The Boudican Revolt: Countdown to defeat

Article (PDF Available) · January 2009with107 Reads
Abstract
This paper reconsiders the chronology of the Boudican rebellion from its outbreak to when it concluded with the defeat of the Iceni and how the chronology of the rebellion can be used to gauge the more probable location of the final battle. It is not the aim of the paper to review the causes leading directly or indirectly to the rebellion as these have been covered extensively elsewhere or to rehearse the scant archaeological evidence used to support arguments for the location of the final battle.
57
Hertfordshire Archaeology and History, Volume 16, 2009, pp. 57-65
Introduction
e purpose of this paper is to reconsider the chronology
of the Boudican rebellion from its outbreak to when
it concluded with the defeat of the Iceni and how the
chronology can be used to gauge the more probable location
of the battlefield. It is not the aim of this paper to review
the causes leading directly or indirectly to the rebellion as
these have been covered elsewhere in greater detail (Webster
1978; Salway 1980; Waite 2007), or to rehearse the scant
archaeological evidence used to support arguments for the
location of the final battle. Nonetheless, where relevant, in
support my thesis that the final battle took place on or near
the Icknield Way, salient details from both archaeological
and historical sources will be brought to bear.
e question of where Boudica was eventually defeated by
the Roman Army under the command of Cauis Suetonius
Paulinus has vexed historians and archaeologists. Both
Tacitus (Annals 14. 29-39) and Dio (Roman History, LXII.
1-12) are mute on the subject, the former providing only
a generic topographical description of the battle site and
the deposition of the troops available to the governor and
e Boudican Revolt: Countdown to Defeat
GRAHAME A. APPLEBY
the Iceni’s tactics (or lack of ). Proposed locations for the
final battle range from the south-east to the west Midlands
(Webster 1978; Fuentes 1983; Waite 2007). e more
widely accepted argument for Mancetter, postulated by
Webster (1978), gained credibility in the mid 20th century,
largely on the basis of the forcefulness of Webster’s person-
ality and presentation of the then known evidence. It is
interesting to note that in Webster’s Boudica. e British
Revolt Against Rome AD60 (1978), only thirteen pages of
the one hundred and thirty are devoted to the actual revolt.
ere is no consideration of the timetable of the revolt,
other than it must have been premeditated. e disposition
of Roman forces presented by Webster (ibid. fig. 4; see Fig
1, below) also only reflects known military sites and is rather
akin to producing a map of former British Army barracks
of the 19th century! Webster’s thesis for Mancetter as the
final battle site is, nonetheless, attractive and seemingly
supported by archaeological evidence; a defensive line, coin
and pottery data of the correct period, evidence of buildings
and the recovery of pieces of Lorica segmentata (Webster
1971, 1978). However, as Webster states, this evidence
only shows the military were present at some stage. It does
Fig 1 Webster’s (1978) figure 4 (re-drawn), showing postulated troop distribution at the time of the rebellion
58
his audience that the legal status and consequence of
Prasutagus’ will (whereby he left half his kingdom to the
emperor and his portable wealth to his daughters), and the
action of the Procurator Catus Descianus and his retinue
(the flogging of Boudica, the rape of her daughters, and
calling in of onerous loans, taxes and confiscation of estates)
provided the motive for the rebellion. Yet it is inconceivable
that the procurator’s officials would have been able to
traverse Icenian territory without molestation, although it
is tempting to consider the lorica segmentata pieces found
in Norfolk (Spratling 1975) as war booty from either a
member of the procurator’s staff, or from one of Cerealis
unfortunate legionaries forming part of a relief force that
was defeated in the early stages of the revolt (I return to
this aspect below). However, it is conceded that when the
rebellion broke out, the procurator was in all likelihood
based in Londinium (see Wacher 1995), thus enabling him
to escape following the destruction of Camulodunum and
defeat of Cerialis.
Although lacking historical evidence to underpin this part
of my thesis, a tripartite organisation of the Iceni, possibly
indicated by the sites at Caistor St Edmunds (Norfolk;
Cunliffe 2005), Quidney Farm, Saham Toney (Norfolk;
Brown 1986; Bates et al. 2000) and possibly Stonea
(Cambs.; Malim 2005), may have resulted in a deliberate
effort by provincial officials to bring Boudica, as regent,
nearer to Camulodunum. is action, rather than a venture
deep into either Suffolk or Norfolk, where the precarious
not necessarily follow that these sites were occupied at the
time of the rebellion, although small units, either drawn
from the XIIII, XX and IX legions and auxilia must have
been situated along the lines of communication between
the east of the province and north Wales. In the following
essay, I will propose a different possible site for the nal
battle, and that this is dependant upon understanding the
chronology of the rebellion and the motives behind its
prosecution once it had started.
