Content uploaded by Ben R. Pears
Author content
All content in this area was uploaded by Ben R. Pears on Dec 24, 2017
Content may be subject to copyright.
The Fields of Britannia: Continuity and
Discontinuity in the Pays and Regions of
Roman Britain
Stephen Rippon, Chris Smart, Ben Pears and Fiona
Fleming
University of Exeter, UK
Abstract
The transition from Roman Britain to medieval England and Wales clearly
saw profound changes in society and landscape, with the large-scale
abandonment of the settlements most closely associated with Romanitas –
villas and towns – and the emergence of new architectural styles, burial rites,
and other material culture of Germanic character. These changes in the
archaeological record suggest profound social dislocation for the higher
echelons of society, but have deflected attention away from what may have
been a very different story for the majority of the rural population. This article
offers a preliminary description of the results of the Fields of Britannia
Project, which is examining the potential for continuity and discontinuity in
agricultural landscapes across the different regions of Roman Britain. Three
strands are explored: the palaeoenvironmental sequences that record how
patterns of land use changed over time, the relationship between excavated
Romano-British field systems and those of the medieval period, and the ways
in which settlement patterns evolved. All three point to considerable
potential continuity and a lack of evidence for large-scale post-Roman
abandonment of the rural countryside in lowland areas.
keywords Roman-medieval transition, palaeoenvironmental analysis, field
systems, settlement patterns
The rural landscape is a complex palimpsest created over several millennia by
many generations of farming communities. In Britain, the distinctive character-
istics of the countryside are numerous, but include its intricate patterns of hedged,
banked and ditched fields, and the way that landscape character varies from
district to district. The traditional view is that the landscape of today essentially
dates back to the medieval period, but the idea that villages and open fields were
landscapes, Vol. 14 No. 1, June, 2013, 33–53
ßOxbow Books Ltd 2013 DOI 10.1179/1466203513Z.0000000005
created in the fifth and sixth centuries by heroic Anglo-Saxon colonists felling
woodland and driving out the surviving native population into the far western
parts of Britain has long been rejected, and has been replaced by a broad consensus
that the origins of villages and open fields lie somewhere between the eighth and
the twelfth centuries. This article will therefore focus on what happened before
that landscape reorganisation, in the period between the fourth and the eighth
centuries. It is easy to assume that the first few centuries of the medieval period
were overwhelmingly different both to what went before (the sophisticated
economy and lifestyles of the Roman period) and to what came afterwards (the
reorganisation of the countryside and re-emergence of urbanism as England
emerged from the chaos of the migration period). But what survived from the
Roman period to influence the landscape of medieval England, and what was the
contribution of the majority non-elite rural population to shaping our countryside?
We are all aware that some of our long, straight roads have Roman origins, and
that some of our towns lie on the site of Roman predecessors, but what was the
legacy of Roman Britain in the medieval and later countryside: what have the
Romans really done for us?
The Fields of Britannia Project is a three-year investigation, generously funded by
the Leverhulme Trust. It explores the transition from Roman Britain to medieval
England and Wales from a broad landscape perspective, reaching beyond the
traditional site-based approach. The focus is on what became of the Romano-British
rural landscape, and to what extent it has determined the character of today’s
landscape. This focus will inform debates over how regional variation in medieval
landscape character came about, but the reason why some parts of Britain but not
others saw the creation of villages and open fields is not considered as this has been
discussed extensively elsewhere (Lewis et al. 1997; Roberts and Wrathmell 2002;
Williamson 2003; Rippon 2008; 2012). There are three principal areas of research:
NLand use: An analysis of palaeoenvironmental sequences in order to
determine patterns of continuity or discontinuity in land management
practices from the late Roman through to the early medieval period at a
variety of scales (pays, regions, and nationally).
NField systems: Studying the extent of possible continuity or discontinuity in
the physical fabric of the countryside by examining the relationship between
late Romano-British landscapes and their medieval successors again at a
variety of scales (pays, regions, and nationally).
NSettlement patterns: An exploration through three case-study counties of the
extent to which there was continuity or discontinuity in settlement patterns in
different regions of Britain from the late Roman through to the early medieval
period.
Regions and pays
Traditional approaches to the sub-division of Roman Britain have been highly
simplistic, with binary distinctions such as civilian and military, upland and
lowland, and native and villa being typical (e.g. Haverfield 1912; Fox 1932; Dark
34 STEPHEN RIPPON et al.
and Dark 1997). In contrast, definitions of medieval and later landscape character
areas have been more sophisticated, and within Roberts and Wrathmell’s (2000;
2002) three-fold division of England into the South Eastern, Central, and
Northern & Western Provinces, for example, they identify a series of smaller
character areas. Agricultural historians have similarly identified relatively
numerous divisions of the landscape, each with their own distinctive character;
in France such regions can be called pays, but in English historical scholarship they
have most commonly been called ‘farming regions’ (e.g. Thirsk 1967; 1984; 1987).
