Castle Hill, Ipswich
Castle Hill, Ipswich
Castle Hill, Ipswich
Ipswich, Suffolk
Archaeological Evaluation
and an Assessment of the Results
Wessex Archaeology
Ref: 52568.02
July 2003
CASTLE HILL, IPSWICH, SUFFOLK
ARCHAEOLOGICAL EVALUATION
AND ASSESSMENT OF THE RESULTS
Prepared for:
By:
Wessex Archaeology
Portway House
Old Sarum Park
SALISBURY
Wiltshire SP4 6EB
July 2003
© Copyright The Trust for Wessex Archaeology Limited 2003, all rights reserved
The Trust for Wessex Archaeology Limited, Registered Charity No. 287786
CASTLE HILL, IPSWICH, SUFFOLK
ARCHAEOLOGICAL EVALUATION
AND ASSESSMENT OF THE RESULTS
Contents
Summary
Acknowledgements
1 Background .......................................................................................................................1
1.1 Description of the site................................................................................................1
1.2 Previous archaeological work ...................................................................................1
2 Methods .............................................................................................................................2
2.1 Introduction ...............................................................................................................2
2.2 Aims and objectives ..................................................................................................2
2.3 Fieldwork methodology ............................................................................................2
3 Results................................................................................................................................4
3.1 Geophysical survey ...................................................................................................4
3.2 Archaeological evaluation .........................................................................................4
Tranmere Grove ........................................................................................................4
Chesterfield Drive .....................................................................................................9
4 Finds.................................................................................................................................10
4.2 Pottery .....................................................................................................................10
Prehistoric ...............................................................................................................11
Romano-British .......................................................................................................11
Post-medieval ..........................................................................................................12
4.3 Ceramic building material (CBM) ..........................................................................12
4.4 Opus signinum, mortar and wall plaster ..................................................................12
4.5 Coins........................................................................................................................13
4.6 Metalwork ...............................................................................................................13
4.7 Other artefacts .........................................................................................................13
4.8 Animal bone ............................................................................................................13
Condition and taphonomy .......................................................................................14
Animal husbandry....................................................................................................14
Consumption and deposition practice.....................................................................15
4.9 Marine shell.............................................................................................................15
5 Paleao-environmental evidence .....................................................................................15
6 Conclusions and potential ..............................................................................................15
7 Further recommendations .............................................................................................18
8 The archive......................................................................................................................18
References
Appendix 1: Pottery records
Figure 1: Site location plan
Figure 2: Plan of Brown’s 1948-50 excavations
Figure 3: Trench location plan
Figure 4: West facing section across Trench 2
Figure 5: Plan and west facing section (A-B) of Trench 4, and north facing section (C-D) of
hypocaust stoke hole
Figure 6: South facing section across Trench 8
i
CASTLE HILL, IPSWICH, SUFFOLK
ARCHAEOLOGICAL EVALUATION
AND ASSESSMENT OF THE RESULTS
Summary
The archaeological evaluation aimed to establish the position of the villa as recorded by Brown
within the present landscape, to establish its full extent, including the presence or absence of
wings, and to evaluate the present condition of the remains. The work included re-evaluation of
existing plans, geophysical survey and 12 trial trenches, some of which were machine
excavated. The work was undertaken over three days in April 2003.
The evaluation produced no evidence for side wings to the villa, so providing support for
Brown’s contention that it was constructed as a single range. However, attempts to fix his
excavation plan to the present landscape were unsuccessful. The evaluation also confirmed
Brown’s conclusion that the villa had been heavily robbed for building stone in the past. A
number of heavily robbed wall trenches were identified on the eastern side the villa, in one case
associated with a layer of burning possibly representing a collapsed wattle and daub wall, while
the evaluation trenches to the south of the villa produced a low density of Roman artefacts.
The work also provided evidence as to the current condition of the site, revealing that most of
the stratigraphy on the site has been removed by the earlier excavations and that only isolated
patches of undisturbed deposit remain. Most of the west end of the villa has been totally
excavated, although unexcavated features do exist at the base of the site.
ii
CASTLE HILL, IPSWICH, SUFFOLK
ARCHAEOLOGICAL EVALUATION
AND ASSESSMENT OF THE RESULTS
Acknowledgements
The geophysical survey was undertaken by John Gater with staff from G.S.B. Prospection, and
survey by Henry Chapman, University of Hull. The excavation strategy was conducted by
Miles Russell (University of Bournemouth) and site recording was co-ordinated by Phil
Harding assisted by Steve Thompson of Wessex Archaeology. The excavations were
undertaken by the Time Team’s retained excavators and members of the Suffolk and Colchester
Archaeological Units. The archive was collated and all post-excavation assessment and analysis
undertaken by Wessex Archaeology including management (Roland J C Smith), report (Phil
Harding), finds (Lorraine Mepham) and illustrations (Marie Leverett). Specialist comment was
provided by Lisa Brown (pot), Stephanie Knight (animal bone), Chris Stevens (plant remains)
and Nick Wells (Roman coins).
The progress and successful completion of the work also benefited from discussion on site with
specialists of Roman archaeology Guy de la Bedoyere and David Neal.
Finally thanks are extended to all the householders in Chesterfield Drive and Tranmere Grove,
who allowed access for geophysical survey and especially those who granted permission to dig.
iii
CASTLE HILL, IPSWICH, SUFFOLK
ARCHAEOLOGICAL EVALUATION
AND ASSESSMENT OF THE RESULTS
1 BACKGROUND
1.1.2 The site lies on Castle Hill, a suburb built in the 1950s in the northwest of Ipswich at
grid reference TM 147 466 (Figure 1). The site was located towards the crest of a
hill, approximately 35m OD. The geology is mapped as Glacial or Kesgrave Sand
and Gravel (BGS Ipswich (1990) Sheet 207), which makes up the crest of the hill,
and Boulder Clay, which comprises the surrounding area. Castle Hill lies
approximately 9.6km southeast from the nearest known Roman town at Coddenham
(Combretovium), and 4.8km east of the Roman Road that passes through that town
(Pye Road).
