The Physical Evolution of The North Avon Levels

The Physical Evolution
of the North Avon Levels

a Review and Summary of the
Archaeological Implications
By Michael J. Allen and Robert G. Scaife

The Physical Evolution of the North Avon
Levels: a Review and Summary of the
by Michael J. Allen and Robert G. Scaife
with contributions from J.R.L. Allen, Nigel G. Cameron, Alan J. Clapham,

Rowena Gale, and Mark Robinson
with an introduction by Julie Gardiner
Wessex Archaeology Internet Reports

Published 2010 by Wessex Archaeology Ltd
Portway House, Old Sarum Park, Salisbury, SP4 6EB
http://www.wessexarch.co.uk/
Copyright © Wessex Archaeology Ltd 2010 all rights reserved
Wessex Archaeology Limited is a Registered Charity No. 287786

Contents
List of Figures
List of Plates
List of Tables
Editor’s Introduction, by Julie Gardiner .......................................................................................... 1
INTRODUCTION
The Severn Levels ............................................................................................................................ 5
The Wentlooge Formation ............................................................................................................... 5
The Avon Levels .............................................................................................................................. 6
Background ...................................................................................................................................... 7
THE INVESTIGATIONS
The research/fieldwork: methods of investigation ........................................................................... 10
The Sites ........................................................................................................................................... 11
1. Awkley Lane ................................................................................................................................ 11
The sequence ........................................................................................................................ 11
Chronology .......................................................................................................................... 12
Sampling the sequence ......................................................................................................... 16
Textural features of the minerogenic sequence, by J.R.L. Allen .......................................... 18
Pollen analysis, by Robert G. Scaife ..................................................................................... 18
Diatom analysis, by Nigel G. Cameron ................................................................................ 30
Molluscs, by Michael J. Allen .............................................................................................. 32
Insect remains, by Mark Robinson ....................................................................................... 32
Waterlogged plant macrofossils, by Alan J. Clapham ......................................................... 34
2. Vimpenny’s Lane ......................................................................................................................... 37
The sequence ........................................................................................................................ 37
Chronology .......................................................................................................................... 40
Textural features of the minerogenic sequence, by J.R.L. Allen .......................................... 40
Pollen analysis, by Robert G. Scaife .................................................................................... 42
Diatom analysis, by Nigel G. Cameron ............................................................................... 49
Insect remains, by Mark Robinson ....................................................................................... 52
Waterlogged plant macrofossils, by Alan J. Clapham ......................................................... 53
3. Hallen Marsh ................................................................................................................................ 54
The sequence ........................................................................................................................ 54
Textural features, by J.R.L. Allen ......................................................................................... 60
Pollen analysis, by Robert G. Scaife ..................................................................................... 60
Diatom analysis, by Nigel G. Cameron ............................................................................... 67
Charcoals, by Rowena Gale ................................................................................................. 68
4. Northwick .................................................................................................................................... 69
The sequence ........................................................................................................................ 69
5. Awkley Interface .......................................................................................................................... 74
The sequence ........................................................................................................................ 74
3
Comment .............................................................................................................................. 78
DISCUSSION AND REVIEW. The Physical and Vegetational Evolution of the Avon Levels
Stratigraphic Overview .................................................................................................................... 79
The Wentlooge Formation .................................................................................................... 79
Sea-level change and palaeo-geographic reconstruction ...................................................... 80
Patterns of consistency: the mapped sediments ................................................................... 82
Environmental change in the Avon Levels ...................................................................................... 85
Pollen taphonomy and interpretation ................................................................................... 85
The late prehistoric vegetation of the Avon Levels ......................................................................... 87
The regional pollen database ............................................................................................... 87
Avon Levels sites ................................................................................................................. 88
The changing wetland/floodplain/estuarine habitat ............................................................. 94
Summary .............................................................................................................................. 96
The mineral, estuarine and freshwater sediments ................................................................ 97
Archaeology and environment: marsh landscape and archaeology……………………………….. 98
Some archaeological implications ..................................................................................... ..98
History of human exploitation patterns and behaviour ...................................................... 99
Why does settlement occur? .............................................................................................. 103
What caused Iron Age abandonment? ............................................................................... 103
A change of use .................................................................................................................. 105
Land claim .......................................................................................................................... 105
Archaeological significance ............................................................................................... 106
Conclusions .................................................................................................................................... 106
Acknowledgements ........................................................................................................................ 106
Appendix: analytical techniques .................................................................................................... 108
Bibliography .................................................................................................................................. 111
4
List of Figures
Figure 1. The Inner Bristol Channel and Severn Estuary showing the distribution of post-glacial
alluvium comprising the Severn Estuary Levels
Figure 2. Sedimentary features comprising the Wentlooge Formation
Figure 3. The geology of the Avon Levels showing the location of the archaeological and
sedimentological investigations
Figure 4. Awkley Lane Wentlooge Formation profiles
Figure 5. Radiocarbon determinations in relation to their OD and the middle Wentlooge wood
and fen peat
Figure 6. Awkley Lane; particle size triangle
Figure 7. Awkley Lane; pollen diagram
Figure 8. Diatom analysis; all sites
Figure 9. Vimpenny’s Lane Wentlooge Formation profiles
Figure 10. Vimpenny’s Lane; particle size triangle
Figure 11. Vimpenny’s Lane pollen diagram
Figure 12. Hallen Marsh: Summary plan of the excavation
Figure 13. Hallen Marsh; full site profiles
Figure 14. Hallen Marsh Wentlooge Formation profiles
Figure 15. Hallen Marsh; particle size triangle
Figure 16. Hallen pollen diagram
Figure 17. Northwick; summary plan of the excavations
Figure 18. Northwick Wentlooge Formation profiles
Figure 19. Awkley Interface composite transect
Figure 20. Relative heights of soils and stasis horizons from various locations
Figure 21. a. Sea-level curve using the existing framework of Haslett et al. (1997) from a
synthesis of data obtained by Kidson and Heyworth; 20b. Seaa-evel curve using the
the corrected RSL graph for Caldicot Pill on the north side of the Severn Crossing
Figure 22. Correlation of the sedimentary facies from the reported sites
Figure 23. Location of archaeological sites in the Avon Levels
Figure 24. Suggested model for the exploitation of the Avon Levels, Neolithic to late Roman
period
List of Plates
Plate 1. Sampling the minerogenic sediment sequence in monoliths at Awkley Lane

Plate 2. Awkley Lane excavation showing the fully excavated and exposed upper and middle
Wentlooge sequence
Plate 3. Vimpenny’s Lane excavation showing the fully excavated and exposed upper and
middle Wentlooge sequence
Plate 4. Round-house 1 at Hallen after excavation
5
List of Tables
Table 1. Radiocarbon determinations and calibrated results from the middle Wentlooge
Formation at Awkley Lane
Table 2. Mollusc data from Awkley Lane
Table 3. Coleoptera and other insects from Awkley Lane
Table 4. The waterlogged plant remains from the wood and fen peat of the middle Wentlooge
Table 5. Radiocarbon determinations and calibrated results from the detrital peat at
Vimpenny’s Lane
Table 6. Mollusc data from Vimpenny’s Lane
Table 7. The waterlogged plant remains from the detrital peat (context 211) from Vimpennys
Lane
Table 8. Mollusc data from Hallen Marsh
Table 9. Charcoal from the round-house at Hallen Marsh
Table 10. Mollusc data from Northwick
Table 11. Mollusc data from Awkley Interface, trench 4
Table 12. Stasis horizons in the middle Wentlooge sequence in relation to their OD height and
radiocarbon determination
6
The Physical Evolution of the North Avon
Levels: a Review and Summary of the
by Michael J. Allen and Robert G. Scaife
with contributions from J.R.L. Allen, Nigel G. Cameron, Alan J. Clapham,

Rowena Gale, and Mark Robinson
with an introduction by Julie Gardiner
Editor’s introduction
by Julie Gardiner
The construction of the new motorway links for the Second Severn Crossing and the resultant infill
and developments which would come in their wake provided the opportunity for a comprehensive
programme of archaeological desk-based research, auger survey, building recording, excavation,
and palaeo-environmental work (The Second Severn Crossing English Approaches project). The
defined area of study covered approximately 54 km², bounded respectively to the north and east by
the M4 and M5 motorways, to the south by Bristol, and to the west by the Severn Estuary. This
work was completed in 1994.
An assessment report was prepared (Wessex Archaeology 1994), in which a Revised Project Design
proposed a series of research aims and themes and the preparation and publication of a monograph
text comprising a series of chronological chapters discussing the development of the Avon Levels
landscape. There followed a lengthy period during and after the construction of the motorway links
and new bridge during which the post-excavation programme was not commissioned. This
happened in 2000, by which time considerable further archaeological work had taken place on both
sides of the Severn Estuary and many of the most important results published (eg, Rippon 1996;
Locock 1997; Locock et al. 1998; Bell et al. 2000). These more recent archaeological works and
publications allowed for a more focused and critical appraisal of the research themes defined in
1994. The relative importance and information potential of some aspects of the English Approaches
project could be seen to have been enhanced, and others diminished, in the light of new information
and ongoing research.
The themes deemed to remain of greatest importance and potential were:

x The characterisation and nature of later Bronze Age to Romano-British settlement and
settlement patterns.
x The characterisation and chronological development of the sedimentological sequence
during the Holocene.
x The reconstruction of the evolution of the palaeo-environment.
New proposals for publication were drawn up (Wessex Archaeology 2000), consisting of a series of
papers, of which the two main ones, destined for appropriate academic journals, would deal
specifically with the first of these themes on the one hand, and with the latter two on the other. The
‘archaeological’ paper was completed in 2001 and published in the Proceedings of the Prehistoric
Society (Gardiner et al 2002). A suitable publication vehicle for the second, much longer, paper was
identified and favourable discussions with the Editors ensued before writing commenced. The draft
1
of the second paper was also completed in 2001. Unfortunately, in the meantime, the journal
underwent an editorial review and altered its remit to the extent that the paper was no longer
deemed suitable for inclusion.
The length and complexity of the paper was such that it was difficult to identify another journal that
could take it in its entirety, yet it was not an attractive proposition for a monograph. Various
protracted negotiations were entered into regarding suitable formats and re-arrangements of the
information, including a multi-media approach to publication. In the end it was decided that the
strength of the paper was in the integration of the detailed analyses with wider thematic discussions
and that its integrity would be unnecessarily compromised by the various proposals to
‘disaggregate’ it that had been considered. The decision has therefore been taken to publish it here,
in full, online.
The text that follows is that which was drafted in 2001. It had already been subject to external
academic refereeing pending submission to the journal and the referees’ comments, and those of its
major contributors, have been largely incorporated. Financial contraints have meant that no new or
additional work has been undertaken and no account has been taken of any archaeological (or
sedimentary) work that has been undertaken since completion of the draft, or of any more recently
obtained sea-level index points. We are confident that none of this will have substantially altered
the arguments and conclusions presented here though some refinements at the detailed level have
certainly been achieved. The only updating that has been undertaken has been in terms of
bibliographic references (several cited ‘in prep’ and ‘forthcoming’ papers have been published in
the interim) and the radiocarbon dates obtained as part of the project have been recalibrated using
Oxcal 4.1.
2
INTRODUCTION
This paper reviews the results of research into the sedimentary sequences examined as a result of
mitigation for the approach roads of the Second Severn Crossing. Although most of the work lay in
strictly archaeological investigations, the brief was broad enough to allow excavation and detailed
examination of long sedimentary sequences with peat horizons that occurred along the road
corridor. This paper provides the first opportunity to present the detailed results of the palaeo-
environmental analysis of the later Holocene sedimentary sequence of the region. These
interpretations are of value in understanding the physical development and evolution of the Avon
Levels and the wider Severn Estuary Levels. They also provide the foundation for wider discussion
of the social development and patterns of human exploitation of the area, particular in later
prehistory. These archaeological elements are outlined here (see discussion) and dealt with in more
detail elsewhere (Gardiner et al. 2002).
There is a wealth of literature concerning the evolution, salt-marsh accretion, geomorphology and
nature of the Severn Estuary. In particular this has concentrated on the sedimentary history of
coastal sites, especially around the inner Severn Estuary, eg, Arlingham (Allen 1990a), Longney
(Allen & Fulford 1990a), Elmore (Allen & Fulford 1990b), Slimbridge Warth (Allen 1986), and the
Welsh Severn Estuary, eg, Rumney Wharf (Allen 1987a; 1996) (see Fig. 1). More recently, the
archaeological significance of these areas has been demonstrated, from early Holocene times
(Mesolithic) to the Iron Age in the Gwent Levels, Wales, at Goldcliff (Bell et al. 2000), and the
Roman to medieval history in both the Welsh and Inner Severn Estuary (Allen and Fulford 1986;
1988; Allen & Rae 1987; Rippon 1991; 1994; 1996; 1997). Most of this work has its origin in the
palaeo-environmental and chronological studies in the Somerset Levels initiated by Godwin (1943).
3
INSET
River Severn Gloucester
Elmore
Arlingham Longney
Awre
Slimbridge INNER
s
e vel ESTUARY
ent L Hill
Gw
Coldicot Level
Oldbury
Wentlooge Level
Goldcliff Redwick
Rumney
Avon Levels
Avonmouth Fig.3
Cardiff MIDDLE
Bristol
ESTUARY
North
Somerset Level
Inner Bristol Channel Brean OUTER

ESTUARY
Somerset
Levels
Porlock
Zones of peat exposed intertidally

Estuarine alluvium
0 10 km
Figure 1. The Inner Bristol Channel and Severn Estuary showing the distribution of post-glacial
alluvium comprising the Severn Estuary Levels
4
The Severn Levels
The Severn Estuary and Inner Bristol Channel lie along the Severn Carboniferous Fault Zone
(Wilson et al. 1988), and bedrock surface forms a broad flat-floored ‘outer valley’, which contains
estuarine sediments through which the Severn and its tributaries cut narrow gorges during the late
Devensian (and probably earlier) low-stand sea(s) (Allen 1990b). The soft Triassic to early Jurassic
mudrock surface is scarred by numerous periglacial features generally ascribed to the Late
Devensian, Dimlington Stadial (Andrews et al. 1984; Allen 1987b).
The catchment of the Severn is c. 11,400 km2 and includes much of mid-Wales and the West
Midlands, and the Severn Estuary. The river valley is overlooked by high ground of Carboniferous
rocks (Mendips and Bristol), mid Jurassic limestones (Cotswolds) and Old Red Sandstones and
Carboniferous strata (Forest of Dean). These bound large areas of low-lying wetlands which occur
on both sides of the estuary.
These wetlands comprise a total of c. 840 km2 of largely disconnected and generally intricate
outcrops of post-glacial alluvium (c. 8 km3) declining from the large inland Somerset Levels to the
very small area of the inner estuary (eg, Arlingham). The main wetlands are the Somerset Levels,
forming the largest inland wetland fringing the Inner Bristol Channel. Along the coast of the outer
estuary are the Gwent Levels, comprising the Wentlooge and Caldicot Levels on the Welsh coast,
and the North Somerset Levels and the smaller Avon Levels marking the northern limit on the east
coast. The Middle estuary comprises much narrower and less extensive wetlands including those
supported by the intertidal rock platforms of the Oldbury Flats and Hills Flats, while the Inner
estuary, extending 15–20 km down stream to Gloucester (ie Slimbridge to Gloucester), includes
important but smaller wetlands such as those studied around Awre, Arlingham, and Longney (Fig.
1).
The Wentlooge Formation
These wetlands are supported by a post-glacial sediment sequence that is, on average, 10–15 m
thick, largely comprising soft estuarine mineral sediments and peats and which arguably contains
some of the most extensive and complete exposures of post-glacial shallow-marine sediments in the
British Isles (Allen 1990c). In the Severn Estuary they are described as the Wentlooge Formation
(Allen 1987a; Allen & Rae 1987; Allen 1990b) and are essentially sediments of high tidal mudflats
and salt-marsh environments with peats overlain with recent estuarine alluvium. This sequence is
remarkably uniform in general character throughout the Avon, Gwent, and Somerset Levels and has
been informally divided into three principal component elements (see Fig. 2):
lower Wentlooge - thin gravels and sands, organic-rich silty sands and ‘peats’ grading into
bluish grey-greenish grey sand to clayey silts. The ‘sands’ (some contain
Ipswichian shells) probably relate to cold climate estuarine sand deposition
(Andrews et al. 1984)
middle Wentlooge - occur at about -1 to +3–4m OD comprise alternation of organic (peat) beds
with often laminated greenish grey sandy to clayey silts (c. 5500–250 cal BC
(6500–2200 BP), with the organic facies representing development from
wetland carr woodland with alder and oak to reed swamp
upper Wentlooge - thick greenish grey sandy to clayey silts which grade into green-brown
mottled silts which are unlaminated and poorly stratified
5
1 Note: Throughout this paper post-glacial refers to post-Devensian glacial and, similarly, ‘alluvium’ is used as
shorthand for estuarine alluvial or intertidal sediments.
The construction of drainage systems and sea defences over the last two millennia has halted the
development of the Wentlooge Formation. On the seaward side of the sea defences three other
estuarine formations have developed: The Rumney, Awre, and Northwick formations (Allen & Rae
1987; Allen 1987a; Allen 1988), which are largely medieval to post-medieval in date.
These post-glacial sediments represent fluvial and marine environments which accumulated during
major marine transgression. Human activity increasingly influenced the character of these
‘marginal’ wetlands being created by a ‘lengthy interplay between human and natural forces’ (Allen
1990c). As such ‘Much can be learned from the sedimentary and archaeological record about the
way in which coasts and estuaries have varied in the past, the way in which man has responded to
or caused such changes and with what consequences for both himself and for the ‘natural’
environment’ (Pye & Allen 2000, 1). The important and significant evidence of prehistoric
communities from the Somerset Level (Coles & Coles 1986) has, with little doubt, distracted
investigators away from the potential and significance of the wetlands as a whole (Allen 1990c, 33).
The Levels form a landscape which superficially appears flat and low-lying (generally less than 7.5
m OD) and uniform, but which may be divided into a number of character areas (cf. Rippon 1994).
The-low-lying nature, fine-grained soils, flatness, and susceptibility to flooding provide wetland
sites of ecological importance. Since 1987 a number of Site of Special Scientific Interest have been
registered in the Levels, and a series of important archaeological sites have been discovered with
well-preserved palaeo-environment material.
The Avon Levels
The opportunity to examine the evolution of the Avon Levels (Fig. 3) was presented as a result of
archaeological work conducted prior to the construction of the new road bridge across the Severn
Estuary (the Second Severn Crossing) and the approach roads from the south (M5) and east (M4).
Initial evaluation work isolated ten archaeological or palaeo-environmental sites which were
excavated in 1992–4. Until that time the Avon Levels were not well known archaeologically
(Barnes 1993, 5), but as the archaeological and palaeo-environmental profile has been raised since
the fieldwork (1992–3), further investigations have started to reveal a more consistent, if
ephemeral, presence of archaeological sites buried within the alluvium (eg, Locock et al. 1998;
Locock 1999; and for the Welsh side of the estuary see, for instance Bell et al 2000; Rippon 1996).
An important medieval landscape exists (see Rippon 1997) in both physical and documentary
evidence as surface features and sites, as exemplified by investigations at, for instance,
Rockingham Farm (Locock & Lawler 2000).
In this paper, we present the results of the most systematic investigation of the middle/upper
Wentlooge Formation undertaken to date. We deal with four main investigations of sediment
profiles up to 5 m thick which are reviewed in relation to previous work and related spatially
through a large-scale auger survey totalling c. 3.6 km. We are primarily concerned with the middle
and upper Wentlooge Formations, the peat within which is dated in the Avon Levels from c. 4500
cal BC (5600 BP), interrupted by episodes of minerogenic sedimentation. Peats were succeeded by
the upper Wentlooge clays at c. 1500–500 cal BC. This work does not deal with the later Rumney,
Awre, and Northwick formations (sensu Allen 1988) which are restricted to the coastal margins and
estuarine fringes seaward of the seabank. The more overtly archaeological aspects of these sites are
6
presented elsewhere (Gardiner et al. 2002), as summaries of the investigations and models of the
social and community use of the landscape.
Background
The importance of understanding the precise nature of the estuarine environment is that it allows us
to develop a picture of the sedimentological, physical, ecological and archaeological evolution of
the Avon Levels and the wider Severn Estuary Levels. Sedimentological and geoarchaeological
approaches have tended to concentrate on defining the origin of the sediment (marine, fluvial,
estuarine), on modelling sedimentation, accretion and salt-marsh growth, and in examining changes
in relative sea-level (RSL) and the effects of autocompaction.
Most of these studies define the nature of the developing landscape and surface morphology as
either high intertidal zone environment (ie, mudflat to salt-marsh) or freshwater fen. There is
certainly a shallow gradient between mudflat and salt-marsh and the definition of each environment
is not as significant in geomorphological terms as it is to archaeology. The precise nature and
location of these types of habitat are directly pertinent to the exploitation, settlement, use, and
longevity of human activity in prehistoric and historic times and the present. It is precisely these
aspects that we attempt to lay the foundations for here, and are explored in archaeological terms by
Gardiner et al. (2002).
7
Vimpenny's Lane
Awkley Interface
W E
Awkley Lane
Northwick
Hallen &
High tides
Upper
Middle
Low tides
Lower
Keuper Marls & Tea Green Marls
Upper Wentlooge Middle Wentlooge Lower Wentlooge

Colluvium Peat Easturine Sands & Gravels
Figure 2. Sedimentary features comprising the Wentlooge Formation
8
353000 354000 355000 356000 357000 358000 359000 360000
Estuarine Alluvium
Head
Keuper Marl (Red Marl)
188000
Tea Green Marl
Limestones & Sandstones
Major Excavation
Environmental Sequence
187000
Auger Transect
Awkley Lane
Awkley Interface
186000
Northwick
el
nn
185000
ha
lC
sto
Bri
184000
ore
sh
ore
183000
nf
do
Vimpennys
se
Lane
po
ex
182000
ta
Pe
181000
Hallen
180000
179000
178000
177000
Figure 3. The geology of the Avon Levels showing the location of the archaeological and
sedimentological investigations
9
THE INVESTIGATIONS
The Research/fieldwork: methods of investigation
The investigations summarised below were conducted as an archaeological and palaeo-

environmental programme of investigations (see Barnes 1993). All investigations were confined to
the route of the M5 and M4 link roads, and their precise siting of was determined through
observations in geotechnical pits, documentary research (Porter 1990; Russett 1990), and evaluation
trenching (Lawler et al. 1992). Archaeological work involved the excavation of 10 sites, a detailed
earthwork survey along the 14 km motor corridor, transplot of aerial photographs covering 54 km2
of the Avon Levels, a summary documentary survey, and a detailed photographic survey of the
village of Redwick.
Investigations which provide information relevant to the physical and social evolution of the Avon
levels are two deep sondages exposing long sediment sequences at Awkley Lane and Vimpennys
Lane, two shallow sequences exposed in relation to excavations of archaeological sites at Hallen
Marsh and Northwick, and sequences on the wetland margins at the foot of Awkley Hill at Awkley
Interface (Fig. 3).
At two of these locations (Hallen Marsh and Northwick), full-scale open area archaeological
excavation was conducted to reveal Iron Age and Romano-British activity in plan. Each was buried
beneath relatively shallow (c. 0.5 m) alluvium. At two locations in the levels deep sondages up to 5
m deep were opened by machine to investigate the middle and upper Wentlooge sequence. These
were conducted by opening trenches c. 10 m square and exposing the sedimentary sequence in a
series of steps (see Plates 2 and 3). A third investigation comprising a series of machine cut
trenches exposed the deposits at the foot of Awkley Hill revealing the alluvial–colluvial interface.
In all cases the sediments were exposed in section, drawn, and recorded, and in most cases full
suites of samples were taken for the investigation of the pollen, diatoms, snails, waterlogged plant
and insect remains, and soils. Where archaeological activity was encountered this provided some
chronology, sparse charred remains, and discarded animal bones. Where organic and peat deposits
were encountered samples were taken for radiocarbon dating. Samples were also taken through two
sequences for archaeomagnetic dating of the fine-grained middle and upper Wentlooge sediments.
Following assessment of the survival and suitability of various microsfossils and datasets to aid in
the understanding of the evolution of the Avon wetland (Wessex Archaeology 1994; 2000), a
selected programme of analysis was undertaken. Full analytical methods adopted are presented in
the Appendix.
In addition to these major field interventions and analyses, the upper 2 m of the Wentlooge
sequence was mapped by a large-scale hand auger survey comprising 411 records. This was
conducted in 6 transects (Fig. 3) with augerholes at 50 m intervals covering an area of 480,000
square metres and the augered transects themselves represent a distance of about 3600 m. The
principal aim was to map and model the upper Wentlooge sequence, as studied in detail in the
excavated exposures, to provide the basis of a three dimensional model of the evolution of the Avon
levels. Understanding of the physical evolution provides the basis for the examination of models of
the archaeological use of the Levels as a whole (see Allen and Gardiner in Gardiner et al. 2002).
The middle and upper Wentlooge Formations were of specific archaeological interest as these later
units are better preserved and more closely integrated with wider scale patterns of exploitation and
use of the Levels in the later Holocene (from c. 1000 BC).
10
The physical and vegetational evolution of the Avon levels provides an important contribution to
the development of the Severn Estuary environment. It also provides the opportunity to examine the
human interpretation and use of the landscape. Because of the three dimensional interpretations
over a relatively wide area (c. 6 km2) afforded by the detailed and dated palaeo-environmental
studies, in combination with the large-scale auger survey and sediment mapping, this information
can be used to examine the resource potential for human occupation. It can help elucidate and
explain the nature of human activity, the onset, and abandonment of both individual settlements and
of phases of settlement activity. In so doing we are able to use this information to provide models of
exploitation of the area in much wider social and geographic terms.
The sites
The sediment sequences and analysis from the five sites are presented individually before the data
is combined and used in wider interpretations of the history of the Avon wetlands as a whole. The
sites are as follows (see Fig. 3).
Awkley Lane: Deep environmental sequence in a back fen location to the east of the Avon levels
near the foot of the rising land of Awkley Hill. Profile almost 5 m deep comprising a
sequence of freshwater peats overlain by estuarine silts.
Vimpenny’s Lane: Deep environmental sequence nearly 1 km from the higher land to the east, but
to the west of locations with deeper back fen freshwater peats. The 2.8 m exposed sequence
included mainly estuarine silts and clays with a dated detrital Phragmites peat lens.
Hallen Marsh: A site situated well into the Avon wetlands to the southern end of the study area
(Fig. 4) at which large scale archaeological investigations of Iron Age (100 BC) settlement
was undertaken, but a short sequence of silts and clays up to 1m were exposed and
examined. It is known that peats occur at depth but were not encountered in the
archaeological investigations.
Northwick: Archaeological excavations of Romano-British (1st century AD) field systems in the
centre of the widest expanse of wetlands. Only the upper 1 m of the sedimentary sequences
was exposed and examined.
Awkley Interface: On the western slopes of Awkley Hill where the fine-grained deposits of the
wetlands meet the terrestrial colluvial deposits relating to the surrounding higher ground.
Limited evidence of localised Roman activity on a bench at the foot of the slope was
recorded.
1. Awkley Lane
The sequence
Evaluation within the vicinity of Awkley Lane (ST 5932 8638) produced evidence of relatively
deep, stratified and well preserved freshwater peats (Lawler et al. 1992, 57). This sequence was
examined in more detail by the excavation of a trench 11 m square and about 4.5 m deep; a further
0.35 m was hand excavated through the base of the trench to sands of the lower Wentlooge
Formation. The deposits recovered can be paralleled with those recorded in evaluation (Lawler et
al. 1992, 57–8), but with greater differentiation in the deposits being recorded during excavation.
11
A series of fine grained, greyish brown silty clay, alluvial deposits (102–108) with very low
magnetic susceptibility (4 to 6 SIu10-8Kg-1) dominated the upper 1.8 m of the profile (6.41 to 4.31m
OD). Between 4.31 and 4.16 m OD two putative stabilisation horizons (107 and 109) were
recognised in the field on the basis of their darker hue (possibly organic content), but this was not
reflected in the magnetic susceptibility profile (Fig. 4), although a gradual rise in enhancement was
seen with depth. Below 4.16 m OD a further 1.4 m of light grey (10YR 6/1) mixed silty clay
alluvium with a greyish and blue hue were present. The magnetic susceptibility profile shows much
higher levels averaging c. 18 SIu10-8Kg-1. Enhancements of up to 38 SIu10-8Kg-1 at about 2.9 m are
not reflected in the sedimentary record and may be due to post-depositional effects of waterlogging.
From 2.30 to 3.65 m OD (110) fine (1–2mm) weak laminations with occasional sand and silt
partings indicate a series of flood events (flood couplets) or possibly indicates a former tidal
boundary, with evidence for the presence of freshwater (J.R.L. Allen, ers, comm.). A large isolated
and waterlogged and weathered ‘bog’ oak log near the base of this unit dated to 2460–2140 cal BC
(3816±50 BP, NZA-12532). This probably represents higher energy flood detritus. Occasional fine
organic matter, possibly Phragmites stem fragments, are distributed throughout this alluvium.
The sediment changes abruptly to a series of humified detrital monocot and wood peats, the surface
of which displayed a distinct ridge in the recorded section (Fig. 4). The ‘ridge’ was clearly marked
by a thin band (20 mm) of loose, black peat, presumably a desiccated surface, beneath which was
80 mm of alluvium hardly distinguishable from that above. This band rises to 3.04 m OD but dips
away to the west, indicating some truncation and modification of the upper peat surface (see Fig. 4).
Nearly 1.4 m of stratified peat occurs below 2.96 m OD. The upper deposit peats are black, well-
humified fen peat (113) over wood peat (121), and a dense black compressed and partially dried
band (120) at 4.10 m OD seals a black well humifed peat (115). Reddish, oxidised humified fen
peat (116 and 118) occurs at the base and is separated by a very thin (20 mm) lens of grey silty
alluvium at 1.99 m OD. The base of the peats were exposed at 1.57m OD by a small hand
excavated pit in the base of the trench revealing an inorganic grey silty sand below.
A full suite of samples was taken from the deposits, including mollusc columns, pollen monoliths,
samples for particle size analysis and samples for investigation of palaeomagnetic dating.
Chronology
The surface of peat at 2.62m had been dated in the evaluation exercise to 2920-2600 cal BC
(Lawler et al. 1992, 115). A series of nine AMS results from the peat and the bog oak log provide a
sequence for the rate of peat accumulation (Table 1; Fig. 5). The start of peat accumulation (1.44m
OD) is dated to 4530–4350 cal BC (NZA-12774 5603±50 BP). The formation of 0.9m of humifed
monoct (?fen) peat occurred steadily over a period of about 1½ millennia until the wood peat (121)
at 3.04 to 2.42 m OD: 3090–2700 cal BC (NZA-12528 4286±55 BP). A depth of c. 1.3 m of wood
peat developed in a period of a couple of centuries between 3090–2700 cal BC to 2860–2470 cal
BC. Alluviation occurred and was then followed by a thin a final peaty horizon (111, dated to 2900-
2600 cal BC) prior to return to salt-marsh and estuarine conditions and a fully alluvial sequence.
The oak log, contained in excess of 40 rings but no match for the pattern could be established in
dendrochronological analysis. It was concluded that that the tree had been subject to considerable
12
WA Awkley Lane GGAT G030
Mollusc & soil
column
6 m O.D.
Pollen Zones
104
5 m O.D.
105
5
106
107
108
109
4 m O.D.
110
3 m O.D. Log
112 111
4190+_ 70BP
113
0 10 20 30 40 50
4190+_ 60BP
113
x10-8 SI/kg
121 3
114
120
2 m O.D. 115
116
2
118 Alluvial silty clays (brownish red hue) Insect sample
1 Alluvial silty clays (bluish grey hue) Plant macro fossils sample
?Stabilisation horizon C14 sample
119
Peat (wood peat vs humified peat)
0 1m
Sandy alluvium
Figure 4. Awkley Lane Wentlooge Formation profiles
13
WA Awkley Lane
110
3816+_ 50BP 3m O.D. 3816+_ 50BP
Log
3991+
_ 55BP
111 3991+
_ 55BP
112 4044+
_ 50BP
4044+
_ 50BP
3885+
_ 45BP
113 3885+_ 45BP
113
4286+_ 55BP
4286+
_ 55BP
121
114
4745+
_ 50BP 4745+_ 50BP
120
115
2m O.D.
116
4683+_ 55BP 4683+_ 55BP
118
5603+
_ 50BP 5603+_ 50BP
Insect sample 119
Plant macro fossils sample
C14 sample
Alluvial silty clays (bluish grey hue)
Sandy alluvium 1m O.D.
0 1m
5000BC 4000BC 3000BC 2000BC
Figure 5. Radiocarbon determinations in relation to their OD and the Middle Wentlooge wood and fen peat
14
ecological stress during growth, and this is why the ring pattern does not fit in with any dated
reference chronologies. It is likely that the distressed growth pattern occurred because local
growing conditions were unfavourable, and this stress may be due to the proximity of trees to saline
water and saline water ingression.
The inorganic fine-grained sediments contain no organic matter to provide a radiocarbon date. Even
the possible stabilisation horizons do not contain sufficient carbon to provide a radiocarbon result.
The sequence was, however, sampled for palaeo-magnetic dating (GeoQuest 1993). Sediments from
this sequence provided a coherent record or geomagnetic secular variation which is particular well
preserved below the gleyed zone (ie, in the weakly laminated silty clay (layer 110) between 4.10
and c. 3 m OD. Demagnetisation tests confirmed that the sediments contain a stable remanent
magnetism and there is a good agreement between replicate samples. Correlation between the
profile and master curve is complicated by changes in the rate of deposition and the possibility of
hiatuses and removed sediment. Absolute dates originally provided (GeoQuest 1993) have been
revised in the light of revision of the ‘master curve’ and radiocarbon determinations from the basal
peats. As a result M. Noel (pers. comm.) is able to indicate that deposits near to top of the profile
(base of layer 103 at 5.7 m OD) are not younger than c. AD 500. Two index points were indicated
as i) about 5.1 m OD at the top of layer 105, c. 0 BC, and ii) about 4.5 m OD at the base of 106 and
top of 107 c. 500 BC. This latter date occurs immediately above the series of stabilisation or flood
event layers (107, 108, and 109), at the base of the alluvium of a reddish brown hue.
Table 1. Radiocarbon determinations and calibrated results from the Middle Wentlooge
context context type depth OD material lab no GC13‰ result BP date Cal BC
Awkley Lane
110 bog oak 3.00 wood NZA-12532 -25.7 3816±50 2460-2140
111 peat band 2.92 humic acids NZA-12590 -26.63 11,350±120 11,840-11,070
111 peat band 2.92 plant matter NZA-12754 -26.63 3991±55 2840-2310
GGAT top of peat 2.62 peat GU-3119 -28.3 4190±70 2910-2580*
GGAT top of peat 2.62 wood alder GU-3120 -28.3 4190±60 2900-2590*
113 top peat 2.78 peat NZA-12533 -27.79 4044±50 2860-2470
113 peat 2.64 peat NZA-12534 -27.48 3885±45 2470-2210
121 peat 2.42 wood alder NZA-12528 -27.68 4286±55 3090-2700
120 band 2.14 wood peat NZA-12529 -27.48 4745±50 3640-3380
117/8 top peat 1.82 twigs NZA-12530 -26.78 4683±55 3630-3360
elm decline 4800 BP
118 base of peat 1.44 female alder NZA-12774 -26.57 5603±50 4530-4350
cones
* GGAT 1992
15
Depth (m) OD Layer Description Comment
0–0.10 6.41.–6.31 100 Dark brown (10YR 3/3) stonefree clay loam Topsoil
0.10–0.25 6.31–6.16 101 Greyish brown (10YR 5/2) stonefree silty clay B horizon
0.25–0.55 6.16–5.86 102 Pale brown (10YR 6/3) stonefree clay, rare medium Upper Alluvium
fleshy roots and vertical voids
0.55–0.70 5.86–5.71 103 Greyish brown (10YR 5/2) stonefree clay with rare Upper Alluvium
mineral (?Fe) staining and gleying along vertical
rootlet channels 3mm wide, diffuse/gradual boundary
0.70–1.20 5.71–5.21 104 Greyish brown (10YR 5/2) stonefree clay with Upper Alluvium
frequent mineral staining, gleyed, no sedimentary
structures evident, diffuse/gradual boundary
1.20–1.70 5.21–4.71 105 Brown (10YR 5/3) stonefree silty clay with patches Upper Alluvium
of light grey material, gleyed, no sedimentary
structures evident,.
1.70–1.80 4.71–4.61 106 Grey (10YR 5/1) stonefree silty clay, no sedimentary Upper Alluvium
structures evident
1.80–1.95 4.61–4.46 107 Dark grey (10YR 4/1) stonefree clay with up to 10% ?Stabilisation
flecks of black materail. horizon
1.95–2.10 4.46–4.31 108 Grey (10YR 5/1) stonefree clay with occasional (5%) Grey Alluvium
flecks of black material
2.10–2.30 4.31–4.10 109 Dark grey (10YR 4/1) stonefree silty clay with up to ?Stabilisation
10% flecks of dark material. horizon
2.30–3.65 4.10–2.75max 110 Light grey (10YR 6/1) mixed clay with very Grey Alluvium
occasional (2%) flecks of dark organic plant material
(?Phragmites). Thinly (1-2mm) and weakly
laminated with occasional sand and silt partings
(flood couplets). A large log (oak) is preserved in
layer near its base.
3.35–3.37 2.92–2.90 111 Black (10YR 2/1) peat. Thin layer of loose black Peat lens
desiccated peat following contour of ridge of peat 113
below.
3.37–3.45 2.92–2.86 112 Light grey (10YR 6/1) stonefree clay. Layer of Grey Alluvium.
alluvial deposit lying over ridge of peat 113
3.45–3.75 2.86–2.56 113 Black (10YR 2/1) humified peat, very little Peat
waterlogged wood/twigs etc. Layer of peat with a
ridged on the south side of the section
3.75–4.05 2.56–2.26 121 Black (10YR 2/1) peat with up to 20% wood Wood Peat
inclusions
4.05–4.12 2.26–2.08 114 Dark reddish brown (5YR 2.5/2) peat with occasional Peat
(5%) wood inclusions, clear smooth boundary.
4.12–4.14 2.08–2.06 120 Black (10YR 2/1) peat. Thin layer of dark peat Peat
running across section, clear smooth boundary
4.14–4.22 2.06–1.98 115 Black (5YR 2.5/1) peat with occasional (up to 5%) Peat
small pieces of wood.
4.22–4.38 1.98–182 116 Dark reddish brown (5YR 2.5/2) peat with occasional Peat
(5%) wood inclusions, one moderately sized piece of
wood.
4.38–4.40 1.82–1.80 117 Grey (10YR 5/1) stonefree clay. Thin alluvial layer.
4.40–4.80 1.80–1.40 118 Dark reddish brown (5YR 2.5/2) peat with occasional Peat
wood inclusions.
4.80+ 1.40+ 119 Grey silty sand. Sandy
Sampling the sequence
The minerogenic and peat deposits below the first step (Plate 1) at c. 3.7 m OD were sampled in
monolith tins to facilitate subsampling for pollen and diatoms and one portion of the upper
sequence encompassing a putative buried soil (107) was also sampled in a monolith tin. The upper
sediment sequence and the gap between the two sets of monoliths was filled with sub-samples from
soil sequences of bulk samples removed for land snail and grain size analyses (see Fig. 4). The
minerogenic deposits below 103 (5.28 m OD) were sampled as contiguous disturbed samples for
16
Plate 1. Sampling the minerogenic sediment Plate 2. Awkley Lane excavation showing the fully
sequence in monoliths at Awkley Lane excavated and exposed upper and middle Wentlooge sequence
Awkley Lane Clay (100%)
1119 1125
1100 1118
50% 50%
y
Cla
Silt (100%) Sand 50% Sand (100%)
Figure 6. Awkley Lane; particle size triangle
17
snails and grain size analysis. In addition three groups of three bulk samples were taken from top,
middle and bottom of the recovered peat for waterlogged plants and insects. Most of the
minerogenic deposits were also sampled for archaeomagnetic dating.
Textural features of the minerogenic sequence

