ArticlePDF Available

Large-scale systematic fluxgate gradiometry at the Roman City of Wroxeter


Abstract and Figures

Over a number of field sessions, the Roman city of Wroxeter was surveyed using fluxgate gradiometers. In the paper the authors describe the strategy for assessing a large site, approximately 0.7 km2 in size, as well as presenting the graphical results. Prior to the survey it was believed that Wroxeter was a ‘garden city’ with areas of open space inside the defensive ramparts. The results of the magnetic survey, however, suggested that almost the whole of the area contains significant archaeological anomalies. The interpretation of the data has provided a wealth of detail concerning the layout and zonation of the city. Although many of the most prestigious buildings were already known, it is the evidence for the lower status occupants of the city that is of particular note. For the first time geophysical survey has defined the settlement areas as well as suggesting industrial zones where the lower status population of Roman Wroxeter are believed to have worked. Copyright © 2000 John Wiley & Sons, Ltd.
Content may be subject to copyright.
Archaeological Prospection
Archaeol. Prospect. 7, 8199 (2000)
Received 13 August 1999
Copyright Þ2000 John Wiley & Sons, Ltd. Accepted 10 December 1999
Large-scale Systematic Fluxgate
Gradiometry at the Roman City of
*, J. A. GATER
GSB Prospection, The Old Sunday School, Kipping Lane, Thornton, Bradford, West
Yorkshire BD13 3EL, UK
Archaeometry Section, English Heritage, Fort Cumberland, Fort Cumberland Road,
Eastney, Portsmouth, Hants PO4 9LP, UK
BUFAU, University of Birmingham, Elms Road, Off Pritchards Road, PO Box 363,
Birmingham B15 2TT, UK
ABSTRACT Over a number of field sessions, the Roman city of Wroxeter was surveyed using fluxgate
gradiometers. In the paper the authors describe the strategy for assessing a large site, approxi-
mately 0.7 km
in size, as well as presenting the graphical results. Prior to the survey it was
believed that Wroxeter was a ‘garden city’ with areas of open space inside the defensive ramparts.
The results of the magnetic survey, however, suggested that almost the whole of the area contains
significant archaeological anomalies. The interpretation of the data has provided a wealth of detail
concerning the layout and zonation of the city. Although many of the most prestigious buildings
were already known, it is the evidence for the lower status occupants of the city that is of particular
note. For the first time geophysical survey has defined the settlement areas as well as suggesting
industrial zones where the lower status population of Roman Wroxeter are believed to have
worked. Copyright © 2000 John Wiley & Sons, Ltd.
Key words: Wroxeter; magnetic; fluxgate gradiometer; interpretation; zonation
The Roman city of Viroconium was, in its heyday
in the late second century
, the fourth largest
town in Britain and its remains, at the village of
Wroxeter in Shropshire, cover some 78 ha of land.
Owing to the abandonment of the city in the early
medieval period and the lack of more recent
development over much of the site, it has pre-
sented an attractive prospect for archaeological
investigation. As has been noted elsewhere in this
issue, Wroxeter has seen a number of excavations
in the past, most notably by Atkinson (1942),
* Correspondence to: Dr C. F. Gaffney, GSB Prospection, The
Old Sunday School, Kipping Lane, Thornton, Bradford BD13
3EL. E-mail:
Barker et al (1997), Bushe-Fox (1912, 1913, 1916),
Kenyon (1940, 1980), Wright (1872), Webster
(1988) and Ellis (2000). These excavations gen-
erally have concentrated on the relatively small
area of high-status buildings around the ruins of
the baths basilica, where an upstanding monu-
ment known as the ‘Old Work’ has survived.
Prior to the Wroxeter Hinterland Project (WHP),
information about the rest of the city was derived
primarily from aerial photographs. Interpretation
of these suggested large open areas within the
city and led to Viroconium being viewed as essen-
tially a ‘garden city’ with areas of open space
inside the defensive ramparts (cf. Lapper’s recon-
structions; Webster and Stanley, 1963).
The principal aim of the WHP was to inves-
tigate the relationship between the city and its
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
hinterland, and to accomplish this a detailed
understanding of the composition and demogra-
phy of Viroconium was required. Hence, it was
decided to test the aerial photographic model
with complementary evidence derived from
geophysical survey. Thus a clear, but ambitious,
goal of the project was to survey all the available
land within the Roman defences. The fluxgate
gradiometer was the chosen instrument as this is
the standard technique used throughout British
archaeology as a first-stage assessment of the
potential of archaeology at a site (Gaffney et al,
1991; English Heritage, 1995). However, a study
of previous work in the area did not give
grounds for optimism as many surveys under-
taken in the locality have been largely unpro-
ductive. The negative results generally have
been attributed to the soils in the area, which can
be described as slowly permeable, seasonally
waterlogged clayey soils that overlie till or river
terraces (SSEW, 1983). Such soils are noteworthy
for producing only slight anthropogenically
derived magnetic anomalies, often around the
1 nT level, which frequently is the value of soil
noise in Britain.
Furthermore, although it might be expected
that the soils in the city would have been altered
significantly as a result of the longevity of the
settlement, previous geophysical surveys at
Wroxeter itself had demonstrated little success
in detecting buried Roman structures, with the
exception of an early magnetometer survey over
a tile kiln on the floodplain (Houghton, 1961). For
instance, Barlett (1975), when surveying a section
of the defensive earthwork for traces of a gate-
house, found that ‘Although strong magnetic
anomalies, suggesting considerable archaeo-
logical disturbances, were detected along the city
wall line, the definition was not clear enough
to say whether a gateway had been present’
(Bartlett, 1975). Other surveys (for instance,
David and Payne, 1990) had had equally indif-
ferent results, suggesting that magnetometry
may not respond favourably on the soils of the
region. Surveys of other Roman sites in Vir-
oconium’s hinterland tended to support this con-
clusion. For example the magnetometer survey at
Whitley Grange (Ovenden, 1995) found that only
heavily burnt features were detected magneti-
cally, whereas excavation showed that wall foun-
dations and other archaeological features are still
well preserved on the site.
The pre-WHP geophysical surveys at Wroxeter
itself had been small in scale and concentrated
almost exclusively on areas either on, or just
outside, the city’s defences. Also, each survey had
attempted only to answer a specific, localized
question with no integration into a wider strat-
egy. Indeed, with the benefit of hindsight, it is
clear that these surveys were not situated in the
places most likely to contain building remains,
and the small size of the survey areas provided
too little of the background required to under-
stand the often complex magnetic response.
Hence, although expectations of the contribution
that geophysics could make were modest, it was
nevertheless felt that its potential could not be
discounted without further trials.
Research design
With the above considerations in mind a two-
stage strategy was envisaged. First, an initial pilot
survey would take place over the ‘core’ of the
site, where aerial photography and excavation
had revealed many elements of the city. Then, if
these results revealed good signal-to-noise levels,
the survey would expand toward the perimeter
of the city.
The initial pilot survey was undertaken in Field
1, which lies to the south of the excavated baths
complex. Air photographic evidence indicated
that the remains of a number of high-status
Roman buildings with stone footings were pre-
sent in this field. Furthermore, it benefited from
proximity to the extensive excavations of the
baths area as well as the availability of other exca-
vation evidence from the field itself, including
Kenyon (1980). Thus it was an ideal test bed from
which to gauge the likely effectiveness of geo-
physical techniques for mapping Roman remains
within the boundaries of the city. In the event the
first survey results immediately highlighted the
enormous potential for magnetic survey. During
this initial survey, which covered the whole of
the field, a remarkably clear plan emerged of the
road and structures to the south of the baths. The
stone structures revealed themselves as negative
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
gradiometer anomalies. This response, although
not unique, is rare on UK soils and results from
magnetically low building material embedded
within settlement soils of high magnetic sus-
ceptibility. Additional detail was identified that
was not visible in aerial photographs, suggesting
that magnetometry would be a valid technique
to assess those areas of the city previously
thought to be devoid of Roman activity. Even at
this early stage, the gradiometer data appeared
to be contradicting the ‘garden city’ view of the
city, as few genuinely open or empty areas were
revealed in the pilot survey area. Assuming that
such conditions prevailed across the whole site
then it was clear that the production of a complete
‘AÐZ street plan’ for the city might be possible.
