MESOLITHIC SETTLEMENT
IN THE NORTH SEA BASIN
A CASE STUDY FROM HOWICK, NORTH-EAST ENGLAND
edited by
Clive Waddington
with contributions by
Geoff Bailey, Alex Bayliss, Alan Biggins, Ian Boomer,
Christopher Bronk Ramsey, Ann Clarke, Jacqui Cotton, Derek Hamilton,
Karen Hardy, Nicky Milner, Kristian Pedersen, Robert Shiel, Tony Stevenson
and Illustrations by Ben Johnson and James Brightman
OXBOW BOOKS
Published by
Oxbow Books, Oxford, UK
© Oxbow Books and the individual authors, 2007
ISBN 978 1 84217 246 9
A CIP record for this book is available from the British Library
This book is published with the aid of a grant
from English Heritage.
English Heritage is now Historic England
This book is available to download from
https://books.casematepublishing.com/Mesolithic Studies in the North Sea basin.pdf
For Marjorie Deakin
CONTENTS
LIST OF CONTRIBUTORS ......................................................................................................................... vii
ACKNOWLEDGEMENTS ............................................................................................................................ ix
SUMMARY.............................................................................................................................................. xi
by Clive Waddington
Résumé ............................................................................................................................................. xiii
Zasammenfassung ........................................................................................................................... xv
Samenvatting .................................................................................................................................. xvii
Dansk Resumé................................................................................................................................. xix
1. INTRODUCTION............................................................................................................................... 1
Clive Waddington
2. GEOPHYSICAL SURVEY .................................................................................................................... 9
Alan Biggins
3. FIELDWALKING ............................................................................................................................. 17
Clive Waddington
4. TEST PITS..................................................................................................................................... 28
Clive Waddington
5. EXCAVATION................................................................................................................................. 34
Clive Waddington
6. ABSOLUTE DATING ....................................................................................................................... 65
Alex Bayliss, Ian Boomer, Christopher Bronk Ramsey,
Derek Hamilton and Clive Waddington
7. CHIPPED STONE TOOLS ................................................................................................................ 75
Clive Waddington
8. BEVELLED PEBBLES, COARSE STONE TOOLS AND OCHREOUS MATERIAL ...................................... 110
Ann Clarke and Clive Waddington
9. RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS ................................................................ 120
Karen Hardy and Robert Shiel
vi
CONTENTS
LEFT-HAND HEADER
10. FAUNAL REMAINS ...................................................................................................................... 137
Geoff Bailey and Nicky Milner
11. PLANT REMAINS AND CHARRED WOOD ..................................................................................... 147
Jacqui Cotton
12. GEOMORPHOLOGICAL AND PALAEOENVIRONMENTAL ANALYSIS ................................................... 163
Ian Boomer, Robert Shiel and Tony Stevenson
13. EXPERIMENTAL CONSTRUCTION................................................................................................... 182
Clive Waddington
14. HOWICK: DISCUSSION AND INTERPRETATION .............................................................................. 189
Clive Waddington, Geoff Bailey and Nicky Milner,
with a contribution by Ann Clarke to the bevelled pebble discussion
15. HOWICK IN ITS NORTH SEA CONTEXT ........................................................................................ 203
Clive Waddington, Geoff Bailey and Nicky Milner
Bibliography .................................................................................................................................... 225
RIGHT-HAND HEADER vii
LIST OF CONTRIBUTORS
GEOFF BAILEY
University of York
Department of Archaeology
Kings Manor
York
YO1 7EP
UK
ALEX BAYLISS
English Heritage
1 Waterhouse Square
138-142 Holborn
London
EC1N 2ST
UK
ALAN BIGGINS
Timescape Surveys
Northumbria
Main Road
Dinnington
Newcastle upon Tyne
NE13 7JW
UK
IAN BOOMER
University of Birmingham
School of Geography
Edgbaston
Birmingham
B15 2TT
UK
CHRISTOPHER BRONK RAMSEY
Oxford Radiocarbon Accelerator Unit
Research Laboratory for Archaeology
and the History of Art
6 Keble Road
Oxford
OX1 3QJ
UK
ANN CLARK
Rockville Lodge
By Kingston
North Berwick
East Lothian
EH39 5JN
UK
JACQUI COTTON
Environment Agency
Rivers House
21 Park Square South
Leeds
LS27 2QG
UK
DEREK HAMILTON
English Heritage
1 Waterhouse Square
138-142 Holborn
London
EC1N 2ST
UK
KAREN HARDY
The Old Toll
Kirkurd
Blyth Bridge
Peebleshire
EH46 7AH
UK
BEN JOHNSON
Archaeological Research Services Ltd
Baltic Business Centre
Saltmeadows Road
Gateshead
NE8 3DA
UK
NICKY MILNER
University of York
Department of Archaeology
Kings Manor
York
YO1 7EP
UK
KRISTIAN PEDERSEN
Archaeological Research Services Ltd
Baltic Business Centre
Saltmeadows Road
Gateshead
NE8 3DA
UK
viii
ROBERT SHIEL
School of Agriculture, Food and Rural Development
King George VI Building
Newcastle University
Newcastle upon Tyne
NE1 7RU
UK
TONY STEVENSON
Newcastle University
School of Geography
Daysh Building
Newcastle upon Tyne
NE1 7RU
UK
LIST OF CONTRIBUTORS
LEFT-HAND HEADER
CLIVE WADDINGTON
Archaeological Research Services Ltd
Angel House
Portland Square
Bakewell
Derbyshire
DE45 1HB
UK
RIGHT-HAND HEADER ix
ACKNOWLEDGEMENTS
The work carried out at Howick and the subsequent
analysis of the data has been a genuine delight to
undertake, providing, during its course, an enjoyable
experience for many. The project has involved people
from all walks of life ranging from professional
archaeologists and students to volunteers, teachers,
the media, retired people and schoolchildren. In all,
over 250 people took part in the fieldwork and it is
estimated that the excavations were visited by around
3000 people. It was a great pleasure to work with so
many people, all of whom added greatly to the project
whether by way of on-site work, enthusiasm or general
support. I am indebted to the landowner, Lord
Howick, who gave permission for the work to take
place and for his support and encouragement through-
out. However, a special mention is due to the farmer,
Mr. Thompson, who helped in virtually all aspects of
the fieldwork and was a generous host, particularly
for all those who camped on the farm for 13 weeks. It
is with pleasure that thanks is given to them and for all
the additional help that included, amongst other
things, arranging for taking aerial photographs from a
helicopter and for a boat ride to see the site from its
seaward side. It is also a pleasure to acknowledge the
co-operation and assistance from the other people of
the Howick-Craster area, who were involved in the
project and made us most welcome.
The erosion of lithics at Howick was first brought
to my attention by Jim Hutchinson, a member of the
Northumberland Archaeology Group, who brought
a collection of flint material from the site to an
artefacts day school being held by the then Centre for
Lifelong Learning of the University of Newcastle.
Although Mesolithic flints had previously been
reported from the erosion scar by John Davies who
referred to it in his 1983 publication, no work had
been carried out there and it is thanks to their
observant discoveries that the site was able to be
recognised and recorded.
The project was funded through the generous
support of the Heritage Lottery Fund, English
Heritage and the University of Newcastle to whom I
am most grateful and who should be recognised for
acting quickly to help record an eroding site. Thanks
is also due to the Society of Antiquaries of Newcastle
upon Tyne and Northumberland County Council
Archaeology Section who provided the funding for
the initial radiocarbon dates. The BBC selected the
site for inclusion in its high-profile ‘Ancestors’ and
‘Coast’ series and supported the project throughout
and I am most grateful to them for this. The BBC
kindly funded the building of the first ‘recon-
struction’ at Milfield and the team of Sophie Robin-
son, Bucy McDonald and Julian Richards deserve
special mention for their help and enthusiasm on the
project. Forest Enterprise and Durham Wildlife Trust
generously supported the project by donating the
timber and grass for the reconstruction. The Air Sea
Rescue team based at RAF Boulmer very kindly
allowed myself and the BBC crew on board a Sea
King helicopter in order to photograph the site and I
am most grateful to them for this. Steve Speake and
Richard Young of Tyne and Wear Museums Service
kindly made available the results of an unpublished
evaluation excavation on the nearby Cushatt Wood
rectilinear enclosure as well as information on the
excavations carried out at Low Hauxley.
The specialist input to this research has been
considerable with all the work being undertaken
notably swiftly which has allowed for its rapid
publication. Thanks is extended to all those who
contributed both in the field and in post-excavation
analyses. In particular Alan Biggins of Timescape
Surveys who undertook two phases of geophysical
survey providing the first survey free of charge when
the project had only limited funds available. The adroit
assistance of Johnny Shipley, Alex Green, Aimee
Lawrence and Dave Henderson during both phases of
the geophysical survey is gratefully acknowledged.
Alex Bayliss of English Heritage has done an out-
standing job on the radiocarbon dating programme,
producing an unrivalled and elegant chronological
sequence for a British Mesolithic site and it was a
delight to work with her on this. Ian Boomer from the
Department of Geography at the University of New-
castle is owed a considerable debt of thanks for the
sediment coring and analysis and pulling together the
environmental components of the project, and for his
generous help in completing the excavation when
resources and time were stretched. Robert Shiel and
Tony Stevenson kindly assisted with the soil and
sediment analyses respectively and I am grateful to
x
them for this. Jacqueline Cotton did a splendid job
identifying and counting the plant macrofossils while
Ann Clarke made a valuable contribution with the
analysis of the coarse stone tools. I would like to thank
Dr Lyall Anderson, Department of Geology and
Zoology, National Museum of Scotland for the
identification of the limestone cobbles and Dr Alison
Sheridan, Department of Archaeology, National
Museum of Scotland for facilitating the geological
identification.
Karen Hardy and Robert Shiel have contributed a
very interesting and informative analysis of residues
and use-wear on a sample of the stone tools and this
important work demonstrates the potential of residue
analysis for future stone tool studies. With regard to
this work, thanks are due to Kristian Pedersen for
making the experimental tools and to Kristian, Andy
Bates, Brian Murray and Victoria Chase for helping
out with the experiments. Oliver Craig is to be
thanked for his illuminating observations and David
Manning is thanked for advice on the analysis of the
residues and comments on the draft text. The SEM
microscope was housed in the Department of Geo-
logy and Geophysics, University of Edinburgh and
Nicola Cayzer and John Craven are thanked for the
training and support they provided.
Peter Forrester of the Design Desk took some
excellent photographs and produced the handsome
information panels and leaflet that can be seen at
Craster Tourist Information Centre. Dave Hall kindly
produced a series of excellent artistic impressions of
the site for use on the panels and interpretive
literature, and again put in far more time on this
work than could have been expected and I am
indebted to him for this. Jo Catling did an excellent
job in devising the schools pack and, together with
Will Graham, Kristian Pedersen, Nicky Milner, Geoff
Bailey, Ian Boomer and myself, undertook a series of
school visits explaining aspects of hunter-gatherer
life and the past environment to Key Stage 2 classes.
English Heritage has been immensely supportive
of this work and deserve special mention for their
help. In particular Kate Wilson, the regional Inspector,
helped with securing support for the project, while
project management was undertaken by Jonathan
Last who helped keep the project running smoothly.
Jacqui Huntley provided support and advice with
regard to plants and macrofossils and Alex Bayliss
and Derek Hamilton have made an important
contribution with the dating mentioned above. I am
also grateful to Chris Scull in Archaeological Com-
missions who made the completion of the analytical
work and the publication of this volume possible.
The professional supervisory staff who worked on
the site during both seasons of fieldwork conducted
themselves with exceptional competence and diligence
and I remain indebted to them for making the
fieldwork such a success, both in terms of recording
ACKNOWLEDGEMENTS
LEFT-HAND HEADER
the archaeology and in training and supervising the
cohorts of volunteers. It is therefore with the greatest
pleasure that I extend my thanks to Richard Chatter-
ton, Will Graham, Claire Henderson, Benjamin
Johnson, Eleanor Johnson and Kristian Pedersen.
Additional professional assistance was provided by
Claire Carey, Gavin Davies, Jo Goodhall, Helena Gray,
Bryan Murray and Richard Smalley. Stuart Winthrope
fastidiously supervised the flotation tank, putting
several tons of sediment through a 2mm mesh as well
as helping with the excavations and treating us all to
his irrepressible humour. Timothy Sandiford endured
endless torment in a dark barn drying out flots with
two hairdryers and, although succumbing to bouts of
cabin fever, his stoic persistence allowed the flots to be
analysed in record time. During the first season of
work Clifford Jones and Ken Stone provided help with
catering, logistics and all round entertainment and I
am very grateful to them for this. During the second
season of work James Whitford organised all cooking
and catering for up to 50 people per day. Together
with James Brightman this pair of redoubtable
troubadours propelled morale to new highs with their
musical talents. Benjamin Johnson and James
Brightman deserve special thanks for their excellent
illustrations and patience in the face of repeated
requests for alterations to figures. Their dependable
help throughout the entire project is also gratefully
acknowledged. Kristian Pedersen, the project officer,
worked above and beyond the call of duty in carrying
out a range of tasks including setting up the project
website (www. ncl.ac.uk/howick) and organising an
international conference on the Mesolithic of the North
Sea Basin at the University of Newcastle. The
proceedings of this meeting are published separately
by Oxbow Books. I would also like to thank my project
co-directors Geoff Bailey and Nicky Milner who helped
from the beginning with the organisation of the project,
fieldwork and specialist analyses of the faunal remains,
as well as dealing with media interviews and
publications and their input to this volume.
I am grateful to Jonathan Last for his editorial work
at English Heritage and to the anonymous referees
who gave their time to reading the text and improving
the book through useful and detailed comment.
Andrew David, also at English Heritage, kindly
undertook editorial work on the geophysical chapter
at a time when the chapter’s author was diagnosed
with a sudden and serious illness. Roger Jacobi kindly
assisted with the identification of some chipped stone
tools and Sönke Hartz kindly drew my attention to the
evidence for the hazelnut roasting at Duvensee. I must
extend a special thank you to Myra Wilkinson van
Hoek, of Fine Line Archaeological Language Services,
who kindly undertook the copy editing of this volume
in remarkably quick time. Finally, the team at Oxbow
Books deserve a special thanks for their patience and
support in bringing this volume to publication.
RIGHT-HAND HEADER xi
SUMMARY
The first indication of a Mesolithic site at Howick
was noticed on separate occasions by two amateur
archaeologists who each discovered flints eroding
from the cliff edge. Subsequently a small test pit,
together with geophysical survey, confirmed the
presence of undisturbed Mesolithic deposits below
the ploughzone. After an excavation season in 2000
to evaluate the deposits, a full-scale excavation was
mounted in 2002 to record, in full, all archaeological
remains that were in danger of further erosion.
The archaeological remains at Howick consist of a
Mesolithic hut site and an Early Bronze Age cist
cemetery located on a modern cliff edge overlooking
a small estuary. This volume is devoted solely to the
reporting and interpretation of the Mesolithic remains
while the Bronze Age material is reported separately
as a journal article (Waddington et al. 2005), as is the
full Holocene environmental sequence obtained from
the sediment core (Boomer et al. 2007). Few Mesolithic
hut sites have been discovered in the British Isles
whereas in other regions, such as southern Scandin-
avia, hut sites belonging to this period are more
common. The relative lack of sites in Britain may be
explained by a combination of reasons including,
most notably, the differential effects of sea level
change and site survival in different areas, the lack of
large open area excavations in coastal areas of the
UK, less focus in recent years on the Mesolithic period
amongst the British archaeological research com-
munity compared to their counterparts in Scand-
inavia, and perhaps a failure to recognise such sites
when they have been revealed during previous
excavations and field evaluations. It is exciting to
note that within months of completing the Howick
excavations, an almost identical site was excavated
80km to the north at East Barns near Dunbar, which
also lies on the North Sea coast (Gooder 2007).
Together with the previously known sites of Mount
Sandel and Broomhill, these four hut sites provide
the least ambiguous evidence for settlements in the
British and Irish Mesolithic to date.
Three huts had been constructed on the Howick
site, all on the same footprint, with no evidence to
indicate a gap between these occupations. This was
confirmed by the radiocarbon dates, which suggest
these different phases followed on continuously one
to another. Whether the site was occupied on a
seasonal basis or as a permanent residence remains
open to question, and this will no doubt form one of
the key outstanding questions relating to this site. A
fourth phase of very short-lived occupation was
identified by the radiocarbon dating programme,
occurring after the hut had been abandoned and
probably representing a transitory use of the hollow
for an overnight camp. The residues from the Howick
site have been subjected to a comprehensive radio-
carbon dating programme and the results have been
statistically modelled using a Bayesian approach to
produce a singularly well-understood chronological
sequence. The initial construction of the site took
place around 7850 cal BC and abandonment took
place in the years around 7650 cal BC.
The remains inside the hut were all consistent with
its use as a habitation site. These consisted of a central
arrangement of hearths that contained burnt animal
bones and masses of charred hazelnut shell frag-
ments, activity areas represented by the distribution
of stone artefacts, and the presence of a hazelnut
roasting pit indicative of strategies associated with
food storage. Although the primary function of the
hut was undoubtedly as a dwelling structure this
does not mean to say that aspects of residential life
did not include ritualised behaviour. Indeed it is
unlikely that domestic and ritual life was separated
out by Stone Age hunter-gatherer groups, who are
more likely to have incorporated ‘ritual’ into all
aspects of life. However, it is important to stress that
there was no obvious physical indication of ritualised
behaviour within the excavated deposits and that no
finds present on the site could be construed as being
specially placed deposits or of intrinsic ritual signif-
icance. Neither was there any sign of burial or the
disposal of human remains. In this way the hut site
was no different from those at Mount Sandel,
Broomhill and East Barns, which produced remark-
ably similar remains to those from Howick.
The lithic material from Howick is the most
accurately dated assemblage from any British Meso-
lithic site and is a classic example of a narrow-blade
industry associated with the micro-triangle techno-
xii
complex, with scalene triangles the most common
form of microlith. Typically for Britain, these sites
date from around 7500 cal BC but the Howick dates
indicate an earlier start for this type of industry in the
centuries immediately after 8000 cal BC. More recent
dates from East Barns accord directly with the earliest
of the Howick dates. Still earlier dates have recently
been acquired for a micro-triangle site at Cramond
near Edinburgh, which now provide the earliest dates
so far for this industry in Britain at around 8400 cal
BC. This raises interesting questions over the origin
and spread of this type of technology, its relationship
with North European industries, as well as the
overlap with what are traditionally termed ‘Earlier
Mesolithic’ artefact industries in southern England.
The chipped stone assemblage from Howick is all
made from locally occurring beach pebble flint and
this fits into the wider pattern of localised raw
material acquisition by groups elsewhere in North-
East England. A wide variety of tool types is present
from within the hut, reflecting the diverse activities
that appear to have taken place there. Evidence for
task diversity is further attested by the presence of a
large bevelled pebble tool assemblage, huge quant-
ities of charred hazelnut shells, some ochre fragments
and the presence of a range of animals.
The Howick site lies in an ecotonal setting with
access to a remarkably resource-rich catchment. With
such a wide range of resources on offer on a year-
round basis, the site is interpreted as a base camp-
type settlement that was used by the same group and
their descendants over a period of several generations
lasting for somewhere in the region of 200 years. The
SUMMARY
LEFT-HAND HEADER
size of the hut indicates its use by a family-sized
group. There was no surviving evidence for any more
than one structure on the site at any one time,
although land on the seaward side of the site has
been lost through erosion and this land could have
contained such evidence. The fieldwalking evidence
suggests that other Mesolithic settlement foci existed
nearby, all in easy reach of the freshwater stream,
which could suggest the presence of other similar
structures nearby. This provides an impression of
family groups occupying sturdy conical-shaped huts
separated by several hundred metres, but all with
easy access to freshwater and the shore, and easy
access to each other. No midden site was found at
Howick and this is almost certainly due to the erosion
of the contemporary Mesolithic shoreline and the
effects of the Main Post-Glacial Marine Transgression,
which would have removed and/or drowned any
midden deposits near to the 8th millennium cal BC
shore.
The Howick excavations have forced a rethink of
the scale and nature of Mesolithic settlement in
North-East England, as well as the relationship
between this and other regions around the North Sea
Basin. It has also demonstrated the potential of
surviving deposits dating to this period as well as the
type of residues and information that such sites can
be expected to yield. Hopefully this work will help
encourage further research into the Mesolithic of the
region and its interactions with adjacent areas of
upland, other North Sea Basin communities, as well
as groups occupying the lands further north and
south.
RIGHT-HAND HEADER xiii
RÉSUMÉ
Les premiers signes d’un site mésolithique à Howick
furent remarqués séparément par deux archéologues
amateurs, qui avaient chacun découvert des silex qui
se détachaient de la paroi de la falaise sous l’effet de
l’érosion. Par la suite, une petite tranchée de sondage,
ainsi que la prospection géophysique confirmèrent la
présence de dépôts mésolithiques non perturbés sous
la zone labourée. Après une saison de fouilles en l’an
2000 pour évaluer les dépôts, des fouilles complètes
furent entreprises en 2002 pour noter, dans leur
intégralité, tous les restes archéologiques qui
risquaient de s’éroder davantage.
Les restes archéologiques de Howick comprennent
un site de huttes mésolithiques et une nécropole de
cistes de l’Age du Bronze situés au bord d’une falaise
moderne qui domine un petit estuaire. Ce volume se
consacre exclusivement à la description et l’inter-
prétation des restes mésolithiques; les restes de l’Age
du Bronze sont décrits séparément dans un article
d’une revue archéologique (Waddington et al. en
cours de publication), de même que la séquence
environnementale intégrale du Holocène obtenue par
carottage des sédiments (Boomer et al. en cours de
publication). Peu de sites de huttes mésolithiques ont
été découverts dans les Iles Britanniques, alors que
dans d’autres régions, comme par exemple le sud de
la Scandinavie, les sites de huttes appartenant à cette
période sont bien plus courants. Le manque relatif de
ces sites en Grande-Bretagne s’explique peut-être par
un ensemble de raisons, notamment, les effets
différentiels du changement du niveau de la mer et la
survie des sites à différents endroits, le manque de
grandes surfaces de fouilles dans les zones côtières
du Royaume-Uni, l’attention réduite portée à la
période du Mésolithique ces dernières années par la
communauté d’archéologues-chercheurs de Grande-
Bretagne, contrairement à leurs homologues
scandinaves, et peut-être le fait que de tels sites
n’aient pas été reconnus en tant que tels au moment
de leurs découvertes lors de fouilles et d’évaluations
de terrain antérieures. Il est encourageant de noter
que dans les mois qui suivirent l’achèvement des
fouilles à Howick, un site presque identique fut mis
au jour à 80km au nord de East Barns, à côté de
Dunbar, qui se trouve également sur la côte de la Mer
du Nord (Gooder en cours de publication). Con-
sidérés conjointement avec les sites déjà connus
de Mount Sandel et Broomhill, ces quatre sites de
huttes fournissent la preuve la moins ambiguë à ce
jour d’établissements au Mésolithique en Grande-
Bretagne et en Irlande.
Trois huttes avaient été construites sur le site de
Howick, toutes sur le même emplacement, sans
preuves qui indiqueraient une interruption entre ces
occupations. Ceci fut confirmé par la datation par le
radiocarbone, qui suggère que ces différentes phases
s’enchaînèrent sans discontinuité. Le débat reste
ouvert pour déterminer si le site était occupé de
manière saisonnière ou comme lieu de résidence
permanente, et cela formera sans aucun doute une
des dernières questions clés liées à ce site. Une
quatrième phase d’occupation très courte fut
identifiée par le programme de datation par le
radiocarbone : elle eut lieu après l’abandon de la
hutte, et représente probablement une utilisation
transitoire de la cuvette laissée par la hutte pour un
campement d’une nuit. Les résidus provenant de
Howick ont été soumis à un programme complet de
datation par le radiocarbone, et les résultats ont été
modélisés grâce à l’analyse statistique Bayésienne
afin de créer une séquence chronologique très bien
comprise. La construction initiale du site eut lieu vers
7850 avant JC, et il fut abandonné aux alentours de
7650 avant JC.
Les restes à l’intérieur des huttes correspondaient
tous à leur utilisation comme site d’habitation. Ceux-
ci se composaient d’un arrangement central de foyers
qui contenaient des os d’animaux calcinés et
quantités de fragments de coquilles de noisettes
carbonisées, de zones d’activité représentées par la
distribution d’artefacts en pierre, et de la présence
d’une fosse pour griller les noisettes qui indique des
stratégies associées au stockage des aliments. Bien
que la fonction principale de la hutte fût sans aucun
doute celle d’une structure d’habitation, cela n’a pas
empêché certains aspects de la vie résidentielle de
comporter une part de rituel. En effet, il est im-
probable que la vie domestique et rituelle ait été
considérée séparément par les groupes de chasseurs-
cueilleurs de l’Age de Pierre, qui auraient
xiv
certainement incorporé le « rituel » à tous les aspects
de la vie. Toutefois, il faut remarquer qu’il n’y avait
pas d’indication physique nette de comportement
rituel au sein des dépôts fouillés et qu’aucune
découverte présente sur le site ne pourrait être
interprétée comme une déposition spéciale ou comme
ayant une signification rituelle propre. Il n’y avait pas
non plus de signe d’inhumation ou de suppression
de restes humains. En ce sens, le site de huttes n’était
pas différent de ceux de Mount Sandel, Broomhill et
East Barns, qui ont révélé des restes remarquable-
ment similaires à ceux de Howick.
Les pièces lithiques provenant de Howick con-
stituent l’ensemble daté de manière la plus exacte de
tous les sites mésolithiques britanniques et elles
forment un exemple classique d’une industrie de
lames étroites associée au techno-complexe de
pointes microlithiques triangulaires, avec des
triangles scalènes, la forme microlithique la plus
fréquente. Typiquement pour la Grande-Bretagne, ces
sites datent d’environ 7500 avant JC calibré, mais les
dates de Howick montrent que ce type d’industrie
avait commencé plus tôt, au cours des siècles qui
suivirent immédiatement 8000 avant JC calibré. Des
datations plus récentes de East Barns concordent
directement avec les dates les plus anciennes de
Howick. On a récemment acquis des dates encore
plus anciennes pour le site de pointes microlithiques
triangulaires de Cramond, à côté d’Edimbourg, qui,
considéré conjointement avec une date ancienne
provenant de Daer Reservoir, aussi en Ecosse,
constituent désormais les dates les plus anciennes, à
ce jour, de la présence de cette industrie en Grande-
Bretagne, aux environs de 8400 avant JC calibré. Ceci
soulève d’intéressantes questions quant à l’origine et
la diffusion de ce type de technologie, ses liens avec
les industries de l’Europe du Nord, aussi bien que le
chevauchement avec ce qu’on appelle tradition-
nellement les industries d’artefacts du « Mésolithique
plus ancien » du sud de l’Angleterre.
L’ensemble d’éclats de pierre de Howick provient
entièrement de galets de silex de la plage locale et
ceci correspond au modèle plus général d’acquisition
de matière première localisée par d’autres groupes
du nord-est de l’Angleterre. Une grande variété de
types d’outils est présente à l’intérieur de la hutte,
représentant les activités diverses qui semblent y
avoir eu lieu. On constate d’autant plus la diversité
des activités grâce à la présence d’un grand ensemble
RÉSUMÉ
LEFT-HAND HEADER
d’outils de pierre à bords biseautés, à la quantité
énorme de coquilles de noisettes carbonisées, aux
quelques fragments d’ocre, et à la présence d’une
variété d’animaux.
Le site de Howick se situe dans un écotone mettant
à disposition une zone remarquablement riche en
ressources. Avec une telle sélection de ressources
disponibles tout au long de l’année, le site a été
interprété comme un habitat de type camp de base
utilisé par le même groupe et leurs descendants
pendant plusieurs générations, sur une période
d’environ 200 ans. La taille de la hutte indique qu’elle
était utilisée par un groupe de la taille d’une famille.
Aucune preuve ne nous est parvenue laissant à
penser qu’il y ait eu plus d’une structure à la fois sur
le site à quelque moment donné que ce soit, quoique
les terres du côté du site face à la mer aient été
perdues à cause de l’érosion et ces terres auraient pu
contenir de telles preuves. Les preuves collectées lors
de ramassages de surface suggèrent que d’autres
lieux d’habitations mésolithiques existaient à
proximité, tous à portée du ruisseau d’eau douce, ce
qui pourrait suggérer la présence d’autres structures
similaires dans les environs. On imagine des groupes
familiaux qui occupaient de solides huttes de forme
conique séparées par plusieurs centaines de mètres,
mais tous ayant un accès facile à l’eau douce et au
rivage, et facilement accessible les uns des autres. On
n’a trouvé aucun site d’amoncellements des détritus
à Howick et cela tient presque incontestablement à
l’érosion du rivage contemporain au Mésolithique et
les effets de la Transgression Marine Post-Glaciaire
Principale, qui auraient enlevé et/ou englouti tout
amoncellement près de la rive du 8ème millénaire avant
JC calibré.
Les fouilles de Howick nous ont forcés à recon-
sidérer l’étendue et la nature des établissements
mésolithiques du nord-est de l’Angleterre, aussi bien
que la relation entre cette région et celles qui bordent
le bassin de la Mer du Nord. Elles ont également
montré le potentiel des dépôts conservés datant de
cette période, aussi bien que les types de résidus et
d’informations qu’on peut attendre de tels sites. Avec
un peu de chance, ces travaux inciteront à plus de
recherches sur le Mésolithique de la région et sur
l’interaction des communautés avec celles des zones
de plateaux adjacentes, celles du Bassin de la Mer du
Nord, aussi bien qu’avec les groupes qui occupaient
les terres plus au nord et au sud.
RIGHT-HAND HEADER xv
ZUASAMMENFASSUNG
Erste Anzeichen eines mesolithischen Fundplatzes in
Howick wurden zu verschiedenen Gelegenheiten von
zwei Amateurarchäologen bemerkt, die beide aus einer
Felskante erodierende Silexgeräte beobachteten. Später
bestätigten eine kleine Sondagegrabung und eine
geophysikalische Prospektion, dass unterhalb des
Pflughorizontes ungestörte mesolithische Schichten
vorlagen. Nach einer Grabungssaison im Jahre 2000,
die die Schichten beurteilen sollte, wurde 2002 eine
umfassende Grabung begonnen, deren Aufgabe es war,
alle archäologischen Hinterlassenschaften, die der
Gefahr weiterer Erosion ausgesetzt waren, zu erfassen.
Die archäologischen Hinterlassenschaften in Howick
bestehen aus mesolithischen Hütten und einem
frühbronzezeitlichen Steinkistengräberfeld. Sie
befinden sich am Rande einer neuzeitlichen Felskante,
die eine kleine, den Gezeiten ausgesetzte Flussmün-
dung überblickt. Dieser Band widmet sich
ausschließlich der Beschreibung und Interpretation
der mesolithischen Hinterlassenschaften. Das bronze-
zeitliche Material (Waddington et al. im Druck), sowie
die gesamte holozäne Umweltentwicklung, die durch
Sedimentbohrungen aufgeschlüsselt wurde (Boomer
et al. im Druck), werden gesondert als Beiträge in
Zeitschriften vorgestellt. Auf den Britischen Inseln
sind nur wenige mesolithische Hüttenfundplätze
bekannt, wohingegen in anderen Regionen, wie
beispielsweise Südskandinavien, Hütten dieser
Zeitstellung häufiger sind. Die relative Seltenheit
solcher Fundstellen in Großbritannien kann durch das
Zusammenspiel verschiedener Gründe erklärt
werden. Die wohl wichtigsten sind die unter-
schiedlichen Auswirkungen der Meeresspiegel-
schwankungen und die Erhaltung von Fundstellen in
verschiedenen Gebieten, die Tatsache, dass bisher
keine großflächigen Ausgrabungen in Küstengebieten
Großbritanniens durchgeführt worden sind, das in
den letzten Jahren geringere Interesse der britischen
Forschungsgemeinschaft am Mesolithikum, vor allem
verglichen mit ihren skandinavischen Kollegen, und
vielleicht auch die Unfähigkeit solche Fundstellen zu
erkennen, wo sie durch vorherige Ausgrabungen und
Voruntersuchungen entdeckt wurden. Spannender-
weise wurde nur wenige Monate, nachdem die
Ausgrabung in Howick beendet war, 80km nördlich
ein fast identischer Fundplatz bei East Barns in der
Nähe von Dunbar ausgegraben, der ebenfalls an der
Nordseeküste liegt (Gooder im Druck). Zusammen
mit den bereits bekannten Fundstellen Mount Sandel
und Broomhill liefern diese Plätze die bisher ein-
deutigsten Nachweise für Siedlungen im britischen
und irischen Mesolithikum.
In Howick waren drei Hütten erbaut worden, alle
nach dem gleichen Schema und ohne Hinweise auf
eine Besiedlungslücke. Dies wurde durch die Radio-
karbondaten bestätigt, die darauf hindeuten, dass die
verschiedenen Belegungsphasen ohne Unterbrechung
aufeinander folgten. Ob dieser Platz saisonal oder als
permanenter Aufenthaltsort genutzt wurde ist noch
offen und wird zweifellos eine der wichtigsten Fragen
für unser Verständnis der Fundstelle bleiben. Die
Radiokarbondatierung stellte auch eine vierte, sehr
kurze Besiedlungsphase fest, die folgte, als die Hütte
schon aufgegeben worden war. Wahrscheinlich
handelt es sich hierbei um eine vorübergehende
Nutzung der Senke als kurzzeitigen Lagerplatz. Für
die Hinterlassenschaften aus Howick wurde eine
umfassende Serie von Radiokarbondaten gemessen,
die durch Anwendung eines bayesischen Verfahrens
statistisch ausgewertet wurden. Dies führte dazu,
dass die chronologische Abfolge außergewöhnlich
gut aufgeschlüsselt werden konnte. Das erste
Gebäude wurde um 7850 cal BC errichtet und der
Platz um 7650 cal BC aufgegeben.
Die Überreste in der Hütte selbst stimmten alle mit
einer Nutzung des Gebäudes als Wohnplatz überein.
Sie bestanden aus zentral angeordneten Herdstellen,
die verbrannte Tierknochen und massenweise verkohle
Fragmente von Haselnusschalen enthielten. Die
Verteilung der Steingeräte ergab Aktivitätszonen und
eine Grube zum Rösten von Haselnüssen weist auf
Strategien zur Lagerung von Nahrungsmitteln hin.
Auch wenn die Hütte zweifellos hauptsächlich als
Behausung fungierte, so heißt das keineswegs, dass
Aspekte des Lebens hier nicht auch ritualisierte
Verhaltensweisen beinhalteten. In der Tat ist es
unwahrscheinlich, dass die steinzeitlichen Wildbeuter-
xvi
gemeinschaften häusliches Leben und Ritualleben
auseinanderhielten. Es scheint plausibler, dass das
„Rituelle“ in alle Lebensbereiche mit einbezogen war.
Es muss jedoch hervorgehoben werden, dass es in den
ausgegrabenen Schichten keine offensichtlichen
Hinweise auf rituelles Verhalten gab, und dass keine
der Funde als intentionell deponiert oder von
intrinsischer ritueller Bedeutung gedeutet werden
können. Noch gab es Anzeichen einer Grablege oder
der Deponierung menschlichen Skelettmaterials. In
dieser Hinsicht war dieser Siedlungsplatz nicht anders
als Mount Sandel, Broomhill und East Barns, die alle
Hinterlassenschaften ergaben, die denen Howicks
erstaunlich ähnlich sind.
Die Steingeräte aus Howick sind der am genauesten
datierte Komplex aus dem gesamten britischen
Mesolithikum. Sie sind ein klassisches Beispiel für
eine auf schmalen Klingen basierende Industrie, die
mit dem Technokomplex der Dreiecksmikrolithen
verbunden ist. Ungleichschenklige Dreiecke sind die
häufigste Mikrolithform. Wie für Großbritannien
typisch, datieren diese Fundstellen ab 7500 cal BC,
aber die Daten aus Howick zeigen, dass diese
Industrie bereits früher, in den Jahrhunderten
unmittelbar nach 8000 cal BC, begann. Die jüngst aus
East Barns gewonnen Datierungen stimmen direkt
mit den frühesten aus Howick überein. Ein noch
höheres Alter wurde kürzlich für einen Fundplatz
mit Dreiecksmikrolithen in Cramond bei Edinburgh
ermittelt. Gemeinsam mit einem frühen Datum aus
Daer Reservoir, ebenfalls in Schottland, sind diese
Datierungen um 8400 cal BC nun die bislang ältesten
für diese Industrie in Großbritannien. Dies wirft
interessante Fragen zu Ursprung und Verbreitung
dieser Technologie auf, sowie zu ihrem Verhältnis zu
nordeuropäischen Industrien und der Über-
schneidung mit dem, was man traditionell die
„frühmesolithischen“ Werkzeugindustrien Süden-
glands nennt.
Die Silexgeräte aus Howick sind alle aus am örtlichen
Strand vorkommenden Knollen hergestellt. Dies fügt
sich gut in das weiter verbreitete Muster einer örtlich
begrenzten Rohmaterialversorgung durch Wild-
beutergruppen in anderen Teilen Nordostenglands.
Aus der Hütte selbst stammt eine große Vielfalt von
Gerätetypen, die die verschiedenen Tätigkeiten, die
offensichtlich hier stattfanden, widerspiegeln.
Weitere Hinweise auf weit gefächerte Tätigkeiten sind
das häufige Vorkommen von Geräten aus schräg
ZUASAMMENFASSUNG
LEFT-HAND HEADER
facettierten Knollen, große Mengen an verkohlten
Haselnusschalen, einige Rötelfragmente und eine
Reihe verschiedener Tierarten.
Die Fundstelle von Howick liegt in einem Ökoton, das
Zugriff auf ein ungewöhnlich artenreiches Gebiet
erlaubt. Da hier eine solche Bandbreite an Ressourcen
das ganze Jahr über zur Verfügung stand, wird der
Fundplatz als Basislager interpretiert, das von der
gleichen Gruppe und deren Nachkommen über
mehrere Generationen hinweg, wohl insgesamt um die
200 Jahre, genutzt wurde. Die Ausmaße der Hütte
deuten auf ihre Nutzung durch eine Gruppe von der
Größe einer Kernfamilie hin. Hinweise auf mehr als
ein gleichzeitiges Gebäude haben sich an diesem Platz
nicht erhalten. Jedoch hat an der dem Meer
zugewandten Seite Landverlust durch Erosion
stattgefunden, wobei solche Hinweise verloren
gegangen sein könnten. Aufgelesene Oberflächenfunde
deuten darauf hin, dass es in der Nähe weitere
mesolithische Siedlungen gab, die alle leichten Zugang
zum Süßwasser führenden Bach hatten. Auch hier
könnten sich Hinweise auf ähnliche Gebäude finden.
Zusammen ergibt dies den Eindruck von
Familienverbänden, die stabile, konische Hütten
bewohnten, welche einige hundert Meter voneinander
entfernt waren. Alle hatten leicht Zugang zu Süßwasser
und zur Küste, sowie leichten Zugang zueinander. In
Howick wurden keine Überreste von Abfallhaufen
gefunden, was höchstwahrscheinlich an der Erosion
der damaligen mesolithischen Küste liegt, sowie an
den Auswirkungen der postglazialen marinen
Transgression, die eventuelle Ablagerungen nahe der
im 8. vorchristlichen Jahrtausend bestehenden Küste
weggespült und/oder überschwemmt hätte.
Die Ausgrabungen in Howick zwingen dazu, Ausmaß
und Charakter des mesolithischen Siedlungswesens in
Nordostengland, sowie das Verhältnis zwischen dieser
und anderen Regionen um das Nordseebecken, neu zu
überdenken. Sie haben auch bewiesen, dass Hinter-
lassenschaften aus dieser Zeit durchaus überdauern
können und gezeigt, welche Überreste und
Informationen man von solchen Fundplätzen erwarten
kann. Hoffentlich wird diese Arbeit dazu beitragen,
weitere Forschungen zum Mesolithikum in dieser
Region anzuregen, sowie ihre Wechselbeziehungen mit
benachbarten Gebieten im Hochland, mit anderen
Gemeinschaften des Nordseebeckens und mit weiter
nördlich und südlich siedelnden Gruppen zu
erkunden.
RIGHT-HAND HEADER xvii
SAMENVATTING
De eerste aanwijzingen voor een Mesolitische
vindplaats nabij Howick kwamen aan het licht toen
twee amateur-archeologen onafhankelijk van elkaar
vuursteen ontdekten dat uit de klifrand erodeerde.
Een kleine testput en geofysische survey bevestigden
de aanwezigheid van onverstoorde Mesolitische
lagen onder het maaiveld. Een evaluatieopgraving in
2000 werd gevolgd door volledige opgraving in 2002,
met als doel het registreren van alle archeologische
indicatoren die onder bedreiging stonden van verdere
erosie.
De vindplaats Howick bestaat uit een Mesolithische
hutplattegrond en een begraafplaats uit Vroege
Bronstijd en bevindt zich op de rand van een
hedendaagse klif die uitkijkt over een kleine
riviermond. Deze publicatie is geheel gewijd aan de
rapportage en interpretatie van het Mesolithische
materiaal; een verslag van het Bronstijdmateriaal zal
als apart artikel verschijnen (Waddington et al. in druk),
evenals de opbouw van het complete Holocene pakket
dat verkregen werd uit de boorstaaf (Boomer et al. in
druk). Op de Britse eilanden zijn tot nog toe maar
weinig Mesolitische hutplattegronden aangetroffen, in
tegenstelling tot gebieden als zuid-Skandinavië, waar
vindplaatsen uit dit tijdperk vaker voorkomen. Dit
relatieve gebrek aan sites in Groot-Brittannië valt te
wijten aan een combinatie van factoren, waaronder
voornamelijk: het differentiële effect van
zeespiegelverandering en de overlevingskans van
archeologische indicatoren in verschillende gebieden;
het gebrek aan grootschalige opgravingen in Britse
kustgebieden; een verminderde recente nadruk op het
Mesolithicum onder Britse onderzoekers vergeleken
met hun Skandinavische collega’s, en wellicht een
onvermogen om zulke sites als zodanig te herkennen
wanneer zij aan het licht komen tijdens opgravingen
en evaluerend veldwerk. Een interessante
ontwikkeling was dat binnen enkele maanden na
afronding van de Howick-opgraving een nagenoeg
identieke site werd opgegraven, 80 km ten noorden
van Howick in East Barns nabij Dunbar, eveneens
gesitueerd aan de Noordzeekust (Gooder in druk). In
combinatie met de twee eerder bekende sites Mount
Sandel en Broomhill vormen deze vier vindplaatsen
de meest overtuigende aanwijzingen voor
nederzettingen in het Britse en Ierse Mesolithicum tot
dusver.
De Howick-site omvat drie hutten, alle met dezelfde
plattegrond; er waren geen aanwijzingen voor
onderbrekingen in de bewoning. Dit werd bevestigd
door de radiokoolstofdateringen, die aangeven dat de
verschillende fasen elkaar ononderbroken opvolgden.
Of de bewoning seizoensgebonden of permanent was
is vooralsnog onduidelijk, en dit zal ongetwijfeld een
onbeantwoorde kernvraag omtrent deze site vormen.
De koolstofdateringen brachten een vierde fase van
zeer korte bewoning aan het licht nadat de hut verlaten
was – waarschijnlijk een kort gebruik van de plek als
overnachtingsplaats. Monsters van Howick werden
onderworpen aan een uitgebreid 14C-daterings-
programma; met de resultaten hiervan werd
vervolgens middels een Bayesiaans statistisch model
een uitzonderlijk duidelijke chronologie opgesteld: de
bouw van de site begon rond 7850 cal BC en de hut
werd verlaten rond 7650 cal BC.
Binnenin de hut wezen alle indicatoren op gebruik
als bewoningsplek. Zij bestonden uit een groep centraal
gelegen haarden die verbrand dierlijk botmateriaal en
enorme hoeveelheden verkoolde resten van
hazelnootdoppen bevatten. De distributie van stenen
artefacten vormde een aanwijzing voor waar in de hut
bepaalde activiteiten werden uitgevoerd, en een kuil
voor het roosteren van hazelnoten wees op strategieën
voor de opslag van voedsel. Alhoewel de hut zonder
twijfel in de eerste plaats als woonplek bedoeld was,
wil dit niet zeggen dat ritueel gedrag geen deel
uitmaakte van het dagelijks leven. Het is
onwaarschijnlijk dat het dagelijks leven en ritueel
gedrag van elkaar gescheiden waren bij deze jagers-
verzamelaars uit de Steentijd; rituelen zullen deel
hebben uitgemaakt van alle aspecten van het dagelijks
leven. Het dient echter benadrukt te worden dat er
geen in het oog springende fysieke aanwijzingen voor
ritueel gedrag werden aangetroffen in de opgegraven
lagen en dat geen enkele vondst geïnterpreteerd kon
worden als zijnde specifiek ritueel geplaatst of van
enig andere rituele betekenis. Ook van begraving of
het anderszins ontdoen van menselijke resten was geen
sprake. In dit opzicht verschilde Howick niet van
Mount Sandel, Broomhill of East Barns, waar
xviii
opvallend gelijksoortige archeologische indicatoren
werden aangetroffen.
Het lithische materiaal van Howick is het meest
precies gedateerde van alle Britse Mesolithische sites
en is een klassiek voorbeeld van een technologie van
smalle klingen geassocieerd met het micro-
driehoekcomplex; de ongelijkbenige driehoek is de
meest voorkomende microlitische vorm. In Groot-
Brittannië dateren zulke sites typisch van rond 7500
cal BC; de Howick-dateringen geven echter aan dat
dit type technologie al eerder begon, namelijk in de
eerste eeuwen na 8000 cal BC. Recentere dateringen
van East Barns zijn in directe overeenstemming met
de vroegste Howick-datering. Nog vroegere
dateringen komen van een micro-driehoek site in
Cramond nabij Edinburgh; samen met een vroege
datering van Daer Reservoir (eveneens in Schotland)
vormen deze de vroegste dateringen tot nog toe voor
deze technologie in Groot-Brittannië: ca. 8400 cal BC.
Dit brengt interessante vragen met zich mee omtrent
de oorsprong en verspreiding van dit type
technologie, de relatie met Noordeuropese en de
overlap met zuid-Engelse, traditioneel ‘Vroeger
Mesolithisch’genoemde, tegenhangers.
De collectie bewerkt steen van Howick bestaat
geheel uit vuurstenen kiezels afkomstig van het
nabijgelegen strand, wat strookt met het patroon van
lokale grondstofverwerving elders in noordoost-
Engeland. De hut bevatte een breed scala van
werktuigen, wat erop wijst dat de aldaar uitvoerde
activiteiten zeer divers waren. Deze interpretatie voor
een grote diversiteit in taken wordt ondersteund door
de aanwezigheid van een grote collectie schuin
afgewerkte werktuigen vervaardigd van kiezels,
enorme hoeveelheden verkoolde resten van
hazelnootdoppen, enkele stukjes oker en
overblijfselen van een grote verscheidenheid aan
diersoorten.
De vindplaats Howick ligt in een ecotoon gebied
dat toegang verschaft tot een regio zeer rijk aan
SAMENVATTING
LEFT-HAND HEADER
voedselbronnen en grondstoffen. Gezien deze
rijkdom, die bovendien het hele jaar door beschikbaar
is, wordt Howick geïnterpreteerd als een basiskamp,
ca. 200 jaar lang gebruikt door dezelfde groep en
diens afstammelingen. Het formaat van de hut wijst
op gebruik door een groep ter grootte van een gezin.
Er waren geen aanwijzingen dat er zich in enige
periode meer dan één bouwwerk tegelijk bevond,
hoewel land aan de zeezijde van de site, door erosie
verloren gegaan, aanwijzingen hiervoor zou kunnen
hebben bevat. Materiaal verzameld tijdens de
veldverkenningen wijst op andere Mesolithische
activiteit in de buurt, steeds in de nabijheid van zoet
water, wat zou kunnen betekenen dat er zich in de
omtrek nog andere gelijksoortige structuren
bevinden. Dit alles geeft de indruk van kleine
groepen, woonachtig in stevige, kegelvormige hutten
die een paar honderd meter uit elkaar liggen maar
alle in buurt van zowel zoet water als de kust en die
makkelijk onderling bereikbaar zijn. Er werden geen
schelpenhopen aangetroffen in Howick, wat vrijwel
zeker te wijten is aan erosie van de Mesolithische
kustlijn en de effecten van de postglaciale mariene
transgressie; schelpenhopen die in het achtste
millenium cal BC dicht bij de kustlijn lagen zullen
hierdoor weggespoeld of verdronken zijn.
De Howick-opgraving dwingt tot een
heroverdenking van de schaal en de aard van het
Mesolithische nederzettingspatroon in noordoost-
Engeland, evenals van de relatie tussen deze en
andere regio’s rond de Noordzee. Tevens is
aangetoond hoeveel informatie indicatoren uit deze
periode kunnen bevatten en wat voor soort
informatie verwacht kan worden van dit type
vindplaatsen. Hopelijk moedigt het hier besproken
werk aan tot verder onderzoek naar het Mesolithicum
van deze regio, de interactie met aangrenzende hoger
gelegen gebieden en met groepen ten noorden en ten
zuiden, alsmede met andere gemeenschappen rond
het Noordzeebekken.
Translated by Myra Wilkinson-van Hoek
Fine Line Archaeolological Language Services
RIGHT-HAND HEADER xix
DANSK RESUMÉ
Howick lokalitetens mulige status som mesolitisk
boplads blev først bemærket af to amatørarkæologer,
som uafhængigt af hinanden havde fundet bearbejdet
flint i stedets eroderende klinter. En prøvegravning,
kombineret med geofysiske undersøgelser, bekræftede
tilstedeværelsen af uforstyrrede mesolitiske strata
under pløjelaget. Efter den første udgravningssæson i
2000, hvis formål var at vurdere karakteren af disse
lag, blev en egentlig udgravning indledt i 2002.
Hensigten med sidstnævnte var at redde alle truede
materielle levn, såvel som at beskrive levnenes
kontekster.
Howick bopladsen ligger på kanten af en nutidig
klint ud mod en mindre flodmunding. Lokaliteten
inkluderer en mesolitisk hyttetomt, såvel som en
stenkiste gravplads dateret til tidlig bronzealder.
Nærværende monografi fokuserer udelukkende på
beskrivelsen og fortolkningen af de mesolitiske levn,
hvorimod bronzealder-fundene vil blive præsenteret
i en selvstændig artikel (Waddington et al. in press).
De naturvidenskabelige undersøgelser af bopladsens
holocæne baggrund vil også blive publiceret
selvstændigt (Boomer et al. in press). Få mesolitiske
hyttetomter er kendt fra De britiske Øer, hvorimod
de er relativt almindelige i andre regioner, såsom
Skandinavien. Der er formodentlig flere grunde til at
hyttetomter fra britisk mesolitikum er så sjældne:
Først og fremmest har havstigninger haft forskellig
effekt i forskellige dele af Nordvesteuropa; få større
fladeudgravninger har været foretaget i de britiske
kystegne; i nyere tid har interessen for mesolitisk
forskning været faldende blandt britiske arkæologer,
sammenlignet med en mere stabil interesse for den
mesolitiske periode i for eksempel Skandinavien; og
antagelig som en konsekvens af sidstnævnte har der
været problemer forbundet med at erkende
karakteren af mesolitiske levn, når de blev frilagt i
forbindelse med udgravninger. Ved et sammentræf
blev en næsten identisk lokalitet udgravet 80 km mod
nord, ved East Barns nær Dunbar (Gooder in press),
kun måneder efter at Howick udgravningen var
fuldført. Også denne boplads lå ved kysten ud mod
Nordsøen. Sammen med de velkendte lokaliteter
Mount Sandel og Broomhill, præsenterer Howick og
East Barns de mest entydige beviser for opførelsen af
hytter i britisk og irsk mesolitikum.
På Howick lokaliteten var der fyldskifter fra tre
hytter, som havde stået mere eller mindre på samme
sted over en kontinuerlig periode. Dette blev bekræftet
af kulstof-14 dateringerne, der i forbindelse med
hytterne udgør en ubrudt sekvens. Det er uvist,
hvorvidt bopladsen blev benyttet sæsonvist, eller om
der er tale om helårsbosættelse. Dette vil antagelig
udgøre et af de mere væsentlige spørgsmål i forbin-
delse med den fortsatte diskussion af bopladsens
karakter. Kulstof-14 dateringerne antyder også en
fjerde kortvarig bosættelsesfase, som fandt sted efter
hyttens opgivelse. Sidstnævnte levn repræsenterer
formodentlig en transitboplads, hvor bopladsen blev
benyttet en enkelt nat eller to. Kulstofprøverne fra
Howick er blevet indgående analyseret, og resultaterne
er blevet statistisk behandlet med henblik på at
producere en logisk kronologisk sekvens. Med bag-
grund i disse analyser antages det, at bopladsen blev
besøgt første gang omkring 7850 cal BC, og at den blev
opgivet omkring 7650 cal BC.
Fundene fra hyttens indre svarer til det, man ville
forvente i forbindelse med en basisboplads. Centralt
i boligen var der en samling ildsteder, som indeholdt
brændte dyreknogler og forkullede hasselnødskaller,
såvel som et aktivitetsområde med stenredskaber og
en grube, hvor hasselnødder blev ristet. Ristnings-
gruben indikerer eksistensen af en strategi til kon-
servering af fødevarer. Selvom hytten givetvis
primært fungerede som bolig, udelukker dette ikke
at rituel adfærd også forekom. Det er således
usandsynligt, at stenalderens jæger-samler samfund
adskilte almindelige dagligdags opgaver og rituelt
liv, men i stedet inkorporerede ’ritualer’ i alle livets
aspekter. Det bør dog understreges, at der under
udgravningen ikke bemærkedes åbenbare fysiske
indikationer på rituel adfærd, og at ingen fund kunne
fortolkes som specielle nedlæggelser eller som
havende egentlig rituel betydning. Der var heller ikke
tegn på gravlæggelse eller anden behandling af
menneskelige levn. På disse punkter adskilte Howick
hyttetomten sig ikke fra hyttetomterne ved Mount
Sandel, Broomhill og East Barns, der generelt svarede
bemærkelsesværdigt til hinanden.
Det lithiske inventar fra Howick er det mest præcist
daterede fra britisk mesolitikum, og repræsenterer en
materiel kultur kendetegnet ved produktionen af
xx
mikroflækker og skævtrekanter. I Storbritannien er de
tidligste dateringer af bopladser med sådanne
inventarer generelt blevet sat til ca. 7500 cal BC, men
dateringerne fra Howick antyder en endnu tidligere
begyndelse for denne materielle kultur, antagelig
umiddelbart efter 8000 cal BC. Nylige dateringer fra
East Barns svarer ret præcist til de tidligste dateringer
fra Howick. Endnu tidligere dateringer er for nylig
blevet opnået for en skævtrekant boplads ved
Cramond nær Edinburgh, og sammen med en tidlig
datering fra Daer Reservoir, også i Skotland, indikeres
der nu en begyndelse for denne materielle kultur i
Storbritannien omkring 8400 cal BC. Dette rejser en
række spændende spørgsmål, såsom oprindelsen og
spredningen af denne form for teknologi, dens forhold
til tilsvarende Nordvesteuropæiske kulturer, og det
tilsyneladende overlap med det, der i det sydlige
England traditionelt er blevet henvist til som ‘tidligere
mesolitikum’.
Det lithiske inventar fra Howick er generelt baseret
på lokalt forekommende kugleflint, hvilket passer fint
ind i det almindeligt accepterede billede af, hvorledes
de mesolitiske grupper i det nordøstlige England
primært erhvervede sig lithiske råmaterialer i deres
nærmiljø. Inde i hytten blev der fundet et vidt
spektrum af redskabstyper, som belyser de
mangeartede aktiviteter, der fandt sted. Beviserne for
varieret adfærd suppleres af tilstedeværelsen af en
samling redskaber baseret på rullesten, store
mængder forkullede hasselnødskaller, fragmenter af
okker, såvel som knogler fra en lang række dyr.
Howick lokaliteten er strategisk velbeliggende
med adgang til et bemærkelsesværdigt rigt catchment
område. Da der vil have været adgang til et stærkt
varieret udbud af ressourcer året rundt, er det
DANSK RESUMÉ
LEFT-HAND HEADER
rimeligt at antage, at lokaliteten var en basisboplads,
som blev benyttet af den samme gruppe over en
række generationer. Formodentlig blev bopladsen
benyttet over en 200-årig periode. Hyttens størrelse
indikerer, at den blev brugt af en familiegruppe.
Fundspredning og kontekster antyder, at der på intet
tidspunkt var mere end en hytte på stedet af gangen,
omend erosionen af den nærliggende kystlinje
betyder, at dette spørgsmål ikke kan afklares entydigt.
Fund fra markvandringer har påvist andre meso-
litiske aktivitetsområder i området, som alle ligger
nær vandløb og dermed ferskvand. Det kan ikke
afvises, at disse lokaliteter kan være forbundet med
lignende hyttetomter. Disse fund antyder, at familie-
grupper boede i velbyggede, koniske hytter, adskilt
af flere hundrede meter, men alle med let adgang til
ferskvand, den nærliggende kyst og hinanden. Der
blev ikke fundet nogen mødding ved Howick, men
dette skyldes antagelig erosionen af den mesolitiske
kystlinje og effekten af postglaciale havstigninger,
som har fjernet eller oversvømmet alle eventuelle
møddingslag nær den daværende kyst.
Udgravningerne ved Howick har bevirket en
revurdering af omfanget og karakteren af den
mesolitiske bosættelse i det nordøstlige England,
såvel som forholdet mellem dette område og andre
regioner omkring Nordsøen. De har også demon-
streret mulighederne for at finde intakte lag fra denne
periode, og den rigdom af information som fund fra
sådanne bopladser har at tilbyde. Forhåbentlig vil
dette arbejde styrke forskningen i regionens meso-
litiske periode, såvel som den mesolitiske befolknings
brug af nærliggende indlandsområder, og kontakten
med grupper nord og syd for Howick, såvel som
omkring Norsøen.
1INTRODUCTION
Clive Waddington
Setting the Scene
Writing 80 years before the excavations at Howick,
Francis Buckley anticipated the occurrence of Meso-
lithic remains there after his discovery of narrow-
blade Mesolithic flints at Budle Crags near Bam-
burgh, Brada Crags near Spindlestone and Chester
Crags near Outchester (Buckley 1922).
“It is likely that the coastland crags south of Craster
have some relics of the Tardenois fishermen; and under
favourable circumstances such relics should in time be
found.” (Francis Buckley 1922, 323).
Buckley likened the flint industry from these sites to
the ‘Tardenois’ industry recognised at the time in
northern France and Belgium. Today these North-
umberland flints are no longer seen as being a direct
extension of the Tardenois industry, but rather fit into
the Later Mesolithic micro-triangle techno-complex
recognised to exist across the British Isles and related
to traditions that extended over much of North-West
Europe at this time (see Jacobi 1976). The drawings
published by Buckley include scalene triangle and
crescent microliths, together with scrapers typical of
this tradition, and these are directly analogous to
those found at Howick. His anticipation of the
Howick site, which lies 3km south of Craster,
demonstrates the regard with which the area was
considered, even then, for hosting Mesolithic re-
mains. This study presents the results of the in-
vestigation of an early 8th millennium cal BC
Mesolithic hut site that was found eroding out of a
cliff that overlooks the Howick Burn estuary in mid-
Northumberland (Fig. 1.1).
Although significant work had been undertaken
along the Northumberland coast by earlier re-
searchers (see Young 2000a for review) who collected
Mesolithic flints from erosion scars (e.g. Davies 1983,
Young 2000a), burnt field surfaces (e.g. Buckley 1922),
old sand dunes (e.g. Raistrick 1934) and ploughed
field surfaces (e.g. Weyman 1984), very little was
known concerning the dating of this material. Indeed
prior to the work at Howick there were no published
radiocarbon dated Mesolithic sites, no excavated
Mesolithic structures and no sequence for Mesolithic
flint assemblages in Northumberland. The oppor-
tunity provided by the Howick site for establishing
firm radiocarbon dates for the Mesolithic, and for
dating particular lithic types from undisturbed
contexts, meant the beginnings of a chronological
framework could be developed.
Against this backdrop of chronological uncertainty
there was very little known about Mesolithic habit-
ation sites. The only exceptions to this are the rock
shelter sites that have been investigated along the
Fell Sandstone escarpments at Goatscrag (Burgess
1972) (Fig. 1.2), Corby’s Crag (Beckensall 1976) and
most recently at Salter’s Nick (John Davies pers.
comm.). Goatscrag and Corby’s Crag produced only
Figure 1.1. View over the Howick excavations looking south with
the Howick Burn estuary in the background. The hut site is located
towards the top of the trench along its left edge.
2
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 1.2. View of the Goatscrag rock shelter site B.
small quantities of lithic material together with some
traces of structural features, including drip gullies,
post holes, pits and hearth pits for what are thought
to have been temporary structures (e.g. Burgess 1972).
However, the discovery of Bronze Age cremation
burials in inverted cinerary urns indicates that later
activity took place in these shelters and it is possible
that this later phase of activity accounts for the pits,
gullies, post holes and fire pits. Until such features
are directly dated, their Mesolithic attribution will
remain contentious. Interestingly though, four carv-
ings of quadrupeds, that most likely represent deer
(Fig. 1.3), have been discovered on one of the vertical
walls within rock shelter B at Goatscrag (van Hoek
and Smith 1988). Although these images cannot be
dated directly, the use of figurative art is a style
common to hunter-gatherer groups, whereas the
Neolithic–Early Bronze Age rock art of the region is
purely non-figurative (see Beckensall 2001; Wadding-
ton 1999a, 108; Waddington 1998). This possible
survival of Mesolithic art opens up an interesting
avenue for future research, particularly in the light of
the recent discoveries of early hunter-gatherer art in
caves at Creswell Crags (Ripoll et al. 2004) and
Cheddar Gorge (Mullan and Wilson 2004).
Although only limited work has taken place to
date, the potential for Mesolithic research in North-
umberland and the Borders, on the coast, along river
valleys and in the uplands, is considerable, part-
icularly given the fact that there are so many
opportunities for acquiring environmental data from
sediment sequences that extend from the Late Glacial
onwards. The study presented here will contribute
new detail to the study of early hunter-gatherer-fisher
groups in the region while providing a platform for a
renewed research impetus. However, beyond the
regional research context, there have been few well-
resourced excavations of Mesolithic sites in the British
Isles as a whole, with study of the period dominated
by a handful of key sites, notably Star Carr (Clark
1971), Thatcham (Wymer 1962), Mount Sandel
(Woodman 1985), Morton, (Coles 1971; 1983), Oron-
say (Mellars 1987), the Southern Hebrides Sites
(Mithen 2000), Kinloch (Wickham-Jones 1990) and the
Obanian sites (e.g. Bonsall 1996) amongst a few
others. In the light of the current research climate the
Howick project has been fortunate to enjoy the level
of support necessary to undertake detailed and full
recording of a relatively intact habitation site. Given
the wealth of information that it has been able to
produce it is hoped that other sites of this period will
benefit from well-resourced study in the future, in
order to help unlock some of the questions relating to
the early settlement of the British Isles.
1INTRODUCTION 3
Figure 1.3. The ‘deer’ carvings on the vertical wall inside Goatscrag rock shelter site B.
Scope of the Volume
This volume provides a full report of the fieldwork
programme directed at understanding the Mesolithic
remains at Howick together with analysis and
interpretation of the data in their wider North Sea
context. The Bronze Age cist cemetery discovered
during the excavation of the Mesolithic hut is not
included in this volume but is reported separately in
a journal article (Waddington et al. 2005). The volume
has been structured to broadly follow the sequence of
fieldwork undertaken at the site, followed by the
analytical chapters and then the wider discussion. As
most of the analyses proved highly revealing, and in
some cases incorporated innovative and ground-
breaking work (e.g. Chapter 6), it was considered
important to present this work as a series of chapters
in their own right rather than relegating them to
appendices. As a result no appendices are included
in this volume. The chipped stone recovered from the
excavation is discussed in Chapter 7, starting with an
assemblage description, followed by metrical, temp-
oral and spatial analysis and concluding with an
overview. Chapter 8 discusses the other lithic mat-
erial produced by the excavation, while Chapter 9
presents a study of residues and use-wear on a
selection of the chipped stone and other lithic
material. It was decided to break the lithic work up
into separate chapters for ease of navigation by the
reader. A review of bevelled pebble tools and
interpretation of the Howick group is provided in
Chapter 14. However, discussion and interpretation
of the chipped stone tool assemblage in relation to its
wider chronological and geographic context is
presented in Chapter 15.
A crucial point for readers to note is that all dates
expressed in the text of this volume are given in
calibrated years BC (i.e. calendrical dates). Although
this is conventional for later prehistory it has not, in
the past, been the custom for studies geared to the
early Holocene. Adopting the same chronology for
Mesolithic as for later sites will allow the real time
difference between them to be articulated; accord-
ingly all the 8th millennium cal BC sites in the British
Isles have been calibrated using the latest curve (see
Chapter 15). Another important point to make
regarding the expression of chronologies in this
volume is the use of age ranges given in italics,
particularly in Chapters 6 and 15. These italicised
date ranges refer to the mathematically modelled
probabilities based on the combined radiocarbon
determinations and do not indicate actual radio-
carbon dates, which are always quoted in standard
typeface. It is recommended that the reader consult
Chapter 6 and its accompanying figure captions for
further clarification.
4
Site Discovery
Flints were first recorded falling out of an erosion
scar at Howick by John Davies (1983), an amateur
archaeologist and member of the Northumberland
Archaeology Group, who noted the flints as being of
Mesolithic date. No further investigation of the site
took place and it was largely overlooked for the next
17 years. Early in 2000 Jim Hutchinson, also a member
of the Northumberland Archaeology Group, brought
a handful of flints he had found at the same erosion
scar and in nearby molehills to an artefacts day school
organised by the then ‘Centre for Lifelong Learning’
at the University of Newcastle. The flints were shown
to the author who was able to confirm that these
flints were of Later Mesolithic ‘narrow-blade’ type.
Subsequent to this Jim Hutchinson took the author
out to the site and showed him the erosion scar and
molehills. More flints were found eroding out of the
section but it was noted that most of the material was
coming from the soil layer below the plough zone,
which, significantly, suggested that in situ remains
survived on the site. This aroused a greater sense of
interest and anticipation as it was known that there
had been very few opportunities to excavate intact
Mesolithic deposits in North-East England before.
Photographs were taken and the farmer and land-
owner were informed. A return visit to the site was
made by the author, together with Nicky Milner and
Ben Johnson, to excavate a single 1m square test pit
immediately behind the erosion scar to test whether
subsurface archaeological deposits survived. The test
pit produced a total of 51 flints and at the top of the
sandy substratum an archaeological feature was
identified that contained quantities of flint and
charred hazelnut shell within a sandy weathered fill.
This generated immense excitement as the survival
of such deposits suggested that in situ Mesolithic
remains did in fact survive on the site. What was
more, these remains were being steadily eroded by a
combination of slippage from the erosion scar and
burrowing moles, and this added a sense of urgency
to dealing with the site.
On the basis of this initial exploratory work, the
author, Nicky Milner and Geoff Bailey, of the then
‘Department of Archaeology’ at the University of
Newcastle, decided to undertake an evaluation of the
site as a student training excavation. This evaluation
work included a close-spaced geophysical survey
around the immediate area of the site by Alan Biggins
of Timescape Surveys (see Chapter 2), sediment
coring around the Howick Burn by Ian Boomer from
the Department of Geography, University of New-
castle, as well as the excavation of an evaluation
trench behind the erosion scar. A sub-circular feature
measuring around 6m across was identified by the
geophysical survey behind the erosion scar and this
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
could be defined on the ground when the archaeo-
logical horizon was exposed by the excavation.
Although large quantities of Mesolithic flints were
obtained from the unstratified soil above the site, it
was considered important to obtain some dating
samples that could be used to demonstrate beyond
doubt that we were dealing with a Mesolithic
structure. Initial sampling of a post hole and burnt
feature took place, together with sampling some of
the upper fill within the sub-circular feature. Two
single entity charred hazelnut shell samples (AA-
41788 and Beta-153650) were submitted for AMS
dating, with assistance from the Society of Anti-
quaries of Newcastle upon Tyne and Northumber-
land County Council, and these returned early 8th
millennium cal BC dates (see Chapter 6). These dates
provided the proof necessary to convince potential
funders that this was a nationally important site that
deserved to be fully recorded before it was lost
through further erosion. A project design was
developed and grants were sought from the Uni-
versity of Newcastle, the Heritage Lottery Fund and
English Heritage, who all contributed to the full-scale
recording of the site in 2002, subsequent analyses,
publication and public dissemination through in-
formation panels, leaflets and a schools pack.
Aims
The project set out to achieve the following over-
arching aims:
1. To record in full the archaeological remains on the site
before they were destroyed through further erosion.
2. To identify more fully the extent of the archaeological
resource at the site and its condition of preservation.
3. To place the site in its wider context by investigating
land use across the surrounding spur of land and by
gaining a more informed understanding of the past
environment at the site and its linkages with the
archaeological record. This was considered particularly
relevant given that coastal erosion and sea-level change
have affected the North-East coast since the early
Holocene, which would have meant that the Howick
site occupied a different setting then than it does today.
4. To improve archaeological understanding of the Meso-
lithic in the region, particularly in a coastal setting, and
to embark on developing a chronology for the period in
the region.
Field Research Strategy
In order to achieve the aims set out above it was
decided to adopt a multi-disciplinary approach.
Firstly it was recognised that a purely site-based
1INTRODUCTION 5
Figure 1.4. Location map showing the Mesolithic site and study area.
approach would not be sufficient to gain an under-
standing of how this site functioned and how it
related to Mesolithic activity in the immediate
environs. Secondly, it was also acknowledged that
keyhole excavations and small-scale sampling were
inappropriate in these circumstances for trying to
understand the true extent and nature of this site. For
these reasons it was decided that both on-site and off-
site techniques should be employed to study the site
and that the excavation trench was to be of a
substantial size in order to be fully sure that all
remains in danger of erosion were investigated and
that any other immediate archaeology associated
with the Mesolithic site be recognised and recorded.
As past human behaviour during the hunter-
gatherer period was generally played out at the
landscape scale rather than focussed on intensive
occupation of a single site, it was considered im-
portant that the research strategy reflected this and
that the site was not investigated in isolation. By taking
an off-site approach to the research strategy (Foley
1981; Zvelebil et al. 1992) it was hoped that a more
6
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 1.5. Aerial view of the Howick Mesolithic site. © Tim Gates.
Figure 1.6. Aerial view of the Cushat Wood triple-ditched rectilinear enclosure (centre). © Tim Gates.
1INTRODUCTION 7
Figure 1.7. Aerial view of the Howick Burn Hillfort. © Tim Gates.
accurate understanding of Stone Age activity might
be achieved.
This approach was put into practice by utilising a
range of off-site prospection techniques prior to the
commencement of the main excavations, although in
actuality some of this work proved to be of only
limited value to accessing the off-site record (see
Chapters 2, 3, 4 and 12). The programme of work
combined assessment of existing aerial photograph
coverage, although this did not add anything new to
what was previously known, together with an
extension of the close-spaced geophysical survey over
an area of 4.8ha (see Chapter 2), and close-spaced
fieldwalking of seven fields extending over 51.5ha
around the site (see Chapter 3). Test pits were also
excavated beyond the limit of the large excavation
trench in order to assess whether other in situ deposits
survived near to the site, and particularly in areas
that may have been prone to future erosion (see
Chapter 4) along the cliff edge. An appraisal of the
soils and geomorphology of the site was also under-
taken, together with sediment coring to obtain a
datable sequence that could be analysed for the
survival of insects and pollen to help with recon-
structing the past environment at the site (see Chapter
12).
Site Location
The site is located on the Northumberland coast at
NU 2585 1657, 8km to the north-east of Alnwick
between Longhoughton and Craster (Fig. 1.4). The
coastline at Howick consists of exposed rock cliffs of
interbedded sandstone, limestone and mudstones,
with sandy beaches in the small bays at Sugar Sands
and Howick Haven. The cliff line is at the 10m
8
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
contour with the ground behind rising to 58m on the
crest of the Whinstone crag at Hips Heugh. The land
is gently undulating and is mantled by a veneer of till
deposits. This area of coastline is drained by the
Howick Burn and an unnamed tributary which
discharges into the sea at Iron Scars. The spur of land
on which the Howick site is located is bounded to the
south and west by the deeply incised valley of the
Howick Burn, while to the east it is bounded by the
North Sea. The only open approach to the site is
along the neck of land from the north and in this way
the site recalls the location of the prolific Mesolithic
flint scatter site at Crimdon Dene on the Durham
coast, which was also located on a spur of land
between the shore and a deeply incised valley
(Raistrick and Westoll 1933). The study area around
the Howick site was defined by the spur of land
extending southwards from Seahouses Farm (Fig.
1.4), as this is a geographically discrete landform that
appears to have been respected through time, as
indicated by the presence of a hillfort and a later
rectilinear enclosure on the spur, both of which
overlook the burn. Two fields outside the spur, to the
south of the Howick Burn, were included in the
fieldwalking as they were available to be walked at
the time, and three test pits were also excavated
behind an erosion scar to check if any archaeological
remains were in danger of erosion. Otherwise, all the
investigations were centred on the spur of land
defined by the burn and the North Sea.
The Mesolithic hut site overlooks the small estuary
of the Howick Burn, which falls into the North Sea
250m south of the site. The site commands extensive
views along the coast to the south-east and out to sea.
Although hindered by the present woodland cover
above and within the valley of the Howick Burn, the
site also has views over the ground to the south
towards Boulmer and there are good views north
along the spur towards the large outcrops of Whin Sill
at Howick Scar and Hips Heugh. Only to the west
does the site have restricted views, where rising
ground leads up to the eminence on which the later
hillfort is located. Today the site is in an exposed
location on a cliff edge (Fig. 1.5) where a public
footpath passes over it, however it is thought that
during the Mesolithic the cliff edge would have been
a hundred metres or so further away (see Chapter 12).
Romano-British (Young 2000a; Steve Speak pers.
comm.) and Early medieval (Keeney 1939) periods.
Previous fieldwork has been predominantly site-
based, including the excavation of a pagan Anglian
cemetery in the area of the Howick Whinstone Quarry
(Keeney 1939). As the quarrying progressed, a further
excavation took place on the quarry site of an Early
Bronze Age cremation cemetery (Jobey and Newman
1975). This excavation identified an enclosed crem-
ation cemetery with a primary double cremation
accompanied by a Collared Urn. The cremation
provided a date of 1840–1690 cal BC (3390 ±90BP, I-
6974). This quarry lies just west of this project’s study
area but other excavations have taken place closer to
the Mesolithic site. Two evaluation trenches were
excavated on the Cushat Wood rectilinear enclosure
(Fig. 1.6), having been positioned to sample two of
the three surrounding ditches. Up to five internal
circular huts were identified on aerial photographs
but none of these were sampled by excavation. A
single sherd of undecorated pottery similar to the
pottery recovered from the Fenton Hill curvilinear
enclosure in the Milfield Basin (Burgess 1984) was
found in the fill of the deeper ditch (Young 2000b;
Steve Speak pers. comm.), but no organic material
suitable for dating was recovered.
Prior to this more recent work, archaeological
remains from the Howick area had long been known
with the discovery of stone axe heads on the farm at
Boulmer (Bateson 1895, 334) and near Longhoughton
(Bosanquet 1934), an Early Bronze Age cist burial at
Longhoughton, and a fragmentary Food Vessel Urn
below a decorated cap stone to the south of the
Howick Burn estuary on Lowstead Farm (Fig. 1.4)
(Maclaughlan 1867; Bateson 1895, 333–4). The Howick
Burn hillfort (Fig. 1.7) was ploughed around 1817 and
pieces of ancient swords and some old coins (reported
to have been Roman) were found (Maclaughlan 1867).
The hillfort was surveyed as part of this project and a
limited geophysical survey conducted to produce a
condition report for the site. This has been completed
and submitted to English Heritage and the County
Sites and Monuments Record.
It is clear that the attractions of the Howick
landscape have been felt across all generations, as
archaeological remains from virtually every period
have been discovered. It should come as no surprise
that in landscape settings such as this, where fresh-
water, marine resources, terrestrial resources, fertile
ground and defendable locations occur together,
human groups will always be close at hand. It is no
exaggeration to state that the area of land around the
Howick Burn and adjacent coast has formed a
‘persistent place’ throughout history with each
successive period leaving traces of its presence on the
landscape. Although the Mesolithic project has been
a substantial undertaking, it is clear that this rich and
complex landscape still has many other stories to tell.
Previous Archaeological Work in the Howick
Area
The Howick area forms a multi-period archaeological
landscape containing remains from the Mesolithic
(Davies 1983), Neolithic (Bateson 1895, 364; Bos-
anquet 1934), Bronze Age (Bateson 1895, 364; Jobey
and Newman 1975), Iron Age (Maclaughlan 1867),
2GEOPHYSICAL SURVEY
Alan Biggins
The Howick geophysical survey consisted of an
exploratory magnetometer survey of 0.3ha, focussed
on the area around the initial test pit that located the
Mesolithic deposits. This was followed by a larger
extensive survey of 4.8ha covering the entire field in
which the Mesolithic site was situated. An additional
exploratory magnetometer survey was undertaken
over the nearby hillfort (Biggins et al. 2002), originally
surveyed by Henry MacLaughlan for the Duke of
Northumberland in 1860 (Fig. 2.1), but is not reported
here. The exploratory survey took place during June
2000 in conjunction with the first season of excavation
at the site. The second survey was undertaken in
advance of the second season of investigation during
May 2002.
The application of magnetometer survey tech-
niques for the identification of early prehistoric
features, especially those dating from the Mesolithic
period, is so far a relatively untested methodology,
certainly in Northern England. The key issue is not
the identification of substantial pit-like features,
which can often generate a positive magnetic an-
omaly, but the identification of such a pit for what it
is amongst features that may provide a similar
response. The attempt to identify a hunter-gatherer
site at Howick by high-resolution magnetometry and
then excavating those features is a demonstration of
the use of intensive geophysical survey for the
prospection and future investigation of hunter-
gatherer sites.
Objectives
The key objectives of the magnetometer survey can
be summarised as:
• Identification of anomalies suggesting the presence of
prehistoric features, together with information regard-
ing constructional configuration, extent and character
within the survey area.
• Identification of any previous agricultural activity,
including features such as structures, routeways, field
boundaries, pits and ditches that might be associated
with the Iron Age and Romano-British occupation of
this area.
• Identification of any potential archaeological features
close to the cliff edge that may be susceptible to future
erosion.
Location and Geology
The survey was centred upon NU 259 167 on land
currently given over to pasture. The survey area was
situated 500m south of Sea Houses Farm on land that
slopes gently towards the north. However, a slight
hollow exists towards the northern field boundary,
which acts as a natural drainage channel. Evidence of
this drainage effect can be seen particularly at the
cliff edge where waterlogged, possibly anaerobic
conditions, may exist. The soils are, overall, relatively
free draining and sandy in composition (see also
Chapter 12). The field boundaries have remained
essentially unchanged since the 1850s, although some
new fencing has been erected to supplement the
previous derelict drystone walling. A farm track
delineates the site towards the west and the eastern
coastal side is typified by recent (and continuing)
erosion of the coastline and defined by a steep cliff
edge.
The Quaternary deposits overlying the solid
geology often influence geophysical responsiveness,
and tend to exhibit a high degree of local variation. It
is recognised that drift geologies show a high degree
of variability and the magnetic response is usually
dependent on the magnetic mineralogy of the parent
solid geology (English Heritage 1995, 10). The area of
investigation is on geology of the Silesian (Millstone
Grit Series) sedimentary formation called Namurian,
comprising sandstones, mudstones, limestones and
coal. To the north, bosses of amygdaloidal Whin Sill
basalt are intruded amongst the bedding planes of
limestone (Johnson 1995). In general, magnetometry
survey can be recommended over any sedimentary
geology, although in some cases overlying drift can
introduce distorting factors. In this case the effects of
the igneous Whin Sill intrusion did not overwhelm
the survey, but some magnetic erratics may be
present. These erratics, only identified during excav-
10
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 2.1. Overview of the second phase of the magnetometry survey comprising some 4.8ha.The much smaller first phase was conducted
at the extreme southern sector of the area indicated. This historic landscape shows the defended settlement surveyed by Henry McLaughlan
in 1860 and the crop mark transcription of a late Iron Age enclosed settlement 150m towards the west.
ation, produced a response indistinguishable from
pits and could be easily mistaken for archaeological
features.
Methods Statement
The surveys were conducted during periods of
generally dry and warm weather conditions. A
Geoscan FM36 fluxgate gradiometer was used,
employing 1m parallel traverses with 0.25m sample
intervals over a pre-surveyed grid of 30m squares
(Fig. 2.2). The grid was set out parallel to the western
farm track and set back 2m from the field boundary
in order to minimise interference from the wire, but
the road itself was surveyed. A digital terrain model
(DTM) was derived from additional height measure-
ments (Fig. 2.3).
The magnetometer survey data were computed
and analysed using Geoplot 3 data processing
software (Geoscan Research). Terramodel and Terra-
2GEOPHYSICAL SURVEY 11