While regional variation in landscape character has been the subject of extensive
research by medieval archaeologists, historians, and historical geographers (e.g. Lewis
et al. 1997; Roberts and Wrathmell 2002; Williamson 2003; Rippon 2008; 2012), this
has not been the case for the Roman period where there remains a strong tradition of
structuring discussion in broadly thematic terms, such as towns, the military
establishment, trade and industry, and the countryside. This leads to an impression
of homogeneity across large parts of the province (e.g. Collingwood 1930; Frere 1967;
Millett 1990). The substitution by Dark and Dark (1997) of the term ‘villa landscape’
for ‘civil zone’, and of ‘native landscape’ for ‘military zone’, only reiterates this existing
over-simplification and adds an additional and misleading dimension by implying that
all lowland areas were characterised by villas when this is obviously not the case, as any
map of the distribution of villas shows (e.g. Jones and Mattingly 1990). In recent years,
however, there has been a greater appreciation of regional variation in the Romano-
British landscape, with simplistic ideas that some areas lacking villas were simply
imperial estates, for example, having been challenged (e.g. Rodwell 1978; cf. Hingley
1989; Millett 1990, 120). David Mattingly (2006) in particular rejects the traditional
thematic approach to discussing Roman Britain (the military, towns, the countryside
etc.) in favour of exploring how three ‘communities’-military, civil (urban), and rural –
interacted with each other across twelve regions. In his discussion of the development
of regionally distinctive societies he prefers the term ‘discrepant experience’ to
‘Romanization’, which he defines as ‘the co-existence of very different perceptions of
history, culture, and relationships between colonizer and colonized’ (Mattingly 2006,
14, 17). Pitts (2008) has also discussed ‘Romanization’, identity, and the affects of
Roman imperialism, while Sargent (2002) has considered how there were different
processes of acculturation in the south/east and west/north of Britain. In the most
recent discussion of rural settlement in Roman Britain, Jeremy Taylor (2007) also
recognises that there may have been significant differences in landscape character, but
his use (Fig 4.15) of twenty-first-century administrative regions is wholly inappropriate
for analysing Roman Britain (e.g. ‘The South West’, according to this definition,
embraces both highly Romanised Dorset, Somerset, Gloucestershire, and Wiltshire,
and the distinctly un-Romanised Devon and Cornwall).
In the Fields of Britannia project, in contrast, an attempt has been made to study
sub-divisions of the landscape that would have been meaningful to people living
there in the past, in order to achieve a more sophisticated understanding of the
Romano-British landscape. For its purposes, it therefore divides Britain into a
series of nine regions based on common physical and cultural characteristics in both
the Roman and the medieval periods: the South East (south of the chalk escarpment
that runs from Dorset through to the Chilterns), East Anglia, the Central Zone, the
THE FIELDS OF BRITANNIA 35
South West, Lowland Wales, the Western Lowlands, the North East Lowlands, the
Northern Uplands and Upland Wales (Figure 1). The Central Zone, for example, is
characterised by its fertile lowland topography, relatively Romanised landscape, and
a reorganisation of the countryside in the late first millennium A.D. that saw the
creation of villages and open fields. The South West (west of the Blackdown Hills on
the Devon/Somerset border), in contrast, has a mix of upland and lowland
topographies, a comparatively un-Romanised landscape, and a medieval country-
side characterised by relatively dispersed settlement and small open fields set within
extensive areas of ‘closes’ (enclosed fields). There is also variation within regions,
however, and in order to reflect the complex interplay of cultural and physical facets
of landscape character, and to achieve a more nuanced understanding of local
landscape development, the project has adopted the concept of pays to subdivide
further the countryside within our nine regions.
Continuity and change in patterns of land use
Before exploring what became of the field systems of Roman Britain – that
inherently involves looking at detailed, local-scale data – it will be useful to
examine the ‘big picture’, with regards to developments in the countryside during
the late Roman and early medieval periods, and in particular broad-scale patterns
of land use. Palaeoeconomic data (faunal and plant-macrofossil assemblages) have
been collected through excavations since the 1970s, and there is now a large
figure 1 Interpretations of
major regions in Roman
Britain and early medieval
England and Wales.
36 STEPHEN RIPPON et al.
corpus of data to study. Individual assemblages only provide information on the
local economy of a particular community, and it may be that a single site is not
typical of its wider area. A comparative analysis of all the assemblages from a
particular pays, in contrast, can be used to reconstruct and compare the animal
and crop husbandry regimes across and between regions. For example, on non-
villa rural settlements within the chalk downland of Wiltshire and Dorset, cattle
comprise 32 percent of the major edible domesticates, compared to 62 percent for
sheep/goat, and six percent for pig (a sample of 14,868 bones from 19 sites), which
compares to 64 percent cattle, 25 percent sheep/goat, and ten percent pig on the
Boulder Clays of Essex (3,526 bones from eight sites), and 20 percent cattle, 73
percent sheep/goat, and eight percent pig from the South West peninsula (1,725
bones from three sites; Rippon 2012; Rippon et al. forthcoming).
Clearly, there were dramatic differences in farming practice across Roman Britain,
but what animal bone and plant macrofossil assemblages (‘on site’, palaeoeconomic)
do not tell us is the relative importance of animal husbandry, arable cultivation, and
woodland within the wider landscape. For, this we must turn to ‘off-site’
palaeoenvironmental sequences, most notably pollen. When Petra Dark (2000) first
mapped the well-dated pollen sequences for Roman Britain, the available data was
noticeably thin, particularly in lowland areas, and especially compared to the data
available for the prehistoric period. Thankfully, this has now changed and the Fields
of Britannia Project has been able to collate the evidence from over 200 Roman to
early medieval sequences that now include significant numbers in all nine regions
within Roman Britain (although within these regions some pays still lack data, most
notably in chalk and limestone areas). At the time of writing, final calculations of
percentage total land pollen across our regions and pays are still not available as new
sites are being added to the database, but the interim statistics are enlightening.
The pollen data can be analysed in a number of ways that embrace variation in
space and change over time at a variety of scales. In Figure 2, for example, the data
from all the sequences within each region have been aggregated to calculate the
average percentage total land pollen (TLP) for the four major categories of land use:
woodland, arable, improved grassland, and unimproved pasture (including heath-
land, moorland etc.). The data can be compared for the Roman and the early
medieval periods. The difference between certain regions is profound: in the Central
Zone, for example, just 14 percent of the TLP during the Roman period was from
woodland, while in East Anglia the figure is 16 percent: these can be compared to the
higher figures of 31 percent in the South East, 27 percent in the South West, 33
percent in the Western Lowlands, and 23 percent in the North East Lowlands. The
trends over time, and how these varied across different regions, show a greater
diversity: in the Central Zone, for example, woodland pollen increased only very
slightly over the early medieval period as a whole (from 14 percent to an average of
15 percent of TLP), while in the South West it increased from 27 percent to an
average of 32 percent (Figure 2). Overall, however, the picture is remarkably stable
particularly in East Anglia and the Central Zone: while some individual sites, pays,
and regions show a rise in woodland pollen, others actually show a decline, with the
aggregate figure for lowland areas of 26 percent TLP for the Roman period and 29
percent for the early medieval period, suggesting a broadly stable area of woodland.
