A Hand-list of Castles Recorded in the Domesday Book

English Hhtonal Rcxnev
O 1991 Longman Group UK limiied
Notes and Documents
A Hand-list of Castles Recorded in the
Domesday Book
1. Sir Henry Ellis, An Introduction to Domesday Book (London, 1833), p. 211. Hereafter referred to
a Ellis.
2. Ellll, p. 22).
3. E. S. Armitage, 'The early Norman castles of England', ante, xix (1904), 209-45 and 417-5J. ID
The Eariy Norman Castla of the British Ida (London, 1912), ihe expanded the scope of these early
papers to incorporate all early Norman castles in Britain. Hereafter this Utter work will be referred
to as Armitage.
4. D. Renn, Norman Castla in Britain (London, 1973), hereafter referred to as Renn. Renn generally
records the earliest documentary reference to a castle as does King; infra, n.5.
5. D. J. Cathcart-King, CastdUrium Anglicanum: An Index and Bibliography of the Castla m England,
Vala and the Islands (New York, 1983), hereafter referred to as King. This volume, in its introductory
chapter, contains an excellent commentary on all major works of castle studies.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
IN 1833 Sir Henry Ellis observed that 'another point on which the Domesday Book throws considerable light is the history of our ancient castles'.1
He then catalogued fifty-one references from Domesday Book which
he identified as recording the presence of fifty-two individual castles.
He concluded: 'Of forty-nine castles here mentioned, one only, Arundel,
is noticed as existing in the time of Edward the Confessor.'2 Amongst
other things Ellis overlooked the reference to Stamford, Lincolnshire,
and incorrectly identified Eastbourne and West Firle, Sussex, as castles.
He appears also to have misinterpreted the entry for Arundel. These
errors, together with the unresolved discrepancy which exists between
his listed references and his concluding remarks, have prompted the present revision and update of the total number of castles recorded in Domesday Book.
There has been no attempt since Ellis published his pioneering work
to discuss this problem at any length. In 1904 Mrs Ella Armitage instigated
the modern study of castles with her innovative work on early Norman
castles in England.3 Armitage was concerned to discuss all available sources
of evidence and did not confine herself to the castles constructed by
the Conqueror. She usually, but not always, noted the fact that a castle
was recorded in Domesday Book. Others have followed Armitage's lead
in compiling castle gazetteers within various geographical and chronological limits: most notably in recent years Derek Renn, who concentrated
specifically on all known Norman castles. Like Armitage he does not
always note whether or not a castle appears in Domesday Book.4 David
Cathcart-King has since published his Castellarium Anglicanum which
lists and provides a bibliography for all known castles in England, Wales
and the islands off their respective shores.5 This must be regarded as
the definitive work on castles within its own geographical boundaries,
although, once again, the fact that a castle is recorded in Domesday Book
is not always noted by King.
372
A HAND-LIST OF CASTLES
April
It is worth considering in general terms why castles should come to
have been recorded in Domesday Book at all. They certainly do not
feature in the instructions to the Commissioners which were recorded
at the time. A number of contemporary commentators relate a few details
of the brief to which the Commissioners were working, but none refers
1. R. A. Brown, 'The Castles of the Conquest', Domesday Book Studies, ed. A. Williams and R.
Erskine (London, 1987), pp. 69-74. A paper with an identical title was written by J. H. Round and
is to be found in Arcbaeologu, lvui (1902), 313-40.
2. H. C Darby and L S.Maxwell, The Domesday Geography of Northern England (Cambridge, 19*2).
3. As has been noted, Ellis concluded that there were forty-nine castles recorded in Domesday Book.
Brown wrote of 'some fifty' and the table of Domesday statistics published in the catalogue, Domesdty
Exhibition 1987, by the Public Record Office states that there were forty-eight castles mentioned in
Domesday Book. The arguments against some of the castles which are claimed by certain authorities
to have been recorded in Domesday Book will be presented after the gazetteer below. Armitage stated
that there were eighty-four castles constructed in England and Wales during the eleventh century:
Armitage, p. 91. Brown, 'Castles', p. 69, accepts this figure without question, although King had already
been able to demonstrate that it should now be revised to ninety-three: King, p. xxxi.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
In 1987, when the novocentenary of Domesday Book presented an ideal
opportunity to review the evidence for castles documented therein, the
late Allen Brown instead discussed the castles of the Conquest in order
to provide a context for the study of Domesday Book, rather than publish
a paper devoted solely to the subject of castles recorded by the Domesday
Commissioners. Indeed the only comment he made concerning the
recording of castles in Domesday Book was to note that 'the Domesday
survey makes casual reference to some fifty castles'.1 The series of studies
of the geography of Domesday England carried out under the guidance
of H. C. Darby generally include amongst their appendices a list of places
within the region covered by each volume at which a castle is recorded
in Domesday Book, although this is not the case with the volume on
northern England.2 There appears to have been no consistent editorial
policy regarding the discussion of castles within the main texts of this
series. No recent study of Domesday Book or eleventh-century British
castles includes any detailed discussion of those castles specifically mentioned in the survey. Nor does there appear to be an agreed figure of
just how many castles are recorded in Domesday Book.3 This hand-list
is intended to fill the gap that has existed since Ellis first launched the
serious historical study of Domesday Book. In addition to the gazetteer
and commentary below, detailing castles directly recorded in the survey,
consideration will be given to those which can be only inferred from
Domesday Book. Castles identified incorrectly by some writers as having
been recorded in Domesday Book will be discussed, and there is a list
of castles documented before 1086 but not mentioned there.
I99 1
RECORDED IN THE DOMESDAY BOOK
373
1
1. These commentaries are to be found, amongst other places, in D. Douglas and G. Grcenaway
(ed.), English Historical Documents, ii (London, 1981), hence EHD n. In particular see the Anglo-Saxon
Chronicle, version 'E', sub anno 1085, EHD 11, doc. 1, p. 168; Robert of Losinga's note added to the
Chronicle 0/ Marian us Scotus, EHD ii, doc. 198, p. 912, which must have been written before Robert's
death in 1095; the annalist of Worcester writing in the early twelfth century, EHD ii, doc. 202,
p. 914; and Florence of Worcester writing before 1118, EHD ii, doc. 203, p. 914.
2. EHD ii, doc. 215, p. 946
3. The purpose of the Domesday Book has been adequately debated elsewhere: for instance S. Harvey,
'Domesday Book and its predecessors', ante, lixivi (1971), 753-73; V. Galbraith, Domesday Book: Its
PUce in Administrative History (Oxford, 1974); P. Sawyer, Domesday Book A Reassessment (London,
198)); and J. Holt (ed.), Domesday Studies (Woodbridge, 1988). It is worth noting here that if the purpose
of Domesday Book was primarily to record military and so<alled 'feudal' obligations then an argument
can be put that all castles should have been recorded by the scribes. The fact that not all castles were
recorded might thus be used to support the contrary view.
4. It was not only the building of defensible castles that led to the destruction of urban property.
In Westgate, Kent (DBi, 3c), twenty-seven dwellings were destroyed between 1066 and 1086 'for the
Archbishop's new lodging (nova hospiuuone)''.
j. These will be discussed in detail individually in the gazetteer.