1.2.1 Castle Hill has been subject to a number of archaeological excavations in the past.
Mosaics were found on the site in the 1850s. More formal excavations first took
place in 1931 (Moir and Maynard 1933), which were continued in 1948-50 by Basil
Brown due to impending housing development (Figure 2). In 1989 Suffolk
Archaeology excavated further buildings south of the original focus.
1.2.2 The results of the previous excavations indicated that the remains of a Roman villa,
dated from at least the early AD 2nd to the late 4th century, lay immediately below
the south facing crest of the hill. Finds from the excavations, including large
mosaics, suggested considerable wealth.
1.2.3 The earliest phase of the building appeared to have been a post-built structure with
some flint footings added. This was completely reconstructed in the late 2nd century,
but the scant excavation records provide no information of any possible later phases.
A bath house, an aisled building and an industrial building (possibly for corn drying)
were identified south of the villa in the excavations of 1989. These excavations
indicated that the main period of occupation was in the AD 3rd to 4th centuries.
1.2.4 The wide range of finds from the site have included mosaics and tessellated floors,
Samian pottery, finger rings, a seal box and stylae for writing, a jet plaque depicting
ATYS (a deity worshipped in Phrygia, and later throughout the Roman empire),
painted plaster, and various coins (ranging between AD 1st and 4th century); also
bone hair pins, personal accessories such as tweezers, mirror fragments, and
household iron tools; and animal bones, chiefly pig and ox.
1
2 METHODS
2.1 Introduction
2.1.1 A project design for the work was compiled by Videotext Communications
(Videotext Communications 2003), providing full details of the circumstances and
methods, as summarised here.
2.2.1 The project provided an opportunity to collate plans from the previous, disparate
archaeological works on the site undertaken by Reid Moir and Maynard, and Brown,
to establish their location in the present landscape and to investigate the extent and
character of the Roman buildings both on and beyond it. This work could be
achieved using a combination of geophysical survey and excavation.
2.2.2 The unfinished excavations from the 1948-50 period, together with the mosaic finds
from the last century suggested that the villa was larger and wealthier than any
others yet known from Suffolk. The lack of large lavish villas from the region has
been considered to reflect differences in the rate at which the Romano-British
lifestyle was adopted compared to the west or north of the country. However many
significant coin hoards have been found in this part of the country, which allude to
extreme wealth being held in the region.
2.2.3 It was hoped that sufficient deposits might remain intact within the area of the
previous excavations, and beyond, to clarify when the villa was first built, whether
there was any evidence of earlier, Iron Age, activity and how the villa developed
structurally and economically within the area.
2.2.4 The evaluation also provided a condition survey of the site and served as a source of
additional archaeological data to that already held by the Suffolk County Council
Sites and Monuments Record for the future management and any redevelopment
within the immediate area.
2.3.1 The fieldwork strategy began with an extensive ground resistance survey across the
site with the aim of locating the extent of the villa and any outlying buildings, and so
guiding the subsequent location of the evaluation trenches. The effectiveness of the
geophysical survey, however, was inhibited by the fact that survey areas were
restricted to individual garden plots.
2.3.2 All the work was undertaken within private gardens, some of which were well
maintained, so limiting the size and location of the trenches. Where access was
restricted, excavation and backfilling were undertaken by hand, but where access
was possible for a tracked mini-digger, and consent granted, trenches were dug by
machine.
2.3.3 Twelve evaluation trenches of varying lengths were dug after consultation with the
on-site director, Miles Russell and associated specialists. Their precise locations
were determined so as to investigate geophysical anomalies, and to re-examine areas
2
of known archaeological features within the areas of previous excavation in order to
answer the specific aims and objectives of the project design.
2.3.4 Seven trenches were excavated using the tracked mini-digger fitted with a 1m
toothless ditching bucket. The remaining five trenches were hand dug. All machine
work was undertaken under constant archaeological supervision and ceased at the
identification of significant archaeological deposits, or where natural bedrock was
encountered first. When machine excavation had ceased all trenches were cleaned by
hand and archaeological deposits were excavated. Where it could be shown that
archaeological deposits were of recent date, especially those of former backfill,
limited machine excavation continued until in situ deposits or the natural geology
were revealed. No trenches were excavated beyond a depth at which it was
considered safe to enter, in accordance with accepted safety procedures.
2.3.5 A sufficient sample of all deposits was examined to allow the resolution of the
principal questions outlined in the aims and objectives above. Other deposits, some
of which were regarded as archaeologically significant, were recorded, preserved in
situ, sealed by sand and backfilled.
2.3.6 A site code (IPS 421), which is compatible with the system used for all
archaeological work currently undertaken by Suffolk County Council, was allocated
to the site before the evaluation began.
2.3.7 All archaeological deposits were recorded using Wessex Archaeology’s pro forma
record sheets with a unique numbering system for individual contexts. Trenches
were located using a Trimble Real Time Differential GPS survey system. All
archaeological features and deposits were planned at a scale of 1:20 and sections
drawn at 1:10. All principal strata and features were related to Ordnance Survey
datum and a photographic record of the investigations and individual features was
maintained.
2.3.8 All spoil was metal detected by Trevor Southgate and Kelvin Gage, as
recommended by Jude Plouviez of Suffolk County Council.
2.3.9 At the completion of the work all trenches were reinstated using the excavated spoil
from the trenches and turf re-laid or replaced. All artefacts were transported to the
offices of Wessex Archaeology where they were processed and assessed for this
report.
2.3.10 The work was carried out from 8th –10th April 2003.
2.3.11 The excavated material and archive, including plans, photographs and written
records are currently held at the offices of Wessex Archaeology at Salisbury under
the code 52568 where they await to be deposited and curated at the Ipswich Museum
under code IPS 421.
3
3 RESULTS
Details of individual excavated contexts and features, a full geophysical report (GSB
2003) and results of artefact and environmental sample analysis are retained in
archive.
3.1.1 The results of the ground resistance survey were of limited value. After equalising
the data to take into account the differing arrays used and the large number of
disconnected survey areas, it became apparent that a concentration of high resistance
anomalies existed in the northern half of the site. However, although these anomalies
appeared to relate reasonably well with the existing plans of the villa, correlation
with the deposits and features exposed in the Time Team excavations remains
unclear. It is believed, therefore, that earth resistance survey succeeded in mapping
the location of former excavation trenches, rather than the archaeological remains
themselves.