Figure 2.2. 30m survey grids and other relevant mapping features were recorded. From these relatively few points a contour map was
produced which showed the major terrain features.
vista mapping and digital terrain software (Spectra
Precision Software) was used to process the topo-
graphical and mapping data.
Results and Discussion
The survey results are illustrated here in both
greyscale and trace plot format (Fig. 2.4). The data
were interrogated and used to produce a colour-
coded anomaly plan, which was also related to
topographical features (Fig. 2.5). Not all possible
features, especially those derived from positive
anomalies (dark on the greyscale plots), have been
indicated in Figure 2.5 and so this interpretation is
subjective in nature and may indicate illusory
alignments or relationships. These anomalies are
enumerated and discussed below.
The pilot geophysical survey conducted prior to
the first trench also included most of that area
excavated in phase 2. One of the objectives of the
12
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 2.3. Digital terrain model viewed from the south-east derived from the 30m survey grids and other relevant mapping features. The
model clearly shows a distinctive re-entrant in the north, which is where the majority of suspected anaerobic anomalies was detected.

Figure 2.4. (A) Grey-scale plot in which dark features indicate positive anomalies (pits, ditches etc.) and white features negative (generally
stone, but may be part of a bipolar, or ferrous response). (B) Trace plot produced from the same data, but which has been clipped at ±25 hTeslas
to remove the more substantial and intrusive ‘iron spikes’.
2GEOPHYSICAL SURVEY 13

Figure 2.5. Interpretation plan highlighting anomalies within the survey area. Many possible features have not been indicated, such as the
possible drainage and anaerobic areas towards the north of the survey transect.
larger survey was to identify potential archaeological
features similar to those found during the phase 1
survey. In this report there will be a general dis-
cussion of those features detected over the entire
survey, followed by a review of those anomalies
detected and subsequently excavated.
The area in general is typified by a large number of
relatively small positive anomalies, not all of which
have been highlighted (Fig. 2.5). The western edge of
the survey area was delineated by a farm track that
showed as a prominent bipolar linear anomaly (1),
giving a response of ±25hTeslas. The track may have
been consolidated by hardcore rubble containing
bricks or other ceramic material. An additional factor
is the compression of the soil substrate, which creates
a more dense medium resulting in a stronger intensity
14
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
of response. The wire fence and a possible water pipe
would all have contributed to this effect.
At the southern terminus of the track a broader
band of anomalies was present. This general area at
the time of survey was used to store bales of hay and
may have been subjected to surface compression.
Another prominent non-archaeological feature was
the route of the footpath (2), which, where it was
surveyed, showed as an intense dark positive an-
omaly (approx +3 to +20hTeslas). The reasons for this
response are similar to those discussed earlier within
this paragraph.
A number of small (c. 1.5–2m diameter) positive
anomalies were detected which appear to form a
linear configuration some 10m in length (5). Because
of their location on a slope, caution must be exercised
in attributing archaeological significance to these
anomalies. A large bipolar anomaly (6), some 3–4m in
diameter, elicited a strong response (c. -10 to +50h
Teslas). Generally a response of this magnitude would
indicate a large ferrous object, although an area of
burning (and/or thermoremanence) may be respon-
sible. Whether this anomaly is a product of ferrous
agricultural debris or the site of a (primitive) kiln/
hearth can only be answered by excavation.
A faint negative linear anomaly (7), possibly
associated with fainter-still parallel positive an-
omalies (unmarked) may be associated with drainage
schemes. Directly to the south of this feature are a
number of sinuous linear positive anomalies (8). On
the eastern side these are some 100m in length and
appear to bifurcate centrally towards the east, with
an orientation south-west to north-east. At the north-
eastern terminus one branch appears to change
direction by almost 90º towards the south-east.
Collectively these linear features appear as a rather
faint attenuated positive anomaly, which in response
is broad and rather intermittent and insubstantial.
This may indicate the path of a former field boundary
and associated ditch. At the eastern edge of this linear
feature (8), towards the edge of the area of landslip
and erosion, is a large group of circular positive
anomalies (9). A number of bipolar anomalies are
also present within this group. Superficially these
anomalies have the characteristics of pits, but the
nature of the slope may create spurious results.
Nevertheless, it may be decided that some of these
anomalies could be worthy of future investigation.
Within the central sector of the survey a number of
circular positive anomalies (10) arranged in a linear
manner were detected. The response from these was
not intense, perhaps caused by archaeological feat-
ures (such as truncated post holes), or by the molehills
observed in this area. To the south of these is a large
linear bipolar anomaly (11) creating a response some
16m in length and 5m in width, with values ranging
between approx. -15 to +50hTeslas. Superficially this
has the appearance of a substantial ferrous pipe,
although there is no surface indication of this feature,
or any suggestion that it is connected to another
modern agricultural device. Large pipes especially
would in general have a response above the threshold
limit of the instrument (±204.7hTeslas) at the sensitiv-
ity setting in use. Whilst apparently unrelated to
modern land usage at the site, this anomaly may
represent a burial trench for ferrous material or a
linear burning trench. It should be noted that the
detectable magnetic response from rust is negligible
compared to the response from less corroded ferrous
debris.
A number of small circular positive anomalies (12)
apparently some 1–2m in diameter, organised in a
linear or curvilinear aspect, were detected. A similar
group of anomalies was detected some 60m further
south (13). The strength of the response was low, in
the region of +2 to +5hTeslas, within the range typical
of many archaeological features. A number of out-
lying small bipolar and positive anomalies are located
nearby. Towards the south-west of these, a broad-
band negative anomaly (14) was detected which
could be geological in origin. At its southern ex-
tremity a linear arrangement of five closely linked
circular positive anomalies were detected, each
approximately 1.5–2.0m in diameter (15). Towards
the south-west terminus, two outlying circular
positive anomalies were detected with similar
dimensions. None of the responses exceeds +5h
Teslas. These anomalies were detected two years
earlier in the pilot survey. The site of the previous
excavation trench (16) was detected in outline as an
extremely magnetically active disturbed area. It had
been backfilled with a plastic lining and contained
ferrous material deliberately left in order to recognise
the extent of the trench when it was reopened for
further excavation.
The initial investigative survey (Fig. 2.6A), which
was conducted in advance of the phase 1 excavation,
located a series of contiguous positive anomalies,
which were identified as possible pits at the time.
Subsequent excavation (see Chapter 5) revealed a
linear burning pit (Fig. 2.5 anomaly 15), some 10m in
length, including two additional ‘pits’, which were
detected at the south-western terminus. The com-
plexity of this composite anomaly only became
apparent after excavation. In brief, it comprised
multiple burning areas containing burnt stones, flint,
charcoal and occasional tiny burnt bone fragments.
The concentration of burnt material was clearly
responsible for generating the magnetic anomalies.
The net positive response (c. +5hTeslas) would have
largely overwhelmed the negative response (c. -
1hTeslas) expected from the stone cist (cist 5), which
was found disturbed below the northern tip of the
burning pit.
The spread of stones and flint located some 2.3m
south of the burning pit was indicated by an area of
2GEOPHYSICAL SURVEY 15