THE FIELDS OF BRITANNIA 37
Where the quality of the radiocarbon dating permits the early medieval period as
a whole to be subdivided into three periods (the fifth, the sixth to mid-ninth, and
the mid-ninth to mid-eleventh centuries), some interesting variations in both time
and space emerge. There is, for example, a more significant increase in woodland
in the South East during the fifth century than in the adjacent Central Zone
(Figure 3). The interpretation of the increase in tree pollen seen in some areas
during the fifth century is, however, far from straight-forward and need not imply
a widespread woodland regeneration, but rather a failure to maintain laid hedges
and coppiced woodland, which allowed the trees there to grow to maturity,
blossom, and so produce pollen (e.g.Hicks 2006), while the fields enclosed by
these hedges remained in use, for pasture if not arable.
Just what happened in the migration period has led to profoundly polarised views
among archaeologists, historians, and place-name scholars with some still arguing
for a mass-folk movement from the continent, while others suggest that the majority
of the population were of British descent and that the Anglo-Saxon contribution was
simply an elite take-over (for contrasting views see Higham 1992; K. Dark 1999;
2000; Ha
¨rke 2002; Pryor 2004; Hills 2009). These debates have focused on the
demographics of early medieval England and social relations between the native and
the immigrant populations, but there is another very important dimension to the
‘continuity debate’: what happened to the physical infrastructure of the late Roman
landscape, and how it relates to the fabric of the early medieval countryside. The
figure 2 Pie charts
demonstrating continuity
and change in the major
types of land use in each of
the regions from the Roman
through to the early medie-
val period (represented by
percentage of total land
pollen).
38 STEPHEN RIPPON et al.
traditional model is one of discontinuity: a late Roman landscape that was relatively
well-wooded to begin with, and from which the native population fled in the fifth
century leading to woodland regeneration. The revisionist view is one of far greater
continuity, with the native population co-existing with the immigrants and
continuing to farm the land, albeit under the control of a new social elite. Not
only is there little agreement over what happened in the fifth and sixth centuries, but
recent archaeological work is showing how some parts of late Roman Britain – such
as northern Kent and southern Essex – appear to show a contraction of settlement
long before the advent of the Anglo-Saxons (e.g. Medlycott et al. 2010; Biddulph
2011). Other areas, however, show continued prosperity well into the fourth
century, and radiocarbon dating is now revealing that occupation on some late
Roman sites clearly continued into the early medieval period (e.g. Evans and
Hancock 2005). The vast number of Romano-British sites found in open
countryside with no later settlement nearby, create the impression of a major
episode of discontinuity, but a more careful examination of the evidence is
increasingly showing that thedecline of these sites was a long drawn out process that
unfolded along different timescales in different areas, and involved evolution and
gradual change as much as simple and sudden abandonment. Herein lies one
example of the significance of scale: that the trajectories of landscape change over
time varied from region to region, and pays to pays, and site to site.
Understanding the latest phases of occupation on late Roman sites also makes us
think about what we mean by ‘continuity’ and ‘discontinuity’. North Shoebury in
Essex serves as a good example (Wymer and Brown 1995). Over the course of the
Roman period, trackways and field systems were reorganised on a number of
occasions, reminding us of how landscapes are constantly evolving. A small
cemetery close by a late Roman enclosure is dated to the fifth century, and
contained a range of distinctively Germanic grave-goods including decorated
cremation urns, which point at the presence of an immigrant population, an
indication of discontinuity. Yet its relationship to late Roman enclosed land
suggests that the early Saxon community was located with reference to an existing
landscape. Whether that existing landscape was still functioning is unclear: the
latest Roman pottery from North Shoebury cannot be dated much beyond A.D.
350 (even though this is a region with a ceramic sequence that continues until the
figure 3 Histogram
illustrating change in
woodland in each of the
regions from the Roman
through to the early
medieval period
(represented by percentage
of total land pollen).
THE FIELDS OF BRITANNIA 39
late fourth century: e.g. Biddulph et al. 2010, 123), which at face value suggests
that the landscape was abandoned in the second half of the fourth century.
This site, however, lies on fertile brick-earth that represents some of the best
soils in the region and it seems illogical that it would have been deserted. Indeed,
two of the major ditches in the late Roman field system produced early Saxon
pottery from their upper fills, which raises two possibilities. One is that the late
Roman settlement and its field system was abandoned – hence the lack of late
fourth-century pottery – but still survived as earthworks to be re-used in the fifth
century. Or, the field system could have continued in use as pasture without
manuring, with the cessation of ploughing leading to a decrease in silting within
the ditch, which will soon have stabilised as an earthwork, with ploughing and
manuring resuming in the fifth century. The first possibility – the complete
abandonment of a lowland landscape – is unlikely as if the land was not actively
managed through grazing it would have been lost to scrub and woodland
regeneration within decades rather than centuries (Harmer et al. 2001). North
Shoebury therefore illustrates the complexities of what we mean by ‘continuity’
and ‘discontinuity’ in the early medieval period: there was clearly a discontinuity
in demographic terms (the native British population being replaced by an
immigrant Germanic community), perhaps a discontinuity of the type of land use
(a shift from arable to pasture, although this would have been in late fourth
century), but broad continuity in the sense that this landscape must have remained
as managed farmland that did not revert to woodland.
figure 4 Hunts Hill Farm,
in Upminster, Essex: a
multi-period landscape in
which the historic
landscape depicted on the
Ordnance Survey First
Edition Six Inch maps of the
1860s is clearly rooted in
the pattern of rectilinear
trackways and fields that
was established in the late
prehistoric period (after
Howell et al. 2011). ß
Crown Copyright and
Landmark Information
Group Limited (2012). All
rights reserved (1872).
