 158), and the case of a tenement in the suburbs of Winchester originally appurtenant to the royal manor of Basingstoke which was apparently annexed by Geoffrey the Chamberlain, ‘but neither the sheriff nor the Hundred have ever seen the King’s seal for it.’ (DB Ham 1,42). This same result could well have been caused by the separation of an urban tenement from its contributing estate through the operation of the land market in the nearly two centuries before Domesday.

A second major premise is that the connections to a particular place, in particular in the majority of cases which were burghal foundations, were established at the same time as the foundation (or re-foundation, or consolidation) of the burh in response to the new organisational initiatives on the part of the king which related to its establishment as a military and administrative centre. This is demonstrated by the ways in which the customary and non-customary tenements were intermingled in the layouts of at least four places which were burhs of late ninth-century origin – at Winchester, Gloucester, Oxford and Wallingford - discussed further in chapter 9. This basic premise is in direct contradistinction to the widespread assumption (for it is no more than this) that the urban-rural connections developed by accretion over the two centuries before Domesday, which is based on a very few recorded exceptions to this basic premise (e.g. Blair 1994, 158). In the case of the foundation of burhs, this would have required the focusing of a new set of obligations of the population within the whole of its burghal territory to its construction and ongoing maintenance as a sustainable institution. While the situation at the time of the formation of burghal institutions in the late ninth and early tenth centuries (and earlier) is largely unknowable, this premise is tested in every case examined below and – as will be shown - is explanatory of the evidence of the pattern of distribution of these contributory estates as it emerges in pre-Conquest charters and Domesday Book. Similar arguments are seen as being appropriate to explain the connections of rural estates with non-burghal centres named above, which are almost invariably the centres of royal ‘multiple’ estates of an earlier age (of which Calne and Droitwich, discussed in this study, are good examples). The social mechanisms by which these connections were established in the first place are discussed below.

It may of course be the case that an urban tenement acquired by a thegn by gift from the king – i.e. a formerly customary tenement – may have been drawn down into the rural estate held by that thegn. But this process would certainly not explain the thousands of such instances from more than 60 boroughs which show heterogeneous tenure. To assume otherwise would of course mean that no unifying explanatory model for the incidence of these connections as a widespread phenomenon at the time of Domesday is possible – though such a ‘model’ could be found in the common paradigm that these connections have grown in number over the nearly two centuries before the time of Domesday Book. Indeed, if Roffe’s reservations were to be couched in terms of a general explanatory model of the origins of heterogeneous tenure, then the substance of this would be that all non-customary tenements originated as gifts by the king from his primary stock of tenements within a burh to thegns at various times and for various reasons. But these hypothesised processes, which Roffe does not examine in detail, are contradicted by the very particular patterns of distribution of these connections on a shire-to-shire basis, which it is the purpose of this study to examine. It is arguable, indeed, that it is these few recorded exceptions which prove the rule. A related issue, raised by Tait, is that the process of commendation could have brought into being the connection of an urban tenement to a lord’s rural estate (Tait 1936, 89-92). This however raises the problem of the prior status of the urban tenement which was so transferred. If it was not held by a lord as part of his rural estate, it would have been held by the king as a customary tenement. It is difficult to imagine the transference of such tenements wholesale to other lords (in spite of Roffe’s reservations on this point mentioned above), which from the king’s point of view would have represented a general haemorrhaging of his assets.

The few exceptions to this, insofar as they can be demonstrated in the documentary evidence, do not in general undermine this basic premise. David Roffe’s view, for instance, is that many of the non-customary burgesses appear to have originated and developed in different ways and with different functions, of which an example is that from Guildford (DB Sur 1,1c and 1,1d) (Roffe 2007, 116, 120-7). Roffe suggests that some non-customary tenements may have been formed since 1066 by the ‘grant of some dues and the wholesale appropriation of many others’ (ibid., 120-1). This may well explain the many instances of the holding of tenements by various individuals which were not obviously tenants-in-chief, such as the long list in the Domesday folios covering Oxford (DB Ox B10), and in the list from Southampton, where the greater proportion of the individuals holding tenements in the borough were not holders of manors in the shire (DB Ham S3). Other instances include that quoted by John Blair, referred to

A more credible working basis for further examination must be that the customary and non-customary burgesses or tenements were created as separate classes with different origins. The underlying hypothesis which this study addresses and tests is that the non-customary tenements as a class originated at the same time as the customary tenements, and for similar (but distinct) reasons. One of the most significant evidential bases of this, mentioned above, lies in the fact that customary tenements paying landgable to the king and non-customary tenements paying dues to rural lords can be shown to have been intermingled in the layout of some of the larger burhs such as Winchester, Gloucester, Oxford and Wallingford, a factor which will be

7

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

discussed in more detail below (chapter 9). In particular, the detailed evidence from Winchester provides an unequivocal demonstration of the validity of this conclusion. This is discussed further in chapter 4.

Midlands (chapters 8 and 9), and in more detail in relation to Oxfordshire, Berkshire and Buckinghamshire (chapter 11). As will become apparent in the following analysis and synthesis of the evidence, the development of the administrative geography of especially the west and central Midlands, where this process is particularly apparent, cannot be understood without considering both how the burghal territories were formed by subdivision of the early regiones, how the initial burghal territories of the ninth century or earlier were divided up into smaller ones in the early tenth, and how the later shires were developed by their subsequent amalgamation and/or fission. A view will be taken in this study that the ‘landscape stratigraphy’ of these relationships, analysed in a horizontal rather than a vertical mode, is a key part of the evidence for the development and function of these political units through time. The patterns of distribution of the urban-rural connections, which were, arguably, formed within them and constrained by them, are a key part of the evidence which can be brought to bear on an understanding of their development. It is only when these developments are examined in detail that a view can be taken concerning the most appropriate historical contexts in which they took place, which in turn has implications for any views about the developing administrative machinery of the late Saxon ‘state’.

A third premise relates to the way in which the shires of Domesday were formed. Most historians have characterised this process, in particular in the west Midlands, as comprising the subdivision of the administrative units of the middle Saxon period (the regiones) to form the shires, each with a principal burh as its administrative centre, as an essentially linear development at some time between the early tenth century and the early eleventh (Stenton 1971, 292-3, 336-8, 502-6; Whybra 1990, 1-15; Gelling 1992, 140-2; Bassett 1996, 153; Hill 1996b, 94; Hill 2000, 174; Keynes 2001, 59). David Hill, for instance, states bluntly that at some point in time ‘Mercia was completely reorganised administratively, shifting from regions to shires’ (Hill 2001, 144). It is argued here that this model is both simplistic and misleading, in that it disregards the processes of the formation of burhs and the setting out of their dependent territories, which is well evidenced in the Burghal Hidage in Wessex, as constituting key developmental stages in this process. Margaret Gelling, for instance, does not appear to recognise burghal territories as distinct administrative units (ibid.). This question is discussed further below, both in relation to the west

8

Chapter 2 Wiltshire Malmesbury

Contributing estates are recorded in Domesday Book from four places in the shire: at Malmesbury, Cricklade, Wilton and Calne. These are tabulated below. Some of the several layers of inferences which have determined the making of this and other tables below are discussed in the section following the tables.

The various tenants-in-chief who were holders of tenements in the borough are given in the head section of Domesday Book, but the manors to which they are appurtenant are in general named only in their entries of their respective fiefs.

Table 1. Malmesbury, Wiltshire – Customary and non-customary tenements Estate / manor

M1 M2

No. of tenements in head section 51 in total ½

M3



8,11

M3



7,11

M4

2

Edward the Sheriff

24,20

M5

3

Ralph of Mortimer

41,1 41,9

M6



-

Durand of Gloucester

[30]

M7



1

William of Eu

[32] 27,10

M8

1

Somerford [1]

M9

1

[Smithcot] [7]

1

Castle Combe [8]

2

Humphrey de l’Isle

27,23

[Stanton St Quintin] – see below

-

Osbern Gifford

[48]

M10

1

Chedglow [9]

½

Alfred of Marlborough

26,19

M11

½

1

Geoffrey the Marshall

?

-

Tovi

Seagry [11]

1

Manor not given

1

[ ] – inferred. Numbers refer to Fig. 1

No. of tenements

Somerford [1]

1

Garsdon [2]

1

Kington Langley [3] Somerford [1] N Wraxall [4] Hullavington [5] Alderton [6]

1 1 2 1 1

Draycot Cerne [10]

Holder in head section in DB

DB section

The king Bishop of Bayeux Abbot of Malmesbury Abbot of Malmesbury Abbot of Glastonbury

[4] 8,3

Humphrey de l’Isle

DB head section

27,11

Humphrey de l’Isle

Customary tenements. held 4 manors in Wilts. Plus 9 cottagers who pay tax with the burgesses. (contr. to Malmesbury not specified).

Held 7 manors in Wilts. (Sheriff of Gloucs Notes). held 17 manors in Wilts. Contr. to Malmesbury not specified. Castle Combe shares app. tenements in Malmesbury & Wilton.

68,21

M12

½

-

M13



Drogo son of Poyntz

49,1

M14

½

Edric’s wife

25,4

M15

1

9

Comments

12 manors in Wilts.

Contr. to Malmesbury not specified, but probably that occupied by Edric’s wife. No manors in Wilts. Contr. to Malmesbury not specified. Can probably be identified with Draycott Cerne [10] above. See discussion.

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Estate / manor [ ] – inferred. Numbers refer to Fig. 1

No. of tenements

Holder in head section in DB

DB section

DB head section

No. of tenements in head section

Foxley [13]

1

Roger of Berkeley

45,1

M16

1

?

-

Arnulf of Hesdin

[25]

M16

1

Comments “of the king’s revenue” – ? received as forfeiture AH held 28 manors “of the king’s revenue” – ? received as forfeiture

1 (Azur’s former ? received as forfeiture holding) Wootton Bassett [14] 1 Miles Crispin 28,1 The following manors are stated as being contributory to wall tenements in a document of c.1283 from Malmesbury (Ballard 1906a; Bateson 1906; Ballard 1906b). Those appearing in the list above are starred **. The others are also shown in fig 1. Alderton ** Not in DB, but belonged to the fee of the king in Broadstone [15] ? 1283 (Ballard 1906a, 101). [Castle] Combe ** West Kington [16] 41,8 ?Kington St Michael Wraxall ** Draycott ** Foxleigh ** Hullavington ** Seagry ** Langley ** Somerford ** Abbot of 8,9 Charlton [17] Malmesbury [?32,5] Abbot of 8,7 Kemble [18] Malmesbury Wodhull (Woodhill) Bishop of Bayeux 4,2 [19] Church of Amesbury 16,6; In Selkley hundred – Winterbourne [Bassett] / unlikely to have been in [20] Humphrey de l’Isle 27,18 Malmesbury’s territory. Chintone in DB Kington Langley [21]

?

-

Bradenstoke [22] Stanton St Quintin [23] Bremhill [24]

The king

M17

Edward of Salisbury Osbern Giffard Abbot of Malmesbury

24,19 48,4 8,12

Table 2 Cricklade, Wiltshire – Customary and non-customary tenements Manor Aldbourne [1] Ramsbury [2] Badbury [3] Purton [4] Cricklade Chiseldon [5] Liddington [6] Lidyard Tregose [7] Clyffe Pypard [8] Earlscourt [9] Calcutt [10]

Number of burgesses / tenements 6 5 1 1 Holds “many burgesses” 6 1 7 3 1 garden 3

Holder of manor

Section in DB / source

The king Bishop of Salisbury Church of Glastonbury Church of Malmesbury Church of Cricklade – (held by Westminster) St Peter’s Abbey, Winchester Church of Shaftesbury Alfred of Marlborough Humphrey de l’Isle Stephen the Carpenter Odo of Winchester

10

1,10 3,3 7,6 8,13 9,1 10,5 12,5 26,7 27,9 66,6 67,1

Comments

Wiltshire

Manor

Number of burgesses / tenements

Holder of manor

Section in DB / source

Clyffe Pypard [8]

1

Wibert

68,24 Charter of 1008 (S 918) [3 holdings in 26,9; 29,6; 43,1 in DB]

Moredon [11]

1

-

Calne

45

The king

1,1

Calne

25

The church (held by the king)

1,1

Comments

No ref to appurtenant tenements in DB Possibly non-customary tenements appurtenant to Cricklade, and living there – see discussion. Also possibly noncustomary tenements at Cricklade

+’many’ others + 1 garden

Most of the connections of rural estates with urban tenements whose owners or lessees were liable to wall work in the document of c.1283 discussed by Ballard, can be traced back to connections evidenced in Domesday. It is likely therefore that the few connections stated in the document which cannot be traced back to Domesday do in fact represent – contrary to Bateson’s views - instances of early connections which are by inference pre-Conquest. These are included in the distribution of these connections in the section on Malmesbury, above, and in fig. 1.

accounts of rural manors alone, together with one preConquest charter (see table 2).

Cricklade

Calne

There is no account of Cricklade in Domesday, and references to appurtenant tenements come from the

All these are plotted in relation to the Domesday shire and other features, including neighbouring pre-Conquest boroughs and marketing centres, in fig. 1 (see table 4).

Wilton As with Cricklade, the lack in Domesday Book of any account of Wilton, the shire ‘capital’, means that its population of burgesses is known only from the mention of holdings appurtenant to several rural manors, both stated and inferred (see table 3).

Table 3. Wilton, Wiltshire – Customary and non-customary tenements Manor Netheravon [1] Salisbury [2]

Number of burgesses / tenements 5 7

Section in DB

Holder of manor The king (formerly Earl Harold) Bishop of Salisbury

1,18 3,4

Dinton [3]

2

Church of Shaftesbury

12,6

Stratford Tony [4] Fifield Bavant [5]

1 2

Earl Aubrey Alfred of Marlborough

23,9 26,14

Castle Combe [6]

1

Humphrey de l’Isle

27,23

Durnford [7] Sutton Mandeville [8] Sherrington [9] Marden [10] Odstock [11]

4

William of Eu

32,1

5

Robert son of Gilbert

40,1

1 1 1

Osbern Giffard Hugh son of Baldric Brictric

48,11 51,1 67,9

Wylye [12]

1

[Church of Wilton

13,11]

11

Comments

Centre not specified; Suggested as contributory to Warminster (VCH Wilts ii, 21). But Warminster was not a borough, and was itself contributory to Wilton.

Manor shares appurtenant tenements in Wilton & Malmesbury Centre not specified

Holding given in charter of 940 (S.469), but not appearing in DB entry.

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Manor

Number of burgesses / tenements

Section in DB

Holder of manor

Warminster [13]

30

The king

1,4

Tilshead [14]

66

The king

1,7

TOTAL -127

Comments Manor contributing non-customary burgesses, arguably living at Wilton. Manor contributing non-customary burgesses, arguably living at Wilton.

Table 4. Calne, Wiltshire – Customary and non-customary tenements Manor

Calne

Number of burgesses / tenements 45

Calne

25

Bishops Cannings [1] Calstone Wellington [2] Calstone Wellington [2]

1 1 2

Holder of manor

Section in DB

The king

1,1

The church (held by the king) Bishop of Salisbury Arnulf of Hesdin Richard Poynant

1,1

Comments Possibly non-customary tenements appurtenant to Malmesbury (or possibly Cricklade), and living there. Probably contributed by the church’s estate(s) to Calne

3,2 25,5 58,1

Discussion – the development of a model The overall model developed in this study is derived from a consideration of the relationship between the distribution of these contributory estates to each other, to the shire or other boundaries, and to other market or central-place settlements which functioned as such in the pre-Conquest and immediately post-Conquest period. Inferences from these spatial relationships can be set within a time frame from the way they relate to the known or inferred development and functions of these places through time. The patterns of relationship between the contributory manors and their centres shown in Wiltshire – even though demonstrably lacking in completeness - provide an important exemplar for a series of inferences which form the basis of this model. Inferences from similar spatial and temporal relationships in the other shires examined in this paper extend and reinforce this model.

1992), and although not technically a Domesday borough was nevertheless probably an urban place with origins as an minster market which subsequently expanded around an early Norman castle (Haslam forthcoming a). Cirencester, sited on the Roman Ermin Street leading straight from Cricklade, was also an important royal and minster site, a late-ninth century Viking stronghold and an incipient market centre in the later Saxon period (Slater 1976; Reece 1976; Gerrard 1994). Any hypothesis which sees the relationship of manors to their centres based on the proximity to the nearest or most convenient market or administrative centre, or as lying within an economically-determined ‘catchment area’, would appear to be confounded by the pattern of connections of manors in SE Gloucestershire to Gloucester, rather than to Cricklade or Malmesbury, and to the absence of any other connections to the north Wiltshire boroughs on the part of manors in South Gloucestershire (see fig. 3) – and of course vice-versa.

Cricklade and Malmesbury are situated on or near the northern borders of the shire, but their contributing manors lie wholly within the shire to the south. Similarly, the contributing manors of Wilton, in the south of the shire, are located around it and to its north within the shire. It is clear that the distribution of these contributory manors bears no relationship to their proximity to a particular Domesday borough or pre-Conquest market. This is particularly emphasised by the proximity of Cricklade and Malmesbury to Tetbury and Cirencester, both just over the shire border within Gloucestershire and the ancient kingdom of the Hwicce (fig. 3). Tetbury lies close to Malmesbury, and was an important minster market from the eighth century (Blair

The evidence from Wiltshire of the distribution of the contributing estates and their relationship to their centres and to others can only be explained by the hypothesis that this was a function of administrative determinants based on their position within the shire, and that the forces which led to the establishment of these connections had little to do with the attraction on the part of an estate-holder to the nearest ‘local’ market, or, in more general terms, to its role as a regional market centre. Malmesbury, Cricklade and Wilton were urban burhs created in Alfred’s scheme of the fortification of Wessex, the constituent places of which are listed in the Burghal Hidage (Biddle & Hill 1971; Hill & Rumble 1996; Brooks 2003; Haslam 2005).

12

Wiltshire

Fig. 1. Orbits of connection of manors contributory to Wilton, Malmesbury and Cricklade, also showing the manors of Wilton Abbey, and the hundreds (named in the list below).

13

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

The hundreds of Wiltshire 1. Chedglow 2. Cricklade 3. Staple 4. Shippen (Scipe) 5. Highworth 6. Dunlow 7. Startley 8. Blackgrove 9. Thornhill 10. Thorngrove 11. Chippenham 12. Kingsbridge 13. Calne 14. Selkley

15. Ramsbury 16. Bradford 17. Melksham 18. Cannings 19. Whorwellsdown 20. Rowborough 21. Studfold 22. Swanborough 23. Elstub 24. Kinwardstone 25. Westbury 26. Warminster 27. Heytesbury 28. Dole

29. Amesbury 30. Mere 31. Dunworth 32. Cadworth 33. Branch 34. Underditch 35. Alderbury 36. Stanford (Chalke0 37. Cawdon 38. Downton 39. Frustfiled 40. Damerham

century was the royal ‘multiple’ estate centre at Bedwyn, adjacent to which was added a new temporary or nonurban burh within the neighbouring Iron-Age hillfort at Chisbury, which is listed in the Burghal Hidage (fig. 2) (Haslam 1984a, 94-102). The extent of the estate centred on Bedwyn is discussed by Simon Draper (Draper 2006, 62-3). This would have created a poly-focal centre at the period of formation of King Alfred’s burghal system, which I have argued can be placed in the period 878-9. I have also adduced evidence that the defensive functions of the non-urban Burghal Hidage fort at Chisbury was replaced by the creation of a new defended burh and market at Marlborough, most probably at a date later in the ninth century (stage 3 in the timeline discussed in chapters 8 and 12; Haslam 2009, 99-100, 103-4). Its burghal status is suggested by the fact that at Domesday the earl held the third penny of Marlborough (DB Wilts B4), a defining characteristic of burghal development, as well as by its topographical characteristics (Haslam 1984a, 94-102) – though this is ignored in the most recent survey of the town (EUS Marlborough 2004). While the earl’s third penny was also derived from the profits of the pleas of the shire, Marlborough is included in a list of other boroughs in Domesday Book (DB Wilts B4). The formation of a burh at Marlborough is also introduced in the alternative model proposed by John Baker and Stuart Brookes of the development of the Wiltshire burhs (Baker and Brookes 2011, 108-10), though the underlying premises behind this are for various reasons arguably unsustainable (see further discussion below, chapter 3).

Other landscape and settlement aspects of these places are discussed elsewhere (Haslam 1984; Draper 2006). Simon Draper has also discussed the meaning of ‘burh’ in a late Saxon context (Draper 2008). That these burhs were organised at a shire level carries the further implication that the connections between contributing manors and these places have resulted from the function of these places as burhs (Hinton 1996; Brooks 1996b). As can be seen from fig. 1, the distribution of the manors connected to particular burhs reflects their position within the burghal territories of these burhs, insofar as they can be reconstructed. This reconstruction is based on a ‘best fit’ estimate of how the hidages of the Burghal Hidage can be made to fit the hundreds of Domesday, as worked out by J & C Thorn (Thorn & Thorn 1975; Thorn 1989; Thorn 2009), combined with an assessment of the local geography and topography. The rationale behind this reconstruction is examined in detail below (chapter 3). A logical inference of this – which is tested further below – is that these connections reflect arrangements put in place at the time of the formation of these burhs as new defended communities set up by the king. This conclusion is further supported by the evidence for the attachment of an urban tenement (vicus) in Canterbury to a rural estate in a charter of 786 (S.125; Tait 1936, 9-10), by the possession of a tenement in Canterbury by the nuns of Lyminge in a charter of 811 (S.160; Tait 1936, 15; Whitelock 1979, 514), and that of 904 in relation to Worcester (S.1280; Baker and Holt 2004, 174-6), which demonstrate that the connection between rural estates and the urban tenements in each place was a feature of the organisational landscape of these places at an early stage in their development as burghal institutions.

The southernmost estates contributing to Cricklade are however considerably nearer to Marlborough than to Cricklade. One of these, contributing 5 burgesses to Cricklade, was Ramsbury, which became the seat of a bishopric from 909, and comprised a 90-hide manor (DB Wilts 3,3) which as part of the royal multiple estate of Bedwyn is likely to have been given by the king to the new bishop on the occasion of the foundation of the see. Aldbourne was also a royal manor (DB Wilts 1,10), and also probably part of this complex of royal holdings in

This preliminary hypothesis can be refined by reference to the spatial relationship of these estates holding properties in Cricklade to other sites in north Wiltshire which are pre-Conquest fortified sites and/or markets or boroughs. I have suggested that the primary organisational focus in the Kennet valley area in the east-central part of the shire (which lay to the south of Cricklade) in the ninth

14

Wiltshire

Fig. 2. Area of Kennet valley, East Wiltshire, showing the relationship of Bedwyn, Chisbury and Ramsbury to Marlborough and the Kennet valley; also showing parish boundaries, Wansdyke and Roman roads (dashed). (From Haslam 1984a, 95). North to top.

E Wiltshire. If the holders of these estates or their subtenants – in particular Aldbourne and Ramsbury - had had a choice, they are likely to have opted for having a tenement or tenements in the new fortified market of Marlborough. These relationships provide a crucial body of evidence for the hypothesis above, to the effect that the observed pattern has been determined not by proximity to a local market even the new fortified one at Marlborough within the shire - but by administrative factors which were a function of the role of Cricklade as a fortified and garrisoned burh of the first stage of King Alfred’s scheme for the defence of Wessex and eastern Mercia. That estates very near Marlborough held tenements in Cricklade rather than within the former’s defences also carries the implication that this pattern was

formed before the secondary burh at Marlborough was set up, arguably in the 890s (see above; Haslam 2009, 103-4) when it replaced Chisbury, and therefore most probably at the time of the creation of the burh at Cricklade (in 8789). In other words, the establishment of Marlborough as a putative burh in probably the 890s appears to have been an intrusion into a landscape in which the territorial obligations of estates within the burghal territories of Chisbury and Cricklade had already been formed. It is, furthermore, no coincidence that Aldbourne was a royal estate, and Ramsbury had been one before 909. As well as generating the inferences already made, the connection of Ramsbury with Cricklade in particular

15

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Axbridge respectively, both of which were burhs listed in the Burghal Hidage (DB Som 1,1 and 1,2). These are, in effect, burgesses attached to the burhs which were appurtenant to the parent royal estate centre, and who could be seen as being contributed by the single royal estate to their respective burhs. In these instances the distinction between ‘customary’ and ‘non-customary’ burgesses breaks down, because they are indeed both – held by the king within the burh, yet owing their dues and services to the king at the rural centre. Similarly, a case can be made for the same relationships shown by the 30 burgesses listed under Warminster and the 66 burgesses listed under Tilshead, which can best be interpreted as being contributed by these royal estate centres to the burh at Wilton. This pattern would appear to be repeated at Calne, where 45 burgesses were held by the king, but possibly living ‘at’ the burh at either Malmesbury or Cricklade, in whose territory it is most likely to have lain (see fig. 1), although it is equally possible that these were in fact more akin to customary burgesses living at Calne itself. The 25 burgesses held there by the church, and thence of the king, could possibly constitute a group distinguished by being contributed to the centre at Calne by the estate or estates which formed the landed endowment of the minster church at the centre, and thus akin to non-customary burgesses. This relationship was also possibly true of the ‘many burgesses’ held by the church at Cricklade, at the time of DB held by Westminster, but formerly of the king. Again, a similar pattern is shown by the 24 burgesses ‘at’ Gloucester held by the royal manor of Kings Barton, and the considerable number of burgesses ‘at’ Bristol held by the royal manor of Barton Regis (below). In all these cases the distinction between customary and non-customary burgesses becomes blurred, and in the last resort somewhat meaningless.

suggests that this link had been put in place by the time the estate was given to the bishop in c.909, and that this date marks the time when this estate was converted from royal loanland to become booked to the bishop of Ramsbury. It also suggests that the connections of these two places with Cricklade have survived because the development of these estates, and the continuation of the holding of Ramsbury by the bishop, had not been subject to the sort of market pressures which appear to have distorted the survival, and therefore the distribution, of these connections in other areas. This is as close a demonstration as it is possible to get with the available evidence for the direct support for the model put forward here – that, as a general pattern, the relationship of the tenements in a burh or borough to the rural estates to which they were attached was established at the time that the burh was set up. An important element in this model is the premise that all or most of the connections between rural manors and their burghal centres were established as a group at the same time and as part of the same process. A primary element in the model advanced here is that these contributory estates were drawn exclusively from the area which was created as the burghal territory of Cricklade, and before they would have been included within the burghal territory of Marlborough which would, in all probability, have been carved out of it. This series of inferences, derived from the evidence of ‘landscape stratigraphy’, is a fundamental aspect of the overall model put forward in this study, and is replicated and supported in other examples discussed in later chapters. This set of inferences also provides an explanation of the suggested connection of the 25 burgesses ‘at’ the royal manor of Bedwyn to the new burh at Marlborough (see fig. 2). The connection of Ramsbury, situated to the north of the valley of the river Kennet, with Cricklade to the north implies that the original burghal territory of Cricklade extended to the south of the river Kennet to encompass all of Ramsbury’s 90-hide estate, and indeed comprised most if not all of the north-east corner of Wiltshire (see fig. 3). The position of Marlborough, also on the north of the Kennet valley, suggests the likelihood that the new burghal territory of Marlborough would have comprised land which had initially formed the burghal territory of Chisbury (the primary non-urban fort adjacent to Bedwyn, which was replaced by Marlborough) to the south of the Kennet, together with an area around it to the north of the river which had formed part of the territory of Cricklade (fig. 2). It seems possible that it would have been at the time of the creation of the new burh at Marlborough that the royal estate centre of Bedwyn would have been allocated one or several tenements within the new burh, of which the 35 burgesses recorded at Domesday were the occupants.

Similar patterns are shown, for instance, in the case of Tewkesbury, Gloucestershire, and Pershore (Worcs), discussed below (chapters 6 and 7), which contributed burgesses to Gloucester and Droitwich respectively, as well as having some of their own (as royal ‘customary’ burgesses held at the estate centre by the king); and by two instances (discussed below) where burgesses ‘at’ Droitwich and Tamworth owed reaping and other services to the nearby royal manors of Wychbold and Drayton Bassett respectively. The same is true of the inhabitants of the 118 tenements ‘at’ the borough of Steyning, Sussex, who ‘worked at the court like villeins’ (DB Sus 5,2) before the Conquest. It might only be an accident of non-survival that burgesses are not mentioned as being ‘at’ the former royal multiple estate of Deerhurst, which contributed burgesses to Gloucester and held a sizeable haga there (discussed below, chapter 6). The same situation is shown by Milborne Port, Somerset, outside the sample area, which arguably had 107 burgesses contributory to the neighbouring burh at Ilchester, and living there, and 56 of its own who lived in Milborne Port (DB Som 1,10) (discussed further in Haslam, forthcoming d). In all these instances the burgesses and their tenements clearly owed their existence to their roles in providing services for the king either at the royal centre or

A parallel situation is apparent in the way that the 34 burgesses listed under the royal manor of Somerton, and the 32 burgesses listed under the royal manor of Cheddar (both in Somerset), are stated as living at Langport and

16

Wiltshire

at the associated burh (or both), by virtue of their ultimate obligation to the king as overlord (Tait 1936, 83-4, and generally at 78-112).

bishop of Salisbury at the time of Domesday (DB Wilts 3,2). Its proximity to Ramsbury to its east suggests that this had been given to the new bishop of Ramsbury in c.909 as part of its original endowment, and that before this it had therefore been part of the royal multiple estate of which Calne was the centre (Draper 2006, 64). The specific connection between the contributing manor and the appurtenant tenement is most likely, therefore, to have been the result of the process by which the manor of Bishops Cannings would have been booked to the new bishop of Ramsbury in 909 (as with the manor of Ramsbury mentioned above), and that this connection therefore represents the fossilised remnants of services rendered to the centre by a constituent part of the royal estate at a rather earlier time. The hypothesis of the origin of the attachment of appurtenant tenements at Calne to both Calstone Wellington and Bishops Cannings at an early stage in the development of the multiple estate is therefore consistent with this arrangement.

The distribution of estates contributory to Malmesbury, Calne and Wilton provide further support for the inferences already adduced. It will be observed that one of these estates appurtenant to Wilton, at Castle Combe, is nearer to Malmesbury than to Wilton, and that its position indeed ‘overlaps’ both the southernmost manors contributing to Malmesbury as well as those contributing to Calne (fig. 1). Furthermore, this estate also shares the contribution of burgesses to Malmesbury, a characteristic also shown by a number of other manors in various shires. The pattern shown in this case seems therefore to contradict the hypothesis that the distribution falls within the primary burghal territory, which seems so clearly indicated in the cases of Malmesbury and Cricklade. An explanation, which is an important aspect of the general hypothesis outlined in this study, is provided by the estates contributory to Calne, and which is already hinted at in the discussion above. The record in Domesday Book of contributory burgesses at Calne, from the holdings of three tenants-in-chief in two manors, was one of the main stumbling blocks in the way of Tait’s acceptance of Maitland’s ‘garrison theory’, discussed above. This was a large unhidated royal (and later hundredal) centre with a minster church, which arguably shows some topographical indications which can be interpreted as vestiges of its early functions as an organic middle Saxon royal tun and market site, the centre of a large multiple estate, perhaps originating as such at an early stage in the formation of the West Saxon kingdom (Haslam 1984a, 102-6; Draper 2006, 64-6). It was however not a burh of the type of Alfredian Cricklade, Malmesbury or even Marlborough, and would have had no ‘public’ defences and no burghal territory, though it might well have had a defensible enclosure around the royal hall and probably also the church (Haslam 1984a, 102-6; Draper 2011). Given the evidence already adduced that the system of contributory estates in Wiltshire was created at least as early as the late ninth century, and that they originated as arrangements put in place to facilitate the performance by estate holders of military and other services at specified centres, it is possible to interpret the presence of the estates contributory to Calne as a survival of arrangements whereby these and other obligations of particular estate-holders in the hundreds nominated by the king were to be discharged at the royal tun in the construction of the king’s hall and its enclosure, the building of bridges (where necessary) and for performing army service and in various other services – all of which were in place in Wessex by at least the middle of the ninth century if not rather earlier (Brooks 1971; Abels 1988, 75-96).

The connections of these outlying estates with Calne can be seen as evidence for the same nexus of tributary relationships between the holders of estates within the orbit of influence of a royal estate centre as can be recognised in the burghal territories surrounding burhs. As Simon Draper has observed, in regard to burghal enclosures of the middle Saxon period, the maintenance of an enclosure around a minster church and a high status dwelling (a ‘burh’) at a royal estate centre would have been an important ‘means of achieving both productivity and social control’, resulting in the formation of ‘elite and non-elite zones of settlement’ (Draper 2011). The very formation of these elite or royal burhs would have created tributary relationships between the centre and the holders of lands within its territory. It is these relationships which were arguably expressed by the development of the same kind of tenurial heterogeneity as can be seen in ‘public’ burhs of military origin. Similar patterns of tenurial heterogeneity at a royal estate centre can be observed in for instance Milborne Port and Bruton in Somerset (Haslam forthcoming d). For the same reasons the existence of the attachment of the manor at Castle Combe to Wilton is also consistent with the development of Wilton as the shire ‘capital’ and preeminent royal centre within the shire (from at least as early as the early ninth century, and probably earlier - VCH Wilts vi 1962, 7-8; Draper 2006, 59), when these duties are likely to have been owed to this centre by estates over a wide area (if not all) of the shire from the earliest period at which the system in which land holders became responsible for these and other services was established. The probability that the original ‘shire’ of Wiltshire was extended northwards from a boundary along Wansdyke to the Thames in the early ninth century (Reynolds & Langlands 2006) does not alter the force of these arguments.

This is further indicated by the dependent relationship of the estate and later hundred of Bishop’s Cannings, which contributed to Calne (Haslam 1984a, 103; Draper 2006, 64-6). Bishops Cannings was a 70-hide estate held by the

The pattern of contributing estates shown at Wilton – in particular the instance of Castle Combe - could be interpreted therefore as representing the vestiges of a

17

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

situation in which the new burghal territory of Malmesbury of the late ninth century overlaid and subsumed survivals of connections which were established before this arrangement was put in place. It can also be argued that the sharing by Castle Combe of contributing burgesses to both Malmesbury and Wilton is a function of these successive roles of these two places – the territory of the late ninth century burh at Malmesbury overlying the larger and earlier regio of the early royal centre at Wilton, which at one stage would have comprised the whole shire. This interpretation is also appropriate to the historical and spatial contexts of the relatively frequent occurrence of manors in other shires which contribute tenements to two different places. Examples of these are described in the case of Gloucester and Winchcombe, and Worcester and Droitwich, discussed in chapters 6 and 7 below, and the manor of Pyrton in Domesday Buckinghamshire with the successive centres of middle-Saxon Oxford and late-Saxon Wallingford, discussed in chapter 10 and 11.

way as to ensure that general obligations for service, which was a concomitant of the holding of land, were directed to these ends. The obligations for defence – either in constructing defences or in garrison duty, or both - would have been only part of the manifold obligations which it would have been necessary to call upon to create and sustain what was a highly complex social institution. Furthermore, these institutional mechanisms appear to be similar to, and inherited from, those of an earlier age in which tenements at a king’s tun were also appurtenant to estates within the jurisdiction or the administrative area of that tun. The houses within a fortified burh (which in essence was merely a special instance of a king’s tun) which were attached or appurtenant to surrounding estates can be interpreted therefore as a facility which would have been added to the assets of each estate by the king as an act of state at the time of the formation of the burh and the creation of the burghal territory. This aspect is discussed in more detail in chapters 8 and 9 below. This would have created a situation which can be seen as being analogous to the ‘tenurial heterogeneity’ shown by some primary royal tuns or multiple estate centres (such as that at Deerhurst or at Berkeley in Gloucestershire), in which a number of tenants-in-chief or other tenants held small portions of the primary estate directly of the king, by inference as partners with the king in the maintenance and upholding of the administrative, economic and social structures and functions which were focussed on these central places.

The details of the temporal and spatial relationships between these places and their contributing manors form an important evidential basis for the overall model given above, which is explanatory in functional terms. This is that the attachment of burgesses to rural estates recorded in Domesday Book and earlier charters are best seen as a relic of the way that a burh or borough was set up from the beginning as a newly-created fortified place which was set out by the king as a permanent settlement, in such a

18

Chapter 3 The Burghal Territories in Wiltshire One test of the basic hypothesis outlined above is to compare the hidage values of the places in Wiltshire which show tenurial heterogeneity, all of which in origin are burhs which are included in the Burghal Hidage, with the hidage of the territories which are arguably defined by the orbits of distribution of the contributory manors to the centre. The Burghal Hidage sets out an arrangement which in Wiltshire initially comprised four burhs (Wilton, Malmesbury, Cricklade and Chisbury). Though an alternative solution has recently been put forward for this arrangement in a recent paper by David Baker and Stuart Brookes (Baker and Brookes 2011, 108-9), I argue here that the dependent territories must have divided the existing shire into four unequal portions as a single planned process. I have set out arguments elsewhere which suggest that the Burghal Hidage is broadly contemporary with the creation of the system which it describes – the specific period 878-9 (Haslam 2005; Haslam 2011 – and see further arguments later in this chapter). From this it can be inferred that the burhs, and the territories which were created to support them, were formed at the same time as part of the same system, notwithstanding the contrary view of Baker and Brookes (Baker and Brookes 2011). This whole question is discussed further in chapters 8 and 9 below.

observations which has been put forward by Peter Sawyer – that the decrease in total hidage values for each shire is the result of some burhs falling out of commission by the time of Domesday Book - and gives reasons for arguing that this is untenable (Brooks 1996b, 134-8; cf Hinton 1996). In considering various possible interpretations of these discrepancies, Brooks has concluded that it is the hidation given in Domesday Book which is ‘the ancient one’, and that the greater hidage values given in the Burghal Hidage would therefore represent an ‘emergency assessment for a military crisis’. This appears to be based on the possibility either that men were moved around from one burh to another as the occasion demanded, or that the Burghal Hidage itself may not have been ‘a military plan composed at one moment in history’ (Brooks 1996, 133-8). These points have been reiterated in similar terms in Brooks’ subsequent discussion of the topic (Brooks 2003, 159-62). Both of these solutions proposed by Brooks are, however, contradicted by other evidence. The idea that the burhs listed in the Burghal Hidage represent anything other than a system which was conceived and put in place at one moment in time, and that this system was created piecemeal over an extended period, is discussed, and decisively rejected, elsewhere (Haslam 2005; and further below in this chapter, and chapters 8 and 9) – although it has been revived by Baker and Brookes (Baker and Brookes 2011). The suggestion that hidages (and therefore men) were transferred from one place to another as the need arose, as suggested by Brooks, goes against the association of particular burhs with particular territories, in which men in one territory would have been responsible for the upkeep of the one burh in the territory which surrounded it. This can be inferred both from the arrangements which are indicated by the Burghal Hidage, and from the arguments which have already been rehearsed to the effect that particular estates would have contributed tenements to one burh in whose territory they lay. This discussion is extended in chapters 8 and 9 below.

The hidages of these burhs are as follows (‘preferred’ figures, from Hill 1996b): Malmesbury 1200 hides Cricklade 1500 Chisbury 700 Wilton 1400

Total

4800

One problem presented by these figures is that the total of 4800 hides is somewhat greater than the number of hides given for Wiltshire in Domesday. Maitland gives this number as 4050 (Maitland 1897, 400), while the figure from the Geld Rolls (Darlington 1955b) totals 3951 (my computation). A neat average of around 4000 hides for Wiltshire in Domesday would make the discrepancy some 800 hides. Various solutions to this puzzle, which is common to all the other Wessex shires except Somerset and Devon (as well as Oxfordshire, Berkshire and Buckinghamshire, which are discussed in chapters 10 and 11 below), have been examined in detail by Nicholas Brooks. He has articulated an explanation for this set of

Some difference of opinion appears to exist in further comments on this question, which, as will be seen, also takes in issues concerning the status of unhidated royal lands in Domesday Book. In his discussion of these estates in ‘core’ West Saxon territories - Hampshire, Dorset, Wiltshire and Somerset - which were subject to the farm of one night, Ryan Lavelle has observed that ‘There may

19

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

have been some relationship between boroughs and the royal lands providing the night’s farm, but a key point is the fact that the figures of the hides assigned to the continuing Domesday boroughs in the Burghal Hidage equate closely with the total hidages from Domesday Book. As the night’s farm lands were not hidated in Domesday Book, it is therefore possible to suggest that . . .these were the successors of lands which had not been hidated at the time of the Burghal Hidage. Such a suggestion depends upon the possibility that the assessments made in the Burghal Hidage were more than a shire could normally bear, as the Burghal Hidage’s provision for most of the boroughs that did not continue up to 1066 took the total hidations for most of the shires some way past their Domesday assessments.’ (Lavelle 2007, 22). Lavelle’s conclusions appear to be based on J H Round’s assumption that the royal one night’s farm manors had never been assessed for hides (Round 1900, 403), combined with Brooks’ discussion of Sawyer’s arguments already quoted, though without taking into account Brooks’ own reservations. In themselves Lavelle’s conclusions are feasible, but the supporting arguments are entirely circular: the inference that unhidated lands were always so is drawn from the assumption that the Domesday hidages were the primary assessments, which in its turn is merely an inference from the assumption that the unhidated lands had in fact never been assessed for hides. All this is called into play by reference to the further assumption (argued below as untenable), that the Burghal Hidage assessments were inflated by the temporary provision for ‘emergency’ burhs whose disuse by the time of Domesday brought the assessment to its base level.

figures, and c) the Burghal Hidage was only composed after this process had taken place (Baker and Brookes 2011, 108-9). As will be made clear in the discussions on this matter below, I will argue that the underlying premises and inferences which form the basis of this model are essentially flawed. It is abundantly clear that in all the treatments of these issues, one commentator’s conjecture or assumption, however illogical or insecure its basis, has become the next commentator’s certainty. There are several arguments which call in question the assumption that the Domesday hidage assessments were the original hidages of the shires, but which support the premise that it is the Burghal Hidage figures which represent their original assessment. Firstly, as Brooks has himself pointed out, hides were used as the basis for the assessment of many services and renders besides boroughwork. The phasing out the use of the smaller ‘emergency’ burhs would therefore have been unlikely to have reduced the general obligations due for services other than boroughwork (Brooks 1996b, 136; Brooks 2003, 160). Secondly, it was not the burhs or the eleventh-century boroughs themselves which were assessed; rather, the assessments were on burghal territories in the Burghal Hidage, and on estates or manors in Domesday Book. It is therefore illogical to argue that the hides attached to the burhs in the Burghal Hidage, and representing their burghal territories, would have fallen out of use or disappeared with the cessation of use of particular burhs as fortified centres by the eleventh century. This would beg the question as to what happened to these hides, considered as representing or being based upon cadastral units or tracts of real land from which many different obligations were due, in the intervening two centuries. Thirdly, this line of argument also ignores the fact that all the so-called ‘emergency’ burhs, which with their hides are assumed to have fallen out of use, were replaced at an early stage in the development of the primary Burghal Hidage system by new and larger fortified burhs on more strategically-placed sites adjacent to the old (Haslam 1984a 263-5; Brooks 1996b, 136; Haslam 2009, 98-100). Each of these would have necessarily have required a territory with a hidage assessment to sustain its upkeep. Reducing the total hidage assessment of the shire in this context would not have been an option before the Conquest.

A further extension of this particular line of thinking has been put forward by Nigel Baker and Stuart Brookes, who in their recent paper have come to similar conclusions concerning the disparity between the Burghal Hidage and the Domesday hidage figures. They appear to have followed Lavelle’s arguments, while again ignoring Brooks’ reservations on the whole issue, pointing out that ‘It must also be appreciated that the Burghal Hidage can only ever provide a distorted picture of the total provision made for civil defence because its principal concern was hidated shire lands, which thereby excluded significant tracts of territory, including royal estates which paid the farm of one night in Domesday Book.’ They add that ‘ these large tracts of land were not counted into the Burghal Hidage arrangements (or Domesday Book shire totals); in most cases the number of hides in the shire in Domesday Book is virtually the same as the number of hides reached by totting up allocations to Burghal Hidage burhs [which continued in use into the eleventh century] in each shire.’ (Baker and Brookes 2011, 105 & n.12). On this basis they have constructed a developmental model of the Burghal Hidage territories of Wiltshire, based on the premises a) that it is the Domesday values which represent the original hidation of the shire, b) that the extra hides were added to this value on the construction of burhs at Cricklade and Marlborough at a later date to give a total hidage which would have been more in accord with the Burghal Hidage

Fourthly, if hides as such did not exist in the unhidated royal lands which rendered the farm of one night, then one would have to invent them in order to establish a means by which both the warland and the inland peasants on these estates would have been called up for work on the burhs, at one man per hide. The suggestion, made both by Lavelle and by Baker and Brookes, that the absence of hidage values in the Domesday royal estates implies that they were unhidated at the time of the Burghal Hidage is no more than an unverifiable assumption – however often it is repeated. The pronouncements of J H Round, illuminating and important as they often are, are not infallible. Furthermore, the creation of the burghal system, with the consequent general conscription of the manpower of the shires to underpin both

20

The Burghal Territories in Wiltshire

its creation and ongoing sustainability, was, indubitably, an essentially royal enterprise. It would therefore stretch the bounds of credibility to suppose that the peasant occupiers of these royal estates, who tilled the land represented by the many hundreds of ploughlands given in the Domesday Book entries, would have been exempt from military duties relating to the construction and upkeep of this royal system, or from participating in the fyrd. They must have been as equally constrained to perform these obligations as others who did not live and work on the royal estates. That the royal estates were drawn into the nexus of tributary relationships which must have supported the burhs is shown by the frequency with which royal estate centres are recorded in Domesday Book as contributing tenements in many of the boroughs. It must be concluded that the hidage assessments in the royal estates would have been as real as those elsewhere, and that the assessments on the inland of these royal estates has simply been ignored in the Domesday survey, in contrast to the untaxed inland hides which were enumerated in great detail as portions of the total geldable hides in almost every estate in each of the hundreds in the Geld Rolls (Darlington 1955b). One reason for the non-recording of the hidage values in these estates has been suggested by Richard Abels as being the result of an arrangement whereby the shire reeves, the men responsible for the collection of the king’s revenues and who farmed these estates, were as a result able to retain a greater share of the income from the shire (Abels 1988, 107).

the rate of hidation’ in unhidated land, and an index which provides ‘a firm datum established by the commissioners themselves . . which is more valid that those normally employed’ (Roffe 2007, 206, 209, 310). Calculations based on the proportion of the numbers of ploughlands to hides given for the estates which were hidated can therefore give an approximate view of the levels of the ‘missing’ hidation in those estates in which no hidation value is given. This ratio can also, by the same token, highlight those estates where beneficial hidation has been at work. Adding these missing hides to those enumerated in the Domesday survey, in ways described below, creates a situation in which the hides on the ground in the later eleventh century and the hidage values given in the Burghal Hidage nearly two centuries earlier are very nearly or exactly equalised. This being so, it would follow that there is no reason for not accepting that the hidages given in the Burghal Hidage are the ‘primary hidages’ of the burghal territories attached to the burhs which operated together within this system, and that the Domesday totals are essentially under-recorded, for the reasons already given. The hides of the Burghal Hidage, in short, have disappeared from the record in the two centuries before the compilation of Domesday Book, but not from the land. The apparent disparity between these ‘primary’ hides and the Domesday hidages, which has formed such a stumbling block to previous commentators, simply disappears. The discrepancy between the figures from the Burghal Hidage and from Domesday Book can be made up from a simple series of calculations of the proportional relationship of the values of ploughlands to the values of the hides in manors where both are given. This can give the value of this lost or invisible hidage of these royal manors where the hides are not given. The rationale behind this is explored further below (chapters 8 and 9). In Wiltshire the unhidated royal estates are Calne, Bedwyn, Amesbury, Warminster, Chippenham and Tilshead (DB Wilts 1,1 – 1,5. 1,7), which together are assessed at 379 ploughs. Maitland’s figures for the hides and ploughlands for the whole of Wiltshire are 4050 hides and 2997 ploughlands (Maitland 1897, 400-1), which gives a ratio of 1.35 hides to 1 ploughland. However, his figure for the ploughlands must include those recorded in the unhidated manors, while leaving out the hides. A more accurate ratio for the rest of the shire should therefore be found by omitting the ploughlands on these estates as well. This ratio would be given by the formula 4050 divided by 2618 (which is 2997 less the 379 ploughlands on the royal estates), which gives a ratio of hides to ploughlands of 1.55 to 1. This is matched, for instance, by the figures from the hides and ploughlands in the first two fiefs given in Domesday Book (out of the total of 68). The twelve estates of the bishop of Winchester, totalling 259.5 hides (DB Wilts 2,1 – 2,12), are assessed at 146.5 ploughs, which represents a ratio of hides to ploughlands of 1.77 to 1. In the bishop of Salisbury’s five estates (DB Wilts 3,1 – 3,5), the total hidage is 267, provided by 174 ploughlands, which gives a ratio of 1.53 hides to one ploughland. Using these general multipliers, the number of ‘missing’ hides in the

The upshot of these arguments is that the discrepancy between the two sets of figures can be more easily explained by suggesting that it is the Burghal Hidage figures which represent the ‘primary’ hidage of the shire, and that these values have become reduced in the nearly two centuries up to the time of Domesday by the twin processes of beneficial hidation, and the exclusion of real but hidden hidation values from the numerous royal and other estates which appear as unhidated in the Domesday figures. Barbara Yorke has recently approached this perception by suggesting that the Domesday hidages could have been ‘topped up’ by extra hides from the unhidated royal estates (Yorke forthcoming), though she has not offered an explanation of how this could be worked out. This pool of potential conscripted labour, at one man per hide, is likely therefore to represent a good part of the shortfall of the ‘missing’ hides in Domesday Book. This raises the question as to how this value can be retrieved – or, in other words, how it is possible to validate the hypothesis that it is the Burghal Hidage figures which are the primary military assessments of the shires. Besides the value in pounds, an assessment which is given in entries in every manor, including in all the royal estates, is that of ploughs or ploughlands (carucis). An alternative means of retrieving these ‘missing’ hides in Hampshire, based on the recorded values of all the estates, has been proposed by Patrick Hase (chapter 4). David Roffe, for instance, has argued that figures for ploughlands provide ‘a measure of

21

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

unhidated royal manors given above would be between 671 (379 x 1.77) and 580 (379 x 1.53). There are, furthermore, three other royal manors - Corsham, Winterbourne Stoke and Collingbourne Ducis - whose hides total 40, but have 107 ploughlands – a clear case of beneficial hidation. The ‘real’ hidages in these manors should therefore be adjusted upwards, using the same multipliers, to between 144 and 189 hides, representing an extra 104 to 149 hides over the figures given in Domesday Book.

some cases the burghal territories as reconstructed appear to have cut across the boundaries of the Domesday hundreds, which are therefore likely to have originated or have been consolidated at a later date. A similar methodology is applied in chapter 11, below, to the reconstruction of the burghal territories of Oxford, Wallingford, Sashes and Buckingham within the later shires of Oxfordshire, Berkshire and Buckinghamshire. The recent reconstruction of the burghal territories of Wiltshire proposed by David Baker and Stuart Brookes is based on the arguably untenable premises that a) the Domesday Hidage represents the original hidage of the shire, b) the defence of Wessex through the use of burhs was a process involving the staged construction of burhs from the heartland of Wessex towards the Thames, and c) Cricklade must therefore be a late addition to an original system of burhs, in Wiltshire represented by Wilton, Malmesbury and Chisbury (Baker and Brookes 2011, 108-11, & esp fig.4).

The addition of the lowest figures in these sums (3951 + 580 + 104) equals 4635 hides; the addition of the highest (4050 + 671 + 149) equals 4870 hides. The latter figure is in excess of the 4800 hides given in the Burghal Hidage. The lower figure would, however, leave room for the fact that Malmesbury, and possibly also Wilton, could possibly have been hundreds (or possibly part-hundreds) in themselves (Tait 1935, 53), which would perhaps add up to 100 or so hides to the Domesday total. Furthermore, a close analysis of the relationships between hides and ploughlands over the whole shire might well bring up other instances of beneficial hidation (as is the case in Hampshire, analysed in chapter 4). It must be concluded that in whatever ways the ‘missing’ hides in Domesday Book are accounted for - which can never be an exact science - the total of the Burghal Hidage figures for the territories appurtenant to the Wiltshire burhs represents a round-figure assessment which in the late ninth century was nearly or exactly matched by the real hides on the ground which sent men at one man per hide for the construction and upkeep of the four burhs in the shire and for general military duties. This conclusion also has an advantage in that what can be concluded as being the ‘primary’ hidage of 4800 hides for Wiltshire in the late ninth century is matched by the figure of 4800 hides in the County Hidage in the eleventh (Maitland 1897, 456). A neat parallel to this situation emerges from an analysis of the figures for Oxfordshire, given in chapters 10 and 11 below.

Wilton The 1400 hides in the territory dependent on Wilton appear likely to have included all or some of the hides of the following hundreds in the southern part of the shire: Warminster 89.5 Heytesbury 137.5 Mere 86 Dole 70 Elstub (73) – part of : estimated 30 hides to W of r. Avon Swanborough (183) – part of: estimated 40 hides to W of r.Avon Amesbury 127.5 Dunworth 121 Cadworth 45.5 Stanford 105 Cawden 59 Alderbury 65 Downton 97 Frustfield 11 Underditch 70 Branch 108.5 Damerham 63 Sub-total 1325

It remains to be seen, therefore, whether these ‘real’ hides on the ground, insofar as they can be reconstructed, fit with territories attached to each burh which are mirrored in the orbits represented by the distribution of the connections of contributory manors to the former burghal centres. This analysis has inevitably to be based on the hundreds and their boundaries which can be reconstructed from the evidence in the Geld Rolls and Domesday Book, in spite of the fact that the Domesday hundreds are unlikely to represent the exact extent of the original ‘proto’- or ‘archaic’ hundreds’ of the later ninth century, by means of which, for instance, the military call-up of the time would have been organised. But this method is perhaps justified by the conclusion that the basis of the hundredal organisation in Wiltshire, based as it was on the extent of royal estates, was by the late ninth century already probably well established in some form (Draper 2006, 66). (For a more detailed map of the Domesday hundreds than that shown in figs 1 and 2, see Draper 2006, 67 fig. 26). It is the aggregate hidage of these hundreds, rather than their exact extent, which is the determining factor in this reconstruction. However, in

To this figure must be added: a) the ‘hidden’ hides from royal estates (calculated at 1.5 times numbers of ploughs) - at Amesbury (60), Warminster (60) and Tilshead (60); and b) extra hides from under-assessed Winterbourne Stoke = 16 hides Total 1521 Chisbury The 700 hides of the dependent territory of Chisbury are

22

The Burghal Territories in Wiltshire

Cricklade

likely to have included the following hundreds and parts of hundreds:

The hides from Cricklade’s dependent territory are likely to have come from the following hundreds and parts of hundreds:

Kinwardstone 196 Elstub (73) – part of : estimated 43 hides to E of r. Avon Swanborough (183) – part of: estimated 143 hides to E of r.Avon Collingbourne 14 Sub-total 396

Cricklade 49 Highworth 60 Staple 52 Shippen (Scipa) 80 Blackgrove 165 Thornhill 170 Selkley` 196 Ramsbury 90 Calne 91 Cannings ` 70 Studfold 94 Plus 60 from Clyffe Pypard [8, fig.1] (in Kingsbridge hundred) Sub-total 1117

To this figures should be added: a) the ‘hidden’ hides from Bedwyn, assessed at 80 ploughs = 120 hides; b) extra hides from under-assessed Colingbourne Ducis = 47.5 hides; and c) hides from Tidworth, Ludgershall and Chute Forest = 13 hides Total 576.5 Malmesbury

To these must be added the ‘hidden’ hides from Calne = 43

Malmesbury’s 1200 hides would have occupied most if not all of the north-west quarter of the shire, including the following hundreds:

Total

Chedglow 173 Dunlow 28 Thorngrove 113 Startley 152 Chippenham 142 Kingsbridge 119 (less 60 hides of Clyff Pypard) = 59 Melksham 86 Bradford 99 Whorwellsdon 92 Rowborough 96 Westbury 40 Studfold 94 Sub-total 1174

1160 hides

Discussion

To these must be added a) the ‘missing’ hides from Chippenham (150); and b) the extra 41 hides from underassessed Corsham;

The figures given above make it quite clear that the spatial distribution of the hidated lands in the shire, combined with the estimated ‘missing’ hides in the unhidated lands on royal estates, makes possible a credible reconstruction of the burghal territories of the four burhs of the shire included in the Burghal Hidage, seen as contemporary elements in a complete system covering the whole of Wessex. However, it is apparent that the hidages of all the hundreds and parts of hundreds which can for various reasons be assigned to the territories of the four burhs do not form an exact fit with the figures given in the Burghal Hidage - though the fact that this reconstruction of the situation in the late ninth century is based on evidence from the late eleventh implies that this would not necessarily be expected. This is summarised in the table below.

Total

The total hidage of the reconstructed territories could,

1365 hides

Table 5: Hidages of burghal territories in Wiltshire. Burh Malmesbury Cricklade Chisbury Wilton Totals

Hidage in the Burghal Hidage 1200 1500 700 1400 4800

Hidage of reconstructed territory

Increase / decrease

1365 1160 576 1521 4622

+165 - 340 - 124 +121

23

Percentage increase / decrease 14 23 18 9

‘Theoretical’ territories 1400 1200 600 1600 4800

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 3. Contributory manors in Wiltshire, shown in relation to the reconstructed burghal territories of Malmesbury, Cricklade, Chisbury and Wilton (based on the hundred boundaries of Domesday).