Figure 2.6. (A) Phase 1 geophysical survey grey-scale plot. (B) Excavated site plan. The location of the burning pit and the Mesolithic hut
are clearly indicated as groups of strong positive anomalies.
reduced magnetic response, but realistically this
could not have been differentiated from other similar
features anywhere on the rest of the site. Cist 4 was
detected as a central positive area with a maximum
response of +1.3hTeslas, with a slight negative (-
0.8hTeslas) periphery. In this instance the internal
positive responsiveness would have been created by
higher levels of organic matter and slow infill of the
cist interior. The negative element would have been
caused by the materials of the cist displacing the
more magnetic medium of the surrounding soil. Cists
1, 2 and 3 produced negligible magnetic responses,
certainly not enough to be recognised for what they
are.
The Mesolithic hut was highlighted as a group of
relatively strong positive anomalies (+5–6hTeslas) set
in a sub-circular arrangement. The magnetic response
was no doubt caused by the presence of magnetically
enhanced burnt debris, associated with the Mesolithic
occupation.
Although the hut was truncated towards the
south-east by the cliff, the dimensions of the magnetic
anomaly associated with it, in the region of 6–7m,
accord well with the excavated size of c.6m (see
Chapter 5). The smaller and less substantial structural
features less than 0.2m in diameter were unlikely to
be detected with a survey resolution of 1.0m x 0.25m
spacings.
Most of the other single ‘pit’ anomalies are almost
certainly caused by magnetic erratics, as a number of
them were detected within the trench set in the
boulder clay but were not specifically plotted. The
effects of igneous intrusions are well documented
and have been observed close to the Whin Sill. Less
easy to distinguish are the effects of relatively small
erratics, such as those of andesite or basalt (for
instance, the response from a football-sized boulder
of andesite is very similar to that measured from a
large pit 1–2m in diameter). For the majority of the
uniformly spread ‘pit’ anomalies at Howick, a similar
effect is probably (but not necessarily) the cause.
Having said that, the sheep burial detected within
one of the test pits, although modern, showed all the
attributes of an archaeological feature, which it would
indeed become in the fullness of time.
In general, larger features, such as hearths, part-
icularly ones that have been reused and are relatively
undisturbed, produce a measurable response. The
combined effects of burning and subsequent infill
between firing episodes are likely to increase local
magnetic susceptibility. The associated pits close to a
hearth will increase the overall effect, whereas larger
pits (c. 0.5–1.0m) of reasonable depth (c. 0.5m) will
probably have a distinctive signature of their own. In
summary, the best indicators for a Mesolithic hut are
likely to be the associated infrastructure of pits and
hearths with their conjoined, accumulated organic
debris, and to a lesser extent the infilled circular or
sub-circular hollows in which they are situated. Once
these features have been recognised, characterised
and tested by excavation at a number of sites, a more
accurate geophysical signature for these types of sites
may be forthcoming.
In summary, the magnetometer survey was suc-
16
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
cessful in picking up the Mesolithic hut site as an
archaeological feature, though it is difficult to ascribe
distinctive morphological characteristics to such
signatures. No strong evidence for previous farming
activity (e.g. medieval ridge and furrow) was de-
tected on the site. Large numbers of bipolar and
positive anomalies were detected and spread almost
uniformly across the entire survey area and some
were subsequently shown by excavation to be glacial
erratics. These, because of their abundance, have not
been specifically identified or enumerated, nor has
the ferrous and ceramic debris that will have caused
some of the responses. Some of course may be
genuine pits. The initial survey revealed the Meso-
lithic hut as a sub-circular anomaly, together with the
location of the Bronze Age cist burials and a linear
burning pit (see Waddington et al. 2006 for a report
on the Bronze Age remains).
Conclusions
The effectiveness of geophysical survey is well known
and magnetic survey in particular has been seen to
work effectively on many sites in Northumberland
(e.g. Biggins et al. 1997; Biggins and Taylor 1994). The
method depends upon recognition of the morpho-
logical exemplars such as pit or post hole alignments,
ditch configurations, hut circles and other relatively
unequivocal features which have been proven by
excavation. Even apparently obvious features such as
a cist burial, some of which are quite small, may not
be recognised unless they are found in the presence
of larger, more diagnostic features. The problem
invariably lies in the magnitude of background
‘noise’, with the effects of later agricultural practices
often masking the relatively weak signals emanating
from small and discrete sub-surface archaeological
sites. Modern intensive deep ploughing has often
disturbed previous horizons by cutting into them and
truncating significant features.
What is less secure upon examination is the study
of Mesolithic sites, which are often only recognised
by artefact concentrations. The response expected
from small and ephemeral episodes of occupation,
usually only recognised by surface lithic scatters, is
not generally considered to provide much scope for
geophysical detection. It is thus perhaps inevitable
that geophysical techniques have rarely been used as
a prospection tool for hunter-gatherer sites in the
British Isles. In view of the results from this site, more
consideration might in future be given to the possible
detection of small clusters of hearths, pits and also
sub-circular hollows of 3–6m diameter (positive
anomalies), which could be indicative of hunter-
gatherer occupation.
To promote confidence in remote sensing in
Mesolithic research, continued feedback from the
results of excavation is crucial, so that the types of
geophysical responses associated with such remains
can be more accurately understood. Further evidence
of the success of geophysical survey in detecting
Mesolithic deposits is provided by the recent survey
close to the cement works at East Barns, Dunbar in
2002, when a remarkably similar Mesolithic hut site
was detected using magnetic survey and sub-
sequently confirmed by excavation (Gooder 2007).
Convincing results from these two Mesolithic hut
sites are to be expected, given the size and sunken-
floored form of these sites, although the recognition
of other types of Mesolithic features remains a future
challenge.
3FIELDWALKING
Clive Waddington
Introduction
As part of the Howick project, a fieldwalking
component was included as one of the strategies for
obtaining evidence of wider Mesolithic land use
around this part of the coast. A total of 244 lithics was
recovered from seven fields covering a combined area
of some 51.5ha. (Figs 1.1 and 3.1). The fields are
located adjacent to the site and the field numbers are
shown on Figure 1.4. The field in which the site is
situated was unable to be walked as it is given over
to pasture, though it has been ploughed in the past.
Method Statement
Each field was line-walked at 5m intervals during
spring 2002 (Fig. 3.3). As the project is focussed on
the prehistoric archaeology of the area, the field-
walking was directed towards the recovery of lithics.
Other finds such as clay pipe and post-medieval
pottery were discarded. The method employed
followed that developed for the Milfield survey in
north Northumberland (Waddington 1999a, 35–44;
Waddington and Passmore in press). Assuming that
each person scans an area 1m either side of their
path, this means that two out of every five metres
were inspected for surface artefacts. This translates
to a 40% coverage rate that can be adjusted by
multiplying by 2.5 to give a notional 100% count (see
Tolan-Smith 1997a). Dividing this notional 100%
figure by the area covered allows a count per hectare
statistic to be arrived at for each field and these
figures can then be compared to other fieldwalking
surveys that have been carried out in the region. The
location of all findspots was recorded using a total
station to ensure accurate point-referenced data, and
these can be tied to the Ordnance Survey grid by
relating the findspots to the field boundaries (Figs.
3.1 and 3.2). All lithics were bagged, labelled, washed
and then analysed with attribute data for each lithic
input into an ‘Excel’ spreadsheet. Each field was
mapped according to slope type at a scale of 1:10,000
to produce a series of morphometric (slope) maps
that could be related to the distribution of surface
lithics.
Taphonomy
In order to make meaningful interpretations from
lithic scatters it is essential to identify, and account
for, the major recovery biases that have affected the
location and visibility of the assemblage. Recently
archaeologists have paid attention to understanding
the processes affecting artefact scatters in the plough-
soil (e.g. Allen 1991; Boismier 1997; Waddington
1999a), as it has become clear that understanding and
interpretation drawn from artefact scatter data are
likely to be erroneous unless the taphonomy of the
landscape has been considered. At Howick, the key
geomorphological process that has affected artefact
scatter recovery has been the loss of land to the sea
due to an eroding coastline. Consequently the lithics
picked up from the coastal edge during the field-
walking were not at the cliff edge at their time of
discard. This also means that we have lost the
material evidence resulting from the activities that
took place at or on the shore, as well as the cliff edge
immediately above. Therefore, the activities rep-
resented by the lithic scatters recorded as part of this
survey should not be assumed to represent the full
range of activities undertaken by these coastal
groups. The lack of shell midden sites on the east
coast of Britain must be directly related to the fact
that it is, in general, an eroding coastline and one
which has experienced submergence of the land
relative to sea level, and this is likely to have resulted
in the obliteration and/or drowning of most shore-
edge sites, and certainly those sites that predate the
Main Postglacial Transgression c.6000 cal BC (Smith
1992, 58). Suggestions that middens may have been
widespread are provided by surviving sites that have
been investigated at Morton, Fife (Coles 1971; 1983),
and the midden-type deposit that was found below a
Bronze Age cairn eroding from sand dunes at Low
Hauxley, Northumberland (Bonsall 1984), which lies
18
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 3.1. Map showing location of fields walked.
only 15km south of Howick. It is noteworthy that in
both cases we are dealing with Mesolithic sites later
in date than Howick; they post-date the maximum
marine transgression. Being situated above the level
of the transgression, these sites have survived intact.
Therefore, sites that pre-date the transgression, such
as Howick, are unlikely to have contemporary
shoreline deposits such as middens surviving.
The other key geomorphological influences af-
fecting the fields along the Howick coastal strip are
slope processes. As slope type has a direct effect on
the taphonomy of surface artefact scatters, and
therefore any subsequent interpretations made from
them (Allen 1991; Waddington 1999a, 85–91), biasing
effects resulting from slope processes must be taken
into account when assessing the distribution of
material. The slope classes devised for the ‘lithic
scatter displacement model’ (Waddington 1999a, 91–
4) formed the basis for the mapping of the Howick
fields (see also Waddington and Passmore in press
for an up to date discussion). At Howick the terrain is
neither flat nor uniform, but of undulating character
3FIELDWALKING 19

Figure 3.2. Plots showing individual lithic findspots within each field.
20
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 3.3. Fieldwalking at Howick.
with areas of short steep slopes, hollows and a deeply
incised, steep-sided valley, now occupied by the
Howick Burn. Many of the fields contain flat areas
together with areas of slope. As these fields are
ploughed and the soil becomes destabilised, this
facilitates the movement of soil and artefacts down-
slope, particularly during wet conditions when
processes such as rilling, hillwash and slumping assist
the transportation of artefacts. A model of lithic
scatter displacement in slope environments has been
devised that identifies the degree of movement of
material that can be expected on different slope types,
and the type of inferences that can be drawn from
artefact scatters affected in this way (Waddington
1999a, 91–4; Waddington and Passmore in press). This
requires the distribution of artefacts in each field to
be described and any patterns likely to be the result
of slope processes identified.
Field 1 slopes down towards its north-east corner,
though the upper areas of the field are flat (see Fig.
3.2). Half of the finds from this field are from the
sloping area and most of these are on the area of
medium slope leading down from a flat platform in
the north-west corner of the field that is occupied by
the Cushat Wood rectilinear enclosure. The enclosure
also extends into Field 2 on the same flat area. As
most surface lithics in medium slope environments
are ‘in transit’ it can be inferred that the string of
flints found running down the slope in this part of
the field has originated from the edge of the flat area
occupied by the enclosure, from where they have
been eroded as a result of plough action and sub-
sequent slope movement. Elsewhere in the field most
lithics are in flat areas; these will have experienced
little in the way of lateral displacement and can
therefore be assumed to be broadly indicative of past
human activity in these locations.
Field 2 contains the greatest concentration of finds
with most located on the area of flat ground occupied
by the Cushat Wood enclosure (Fig. 3.2), but also on
the same medium slope that slopes down from the
enclosure. It is important to note that the act of ditch
digging for the construction of the enclosure has
probably resulted in the disturbance of earlier
artefacts and features, and it is these artefacts that
later ploughing has dispersed across the slopes. This
disturbance may also account for field 2 having by
far the highest density of finds. Bearing in mind that
virtually the entire depth of the Mesolithic hut
deposits survived intact below the plough zone (see
Chapter 5), it means that other features belonging to
this period should also survive below the plough
zone across this area. If the widespread survival of
Mesolithic features can be identified, it raises im-
portant questions regarding future research, manage-
ment and conservation priorities in the area, and in
particular the need to avoid deep ploughing.
Although Field 3 slopes away steeply on its south-
eastern flank, most of the lithics recovered from this
field were located on an area of gentle slope above,
where only limited downslope movement can be
expected. Field 4 produced a high density of lithics
and virtually all of these were located on the gentle
slope that falls away from the western side of the
field. Here again some limited downslope displace-
ment can be expected, but this will not have moved
the artefacts very far. Therefore, in the case of Field 4,
most of the lithics will be broadly representative of
their position of discard. Field 5 is relatively flat
except for some areas of gentle slope and a hollow.
Most of the artefacts were recovered from the flat
area (Fig. 3.2), although one pattern worthy of note is
the tendency, observed elsewhere in the Milfield basin
(Waddington 1999a, 85), for the incidence of artefact
recovery to increase close to breaks in slope. In this
case a string of lithics can be identified running along
the break in slope down to the cliff on the east edge of
the field. In Field 6 all but two of the lithics were from
the area of flat ground on the south side of the field.
To the north, the field slopes steeply down to the
valley containing the Howick Burn. The field bound-
aries along this edge of the field sit astride a positive
lynchet – lynchets are known to obscure visibility of
artefacts. It is likely that in this case any flints
discarded on the slopes above the burn have been
moved downslope and become trapped and buried
in the lynchet. An Early Bronze Age barrow was
located on the knoll in the north-east area of the field
investigated by antiquaries during the 19th century.
A small Food Vessel Urn was found covered by an
ornamented stone (Bateson 1895, 330), the latter
probably being a reference to a cup-and-ring marked
capstone. Subsequently the mound has been plough-
ed virtually flat with only a slight surface expression
left. It could be expected that lithics would be
discovered in proximity to the site of this monument
but none were found. This may again be because the
barrow was located at the head of the steep slope
down to the burn, and any artefacts disturbed by the
ploughing are likely to have moved downslope. Field
3FIELDWALKING 21
Field Area (ha.) Actual No. Lithics Adjusted No. Lithics (x 2.5) No. Lithics per ha.
1 7.02 31 77.5 11.0
2 7.07 77 192.5 27.2
3 5.37 32 80.0 14.9
4 6.75 52 130.0 19.3
5 9.14 25 62.5 6.8
6 7.99 14 35.0 4.4
7 7.81 13 32.5 4.2
Total 51.15 244 610 av.11.9
Table 3.1. Lithic counts per hectare from the walked fields around Howick.
Project/Location Average (100%) density per ha. Reference
Coastal
Maiden’s Hall (Northumberland) 51.8 Waddington 2001a
East Durham and Cleveland Coast 13.0 Haselgrove and Healey 1992, 6
Howick (Northumberland) 11.9
Middle Warren (Hartlepool) 11.8 Archaeological Practice 1996, 5
Turning the Tide (Durham Coast) 10.9 ASUD 1998
Inland
Lower Tyne Valley (Northumberland) 10.0 (calculated from) Tolan-Smith 1997a, 82
Milfield Basin (Northumberland) 5.5 Waddington 2001b
Middle Tees Valley (Durham) 3.1 Haselgrove and Healey 1992, 14
East Durham Plateau (Durham) 0.6 Haselgrove and Healey 1992, 4
Tees Lowlands (Durham) 0.3 Haselgrove and Healey 1992, 13
Wear Lowlands (Durham) 0.3 Haselgrove and Healey 1992, 3
Table 3.2. Average lithic counts per hectare from other North-East fieldwalking surveys.
7 is mostly an area of flat ground where little
movement of artefacts in the ploughsoil can be
expected. Only one lithic was recovered from the
north end of the field where there is a steep slope
down to the burn.
Visibility was not consistent across the fields, as
they were walked on different days under different
weather conditions, and with differential cover of
sprouting crop, which meant that there were vari-
ations in the percentage of ground exposed. Fields 1,
2, 3 and 5 were walked in fairly good conditions with
sprouting crop, damp ground and not too intense
sunlight. The level of sprouting crop cover was
estimated to have obscured the ground surface by up
to 25%, implying that the lithic recovery for each field
could have been up to 25% higher. Field 4 was walked
after it had been ploughed, allowing for full observ-
ation of the field surface. However, when it was
walked the sun was extremely strong and this had
resulted in the baking of the ploughed surface. This
renders the surface more dusty and lighter in colour,
making observation of flint more difficult. In add-
ition, the glare that reflects from the ground surface
also hinders visibility. Nonetheless, the field did
produce a high lithic count, indicating that this field
is probably extremely rich in lithics and should be
classed on a par with Field 2. Fields 6 and 7 were
walked in overcast conditions conducive to the
spotting of artefacts, but in the case of both these
fields the sprouting crop cover had increased dram-
atically due to the mild winter and warm spring so
that up to 75% of the ground surface was obscured in
places. Therefore, it is considered reasonable to
assume that a more accurate indication of the lithic
density in these latter fields would be achieved if the
recorded total for each field was multiplied by four.
This produces lithic densities comparable to Fields 3
and 4 and it is likely that this is a more accurate
indication of the level of past human activity around
Fields 6 and 7 than the actual recorded counts.
Density
Table 3.1 shows that the highest density of lithics is
found in Field 2, which has a notional 100% lithic
count per hectare of 27.2. Field 4 also has a high count
of 19.3 per hectare and it should be considered that
this is probably under-representative (see above). The
counts for Fields 1 and 3 are also high and indicate
significant densities of chipped stone in these areas.
The low densities recorded for Fields 6 and 7 are
almost certainly related to the poor recovery cond-
itions. Field 5, however, has a low count and this field
did not have particularly poor visibility so this is
probably a realistic indication of a reduced level of
activity in this area.
In comparison to other fieldwalking surveys in the
North-East region (Table 3.2), the results from Howick
22
No. Lithics
250
200
150
100
50
0
Flint Chert Quartz Agate
Rock Type
Figure 3.4. Lithic Assemblage by Raw Material.
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Breakdown of Assemblage by Raw Material
resulting in higher than average lithic counts. This
was seen to be the case at Maiden’s Hall where 36% of
the material belonged to the primary stage of the
reduction sequence (Waddington 2001a) and similarly
at Howick, 32.8% of the assemblage belonged to this
stage. In contrast, however, the primary-stage material
in the Milfield survey accounted for only 23.5% of the
assemblage (Waddington 2001b) which is more typical
for inland areas. Otherwise, these three surveys for
which such information is available revealed remark-
able consistency in the percentage of tools, or tool
fragments, that represent the tertiary stage of the
reduction sequence or chaîne opératoire. In the case of
the Howick survey 10.6% of the assemblage belonged
to this stage, while the Milfield survey produced
10.4% and the Maiden’s Hall survey 10%. This
percentage is significant as it reflects perhaps a similar
level of activity associated with settlement for all three
areas because, as Schofield has noted (1991, 119), high
proportions of tools are usually associated with settle-
ment areas.
fit comfortably with the lithic counts recorded for
other coastal areas. For example the Middle Warren
fieldwalking project outside Hartlepool recorded an
average lithic density of 11.8 per hectare compared
with the 11.9 per hectare recorded from Howick.
Similarly the East Durham and Cleveland Coast
survey recorded a figure of 13.0 per hectare. The
incredibly flint-rich Maiden’s Hall survey near
Chevington, Northumberland, recorded an average
lithic count of 51.8 per hectare, but this was found to
be primarily the result of the area itself being used as
a source of glacial flint, containing large quantities of
primary and secondary flaking debris, and because it
was walked in ideal circumstances by an experienced
fieldwalking team. It is clear from the Howick survey
that if more suitable ground conditions had been
encountered, and a more experienced team had been
employed, then this survey would have yielded a
significantly higher count per hectare. Therefore,
when comparing with these other surveys it should
be considered that the Howick flint count is very much
a minimal figure. It is also interesting to note that
lithic counts for inland areas are, in general, sig-
nificantly lower than those recorded for the coastal
areas, even in areas known to have been heavily
exploited during the Stone Age, such as the Milfield
Basin for example. Although a suite of taphonomic
processes has been identified that has served to under-
represent lithic counts in the Milfield area (Wad-
dington 1999a, 55–7), it can still be seen that the North-
East coastlands were particularly favoured by Stone
Age hunter-gatherer groups.
However, it must also be noted that most of the
flint sources are located on and around the coastal
strip as secondary deposits in the tills, and as beach
pebbles washed on to the shore. As a result, the
availability of flint would have attracted Stone Age
groups to the coast, and primary testing, flaking and
dressing would have undoubtedly produced a greater
quantity of waste material near these ‘quarry’ areas,
Distribution
Consideration of the densities of lithics from the
different fields shows that those fields with the
highest densities are those that immediately fringe
the valley of the Howick Burn, such as Fields 1, 2 and
4 (see Table 3.1). In contrast, Field 5 lies furthest from
the burn and has one of the lowest densities. When it
is considered that recovery conditions on this field
were fine and that those for Fields 6 and 7 significant-
ly under-represented their actual lithic population, it
appears that Field 5 probably has by far the lowest
lithic count in actuality. Therefore, on the basis of the
data recovered so far it can be concluded that the
lithic concentrations tend to cluster close to the
Howick Burn and that the lithic counts appear to fall
off with distance from it. This observation gains
further support when the position of the excavation
site is taken into account, as this is situated close to
the mouth of the burn.
Observing the distribution of lithic types across
the various fields (Table 3.3) reveals some areas of
interest. In Field 2 the cores fall into two distinct
groups: one on the flat area partly occupied by the
Cushat Wood enclosure and another larger group in
the south-east corner of the field, on the area of flat
and medium slope running down from the enclosure
platform. In addition, three of the four scrapers
recovered from this field are situated amongst the
group of cores in the south-east corner and the fourth
scraper is situated with the cluster of cores on the
west side of the field. Two possible awls are also
associated with the concentration of material in the
3FIELDWALKING 23
Field No. NGR Parish Size
(ha.)
Nodules and
Bashed Lumps
Flakes and
Cores Retouched
Scrapers Microlith Other Total
Blades
Flakes and
Blades
1 NU257168 Longhoughton 7.02 24 3 2 2 31
2 NU258172 Longhoughton 7.07 2 50 14 5 4 1 1 awl 77
3 NU260171 Longhoughton 5.37 5 20 4 2 1 32
4 NU255166 Longhoughton 6.75 5 34 10 1 1 1 52
5 NU260175 Longhoughton 9.14 5 14 4 2 25
6 NU257160 Longhoughton 7.99 2 9 1 2 14
7 NU256157 Longhoughton 7.81 11 1 1 13
Totals 19 162 37 13 10 2 1 244
Table 3.3. Summary of lithic finds by field.
south-east corner. Bearing in mind that much of the
material on the medium slope in this corner of the
field will have been transported downslope from the
flat area above, it is the ground in the southern half of
the field that can be identified as having formed a
focus for Stone Age activity and may well conceal
remains of sites similar to that excavated as part of
this project. This pattern is confirmed by the distrib-
ution in Field 1 where two of the three cores and one
of the two scrapers are also located on the enclosure
platform and area of slope that continues into this
field from Field 2. Elsewhere, in Field 3 a concen-
tration of material on a gently sloping bluff in the
southern half of the field that projects out towards
the coast suggests that this locale formed a focus for
activity. This gains further support when the position
of two cores and two possible awls on the steep slopes
running down from this bluff are considered, and
which should be interpreted as representing activity
that had actually taken place further upslope. A
scraper was also recovered from on top of this bluff.
In Field 5, half the pieces were found on or close to
the break in slope that leads to the present cliff edge.
All four cores and one of the two scrapers were found
in this area, suggesting that the small amount of
activity represented in this field tends to be clustered
on the flat ground.
Assemblage Chronology
As is typical for fieldwalking assemblages in North-
East England, the majority of the diagnostic lithics
identified in this assemblage belong to the Mesolithic.
A total of 34 lithics could be characterised as Meso-
lithic, representing 13.9% of the total assemblage,
while a further four pieces could be either Mesolithic
or Early Neolithic in date, and the thumbnail scraper
142 (Fig. 3.6) could belong to either the Mesolithic or
Early Bronze Age. Of the Mesolithic pieces none could
be definitely identified as Earlier Mesolithic on
typological grounds, although the reworking of
already chipped and heavily patinated pieces does
provide indirect evidence for Earlier Mesolithic and
perhaps Palaeolithic activity (see below, Chapters 7
and 14). Most of the diagnostic pieces were the
product of a narrow-blade manufacturing tradition
indicative of Later Mesolithic flintworking. The other
pieces that showed diagnostic traits included four
retouched blade tools, including end scrapers, that
could belong to either the Mesolithic or Early Neo-
lithic, two very heavily patinated pieces that could
conceivably be Palaeolithic or Earlier Mesolithic, and
the thumbnail scraper that could be Mesolithic or
Early Bronze Age. However, all the other material
showed direct similarities with the lithic assemblage
recovered from the Mesolithic hut site (Chapter 7)
including varying degrees of patina development,
reliance on locally available material and flaking
characteristics indicative of small narrow-blade
production. There were no indications of later material
in the assemblage other than the possible scrapers
referred to above. Therefore, this assemblage is treated
as being of predominantly Later Mesolithic date. The
formal implement types are also directly comparable
to those recovered from the Mesolithic hut site
(Chapter 7) and so the fieldwalking assemblage is
considered to have accumulated, for the most part at
least, during the Later Mesolithic, c.8000–4000 cal BC.
The assemblage included some heavily patinated
and rolled pieces that are probably significantly older
than the more common Later Mesolithic material, as
well as a few end scraper and blade forms that could
be either Mesolithic or Early Neolithic in date. Nearly
half of the pieces are broken (117) and this includes
pieces broken in antiquity, perhaps at their time of
use, as well as more recent plough-damaged pieces.
There are also some burnt pieces (19), some of which
were finished tools, such as the fragmentary end
scraper 302 from Field 5 that may have been brought
to the surface from disturbed burnt/hearth deposits.
Beach-rolled, previously chipped flints were collected
from the shore and brought on to the higher ground
for reworking, and flints 149, 103, 144 from Field 1
are typical examples. Another feature common to this
assemblage, and evidenced widely in flint collections
from the North-East coast (e.g. Waddington 2001a),
is the reworking of previously chipped pieces that
have then been rolled in the sea and have usually
24
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 3.5. Selected cores and flakes recovered by fieldwalking.
developed a heavy patina. For example flints 178,
208, 225, 244 and 399 from Fields 1–4 all show clear
evidence of having been re-chipped in this way. These
flints provide a proxy indicator of Earlier Mesolithic,
and more likely Palaeolithic, occupation in this now
submerged area. Still more flints show evidence of
recycling as patinas have developed over the original
flaking scars; these pieces were then re-chipped
during the Mesolithic, such as the core 408 and awl
251. However, in these cases there is no evidence of
them having been beach-rolled before recycling,
which indicates that these pieces were collected from
the terrestrial environment. These flints therefore
indicate the presence of earlier activity on the land
3FIELDWALKING 25

Figure 3.6. Selected implements recovered by fieldwalking.
26
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
that now forms the coastal strip. Therefore, by
observing the ‘stratigraphy’ apparent on some of the
pieces in this lithic assemblage, it is possible to posit
evidence for pre-Later Mesolithic activity along this
stretch of coastline in what are now both the terres-
trial and marine zones (see Chapters 7 and 16 for
further discussion in relation to the excavated
material).
that is common in other coastal assemblages from
Northumberland. Similarly there were 12 examples
(4.9%) of the orange beach pebble flint, together with
10 pieces (4.1%) of orange-grey and 7 pieces (2.9%) of
orange-brown flint, which are also frequently seen in
other North-East assemblages. More than half of the
chert (8 of the 14) is of the blue-grey variety common
in assemblages across Northumberland. The other
types of chert belonged to a mixture of oranges and
greys with one piece of red-brown chert.
The provenancing of flint by visual inspection only
is a subjective art, and as flint nodules can have great
colour variation within a single nodule as well as
throughout flint-bearing strata, colour is not always a
reliable indicator of provenance. However, in this
case the colour and texture of most of the flint, and
the survival in many cases of areas of cortex, were
sufficiently distinctive and consistent to indicate that
most of the material was coming from the beach, and
this gains further support by the consistencies it
shares with other coastal assemblages examined by
the author in the Museum of Antiquities of Newcastle
and Sunderland Museum.
Raw Material
The lithic assemblage is dominated by flint, which
accounts for 227 of the 244 pieces (93%) (Fig. 3.4). The
other material that is relatively common in the
assemblage is chert, which accounts for 14 pieces
(6%). The remaining one percent is made up of two
quartz and one agate pieces.
As many of the flints retained small areas of cortex
this has allowed examination of the type of raw
material being utilised. Of the 119 pieces that could
be identified to type, 118 (99%) were beach flint while
one piece was from nodular flint. No boulder clay
flint could be recognised with certainty, although it is
likely that some of the pieces were produced from
this type of secondary flint source. As nodules of the
same kind of flint as those found during the field-
walking can still be picked up on the Howick beaches
today, it can be concluded that most of the utilised
flint was originally obtained from the beach. This rich
source of flint, in an otherwise relatively flint-sparse
region, was no doubt one factor that attracted Stone
Age groups to this area. The virtually complete
reliance on locally available flint indicates a self-
sufficient community that did not rely on the
importation of stone from distant sources.
A wide variety of flint colours was evidenced in
the assemblage (see Table 3.4), though for some pieces
it was not possible to identify their true colour as
they had developed thick patinas of a different colour.
Otherwise it was flints of various shades of grey that
tended to dominate the assemblage, with 29.5% being
light grey material and 19.3% medium grey. However,
there was also 11.9% red-brown flint, often speckled,
Flint Colour No. Pieces % of Entire
Assemblage
Light Grey 72 29.5%
Medium Grey 47 19.%
Red-Brown 29 11.9%
Orange 12 11.9%
Orange-Grey 10 4.9%
Fawn 8 3.3%
Orange-Brown 7 2.9%
Dark Grey 6 2.5%
White 4 1.3%
Brown 3 1.2%
Table 3.4. Classification of flints by colour (only includes those flints
for which a colour could be assigned).
Types
As is typical, flakes dominate the lithic assemblages,
accounting for 54.9% of all pieces (see Table 3.5).
There is a significant percentage of cores, 15.2%,
which compares with 17% from the Milfield survey
(Waddington 2001b) and 12.7% from the Maiden’s
Hall survey (Waddington 2001a).
Scrapers accounted for 4.1% of the Howick field-
walking assemblage, whereas they only account for
0.2% of the Maiden’s Hall assemblage. The Maiden’s
Hall assemblage also differs in having higher
frequencies of primary material, as is indicated by
the counts for test-pieces; 9.6% for Maiden’s Hall
compared with 5.7% for Howick. However, if the
frequencies of all tools are combined, they form 10.6%
of the Howick survey compared with 10% of the
Maiden’s Hall survey, making them broadly similar.
According to Schofield’s model of expected assem-
blage characteristics (reproduced as Table 3.6 below)
this would suggest that the Maiden’s Hall area was
used primarily for industrial processes, which in this
case were acquisition and preliminary dressing of
flint, while the Howick area was used more for
settlement purposes. However, the contrast is cer-
tainly not as clear-cut as this, as there is evidence for
activities associated with primary flint working and
settlement activity in both areas. Rather, the difference
is one of degree. The Howick landscape, particularly
those areas located near to the freshwater burn,
appears to have been an attractive settlement focus
for hunter-gatherer groups, while the nearby coast
Type No. % of Assemblage
Nodule 5 2%
Bashed Lumps/Test Pieces 14 5.7%
Flakes 134 54.9%
Blades 28 11.5%
Cores 37 15.2%
Modified Flakes 6 2.5%
Modified Blades 4 1.6%
Scrapers 10 4.1%
Microliths 2 0.8%
Awls (inc. possibles) 4 1.6%
Total 244
Table 3.5. Breakdown of lithics by type.
3FIELDWALKING 27
where their specific purpose remains problematic,
though specialist activities associated with catching
birds and/or fish are possible.
Activity Density Primary Waste Tools Cores
Settlement Low Low High High
Industrial High High Low Low
Table 3.6. Schofield’s model of expected assemblage characteristics
(1991, 119).
(now eroded back) provided a source area for flint
extraction. In the case of the Maiden’s Hall area, the
boulder clays were intensively exploited as a source
of flint but settlement also appears to have taken place.
Most pieces in the assemblage are relatively small
and this is probably a direct result of the nature of the
raw material used, as beach flint is not usually as
large as flint nodules from primary geological
deposits.
Most of the cores are blade cores with single
platform cores present, indicating the reliance on a
narrow-blade technology typical of Later Mesolithic
flint-working traditions. In addition, there are
examples of pyramidal, bipolar and flake cores which
are also typical of Mesolithic assemblages from this
region (Fig. 3.5). A range of scrapers was evident in
the assemblage, including Mesolithic ‘tiny’ scrapers
as well as end scrapers and an excellent specimen of
a thumbnail scraper (Fig. 3.6). These types of tool are
usually thought to be indicative of hide working. The
presence of awls in the assemblage is also of note as
these are usually associated with stitching hides and
skins. The working of hides, a processing activity
usually associated with domestic (i.e. settlement)
occupation, is also indirect evidence for hunting and
exploitation of game and terrestrial fauna, even
though this is a coastal setting. Little can be said of
the one certain microlith other than that it most
closely resembles a crescentic form. Microliths of this
type have been found on the excavated site (see
Chapter 7) and also in abundance at the Fife Ness
coastal site in Fife (Wickham-Jones and Dalland 1998)
Activity in the Wider Landscape
The Howick fieldwalking assemblage is informative
on several counts. The artefact types and their stages
in the reduction sequence indicate that a wide range
of activities took place across the area. Most notably
this included the acquisition and preliminary working
of flint from the nearby beach. This comprised the
flaking of flint nodules as well as the reworking of
previously chipped and patinated pieces washed on
to the shore. The large proportion of cores and tools in
the assemblage is indicative of processing activities
that are generally thought to be associated with
settlement sites, suggesting that residential occupation
took place across this coastal strip and was not just
confined to the single structure excavated as part of
this study. As nearly all the diagnostic pieces present
in the assemblage can be associated with Later
Mesolithic flint-working traditions, and this also
corresponds with the radiocarbon dates returned from
the excavation site, it is likely that most of the material
in the assemblage dates from this broad period.
Therefore, it can be argued that during the Later
Mesolithic period the coastal strip around Howick,
and particularly the relatively flat elevated areas close
to the Howick Burn, were used extensively by
Mesolithic groups visiting the coast, both for the
acquisition of flint and also for habitation. It is likely
that the availability of fresh water from the burn was
an important factor affecting the location of settlement
sites. The steep-sided burns that discharge into the
North Sea along the Durham coast are well known for
their extremely rich assemblages of Mesolithic flints,
such as the site at Crimdon Dene (Raistrick and
Westoll 1933; Raistrick et al. 1939), and the Howick
landscape appears to have been similarly exploited.
The fieldwalking survey has been crucial in demon-
strating that Mesolithic exploitation of the landscape
was not confined to the area immediately around the
excavated site, but rather that settlement activity
appears to be spread around this coastal margin on
areas of free-draining flat land within easy reach of
freshwater. Therefore, it can be expected that other
Mesolithic deposits survive in situ below the plough
zone in the fields around the Howick site. In particular
areas of Fields 1, 2, 3 and 4 have been identified as
being of interest (see above).
28
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
4TEST PITS
Clive Waddington
Methods Statement
The test pits were excavated primarily to check if any
other Mesolithic structures or deposits were at risk
from current or future erosion from the cliff edge. A
total of 73 test pits was excavated around the Howick
site (Fig. 4.1). Each pit measured 1m2 and was
excavated down to, or below, the natural substratum
horizon, which consisted of glacial till deposits. After
removal of the turf, each pit was excavated using
trowel, mattock and spade, and the entire contents of
the pit passed through a 5mm sieve to maximise
finds recovery. The pits were positioned to radiate
out from the excavation trench in order to try and
locate any other surviving structures. In particular,
the eroding cliff edge was targeted so that any
remains in danger of immediate erosion could be
recorded. Lines of pits were excavated at varying
intervals of 2, 5, 10, 20 and 50m in order to assess
what spacing interval was required in order to pick
up areas of high artefact densities. The pits aligned
along the cliff edge were excavated at 5m intervals so
that if another hut feature of similar dimensions (c.6m
diameter) to the one already discovered was under
immediate threat of erosion, this sampling interval
would allow for picking up traces of it. The pits were
excavated at the same time as the Mesolithic hut
during summer 2002 by a combination of professional
staff and supervised volunteers and school groups.
Distribution and Artefact Density
A total of 246 lithics was recovered from the pits,
together with 309 modern shells (as they show little
sign of erosion by soil acidity), eight sherds of modern
pottery, two fragments of modern animal bone
(probably sheep), five clay pipe fragments and one
piece of modern glass. The lithic total is substantial,
giving an average count of 3.3 lithics per pit –
although without including pit 1a, which was located
on top of the Mesolithic hut –; this gives a total count
of 195 and an average of 2.7. This compares with 0.7
lithics per pit recovered during the excavation of 146
test pits around the Milfield Basin (Waddington
1999a, 62–71). The Howick results, therefore, indicate
a focus of Stone Age activity around the area of the
Mesolithic hut. No cut archaeological features were
found in any of the test pits.
As all but one of the diagnostic lithics are Meso-
lithic, and as they are directly analogous with the
material from inside the hut, this implies that it is
largely Mesolithic activity that is represented by the
test pit lithic finds. Assuming some degree of
contemporaneity with the hut site, this high-density
artefact spread indicates that activity took place
beyond the immediate confines of the hut. Indeed,
the concentration of lithics in pits 67, 68 and 64
suggests an activity area near to the hut on its west
side, and this fits in with the presence of undated pit
features found between the test pits and the hut (Fig.
4.2). Other concentrations were noted on the cliff in
pits 1–6, before the ground slopes away to the north-
east, suggesting that this area may have been utilised
for a range of outdoor processing tasks. It is apparent,
however, that the density of lithics falls off with
distance from the hut along the north-east line of pits.
Another localised activity focus is suggested by pits
38, 44 and 47 on the north-west edge of the trench,
which produced 19 lithics between them. The concen-
tration of material on what is currently the cliff edge
would have been located in a more sheltered position
set back from the Mesolithic cliff by perhaps as much
as a few hundred metres. Test pit 69, cut into the
collapsed cliff deposits that had eroded out from the
hut site, was not excavated to full depth (i.e. down to
the substratum) as it became heavily waterlogged
during excavation. However, it is likely that most of
the 10 lithics from this pit had eroded out of the
Mesolithic structure above and this would account
for the high concentration in the levels that could be
excavated. The three test pits excavated on the
headland to the south of the Howick Burn around an
erosion scar (pits 71–73) yielded only a low quantity
of material (six pieces from the three pits), and so no
more evaluation took place in this area.
Comparing the results from the lines of pits set at
different spacings aligned inland from the north-west
end of the trench shows an interesting pattern (see
4TEST PITS 29

Figure 4.1. Plan showing the location of all test pits in relation to the excavation site and the cliff edge.
30
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 4.2. Detailed plan showing numbers of lithics per test pit.
4TEST PITS 31
Test Pit No Lithics Shells
Modern
Pot
Modern
Bone Clay Pipe Glass
1a (over meso hut) 51
1 5 8
2 8 1
3 10 8
4 8 5
5 8 2 3
6 5 8
7 1 11
8 3 2
9 1 3
10 3 7
11 3 1
12 4 1
13 5 2
14 10
15 2 4
16 13 1
17 2 5
18 2
19 3
20 1
21
22 1 1
23
24 3
25 3
26 1 1
27 2 4
28 7
29 8 3
30 6
31 1 4
32 4 12 1
33 2 3
34 15
35 4 2
36 2 3
37 4 5
38 7 4
39 1 1
40 4 5
41 1 4
42 1 8
43 1 1
44 7 2
45 1 1
46 3 1
47 5 5
48 3 4
49 1 10
50 4 5
51 1
52 3 3
53 1 5
54 2 3
55 1 3
56 1 5
57 4
58 2
59 4
60 3 9
61 1 7
62 3
63
64 4 2
65 15 1
66 6 1
67 6 2 3 1
68 10 17 2
69 10 2 1
70 4 3 1
71 2 1
72
73 4
Total 246 309 8 2 5 1
Table 4.1. Numbers and types of finds from each test pit.
32
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Interval No. pits No. lithics Average no.
lithics per
pit
2m 10 24 2.4
5m 14 30 2.1
10m 6 15 2.5
20m 5 18 3.6
50m 2 8 4.0
Table 4.2. Lithic counts of pits set at different spacings.
Type Number % of Test Pit Assemblage
Nodule 4 2
Test Piece 9 4
Flakes 156 63
Blades 39 16
Cores 19 8
Retouched Flakes 9 4
Retouched Blades 1
Utilised Flakes 3 1
Utilised Blades 3 1
Microliths 2
Scraper 1
Total 246
Table 4.3. Lithic types recovered from the test pits.
Table 4.2). The average number of lithics per pit tends
to rise if the interval spacing is larger and if there are
fewer pits. This is perhaps the opposite of what may
have been expected, as pits with a greater spacing are
less likely to provide an accurate indication of
densities and fall-off than more closely spaced pits.
The 50m spacing result is therefore probably skewed
by pit 44 which produced 7 lithics and, given that
only one other pit was excavated in this line, the
average density has inevitably remained high. The
same is true for the 20m spaced pits where the high
count from one pit, in this case pit 38 with 7 lithics,
has raised the average density count for the few other
pits included in this line that usually only contained
just one lithic. However, it is telling that the results
from pits positioned at 2m, 5m and 10m intervals are
remarkably consistent, ranging from 2.1 to 2.5 lithics
per pit. Given that the 2m spaced pits are likely to
give an accurate indication of lithic densities, as it
represents a 50% sample of the entire plough zone
assemblage in any given line.These results suggest
that pits positioned at 5m and 10m intervals can also
provide a reasonably accurate reflection of the
patterns. However, interval spacings greater than this
tend to give results inconsistent with the 2m ‘control’.
Another way of undertaking this experiment would
have been to excavate a long line of 2m spaced pits
and then record every 2nd, 5th, 10th, 20th and 50th pit
and compare those results. This would ensure ab-
solute consistency across the ground being examined.
The Lithic Assemblage
The test pit assemblage reveals a wide range of pieces,
including primary material, such as nodules and test
pieces, which together account for 6% of the total, as
well as the usual preponderance of flakes (63%),
blades (16%) and cores (8%), but also a significant
proportion of tools (7%). The presence of primary
waste, together with the cores, blade blanks, flake
debitage and finished tools, shows that the entire
reduction sequence is manifested in the assemblage.
This is consistent with the findings from both the
fieldwalking and the excavated lithic assemblages,
reinforcing the view that flint was collected locally
and worked into tools at and around the site. As 55%
of the test pit lithics are broken, including most of the
tools, many of these represent pieces discarded in
antiquity, although some breakages have evidently
resulted from plough damage. The former is con-
firmed by a number of the broken lithics having
patination development over the broken area. The
important point here is that if some of the tools were
broken and discarded in antiquity then not only is
the entire reduction process evidenced in this assem-
blage but also the discard of tools after they have
been used. In other words, the tools were also used
here in processing activities of various kinds. A single
scraper was found, hinting towards hide preparation,
while the two broken microliths, although unclass-
ifiable, may have formed part of a hunting tool kit, or
as some commentators have suggested, been used in
plant processing tasks. The retouched and utilised
flakes and blades could have been used for a wide
variety of purposes. The presence of a range of
processing tools indicates that activities consistent
with domestic occupation took place in the open air
in the immediate environs of the hut. Some of the
lithics were patinated (38%), but most of these only
lightly, while 9% of the lithics were burnt.
As with the fieldwalking and hut flints, most of
the cores were platform varieties of one type or
another, but as most of them were small they had
evidently been struck wherever there was an oppor-
tunity for producing a blade. On the whole the
assemblage is geared around a blade-based tradition,
with the use of bipolar flaking evident in a few
instances. Many of the retouched pieces have areas of
cortex remaining on them, revealing the small size of
the raw materials generally available for chipping.
The variation in the size of the bulb of percussion and
evidence for crushed striking platforms in some cases
shows the use of both hard and soft hammers. Only
one piece in the assemblage has been pressure-flaked
to produce some fine invasive retouch. This broken
retouched flake from pit 22 is likely to be Neolithic in
date and, together with the leaf arrowhead recovered
in the nearby excavation trench (Chapter 7), shows
some evidence for activity on the site in a later period.