6. H. C. Darby, Domesday England (Cambridge, 1977), pp. 3*4-8.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
to castles. Indeed, the only contemporary commentator whose account
of the Domesday inquiry mentions castles is Henry of Huntingdon who
wrote sometime before 1129 that 'he [King William] also made inquiry
what each city, castle, township, village, marsh and wood was wont to
render each year', a statement somewhat removed in time from the actual
commissioning of the work and as such, perhaps, of doubtful authority.
The Prologue to the Inquisitio Eliensis, written soon after 1086,2 details
exactly what the Commissioners were to enquire into, and in doing so
notes that answers should be provided as to 'how much has been added
to, or taken away from, the estate'. This presents at least a clue why
some castles may have been recorded, although its very proximity to
the actual compilation of Domesday Book and the absence of any specific
reference to castles in the prologue would support an argument that
William made no mention of castles when he ordered the inquiry.3
In building a castle it was necessary, in some instances, to destroy valuable agricultural land and urban assets. Such an action would certainly
come within the scope of an inquiry into how much useful land had
been taken away from an estate. In thirteen instances Domesday Book
records that land or houses had been laid waste because of the construction
of a castle. Two of these references are rural examples in which the
waste is reflected in reduced geld assessments, and the remaining eleven
are to be found in urban contexts/ The urban references are to be found
at Cambridge, Canterbury, Gloucester, Huntingdon, Lincoln, Norwich,
Shrewsbury, Stamford, Wallingford, Warwick and York.5 Typically the
number of houses destroyed to make way for the castle is recorded.
The numbers vary considerably from four houses at Warwick to ninetyeight at Norwich. Darby has demonstrated that of 112 boroughs listed
in Domesday Book, thirty (26 per cent) had recorded waste6 amounting
to a total of 3,558 houses destroyed. Of these thirty boroughs the eleven
with castles recorded accounted for a total of 1,474 destroyed houses,
with 410 (27 per cent) of these being directly attributed to the construction
374
A
HAND-LIST OF CASTLES
April
THE GAZETTEER
In the gazetteer which follows, each county is dealt with alphabetically,
and each reference within the county likewise. No significant differences
or patterns were discovered when the way in which castles were recorded
within each circuit was examined. At the end of the gazetteer an alphabetical list of castles will be presented as an alternative means of easy reference.
1. Both these sites have produced excavated evidence for substantial ringworks which were in existence
prior, not only to the Domesday survey, but also to the Conquest. At some time after the Conquest
the ringwork at Goltho, which enclosed the manorial estate centre, was converted to a motte and
bajley. G. Beresford, Goltbo: The Development of an Early Medieval Manor (London, 1987), and B. K.
Davison, 'Excavations at Sulgrave, Northamptonshire, 1960-76', Archaeological Journal, exxiv (1977),
105-14.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
of the castle. Only n per cent, therefore, of all houses recorded as vastae
in the thirty boroughs were as a result of castle-building. It has been
argued, as will be shown below, that the mentioning of urban waste
in Domesday Book may indicate the presence of a castle in 1086 even
though no castle is recorded, but clearly such arguments cannot account
for all urban waste.
There are just two castles which can be considered to have found their
way into Domesday Book directly as the result of being associated with
an economic asset. The collection of a toll is noted amongst the miscellaneous assets of the estate at Chepstow and a baronial market is recorded
for the estate of William Malet at Eye, albeit in the entry for Hoxne
on episcopal lands. Markets are also associated with entries in the Exon
Domesday text for Launceston and Trematon, although the main Exchequer text entries for these castles make no mention of this fact, or
else appear to be ambiguous about any such associations. Seven entries
treat the castle as just another manor, for instance the entries for Rhuddlan
and Arundel where the format is no different from any other manorial
estate entry. Fourteen other references go no further than to report,
quite blandly, 'ibi est castrum . There is no consistency in the way in
which castles are recorded in Domesday Book, nor is any specific inquiry
formula apparent for the noting of such structures. As will be shown
below, twenty-one castles which were documented before 1086 are not
recorded in Domesday Book. When manorial estates such as Sulgrave
are recorded with no mention of the substantial ringwork, when there
is no mention of manors such as Goltho at all1, and when the castle
at Stamford is recorded because it no longer existed in 1086, then perhaps
it is best to conclude that any reference to a castle in Domesday Book
is no more than incidental and exists because the scribe thought he ought
to record that particular castle at the time. These specific and individual
reasons will become apparent in the detailed gazetteer.
I99 1
RECORDED IN THE DOMESDAY BOOK
375
Berkshire
There are two references to castles in Berkshire in 1086. The first occurs
under the entry for the borough of WALUNGFORD (DBi^tfb)1 where
eight dwellings (hagae) were destroyed to make room for the castle within
the Saxon burh (pro castello sunt viii destructae hagae). In the entry for
the manor of Clewer, held by Ralph, son of Siegfried, in 1086, the manor
'was then assessed at five hides, now at four and a half hides; and the
castle of Windsor (castellum de Windesores) is on the half-hide' (DBi,62d).
Cheshire
In the marcher territories beyond the Dee estuary the focus of settlement
lay round the newly built castle (factum novum castellum) of RHUDDLAN
(DBi,269a), which was jointly held by Earl Hugh of Chester and Robert
of Rhuddlan:
Earl Hugh holds Rhuddlan of the King. ... Now he has in demesne half
of the castle which is called Rhuddlan and which is in the centre of the district
(caput est hujus terre)... (DBi.Kjo.a).
There is a similar entry for Robert of Rhuddlan which demonstrates
the joint responsibility and reward shared by Earl Hugh and Robert:
Robert of Rhuddlan holds half of this castle and borough from Earl Hugh,
wherein Robert himself has ten burgesses; half of the church and the mint,
and of the iron mines found there; half of the waters of the Clwyd; of the
fisheries and mills there, made or to be made; half of the toll; and of the
forests ... (DBi^a).
Cornwall
Robert of Mortain held both the castles recorded in Cornwall. The entries
in the survey of his lands are succinct and to the point. At Dunhevet
(LAUNCESTON) ibi est castrum comitis (DBi,i2id), and at TREMATON ibi habet
comes unum castrum. Here it goes on to mention that there was a market
on the manor which paid three shillings (DBi,ma). The Exon Domesday
reveals slightly more information regarding Trematon because it records
on the bishop of Exeter's estates at St Germans (DBi,i20c) that the bishop's
Sunday market had been reduced almost to nothing by the Count of
Mortain's market which was held in his nearby castle on the same day.2
Elsewhere it is noted that the canons of St Stephen's held St Stephen'sby-Launceston (DBi,i2od). The entry goes on to record that 'the Count
of Mortain took away a market from this manor, which lay there before
1066, value 20 shillings'. The Exon text explains why. The Count took
1. Throughout the gazetteer the system of folio identification deviled by Vinogradoff and employed
in the Phillimore series of Domesday Surveys (general editor J. Morris) wil] be used.
1. In C Thorn and F. Thorn (ed.), Domadxy Book- Cornwall (Chichener, 1979), the Eion ten
variations are set out alongside the main Exchequer text.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
Cambridgeshire
The castle (castro) at CAMBRIDGE is mentioned (DBi,i89a) only because
twenty-seven houses were laid waste during its construction.
376
A HAND-LIST OF CASTLES
April
the market away from the monastic house and 'put it in his castle'.
