Wittenham Clumps
Wittenham Clumps
Wittenham Clumps
Oxfordshire
Archaeological Evaluation and an
Assessment of the Results
Wessex Archaeology
Ref: 52568.09
March 2004
ROUND HILL, WITTENHAM CLUMPS,
OXFORDSHIRE
Prepared for:
By:
Wessex Archaeology
Portway House
Old Sarum Park
SALISBURY
Wiltshire
SP4 6EB
© Copyright The Trust for Wessex Archaeology Limited 2004, all rights reserved
The Trust for Wessex Archaeology Limited, Registered Charity No. 287786
ROUND HILL, WITTENHAM CLUMPS,
OXFORDSHIRE
Contents
Summary ................................................................................................................. 4
Acknowledgements ................................................................................................. 5
1 BACKGROUND ........................................................................................... 6
1.1 Description of the site .................................................................................... 6
1.2 Previous archaeological work ....................................................................... 7
2 METHODS .................................................................................................... 8
2.1 Introduction.................................................................................................... 8
2.2 Aims and objectives ....................................................................................... 8
2.3 Fieldwork methods......................................................................................... 9
3 RESULTS .................................................................................................... 10
3.1 Introduction.................................................................................................. 10
3.2 Geophysical survey ...................................................................................... 10
3.3 Archaeological evaluation............................................................................ 11
3.4 Trenches 1, 2 and 3 ...................................................................................... 11
3.5 Trench 4 ....................................................................................................... 12
3.6 Trench 5 ....................................................................................................... 12
3.7 Trench 6 ....................................................................................................... 15
3.8 Trenches 7, 8 and 10 .................................................................................... 16
3.9 Trench 9 ....................................................................................................... 18
3.10 Trench 11 ..................................................................................................... 19
4 FINDS .......................................................................................................... 20
4.1 Introduction.................................................................................................. 20
4.2 Pottery .......................................................................................................... 20
4.3 Ceramic building material............................................................................ 22
4.4 Opus Signinum, wall plaster and mortar ...................................................... 22
4.5 Stone ............................................................................................................ 22
4.6 Coins ............................................................................................................ 22
4.7 Metalwork .................................................................................................... 25
4.8 Marine shell.................................................................................................. 25
4.9 Other finds.................................................................................................... 25
4.10 Animal bone................................................................................................. 25
2
6 DISCUSSION .............................................................................................. 30
9 REFERENCES............................................................................................. 35
3
ROUND HILL, WITTENHAM CLUMPS,
OXFORDSHIRE
Summary
The project was designed to provide archaeological information about the area
surrounding the Iron Age hill fort of Castle Hill and to supplement the results of work
currently being undertaken on the hill fort jointly by the Northmoor Trust and Oxford
Archaeology.
The work, which involved the use of an extensive geophysical survey supplemented by
trial trenching, aimed to examine any archaeological remains on Round Hill and in the
area to the south-west. Previous archaeological work in 1947 had established that a
Roman building and well-preserved Iron Age deposits were present in this area.
The project was restricted in extent by woodland on Round Hill and was unable to
produce evidence for any activity on the hill top. The results of the work on the lower
slopes, however, were able to place the findings of 1947 in a more precise context. The
geophysical survey indicated that this area contains the remains of a large open
settlement, characterised by dense groups of pits. Evaluation demonstrated that these
pits were of Early Iron Age date and that well-preserved, stratified remains of
contemporary Early Iron Age occupation were sealed beneath plough soil derived from
higher up the slope. A number of post-holes may indicate the presence of timber
buildings. Small square enclosures, which also appeared to be of Early Iron Age date,
were identified below the slopes of Castle Hill. The evaluation produced a small
assemblage of Early Iron Age pottery with sheep/goat and cattle bones that provided
evidence of husbandry and meat processing.
The work has established that there appears to be no evidence for Middle and Late Iron
Age occupation and suggests that the land may have been cultivated at this time.
The spread of Roman demolition material identified in 1947 was relocated and shown to
represent the remains of a farmstead within a ditched enclosure and approached by an
entrance to the east. Pottery suggests that activity occurred throughout the Roman
period, however the majority of the finds, including coins, indicated a large scale
redevelopment of the site in the mid 3rd to 4th century AD. This phase of activity
included the construction of a high status residence, possibly of flint construction and
with a tessellated floor.
4
ROUND HILL, WITTENHAM CLUMPS,
OXFORDSHIRE
Acknowledgements
The geophysical survey was undertaken by John Gater and Chris Gaffney, with staff
from GSB Prospection, and survey by Henry Chapman, University of Hull. Evaluation
strategy was conducted by Professor Mick Aston (Bristol University) and site recording
was co-ordinated by Phil Harding, assisted by Nick Best of Wessex Archaeology. The
evaluation was undertaken by the Time Team’s retained excavators with help from
members of Oxford Archaeology. The archive was collated and all post-excavation
analysis and assessment undertaken by Wessex Archaeology including management
(Roland J C Smith), report (Phil Harding), finds (Lorraine Mepham) and illustrations
(Mark Roughley). Specialist comment was provided by Stephanie Knight (animal
bone), Sarah Wyles and Chris Stevens (plant remains) and Nick Cooke (Roman coins).
The progress and successful completion of the work also benefited from discussion on
site with specialists of Iron Age archaeology especially Francis Pryor and Paul Booth
(pottery).
5
ROUND HILL, WITTENHAM CLUMPS,
OXFORDSHIRE
1 BACKGROUND
1.1.2 Round Hill and Castle Hill are two very prominent Chalk outliers lying north
of the main Chalk escarpment in south Oxfordshire. The two hills and their
environs form the Little Wittenham Nature Reserve, which is owned and
managed by The Northmoor Trust. Castle Hill is capped by an Iron Age hill
fort, a scheduled ancient monument (No. 208), which is the focus of a long-
term project by Oxford Archaeology and the Northmoor Trust, in
collaboration with English Heritage. The Time Team project at Round Hill
was designed to run independently of the Oxford Archaeology project but to
supply results that would be of value to the work being undertaken on Castle
Hill.
1.1.3 Round Hill lies less than 350 m north-west of Castle Hill, with a saddle in
between the two. Round Hill is the higher, with an elevation of 120m OD,
while Castle Hill rises to only 107 m OD. There are no earthworks apparent
on the slopes of Round Hill, whereas Castle Hill retains clear and impressive
earthwork defences, with the entranceway to the hill fort being particularly
clear.
1.1.4 From the north, the two hills are prominent features within the landscape and
form a clear focal point as they rise above the River Thames valley, which
flows in a broad meander some 1.5 km to the north of Round Hill. To the
south, a broad spread of land is drained by the Mill Brook, a minor tributary
stream of the River Thames. The ascent to the hills is more gradual from this
direction and their appearance less prominent.
1.1.5 Round Hill and Castle Hill form part of the Sinodun Hills, outliers of Lower
Chalk that are capped by a veneer of glacial sands and gravels of the Fourth
Terrace. The lower slopes of the hill, which coincided with the location of
most of the Time Team trenches, lie at the boundary of the Lower Chalk and
the underlying Upper Greensand (BGS 1:50,000 Solid and Drift Series Sheet
254).
6
1.2 Previous archaeological work
1.2.1 The most conspicuous features at Wittenham Clumps are the defensive
earthworks of the Iron Age hill fort on Castle Hill. Excavations by Oxford
Archaeology, in collaboration with the Northmoor Trust, in 2003
(www.oxfordarch.co.uk/wittenham) showed that occupation of the hill top could be
traced almost continuously from the Mesolithic period to the late Roman
period, with medieval pits suggesting 12th century occupation. The site was
initially defended by a ditch during the Late Bronze Age, which was
superseded by the ramparts of the Iron Age fort.
1.2.2 The 1st edition OS map of 1887 indicates that human remains were found in
the early 19th century immediately north and west of the defences. In 1950, a
bronze bracelet was recovered from the centre of the clump on Castle Hill
(SMR 3163). Coins of Domitian, Gratian and Arcadius have also been
recovered from the vicinity. In 1984, further burials were found and
excavated on the north-east of Castle Hill (Chambers 1986). These were
judged to be of a late Roman date, although there were no finds.
1.2.3 The Abingdon Chronicle (written in the 12th century), state that Offa of Mercia
established a frontier on the Berkshire Downs in the 8th century and from this
it has been suggested that Castle Hill was taken and refortified as part of this
frontier (Parker 1885). A charter of 862 (BL. Cotton Claudius c.ix.108v;
Kelly 2000, 15) lists the estate as 10 hides and gives the bounds, but does not
mention Castle Hill.
1.2.4 By 1048, the manor of Little Wittenham was owned by Abingdon Abbey, and
is listed in Domesday. There is at that time no mention of occupation on either
hill.