24

The Burghal Territories in Wiltshire

however, be increased by a further 100 hides or more by increasing the multiplier for converting ploughs to hides of the unhidated royal estates from 1.5 to 1.7, and by factoring in other hidden and therefore unquantifiable instances of beneficial hidation in other royal and non-royal estates by the time of Domesday. However, given an original hidage figure of 4800 hides for the shire, there is not very much room for flexibility in assigning areas to particular burghal territories. The common boundary of the territories of Chisbury and Wilton is taken here as following the upper part of the course of the river Avon, dividing the hundreds of Elstub and Swanborough to give parts of each to both territories. This is based on the incidence of connections of the manors of Marden [10] in Swanborough hundred and Netheravon [1] in Elstub hundred to Wilton. Further work on the relationships of the early development of the pattern of parishes in this area might elucidate these potential anomalies. Alternatively, the balance between the hidages of Wilton and Chisbury might be righted by the inclusion of part of the hundred of Amesbury to the south within the territory of Chisbury. This is unlikely, however, since most of this area is closer to Wilton than Chisbury. Chisbury’s territory is unlikely to have extended further into Cricklade’s territory to the north (say to the line of the river Kennett), as reconstructed here, since this would take even more hides from the latter’s seemingly underresourced allotment. None of Malmesbury’s hides could be given to Chisbury, nor Wilton’s hides to Cricklade, since the territories of these are not contiguous. It is possible too that the vast hundred of Kinwardstone, containing the royal manor of Bedwyn and the burh of Chisbury, was rated at rather more hides in the late ninth century than can be recovered from the Domesday evidence. There is some room, however, for some adjustment between the territories of Malmesbury and Cricklade to give the latter more of the former, but this cannot be done with any certainty on the available evidence either of the distribution of the connections of rural manors to each centre, or the disposition of the Domesday hundreds. The 60-hide manor of Clyffe Pypard with its 3 burgesses contributory to Cricklade (DB Wilts 27,9) [8], and which is made up of 14 separate components in as many fiefs in Domesday Book, is situated in Kingsbridge hundred, in which are also two manors which are contributory to Malmesbury (Wootton Bassett [14] and Woodhill [19]), which implies a split along a line which is not represented by the hundred boundary.

territory of Chisbury. This would bring the hidage of the latter up to 670, but would diminish that of Cricklade from 1160 to 1066, which is even further from the Burghal Hidage figure of 1500. This is rendered even more unlikely by the early development of the two hundreds of Cannings and Studfold, which appear to have been two parts of the same territorial block with a hundred meeting place near their common border (Pitt 2003, 80-1), by inference before Bishops Cannings was given by the king to the new bishopric of Ramsbury in 909. The incidence of connections of Bishops Cannings with Calne (table 4 above) suggests that the hundreds of Cannings and Studfold were part of a larger territory dependent on the royal centre of Calne, and that these connections – like that of Ramsbury with Cricklade – had already been in place by the time that the new bishop of Ramsbury was given Bishop’s Cannings estate in 909. In relation to this hundred, a further problem is shown by the possible connection of the manor of Etchilhampton (DB Wilts 25,5) in Studfold hundred or of Calstone Wellington (DB Wilts 25,5) in Calne hundred, with Malmesbury. These manors were the only two in the shire held by the wife of Edric, whose lands were taken over after 1066 by various tenants-in-chief. Since she also held one tenement in Malmesbury (DB Wilts M15), it might be reasonably inferred that this tenement was appurtenant to one or other of these manors. If so, this would place the hides in these hundreds as being within Malmesbury’s burghal territory, on the model argued in this study. But the topographical and other reasons for placing Calne, Cannings and Studfold hundreds in the territory of Cricklade are compelling. A solution is provided by the notice under the manor of Draycott Cerne [10] (see figs. 1 and 2], which was held by Geoffrey the Marshall as successor to Edric (DB Wilts 68,21). Geoffrey held one burgess in this manor which was not assigned to a centre, while Edric’s wife held one burgess in Malmesbury which was not assigned to a manor. It would be reasonable to infer therefore that the wife’s tenement in Malmesbury was held as an appurtenance of her (?late) husband’s manor at Draycott, since Draycott, in Startley hundred, must have been within Malmesbury’s burghal territory. This is notwithstanding the entry in section M12 which assigns Geoffrey the Marshall only half a tenement – perhaps shared with Alfred of Marlborough (section M11). One solution to the disparity in the figures analysed above over against those of the Burghal hidage would be to consider that the original figures given to each burh in the Burghal Hidage for Wiltshire are themselves incorrect. This is demonstrated in chapters 10 and 11 below in the case of Oxford, where the disposition of the hides on the ground clearly indicate that its burghal territory would most probably originally have been assessed at 2400 hides rather than the 1500 hides given to it in the Burghal Hidage. It might in this case be possible to reconstruct the ‘theoretical’ sizes of the territories, as in the last column of the table above, based on the premise that the original hidage of the whole shire was 4800, the same as the figure of the later

A particular puzzle is the status of Studfold hundred (of 94 hides), which lies right in the centre of the shire and occupies some of the highest ground over the major watershed which divides the south-eastern part of the shire from the north-western part (and which posed such a problem for the builders of the Kennet and Avon canal in the 19th century) (figs. 1 and 2). This has been placed in the territory of Cricklade, together with Calne and Cannings hundreds, as forming the best-fit solution to the distribution of the hides appurtenant to Cricklade and Malmesbury in the Burghal Hidage. However, from a topographical point of view Studfold hundred could have formed part of the

25

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

County Hidage (above). My own arguments which were used to demonstrate that the ‘preferred’ Burghal Hidage value of Cricklade should be 1500 hides rather than 1400 hides were based on the then prevailing paradigm that the sizes of defences could be predicted from the hidage - in this case with the linked assumption that the known size of the defensive circuit was therefore indicative of the hidage where ambivalent (Haslam 1986). I have subsequently argued that this whole paradigm is untenable (Haslam 2009, 111-4). There is therefore no evidence which necessitates the conclusion that Cricklade’s hidage assessment in the Burghal Hidage must be 1500 rather than 1400 hides, or indeed any other figure. However, the main reason for accepting the figure of 1500 hides for Cricklade as being the ‘original’ figure is that this makes up the round duo-decimal total of 4800 for the assessments in the shire as a whole.

This being so, this would then have been divided up between the four burhs by a complex process of accommodation between the pre-planned or ‘theoretical’ hidages which are listed in the Burghal Hidage and various other factors, including the already-existing hidated values of royal estates and / or original ‘proto-hundreds’, the topographical relationships of territories to each other and to landscape features such as rivers, river-systems and watersheds, and possibly already-existing patterns of tenurial holdings. It appears to have been an exercise in top-down planning in which the assessments of the hidages which were initially given to particular places were adjusted to fit these constraints on the ground, seemingly without exact regard to what had been planned from the start. There is every reason to hold that one important aspect of these constraints would have been the irregular pattern of hundreds which to some extent are already likely to have been well established by the late-ninth century (Draper 2006, 66-8; see further discussion on these points in Lavelle 2007, ch. 4). As such, it appears to represent a compromise solution between the hidages (and therefore men) which were thought necessary to construct and maintain the burhs as defended places, which were possibly assigned in a rough-and-ready way in proportion to their sizes and possibly their perceived strategic importance, and both the extent and disposition of the territories from which these men would be drawn and their perceived significance as settlements at the time. It is not improbable that the actual boundaries of the shire (and of other contiguous shires) were settled in their final form through this process of adjustment and accommodation.

It might seem reasonable to suggest that Cricklade should have been given fewer hides than Malmesbury, which in the late-ninth century was a well-established place of some historical importance (Haslam 1984, 111-7), and which at the time of Domesday (and doubtless earlier) was possibly reckoned as a half or a whole hundred in itself (Tait 1936, 45), whereas Cricklade, though of considerable strategic significance at the time it was built as a burh (Haslam 2005, 130), was historically of lesser importance as a settlement (Haslam 1984, 106-10). By the same account, Wilton might be expected to have been given more hides than either Malmesbury or Cricklade, given its earlier importance as the centre of an early regio and the chief place in the shire (Darlington 1955a, 1-5; VCH Wilts vi 1962, 7-8; Haslam 1984a, 122-8; Draper 2006, 59). It might be of some significance, therefore, that the reconstructed hidage on the ground matches most closely the ‘theoretical’ hidage of their territories as reflecting the perceived hierarchy of the four places as settlements.

Various conclusions flow from this assessment. Firstly (and in contrast to the reconstruction put forward by Baker and Brookes), it is clear that the four burhs and their territories in Wiltshire were seen as a mini-system of interlocking elements which was conceived and applied to the whole of the shire at one point in time. The concept of a ‘bestfit solution’ was as appropriate to this context in the late ninth century as it is to the reconstruction of this pattern in the twenty-first. Insofar as this mini-system in Wiltshire necessarily interlocked with those in other adjoining shires, and with shires adjoining them, it must be inferred that this was part of a wider system which covered the whole of greater Wessex in one process, and that therefore all the burhs in the system as a whole – whatever their immediate strategic or local function – must be considered as being contemporary elements. On the evidence presented here it is not possible to envisage Cricklade, for instance, as having been formed at a later date than the other three burhs, or that the system as a whole developed in stages by degrees to meet disparate strategic needs, both of which premises are important aspects of the alternative model put forward by Baker and Brookes. This issue is discussed further below, with regard to the interlocking burghal territories of Oxford, Wallingford, Sashes and Buckingham, in chapters 10 and 11.

This is, admittedly, somewhat speculative, but has merit in being able to account for the disparities between the figures of hides for the burhs in the Burghal Hidage and disposition of the hides comprising their reconstructed territories on the ground. However, an alternative interpretation, which involves rather fewer undemonstrable assumptions, would be to suggest that the Burghal Hidage figures are those which were indeed planned from the start, with the intention that they would represent the hides (and therefore men) available for the maintenance of each burh, while the ‘theoretical’ figures are those of the territories which for various reasons came to be assigned to them for their support, and which perhaps reflected their perceived importance as places at a local level. In this case the ‘theoretical’ hidage (table 5 above), becomes the ‘actual’ hidage on the ground, while the Burghal Hidage figures could be seen as the ‘theoretical’ hidage, which was not achieved on the ground. It is apparent from this exercise in reconstruction that the territories assigned to the four burhs were based on assessments of hidage values which there is no reason to suppose did not originally add up, nearly or exactly, to the total hidage value of the shire of 4800.

Secondly, the evidence presented here provides direct support for the hypothesis I have presented elsewhere that

26

The Burghal Territories in Wiltshire

the Burghal Hidage as a document must originally have been produced as a prescriptive list of burhs and their hidages which was set out centrally during the process of the planning and construction of the system of burhs as a whole, covering all of the shires of greater Wessex over which King Alfred exercised hegemony. As I have recently pointed out,

Whether the production of the document was part of the initial phase of construction of the system in 878-9, or whether it played a part in the consolidation of the system in the early 890s (as argued elsewhere – see Haslam 2009, 98-100, 103-4), it would be not unreasonable to see this as evidence for a duality in the administrative procedures of the time, in which a centrally-planned scheme – represented by the figures and the structure of the Burghal Hidage – was worked out and implemented on the ground on a shire-toshire basis by the accommodation of the requirements of the scheme to the realities of the patterns of local administration and local topography by the ealdorman and thegns in the shire. That the prescriptions handed down from the centre were not exactly met in these arrangements in the shire need not necessarily have affected the efficacy of the system as a whole.

‘. . . the list of the Burghal Hidage and its assessments can best be interpreted as a ‘top-down’ set of summary prescriptions from the originators of the system to those responsible for setting it up in the field. The assessments therefore reflect moreor-less closely the decisions already taken about which defended places should be placed at which sites, about their character and size, and about the ways in which such a complex system with so many variables was to be facilitated and supported by the resources available on a shire-to-shire basis. The Burghal Hidage can be interpreted therefore as a follow-on from an initial directive which determined the most suitable sites that matched the strategic objectives of the system as a whole. This was then circulated from the centre to the shires as an aid to marshalling the resources necessary to ensure that all the elements of the system would be put in place. It clearly leaves many decisions about how to implement these assessments on the ground to the ealdorman, reeves and thegns in each shire, who would have had the responsibility of conscripting the men required to do the work.’ (Haslam 2009, 113).

Thirdly, as I have suggested elsewhere, one of the main motivating factors for the production of the document in the context of the initial organisation of the burghal system - presumably by the royal administrative machine in Winchester - could perhaps be seen in its function as an essentially political aid, or even a goad, to the ealdormen of the shires, to get them to see the military organisation of their particular spheres of influence as part of a scheme of national regeneration, and to act on this imperative (Haslam 2005, 129-33 - see further discussion in chapters 9 and 14 below). This might provide the context for the fact that the Burghal Hidage lists all the burhs in the whole of Wessex as a complete system, rather than on a shire-to-shire basis. It also reflects the function of the connections of rural manors to their centres as encapsulating one element of the control of the loyalties and submission of the tenants-in-chief of the shires by the king through the building of burhs and the setting out of their ‘territories of obligation’, as elements in the wider development of tributary relationships. This being so, it provides a credible context for the production of the Burghal Hidage document, linking it unequivocally to the time-frame in which the burghal system as a whole was constructed. I have elsewhere produced arguments supporting a date range between mid 878 and late 879 for the construction of the system (Haslam 2005, 133).

Given that the disparities between the Burghal Hidage figures and those of the reconstruction of the burghal territories represent a real anomaly (rather than one based on a misinterpretation or misreading of the evidence, such as it is), the evidence adduced in this study provides essential support for this hypothesis, to the effect that this process must have been initiated, and the document drawn up, after the most suitable sites of the burhs had been chosen and hidages assigned to them, but before the detailed arrangements for their support had been put in place on the ground.

27

Chapter 4 Hampshire Hampshire provides a particularly good instance of the way in which the recorded connections between rural manors and urban tenements appear to fit with the disposition of the burghal territories of the four burhs in the shire which were included in the Burghal Hidage – Winchester, Twynham (Christchurch), Southampton and Portchester. The evidence of the orbits of these connections thus provides an important test of the hypotheses set out in the previous chapters. Furthermore, the way in which the burghal territories of the four burhs can be reconstructed also provides a test – and arguably a validation - for the relationship between the hidages in the Burghal hidage and the hidages recorded in Domesday which is put forward in relation to Wiltshire in the preceding chapter. The pattern of these connections shown in Winchester also exhibits regularities which are characteristic of places in Wiltshire, examined in the previous chapters, as well as of those of other places in the chapters following.

with some additions, for which list see further details of values, names of types of tenements and folio numbers of the Domesday entries. The orbit of distribution of the estates given above which are contributory to Winchester occupies much of the area of what can be inferred from other evidence as comprising its burghal territory (fig.5), which is discussed below. The only anomalous instance is Minstead, in the heart of the New Forest, which must have been within the burghal territory of Twynham. Minstead was the main manor of the sons of Godric Malf, whose other estates – all small manors – lay within the area of the New Forest. Patrick Hase (pers.com.) has suggested that Godric was the chief forester of the New Forest, and that his dependent tenement in Winchester was, by implication, a gift of the king to facilitate his business in Winchester to discuss Forest business. It is perhaps significant that the manor of Minstead lay in the area of the primary (arguably pre-Conquest) Foresta Regis, rather than the area of the Nova Foresta to the south, which was probably its post-Conquest extension (Mew 2000, 158-60).

There is, as might be expected, a relative abundance of evidence for these connections in Winchester, which can be derived from the pages of Domesday Book (and which might well have been augmented in the ‘missing’ entry for Winchester itself), from pre-Conquest charters and from several entries in the Winchester surveys I and II of c.1100 and c.1148 (Biddle (ed.) 1976). There are a few references to connections with Twynham and Southampton, but none to connections with Portchester. Neither are there any connections of rural estates and urban tenements recorded in the case of other markets mentioned in Domesday Book at Basingstoke, Neatham and Tichfield (Finn 1971, 353), or in other probably late Saxon urban centres at Kingsclere, Stockbridge, Andover, Fareham or Romsey (Hinton 1984, 151-2), although Basingstoke itself had both a market and contributed four tenements in Winchester (DB Ham 1,42). The complexity of the patterns of connections shown for instance by Milborne Port and Ilchester in Somerset, where the royal centre of Milborne itself had burgesses and showed heterogeneous tenure, while also having burgesses in neighbouring Ilchester (see further in Haslam forthcoming d), is not shown by any of the places of similar status in Hampshire.

The cases of the tenements held by the abbeys of Romsey and Wherwell, two of the great fief-holders in Winchester in the two early medieval surveys discussed by Biddle, are important in establishing some of the mechanisms by which heterogeneous tenure became so widespread in early medieval towns. Biddle has suggested that ‘by virtue of their size [these properties] seem to represent urban estates of an altogether different character [to other tenements in Winchester attached to urban estates]’ (Biddle (ed.) 1976, 384 n.1). These differences are, however, arguably only in scale, but not in kind, since their characteristics are similar in many respects to other tenements or hagae in other towns held by monastic institutions. Furthermore, the origins of the connections of their urban tenements with their holdings in the shire can be explained in very similar ways, as arising from gifts to the religious house of estates which had urban tenements already attached at the time of their gift. Wherwell Abbey held six estates, including the vill of Wherwell itself (DB Ham 16,1-6). Its 31 urban tenements are enumerated in section 16,7 as a group without any indication of whether these were attached to its rural estates. This way of recording these connections is, however, similar in many respects to the way in which, for instance, the 28 tenements held by Evesham Abbey in Worcester are

Winchester The following table is based on that given in Winchester in the Early Middle Ages (Biddle (ed.) 1976, 384 table 28)

28

Hampshire

Table 6. Winchester – non-customary tenements Number in fig. 4

Number of tenements

Nether Wallop Clatford Basingstoke Faccombe West Meon Houghton

1 2 3 4 5 6

2 7 4 6 8 3

Mottisfont

7

1

Eversley Preston Candover Bramley Corhampton

8 9 10 11

Headbourne Worthy

12

King’s Somborne Stratfield Saye Bramdean Norton Dummer Minstead

13 14 15 16 17 18

1 1 3 1 8 1 9 1 1 5 3 1

William de Eu Hugh son of Baldri Milo the porter Odo of Winchester Odo of Winchester Sons of Godfric Malf

Awbridge & Houghton

19

3

Hugh de Port

West Tistead

20

1

Kilmeston

21

1

Hurstbourne

22

13

Overton

23

1

King

Polhampton

24

1

King

Lasham

25

2

King

Odiham

26

-

King

Wolverton

27

1

Alresford

28

Romsey

29

Wherwell

30

Manor

Holder of manor King King King King Bishop of Winchester Bishop of Winchester Archbishop Thomas of York Westminster Abbey Earl Roger of Shrewsbury Hugh de Port Hugh de Port Ralf de Mortimer

K Eadred granted to Aethelgeard K Edgar granted to Aethelwulf K Edgar granted to monks of Abingdon

Romsey Abbey Wherwell Abbey

totted up at the end of the section in which its estates are enumerated (DB Worc 17 – see chapter 7 below). As is pointed out, the mention of the fact that one of these estates, at Bengeworth, was contributory to four of its tenements in Worcester in a separate document of 1077 carries the implication that all the other 31 tenements were appurtenant to most or all of the other estates recorded as being held by Evesham Abbey in section 17. Analogous situations can be inferred in the cases of the estates of Bath Abbey and its tenements in Bath (where the tenements are given in a

Comments

1,19 1,25 1,42 1,46 2,11 2,20 4,1 8,1 21,2 23,5 23,19 29,3 39,3 32,1 44,3 68,7 69,2 69,7 NF 9,37 23,16

Bishop of Winchester 14 31 1 mill

Section in DB / source

S.488 (946x55) S.693 (961) S.689 (961) Survey I, 152 Survey I, 156 Survey I, 157 Survey II, 387, 392, & ?830 Survey II, 435 Survey II, 742 15,1 16,7

Winchester not mentioned. Biddle (ed.) 1976, 384 n.1 suggests these were not urban dwellings Connection not given in DB Connection not given in DB Connection not given in DB Connection not given in DB Connection not given in DB Connection not given in DB Connection not given in DB Connection not given in DB Connection not given in DB Abbey held 5 estates in DB Abbey held 6 estates in DB

separate section before its manors - DB Som 7,1; 7,2-15); the estates held by Coventry Abbey with its 36 tenements held in Warwick (discussed in chapter 5 below); and the estates of Abingdon Abbey in later Berkshire and its 14 tenements in Oxford (discussed in chapter 10). A similar situation occurs at Wilton, Wiltshire, where the account of the abbey’s 21 estates in the shire (DB Wilts 13,1-21 – see fig. 1, chapter 2) are completed with the separate statement in the following section to the effect that ‘All the payments which the church has from the Borough of Wilton

29

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

17), which implies not only that its principle haga was located in this street, but also, like that of Wherwell, this had become subdivided by the time of Domesday. As with the case with Wherwell Abbey, its other properties in three other streets can reasonably be inferred to have come to it by the gift of estates which already had appurtenant urban tenements attached to them. In the cases of both Wherwell and Romsey it is also reasonable to infer that the principle tenements of both abbeys in Winchester were part of their original endowment, as being a direct gift to each abbey from the king of a single core estate, to which an urban tenement was already attached.

are valued at £10 17s 6d.’ (DB Wilts 13,22). The size of this payment suggest with little room for doubt that this refers to, or includes, the income from what must have been a considerable number of tenements within the borough which were appurtenant to most or all of its rural estates in the southern part of the shire. It can be reasonably inferred, therefore, that Wherwell Abbey’s 31 tenements in Winchester were appurtenant to most if not all of its five estates in the shire, all of which lay in Welford hundred, in which Wherwell itself also lay. These are marked in figs. 4 and 5. That these tenements were distributed around four of the streets of late Saxon Winchester appears to confirm the conclusions that these estates would have been given to the abbey subsequently to, or even perhaps on the occasion of, the creation of the abbey in c.986 by Elfrida, widow of King Edgar (VCH Hants ii, 1903, 132-7). It is likely to have been on this occasion that the king would have granted the haga in Winchester to the abbey, together with the customary dues on it (DB Hants 16,1) which would otherwise have gone directly to the king. Houses in Winchester were mentioned in the charter of confirmation to the abbey by Aethelred (ibid.). However, the concentration of half of the rents of the abbey in one street (Biddle (ed.) 1976, table 11) would appear to signify that some of the abbey’s 31 tenements in Domesday could well have comprised subdivisions of the abbey’s principle haga in this street – a situation which appears to be directly reflected in the topographical evidence (ibid. 341-2 & fig. 12). It would follow that the rents in other streets were derived from urban tenements which would have come to the abbey in association with the other rural estates given to the abbey either at the time of its foundation or subsequently. The subdivision of initially large hagae can be seen in other towns, and for instance in the subdivision of some of the properties of the king and other landlords in Winchester (Ibid. 341, 351). This process might also explain the holding of 13 tenements in Winchester which were appurtenant to the 50-hide manor of Hurstbourne, given by the king to Abingdon Abbey in 961 ([22], table 6), as being a single haga which already by 961 had become subdivided.

Little can be said of the urban tenements of the other great fiefs in Winchester. It is a reasonable inference that many of these tenements had once been appurtenant to the rural estates of the fief-holders, including the king, though there is no evidence which has survived relating to this issue. However, one important aspect of the distribution of these tenements is that they were all intermingled on all the streets of the late Saxon burh in ways which demonstrate that there was no systematic allocation of tenements of the great fief-holders (and other estate-holders) to particular streets or blocks of streets, although some, such as those of the bishop, tended to be concentrated in areas. In particular, the overall pattern of distribution of the tenements held by the king, on which landgable was payable, was spread out over nearly the whole of the inhabited area, including the western suburb (Biddle (ed.) 1976, 35-3; table 12a & fig. 14). That these tenements can be assumed to have formed the customary tenements, from which dues were paid to the king alone, carries the implication that these would have been laid out as part of the allocation of land at the time of the initial (re)organisation of the streets and tenements at the time of the formation of the burh by King Alfred. It would follow from this that the tenements of the other fief-holders, including those not amongst the seven great fiefs characterised by Biddle, which were intermingled in all of the intra-mural streets and certainly the western suburb with the king’s tenements (Ibid. 379-80; especially fig. 15, 18 & 19; tables 18 & 19), were allocated as the non-customary tenements to the tenants-in-chief of the shire at the same time. It is this conclusion (also put forward by Biddle – ibid. 349, 358), which is drawn from the virtually unique and detailed evidence from the two twelfth-century surveys of Winchester, which underpins the underlying premise in this study that the attachment of these tenements to rural estates would have been put in place at this time as well, rather than being an accretion or a product of evolution over the two centuries before this pattern can be recognised as being universal at the time of Domesday. This is discussed further in chapters 9 and 14 below.

The evidence in Domesday Book relating to Romsey Abbey, another of the seven great fiefs of Winchester, can be interpreted in a similar way. The abbey was founded by King Edward the Elder in c.907 (VCH Hants ii 1903, 126), and its 5 estates in Domesday were scattered over the whole of the shire. Three of these - Romsey itself, Itchen Stoke in Bountisborough hundred, and Sydmonton in Kingsclere hundred, were presumably contributory to the 14 tenements in Winchester described in the section on Romsey (DB Hants 15,1). Its other two estates, at Totton in Redbridge hundred and Sway in Boldre hundred, possibly originally held tenements in Southampton and Christchurch respectively, though there is no evidence from either place of an association with Romsey. Romsey Abbey drew 70% of its rents in Winchester from Gerestret to the south of the High Street in 1148 (Biddle (ed.) 1976, tables 11 &

Southampton Southampton is pictured in Domesday Book (DB Ham S1-3) as a complex borough with heterogeneous tenure as intricate as any. However, only two instances of

30

Hampshire

Fig. 4. Orbits of connections of rural estates in Hampshire to Winchester, Christchurch (Twynham), and Southampton. North to top.

contributory manors are given in the body of the text under the rural manors, and these references are probably also repeated in the borough entry (see table 7 below). There are a number of other examples of the holding of tenements by tenants-in-chief in section S3 which it is not possible to connect with any rural manor, either because the individual had no holdings in the shire or because their holdings would not have been in the burghal territory which on other grounds could be assigned to Southampton.

Out of 22 holders of tenements in section S3, 16 either did not hold any estates in the shire, or held estates in hundreds other than Redbridge or Mansbridge, which are identified below as the two hundreds which must have formed the burghal territory of Southampton. Of the remaining six, two (Reginald son of Croc and Richard Povnant) held only one manor each in the shire (Woolstone and Netley respectively), both in Mansbridge hundred. It a reasonable inference, therefore, that the manors held by these two were contributory to their tenements in the borough. A slight

31

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Table 7. Southampton – customary and non-customary tenements Manor

Number of tenements

Number in fig. 4

Holder of manor

76

The king

Section in DB / source S1

65 Frenchmen

S2

31 Englishmen

S2

Shirley

1

4

Ralph of Mortimer

29,5

Chilworth

2

3

Bernard Pancevolt

39,4

Comments Customary tenements All paying customary dues to the king All paying customary dues to the king Probably the same as the 3 exempt houses held by RM in S1 Probably the same as the 3 exempt houses held by BP in S1

Holders of tenements from section S3 Holder of tenement

Number of tenements

Bishop G Abbot of Cormeilles Abbot of Lyre Count of Evreux

1 1 1 2

Ralph of Mortimer

2

Gilbert of Bretteville William son of Stur Ralph of Tosny Durand of Gloucester

2 2 2 2

Hugh of Port

1

Hugh Grandesmnil Count of Mortain Aiulf the chamberlain Humphrey his brother Osbern Giffard Nigel the doctor Rochere of les Andelys Richard Poynant Stephen the steersman Turstin the chamberlain

2 5 5 1 1 4

Number in fig. 4 [connections inferred]

Manor in Redbridge or Mansbridge hundred

3 4

Botley N Baddesley none none none none

29,6 29,7

Lestred

23,67

unlocated

4

6

1 2

7

Sect in DB

none none none South Stoneham church Netley -

2

Comments ?Not an estate-holder Not an estate-holder in the shire Not an estate-holder in the shire Not an estate-holder in the shire 16 manors held in Hants

Not an estate-holder in the shire Only 1 out of 69 manors within the hundreds of Redbridge and Mansbridge Not an estate-holder in the shire Not an estate-holder in the shire Not an estate-holder in the shire 3,16 42,1

Held 2 other churches which belonged to this mother church Only one manor held in Hants Not an estate-holder in the shire

none

Ansketil son of Osmund

2

8

Northam

68,4

Reginald son of Croc Abbess of Wherwell

2 1

9

Woolstone

59,1

oddity is the appearance of the tenement held by Rochere of les Andelys, who was clearly the priest of the minster church of Stoneham (as well as two others in its parochia), and who can be inferred to have held his four tenements in respect of his position, notwithstanding the fact that the estate as a whole was held by the monks of the Old Minster in Winchester. It might be that the tenement held by Bishop G could also have been appurtenant to this estate as a whole. As for all the other holders of tenements, it is difficult to find a satisfactory reason why many of them should have

Only 1 virgate in size. Given as in Redbridge hundred, but in fact in Mansbridge Only one manor held in Hants 1 mill held

been holding tenements, if they were not attached to a rural manor. It is however possible that these urban tenements might have been direct gifts from the king to individuals, connected in some way with services they might have provided to the king at the port after the Conquest, and thus not connected with the role of Southampton as a burh. A further factor which has certainly influenced to structure of the tenurial holdings of Southampton immediately after the

32

Hampshire

Conquest (and therefore in its account in Domesday Book) will have been the formation of a ‘quasi-rape’ by William fitz Osborne, which appears to have included the Isle of Wight and part of the adjacent mainland – specifically the areas forming the former burghal territories of Southampton and Christchurch - as a strategic initiative to secure the area for the king immediately after the Conquest (Mew 2000, 162-5 – for Christchurch see n. 46). As Karin Mew states, in relation to the situation at Southampton:

set aside from the royal estate for the provision of the minster. It is probable, therefore, that the manor of 1h 3v at Bashley, held by Alsi the priest (DB Hants 17,3), which formed part of the estates of Christchurch, would also have contributed a tenement in the borough. The holding of Bosley by Alnoth the priest puts in context the reference, above, to the holding by Rochere of les Andelys of the minster at Stoneham, which was held by the monks of the Old Minster at Winchester for their supplies, yet contributed a tenement in Southampton. It also provides an analogy to the context for the holding of 25 tenements in Calne, Wiltshire, by the church, already discussed (table 4, chapter 2), presumably in respect of the estates which the minster held within the royal estate of Calne itself.

‘The entry for the borough of Hantune was recorded in fo.52 immediately before the section on the Isle of Wight, and the connection between the lord of the Isle and the borough takes on greater significance given that in Domesday Book the Sussex Rapes, which were also military lordships, each had a castle and borough town. Not only does a charter issued in 1070 in favour of Earl William’s Norman abbey of Lyre provide further evidence of Earl William’s lordship of the isle of Wight, but it also shows that the earl exercised authority over the borough as he granted the abbey property, land, income and concessions in the borough of Southampton. The Domesday entry for Hantune confirms that the abbots of Lyre and Cormeilles, the earl’s other Norman foundation, each had a house in the borough, and other men associated either with the earl or with the Island also had houses there. His brother-in-law Ralph of Tosny had a house, for example, and the Island sheriff William son of Stur had two houses as did Gilbert of Breteuil [the central town in William’s Norman estate]’.(Ibid., 163).

The burghal territories of Hampshire The burghal territories in Hampshire can be relatively easily inferred from the hidages given in the Burghal Hidage. These territories can be seen to have conformed with the pattern of the long-lived major divisions of the shire into royal estates centred on villae regales, each of which became the parochia of an old minster located at the estate centre from the later seventh century, and which have been reconstructed through the work of Patrick Hase (Hase 1988; Hase 1994). The territories of each of the burhs in the shire must have been assigned in such a way as to form an interlocking system or network, which was laid out at one moment in time within the envelope formed by the already-existing shire of Hampshire. It is possible, indeed likely, that the boundaries of the shire became finally fixed in a manner that they had perhaps not been before by the necessity to closely define the territories which were to owe service to a particular burh. The conclusion follows from this that the burhs themselves were contemporary elements of a system or network, and that this network extended beyond the borders of the shire to encompass the whole of Wessex. This aspect is discussed further in chapters 8 and 9 below.

Christchurch (Twynham). The record of manors contributing to Christchurch in the pages of Domesday Book is minimal. It is possible, however, that such connections as may be assumed to have once existed in the early years of the development of the burh could well have been lost in the processes of the creation of the King’s Forest before the Conquest, and of the New Forest immediately after, discussed above.

The exercise in reconstruction which is based on these comparisons is at once faced with the problem of the disparity between the total recorded hides in Domesday Book and the higher figure derived from the total hidages of the four burhs in Hampshire recorded in the Burghal Hidage, as shown in table 9 below.

The record of non-customary tenements comes solely from the relatively small estates held by the canons of the minster, which may be reasonably regarded as the manors

Table 8. Christchurch – customary and non-customary tenements Manor [ ] inferred

Number in fig. 4

Number of tenements

Christchurch (Twynham)

31

Christchurch

6

Bosley

1

2

[Bashley]

2

-

Holder of manor The king The canons of Christchurch The canons of Christchurch The canons of Christchurch

33

Section in DB 1,28

Comments Customary burgesses

17,1 17,2 See discussion below

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 5. Reconstruction of the burghal territories in Hampshire of Winchester, Christchurch (Twynham), Southampton and Portchester, with contributory rural estates (from fig. 4) and hundreds (see key below). North to top.

The hundreds of Hampshire A – Andover As – Ashley B- Barton Ba – Basingstoke Be – Bermonspit Bo – Bountisborough Bol - Boldre Bos – Bosbarrow

Bu – Buddlesgate Br – Broughton C – Chalton Ch – Chutely Cr – Crondall D – Droxford E Me – East Meon E – Edgegate

34

Ev – Evingar F – Falemere Fa – Fareham Fo – Fordingbridge H – Hurstbourne Ho – Hoddington Hol – Holdshot K – Kingsclere

Hampshire

M – Micheldever Ma – Mainsborough Me – Meonstoke N – Neatham O – Overton

Od – Odiham P – Portsdown R – Redbridge Ro – Rowditch S – Somborne

T - Titchfield W – Waltham We – Welford

As with the case in Wiltshire, discussed at length in chapter 3 above, the disparity can be accounted for by the incidence of beneficial hidation which can be recognised in the Domesday entries for many manors (some of which are explicitly stated, some inferred), combined with the evidence of ‘missing’ hidation values for military service in the unhidated royal manors, many of which, as in Wiltshire, formed estates assessed in renders for the farm of one day or one night. The extent of the operation of these two factors has been examined by Patrick Hase, in an unpublished study which has formed the basis of all the hidation figures which are given in this chapter. This explanation is, as is discussed in relation to the figures for Wiltshire in chapter 3, at variance with those put forward by various commentators who see the Domesday hidation figures as the original assessment, with the implicit and sometimes explicit implication that the Burghal Hidage figures were inflated from this base level as an emergency measure (Sawyer 1978, 227-8; Brooks 1996b, 133-8; Brooks 2003, 159-62; Lavelle 2007, 22; Baker and Brookes 2011, 105 & n.12).

of earlier assessments almost exactly equals the total hides recorded in the Burghal Hidage. It is these hides which would have been used as the basis for the conscription at one man per hide for work on the burhs. It remains to be determined in a systematic manner whether the use of ploughland figures, which I have used to draw the same conclusion in Wiltshire, could also be applied to the evidence from Hampshire, and whether the valuation figures can be used for the same ends in Wiltshire. In Hampshire, however, values for ploughlands are not given on four out of thirteen unhidated estates, which would make a direct comparison with the situation in Wiltshire somewhat less than exact. Hase has suggested (information in this paragraph provided by Patrick Hase – pers com) that the territory of Christchurch (Twynham) was formed by the territory of the major villa regalis at Twynham itself, and another centred on Rockbourne and Breamore to its north, which area was divided at some time between two hundreds (Hase 1988, 65 n.44). The territory of Southampton appears to have been formed from three villae regales (Southampton, Eling and Bishop’s Waltham), though Romsey, which lies between Eling and Southampton, appears to have been within the burghal territory of Winchester, as does the northern part of the Bishop’s Waltham estate (Waltham hundred). The burghal territory of Portchester appears to have been formed from the five villa regalis territories of Portchester, Wymering, Titchfield, Meonstoke and Chalton. Some uncertainty arises over whether Droxford should be assigned to Winchester’s territory (as being held by the monks of the Old Minster (DB Ham 3,2)), or to Portchester. If the former, then this would make up its hidation value from 2354 to 2393, almost exactly its Burghal Hidage figure, but would correspondingly decrease that of Portchester from 490 to 451 hides, even further below its Burghal Hidage figure of 500 hides However, since both figures of 2400 and 500 hides are anyway probably rounded up (or down) to create round-figure totals, to chose between one alternative and another would be somewhat meaningless.

The results of Hase’s researches are summarised in table 10 below. His figures for beneficial hidation, and for the ‘missing’ hidation values in unhidated estates, are derived from a systematic comparison of the recorded hidages with the valuation figures recorded in pounds before the Conquest. On this basis he gives an overall hidation value of the ‘missing’ hides from unhidated royal estates as 517 hides or thereabouts. This is a somewhat different way of achieving the same end of recovering the original hidage figures for military service in the late ninth century from the Domesday evidence as is given for the Wiltshire figures in chapter 3 above. In this shire it is the relationship of the ploughland figures to the hidage figures which is taken as the base indicator both of beneficial hidation and of the ‘missing’ military hidages figures from unhidated royal estates. Hase’s results show, however, that the ‘original’ hidage values of all estates in the shire which can be reconstructed by using the valuation figures as a indicator

Table 9 – hidages of burhs in Hampshire Burh

Hidage in the Burghal Hidage (from Hill 1996a) (not including IoW)

Winchester Twynham (Christchurch) Southampton Portchester

2400

Total

3520

Total Hidage in DB (From Brooks 1996a) (? including IoW)

Total Hidage in DB (Hase – table 10 below) (not including IoW)

470 150 500 2588

35

2240

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Table 10 (see fig. 5). Hidages of Hampshire hundreds Hides given ,TRE (to nearest hide) (not including hides on the Isle of Wight)

Total original Hidage – reconstructed from extent of beneficial hidation and from unhidated royal estates

Winchester burghal territory Romsey Meonstoke Waltham Hurstbourne Evingar

18 52 24 32 108

22 61 34 39 120

Kingsclere

54

60

Basingstoke

55

145

Overton Chuteley Holdshott Hoddington Odiham Crondall Andover Welford Buddlesgate Barton Micheldever Mainsborough Bermonspitt Falmer Fawley Broughton Somborne Bountisborough Bentley liberty Ashley Neatham

76 88 24 18 65 59 105 67 23 42 145 39 43 1 128 90 79 48 13 20 84

88 88 27 18 107 69 213 67 60 109 152 39 50 78 149 169 111 55 13 37 174

Totals

1600

2354

Twynham burghal territory Fordingbridge Ringwood Shirley Edgegate Rowditch Boldre

40 28 28 50 23 48

79 28 28 70 27 88

Totals

217

320

Southampton burghal territory Mansbridge Redbridge

46 29

92 80

Totals

75

172

Portchester burghal territory Bosmere Fareham Droxford

54 31 18

54 41 29

Hundred

36

Comments

North part of hundred

Manor of Kingsclere (unhidated) included with Basingstoke Includes Hurstbourne Tarrant and Kingsclere (unhidated)

Includes Hinton Ampner

Including Nursling

Hampshire

Hundred

Hides given ,TRE (to nearest hide) (not including hides on the Isle of Wight)

Total original Hidage – reconstructed from extent of beneficial hidation and from unhidated royal estates

E Meon Chalton

41 89

78 89

Titchfield

39

45

Meonstoke Portsdown

28 48

63 91

Totals

348

490

Totals for the shire

2240

3336

Comments

Titchfield (unhidated) included in Meonstoke, below) South part of hundred

Table 12: Hidages from Burghal Hidage burhs, and recovered hidages from manors in Hampshire. Burh

Hidage in the Burghal Hidage (from Hill 1996a)

Winchester Twynham (Christchurch) Southampton Portchester

2400

Total

Total hidages recovered from beneficial-hidated and unhidated manors 2354 (2362 with Nursling)

470

320

150 500

172 (164 without Nursling) 490

3520

3336

the value of the Burghal Hidage figure. The probable explanation for this discrepancy lies in the process suggested by Karin Mew. She has made a strong case that the preConquest Foresta Regis in the area around Lyndhurst was extended to the south to form the Nova Foresta immediately after the Conquest by King William, with a resulting reduction in hidage assessment of the affected area (Mew 2000, 159 map 3, 161; see also Finn 1971, 331). As David Roffe has pointed out, the geld was no longer paid on land taken into the forest, so ‘only TRE details of the estate are recorded unless part remained outside the forest.’ (Roffe 2007, 199). It is clear, however, that even these TRE assessments had become adjusted downwards since the time of the assessment given in the Burghal Hidage, very likely for the same reasons. A similar reduction in assessment appears to have been applied to at least two of the manors in the southern part of Redbridge hundred (Gatewood and Stanswood), in the burghal territory of Southampton, which was also included in the area of the Nova Foresta. There are, therefore, good reasons for suggesting that the sum of the hidages of the estates in the six Domesday hundreds in Twynham’s burghal territory would very likely have equalled the 470 hides given to Christchurch in the Burghal Hidage in the late ninth century.

The figures in the table above give a remarkable approximation to the Burghal Hidage figures (table 12 above). The differences in the totals for the territories of Winchester, Southampton and Portchester can probably be accounted for by the uncertainties in the estimation of the precise extent of beneficial hidation, and are insignificant. A more obvious uncertainty is the status of Nursling, included in Southampton’s territory in table 11 above. This was, however, an estate belonging to the Old Minster, and as such included in Buddlesgate hundred in Domesday Book (DB Hants 3,2). Its (reconstructed) 8 hides, although surrounded by Mansbridge hundred, might well therefore have been included in Winchester’s territory. If so, this would balance the totals in the table above to nearer their Burghal Hidage values. Whatever the case, these values are sufficiently near to the Burghal Hidage figures to suggest strongly that the reconstruction of these territories on the map has some basis in the reality of the situation as it existed in the late ninth century. The exception is Twynham, the recorded hidages of whose territory in Domesday Book come to around two-thirds

37

Chapter 5 Warwickshire Warwick and Tamworth

landscape and administrative development, is shown by Warwick (fig. 6), a new burh of the early tenth century. The case of Tamworth, a borough shared by both Warwickshire and Staffordshire to the north, is also relevant to the thesis discussed here.

A further demonstration of the early existence of the pattern of the attachment of urban properties to rural estates, and an example of the use of the stratigraphical techniques in

Table 13. Warwick – customary and non-customary tenements Named estate / manor

No. of tenements

Holder in head section in DB

DB section

The king The king’s barons Hampton Lucy [1] Alveston [2]

3 4

[?Fillongley] [14]

Butlers Marston [3]

2 (+2 frenchmen)

Pillerton Hersey [4]

1

Billesley [5]

1

Bishop of Worcester Bishop of Chester Abbot of Coventry

3,3 [2] [6]

Bishop of Coutances Count of Meulan Earl Aubrey

5,1 [16] [14]

3 1

Budbrooke [8] [?Newbold Comyn] [9]

7

Wolverton [10] Bearley[11]

1 1

No. of tenements in head section

B1 B1

113 112

B2

9

B2 B2 B2 B2 B2 B2

7 36 4 1 12 4 4

18,2 Hugh of Grandmesnil

[19]

B2

1 held from HG by monks of Pillerton 2

[38]

B2

2 (+ 2)

B2

6

B2 B2

2 1 9

18,3

B2

18,14 Henry of Ferrers Harold son of Earl Ralph

Tysoe [6] Edstone [7]

3,1

DB head section

Robert of Stafford Roger d’Ivry Richard the hunter Ralph of Limesy Abbot of Malmesbury William Bonvallet William son of Corbucion Geoffrey de Mandeville Geoffrey de la Guerche

22,4 22,8 [20] [44,7-8] 26,1 [9]

1

[29] 28,17 28,18

B2

1

B2

2

[30]

B2

1

[31]

B2

1

38

Comments

Manors shown in fig 6 4 waste because of the castle

Tenement in Pillerton unassigned

Second listing a mistake – see Notes

Warwickshire

Named estate / manor

No. of tenements

[? Willington] [12]

Coughton [13]

4

Bishopston [15]

1

Long Itchington [16]

1

Holder in head section in DB

DB section

DB head section

No. of tenements in head section

Gilbert of Ghent

[32]

B2

2

Gilbert of Bouille Nicholas the bowman Stephen the steersman Thorki of Warwick

[40]

B2 B2 B2 B2

1 1 1 4

Comments

Charter of 1016 (S.1388) – lay in manor of Stratford (Bassett 2009, 126 n.12) Charter of 1001 (S.898): (Bassett 2009, 126 n.12)

42,3

Osbern son of Richard B2 1 Christina B2 1 The nun Leofeva B2 2 “These messuages belong to the lands which these barons hold outside the Borough and are there valued” 19 burgesses

B3

19

Table 14. Tamworth – non-customary tenements Coleshill (WAR) [1] Wiggington [2] Drayton Bassett [3]

10 4 8

The king The king The king

WAR 1.5 STS 1,9 STS 1,30

Burgesses work at DB “like other villagers”

1971, 336-8; Gelling 1992, 126-31), though arguments are given below to suggest that this happened somewhat later (see discussion in chapters 8 and 13). Working with the hypothesis already formulated in regard to Wiltshire – that the orbits of the contributing estates of Cricklade and Malmesbury reflect the extent of the original burghal territories - the pattern of the contributing estates of Warwick can be seen as indicating the extent of its burghal territory of 914, before it became subsumed by the later shire. On this evidence, this was contained within the old kingdom of the Hwicce and defined by the latter’s borders to its north and east. The burhs at Tamworth and Stafford had been built by Aethelflaed the year before, at which time the burghal territory of Tamworth would have been established to extend within the earlier province (and the see of Lichfield) to its southeast, in the area represented by the later extent of Warwickshire to its northeast (see chapter 8, period 4b in the timeline of development of the territories, and fig. 13). It would be logical to conclude, therefore, that the burghal territory of Warwick was fixed in 914, a year later, to include the area up to its boundary with that of Tamworth on its northern and eastern sides.

As with the case of Malmesbury (chapter 1) the information relating to contributing manors and appurtenant tenements for Warwick is given both in the main entry (DB War B13) as well as in the entries for the rural manors, though these lists do not match up, except incidentally. The tenements appurtenant to the two manors of Bishopstone and Itchingham in early eleventh-century charters, noted by Steven Bassett, are also included (Bassett 2009, 125-6, & n.12). The distribution of the manors which held appurtenant tenements in Warwick is particularly significant (fig. 6). Warwick is placed approximately centrally in relation to the Domesday shire, but – as other commentators have noticed - its 13 named contributing estates recorded in DB are for the most part spread out to its south and west, and contained within the boundaries of the north-eastern part of the more ancient kingdom of the Hwicce which was included in the later shire (Slater 1982, 176 fig 8.1; Slater 1983; Gelling 1992, 156-8; see also Bassett 2009, 153-5). According to the Chronicle (Mercian Register) the burh of Warwick was founded in 914 as one element in Aethelflaed’s defensive provisions for western Mercia, arguably at a royal and minster site of central place and strategic significance (Slater 1983; Bassett 2009). The shire of Warwickshire cannot have been created before Edward the Elder’s supposed reorganisation of c.920 (Stenton

This interpretation is somewhat at variance with that of Steven Bassett, who allocates a rather smaller area to the original burghal territory of Warwick (Bassett 1996, 153 fig. 12). Bassett’s interpretation, however, overlooks the

39

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 6. Contributory manors in Warwickshire, shown in relation to some of the manors contributory to Winchcombe, as in Fig.8.

40

Warwickshire

evidence of the distribution of the contributory manors to the south of Warwick lying to the east of the Avon (see fig. 6). A space also has to be found for the burghal territory of Bridgnorth, built in 912, and situated on the Severn to the north of Worcester on the borders of the later shires of Staffordshire and Shropshire, but within the diocese (and arguably the earlier burghal territory) of Hereford. This is, however, another story. The situation of Warwick on the border of its burghal territory is exactly analogous to the situation of Cricklade on the northern border of Wiltshire, with its contributing manors spread out to its south, discussed in chapter 3 above (see fig. 1). This situation at Warwick is further evidence which supports the model already put forward - that the connections of appurtenant tenements to contributing estates are a relic of arrangements put in place at the initial stages of the formation of the burh by the king - or in the case of Warwick and other burhs in Western Mercia, by Aethelflaed, the First Lady, King Edward’s sister.

126). Unless some reason can be found for the fact that they did not record those manors to the north and east of the earlier boundary of the Hwicce, but only recorded those to the west and south, this reasoning must be regarded as being somewhat tendentious. This is arguing from inferences based on negative evidence, and appears to be contradicted by the evidence of those manors in this area which have survived to be recorded. Bassett’s interpretation of this distribution, and in particular of the connections shown by the evidence of the two early eleventh-century charters, is that at this time Warwick was ‘a thriving commercial centre in which rural landowners saw value in having a tenurial foothold’ (ibid.). This may well have been so, but does not in itself provide an adequate explanation for the origins of these connections. This represents a restatement of the long-held paradigm originating from the views of Tait (above), which it is the purpose of this study to reassess. The origin of these connections outside the original area of the Hwicce must be as early as the gift of these manors to the abbey of Coventry in 1043, but it could be argued that this pattern of distribution reflects connections between Warwick and the various manors acquired by Coventry Abbey which had a somewhat earlier origin. In support of this is the strong probability, argued by Stephen Bassett, that Coventry itself was the site of an early minster which had developed its own market well before the Conquest (Gelling 1992, 158-9; Bassett 2001, 31-3). That the earl’s, and latterly the abbey’s, manors held appurtenant tenements at Warwick at the time of Domesday, rather than at its own market in Coventry itself, emphasises that these connections cannot have been established as a result of the proximity of these manors to the nearest market, as would be suggested in the current explanatory paradigm, especially in view of the strong drawing power of the abbey of Coventry as an administrative and marketing centre for its area.

While there is every reason to hold that this was a primary pattern relating to the foundation of the burh in 914, the distribution of other manors held by the abbot of Malmesbury ([9] in fig. 3), the bishop of Coutances ([14] in fig. 6), the manor of Long Itchington [15] which held a tenement in Warwick in 1001, and those of the abbot of Coventry (un-numbered in fig. 6), all of whom held tenements in the borough, suggest that these had a different origin (for Long Itchington, see S.898). The connections between these urban tenements and manors are inferred from the fact that the abbot of Malmesbury and the bishop held only one manor each, and are stated in the description of the borough to have each held one tenement, while the abbot of Coventry held the unusually large number of 36 tenements in Warwick. These latter can only have been appurtenant to some or most of the abbey’s 20 estates within the later Domesday shire. (The manor of Clifton-uponDunsmore, only given to the abbey in the mid-eleventh century (DB War 6,9) should be excluded from this total, though it is possible, indeed likely, that this manor came into the abbey’s possession with an urban manor already attached). However, the fact that all of these (with the exception of only two of the abbot of Coventry to the south and south-west of Warwick) are outside the primary burghal territory in the area of the Hwicce postulated above, can be arguably best interpreted by the suggestion that they were added to an original pattern represented by those within the area of the Hwicce to the south and west of Warwick. This is a further example of layering, though in a horizontal rather than a vertical plane, as noted above in the case of Wiltshire, and in the case of Gloucester and Winchcombe described in detail below (chapter 6).

The logic of these arguments implies that their origin can most reasonably be placed within the context of the occasion when Warwick became the military and administrative centre of the enlarged shire, which process is argued below (in chapter 8, see also chapter 13) as having taken place in the third quarter of the tenth century. It must have been at this time that the estates in the enlarged shire would have been confirmed to their holders as bookland by the king, with the consequent responsibilities to perform obligations at the administrative centre of the new shire. If this is so, then the connections of these manors with Warwick had already been formed when they were acquired by the earl or his predecessor, and before he gave them to Coventry abbey. The implication is that any attachment of manors in this area to the earlier burh at Tamworth had been superseded in the process of the political reorganisation represented by the formation of the new shire centred on Warwick. In spite of this process, however, the connection between the royal manor of Coleshill within this enlarged shire with Tamworth still survived (fig. 6). This process is discussed further below.

Steven Bassett has, however, suggested an alternative explanation for this distribution, to the effect that the supposed connections to Warwick of manors originally lying within the diocese of Lichfield, outside the territory of the Hwicce, have not survived through the failure of the Domesday commissioners to record them (Bassett 2009,

41

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

The case of neighbouring Tamworth can also be interpreted in the same way. The contributing estates are located to the north, now in Staffordshire, and to the south, now in Warwickshire, with the shire border famously bisecting the original burh (Gelling 1992, 152 & fig. 58). Since the burh at Tamworth was founded or refounded in 913, before the creation of the two shires, the pattern of distribution of the contributing estates can be interpreted in the light of the hypothesis already stated – that they represent the surviving vestiges of tenurial connections within a burghal territory which existed before the formation of the shires. As argued above, this must have extended over north Warwickshire and south Staffordshire until it was subsumed by the later arrangements attendant on the formation in their final form of the shires of Warwickshire and Staffordshire, which led to the boundary between them being cut through the town. That these burgesses were all appurtenant to royal estates is consistent with the suggestion that they are vestiges of a much earlier arrangement in which connections were established between outlying royal manors and the central tun, possibly before Tamworth was first created as a burh, very much in the manner in which connections were formed between estates in Wiltshire between the central tun at Wilton and outlying royal manors, discussed above. As with the case with Warwick, the inference can be made that the attachment of estates to the burh of Tamworth of 912 within the later Staffordshire had been subsumed in the creation of the shire and transferred to Stafford as the new shire capital.

In the case of Warwick and Tamworth, this evidence is best explained, at least in the West Midlands, by the hypothesis that the burghal territories represented an intermediate state of territorial and administrative development and organisation between the provinces or regiones and the shires. The question of the origin of the shires is complex and contentious, and will be discussed in chapter 8 below, and more fully in chapter 13. One strand of ideas – a constantly-recurring paradigm in any discussion of this issue – is that the shiring did not take place until the early eleventh century (Taylor 1898). This is, however, premised on the fact that this time marks the first documentary mention of shires. It also does not take account of the likely development of institutions in the previous decades. Another seemingly more persistent strand of ideas is that Edward the Elder reorganised the shires of the western Midlands soon after the death of his sister Aethelflaed in 919 (Stenton 1971, 336-8; Gelling 1992, 137-45; Whybra 1990; Hill 1996b, 94; Hill 2000, 174; Hill 2001; see also the discussion in Bassett 1996). However, the overall conclusion from a range of evidence is that the most likely period in which the shires were brought into being was in the 960s, immediately consequent on the division of the southern part of England between the two kingdoms of Mercia and Wessex in 957, and put in place by Aelfhere, ealdorman of Mercia from 956 to 983. This is discussed in more detail in chapter 13, taking into account the evidence for the divisions in the upper Thames area.

42

Chapter 6 Gloucestershire The cases of Gloucester and Winchcombe, the two centres in the shire (apart from Bristol in the south) with tenements appurtenant to rural estates, illustrate both the apparently haphazard way in which the orbit of contributory estates of the one overlaps that of the other, as well as the way in which some estates contribute tenements to both centres. The relationship of the respective contributing manors to the shire boundaries of Gloucestershire and Winchcombeshire (fig. 7), and of Warwickshire to the north-east (discussed in chapter 5 – see fig. 6), are also particularly complex. These

patterns are, however, explicable by reference to the model already put forward in relation to the situation in Wiltshire. Gloucester The haga on port appurtenant to the manor of Dumbleton, which is mentioned in the bounds of a charter of 1002 (S.901), without reference to which port, has been assigned to Gloucester rather than Worcester on the basis of its proximity to the former, as well as its inclusion within the

Table 15. Gloucester – non-customary tenements Manor

Number of burgesses / tenements

Holder of manor

DB Section

The king The king Ch of Worcester Ch of Glastonbury

1,24 1,47 3,5 8,1

-

St Peter’s Gloucester

10,14

Broadwell [5]

4

12,4

Deerhurst [6]

30

Horsley [7] Bisley [8]

1 11

St Mary’s Evesham St Denis, Paris (part of larger royal multiple estate before 1059) Church of Troam Earl Hugh

Oxenhall [10]

3

Roger de Lacy

39,2

Temple Guiting [9] Quenington [11] Gt Rissington [12] Brimpsfield [13] Frampton on Severn [14] Bully [15] Lechlade [16] Kempsford [17]

2 1 1 5 1 1 1 7

39,6 39,12 46,1 50,3 54,1 58,1 59,1 60,1

Woodchester [18]

1

ditto ditto Robert of Tosny Osbern Giffard Drogo son of Poynz Walter the Bowman Henry of Ferrers Arnulf of Hesdin The king – held by Brictric (also holds Leckhampton – 78,9

Kings Barton [19]

24

Dumbleton, (Glos) [20]

1 haga

Tewkesbury [1] Thornbury [2] Withington [3] Pucklechurch [4]

8 A fishery 4 1

The king

20,1 24,1 28,1

78,14 1,2 13,1. 34,13.

King Aethelred

43

Comments ‘served at the court’

Held 52 burgesses in Ev K1. Render of `6 salmon + 50s from the burgesses – for the monks’ supplies 1 also in Winchcombe 2 also in Winchcombe. Also noted in EvK1 as 36 burgesses Also noted in EvK1 Successor Hugh held 28 tenements in Ev K1 3 also in Winchcombe Paid in ploughshares

2 also in Winchcombe

Paid in horseshoes [pieces of iron] From EvK 1 (below) (burgesses not mentioned in DB) S 901 (1002) (? in Gloucestershire)

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Table 16. Tenements and manors from Gloucester in Evesham K1, additional to the manorial entries in table 15 Holder of tenements Archbishop of York The abbot Abbot of Pershore Hugh of Lacy [brother of Roger – Hugh received Roger’s fee in 1096] William son of Baderon

Number of tenements

Number of manors held in DB

section ref in DB

60 52 1

13 14 2

2 10 14

28

21

39 [Roger]

17

13

32

Domesday shire. Its connection with Worcester is however suggested by Baker and Holt (Baker & Holt 2004, 264).

amongst others, of a named holding in Longborough, and of the 40 tenements held by the abbot of Winchcombe (Moore 1982, Appendix). The abbot’s tenements can be inferred to have been appurtenant to many or most of his 14 manors in the area of Winchcombeshire, which are therefore included in fig. 8a.

The information in the DB folios concerning Gloucester is considerably expanded in a list of tenements in Gloucester in one of the Evesham satellites (Evesham K1) of c.1100 (Moore 1982, Appendix). This information not only supplies the number of customary burgesses (‘300 burgesses in lordship’), but also appears to enlarge the list of non-customary burgesses to a possible total of 301. Entries relevant to the present thesis are included in the table 16, above.

Bristol Bristol lies on the southern border of Gloucestershire (fig.4), but has burgesses appurtenant to manors in both Gloucestershire and Somerset.

Since those in the list in Evesham K1 who held most of the tenements also held the greatest number of manors in the shire, it is reasonable to infer some correspondence between the two. With this inference in mind, the manors held by these five tenants-in-chief the are given in figs. 8a & b, in order to give some idea of the possible spread of other un-named manors contributing to Gloucester. The geographical spread of manors held by the abbot in the shire also appears to reflect the widespread distribution of the abbey’s tenements which are intermingled amongst those paying landgable to the king within Gloucester itself, as shown in the survey of 1455 (Baker & Holt 2004, 279-81), giving some support to the hypothesis that these Gloucester tenements were appurtenant to some or all of these manors. This is discussed further in chapter 9, and shown in fig. 17b. The reference to the 52 tenements held by the abbot in EvK1 is apparently reflected in Domesday Book, to the effect that ‘Before 1066 St Peter’s had 19s 5d and 16 salmon from its burgesses in Gloucester; now it has as many salmon and 50s’, as well as having four fisheries of its own (DB Glouc 10,14) (Moore 1982, Appendix). To this sum might be compared the sum of £10 17s 6d which the abbess of Wilton had from the borough of Wilton (DB Wilts 13,22), much of which must have come from the rents of tenements appurtenant to the abbey’s estates (chapter 3 and fig. 1).

Discussion - the pattern in Gloucestershire There are a number of significant aspects of the distribution of the manors with appurtenant tenements in Gloucester, Winchcombe and Bristol. These can best be interpreted in the light of the model formulated above. The first is that the distribution of these manors in relation to their centres is not consistent with what may be called the ‘market-proximity’ hypothesis – that the holders of the manors involved will have sought to acquire tenements in the nearest or the most conveniently-situated borough, for economic or other reasons. This is demonstrated, in the first place, by the way that many of the manors contributing to Winchcombe are overlapped by those contributing to Gloucester. Those with connections with Gloucester are spread more-or-less evenly around the shire, whereas those with connections to Winchcombe are distributed around all sides of the borough but limited to the area of the tenth- and eleventh-century shire of Winchcombeshire which occupied the north-eastern part of Gloucestershire. Only Lechlade to the south-east [12] falls outside the latter (see Wybra 1990 for the shire boundaries; compare with Hill 1981, 99). This lack of any correlation between the distance and the connection of manors to their centres is also emphasised by the fact that almost all the manors contributing to Gloucester which lie to its south are nearer to other boroughs than to Gloucester. Three in the south-east - Quenington [11], Kempsford [17] and Lechlade [16] - are considerably closer to both Cirencester and Cricklade; and Woodchester [18] and Horsley [7] are nearer to both Malmesbury and Tetbury, as well as the two Domesday boroughs at Thornbury and Berkeley. Similarly, Pucklechurch [4] to the south is far closer to Bristol, Bath and to the Domesday borough of Thornbury, as well as to Malmesbury and Tetbury.