This provides as clear an example as could be wished for of a local lord
reinforcing his political and social influence by seizing effective control
of local economic interaction.1
Entries amongst the Devon folios on the manors of Benton and Haxton
show the Count exchanging land for the castle of Cornwall and this
may well refer to Launceston: istas ii predictas mansiones dedit comes
deMoritonio episcopo propter escanbium castell Cornugallie (DBi.ioib).
William Peverel's castle is indirectly referred to in the recording of the
Tfempore] R[egis] E[dwardi] land-holders: 'In PEAK'S ARSE Arnburn and
Hunding held the land of the castle {terra castelli, DBi,276b). Authorities
variously refer to this castle as Peak, Peverel or its Domesday name,
Pechefers.
Devon
Devon has only one castle recorded in Domesday Book, at OKEHAMPTON,
where a mill and a market are also noted, and then only in the briefest
of entries matching those for Cornwall: 'Baldwin [the sheriff] holds Okehampton himself from the King; [his] castle (castellum) is sited there'
pBi,iojd).
Dorset
The only Dorset castle recorded in Domesday Book comes under the
entry for the lands of the abbey at Shaftesbury which had received as
a royal gift the manor of Kingston on the Isle of Purbeck. 'Of the manor
of Kingston the King has one hide on which he built Wareham castle
(castellum Warebam)' (DBi,78d). This is generally thought to refer to
CORFE CASTLE near Kingston rather than the motte-and-bailey built in
the south-west corner of the Saxon burh at Wareham itself.
Essex
On the manor of RAYLHGH, Rochford, Essex, it is recorded that Sven
built his castle {fecit suum castellum, DBii^b). With such a sound AngloScandinavian name the lord of this estate might initially be mistaken
for a native, especially as much of his land was held TRE by his father.
But Sven's father was Robert fitz Wymarc, one of the Confessor's Norman favourites. This is not an example of an Anglo-Saxon retaining preConquest tenurial possessions and holding a castle under William I,
merely a man of Norman descent whose family's landed interests in
England commenced earlier than most.
Gloucestershire
At CHEPSTOW (DBi,i62a), recorded as the castellum de Estrighoiel, Domesday Book reveals that the castle was built by Earl William fitz Osbern
and that boats crossing the River Wye to the woods on the opposite
bank had to pay a new toll to the castle.
i. A similar instance ii to be found in the entry for Eye, Suffolk, tft/ra, p. 381.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
Derbyshire
1991
RECORDED IN THE DOMESDAY BOOK
377
The city of GLOUCESTER provides another example of urban waste caused
by the construction of a castle. Sixteen houses were knocked down (quae
modo desunt, DBi,i62a) between 1066 and 1086 for this purpose.
On the royal estate at SHARPNESS Earl William fitz Osbern built unum
castellulum (DBi.i^b). This is also known as Berkeley Castle.
Hampshire
The King also holds Alvington. Dunn held it. Then for two and a half hides;
now for two hides because a castle (castellum) stands on one virgate... (DBi,52c).
The placing of this castle on agriculturally productive land is reflected
in the reduced geld assessment for the estate. This is an example of rural
waste as a consequence of castle-building.
Herefordshire
The Herefordshire Domesday record has more references to castles than
any other county. It also includes two references to strong or defensible
houses. This reflects the tense political climate and the greater need for
social and economic security which existed in the marcher territories.
Some of the entries introduce us to the jurisdictional area of a castle
termed a castellany or castelry. One such example is CAERLEON Castle
where the reference comes in the form of an indirect mention of the
castellany:
William de Scohies holds eight carucates of land in the jurisdiction of Caerleon
castle (castellaria), and Turstin holds of him. ... The land was waste TRE
and when William received it (DBi,i85c).
The entry for CLIFFORD'S CASTLE can be interpreted as equating the unit
of the castellany with that of the manor, thus linking an administrative
with an economic unit:
Ralph of Tosny holds the castle (castellum) of Clifford. Earl William built
it on waste land. ... This castle is in the kingdom of England and not subject
to any Hundred or customary dues. Gilbert the Sheriff holds it at a revenue
... he pays sixty shillings. In this castellany {castellaria) Roger holds land for
four ploughs ... (DBi,i83b).
was one of the pre-Conquest castles built in this marcher
territory by the Norman favourites of the Confessor and its relative
antiquity in 1086 is hinted at in Domesday Book:
EWYAS HAROLD
Alfred of Marlborough holds the castle (castellum) of Ewyas from King William.
The King himself granted him the lands which Earl William, who had refortified (refirmauerat) the castle, had given him (DBi,i86a).
It is interesting to note that Domesday Book apparently records the
refurbishing'-of one of the pre-Conquest castles after some thirty years.
It is not clear from this entry whether Earl William fitz Osbern had
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
Hampshire has one entry, CAJUSBROOKE on the Isle of Wight, where the
extensive Norman earthworks lay in the royal manor of Alvington. Two
mills were attached to this castellany:
378
A HAND-LIST OF CASTLES
April
carried out necessary repairs resulting from the natural wear and tear
of occupation or whether he had engaged in the up-grading and strengthening of the defences as a response to political and military needs after
1066.
On royal estates listed at MONMOUTH a church (or chapel) and a castle
are uniquely linked in a single entry:
At first sight the entry for RICHARD'S CASTLE suggests that the manor
was economically insignificant. In its entirety it reads:
Osbern also has twenty-three men in castello Avretone (Richard's Castle). They
pay ten shillings. Value to him of this castle, twenty shillings (DBi,i86d).
But on folio 185b an entry under the lands of Robert Gernon shows
that he held in demesne five and a half hides in 'castellaria de Avretone
which did not pay tax and on which were to be found a mill and a
smith. This demonstrates that Richard's Castle was not, after all, set
in an economic wilderness, and that the principal land-holder within
the jurisdiction of the castle may not necessarily be the holder of the
castle as well. Before strong conclusions are drawn from this it should
be remembered that the unique political situation in the region may
well account for any number of apparent anomalies.
The final castle listed in Herefordshire is that of WIGMORE which is
also the subject of a brief and terse entry. 'Ralph of Mortimer holds
Wigmore castle (castellum). Earl William built it on waste land which
is called Merestun ...' (DBi,i83c).
Only within the Herefordshire Domesday survey are fortified or defensible houses recorded. A mesne tenant named Robert held EARDISLEY
from Roger of Lacy. The manor is described thus: 'It is situated in the
middle of a wood, a fortified house (domus una defensabilis) is there'
(DBi,i84d). The other domus defensabilis is at AILEY (DBi,i8/a). There
is nothing in the entries to indicate how these strong houses differed
from castles or why the Domesday Commissioners chose to use a distinct
terminology for those two manors, but it seems safe to conclude that
both locations were capable of holding out against some form of military
action if necessary.
Huntingdonshire
Amongst the HUNTINGDON borough customs it is unequivocally stated
that 'on the castle site (in loco castri) there were twenty residences liable
for all customary dues, which paid sixteen shillings and eight pence a
year. They are not there now' (DBi,2O3a).
Kent
The only direct reference to a castle in the Kent Domesday is to be
EHRApr. 9i
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
In Monmouth castle the King has in demesne four ploughs. ... What the
King has in this castle (castello) is worth one hundred shillings ... three mills
worth twenty shillings. ... The Church of this castle and all the tithe with
two carucates of land St Florence of Saumur holds (DBi,i8od).