1.2.5 Rocque’s map of Berkshire 1761 shows both hill tops as bare, and the tithe
map of 1843 indicates that the interior of Castle Hill was ploughed for arable.
The ‘clump’, which refers to the beech plantations at the summit of each hill,
is thought to date from the 18th century.
1.2.6 No formal archaeological work had ever taken place on the summit of Round
Hill, although there has been a number of small-scale investigations on the
neighbouring Castle Hill and the slopes to the south of both sites. Antiquarian
references mention the discovery of two burials on the summit of Round Hill
(Hearne 1716).
1.2.7 In 1947 a field walking exercise on the southern slopes of Round Hill was
carried out (Rhodes 1948) and revealed a Roman building defined by a dense
scatter of surface finds including pottery, painted wall plaster, tesserae and
tegulae. During subsequent excavations Rhodes found, beneath the Roman
building debris, a dark occupation layer between 0.3 and 0.5m thick, within
which was a rectangular chalk and stone platform. The layer was associated
with large quantities of Early Iron Age pottery and animal bones. This layer
could have been a midden.
7
1.2.8 Rhodes recovered a wealth of Iron Age pottery within occupation layers from
the excavation. Beneath the Iron Age layer was a possible buried turf line, and
below it a grey clay containing occasional sherds, including Beaker pottery
and a struck flint, which implied possible early prehistoric occupation on the
site.
1.2.9 Near by to the east, two cinerary urns and part of a third were found within a
pit in the corner of a rectangular enclosure (SMR 7904), aligned north to
south, that was revealed by geophysics using a fluxgate gradiometer (Price
1995). The survey was undertaken on an area just over 1 hectare to the south
and east of the site of Rhodes’ 1947 excavation.
2 METHODS
2.1 Introduction
2.1.1 A project design for Time Team’s evaluation was compiled and provided by
Videotext Communications (Videotext Communications 2003). Full details of
the circumstances and methods are contained in the project design and are
summarised below.
2.2.1 The project provided an opportunity to examine the archaeology of Round Hill
and its lower slopes and place the results in context relative to the large,
ongoing Castle Hill project. The results would include any previously
unrecorded archaeological remains and establish their chronology,
distribution, function, status and state of preservation.
2.2.2 It was proposed to examine an undated earthwork feature inside the clump and
to re-locate and re-assess the Iron Age and Roman remains, including the
chalk building platform identified by Rhodes, south of Round Hill. This would
help to determine its function and establish whether it related to the
geophysical anomaly highlighted by Price (1995).
2.2.3 Rhodes (1948) suggested that the occupation on the southern slope between
the two hills is the settlement of the people who built the Castle Hill hill fort.
He also suggested that the large spread of material within which the building
platform was identified might be evidence of a small Roman villa.
2.2.5 Evaluation would make it possible to examine in greater detail the earliest
deposits, including the chalk structure relating to Iron Age activity, and to
8
establish and understand the extent and relationship of Rhodes’ Roman
building to the rectangular enclosure. The results would provide dating
evidence for the enclosure, its phasing and function and investigate its position
in the wider landscape.
2.2.6 The results would also provide a condition survey of the archaeological
remains along the lower slopes of the site. The site has been vulnerable to
ploughing in this area up until 1947 and to tree planting at the top of Round
Hill.
2.3.2 Eleven machine-excavated trenches of varying sizes were dug (Figure 1) after
consultation with the on-site director, Professor Mick Aston and associated
specialists from Oxford Archaeology, particularly Tim Allen and Paul Booth.
The precise location of individual trenches on Round Hill was guided by the
availability of open space within the tree canopy. Trenches in the open pasture
were located across geophysical anomalies from the results of the geophysical
survey to answer specific aims and objectives of the project design.
2.3.3 The trenches were excavated using a wheeled JCB mechanical digger and
back hoe fitted with a toothless ditching bucket. A small tracked mini-digger
was also used for the removal of topsoil in the wooded area on Round Hill. All
machine work was undertaken with constant archaeological supervision and
ceased at the identification of significant archaeological deposits, or where
natural deposits were encountered first. When machine excavation had ceased
all trenches were cleaned by hand and archaeological deposits were excavated.
2.3.4 A sufficient sample of all deposits was examined to allow the resolution of the
principal questions outlined in the aims and objectives above.
2.3.5 All archaeological deposits were recorded using Oxford Archaeology’s pro
forma record sheets with a unique numbering system for individual contexts.
This ensured a compatible recording system with that being used on the
excavation at the adjacent Castle Hill. Trenches were located using a Trimble
Real Time Differential GPS survey system. All archaeological features and
deposits were planned at 1:20 or 1:50 and sections drawn at 1:10 or 1:20,
whichever was appropriate for the circumstances. All principal strata and
features were related to Ordnance Survey datum and a photographic record of
the investigations and individual features was maintained.
2.3.6 The work was carried out over three days on 29th-31st July, 2003. All spoil was
metal detected by an individual recommended by Oxford Archaeology.
9
2.3.7 At the completion of the work all trenches were reinstated using the excavated
spoil from the trenches. All artefacts were transported to the offices of Wessex
Archaeology where they were processed and assessed for this report.
3 RESULTS
3.1 Introduction
3.1.1 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.
3.2.1 Two areas comprising a total of 7.5 ha were investigated. Area 1 lay south of
Round Hill and Area 2, to the south of Castle Hill were investigated using
magnetic survey (Figure 1). A small area across the centre of the rectangular
enclosure in Area 1, which coincided with the distribution of Roman building
debris recorded by Rhodes, was also surveyed using resistance.
3.2.2 The magnetic survey in Area 1 revealed a series of rectilinear anomalies in the
west part of the survey area, which may be part of an enclosure.
3.2.3 Numerous anomalies, some in clusters, were apparent in the west part of the
survey area. Similar better-defined examples in the east suggest that these
features are likely to represent pits. Elsewhere there were numerous, more
isolated and larger pits. These features are likely to indicate the distribution of
storage and rubbish pits, although others may be associated with burning and
represent hearths, ovens or small-scale industrial/workshop activity.
3.2.4 The clearest anomaly in the centre of Area 1 was the rectilinear enclosure with
an entrance on the east side. The resistance survey also indicated a high
resistance anomaly that was also visible on the results of the magnetic survey
and which coincided with the Romano-British building investigated by
Rhodes.
3.2.5 Several other anomalies were present within and immediately outside the
enclosure, including three well-defined pits in the north-west corner of the
enclosure. The southernmost feature was sampled with the ditch in trench 6.
3.2.6 A possibly semi-circular feature was also detected immediately north of the
enclosure, however it was possible that this represented an interconnecting
group of pits.
3.2.7 Two small sub rectangular enclosures, less than 20 m across and with internal
features, were identified in the south-east of survey Area 1. A noticeably
strong magnetic anomaly, which may represent interconnecting pits was
located between the two enclosures.
10
3.2.8 Two strong parallel linear anomalies that coincided with a modern track were
located in the north-east of the survey area. They lay immediately west of
three curvilinear responses that may be associated with the defensive ditches
of Castle Hill.
3.2.9 A large area of magnetic response was detected on the higher slopes of Round
Hill. It was considered that although these responses may represent ploughed
out defences it was thought more likely that they represented thinner, plough
damaged soils on the upper slopes.
3.2.10 The entire survey area was overlain by a series of parallel linear anomalies
that could be attributed to ridge and furrow agriculture.
3.2.11 Area 2 lay in land below the ramparts of Castle Hill (Figure 1). The results
indicated that there was a dense concentration of probable pits across most of
the area, although there was a clear eastern boundary. Two curvilinear features
were also detected, one in the west of uncertain interpretation. The other, in
the central part of the area, measured approximately 12 m across with a break
on the south-east part of the circuit that may be an entrance, which suggests
that it may be a round house (or ploughed barrow).
3.2.12 Two linear anomalies beyond the main concentration of pits may indicate
outworks to the main defences of the hill fort with a hollow way leading to the
entrance.
3.4.1 Three trenches were excavated within small open areas between the tree-
canopy on the heavily wooded summit of Round Hill (Figure 1). They were
dug to re-examine the possibility that human burials were present on the hill,
to identify any evidence of settlement and section the earthwork within the
clump.
3.4.2 Trenches 1, 2 and 3 were excavated using a tracked mini digger and measured
5 m, 8.5 m, and 2 m long respectively and a bucket width, 1 m wide. Trench 2
was aligned obliquely across the slight bank that enclosed the more mature
trees.
3.4.3 The evaluation indicated that there was a consistent covering of heavily
rooted, dark brown humic silt topsoil (101, 201, 301), which varied from 0.15
m to 0.4 m thick, and which overlay the weathered Chalk bedrock. In places
11
residual patches of mid brown chalky silt subsoil (102, 302, 303) or pebbly
gravel (103) filled pockets in the weathered Chalk surface (105, 203).