40 STEPHEN RIPPON et al.
The legacy of Roman farming: the survival and loss of field systems
The mapping of Roman field systems
The palaeoenvironmental and palaeoeconomic evidence briefly discussed above
clearly demonstrates that in discussing the landscape of Roman Britain we must go
beyond simplistic binary distinctions such as civilian and military, upland and
lowland, and native and villa landscapes, and explore the differing experiences of
individual regions and pays. The same is true when studying the physical fabric of
the Roman countryside, and how it fared in the early medieval period. A second
strand of research within the Fields of Britannia Project is therefore looking at
what happened to Romano-British field systems and trackways.
Data collection was carried out in two stages: first, across the whole study area
(Roman Britain south of Hadrian’s Wall) sites were identified by searching
published sources and grey literature available through the Archaeology Data
Service (ADS), and second, the Archaeological Investigations Project (AIP) was
used to identify counties with a large amount of other grey literature, and a sample
of Historic Environment Records were contacted. For each excavation, the plan of
the Roman ditches and associated features transcribed from aerial photographs and/
or geophysical surveys were scanned and plotted against the OS 1
st
Ed. Six Inch map
using GIS (eg. see Figure 4). Historic landscapes were divided into four simple
categories (based on the morphology of the OS 1
st
Ed. Six Inch map): closes (i.e.
enclosed fields), former medieval open field, late enclosure (Figure 7), and
‘indeterminate’. For each region, the data was tabulated so that comparative
statistics could be prepared (see below).
This process allowed the project to explore the post-Roman life of fields in detail
on multi-period sites in regions with a continuous or near continuous ceramic
sequence. For example, excavation in advance of aggregate extraction at Hunts
Hill Farm, in Upminster, Essex, situated on the Thameside gravel terraces, revealed
a multi-period landscape complex of settlements and field systems across an area
of 16 hectares (Howell et al. 2011). A series of rectilinear enclosures dating to the
late Bronze Age, Iron Age, Roman, Saxon, and medieval periods all share the same
orientation, and this rectilinear pattern clearly formed the foundation of the
present historic landscape (Figure 4). The implication, therefore, is that the
fieldscape of today is the product of sequential development within a common
framework over several millennia, without significant periods of abandonment.
This process may account for the origins of ‘co-axial’ field systems, which in areas
such as East Anglia have been proposed to pre-date Roman roads (e.g. Williamson
1987; 1998; cf. Hinton 1997), but unfortunately there is no new excavated
evidence with which we can further the arguments with regard to their date (but
see Bryant et al. 2005 for an excavated example in Hertfordshire).
Although large-scale excavation of multi-period sites like that at Hunts Hill
Farm allow detail site biographies to be written, such cases are relatively few in
number. Far more common are smaller scale excavations that reveal only
fragments of Romano-British field systems and no medieval field boundary ditches,
but where the relationship between the excavated features and later landscape can
still be explored by placing it in the context of historic cartographic sources such as
THE FIELDS OF BRITANNIA 41
the First Edition of the Ordnance Survey Six Inch to the Mile maps of the mid to
late nineteenth century (hereafter ‘OS 1
st
Ed. maps’).
The same can be done with geophysical surveys and features plotted from aerial
photographs, where these field systems can be dated. Figure 5, for example, shows
firstly a transcription of an aerial photograph showing a cropmark complex at
Podimore, near Yeovilton in Somerset (Leech 1975), secondly the Iron Age and
Roman features redrawn alongside those of late Roman date identified through
geophysical survey and excavation at an adjacent site (Lovell 2006), and thirdly, the
composite of these brought together in the context of the landscape shown on the OS
1st Ed. map. Some Romano-British features can be seen to share the same alignment
as the earthworks of later ridge and furrow, as well as the orientation of the present-
day fields derived from the enclosure by agreement of medieval open field strips,
suggesting that elements of the Roman landscape remained into the tenth century
when open fields in this region were probably established (Aston and Gerrard 1999;
Rippon 2008; 2012). Although most elements of the Romano-British landscape
were thereafter lost (replaced by strips within an open field that itself was then
enclosed), a small number of features, such as parts of the trackway, survive into the
historic landscape of today (e.g. east of Orchard Farm: Figure 5B and D).
Core to the Fields of Britannia Project has therefore been the superimposition of
excavated and dated, Romano-British features onto the OS 1
st
Ed. map, and then
establishing one of three relationships between these two landscapes: i) Romano-
British fields on a completely different orientation, thus no evidence for continuity
whatsoever; ii) Romano-British field boundaries with the same orientation as those
on the OS 1st Ed. map, which raises the possibility of continuity; and iii) excavated
Romano-British ditches sharing the same specific alignment as a feature within the
later landscape, which implies a greater likelihood of a form of continuity
(Figure 6). This latter phenomenon has been recognised for some time, for
example in Taylor and Fowler’s (1978) short paper on ‘Roman fields into medieval
furlongs’ that identified a number of examples of Roman ditches under the
headlands of medieval open fields. More recently, Stephen Upex (2002) has
collated similar examples elsewhere within the East Midlands, and an example has
also been excavated at Shapwick in Somerset (Gerrard with Aston 2007). In areas
beyond the open fields, there are also published examples of Roman or earlier
ditches exactly aligned with medieval or even extant field boundaries (e.g. Percival
and Williamson 2005), and a major aim of the Fields of Britannia Project has been
to determine just how common this phenomenon is within local pays and wider
regions. While this close physical relationship between Romano-British and
medieval field boundaries does not prove that a field system was in continuous use
throughout this period, as an abandoned earthwork may have been re-dug in the
medieval period, it does suggest that the landscape continued at least to be grazed
by livestock and that there was no woodland regeneration. There is, of course, the
possibility that the character of the physical environment – topography and
drainage – may have led to similar field patterns at different periods. We have been
careful to identify such sites (those adjacent to a wetland edge for example) but
that it is also common to find field systems of different orientations laid out across
the same hillside does imply that this deterministic view may be too simple.