Chronological problems
It is clear from my review of previous assessments of the
rebellion that there has not been a critical or systematic study
of the chronology of the rebellion and the logistical elements
underpinning its motivations, execution, management and
direction. e notable exception is Carroll’s (1979) sum-
mary of the timing of the revolt, included as part of his
discussion on the question of the date of the rebellion,
and Fuentes’ (1984) consideration of the tactics employed
by Paulinus. Waite (2007) has provided a much more de-
tailed assessment for his proposal of High Cross for the
final battle site (Fig 2), but again is too dependent upon
Classical sources. In addition, he admits that his timings
for the speed of the Roman Army is based onstraight line’
distances between localities due to a lack of certainty of
the then existing road network (ibid. Appendix A). It is
necessary, therefore, to begin at first principles; the point
where the rebellion began. Tacitus is useful in informing
Fig 2 Sites mentioned in the text: 1 Caistor; 2 Camuludonum (Colchester, Essex); 3 Londinium (London); 4
Verulamium (St Albans, Herts.); 5 Stonea (Cambs.); 6 Saham Toney (Norfolk); 7 Longthorpe (Cambs.); 8
Brentford (Middx.); 9 North Wales/Anglesey; 10 Baldock (Herts.); 11 Sandy (Beds.); 12 Staines (Middx.); 13
Mancetter (Warwickshire); 14 High Cross (Leics.); F Arbury Banks (Herts.) – after Chevallier (1976)
59
nature of the relationship between Rome and her ally’
were at best respectful, and at worst outwardly hostile (see
Creighton 2006 on the question of succession within client
kingdoms), would have negated the need for a large military
force. e memory of the Iceni’s rebellion of AD 47 may
still have had currency in the provincial capital, especially
for those veterans settled in Camulodunum who may have
taken part in the earlier suppression. Certainly, with the
bulk of the army campaigning in the west, any attempt at
a wholesale acquisition of the kingdom from deep within
Icenian territory without sufficient support of the military
would have been very risky. A further consideration is the
position of the border between the Iceni and the province
of Britannia. Branigan (1987) and Millett (1990), among
others, place the southern border essentially along the rivers
Orwell and Gipping through Suffolk. A larger natural
barrier, however, would be the River Stour, much closer
to Camulodunum, a frontier already proposed by Cunliffe
(2005, 198 & fig 8.15).
Accepting a geographically closer encounter between the
Iceni and the procurator and or his sta would provide
an explanation for the rapidity of the initial phase of the
rebellion and the seeming lack of intelligence afforded to
the Romans in Camulodonum. ere simply wasnt time to
organise a defence and a stand until relieved’ policy was
adopted. at there was time to send messengers to Cerealis
at Longthorpe is undisputed, but what is questionable, is
the scale and speed of the response. A double cohort with
accompanying cavalry under the command of Cerealis is
a plausible suggestion, rather than a larger contingent or
vexillation. By this stage, news of the rebellion would have
begun to spread westward towards Paulinus, with inform-
ation reaching him within a matter of a few days. e time
to assess the significance of the news of the rebellion, its
scale and the then available intelligence may have resulted
in an unfortunate pause for the east of the province. In
the meantime, increasingly urgent and more reliable news
will have been essential, with the army needing time to
disengage from its active campaign in Anglesey, and pre-
parations made prior to moving eastward.
Early chronology of rebellion
It is with this framework in mind that an assessment of
the chronology of the rebellion can be proposed. With the
closeness of the Iceni to Camulodunum and the lack of
time to organise a defensive position, these account for
the easy penetration of the city. Such ease would explain
both the historical account and archaeological evidence
for the lack of time available to prepare proper defences
and the decision to take up a defensive position within the
temenos of the Temple of Claudius. e archaeological
evidence demonstrates that the earlier legionary fortress
and its defences had been levelled (Crummy 1977, 1997);
London was similarly unwalled at this time (Wacher 1995,
90). Providing the only significantly walled and defensible
area, the tactical sense of attempting to defend the temple
precinct is clear. Yet there was time for the procurator to
escape the initial phase of the rebellion, probably within the
first few hours of it breaking out, thus further undermining
the assumption the rebellion started in Norfolk. It is unclear
in which direction the procurator ed, but at the same
time, a request for assistance (rescue?) would have been
despatched to the IX Legion, being the nearest and largest
contingent of troops (Frere et al. 1974). ese opening
events may have occurred over a short a period of time
of less than twenty-four to forty-eight hours. Interestingly,
no mention is made by Tacitus or Dio of the numbers
of veterans at Camulodunum, nor disposition of auxilia
in the eastern part of the province. Nonetheless, troops
from this latter group must have been stationed along
lines of communication within the province and most
probably at Longthorpe (ibid.). Such dispositions would
potentially account for the speed at which news of the
rebellion reached Cerealis, and the rapidity of his response.