3.4.4 The bank was very slight but was visible as a deposit of pale grey-brown
chalky material (202), 0.20 m thick that was indistinguishable from the
subsoil. The spread of the bank coincided with a slight rise in the level of the
underlying bedrock. There were no datable finds and it was concluded that the
bank dated to the establishment of the 18th century plantation.
3.5 Trench 4
3.5.1 This trench was located from the results of the magnetic survey to examine the
entrance of the rectangular enclosure (Figure 1). The trench (Figure 2)
measured 4.70 m long and 1.90 m wide and was aligned approximately north-
west to south-east across the south terminus.
3.5.2 The topsoil horizon (401) overlay a soil accumulation of grey brown sandy silt
(402, 403, 404). This undifferentiated subsoil deposit contained large
quantities of archaeological material and was excavated by hand in two spits
(402, 404) 0.10 m thick at the south end but only one spit (403), 0.10 m thick
at the north end. The two ends of the trench were characterised by pottery
assemblages of widely differing dates. Large quantities of Iron Age pottery
were common towards the south end (402, 404) of the trench with a
concentration of Roman material (403), including pottery, bone, tile, flint and
metal at the north end. There was no definable boundary between them.
3.5.3 A second spit (404) was removed from the south end of the trench, which
included a high concentration of rounded pebbles. However at the north end it
was possible to define the edges of a ditch (406), 2 m wide and 0.70 m deep
with moderately sloping sides and a rounded base. It became apparent that
most of the material excavated as spit (403) formed the tertiary fill of this
ditch with an underlying soft grey brown silty clay (405) and a firm mid to
dark yellow brown primary fill (407).
3.5.4 The excavation of ditch (406) revealed that it cut the terminus of the main
enclosure ditch (409) as defined by geophysics, although the written archive
suggests that terminus (409) was later. The terminus extended approximately
0.40 m from the south section of the trench, was cut with vertical sides and a
flat base and penetrated the natural bedrock by 0.80 m. It was filled with mid
grey brown clay silt (408).
3.6 Trench 5
3.6.1 Trench 5 (Figures 1 and 3) was located from measurements calculated from
the excavation report of Rhodes to re-locate the chalk floor, the associated
Iron Age levels and any evidence for a Roman building. It was dug as a trench
3 m by 3 m but was subsequently extended to a pair of interlocking squares,
12
with an extension to the north to provide a trench with a total area of 30 square
metres.
3.6.2 Following the removal of the topsoil (501) and plough-disturbed subsoil (502),
Rhodes’ excavation could be defined within an area of unexcavated deposit.
The former excavation (516) measured 3.10 m north-west to south-east and
2.90 m north-east to south-west. The depth and extent of the archaeological
work undertaken by Rhodes varied. Some deposits (503 and 513) had been left
in situ, while others had been sectioned.
3.6.3 The complete stratigraphic sequence of the trench was evaluated in two slots,
0.50 m wide, that were excavated along the north-east and south-west sections
of the trench. A firm grey brown chalky clay (514) was exposed in the base of
the slot in the north-east of the excavated area. It was considered likely that
this deposit, at 99.17 m OD formed the natural bedrock surface in the trench.
3.6.4 This deposit was overlain by friable grey brown chalky silt (515), which was
similar in composition and stratigraphy to material (504) sectioned in the slot
along the south-west edge of the trench.
3.6.5 These basal soils were overlain by a cobbled surface (503), composed of
quartzite pebbles up to 0.12 m long. The surface, into which fragments of
Early Iron Age pottery and bone were impressed, was most well-defined
towards the south-east corner of Rhodes’ trench. Here it was concentrated
although it thinned away from the former excavation to the south-east.
3.6.6 It was sealed by a chalk floor (513) towards the north-east, which is likely to
be the chalk floor described by Rhodes. It was of varying thickness but
reached up to 0.10 m thick in places. Rhodes did not remove either of these
surfaces.
3.6.7 The chalk surface appeared to have been cut into by two possible post-holes
(510 and 512), which lay within the area of Rhodes’ excavation but which are
not shown on his plan. Post-hole (510) was approximately 0.45 m in diameter
and survived to 0.15 m deep, while post-hole (512) was smaller,
approximately 0.20 m in diameter and depth. Both post-holes had irregular
sides, flat bases and were filled with dark brown silt (509 and 511) and both
contained Early Iron Age pottery.
3.6.8 This occupation surface was sealed across the trench by a deposit of dark
brown silt (517), 0.20 m thick, which is likely to represent a soil horizon.
3.6.9 The soil was covered by a thin off-white chalk surface (506), which was most
prevalent in the north-west corner of the area where it had been truncated by
Rhodes’ trench. However it was also traced across the site to the north-east
corner of the trench where it was recorded in the section. This surface was
overlain by a deposit (505) containing Roman material, including tesserae.
3.6.10 An additional undated possible post-hole (508) appeared to be cut through this
chalk surface. The post-hole was approximately 0.20 m in diameter and 0.15
13
m deep, with steep sides and a flat base. It is unclear whether it had any direct
relationship to post-holes (510) and (512), which contained Iron Age pottery
but which are likely to have been truncated by Rhodes’ excavation. They are
all of similar size and lie on a similar arc.
3.6.11 This trench re-established the precise location of Rhodes’ trench and
confirmed the date and accuracy of his observations. It relocated the Iron Age
cobbled surface and overlying chalk floor although there was little to
demonstrate conclusively that they lay within a building, apart from two post-
holes and it is possible that they formed part of an exterior yard. However this
part of the site clearly lay within an occupation area that was not set aside for
pits. The results of the geophysical survey indicated that similar areas could be
defined which may make it possible to reconstruct the settlement plan in more
detail. The stratigraphy of the trench is very reminiscent of that seen at the
base of trench 8 where a sequence comprising a cobbled surface and soil
containing Iron Age material were sealed by a chalk surface capped by Roman
demolition rubble.
3.6.12 The Roman material lay towards the south-east extent of the spread of
demolition material defined by the geophysical survey. This material was
derived from a building represented by the substantial wall foundations
observed in trench 7. However the chalk surface (506) may have lain within a
timber outbuilding within the enclosure.
3.6.13 The results of the Time Team excavation reveal both similarities and
variations to the account published by Rhodes (1948). There are clear
correlations between Rhodes’s topsoil horizon, his Romano-British occupation
(layer 1) and the ‘upper’ part of his black soil (layer 2) with Time Team’s 501-
2, 506 and 517. His published plan of the floor of the Iron Age hut also shows
remarkable similarities to the extent of the compacted chalk floor (513) with
large Bunter pebbles to the south (503) as re-excavated by Time Team. The
inclusion of the post-hole excavated by Rhodes to the plan shown by Time
Team does nothing to clarify the plan of any hut superstructure. The post-hole
is recorded as being 2 feet (0.61 m) deep (1948, 22) but was not relocated in
the recent excavations.
3.6.14 The ‘floor of the hut’ was ‘left intact’ by Rhodes (1948, 20), which accounts
for the remarkable similarities in the two plans. It is apparent from his
illustration that although he excavated around the hut floor he only
reconstructed the complete stratigraphic sequence in a test pit dug in the north
west corner of the trench. This test pit, which measured approximately 4 feet
long (1.22 m) and 1 foot 6 inches (0.46 m) wide, penetrated approximately 2
feet 6 inches (0.76 m) below the level of the chalk floor (513) to the natural
bedrock, which he identified as a chalky clay overlain by yellow clay.
3.6.15 The correlation of the stratigraphy between the two excavations can be further
matched in the ‘lower’ part of Rhodes’s black soil (layer 2) with 515 and his
grey clay (layer 4), which measured approximately 1 foot (0.30 m) thick, with
514. The yellow clay horizon lay approximately 3 feet 2 inches (0.96 m)
below the ground surface. There is no record of any yellow clay from the
14
Time Team excavations in trench 5, which reached a maximum depth of 0.80
m; however clay deposits, which may equate to Rhodes’s yellow clay, were
detected at the base of trenches 7 and 8. In the light of these results it is
apparent that the Time Team excavations did not reach the natural bedrock in
trench 5. Strangely no trace of Rhodes’ test pit was detected in the north west
corner of his former excavation.
3.6.16 The re-excavation of Rhodes’s Iron Age hut floor provided an opportunity to
reassess the extent and condition of the deposit. Rhodes’s plan (1948, fig 8)
showed a clearly defined compacted chalk surface with straight edges,
bounded by a band of Bunter pebbles to the south. The Time Team
excavations were able to confirm the general accuracy of this plan. It is
possible that minor variations, including post holes 510 and 512, which were
not noted by Rhodes, have resulted from bioturbation or surface weathering.