42 STEPHEN RIPPON et al.
Scales of analysis: the local, regional and national picture
The biographies of individual sites present fascinating insights into continuity and
discontinuity in particular locales. At Saxon’s Lode Farm, Worcestershire
(Figure 8), in the Worcestershire Plain pays of the Western Lowlands region, for
example, the boundary features associated with sunken featured buildings, one of
which yielded mid sixth- to mid seventh-century radiocarbon dates (and which
also respected the position of the Roman features), followed the orientation and
alignment of a series of Romano-British enclosures and field boundaries of second-
to early third-century date (Barber and Watts 2008). The boundary sequence along
the northern edge of the site, where mid-Roman and early medieval ditches were
cut on the same alignment and were sealed beneath the line of the historic
hedgebank, implies that elements of the mid-Roman fields remained visible beyond
the end of the fourth century to influence the layout of the early Saxon farmstead
and subsequently the pattern of medieval furlongs that are still fossilised in the
historic landscape of today. Exemplfying potential continuity from the Roman
through to the early medieval period, Saxon’s Lode Farm is typical of its pays (the
figure 5 Podimore and Yeovilton, Somerset. A wide range of archaeological evidence (aerial
photography, geophysical survey, and excavation) can be used to study the relationship of
the Romano-British countryside and the overlying medieval landscape: perpetuation of the
line of many Romano-British features in the later open fields suggests broad continuity in the
management of this landscape (after Leech 1975; Lovell 2006).
ßCrown Copyright and Landmark Information Group Limited (2012). All rights reserved
(1888–89).
THE FIELDS OF BRITANNIA 43
Worcestershire Plain), where six out of nine Romano-British sites share a common
orientation or alignment with the historic landscape, a figure (67 per cent) that is
just above the average for the Western Lowlands Region as a whole (c.58 per
cent).
At Brizen Farm, in Shurdington, Gloucestershire, in the Vale of Gloucester pays
of the Central Zone region, in contrast, there is a clear difference between the
layout of Roman ditches and the enclosed common field strips that are fossilised
within the later landscape. When, however, was there ‘discontinuity’? Three
northwest to southeast aligned late Roman ditches were recorded in one area
during an extensive evaluation, and running parallel with these was a ditch whose
fill contained two sherds of hand-made early to middle Saxon pottery (c. A.D. 450-
850), six sherds of late ninth- to early thirteenth-century ‘Cotswold-type’ ware,
and three residual Roman pieces. Significantly, a fifth ditch, on a different axis to
figure 6 Schematic representation of how excavated Romano-British field boundaries
share a common orientation, share a specific alignment, or show no relationship with the
later historic landscape. ßCrown Copyright and Landmark Information Group Limited
(2012). All rights reserved (1890).
44 STEPHEN RIPPON et al.
figure 7 Three major historic landscape types. ßCrown Copyright and Landmark
Information Group Limited (2012). All rights reserved (1869 (top), 1906 (middle and bottom)).
THE FIELDS OF BRITANNIA 45
these but oriented with ridge and furrow and the historic boundaries, contained
residual Roman pottery with a sherd of ‘Malvernian’ ware current between the
late thirteenth and seventeenth centuries (Meara 2008). The evidence from this
sequence suggests that the axis of the Roman landscape continued until as late as
the ninth century, after which the landscape was reorganised as the open fields
were set out. The negative relationship between Roman fields and later landscape
features is not normal for this pays, where 79 percent of nineteen Romano-British
sites within countryside characterised by enclosed common field strips share a
common orientation or alignment with it, which compares with 61 percent for the
Central Zone region as a whole.
What was happening at an individual site may not have been typical of the
surrounding area, and so if a particular pays has produced just one or two sites
with useful data, we cannot use that data to generalise about that pays as a whole.
Such individual sites do become valuable, however, when their data is aggregated
across the region as a whole in order to produce a larger sample size. As the Fields
of Britannia Project is ongoing at the time of writing, it is not possible to give any
overall statistics for Roman Britain as a whole, but for those regions for which
data collection is complete (subject to additional sites being published) there are
clear differences across each region. To begin with, if we look at the relationship
between excavated Romano-British field systems and later landscapes depicted on
the OS 1st Ed maps whose character is suggestive of former medieval open fields or
closes held in severalty, then in East Anglia 69 percent share a specific alignment or
figure 8 Saxon’s Lode
Farm, near Ripple,
Worcestershire: a mid-
Roman agricultural
settlement and field system
whose location, as well as
orientation and alignment,
was adopted in what the
excavators refer to as the
‘early and middle Saxon’
periods, and remained to
influence the shape of the
later historic field pattern
(after Barber and Watts
2008). ßCrown Copyright
and Landmark Information
Group Limited (2012). All
rights reserved (1887).
46 STEPHEN RIPPON et al.
general orientation, with a similar figure (66 percent) in the South East, falling to
58 percent in the Central Zone, 57 percent in the Western Lowlands, and just 29
percent in the South West. Before we move on to consider the changing patterns of
settlement, which inevitably includes discussion of ‘native’ versus ‘Anglo-Saxon’,
and how these two communities co-existed, it is worth highlighting the fact that of
the areas with the greatest degree of continuity in both land use and land division,
some (i.e. East Anglia, and parts of Kent and the East Midlands) were subject to
the strongest early Germanic influence, while others were not (for example, parts
of Gloucestershire and Somerset). The conclusion can be drawn, therefore, that
long-term continuity on a landscape-scale appears to have little to do with changes
in the politics, ethnic composition, or material culture of early medieval England.