by J.R.L. Allen
A series of 26 samples was analysed for grain-size distributions and the triangular diagram (Fig. 6)
shows values of the clay-silt-sand ratio. The methods used are detailed in the Appendix. The
triangular diagram (Fig. 6) show that the sediments range from clayey-sandy silts to clayey silts.
Falling into a finer (top of layer 106–105; 4.68–5.28 m OD) overlying a coarser (layers 112–106;
2.68–4.68 m OD) subset, they form a distinct trend on the triangular diagram, similar to trends
noted at Hallen Marsh and Vimpenny’s Lane (see below).
The silts of the stratigraphically lower subset are surprisingly coarse grained for such an inland site.
The grain-size distributions show strongly defined modes which typically exceed the means in
value, well-developed coarse tails, and noticeable proportions of sand. The resolution provided by
the sampling interval is not great, but the presence of three textural trends may be suggested: fining-
up (112 to middle of 110; 2.68–3.18 m OD), coarsening-up (middle 110; 3.18–3.58 m OD) and,
finally, fining-up (top of layer 110 to 106; 3.58–4.59 m OD). There is nothing texturally to
distinguish the two soils considered to be present in this part of the sequence.
A sharp and substantial reduction in grain size occurs between samples in 107 at 4.49–4.68 m OD,
interpreted as a possible soil reminiscent of that reported from Vimpennys Lane. The overlying
silts, belonging to the finer subset, display a strong coarsening-up trend, as is also seen at
Vimpennys Lane. At the latter site, however, where a longer sequence terminating in a soil was
sampled above the textural break, the coarsening-up pattern is followed by a short fining-up trend.
The unusually coarse texture of the lower silts suggests either that they were deposited close to a
major tidal creek which reached deeply back into the marsh; that they formed near a no-longer
recognisable stream issuing from the hinterland; or that they include hillwash from nearby slopes
underlain by head/bedrock. Under the first possibility, the textural trends may suggest changes in
either the position of the distant marsh edge from which the creek extended and/or the rate of sea-
level change. Under the second and third possibilities, the trends could register variations in
sediment yield from the nearby slopes induced by changes in climate and/or vegetation cover.
The dramatic reduction in grain size between samples at 4.59 m OD points to a sharp change in
sedimentary conditions, and the sediments are not what would be expected at such an inland site.
Either the postulated creek/stream suddenly ceased to be active, leaving only distant sources of
sediment, or the site became isolated in some way from the neighbouring slopes. The subsequent
coarsening-up trend suggests an episode of marsh- edge retreat and/or positive sea-level tendency.
Pollen analysis
by Robert G. Scaife
The Awkley Lane profile is the deepest at c. 4 m and acts as a key to establishing a local and
regional pattern of vegetation and environmental change. A total of 65 pollen samples, of which
62 contained pollen, was examined from the profile. Material was obtained in the field using
18
monolith profiles taken directly from the excavated section which were sub-sampled for analysis
and stratigraphical description in the laboratory of Wessex Archaeology.
A total of five pollen assemblage zones (p.a.z.) and a number of pollen assemblage subzones
(p.a.s.z.) has been recognised at Awkley (Fig. 7). These have been delimited and characterised
from the base of the profile at 1.44m OD upwards as follows.
AWK:1 1.44–1.66 m OD, detrital peat/lower fen peat. Ulmus-Tilia-Quercus-Corylus avellana type-Alnus. APF
are from 138,000 grains/ml at the base to 215,000 grains/ml at the top. Tree, shrub pollen are dominant (to 75% and
40 % resp.) along with fen/marsh taxa which are dominated by Alnus (to 65%). This basal zone is characterised by
highest values of Tilia (to 33%), Ulmus (to 14%), and Fraxinus (4%). Other trees include sporadic occurrences of
Betula, Pinus Populus, Fagus, and Taxus. Shrubs are dominated by Corylus avellana type (declining from 40%).
There are few herbs in this zone compared with subsequent zones (10% NAP) with only small percentages of
Poaceae (6%) being of note. A single cereal type grain is noted. Alnus is the dominant wetland taxon (to 65%) with
Cyperaceae (to 20%). Spores comprise Dryopteris (monolete) type (c. 20–25%) and small numbers of Polypodium
vulgare (<5%).
Sampling gap in profile from 1.54 to 1.78 m OD.
AWK:2 1.66–2.53 m OD, detrital lower fen &wood peat. Quercus-Corylus avellana type-Dryopteris type. APF
values attain highest values of up to 1,730,000 grains/ml in the lowest level of this zone and subsequently decline
upwards. In the intervening gap in the sample profile between 1.54 m and 1.78 m OD there are reductions in Ulmus,
and Tilia and an increase in Corylus avellana type. Fraxinus (<5%) is present throughout. Three pollen assemblage
sub zones have also been recognised. Sub-zone-b (1.86–2.21 m OD) shows further reduction in Ulmus (to absence),
Tilia (to 5%) and Corylus type (from highest values of 70% to c. 25%). Sub-zone-c wood peat (2.21–2.53 m OD)
shows some expansion/regeneration of the these elements; especially Ulmus to 10–12% and Corylus to c. 40%.
There is a minor expansion of Fraxinus and Taxus is also noted. Herb pollen increases markedly with increased
taxonomic diversity and particularly expansions of Poaceae (to 59%) peaking in sub-zone-a (to 59%), Asteraceae
types-Aster type in sub-zone-a (peak to 17%) and Bidens type in sub-zone-b (6%). Plantago lanceolata (<5% in
sub-zones a and b), Chenopodiaceae (8%) and Apiaceae type 3 are of note. Alnus values are highest in sub-zone-a
(to 90%) but decline in sub-zone-b (av. 20–25%) with a peak of 65% at 2.38 m OD in subzone c). There is a minor
expansion of Salix and also peaks of Cyperaceae in sub-zones a and c (40% and 30% respectively). Typha
angustifolia/ Sparganium and T. latifolia become more important in this zone. Aquatic megaphytes include a peak of
Lemna at the top of sub-zone-a. Spores of ferns are dominated by Dryopteris type (monolete spores) which increase
to high values (80–90%) in this zone. Of note is a peak of Thelypteris palustris at the top of sub-zone-c (to 20%).
AWK:3 2.53–2.82 m OD, humified detrital peat (context 113). Betula-Corylus avellana type-Ericales. APF
Values increase to between 220,000 and 855,000 gains/ml. This zone is characterised by a sharp decline in herb
percentages (especially Poaceae) and a further decline of Tilia (to <5%) and a sharp reduction in Dryopteris
(monolete) type. Quercus also declines to 10–15%. These are in response to a very marked expansion of Betula
which is dominant peaking to 80%. Of note are peaks of Erica (5%) and Calluna (to 18%). Pinus, Ulmus, Tilia, and
Fraxinus remain consistent with small values (<5%). Sporadic Fagus is noted. A possible sub-zone (AWK:3-b)
change occurs between 2.64 and 2.58m OD where Quercus and Corylus type expand in contrast to a sharp declines
in Betula, Erica, and Calluna. Betula re-expands to high values but less than earlier in this zone. NAP decline with
sharp reduction in Poaceae, to low levels. There is, however, a minor peak of Poaceae between 2.64 m OD and 2.60
m OD (sub-zone 3-b). There are also declines in Plantago lanceolata, Apiaceae, and Asteraceae types noted in the
preceding zone. Alnus is the principal marsh autochthonous taxon (20–30%) but reduced over the preceding zone
AWK:3. Other wetland taxa are somewhat reduced over AWK:3 with only sporadic occurrences of Salix, Typha
angustifolium/Sparganium type, and other aquatic megaphytes. Of note are peaks of Osmunda regalis and
Sphagnum. Spores of Dryopteris type are markedly reduced over the high values present in AWK:2.
AWK:4 2.82–3.95 m OD, upper Wentlooge sediments. Quercus-Corylus avellana type -Chenopodiaceae -
Pteridium aquilinum-Polypodium-Pre-Quaternary Palynomorphs. APF values decline with the stratigraphical
change from peats to minerogenic sediments. This local pollen assemblage zone is defined by sharp reductions in
Betula, and expansions of Cyperaceae, spores of Pteridium aquilinum, Monolete/Dryopteris type. Polypodium
vulgare and miscellaneous microfossils (Dinoflagellates and derived, pre-Quaternary palynomorphs). Absolute
pollen frequencies decline sharply with change to minerogenic sediments from underlying peats. High values of
Betula in AWK 3 decline very sharply to <5% and subsequent absence. There are also substantial reductions in
Erica and Calluna and Fraxinus. Quercus attains highest values (65%) with some expansion of Pinus (to 8%). Tilia
remains (low values) but there is an increase in its degraded grains from the base of this zone associated with the
19
change to minerogenic sediments. Quercus is dominant (to 65%) with Corylus avellana type (20–30%). NAP
expand after the low values of AWK:3 peaking to 75% of tdlp. This expansion comprises increases in
Chenopodiaceae (15–20% with a peak to 68%), Poaceae (to 30%), Plantago lanceolata (<5%), Cereal type (<5%).
Within the marsh category, Alnus (to c. 20%) is consistent throughout. Sedges show some increase along with
Potamogeton type. Spores become important with substantial expansions of Pteridium aquilinum, Dryopteris type
(monolete) and Polypodium vulgare. Pre-Quaternary palynomorphs are incoming and attain high values.
Dinoflagellates of Holocene or possibly pre-Quaternary origin also become important.
AWK:5 3.95–5.24 m OD, upper Wentlooge sediments. Quercus-Corylus avellana type-Poaceae-Cyperaceae.

APF values remain low (especially at the lower zone boundary where some levels did not produce pollen) but
slightly increased over AWK:4 (ranging from c. 4000 to c. 40,000 grains/ml.). This zone is delimited by reduction in
Pinus to sporadic occurrences and some expansions of Betula (including a peak at 4.66 m OD) and Cyperaceae (to
50%). Quercus (20–30%) and Corylus avellana type (to 30%) remain the dominant trees and shrubs with sporadic
occurrences of Fraxinus, Fagus, Taxus and Ericales after absence in the preceding zone. Herbs remain important (to
70%). Chenopodiaceae remain consistent from the preceding zone. Poaceae becomes dominant (to 50%) with an
increase in the diversity of types – Cereal type, Secale cereale (single grain), Fabaceae spp., Polygonaceae spp.,
Plantago lanceolata (to 10%), and Asteraceae types (Lactucoideae, Centaurea spp., Bidens, Anthemis type, Cirsium
type. Of note are halophytes including P. maritima type (1%), Armeria ‘A’ line, Armeria ‘B’ line, and possibly
Aster type and large (non-cereal Poaceae). Within the fen/aquatic category, there is a significant increase in
Cyperaceae (to 50%). Alnus remains consistent with small values (average 10%). There are also sporadic
occurrences of Salix, Caltha type, Littorella uniflora, cf. Butomus umbellatus, and Alisma type, Typha latifolia,
Typha/Sparganium, and the spores of Osmunda regalis. Aquatics are present include-Myriophyllum, Callitriche,
Lemna, Potamogeton type (possibly Triglochin) and occasional cysts of algal Pediastrum. Spores of Pteridium
aquilinum, Dryopteris (monolete type), and Polypodium vulgare and Miscellaneous types (Palynomorphs and
Dinoflagellates) remain as AWK:4. There is some increase in Sphagnum although percentages remain <5%.
AWK:5-a. At a depth of 4.66 m OD there is a shallow more organic horizon. This appears to be a stabilisation
horizon (? soil) within the otherwise salt-marsh silts. Palynologically this is characterised by a small expansion of
Alnus (30%) and slightly above by increase of Betula (15%).
Prehistoric vegetation and environmental changes

Interpretation of the pollen data at Awkley must take into account the taphonomy of the pollen
recovered from the lower organic peats which contrast with the largely inorganic and fluvially
derived upper sediments. Furthermore, differentiation between the on-site (autochthonous)
vegetation and the vegetation communities of the adjacent well drained terrestrial soils is
important.
The Lower Fen Peat: In the basal and, therefore, earliest pollen zone (AWK:1), the peat
accumulated under a typical alder fen carr woodland (Alnetum glutinosae). This somewhat dry
fen carr supported a ground flora of ferns (eg, Dryopteris felix femmina and D. felix mas-female
and male fern) and sedges such as Carex paniculata (tussock sedge). Taxus baccata, although
only a single record, was typically a tree of such dry fen carr woodland (Godwin 1975a; Scaife
2000). This fen carr apparently developed on a previously wetter, fluvial or marine sandy deposit.
This is comparable with the typical Somerset Levels hydroseral succession and is seen
throughout low-lying regions of the Bristol Channel.
The location of the Awkley Lane site in relatively close proximity to higher, dry-land areas has
resulted in a clear representation of the vegetation of these areas. In zone AWK:1, Tilia (Tilia
cordata; small leaved lime) attains high pollen values which suggest local dominance of this
largely under-represented pollen taxon (Andersen 1970; 1973). Its proximity and importance is
also attested by the presence of fruits in the peat (see below). The importance of lindens in the
landscape is discussed below. Other elements of the dry-land flora comprise especially Quercus
(oak), Ulmus (elm), Fraxinus (ash), and Corylus avellana (hazel). Whilst Tilia was clearly the
dominant woodland element, these tree and shrub taxa were also of importance in the local
environment. It is possible that oak, ash, and hazel formed parts of the dry carr woodland or
woodland fringing the edge of the mire – perhaps on the thicker, down-slope soils.
20
A
3816 + 50 BP
3991 + 55 BP
4044 + 50 BP
3885 + 45 BP
4285 + 55 BP
4745 + 50 BP
4683 + 55 BP
5603 + 50 BP
Figure 7a. Awkley Lane; pollen diagram.
21
B
Figure 7b. Awkley Lane; pollen diagram.
22
C
Figure 7c. Awkley Lane; pollen diagram.
23
D
Figure 7d. Awkley Lane; pollen diagram. Note that alder (Alnus sp) is included with fen species.
24
E
Figure 7e. Awkley Lane; pollen diagram.
25
F
Figure 7f. Awkley Lane; pollen diagram.
26
Ilex aquifolium (holly) is recorded. It is usually extremely poorly represented in pollen spectra
and thus, as with lime, implies some importance. Alternatively, and very likely, these woodland
trees formed lesser elements within the dominant lime woodland. Radiocarbon dating confirmed
the age of this basal peat as 4530-4350 cal BC (NZA-12774, 5603±50 BP), ie, late Atlantic age,
middle Holocene (Flandrian Chronozone II) or, at latest, very early Neolithic (Flandrian
Chronozone III). In either case, the pollen is a representation of the dominant and maximum
extension of middle Holocene woodland prior to any extensive human influence and
deforestation. This is also reflected by the small representation of non-arboreal pollen in the
Awkley profile as a whole.
In pollen zone AWK:2, the importance of lime and elm is reduced and there are increasing
numbers of herbs. It is very unfortunate that there is a gap in the sample record, since evidence of
a phase of important ecological change has been lost. However, it is suggested that this zone
AWK:l/2 transition is the early part of the prehistoric elm decline at c. 5000 BP, or slightly later
if dates for the Somerset Levels pertain. This is confirmed by the radiocarbon date of 3630-3360
cal BC (NZA-12530, 4683±55 BP) near the base of this zone. Further reduction in elm pollen at
1.86 m OD (sub-zone a and sub-zone b) and sharply expanding values of Poaceae (grasses) with
sporadic occurrences of cereal type pollen, Plantago lanceolata (ribwort plantain), and peaks of
other herbs including Asteraceae (Bidens type and Aster type), Urtica (nettle and pellitory),
Chenopodiaceae (goosefoots and oraches) are all indicative of changes in the local dry-land
vegetation. The reduction in lime and elm and the increase in herbs is strongly indicative of the
first impact of human disturbance in the environment and of agriculture attributed to a local
Neolithic economy.
Associated with this event are also changes in the status of the mire at the sample site. In sub-
zone b, there is evidence of an ephemeral phase of increasing wetness as evidenced by increasing
pollen frequencies of reed swamp taxa including Cyperaceae (sedges), Typha latifolia, and Typha
angustifolia/ Sparganium type (bulrushes) and bur reed, and aquatic megaphytes such as
Potamogeton (pond weed), Callitriche (water starwort), and Lemna (duckweed). Alnus (alder)
declines sharply in sub-zone b in response to this increasing wetness. Local woodland clearance
on the nearby interfluve may have caused a reduction in evapotranspiration, and increased
surface run-off resulting in a higher ground water table. The fragile balance of alder in fen carr
would clearly have been disturbed resulting in the demise of carr. Typically, alder will tolerate
flooding of the peat surface around the root boles for some 3–4 months of the year during winter
(Tansley 1949; McVean 1953; 1956). However, with the exception of a single peak of Alnus at
2.38 m OD, values of alder remain much reduced over its earlier dominance. Fluctuation of the
reed swamp taxa (see Cyperaceae and Typha latifolia) may indicate a rather unstable/variable
wetland habitat. Small increase in Salix (willow) pollen may be also significant since willows are
largely under-represented in pollen spectra. Because of the very substantial numbers monolete
Dryopteris spores (typical ferns) it is also likely that these are ferns growing within the wet, fen
carr marsh habitat-although subsequently the diagnostic marsh fern (Thelypteris palustris is well
represented.
By pollen assemblage sub-zone c, the numbers of grasses and weeds (see for example Plantago
lanceolata) are much reduced. There is an accompanying increase in arboreal and shrub pollen.
By the top of local pollen assemblage zone 2, there is some evidence of woodland regeneration
with expansions of Tilia and Ulmus to c. 50% of previous values and some expansion of
Fraxinus (ash), the latter being a typical secondary woodland element. These changes may
suggest either local abandonment or change in agricultural or land use on the adjacent dry land.
Although only a single grain/record, Fagus in this zone may be significant as this tree and also
ash are very poorly represented in pollen spectra.
27
Pollen zone AWK:3 exhibits a very marked change in the character of local woodland. Betula
(birch) becomes the dominant local vegetation. Alnus remains consistent at levels seen in
subzone a and b of the preceding zone 2. Quercus, however, declines sharply in response to the
increasing importance of birch. Corylus (hazel) remains broadly similar with some reduction.
Clearly, there was a marked increase in growth of birch woodland. The decline in oak noted
above may be both a real decrease in the area of growth and also a statistical function of the
expansion of birch within the same pollen sum. The relatively small change of hazel implies that
hazel (scrub?) also expanded significantly. Also associated with this change is the expansion of
acidophilous elements including Ericales (ling and erica), a small peak of Osmunda regalis (royal
fern), and Sphagnum (bog moss).
Birch has been recognised particularly as a component of fen carr woodland in the Somerset
Levels (Beckett 1979) and these changes and high values present here may suggest a phase of
drier fen wood or scrub development. Certainly, values of Cyperaceae are significantly reduced
and the expansion of the acidophiles noted may represent the development of a different ground
flora to that which existed in the alder carr with sedge and fern understorey. This does not,
however, explain the continuance of alder at preceding levels unless differing areas of the
mire/low lying ground were colonised rather than one community replacing another. A further
alternative is the possibility that oak and lime woodland growing on the near terrestrial zone was
cleared for its wood (for there is no real indication of agricultural expansion). The sandy
character of the local geology would have been susceptible to leaching and creation of
podzolic/acid soils suited to colonisation by ericaceous communities and in wetter areas on the
fringes of the mire by Sphagnum. Plant macrofossil (seed and wood) data and Coleoptera indicate
that the autochthonous vegetation communities of the fen were Sphagnum bog and densely
vegetated marshes.
The Upper Clay-Silt: From 2.82 m OD there is a marked change in stratigraphy from the fen carr
wood and detrital peat of pollen zone AWK:1, 2 and 3 into the more or less homogeneous blue-
grey clay and silts of zone 4 upwards. Clearly, this change implies a major environmental change
to alluvial and/or salt-marsh conditions. This changing environment and resultant
lithostratigraphy importantly will have also changed the taphonomy of the pollen. Whilst the peat
horizons discussed above (zones 1–3) will follow the typical pollen models for localised basins in
a wooded environment (eg, Tauber 1965; 1967), interpretation of the minerogenic sequences
must account for other taphonomic aspects such as fluvial transport and deposition. This includes
the reworking of older soil/sediment reservoirs containing pollen and an extended pollen
catchment through increased ‘openness’ as well as microfossil transport from wider parts of basin
along with river-borne sediments. Furthermore, the sediments discussed here in pollen zones
AWK:4 and AWK:5 are largely in part formed in marine/brackish water salt-marsh caused by
regional positive RSL causing the widespread estuarine inundation of the fen peat/marsh zone.
In zone AWK:4, Quercus (oak) and Corylus avellana (hazel) are the dominant pollen taxa
reflecting the broader importance of this woodland within the region. Tilia cf. cordata (not
degraded) remains in small numbers throughout indicating continued growth within the
catchment on drier better drained soils. Ash percentages decline although this does not
necessarily imply that there was a real reduction in its growth/ importance since it is very likely
that the pollen sources became farther away from the sample site. This frequently occurs with
Tilia (Waller 1994; Scaife 2000). In zone 4, this is exemplified by the reduction in not only
Fraxinus but also of the Ericales noted in zone AWK:3. In contrast to these declining responses
to alluviation/flooding, the pollen spectra give a wider picture of the regional vegetation. In
addition to the oak and hazel noted, which maintain importance due to their high pollen
productivity and anemophily (wind pollination), evidence for regional agriculture is more
pronounced. Opening of the environment allows increased airborne input of the airborne pollen
28
catchment as well as pollen fluvially derived from further afield. Here, there is an increase in
cereal pollen especially in zone 5 (note Secale cereale – rye) along with typical weeds of
cultivation such as Plantago lanceolata, Polygonaceae spp., and possibly other taxa which are
not differentiable to species or even genus (eg, Fabaceae).
Marshland: The change to grey minerogenic sediments represents a shift from anaerobic peat
formation to freshwater and/or more likely brackish estuarine water and salt-marsh conditions.
From the base of pollen zone AWK:4, and the start of the grey sediments, there is an expansion
of Chenopodiaceae which is a strong indication of salt-marsh development typified by Salicornia
(glassworts) and Atriplex (oraches). Other halophytes are present and become increasingly
important up the profile and especially in the uppermost zone (AWK:5). These taxa include
sporadic but nevertheless significant presence of Armeria ‘A’ and ‘B’ line (thrift and sea
lavender), Plantago maritima (sea plantain), Aster type (including sea aster), and large Poaceae,
the latter being some halophytic grasses with large diameter pollen grains but not of cereal type.
Potamogeton type may be Triglochin maritima but also includes Potamogeton (pond weed)
which has similar pollen morphology preventing identification to a lower taxonomic level. These
taxa are clear evidence for marine/brackish water incursion of the previously freshwater fen
habitat. This is not unexpected since there is much evidence for this in other low lying areas
along the Severn/Bristol Channel. In the Somerset levels this is evidenced clearly at Meare Heath
(Hibbert 1980) where the raised Sphagnum ombrogenous mire was inundated at c. 910–760 cal
BC (SRR-914, 2624±45 BP) lasting for c. 150 years.
Freshwater marsh/aquatic vegetation elements are present in the sediments since the out-flowing
river catchment was transporting pollen from upstream sources. Cyperaceae are dominant with
other fen marsh elements present including Caltha type (marsh marigold), Butomus umbellatus
(flowering rush), Alisma plantago-aquatica (water plantain), Typha latifolia (bulrush),
Typha/Sparganium (lesser bulrush and bur-reed), Myriophyllum spicatum (spiked water-milfoil),
Callitriche (water starwort), Lemna (duckweed), and cysts of algal Pediastrum. Potamogeton
type is also present but as noted above may also derive from Triglochin maritima a halophyte or
from water pondweed (Potamogeton). Alder from fringing or valley fen carr is consistent
throughout the two upper zones but values suggest that its importance in the region was
diminished compared with the earlier and more extensive areas of fen peat carr in the Avon
Levels.
Summary of environmental change

i) Pollen zone AWK:1. Late Atlantic/middle Holocene (Flandrian Chronozone II; c. 4500
cal BC): Alder fen carr with a ground flora of sedges, grasses, and ferns occupied the
sample site. Adjacent dry-land was dominated by lime woodland, with oak, elm, hazel,
and ash. These latter may, however, have been constituents of drier parts of the fen carr
and on heavier soils at the base of the valley sides. The pollen catchment is likely to have
been of small extent being in the order of hundreds of metres surrounding the site.
ii) Zone AWK:2a Lime, elm, and oak remain important but hazel expands sharply. In sub-
zone b, lime and elm woodland is reduced and this phenomenon is thought to be the
Neolithic ‘Primary Elm Decline’. Elm reduction was most possibly due to the spread of
elm bark beetle and disease (Girling 1988) and the clearance of the lime woodland by the
incoming of Neolithic/human activity. This is evidence by expansion of grasses, cereals
and weeds of cultivation and human disturbance. Date c. 5000–4500 BP. Alder remains
important on the site but there phases of wetter sedge fen and reduced alder. These phases
may be due to the human activity increasing the ground water table locally (woodland
clearance, increased run-off, decreased evapotranspiration).
iii) Zone AWK:2c there was a reduction in human activity – at least cultivation, allowing
secondary woodland regeneration of elm, lime, and possibly ash. Hazel remained
29
important and the carr community saw the stabilisation again of alder carr woodland with
a ground flora dominated by marsh type ferns.
iv) A substantial ecological change in zone AWK:3 (c. 4200 BP) may be due to clearance of
oak, lime, and possibly elm woodland for timber on nearby drier areas, colonisation by
pioneer birch scrub and possible extension or increase in flowering of local hazel. Sandy
soils became degraded giving pedalogical conditions suited to the establishment of an
acidophilous heathland community (Erica, Calluna, and Sphagnum moss) c. 2550 cal BC.
An alternative interpretation is that drying out of the fen carr dominated by alder gave
way to birch as the dominant carr woodland also with development of an acidophilous
bog. The latter is in accord with the data from the Somerset Levels but is not at the same
time.
v) Erosion and disturbance of the upper peat occurred due to inundation by brackish water or
freshwater ponded back in the river systems due to increasing RSL (relative sea-level)
after c. 2550 cal BC. A date from the upper peat/sediment contact will produce an age for
this event. This will however, be a terminus post-quem because of the erosion of the top
of the peat and/or disturbance which has occurred.
vi) Change to minerogenic sediment accumulation and presence of halophytic (salt tolerant)
plants in pollen zone AWK:4 and 5 marks the inundation of the freshwater fen carr
communities and replacement by progressively important salt-marsh habitats. The
openness of the salt-marsh extended the pollen catchment which resulted in greater long
distance and regional pollen input and also from freshwater fluvial discharges. The dating
of the upper part of this profile is at present conjectural (Late Bronze Age–Iron Age–
Romano-British?) but it seems that oak and hazel woodland remained the primary
woodland elements throughout the late-prehistoric period. Secale cereale (rye) is present
and is typically a Roman cereal crop but is now also occasionally found in Iron Age
contexts (Chambers 1989) and even in Late Bronze Age ones (Grieg 1991). There are
unfortunately no further pollen indicators.
vii) Ephemeral phases of salt-marsh drying (for example AWK:5-a) possibly caused short
lived phases of small negative sea-level tendency. This allowed partial pedogenesis and or
organic formation in the grey salt-marsh sediments.
Diatom analysis
by Nigel G. Cameron
The results of diatom analysis are presented in Figure 8. On the diagram the samples are identified
by site name, AWK: Awkley Lane; HAL: Hallen Marsh; VIM: Vimpenny’s Lane. The subsequent
numbers in the sample names refer to depths above Ordnance Datum. Diatom species salinity
preferences are classified using the halobian groups of Hustedt (1957, 199):
1. Polyhalobous >30 g l-1

2. Mesohalobous 0.2-30g l-1
3. Halophilous – optimum in slightly brackish water
4. Oligohalobous indifferent – optimum in freshwater but tolerant of slightly brackish water
5. Halophobous – exclusively freshwater
6. Unknown – diatoms of unknown salinity preference
The principle source used for species salinity classification was Denys (1992).
30
Figure 8. Diatom analysis; all sites
31
+4.42m OD (context 107; within pollen zone AWK5): Of a series of samples assessed in 1994, and a further six
examined in 2000, only one was countable. The diatom assemblage of this sample is composed of 55% marine species
and almost 35% brackish water taxa. Freshwater diatoms comprise just over 5% of the assemblage and approximately
5% of the species were of unknown salinity preference. The marine planktonic diatom Paralia sulcata was dominant,
comprising almost 40% of the assemblage and the mesohalobous diatom Nitzschia navicularis accounted for almost
30% of the assemblage. Both of these taxa are typical of estuarine sediment assemblages. The latter species is not
common in the other samples analysed for diatoms, although it is common in estuarine sediments analysed in other
geoarchaeological investigations in the Severn Estuary (eg. Cameron 1997). Vos & de Wolf (1993) consider Nitzschia
navicularis to be part of the marine-brackish epipelic (motile mud-dwelling) diatom community and the species is an
autochthonous component of the fossil assemblage of tidal mudflats. The context (107) description for this sample
indicates a grey clay, with the comment that the sediment may be a buried topsoil. The diatom assemblage represents a
highly saline, tidal, diatom habitat.
Molluscs
by Michael J. Allen
Preservation of shells in these deposits is known to be poor and patchy (cf. Allen & Ritchie 2000;
Bell & Johnson in Lawler et al. 1992). Nevertheless, the significance of determining brackish
water, freshwater, or terrestrial facies from limited contexts was high. Contiguous samples were
taken through the main sedimentary sequences (Awkley Lane, Vimpenny’s Lane, Awkley
Interface), as well as from ditches on the occupation sites (Hallen and Northwick). In each case
large samples (nominally 1750 g) were processed after the removal of a sub-sample suitable for the
measurement of magnetic susceptibility. Samples were disaggregated in water and washed through
a nest of sieves down to 0.5 mm (cf. Evans 1972). The results are presented in Table 2 and the
nomenclature follows Kerney (1999).
Shells only survive in a limited portion of this 4 m section, essentially that coincidental with the
two putative buried soils or temporary stabilisation horizons indicating and confirming temporary
stasis. The assemblages are predominately dominated by fresh and brackish water species.
Insect remains
by Mark Robinson
Insects were analysed from the Neolithic peats at the base of the alluvial sequence at Awkley Lane.
They have been sampled in sequence, with three groups of three samples from within the Middle
Wentlooge peat. Sub-samples from all of them were assessed for insect remains, which were found
to be sparse and poorly preserved. It was therefore decided to limit analysis to 2 kg of the middle
sample from each group. Each sample was subjected to paraffin flotation to recover insect remains.
The flots were caught on a 0.2 mm sieve, washed in detergent and sorted under a binocular
microscope. Identifiable insect remains were absent from context 114 (2.09 m OD). The results
from the other two samples are given in Table 3. Nomenclature for Coleoptera follows Kloet and
Hincks (1977).
The low concentration of insect remains makes detailed interpretation difficult, although the
majority of the taxa are characteristic of wet habitats. Just under half the beetles in context 118
(1.49 m OD), the lowest peat layer, were aquatic, Ochthebius minimus being the most numerous.
Aquatic larvae of Trichoptera (caddis fly) and Chironomidae (midges) were also present. There
were fewer aquatic insects in context 121 (2.49 m OD), the top peat layer, but they included the
water beetle Hydrobius fuscipes. The aquatic insects are more suggestive of the faunas of small
pools and swamp areas on the peat rather than insects introduced in floodwater.
32
Table 2. Mollusc data from Awkley Lane (note: only samples with shells tabulated: totals exclude C. acicula)
Sample 2 10 3 4 5 6 7 9 11 12 14 15 17 18 20 21 23 25 8 1 16 22 13 26 19 24
Sample 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
Context 112 110 110 110 110 110 110 110 110 110 110 110 110 110 109 108 108 107 106 105 105 105 105 105 105 104
Depth (mOD) 2.78- 2.88- 2.98- 3.08- 3.18- 3.28- 3.38- 3.48- 3.58- 3.69- 3.79- 3.89- 3.99- 4.09- 4.19- 4.29- 4.39- 4.49- 4.59- 4.68- 4.78- 4.88- 4.98- 5.08- 5.18- 5.28-
2.68 2.78 2.88 2.98 3.08 2.18 3.28 3.38 3.48 3.59 3.69 3.79 3.89 3.99 4.09 4.19 4.29 4.39 4.49 4.58 4.68 4.78 4.88 4.90 5.08 5.18
Wt (g) 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1500 1400 1750 1750 1750 1750 1500
Succinea putris (Linnaeus) - - - - - - - - - - - - - - - - - - 4 - - - - - - -
Succinea cf putris/Oxyloma - 1 - - - - - - - - - - - - - - - - - 5 - - 1 - - -
Oxyloma pfeifferi (Rossmässler). - - - - - - - - - - - - - - 1 - - - 5 - - - - - - -
cf Catinaella arenaria/S. oblonga - - - - - - - - - - - - - - - - - - 1 - - - - - - -
Vallonia costata (Müller) - - - - - - - - - - - - - - 1 (b) - - - - - - - - - - -
Cepaea/Arianta spp. - - - - - - - - - - - - - - 1 - - - - - - - - - - -
Land mollusc Taxa 0 1 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 3 1 0 0 1 0 0 0
LAND MOLLUSC TOTAL 0 1 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 10 5 0 0 1 0 0 0
Bithynia tentaculata (Linnaeus) - - - - - - - - - - - - - - 15 12 3 - - 2 - - - - - -
Bithynia tentaculata operculum - - - - - - - - - - - - - - (16) (26) (7) - (6) (1) - - (2) - - -
Aplexa hypnorum (Linnaeus) - - - - - - - - - - - - - - - 1 - - - - - - - - - -
Lymnaea truncatula (Müller) - - - - - - - - - - - - - - 9 13 - 17 5 - - - - - - -
Lymnaea peregra (Müller) - - - - - - - - - - - - - - 11 6 - - 1 - - - - - - -
Lymnaea sp. - - - - - - - - - - - - - - - 4 - - - - - - - - - -
Anisus leucostoma (Millet) - - - - - - - - - - - - - - 10 5 - 1 2 1 - - - - - -
Anisus vortex (Linnaeus) - - - - - - - - - - - - - - 17 1 - - - - - - - - - -
Gyraulus albus (Müller) - - - - - - - - - - - - - - 5 1 - - - - - 1 - - - -
Hippeutis complanatus (Linnaeus) - - - - - - - - - - - - - - 7 1 - - - - - - - - - -
Pisidium (valves) y 2
Pisidium personatum Malm - - - - - - - - - - - - - - 1 - - - - - - - - - - -
Pisidium obtusale (Lamarck) - - - - - - - - - - - - - - 2 - - - - - - - - - - -
Pisidium pulchellum Jenyns - - - - - - - - - - - - - - 3 1 - - - - - - - - - -
Hydrobia ventrosa (Montagu) - - - - - - - - - - - - - - 1 - - - 2 1 3 - - - - -
Hydrobia ulvae (Pennant) - - - - - - - - - - - - - - - - - - 1 1 1 - - - - -
Aquatic Taxa 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 9 1 2 4 4 2 1 1 0 0 0
FRESHWATER TOTAL 0 0 0 0 0 0 0 0 0 0 0 0 0 0 79 58 7 18 14 3 0 1 2 0 0 0
MARINE TOTAL 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 1 0 0 3 2 4 0 0 0 0 0
Ostracods (valves) - - - - - - - - - - - - - - 54 - - 19 17 - - - - - - -
Marine Gastropod - - - - - - - - - - - - - - - - - - - 1 - - - - - -
TOTAL 0 1 0 0 0 0 0 0 0 0 0 0 0 0 86 59 7 18 27 10 0 1 3 0 0 0
33
Both samples contained ground beetles of marsh or peaty habitats, for example Elephrus cf.
cupreus, which occurs at the edge of standing water where some vegetation is present and in forest
swamps, and Agonum obscurum, which occurs in damp deciduous forests and densely vegetated
marshes (Lindroth 1974, 33, 82). Evidence for carr (woodland growing on peat) was given by a
scolytid beetle, possibly Lymantor coryli which lives under the bark of Alnus glutinosa (alder) and
Corylus avellana (hazel), in context 118 and Rhynchaenus sp. (not pratensis), a weevil which feeds
on tree leaves, in context 121. However, some open meadow-like areas were suggested by a couple
of examples of the weevil Apion sp., which mostly feed in clovers and vetches, context 118.
Table 3. Coleoptera and other insects from Awkley Lane
Min. no. indiv