The Time Team excavations traced the chalk surface to the east, beyond
Rhodes’s trench, but suggest that the floor thinned to the north, where it was
recorded in section as a ‘thin cobble spread’.
3.7 Trench 6
3.7.1 A trench, 12.5 m long and 1.8 m wide (Figures 1 and 4) was opened by
machine to locate and examine the ditch on the west side of the enclosure and
included a pit-like anomaly detected by geophysics.
3.7.2 The friable dark brown loamy topsoil horizon (6001) measured 0.2 m thick
and overlay a brown silt loam subsoil (6002), 0.18 m thick that contained
mixed fragments of roof tile, animal bone, oyster shell and Roman pottery.
This disturbed deposit is likely to represent plough soil including additional
material from further up the hill.
3.7.3 The natural pale yellow brown clay bedrock (6008) was overlain and diffused
into a veneer of stiff brown silty clay (6020), 0.08 m thick. This horizon,
which was only recorded to the east of the ditch produced no finds but was cut
through by the enclosure ditch (6015). It is most likely that this deposit
represents the remains of a truncated pre-Roman old ground surface and may
equate with contexts (515) and (504) as seen in trench 5.
3.7.4 Pre-Roman activity in the trench was represented by a steep sided pit (6016),
which extended from the south trench edge immediately west of the outer lip
of enclosure ditch (6015). It measured approximately 1.5 m north to south and
was excavated to 0.5 m deep but was not bottomed due to lack of time. It was
filled with a series of loose mid orange grey silt loam deposits (6017, 6018),
which became more grey brown (6019) towards the base. These upper pit fills,
which are likely to represent only the tertiary deposits of the feature, produced
considerable quantities of Early Iron Age pottery including some very large
pieces. This pit represents the most westerly extent of the Early Iron Age
occupation yet excavated on the site.
3.7.5 The enclosure ditch (6015) was 2.8 m wide and was cut into the chalky marl
bedrock. It measured 1.08 m deep with moderately sloping sides to a slightly
15
rounded base 0.20 m across. The ditch appears to have silted naturally with a
primary fill of light grey brown chalky silt (6014), derived from the ditch
sides, which was replaced by darker less chalky silts in the secondary and
tertiary fills (6012, 6011, 6010). There were no intervening lenses of domestic
refuse, although small quantities of residual Iron Age pottery were
incorporated with the Roman material, nor were there indications of re-cutting,
or of turf lines. The sedimentation was slightly off centre to the east,
suggesting that some of the filling was derived from an internal bank.
3.7.6 The ditch fill was capped by a dump (6009) of angular Chalk rubble and flint
nodules, up to 0.12 m across, that was mixed with Roman material,
particularly roof tile.
3.7.7 There was very little conclusive evidence for an internal bank apart from a
deposit of dark grey-brown silty clay (6005), up to 0.25 m thick, which
extended across the trench to the east of the inner ditch edge. This material
contained a mixed assemblage of Early Iron Age and Roman pottery and must
therefore be at least Roman in date. It was considered by the excavator to
represent the remnants of a spread bank; however its composition and pottery
assemblage was remarkably similar to the overlying subsoil 6002. The
frequency of Roman pot and tile was noticeably less frequent towards the
ditch. Regrettably the stratigraphic relationships between deposits in sections 1
and 2 were not included in a single drawing.
3.7.8 The bank spread was cut into by a pit (6006), although the evaluation record
indicates that the relationship was indistinct. The pit extended approximately
1.5 m from the north edge of the trench and was 0.65 m deep with steep sides
and rounded base. The dark grey silty clay (6013) primary fill was overlain by
a deposit of similar composition (6007), but which contained large quantities
of Roman pottery, animal bone, tile and oyster shell. The pit was sealed by
spreads of chalk rubble and flint nodules (6004, 6003) similar to that in the top
of the enclosure ditch. It sees likely that these deposits, which represent
imported materials, represent spreads of demolition material that have been
heavily truncated by ploughing from buildings of chalk and flint construction
that once stood within the enclosure. It is possible that the rubble was
introduced deliberately as make-up or merely filled weathering hollows in the
tops of earlier features.
3.8.1 Three trenches (Figures 1 and 5) were excavated in a group to define and
interpret an approximately rectangular disturbance that was located by
resistance in the centre of the enclosure. The anomaly, which measured
approximately 25 m long and 11 m wide, was aligned approximately south-
east to north-west with a rounded apsidal north end. It was thought that this
anomaly might represent the plan of the principal building within the
enclosure.
3.8.2 Trench 7 measured 4 m long and 2 m wide and was aligned north-east to
south-west across the south side of the anomaly, at a point at which the edge
16
was most clearly defined. Trench 8 was added 3 m to the north-east and
measured 3 m long and 2 m wide to trace the north side of the anomaly, while
trench 10 was inserted to locate the west apsidal end. It measured 3.5 m north-
west to south-east and was 1.8 m wide. All three trenches were opened by
machine and excavated to their full extent but were later sectioned using hand
excavated slots 0.70 m wide to locate the natural bedrock and establish the
overlying sequence of deposits.
3.8.3 A mottled grey green compact clay (811) was observed in a small sondage at
the north end of trench 8 at 100.70 m OD. A similar deposit (709) was also
recorded in the base of a robber trench in trench 7. It was thought likely that
these two units formed part of the natural bedrock on the site; however
elsewhere on the site features were cut into chalky marl, which may indicate
that the clay is an archaeological deposit.
3.8.4 The clay was overlain, in trench 8, by a cobbled surface (810) which was also
traced to trench 10 (1004). It comprised ochreous stained, sub rounded flint
cobbles, approximately 0.05 m across, with some mixed quartzite pebbles.
They were present throughout the excavated section in trench 8 but thinned to
the north-west of trench 10.
3.8.5 The cobbles in trenches 8 and 10 were undated but were overlain by a soil
(809, 1003). A few fragments of Early Iron Age pottery from (809) were
found with a sherd of Roman material, which is likely to be intrusive, and
suggest that this is a pre-Roman horizon.
3.8.6 A thin, continuous chalky surface (808, 1008) spread across trenches 8 and 10
but thinned to the north and west. It was particularly well defined at the east
end of trench 10 and the south end of trench 8, where it butted a compact flat
creamy yellow mortar surface, (807) that extended 0.50 m north-east from the
end of trench 8.
3.8.8 The clearly defined sequence of individual soil and demolition layers could
not be traced to trench 7. They appeared to be represented by a single deposit
of dark grey silt (706, 705, 702) that included quantities of Roman roof tile.
This horizon corresponded to (1002) and was present above the natural clay
exposed in the section of robber trench (708).
3.8.9 At the south end of trench 7 a robber trench (708) aligned east to west was cut
through the demolition deposits and through a dark grey brown plough soil
(704) to the south. The robber trench, which was also filled with demolition
17
rubble (703, 707), measured 0.70 m across and was 0.65 m deep with steep
sides and slightly rounded base.
3.8.10 The excavation of these three trenches demonstrated that the anomaly detected
by geophysics coincided with the extent of demolition debris. Its southern
edge was marked by a robber trench, which ran approximately parallel to the
north edge of the enclosure. This feature provided evidence for at least one
substantial building on the site, which may have formed the principal structure
on the site and have included a tessellated floor. However it was not possible
to confirm the orientation, extent or precise location of the building nor was it
possible to provide evidence for floor levels or of domestic occupation. It is
possible that the demolition debris merely marked an accumulation of debris
from a range of buildings, incorporating at least one of flint construction with
others of clay wall or timber construction.
3.8.11 The presence of a range of roofing materials suggests that buildings with both
stone and ceramic tiles were present on the site. These buildings may have
been located around a central open yard, set inside the perimeter bank or
around a principal building in the middle of the enclosure. The bulk of the
Roman pottery is of mid 3rd to 4th century AD date and probably relates to the
main phase of occupation and its later use including the demolition of
buildings within the enclosure. However the stratigraphic record, indicated by
the re-cut ditch in trench 4 across the entrance, indicates prolonged Roman
activity on the site, which is confirmed by the presence of Early Roman
pottery from the site.
3.9 Trench 9
3.9.1 A small trench 2.5 m north to south and 2 m east to west was excavated over
an anomaly detected by geophysics that was believed to indicate a pit (Figure
1). The removal of the topsoil (901) and subsoil (902) horizons revealed that
the trench area contained three inter-cutting pits (Figure 6) that were cut into
the chalky marl bedrock of the Lower Chalk.
3.9.2 The initial interpretation suggested that it post-dated pit (910) to the north and
pit (906) to the south. However a reassessment of the dipping stratigraphy that
was visible in the section and a reconsideration of the excavated edge on the
south side of pit (904) cast some doubt on this sequence.