Figure 4.3. Number of lithics made from different raw materials.
4TEST PITS 33

Figure 4.4. Number of lithics made from different coloured raw
material.
It is notable that a lower proportion of cores was
recovered from the test pit assemblage (8%) com-
pared with the fieldwalking assemblage (15.2%),
which suggests that cores are perhaps over-rep-
resented in surface collections, no doubt on account
of their larger size and greater visibility.
The spread of raw materials used for making stone
tools during the Mesolithic is skewed towards the
use of flint (86%). However, other locally occurring
materials were also used from time to time, probably
as washed-up beach material, and this included chert
(7%), quartz (6%) and one piece of agate. Of all the
flint that could be provenanced, every single one was
of beach flint (14% of the test pit assemblage). As
with the fieldwalking and excavation results, the test
pit lithics demonstrate a reliance on locally available
material from the nearby beach and cliff-sections.
The range of flint colours present in the test pit
assemblage was directly analogous to that in the
fieldwalking, and what is more, the colours occurred
in the same proportions (Fig. 4.4). All the types of
flint common on the North-East coast are represented
in the assemblage, with varieties of light grey the
most common, followed by medium grey and red-
brown (see Fig. 4.4). These types of flint are common
in other coastal collections such as those from New-
biggin, the Bamburgh area and from areas on the
Durham coast such as Crimdon Dene and Middle
Warren. Thirty pieces could not be attributed a colour
as they were either too heavily patinated or burnt.
Comparison of the lithic density from the test pit
above the Mesolithic structure with those from the
test pits away from the site provides a rough in-
dication of the sort of densities that can be considered
indicative of a dwelling site. A total of 51 lithics was
recovered from test pit 1a that was excavated above
the inside of the Mesolithic hut, while the highest
counts in any of the other test pits was 10, though the
average was 2.7. During the initial stripping of the
site in 2000, all the topsoil and subsoil was removed
in 1m squares and sieved, so in effect this soil was
removed as a series of 1m square test pits. The highest
density of lithics occurred immediately over the hut
area (up to 51 lithics in a single metre square), with
the densities gradually falling off with distance from
the hut. Although these numbers can only be taken as
guides, these figures provide baseline data for
assessing how many lithics can be expected in a pit if
occupation deposits had accumulated in that position.
34
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
5EXCAVATION OF THE HOWICK HUT
Clive Waddington
Methods Statement
Further to the observation of flints falling out of the
erosion scar (see Chapter 1), the site was first
investigated with a single 1m square test pit during
May 2000. This revealed a stratigraphic sequence
consisting of a topsoil, subsoil and sand substratum,
the latter containing a cut archaeological feature.
Lithics and shells were recovered from all these layers
and charred hazelnut shells from the top of the
archaeological feature. Based on these findings a
strategy was adopted for an evaluation excavation
that took place during June and July 2000. This
involved removing the overburden in metre squares
and bucket-sieving all the soil through a 5mm mesh
in order to maximise finds recovery and to gain some
sense of flint distribution and density. The topsoil of
a trench covering 195 square metres was stripped
away and sieved, followed by removal of 90 square
metres of the subsoil, repeating the same metre by
metre sieving, down to the beginning of the archaeo-
logical horizon (Fig. 5.1). Although the results of
sieving the subsoil were to some extent informative
(see below), it was reasoned that the time taken to
remove the overburden by this method did not justify
the meagre results. The full excavation of the site,
which took place during summer 2002, involved
opening a much larger area than the evaluation,
measuring 68m by 20m and extending over some
1120 square metres. The layout and extent of this
trench was based on the results from the initial
evaluation and the geophysical survey. In order to
ensure the full excavation of all the deposits at risk
from erosion, and to maximise the time available to
record intact deposits, the topsoil across the entire
area was removed by a machine, leaving the subsoil
to be removed by hand. The start of the archaeological
horizon was trowelled back to reveal the surviving
features.
All features were excavated in their entirety with
full records made using plans, sections, photographs
and recording forms. A single-context planning
system was employed to record the archaeology and
this proved useful in helping to unpick the complex
archaeology of the Mesolithic hut deposits. All small
finds were spatially located using a total station. The
area of Mesolithic deposits was situated in a localised
area of sand inset within the till deposits and all
features within this unit required delicate excavation.
This meant that small tools such as plasterer’s leafs
and spoons were employed for most of the excavation
work rather than trowels. A 100% sample of most
deposits from the Mesolithic structure was passed
through a flotation tank which had graduated brass
sieves to collect the flot and a 2mm mesh to catch
non-organic material (Fig. 5.2) (see also Chapter 11).
This not only ensured collection of organic residues
but it also maximised the recovery of small finds, and
lithic material in particular. The Bronze Age remains,
which consisted of a cist cemetery together with a
later linear burning pit, were cut into the heavier clay
till deposits to the north of the Mesolithic hut. The
excavation of these features is reported in full in a
separate publication (Waddington et al. 2005).
Site Stratigraphy
The Mesolithic site is situated in grassland given over
to pasture for livestock. The sandy topsoil, which is a
plough zone, varies between 0.22m and 0.4m in
thickness and overlies a red-brown sandy subsoil
which varies between 0.1m and 0.2m thick. Below the
subsoil was the sand substratum, although in the
north-eastern half of the site this changed to glacial
till. A schematic summary of the geoarchaeological
setting of the Howick site is provided in Figure 5.3.
There were no traces of archaeological features in the
topsoil as this ground has been ploughed on several
occasions during living memory. The occurrence of
clay pipe stems and bowls in the topsoil suggests the
ground has been ploughed as far back as the 18th
century, perhaps in association with the enclosure of
land. The subsoil and topsoil have also been affected
by bioturbation (including moles, worms, soil chem-
istry and so on) over the last 10,000 years. However,
in some instances an extremely vague discolouration
in the subsoil horizon above an archaeological feature
5EXCAVATION OF THE HOWICK HUT 35

Figure 5.1. Trench plan showing the location of the 2000 and 2002 trenches and archaeological features.
could be observed, representing the ‘ghost’ of a
feature in this layer. The bioturbation-driven trunc-
ation in the subsoil has not actively broken through
features, as is the case with plough truncation, but
has had the effect of breaking down the integrity of
deposits and effectively ‘smudging’ them into the
surrounding matrix. Therefore, it was only when the
subsoil was removed, and the natural substratum
exposed, that intact archaeological features could be
distinguished. It is evident, then, that the top lens of
the Mesolithic archaeology, and much of that from
later periods, has been lost and it is only the features
that were cut into the underlying substratum that
have survived.
Mesolithic archaeological features were only distin-
guishable in the lowest 6cm of the subsoil horizon.
The survival of a fallen timber associated with the
Mesolithic structure, with only the top few centimetres
disturbed by bioturbation and soil weathering,
indicates that only the top few centimetres (i.e. 6cm)
of this archaeology has been truncated. This means
that the Mesolithic land surface must be located in
what is now the lower subsoil horizon. Therefore, the
Mesolithic soil appears to have consisted of a thin soil
36
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
cover immediately overlying the sandy sediments,
with the bulk of the modern soil having built up since
then. In contrast, the Bronze Age remains survived
through the entire subsoil horizon, indicating that they
had been cut down from a higher, truncated, level
within what is now the plough zone horizon. There-
fore, the Bronze Age land surface was considerably
higher than in the Mesolithic and was situated
somewhere in what is now the topsoil. It can be
concluded that a considerable build-up of soil took
place on the site between the Mesolithic and Bronze
Age occupations, a period of some 5800 years.
During the 2002 excavation the topsoil was
surface-stripped and the subsoil cleaned back. This
produced a quantity of unstratified lithic material
from across the site (Fig. 5.4). A clear distinction was
evident between the north-east and south-west ends
of the site. The north-east half of the trench lay over a
heavy clay till and contained remarkably few finds
compared with the south-west half of the trench that
contained a sandy hollow set within the surrounding
till and had a distinct abundance of finds. The
Mesolithic hut was situated in the soft sandy deposits,
demonstrating a clear preference for site location on
this free-draining substratum. The cists were situated
further to the east with some cut into sandy
sediments and others cut into heavier till sediments.
The overburden
Table 5.1 summarises the finds recovered from both
seasons of excavation in the overburden. During the
2000 season, the entire topsoil and subsoil within the
area of the evaluation trench were sieved and counts
recorded on the basis of a 1m square grid system. The
distribution of lithic material in the topsoil showed
little patterning across the site other than concen-
trations in the general area of the hut; however, a
concentration of lithics was found in the subsoil
directly on and around the hut area. This finding is
considered important as it indicates that there has
been relatively little horizontal movement of lithic
material in the subsoil, whereas that in the overlying
topsoil has been moved both horizontally and
vertically through the stratigraphic sequence. This is
not surprising given that the field has been cultivated
and ploughed, probably over many hundreds of
years, resulting in the displacement of material in the
plough zone. The spread of lithics around the hut is
likely to indicate that processing and domestic
activities took place in the immediate environs of the
building, however this assumption must be viewed
with caution, as the subsequent ploughing of the site
must account for the spread of much of this material
in the topsoil. The concentration of lithics immed-
iately above the hut deposits in the unploughed
subsoil appears to reflect a remarkably concentrated
pattern of lithic discard in the hut, suggesting that
most processing tasks took place indoors.
Mesolithic Hut
The only radiocarbon-dated Mesolithic feature in the
trench was the remains of a substantial sub-circular
hut situated along what is now a cliff edge (Fig. 5.5).
It is estimated that about a quarter of this structure
has been truncated on its south-east side as a result of
slippage down the cliff and modern sheep burials.
The cliff-top fence used to run along the truncated
edge of the structure, and the remains of a modern
fence post were found driven into the truncation line
on the east side of the structure. However, the rest of
the hut survived in a remarkable state of preserv-
ation, with clear structural evidence, including the
stains of timbers, as well as post holes, post sockets,
stake holes, pits and hearths, together with an
abundance of artefactual material including over
13,000 flints (Chapters 7 and 8), 32 bevel-ended tools
and tool blanks, several hundred thousand charred
hazelnut shells (Chapter 11), together with burnt bone
fragments (Chapter 10) and occasional fragments of
mollusc shells (Chapter 10) and ochre (Chapter 8).
The site has experienced considerable mole damage
but the mole holes were clearly visible and could be
distinguished from the archaeological features; this
meant care had to be taken when sampling features
for dating material.
Deposit Lithics Shell Pottery Clay Pipe Other Flints previously found at the
surface by John Davies and Jim
Hutchinson
Topsoil 604 794 109 7 ochre, bone, glass, plastic,
iron, brick, slag,
Subsoil 239 63 7 1 ochre, glass, lead, iron,
bone, slag
Sand Substratum 3 3 small burnt stones
Till Substratum 55
233
Total 901 857 116 8
Table 5.1. Summary of finds from the overburden and underlying natural sediment surfaces.
5EXCAVATION OF THE HOWICK HUT 37
The hut had three distinct constructional phases,
indicated by the stratigraphic sequence and sup-
ported by the chronological spread of radiocarbon
dates throughout these deposits (see Chapter 6). No
turf lines or hiatuses could be identified in the
stratigraphic sequence that would indicate any
substantial period of abandonment. Many of the
archaeological features survived as ephemeral re-
mains with the constituent make-up of most fills
comprising a sand of varying colour and compaction.
All unburnt organic content had decayed, leaving
features such as timbers only visible as stains and 3-
dimensional shapes in what was otherwise an acidic
sand matrix (see Chapter 12). Features could be
identified and delimited through distinctions in
colour, texture and compaction, and in the case of
heat-affected features, by fire-reddening and the
presence of black charred material. It also became
apparent while excavating that a number of features
had developed a slight iron/mineral stain around
their edges where the iron/mineral fraction in the
soil and sediments had accumulated at the edge of
archaeological cuts. This created a ginger halo effect
around a number of features, particularly those cut
into the sand substratum. In addition to this, three
ferruginous layers were identified running across the
entire hut deposits and beyond into the natural sand.
At first this complicated the recording and inter-
pretation of the archaeological deposits, but once it
was observed that these lenses continued into the
surrounding sand it was realised that they had
formed as a result of natural soil processes.
The stratigraphy of each phase is discussed in turn
below and full descriptions of each context are
provided in the accompanying phase tables. There

Figure 5.2. On-site flotation of sediments from the Howick hut.
are accompanying plans for each phase together with
section drawings. The sections of the two baulks,
shown on the Phase 1a plan, are also included in the
Phase 1a section. An exploded view of the phase
plans is provided in the concluding section.

Figure 5.3. Schematic section of the archaeological remains and surrounding sediments at Howick.
38
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.4. Trench plan showing lithic findspots.
Phase 1a Hut (see Figs. 5.6–5.9)
Within the first structural phase of the hut was a
succession of internal features and debris that had
accumulated during this first phase of occupation.
These were excavated layer by layer and form the
subdivisions a and b of the Phase 1 hut. Whether they
represent different phases of occupation, or rather
the successive build-up of debris from one single
permanent occupation remains unclear. However, the
Phase 1b occupation may be related to minor struc-
5EXCAVATION OF THE HOWICK HUT 39
tural alterations suggested by some flat stone pads,
possibly for supporting timbers, and some stake holes
identified in the Phase 1b deposits.
Structural Features
The first phase of the hut consisted of a sunken-floored
circular structure cut into sand deposits on a flat area
of land (Fig. 5.6). The sunken-floor area measured 6m
across and averaged 0.4m–0.45m deep below the start
of the archaeological horizon, except in the centre
where some of the hearth features were cut to a depth

Figure 5.5. Aerial view of the Mesolithic hut during the early stages of excavation.
Figure 5.6. View of the Phase 1a hut after complete excavation looking north-east (2m scale).
40
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.7. Plan of the Phase 1a hut.
of 0.65m. None of the ‘occupation deposits’ (see Phase
1b section) could be directly linked to this first phase
of construction, although some parts of these deposits
must relate to this initial occupation.
A series of circular features [318, 345, 343, 330, 373,
311, 353, 349, 314, 272, 369], consistent with the
appearance and form of sockets for timber posts, was
identified set in a ring within the sunken-floored area
(Fig. 5.6). These features were circular or sub-circular
in shape and consisted of shallow, usually flat-
bottomed, sockets with the bases averaging 0.15m in
diameter and averaging between 0.12m and 0.2m
deep. None of these sockets had any evidence of
packing material and it is thought that they were
formed by stout timbers being driven into the sand
substratum with the weight of the roof used to hold
them in position.
Set at a higher level, 0.25m above the floor of the
structure around the perimeter edge, was a ring of
evenly-spaced stake holes [240, 242, 244, 247, 249,
251, 255, 257] that contained the degraded charred
ends of sharpened poles. These stake holes were
covered by the upper deposit [210], into which the
Phase 2 hut post sockets were set. Therefore, it is
evident that these stake holes form part of the Phase
1 hut. The stake hole slots averaged 0.04m in diameter
but as these were only the sharpened ends the actual
thickness of the poles is thought to have been in the
order of 0.07m in diameter. The poles were set in the
ground at an angle close to 65° and aligned towards
a central apex, thus forming a steep angle of pitch.
Two linear features [359] and [377] may have had
a structural purpose. Both of these features survived
as ginger-brown stains set in the natural yellow-
brown sand and had curved profiles. Feature [359]
contained occasional flints, charred hazelnut shells,
mollusc shell fragments and charcoal. Feature [377]
contained no small finds. These contexts were
considered to be structural components of the hut as
they were positioned around the edge of the structure
and a stake hole, [375], was cut into the fill of [359].
Precisely how these linear sockets were utilised
remains unclear although it is possible that feature
[359] is the remains of a fallen timber post surviving
only as a stain in the sand, as it had a round profile
and even appeared to have had a bough removed
from it on it one side. As unburnt organic material
decays in this acidic environment, leaving only a sand
matrix, it is not surprising that occasional lithics and
other small finds have found their way into such
deposits as a result of earthworm action and other
mixing processes.
5EXCAVATION OF THE HOWICK HUT 41

Figure 5.8. Phase 1a feature sections.
42
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.9. The baulk sections.
Hearths
The primary central hearth [355] measured 2.3m long
at its maximum extent and in places measured up to
0.58m deep. This is a very large pit that contained
evidence within its fill for multiple burning events.
Indeed this sizable hearth appears to have formed as
a result of numerous re-cuts resulting in a large
irregular pit with cuts at different levels. The ground
around the pit was heavily scorched and fire-
reddened and the fill was blackened as a result of the
burnt and charred material. Grey patches of ash were
also present. The fill contained large quantities of
burnt flints, charred hazelnut shell and charcoal,
together with occasional fragments of burnt bone.
Two radiocarbon dates were obtained from in situ
single-entity charred hazelnut shell samples (see
Table 5.2 and Chapter 6). This hearth had an unusual
burnt feature resembling an angled post hole, aver-
aging 0.22m in diameter, projecting below the base of
the hearth. This is thought to be the likely remains of
a degraded charred post; probably associated with
fire-side activities such as cooking. This burnt feature
had been set deep into the hut floor, reaching a depth
of 1.2m from the start of the archaeological horizon,
and 1.6m below the modern ground surface, making
it the deepest archaeological feature on the site.
Lying to the east of the central hearth area [355]
were two other large hearth pits, [357] and [379]. Of
these, [379] was the earlier as it was cut by [357]. As
with hearth [355] both pits were somewhat irregular
in shape, but substantial, with [379] measuring up to
1.25m across and 0.67m deep while [357] measured
up to 0.33m across and 0.12m deep. Both had dark
charred fills with areas of ash together with fire-
reddened edges and patches of burnt, fused sand.
Both hearth pits also contained many pieces of flint,
some of them burnt, together with abundant charred
hazelnut shells, charcoal, and in the case of [357],
occasional fragments of burnt bone. Single-entity
charred hazelnut shell fragments, from in situ burnt
residues from pits [379] and [357], were submitted
for radiocarbon dating. Intriguingly, as with hearth
[355], evidence for a dark charred circular post was
observed on the north-west side of hearth pit [357].
This small post socket [369] measured 0.18m in
diameter and had a depth of 0.14m. Two opposed
slots were extant at the edges of the hearth [379], one
of them also positioned on the north-west side. The
suggestion here is that these three major hearths,
occupying the central area of the hut, may have had
timber supports associated with them.
Two much smaller burning pits, [383] and [381],
were also observed cut into the natural sand at this
level. Pit [383] contained a black charred sandy fill
surrounded by fire-reddened sand indicative of in
situ burning. Two single-entity charred hazelnut
shells from an in situ burning event within pit [383]
5EXCAVATION OF THE HOWICK HUT 43
were submitted for radiocarbon dating. The fill
contained charred hazelnut shell and burnt bone
fragments. Pit [381] was slightly larger and had a fill
similar to that of [383] but also contained charred
hazelnut shell fragments. These pits lay beyond the
large central hearth pits but still within the central
area of the hut space. Neither had evidence in its fill
for multiple burning events, as in the case of the three
large hearths.
Two later hearths, [367] and [291], were located
cut into the primary hearth pit [355]. Both of these
were small; [367] measuring 0.33m by 0.23m by 0.12m
deep and [291] 0.41m by 0.24m by 0.055m deep. Both
had black charred fills surrounded by fire-reddened
sand indicating in situ burning. Both contained flint,
charred hazelnut shell and occasional burnt bone
fragments and charcoal. Two single-entity charred
hazelnut shells, resulting from in situ burning in
hearth [291], were submitted for radiocarbon dating.
Other Internal Features
The other interior features discovered in the Phase 1a
hut consisted of four pits of varying form and no doubt
function. Two small pits, [274] and [351], were located
closest to the central hearth area, suggesting they may
have been associated with domestic activities such as
food preparation or storage. Pit [274] measured 0.32m
across by 0.13m deep and was filled by dark grey-
brown loose sand that contained the occasional flint.
Pit [351] measured 0.22m by 0.17m across by 0.07m
deep and was also filled by dark grey silty sand with
occasional fragments of charred wood but no evidence
for in situ burning. Pit [365], located on the south-east
side of the structure, was oval in shape and somewhat
larger than pits [274] and [351], measuring 0.75m by
0.61m by 0.42m deep. It was filled by a grey-brown
silty sand that contained occasional flints. However,
by far the largest pit, [270], had an irregular arc-shape
and was situated on the north-east side of the hut. It
had a maximum length of 2m, varied considerably in
width from 0.4m to 0.8m and averaged 0.13m deep.
The fill was generally darker than the other pits as it
contained black charred material and loose brown sand
towards the edges of the feature. The fill contained
occasional flints and charred hazelnut shells, though
there was no definite evidence for in situ burning. The
purpose of this pit remains uncertain although it
appears to be associated with the arc of elongated
pebble tools situated in this part of the hut (see also
Chapters 8 and 14) and it is worth noting that the
deposit was cut by post socket [272]. This pit is early
in the structural sequence of the Phase 1a hut and
could be associated with possible construction slot
[377] and slot/timber [359] that continue around the
north and west sides of the hut.
44
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14 CDates bp
(uncal.)
Hearths
355 Large and
complex irregular
hearth area with
multiple burning
episodes within
2.3 x 1.1 0.58 Black with fire-
reddening and
brown patches,
surrounded by
scorched sand
Charred sandy silt
with occasional
Small stones
557 flints, burnt
bone, c.43000
charred hazelnut
shell fragments,
charcoal, charred
wood, sandstone
stone fragment
8734±37 OxA-11801
8725±39 OxA-12327
379 Large, deep
irregular hearth
1.25 x 1.0 0.67 Black with fire-
reddening and
brown patches,
surrounded by
scorched sand
Charred sandy silt 262 flints, over 5000
charred hazelnut
shell fragments,
charcoal, charred
wood, beveled
pebble tool blank
8785±45 OxA-11856
8750±45 OxA-11857
357/371 Large, irregular
hearth
1.0 x 0.8 0.3 Black with fire-
reddening,
surrounded by
scorched sand
Burnt sand 147 flints, burnt
bone, c.4500 charred
hazelnut shell
fragments, charcoal,
charred wood, a
beveled pebble tool
blank
8754±38 OxA-11802
383 Sub-circular small
hearth
0.27 diameter 0.12 Black surrounded
by scorched sand
Burnt sand 1 flint, burnt bone,
charred hazelnut
shell, charcoal,
charred wood
8885±65 OxA-12402
8785±40 OxA-12292
381 Sub-circular
hearth
0.47 x 0.37 0.25 Black surrounded
by scorched sand
Burnt sand Charred hazelnut,
charred wood
367 Hearth 0.33 x 0.23 0.12 Black with fire-
reddening,
surrounded by
scorched sand
Charred sandy silt 12 flints, burnt bone,
c.80 charred hazelnut
shell fragments,
charcoal, charred
wood
291 Small oval hearth 0.41 x 0.24 0.055 Black with light
brown sand
lenses
Burnt silty sand 7 flints, burnt bone,
c.100 charred
hazelnut shell
fragments, charcoal,
charred wood
8763±38 OxA-11803
8765±40 OxA-12326
Post Sockets
318 Ovoid post socket 0.28 x 0.14 0.2 Grey brown Silty sand
345 Circular post
socket
0.15 x 0.13 0.21 Dark grey brown Silty sand 3 flints, 45 charred
hazelnut shell
fragments
343 Subcircular post
socket
0.17 x 0.14 0.09 Grey brown Silty sand 30 charred hazelnut
shell fragments
330 Ovoid post socket 0.15 x 0.13 0.1 Medium brown Silty sand 1 flint, 16 charred
hazelnut shell
fragments
373 Ovoid post socket 0.17 x 0.14 0.12 Light grey brown Silty Sand 1 flint, 25 charred
hazelnut shell
fragments
311 Post socket, with
disturbed fill
caused by
extraction
0.22 diameter 0.1 Medium brown Silty sand 1 flint and charred
hazelnut shell
353 Ovoid possible
post socket (3
phases)
0.17 diameter
0.14 diameter
0.08 diameter
0.1
0.2
0.28
Medium brown Sand 7 flints, c.40 charred
hazelnut shell,
fragments
349 Post socket 0.25 x .33 0.21 Grey brown Silty sand 16 flints, c.120
charred hazelnut
shell fragments
314 Large ovoid post
socket
0.4 x 0.3 0.2 Medium brown Silty sand 1 flint
272 Circular post
socket
0.17 diameter 0.1 Medium yellow-
brown
Loose silty sand 20 charred hazelnut
shell fragments
369 Post socket 0.18 diameter 0.14 Dark Brown Silty sand Over 200 charred
hazelnut shell
fragments
375 Stakehole 0.07 diameter 0.12 Black Charred wood
Stakeholes at
higher level
240 Circular stakehole
fill
0.05 diameter 0.04 Black Charred wood
degraded into sand
matrix
242 Circular stakehole
fill
0.045 x 0.04 0.015 Black Charred wood
degraded into sand
matrix
245 Circular stakehole
fill
0.04 x 0.04 0.05 Black Charred wood
degraded into sand
matrix
5EXCAVATION OF THE HOWICK HUT 45
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14 CDates bp
(uncal.)
242 Circular stakehole
fill
0.045 x 0.04 0.015 Black Charred wood
degraded into sand
matrix
245 Circular stakehole
fill
0.04 x 0.04 0.05 Black Charred wood
degraded into sand
matrix
247 Circular stakehole
fill
0.04 x 0.03 0.015 Black Charred wood
degraded into sand
matrix
249 Circular stakehole
fill
0.06 x 0.05 0.03 Black Charred wood
degraded into sand
matrix
251 Circular stakehole
fill
0.045 x 0.045 0.035 Black Charred wood
degraded into sand
matrix
255 Circular stakehole
fill
0.035 diameter 0.035 Black Charred wood
degraded into sand
matrix
3 flints
257 Circular stakehole
fill
0.05 x 0.04 0.02 Black Charred wood
degraded into sand
matrix
Linear
Features
377 Linear, probably a
timber
1.0 x 0.3 0.4 Medium ginger
brown
Compacted sand
359 Linear, probably a
timber
2.5 x 0.4 (av.
0.25)
0.26 Medium ginger
brown
Sand c.250 charred
hazelnut shell
fragments, 36 flints,
charcoal
Pits
270 Linear pit 2.0 x 0.8 0.13 Black with brown
areas towards
edges
Loose sand 113 flints, over 3000
charred hazelnut
shell fragmentst, 2
bevelled pebbles, 2
sandstone cobbles
and a sandstone
fragment
274 Sub-circular Pit 0.32 x 0.32 0.13 Dark grey brown Loose silty sand 32 flints, c.200
charred hazelnut
shell fragments
365 Ovoid pit 0.75 x 0.61 0.42 Medium grey-
brown
Silty sand 142 flints, over 2000
charred hazelnut
shell fragments
351 Ovoid shallow pit 0.22 x 0.17 0.07 Dark grey Silty sand Charred wood, 4
flints, c.25 charred
hazelnut shell
fragments
Floor Level
005 Sand spread Medium brown Sand Charred hazelnut
shell fragments, 263
flints, 4 bevelled
pebble tools on the
original hut floor and
a beveled pebble
blank
Table 5.2 Phase 1a Context Descriptions
Phase 1b Hut (see Figs. 5.10–5.12)
Immediately overlying the primary Phase 1a deposits
was a series of later features that comprised internal
occupation deposits but with little evidence for a full
structural rebuild of the hut, though some building
alterations and/or maintenance may be inferred.
Post Sockets and Stake Holes
A number of stake holes were observed in this
stratigraphic phase, although most of them were in a
localised area running along the south and west edges
of the hut, suggestive of repairs or maintenance in
this area (Fig. 5.10). Stake holes [300], [332], [334],
[336], [338] and [307] were grouped around a small
flat stone next to the south edge of the structure cut
into occupation deposit [326]. These stake holes
averaged 0.06m in diameter and varied in depth from
0.04m to 0.12m. They were filled with a dark grey
charred deposit of degraded wood that had decayed,
leaving a sand matrix which contained occasional
charred hazelnut shell fragments. A larger stake hole
or post socket, [310], also lay along the south edge of
46
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.10. Plan of the Phase 1b hut.
the structure. Stake hole [238] was an isolated feature
located further into the hut, cut into burnt spread
[264]. Two further stake holes, [289] and [287], were
located round the south-western edge of the structure
and had similar dimensions to the stake hole group.
Possibly associated with post socket [310] were the
stone pads/wedges [309], [386], [306] and [385] that
were also set inside the hut edge running along the
south and west sides. If timbers had been placed on
these stone pads then they form a continuation of the
structural features represented by the stake hole
group and post socket [310]. If this is the case, as
seems likely, then it suggests that a repair of the
south-west arc of the hut took place during this phase.
Two groups of small stone fragments were also found
at this level, set in the yellow sand occupation deposit
[316]. These small stones were too small to have had
any structural purpose or than as wedges to stabilise
timber posts and small caches of flint cores were
found close to each of these groups of stones. Maybe
flint knappers had sat with their backs against the
timber posts while they produced flint tools.
was located next to, and partly overlying, burnt spread
[340] in the south sector of the hut. This was a very
shallow spread of distinctive medium brown silty
sand. It contained occasional flints, charred hazelnut
shell fragments and charcoal. Its slight form and
unusual shape suggests it was not a structural feature,
though it could have had a use associated with
activities specific to the south side of the hut interior.
Linear Feature
An irregular linear feature in the shape of an ‘r’, [298],
Occupation Deposits
Two occupation deposits, [316] and [326], surrounded
the burning spreads [264] (see below) and [320].
Deposit [326] consisted of medium brown sand that
had a maximum extent of 2.7m by 1.78m and
contained flint and charred hazelnut shell fragments.
Deposit [316] consisted of light brown sand and
extended over the rest of the hut floor around the
burnt spreads with a width of up to 5.5m and
extending up to the truncated edge. The deposit
contained flint, charred hazelnut shell, occasional
mollusc shell fragments and charcoal. It also con-
tained five small pieces of sedimentary rock, each
containing copper mineral making the stone heavy.
Unfortunately these samples were lost during transit
for analysis.
5EXCAVATION OF THE HOWICK HUT 47

Figure 5.11. Phase 1b feature sections.
48
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.12. Hearth 357 (Phase 1a) after half-sectioning looking north.
Hearths
A sequence of well preserved stratified hearths was
found stacked over the primary hearths [355], [379]
and [357] of Phase 1a. Lying immediately above the
primary hearths was an extensive burnt spread [340]
that covered the central area of the hut. It measured
2.8m wide and continued all the way to the truncated
edge and was, therefore, over 3.75m long. Patches of
discrete in situ burning were observed within the
spread and single-entity charred hazelnut shell
samples from one of these patches were radiocarbon
dated. The spread varied in colour and texture from
black charred patches through various shades of grey
and brown to fire-reddened and ginger scorched
sand. Large quantities of flint and charred hazelnut
shell, together with some fragments of burnt bone,
were recovered from the fill. Two discrete spreads of
burning, [264] and [320], were noted overlying
burning spread [340]. Burnt area [264] lay slightly to
the west of centre in the hut and extended over an
area 2.7m by 2.2m with a largely uniform thickness of
0.1m. Again, patches of in situ burning were observed
across this deposit that consisted of black and grey
charred material in a scorched sand matrix. This
deposit also contained flint, charred hazelnut shell
and burnt bone fragments. Abutting [264] was
another burnt spread, [320], that contained patches of
in situ burning set in a charred black and burnt sand
deposit. This spread measured 2.54m by 2.5m and
had a thickness of 0.1m. It contained flints and
charred hazelnut shell.
A hearth pit, [293], cut into the north-east side of
burnt spread [264], measured 0.5m by 0.38m by 0.2m
deep. The fill of this hearth pit consisted of black
charred material and ginger scorched sand and
contained flints and charred hazelnut shell. Two in
situ single-entity charred hazelnut shell fragments
from this hearth were radiocarbon dated (see Table
5.3 and Chapter 6), although the first of these two
dates is clearly residual from earlier activity at the
site (see Chapter 6). A linear sinuous burning deposit
[283] lay above burnt spread [264] and measured
0.68m long by 0.13m wide by 0.08m deep. This
deposit had the texture of charred wood and is
thought to most likely represent the remains of a
burnt log. The sand immediately around it was
heavily burnt indicating that it had been burnt in situ.
Occasional flints and charred hazelnut shell were
found around the edges of the deposit and these are
thought to have become worked into the sandy fill
after the wood had degraded. Truncating hearth pit
[293] and the probable burnt log [283], and cut into
burnt deposit [264], was a well-preserved hearth pit,
[268] (Fig. 5.12). This latter hearth had a broadly oval
shape and measured 1.3m by 0.61m by 0.22m deep. It
was filled with the characteristic black and dark grey
charred material in a heat-affected red-ginger sand.
The fill contained flint, charred hazelnut shell and
occasional burnt bone fragments. Two single-entity
charred hazelnut samples from in situ burning within
hearth pit [268] were radiocarbon dated (see Table 5.3
and Chapter 6).
Pits
Three pit features belonged to this phase of occu-
pation. Pit [285] was cut into burnt spread [264] close
to hearth pit [268] and probable burnt log [283]. This
5EXCAVATION OF THE HOWICK HUT 49
pit measured 1.0m by 0.62m by 0.12m deep. It was
filled by a dark to medium grey silty sand and ash,
which contained occasional flints and charred hazel-
nut shell. The fill of this pit is consistent with hearth
debris but there was no evidence for in situ burning
and therefore may have functioned as a raking pit.
Pit [214], although also cut into burnt spread [264],
was significantly smaller, measuring 0.54m by 0.22m
by 0.06m deep. The pit was filled by yellow-brown
sterile sand that contained occasional charred hazel-
nut shells. Pit [324] was situated on the east side of
the structure next to the truncated edge. It had an
oval shape and measured 0.7m by 0.46m and was
shallow, with a maximum depth of just 0.04m. The
fill was clean orange-brown sand with occasional
flecks of charcoal and charred hazelnut shell frag-
ments. The purpose of pits [214] and [324] remains
uncertain although a use associated with storage or
other activities that could be undertaken next to the
hut edge seems most likely.
was an irregular layer of yellow-brown sand with a
maximum extent of 1.7m by 1.2m and a maximum
depth of 0.1m. Otherwise it was the same as layer
[210]. This deposit also contained charred hazelnut
shell and flint, the latter including vertically set
examples. Layer [111] partly overlay the edge of [118]
and was itself partly overlain by what is interpreted
as a hazelnut-roasting pit [109] (see below). This small
levelling layer was also discerned on account of its
colour (medium brown). However, as with [210] and
[118], it also contained vertically set flints and
occasional charred hazelnut shell. The upper lenses
of these levelling layers have undoubtedly accrued
occupation debris as a result of activities taking place
within the hut.
Phase 2 Hut (see Figs. 5.13–5.16)
The Phase 2 hut was identified on the basis of two
defining characteristics. First, all the Phase 1 features
had been covered by layer [210] that extended across
the full area of the hut floor (Fig. 5.9). As this material
covered all earlier deposits, including the post sockets
and stake holes, the preceding structure must have
been dismantled and this layer of material spread
over it, probably as a levelling layer. Second, a new
arc of stake holes was inserted around the edge of the
structure, cut into layer [210], indicating the con-
struction of a new superstructure.
Levelling and Occupation Layers
Levelling layer [210] extends over the full area of the
hut floor, which in this phase was slightly smaller
than the Phase 1 hut, with a maximum diameter of
5.9m and an average depth of 0.1m. The layer
consisted of pale yellow-brown sand with occasional
patches of in-situ burning near to the centre of the
structure. The deposit was exceptionally rich in flint,
though it also contained hazelnut shell fragments and
charcoal. Of particular note, however, was the fact
that a few of the flints encountered in this layer were
vertically set. As most of the pieces are from a blade-
based industry they usually fall flat if dropped on to
a surface. In order to explain these vertical settings it
is considered most likely that at least some of the
sandy deposit [210] was tipped into the hut floor,
levelled and compacted (see Chapter 7 for further
discussion).
Layer [118] was set within levelling layer [210] but
was distinguished by its slightly different colour. It
Post Sockets and Stake Holes
An arc of stake holes that includes [260], [262], [180],
[182] and [184] was located along the south edge of
the hut. These stake holes were generally small,
ranging from 0.035m–0.06m diameter and 0.02m–
0.045m deep. They were either set vertically or
inclined slightly towards the centre of the structure.
A post socket, [113], was positioned immediately next
to this group. It measured 0.4m in diameter and
0.02m deep and was filled by a dark to medium grey
silty sand that contained occasional flints, charred
hazelnut shell fragments and charcoal flecks.
Stake holes [180], [182] and [184] were filled with
dark brown loose sand while stake holes [260] and
[262] contained black, charred, degraded wood that
had decayed into a sand matrix. Set inside the group
of stake holes [180], [182] and [184] was post socket
[224] that had been inserted into the fill of a pit [206].
This post socket measured 0.17m by 0.11m across
and was 0.24m deep. It was filled with a distinct grey
sandy fill with occasional small stones, flints and
charred hazelnut shell. A group of three stake holes,
[169], [170] and [171], was located along the north
edge of the hut. These all had a diameter of 0.05m
and were between 0.04m and 0.05m deep. All had a
dark brown-grey fill of charred wood that had also
decayed into a sand matrix. They were either
vertically set or inclined slightly inwards.
A group of four stake holes, [194], [196], [198] and
[200], was located close to the centre of the hut around
a high concentration of hazelnut shells and patches of
in situ burning. These circular stake holes, averaging
0.05m in diameter and 0.04m in depth, contained the
remains of charred stakes in their fill, although this
wood had decayed leaving a sandy texture. The
stakes appear to have formed some kind of small
internal structure and, given their position in the
centre of the hut next to the central burning area and
charred food debris, it is considered likely that they
may have formed some kind of cooking or smoking
facility.
50
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14 C Dates bp
(uncal.)
Hearths
340 Large burnt spread,
overlain by 264
3.75 x 2.8 0.1 Various,
scorched ginger
sand, black and
dark grey burnt
organics
Scorched sand and
charred organics
682 flints, 0ver
15000 charred
hazelnut shell
fragments,
occasional burnt
bone and charcoal
8802±38 OxA-11804
8739±39 OxA-12325
264 Large burnt spread 2.7 x 2.2 0.1 Black and dark
grey burnt
material and red-
ginger sand
Compacted burnt
charcoal and burnt
sand
364 flints, over
11000 charred
hazelnut shell
fragments,
occasional burnt
bone and charcoal, 2
bevelled pebble tool
blanks
320 Large burnt spread 2.54 x 2.5 0.1 Black and dark
grey burnt
material and red-
ginger sand
Burnt silty sand 78 flints, over 2700
charred hazelnut
shell fragments and
charcoal, beveled
pebble tool, a
limestone possible
smoothing stone
268 Large hearth 1.3 x 0.61 0.22 Black and dark
grey burnt
material and red-
ginger sand
Burnt silty sand 259 flints, over 8000
charred hazelnut
shell fragments,
occasional burnt
bone and charcoal
8710±45 OxA-11854
8650±45 OxA-11855
283 Linear burnt deposit
in-situ
0.68 x 0.13 0.08 Black Mineralised
charred wood in
silty sand
36 flints, c.1000
charred hazelnut
shell fragments
293 Hearth area next to
268
0.5 x 0.38 0.20 Black burnt
organics and
ginger brown
scorched sand
Fused, hard, sand 17 flints, c.850
charred hazelnut
shell fragments
8890±45 OxA-11829
(residual from earlier
activity)
8785±45 OxA-11828
Pits
285 Large irregular pit 1.0 x 0.62 0.12 Dark to medium
grey
Silty sand and ash 51 flints, c.4000
charred hazelnut
shell fragments, a
sandstone cobble
214 Small sausage-
shaped pit
0.54 x 0.22 0.06 Yellow-brown Sand Charred hazelnut
shell
324 Sallow ovoid pit 0.7 x 0.46 0.04 Orange-brown Sand 11 flints, c.70
charred hazelnut
shell fragments,
charcoal flecks
Post
Sockets
238 Circular stakehole 0.06 x 0.05 0.03 Black, charred Charred organic
material
310 Circular post socket 0.28 x 0.22 0.1 Medium brown Silty sand 5 flints, c.100
charred hazelnut
shell fragments and
birch
307 Circular stakehole 0.06 x 0.05 0.06 Medium brown Silty sand
332 Circular stakehole 0.06 diameter 0.04 Dark grey Silty sand and
charred wood
14 charred hazelnut
shell fragments
334 Circular stakehole 0.08 x 0.05 0.12 Dark grey Silty sand and
charred wood
10 charred hazelnut
shell fragments
336 Circular stakehole 0.10 x 0.06 0.12 Dark grey Silty sand and
charred wood
15 charred hazelnut
shell fragments, 1
flint
338 Circular stakehole 0.05 x 0.06 0.05 Dark grey Silty sand and
charred wood
289 Circular stakehole fill 0.04 diameter 0.02 Dark
brown/black
Loose sand
287 Circular stakehole fill 0.055 diameter 0.025 Dark
brown/black
Loose sand
300 Circular stakehole fill 0.06 diameter 0.03 Dark brown Loose sand
Stone
pads
385 Ovoid flat stone 0.18 x 0.14 0.035 Yellow brown Sandstone slab
306 Sub-rectangular flat
stone
0.18 x 0.16 0.04 Yellow brown Sandstone slab
386 Sub-rectangular flat
stone
0.09 x 0.08 0.02 Yellow brown Sandstone slab
309 Sub-circular flat
stone
0.12 x 0.08 0.03 Yellow brown Sandstone slab
5EXCAVATION OF THE HOWICK HUT 51
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14 C Dates bp
(uncal.)
Irregular
linear
Feature
298 Long ‘r’ shaped
linear
2.34 x 0.45 0.03 Medium brown Silty sand c.1200 charred
hazelnut shell
fragments, 61 flints,
charcoal
Occupation
Deposits
326 Sand spread 2.7 x 1.78 0.05 Medium brown Sand Charred hazelnut
shell fragments, 5
flints, 3 bevelled
pebble tools and a
bvelled pebble blank
316 Sand spread 3.0 x 1.0 0.04 Light brown Sand Over 2000 charred
hazelnut shell
fragments, 325 flints,
shell, charcoal, 5
small pieces copper
mineral in
sedimentary rock, a
sandstone cobble, a
sandstone faceted
pebble, abevelled
pebble tool and a
beveled pebble tool
blank
232 Sand spread 0.05 Medium brown Sand 2548 flints, a faceted
sandstone pebble,
c.30000 charred
hazelnut shell
fragments
Table 5.3 Phase 1b Context Descriptions
Figure 5.13. Plan of the Phase 2 hut.
52
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.14. Phase 2 feature sections.
A second group of five stake holes, [212], [216],
[218], [220] and [222], was positioned on the west
side of the central burning area in a linear configur-
ation. These circular stake holes averaged 0.07m
across and 0.04m deep and contained the remains of
charred stakes, though this dark brown residue had
decayed leaving only a sandy matrix. As with the
previous stake hole concentration, this group appears
to have formed part of an internal structure possibly
associated with cooking activities in the centre of the
hut. A single isolated stake hole [190] was recorded
0.7m to the south of the previous group of stake
holes. It measured 0.04m in diameter and 0.03m deep
and contained a dark brown loose sand fill. It may be
related to the linear group of stake holes, and if so
could suggest some kind of internal partition.
Hearths
Although there had clearly been a succession of open
fires in the central burning area that no doubt served
as cooking hearths on occasions, only one discrete
hearth pit, [158], was identified in the Phase 2 hut. It
was sub-circular in shape and located in the north-
5EXCAVATION OF THE HOWICK HUT 53

Figure 5.15. The Phase 2 hut during excavation.