Context 118 114 121
Sample 1036 1032 1030
OD 1.49 2.09 2.49
Coleoptra
Elephrus cf. cupreus 1 - -
Pterostichus diligens - - 1
Agonum obscurum - - 1
Hydrobius fuscipes - - 1
Helochares or Enochrus sp. - - 1
Ochthebius minimus 1 - 1
O. cf. minimus 5 - -
Hydraena sp. (not testacea) 1 - -
Ptenidium sp. 1 - 2
Olophrum cf. fuscum - - 1
Stenus sp. 1 - -
Lathrobium sp. (not rufipenne) 1 - 2
Medon sp. - - 1
Philonthus sp. - - 1
Aleocharinac indet 1 - -
Pselaphidae indet. 1 - -
Corylophus cassidoides 1 - -
Apion sp. (not Malvaceae, Urtica or Salix-dependent) - - 2
Rhynchaenus sp.(not pratensis) - - 1
cf. Lymantor coryli 1 - -
Other insects
Trichoptera indet. – larva 1 - -
Chironomidae indet – larva 1 - -
Diptera indet – puparium - - 1
Waterlogged Plant Macrofossils

by Alan J. Clapham
Seven samples were examined from the Middle Wentlooge peats; three contiguous samples from
the dark reddish brown basal wood peat (context 118), three contiguous samples from peat contexts
115 (1.99 m OD), 120 (2.04 m OD), 114 (2.09 m OD), and a final sample from wood peat context
121 (2.49 m OD). Waterlogged plant macrofossils were, in most cases, well preserved (Table 4)
Samples from the three zones of peat were analysed. As the volume of each sample was the same it
is possible to produce an equivalent diagram for seeds as for pollen, this is presented in Figure 7. In
order to aid the discussion of the waterlogged plant remains the results are presented as groups
corresponding to the peat horizons which they represent.
Context 118 (1.44m OD, 1.48m OD and 1.54m OD; polen zone AWK:1) The dominant tree remains within the samples
were of alder (Alnus glutinosa), and included fruits, female cones, and male catkin fragments and scales. In some cases
it was possible to detect the presence of anthers within the latter. Alder buds and budscales were also identified from
34
1.44 m OD. Hazel (Corylus avellana) nutshell fragments were identified from the lower two samples, including one
charred. Two immature small-leaved lime (Tilia cordata) fruits were identified.
Other species identified included those which indicate wet/marshy conditions and prefer to have their feet in water.
These include celery-leaved buttercup (Ranunculus sceleratus), creeping yellow-cress (Rorippa sylvestris), alternate-
leaved golden saxifrage (Chrysosplenium alternifolium), gipsywort (Lycopus europaeus), marestail (Hippuris vulgaris),
hemp agrimony (Eupatorium cannabinum), water-plantain, rush (Juncus sp.), great fen-sedge (Cladium mariscus),
sedges (Carex spp.), and reedmace (Typha sp.).
A drier habitat is indicated by the presence of buttercups (Ranunculus subgenus Ranunculus), whilst a scrubbier
element is represented by the presence of stinging nettle (Urtica dioica), bramble (Rubus Section 2 Glandulosus) and
woody nightshade (Solanum dulcamara). Charcoal fragments were found in each of the samples.
As the samples contain the same plant taxa representing similar habitats it is possible to treat the
three samples as one with regards to environmental reconstruction. The dominance of alder
confirms the pollen evidence for the dominance on-site of alder carr within which, on the drier
areas, hazel was present. The undergrowth consists of a wetland/marshy flora with tall herbs such
as reedmace, hemp agrimony, and great fen sedge associated with stinging nettle. Bramble and
woody nightshade can be found scrambling over this vegetation. Wet/marshy species grew beneath
while, in the open water, which may be present as pools within the alder carr, marestail can be
found growing along with waterfleas, the eggs of which were common in sample 1.54 m OD. The
occurrence of fruits of small-leaved lime suggests that drier land is not too far away. The presence
of the charcoal is difficult to interpret, it may represent in situ burning, whether natural or
anthropogenic, or may represent an allocthanous element. Caddis fly larvae were also recorded
from sample at 1.44 m OD indicating an aquatic environment.
Context 115 (2.06 m OD; pollen zone AWK:2) Context 115 was a woody peat with common monocotyledonous roots.
There were no tree remains identified from this sample, most of the species recovered are representative of wetland
habitats and in this case it most likely indicates the presence of reedswamp/fen. The sample was dominated by seeds of
the square-stalked St John’s-wort (Hypericum tetrapterum), watermint (Mentha aquatica), and hemp agrimony cypsela
fragments. Rhizome fragments of reed were also very common as were small grass caryopses. Other species include
buttercup, stinging nettle, orache (Atriplex sp.), bramble, woody nightshade, rush (Juncus sp.), and sedge (Carex sp.)
along with reedmace (Typha sp.). Charcoal fragments were also recorded, the origin of which is difficult to determine.
The presence of one seed of elder (Sambucus nigra) was also identified. Water flea eggs were present.
Context 120 (2.14 m OD; pollen zone AWK:2) Context 120 was described in the field as a black peat (most likely
detrital in origin). This sample produced a radiocarbon date of 3640–3380 cal BC. Again there were no tree species
present. The dominant species in this sample were, stinging nettle, fragments of hemp agrimony fruits, and fragments
of Phragmites rhizomes. Other species included celery-leaved buttercup, square-stalked St John’s-wort (although not as
common as in context 115), bramble, woody nightshade, gipsywort, watermint, sedges, small grasses, and reedmace.
waterflea eggs were also present which suggests some pools of open water. This represents a tall herb fen/reedswamp.
Context 114 (2.09m OD, pollen zone AWK:2) Of the three samples from this peaty horizon, this was the richest sample
in terms of numbers and taxa of waterlogged plant remains and again was dominated by taxa which are found growing
in reedswamps. The sample was dominated by the presence of fragments of hemp agrimony fruits and rhizome
fragments of Phragmites. Other common species included celery-leaved buttercup and watermint. A species which is
often found in fens, common meadow-rue (Thalictrum flavum) was also recorded. Other species included, stinging
nettle, red goosefoot (Chenopodium rubrum), orache (Atriplex sp.), golden dock (Rumex maritimus), square-stalked St
John’s-wort, bramble, hairy willowherb (Epilobium hirsutum), woody nightshade, gipsywort, sedges, small grasses,
and reedmace.
In this sample the aquatic habitat was represented by the presence of rigid hornwort (Ceratophyllum demersum var
apiculatum), fool’s watercress (Apium nodiflorum), and narrow-leaved water-plantain (Alisma lanceolatum).
Again this sample can be said to indicate the presence of reedswamp/tall herb fen. The presence of a more aquatic
environment is more prevalent in this sample than in those from contexts 1145 and 115.
These three samples discussed indicate the presence of a tall herb fen/marginal
aquatics/reedswamp, with no arboreal elements apart from the find of one elder seed. The habitat
35
Table 4. The waterlogged plant remains from the wood and fen peat of the Middle Wentlooge
Context 118 118 118 115 120 114 121

Sample 1037 1036 1035 1034 1033 1032 1030
O.D. 1.46-1.41 1.51-1.46 1.56-1.51 2.01-1.96 2.06-2.01 2.11-2.06 2.51-2.46
(1.44) (1.49) (1.54) (1.99) (2.04) (2.09) (2.49)
14C yrs BP(uncalibrated) 5603±50 4745±50 4286±55
3
Volume processed cm 300 300 300 300 300 300 300
Chara oogonia - - - 2 - - -
Cenococcum geophilum 41 15 33 - - - 133
Musci common common occasional - - occasional -
Ceratophyllum demersum var apiculatum - - - - - 2+14f -
Ranunculus sceleratus - 1 - - 7 63+60(½s)+1 -
8f
Ranunculus subgenus Ranunculus 2+7f 21+35f 4+8f 2f - - -
Thalictrum flavum - - - - - 3f -
Urtica dioica 2 - 1+2f 18 301 18 -
Alnus glutinosa fruits 33+68f 35+8f 63+42f - - - -
Alnus glutinosa female cones 14+9f 18+19f 15 - - -
Alnus glutinosa male catkin frags (stem) 28 100+ 26 - - - -
Alnus glutinosa male catkin scales 285 - - - - - -
Alnus glutinosa budscales +buds 5+2 - - - - - -
Corylus avellana nutshell fragments 19+1char 3 - - - - -
Tilia cordata immature fruits - 2 - - - - -
Rorippa sylvestris 1 - - - - - -
Chenopodium rubrum - - -- - - 1 -
Atriplex sp. - - - 1 - - -
Rumex maritimus valves - - - - - 7 -
Rumex maritimus nutlets - - - - - 1+1f -
Hypericum tetrapterum - - - 335 12 2 -
Chyrsosplenium alternifolium 144+20f 54 - - - - -
Rubus Section 2 Glandulosus - 2f 3+1f 5f 2+5f 5f -
Epilobium hirsutum - - - - - 2f -
Apium nodiflorum - - - - - 1 -
Solanum dulcamara - 2f - 1+8f 4f 3+11f -
Lycopus europaeus 9 - - - 1 14 -
Hippuris vulgaris 1 - -- 149+6f 4 145+46f -
Mentha aquatica/arvensis 1 - - 1 - - -
Sambucus nigra 1 - - - - - -
Eupatorium cannabinum 8f 3f - 257f 1000f+ 1000f+ -
Alisma lanceolatum seedcase - - - - - 1+1f -
Alisma sp. embryo - - - - - 1 -
Alisma sp. seedcase 1 - - - - - -
Juncus sp. 2 - - 1 - - -
Cladium mariscus 1 - - - - - -
Carex sp. (lenticular) 105+127f 75+21f 1+6f 1 - - -
Carex sp. (trigonous) 16+17f 7+10f 2f - 13f 36f -
Carex sp. utricule fragments - - - - - 16 -
Typha latifolia/angustifolia 6 1 - - - - -
Sphagnum spp. leaves - - - - - - 1000+
Betula sp. seed - - - - - - 5+2f
Phragmites australis rhizome fragments - - - 100+ 100+ 100+ -
Small Poaceae - - - 170 13 12 -
Typha latifolia/angustifolia - - - 22 7 55 -
Leaf fragments 2 - - - - - -
Charcoal 83f 14f 2f 36f - - -
Culm node - 1 - - - - -
Budscales 1 - - - - - -
Bud bases - 1 8 - - - -
Leaf abscission pads 39 - - - - - 2
Miscellaneous anther 1 1 - - - - -
Worm cocoons - - - - - - 16
was dominated by the tall herbs such as reeds, reedmace, stinging nettle and hemp agrimony, which
supported the scrambling woody nightshade and bramble. The undergrowth supported plants such
as gipsywort, celery-leaved buttercup, square-stalked St John’s-wort, watermint, and golden dock.
It can perhaps be said that the samples show a steady progression from carr to fen/reedswamp. In
context 114, the presence of aquatic species such as rigid hornwort, narrow-leaved water-plantain
and fools watercress may indicate a rising in the watertable creating more pools of open water.
36
Context 121 (2.41 m OD, pollen zone AWK2-c) This sample contained large chunks of wood and was very humified, in
the field it was described as a black peat with up to 20% wood inclusions. This context produced a radiocarbon date of
3090–2700 cal BC (NZA-1258, 4286±55 BP). This sample was the middle sample from the uppermost peat horizon.
There were very few plant remains, the sample being dominated by large quantities of Sphagnum moss leaves
(Sphagnum spp.) The only other remains were birch (Betula sp.) fruits. Birch bark and moss stems, presumably of
Sphagnum were also present in large quantities. One noticeable feature of this sample was the sclerotia of Cenococcum
geophilum, although numerous the striking feature about them was that some were up to 3 mm in diameter which is
large for fungal sclerotia.
This sample represents a Sphagnum bog, on which birch trees were growing. The small numbers of
birch fruits recovered suggest that the trees were growing some distance from the sample, but,
conversely, the presence of the birch bark suggests that they may be more local.
Overall, there seems to be an agreement between the different sources of environmental evidence.
The lack of beetle and other insect remains with the samples makes it difficult to carry out an
accurate comparison, which is not so with the pollen record. Altogether the different environmental
evidence suggests that there is a change in habitat through time, with alder carr giving way to tall-
herb fen/reedswamp, which finally gives way to a birch, Sphagnum dominated woodland.The lack
of archaeological evidence from the area means that it is difficult to determine if these changes
were entirely natural, anthropogenic, or a combination of both.
2. Vimpennys Lane
The sequence
Two trenches were excavated off Vimpenny’s Lane (ST 5561 8211) approximately 1.5 km to the
west of Easter Compton (Fig. 9). Apart from providing a sedimentary sequence these trenches were
excavated specifically to investigate a lens of freshwater peat at 3.5 m OD (2.7 m below ground
level) previously identified by Lawler et al. (1992, 72). The objective was to provide a full analysis
of the sedimentary sequence including palaeo-magnetic dating. A full range of samples including
mollusc columns, pollen monoliths, samples for particle size analysis, was taken from trench 2.
Both trenches were c.10 u 10 m, and were excavated to a maximum depth of 3 m. The north-east
side of each trench was stepped to give a running section. The profile in trench 2 was recorded and
sampled (Fig. 9) and an attempt made to obtain palaeo-magnetic dates from the fine-grained
sequence. An auger hole in the base of trench 1, using a dutch auger, extended the profile to 5.8 m
depth (-0.15 m OD), but did not enable sample retrieval. The sediments were described by the
archaeologists in the field and these subsequent descriptions augmented by examination of
monoliths in the laboratory (see descriptions).
The sequences in trenches 1 and 2 were virtually identical. A series of low-energy greyish brown
silty clay alluvial deposits 1.25 m thick (to 4.2 m OD) with low or suppressed magnetic
susceptibility (11 to14 SIu10-8Kg-1) rising in the lower 0.5 m (to 16 SIu10-8Kg-1), were present
below the present day soil. These overlay a 0.2 m thick, very dark grey silty clay putative
stabilisation horizon (layer 207, 4.19 m OD), with slightly suppressed rather than enhanced
magnetic susceptibility, developed over a greyish brown silty clay similar to that above. The
undulating disturbance patterns seen in section and irregular nature of this layer as seen elsewhere
(eg, GGAT layer 223, see Fig. 9), may suggest turbulent water erosion, flooding, and mixing.
Below c. 3.90m are a further 0.95m of grey and dark grey silty clay alluvial deposits which are
distinctly laminated towards the base. This zone of very fine slightly siltier laminations appear to be
remnants of depositional strata, probably flood couplets (very fine layers of alternating silt and silty
37
clay, each pair relating to a single flood event), resulting from over-bank deposition during periods
of flood of MHW, or possibly tidal movement. This unit overlies a narrow band of dark grey/black,
highly humified, silty peat (layer 211 at 3.04 m O.D.), 0.05m thick which overlies a heavily rooted
silty sand/silt loam grading into 1.15m of light grey silty clay alluvium. This in turn overlies a
second band of peat (1.74 m O.D.), 0.4 m, thick and only recorded in a auger hole through the base
of Trench 1. The final 1.35 m at the base of the auger hole comprise a series of grey silty clay and
grey sands.

0-0.15 5.80-5.65 200 Dark greyish brown (10YR 4/2) silty loam, occasional pebbles, rare
small pebbles.
0.15-0.25 5.65-5.55 201 Brown (7.5YR 5/6) stonefree silty clay, with thin lenses of
lighter grey matrix.
0.25-0.55 5.55-5.25 202 Light brown (7.5YR 6/4) stonefree masive (unstructured) Upper Alluvium
silty clay, common mineral (iron and mangenese) staining,
occasional gritiness caused by concretions of iron and
oxyhydroxides, substantial rootlets (Phragmites), and stems
marked with orange traces, diffuse boundary.
0.55-0.80 5.25-5.00 203 Greyish brown (10YR 5/2) stonefree silty clay loam, Upper Alluvium
occasional mineral staining.
0.80-0.90 5.00-4.90 204 Grey (5Y 6/1) stonefree silty clay loam; 30mm zone of grey Upper Alluvium
silty clay at top of band of fine gleyed material.
0.90-1.30 4.90-4.50 205 Brown (10YR 5/3) stonefree silty clay, occasional mineral Upper Alluvium
(iron) staining.
1.30-1.60 4.50-4.20 206 Dark greyish brown (2.5Y 4/2) stonefree silty clay, common Upper Alluvium
mineral (inc manganese) staining, very rare charcoal, some
relict fine (1mm) laminations towards the base of this unit.
1.60-1.75 4.20-4.05 207 Very dark grey (10YR 3/1) stonefree slightly silty clay, humic/ ?Buried soil/
organic layer [but not so in sampled core] frequent mineral stabilisation horizon
(iron) staining, coarse matrix, slightly gritty due to the
presence of sand element and a number of shells (Hydrobia). Some
‘swirling’ or undulating disturbance pattern seen in section. Distinctly
less organic than the same horizon recorded in evaluation trench,
described as with rare snail shell fragments and partly
decayed reed/grass stems and rootlets.
1.75-1.90 4.05-3.90 208 Grey (2.5Y N/5) stonefree silty clay, mottled, common Grey Alluvium
mineral staining and occasional vegetable matter and r
rootleting indicated by orange-brown staining, some
ubvertical zones and localised patches.
1.90-2.05 3.90-3.75 209 Grey (10YR 5/1) stonefree silty clay, occasional mineral Grey Alluvium
staining and very occasional vegetable matter marked by
orange brown staining.
2.05-2.65 3.75-3.15 210 Grey (2.5Y 5/0) silty clay, occasional vegetable matter, Grey Alluvium
eroded peat and some intact root stems pass through this
unit, diffuse silty rhythmical laminations towards base with
alternating very fine laminae of silt and silty clay – distinct
flood couplets in lower 0.2m (2.45-2.65m OD), clear to
gradual wavy boundary
2.65-2.75 3.15-3.05 211 Dark grey/black (10YR 3/1–2/1) highly humified silty detrital Phragmites
peat with Phragmites, mottled with light grey silty clay ?backfen peat
patches, gradual smooth boundary and peat grades
downwards into heavily rootleted, light grey silty sand
2.75+ 3.05+ 212 Grey (10YR 5/1) stonefree silty sand/silt loam, heavily Grey Alluvium
rootleted, common vegetable matter.
Auger hole through trench 1
3.10-4.10 2.60-1.50 Light grey (7.5YR N7/0) silty clay, common plant matter. Probably the same
layer as 210.
4.10-4.25 1.50-1.45 Very dark brown (10YR 2/2) peat. Peat
4.25-4.50 1.45-1.20 Black (10YR 2/1) peat. Peat
4.50-4.55 1.20-1.15 Dark grey ((10YR 4/1) silty clay, common plant matter. Alluvium
4.55-4.90 1.15-0.80 Grey (2.5Y N5/0) sand, occasional vegetable matter. Alluvium
4.90-5.80 0.80- -0.10 Grey (10YR 5/1) silty sand. Alluvium
5.80-5.85 -0.10- -0.15 Grey (7.5YR N5/0) silty sand clay. Alluvium
5.85+ -0.15+ Grey (2.5Y N5/0) sand. Alluvium
In general, the sequence was the same as that identified by Lawler et al. (1992, 72–3). The upper
band of silty peat at 3.14 m OD was dated to 2900-2620 cal BC (4182±55 BP, NZA 12527) and
appears to correspond with that found in evaluation (trial pit G009, context 227 at 2.8 m OD), dated
38
WA Vimpenny Tr.1 WA Vimpenny Tr.2 GGAT G009
Mollusc & soil
column
Pollen Zones
201
202
203
5 m O.D. 2b 204
205
2
206
2a 207
4 m O.D.
208
209
1b
0 10 20 30
x10-8 SI/kg
210
1
4182+_ 55BP
211
3 m O.D.
1a 212 4420+_ 90BP
2 m O.D.
Insect sample
1 m O.D. Plant macro fossils sample
C14 sample
Alluvial silty clays (brownish red hue)
?Stabilisation horizon
Sandy alluvium
0 m O.D.
0 1m
Figure 9. Vimpenny's Lane Wentlooge Formation profiles
39
to 3350–2900 (4420±90 BP, GU3121) (Lawler et al. 1992, 115). This would categorise it as a
fourth peat as defined by Lawler et al. (1992, 8). The lower band of peat would also, by virtue of its
relative height OD, fall into this fourth peat category and may represent an area where the fourth
peat development was interrupted by inundation, with further development resuming at a later date.
The sequence represents low-energy estuarine deposition with two obvious episodes of stasis and
peat formation and one possible non-organic stasis or stabilisation horizon. The base of the sampled
sequence is dated to about 3000 cal BC, and no other chronological references can be provided for
the overlying 2.7 m of minerogenic sediments.
Chronology
As most of the exposed sequence was minerogenic there was little opportunity for absolute dating.
Palaeo-magnetic intensities and directions were very scattered within the minerogenic sequence.
Demagnetistaion revealed stable remnance, suggesting that the scatter in vectors was due to
inhomogeneous magenetisation composed of two stable components of differing age: the primary
depositional vector overprinted by secondary chemical magnetisation arising during diagenesis. The
palaeo-magnetic record was, therefore, unsuitable for absolute dating.
Only one horizon, a well humified Phragmites peat band 0.1 m thick, was suitable for radiocarbon
dating. A result of 4420±90 BP was obtained from the evaluation sequence, and a slightly latter
determination of 4182±55 BP from the analysed sequence (Table 5). These give calibrated results at
the end of the 4th–beginning of the 3rd millennium BC.
Table 5. Radiocarbon determinations and calibrated results from the detrital peat at
Vimpenny’s Lane
context context type depth OD material lab no GC13‰ result BP calib

Vimpenny’s Lane
211 ?OLS 3.14m plant matter NZA 12527 -25.98 4182± 55 2900-2620
GGAT 2.80m GU-3121 -27.2 4420± 90 3350-2900*
* GGAT 1992
As with the Awkley Lane sequence the full minerogenic sequence, excepting the upper 0.5 m, was
sampled in monolith tins to facilitate subsampling for pollen and diatoms. Again gaps in this
sequence were filled by sub-sampling disturbed samples taken for land snails (Fig. 9). A series of
19 contiguous samples was taken for snails from which sub-samples were removed for grain size
analysis. Bulk samples were was taken for archaeo-magnetic dating through the minerogenic
deposits, but these failed to produce consistent results.
Textural features of the upper sedimentary sequence

by J.R.L. Allen
The 19 samples from the upper minerogenic sequence were analysed and described. The triangular
graph (Fig. 10) shows the scatter of values of the clay-silt-sand ratio. Mean values for some of the
finer-grained samples should be treated with caution. Although very small in quantity,
40
Plate 3. Vimpenny's Lane excavation showing the fully excavated and exposed upper and
middle Wentlooge sequence
Vimpenny's Lane Clay (100%)
1108 1117
1100 1107
50% 50%
y
Cla
Silt (100%) Sand 50% Sand (100%)
Figure 10. Vimpenny's Lane; particle size triangle
41
concretionary material was present which could not be dispersed. This has exaggerated the coarse
tails of the distributions, inflating values for the mean.
The triangular graph, in addition to the partical size proportions, shows that the samples divide
between a lower, coarser (layers 209–206 at 3.81–4.61m OD) and an upper, finer (layers 205–200,
4.61–5.75m OD) sub-set. Taken together, however, the two subsets define a single trend similar to
that recognised at Hallen Marsh (below).
The sediments of the coarser sub-set are texturally similar to all but the finer deposits at Hallen
Marsh and similar to deposits typical of salt-marshes in the Severn Estuary Levels. Given the
resolution permitted by the chosen sampling interval, the particle size proportions reveals no clear
stratigraphic trends in this sub-set. One sample from the middle of layer 206 at 4.45 m OD is,
however, noteworthy for its unusually coarse mode.
There is a major textural change from the coarser to the finer sub-set between samples within layer
206 at 4.51 m OD. The latter sub-set is exceptionally fine-grained for Severn Estuary Levels salt-
marshes, although not inconsistent with the interior position of the site. No stratigraphic trends
within it can be recognised with any great confidence, but it may be noted that samples from the top
of 203–201 at the top of the sequence (5.25–5.55 m OD) yield modes and curve shapes suggesting
that these sediments are coarser than those below and above. It is possible that a weak coarsening-
up trend is followed by a weak fining-up sequence.
The stratigraphically lower, coarser sub-set suggests a stable marsh that is comparatively near the
marsh edge and/or major creeks. A sudden, major change in conditions is suggested by the sharp
textural break within layer 206 above sample at 4.41–4.51 m OD and the appearance of the finer-
grained subset. The new environment appears to involve a distant marsh edge and/or remote major
creeks (?sudden diminution in creek density). These conditions were either stable or, possibly,
included an earlier episode of slight marsh-edge retreat and/or positive sea-level tendency, followed
by a later episode of slight marsh-edge advance and/or negative sea-level tendency.
Pollen analysis
by Robert G. Scaife
Pollen extraction procedures used at Vimpenny’s Lane followed those detailed for the analysis of
Awkley Lane. Samples of 5 ml were used and absolute pollen frequencies were calculated.
The Vimpenny’s profile can be divided into two principal pollen assemblage zones (Fig. 11) which
relate to regional vegetation. A number of local pollen assemblage sub-zones are also defined
which relate to changes in the on-site vegetation communities. These are also manifested in the
changing stratigraphy of the site. These pollen zones are defined and characterised as follows.
VIMP:1 (212, 211, 210, 209) 2.90–3.98 m OD, Pinus-Quercus-Tilia-Corylus avellana type
This zone is characterised by highest values of Quercus (to 60%), Tilia (to 6%), and Corylus avellana type (to 28%).
Also present are Pinus (to 10%) and sporadic/non-continuous occurrences of Betula, Picea, Ulmus, and Fraxinus.
Herbs (30–40%) are dominated by Chenopodiaceae (to 24% at the top of VIMP 1.a.) and Poaceae (peak to 30%) with
sporadic occurrences of other taxa including halophytes – Armeria ‘A’ and ‘B’ line. Marsh taxa include Alnus (10%),
Cyperaceae (to 18%), and some aquatic types (?Potamogeton) and reed swamp taxa (Typha and Sparganium). There
are substantial numbers of spores comprising monolete Dryopteris type (20%), Pteridium aquilinum (peaking to 40%),
and Polypodium vulgare (basal level = 22%). Of note are higher values/numbers of pre-Quaternary palynomorphs in
VIMP 1.a and Sphagnum incoming at the top of the zone from 3.58 m OD.
42
Two local pollen assemblage sub-zones have been recognised. These are characterised as follows.
VIM:1a, (212, 211) 2.90–3.16 m OD: A radiocarbon date of 2900–2620 cal BC (NZA-12527, 4182± 55BP) from a
detrital organic horizon (211) delimits the top of this sub-zone. Arboreal pollen is defined as above but with some
reductions in Quercus and Pinus relating to the thin peat horizon, that is, reflecting increased percentages of
Chenopodiaceae (expanding to 26%). Derived pre-Quaternary palynomorphs are more important. Absolute pollen
frequencies increase from typically low values in these sediments to >100,000 grains/ml.
VIMP:1b, (210, 210) 3.62–3.98 m OD: In this sub-zone, there is an expansion of acidophilous taxa including
Ericaceae (Erica and Calluna) and Sphagnum (to 15%). Some herbs are also incoming in this sub-zone-Plantago
lanceolata, Lactucoideae, and Plantago coronopus type.
VIMP:2: (108, 207, 205, 204, 203, 202) 3.98-5.42 m OD, Chenopodiaceae-Herbs-Pteridium aquilinum.
This zone is delimited by a sharp reduction in tree pollen at 184cm with marked declines in Quercus (down to 20–
25%), Pinus (2–3%) and Tilia (to occasional occurrences). Corylus avellana type remains consistent from the middle of
zone 1 (zone 1b) at c. 20%. Herbs become dominant (to 75–80%) with Poaceae (to 50%), Chenopodiaceae (25%) and
increased herb diversity, including Plantago lanceolata, Cereal type and Asteraceae types. Incoming of Ericaceae
(Erica and Calluna) in zone 1b continues throughout zone 2. Within the marsh/aquatic category, Alnus remains but
with decrease in diversity. Ruppia maritima is present. Of note is a peak of Potamogeton type (cf. Triglochin) along
with the halophytes Plantago maritima, Aster type and Chenopodiaceae (zone 2b). Spores become more important with
increasing percentages of Pteridium aquilinum (to 55%), Polypodium vulgare declining from values of zone along with
other spores showing some further expansion after zone 2a.
Two local pollen assemblage subzones have been recognised.
VIMP:2a, (208, 207 lower 206) 3.98–4.26 m OD: Immediately above the lower zone boundary (VIMP:1/2) at 3.94 m
OD absolute pollen frequencies attain highest values (to c. 200,000 grains/ml) after which values decrease sharply to
typical site values of 5–10,000 grains/ml. In this sub-zone there is a peak of Poaceae (highest values of 50%) and
Chenopodiaceae (27%). There are, however, reductions in Pinus (to absence), Quercus (to 15%) and especially spores
– Pteridium aquilinum, Dryopteris type, Polypodium vulgare, and miscellaneous pre-Quaternary palynomorphs (to
absence) and Dinoflagellates.
VIMP:2b, (205, 204, 203) c. 4.70–5.08 m OD: Defined by a marked peak of Potamogeton type (38%) (includes
Potamogeton and Triglochin maritima). This is associated with expansions of Plantago maritima (17%), Plantago
media/major type, Aster type ( 5%), and Chenopodiaceae (to 28%). There are corresponding reductions in
dinoflagellates and pre-Quaternary palynomorphs.
As discussed above the lithostratigraphical sequence at Vimpenny’s Lane comprises largely

minerogenic, grey silt, and silty clay. This has a texture and appearance typical of salt-marsh
sediments, a fact largely confirmed by the pollen data obtained for this site. Although salt-marsh
was important throughout the time-span of this sequence, changes in relative sea-level have given
subtle changes in the on-site vegetation and consequent stratigraphy. Of particular note is a thin but
highly organic detrital peat at 3.18 m OD which has furbished a radiocarbon date demonstrating
that the base of the pollen profile is of Middle–Late Neolithic age 2900–2620 cal BC (NZA-
12527,4182±55 BP). The regional vegetation of the late prehistoric period was dominated by oak,
lime, and hazel (seen in the lower pollen zone). A cessation in sedimentation at a depth of 3.98 m
OD (184 cm in the pollen profile) possibly delimits the later prehistoric sediments from sediments
of later historic (?post-Roman–medieval) age. During the interval of this hiatus, woodland
clearance had taken place. The pollen data overall can be correlated with the analysis of Awkley
Lane and Hallen and from further afield from the Somerset Levels which provide a detailed and
established framework for this geographical region.
43
4182+55 BP
Depth cm
96
80
64
48
32
112
288
272
256
240
224
208
192
176
160
144
128
Strat.
BE
T
PI UL
N A
U
For Key See Figure 7a

PC
I S
U EA
LM
Q US
U
ER A: Vimpenny's Lane
C
U
S
20 40 60
FR
A
PO XIN
TI PU US
LI L U
TIL A S
I
JU A (
N de
TA IP gra
X E d
C US R U e d
O B S )
20
RY A C
LU CC OM
ER S AT M
AV A UN
Figure 11a. Vimpenny's Lane pollen diagram.

IC IS
C A EL
AL LA
EM LU N
N A
AN PE A TY
E TR PE
R MO U M
AN
TREES & SHRUBS
SI UN NET
N
H AP CU YP
O SI LU E
cf RN TY S
. V U P TY
44
H IO NG E PE
YP LA IA
M ER TY
A I PE
D LVA CU
IA C M
C N T EA
ER H E
ST AS US
E T T
SP LL IUM YP
E A E
C RG RIA TY
H
EN U TY PE
20
O LAR PE
PO I
FA D AT
BA IU Y
TR C M PE
IF EA TY
M O E PE
ED LI U
R IC UM ND
O A T I
% = Sum Total Dry Land Pollen

FI SAC GO YP FF.
L T E
POIPE EAE YP
T N E
AP EN DU UN
I T L D
H AC ILL A IFF
ED E A .
PO ER AE TY
L A U P
PO YG H ND E
O E IF
HERBS
R LYG NA LIX F.
U O C
M
AR E NU EA
M X M EU
AR ER AV N
C ME IA ' IC DI
AL R A' U FF
Y IA L
LA .
SC S ' IN
B R
R TE ' L E ET
O G IN YP
PH IA E
U E
LA
R
IA
Zone
C
VIMP 1
VIMP 2
VIMP 1.a
VIMP 1.b
VIMP 2.a
VIMP 2.b.
EA
E
U
N
D
IF
F.
Depth cm
96
80
64
48
32
112
288
272
256
240
224
208
192
176
160
144
128
PL
A
PL NT
A A
PL NT GO
AN AG M
20
O A
PL TAG LA JO
A O N RT
cf NT M CE YP
L A A O
G ITT GO RIT LA E
AL O C IM T
VA IU RE OR A A
L M L LA O TY
SC ER N
A IA U O P
N P E
SU BI NA IF U
C OS O LO S
BI CI A FF R TYP
D SA TY IC A E
E
AS N T P IN
TE S T YP E AL
IS
B: Vimpenny's Lane cont.
AN R Y E
T P
AR HE TYP E
% Sum t.d.l.p.
TE M E
C M IS
IR I T
C SIU SIA YP
EN M E
LA TA T
C U Y
PO TU RE PE
AC CO A N
EA ID IG
20 40
E EA RA
E T
YP
C E
ER
LA EA
R L
U GE TY
HERBS Cont.
N P P
AL IDE O E
N NT AC
20
U IF E
SA S IE AE
L D
C IX /D
AL EG
M H T R
YR A AD
M I O TY ED
YR P P
45
cf IO HY E
.H P L
PO IP HY LUM
P
TA UR LLU A
M IS M LT
O V S ER
20 40
G U P N
LE ET L IC I
M O GA AT FL
N R U O
R NA
U TY IS M RU
% sum tdlp + Marsh

TY I
PP PE M
PH A M
TY A A
P R
C HA LAT ITI
YP A IF M
20
ER N OL A
O AC GU IA
S S T
PT MU EAE TIF YP
ER ND O E
A LI
MARSH
ID R A
IU E TY
M G PE
AQ AL
I
20 40 60
U S
D IL
RY IN
O U
20
PT M
PO ER
LY IS
T
20
PO YP
D E
SP IU
H M
A
20
VU
LI GN L
% sum tdlp + Spores

VE U G
AN RW M AR
TH O E
O T R
C S
ER
AS
Zone
PU
VIMP 1
VIMP 2
N
VIMP 1.a
VIMP 1.b
VIMP 2.a
VIMP 2.b.
C
SPORES
TA
TU
M
Figure 11b. Vimpenny's Lane pollen diagram. Note that alder (Alnus sp) is included with fen species in Fig 11b.
C: Vimpenny's Lane cont.
MISC. TOTALS
N ES
ER E R
Y
E
AR
l
E CH E A
/m
AT H
U SP
ns
C
H EM UM
M
TR ATA
-Q O
ai
U
G TR
gr
B ES
S
BS
ES
EN
0
ZY AS
ER L
PR I
H
S
x1
S
H CA
RS
N
R
EE
LL
C
I
O
D
R
YS
F
IS
I
PO
TR
SH
ER
PE
SP
AP
Zone
M
32
48
64
80
96 VIMP 2.b.
112
128
VIMP 2
144
Depth cm
160
VIMP 2.a
176
192
VIMP 1.b
208
224
240 VIMP 1
256
272
VIMP 1.a
288
20 40 20 40 60 20 20 40 60 80 20 40 20 40 60 80 20 40 200
4 00
6 00
800 500 100015002000
% = Sum + Misc.
Rob Scaife
Figure 11c. Vimpenny's Lane pollen diagram.
46
Vimpenny’s Lane: the environment
Peat and mineral sediments are present at a depths below 2.90 m OD, underlying the base of this
pollen profile which starts within the grey minerogenic sediment of layer 212. Sampling was not
undertaken from deposits below the exposed section, and recovery of suitable samples from the
auger hole (Fig. 9) was not feasible. Pollen was not present in the mineral sediments of unit 212
(below 2.90 m OD.). It seems likely, however, that the lower peats are of similar age and character
to those described for Awkley Lane (above). Given the broad topographical extent of the Avon
Levels, formation of the principal stratigraphical units is likely to have been autogenically
controlled. The overall character of the levels is undoubtedly due to the status of, and changes in,
the relative sea-level which affected the base level and extent of estuarine, salt-marsh, and
freshwater ecosystems.
The lowest pollen spectra at the base of zone VIMP:1 suggests that the sediment (context 212) was
deposited in a salt-marsh (possibly a mudflat, environment). This is evidenced by halophytes,
especially the Chenopodiaceae which include goosefoots, oraches, and glassworts. Values of these
increase substantially in zone 1a along with pre-Quaternary palynomorphs. This phase culminated
in accumulation of detrital peat (context 211) possibly caused by a negative change in relative sea-
level allowing the colonisation of marsh vegetation. Clapham suggests (below) that this was
Phragmites autralis fen. However, the numbers of halophytic pollen taxa also suggests periodic
estuarine incursions. This organic horizon has provided a valuable, later Neolithic, date for this
phase of 2900–2620 cal BC (NZA-12527, 4182± 55BP). Subsequently (from 3.22 m OD), there is a
change back to salt-marsh and/or mudflat sedimentation (context 210). Whilst there remain
relatively high values of derived geological palynomorphs, these are in reduced numbers which
may
indicate a different sediment source to that of layer 212/211 (zone VIMP:1a). There is, however, a
return to the higher values of tree pollen possibly derived from the fluvial catchment with Quercus
(oak), Tilia (lime), Ulmus (elm), and Corylus avellana type (hazel and possibly sweet-gale) being
the most important tree types. Pinus (pine) values (to 10%) typically reflects the over-
representation of its saccate pollen grains in fluvial environments (Groot 1966) and are not
considered important. This similarly applies to the occasional occurrences of Picea (spruce) for
which there is an increasing number of records for long distance marine transport (Scaife 2000).
Both Tilia and Fraxinus (ash) may be markedly under represented in pollen spectra (Andersen
1970; 1973). Values of Tilia, especially, are higher in sediments deposited during the Neolithic at
Awkley Lane and this undoubtedly reflects the closer proximity to the site of drier and more
favourable soils. At Vimpenny’s Lane, such land was over 0.5 km away, probably reflected in the
smaller pollen frequencies of lime and ash. The pollen here was most likely fluvially transported
and the occurrence of degraded pollen is a further indication of this and perhaps of sediment
reworking.
At the top of VIMP:1 (at 3.74 m OD) there is a clear reduction in Tilia to low levels. This reduction
falls below (by 20 cm) the zone VIMP:1/2 boundary which is delimited by a major reduction of
Quercus and a possible hiatus in sedimentation. This reduction in Tilia pollen is also associated
with an increase of Ericaceae (Calluna and Erica) and also herbs including Plantago lanceolata
(ribwort plantain). It seems plausible that this is the often described ‘lime decline’ which has been
similarly evidenced at Awkley Lane (above) and Hallen (below) in the Avon Levels and from the
Somerset Levels (see discussion section). This event may provide a useful datum since the lime
decline although an asynchronous phenomenon, due to its usual anthropogenic causation (Turner
1962), has been dated to the Middle–Late Bronze Age in many areas as discussed below. At
Vimpenny’s Lane there is clear evidence of landscape changes associated with human activity at
this time. Whilst waterlogging of the landscape may be responsible for reductions in lime pollen in
certain circumstances (Waller 1994), here, the decline of lime seems to be associated with
increasing evidence of human activity. Apart from the reduction of trees and increase in grassland,
47
there is some evidence of soil deterioration with increasing numbers of acidophiles (Erica, Calluna,
and Sphagnum) suggesting woodland clearance (for agriculture?) and subsequent deterioration of
the more sandy soils, possibly glacial sands of the lower Wentlooge Formation.
At 4.02 m OD there is significant evidence of environmental change (contexts 209/208) with

marked reductions in trees (especially oak) and expansion of herbs (esp. Poaceae). There are also
high absolute pollen frequencies at this horizon (to 200,000 grains/ml in context 208/207) but
which subsequently decline sharply. This horizon is somewhat enigmatic and was not recognised in
the field as a stabilistaion horizon. Two possible explanations may be suggested, first that the
sediment sequence is broadly continuous and that there is a ‘real’ decline in trees at this time (that
is, forest clearance) or second, that there is in fact a hiatus in the sediments so that these changes in
woodland may have occurred during the intervening period. If the underlying lime decline was of
Late Bronze Age date this cessation of sedimentation would apparently be of Late Bronze Age or
Iron Age date. The reason for the highest absolute pollen frequencies/values at this event may be
significant. If there was a palaeosol developed in this horizon (context 208), APF values would be
expected to decline down the profile (Dimbleby 1988) rather than upwards as happens here.
Interestingly, there is a reduction and absence of derived geological microfossils which suggests
less sediment derived from marine sources and/or from bedrock erosion. It is possible that there was
a change in taphonomy with perhaps freshwater sources introducing the Poaceae (and other herbs)
from the inland zone.
Subsequently, the sediments and the pollen spectra continue to show that the on-site environment of
deposition was one of salt-marsh and mudflat as demonstrated by the abundance of halophytes of
which Chenopodiaceae are the most important. Between 4.80 and 5.04 m OD (zone 2b; context
205) there is a significant increase in Potamogeton type pollen. This taxon includes both Triglochin
maritima (sea arrow grass) and Potamogeton (pond weed). In this sub-zone, there are also
expansions of Plantago maritima (sea plantain) and Aster (various pollen taxa but also including
sea aster) which suggests that the former (Triglochin) was most the likely taxon forming part of a
middle–upper salt-marsh community. Thus, it is suggested that this sub-zone represents a partial
drying out of the marsh with change from lower salt-marsh and mudflat to the middle and upper
salt-marsh suggested.
After this transient drier phase, there was a return to mudflat conditions (contexts 203/202) and
deposition of grey sediments containing much reworked geological pollen, dinoflagellates and
increases of Pinus and Betula from fluvial sources. It is thought that these upper levels are of
medieval or post medieval age.
A summary of the ecological and environmental changes