3.1.2 Several areas beyond the main area of the villa, which were initially thought to be of
interest, proved upon excavation to be natural gravel-type deposits or related to
modern landscaping and ground disturbance.
3.2.1 Details of individual excavated contexts and features are retained in archive.
Archaeological features were overlain by mid grey-brown well-sorted friable silty
clay topsoil, 0.20-0.30m thick, that was normally preserved within the turf of
individual lawns. Although the underlying stratification of individual trenches
varied, a distinction can be made between those to the north, in properties along
Tranmere Grove, and those to the south, in properties along with Chesterfield Drive.
In the former, in the main area of the Roman villa, the underlying subsoil
accumulation was heavily disturbed, and included backfill and refuse from former
excavations with fragments of mixed Roman material. In the latter, where there was
less Roman and later activity, the soil profile was generally undisturbed and
represented a subsoil accumulation, up to 0.60m thick, which may have resulted
partially from agricultural activity that preceded housing in the 1950s.
Tranmere Grove
3.2.2 Eight evaluation trenches were located in gardens to the rear of houses in Tranmere
Grove (Figure 3), in the known area of the Roman villa.
Trench 1
3.2.3 A hand-dug excavation was located in the lawn of 21 Tranmere Grove. This garden
was believed to contain a series of intersecting wall alignments as recorded by
Brown. It was hoped that if they could be found it would be possible to plot the
position of the villa more accurately, reconstruct the wall alignments, and reconsider
the phasing of the site. It would also help establish how much of the site Brown had
excavated and predict more reliably where undisturbed deposits might be located.
4
3.2.4 The trench was laid out to cover as large an area as possible in order to maximise the
number and extent of wall lines revealed, in the belief that they would lie close to
the surface. However, it was soon clear that the backfilled overburden was deeper
than anticipated, and the trench was reduced in size to 3.5m east-west and 3.0m
north-south.
3.2.5 No wall lines were identified in the trench. The thin (0.10m) topsoil overlay a mid
grey brown sandy clay layer (101), 1.25m deep containing fragments of re-deposited
stone, mortar, Roman brick and tile, but also modern metal fragments and tin
sheeting. This deposit was excavated to a depth of approximately 0.50m below the
turf line across the entire evaluation trench, but was then excavated in an ‘L’ shaped
slot, 0.80m wide, to reduce the quantity of material to be removed.
3.2.6 The natural clay was exposed along the north edge at 34.97m OD but to the south
the backfill continued into the top of a large, possibly square, pit (103), which
contained in situ Roman deposits. This feature was filled with yellow brown sandy
silt (105) and contained fragments of Roman tile. The upper 0.40m of this feature
were removed, but due to insufficient time and the fact that the slot was too narrow
and deep to be excavated safely, excavation of the trench was terminated. The
presence of the pit demonstrated, however, that undisturbed deposits remain within
the area of Brown’s excavation, although they lie at a considerable depth.
Trench 2 (Figure 4)
3.2.7 This trench, measuring 9m long and 1m wide, was hand dug through the lawn of 17
Tranmere Grove. Its purpose was to locate projected wall lines, indicated on
Brown’s plan, on the same line as those anticipated, but not found, in Trench 1
(above).
3.2.8 The excavation revealed 0.44m of dark brown silty topsoil (201) overlying
archaeological deposits. An east-west linear feature (205), 1.5m wide and 0.43m
with moderately sloping sides and a rounded base, filled with brown loam (204),
crossed the trench c. 3m from its south end in the approximate position of an
anomaly detected by geophysical survey. It was visible immediately below the
modern turf line and cut through a number of underlying archaeological deposits. It
is likely to date from the 1950s, although the absence of backfilled demolition rubble
makes it impossible to determine whether it represented the earlier excavation of a
former robbed wall trench.
3.2.9 There was a clear distinction in the character of the deposits to the north and south of
the feature. Those to the north comprised a compact layer of dark brown silty loam
(202), approximately 0.20m thick, containing fragments of mortar, overlying a layer
of mid brown silty loam (209), approximately 0.16m thick. These deposits were
revealed in the edge of ditch 205 and recorded but were not excavated.
3.2.10 The three layers to the south, however, which were sampled in a 1m square test pit at
the south end of the trench, provided possible evidence of the abandonment and
demolition of the villa (although none about its construction). A thin layer of yellow
brown sandy loam (203), 0.08m thick, overlay green-brown clay loam (206), 0.12m
thick, containing relatively large quantities of domestic refuse. This layer capped an
irregular mortar surface (207), possibly a demolition layer, containing further
5
domestic refuse. A basal layer of green grey sandy clay (208), approximately 0.20m
thick, overlay the natural gravel.
Trench 2A
3.2.11 Trench 2A lay southeast of Trench 2 in the same garden plot and was positioned to
relocate and trace the extent and alignment of a fragment of flint wall foundation
discovered accidentally during the erection of the property boundary fence. The
excavation, which measured 3m east-west by 1.6m north-south, revealed the wall
(213) immediately beneath the turf. It measured 0.50m wide and extended 2m west
from the fence before turning north. It was constructed of unmodified flint nodules,
approximately 0.10m long, in a compact yellow mortar. No attempt was made, in the
time available, to establish the depth of the wall foundation.
Trench 4 (Figure 5)
3.2.12 A machine dug trench, 16m long and 1m wide was dug in the garden of 13
Tranmere Grove. It was aligned north-south and was positioned to investigate an
area of high resistance recorded in the south of the plot, as well as to provide a long
transect across much of the main block of the villa.
3.2.13 The excavation revealed that this part of the site had also been heavily robbed, as
well as lying within an area of previous excavation. Many of these previous
interventions could not be dated accurately, although they probably include some
dug in antiquity when the site provided a source of building stone, and other more
recent investigations by antiquarians and early archaeologists (such as feature 424).
Some, such as 430 and 444, were cut from either immediately or just below the
present topsoil through deposits (407, 423 and 443) containing modern material,
including metal sheeting, into features that had already been robbed.