The evolution of settlement patterns
Across Roman Britain there are examples of extensively excavated settlements that
appear to have seen continued occupation by the native British population into the
early medieval period, such as Frocester in Gloucestershire (Price 2000). There are
equally well-known sites that appear to have seen the replacement of the native
community by Germanic immigrants, such as at Barton Court Farm in Oxford-
shire (Miles 1986), and Rivenhall in Essex (Rodwell and Rodwell 1986), or which
appear to have been abandoned at the end of the Roman period, such as the villas
at Chignal St James in Essex (Clarke 1998) and Gadebridge Park in Hertfordshire
(Neal 1974). There are also examples of landscape surveys – most notably based
upon fieldwalking – that have indicated how certain parts of the countryside were
abandoned in the early medieval period (e.g.in the heavy clay soils of the
interfluvial plateaux in Suffolk: Newman 1992; West and McLaughlin 1998;
Laverton 2001). The Fields of Britannia Project has therefore also examined the
development of settlement patterns across adjacent and contrasting pays in three
county-based case studies: Norfolk, Kent, and Somerset. Norfolk proved the
easiest to study due the large number of fieldwalking surveys and the continuous
ceramic sequence, with identifiable late Roman, early Saxon, middle Saxon, and
late Saxon wares. In Kent, there was a particularly large number of excavations,
but little fieldwalking, which also provided a good sized database, although in
Somerset the scarcity of early medieval ceramics meant that the most that could be
achieved was a comparison between the locations of late Roman settlements,
medieval parish churches, and places referred to in Domesday Book.
The results from Norfolk provide the clearest evidence for how settlement
patterns developed during this period. The two pays with the greatest contrast in
terms of their soils are the Boulder Clay, which covers most of central Norfolk,
and the sandy Breckland to the south-west. There are two ways of approaching the
analysis: firstly, to examine what became of the Romano-British landscape (e.g.
how many Romano-British settlements are associated with early, middle or late
Saxon sites), and secondly, how far the well-documented and archaeologically
attested landscape of the eleventh century owes its origins to the Roman period.
The contrasts between these pays are marked. In a sample of parishes on the
Boulder Clay – that have seen well-recorded fieldwalking – 31 percent of the 62
Romano-British settlements have fifth- to seventh-century occupation within
THE FIELDS OF BRITANNIA 47
500m, compared to 52 percent of the 35 Romano-British settlements in Breckland
(Figures 9 and 10). This pattern of greater continuity in the Breckland is also seen
from the perspective of the eleventh-century landscape: on the Boulder Clay, 55
percent of the eleven parish churches have Romano-British settlement within
500m, compared to 73 percent of the fifteen churches in Breckland. Within these
pays, however, there are important differences in terms of where there was greater
and lesser apparent continuity of settlement. The Boulder Clay pays, for example,
actually includes three distinct topography/soil types: the damper valley bottoms,
the well-drained valley sides, and the heavy soils of the inter-fluvial areas. In the
Roman period, settlement focussed on the well-drained soils of the valley sides,
although it had also pushed up into the inter-fluvial areas. In the early Saxon
period the contraction of settlement was greatest in these inter-fluvial areas. Their
abandonment was brief, however, as by the time that middle Saxon Ipswich Ware
was in use, settlement had once again expanded into these areas, although this
expansion was most pronounced during the tenth to eleventh centuries. In the valleys,
there was more continuity in the areas that were occupied, although this is greatest for
non-villa settlements. In Breckland, the valleys were once again the favoured
settlement locations in the Roman period, although it was no doubt because the inter-
fluvial areas were so dry (in contrast to the Boulder Clay pays where the interfluves
suffered from waterlogging) and it was these same areas that saw the greatest
continuity of settlement into the early medieval period (Figures 9–10).
Conclusions
The Fields of Britannia Project is trying to shed new light on landscape change at a
critical point in British history, the transition between the Roman and medieval
periods. Through the use of different scales of analysis – such as regions and pays –it
has recognised that the landscapes of Roman and medieval Britain were far from
uniform in their character. This is borne out by the very marked differences in the
potential continuity between these two periods in different regions and pays, both in
terms of the palaeoenvironmental evidence for land use and the archaeological
evidence for the potential survival of field systems. Extensive use has been made of
the growing body of unpublished ‘grey literature’, which is now a major source of
research data. The traditional approach to studying this transition period is through
the excavated sequences on a relatively small number of individual sites that may, in
fact, not be typical of the local area or wider region. In contrast, the Fields of
Britannia Project has looked at detailed local site biographies within the broader
context of aggregated data across both relatively small pays and far wider regions,
which allows us to establish the normal pattern of landscape evolution. The
palaeoenvironmental evidence shows very clearly that in lowland areas there was no
widespread woodland regeneration in the post-Roman period, and although some
pays did see a slight increase in tree pollen in the fifth century, at least some of this
could be accounted for by a failure to maintain hedgerows and coppiced woodland.
The excavation evidence shows the previously unexpected degree to which Romano-
British field systems are on the same general orientation, and sometimes even share
common alignments, as their medieval successors, and so still affect the character of
48 STEPHEN RIPPON et al.
today’s landscape. This article is, however, just an interim report on the project, the
final results of which will be published by Oxford University Press in 2014 in a book
titled The Fields of Britannia: Regional Landscapes in Transition A.D. 400-1000.
figure 9 The Breckland pays in Norfolk, showing that a broad continuity can be seen (in
places which have seen extensive fieldwalking, such as Barton Bendish, Illington, and
Hargham), between the areas that were settled in the Roman and in the early medieval
periods. (Based on data obtained from the Norfolk Historic Environment Record, and
Mackney et al. 1983 for soils).
THE FIELDS OF BRITANNIA 49
Acknowledgements
The authors wish to thank the Leverhulme Trust for funding this research (award
F/00 144/BI), and all of the Historic Environment Records, archaeological units,
and individuals who have supplied data.
figure 10 Fieldwalking around Ingoldsthorpe and Snettisham shows that the majority of Roman-
British settlements have early Saxon occupation in the vicinity, but that in the middle Saxon period
this was no longer the case. (Based on data obtained from the Norfolk Historic Environment Record).
ßCrown Copyright and Landmark Information Group Limited (2012). All rights reserved (1890).