Peat and sediment underlying the lowest levels of this profile were not available for analysis. It
seems plausible that these underlying peats and mineral sediments are of similar age and character
to those described and dated from Awkley Lane (above) forming in carr and/or sedge fen. At
Vimpenny’s Lane the first pollen data from 2.90 m OD (base of VIMP:1; context 212/211) is of
middle Neolithic age. From this point in the stratigraphy, the ecology and environmental changes
are summarised as follows.
i) Salt-marsh (VIMP:1a, 2.90–3.16 m OD.) (context 212) sedimentation as demonstrated by

high Chenopodiaceae values and sediment character. This context also contains many
reworked pre-Quaternary pollen/spores and dinoflagellates. The background, terrestrial
vegetation comprised oak, lime and hazel woodland but possibly at some distance. Alder
and sedges were fluvially transported from river valleys flowing into the estuary.
ii) Stabilisation of the sediments caused by negative eustatic change (top of VIMP:1a, 3.14–
3.20 m OD) at c. 2750 cal BC. Middle salt-marsh habitat and formation of detrital peat from
48
marsh and fen vegetation (Chenopodiaceae, grasses, sedges)(context 211). Background
vegetation remained one of oak, lime, and hazel – at some distance along with alder and
Phragmites with occasional freshwater aquatic and marsh taxa (possibly sedges and bulrush
and bur reed).
iii) Return to lower salt-marsh, ?mudflat conditions and deposition of grey estuarine sediments
(3.20 m OD; context 210). Evidence of minor freshwater input from in-flowing rivers comes
from occasional pollen of aquatic macrophytes.
Background woodland remained one of oak, ash, lime, and hazel. There is, however,
evidence of reduction in lime woodland (possible the lime decline) and increasing
acidification of local acid sub-strates with increases in Ericaceae (Calluna and Erica) and
expansion of Sphagnum. Whilst this may suggest changing taphonomy, ie, pollen
transported from different sources, it is also plausible that this is evidence of increasing
human activity and consequent soil degradation – possibly due to increased human activity
in? the Bronze Age. This is to some extent evidenced by the expansion of ‘typical;’
anthropogenic herbs such as ribwort plantain (Plantago lanceolata)
iv) Possible Hiatus and/or truncation of the sediments at 3.98 m OD (context 208/209).
v) Woodland clearance: At this point in the pollen profile there is a marked reduction in tree
(oak) and shrub (hazel) pollen.
vi) Continuation of salt-marsh (above 3.98 m OD; contexts 207–202). Possible changes in
taphonomy give, initially, markedly high APF values above this interface but declining to
the low values which are typical of these salt-marsh sediments. The background vegetation
shows a general reduction in trees and in particular oak and hazel (lime had been removed
prior to hiatus). There is an increase in herb diversity including evidence of human activity.
The acid loving taxa (Ericaceae and Sphagnum) remain. The high APF values (context
208/207) also correspond to a reduction in oak percentages and especially low levels and/or
absence of pre-quaternary palynomorphs and dinoflagellates. This suggests changes in
sedimentary regime and taphonomy. It is possible that this may be a stabilisation phase
(although not considered to be a soil) with build up pollen on the surface whilst there was
less input from marine/estuarine sources with contributory derived geological microfossils
and oak pollen from the regional vegetation.
vii) Drier (?middle-upper) salt-marsh phase. 4.80–5.04 m OD (context 205). A peak of
Potamogeton type is most probably Triglochin (sea arrow grass) since there are also
expansion of allied taxa such as the Chenopodiaceae (oraches, glassworts etc.) sea plantain
(Plantago maritima type) and sea aster (Aster type.).
viii) Return to lower salt-marsh/mudflat conditions. Increase in pine, birch, and fern spores from
increase fluvial sediment input/transportation.
Diatom analysis
by Nigel G. Cameron
The two diatomaceous samples counted from Vimpenny’s Lane, the detrital humic peaty lens (211)
at 3.18 m OD and laminated grey alluvium (210) at 3.26 m OD respectively, have similar diatom
assemblages. Marine diatoms are dominant, comprising over 65% of the assemblage in the lower
sample of the detrital humic peat (211) and almost 85% in the laminated grey alluvium (at 3.26 m
OD). In both samples marine-brackish (polyhalobous to mesohalobous) and brackish
(mesohalobous) groups each account for approximately 5% of the diatom assemblages. There is,
however, a significant difference in the size of the freshwater (oligohalobous indifferent)
component of these samples. In the detrital peat (210) layer, freshwater diatoms comprise almost
20% of the total, whereas in the laminated grey alluvium (210) freshwater diatoms account for only
about 2% of the assemblage. Fragilaria pinnata, Fragilaria construens var. venter, and Fragilaria
brevistriata are the main freshwater components at +3.18 m OD. Paralia sulcata and the marine
49
tychoplanktonic (the tychoplankton spend part of their lifecycle associated with non-planktonic or
bethic habitats, in contrast to the plankton which complete their entire life-cycle in the open water)
diatom Cymatosira belgica are the dominant marine species in both samples.
Diatom assemblages like that of Vimpenny’s Lane in the humic peat, which are composed of
disparate marine and freshwater elements, and with a relatively small brackish water element, have
been observed elsewhere in the Severn Estuary (Cameron 1997). In this case tidal waters are the
most important source of diatoms. However, the appearance of a significant percentage of
freshwater species can be attributed to a number of factors. For example, freshwater diatoms can be
transported from their life habitats a short distance inland. The reason for the poor representation of
mesohalobous taxa is unclear but the hyper-tidal nature of the Severn may have resulted in erosion
and loss of brackish water habitats during transgressive and regressive phases. Further, considering
the ecology of the dominant freshwater diatoms, those Fragilaria species present are opportunistic
taxa which, despite their optimal growth in freshwater, appear to have wide salinity tolerances and
can be found in significant numbers in habitats periodically exposed to higher salinities. However,
consistent with its estuarine flora, the sample 3.28 m OD is a grey laminated alluvial clay (210).
The sample from +3.18 m OD is a humic/peaty stratum (211), which is consistent with the
significant component of freshwater diatoms in the assemblage. The dominant environmental factor
in both cases is the influence of water from the estuary.
The diatom assemblages examined are predominantly estuarine with a large marine planktonic
component typical of the Severn Estuary. Marine diatom plankton is commonly transported by tides
to sites of lower salinity than that for the species optimal growth. The groups are nevertheless
useful in indicating the predominance of tidal conditions.
The diatom assemblages show that detrital peat layer (context 211) has a mixed marine to
freshwater diatom assemblage. Although it is dominated by marine plankton, this sample has a
significantly greater percentage of freshwater taxa when it is compared with the other diatom
assemblages that have been analysed from the minerogenic sediments. The freshwater component
of this sample matches the context description which shows that the sediments have a high organic
content. Diatom assemblages composed of taxa with widely different salinity preferences are
common in coastal and estuarine sediments in general and in the hypertidal Severn Estuary in
particular (Cameron 1993; 1997; Cameron & Dobinson 1998; 2000; Walker et al. 1998a; 1998b)
and this mixing may be particularly marked at zones of transition between marine and freshwater
sediments.
In summary, the diatom assemblages of all four samples are dominated by a flora common
elsewhere in the Severn Estuary. The marine planktonic diatoms Paralia sulcata and Cymatosira
belgica are dominant, but a large number of other marine and estuarine taxa are common, for
example: Rhaphoneis spp., Podosira stelligera, Thalassiosira decipiens, and Pseudopodossira
westii. The sample from Awkley Lane, 4.42 m OD has a greater percentage of the mesohalobous
(brackish) non-planktonic taxon, Nitzschia navicularis. Whilst this species is rare at the other sites,
where the marine tychoplanktonic species Cymatosira belgica is more common, it is an estuarine
diatom associated with high salinity levels. This difference may be a reflection of slightly less
saline conditions compared with the other samples. Alternatively the difference in diatom
assemblages may also reflect a difference in depositional environment or suitable substrate for
diatom growth as much as a reduced salinity.
It should also be noted that the freshwater diatoms that are present in these samples, for example
several oligohalobous indifferent Fragilaria spp. and Navicula spp., are species tolerant of high
50
Table 6. Mollusc data from Vimpenny’s Lane
Sample 1000 1001 1002 1003 1004 1005 1006 1007 1008 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
b
Context 209 208 208 207 206 206 206 206 206 205 205 204 203 203 203 202 201 201 200
Depth (mOD) 3.91- 4.01- 4.11- 4.21- 4.31- 4.41- 4.51- 5.61- 5.71- 6.85- 4.95- 5.05- 5.15- 5.25- 5.35- 5.45- 5.55. 5.65- 5.75-
3.81 3.91 4.01 4.11 4.21 4.31 4.41 5.51 5.61 5.75 4.85 4.95 5.05 5.15 5.25 5.35 5.45 5.55 5.65
Wt (g) 1750 1750 1750 1750 1750 1750 1750 1750 1750 1700 1750 1750 1750 1750 1750 1750 1750 1750 1750
Vallonia pulchella (Müller) - - 1 - - - 1 2 1 - - - - - - - - - -
Land mollusc Taxa 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0
LAND MOLLUSC TOTAL 0 0 1 0 0 0 1 2 1 0 0 0 0 0 0 0 0 0 0
Bithynia tentaculata (Linnaeus)
Lymnaea truncatula (Müller) - - - - - 1 - - - - - - - - - - - - -
Lymnaea palustris (Müller) - - - - - 1 - - - 1 - - - - - - - - -
Lymnaea spp. - - - - - - - - 1 - - - - 1 - - - - -
Gyraulus albus (Müller) - - - - - - - - - - - - - - - - 3 - -
Hydrobia ventrosa (Montagu) - 1 - 106 117 - - 21 6 36 31 4 11 7 - - - - -
Hydrobia ulvae (Pennant) - 2 - 2 6 - - 1 - - 5 5 - 3 - - - - -
Hydrobia spp. - - - 7 5 - 2 3 - - - 4 - - - - - - -
Ovatella myosotis (Draparnaud) - - - - - - - - - - 1 - - - - - - - -
Aquatic Taxa 0 2 0 2 2 2 1 2 2 2 3 2 1 3 0 0 1 0 0
FRESHWATER TOTAL 0 0 0 0 0 2 0 0 1 1 0 0 0 1 0 0 3 0 0
MARINE TOTAL 0 3 0 115 128 0 2 25 6 36 37 13 11 10 0 0 0 0 0
Ostracod (valves) - - - 17 46 - 2 3 23 17 2 - - - - - 28 - -
TOTAL 0 3 1 115 128 3 3 27 8 37 37 13 11 11 0 0 3 0 0
(note: only samples with shells tabulated: totals exclude C. acicula)
51
salinities and are often associated with salt-marsh and high tidal mudflats (especially in areas with a
strong freshwater influence) despite their optimal growth in freshwater (Denys 1988).
Molluscs
by Michael J. Allen
Only the upper 2 m of the alluvial profile (above 3.8 m OD) was sampled for molluscs in a series of
19 contiguous samples (Table 6). Shells below this level seemed absent in the field, but a low
presence of Hydrobia ventrosa and of the estuarine species Phytia mytosis were present in the
evaluation trench at this level (Bell & Johnson, in Lawler et al. 1992, 109). The putative stabilisa-
tion horizon at 3.9–3.8 m OD is devoid of shells, but again Hydrobia ventrosa and cf. Macoma
were present in evaluation trench (op. cit.). The only point in the sequence where shell numbers are
moderate is at 4.3–4.1 m OD, at the interface of two visually different layers (207, very dark grey
(10YR 3/1) slightly humic silty clay and, 206, a grey (2.5Y N/5) silty clay with some organic
matter). Only Hydrobia species are present, mostly H. ventrosa. Although not common today, there
are old records of its existence in drainage ditches, rhines, and pills (Boyden et al. 1977, 524). This
suggests a quite brackish creek or lagoon environment, and the higher numbers may indicate
ponding of salty brackish water, perhaps in a local backfen environment, or washed in by flood
events. Notably, this high occurrence of solely Hydrobia species is precisely coincident with the
only occurrence of the rare salt-marsh pool plant Ruppia maritima in the two levels of the pollen
diagram.
Immediately above this level (layer 206) shell survival is poor but the presence of Lymnaea species
indicates the presence of freshwater discharging into this area and may suggest a changing
environment. The presence of a few terrestrial species might also indicate drier salt-marsh
conditions. Higher up the profile again (4.85–570 m OD; essentially layer 205) shell numbers
increase, dominated by Hydrobia ventrosa indicating wetter conditions and probably mudflats. Rare
instances of Lymnaea species indicate some freshwater conditons and Ovatella myosotis may
suggest rotting debris on mudflats (Macan 1977; Janus 1979, 66). Above 5.15 m shells are largely
absent. Three freshwater specimens of Gyraulus albus just below the present day soil profile are
probably washed in, though significantly, the occurrence of ostracod valves increase at this point
too (Table 6).
Insect remains
by Mark Robinson
A sample of 2 kg from the detrial peat deposit (211) within the Vimpennys Lane alluvial sequence
was subjected to paraffin flotation to recover insect remains. The flot was caught on a 0.2 mm
sieve, washed in detergent and sorted under a binocular microscope. The results are listed below:
Coleoptera Min. no. indiv.

Megasternum obscurum 1
Ochthebius minimus 1
Micropeplus caelatus 1
Stenus sp. 1
Pselaphidae indet. 1
The concentration of insect remains was very low and their preservation poor so no more of the
sample was analysed. However, one of the species of Coleoptera (beetles) is of particular interest.
Micropeplus caelatus is now extinct in England although it is still to be found in the British Isles in
the south-west of Ireland. It inhabits swampy or marshy ground, where it occurs in decaying plant
52
refuse. Several examples of M. caelatus were, moreover, found in Neolithic reedswamp deposits at
the base of the peat sequence of the Somerset Levels at the Rowland’s Track site (Girling 1977,
53). The sparse insect assemblage from Vimpenny’s Lane would be appropriate to vegetation debris
in a reedswamp.
Waterlogged plant macrofossils

by Alan J. Clapham
Two sub-samples of 300 and 250 cm3 from the organic detrial humified peat horizon (211), were
analysed for waterlogged plant macrofossils. This layer has been dated to 2900-2620 cal BC
(4182±55BP, NZA 12527). The sample was found to contain only a limited number waterlogged
plant macrofossils (Table 7).
The waterlogged plant macrofossils were well preserved and allowed positive identifications in
most cases. A total of 22 taxa was recovered; there were very few large plant macrofossil remains.
The largest were fragments of reed (Phragmites australis). Charcoal fragments were also recorded
with the majority being pithy in nature suggesting that some stems of reed had been burned.
Charred culm nodes which can be assumed not to be of a cereal were also identified supporting this
interpretation.
Tree species were represented by a single find of a birch (Betula) fruit. As the fruits of this tree
have wings and only one example was recorded it may be that the origin of this fruit were extra-
local, although the majority of birch seeds are known to fall within a short distance of the parent
tree (Gimingham 1984). The dominant habitat was that of wetland/fen/reedswamp – taxa recorded
include celery-leaved buttercup (Ranunculus sceleratus – which is also common in shallow
brackish water in coastal marshes), stinging nettle (Urtica dioica), woody nightshade (Solanum
dulcamara), gipsywort (Lycopus europaeus), watermint (Mentha aquatica), reed, and reedmace
(Typha sp.). The presence of mosses also suggests damp conditions. Oak-leaved goosefoot
(Chenopodium cf glaucum), common orache (Atriplex patula), and the Chenopodiaceae in general,
along with docks (Rumex sp.) and cleavers (Galium aparine) are most often associated with
disturbed, arable, or waste ground, but their presence here may be due to areas of the marshy/fen
ground having been disturbed by some activity; whether this was natural or anthropogenic in nature
is difficult to determine. The same can be said for the grasses (bent (Agrostis type) and couch
(Elytrigia type) although, again, these grasses can be found in natural habitats. A scrub type habitat
may be indicated by the presence of bramble (Rubus Section 2 Glandulosus) although, like the
cleavers, these can be found in fen/reedswamp conditions growing over the other vegetation along
with woody nightshade.
53
Table 7. The waterlogged plant remains from the detrital
peat (context 211) from Vimpennys Lane
Context no. 211 211 Overall, the type of habitat indicated

Sample no. 1019 1019
is tall fen/reedswamp vegetation
O.D. 3.14 3.14
dominated by reedmace and reeds,
14C yrs BP(uncalibrated) 4182 ± 55 4182 ± 55
Volume processed cm
3
300 250
with some stinging nettle over which
the bramble, woody nightshade, and
Musci common -
cleavers are scrambling. Below this
Ranunculus sceleratus 99+99f 46+69f tangled mass of vegetation, the celery-
Urtica dioica 5 - leaved buttercup, gipsywort, and
Betula sp. seed 1 - watermint can be found growing. In
Chenopodium cf glaucum - 5 some areas where there has been some
Atriplex patula 5+18f 8+19f disturbance oak-leaved goosefoot,
Chenopodiaceae 5 61
common orache, and the grasses
Rumex sp. 1 -
occur. This type of reedswamp
Rubus Section 2 Glandulosus 1f -
Rosaceae thorn/prickle 1 -
vegetation is nowadays often found
Solanum dulcamara 36f 23f around ponds, lakes, and at the sides
Lycopus europaeus - 3 of rivers where the water-flow is slow
Mentha aquatica/arvensis 43+32f 52+47f or standing and in low lying areas
Galium aparine 14+7f 7+19f where the watertable is high. This
Asteraceae indet 14 3 type of vegetation can cover large
Agrostis type caryopsis - 10
areas. The presence of the pithy
Elytrigia sp. caryopsis 1 100+
charcoal and the charred culm nodes
Phragmites australis rhizome fragments - 2
Typha latifolia/angustifolia 21 30
suggests that some of the taller
Charcoal 80f - vegetation had been burnt at some
Charred culm nodes - 26 stage. Whether this is a natural or
Miscellaneous 17 - anthropogenic event is difficult to say,
although the presence of the disturbed
habitat plant species and the grasses may suggest that there was some limited human activity in the
area, possibly to clear some of the tall, tangled vegetation to gain access to more open water for
fishing or other hunting activities.
3. Hallen Marsh
The sequence
Several sherds of Iron Age pottery and other occupation remains were found in a ‘humic’ layer 0.9
m beneath the surface during observations of a geotechnical pit at Hallen Marsh (ST 3543 1804 and
6.4 m OD) (Russett 1990, 5). A number of machine excavated trial trenches were cut to clarify the
nature of the remains (Lawler et al. 1992, 79–84). This evaluation identified a series of
archaeological features comprising a rubble spread and deposits rich in artefacts, in particular Iron
Age pottery, sealed beneath the ‘humic’ layer. Given the undoubted potential of this, one of the first
prehistoric occupation site identified on the Avon Levels, a full archaeological excavation was
conducted clearing an area of 1600 m² for examination (Fig. 12).
Within this cutting, detailed archaeological excavation was restricted to two areas (Fig. 12) where
remains of Iron Age occupation had been identified, including two round-houses with surrounding
drip-gullies that had been recut/re-established on several occasions (Plates 4 & 5); a third similar
54
area was recorded but not investigated. The archaeological remains and associated cultural debris
are reported elsewhere (Barnes 1993; Gardiner et al. 2002).
The sequence at Hallen Marsh was exposed to a depth of only 2.5 m max (4.10 m OD; Fig. 13) but,
by virtue of the archaeological excavations, a long section of some 40 m+ was exposed (Fig. 14).
This provided an indication of lateral variation in the sediment sequence. Over much of the site
only c. 1.4 m (to 4.36 m) of the deposits were exposed. Although the present surface topography is
essential flat and level, the occupation deposits occurred under the upper reddish-brown alluvium
on slight natural rises separated by stretches of alluvial infill. Because of this variation, two main
sequences were described (Figs 13 & 14); one where occupation deposits were encountered and a
second through the stratigraphy between two major foci of Iron Age occupation.
Non-occupation sequence (Fig. 14)

Beneath the topsoil, uniform deposits of alluvial clay loam (100 and 101) with frequent ferric and
manganese flecking, which providing the reddish brown hue, occurred across the entire area
regardless of the occurrence of archaeological deposits. This alluvium was up to 1.1 m thick, but
only 0.5 m thick over the archaeological deposits. At the base of this ‘upper alluvium’ a zone of
more intense mottling occurred (102 and 151) in a greyish brown (10YR 5/2) silty clay. Where this
overlay the archaeological deposits it was only 0.1 m thick (102) but was appreciably thicker where
no archaeological deposits were encountered. Between the ‘islands’ of archaeology, thicknesses of
up to 0.7 m of gleyed grey (2.5Y N7) and pale brown (10YR 6/3) alluvium (152 and 186
respectively) with lighter grey veins occurred (151). Between the ‘islands’ of archaeology, this
local ‘channel-fill’ alluvium was up to 1.2 m deep.
Beneath both the archaeological deposits and the local channel-fill alluvium was a dark grey gleyed
mottled and veined clay (184) which was at least 0.75 m thick (exposed to c. 4.1 m OD). It was
laminated in some places and blocky elsewhere and, was typically mottled with diffuse orange,
ferruginous colours, and streaked and veined with predominantly light grey colours. This alluvium
(183 and 184) occurred at c. 4.65 m OD between the occupation areas, but was elevated and nearly
1 m higher (c. 5.4 m OD) where the occupation occurred. This was the ‘grey alluvium’ recorded
elsewhere. The upper 0.2 m of this (183) was sandier (sandy clay) and a distinctly browner hue was
probably the oxidised surface of the upper Wentlooge Formation hue.
Occupation sequence
The occupation deposits were identified beneath the upper reddish brown alluvium as three areas of
darker, greyer deposits rich in fragments of pottery, burnt clay, and animal bone. These were light
grey (5Y 7/1) to grey (5Y 6/1) and dark grey (5Y 4/1) silty clay up to c. 0.15 m thick, with iron and
manganese staining and weak medium blocky structure indicating a post-occupation soil formation.
This layer sealed the surfaces, structures and ditches of the settlement and is thought to represent
the development of a soil above the abandoned site. The ‘occupation’ deposit, as defined by the
field archaeologists, was a light grey (10YR 7/2) clay up to 0.12 m thick with a compact surface
(trampling and ?crusting) and distinctly greenish and olive (5Y 6/4) hues and was restricted to the
areas of settlement features.
The surface beneath the occupation deposits was not obviously eroded and it is assumed that the
lower levels between the occupation foci, although possibly exacerbated by later erosive and
flooding events, are essentially the true form of the Iron Age topography. The occupation foci were
located on slightly higher rises within an essentially flat and level alluvial plain. As there is no
stratigraphic relationship between the channel and the settlement area, and a lack of artefacts within
the channel the channel fills there is no way of telling whether the form of the ‘channel’ between
the settlement areas has been deepened subsequent to settlelment and prior to infill by alluviation.
55
Occupation sequence
0-0.14 6.43-6.29 Topsoil
0.14-0.30 6.29-6.13 100 Greyish brown (10YR 5/2) stonefree clay loam, common Upper alluvium
fine and medium fleshy rots, rare charcoal flecks, and rare
small limestone fragments
0.30-0.59 6.29-5.84 101 Light yellowish brown (10YR 6/4) clay loam with rare fine Upper alluvium
fleshy roots, - a reddish appearance is given to these layers
by the frequent ferric and manganese flecking (gleying),
gradual to diffuse smooth boundary to layer below (102)
which is distinctly greyer.
0.59-0.66 5.84-5.77 102/ Light brownish grey (10YR 6/2) silty clay with rare small post occupation
151 limestone fragments, flecked with Fe ad Mn staining. Seals alluviation
the archaeological occupation deposits.
0.66-0.85 5.77-5.58 105 Light grey (5Y 7/1) to grey (5Y 6/1) silty clay, weak Occupation
medium to large blocky peds with moderate Fe and Mn
staining, becoming darker in colour where highest
concentration of artefacts occur and conversely distinctly
lighter away from these three defined foci (buildings). It
contains patches and spreads or lumps of buff brown clay
indicating mixing and post possibly deposition colour
changes. Occupation layer
0.85-0.91 5.58-5.52 321 Light grey (10YR 7/2) clay with greenish hue and compact Occupation surface
surface
0.91+ 6.52+ 183 Brown (10YR 5/3) gritty gleyed clay, mottled with grey and Grey alluvium
orange streaks. It occurs as a thick band which is laminated
in some laces and blocky elsewhere; pre-dates the
occupation, diffuse boundary
Non-occupation sequence
0-0.11 6.56-6.45
0.11-0.51 6.45-6.05 100 Greyish brown (10YR 5/2) stonefree clay loam, common Upper alluvium
fine and medium fleshy rots, rare charcoal flecks, and rare
small limestone fragments
0.51-1.15 6.05-5.41 101 Light yellowish brown (10YR 6/4) clay loam with rare fine Upper alluvium
fleshy roots, - a reddish appearance is given to these layers
by the frequent ferric and manganese flecking (gleying),
gradual to diffuse smooth boundary.
1.15-1.33 5.41-5.23 151 Greyish brown (10YR 5/2) silty clay with lighter grey veins post occupation
and mottling, diffuse alluviation
1.33-1459 5.23-5.07 152 Light grey (2.5Y N7) gleyed clay, rare small stones; gleying Upper alluvium
creates a blueish hue. The relationship with layer 105 is not channel fill;
clear. Gleyed channel infill
1.49-1.72 5.07-4.84 186 Pale brown (10YR 6/3) clay with some mottling and orange Upper alluvium
and grey streaks channel fill;
1.72-1.85 4.84-4.71 183 Brown (10YR 5/3) gritty gleyed clay, mottled with grey and Grey alluvium
orange streaks. It occurs as a thick band which is laminated
in some places and blocky elsewhere; pre-dates the
occupation, diffuse boundary
1.85-2.46 4.71-4.10m+ 184 Dark grey (10YR 4/1) gleyed clay, slightly mottled and Grey alluvium
veined. Wentlooge series
Two main locations within the sequence were sampled. One was through the northern occupation
area and building (Fig. 14), where disturbed bulk samples were taken through the sequences of land
snails. The second location was within the ‘palaeochannel’ between the occupation areas. Two
sample sequences here comprised a monolith sequence through the alluvium and the deposits
underlying all settlement activity, and a comparable sequence of bulk samples at 0.1 m intervals
taken 5.5 m to the north (see Fig. 13).
56
WA Hallen Marsh
Section
Fig.7
House 1
House 2
0 10 m
Figure 12. Hallen Marsh: Summary plan of the excavation
57
WA Hallen GGAT G018
100
100
6 m O.D.
101
102/151
183
0 10 20 30
x10-8 SI/kg
5 m O.D.
WA Hallen
WA Hallen
100
Pollen Zones
6 m O.D.
101 101
151 151
2
152
5 m O.D.
186 152
183
186
184 184
1
0 10 20
x10-8 SI/kg
4 m O.D.

Alluvial silty clays (colour unknown)
0 1m
Figure 13. Hallen Marsh; full site profiles
58
A. WA Hallen Summary
S N
B1 B2
Vertical scale
1m
Section break
see below
5m
Horizontal scale
B. Detail
7 m O.D.
S N
B1
100
6 m O.D.
101
181
5 m O.D.
182
183
184
Continued below
4 m O.D.
7 m O.D.
S N
B2
100
6 m O.D.
101
181
5 m O.D.
182
183
184
Continued above
4 m O.D.
Vertical scale
Main post-occupation alluvial sequence 1m

Primary post-occupation alluvial sequence
Humic land surface above abandoned settlement Alluvial silty clays (brownish red hue)
5m
Iron Age occupation deposits
Horizontal scale
Alluvial silty clays (bluish hue)
Figure 14. Hallen Marsh Wentlooge Formation profiles
59
Textural features
by J.R.L. Allen
Sixteen sediment samples from the channel between the settlement foci were analysed and grain-
size patterns are shown graphically (Fig. 15). The triangular graph shows the scatter of values of the
clay-silt-sand ratio and shows the sediments to range from clayey silts to sandy-clayey silts and to
be broadly similar to other, similarly situated salt-marsh deposits from the outer Severn Estuary.
The stratigraphic sequence of samples reveals three textural patterns. The earliest few samples
(layers 184–183; 4.46–4.76 m OD) display no overall textural trend. From top of layers 184–151
(4.86–5.33 m OD) there is a marked fining-up trend. A strong coarsening-up trend is evident from
the top of the sampled sequence (layer 101; 5.43–5.73 m OD), the two topmost samples being
coarser-grained than any others in the profile.
Textural patterns from Holocene salt-marsh sequences cannot be given unequivocal interpretations
as two reinforcing factors are at work (Allen 1995). Firstly, the mud deposited over a marsh is
advected by the tide, which enters the marsh across its edge and along the major creeks. Because
the heavier particles settle out earliest, the deposit consequently becomes finer-grained with
increasing distance inland. If the marsh-edge is retreating, the successive deposits formed at a point
on the marsh will increase in coarseness over time. The opposite trend is expected if the marsh edge
is building out. The latter is favoured by a negative sea-level tendency – a decline in the rate of sea-
level rise – whereas a positive tendency – an increase in the rate of sea-level rise – encourages
marsh-edge retreat. Secondly the sea-level tendency influences the height of a marsh relative to the
tidal frame and, therefore, the general coarseness of the sediment advected over the marsh. During
times of positive tendency, the marsh surface is depressed, allowing coarser sediment from faster-
moving water to enter, The opposite effects are expected during a period of negative tendency. in
the case of the Hallen Marsh profile, the site seems to have experienced (1) a period of stability of
the marsh edge and surface, followed by (2) a period of negative sea-level tendency and/or marsh
outward growth, succeeded in turn by (3) a period of marked positive sea-level tendency and/or
marsh-edge retreat.
Pollen analysis
by Robert G Scaife
A number of questions were posed by the archaeology which it was anticipated that pollen analysis
could answer. These questions relate to the environment of deposition of the lower grey sediments
(middle Wentlooge Formation) and an overlying alluvial deposit. Furthermore, it appears that the
Iron Age settlement was abandoned and lies under thick alluvial silts. It was anticipated that multi-
disciplinary analyses might demonstrate the environment and causes of this settlement
abandonment and the basic elements of environmental and habitat change.
Pollen was extracted from all 17 samples examined. Absolute pollen frequencies were
generally low in the largely minerogenic sediments and this dictated the pollen sum counts
made. However, with the exception of the uppermost level, counts of greater than 300 grains
per level were made throughout thus providing statistically adequate data. These pollen data
are presented in standard pollen diagram form (Fig. 16).
Absolute pollen frequencies were variable, ranging from lowest values at the top of the
profile of 1700 grains/ml to 67,500 grains/ml at 5.24 m OD – values are higher in the
60
Hallen Marsh
Clay (100%)
50% 50%
y
Cla
Silt (100%) Sand 50% Sand (100%)
Figure 15. Hallen Marsh; particle size triangle
61
uppermost of the two recognised pollen assemblage zones. In pollen zone HALL:1, APF
values are in the order of 10–15,000 grains/ml. Thus, the values and available pollen for
counting is, in general, small. Pollen preservation was also variable with some deteriorated
grains of Lactucoideae especialy in zone HALL:2 indicating longer and differential
preservation of this taxon. However, the majority of pollen grains were moderately well
preserved in spite of the low absolute pollen frequencies and the minerogenic character of the
sediments. This is possibly a product of the taphonomy and alluvial preservational
environment.
Two distinct local pollen assemblage zones have been delimited in the Hallen Marsh pollen
sequence. These are delimited and characterised from the base of the profile as follows.
HALL:1 4.20–4.80 m OD, Quercus-Corylus avellana type-Chenopodium type-Plantago maritima-Poaceae.

This zone is characterised by higher values of tree pollen than in zone HALL:2 with substantially higher values
of halophytes-Chenopodiaceae and Plantago maritima. Trees are dominated by Quercus (to 43%) especially in
the lower levels of the zone. Betula (7%), Pinus, Ulmus, Tilia, and Fraxinus are also more abundant but values
are relatively small. Corylus avellana type is the principal shrub (18%). Herbs are important (to 60% of pollen)
dominated by Poaceae (20–30%) with Chenopodium type (to 16%), Plantago maritima (peak to 19%) and
Plantago lanceolata (10%). Lactucoideae are incoming mid-way through the zone. Freshwater marsh taxa
comprise Alnus (to 8%) with Cyperaceae (8%). There are sporadic records of other fen taxa including Typha
angustifolia/Sparganium type, Typha latifolia and Littorella uniflora at the top of the zone. Halophytes are
Chenopodiaceae, Plantago maritima type, Spergularia type, Aster type and Potamogeton type. The latter,
Potamogeton type may be Potamogeton and/or Triglochin. Spores of ferns include dominant Pteridium
aquilinum (peaking to 50% at 4.60 m OD). Dryopteris type (monolete) spores (average 10–12%) and
Polypodium vulgare (6%) are most important. There is a small but consistent presence of Sphagnum moss
spores. Pre-Quaternary palynomorphs (including Jurassic-Cretaceous Cycadopites) are consistent (10%). Of
note is a peak of Dinoflagellates at the top of the zone.
HALL: 2: 4.80–5.48 m OD, Plantago-lanceolata-Lactucoideae-Poaceae. This upper zone is delimited by

reductions in Quercus (to 10–15% av.), and Betula, Pinus, Tilia, and Alnus and the herbs Plantago maritima
type (to <2%), Chenopodiaceae (5%), and possibly Potamogeton type (if Triglochin). Contrasting are significant
expansions of Plantago lanceolata (to 35%), Plantago coronopus type (8%), Lactucoideae (8%), Poaceae
(50%), and some increase in Cereal type. Plantago lanceolata and Poaceae are dominant. Within the trees and
shrubs, Quercus and Corylus remain the most important and consistent taxa. Fagus is present in the lower half
of the zone. Although there is a substantial reduction in the halophytes Chenopodiaceae and Plantago maritima
remain. Marsh taxa comprise Alnus reduced to (4–5%) with Cyperaceae expanding to a peak of 15% at the top
of the profile (5.32 m OD). There are occasional freshwater aquatic/megaphytes (Myriophyllum alterniflorum,
Lemna, and Typha/Sparganium type and algal Pediastrum). Spores remain dominated by Pterdium aquilinum
(expanding to 55%) with Dryopteris type (to 14%) and Polypodium vulgare (values increasing upwards). There
is a peak of Dinoflagellates/Hystrichospheres at the lower zone boundary (4.84–4.68 m OD). There is also an
expansion of Pre-Quaternary palynomorphs (Jurassic).
Vegetation and Environmental Change at Hallen Marsh

The two pollen assemblage zones represent a distinct change in environment and pollen
taphonomy. The Iron Age archaeology of the Hallen site corresponds with this horizon of
change at c. 4.80 m OD.
Pollen zone HALL:1 contains substantial quantities of halophytes including Chenopodiaceae

(includes Chenopodium, Atriplex, Salicornia), Plantago maritima (sea plantain) and possibly
Aster type (?sea aster) and Potamogeton type (?Triglochin-sea arrow grass). These halophytes
and the character of the sediments suggest that the on-site environment was one of salt-marsh
(Upper Wentlooge Formation). This would have been subject to brackish water/marine
inundation at mean high water spring tide.
62
Depthcm
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
115
110
125
120
105
100
Strat.
BE
PI T UL
N A
PI U S
C
A: Hallen
U EA
LM
For Key See Figure 7a

Q U
U S
ER
C
US
20 40
TI
L
TI IA
L
FRIA (
d
FA AXI egr
G N a
PR US U S ded
U )
C N
O U
RY S
20
VI LU TYP
B S E
ER UR AV
N E
C IC A U M LL
AL AN
A
Figure 16a. Hallen pollen diagram.

R LU
AN N TY
SI U A PE
TREES & SHRUBS
NA NC
20
PI UL
H
O S T US
c f RN YP T Y
.H U E P
M YP N G E
A I
D LVA ERI A T
IA C C Y
SP N T EA U M PE
H
C ER US E
HE G T
N UL Y
20
Sum = % Total Dry Land Pollen
F A OP AR P E
B O IA
M AC DI T
E E U Y
T R DI C AE M T PE
I A YP
PO FO GO UN
L D E
AP T E I UM TY I FF
I N P .
PO AC T IL T Y E
E L P
R L YG AE A T E
U T Y
63
S C M E ON YP P E
X UM E
PL RO AV 1
A N PH IC
PL T U
AN A G LA U
LA
TA O R IA RE
20
G MA C E TY
O J A
LA O E PE
NC T NR U
PL E O YP D I
AN LA E F F.
40 20
TA
PL TA
A G
LI N T O M
TT A
G O GO A R
A R IT
SU LI U ELL CO IM
M A R AT
BI CC UN ON YP
DE IS IF OP E
AS N A T LO U
S Y
ANT ER T Y PE RA S T
HERBS
P YP
AR TH TY E
EM P E
T
C EM I S E
IR
C SIU ISI AT YP
EN M E
LA T T
C A Y
LI TU UR PE
LI C EA
A
PO C I N O
AC EA DE IGR
EA E U AE A
E ND TY
IF P
E
20 40
F.
C
E
LA RE
R A
U G LT
NI E YP
DE PO E
N AC
TI E
FI A
ED E
Zone
Zone 1
Zone 2
/D
EG
R
AD
ED
Depthcm
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
115
110
125
120
105
100
AL
N
US
20
SA
L
M IX
Y
LE RI O
M P
M N HY
E A L
PO NYA LU
M
LI T A N TH AL
TT MO E TE
R
B: Hallen cont.
H OR G S NI
YD E E
T Y R LL O T FL
P O A N O
T Y HA CO UN TY R
UM
P T P
C HA LAT YL IF L E
YP I F E T OR
A
20
ER N OL Y A
IS G
AC U I A E P
O S T
EQ ET EA TI F YP
E E O E
O U IS S L LI
SM E A A
PT U T U U C TY
ER N D M ST PE
MARSH
R
ID A R IS
IU E
M GA
AQ L
20 40 60
U IS
Figure 16b. Hallen pollen diagram.