Leaving almost immediately and travelling at or quicker
than the military pace (see Benario 1986 for a discussion
on legionary marching speed during the revolt), the IX
would have expected to reach Camulodunum within two
to three days. In this respect, I agree with Fuentes’ (1984,
312) and Waite’s (2007, 121) assessment of the time it took
for Cerealis to respond.
e relief force, as Tacitus informs us, failed to reach
Camulodunum. Having to cover a distance of c. 120km (77
miles), with a full-scale rebellion in progress, the western
pagi of the Iceni, (possibly based around the Stonea - March
area; the coin hoard found at Field Baulk, March, may reflect
the uncertainty of the revolt; Malim 2005, 93) would have
been ideally situated to outflank and either delay or wipe-
out the Roman force, possibly waiting until the terrain
and circumstances would favour the Iceni; echoes of the
Varian disaster of AD 9 spring to mind. With news of the
rebellion reaching the IX the same day the rebellion began,
or shortly afterwards, there was no prospect of relieving
Camulodunum. Cerealis thus had only one option left
open to him; preventing the Iceni reaching Londinium. I
return to this point below. Keeping with Tacitus’ account
and having lost his infantry contingent, Cerealis was faced
with the very real possibility of losing the military initiative
to the rebellion and any further fighting on his part would
have been futile. His strategic and intelligence assessment
of the situation would have been vital to Paulinus and the
Legatus of the II Augustus, then stationed in the south-west.
Cerealis’ retreat would appear to be marked by the rapid
construction of a smaller fort within the south-west corner
of the legionary fortress at Longthorpe (Frere et al. 1974),
indicating a real or perceived threat of attack. e defeat of
the IXs contingent before reaching Camulodunum would
not only have provided impetus to the Iceni, but may
have encourage disaffected Britons to join the rebellion,
with military outposts, with a minimal military presence,
targeted for punitive attacks. Evidence for such ‘local’
attacks may be seen in destruction levels at sites such as
Baldock (Herts.; Armour pers.comm), Staines (Branigan
1987, 97) and Brentford (Middx.; Fuentes 1984, 311),
and possibly the hiatus in the settlement evidence seen at
Gorhambury, (Middx.; Salway 1980, 125). What began as
an initial response to the actions of the procurator and his
staff, was now a full-blown revolt exceeding the scale and
scope of the earlier Icenian rebellion.
Iceni or British?
It is at this point that a brief digression is necessary for
the latter part of this thesis. is is a consideration of the
political aims of the Iceni and their allies once the revolt
was at its height; was there a real intention to ‘rid Britain of
Rome’? Attributing political and military objectives to the
60
revolt is attractive, not least of all because once large-scale
conflict and the wholesale destruction of the provincial
capital had occurred, there was no turning back . However,
the notion that Boudica and the Iceni aristocracy held
planning meetings as a direct result of Roman provocation
and alleged mistreatment of their allies such as suggested
by Waite (2007, 121) relies too heavily upon Tacitus’
and Dios accounts of the events leading to the rebellion.
ese accounts must be dismissed as unreliable. We must
recognise that until the events that led to the revolt took
place, the Iceni were allied to Rome, even if uncomfortably;
the Catevallunni were not a conquered people, unlike
the Trinovantes. e Atrebates, under Cogidubnus, were
certainly allied to Rome on a more formal basis; the status
of the Cantiaci, Corvetti and Corialtani are impossible to
assess, but it is disingenuous to assume these polities would
naturally ally themselves to the Iceni. As Webster observed,
Cogidubnus may have exerted influence on the Dubonni to
surrender during the westward expansion of the II Augusta
(Webster 1978, 55). It is distinctly possible that the Atrebates
would actively oppose the Iceni and even possibly aided the
suppression of the revolt. Certainly, there is no evidence for
destruction levels at Silchester and the possible provision
of earthwork defences may relate to attempts to protect
the city from any potential eastward attack following the
earlier revolt (Wacher 1995, 271-275). Intriguingly, the
argument put forward by Cunliffe (2005, 229) that the II
Augusta remained in the west to ensure compliance of the
Durotriges, on closer inspection, may have merit to it. If
this was the case, why did Poenius Postumus, temporary
commander of the II Augusta, subsequently commit suicide
for, in effect, preventing a second revolt breaking out to
the rear of Paulinus (see below)? We should then, probably
see the subsequent events of the revolt as not a deliberate
aim to evict Rome from Britain under a policy of national
liberation, but more as the consequence of actions and the
need to prosecute the war to a natural conclusion; victory
or defeat. We should therefore dismiss any ideas of a pan-
polity identity and common aims (see Creighton 2006 for
a discussion on the politics of late pre-Roman Iron Age
and early Roman Britain). Being aware of the geo-political
situation in southern Britain, this would have affected the
decision making of the Iceni, influenced the course of the
revolt, and ultimately where the final battle took place.