Winchcombe The estates contributing to Winchcombe are noted in the entries under the manors. As with Gloucester, the meagre information concerning Winchcombe is supplemented by further information in the Domesday satellite Evesham K116, which gives details,

44

Gloucestershire

Table 17. Winchcombe – non-customary tenements Manor

Number of burgesses / tenements

Holder of manor

DB Section ref.

Comments

Tenements also in Gloucester

Oxenton [1] Alderton / Dixton / Hentage [2] Withington [3] Prestbury [4]

3

king

1,25

1

king

1,43

1 1

3,5 4,1

4

Broadwell [5]

1

12,4

4

Deerhurst [6]

2

20,1

36

Clopton [7]

1

Guiting Power [8]

2

Temple Guiting [9] Hampnett [10] Childswickham [11] Lechlade [12] Pinnock [13]

2 10 1 2 1

Ch of Worcester Ch of Hereford St Mary’s of Evesham St Denis, Paris Willian Goizenboded Willian Goizenboded Roger de Lacy Roger d’Ivry Robert the Bursar Henry of Ferrers The king

36,9 41,1 47,1 59,1 78,10

From Ev K116 Longborough [14]

3

Count of Mortain

29,1

40

The abbot of Winchcombe

34,3 34,8 3

1

Distribution of manors shown in Fig. 8a

11

Table 18. Bristol – customary and non-customary tenements Manor

Barton Regis [1] Westbury on Trym [2] Bishopsworth [3]

Number of tenements / burgesses Not stated – but a considerable number 2 10

Holder of manor

DB section

Roger of Berkeley / the king

1,21

Ch of Worcester

3,1

Bishop of Coutances

Thornbury itself is nearer to Berkeley, Bristol, Bath, Tetbury and Malmesbury than to Gloucester. A different kind of anomaly is represented by Thornbury [2], itself a market in Domesday yet with an appurtenant holding (albeit a fishery) in Gloucester, and also by Tewkesbury [1] to the north of the shire, a Domesday borough with 13 of its own burgesses yet with 8 burgesses in Gloucester. And not least, the two manors of Oxenhall [10] and Broadwell [5] are on the eastern side of the shire on the other side of Winchcombe itself. Any explanation for the origin of these connections as resulting from the attraction of estate holders to ‘neighbouring’ markets or burghal centres is, quite simply, untenable.

SOM 5,20

Comments Probably customary burgesses in Bristol appurtenant to the royal manor. Large multiple estate Manor also has 2 houses in Bath

Westminster and the St Denis in Paris, in approximately equal proportions (Taylor 1902, 230-6). The part given to Westminster contained the capital manor, but was less valuable than that given to St Denis (59 hides and 64 hides respectively). It is significant therefore that the tenements in Gloucester, which would have been appurtenant to the manor as a whole before its subdivision, were included not in the portion given to Westminster which had the capital manor, as would be expected, but with the portion given to St Denis. This implies a definite decision to divide the estate in this way. It also demonstrates that the 30 tenements recorded in Domesday were not appurtenant to individual dependencies within the original multiple estate of Deerhurst, but had originally been appurtenant as a single group to the capital manor itself, from which they were divided. In the ninth century, however, this group would have occupied a single haga, subsequently subdivided. This has implications for the interpretation of other similar or analogous situations in other shires. It is also shown by

Also of some interest is the holding of 30 burgesses recorded in Domesday Book (which had grown to 36 in EvK1) in Gloucester which were appurtenant to Deerhurst. This was an extensive multiple estate which had been granted by the king in c.1059 to two monastic institutions – St Peter’s

45

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 7. Gloucestershire (including Winchcombeshire) showing manors contributory to Gloucester, Winchcombe and Bristol. These are shown in relationship to those contributory to Cricklade and Malmesbury in N Wiltshire (from Fig.1).

hands (Heighway 1984, 364, & fig. 118; Baker & Holt 2004, 116). The position of this haga is compared to that of Staeningahaga and Basingahaga in London (Baker & Holt 2004, 232). It is suggested that this can best be explained by the general hypothesis argued in this study, that this area, as probably a single haga, was given over as an tenement appurtenant to Deerhurst as part of the primary organisation of the new burh – a conclusion which is consistent with the arguments of Baker and Holt against Heighway’s unlikely proposition that this haga had even earlier origins (Baker &

later evidence that the group of 30 burgesses occupied the north-east corner of the walled area of Gloucester, which area later became the parish of St Aldate, whose church belonged to Deerhurst Priory which was itself in the hands of St Denis (Heighway 1988, 11; Baker & Holt 2004, 1157). That this area had no plots paying landgable in the 1455 rental, (apart from four on the Northgate Street frontage) shows that it originally comprised a single urban haga held by Deerhurst rather than the king, notwithstanding the fact that Deerhurst itself had been a multiple estate in royal

46

Gloucestershire

Holt 2004, 116-7). It is also consistent with Baker & Holt’s suggestion that the church of St Aldate was a private chapel attached to a ‘prominent house’, which may be inferred to have been the abbey’s capital messuage, arguably formed at the time of the setting out of the burh.

estates contributing to Winchcombe can be seen as evidence which helps to define the lost shire of the early eleventh century (Whybra 1990, 24-30). A further inferential step would suggest that the burghal territory of the burh of the late ninth century, which the archaeological evidence would imply, extended to the north-east to the borders of the Hwicce (shown in fig. 11), and that the north-eastern part of this territory was subsequently decreased in extent by the formation of the territory of the new burh of Warwick in the early tenth century (shown in fig. 13).

Less clear-cut, however, are the 24 burgesses mentioned in EvK1 (but not mentioned in Domesday) who were appurtenant to the 9-hide royal manor of Kings Barton. This manor or later liberty had its origin as an estate which probably originally included the royal palace of Kingsholme, immediately to the north of the borough, in the late Saxon period (Heighway 1988, 9). The manor, together with Kingsholme itself, was divided between St Mary de Lode and the original St Oswald’s parish, which area ‘may represent the survival of the original territory with which the Old Minster was endowed in 679’ (Heighway 1988, 9-11; see also comments on the royal palace at Kingsholm in Baker & Holt 2004, 19-21). It might reasonably appear at first sight that the 24 burgesses appurtenant to this manor could represent the occupants of an early haga or soke within the walls which was in royal hands, in a similar relationship to the haga which was appurtenant to the former royal centre at Deerhurst, and which may therefore have been created at an early stage – or indeed the primary stage - in the formation of the burh at Gloucester. This arrangement would mean that these burgesses were noncustomary, but this would imply that they paid their dues to the same royal centre as the customary burgesses. These burgesses must therefore be seen as customary burgesses, even though appurtenant to the royal centre outside the burh itself. This situation has a direct analogy to the relationship, already discussed, of the burgesses of the royal manors of Cheddar and Somerton, Somerset, who lived at the burhs of Axbridge and Langport respectively. This being so, it may be that the tenements appurtenant to this manor were distributed around the area of the burh, forming some of the customary tenements paying landgable in the survey of 1455 (Baker and Holt 2004, 277-81, 326-8).

This conclusion is lent some support by the distribution of the manors of Gloucester Abbey, two of which lie within Winchcombeshire to the north of Winchcombe, which are overlapped by those held by Winchcombe Abbey which are limited to the area of Winchcombeshire (see fig. 8a). This overlapping is also seen in the case of Tewkesbury, which had 13 burgesses of its own, yet also had eight others contributing to Gloucester (DB Glouc 1,24). It would be appropriate to see this as a close parallel to Calne (Wilts) discussed above, in which the Domesday pattern had developed in two stages. This evidence is consistent with a sequence in which the burgesses ‘at’ Tewkesbury were the tangible sign of earlier connections which reflect its early functions as the centre of a large multiple estate, which functions had been overlain by those newly created on the occasion of the foundation of the burh at Gloucester in c.880. The patterns at Calne and Tewkesbury also seem very similar in this regard to that shown by Milborne Port in Somerset. It is clear from the wording of the Domesday entry (DB Som 1,10) that the royal manor at Milborne Port showed heterogeneous tenure in that it had a flourishing market of its own with its own contributing manors, but also had 107 burgesses who were ‘at’, or were contributed to (i.e. living at) the neighbouring burh of Ilchester, with the remaining 56 burgesses living ‘at’ Milborne Port itself (Haslam forthcoming d). The underlying premise which can be deduced from these relationships is that these connections were put in place at the time of, and as a result of, the establishment of Gloucester and Winchcombe as significant administrative centres and / or as fortified burhs (the latter implying the former); and that the connections shown in the Domesday record are the survivals of a pattern of distribution of manors whose connections with the centres related to the functions of these places within territories which looked to these centres for defence and other administrative and marketing functions – i.e. their burghal territories. The distribution of connections of estates contributory to Winchcombe and to Gloucester can be readily interpreted by reference to the overall developmental model discussed further in chapter 8 below. It is argued that both Winchcombe and Gloucester were developed as burhs in the phase of burh formation of c.880 (characterised as stage 2b in the timeline, discussed in chapter 8 below), replacing earlier burghal functions at Winchcombe which had been put in place in the late eighth or early ninth century in stage 1b. The extension of the orbit of connections to Gloucester over

The explanation of these spatial patterns relating to the manors contributing to Gloucester, Winchcombe and Bristol, as well as Warwick to the north-east, discussed in detail above, must lie in the overlapping and superposition of the orbits of connection which had been formed at different times. These can most easily be envisaged as separate layers which represent the different stages in which the attachment of estates to these four centres occurred in the development of the urban, administrative and strategic foci in this area of western Mercia. The basic pattern of distribution might be taken as suggesting that the orbit of manors contributing to Gloucester, which covers the greater part of the shire, has been overlaid by a less extensive and therefore secondary orbit of those contributing to Winchcombe. However, the evidence of the historical primacy of Winchcombe, and of its development as a burh from the late eighth or early ninth century (stage 1b in the timeline set out in chapter 8 below; see also Bassett 2008a; Bassett 2011) would seem to indicate that the opposite was the case. The orbit of the

47

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

much of the area of later Winchcombeshire can be seen as the result of the development of Gloucester as the ‘shire capital’ and the final disappearance of any administrative functions of Winchcombe in the early eleventh century, which is also argued in chapters 8 and 13. As an alternative reconstruction, it may well have been the case that the earlier burghal territories of Gloucester and Winchcombe had remained intact in the process of shire formation in the 960s, the two areas effectively operating as ‘shires’ in the developed sense, and that Winchcombe had lost these functions to Gloucester only in the early eleventh century through the activities of Earl Streona. In this process the connections of estates to Gloucester would have been augmented by new connections over the former burghal territory and shire of Winchcombe, in very much the same way as has already been suggested was the case with the extension of the early burghal territory of Warwick to form the later shire (see chapter 5). This at once supplies the explanation for the fact that, in this instance, five manors – Broadwell [Gl 5 / W 5], Deerhurst [Gl 6 / W 6], Withington [GL 3 / W 3], Temple Guiting [Gl 10 / W 9] and Lechlade [Gl 16 / W 12] - share appurtenant tenements in both places (see tables 16 and 17), which can in each case be interpreted as the development of connections to Winchcombe to which were added connections to Gloucester in perhaps the early eleventh century (stage 6).

distribution of these manors is emphasised by the fact that the easternmost manors contributory to Gloucester are in fact somewhat nearer to Oxford (Blair 1994, 118 fig. 69, and see chapters 7 and 8 below). Following the timeline given in chapter 8 below, it can be inferred that this border must therefore mark the common boundary of the original burghal territories of Oxford and Gloucester. A similar and perhaps more informative set of relationships is shown by Bath, lying to the north of the river Avon but with contributory manors on both the north and the south of the river, as well as manors to the north of the river held by Bath Abbey (Finn & Wheatley 1967, 198 fig. 47). The area to the north of the river had been in Hwiccian territory probably until King Alfred requisitioned Bath as a burh in his defensive system for Wessex of 878-9 (Aston 1986, 54 fig. 7.3; Manco 1998; Haslam 2005, 130 & n. 52), which is described in the arguably contemporary Burghal Hidage (Haslam 2005, 135-48; Haslam 2009, 967). These relationships can be interpreted according to the timeline given in chapter 8 below, to the effect that the burghal territory of Gloucester, formed soon after in c.880, would not have included the area which already formed the burghal territory of Bath. However, the burghal territory of the secondary burh at Bristol, comprising the south part of later Gloucestershire which is arguably indicated by the area known as the seven hundreds of Grumbalds Ash (Cam 1963, 96), must have been carved out of that of Gloucester to its north and that of Bath to its east at a later date (arguably in the secondary phase of burghal formation in the 890s - Haslam 2009, 98-100, 103-4; Manco 2008 – stage 3 in the timeline set out in chapter 8 below). This would have comprised an area on both sides of the Avon (of which its contributing manors give some indication), which would appear likely to have been swallowed up in the extension of the shire from the original burghal territory of Gloucester either in the mid tenth century (stage 5) or the early eleventh (stage 6).

The orbit of manors contributing to Winchcombe also suggest the possibility that some of the connections may well have been inherited from the earlier phase of burghal formation in the late eighth or early ninth century (stage 1b in the timeline set out in chapter 8 below). Lechlade [12] is the clearest example; others might include Deerhurst (its connection with Winchcombe formed at this time by the evidence of its inclusion within Winchcombeshire – Whybra 1990, 24-7) which was overlain by a new connection with Gloucester in stage 2b. That the haga contributing to Deerhurst occupies a primary position in the layout of the burghal space of Gloucester, discussed above, suggests that this connection was formed as a result of the primary phase of the internal organisation of the burh. This methodology of landscape stratification has been shown to be explanatory on a number of different levels for similar overlaps in the orbits of distribution of these connections in Wiltshire and Warwickshire, discussed above, and in Worcestershire (between Worcester and Droitwich) examined below. Similar evidence has been used by David Roffe to indicate the burghal territories of Wallingford and Sashes at the time of the King Alfred’s provision of the system of fortified burhs in Wessex of the late ninth century (Roffe 2009, 424). This is discussed further in chapters 10 and 11 below.

A significant implication of the general inference, that the distribution of the contributory manors gives an indication of the minimal extent of the original burghal territory of the centre, is also shown by the presence of two named manors contributory to Gloucester which lie to the west of the Severn - Oxenhall [10] and Bully [15] (fig. 7). Their presence in this area is consistent with the inference that some or all of the territory of the later shire to the west of the Severn, which has been suggested as being transferred from Herefordshire in c.1007 (Finberg 1972b; Moore 1982; Whybra 1990, 87 map XVI), possibly belonged to the original (late ninth-century) territory belonging to Gloucester. This is perhaps supported by the distribution of the manors of both Roger de Lacy and William son of Baderon, as well as those of Gloucester Abbey and the Archbishop of York (representing the earlier holdings of St Oswald’s Priory), who between them held 15 manors west of the Severn (Figs. 5a & 5b), some or all of which are likely to have contributed some of their respective holdings of 28,17, 52 and 60 appurtenant tenements in Gloucester.

It is also consistent with this overall model that several manors contributory to Oxford, which are approximately equidistant from Oxford and Winchcombe, lie in Oxfordshire just to the east of the Gloucestershire / Winchcombeshire border, and therefore to the east of the border of the kingdom of the Hwicce (see fig. 3). The importance of the shire boundary in determining the overall

48

Gloucestershire

Fig. 8a. Gloucestershire (including Winchcombeshire) showing manors of major ecclesiastical tenants-in-chief, which were possibly connected with their large holdings of tenements in Gloucester and Winchcombe.

Fig. 8b. Gloucestershire (including Winchcombeshire) showing manors of lay tenants-in-chief, which were possibly connected with their large holdings of tenements in Gloucester.

49

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Although consistent with this inference, the distribution of these is however of limited evidential value. All of these tenants-in-chief, apart from the archbishop of York, held manors in Herefordshire (Roger de Lacy 75 manors (DB Her 10,1-75); Wm son of Baderon 10 manors (DB Her 15,1-10); Gloucester abbey 2 manors (DB Her 5,1-2), of which one, Lea, had been a gift after 1066. However, the fact that the archbishop of York held no manors in either Worcestershire or Herefordshire at the time of Domesday does suggest that his one manor to the west of the Severn would have contributed to at least one of his appurtenant tenements in Gloucester.

5) or the early eleventh (stage 6), to include the eastern part of the former burghal territory and shire of Hereford, which lay in the area of the former Magonsaetan and diocese of Hereford (Hill 1981, 81 map 143). This would be analogous to the formation of connections with Warwick of those manors to its north and east in possibly the mid tenth century, discussed above. If this is so, it is perhaps some support for this hypothesis that the process of the reorganisation of the shires in the early eleventh century – if this was indeed the case - is likely to have been associated with a general military reorganisation (including the reformation of earlier burghal territories) in Aethelred’s reign, in perhaps the first decade of the eleventh century (Haslam 2011b). It is reasonable to suggest, therefore, that as a result of this process estates to the west of the Severn became attached to Gloucester rather than Hereford, for the same reasons which gave rise to the attachment of others to Gloucester in the late ninth century.

Rather than seeing this pattern as the development of one period, however, it is perhaps more likely that these manors to the west of the Severn acquired their appurtenant tenements in Gloucester on the occasion of the probable extension of the shire in either the mid tenth century (stage

50

Chapter 7 Worcestershire Worcester and Droitwich

as non-customary burgesses. The 31 burgesses included in the lands of Pershore in Droitwich, held by the abbey of St Peter’s at Westminster (DB Wor 8,13), would have performed a very similar function, though these would have been probably more specialised than the 28 burgesses attached to the market at Pershore itself (DB Wor 8,1). This pattern at Pershore is very similar to the examples of Tewkesbury (Glos) and Calne (Wilts) noted above.

As with those of Gloucester and Winchcombe, the orbits of the distribution of the estates contributing to Worcester and Droitwich overlap (fig. 9). As is apparent from fig. 9, those contributing to Droitwich occupy a more restricted distribution in the northern part of the shire of Worcestershire, while those of Worcester are spread over its entire extent, completely overlapping the orbit of the estates which are contributory to Droitwich.

Three of the contributing estates of Worcester and Droitwich, like some of those of Winchcombe and Gloucester, are shared between these two places. There is, however, no evidence which would indicate that Droitwich was developed as a burh of the late Saxon model at any time (see an analysis of the development of the borough in Bassett 2008b). The estates which contributed burgesses to Droitwich in Domesday Book appear to have been of a kind which have already been discussed as contributing to other early royal estate centres, such as Calne (chapter 2 above). These can be distinguished from the far more numerous connections which represent holdings of rights in, or shares of, salt-works by various estates in the west and central Midlands. Many of the latter were purely commercial arrangements which were usually set up as mechanisms whereby salt was exchanged for a supply of wood (Hurst 2004), and were evidently regarded as assets which were granted by the king to various interests, including the bishop of Worcester, from the late seventh century onwards (Hooke 1981, 137-42, 154-9; Hopkinson 1994, 30-7; Maddicot 2005, 28-31). However, in the case of Droitwich, as contrasted with other instances, it is difficult to separate the holdings of those in rural manors, who at an early stage in their development would have performed services at the centre, from those which appear to have acquired a purely commercial function.

The holdings in Droitwich are somewhat different in form (but arguably not in kind) to other places. Except for those mentioned in sections 1,3 and possibly 11, the customary burgesses of the king, so evident in other places, appear to have been acquired by various tenantsin-chief – presumably as grants from the king – together with parcels of land ‘in’ Droitwich (half a hide or 1 hide each), as well as various salt-works. Even the former royal manor of Wychbold had by the time of Domesday Book been privatised (or mediatised) in this way, being held before the Conquest by Earl Godwine (Hooke 1981, 129). These burgesses appurtenant to Wychbold, who reaped and mowed for the lord and served in the court, are similar to those other burgesses in Hereford, Steyning (Sussex), and Drayton Bassett near Tamworth (above) who also performed the same services. They must also be very similar to the free men and others who performed services of one kind or another, including reaping and mowing, at the court of the multiple estate of Pershore (e.g. DB Wor 8,9b; 8,11; 8,17; 8,23). The three salinarii from the large multiple estate of Bromsgrove (Hooke 1981, 129) have been identified as including one salinarius from the king’s manor at Princes Risborough - though on no good reason except an identity of title (Hopkinson 1994, 32). More probably, however, these were acting as agents of the king in handling the distribution of the salt from his saltworks. There is no reason, therefore, to look for any other explanation than that they were specialised burgesses in the service of the king at the manor of Bromsgrove who were no different in kind to burgesses who lived in Droitwich performing services (doubtless variously as agents, procurers or middle men of one kind or another, or possibly even as salt producers) on behalf of the tenants-in-chief of other manors – i.e.

Droitwich was an early royal centre, which developed in relation to a royal vill at nearby Wychbold (Hooke 1981, 129; Cambell 2003; Maddicot 2005, 32-3), which seems likely to have functioned in establishing and consolidating royal control over the salt workings from the seventh century. A large share of the rights to salt-working was still in royal hands by the time of Domesday, in spite of several royal grants (Hooke 1981, 137). The connection is further evidenced in the mention under Wychbold of 13 burgesses contributing to Droitwich, who had to do service at the lord’s court at Wychbold. The distribution

51

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig 9. Worcestershire, showing estates attached to both Worcester and Droitwich in the 10th and 11th centuries.

of the contributing estates around Droitwich would be consistent with the hypothesis set out above which sees the connections of these contributing estates to the centre being established at an early date in its development. Several charters of the seventh to the early ninth centuries record transactions in which salt pits or salt houses appear to have been appurtenant to estates held by the king. These include a reference to salt pits appurtenant to a 50-hide estate at Hanbury in 657x75 (Hooke 1981, 129, 152; Maddicot 2005, 28-9), a reference to a shed and salt pits appurtenant to Fladbury given by the king to the bishop of Worcester in 691 (Maddicot 2005, 29), and the charter of Aethelbald of

716x717 (S. 97) granting part of a building in Droitwich to Evesham Abbey (Hooke 1981, 125; Maddicot 2005, 39-41). That salt works and/or houses were appurtenant to these estates could be interpreted as early instances of the obligations of estates in the Droitwich area to maintain operations at the central saltworks as part of their provision of services and renders to the king, notwithstanding the fact that the provision of salt-rights were an aspect of royal patronage for monastic communities (to which these estates were given) for whom the provision of salt would have constituted an essential aspect of their economies (Maddicot 2005, 31). This general process which sees the connections

52

Worcestershire

Table 19. Droitwich – non-customary tenements Manor

Number of burgesses / tenements

Bromsgrove [1]

3

Holder of manor

DB Section ref.

Comments

The King

1,1a

3 salinarii (+ 13 salt-houses) (see below)

Droitwich

18 (with 1 hide in D)

St Denis

4,1

burgesses

Pershore [13]

31

St Peter’s Westminster

8,13

burgesses

St Guthlac

12,1

burgesses

Roger de Lacy

18,6

burgesses

Harold son of Earl Ralph

22,1

burgesses

The king

1,3a

Houses – probably the equivalent of customary burgesses

5,1

burgesses

Droitwich Droitwich Droitwich Droitwich

9 (with 1 hide in D) 11 (with ½ hide in D) 20 (1 hide held from the king) 11

Salwarpe [2]

4 (with 1 hide in D)

St Mary’s Coventry

Elmley [3]

5 houses

Ralph Tosney

15,13

houses (separate entry from salthouses) burgess (centre not specified, but also holds 1 salt-house burgesses “who reap (?mow) for 2 days in August and March and who serve the court.” burgess (centre not specified, but also holds 2 salthouses) burgesses burgess

Morton Underhill [4]

1

Robert of Stafford

17,1

Wychbold [5]

13

Osbern son of Richard

19,12

Crowle [6]

1

Osbern son of Richard

19.14

Witton [7] Cookhill [8]

7 1

Urso d’Abetot Urso d’Abetot

26,16 26,1

Hallow [9]

10

Church of Worcester

2,68(-71)

houses

Hartlebury [10]

5

Church of Worcester

2,82

houses

Kidderminster [11]

1

The king

1,2

house

Northwick & Tibberton [12]

3

2,50

houses

Dunclent [**]

9

12,1

burgesses

Broadwas Grimley Moseley

Church of Worcester St Guthlac’s ch Church of Worcester Church of Worcester Church of Worcester

53

Tenements also in Worcester

s.1370

Salt-boiling

s.1370

Salt-boiling

s.1370

Salt-boiling

28 tenements in Worcester in 1090s 1 tenement in Worcester 90 tenements in Worcester

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

between rural estates and royal estate centres as a tangible expression of obligations of the estate-holders to the centre has been put forward in explanation for the formation of contributory estates to royal estate centres such as Calne in Wiltshire (see chapter 3). A more specific instance of the connections of tenements in Droitwich to outlying estates is the record which claims that 10 houses with salinae (shares in a salt-works) were granted to the church at Worcester, together with 5 different estates, by King Cenwulf (798822) (S.179, S.180; Hooke 1981, 137-8, 153). That one of these estates, at Hallow, is amongst those contributing to Droitwich in Domesday Book (with the comparatively large number of 10 burgesses, exceed only by the 13 burgesses contributing to the royal vill at Wychbold) suggests strongly that this reference does indeed record connections between these contributory estates and the centre which were already established in the early ninth century. The manor of Hallow is further discussed by Baker and Holt (Baker and Holt 1996, 138; Baker & Holt 2004, 264).

tenements within burhs which had become attached to the centre by an act or royal prerogative. The particular case of the houses in Droitwich attached to Wychbold which owed service to the king at the royal tun itself can be seen as merely another instance of similar arrangements of obligations and service at other royal estates, already discussed. The fact that they performed these services at the king’s hall, even though ‘attached to’ (and probably living at) Droitwich rather than Wychbold, entitled them to privileges and protection which gave them an enhanced status, dignified by the term ‘burgess’ in Domesday Book. In this case it is not difficult to surmise that a proportion of the services required would have been to make salt for the king at the king’s saltworks at Droitwich, while those contributed from Hallow and other places held by the bishop would have had to have spent an additional chunk of their time making salt for the bishop as well. Further implications of these arrangements are discussed below. This model therefore provides an exactly analogous situation, though at a possibly even earlier period, to the way in which the customary burgesses at some burhs were ‘appurtenant to’ a neighbouring royal manor. Analogous examples are the burgesses in Bristol who were appurtenant to the royal manor of Barton Regis, and or the burgesses appurtenant to the royal manor of King’s Barton at Gloucester, discussed above (chapter 6). The same set of relationships is also found, for instance, in the burgesses at Axbridge and Langport, Somerset (also quoted above), which were ‘appurtenant to’ the royal manors of Cheddar and Somerton respectively.

This evidence also raises the question as to the status of the burgesses who were ‘at’ Droitwich, but were nevertheless contributed by or appurtenant to the royal villa or tun at Wychbold, where they had to perform mowing service and other unspecified services at the lord’s court. These represented about a quarter of the total number of burgesses or houses contributed from 11 estates. It would be consistent with this evidence to suggest that the connection between the two places shows that Droitwich was a specialised appendage of the royal tun, acting as an industrial and marketing centre, many of whose inhabitants (who were all presumably connected with the salt-producing industry) were drawn from estates comprised within the administrative area or regio of the tun, to perform services required by the king at the secondary centre rather than the primary one. It is quite clear from the early documentary evidence, discussed in detail by John Maddicot, that the development of the industrial resources of Droitwich was facilitated by the king from the time of the origins of the Mercian kingdom in the seventh century as a royal prerogative (Maddicot 2005, 26-33). A functional model which would accommodate this evidence would be to suggest that from the early days of the development of salt-working, both the king and perhaps the bishop laid out hagae within Droitwich, to which were attached rights in one or other of the various saltworks in the area, and which were attached as an appurtenance to various manors held by thegns. These holders were thereby obliged to perform various services for the king, the majority of which would doubtless have involved the production or handling of salt. The large sizes of these holdings in Domesday Book, measured in half-hides and hides, suggests that each individual holding would have accommodated subtenants whose business it was to work in the saltworks on behalf of the chief tenants. These hides and half hides – if they can be interpreted as actual holdings or hagae rather than being a measure of an assessment towards a contribution to a total value which was farmed by the shire reeve - therefore appear to have been closely analogous to the non-customary

Worcester Tenements in Worcester attached to rural manors are given both in the folios of Domesday Book as well as in numerous pre-Conquest charters. These have been described and discussed by a number of commentators (Dyer & Clarke 1968-9, 30-3; Hooke 1980, 39-40, 44, 48 (tabulated); Baker & Holt 1996, 136-40; Baker and Holt 2004, 133-4; Holt 2009, 67-9). Discussion - the pattern in Worcestershire The estates contributory to Worcester are spread over the whole shire, forming an area which completely overlaps the orbit of those contributory to Droitwich. As with the case of Winchcombe and Gloucester, discussed above, many of the more distant estates are somewhat nearer to other late Saxon market centres over the various shire borders, reinforcing conclusions drawn in relation to other centres that the proximity to local markets would have played no part in the genesis of this particular pattern of distribution. Some of those in the south of the shire, such as Bushley [28], and Teddington [15] are nearer to Pershore, a market centre on a large multiple estate with 28 burgesses of its own, and with burgesses ‘in’ Droitwich. Bushley is only 2.3 km (1 mile) from Tewkesbury, another multiple estate market in the south-east of Gloucestershire on the east bank

54

Worcestershire

Table 20. Worcester – non-customary tenements Contributory manor

Number of burgesses / tenements

Astley [1]

2

Bushley [2]

1

Chaddesley [Corbett] [3]

2

Coddington [4]

3

Halesowen [5]

1

Hollow Court [6]

1

Martley [7]

3

Northwick [8] (Tibberton is an appendage)

90

Osmerley [9]

1

Pedmore [10]

2

Suckley [11]

1

Upton Warren [12]

1

Witton [13]

1

Bengeworth [14] Oddingley & Laughern [15]

Cotheridge [16]

Battenhall, Perry [17] Clopton [19] Bredicot [20] Bentley in Holt [21] Wolverley, (Worcs) [and Blackwell, (Warwicks] [22] Bromsgrove [23] Hallow [24] Teddington [25]

4 1 haga in S of burh, 12 perches long, 7 broad 8 in 1086 1 haga, (by the S wall) 1 haga 1 haga ‘by the gate’ 1 haga ‘within the port’ 1 haga ‘in the port’ 3

Holder of manor

DB Section ref.

Ralph of Tosny Church of Worcester Edeva Church of Hereford Earl Roger

Comments

Tenements also in Droitwich

15.9 2,30; E4 28,1 HEF 2,32

In DB Herefordshire

14,1 HEF 1,41 X2.E3 HEF 1,39 WOR 18,5; 21,4; X3; E1

Bishop (25 others in the market place held by Urso the sheriff) Urso d’Abetot William son of Ansculf The king

2,49(-61)

3 (sect 2,50)

26,2 23,12 HEF 1,47

Urso d’Abetot (formerly Evesham Abbey) William s. of Corbucion (24,1) Urso d’Abetot (26,16)

Same as Barbourne, N Clines in charter of 904 (S.1280)

Earl William transferred to Hereford. (Part of Bromsgrove)

26,15

24,1; 26,16 7 held by Urso

Abbot of Evesham

2,75; 10,12

Bishop

Bishop, held by Osbern

S 1590 1077 – ‘Worcester H’

2,56

S 1297 (943 for 963)

2,4

S. 1303 (by 963) – with the advowson of All Saints (B&H 205, 264)

Bishop

Not given

S. 1327 (969)

Bishop

2,10

S. 1352 (985)

Bishop

2,60

S. 1369 (983x5)

Bishop

26,4

Bishop (Earl Leofric in DB) The king

2,83; 2,46 1,1a

28 (in 1090s)

The Priory

2,68(-71)

1 curtem

The Priory

2,23

55

S. 1384 (1042); S. 1394; S. 1395 (1042) S.1232 (1052-7) (inferred as Worcester) 3 saltworkers In 1090s (B&H 264) S. 1408 (since 969) (B&H 264)

10

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Contributory manor Perry (Wood) [26]

Number of burgesses / tenements 1 haga

Bushley [28] Kidderminster [29] Cookhill [30]

2 hagas (within the gate) 1 1 1

[Evesham church]

28

Tapanhall (in N Claines) [27]

Feckenham [31] Himbleton

2 1 25 ‘in the market place’

Holder of manor

DB Section ref.

Bishop

2,61

Bishop

Not given

Bishop The king Urso d’Abetot

HEF 1,44 1,2 26,1

Evesham church The king Bishop

Comments

Tenements also in Droitwich

S. 1385 (1003x1023) S. 1393; Hooke 1980, 48 1 Burgess Includes 4 app. to Bengeworth. Distribution of manors shown in fig. 9

HEF 1,40-41 S. 1593 (nd)

Urso d’Abetot

Notes a) the charter relating to land at Oddingley and Laughern [15] (S 1297) does not specify to which manor the haga mentioned in it is appurtenant; attachment lines are drawn to both. b) In the case of Wolverley and Blackwell (the latter in Warks) [22], the haga is inferred as being appurtenant to Wolverley, the largest manor; c) Osmerley [9] is not mapped in the Phillimore ed., but is here placed centrally within the hundred. d) Battanhall [17] in the charter S 1327 is not given in DB, but is associated with Perry [26] (Finberg 1961, 115). e) Cookhill [30],held by Urso, gives only I burgess without specifying the centre to which it is contributory; this could equally be Droitwich. B&H refers to Baker & Holt 2004. The phrases in the charters describing the types of urban tenements and their locations within the burh are given in Hooke 1980, 48.

of the Severn and Avon, yet is contributory to Worcester 18 km to the north. It is also nearer to both Gloucester and Winchcombe. The only apparent anomaly is Coddington [4]. The most natural explanation for this is that this estate, like Mathon immediately to its north (Humphrey-Smith 1984, map 38), was at an early period within the diocese of Worcester and therefore within the primary burghal territory of Worcester. Just as many of the estates in the putative early burghal territory of Gloucester which were held by the church of Worcester were included within the Domesday shire of Worcestershire, so Coddington, held by the church of Hereford (DB Hef 2,32), appears to have been included within the later (rearranged) shire of Herefordshire.

59-60, 65-6) would, on this interpretation, merely have developed and consolidated an arrangement which had been created at an earlier stage. (This would be a component of stage 3 of the timeline given in chapter 8 below - see fig. 12). I have also argued elsewhere (above) that the new burh at Gloucester, together with its burghal territory (the later shire of Gloucestershire) and its connections with rural estates, was also set up as a result of the same political developments (stage 2b, chapter 8). This original burghal territory of Gloucester would have been defined in relation to those of both Worcester to its north-west and Winchcombe to its east, the three together forming divisions of the former kingdom of the Hwicce (Hill 1981, 81; Hooke 1985; see also Bassett 1989a; Bassett 1989b; Bassett 2007; Bassett 2008a) – see fig. 11. It would therefore follow that the burghal territories of Worcester and Winchcombe must have been defined at the same time. This process was arguably also extended to include Hereford, with the formation of the redefined burghal territory. The archaeological evidence, analysed for instance by Steven Bassett (Bassett 2008a, 182-91), is consistent with the construction (or reconstruction) of an enlarged burh there at this time (c.880). The burghal territory of Winchcombe would have formed the template for the later shire of Winchcombeshire.

It would be consistent with the model already advanced to interpret this evidence of the spatial patterning of these contributing estates - as other examples discussed above as indicating the extent of the primary burghal territory of the burh of Worcester. I have argued elsewhere, on quite independent grounds, that Worcester is likely to have been the site of a new burh formed by King Alfred in c.880 after his takeover of Mercia in probably late 879 or early 880 (Haslam forthcoming b; see also the discussion in the following chapter). The documented episode of burhbuilding at Worcester in the 890s which is the subject of Aethelred’s well-known charter (S.223 – Tait 1936, 19-21; Whitelock 1979, 540-1; Brooks 1996b, 143-4; Baker & Holt 1996, 130-2; Baker & Holt 2004, 176-7; Holt 2009,

The most natural conclusion from this suggested developmental model would be that the pattern of

56

Worcestershire

distribution of the estates contributory to Worcester was established on the creation of this putative new burh of c.880, overlapping and in part subsuming, but also extending and adding to, the connections already established at Droitwich. Entirely consistent with this interpretation is the distribution of the contributing estates to the west, north-west and north of Worcester, which lie on the outer fringes of the old boundary of the Hwicce and the diocese, but are not found within the north-western area of the later shire which was part of the earlier diocese of Hereford and the area of the Magonsaetan (see fig. 11). It would appear likely that this area would have been added to Worcestershire as part of the same process of reorganisation in which the area to the west of the Severn (and also within the diocese of Hereford) had been added to Gloucestershire, in probably the phase of formation of the shires of the west Midlands, argued below as falling into the third quarter of the tenth century and into the early eleventh. This is discussed further in the following chapter (stage 5 and/or stage 6).

theory’, Baker & Holt’s model has lost sight of the obligations of the thegn to the king, and of the king’s role in determining these connections, which is argued in this study as being the fundamental dynamic in the creation of these urban-rural connections at a rather earlier stage than the first evidence for these connections in tenth-century charters. The partiality and particularity of this explanation by Baker and Holt for a ubiquitous phenomenon also overlooks the fact that already by 904 the haga and its associated water meadow which were given to be held by Aethelred by the bishop (S.1280 – discussed below) was appurtenant to the manor of Barbourne. This relationship between the urban appurtenances and the manor was clearly already in existence at the time of the grant, and being earlier than 904 takes the connection between manor and urban tenement to very near the beginning of the life of the burh. Baker and Holt’s interpretation is also not consistent with the fact that in every single charter which mentions this connection it is the urban tenement which is given as being an appurtenance of the estate, rather than the other way round. The bishop may well have been minded to make provision for his more important tenants by granting these estates to them, as Baker and Holt suggest (Baker & Holt 2004, 263-4), but this does not necessarily imply that these connections were first made at the time of these particular grants. It is just as likely that the urban tenement would have been appurtenant to the rural estate before their acquisition by the bishop, and therefore before the bishop’s gift of the estate to his retainer.

This overall developmental model is somewhat at variance with that put forward to explain the context of the connections of rural manors with Worcester by Nigel Baker and Richard Holt, and subsequently by Holt (Baker & Holt 2004, 261-7, 368-9; Holt 2009, 67-9). Holt’s model is an elabouration of Tait’s original thesis (discussed in chapter 1 above), and is in line with those put forward to account for these connections in most other places by almost every historian since Tait’s time, although given a new twist. The underlying premise of this interpretation is that the tenements of the church within the area of the burh ‘came to be attached’ to estates also owned by the bishop by a process of addition and accretion. They are interpreted as being ‘established for the convenience of the bishop’s administration . . . to provide [his] retainers both with a means of support and a base within the shire town’ (Baker & Holt 2004, 263), an interpretation also extended to those manors held by the king. This is seen as a more general process whereby the members of the ‘emerging landholding aristocracy’ – in effect the pre-Conquest thegns who owed service to either the bishop or the king - had become an essentially urban class, and who in consequence required the support of a rural manor. In explanation for the fact of these connections Baker and Holt suggest that ‘It was not the town house or burgage that was initially appurtenant to a manor, but rather the manor that had been assigned to support an urban household.’ (Baker & Holt 2004, 263, 265). This neat relationship is, however, confounded by the fact that two of the bishop’s seven estates in neighbouring Warwickshire, at Hampton Lucy and Alveston (DB War section 3 – see table 13), were contributory to three and four tenements respectively in Warwick. This relationship, which is clearly a function of their position within the shire (or, more accurately, within the early burghal territory), tells against the idea that the bishop was able to attach any of his estates to any tenement in Worcester.

Holt’s conclusion from other instances of these urban-rural connections that thegns owing service to the king as well as the bishop had been generally based within the burh during the tenth century is undoubtedly correct. But this raises the question as to whether a thegn derived his status from his holding of a town tenement (to which an estate may or may not have been attached), or his holding of an estate above a certain size (to which a town tenement may or may not have appurtenant). The bishop’s gifts of estates to his retainers can therefore be seen as a mechanism whereby he provided both for their residence in the town as well as an income from the estate to which is was appurtenant. On the alternative interpretation offered here, this pattern would have originated in arrangements which had nothing to do with the provisions of one tenant-in-chief for his retainers, but rather with provisions made within a nexus of tributary relationships which bound estates and their holders to the king (whoever the tenant-in-chief was at the time) for the upholding of the king’s works. This is discussed further in chapter 9. These connections were arguably therefore set up at a rather earlier time than Baker and Holt would allow – i.e. in the later ninth century on the foundation of the burh as an institution. The creation of the burghal territory of Worcester at this time can most reasonably be considered the explanation for the fact that four estates, at Kidderminster, Northwick /Tibberton, Witton and Hallow, held burgesses or houses at both Droitwich and Worcester. This also holds for other estates in other shires already discussed. In this case, the

However, as in Tait’s initial reaction to Maitland’s ‘garrison

57

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

between the bishop and Earldorman Aethelred. This is evidenced directly in the fact that at the time of Domesday 45 out of the 90 houses in this haga belonged to the bishop in lordship. His tenants, who are named, paid nothing ‘except work in the bishop’s court’ (DB Wor 2,49). The arrangement in which the ealdorman took the manor of Northwick, to which this haga was appurtenant, therefore did not affect the tenurial arrangements of the holding of the haga itself, which would anyway have been held by both Aethelred and the bishop from the time of the formation of the new burh in the 890s. Such a division of assets is also shown in the somewhat atypical arrangement in which the bishop had the third penny of the borough before 1066 which he shared with the king and the earl (presumably at one third share each) thereafter (DB Wor 2,49). This is also a particularly striking example of how one large tenement which is known to have been appurtenant to a rural estate at a particularly early stage in the development of the burh became subdivided into many smaller ones by the time of Domesday, a process described in detail by Baker & Holt (Baker and Holt 2004, 174-7). A similar process must have occurred in the history of the development of the probably single haga in Gloucester appurtenant to Deerhurst, which became subdivided into 30 tenements by the time of Domesday, described above.

connections which had been established at Droitwich possibly from at least as early as the beginning of the ninth century, and which possibly represented the area of an early regio with its centre at the royal site of Wychbold, would have been overlain by those established with the putative new burh at Worcester in c.880. This provides an exactly parallel instance of the same situation pertaining between Winchcombe and Gloucester, discussed in the preceding chapter. The hypothesis that the connections between these contributory estates and the burh at Worcester were established at the time of its foundation receives some support from the charter of 904 (S.1280), granted by the bishop for three lives (ie Aethelred, Aethelflaed and their daughter Aelfwyn) already mentioned above. Both the haga in the town and an area of meadow on the west side of the river were appurtenant to the manor of Barbourne, which was a part of the large manor of Northwick which was in the hands of the bishop, and part of the parish of Claines, a chapelry of the early church of St Helen’s Worcester (Baker 1980; Bassett 1989a; Bailey 2001, 117-8; Baker & Holt 2004, 197-8). As suggested above, it is clear that already by 904 both the urban haga and the associated water meadow were considered as an appurtenance of the rural manor, and were not created as such on the occasion of the grant. The haga appurtenant to Northwick in Worcester held 90 houses in Domesday (DB Wor 2,48-9) – more than three times the size of the holdings of Evesham (23 houses), and many more times the size of the other 12 estates (8 with one tenement or burgess, 2 with two, and 2 with three). This haga in Worcester was situated along the waterfront to the south of the bridge, possibly straddling the defences of c.890, where activity before the foundation of the burh is indicated by archaeological evidence of middle Saxon occupation (Baker & Slater 1992; Baker et al 1992, 73; Baker and Holt 2004, 174-7). The particular layout of this haga in relation to other primary elements of the topography of the late ninth-century burh, in particular the line of the northern defences (Baker & Slater 1992, 60-1; Baker & Holt 1996, 134-5; Baker & Holt 2004, 174-7, 262-3), and its ownership by the bishop in 904, shows that it is likely to have been set out and defined during the very beginning of the formation of the burh. There is every reason to infer, therefore, that this haga was from the beginning attached to or appurtenant to Barbourne, part of the bishop’s manor of Northwick. The evidential importance of this is of a similar order to the inference which can be made in relation to the existence of the connection of the manor of Ramsbury to the burh of Cricklade (Wilts) before 909, when it was given to the bishop of Ramsbury, which is discussed in chapter 2 above. This being so, this evidence provides a basis for a new paradigm which establishes the circumstances in which other connections between urban tenements and rural manors are likely to have been formed.

It would be consistent with the evidence as a whole to suggest that long before the foundation of the burh this haga was a discrete area adjacent to the cathedral precinct (but outside the primary defended enceinte within the former Iron-Age and Roman hillfort), which was controlled by the bishop and which had been given over to river-borne trade, and that this was the primary locus for the redistribution of salt from the bishop’s commercial interests in Droitwich from the late seventh or early eighth century. It is of interest that this haga in Worcester is very similar in its type, dimensions, siting and date to the haga at London given to the same bishop in 889 by Alfred and Aethelred, which was situated at Queenhithe on the Thames in Alfred’s new burh (Dyson 1978; Dyson 1990). This haga was the latest in a series of holdings and privileges in London given to the bishop by various Mercian kings from the eighth century onwards (Kelly 1992). I have suggested elsewhere – taking forward ideas put forward by John Maddicot (Maddicot 2005) - that the development of this haga was given to be held by the bishop by King Alfred and Aethelred to facilitate the promotion of the bishops’ commercial interests in the trade in salt from Droitwich within the developing burh at London and probably further afield, through a transhipment point at Lechlade on the Thames (discussed above) (Haslam 2010b, 128-30; Haslam 2011a, 133). It is possible, therefore, that the gift in 904 by the bishop to Aethelred was, as already suggested by Nigel Baker and Richard Holt, some sort of quid pro quo for the favours given to him in London (Baker & Holt 2004, 263).

In the process of the development of the burh in the 890s the assets of the burh, doubtless with the burghal space itself, were divided – as the charter S.1280 indicates -

The evidence of the connections of contributory estates to Worcester is extended by the case of the church of Evesham, which held a total of 28 tenements in Worcester (DB Worcs

58

Worcestershire

10,17). This entry is the last in the list of 16 estates it held in the shire, including part of Evesham itself, but in none of these entries in Domesday Book is any mention made of holdings of appurtenant burgesses or tenements. However, in a dispute between the abbot of Evesham and the bishop over rights relating to two of these estates in 1077 mention is made of 4 tenements in Worcester appurtenant to Bengeworth (DB Worcs Appendix – ‘Worcester H’). It is a likely inference, therefore, that the 28 tenements contributed by Evesham in Worcester represents the sum of separate tenements which were appurtenant to some if not most or all of the 16 estates held by Evesham in the shire. The distribution of these, all located in the south-east corner of the shire, is therefore shown in fig. 9.

estates and tenements by the abbey carries the implication that the estates given to the abbey at various times would have come with urban tenements already attached. Other comparable examples already discussed include the former holdings of the earl of Warwickshire, given to Coventry Abbey before the Conquest, which held 36 tenements in Warwick at the time of Domesday (fig. 3); the abbot of Gloucester, with 14 manors in the shire and 52 tenements in Gloucester (fig. 8a); the archbishop of York, as successor of the lands of St Oswald’s Priory, who held 13 manors in Gloucestershire and 60 tenements in Gloucester itself (fig. 8a); and the abbeys of Wherwell and Romsey in Hampshire with urban tenements in Winchester (see chapter 4). A similar conclusion can be drawn – that the estates held by these institutional bodies, and perhaps acquired by them at various times, were from the time of their acquisition already provided with an appurtenant urban tenement. A similar conclusion has already been made above with regard to the estates held by the bishop of Worcester and his urban tenements.

A somewhat different situation appears to be represented by the holdings of the lands belonging to St Peter’s Westminster (formerly held by King Edward) at Pershore. This estate held a sizeable holding in Droitwich, which comprised four salt houses and a church with 2 priests, together with 31 burgesses, all presumably in one haga (DB Worcs 8,13). This could be taken as arising from a situation in which the royal estate at Pershore was contributory to Droitwich at an early period – possibly in the eighth or ninth century - in the sense that its holder, or the agent appointed by the holder, had obligations to perform services at the king’s salt workings at Droitwich, where it would have acquired an appurtenant tenement. This is a vivid illustration of the principle, elabourated above, in which the commercial interests of the king, or perhaps a large ecclesiastical body such as the bishopric of Worcester, were facilitated by obligations for service as one aspect of a tributary relationship.

A parallel situation is also perhaps represented by the 25 houses held ‘in Worcester market place’ by Urso the sheriff (DB Worc 2,51). Urso also held 24 houses which represented subdivisions of the original haga appurtenant to Northwick, held by Aethelred in the charter of 904 (above). Although Urso’s 25 houses in the market place in section 2,51 are described as belonging to Northwick, it would seem more likely that these are different from his 24 within the area of the haga of 904, and perhaps represent the sum of those appurtenant to his other 17 manors in the shire. If this is so, the distribution of these is of interest in covering much of the north-eastern quarter of the shire. This distribution does not, however, give any further evidence for determining the earliest territory of the burh beyond the pattern of distribution of the named estates which are documented as being contributory to tenements in Worcester. Both these estates held by Urso and the estates held by Evesham and Pershore merely increase by a factor of two or more the documented incidence of these estates over the whole shire which may be reasonably held to have had appurtenant tenements in the central burh of Worcester and the royal multiple estate of Droitwich.

An analogous situation to that of Evesham is that at Bath, where the abbey church of St Peter held 24 burgesses within the borough (DB Som 7,1). These may reasonably be inferred to have comprised the sum of individual tenements appurtenant to its holdings of 14 manors around Bath. The fact that the abbey had acquired many if not most of these estates by gift during the tenth and early eleventh centuries (Cunliffe 1984, 352-3 & n. 10) does no affect this conclusion. The fact of these holdings of

59

Chapter 8 Discussion – Urban-rural Connections and the Development of Territories Discussion – urban-rural connections and the development of territories

as of archaeological work on the investigation of burghal defensive systems. This timeline incorporates the basic premise, already discussed, that the middle Saxon regiones were carved up, sometimes successively, into burghal territories at particular periods of burghal formation in the ninth and tenth centuries, and that these territories were themselves subsequently modified by both splitting and amalgamation to form the later shires. (The development of the west Midland shires is discussed in Whybra 1990, and in Bassett 1989b and 1996. The development of burghal systems, and archaeological work in particular at Winchcombe and other West Midlands burhs, is reassessed in Bassett 2007, 2008a and 2011, though for a critique of some of Bassett’s conclusions see Haslam 2011b, 210-14).

In this chapter an attempt will be made to articulate a developmental and functional model for the whole phenomenon of the connections between rural manors and urban or other royal centres in the early medieval period, based on the evidence relating to the shires so far discussed. This discussion will be extended to cover the area of the upper Thames – effectively Oxfordshire, Berkshire and Buckinghamshire – in the following chapters. This will act as a further test of the general model in an area which covers the southern part of former Mercia as well as Wessex. The developmental model which has been put forward is based initially on the evidence from Wiltshire. The distribution of manors or estates with appurtenant tenements in Malmesbury, Cricklade and Wilton is seen as a function of their position within the respective burghal territories of these burhs, or of the even earlier administrative areas of the royal estate centres of for instance Calne and Wilton. It is argued that the best explanation in functional terms for the spatial distribution of the pattern of the contributing estates to these and other centres is that these connections were for the most part formed at the same time as these centres and their dependent territories. This can be applied to the evidence from the other three shires in West Mercia considered here. Since the orbits or patterns of interrelationship of the estates contributing tenements to the centres in Gloucestershire, Warwickshire and Worcestershire show similar patterns of juxtaposition and overlapping (layering in a temporal sense), as well as broad distributions which can be seen as reflecting the boundaries of early burghal territories, this model also serves to provide the basis for the interpretation of their inter-relationship and function. This overall model is also consistent with recent research into the development of burhs, burghal territories and shires, in ways which will be brought out below.

This same temporal sequence will be shown in the following chapters (10, 11 and 13) to be applicable to the areas of Domesday Oxfordshire, Berkshire and Buckinghamshire, with the difference that the sub-stages 1a and 1b cannot be readily separated in the Oxford area by reference to unassailable archaeological evidence for the existence of a phase of the creation of defences of a burh at Oxford in the middle Saxon period. The general historical contexts of these phases of burghal formation differ in some important respects from the model put forward by Steven Bassett in several papers (particularly Bassett 2007; Bassett 2008a; and Bassett 2011) and by Julian Whybra (Whybra 1990). In my view the archaeological evidence, which Bassett has so carefully teased out of sometimes opaque records of earlier excavations, is susceptible to somewhat different interpretations when viewed in tandem with the documentary records of the overarching political and strategic contexts. A part of the reassessment of this evidence, as it relates to the development of Tamworth, is given in Haslam 2011a, 214-5. Stage 1 - Late 7th – late 9th centuries. Period 1a - The royal administrative centre at Winchcombe can be seen as a complement to the ecclesiastical site at Worcester; and the whole of the territory of the Hwicce could be seen as the former’s administrative area (province) as well as the latter’s diocese (Slater 1982; Slater 1983; Hooke 1985; Bassett 1985; Brooks 1989). The original connection of Lechlade to Winchcombe (as well as the fact of the survival of this connection) can be seen as the result of the probable administrative role of Winchcombe in

These orbits can be explained, in terms of both known historical processes as well as functions, by the overlapping of phases of burghal and territorial formation which can be characterised by the following timeline. This is put forward as a preliminary model which attempts to accommodate the processes of formation of the spatial patterns discussed above to the considerable amount of accumulated work on the origins of kingdoms, territories and burhs, as well

60

Discussion

Fig. 10. Reconstruction of the territorial development in the West Midlands – Stage 1, period 1b (late eighth – early ninth century) – development of primary burghal territories within the regiones of Winchcombe and Hereford.

the distribution and trade of salt from Droitwich along the Thames through a distribution point at Lechlade at this and later periods (Maddicot 2005, 44-5; Kelly 1992, 12; Blair 1996; Haslam 2011a, 133). An analogous situation of the role of Oxford as the centre of a middle Saxon territory or regio, which is contiguous to that of the Hwicce to its east, is argued in chapter 13 below.

of the Hwicce. The burghal territory of Tamworth would have comprised at least the southern part of the province of the Mercians (Bassett 1996, 152, fig. 11). The dependent territory of the latter probably extended south and eastwards over the area defined by Watling Street to the north-east (with the Viking-held area dependent upon Leicester beyond it to the north-east), the boundary of Lichfield and Dorchester dioceses to the south-east, and northwards over much of Staffordshire. The issue of the northwards extension of Tamworth’s primary territory is of importance in this general historical model, but is outside the scope of this study. Some of the connections of estates with both Winchcombe and Droitwich are likely to have belonged to this phase. Evidence for early defences at Hereford also suggests the existence of a burh there at this time (Bassett

Period 1b – late eighth / early ninth century (see fig. 10). The archaeological and other evidence for the formation at this time of defended burhs at Winchcombe and Tamworth, as well as at Hereford, is set out by Steven Bassett (Bassett 2008a; see also Bassett 2011, 5-10). The burghal territory, or dependent ‘territory of obligation’ of Winchcombe seems likely to have comprised all of the area of the province

61

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

2008a; Bassett 2011), its burghal territory coextensive with the regio of the Magonsaetan (Hill 1981, map 143; Hooke 1985, 7). It seems likely that the early territory dependent on the burh at Winchcombe would have extended eastwards only as far as the Severn (Finberg 1972, 45), and thus coextensive with the diocese of Worcester. This situation would appear to have formed the context of the mention of Wincelcumbe scire in 803 (Taylor 1898, 43-4), although this could equally perhaps have referred to its role as a unit of royal estate management at an earlier period (stage 1, period 1a above), rather than to its military role as the centre of a burghal territory. A similar ambivalence exists in the context of the annual food render payable from the estate at Aldingham, near Evesham, to Winchcombe in the reign of Coenwulf (Finberg 1972, 229).

as a new burh at the time (Haslam 2010b; Haslam 2010c; Haslam 2011a). Steven Bassett and others have however consistently attributed this phase of burh-formation in the western Midlands to Aethelred and Aethelflaed in the years around 900 (Heighway 1984, 366; Blair 1994, 99-101; Bassett 1996, 155-7; Bassett 2008a; Bassett 2011, 10-13; Baker and Brookes 2011, 111-4). I have, however, set out arguments against this scenario elsewhere (Haslam 2010a; Haslam 2011a, 133-5; and further in chapters 9 and 12 below). The central basis for this shift in understanding is that this general hypothesis of the origin of these burhs in the late 890s has ignored the reality of the political developments in the early 880s, when Mercia became absorbed into a ‘wholly new and distinctive polity’ which contemporaries called the ‘Kingdom of the Anglo-Saxons’, with King Alfred as its overlord, alluded to above. In this context, the notion of burghal formation initiated by Aethelred, who was subordinate to King Alfred in his dealings within Mercia, is untenable, since the necessary submission and oath of allegiance which burghal formation would have entailed would have been to Alfred, not to Aethelred. It is for this reason that it is possible to postulate a process of burghal formation in western Mercia by King Alfred soon after this momentous political development, in c.880 (or possibly even in late 879), as part of a programme which included the development of a burh at London at this time. This can be seen as perhaps the most crucial aspect of the way Alfred was able to extend his own control over this territory, even though placing its management in the hands of the Mercian Aethelred as its ealdorman. This programme arguably included the formation of burhs at other centres besides London in western Mercia, at Worcester, Hereford, Tamworth, Winchcombe and Gloucester. To these might well be added Chester and Shrewsbury, by the evidence adduced by Bassett (Bassett 2011, 10-13, and map fig. 1), although discussion on this is beyond the scope of this study. It seems likely that Bristol would also have been formed in this phase. In the reconstruction in fig. 12 the development of Shrewsbury is however depicted as being a secondary development, belonging to the subsequent stage 3, which may well have been initiated by Aethelred and Aethelflaed as a new political initiative in the 890s – though still under the ultimate control of King Alfred. The defences of Hereford were also arguably extended to the east of an initial sub-rectangular enclosure (of stage 1, period 1b, above) at this time (stage 2b).