D IL
RY IN
O U
20
PO P T M
LY ER
SP P IS
H O TY
LI AG D I
V U PE
AN ER NU M V
TH W O M UL
PE O R G
T AR
D D IA CER S
IN S A E
PR OF T R S
E- LAGU M PU
Q E NC
UA L TA
20 40
T LA
SPORES
ER T TU
cf N ES M
.A AR
T R SC Y
EE AR PA
S IS LY
NO
M
O
20 40 60
RP
64
SH H
S
20
RU
ER BS
IC
H A
MISC.
E R LE
BS S
20 40 60 80
M
AR
20
SH
SP
O
R
ES
20 40 60 80
M
IS
C
TOTALS
20 40 60 PO
LL
E N
TO
TA
400
A.
P.
F .X
10
G
ra
in
s/
m
l.
200 400 600 800
Zone
Zone 1
Zone 2
Note that alder (Alnus sp) is included with fen species in Fig 16b)
This zone also shows evidence of woodland dominated by Quercus (oak) with Corylus
avellana type (hazel or possibly sweet gale) with some lesser representation of Tilia and
Fraxinus (ash). The latter are, however, usually very under represented in pollen spectra
(Andersen 1970; 1973) and as such, may have been more important on drier soils. However,
given that even these values are small and the possible taphonomy of the pollen it seems
likely that we are seeing the last vestiges of areas of previous important lime dominated
woodland such as described for Awkley and Vimpenny’s Lane. Betula (birch), Pinus (pine),
and Ulmus (elm) are considered here to be of extra-local origin.
The change to zone HALL:2 at 4.80 m OD shows a change in environment from upper salt-
marsh to apparently one of pasture ‘on-site’. Poaceae (grasses) are dominant with substantial
quantities of Plantago lanceolata (ribwort plantain) and sporadic herbs typical of pasture
(Fabaceae) and Lactucoideae (dandelion types). The pasture was probably not short
cropped/grazed or tall herb pasture but was probably wet meadow of medium length pasture
(based on the presence of Plantago lanceolata flowering). Plantago coronopus is also present
and increasing in zone HALL:2. This is more typical of coastal grassland swards on sandy
soils and dunes and unstable ground. It is possible that the pollen source was from
path/trackways on the more sandy sediments of the floodplain or from near coastal grassland
habitats (dunes?).
Reduction in tree pollen from zone HALL:1 represents an important taphonomic change.
Zone HALL:1 saw the accumulation of salt-marsh sediments along with the typical salt-
marsh/marine silts and included pollen. This would have been from periodic (possibly
monthly) high water inundation. However, in zone HALL:2 there is evidence of a change in
taphonomy from salt-marsh/brackish water and marine derived pollen to probable freshwater
alluvium containing pollen from different areas of the catchment. In zone HALL:2, there is
evidence of sediment sources derived from the local and/or regional (Jurassic) geology;
derived palynomorphs are more abundant in zone HALL:2. The reduction in oak and hazel
may not, therefore, be a real decrease in the extent of such woodland at this time period (ie,
pollen zone boundary).
The cause of this change from salt-marsh to rough pasture is not yet clear. It seems most
likely that it is the result of some negative eustatic tendency (decline in sea-level relative to
land), though much more localised factors cannot be ruled out. This would have allowed
colonisation of the grassland on the subsequently drier soils and as a consequence, Iron Age
settlement/activity. However, there is clear evidence for continued sedimentation, possibly
caused by seasonal flooding (and overbank sediment deposition).
Some remaining halophytes (eg, Chenopodiaceae) indicate that salt-marsh remained within
the near region. Alternatively, there may have been occasional marine/brackish water
inundation depositing these halophytic taxa. Diatom analysis (Cameron, below) may also
demonstrate this and provide a clue as to whether the halophyte pollen is of airborne
derivation or fluvially transported (ie, if marine/brackish water diatom taxa are found). Such
periodic inundation and sedimentation may have been responsible for abandonment of the
archaeological site.
Some cereal pollen and possible associated weeds are present throughout both zones, but
especially in zone HALL:2. This is clear evidence that cereal cultivation was taking place.
However, values are small and it is suggested that the pollen is derived from some distance
(fluvially or airborne) from suitable interfluve soils in the region.
65
Summary of environment and habitat changes
i) Bronze Age/Iron Age deposition of typical grey, marine salt-marsh sediments (Wentlooge
Formation) as a consequence of progressive rise in relative sea-level (positive tendency).
Habitat ecology was upper salt-marsh dominated by Chenopodiaceae (eg, oraches but also
transported glassworts from the lower salt-marsh?) and Plantago maritima. Dry land
vegetation was apparently oak and hazel although there were perhaps the last vestiges of
lime woodland (ie, just post-Late Bronze age lime decline) and possibly some secondary
ash. Pine and birch are present but are probably long distance transported components.
ii) Sedimentological and pollen taphonomic changes at c. 4.80 m. The previous salt-marsh
changes to a grassland pasture environment dominated by grasses and ribwort plantain. Also
at this horizon there is archaeological activity and habitation. This may have been a
response to the availability of this new pasture/land surface. The cause of this may have
been drainage by the Iron Age communities (but there is no archaeological evidence to
confirm this) but more likely due to a negative eustatic tendency (ie, small lowering in
relative sea-level). Unexplained at present is a peak of Hystrichospheres across this zone of
change. Was this flooding? The sediment source changed from marine/salt-marsh to
probably freshwater alluvial material. This deposited sediments containing geological,
derived pollen, eroded and transported from the river catchment. This was possibly seasonal
and as such, the question of seasonal occupation of the floodplain and grazing activity
should be considered (such as by Mark Robinson in his discussion of the Thames floodplain
at Farmoor, Oxfordshire). Furthermore, introduction of winter sown wheat during the Iron
Age would have caused greater soil instability, soil erosion and thus sediment supply within
the fluvial catchment.
There is some evidence for cereal cultivation at distance on the better drained
interfluves and island soils. Oak and hazel woodland probably remained with reduced pollen
values due to the taphonomic changes described. Alternatively, if there was increased Iron
Age activity and intensification of land use, there may also have been woodland clearance.
Charred plant remains and querns found on-site and the small values of cereal pollen in
section would tend to indicate that pre-processed cereals were imported for domestic use on
the site. This was an apparent phenomenon of emmer and spelt cultivation during the Iron
Age (producer and consumer sites) with these cereals being transported as whole ears to
consumer sites. Areas of marine influence (inlets or rivers) remained within the proximity of
the site: periodic inundation of nearby salt-marsh?
iii) Expansion of pine, birch and hazel and Cyperaceae may indicate increasing wetness of the
site; the latter indicating fen, the former from alluvial transport? There is a reduction of
ribwort plantain, hoary plantain and to some extent, grasses which may also be due to
waterlogging of the soils. This appears to have become mudflat, with sediments derived
from fluvial sources and relating to the local channel infill; there is no indication of marine
inundation.
Conclusion
There is clear evidence of environmental change from salt-marsh to grassland pasture. The
transition is also at the point of Iron Age activity and confirms suggestions that the occupation was
sited on an alluvial plain. This may, however, have been drying out salt-marsh in transition to
pasture. Whether this activity was causative or a response to change is at present enigmatic. It is
realised that some of the suggestions made may be a bit tenuous and it is expected that integration
with other environmental aspects in the future may produce a more definite conclusion.
66
Diatom analysis
by Nigel G. Cameron
One diatom assemblage, from the grey alluvium, 184 (4.20 m OD), was counted from Hallen Marsh
and was composed of approximately 85% marine species, almost 10% marine-brackish species and
less than 5% brackish water species. Halophilous and freshwater diatoms comprise less than 1% of
the diatom assemblage. The marine planktonic diatom Paralia sulcata comprises about 50% of the
assemblage, whilst the marine tychoplanktonic diatom Cymatosira belgica accounts for over 25%
of the diatom flora of this sample. This assemblage clearly represents a diatom flora deposited
under estuarine conditions, perhaps with a high proportion of allochthonous, marine planktonic
diatoms transported from the outer estuary. However, the local habitat is exposed to high salinity
and tidal conditions. The diatom assemblage is consistent with the context (184) description for this
sample which indicates a dark grey clay interpreted as part of the Wentlooge Formation.
Molluscs
by Michael J. Allen
Two columns of samples were taken through the relatively short/shallow sequence at Hallen Marsh.
The first, a sequence of ten samples through House 1 and the occupation deposits, containsed very
few shells. The robust Hydrobia species is only present in one sample and Ovatella mytosis, which
does not occur in the second sequence, is also present in one sample. This is an estuarine species,
but one common under driftwood and rubbish (Macan 1977); whether this has washed in, or was
living in brackish pools and rubbish associated with settlement cannot be determined. The
terrestrial assemblage is sparse, but significantly richer than that from the second sequence (Table
8). The presence of Pupilla muscorum and Vallonia costata tends to indicate some drier habitats,
though both are recorded in floodplain meadows (Davies 1996; 1998). The molluscan evidence
does not amplify that provided by the pollen record.
A second sequence of 16 samples from the same section as the pollen monoliths (Fig. 13) and grain
size analysis was through the deposits between the two structures, and encompassed pre-
occupation, and immediately post-occupation deposits. Shells are absent in the lower, pre-
occupation, and upper (later) profile and not abundant throughout. The terrestrial component is
restricted to only two samples with Pupilla muscorum and one fragment of Cepaea sp. The aquatic
assemblage is dominated by Hydrobia, especially H. ventrosa. Most occur in the immediate post-
occupation deposits (layers 152 and 151) and all relate to pollen zone HALL2. This species is
predominant in brackish salt-marsh creeks away from the marine mudflat edge. The lack of any
other species may, however, indicate that these were washed in on tidal flooding. Numbers of shells
fluctuate through the profile; whether this is a factor of preservation or of episodic inwash cannot
be determined. This does, however, tend to indicate marine flooding or ingression in localised
creeks and reens.
This may represent a flooding episode at soon after 390–110 cal BC resulting from rising sea-level
and equating with the abandonment of the Iron Age settlements in peats marginal to marsh areas at
the Mere and Glastonbury lake villages.
67
Table 8. Mollusc data from Hallen Marsh
Between houses In house 1

Sample pre-occupat post occupation later pre- floor post occupa later
Sample 4 5 17 16 12 1 3 8 7 6 10 15 21 19 20 20 25
Sample 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1115 1116 1030 1031 1032 1033
Context/Feature 183 183 152 152 152 152 152 151 151 101 101 183 321 105 105 102 101
Depth (cm) 100- 90- 80- 80- 75- 70- 60- 50- 40- 30- 20- 80- 70- 60- 50- 40- 30-
110 100 90 85 80 75 70 60 50 40 30 90 80 70 60 50 40
Wt (g) 1750 1750 1750 1550 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750 1750
Pupilla muscorum (Linnaeus) 1 - - - - 3 - - - - - 1 2 1 - - 3
Vallonia costata (Müller) - - - - - - - - - - - - 1 - - - -
Cepaea spp. - - - + - - - - - - - - - - - - -
Land mollusc Taxa 1 0 0 0 0 1 0 0 0 0 0 1 2 1 0 0 1
LAND MOLLUSC TOTAL 1 0 0 0 0 3 0 0 0 0 0 1 3 1 0 0 3
Hydrobia ventrosa (Montagu) 1 1 5 4 61 13 58 19 57 18 - - 2 - - - -
Hydrobia ulvae (Pennant) - - - - - - 1 4 - - - - - - - - -
Hydrobia spp. - - - - - - - - - - 1 - - - - - -
Ovatella myosotis (Draparnaud) - - - - - - - - - - - - - - 2 - -
Aquatic Taxa 1 1 1 1 1 1 2 2 1 1 1 0 1 0 1 0 0
FRESHWATER TOTAL 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
MARINE TOTAL 1 1 5 4 61 3 59 23 1 18 1 0 2 0 2 0 0
TOTAL 2 1 5 4 61 16 49 23 57 18 1 1 5 1 2 0 3
Magnetic Susceptibility
Charcoals
by Rowena Gale
Charcoal was sparsely present in Iron Age contexts at Hallen Marsh. Two samples collected during
the excavation in 1992 were examined and identified to genus (Table 9). Samples were from an
oven and floor within the occupation area.
A few thin flakes of oak (Quercus sp.) charcoal were recovered from the floor occupation area of
House 1. The charcoal probably contained sapwood although it was too fragmented to be certain.
Charcoal fragments from an irregular patch north of the oven were exceedingly small and difficult
to examine. Oak formed the major component, with variable growth rates from average to slow.
Extraneous materials deposited throughout the wood cells prevented an assessment of maturity.
Table 9. Charcoal from the roundhouse at Hallen Marsh
Sample Context Description Quercus Other

1086 195 Floor occupation area 19 -
1089 311 Irregular patch north of oven 9 1 (see text)
The number of fragments identified is indicated.
Both samples were recovered from the House 1 and, by inference, the charcoal represents fuel
debris from a hearth within the round-house, and indicates the domestic use of oak. The settlement
at Hallen suggests that, during the Midddle–Late Iron Age, local woodland resources were
sufficiently plentiful to provision the building of round-houses and other structures. Although
speculative, these would almost certainly have incorporated oak timbers.
Evidence from mid-Flandrian levels at sites somewhat closer to the estuary than Hallen Marsh
demonstrates that dense woodland once flourished on the peat-levels bordering the coastline (Allen
1992a; see also Bell et al. 2000). At Grange Pill, Hills Flats, Oldbury Flats, Gold Cliff, and other
sites their remains are now exposed daily, lying in situ where they fell on the present-day foreshore
of the Severn Estuary. The well preserved slender trunks, crowns and root-systems estimated to be
68
roughly 50 years or so in age before succumbing to windblow, suggest successional woodlands of
relatively young trees. At this time the marginal wetlands and moist soils were probably dominated
by oak, alder (Alnus glutinosa), birch (Betula sp.), and willow/poplar (Salix sp./Populus sp.).
Evidence from Late Bronze Age/ Early Iron Age charcoal deposits from Avonmouth at Bristol
Sewage Works (Gale unpubl.) and Kites Corner (Walker et al. 1999a), and pollen from Cabot Park
(Walker et al. 1999b) and Rockingham Farm (Walker et al. 1998b), attest to local woodland
composed of oak, ash (Fraxinus excelsior), alder, yew (Taxus), holly (Ilex aquifolium), birch, the
hawthorn/Sorbus group (Pomoideae), maple (Acer campestre), elm (Ulmus sp.), hazel (Corylus
avellana), willow/poplar, and blackthorn (Prunus spinosa).
The Roman settlement at Rumney Great Wharf on the Welsh coastline of the estuary produced
evidence of wetland reclamation and a small-scale metalworking industry (Fulford et al. 1992).
Charcoal deposits were indicative of woodlands somewhat similar in character to those from the
Late Bronze Age/Early Iron Age sites discussed above, with oak, alder, hazel, elm, gorse/broom
(Ulex/Cytisus), spindle (Euonymus europaea), Prunus, and probably maple, member/s of the
hawthorn/Sorbus group, and willow/poplar (Gale 1992).
4. Northwick
The sequence
Observations of geotechnical pits during advance works in 1990 revealed the presence of ditches
containing Iron Age and Romano-British pottery (Russett 1990, 12). Evaluation test pits at these
locations recorded a Romano-British ditch (Lawler et al. 1992, 51–6) and postulated the presence
of a more extensive and long-established archaeological site. As with the archaeological
investigations at Hallen, large areas were stripped of topsoil and alluvium to expose the
archaeological site and, consequently, an area of c. 19,325 m² spread over three fields was
excavated in plan (Fig. 17). The site was preserved beneath extant ridge and furrow, so a
representative portion of the section has been described, though the surface undulation thus created
fluctuations in the depth of deposits.
The Romano-British site was only buried under less than 0.5 m of alluvium so the sequences
exposed are relatively shallow, but they relate well to the upper sequences recorded elsewhere.
Beneath the humic alluvial gley topsoil uniform brown and yellowish brown silty clay alluvium
occurred to depths of c. 0.4 m. This ‘upper alluvium’ (251) was lightly gleyed and mottled and
locally contained indistinct laminae (possibly flood couplets). A gleyed horizon (354)
approximately 0.1 m thick and with slightly higher susceptibility levels occurred at the base of this
sequence at c. 5.90 m OD and lay unconformably on the deposits into which the Romano-British
features archaeological features were cut. Higher susceptibility levels probably reflect enhancement
due to hydromorphism rather than pedogenesis (cf Allen & Macphail 1987). The archaeological
features were cut into a dark brown–brown silty clay with clear prismatic structure typical of the
‘upper alluvium’.
Although deep Romano-British ditches (up to 1.2 m depth) were excavated, these contained
sedimentary sequences largely local to the features and are not described here. A deeper sondage
was, however, recorded through the sediments to 2.0 m OD during the evaluation (Lawler et al.
1992, 51–4, fig. 9). This revealed a very narrow band of humified peat at c. 4.6 m OD with very
low pollen counts, beneath which were low energy estuarine alluvial deposits similar to those
reported elsewhere. At depth, however, c. 2.20 m OD was lower organic horizon overlying beach
69
WA Northwick
530
Trench 5
254
327
Trench 8
Trench 27
328
0 10 m
Figure 17. Northwick; summary plan of the excavations
70
WA Northwick WA Northwick GGAT G024
6 m O.D.
0 10 20
x10-8 SI/kg
5 m O.D.
4 m O.D.
3 m O.D.

Alluvial silty clays (colour unknown)
0 1m
2 m O.D.
Figure 18. Northwick Wentlooge Formation profiles
71
0-0.12 6.26-6.14 250 Dark brown (10YR 4/3) dark brown, silty clay loam Topsoil
with high organic matter, medium granular to
subangular blocky structure, many fine roots, clear
boundary, Topsoil
0.12-0.38 6.14-5.88 251 Yellowish brown (10YR 4/5) to brown (10YR 5/3) Upper alluvium &
silty clay loam, massive and structureless, but locally ridge and furrow
both weak medium blocky structure and very faint
laminae are present, and common diffuse ferruginous
mottling, rare small local stones, rare inclusion of
weathered and worn pottery, bone and charcoal with
clear to gradual wavy boundary
0.38-0.46 5.88-5.80 354 Grey (5Y 5/1) to brown (10YR 5/3) compact gleyed Upper alluvium
silty clay with coarse angular blocky to prismatic occupation interface
structure, 0.5% fine macropores, common to many
diffuse ferruginous yellowish brown (10YR 5/6)
mottles and common fine to very fine manganese
flecking, clear smooth boundary
0.46+ 5.80+ 316 Dark brown (7.5YR 4/4) to brown (10YR 5/3), coarse Upper alluvium
prismatic silty clay with common greyish hues and
grey (5Y 5/1) interped surfaces due to gleying, and
rare fine manganese flecking
sands and gravels. The relatively shallow sequence replicates the upper profiles at Awkley Lane,
Vimpenny’s Lane, and Hallen Marsh so sampling was analysis was restricted to land snails,
charcoals, and charred plants, largely from archaeological features.
Molluscs
by Michael J Allen
The sedimentary sequence was not sampled for snails at Northwick. Small samples were processed
but were devoid of shells. Instead two deep, alluvial filled Romano-British ditches associated with
field boundaries and land tenure were sampled.
Ditch 237
A sequence of seven samples was taken through the shallow ditch (0.6 m deep) and into the natural
alluvium through which it was cut. Although shells only survived in the upper profile, Romano-
British pottery is still prevalent in these contexts, but this occur above a significant peak in
magnetic susceptibility indicating deposits directly contemporary with human activity. Shell
numbers are typically low (Table 10) in the three samples and only one aquatic species (Ovatella
mytosis) is present which is more terrestrial than aquatic in its habitat (Kerney 1999, 46). The
remaining impoverished terrestrial assemblage includes Pupilla muscorum, Cepaea spp. and
rupestral Clausiliidae. It seems likely that the rupestral species represent damp habitats, possibly
even wet wood and leaf mould collected in the ditch and this is the niche exploited by O. mytosis.
Certainly C. bidentata is also tolerant to human disturbance.
72
Table 10. Mollusc data from Northwick
Feature ditch 254 dicth 327

TEMP Sample 12 13 10 11 1038 1039 1040
Sample 1008 1009 1010 1011 1038 1039 1040
Context 265 265 257 257 333 835 325
Depth (cm) 110- 110- 90- 80- 70- 60- 50-
120 110 100 90 80 70 60
Wt (g) 1400 1500 1600 1425 1750 1750 1750
Pupilla muscorum (Linnaeus) - - - - 2 13 4
Cochlodina laminata (Montagu) - - - - - - 5
Clausilia bidentata (Ström) - - 1 - 1 - 7
Helicella itala (Linnaeus) - - 1 - - - -
Trichia hispida (Linnaeus) - - 1 - - - -
Cepaea spp. - - + + + 2 2
Land mollusc Taxa 0 0 4 0 2 2 4
LAND MOLLUSC TOTAL 0 0 4 0 3 15 18
Bithynia tentaculata (Linnaeus) - 1 - - - - -
+ Lymnaea truncatula (Müller) - - 1 - - - -
Hydrobia ventrosa (Montagu) 9 - 1 - - - -
Hydrobia ulvae (Pennant) 1 16 - - - - -
Hydrobia spp. - 17 - - - - -
Ovatella myosotis (Draparnaud) - - - - - 3 -
Aquatic Taxa 2 2 2 0 0 1 0
FRESHWATER TOTAL 0 1 1 0 0 0 0
MARINE TOTAL 10 33 1 0 0 3 0
TOTAL 10 34 6 0 3 18 18
note: only samples with shells tabulated: totals exclude C. acicula
Ditch 254
In contrast, ditch was 1.5 m deep and 17 samples were taken in columns through the ditch deposits
and into the natural. Shells occur in low numbers in four samples (Table 10). These are at the top of
the primary fill (layer 265), and within the main ‘occupation’ related deposit; ie, that containing
most artefacts. The assemblages are also distinctly different from the shallow ditch 327. The
primary fill contains only aquatic species; the majority of which are brackish species (Ventrosa sp.)
with only one specimen of a freshwater taxa (Bithynia tentaculata). The presence of H. ulvae as the
dominat Hydrobia species in this feature is particularly interesting as this has a much more
estuarine and coastal preference than H. ventrosa. We may infer that this either represents near
coastal conditions or, more likely, that these are derived from habitats and washed in by local flood
events relating to rising relative sea-level (RSL) at this time. In the secondary fill, relating to
‘occupation’, however, aquatic species are minimal and a low number of terrestrial species not
previously present indicate localised open drier habitats locally.
Comment
There is a distinct variation between the assemblages from the two ditches; whether this is a factor
of depth (4.90 m OD vs 5.40 m OD), or time cannot be determined. As occupation and the ditches
seem, however, to cover a relatively short time span in the first century AD and probably not
extending much beyond a century, it is most likely that the local variation in depth, and use or
function of the ditches, rather than differences in time, account for the variation between the two
groups of assemblages.
73
5. Awkley Interface
The sequence
Four large machine cut soil pits (and one evaluation pit) were excavated along a transect of c. 125
m on the footslope of Awkley Hill (ST 895 863) to illucidate the relationship between estuarine
innundation and colluvial deposits on the north-west edge of the Avon Levels (Fig. 3). Trench 1,
almost at the foot of the slope, was only 130 m east of the deeply stratified mineralogenic silts and
peat sequence at Awkley Lane. The composite section (Fig. 19) can be extended into the back fen
to the Awkley Lane sequence by virtue of an evaluation trench and augerholes conducted by GGAT
(Lawler et al. 1992, 58–62).
The excavations revealed alluvial and colluvial deposits overlying Mercian mudstones (Kueper
Marl and Tea Green Marl). Evidence of limited Romano-British activity was present near the base
of the footslope. The basal deposits are highly variable forms of Kueper Marl and Tea Green Marl
bedrock with many mudstone fragments (Fig. 19). Weathered Mudstone occurred beneath the
Keuper Marl on the slopes of Awkley Hill itself. Deep freshwater fen peats occur below Awkley
Hill (Awkley Lane) but, as the sequence and bedrock rise above c. 4.2 m OD, they are absent.
Overlying the peat are sandy alluvial depoists (cf Crowther in Lawler et al. 1992, 98), with
moderate magnetic susceptibility readings (c. 20 SIu10-8Kg-1) with an increasingly colluvial
component both upslope and up profile. The higher magnetic susceptibility readings here than in
other alluvial sequences probably reflect the colluvial and terrestrial soil input. The Keuper Marl
bedrock forms a bench at the base of the slope and marks the location of former Romano-British
activity in the form of field boundary ditches, pottery and charcoal. This is also the location of the
former Old Awkley Lane and Awkley Road.
To the east, and upslope of this point, alluvial and estuarine deposits give way to coarse colluvial
soils derived from Awkley Hill. Magnetic susceptibility values generally remain similar to that of
the alluvium, but peaks (up to almost 60 SIu10-8Kg-1) clearly define horizons of archaeological
activity
Samples were removed from trenches 1 and 4 in columns of contiguous samples at 0.1m intervals
for soil magnteic susceptibility and land snail analysis. These sequences were nearer the base of
Awkley Hill, with the aim of examining any alluvial-colluvial interface. Bulk samples from the
archaeological features produced no charred remains.
74
Trench 1; base of slope
Depth (m) approx OD Layer Description Comment
0-0.14 7.25-7.11 100 Dark greyish brown clay loam Topsoil.
0.14-0.70 7.11-6.55 101 Weak red silty clay. One small sherd of Romano-British Alluvial /colluviual
pottery present
0.70-1.60 6.55-5.65 102 Grey silty clay, a single limestone slab recorded in layer. [GGAT E]
Layer sits in what appears to be a slight hollow
1.20-1.35 6.05-5.90 118 Light grey silty clay, occasional flecks of brown sand Alluvial
[GGAT D1].
1.35-1.60 103 Brown sandy clay, common mineral staining. Alluvial
[GGAT C1]
1.60-1.70 104 Grey/pinkish grey mottled sandy clay, very occasional ?occupation
charcoal flecks and small pebbles [GGAT C1]
1.70-1.75 105 Light grey sandy clay Alluvial
[GGAT C1]
1.75-1.85 106 Pinkish grey/brown mottled sandy loam Alluvial
[GGAT C1]
1.85-2.00 107 Red silty clay, thin striations of pinkish grey clay throughout Alluvial
layer [GGAT C1]
2.00-2.05 108 Greyish brown silty clay Alluvial
[GGAT C1]
2.05-2.65 109 Dark reddish brown sandy clay, occasional lenses of orange Kueper Marl
brown sand [GGAT A]
2.65-2.70 110 Weak red silty clay, layer is involuted and is not present Keuper Marl
across whole trench [GGAT A]
2.70-2.90 111 Light greenish grey silty clay with lenses of weak red clay, Tea Green/Keuper
layer is involuted and is not present across whole trench Marl
[GGAT A]
2.75-2.80 112 Weak red silty clay, thin lense of material only present in Keuper Marl
part of trench [GGAT A]
2.80-2.82 113 Light greenish grey gleyed silty clay, thin lense of material Tea Green Marl
only present in part of trench [GGAT A]
2.90-3.20 114 Reddish brown (5YR 3/4)/weak red sandy clay mottled Keuper Marl
material, layer is involuted [GGAT A]
3.20-3.25 115 Light greenish grey (5G 6/1) gleyed silty sand, layer is Tea Green Marl
slightly involuted [GGAT A]
3.25+ 116 Weak red silty clay Keuper Marl
[GGAT A]
Trench 3
0-0.10 300 Very dark greyish brown loam, occasional limestone, chalk Topsoil.
and charcoal flecks.
0.10-0.25 301 Reddish brown sandy loam, occasional charcoal flecks.
0.25-0.60 302 Reddish brown sandy loam, veins of crushed limestone and colluvial
some mineral staining.
0.60-0.80 303 Brown silty loam, occasional charcoal flecks, only appears buried soil.
in N-W end of trench.
0.60-1.10 304 Weak red silty clay loam, lenses of pink coarse silty loam. colluvial
0.75-1.10 305 Dark red silty clay loam, occasional lenses of pink sand. colluvial
Layer only occurs occasionally through section.
F312 Small cut 0.35m deep only seen in section, filled with ?post-hole.
reddish brown silty clay loam with occasional limestone and
mineral staining
1.10-1.20 306 Reddish brown silty clay loam, lenses of sandy material and colluvial/weathered
occasional mineral staining. Keuper Marl.
1.20-1.40 308 Contorted reddish brown silty clay loam, thick veins of Keuper Marl.
greenish grey material and veins of crushed limestone run
diagonally through layer into 309 below.
1.40-2.10 309 Contorted reddish brown silty clay loam, common veins of Keuper Marl
greenish grey material and crushed limestone running
diagonally through layer.
2.10-2.35 310 Reddish brown/dark red coarse dandy loam with up to 50% Keuper Marl &
mudstone. Mudstone.
2.35+ 311 Dark red sandy clay loam with up to 70% mudstone. Mudstome
75
Trench 4
0-0.20 8.18 409 Very dark greyish brown (10YR 3/2) loam, occasional Topsoil
manganese frags and crushed limestone pieces. Diffuse
wavy boundary
0.20-0.50 410 Brown (5YR 5/2) fine sandy loam, occasional manganese
frags and crushed limestone pieces. Gradual boundary
0.50-0.80 411 Reddish brown (5YR 4/3) fine sandy loam, very occasional Largely colluvial
manganese frags and crushed limestone and small stones. [GGAT E]
0.80-0.95 400/ Dark reddish grey (10YR 4/2)/brown (7.5YR 4/4) silty loam Occupation layer?
412 to to silty clay loam, common medium limestone pieces, 13 = GGAT 396 [D1]
sherds of Romano-British pottery recovered. Gradual to
diffuse boundary
0.95-1.20 413/ Reddish brown (10YR 4/3 - 10YR 5/3) coarse sandy clay Occupation deposit ?
423 loam. Occasional manganese staining rare limestone pieces [GGAT C1]
and occasional charcoal and fired clay, with occasional
lenses of olive green – pale yellow (2.5YR 7/4) towards at
base. Sandier than 412 above
1.20-1.30 422 Reddish brown (5YR 5/3) fine sandy loam, occasional [GGAT C1]
mineral staining, mottled with pink and yellow sand.
F405 Irregular cut feature up to 0.75m deep only seen in section. ?Ditch
Probbaly a ditch, filled with reddish brown silty and sandy
loams over a dark reddish brown silty clay loam with olive
brown lenses.
F428 Irregular cut feature up to 0.70m deep only seen in section. ?Ditch
Probably a ditch, filled with yellowish red coarse silty loam
with occasional charcoal and cut by F405.
1.30-1.35 407/ Brown silty loam with frequent limestone and charcoal Occupation layer
425
F401 Broad shallow 0.50m deep ditch only seen in section, filled ?Ditch
with dark reddish brown silty clay loam with occasional
charcoal and limestone fragments, cut by F428
F402 Irregular cut up to 0.55m deep only seen in section, filled ?Pit
with dark brown silty clay loam with occasional charcoal
1.30-1.50 415 Light brown (5YR 6/4) loamy sand, mottled with reddish alluvial
yellow (7.5YR 6/4) sand, occasional mineral staining and =GGAT 401.
charcoal. Gradual boundary [GGAT C1]
1.50-1.70 416 Reddish brown (2.5YR 4/4) sandy clay loam, very alluvial
occasional small stones and mineral staining, distinguished = GGAT 403
by presence of manganese and crushed limestone fragments. [GGAT C1]
Gradual boundary
1.70-2.00 426 Reddish brown silty clay with veins of grey green clay alluvial
= GGAT 404 [GGAT
C1]
2.00-2.10 417 Red (2.5YR 4/6) fine silty clay loam with veins of grey alluvial
green clay and lenses of pinker material. = GGAT 404
2.10-3.00 418 Reddish brown (2.5YR 4/4) /dark red (2.5YR 3/6) coarse Mudstone,
sandy loam: mixed interleaved layers of mudstone and Keuper Marl [GGAT
coarse loamy sand. A]
3.00+ 419 Greenish grey (5GY 6/1) sandy clay. Tea Green Marl
[GGAT A].
76
WA Awkley Lane WA Awkley Interface
NW SE
Awkley Hill
12 m O.D.
Occupation Zone
11 m O.D.
10 m O.D.
Tr.2
Old
9 m O.D. Awkley Lane
Tr.3
Tr.4 A
8 m O.D. Awkley Lane
Tr.1 C
A
7 m O.D.
B D
6 m O.D. C
A
5 m O.D. D
A Keuper Marl Mudstone

4 m O.D.
B
3 m O.D.
Easturine Sediment
2 m O.D. Peat
1 m O.D.
0 100 m
Figure 19. Awkley Interface composite transect
77
Molluscs
by Michael J. Allen
Shells were absent from all samples of alluvium from trench 1 and were were sparse from those in
trench 4, only being present in six samples (Table 11) and were all restricted to the upper 0.8 m of
the colluvial profile. The depauperate assemblages were wholly terrestrial, with only one specimen
of fresh or brackish water species present, confirming the absence of developed salt-marsh and
inundation in the sequence with surviving shells. The low numbers of shells make any detailed
palaeo-environment comment impossible. Shells were recovered from the latest and uppermost
mixed layer of the putative Romano-British occupation deposits (layer 412) and from sandy loam
colluvial sealing it. The main occupation layer (413) and all deposits beneath it were devoid of
shells. Most of the assemblage is typical of open long, vegetation with a dense sward proving local
mesic habitats at ground level. Some species tolerate marshes or damper habitats such as Vallonia
pulchella and Vitrea crystallina. One specimen in particular, however, is worthy of note; Vertigo
lilljeborgi. This rare species, which has never before been recorded in southern England (Kerney
1999, 97), is tolerant of acid conditions and is restricted to ‘saturated decaying vegetation in Carex
and Juncus swamps, sometimes shaded by alders’. Although it has mainly been found on the
margins of highland lakes, it has occasionally been found in small isolated mires.
The limited assemblages here may indicate essentially terrestrial land on the margins of damper
land with mires and possibly salt-marsh in Roman and post-Roman times.
Table 11. Mollusc data from Awkley Interface, trench 4
Sample 1035 1036 1037 1038 1039 1040 1041 1042 1043
Context 412 411 411 411 411 410 410 409 409
Depth (cm) 75- 70- 60- 50- 40- 30- 20- 10- 0-
80 75 70 60 50 40 30 20 10
Wt (g) 1750 1750 1750 1750 1750 1750 1750 1750 1750
Vertigo lilljeborgi (Westerlund) - 1 - - - - - - -
Vallonia costata (Müller) - - - - 3 2 2 - -
Vallonia pulchella (Müller) no shells cf 1 - - 1 - - - - -
Vallonia spp. below 80cm 1 - - - - - - - -
Discus rotundatus (Müller) - - - - - - + - -
Vitrea crystallina (Müller) - - - - - - 1 - -
Aegopinella nitidula (Draparnaud) - - - - - - + - -
Limacidae - - - 4 - - - - -
Cecilioides acicula (Müller) - - - - - - 2 - -
Trichia hispida (Linnaeus) - - - - - - + - -
Land mollusc Taxa 2 1 0 2 1 1 2 0 0
LAND MOLLUSC TOTAL 2 1 0 5 3 2 3 0 0
Bithynia tentaculata (Linnaeus) - - - - - 1 - - -
Aquatic Taxa 0 0 0 0 0 1 0 0 0
FRESHWATER TOTAL 0 0 0 0 0 0 0 0 0
MARINE TOTAL 0 0 0 0 0 1 0 0 0
TOTAL 2 1 0 5 3 3 3 0 0
Comment
Most of the Holocene sedimentary sequence exposed on the slopes of Awkley Hill is Romano-
British or later in date. Only at the flootslope do these layers seal colluvial or alluvial deposits
which remain undated. Despite sampling these deposits, few microfossils were found and no real
evaluation of the environment of the colluvial–alluvial interface could be determined, except that
marsh (snails) existed locally.
78
DISCUSSION and REVIEW
The physical and vegetational evolution of the Avon Levels
Stratigraphic overview
The Wentlooge Formation
The Wentlooge sequence, as described by Allen and Rae (1987), provides a basis for the
interpretations presented here. Detailed analyses at a number of sites has generated local palaeo-
environmental interpretations which, in conjunction with the large-scale auger survey of the
superficial deposits, provide the opportunity to extend these interpretations laterally. In so doing an
attempt is made to describe the development of the Avon Levels as a whole.
The excavations described have provided detailed local palaeo-environmental reconstructions. In

general terms some consistency in interpretations can be seen within the major facies of the
Wentlooge Formation between sites, although some local variation is clearly evident. Because of
the nature of both of the two deeper excavations (limited to 5 m depth and exposing the middle and
upper Wentlooge Formations), the archaeological excavations and the data from the auger survey
(the latter limited to 2 m depth), only the upper facies of the Wentlooge Formation were examined
(ie, middle and upper Wentlooge). Nevertheless, this information includes geomorphological,
sedimentological, biological, and archaeological elements with which to examine the Late
Holocene evolution and use of the Avon Levels and to review and re-examine the nature of the
Wentlooge Formation and the development of the Severn Estuary Levels. When the palaeo-
environmental data is combined with the auger survey a much more comprehensive picture of the
entire Avon Levels can be suggested.
The lower Wentlooge Formation (Mesolithic)

Deposits relating to this facies were at great depth and were only encountered as basal deposits
(Awkley Lane), or in exploratory auger holes through the base of deep excavations (Vimpenny’s
Lane, trench 1). These facies do, however, occur more accessibly as peat ledges (Allen 1990c)
along the coastal margins of the Avon Levels (eg, Gravel Banks) and have been studied and
reviewed by Druce (2001).
The middle Wentlooge Formation (later Mesolithic to Bronze Age)

Interleaved deposits of peats and minerogenic clays were only truly encountered at Awkley Lane.
The lower minerogenic sequence and narrow detrital peat bands at Vimpenny’s Lane also belong to
the very top of the middle Wentlooge Formation. These represent freshwater fen peats and reed
swamp which are inundated by minerogenic sediments in flunctating salt-marsh and mudflat
conditions.
Boundary of the middle and upper Wentlooge Formations

A clear boundary exists between the humic peats and lower minerogenic deposits, and the upper
minerogenic deposits. This can be detected on physical criteria (characteristics such as colour and
structure), and in terms of pollen zone boundaries (Fig. 22, below). The humic peats developed
under carr (birch woodland) with local acid peat communities. The lower minerogenic deposits
represent brackish/estuarine inundation at c. 2250 cal BC (4000 BP) in a variety of local salt-marsh
and mudflat regimes. Although the upper minerogenic deposits are also silts and silty clays
deposited under salt-marsh and mudflat conditions they indicate deposits which contain more
fluvially derived sediments, with ferrous and manganese minerals giving rise to strong reddish and
brown hues on gleying. Unlike the lower minerogenic deposits, they rarely contain any fluvial
structures such as flood couplets or laminae, indicating that such features have been destroyed by
79
pedogenesis and biotic reworking. Thus, although this boundary is physically clearly defined, the
change of environmental deposition between the upper and lower facies is not as clear. Its
stratigraphic and chronological location is, however, extremely significant in relation to occupation
of Iron Age and Romano-British date. It is discussed below as it relates to one of the main phases of
late Holocene positive relative sea-level rise occurring in the later Bronze to Romano-British
periods.
Upper Wentlooge Formation