e Icenian chronology
Returning to the matter in hand, the chronology of events,
the defenders of the Temple of Claudius were annihilated
within a couple of days, with the city ransacked and razed to
the ground. e archaeological evidence for the destruction
of the city is unequivocal (Hawkes & Crummy 1995;
Crummy 1977, 1997), but in the absence of the historical
account of the revolt, this data may easily have been
attributed to other non-violent causes, such as in the case
of the mid 2nd century fire that destroyed large sections of
Verulamium (Niblett 2001). It is at this point that my thesis
further diverges from the accepted chronology of the revolt.
Cerealis, unhindered, may have reached Camulodunum in
time to relieve the defenders (although this may simply
be semantics), engaged the Iceni and stalled any further
advance or spread of the revolt. Obviously, we know this
was not the case, but the subsequent events between the
loss of Cerealis infantry, the destruction of Camulodunum
and the advance on Londinium are problematic. e reason
why the timing of events at this point presents a challenge
is the unknown size of the Iceni war band and the speed at
which they moved. Accepting (somewhat cautiously) that
the war band now included women and children, this has
more the appearance of a tribal migration than military
campaign; however, such a migration is unlikely due to
the circumstances that led to the revolt in the first place.
Nonetheless, we do have to account for the speed at which
the Iceni advanced towards Londinium and consider the
possibility that Cerealis was actually defeated between here
and Camulodunum.
Assuming a war band of several thousand, the speed at which
this body of people could move would be severely restricted
and moving as an undisciplined entity probably limited to
no more than 16km (10 miles) per day. As has been observed
elsewhere, during the 18th century covering 13 - 16km
(eight miles to 10 miles) a day was considered acceptable
(Chandler 1981, 48). e city of Londinium is 80km (56
miles) from Camulodunum. is gives a time-span for the
attack on Londinium of between five and six days after the
departure of the Iceni from Camulodunum, and more than
sufficient time for Cerealis to have positioned his forces in
front of the Icenian advance. Unfortunately, his force was
overwhelmed. Retreating to Longthorpe he, nonetheless,
would have been able to provide intelligence and an up to
date situation report to Paulinus, already advancing from
north Wales. Aware of the loss of Camulodunum and the
defeat of Cerealis, verification of the strategic situation
was needed by the governor. Taking a cavalry detachment
(Equites Singulares?), Paulinus reached Londinium ahead of
the Iceni, but with insufficient forces to place into the field
and, more importantly, into a position that provided the
greatest advantage to the number of troops that could be
brought quickly together. Consequently, Londinium was
abandoned to its fate.
e Icenian attack on Londinium was a foregone conclusion.
As at Camulodunum, Londinium was looted and burned
(Marsden 1980; Perring et al. 1991). From a strategic
perspective, and with hindsight, Paulinusdecision not to
defend the city was sensible. e city was not provisioned
with a wall or a substantial garrison, and was at this time
a largely mercantile centre (Wacher 1995; Millett 1990).
e conclusion to be drawn from Paulinus’ actions was that
the time from the receipt of the news of the revolt and the
attack on Londinium could have been no more than a week.
is was, however, sufficient time for him to be within
reconnaissance range of the enemy, but in no position to
oppose them. Furthermore, it indicates the XIIII Legion
and elements of XX Legion was still too far away to prevent
the Iceni moving westward towards Verlamium. e defeat
of Cerealis, the loss of Londinium and the deposition
of the XIIII/XX seems also to have influenced Poenius
Postumus(temporary commander of the II Augusta) dec-
ision to ignore the governor’s order to march eastward.
Whether this decision was made on the basis that victory
seemed unlikely, or that incomplete intelligence and a lack
of appreciation of the strategic situation that led to a fatal
hesitation (for the commander), the II Augusta played no
part in suppressing the revolt. Again, there was sufficient
time for the Iceni to advance upon and attack Verulamium,
meting out the same fate as at the previous two centres.