Stage 2 – late 9th century Period 2a - 878-9 - King Alfred’s first phase of the redefence of Wessex, which involves the creation of the burhs listed in the Burghal Hidage, with burhs in Wessex at Malmesbury, Cricklade and Bath just over the borders of Gloucestershire to the south. (For the dating, and description of some of the strategic factors, see Haslam 2005; Haslam 2009; Haslam 2011a; see also further discussion in chapters 2 and 3 above, and chapter 12). At this time Guthrum’s Viking army controlled West Mercia, occupying Gloucester and Cirencester in the south. The western part of the burghal territory of the burh of Oxford, created at this time within an area which was subject to the overlordship of King Alfred, was defined by the eastern boundary of the Hwicce (see chapters 10 and 11 below). A portion of the southern part of the Hwicce was also subsumed as part of the burghal territory of the formerly Mercian centre of Bath, which was also one element of the system of burhs listed in the Burghal Hidage (Haslam 2005, 141-4). Period 2b - c.880 (see fig. 11). The developments in Wessex were followed soon after by King Alfred’s absorption of western Mercia under a new polity, which contemporaries termed the ‘Kingdom of the Anglo-Saxons’ (Keynes 1998, 24-6, 34-9, 43-4; Keynes 1999, 460-6; Keynes 2001), which was made possible by the death of the Mercian King Ceolwulf and the retreat of the Vikings under Guthrum to East Anglia in late 879. I have argued elsewhere that this would have led to a reorganisation of the burghal system in western Mercia on the instigation of King Alfred, with the formation of new burhs at Gloucester and at Worcester, as well as the refurbishment of the burhs at Hereford (above), Winchcombe and Tamworth (Haslam 2005, 143-4; Haslam 2011a, 133-5). This is consistent with the archaeological evidence, as set out by Steven Bassett (Bassett 2008a; Bassett 2011), though the divergence of this view stated above with his own cannot be pursued in more than outline here. As I have argued elsewhere, this absorption or conflation of the kingdom of Mercia with Wessex under a single polity also formed the historical context for King Alfred’s resumption of control of London and the minting operations therein, and his development of the Roman city

The rationale of the necessity for the development of this system of c.880 – for such it must have been – is provided by the consideration that the formation of burhs would have required the formal submission to the king – in this case King Alfred - of populations within the respective burghal territories of each of the burhs. The formation of this new burghal system at this time would therefore have been one of the principle instruments through which Alfred would have ensured the political and military control of this area of (former) Mercia, through the submission of people within

62

Discussion

Fig. 11. Reconstruction of the territorial development in the West Midlands – Stage 2, period 2b (c.880) – development of burghal territories of Hereford, Worcester, Winchcombe, Tamworth and Gloucester (north-western extent of the territory of Tamworth uncertain). This period also possibly saw the creation of a new burh at Bristol (see fig. 12).

each of the burghal territories to himself. It is this process which he was able to consolidate in the general submission to his overlordship in 886 (Whitelock 1979, 199). This is discussed further in chapters 9 and 14. This programme of burh-building would arguably also be entirely consistent with the archaeological evidence for the development of the defences of these places, as set out by Bassett (in particular in Bassett 2008a and Bassett 2011), though this cannot be discussed further here.

of Worcester, Gloucester and Winchcombe. These would have been the progenitors of the later shires, and possibly equivalent to them in many of their functions in all but name. A similar process would have taken place at Hereford, with the formation of its burghal territory which would have been the forerunner of the shire of Herefordshire. The possibility, mentioned above, that the development of a burh at Shrewsbury and the formation of its burghal territory at this time would have altered the earlier pattern must, however, be borne in mind. The burghal territory of Worcester covered most (but not all) of later Worcestershire; that of Gloucester covered the southern part of later Gloucestershire east of the river Severn. Both, in other words, would have been bounded by

The administrative geography of the earlier regio or kingdom of the Hwicce, which had Winchcombe as its central royal focus, would thus have been redefined at this time in terms of the formation of burghal territories

63

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 12. Reconstruction of the territorial development in the West Midlands, with the formation of new burghal territories of Bristol and Shrewsbury. (That of a possible burh at Cirencester is not shown.) Stage 3 - early-mid 890s. It is however more likely that the formation of a burh at Bristol belongs to the preceding period (stage 2, period 2b).

connection of Winchcombe with Lechlade, and probably with other places, had possibly survived from an earlier period.

the earlier boundaries of the Hwicce. The burghal territory of Winchcombe would have covered the north-eastern part of later Gloucestershire as well as southern Warwickshire up to the eastern border of the Hwicce.

Stage 3 - early-mid 890s (fig. 12).

However, an uncertain aspect of this reconstruction is the inclusion of an area to the east of later Worcestershire and to the north of the river Avon (which later became the northern part of the burghal territory of Warwick – see fig. 12) within the burghal territory of Winchcombe, rather than that of Worcester. This process of burghal formation is arguably the most appropriate and likely occasion for the formation of a new set of connections of estates within these territories to the central burhs of Worcester, Gloucester and Winchcombe (see discussion below), though, as already suggested, the

It is to this time that can be assigned the phase of King Alfred’s secondary consolidation of the defences of both Wessex and western Mercia, which arguably involved the creation of new burhs in Wessex replacing the temporary burhs which are listed in the Burghal Hidage, and the replacement of defences of earth and turf with stone walls, in response to the new Viking incursions of the time (Haslam 2009, 98-100, 103-4, Haslam 2011b, 210-15). It can be argued that this process is paralleled by exactly

64

Discussion

similar developments in western Mercia. It is to this phase that the creation of a burh at Bristol, and possibly another at Cirencester, amongst others, is most likely to belong, as well as the development of a new burh at Shrewsbury. This context is usually seen as the occasion for the formation of the enlarged burh at Worcester as part of this process, though is now considered (above) to have been formed in the preceding stage. (For Worcester, see Tait 1936, 19-21; Brooks 1996, 143-4, Baker & Holt 2004, 147-95, 34750; Bassett 2008a, 226-30; Holt 2009, 61-6. Cirencester is discussed in Williams 1989, 9 and Gerrard 1994.) The case for the formation of a burh at Bristol at this time has yet to be argued, but is suggested as having been formed as a burh in the late ninth century (Manco 2008), with a dependent territory probably comprising the seven hundreds of Grumbalds Ash (Cam 1963, 96) being formed around it. This was clearly carved out of the burghal territory of Bath to the south of the river Avon as well as that of Gloucester to the north. This is reflected in the pattern of tenements in Bristol which were appurtenant to estates on both sides of the river Avon (see fig. 6).

chapters 10 -13 below. In the west Midlands it is represented by two sub-periods: Period 4a – This period involved the reconstitution of the defences of the burh at Tamworth and the creation of a new burh at Stafford in 913 (Anglo-Saxon Chronicle, Mercian Register – Whitelock 1979, 212). The early tenth-century burghal territory of Tamworth can be reconstructed so as to include all of later Warwickshire in the diocese of Lichfield up to its border with Leicestershire, as well as the southeastern part of later Staffordshire, though this was probably not much different to its burghal territory established in c.880 (see the map in Bassett 1996, 152-4, figs. 11-13). Bassett has suggested that before this time Tamworth would have been occupied by the Vikings from Leicester and Northampton (Bassett 2011, 15-17). However, the fact that the excursion of the Viking army which ‘broke the peace’ in 913 was aimed to the south around Hook Norton in north Oxfordshire (Whitelock 1979, 212), suggests rather that the defences of Tamworth remained intact, and still functioned as part of the defences of western Mercia. The timing of these developments (raiding by the Vikings in late March, with the new defensive arrangements put in place in the summer) suggests that the strengthening of the burghal defences of Tamworth and the creation of Stafford at this time was a response to the insecurities brought about by these Viking initiatives. The spatial inter-relationship of the burghal territories of Stafford and Tamworth at this time, though an important aspect of the general historical model outlined here, cannot be analysed further, as being outside the scope of this study.

As I have suggested elsewhere, the extension of the phase of burghal formation at this time also probably included the consolidation of the defences of both Hereford, Worcester, Tamworth and Winchcombe with stone walls, which is shown in the archaeological evidence from all four places (Haslam 2011b, 210-12). Steven Bassett has, though, consistently assigned this process to the early eleventh century, his third phase of the development of defensive provision of the burhs (Bassett 2008a; Bassett 2011, 5, 20). However, there are reasons, which are based on the archaeological evidence, for concluding otherwise, some of which are discussed elsewhere (Haslam 2011b, 210-12). Arguments for regarding the burh of Worcester, documented in the charter S.223 , as a secondary extension of a primary phase of burh-building of c.880 will also be set out elsewhere (Haslam forthcoming b). It might well be, as Bassett has argued, that a burh at Shrewsbury could have been formed in this period (Bassett 2011, 11-12), rather than as part of the initial ‘Alfredian’ programme of c.880. In this case its burghal territory would probably have been extended, as described above, to absorb the northern part of the earlier regio and burghal territory of Hereford, extending to the boundary of the latter (which was also the diocesan boundary) on its south-eastern side, but extending beyond this to incorporate part of the burghal territory of Tamworth to its east and northeast. This reconstruction is, however, beyond the scope of this study.

Period 4b – This is represented by the creation of a burh at Warwick in 914, at a royal and probable early minster site with established central place and strategic significance (Slater 1982, 178-9; Slater 1983; Bassett 2009). There are good arguments for suggesting that its burghal territory would have been carved out of the north-eastern part of the putative burghal territory of late ninth-century Winchcombe, which had earlier extended up to the north-eastern boundary of the Hwicce (which must have formed the south-western boundary of the burghal territory of Tamworth re-formed in 913). This would have left a smaller territory dependent on Winchcombe, which was to become the forerunner of the tenth- and eleventh-century Winchcombeshire. (For the boundaries, see Hill 1981, 99; Whybra 1990 redefines the boundary on more detailed evidence). The common boundary formed in consequence between the territories of Winchcombe and Warwick are, however, not easily demonstrated in the surviving evidence. A similar process is shown by the creation of the burhs at Bricg (probably Bridgnorth: see discussion in Bassett 2011, 14-15), and Chirbury, their burghal territories formed from the southern part of that of Shrewsbury (and within the diocese of Hereford). This development, and those involving the areas to the north (later Shropshire, Staffordshire and Cheshire), though an integral part of this process, are beyond the scope of this discussion. This is not therefore shown in fig. 14. As has already been suggested, it is therefore a reasonable

Stage 4 - early 10th century (fig. 13). This phase is represented by the staged expansion of the burghal system and the creation of dependent territories throughout western and eastern Mercia by Aethelred and Aethelflaed, and from c.910 by Edward the Elder and Aethelflaed. In western Mercia this process has been described most recently by Steven Bassett (Bassett 2011, 14-19). The case of the central Midlands is described in

65

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 13. Reconstruction of the territorial development in the West Midlands - Period 4b (c. 914). Development of new or redefined burghal territories of Warwick, Winchcombe, Stafford and Shrewsbury, as well as Chirbury and Bricq (Bridgnorth). The extent of the territory of Stafford in relation to that of Tamworth at this stage is uncertain; those of Chirbury and Bridgnorth are also uncertain, being probably carved out of the territories of Hereford and Shrewsbury.

in chapter 13 below. This can be seen to have involved the formation of new administrative units which were formed by the amalgamation and fission of the former burghal territories around some of the already-existing burhs, and by the loss of functionality and of administrative status of Tamworth, as well as of some of the smaller burhs, such as Chirbury, Bridgnorth and Bristol. This reflects the archaeological and some historical evidence, discussed in chapter 13, for the abandonment of the defences of some of the smaller burhs in Wessex and the probable disuse of their burghal territories (Haslam 2011a). This process seems likely to have involved the following processes:

inference that the pattern of the connections of manors with Warwick, which in general lie to its south and east, reflect the event of the creation of its burghal territory at the same time in 914. Those manors contributory to Tamworth were also formed in 913 or earlier. Stage 5 - mid-late 10th century (fig. 15). There are a number of considerations which combine to suggest that the west Midland shires were reorganised at this time – most probably in the 960s. This general model, together with the dating, and previous discussions on the issue, are more fully treated in relation to the development of the shires of Oxfordshire, Berkshire and Buckinghamshire

• The expansion of the burghal territory of Warwick to

66

Discussion

Fig. 14. Reconstruction of the territorial development in the West Midlands - stage 5 (mid-late tenth century) – the first stage of formation of the shires, involving the development of Warwickshire and Staffordshire at the expense of the burghal territory of Tamworth, the expansion of Shropshire at the expense of the burghal territories of Chirbury and Bridgnorth, and the expansion of Gloucestershire at the expense of the burghal territories of Bristol and (possibly) Cirencester.

the north and east, taking in the south-east part of the area of the burghal territory of Tamworth to form the Warwickshire of Domesday. As argued above, this would have been accompanied by the formation of new connections of manors in the area of the expanded shire with tenements in Warwick (many of which connections were probably preserved by being granted by the earl to Coventry Abbey in 1043). The south-western boundary of this new shire, as shown in fig. 14, is however somewhat notional. • The burghal territories of Tamworth and Stafford were in this process divided between Staffordshire and Warwickshire (with new administrative centres

at Stafford and Warwick), with the common boundary between these two new shires passing through Tamworth itself. • The southern part of the new shire which was based on Shrewsbury (Shropshire) would have incorporated the earlier burghal territories of Chirbury and Bridgnorth (Bassett 2011, 4 fig. 1), probably being re-established in relation to the boundaries of its earlier burghal territory established in stage 3, as discussed above. • The absorption of the burghal territory of Bristol to the north of the Bristol Avon (and that of the possible burh at Cirencester) into the new shire of Gloucestershire. • The extension of the new shire of Gloucestershire to

67

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig 15. Reconstruction of the territorial development in the West Midlands - stage 6 (early eleventh century). This is represented by a) the absorption of Winchombeshire by Gloucestershire and the expansion of Gloucestershire to the west of the Severn, b) the consolidation of areas of Gloucestershire and Winchcombershire south of the Avon to form Worcestershire in its final form, and c) the expansion of Worcestershire to the north-west at the expense of Herefordshire.

include a chunk of territory to the west of the river Severn, which would have been carved out of the territory of the former burh at Hereford. This might, however, have taken place in the early eleventh century – stage 6, below. • The creation of the new shire of Worcestershire would at the same time have involved the extension of the earlier burghal territory to the west and north-west, again at the expense of the territory of Hereford. Similarly, the new shire may well have been pushed to the south of the river Avon to incorporate part of the former burghal territory of Winchcombe. It is to this period, therefore, that the formation of the connections of two

estates in this area (Bengeworth [14] and Teddington [25]) can be most appropriately assigned – parallelling similar processes which appear to have taken place on the formation of Warwickshire at this time. It is also possible that the attachments of all the estates of Evesham Abbey (including Bengeworth) which lie to the south-east of the river Avon (see fig. 9) could have been transferred from a primary association with the burh at Winchcombe, being included within its burghal territory, to become contributory to the new shire capital at Worcester, in an exactly similar way to the realignment of the associations of the estates of Coventry Abbey from the old burghal territory of

68

Discussion

Tamworth to become contributory to the new shire capital of Warwick, at the same period (see chapter 5).

over time. One of the important implications of this model is that these connections, as they have survived to be recorded in Domesday Book, earlier charters and later historical sources, were brought into being as a result of the position of the centre as the head place of a territorial unit – an early regio or province, a burghal territory, a large royal multiple estate, or a later shire. An essential corollary of this is that these connections were formed at the same time and by the same acts of state by which these places were constituted as the central administrative places of these territorial units. This arguably supplies the best explanation for a number of aspects of their spatial relationships, detailed above. The alternative view that these connections of manors to their centres developed by a process of accretion, through the acquisition by thegns with rural estates wanting to acquire a town house at their nearest centre for whatever reason, or by a king or bishop who wished to reward a thegn or retainer with land, is not consistent with the relationships shown by this spatial patterning. These processes are seen as inadequate as an explanation of the historical and spatial attributes of groups of these connections examined using the same methods of enquiry over a number of shires.

Stage 6 - early 11th century (fig. 15). This time saw a further reorganisation of the West Midland shires, involving the absorption of Winchcombeshire into a new enlarged shire of Gloucestershire. This may reasonably be attributed to the activities of Earl Edric Streona in c.1007 (Taylor 1898, 43; Whybra 1990, 5-6). It is possibly at this time that large chunks of territory, involving the amalgamation of manors in one fief into separate hundreds, in particular in the area of the former shire of Gloucester, were incorporated into the administrative region of Worcestershire, forming the confused pattern of interrelationships of parts of one shire contained within another, which is so characteristic of the arrangements in the medieval period and later (e.g. Hill 1981, 99). As C S Taylor has suggested, this phase saw a probable readjustment of the boundary between the mid tenth-century shires pertaining to Hereford and Gloucester, such that Gloucestershire gained parts of Herefordshire west of the Severn, and Worcestershire was expanded to the north-west, again at the expense of Herefordshire (Taylor 1898, 44). It is as possible, however, that these processes of readjustment had already taken place at an earlier stage in the 960s (stage 5, above). It is to this period that can be attributed the formation (or perhaps realignment) of the connections of some of the manors which lay within the former (late-ninth century) burghal territory of Winchcombe, but which are shown to be contributory to Gloucester in Domesday Book. This process would also account for the incidence of shared connections between some manors between Winchcombe and Gloucester, in the same way and for the same reasons which would account for similar shared connections between Droitwich and Worcester (discussed in chapter 7), and between Oxford and Wallingford (chapters 10 and 11).

The functions and origins of these connections are discussed in the chapters which follow (chapter 9 below, and further in chapter 14). In all of the cases analysed in this and preceding chapters, the patterns or orbits of distribution of estates with connections to burhs or boroughs can be seen to reflect the various ways in which the territorial administrative units grew and developed throughout the period from the late eighth to the early eleventh centuries, their remnants becoming fossilised in the folios of Domesday Book and the occasional pre-Conquest charter. Since these orbits are in every case constrained (and by inference determined) by the boundaries of these territorial units, there is little reason to hold that the distribution is in any way a reflection of the commercial hinterlands of the centres themselves regarded as urban institutions, as is often suggested (see chapter 1). As is shown in the next chapter, and as already indicated above, their distribution rather reflects the sometimes complex ways in which the central settlements were meshed into the nexus of tributary relationships of all estates and estate holders to the king, in ways which were determined by him in the pursuit of his military, economic and other agendas. These patterns of distribution can, therefore, if interpreted in the right way, be regarded as a tangible sign of the process of state-formation itself.

Summary and conclusions It has been the purpose here to show that the analysis of the spatial relationships of contributory manors to their centres, which are a ubiquitous feature of the Domesday account of boroughs, can help elucidate issues not only about their origin and function, but also about the disposition of regiones, burghal territories and shires, and the ways in which these administrative units in the landscape developed

69

Chapter 9 Urban-rural Connections – Function and Origins: a Model Urban-rural connections – function and origins: a model

fiefs in the shire, as Ballard has so clearly demonstrated (Ballard 1904, 31). In some areas, such as the shires of Devon, Dorset and Somerset, too few of these connections have survived to enable meaningful conclusions to be drawn from the distributions of these connections on their own. At the same time, the number of tenements in any borough recorded as being appurtenant to rural manors is probably only a fraction of their original tally, based on the admittedly somewhat uncertain premise that a burh or borough would have had approximately equal numbers of customary and non-customary tenements. Instances in the shires in the study areas given above are Gloucester, with 300 customary burgesses, and 301 non-customary (from Evesham K1); and Warwick, with 113 customary, and 112 non-customary. The original proportions are, however, debateable, and because of the inconsistencies and deficiencies in the Domesday record, are probably unrecoverable. David Roffe, for instance, notes that the number of urban liberties or sokes is probably under-represented in the Domesday record (Roffe 2007, 122). There is anyway unlikely to be any relationship of equivalence between the numbers of non-customary tenements in Domesday Book and the original non-customary tenements, since in many cases it is probable that one haga or tenement in a late ninthcentury burh, for instance, will have become subdivided into many through population expansion and consequent subdivision into tenement in the intervening two centuries. Such was certainly the case with the single large haga at Worcester of 904, discussed above, which came to be subdivided into 90 tenements at the time of Domesday. A similar development appears to have taken place with the haga in Gloucester appurtenant to Deerhurst, which by the time of Domesday had been divided into 30 tenements, and in the subdivision of the principal haga of the abbey of Wherwell in Winchester (see further in chapter 4) . Many of the connections included in charters in the tenth and early eleventh centuries have also been lost by 1068, or are simply not mentioned in Domesday Book. Instances of these in the four shires studied here include S.918 of 1008 relating to Cricklade; S.469 of 940 relating to Wilton; S.898 of 1001 relating to Warwick; and many others in Worcester (see table 20). It has also been pointed out above, in relation to all three centres of Worcester, Warwick and Gloucester, that the number of connections between rural manors and urban tenements could be more than doubled by

The issue as to why and how these connections between rural estates and the various centres were formed in the first place still remains. Any explanatory model must comprehend issues of functionality, as well as the evidence of spatial patterning and, not least, the constraints and limitations of the documentary evidence. As is argued above, a new way of seeing the origin of these connections is to suggest that they were brought about through the agency of the king and the earl of the shire working in partnership to create the burhs as sustainable communities, in ways which underpinned the king’s intentions to establish institutions which were designed to further his military, strategic, economic, social and religious agendas. At distinct historical periods the tributary functions of the population were focussed upon the royal burhs, and their enforcement of these obligations became the essential means by which the logistic and other aspects of the organisation of the burhs were maintained. This was effected through the allocation of these tributary functions on a local basis by the setting out of burghal territories, or what may be called ‘territories of obligation’. It was thus through the imposition of these burhs and their dependent territories onto pre-existing historical landscapes that the king was able to exert the political as well as the military control which he needed to pursue these agendas. The most clear-cut of these episodes of burghal formation as a controlling network is that covering Wessex, which is described in the Burghal Hidage. The existence of a similar network in western Mercia formed in the late eighth or early ninth century, which was developed further by King Alfred in c.880 (and which included a burh at London), can be argued from a broad range of archaeological, topographical and historical evidence. Similar processes can be inferred in those instances in which royal multiple estate centres, such as Calne and Droitwich, held tenements or burgesses appurtenant to nearby manors. However, one problem with this interpretation is that the number of tenants-in-chief in any shire who can be shown from the evidence in Domesday Book to have held properties in boroughs (even with the addition of those given in earlier charters) is only a fraction of those with

70

Urban-rural Connections

including inferred connections between the estates of large land-holders in the shires and their documented holdings of tenements in all these places.

connections between the landholders of the shire or burghal territory and the central burh, King Alfred (and Aethelred, Aethelflaed and King Edward after him) would have ensured the maximum degree of both control and loyalty in a situation where he would have needed all the support he could get. These territories can therefore be seen as ‘territories of obligation’, from which services to the king at the burh or estate centre would have been due from all the landholders in its territory.

There are, therefore, a number of reasons for inferring that, at the time of the creation of a burh and the formation of its burghal territory, every estate held by landholders above a certain standing within the area of the burghal territory would have been assigned a tenement in the burh, either as a large haga or soke or as a smaller burgage plot, according to the size and/or value of the estate. On this interpretation, therefore, the pattern of these connections as shown in Domesday represents an attenuated survival of a once notionally complete tally. As emphasised above, this is the only way in which the distribution of the contributing estates in relation to the shire and other boundaries can be explained. These appurtenant tenements would have formed a significant proportion of the population of the new burghal institution, in which the tenements held by the king (the customary tenements or burgesses of Domesday) would appear to have been roughly balanced by those held by the thegns and/or the tenants-in-chief who held manors in the burghal territory. This process would have created a situation in which these landholding thegns would have been bound to the king not only by their normal obligations of service, but also by the fact that the king had made available to them an asset (a tenement) within the centre of administration of the shire, burghal territory or multiple estate, thereby creating a situation of mutual and beneficial dependence. This act of state on the part of the king in giving tenements to be held by the thegns and tenants-in-chief of these territories can in this way be interpreted as one aspect of the long-established principle of reciprocity, in which the gift of land by the king constrained the thegn who received it to respond with a counter gift of loyal service (Abels 1988, 30-4). Richard Abels suggests that ‘In this way [a gift of] land cemented a man to his lord, compelling him to render whatever aid the lord might require’ (ibid., 31).

The establishment of these connections can be seen in a more general sense as an essential way in which the kings of the ninth and tenth centuries developed the bonds of lordship, to the end of the consolidation of their hold on territories and their access to the resources which this gave them. This theme of lordship has been discussed in detail by Richard Abels (Abels 1988, esp. 79-130). As he has shown, ‘the spread of book-right and the creation of bookland [in the ninth century] had altered the nature and composition of the fyrd, transforming it from the king’s retinue arrayed for war into an assembly of landowners and the contingents they owed in respect of their bookland’ (Ibid, 89; see also Abels 1984). This process is seen as the essential means whereby all ‘moot-worthy’ men – ie thegns or other landholders – were obliged to regard the king as their personal lord. The three common obligations of fyrdservice, borough-work and bridge-work, all of which were required in the granting of land by book (Brooks 1971; Abels 1988, 116-31,146-59), established the foundations of military obligation firmly on the lordship tie to the king, which was consolidated through the tenure of land by book-right (ibid., 96, 116-8). Thus all landholders were responsible for the maintenance and manning of a burh, as well as participating in the fyrd, their liability of service arising out of the tenures of their estates. However, the clear evidence from the creation of dykes by means of (it must be presumed) a general conscription in earlier centuries suggests that the creation of bookland merely formalised more ubiquitous obligations for service to the king which was based on folkland.

It would not be distorting the evidence of this practice at a rather earlier period to suppose that this principle of reciprocity would have underlain the practice whereby the thegns of the burghal territory received their grants of land to be held by book on the creation of the burh, with which was associated a new appurtenance of a tenement within the burh which became part of their estate. In this way the king was able to create new bonds of loyal service or to reinforce ones already established. Similar considerations must have underlain the origins of tenurial heterogeneity at royal multiple estate centres and of the attachment of contributing estates to them. In terms of defence, Asser himself is witness to the difficulties experienced by King Alfred in encouraging or coercing the cooperation of the thegns of the shire to fulfill their military obligations (Asser, chap. 91; Abels 1988, 76-7). By physically involving the thegns and lords in the setting out, functioning and upkeep of the burhs in this direct way by creating new tenurial

Richard Abels has also emphasised how, during the reconquest of the Danelaw by Edward the Elder in the early tenth century, the development of royal authority through the enforcement of personal lordship to the king was achieved through the submissions of populations with their lands to the king himself (ibid., 89-96). This is seen as the basis for the organisation of the layout of the Chronicle entries in this period (Haslam 1997). The veracity of this account as reflecting a political agenda which originated directly from the king is supported by arguments which suggest that this section of the Chronicle was written by the king’s brother while on campaign (Pelteret 2009). As Abels has emphasised, King Edward ‘not only demanded that the Danes “take him to lord” but required that that they submit to him with their lands.’ (Abels 1988, 89 – his emphasis). Through this submission, land tenure thus became the basis on which military obligation was enforced (ibid., 89-96). The case of the landholder of Huntingdon

71

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

in the early tenth century who lost his lands when he failed to submit with his estate at the right time and place (ibid., 89-90) emphasises the importance not only of the crucial role of bookland in this process and in the act of submission itself, but also of the essential connection between the burh and its dependent territory as the fundamental means by which political control was exercised.

sight that these arguments might not apply to the residents or tenants of royal estates, it is noteworthy that many royal estate centres in the shires already discussed are recorded in Domesday Book as contributing tenements to the various burhs at the centre of the burghal territories in which they lay. The survival of these connections to the time of Domesday is of course a function of the general removal of these estates from the late Saxon land market, but the fact of their existence emphasises that such connections were an attribute of all estates, whether royal or not. This in turn underlies the arguments made in chapters 3 and 4 above for the inclusion of the manpower on unhidated royal estates in the general conscription of the population for work on the burhs as well as for army service. As has been pointed out above (chapter 8), it is this process of general submission to the king which forms the essential back story to the implementation of a new polity in Mercia in c.880, and which arguably formed the context for a phase of burghal formation in western Mercia at that time (stage 2b in the timeline set out in chapter 8 above) as the principle instrument by which this submission could be enforced in practice.

Abels also recognises that the scheme of burhs outlined in the Burghal Hidage ‘points to a reorganisation and enlargement of Anglo-Saxon administration’ (ibid., 79). It would therefore be entirely appropriate to accept that the process of the creation of bookland and the submission of landholders within a burghal territory to a burh was as much a king-pin of royal policy in Wessex in the late ninth century, or even in Mercia in the eighth or early ninth century, as it so clearly was in the reconquest of the Danelaw areas in the early tenth. One must suppose that the granting or confirming of estates of thegns by book-right by the king, with which went the requirement to perform the three military obligations, was systematically applied as an act of royal policy by the king at the same time as the organisation of the landscape into burhs within their burghal districts. As pointed out above, these districts can for this reason be characterised as ‘territories of obligation’. It also follows that the holding of a tenement within a burh attached to these estates was established at this time as a necessary and universal adjunct to the act of submission to the king at the burghal centre by which the thegn acquired, or was confirmed in, the tenure of his estate by bookland. It seems highly unlikely, in view of the universality of these military obligations, that a situation in which some thegns were exempt from these obligations, while some were not, could have been allowed by the king. This process of the allocation of tenements within burhs to landholders can be most reasonably seen, therefore, as a means of binding the general military obligations of all thegns to a particular burh, so that representatives of each thegn’s estate, at one man per hide, were available to act as manpower for its proper and appropriate function and upkeep.

It is argued that this process of the establishment of the physical connections between the holdings of the estates of thegns within the dependent territory to a particular burh must be recognised for what it clearly was – perhaps the most fundamental and important factor in the way in which burhs of the late Saxon period were not only set up, organised and manned by the king, but were also given the essential support system which was designed to maintain them as institutions which were both effective and sustainable. The assignment of tenements in the burhs to all the thegns of the burghal district can, however, be seen as one aspect of a wider logistical issue, in that this association must have formed a very practical link between the provisioning of the men draughted in to construct and garrison the burhs and the rural estates which must have been required to provide their sustenance through a system of renders. This was arguably achieved by means of the assessment of all estates in ploughlands to provide provisions which enabled the men and families to perform these obligations, which factor is discussed below (this chapter). At one stroke these two factors created a new set of conditions by which both the new inhabitants of these new burhs who occupied the customary tenements, and the landholders of the burghal territory, were obliged by their terms of service to the king, and through the new lordship bonds which were created, to participate in and uphold the king’s overall military, economic and social agendas. A king who had – and frequently exercised – the power to take these assets from individuals if these terms were not met was in a very powerful position indeed.

It would fit the evidence of the urban rural connections adduced above in relation to the West Midland shires that the creation of these Mercian burhs would have involved a similar if not identical process. These are discussed by Steven Bassett in terms which, quite rightly, emphasise their role as key elements in the process of royal state formation, rather than any role as a systematic defence against Viking depredation, as originally argued by the writer (Bassett 2007; see Haslam 1987). Although Bassett recognises the importance of burhs as instruments of control (Bassett 2007, 82-3), he has not, however, considered either the formal process of submission of the population of the burghal territory to the king or the consequent formation of bookland as a part of this process. In Mercia, as in Wessex in the late ninth century and in the Danelaw in the early tenth, both of these aspects were arguably designed to consolidate the relationship of the thegns of the burghal territories to the king as lord. While it might appear at first

It can also be argued that the formation of these new connections in burhs can be seen as one aspect of the practice of the king in creating partnerships, as with the earl of the shire (through which he acquired his ‘third penny’), to facilitate and ensure the ongoing maintenance of the new

72

Urban-rural Connections

burh as a sustainability community (Tait 1936, 30, 61-5, 141-8; Stenton 1971, 534-5), and to bind the obligations of the king’s subjects to himself the more strongly. This aspect of partnership can be recognised in for instance the process of the creation of the new enlarged burh at Worcester in probably the early 890s, in which Ealdorman Aethelred and his wife Aethelflaed granted the bishop half of most of the assets of the new burh (S. 223; Whitelock 1979, 540-1; and discussed in for instance Tait 1936, 19-21; Brooks 1996, 143-4; Baker and Holt 2004, 175-7; Holt 2009, 59-60; Haslam forthcoming b). In this case, the charter makes it clear that Aethelred (acting as King Alfred’s regent, and at the request of the bishop) is the ‘senior’ partner, ensuring that the bishop was not able to short-circuit the process of the creation of these mechanisms by which the obligations of service to the king were set up and maintained.

order within the burghal space in places such as Winchester, Gloucester, Wallingford and Oxford. For Winchester, the detailed evidence for this is derived from the two twelfthcentury surveys as well as from Domesday (Biddle (ed.) 1976, 349-69; see also diagrams and plans in tables 11-19, 21, 22, & figs. 14-16, 19). This is discussed in more detail in chapter 4 above. This pattern has been ascribed by Biddle to the processes involved in the ‘apportionment of land in the city at the time of the reorganisation under Alfred’ (ibid., 349). Since it is virtually axiomatic that the customary tenements of the king (on which landgable was due) were those formed at the initial stages of the foundation of the burh, it follows that this must also have been true of the origins of the non-customary tenements which were appurtenant to rural manors, which in every street were intermingled with the customary tenements. This is shown graphically in fig. 16 reproduced here.

The same motives also appear to have been at work in the creation by King Alfred of the substantial soke in London in 889 for the use of the same bishop, discussed in chapter 7 above, which arrangement was arguably set up to facilitate the trade in salt from Droitwich within the king’s new burh, on the proceeds of which the king and the bishop had a virtual monopoly (Haslam 2010b, 128-30; the trade in salt is discussed in Maddicot 2005). The division of assets between the king and earl was a process described by F M Stenton as ‘a normal feature of Old English borough finance’, and by Nicholas Brooks as ‘a financial carve-up between the king . . . and the interested great lords’ (Stenton 1971, 534-5; Brooks 1996b, 143). It would be entirely reasonable to infer that this same process was extended to include all landholders of the burghal territory above a certain status at the time of the formation of the burh, and for the same reasons. The process in which the thegns and tenants-in-chief of the burghal territories would have been allocated tenements which comprised significant proportions of the occupied areas of the new burhs is also, from a purely logistical point of view, probably the best way in which the king could have populated the new burhs to an extent which ensured their strategic and economic viability as sustainable military and social institutions. As has been emphasised above, this also had a political dimension, in that this process would also have been arguably the most effective means by which he ensured that the bonds of lordship of the thegns to himself could be enforced, to the end of consolidating his political and economic control over territories and their landed resources in the most advantageous way. This aspect is further discussed in chapter 9, below.

A similar intermingling of tenements in Gloucester paying landgable to the king (which appear to be the successors of the customary tenements), with those held by the abbey (amongst others), is shown clearly in the survey of 1455 (Baker & Holt 2004, 279-81 and esp. figs. 10.1 & 10.2 – see fig. 17a and b reproduced here). The same pattern is shown in Oxford (fig. 18b), where the tenements paying landgable rents are distributed around the whole of both the western and eastern burhs – though the few recorded, mainly from the Hundred Rolls and abstracted by H E Salter (Salter 1936), must represent a small proportion of the original customary tenements held by the king in the pre-Conquest period. Similar conclusions can be drawn from the distribution of the non-customary tenements in Oxford belonging to Abingdon Abbey, all of which were probably appurtenant to some or many of their various estates. As is shown graphically in fig. 18a, these were intermingled with both the customary tenements and other non-customary tenements over the whole of the area of the primary burh (see also chapters 10-11 below). Evidence relating to similar patterns at Wallingford, derived from Domesday Book as well as from earlier and later sources, is described by David Roffe. The pattern of distribution of the individuals who paid landgable in a series of thirteenth-century rent rolls shows that these were distributed more or less evenly around the four wards of the borough, and the non-customary tenements show a similar distribution (Roffe 2009, 35-6). This evidence is extended by the pattern of distribution of tenements paying quit rents in the later medieval and post-medieval periods, which were the successors of landgable rents after the borough received its royal charter in 1155 (Pedgley forthcoming). The landgavel (hawgavel) rents at Cambridge also appear to have been evenly distributed around the borough in the 13th century, a situation which can probably be taken back to the pre-Conquest period (Maitland 1898, 180-2).

It follows therefore that this process – the results of which survived, albeit in attenuated form, to become manifested as the tenurial heterogeneity of the Domesday borough - was accomplished not by gradual accretion but by royal fiat as a single ‘act of state’ (if that term is not too anachronistic – since in practice the ‘state’ did not exist as a separate entity from the king and his witan). This is shown most strikingly by the way in which non-customary and customary tenements were intermingled in no apparent

It can be concluded from this evidence relating to these five places that the way that the customary and non-customary

73

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 16. The percentage composition of each street by holdings and rents of the seven major fiefs in Winchester in 1148. The left-hand columns show holdings of tenements; the right-hand columns show rents. (Low res. image). [Permissions]

74

Urban-rural Connections

Fig. 17a. Gloucester properties paying landgable in 1455 (reproduced from Baker and Holt 2004, 278). (Low res. image). (Reprinted by permission of the Publishers from ‘The major religious institutions: their lands and their role in urban growth’, in Urban Growth and the Medieval Church by Nigel Baker and Richard Holt (Farnham: Ashgate, 2004), fig. 10.1, p. 278.  Copyright © 2004).

Fig. 17b. Gloucester - St Peter’s Abbey properties in 1455 (Low res. image). (Reprinted by permission of the Publishers from ‘The major religious institutions: their lands and their role in urban growth’, in Urban Growth and the Medieval Church by Nigel Baker and Richard Holt (Farnham: Ashgate, 2004), fig. 10.2, p.280.  Copyright © 2004).

75

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 18a. Abingdon Abbey tenements in Oxford. Note the distribution of these solely within the area of the primary burh and its southern suburb. (Drawing copyright Dr James Bond, who kindly supplied this version).

76

Urban-rural Connections

Fig. 18b. Distribution of tenements paying landgable in Oxford, from information in Salter 1936, derived from the 13th century Hundred Rolls and other sources. The hached block in the centre represents the area in St Martin’s parish of an otherwise unlocated tenement. From information kindly supplied by James Bond. North to top.

77

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

tenements were intermingled within the layout of these burhs shows that in origin these two classes of tenement were contemporary, and that both classes were laid out in the space available within the burhs as part of the process of burghal formation. It is not possible to envisage a scenario in which spaces between the primary customary royal tenements, which would have been laid out along the new streets as part of the process of the initial spatial organisation of the new burh, were then subsequently filled in by tenements acquired by landholders of the burghal territory of the shire by accumulation and accretion over the long period before the time of Domesday. This particular spatial arrangement must have been by means of a process of allocation by a central and controlling authority, in which customary and non-customary tenements were laid out together as part of a single initiative of settlement-formation. This is of course not to say that this process would have involved a full-scale and immediate development of all the tenements with houses fronting onto the streets. That this did not happen is indicated, for instance, by the archaeological evidence from Oxford (Blair 1994, 152-67). Such developmental processes would have been separate to those which must have been involved in the initial allocation.

burh at Worcester was set up (Tait 1936, 19-21). Richard Holt’s alternative model, which holds that the new burhs were not set up as new markets but merely as fortified refuges (Holt 2009), can be questioned on a number of grounds (Haslam forthcoming b). Not the least of these reasons are the arguments set out above which suggest that the burhs were laid out as new settlements with customary and non-customary tenements intermingled along streets which would have been contemporary with the formation of the burh and its defences. It would for this reason have decidedly been in the king’s interests to have brought into being the new burghal market and its associated institutional and social structures, which would have had the effect of channelling the sale or exchange of the goods or agricultural surpluses from the thegns’ estates through this new market. In this way the king would have gained a new degree of control over access to particular markets, and the increased revenues from taxes and tolls which this would bring him. As Nicholas Brooks has argued, ‘the detailed economic provisions that accompanied the building of borough defences . . . involved the direction of trading activities into the boroughs so that market tolls, burgage rents and the fines imposed by a court would all be profitable to those lords [the king most of all]. Both the market and the borough court are essential components of the plans from the start.’ (Brooks 1996b, 143-4). The laws of the early tenth century confining trading transactions to burhs can therefore be seen as the development and reinforcement of a set of conditions which was built into the way a burh was set up from the beginning. Another aspect of the value to the king of these partnerships between himself and the thegns would have been to provide them with ready accommodation for attendance at the shire or borough courts (as Stenton has so eloquently pointed out in the quotation given in chapter 1), and to ensure the involvement of the thegns in the religious observances and ceremonies of the great churches of the burh, which processes are examined at length by Robin Fleming (Fleming 1993).

This is indeed a model which is applicable to the way in which every burh of the middle and late Saxon period would have been set out from the start. It is also gives some support for arguments I have recently put forward to the effect that the wall tenements of Oxford (the socalled ‘mural mansions’ of Domesday Book) were not only equivalent to the non-customary tenements distributed around the whole of the burh, but were also all appurtenant to rural manors. Burgesses in these tenements were distinguished as a group from the king’s customary burgesses by the fact that they alone were obliged to do wall work. I have made a case that it was on this group, and not on the customary burgesses, that these new obligations were imposed on the occasion of the extension of the primary Alfredian burh in c.911 (Haslam forthcoming c). If this is so, this provides a generally applicable terminus ante quem for the existence of the connections between rural manors and urban tenements as a group, a conclusion which is reinforced by the distribution of the tenements belonging to Abingdon Abbey in Oxford within only the primary burh (fig. 18a).

The importance of the trading and marketing functions of the burh is put in perspective by the case of non-burghal site of Droitwich (Worcs), discussed above. Like any burh this showed tenurial heterogeneity in having assets in the form of both salt works and tenements appurtenant to a number of rural estates, as well as to the nearby royal manorial centre at Wychbold. This Droitwich – Wychbold axis may be considered to be a special case of an early dispersed estate structure (which it has been found convenient to call a multiple estate), some of which in other places developed significant markets of their own. Some of these connections at Droitwich may well have arisen as early as the beginnings of the involvement of the Mercian kings and of the bishop of Worcester in its development from the seventh century, a process which has been examined in detail by John Maddicot (Maddicot 2005). Although the borough of Droitwich has been considered as anything but normal (Bassett 2008b), the structure and disposition of these tenements can be interpreted as being little different

Although one aspect of these obligations would have been to ensure the construction and garrisoning of the burhs (the essence of Maitland’s garrison theory), as well as their provisioning, another powerful engine which would have sustained the king’s objectives would have been the creation of new markets in which transactions could be concentrated, to the king’s ultimate profit, through the increased control which this gave him in the levying of tolls on buying and selling and the movement of goods. As historians have consistently pointed out, from the time of Tait onwards, this is an aspect of burghal formation which is demonstrated so clearly in the details of the way that the

78

Urban-rural Connections

in a functional sense to the customary and non-customary tenements seen in ‘normal’ Domesday boroughs.

no reason to suppose, therefore, that there would have been insufficient food available to support working parties, and then garrisons, in the burhs in the late spring and summer of 878, and into 879, after Alfred’s victory at the battle of Edington, as Baker and Brookes imply. There is no reason to believe that estate-holders who submitted to Viking control would have seen the productive capacity of their lands seriously reduced. This is not, therefore, an argument which necessarily demonstrates that insufficient resources would have been available in this crucial window of opportunity, and that the burhs could not as a result have been built within the time frame suggested.

The logistics of burh construction What must not be forgotten in all these processes are the ways in which the manpower which was diverted to the construction, maintenance and garrisoning of the burhs was supported with enough provisions – both of food and of working materials - to make the whole enterprise workable. In their recent paper, David Baker and Stuart Brookes quite rightly observe that ‘Calculations of the nutritional and supply requirements associated with medieval military and building enterprises show that the obvious, immediate manpower needs only constituted one part of the total drain on resources.’ (Baker and Brookes 2011, 112-3). Important issues concerning the logistics of the supply of food, horses and other requirements for the upkeep of armies and fortresses have been examined closely by Richard Abels (Abels 1996). Clearly, sandwiches and flasks of tea would not have been enough to have sustained the building or reconstruction of nearly 40 km of burghal defences by the labour of men from nearly 30,000 hides in Wessex, at one man per hide, as well as the construction of ancillary works such as the laying out of streets, and the building the bridges with which so many burhs appear to have been associated. At a slightly later date the earth and timber defences of many if not most of the burhs around both Wessex and western Mercia were reinforced with stone walls, which, as I have pointed out, would have been an exercise in logistics, procurement and the allocation of labour which must have been every bit as ambitious as the construction of the original defences of earth and turf (Haslam 2009, 104).

Clearly, though, the whole enterprise had to be supported by the agricultural and other resources of the land. The division of shires into burghal territories, which is so neatly indicated by the Burghal Hidage, carries the implication that each territory would have been responsible for sustaining the building and garrisoning of the burh of which it was the centre. This conception is reinforced by the arguments given above for the importance of obligations, based on the bonds of lordship, of all land-holders to support and provide services for the upkeep of the particular burh in whose burghal territory their estate lay. This is in itself an argument which militates against the possibility, suggested by Nicholas Brooks, that working parties could have been moved around from burh to burh at will to facilitate their construction and upkeep, irrespective of the allocations of the hides of particular estates to particular burhs (Brooks 1996b, 137-8). It is therefore almost axiomatic that each estate would have been required to provide renders in kind – whether of arable origin or as livestock, and possibly even as raw materials – for the daily support of the men and families working in them and defending them.

One factor which is open to question, however, is Baker and Brookes’ assumption that these factors would have been so onerous as to show that the system of 31 burhs listed in the Burghal Hidage could not have been built in the 18 months or so between early 878 and late 879, which I have suggested was the most appropriate window of opportunity for their construction – in spite of clear evidence in the Chronicle of various episodes of burh-building in far shorter time spans (Haslam 2005, 132-3). They have argued that the Viking occupation of Chippenham in early 878 would have resulted in a ‘far-reaching and hostile subjugation of Wessex, with many people driven into exile’, a series of events which would have led to Wessex as a whole becoming ‘war-ravaged’ as a result (Baker and Brookes 2011, 113). This has to be put in perspective, however. Firstly, nothing in the Chronicle account suggests that anything more than an area around Chippenham – perhaps north Wiltshire and parts of Somerset – was subjugated as a result of the Viking occupation. Secondly, it is questionable whether any part of Wessex would have been ‘war-ravaged’ or at all devastated as a result. The aim of the Vikings would have been to control territory, and to live off the agricultural or other resources they found, rather than to destroy them. There is no need to envisage that agricultural production would have been decimated, and

Mention has been made above (chapter 3) of the universality of the assessment of estates in Domesday Book by ploughs or ploughlands, and that this assessment appears, at least in Wiltshire, to be proportional to the hidage of each estate. The interpretation of the significance of the ploughland has been described as ‘the most complex of Domesday problems’ by David Roffe (Roffe 2007, 203), who has summarised the widely disparate views as to its origin (ibid., 203-9). He has suggested that the ploughland should be regarded as a ‘non-fiscal measure of fiscal land’, ‘a fiscal hide’ rather than a statement of an area of land per se , or ‘a measure of the capacity of the hidated land to pay the geld assessed on it’ (Ibid., 207). While the details of this do not appear on the face of it to account for the ploughlands on unhidated land, for instance in the royal estates paying the farm of one night in Wiltshire and Hampshire and elsewhere mentioned above (chapters 3 and 4), the sense of the ploughland being a unit of assessment of some kind, rather than a measure of land, appears to remain a common denominator in Roffe’s analysis. A similar conclusion, that ploughlands were an assessed measure of agricultural productivity, has been drawn by Michael Costen in the relation to the analysis of the Domesday figures from Somerset (Costen 2011, 128).

79

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Roffe does not discuss the origins of this assessment, if such it be, or the possibility that ploughlands represented a form of assessment at an earlier date, except to suggest that the ploughland figures would have been used in the process of negotiation by which hitherto untaxed demesne lands were to be subject to the geld in probably the 1080s or 1090s. However, Cyril Hart, following J H Round and F H Baring, has argued that the ploughland was a unit of assessment in the eastern Midlands prior to the institution of the assessment in hides – i.e. that this was probably used as such in the ninth century. He does not, however, discuss the possibility that ploughlands were an early assessment for renders to support the logistics of work on the king’s works (Hart 1970, 24-32; Hart 1974, 38-45; Darby 1977, 117). Without recapitulating the many discussions on this topic which have fuelled what is clearly an unresolved debate, it could be suggested as an alternative model that the ploughlands recorded for most estates in Domesday Book are the developed (and perhaps changed) forms of an original assessment of agricultural productivity which was put in place to determine the obligations of all estates to provide renders in kind, in general to the king, but specifically to support the king’s enterprises such as the building of burhs and/or other works and the support of the fyrd in the field. This was perhaps originally the function of the hide, but the assessments on land in hides, at least in Wessex, for the purposes of service and obligation might well have become separated from a more-or-less realistic assessment of the productive capacity of land for the purpose of determining renders for the supply of provisions for the labourers on the king’s works.

assessment would, for purely pragmatic reasons, have been necessary at an earlier date than the time of the formation of the Burghal Hidage system in the late ninth century. This is perhaps supported by the record, for instance, in the early ninth century of a render of cows and honey from an estate near Evesham which is payable at the royal centre (and burh) at Winchcombe (Finberg 1972, 229). Just such a system of renders from estates would also have been necessary as an essential aspect of the logistics of the building of early or middle-Saxon dykes around the country. The Middle or Late Saxon burh can for all these reasons be best seen as an instrument of political, economic, social and military consolidation, whether in late eighth- or early ninth-century Mercia or in Canterbury, in late ninthcentury Wessex, or in early tenth-century Mercia or the Danelaw, which enabled to the king to exert a new level of ‘institutional coercive power’ to achieve his aims, by the consolidation and enforcement of the bonds of lordship (Abels 1988, 80). The role of these burhs as markets and jurisdictional centres can likewise be seen as essential aspects of the ways in which this political control in all its ramifications was exercised in practice. As Richard Abels has put it, the burhs were ‘islands of royal power through which the king and his agents, earldormen, bishops, and reeves, were able to dominate the countryside’, and the means by which they ‘reinforced the traditional connection between landholding and military obligation to the crown’ (ibid., 80). Abels goes on to make the point that ‘the burghal system’s most lasting consequence may have been the enhancement of the institutional power of the West Saxon monarchy over its subjects’ (ibid., 208). The case has been made above that this was, in essence, achieved through the greater degree in which the king was able to consolidate the bond of lordship of his subjects to himself. This coercion would have been consolidated through the gift of a tenement within the burh, which enabled the king to set up the burh as a functioning institution with either the willing or the enforced support of the thegns and lords of the burghal territory in all of its various functional aspects discussed above. The bonds of lordship would also have underpinned the way in which the supply of food renders from each estate would have been enforced, and the physical location of the estate’s tenement within the burh would have facilitated their distribution.

As a working hypothesis, therefore, it is possible to see the ploughland as an assessment on every estate to determine its obligations to provide renders in kind in the context of the new administrative exigencies represented by the creation the system of burhs which are listed in the Burghal Hidage. These renders would have been supplied directly to the burhs themselves, or to the royal estate centre in such cases as Langport or Axbridge in Somerset where the burhs were so clearly the dependencies or appendages of the neighbouring royal estate. Furthermore, it is possible to see the institution of the connections of urban tenements to rural estates as falling within the same nexus of tributary relationships, in which context the thegns holding estates within a particular burghal territory were welded even more closely to the support of the central burh through the fact of their being granted tenements within it by gift of the king.

The logic of these arguments implies that these connections between urban tenements and rural manors, as they have survived to be recorded in much later sources, are but a pale vestige of patterns of connection and obligation which originally involved all landholders of the territory of the burh to which they would have owed allegiance. On the interpretation of the development of burhs and their territories in western Mercia given above (chapter 8), the creation in c.880 by King Alfred of his new polity by which Wessex and Mercia were merged to form the ‘Kingdom of the Anglo-Saxons’, a new burghal system (with burhs at

Whether this assessment was new at the time, or whether the ploughlands represented old assessments which were utilised, and perhaps modified, anew for a new purpose, is probably unknowable. However, the widespread use of ‘royal’ fortifications before the late ninth century in Wessex (Baker and Brookes 2011, 106-8; Draper 2011; Haslam 2011b, 199-204), and the obligations for the holders of estates to perform the three common obligations since at least the late eighth century in Mercia (Brooks 1971; Abels 1988, 75-96), would suggest that such a system of

80

Urban-rural Connections

Gloucester, Winchcombe, Worcester, Hereford, Tamworth and London, and possibly Bristol) can be seen as an essential part of the way in which this new polity would have been established on the ground. The creation of new bonds between estate-holders and the burhs which this would have entailed would have been seen as an essential part of the way that the lordship of King Alfred was established and consolidated within western Mercia, as it had been in Wessex.

partnerships would have been built in to the processes by which the burh was set up from the beginning to reflect and facilitate the intentions and the agendas of the king, and in particular, as discussed above, to enhance the bonds of lordship. Indeed, the need to consolidate these bonds of lordship can be seen as the driving force which underpinned the creation of the burghal systems in both Wessex and Mercia, from the late eighth or early ninth century onwards, and which moulded the patterns of obligation of the men and the thegns of the shire to the king. This model provides a new view of the way in which new burghal institutions were created as effective military and political instruments, as successful new markets, as centres of justice often with their own courts, and, in many cases, as new religious centres. It also focuses attention once again on one of Maitland’s insights into the nature of the Anglo-Saxon burh - a view lost in Tait’s critique of Maitland’s position:

The model outlined above, therefore, embraces the essence of Maitland’s ‘garrison theory’, in the sense that one of the objectives of these arrangements would have been to bring into play the obligations for the defence and the ongoing garrisoning of the burhs, which the creation of the connections between these manors and their urban holdings would have ensured at a practical ‘hands-on’ level. This also applies to the patterns of obligation of landholders to pre-burh royal estate centres such as Calne, which is demonstrated, for instance, in the complex of services owed to the early royal centre of Taunton (Somerset) by its dependencies which is shown in Domesday Book (DB Som 2,2-3,9) (Roffe 2007, 274-5). It also embraces Tait’s viewpoint that wider marketing and other interests of the thegns and tenants-in-chief would have played a significant part in the survival of the patterns of the connections which have come to be recorded in Domesday Book. The main difference between the model proposed here and that of Tait and his legion of followers is that these multi-faceted

“In the ancient boroughs there is from the first an element that we must call both artificial and national. The borough does not grow up spontaneously; it is made; it is ‘wrought’; it is ‘timbered’ [defended]. It has a national purpose; it is maintained ‘at the cost of the nation’ by the duty that the shire owes to it. This trait may soon have disappeared, may soon have been forgotten, but a great work has been done. In these nationally supported and heterogeneously peopled towns a new kind of community might wax and thrive.” (Maitland 1897, 219).

81

Chapter 10 Oxfordshire, Buckinghamshire and Berkshire pattern of connections with Oxford. The distribution of the connections of rural manors with Wallingford, which include a number in the south-east part of Domesday Oxfordshire, has also been examined by David Roffe, who has used this evidence to draw important conclusions as to how the original burghal territories of Wallingford and Sashes in the ninth century spanned the Thames in a way which was very different to the arrangement of the later shires (Roffe 2009, 39-45). Keith Bailey has made a similar inference regarding the burghal territory of Sashes as covering the southern half of later Buckinghamshire and the eastern part of the later Berkshire (Bailey 1994, 89).

In the following chapters the basic model of the relationships between the contributory manors and their centres which has been formulated in the preceding chapters is tested against the pattern of distribution of the manors which were contributory to the three centres of Oxford, Wallingford and Buckingham, and the relationships of these patterns to the shires of Oxfordshire, Berkshire and Buckinghamshire and their progenitors at various periods from the middle Saxon to the early post-Conquest period. These three shires are considered together, for reasons which will become clear. The folios of Domesday Book for Oxfordshire and Berkshire contain information relating to the contributing estates and their appurtenant tenements in Oxford and Wallingford both in the sections on the boroughs and in various entries for the rural manors. Connections recorded in Domesday of rural manors with Buckingham are not directly stated, but can be inferred with various degrees of certainty for the reasons given in the table below. The two Buckinghamshire manors with appurtenant tenements in Oxford (Princes Risborough and Twyford) tell a different story, and with other apparently anomalous instances generate important inferences which are supportive of the general thesis discussed in this study. As is well known, the orbits of the urban-rural connections of both Oxford and Wallingford not only overlap to some degree, but also extend beyond the borders of the Domesday shires. These overlapping orbits are arguably peculiarly indicative of functional developments at an early stage in the formation of burhs and shires, which patterns have been discussed in relation to the development of the burghal territories and shires of Wiltshire, Hampshire and the West Midlands. The analysis which follows will serve to test further the basic hypotheses which have already been articulated in relation to the evidence from these shires in previous chapters.

These conclusions have formed the basis for a fresh examination of the evidence from the three shires together, from which it is possible to reconstruct the extent of the burghal territories of Wallingford, Oxford, Sashes and Buckingham - all original burhs of the late ninth-century Burghal Hidage system – in a way which extends and modifies these insights. The patterns of distribution will be discussed after the evidence from Oxford and Wallingford is presented. Together with other evidence, in particular the development of minting and the coinage of King Alfred and Kings Burgred and Ceolwulf of Mercia in the 870s, this reconstruction generates important conclusions concerning the political status of this area of the central Thames region in the late ninth century. This has an important bearing not only on the interpretation of the historical context of the Burghal Hidage and of the burhs it lists, but also on the way in which the West Saxons under King Alfred and his son and successor King Edward the Elder were able to extend their hegemony over southern Mercia and into the Danelaw to the north in the late ninth and early tenth centuries. The connections set out in the accompanying tables are shown graphically in fig. 20.

These patterns relating to Oxford and its neighbours have been examined by E M Jope (Jope 1956, 245), as well as by John Blair (Blair 1994, 117-9, 155-8). Jope’s map, though a pioneering statement which emphasised connections seen on a regional basis perhaps for the first time, is based only on connections directly stated in the documented evidence. It does not include reasonable inferences about such connections, such as most of the connections of rural manors with Buckingham. Both Blair’s list and his drawn diagram include several inferred instances which are drawn from various non-Domesday sources, which extends the

Oxford One of the notable features of the Domesday entries relating to Oxford is the incidence of the wall tenements (‘mural mansions’). These tenements are those which, in the words of the Domesday scribe, “if there is need and if the King commands, they repair the wall.” They were, furthermore, “free from all customary dues except [military] expeditions and wall repair”. As John Blair has remarked, it appears that this function was shared by most if not all of the

82

Oxfordshire, Buckinghamshire and Berkshire

Table 21. Customary and non-customary tenements in Oxford Contributory estate [ ] = inferred

No of estate in fig. 20

Number of tenements in Oxford 243 (+478 waste) 50

LORD / holder of tenements The king

DB section B4

Robert d’Oilly

28,8; 28,28

Number of manors in Oxon

Notes & Comments Customary tenements Customary tenements belonging to a separate urban fee (extra-mural and intramural) – Haslam 2010c, 29-30.

Section B5 (wall tenements) [Broadwell]

1

20

The king (formerly held by Earl Algar)

Blair 1994, 157. Non-customary tenements TRE – appurtenant estate(s) not given

B5. Probably also former comital 1,9. holdings. 58,15 Bloxham 3 1 The king 1,7a-b Harold’s house TRE Bucks Princes Risborough 4 1 The king 1,3 (Bucks) Bucks Countess Godgifu’s houses Twyford (Bucks) 5 2 The king 37,1 TRE “The reason that they are called Wall-messuages is that if there is need and if the King commands, they repair the wall.” [A scribal gloss, but not postscriptal] Section B6 18,1 Iffley 6 2 (+ St The king Blair 1994, 158 n. 63. Mary’s ch) (formerly held by Earl Aubrey) B6; Burford 7 1 7,36 Section B7 The king ? 9 (formerly held by B7 Earl William) Section B8 Jope & Terrett 1971, 231; ?Newington Blair 1994, 157 favour Archbishop of ?Monks Risborough 2,1 1 8 7 Newington – but see Canterbury (Buk) discussion below. 3,1 9 Bishop of [Witney] and/or + 2 9 Winchester [Adderbury] 3,2 10 65; none in ? 18 Bishop of Bayeux 7 Berks [Banbury], 41 Formerly bishop of [Cropredy] 42 30 Bishop of Lincoln 6 Dorchester. [Gt Milton] 43 Bishop of Berks None in Ox, [Compton] (Brks) 11 2 Coutances 6 1 in Berks Bampton formerly held by [Bampton] Bishop of 5 None Bishop of Exeter. 12 3 Hereford Blair 1994, 157 n. 56. Taynton St Edmund’s 13,1 1 Blair 1994, 158 n.60. 13 1 Abbey Salter 1936, 16; Blair 1994, 151, 157. Salter 1933, 301; Blair 1994, Lyford, Brks (1) 14 157. 14 (+ St Abingdon Abbey 9 7 Distribution of manors in Martin’s ch Tadmarton (1) 15 both Oxon and Berks shown in Bond 1979, 61; Kelly 2001, clv. Shipton under Whychwood

2

1

The king

83

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Contributory estate [ ] = inferred

Eynsham

Section B9 [Horley] Pyrton [Buscot] and/or [Drayton] (Brks)

No of estate in fig. 20

Number of tenements in Oxford

16

13 + St Ebbe’s ch

Eynsham Abbey

17

10

Count of Mortain

18

7

19 40

LORD / holder of tenements

Earl Hugh (of Chester)

DB section

6,6-8

16 15,1-5 Berks 18,1-2 Oxon 17; Berks 13 Oxon 24; Berks 21 23,1 23,2 41,1

?