3.2.14 At the north end of the evaluation trench, there was a recently backfilled robber
trench (426), possibly the product of ‘wall chasing’ by Brown or other early
excavators. It was 0.60m deep and approximately 1.5m wide at the base, although it
had been extended north by a further 0.80m where, on the east side, it cut through a
layer of Roman demolition rubble (427, below). Its position corresponds
approximately with the line of a former wall shown on Brown’s site plan, and its
east-west alignment reflected the general orientation of other wall lines attributed to
the villa.
3.2.15 In the south side of this feature was exposed the entrance to a channel hypocaust
(448), the western side of its stoke hole being marked by a stack of mortared tiles
(438), and the natural clay bedrock (428) near its mouth being heavily heat-affected.
There was a deposit of ashy material (439) from the stoke hole on the east side of the
evaluation trench, and this was partly overlain by a layer of large flint nodules and
Roman tile fragments (427) extending to the north, which may relate to the
demolition of the east wall of the hypocaust.
3.2.16 The hypocaust itself, which was cut 0.2m through subsoil into the natural clay, was
approximately 0.60m wide with a mortared floor defined by two parallel walls,
0.20m across. The wall exposed on the east side (418) was constructed of mortared
flint nodules. The hypocaust extended 2m southwards to a robbed wall trench (447),
and its fill, which was not removed, comprised light grey brown sandy silt (417),
with mixed flecks of mortar, and pieces of flint and tile. Traces of two clay floors,
6
each 0.05m thick and both heat affected, were preserved over the hypocaust wall on
the west side. The lower floor (416/419), a light yellow brown with patches of pink,
was separated from the upper (421) by a thin (0.03m) layer of burnt material (406),
which contained seven Roman coins of mid to late 4th century AD.
3.2.17 The wall at the south end of the hypocaust had been robbed, possibly in antiquity,
the series of deposits in the area suggesting several phases of robbing activity (447
and 442), as well as being re-excavated in recent times (444). It was not possible in
the time available to fully excavate these deposits in order to establish the precise
relationships between the different phases of robbing.
3.2.18 Immediately to the south, a possible corridor, 2m wide and defined by fragments of
two walls (440 and 432), ran east-west through the apparent centre of the villa. On
its north side, wall 440 (in wall trench 441) had been partially robbed and survived
only 0.35m wide and 0.19m deep. It comprised approximately four courses of
mortared chalk or limestone blocks. Wall 432, to the south, had been ‘sectioned’ in a
previous excavation and survived only on a narrow ‘baulk’ across the excavated wall
trench (430), where it was 0.56m wide, 0.25m deep and constructed of flint.
3.2.19 The construction trenches for these walls were cut through the subsoil to the surface
of the natural clay. A veneer of charcoal-rich silt (422), 0.03m thick, which overlay
the subsoil, was preserved in remnant patches between the robber trenches and
excavations, but all stratigraphic relationships with the walls had been destroyed by
previous activity, so that it could not be determined how the burning related to the
phasing of the villa. (A similar layer of burning was found to the east in Trench 10,
below.) The apparent absence of any occupation, construction or demolition material
beneath the burning suggests that it may relate to Brown’s phase of timber buildings,
which preceded the construction of the stone villa.
3.2.20 The east end of wall 432 intersected, at the edge of the excavation, with a heavily
robbed wall trench (435) aligned approximately north-south. Only remnant
fragments of mortar and flint (436) of this major structural wall survived at the base
of the wall trench, which was 0.57m below the surface of wall 432.
3.2.21 The south end of the evaluation trench crossed an area of high resistance detected in
the geophysical survey. The excavation showed that the anomaly represented a
feature approximately 4m across (408), possibly a cellar or a large pit, the limited
investigation at its south end suggesting a depth of least 2m. The lower fill (434),
most of which was not removed, comprised mid brown silty clay mixed with
demolition rubble including tile, mortar and stone fragments. The upper fill (407)
was of a similar composition, but contained less rubble and may have been
redeposited backfilled material from a former excavation; the upper layer extended
beyond the feature in both directions and was cut by robber trench/excavation 430.
3.2.22 An additional shallow ditch (402), which cut layer 407, ran approximately east-west,
but at a slight angle to the axis of the villa, at the south end of the trench. It
measured 0.70m across and 0.35m deep with moderately sloping sides and a flat
base, and was filled with mid brown garden loam (403) with fragments of pot and
tile. There was no demolition rubble to suggest that it represented the previous
excavation of a robbed wall line.
7
Trench 4A
3.2.23 A small trench, approximately 2m long and 1m wide, was dug by machine to the
east of the southern part of Trench 4, in the same garden plot, in order to attempt to
establish the eastern extent of cellar/pit 408 (above). This was not found. Instead, the
excavation revealed 0.81m of disturbed ground, similar to that capping Trench 4,
above the natural clay. Two clay filled pits, 409 and 411, were exposed cutting the
natural in the base of the trench. Pit 409 measured approximately 0.70m in diameter
and was 0.55m deep with sloping sides and a rounded base. It was cut by pit 411,
most of which lay under the baulk on the south side.
Trench 8 (Figure 6)
3.2.24 A small area excavation, 4m north-south and 3m east-west, was stripped by machine
in the garden of 7 Tranmere Grove to locate any evidence of the eastern part of the
villa, or to define alternative Roman activity beyond its eastern side. The results
showed a low level of archaeological activity. A slightly curving ditch (803), 0.45m
wide and 0.47 deep with steep sides and a rounded base, was revealed immediately
below the topsoil. It was aligned approximately north-south, curving slightly
eastwards towards the north, and was filled with mid brown silty soil containing
Roman pot, tile and bone. It was cut into an underlying well sorted, moderately
compact, dark brown gravely silt ploughsoil/subsoil (804) which produced Roman
pot, hypocaust tile, tile and bone. This layer sealed a second ditch (806), 1m west of
and parallel to ditch 803. This was 0.43m wide and 0.20m deep with a similar, but
shallower, profile to ditch 803 and a similar fill. It contained fragments of Roman
pottery of 1st-2nd century AD date.