What is of interest following the attack on Londinium is
the logic behind attacking Verulamium. Was this part of
the Icenian strategy, revenge exacted upon Roman allies,
61
notably by the conquered Trinovantes, almost as a sideshow,
or more simply a means to reprovision the war band? e
historical account tells us that the Iceni attacked official
centres and where grain was stored and this may account
for this western diversion. Unfortunately, with very few
exceptions, the archaeological evidence for widespread des-
truction dating to the same narrow period outside these
urban areas is somewhat lacking, and there remains some
doubt as to the size and importance of Verulamium at this
time (Hall pers. comm). e possible exceptions to this
lack of evidence are the locations cited above. e need,
therefore to reprovision a large war band may well have
determined the tactical objectives of the Iceni after the
destruction of Londinium. e destruction of Verulamium
is where we effectively ‘lose’ the Iceni. Rather than suggest
they continued along Watling Street towards Paulinus, a
more logical direction, and strategic decision, would be to
utilise the Icknield Way as a direct route back to Norfolk.
e Icenian chronology for the revolt would thus adhere to
the following timetable (Fig 3):
Day 1 revolt begins
Day 2/3 destruction of Camulodunum and siege of
defenders
Day 3 destruction and razing of Camulodunum
Day 4 advance on Londinium
Between Day 4 and Day 9, Cerealis defeated
Day 8/9 destruction of Londinium
Day 10/12 advance on Verulamium
Day 13 destruction of Verulamium
Day 14/15 swing north onto the Icknield Way
Day 16 final battle
Chronology of response
e loss of Camulodunum, Londinium and Verulamium
was a severe blow to Roman prestige, especially the cult
centre to the deified Claudius, and the defeat of a significant
fraction of the IX Legion; effectively temporarily ‘knocked’
out of action. It is at this point that the chronology and
historical account of the revolt need to be critically re-
considered. is reconsideration stems from the suggestion
that the Iceni were not following a nationalist agenda to
defeat and eject Rome, and the reaction and speed of the
governor to the news that a serious revolt had erupted in the
east of the province.
Paulinus response to the revolt was clear if not immediately
effective; to intercept the Iceni. e problem with the
historical account, the speed and movement of the Iceni,
and the speed of the disengagement and march east by
the XIIII/XX, cannot be reconciled for several reasons.
Firstly, the decision of Paulinus to abandon Londinium
and Verulamium clearly illustrates he was in no position
Fig 3 Icenian movement rate: A Days 2–3, destruction of Camuludonum; B Days 4–8, defeat of Cerialis; C Days
8–9, advance on Londinium; D days 10–13, advance on Verulamium; E Days 14–15, turn on to Ickniled
Way; F Day 16, Arbury Banks - Final battle)
62
to bring his forces to bear. Allowing for the receipt of news
of the revolt at the earliest possible moment, the loss of
Londinium and Verulamium indicate that even marching
at 50kms a day, covering the 370 - 400km from north
Wales would have taken seven to eight days, and would
have left the army severely fatigued. Secondly, news of
the revolt and defeat of Cerealis must have taken at least
two, possibly three days to reach north Wales. irdly, the
news that the IX Legion had despatched a detachment may
have been considered a sufficient military response that led
to a delay in Paulinus pulling the XIIII/XX back east at
speed. Finally, having learnt of the defeat of the IX Legion,
Paulinus needed real-time (first-hand?) intelligence, thus
leading to his rapid advance on Londinium with only a
light covering force of cavalry.
is sequence of events, accepting the timetable outlined
above, enables an assessment of the speed or direction of
the XIIII/XX to be made, and thus the most likely area
where they intercepted the Iceni. With the XIIII/XX on
active campaign in north Wales, it is reasonable to assume
that this expedition would not have started until the
spring. It is therefore more likely to be propaganda that the
Iceni are recorded as not having sown their crops, a view
recently reiterated by Frere (2000) and Waite (2007). With
Paulinus on campaign and the inference that the Iceni had
sown their crops, this provides a clue as to where and at
what stage in the revolt the final battle took place. What is
this clue? Shorn of a pan-nationalist agenda to ‘rid Britain
of Rome’, having destroyed three urban centres (and an
unquantified number of rural sites), and camped near Ver-
ulamium, the less prosaic decision the Iceni could make
was to simply return home via the Icknield Way. Reaching
this decision may have been difficult, and in all probability
opposed, but it has merit. e Iceni still held the strategic
initiative, they were confident, they knew the terrain and,
finally, a retreat to Norfolk would provide further military
advantages. We know, of course this did not happen and
it is the tactical advantage of a small, well disposed and
disciplined military force that swung the strategic advantage
in Paulinus’ favour.