3.9.3 The revised interpretation considered that the earliest pit (910) lay to the north
and was approximately 0.60 m in diameter and 0.30 m deep. It was filled with
dark brown friable clay silt (911), which was exposed in the edge of pit (904)
18
but was not excavated. Pit (904), which was cut with a slightly belled profile
was subsequently cut by pit (906) to the south. This pit measured
approximately 1 m in diameter and at least 0.70 m deep. It was not fully
excavated but was filled by a series of friable mid to dark grey brown clay silts
(907, 908, 909) with faint, but visible, tip lines that trended to the south.
3.9.4 It seems most likely that this small trench was cut within a concentration of
Early Iron Age storage/rubbish pits that were dug in sequence and migrated to
the south.
3.10 Trench 11
3.10.1 This trench was located towards the south-east corner of the lower slopes of
Round Hill. It was excavated to examine one of two small square enclosures
with central anomalies that were revealed in that area by geophysics (Figure
1).
3.10.2 The geophysics plot revealed that the enclosure measured approximately 16 m
across with a central anomaly. A trench 12 m long and 2 m wide was therefore
excavated from the centre northwards across the enclosure ditch (Figure 7). A
representative sample of features that were cut into the chalk marl bedrock
were excavated by hand.
3.10.3 The trench demonstrated that the enclosure ditch (1113) measured 1 m across
and was 0.50 m deep with moderately sloping sides and a narrow, slightly
concave base 0.15 m across. It had silted with, an upper, dark grey brown
(1114) and, lower, grey brown silty clay (1115). A veneer of primary grey clay
(1116), containing two sherds of Early Iron Age pottery, lay across the base.
There was nothing to indicate an inner earthen bank or that the feature had
been recut.
3.10.4 The central feature comprised a large Early Iron Age storage/rubbish pit
(1103). It projected from the east edge of the trench but was 1.80 m across and
0.85 m deep with steep vertical sides that were slightly undercut in places, and
a flat base. It was filled with homogeneous grey brown silty clay (1104, 1105)
that was marginally darker and chalkier towards the base. It contained Early
Iron Age pottery, and animal bone.
3.10.5 A small rounded extension (1123) was situated on the west edge of the pit.
This may have been the arc of a smaller, earlier pit that was cut by the main
feature or was a step to provide access to the main pit.
3.10.6 A shallow ditch (1106) ran north-west to south-east across the site
immediately south-west of pit 1103. The ditch was 0.90 m across and 0.50 m
deep with sloping sides and a rounded base. It was filled with dark grey silty
clay (1107) that overlay lighter material (1109), which were separated by a
lens of charcoal (1108), which contained Early Iron Age pottery.
3.10.7 An additional pit (1110), with vertical sides and a flat base, was exposed at the
base of the excavated ditch. It measured 0.6 m in diameter and was 0.30 m
19
deep. It was filled with a layer of grey-green clay (1112), 0.20 m thick that
also contained Early Iron Age pottery and which was overlain by silty clay
(1111).
3.10.8 A shallow post-hole (1117) 0.26 m in diameter and 0.10 m deep lay between
pit (1103) and the enclosure ditch. It was poorly cut into the underlying natural
rubble with sloping sides and a rounded base. A possible post-hole (1126),
which was unexcavated but of similar diameter extended from the edge of the
trench approximately 1.30 m west of this post-hole.
3.10.9 There were also two other miscellaneous features at the south end of the
trench. These features (1119, 1121) were recorded but could not be excavated
in the time available.
4 FINDS
4.1 Introduction
4.1.1 Finds were recovered from eight of the eleven trenches excavated; no finds
were recovered from Trenches 1 – 3, and relatively little material came from
Trenches 9 and 11. All finds have been quantified by material type within each
context. 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 has been subjected to more formal scanning,
including quantification by ware type (details below). Spot dates have been
recorded for selected material types as appropriate. All finds data are currently
held on an Access database.
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, with particular
reference to the long-term Oxford Archaeology/Northmoor Trust project at
Castle Hill. The assemblage is divided into two chronological groups, Early
Iron Age and Romano-British, with a small quantity of worked flint of earlier
date.
4.2 Pottery
4.2.1 The entire pottery assemblage has been scanned, and quantified by broad ware
group (e.g. sandy wares) or known type (e.g. samian) within each context.
Spot dates have been recorded on a context by context basis, and the presence
(but not quantity) of diagnostic forms noted. Results are summarised by trench
in Table 2.
4.2.2 The Early Iron Age assemblage is characterised by a high proportion of sandy
wares, with a low frequency of shelly and flint-tempered wares. Sandy wares
20
cover a wide range of coarseness, from fine, silty fabrics, well finished and
frequently burnished (and sometimes red-finished) to fabrics containing very
coarse (10mm+) rounded pebble inclusions, and others containing prominent
angular quartz. Identifiable forms include fineware carinated bowls, frequently
burnished and sometimes red-finished; and coarseware angular jars with finger
impressed shoulders (an almost complete profile was recovered from pit
(6016) in trench 6), and convex jars. There are also two expanded, finger
impressed rims from coarseware jars of uncertain form, and three lug handles.
Two body sherds with furrowed decoration almost certainly derive from
fineware bowls.
4.2.3 This assemblage finds numerous parallels within Early Iron Age assemblages
from sites in the Upper Thames Valley. Angular bowls and jars comparable to
the Round Hill vessel forms can be seen in phase 1 at Farmoor (Lambrick and
Robinson), period 2 at Ashville (de Roche 1978), and at Appleford (Hinchliffe
and Thomas 1980). The emphasis on sandy wares at the expense of shelly and
flint-tempered wares contrasts with the ceramic sequence seen at Farmoor and
Ashville, but sandy wares are now recognised as a common element of
Earliest Iron Age assemblages at sites such as Yarnton. One or two convex
forms are present here, but the sandy wares are definitely dominated by
angular forms.
4.2.4 The bulk of the Early Iron Age assemblage was recovered from Trenches 4, 5
and 6, with smaller quantities from Trenches 9 and 11, and only a handful of
sherds from Trenches 7, 8 and 10. The largest context group came from soil
accumulation deposit (402/403/404) (129 sherds), but this was mixed with
Romano-British material. In fact, only 38% of the Early Iron Age assemblage
(by number of sherds) occurred in contexts unassociated with Romano-British
pottery (none of these from trench 4).
4.2.6 The majority of the Romano-British pottery came from trenches 4 and 6, with
small quantities from trenches 5, 7, 8 and 10; trenches 9 and 11 produced no
Romano-British pottery. The largest context group came from possible bank
material (6005) (113 sherds).
21
4.3 Ceramic building material
4.3.2 One piece of tile from topsoil in trench 6 (6001) has been neatly trimmed to a
disc approximately 80mm in diameter.
4.4.1 Other building materials in the form of opus signinum and mortar were
recovered in small quantities (trenches 4 and 7), and a single piece of painted
(monochrome) wall plaster was found unstratified. All these are consistent
with the existence of a substantial, high status building on the site.
4.5 Stone
4.5.1 Nearly all the stone represents building material, consisting largely of
fragments of limestone roof tiles, some with surviving nail holes. There are
also a few (white) limestone tesserae, in similar sizes to the ceramic tesserae
(see above). One large piece of quartz conglomerate (pit 6016) could be from
a quern, but otherwise portable objects (querns and whetstones) are entirely
absent.
4.6 Coins
4.6.1 Seven coins were recovered. All of these are Roman, and all date to the late 3rd
or 4th centuries AD. All were sufficiently legible to be assigned to period,
although their general condition is poor, with some showing signs of heavy
corrosion, possibly indicating that they had been in topsoil or ploughsoil
deposits for some time. One of the coins, a contemporary Barbarous copy of a
radiate antoninianus dates to the late 3rd century, whilst the remaining six date
to the 4th century. None of the latter are particularly rare or unusual, with most
being fairly common 4th century types. Barbarous Radiates are common
throughout the Western Empire in the late third century and are thought to be
poor contemporary copies of ‘official’ coinage, although opinion is divided as
to whether these ‘barbarous’ copies were officially sanctioned. The 4th century
coins also include some possible contemporary copies, evident from their
stylised engraving. Only one of the coins was recovered from a stratified
context (ploughsoil (704) in trench 7) This was a badly corroded coin of the
early 4th century.