Heavy gleying with strong iron and manganese staining gives the upper Wentlooge Formation its
distinctive reddish-brown hue, making it easily distinguishable in the field and in auger survey. It is
uniformly later Iron Age and later, but its inception, although undated, is probably asynchronous
across the region. Nevertheless, by the later Romano-British period, the upper Wentlooge
Formation is recognised throughout the wetlands. This formation overlies later Iron Age features at
Hallen Marsh, and encompasses later Romano-British ones at Northwick. Later medieval and post-
medieval ridge and furrow are the latest ubiquitous topographic forms within this deposit.
Sea-level change and palaeo-geographic reconstruction
The peat and mineral sediment sequences of the lowlands surrounding the Bristol Channel contain a
valuable archive of the local and regional changes in Holocene sea-level. The sediment sequences
were influenced and/or accreted directly in response to changing RSL and the inter-relationships of
peat, sediments, and their heights relative to present Ordnance Datum can provide data on the sea-
level changes which have taken place during the Holocene period. Radiocarbon dating and
microfossil analysis of the peat and microfossil analysis (especially pollen and diatoms) of the
estuarine silts/clays has allowed some deductions regarding the position of the palaeo-environment
and palaeo-geography.
Whilst the late-Devensian glacial and early Holocene period to c. 7000 BP globally saw sharply
rising sea-levels from -120 m OD (Fairbanks 1989), the rate subsequently slowed, resulting in many
and more local factors such as local neotectonics and variations in coastal dynamics (Long et al.
2000) having a greater impact on the spatial aspects and on the middle and late Holocene sea-level
and shoreline position. It is, however, clear that general expansion of RSL to the present day has
altered the configuration of the coastline and as such, the character and distribution of human
resources along the coastal and estuarine fringe.
Reviews of sea-level change in this region have been given by Hawkins (1971; 1973) Work has
concentrated on the Bristol Channel region on both the English and Welsh sides and has a long
history dating back to Reid (1913, 129), Godwin (1945), and Churchill (1965). More recent studies
include Balaam et al. (1987), Kidson and Heyworth (1973; 1978), and Heyworth and Kidson
(1982). Published sea-level curves for the region are available and include those of Churchill
(1965), Balaam et al. (1987), Heyworth and Kidson (1982), Scaife and Long (1994), Canti et al.
(1993), Haslett et al. (1997), and Haslett et al. (1988). Figure 21a is a synthesis of data obtained by
Kidson and Heyworth which incorporates sea-level change with datum points obtained from the
Welsh side of the Second Severn Crossing at Caldicot Pill. Shennan (1989), examining neotectonics
around Britain’s coastline, suggested that the Bristol Channel is/was experiencing crustal
subsidence at a rate of between c. -0.2 mm to -0.5 mm OD per year; that is, substantially less than
other regions of the country with postive changes (the uplifting North) or areas of greater tectonic
subsidence such as the southern North Sea Basin and Thames Estuary (Devoy 1977; 1979; 1980;
1982) where net sea-level rise has been greater than the study region. For the purposes of this study,
the work of Kidson and Heyworth (1978) and especially Heyworth and Kidson (1982) are most
80
relevent to the early and middle Holocene changes which resulted in the formation of the basic
coastal morphology, whilst late Holocene changes have been addressed by Allen (1987a) and Allen
and Rae (1988). After these sharply rising sea-levels, and sedimentation seen in the Somerset
Levels, there was a retreat evidenced at many sites along the Bristol Channel (eg, at Caldicot Pill;
Scaife & Long 1994; and Westward Ho! Balaam et al. 1987). This retreat resulted in the
establishment of a peat forming reed-swamp and fen carr. It is from this point in time that studies
here commence with the basal peat sequence dated to c. 3800 cal BC (5000 BP) at Awkley Lane.
There has, however, been some debate, with Heyworth and Kidson (1982) arguing in favour of a
smooth exponential change in sea-level through the Holocene in this and other regions (Jelgersma
1961; 1966; Jardine 1976). Heyworth and Kidson (1982) suggest that after rapidly rising sea-level
during the early and middle Holocene, there was a substantial slowing down between 4000 and
3000 BP, after which any increase has been minor. They suggest that any such oscillations of sea-
level are of a lesser magnitude than the inherent uncertainties associated with typical studies of
relative sea-level change. This view contrasts with other workers favouring the evidence for an
oscillating sea-level which gave rise to alternating sequences of peat and mineral sediment as for
example seen in the Lower Thames (Devoy 1979; 1980; 1982). Whilst this is, perhaps, the case
when studying the typical inter-relationship between the principal peat facies intercalated within
estuarine/marine sediment units here, examination of the latter in more detail suggests that some
further resolution of sea-level and environmental fluctuations may indeed be represented.
Allen (1990a; 1990b), Allen and Rae (1988) and Haslett et al. (1998) argue that RSL continued to
rise throughout the historic period at 0.41–0.82 mm per year after 3000 BP. Woodworth (1987)
shows a continuing annual rate approaching 2 mm based on tide gauge evidence. This continuous
change resulted in continued sedimentation in areas not subject to Romano-British or later
reclamation or land division schemes of 1.58–1.92 mm per year maximum and 0.8–0.96 mm per
year minimum. Allen and Rae (1988) also suggest rates of accumulation of 1.22 m since the
Romano-British period, 1.05 m since the medieval period, and 0.21 m since 1945. Where
reclamation and land division has occurred (?Nyland Hills in the Somerset Levels) an
anthropogenically induced negative RSL has resulted.
It has been argued that the typical grey silts accumulating in salt-marsh and estuarine habitats may
be unsuitable for establishing past sea-levels because salt-marsh halophytes with the ability to grow
above the ground water table can be influenced by changes in RSL and local ground water
conditions (Heyworth & Kidson 1982), which may themselves reflect very local changes in coastal
morphology due to, for instance, the formation and breakdown of sand barriers, and the substantial
effects of sediment compaction (Haslett et al. 1998). Thus, there are clearly problems in using
sediment data as a proxy for establishing past relative sea-level. Allen (1992b), however, points to
the undue emphasis placed on the intercalated peats whereas little is known about the tidally
influenced marsh deposits of the Wentlooge Formation. The standard sea-level curves referenced
above are based on radiocarbon dated peat/sediment contact horizons which mayor may not have
had biostratigraphical/multi-proxy verification of their eustatic relevance. Within this continuously
rising sea-level change curve is hidden evidence of minor RSL oscillations (Haslett et al. 1997).
This is similarly in evidence in this study. Problems of sediment compaction, lack of understanding
of sedimentation deposition processes in salt-marshes, and problems of obtaining absolute
(radiocarbon measurements) from the predominantly minerogenic sediments negates the
reconstruction of such minor fluctuations from the pollen data. This is further compounded by the
fact that the Severn has the largest tidal range in Britain (Shennan 1982).
In this study, major changes brought about by positive RSL are clearly in evidence in the change
from the peats of the middle Wentlooge at Awkley to the overlying upper Wentlooge sediments
deposited under marine/estuarine conditions. Two sea-level data points have been obtained: at
81
Awkley Lane the determination of 4044±50 BP (NZA-12533) at a height of +3.14 m OD and at
Vimpenny’s Lane a determination of 4182±55 BP (NZA-12527) at a height of +2.78 m OD. When
plotted within the existing framework of Haslett et al. (1997) it can be seen that the datum points
clearly fall the general curve of RSL change (Fig. 21a). When corrected to mean high water spring
tide (based on data from Minehead) the former gives an OD of -2.42 m OD and the latter at -2.06 m
OD. When plotted on the corrected RSL graph for Caldicot Pill (Scaife & Long 1994) on the north
side of the Severn Crossing (Fig. 21b) it can be seen that the points lie above the general curve of
Heyworth and Kidson. These points however, compare with data added for Caldicot Pill which are
also on the high side. However, both datasets, when added to the Kidson and Heyworth curve,
mirror the envelope of change. This phenomenon has also been noted by other recent workers in the
region (A. Long pers. comm.).
Detailed pollen analysis of the upper Wentlooge sediments at the three sites (Vimpenny’s, Awkley,
and Hallen; below) provide evidence for fluctuations between salt-marsh and mudflat conditions
after the principal late Holocene marine incursion. In terms of vertical displacement this equates to
some 1–2 m difference between mudflat and lower salt-marsh typically separated by a cliff (Allen
1989 suggests an alarming 5–10 m in places!). Given that human utilisation of a stabilised salt-
marsh is now in evidence at Hallen, any such minor oscillations may have had a significant effect
on these human activities.
Patterns and consistency; the mapped sediments
The auger survey successfully mapped a reddish brown alluvial unit over a greyish alluvial unit.
These were clearly distinguishable by colour in the field. Although the survey recognised and
recorded distinct peat and buried soil or stasis horizons, the survey was less able to map these as
they were not extensive surfaces which could be linked and modelled (in 2001) using QuickSurf.
Stabilisation surfaces and deep peats were highly localised in their occurrence as pools or shallow
mires within the salt-marsh, or fringing the higher dryland at the east of the Avon Levels. The
extent of deeper peats was not mapped by the auger survey which was restricted to the 2 m of the
sedimentary profile. Nevertheless, peats of varying thickness (middle Wentlooge Formation) exist
over much of the North Avon Levels as determined from numerous commercial bore holes (Russett
1990; Druce 2001; Allen 1990c)
Although the main area of the wetland is both low lying and relatively flat, subtle variations in the
present day and prehistoric surface altitude can be seen from the records of excavated sites. Buried
soils and sub-surface stabilisation horizons at depth are obviously related to this and as can be seen
from Figure 20. Furthermore, some individual surfaces have a more complex buried topography
than previously envisaged. The most obvious of these is at about 2 m below the present surface and
fluctuating about the 4.5 m OD height. Indeed, an organic stabilisation horizon can be detected at
this depth at a number of locations and is sometimes referred to as the ‘BARAS layer’ after its first
record by the Bristol & Region Archaeological Services (BARAS 1998). Examination of Figure 20
shows that sections at Awkley Lane, Vimpennys Lane, Avonmouth Sewage Treatment works,
Hallen Marsh, and Rockingham Farm all indcate stasis horizons at about this height. These vary in
their nature and superficially may be considered to be a single unit. Radiocarbon determination on
horizons at all of the noted sites, however, contradict this. Dates vary from 5880–5660 cal BC at
Avonmouth to 2900–2620 cal BC at Vimpenny’s Lane, to 1190-810 cal BC and 1100-810 cal BC at
Rockingham Farm and Avonmouth respectively (see Table 12 and Fig. 22). Indeed, the
82
Awkley Vimpenny Vimpenny Avonmouth Hallen Rockingham Cabot
Tr.1 Tr.2 Farm Park
6 m O.D.
2192+_ 55BP
2778+_ 55BP 2810+_ 70BP
5 m O.D. 2957+_ 55BP 3040+_ 60BP
5879+_ 70BP
6380+_ 50BP
4170+_ 70BP
4 m O.D.
3 m O.D. 4182+_ 55BP

3816+_ 50BP
3991+_ 55BP
4044+_ 60BP
4745+_ 50BP
2 m O.D.
_
5603+50BP
1 m O.D.
?Stabilisation horizon
0 m O.D.
Figure 20. Relative heights of soils and stasis horizons from various locations
83
Table 12. Stasis horizons in the Middle Wentlooge sequence in relation to their OD height
and radiocarbon determination
Site layer mOD result date Cal description

BC
Rockingham 5.7 2850±40 BP, Beta-134902 1190–910 upper gleyed layer
Farm
Later Bronze Age soils with archaeological activity

Avonmouth 523 5.3-5.25 2778±55 BP, NZA-12725 1110–810 Pale blue clay, incipient
2957±55 BP, NZA-12726 1370–1010 stabilisation horizon, arch
Rockingham 729 5.20-5.10 2810±70 BP, Beta-118379 1190–820 thin bands of organic material
Farm 3040±60 BP, Beta-118378 1430–1120
Kites Corner 462 5.1 2610±70 BP, Beta-129554 920–520 charcoal patch archaeological
activity
Little Googs/ 5.1-5.2 2970±60 BP, Beta-134900 1390–1020 non humic soil
Kites Corner 3350±60 BP, Beta-134901 1870–1500
BARAS layer
Cabot Park 162-4 4.5 3970±60 BP, Beta-125795 2830–2290 organic clay 'BARAS' charcoal
4170±70 BP, Beta-125794 2900–2570 - activity in vicinity
Seabank site ‘BARA 4.7 3930±50 BP, Wk-5804 2580–2280
S’
Avonmouth 525 4.85-4.75 5879±70 BP, NZA-12478 4930–4550 Dark bluish grey humic clay
6866±50 BP, NZA- 12495 5880–5660
Other pre-Late Bronze Age stasis horizons

Awkley Lane 107 4.61-4.46 R26327/2 no result Dark grey (10YR 4/1) stonefree
clay with up to 10% flecks of
black material.
Vimpennys Lane 207 4.20-4.05 4182±55 BP, NZA-12527 2900–2620 Very dark grey (10YR 3/1)
stonefree slightly silty clay,
humic/organic layer
organic horizon at Awkley Lane at c. 4.5 m OD which may appear equivalent to that at Vimpenny’s
Lane, lies 1.5 m above a surface dated to 2840–2310 cal BC and is probably Roman or later in date.
Two major alluvial units were recognisable in most of the excavations described (named ‘Upper
Alluvium’ and ‘Grey (lower) Alluvium’). One of the major concerns was whether this colour
differentiation was a post-depositional effect of gleying and reduction creating ‘false’ stratigraphy.
If this was the case then the significance of its mapping and of any lateral interpolation of palaeo-
environmental information that may derive from it is greatly reduced. In order to verify the nature
and significance of the two units, the general physical and biological characteristics of both were
examined.
Characteristics
Visually in the field the two units are differentiated by colour, rarely by structure, and by the
occasional presence of weak laminae and possible flood couplets in the lower, grey alluvial unit.
Both horizons were largely massive (unstructured), excepting the localised occurrence, or
preservation, of laminae and flood couplets. In the field no major textural difference was observed
between these two units and this was verified by grain size analysis (most variation occurs within
them). The colour change is not related to depth of burial (compare Hallen Marsh and Awkley Lane
or Vimpenney’s Lane), nor are the observed colours likely to be entirely a result of post-
depositional redox processes that have overprinted the depositional strata. During pollen analysis,
major zone boundaries were defined independent of the stratigraphy. In every case the main pollen
zone boundaries lie on, or close to, this colour change. We can conclude that the observed colour
84
change may reflect some consistent change in the depositional environment. Although this change
occurred at between c. 4.1 m OD (Awkley Lane and Vimpenny’s Lane), and 4.5–4.7 m OD (Hallen
Marsh), and 5.5 m OD (Avonmouth), in every case they occur stratigraphically at about dated, or
assumed, Iron Age–Romano-British horizons. The informal nature of the ascription of the divisions
of upper, middle and lower Wentlooge Formation do not necessarily make it possible to define
precise boundaries between these divisions. However, in this case the colour change might relate to
the boundary between the middle/upper Wentlooge Formation which would occur at about this
point in the sedimentary sequence (though it should not be taken as a definite chronostratigraphic
unit). Thus, this boundary may map both a broad chronological marker and ecological change
through large area of the Avon Levels. This boundary, therefore, seems to relate to the main phase
of post-glacial/late Holocene positive RSL; ie, Late Bronze Age–Romano-British innundation.
Environmental change of the Avon Levels
Pollen taphonomy and interpretation
In the past, attention has tended to be been focused on the pollen analysis of peat intercalated in
estuarine sediments (Allen 1992b) and the radiocarbon dating of the upper and lower transitions to
silt for the purpose of establishing sea-level age/datum points. Because of the generally low APF
values of the mineral sediments and the complex pollen taphonomy, the majority of studies have
ignored these sediments and their palynological potential for environmental reconstruction. Allen
(1992b) similarly points to the undue emphasis placed on peat whereas there is little data for the sub-
fossil, tidally influenced marshes. However, study of these mineral sediments in the Avon Levels,
and also recently from the Solent (Scaife in Tomalin et al. forthcoming) and North Kent (Scaife
2001) suggests that useful environmental data can indeed be gained. This is especially pertinent
where there is a need to understand the on-site/local environment in relation to human activity. This
is becoming increasingly relevant with recent studies of prehistoric activity and impact on the coastal
zone.
The predominantly minerogenic character of the Upper Wentlooge sediment sequences (largely grey-
estuarine silts) clearly implies that any pollen contained will have a complex taphonomy having
come from a variety of different sources along with the sediments. (Allen 1990b; 1991). These
sources may include the following.
x Pollen fluvially transported from some distances up the river catchments entering the Avon
Levels with the sediment (Grichuck 1967; Hall 1981; Scaife & Burrin 1992).
x Reworking of older sediments (from soils of the interfluves and/or on-site sediment) containing
pollen. The more robust forms may remain in sediments for a long time. Pre-Quaternary
palynomorphs are a good indication of this, although the presence of degraded/reworked
Holocene Tilia is perhaps a better indication of the problem
x ‘Natural’ airborne pollen sources (cf. Tauber 1965; 1967; Moore et al. 1991; Scaife & Burrin
1992)
x Pollen from ‘on-site’ ie, autochthonous salt-marsh and fen mire vegetation
x Wetland taxa also fluvially transported from the river catchments including pollen evidence of
carr woodland and aquatic megaphytes such as Myriophyllum and marginal aquatic fen taxa, also
algal Pediastrum.
As a consequence of the taphonomic complexity, the generally small APF values and the rigorous
pollen extraction procedures required, there have been few attempts to carry out full/detailed pollen
analysis on the typical grey inorganic sediments. Discussion regarding the taphonomy of pollen,
spores and other microfossils (diatoms, dinoflagellates, pre-Quaternary palynomorphs) in fluvially-
85
derived sediments have attempted to produced models as an aid to understanding this complexity (see
for example Burrin & Scaife 1984; Brown 1985; Cundill & Whittington 1987; Scaife & Burrin
1992). In spite of these taphonomic problems and lack of research data associated with the pollen in
salt-marsh sediments and alluvial deposits, the presence of pollen can provide data on the local and
regional vegetation changes (Clarke & Patterson 1984; Jennings et al. 1993; Long et al. 1999). The
pollen data obtained from the Vimpenny’s Lane site (and at Awkley Lane and Hallen) further
demonstrate that useful information pertaining to local and regional environmental changes can be
gained from areas where perhaps, more typical or satisfactory peat sequences are not available to the
analyst.
Clearly, these factors have to be considered when interpreting the pollen spectra obtained from these
sediments and similarly at Awkley Lane and Hallen Marsh. Because fluvial transport (freshwater,
estuarine, and ?marine) may have transported pollen for substantial distances, the term ‘local pollen
assemblage zone’ is not used and primary pollen zones recognised have been termed ‘pollen
assemblage zones’. However, on-site (autochthonous) changes in vegetation ecology and
environment of deposition are also present and discussed where relevent (local pollen assemblage
sub-zones). These aspects relate principally to the changing status of the local marsh habitat. The
stratigraphical change between peat formed under wet fen and fen carr conditions such as at Awkley
Lane is reflected strongly in the pollen spectra first, by a marked reduction in absolute pollen
frequencies in the sediments contrasting with the very high APF values of the fen peat and, secondly,
there is an expansion of taxa which are typically over-represented or more abundant in fluvially
derived sediments. These include:
i) Pinus (pine) which is frequently over-represented in marine sediments.

ii) Degraded grains of Tilia (lime) most probably derived from soils on the interfluves and
eroded sediments.
iii) Fern spores, especially Pteridium aquilinum (bracken), but also monolete Dryopteris type
(typical ferns) and Polypodium vulgare (common polypody). The former is diagnostic (Peck
1973).
iv) There are also substantial increases in pre-Quaternary palynomorphs and Dinoflagellates, the
latter which may be marine/brackish water Holocene and as well as from older, derived
geological microfossils.
Furthermore, variations in the pollen spectra within the minerogenic marsh sediments have become
apparent from the sequential analysis of long profiles. In addition to the essential changes noted
above, are variations in the autochthonous component from which data on the changing environment
(and thus RSL) can be suggested. These changes relate to the accretion of sediments under salt-marsh
(upper, middle, and lower) and mostly unvegetated mudflat. Such division between salt-marsh and
mudflat is of importance since there may have been several metres difference between them due to
typical stepped/salt-marsh cliff profiles along channels and the estuarine fringe. Allen (1989) notes
that such salt-marsh cliffs are ‘strong and tall’ and in some cases up to 5–10 m in places. This will
also have been a significant factor in human activity as suggested from the archeology and
abandonment of the Iron Age site at Hallen.
Pollen is present in these sediments coming from the on-site (autochthonous) plant communities
which may, as noted, enable differentiation between mudflat, salt-marsh and shingle communities.
Mudflat: This appears to be characterised by substantial numbers of reworked pre-Quaternary

palynomorphs derived from eroded and/or reworked geological strata and/or Holocene alluvium.
There are also frequently substantial numbers of Holocene spores, eg, bracken and monolete spores.
All of these taxa are very resistant to decay. Also within the mudflat sediments are freshwater aquatic
taxa and algal Pediastrum derived from rivers discharging into the estuaries (frequently also
86
evidenced in diatom assemblages). Numbers of Pinus grains usually increase because of its over-
representation in fluvial environments (it has air sacs which give buoyancy). The dominant
autochthonous pollen comes from Chenopodiaceae which produce substantial numbers of pollen
grains and may come from on-site glassworts (Salicornia).
Salt-marsh: Pollen spectra show smaller percentages/APF values of all elements noted above for
mudflats. In contrast, there is a better representation of salt-marsh taxa with poorer pollen production.
Thus, typically Spergula type, Plumbaginaceae (Armeria and Limonium), Triglochin maritimum,
Aster type (ie, Aster tripolium), Plantago maritima and pollen of halophytic grasses (eg, Spartina)
with large (>45 m with thin exine and small pori, ie, not cereal).
Upper Salt-marsh, dune, and shingle habitats: These are less well represented but with occasional
Hippophae rhamnoides, Calystegia sepium, and possibly Liliaceae (possibly coming from Scilla
verna, S. autumnalis Hyacinthoides non-scripta). Some Chenopodiaceae may come from
Chenopodium (goosefoots) and Atriplex sp. (oraches).
It should, however, be noted that many of the pollen taxa may contain more than one taxon and
interpretations may be based on suites/assemblages of such types to differentiate plant
communities/habitat. Many pollen taxa recorded are also of catholic distribution and thus not
definable to specific habitats.
The late prehistoric vegetation of the Avon Levels
There is now a substantial corpus of palynological data from the Bristol Channel and Severn regions
of North Somerset, Avon, and south Wales from which much information on the development of late
prehistoric vegetation and environment has been gained. This is unusual for lowland England as a
whole but, as with the Fenlands of East Anglia, is attributed to the extensive lowland fen marshes of
these areas and the extensive spreads of peat and mineral sediments which have accumulated. These
deposits have offered tremendous potential for pollen studies dating back to the inception of the
discipline (eg, Godwin 1941; 1943). Interest in this area and again the Fenlands is also enhanced by
the potential for preservation of wetland archaeology (Coles et al. 1975–8) including notably, the
well documented Neolithic and Bronze Age trackways of the Somerset Levels and the Iron Age lake
dwellings of Glastonbury (Bulleid & Gray 1911; 1917; Coles & Minnett 1995) and Meare (Bulleid &
Gray 1948; Coles & Coles 1986). In contrast, the Avon Levels lying to the east, whilst offering
potential for palaeo-environmental studies (vegetation, environment, human impact, and Holocene
sea-level change) had not been examined in any detail until the Second Severn Crossing English
Approaches Project necessitated archaeological and palaeoenvironmental investigations.
The regional pollen database
Whilst the Somerset Levels are some distance away from the Avon Levels study area, the wealth of
pollen data and diagrams from the region as a whole offers a valuable framework within which the
information gained pollen studies at Awkley Lane, Vimpenny’s Lane, and Hallen may be compared.
The substantial depths of sediment that has accreted in the low-lying areas adjacent to the Severn
Estuary and Bristol Channel are clearly a response to post-(Devensian) glacial eustatic change.
Increase in RSL and thus transgression from the Late Devensian was responsible for the deposition
of basal marine clays (lower Wentlooge) seen in the Severn Estuary Levels. This transgression was
recognised at Shapwick Heath by Godwin (1943) and in the later extensive excavations of Coles in
the Somerset Levels (Coles et al. 1973). Subsequent, negative changes in RSL resulted in coastal
87
emergence, development of extensive fen carr and peat accumulation over the underlying clays, and
some incipient pedogenesis, as seen at Hallen. During the Late Bronze Age–Iron Age marine
inundation saw the re-incursion of marine influence, further deposition of marine clays over the fen
peat, and, locally, ponding back of the perimarine zone and fluvial systems causing flooding as
evidence at Glastonbury and Meare Lake dwellings. Changes in RSL are viewed as the principal
regional controlling factor in the character of sediment accretion in these low lying areas adjacent to
the Bristol Channel/Severn Estuary (Kidson & Heyworth 1976; Allen 1987a; Allen & Rae 1987;
Hibbert 1980; Scaife & Long 1994).
The north Somerset region has been extensively studied for many years. Of particular significance
was Godwin’s work (1940; 1941; 1943; 1946; 1948; 1956; 1960). This work established the
relationships between the rising post-glacial sea-level, consequent environmental changes, and the
importance and effect on human populations in the region. For example, Dewar and Godwin (1963)
recognised two primary phases of arable and pastoral agricultural activity. Without the aid of
radiocarbon dating at this time, he attributed these phases to the Middle–Late Bronze Age and
secondly to the Iron Age and Romano-British periods. This primary data has been much extended by
archeological excavation of the Somerset Levels trackways and associated environmental studies
(plant macrofossil, pollen and diatom analyses, and radiocarbon dating (Coles et al. 1973; 1975–8;
Beckett & Hibbert 1979; Hibbert 1978). With detailed and multiple (three dimensional) pollen
diagrams constructed in association with the prehistoric trackways, palaeo-geographical
reconstruction, and landscape history has been made possible demonstrating for example, localised
Neolithic clearance on the sandy islands of the Burtle Beds (Hibbert in Coles et al. 1970; Coles et al.
1973). Typical features such as the prehistoric ‘elm decline’ (eg, Chilton track (Coles et al. 1970);
Sweet Track, Factory Site (Beckett 1979); and Abbot’s Way (Beckett & Hibbert 1976)) have also
been recorded and radiocarbon dated. From the now large number of studies, regional pollen
assemblage zones and overall vegetational history has been established (Beckett & Hibbert 1978,
86). This has proven to be especially important providing useful comparative data for the pollen
study presented here from Awkley Lane.
The area of the Avon Levels has been less well studied than the Somerset Levels in spite of the
obvious potential of this low lying region adjacent to the Bristol Channel. Thus, there is little data
with which to compare directly, the local pollen sequences. Where data do exist, these are
preliminary investigations or assessments undertaken in response to commercial developments and
have not yet been fully analysed. These Avonmouth sites include Rockingham Farm (Walker et al.
1998b), Cabot Farm (Walker et al. 1999a), and Kites Farm (Walker et al. 1999b).
Relevent data are also available from south Wales which also date back to earlier periods of pollen
analysis (Godwin & Newton 1938; Godwin 1940; 1943) and associated with the foreshore
archaeology and peat accumulations in south Wales in the Gwent Levels. The Welsh side of the
Second Severn Crossing has also been subject to palaeo-environmental studies including diatoms
(Cameron 1993). Pollen analysis (Scaife 1993; 1994), and sea-level change (Scaife & Long 1994).
Clearly, comparisons and relationships can be drawn between the south Wales and Avon Levels
stratigraphical and biological sequences. This has previously been attempted by Smith and Morgan
(1989; and see Bell et al. 2000).
Avon Levels sites
Although only at three sites has pollen been analysed, the data obtained provide clear indications of
the environment of the Avon Levels and its fringing areas and how this changed throughout the later
prehistoric period from the c. 4500 cal BC (5600 BP), the late middle Holocene Atlantic period
(Flandrian II), until the Romano-British period in the 1st century AD. Godwin’s discussions
88
concerning the relationships between rising post-glacial sea-levels, environmental changes, and the
importance of and effect on human populations in the Somerset Levels, discussed above, similarly
apply to the Avon levels and are reflected in the environmental analyses which have been undertaken
at the sites of Awkley Lane, Hallen, and Vimpenny’s Lane.
Sea-level change
As discussed above, the substantial depths of sediment which accreted at these sites was a response
to rising post-glacial RSL, the principal controlling factor in the character of sediment accretion in
these low lying areas adjacent to the whole of the Bristol Channel/Severn Estuary. It is
acknowledged that, because of the sensitivity of organic sedimentation facies to forcing factors,
combined with lag effects, substantial intrinsic uncertainties may be expected in the sea-level curves
discussed above constructed from radiocarbon dated peats (Allen 1995, 42–3).
Because of the closeness of the Somerset and Avon Levels, it is highly likely that the same
controlling factors will have been responsible for the middle and late Holocene development of the
latter. The Late Bronze Age–Iron Age marine inundation of the Somerset Levels which caused
deposition of marine clays over the fen peat, ponding back of the perimarine zone and fluvial systems
and ultimately the flooding of the Glastonbury and Meare Lake dwellings, may also have been a
major cause of human abandonment of the Iron Age site at Hallen Marsh although the on-site
stratigraphy reflects the position of the site in the centre of the Avon Levels floodplain. Certainly, the
transition from fen peat/poor fen to estuarine sedimentation at Awkley Lane bears close comparison
with the Somerset (Shapwick Heath) model and fall within the framework of RSL datum points
presented by Haslett et al. (1997) and Scaife and Long (1994; see Figs 21a and b).
The changing vegetation environment

The palynological study of the three sites here embraces the late middle Holocene period (Flandrian
chronozone II) the sub-boreal and sub-Atlantic period (Flandrian chronozone III); that is, the late
prehistoric and historic period until the medieval period.
Pollen data presented provide evidence for the changing vegetation and environment of the better
drained soils of the terrestrial zone and the on-site (autochthonous) development of fen wood, poor
fen, and salt-marsh and mudflat habitats, the latter in response to fluctuations in RSL. The former
demonstrates the demise of the dominant middle Holocene woodland through progressive late
prehistoric woodland clearance. An attempt has been made to correlate these changing vegetation and
environment of these sites/periods using this data and the available radiocarbon dates and
stratigraphical architecture recorded from the three profiles examined. Local pollen assemblage zones
described from the three sites have been collated (Fig. 22).
The terrestrial vegetation: Awkley Lane has the longest temporal/stratigraphical sequence being the
only one which provides evidence for the middle Holocene environment (pollen zone AWK:1). and
first effects of human impact on the climax woodland communities. The determination of 5603±50
BP (NZA-12774) (4530–4350 cal BC) places the basal peat at Awkley Lane (AWK:1) at the end of
the Atlantic, Flandrian Chronozone II. From the pollen data this is clearly just prior to the ‘primary
(Neolithic) elm decline’ (see below) which is evidenced in AWK:2. This phase represents the
culmination of early Holocene vegetation migration and establishment and stability in the middle
Holocene. Typically, the woodland of this period comprised dominant oak, elm, lime, ash, and hazel
(Birks et al. 1975) with alder growing on-site in carr woodland (see below).
The importance of lime/linden woodland: Tilia (Tilia cordata), with the high pollen percentages and
plant macrofossils (Clapham above) seen in AWK:1, demonstrates that this was locally dominant on
the nearby well drained soils. This adds further to the now numerous data from southern Britain
which has demonstrated that this was the dominant or at least co-dominant taxon (with oak and/or
89
elm) at this time (eg, Moore 1977; Scaife 1980; 1988; 2000; Waller 1994). After its arrival in the late
Boreal (Godwin 1940; 1975a; 1975b) or early Atlantic period, Tilia cordata (lime/linden) became an
important and, in some areas, the dominant woodland component of the middle Holocene (Atlantic
period; Flandrian Chronozone II). This is in fact the most characteristic phenomenon of the late
prehistoric woodland in many areas of southern England (Birks et al. 1975; Birks 1989; Moore 1977;
Scaife 1980; Greig 1982) along with its decline at various times during the late prehistoric period,
particularly during the Middle–Late Bronze Age. Whilst Tilia has been recorded many times in
studies of the Somerset Levels, at Awkley, the proximity of the sample site to well drained land has
recorded percentages to 33%. Given the very great under-representation of Tilia in pollen spectra
(Andersen 1970; 1973), these values must be regarded as significant. This, plus the presence of the
fruits, provides clear evidence for such domination on suitable soils in the west country. Lime growth
was considered to be a characteristic only of the well drained soils of the interfluves and here, on the
localised islands which occur on the floodplain. However, it should also be considered that some
recent studies are now, indicating that far from being restricted to such ‘good’ soils, lime may have
also been a constituent of poorer sandy soils (Waller 1994, 96) and even very ‘damp’ woodland on
peaty substrates (Scaife in Tomalin et al. forthcoming) as shown by studies at Wootton-Quarr, Isle of
Wight (Scaife 2000).
Whilst Tilia was clearly the dominant woodland element, other tree and shrub taxa were also of
importance in the local environment. It is possible that oak, ash, and hazel formed parts of the dry
carr woodland or woodland fringing the edge of the mire – perhaps on the thicker, down-slope soils.
Ilex aquifolium (holly) is recorded at Awkley (AWK: 1) and is usually extremely poorly represented
in pollen spectra and thus, as with lime, implies some importance. Alternatively and very likely,
these woodland trees formed lesser elements in the dominant lime woodland during the middle
Holocene (AWK:1).
The first evidence of human disturbance: It is suggested that the zone AWK:l/2 transition is the early
part of the prehistoric elm decline at c. 5000 BP or, slightly later if c. 3550 cal BC (c. 4800 BP) if
dates for the Somerset Levels pertain. This is confirmed by the radiocarbon date of 3630–3360 cal
BC (4683±55 BP, NZA-12530) near the base of zone AWK:2. Further reduction in elm pollen at 1.86
m OD (AWK:2 sub-zone a and sub-zone b) and sharply expanding values of Poaceae (grasses) with
sporadic occurrences of cereal type pollen, Plantago lanceolata (ribwort plantain) and peaks of other
herbs including Asteraceae (Bidens type and Aster type), Urtica (nettle and pellitory), and
Chenopodiaceae (goosefoots and oraches) are all indicative of changes in the local dry-land
vegetation. This reduction in lime and elm at the zone AWK:l/2 transition is the prehistoric elm
decline at c. 4400–380 cal BC (5500–5000 BP) (Smith 1970; Smith & Pilcher 1973). The increase in
herbs which occurs is also strongly indicative of the first impact of human disturbance and
agriculture on the environment and is attributed to the establishment of a Neolithic economy with
localised cereal cultivation. Whilst the decline in elm was likely to be the result of spread of disease
instigated by Neolithic opening of the forest (Girling 1988), the reduction of other areas of woodland
was perhaps due to clearance for occupation and agriculture. Subsequently, oak, lime, ash, and hazel
woodland remained the most woodland trees on the better drained soils. Typically, there is evidence
of a secondary regeneration of elm which is only seen at Awkley (AWK:2c) which has similarly been
evidenced in other areas of southern England (Scaife 1988). Here, this is also associated with a
regeneration of lime and ash (both as noted, are poorly represented in pollen spectra). This occurred
at c. 2900–2800 cal BC (4300–4200 BP), the late Neolithic. Whilst there have been arguments that
Late Neolithic woodland expansion resulted from depopulation (Whittle 1978), there is also
90
5
A
Vimpenny's
Lane
Awkley
0 Lane
-5
Altitude (mOD)
-10
-15
-20
-25
0 2 4 6 8 10
Calender years BP (thousands)
Figure 21a. Sea-level curve using the existing framework of Haslett et al. (1997) from a synthesis
of data obtained by Kidson and Heyworth
0
B
Sea-level (metres below present sea level)
Site 3
5 Site 11 Vimpenny's Lane
Oscar 3 Awkley Lane
10 Context 333-340
15
20
25
30
35
10 9 8 7 6 5 4 3 2 1 0
Radiocarbon years BP (thousands)
Figure 21b. Sea-level curve using the the corrected RSL graph for Caldicot Pill on the north side
of the Severn Crossing
91
E 100 W
100
6 m O.D.
101
101 102/151 WA Northwick