Figure 5.16. The fused lining of the hazelnut roasting pit 109.
east part of the hut cut into sandy levelling layer
[118]. It measured 0.48m in diameter with a depth of
0.17m. The hearth fill consisted of black charred
material and a fire-reddened silty sand that contained
flints and charred hazelnut shell. The surrounding
scorched sand again indicated in situ burning. Two
fragments of in situ single-entity charred hazelnut
shell from this hearth were radiocarbon dated (see
Table 5.4 and Chapter 6). Within the central burning
area, discrete patches of burning could be identified
although they could not be accurately delimited.
Nonetheless in situ burnt residues adhering to
scorched sand could be identified and single-entity
charred hazelnut shells from two of these patches
were submitted for radiocarbon dating (see Table 5.4
and Chapter 6).
Pits
Six pits of varying shape and size were located in the
54
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14C Dates bp
(uncal.)
Hearths
158 Circular burning pit
containing charred
hazelnuts but cut by
later post pipe
0.48 diameter 0.17 Black charred
material
Silty sand,
scorched
5 flints, charred
hazelnut shell
8780±45 OxA-11832
8790±45 OxA-11853
Central
burning
area
forming
part of
210
Central burning area
with individual areas
of burning too
indistinct to
accurately delimit
2.6 x 1.65 0.1 Black, grey and
red-brown
scorched and
charred material
Silty sand See 210 8739±39 OxA-12324
8710±38 OxA-12347
8355±39 OxA-12323
8310±40 OxA-12322
Pits
160/276 A truncated linear pit
on E side of structure
2.4 x 0.6 0.21 Dark brown-
black, burnt
material
Silty sand 150 flints, c. 2300
charred hazelnut
shell fragments, 2
beveled pebble tools
206 Sib-circular steep-
sided pit
0.58 x 0.48 0.24 Yellow brown Sand 126 flints, c.1400
charred hazelnut
shell fragments,
small burnt bone
fragments, charcoal
flecks
192 Irregular pit
containing much
charred hazelnut,
possibly extension of
109
0.75 x 0.70 0.12 Black grey
charred material
Sandy silt 23 flints, c.200
charred hazelnut
shell fragments
214 Small oblong flat-
bottomed pit
0.22 x 0.11 0.055 Yellow brown Sand 2 flints, c. 40 charred
hazelnut fragments
155 Small, shallow, sub-
circular pit
0.36 x 0.39 0.05 Light yellow
brown
Loose sand 4 flints
313 Small irregular pit
that may not be
archaeological
0.15 x 0.17 0.04 Dark
brown/black
Sandy silt loam
Roasting
Pit
109/236 Hazelnut roasting pit
lining
1.8 x 1.37 0.03 Burnt red-brown Heat affected sand
and charred
material
Over 7100 Charred
hazelnut shell
fragments, 309 flints
8715±50 OxA-11830
8715±45 OxA-11831
107 Fill of roasting pit
109
1.8 x 1.37 0.03 Brown grey Silty sand and ash Over 2000 Charred
hazelnut shell
fragments, 127 flints
and charcoal
Post
Sockets
113 Circular post socket
fill
0.4 diameter 0.2 Dark – medium
grey
Loose silty sand 24 flints, 113 charred
hazelnut shell
fragments, charcoal
flecks
169 Circular stakehole fill 0.05 diameter 0.05 Dark brown-grey Charred wood
degraded into sand
matrix
170 Circular stakehole fill 0.05 diameter 0.04 Dark brown-grey Charred wood
degraded into sand
matrix
171 Circular stakehole fill 0.05 diameter 0.05 Dark brown-grey Charred wood
degraded into sand
matrix
180 Circular stakehole fill 0.04 x 0.045 0.02 Dark brown Loose sand
182 Circular stakehole fill 0.045 diameter 0.02 Dark brown Loose sand
184 Circular stakehole fill 0.035 diameter 0.025 Dark brown Loose sand
190 Circular stakehole fill 0.04 diameter 0.03 Dark brown Loose sand
194 Circular stakehole fill 0.05 diameter 0.045 Black charred Burnt organics and
sand
196 Circular stakehole fill 0.06 diameter 0.04 Black charred Burnt organics and
sand
198 Circular stakehole fill 0.04 diameter 0.04 Black charred Burnt organics and
sand
200 Circular stakehole fill 0.06 diameter 0.045 Black charred Burnt organics and
sand
212 Circular stakehole fill 0.04 x 0.07 0.04 Dark brown-grey Charred wood
degraded into sand
matrix
216 Circular stakehole
fill, inclined at c.65°
0.07 x 0.05 0.03 Dark brown-grey Charred wood
degraded into sand
matrix
218 Circular stakehole fill 0.09 x 0.07 0.12 Dark brown-grey Charred wood
degraded into sand
matrix
5EXCAVATION OF THE HOWICK HUT 55
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14C Dates bp
(uncal.)
220 Circular stakehole fill 0.08 x 0.07 0.04 Dark brown-grey Charred wood
degraded into sand
matrix
222 Circular stakehole fill 0.13 x 0.09 0.04 Dark brown-grey Charred wood
degraded into sand
matrix
224 Post socket fill 0.17 x 0.11 0.24 Grey Sand with
occasional stones
4 flints, c. 30 charred
hazelnut shell
fragments
260 Circular stakehole fill 0.035 diameter 0.035 Black Charred wood
degraded into sand
matrix
262 Circular stakehole fill 0.06 x 0.04 0.045 Black Charred wood
degraded into sand
matrix
112 flints, c.1500
charred hazelnut
shell fragments
Clay
Lumps
208 Irregular clay spread 0.48 x 0.30 0.14 Pink brown Clay lump
(imported into the
hut)
209 Irregular clay spread 0.33 x 0.18 0.10 Pink brown Clay lump
2 flints, 8 hazelnut
shell fragments
(imported into the
hut)
Levelling/
Occupati
on Layers
210/115/1
61/244
Sand occupation
deposit into which all
other features in this
phase are cut or
overlie
Full width of
structure (5.9)
0.1 Pale yellow
brown
Sand 1280 flints, much of
it set vertically as
though tipped in,
c.17500 charred
hazelnut shell
fragments, charcoal,
2 bevelled pebble
tools, 1 bevelled
pebble blank and a
sandstone fragment
118/159/2
34
Irregular sand deposit 1.7 x 1.2 0.1 Yellow brown Sand spread 610 flints, c.4000
charred hazelnut
shell fragments,
beveled pebble tool
blank
111/259 Irregular sandy
deposit
0.75 x 0.65 0.1 Medium brown Sand spread 38 flints, c. 100
charred hazelnut
shell fragments
Table 5.4 Phase 2 Context Descriptions
Phase 2 deposits, all cut into levelling layer [210]. Pit
[160] was the largest, located on the east edge of the
surviving structure where part of its east side had
been truncated by the cliff edge erosion. The pit was
linear in shape, measured 2.4m long and had a
maximum depth 0.17m. The profile was unusual as
its west side was near-vertical while its east side
sloped more gently. The fill consisted of dark brown-
black silty sand with much burnt material, and
contained flints, charred hazelnut shell and an
elongated pebble tool. Pit [206] was located in the
southern area of the hut, also cut into layer [210], and
was sub-circular in shape. The pit measured 0.58m
by 0.48m and was 0.24m deep. It had a curved profile
and the fill consisted of yellow-brown sand that
contained flints, charred hazelnut shell and oc-
casional fragments of burnt bone and charcoal. A
later post socket, [224], had been inserted into the fill
of this pit (see above). An irregular pit, [192], was
located near to the west edge of the hut, cut into layer
[210]. This pit was filled by a black and grey charred
sandy fill that contained a large quantity of charred
hazelnut shell, together with flint, and may be related
to what is thought to be a hazelnut-roasting pit [109]
(see below). The pit measured 0.54m by 0.22m by
0.055m deep. Pit [214] was a small oblong feature
located in the approximate centre of the hut. This was
a flat-bottomed pit that measured 0.22m by 0.11m by
0.055m deep. The pit was filled with sterile yellow-
brown sand that did not contain any small finds. Pit
[155] was a shallow circular pit located next to layer
[118], 0.3m from hearth pit [158]. The pit had a
maximum diameter of 0.39m with a shallow dish
profile and a maximum depth of 0.05m. It was filled
with light yellow-brown loose sand that contained
the occasional flint. The final pit, [313], was a small
irregular feature that is likely to be a modern
intrusion on account of it being located next to a
sheep burial on the truncated edge of the hut and
having a black sandy loam fill identical to the modern
plough soil. The fill and edges of this feature were
not weathered like the other archaeological deposits
56
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
and it contained no small finds. It probably formed
part of the old fence line that ran along the cliff edge.
structure. These substantial timbers indicate that this
final-phase hut was of a different constructional form
to the previous two structures, with vertical set
timbers placed immediately inside the outer edge. It
also had one, or perhaps more, central posts, though
whether these were for roof support or for a fireside
feature remains equivocal. This structural evidence
indicates another very robust timber-framed struc-
ture but based on a different structural template to
the two previous huts.
Roasting Pit
An irregular heat-affected dished feature, [109], was
located in the north-west area of the hut, abutting one
of the edges (Figs. 5.13, 5.14 and 5.16). This shallow
feature, measuring 1.8m by 1.37m and 0.03m deep,
was filled by an extremely high density of charred
hazelnut shell (see also Chapter 11) and heat-affected
sand [107]. However, there was no evidence for in
situ burning as there was no black charred debris
other than the hazelnut shell. The heat-fused lining of
pit [109] was scorched rather than burnt and had
uncountable quantities of charred hazelnut shell
fragments incorporated into its surface that must
have numbered many tens of thousands. In addition
to the mass of hazelnut shells, the fill also contained
occasional flint and charcoal flecks. The clear evi-
dence for in situ heating of the pit surface, but with
no clear sign of actual burning, together with the high
concentration of charred hazelnut debris within it,
suggests this was a pit used for the roasting of
hazelnuts, presumably to preserve them for later
consumption (see also Chapter 11). Two samples of
single-entity hazelnut shell from the heat-affected
surface of the pit lining were submitted for radio-
carbon dating (see Table 5.4 and Chapter 6).
Clay Lumps
Two discrete clay lumps, [208] and [209], were found
in the Phase 2 occupation horizon overlying layer
[210]. Both were irregular in shape and consisted of
pure clay with no inclusions or associated small finds
or structural features. Lump [208] measured 0.48m
by 0.3m and was 0.14m thick while lump [209]
measured 0.33m by 0.18m by 0.1m thick. Both were a
pink brown colour, suggestive of a common source,
and had evidently been brought into the hut. Glacial
clay can be easily obtained from nearby cliff and river
sections. The use and implications of these clay
lumps, as yet, remain uncertain.
Levelling and Occupation Deposit
Layer [49] extended over the full area of the hut floor
covering all the Phase 2 deposits. This yellow-brown
fine sand layer averaged 0.1m thick and contained a
range of features cut into or laid on top of it. The
layer itself was not entirely homogeneous and it is
thought that these variations reflect different dumps
of material as part of a levelling layer spread over the
previous hut deposits once the Phase 2 structure had
been dismantled. However, as with layer [210], the
upper lenses of this layer are likely to have accum-
ulated in situ as a result of the build-up of occupation
debris. Some burnt patches towards the centre of the
layer were more discrete and were indicative of a
central burning/hearth area. The central area of [49]
was noted for its higher proportion of charred
material, particularly around hearth features [47] and
[173]. As with layer [210] this levelling/occupation
horizon [49] was very rich in material, producing
high quantities of flints and charred hazelnut shell, as
well as occasional fragments of mollusc shell, red
ochre and some burnt bone fragments. In addition,
the flints encountered in this layer were vertically set,
indicating that, like [210], much of this deposit was
probably tipped in, perhaps to level off and cover the
previous hut floor.
Phase 3 Hut (see Figs. 5.17–5.20)
The Phase 3 hut was constructed directly over the
previous huts although it was less circular in plan
and tended to have straighter sides. The Phase 3 hut
was not quite as large as the previous huts, having an
internal width of 5.4m compared with 6m and 5.9m
for the earlier two. Furthermore, when this final
structure was ultimately abandoned, two of the fallen
timber uprights survived in the form of stains that
could be seen on the south and west sides of the
Post Sockets
The post sockets occurred in two distinct groups:
there was a group of three positioned in the central
area of the hut, [148], [153] and [67], and the other
four, [51], [35], [33] and [267], were positioned around
the hut edge. The latter group would become five if
probable post socket [106] is included (see below).
The central post sockets were all of similar size, shape
and form, ranging from 0.19m to 0.23m in diameter
and 0.1m to 0.14m deep. They were all of broadly
circular shape and were vertically set. All the sockets
were filled with dark brown silty sand that contained
occasional flints and charred hazelnut shells. Some
packing stones were also noted in pit [148]. The pits
form a triangular arrangement to the east of the
hearth area and are certainly for substantial timbers
that could have formed structural supports for some
kind of cooking/smoking/drying structure. It is also
5EXCAVATION OF THE HOWICK HUT 57

Figure 5.17. Plan of the Phase 3 hut.
possible that they formed structural supports for the
hut superstructure; however, whether they were used
contemporaneously as a triangular arrangement or
whether they represent subsequent replacements of a
central post remains entirely hypothetical.
On the north-east side of the structure three sockets
were identified that were all thought to have been
structural. Socket [267] was clearly seen in the baulk
section during excavation where it had cut down into
earlier deposits. This vertically set circular socket had
held a post of 0.17m in diameter and had been cut
down 0.15m. It was filled by light grey-brown silty
sand. Post socket [33] was slightly larger, measuring
0.2m in diameter and had been cut down 0.16m. This
socket was also circular and had been vertically set. It
had a similar dark brown silty sand fill that contained
occasional flints, charred hazelnut shell and some red
ochre. Socket [35] was more elongated than the
previous two, measuring 0.7m by 0.37m by 0.2m deep.
This was also filled with dark brown silty sand that
contained occasional flints and charred hazelnut shell.
Although no obvious post pipe could be identified in
this socket, its location on the edge of the structure
close to [33] implies that it had a structural purpose.
Post hole [51] was a substantial feature clearly
observed in plan and section, measuring 0.29m in
diameter and 0.31m deep. It was filled by a distinctive
dark brown silty sand that contained occasional flints
and charred hazelnut shell. The post hole was
vertically set and had vertical sides. A single-entity
charred hazelnut shell fragment from within the fill
of this post hole was submitted for radiocarbon
dating after the initial evaluation in 2000 in order to
obtain a provisional date for the structure (see
Chapter 6). This post hole was located next to timber
stain [29] and its associated stone post pad or wedge
[387] (see below).
Timber Stains
Two stains of unsplit timbers were evident in the Phase
3 horizon (Figs. 5.17 and 5.20). Timber [41] was
situated inside the south edge of the hut and measured
2.3m long by 0.34m wide with a surviving depth of
58
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.18. Phase 3 feature sections.
0.14m. The profile of the timber was rounded and had
a consistent curvature along its length, suggesting that
it was an undressed segment of tree trunk. The organic
wood component had decayed to leave a dark brown
stained silty sand matrix. Occasional flints, charred
hazelnut shell fragments and charcoal were found
within this matrix; they are thought to have worked
into the sand as the wood degraded, as a result of
earthworm activity and other processes of bio-
turbation. It is possible that this timber formed a
revetment against the south edge of the hut, but as it
does not lie flush with the edge at its west end this is
unlikely. It is more likely that this was an upright that
fell to the floor when the site was abandoned. This
5EXCAVATION OF THE HOWICK HUT 59

Figure 5.19. The Phase 3 hut during excavation.
gains support from the fact that a flat-bottomed pit,
[106], lies at the east end of the timber, forming its
post-socket. A second timber stain, [29], was located
on the west side of the hut. At the east end of the
timber, immediately inside the hut edge, was a flat
stone which is thought to have been a post pad, but
bearing in mind the experience gained from building
the second reconstruction (see Chapter 13), this is
interpreted as a wedge inserted beneath the post to
raise it up and provide stability. Again the timber has
fallen towards the west, which is consistent with the
direction of the prevailing wind. Immediately next to
the post pad, also situated immediately inside the hut
edge, was a substantial post hole [51]. The post that
had sat in the post hole was at some time replaced by
the timber that sat on the post pad. This suggests that
the Phase 3 hut was probably repaired before it was
finally abandoned. Timber stain [29] measured 1.6m
long by 0.36m wide, though as it had been more
severely truncated than timber [41] during the initial
subsoil stripping, it only survived to a maximum
thickness of 0.05m; its inconsistent shape is also a
result of this. As this timber had lain above the hut
deposits and must have lain partly on the Mesolithic
land surface it is fortunate that any trace survives of
it. Timber [29] had a dark brown sand fill containing
the occasional flint that is thought to have become
worked in as the timber decayed. The stone pad, [387],
comprised a flat yellow-brown sandstone slab that
measured 0.22m by 0.18m by 0.04m thick. Both
timbers and the stone pad lay directly on occupation
deposit [49].

Figure 5.20. Close-up view of one of the truncated timber stains
[29] from Phase 3 with the post pad at its base and post hole [51]
next to it.
60
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds 14C Dates bp
(uncal.)
Hearths
47 Hearth near centre of
structure
0.49 x 0.29 0.25 Black, burnt Heat fused sand
and charred
organic material
34 flints, c.900
charred hazelnut
shell fragments,
burnt bone
fragments, charcoal,
ockre
8555±60 AA-41788
8324±37 OxA-11805
8690±40 OxA-12294
173 Small burning pit 0.26 x 0.19 0.035 Black, burnt Heat fused sand
and charred
organic material
150 charred hazelnut
shell fragments,
charcoal
8630±40 OxA-11826
8700±45 OxA-11827
Pits
45 Large pit next to
hearth area
1.15 x 0.84 0.23 Ginger brown Heat affected sand
and a degraded
charred log
75 flints, c.1650
charred hazelnut
shell fragments,
charcoal
53 Pit 0.50 x 0.43 0.21 Dark brown Silty sand 31 flints, c.450
charred hazelnut
shell fragments
65 Pit 1.04 x 0.80 0.17 Grey brown Silty sand 49 flints, c.230
charred hazelnut
shell fragments
106 Pit 0.73 x 0.5 0.08 Dark brown Sand 107 flints, over 500
charred hazelnut
shell fragments,
charcoal
Post sockets
33 Post socket at edge of
structure
0.20 diameter 0.16 Dark brown Silty sand 10 flints, charred
hazelnut shell, ochre
35 Structural socket on
edge of structure
0.7 x 0.37 0.2 Dark brown Silty sand 23 flints, c.150
charred hazelnut
shell
51 Post hole at edge of
structure
0.31 diameter 0.29 Dark brown Silty sand 13 flints, 62 charred
hazelnut shell
fragments, a beveled
pebble tool blank
8730±40 Beta-
153650
This date probably
on residual material
67 Post socket near
centre of structure
0.19 diameter 0.14 Dark brown Silty sand 9 flints, 10 charred
hazelnut shell
fragments, mollusc
shell
148 Post socket near
centre of structure
0.23 0.1 Dark brown Silty sand Three packing
stones, 6 flints, 60
charred hazelnut
shell fragments,
charcoal
153 Post socket near
centre of structure
0.22 diameter 0.13 Dark brown Silty sand
267 Post hole 0.17 diameter 0.15 Grey brown Silty sand 71 flints, c.450
charred hazelnut
shell fragments
Timbers
29 Linear stain of split
timber, rectangular
profile
1.6 x 0.36 0.05 Dark
brown/black
Silty sand 1 flint and a
sandstone chocking
pad
387 Stone pad for timber
29
0.22 x 0.18 0.04 Yellow brown Flat sandstone slab
41 Linear stain of
timber, rounded
profile indicating a
tree trunk
2.3 x 0.34 0.14 Dark brown Silty sand 277 flints, over 2000
charred hazelnut
shell fragments,
charcoal
Leveling/
Occupation
Deposit
49 All phase A features
cut into this
occupation layer
containing vertically
set flints
Full width of
structure 5.4
0.10 Yellow brown Fine silty sand 3573 flints, over
44000 charred
hazelnut shell
fragments, burnt
bone fragments,
ochre fragment,
mollusc shell, small
burnt stones, 2
bevelled pebble
tools, 2 bevelled
pebble blanks, a
sandstone
hammerstone and a
sandstone fragment
Table 5.5 Phase 3 Context Descriptions
5EXCAVATION OF THE HOWICK HUT 61
Context Decsription Max
Dimensions
(m.)
Max
depth
Colour of fill Texture of fill Small Finds C14 Dates bp
(uncal.)
Hearths
63 Small burning pit 0.31 x 0.10 0.09 Dark red-brown Heat fused sand
and charred
organic material
4 flints, charred
hazelnut shell,
charcoal
8278±35 OxA-11806
8233±36 OxA-11807
55 Small burning pit 0.40 x 0.12 0.18 Dark-
brown/black,
burnt
Heat fused sand
and charred
organic material
9 flints, c.170
charred hazelnut
shell fragments,
charcoal
8330±45 OxA-12408
8280±40 OxA-12293
210 Patch
Small burnt patch 0.2m diameter ? Dark-
1
brown/black,
burnt
Black charred
organic material
Table 5.6 Phase 4 Context Descriptions
8355±39 OxA-12323
8310±40 OxA-12322
Context Decsription Max
Dimensions
(m.)
Max
depth
(m.)
Colour of fill Texture of fill Small Finds
13 Subcircular pit 1.1 x 0.4 0.14 Light brown Silty sand Flint, charcoal
21 Subcircular pit 0.7 x 0.52 0.38 Red brown Silty sand Bone, iron, bevelled
tool, flint
61 Circular pit 0.49 x ? 0.18 Dark brown Silty sand Flint, shell, charcoal
Table 5.7 Context Descriptions of features to the south of the Mesolithic hut
Hearths
Two hearth-type features were located in the central
area of the hut, cut into deposit [49]. The largest
hearth pit, [47], had an irregular elongated shape and
measured 0.49m by 0.29m with a maximum depth of
0.25m. It was filled with black charred debris and
scorched sand, with the sand around its edges fire-
reddened and fused as a result of in situ burning. The
fill contained flints, charred hazelnut shell, burnt
bone fragments, charcoal and red ochre. Two single-
entity samples of charred hazelnut shell were radio-
carbon dated (see Table 5.5 and Chapter 6). Hearth
pit [173] was located 0.4m to the east of hearth [47],
cut into layer [49], measuring 0.26m by 0.19m with a
depth of 0.035m. The fill consisted of black burnt
material surrounded by heat-fused sand, again
indicating in situ burning. Within the fill were flint,
charred hazelnut shell and charcoal. Two single-entity
charred hazelnut shell fragments were submitted for
radiocarbon dating (see Table 5.5 and Chapter 6).
These hearths were surrounded by an abundance of
charred material in occupation deposit [49] that had
probably accumulated as a result of open fires and
fireside activities.
Pits
Pits [45] and [53] were located to the south and south-
east of the central hearth area. Pit [45] was sub-
circular in shape with a bowl-shaped profile, meas-
ured 1.15m by 0.84m and had a maximum depth of
0.23m. The fill was a distinctive ginger brown heat-
affected sand that contained a considerable quantity
of flint and charred hazelnut fragments. A small
charred log was found at the base of the pit but the
surrounding sandy deposits did not appear to be
heat-affected, and so it is unlikely that this piece of
wood was burnt in situ. It was this pit that was
encountered in the first exploratory test pit on the
site in 2000. Pit [53], measuring 0.5m by 0.43m and
0.21m deep, had an irregular shape and a stepped
profile. It was filled by dark brown silty sand that
contained flints and charred hazelnut shell. It is likely
that these pits were used for some type of hearth-
associated activity as they contained much charred
debris, charcoal and heat-affected sand in their fills,
though evidence for in situ burning in these pits
remained equivocal. They may have served as pit
ovens or as receptacles for raked ashes.
Two further pits were noted in this phase of
occupation. Pits [65] and [106] were located in the
south corner of the structure, next to the truncated
edge. Pit [65] had a slightly irregular shape with a
shallow flat base and measured 1.04m by 0.8m by
0.17m deep. It was filled with grey-brown silty sand
that contained flints and charred hazelnut shell. No
in situ burning was evident and the use of this pit
remains indeterminate. Pit [106] was ovoid in shape
with vertical edges and was situated on the edge of
the hut at the foot of the timber stain [41]. It measured
0.73m by 0.5m and was 0.08m deep. It was filled by
dark brown sand that contained occasional flints,
charred hazelnut shell and charcoal. It is thought
likely that this was a post socket for the timber
represented by stain [41].
62
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 5.21. Plan of the Phase 4 deposits together with sections.
Phase 4 Hut Re-Occupation (see Fig. 5.21)
This phase of re-use of the hut site was only recognised
as a result of the radiocarbon dating programme and
could not otherwise have been identified from the
surviving site stratigraphy. All the Phase 3 and Phase
4 deposits lay immediately below the same subsoil
horizon and did not share any horizontal stratigraphic
associations, which meant that a relative stratigraphy
could not be observed.
The features associated with this phase of use
include a small burnt ovoid feature, [63], measuring
0.31m by 0.1m and 0.09m deep, which was located
next to hearth [47] that is dated to Phase 3. Although
too small to be characterised as a hearth, this feature
had evidently experienced burning in situ as it had a
dark charred and fire-reddened fill and was sur-
rounded by heat-fused sand. This small burning pit
also contained flints, charred hazelnut shell and
charcoal. Two samples of single-entity charred
hazelnut shell were radiocarbon dated (see Table 5.6
and Chapter 6). A second small burning pit, [55], was
located to the south-west of hearth [47], forming a
slightly arcing pit. It measured 0.4m by 0.12m and
was 0.18m deep. It was filled with charred, dark
brown and black organic material and heat-affected
sand, with the surrounding sand fire-reddened and
fused by in situ burning. The fill also contained flints,
charred hazelnut shell and charcoal. A discrete patch
of in situ burning identified in the burnt central area,
cut into levelling layer 210 and referred to as [210]
burnt patch 1, has also been dated to this phase,
having clearly been cut down through the earlier
deposits during the Phase 4 activity at the site.
Likewise an intrusive charred hazelnut shell of Phase
4 was dated from hearth [47], in the Phase 3 deposits.
This could easily have been worked into the adjacent
Phase 3 feature either as a result of bioturbation or
ground disturbance during the Phase 4 occupation.
No other features could be attributed to this phase.
5EXCAVATION OF THE HOWICK HUT 63

Figure 5.22. Phase 4 feature sections and plans and sections of external features outside the Mesolithic hut.
Features and Activity Outside the Mesolithic
Hut
A series of pit and scoop features was identified
which clustered at the south-west end of the trench
(Fig. 5.1). The following table summarises the
characteristics of each of these features.
Pits [11] and [15] could be modern pits/scoops
associated with the sheep burials, as they contained
fragments of brick and iron objects respectively. Pit
[27] was extremely shallow and was so heavily
truncated that it was not possible to confidently
describe this feature as being prehistoric in date.
However, the other pit features, [13], [21] and [61],
are considered to be archaeological on the basis of
their shape, form, the texture of the fill and the
weathered nature of the pit sides as well as their
contents.
The function of pit features [13], [21] and [61] can
only be guessed at but the presence of flints, charcoal
and shell in [61] and of a coarse stone tool and flint in
[21] is consistent with domestic/processing activities.
It is possible that the contents of these pits are residual
material from the Mesolithic ground surface, but the
absence of the modern materials found in the sheep
burial pits favours the view that they are more likely
to be contemporary with the occupation of the
Mesolithic hut. These pits are located downwind
from the main structure and could, therefore, be
rubbish pits or pits associated with outdoor activities.
A pit with a clay-lined base, [9], was located in the
south corner of the trench and had been truncated on
three of its four sides. It survived as an irregular
outline in plan, although its intact side suggested it
had originally been of sub-circular shape. It survived
to a maximum width of 0.9m and had a maximum
thickness of 0.11m. The feature consisted of a shallow
pit cut into the natural sand [5] that had up to 0.07m
thickness of clay spread across its base. The base of
64
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
the pit was slightly convex so that its centre projected
as a low dome. The clay was blue-grey with some
fire-reddening on the interface with the sand fill, and
was flecked throughout with charcoal. The sand lying
on the clay was reddened suggesting in situ heating.
Situated within the clay lining a thin spread (0.04m
thick) of fire-reddened sand survived, representing
presumably the final contents of the pit before it went
out of use. A number of small shallow holes were
observed across the clay lining of the pit. On
excavation these gave the appearance of stake holes;
given the action of moles on the site this cannot be
stated with certainty, although the mole damage
tends to be at a higher level in the trench.
The function of this clay-lined pit remains prob-
lematic, especially as it cannot be ascertained that it
is Mesolithic. Speculation as to its purpose has
included a fish-smoking structure or an oven pit.
Sadly, the area to the south-west of the hut where
the evidence for external activity was located had
been heavily truncated and disturbed by a combin-
ation of sheep burials and old fence lines. This meant
that it was hard to disentangle true Mesolithic
features contemporary with the hut from those of
more modern origin. As a consequence, very little
can be said with any certainty concerning activities
external to the hut, other than that it was clear from
the distribution of lithics in the overburden that
processing activities requiring the use of flint tools
occurred in the immediate environs of the building.
The scorched clay in pit [9] implies some activity
associated with burning. Attempts to answer the
question of whether this hut formed part of a larger
settlement or was an isolated building are hindered
by the fact that the land to the immediate east of the
hut has eroded away, destroying any other features
that may have existed in this area. However, the
extensive size of the trench, the test-pitting and
geophysical survey all go to show that there is no
evidence for another hut to the north, south or west
of the known hut and so, on balance, it is most likely
that this is an isolated hut. The results from the
fieldwalking, however, suggest that other huts, or
areas of occupation, are located nearby and probably
comprise similarly isolated hut foci.
Conclusions
The stratigraphy of the Mesolithic hut deposits at
Howick represents one of the best-preserved se-
quences of Mesolithic activity so far discovered in
the British Isles. The careful excavation, flotation and
sampling of in situ material demonstrate the potential
of such sites to reveal complex occupation sequences
and structural forms. Despite the deposits surviving
in a fragile sand matrix that had been subjected to
considerable mole damage, and despite cliff collapse,
together with weathering of the uppermost deposits,
identification and delimitation of discrete features
was possible. The sequence recorded during the
excavation has been supported by a comprehensive
dating programme (see Chapter 6), providing an
independent test for the observed stratigraphic
sequence of occupation. Summaries of each phase are
provided in Chapter 14.
6ABSOLUTE DATING 65
6ABSOLUTE DATING
Alex Bayliss, Ian Boomer, Christopher Bronk Ramsey,
Derek Hamilton and Clive Waddington
Introduction
Fifty-nine radiocarbon measurements have been
obtained from Howick; 33 from archaeological
features and 26 from sediment cores. Fifty-five
samples of carbonised or waterlogged plant material
were dated by the Oxford Radiocarbon Accelerator
Unit between 2002 and 2004. Two further samples of
charred hazelnut shell were dated in 2001, one by the
Scottish Universities Research and Reactor Centre
(SURRC), and one by Beta Analytic Inc. Two samples
of waterlogged plant macrofossils were dated by
SURRC, also in 2001.
and the more specific Gibbs sampler (Gelfand and
Smith 1990; Gilks et al. 1996;). Details of the algo-
rithms employed by this program are available from
the OxCal on-line manual or in Bronk Ramsey (1995;
1998; 2001), and fully worked examples are given in
the series of papers by Buck et al. (1991; 1992), Buck,
Litton et al. (1994), and Buck, Christen et al. (1994).
The algorithms used in the models described below
can be derived from the structure shown in Figures
6.1–6.2*.
Replicate radiocarbon measurements on the same
sample have been combined before calibration by
taking a weighted mean, and the consistency of
groups of results which are, or may be, of the same
actual age has been tested using methods outlined by
Ward and Wilson (1978).
The following section concentrates on the archaeo-
logy – particularly on the reasoning behind the
interpretative choices made in producing the models
presented. These archaeological decisions fund-
amentally underpin the choice of statistical model.
General Approach
A Bayesian approach has been taken to the interpret-
ation of chronological data from this site (Buck et al.
1996). This is a mathematical modelling technique
that combines the radiocarbon dates with the strat-
igraphic sequence. This allows more precise dating to
be provided by determining which parts of the simple
calibrated radiocarbon dates are unlikely because of
the relative dating provided by the stratigraphic
relationships between samples. The process results in
a reduced date range, known as a posterior density
estimate (shown in black in Figs 6.1–6.2*). These
distributions are based on probability, and are shown
in italics when expressed as date ranges in the text.
The posterior density estimates are not absolute, they
are interpretative estimates, which can and will change
as further data become available and as other
researchers choose to model the existing data from
different perspectives.
The technique used for the Bayesian analysis is a
form of Markov Chain Monte Carlo sampling, and
has been applied using the program OxCal v3.5
(http://units.ox.ac.uk/departments/rlaha/), which
uses a mixture of the Metropolis-Hastings algorithm
Objectives
The principal aims of the dating programme were:
• to determine the absolute dates of the structure;
• to estimate the duration of occupation;
• to test whether the identified stratigraphic phases
represent an accumulation of in situ material that forms
a temporal sequence of separate structural episodes;
• to provide absolute dating for the associated flint
assemblage (the first dated Mesolithic assemblage from
North-East England);
• to determine whether the lower horizons of the sed-
iment core from the Howick Burn were contemporary
with the Mesolithic structure;
• to provide absolute dating for the environmental
changes shown in this sediment record.
66
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 6.1. Probability distributions of simulated dates from the Mesolithic structure at Howick: each distribution represents the relative
probability that an event occurred at some particular time. For each of the simulated radiocarbon measurements two distributions have been
plotted, one in outline, which is the result of simple radiocarbon calibration, and a solid one, which is based on the chronological model used.
The other distributions correspond to aspects of the model. For example, the distribution ‘construct_hut’ is the estimated date for the
construction of the building. The large square brackets down the left hand side along with the OxCal keywords define the overall model
exactly.
Sampling
The initial step in sample selection was to identify
non-residual short-lived material. In the case of the
Mesolithic hut deposits, this meant that sampling was
largely restricted to material that had been burnt in
situ, and so had a demonstrable functional relation-
ship with the context from which it was recovered.
All samples consisted of single fragments of carbon-
ised or waterlogged material (Ashmore 1999). Those
from the structure consisted of charred hazelnut
shells recovered from features that showed evidence
of in-situ burning, with the exception of one sample
of hazelnut shell from a post-hole fill, which was
dated in 2001 as part of the assessment of the site
(Beta-153650).
Once a pool of potentially suitable samples had
been identified, a number of models were built
simulating the results of the dating programme (e.g.
Fig 6.1). These models included the stratigraphic
order of samples and phases, and archaeological
6ABSOLUTE DATING 67

Figure 6.2. Probability distributions of dates from the Mesolithic structure at Howick during Phases 1A, 1B, 2, and 3. The format is identical
to that of Figure 6.1. Radiocarbon dates marked with a “?” have been excluded from the analysis. The large square brackets down the left
hand side along with the OxCal keywords define the overall model exactly.
Laboratory
Number
Mesolithic Structure
Sample Material and Context 13C
(‰)
Radiocarbon
Age BP
Calibrated Date (95%
confidence)
Posterior Density Estimate (95%
probability)
OxA-11801 HEX02 355 hazelnut shell from hearth [355] -25.7 8734±37 cal BC 7960-7600 cal BC 7940-7930 (2%) or cal BC
7920-7750 (93%)
OxA-11802 HEX02 357 hazelnut shell from hearth [357] -25.1 8754±38 cal BC 8160-7600 cal BC 7880-7740 (95%)
OxA-11803 HEX02 291 hazelnut shell from hearth [291] -24.0 8763±38 cal BC 8160-7610 cal BC 7890-7740 (95%)
OxA-11804 HEX02 340 hazelnut shell from hearth [340] -23.8 8802±38 cal BC 8200-7680 cal BC 7840-7730 (95%)
OxA-11805 HEX02 047 hazelnut shell from burnt feature/possible hearth [047] -23.7 8324±37 cal BC 7520-7190 -
AA-41788 HEX02 047 hazelnut shell from same context as OxA-11805 -23.4 8555±60 cal BC 7740-7480 cal BC 7750-7580 (95%)
OxA-11806 HEX02 063A hazelnut shell from burnt feature/possible hearth [063] -23.4 8278±35 cal BC 7520-7140 cal BC 7460-7310 (95%)
OxA-11807 HEX02 063B hazelnut shell from same context as OxA-11806 -26.6 8233±36 cal BC 7450-7080 cal BC 7460-7310 (95%)
OxA-11826 HEX02 173A hazelnut shell from hearth [173] -23.4 8630±40 cal BC 7750-7580 cal BC 7750-7590 (95%)
OxA-11827 HEX02 173B hazelnut shell from same context as OxA-11826 -22.6 8700±45 cal BC 7940-7590 cal BC 7740-7600 (95%)
OxA-11828 HEX02 293A hazelnut shell from hearth [293] -22.8 8785±45 cal BC 8200-7650 cal BC 7830-7720 (95%)
OxA-11829 HEX02 293B hazelnut shell from same context as OxA-11828 -23.9 8890±45 cal BC 8240-7830 -
OxA-11830 HEX02 109A hazelnut shell from roasting feature [109] -25.2 8715±50 cal BC 7960-7590 cal BC 7780-7650 (95%)
OxA-11831 HEX02 109B hazelnut shell from same context as OxA-11830 -28.2 8715±45 cal BC 7960-7590 cal BC 7780-7650 (95%)
OxA-11832 HEX02 158A hazelnut shell from hearth/hazelnut pit [158] -25.0 8780±45 cal BC 8200-7610 cal BC 7780-7650 (95%)
OxA-11853 HEX02 158B hazelnut shell from same feature as OxA-11832 -23.1 8790±45 cal BC 8200-7650 cal BC 7780-7650 (95%)
OxA-11854 HEX02 268A hazelnut shell from hearth [268] -23.1 8710±45 cal BC 7950-7590 cal BC 7810-7700 (95%)
OxA-11855 HEX02 268B hazelnut shell from same context as OxA-11854 -22.4 8650±45 cal BC 7800-7580 cal BC 7810-7700 (95%)
OxA-11856 HEX02 379A hazelnut shell from hearth [379] -26.5 8785±45 cal BC 8200-7650 cal BC 7930-7750 (95%)
OxA-11857 HEX02 379B hazelnut shell from same context as OxA-11856 -23.6 8750±45 cal BC 8160-7600 cal BC 7920-7750 (95%)
Beta-153650 HEX02 051 hazelnut shell from post hole fill [051] -26.1 8730±40 cal BC 7950-7600 -
OxA-12292 HEX02 383B hazelnut shell from hearth [383] -25.8 8785±40 cal BC 8200-7650 cal BC 7920-7750 (95%)
OxA-12293 HEX02 055B hazelnut shell from burning pit [055] -23.7 8280±40 cal BC 7520-7140 cal BC 7480-7320 (95%)
OxA-12294 HEX02 047A hazelnut shell from hearth [047] -24.4 8690±40 cal BC 7940-7590 cal BC 7740-7600 (95%)
OxA-12322 HEX02 210(1)A hazelnut shell from levelling layer [210] -23.5 8310±40 cal BC 7520-7180 cal BC 7450-7300 (93%) or cal BC
7220-7200 (2%)
OxA-12323 HEX02 210(1)B hazelnut shell from levelling layer [210] -23.8 8355±39 cal BC 7540-7200 cal BC 7450-7310 (93%) or cal BC
7220-7200 (2%)
OxA-12347 HEX02 210(2)A hazelnut shell from levelling layer [210] -23.1 8710±38 cal BC 7940-7590 cal BC 7780-7650 (95%)
OxA-12324 HEX02 210(2)B hazelnut shell from levelling layer [210] -23.0 8739±39 cal BC 7970-7600 cal BC 7780-7650 (95%)
OxA-12325 HEX02 340A hazelnut shell from hearth [340] -23.0 8739±39 cal BC 7970-7600 cal BC 7840-7730 (95%)
OxA-12326 HEX02 291A hazelnut shell from hearth [291] -22.6 8765±40 cal BC 8160-7600 cal BC 7890-7740 (95%)
OxA-12327 HEX02 355A hazelnut shell from hearth [355] -24.8 8725±39 cal BC 7960-7600 cal BC 7940-7930 (2%) or cal BC
7920-7760 (93%)
OxA-12402 HEX02 383A hazelnut shell from hearth [383] -24.2 8885±65 cal BC 8260-7750 cal BC 7930-7750 (95%)
OxA-12408 HEX02 055A hazelnut shell from burning pit [055] -23.7 8330±45 cal BC 7530-7190 cal BC 7510-7500 (1%) or cal BC
7490-7320 (94%)
68
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Laboratory
Number
Sediment Core
Sample Material and Context 13C
(‰)
Radiocarbon
Age BP
Calibrated Date (95%
confidence)
Posterior Density Estimate (95%
probability)
OxA-12944 HEX02-11007-11 phragmites from 137cm -26.3 2982±34 cal BC 1380-1050 cal BC 1380-1330 (8%) or cal BC
1320-1110 (87%)
OxA-12945 HEX02-11007-12 unidentified herbaceous material from 178cm -27.2 3755±35 cal BC 2290-2030 -
OxA-12967 HEX02-11007-13 phragmites from 200cm -25.8 3311±35 cal BC 1690-1510 cal BC 1690-1510 (95%)
OxA-12946 HEX02-11007-14 alder (Alnus glutinosa) roundwood including bark from
-27.7 3820±35 cal BC 2410-2140 cal BC 2410-2370 (5%) or cal BC
213 cm
2360-2130 (90%)
OxA-12947 HEX02-11007-15 alder (Alnus glutinosa) roundwood from 297cm -29.1 5846±38 cal BC 4810-4600 cal BC 4810-4600 (95%)
OxA-13028 HEX02-11007-16 oak (Quercus sp.)/ash (Fraxinus excelsior), knotty
-24.3 6661±38 cal BC 5660-5480 -
fragment from very degraded roundwood from 366cm
OxA-12948 HEX02-11007-16 oak (Quercus sp.)/ash (Fraxinus excelsior), knotty
-25.7 6473±37 cal BC 5490-5360 -
fragment from very degraded roundwood from 366cm
OxA-12949 HEX02-11007-17 woody stem from 406cm -31.3 5985±36 cal BC 4950-4730 cal BC 4960-4780 (95%)
OxA-12950 HEX02-11007-18 woody stem/root from 437cm -30.1 6290±45 cal BC 5370-5080 cal BC 5310-5200 (58%) or cal BC
5180-5070 (37%)
OxA-12951 HEX02-11007-19 hazel (Corylus avellana) roundwood from 444cm -26.6 6274±38 cal BC 5320-5080 cal BC 5340-5200 (92%) or cal BC
5170-5140 (3%)
OxA-13029 HEX02-11007-20 alder (Alnus glutinosa) fragment from roundwood or
-27.9 6586±39 cal BC 5620-5470 cal BC 5620-5580 (22%) or cal BC
largewood from 485cm
5570-5470 (73%)
OxA-12952 HEX02-11007-21 hazelnut shell from 580cm -26.5 6988±37 cal BC 5990-5740 cal BC 5980-5940 (7%) or cal BC
5930-5740 (88%)
OxA-12953 HEX02-11007-21 hazelnut shell from 580cm -26.1 7117±39 cal BC 6070-5890 cal BC 6000-5870 (91%) or cal BC
5860-5840 (4%)
OxA-12954 HEX02-11007-22 sliver of wood bark from wide roundwood/trunk from
-30.7 7075±37 cal BC 6020-5840 cal BC 6020-5900 (95%)
583cm
OxA-11859 HEX02-11007A/09 Corylus avellana, ?carbonised wood from 627cm -26.4 7174±35 cal BC 6160-5920 cal BC 6070-5980 (95%)
OxA-11860 HEX02-11007A/10 Corylus avellana twig from 630cm -27.3 7160±40 cal BC 6160-5920 cal BC 6160-6140 (2%) or cal BC
6090-5990 (93%)
OxA-11858 HEX02-11007A/08 hazelnut shell from 683cm -25.6 7308±40 cal BC 6240-6030 cal BC 6200-6050 (95%)
OxA-11833 HEX02-11007A/07 hazelnut shell from 684-685cm -24.9 7269±39 cal BC 6230-6020 cal BC 6230-6080 (95%)
OxA-11936 HEX02-11007A/05 small twig from 764-765cm -29.4 10035±45 cal BC 10140-9310 cal BC 9750-9300 (95%)
OxA-13370 HEX02-11007A/06 indeterminate herbaceous stem from 770 cm -29.1 10000±90 cal BC 10150-9250 cal BC 10020-9840 (15%) or cal BC
9830-9390 (80%)
OxA-12825 HEX02-11007A/04 bulk sediment (humin) from 783-784cm -24.6 10265±70 cal BC 10790-9740 cal BC 10390-9740 (95%)
OxA-11852 HEX02-11007A/04 bulk sediment (humic acid) from 783-784cm -25.8 8465±45 cal BC 7590-7380 -
OxA-12824 HEX02-11007A/03 bulk sediment (humin) from 791-792cm -23.9 10430±140 cal BC 10970-9740 cal BC 10860-10760 (4%) or cal BC
10730-9810 (91%)
OxA-11870 HEX02-11007A/03 bulk sediment (humic acid) from 791-792cm -26.4 8250±45 cal BC 7590-7380 -
AA-45401 HOW4-439 hazel (Corylus avellana)/willow (Salicaceae) from 439cm -27.3 6595±55 cal BC 5630-5470 -
AA-45402 HOW4-720 hazelnut shell from 720cm -26.3 6865±55 cal BC 5840-5630 -
Table 6.1: Radiocarbon determinations from Howick
6ABSOLUTE DATING 69
70
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
estimates of the likely age of the site. Radiocarbon
results were simulated using the R_Simulate function
in OxCal, with error terms estimated on the basis of
the material available. Once the first series of results
had been received, further simulation models were
constructed and further samples submitted to refine
the dating.
The first dates from the sediments in the Howick
Burn consisted of two samples of plant macrofossils
dated from core HOW4 in 2001, as part of the
assessment of the palaeoenvironmental potential of
these sediments. These results suggested that part of
the sequence might be contemporary with the
Mesolithic structure, and so a second core (11007)
was taken for detailed analysis.
Two series of radiocarbon samples were submitted
from core 11007. The first set of samples concentrated
on the base of the core to determine whether this was
contemporary with the Mesolithic structure, and the
second series aimed to date particular episodes in the
pollen record.
Unfortunately, the stream at Howick runs across
exposed coal measures, potentially contaminating
any samples of bulk sediment with carbon of geo-
logical age (thus making the resultant dates anom-
alously old). For this reason, samples of single
waterlogged plant macrofossils, which had been
extracted from known levels in the core, were dated
wherever possible. Preservation was not ideal, and
so the range of suitable material was limited.
It was particularly disappointing that no macro-
fossils could be recovered from the lowest organic
sediments, which were thought most likely to
preserve a record of the environment contemporary
with the use of the Mesolithic structure. Because of
the desirability of dating these lowest levels, bulk
sediment was submitted for dating. Both samples
consisted of 10mm of sediment, and repeat fractions
(humic acids and humin) were extracted from each.
This is important because any coal contamination
would make the humin fractions anomalously old. If
consistent results are returned from different fractions
of the same sample, then more confidence can be
placed in the reliability of the date (Shore et al. 1995).
by the Scottish Universities Research and Reactor
Centre were processed as described by Stenhouse
and Baxter (1983) and Slota et al. (1987), and dated
using methods outlined in Donahue et al. (1997). The
sample dated at Beta Analytic Inc was dated by AMS
using methods outlined at http://www.radiocarbon.
com/.
All three laboratories maintain continual pro-
grammes of quality assurance procedures, in addition
to participation in international inter-comparisons
(Scott 2003). These tests indicate no laboratory offsets
and demonstrate the validity of the precision quoted.
The Results
The results are given in Table 6.1, and are quoted in
accordance with the international standard known as
the Trondheim convention (Stuiver and Kra 1986).
They are conventional radiocarbon ages (Stuiver and
Polach 1977).
Calibration
The calibrations of these results, which relate the
radiocarbon measurements directly to the calendrical
time scale, are given in Table 6.1 and in outline in
Figures 6.2, 6.3, and 6.5. All have been calculated
using the datasets published by Stuiver et al. (1998)
and the computer program OxCal (v3.5) (Bronk
Ramsey 1995; 1998; 2001). The calibrated date ranges
cited in the text are those for 95% confidence. They
are quoted in the form recommended by Mook (1986),
with the end points rounded outwards to 10 years.
The ranges in Table 6.1 have been calculated ac-
cording to the maximum intercept method (Stuiver
and Reimer 1986); all other ranges are derived from
the probability method (Stuiver and Reimer 1993).
Those ranges printed in italics in the text and tables
are posterior density estimates, derived from the
mathematical modelling described below.
Radiocarbon Analysis and Quality Assurance
Samples processed at the Oxford Radiocarbon
Accelerator Unit were prepared using methods
outlined in Hedges et al. (1989). Those with OxA-
numbers below 12000 were dated as described by
Bronk Ramsey and Hedges (1997), those with OxA-
numbers above this were measured as described by
Bronk Ramsey et al. (2004). All targets were graphite,
except those for OxA-12402, OxA-12824-5, and OxA-
13370 which were carbon dioxide. The samples dated
Analysis and Interpretation
Mesolithic Hut
The model for the chronology of the Mesolithic
structure at Howick is shown in Figures 6.2–6.4.
Hearths 355, 379, and 383 were dated, and are
among the earliest identifiable features inside the
structure (OxA-11801, OxA-12327, OxA-11856-7,
OxA-12292 and OxA-12402). Directly above hearth
355 lay hearth 291. Two fragments of hazelnut shell
6ABSOLUTE DATING 71
were also dated from this deposit (OxA-11803 and
OxA-12326). Directly above hearth 379 was hearth
357. One sample was dated from this feature (OxA-
11802). There was no direct relationship between
hearth 291 and 357. All these samples fall in strat-
igraphic Phase 1a.
Later than all these samples was the large burnt
spread 340, which produced two radiocarbon dates
(OxA-11804 and OxA-12325). Above this is hearth 293
(OxA-11828-9), and above this in turn was a large
hearth 268 (OxA-11854-5). These three features
provide dates for stratigraphic Phase 1B. The results
are consistent with the stratigraphic positions of these
contexts, except for OxA-11829, which has only a 0.3%
probability of dating this phase of activity. This
hazelnut fragment is clearly earlier reworked material.
The first rebuilding of the hut is evidenced by
levelling layer 210 and associated structural features.
A number of discrete features were dated on, and cut
into, layer 210. This included an unnumbered burnt
patch (OxA-12347 and OxA-12324), hearth 109 (OxA-
11830-1), and circular burning pit 158 (OxA-11832
and OxA-11853). All these features fall in strat-
igraphic Phase 2. The results are consistent with the
stratigraphic positions of these contexts.
The second rebuilding of the hut is recognised by
levelling layer 049, and associated structural features.
Two hearths on top of this layer (047 and 173)
produced radiocarbon dates (OxA-11805, AA-41788,
OxA-12294 and OxA-1186-7). Following the initial
assessment excavation in 2000, a sample from post
hole 051 was dated (Beta-153650). Although this
measurement provided a useful preliminary in-
dication of the age of the structure, it has not been
included in the model described here because the
taphonomy of a charred hazelnut within a post hole
is uncertain. These features provide absolute dating
for stratigraphic Phase 3.
A number of features appear to have been cut into
the uppermost archaeological deposits within the hut.
Two hearths (055 and 063) produced radiocarbon
dates that are significantly later than those from the
hut itself (OxA-12293, OxA-12408 and OxA-11806-7).
It appears that another unnumbered burnt patch from
context 210 may be part of this activity, since this
feature produced two similar radiocarbon measure-
ments (OxA-12322-3). An anomalously late sample
from Phase 3 (OxA-11805 from hearth 047) appears to
be intrusive in its context and may also relate to this
late use of the site. This final phase of activity was
only recognised as a result of the radiocarbon
programme, and forms a fourth phase of activity at
the site.
The model explained above, and shown in Figures
6.2 and 6.3, can now be used to address the objectives
of the dating programme outlined above.
The hut was first occupied in cal BC 7970–7760
(95% probability; start structure; Fig 6.2). The first
rebuild occurred in cal BC 7800–7700 (95% probability;
rebuild 1; Fig 6.2), and the second rebuild occurred in
cal BC 7750–7630 (95% probability; rebuild 2; Fig 6.2).
Occupation of the hut appears to have ended in cal
BC 7740–7560 (95% probability; end structure; Fig 6.2).
Phase 1 occupation lasted for 20–230 years (95%
probability) or 40–160 years (68% probability; use 1; Fig
6.4). Phase 2 occupation lasted for 1–110 years (95%
probability) or 10–80 years (68% probability; use 2; Fig
6.4). Phase 3 occupation lasted for 1–100 years (95%
probability) or 1–60 years (68% probability; use 3; Fig
6.4). Overall the hut was occupied for 40–380 years
(95% probability) or 100–300 years (68% probability; use
hut; Fig 6.4). It is apparent that each phase of
occupation becomes progressively shorter. This is
shown clearly in Figure 6.4, where the skewed
probability distributions for the duration of each
phase of occupation suggest that shorter periods of
use are more likely within each range. In contrast the
probability distribution for the duration of use of the
entire structure is less skewed, and it is likely to have
had an overall occupation of a couple of centuries.
The dating of Phase 4 is shown in Figure 6.3. The
re-use of the site began in cal BC 7560–7330 (95%
probability; start re-use; Fig 6.3), and according to the
model the site was finally abandoned in cal BC 7450–
7240 (93% probability) or cal BC 7220–7120 (3%
probability; end re-use; Fig 6.3). The three features
assigned to Phase 4 produced six radiocarbon results
which are statistically consistent (T’=6.4; T’(5%)=11.1;
˜=5; Ward and Wilson 1978). This means that this re-
occupation could simply represent a single visit to
the site. Alternatively, these features could span a
period of 1–280 years (95% probability) or 1–120 years
(68% probability; use 4; Fig 6.4), although the shape of
the probability distribution suggests that a shorter
period of re-use is most likely.
The site appears to have been abandoned between
Phases 3 and 4, for a period of 50–360 years (95% prob-
ability) or 130–280 years (68% probability; gap; Fig 6.4).
The model is in good overall agreement with the
stratigraphic sequence (Aoverall = 69.0%; Bronk Ramsey
1995), and with the structural phasing identified by
excavation. Whilst there is clearly a gap between
Phases 3 and 4, the model suggests that occupation
from Phase 1 to Phase 3 is continuous. This is also
reflected in the stratigraphic sequencing, which
provided no indication of a break between phases.
The dating of this site to the earlier eighth
millennium cal BC provides the first absolute dating
for a discrete Mesolithic flint assemblage in North-
East England. Together with the dates recently
published for Crammond (Lawson 2001; see Table
6.2), the Howick excavations have produced early
dates for scalene triangle microliths. Indeed, the dates
from these sites have pushed back the chronology of
narrow-blade industries in northern Britain to the
centuries before and around 8000 cal BC.
72
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 6.3. Probability distributions of dates from the Mesolithic structure at Howick during Phase 4, which is a later reuse of the area. The
format is identical to that for Figure 1. The large square brackets down the left hand side along with the OxCal keywords define the overall
model exactly.