1

Count of Evreux

?

2

Henry of Ferrers

4

William Peverel

2

Edward the sheriff

3

Arnulf of Hesdin

1 2

Berengar of Tosny Miles Crispin

3

Richard of Courcy

32

12

Robert d’Oilly

28

Crowell and/or Emmington North Aston and/or Hempton [Black Bourton] [Ludwell] [Chipping Norton] ? ? [Nuneham Courtnay] [Sarsden] [Foscot] ? Includes Whitehill

20 21 22 23 24 25 26

27 28 29

41,2 40, 1 40,2 40,3 34 35

Number of manors in Oxon Manors held by Columban from the bishop of Lincoln 1

5 (before 1066)

Notes & Comments Blair 1994, 151 n.21; 158, n.58. These probably originated as one grant in 1005.

Blair 1994, 158 n.65 Salter 1929, 13-14; Salter 1933, 301-2; Blair 1994, 157 n.55; Sturdy & Munby 1985, 50. See discussion below See discussion below

8 in Ox; 13 in Berks 7 in Ox; 22 in Berks 2 2 3 3 34 3 in Ox; none in Berks

29 23 in Ox; 5 Blair 1994, 158 n.59 30 15 Roger d’Ivry 29 in Berks 1 (from the [Milton-under31 1 Ranulf Flambard 14,6 king) Wychwood] 1 in Ox; Guy of 36 none in [Wroxton] 32 2 Raimbeaucourt Berks 10 in Ox; ?Includes [West See discussion below 38 17 Walter Giffard 20 2 in Berks Hanney] Hampton 33 1 Gernio 58,16 1 Bletchingdon 34 1 Manasseh’s son none “All the above [in sections B8 & B9] hold the aforesaid messuages free, because of repairing the wall. All these called Wallmessuages before 1066 were free from all customary dues except [military] expeditions and wall repair.” Section B10 14,1 [Manor ‘near Canons of St 2 Oxford’] 15 Frideswide’s 14,2 [Cutteslow] 35 Priests of St none ? 2 Michael’s 58,19; Saewold of 58,315 ? 9 Rofford 4. Probably non- customary 54 Others tenements, i.e. wall tenements – Blair 1994, 155-8.

84

Oxfordshire, Buckinghamshire and Berkshire

Contributory estate [ ] = inferred

No of estate in fig. 20

Number of tenements in Oxford

LORD / holder of tenements

DB section

Number of manors in Oxon

Notes & Comments

“If the wall is not repaired when needed by him whose job it is, either he shall pay the King 40s, or he loses his house.” The king (formerly held by Berks Steventon, Berks 36 13 Earl Algar, now B1,39 by R d’Oilly) Geoffrey de Berks Streatley, Berks 37 1 Mandeville 38,6 Borough not specified, Berks but Oxford inferred (see Faringdon, Berks 39 9 The king 1,34 discussion)

non-customary tenements (Blair 1994, 157). The wider importance of these wall tenements was highlighted by Maitland, who saw this class as the principal evidential base for his so-called ‘garrison theory’, by which he sought to explain the origin of heterogeneous tenure of the Domesday borough, discussed above (chapter 1). I have argued, however, that these wall tenements were a feature of Oxford alone, and cannot therefore be used as evidence for more general processes relating to either the development of heterogeneous tenure or the formation of burhs or other estate centres, which is the subject of this study. I have also argued that these tenements were distributed around the whole of the burh, intermingled with the customary tenements, and that this service of wall-repair was placed upon all the landholders of the original burghal territory of Oxford who held tenements within the burh, most probably on the occasion of the extension of the burghal defences to the east (Haslam 2010a, 23-8). The most likely date for this is the occasion in 911 when Edward the Elder is recorded in the Anglo-Saxon Chronicle as having ‘succeeded to London and Oxford and all the lands which belonged to them’ (Whitelock 1979, 211; Haslam 2010a, 25-8). One important outcome of this conclusion is that it can be inferred that the connections between the wall tenements and the rural manors to which they were appurtenant were already in place by the time of the extension of the new burh to the east in c.911, and are, therefore, likely to have been set out as part of the urban landscape of the original burh from its inception.

document of 1227 as well as those already identified by H E Salter. To this total may be added two identified by John Blair: a) a ‘court’ (curia) around St Ebbe’s church, given to Eynsham abbey by Ealdorman Aethelmaer, its founder, in 1005; this may well have been an Oxford tenement (in this case a comparatively large haga) which had already been appurtenant to Eynsham; and b) a ‘little estate’ (praediolum or haga ) around St Martin’s church at Carfax to the north of Queen Street which was appurtenant to Lyford (Berks), which was given to Abingdon Abbey in 1032 by King Cnut (Blair 1994, 151-2). The latter manor looks very much like a forfeit to the king by some unknown thegn, rather than an earlier royal estate. Another instance is the manor of Tadmarton in N Oxfordshire, the appurtenant tenement(s) to which must have been amongst the 21 belonging to Abingdon Abbey, all of which lay in the central part of Oxford (fig. 18; see Bond 1979, fig. 7 – reproduced in Blair 1994, 160 fig. 92 – and further discussion below). Three others are represented by two tenements plus St Mary’s church (near or on the original east gate of the primary burh) held by Earl Aubrey, which ‘lie with the lands of St Mary’s church’ (which lands are identified by Blair as Iffley manor), with a third stated as being appurtenant to Burford. These six non-customary tenements or hagae are not specifically stated as being wall tenements, and in only a single case are the wall tenements cited by Turner stated to be appurtenant to a particular rural manor. However, since the identity of the two kinds of tenements can be inferred from the other entries in the Domesday account, these six can be reasonably added to Turner’s 41 to make a total of 47 identifiable tenements which were, or had been in the past, in all probability appurtenant to rural manors. Of these, however, only 23 can be shown to lie east of Catte St / Magpie Lane, the line which marks the beginning of the new eastern extension (fig. 18). Furthermore, this total of 47 identifiable tenements in the area of the whole of the defended and extra-mural area of Oxford is just under 16% of the total of 298 enumerated in the Domesday entry (above) – not including the extra 19 which may be reasonably considered to have belonged to un-named manors belonging to Abingdon Abbey, as argued below. Furthermore, the total of only 23 tenements lying to the east

This is supported by a number of other considerations. The distribution of known wall tenements has been mapped from the available evidence - sparse though it is – by Hilary Turner (Turner 1990), who has attempted to demonstrate that most of them lie within the eastern extension of the burh at Oxford, in support of the hypothesis that the system of allocating the duties of the repair of the wall was occasioned by the process which brought this new defended enceinte into being (Blair 1994, 158; Brooks 1996b, 142; Munby 2003). Although this conclusion may well be correct, this is not demonstrated by this evidence or its interpretation. Turner has listed and mapped a total of 41 wall tenements known to her, including those given in a

85

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

of Catte St / Magpie Lane is less than half of the 47 which are identifiable in the two burhs put together.

These will, however, be discussed after the evidence from Wallingford and Reading has been presented.

This is therefore hardly a statistically valid sample to support, let alone demonstrate, Turner’s conclusion that ‘This concentration in the NE ward of properties liable to wall-service may show the transference of the duty from the men of the shire (and the manors in the county) to the town dwellers…’ (Turner 1990, 77). Nor is it consistent with Julian Munby’s conclusion that ‘a significant majority of possible “mural mansions” occur within the eastern half of Oxford, possibly implying an imposition on landowners within the extended area of the borough’ (Munby 2003, 24). These figures demonstrate, on the contrary, that the 298 known non-customary wall-tenements in Domesday are likely to have been distributed more or less evenly around the whole of the two parts of the borough, intermingled in no apparent order with the 721 customary tenements of the king. This is demonstrated by the distribution of the tenements belonging to Abingdon Abbey, all presumably non-customary tenements appurtenant to many or most of the abbey’s manors, which are dispersed around the whole of the primary burh of Oxford (fig. 18a), and which are discussed further below. This is paralleled by the distribution of the few tenements paying landgable in the medieval period, shown in fig. 18b, which are the presumed remnants of the king’s customary tenements of the pre-Conquest period. There is no reason to believe that the arrangement at Oxford would have been any different from similar patterns shown for instance at Winchester, Gloucester and Wallingford, discussed further in chapter 9 above. A further corollary of this is that there is on this evidence no demonstrable spatial and therefore functional connection of the wall tenements with the construction of the eastern extension of the burh in c.911, although I have concluded on other grounds that the imposition of this service was indeed occasioned by the new obligations for labour services which were required by this event (Haslam forthcoming c).

Wallingford and Reading, Berkshire The tenements in Wallingford appurtenant to rural manors have been listed most recently in David Roffe’s examination of the Domesday entries for Wallingford, though he omits from his list the urban tenements and their holders which can be inferred, but not directly stated, as being connected to rural manors (Roffe 2009, 42-5). All these are given here for completeness and in order to relate them to the pattern of distribution given in fig. 20. As with the case of Oxford, it is possible to make a number of reasonable inferences about these connections beyond what is directly stated in the text. Other possible connections from later sources are also discussed by Roffe, and are included here. Discussion – the pattern in Oxfordshire and Berkshire The basic premise that the burghal territories of Oxford, Wallingford and Buckingham, as well as the non-urban Sashes, interlocked in an arrangement which was somewhat different from the Domesday shires of Oxfordshire, Berkshire and Buckinghamshire, has already been put forward by a number of commentators. The possibility that the territory of Sashes extended over the Thames into southern Buckinghamshire has been examined by Arnold Baines and Keith Bailey (Baines 1984; Bailey 1994, 89). This arrangement has also been examined by David Roffe, who also suggested that the original burghal territory of Wallingford covered southern Oxfordshire (Roffe 2009, 42-5). The premise that the burghal territory of Oxford extended into northern Berkshire is examined in this chapter. It is important to recognise, therefore, that there will have been connections, both stated and inferred, of manors within Domesday Oxfordshire to Wallingford, and connections of manors within Domesday Berkshire to Oxford. There were also connections of manors within Domesday Buckinghamshire with Oxford, though not with Wallingford. Based on the model already articulated for other shires in previous chapters, it can be argued that the pattern of distribution of these connections across the boundaries of the three shires is of considerable evidential value in determining these earlier arrangements which predated the formation of the shires in their final form. It is also argued here that some of the more anomalous instances of these connections could be the survivals of connections within an even earlier pre-burghal territory or regio or provincia centred on Oxford, which can be taken back to the early ninth century if not earlier. This is discussed in chapter 13.

In this regard, the pattern of distribution of tenements held by Abingdon Abbey within the primary burh of Oxford, shown in fig. 18a (after Bond 1979, 72 fig. 7), is particularly significant. In the first place, of the 21 tenements or groups of smaller tenements in Oxford, two are documented as being appurtenant to two rural manors - Tadmarton in Oxfordshire and Lyford in Berkshire. It can be inferred from this that all the other tenements of the abbey situated in Oxford were appurtenant to some or most of the rural manors held by the abbey in Oxfordshire and the northern part of Berkshire. This distribution, therefore, not only extends the number of these connections in Oxford as a whole by nearly half as much again, but also demonstrates how these non-customary tenements were distributed moreor-less evenly around the burghal space. This aspect has been discussed in chapter 9 above.

It is the intention in the following discussion to determine how the evidence of these connections, and inferences which can be drawn from them, can validate or modify the basic premise put forward by Baines, Bailey and Roffe. In many cases, because the evidence is either incomplete or lacking, such inferences can only remain as unverifiable

There are other features of this list which require comment.

86

Oxfordshire, Buckinghamshire and Berkshire

Table 22. Customary and non-customary tenements in Wallingford, Berkshire Contributory estate and shire O = Oxfordshire [ ] = inferred

No of estate in Fig. **

-

Number of tenements in Wallingford 276 5

Long Wittenham

1

1 acre

[Willington]

43

1

Non-customary tenements Section B2 [Lewknor] 44 (‘Oxford’)

7 in 2 acres

Brightwell

2

27

Newnham Murren (O)

3

20

Gt Haseley (O) North Stoke (O) Chalgrove (O)

4 5 6

6 in 1 acre 6 1 1

Sutton Courtenay

7

Bray

8

?

LORD / holder of tenements

DB section

The king

Held by Miles Crispin

B1; 1,22

Nigel, held from Henry of Ferrers

B1; 21,8

Abbot of Abingdon Bishop Walkelin (of Winchester) Miles Crispin

11

? [Rycote] (O)

12

? Chilton

13

?

7 in Ox; 40 in Brks Bks 2,3

See discussion below Roffe 2009, 31, 38, 42.

Ox 35,11 9 in Brks; 33 in Ox

Miles Crispin Miles Crispin Miles Crispin Miles Crispin

Ox 35,2 Ox 35,10 Ox 35,6

6 in 1 acre

Miles Crispin

1,37

11 in 1 acre

Miles Crispin

1,22

1

Gilbert of Ghent

2

Ralph son of Seifrid

33,9

1

Hugh of Bolbec

1

Ranulf Peverel

6

Walter son of Other

31,2

1 piece of land

William Lovet

26,1-3

49 Ox 25,1

Section B4 East Ilsley

14

3

Henry of Ferrers

21,2

Brightwalton

15

5

Battle Abbey

15,1

87

Notes & Comments Customary tenements Customary tenements – probably gifts from the king Long Wittenham is a manor of Walter Giffard – 20,3 This is the only one of Henry’s manors held by Nigel

Various tenants in chief

“All this land belongs to Oxfordshire; however, it is in Wallingford lands’ Section B3 11 Rainald 1,35 Albury (O) 9 in 1 acre Archbishop of Ox 2,1 ?[Newington] (O) 10 6 Canterbury 10 Walter Giffard in 1 acre ? 4 Robert d’Oilly [Langley]

Number of manors in Berks / Oxon

Manor held by the king, but Miles Crispin implied in the text Manor held by the king, but Miles Crispin implied in the text

None in Brks; 1 in Ox 2 in Brks; 9 in Ox 1 in Brks; 2 in Ox

See discussion below See discussion. Manor identified in Notes. Tenement could however be appurtenant to Ewelme (Ox 38,2)

3 in Brks; none on Ox None in Brks; 1 in Ox None in Brks or Ox Connection of the 6 tenements with the manor given only under the manor 3 in Brks; none in Ox

Connection specified in the entry of the manor

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Contributory estate and shire O = Oxfordshire [ ] = inferred

No of estate in Fig. **

Number of tenements in Wallingford

LORD / holder of tenements

DB section

None in Brks or Ox

3

Bishop Peter of Chester The king

6

Henry of Ferrers

22 in Brks

?

1

? ?

none in Brks; 6 in Ox,

[Thame] (O)

16

1

Bishop of Lincoln

[S Weston] (O)

17

1

Earl Hugh

1 1 1 5

Godric Doda Algar The smiths

2 2

The king Count of Evreux

1

Hugh of Bolbec

Ox 25,1

1

Roger de Lacy

45,1-2

1

Robert d’Oilly

1

3,1

2

The king Bishop of Salisbury Robert d’Oilly

5

Roger de Lacy

45,1-2

7

Ralph Piercehedge

-

1

Rainbald the Priest

61,2

? ? ? ? Section B5 Aldermaston ? [?Rycote] (O) [Enborne] and/or [Childrey]

18 12 19 20

? Section B6 [Sonning] [Enborne] and/or [Childrey]

21 19 20

? [Aston]

21

7

18,1-2

22

1

[Hendred]

22

1

Abbot of St Albans

Ewelme (O) Pyrton (O) Caversham (O) Watlington (O) Waterperry (O)

25 26 27 28 29

? 1 3 2 1

Abbot R Earl Hugh Walter Giffard Robert d’Oilly Robert d’Oilly Ilbert of Lacy, Roger son of Seifrid, Ordgar

Crowmarsh Giffard (O) ?

3 30

3

Hugh de Boldec

1

Hugh the large of Scoca

88

2 in Berks; 5 in Ox none none none none

Notes & Comments

Formerly Godric the Sheriff’s land – Roffe 2009, 43 Former bishopric of Dorchester before c.1072. See discussion. See discussion below.

17 in Brks None on Brks; 1 in Ox 2 in Brks; none in Ox Many

See discussion below. Many

Abbey of St 12,1 Albans ? 1 Brictric ? 1 Leofeva ? 1 Godwin ? 1 Alwin Section B9 “The undermentioned Oxfordshire thanes had land in WALLINGFORD” Newington (O) 10 4 Archbp Lanfranc Ox 2,1 Dorchester (O) 23 1 Bp of Lincoln Ox 6.1 [Hendred]

?

Number of manors in Berks / Oxon

12,1

2 in Brks; none in Ox None in Brks or Ox 1 in Brks; none in Ox 1 in Brks; none in Ox none none none none

1 in Brks; None in OX See discussion below.

-

none Crowmarsh Giffard held from Walter Giffard. none

Oxfordshire, Buckinghamshire and Berkshire

Contributory estate and shire O = Oxfordshire [ ] = inferred

No of estate in Fig. **

Number of tenements in Wallingford

Shirburn (O)

31

3 (inc. S Weston)

South Weston (O)

17

Ewelme (O)

25

Sotwell

33

1 1 Church 8

Long Wittenham

1

8

North Moreton

34

5

Basildon

35

3

Chilton

36

6

Harwell

37

3

Thatcham

39

12

DB section

Drogo

Ox 28,9

Drogo

Ox 1,88

Robert d’Amentieres Wazo

1

Ewelme (O) 25 Entered under rural manors: Sutton Courtenay 7 Sonning 21

LORD / holder of tenements

Number of manors in Berks / Oxon

Notes & Comments Shirburn held by Robert d’Oilly.

Ox 1,89 Ox 1,90

The king Bp of Salisbury St Peter’s Abbey Winchester Walter Giffard William FitzCorbucion The king Walter son of Other Bishop of Winchester The king

1,37 3,1 10,2 These tenements are presumably different to the 1 acre appurtenant to W Hannay, above.

20,3 27,2 1,8 31,2 2,2

Borough not specified, but Wallingford inferred.

1,2

Relationships between manors and the borough in later evidence (Roffe 2009, 34) [Aston Rowant] 40 (O) [Clapcot] 41

Manor in S Oxon

[Hungerford]

Part of the Domesday manor of Eddeventone (Eddington).

42

Table 23. Customary and non-customary tenements in Reading, Berkshire

-

No of estate in Fig. ** -

-

-

Contributory estate

Reading Earley

1

Finchampstead

2

Number of tenements

LORD / holder of tenements

DB section

28

The king

29

Abbot of Battle

1

Henry of Ferrers

1

The king

1,21; 1,42

The king

1,19

89

1,42

Number of manors

Notes & Comments Customary tenements ? Non-customary tenements, part of gift of a separate manor by the king Held with ½ virgate Held in freehold by Almer, who transferred it to his manor (1,42) Royal manor paying dues in Reading

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

conjectures. Some, however, are more certainly indicative of earlier arrangements; cumulatively, they generate a detailed model of the territorial development of the central Thames area from the early ninth century onwards (see fig. 21). This model is based on the premises already established in previous chapters, particularly in relation to the West Midlands area, as providing the best explanation of the spatial patterning of these connections. These patterns of distribution are seen as indicative of what I have described above (chapter 6) as ‘territories of obligation’, where populations within spatially-determined areas of land were required to perform labour and other services at burghal or other royal centres to facilitate and maintain the ‘king’s works’, in the broadest sense of the term.

These manors were distributed over all of Oxfordshire (DB Ox 6,1-8), with none in Berkshire, with a group of his principal holdings forming the (later) hundred of Dorchester (VCH Oxon vii 1962, 1-4). However, one tenement held by the bishop in Wallingford is given as being appurtenant to Dorchester in section B9, amongst the list of Oxford manors with tenements in Wallingford. He also held another one in Wallingford in section B4 where the connection is not stated. The distinction between this connection and the inferred connections of his other manors to his 30 tenements in Oxford carries certain implications. It could be that the greater proximity of Dorchester to Wallingford than to Oxford – as with the archbishop’s manor of Newington meant that he preferred to have one tenement in the nearest centre, for whatever reason. However, the distribution of other manors in the southern part of Domesday Oxfordshire which are contributory to Wallingford (fig. 20) has suggested to David Roffe that this area once formed part of the burghal territory of Wallingford on the eastern side of the Thames (Roffe 2009, 42-5). Although Roffe does not include the hundred of Dorchester in this territory – probably because most of the hundred lies to the north of the river Thame - this connection of the tenement in Wallingford which is appurtenant to Dorchester suggests that the area of this hundred should also be included within the territory of Wallingford.

In the cases of both Oxford and Wallingford, a number of inferences can be made concerning the connections of rural manors to urban tenements beyond what is explicitly stated in Domesday Book. One of the clearest and most interesting is the case of the Archbishop of Canterbury. He held only one manor in Oxfordshire - at Newington [Ox 10] in the hundred of Benson and therefore within the putative burghal territory of Wallingford - and none in Berkshire. The connection of this manor as being contributory to four tenements of his in Wallingford is stated in the description of Wallingford (DB Bks B9) amongst the list of other manors from Oxfordshire which are also contributory to it. The proximity of Newington to Wallingford could be taken as implying, under the prevailing paradigm of interpretation, that this was merely the nearest town to which he would have found it convenient to have had a town house. However, he also held seven tenements in Oxford which are unassigned (DB Ox B8), as well as another six tenements in Wallingford which were also unassigned (DB Bks B3). One possibility would be to infer that the archbishop’s single manor of Newington was contributory to his tenements in both Oxford and Wallingford. This would constitute one of two instances in which a single manor is contributory to two centres, the other being Pyrton (below). Another possibility, however, would be to suggest that the archbishop’s seven unassigned tenements in Oxford could have been appurtenant to one, two or all three of his manors in the central part of Buckinghamshire, at Monk’s Risborough, Haddenham and Halton (marked as MR, Hd and Ha in fig. 20). These are located in an area which it is argued below could well have been included within a primary pre-burghal (late-eighth to early-ninth century) territory or regio centred on Oxford (which is discussed below in chapter 13). This conclusion is perhaps strengthened by the fact that Monks Risborough is adjacent to Princes Risborough, an ancient royal manor, which also held an appurtenant tenement in Oxford (DB Ox 1,3; DB Buk 1,3). This will be discussed further below.

If this is so, it is probable that the hundred of Thame, formed by an amalgamation of several of the bishop’s estates, which lies to the south-east of the river Thame (VCH Oxon vii 1962, 113-6 & map), was also included within the burghal territory of Wallingford rather than Oxford, and that the unattached tenement held by the bishop in Wallingford in section B4 was appurtenant to the head manor of Thame. This connection is included in fig. 20. This would therefore represent a connection which can be taken back to the early stages of the foundation of the burh. As with other cases – such as the connections of Ramsbury with Cricklade in Wiltshire, and Bishops Cannings with Calne (see chapter 2 above) – this connection has survived by virtue of the fact that the manor remained in ecclesiastical hands for the nearly two centuries up to the time of Domesday. A similar inference could reasonably be made concerning the connection of some or all of the bishop’s 30 tenements in Oxford with his manors in north Oxfordshire which made up the hundred of Banbury, the principle ones of which were Banbury [41] and Cropredy [42] (VCH Oxon x 1972, 1-4 & map). As well as these, the bishop held six estates in Buckinghamshire (DB Bucks 3a,1-6), and although it is possible that one or more of these were contributory to Oxford (as was Twyford), the probability remains that his 30 houses in Oxford are most likely to have been appurtenant to his north Oxfordshire manors, as well as to Great Milton (DB Ox 6,3). His other holdings at Eynsham, Shifford and Little Rollright (DB Ox 6,6-8) were evidently endowments to Eynsham monastery, being tenanted by the monk Columban, its first abbot. These manors may well have been contributory to the bishop’s 30 tenements, though equally probably would have been listed as being held by

The distribution of the tenements of Bishop Remegius of Lincoln in both Oxford and Wallingford is also of some significance. He held eight manors, in all probability since the time of the foundation of the see of Dorchester, from which it was transferred to Lincoln between 1072 and 1086.

90

Oxfordshire, Buckinghamshire and Berkshire

Columban. Links to Banbury [41], Cropredy [42] and Great Milton are therefore shown in fig. 20, but links to the latter three manors are not. The presence or absence of these connections, while interesting in itself, does not, however, affect any overall conclusions which can be drawn from their distribution in Oxfordshire.

with Long Wittenham in north Berkshire and Caversham in south Oxfordshire are consistent with the premise that they were originally located within the burghal territory of Wallingford, and therefore help to define its extent. There are a number of other instances where reasonable inferences can be made concerning connections of rural manors with urban tenements in Wallingford. In section B6 the Bishop of Salisbury is stated as holding seven tenements, though the contributing manor is not specified. However, under his manor of Sonning [21] it is stated that ‘Roger the priest holds 1 church in Wallingford which rightfully belongs to this manor’ (DB Brk 3,1). Since it may be reasonably inferred that the church and the 7 tenements formed a single haga – as in the cases of St Ebbe’s and St Martin’s in Oxford (above), and St Lucian’s church and eight associated tenements in Wallingford (Roffe 2009, 38) – the bishop’s tenements can be reasonably inferred to have been appurtenant to Sonning as well. One other instance of such an inference is the possible connection between Roger de Lacy’s two manors in Berkshire (Enborne [19] and Childrey [20] (DB Brk 45,1-2) with his one tenement mentioned in section B5 and/or his 5 tenements in section B6 in Wallingford. Similarly, both Rainbald the Priest and the abbey of St Albans each held one manor in Berkshire and none in Oxford, which can therefore be connected with their single tenements in Wallingford (respectively Aston [21] and Hendred [22]). A similar inference can also be made in relation to the holding of one tenement in Wallingford by Gilbert of Ghent (DB Brk B3). He held one manor in Berkshire at Langley in Reading Hundred [11], as well as two in Oxfordshire, at Hanborough in Wootton Hundred to the north of Oxford, and another 8-hide manor at Ewelme in S Oxfordshire [25]. It is possible therefore that this single tenement was appurtenant to Ewelme rather than Langley. However, since there are other stated connections of Ewelme in Wallingford with the tenements of various subtenants (section B9), this inference does not extend the number or the spatial distribution of connections of Wallingford with south Oxfordshire.

A number of further inferences can be made about the connections of tenements in Oxford to the manors of particular owners beyond those which are stated in Domesday Book, by virtue of the fact that a connection can often be reasonably inferred between tenants-in-chief with few manors and their holdings within the borough. The clearest examples are those of the connection of the Count of Mortain’s single manor of Horley [Ox 16] with several or all of his 10 tenements in Oxford; the connection of the single manor of Milton-under-Wychwood [Ox 32] held by Ranulf Flambard with his single tenement in the borough; and the single manor of Wroxton [Ox 28] held Guy de Raimbeaucourt with his two tenements. Many of these instances have been brought out in the detailed analyses of John Blair (references in the table). Although cases such as these are inferences rather than being directly documented - and are indicated as such in the table - these are shown as connections in the drawn figure (fig. 20). Other examples of the importance of such inferences in other shires have been usefully discussed and tabulated by A Ballard (Ballard 1904, 20-31). The distribution of the manors of Walter Giffard are of interest in demonstrating both the potential and the limitations of the inferences which can be drawn about the relationship between the urban tenements, their contributing manors and the extent of the early burghal territories. Walter held nine manors in Oxfordshire and two in Berkshire. He also held 17 tenements in Oxford (DB Ox B9) which were unassigned to any rural manor, as well as 10 tenements in 1 acre in Wallingford (all therefore arguably appurtenant to one manor), three other tenements in Wallingford appurtenant to Caversham in south Oxfordshire (DB Brk B9) [27], and one further tenement in Wallingford appurtenant to Long Wittenham in north Berkshire (DB Brk 20,3) [8]. As well as his manor in Long Wittenham he also held one other manor in Berkshire at West Hanney in Wantage Hundred (DB Brk 20,1-2) [Ox 17]. West Hanney could on the face of it have been contributory to his ten tenements in one acre in Wallingford, but since it lies within what can be reasonably interpreted (below) to have been the putative burghal territory of Oxford, it would more probably have been contributory to one of his 17 tenements in Oxford. This is shown in fig. 20. The ten tenements in their acre in Wallingford could equally well have been appurtenant to any one of Walter’s four other manors in south Oxfordshire apart from Caversham, already stated as being contributory to a tenement in Wallingford. The possibilities are intriguing, though the lack of certainty means that the underlying hypothesis itself is not directly validated. What can be said, however, is that the stated connections of tenements in Wallingford

A particularly instructive instance is the holding by Earl Hugh of Chester of one tenement in Wallingford (DB Brk B4), with no contributory manor given, and another tenement amongst those Oxfordshire thegns holding tenements in Wallingford (section B9), where the contributing manor is given as Pyrton [26]. He held one other manor in south Oxfordshire at South Weston [17], so it might seem reasonable to infer that the unassigned tenement in Wallingford in section B4 was appurtenant to this manor. Pyrton was, however, also connected with Oxford as well as Wallingford. John Blair has argued that Earl Hugh’s seven tenements in Oxford (DB Brk B9) were appurtenant to Pyrton as the head manor (Blair 1994, 157, & n. 55), and David Sturdy and Julian Munby have shown that several properties in Cornmarket Street, in the centre of Oxford, were appurtenant to this manor (Sturdy & Munby 1985, 50; cf Dodd 2003, 30). This is, therefore, the only well-evidenced example in all three shires of

91

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Buckingham, Buckinghamshire

Oxon, Berks and Bucks where one manor contributed to two centres, a situation which is shown in relation to Droitwich and Worcester, and Winchcombe and Gloucester, already discussed in chapters 6 and 7. As will be shown below, this fits into the later pattern of the development of burghal territories and shires in the same way as these other examples.

While several tenants-in-chief and under-tenants are recorded in the section on Buckingham as holding burgesses, none of their associated manors is recorded. Nor is there any mention under the entries of the manors themselves of appurtenant burgesses or houses. However, as Ballard has pointed out, in several cases ‘the same person succeeded to both the urban and the rural properties of the evicted Saxon’ (Ballard 1904, 24-5), allowing reasonable inferences to be made concerning these connections of the holders of most of the manors with the holders of the burgesses, mostly through a process of elimination. The basis on which the inference is made in each case is given here under each entry. These inferences are also made by Arnold Baines, who gives more background historical information (Baines 1985, 56-61), though in all other respects his interpretation of the origins of Buckingham itself must be reassessed in the light of other considerations brought out in this study.

Earl Hugh also held two other manors in north Berkshire, at Buscot in Wyfold Hundred [Ox 19] and Drayton in Sutton Hundred [Ox 20] (DB Brk 18,1-2), which are argued below as lying within the putative burghal territory of Oxford in the Vale of the White Horse. It is possible, therefore, that one or several of his seven tenements in Oxford were appurtenant to either or both of these manors. These connections are included in the distribution in fig. 20, though are more inferential than most. This is a particularly apt example where connections can be postulated as being consistent with the particular hypothesis of the disposition of the burghal territories of Oxford and Wallingford, without constituting corroborative evidence for the hypothesis itself, which is based on wider considerations. A more certain inference is the holding of Hugh of Bolbec of one tenement in Wallingford (DB Brk B3), and none in Oxford. Hugh held no manors in Berkshire in his own right, and only one in Oxfordshire, at Rycote in Thame Hundred [12] (Db Ox 25,1). This lay within the putative burghal territory of Wallingford, and perhaps therefore provides some supporting evidence for the extent of this territory, without the force of an unequivocal demonstration.

Taken as a group, these connections are distributed around all sides of Buckingham, and can therefore be taken as being indicative – as in other cases already discussed, and for the same reasons – of the minimal extent of the original burghal territory of the burh of the Burghal Hidage. This pattern is discussed further below.

92

Oxfordshire, Buckinghamshire and Berkshire

Table 24. Customary and non-customary tenements in Buckingham, Buckinghamshire Contributory estate (All are inferred to some extent)

No of estate in Fig. **

Number of tenements in Buckingham

LORD / holder of tenements

26

DB section

Number of manors in Bucks

Notes & Comments

?The king B1 Customary burgesses Bishop of B3 Stewkley 1 3 Coutances The Bishop of Coutances has 3 burgesses, whom Wulfward son of Edeva held. Of the bishop’s 21 manors in Bucks, Stewkley is the only one formerly held by Wulfward. (5,5) (Baines 1985, 56). Shenley 2 1 Earl Hugh B4 Earl Hugh has 1 burgess who was Burghard of Shenley’s man. Of the earl’s 3 manors, Shenley (Church End) is the only one to have been held by Burghard of Shenley (13,2-3). Shalstone 3 B5. or 1 Robert d’Oilly 19,6 7 Water Stratford 4 19,7 Robert d’Oilly has 1 burgess who was Azur son of Toti’s man. Of Robert’s 7 manors, only two were held by Azur of Toti Shalstone (19,6) and Water Stratford (19,7). 5 Nashway, Radclive, 6 4 Roger d’Ivry B6 7 Thornton 7 Roger of Ivry has 4 burgesses who were also Azur [son of Toti]’s men. Of Roger’s 7 manors, these three had been held by Azur son of Toti – Nashway (41,1), Redclive (41,4) and Thornton (41,6). ?Linford or ?Whitchurch Bourton 16 4 Hugh of Bolbec B7 Hillesden 17 Beachampton 18 Akeley 19 Hugh of Bolbec has 4 burgesses who were Alric’s men. Alric is not mentioned as having held any of Hugh’s 11 manors, though Linford had been held by three un-named thanes (26,9), and Hugh held Whitchurch of Walter Giffard which had been held by two thanes, again un-named (14,16). However, Baines (1985, 56) identifies these manors as Bourton, Hillesden, Beachampton and Akeley, held by Alric’s successor Walter Giffard, of whom Hugh of Bolbec was undertenant. ?Padbury 20 4 Mainou B8 Mainou the Breton has 4 burgesses who were men of Edeva, wife of Sired. None of Mainou’s 11 manors (43) appears to have been held before 1066 by Edeva. However, Baines (1985, 56-8) suggests Padbury, where Mainou’s antecessor is not named. Quainton 8 1 Hascoit Musard B9. 49,1 Hascoit Musard has 1 burgess who was Azur son of Toti’s man. HM held Quainton (49,1), his only holding in the shire, and this had been held by Azur son of Toti. Lenborough 9 or 1 Arnulf of Hesdin B10; Barton Hartshorne 10 Arnulf of Hesdin has 1 burgess who was Wiglaf of Barton’s. Arnulf of Hesdin held two manors of the Bishop of Bayeux which had been held by Wiglaf – Lenborough (4,34) and Barton [Hartshorn] (4,37). William of B11 none ?Leckhampstead 21 2 Castellion William of Castellion has 2 burgesses, of the Holding of the Bishop of Bayeux, who were Earl Leofwin’s men. William is not recorded as holding or subletting any manors. 18 of the 43 bishop’s manors had been held by Earl Leofwine. Baines (1985, 58) suggests the 2 burgesses would have belonged to Leckhampstead. Biddlesden 11 1 Earl Aubrey B12 1 From Earl Aubrey’s Holding, 1 burgess pays 2d to the King. Earl Aubrey had held Biddlesden (1,7) (held in 1086 by the king), the only holding of his in Bucks. Some or all of: Salden 12 Leofwine of B13 Mursley 13 5 Nuneham Maids Moreton 14 Beachampton 15 Leofwin of Nuneham has 5 burgesses, and had them before 1066. Leofwine held 5 manors and no others (57,1-5). The fifth – Waveden (W in fig. 20) – can probably be discounted on topographical grounds, and because it only comprised one virgate. See further discussion below.

93

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 19. Oxfordshire, Berkshire and Buckinghamshire: relief and rivers with Domesday shire boundaries. Land over 400 ft (120 m) above OD stippled. North to top.

94

Oxfordshire, Buckinghamshire and Berkshire

Fig. 20. The orbits of connections of manors contributory to Oxford, Wallingford, Buckingham and Reading, both stated and inferred. Numbers refer to table 21 (Oxford), table 22 (Wallingford), table 23 (Reading), and table 24 (Buckingham). North to top.

95

Chapter 11 The Burghal Territories of Oxford, Wallingford, Sashes and Buckingham. royal estate centre, and which involved every estate or landholder above a certain status. This premise has been established in relation to other centres already discussed in preceding chapters. It provides an interpretive framework for these patterns which in all cases is consistent with the evidence as it can be reconstructed, fragmentary and unsatisfactory though it is. It is the internal coherence of this framework of interpretation which in turn arguably provides a validation for the basic hypothesis of the origin and function of these connections underlying this study. As already discussed, this hypothesis posits that all holders of manors above a certain status within defined territories were given tenements in the new burhs by the king, which thereby formed a holding of the manor rather than of the centre, to facilitate the organisation of the burhs as centres of defence and/or of regional administration at the time of the origin of the burhs as new institutions, and of other royal centres at probably an even earlier time. This was effected through the establishment and reinforcement of lordship bonds which arose out of obligations to the king derived from the holding of land. The spatial distributions of these connections should therefore be interpreted as helping to define what might be termed ‘territories of obligation’, which included the services due from the holders of land in these territories to fulfill the three common military obligations, which are documented from the mid eighth century in Mercia (Brooks 1971; Abels 1988, 61-2. 75-8, 91-3). This aspect has been examined in chapter 9 above, and more closely in chapter 14 below.

In the previous chapter a number of points have been made concerning the significance of the distribution of the manors showing connections with their various centres as evidence from which inferences can be made about the disposition of the burghal territories of the primary late ninth-century burhs of Oxford, Buckingham, Sashes and Wallingford. All of these were included in the Burghal Hidage document (Hill and Rumble 1996), and can be seen as providing for the overall strategic control of the upper Thames region and its defence against further Viking incursions. The basic premise that these distribution patterns constitute primary evidence from which these territories can be inferred has already been established in relation to the burghal territories in Wiltshire and Hampshire, as well as the west Midlands, discussed in previous chapters. It has also been established that some elements of the distribution of contributing manors to their centres are a function of the roles of particular places as centres of regional administration, and of royal control of economic activity, which can be taken back to a time before the creation of burhs and their territories from the late 870s in Wessex and the early 880s and the early tenth century in the west Midlands. The evidence from which these burghal territories can be reconstructed is examined in this chapter. The distribution of the manors showing connections (whether inferred or stated) with the centres in Oxfordshire, Berkshire and Buckinghamshire are particularly complex, and at first sight show a somewhat confused pattern. These patterns of distribution also include connections which are not directly stated, but based on inferences which are very much a matter of judgment. In all cases, however, these different levels of inference have been distinguished and characterised. The more certain connections have been used to establish the basic model of the distribution of the primary burghal territories, while the less certain inferences are merely seen to be consistent (or in some cases not consistent) with the model, without giving it direct validation.

As already pointed out, a basic hypothesis that the burghal territories of the two burhs of Wallingford and Sashes, both included in the Burghal Hidage, comprised areas which straddled the Thames in the three Domesday shires of Oxfordshire, Buckinghamshire and Berkshire has been articulated by Arnold Baines, Keith Bailey and David Roffe (Baines 1984; Bailey 1994, 89; Roffe 2009, 39-45). Baines’s and Bailey’s discussion, however, dealt only with Sashes, and Roffe’s with only Sashes and Wallingford. Their reconstructions are seen as providing an explanation for the disparity between the hidage allocations given in the Burghal Hidage and the hidage totals for the three shires, as well as for the presence of a number of manors within the area of southern Oxfordshire which were contributory to tenements in Wallingford. Their recognition that the

Nevertheless, the underlying premise is that the urbanrural connections as shown in Domesday Book are the attenuated remains of a more complete pattern of distribution which reflected connections made at the time of the initial foundation of a burh, or of an earlier

96

The Burghal Territories

burghal territories of these burhs straddled the Thames, rather than being defined by it, as with the Domesday shires, establishes a basis for a new way of seeing these connections in terms of their function at an early stage in the territorial development of the area, which is not constrained by the apparent fixity of the shire boundaries as they had clearly been established at the time of Domesday. This reinterpretation of the development of the Domesday shires from earlier arrangements of burghal territories, and even earlier regiones, is paralleled by the historical evolution of similar territories in the west Midlands, which is presented graphically in chapter 8 above. The evidence for this reconstruction, though alluded to in several instances above, is set out in what follows.

3.

One of the important outcomes of this interpretive framework is that these developments can give a new perspective on the relationship between political power, the control of territory and the deployment of military strategies in the late ninth century. In particular, they give some insight into the role of the king as originator of the system of burhs in Wessex and elsewhere, and help to define both the political and strategic circumstances which led to the development of the burhs of the Burghal Hidage, in both their geographical and functional aspects. They also throw some illumination on a particular phase of the development of relations between King Alfred and the West Saxons with both the Mercians and the Vikings, at what can be recognised as a crucial stage of the beginnings of the development of the West Saxon hegemony over the rest of the former kingdom of Mercia in the following decades. This is discussed in more detail in the next chapter.

4.

The development of burghal territories and shires It is argued here that the development of the burghal territories and the shires of this region of the central Thames valley occurred in several main stages from the ninth to the early eleventh centuries. These stages are exactly parallel to those already set out in relation to the west Midlands in chapter 8 above. The evidence for each of these stages will be discussed more fully in turn, but can be summarised as follows:

5.

1. Late eighth to late ninth centuries. The development from earlier ‘tribal’ units of a large ‘proto-shire’, regio or provincia, of probably Mercian origin, covering much or most of later Oxfordshire, Berkshire and Buckinghamshire, as a dependent ‘territory of obligation’ with Oxford as its centre. 2. Late ninth century. The organisation of the primary territories around the four Burghal Hidage burhs of Oxford, Wallingford, Sashes and Buckingham, each of which would have been given a territory which interlocked with its neighbours, and which can therefore be seen as being formed at the same time. The territories of the first three spanned the Thames. That of Oxford included the area of north-west Berkshire to the south of the Thames in the area of the Vale of

6.

the White Horse; that of Wallingford included the area of what became south-east Oxfordshire between the rivers Thame and the Thames; that of Sashes included southern Buckinghamshire and eastern Berkshire; and that of Buckingham occupied most (but not all) of the northern part of Buckinghamshire, as well as part of later Northamptonshire to the south-west of Watling Street. It is argued further below, and in chapter 12, that these burghal territories were formed and laid out as part of a general administrative arrangement by which the system of burhs listed in the Burghal Hidage was brought into being by King Alfred in the period immediately after his victory over the Viking army at the battle of Edington in early 878. Late ninth century – probably the early 890s. A phase of the secondary consolidation of the initial arrangement which is recorded in the Burghal Hidage. As I have already argued elsewhere, this phase can be best be placed within the context of the new wave of assaults against the Anglo-Saxon kingdoms by the Vikings from c.891(Haslam 2009, 98-100, 103-4; Haslam 2011b, 211). The replacement of the non-urban Burghal Hidage fortress of Sashes can be argued as being replaced by two new ones – the first at Aylesbury in central Buckinghamshire to the north of the Thames, and the second by Reading in eastern Berkshire, and the division of the territory dependent on Sashes between them. It is very probably in this phase that the earth and turf defences of both Oxford and Wallingford were consolidated and strengthened with a stone wall. Early tenth century. The creation of new burhs at the start of Edward the Elder’s campaign of burh-building in central and eastern Mercia, the first at Oxford itself (in c.911), the second at Newport Pagnell, extending West Saxon control into the territories belonging to Bedford and Northampton (c.914), the third at Wigingamere, on the eastern border of Buckinghamshire (917), and a fourth at Towcester (917). These developments can be seen as part of a more widespread policy of the containment of Viking interests and influence in Northamptonshire and beyond, which is represented in western Mercia by the creation of new burhs at Warwick and Tamworth in 913 (examined in chapter 8 above). Third quarter of the tenth century – probably the 960s. It is to this time that the ‘shiring’ of the central Thames area can be most reasonably be assigned, a process which led to the formation of the direct precursors of the Domesday shires. Early eleventh century. Minor adjustments were made to the mid tenth-century pattern, perhaps associated with the new military preparedness and the development of the Honour of Wallingford.

Stage 1 The development of a Mercian ‘proto-shire’ covering the area of the three Domesday shires (stage 1) must be considered later (chapter 13), since the hypothesis of its existence depends in part on the interpretation of anomalies

97

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

in the patterns of distribution of contributing manors to Oxford which can only be brought out by reference to the more general distribution of the relationship all the manors to all the centres, which in most cases can be taken as indicative of the second phase.

therefore a greater degree of uncertainty) than perhaps either Bailey or Roffe has been prepared to allow. The basis for the interpretation given here is, however, primarily that of the evidence of the patterns of distribution of the manors contributory to each centre at the time of Domesday, an interpretation already set out in relation to Wiltshire and the west Midland shires discussed in previous chapters. This is then interpreted in the light of other topographical evidence, in particular in the ways these suggested territories fitted in with, or developed from or cut across, earlier territorial arrangements. This in turn is then tested against an analysis of the comparison of the hidages of these areas at the time of Domesday with those stated in the Burghal Hidage. Since each of these burghal territories has two neighbours (amongst the group of four), these hypotheses must in their turn be tested or modified by examining ways in which topographical details and hidage values for the territory for each burh fit in with the spatial and other constraints represented by the territories of the three other burhs in the region.

Stage 2 - the burghal territories of the late ninth century The basic hypothesis of Keith Bailey and David Roffe, already alluded to, is that the primary burghal territory of Wallingford included all of southern Oxfordshire, and that the burghal territory of Sashes comprised an area of eastern Berkshire and southern Buckinghamshire, both straddling the Thames which they ignored as a boundary (Bailey 1994, 89; Roffe 2009, 39-45 & fig. 5.11). This was based in part on the need to find some correspondence between the hidages attached to each burh in the Burghal Hidage and the distribution of hidages in the Domesday hundreds of the three shires. It is also an inference from the premise that the distribution of the manors in southern Oxfordshire contributing to tenements in Wallingford in Domesday Book gives some indication of the original territory attached to Wallingford in the late ninth century. In terms of the evidential force of each of these premises, however, the latter was assumed by Roffe rather than argued from independent evidence, though this has of course formed the basis for much of the discussion in this study.

The burghal territory of Oxford Roffe’s analysis of the extent of the burghal territory of Wallingford includes within it all of north Berkshire, comprising the geographical area north of the Berkshire Downs to the river Thames to the north (known as the Vale of the White Horse), though he did suggest the possibility that this could have formed part of the primary burghal territory of Oxford (Roffe 2009, 44). There are, indeed, definite grounds for including part of this area within the latter. The Vale forms a distinct and distinctive geographical region which in several different ways geomorphological as well as in its settlement and land-use

The exact correspondence of the Domesday Hundreds with the original ‘proto-hundreds’ of the late ninth century, which themselves may well have formed the basis for the assessments in the Burghal Hidage, is less certain, and requires rather more flexibility of interpretation (and

Table 25. Key to the hundreds of Oxfordshire, Berkshire, Buckinghamshire and south-western Northamptonshire Oxfordshire

Berkshire

Buckinghamshire

Bb - Banbury Bi - Binfield Bp - Bampton Bs - Benson Bx - Bloxham D - Dorchester H - Headington K - Kirtlington La - Langtree Le - Lewknor Pr - Pyrton S - Shipton

Be - Beynhurst Br - Bray Bu - Bucklebury Bw - Blewbury Ch - Charlton Co - Compton G - Ganfield E - Eagle Hi - Hillslow Ho – Hormer K - Kintbury L - Lambourne M - Marcham Re - Reading Ri - Ripplesmere Ro - Rowbury Sf - Slotisford Sh - Shrivenham Su - Sutton T - Thatcham Wa - Wantage Wy – Wyfold

As - Ashenden Ay - Aylesbury Bh – Burnham Bs – Bunsty C - Cottesloe D - Desborough Ix – Ixhill L – Lamua Mo – Mousloe Mu - Mursley Ri – Risborough Ro – Rowley Se – Seckloe Sf - Stotfold Sk - Stoke Sn – Stone W – Waddesden Y - Yardley

T - Thame W - Wootton

98

Northamptonshire (south-west of Watling Street) Ab - Alboldstow Aw - Alwardsley Cy - Cleyley F - Foxley Gr - Gravesend Su – Sutton Wa - Warden

The Burghal Territories

Fig. 21. Reconstruction of the primary burghal territories of Oxford (north-west), Wallingford (south-west), Buckingham (north-east) and Sashes (south-east), in relation to later shire and hundred boundaries. For symbols see fig. 20 and tables, for the hundreds see fig. 22 and table. North to top.

99

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 22. The boundaries of Domesday shires and names of hundreds in Oxfordshire, Berkshire, Buckinghamshire and south-west Northamptonshire,. Based on maps in the Phillimore editions of Domesday – Berkshire (Morgan 1979); Oxfordshire (Morris 1978a); Buckinghamshire (Morris 1978b). North to top.

100

The Burghal Territories

contributory to Wallingford, within its territory. The boundary here must follow a small stream which divides Sutton Courtney on the east from Steventon [Ox 36] on the west, which is contributory to Oxford (see map in Hooke 1988, 126 fig. 6.2). The analysis of the totals of the hidages within the putative burghal territory of Oxford, discussed below, would suggest that the boundary is most likely to have run westwards along the river Ock, placing the hundreds to its north within Oxford’s territory. The manor of West Hanney [Ox 38] just to the south of this line, which was held by Walter Giffard, could have been contributory to one of his tenements in either Wallingford or Oxford, and so is not indicative of the boundary between the two territories. The clearly anomalous instance of Streatley to the south of Wallingford [Ox 37], held by Geoffrey de Mandeville (DB Brk 38,6) and contributory to Oxford, which does not fit with this pattern, is discussed in the next chapter.

history - has more in common with the central claylands of Oxfordshire on the other side of the Thames (compare Campbell 1971, 282-3 and Jope & Terrett 1971, 235-6). Keith Bailey has suggested that the natural alignment of the chalk escarpment was followed by the boundaries of early sub-kingdoms formed under the Mercian hegemony from the seventh century (Bailey 1996, 133-5). From a topographical perspective, the southern boundary of the suggested burghal territory of Oxford which includes the Vale of the White Horse in Berkshire is aligned with the southern boundary of this territory to the east in southern Oxfordshire, as well as the boundary between the burghal territories of Buckinghamshire and Sashes in southern Buckinghamshire. The two latter are, however, arguably formed for the most part by the length of the river Thame, which flows in a generally south-westerly direction some way to the north of the chalk escarpment, rather than by the Ridgeway along its crest. The river Thame is also used in part as the boundary between the burghal territories of Wallingford and Oxford immediately to the east of the Thames. In a precisely similar way, the putative boundary between the burghal territories of Berkshire and Oxfordshire to the west of the Thames is argued below as following the river Ock, which flows in a generally easterly direction within the Vale of the White Horse to reach the Thames at Abingdon.

The hypothesis of the inclusion of at least the northern part of the Vale of the White Horse within the burghal territory of Oxford is further strengthened by the relationship of the manors of Abingdon Abbey with its holdings in both Oxford and Wallingford which were appurtenant to them. Abingdon is located within the Domesday shire of Berkshire, yet the abbey held a total of 21 tenements or groups of smaller tenements in Oxford (Bond 1979, 72 fig. 7 – see fig. 18a in chapter 9 above), but only one haga, comprising seven tenements on two acres, in Wallingford (see tables 21 and 22 above). However, the abbey held only seven manors in Oxfordshire, but 40 in Berkshire, most of which were situated in the area to the north of the river Ock (see figs. 23a & b) (Bond 1979, 61 fig. 1; Kelly 2001, cliiiclxv esp. maps 1 & 2). The Wallingford haga can anyway be reasonably inferred to have been linked with Lewknor, to the east of the Thames (see below). These observations give grounds for suggesting that a good proportion of the total holdings of the abbey in the two shires were located within the former burghal territory of Oxford rather than that of Wallingford, and that the tenements in Oxford were appurtenant to some or many of them. The fact that one of the abbey’s tenements within Oxford (though given to it at a comparatively late date in 1032 – see above) is documented as being appurtenant to Lyford [Ox 14] in N Berkshire strengthens this general inference. There is no reason to believe that the connection of this tenement in Oxford with Lyford was only established on the occasion of this gift. Amongst all the abbey’s manors, the other documented connection with Oxford is Tadmarton [15] in N Oxfordshire. Since Abingdon is equidistant from both Wallingford and Oxford, connected to both by the river Thames, a distribution of these connections which might have resulted from the operation of the commercial or other interests of the landholders would be expected to have been more-or-less equally divided between the two. These observations therefore support the conclusions drawn from similar patterns in other instances that these connections arose at an early stage in the formation of burhs and burghal territories.

There are independent grounds, therefore, for interpreting the patterns of distribution of manors contributing to Oxford within the Vale of the White Horse as indicating the extent of the primary burghal territory of Oxford. As shown in fig. 20, there are no connections in this area of manors with Wallingford, but a significant number of connections of manors with Oxford. Of the seven manors in this area which are stated or inferred as contributing to tenements in Oxford, two are certainly documented – Lyford [Ox 14] held by Abingdon Abbey, and Steventon [Ox 36], held by Robert d’Oilly of the king. As discussed above, the connection of Compton [Ox 11] with Oxford is a reasonably safe inference; either Buscot [Ox 19] or Drayton [Ox 20], held by Earl Hugh, can also be reasonably inferred, though less certain are the connections with both together (though the absence of either or both would not affect the conclusion derived from the distribution of the others); and West Hanney [Ox 38] is a less certain inference. The connection of the nine tenements contributed by Faringdon [Ox 39] to Oxford (DB Brk 1,34) is an inference from its position, as well as from the fact of its early tenurial connections with manors on the northern side of the Thames (Gelling 1973-6, ii, 366; Hooper 1988, 3, 11 - but see reservations in Blair 1994, 190 n.65). The establishment of the boundary between the putative burghal territories of Oxford and Wallingford is very much dependent on the weighing up of probabilities. A reconstruction which appears to fit the location of manors contributory to either centre would cut across the Domesday hundred of Sutton, to include the two manors of Sutton Courtney [Bk 7] and Long Wittenham [Bk 1], which are

101

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Fig. 23a. Abingdon Abbey’s Domesday estates in Oxfordshire and Berkshire. Reproduced by kind permission of James Bond.

102

The Burghal Territories

Fig 23b. Abingdon Abbey’s Central estates in Domesday Book. Reproduced by kind permission of James Bond.

103

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

This set of relationships provides an explanation for the otherwise somewhat puzzling reference in the section of the Berkshire Domesday to the fact that ‘The Abbot of Abingdon has 2 acres [in Wallingford], in which are 7 messuages at 4s; they belong to Oxford’ (DB Brks B2). David Roffe includes ‘Oxford’ amongst other rural manors (Roffe 2009, 39 fig. 5.9), but the notion that tenements in one borough can ‘belong to’ or be appurtenant to another borough is meaningless. It would appear most likely, therefore, that ‘Oxford’ here implies the shire, as suggested in Roffe’s commentary (ibid. 31) and by Susan Kelly (Kelly 2001, clxii). Apart from Culham, which may have been included in the Berkshire hundred of Hormer, held in its entirety by the abbey (Kelly 2001, xclxiii-xclxiv), the only manor held by the abbey which lay in the south of Oxfordshire to the south of the River Thame and east of the Thames, and therefore within the putative burghal territory of Wallingford (above), was Lewknor [Bk 44] (Kelly 2001, clxiii) (see also fig. 23a). It can be reasonably inferred, therefore, that the 2 acres with its seven tenements in Wallingford were appurtenant to this manor. This connection is included in fig. 20.

fact of their being held by a monastic house, rather than being subject to fission and dispersal in the land market before the Conquest. To the west and north of Oxfordshire, the distribution of the manors contributing to Oxford suggest that the boundaries of the Domesday shire largely represent those of the primary burghal territory. The inclusion of inferred connections, as well as those directly evidenced in Domesday Book (shown in Blair 1994, 118 fig. 69), considerably extends the incidence of manors, especially on the western borders of the territory which marched with the ancient boundary of the Hwicce to the west, and strengthens the inferences which have already be made above that this represented a well-established boundary which was already of some importance by the late ninth century. This is also the implication of the pattern of provincial groupings of estates noted by Glenn Foard (Foard 1985, 196 fig. 5). The northern boundary of Oxford also marches with the boundary between the dioceses of Lichfield and Dorchester, which may, like that of the Hwicce, be considered to have been set out at an early stage in the processes of territorial demarcation in the middle Saxon period.

The distribution of Abingdon Abbey’s tenements solely within the area of the primary burh of Oxford (see fig. 18a) makes possible another highly significant conclusion - that the connections of these tenements with these estates, many of them presumably coming into the abbey’s hands after its re-formation in the middle of the tenth century, had already been formed by the time of the construction of the secondary burh to its east in c.911 (Haslam 2010a). This is the same conclusion as has been arrived at by another route, above. Kelly has indeed inferred that the position of these tenements within the area of the primary burh ‘may be consistent with relatively early acquisition’ (Kelly 2001, clxiii). If these connections had been formed as part of a gradual and cumulative process up until the time of Domesday, as John Blair has suggested (Blair 1994, 160-1), it would be expected that at least some of the abbey’s estates would have had appurtenant tenements within the area of the secondary burh. Kelly has argued that many of the abbey’s estates came into its possession after its refoundation in the mid-tenth century as a result of exchange of some of its more distant possessions for local estates (Kelly 2001, clxiv-clxv).