50 STEPHEN RIPPON et al.
References
Aston, M. and Gerrard, C. 1999. ‘Unique, traditional and charming: The Shapwick Project, Somerset’, in The
Antiquaries Journal 79, 1–58.
Barber, A. and Watts, M. 2008. ‘Excavations at Saxon’s Lode Farm, Ripple, 2001–2: Iron Age, Romano-
British and Anglo-Saxon rural settlement in the Severn Valley’, in Transactions of the Worcestershire
Archaeological Society 21, 1–90.
Barker, P., White, R., Pretty, K., Bird, H., and Corbishley, M. 1997. The Baths Basilica Wroxeter: Excavations
1966-90, English Heritage Archaeological Report 8.
Biddulph, E., Brady, K., Ford, B.M., and Murray, P. 2010. ‘Roman settlement, pottery production, and a
cemetery in the Beam valley, Dagenham’, in Transactions of the Essex Society for Archaeology and History
1, 109–65.
Biddulph, E. 2011. ‘Roman settlement of the Ebbsfleet valley in a regional perspective’, in Settling the
Ebbsfleet Valley: High Speed I Excavations at Springhead and Northfleet, Kent. The Late Iron Age,
Roman, Saxon, and Medieval Landscapes, ed. P. Andrews, E. Biddulph, A. Hardy and R. Brown, Oxford
Wessex Archaeology, 243–8.
Bryant, S., Perry, B. and Williamson, T. 2005. ‘A ‘‘relict’’ landscape in South-East Hertfordshire’, in Landscape
History 27, 5–16.
Clarke, C.P. 1998. Excavations to the South of Chignall Roman Villa, Essex, 1977–81, East Anglian
Archaeology 83, Essex County Council Archaeology Section and Scole Archaeological Committee.
Collingwood, R.G. 1930. The Archaeology of Roman Britain, Methuen.
Dark, K. and Dark, P. 1997. The Landscape of Roman Britain, Sutton Publishing.
Dark, K. 1999. Civitas to Kingdom: British Political Continuity 300–800, Leicester University Press.
Dark, K. 2000. Britain and the End of the Roman Empire, Tempus.
Dark, P. 2000. The Environment of Britain in the First Millennium A.D., Duckworth.
Evans, D. and Hancocks, A. 2005. ‘Romano-British, late Saxon and medieval remains at the Old Showground,
Cheddar: excavations in 2001’, in Proceedings of the Somerset Archaeological and Natural History Society
149, 107–22.
Fox, C. 1932. The Personality of Britain: Its Influence on Inhabitant and Invader in Prehistoric and Early
Historic Times, National Museum of Wales.
Frere, S. 1967. Britannia: A History of Roman Britain, Kegan and Paul.
Gerrard, C. with Aston, M. 2007. The Shapwick Project, Somerset: A Rural Landscape Explored, Society for
Medieval Archaeology Monograph 25, Society for Medieval Archaeology.
Gray, H.L. 1915. English Field Systems, Harvard.
Ha
¨rke, H. 2002. ‘Kings and warriors’, in The Peopling of Britain, ed. P. Slack and R. Ward, Oxford University
Press, 145–75.
Harmer, R., Peterken, G., Kerr, G. and Poulton, P. 2001. ‘Vegetation changes during 100 years of development
of two secondary woodlands on abandoned arable land’, in Biological Conservation 101, 291–304.
Haverfield, F. 1912. The Romanization of Britain, Clarendon Press.
Hicks, S. 2006. ‘When no pollen does not mean no trees’, in Vegetation History and Archaeobotany 15, 253–
61.
Higham, N. 1992. Rome, Britain and the Anglo-Saxons, Seaby, London.
Hills, C. 2009. ‘Anglo-Saxon DNA’, in Mortuary Practices and Social Identities in the Middle Ages: Essays in
Burial Archaeology in Honour of Heinrich Ha¨rke, ed. D. Sayer and H. Williams, University of Exeter Press,
123–40.
Hingley, R. 1989. Rural Settlement in Roman Britain, Seaby.
Hinton, D. 1997. ‘The ‘‘Scole-Dickleburgh field system’’ examined’, in Landscape History 19, 5–13.
Howell, I., Swift, D. and Watson, B. 2011. Archaeological Landscapes of East London: Six Multi-period Sites
Excavated in Advance of Gravel Quarrying in the London Borough of Havering, Museum of London
Archaeology Monograph 54, Museum of London Archaeology Service.
Jones, B. and Mattingly, D. 1990. An Atlas of Roman Britain, Blackwells.
THE FIELDS OF BRITANNIA 51
Laverton, S. 2001. Shotley Peninsula: The Making of a Unique Suffolk Landscape, Tempus.
Leach, P. 1994. Ilchester Volume 2: Archaeology, Excavations and Fieldwork to 1984, Sheffield Academic
Press.
Leech, R. H. 1975. ‘An Iron Age and Romano-British settlement at Podimore, Somerset’, in Proceedings of the
Somerset Archaeology and Natural History Society 119, 72–6.
Lewis, C., Mitchell-Fox, P. and Dyer, C. 1997. Village, Hamlet and Field, Manchester University Press.
Lovell, J. 2006. ‘Excavation of a Roman-British farmstead at RNAS Yeovilton’, in Proceedings of the Somerset
Archaeology and Natural History Society 149, 7–67.
Mackney, D., Hodgson, J.M., Hollis, J. M., and Staines, S. J. 1983. Legend for the 1:250,000 Soil Map of
England and Wales, Soil Survey of England and Wales.
Mattingly, D. 2006. An Imperial Possession: Britain in the Roman Empire, Penguin.
Meara, H. 2008. ‘Brizen Farm, Shurdington, Gloucestershire: archaeological evaluation report’, Oxford
Archaeology Report 4118.
Medlycott, M., Weller, S., and Benians, P. 2010. ‘Roman Billericay: excavations by the Billericay
Archaeological and Historical Society 1970–77, in Transactions of the Essex Society for Archaeology
and History 1, 51–108.