22
Table 1: Finds totals by material type (number / weight in grammes)
23
Table 2: Pottery totals by ware group (number / weight in grammes)
Ware Tr 4 Tr 5 Tr 6 Tr 7 Tr 8 Tr 9 Tr 10 Tr 11 TOTAL
IRON AGE
Calcareous ware - 1/11 1/20 - - - - 2/31
Flint-tempered ware 3/26 1/2 3/18 1/23 - 4/48 - 2/14 14/131
Sandy ware 155/1757 171/2350 165/2838 7/116 6/33 48/408 1/10 39/320 592/7832
Shelly ware 13/138 7/142 2/12 - - 2/17 - 4/34 28/343
sub-total Iron Age 171/1921 180/2505 171/2888 8/139 6/33 54/473 1/10 45/368 636/8337
ROMANO-BRITISH
Black Burnished ware 14/103 - 3/31 6/77 - - 3/12 - 26/223
Greyware 95/985 17/67 152/1964 21/217 29/215 - 26/285 - 340/3733
Grog-tempered ware 3/59 - 7/115 1/44 1/37 - 4/92 - 16/347
Misc. colour coat 1/8 - 9/85 - - - - - 10/93
Misc. whiteware 4/9 1/1 6/47 - - - 1/8 - 12/65
Oxidised ware 3/7 1/2 3/20 1/10 1/5 - 1/4 - 10/48
Oxon colour coat 22/118 2/3 3/170 9/73 6/48 - 8/56 - 50/468
Oxon whiteware 1/41 1/5 2/69 - 1/22 - - - 5/137
Samian 7/76 1/1 9/20 - - - - - 17/97
sub-total RB 150/1406 23/79 194/2521 38/421 38/327 - 43/457 - 486/5211
OVERALL TOTAL 321/3327 203/2584 365/5409 46/560 44/360 54/473 44/467 45/368 1122/13,548
24
4.6.2 The small number of coins recovered from the site makes intra-site
comparisons invalid. The date range of these coins (the earliest dated to AD
270 – 90, whilst the latest dates to AD 364 – 378) indicates that activity on the
site continued into the later 4th century. The absence of any 1st or 2nd century
coins or post-Valentinian coins need not be significant, as the coin assemblage
is small, and coins of these dates are rarer as site finds
4.7 Metalwork
4.7.1 This category includes objects of copper alloy, iron and lead. Apart from the
coins (see above), the copper alloy includes one pin (Roman) and one button
(post-medieval). The remaining two objects are unidentified and of uncertain
date. None of these objects were stratified.
4.7.2 The iron consists largely of nails (54), with one hobnail, one harness ring
(402), one socketed spearhead (404) and two strip fragments. Of these objects
the hobnail, one of the strips and 21 of the nails were unstratified finds. The
spearhead and harness ring both came from soil accumulation deposit
(402/403/404) in trench 4. The spearhead, which is incomplete, appears to be
an unusual pierced form, perhaps a ceremonial rather than purely functional
object.
4.8.1 This consists entirely of oyster, and includes both left and right valves, in other
words, both preparation and consumption waste. Most of the shell came from
trenches 4 and 6.
4.9.1 Other finds comprise small quantities of ironworking slag, worked and burnt
flint, fired clay (one piece with a probable wattle impression, plus six
featureless fragments) and vessel glass (post-medieval bottle).
4.10.1 The species present in each context was recorded, in order to give a species
frequency (O’Connor 1985) rather than absolute numbers of fragments
(NISP). The potential of the assemblage to provide information about
husbandry patterns, population structures and consumption practices was
ascertained from estimating the proportion (high over c.10%, medium c.4-10%
or low c.1-3%) of bones that could give information on the age, size and sex of
animals, butchery, burning and breakage. Using species frequencies and rough
estimations, although obviously less accurate, is more rapid than counting
exact numbers, and is considered adequate for the purposes of this assessment.
25
4.10.2 The extent of mechanical or chemical attrition to the bone surface was
recorded, with 1 indicating poor condition, 2 fair and 3 good. The proportions
of gnawed bone were noted as high, medium or low in the same manner as
outlined above. Any unusual combinations of bones that may have resulted
from a specific activity or could have been articulated were recorded.
4.10.3 1594 fragments of animal bone were recovered from 49 contexts. 548 bones
were from the 12 contexts that contained only pottery dating to the early Iron
Age, and 37 bones were from the four contexts that contained only Roman
pottery. The remaining deposits were undated or mixed (containing both early
Iron Age and Roman pottery), the latter are probably Roman in origin but
contain an undefined quantity of reworked earlier material. Only the bone
from contexts containing ceramics of just one period was included in this
assessment, as there is no way of knowing what proportion or which bones in
the mixed deposits was residual. The assessed sample has also excluded bone
from undated contexts that form part of the fills of dated features (such as the
enclosure ditch).
4.10.5 Bone was recovered from samples taken from five of these deposits, totalling
205 fragments, a large proportion of the 548 bones assessed. Seven different
species were noted in the samples, with a greater proportion of smaller species
(Table 3). Bird, fish, amphibian and small mammal bones were found only in
the coarse residue of samples. However, hand recovery was generally good, as
small bones such as sheep phalanges and hyoids were collected. Unidentified
fragments of bone from larger animals made up the majority of bones from
samples, and the potential for sample material to provide information on
butchery, age and size of animals was generally limited.
Table 3: Animal bone species list and frequency for Early Iron Age contexts.
4.10.6 Cattle and sheep/goats (no positive identifications of goat) were found in the
same number of contexts, although the bones of these two species were not
26
always found in the same deposits. One would expect more sheep/goats than
cattle in the Iron Age, and the apparently equal frequency may be a symptom
of the method of quantification rather than a true representation. For example,
cattle provide large quantities of meat and their carcasses may have been
divided into more ‘portions’ than sheep/goat, resulting in a wider dispersal of
skeletal elements, which might then be deposited in more contexts per
individual. Pigs were found in fewer contexts, and horse was even less
frequent, as only one bone was represented. Only one context did not contain
any unidentified fragments.
Table 4: Number of Early Iron Age contexts containing bones with high,
medium or low potential to inform on husbandry, butchery and disposal
practice.
4.10.7 Ageing information was relatively common, with most contexts containing
numerous bones that could be aged (Table 4). Very young or neonatal
sheep/goat and pigs were in evidence, suggesting breeding on or near site. Size
information was more limited, as many of the bones were fragmentary, but
most contexts contained some measurable bones. One pig tooth with furrows
in the enamel suggests that some animals may have been through periods of
malnourishment (Hillson 1986, 129), and a porous cattle rib head with
exostoses indicates at least one animal suffered from injury or disease
associated with old age.
4.10.8 Butchery marks and burning were infrequent in most contexts, although the
majority of contexts contained bones with some evidence of butchery, often
including helical fractures (breaks made when the bone was fresh). Burning
was seen in six contexts but the majority of burnt fragments were from
samples. No pattern was noted, suggesting that burning resulted from the
inclusion of bone in fires rather than from the cooking process.
4.10.9 No unusual combinations of bone elements were seen, although it the large
number of large mammal rib bone fragments in pit (6016) can be noted. A
mouse-sized rodent in pit (904) was represented by three bones, two of which
may have been articulated, suggesting that this animal was buried (but not
excavated) whole, perhaps an accidental inclusion or intrusive individual.
4.10.10 Working or wear was noted on a sheep metatarsal, which had been worn in
several places across the shaft of the bone, especially just above the distal
epiphysis. Similar wear, although more pronounced, was noted on Iron Age
sheep metapodials at Danebury (Sellwood 1982: 389).
Roman
4.10.11 The bone in Roman contexts is in fair-good condition, with gnawing frequent
in two contexts, rare in another and absent in a fourth.
27
Table 5: Animal bone species list and frequency from Roman contexts.
4.10.12 The species frequency differs from that in the early Iron Age, with horse and
bird (probably chicken) bone more frequent (Table 5). Sheep/goat appear to
be more frequent than cattle, often regarded as evidence of an ‘unromanised’
settlement, although the numbers are too small to provide confirmation of this.
4.10.13 The proportion of bones that could provide evidence of age and size is
generally high, although butchery marks are less frequent and burnt bone is
not present.
5 ENVIRONMENTAL SAMPLES
Flot Residue
Feature type/ Context Sample size flot size Grain Chaff Weed Seeds Charcoal Other Charcoal analysis
No litres ml Uncharred Charred >5.6mm >5.6mm
Trench 4
15
Roman 405 1 ? 40 A C - C C Moll-t (A*) A* C
5
ditch 406 407 4 33 60 B B c B C Moll-t (A*) -
Trench 5
200
EIA surface 506 6 37 400 C B c C C Moll-t (B) -
100
EIA silt layer 515 5 36 200 C - c C C Moll-t (A) -
500
IA soil 517 7 33 700 B B c C C Moll-t (C) -
Trench 6
40
Roman pit 6007 600 35 100 B B c A C Moll-t (A*) - P
6006
Trench 9
15
EIA pit 904 903 2 ? 30 C C c B C - C C
15
905 3 ? 30 B C c B C Moll-t (C) -
Trench 11
50
EIA pit 1103 1105 1100 16 90 A A c B C - - P
KEY: A* = 30+ items, A = t10 items, B = 9 - 5 items, C = < 5 items, Moll-t = terrestrial
molluscs. Analysis: P = plant. NOTE: 1flot is total, but flot in superscript = ml of rooty material.