It is of some support for the general model examined in this study that many of the manors in the southwestern part of the shire, including some of those in the western part of the suggested burghal territory of Oxford in the Vale of the White Horse to the south of the Thames, are nearer to Cricklade to the west than to Oxford. Similarly, a number of manors in the western part of the shire are nearer to Winchcombe than to Oxford, and several of those in the northern corner are nearer to Warwick to the north-west. All nine of the latter are, furthermore, nearer to Buckingham than to Oxford. This pattern of distribution is therefore not consistent with the prevailing paradigm that the connections of these manors to their centre would have been created as a result of their proximity to the most convenient market or administrative centre by the exercise of choice on the part of their holders. The north-eastern and eastern borders of Oxfordshire with Northamptonshire and Buckinghamshire appear to be artificial lines which cut across already established boundaries of early groupings, both civil and ecclesiastical. The relevant evidence has been explored in some detail by Glenn Foard and John Blair (Foard 1985; Blair 1994, 52; cf Morley Davies 1947-52, 238; Taylor-Moore forthcoming). Keith Bailey has given plausible reconstructions of middle Saxon regiones within the area north of the Thames which is concordant with the evidence of later boundaries (Bailey 1996), and which appear to have influenced their course (bearing in mind the dangers of circularity of argument here). The incidence of detached pasture and woodland rights, as well as ecclesiastical connections, between northern Oxfordshire (within the area of the hundred of Ploughley) and southern Northamptonshire in Domesday Book, has suggested to Foard the existence of an early territory or regio spanning the two (Foard 1985, 196-8

The concentration of its urban tenements within the primary burh of Oxford carries the implication that the estates which the abbey acquired in the Vale of the White Horse, at whatever period, already came with an appurtenant tenement in Oxford, and that these would not have been acquired as separate transactions. This conclusion forms an important strand in the evidential basis for the overall model advanced in this study – that the urban-rural connections were formed as part of the process of, and at the same time as, the formation of burhs as well as royal estate centres. These connections in Oxford, even though not directly associated in surviving documentation with particular estates (except in two cases), have been preserved by the

104

The Burghal Territories

Hidage values

& esp. fig. 5). Similarly, Blair has for instance pointed to the division of an early minster territory focused on King’s Sutton by the boundary between Oxfordshire and Northamptonshire (Blair 1994, 52; cf Foard 1985, 199200). Though the details are intriguing and suggestive, they cannot be fully discussed here. However, in the reconstruction of the burghal territories of Oxford and Buckingham analysed below, based on the likely distribution of the hides given in the Burghal Hidage (see fig. 20), it is suggested that the early territory based on King’s Sutton could well have been incorporated in the former, and that the territory of Buckingham lay to its east within later Northamptonshire. A discrete administrative entity, which Blair has suggested was a Middle Anglian territory or regio (Blair 1994, 51), and perhaps centred on King’s Sutton, appears to have preceded the formation of the burghal territories discussed here, and to have played some part in their later delineation.

The model formulated in this study – that the burghal territories of the four burhs of the Burghal Hidage occupied the combined areas of Buckinghamshire, Oxfordshire and Berkshire in such a way as to ignore the Thames as a boundary – can be refined by comparing the hidages recorded in the Burghal Hidage document with the total numbers of hides and/or the total numbers of the hundreds into which they were grouped, which are recorded in Domesday Book. Uniquely for the Midlands, these Burghal Hidage figures provide an independent guide or bench mark to the likely extent of the territorial units which preceded the formation and re-formation of the shires to the north of the Thames in the tenth and early eleventh centuries, which established the pattern seen in Domesday.

Whatever the case, however, it is clear that the very northern boundary of the shire, which forms a salient into later Northamptonshire, has been determined by the need to include within it the estates of the bishop of Lincoln (formerly of Dorchester) around Banbury and Cropredy, which were later to form the hundred of Banbury (VCH Oxon x 1972, 1-4 & map). This conclusion would be entirely consistent with the suggested origin of these estates as holdings acquired by the bishop in a Mercian (i.e. eighth or ninth century) context (ibid., 1). It is suggested below that this awkward shape has been formed by the later imposition of arbitrary shire boundaries on an earlier arrangement. Furthermore, since the bishop did not hold any tenements in either Warwick, Northampton or Buckingham, the nearest large pre-Conquest centres, it would seem a reasonable inference that at least some of his 30 tenements in Oxford were appurtenant to the constituent manors of this hundred (Banbury and Cropredy), and that this relationship is most likely to have been formed at the time of the formation of the burh of Oxford with its burghal territory, or possibly earlier.

Buckingham 1600 (version A only); Hill (1969, Table I) gives a value of 1000 in others, but this is an inference only; the 600 hides given by Gale, and noted by Hill, cannot be correct. Sashes 1000 (all versions) Wallingford 2400 (all versions) Oxford 1500 (version A); 1300 (version B).

An analysis of the hidage values of the area suggested as comprising the extent of the burghal territory of Oxford immediately exposes a disparity between the value given in the Burghal Hidage and the actual hides on the ground which need to be included within its territory, as it can be reconstructed using other evidence. The totals of the hides attached to the four burhs in the Burghal Hidage (Rumble 1996) are as follows:

The total number of hides within these four territories in the Burghal Hidage is therefore a maximum of 6500. This at once creates an imbalance with the recorded number of hides in Domesday for the three shires of Oxfordshire, Buckinghamshire and Berkshire of 7060 (Roffe 2009, 44). In other shires in Wessex (with the exception of Somerset) the Burghal Hidage totals exceed the Domesday hides by a factor of between 16 and 27% (Haslam forthcoming d; see figures in Brooks 1996b, 133-8). This discrepancy could therefore be explained by the suggestion that the figure for Oxford is more likely to have been 2400 hides rather than 1500.

Similar observations have been made relating to the boundary between Buckinghamshire and Oxfordshire as being ‘ragged and artificial’ (Blair 1994, 51), and as dividing an early minster territory around Brill. Foard suggests that these boundaries were not defined ‘until as late as the tenth century’ (Foard 1985, 198). As suggested below, this should, however, be seen in the context of the imposition of West Saxon hegemony over a territory which had had long-standing Mercian antecedents, but which was, from the evidence of the coinage of the time, already being drawn into the West Saxon orbit before the creation of the burghal system of the Burghal Hidage, perhaps from as early as the 860s. An alternative scenario could be that this boundary to the west and north of Oxfordshire was formed as a result of the creation of the suggested Mercian provincia of stage 1 (above). This is discussed in detail below (chapter 13).

This is indeed supported by other considerations. In the late ninth century there is no doubt that Oxford was of considerably greater importance than Buckingham, having an established presence as the focus of religious and trading activity within a wide region for a considerable time beforehand (factors which are discussed in detail below, chapters 12 and 13). Yet Buckingham is rated at 1600 hides (A Version only), while Oxford at only 1300 or 1500 (Hill 1969, table II; Hill 1996a, table 4.2, 78). It is clear that there is a general, though not an exact, relationship between the sizes of burhs and their hidages (Hill 1981, 85-6; but see caveats in Brooks 1996b, 129-32; Haslam 2009, 111-4),

105

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Table 26. Lengths of defences and areas of Buckingham, Oxford, Wallingford and Winchester Burh

Hidage in Burghal Hidage

Buckingham Oxford

1600 1300 / 1500

Wallingford

2400

Winchester

2400

Length of defensive circuit (m) 760 2000 (primary burh) 2660 (4 sides) 1860 (3 sides) 3034

which would cast some doubts on the veracity of the figures of either Oxford or Buckingham. The length of the defences of the primary burh of Oxford, and therefore the size of its defended area, was considerably greater than those of Buckingham. Recent archaeological work at Oxford, which demonstrate that the defences of the primary burh extended westwards to the edge of the gravel terrace under the site of the castle (Poore et al 2009, 8 fig. 4), would indicate that its original burghal defences were around 2000m in length and enclosed an area of approximately 26 hectares (measurements here and elsewhere determined in Edina Digimap). The defensive circuit of Buckingham (although nowhere confirmed by excavation, and estimated from the topography of the town by inspection and as shown on early maps) is around 760m in length, enclosing an area of approximately 3.3 hectares. There is some equivalence in size of the defences of Oxford to the size of defensive circuit of Wallingford, also assessed at 2400 hides (2660m including the river side, or 1860m excluding it, enclosing an area of 39.6 hectares). The circuits of both these places are however smaller than that of Winchester, also assessed at 2400 hides (3318 yards or 3034m, enclosing an area of 59.9 hectares) (Brooks 1996b, 129-30; cf further discussion in Haslam 2005, 145-6; Haslam 2009, 111-4). These figures are presented in the table 26 above.

Area enclosed (hectares)

Suggested original hidage

3.3 26

1600 2400

39.6

2400

59.9

2400

territories are also defined by natural features or by earlier boundaries. In relation to Buckinghamshire, for instance, Frank Thorn has observed that ‘The number of hundreds, their differences in size and their sometimes irregular shape suggests that Domesday does not describe the primordial arrangements, but records a state evolved from an earlier, simpler pattern’ (Thorn 1988, 38), a conclusion supported by Bradbury (Bradbury 1988, 20). Keith Bailey, however, describes how the hundreds were associated in groups of three (Bailey 1994). In Oxfordshire, the Domesday hundreds of 250 or 300 hides, which appear to have been the original arrangement, appear to have evolved directly from territories dependent upon early royal estate centres (Thorne 1990, 26-7; Blair 1994, 107-111). In Berkshire, the development of hundreds of a wide variety of different sizes (Baring 1909, 45) also appears to suggest that this pattern does not reflect arrangements put in place nearly two centuries earlier. It is therefore the sum of the hidages of individual manors which make up the hundreds which can be considered the most reliable guide to the earlier pattern. These assessments have been shown to have been remarkably stable for a long time (Blair 1994, 77-9; Bailey 1994, 87-8; Brooks 1996b, 130-2). This in turn has to be tempered by the premise that while the Burghal Hidage figures are likely to represent the original total of the hides available within each district on which military obligations were due, the Domesday hidage totals will in many cases omit some of these ‘primary’ or inland hides in royal or other demesne lands which were not subject to geld (Finn 1963, 258-61; Roffe 2007, 200-3). This observation has also been argued in detail in relation to the hidage of Wiltshire in chapters 2 and 3 above, and of Hampshire in chapter 4, and by Cyril Hart in relation to Northamptonshire (Hart 1970, 17-32), and is discussed further in chapter 9 above. This factor is possibly in some cases further exaggerated by the incidence of beneficial hidation before the time of Domesday, in which process the original hidages of geldable lands have for one reason or another been reduced. This factor is particularly noticeable in the case of Hampshire, discussed in chapter 4. As has been already been argued in chapters 2, 3 and 4, these factors have not been taken into account in the attempt by Brooks and others to account for the discrepancies between the Burghal Hidage figures and the hidage figures recorded in Domesday Book (Brooks 1996b, 134-8; Brooks 2003, 158-62; Lavelle 2007, 22; Baker and Brookes 2011, 105 & n.12).

It would be unreasonable, therefore, to accept the Burghal Hidage figure of 1500 hides for Oxford, which is smaller than that given to Buckingham, at face value. The suggested figure of 2400 hides for Oxford does in fact provide a workable solution which, as is demonstrated below, neatly fits the hidage values of the area which can be reasonably suggested as comprising Oxford’s territory, and which can therefore be tested by fitting it to the areas indicated by the hidages of other probable burghal areas. General support for this conclusion is, furthermore, provided by the fact that Oxfordshire is rated at 2400 hides in the County Hidage. While it seems likely that this and other totals in the document refer to actual hundreds (though with some exceptions), the development and changes of these hundreds on the ground in the two centuries between the date of the Burghal Hidage and Domesday Book makes the Domesday hundreds (insofar as they can be reconstructed) an unreliable guide to use on their own. Notwithstanding this, however, the boundaries of the Domesday hundreds have been used in reconstructing the burghal territories in the following analyses, since there is no other means of doing so. Most of the boundaries of the burghal

106

The Burghal Territories

The hidages of the hundreds in the northern and central parts of Oxfordshire, including Ploughley and the northern part of Headington (Bullingdon) hundred as suggested above, are as follows (figures from VCH Oxfordshire): Hundreds (fig. 11): Hides Banbury 100 Bampton 201 Bloxham (and Adderbury) 212 Shipton 283 Wootton 350 Ploughley 284 Bullingdon (Headington) 228 (less10 Waterperry) Sub-total - 1648.

hides

The best-fit solution to the northern extent of the burghal territory of Oxford would see it as being more-or-less coterminous with the northern boundaries of later Oxfordshire. However, the addition of the Northampton hundred of King’s Sutton to this territory would remove the peculiarity of the rather awkward northern extension of Oxfordshire into later Northamptonshire. This is largely comprised of the bishop of Lincoln’s manors, and looks very much like a remnant formed in a later phase of reorganisation. It is of some significance, therefore, that the very north-western boundary of Oxfordshire marches with the boundary between the dioceses of Lichfield and Dorchester (Hill 1981, 161 map 255), which, like that of Worcester which defined the kingdom of the Hwicce, must have belonged to an earlier strand of middle Saxon territorial division in the area. As will be argued below, the boundary of the territory dependent on Buckingham, arguably including all of south-west Northamptonshire, would also have marched with that of the north-east boundary of that of Oxford (see fig. 20). The close approximation of this territory (albeit in part reconstructed) to the round-figure assessment of 2400 hides suggests that the actual area of land on the ground which this hidage value represents was imposed from above as part of a substantial local government reorganisation in which West Saxon hegemony was firmly exerted and defined, perhaps for the first time. It also brings the hidage assessment of Oxford firmly into line with that of Wallingford, to which it was certainly equal in historical importance and strategic significance.

for

To this total must be added the sums of the hundreds and manors from the northern part of the Vale of the White Horse in Berkshire, as indicated in figs. 21 & 22 and discussed above (figures from Baring 1909, 45): Hormer 110 Marcham 141 Ganfield 185 Wyfold 144 Sutton (W part) 83 Sub-total - 631 The total for these areas is 2279 hides for the whole of the burghal territory, which is nearly identical with the 2400 hides postulated above. The hundred of Shrivenham, in which the manor of Compton [11] is inferred as being connected with Oxford, is omitted here on topographical grounds – it is not on the Thames, and therefore perhaps more likely to have been part of the territory belonging to Wallingford. It might be considered that the figure of 1648 hides for the area to the north of the Thames, which is near the total for Oxford in version A of the Burghal Hidage of 1500 hides, could be taken as indicating that the area in the Vale of the White Horse to the south of the Thames might not have been included in the original burghal territory of Oxford. However, the evidence in this area of both known connections of rural manors to Oxford, and the distribution of most of the holdings of Abingdon Abbey in this area (figs. 23a & b) and their likely correspondence with its holdings in Oxford, discussed above, are strong independent reasons for its inclusion within the territory of Oxford. Furthermore, if the sum of the hides in the Vale of the White horse is added to the total hidage for the burghal territory of Wallingford, which is examined below, it would inflate this to a level which would show no acceptable correspondence with the total stated in the Burghal Hidage. However, as is suggested below, it is likely that the Northampton hundred of King’s Sutton, with a pre-Conquest hidation of 100 hides, and which represents the northern half of an earlier regio based on King’s Sutton, could have been included in Oxford’s territory. This would make up a figure of 2379 hides for the area dependent upon Oxford, very near the postulated figure of 2400 hides.

The burghal territory of Wallingford The evidence for the extension of the burghal territory of Wallingford to the east of the Thames, pointed out by David Roffe, comprises the connections of several south Oxfordshire manors with Wallingford which are documented in Domesday Book. Roffe, however, limits this area to the extent of the five Chiltern Hundreds of Benson (or Ewelme), Pyrton, Lewknor, Langtree and Binfield (fig. 22), primarily on the basis of the need to match the sums of the hidages of the constituent hundreds of the postulated burghal territory as a whole with the Burghal Hidage figures (Roffe 2009, 42-5). As already suggested, the evidential force of these documented connections must be extended to include others in the adjacent hundreds of Dorchester and Thame to the north, both episcopal hundredal manors (VCH Oxon vii 1962, 1-4 & map, 1136; Thorn 1990, 28). As such, these hundreds are likely to have been artificial creations, made up of the bishop’s lands, subsequent to the formation of their surrounding hundreds. Dorchester [Bk 23] is documented as having had an appurtenant tenement in Wallingford, as did Albury [Bk 9] and Gt Haseley [Bk 4]. A further case is Waterperry [Bk 29] within Headington (Bullingdon) hundred on the north-western side of the river Thame, which elsewhere marked the boundary of Thame hundred to the south-east. The complexity of the interlocking hundred boundaries at this point, with several areas of Thame, Bullingdon and

107

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Ewelme hundreds surrounding each other on both sides of the river Thame (VCH Oxon vii 1962, 113-6 & map; Thorn 1990, 29; Blair 1994, 51 fig. 39 and 109 fig. 62), suggests that Waterperry could well have been included at one time within an original ‘proto-hundred’ which existed before the definition of the amalgamated Thame hundred as part of the bishop’s triple hundred in south Oxfordshire. The fact that Waterperry was not held by the bishop would have provided a reason for its exclusion from the later Thame hundred, which only comprised three detached estates belonging to the bishop. The connection of Rycote [Bk 12] in this area to Wallingford is based on reasonable inference, as is the connection of Thame itself [Bk 16] to Wallingford, for reasons given above.

the original territory of Wallingford would have marched with the southern boundary of that of Oxford within the Vale of the White Horse to the south of the Thames. It would appear significant that this boundary follows the river Ock, which drains the centre of the Vale eastwards. Similar relationship of the common boundary of Oxford and Wallingford’s territories to the east of the Thames is shown by its use of the river Thame as part of its boundary, which is also followed further to the east by that between Sashes and Buckingham. In all these areas it is noticeable that these common boundaries, as reconstructed from other evidence, follow neither the line of the Chiltern escarpment (followed by the Ridgeway) nor the Icknield Way at its foot. Roffe includes the hundred of Sonning to the south of the Thames within the burghal territory of Sashes, based on the hidage totals for the two parts of the latter, but a consideration of disposition of the hides supporting Sashes, discussed below, suggest that this area lay within the territory of Wallingford, and the rest of eastern Berkshire (the hundreds of Beynhurst, Bray and Ripplesworth) within the territory of Sashes. The connection of Wallingford with Bray [8], on the far eastern side of the shire, would in any interpretation appear to be anomalous. But this connection could have been put in place in a later phase of reorganisation of the shire in the mid tenth or the early eleventh century, a process possibly connected with the formation of the Honour of Wallingford. As discussed above, the only other area where there is room for interpretation is the north-eastern part of the Berkshire, lying to the north of the Berkshire Downs within the Vale of the White Horse, which is argued here lay within burghal territory of Oxford.

One apparent anomaly is that of Pyrton [Bk18], which is contributory to tenements in Oxford as well as in Wallingford (above). This represents the only documented example in all three shires of a manor which is contributory to two centres. Other instances of this are discussed in preceding chapters in relation to Wiltshire (chapters 2 and 3, as well as to Worcester and Droitwich (chapter 7), and Gloucester and Winchcombe (chapter 6), and are explained by the successive superimposition of the orbit of connections to two different centres. The instance of Pyrton is one strand of evidence which could be explained by the hypothesis that a large territory dependent on a Mercian centre at Oxford was divided between the four burghal territories, the orbit of connections of Wallingford being imposed on that of Oxford. This would provide the tenurial context for the connections of Buscot [Ox 19] and Drayton [Ox 39] with Oxford, though these are inferred rather than stated, and of Princes Risborough [Ox 4], just over the border in Buckinghamshire. These will be discussed further in chapter 13 below.

It might also be considered as significant that the carrying duties of the burgesses of Wallingford, spelled out in section 1 of the Berkshire section of Domesday Book, are to four places on or near the Thames (Hooper 1988, 25-6 & map 3). Three of these were royal manors within Berkshire (Blewbury, Reading and Sutton Courtnay), while the third was to Benson, an early royal and territorial centre on the eastern side of the Thames, in Oxfordshire but within the putative burghal territory of Wallingford. These carrying duties of the burgesses can be reasonably be seen as a survival of duties of service of the burgesses to the king which were arguably put in place at the time of the foundation of the burh. This is a further indication that its ‘territory of obligation’ extended to the eastern side of the Thames, discussed above, of which the connections between rural manors contributory to urban tenements was one aspect.

In the reconstruction argued above, the north-eastern boundary of the burghal territory of Wallingford with that of Oxford is formed in the main by the river Thame, though with Waterperry just to its north included in the former. Dorchester, however, lies to the north of the confluence of the Thame with the Thames, and most (but not all) of the hundred of Dorchester lies to the north and west of the river Thame (VCH Oxon vii 1962, 1-4 & map). It is significant that this boundary reflects the earlier boundaries between the middle Saxon regiones of the Hendrica (to the north-west) and the Cilternsaetan (to the south-east), as reconstructed by Keith Bailey (Bailey 1996, 134). The eastern boundary of the putative territory dependent on Wallingford abuts that of Sashes to its east. However, the presence of detached portions of Lewknor hundred within later Buckinghamshire (VCH Oxon viii 1964, 1-3; VCH Bucks iii 1925, 96, 99; Thorn 1988, 40), suggests that the Oxfordshire / Buckinghamshire boundary (which marks the boundary of the Wallingford and Sashes burghal territories) has perhaps been cut through an earlier arrangement in which the original regio of the early royal centre at Benson perhaps extended further to the east. As argued above, the northwestern boundary of

This reconstruction is supported by an analysis of the hidages of the area. All versions of the Burghal Hidage give a value of 2400 hides for the territory attached to Wallingford (Hill 1969, table 1; Hill 1996a, 75-6). The sums for the hidages in Berkshire which would have comprised part of the burghal territory of Wallingford are as follows (figures from Baring 1909, 45-6). These

108

The Burghal Territories

omit hidages from the eastern hundreds of Benhurst, Bray and Ripplesmere, which are assigned to Sashes, and the hundreds of Hormer, Marcham, Ganfield and Wyfold, and the western part of Sutton, which for reasons given above are assigned to the burghal territory of Oxford.

The assumption that its boundary with the territory of Wallingford to the west marched with the Buckinghamshire / Oxfordshire boundary would fit with the disposition of the hidages given to both. The line of the boundary to the north, where it marches with the putative burghal territory of Buckingham (below), is, however, again somewhat equivocal and open to interpretation. Keith Bailey has suggested that it lay along the crest of the Chiltern escarpment (Bailey 1994, 89), though this is not marked by the boundaries of later hundreds. This solution reflects his view that the northern boundary of the earlier regio of the Cilternsaetan followed this line as well (Bailey 1996, 131). David Roffe, on the other hand, has given reasons for suggesting that the territory extended northwards to the river Thame, to include the hundreds surrounding Aylesbury, as well as Aylesbury hundred itself (Roffe 2009, 44).

Hundreds (fig. 22): hides: Charlton 145 Reading 154 Hesletsford 117 Blewbury 127 Nachededorne 103 Bucklebury 47 Tacham 74 Roeberg 128 Kintbury 133 Eagle 96 Lambourne 73 Wantage 239 Hilleslau 140 Shrivenham 71 Sutton (E part) 38 Sub-total 1685

There are several reasons for following Roffe’s interpretation. In the first place, a boundary along the line of the Thame would continue the line of the boundary between the north-east of the burghal territory of Wallingford east of the Thames and that of Oxford to its north-west, for at least some part along the river Thame, which is argued above. Secondly, this line appears to respect that of the earlier district which was dependent on Aylesbury, an Iron-Age hillfort, an early and mid-Saxon territorial focus and an important middle Saxon minster church (EUS Aylesbury). Aylesbury was one of the tunas captured by the W Saxons in ‘571’, which, as Keith Bailey has pointed out, defines the northern limit of the Cilternsaetan (Bailey 1996, 134), This suggests that already in the seventh century it was the head place of a dependent territory. Bailey has even suggested that a minster at Aylesbury (as at Benson, another of the tunas captured by the West Saxons in ‘571’) was created by the West Saxons before being absorbed by the Mercians towards the end of the seventh century (Bailey 1991, 45; but see Blair 1994, 60-1). At the time of Domesday Aylesbury was supported by eight hundreds ‘in circuito’. An interpretation of this phrase has suggested that this referred to a circuit of eight hundreds used by the Domesday commissioners to record their findings, (comprising Aylesbury, Stone, Risborough, Burnham, Desborough, Ixhill, Ashendon, Waddesdon, and probably also Stoke) (Bradbury 1988, 14), though Frank Thorn regards this phrase as meaning merely the eight surrounding hundreds, which comprise a somewhat different grouping (Thorn 1988, 38). Although this phrase refers to the minster parochia of Aylesbury, rather than a civil territory - as J H Round pointed out some time ago (Round 1905, 223) both Bradbury and Thorn recognise that this grouping, in whatever way it can be constituted, represents a significant territory which comprised most or all of the southern half of the later shire, and which may well be of early origin. Comparisons may be made with the eight hundreds of Oundle, the seven hundreds around Cirencester, and others (Cam 1963; Thorn 1988, 38, n.5).

To be added to this sub-total are the following hundreds to the east of the Thames, now in the southern part of Oxfordshire (the so-called 4½ hundreds of Benson, plus the two hundreds of the bishop of Lincoln): Lewknor 131 Binfield 65 Langtee 114 Pyrton 103 Ewelme 85 Dorchester 100 Thame 100 Sub-total 698 This gives a total for the putative burghal territory of Wallingford as 2383 hides – very near the Burghal Hidage figure of 2400. It is probable that this total would have been made up from unrecorded inland hides at a several royal manors in which beneficial hidation is shown by the fact that the number of ploughlands appears to be rather greater that the number of recorded hides. These are Thatcham (1,2); Cookham (1,3); Waltham (1,4); Blewbury (1,5); Wantage (1,9); possibly Bucklebury (1,23); Kintbury (1,26); Lambourne (1,29); and Aldermaston (1,23). The burghal territory of Sashes As both Keith Bailey and David Roffe have already pointed out, the hidages given to Buckingham, Wallingford and Sashes in the Burghal Hidage requires that the burghal territory of Sashes would have comprised a large territory on the northern side of the Thames as well as on the southern side, occupying the southern half of the later shire of Buckinghamshire as well as the eastern part of later Berkshire. This is worked out in more detail below.

Keith Bailey sees this territory dependent upon Aylesbury has having changed somewhat over the centuries before the

109

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

time of Domesday, and suggests that it gained territory from other minsters such as Wing as a result of ‘the upheavals in the later ninth and tenth centuries associated with the Danish wars’ (Bailey 1991, 46). I would argue, however, that it is more in accordance with the evidence of the topography of the area to accept that the area dependent on Aylesbury has not in fact changed since its formation as a small regio in the middle Saxon period, and that this was perpetuated as its minster parochia which was then assigned to Sashes as part of its burghal territory in the later ninth century, with little alteration of long-established boundaries. As indicated by the Domesday evidence, its northern boundary is most likely to have been formed by the north-western boundaries of Aylesbury and Stone hundreds (fig. 22), with Aylesbury sited at its northern edge. This boundary formed a relatively straight line from south-west to north-east following the river Thame, and suggests that the burghal territory of Sashes occupied all the territory to its south and south-east to the river Thames (fig. 21). It therefore included part of the claylands to the north of the Chilterns, and thus extends beyond the area of the early regio of the Chilternsaetan identified by Bailey, which he suggests was bounded by the Chiltern scarp and the Icknield way which followed it (Bailey 1994, 93; Bailey 1996, 131-2). However, the absence of any evidence of the extent of the minster territory at any time, apart from the elusive reference in Domesday Book, makes it difficult to reach any certainty. It is of course possible, if not likely (as against one interpretation of Bailey), that this boundary had formed the northern boundary of the Chilternsaetan at an earlier period, and that there was a continuity between the two which was perpetuated into the later ninth century. As will be shown, this interpretation is consistent with, and indeed supported by, the evidence from the connections of manors contributory to Buckingham, discussed below.

a newly negotiated boundary to the east of London in late 879 as a result of the accord recorded in the Treaty between Alfred and Guthrum. I have argued that the boundary of this new political unit to the west of London, which the Vikings were able to establish as a result of their military and political ascendancy at this time, ran up the river Colne and northwards along the river Gade, crossing the Chilterns to meet the river Ouzel which it followed northwards until it met the line of Watling Street, which then formed the boundary further to the north (fig. 24). This alignment is supported not only by its topographical appropriateness, but also by various other considerations concerning its later history and development (Haslam 2011a, 125-7). Given the reality of this boundary at a crucial time in the strategic powerplay between King Alfred and the Vikings at this time, it is necessary to accept that the burghal territories of both Sashes and Buckingham north of the Thames would have comprised lands which lay to the west of this line, since they would have been formed after its imposition and before Alfred gained London and the area to its north in late 879. The northern part of this line, as it defined the eastern boundary of the burghal territory of Buckingham, is discussed below. An analysis of the hidages of the putative area dependent on Sashes might, however, suggest that the inclusion of this area of later Hertfordshire would inflate the hidage value given to Sashes well beyond its allotted 1000 hides. It is possible therefore that this boundary of c.877 to the west of London could have followed a more westerly line more closely approximating to that of the later shire boundary of Buckinghamshire and Hertfordshire. Though this follows no obvious topographical feature from south to north it would form an alignment leading from the River Colne, which marked the western alignment of the ancient border of Middlesex, directly northwards to the river Ouzel flowing into the Great Ouse to the north (fig. 24). This being so, the whole of the area of the hundred of Yardley, which forms a salient into later Hertfordshire, should perhaps be excluded from the areas of the Burghal territories of both Sashes and Buckingham. There is, however, no independent evidence for any of these alternatives, other than that this would more easily accommodate the reconstructions based on the hidages stated in the Burghal Hidage.

The eastern boundary of the territory dependent on the burh at Sashes is somewhat more hypothetical, and open to different but equally plausible reconstructions. There are some grounds for suggesting that this territory might well have included the western part of later Hertfordshire to the west of the river Gade, a small south-flowing tributary of the Colne to the south (fig. 24). As the western boundary of Middlesex, the Colne would have formed the southern part of the eastern boundary of the territory of Sashes. As such, the Hertfordshire hundreds of Tring, which forms a noticeable salient into Buckinghamshire on its eastern side, as well as a part of Dacorum, could well have been included within it. This conclusion is dependent upon an interpretation of the course of a Viking / Mercian boundary to the west of London, which I have examined elsewhere, which is suggested as having been set out as a result of a Viking occupation of London consequent upon the ‘sharing out’ of parts of Mercia by the Viking army in 877 (Haslam 1997, 118-21; Haslam 2011a, 125-7) (see fig. 25). This process is seen as a considerable territorial gain for the Vikings, by means of which they were able to gain control of London and its territory, which was defined on its western side by this boundary, until forced eastwards to

As argued above, the 1000 hides accorded to Sashes in the Burghal Hidage comprise areas on both sides of the Thames. To the south of the Thames, the following hundreds are included (pre-Conquest values for hides from Baring 1909; 45): Hundreds (fig. 22): Hides Bray & Benhurst 93 Riplesmere 60 To these figures are to added those of the following hundreds to the north, for reasons given above. These

110

The Burghal Territories

Fig. 24. Area of eastern Mercia, including London, in relation to Alfred and Guthrum’s Treaty boundary (c.880) and later shires. This map extends the range of figs. 19-22 to the east. North to top.

111

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

formed two triple hundreds (figures from Bradbury 1988, 20 table 2):

imposed upon an administrative landscape which had been arranged rather differently before the burghal territories were laid out.

Stoke 95 Burnham 104 Desborough 148 Stone 150 Aylesbury 151 Risborough 99 Sub-total 747

Buckingham itself lies on a prominent spur on the northern side of a bend in the river Great Ouse, and was a site of earlier settlement and a royal vill which possessed a minster church (Bailey 1996, 134; EUS Buckingham 2008, 15, 35-40). These factors appear to have determined the inclusion of a chunk of territory to its north, which occupied the hundred of Stotfold, and which at first sight appears as a salient which was pushed into the Viking territory of Northampton northwards from an earlier regional boundary which possibly followed the Great Ouse at this point (Foard 1985, 196 fig. 5; Bailey 1996, 134; TaylorMoore 2012). This is shown particularly clearly by Morley Davies (Morley Davies 1952, 239). As Taylor-Moore has shown, the northern boundary of this hundred follows the watershed between the Great Ouse and the Nene, which comprises the forest of Whittlewood, and she has suggested (ibid.) that its boundaries were an artificial creation made with local knowledge. This would be consistent with an origin as a new hundred created at the time of the suggested formation of the burh of Buckingham in 878-9 within Viking-controlled territory which was not at the time divided into hundreds. It is significant therefore that four of the connections of manors to Buckingham at the time of Domesday lay in this area. These comprise Biddlesden [1], Shalstone [3], Radclive [6] and Maids Moreton [14], though all except the first are less than certain inferences. This process of the annexation of Mercian / Viking territory by a new burh of the Burghal Hidage system is also shown in the case of Bath, which occupies a very similar topographical position to Buckingham, lying on the north bank of the river Avon, with its burghal territory extending both to its north and south and clearly carved out of the territory of the Hwicce, which at the time was held by the Mercian King Ceolwulf (Manco 1998).

This gives a total of 900 hides for the hundreds within Buckinghamshire and Berkshire. However, by including the hidages of manors in Tring hundred in Hertfordshire, which forms a noticeable salient into Buckinghamshire, the total could be brought to nearer the figure of 1000 hides indicated in the Burghal Hidage. Alternatively, the 10% shortfall could be considered as a normal example of a pattern seen in all the other Wessex shires (except Somerset), where the Burghal Hidage totals exceed the Domesday totals by amounts varying between 12% and 27%. As with Berkshire, discussed above, this shortfall could be explained by the omission in Domesday Book of demesne or inland hides on many estates (Finn 1963, 261; Bradbury 1988, 19), a subject discussed in more detail in relation to Buckinghamshire by Keith Bailey (Bailey 2010), and to Wiltshire and Hampshire, discussed in chapters 3 and 4. It should be noted that to include the Berkshire hundred of Charlton (145 hides) to the south of the Thames in the burghal territory of Sashes, as does David Roffe (Roffe 2009, 44), would introduce an unacceptable imbalance in the totals of the hidages of the territories of both Wallingford and Sashes. The burghal territory of Buckingham Even as recently as 1988 it was assumed that the formation of Buckinghamshire was a simple linear process in which the territory was determined, and the area first hidated, on the occasion of Edward the Elder’s reconquests of the Danelaw in the second decade of the tenth century (Bradbury 1988, 7-8, 17). The reality, as should be apparent from the discussion in this chapter, is somewhat more complex. This view ignores both earlier and later developments, and is of course based on the assumption that the burh at Buckingham in 914 was the first on its site – a deeply-entrenched paradigm which I have argued elsewhere is unsustainable (Haslam 2005, 141-4). This is also discussed further below (this chapter). From the broad distribution of the manors which were, or can be reasonably inferred as being, contributory to Buckingham, it is clear that its burghal territory would have occupied an area which comprised much (but not all) of the northern half of the later shire (fig. 21). While other alternatives are possible, an analysis of the hidages, below, suggests that the western boundary of the burghal territory marched with that of Oxford along the boundary of the Domesday and later shire. Here, as elsewhere, this appears to have been

There is no reason to doubt that in the few years before the restructuring of the political geography of the area around and to the north of London as a result of the withdrawal of the Vikings in London and the rest of Mercia to East Anglia, recorded in the Chronicle under 879, the area of north Buckinghamshire to the west of Watling Street and to the south of the Ouse was held by King Alfred, with territory to the north and to the north-east of Watling Street held by the Vikings based at Northampton. If this were not so, then the boundary between the territory around London which the Vikings annexed as a result of the ‘Partition’ of 877 would have had to have made an awkward salient westwards to include this area, doubling back on itself along the Great Ouse, until resuming its well-established course northwards up Watling Street. As I have already argued, the fact that the course of the boundary in Alfred and Guthrum’s Treaty is not specified to the north of its ending point at Stony Stratford (at the crossing of the Great Ouse by Watling Street) implies that this was a renegotiation of an earlier boundary which lay to the west, which also came to this

112

The Burghal Territories

point (as shown in fig. 24), and therefore that the de facto east-west West Saxon / Viking boundary along the Great Ouse had continued northwards up Watling Street (Hart 1970, 13; Haslam 2011a, 125-7). Evidence presented below (this chapter) can be taken as suggesting that all the land to the south-west of Watling street which lies in the southwestern part of later Northamptonshire comprised part of the burghal territory of Buckingham.

To this total Keith Bailey would add 235 hides comprising the southern part of his putative burghal territory which he has suggested extended southwards onto the crest of the Chiltern Hills, to include parts of the hundreds of Aylesbury, Stone and Risborough (Bailey 1994, 89). This would give a total of 1371 hides. He also adds the totals from the hundreds of Bunsty and Moulsoe (98 + 113) to arrive at a total of 1582 (1615 according to Bradbury’s totals), which is very similar to the total of 1600 in the Burghal Hidage.

The distribution of other contributory manors helps define the extent of the rest of the burghal territory of Buckingham, which pattern is consistent with the delineation of an area extending to the south to the northern boundary of the territory of Sashes, which is defined here by the extent of the putative minster parochia of Aylesbury (above). The eastern boundary of the burghal territory was arguably defined by the Viking / Mercian boundary established in c.877, described above, which can be reconstructed as followed the river Ouzel and the line of Watling Street from south to north (fig. 24). It is argued below that the extension of the later shire to the northeast to include the hundreds of Bunsty and Moulsoe, as well as much of the hundred of Seckloe, was the result of the incorporation of an area defined by these hundreds north-east of Watling Street to form the burghal territory of a new burh at Newport Pagnell in the early tenth century (stage 4 in the timeline outlined above, this chapter).

There are, however, inherent difficulties with this arrangement. The common boundary of the burghal territories of Buckingham and Sashes is, for reasons given above, more likely to have run along the river Thame to the north of the Chiltern scarp, rather than along an otherwise unevidenced alignment along the top of the Chilterns, which is not marked by any later boundary. This latter alignment would have cut through the early territory or regio of Aylesbury, and of its later minster, which must have been bounded by the river Thame (which was followed by the north-western boundaries of Aylesbury and Stone hundreds). This is more likely to have comprised the burghal territory of the later and secondary burh at Aylesbury itself (see below). This pattern of the use of rivers is seen in a number of the early land divisions in the area of Buckinghamshire and Oxfordshire (Thorn 1990, 25), in contrast to the pattern in western Northamptonshire, where they were determined more by the alignments of watersheds between rivers (Taylor-Moore 2012, ch. 4). The drawing of the southeastern boundary of the territory along the Thame would also be more consistent with the orbit of distribution of the contributory manors to Buckingham (fig. 20), discussed above. It would also mean that these 235 hides would be removed from the burghal territory of Sashes to the south, leaving it an area which was too restricted to cover its recorded 1000 hides.

There is no doubt that the burghal territory of the burh at Buckingham would have occupied most of the northern half of the later shire. As argued above, this will have included the hundred of Stotfold to the north of the river Great Ouse, but not the hundreds of Bunsty and Moulsoe to the north-east of the Domesday shire, which were grabbed from the territories dependent upon Viking Northampton and Bedford respectively in c.914 (as argued below). As is shown in the hidages of the various hundreds in the area, this is however too small an area to make up the total of 1600 assigned to Buckingham in the A Version of the burghal Hidage. The sum of the hides in this area, arranged by Domesday hundreds, are as follows (figures from Bailey 1994).

A possible solution which would give to Buckingham its quota of 1600 hides would be to include within its burghal territory the triple hundred of Ploughley in Oxfordshire to its west, plus the northern outlier of Headington hundred. This area would have covered the whole of the catchment area of the eastern side of the river Cherwell, which would have formed its western boundary (see figs. 19 and 20). The inclusion of this area might appear to form a more natural land division than the boundary of the later shire to the east, which is entirely artificial. The hidages of these add up to 284 for Ploughley hundred, plus 29 for Headington, totaling 313, thereby making up one of the triple hundreds seen elsewhere in Oxfordshire and Buckinghamshire (Thorn 1990; Blair 1994, 107-11 & map 62). The addition of this sum to the 1136 hides for the 10 hundreds surrounding Buckingham (above) totals 1449, which is nearer the total of 1600 given in the Burghal hidage, though even this figure falls somewhat short. It could also be suggested that this shortfall could have been made up of one or two of the hundreds which later made up the southwestern part of Northamptonshire, in particular the hundreds of King’s Sutton and Aldboldestowe (fig. 22). Each of these would

Hundreds (fig. 22): Hides Stotfold 98 Rowley 101 Lamua 127 Ashenden 112 Waddesden 91 Ixhill 121 Cottesloe 145 (Yardley 118) Mursley 123 Seckloe 100 Total 1136 (Bradbury (1988, 20) gives a total of 1169 hides for these hundreds -1018 hides excluding Yardley)

113

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

have had an original assessment of 100 hides, although subject to beneficial hidation by a factor of two-fifths by the time of Domesday (Hart 1970, 25-6). The extension of the burghal territory to include King’s Sutton would explain the inclusion of lands belonging to Northamptonshire within Ploughley hundred in Oxfordshire (as successor at a later date to the putative burghal territory of Buckingham), which Kim Taylor Moore has explained as areas of pastureland appurtenant to the royal estates especially of Kings Sutton and Kirtlington (the hundredal manor of Ploughley) and perhaps also Buckingham, which had been given to royal administrators before the Conquest (Taylor-Moore 2012, ch. 4).

reading of the text - that a hidage figure for Buckingham has merely been omitted by a careless copyist in the archetype of the B Version (which situation is shown so clearly in Hill’s table of parallel texts) – offers a perfectly adequate explanation for the lack of a figure for its hides in version B. This implies that Buckingham should take its rightful place in the list of all the other burhs as an original part of the system which they comprise. There is also every reason for accepting that the figure of 1600 hides is no less reliable than other figures given in the document. Reasons for suggesting that it is less reliable, or that it is an interpolation, have more to do with fitting the evidence to suit the hypothesis that Buckingham was not part of the system of burhs listed in the Burghal Hidage, which itself is based on inadequate or ill thought-out premises.

There are however clear disadvantages in this arrangement, in that this would give to Oxford a disproportionately small and awkward area compared to that of Buckingham, and would decrease its hidage total to well below the 2400 hides postulated for it – though rather more than the alternative of 1500 hides. Neither would this fit with the evidence of the connection to Oxford of two manors within the hundred of Ploughley, at Hampton [33] and Bletchingdon [34]. While there is a conspicuous absence of any other connections with Oxford of manors in the rest of the hundred, the same is also true for any connections with Buckingham.

There is, however, another solution which could account for the apparent shortfall in the hidage values in the area. This is that the full complement of Buckinghamshire’s 1600 hides could have been made up by the inclusion of most if not all of the area forming part of later Northamptonshire to the south-west of Watling Street. This area comprises somewhat under 700 hides (based on figures in Hart 1970), as follows: Hundreds (fig. 22): Hides (King’s) Sutton 100 Aldboldstow 100 (Chipping) Warden 100 Foxley 100 Cleyley (part of) 40 (by estimation) Towcester (part of) 40 (by estimation) Fawsley 100 Alwardsley 100 Sub-total 680

One solution to this puzzle, therefore, would be to allow to Buckingham only 1000 hides (represented by the hidage values of 1018 hides (or 1051 according to Bradbury) from nine hundreds enumerated above (excluding Yardley). Again, however, there are difficulties with this solution. It has been suggested that the B version of the Burghal Hidage implies a figure of 1000 hides for Buckingham (Hill 1969, table 1), but it is clear that this is merely an inference from a gap in all the MSS, where the figure of 1000 hides is read from that of Sashes which follows it. It does not, therefore, relate to any possible original hidage value for Buckingham’s territory. Various possibilities arising from the interpretation of the texts of the different versions of the Burghal Hidage at this point have been discussed by Nicholas Brooks, who has suggested that the reference to Buckingham should be seen as a later interpolation (Brooks 1996b, 89-90).

All except Cleyley and Towcester lie to the south-west of Watling Street; only these hundreds, surrounding and adjacent to Towcester, straddle Watling Street. Together they make up what Cyril Hart has called the ‘Eight hundreds of the South-west’, which are seen as a distinct and distinctive unit which had had a different history of development from the rest of Northamptonshire (Hart 1970, 13). The inclusion of this area, together with the 1018 hides for the hundreds mentioned above, would give a total of 1698 hides for Buckingham’s putative burghal territory. If the hundred hides of Egelweardesle (Alwardsley) in the northern corner of this grouping is omitted, then the total can be brought to within just a few hides of the documented figure of 1600 hides. Alternatively, as already suggested, the hundred of (King’s) Sutton could have been included with the territory of Oxford, which would provide a neat equivalence for its total hidage assessment of 2400 hides. From the point of view of its neatness on the map, this might seem a more acceptable solution. This would also mean that the earlier regio and minster territory of King’s Sutton would not have been divided from most of the rest of its territory (Foard 1985, 195-6).

I have however shown elsewhere that for various reasons these arguments are unsustainable (Haslam 2005, 141-4). The evidence from the spatial analysis of the relationship between the burghal territories of the burhs of Buckingham, Oxford, Wallingford and Sashes discussed in this study should make it quite clear that the formation of the burh of Buckingham and its territory must be part of the same formative process as the other three, and therefore contemporary with them. The figure of 1600 hides for Buckingham in the A version of the Burghal Hidage must therefore be considered to be the original and true assessment, unless it can (as with the case of Oxford, above) be demonstrated on independent evidence to be otherwise. Furthermore, and in spite of Brooks’ complex arguments to the contrary, it must be said that the simplest

114

The Burghal Territories

The grouping of the areas of these hundreds with the territory dependent on Buckingham in c.878-9 would make considerable strategic sense, in that it would have given control of a swathe of territory lying to the south-west of Watling Street to King Alfred. This would have represented the annexation of an area of upland, on the northern fringes of the Cotwolds, whose political affiliations had perhaps not been particularly closely defined, lying as it did some distance from the Viking territorial centre at Northampton, the Hwiccian centre at Gloucester, the Mercian centre at Tamworth, and the West Saxon centre at Oxford. As I have already discussed elsewhere, the imposition of West Saxon control on this area would have been put in place at a time in which the burh at Buckingham, and the control of its dependent territory which this would have entailed, would have represented the first phase of a major putsch by Alfred against Viking-controlled territory to the northeast of Watling Street (Haslam 2005, 141-4). In particular, this would have given Alfred some measure of control over the strategically-important Watling Street north of the Great Ouse. As is clear from developments only a year or two later, this control over Watling Street was extended to London itself and consolidated by the circumstances in which Alfred regained control of London and its immediate hinterland, of which one important outcome was the reversion of the Viking / West Saxon boundary to the river Lea to the east of London, which is set out in Alfred and Guthrum’s Treaty of c.880 (see fig. 24) (Kershaw 2000; Haslam 2005, 122-7; Haslam 2011a, 124-7).

manner described above, together with the assignment to it of 2400 rather than 1500 hides, makes it possible to arrive at a relatively neat fit between the hidages in the Burghal Hidage given to the other three burhs at Wallingford, Sashes and Buckingham and the disposition of these hides on the ground, in such a way as to best accommodate all the diverse topographical and other factors which must have played a part in the delineation of the territories of these other burhs. This is shown, firstly, by the incidence of the stated and inferred connections of rural manors lying in the Vale of the White Horse in northern Berkshire to Oxford, combined with the absence of connections to Wallingford; secondly, by the figure of 2400 hides which comprises the hidage of this area plus that of the rest of Oxfordshire to the north of the Thames, which is as might be expected from the territory of a burh of Oxford’s regional importance; and thirdly, by the fact that the subtraction of the hidage of this area, in northern Berkshire, together with a chunk of territory in the eastern part of Berkshire, gives a total hidage figure for Wallingford (including the southern part of Oxfordshire) of very nearly 2400 hides, its hidage value in all versions in the Burghal Hidage. The addition of this area in eastern Berkshire to a topographically-distinct territory to the north of the Thames in southern Buckinghamshire in turn gives a hidage value which approximates to the 1000 hides for Sashes in the Burghal Hidage (give or take a bit of later Hertfordshire to its east). This in its turn leaves an area of most of northern Buckinghamshire, as the area would have been defined in the late ninth century, as the burghal territory of Buckingham. The acceptance of the figure of 1600 hides for Buckingham in the A version of the Burghal Hidage makes it necessary to look for the balance of its hidage total in the area of south-western Northamptonshire to the south-west of Watling Street.

This area fits like a hand into a glove within the space between Watling Street to the north-east, part of the presumably already-established north-western boundary of the diocese of Dorchester (with the diocese of Lichfield and the putative burghal territory of Tamworth, now Warwickshire) to the north (fig. 21), and that of the territory of Oxford to the west, and represents a continuation to the north of the block of territory comprising the northern part of later Buckinghamshire. It would also place the burh of Buckingham near the geographic centre of its burghal area, rather than in the strategically rather awkward and exposed position in its northern edge. The exclusion of this area from the later shire of Buckinghamshire is likely to have been due to the fact that the establishment in 917 of a burh at Towcester, on Watling Street itself, would have swallowed up this area to incorporate it into its own burghal territory (stage 4 in the timeline, above), which would in turn have been incorporated into the later shire of Northamptonshire. In this process the single hundred of Stotfold, in which Buckingham itself was situated (fig. 22), would have been left to Buckinghamshire.

This figure of 2400 hides for Oxford is thus determined to a large extent by the disposition of the hidages of the areas of the other three burghal territories. Given that all these contiguous areas would have had to have been determined at one point in time not only to fit together but also to be concordant with pre-existing administrative boundaries, whether secular or ecclesiastical, this seems to be the only viable solution which accommodates all the variable factors and constraints which would have had to have been taken into consideration. There is, in short, little room in which to manoeuvre these territories in any other way around the mental map of the area as it was perceived in the late ninth century, and as it can be reconstructed in the twenty first. For these and other reasons there is no way that the figure for Oxford in the Burghal Hidage, of either 1300 or 1500 hides, can be accepted as being correct.

Summary and conclusions

These reconstructions, and the evidence on which they are based, bring to the fore an important conclusion. This is that the construction of the burh at Buckingham, with its dependent territory of 1600 hides, must be seen as an integral part of the programme of administrative and strategic reorganisation in the Upper Thames region of which the system of burhs recorded in the List of the

In conclusion, there are perhaps three strands of evidence which together determine the extent of the late ninthcentury burghal territory dependent on Oxford, and which in turn validate the general hypotheses set out above. The reconstruction of the burghal territory of Oxford in the

115

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Burghal Hidage document is the outcome. The fact that this territory interlocks with those of Oxford, Wallingford and Sashes, whose territories straddle the Thames and which in their turn interlock with the territories of all the other burhs comprising this system, shows this programme to have been conceived and implemented over a short period of time as part of a high-level administrative restructuring designed to enhance royal control of the area as a whole. In other words, all the four burhs discussed here - including,

emphatically, Buckingham itself – must have formed elements of a single system. It must be pointed out that this conclusion is at odds with the reconstruction recently put forward by David Baker and Stuart Brookes (Baker and Brooks 2011), which holds to the old paradigm that the burh of Buckingham can be no earlier than 914. The wider strategic implications of this arrangement are explored in the next chapter.

116

Chapter 12 Discussion – the Development of the Upper Thames Area, 9th to 11th Centuries of the Burghal Hidage burh of Oxford incorporated a significant area of northern Berkshire; that the primary burghal territory of Wallingford incorporated the area to the north and east of the Thames which later became the southern part of Oxfordshire; that the territory of Sashes also straddled the Thames, comprising the eastern part of Berkshire and the southern part of Buckinghamshire; and that the territory of Buckingham comprised not only the northern half of the later eponymous shire but also the whole area to the south-west of Watling Street which was included in later Northamptonshire. It is also quite clear that the three Domesday shires developed at a later date (in tandem also with Northamptonshire to the north) by a process of the rearrangement of these territories through probably two stages, which involved the modification and readjustment of the initial defensive arrangements in ways which represented administrative rather than strategic requirements. This in turn carries the implication that the establishment of the connections of manors to the centres in the areas affected by these changes can be seen as predating these changes.

The delineation of the burghal territories of the four burhs within the area of the upper Thames under discussion – Oxford, Wallingford, Sashes and Buckingham – in stage 2 of the timeline reconstructed above, provides crucial evidence for the ways in which strategic control of the wider area immediately to the north of the Thames was exercised successively by the Mercians, Vikings and West Saxons, all of whom were vying for territorial gains and, ultimately, overall political domination and the control of resources of southern and Midland England. The earlier context of these late ninth-century developments is explored in the following chapter. The establishment of these burghal territories carries important implications relating to the historical and strategic context of the Burghal Hidage document and of the creation of the burhs which it lists, which are generally seen as forming an important stage in these developments in the later ninth century. The details of these developments can be seen to have been crucial first steps in the subsequent extension of the hegemony of the West Saxon state, in regard to the lordship of King Alfred and his son King Edward and daughter Aethelflaed of Mercia, over what had been a formerly more powerful Mercian kingdom. It has been the intention here to argue that the record of the connections of rural manors to the burhs comprise a hitherto unnoticed class of evidence from which the spatial reconstruction of their primary dependent territories can be reasonably inferred, and thus gives important information about how these burhs functioned in terms of the strategic and political developments of the times. This in turn carries important implications for any interpretation of the ways in which events and processes on a wider historical stage unfolded at this crucial juncture.

This conclusion has a considerable bearing in particular on the interpretation of the context of the Burghal Hidage document, and in general on the ways in which King Alfred, and King Edward the Elder after him, were able to exert hegemony over an area which only a little while earlier had been part of the kingdom of Mercia. In the first place, the disposition of the burghal territories of the four burhs under discussion shows quite clearly that these burhs and their territories were formed and laid out at the same time and as part of the same process of strategic consolidation. It is not possible to envisage Buckingham and its territory, for instance, as being formed as part of a process which was later in date that the formation of its nearest neighbours of the territory of Oxford to its west and of Sashes to its south. The territories of Oxford and Wallingford were also clearly laid out at the same time with a carefullyconceived common boundary which respected not only obvious landscape features, but also earlier divisions between hundredal manors on both sides of the Thames, and which did not utilise the Thames itself except over a very short stretch. Similarly, the territory of Sashes was laid out to incorporate territories on both sides of the Thames which can be reasonably inferred to have reflected earlier administrative units or regiones. The northern boundary of this territory along the river Thame would have been

Perhaps the most important aspect of the delineation of the burghal territories of the four burhs of the Burghal Hidage discussed here is that it can be established beyond reasonable doubt that these primary territories interlocked in ways which disregarded the river Thames as a boundary. Some of the details of the connections of particular manors or estates to any particular centre might be open to discussion and reinterpretation, which could well modify some of the details of the wider patterns. The delineation of the burghal territories suggested below from evidence of hidage figures in the Burghal Hidage and Domesday might also be modified by further scrutiny. But the evidence taken as a whole seems to point unequivocally to the conclusion that the original or primary burghal territory

117

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

the true hidage figure for Oxford is more likely to have been 2400 rather than 1500 (or 1300 or 1400) hides, and by the identification of a burh at Ilchester in Somerset as part of the same system, which with possibly around 1300 hides appears to have been omitted from the list of burhs in both versions A and B of the Burghal Hidage (Haslam, forthcoming d).

merely a continuation eastwards, from a topographical point of view, of that to the west between the territories of Oxford and Wallingford. All four of these territories as reconstructed, furthermore, fit the known hidages given in the Burghal Hidage to a remarkable degree. In short, the organisation and laying out not only of the four burhs, but also the territories of which they were the centres, must have been contemporary manifestations of the same phase of strategic reorganisation and consolidation. An exactly analogous situation must have pertained with the formation of the burghal territories of Malmesbury, Cricklade, Chisbury and Wilton within the already-existing envelope formed by the boundaries of Wiltshire (analysed in chapters 2 and 3 above), and formation of the territories of Winchester, Christchurch (Twynham), Southampton and Portchester within the envelope formed by the boundaries of the shire of Hampshire (discussed in chapter 4).

On this slender and essentially misconceived (and illogical) basis the Burghal Hidage itself has been invariably dated to 914 or later. Steven Bassett has remarked on the ‘trap which we often meet with as historians – of presuming that something had only recently come into being when we first find it mentioned in our sources’ (Bassett 1996, 149). The whole question of the origin of the primary burh at Buckingham is just such a trap, if not a conceptual black hole into which all semblance of critical and logical deductive thinking down the decades has inexorably disappeared. I have, however, argued at length, and against the tide of received opinion, that the date of 914 for the formation of a burh at Buckingham is in reality only a terminus ante quem for the existence of a burh there – it would not be later in date, but could be earlier (Haslam 2005, 141-4). This date, on which so much uncritical weight has been placed, therefore gives absolutely no information which can be used as hard historical evidence about the date of the origin of the burh, and therefore of the context of the Burghal Hidage document or of the origin of the system of burhs which it lists. The determination, from the extent and the spatial interrelationship of the burghal territories, that all of the four central Thames burhs are contemporary has consolidated the evidential basis for this reassessment, and has established with little room for doubt that Buckingham must have formed a component of the original system of burhs which is listed in the Burghal Hidage.

It is necessary to emphasise this conclusion in view of the long-standing controversy regarding the status and the date of the burhs at Buckingham and Oxford - the only burhs included in the Burghal Hidage which lay to the north of the Thames - in the strategic development of the times. Since the process by which the formation of the territories which were assigned to them was contemporary with the laying out of the territories of burhs to the south of the Thames, it must be concluded that they were part of the same system, and built to address the same strategic concerns. Most commentators on the Burghal Hidage since the time of F W Maitland have assumed without question that a burh at Buckingham was first built in c.914, the date of its first mention in the Anglo-Saxon Chronicle, and that it was therefore an addition to a system which was probably already in existence (e.g. Brooks 1964; Hill 1969, 88; Hill 1996, 79; Brooks 1996b, 90; see also Haslam 2005, 141-4) . As I have pointed out, however, this begs the question as to why Buckingham was included in this earlier ‘system’ to the exclusion of burhs at Hertford (2 burhs), Maldon (first phase burh) and Witham, all constructed in 912 (ibid.). A contrary and reactionary restatement of this long-established paradigm has, however, been put forward by Baker and Brookes (Baker and Brookes 2011). It has clearly been necessary to hold this paradigm of interpretation because Buckingham (with Oxford) was self-evidently not in Wessex, and because all the other burhs are clearly so. Another reason for its exclusion from the sacred canon is that supposed or real anomalies in the final totals for the number of burhs in the system and the number hides in Wessex, given in the Appendix, are thereby thought to be eradicated – or at least smoothed out (Hill 1969, 86-8). However, these totals are themselves subject to uncertain and mutually-conflicting interpretations (Hill 1996a; Rumble 1996, 72; Brooks 1996a), in such a way as to show that the removal of Buckingham from the canon for these reasons is merely reinterpreting and rejigging the evidence to suit the hypothesis. The neat conclusions espoused by David Hill and Nicholas Brooks are also thrown into disarray by the demonstration, above, that

It must also be stressed that this forms the only possible context for the fact that Buckingham is listed in the Burghal Hidage in its expected, rightful and logical place in the circuit of all the burhs in Wessex. The compromise solution, that while the burhs considered as a system (without Buckingham) may well be ‘Alfredian’ in origin, the origin of the Burghal Hidage as a document must be placed after 914 when the burh at Buckingham was added to the system, completely ignores any development of the system by the rearrangement of burhs and burghal territories in the intervening decades which the wider landscape evidence seems to suggest. The more recent views of Baker and Brookes, however, do away entirely with the notion of the burhs of the Burghal Hidage as a system, allowing different dates of origin to be assigned to the burhs (including Buckingham) to suit their hypotheses (Baker and Brookes 2011, 111-4) with little regard to the strategic contexts of their formation, and with no regard to the importance of the burhs of the Burghal Hidage as forming an interlocking network of functionally-differentiated nodes, which covered an area which at the time of the formation of the system was a single polity controlled by King Alfred.