Having assessed the situation and deciding to bring the
XIIII/XX up as quickly as possible, Paulinus was presented
with a dilemma. Either allow the Iceni to advance on a
well-prepared position or attack the Iceni using standard
field manoeuvres. e absence of significant forces in the
area around the Iceni may have encouraged them to believe
that their objectives had been achieved and that any further
advance west would inevitably bring them closer to two
unmolested legions, a potentially equal number of auxilia
and additional cavalry. A decision by the Iceni to return to
Norfolk, however, permitted Paulinus to opt for a defensive
action in a prepared position. e travelling speed of the
XIIII/XX now makes its mark on the chronology of the
revolt. Covering the 50km per day, the requirement to find
a suitable location to prepare and position his forces, and at
a distance sufficiently far away from the enemy to prevent
early discovery of the plan, and at a suitable interception
point, allow the following argument to be proposed for the
location of the final battle and defeat of the Iceni.
Using the networks of roads to bring the XIIII/XX east
to a point near Baldock, and in the possession of accurate
intelligence as to the Iceni’s position, numbers and their
somewhat slower progression along the Icknield Way, a
suitable interception point could be selected. is position
needed to provide protection to the rear and anks, but
sufficiently wide enough to deploy the legion with the
cavalry on the flanks. e location for an interception also
needed to be sufficiently far north to permit the preparation
of the position unmolested. e high stakes of the operation
and subtle niceties of a carefully laid-out camp make the
construction of such a camp unlikely and the more likely
scenario was the need for a suitable redoubt. On day fifteen
or sixteen of the revolt Paulinus was in position; time was
on his side and aided by the tactical error committed by
the Iceni. e chronology for the Roman army would thus
keep to the following proposed schedule (Fig 4):
Day 1 revolt begins
Day 2/3 destruction of Camulodunum and siege of
defenders
news of the rebellion received by Cerealis;
preparation for movement of a detachment of
the IX Legion
Day 3 news reaches Paulinus; disengagement from
campaigning in north Wales and preparation
for advance eastward
Day 4/9 Cerealis defeated between Camulodunum
and Londinium; retreat to Longthorpe and
messenger despatched to Paulinus
Day 6/7 XX Legion increases marching speed; Paulinus
uses small cavalry force to reconnoitre the
Iceni’s advance on Londinium
Day 8/9 destruction of Londinium; Paulinus unable to
bring the XIIII/XX forward quickly enough
Day 10/12 Iceni advance on Verulamium
Day 13 destruction of Verulamium
Day 14/15 intelligence indicates Iceni progressing along
Icknield Way
Day 15 Paulinus positions XIIII/XX in defensive
position
Day 16 final battle
Locating the final battle
Considering together the time for the XIIII/XX to march
east, preparation of a suitable defensive position and
exploiting the topography to its full advantage, the route
along the Icknield Way near Ashwell (Herts.) provides
the ideal location to block the Iceni’s route and force an
engagement. Significantly, the oval earthwork at Arbury
Banks (SAM 20759; TL 2612 3870) is located within a
dry valley, broad enough to deploy the infantry and cavalry,
sufficiently steep-sided and wooded along the ridge line,
thus protecting the flanks and rear. Importantly, although
Iron Age in origin, elements of the rampart and ditches
survive, respectively, to a height and width of 1.2m and 5m
to the present day, attesting to its potential use as a fortified
redoubt. Unfortunately, ploughing of the site has reduced
the outward-facing scarp or has completely destroyed it.
Excavation in the mid-19th century revealed the external
ditch reached a depth of 4.5m and features within the
earthwork produced several human skulls and Roman mat-
erial (albeit spanning the whole Romano-British period)
and a reputed coin hoard (Beldam 1859; Aylott 1912). e
earthwork, itself, is located below the head of the valley
and the western aspect, facing the Icknield Way, presents
a curving frontage; the eastern side is straight due to
more recent alteration. Selecting this conveniently located
63
earthwork makes military and tactical sense. Crucially,
minimal effort would have been required by the Roman
army to prepare the site as a defensive redoubt. Additionally,
the eastern side of the site is characterised by a plateau that
remains out of site from the lower valley and would have
provided an ideal vantage point for cavalry elements of
Paulinus forces. e frontage presented to the Iceni would
therefore have been narrow, approximately 250-300m. In
opting for this defensive position, or a similar location
nearby, there would have been the added advantage of
effectively ‘funnelling’ the war band into and up the valley.
is would have compacted and restricted the ability of
the Iceni to use their weapons effectively, and screened’
the cavalry until a large part of the Iceni war band was
committed to an attack. e tactical odds would have been
significantly in Paulinus’ favour.
ere are also two further features of the site to take into
consideration when proposing Arbury Banks as a more
likely contender for the location of the final battle. Firstly,
the remnants of the IX Legion would have been deployable
to the area, meeting the XIIII/XX around Sandy (Beds.),
and providing rst-hand intelligence. e addition of the
remaining elements of the IX Legion would have bolstered
the XIIII/XX with up to two thousand men, if not more.