Miles, D. 1986. Archaeology at Barton Court Farm, Abingdon, Oxon, Council for British Archaeology
Research Report 50, Oxford Archaeological Unit and the Council for British Archaeology.
Millett, M. 1990. The Romanization of Britain, Cambridge University Press.
Neal, D.S. 1974. The Excavation of the Roman Villa in Gadebridge Park Hemel Hempstead 1963-8, Society
of Antiquaries of London Research Report XXXI.
Newman, J. 1992. ‘The Late Roman and Anglo-Saxon Settlement Pattern in the Sandlings of Suffolk’, in The
Age of Sutton Hoo, ed. M. Carver, Boydell and Brewer, 25–51.
Percival, S. and Williamson, T. 2005. ‘Early fields and medieval furlongs: excavations at Creake Road’, in
Landscapes 6 (i), 1–17.
Pitts, M. 2008. ‘Globalizing the local in Roman Britain: an anthropological approach to social change’, in
Journal of Anthropological Archaeology 27, 493–506.
Price, E. 2000. Frocester: A Romano-British Settlement, its Antecedents and Successor, Gloucester and District
Archaeological Research Group.
Pryor, F. 2004. Britain AD, Harper Collins.
Rippon, S. 2008. Beyond the Medieval Village, Oxford University Press.
Rippon, S. 2012. Making Sense of an Historic Landscape, Oxford University Press.
Rippon, S., Smart, C. and Pears, B. (forthcoming). The Fields of Britannia: Regional Landscapes in Transition
AD 400–1000, Oxford University Press.
Roberts, B. and Wrathmell, S. 2000. An Atlas of Rural Settlement in England, English Heritage.
Roberts, B. and Wrathmell, S. 2002. Region and Place, English Heritage.
Rodwell, W. 1978. ‘Relict landscapes in Essex’ in Early Land Allotment, ed. H.C. Bowen and P.J. Fowler,
British Archaeological Reports, British Series 48, 89–98.
Rodwell. W.J. and Rodwell. K.A. 1986. Rivenhall: Investigation of a Villa, Church and Village, 1950–77,
Council for British Archaeology Research Report 55, Council for British Archaeology.
Sargent, A. 2002. ‘The North-South divide revisited: thoughts on the character of Roman Britain’, in Britannia
33, 219–26.
Taylor, C.C. and Fowler, P.J. 1978. ‘Roman fields into medieval furlongs?’, in Early Land Allotment, ed. H.C.
Bowen and P.J. Fowler, British Archaeological Reports, British Series 48, 159–62.
Taylor, J. 2007. An Atlas of Roman Rural Settlement in England, Council for British Archaeology Research
Report 151, Council for British Archaeology.
Thirsk, J. 1967. ‘The farming regions of England’, in The Agrarian History of England and Wales, vol. 4:
1500–1640, ed. J. Thirsk, Cambridge University Press, 1–112.
Thirsk, J. 1984. ‘Introduction’, in The Agrarian History of England and Wales, vol.5: 1640–1750, I: Regional
Farming Systems, ed. J. Thirsk (ed.), Cambridge University Press, xix-xxxi.
Thirsk, J. 1987. England’s Agricultural Regions and Agrarian History, 1500–1750, Palgrave Macmillan.
52 STEPHEN RIPPON et al.
Upex, S. 2002. ‘Landscape continuity and the fossilization of ancient fields’, in Archaeological Journal 159,
77–108.
West, S.E. and McLaughlin, A. 1998. Towards a Landscape History of Walsham le Willows, Suffolk, East
Anglian Archaeology 85, Suffolk County Council Archaeology Service and Scole Archaeological
Committee.
Williamson, T. 1987. ‘Early coaxial field systems on the East Anglian Boulder Clays’, in Proc. of the
Prehistoric Soc. 53, 419–31.
Williamson, T. 1998. ‘The ‘‘Scole-Dickleburgh field system’’ revisited’, in Landscape History 20, 19–28.
Williamson, T. 2003. Shaping Medieval Landscapes, Windgather Press.
Wymer, J. and Brown, N. 1995. North Shoebury: Settlement and Economy in South-East Essex 1500 BC-AD
1500, East Anglian Archaeology 75, Essex County Council Archaeology Section and Scole Archaeological
Committee.
Notes on contributors
Stephen Rippon is Professor of Landscape Archaeology, and Dean of the Faculty of
Graduate Research, at the University of Exeter. His research interests include the
origins and development of regional variation in landscape character, the Roman
to medieval transition from a landscape perspective, and the exploitation,
modification and transformation of coastal wetlands. His most recent books are
Making Sense of an Historic Landscape (OUP, 2012) and the Second Edition of
Historic Landscape Analysis (CBA Handbook 16). Contact: s.j.rippon@ex.ac.uk
Chris Smart is a landscape archaeologist specialising in the Roman and medieval
periods in Britain, and was Associate Research Fellow on the Fields of Britannia
project. His doctoral thesis (2008: Exeter) examined Roman to medieval
transitions in western Britain, particularly in the rural landscape. Chris has
research interests in subjects as diverse as the impact of the Roman army in Britain
(directing ongoing excavations at Calstock Roman fort, in Cornwall) and the role
of rural industries in shaping historic landscapes. Contact: c.j.smart@ex.ac.uk
Ben Pears is a geo-archaeologist with research interests in palaeoenvironmental
archaeology, agrarian landscape history, and settlement development of north-
west Europe during the Roman and medieval periods. He was an Associate
Research Fellow on the Fields of Britannia Project. His recent work has focussed
on the development of anthropogenic soils on rural sites in England, Scotland,
Ireland and the Netherlands. Contact: benrpears@gmail.com
Fiona Fleming is a postgraduate student at the University of Exeter, where her
doctoral research looks at the relationship between late Roman and early medieval
settlement patterns. Her research interests are rooted in the archaeology of
landscapes and the social and cultural influences that have shaped regional
character across these two periods. Contact: fjf201@ex.ac.uk
THE FIELDS OF BRITANNIA 53