All of the surveys were undertaken using Geo-
scan Research’s FM36 gradiometers, with the
data collected at a density of four samples per
metre along traverses 1.0 m apart. The data were
collected on the National Grid, using 100 m inter-
sections surveyed by Birmingham University
Field Archaeology Unit (BUFAU) from key points
established by a global positioning system (GPS).
The gradiometer survey covered approximately
70.2 ha of ground, in 15 fields of various sizes.
Approximately 2.8 million data points were col-
lected during fieldwork, analysed, displayed and
interpreted over a period of approximately 4
years. The majority of the survey was undertaken
in excellent survey conditions — the grass being
maintained short by sheep grazing. Some of the
larger fields were either permanently or tem-
porarily divided by metal and electric fences.
There were few other physical obstacles within
the town apart from the obvious modern features,
such as roads and the few buildings that exist at
the centre of the site. However, Field 10, which
encompasses the northeast corner of the city con-
tained an area of bushes and very steeply sloping
ground. In places the slopes are almost sheer
drops and as a result this block had to be excluded
from the survey.
The entire geophysical survey is depicted as a
linear grey-scale plot in Figure 1. In Figures 2Ð5
and Plates 1Ð4 the same data have been displayed
at a slightly larger scale by splitting the survey
area into four rough quadrants.
The interpretation of such a large geophysical
data set from an archaeological site presents a
significant challenge. It is clear from a cursory
inspection of the grey-scale plot in Figure 1 that
almost the entire 0.7 km
covered by the survey
is densely packed with anomalies of potential
archaeological significance and it was intended
that this entire area should be scrutinized with the
rigour applied to smaller, more routine, surveys.
The scale of the task dictated that geophysicists
from both the Archaeometry Section of English
Heritage and GSB Prospection would have to
contribute to the overall interpretation if it was
to be completed within the lifetime of the WHP.
This posed difficulties in ensuring that the final
interpretation plan was consistent, because the
two groups had slightly different working prac-
tices. One group preferred to draw interpretation
on paper plots and to digitize the results into a
CAD system, whereas the other interpreted the
digital data directly on-screen using a GIS system.
Three required levels of consistency were
identified: that the individual interpreters should
be consistent in the types of anomalies they ident-
ified as archaeologically significant; that similar
significant anomalies, identified by different
interpreters should be classified in the same way;
and that the results of interpretation by different
individuals should all have a similar appearance.
Additional to these criteria of consistency was the
requirement that the output of the geophysical
interpretation should be easily combined and
compared with the other georeferenced data sets
generated by the WHP.
It was agreed immediately that the only con-
venient way in which to divide the data set into
smaller units for interpretation was to use the
present land divisions and interpret the survey
field by field. As the survey work had also been
carried out in this manner it was further decided
that each field would be interpreted by the group
that surveyed it. To address the problems of con-
sistency identified above, grey-scale plots of the
data were produced by both GSB and AML and
compared to ensure that inconsistencies of
interpretation would not be introduced by dif-
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Figure 1. Grey-scale plot showing fluxgate gradiometer survey of the Roman city of Wroxeter.
ferences in data display methods. Then, prior to
any detailed post-survey interpretation, a list of
terms was agreed into which it was felt that all
the anomalies within the survey could be classi-
fied (see Table 1). The way that each classification
would be applied was also agreed, so that the
appearance of the interpretations would, as far as
possible, be consistent. It was agreed that the
resulting interpretations should all be produced
in the DXF vector format, using only lines (not,
for instance, area fills) and assigning each classi-
fication its own layer number. This format could
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Figure 2. Grey-scale plot of northwest quadrant (Fields 3, 6, 7 and 12).
be imported easily into the BUFAU GRASS GIS
system so that the overall interpretation plan
could be assembled from the individual field
A further check on the consistency of the indi-
vidual interpretations was also proposed. It was
intended that each group should provide a
second independent interpretation of one of the
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Table 1. List of anomaly classification terms
Type of archaeology Description
Wall (positive) Linear anomaly with increased magnetic signal (ca. 5Ð7 nT), sharply defined
edges and forming a rectilinear pattern with other such anomalies. Judged to
represent the stone footings of a wall forming part of a Roman building, the stone
being more magnetic than the surrounding soil
Wall (negative) As above but the anomaly exhibits lower magnetic signal than the surrounding
background. The stone footing being less magnetic than the surrounding soil
Road A broad (ca. 5 m) linear area interpreted as representing a Roman road. Either a
distinct negative linear anomaly flanked by two positive ditches, or simply a linear
absence of anomalies in an area otherwise densely packed with features. Only
marked where there is clear geophysical evidence. Isolated stretches on the
same alignment are not joined with tentative continuations
Ditch (positive) Linear anomaly with higher magnetic gradient than the surroundings (ca. 5 Ð7
nT). Judged to be a Roman ditch with a fill more magnetic than the surrounding
Ditch (negative) As above, but the anomaly has a lower magnetic gradient than the surrounding
soil. Represents a ditch filled with material less magnetic than the surrounding
Defences Magnetic anomalies associated with the remains of the Roman town defences.
Indicates general areas of raised or lowered magnetic gradient caused by the
topographic effects of the still extant earthwork remains. Where more specific
anomalies can be identified as components of the defence, these are used instead
Pit (positive) A small area (ca. 1Ð2+ m diameter) of increased magnetic response judged to
be caused by a pit-type feature with a fill more magnetic than the surrounding
Pit (negative) As above, but with a lower magnetic response than the surrounding soil,
representing a less magnetic fill
Industrial Small area, similar in extent to pit-type anomalies but with a very strong positive
magnetic gradient (×30 nT) judged not to be due to surface iron. Typically caused
by the remains of a fired clay structure such as a kiln or furnace. Where associated
with stone structures they may indicate the remains of a domestic hearth or
hypocaust system
Disturbed area (structural associations) Archaeological anomalies that appear to represent the remains of a Roman stone
building but where no clear building plan can be discerned. Usually identified by
a concentration of pit-type anomalies in an approximately rectilinear area. This
category is then used to indicate the estimated perimeter of the building
Disturbed area (archaeological) Denotes an area of disturbance, usually indicated by increased soil noise and
judged to be of archaeological significance rather than of modern origin
Modern disturbed area Area where the ground has been disturbed in the recent past. Characterized by
very large magnetic gradients and a high level of noise often accompanied by
concentrations of dipolar, near-surface ferrous responses
Modern pipe Straight, linear anomaly with very large magnetic gradients alternating regularly
between positive and negative polarity
Geological Indicates anomalies of possible geomorphological origin
Previous excavation? Area of uniform magnetic signal contained within a well-defined boundary in
regions otherwise densely covered with archaeological anomalies
fields interpreted by the other. These alternative
interpretations would then be compared to
ensure that any discrepancies were acceptable.
Unfortunately, the time constraints of the WHP
eventually ruled out this approach. However, it
is a step that might be profitably considered for
any similar project in the future.
Classification layers from the final detailed
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
interpretation are presented in Plates 1Ð4, and
the geophysical evidence from some of the more
interesting classifications are discussed in the
next section. Images of the data that comprise this
survey may be seen on the BUFAU World Wide
Web site (BUFAU, 1999). At the completion of the
project the whole data set will also be available
for downloading on the Internet and details will
be posted on this Web site.
Results of the fluxgate gradiometry data
The defences of Wroxeter are well known from
fieldwork and aerial photographs (APs) and had
been subjected to previous geophysical inves-
tigation (Bartlett, 1975). The geophysical evidence
for the defences includes that of the main town
and the different phases in its development (Fig-
ure 1; White and Barker, 1998).
Northwest quadrant (Figure 2 and Plate 1)
Field 7, which lies at the extreme northwest of
the city, incorporates a long stretch of the still
extant outer Roman defensive earthwork that
forms its northern perimeter. The most striking
anomaly is situated where the ramparts turn to
the southwest, and probably can be discounted
as a Roman feature owing to the high magnitude
of its response. It is most likely the remains of a
recent structure, possibly an emplacement from
World War II utilizing the defence’s locally stra-
tegic position.