3.2.25 The function of these ditches is unclear, although they are more likely to represent
boundary ditches rather than timber beam slot foundations. The presence of early
Roman pottery adds weight to Brown’s conclusion that occupation on the site dates
from the 1st century AD. Ditch 803 is assumed to be of Roman date, although there
is no way of knowing whether the feature is intrusive and contains residual material.
The absence of covering deposits may indicate that the surface has been truncated by
ploughing or landscaping during house construction.
Trench 9
3.2.26 This 1m wide machine-excavated trench in the lawn of 11 Tranmere Grove
measured 3m east-west and was positioned in an attempt to locate walls in the
eastern part of the Roman villa. A modern soak-away was discovered beneath the
topsoil and the trench was abandoned.
Trench 10
3.2.27 Another 1m wide machine-dug trench, measuring 4.5m east-west, was positioned
spanning the gardens of 11 and 13 Tranmere Grove, between Trenches 4 and 9,
again with the purpose of locating walls in the eastern part of the Roman villa.
Approximately 0.40m of topsoil and subsoil overlay a series of well-preserved, in
situ layers apparently relating to the demolition of the villa.
3.2.28 The east edge of a feature (1004), which was thought to be a robber trench, was
aligned north-south across the evaluation trench, the west edge being not visible. It
turned east (1005) and ran parallel to the north edge of the evaluation trench. The
upper fills of these features (1003, 1006) were dark brown with quantities of ceramic
building material, mortar and possible fragments of opus signinum or plaster. Layer
8
1003, in feature 1004, overlay a similar deposit (1008), containing large quantities of
tile, brick and charcoal that appears to have been derived from a collapsed roof that
had burnt in situ. The west end of the evaluation trench was also characterised by
very dark brown/black material with areas of red fired clay. A zone of alternating
thin bands of charcoal and fired clay suggested that this part of the evaluation trench
contained evidence of a collapsed wattle and daub wall (1009).
3.2.29 The quality of the preserved deposits in this evaluation trench, their rarity elsewhere
on the site and the small size of the excavated area made it preferable to restrict the
excavation and to preserve the deposits for more detailed systematic future research.
The evaluation trench was therefore photographed, described and backfilled, sealing
the collapsed wall beneath a layer of sand.
Chesterfield Drive
3.2.30 Four evaluation trenches were located in lawns to the rear of properties in
Chesterfield Drive. They were positioned in order to examine the possibility that the
villa might extend to the south. However, no evidence for the villa was found and
very little additional archaeological evidence recovered.
Trench 3
3.2.31 This machine excavated trench, 4m long aligned north-south and 1.8m wide, was
excavated in the lawn of 14 Chesterfield Drive in order to investigate a strong linear
geophysical anomaly that was thought might mark the line of an east wing of the
villa. However, no archaeological features and no evidence of any wing was present.
3.2.32 The topsoil horizon overlay a grey-brown silty clay subsoil, 0.50m thick, which
became increasingly more gravely and lighter in colour towards the natural gravel
(303) at the base. The subsoil (302, 304) showed no clear stratification, although a
horizon containing fragments of Roman roof tile and pottery was clearly present
near the base. The natural gravel (303) at the base of the trench included linear
structures containing chalk. It is likely that these geological features represent
periglacial involutions within the upper part of the gravel and may in some way be
responsible for the strong geophysical anomalies detected in the survey, although
they occurred at the limit of the geophysical detection (Gater pers. comm.).
Trench 5
3.2.33 A hand dug trench, 3m long aligned east-west and 1m wide, in the lawn of 18
Chesterfield Drive produced a similar stratigraphic sequence to that recorded in
Trench 3, comprising topsoil (501) overlying a subsoil (502), 0.58m thick, which lay
on natural gravel (503). No archaeological features were recorded.
Trench 6
3.2.34 To the north of Trench 5 in the same garden plot, a trench of similar dimensions was
aligned approximately northeast-southwest. Again, the stratigraphic sequence was
almost identical, although a lens of refuse, 0.03m thick containing Roman tile,
animal bone and oyster shell extended from the baulk across the southwest end of
the trench.
Trench 7
3.2.35 A machine dug trench, approximately 3m long aligned north-south and 1m wide,
was excavated to the rear of 32 Chesterfield Drive, at the location marked by the
9
Ordnance Survey as ‘Roman Villa (Site of)’. The section revealed, below the topsoil
(701), a layer of yellow brown sandy silt subsoil (702), 0.25m thick, capping a large
void filled with modern rubbish. This intrusion had penetrated and removed any
archaeological stratigraphy and the trench was abandoned.
4 FINDS
4.1.1 Finds were recovered from eight of the ten trenches excavated; no finds were
recovered from Trenches 7 or 9. The assemblage comprises mainly bulk finds, with
a smaller proportion of individually recorded Objects (‘small finds’), mainly
metalwork. All finds have been cleaned (with the exception of the metalwork) and
have been quantified by material type within each context. There is also a register of
individual Objects. Quantified data form the primary finds archive for the site and
these data are summarised by trench in Table 1.
4.1.2 Subsequent to quantification, all finds have been at least visually scanned in order to
gain an overall idea of the range of types present, their condition, and their potential
date range. Pottery and ceramic building material have been subjected to more
formal scanning, including quantification by ware group/type (details below). Spot
dates have been recorded for selected material types as appropriate. All finds data
are currently held on an Excel spreadsheet.
4.1.3 This section presents an overview of the finds assemblage, on which is based an
assessment of the potential of this assemblage to contribute to an understanding of
the site in its local and regional context. The assemblage is largely of Romano-
British date (late AD 1st to 4th century), with small quantities of prehistoric and post-
Roman material).
4.2 Pottery
4.2.1 Pottery and coins provide the primary dating evidence for the site. The
overwhelming majority of the pottery assemblage consists of Romano-British
10
material, but there are also very small quantities of later prehistoric and post-
medieval sherds. The whole assemblage has been quantified, within each context, by
broad ware group or known type (e.g. coarse greywares, samian). The presence of
identifiable vessel forms and other diagnostic features has been recorded, along with
spot dates. Summary totals by ware group are presented in Table 2. Details by
trench and context are presented in Appendix 1.