201
183
202
104 151 Upper Wentlooge
t t WA Hallen
203 fla fla
152
5 m O.D. ud ud el
204 M M nn
a
186 ch
105 sh t 183
ar 205 fla sh
ltm ud ar
106 M
Sa 184 tm
107 Sal
sh
206 ar t
108
ltm fla
109 Sa ud
207
4 m O.D. sh M
208 ar
l tm
209 Sa WA Hallen
sh
110 t ar
m
fla 210 a lt
S s
ud as
M gr
4182+_ 55BP g e
211 d
3 m O.D. Log Se
212 t
112 fla
113 ud
M
113 rr
ca WA Vimpenny Middle Wentlooge
121
h
rc
Bi
114 d ge
120 se
2 m O.D. 115
d er en
116 Al d f
ee
R
118 rr
ca
119 d er
Al
WA Awkley Lane
1 m O.D.
Lower Wentlooge
0 m O.D.
+ve RSL
-ve RSL 0 1m
=equal
Figure 22. Correlation of the sedimentary facies from the reported sites
92
substantial evidence that the woodland regeneration occurred through changing agricultural practice
to woodland based pastoralism (Scaife 1980; 1988).
After the Late Neolithic regeneration phase there is progressive reduction in woodland which
ultimately left oak and hazel as the principal arboreal constituents but with some remaining lime, ash,
and beech. Evidence for these changes is complicated by the changing environment of deposition and
thus sedimentology and pollen taphonomy. Clearly, however, the next significant event was the
widespread removal of lime woodland from its areas of dominance. This was a result of increasing
human pressure on available land and reflects the expansion of Bronze Age activity in the region.
This decline in lime is evidenced at all three sites examined at zone boundaries AWK:4/5; VIMP:1/2
and to a lesser extent at Hallen Marsh HAL:1.
Later Holocene deforestation: the lime decline: The decline of lime from southern England was
initially attributed to climatic change by Godwin (1940; 1956) and, without radiocarbon dating, it
was previously assumed that the lime decline was synchronous and represented climatic worsening
from the sub-Boreal to Sub-Atlantic period at c. 1000–500 BC. Turner (1962), however,
demonstrated the possibility that prehistoric clearance for agriculture may have been a major factor
in its decline by identifying the expansion of agricultural weeds which occurred along with, or
shortly after, declining lime pollen percentages. This has subsequently been widely illustrated from
sites in southern England (eg, Scaife 1980; Waton & Barber 1987; Waller 1993). Furthermore, the
development of radiocarbon dating has demonstrated that the decline was not synchronous but took
place from the Neolithic period onwards. Whilst the majority of dates come from the Late Bronze
Age, at Awkley Lane, the late Atlantic dominance is reduced at the time of the elm decline possibly
representing clearance of lime for agriculture on the locally well drained soils or even for use of its
bast fibres. Neolithic decline is relatively rare but see on the Isle of Wight (Scaife 1980), Somerset
Levels sites (eg, Beckett 1979; Beckett & Hibbert 1976; 1978) and London (Scaife 2000). Other
important explanations for the lime decline have also been forthcoming in recent years. Waller
(1993; 1994) has suggested that expanding wetland (in Sussex) may have forced lime dominated
woodland farther away from basin sample sites, thus causing reduction in pollen percentages at that
site.
Taphonomic causation may be foremost and relevent to the decline in Tilia at the Avon Level sites
where pollen assemblage zone boundaries are associated with major changes in sediment regime and
thus pollen taphonomy. However, superimposed on these taphonomic changes, lime must have been
removed during the late prehistoric period; since it no longer exists and there is no regional evidence
for its extensive growth during the historic period. From the discussion presented, it seems plausible
that this was during the Bronze Age and associated with the overall reduction in trees seen in the
three pollen profiles. During the span of pollen zone AWK:4; VIMP:1a–c and HALL:1 the low
percentages of Tilia in the mineral (estuarine) sediments possibly represents growth but at distance
from the sample site. However, above the critical pollen/stratigraphical zone boundaries (see Section
B, above) of AWK:4/5, VIMP:1/2 and HALL:1/2, numbers are substantially reduced and it seems
likely that the final removal occurred during the temporal hiatus suggested for this stratigraphical
boundary.
From the evidence at Hallen we can see that this boundary occurs in the later Iron Age, with
incursion and sedimentation sealing an Iron Age seasonal (summer) occupation dated to 390–110 cal
BC (2195±55 BP, NZA-12727). If this boundary is consistent and contemporaneous across the
region and is marked by the colour changes discussed above which has been postulated to relate to
the middle/upper Wentlooge boundary, then we can examine the dated evidence at Northwick.
Although no palaeo-environmental studies were undertaken on this relatively shallowly buried
Romano-British site (of the 1st century AD), it is clear (see Fig. 18) that the features of this site lie
93
above the blue (lower) alluvium of the middle Wentlooge and are wholly within the upper alluvium
(upper Wentlooge). This would enable us to postulate this boundary occurring at the beginning of the
1st millennium AD.
Awkley provides the best evidence of terrestrial vegetation changes being in closer proximity than
Hallen and Vimpenny’s Lane. In zone AWK:3 there is evidence of soil deterioration perhaps caused
by such woodland clearance and subsequent acidification/podzolisation of the sandy soils of the
islands. This is shown by the expansion of acidophilous communities (Erica, Calluna, and
Sphagnum: see below) and the possible presence of charcoal in the macrofossil samples during the
period c. 2800–2400 cal BC (c. 4300–3900 BP) (based on the dating of the organic sediments at
Awkley Lane zone AWK:3; see Table 1 and Fig. 5). Similarly at Vimpenny’s Lane there is an
expansion of such acidophiles although here, their expansion is associated with the change from
organic to mineral sediments suggesting that pollen was derived fluvially from some distance.
Overall, it appears that significant environmental changes were taking place during the late
prehistoric period brought about by a positive RSL tendency and the increasing intensity of human
activity throughout the Bronze Age and into the Iron Age. Thus, whilst the sedimentological and
taphonomic changes caused by marine inundation of the fen community at Vimpenny’s Lane and
Awkley Lane at c. 2550 cal BC (4000 BP) resulted in the reduction of some elements such as Tilia
(see above), it may be argued that the extension of the pollen catchment through fluvial transport
(rather than largely airborne) might provide a more realistic picture of vegetation in the wider
geographical area; that is, assuming that pollen had become incorporated into sediments in the
estuary directly or from rivers exiting into the estuary.
From this period (ie, post-lime decline), there is consistent evidence of woodland dominated by
Quercus (oak) with Corylus avellana type with some lesser representation of other and less well
represented pollen taxa including Tilia, Fraxinus, Fagus, Populus, Ilex aquifolium, Taxus baccata.
Such importance of oak and hazel woodland during the later prehistoric period until present is
characteristic of southern Britain and is seen in the majority of later pollen sequences cited from this
region. The importance of oak and hazel pollen during the later/historic period results from the high
pollen productivity of these taxa and the fact that they were maintained in managed (coppice with
standards) woodland. Populus and the more common taxa (Betula and Pinus) which are likely to be
from airborne transport represent growth within the region as a whole, and from longer distances.
Remaining Tilia as well as Ilex aquifolium and Fraxinus and Fagus sylvatica are taxa which are
generally substantially under-represented in pollen spectra and, if not fluvially transported from
elsewhere in the catchment, probably indicate local growth. Associated with the demise of woodland
is the continued expansion of agriculture giving increased herbaceous diversity and pollen
percentages from pasture and arable habitats (ruderal/segetals). These are superimposed on the pollen
from the autochthonous communities.
The changing wetland/floodplain/estuarine habitat
Increasing RSL during the early Holocene and the early middle Holocene periods was responsible for
the deposition of the lower Wentlooge Series, the basal estuarine and marine clays seen of the
Somerset Levels and Avon Levels. This initial transgression was recognised in the Somerset Levels
at Shapwick Heath by Godwin (1943) and later by Coles et al. (1973). Subsequent negative tendancy
in RSL resulted in coastal emergence and development of extensive fen carr and peat accumulation
(middle Wentlooge) over these underlying clays. The middle Wentlooge peats have been analysed at
Awkley where they formed under a fen carr woodland community along the fringes of a more
extensive reed swamp fen evidenced in the lower (analysed) levels at Vimpenny’s Lane. The former
94
dating to 4530–4350 cal BC (5603±50 BP, NZA-12774) to 2840–2310 cal BC (3991±50 BP, NZA-
12754) and the latter from 3350–2900 (4420±90 BP, GU-3121) to 2900–2620 (4182±55BP, NZA-
12527).
This lowest peat (AWK:1) is only seen in this study at Awkley Lane where it accumulated under a
typical alder fen carr woodland (Alnetum glutinosae) from c.5600–c. 5000 BP. The best
environmental data for this community comes from plant macrofossil analysis. Numerous macro-
remains and large pollen frequencies of Alnus glutinosa were recovered along with evidence of a
typical fen carr ground-flora (Chrysosplenium alternifolium, Lycopus europaeus, Eupatorium
cannabinum). One of the characteristic species of this type of woodland is the alternate-leaved
golden saxifrage which was found in some abundance in samples (Awkley) at 1.44 m OD and 1.48 m
OD. Typical of this habitat were localised areas of wetter habitats/pools including for example,
Hippuris vulgaris, Alisma plantago-aquatica, and Cladium mariscus and the eggs of water-fleas.
There were also drier areas of fen carr with a ground flora of ferns (eg, Dryopteris felix femmina and
D. felix mas-female and male fern), sedges such as Carex paniculata (tussock sedge), Urtica dioica
(stinging nettle), Solanum dulcamara (woody nightshade), and Rubus sp. Taxus baccata, although
only a single (pollen) record, was typically a tree of such dry fen carr woodland (Godwin 1975a;
Scaife in Sidell et al. 2000).
In stratigraphical contexts 114, 120, 115 (zone AWK:2.b at c. 3500 cal BC (4700BP)), there is
evidence of an ephemeral phase of increasing wetness as indicated by increasing pollen frequencies
of reed-swamp taxa including Cyperaceae (sedges), Typha latifolia and Typha angustifolia/
Sparganium type (bulrushes and burreed), and aquatic megaphytes such as Potamogeton (pond
weed), Callitriche (water starwort), and Lemna (duckweed). Plant macrofossils similarly demonstrate
an expansion of a wet fen with Phragmites australis reedswamp with for example Mentha aquatica,
Thalictrum flavum, Apium nodiflorum, Alisma lanceolatum, and Ceratophyllum demersum (see Table
4).
In contrast, Alnus (alder) displays a sharp reduction in sub zone AWK:2b in response to this
increasing wetness. It is possible that local woodland clearance (charcoal was present) on the nearby
interfluve may have caused a reduction in evapotranspiration, an increase in surface run-off and a
resultant higher ground water table. Alternatively, a fluctuation in ground water table through more
regional change in RSL may have been responsible. The fragile balance of alder carr and water table
along the fen margins was clearly disturbed resulting in the demise of the carr woodland. Typically,
alder will tolerate flooding of the peat surface around the root boles for some 3-4 months of the year
during winter (Tansley 1949; McVean 1953; 1956). With the exception of a single peak of Alnus at
2.38 m OD, values of alder were and remained much reduced from its earlier dominance (seen in
zone AWK:1; context 118).
Fluctuation of the reed-swamp taxa as indicated in the pollen record (see for example the Cyperaceae
and Typha latifolia in Fig. 7) indicates a rather unstable, variable wetland habitat. The small increase
in Salix (willow) pollen may be also significant since willows are largely under-represented in pollen
spectra and relate to fringing willow carr. Because of the very substantial numbers monolete
Dryopteris spores (typical ferns) it is also likely that these were ferns growing within the wet, fen
carr marsh habitat-although subsequently the diagnostic marsh fern (Thelypteris palustris) is well
represented.
From c. 3550 cal BC (4800 BP) the peat at Awkley (context 121; zone AWK:3) contains substantial
fragments of wood dated to 3090–2700 cal BC (NZA-12528, 4286±55 BP). This sample was the
middle sample from the uppermost peat horizon. Pollen in zone AWK:3 exhibits a very marked
change in the character of local woodland with Betula (birch) becoming dominant while Alnus
remains consistent at levels seen in earlier pollen sub zones. Macrofossil evidence (wood, bark,
95
fruits, bud-scale) also attests to the development of birch carr woodland. This is comparable with
similar birch dominance in some phases of the Somerset Levels (Beckett 1979). This clearly
represents a major environmental change. Apart from the birch plant remains few other macrofossils
were present except for substantial quantities of Sphagnum. This is similarly attested in the pollen
analysis with peaks of Sphagnum spores and Ericales.
At Awkley Lane, there was a developing hydrosere where reed-swamp/tall-herb fen raised the
surface above the watertable, via the accumulation of peat leading to the development of alder carr
(AWK:1 and AWK:2) and birch carr (AWK:3). At Awkley, however, removal of trees from the
adjacent sand island may also have caused local acidification of the soils (with heath communities)
and increased the acidity of the mire (the birch carr of AWK:3) giving local Sphagnum bog. A thin
peat facies at Vimpenny’s Lane allowed radiocarbon dating allowing correlation with the primary
sequence examined at Awkley. Here, at 2900–2620 cal BC (4182±55 BP, NZA-12527 at 3.24 m
OD), no evidence of similar birch (or alder) carr was found. Although there is evidence of
acidophilous communities, these occur in the overlying mineral sediments and as such may reflect
fluvial transport from other sites such as Awkley at some distance. This suggests that vegetation
development at Awkley represents the areas fringing the floodplain and adjacent to higher
ground/terrestrial environment. Vimpenny’s is more clearly situated in the outer floodplain where an
overall wetter habitat was present supporting a wet, poor fen.
Summary
Overall, therefore, there is evidence of change in habitat types through the Awkley (middle
Wentlooge) peat sequence and evidence of regional variation as seen at Vimpenny’s lane. At
Awkley, the lower peat horizon was dominated by an alder carr woodland with an undergrowth of
wet/damp loving species. The middle peat horizon, (contexts 115, 120, and 114) shows a tall herb
fen/reedswamp was the dominant vegetation with very little or no indication of woodland vegetation
being present within the local mire habitat. This was then followed by a change in vegetation to one
dominated by Sphagnum moss with strong evidence of birch growing on and near the site. The
provision of radiocarbon dates makes it is possible to give a rough idea of the time span of these
developments at Awkley Lane. The date for the bottom of the sequence is 4530–4350 cal BC (NZA-
12774, 5603±50 BP), the date for detrital peat (context 120) from the middle peat horizon is 3640–
3380 cal BC (NZA-12529, 4745±50 BP), which suggests that the change from alder woodland to
reed-swamp/tall herb fen is approximately 900 years. The date for the topmost sample (context 121)
of peat is 3090–2700 cal BC (NZA-12528, 4286±55 BP) at 2.42 m OD which gives the time span for
the change from reedswamp/tall herb fen as approximately 500 years.
Caution must, however, be used when expressing the time span between successions in this way. It
should be considered that the peat bands are not contiguous and are separated by thin bands of
alluvium and, therefore, it can be very difficult to give a more accurate time span for these vegetation
changes to occur. However, in general terms it can be said that a period of approximately 1500 years
saw a vegetation change from an alder woodland to one dominated by Sphagnum moss at Awkley
whereas, further out in the floodplain at Vimpenny's Lane (context 211), such birch carr woodland
was not present and instead, a poor fen (reed swamp) dominated by Phragmites australis was present
(also confirmed by evidence of the insect remains). This fen was perhaps periodically inundated by
brackish water (see also diatom evidence). These data may be comparable with the Somerset levels
(Godwin 1975a, 31) where the development of ombrotrophic mire was often preceded by a stages of
Phragmites and Cladium fen and fen carr woodland. At Meare Heath (Beckett and Hibbert 1979;
Hibbert 1980) for example, a raised Sphagnum ombrogenous mire had developed and was inundated
by estuarine conditions at c. 910–760 cal BC (SRR-914, 2624±45 BP).
96
The mineral, estuarine and freshwater sediments
Radiocarbon dates of 2900–2620 cal BC (4182±5 5BP, NZA-12527 @ 3.14 m OD) at Vimpenny's
Lane and 2840–2310 cal BC (3991±50 BP, NZA-12754 @2.92 m OD) at Awkley Lane represent the
uppermost (and thus datable) peat at these sites prior to estuarine sedimentation. These dates
represent an earliest commencement (terminus post-quem) for this major environmental
change/phase since it is highly likely that some of the uppermost organics will have been eroded.
Clearly, however, this marked stratigraphical change, after which there are only ephemeral phases of
organic accumulation during the major phase of late Holocene sedimentation – the Upper Wentlooge
Formation. Allen (1990b; 1992b) has noted the little understood nature of the tidally influenced
marshes preserved within the Wentlooge formation, and, whilst it appears unlikely that the sediments
will provide a useful record of dating and relative sea-level change (see above) there is clear
evidence of the changing environment of the salt-marsh as evidenced by the changing palynological
assemblages.
The lowest levels analysed at Vimpenny’s correlates temporally with the uppermost peat/sediment
contact at Awkley Lane at c. 2550 cal BC (4000 BP). At Vimpenny’s Lane, the position of this site
further out on the floodplain/marsh suggests that there will be less of the fen and carr peat which in
general thickens towards the terrestrial zone (Allen 1992b). This appears to be the case although the
full stratigraphical extent/depth was not examined. At Vimpenny’s the lowest analysed sequence at
2.90 m OD (context 212) shows strong evidence of halophytic vegetation and reworked
palynomorphs which suggest a salt-marsh or possibly mudflat environment. The organic facies of
context 211 (zone VIMP 1a) has been dated at 2900-2620 cal BC (4182±55 BP, NZA-12527)
appears represent a phase of negative sea-level tendency which allowed the development of reed-
swamp dominated by Phragmites australis and other rooting, poor fen taxa. However, it appears that
this habitat was periodically inundated since salt-marsh taxa remain important in the pollen spectra.
This was, however, an ephemeral phase of reed-swamp which culminated with further deposition of
grey estuarine/marine clays and silts (upper Wentlooge). This was caused by a positive sea-level
tendency from c. 2550 cal BC (4000 BP) and correlates with a similar event at Awkley Lane
(AWK:4 context 110). At Vimpenny’s Lane, this initially formed salt-marsh (context 210) and then
mudflat (209) (VIMP:1b). These arguments are based on specific characteristics and fluctuations of
the palynological assemblages as described in above; that is, the prevalence of pre-Quaternary
palynomorphs, Holocene spores derived from fluvial sources and deposited with sedimentary fines
on mudflats and presence and/or absence of autochthonous halophytic taxa from the more stable salt-
marsh habitat.
At Awkley Lane, the upper Wentlooge Formation is similarly characterised by estuarine silts resting
on the middle Wentlooge peat (zones AWK:1–3). The upper Wentlooge transgressive event as with
Vimpennys occurred at c. 2550 cal BC (4000 BP). Palynologically these events seem to suggest that
at Awkley, the minerogenic sediments accreted in a mudflat environment rather than salt-marsh. The
sediment source was, derived largely from rivers (Allen 1990b; 1991) exiting and depositing their
load into the now incumbent estuarine conditions. Such transported material is apparently better
represented in the sediments of mudflats. Thus, and unsurprisingly, there is evidence from the sites
analysed of significant variations in the local character of the floodplain and estuarine environment
with mudflat, salt-marsh and incised channels. As Allen (1990b) points out, theoretical, mudflat-
marsh growth is determined by rates of minerogenic and organogenic sedimentation, the rate of
change and tendency of relative sea-level and the rate of ‘long-range’ sediment compaction. The
fringing fen and accumulating peat would have been moved transgressive inland with rising RSL
giving higher ground water table and thus potential for anaerobic conditions along river valleys and
along the terrestrial fringes. Negative tendency may also give rise to peat accumulation in areas
previously subject to salt-mash/tidal influence; effectively an off-lapp situation such as indicated in
97
context 211 at Vimpenny’s Lane. This reflects the complex causes and responses to a range of
controlling factors in estuarine environments/habitats (Long et al. 2000).
At Hallen although the lower levels (zone HALL:1) are not clearly dated, it is thought that they
correlate with the lower part of the upper Wentlooge Formation. As at Awkley, it appears that
mudflat developed into salt-marsh. Whether this was positive or negative sea-level tendency is not
clear but its seems more plausible that the on-site changes reflect a critical balance between sediment
supply and vegetation dynamics.
A significant stratigraphical break has been detected which occurs within the upper Wentlooge Series
at three of the sites: at 3.98 m OD at Vimpenny’s Lane, 3.95 m OD at Awkley Lane, and 4.80 m OD
at Hallen. This horizon appears to represent a stasis horizon resulting in a hiatus in sediment
deposition caused by a significant negative tendency of sea-level change during the Iron Age period.
This horizon is visible in many sections of the Avon Levels and appears to be delimited by change
from lower, typical grey, estuarine/marine silts to brown oxidised silts/clays (see above). Once
thought to be a post-depositional colour change, this is now shown to be due to differing sediment
sources. This stasis horizon has important archaeological implications since at sites such as Hallen
this standstill or negative tendency allowed increased human activity on the salt-marsh. At Hallen
this horizon is intricately linked with Iron Age occupation dated to 390-110 cal BC (2192±55 BP,
NZA-12727) where there is some evidence that the marsh developed into a drier and possibly rough
pasture community on which Iron Age occupation existed at this time.
Our model indicates that this stasis event was terminated by a final (Iron Age–Romano-British)
positive tendency which saw progressive waterlogging, sediment accretion, and a return to estuarine
salt-marsh and mudflat conditions. At Hallen and Vimpenny’s Lane, this became initially mudflat.
However, there was an ephemeral phase of salt-marsh development at Awkley at c. 4.88 m OD
(context 204; VIMP:2b). This may represent an oscillation of RSL (negative) or some other local
factor (drainage?) which allowed development of middle/lower salt-marsh dominated by halophytic
plants. Unfortunately there was too little organic material to allow radiocarbon dating of this horizon
and the possibility of an anthropogenically (post Iron Age) induced negative RSL tendency through
drainage cannot be precluded. This event was, however, short lived and there was a return to mudflat
(context 203).
Archaeology and Environment; marsh landscape and archaeology
Some archaeological implications
The detailed and dated environment studies have provided information of the evolution of the Avon
Levels (presented above). They also provide a framework for understanding the patterns and nature
of human action and intervention in this landscape which seems to be localised and both specialised
and intermittent. Obviously human action itself is also an important factor in the development and
modification of this landscape. The changing wetland environment facilitates and curtails
exploitation of some resources and provides strict temporal, seasonal, and spatial parameters for
settlement or occupation. The archaeological evidence from the sites discussed here (especially
Hallen Marsh and Northwick) is reviewed in relation to other recent archaeological finds to provide
an explanation of human behaviour which is integrated with the changing nature of the landscape
and its potential resources. It is recognised that throughout prehistory and the early historic period
the Levels provided but one small area (physically and economically) of larger social and economic
systems.
98
Limited evidence of changing RSL has been obtained, but detailed fluctuations of positive and
negative tendency have not been isolated, and it is these that might have a significant effect on the
parameters of human behaviour and patterns of activity. The importance of understanding the
evolution and changing nature of these small Levels cannot be over-emphasised in terms of
providing a framework and setting for human action. In particular, changes in the environmental
conditions are closely examined for the role they may have played in enabling and curtailing
utilisation, exploitation and settlement within the Avon wetlands. Obviously the earlier Holocene
(Mesolithic) activity, although sparse, is not considered here, as no deposits relating to this epoch
were examined. Instead the reader is directed to research and review by Druce (1997; 1998; 2001)
which covers this aspect more adequately from peat data at minus OD heights exposed on the
foreshore around Gravel Banks.
History of human exploitation patterns and behaviour
Conventionally the Avon Levels is essentially devoid of any evidence of Neolithic activity (Darvill,
in Aston & Iles 1986). Unlike the Somerset Levels, there is little evidence for the Neolithic
occupation of the Avon Levels. The Middle Holocene environment comprised patchy discontinuous
fen carr woodland fringing more extensive reed-swamp fen with localised dryer marshes and
woodlands of oak, elm, ash and hazel. Woodland cover that existed seems to have persisted later
than in many other areas. Although substantial clearances were made in parts of the Somerset
Levels at about 3500 cal BC (Beckett & Hibbert 1978), much of the lower Severn Valley was not
seriously deforested until the later part of the Neolithic (Brown 1982). Evidence of regional and
climatic vegetation succession is recorded during the Neolithic and the elm decline is present at
Awkley Lane (c. 3550 cal BC, 4800 BP). In addition limited human interference of the climax
woodland can be detected, and there clear evidence of disturbed ground (nettles, ribwort plantain,
etc).
Although Neolithic artefacts are present at Hills Flats as stray finds on the foreshore to the north in
the inner estuary (Allen 1997; 1998); they are absent from the main Avon Levels. It is possible that,
if present, they are buried by several metres of later Holocene minerogenic sediments, however, the
Holocene environment does not seem to be conducive to settlement, and as yet no ‘islands’ of
archaeology, or trackways such as are present in the Somerset Levels, are known. Despite this,
there is some cultivation on the edge of, or adjacent to, the Avon Levels as indicated in the presence
of cereal pollen at Awkley Lane from about 4530–4350 cal BC (c. 1.5 m OD), with sporadic
occurrences throughout the Neolithic peat. From this sparse evidence, enhanced by the very limited
archaeological intervention in the Levels with information at about this date, we can envisage the
area as one cautiously and periodically exploited for a range of fish, fowl and plants (reeds,
Phragmites) to compliment the resources from the few Neolithic sites on the surrounding higher
land (flint scatters at Almondsbury (unpublished SMR data)). Whether cereal cultivation look place
on the Levels or slightly higher and drier land of the narrow scarp foot bench on its margins is
undetermined. Nevertheless, such activity took place within the wider context of the Levels
landscape.
Within the Avon Levels themselves the only artefacts are occasional Neolithic axes (three) from
intertidal and near wetland contexts along the coastal margins (Allen 1990d), and Neolithic flint
knapping sites on the coastal margins in the inner estuary at Oldbury Flats (Allen 1990c, 34) and
Hills Flats (Allen 1997). These may represent localised flintworking, and perhaps the exploitation
of flints from mixed, now almost offshore, gravels in the fashion of exploitation regimes postulated
for the river gravels in Langstone Harbour, Hampshire (Allen and Gardiner 2000; 2007).
99
Very sparse evidence of later Neolithic and earlier Bronze Age activity, save a few localised flint
scatters which have become exposed in coastal locations such as Hills Flats (Allen 1997; 1998),
tend to suggest that the area was a backwater, perhaps providing less important resources on rare
forays into the area.
Local stasis horizons are noted within the middle Wentlooge Formation comprising varyingly
organic surfaces. One such weakly organic horizon at about 4.5–4.7 m OD at the Cabot Park and
seabank sites (see Table 12 above and Locock 1999) indicates localised soil ripening and gave
consistent dates around 2700 cal BC. This might have provided an opportunity for increased
human exploitation. But to date, this horizon has been devoid of evidence of human exploitation
wherever it has been encountered, though fragmentary charcoal has been recovered from samples
(Locock 1997a; 1999) possibly indicating anthropogenic activity in the vicinity.
It is not until the later Bronze Age that a series of ‘settlements’ can be detected in the Levels. A
cluster of similar sites occurs around Cabot Park (Locock et al. 1998) and Avonmouth but this
concentration is largely an artefact of recent industrial development at this location which has led to
archaeological exploration, and we can expect similar sites to be widely dispersed over much of the
Levels.
Fen carr and swamps had been inundated by fluvial and intertidal sediment creating a more uniform
environment of salt-marsh; freshwater marsh occurred locally, fringing the higher ground and
associated with streams draining from the hills to the coast. This open salt-marsh with some higher
and drier patches enabled limited soil formation; typically of a saline alluvial gley soil (Avery
1990) with some soil ripening. This is reflected in a stasis horizon observed at 5.1–5.2 m especially
around Cabot Park and Avonmouth (Locock et al. 1998, table 1; Locock 1999), consisting of a
stabilisation horizon with incipient gleying, evidence of soil ripening and visually appearing
slightly more humic. Evidence of human occupation has not been found at every location that this
‘Kites Corner’ horizon occurs, but its present indicates some localised high spots or drier patches
within a wider salt-marsh which graded gently to mudflat at its estuarine and river margins within
the salt-marsh.
The archaeological sites are ‘set back’ from the present coastal margins (Fig. 23), in contrast to
sites of the same date in the Gwent Levels, where there has been a 1–2 km loss of the Levels over
the last 2000 years (Bell et al. 2000), which lie on and below peat shelves on the present day tidal
edge, and are appreciably higher than those in Wales (5.1 vs 2.5–3.9 m OD). The Avon sites are
ephemeral and comprise charcoal patches, and concentrations of animal bone, burnt stone, imported
stone, and pottery, with no, or scant evidence of structures (Locock et al. 1998). These sites are not
distinguishable on the basis of the physical evidence and tend to indicate localised temporary
activities on the salt-marsh edge, perhaps near to former small water courses. Locock (2000) has
suggested that these might be temporary seasonal ‘camps’ associated with summer grazing,
saltmaking and possibly even potting. The sites were not long lived, nor re-visited, even though the
general vicinity was. What is also evident from the radiocarbon determinations is that this activity
persisted almost unchanged for about one millennium from 1780–1510 cal BC to 930–520 cal BC.
These locations seem to have operated in spatial and temporal isolation from each other; there is no
evidence of any infrastructure or route-, or track-, ways. This contrasts with Caldicot Levels where
brushwood trackways have been found at Cold Harbour Pill (Locock 1997b) that probably date to
around 1300–450 cal BC. Evidence of trackways and more permanent infrastructures in the Gwent
Levels fits with evidence of later Bronze Age round-houses and more substantial settlements at
fig 23
100
Chapel Tump (Whittle 1989), Rumney Great Wharf (Allen 1996a), and Collister Pill (Bell et al.
2000, 309–11). Similarly short-lived trackways existed in the Somerset Levels (Coles & Coles
1986; 1998) linking foci of semi-permanent Bronze Age activity on islands of slightly higher
ground. These more complex social systems do not seem to have developed, or have not yet been
discovered, in the Avon Levels.
Kites Corner dense charcoal patch 3m across, and wider spread of charcoal, some stake-holes
5.1 m OD 2610±70 BP (920–520 cal BC)
Avonmouth 1 scatter, stone and charcoal, some pottery
5.3 m OD 2778±55 BP (1110–810 cal BC)
5.3 m OD 2957±55 BP (1370–1010 cal BC)
Little Googs dense scatter of burnt stone, bone, pottery, and charcoal (6 m area)
5.1-5.2 m OD 2970±60 BP (1390–1020 cal BC)
5.1-5.2 m OD 3350±60 BP (1870–1500 cal BC)
Stinkums 3 sites, burnt stone, animal bone, and pottery
By about 450 cal BC all of the later Bronze Age sites known in the Avon Levels were inundated by
minerogenic sediments relating to salt-marsh or mudflat environments. Abandonment is
asynchronous, in keeping with the activity, and suggest a general diminishing of exploitation rather
than large- and wholesale changes in the environment rapidly curtailing these summer forays. This
fits with the environmental evidence of gradual changes and fluctuations in positive and negative
RSL at this time. So, as environmental change gradually lead to the lessening of human activity, a
period of seldom/sparse human invention followed.
By the later Iron Age (350 cal BC), whereas mudflats were more widespread previously, now they
only fringed open salt-marsh along alluvial or estuarine inlets and much of the sedimentation was
essentially marine. The subtle transition to increased fluvial sedimentation, slightly longer and drier
summer spells and increased freshwater fluvial input into the wetland system led to subtle changes
in the observed sedimentary sequence and prescribed the change from the middle to upper
Wentlooge Formation, as described earlier. Unfortunately we have no firm dates for this change,
but is likely to asynchronous across the region, varying in accordance to proximity to the shore,
tidal inundation sources and freshwater streams.
Very little Iron Age evidence is known from the Levels as a whole (Fig. 23) but at Hallen Marsh,
by the Late Iron Age (200 cal BC), relatively substantial settlement is present, and to date there is
no evidence of any earlier Iron Age activity in the Avon Levels. By the later Iron Age complex
societies had developed with both permanent and specialised sites. Hillforts are present on the
upland overlooking the Avon Levels; the Hallen site is situated within the Levels, in contrast to
those in north Somerset which seem to be specifically located on the fen edge or in the Brue Valley,
where the freshwater peats have produced the famous sites of Glastonbury and Meare (Coles &
Coles 1996; Coles & Orme 1983; 1985).
Although the Iron Age activity at Hallen Marsh is considerable more substantive than that of the
later Bronze Age sites, it nevertheless, does not indicate permanent settlement such as those known
on the higher lands. The settlement at Hallen Marsh comprises at least two complex round-house
structures (Plate 4) sited on either side of a small stream. All of the artefacts were readily portable
and it seems that all the necessities of basic life, including staple foods, were brought to the site,
either in containers or on the hoof. The site seems to have been occupied on a seasonal basis,
specifically for the grazing of sheep, with some cattle, horses, pigs, and dogs present. The animal
bone assemblage indicates a ‘normal’ herd structure, with animals being killed as required for
immediate consumption. It seems likley that Hallen represents evidence for transhumance between
101
353000 354000 355000 356000 357000 358000 359000 360000
188000
l
187000
ne
n
ha
lC
sto
186000 Northwick Awkley Interface

Green
Bri
Lane
185000
Rookery Farm
Ellinghurst Farm
184000
Caltybrook Farm
Brynleaze Farm
Almondsbury Fort
183000
Crookmarsh
Farm
182000
Elmington Manor
181000
Rockingham
Farm
Cabot
Park Hallen
180000
Katherine’s
Farm
179000
Lawrence Blaise Castle

Weston Land over 10m
Long Cross
Bronze Age Barrow/findspot
178000 Kings
Late Bronze Age Site
Weston
Kings Weston Hill Middle-Late Iron Age Site
Romano British Site
177000
Figure 23. Location of archaeological sites in the Avon Levels
102
the limestone uplands and the rich summer pasture of accessible parts of the Levels (see Gardiner et
al. 2002 for a full discussion and Fig. 24).
Why does settlement occur?
There is no major change in the local environment, but then there need not be large changes in the
‘marginal’ landscape to enable it be utilised for exploitation and grazing. As we have discussed
there is a minor change in local conditions (the middle/upper Wentlooge boundary), but this may
not in itself have been great enough to engender the onset of human exploitation and settlement. As
we cannot invoke environment deterministic ideals, we may examine this in terms of environmental
possibilism. It is the combination of a subtly and slowly changing landscape enabling some pasture
and seasonal occupation, in combination with increasingly complex and larger societies (the Avon
Levels in the later Iron Age formed part of the territory of the Dobunni tribe; Cunliffe 1991) which
enabled and facilitated the necessity to utilise the Avon wetlands (Fig. 24).
The site at Hallen Marsh is specialised and seasonal, undoubtedly concentrating on summer grazing
and production of local and portable crafts. The lack of charred grain from the settlement (see
Clapham in Gardiner et al. 2002) is likely to be a real, rather than taphonomic, phenomenon. Grain
was present for preparation and consumption as a stored food stuff, but agriculture was not part of
the Hallen economy.
The two round-houses lie either side of stream, which itself seems to be upper profile of a much
early palaeochannel that cuts through the earlier, middle Wentlooge peats, at this location (Druce
2001; Juggins 1982). This situation was probably important for the welfare of both humans and
livestock. Although not clearly ascertained from analysis, this is likely to have been largely
freshwater, but inundated with brackish water on higher and storm tides.
What caused Iron Age abandonment?
Cessation for settlement is not due to environmental catastrophe or rapid change. There is evidence
at Hallen of increased brackish and marine sediment input, probably from overbank flooding of the
stream during the period of later Holocene positive RSL. This led to increasingly wetter conditions
prevailing during longer periods of the year. Summer grazing of stock, especially sheep, became
unfeasible. Ultimately the settlement itself was inundated with mineral sediment and the artefacts
washed over the site and incorporated in the overlying, newly forming alluvial gley soil in the salt-
marsh. This was inevitable and was understood by the inhabitants of Hallen, for not only were the
buildings located on slightly higher land but attempts were made to consolidate the entrances-ways
of each by adding imported stone, and discarded bone and broken pottery to the paths. The flooding
was exacerbated locally by the proximity of the stream, but ultimately represents a wider
phenomenon. Thus, although ultimately we can see that the local environmental conditions were
responsible for abandonment of the Late Iron Age summer grazing regimes at Hallen Marsh, this
was a gradual retreat of Iron Age use of the area. A continued cycle of mudflat/salt-marsh – salt-
marsh/mudflat during positive RSL made the region unpredictable and thus, in economic terms,
unviable. This rise in RSL is seen as the cause of the abandonment of a number of Iron Age–
Romano-British sites throughout the Somerset (Hibbert 1980) and Gwent Levels (see Boon 1980),
and was first defined by Godwin (1940; 1943).
103
Topographic Neolthic Later Early Mid-Late Early Late
location Bronze Age Iron Age Iron Age Roman Roman
Upland Monuments Barrows No Hillforts Rural Villas and

evidence settlements temples
Some Development
settlement Development
of of settlement
tribal territory hierarchy and
communications
Flint &
Flint Farmsteads
stone axes,
scatters Farmsteads
flint scatters
Lower Possible No Small towns

Slopes clearance Some evidence
some arable clearance ?Fields
arable Rural
Fen Edge Fish and No Mixed farmsteads,
fowl evidence farming mixed farming
Reed Open Mudflat/ Increasing seasonal flooding

Fen swamp salt-marsh salt-marsh leading to permanent waterlogging
Pot Increasing
making Seasonal seasonal
Probably No Seasonal
grazing with
Fen not evidence grazing flooding
field systems leading to
Islands present for presence and temporary
Summer and temporary permanent
settlement
grazing settlement waterlogging
Intertidal Mudflat/ Saltworking Mudflat/

salt-marsh salt-marsh Increasing seasonal flooding
Zone
Figure 24. Suggested model for the exploitation of the Avon Levels.
105
A change of use
Some human exploitation also occurs in the Levels in the earlier Romano-British period. Field
systems seem to have been established in the salt-marsh in a few locations such as Northwick,
bounded by relatively deep, but well-defined ditches. We have no evidence of major environmental
change at this time though local variations in ground-water regimes will have been important
factors; minor and subtle negative changes in RSL would make these developments feasible. This
activity will have been accompanied by some occupation in the Levels – quantities of pottery,
animal bone, and charcoal occur in the ditches at Northwick, but the nature and extent of such
activity was not elucidated in excavation here. The limited evidence that so far exists for Romano-
British occupation in the Levels themselves seems to be specifically early Roman in date (1st–2nd
century AD). Crook’s Marsh is, to date, is the only later Roman site to have been identified within
the Levels (Everton & Everton 1980; Allen & Fulford 1986; Juggins 1982, 21–31). Here a series of
ditches was identified, not dissimilar to Northwick. Both probably represent paddocks, rather than
arable fields, and more formal and controlled stock husbandry; presumably based on summer
grazing. These fields, with often water-filled ditches, were probably used for seasonal graze and,
therefore, as an economic activity were less jeopardised by gradual rising RSL and increased
periodic inundation.
Other Romano-British settlements are sited on the fen-edge and on the slopes of the higher ground
overlooking the Levels. A number of sites north of the Avon and centred around Lawrence Weston
and Kings Weston are particularly notable, with habitation spanning the 1st–4th centuries AD but
the majority of dated artefacts are of later Romano-British date (3rd–4th century AD).
In general, although we can see formalised control of stock in the Levels, the utilisation and
occupation of the area is at a lower and more ‘rural’, nature than that seen in both the Gwent and
the Somerset Levels. Whether there is continuity of occupation at sites like Northwick, where Iron
Age pottery also occurs, cannot yet be determined but certainly by the later Romano-British period
settlement was sited on the distinctly higher land peripheral to the Levels themselves. Ultimately,
even the fen-edge fields too, were inundated by further minerogenic sedimentation. Agricultural
development of the area resumed in the medieval and post-medieval periods with large-scale ridge
and furrow agriculture (Fig. 24).
Land claim
Much has been made previously of Roman land claim of coastal margins of the Severn Levels (eg,
Rippon 1997), and this is well demonstrated at Rumney Great Wharf and along the Wentlooge
Levels (Allen & Fulford 1986). Similar piecemeal land claim is also evident along the narrow inner
estuary at Longney (Allen & Fulford 1990a), Elmore (Allen & Fulford 1990b), and Oldbury-on-
Severn in Gloucestershire (Allen & Fulford 1990b; Riley 1998). In keeping with the lower level of
activity, along the Avon Levels land claim is, however, less well testified, and we can argue that
such activity did not take place along this stretch of coastline. This has implications for the nature
and intensity of Romano-British activity. Unlike the Wentlooge Levels where major settlement
occurs within the former fens and industrial activity (smelting) occurs, a much smaller scale of
operation seems evident in the Avon Levels.
105
Archaeological significance
Here we have summarised some of the main events defined by the archaeological evidence and
relating to human behaviour. In this landscape, which is subject to distinctive changes resulting
from minor fluctuations in RSL and runoff, changes in the vegetation and local environment are
inextricable tied to the nature and phases of human activity. The analyses reviewed in this paper
provide a more detailed geoarchaeological picture of the Avon Levels than was previously possible.
A full discussion of the rchaeological evidence and a model for the prehistoric exploitation and
georachaeology is presented elsewhere (Gardiner et al. 2002).
Conclusions
The analytical programme discussed here has enabled us to define the middle and upper
Wentlooge Formations in the field and analytically in terms of subtle changing vegetation
patterns. It has also provided the basis for re-examination of patterns of human behaviour
form the Neolithic to Romano-British periods which are summarised here and explored in
more detail elsewhere (Allen & Gardiner, in Gardiner et al. 2002). We can, therefore,
summarise the main events in terms of the Wentlooge Formations as follows.
Formation deposit Period Environmental conditions Archaeology

Upper Wentlooge reddish Romano-British and return to saltmarsh, some ditched field systems &
estuarine later (500 BC +) gradual flooding & drained landscape
silts/clays inundation especially
associated with channels.
Middle blue & grey Bronze Age–Iron salt-marsh & intertidal Human exploitation of
Wentlooge estuarine silts/ Age (4500–800 BC) mudflats the saltmarsh: Cabot
clay over peat park & Avonmouth
(LBA), Hallen Marsh
(IA), Northwick (RB)
Lower Wentlooge peats or Mesolithic–Neolithic salt-marsh, wet alder carr Potential for fishing and
blue/grey (5500–4500 BC) with local Phragmites reed fowling. Limited
estuarine silts swamps, raised bogs, minor artefacts & occupation
over sands rivers & numerous small
streams
Acknowledgements
The authors would like to thank all of those who contributed in the field and to the assessment and
post-excavations programmes of the English Approaches Project, and especially to those who
contributed to this paper. Their written words are produced here but the text also includes their
thoughts and ideas from long and detailed discussions. The principal authors would like to thank all
the specialists who contributed to the programme and to this report. In addition we thank Mark
Noel and associates (archaeomagnetic dating) and particularly Karen Nichols for interpreting our
rough drawings and workings-out to produce the illustrations.
This paper has benefitted from discussions with and information provided by numerous colleagues
and specialists, of whom Julie Gardiner, Anthony Long, Martin Locock, and Steve Rippon deserve
special mention. Steve Juggins allowed us to use and quote from his long-lost undergraduate
dissertation and Denise Druce was especially generous in discussing the results of the PhD research
and supplying extracts of her thesis. Julie Jones and Vanessa Straker provided information and
106
copies of various reports unavailable to us. Thanks are also due to Richard Newman whose
foresight prior to and during fieldwork in 1992–4 has finally been realised by this and its
complimentary papers. The paper was edited for publication by Julie Gardiner, who would like to
thank all the contributors and referees for their patience and assistance. The post-excavation project
was funded by Highways Agency, managed for them by Ian Smith, and monitored on their behalf
by Janet Miller of W.S. Atkins Heritage.
107
Appendix: Analytical Techniques
Pollen analysis
by Robert G. Scaife
Standard techniques were used for the extraction of the sub-fossil pollen and spores (Moore &
Webb 1978; Moore et al. 1991). Micromesh sieving (10 m) was used to aid the removal of the
substantial clay fraction in the grey sediments. Samples of 1.5–2 ml were used for the peat and 5
ml for the overlying grey minerogenic sediments. Absolute pollen frequencies were calculated
using added exotics to known volumes of sample (Stockmarr 1971). Pollen was identified and
counted using an Olympus biological research microscope fitted with Leitz optics. A pollen sum
of 400 grain of dry land taxa plus all extant spores and pollen of marsh taxa (largely Cyperaceae
and Alnus), fern spores, and miscellaneous pre-Quaternary palynomorphs was counted for each
level where preservation made this possible. This was generally achieved except in some levels
of the upper minerogenic sediments. Very substantial total pollen numbers total were counted for
the lower peat deposits where pollen of Alnus, Betula, and fern spores were extremely abundant.
An extensive reference/comparative collection of modern taxa was available to aid identification.
Pollen diagram(s) have been constructed and plotted using Tilia and Tilia Graph. Percentages
have been calculated as follows:
Sum = % total dry land pollen (tdlp)

Marsh/aquatic = % tdlp+sum of marsh/aquatics
Spores= % tdlp+sum of spores
Misc.= % tdlp+sum of misc. taxa.
Alnus, with its extremely high values and on-site dominance in the lower pollen assemblage zone,
has been excluded from the pollen sum and is incorporated into the marsh/aquatic category as
originally detailed by Janssen (1969). Taxonomy in general follows that of Moore and Webb
(1978) modified according to Bennett et al. (1994) for pollen types and Stace (1997) for plant
descriptions. These procedures were carried out in the Palaeoecology Laboratory of the
Department of Geography, University of Southampton.
Diatoms
by Nigel G. Cameron
Diatom preparation followed standard techniques. Two coverslips, each of a differing concentration
of the cleaned solution, were prepared from each sample. Where the dilutions were not satisfactory
for counting a second set of slides was prepared. Slides were scanned thoroughly at magnifications
of x400 and at x1000 under bright-field, phase contrast, and differential interference contrast
illumination.
Waterlogged plant macrofossils

by Alan J. Clapham
3
Sub-samples of 300 cm from each sample were processed via wet sieving through a series of
granulometric sieves with mesh diameters of 1 mm, 500 μm and 300 μm. The resulting fractions
were sorted under water using a low-powered stereomicroscope (x6–x40) and the critical taxa were
compared with the modern seed reference collections housed in the George Pitt-Rivers Laboratory,
department of Archaeology, University of Cambridge. The taxa identified were then counted. All
nomenclature follows that of Stace (1997).
108
Molluscs
by Michael J. Allen
Bulk samples of up to 2 kg were processed by standard wash-over flotation methods (Evans