I99 1
RECORDED IN THE DOMESDAY BOOK
379
Lancashire
There was just one castle recorded in the Domesday survey which covered
part of Lancashire, at PENWORTHAM. This estate was held by the King
and was assessed at two carucates worth ten shillings. Amongst the seigneurial assets are listed half a fishery and hawks' eyries which emphasize
the importance of hunting. The entry itself is brief and to the point:
'modo est ibi castellum' (DBi,27oa).
Lincolnshire
Castles are recorded in Domesday entries for Lincoln and Stamford. Both
these entries are in the preamble concerning boroughs which prefaces
the main record for the county, and both constitute further examples
of urban waste as a consequence of castle construction.
At LINCOLN it is recorded that:
Of the said unoccupied residences 166 were destroyed because of the castle
(castellum); the remaining seventy-four are unoccupied outside the castle's perimeter, not because of the oppression of the sheriffs and officers, but because
of misfortune, poverty and the ravages offire(DBi,}}6c).
It is noteworthy that the scribes felt it necessary to distinguish so carefully
between those houses destroyed to make way for the castle and those
which lay devastated through economic distress and fire. The fact that
such an entry is so meticulous renders it all the more peculiar that in
many places with castles the presence of the fortification was not recorded
at all, even though urban waste which is recorded may well have been
attributable to castle-building. This supports the argument that if urban
waste is not specifically attributed to castle-building then in fact it may
well have been due to other factors.
The number of houses lost through construction of the castle at STAMFORD was considerably less:
In thesefivewards before 1066 there were 141 residences and half a mill which
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
found in the borough entry for CANTERBURY. Fourteen houses were destroyed to make way for the castle (castelli, DBi,2a),whilst another eleven
were destroyed during the construction of the city ditch (fossato rivitatis,
DBi,ia).
There is an indirect reference to the castle at ROCHESTER in the entry
for Aylesford. The bishop of Rochester held land there valued at £o
17s 4d in exchange for land on which Rochester Castle had been built
{pro excambio terrae in qua castellum dedet, DBi,2c). Amongst the several
examples of persons forfeiting land or dwellings to make way for a castle
this is the only example of compensation. Prominent priests in the Church
could exchange their land when it fell foul of progress. Ordinary parishioners in the towns are not recorded as being compensated, but then,
as they were not land-holders of any note, it is unlikely that Domesday
Book would record any such payments which would only have been
small.
380
A HAND-LIST OF CASTLES
April
paid ill the customary dues: but now there are as many except five which
are unoccupied because of the construction of the castle (castri, DBi,336d).
Norfolk
Northamptonshire
In Northamptonshire only one castle is specifically recorded, on royal
demesne assessed at one hide at ROCKINGHAM: 'It was waste when King
William ordered a castle (castellum) to be built there' (DBi,22oa).
Shropshire
The palatine earldom held by Roger of Montgomery in the marcher
county of Shropshire boasts the second largest number of entries directly
recording castles in Domesday Book.
MONTGOMERY Castle is recorded as a manor in its own right (DBi,253c).
'At Montgomery castle (castellum) the Earl [Roger of Montgomery] has
four ploughs himself; he has six pounds of pence from one district of
Wales which belongs to this castellany.' There is further information
on folio 254a which shows that the Earl built the castle himself and
that fifty-two and a half hides attached to it were exempt from tax and
were used for hunting TRE.
AT SHREWSBURY English-born burgesses complained to the commissioners:
It is very hard on them that they now pay geld as they used to pay in the
time of King Edward, although the Earl's castle has occupied the place of
fifty-one burgages (castellum comitatis occupaverit li mansuras) and another
fifty are waste ... (DBi,252a).
Two mesne tenants of Earl Roger held castles of him. At STANTON (Holdgate) the entry reads 'Ibi habet Helgot castellum' (DBi^58c) whilst an almost
identical entry for Maesbury records the castle of OSWESTRY held by sheriff
Reginald:' ibi fecit Rainald castellum Luvre (DBi,2j3c).
Somerset
There are two very brief entries for castles in Somerset. Under the lands
of William of Mohun is listed the estate at DUNSTER. The entry reads:
'He holds Dunster himself, his castle (castellum) is there' (DBi,9jc). For
the Count of Mortain's land at Bishopstone it is recorded that 'his castle
(castellum), called MONTACUTE, is there' (DBi,93a).
Staffordshire
Castle, strictly speaking, did not exist in 1086. It appears to
be unique in that both its construction and destruction are recorded
STAFFORD
EHRApr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
In the long introduction to the Norfolk survey there are two references
to houses laid waste in NORWICH because of the castle. The first states
that seventeen houses were on the castle site, and secondly it is recorded
that eighty-one houses lay within the boundaries of the castle grounds,
suggesting that a total of ninety-eight houses had been laid waste by
the construction of the castle out of a total of 297 houses recorded as
destructae or vastae (DBii,n6b).
1991
RECORDED IN THE DOMESDAY BOOK
381
in Domesday Book. It was located on Henry de Ferrers' estate at Chebsey,
on which Humphrey was a mesne tenant. King William ordered it to
be built, but the entry shows that it had been destroyed by 1086:
Ad hoc Manerium pertinuit terra de Stadford in qua rex praecepitfiericastellum
quod modo est destructum (DBi,248c).
Suffolk
The single entry for a castle in Suffolk, at EYE, is unique amongst the
Exchequer Domesday records for castles and markets, although an entry
of similar significance from the Exon Domesday has already been noted
above:
In this manor [Hoxne] there used to be a market TRE ... and it was set
up on Saturdays. And Williams Malet made his castle at Eye, and on the
same day as the market used to be held on the bishop's manor [i.e. of Hoxne]
William Malet made another market in the castle and thereby the bishop's
market has been so spoiled that it is of little worth; and now it is set up
on Friday. But the market at Eye is set up on the Saturday (DBii,379a).
The castle at Eye, therefore, became a tool of coercion used by William
Malet to ensure his economic domination of the immediate district. The
Count of Mortain obviously adopted similar economic tactics at Trematon, Cornwall.
Sussex
Three castle are listed amongst the palatine rapes held by the Conqueror's
most trusted barons in Sussex. A fourth is mentioned only in the Norfolk
survey.
At ARUNDEL the Domesday place-name presents the evidence for the
existence of a castle. Ellis argues that the entry is proof of an Anglo-Saxon
pre-Conquest castle at Arundel,2 and apparently ignores the Herefordshire
evidence when he further states that this is the only pre-Conquest castle
recorded in Domesday Book. There are two possible interpretations of
the opening line of this entry, however, and it is assumed here that
1. H. C. Darby and I. Terrett, The DomatUy Geography of Midland England (Cambridge, 1971),
p. 210. King, p. 452. Renn, p. 135, identifies Burton as Stafford Castle. He does not discuss the Domesday
entry for Tutbury, referring instead to an earlier reference in Ordenc Vitalis to the castle being granted
by William I in 1071 'to a newcomer', p. 33J.