Prehistoric
4.2.2 The assemblage included a single prehistoric sherd in a coarse sand-tempered fabric
with flint and quartz inclusions, probably dating to the Early or Middle Iron Age.
Romano-British
4.2.3 The Romano-British assemblage consists of 133 sherds weighing 1032 grammes and
is dominated by coarse greywares. The few diagnostic sherds present suggest a late
Roman date. These include a small necked globular bowl similar to a type found in
fourth century levels at Portchester and a number of late black-burnished derived
forms. A proportion of the greywares are probably products of the Hadham kilns,
manufactured alongside a range of oxidised, generally red-slipped, wares at Little
Hadham and Much Hadham in Hertfordshire from the mid AD 3rd century but
distributed more widely during the 4th century. Four sherds of Hadham oxidised
ware were recovered. The two small sherds of rilled South Midlands shell-tempered
ware, produced at Harrold (Beds.) and possibly Lakenheath (Suffolk), would also
have reached the site during the 4th century.
11
4.2.4 The small range of finewares provides evidence for somewhat earlier activity on the
site. Colchester colour-coated ware and Trier black-slipped ware (‘Moselkeramik’)
were in production up to the mid to late 3rd century, and Nene Valley colour-coated
ware up to the end of the 4th century. The two amphora sherds, one from Gaul, the
other unsourced, could also belong to a 3rd century phase of occupation. The single
Colchester samian vessel (a copy of a Drag 35 cup), however, is the product of an
industry which operated only during the 2nd century and a Central Gaulish samian
sherd is of similar or earlier date.
Post-medieval
4.2.5 The remaining 32 sherds are of post-medieval date and comprise coarse redwares,
which are not closely datable within the post-medieval period, and modern industrial
wares (19th/20th century).
4.3.1 This category includes fragments of brick and tile. The assemblage has been
quantified by type within each context, and this information is summarised by trench
in Table 3.
4.3.2 The overwhelming majority of the fragments are of Romano-British date, and
include identifiable imbrex and tegula roof tiles, box flue tiles and tesserae, the latter
re-used from tiles. No attempt has been made at detailed fabric analysis at this stage,
but the scan showed that several visually distinct fabric types are present,
presumably indicative of the exploitation of more than one source for the CBM.
4.3.3 A small quantity of post-medieval CBM was also recovered, mainly from topsoil
contexts, comprising modern brick fragments, and two pantiles.
4.4.1 Other building material was recovered in the form of opus signinum (concrete-like
substance used to line walls, floors and tanks), wall plaster and the mortar backing to
the plaster. The wall plaster includes monochrome white and red fragments as well
as two polychrome fragments. There is insufficient evidence on which to base any
reconstruction of colour schemes or designs.
12
4.5 Coins
4.5.1 Ten coins were found ranging in date from the late 1st century BC to the late AD 4th
century. The earliest of the coins (Object 2) is of potin (a tin rich copper alloy) and
although of Late Iron Age date could have circulated into the Roman period – but
certainly not much after AD 64. The remainder of the coins are copper alloy – all but
one (Object 3) types dating from the mid to late 4th century AD. The exception is a
fine but worn sertertius of Faustina the Younger (wife of Marcus Aurelius) struck
between AD 161 and 180.
4.5.2 Table 4 provides a basic spot date for each of the coins, listed in order of Object
number;
4.6 Metalwork
4.6.1 Other metalwork includes objects of copper alloy (2), lead (32) and iron (34). None
are typologically distinctive, and none are of definite Romano-British date; most if
not all are likely to be post-medieval. Identifiable objects include iron nails and other
structural items, and a possible copper alloy buckle fragment. The lead consists
entirely of waste/offcuts. All the metalwork has been X-radiographed and this record
is contained within the project archive.
4.7.1 Other artefacts recovered, all in small quantities, comprise vessel glass (one small
Romano-British fragment from 603, with the rest being modern bottle/jar and
window glass), clay pipe stem fragment (post-medieval), and worked flint flakes
(two prehistoric, two possibly from Romano-British or later walling). The two
pieces of shale were both from armlets, worn on the wrist or arm of adults and
children alike as personal ornaments (jewellery). Both are plain, D-shaped in cross-
section and were made using a lathe, most probably in the Wareham/Poole Harbour
region of Dorset. Although both almost certainly belong within the Roman period
(AD 1st – 4th centuries), they cannot be dated any more precisely.
4.8.1 The potential of the assemblage to provide information about husbandry patterns,
population structures and consumption practices was ascertained from the number of
bones that could give information on the age and sex of animals, butchery, burning
13
and breakage patterns. The numbers of bone that could provide metrical information
were also counted. The extent of mechanical or chemical attrition to the bone surface
was recorded, with 1 indicating poor condition, 2 fair and 3 good. The numbers of
gnawed bone were also noted. Conjoining fragments that were demonstrably from
the same bone were counted as one bone in order to minimise distortion. No
fragments were recorded as ‘medium mammal’ or ‘large mammal’; these were
instead consigned to the unidentified category.
4.8.2 Of the fifteen contexts from which animal bones were recovered, only four could be
dated to the occupation of the villa (406, 602, 603 and 604), and another three to the
period immediately after the villa was abandoned (203, 206 and 208). The remainder
comprised robber trenches, topsoil and the fill of robber trenches. Unfortunately
numbers were not large enough to enable phases or feature types to be assessed
separately in any detail.
4.8.3 A total of 107 animal bones were recovered, of which 45 (42%) could be identified.
Of the identified bone, a fairly high proportion, 56% (25 bones) was from the
occupation layers and 7% (3 bones) from the period of abandonment.
4.8.4 83% of the fragments were in fair condition (some abrasion or root etching present),
and 17% were in poor condition (bone surface obscured). Gnawing was noted on 17
bones (16%). The assemblage is therefore in reasonably good condition.
Animal husbandry
4.8.5 Cattle were the most commonly represented species, followed by horse (Table 5).
Pig, deer and bird (probably chicken) were also present, as were sheep/goat to a
lesser extent. There were no positive identifications of goat. Deer were represented
by postcranial elements, indicating that they were exploited for meat. In addition,
numerous small mammal bones and some fish bones and scales have been recovered
from the bulk samples. The full range of animals were found in the occupation
deposits, suggesting a relatively wide range of species were exploited, and that the
robber contexts probably contain mainly redeposited Roman material, thus they have
not distorted the species list.