To the south, Field 12, which lies at the western
periphery of the town, is potentially the least
informative from the whole survey. Few ano-
malies of note can be seen in the data. Even the
anomalies associated with the defences, which
run approximately northÐsouth through this
field, are barely above the limit of detectability.
Linear anomalies associated with the canabae (a
civilian area immediately adjacent to a fort) and
the later Roman defences can be seen towards the
southern corner of Field 6 and extending in an
approximately eastÐwest orientation through
Field 3. They consist of the ditch, detected as a
positive magnetic anomaly, and, 20 m further
south, a parallel rampart, the footings of which
are visible as a negatively magnetized feature.
The magnetic data strongly suggests a significant
change of character either side of the defences.
To the south of the defences are the traces of the
primary street grid, whereas to the north is a
pattern of anomalies suggestive of industrial or
other concentrated usage and set within a sec-
ondary street system.
Northeast quadrant (Figure 3 and Plate 2)
The main town defences cut a broad swathe
through the northeastern part of the survey. Field
9, which lies in the central northern region of the
town, has the city defences as its northern limits.
The magnetic responses associated with the
defences comprise a broad band of anomalies that
correspond with the ditch and bank features. It
is interesting to note that medieval ridge-and-
furrow ploughing visible on aerial photographs
as extending into this field does not appear to
be recorded in the magnetic data. This negative
result must be a product of both the natural soils
and the non-settlement activities during this
period. The results are similar to those seen over
the northwest defences in Field 7 and clearly con-
tinue south through Field 10.
The response from the defences in Field 10,
which are quite well preserved as earthworks,
is dwarfed by the very strong anomalies at the
northern edge. These are modern in origin,
coinciding with a piece of agricultural machinery
that could not be moved at the time of the survey.
Another ‘blank’’ area coincides with hay bales in
the field. Field 10 was the only field where a
reasonable area outside of the main defences was
surveyed. The absence of anomalies outside of
the defences is quite stark in comparison to the
remainder of the responses within the field. A
continuation of the canabae defences from Field 6
in the northwest quadrant can clearly be seen
within Field 5.
The road system into and within the town is well
documented, including Watling Street approach-
ing from the northeast. Naturally the modern
road system utilizes some of the Roman
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Figure 3. Grey-scale plot of northeast quadrant (Fields 5 and 8Ð12).
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
thoroughfares, which therefore were inaccessible
for survey.
Northwest quadrant (Figure 2 and Plate 1)
The northwest quadrant provides some evidence
for the secondary road system in the town. An
eastÐwest road alignment is visible just within
the defences towards the north of Field 7. This
has two clear ditches showing as positively mag-
netized linear anomalies bounding its north and
south edges. Linear ditch-like anomalies radiate
orthogonally from this road on both sides, per-
haps marking the boundaries of land plots. The
ends of the road do not appear to respect the
defensive earthwork but run into it, particularly
at the western end. This may suggest that it pre-
dates the construction of the defences, the visible
segment being incorporated into the city street
plan when it was fortified.
There is fragmentary evidence for another
street south of the Bell Brook, apparently associ-
ated with an increased density of pit- and ditch-
like anomalies. This street would not have been
parallel to the road described above, but seems
to be an extension of the street layout north of the
Bell Brook.
On the eastern side of Field 6 an interesting
Roman eastÐwest street can be seen, which con-
tinues into Field 5 to the east. Facing on to this
on its southern side, just west of the junction is
the only building with stone footings visible in
this field. The street continues 100 m into the field
and then alters alignment to run roughly parallel
to the Bell Brook, which forms the northern field
boundary, some 100 m away and may relate to
the earlier alignment of Watling Street.
Three northÐsouth streets are visible sub-
dividing this area and running from the southern
edge of the field to join the realigned section of
the street described above. It is interesting that
these streets are not detected as negative mag-
netic anomalies as they are in the fields to the
south, but rather are defined simply as linear
areas in which pit-like anomalies are absent. This
may indicate that the streets were not metalled in
this lower status quarter. A possible earlier road
alignment, also detected in aerial photographs,
can be discerned as two faint parallel linear ano-
malies running southwest through this area of
dense pit features from a point on the eastÐwest
street close to where it changes alignment. Unlike
the northÐsouth streets just described, pit fea-
tures are superimposed on it, suggesting that this
thoroughfare was abandoned as use of this area
grew more intense.
The two short stretches of road noted in Field
3 are clearly more aligned with the system noted
in the eastern part of the site.
Northeast quadrant (Figure 3 and Plate 2)
Field 5, the largest within the city, has a topo-
graphically elevated position to the northeast of
the baths. This area exhibits all of the charac-
teristics that appear to define the central part of
the site. That is, a clear regular gridiron pattern
of roads infilled with structures and settlement.
Significantly this pattern is continued to the north
of the canabae defences, which sweep round in an
arc in the northern part of the field. Although
the land descends toward the Bell Brook stream,
archaeological features are present in this field to
the limit of the survey. In comparison, the fields
to the north of the Bell Brook are less regular. In
Field 9 there is one clear eastÐwest street that
divides the field; it follows an erratic course from
close to the Watling Street gateway in the
defences towards Whitchurch Road, where the
anomalies fade in strength. In the southeast cor-
ner of the field there are suggestions of two fur-
ther intersecting streets that are also visible on
aerial photographs. However, the anomalies are
unclear and may represent ditched field systems
or land plots.
As noted above the main town defences occupy
the majority of Field 10. However, a NÐS street,
the continuation of the fortress via principalis,is
apparent in the data. It extends northwards from
Field 5 to the south of the Bell Brook and heads
towards the Watling Street gateway. An arm
extending eastwards represents another street.
The overall picture matches closely with the street
plan visible on aerial photographs. There is a
concentration of anomalies alongside these
streets, reflecting pits and some stone/brick foun-
dations. However, there are no clear building
plans visible.
Southeast quadrant (Figure 4 and Plate 3)
The southeast quadrant contains the best view
of the fossilized town. The street plan is clearly
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Figure 4. Grey-scale plot of southeast quadrant (Fields 1, 4 and 11).
visible in Fields 1 and 4, the roads showing up
as anomalies of lower magnetization, suggesting
perhaps that some metalling still survives.
Towards the western end of Field 1 there is a
departure from the regular arrangement of the
northÐsouth streets visible further west. This
reflects the organization of the underlying
fortress. The ditches to either side of the main
street in Field 4 are also particularly prominent,
showing up as linear anomalies of high mag-
netization. This may reflect the character of occu-
pation in this quarter of the city and is discussed
further below.
Although the Roman street layout in Field 4
appears to have been highly regularized, forming
a pattern of square insula blocks, the geophysical
evidence, nevertheless, suggests some modi-
fications were made to this plan. The easternmost
insulae, bordering the city defences, appear to
have been subdivided, with additional streets
running eastÐwest. Also the line of the longest
stretch of eastÐwest street running through this
field, just north of its centre, appears to have
been adjusted at some point in time. Two slightly
different alignments are visible, apparently con-
verging towards the western edge of the field.
In Field 11 the land is subdivided into three
smaller fields, as is evident in the plot of the
survey. The magnetic response has been dis-
rupted in many places owing to the presence of
extensive modern drainage works and steel live-
stock feeders. Potential archaeological anomalies
are fainter and less abundant here than in Field 1
to the north. However, the southern con-
tinuations of two streets from that field are just
visible and the whole area appears to be sub-
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
divided with linear ditch-like anomalies on the
same alignment, perhaps representing the
boundaries between plots. Some discrete posi-
tively magnetized pit-like anomalies are visible
in the data, although not in the density that is
observed in surveys of other parts of the town.
Southwest quadrant (Figure 5 and Plate 4)
The data collected in Field 13 clearly suggest that
there is considerable archaeological potential. An
Figure 5. Grey-scale plot of southwest quadrant (Fields 2
and 13).
anomaly interpreted as a road leaves the town at
its southwest corner and leads to a ford across
the River Severn. The most likely place for struc-
tures is alongside this road, adjacent to the east-
ern end, where there is an increase in noise.