Figure 6.4. Probability distribution showing the length of time during which the structure at Howick was in use. The distribution is derived
from the model defined in Figures 6.2 and 6.3.
Sediment Sequence
The chronological model for Core 11007 from the
Howick Burn is shown in Figure 6.5. A full discussion
of this core in relation to its wider Holocene context is
published elsewhere (Boomer et al. 2007). In general,
the relative dating provided by the height in the
sediment column is consistent with the radiocarbon
measurements. The first exception to this is the two
results on the humic acid fractions of samples 11007-03
and 11007-04. These were from near the base of the
organic deposits, and in both cases the humic acid
fractions are several thousand years younger than the
replicate measurement on the humin fractions of the
same samples. The results from these humin fractions
(OxA-12824-5), however, are consistent with the results
from the plant macrofossils in the immediately
overlying levels (OxA-13370 and OxA-11936). As these
macrofossils cannot have been contaminated with coal
fragments, it seems likely that it is the measurements
6ABSOLUTE DATING 73

Figure 6.5. Probability distributions of dates from the sediment core 11007 at Howick Burn. The format is identical to that of Figure 6.1.
Radiocarbon dates marked with a “?” have been excluded from the analysis. The large square brackets down the left hand side along with
the OxCal keywords define the overall model exactly.
on the humin fractions that are nearer in age to the
deposition of the sediment at these levels, and the
humic acid results are anomalously young. It appears
that younger humic acids may have been translocated
down the sediment profile. For this reason the results
from the humic acid fractions have been excluded from
the model.
There also appear to be re-worked plant macro-
fossils present at two levels higher up the profile.
Both samples at 366cm (OxA-13028 and OxA-12948)
are older than their relative stratigraphic position
would indicate, and may have been washed into the
lower valley from higher up in the catchment. A
single macrofossil collected at 178cm (OxA-12945)
74
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 6.6. Age-depth relationships for the three sections of Core 11007. Trend lines indicate mean accumulation rates.
also appears to be too old for its position in the
column. Consequently, these three measurements
have also been excluded from the analysis.
Figure 6.5 demonstrates that Core 11007 contains
sediments that span at least 12,000 years. The lowest
organic deposits measured, between 791–2cm, date
to cal BC 10860–10760 (4% probability) or cal BC 10730–
9810 (91% probability; OxA-12824). Sediment appears
to have accumulated relatively constantly until at
least cal BC 9750–7300 (95% probability; OxA-11936;
765–4cm). If we assume that the posterior density
estimates are normally distributed and calculate their
means, simple linear regression can be fitted, thus
providing an estimate of the accumulation rate of this
part of the core of 0.34m ka-1 (0.3m per thousand
years; Fig 6.6a).
There appears to be an erosional event, represented
by the deposition of a coarse sand and gravel unit,
between 750cm and 705cm. It is likely that this event
would have eroded pre-existing sediments. The next
datable horizon, at 684cm, is dated to cal BC 6230–
6080 (OxA-11833), suggesting that around 3000 years
of sediment have been lost. Given the relatively rapid
sediment accumulation of the sediments overlying
this sand and gravel, it is likely that the erosional
event occurred close to this date.
The sedimentation rate appears to be reasonably
constant between 684cm and 297cm, at approx-
imately 2.63m ka-1 (Fig 6.6b). The last datable level
within this unit dates to cal BC 4810–4600 (95%
probability; OxA-12947). This horizon also marks the
end of a period of sedimentation under permanent
aquatic conditions, as witnessed by the microfossil
record (see Chapter 12). A 45cm thick deposit of silty
sand represents more rapid accumulation within this
phase, an event dated to cal BC 4920–4650 (95%
probability; sand).
The next datable level occurs at 213cm, dated to
cal BC 2410–2370 (5% probability) or cal BC 2360–2130
(90% probability; OxA-12936). The accumulation rate
of the overlying material reduces to approximately
0.74m ka-1 (Fig 6.6c). This rate suggests that there
may well be no gap in the sediment sequence, just a
slowing in the accumulation rate above 297cm. The
youngest sample, at a depth of 137cm, dates to cal BC
1380–1330 (8% probability) or cal BC 1320–1110 (87%
probability; OxA-12944), and so the upper 1.3m of
sediment represents the last three thousand years.
Landscaping in the valley during the nineteenth
century is represented by disturbance in the upper
30cm of the sediment profile. This makes it impossible
to extrapolate the sedimentation rate to the modern
floodplain surface.
The posterior density estimates for the sediment
core are plotted against their absolute core depths in
Figure 6.7.

Figure 6.7. Age-depth relationship for Core 11007.
7CHIPPED STONE TOOLS 75
7CHIPPED STONE TOOLS
Clive Waddington and Kristian Pedersen
The Lithic Assemblage
Clive Waddington
Introduction
The archaeological deposits at Howick produced lithic
material from most contexts. This included 13,219
lithics from the Mesolithic hut, 58 from the external
pits near to the hut, 64 from the Bronze Age cists, 240
from a later linear feature and 901 from unstratified
deposits. The lithics from the fieldwalking and test
pits are dealt with separately in their respective
chapters, while those from the cists and the later linear
burning pit are reported in a separate publication
(Waddington et al. 2005). The lithics from the pits and
unstratified locations are reported under a separate
section towards the end of this chapter, otherwise the
rest of this chapter is concerned solely with those lithics
recovered from the stratified deposits within the
Mesolithic hut. Every piece was examined and
catalogued and this information recorded into a
database, which forms part of the archive. The
following report provides an account of the raw
material, artefact types, technology, assemblage types
and stratigraphic and spatial distribution, together
with a discussion of the results. A selection of the
various implement types is illustrated in Figures 7.1
and 7.2. The following chapter reports the results of a
limited programme of residue and use-wear analysis
on a sample of the lithic assemblage.
The lithic assemblage is remarkable in several
respects. The lithics from within the Mesolithic hut
form a discrete stratified group that is not disturbed
by any intrusive or overlying later material. This
means it is the most coherent and stratigraphically
intact Mesolithic assemblage so far recovered in
North-East England. Thanks to the high-resolution
radiocarbon dating of the hut deposits, the assem-
blage is also tightly defined in terms of its chron-
ological span, which, as more assemblages are dated,
will allow for the identification of typological
progression in the North-East during the Mesolithic.
All the lithic materials from the Mesolithic deposits
are from local sources and this has allowed the
assemblage to be related directly to local patterns of
resource procurement, as well as other comparable
North-East lithic assemblages. The hut deposits have
produced a substantial collection of 276 microliths
that will allow North-East traditions to be fitted into
the wider milieu of Mesolithic industries from
adjacent areas of England and Scotland, as well as
further afield across the North Sea and Irish Sea.
The proportion of tools in the hut assemblage is
4.6%, which is low when compared with other sites
such as Star Carr with 14% (Clark 1971; 1972) and,
closer to Howick, South Shields with 8% (Waddington
2001a, 78). However, the Howick figure is skewed
in relation to many other excavated Mesolithic
assemblages in the UK, as the entire hut deposits were
passed through a 2mm mesh in a flotation tank to
maximise the retrieval of microdebitage, which
increased the recovery of flint from the hut by around
8200 pieces. Without this material the assemblage
would have only numbered around 5000 and the
proportion of tools would have been significantly
higher, around 12%, which is directly comparable to
other Mesolithic settlement sites where sieving or
flotation have not taken place. All phases of the chaîne
opératoire are represented in the assemblage, indicating
not only stone-tool working on the site but also the use
of tools for a wide variety of processing tasks. This
conclusion is strengthened by the fact that many of the
finished tools in undisturbed deposits are broken as a
result of their use. The actual amount of utilised pieces
in the assemblage is likely to be considerably higher
than the 4.6% figure for visually recognised tools, as
the use-wear and residue analysis (see below, Chapter
8) has shown that many pieces that retain no visible
surface traces of retouch or utilisation have, on
microscopic inspection, clear traces of use.
All the diagnostic pieces in the assemblage from
the hut and nearby pits are Mesolithic types. The
only diagnostic piece in the entire assemblage that is
later than the Mesolithic is a broken leaf-shaped
arrowhead, typical of Early Neolithic assemblages,
from the unstratified subsoil to the north of the hut
(ill. in Fig. 7.2).
76
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 7.1. Cores, scrapers and burins from Howick.
7CHIPPED STONE TOOLS 77