4.8.6 The potential for reconstructing husbandry practice is limited: only 17 bones (16%)
provided evidence for age, one (1%) had pathological bone growth and 8 (7%) could
provide useful metrical data.
14
Consumption and deposition practice
4.8.7 Five bones (5%) had been marked by chops or showed evidence of fracture whilst
fresh for marrow extraction, but there is very little that can be said about
consumption and deposition from such a small sample.
4.9.1 The small quantity of shell consists entirely of oyster, and includes both left and
right valves, i.e. both preparation and consumption waste.
5 PALAEO-ENVIRONMENTAL EVIDENCE
5.1.1 A single bulk sample of 28 litres was taken from burnt layer 406 probably dating, on
coin evidence, to the 4th century AD. The sample was processed for the recovery and
assessment of charred plant remains and charcoals to see if this could help
understand the burning event.
5.1.2 The bulk sample was processed by standard flotation methods. The flot was retained
on a 250Pm mesh and the residues fractionated into 5.6mm, 2mm and 1mm fractions
and dried. The coarse fractions (>5.6mm) were sorted, weighed and discarded. The
flot was scanned under a x10 - x30 stereo-binocular microscope and presence of
charred remains quantified in order to present data to record the preservation and
nature of the charred plant and charcoal remains.
5.1.3 Relatively little charred plant material other than wood charcoal was recovered.
Most of the remains were fragments of hazelnut shell (Corylus avellana). Two
cotyledons of pea/bean were also found (Pisum sativum/Vicia faba) more probably
the former than latter. No other remains were present. Charcoal was noted from the
flot. Land snails were also noted. About 50 plus shells were present, and covered
open country, Helicella itala and Vallonia spp.; shaded conditions, Carychium spp.,
Discus rotundatus, Oxychilus/Aegopinella spp. and catholic species, Cochlicopa spp.
Small mammal bones, fish bones, as well as scales, were also noted.
5.1.4 The charred plant remains are unusual in some respects in that cereal remains are
totally absent. The presence of only hazelnut remains, would tend to indicate that
either cereals were brought to the site fully processed e.g. as cleaned grain ready for
milling, flour or perhaps bread/barley or that such activities were absent from the
site. The presence of hazelnuts does indicate the probable utilisation of this species.
Hazelnuts have been collected for food on a reasonable scale since prehistoric times;
their presence may indicate their use as a snack food.
15
unfinished 1948-50 excavations indicated that the building was occupied mainly
during the 3rd to 4th centuries AD, but with its origins in at least the early 2nd century
AD, and with finds indicating 1st century activity on the site.
6.1.2 The evaluation sought to define more precisely the extent, location and character of
the villa, to add to the understanding of its structural development, and of its place
within the local and regional economy, and to assess the nature and condition of the
surviving archaeological deposits in order to inform future decisions on the
management and development of the site.
6.1.3 Attempts to trace the extent of the villa using geophysical survey were largely
unsuccessful. So was the attempted matching of archaeological features recorded in
the trenches with Brown’s plan of the building, due mainly to the plan’s apparent
inaccuracy and the likelihood that Brown had failed to record the positions of all
walls. However, the evaluation appeared to confirm Brown’s conclusion that the
building consisted of a single range, there being no indications in any of the trenches
that the villa had wings at its east and west ends.
6.1.4 Dating evidence, provided by coins and pottery, have provided additional
information on the chronology and nature of the site, although it should be noted that
most of coins came from a single context (406), the layer of burnt material between
the floors above the channel hypocaust in Trench 4. Moreover, the dating potential
of the pottery is limited by the small size of the assemblage, and by the
predominance of coarsewares, most of which are not closely datable. Nonetheless,
the finds confirm that the site was occupied from the 1st to 4th century AD, with the
most concentrated activity occurring between the late 3rd and 4th century AD, a phase
broadly contemporary with the use of the bathhouse and aisled building uncovered
to the south by the Suffolk Archaeological Unit in 1989.
6.1.5 However, the quantity and range of other material types was not great, and because
of the disturbed nature of many of the contexts their potential for further analysis is
correspondingly limited. Some structural information can be gained from the
ceramic and stone building material, but insufficient to warrant further analysis, and
no evidence was found for on-site activities such as grain processing or textile
making, etc.
6.1.6 Although the structure of the animal bone assemblage from disturbed contexts (such
as robber trenches and the backfill from earlier excavations) may have been distorted
by fragmentation and modern additions, it probably reflects the species present in
securely dated Roman deposits, and the exploitation of a range of species is
interesting and suggests a high status site. However, the small size of the assemblage
means that the potential for reconstructing aspects of animal utilisation, such as kill
patterns, breed types or butchery practices, is very limited. A quantity of small
mammal bones, the majority of them possibly from a single rabbit, were recovered
from burnt layer 406 in Trench 4. Unless there is evidence of post-depositional
bioturbation, these date to the period of the villa’s occupation, but they have no
potential for indicating the surrounding environment. The small number of fish
bones from the same context could be from a local source or could indicate trade.
16
6.1.7 In general, the range of finds recovered is consistent with that expected from a
relatively high status site, supplied with goods from various local, regional and
continental sources.
6.1.8 The environmental materials recovered provide data also relating to the same single
context (406), and the presence of moderate quantities of charcoal may enable the
nature of the burning to be determined. The small amount of charred plant remains,
however, leaves little potential for further investigation. The land snails present are
typical of open and garden habitats, and there is little potential for them to add any
further significant information.
6.1.9 The evaluation has provided a good indication as to the current condition of the
monument and level of preservation of the surviving archaeological deposits. While
it is clear that the site has been damaged both by stone robbing in antiquity and by
the previous excavations, archaeological deposits survive at three levels.