The other field in this quadrant shows a less
regular road grid than observed in the eastern
part of the site. It is possible that this may be a
product of the legionary fortress’ annexe and
local topography.
The evidence for settlement features and struc-
tures is the most obvious element within the
gradiometer data set. The pilot study in Field 1
had indicated that the magnetic response over
stone buildings is likely to be negative, the result
of lower susceptibility stone embedded within
highly enhanced soil.
Northwest quadrant (Figure 2 and Plate 1)
The eastern edge of Field 6 is bounded to the east
by the modern road to Norton, which overlies the
main northÐsouth street of the Roman city. This
street is lined on the western side by an area of
dense archaeological anomalies that presumably
represent the timber strip housing of premises
exploiting this busy thoroughfare. The area
extends some 40 m west into the field and a
northÐsouth ditch running parallel to the street
defines its western edge.
Within Field 3, which lies to the north of the
forum field (2), the data exhibit a high amount of
zonation. The evidence for structures is almost
all within the southern part of the area and is also
constrained by the defences of the early (mid-
first century) legionary fortress. This area shows
a highly regular response indicating both struc-
tures and roadways.
Northeast quadrant (Figure 3 and Plate 2)
Perhaps of most interest in Field 5 is the apparent
reversal of anomaly signal over some of the sup-
posed stone structures. As noted earlier, the wall
foundations gave a negative signal in the pilot
survey area. This clearly is a result of the low
magnetic material embedded within the highly
susceptible soil associated with centuries of
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Figure 6. Comparison of grey-scale plots over burnt and non-burnt buildings in Field 5.
settlement and industrial activity at the site. The
fact that in some areas of Field 5 the structures
are defined by positive rather than negative ano-
malies (Figure 6) is unusual in Britain. It is
believed that the reversal in response is probably
the result of fire-induced enhancement of the sur-
face of the stone structures. Although it may be
suggested that the response could be a result of
the use of a magnetic building material, theor-
etical (Bevan, 1994) and empirical (Hesse et al,
1997) studies over structures made with highly
magnetic material show that this response is
unlikely. Indeed, the results are consistent with
unpublished work on a brick structure that was
destroyed by fire (Gaffney and Gater, in prep-
aration). It would appear that several hectares of
structures were burnt down in this zone.
Although it is not possible to be certain if the
burning reflects a single cataclysmic event, the
fact that the response is discrete suggests such a
conflagration. It has been suggested that there
may be some link with the fires recorded by
Atkinson (1942) in the Forum, although this is
highly speculative (Gaffney et al, 1998).
Field 8 occupies a wedge of land between
Watling Street and the main northÐsouth
thoroughfare of the Roman city, and to the
immediate north of the junction between the two.
The latter road is visible at the western edge of
the field where the line of the modern road devi-
ates from it. It would appear that strip housing
lined the eastern side of this road, similar to that
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
on the opposite side, detected in Field 6. Towards
the centre of Field 8, characteristic linear ano-
malies of negative magnetization suggest a build-
ing with masonry wall footings. It abuts the street
at its western end and its long axis runs per-
pendicular to it. It appears to have a semi-circular
apse at its eastern end. East of this structure, evi-
dence for modern intervention has prevented the
magnetic detection of Roman remains over a
large area at the eastern corner of the field.
By contrast, to the north there is a concentration
of anomalies along the eastern flank of Whit-
church Road (Field 9), including evidence in the
data for a cluster of stone foundations just inside
the city defences. Further stone foundations are
visible on either side of the eastÐwest street
referred to above. Although most of the ano-
malies are fragmentary, there is one clear struc-
ture. This building measures approximately 8 m
by 4 m and has internal divisions; the evidence
indicates a high-status building similar to others
found closer to the centre of the city. It is unlikely
to date before the mid-second century. The
results equate with a structure visible on aerial
photographic transcriptions (e.g. Webster and
Barker (1991, p. 23) but not in White and Barker
(1998, p. 30, Figure 12).
The importance of the results in Fields 9 and
10 is that they demonstrate that residential settle-
ment extends to the northern limit of the defences
on the eastern side of the city.
Southeast quadrant (Figure 4 and Plate 3)
An initial indication of the success of the gra-
diometry was provided by the remains of a large
building immediately south of the baths them-
selves (Field 1). This had been identified and
mapped by aerial photographs taken by J. K. S.
St Joseph in the late 1940s. Further information
was supplied by Kenyon and Webster when they
dug a number of box trenches across the building
during 1952Ð53 (Kenyon, 1980). They interpreted
the structure as a high-status courtyard house
and showed that initially it had been constructed
in the second-century
and was then rebuilt in
the fourth century with the addition of a pro-
jecting room with a curved apsidal wall on the
southwest corner. This room was thought to have
been a triclinium (dining room), from which the
garden around the house could be admired.
Geophysically, this building was first ident-
ified in the 1995 gradiometry survey; it was then
resurveyed with traverses 0.5 m apart to improve
the definition of the buried walls. Figure 7 shows
that the plan of remains detected by the mag-
netometer correlates well with that derived from
the excavation and air photographs. It also con-
firms that masonry wall footings are being
detected as negative magnetic anomalies, having
a lower magnetization than the surrounding
soils. A second, similar, structure may be dis-
cerned immediately to the east. Comparison with
that just discussed suggests that this is another
courtyard house. This building also appears to
have a projecting room with an apsidal wall at
its southeast corner. Ground-penetrating radar
results (Goodman and Nishimura, 2000, this
issue) from the same area indicate that the
masonry remains here are closer to the ground
surface, suggesting that, as with the first house,
this apsidal room was a later modification. It is
notable that over both buildings the magnetic
survey has responded primarily to the final
layout, with little evidence of the deeper, earlier
phases being visible.
To the east further evidence for masonry struc-
tures is evident, although no clear building plans
can be discerned. To the northeast of the court-
yard buildings described above, the outline of an
‘L’ shaped building with masonry footings has
been detected. This ‘corridor house’ has a curving
apse on its northern end and in plan it is identical
to the simpler country villas known throughout
the northwestern provinces of the empire.
Some 90 m south of the apsidal building, the
remains of a large building some 27 m in length
can be seen. This is oriented eastÐwest with the
suggestion of a curved apse at its eastern end. The
gradiometer reveals little evidence for internal
subdivision, except perhaps one northÐsouth
wall in the eastern half of the structure. Given this
shape and orientation it is tempting to suggest the
presence of a church, which would have been a
late construction in the life of the Roman town.
Without excavation this inference must remain
conjecture, but as has been noted in the intro-
ductory paper, ground-penetrating radar has
revealed that these remains are amongst the least
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Figure 7. Comparison of high sample density survey with excavation plan of Kenyon’s courtyard villa. (Kenyon 1980)
Illustration reproduced with permission of the Shropshire Archaeological and Historical Society
deeply buried in the vicinity. Further evidence
may be gleaned from the iron pipeline running
southeast across this field, which clips the edge
of the building. The excavation of the trench for
this was observed by Calvert et al, (1902), who
reported at this point the remains of a concrete
floor some 14 inches (35 cm) thick, the top of
which was 1 ft 8 in (50 cm) below the surface. The
magnetic character of the interior of this structure
is quite distinct from the general background
response, suggesting that the concrete floor may
survive throughout the building.
About 90 m southeast of the putative church
is a large structure consisting of two conjoined
courtyards. It is interesting that this building is
orientated at a slight angle to the street plan just
described. This alignment matches that of the
Roman fortress that preceded the city, part of
which was located in this field. Its plan is known
from aerial photography of cropmarks and it is
conjectured that these buildings represent a sur-
vival of structures from this earlier military
phase, perhaps continuing in use as Wroxeter’s
mansio (White and Barker, 1998, pp. 76Ð76).
Some 60 m west of this building are the remains
of another rectangular structure consisting of a
series of identical rooms arranged around a cen-
tral courtyard. This building has been further
examined using a multiplexed resistivity array
(Walker, 2000, this issue) allowing measurements
of earth resistance from a range of mobile elec-
trode separations to be made simultaneously.
This technique suggests that, as with the above
building, the foundations of this structure are
likely to be at a similar depth to those of the
fortress. Thus it has been conjectured that this too
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
is an early building that survived into the period
of the later city, perhaps seeing use as an inn or
hotel (White and Barker, 1998, p. 75).