2. Ellis, p. 212. For an early criticism of this see Round, 'Castles'.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
On folio 248b the scribe records castles at TUTBURY and Burton held
by Henry de Ferrers. It is possible, however, that the scribe has mis-copied
Burton for 'burh'. The first entry reads: 'Henry de Ferrers has the castle
of Tutbury.' The very next entry reads: 'In Burton he had half a hide
on which his castle was situated (in qua sedet ejus castellum).' Darby argues
that the second entry should be read as a continuation of the first entry
and merely describes the burh of Tutbury, although King identifies it
as a castle recorded in Domesday Book which has subsequently vanished.1
382
A HAND-LIST OF CASTLES
April
the place-name 'Arundel Castle' is post-Conquest even though it is used
in a sense which also describes pre-Conquest customs. There is no other
evidence of any sort to suggest a pre-Conquest castle at Arundel. The
estate was held by Earl Roger. The emphasis of the entry is primarily
economic:
Arundel castle (castrum) before 1066 paid forty shillings from a mill; twenty
shillings for three banquets; and twenty shillings for one entertainment. Now
the borough and the harbour and the ship customs between them pay twelve
pounds (DBi^a).
Warwickshire
In the entry for the borough of WARWICK it is recorded that four houses
were laid waste to make room for the castle (iiii sunt vastas propter situm
castelli, DBi,238a).
Worcestershire
'Ibi est castellum eius' records the castle at DUDLEY held by William fitz
Ansculf (DBi,i77b). On the same folio it is noted that 'the same William
holds Bellington in the jurisdiction of his castle (in castellaria sua)'. Bellington is nine miles south-west of Dudley and this entry therefore gives
a small hint at the area of jurisdictional influence.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
The entry for the manor of Washington held by William de Braose and
assessed at fifty-nine hides is shorter but more specific: 'BRAMBER castle
(castellum) is situated in one of these hides' (DBi,28a).
The castle at HASTINGS is mentioned in passing when it is recorded
that the Conqueror gave the castellaria de Hastinges to the Count of
Eu (DBi,i8b).
Neither of the castles at LEWES is mentioned in the thirty-two entries
in which the town is recorded in the Sussex folios. There are numerous
indirect references to a castle at Lewes in the Norfolk Domesday on
the lands held by William de Warenne. The castellum de lauves (DBii.^b)
is also listed as castellione aquarum (DBii,i62b) and the castello de laquis
(DBii,i64a; 165b). There is also reference to the castellany (castaellatione)
(DBii,i63a).
Clerical pruning took place between the production of the preparatory
circuit returns such as Little Domesday, in which the East Anglian surveys
survive, and the final compilation of Great Domesday. The record of
William de Warenne's Norfolk estates still carried much superfluous
material which the Winchester scribe would probably have omitted and
this may account for the recording of a Sussex castle amongst the Norfolk
entries.
Besides providing evidence of the administrative associations between
a magnate's widely dispersed estates these entries also hint at the possibility that perhaps far more information was collected about castles in
the preparatory surveys only to be edited out for the final version. The
Exon Domesday entry for Trematon, Devon, provides similar evidence
of such editorial practice.
RECORDED IN THE DOMESDAY BOOK
383
Yorkshire
Table I
ALPHABETICAL LIST OF CASTLES AND STRONG HOUSES RECORDED IN
DOMESDAY BOOK
Castle
AILEY
ARUNDEL
BRAMBER
CAERLEON
CAMBRIDGE
CANTERBURY
CARISBROOKE
CHEPSTOW
CLIFFORD'S
CORFE
DUDLEY
DUNSTER
EARDISLEY
EWYAS HAROLD
EYE
GLOUCESTER
HASTINGS
HUNTINGDON
ILBERT'S
LAUNCESTON
LEWES
LINCOLN
MONMOUTH
MONTACUTE
MONTGOMERY
NORWICH
OKEHAMPTON
OSWESTRY
PENWORTHAM
Domesday County
Herefordshire
Sussex
Sussex
Herefordshire
Cambridgeshire
Kent
Hampshire
Gloucestershire
Herefordshire
Dorset
Worcestershire
Somerset
Herefordshire
Herefordshire
Suffolk
Gloucestershire
Sussex
Huntingdonshire
Yorkshire
Cornwall
Sussex
Lincolnshire
Herefordshire
Somerset
Shropshire
Norfolk
Devon
Shropshire
Lancashire
1. King, p. 52}, confirms the identification made by Armitage, p. 187.
EHR Apr. 91
DB Reference
1,187a
1,23a
i,28a
1,185c
1,189a
1,2a
i,52C
i,l62a
1,183b
i,78d
i,i77b
i.95c
i,i84d
i,i86a
».379»
1,162a
1,18b
1,203a
i.373c
i,i2id
11,162b etc.
i.336c
i,i8od
i.93 3
M53C
ii,n6b
i,io;d
1.253c
ij7oa
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
Within the clamores at the end of the Yorkshire survey it is recorded
that 'all Thorner is situated within the bounds of ILBERT'S CASTLE (castelli
Ilberti, D B i ^ c ) . Another entry on the same folio notes that in Saxhale
one manor and two carucates of Ketel's land lay within the castle boundary (infra metam castelli). This would appear to refer to the same castle
which has been identified from later documents as Pontefract Castle held
by the de Lacy family.1
The city of YORK was divided into seven wards, six of which were
held by the King. Of these six, one was laid waste for the construction
of two castles (castellis). In fact the two castles had been built and one
destroyed by siege as early as 1069 (DBi^98a). The destroyed castle was
rebuilt. Some dwellings were also laid waste to make way for the city
ditch (fossato urbis).
384
Derbyshire
Essex
Cheshire
Herefordshire
Kent
Northamptonshire
Gloucestershire
Shropshire
Staffordshire
Lincolnshire
Shropshire
Cornwall
Staffordshire
Berkshire
Warwickshire
Herefordshire
Berkshire
Yorkshire
April
M76b
11,43b
1,269a
i,i86d
i,2C
i ,220a
1,163b
i^52a
1,248c
'.336c
i^jjSc
1,122a
1,248c
i,j6b
i^3«a
1,183c
i,62d
M5»8a
It will be seen from this list that forty-seven places are mentioned in
Domesday Book as having castles and, as the entry for York appears
to indicate the presence of two castles, we may conclude that Domesday
Book directly records a total of forty-eight castles.
In addition there are two further references in Domesday Book from
which castles can be inferred. The first of these is Pevensey which can
be deduced from the entries for Eastbourne (DBi,2Oc) and West Firle
(DBi,2ia). Both of these manors were held by the Count of Mortain in
demesne and were assessed at forty-six hides and forty-eight hides respectively. A minimal amount of land from each of these estates fell within
the administration of the Rape of Hastings but the greater portion of
each manor lay within the Count's Rape of Pevensey. On each of these
estates persons referred to as custodes castelli held land as mesne tenants.
Clearly these persons form part of a castle-guard which was probably
permanent. It could be argued that they held their land in return for
service as custodians of the castle, but Domesday Book does not specifically state this to be the case.1 The entries do not record which castle
these persons were custodians of, but logic dictates that they should have
served at the nearest castle, which would have been the converted Saxon
shore fort of Pevensey.2 It would also be sensible to assume that these
entries refer to the castle of the rape in which the estates were located.
The second inference may be drawn from the entry for Richmond,
Yorkshire. The entry reads: 'Count Alan has in his castellany (castellariuni) two hundred manors less one. ... Besides the castellany he has
1. Elsewhere Domesday Book does record instances of land held directly for service. Examples of
conditional tenure include the owing of military service such as that recorded for the tenants of the
bishop of Winchester on hij estate at Taunton, Somerset (DBi^r/c). At Amport, Hampshire (DBi^jd),
a native thegn was subinfeudated by the mesne tenant 'as long as he behaved well towards him'.