28
5.2 The bulk samples were processed by standard flotation methods and the results
are quantified in Table 6.
5.3 Some of the flots were very large and contained large quantities of rooty
material. While few of the samples were highly rich, most contained several
grains, glumes and spikelet forks of emmer and spelt type wheat (Triticum
dicoccum, T. spelta). The former, emmer, is of some interest as it is rarely
recorded in the Upper Thames sites to the west (Robinson and Wilson 1987;
Jones 1984; Stevens 1996). Several weed seeds were also recorded in the
samples. These were mainly of large seeded species, oats (Avena sp.), brome
grass (Bromus sp.), knotgrass (Polygonum aviculare), cleavers (Galium
aparine), corn gromwell (Lithospermum arvense) and vetch/tare
(Vicia/Lathyrus sp.). These are commonly recorded from other sites along
with some of the smaller seeded species found in the samples, annual meadow
grass (Poa sp.), goosefoot (Chenopodium sp.), dock (Rumex sp.), clover and
medick (Trifolium /Medicago sp.). Most of these species are associated with
drier soils although a possible seed of rush (Juncus/Lathyrus) was recorded.
5.4 A large range of Iron Age and Roman Thames sites have seen archaeo-
botanical studies to the west of Wittenham Clumps, especially in the area
around Abingdon and the Windrush confluence. However, fewer sites have
been studied within the Dorchester-upon-Thames region so the site is of some
importance. A point noted from previous work in the region is the difference
between sites located predominately in the west of the Upper Thames to those
lying to the east. Those around the Windrush, Mingies Ditch (Jones 1993),
Gravelly Guy (Moffett 1989), Yarnton (Stevens 1996) and to the west,
Groundwell West (Stevens 2001) had evidence for both earlier and later stages
of processing indicative of storing crops in a relatively unprocessed state,
perhaps as partially threshed ears (Stevens 2003). Those from the south-
eastern part of the Upper Thames including Ashville (Jones 1978), Mount
Farm (Jones 1984) and Whitehorse Road (Letts 1993) contained evidence for
the later stages of processing. These were characterised by the predominance
of seeds of large seeded species. Such seeds are only removed in the final
stages and so can be seen as indicative of the storing of crops in a relatively
clean state, probably as semi-cleaned spikelets or grain. The samples from
Wittenham Clumps, through the predominance of larger seeds, would appear
to represent only the final stages of processing and so in keeping with this
pattern.
5.5 One sample from Romano-British pit (6006) contained quite high numbers of
seeds of vetch (Vicia sp.). This species has been seen by Jones (1981) to
increase during the later Iron Age into the Roman period. Given the Roman
date of this sample the high numbers are then in keeping with this pattern.
29
5.7 Charcoal was noted from the flots of the bulk samples and is recorded in
Table 6. Land snails were also noticed in a number of the flots. The species
were typically open country and included Helix aspersa, Hellicella itala and
Vallonia sp. These are typical assemblages for late prehistoric and Roman
contexts. The presence of H. aspersa a Roman introduction (Kerney 1966)
confirms a Roman or later date for the sampled features.
6 DISCUSSION
6.1 Time Team’s evaluation at Round Hill has produced significant new
information on a site of regional importance. The value of the results is
enhanced by their association with the on-going work of Oxford Archaeology
and the Northmoor Trust at Castle Hill and its environs. The results provide
information on the location, extent and condition of buried remains that can be
used by the Northmoor Trust in their plans for the long-term management of
their Estate and contribute to education and outreach on the archaeological
heritage to visitors.
6.2 This discussion does not attempt to place the results of Time Team’s
evaluation within the framework of the on-going research project, as this is
best undertaken by others, and forms part of the recommendations for further
work set out below. However statements on the nature, character and condition
of the archaeology can be made as follows.
6.3 While the limited trial trenching on the summit of Round Hill failed to
produce evidence for archaeological remains, the geophysical survey and trial
trenching evaluation on the southern slopes of the Hill produced excellent
results.
6.4 The geophysical survey produced results of clarity that have enabled an
interpretative plan to be proposed. Trial trenching has tested the survey results
and confirmed their reliability. The survey results can therefore be viewed
with some confidence and provide a landscape wide picture of the archaeology
to the south of Round Hill and Castle Hill. The survey has corroborated,
enhanced and extended the results of English Heritage’s survey in 1995 (Price
1995).
6.5 The evaluation results indicated the remains of a widely spread, open
settlement, characterised by dense groups of pits. Finds of pottery and animal
bone indicate an Early Iron Age date. On at least the lower slopes of Round
Hill, well-preserved, stratified remains of contemporary occupation were
sealed beneath plough soil. The evaluation re-establishing the precise location
of Rhodes’ 1947 trench in trench 5 and confirmed the date and accuracy of his
observations. It relocated the Iron Age cobbled surface and overlying chalk
floor, although there was little to demonstrate conclusively that they lay within
a building. It is possible that they formed part of a yard. However this part of
the Early Iron Age settlement lay within an area that was not set aside for pits.
A comparable sequence comprising a cobbled surface and soil containing Iron
Age material, sealed by a chalk surface, was recorded in trench 8.
30
6.6 A number of post-holes indicated the possible presence of timber buildings,
while geophysical survey in Area 2 suggests at least one possible roundhouse.
Small square enclosures of unknown function, were also identified and were
likely to be of Early Iron Age date. A small assemblage of Early Iron Age
pottery was recovered, along with sheep/goat and cattle bones that provided
evidence of husbandry and meat processing.
6.7 No discernible evidence for Middle and Late Iron Age occupation or activity
was recorded and suggests that the land may have been in long-term pasture or
limited cultivation at this time.
6.8 The next phase of activity occurred in the Roman period. Known Roman
activity on the southern slopes of Round Hill was identified in 1947 by Rhodes
and Time Team’s evaluation was aimed at re-assessing the results of that
earlier work. The spread of Roman demolition material identified in 1947 was
relocated and shown to represent the remains of a farmstead within a ditched
enclosure, approached by an entrance to the east.
6.9 A robber trench provided evidence for at least one substantial building on the
site, which may have formed the principal structure, and have included a
tessellated floor. However it was not possible to confirm the orientation, extent
or precise location of the building nor was it possible to provide evidence for
floor levels or of domestic occupation. It is possible that the demolition debris
merely marked an accumulation of debris from a range of buildings,
incorporating at least one of flint construction with others of clay wall or
timber construction. Pottery suggests that activity occurred throughout the
Roman period, however the majority of the finds, including coins, indicated a
large-scale redevelopment of the site in the mid 3rd to 4th century AD.
7.2 The results of the project should be made available to the wider archaeological
and academic community through publication of the results in an appropriate
form. It is proposed that the results of Time Team’s evaluation form an
adjunct to Oxford Archaeology’s future academic publication of the results of
their long-term project.
7.3 The site data from Time Team’s evaluation will be condensed from this
evaluation and assessment report and the geophysical survey report. This will
31
include the background to the project and a summary of the principal findings.
Plans, sections and photographs will be incorporated as appropriate.
7.4 The following finds and environmental analyses are proposed. Appropriate
specialists, agreed in consultation with Oxford Archaeology, will undertake
the analysis with the aim of ensuring the assemblages are placed within their
local and regional context and particularly in relation to the on-going finds and
environmental analyses for the Wittenham Clumps project.
7.5 The pottery will be analysed following nationally recommended guidelines for
the recording and analysis of prehistoric pottery (PCRG 1997) and using the
standard Oxford Archaeology pottery recording system for the Romano-
British pottery. Text will present the range of types present and set the
prehistoric and Romano-British assemblages in their local and regional
context. Any implications for patterns of production and distribution will be
discussed. A representative selection of vessel forms (mainly Iron Age) will be
illustrated.
7.6 Ceramic building material will be quantified by type (e.g. tessera, tegula, etc)
and by fabric (using relevant type series created for other Romano-British sites
in Oxfordshire, held by Oxford Archaeology). Text will set out the range of
types present (including opus signinum, mortar and wall plaster), and discuss
with reference to the nature of any structure(s) present, the implications for
site status, and any links with identifiable sources of the building material.
Stone will be identified to geological source with the text summarising the
implications for usage and supply.