Located 45 m further west, the footings of a
small building consisting of two concentric
squares may be discerned, the outer side of the
larger square measuring about 9 m. This building
was partially detected by aerial photography, but
the gradiometer survey allows its full plan to
be seen clearly. This plan is characteristic of a
Romano-Celtic temple and although small, other
examples of comparable size have been found in
Britain (Lewis, 1965).
Whereas all the buildings so far described from
this field have been of relatively high status, some
evidence for the dwellings of the less prosperous
citizens of Roman Wroxeter has been detected
along the western edge of the field. Here a num-
ber of rectangular structures have been detected
in close proximity to each other and with their
narrower sides fronting on to Watling Street. The
latter is now marked by the modern road through
the village that forms the western boundary of
this field. These structures are similar in plan to
the timber strip buildings excavated by Bushe-
Fox nearby (1916), on the western side of this
road, which he interpreted as being shops of craft-
smen lining the principal thoroughfare through
the town. However, unlike his buildings these
must have possessed stone footings, given the
negative gradiometer response.
Linear, low magnetic anomalies, similar to
those in Field 1, were detected in Field 4.
However, these are far fewer in number and they
do not form coherent floor plans, suggesting that
stone was not the predominant building material
in this part of the city. This would tend to suggest
that, moving east from the centre of the city, a
transition occurs from the high status residences
seen in Field 1 to lower status housing occupied
by artisans just within the eastern city defences.
It may be inferred that this latter area was densely
occupied but that the structures here were of less
imposing timber construction, leaving little geo-
physical evidence. This concentration of indus-
trial activity here might explain the highly
magnetized fills of the street ditches as being
occupation soils magnetically enhanced by the
processes taking place in the vicinity.
In the extreme northwest corner of the field a
rectangular building has been detected, com-
posed of linear anomalies of positive magnet-
ization. It is possible that this building had
masonry footings and is exhibiting the reversed
anomaly effect noted immediately to the north in
Field 5 (see above). This then might mark the
southern extent of the area affected by the hypo-
thesized fire.
A clear contrast can be found in Field 11, where
only one building plan may be discerned. This is
also the only building known from aerial photo-
graphs of the area. It appears as a complex of
faint positive linear anomalies in contrast to the
buildings in the field immediately to the north.
This house was the location of an unpublished
excavation by Houghton and this probably
explains the low peak magnitude of the response.
The absence of evidence for high-status struc-
tures and the sparseness of archaeological ano-
malies tempt the inference that this was perhaps
an area containing the residences of Viroconium’s
poorer inhabitants, their less durable structures
leaving few detectable traces. A second interpret-
ation might be that much of this area was utilized
for agriculture. This was certainly the case in the
medieval period, as attested by surviving ridge
and furrow in the western two fields.
Southwest quadrant (Figure 5 and Plate 4)
The data gives the appearance of a number of
zones within Field 2. In general the area at the
eastern edge of the field, i.e. toward the centre of
the Roman town, contains the greatest con-
centration of structural type responses. This cor-
relates with Wilson’s (1984) interpretation of the
aerial photographic evidence, although there are
some differences in the detail. Potentially, some
of the variance may be the result of excavations
that are known to have occurred in this field. A
northÐsouth road is found within the western
part of the field and a number of structures lie to
the west of it.
The evidence for boundaries is ambiguous owing
to the overall nature of the archaeology. By and
large the information on boundaries, i.e. linear
and sublinear anomalies not designated as roads
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
or defences, can be best identified in the mag-
netically quiet areas on the site. For example, in
Field 2 in the southwest quadrant, linear ano-
malies that may indicate drainage ditches, of
unknown date, have been found. These zones
correlate largely with the cropmark-free areas
noted by Wilson (1984).
Industrial features are often amongst the most
characteristic of magnetic anthropogenic
responses. However, such responses are poten-
tially more difficult to identify in an urban
context, where significantly enhanced stra-
tigraphy exists. It could be expected that the more
peripheral parts of the town are likely to contain
industrial activity on account of the potential fire
risks and the inevitable localized air pollution.
Northwest quadrant (Figure 2 and Plate 1)
In Field 6 in an area south of, and bounded by,
an eastÐwest road is a very dense area of discrete,
positive pit-like anomalies. Few linear features
are visible within this area, and its location and
the lack of evidence for residential structures
imply that these putative pits might represent
evidence for an industrial quarter utilizing the
nearby water supply. Given the likely importance
of livestock to the local economy it is tempting
to suggest that fulling and tanning were major
activities here (White and Barker, 1998, p. 79). It is
also possible that some of the discrete anomalies
detected are kilns, as the remains of one such
structure have been excavated outside the nor-
thwest gate of the Roman city, just beyond the
western edge of this field (Morris, 1935, p. 15).
Considerably enhanced magnetic responses
can be seen within the eastern extremity of Field
3, which, although archaeological, do not show
much regularity. It is likely that they indicate
significant settlement deposits. By way of
contrast, the land in the western part of the area
is magnetically quiet. Of some interest is the fact
that this part of the survey exhibits very little
overall structure, lacking even a street grid, per-
haps indicating a different use. Evidence for indi-
vidual industrial-type anomalies is also clearly
visible in Field 3. In particular there are a number
of kiln-type anomalies in the northern part of
the survey. Also of some significance are a great
number of strongly magnetized responses in this
region (paralleled in the southern part of Field 6),
which may be the result of burnt deposits along
the northern boundary of Field 3. Clear evidence
is, therefore, available in this field for zonation of
activities, although the exact function of each
zone is highly conjectural.
Northeast quadrant (Figure 3 and Plate 2)
The northeast quadrant is largely devoid of
industrial-type anomalies. However, there is an
isolated and very strong industrial-type magnetic
anomaly in Field 9, although its precise archaeo-
logical interpretation remains uncertain.
Southeast quadrant (Figure 4 and Plate 3)
It was noted earlier that the southeast quadrant
was heavily utilized; the shear number of closely
packed magnetic anomalies indicates that this
area was densely occupied. In Field 4 these ano-
malies are typically discrete and are often highly
magnetic. Some are small and of extremely high
magnitude, suggesting thermoremanent struc-
tures such as kilns or furnaces, whereas others
are of more amorphous shape, perhaps indicating
pits filled with industrial debris or burnt material.
Southwest quadrant (Figure 5 and Plate 4)
Field 2, which is known as the ‘forum field’, con-
tains many strong magnetic responses. In fact the
dominant response in this data set is from a pipe
that crosses the field in an approximately eastÐ
west orientation. This is a known modern pipe
that was placed deliberately within the main
drain of the forum. Of more interest is the strong
contrast at the eastern edge of the field. Although
this response may be the result of modern intrus-
ive work at the site, an alternative explanation is
that the marked variation exhibited in this region
could be the result of intense industrial activities
associated with shops/workshops facing the
road to the east.
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Market gardens
The pre-survey model for the site suggested that
the town was underpopulated and that market
gardens were likely to fill much of the peripheral
area within the ramparts.
Northwest quadrant (Figure 2 and Plate 1)
In the north of Field 6 the concentration of dis-
crete anomalies becomes far less dense, par-
ticularly north of an eastÐwest linear ditch
feature that can be seen running parallel to the
line of the Bell Brook some 50 m south of it. It has
been suggested that the areas to either side of the
brook were divided into plots for market gard-
ening (White and Barker, 1998, p. 67) and the
weak parallel, linear anomalies running between
the brook and this ditch might represent associ-
ated drainage channels or simply later ridge and
In general, Field 7 is less densely packed with
anomalies than those further south and this may
be because the area was used primarily for mar-
ket gardening, as has been conjectured above. A
scatter of discrete pit-like anomalies is apparent,
which might be interpreted as refuse pits, or per-
haps as indicating that some industrial activity
spilled over into this area from the southern side
of the brook.
Northeast quadrant (Figure 3 and Plate 2)
As with Field 7 there is less of a concentration of
pit-like responses in parts of Field 9, a fact that
may lend support to a view that the area had an
agricultural use in the past; perhaps garden plots.