2. Eastbourne is 3.9 miles as the crow flies from Pevensey, and West Firle is 10.9 miles. The next
nearest castle would have been Hastings, more than twice the distance of Pevensey from each of these
manors.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
PEAK'S ARSE
RAYLEIGH
RHUDDLAN
RICHARD'S
ROCHESTER
ROCKJNGHAM
SHARPNESS
SHREWSBURY
STAFFORD
STAMFORD
STANTON HOLDGATE
TREMATON
TUTBURY
WALLINGFORD
WARWICK
WIGMORE
WINDSOR
YORK
A HAND-LIST OF CASTLES
1991
RECORDED IN THE DOMESDAY BOOK
385
Various authorities have put forward arguments to suggest that other
castles are recorded in Domesday Book. Ellis identified Eastbourne and
West Firle as castles in their own right on the basis of the entries discussed
above in connection with Pevensey.5 He also suggested that there were
castles recorded at Havstone and Botintonei in Devon, although careful
examination of the Domesday text for Haxton and Benton (DBi,ioib)
reveals that these estates formed part of land exchanged for the 'castle
of Cornwall', which is probably a reference to Launceston.
Renn suggests a number of castles which he identifies as being referred
to in Domesday Book. The first of these is Berkhamstead, Hertfordshire,
where the presence of zfossarius is recorded amongst the mesne tenants
(DBi,i36c). Renn translates fossarius as 'dyke-builder' and interprets the
term to mean castle-builder. Latham confirms the translation as a trencher
or digger and cites the Domesday reference as an example.7 The link
between the presence of a skilled civil engineer amongst the mesne tenants
and the motte and bailey castle at Berkhamstead is nothing more than
tenuous and King suggests that the castle was only 'probably' built during
1. Supra, pp. 377, 3&Z.
2. The argument for this is to be found in M. Faull and M. Stinson (ed.), DomaeUy Book: Yorkshire
(Chichester, 1986), note <Ni9, and A. H. Smith, The Piact-ruana of the North Riding of Yorkshire (Cambridge, 1928), p. 2S7.
y Ellis, p. 221, argues that thu entry 'undoubtedly alludes to the castle of Richmond'. Renn,
p. 295, cites documentary evidence to support the 1089 date, and King, p. 524, sayj that the castle
is referred to 'obliquely by Domesday Book'. Later sources confirm that Count Alan's castellany was
based at Richmond.
4. D. C. Douglas (ed.), The DomatUy Mmutchorum of Christ Church, Cknttrimry (Public Record
Office, London, 1944).
5. Ellis, p. 212.
6. Ibid-, p. 214.
7. Renn, p. 109. RE. Latham, Revised Medirad Lttin Word-list (London, 1963), p. 199.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
43 manors' (DBi,38ib). This differs from the entries for Caerleon and
Hastings,1 in which the jurisdiction^ areas of castles are also mentioned
(castellaria), in that there is no actual mention of the castle itself. The
entry refers to the castellany of Count Alan, rather than linking the
jurisdiction with a specific place. At the time of the survey the settlement
now known as Richmond was still referred to as Hindrelag and it has
been argued that the place-name did not change to Richmond until after
the castle was built.2 The mention of a castellany seems at odds with
this argument, but it may be the case that the jurisdictional area and
powers of the 'castle' were defined before the castle itself was constructed.
The actual date of construction is not known, but there is evidence to
suggest that Count Alan had begun to build it before 1089.3
In addition to these two references a castle can also be inferred in
a Domesday satellite document, the Domesday Monachorum.* The castle
at Tonbridge held by Richard of Tonbridge is linked to an entry in
which it is recorded that Richard received the sum of ten shillings a
year from his castellany.
386
A HAND-LIST OF CASTLES
April
1
1. King, p. 119.
2. Renn, p. 220.
3. The compilers of the VIOOTUL County History of Ox County of Lancaster, vol. ii (London, 1908),
W. Farrer and J. Brownbill, disagree on whether or not Roger of Poitou actually held Lancaster at
the time of the survey: p. 181; see also their footnotes 44 and 46. They offer the alternative suggestions
that these references might be to the jurisdictions] areas of the castles at Penwortham or Clitberoe.
Renn, personal communication, 7 July 1990, confirms that his remarks on the Lancaster case should
betaken as a suggestion rather than a positive identification. B.J. N. Edwards, 'George Venue's Engraving
of Clitheroe Castle', Antiquaries Journal, btiv (1984), 366-71, explains why these references cannot be
to Clitheroe.
4. Renn, p. 235.
j. P. Addyman, 'Medieval Britain', Medieval Archaeology, ut (196$), 194; x (1966), 196-8.
6. Supra, p. 379.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
the reign of the Conqueror. It is argued here that this entry is insufficient evidence from which to infer the recording of Berkhamstead Castle
in Domesday Book.
Renn also argues the case for a castle at Lancaster in Domesday Book,,
suggesting that the 'castellatu Rogerii pictaviensis' refers to the castle at
Lancaster where the keep 'is traditionally attributed to Roger of Poitou'.2
Here, again, this reference is to the castellany rather than the castle itself.
It occurs twice in brief entries in the Yorkshire folios on the land of
Roger of Poitou close to the county boundary at Barnoldswick and Calton
(DBi,332a). It is by no means certain that Roger held Lancaster as early
as 1086, so these references to a jurisdictional area are not conclusive
proof of Lancaster Castle's being recorded in Domesday Book.3
With Lydford, Devon (DBi,ioob), Renn argues that the recording of
the destruaion of forty houses within the burh between 1066 and 1086
indicates the presence of a castle.4 The excavation evidence from both
the castle and the ringwork at Lydford suggests that both these structures
were built in the twelfth century.5 In the introduction to this paper
the actual amount of urban destruaion caused by the building of castles
was put in context. Ninety per cent of urban destruaion was due to
causes other than castle-building and so the inference drawn in respea
of the Domesday entry for Lydford is not supported on this basis alone.
There was similar destruction recorded at Barnstaple in Devon (DBi,ioob),
and at Dorchester, Shaftesbury and Wareham in Dorset (DBi,75a), to
name but a few instances of urban destruaion which could possibly
be associated with the building of castles by the Normans. Twenty houses
at Bridport, Dorset, however, had 'been so negleaed' that they too warranted the description of vastae (DBi,7ja). The simple recording of urban
destruaion without any explanation is insufficient evidence from which
to draw conclusions about the presence of a castle at the time of the
Domesday survey. Indeed, as has been shown at Lincoln,6 the absence
of a specific reason may militate against such an interpretation.
At Quatford, Shropshire (DBi,254a), a nova domus is recorded. Again
the existence of a castle at Quatford has led some scholars to interpret
1991
RECORDED IN THE DOMESDAY BOOK
387
1. Renn, p. 289. Brown,'Castles', p. 71.
2. Renn, p. 518.
y. King, p. 237.
4. The evidence to juppon thij assertion will be found in Chapter z, section 5 of the present author's
M.Phil, thesis submitted in March 1990 to the University of Southampton and entitled 'Feudal Economies'.
j. The nngwork is the only identified medieval fortification of any description in the vicinity.
6. The original article upon which this argument is based is to be found in J. H. Round, 'The
origin of Belvoir castle', ante, xxii (1907), J08-10. Sir Frank Stenton used Round's argument to make
the cate for Wejt Bitham in his book The First Century of English Feudalism (Oxford, 2nd edn., 1961),
p. 144, n. 4, stating that 'similar groups of tenants are found at many plica where the residence of
a great French lord is proved by the earthworks of an early Norman castle'. Renn, p. 126, and King,
p. 260, both concur.