118

Discussion – the Development of the Upper Thames Area, 9th to 11th Centuries

These views concerning Buckingham are also reflected in interpretations about the origins of Oxford. John Blair in particular has argued that its origins as a burh should be sought in a Mercian context – i.e. as a creation by Ealdorman Aethelred in the late 890s (Blair 1994, 99101). This is based in part on an interpretation by David Dumville of the course of the boundary of the Treaty between Alfred and Guthrum, which he sees as defining the boundary between Mercia and Wessex from c.878 along most of the course of the Thames (Dumville 1992). I have given reasons for questioning Dumville’s thesis on strategic grounds (Haslam 2005, 122-4). Simon Keynes, amongst others, has also dismissed this interpretation, commenting that the course of the boundary as proposed by Dumville ‘seems, quite simply, to offend against common sense’, pointing out that objections to it ‘are several and obvious’ (Keynes 1998, 32 fig. 1, 33 & n.145), a view endorsed by Paul Kershaw (Kershaw 2000, 46). There is no basis, therefore, on which to reject the traditional view that the boundary defined Viking lands to the east of the River Lea in Essex and lands held by King Alfred in and around London and to the west. I have also given other reasons for suggesting that Blair’s interpretations about the origins of Oxford as a burh are not sustainable, based in part on a more recent re-evaluation of the evidence of the early coinage from Oxford (Haslam 2010a; see esp. Blackburn 1998, 110-1; Blackburn 2003, 214). These somewhat entrenched paradigms of an earlier age have, however, been given a new lease of life in the revisionist analysis Baker and Brookes, who have essentially restated Blair’s position, and for the same reasons and using the same logic, to the effect that the boundary between Mercia and Wessex in the later ninth century lay at this time along the Thames, and that both Oxford and Buckingham must therefore have been later additions to the series of burhs in Wessex proper (Baker and Brookes 2011, 106-114).

which all were conceived at the same time and constructed to meet the same range of strategic and political concerns, and which were all therefore complementary in function. As has been pointed out above, this understanding is in direct contradiction to the thesis propounded by Baker and Brookes, who have argued that at least the northern burhs of the Burghal Hidage were built to meet various disparate and non-contemporary strategic needs, and therefore did not in fact form part of a system. These conclusions open up a range of further issues. If the dependent territories of all four of the burhs at Oxford, Wallingford, Sashes and Buckingham shared land on the northern side of the Thames, their combined areas must have formed a de facto part of the West Saxon kingdom, over which King Alfred had effective control at the time that they, and the system of which they were a part, were constructed. It would follow that the areas to the north of the Thames to the east (London and its surroundings), to the west (essentially the kingdom of the Hwicce), and to the north (all of Northamptonshire north-east of Watling Street), were not in Alfred’s control, and that the outer boundaries of this Oxfordshire/Buckinghamshire unit - i.e. the area to the north of the Thames - were defined in relationship to these political units which were controlled by competing if not actively hostile polities. I have already given detailed arguments for holding that the ‘best-fit’ solution to establishing an appropriate historical situation in which the establishment of King Alfred’s hegemony over this area would have been possible, and in which these areas on its fringes would have represented potentially hostile territories, was the short period between Alfred’s defeat of Guthrum’s Viking army at Eddington in May 878 and the death of King Ceolwulf and the general Viking retreat from Mercia and the London area in or by late 879 (Haslam 2005; Haslam 2011a). This is supported (in general terms) by the analysis of Richard Abels (Abels 1988, 62-5). In this particular window of time the area to the west and to the north of Oxfordshire would have been under the control of Ceolwulf and his Viking overlords; and that to the east – represented by London and its associated area, with a boundary along the Colne northwards along the river Gade and over the Chilterns to Watling Street - still part of the same polity. As I have argued elsewhere, this situation would have been blown apart by events which led to the death of Ceolwulf and the general retreat of the Vikings to East Anglia from southern Mercia and the London area in late 879, and the subsequent assumption of control of the whole of Ceolwulf’s former territory by King Alfred under a new polity that contemporaries called ‘the kingdom of the Anglo-Saxons’ (Haslam 2005; Haslam 2011, 127-31; pace Keynes 1998, 24-6, 34-9, 43-4; Keynes 2001, 44-8), which Simon Keynes has described as a ‘union of two peoples for the purposes of political organisation’ (Keynes 2001, 45). It is this control which is referred to by the adoption of the ‘buzz-word’ Angelcynn to refer to the peoples of both Wessex and southern Mercia which is used in both the Treaty between Alfred and Guthrum of

The evidence adduced in this study, however, opens up a way of reinterpreting these issues from a different perspective. It is important to have established, in the first place, that the burhs at Buckingham and Oxford to the north of the Thames are contemporary with the burhs at Wallingford and Sashes which lay to the south, with whose territories they interlocked, and are therefore contemporary with other burhs with contiguous and interlocking territories over the rest of Wessex. The evidential basis of this is that the burghal territories of Oxford, Wallingford and Sashes extended across the Thames, thereby ignoring it as a frontier or boundary. This hypothesis of the boundary between Wessex and Mercia in the ninth century along the Thames, espoused by Dumville and by Baker and Brookes, is also rendered untenable by the considerable body of evidence, discussed below (this chapter), which demonstrates that King Alfred exercised hegemony over the central area, extending eastwards to include London and its region, from the early 870s. In the second place, the interlocking spatial relationships of the burghal territories are prima facie evidence for the existence of the burhs which are listed in the Burghal Hidage as a system, in

119

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Mercia and Aethelwulf of Wessex, Alfred’s predecessor, demonstrates a degree of political accommodation between the two kingdoms, a situation which Alfred inherited on his accession in 871 (Keynes 1998, 34-8). The Mercian hegemony over Berkshire is shown in the grant of land at Pangbourne (to the south of Wallingford) by the bishop of Leicester to the Mercian king Berhtwulf in 844 to secure immunities for Abingdon Abbey. This estate was then passed on to Ealdorman Aethelwulf (Stenton 1913, 25-6). However, the fact of the birth of Alfred in Wantage in 849, and the reversion of the loyalties of Ealdorman Aethelwulf to the West Saxons from the early 850s (Keynes 1998, 6 and n.19 & 20; Brooks 2003, 157), shows that by this time the area to the south of the Thames seems to have reverted to West Saxon overlordship (For a contrary view and useful discussion, see Smyth 1995, 3-9). In 862 an estate of 10 hides at Wittenham in central Berkshire, just to the south of Abingdon, was given to Ealdorman Aethelwulf by Aethelred of the West Saxons as a reward for service (S.335; Brooks 2003, 157), which shows that by this time probably all of Berkshire was regarded as West Saxon.

c.880, and in the reference to the general submission of “all the English people that were not under subjection to the Danes” to Alfred in 886 (Whitelock 1979, 199; Kershaw 2000, 47-50, 58-9) It must be concluded that since the ‘envelope’ of territories dependent upon Oxford, Wallingford, Sashes and Buckingham fits into the space on the upper Thames which is defined by these hostile polities, then it must have been created within the time frame in which these polities were a determining factor in the development of the political and strategic landscape. The initial phase of development of these burhs and their interlocking territories, which are defined both by their hidages and by the orbits of distribution of the urban-rural connections, cannot be placed at any time either before or after this short time-frame. What still remains, however, is the issue as to how King Alfred of Wessex was able to gain the control of this area to the north of the Thames by the late 870s, which perhaps only a generation or two earlier had been securely part of Mercian territory. The process by which the area of later Berkshire became West Saxon territory, perhaps during the 840s, has been examined from the point of view of the history of Abingdon Abbey in particular (Stenton 1913, 1726), and of Berkshire in general (Gelling 1973-6, 838-47). Neither of these commentators, however, has considered the status of either Oxfordshire or Buckinghamshire. The development of the Oxford area until the early tenth century has been analysed by John Blair in terms of its status solely as a Mercian province (Blair 1994, 92-105). This is in part based on the now exploded interpretation by David Dumville of Alfred and Guthrum’s Treaty as defining the area of Guthrum’s hegemony as lying in ‘Mercia’ in the area north of the Thames in the upper Thames region, referred to above. It thus clearly informs his conclusions about the development of Oxford as a burh created in a Mercian context. This interpretation is, furthermore, perpetuated (for apparently the same reasons) in the more recent analysis of Baker and Brookes (Baker and Brookes 2011, 110-1). This question, however, has to be addressed from a different angle, since the evidence adduced above of the extension of the burghal territories of Wallingford and Sashes in particular to the north of the Thames points quite clearly to a different course of development of the Oxfordshire / Buckinghamshire area in the late 870s.

It is not difficult to surmise that the cooperation between Wessex and Mercia would have hugely facilitated by the marriage of Burgred of Mercia to Aethelswith, daughter of King Aethelwulf of Wessex (and Alfred’s sister), in 853 (as Stenton has indeed suggested – Stenton 1913, 28-9), and by the marriage of Alfred himself to Ealswith, daughter of a Mercian ealdorman, in 868. Already by c.867 Aethelred of the West Saxons had established a ‘monetary union’ between Wessex and Mercia (Blackburn 2003, 204-5). These inter-dynastic alliances are likely to have provided the essential context for the apparent expansion of Alfred’s hegemony to include London and its area in the mid 870s, where he was able to control minting operations to the extent of issuing coinage in his own name through London moneyers, and in generating a significant and wide-ranging monetary reform. This reform involved the production of new types struck to a Carolingian weight standard and a restoration of the silver content of the earlier coins (Blackburn 1998, 106-9, 112; Blackburn 2003, 205-7). As Blackburn comments, ‘Political historians should thus be aware of the strength of government exercised by Alfred in implementing this recoinage, as early as the mid 870s’ (Blackburn 2003, 206). Furthermore, moneyers in London were issuing coins in Alfred’s name during the last years of Burgred’s reign. After the latter’s deposition in the coup of 874, Alfred shared minting operations in London for a time with Ceolwulf, the Viking appointee. Most of the latter’s coins, however, can be ascribed to the period after 877, when those of Alfred ceased to be produced (Keynes 1998, 17-19; Blackburn 1998, 119-20). This evidence requires, in the words of Mark Blackburn, ‘some radical rethinking about the course of events following Burgred’s exile in 873/4’ (Blackburn 1998, 119). At the very least, it shows that Alfred had an active interest in maintaining a hegemony over both the upper Thames area as well as London in the years before

Of particular relevance is the more recent evidence of King Alfred’s early coinage, which has been examined by both Mark Blackburn and Simon Keynes (Blackburn 1998; Keynes 1998; Blackburn 2003). I have discussed this evidence elsewhere (Haslam 2005, 128-9; Haslam 2011a, 123-4), but the conclusions which flow from it need to be re-emphasised in view of the fact that the analysis of this crucial evidence for the period by Blackburn and Keynes was not available in Blair’s discussion of 1994, and, furthermore, has been largely ignored in the more recent discussion of Baker and Brookes (Baker and Brookes 2011). Already in the 860s the coinage of Burgred of

120

Discussion – the Development of the Upper Thames Area, 9th to 11th Centuries

the Viking ‘Partition’ of 877 and his rout at Chippenham in early 878. The few coins of Ceolwulf of this early period are particularly indicative of the political dynamics between Alfred and Ceolwulf. Blackburn has suggested that the production of coins in the name of Ceolwulf from a London mint in c.875 suggests that Ceolwulf was ‘a partner with Alfred in the original recoinage’ (Keynes 1998, 17-18; Blackburn 2003, 213). I have however, argued that a more reasonable characterisation of the political dynamic of the time would be to see this as evidence of rivalry, in which Ceolwulf was attempting, with the support of his Viking sponsors, to establish his power base in London by the issue of an alternative coinage (Haslam 2005, 128-9). It may even have been the case that even before Burgred’s exile in 874 King Alfred saw himself as the natural successor of Burgred as king of Mercia as well as of the West Saxons, and that the extension of his power – for that is what minting implies – over southern Mercia was a course of action in which he saw himself as fulfilling a well-considered option in agreement with the wishes of Burgred himself. This is admittedly somewhat speculative, but would be consistent with, and provide a reason for, the development of Alfred’s power in the area north of the Thames in the mid 870s which can be inferred from the evidence of the coinage.

have remained, any political and/or economic control of the area by Alfred was no longer possible. As has so often been recounted, this situation was changed dramatically with the defeat of Guthrum’s Viking forces by those of King Alfred at Edington in May 878 (Stenton 1971, 255-7; Whitelock 1977), which resulted in the submission of Guthrum and his commanders to Alfred, their baptism, and their subsequent retreat to Cirencester just across the border of Wessex within the lands of the Hwicce. There is no reason to suppose, however, that this process resulted in the removal of the Viking presence from London, since a new Viking army moved in to a base at Fulham to the west of London at this time or soon after – possibly even as a protection of London and its region from incursions by West Saxon forces from the west. On the contrary, Ceolwulf still remained in power, evidently producing his own coins at a mint in London, until removed from the scene (possibly in a coup d’etat at the instigation of Alfred himself) in late 879. It was at this time that Guthrum and his Vikings retreated to a new sovereign state in East Anglia, and the Fulham Vikings to the continent, arguably through the forceful political (and perhaps military) pressures exerted by Alfred. It is argued, however, that in the intervening period between his victory at Edington and this retreat of the Vikings in late 879, and as a result of his increased standing in the eyes of the Mercians (not to mention the West Saxons) earned by his victory over the Vikings at Edington, Alfred was able to regain his hegemony over the area to the north of the Thames from early or mid 878, a process no doubt facilitated by the connections of his family with the Mercians and in particular of his sister as wife of the illfated Burgred, and by the memories of his former control over the area (Haslam 2011a). This window of opportunity would have allowed King Alfred to have built the burghal system which is described in the Burghal Hidage (Haslam 2005), and which explicitly extends over the area to the north of the Thames which King Alfred had regained as a result of his victory at the battle of Edington. This particular historical juncture appears to correspond exactly with the revival of minting operations and a second reform of the coinage in c.879-80, which seems to have been associated with the new burghal foundation at Winchester and probably other new burhs in Wessex, including perhaps Oxford, which process can be independently adduced from the evidence of the coinage of the period (Blackburn 2003, 207-8, 211-2). It seems highly likely, therefore, that it was the development of the burhs which provided the essential context for the reform of the coinage and the extension of minting activity at this time (Blackburn 1996; Blackburn 2003, 207-8). In short, it is the dating of the coinage and the mints which provides a particularly strong, if not conclusive, support for the general hypothesis of the dating of the construction of the burghal system, as listed in the Burghal Hidage, to this particular period.

I have, furthermore, argued that the sequence in which Ceolwulf’s minting activity in London succeeded that of Alfred, which Blackburn has demonstrated from the evidence of the coinage itself, can best be explained as being the direct result of the extension of Viking control over London and its area as a result of their ‘sharing out’ of eastern Mercia in 877, in which Ceolwulf, as a Viking appointee, appears to have gained sole control of the output of the London mints at the expense of the interests of King Alfred (Blackburn 1998; Keynes 1998, 17; Blackburn 2003, 212-4; Haslam 2005, 128; Haslam 2011a, 123-4). This ‘partition’ is arguably the context for the imposition of a boundary to the west of London (fig. 24), which has been discussed above (and in more detail in Haslam 1997, 118-23; Haslam 2005, 128; Haslam 2011aaa, 126-7). As a result of the deposition of Alfred himself by the Vikings at Chippenham in early 878 very soon after the ‘partition’ of 877, Alfred would have lost all control in the London area, if not also the area to the north of the Thames around Oxford. This appears to be supported by the evidence of coin dies supplied from a moneyer named Dunna in Winchester to a probable mint in Oxford in the 870s. Three of these are in the name of Alfred (styled REX SAX), but the last in the succession of four was in the name of Ceolwulf (styled R) (Blackburn and Keynes 1998, 143; Blackburn 2003, 217). This reflects the sequence noted in the coinage from London (Blackburn 2003, 212-4). From Alfred’s perspective this loss of political control would indeed have been a double whammy of disastrous proportions, though for Ceolwulf and the Vikings it would have given them everything they would have wanted – control of London and southern Mercia without Alfred’s ‘interference’. While political and indeed military loyalties of some Mercians to Alfred may

It is therefore this political situation which appears to provide the best context for the imposition of the burghal

121

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

system not only over this particular corner of the former Mercian kingdom, but also over Wessex as a whole. I have given reasons for holding that this process was, furthermore, extended to King Alfred’s assumption of control over the London area after the Vikings retreated to East Anglia in late 879 (with the common boundary between the two kingdoms to the east of London established in Alfred and Guthrum’s Treaty), and by his construction at London of a new burh at this juncture or very soon after (Haslam 2010b; Haslam 2010c; Haslam 2011a, 131-40). This event was indeed marked not only by the reform of the coinage, based on the mints in London, but also by the issue in particular of the London Monogram coinage, which, with the issue of coins from mints in Oxford and Gloucester, was struck to commemorate Alfred’s newly-acquired hegemony over southern Mercia, including London. I have also argued that one of the contributing factors which resulted in the removal of the direct influence of the two Viking armies from at least southern Mercia in late 879 is likely to have been the new military, strategic and political clout which the construction of the burghal system in the period 878-9 gave to Alfred. The retirement of Guthrum and his army to East Anglia can in this light be seen as a face-saving exit strategy in response to the overwhelming odds which Alfred had placed in his way (Haslam 2005, 129; Haslam 2011a, 128). At the same time, and clearly as a consequence of these developments, Alfred was able to gain control over the whole of southern Mercia, including the former kingdoms of the Hwicce and the Magonsaeten, which from this time

effectively became part of a single province of Wessex under King Alfred’s control (Keynes 1998, 24-6, 34-9, 43-4; Keynes 2001, 45-8). It is these developments which must have formed the essential precursors to the general submission to King Alfred in 886 of “all the English people (angelcynn) that were not under subjection to the Danes” recorded in the Chronicle. It is necessary to restate these realities concerning Alfred’s control over central, southern and south-eastern Mercia in the 870s (albeit with a brief hiccup in early 878) – which was extended to include western Mercia after the Vikings’ retreat to east Anglia in late 879 - in order to establish the political context for the formation of the burghal system of the Burghal Hidage, and to counter some of the revisionist thinking in the recent paper by Baker and Brookes (2011). A detailed response to their thesis must, however, await a separate treatment. Since, as has been so clearly demonstrated, the formation of the burghal territories of two burhs of Buckingham and Oxford to the north of the Thames interlock spatially across the river Thames with the territories of Wallingford and Sashes to the south of the river, and thereby with the burhs in the rest of Wessex to the south, the assignment of this particular political and strategic context to their formation also gives an appropriate context for the formation of all the other burhs in the rest of Wessex which are listed in the Burghal Hidage, conceived as part of the same system.

122

Chapter 13 Before and After the Burhs of the Burghal Hidage Stage 1 – the development of a Mercian proto-shire or regio

Another anomaly is the connection of Pyrton [Ox 18, Brks 26] with both Oxford and Wallingford, in both of which it had tenements which were appurtenant to it. This is the only example of this dual relationship in the three shires. Pyrton lay in south Oxfordshire and, as already determined, therefore within the putative burghal territory of Wallingford. Other examples of this dual connection to two centres is explained in other shires, such as between Worcester and Droitwich, and Gloucester and Winchcombe, as the overlapping of orbits of the ‘territories of obligation’ of each centre as successive stages of development. Given this interpretation, this spatial anomaly can best be explained by the connection of Pyrton with Oxford having been overlain by a connection with Wallingford, the former being considered as being primary in both space and time. This would place Pyrton, and the area of southern Oxfordshire, within a wider area which had originally been dependent on Oxford at the time of the formation of this connection. This inference is strengthened by the particular status of Pyrton as a minster, the subject of a large land grant by Aethelred in 887 (S.217; Blair 1994, 112; Hammond 1998), and as a 40-hide manor held in 1086 by Earl Hugh of Chester as successor to Archbishop Stigand (Ox 15,2) (Sturdy & Munby 1985, 50), all of which would indicate this as being an ancient centre of royal administration of some antiquity and importance in the ninth century and perhaps earlier.

The initial development of a large Mercian territory, with Oxford as its centre, covering much of Oxfordshire, Buckinghamshire and Berkshire, as a distinct stage in the development of the area from the middle Saxon regiones to the late ninth-century burghal territories can be postulated, if only somewhat tentatively, as an explanation of some of the anomalies which are brought out in the examination of the patterns of connections of rural estates which were contributory to Oxford, described above. A particularly instructive anomaly exists in the holding of Streatley in Berkshire [Ox 37] by Geoffrey de Mandeville as contributory to a tenement in Oxford. As is shown in the map (fig. 20) this lies due south of Wallingford, well into what would have been Wallingford’s burghal territory. There is no possibility that this could have been included within Oxford’s territory at any time after the formation of both Oxford and Wallingford as burhs. There is no evidence that Streatley had been added to one of Geoffrey’s manors at a comparatively late date, perhaps sometime soon before the Conquest (as was the case with a tenement in Reading). Even if it had been, its anomalous position in relation to both Oxford and Wallingford would require an explanation. Nor is it possible that this was included in Oxford’s territory in any reshuffle of the respective territories of Oxford an Wallingford in the mid tenth or the early eleventh century, which is discussed further below. Its spatial relationship to Wallingford suggests strongly, therefore, that its connection with Oxford would have been earlier than the foundation of the burh at Wallingford, and that it therefore represents a survival of a connection with Oxford which existed before the late ninth-century imposition of burghal territories on this part of the upper Thames region. Its inferred significance at this time would be consistent with the mention of Streatley, with its appurtenances, as an estate given to the early abbey at Abingdon by King Ine in the late seventh or early eighth century (Stenton 1913, 12-14). As such, it would be comparable to the connection between Wilton in the south of Wiltshire with Castle Combe in the north, which has been argued above (chapter 3) as being the survivor of a connection with Wilton as the centre of an early regio, or between the early centre at Winchcombe and Lechlade in Gloucester (discussed in chapter 6).

Two other anomalies in the pattern of distribution of connections to Oxford can be interpreted in the same way. In Buckinghamshire, the two manors of Princes Risborough and Twyford, both owned by the king, had appurtenant tenements in Oxford (DB Ox B5; DB Buk 1,3; 37,1). Princes Risborough lies within the putative burghal territory of Sashes, and Twyford within that of Buckingham. The latter is considerably nearer to Buckingham than to Oxford, which is consistent with the pattern of a primary connection to Oxford being overlain by the orbit of connections to Buckingham at a later stage. A similar conclusion can be applied to Princes Risborough, which is nearer to Aylesbury (arguably a burh of stage 3, above) than to Oxford. The antiquity of Princes Risborough as a regional centre is indicated by its status as a villa regalis and an early minster, and as the terminus of a salt way from Droitwich of probably Mercian origin (Blair 1994, 84-7). As well as the documented connection of Oxford

123

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

with Princes Risborough, it is possible, as suggested above, that other manors in the same area of Buckinghamshire held by the Archbishop of Canterbury at Haddenham (DB Bks 2,1) and Halton (DB Bks 2,2) (Ha and Hl in fig. 20) could have contributed to one or several of the archbishop’s seven tenements which he held in Oxford. This, however, must remain within the realms of speculation. Similarly speculative are the possible connections of Crowell [Ox 20] and/or Emmington [Ox 21], both within the putative burghal territory of Wallingford, with William Peverel’s four tenements in Oxford. These connections are inferred from the fact that these are the only two manors held by William Peverel in Oxfordshire.

exchange systems in the distribution of the coinage of the period (Blair 1994, 80-87), would give further support for this suggestion. It also seems to be a strong candidate for the location of a mint in the eighth and/or ninth centuries (Blair 1994, 84; Blackburn 2003; 207, 212). The role of Oxford as an administrative centre of an early regio which included the area of later Berkshire to the south of Wallingford would fit a historical and geographical setting in the early ninth century, at the zenith of Mercian power and its territorial extent before Berkshire came under the control of the West Saxons (Stenton 1913; Gelling 1973-6, 838-43; Blair 1994, 54-6). At this time this area would have comprised a constituent part of what John Blair has called ‘a great territorial federation’ created by the Mercian King Offa both to the north and south of the Thames (Blair 1994, 55).

It is of course quite possible that other connections with Oxford, whether stated or inferred, might have originated in this earlier stratum of connections, rather than with those which were consequent upon the formation of the burh in the late ninth century, as argued here. Connections with some of the more distant manors to the west are possible candidates, though those in a group of nine on the northern edge of the shire would, as discussed above, appear to be contemporary with the imposition of the boundaries of the burghal territory over earlier regiones. Another possible candidate is Steventon [Ox 36], a former comital manor, which, as in the case of Streatley, is given in the Berkshire folios of Domesday Book (DB Brk 1,39). The significance of this is, however, not particularly diagnostic.

It is of some diagnostic importance for this interpretation that Oxford was considered the centre of a wide area comprising the southern part of the former kingdom of Mercia when, together with the territory dependent on London, it was taken under the control of Edward the Elder in 911 on the death of the sub-regulus Aethelred. It was arguably this set of circumstances which occasioned the building of a second burh at Oxford as an extension of the primary burh of King Alfred (Haslam 2010a). All of this evidence, together with that of the connections to Oxford of the rural manors just discussed, is consistent with the status of Oxford as the centre of a regio of Mercian origin which covered a large area which spanned the upper reaches of the Thames, and which was the precursor of the burghal territories of the late ninth century into which it was divided.

There is therefore evidence of several connections between rural manors and appurtenant tenements in Oxford which can arguably only be satisfactorily interpreted as survivals from a nexus of such connections which pre-date the creation of the burghal territories of Wallingford (at Streatley), Sashes (at Princes Risborough) and Buckingham (at Twyford). None of these connections can be explained by any reference to its proximity to an adjacent urban centre. All of them are high status sites of inferred or documented antiquity, either as royal tunas or possessing a minster, or both, which would be expected to have played a key part as centres of royal administration of their areas in the middle Saxon period.

Stage 3 – The development of the Burghal Hidage system in the later ninth century I have discussed elsewhere the existence of a phase of the development of the system of burhs in Wessex as a whole which is listed in the Burghal Hidage, which involved the replacement of many if not all of the small and temporary burghal fortifications, all within fortifications of earlier origin, by new and larger burhs on more strategicallylocated sites. The replacement of Pilton by Barnstaple, and Halwell by Totnes, both in Devon, are text-book examples of this process (Haslam 1984b). This process was arguably contemporary with the replacement of the unstable timber and/or turf palisade or revetment at the front of the defensive banks of many of the larger burhs with stone walls, which can be recognised in sites all over Wessex as well as in western Mercia. Both these processes can be most satisfactorily assigned to the early 890s in the context of a concerted response to the new threats posed by renewed Viking incursions at the time (Haslam 2009, 98-100, 103-4; Haslam 2011a, 211-5). This process can also be recognised in the suggested replacement of the rather awkwardly-placed island fortress of Sashes on the Thames with two new burhs - at Reading to the south of the Thames, and at Aylesbury to the north. In this process

Although the writer’s hypothesis of the existence of a defended burh at Oxford from the time of Offa in the late eighth century (Haslam 1987) has not been supported by tangible archaeological evidence of defences along the line which they followed later, there is growing evidence for the importance of Oxford as both a regional administrative centre as well as a minster from this time or even earlier, possibly replacing Benson as a regional royal administrative centre. This includes archaeological evidence for an important river crossing on the line of the main south-tonorth highway, comprising an early-Saxon causeway and mid-Saxon timber bridge and other features (Dodd 2003, 12-16), as well as a late seventh-century minster with midand late-Saxon burials (Blair 1994, 87-92; Dodd 2003, 1718). Evidence for the development of Oxford as a focus of both a flourishing market economy as well as long-distance

124

Before and After the Burhs of the Burghal Hidage

the burghal territory of Sashes, which spanned the Thames, would have been divided between the two new burhs, with their common boundary probably now lying, for the first time, along the Thames.

manors contributory to either Aylesbury or Reading, but this development would put into context the archaeological and topographical evidence from both places which could be reasonably be interpreted as indicating the existence of new burghal foundations at these places.

Though there is little absolute dating evidence for this phase, it is nevertheless a hypothesis which provides a best-fit explanation, in terms of known developments of the period, for a wide range of both historical and archaeological observations. As such it puts various features of the wider historic landscape into an appropriate and meaningful historical context. The archaeological evidence from Aylesbury, in particular, has demonstrated the existence there of an Iron-Age hillfort which provided the locus not only for an important minster church, with which were associated middle Saxon burials, but also for a period of refortification in the late Saxon period (EUS Aylesbury 2009, 52-7). The scale of this episode of refortification suggest that this was the result of the construction of a burh – as originally suggested by Michael Farley nearly forty years ago (Farley 1974; see also Farley 2007). The evidence of the extent of the minster lands of Aylesbury and of its relationship to the burghal territories of Buckingham and Sashes, discussed above, shows that the new arrangements entailed in the replacement of one burh by another would have would have fitted in well with the existing administrative landscape.

Stage 4 – the ongoing development of the burghal system in the early tenth century. Soon after King Edward the Elder “succeeded to London and Oxford and to all the lands which belonged to them” in 911 (Whitelock 1979, 211; and see above), he started on a long campaign to extend the control of the West Saxon kingdom over Viking-held territory to the north and east, while at the same time his sister Aethelflaed was pursuing the same goal in the west Midlands (chapter 8 above). Both these strategies were effected by long drawn-out campaigns which involved the building of burhs at strategic locations. These were arguably designed both as offensive military instruments, as well as staged components of a more politically-motivated strategy which was designed to gain the control of Viking-held territories by forcing the submission of the separate armies and peoples and the recognition of the lordship of Aethelflaed and Edward. This aspect is discussed further below. The probability that the detailed record of Edward’s exploits which has survived in the Chronicle was written while on campaign by Edward’s youngest brother, argued recently by David Pelteret (Pelteret 2009), implies that this account reflected events in a way which approaches a more-or-less objective historical ‘narrative’, in the sense that the actual progress of events was recorded to express the royal perception of their significance more closely than had been the case up to that time. I have already explored the internal structure of this narrative as expressing the political need to emphasise the process of submission to Edward by the Viking armies and the people of their territories (Haslam 1997, 111-8). This process had already been established by King Alfred in 886, when as a result of his ‘occupation’ of London, when “all the English people that were not under the subjection to the Danes submitted to him” (Whitelock 1979, 199) - even though the chronology and perhaps the significance of this whole process can be reinterpreted in a somewhat different way that might appear from a casual reading of the annal (Haslam 2011a, 139-40).

The existence of a burh at Reading is not so well attested from archaeological evidence. Its siting on a spur of land which divided the Thames and its tributary the river Kennet would, however, have provided an ideal site, which would have been strategically placed to control movement both across and along the two rivers. These same strategic reasons clearly influenced the choice of this site as a Viking camp in c.871, at which time it is mentioned as being a royal vill (Asser ch. 35, in Keynes & Lapidge 1983, 78; Yorke forthcoming). While no sustained case has been made for the construction of a West Saxon burh at Reading, and no line of defences demonstrated archaeologically, this development would certainly be consistent with the compact topography of its historic core, which was centred around the church of St Mary with its adjacent market area and sited in a significant strategic relationship to the crossings of the Kennet and Thames rivers. This is shown in Grenville Astill’s reconstruction of the topography (Astill 1984, 703), although there are grounds for suggesting the existence of a burh of rather larger size, based on the topographical evidence of the lie of the land. The existence of a burh here would also provide the historic context for the slight indications of heterogeneous tenure in Domesday Book (table 23). This being so, it is likely that the development of a burh at Reading would have entailed the transfer to it of the portion of the former burghal territory of Sashes to the south of the Thames, to which were probably added the areas which became Reading and Charlton hundreds from the primary territory dependent on Wallingford (figs. 20 & 22). This is, admittedly, somewhat conjectural; there is little evidence of this process in the distribution of

In 914, King Edward, in the words of the Chronicle, “went to Buckingham with his army, and stayed there four weeks, and made both the boroughs, on each side of the river, before he went away.” (Whitelock 1979, 213). This reference is persistently and misleadingly interpreted by almost every commentator as indicating the presence of a ‘double burh’ at Buckingham (e.g. EUS Buckingham 2009, 35-9), whereas the Chronicle entry merely states that two burhs were built while Edward stayed at Buckingham – one of them presumably at Buckingham itself. I have given reasons for suggesting not only that this episode represented a refurbishment of an earlier burh at Buckingham which had already been established since the late 870s and which

125

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

coercion implied in the annexation of the areas of these two hundreds to the control of the West Saxon king would have led directly to the outcome of the submissions of men from both Viking armies from Bedford and Northampton, which is recorded in detail in the Chronicle account of 914 quoted above.

was part of the original Burghal Hidage system of 878-9 (Haslam 1997, 125-6, n.23; Haslam 2005, 131, 141-4), but also that the second burh was sited at Newport Pagnell to the east, which is indeed on the other side of the river Ouse to Buckingham (Haslam 1997, 124-5 n.22). This identification (although contrary to the views of A H J Baines) is consistent with details of its built topography, its early development, and the scraps of archaeological evidence from the town (Baines 1986; Robinson 1975; Beamish and Parkhouse 1991; EUS Newport Pagnell 2010). It also provides a solution to the conundrum that, contrary to the views expressed in the recent Extensive Urban Survey, there is no archaeological or topographical evidence at or near Buckingham to indicate the presence or the location of this hypothesised second burh (EUS Buckingham, 359). The location of a second burh on the south side of Buckingham would not only be highly improbable from a topographical or landscape point of view; it is also unlikely to have acted in any meaningful way as a defence against navigation on the Great Ouse, which is at this point only a shallow stream. It would also make strategic nonsense, in that the burghal territory of this supposed second burh at Buckingham on the southern side of the river would inevitably have been cut out of the already-existing territory of the burh on the northern side, and which would have extended to the south of the river as well as to the north, and which therefore represented territory which was already in the control of the West Saxons.

These two hundreds can therefore be seen as forming part of the new burghal territory of Newport Pagnell, formed by the taking in of some of the former Vikingheld territory dependent on Northampton to the north and on Bedford to the east. Seckloe hundred appears to be a later amalgamation of the area to the north-east of Watling Street south of the Ouse and some territory south-west of Watling Street within the area of the primary burghal territory of Buckingham. This is as clear a case as any which demonstrates the formation of these hundreds at a particular historical context, as former Viking lands which now became dependent upon the new West Saxon burh, to which the holders of their constituent manors now owed allegiance and, doubtless, service. This is also arguably a clear example on the ground of the outcomes, in terms of the reorganisation of the landscape, which were a consequence of the processes of submission of landholders of the conquered territories of the Danelaw, which has been examined in detail by Richard Abels (Abels 1988, 89-96), and further examined in the next chapter. The refurbishment of the Burghal Hidage burh at Buckingham and the creation of the new burh at Newport Pagnell resulted in the capitulation of several Danish earls, and were arguably the proximate causes of the capitulation and submission of the Viking army at Bedford in 915. These outcomes have been explored in similar terms by Cyril Hart in his studies of the early development of Northamptonshire and Cambridgeshire (Hart 1970; Hart 1974).

The development of a burh at Newport Pagnell at this time provides an explanation for the shape of the three surrounding hundreds of Bunsty, Moulsoe and Seckloe, of which Newport was the centre, which together form a noticeable salient of the shire to the north of the Ouse and to the east of Watling Street (see figs. 21 & 22). The Chronicle gives the information that as a result of the building of the two burhs, “Earl Thurcel came and accepted him as lord, and so did the earls and the principal men who belonged to Bedford, and also many who belonged to Northampton.” (Whitelock 1979, 213). Bunsty hundred lies to the north of the river Great Ouse, in what would be expected to have been Northampton territory controlled by the Vikings at the time. On this area, Kim Taylor-Moore remarks that it occupies ‘part of the northern Ouse valley, [and] retained strong links to Northamptonshire into the eleventh century, whilst there were no apparent links to Buckingham[shire]. Furthermore . . . the part of the county border separating Bunsty hundred from Northamptonshire was of a different type to that further west. Here the border had the characteristics of a boundary drawn for administrative purposes with little thought paid to existing local links, passing as it did along a watershed and separating vills from their woodlands. It seems, therefore, that Bunsty was once a part of Northamptonshire’ (Taylor-Moore 2012, ch. 4). Moulsoe hundred lay to the south of the Great Ouse and east of the river Ouzel, which had marked the boundary of the territory dependent on Bedford, to which it had reverted after the establishment of Alfred and Guthrum’s boundary in c.880 (Haslam 1997, 123). The military and political

Another burh in this area which was constructed as part of King Edward’s campaigns in this area was at Wigingamere, built in 917, which the writer has argued was located at Linslade, to the east of Wing (Haslam 1997). It was sited on the west bank of the Ouzel at a significant historical boundary between the Wigingas on the west and the Yttingas to the east, which had also been the boundary between Viking-held territory to the east and West Saxon territory to the west, and was near a routeway which was already ancient in the early tenth century (Haslam 1997, 128-9). As such, however, this burh would have consolidated the control of territory already in West Saxon hands, rather than extending it into Viking-held territory. It too would have been allocated a burghal territory, which in this case would probably have comprised some or all of the parochia of the minster at Wing, although this is not shown in any distribution of connections with rural manors. The burh at Towcester on Watling Street, also built in this year, was clearly part of Edward’s forward movement of consolidation to exert control of Danish territories, which was begun in this area by the construction of the two burhs when Edward stayed at Buckingham in 914.

126

Before and After the Burhs of the Burghal Hidage

Towcester lay on Watling Street in the southwestern part of Northamptonshire (fig 20). In spite of the ‘softening-up’ process on the Danish earls of Northampton and Bedford recorded in 914, the army in Northampton, together with reinforcements from points north, reacted strongly, which presumably led to the reinforcement of Towcester by Edward’s army in the autumn of that year. The strategic importance of this burh shows that it must have been given a burghal territory, which appears to have comprised much of the area of later Northamptonshire to the south-west of Watling Street (Hart 1970, 13 and end map; Taylor-Moore 2012). It is argued above that this process would have redefined the military obligations of the territory in this area which had formerly been part of the burghal territory of Buckingham since 878-9, thus setting the stage for its subsequent inclusion in Northamptonshire in the following stage.

tenth century had dependent territories or ‘administrative districts’, in spite of his quoting passages from the Chronicle of the early tenth century which clearly demonstrates the existence of ‘army-districts’ and the like (Taylor 1897, 37-8). He goes on to say that ‘The men [the fyrd] must rally to their own centre, but the land was yet an undivided whole’ (ibid., 49), ignoring the need of the time to define the districts from which men were obligated to ‘rally’ to a particular centre. This particular paradigm might have been consigned to the waste bin of historical studies long ago were it not for the fact that Taylor’s views are quoted on every occasion on which the issue of the origin of the shires is discussed. This particular model has been espoused, for instance, by John Blair in relation to the formation of Oxfordshire, though with some acknowledgement of the significance of earlier developments (Blair 1994, 102-5). It was possibly to fill the yawning gap between the early eleventh century and the time when any pre-Danish arrangements had been ‘wiped away . . . by the sponge of the Danish Conquest’ (Taylor 1897, 36) that Stenton and a number of his followers formulated an alternative view of the origin of especially the west Midland shires as being formed in the early tenth century (Stenton 1971, 337-8). This was based on the argument that only Edward the Elder would have been strong enough to ignore Mercian resentment and to impose a new system from above. In this he is followed by Margaret Gelling and David Hill (Gelling 1992, 141; Hill 2001), and a similar conclusion is drawn in relation to the east Midland shires by Cyril Hart (Hart 1970, 12-14). The espousal of this idea in particular by David Hill is, furthermore, predicated on his hypothesis of the dating of the Burghal Hidage to c.919, and its creation in a Mercian context as a ‘blue-print’ for the underlying arrangements for the shiring of Mercia (Hill 1969, 92; Hill 1996b), an idea taken up by Richard Abels (Abels 1988, 745). I have elsewhere given detailed reasons for regarding this hypothesis as untenable (Haslam 2005, 135-41).

Stage 5 - The ‘shiring’ of the central Thames area – mid tenth century The issue of the origins of the shires of the west, central and east Midlands has exercised the minds of historians for more than a century. It is clear from the foregoing analysis of the organisation of burhs and burghal territories in the late ninth and early tenth centuries that the shires which were in place by the time of Domesday were created by the amalgamation and fission of earlier administrative arrangements which had served somewhat different purposes, and which were still being created into the second decade of the tenth century. This process has been discussed in chapter 8 in relation to the formation of the burghal territories and shires in the west Midlands. In a paradigm of interpretation which has had a long innings, the shires in particular in western and central Mercia are regarded as having originated in the early eleventh century, a view which is based almost entirely on the general incidence of their first mention in documents in this period (Taylor 1897; Whybra 1990, 4-5, 11-12). C S Taylor based his original conclusions both on the absence of any reference to shires as such before the early eleventh century, and on the reference to the amalgamation of Winchcombeshire with Gloucestershire by Ealdorman Eadric Streona in c.1007 (Taylor 1897, 34-5, 39, 41-2), ignoring the fact that this reference marks only a terminus ante quem for the formation of these shires rather than evidence of their creation. In particular, this view does not allow for the fact that already before 1007 the shire of Winchcombeshire must already have had a long existence to account for its being remembered in 1007 with such nostalgia for its passing – a point made by Julian Whybra (Whybra 1990, 12). This process must demonstrate the existence of the shires at, and before, this time.

Stenton, however, was insistent on the idea that the division of Wessex into districts in the Burghal Hidage was a military rather than an administrative expedient, with the implication that this was also the case in later developments in both the west and the east Midlands. He states that ‘The organization which it [the Burghal Hidage] records was brought into being for the sole purpose of providing garrisons for the fortresses. It gives no ground for any theory that the districts assigned to the fortresses were used as administrative as well as military units’ (Stenton 1971, 265 n.2, see also 337). This view appears to be echoed in Steven Bassett’s somewhat more nuanced discussion of the subject (Bassett 1996). It is quite clear, however, from more recent work on several fronts (and summarised in various places in this study), that the burhs and their associated territories were crucial instruments in the royal drive to exert hegemony over territories, from the late eighth to the mid tenth centuries, in matters pertaining to the exercise of lordship by the king. Bassett has rightly emphasised this aspect of the creation of burhs in the late eighth and

Taylor’s position was, however, the almost inevitable result of his belief that before the early eleventh century there were in fact no administrative divisions in Mercia north of the Thames, and that none of the burhs of the early

127

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

early ninth centuries (Bassett 1996, 149-50, 151-6; Bassett 2007). These might be termed ‘territories of obligation’, in that their creation defined obligations of landholders to the central burh in military matters, as well as the exercise of the law, the development of the economy through the establishment of markets and mints, and the development of burhs as ecclesiastical centres. To hold that these functions were not in any way to do with ‘administration’ would be simplistic, to say the least. A similar dismissal of the importance of burghal territories as units of local administration informs Julian Whybra’s conclusion that ‘the west Midland shire system developed as the formal embodiment in 1007 of an informal early tenth-century arrangement of attaching districts to boroughs for military purposes. These ‘districts’, in some instances, may have developed on more ancient, existing Mercian administrative divisions.’ (Whybra 1990, 13-15). This view is reinforced by Bassett’s observations of the development of burghal territories and shires from the earlier provinciae, and his characterisation of these territories and divisions as having military rather than administrative functions (Bassett 1996, 157).

this programme of reorganisation of the west Midlands into shires (e.g. Gelling 1992, 158; Bassett 1996, 155-6), must have occurred after the last date at which Tamworth was functioning as a burh with a dependent territory. Tamworth was seen as a particularly significant political and military centre in both 918, from where Aethelflaed had been able to exercise political domination over Leicester and its territory and where she died in that year, as well as in 926, when King Athelstan married off his sister to King Sihtric of York at Tamworth. It was also clearly a centre of military importance with a dependent burghal territory which were together worth capturing and controlling in c.940, when it was stormed by the Norse forces of Olaf Guthfrithson (Whitelock 1979, 221). By controlling this burh and its territory Olaf would have established a crucial strategic foothold to the south-west of Watling Street, directly confronting Mercian territory of the west Midlands and challenging the forces controlling its burhs. It is quite clear that the shires of Staffordshire and Warwickshire, in their Domesday forms, which resulted in the division of the burh of Tamworth between the two, could not have been formed until after this date.

The currency of these dominant yet competing paradigms has meant that the existence and importance of the series of interlocking burghal territories as crucial administrative arrangements which lasted well into the middle years of the tenth century, but which preceded the developed shires of Mercia seen in Domesday Book, has been somewhat overlooked. As already pointed out, they were very much more than ‘informal’ arrangements, and were, furthermore, real instruments through which the royal administration of the day was implemented, largely through the obligations of renders of service and provisions, and the imposition of royal lordship, which were imposed on the populace and which were channelled through them. It is one of the purposes of this study to emphasise and to develop the concept of the significance of the burghal territories as providing a natural and logical stage which ties the disparate strands of evidence for the development of all these aspects into a coherent and rational developmental historical model.

Secondly, as is clear from the extended account in the A version of the Chronicle, Edward the Elder was throughout the later part of his life engaged in, and apparently consumed by, the conquest of the Scandinavian-held territories and the creation of burhs and burghal territories, through which he exerted control of the reconquered kingdoms, a policy (and practice) which his sister Aethelflaed appears to have mirrored in the west Midlands at the same time (Wainwright 1975; Stafford 2008). The arguments put forward to the effect that the annals relating to King Edward’s exploits in this section of the A version of the Chronicle were composed by his brother, who accompanied him while on campaign (Pelteret 2009), lends this section a degree of historical veracity as representing the King’s own perception of the significance of the events described, and thus of his general strategies. These burhs which were created as a result of this strategy from 911 onwards comprised both military and political instruments which enabled him and his sister Aethelflaed to enforce the submissions of the populations of the burghal territories to their overlordship, and thereby extend and consolidate royal control over former Scandinavian territories. The creation of shires as larger units of administration which subsumed the burghal territories would have run entirely counter to these strategies, which up to the end of Edward’s life had been the mainstay of his political and military agendas. As is remarked above, it is a feature of previous discussions on the origin of the Midlands shires, both east and west, that the existence and importance of the burghal territories as instruments of local administration has been generally unrecognised. The role of these territories and of the burhs which they were set up to support is for instance entirely overlooked in Hart’s analysis of the development of Northamptonshire, which he sees as the creation in its entirety of Edward the Elder in c.917. That this subsumed the burghal territories both of Towcester in the south-west

With this in mind, there are a number of reasons for suggesting that a significant phase of the formation of the shires as they appear in Domesday Book can be most satisfactorily placed in the middle or third quarter of the tenth century, rather than the early tenth or the early eleventh, as is commonly assumed. This view recognises (as do those of Bassett) that these new shires were formed by fission and fusion of blocks of territory inherited from earlier arrangements. Quite apart from any supposed but unsustainable association of the formation of shires in western Mercia with the context of the formation of the Burghal Hidage document argued by David Hill, there are three main arguments which tell against the particular model of the formation of the shires in the early tenth century. Firstly, the division of Tamworth between Warwickshire and Staffordshire, which represented part of

128

Before and After the Burhs of the Burghal Hidage

(which had itself arguably been carved out of the initial burghal territory of Buckingham lying to the south-west of Watling Street) and that of the southern burh at Stamford in the north does not play much of a part in the development of Hart’s historical model (Hart 1970, 11-14).

which would have been continued by Edgar after 959, and implemented by the powerful Ealdorman Aelfhere, who became ealdorman of the whole of Mercia from c.960 (Williams 1982). It is possible – though this cannot be pursued further here - that this process of reorganisation of the western Mercian shires was a development of a similar process which had perhaps been started in the area of the Five boroughs, after their capture by King Edmund in 944.

Thirdly, it is clear from the laws issued by Edward’s successor King Athelstan (924-39) that the main thrust of his political and military policies was to strengthen and consolidate the defensive as well as the administrative and economic functions of the burhs and the burghal systems which had been created in previous decades, and by implication to maintain the contributory functions of the burghal territories, rather than setting up a new system which would essentially have superceded them. Cap. 13 in Athelstan’s Grately code of c.926 states that “every borough [port] is to be repaired by a fortnight after Rogation days”, and, similarly, trading is to be confined to a port (repeating an ordinance of Edward the Elder) in cap. 12 (Whitelock 1979, 419). A similar concern is also shown in cap. 14 for the provision of mints within some of the burhs [burgum] of southern Wessex, all of which, with the exception of Dorchester, Dorset, were included in the List given in the Burghal Hidage (Blackburn 1996). Blackburn has, furthermore, suggested that cap. 14 may well be an interpolation from an earlier administrative document (Ibid. 172). The presence of these laws originating from Athelstan’s reign is thus indicative of the continuation of a policy which preserved practices from the previous reign relating to the administrative, logistic and defensive roles both of the burhs and of their dependent territories. In this regard it is worth pointing out that David Hill’s hypothesis of the formation of an updated series of burhs around Wessex and Mercia by Athelstan (Hill 2001) runs entirely counter to his own hypothesis of the shiring of Mercia in the reign of the previous king, Edward the Elder. The extension of the burghal system, and the consequent formation and re-formation of the dependent ‘territories of obligation’ within this developed system which this would have entailed, would have represented processes which must have come before the reorganisation of Mercia into shires, as is shown graphically in the case of the division of Tamworth.

The ascription of this process to this time would also fit the wider context. The peaceful conditions in the third quarter of the tenth century would have been more conducive to what amounted to a wholesale administrative reorganisation of earlier arrangements which had revolved around the formation of burhs and their dependent territories. As I have argued elsewhere, the burghal system, together with the territories which supported it - both in the west and central Midlands as well as Wessex – had probably become obsolete by the 960s (Haslam 2011b, 208). Richard Abels has pointed out that in an era of peace, their defences would have become too expensive to maintain in terms of the provision of manpower and resources by the state. He has argued that the burghal systems of the late ninth and early tenth centuries were ‘gradually replaced by one less costly and better suited for peacetime’, and that these systems were made redundant through their own success. Town defences were eroded as a result of the replacement of permanent garrisons by ad hoc levies, and by the formation of private armies by lay and ecclesiastical magnates in the reign of Edgar (959-75), so that by the 980s ‘the very memory of Alfred’s burghal system had been forgotten’ (Abels 1996; Abels 2001, 21-3, 30). Developments such as the formation of ship sokes in the 960s would also have diminished the administrative importance of the burhs, and would probably have diverted manpower resources away from their continued upkeep. Archaeological evidence shows that many of these defences appear to have been abandoned and/or dismantled, and in some cases had begun to collapse. This phase of abandonment is clearly seen, for instance, in the evidence from the defences of Cricklade (Haslam 1984a, 106-11, 137; Haslam 2003, period 2 B), as well as in several cases in the west Midlands (Haslam 2011b, 211-5). The territory-based obligations for the upkeep of these burhs of earlier generations, which had provided the initial impetus for their construction and which had sustained their maintenance well into the tenth century (as is shown in cap. 13 of Athelstan’s Grately Code, quoted above), were no longer relevant to the administration of wider areas, except insofar as these earlier territories in part formed the basis for the amalgamations into larger units at this time.

These arguments show quite clearly, therefore, that the shiring of Mercia must have occurred after the end of Athelstan’s reign (d. 939). The situation is complicated, but from an evidential point of view perhaps clarified, by the subsequent division of Wessex and Mercia as separate kingdoms in the reigns of Eadwig (king in Wessex, 957-9), and Edgar (king in Mercia at the same time, and after 959 of a reunited kingdom). It is at this juncture that the boundary between the new Wessex and the new Mercia is most likely to have been redrawn along the Thames, thus effectively realigning areas on both sides of the river dependent upon Wallingford and Oxford, and thereby creating the new shires of Berkshire (as part of Wessex) and Oxfordshire (as part of Mercia). The shiring of both western and central Mercia could, therefore, be reasonably seen as a process

This time, therefore, could be seen as providing the most appropriate opportunity for the reorganisation of these earlier burghal territories into larger units based on the more important burhs of an earlier generation, in which considerations of the needs of the defensive capability of sites within a mutually-supporting system or series was less important than the demands of royal administration and

129

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

the territorial basis of the judicial system. In this process it would, however, be difficult to determine whether the abandonment of the burghal systems was the cause or the consequence of this development. John Blair and others have preferred to see this process as being the result of a reorganisation of the territorial landscape in the early eleventh century, in which the administrative and legal functions of several former burghal territories were consolidated into one place by a process of restructuring and amalgamation, though he recognises that the foundations for this restructuring, in terms of the organisation of hundreds, was part of an evolving process which started rather earlier (Blair 1994, 102-5). In relation to the area of the middle Thames, Katharine Keats-Rohan has more recently suggested that this time – the 960s - was marked by a period of ‘major reorganisation’ (Keats-Rohan 2009, 62).

for the 698 hides of southern Oxfordshire (comprising part of the burghal territory of Wallingford), which process to all intents and purposes would have maintained the figure of Oxfordshire’s 2400 hides. It is of some significance for the reconstruction of Oxford’s original territory of 2400 hides in the Burghal Hidage that the County Hidage gives this figure as the hidage for early eleventh-century Oxfordshire (Maitland 1897, 455-60). (There are unfortunately no figures for Berkshire or Buckinghamshire). This process would also have involved the absorption of Reading’s burghal territory into the administrative area of Wallingford to create Domesday Berkshire, just as Aylesbury’s burghal territory was absorbed by the shire now centred at Buckingham, less the territory to the south-west of Watling Street which had been absorbed by the burghal territory of Towcester, to create Domesday Buckinghamshire. The awkward shapes of all three shires is indicative of the threeway amalgamation of somewhat smaller territories which had served rather different purposes in the past.

It seems plausible therefore that one of the proximate causes of this restructuring, at least in the upper Thames area, could have been the temporary separation of Wessex and Mercia as separate political entities during the period 957-9, and the enhancement of a sense of Mercian identity on the appointment of Aelfhere as Earldorman of Mercia (Williams 1982; Stafford 1985, 122-3; Blair 1994, 102). In this process the Thames would have been perceived as a natural administrative boundary of demarcation rather than a means of communication which was ignored by boundaries of an earlier age. This seems to be supported by the existence of separate reeves for Oxfordshire and Buckinghamshire which are mentioned in a charter of 995 (S.883; quoted in Blair 1994, 103), somewhat before the time of the assumed formation of the shires in the early eleventh century. The formation of the new shires at this time also appears to be consistent with evidence from a charter of 969 for the existence of boundary stones marking the meeting place of several shires (S.1325; Hooke 1985, 62; Hooke 1996, 100; Whybra 1990, 89 n.82) – though the topography of this presents some problems which require further elucidation. This process can in general terms be recognised as arising from the development in King Edgar’s reign of the concept of the ‘kingdom of the English’ as an enlargement of the ‘kingdom of the Anglo-Saxons’ (Keynes 1998, 39).

A similar instance of this wholesale amalgamation and re-formation of burghal territories to form shires can be recognised in the west Midlands. It has already been argued in chapter 5 that the extension of the old ‘territory of obligation’ centred on Warwick was formed into the new shire of Warwickshire by amalgamation of the former burghal territory of Warwick to its south and part of the former – and now redundant - burghal territory of Tamworth to its north. This appears to have led to the creation of new tenements in Warwick which were appurtenant to estates in this newly-added territory to the north, a good proportion of which connections were preserved by the fact that the estates were given by the earl to Coventry Abbey in the early eleventh century. This suggests that connections of estates with the central burh or borough in the new shire were still seen as important in its administrative functionality at this time. In the case of Warwick the connections of these estates to Warwick in the added territory may have been created anew, or, more likely, merely transferred from Tamworth. Being now divided between Staffordshire and Warwickshire, Tamworth was now no longer a centre with a dependent territory, and therefore clearly had no administrative role of any importance in the new arrangements. It has already been pointed out that the dual connections shown by other estates in Worcester and Gloucestershire can be explained by the addition of a connection of an estate to a new centre with that to an older one, which is quite consistent with the proposition that the connection of estates in north Warwickshire could be changed from Tamworth to Warwick.

As can be recognised from the previous discussion, however, the shires which did develop were formed as a result of the rearrangements of earlier burghal territories. This pattern seems to be clearly evidenced in the west Midlands (discussed in chapter 8). It seems most likely that in this process the part of Wallingford’s territory lying to the east of the Thames became included in the new Oxfordshire, and the southern part of Oxfordshire’s territory on the south side of the Thames in the vale of the White Horse was included in the new Berkshire centred on Wallingford, with Oxford and Wallingford, on opposite sides of the Thames, maintained as the administrative centres of the new shires. It would appear to be significant that in this process the 734 hides of northern Berkshire (comprising part of the burghal territory of Oxford) would have been exchanged

The same process can be recognised in a few of the connections with both Wallingford and Oxford. The connection of the royal estate at Bray [8], on the far eastern end of Berkshire, must have been formed with Wallingford after the burghal territories of Sashes, and then Reading, had been absorbed by the new shire based on Wallingford. It is possible that this estate had originally been contributory to Sashes, although its nature as a fortress which was not

130

Before and After the Burhs of the Burghal Hidage

permanently occupied would make this unlikely. There are however other alternatives – the connection could have been formed a) originally with Reading, on the occasion of its foundation as a burh; b) with Wallingford, on the occasion of the formation of the new shire; or c) with Wallingford in the early eleventh century, perhaps on the occasion of the creation of the honour of Wallingford, discussed below. This range of possibilities, though, emphasises the importance of particular episodes of administrative reorganisation in the creation of shires from ‘territories of obligation’, in which these connections between rural estates and the centres are argued here as having been formed. It is possible, too, that the dual connections of Pyrton with both Oxford and Wallingford (above) could have been established at this period – a primary connection with the burh at Wallingford being augmented with a secondary connection with the new shire town at Oxford. However, the central position of the tenements in Oxford which were appurtenant to Pyrton, their inferred topographical primacy in being sited on the middle Saxon routeway leading northwards from the old ford, and the antiquity and importance of the Pyrton estate pointed out above, are all aspects which would suggest that the reverse is more likely to have been the case.

in the late tenth and early eleventh centuries. It is not possible to recognise particular instances of alterations to the putative mid-tenth century shires, as can be reasonably suggested for Gloucestershire and Worcestershire, above. It is quite possible, however, that some of the connections of estates with tenements in the principle burhs of the lateninth century could have been formed at this period. The connection of Wallingford with the contributing estate at Bray is one of the more obvious examples, though this may equally have been formed in the mid-tenth century, as suggested above. In other cases this possibility can only be conjectural. There is, however, considerable archaeological evidence that former burghal defences in both Wessex and Mercia were refurbished at this time, a topic which has been discussed in detail by the writer (Haslam 2011b, 204-17). There appears to be some archaeological evidence to support the hypothesis that Oxford itself was redefended at some point in the years around 1000 with a new stone wall which replaced the probably ruined remains of an earlier wall of the late ninth century (which itself had been added to a timber-revetted bank in stage 2, above) (Haslam 2009, 98-100, 103-4; Haslam 2011b, 211-2). This reflects a wider concern with the defence of the realm against the new Viking onslaughts against the Anglo-Saxon state at the time, and appears to reflect the importance of Oxford as a new regional centre of both defence and administration at this period (Blair 1994, 159-70). This is also evidenced in the development of new defended sites at hillforts such as South Cadbury and Daws Castle near Watchet (Alcock 1995, 154-70; Haslam 2011b, 204-8).

It would seem most likely that the northern parts of both Oxfordshire and Buckinghamshire were reorganised at this time. As has been argued above, the original burghal territory of Buckingham most probably extended northwards from Buckingham itself to encompass the swathe of territory comprising the south-west of later Northamptonshire to the southwest of Watling Street (fig. 20). This area was then absorbed into the burghal territory of Towcester in c.917 (stage 4 in the timeline above). This development would provide the natural precursor to the development of the new shire based on Northampton, subsequent to the cessation of the local defensive functions of the burh of Towcester, here argued in the years around 960. In a similar way, the new shire of Northamptonshire would have absorbed the burghal territory of the southern burh at Stamford (Hart 1970, 13 & n.3), though this development is beyond the reach of this discussion. The development of the shire of Northamptonshire seems likely to have encroached on the original burghal territory of Oxford, dividing the old regio dependent on King’s Sutton, formerly for the most part in Oxford’s territory, between the new Oxfordshire and Northamptonshire. In this process, the delineation of the new shire of Northamptonshire appears to have left to Buckinghamshire the hundred of Stotfold, which lies to the north of the Great Ouse, as comprising the hundred in which Buckingham itself was situated. It also left the bishop of Lincoln’s manors around Banbury, which at this time were probably constituted into a formal hundred, to Oxford, to which they had had a primary attachment.