Figure 7.2. Microliths and blade tools from Howick.
78
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Raw Material
The lithic assemblage is composed almost entirely of
flint (13,073 pieces = 98.9%) with a few pieces of chert
(94 pieces = 0.7%), quartz (49 pieces = 0.4%) and one
piece each of chipped agate, sandstone and an
unidentified stone. Most of the flint pieces have been
struck from small nodules and as a result, areas of
cortex survive on many pieces, which has allowed
the likely origin of 14.4% of the assemblage to be
assessed. Beach flint – that is, nodules that have been
rolled in the sea and redeposited on the shore –
account for 97.8% of the provenanced flint. Glacial
flint from till deposits, which forms the drift cover
along this stretch of the coastal plain, accounts for
2.1%. Possible pieces of nodular flint with a chalky
cortex accounted for just 0.1%.
The flint comes in a wide range of shades, shapes
and sizes, although the most common types are
various shades of light grey material (52%), medium
and dark grey (15.4%), a distinctive speckled red-
brown flint (5.8%), an off-white flint (5.7%) and brown
flint (2.8%), together with small quantities of honey-
coloured, purple, pink, orange and translucent flints
(0.7%). The remaining 17.6% of the assemblage was
unable to be attributed a colour as these pieces were
either patinated or burnt. The beach material tends to
have a very rounded and smooth cortex that is often
a very different colour from the flint within the
nodule. Although impurities are fairly common, both
the beach pebble and glacial flint are generally of
high quality. The main limitation of this local material
is its small size, which has evidently constrained, to
some extent, the size and range of tools that were
made. The beach flint pieces from the site show that
nodules ranging from thumb to fist-sized were
typically used. As beach and glacial flint can come
from a wide variety of sources and still end up in the
same area it is not surprising that a wide range of
flint material is represented in this assemblage.
The origin of the beach flint is an intriguing
question that requires attention here. Some of this
material derives from the glacial till deposits of the
coastal plain that are eroded by the sea and then
redeposited on the shore. Others are from till deposits
eroded from the submerged sea floor and this
material no doubt has a wide catchment, ranging
from inshore waters to deep North Sea waters where
the tills from the Scandinavian ice sheets are eroded
during storm events. Still another type of beach flint
is that which comes from the erosion of submerged
prehistoric sites that now lie under the North Sea.
This material is distinctive: although evidently
chipped – usually to form blades and what for
Northumberland are large blade tools – these pieces
have developed very heavy patinas over the chipped
areas and have subsequently become heavily beach-
rolled along all edges. These pieces appear to be
mostly of Palaeolithic origin, although some may
belong to the early Mesolithic. When found on site at
Howick they have a distinctive flaking stratigraphy
evident on their surfaces, including: fresher Meso-
lithic flaking facets that go with the occupation at
Howick, earlier flaking facets that are very heavily
patinated and beach-rolled, dating from the Palaeo-
lithic/Early Mesolithic, and areas of original cortex.
A particularly good example is scraper [3050], which
has been re-chipped from a previously struck patin-
ated piece that had been beach-rolled. There are many
other examples where beach-rolled and heavily
patinated previously struck pieces have been col-
lected as nodules for reworking on site, usually as
cores. This includes cores [2319] and [3737], together
with flakes and blades struck from recycled pieces
such as [793], [1748] and [3641]. This is an important
phenomenon, as it not only serves as proxy evidence
for Palaeolithic and Early Mesolithic occupation in
areas now submerged below the North Sea, but the
earlier flaking of these pieces cannot have gone
unnoticed by the Howick inhabitants. It is tempting
to speculate how the Howick people regarded these,
by then, ancient artefacts, although they clearly had
few concerns when it came to recycling this material.
Howick is not the first place on the North-East coast
where the re-use of already ancient chipped flints has
been evidenced in Mesolithic collections. Lithic
assemblages assessed by the author from Middle
Warren near Hartlepool (Waddington 1996), various
locations on the Durham coast (Waddington 1998a),
and the Trechman collection from the North-East
coast (personal inspection at Sunderland Museum),
have all produced evidence for this recycling.
Therefore, the behaviour witnessed at Howick,
although more firmly dated than at the other sites, is
part of a pattern of what appears to be opportunistic
acquisition of flint during the Mesolithic.
The glacial flint, which can be collected from
exposures and scars in the local till such as cliff and
stream sections, tends to have a thinner cortex than
the beach flint and has a rougher, usually non-chalky
texture, sometimes with a slight honeycomb effect.
The cortex is generally dull and opaque and can range
in colour from a dirty grey to sandy brown. It is
common for pieces of glacial flint to contain im-
purities and sometimes these can be observed to have
affected the flaked shape of a piece. In contrast, the
nodular flint can be recognised by the presence of a
more or less uniform and thick chalky cortex and
with fewer impurities evident in the flint. There is no
history of ballast dumping in the Howick area and
only a couple of possible pieces of nodular flint were
recorded.
The chert pieces are more opaque and less fine-
grained than the flint. The colours of the chert pieces
are also distinctive, in particular the blue-grey variety
7CHIPPED STONE TOOLS 79
that accounts for 14.8% of the 97 chert pieces. There
are also various shades of grey, accounting for 49%,
together with brown (13.8%), white (7.4%), red-brown
(4.3%), purple (3.2) and other colours (7.4%). The
chert, like the flint, comes from beach and glacial
sources; 18% could be provenanced as beach material
on account of the surviving areas of cortex. Chert can
be collected today both from the beach and from the
boulder clay, although in much less abundance than
the flint.
The quartz component was easily recognised,
being crystalline and having a distinctive sparkle.
Most of the quartz was either light grey (40.8%) or
white (34.7%), with occasional pieces of translucent
material (12.2%) and the rest a mixture of pink, red
and brown (12.2%). Some quartz nodules can be
found in the coastal boulder clays and it is likely that
these struck materials represent the opportunistic
collection of pieces that were suitable for knapping.
The single piece of agate is of the banded variety and
is also likely to have come from a boulder clay source,
while the single piece of flaked sandstone no doubt
came from the exposed solid geology at the base of
the cliffs, which consists of inter-bedded sandstones,
limestones and mudstone.
Technology
This blade-based assemblage is typical of Later
Mesolithic industries of the British Isles, even though
chronologically it occurs at the very beginning of this
phase. It is characterised typologically as a narrow-
blade industry, similar to those from Filpoke Beacon
(Jacobi 1976), and fits well with a date of around 7800
cal BC for the initial occupation of the site (see below
Chapter 6). This type of narrow-blade assemblage has
been characterised by Jacobi as comprising ‘micro-
triangles’ with elongated scalene microliths diagnostic
of this techno-complex (Jacobi 1976, 71–2). No axes
were present in the assemblage, which is consistent
with Jacobi’s observation that core axes do not occur
on Later Mesolithic northern British sites (Jacobi 1976,
73), although this should probably be seen as a
function of the small beach nodules precluding
widespread axe head manufacture, and those axe
heads that were used being recycled into smaller tools
rather than being discarded. Flint from the entire
knapping sequence is present in the hut assemblage,
revealing a focus on the production of small parallel-
sided blade-based tools. As the raw material was
already small when collected, some of the blade forms
tend to be fairly squat although more slender blades,
typical in more southern collections, also occur.
However, the squat forms are blades nonetheless and
this is further evidenced by the flaking scars on cores
which all reveal traces of blade detachments. Bipolar
flaking is a common feature of the assemblage,
apparent on both cores and detached flakes. This is
almost certainly due to the nature of the raw material:
being small, the pieces could be most easily worked
by placing them on an anvil stone. Indeed one of the
reasons why there are no large flint pieces in the
assemblage, such as axes, large blades and cores, is
that pieces of this size are unlikely to have been
discarded on account of their ability to be recycled
into further tools. This would also explain why most
of the finished and broken tools are so small, as they
had reached the size where they could not be recycled
further. Indeed direct evidence for recycling is
apparent on a number of pieces, such as the broken
flake [1732] that was reworked into a scraper. Overall
this indicates a parsimonious attitude towards the
discard of re-usable pieces, reflecting the size con-
straint of an otherwise abundantly available source of
raw material.
Direct flaking with hard hammers is evident
throughout the assemblage, as witnessed by the
abraded surfaces and crushed striking platforms on
primary and secondary flakes together with the
common occurrence of eraillure scars, bipolar flakes
made on anvil stones and the presence of a quartzite
hammerstone within the hut deposits. Visual in-
spection indicates the use of hard and soft hammers
and possibly some indirect percussion. A curious
occurrence in the hut assemblage is the presence of
three small flint blades (ills. 3271, 3842, 6172 in Fig
7.2) that appear on first inspection to have been finely
cut in a straight line; however, these are in fact a
natural phenomenon resulting from the erosion of
echinoids within the flint (Jacobi pers. comm.).
As said, all stages of the chaîne opératoire are present
in the hut assemblage, indicating that nodules were
collected locally, brought to the hut, knapped and
fashioned into tools where they were also used and,
ultimately, discarded. The assemblage has been
classified into primary (nodules, test pieces and
primary flakes), secondary (secondary flakes, cores
and blanks) and tertiary material (finished tools only,
not including the debitage from tool production,
which probably makes up most of the unascribed
material). Primary material accounts for 13.2% of the
assemblage; secondary material 75.1% and tertiary
pieces 4.6%, leaving 7.2% that could not be ascribed
to one of these stages in the reduction sequence. The
relatively low proportion of primary waste, which
could be expected to form more of the assemblage if
all of this work was taking place on site, suggests that
much of the primary working took place elsewhere,
no doubt closer to the source of the material.
Therefore, working of stone at the primary stage was
only undertaken on a relatively small scale in and
around the hut. The high proportion of secondary
material is typical for settlement sites and indicates
the manufacture of blanks, although the number of
80
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
cores recovered from the hut is relatively small (see
below). The high counts of debitage and micro-
debitage and the presence of blade blanks and
finished tools indicate that the manufacture of flakes
and blades into stone tools took place on a consider-
able scale in the hut. As the use-wear study makes
clear (see below, Chapter 9) many of what appear to
be blanks, based on visual inspection alone, are likely
to have been utilised, and therefore classified as
secondary. The fact that many of the identifiable tools
were broken in antiquity (see above) and show visible
signs of wear, as well as microscopic residues,
indicates that most of the tools were used in and
around the hut for processing activities, only to be
discarded when they broke or were of no further use.
Some, however, were curated and recycled after their
initial use. This paints a picture of an organised
sequence of stone tool production from the collection
strategy to tool manufacture and use around the hut,
as well as a sparing approach to tool production, use,
curation and discard.
Final flaking is dominated by abrupt retouch,
which, in virtually all cases, is applied to only one
side of the piece. This is consistent with the flaking
apparent in most Mesolithic assemblages. Such
unifacial working, however, applies to fully re-
touched pieces, whereas some of the edge-trimmed
pieces can have trimming on both sides depending
on the shape and form of the piece concerned. In
those few instances where the piece has working on
more than one side, this is usually due to the odd
shape of the piece and/or the presence of impurities.
In other words, it is a strategy that appears to have
been employed expeditiously to cope with the
irregularities of the raw material. Other tool manu-
facturing techniques include the widespread practice
of snapping the ends of blades as part of the microlith
production strategy. There are many examples of
snapped blade segments in the assemblage, with
snapping at an angle across the proximal end most
common, although sometimes it occurs across the
distal end. This technique is particularly well suited
to the production of triangle forms and it is precisely
these types which dominate the microlith assemblage
(see below).
The Assemblage
The Howick assemblage has produced a range of
different artefact types, summarised in Table 7.1
below, testifying to a wide range of activities at the
site. As is typical of assemblages of this sort, it is
dominated by flake and blade debitage, although
there is a significant collection of microliths of
different types. Many pieces show evidence for very
light patina development but otherwise the lithics are
remarkable for surviving in a very fresh state. On
those pieces where patina development is more
advanced it tends to be milky white in colour. This
contrasts with the heavier patina visible on the re-
used beach-rolled material that is usually an opaque
orange-brown colour. In fact the patina on the re-
used pieces is so thick that they have in effect become
entirely recorticated. It is not just the flint, though,
that shows signs of patina development. Occasional
pieces of chert and agate do so too, such as a re-
struck piece of quartz, [4068], used as a core. A total
of 163 pieces had evidence for some patina develop-
ment, which accounts for just 1.2% of the assemblage.
Those flints with heavy patina development are, by
and large, likely to be extremely old and in some
cases potentially of Palaeolithic age.
Measurements were only taken on whole pieces
with dimensions recorded to the nearest half milli-
metre. Mean average measurement calculations
exclude the broken pieces, which were not measured.
All measurements given are the maximum dimen-
sions of length, breadth and thickness based on the
recording technique advocated by Saville (1980).
Cores, test pieces and nodules had measurements
taken along their two longest axes, while all other
pieces were recorded across three axes. Micro-
debitage has been characterised as waste material
that has maximum dimensions of 5mm or less, while
debitage has been characterised as waste material
with dimensions greater than 5mm. As we are dealing
here with a blade-based assemblage, the terms
‘primary’, ‘secondary’ and ‘tertiary’ have been used
to describe pieces in relation to different stages of the
reduction sequence. For those pieces where colour or
stage in the reduction sequence could not be as-
certained no data were recorded for these attributes.
This was particularly frequent for broken, burnt and
patinated pieces. A feature of this assemblage is the
use of multi-directional flaking in order to gain as
many flakes and blades from a piece as possible. This
has resulted in flake facets from previous core
removals giving some flakes the appearance of being
retouched.
A significant proportion of the assemblage is
broken (62.7%), presumably through use, which
indicates that this is an assemblage directly related to
functional activities that took place at the place of
discard. The breaks have occurred due to use, the
knapping process and perhaps some trampling
underfoot, although the floor surface was soft sand.
As the assemblage was retrieved from occupation
deposits within a defined hut that had a sequence of
hearth pits at its centre, it is not surprising that a
significant number of the lithics are heavily burnt.
Indeed most of the 2045 burnt pieces were from the
hearth fills and in total the burnt pieces account for
15.5% of the assemblage.
Type Count % of
Assemblage
Hammerstone 1 -
Nodules 11 0.1%
Test Pieces 21 0.2%
Unretouched Flakes 10,093 76.4%
Unretouched Blades 2,376 18%
Cores 107 0.8%
Retouched Flakes 92 0.7%
Retouched Blades 51 0.4%
Utilised Flakes 14 0.1%
Utilised Blades 39 0.3%
Notched Flakes and Blades 4 -
Scrapers 109 0.8%
Awls 10 + 1 poss. 0.1%
Burins 4 + 1 poss. -
Microliths 275 2.1%
Microburins 9 0.1%
Crested Bladelet 1
Total 13,219
Table 7.1. Summary of artefact types from the Howick hut.
7CHIPPED STONE TOOLS 81
Phase Lithic Count
Phase 1A 1609
Phase 1B 4462
Phase 2 2883
Phase 3 4253
Phase 4 12
Total 13219
Table 7.2. Summary of lithic counts by Phase.
have been struck no more than twice before being
discarded. Their mean maximum dimensions are
36.4mm by 28.7mm, which corresponds with the
small sizes of the nodules found on the site. Of the 21
test pieces found in the hut, 17 were flint, 3 quartz
and 1 chert. Apart from one piece of glacial flint all
the other pieces were from the beach. The presence of
these test pieces confirms that some preliminary
working and testing of pieces took place at the hut.
Hammerstones
One hammerstone was recovered from the Mesolithic
hut assemblage together with one from an un-
stratified context [87]. Both were made of quartzite,
being red-brown and brown respectively. Both were
fist-sized pieces, the one from the hut having
maximum dimensions of 45mm by 42mm and the
unstratified one 67mm by 61mm. A coarse stone
elongated tool was classified as having potentially
been used in a hammering motion (see below,
Chapter 8), but as it is made from relatively soft
sandstone it is unlikely to have been used for the
knapping of flint.
Nodules and Test Pieces
Unworked nodules comprise 0.1% of the assemblage
and their presence indicates the acquisition of
material for chipping into tools. A total of eleven
nodules was recovered of which nine were flint and
two quartz. Their mean average dimensions are
33.7mm by 24.8mm, indicating that the flint collected
from nearby sources was generally small. All six of
the pieces that could be provenanced were beach
flint. As nodules only account for a very small
proportion of the assemblage it can be assumed that
most nodules underwent initial working closer to the
place of acquisition and that those brought to the site
were chipped to produce tools, resulting in only a
few nodules being lost or discarded in the hut. Test
pieces, sometimes referred to as ‘bashed lumps’,
account for 0.2% of the assemblage and their presence
indicates the working of unprepared nodules on the
site. Test pieces are characterised here as nodules that
Unretouched Flakes and Blades
The Howick assemblage is dominated by debitage
and unretouched flakes and blades, which together
account for 96.4% of all pieces. The debitage results
from all stages of the reduction sequence. Out of a
total of 10,093 flakes, 2617 (25.9%) were classed as
microdebitage, 7107 (70.4%) as debitage and 369
(3.7%) as blanks or possible debitage. The micro-
debitage mostly results from the tertiary stage of tool
production and has mean average length, breadth
and thickness measurements of 4.1mm, 3.5mm and
1.0mm respectively. The debitage has mean average
length, breadth and thickness measurements of
10.7mm, 8.6mm and 2.5mm respectively while the
remaining blanks have mean average measurements
of 19.8mm, 15.7mm and 4.8mm. These diminutive
average sizes result partly from the use of sieving to
recover small lithics, and partly from the small size of
the raw materials available. Many pieces still retain
areas of cortex on one or more facets, which further
indicates that the parent nodules from which they
were derived were of no great size. Given the small
size of most of the pieces it was difficult to ascribe
with any precision which stage of the reduction
sequence many of the pieces were associated with,
particularly as the presence of cortex is not an instant
indication of primary flaking in assemblages such as
this. This is confirmed by the fact that many of the
finished tools still have cortex surviving on their
surfaces.
Of the 2376 pieces characterised as blade forms,
113 (4.6 %) were classed as microdebitage, 1288
(54.2%) as debitage and 975 (41%) as blanks or
possible debitage. When the blank components of the
unretouched flakes and blades are compared there is
82
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
a striking contrast: there are only 3.7% flake blanks
compared with 41% blade blanks. This is indicative
of the flaking industry on the site which is evidently
geared towards the production of blade-based tools
of which the blanks, or perhaps more accurately
‘unretouched blades’, are a key component. The low
number of flake blanks results from the rest of the
flakes being largely the result of debitage from the
knapping process. The unretouched-blade com-
ponent demonstrates an industry reliant on narrow
parallel-sided small-blade technology for the pro-
duction of a range of tools. The use of a blade-based
industry has been associated with non-wasteful
production strategies and a light-weight tool kit
(Bradley 1987, 183).
that the curation of cores took place. These features
of the cores, together with the recycling of pieces and
the blade technology itself, point towards a par-
simonious attitude to discard. The small size of the
blade scars evident on most cores also reflects the
concern for producing microlith-sized blade blanks,
which is consistent with other northern Mesolithic
assemblages (e.g. Weyman 1984; Waddington 1999a).
Cores
A total of 107 cores was retrieved, which accounts for
0.8% of the hut total. Multi-faceted cores pre-
dominate, accounting for 35.6% of all cores, with
platform cores accounting for 27.1% and flakes re-
used as cores accounting for 24.3%. A further 6.5%
were pebble cores and the remaining 8.6% were core
fragments that could not be accurately attributed.
The detachment scars on virtually all cores showed
the intentional production of blades. Bi-polar work-
ing was apparent on a number of the cores, indicating
the use of anvil stones. Together with the presence of
the pebble cores these flaking techniques demonstrate
the adoption of strategies for coping with the small-
sized beach pebble flint. It is pertinent that only one
pyramidal-type core was found in the assemblage,
which could suggest that such core types are assoc-
iated with later developments in the Mesolithic flint-
working traditions of the region, or perhaps the raw
material again made flaking in this way difficult.
Indeed the nature of the raw material accounts for
the high percentage of multi-faceted cores, which
were clearly being struck this way to maximise the
amount of detachments from what are otherwise
irregularly shaped small pieces. Again, this is likely
to be a response to dealing with the type of raw
material available and hence explains the shortage of
regular-looking cores in the assemblage. The use of
flakes as cores is by no means restricted to northern
sites but the high proportion of flake cores (i.e. flakes
re-used as cores) (24.3%) in the assemblage reveals a
careful approach to husbanding this precious com-
modity. The mean average size of the cores is
consistent across all types; multi-faceted cores
averaging 29.4mm by 24.2mm, pebble cores 27.1mm
by 25.4mm, platform cores 26.0mm by 25.7mm and
flake cores 24.4mm by 22.3mm. Most of the cores are
very small and this is related to the paring down of
the material to the point where no further useful
detachments could be made. Core rejuvenation flakes
were identified in the assemblage, which confirms
Retouched Flakes
A total of 92 retouched flakes was identified, account-
ing for 0.7% of the assemblage. Of these, four were
edge-trimmed, one was microlithic in size, while
another was a possible scraper. Their mean average
length, breadth and thickness measurements are
18.5mm, 15.8mm and 5.6mm respectively, indicating
that most of these pieces are small; indeed the largest
only had measurements of 31mm, 38mm and 15mm.
As 64 (70%) of the retouched pieces are broken, the
functions of most of these pieces can only be guessed
at, although it is clear from their diversity that a wide
range of processing tasks was involved (see also
Chapter 9). Most of the pieces show signs of util-
isation visible to the naked eye, and together with the
high proportion of broken pieces, this indicates that
retouched flakes were utilised and discarded when
of no further use.
Retouched Blades
Fifty-one retouched blades were identified in the
assemblage, accounting for 0.4% of the total. Of these,
four were edge-trimmed, one was a possible awl and
the rest could not be ascribed to a specific tool type.
However, it is of some interest that most are only
retouched along one of their long edges. One piece,
[1027], has been retouched around its entire surviving
perimeter and may have been a knife but as it is
broken this is not certain (see Fig 7.2). Their mean
average length, breadth and thickness measurements
are 24.7mm, 10.5mm and 3.9mm respectively. Like
the retouched flakes, 67% (34) of the retouched blades
are broken. This makes understanding their func-
tional capabilities more difficult although undoubt-
edly some were cutting tools. Similarly, most of the
pieces show signs of utilisation visible to the naked
eye. The blades are all of regular parallel-sided form
and are typical of Later Mesolithic assemblages.
Utilised Flakes and Blades
Flakes and blades that were not formally retouched
but nevertheless had their edges worn as a result of
use have been categorised under the term ‘utilised
flake/blade’. A total of 14 utilised flakes was present
in the assemblage, accounting for 0.1% of the total.
Their mean average length, breadth and thickness
7CHIPPED STONE TOOLS 83
measurements are 22.6mm, 21.3mm and 6.4mm
respectively. Although these pieces have evidently
been used as tools it is not possible to suggest a use
on the basis of their visual appearance alone.
However, the use-wear and residue analyses have
shown that these pieces can be used for a range of
purposes (see below, Chapter 9). A total of 39 utilised
blades was present in the assemblage, accounting for
0.3% of the total. Their mean average length, breadth
and thickness measurements are 26.7mm, 11.7mm
and 3.7mm respectively. As with the utilised flakes,
although these pieces have evidently been used as
tools it is not possible in most cases to suggest a use
based on visual inspection alone, but the use-wear
indicates that some such pieces were used for cutting
fibrous plants, such as [963] [2151], and these have
been illustrated (Fig. 7.2).
they must have been hafted in some way although
quite what their use was remains unknown. Most of
the scrapers are abraded along their retouched edges,
indicating that they were used as functional tools
before being discarded. In all cases the retouch is
abrupt and unifacial, which produces a slightly
blunted edge. Although usually associated with the
preparation of hides, which is confirmed to some
extent by the residue and use-wear analyses, a more
diverse range of activities may also be associated
with the Howick implements, such as the working of
hard materials like wood, bone and antler (see below,
Chapter 9). If hide preparation was a key function of
scraper tools at Howick then their relatively low
number in the hut may indicate that this was a task
that mostly took place outside.
Notched Flakes
Four notched pieces were present in the assemblage,
accounting for less than 0.1% of the total. The mean
average length, breadth and thickness measurements
are 14.5mm, 11.0mm and 3.0mm respectively. Three
of the notched pieces are notched flakes, the other
being a notched blade. These pieces are thought to
have been notched for some kind of use that required
the angle of a trimmed indentation, though it is
possible that in some cases the notch may have been
associated with hafting the implement or the pro-
duction of microliths.
Awls
A total of 10 awls was present in the assemblage,
together with another possible example, which
together account for 0.1% of the total. Their mean
average length, breadth and thickness measurements
are 29.0mm, 13.0mm and 5.0mm respectively. These
pieces were some of the larger tools identified in the
assemblage and usually had marginal retouch along
all or part of one long edge up to the point (ills 2760
and 4793, Fig 7.2). These tool types are usually assoc-
iated with hide working, in connection with piercing
and stitching. Alternatively they may have been used
on other materials such as wood, bark or bone.
Scrapers
For an assemblage of this size the number of scrapers
is relatively low, with 109 pieces confidently assigned
to this category, which accounts for just 0.8% of the
assemblage total. Their mean average length, breadth
and thickness measurements are 16.9mm, 14.1mm
and 5.8mm respectively. A range of scraper types
could be discerned (see Fig 7.1) and these include:
tiny scrapers (28); small chunky end scrapers made
on short bulbous blades (19); abruptly retouched
thumbnail scrapers (3); pebble scrapers (2) and an
assorted group of pieces (51) that did not fit any
particular category, either because they were broken
or because they were opportunistically made on an
available flake or blade blank. A type of scraper that
has not previously been recognised in North-East
assemblages was a small ‘tongue’ or ‘tablet-shaped’
scraper made on a blade with abrupt retouch, regular
shape and measuring around 10mm long by 8mm
wide and 3mm thick (ills 4014, 4864, 3174 and 5081,
Fig. 7.1). Six examples of this type were identified in
the assemblage and are distinctive not only on
account of their shape and size but also because they
are retouched around virtually all their sides in order
to achieve the intended shape. They are so small that
Burins
Only four burins could be confidently assigned to
this category out of the entire assemblage, plus one
possible burin, which together account for less than
0.1% of the assemblage. Their mean average length,
breadth and thickness measurements are 26.0mm,
10.3mm and 3.0mm respectively although only three
survived unbroken. They took a variety of forms with
one broken blade having had what appears to be a
burin splinter detached from one of its long sides that
terminated in a hinge fracture. This same piece,
[2142], had been retouched along its opposite long
edge. Two of the other burins were more typical small
blades that had been snapped or struck at one end to
achieve a diagonal edge. One piece, [14695] (Fig. 7.1),
had been retouched along this edge while the other
piece, [13519], although it may just be an accidental
snap, makes a classic engraving piece.
Microliths
The microliths form the largest tool component in the
assemblage, totalling 275 pieces and accounting for
2.1% of the assemblage. This is perhaps not surprising
as microliths are usually regarded as forming parts of
84
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
composite implements that by their very nature
would require a greater number of lithic pieces to
render them effective. A range of microlith types was
present in the hut deposits and includes several
commonly associated with Later Mesolithic in-
dustries as well as an unusual form that has been
termed here ‘thick-edged’. A few of the microliths are
so small, such as [6324] and [9314], which have
maximum lengths of 7.5mm, that they must have
been used for very detailed work. However, being so
small these pieces do not conform to any typical
typological category. The distribution of the micro-
liths and their functional associations are dealt with
later in this section. It is notable that most are made
from flint although there are occasional examples
made from other stone such as scalene triangle
[13485], which is made from agate.
Scalene Triangles
As with the contemporary Mesolithic hut site at
Mount Sandel (see Chapters 6 and 15), the dominant
form of microlith in the Howick assemblage is the
scalene triangle. A total of 102 could be attributed to
this class, which forms 37.1% of the microlith
component. They form a cohesive typological group
with a narrow elongated form evidenced by the mean
average length, width and thickness measurements
of 17.8mm, 5.1mm and 1.7mm respectively. The angle
between the two retouched edges varies from very
acute to obtuse, though in most cases this angle is
close to 45°. Unlike the Mount Sandel pieces (Wood-
man 1985), these scalenes rarely have a concavity on
the short edge, this being straight in most cases. The
base of some of the scalene pieces appears to have
been snapped off to give them a squared profile at
the opposite end to the retouched angle. The dis-
covery of a scalene triangle still hafted to an arrow at
Lilla Loshult, Sweden (Petersson 1951), indicates that
in some cases this type of microlith was used as a
barb on projectiles. The proportion of left-shouldered
and right-shouldered scalenes is roughly equal,
which makes an interpretation of the scalenes as
armatures for projectiles with bi-serial rows of
microliths rather tempting. However, there is a
considerable range in size for the scalene triangles,
with the smallest measuring 9.5mm long and the
longest 30.5mm long. Indeed the scalene component
could probably be subdivided into different cate-
gories based on size and angle made by the two
retouched edges. Whether this would be instructive
is debatable but it would reinforce the point that
there is considerable variation in the size of these
pieces. It is likely that microliths were used in tools
that served a variety of purposes and not just in
hunting weapons (see also Chapter 9 below).
Backed Blades
A total of 54 backed blade microliths was present in
the assemblage, which accounts for 19.6% of the
microlith total. These small, generally narrow pieces
show some variation in shape and size and are
something of an eclectic group. They tend to have a
slender, somewhat rectangular or narrow leaf-shape
form, but they are not triangles or crescents and nor
do they form obvious points. Occasionally they may
have a crescentic shape but the retouch is along the
back edge and not the curving edge, which precludes
them from the latter category. These pieces have
retouch usually along all, or part, of the length of the
thickest edge and it is usually abrupt, although light
edge-trimming is evident on some pieces. One piece
of particular interest is [4410], which has a deliberate
notch towards the proximal end that may have served
a hafting purpose. The mean average length, width
and thickness measurements are 18.4mm, 5.7mm and
1.9mm respectively.
Crescents
A total of 21 crescent microliths was present in the
assemblage, which accounts for 7.6% of the microlith
total. These pieces are typically small in size and
have abrupt retouch along their curving edge. In
some cases, such as [9137] (Fig. 7.2) and [7814], there
was retouch on both the straight and curving edge
but these are the exception rather than the norm. The
crescents have mean average length, width and
thickness measurements of 15.4mm, 4.6mm and
1.7mm respectively. There is some variation in size
with the smallest piece measuring just 10.5mm in
length and the largest 25.5mm. They probably
compare most closely with the 20 crescentic micro-
liths from the Scottish east coast site at Fife Ness,
which are very similar in size and appearance and
come from a radiocarbon-dated site that is contem-
porary with the later phases of occupation at Howick
(Wickham-Jones and Dalland 1998). The authors of
the Fife Ness report suggested that the focus on
crescentic microlith forms at this site could be
associated with a particular specialist activity,
although there was no clear evidence for what this
might have been.
Points
Fifteen points were present in the assemblage, which
account for 5.5% of the microlith total. The most
striking type are the narrow elongated forms, re-
touched on both long edges, which can be equated
with Woodman’s ‘type A needle points’ from Mount
Sandel (Woodman 1985, 45). These needle points have
mean average length, width and thickness measure-
ments of 22.2mm, 6.1mm and 2.3mm respectively.
There are also smaller points that tend to be wider
and shorter but may also have retouch on one or both
edges. These more squat points have measurements
of 21.3mm, 6.3mm and 2.4mm respectively. The
retouch on all the points is usually abrupt, although
7CHIPPED STONE TOOLS 85
some light edge trimming has been used on certain
pieces, such as [2748] (Fig. 7.2), to achieve the desired
shape along the thinner long edge.
Other Triangles and Geometrics
Fifteen triangles and geometric microliths were
present in the assemblage, which account for 5.5% of
the microlith total. This group could be subdivided
into three types: isosceles triangles, trapezes and a
lanceolate. Twelve isosceles pieces were identified,
accounting for 4.4% of the microlith assemblage.
These pieces generally had a regular shape with two
sides of equal length and mean average length, width
and thickness measurements of 12.6mm, 5.1mm and
1.4mm respectively. However, there appeared to be a
divide between a small group measuring less than
10mm in length and a larger group measuring 14mm
or more. The retouch was abrupt in all instances and,
notwithstanding their more uniform shape, they were
very similar in form to many of the scalene triangles.
The two trapezes were both small and very similar in
size, with mean average length, width and thickness
measurements of 13.5mm, 5.0mm and 1.3mm respect-
ively. The single lanceolate piece was small, having
mean average length, width and thickness measure-
ments of 21.0mm, 6.0mm and 4.0mm respectively.
Geometric forms such as these are typical in Later
Mesolithic flint assemblages across the British Isles,
such as the two isosceles triangles recovered from
Filpoke Beacon on the County Durham coast (Jacobi
1976). Nearly all of these small geometric pieces were
recovered by sieving and if this had not taken place,
no strictly geometric component would have been
recognised in the microlith assemblage from Howick.
Other types
A single rod microlith was recovered from the site
and being less than 5mm in width would fall into the
‘narrow rod’ category as defined by Woodman (1985,
43). This piece, [10876], is retouched along the full
length of one side parallel to the main axis of the
blade. Both ends of the piece are squared off,
suggesting they were deliberately snapped to pro-
duce this effect or that they are broken.
A single example of a lamelles à cran microlith,
[3630], was found, with retouch that has thinned the
blade along one edge but, unusually, on the ventral
side. Further down the same edge on the surviving
original blade edge is a section of light edge trim-
ming, but this time on the dorsal side. These types
have been suggested as representing a stage in the
manufacture of scalene triangles (Brinch-Petersen
1966), but as there is additional edge trimming, this
suggests it could have served as an implement –
possibly used as a point. It measures 24mm long by
7mm wide by 2mm thick.
Two obliquely blunted microliths were present in
the assemblage. These types of microliths are usually
associated with Early Mesolithic assemblages al-
though they do sometimes occur in later assemblages
(e.g. Filpoke Beacon, Jacobi 1976, 71). These pieces
appear fresh and are likely to be contemporary with
the other Howick material. The mean average length,
width and thickness measurements for these pieces
are 15.0mm, 5.5mm and 1.3mm respectively.
A total of 65 unclassified microliths was present in
the assemblage, which account for 23.6% of the
microlith total. Most of these pieces were broken and/
or burnt and so could not be reliably attributed to
any particular microlith class. However, all were
made on narrow blades and were evidently of the
same narrow-blade tradition as the rest of the
assemblage. Retouch was abrupt in all instances and
some light edge trimming could be observed on a
few pieces. Two of the pieces may be examples of
Mèche de foret, resulting from re-pointing a tool such
as an awl.
Microburins
Nine microburins were present in the assemblage,
accounting for 0.1% of the total and therefore rare.
Being the snapped off ends of microlithic blades, their
mean average length, breadth and thickness measure-
ments are small: 7.9mm, 5.2mm and 1.4mm respect-
ively. For most of the pieces it is not clear from which
side they have been notched, as the bulbar end does
not survive, but in at least one unambiguous case the
notch had been made from the right when viewed
from the dorsal side. Given that there are equal
numbers of right and left-handed scalene triangles it
would be unlikely that a preference for notching one
side was found.
External Pits
A total of 58 lithics was recovered from the pits to the
south of the Mesolithic hut. Since they may be later
features with residual finds (Chapter 5), this material
is only briefly described as it is not considered to be
of particular interpretive value other than to docu-
ment its presence.
The lithic summary shows the usual predominance
of flakes and blades and both the platform core and
the backed blade microlith indicate the presence of
diagnostic Mesolithic material. The blade component
is large and shows a concern for small narrow
parallel-sided blade production, some showing
evidence of bi-polar flaking. All these flints compare
directly with the Mesolithic material from the hut
deposits and not one piece of possibly later material
was in evidence. The same variety of flint colours
was present, together with the same types of beach
and glacial flint; 79.3% of the material is broken and
8.6% shows patina development.
86
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Type Actual
As % of
Number
Assemblage
Unretouched Flakes
34 58.6%
and Blanks
Unretouched Blades
22 37.9%
and Blanks
Core 1 1.7%
Microlith 1 1.7%
Total 58
Table 7.3. Summary of artefact types from the pits outside the hut.
Context Flakes Blades Core Microlith Total
Pits
013 3 2 5
014 17 12 29
015 1 1
021 2 1 3
061 1 1
Suspected
Sheep
Burial Pits
023 2 2
037 2 2
039 1 1 2
081 4 1 1 6
083 1 4 1 6
085 1 1
Table 7.4. Summary of pit lithics by context.
Unstratified Material
A total of 901 lithics was recovered from unstratified
contexts around the Mesolithic site, mostly from the
topsoil (001) and subsoil (003) layers. Importantly,
the presence of a broken leaf-shaped arrowhead
demonstrates Neolithic activity on the site for which
no other trace has survived, although this piece may
reflect no more than casual loss. This is an invasively
retouched specimen made from a red-coloured flint.
Otherwise, the rest of the unstratified assemblage is
directly comparable to the Mesolithic material
recovered from the hut deposits. The same range of
raw materials is present, which includes 3 pieces of
chipped agate (0.3%), 8 chert (0.9%), 14 quartz (1.6%),
and the remaining 876 (97.2%) flint. Of all the flint
that could be provenanced, 185 pieces (84.1%) were
beach flint and 35 (15.9%) glacial flint. The same range
of colours is found in the unstratified material
including the predominant light grey shades, together
with dark grey, red-brown, white, orange, brown,
fawn, purple and pink. The range of tool types
accurately reflects the range of types recovered from
the stratified hut deposits, including scrapers, burins
and microliths (see Table 7.5 below). Of the eight
microliths recovered, two are scalene triangles, one a
crescent, one a backed blade and the remaining four
could not be classified. Six of the scrapers were tiny
scrapers and the other six abruptly retouched flakes
and blades, not able to be ascribed to any particular
classic scraper type. The retouch is abrupt on most of
the finished pieces although occasional edge trim-
ming was attested. One of the burins (1388) is worthy
of particular note as it is heavily patinated and made
on a broad blade and is likely to be of Late Upper
Palaeolithic or Early Mesolithic date. Being un-
stratified, this piece could be present as a result of
earlier activity on the site and together with a heavily
patinated retouched blade (2291) hint at an earlier
human presence on this site.
The lithic summary table shows the usual pre-
dominance of flakes and blades in addition to
Table 7.5. Summary of artefact types from unstratified contexts.
primary pieces, including collected beach nodules
and test pieces. The blade component (19.2%) com-
pares closely with the blade proportion in the
stratified hut assemblage (18.0%). This substantial
proportion shows a concern for the production of
small, narrow, parallel-sided blades directly an-
alogous with those from the hut. The flakes, blades
and associated debitage display working at all stages
of the reduction sequence, indicating the working of
nodules through to tools in the immediate environs
of the site. Bi-polar flaking is attested on a number of
pieces, suggesting the use of stone anvils and the
need for strategies that allow small nodules to be
utilised. Of the unstratified total, 64.9% was broken
although some of this is no doubt a result of plough
damage and other later disturbance. The truncation
of some of the hut deposits, particularly the upper
hearth pits, would no doubt account for at least some
of the 13.4% burnt lithics in the unstratified assem-
blage. One distinct difference with the stratified lithic
assemblage is the proportion of patinated material: in
7CHIPPED STONE TOOLS 87
the stratified assemblage, patinated pieces account
for 1.2% of the assemblage (see above), whereas in
the unstratified assemblage they account for 32% of
the assemblage. As the two assemblages are both
clearly Mesolithic and directly associated, it raises
the question why there should be such a profound
difference in the degree of patina development
between these two groups of material. Whatever the
specific reasons for this difference, it is evident that
the conditions in the sandy stratified deposits did not
encourage patina development, whereas those pre-
vailing in the humic sandy soils above appear to have
facilitated a distinctly more rapid development.
Stratigraphic and Spatial Distribution
Kristian Pedersen
Introduction
This study seeks to investigate the temporal and
spatial patterns in the lithic assemblage from the
Mesolithic hut deposits. It consists of three discrete
inquiries: firstly, a study of the metrical attributes of
the blades and microliths; secondly, an inquiry into
the composition of the assemblage by phase; thirdly,
an investigation of the spatial patterning in the
artefacts. As the dwelling comprised three structural
phases (1a–3), these studies are arranged in three
corresponding sections, followed by a comparative
analysis and then a summary of the main points.
Each of the inquiries is prefaced by a discussion of
the methods, principles and assumptions that under-
pin the analysis. The assumptions are based partly on
observations made at other Mesolithic sites in North-
West Europe, partly on ethnographic studies of
forager populations, and partly on analyses of
patterns produced in flint-knapping experiments.
All small finds encountered during the excavation
were labelled according to the deposit in which they
occurred and digitally plotted. All excavated sed-
iment was subjected to flotation and wet sieving
through a 2mm mesh and this produced a substantial
quantity of further lithic material (see above, this
Chapter), most of which was too small to have been
identified during excavation. The deposits from
which the latter material was recovered are therefore
known, but its precise location within these deposits
is not. The method by which the lithic material was
recovered did not affect the study of the metrical
attributes of the artefacts, nor the investigation of the
lithic assemblage by phase. It did, however, affect the
selection of material for the spatial analysis, since
debitage, given its small size, is not well represented
in the material that was digitally plotted and there-
fore those that were recorded in such a fashion are
unrepresentative as they form only a small portion
the overall debitage assemblage. This is also true for
the burnt flint, which was also largely recovered
through the flotation process as it tended to be small
in size.
Not all classes of the larger artefacts are sufficiently
prevalent for a statistically meaningful inquiry based
on their spatial patterning to be conducted (e.g.
scrapers, awls, burins). Furthermore, some classes of
artefact are so poorly represented that their occur-
rence in an assemblage has little significance (e.g. the
single ‘rod’ microlith). This problem is to some extent
reduced as not all artefacts have the same significance
for the reconstruction of past activities. The blades
and microliths are two classes of artefact of foremost
importance in the study of Mesolithic assemblages.
Because the lithic industry at Howick was blade-
based, it is assumed that blades, and those tools based
upon them, were accorded special attention in order
to achieve consistent form and proportions, as well
as in their placement and handling during their
production and use. The microliths are of special
interest because of the social significance that they
may contain: ethnographic studies provide examples
of hunter-gatherer populations using the idiosyn-
cratic stylistic attributes of their projectile points to
express kinship and tribal affiliations (Sackett 1990;
Wiessner 1983; 1990). Based on idiosyncrasies in
microlithic armatures observed in the Low Countries
(Gendel 1984), and subtle changes in the shape of
flake axes noted along the Øresund littoral in
Denmark (Vang Petersen 1982), the existence of
discrete social territories has been inferred. It might
also be expected that because of the potential
significance of these artefacts in social life they were
fashioned with distinctive attributes, and thus
conformed to a stylistic template more closely than
any other types of tool preserved on the site. The
identification of such patterns cannot be achieved
through the study of a single site, as the idiosyn-
crasies only become apparent when contempor-
aneous assemblages throughout a given region are
compared. It is however worth bearing in mind that
these artefacts are likely to have had a significance
beyond that of armatures for hunting weapons and
that their production was governed by more than
merely functional considerations. It has already been
referred to above how this assemblage fits into the
‘micro-triangle’ technocomplex identified at other
contemporary North-East British coastal sites such as
Filpoke Beacon (see also Jacobi 1976).
An assumption applied to the spatial analysis
undertaken here is that blades are laid out in either
concentric patterns around the artisan whilst he is in
the process of producing these implements, or
arranged to one side (according to whether the artisan
was right- or left-handed), in arc-shaped rows:
88
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
‘One phenomenon often seen in flint knapping exper-
iments is that knappers behave differently with respect
to blades than to the flint waste of flakes and chips …. the
knapper may collect the blades in front of him, placing
them more or less in a semi-circle. Or he may place them
to his left or to his right, often in arc-shaped rows. Small
flakes and chips, on the other hand, just fall to the ground,
or are brushed off the knapper’s lap. Blades are what the
knapper is aiming to produce, so it is only natural that
they should be treated with special care and attention.’
(Johansen and Stapert 1998, 34).
The patterning in the distribution of blades shall
therefore serve as the cornerstone of the attempt
made here to reconstruct the locales at which flint
knapping was undertaken within the hut, whereas
the distribution of flakes and other types of lithic
material is of only subsidiary significance in this
respect. The microliths are, however, important since
they were treated similarly to blades; their arrange-
ment is similar, as they are subject to retouch and
arrayed for hafting into bone or wooden points. The
distribution of blades and microliths is considered
concurrently with cores, which are customarily
deposited in the immediate vicinity of flint-knapping
activity. The distribution of formal tools may not be
informative in connection with the location of flint-
knapping events, but they may provide indications
of where activities associated with their use took
place. Only scrapers occurred in sufficient quantity
to render such an inquiry meaningful, and because
these tools are typically associated with the prepar-
ation of hides, this affords important evidence since
the faunal remains indicate the presence of seal and
other mammals such as pig and fox (see Chapter 10),
which, in addition to providing meat, would have
been a valuable source of skin and fur.
Although the microliths are an important com-
ponent of the assemblage, they are not considered
separately according to type, as all types of microliths
occurred throughout all stratigraphic phases and so
no typological difference through time is apparent.
When the different types of microliths were plotted
spatially, no difference could be observed between
the spatial patterning of microliths of different types.
Apart from the scalene triangles and backed blades,
the rest of the microliths occurred in such small
quantities that they were too few to allow for any
statistically significant results to be obtained anyway.
As a consequence, the microliths were considered
together throughout this analysis.
The spatial analysis also relies on studies of the
distribution of flint produced by modern flint
knappers. A few remarks concerning the nature of
the flint resources available to the inhabitants of the
Howick hut are necessary by way of preface to this
discussion, as the quality and quantity of the raw
material imposed constraints on both the metrical
attributes of the artefacts and on the distribution of
lithic waste. Many of the models derived from the
distribution of flint waste from tool production have
been undertaken by flint knappers in flint rich
regions, or by flint knappers who have access to flint
and chert nodules of good quality and reasonable
size. The inhabitants at Howick did not have such
resources: they had to rely on small pebbles found on
the shore, which were often riddled with flaws. The
size of the pebbles renders it difficult to use the antler
punch for the manufacture of blades and therefore
alternate methods, less likely to produce regular
blades and discrete waste patterns, were employed.
Amongst these methods was direct percussion; the
prevalence of this technique in the production of
blades is indicated by the crushing on the platform
and by the absence of a pronounced lip on the bulb of
production (see also above, this Chapter). Another
method employed was the ‘hammer-and-anvil’, or
‘bi-polar’ technique, which consists of placing the
flint pebble on an anvil stone and striking the top
with a stone or antler billet and thereby detaching
blades. This is an effective method for the production
of blades from small pebbles, but it does not result in
the manufacture of regular blades with the same
consistency as the use of an antler punch. Moreover,
the distribution of flint waste is also less regular.
It is also important to bear in mind that, given the
paucity of good quality flint, the tools might have
been resharpened and transformed into other types
of tools as they broke and became increasingly worn
(see above, this Chapter). The metrical attributes of
the microliths and the blades might, therefore, not be
directly comparable to those in contemporaneous
assemblages from regions benefiting from more
copious quantities of flint of good quality. In regions
where high-quality flint was prevalent, a more
profligate use of tools is observed. For instance, in
Denmark during the Later Mesolithic, micro-wear
analysis has determined that long and regular blades
were often used on only one or two occasions and
then discarded as waste (Juel Jensen 1986). Since most
flint knappers involved in experimentation work in
flint-rich regions, a perfect correspondence in waste
patterns between that observed on the occupation
floors at Howick and that replicated in experiments
cannot be expected. Nevertheless, the general pat-
terns are likely to be comparable, which is the
justification for the comparisons below.
The main assumption regarding the use of domestic
space in this study is that the hearth was the focus for
internal activities, including flint knapping. The hearth
as a focus of domestic life is partly explicable by
practical considerations, such as the warmth, light and
cooking requirements, which it affords to inhabitants
whilst they undertake various tasks. Nevertheless,
symbolic considerations must also be entertained, as
the ethnographic record is replete with examples of
houses representing microcosms, and with specific
7CHIPPED STONE TOOLS 89
features bearing religious significance (e.g. Eliade 1959;
1964). Other aspects pertaining to the use of domestic
space in small structures amongst hunter-gatherers
involves the segregation of families on either side of
the hearth (Grøn 1989; 1999). The segregation of
activities, in accordance with the observance of taboos,
must also be considered because some populations,
such as the Netsilik Eskimo, strictly avoided the
processing of animals living in different environments
(such as terrestrial and marine beasts) in the same area
(Balicki 1962). An even more complex series of taboos
occasioning the spatial segregation of activities, and
even the sex of those permitted in certain precincts or
structures, is encountered amongst the Pacific peoples
like the Kwaio (Keesing 1982). Although the inferences
ventured in this report principally concern industrial
and economic behaviour, some profitable courses of
inquiry are suggested that might elucidate social and
symbolic practice.
Methods
The principal objective of the metrical analysis is to
attempt to identify distinctions in the blades and
microlithic armatures from each phase of the Meso-
lithic occupation, as this might provide a basis for
observing technological developments and/or
changes in activities carried out during succeeding
phases. A metrical characterisation of these im-
plements will facilitate a comparative analysis of the
assemblage from Howick with those from other
contemporaneous sites in the British Isles and North-
West Europe. The attributes of blades and microliths
are those most commonly calculated from Mesolithic
assemblages, and indeed, the study of the typological
attributes of microliths has traditionally formed the
foundation for chronological studies, as well as the
arrangement of material into discrete industries,
typically regarded as revealing the degree of social
interaction or cultural affinity amongst regions.
The availability of raw material from which the
lithic tools were produced did not change through
the period in which the dwelling at Howick was
occupied, nor did the chaîne opératoire. It is therefore
assumed that there would be little probability of a
change in the metrical attributes of the debitage; the
material most likely to produce any diachronic
change would be blades and microlithic armatures,
since their form was governed by stylistic consider-
ations more than the technical constraints imposed
by the reduction methods and the characteristics of
the raw material. However, no such changes are
evident.
A clear pattern emerges in the length-width
relationship of the blades. However, as this class of
artefact is customarily defined by the length ex-
ceeding the width, it is necessary to tease out patterns
which might have stylistic and industrial significance,
rather than those that merely express the class-
ification criterion employed (Andrefsky 1998, 69).
Moreover, it is also necessary to contend with the
variability in the blade assemblage introduced by the
decision not to adhere to a strict length-width ratio
for the classification of the artefact as a blade. Instead,
the intention of producing a blade was deemed most
important. This approach seemed justified because of
the nature of the lithic raw material available to the
Mesolithic artisans, which consisted overwhelmingly
of small, flawed beach pebbles. Such material is most
effectively worked by the hammer-and-anvil tech-
nique which, when used on lithic material of this
quality, results in the production of irregular blades.
The ability to consistently produce blades from this
material with a length-width ratio of 2:1 or more (the
traditional metrical criterion used to distinguish
blades from flakes) is thus less likely than with better-
quality flint resources. With regard to these points,
the decision to make intuitive judgements on the
intention to produce blades is no more arbitrary than
an insistence that blades might be distinguished from
flakes by a minimum length-width ratio. Further-
more, it is more likely to approximate the actual
quantity of blades attempted whilst flint knapping at
Howick than the imposition of metrical criteria
developed to define this class of artefact from regions
that provide better flint resources.
Less equivocal is the identification of microliths,
for the retouch applied to one or more of the sides
clearly distinguishes this class of artefact from
ordinary blades and flakes. Nevertheless, in rare
instances there is scope for uncertainty in dis-
tinguishing these artefacts from scrapers, awls, and
so forth because of discontinuous retouch or the
anomalous shape of the blade upon which they were
produced. A key assumption of this study is that not
all blades were regarded as suitable for the production
of microliths – only those with specific proportions
were typically selected, and therefore the microlithic
armatures will tend to aggregate in clusters with
metrical attributes diverging from the main body of
the blades. The degree of this divergence is pre-
sumably governed by the difference in the blades
commonly produced and the characteristics regarded
as desirable for the production of microliths.
The statistical methods employed to identify
significant metrical attributes involves the calculation
of the correlation coefficient, mean lengths and
widths along with standard deviations, and cluster
analysis. The calculation of correlation coefficients
involves the use of the commonly applied regression
formulae such as the Pearson correlation (Shennan
1997), which indicates whether the correlation is
significant. This is undertaken on the basis of bi-
variate correspondences, since only two variables
90
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
(length and width) are relevant in this investigation.
The Pearson correlation assumes that a linear rel-
ationship exists between length and width; this is a
justifiable assumption in the study of blades, as this
class of artefact is usually defined by this linear
relationship. At a 0.05 level of significance, the nearer
the correlation coefficient is to 1, the more significant
the relationship (Shennan 1997, 127ff.).
The second method that is employed is that of
cluster analysis. This seeks to arrange the material
into discrete clusters, or groups of finds. A more
detailed discussion of the assumptions underpinning
the use of the K-means algorithm occurs in the section
pertaining to spatial analysis (see below), but it is
apposite here to mention the use of hierarchical cluster
analysis. This method seeks to arrange the material
into groups, and is therefore ‘agglomerative’; the
clusters are most clearly identified by using dendro-
grams. These plots are typically extremely large, and
therefore they are not presented in either the text or
appendices; however, the agglomerative schedules are
offered in tabular form as these represent visual
depictions of the clusters on a manageable scale. The
principal concern in using the hierarchical cluster
method is at what scale the clusters are significant,
since many small clusters with a population of only
one or two cases are identified. A certain subjectivity
must come into play here – the visual depictions in
the dendrograms permit the selection of the most
inclusive clusters. A more objective selection of
clusters is not possible for two reasons: firstly, and
most important for the purposes of this inquiry, is
that there is no model of blade metrics against which
the material might be tested. The experimental studies
of blade production have chiefly been focussed on the
metrics and spatial distribution of this class of artefact
produced on flint of much better quality than that
available to the artisans at Howick, and also on large
nodules. Secondly, the size of the pebbles available at
Howick precludes the use of the antler-punch tech-
nique, which results in the manufacture of long,
regular and elegant blades. The blade clusters selected
were therefore those that encompassed groups that
formed a statistically significant sample of the overall
assemblage for which metrical attributes could be
determined. Small clusters of no statistical significance
that remained were nevertheless kept, since these rep-
resent outlying examples and to seek to subsume these
in the other clusters would skew the results.
Phase Number of
Number of
Percentage of Blade
Blades
Complete Blades
Assemblage
1 1164 181 6.43
2 511 63 12.32
3 649 101 15.56
Table 7.6. Complete and fractured blades by phase.
phase of the occupation, as most of them were broken
and their original dimensions are unknown. There-
fore, the blades investigated here only form a portion
of the total blade assemblage in each phase, ranging
from 6.43% in Phase 1 to 15.56% in Phase 3 (Table
7.6). The largest quantity of artefacts was recovered
from Phase 1 (a and b) and this also provided the
largest quantity of complete blades. The proportion
of complete blades expressed as an overall percentage
of blades within the phase is, however, lowest in
Phase 1. No obvious reason for this suggests itself,
nor is it apparent why Phase 3 provides the largest
proportion of complete blades. The tendency for the
percentage of complete blades to increase through
time, as the duration of the structure’s use becomes
shorter (see Chapter 6), does imply that a correlation
between the fracturing of blades and length of
occupation exists.
The metrical attributes of all the blades from
Howick are depicted in Figure 7.3. A positive
relationship between length and width is immed-
iately apparent, but it is not apparent whether this
merely expresses the selection criterion. Moreover,
another salient aspect of the assemblage is that there
is considerable overlap, with a tendency for the
blades to be short (<20mm in length) and thick
(<15mm in width). The clustering of blades into
groups with specific metrical attributes is explored
by employing cluster analysis, of which the K-mean
algorithm shall assume a prominent role. For the
moment, however, it is worth remarking on the
characteristics of the overall blade assemblage. The
Metrical Analysis
Blades
The number of blades suitable for metrical analysis is
less than the total quantity of blades present in each
Figure 7.3. Length-width relationship of blades from all phases.
7CHIPPED STONE TOOLS 91
Pearson Correlation Coefficient of 0.618 demonstrates
a significant linear relationship, calculated at the 0.01
level, between the length and width of the blades.
The blades have a mean length of 16.3971 mm and a
mean width of 8.4043 mm; the standard deviations
are 9.34156 and 5.11577, respectively. There is, then,
considerable variability in the metrical attributes of
this assemblage. The most appropriate method for
identifying significant similarities is therefore by
arranging the material from each phase into clusters
and then comparing them.
Phase 1
As discussed above, the number of complete blades
expressed as a percentage of the overall blade
population is least here of all the assemblages. A
cursory glance at the length-width relationship of the
blades presented in the scatter plot in Figure 7.4
reveals that many of the blades are short (<20 mm)
and wide (<10 mm). The Pearson Correlation Co-
efficient (0.615) shows, firstly, that there is a signif-
icant linear relationship between length and width;
secondly, that this assemblage closely conforms to
the attributes that characterise the blades from all the
phases of occupation.
The hierarchical cluster analysis, which seeks to
group the material in hierarchical groups, identified
six clusters. A K-means test was then run on the basis
of the existence of these six clusters. The number of
blades in each of the clusters is presented in Table 7.7,
whereas the metrical centres of these clusters is
presented in Table 7.8. The most striking aspect of the
blade clusters is that the largest cluster (4) is
characterised by short and wide blades, with a length-
width ratio of just slightly under 2:1. Indeed, half of
the clusters (4, 5, and 6) are distinguished by a length-
width ratio of less than 2:1. This does not seem to
reflect a stylistic preference, as the metrical attributes
of the microlithic armatures, which were based on
selected blades, differs slightly. Rather, this pattern
was most likely governed by the raw material
availability because it is essentially repeated in each
phase. Moreover, the largest clusters fall within the
same general length-width range; these could hardly
be regarded as affording the most desirable blade
metrics, given they were so short and thick and that
the microlithic armatures seemed to have been, as a
rule, selected from longer and more narrow blades.
Phase 2
The metrical attributes of the blades in Phase 2 are
fundamentally similar to those in Phase 1, although
there is some uncertainty concerning the provenance
of those blades occurring in levelling deposits [115],
[118] and [210]. This uncertainty occurs because the
levelling deposits might have introduced material
from the preceding occupation horizon, or from
Figure 7.4. Length-width relationship of blades from Phase 1
(Pearson Correlation = 0.615).
Phase Mean
Length
Mean
Width
Standard
Standard
Deviation
Deviation
(Length)
(Width)
1 15.5939 8.5387 8.88848 5.46439
2* 15.1296 6.5000 9.92917 4.60769
2 16.2222 7.5952 9.36008 4.69974
3 18.0594 8.7129 10.15906 4.69007
All 16.4275 8.4203 10.15906 4.69007
Table 7.7. Mean length and width of blades from all phases (* Denotes
calculations made without material from the levelling deposits).
material introduced into the hut (see Chapter 5). If
material was introduced to the hut, the metrical
attributes would presumably be largely similar since
this material is contemporaneous with the occu-
pation; if the material was chiefly introduced from
the former source, the metrical characteristics of the
blades from these deposits might more closely
approximate to those from Phase 1.
In order to address this problem the metrical
attributes of the assemblage were characterised in the
following way. Firstly, by calculating the mean length
and width of the blades from the overall assemblage,
then secondly calculating the same measurements
from the material that does not derive from the
levelling deposits, and then thirdly calculating the
same measurements from the material within the
levelling deposits. An attempt to identify groupings
in the results from these analyses could then be made.
The blade assemblage as a whole is broadly similar to
that in Phase 1, and, indeed, with the material from
the Phase 3 occupation horizon. This can be seen in
Table 7.9, which presents not only the calculations for
mean length and width but also the standard dev-
iations in these values.
The cluster analysis of the blades occurring in only
those deposits that were not part of the levelling
spreads reveals five groups. The largest group
92
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 7.5. Length-width relationship of all blades in Phase 2
(Pearson Correlation = 0.650).
contains sixteen specimens, followed by a group
containing only seven, whereas the remaining groups
consist of only one or two specimens. As with all the
other phases, the richest cluster is distinguished by
blades that are relatively short and thick, with a
length-width ratio of slightly less than 2:1. The second
largest cluster is centred on material that has a ratio
of slightly over 2:1. The remainder of this material
represents outlying artefacts with metrical attributes
that diverge substantially from the main body of
finds.
A very strong correlation between length and
width is observed when only the material from the
levelling deposits in Phase 2 is considered. This is,
indeed, the strongest linear correlation seen in any of
the material subjected to a metrical analysis. Never-
theless, the grouping of blades into discrete clusters
follows the pattern already established, that is to say,
the largest cluster consists of material that is distin-
guished by a length-width ratio of roughly 2:1 and is
generally short and thick. Therefore, it is not possible
to determine whether this material consists largely of
material from the preceding phase or was instead
deposited whilst the levelling deposits served as an
occupation floor.
Figure 7.6. Length-width relationship of blades from Phase 2,
excluding those from the levelling deposits (Pearson Correlation =
0.589).
Phase 3
The visual impression given from the length-width
relationship of the blades in this phase is that several
discrete groups exist and that there is considerable
variability in the metrical centres of the groups. This,
however, is not immediately suggested by the
Pearson Correlation Coefficient, which displays a
strong linear relationship between length and width.
Again, three clusters were identified; the first two of
these contained the largest amount of material,
whereas the third represents an outlying metrical
centre. Some metrical variation is observable in Phase
3, as the largest cluster is centred on longer blades
than in the preceding occupation phases. Never-
theless, the length-width ratio of these blades remains
centred on a point that is still slightly less than 2:1.
The arrangement of the material from levelling/
occupation deposit [049] into clusters shows the same
tendency. Only three clusters were identified, the
largest of which consisted of material centred on
blades that were longer than those in the preceding
phases, but still provided a length-width ratio of
slightly less than 2:1.
Figure 7.7. Length-width relationship of blades from levelling
deposits in Phase 2 (Pearson Correlation = 0.726).
Microliths
The microlithic armatures display relatively little
metrical variability, a tendency which is obvious both
in the calculations of their mean length and width and
the concomitant standard deviation in both these
attributes, as well as their distribution in the scatter
Figure 7.8. Length-width relationship of blades in Phase 3 (Pearson
Correlation = 0.633).
7CHIPPED STONE TOOLS 93
diagram. One immediate impression is that the
standard deviation in the length of the armatures is
roughly half that of the blades, whereas the deviation
in width is only slightly less than that observed in the
blades. Moreover, the mean length is significantly
greater amongst the microlithic armatures. The
cumulative effect of these observations is that the arm-
atures were fashioned with greater metrical consist-
ency than the blades. This might be explained by the
fact that only specific blades were selected for further
manufacture into microliths; thus, the main cluster of
blades does not represent those with the most desir-
able characteristics for the production of armatures.
Before describing the metrical analysis, it is worth
considering the possible causes of the breakages in
the microlithic armatures. The fractures probably
occurred because of use, being trodden upon, or
during their manufacture. Breakage during hunting
has been investigated in other studies of fracture
patterns in microlithic armatures (e.g. Fischer et al.
1984; Friis Hansen 1990), but the microliths were
obviously not used for hunting within the precincts
of the dwelling. The high occurrence of fractured
implements cannot therefore be attributed to this
Figure 7.9. Length-width relationship of blades in levelling deposit
[049] from Phase 3 (Pearson Correlation = 0.692).
Figure 7.10. Length-width relationship of all microliths (Pearson
Correlation = 0.310).
Phase Number of Microliths Number of Complete Microliths Percentage of Microlith Assemblage
1 110 31 28.18
2 62 18 29.03
3 96 33 34.37
Table 7.8. Number of complete and fractured microliths.
Phase Mean Length Mean Width Standard Deviation (Length) Standard Deviation (Width)
1 20.1613 8.8387 4.45419 4.44107
2 16.9722 7.0000 4.34021 3.97418
3 17.6061 7.5152 4.75125 3.57833
All 18.4329 7.9024 4.70725 4.03185
Table 7.9. Mean length-width relationship of microliths.
94
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
activity. The most likely explanation of the fracturing
must be that they broke whilst being retouched or
repaired, that broken projectiles were brought into
the dwelling in the wooden or bone tools in which
they were hafted for subsequent replacement, or that
they were trodden upon. Little indication that they
were transported into the hut lodged within animal
bone is forthcoming, but this does represent a
possibility – consider the projectiles embedded in the
bone of aurochs from Vig and Prejlerup (Aaris-
Sørensen 1984), and that lodged in the elk at Skotte-
marke and Favrbo (Møhl 1980). Such an explanation
would not necessarily account for all the broken
microlithic armatures, but it is certainly possible that
some were introduced to the structure in this manner.
Being trodden upon remains the most likely explan-
ation, but a resolution of this matter must await a
study of the fracture patterns.
microlithic armatures. To be sure, some relatively
broad blades were selected, but the vast majority of
those selected for manufacture into armatures were
long and narrow.
Phase 1
The microlithic armature assemblage from Phase 1
was the only one of the three that revealed a
statistically significant linear correlation between
length and width, calculated by the Pearson Correl-
ation Coefficient at a 0.05 level of significance.
Although the relationship is significant, the sample
size for each cluster is quite small.
Four clusters are apparent in this material. Three
(2, 3 and 4) contain a similar quantity of material,
whereas the outlying one (1) consists of an anomalous
broad armature with a length-width ratio close to 1:1.
The armatures in the largest cluster have a relation-
ship of nearly 3:1; the others range between 2:1 and
nearly 3:1. This underscores the disparity in the
metrical attributes between the main body of the
blades and those selected for further production of
Phase 2
The levelling deposits in Phase 2 contained most of
the microlithic armatures; only 5 of the 18 complete
microliths were encountered in other deposits. Such
a small sample cannot be profitably investigated by
cluster analysis and, moreover, any metrical char-
acteristics identified would not be statistically
significant. All the microlithic armatures, regardless
of the deposit from which they derived, have
therefore been collapsed together in this inquiry. One
aspect of this assemblage is immediately apparent:
there is no statistically significant linear correlation
between the length and width of the microlithic
armatures in Phase 2, when calculated by the Pearson
Correlation Coefficient at a 0.05 level of significance.
This might represent two phenomena, the first being
that there is simply no correlation, the second being
that two discrete clusters distinguish this assemblage.
Both of these possibilities can be resolved through an
analysis of the clusters into which the microlithic
armatures might be arranged.
Of the four identified, only two clusters (2 and 3)
are significant. The other clusters represent outliers
that could not be encompassed in the first clusters
mentioned above without skewing the results. It is
apparent that the microlithic armatures in this
assemblage tend to be long and narrow, with a ratio
of between 4:1 and 7:1. Although no linear relation-
ship could be seen in this assemblage, it corresponds
well with the metrical attributes of clusters identified
in the microlithic armatures from the other phases.

Figure 7.11. Length-width relationship of microliths, Phase 1
(Pearson Correlation = 0.429).
Figure 7.12. Length-width relationship of microliths, Phase 2
(Pearson Correlation = 0.212).
7CHIPPED STONE TOOLS 95
Phase 3
This phase is noteworthy for the lack of a statistically
significant linear relationship between the length and
width of the microlithic armatures. Since no micro-
lithic armatures were forthcoming from Phase 3 apart
from levelling deposit [049], this study concerns only
what might possibly be a redeposited layer. The
cluster analysis does differ significantly from that
undertaken for the other phases: the largest cluster is
centred on much broader armatures, with a length-
width ratio of only slightly over 2:1.
Temporal Analysis
In this section, the composition of the assemblages
from each phases of occupation is described and
compared. This analysis relies chiefly on the present-
ation of the percentages of certain classes of artefact
expressed as a proportion of the overall assemblage,
or calculated against primary and secondary debit-
age. The artefact frequencies from each phase are
presented together, to facilitate comparison in tabular
form.
The most striking feature of the assemblage com-
position from each phase of occupation is their
fundamental similarity. All the assemblages consist
overwhelmingly of debitage, followed distantly by
microliths, scrapers, cores and retouched flakes. The
high counts of debitage are clearly related to the
Figure 7.13. Length-width relationship of Microliths, Phase 3
(Pearson Correlation = 0.138).
Artefact Type Number Percentage of
Assemblage
Nodules 1 0.03
Test Pieces 7 0.24
Cores 15 0.44
Debitage 2566 94.42
Retouched
21 0.74
Flakes
Retouched
3 0.13
Blades
Utilised
Scrapers 20 0.73
6 0.22
Blades
Utilised Flakes 17 0.63
Awls 1 0.03
Microliths 62 2.28
Burins 3 0.11
Table 7.11. Occurrence of artefact types, Phase 2.
Artefact Type Number Percentage of
Assemblage
Nodules 8 0.13
Test Pieces 5 0.08
Cores 61 1.02
Debitage 5616 94.27
Retouched
40 0.67
Flakes
Retouched
26 0.44
Blades
Scrapers 62 1.03
Utilised
16 0.27
Blades
Utilised Flakes 6 0.1
Awls 6 0.1
Microliths 110 1.84
Burins 3 0.05
Table 7.10. Occurrence of artefact types, Phase 1.
Artefact Type Number Percentage of
Assemblage
Nodules 2 0.08
Test Pieces 9 0.21
Cores 27 0.62
Debitage 4012 94.39
Retouched
29 0.68
Flakes
Retouched
18 0.44
Blades
Scrapers 25 0.62
Utilised
16 0.38
Blades
Utilised Flakes 2 0.07
Awls 2 0.07
Microliths 96 2.25
Burins 7 0.19
Table 7.12. Occurrence of artefact types, Phase 3.
96
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
comprehensive sieving strategy that was implemented
at Howick. Comparisons with other assemblages are
thus inappropriate, leaving no directly suitable
analogues for this particular study. The most obvious
difference amongst the assemblages, showing a clear
diachronic trend, is the proportion of primary and
secondary debitage, calculated as a total of overall
debitage. The association of the highest proportion of
primary debitage with the highest occurrence of cores
cannot be fortuitous. This is puzzling, principally
because there is no other indication (apart from the
proportion of the debitage assemblage comprised of
primary and secondary) that the primary reduction
sequence is more prevalent in Phase 1 than in other
phases. Indications of this must be sought in the spatial
analysis, since none are forthcoming in the metrical
inquiries. The proportionate increase in secondary
debitage is similarly explained, as the decline in
primary flint working was accompanied by a corres-
ponding increase in secondary reduction work. The
only interpretation of these trends that immediately
suggests itself is that primary reduction was under-
taken more often in Phase 1 and that this subsequently
declined. The only corresponding change is that
microliths constituted an increasing proportion of the
overall assemblage.
Phase 1 contains more material than any other
structural phase, totalling 5964 lithic artefacts, the
vast majority of which (94.27%) consisted of lithic
debitage. This is consistent with the debitage propor-
tions of the assemblages from the succeeding phases
of occupation. The remainder of the assemblage
consisted of tools and cores; only the microliths,
scrapers and cores exceeded 1% of the total.
Figure 7.14. Relationship of primary and secondary debitage amongst phases.
Artefact Type Phase 1 Phase 2 Phase 3
Nodules 0.13% 0.03% 0.08%
Test Pieces 0.08% 0.24% 0.21%
Cores 1.02% 0.44% 0.62%
Debitage 94.27% 94.42% 94.39%
Retouched Flakes 0.67% 0.74% 0.68%
Retouched Blades 0.44% 0.13% 0.44%
Scrapers 1.03% 0.73% 0.62%
Utilised Blades 0.27% 0.22% 0.38%
Utilised Flakes 0.1% 0.63% 0.07%
Awls 0.1% 0.03% 0.07%
Microliths 1.84% 2.28% 2.25%
Burins 0.05% 0.11% 0.19%
TOTAL 100 100 100
Table 7.13. Comparison between the three phases showing the percentage composition of the flint assemblages.
Spatial Analysis
Intra-site spatial analysis has developed largely
through the investigation of Late Palaeolithic and
Mesolithic sites in North-West Europe (e.g. Blank-
holm 1985; 1987; 1993a; 1993b; Grøn 1983; 1987; 1988;
1989; 1990; 1992; 1994; 1995; 1999; Grøn and Sørensen
1993; 1995; Karsten and Knarrström 2001; Lass Jensen
2001; Stapert et al. 1986; Stapert and Krist 1990;).
Studies of forager sites in the Americas, Australasia
and Africa, however, have also contributed signif-
icantly to the elaboration and refinement of methods
and principles used in these analyses (e.g. Whallon
1973; 1974; Yellen 1977; Binford 1978a; 1978b; Price
1978; ), as have those pertaining to the Lower and
Middle Palaeolithic. The development of methods
appropriate for the identification of patterns in the
distribution of artefacts represents a preliminary
objective, whereas the principal task of interpretation
typically involves recourse to anthropological and
psychological theories that might elucidate the
behavioural significance of the patterns observed
(Grøn 1989; 1995). It is worth emphasising that intra-
site spatial analysis is not concerned exclusively with
the interpretation of the spatial patterning of material
occurring within domestic structures: it is more
generally concerned with the identification of areas
on sites given over to specific activities (e.g. Brinch
Petersen 1989; Houtsma et al. 1996; Baales 2001;).
Much of the discussion concerning the intra-site
spatial analysis of dwellings has, however, been
concerned with identifying huts through the distrib-
ution of lithics and the arrangement of this material
around features such as hearths (Whallon 1978; Grøn
1983; Blankholm 1985; 1987; 1993b). The methods
used in such analyses are of only tangential interest
and relevance to the inquiries here because the
presence of discrete features within a clearly de-
limited dwelling removes any need to use methods
designed to infer the existence of a hut.
Ethnographic studies have, of course, figured
prominently, but these are often of only limited value
for the understanding of the use of space in domestic
structures amongst hunter-gatherers because this
topic has been of only marginal interest to ethno-
graphers, with only occasional studies, such as those
undertaken in Siberia (Jordan 2001; 2003). The
ethnographic analogues that do exist are principally
derived from the study of populations in the North-
ern Hemisphere, although an argument might be
made that the traditions of people occupying the
same basic environments are more relevant to the
study of the North-West European Mesolithic than
those of hunter-gatherers resident in the Kalahari or
the Amazonian rain forest.
7CHIPPED STONE TOOLS 97
Statistical Methods
The discrimination of artefact clusters occurring in
each phase of the occupation at Howick is attempted
by using the K-means clustering algorithm. This
method is appropriate for the analysis of large
quantities of data that are distinguished by few
variables, and when only a small number of clusters
are expected (Hodson 1971, 31). As the distribution
of lithic material is plotted according to only four
variables (x,y,z co-ordinates and nominal scale
identifier), few clusters are anticipated given that the
domestic area is relatively small, and the range of
activities undertaken within the structure is likely to
have been limited and overlapped, this method was
adopted for the following analyses.
In practice, the identification of clusters with
regard to specific artefact types in some of the phases
is quite simple, as these can be spotted visually with
relative ease. Furthermore, the concentration of
artefacts in some of the small and clearly delimited
deposits provides a clear indication of their assoc-
iation with other artefacts. Material from the base of
hearths, for instance, cannot justifiably be included in
clusters of lithic material occurring in the occupation
floors surrounding them at a higher elevation. The
discrimination of discrete clusters in the dispersed
material from the levelling/occupation deposits,
however, requires the application of statistical
techniques, as here there is no definite contextual
association.
The K-means algorithm employed in this study
consists of the following elements. Firstly, the ‘split’
phase at which clusters are formed, when the initial
assemblage is divided into increasingly circum-
scribed clusters. Secondly, the clusters are then
combined if their centres are close enough. This
obviates the problems involved in producing many
clusters with only small populations (Shennan 1997).
Phase 1a
A prominent central hearth complex is characteristic
of Phase 1a. Around this complex formal tools are
arranged in three clusters. The most obvious inter-
pretation of these clusters is that they represent
discrete activity areas. An anomalous feature [270] at
the south-eastern extremity of the structure is also a
focus for activity, and the distribution of the lithic
material within it, and in its immediate vicinity, is
quite similar to that at the hearth complex. The key
aspect of the lithic distribution in this phase is that
the two recurrent clusters seen in most classes of
artefact are opposite one another.
Blades
The occurrence of blades, for reasons outlined earlier,
are regarded as the best proxy indicator of the
location of discrete flint knapping locales within the
98
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Fig. 7.15. Blade clusters in Phase 1a.
dwelling. The blades occur in three clusters, assoc-
iated with specific features (Fig. 7.15). The first cluster
is on the immediate periphery of the central hearth
complex, the second occurs within and around a
shallow pit [270], whereas the third is a modest
cluster focussed on the circular feature [274]. Because
of the obvious association of the blades with features,
there was little need to employ statistical methods to
identify patterns in their distribution. Instead, it is
possible to move directly to interpretation of this
pattern.
The concentration of blades around the central
hearth complex is not altogether unexpected since
the artisans undoubtedly sought to sit near the fire
for reasons such as the light and the warmth that the
fire would have provided while they produced the
blades. It is important to remark on the concentric
arrangement of the blades at the immediate margins
of the hearth in Cluster A. The arrangement of the
blades in this cluster suggests that an artisan sat
facing the hearth complex, and laid out the blades as
close to this feature as possible. There is no way of
establishing which of the hearths was in use whilst
these small concentric arrangements of blades were
deposited. It might be inferred that each of these
arrangements represents an individual knapping
episode, focussed on different hearths, or perhaps
these blades were all deposited at once.
The occurrence of blades in Cluster C is more
puzzling. This is principally due to the uncertainty
concerning the function of feature [270], in which
they occur: this shallow pit is not a hearth, but the
pattern of concentric arrangements of blades is quite
similar to that seen at the hearth complex. The same
concentric arrangement of blades on the margin of
the feature occurs, but it is significant that a linear
arrangement is also evident. Both of these patterns
are typical of lithic artisans laying out blades during
the knapping process. At the hearth complex, the
heat and light provided by the fire were undoubtedly
an attraction to the flint knapper, but no such
utilitarian considerations are apparent for feature
[270]. This matter is considered in more detail in the
discussion of debitage distribution (see below).
Microliths
The microliths are more widely dispersed than the
blades, although their distributions broadly agree
(Fig. 7.16). Only a small segment of the overall
microlith assemblage from Phase 1a had its location
digitally plotted (as the rest were recovered from the
sieving), which has reduced the utility of this class of
artefact for identifying flint knapping locales. Never-
theless, the material recovered from flotation derives
from the same deposits, that is to say, from the hearth
complex and feature [270], so the material recorded
7CHIPPED STONE TOOLS 99