Added to the five thousand legionaries and a probable
similar number of auxilia of the XIIII/XX, Paulinus would
potentially have had between ten to twelve thousand men
at his disposal. However, it is recognised that although dis-
engaged from hostilities in Wales, a sufficient number of
legionaries and auxilia would have remained deployed in that
theatre. e absence, therefore, of the II Augusta makes the
choice of battle site all the more important; Tacitus states
two thousand legionaries were deployed from Germany to
bring the IX Legion back up to strength, in addition to a
further eight auxilia infantry cohorts and a thousand cavalry
added to the reinforcements sent to the province after the
revolt was suppressed. It is impossible to accurately estimate
the number of combatants within the ranks of the Iceni
(we can dismiss Dios estimate of 230,000 as propaganda),
but a number twice or three times greater than the Roman
forces does not seem unreasonable. Funnelled into the dry
valley, at Arbury Banks there was, literally, no room for
the Iceni to manoeuvre and they faced a buttressed and
well-positioned Roman infantry line, auxilia and cavalry
across a narrow front. A further piece of evidence that I
wish to consider in support of my thesis is the siting of
a substantial building south-east of Arbury Banks. is
building measures c. 150m x 115mm (Fig 5) and following
a review of the building’s position, layout and size, these
indicate the site of a probable temple (correspondence
2000), contrary to St Josephs initial interpretation of a
farmstead (1977 160 & Plate XVII). e construction of
a temple to mark the location and victory of the XIIII/XX
would not be unreasonable. Only a survey of the site and
excavation of this building will confirm its function.
Fig 4 Roman movement rate: A Days 2–3, disengagement; B Days 4–8, Cerialis defeated; C Days 6–9,
reconnaissance line, and disposition of Legio XIV and XX; D days 10–13, Redirection of combined force,
including remnant of Legio IX; E Days 14–15, intelligence report on Iceni position; F Day 16, Arbury Banks
- Final battle)
64
Conclusion
I accept that the thesis outlined above will face the same and
equally valid criticisms that others have in that the evidence
to support the nal battle location is based on no more than
a coincidence of topography. I do believe, however, and
have cogently argued, that previous arguments have failed
to sufficiently take into consideration the disparity in the
chronology of the revolt, recounted both in the historical
account and from the known speed at which a large body
of unorganised groups move. Importantly, the ability of the
procurator to escape has not been properly considered, nor
the impact and possible location of the defeat of Cerealis
and the IX Legion. Similarly, the speed at which the XIIII/
XX marched eastward needs to be considered in the light
of the intelligence available and the requirement to rest and
prepare a defensive position. It is also necessary to reconsider
the relevance of the II Augustas legate’s decision to ignore
Paulinus order. is clearly reveals that the Iceni were not
moving westward as suggested as such a movement would
inevitably lead to the II Augusta having to contend with
the revolt; why wait? e legate had already concluded the
situation in the east was untenable and that to hold the
west, unmolested, was the correct medium term stance to
take. As we know, Paulinus had other views on the matter.
e selection of Arbury Banks as the site for the final
battle between the Iceni and Paulinus is predicated on the
assumption that we have poorly understood the chronology
of the revolt from its initial outbreak and the speed at
which the respective forces would have been able to move.
e Iceni much more slowly, Paulinus disciplined and
fast, but requiring an element of pre-planning, defensive
positioning and the marshalling of his forces. My thesis also
hinges on the assumption that the Iceni were not following
a nationalistic’ agenda, but were pursing their own aims. It
remains to be seen by systematic fieldwork whether further
archaeological evidence from Arbury Banks, or nearby, will
vindicate the argument discussed above, but I do believe
we should cease looking for the battle site in the Midlands,
conveniently halfway between Wales and Londinium and
turn attention closer towards the Icknield Way.
Fig 5 e Romano-British building, Claybush Hill, Ashwell (TL265384),
looking north-west (reproduced with permission of the Unit for Lanscape
Modelling, University of Cambridge).