6.1.10 Firstly, there are small areas of intact deposit, these increasing in frequency towards
the eastern end of the villa; for instance, parts of the floor level appear to have been
preserved in Trench 4. However, these are limited in extent and depth, some of them
having been reduced to small patches of material with no stratigraphic relationship
to other contexts or to the structure of the building. There is limited surviving
masonry in Trenches 2A and 4 (although the full depth of foundations were not
established) reflecting the fact that the site has been extensively robbed, and
suggesting also that the upper levels may have been truncated by ploughing or
landscaping. Other surviving deposits included a small area of a channel hypocaust,
and some well preserved burnt levels that appear to relate to the demolition of one
phase of the villa. As the stone robbing cut through these layers it is not possible to
be sure whether construction levels also cut through them.
6.1.11 Secondly, some of the site contains areas of intact robbed deposits, although the
evaluation did not attempt to examine much of this material, and it has been
impossible to date any of the robbing episodes.
6.1.12 Finally, there are deposits resulting from previous excavations of the site, both
recorded and potentially unrecorded. Large parts of the upper levels of the site were
covered by this material, including most of the upper levels of Trench 4.
6.1.13 On the basis of the small sample of the site evaluated, and the lack of reliable results
from the previous excavations, the above conclusions are necessarily provisional.
There remain significant areas of uncertainty as to the precise location of the
features uncovered by Brown, the layout and extent of the villa, and the history of
construction and development on the site over four centuries. However, providing
answers to these questions has certainly been aided by the identification of
significant features, deposits, and finds. Combined with a better understanding of the
current condition of the site, this will facilitate the future systematic research and
management of an important site.
17
7 FURTHER RECOMMENDATIONS
7.1.1 The project has been successful in provide useful data on the survival and condition
of the Roman villa at Castle Hill and some information towards confirming its
development. Further detailed analysis of all the results of this project, however, are
not considered to be appropriate in view of the limited excavation of stratified
deposits, their disparate character and dispersed nature, and the limited size and
potential of the finds and environmental data. Some limited further work is
proposed, however, and is set out below.
7.1.2 A basic archive level of finds recording has already been achieved, and where
appropriate this meets minimum recording standards (e.g. SGRP 1994).
7.1.3 Other proposed work concerns the requirements for long-term curation of the finds
archive. Most of the artefacts are in a stable condition and have been packaged in
accordance with national guidelines on the preservation of artefactual archives for
long-term curation (e.g. UKIC 1983). The metalwork, however, is inherently
unstable and this has certain implications for long-term curation. It is recommended
that conservation treatment, involving partial or total cleaning by a skilled
conservator, is undertaken for at least the majority of the coins.
7.1.4 While the environmental sample is of some interest, the general lack of other
contextual evidence from the site suggests that there is little value at present in
undertaking further analysis than that set out in this report. The results of this
assessment should be noted for further work that may be conducted at this site.
7.1.5 This report will be deposited with the Suffolk Sites and Monuments Record so as to
be available to future researchers. A summary report will be distilled from this report
and published as a note in the Proceeding of the Suffolk Institute of Archaeology &
History.
8 THE ARCHIVE
8.1.1 The archive, which includes all artefacts, and all written, drawn and photographic
records relating directly to the investigations undertaken, is currently held at the
offices of Wessex Archaeology under the site code IPS421 and Wessex Archaeology
project code 52568. It is intended that, in accordance with the wishes of the
landowner, the excavated material and records will eventually be deposited and
curated at the Ipswich Museum, and a copy of the archive sent to Suffolk County
Council Archaeological Service.
18
References
GSB, 2003, Castle Hill, Ipswich. Unpublished Client Rep. 2003/27. GSB Prospection.
Moir, J.R. and Maynard, G., 1933, ‘The Roman villa at Castle Hill, Whitton, Ipswich’.
Proceedings of the Suffolk Institute of Archaeology & History XXI Part 3, 240-262
SGRP, 1994, Guidelines for the Archiving of Roman Pottery, Study Group for Roman Pottery
Guidelines Advisory Document 1
UKIC, 1983, Packaging and storage of freshly excavated artefacts from archaeological sites,
United Kingdom Institute for Conservation, Conservation Guidelines No. 2
19
Appendix 1: Pottery records
20
Tr./Cxt. Fabric Vessel type No Wt ST Dec/S MT Date / comments
Tr6/602 Coarse greyware Bowl 7 47 WT RB / 2 probably
Hadham
Shell-tempered Bowl 2 2 Rilled WT C4/ South
Midlands type
(Harrold)
Hadham oxidised Bowl 1 25 SL WT C3-C4
Industrial ware 1 3 Post-med
Tr6/603 Coarse greyware Bowl 1 20 WT RB
Tr6/604 Coarse greyware Bowl 2 18 WT RB
Jar with everted, 1 9 WT Late RB?
triangular rim
Nene Valley c-c Beaker base 7 104 C-C Roulette + white WT C3-C4
paint
Tr8/802 Coarse greyware Bowl 6 82 WT RB
Flint-tempered Bowl 1 10 HM Later prehistoric
(probably E-MIA)
Tr8/804 Coarse greyware Bowl 2 12 WT RB/ probably
Hadham
Tr10/1007 Coarse greyware Bowl 1 17 WT RB
21
Supplied by Time Team Licence number AL 100018665
Date: 16/06/03 Revision Number: 0
Digital data reproduced from Ordnance Survey data © Crown Copyright 2003. All rights reserved. Reference Number: 100020449.
This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Scale: N/A Illustrator: MCL
Wessex
Archaeology Path: X:\projects\52568\...\Castle Hill Ipswich\Figures\Figures.dwg
Topsoil
203
206
202 207
204
209 208
205
This material is for client report only © Wessex Archaeology. No unauthorised reproduction.
D
440 421 B
NATURAL CLAY
A
434 422 430 418 416/419
408
402 422 424 441
432 426
428
417
435 436
433
439 427
C
416/419
0 1 2m
Plan at 1:50
430
A
TOPSOIL
438
417
439
428
0 0.5m
Section at 1:10
Plan and west-facing section (A-B) of Trench 4, and north-facing section (C-D) of hypocaust stoke hole Figure 5
W E
Topsoil
804 804
802
807
805
806
803
This material is for client report only © Wessex Archaeology. No unauthorised reproduction.