Other features
Northeast quadrant (Figure 3 and Plate 2)
In the southern part of the Field 5 and on a topo-
graphic high point is a comparatively quiet mag-
netic zone. This equates with the area of the
suspected forum boarium (cattle market) (White
and Barker, 1998, p. 92). Although elements of
subdivision exist within this zone, it is uncertain
if it is genuinely blank or if more ephemeral struc-
tures, perhaps constructed from timber, are
present. The geophysical data, although indi-
cating more activity within this zone than is sug-
gested by the aerial photographs, could still
substantiate the forum boarium theory.
Southeast quadrant (Figure 4 and Plate 3)
One further piece of evidence suggesting where
lower status Roman structures might be located
is the relative lack of the characteristic low mag-
netization linear wall-footing anomalies visible
towards the southeastern corner of Field 1. As the
road system shows with the same clarity here
as elsewhere, it might be conjectured that the
structures here were constructed of timber rather
than stone. This apparent change in building
material coincides with an increase in discrete
circular anomalies characteristic of pits and fired
features, such as potential hearths or kilns. These
same features are abundant in Field 4 immedi-
ately to the east.
Another anomaly of note is a circular ditch
feature some 50 m in diameter towards the west-
ern edge of the field, just to the north of the centre.
This appears to contain a high density of discrete
anomalies and it has been suggested elsewhere
in this volume (Buteux et al, 2000, this issue) that
it represents the remains of a Bronze Age barrow
cemetery of a type typical in this part of the coun-
try. However, the obvious density of Roman
features in this area renders a definitive interpret-
ation impossible without the support of exca-
vation evidence.
Although geophysical survey of such a large site
seemed a daunting prospect at the outset, data
collection was relatively straightforward once the
prospecting programme was underway.
However, the resulting volume of magnetic
measurements made interpretation a monu-
mental task. This was exacerbated by the density
and complexity of the archaeological anomalies
detected, which, as is often the case with such
data sets, lent themselves to continual rein-
terpretation each time they were examined. The
collaboration of two groups with subtly different
working practices was an additional obstacle that
was overcome without much difficulty. These
problems were resolved by establishing an
C. F. Gaffney et al.
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
agreed framework for the interpretation at the
beginning to ensure consistency between the two
teams. Use of the BUFAU GIS system to collate
all the data collected and co-ordinate the survey
and interpretation work, was also a major influ-
ence on the successful outcome of the project.
Perhaps the most important result of the gradi-
ometer survey at Wroxeter is its comprehensive
nature. It provides the best approximation we
have to a complete plan of a Romano-British civ-
itas capital. The results of the magnetic survey
have revealed a well developed city, char-
acterized by clearly defined functional zones.
Such zones reflected the social, economic and
functional divisions within the town. The ability
to interpret such zonation is important. It will
allow a more sophisticated approach to the analy-
sis of the remains at Wroxeter, and allow reliable
comparison with better preserved towns else-
where in the empire.
Obviously, the interpreted maps offer few
clues to the temporal sequence of development
in Viroconium, and only a suggestion of the
maximum extent of settlement and activity
within the defences. Much of the information that
the magnetometer survey has revealed appears
to date from the most prosperous phase of the
city’s history. However, it is also apparent that
some earlier structures survived in use into this
phase (for example the mansio in Field 1). Aside
from a short section of the northern defences,
such buildings are the best evidence for the for-
tress and early city that preceded the Hadrianic
and later city. Few buildings conclusively associ-
ated with the later city are visible either, with
possible exceptions being the putative early
Christian church in Field 1 and the line of the
defences of the shrunken town. No clear evidence
for the type of excavated structures known to
have existed in the latest phases of the settlement
can be discerned (Barker et al, 1997), suggesting
that the traces of these were too faint to be
detected above the signals produced by the
deeper, more substantial remains.
The above notwithstanding, the magnetometer
survey has produced a wealth of new information
about the layout and zonation of the Roman city.
It has clearly mapped the layout of many of the
more prestigious buildings in Viroconium as well
as evidence for the probable types of structures
present in other areas within the city defences.
Paradoxically, it is those areas, which are the most
confused magnetically, that may yet have much
more to say about the story of the city. These
areas, most clearly associated with the lower
status urban population, are the least understood
in any Romano-British city because few exca-
vations have taken place within such zones. The
Wroxeter survey has, for the first time, defined
where such people lived and, through the evi-
dence for industrial activities, where they pre-
sumably worked. The survey, therefore, provides
a firm basis either for further geophysical work
or excavation based on the interpretation of the
geophysical data.
The authors would like to thank all of the sur-
veyors who have contributed to the collection,
processing, interpretation and display of this data
set — their help has made an onerous task much
easier. Apart from the authors a great number of
people have contributed to this project. From GSB
Prospection the following contributed to both
data collection and analysis: Claire Adam, Louis
Harvey, Nadia Nemcek, Susan Ovenden, Dan
Shiel, Anne Shields, Claire Stephens and David
Weston. From EH the authors wish to thank Alex
Bayliss, Peter Cottrell, Sarah Hill, Tim Horsely,
Lynn Keys and Andy Payne for data collection.
Mark Cole and Neil Linford contributed both in
the field and during data analysis. Andrew David
from EH made helpful comments on initial drafts
of the text.
English Heritage paid for the reproduction of
the colour diagrams used in this article.
Atkinson A. 1942. Report on Excavations at Wroxeter (the
Roman City of Viroconium)in the County of Salop 1923Ð
1927. Oxford University Press: Oxford.
Barker PA, White RH, Pretty KB, Bird H, Corbishley
MJ. 1997. The Baths Basilica Wroxeter. Excavations
1966Ð90. Archaeological Report 8, English Heritage.
Bartlett ADH. 1975. Geophysics:Wroxeter. Unpublished
Old Series Report 1711, Ancient Monuments Lab-
oratory London.
Large-scale Systematic Fluxgate Gradiometry
Copyright Þ2000 John Wiley & Sons, Ltd. Archaeol. Prospect. 7, 81Ð99 (2000)
Bevan B. 1994. The magnetic anomaly of brick foun-
dation. Archaeological Prospection 1:93Ð104.
Bufan 1999.
Bushe-Fox JP. 1913. Excavations on the Site of the Roman
Town at Wroxeter, Shropshire, in 1912. Research Report
1, Society of Antiquaries of London.
Bushe-Fox JP. 1914. Second Report on the Excavations on
the Site of the Roman Town at Wroxeter, Shropshire, in
1913. Research report 2, Society of Antiquaries of
Bushe-Fox JP. 1916. Third Report on the Excavations on
the Site of the Roman Town at Wroxeter, Shropshire, in
1914. Research report 4, Society of Antiquaries of
Buteux S, Gaffney V, White R, Van Leusen M. 2000.
Wroxeter Hinterland project and geophysical survey
at Wroxeter. Archaeological Prospection 7:69Ð80.
Calvert E, Southam HHR, Phillips W, Beacall W. 1902.
Some recently discovered remains on the site of
Uriconium. Transactions of the Shropshire Archaeo-
logical Society 25: 163Ð165.
David AEU, Payne AW. 1990. Wroxeter Geophysical
Survey, 1990. Unpublished Report 34/90, Ancient
Monuments Laboratory: London.
Ellis P. 2000. The Roman Baths and Macellum at Wroxeter.
Excavations by Graham Webster, 1955Ð85. Archaeo-
logical Report, N.S. 9 English Heritage: London.
Ellis P. (ed.). in press. Excavations within Wroxeter
Roman City, 1970Ð91. Transactions of the Shropshire
Archaeological Society.
English Heritage. 1995. Research and Professional Services
Guideline No. 1. Geophysical Survey in Archaeological
Field Evaluation. English Heritage: London.
Gaffney C, Gater J, Ovenden S. 1991. The Use of Geo-
physical Techniques in Archaeological Evaluations.
Technical Paper Number 9, Institute of Field
Archaeologists: Birmingham.
Gaffney VL, Gaffney CF, Corney M. 1998. Changing
the Roman landscape: the role of geophysics and
remote sensing. In Science in Archaeology:Agenda for
the Future, Bayley J (ed.). English Heritage: London;
Nishimura Y, Goodman D. 2000. Ground-penetrating
radar survey at Wroxeter. Archaeological Prospection
7: 101Ð108.