I have also argued that this period was marked by a renewed interest in the Burghal Hidage document as an exemplar of the kind of arrangements in which burhs which were supported by territories, which were perceived to have been so successful in countering Viking attacks more than a century earlier. I have suggested, for instance, that the Calculation appended to the List of burhs in the Burghal Hidage, which appears in a document of the early eleventh century, was first composed at the time of this revival of interest in earlier arrangements. A plausible context for its compilation would be to see it as an attempt to re-establish the extent of the earlier burghal ‘territories of obligation’, by calculating the numbers of hides which would have been required from lengths of burghal defences in the main List which already existed and which could be measured on the ground (Haslam 2011b, 217-21). The chances are, however, that this would have been more confusing than useful in the cases of the burhs in three shires of Berkshire, Oxfordshire and Buckinghamshire, where the earlier arrangements had been altered out of all recognition in the 120 years or more since King Alfred’s day.

Stage 6 – the early eleventh century

It is this period, too, which could be seen as the most appropriate historical context for the formation of the Honour of Wallingford in the form in which it can be recognised in the eleventh century, particularly at the

As suggested above, the shires in their later form are likely to have been already established by the time of the developments which can be attributed to King Aethelred

131

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

time of Domesday. In describing this Honour, Katharine Keats-Rohan has suggested that it was ‘a deliberatelycreated Anglo-Danish ‘castelry’ placed under the command of trustworthy kinsmen, possibly one that formalised an existing arrangement, originally based on defence and

communication, that went back to the founding of the burh’ (Keats-Rohan 2009, 62). As such it appears to have incorporated lands on both sides of the Thames, reflecting not only the original arrangements of the late ninth century, but also the new arrangements of the mid tenth.

132

Chapter 14 The Function of the Burghal System of the Burghal Hidage If, as is argued in this study, all the burhs in Wessex – including Buckingham and Oxford – are to be conceived as a contemporary and unitary system, their purpose as such a system is likely to have been greater and more complex than the sum of the functions of all its components. It would otherwise be necessary to fall back on the position favoured by Baker and Brookes, to the effect that individual burhs were constructed at different times to meet different and perhaps disparate strategic situations, or were individually built, one by one, with hindsight to ‘defend’ a particular area which was shown to have been vulnerable to particular hostile attentions by the Vikings (Baker and Brookes 2011, 112; see also Lavelle 2011, 234). It is certainly true that individual burhs were built to guard or defend particular nodal points in the landscape, such as the crossings of rivers by significant Roman or other routeways, or the locales of important royal estate centres, or more complex systems of routeways and look-out points (ibid., 109-11), or access points on major estuaries on the south coast. I have indeed analysed some of these factors which governed their siting (Haslam 2005, 129-33). The association of burhs with beacon systems, roads and bridges is also an important aspect of their functionality (Lavelle 2010, 217; Baker and Brookes 2011, 110;

observation by Richard Abels that the burhs performed an essential aspect of the support of the mobile fyrd or field army, another of King Alfred’s innovations at the time (Abels 1989, 63-8). Important observations concerning the relationship between the three common obligations and the defensive functions of a burghal system such as that in Wessex have been made by Gareth Williams (Williams 2005, 105) It is suggested, however, that to concentrate on these aspects of internal defence and functionality and the desire to address local strategic needs actually misses the most crucial aspect of the creation of the 31 burhs around Wessex (including the contemporary burhs at Oxford and Buckingham) as a system. It is argued here that one of the most important reasons for the construction of this system at one point in time was the political need to enhance the control of the territory of the kingdom by the king, through the enforcement of lordship bonds of all landholders to himself, in ways which have already been analysed in chapters 8 and 9 above. As the annals of the A version of the Chronicle make clear, this policy is reflected so clearly in the staged extension of the burghal system by Edward the Elder over the Danelaw from 911. This could certainly not have been achieved by the piecemeal construction of the various burhs at different times around Wessex to meet local threats, as suggested by Baker and Brookes. Bacharach and Aris, in characterising this process of the construction of the burghal system in Wessex as the allocation of ‘incomeproducing landed resources’ from the burghal territories to the support of the burhs, have also missed this crucial point (Bacharach and Aris 1990, 2). By characterising the landed resources as producing ‘income’, presumably in monetary form, they have also missed the real nature of the support system as obligatory renders supplied by the estates within the burghal territory, as discussed further in chapter 9 above. Whatever the defensive strategies, the deployment of these resources, and the administration of the human conscription through the obligations due from the territories of each burh, must have been an important and logistically-complex enterprise which cannot have been achieved without an underlying political rationale.

Some of the burhs were also clearly offensive in intent, as directly challenging the Viking presence in adjacent territories. Examples of such are Southwark, challenging the Vikings who were in control of London (Haslam 2010b, 130-7); Cricklade, which confronted Guthrum’s Viking army at Cirencester just a short march up Roman Ermin Street; and Buckingham itself, which directly challenged the Vikings in control of Northamptonshire to its north and north-east (above), as well as the Viking control of Watling Street (Haslam 2005, 130-3; Haslam 2011a, 125-7). Much has also been made of the fact that these burhs were spaced more or less evenly around the whole of Wessex in such a way as to be reachable by most of the population in not more than a day’s journey (Tait 1936, 22-3; Stenton 1971, 264; Wormald 1991, 152-3; Hall 2011, 606), and as reflecting a strategy of ‘defence in depth’ in which garrisons in individual burhs could support each other in the event of attack (Abels 1989, 74; Bacharach and Aris 1990, 2-5; Lavelle 2003, 16, 29-32; Lavelle 2010, 209-10; Baker and Brookes 2011, 105). An additional function of the burghal system as a unitary whole can be recognised in the

Reasons have been given in chapters 8 and 9 for regarding the connections between urban tenements and rural estates, as recorded in pre-Conquest charters and in the folios of

133

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Domesday Book, as the remnants of a once-universal arrangement whereby all the landholders within the various territories which were set out as providing for the upkeep of the burhs were to owe service to that burh for its physical construction, maintenance and upkeep as a sustainable community. It is argued that this was achieved by the assignment by the king of a tenement or haga within the burh to be held as part of the estate of every landholder above a certain status, as an integral part of the way the burh was set up from the beginning. By this means the king established and reinforced reciprocal obligations on the part of landholders for service to the king or his reeves at the particular burghal centres in whose territories they held their estates. This is seen as an extension of an arrangement which had been common practice from much earlier times, in the need to enforce the performance of services at royal estate centres. This must have been not only one of the most important instruments which the king had in his power to ensure the stability of his new burhs as sustainable institutions; it would also have led inevitably to the enhancement of the bonds of lordship of every landholder to the king, and thus to the internal cohesion of the ‘state’, in the person of the king, in its social, economic and military aspects. It is this aspect which arguably underpins the whole enterprise, in the sense of providing a fundamental political and therefore functional rationale for the construction of the system as a whole.

and the Oxford region, and the minting activities which were the outcome and the sign of the exercise of this power and control. It was as a result of his decisive victory at Edington against Guthrum’s Viking forces in May 878 that King Alfred was presented with the opportunity to exert for the first time the degree of power which enabled him to put in place the system of burhs all over Wessex, through which he effectively compelled the submission of the whole population to himself as lord. As already argued, the inclusion in this system of the area of later Oxfordshire and Buckinghamshire to the north of the Thames was a device which also enforced the submission of the population of the area to himself, thereby ensuring the support and loyalty of estate-holders in an area surrounded to the west, north and east by Viking-controlled lands. As I have argued elsewhere, this policy of the building of burhs to enforce submission of populations was also extended to London immediately after the retreat of the Vikings to a new state in East Anglia in late 879, a forerunner of the general submission to himself recorded in the Chronicle in 886 (Haslam 2010b, Haslam 2010c; Haslam 2011a). As argued in chapters 8 and 9 above, this policy of the creation of burhs and their ‘territories of obligation’ can also be seen as underpinning King Alfred’s extension of his control of western Mercia under a new polity which contemporaries called the ‘kingdom of the Anglo-Saxons’, which represented a new political order which subsumed the identities of the people of Wessex and of Mercia as separate kingdoms (Keynes 1998, 24-6, 34-9, 43-4; Keynes 2001, 44-8). There is reason to hold that Alfred’s overall control of this new polity was achieved by means of the construction of new burhs, or reconstruction of earlier burhs, and their territories at Gloucester, Worcester, Winchcombe, Tamworth and Hereford in the west, and of London in the east, in c.880 (but as a strategic enterprise possibly begun in late 879), which would have extended the range and the political force of the system in Wessex built only a little while earlier (Haslam 2011a, 133-5). As in the cases of Wessex and central Mercia already discussed, the establishment of the burhs and their burghal territories in western Mercia must have been the essential means by which Alfred was able to extend his control and establish the allegiance of the populations to his lordship. As already discussed (chapter 8), this system overlay, and was based upon, an earlier system of burhs and their dependent territories which had been put in place in the late eighth or early ninth century by one of the kings of Mercia, undoubtedly for the same reasons (Bassett 2007; Bassett 2011). The basic conclusion from the distribution of the connections of rural estates to these centres which has already been examined is that these are the surviving evidences of the operation of this process, both in Mercia and in Wessex.

The creation of the system of the burhs of the Burghal Hidage all over Wessex, from Sussex in the east to Devon in the west, can in this light be seen as a deliberate and well-thought out strategy aimed at both guaranteeing the loyalty of all landholders and their men to the king in person, as well as ensuring that the universal obligations for service on the king’s works could be effectively enforced, and channeled towards measures which addressed the exigencies of the time. As a political tool of considerable sophistication and boldness, it combined the political needs of the consolidation of the powers of the ‘state’, in the person of the king, as well as the pressing strategic needs of the moment to combat external military pressures from the Vikings. The deployment of this grand strategy at the particular juncture argued above (in the period 878-9) can be seen as an urgent response to recent military and political developments. Through the Vikings’ enforced partition of eastern Mercia in 877, King Alfred had arguably lost control of London to his rival Ceolwulf, who was supported by the Vikings. And the Viking near-miss coup against Alfred at Chippenham in early 878 had resulted in the control and submission to the Vikings of a part of northern Wessex (of undefinable extent, but perhaps covering northern Wiltshire and Somerset), in which process the earldorman of Wiltshire, together with probably as much of the population in this area who wished to keep their lands, had had to submit to the Viking controllers as their lord (Anglo-Saxon Chronicle, A version, sa 878; Darlington 1955a, 6-7). Alfred had also lost his ally King Burgred of Mercia, whose cooperation must to a large extent have facilitated Alfred’s former exercise of hegemony in London

The attempts by various commentators to place the development of burhs at Gloucester, Oxford and Worcester

134

The Function of the Burghal System of the Burghal Hidage

by Ealdorman Aethelred in the late 890s in a Mercian context (above; and Blair 1994, 92-105; Baker and Brookes 2011, 110-1) are in this sense self-contradictory. As Simon Keynes has so graphically shown, Aethelred (together with his wife Aethelflaed after c.886), was Alfred’s protégé, acting thereby as only the sub-regulus of Mercia, and was decidedly under the political control of King Alfred (Keynes 1998, 19-34). Any burh and its associated territory which would have been brought into being by Aethelred, or which would have been nominally controlled by him, would therefore have required the submission of the population to King Alfred, not to himself. This is indeed the implication of the twice-repeated phrase in the charter detailing arrangements for the (re)fortification of Worcester in probably the early 890s, that the grant was made ‘in the witness of King Alfred and of all the councillors of the Mercians’ (Whitelock 1979, 540-1). Very similar phraseology is used in the known charters of Ealdorman Aethelred in the 880s (Keynes 1998, 20-1, 27-8). This political device of the consolidation and extension of King Alfred’s lordship clearly underlay the submission in 886 of ‘all the English people that were not under subjection to the Danes’ to King Alfred, as stated in the Chronicle (Whitelock 1979, 199). The notion espoused by Baker and Brookes of both Oxford and London being under ‘Mercian authority’ in the late ninth and early tenth centuries, and having ‘changed hands at least once’ during the period c.880 and c.910 thereby constituting one reason for excluding Oxford from a West Saxon system of burhs of c.879 (Baker and Brookes 2011, 110-1), is meaningless, since the hands it supposedly ‘changed’ between were those of a single unified political body which had King Alfred as its nominal – and very effective - head. This is amply demonstrated in the conclusion, examined above, that the primary burghal territories of Oxford, Wallingford and Sashes straddled the Thames, as well as by the submission of all the non-Danish population to King Alfred in 886.

a new appreciation of the significance of the Burghal Hidage as a document. As I have discussed elsewhere, and elabourated on above (chapters 8, 9 and 12), this was arguably produced at one moment in time during these crucial and seemingly fast-moving developments of the later ninth century. I have made a case that the most plausible strategic circumstance in which all the burhs and their territories (taken as a system) would have been created would best fit the historical context of the period between the defeat of Guthrum’s Vikings by King Alfred’s forces at Edington in May 878 and their final retreat to East Anglia in late 879 (Haslam 2005; Haslam 2011a). This is supported by the independent evidence from the development of the coinage of the period analysed by Mark Blackburn (Blackburn 2003). I have also argued that the internal characteristics of the document itself show that it is likely to have been produced at the beginning of a long process of development of the system which it sets out, in which the smaller burhs which were less viable as settlements were replaced by larger burhs on new sites (Haslam 2005, 129, 137 & n.103; Haslam 2009, 103-4). This is a quite different proposition to that put forward recently by Baker and Brookes, to the effect that the Burghal Hidage itself is the record of a system which was the end product of a long process of internal development which affected the burhs given in the List, in Wessex as a whole (Baker and Brookes 2011, 111-13). They have presented arguments to the effect that the burhs along the Thames, including Cricklade, Wallingford and Sashes (but excluding Oxford – and of course Buckingham) were a secondary development to forts such as Bath, Malmesbury and Chisbury, and that this arrangement was itself superceded in the second decade of the tenth century by burhs which acted as defensive bulwarks in the area to the north of the upper Thames, such as Towcester and Buckingham. The Burghal Hidage document is thus seen as a description of the last stages of a completed system of which many elements in the heartland of Wessex may already have been anachronistic (ibid., 112).

As Richard Abels has so clearly shown, and as I have argued elsewhere, the creation of burhs and their ‘territories of obligation’, and the expansion and consolidation of the personal allegiance of the estate holders within them to the king which this entailed, were also crucial aspects of the subsequent annexation of territories in the early tenth century by Alfred’s son Edward the Elder and daughter Aetheflaed in eastern and western Mercia respectively (Abels 1988, 82-96; Haslam 1997, 114-8). This is emphasised by the nature of this process as a rolling campaign in which this submission was achieved territory by territory over a period of nearly 20 years, starting with the construction of a secondary burh at Oxford in 911 (Haslam 1997; Haslam 2010a). In this respect, as Simon Keynes has so clearly shown, the new kingdom forged by these means, which led to the extension of Edward the Elder’s lordship over Scandinavian territory, was essentially a continuation of the very successful policy put in place by King Alfred (Keynes 2001).

Many of the arguments which are called in support of this model can be fundamentally challenged. This will be discussed more fully elsewhere. Baker and Brookes argue, for instance, that the secondary status of Cricklade on the Thames is suggested by the proposition that its hidage, as recorded in the Burghal Hidage, must be an addition to a sum represented by the hidages in the rest of Wiltshire for Wilton, Malmesbury and Chisbury (ibid., 108-9), which together approximate to the hidages recorded in Domesday. As is shown above (chapter 3), this proposition is flawed; the original hidages of the burghal territories of all four burhs in the shire reflect very closely the total given in the Burghal Hidage, which shows that they must have been contemporary. Baker and Brookes’ reconstruction of the Wiltshire burghal territories also ignores the orbits of connections of contributory manors to the burhs, which they have suggested themselves may originate in the defensive

These strategic and other considerations also make possible

135

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

arrangement of the late ninth / early tenth century (Ibid., 106).

of the initial system in Wessex in ways which utilised established methods to meet the new strategic and political developments of the later ninth and early tenth centuries, in both Wessex and former Mercia. The analysis of Baker and Brookes, in attributing the foundation of the burhs of the Burghal Hidage and their territories to differing circumstances within a broad time scale, and in ignoring the political and strategic circumstances which prevailed to the north of the Thames between the 870s and c.911, has missed the significance and importance of any of these developments within the context of any credible historical narrative of King Alfred’s reign. In particular, Alfred’s political and strategic interest in London - from the mid 870s as a minting place (by the evidence of the coinage), from at least as early as 886 as a defended burh (by the witness of the Anglo-Saxon Chronicle), and from c.880 by witness of the archaeological evidence (Haslam 2010b; Haslam 2011a) - is written out of the model put forward by Baker and Brookes as though these developments had just not happened. Similarly, the internal developments of the Burghal Hidage system, shown for instance by the replacement of Pilton and Halwell by Barnstaple and Totnes respectively, and which can be seen as elements of a secondary development in a way which is paralleled in other sites all over Wessex, are ignored, as is the context of the major logistical programme of the replacement of turf and timber defences with stone walls in a wide range of sites.

Most importantly, however, Baker and Brookes’ analysis ignores the fundamental role of the burghal territories in defining the obligations of conscripted men and their families for the logistics of burghal construction and maintenance. The fact that these territories were intermeshed in a complete network around the whole of Wessex –– in ways which imply the existence of all as elements in a contemporary network of such territories (as is demonstrated for Wiltshire in chapter 3, Hampshire in chapter 4, as well as for Oxfordshire, Berkshire and Buckingham in chapters 10 and 11) - is one of the principal evidential bases for the conclusion that the burhs listed in the Burghal Hidage were contemporary. As already discussed in chapter 12, the clear historical and numismatic evidence for King Alfred’s control of the central lands of (former) Mercia north of the Thames from the early 870s, and of the London area from the c.879, when Guthrum’s Vikings settled in East Anglia, demonstrates that the existence of the so-called ‘Thames frontier’, which forms a key part of Baker and Brookes’ model (Ibid., 109-11), is, quite simply, unsustainable. Not only do the burghal territories of Oxford, Wallingford and Sashes straddle the Thames; but Oxford itself cannot ever have been a ‘Mercian’ burh, even though under the nominal control of Ealdorman Ethelred, because, as already pointed out above, the submission of its inhabitants and of the population of its burghal territory must always have been to King Alfred himself.

The conclusions in this study give a new dimension to the significance of the Burghal Hidage as a political statement. David Hill has suggested that ‘the document may not have arisen for any other purpose than as a stock-taking of the situation and not linked to a particular crisis or decision’ (Hill 1996b, 96). It must be said, however, that all the considerations discussed above combine to indicate that the contrary was the case. This was certainly not the view of the early eleventh-century compiler of the collection of documents known as BL, Cotton MS. Otho B. xi of c.1006, in which the Burghal Hidage and its associated Calculation were included as important administrative documents in a collection which was ‘brought together to foster and celebrate the traditions of the English people’ (Wormald 1996, 64; see further discussion in Haslam 2011b, 217-21). Just as the system of burhs itself was the primary instrument through which the king was able to exert control of the populations of ‘territories of obligation’, so can the production and ‘publication’ of the document be seen as a political statement of the first order, whose purpose was to announce, and thereby consolidate and substantiate, King Alfred’s hegemony and control over the whole of the territory to which he laid claim, over against the territories which at the time were in the hands of competing if not actually hostile polities. This is particularly appropriate to the area represented by the burhs of Oxford and Buckingham, the control of whose territories would have formed an essential strategic platform for Alfred’s subsequent assumption of control of London and western Mercia on both sides of Watling Street from

While it is becoming increasingly clear that many burhs of various sizes and types, some on new sites but most utilising already-established royal or other sites, were built or utilised in Wessex in the century or more before the late ninth century (Brooks 1996b, 129; Reynolds 2003, 130; Draper 2008; Draper 2011; Baker and Brookes 2011, 106-8; Haslam 2011b, 199-204; Yorke forthcoming), and of course from possibly an earlier date in Mercia (Bassett 2007; Bassett 2011), the main thrust of the arguments which I have presented in this study is indicative of a system which was imposed in Wessex on these earlier arrangements as a new overaching system of strategic and political control which, though it incorporated earlier elements, was entirely new at the time. This was followed by a process of expansion and internal reorganisation reflecting similar political and strategic agendas, which was arguably carried forward, firstly by the inclusion of new burhs at London and several in western Mercia in c.880 (Haslam 2010b; Haslam 2011a, 131-5; and further in chapter 8), and secondly by the strengthening and internal expansion of this new system in the early 890s in response to the new wave of Viking incursions at the time (see chapter 11). This was enlarged a generation later by the addition of new burhs over the rest of the Danelaw by King Alfred’s successor Edward the Elder, and by Aethelflaed in western Mercia (e.g. Haslam 2010a). All of these can be seen as the development and extension

136

The Function of the Burghal System of the Burghal Hidage

late 879, when Guthrum’s Viking forces were forced to retreat to East Anglia. In this sense the original Burghal Hidage document – perhaps ‘circulated’ by the king to all the earldormen and reeves of the shires and territories – could be seen as a powerful tool of political propaganda which would have been perceived at the time as being equal in force to the minting of the London Monograph coins, which were designed to celebrate and substantiate King Alfred’s newly-won control of London and its territory, and perhaps of Mercia as a whole, in c.880 (Blackburn 1998, 110-12; Keynes 1998, 31; Blackburn 2003, 205-7; Haslam 2011a, 130).

assessment at the time, or an old-established assessment revived, to provide a logistical support network for the manpower conscripted through obligations which were expressed in hides. This is discussed further in chapters 3 and 9 above. In Wessex, these burghal territories can be seen to interlock in space, and therefore in time, in such a way as to demonstrate the function of the burhs and their territories as both strategic and political instruments, established at the same time as a unitary system. One of the main purposes of this was to enhance and consolidate the control of territories and populations, and thereby the ‘state’ as a whole, by the king. This development can be closely observed in the details of the delineation of the burghal territories of Buckingham, Oxford, Sashes and Wallingford in particular, which can be best seen as instruments in the process of the consolidation of political and strategic control over a key part of an area over which King Alfred regarded himself as the rightful lord, and which he wished to establish and reinforce over against opposing forces on three sides. This same process can be seen in the west Midlands, where King Alfred arguably constructed several burhs a short time later to consolidate his newly-won control of the area - forming thereby a system in itself - subsequent to the abrogation of control by King Ceolwulf and the Vikings in late 879. The formation of these burghal territories can, furthermore, shed light on aspects of landscape development, in particular on the metamorphosis of the political landscapes of regiones, provinciae, shires, sub-kingdoms and kingdoms, in new and unexpected ways.

In conclusion, there is every reason to hold that the formation of the burghal system around Wessex, and in western Mercia in the early 880s, was designed and put in place by King Alfred to consolidate his political and strategic control of the kingdom in a way which is directly evidenced throughout the pages of the Chronicle and in other sources in succeeding decades. A case has been made in this study that the physical manifestations of the application of this series of strategies can be recognised in the definition of the dependent territories of the burhs as ‘territories of obligation’, for which the urban-rural connections discussed in this study are a functional attribute, and for whose extent they are, therefore, a principal source of evidence. A further case has been made that estates within each of the burghal territories were connected to the burhs by a system of assessment whereby they were obliged to provide renders of food supplies, which was measured in ploughlands. This may have been either a new type of

137

Acknowledgements I am grateful to both David Roffe and Steven Bassett for reading an earlier draft of part of this study (notwithstanding the fact that neither would - on published evidence - agree with all the details of the hypotheses advanced herein), and for suggesting a number of corrections and improvements as well as asking pertinent questions about key issues. James Bond has kindly provided me with an updated version of fig. 18a which shows the distribution of tenements in Oxford appurtenant to Abingdon Abbey. He has also gathered the data used in fig. 18b, and has allowed me to use the maps of the estates of Abingdon Abbey in figs. 23a and 23b. I am also grateful to the publisher of Urban Growth and the Medieval Church: Gloucester and Worcester by Nigel Baker and Richard Holt who has allowed me to use two of the illustrations relating to the distribution of tenements in

Gloucester in figs. 17a and 17b. Similarly, I am grateful to Oxford University Press for allowing me the use of fig. 16 relating to the distribution of tenements and rents in Winchester prepared by Martin Biddle. I am also indebted to Patrick Hase for not only discussing the thesis of the book with me, but also allowing me to use his unpublished researches on the hidation of the manors and the formation of the burghal territories of Hampshire. This study has been written during my tenure of a Visiting Research Fellowship at the Centre for English Local History at the University of Leicester, and I am appreciative of the chance to use the resources at the Centre as well as those online which this Fellowship has made possible. Last, but not least, I owe Penny a debt of gratitude for her forbearance during the writing of this study.

138

Bibliography

Abels R (1984), ‘Bookland and fyrd service’, AngloNorman Studies 7, 1-25. Abels R (1988), Lordship and Military Obligation in AngloSaxon England (London: British Museum Publications). Abels R (1996), ‘English logistics and military administration, 871-1066’ in A N Jorgensen & L C (Birthe eds.), Military Aspects of Scandinavian Society in a European Perspective, AD 1-1300 (Papers from an international research seminar at the Danish National Museum, Copenhagen, 2-4 May 1996), Publications from the National Museum. vol. 2. Studies in archaeology & history. Online at < http://www. deremilitari.org/resources/articles/abels1.htm > Abels R (2001), ‘From Alfred to Harold II: The Military Failure of the Late Anglo-Saxon State’ in R Abels and B S Bachrach (eds.),The Normans and their Adversaries at War, (Warfare in history series; v. 12) (Woodbridge: Boydell Press), 15-30. Alcock L (1995), Cadbury Castle, Somerset: the Early Medieval Archaeology (Cardiff: University of Wales Press). Asser: Life of King Alfred, in Alfred the Great, trans. & ed. S Keynes & M Lapidge, (London: Penguin Books, 1983). Astill G (1984), ‘The towns of Berkshire’ in J Haslam (ed.) Anglo-Saxon Towns in Southern England, (Chichester: Phillimore), 53-86. Astill G (2006), ‘Community, Identity and the later AngloSaxon town: the case of southern England’, in W Davies, G Halsall, A Reynolds and A Langlands (eds.) People, Space in the Early Middle Ages, 300-1300, (Turnhout: Brepols), 233-54. Aston M (1986), ‘Post-Roman central places in Somerset’, in E Grant (ed.) Central Places, Archaeology and History, ( Sheffield: Sheffield University Dept. of Archaeology), 49-78. Bacharach B S and Aris R (1990), ‘Military technology and garrison organization: some observations on AngloSaxon military thinking in light of the Burghal Hidage’, Technology and Culture 31.1, 1-17. Bailey K (1991), ‘The hidation of Buckinghamshire’, Records of Buckinghamshire 32, 1-34. Bailey K A (1994), ‘The hidation of Buckinghamshire, Pt II: before Domesday’, Records of Buckinghamshire 34, 87-96. Bailey K A (1996), ‘Early Anglo-Saxon territorial organisation in Buckinghamshire and its neighbours’, Records of Buckinghamshire 36, 129-43.

Bailey K A (2010), ‘Manorial Demesnes in Domesday Buckinghamshire: a Reappraisal’, Records of Buckinghamshire 50, 25-33. Bailey M (2001), ‘Aelfwynn, second lady of the Mercians’, in N J Higham & D H Hill (eds.), Edward the Elder (London: Routledge), 112-27. Baines A H J (1984), ‘The Danish wars and the establishment of the Borough and County of Buckinghamshire’, Records of Buckinghamshire 26, 11-27 Baines A H J (1985), ‘The development of the Borough of Buckingham’, Records of Buckinghamshire 27, 53-64. Baines A H J, (1986), ‘The origins of the Borough of Newport Pagnell’, Records of Buckinghamshire 28, 128-137. Baker J and Brookes S (2011), ‘From frontier to border: the evolution of northern West Saxon territorial delineation in the ninth and tenth centuries’, Anglo-Saxon Studies in Archaeology and History 17, 104-119. Baker et al (1992), N Baker, H Dalwood, R Holt, C Mundy & G Taylor, ‘From Roman to medieval Worcester: development and planning in the Anglo-Saxon city’, Antiquity 66, 65-74. Baker N J (1980), ‘Churches , parishes and early medieval topography’, in M O H Carver (ed.), Medieval Worcester: an Archaeological Framework, Trans Worcestershire Arch Soc (3rd series) 7, 31-7. Baker N J & Slater T R (1992), ‘Morphological regions in English medieval towns’, in J W R Whitehand, and P J Larkham (eds.), Urban Landscapes: International Perspectives (London: Routledge). Baker N and Holt R (1996), ‘The city of Worcester in the tenth century’ in N Brooks, C Cubitt (eds.), St. Oswald of Worcester: Life and Influence (Leicester: University Press), 129-46. Baker N and Holt R (2004), Urban growth and the medieval church: Gloucester and Worcester (Aldershot: Ashgate). Ballard A (1904), The Domesday Boroughs (Oxford: Clarendon Press). Ballard A (1906a), The walls of Malmesbury’, The English Historical Review 21, 98-105. Ballard A (1906b), ‘The burgesses of Domesday’, The English Historical Review 21, 699-709. Baring F H (1909), Domesday Tables - Surrey, Berkshire, Middlesex, Hertford, Buckinghamshire, Bedford and the New Forest, (London: St Catherine Press). Bassett S (1985), ‘A probable Mercian royal mausoleum at Winchcombe, Gloucestershire’, Antiquaries Journal 65, 82-100.

139

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Blair J (1994), Anglo-Saxon Oxfordshire (Stroud: Alan Sutton). Blair J (1996), ‘The minsters of the Thames’, in J Blair & B Golding (eds.), The Cloister and the World: Essays in Medieval History in Honour of Barbara Harvey (Oxford: Clarendon Press), 5–28. Bond C J (1979), ‘The reconstruction of the medieval landscape: the estates of Abingdon abbey’, Landscape History 1, 59-75. Bradbury J (1988), ‘An introduction to the Buckinghamshire Domesday’, in A Williams and R W H Erskine (eds.), The Buckinghamshire Domesday (London: Alecto), 1-36. Brooks N P (1964), ‘The unidentified forts of the Burghal Hidage’, Medieval Archaeology 8, 74-89. Brooks N P (1971), ‘The development of military obligations in eighth- and ninth-century England’, in P Clemoes & K Hughes (eds.) England Before the Conquest (Cambridge: University Press), 69-84. Brooks N P (1989), ‘The formation of the Mercian kingdom’, in S Bassett (ed.) The Origins of the AngloSaxon Kingdoms (Leicester: University Press), 159-70. Brooks N P (1996a), ‘The West Saxon hidage and the ‘Appendix’, in in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and AngloSaxon Fortifications ((Manchester : University Press), 87-91. Brooks N P (1996b), ‘The administrative background of the Burghal Hidage’, in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and AngloSaxon Fortifications (Manchester : University Press), 128-50. Brooks N P (2003), ‘Alfredian government: the West Saxon inheritance’, in T Reuter (ed.) Alfred the Great (Aldershot: Ashgate), 153-73. Cam H M (1963), ‘Early groups of hundreds’, in Liberties and Communities in Medieval England (London: Merlin Press). (Reprinted from Historical Essays in Honour of James Tait, Manchester, 1931), 91-105. Campbell E (1971), ‘Berkshire’ in H C Darby and E J Campbell (eds.), The Domesday Geopgraphy of Southeast England (Cambridge: Cambridge University Press), 239-286. Campbell J (2003), ‘Production and distribution in early and middle Anglo-Saxon England’ in T Pestell and K Ulmschneider (eds.), Markets in Early Medieval Europe: Trading and Productive Sites, 650-850 (Macclesfield: Windgather), 12-19. Costen M (2011), Anglo-Saxon Somerset (Oxford: Oxbow). Cunliffe B (1984), ‘Saxon Bath’, in J Haslam (ed.) Anglo-Saxon Towns in Southern England (Chichester: Phillimore), 345-358. Darby H C (1977), Domesday England (Cambridge). Darlington R R (1955a), ‘Anglo-Saxon Wiltshire’, in VCH Wiltshire ii, 1-34. Darlington R R (1955b), ‘Wiltshire Geld Rolls’, in VCH Wiltshire ii, 169-217. Dodd A (ed.) (2003), Oxford before the University, (Oxford: Oxford School of Archaeology, Thames Valley Monograph series).

Bassett S (1989a), ‘Churches in Worcester before and after the conversion of the Anglo-Saxons’, Antiquaries Journal 69, 225-56. Bassett S (1989b), ‘In search of the origins of Anglo-Saxon kingdoms’, in S Bassett (ed.), The Origins of AngloSaxon Kingdoms (Leicester: University Press), 3-27. Bassett S (1996), ‘The administrative landscape of the diocese of Worcester in the tenth century’, in N Brooks & C Cubitt (eds.), St Oswald of Worcester; Life and Influence (Leicester: University Press), 147-73. Bassett S (2001), Anglo-Saxon Coventry and its Churches, Dugdale Society Occasional Paper 41. Bassett S (2007), ‘Divide and rule? The military infrastructure of eighth- and ninth-century Mercia’, Early Medieval Europe 15:1, 53-85. Bassett S (2008a): ‘The Middle and late Anglo-Saxon defences of western Mercian towns’, Anglo-Saxon Studies in Archaeology and History 15, 180-239. Bassett S (2008b), ‘Sitting above the salt : the origins of the borough of Droitwich’, in O J Padel, D N Parsons (eds), A Commodity of Good Names : Essays in Honour of Margaret Gelling (Donington: Shaun Tyas), 3-27. Bassett S (2009), ‘Anglo-Saxon Warwick’, Midland History 34: 2, 123–155. Bassett S (2011), ‘Anglo-Saxon defences in Western Mercia’, Midland History 36:1, 1-23. Bateson M (1905), ‘Review’ of A Ballard’s Domesday Boroughs, The English Historical Review 20, 143-51. Bateson M (1906), ‘The burgesses of Domesday and the Malmesbury wall’, The English Historical Review 21, 709-723. Beamish H and Parkhouse J (1991), ‘New archaeological evidence for Anglo-Saxon activity at Newport Pagnell’, Records of Buckinghamshire 33, 145-147. Biddle M (ed.) (1976), Winchester in the Early Middle Ages (Winchester Studies 1) (Oxford: University Press). Biddle M and Hill D (1971), ‘Late Saxon Planned Towns’, Antiquaries Journal 51, 70-85. Blackburn M (1998), ‘The London mint in the reign of Alfred’, in M A S Blackburn and D N Dumville (eds.) Kings, Currency and Alliances: History and Coinage in Southern England in the Ninth Century (Woodbridge: Boydell Press), 105-23. Blackburn M (2003) , ‘Alfred’s coinage reforms in context’, in T Reuter (ed.), Alfred the Great: Papers from the Eleventh-centenary Conferences (Aldershot: Ashgate), 199-217. Blackburn M and Keynes S (1998), ‘A corpus of cross-andlozenge and related coinages of Alfred, Ceolwulf II and Archbishop Aethelred’ in M A S. Blackburn and D N Dumville (eds.) Kings, Currency and Alliances: History and Coinage in Southern England in the Ninth Century, 105-23, (Woodbridge: Boydell Press), 125-50. Blair J (1990), ‘An introduction to the Oxfordshire Domesday’ in A Williams and R W H Erskine (eds.), The Oxfordshire Domesday (London: Alecto), 1-19. Blair J (1992), ‘Anglo-Saxon minsters: a topographical review’, in J Blair & R Sharpe (eds.), Pastoral Care before the Parish (Leicester: University Press), 226-266.

140

Bibliography

Draper S (2006), Landscape, Settlement and Society in Roman and Early Medieval Wiltshire, BAR (British ser.) 419, (Oxford: BAR Publishing). Draper S (2008), ’The significance of Old English Burh in Anglo-Saxon England’, Anglo-Saxon Studies in Archaeology and History 15, 240-253. Draper S (2011), ‘Burh enclosures in Anglo-Saxon settlements: case studies in Wiltshire’ in R Jones and S Semple (eds.), A Sense of Place in Anglo-Saxon Wiltshire (Donnington: Shaun Tyas). Dumville D N (1992), ‘The treaty of Alfred and Guthrum’, in Wessex and England from Alfred to Edgar (Woodbridge), 1-27. Dyer C C and Clarke H (1968-9), ‘Anglo-Saxon and early Norman Worcester – the documentary evidence’ in P Barker (ed.) The Origins of Worcester. Trans Worcestershire Arch Soc (3rd Ser) 2, 27-34. Dyson T (1978), ‘Two Saxon land grants for Queenhythe’, in J Bird, H Chapman, and J Clark (eds.), in Collectanea Londonensia: Studies in London Archaeology and History Presented to Ralph Merrifield (London and Middlesex Archaeological Society Special Paper 8), 200-215 Dyson T (1990), ‘King Alfred and the Restoration of London’, London Journal 15.2, 99-110. EUS Aylesbury (2009), Aylesbury Historic Town Assessment Report (Draft June 2009) (Buckinghamshire County Council & English Heritage), online at < http://www. buckscc.gov.uk/assets/content/bcc/docs/archaeology/A_ HT_Aylesbury_Draft_Report_Part_1.pdf EUS Buckingham (2009), Buckingham Historic Town Assessment Report (Draft June 2009) (Buckinghamshire Co. Council & English Heritage), online at < http://www. buckscc.gov.uk/assets/content/bcc/docs/archaeology/A_ HT_Buckingham_Report_Part_3.pdf> EUS Marlborough 2004, Marlborough Extended Urban Survey (Wilts Co. Council and English Heritage), online at EUS Newport Pagnell (2010), Newport Pagnell Historic Town Assessment Report (draft Sept 2010) (Buckinghamshire Co. Council & English Heritage), online at Farley M (1974), ‘Aylesbury – a defended town?’, Records of Buckinghamshire 19, 429-448. Farley, M (2007), ‘Saxon Buckinghamshire’, in Bucks CC. Solent Thames Research Framework: Buckinghamshire. Unpublished. Finberg H P R (1972), ‘The ancient shire of Winchcombeshire’, in H P R Finberg, The Early Charters of the West Midlands (2nd ed.) (Leicester: University Press), 228-36 Finn R W (1963), An introduction to Domesday Book (London: Longmans). Finn R W (1971), ‘Hampshire’, in H C Darby and E M J Campbell (eds.) The Domesday Geography of SouthEast England (Cambridge: University Press), 287-338.

Finn H W & Wheatley P (1967), ‘Somerset’ in H C Darby & H W Finn (eds.), The Domesday Geography of SouthWest England (Cambridge: University Press), 132-222. Fleming R (1993), ‘Rural elites and urban communities in late-Saxon England’, Past and Present 141, 3-37. Foard G (1985), ‘The administrative organisation of Northamptonshire in the Saxon period’ Anglo-Saxon Studies in Archaeology & History 4, 185-222. Gelling M (1973-6), The Place-Names of Berkshire (3 vols), (English Place-Name Society, 49, 50, 51). Gelling M (1992), The West Midlands in the Early Middle Ages (Leicester: University Press). Gerrard C (1994), ‘Cyrnceastre: a royal/ecclesiastical centre in the early medieval period’, in T Darvill and C Gerrard (eds.), Cirencester: Town and Landscape. An Urban Archaeological Assessment (Cirencester: Cotswold Archaeological Trust), 87-97. Hall R A (2011), ‘Burhs and boroughs: defended places, trade, and towns. Plans, defences and civic features’, in H Mamerow, D Hinton and S Crawford (eds.), The Oxford Handbook of Anglo-Saxon Archaeology (Oxford: University Press), 600-21. Hammond M (1998), ‘The Anglo-Saxon estate of Readonora and the manor of Pyrton, Oxfordshire’, Oxoniensia 63, 23-42. Hart C (1970), The Hidation of Northamptonshire, (Leicester: Leicester University, Department of English Local History, Occasional Papers, 2nd ser., 3). Hart C (1974), The Hidation of Cambridgeshire (Leicester: Leicester University, Department of English Local History, Occasional Papers, 2nd ser., 6). Hase P (1988), ‘The mother churches of Hampshire’, in Blair J (ed.), Minsters and Parish Churches: the Local Church in Transition 950-1200 (Oxford: Oxford University Committee for Archaeology), 45-66. Hase P H (1994), ‘The Church in the Wessex heartlands’ in M Aston, C Lewis (eds.), The Medieval Landscape of Wessex (Oxbow monograph, 46) (Oxford: Oxbow Books), 47-81. Haslam J (1984a), ‘The towns of Wiltshire’ in J Haslam (ed.) Anglo-Saxon Towns in Southern England (Chichester: Phillimore), 87-148. Haslam J (1984b), ‘The towns of Devon’ in J. Haslam (ed.) Anglo-Saxon Towns in Southern England, (Chichester: Phillimore), 249-283. Haslam J (1986), ‘The metrology of Anglo-Saxon Cricklade’, Medieval Archaeology 30, 99-103. Haslam J (1987), ‘Market and fortress in the reign of Offa’, World Archaeology 19 no. 1, 76-93. Haslam J (1997), ‘The location of the burh of Wigingamere – a reappraisal’, in A R Rumble & A D Mills (eds.), Names, People and Places (Stamford: Paul Watkins), 111-30. Haslam J (2003), ‘Excavations at Cricklade in 1975’, Internet Archaeology 14: online at Haslam J (2005), ‘King Alfred and the Vikings: strategies and tactics, 876-886AD’, Anglo-Saxon Studies in Archaeology and History 13, 122-154.

141

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Haslam J (2009), ‘The development of late Saxon Christchurch, Dorset, and the Burghal Hidage’, Medieval Archaeology 53, 95-108. Haslam J (2010a), ‘The two Anglo-Saxon burhs of Oxford’, Oxoniensia 75, 15-34. Haslam J (2010b), ‘The development of London by King Alfred: a reassessment’, Trans London & Middlesex Arch Soc. 61, 109-43. Haslam J (2010c), ‘King Alfred and the development of London’, London Archaeologist 12 no. 8 (Spring 2010), 231-5. Haslam J (2011a), ‘King Alfred, Mercia and London, 874-886: a reassessment’, Anglo-Saxon Studies in Archaeology and History 17, 120-46. Haslam J (2011b), ‘Daws Castle, Somerset, and civil defence measures in southern and midland England in the ninth to eleventh centuries’, Archaeological Journal 168, 195-226. Haslam J (forthcoming a), ‘The origins and early development of Tetbury, Gloucestershire’. Haslam J (forthcoming b), ‘Planning in Late Saxon Worcester’. Haslam J (forthcoming c), ‘The wall tenements of Oxford in Domesday Book, and the process of burghal formation’. Haslam (forthcoming d), ‘A probable late Saxon burh at Ilchester’, Landscape History Heighway C (1984), ‘Saxon Gloucester’ in J Haslam (ed.), Anglo-Saxon Towns in Southern England (Chichester: Phillimore), 359-84. Heighway C (1988), ‘Anglo-Saxon Gloucester: c.680-1066’ in VCH Gloucestershire iv: The City of Gloucester, 5-12. Hill D (1969), ‘The Burghal Hidage - the establishment of a text’, Medieval Archaeology 13, 84-92. Hill D (1981), An Atlas of Anglo-Saxon England (Oxford: Blackwell). Hill D (1996a), ‘The nature of the figures’, in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and Anglo-Saxon Fortifications (Manchester: University Press), 74-87. Hill D (1996b), ‘The Calculation and the purpose of the Burghal Hidage’, in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and AngloSaxon Fortifications (Manchester: University Press), 92-7. Hill D (2000), ‘Athelstan’s urban reforms’, Anglo-Saxon Studies in Archaeology and History 11, 173–185. Hill D (2001), ‘The shiring of Mercia - again’, in N J Higham & D Hill (eds.), Edward the Elder (London: Routledge),144-159. Hill D and Rumble A (1996), The Defence of Wessex the Burghal Hidage and Anglo-Saxon Fortifications (Manchester: University Press). Hinton D A (1984), ‘The towns of Hampshire’, in J. Haslam (ed.) Anglo-Saxon Towns in Southern England, (Phillimore, Chichester), 149-66. Hinton D (1996), ‘The fortifications and their shires’ in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and Anglo-Saxon Fortifications (Manchester: University Press), 151-9.

Holt R (2009), ‘The urban transformation in England, 9001100’, Anglo-Norman Studies 32, 57-78. Hooke D (1980), ‘The hinterland and routeways of late Saxon Worcester: the charter evidence’, in M O H Carver (ed.) ‘Medieval Worcester’, Trans Worcestershire Arch Soc (3rd Ser.) 7, 39-49. Hooke D (1981), ‘The Droitwich Salt Industry: an Examination of the West Midland Charter Evidence’, Anglo-Saxon Studies in Archaeology and History, 2 ( BAR (British ser.) 92), 123-69. Hooke D (1985), The Anglo-Saxon Landscape: the Kingdom of the Hwicce (Manchester and Dover, N H. Reprinted 2009 Manchester University Press). Hooke D (1988), ‘Regional variation in southern and central England in the Anglo-Saxon period and its relationship to land units and settlement’, in Hooke D (ed.) (1988), Anglo-Saxon Settlements (Oxford: Blackwell), 123-51 Hooke D (1996), ‘Reconstructing Anglo-Saxon landscapes in Warwickshire’ Transactions, Birmingham and Warwickshire Archaeological Society 100, 99-116 Hooke D (ed.) (1988), Anglo-Saxon Settlements (Oxford: Blackwell). Hooper N (1988), ‘An introduction to the Berkshire Domesday’ in A Williams and R W H Erskine (eds.), Berkshire Domesday (London: Alecto), 1-28. Hopkinson B. (1994), Salt and the Domesday Salinae at Droitwich AD 674 to 1690 (Droitwich: Droitwich and Brine Springs Archaeological Trust). Hurst J D (2004), ‘Fuel supply and the medieval salt industry in Droitwich’, Trans Worcestershire Archaeological Society, 3rd Series 19, 111-32. Humphrey-Smith C R (1984), The Phillimore Atlas and Index of Parish Registers, (Chichester: Phillimore). Jope E M (1956), ‘Saxon Oxford and its region’ in D B Harden (ed.) Dark Age Britain (London), 234-58. Jope E M and Terrett I B (1971), ‘Oxfordshire’ in Darby H C and Campbell E M J (eds.), The Domesday Geography of South-East England (Cambridge: University Press). Keats-Rohan K S B (2009), ‘The genesis of the honour of Wallingford’, in The Origins of the Borough of Wallingford: Archaeological and Historical Perspectives,. K S B Keats-Rohan and D R Roffe (eds.), BAR 494, 52-67. Kelly S (1992), ‘Trading privileges from eighth-century England’, Early Medieval Europe, 1, 3-28. Kelly S E (ed.) (2001) Charters of Abingdon Abbey, AngloSaxon Charters, 7-8 (Oxford: University Press). Kershaw P (2000), ‘The Alfred-Guthrum Treaty: scripting accommodation and interaction in Viking-age England’, in D M Hadley & J D Richards (eds.) Cultures in Contact: Scandinavian Settlement in England in the Ninth and Tenth centuries (Turnhout: Brepols), 43-64. Keynes S (1998), ‘King Alfred and the Mercians’, in M Blackburn & D Dumville (eds.) , Kings, Currency and Alliances: History and Coinage in Southern England in the Ninth Century (Woodbridge: Boydell Press) 1-45. Keynes S (1999), ‘England, 900-1016, in T Reuter (ed.), New Cambridge Medieval History III (Cambridge: University Press), 456-84.

142

Bibliography

(eds.), Wallingford: the Castle and Town in Context, BAR Publishing forthcoming. Pelteret D A E (2009), ‘An anonymous historian of Edward the Elder’s reign’, in S D Baxter, C E Karkov, J L Nelson and D A E Pelteret (eds.), Early Medieval Studies in Memory of Patrick Wormald (Studies in Early Medieval Britain) (Farnham: Ashgate), 319-36. Petit-Dutaillis C (1930) Studies and Notes Supplementary to Stubbs’ Constitutional History (Manchester : Manchester University Press). Phythian-Adams C (1986), ‘Leicester and the emergence of its county’, in C Phythian-Adams (ed.), The Norman Conquest of Leicestershire and Rutland (Leicester: Leicestershire Museums), 9-11. Pitt J (2003), ‘Malmesbury Abbey and late Saxon parochial development in Wiltshire’, Wiltshire Archaeological Magazine 96, 77-88. Poore D, Norton N and Dodd A 2009, ‘Excavations at Oxford Castle: Oxford’s western quarter from the midSaxon period to the late eighteenth century’ [Interim report], Oxoniensia 74, 1-18. Reece R (1976), ‘From Corinion to Cirencester – models and misconceptions’ in A McWhirr (ed.), Archaeology and History of Cirencester, (Oxford: British Archaeological Reports 30), 61-79. Reynolds A (2003), ‘Boundaries and settlements in later sixth to eleventh-century England’, Anglo-Saxon Studies in Archaeology and History 12, 98-135. Reynolds A J & Langlands A (2006), ‘Social identities on the macro scale: a maximum view of Wansdyke’, in W Davies, G Halsall, A J Reynolds, (eds.), People and Space in the Middle Ages, 300-1300 (Studies in the Early Middle Ages, 15) (Turnhout: Brepols). Reynolds S (1987), ‘Towns in Domesday Book’ in Holt, J C (ed.), Domesday Studies : Papers read at the Novocentenary Conference of the Royal Historical Society and the Institute of British Geographers : Winchester, 1986 (Woodbridge: Boydell Press), 295-309. Robinson J (1975), The Evolution of the Townscape of Newport Pagnell, (Bradwell Abbey Field Centre). Roffe D (1977), ‘Rural manors and Stamford’, South Lincolnshire Archaeology 1, 12-13. Roffe D (2007), Decoding Domesday (Woodbridge: Boydell Press). Roffe D (2009), ‘Wallingford in Domesday Book and beyond’, in The Origins of the Borough of Wallingford: Archaeological and Historical Perspectives, K S B Keats-Rohan and D R Roffe (eds.) (BAR British Series 494), 27-51. Round J H (1900), ‘Introduction to Domesday Book’, VCH Hampshire i, 399-447. Round J H (1902), ‘Introduction to Domesday Book’, VCH Surrey i, 275-94. Round J H (1905), ‘Introduction to Domesday Book’, VCH Buckinghamshire i, 207-29. Round J H (1908), ‘Introduction to Domesday Book’, VCH Herefordshire i, 263-307. Round J H (1930), ‘Burh-bot and Brig-bot’, in W Page (ed.), Family Origins and other Studies (London: Constable), 252-65.

Keynes S (2001), ‘Edward, king of the Anglo-Saxons’, in N J Higham and D H Hill (eds.) Edward the Elder, (London: Routledge), 40-66. Keynes S (2005), ‘The kingdom of the Mercians in the eighth century’, in D Hill and M Worthington (eds.) Aethelbald and Offa – two Eighth-Century Kings of Mercia, (Oxford: BAR Publishing, British Series 383), 1-26. Keynes S & Lapidge M (1983), Alfred the Great (Penguin Classics). Lavelle R (2003a), Fortifications in Wessex c.800-1066 (Botley: Osprey). Lavelle R (2003b), ‘The ‘farm of one night’ and the organisation of royal estates in late Anglo-Saxon Wessex’, Haskins Society Journal 14, 53-82. Lavelle R (2007), Royal Estates in Anglo-Saxon Wessex : Land, Politics and Family Strategies, BAR, British series. (Oxford: BAR Publishing). Lavelle R (2010), Alfred’s wars : Sources and Interpretations of Anglo-Saxon Warfare in the Viking Age (Woodbridge: Boydell Press). Maddicott J R (2005), ‘London and Droitwich, c. 650–750: trade, industry and the rise of Mercia’, Anglo-Saxon England 34, 7-58. Maitland F W (1896), ‘The origin of the borough’, English Historical Review XI, 13-19. Maitland F W (1897), Domesday Book and Beyond (Cambridge: University Press). Maitland F W (1898), Township and Borough (Cambridge: University Press). Manco J (1998), ‘Saxon Bath, the legacy of Rome and the Saxon rebirth’, Bath History 7, 27–54. Manco J (2008), ‘The Saxon origins of Bristol,’ online at http://www.buildinghistory.org/bristol/index.shtml Martin G H (1985), ‘Domesday Book and the boroughs’ in P H Sawyer (ed.) Domesday Book: a Reassessment (London: Arnold), 143-63. Martin G H (1987), ‘The Domesday boroughs’ in A Williams and R W H Erskine (eds.) Domesday Book Studies (Cambridge), 56-60. Mew K (2000), ‘The dynamics of lordship and landscape as revealed in a Domesday study of the Nova Foresta’, Anglo-Norman Studies 23, 155-66. Moore J S (ed.) (1982), Domesday Book: Gloucestershire (Chichester: Phillimore). Morgan P (ed.) (1979), Domesday Book: Berkshire (Chichester: Phillimore). Morley Davies A (1947-52), ‘The hundreds of Buckinghamshire and Oxfordshire’, Records of Buckinghamshire 15, 231-49. Morris J (ed.) (1978a) Domesday Book: Oxfordshire (Chichester: Phillimore). Morris J (ed.) (1978b) Domesday Book: Buckinghamshire (Chichester: Phillimore). Munby J (2003a), ‘The eastern extension’ in Dodd A (ed.) 2003, 24-5. Pedgley D E (forthcoming), ‘Rents as indicators of the medieval settlement pattern in Wallingford, Berkshire’, in N. J. Christie, K. S. B. Keats-Rohan, and D. R. Roffe

143

Urban-Rural Connections in Domesday Book and Late Anglo-Saxon Royal Administration

Rumble A R (1996), ‘Diplomatic sub-sections’, in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and Anglo-Saxon Fortifications (Manchester: University Press), 69-74. Salter H E (1933), ‘An Oxford mural mansion’ in J G Edwards, V H Galbraith and E F Jacob (eds.), Historical Essays in Honour of James Tait, (Manchester), 299-303. Salter H E (1936), Medieval Oxford (Oxford: Oxford Historical Society). Sawyer P (1978), From Roman Britain to Norman England (London). Slater T (1976), ‘The town and its regions in the AngloSaxon and medieval periods’, in A McWhirr (ed.), Archaeology and History of Cirencester, (Oxford: British Archaeological Reports 30), 80-108. Slater T R (1981), A History of Warwickshire (Chichester: Phillimore). Slater T R (1982), ‘Urban genesis and medieval town plans in Warwickshire and Worcestershire’, in T R Slater & P J Jarvis (eds.), Field and Forest: the Historical Geography of Warwickshire and Worcestershire (Norwich), 173-202. Slater T R (1983), ‘The origins of Warwick’, Midlands History 8, 1-12. Slater T R (2000), ‘The South-West of England’, in D M Palliser (ed.)The Cambridge Urban History of Britain vol.1, 600–1540 (Cambridge: University Press), 583-607 Stafford P (1985), The East Midlands in the Early Middle Ages (Leicester: University Press). Stafford P (2008), ‘“The Annals of Æthelflæd” : Annals, history and politics in early tenth-century England’, in J Barrow and A Wareham (eds.), Myth, Rulership, Church and Charters : Essays in Honour of Nicholas Brooks (Aldershot: Ashgate) Stenton F M (1907), ‘Introduction to Domesday Book’, VCH Leicestershire i, 227-305. Stenton F M (1913) The Early History of the Abbey of Abingdon (Reading: University College, Reading, Studies in Local History). Stenton F M (1939), ‘Introduction to Domesday Book’, VCH Oxfordshire i, 373-95. Stenton F M (1971), Anglo-Saxon England (3rd ed.) (Oxford: Clarendon Press). Stephenson C (1930), ‘The Anglo-Saxon Borough’, English Historical Review 45, 177-207. Sturdy D and Munby J (1985), ‘Excavations in Cornmarket and Queen Street 1959-62’, Oxoniensia 50, 47-94. Tait J (1897), ‘Review of Maitland’s Domesday Book and Beyond’, English Historical Review 12, 772-7. Tait J (1936), The English Medieval Borough (Manchester: University Press). Taylor C S (1898), ‘The origin of the Mercian shires’, Transactions of the Bristol and Gloucester Arch Soc 21, 32-57, reprinted in H P R Finberg (ed.), Gloucestershire Studies (1957), 17-45. Taylor C S (1902), ‘Deerhurst, Pershore and Westminster’, Transactions of the Bristol and Gloucester Arch Soc 25, 230-50.

Taylor-Moore K (2012), ‘The origins of Buckinghamshire and Northamptonshire’, PhD thesis, Centre for English Local History, University of Leicester. Thorn F R (1988), ‘Hundreds and Wapentakes’, A Williams and R W H Erskine (eds.), The Buckinghamshire Domesday (London: Alecto), 37-41. Thorn F R (1989), ‘Hundreds and Wapentakes’, in A Williams and R W H Erskine (eds.), The Wiltshire Domesday (London: Alecto), 31-45. Thorn F R (1990), Hundreds and Wapentakes’, A Williams and R W H Erskine (eds.), The Oxfordshire Domesday (London: Alecto). Thorn F R and Thorn C (?2009), ‘Wiltshire hundreds, introduction’ online at http://edocs.hull.ac.uk/muradora/ objectView!getDataStreamContent.action?pid=hull:524 &dsid=externalDatastream__0.8985829049168778&mi meType=application/rtf Turner H L (1990), ‘The mural mansions of Oxford, attempted identifications’, Oxoniensia 55, 73-9. VCH – Victoria County Histories (London: Institute of Historical Research). Wainwright F T (1975), ‘Aethelflaed, lady of the Mercians’, in H P R Finberg (ed.) Scandinavian England: Collected Papers by F T Wainwright, (Chichester: Phillimore). Whitelock D (1977), ‘The Importance of the Battle of Edington, AD 878’, in From Bede to Alfred: Studies in Early Anglo-Saxon Literature and History (London: Variorum Reprints, 1980). Whitelock D (1979), English Historical Documents, I, c.500-1042 (2nd ed.) (London: Eyre Methuen). Whybra J J (1990), A Lost English County: Winchcombeshire in the Tenth and Eleventh Centuries (Woodbridge). Williams A (1982), ‘Princeps Merciorum gentis: the family, career and connections of Ælfhere, ealdorman of Mercia 956-83’, Anglo-Saxon England 10, 143-72. Williams A (1989), ‘An introduction to the Gloucestershire Domesday’, in A Williams (ed.) The Gloucestershire Domesday (London: Alecto). Williams G (2005), ‘Military obligations and Mercian supremacy in the eighth century’, in D Hill, M Worthington (eds.) Æthelbald and Offa : Two EighthCentury Kings of Mercia (British Archaeological Reports, British series, 383) (Oxford: BAR Publishing), 103-10. Wormald P (1991), ‘The burhs’, in J. Campbell (ed.), The Anglo-Saxons, (London: Penguin) 152–3. Wormald P (1996). ‘BL. Cotton MS. Ortho B. xi: a supplementary note’, in D Hill & A R Rumble (eds.), The Defence of Wessex - the Burghal Hidage and AngloSaxon Fortifications (Manchester: University Press), 59-68. Yorke B (forthcoming), ‘West Saxon fortifications in the ninth century: the perspective from the narrative sources’.

144