1972). Because of the sparse and fragile nature of shells from these deposits, extreme care
was taken in their dissaggregation; in cases where the silts did not disaggregate upon
immersion in warm water, they were subject to up to to three cycles of freeze-thaw (see Bell
in Balaam et al 1987) and re-immersed in warm water. After disaggregation the flots were
retained on a 0.5mm mesh and the residues passed through a 0.5 mm mesh. The lack of any
residue negated the necessity of passing the minerogenic fraction through a stack of sieves.
Insects
by Mark Robinson
Samples of 2 kg were subjected to paraffin flotation to recover insect remains. The flot was caught
on a 0.2 mm sieve, washed in detergent and sorted under a binocular microscope. Nomenclature for
Coleoptera follows Kloet and Hincks (1977).
Sediments
by J.R.L. Allen
Samples were analysed for grain-size distributions and plotted on triangular diagrams to show
values of the clay-silt-sand ratio. Each sample was coarsely but non-destructively crushed and
quartered to yield a sub-sample. This sub-sample was crushed more finely and reduced to a further
sub-sample weighing a few grams. This further sub-sample was then mixed with a weak aqueous
solution of Calgon to form a paste and spread out over a glass plate. From the ‘biscuit’ so formed, a
final sub-sample of a few tenths of a gram was removed by jabbing a tiny spatula into the biscuit at
numerous, well-scattered points. The final sub-sample was then introduced to a Coulter LS laser
granulometer and a count allowed to proceed for 1–2 minutes. Using the Fraunhofer with PIDS
model as the basis for calculating the results, the resulting machine output chiefly lists measures of
central tendency (especially mean and modal grain sizes), selected percentile values, and the clay-
silt-sand ratio.
Charcoal
by Rowena Gale
Bulk samples were processed by floation with flots and residues retaiend on a 0.5 mm mesh.
Samples of charcoal were prepared for examination using standard methods. Fragments from each
sample were fractured to expose fresh transverse surfaces (TS) and sorted into groups based on the
anatomical features observed using a x20 hand lens. Representative fragments from each sample
were selected for detailed study at high magnification. Additional surfaces to show the wood
structure in the tangential (TLS) and radial planes (RLS) were also prepared. The fragments were
supported in washed sand and examined using a Nikon Labophot microscope at magnifications of
up to x400. The anatomical structures were matched to prepared reference slides. When possible the
maturity (ie, heartwood/ sapwood) of the wood was assessed. Classification follows that of Flora
Europaea (Tutin, Heywood et al 1964–80).
109
Magnetic susceptibility
by Michael J. Allen
Samples for magnetic susceptibility analysis were extracted from the sample columns at 10 cm
intervals through the upper alluvial sequences at each site, and from the Romano-British ditches at
Northwick. These were air-dried, crushed and passed through a 2 mm sieve with 10 g of the <2 mm
fraction measured using a Bartington MS2 meter coupled to a MS2b dual frequency sensor.
Measurements were made at 0.46kHz (low frequency) and at higher frequency (4.6kHz).
110
Bibliography
Allen, J R L, 1986 A short history of salt-marsh reclamation at Slimbridge Warth and neighbouring areas,
Gloucestershire. Trans Bristol Gloucestershire Archaeol Soc 104, 139–55
Allen, J R L, 1987a Late Flandrian shoreline oscillations in the Severn Estuary: the Rumney Formation at its type site
(Cardiff area). Phil Trans Roy Soc London B 315, 157–84
Allen, J R L, 1987b Dimilington Stadial (Late Devensian) ice-wedge casts and involutions in the Severn Estuary,
southwest Britain. Geol J 22, 109–18
Allen, J R L, 1988 Modern-period muddy sediments in the Severn Estuary (south western UK): a pollutant-based
model for dating and correlation. Sediment Geol 58, 1–21
Allen, J R L, 1989 Evolution of salt-marsh cliffs in muddy and sandy systems: a qualitative comparison of British west-
coast estuaries. Earth Surface Processes & Landforms 14, 85–92
Allen, J R L, 1990a Late Flandrian shoreline oscillations in the Severn Estuary: change and reclamation at Arlingham,
Gloucestershire. Phil Trans Royal Soc London A330, 315–34
Allen, J R L, 1990b The Severn Estuary in southwest Britain; its retreat under marine transgression, and fine-sediment
regime. Sediment Geology 66, 13–28
Allen, J R L, 1990c The post-glacial geology and geoarchaeology of the Avon wetlands. Proc Bristol Natur Soc 50
(special issue), 28–46
Allen, J R L, 1990d Three Neolithic axes from the Severn Esturay. Trans Bristol Gloucestershire Archaeol. Soc. 108,
171–74
Allen, J R L, 1991 Fine sediment and its sources, Severn Estuary and inner Bristol Channel. Sediment Geol 75, 57–65
Allen, J R L, 1992a Trees and their response to wind: mid Flandrian strong winds, Severn Estuary and inner Bristol
Channel, southwest Britain. Phil Trans Roy Soc London B, 338, 335–64
Allen, J R L, 1992b Tidally influenced marshes in the Severn Estuary, southwest Britain, in Allen, J R L and Pye, K
(eds), Saltmarshes; morphodynamics, conservation and engineering significance. Cambridge: Univ Press,
123–47
Allen J R L, 1995 Salt-marsh growth and fluctuating sea level: implications of a simulation model for Flandrian coastal
stratigraphy and peat-based sea-level curves. Sediment Geol 100, 21–45
Allen, J R L, 1996 Three final Bronze Age occupations at Rumney Wharf on the Wentlooge Level, Gwent. Studia
Celtica 30, 1–16
Allen, J R L, 1997 A scatter of Neolithic–Bronze Age flintwork from the intertidal zone at Hills Flats, South
Gloucestershire. Trans Bristol Gloucestershire Archaeol Soc 115, 265–276
Allen, J R L, 1998 A prehistoric (Neolithic–Bronze Age) complex on the Severn Estuary Levels, Oldbury-on-Severn,
South Gloucestershire. Trans Bristol Gloucestershire Archaeol Soc 116, 93–115
Allen, J R L and Fulford, M G, 1986 The Wentlooge Level; a Romano-British saltmarsh reclamation in southeast
Wales. Britannia 17, 91–117
Allen, J R L and Fulford, M G, 1988, Romano-British settlement and industry on the wetlands of the Severn Estuary.
Antiq J 67, 237–89
Allen, J R L and Fulford, M G, 1990a Romano-British wetland reclamations at Longney, Gloucestershire, and evidence
for the earlier settlement of the inner Severn Estuary. Antiq J 70, 288–326
Allen, J R L and Fulford, M G, 1990b Romano-British and later reclamations on the Severn salt-marshes in the Elmore
area, Gloucestershire. Trans Bristol Gloucestershire Archaeol Soc 108, 17–32
Allen, J R L and Rae, J E, 1987 Late Flandrian shoreline oscillations in the Severn Estuary: a geomorphological and
stratigraphical reconnaissance. Phil Trans Roy Soc London B 315, 185–230
Allen, J R L and Rae, J E, 1988 Vertical salt-marsh accretion since the Roman period in the Severn Estuary, southwest
Britain. Marine Geol 83, 225–235
Allen, M J, and Gardiner, J P, 2000 Our Changing Coast; a survey of the intertidal archaeology of Langstone Harbour,
Hampshire. York: Counc Brit Archaeol Res Rep 124
Allen, M J and Gardiner, J P, 2007 The Neolithic of Langstone Harbour: dryland resources preserved by a wetland
environment, in Haughey, F and Siddell, J (eds), Current Research in Neolithic Intertidal Archaeology.
Oxford: Neolithic Studies Group, 59–70
Allen, M J, and Macphail, R I, 1987 Micromorphology and magnetic susceptibility studies; their combined role in
interpreting archaeological soils and sediments, in Fedoroff, N, Bresson, L M, and Courty, M A, (eds), Soil
Micromorphology. Paris: Association Francaise pour l'etude du sol, 669–76
Allen, M J and Richie, K, 2000 The stratigraphy and archaeology of Bronze Age and Romano-British deposits below
beach level at Brean Down, Somerset. Proc Univ Bristol Spelaeol Soc 22, 5–52
Andersen, S Th, 1970 The relative pollen productivity and pollen representation of North European trees, and
correction factors for tree pollen spectra. Danm. Geol. Unders. Ser I 196, 99
111
Andersen, S Th, 1973 The differential pollen productivity of trees and its significance for the interpretation of a pollen
diagram from a forested region, in Birks, H J B and West, R G, Quaternary Plant Ecology. Oxford: Blackwell,
109–115
Andrews, J T, Gilbertson, D D and Hawkins, A B, 1984 The Pleistocene succession of the Severn Estuary: a revised
model based upon amino acid racemiztaion studies. J Geol Soc London 141, 967–74
Aston, M and Iles R, 1986 The Archaeology of Avon. Bristol: Avon County Council
Avery, B W, 1990 Soils of the British Isles. Wallingford: CAB (Commonwealth Agricultural Bureau) International
Balaam, N D, Bell, M G, David, A E U, Levitan, B, Macphail, R I, Robinson, M, and Scaife, R G, 1987 Prehistoric and
Romano-British sites at Westward Ho! Devon: archaeological and palaeoevironmental surveys 1983 and 1984,
in Balaam, N D, Levitan, B, and Straker, V, Studies in palaeoeconomy and environment in southwest England.
Oxford: Brit Archaeol Rep 181, 163–259
BARAS, 1998 Archaeological watching brief on the laying of an effluent pipeline between Avonmouth sewage works
and Saebank power station, Severnside, Bristol for Wessex Water plc. BARAS report 265/1998
Barnes, I, 1993. Second Severn Crossing: English Approaches an interim statement on the 1992/3 fieldwork. Ann Rep
Severn Estuary Levels Comm, 3–30
Beckett, S C, 1979, The environmental setting of the Sweet Track Drove site, in Coles, J M (ed), Somerset Levels Pap
5, 75–83
Beckett, S C and Hibbert, F A 1976, An absolute pollen diagram from the Abbot’s Way, in Coles, J M et al Somerset
Levels Pap 2, 24–27
Beckett, S C and Hibbert, F A, 1978 The influence of man on the vegetation of the Somerset Levels – a summary, in
Coles, J M (ed) Somerset Levels Pap 4, 86–90
Beckett, S C and Hibbert, F A, 1979 Vegetational change and the influence of prehistoric man in the Somerset Levels.
New Phytol 83, 577–600
Bell, M G, Caseldine, A, and Neumann, H, 2000 Prehistoric Intertidal Archaeology in the Welsh Severn Estuary.
London: Counc Brit Archaeol Res Rep 120
Bennett, K D, Whittington, G and Edwards, K J, 1994 Recent plant nomenclatural changes and pollen morphology in
the British Isles. Quat Newslet 73, 1–6
Birks, H J B, 1989 Holocene isochrone maps and patterns of tree-spreading in the British Isles. J Biogeog 16, 503–40
Birks, H J B, Deacon, J and Peglar, S, 1975 Pollen maps for the British Isles 5000 years ago. Proc Royal Soc London
B189, 87–105
Boon, G C, 1980 Caerleon and the Gwent Levels in early historic times, in Thompson, F H, (ed), Archaeology and
Coastal Change. London: Soc Antiq London Occas Pap NS 1, 24–36
Boyden, C R, Crothers, J H, Little, C, and Mettam, C, 1997 The intertidal invertebrate fauna of the Severn Estuary.
Field Stud 4, 477–554
Brown, A G, 1982 Human impact on the former floodplain woodlands of the Severn, in Bell, M, and Limbrey, S, (eds),
Archaeological Aspects of Woodland Ecology. Oxford: Brit Archaeol Rep S146, 93–104
Brown, A G, 1985 The potential use of pollen in the identification of suspended sediment sources. Earth Surface
Processes and Landforms 10, 27–32
Bulleid, A and Gray, H S G, 1911 The Glastonbury Lake Village. Vol 1. Taunton: Glastonbury Antiq Soc
Bulleid, A and Gray, H S G, 1917 The Glastonbury Lake Village. Vol 2. Taunton: Glastonbury Antiq Soc
Bulleid, A and Gray, H S G, 1948 The Meare Lake Village. Vol 1. Taunton, Taunton Castle
Burrin, P J and Scaife, R G, 1984 Aspects of Holocene valley sedimentation and floodplain development in southern
England. Proc Geol Assoc 95, 81–96
Cameron, N G, 1993 Diatom analysis of the Oscar3 Profile, in Godbold, S and Turner, R C, Second Severn Crossing
Archaeological Response, Phase 1 – The Intertidal Zone in Wales. Cardiff: Cadw, 61–4
Cameron, N G, 1997 The diatom evidence, in Nayling, N & Caseldine, A, Excavations at Caldicot, Gwent: Bronze Age
palaeochannels in the Lower Nedern Valley. York: Counc Brit Archaeol Res Rep 108, 117–28
Cameron, N G & Dobinson, S J, 1998 Diatom analysis of sediments and caulking material, in Nayling, N, The Magor
Pill Mediaeval Wreck. York: Counc Brit Archaeol Res Rep 115, 38–40
Cameron, N G & Dobinson, S J, 2000 Diatom analysis of sediments, in Bell, M., Caseldine, A., & Neumann, H,
Prehistoric Intertidal Archaeology in the Welsh Severn Estuary. York: Counc Brit Archaeol Res Rep 120,
262–6
Canti, M, Heal, V, Jennings, S, McDonnell, R, and Straker V, 1993 Archaeological and paleoenvironmental evaluation
of Porlock Bay and Marsh. Archaeol Severn Estuary, 49–69
Chambers, F, 1989 The evidence for early Rye cultivation in North West Europe, in Milles, A, Williams, D and
Gardner N (eds), Beginnings of Agriculture. Oxford: Brit Archaeol Rep S496, 165–75
Churchill, D M, 1965 The displacement of deposits formed at sea level 6500 years ago in Southern Britain.
Quaternaria 7, 239–49
Clarke, J C and Patterson, W A, 1984 Pollen, Pb-210 and opaque spherules: an integrated approach to dating and
sedimentation in the intertidal environment. J Sediment Petrol 54, 1249–63
Coles, B and Coles, J M, 1986 Sweet Track to Glastonbury; the Somerset Levels in Prehistory. London, Thames &
Hudson
112
Coles, B, and Coles, J, 1998 Passages of Time. Archaeol Severn Estuary 9, 3–16
Coles, J and Minnitt, S, 1995 Industrious and Fairly Civilized; the Glastonbury lake Village. Taunton: Somerset Levels
Proj/Somerset Co Counc Mus Serv
Coles, J and Orme, B, 1983 The Sedgemoor survey, 1982. Somerset Levels Pap 9, 6–8
Coles, J and Orme, B, 1985 Archaeology in the Somerset Levels 1984, Somerset Levels Pap 11, 5–7
Coles, J M, Hibbert, F A and Clements, C F, 1970 Prehistoric roads and tracks in Somerset – Neolithic. Proc Prehist
Soc 36, 125–50
Coles, J M, Hibbert, F A, and Orme, B J, 1973 Prehistoric roads and tracks in Somerset, 3 – the Sweet Track. Proc
Prehist Soc 39, 256–93
Coles, J M, et al. 1975–1988 Somerset Levels Papers 1–14.
Cundill, P and Whittington, G, 1987 Flood plain sedimentation in the Upper Axe Valley, Mendip, England: a comment.
Trans Inst Brit Geog 12, 360–2
Cunliffe, B W, 1991 Iron Age Communities (3rd edn). London: Routledge
Davies, P, 1996 The ecological status of Pupilla muscorum (Linné) in Holocene overbank alluvium at Kingsmead
Bridge, Wiltshire. J Conch Soc London 35, 467–71
Davies, P, 1998 Numerical analysis of subfossil wet-ground molluscan taxocenes from overbank alluvium at
Kingsmead Bridge, Wiltshire. J Archaeol Sci 25, 39–52
Denys, L, 1988 Fragilaria blooms in the Holocene of the Western Coastal Plain of Belgium, in Simola, H (ed), Tenth
International Diatom Symposium, Koenigstein: Koeltz, 397–405
Denys, L, 1992 A check list of the diatoms in the Holocene deposits of the Western Belgian Coastal Plain with a survey
of their apparent ecological requirements: I. Introduction, ecological code and complete list. Service
Geologique de Belgique. Professional Pap 246, 41
Devoy, R J, 1977 Flandrian sea level changes in the Thames estuary and the implications for land subsidence in
England and Wales. Nature 270, 712–15
Devoy, R J N 1979 Flandrian sea level changes and vegetational history of the Lower Thames estuary. Phil Trans Roy
Soc London B285, 355–407
Devoy, R J N 1980 Post-glacial environmental change and man in the Thames estuary: a synopsis, in Thompson, F H
(ed), Archaeology and Coastal Change. London: Soc Antiq London Occas Pap NS 1, 134–48
Devoy, R J N 1982 Analysis of the geological evidence for Holocene sea-level movements in south-east England. Proc
Geol Assoc 93, 65–90
Dewar, H S L and Godwin, H, 1963 Archaeological discoveries in the raised bogs of the Somerset Levels, England.
Proc Pehist Soc 29, 17–49
Dimbleby, G W, 1988 The palynology of Archaeological sites. London: Academic
Druce, D, 1997 Palaeoenvironmental constraints on Mesolithic to Romano-British communities of the Severn Estuary.
Archaeol Severn Estuary 8, 99
Druce, D, 1998 Late Mesolithic to early Neolithic environmental change in the central Somerset Levels: recent work at
Burnham-on-sea. Archaeol Severn Estuary 9, 17–30
Druce, D, 2001 Mesolithic to Romano-British Archaeology and Environmental Change of the Severn Estuary England.
Unpubl PhD thesis, Univ Bristol
Evans, J G, 1972 Land Snails in Archaeology. London, Seminar
Everton, A, and Everton, R, 1980 Romano-British occupation at Crooks Marsh Farm, Avonmouth. Bristol Archaeol
Res Group Rev 2, 57–8
Fairbanks, R W, 1989 A 17,000 year glacio-eustatic sea-level record: influence of glacial melting rates on Younger
Dryas event and deep ocean circulation. Nature 342, 637–42
Fulford, M, Allen, J R L, and Rippon, S J, 1992 The settlement and drainage of the Wentlooge Level, Gwent:
excavation and survey at Rumney Great Wharf 1992. Britannia 25, 175–211
Gale, R. 1992 Wood and charcoal, in Fulford et al 1992, 205–6
Gardiner, J, Allen, M J, Hamilton-Dyer, S, Laidlaw, M, and Scaife, R G, 2002, Making the most of it: Late prehistoric
pastoralism in the Avon Levels, Severn Estuary, Proc Prehist Soc 68, 1–40
GeoQuest Associates, 1993. Palaeomagnetic analysis of sediments from sites R10 [Vimpennys Lane], R11 [Hallen]
and R15 [Awkley Lane], English Approaches. Unpubl rep for Wessex Archaeol
GGAT = Glamorgan Gwent Archaeological Trust, 1992, Archaeology of the Second Severn Crossing: assessment and
recommendations for the English Approaches. Unpubl Client Rep
Gimmingham, C H, 1984 Ecological aspects of birch. Proc Roy Soc Edinburgh 85B, 65–72
Girling, M A, 1977 Fossil insect assemblages from Rowland’s Track. Somerset Levels Pap 3, 51–60
Girling, M A, 1988 The bark beetle Scolytus scolytus (Fabricius) and the possible role of elm disease in the early
Neolithic, in Jones, M. (ed), Archaeology and the Flora of the British Isles. Oxford: Oxford Univ Comm
Archaeol, 34–8
Godwin, H, 1940 Pollen analysis and forest history of England and Wales. New Phytol 39, 370–400
Godwin, H, 1941 Studies of the post-glacial history of the British vegetation. VI Correlations in the Somerset Levels.
New Phtyol 40, 108–32
Godwin, H, 1943 Coastal peat beds of the British Isles and North Sea. J Ecol 31, 199–247
113
Godwin, H, 1945 Coastal peat-beds of the North Sea region, as indices of land and sea-level changes. New Phytol 44(1)
29–69
Godwin, H, 1946 The relationship of bog stratigraphy to climatic change and archaeology. Proc Prehist Soc 12, 1–11
Godwin, H, 1948 Studies of the post-glacial history of the British vegetation. X Correlation between climate, forest
composition, prehistoric agriculture and peat stratigraphy in Sub-boreal and Sub-atlantic peats of the Somerset
Levels. Phil Trans Roy Soc London B233, 275–86
Godwin, H, 1956 History of the British Flora. Cambridge: Univ Press
Godwin, H, 1960 Prehistoric wooden trackways of the Somerset Levels: their construction. age and relation to climatic
change. Proc Prehist Soc 16, 1–36
Godwin, H, 1975a, History of the British Flora. A Factual Basis for Phytogeography (2nd edn). Cambridge: Univ Press
Godwin, H, 1975b History of the natural forests of Britain: establishment, dominance and destruction, Phil Trans Roy
Soc London B 271, 47–67
Godwin, H and Newton, L, 1938 The submerged forest at Borth and Ynyslas, Cardiganshire. New Phytol 37, 333–44
Greig, J R A, 1982 Past and present lime woods of Europe, in Bell, M and Limbrey, S (eds), Archaeological Aspects of
Woodland ecology. Oxford: Brit Archaeol Rep S146, 23–55
Greig J R A, 1991 The prehistoric riverside settlement at Runnymede: the botanical story, in Needham S and Longley
D, Excavation and Salvage at Runnymede Bridge, Berkshire, 1978. London: British Museum, 134–212
Grichuk, M P, 1967 The study of pollen spectra from recent and ancient alluvium. Rev Palaeobotany Palynology 4,
107–12
Groot, J J. 1966 Some observations on pollen grains in suspension in the estuary of the Delaware River. Marine Geol 4,
409–16
Hall, S A, 1981 Deteriorated pollen grains and the interpretation of Quaternary pollen diagrams. Rev Palaeobotany
Palynology 32, 193–206
Haslett, S K, Davies, P, and Strawbridge F, 1997 Reconstructing Holocene sea-level change in the Severn Estuary and
Somerset Levels: the foraminifer connection. Archaeol Severn Estuary 8, 29–40
Haslett, S, Davies, P, Curr, R H F, Davies, C F C, Kennington, K, King, C P, and Margetts, A J, 1998 Evaluating
Holocene Late Holocene relative sea-level change in the Somerset Levels, southwest Britain. Holocene 8,
197–207
Hawkins, A B, 1971 The Late-Weichselian and Flandrian transgression of South-West Britain. Quaternaria 14, 115–30
Hawkins, A B, 1973 Sea level changes around South-West England, in Blackman, D J, Marine Archaeology. London:
Butterworths, 67–87
Hewlett R and Birnie J, 1996 Holocene environmental change in the inner Severn Estuary UK; an example of the
response of estuarine sedimentation to relative sea level change. Holocene 6, 49–61
Heyworth, A, and Kidson, C, 1982 Sea-level changes in southwest England and Wales. Proc Geol Assoc 93, 91–111
Hibbert, F A, 1978 The vegetational history of the Somerset Levels, in Limbrey, S and Evans J G, (eds), The Effect of
Man on the Landscape: the Lowland Zone. London: Counc Brit Archaeol Res Rep 21, 90–5
Hibbert, F A, 1980 Possible evidence for sea-level change in the Somerset Levels, in Thompson, F H, (ed),
Archaeology and Coastal Change. London: Soc Antiq London Occas Pap NS 1, 103–5
Hustedt, F, 1957 Die Diatomeenflora des Fluss-systems der Weser im Gebiet der Hansestadt Bremen. Ab. naturw. Ver.
Bremen 34, 181–440
Janssen, C R, 1969 Alnus as a disturbing factor in pollen diagrams. Acta Bot. Neere. 8, 55–8
Janus, H, 1979 The Young Specialist Looks at Land and Freshwater Molluscs. London: Bourke
Jardine, W G, 1976 Some problems in plotting the mean surface level of the North Sea and the Irish Sea during the last
15000 years. Geol. Foreningens 98, 78–87
Jelgersma, S, 1961 Holocene sea-level changes in the Netherlands. Meded geol. Sticht ser. (6)7, 1–100
Jelgersma, S, 1966 Sea level changes during the last 10,000 years, in Sawyer, J S, (ed), World climate 8000–0 BC. Roy
Met Soc, 54–71
Jennings, S, Carter, R W Gm and Orford, J D, 1993 Late-Holocene salt-marsh development under a regime of rapid
relative sea level rise: Chezzetcook Inlet, Nova Scotia. Implications for the interpretation of palaeo-marsh
sequences. Canadian J Earth Sci 30, 1374–84
Juggins, S, 1982 The Geoarchaeology of Hallen Marsh. Unpubl undergrad dissert, Univ Bristol
Kerney, M P, 1999 Atlas of the Land and Freshwater Molluscs of Britain and Ireland. Colchester: Conch Soc Grt
Britain Ireland
Kidson, C and Heyworth, A , 1973 The Flandrian sea-level rise in the Bristol Channel. Proc Ussher Soc 2, 265–84
Kidson, C and Heyworth, A, 1976 The Quaternary deposits of the Somerset Levels. Quart J Engin Geol 9, 217–35
Kidson, C and Heyworth, A, 1978 Holocene eustatic sea level change. Nature 272, 748–50
Kloet, G S and Hincks, W D, 1977 A Check List of British insects, (2nd edn, rev): Coleoptera and Strepsipter, London:
Roy Entomol Soc Handbk Ident Brit Insects 11(3)
Lawler, M, Parkhouse, J and Straker, V, (eds), 1992 Archaeology of the Second Severn Crossing: assessment for
recommendations for the English Approaches. Swansea: Glamorgan-Gwent Archaeol Trust
Lindroth, C H, 1974 Coleoptera: Carabidae. London: Roy Entomol Soc Handbk Ident Brit Insects 4(2)
114
Locock, M, 1997a Rockingham Farm, Avonmouth: mouted enclosures on the North Avon Levels. Archaeology in the
Severn Estuary 8, 83–8
Locock, M, 1997b A prehistoric trackway at Cold Harbour Pill, Redwick, Gwent. Archaeol Severn Estuary 8, 9–12
Locock, M, 1999 Buried soils of the Wentlooge Formation. Archaeol Severn Estuary 10, 1–10
Locock, M, 2000 A Later Bronze Age landscape on the Avon Levels: settlement, shelters and saltmarsh at Cabot Park,
in Brück, J, (ed), Landscape and Settlement in Bronze Age Britain. Oxford: Oxbow, 121–8
Locock, M, and Lawler, M, 2000 Moated enclosures on the North Avon Level: survey and excavation at Rockinhgham
Farm, Avonmouth, 1993–7. Trans Bristol Gloucestershire Archaeol Soc 118, 93–122
Locock, M, Robinson, S, and Yates, A, 1998 Late Bronze Age sites at Cabot Park, Avonmouth. Archaeol Severn
Estuary 9, 31–6
Long, A J, Scaife, R G, and Edwards, R J, 1999 Pine pollen in intertidal sediments from Poole Harbour, UK;
implications for late-Holocene sediment accretion rates and sea-level rise. Quat Inter 55, 3–16
Long, A J, Scaife, R G, and Edwards, R J, 2000 Stratigraphic architecture, relative sea-level, and models of estuary
development in southern England: new data from Southampton Water, in Pye, K, and Allen J R L (eds),
Coastal and Estuarine Environments: sedimentology geomorphology and geoarchaeology. London: Geol Soc
London Spec Publ 175, 253–79
McVean, D N, 1953 Biological flora of the British Isles. Alnus glutinosa (L) Gaertn. J Ecol 41, 447–66
McVean, D N, 1956 Ecology of Alnus glutinosa ( L) Gaertn. VI Post-glacial history. J Ecol 44, 331–3
Macan, T, 1977 A Key to the British Fresh- and Brackish-water Gastropods. Kendall: Freshwater Biol Assoc Sci Publ
13 (4th edn)
Moore, P D, 1977 Ancient distribution of lime trees in Britain. Nature 268, 13–14
Moore, P D and Webb, J A, 1978 An Illustrated Guide to Pollen Analysis. London: Hodder & Stoughton
Moore, P D, Webb, J A and Collinson, M E, 1991 Pollen Analysis (2nd edn). Oxford: Blackwell
Peck, R M, 1973 Pollen budget studies in a small Yorkshire catchment, in Birks, H J B and West, R G (eds),
Quaternary Plant Ecology. Oxford: Blackwell, 43–60
Porter, D, 1990 The Archaeology of the Second Severn Crossing (preliminary survey). Unpubl, Avon County Planning
Dept
Pye, K, and Allen, J R L, 2000 Past, present and future interactions, management challenges and research needs in
coastal and estuarine environments, in Pye, K, and Allen J R L (eds), Coastal and Estuarine Environments:
sedimentology geomorphology and geoarchaeology. London: Geol Soc London Spec Publ 175, 1–4
Riley, H, 1998 Intertidal survey at Avon mouth and Oldbury-on-Severn 1998. Archaeol Severn Estuary 9, 79–80
Reid, C, 1913 Submerged Forests. Cambridge: Univ Press
Rippon, S, 1991 The Somerset Levels in the Roman period. Ann RepSevern Estuary Levels Res Comm, 43–6
Rippon, S, 1994 The evolution of a historic wetland landscape: the Gwent Levels historic landscape study. Archaeol
Severn Estuary, 21–34
Rippon, S, 1996 Gwent Levels: the evolution of a wetland landscape. York: Counc Brit Archaeol Res Rep 105
Rippon, S, 1997 The Severn Estuary; landscape evolution and wetland reclamation. Leicester: Univ Press
Russett, V, 1990 The Archaeology of the Second Severn Crossing (observations and geotechnical trial pits). Unpubl.
Avon County Planning Dept
Scaife, R G, 1980 Late-Devensian and Flandrian Palaeoecological Studies in the Isle of Wight. Unpubl PhD thesis,
Univ London
Scaife, R G, 1988 The elm decline in the pollen record of south east England and its relationship to early agriculture, in
Jones, M, (ed), Archaeology and the Flora of the British Isles. Oxford: Oxford Univ Comm Archaeol, 21–33
Scaife, R G, 1993 The palynological investigation of the peats and sediments, in Godbold, S and Turner, R C, Second
Severn Crossing. Archaeological response Phase 1–The intertidal zone in Wales. Cardiff: Cadw, 51–9
Scaife, R G, 1994 Pollen analysis and radiocarbon dating of the intertidal peats at Caldicot Pill. Archaeol Severn
Estuary, 67–80
Scaife, R G, 2000 Holocene vegetation development in London, in Sidell et al 2000, 111–17
Scaife, R G 2001 Pollen Analysis of Peat and MineralSsediments at Motney (W4358), Gravesend (W1847) and
Queenborough (W1872), North Kent. Unpubl rep for Wessex Archaeol
Scaife, R G and Burrin, P J, 1992 Archaeological inferences from alluvial sediments: some findings from southern
England, in Needham, S, and Macklin, M G, (eds), Alluvial Archaeology in Britain. Oxford: Oxbow Monogr
27, 75–91
Scaife, R G, and Long, A, 1994 Evidence for Holocene sea-level changes at Caldicote Pill. Archaeol Severn Estuary,
81–5
Shennan, I, 1982 Interpretation of Holocene sea-level data from the Fenland, England. Proc Geol Assoc 93, 53–63
Shennan, I, 1989 Late Quaternary sea-level changes and crustal movements in the British Isles. J Quat Sci 4, 1–94
Sidell, J, Wilkinson, K, Scaife, R G, and Cameron, N, 2000 The Holocene Evolution of the London Thames,
Archaeological Excavations (1991–1998). London: Mus London Archaeol Serv Monogr 5
Smith, A G, 1970 The influence of Mesolithic and Neolithic man on British vegetation: a discussion, in Walker, D and
West, R G (ed), Studies in the Vegetational History of the British Isles. Cambridge: Univ Press, 81–96
115
Smith, A G and Morgan, L A , 1989 A succession to ombrotrophic bog in the Gwent Levels, and its demise: a Welsh
parallel to the peats of the Somerset Levels. New Phytol 112, 145–67
Smith, A G and Pilcher, J R, 1973 Radiocarbon dates and the vegetational history of the British Isles. New Phytol. 72,
903–14
Stace, C, 1997 New Flora of the British Isles (2nd edn). Cambridge: Univ Press,
Stockmarr, J, 1971 Tablets with spores used in absolute pollen analysis. Pollen et Spores 13, 614–21
Tansley, A G, 1949 The British Isles and their Vegetation. London: Cambridge Univ Press
Tauber, H, 1965 Differential pollen dispersion and the interpretation of pollen diagrams. Danm. Geol. Unders. II 89, 1–
69
Tauber, H, 1967 Investigation of the mode of pollen transfer in forested areas. Rev Palaeobot Palynol 3, 277–87
Tomalin, D J, Loader, R, and Scaife R G, forthcoming Wootton Haven: Coastal and port archaeology in a dynamic
envirnment. Oxford: Brit Archaeol Rep
Turner, J, 1962 The Tilia decline: an anthropogenic interpretation. New Phytol 61, 328–41
Tutin, T.G., Heywood, V.H. et al 1964–80 Flora Europaea 1–5, Cambridge: Univ Press
Vos, P C and Wolf, H, de 1993 Diatoms as a tool for reconstructing sedimentary environments in coastal wetlands:
methodological aspects. Hydrobiologia 269/70, 285–96
Waller, M, 1993 Flandrian vegetational history of south-eastern England. Pollen data from Pannel Bridge, East Sussex.
New Phytol 124, 345–69
Waller, M, 1994 The Fenland Project, No 9: Flandrian Environmental Change in Fenland. Cambridge: E Anglian
Archaeol 70
Walker, M J C, Bell, M, Caseldine, A E, Cameron, N G, Hunter, K L, James, J H and Smith, D N, 1998a
Palaeoecological investigations of middle and late Flandrian buried peats on the Caldicot Levels, Severn
Estuary, Wales. Proc Geol Assoc 109, 51–78
Walker, M J C, Caseldine, A E, Barrow, K, James, J H, Cameron, N G, and Kreiser, A, 1998b Palaeoenvironmental
Assessment of Late Flandrian Sediments from Rockingham Farm, Avonmouth, near Bristol. Unpubl Rep,
Palaeoenviron Res Centre, Univ Wales, Lampeter
Walker, M J C, Caseldine, A.E., Crowther, J, Johnson, S, James, and Macphail, R I, 1999a,
PalaeoenvironmentalAassessment of Late-Flandrian Sediments from Cabot Park, Avonmouth, near Bristol.
Unpubl Rep Palaeoenviron Res Centre, Univ Wales, Lampeter
Walker, M J C, Caseldine, A E, Barrow, K, Cameron, N G, James, J H, Kreiser, A, and Macphail, R I, 1999b
Palaeoenvironmental Assessment of Late-Flandrian Sediments from Kites Corner, Avonmouth, near Bristol.
Unpubl Rep Palaeoenviron Res Centre, Univ Wales, Lampeter
Waton, P V, and Barber, K E, 1987 Rimsmoor, Dorset; biostratigraphy and chronology of an infilled doline, in Barber,
K E, (ed), Wessex and the Isle of Wight, Field Guide. Cambridge: Quat Res Group, 75–80
Wessex Archaeology, 1994 The Second Severn Crossing: English Approaches Archaeological Survey and Landscape
Analysis: Post-excavation assessment. Unpubl Client rep 35745
Wessex Archaeology, 2000 The Second Severn Crossing English Approaches Archaeological Post-Excavation
Programme. Unpubl client rep T5512
Whittle, A W R, 1978 Resources and population in the British Neolithic. Antiquity 52, 34–42
Whittle, A, 1989 Two later Bronze Age occupations and an Iron Age channel on the Gwent foreshore. Bull Celtic Stud
36, 200–23
Wilson, D, Davies, J R, Smith, M and Waters, R A, 1988 Structural controls on Upper Palaeozoic sedimentation in
south-east Wales. J Geol Soc London, 145, 901–14
Woodworth, P L, 1987 Trends in U K mean sea level, Marine Geol 11, 57–87
116
Wessex Archaeology
Internet Report
Registered Head Office:
Portway House
Old Sarum Park
Salisbury,
Wiltshire
SP4 6EB
Tel: 01722 326867

info@wessexarch.co.uk
www.wessexarch.co.uk
Registered Charity No. 287786

A company with limited liability
registered in England No. 1712772

The Physical Evolution of The North Avon Levels

Uploaded by

Copyright:

Available Formats

The Physical Evolution of The North Avon Levels

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Physical Evolution of The North Avon Levels

Uploaded by

Copyright:

Available Formats

What is the purpose of the document?

What is the purpose of the document?

What areas does the document cover?

What areas does the document cover?

The Physical Evolution

of the North Avon Levels

By Michael J. Allen and Robert G. Scaife

with contributions from J.R.L. Allen, Nigel G. Cameron, Alan J. Clapham,

with an introduction by Julie Gardiner

Wessex Archaeology Internet Reports

Copyright © Wessex Archaeology Ltd 2010 all rights reserved

Wessex Archaeology Limited is a Registered Charity No. 287786

Editor’s Introduction, by Julie Gardiner .......................................................................................... 1

Plate 1. Sampling the minerogenic sediment sequence in monoliths at Awkley Lane

with contributions from J.R.L. Allen, Nigel G. Cameron, Alan J. Clapham,

with an introduction by Julie Gardiner

The themes deemed to remain of greatest importance and potential were:

Inner Bristol Channel Brean OUTER

Zones of peat exposed intertidally

The Wentlooge Formation

The Avon Levels

Upper Wentlooge Middle Wentlooge Lower Wentlooge

Figure 2. Sedimentary features comprising the Wentlooge Formation

The investigations summarised below were conducted as an archaeological and palaeo-

Figure 4. Awkley Lane Wentlooge Formation profiles

Sampling the sequence

Awkley Lane Clay (100%)

Silt (100%) Sand 50% Sand (100%)

Figure 6. Awkley Lane; particle size triangle

Textural features of the minerogenic sequence

Sampling gap in profile from 1.54 to 1.78 m OD.

AWK:5 3.95–5.24 m OD, upper Wentlooge sediments. Quercus-Corylus avellana type-Poaceae-Cyperaceae.

Prehistoric vegetation and environmental changes

Figure 7a. Awkley Lane; pollen diagram.

Figure 7b. Awkley Lane; pollen diagram.

Figure 7c. Awkley Lane; pollen diagram.

Figure 7e. Awkley Lane; pollen diagram.

Figure 7f. Awkley Lane; pollen diagram.

Summary of environmental change

1. Polyhalobous >30 g l-1

Table 3. Coleoptera and other insects from Awkley Lane

Min. no. indiv

Waterlogged Plant Macrofossils

Context 118 118 118 115 120 114 121

Depth (m) OD Layer Description Comment

Figure 9. Vimpenny's Lane Wentlooge Formation profiles

context context type depth OD material lab no GC13‰ result BP calib

Sampling the sequence

Textural features of the upper sedimentary sequence

Vimpenny's Lane Clay (100%)

Silt (100%) Sand 50% Sand (100%)

Figure 10. Vimpenny's Lane; particle size triangle

Two local pollen assemblage subzones have been recognised.

As discussed above the lithostratigraphical sequence at Vimpenny’s Lane comprises largely

For Key See Figure 7a

Figure 11a. Vimpenny's Lane pollen diagram.

% = Sum Total Dry Land Pollen

% sum tdlp + Marsh

% sum tdlp + Spores