Hesse A, Barba L, Link K, Oritz A. 1997. A magnetic
and electrical study of archaeological structures at
Loma Alta, Michoacan, Mexico. Archaeological Pros-
pection 4:53Ð67.
Houghton AWJ. 1961. The Roman tilery and brickfield
at Ismore Coppice, Wroxeter. Transactions of the
Shropshire Archaeological Society 57(1): 7Ð12.
Kenyon KM. 1940. Excavations at Viroconium 1936Ð37.
Archaeologia 88: 175Ð227.
Kenyon KM. 1980. Excavations at Viroconium in insula
9, 1952Ð3. Transactions of the Shropshire Archaeological
Society 60:5Ð73.
Lewis MJ. 1965. Temples in Roman Britain. Cambridge
University Press: Cambridge.
Morris JA. 1935. A Guide to the Roman City of Uriconium
at Wroxeter, Shropshire. Shropshire Archaeological
Society: Shrewsbury.
Ovenden SM. 1995. Report on Survey at Whitley Grange.
Unpublished report 95/55, GSB Prospection: Brad-
Soils of England and Wales. 1983. Sheet 3, Midland and
Western England. Harpenden.
Walker AR. 2000. Multiplexed resistivity survey at the
roman town of Wroxeter. Archaeological Prospection
7: 121Ð134.
Webster G. 1988. Wroxeter (Viroconium). In Fortress into
City, Webster G (ed.). London; 120Ð144.
Webster G, Barker PA. 1991. Wroxeter Roman City. Site
Guide, English Heritage: London.
Webster G, Stanely B. 1963. Viroconium: a study of
problems. Transactions of the Shropshire Archaeological
Society 57(2): 112, 131. Sutton, Stroud.
White R, Barker P. 1998. Wroxeter:Life and Death of a
Roman City. Tempus Publishing: Stroud.
White T. 1872. Uriconium. A Historical Account of the
Ancient Roman City at Wroxeter. Sandford’s, Shrews-
bury & London.
Wilson DR. 1984. The plan of Viroconium Corno-
viorum. Antiquity 58: 117Ð120.
... Anomaly classification used to interpret Earth Resistance data After Gaffney & Gater (2003) and Gaffney et al. (2000). ...
... After Gaffney & Gater (2003) and Gaffney et al. (2000). ...
Technical Report
Full-text available
Beglane, F., Meehan, H. and Nugent, L. 2016. Conservation Management Plan for the Ecclesiastical Enclosure and Pilgrim Landscape of Disert, Inver Parish, Co Donegal. Unpublished report prepared for the Heritage Council.
... Some of the various successful geophysical applications achieved in the last decade are presented by Nuzzo et al. (2009) Yılmaz et al. (2018), and Yılmaz et al. (2019a and2019b). The electrical resistivity method is one of the most widely used geophysical measurement techniques (Drahor et al. 2008;Tsokas et al. 1994;Gaffney et al. 2000;Yılmaz et al. 2019aYılmaz et al. , 2019b. This method results in electrical indications of a wide variety of materials such as groundwater and mineral exploration, geotechnical and environmental research, and archeological studies. ...
Nif Mountain is located approximately at the borders of Kemalpaşa, Torbalı, and Buca districts in the east of Izmir which is located in the west of Turkey. Nif Mountain has three investigated sectors, namely as Dağkızılca, Karamatepe, and Başpınar. Surface investigations were started in 2004. The areas include the surrounding area of the monastery complex in Dağkızılca that belongs to the classic Hellenistic period. Since official excavations can only be carried out in the summer period, some illegal excavations were also encountered in the study area, and it was seen that these illegal excavations destroyed historical remains. Despite the illegal excavations, it is estimated that some ancient ruins and burial areas (Tumulus, chamber tombs, cist graves, terracotta sarcophagus, etc.) expected to be in the Dağkızılca area are still buried. For this purpose, archeo-geophysical research was carried out in the Dağkızılca area using the electric tomography method. We aimed to determine the location and depth of the buried structures and graves, as the geo-electric survey was carried out via Wenner-Alfa array along 8 profiles 1 m apart in two areas. One of them was on top of the illegally excavated grave to test the results of the electrical method. It was observed in the resistivity images that our results were successful in getting an idea of the electrical resistivity anomaly on the grave. To detect a similar anomaly in other untouched parts of the study area, measurements were taken in 7 profiles. An anomaly, similar to obtained from our test profile, was observed in one profile. To test the accuracy of this anomaly, a new measurement was taken by shifting the beginning of the measurement point on the profile. Thus, the findings were tested once again. The exact location of the buried tomb was eventually found. Consequently, this study reveals possible buried remains that cannot be observed from the surface. It was made to quickly display the location, shape, depth, etc. features. Thus, it has been shown that the ruins can be reached more quickly and at a lower cost without deterioration by guiding the archeological studies.
... The utilisation of geophysical information in this role within a GIS is also appreciated by the archaeological computing community (e.g. Buteux et al. 2000, Gaffney et al. 2000b). This has proved invaluable when monitoring landscapes that contain poorly understood archaeological resources (Chapman et al. 2001). ...
Full-text available
... Magnetic survey is usually used in archaeological prospecting to detect features such as buried walls, pottery, bricks, fire pits, buried pathways, tombs, and numerous archaeological objects. These features are detected because they are more magnetic than the host materials [37,38]. In magnetic field survey, measurement of the total magnetic field (TMF) is widely used; however, the magnetic gradient is more effective in archaeological prospection [30,39,40]. ...
Full-text available
Among all geophysical techniques, electrical resistivity and magnetic surveying as an integrative approach has been used widely for archaeological prospection at different scales of investigations. In this study, DC resistivity (1D vertical sounding and/2D/3D ERT) and magnetic surveys (total and gradient) as a multi-scale approach was applied in a highly terraneous archaeological site (Tell) with a case study to characterize and image the various archaeological assets at different depths with different spatial resolutions. Four critical zones of great interest within the considered Tell were surveyed. At the heart of the study area, three layers were depicted clearly from 1D resistivity sounding. A thick conductive zone of mostly clay is sandwiched between two resistive layers. The topmost layer contains construction debris (dated back to the Islamic Era), whereas the deeper layer could be related to Gezira sand on which the probable Pharaonic temple was constructed. A long 2D ERT profile using Wenner Beta (WB) and Dipole–Dipole (DD) arrays with a 5-m electrode spacing identified shallow high resistivity anomalies that could be related to construction ruins from fired bricks. Additionally, it succeeded in imaging the turtleback-shaped deeper resistive layer of mostly sand. At an elevated rim to the east and west of the Tell, total and vertical magnetic gradient maps clearly delineated different archaeological structures: the walls of the rooms of ancient Islamic settlers and the walls of water tanks from the Byzantine Era. Magnetic modeling assuming 2.5-dimensional magnetic models constrained by the 2D ERT inversion models could be used to create a realistic representation of the buried structures. Toward the northern part of the Tell, the joint application of the quasi-3D ERT inversion scheme and the magnetic survey revealed an anomaly of a well-defined geometric shape of an archaeological interest thought to be a crypt or water cistern based on nearby archaeological evidence. The overall results of the geophysical survey integrated with the image of some partially excavated parts provided the archaeologists with a comprehensive and realistic view of the subsurface antiquities at the study area.
... Магнитометры. Существует широкий спектр доступных, но весьма разных инструментов для измерения магнитного поля Земли (Lenz, 1990), однако для специального использования при археологическом обследовании требуется прочный и одновременно чувствительный прибор, которым можно было бы легко и быстро оперировать в полевых условиях (Becker, 1995;1997;Gaffney et al., 2000). ...
... Many years after the first applications in the late 1940s (Atkinson, 1952) geophysical methods started to be used frequently in archaeological sites around the world in the 1980s (e.g., Hesse et al., 1986;Scollar et al., 1986;Griffiths and Barker, 1994;Gaffney et al., 2000;Leucci et al., 2007). These methods that enable detecting structures hidden in the subsurface by measuring some physical properties of the subsoil such as electrical conductivity, dielectric constant, and magnetic susceptibility have become an essential tool in archaeology (Cocchi et al., 2012;Ekinci et al., 2012;Garrison, 2016;Cozzolino et al., 2020a). ...