65
Bibliography
Aylott, G. 1912 Arbury Banks, Ashwell. In East Hertfordshire
Archaeological Society Transactions 4
Bates, S., Ashwin, T., Davies, J., Gale, R., Lyons, A., Percival, S.,
Starley, D., Turner-Walker, C., Mortimer, C. and Davies,
J.A. 2000 Excavations at Quidney Farm, Saham Toney,
Norfolk 1995. In Britannia 31: 201-237
Bedlam, J. 1859 Excavations at Arbury Banks, near Ashwell,
Herts. In Proceedings of the Society of Antiquaries 4: 285-
291
Benario, H.W. 1986. Legionary Speed of March before the Battle
with Boudicca. In Britannia 17: 358-362
Branigan, K. 1987 e Catuvelluani. Gloucester: Alan Sutton
Brown, N.A. 1986 e Iron Age and Romano-British Settlement
at Woodcock Hall, Saham Toney, Norfolk. In Britannia
17: 1-58
Carroll, K.K. 1979 e Date of Boudiccas Revolt. In Britannia
10: 197-202
Chandler, D.C. 1981 e Logistics of Military History. In History
Today. February: 48
Chevallier, R. 1989 Roman Roads. London: Batsford
Creighton, J. 2006 Britannia. e creation of a Roman province.
London: Routledge
Cunliffe, B.W. 2005 Iron Age Communities in Britain. London:
Routledge
Crummy, P. 1977. Colchester: e Roman Fortress and the
Development of the Colonia. In Britannia 8: 65-105
Crummy, P. 1997. City of Victory: e Story of Colchester Britains
First Roman Town. Colchester: Colchester Archaeological
Trust
Frere, S. 2000. A ‘Limitatio of Icenian Territory? In Britannia
31: 350-355
Frere, S., St Joseph, J.K., Charlesworth, D., Goodburn, R.,
Hartley, B.R., Marples, B.J., Myres J.N.L., Reece, R. and
Wilson, M. 1974. e Roman Fortress at Longthorpe. In
Britannia 5: 1-19
Fuentes, N. 1984. Boudicca Revisited. In London Archaeologist
4: 311-317
Hawkes, C.F. and Crummy, P. 1995. Camulodunum 2. Colchester
Archaeological Reports 11. Colchester: Colchester
Archaeological Trust
Malim, T. 2005 Stonea and the Roman Fens. Stroud: Tempus
Marsden, P. 1980 Roman London. London: ames & Hudson
Millett, M. 1990 e Romanization of Britain. Cambridge:
Cambridge University Press
Niblett, R. 2001 Verulamium: the Roman city of St Albans. Stroud:
Tempus
Perring, D., Roskams, S. and Allen, P. 1991 Early Development
of Roman London west of the Walbrook. e Archaeology
of Roman London Volume 2. CBA Research Report 70.
London: Museum of London & Council for British
Archaeology
Salway, P. 1980 Roman Britain. Oxford: Oxford University Press
St Joseph, J.K. 1977 Air Reconnaissance in Roman Britain, 1973-
7. In Journal of Roman Studies 61: 125-161
Spratling, M. 1975 Fragments of a Lorica Segmentata in the
Hoard from Santon, Norfolk. In Britannia 6: 206-207
Wacher, J. 1995 e Towns of Roman Britain. London: Batsford
(BCA)
Waite, J. 2007 Boudica’s Last Stand: Britain’s Revolt Against Rome
AD60-61. Stroud: Tempus
Webster, G. 1971 e Military Situations in Britain between A.
D. 43 and 71. In Britannia 1: 179-197
Webster, G. 1978 Boudica : the British revolt against Rome, AD
60. London: Batsford
Webster, G. 1984 Boudica - A Reply. In London Archaeologist 4:
411-412
66
  • Article · Nov 1986
  • Article · Nov 1979
  • [Show abstract] [Hide abstract] ABSTRACT: The Roman military site at Longthorpe, west of Peterborough, was discovered by observation from the air in 1961: no traces now remain on the surface. Excavations between 1967 and 1973 have shown that a 27-acre fortress with timber buildings, accommodating a legionary vexillation together with auxiliary troops, was succeeded by an 11-acre fort. The period of occupation extended from c. 44/8 to c. 62. Crop marks observed from the air at Orton Stanch, 400 m. south-east of the fortress, have led to the discovery of a native farmstead. Excavations there by Messrs. A. Challands, G. B. Dannell and Dr. J. P. Wild have shown that agricultural activities were intermitted for the period of the occupation of the fortress, the settlement being replaced by a large industrial establishment or worksdepot involved in pottery manufacture and possibly in bronze-working with clear military connections. The present report describes the excavation of the fortress: the industrial site is to be published in a future volume of Britannia.
    Article · Nov 1974
Show more

    We use cookies to give you the best possible experience on ResearchGate. Read our cookies policy to learn more.