7.7 Selected metal objects (coins, pin, spearhead) will be submitted for
investigative conservation treatment, to aid identification and to stabilise the
objects for long-term storage. The existing coin catalogue will be updated if
necessary following conservation, as will existing catalogue-style archive
descriptions of other objects. The significance of the unusual spearhead type
will be briefly discussed, citing relevant parallels. The spearhead, and possibly
the pin, will be illustrated.
7.8 Although the Early Iron Age animal bone assemblage is not large, it is
relatively well preserved. Formal identification of species, age and size of
animals should be carried out to enable animal husbandry and consumption
patterns to be characterised and compared to similarly dated sites. Some
possibly unusual deposits of bone should be quantified to assess whether some
deposits contain the waste from specific activities or an amalgamation from
several activities. Comparison of taphonomic indicators, such as gnawing and
fragmentation, and bone element representation of bones from pits and
‘occupation’ deposits can also be carried out on a limited scale to try to
understand the different deposition practices involved.
7.9 Despite its good condition and the relatively high proportion of bones with the
potential to inform on the age and size of animals, the Roman animal bone
assemblage is too small to provide any useful information about animal
32
husbandry or consumption and deposition practice. No further work is
proposed.
7.10 Other categories of material do not warrant further analysis, but data gathered
as part of this assessment may be utilised in the final report.
7.11 The charred plant remains are of value as they provide information from one
region that has received less study within the well-studied Upper Thames
corridor. The remains provide the potential to determine the function and role
of the site (cf. Stevens 1996; 2003) within the broader Iron Age and Roman
landscape. Charcoal is, surprisingly, relatively sparse, excepting a dump in a
Roman ditch in trench 4. This paucity may indicate the excavated area does
not lie within the main foci of domestic and burning activity. Nevertheless, the
charcoal from Early Iron Age and Roman contexts has the potential to provide
some information about the character of local woodland, and evidence of
management.
7.12 Two plant remains samples have been selected for analysis (Table 6). It is also
recommended that this analysis be reported along with data from this
assessment. Two charcoal samples have also been selected from an Early Iron
Age and Roman context to examine the character of the local woodland and
evidence of any woodland management. Although preservation is good, the
snail assemblages have little information to provide any further detailed
information about the site and its environment.
7.13 A consideration of the overall results of the evaluation in relation to the wider
evidence for landuse and settlement at Wittenham Clumps should be
undertaken by an appropriate authority in the archaeology of the region.
7.14 A copy of this assessment report, along with a copy of the geophysical survey
report, will be submitted to the Oxfordshire Sites and Monuments Record.
33
8 THE ARCHIVE
8.1 The archive, which includes all artefacts, written, drawn and photographic
records relating directly to the investigations undertaken, is currently held at
the offices of Wessex Archaeology under the site code RH 03 and Wessex
Archaeology project code 52568.
8.2 It is intended that, in accordance with the wishes of the landowner, the
excavated material and records will be deposited with the Oxfordshire
Museums Service in due course. In consultation with Oxford Archaeology, the
project archive may by integrated into that of the overall Wittenham Clumps
project and deposited as a single entity.
8.3 The paper archive is contained in a lever arch ring binder file. It includes:
Project Design
Finalised Assessment Report
The geophysics report includes a record of all data, plots of the results,
interpretation with detailed comments and conclusions.
34
9 REFERENCES
Jones, M., 1978 ‘The plant remains’ in The excavation of an Iron Age settlement,
Bronze Age ring ditch and Roman features at the Ashville Trading Estate,
Abingdon, Oxon (ed. Parrington, M), London, Council for British Archaeology
Research Report 28, 93-110
Jones, M, 1981 ‘The development of crop husbandry’ in The environment of man: the
Iron Age to the Anglo-Saxon period (eds Jones, M and Dimbleby, G) Oxford.
British Archaeological Report 87, 95-127
Kelly, S E, 2000 Charters of Abingdon Abbey Part 1 The British Academy, Oxford
University Press
Kerney, M P, 1966 ‘Snails and man in Britain’ Journal of Conchology 26, 3-14
Lambrick, G, 1979 ‘The Iron Age pottery’, in Lambrick, G and Robinson, M, Iron
Age and Roman Riverside Settlements at Farmoor, Oxfordshire, Council for
British Archaeology Research Report 32, 35-46
35
Moffett, L C, 1989 ‘The evidence for crop processing products from the Iron Age and
Romano-British periods at Gravelly Guy and some earlier prehistoric plant
remains’ Ancient Monuments Laboratory rep, 46/89
Parker, J, 1885 ‘The Early History of Oxford’ Ox. Hist. Soc. Vol. 3, 109-111
PCRG 1997, The Study of Later Prehistoric Pottery: General policies and guidelines
for analysis and publication, Prehistoric Ceramics Research Group Occas.
Papers 1/2 (revised edition)
Sellwood, L, 1984 ‘Objects of bone and antler’ in Cunliffe, B, Danebury: An Iron Age
Hillfort in Hampshire; Vol. 2, The Excavations 1969-1978: the Finds.
London: Council for British Archaeology Research Report No. 52, 371-395
Stevens, C J, 1996 Iron Age and Roman Agriculture in the Upper Thames Valley
Archaeobotanical and Social Perspectives Unpublished Ph.D. thesis,
Cambridge University
36
Mapping supplied by Time Team
Licence number AL 100018665
99.798mOD
401
406 402
404 403
405 405
409
408
406
Section 2
NW SE
409 406
99.892mOD
401
402
403
404
405
407
406
0 2m
Scale for plan 1:50
Section 1 Section 2 0 1m
Scale for section 1:25
Section 2
This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Date: 06/02/04 Revision Number: 0
Wessex Scale: Plan 1:50, Sections 1:25 Illustrator: MR
Archaeology
Path: Y:\PROJECTS\52568\DRAWING OFFICE\Round Hill\Drawing Office\Report Figures (y-m)
504
512
one
503 513
Do 506
508
one View of Trench 5 looking north
one
510
Section 1
one
Section 1 Son e
one
Section 1a Section 1b
NW SE NE SW
99.89 mOD
0 1m
Scale for section 1:25
This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Date: 06/02/04 Revision Number: 0
Wessex Scale: Plan 1:50, Sections 1:25 Illustrator: MR
Archaeology
Path: Y:\PROJECTS\52568\DRAWING OFFICE\Round Hill\Drawing Office\Report Figures (y-m)
6015
6016 6006
Section 2
Section 1
Section 1
SE NW
99.89 mOD
6001
6002
6017
600
6016
6010 6011
Stone
Section 1 (before excavation of pit 6016) 6012
6015 601 6018 Flint
6014 Chal
Tile
Pot
Section 2
NW SE one
6004 6003 99. mOD
6005 6005
This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Date: 06/02/04 Revision Number: 0
Wessex Scale: Plan 1:50, Sections 1:25 Illustrator: MR
Archaeology
Path: Y:\PROJECTS\52568\DRAWING OFFICE\Round Hill\Drawing Office\Report Figures (y-m)
810
701
1004
808 702
1001
100. 0mOD
703
705
704 707
706
Section
708
1003
1002 1005 807 709
TRENCH 10 1007
Section 2
Section 2
SE NE
NW SW 101.1 8mOD
Section 1
801 802
704 803
TRENCH 7
804
810 811
807
805 806 808 809
Section 3a Section 3b
708
Trench 8 - looking south-west NW SE NE SW
1005
100.89 mOD
1000
1000
1001
1001
Stone 1002
Flint 1002
1006
imestone slates
1003
1006
1008 1005 1007
0 2m 1004
Scale for plan 1:50
0 1m
Trench - Section 1
Scale for section 1:25
This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Date: 06/02/04 Revision Number: 0
Wessex Scale: Plans 1:50, Sections 1:25 Illustrator: MR
Archaeology
Path: Y:\PROJECTS\52568\DRAWING OFFICE\Round Hill\Drawing Office\Report Figures (y-m)
10
11
03
04
Section 1
07
06
Section 1
NW SE
100.70 mOD
01
02
10 03 07
08
04
05
0
06
0 1m
Scale for plan section 1:25
This material is for client report only © Wessex Archaeology. No unauthorised reproduction.
1126
1119
1113
1123 Section
1106
1121
1103 1117
Section
Section 1 Section 2
1110
Section 1 Section 2
NE SW NE SW
1101 1102
98.9 mOD
99. mOD
1101
1104
1114
1113
1103 1115
1105 Section 2
Section 1 1116
Section 3
Section
SW NE
1107 NW SE
1108
98. 7mOD
99. 2mOD
1106
This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Date: 06/02/04 Revision Number: 0
Wessex Scale: Plan 1:50, Sections 1:25 Illustrator: MR
Archaeology
Path: Y:\PROJECTS\52568\DRAWING OFFICE\Round Hill\Drawing Office\Report Figures (y-m)