7. Supra, p . 384.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
this entry as referring to the castle itself. The entry goes on to mention
that there was a borough at Quatford which paid nothing in dues. Unlike
the defensible houses at Ailey and Eardisley, Herefordshire, there is
nothing in this reference to indicate the castle, although Renn and Brown
accept it as such.1
The Domesday spelling of Swanscombe, Kent, provides the basis for
Renn's argument that here, too, Domesday Book records a castle. The
Domesday scribe has recorded the name as Swinescamp (DBi,6a) which
Renn interprets as 'Swein's Camp'.2 Elsewhere in the Kent survey the
place-ending 'coombe/combe' is invariably spelt cumbe, and so the departure from this formula for the Domesday version of Swanscombe is
noteworthy, yet it is not enough upon which to argue for the presence
of a castle in 1086. A ringwork at Swanscombe was destroyed by quarrying
in 19283 so there is no surviving archaeological record to resolve this
problem. The entry also records six fisheries, of which one served the
hall. There are numerous references to halls in Domesday Book, but
almost without exception the term 'hall' is synonymous with the term
'manor'.4 There simply is not enough evidence to link conclusively this
Domesday entry with the ringwork which was destroyed at the beginning
of this century.5
At West Bitham, Lincolnshire, a number of authorities have argued
(citing each other) that the reference to seven francigenae indicates the
presence of a castle (DBi,36od). Both King and Renn follow the argument
set out by Stenton that this group 'closely resembles' a similar group
of Frenchmen identified by Round as the castle-guard for Belvoir living
at Bottesford, Leicestershire.6 As was argued with reference to Pevensey7,
at best it is possible only to infer the presence of a castle from the noting
of certain people in the Domesday Book. The case for Pevensey seems
well supported by the references to the custodes castelli. The case for
Berkhamstead is much weaker, founded as it is upon a fossarius. Given
that a number of other places recorded without castles in Domesday
388
A HAND-LIST OF CASTLES
April
Book have groups of Frenchmen documented1, the existence of a castle
at West Bitham in 1086 cannot necessarily be concluded from the Domesday Book entry, whatever other documentary sources may indicate.
It is not suggested here that these authorities are necessarily wrong
in attributing a pre-1086 date for those castles discussed immediately above.
There is often other evidence to support such early dates. But the evidence
from Domesday Book is insufficient in itself to warrant the claims that
these castles are actually recorded, or referred to, by the scribe. As such,
then, they should not be counted amongst those mentioned in Domesday
Table II
CASTLES RECORDED IN OTHER DOCUMENTS BEFORE 1086" AND WHICH DO
NOT OCCUR IN DOMESDAY BOOK
Castle
ALDRETH Cambs.
BARNSTAPLE Devon
BELVOIR Leics.
BISHOP'S STORTFORD
Herts.
BOURN Cambs.
CARDIFF Glim.
CASTLE ACRE Nfk.
CHESTER Cheshire
DOVER Kent
DURHAM Co. Durham
ELY Cambs.
EXETER Devon
NEWCASTLE N'humb.
NOTTINGHAM Notts.
OLD SARUM Wilts.
OXFORD Oxon.
Date
1071
In the lifetime of the DB tenant.
preio88
first mentioned tempore
William.
In the lifetime of the DB tenant.
1081
1078
1070
1067
1072
1071
1067
1080
1068
1069-70
1071
Source
Liber EJtensis2
Monasticon'
Monasticon, u'i. 288
Regesta'
Monasticon, vi. 86
A nnales de Margin*
Monasucon, v. 12
Orderic Vitalis*
Ordenc Vttalis, iv. 205
Symeon of Durham7
Liber Eliensis, p. 194
Orderic Vitalis, iv, cap. 4
Symeon of Durham, ii. 211
Orderic Vitalis, iv, cap. 4
Regesta, i. 13, no. 46
Abingdon Chronicle1
1. Ellis, pp. 317-18, and H. C. Darby, Domesdty EngUnd (Cambridge, 1977), pp. 340-1, both provide
readily accessible lists of persons specifically recorded in the Domesday Book as being French. To
cite two examples: four Frenchmen held Grafftum, Sussex, from Earl Roger (DBi^jc), and four French
'men-at-arms' (milues) held land valued at twelve pounds a year at Chislet, Kent (DBi^zb) from St
Augustine's Abbey. In neither case is there any evidence to link these groups with any specific castle
as castle-guard, and neither of these locations had a castle.
2. E. O. Blake (ed.), Liber Eliensis (Camden Society, 3rd ser xcii, London, 1962), p. 194.
3. Sir William Dugdale, Monasticon Angiicanum, ed. J. Caley, H. Ellis and B. Bandinel (6 vols, London,
1817-30), v. 197.
4. Regesu Regum Angh-Nornumnorum 1066-1114, ed. H. W. C. Davis, C. Johnson, H. A. Cronne
and R. H. C. Davis (3 vols., Oxford, 1913-68), i. 72-3.
5. H. R. Luard (ed.), 'Anrula de Margin'. Amudes Morustki (Rolls Series, xxxvi, London, 18*4),
i p. iv {villa).
6. A. le Prevon (ed.), Ordenci Vitalis Historux EalesUsaa (Paris 1838-55), iv, cap. 5.
7. T. Arnold (ed.), Symeoms Monacbi Open Omnia (Rolls Series, lxxv, 2 vols., London, 1882-5),
ii. 199.
8. J. Stevenson (ed.), Cbroniam Mowtsterii de Abingdon (Rolls Series, ii, 2 vols., London, 1858), ii.
3.'33EHRApr.91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
Book.
I99I
RECORDED IN THE DOMESDAY BOOK
SKIPSEA Yorks.
TICKHILL Yorks.
TOTNES Devon
WISBECH Cambs.
WORCESTER Worcs.
tempore William.
built by DB tenant.
In the lifetime of the DB
tenant.
1071
1069
389
1
Cbron. deMdsa
Orderic Vitalis, x, cap. 7
Monasticon, iv. 630
Matthew Paris1
William ofMalmobur/
There are two specific points to be made when drawing conclusions
about the entries in Domesday Book which mention castles. The first
of these concerns an aspect of local economics and the second concerns
the work of an individual. Two of the estates discussed in the gazetteer
have demonstrated a positive link between castles and markets. Indeed,
the markets were held within the castle, and this presumably means within
the bailey area of the castle. These examples prove that a castle could
be more than just a military base and seigneurial residence. In these two
instances at least the castle also played a crucial role within the local
economy by providing, a setting for the hub of the economic exchange
network. For the lord within whose castle the market was held there
was the added bonus of having complete control over the organization
1. E. A. Bond (ed.)t Qrronica de Meh* a fundatione usque ad annum 1396, auctore Thorn* de Burton,
Abhue(Ro\\s Strict, xlni, 3 vok, London, 1866-8), i. 89.
2. H. R. Luard (ed.), MtUtbaei Pansicnsts, moTucbt umcu Albuni, Chronica Mdjora (Rolls Series, lvii,
7 vols., London, 1872-84), ii. 7.