Full-text available
Doliche is an ancient city located at the northern fringe of Gaziantep in SE Turkey. The settlement spreads over a shallow hill called Keber Tepe. However, the lack of visible remains on the surface severely limits the comprehension of the spatial organization of the site. To understand the ancient city layout, but also to identify promising starting points of archaeological investigations, geophysical exploration plays a substantial role. In 2019 and 2020 campaigns, ground-penetrating radar (GPR) and electrical resistivity tomography (ERT) surveys were conducted to determine anthropogenic structures in selected three study areas (SA) of the hill. Among them, SA-1 is adjacent to an excavation area on the south slope of the hill, where a large three-aisled Christian basilica had been discovered. SA-2 is in the eastern part of the city area, next to a sizeable civic bath building of the Roman period. Lastly, SA-3 covers the mosaic floor of the central nave and the terrace of the basilica where archaeological excavations have been carried out. Archaeo-geophysical survey presents a challenging task as Keber Tepe is mainly composed of clayish-limestone and the same material was used for architectural construction. Despite the disadvantages, GPR provided partial traces of the nave walls conforming with the existing basilica structures. The archaeological trenchs in SA-3 confirms the partial reflection properties observed in the GPR sections. ERT surveys in this area provided subsurface models containing traces of anthropogenic structures unearthed. Since the findings obtained from both methods support each other very well in SA-2, we assumed that the field condition in this region of the hill is different or buried materials have lesser clay content in comparison to SA-1 and SA-3. As a result, the archaeo-geophysical survey contributed to the understanding of the archaeological structures and help to identify promising areas for further archaeological excavations.
Full-text available
The site of Regina Turdulorum is located in the «Campiña Sur» region, south of Badajoz province, 1,5 kilometres from the village of Casas de Reina. This municipality is placed in the lowlands beside the hill called «Cerro de las Nieves». In this hill was located the oppidum which was the first roman settlement linked to Regina and its origin. This roman town belonged to the Conventus Cordubensis, in the province of Baetica. The main purpose of the foundation of the city is managing a territory defined by its agricultural resources, but especially by the presence of mines. The excavations made in the site have discovered several aspects of the urban morphology, such as the theatre, a part of the sewers network and the central area of the town, dedicated to public buildings. These studies allow determining a foundation and initial apogee phase from Julio-Claudian dynasty to late second century AD. At the end of Julio- Claudian stage or early Flavian period is observed a period of intense construction activity at the town. In this period are built the theatre around 95 A. D. in Diocletianus princedom. In the third and fourth centuries AD there was a decline phase in the settlement of the township, indicated by the amortization of some buildings in the second half of the fourth century AD. The Project presented in this paper aimed to increase knowledge about the site, considering its extension, perimeter and organization of urban space. Geophysical surveys were used to achieve this objective. Two seasons of magnetic surveys, in 2009 and 2010 and one electrical surveys campaign were conducted in the site. Some archaeological trenches were made to verify the results from geophysical surveys. The geophysical studies have been complemented by the unification of mapping and topographic data and georeferencing of planes from past excavations. Both, geophysical and planimetric information, have been integrated into a GIS.
This report presents the findings of the survey of 75 acres and the excavation of 28 cubic meters of site 41TR198 (Crooked Oxbow Site) within the Riverside Oxbow Project sponsored by the U.S. Army Corps of Engineers, Fort Worth District, in partnership with the City of Fort Worth and the Tarrant County Water District. Planned impacts from this proposed project include habitat restoration, channel reestablishment, vegetation plantings, new roads, and sports field construction. The deepest impacts planned for the Area of Potential Effects are one meter and involve the excavation of a shallow lake utilizing the relict oxbow bordering site 41TR198. Impacts planned for the remainder of the project area will be less than one-half meter deep. Overall, the project will attempt to use the existing landscape as much as possible in order to reduce impacts. As a federal agency the USACE is required to undertake cultural resource investigations for their projects in accordance with the National Historic Preservation Act of 1966, as amended through 2001. Since the Tarrant Regional Water District is the landowner and co-sponsor of the project, and a political subentity of the state of Texas, this project was also conducted under Texas Antiquities Permit No. 5040.
Full-text available
This paper provides the overall archaeological context of a geophysical survey carried out at the Roman town at Wroxeter, including both excavation carried out within the town and survey and excavation within its hinterland. The geophysical survey at Wroxeter was carried out within the remit of the Wroxeter Hinterland Project. The archaeological rationale behind this project is explained and the contribution of geophysical survey data to achieving the goals of this project is discussed. Copyright © 2000 John Wiley & Sons, Ltd.
Most of the Roman city of Viroconium Cornoviorum at Wroxeter (Shropshire) was purchased for the nation in 1973 to protect its remains, which mainly lay in arable land, from further damage by the plough. This part of the city, beyond the limits of the Department of the Environments's Guardianship site, is known very largely from crop-mark photography, tracing the lines of buried foundations by corresponding patterns in a growing cereal crop (Frere & St Joseph, 1983, 162–6). This technique continues to yield results, for ploughing at Wroxeter is regulated rather than prohibited, and never to better effect than in the dry summer of 1975, when excellent marks developed to west and north of the Baths (PLS. XIV, XVI), and in 1976, which afforded new details to south and east of them (PL. xv). At the request of the Inspectorate of Ancient Monuments and Historic Buildings vertical photographs were taken by the University of Cambridge to form the basis of a new plan of the known remains. The potential of air photography for reconstructing such a plan had been shown by Webster and Stanley (1964), using the photographs of Mr Arnold Baker, but much more had been learnt about certain areas both by Mr Baker and by Cambridge University since then.
This experiment was done at the foundation of a chapel that was constructed about 300 years ago at St Mary's City, Maryland, in the USA. The magnetic anomaly of this foundation is irregular and lumpy. The magnetic map does not show a uniform band with a high magnetic field along the line of the foundation; instead, it shows a series of high readings spaced unevenly near the foundation. This irregularity is caused by the addition and subtraction of the remanent magnetization of randomly oriented bricks in the foundation. Most brick foundations will probably show this irregularity, which will increase where the bricks are larger, the wall is smaller, or the magnetic sensor is closer to the foundation. This irregularity can make it difficult to delineate a buried foundation by magnetic survey. Several bricks were removed from the chapel's foundation and detailed magnetic maps were made of them. An analysis of these maps allowed the remanent and induced magnetization of the bricks to be determined. Having determined typical magnetic parameters for the bricks, simulations of the magnetic maps of brick foundations were constructed. These explain the irregularity of the anomalies.
This paper describes the results of high spatial resolution magnetic and resistivity surveys of the Loma Alta site in Michoacan, Mexico. This site is the largest of a series of man-made earthen islands built up within an ancient freshwater basin. Occupied during three main phases from 100 BC to AD 850, the site now shows no surface detail of the underlying structural complexity. Test pits and trenches excavated during several field seasons revealed a fraction of the large ceremonial complex and provided isolated archaeological data difficult to interpret.Geophysical prospection aimed to investigate the so far unknown settlement pattern and detect architectural remains. Magnetic prospecting of the stone architecture turned out to be excellent due to the high contrast between the volcanic rock and the sedimentary fill. A fluxgate gradiometer survey of the entire mound surface defined a large number of small aligned positive and negative magnetic anomalies. Electrical surveys complemented and verified the magnetic results. A series of verification excavations exposed structural elements predicted by the prospecting and aided a further interpretation of the data. The major results of the study are discussed, as they provide a clear image of the orthogonal layout of the highly structured ceremonial site. © 1997 John Wiley & Sons, Ltd.
Two seasons of ground-penetrating radar were undertaken at selected locations within the Roman town of Wroxeter. Both surveys used a GSSI system incorporating a 300 MHz antenna. A strategy has been developed to collect large data sets in an efficient manner. The results have been processed and displayed as time-slice images to facilitate interpretation of complex data collected over part of this Roman urban landscape. Copyright © 2000 John Wiley & Sons, Ltd.