3. W. Stubbs (ed.), De gcau regum Anglorum libri quinque: hiaorue noveiUe Ubn tres (Rolls Series,
xc, 2 vols., 1887-9), i- 2534. This list is based on references provided by King. Belvoir, Bishop's Stortford, Bourn, Skipsea,
Tickhill and Totnes have been included on the balance of probabilities. Of these twenty-one castles,
Renn notes only thirteen as being recorded in other sources before 1086. The accuracy of this list
therefore depends upon editorial decisions made by other authorities and, in the light of their treatment
of Domesday Book references, it cannot be regarded as definitive. It falls outside the scope of this
paper to discuss such references in depth.
5. King, p. 206, and Armiuge, p. 161. There is no confirmation of this.
6. EHD ii, doc. 1, p. 128.
7. Round, 'Castles', p. 328; King, p. 143. One of the Herefordshire pre-Conquest castles jeems a
more likely identification.
8. DBi, 293d. King. p. 417.
EHR Apr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
These twenty-one references demonstrate that not all castles built before
1086 were necessarily recorded in Domesday Book. Of the castles recorded
in Domesday Book, Cambridge, Lincoln, Tutbury and Wallingford were
also documented in earlier sources.'*
King notes that Hereford was 'probably built before the Norman Conquest'.5 Round has argued that the castle recorded as belonging to
'Robert' in the Anglo-Saxon Chronicle 'E' 6 can be identified as Clavering,
Essex, but King considers this identification to be unlikely, as Clavering
was built in stone from the outset and as such would probably have
come to the notice of other sources besides the Anglo-Saxon Chronicle.7
The Domesday entry for Oakham, Rutland, records an aulam for the
estate, and King suggests that the castle there was 'probably fortified
soon after 1075', but once again there is little documentary evidence to
confirm this.8
390
A HAND-LIST OF CASTLES
April
In all, five castles are associated with markets in Domesday Book, and
the remaining three are to be found in the entries for the towns of Okehampton, Tutbury and Wallingford. There is also a market mentioned
in York, but the Domesday entry does not directly link this market
with the castle. The principle behind the association of an urban market
with an urban castle is different only in degree from the rural market
examples discussed above. Urban castles and markets were generally under
royal control. Their economic influence was considerable, so much so
that in the twelfth century lords attempted to create new towns and
markets in order to benefit from the customs and tolls which could
be levied on their activities. To place these references in perspective it
should be noted that a total of sixty markets are mentioned in Domesday
Book.1 Local medieval markets were probably on average no more than
six miles apart, given the nature of transport conditions and travel logistics, so Domesday Book clearly records only a small minority of the
actual number. It is likely that the actions of Malet and Mortain were
repeated by lords of varying economic and social rank all over the country
for the same reasons.
i. Darby, Domesday England, pp. 369-70.
EHRApr. 91
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
of the market, and therefore, by logical extension, considerable control
over the local economy as well.
The actions of William Malet at Eye are those of aggressive seigneurial
economics. Malet set out to curb the influence of the ecclesiastical lords
in this area by seizing the initiative in the local economy. In establishing
within his own castle a rival market to that already established by the
local bishop, Malet presented a clear message to his mesne and sub-mesne
tenants about whose market they should patronize. In order to continue
showing their seigneurial lord due deference and servility, Malet's tenants
had no option but to use his market. There being insufficient economic
activity to sustain two competing markets, it was inevitable that one
should suffer a decrease in trade and go out of business. It is fortunate
that Domesday Book records this minute detail of local economic competition. The motives of Malet are clear. The fact that he succeeded is
proof that in the area around Eye the secular lord commanded a greater
influence than the ecclesiastical lord, even though the latter's influence
was established in antiquity. The actions of William Malet were mirrored
by those of the King's brother, the Count of Mortain, at both Launceston
and Trematon. If Robert engaged in similar actions at his other castles,
then Domesday Book does not record the fact. It is likely that, in order
to ensure his economic predominance in other areas, the Count would
have closed other markets besides that held by the canons of St Stephen's
at Launceston, and relocated them within his own castle walls, where
he could exert maximum influence and control.
I99 1
RECORDED IN THE DOMESDAY BOOK
391
William fitz Osbern is described as building the castles of Chepstow,
Clifford, Sharpness and Wigmore and is recorded as refortifying the preConquest castle of Ewyas Harold.4 Only seven other persons
are mentioned specifically as building castles. They are King William
(Corfe, Monmouth, Rockingham, Stafford), the Count of Mortain (Launceston, Trematon), William Malet (Eye), Earl Roger (Montgomery), Sheriff Baldwin (Okehampton), William Peverel (Peak's Arse), and Sven, son
of Richard (Rayleigh). Uniquely, in this instance, Domesday Book supports the chroniclers and other sources which have been used to demonstrate castle-building campaigns as a means of conquest and settlement.
Also uniquely, it provides an insight into one of the principal individuals
who helped make the Conquest and settlement successful and yet who
had died some fifteen years before Domesday Book was commissioned.
Had he survived, William fitz Osbern would no doubt have ranked
amongst the very greatest of the tenants-in-chief.
A source of 'considerable light' on the history of English castles Domesday Book is not. It is impossible to concur with Sir Henry Ellis's remarks
quoted at the outset. There are no useful or significant general conclusions
1. D. C. Douglas, William the Conqueror (London, 1969), p. 138; see also p. 189 for comments on
his successful attempts to persuade doubters to join the expedition to England in 1066.
2. Ibid. p. 272.
3. 'With one doubtful exception no Norman count (comes) became during the Conqueror's reign
at any time an earl (comes) in England.' Ibid. p. 294.
4. The texts of these entries are described in the gazetteer above. For immediate reference the Domesday folio references are DBi, 162a, 193b, 163b, 183c and 186a respectively.
EHRApr.9i
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
The second point worth commenting upon is the alternative source
of information which Domesday Book provides on the castle-building
activities of Earl William fitz Osbern in the Welsh Marches. Since the
Earl was killed in the battle at Cassel, Flanders, in 1071 his appearances
in Domesday Book are posthumous. Yet his very considerable influence
is more than adequately demonstrated. Earl William was the son of
Osbern, the Steward of Normandy to Duke Robert, the Conqueror's
father. Osbern was murdered in the service of the Duke, and his son
inherited both his position and influence when William the Bastard
assumed the ducal title. Indeed William fitz Osbern was perhaps the
Conqueror's closest adviser and because of the position of responsibility
he held after the invasion and during the settlement, he has been described
as 'one of the architects of the Norman conquest of England'.1 Initially
he was entrusted with control of the Isle of Wight,2 but in 1067 he was
selected by King William to control one of the palatine marcher earldoms
and was appointed Earl of Hereford. The earldoms of Shrewsbury and
Chester were held by Roger Montgomery and Hugh fitz Richard respectively. It was a position of great trust.3 Earl William's task was to subdue
and secure the Welsh border. The way he set about this can be seen
in the Domesday references to five castles within his earldom.
392
A HAND-LIST OF CASTLES
that can be drawn about the castles mentioned in its folios, or the manner
in which they are recorded; and the variety of the entries illustrated
above amply supports this assertion. Only just over half the castles which
existed in 1086 are recorded, as are only two defensible manor houses,
of which there must have been perhaps several hundred of similar appearance to Sulgrave or Goltho. In specific and individual cases, however,
it is an unequalled source of minute detail.
University ofSouthampton
c. G.
HARFIELD
Downloaded from http://ehr.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 16, 2016
EHR Apr. 91