Figure 7.16. Distribution of microliths in Phase 1a.
is broadly representative of the location of the other
microliths. The location of the microliths in the
hearths causes some interpretive difficulties. This
hearth pit complex was re-dug on several occasions,
and therefore it is possible that they were indeed
placed at the margins of the hearth during the
knapping. Alternatively, once the microliths had been
used some may have been discarded in the fire or
become incorporated into the hearth deposits by
accidental loss resulting from use of the tool in this
locale.
Scrapers
Only seven scrapers were digitally plotted from this
phase of occupation, however the total occurrence of
this class of artefact amounted only to sixteen (Fig.
7.18). The deposits from which the other scrapers
derived are the same as those in which the digitally
plotted artefacts occurred, so the distribution of the
small amount of digitally plotted material seems to
accurately reflect the distribution of this class of
artefact. This material, like the microliths and blades,
also seems to aggregate in, or lie immediately
adjacent to, the central hearth complex and the
Cores
shallow pit feature [270].
The distribution of cores is almost identical to that of
blades and microliths (Fig. 7.17), although the small
quantity only permits generalisations based on their
association with features in the hut. Nevertheless, the
strong spatial associations amongst blades, micro-
liths, cores and the debitage suggest that their
distribution is a function of their involvement in the
flint knapping process rather than the dumping of
material in baskets or skins, as was documented, for
instance, outside a quite similar Mesolithic structure
at Tågerup in Scania (Karsten and Knarrström 2001).
It is, however, important to remark on the relatively
large concentration of cores in [270]: this is not
associated with any correspondingly large concen-
tration of either blades or microliths.
Debitage
The distribution of the primary debitage corresponds
almost exactly with the distribution of blades and
closely conforms to the distribution of cores (Fig.
7.19). This correspondence is largely explicable by
the knapping process. Thus, the production of
debitage would fall in the immediate vicinity of the
blades and cores. The distribution of secondary
debitage, being lighter, is presumably more dis-
persed, around the knapping zone.
This supposition agrees well with the distribution
of the secondary debitage (Fig. 7.20). The material is
only slightly displaced in relation to the primary
debitage; the displacement occurs towards the south
100
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 7.17. Distribution of cores, Phase 1a.
Figure 7.18. Distribution of scrapers, Phase 1a.
7CHIPPED STONE TOOLS 101

Fig. 7.19. Distribution of primary debitage, Phase 1a.
Figure 7.20. Distribution of secondary debitage, Phase 1a.
102
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 7.21. Distribution of blades, microliths and cores, Phase 1b ( = blades, = cores, = microliths).
of the structure, which suggests that the artisan(s)
faced southward whilst producing the implements.
The orientation of the artisan(s), moreover, corrob-
orates the assumption that the hearth was the focus
of the activity, insofar as it provided light and
warmth. It must be remembered that the hut was
relatively dark throughout the day – the reconstruc-
tion of the hut, in which experimental flint knapping
was undertaken, clearly revealed this. Perhaps the
entrance was also situated towards the south, which
might have provided even more light, but this must
remain speculative. A far more intractable problem is
presented by the clustering of material in [270]: this
feature does not seem to be associated with any
source of heat or light, nor does its form suggest any
obvious function. All that might be said at the
moment is that it was a focus for lithic production
and the deposition/discard of tools.
Phase 1b
The distribution of material in Phase 1b is more
difficult to interpret because of its quantity and the
association of much of it with a wide range of
features. Unlike Phase 1a, which was characterised
by very discrete concentrations immediately visible
on the plans, the material recovered from Phase 1b
requires cluster analysis in order to identify any
patterning in its distribution.
Blades, Microliths and Cores
The blades, microliths and cores occur in the central
portion of the structure in this phase (Fig. 7.21). When
compared with the distribution of the same artefacts
in Phase 1a it is immediately apparent that there is a
slight displacement towards the west, with the central
hearths forming the sole focus for knapping and
associated activities. Small concentric groups are
visible in each of the identified groupings, which
might reflect the deliberate arrangement of blades
during multiple knapping events.
In each of the clusters identified (A, B and C),
there were small concentric arrangements of blades
and microliths (Fig. 7.21). These are most clearly seen
in Cluster C, as the quantity of material occurring
within this is not so high as to obscure the arrange-
ments. Two arrangements might be noted here.
Although concentric arrangements of blades are
apparent in Cluster A, it cannot be determined
whether these represent deliberately placed tools or
fortuitous arrangements resulting from the volume
of material deposited in this area. Nevertheless, it
does seem apparent that there are at least three main
knapping areas where blades were produced, but the
main concentration of this activity seems to lie on the
7CHIPPED STONE TOOLS 103

Figure 7.22. Distribution of blades, Phase 2.
south side of the central hearth pit area. The dis-
placement of activity compared to the preceding
phase is interesting. Does it suggest that domestic
space was used differently when the hut underwent
the structural modifications that separate these two
phases? It is possible that as the position of industrial
activities changed when the hut was modified this
could be related to the repositioning of key structural
features that may have affected task location, such as
the doorway. As the knapping locale changed, this
suggests that it is unlikely that a cosmological belief
or taboo underpinned this positioning: if this were
the case the activity could be expected to have
remained in the same place. This does not of course
preclude the latter interpretation but it is considered
the less likely option here.
Phase 2
The spatial distribution of material in Phase 2 is more
complex than in Phase 1a or Phase 1b (Figs 7.15–
7.21), principally because of the existence of level-
ling/occupation deposits that might contain material
that was introduced to the hut from outside. The
confusion and problems arise because deposits [115],
[118] and [210] were also occupation floors and some
material appears to have been deposited whilst it
served as such (see Chapter 5). Earlier it was
mentioned that the metrical attributes of the blades
and microliths could not resolve the issue of how
much of this material was redeposited, and how
much was likely to be in situ. The spatial analysis of
the blade, microlith and core distribution does,
however, strongly suggest that the majority of this
material was indeed deposited whilst the afore-
mentioned deposits served as occupation floors. The
manner in which the blades surrounded the central
hearth, rather than being spread across it, suggests
that these features were respected. Of course, it is
possible that small amounts of material were dumped
around the burnt deposit deliberately, but such an
argument would involve special pleading as it is a
striking coincidence.
A slight displacement of activity to the east is
apparent in this phase when compared to the central
activity locales in Phase 1b. This pattern conforms
more closely to that seen in Phase 1a, although there
are some differences, namely, that all the clusters are
arrayed around the central burning/hearth area. The
hearths again form the focus of activity, but there is
no secondary focus as in Phase 1a.
Microliths
A curious phenomenon is that the microliths are all
arrayed to the immediate east of the central hearth
area, and thus their distribution does not correspond
104
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 7.23. Distribution of microliths, Phase 2.

Figure 7.24. Distribution of scrapers, Phase 2.
7CHIPPED STONE TOOLS 105
with the general spread of blades. Moreover, it is also
not in agreement with the distribution of the cores
and test pieces that would, presumably, have been
deposited in the immediate vicinity of flint knapping
locales. The general distribution of cores and test
pieces conforms better to that witnessed for the
blades (Fig. 7.22). Interestingly, the distribution of
microliths (Fig. 7.23) is broadly similar to that of
scrapers (Fig. 7.24). This might suggest that the
distribution of these two classes of artefact in Phase 2
does not represent production, but instead represents
the location at which they were discarded after use. It
is noteworthy that the residue and use-wear analysis
showed that some scrapers and microliths may have
been used on similar materials, ranging from raw
meat and skin products to hard materials such as
wood, bone, antler and horn (see Chapter 9).
Phase 3
Much of the material from Phase 3 occurred in
levelling/occupation deposit [049], so the same
problems that encumbered a consideration of the
patterns of artefact distribution in Phase 2 are
repeated here.
A large quantity of blades (198) was digitally
plotted during the excavation of Phase 3. Their
distribution generally agrees more with that seen in
Phase 1b than in Phase 2, so there is a curious
alteration in the main concentrations of blades in
every horizon of occupation: from the east side of the
central hearth, to the west, back to the east, and finally
to the west again in Phase 3. This might have been
governed by symbolic considerations, but it might
equally have been a function of the situation of the
entrance during these different phases. Similarities in
the general distribution and quantity of blades and
other tools are seen in Phase 3 (Figs 7.25–7.28) and
are comparable with the patterns from the earlier
phases, except that there is a displacement in the
location of clusters. If the displacement is left aside
the Phase 3 distribution most closely resembles that
of Phase 2 where one large, dense cluster represents
the main focus of activity around the centre of the
hut, with other subsidiary clusters overlapping.
The scrapers do not occur in sufficient numbers to
make a statistical analysis meaningful, but their
distribution in three small clusters can be easily seen
(Fig. 7.27). A central cluster is the only one of any
size, containing nine specimens. Despite the small
size of the assemblage, the distribution of scrapers is
fundamentally similar to that identified for the
microliths. This suggests that the centre of the
structure was an area in which a variety of activities
was pursued as well as at the margins on the north
and west sides.
Summary
This analysis has presented the results from three
discrete, but complementary, investigations. The
metrical analysis confirms that although the assem-
blage belongs to a blade-based industry the attributes
of the blades were also governed by the qualities of the
available raw material, which tended to result in short,
squat blade forms. The stability of the assemblage
composition through time is also noteworthy. The
assemblages from all phases are overwhelmingly
dominated by debitage, with the remaining categories
of artefact constituting less than 6% of the material.
The majority of the latter portion of the assemblage is
microliths, cores and scrapers; other classes of artefact
are only nominally represented by several pieces and
form less than 0.5% of the assemblage in each phase.
Nevertheless, amidst the evidence for such con-
siderable stability are indications of subtle temporal
trends. The most noteworthy is the gradual rise in the
proportion of secondary debitage through time, and a
corresponding decrease in primary debitage. This
phenomenon cannot be correlated with either an
increase or decline in the quantity of cores present in
each phase. When considered concurrently with the
evidence from the metrical analysis, which suggests
no fundamental difference in the production of blades
or microliths through time, there is no basis for
suggesting that a change in the chaîne opératoire
occurred. It may simply indicate that there was a trend
to undertake the curation and maintenance of tools
within the setting of the hut, with the primary flaking
taking place elsewhere outside the hut, presumably
closer to the place of raw material acquisition. This
may have arisen from practical necessity or could be
related to a slight change in the organisation of the
knapping routine as a result of personal preference.
The spatial analysis revealed clear knapping and
activity zones throughout all phases, with the central
hearths providing the key focus throughout the
different phases of the hut. During Phase 1a there was
a secondary activity focus centred on a shallow linear
pit feature, [270], in the north-east quadrant of the
structure. Both the linear feature and the central hearths
served as flint knapping locales, but they also contained
evidence for scrapers, suggesting the pursuit of other
activities. The use of the domestic space seems to alter
in the succeeding phase when a displacement of
activity to the western side of the central hearth
complex can be observed. A change in the spatial
distribution of the artefacts occurs in Phase 2 where
the material clusters around the central burning/hearth
area, and presumably reflects activities undertaken by
people sitting around a fire. The interpretation of this
material is, however, more problematic than that
occurring in Phases 1a and 1b, due to the presence of
the levelling/occupation deposits that, probably in
part, have been introduced from outside the hut.
106
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 7.25. Distribution of blades, Phase 3.

Figure 7.26. Distribution of microliths, Phase 3.
7CHIPPED STONE TOOLS 107
Figure 7.27. Distribution of scrapers, Phase 3.

Figure 7.28. Distribution of burnt lithics, Phase 3.
108
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Indeed, the distribution of the artefacts in Phases 2 and
Phase 3 renders it possible that the material was
dumped. Militating against this, however, are the
general uniformity of artefact types in the composition
of the material and the clear focus of material in and
around the central hearth zone. Consequently this
patterning is most logically explained as being formed,
for the most part, by material discarded in situ during
activities around the hearths. The existence of small
concentric arrangements of blades, which are typically
associated with discrete knapping episodes, supports
the view that much of the material is from in situ discard
and that the patterns observed for these two phases is
largely representative.
Overall Summary and Discussion
Clive Waddington
The lithic assemblage from Howick is one of the
largest discrete assemblages of stone tools so far
recovered from North-East England. The date cor-
relates well with those from two other dated North-
East coast sites at Filpoke Beacon (Jacobi 1976, 71)
and East Barns (Gooder in press) that produced
narrow-blade assemblages typified by the presence
of scalene triangles. However, these are not the
earliest dates for the occurrence of what Jacobi has
termed the ‘narrow-blade micro-triangle techno-
complex’ in Britain, as dates from the recently
excavated site at Cramond place this scalene triangle
site in the last quarter of the 9th millennium cal BC
(Lawson 2001; Saville 2004).
A key observation resulting from the examination
of this assemblage is that it includes previously
chipped, patinated and beach-rolled pieces which have
been later modified into artefacts used in the Howick
hut. This proxy evidence for Palaeolithic/Early
Mesolithic occupation on what were likely to be coastal
sites along the western shore of the North Sea Basin
provides an important indication that North-East
Britain experienced episodes of occupation during the
Palaeolithic and early Post-Glacial periods. Not only is
this Palaeolithic presence supported by Saville’s recent
survey of potential Palaeolithic hand-axes in Scotland
(Saville 1997), but it also provides a logical counterpart
to the early dates for initial occupation along the arctic
coast of Norway (Bjerck 1995; Bang Andersen 1996;
Waraas 2001). Evidence for what is likely to have been
a fleeting Early Mesolithic presence on the site is
provided by two tools: a heavily patinated obliquely
blunted broad-blade microlith [1388] from the subsoil
and a heavily patinated backed blade [2291], more than
41mm in length. Although these two pieces cannot be
directly dated they do not fit with the rest of the
assemblage and are conspicuous by their size, form
and patinated condition. If nothing else they provide
an indication of a much earlier presence in the area,
just as the Neolithic leaf-shaped arrowhead, recovered
from the subsoil to the north of the hut, is indicative of
a post-Mesolithic presence.
The Howick lithic assemblage is characterised by
the exclusive use of locally available material, in
particular the small beach pebbles that can be picked
up from the shore. This is indicative of opportunistic
collection of an easily acquired resource with no need
for wider collection or acquisition through organised
exchange networks with more distant groups. The raw
material, as we have seen, is a key determinant that
has conditioned to some extent the size and shape of
formal tools as well as the primary flaking of nodules
by the use of bipolar (hammer and anvil) flaking. The
net result is that although there is clearly a concern to
maintain a narrow-blade tradition, the blades are
generally squat and often retain areas of cortex due to
the small size of the raw material. One possible reason
for the lack of core-axes found in northern Britain at
this time may be that such large pieces would not have
been discarded at the end of their use life as they could
have been more usefully remodified into new smaller
tools. It is clear from the small size of all the lithics in
the Howick assemblage that any large pieces of flint
would have been at a premium and are unlikely to
have been knowingly discarded.
Although all phases of the reduction sequence are
present in the Howick assemblage it is clear that most
primary working took place away from the site and it
is most likely that this occurred close to the collection
locales on the shore. However, there is a remarkable
lack of microburins given the size of the microlith
component and this suggests that the production of
microliths, and probably other formal tool types,
mostly took place at another location outside the hut.
Based on ethnographic studies it is most common for
flint knapping to take place outdoors, which is not
surprising given the flying slivers of flint and
production of large quantities of sharp debris littering
the floor. In contrast, the Howick assemblage can
perhaps be more accurately viewed as a ‘working’
assemblage that has accumulated as a result of
processing tasks taking place in the hut, with only
occasional curation and replacement of tools taking
place as an adjunct to the processing tasks being
undertaken. The arcs of microliths and blades, the
clusters of scrapers, and the arcs of bevelled pebble
tools (see Chapter 8) suggest the use of these tools for
tasks undertaken in a sitting position. If the full
production cycle of all the tools in the Howick
assemblage took place within the hut then there
would have been vastly more debitage recovered by
the wet sieving, but as the volume of such material is
simply not present it points to the conclusion that the
debitage assemblage from the hut is largely that for
the curation, and occasionally replacement, of tools.
Moreover, these tools appear to have been used for a
7CHIPPED STONE TOOLS 109
wide-range of processing tasks (see also Chapter 9),
situated in a designated part of the hut, in front of a
central hearth and carried out in a sitting position.
The presence of tool types that could be clearly
differentiated on morphological grounds, together
with the results from the residue and use-wear
analyses, indicate that a wide range of specialist tasks
took place in the hut. Based on the tool forms, residue
and use-wear analyses and the types of material
found within the hearths, we can envisage these tasks
including the softening and/or smoothing of skins
and furs, the stitching of these materials into clothes,
bedding, coverings and containers, the processing of
plant materials, such as nettles for making string, the
processing of seaweed, hazelnuts, wood, antler and
bone to make other required items, and the prepar-
ation of meat, fish, birds and plant foods for cooking.
Although the assemblage from Howick may be
seen to contain a little bit of everything when it comes
to the reduction sequence, it is clear that it is an
assemblage that has accumulated over a prolonged
period of occupation (see Chapter 6) and which,
bearing this in mind, has a relatively low incidence of
knapping debris. Consequently the assemblage can
be characterised as having:
1. a disproportionately large number of formal tools and
utilised pieces in relation to the amount of primary and
secondary debitage;
2. a general absence, or very low incidence, of micro-
burins;
3. a wide range of tool types (scrapers, awls, burins, blade
tools, etc.);
4. a wide range of microliths (scalene triangles, backed
blades, crescents, points, and occasional rods, Lamelle á
Cran, obliquely blunted) and
5. a large proportion of broken tools as a result of use
before discard.
These characteristics provide a tentative signature
for lithic assemblages acquired from Mesolithic
settlement sites. When the Howick assemblage is
compared with the lithic assemblage from similar
sites such as Mount Sandel, parallels are instantly
recognisable, such as the wide range of tool types
and narrow-blade microliths present (Woodman
1985), although differences between the Irish material
and that from North-East Britain is evident, such as
the presence of core, flake and polished stone axe
heads on the Irish sites and their absence from the
British sites. The Howick awls are broader than the
Mount Sandel equivalents and the microlith points
are also slightly different. Overall though, the
similarities are greater than the differences and the
lack of axes on Northern British sites is most likely
accounted for by them being recycled during the
Mesolithic rather than them not being used. This is
supported by the fact that occasional Mesolithic axe
heads and axe head fragments have been found in
North-East Britain, including tranchet axes from
Monkwearmouth and Seaton Carew (Weyman 1984,
39) and the Milfield basin (Waddington 1999).
However, it remains to be seen if other dwelling
structures, such as that at East Barns, produce a
similar ‘signature’ for a hut site. By attempting to
characterise the assemblage in this way, further
refinement of this suggested signature is left open for
future studies, which may look in more detail at the
assemblages from hut sites. It is hoped that this
characterisation of the Howick assemblage may aid
the interpretation of Mesolithic lithic scatter sites
where few, if any, structural remains survive.
Although it did not reveal a dwelling structure like
that at Howick, and there was clearly evidence for the
full quantity of knapping debris associated with all
stages of the reduction sequence, the site at Kinloch,
Rhum, holds other similarities. Not only does this site
have comparable dates to Howick but the range of
microliths present is also the same as those from
Howick (Wickham-Jones 1990, 99). In the case of
Kinloch we are also dealing with a coastal site whose
inhabitants relied exclusively on locally available raw
material in much the same way as those at Howick. As
only a proportion of this site was excavated, the
structural remains of a hut could yet be preserved there.
However, an interesting distinction arises when it is
considered that at Kinloch it was the chief knapping
area that was excavated rather than perhaps the main
dwelling area, whereas at Howick it was the dwelling
area and not the main knapping area that was
excavated. As the lithic assemblage at Kinloch was
vastly larger, comprising 138,043 pieces of worked
stone – which was only a fraction of the other debitage
that littered the site (Wickham-Jones 1990, 57) – it is
evident that this material represents an area where
tool production took place. It is this stone tool
production locale that is missing from the Howick site
and, therefore, it is reasonably presumed that in the
case of Howick this is likely to have been situated
outdoors and closer to the raw material source. As the
cliffs at Howick have eroded considerably since the
Mesolithic, the survival of such a locale is unlikely.
The microlith types present in the Howick hut are
conspicuous by their varied forms which contrasts
with those sites where just one or two forms pre-
dominate, such as at the small temporary site at Fife
Ness where crescents were by far the most common
(Wickham-Jones and Dalland 1998). In the case of
Fife Ness it was thought that this may indicate that
the site was used for specialist activities such as the
collecting of a particular resource (Wickham-Jones
and Dalland 1998). In contrast, the wide range of
microliths present at Howick, together with the
evidence from the faunal, botanical, residue and use-
wear analyses, make it clear that a much wider range
of tasks took place, which is consistent with the view
that the site was occupied for a sustained period, by
a group numbering several people.
110
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
8BEVELLED PEBBLES, COARSE STONE TOOLS AND
OCHREOUS MATERIAL
Ann Clarke and Clive Waddington
Introduction
A total of 19 bevelled pebbles (including four pieces
that refit into two complete ones) and 15 bevelled
pebble tool blanks was recovered from the Howick
excavations. These bevelled tool forms, all made from
the locally outcropping sandstone, dominate the
coarse stone assemblage from Howick (see Table 8.1).
The location of each of these tools can be seen on the
phase plans that form part of the discussion in
Chapter 14 (Fig. 14.2). A few other associated tool
types were also present, including two faceted
pebbles, two hammerstones and what appears to be a
smoothing stone. A stone pad utilised as a timber
support was recovered from the foot of timber stain
[29] in the Phase 3 horizon and eight unused sand-
stone cobbles and burnt stone fragments were also
found. Two quartz flint-knapping hammerstones
were discovered within the hut deposits but these are
discussed elsewhere in Chapter 7 as the form, ma-
terial and crushing indicates these are knapping tools.
Bevelled Pebbles and Associated Tool Blanks
The bevelled pebbles form a discrete group of tools
(Fig. 8.1) that share a number of common character-
istics summarised here and discussed below:
i) They are made on narrow, elongated or finger-like
pebbles.
ii) The bevels are usually formed on just one end, normally
the broadest end of the pebble.
iii) The bevel is larger on one face and is convex in section.
It has been smoothed and often striations are visible
running longitudinally.
iv) More often than not this bevel is worn diagonally across
the face, usually from left to right. Where the wear is
light the left side has been more heavily used.
v) On the opposite face the bevel is narrower and rougher
in character with some smoothing over chipping
damage.
vi) Occasionally the pebble is turned over and both faces
are worn.
vii) Occasionally both ends of the pebble are used.
All the pebbles are made on sandstones that vary in
texture from fine- to medium-grain and these would
have been available locally on the beach. The finer-
grained sandstones exhibit more smoothly worn faces
and the striations are more visible than on the coarser
sandstones, but this is most likely to be related to rock
hardness and homogeneity rather than any difference
in their use. They all appear to have been used for the
same activity since the wear traces are similar.
By shape and size the bevelled pebbles form a
discrete group ranging from 95mm–140mm in length
and 28mm–48mm in width (Fig. 8.2 and Table 8.2).
The selection of a specific size for use is confirmed by
the presence of the tool blanks, which are pebbles of
the same size, shape and raw material as the bevelled
pebbles but which bear no traces of wear (Fig. 8.2).
These unused pebbles were undoubtedly collected
from the beach and brought into the house with the
intention of using them in a similar manner to the
bevelled pebbles.
From observations of the 13 complete tools, the
broader end of the pebble is preferred for use in most
cases whilst the narrower end is used in only one
case. A number of the tools have been used on both
ends and the wear traces on four pieces indicate that
the tool has been turned over so that a large bevel is
worn on both faces at one end.
There is variation in the level of wear as some
tools appear to be more lightly worn than others,
indicating that the alteration to the surface of the tool
was through use rather than manufacture. The bevels
have been worn smooth by a grinding action which,
given the unidirectional nature of the striations, was
most probably in a forward or downward motion.
The bevel itself is slightly convex in cross-section,
which may suggest that the material upon which the
tool was used was slightly concave in section or had
a degree of elasticity allowing the bevel to form a
rounded surface. The formation of the bevel on the
opposite face is different in that it is usually narrow
and there is often additional damage to the end of
this bevel in the form of light chipping or pecking
which has then been smoothed over.
The larger bevel is worked at an angle across the
face, with most of the tools exhibiting a clear diagonal
8BEVELLED PEBBLES, COARSE STONE TOOLS AND OCHREOUS MATERIALS 111
Phase Bevelled
pebbles
Bevelled
pebble
blank
Faceted
Pebble
Hammer-
Stone
Smoothing
Stone?
Timber
Pad
Unused
Cobble/
Fragment
Total
Phase 1a
Pit 3 2 - - - - 3 8
Hearth 1 3 - - - - 1 5
Sub Total 4 5 - - - - 4 13
Phase 1b
Occ. Layer 5 2 2 - 1 - 1 11
Linear
Spread/Pit
- 1 - - - - 1 2
Hearth - 2 - - - - - 2
Sub Total 5 5 2 - 1 - 2 15
Phase 2
Occ. Layer 2 2 - - - - 1 5
Burnt Material 2 - - - - - - 2
Sub Total 4 2 - - - - 1 7
Phase 3
Occ. Layer 3 1 - 1 - - 1 6
Post Socket - 2 - - - 1 - 3
Sub Total 3 3 - 1 - 1 1 9
Others
Unstratified 1 - - 1 - - - 2
Outside Hut 2 - - - - - - 2
Sub Total 3 - - 1 - - - 4
Totals 19 15 2 2 1 1 8 48
Table 8.1. Coarse stone tools by phase and feature type (including the broken segments of two refits in their separate phases, as in both cases
they had been re-used).
bevel from left up to the right, and on an additional
three pieces where the bevel is less developed, there
is a tendency for the left side to be more heavily
worn. On just two pieces the bevels are angled from
right up to left and this most probably demonstrates
at least a degree of left-handedness in the population
at Howick.
Given the angle of the wear and the direction of
striations running parallel with the length of the tools,
they could have been held like a knife, angled
towards the user (which would allow for the develop-
ment of the angled bevel) and used in a downward
motion (see Fig. 8.2). Possible uses for these tools are
explored in Chapter 14.
There is a refit between broken fragments of a tool
blank from hearth [357] in Phase 1a and burnt spread
[320] in Phase 1b. This piece is badly burnt and
abraded and the two fragments were clearly de-
posited separately in antiquity. Another refit, this
time of a bevelled pebble, links two fragments from
an occupation layer in Phase 1a, [232], and the Phase
3 occupation layer [49]. Abrasion over the broken
ends indicates, again, that the tool was broken, each
piece re-used and then discarded separately in
antiquity.
It is clear from the Phase 1a and 1b plans (Fig. 14.2)
that the bevelled pebbles were clustered around
specific sectors of the hut and that the location of
these activity areas was different in each of these
phases. Perhaps this change in location is related to a
change in the position of the hut entrance (see also
Chapter 14). In general the tool blanks follow the
same contexts of deposition as the bevelled pebbles
and are present in all phases of the hut. A review of
bevelled pebble tools is included in Chapter 14,
together with a full discussion of the Howick pieces,
including their relationship with the other excavated
small finds and analyses.
112
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 8.1 Bevelled pebble tools, tool blanks and a faceted pebble from the Howick hut.
8BEVELLED PEBBLES, COARSE STONE TOOLS AND OCHREOUS MATERIALS 113

Figure 8.2 Reconstruction of the way in which the Howick bevelled pebble tools are thought to have been used.
Faceted Pebbles
There were two faceted pebbles recovered from
Mesolithic contexts. One, (3913), has a rounded
pecked facet on one end of a finger-like pebble which
resembles the rounded pecked facet on the end of the
unstratified bevelled pebble (15107), suggesting that
the latter piece was selected twice for different jobs.
A tabular piece of sandstone, (4351), is badly heat-
damaged but narrow ridged facets may have been
worn on opposite faces.
Unused Cobbles and Stone Fragments
Four unused sandstone cobbles and four unused
sandstone fragments were found in the Mesolithic
phases with six of the eight showing obvious signs of
having been heat-affected. These rocks could have
been used as cooking stones. As three of these stones
were found in pit [270] from Phase 1a, it suggests that
this pit may have been used for heating activities. It is
perhaps notable that an arc of bevelled pebbles was
also found in and around this shallow pit, implying a
connection with an activity that required heat.
Hammerstones
Both hammerstones have non-specific spreads of
pecking on their surface. One, a limestone cobble,
(15108), may also have a face which has been worn
smooth, though like the other specimen it is not a
confident assignation of type. The sandstone ham-
merstone fragment (4923) has the dimensions of a
bevelled pebble and this may have been its original
function.
Smoothing Stone?
The possible smoothing stone (4316) is simply a small
oval limestone cobble with one face that may have
been worn smooth from use. Such tools are difficult
to identify confidently unless the cross-section of the
surface has been altered and in this case it has not.
Ochreous Material
By Clive Waddington
The ochreous material from Howick was catalogued
and briefly assessed to establish its possible uses with
regard to the Mesolithic occupation of the hut. A total
of 38 fragments of ochreous material was retrieved
from the Mesolithic hut deposits, together with a
further two from the subsoil horizon [3] above. The
stratified fragments came from 11 different contexts
across Phases 1a, 1b, 2 and 3. A variety of ochreous
material was identified, including micaceous sand-
stone and haematised shale, which may have been
available locally.
The ochreous material came in a variety of shades
including red-browns and yellows. Ochreous ma-
terial does not occur naturally in the sand substratum
at Howick and therefore the material has evidently
been collected and brought into the hut. The frag-
114
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context
No.
Small
Find No.
Phase Feature Type Material Max
Length
(mm)
Max
Width
(mm)
Max
Thickness
(mm)
Remarks
270 3487 1a Pit Stone
fragment
Sandstone 130 110 43 Fragment from block of
sandstone. Burnt.
270 3519 1a Pit Cobble Sandstone Fragment of unused sandstone
cobble. Heat-cracked. No sign of
use wear.
270 3572 1a Pit Cobble Sandstone 66 53 27 Oval cobble of sandstone.
Weathered. No sign of use wear.
5 (hut
floor)
3407 1a Pit Tool Blank Sandstone 115 31 20 Finger-like pebble of sandstone.
Fragmented and very abraded,
probably heat-damaged.
Abrasion has obliterated wear
traces.
5 (hut
floor)
3483 1a Pit Tool Blank Sandstone 105 32 16 Flat elongated pebble of
sandstone. No clear signs of use.
May be a blank for a bevelled
pebble.
5 (hut
floor)
3564 1a Pit Bevelled
Sandstone 113 28 20 Finger-like pebble of sandstone.
pebble
Heat damage, fragmented. Bevel
worked on broader end. Larger
bevel (13mm broad) on one face.
Bevel appears smooth and is
rounded. On opposite face bevel
is narrower (6mm broad) and
less worn.
270 4814 1a Pit Bevelled
Sandstone 113 31 18 Finger-like pebble of sandstone.
pebble
Bevel worked at broader,
spatulate end. Larger bevel
(11mm broad) is convex in
section and very smooth as if
from being rubbed. In plan the
bevel is angled down from left
up to right. On opposite face the
bevel (10mm broad) is more
rounded and smoothed over
chipping damage. Width at bevel
25mm.
270 4984 1a Pit Bevelled
Sandstone 95 32 17 Flat finger-like pebble of
pebble
sandstone. Bevel worked on
broader end. The bevel (9mm
broad) is only lightly worked
with smoothing and longitudinal
striations visible. Similar bevel
(7mm broad) on opposite face
with some chipping damage.
Light bevelling on both faces of
opposite end too. Some pecking
on one side also towards narrow
end. The bevelling is less
extensive than on other pieces.
Width at bevels 22mm and
18mm.
355 4957 1a Hearth Stone
Sandstone 140 115 50 Fragment from block of
sandstone. Burnt.
5 (hut
floor)
fragment
4732 1a Hearth Bevelled
pebble
Sandstone 44 32 17 Elongated flat pebble of
sandstone. Broken across width.
Cracked from heat damage.
Surviving end has a bevel worn
on opposite faces. They are both
15mm broad and smoothed with
longitudinal striations. They are
both worked at an angle from
left up to right. This tool has
clearly been turned to use
Table 8.2. Catalogue of coarse stone artefacts. (Continued over the next four pages).
Context
No.
8BEVELLED PEBBLES, COARSE STONE TOOLS AND OCHREOUS MATERIALS 115
Small
Find No.
Phase Feature Type Material Max
Length
(mm)
Max
Width
(mm)
Max
Thickness
(mm)
Remarks
opposite face. Width at bevel
26mm.
5 (hut
floor)
4813 1a Hearth Tool Blank Sandstone 83 44 20 Flat finger-like pebble of
sandstone. Broken across width.
Heat-cracked. No sign of wear
on surviving end.
357 5158 1a Hearth Tool Blank Sandstone 105 37 35 Elongated oval cobble of
sandstone. Weathered with
removal of cortical surface. No
sign of wear.
320 and
5114/4812 1b
379
Hearth Tool Blank Sandstone 150 44 19 Refit. Elongated flat pebble of
and
1a
sandstone. Broken across width
with each half having different
find no and context. Clearly been
burnt and then abraded
afterwards. Broad end from 320,
narrow end from 357. No
definite signs of use wear but
this may have been obliterated
by weathering.
320/49 3911/1816 1b
and 3
Occ.
Layer
Bevelled
pebble
Sandstone 65 27 16 Refit. Finger-like pebble of
sandstone. Broken across width.
Each half has different find no
and context. On broader end a
bevel (14mm broad) has been
smoothed with visible
longitudinal striations. Bevel
worked at an angle from left
down to right so held differently
to other bevelled pebbles.
Narrower bevel (5mm broad)
slightly rougher worked on
opposite face. Piece from C.49 is
unworked end and
abraded/weathered over broken
end. Width at bevel 21mm.
118 4988 2 Occ.
Tool Blank Sandstone Flat elongated pebble of
Layer
sandstone. Fragment only, burnt.
Breakage has obliterated any
sign of wear.
160 4021 2 Burnt
material
Bevelled
pebble
Sandstone 118 43 26 Flat finger-like pebble of
sandstone. Bevel worked on
broader, spatulate end. On one
face the bevel is narrow (9mm
broad) and very smooth as if
produced by rubbing. It is
slightly convex in section and
worked more on left side in plan
and angled down from left up to
right. Possibly some light
insignificant bevelling on
opposite face. Width at bevel
35mm.
160 614 2 Burnt
material
Bevelled
pebble
Sandstone Broken, bulbous segment with a
heat-shattered end. An angled
facet survives, indicating this
bevelled tool has been used.
Table 8.2. Continued.
116
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context
No.
Small
Find No.
Phase Feature Type Material Max
Length
(mm)
Max
Width
(mm)
Max
Thickness
(mm)
Remarks
210 2903 2 Occ.
Layer
Bevelled
pebble
Sandstone 135 48 27 Elongated flat pebble of
sandstone. Bevel worked on
broader end. Bevel (13mm
broad) convex in cross-section
and smoothed with visible
longitudinal striations. Worn
more heavily on left side of
bevel. On opposite face the bevel
(7mm broad) is more lightly
worn over some chipping
damage. Width at bevel 37mm.
210 2852 2 Occ.
Stone
Sandstone 83 80 48 Fragment from block of
Layer
fragment
sandstone. Burnt.
210 2929 2 Occ.
Bevelled
Sandstone 111 40 20 Elongated flat pebble of
Layer
pebble
sandstone. Possibly heat-cracked
and surface abrasion. Bevel
(11mm broad) on one end is
smooth with visible striations
and worked at an angle from left
up to right. On tip and opposite
face there is some chipping
under the smooth bevelling.
Possible light bevel worked on
left side of one face at opposite
end. Width at bevel 25mm.
210 3587 2 Occ.
Tool blank Sandstone 140 50 23 Flat oval pebble of sandstone.
Layer
Weathered with onion-peel
removal of cortex, obliterating
any signs of use wear that may
have existed. Possible tool blank.
232 3913 1b Occ.
Faceted
Sandstone 132 23 27 Finger-like pebble of sandstone.
Layer
Pebble
Rounded facet pecked on one
end possibly from two faces.
Some flaking from the edge of
facet.
264 4361 1b Burnt
Tool Blank Sandstone 150 54 17 Flat elongated pebble of
Spread
sandstone. Heat-cracked. No
clear signs of use. May be a
blank for a bevelled pebble.
264 3673 1b Burnt
Tool Blank Sandstone Finger-like pebble of sandstone.
Spread
Fragment surviving. Heavily
burnt. No sign of wear on
surviving surface.
285 3856 1b Pit Cobble Sandstone Fragment of unused sandstone
cobble. Heat-cracked. No sign of
use wear.
316 5085 1b Occ.
Cobble Sandstone Fragment of an unused
Layer
sandstone cobble. No sign of
wear on surviving surface.
316 4351 1b Occ.
Faceted
Sandstone 102 43 17 Tabular pebble of sandstone.
Layer
Pebble
Fragmented, heat-damaged,
abraded. Surviving broad end
may have narrow ridged facets
worked on opposite faces.
Difficult to tell with surface
damage.
316 4095 1b Occ.
Bevelled
Sandstone 119 35 22 Flat elongated pebble of fine-
Layer
pebble
grained sandstone. Bevel worked
on broader spatulate end. On one
face the bevel (13mm broad) is
very smooth over some flaking
damage with visible longitudinal
striations. On opposite face the
bevel (12mm broad) is smooth
with visible longitudinal
striations. Smoothing over of
chipping damage. Bevel in plan
angled left up to right. Width at
bevel 32mm.
Table 8.2. Continued.
8BEVELLED PEBBLES, COARSE STONE TOOLS AND OCHREOUS MATERIALS 117
Context
No.
Small
Find No.
Phase Feature Type Material Max
Length
(mm)
Max
Width
(mm)
Max
Thickness
(mm)
Remarks
316 3330 1b Occ.
Tool blank Sandstone 125 37 21 Elongated oval pebble.
Layer
Fragmented. Burnt/abraded. No
sign of use wear.
320 4316 1b Occ.
?Smoothing
Limestone 85 46 29 Oval cobble of limestone. One
Layer
Stone
face may have been worn
smooth from use.
326 4249 1b Linear Tool Blank Sandstone 105 26 18 Finger-like pebble of fine-
grained sandstone. Heat-
damaged. Broken down length,
destroying one end. No sign of
working but this may have been
obliterated by breakage.
Probable blank for bevelled
pebble.
326 4061 1b Occ.
Layer
Bevelled
pebble
Sandstone 75 42 20 Flat elongated oval pebble of
sandstone. Broken across width.
Bevel end surviving. Larger
bevel (14mm broad) on one face
is very smooth as if produced by
rubbing and convex in section
with visible longitudinal
striations. In plan the bevel is
angled down from left up to
right. On opposite face the bevel
is narrower (7mm broad) and
less heavily worn and is rubbed
smooth with some chipping
damage. Width at bevel 30mm.
326 4062 1b Occ.
Layer
Bevelled
pebble
Sandstone 140 44 24 Flat elongated pebble of fine-
grained sandstone. Bevels
worked on narrower spatulate
end. On one face bevel (11mm
broad) is smooth and convex in
section with visible longitudinal
striations. The bevel in plan is
angled left up to right. On
opposite face the bevel (12mm
broad) is smooth with visible
longitudinal striations and
worked more heavily on the left
side. Width at bevel 29mm.
326 4063 1b Occ.
Layer
Bevelled
pebble
Sandstone 111 40 13 Flat elongated oval pebble of
sandstone. Bevels worked on
broader, spatulate end. Larger
bevel (12mm broad) is convex in
section and very smooth as if
from being rubbed and
longitudinal striations are
visible. In plan bevel angled
down from left up to right. On
opposite face the bevel is
narrower (6mm), less heavily
worn and rubbed smooth with
some chipping damage. Width at
bevel 34mm.
29 15109 3 Timber
Stone Pad Sandstone 190 160 63 Fragment from block of
Stain
sandstone. Burnt. Used as timber
support in base of C.29.
49 859 3 Occ.
Layer
Stone
fragment
Sandstone Fragments of heat-cracked and
surface damaged stone.
Table 8.2. Continued.
118
Context
No.
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Small
Phase Feature Type Material Max
Max
Max
Remarks
Find No.
Length
Width
Thickness
(mm)
(mm)
(mm)
49 862 3 Occ.
Layer
Bevelled
pebble
Sandstone 122 40 22 Flat elongated pebble of fine-
grained sandstone. Bevel worked
at broader end. Larger bevel
(15mm broad) is convex in
section and very smooth with
visible striations running parallel
to the length of the tool. In plan
the bevel is angled from left up
to right. Some chipping damage
on bevel ridge and over the
opposite face where there is a
very narrow bevel (4mm broad).
The tool has been turned to use
again. On opposite narrower end
(on opposite face to big bevel)
there is a rounded, smoothed
bevel with visible striations and
some chipping damage on bevel
ridge – not so developed as
opposite end. Width at bevels
38mm and 20mm.
49 1935 3 Occ.
Layer
Bevelled
pebble
Sandstone 82 43 19 Flat elongated pebble of
sandstone. Broken across width.
Heat-cracked and surface
weathered. Possibly some light
bevelling on surviving end but
this is not clear through the
surface damage.
49 1939 3 Occ.
Tool Blank Sandstone 60 23 15 Small finger-like pebble of
Layer
limestone. No clear sign of wear
but may be a tool blank.
49 4923 3 Occ.
Hammerstone Sandstone 103 43 29 Elongated pebble of sandstone.
Layer
Broken across width. Heat-
cracked. Small patch of light
pecking on surviving narrow
end.
49 2945 3 Post
Tool Blank Sandstone 113 31 15 Flat elongated pebble of
Socket
sandstone. Heat-cracked. No
clear signs of use. May be a
blank for a bevelled pebble.
51 15106 3 Post
Tool Blank Sandstone 105 34 18 Elongated flat pebble of
Socket
sandstone. No sign of wear.
21 15105 Meso Pit Bevelled
Sandstone 118 42 26 Elongated pebble of sandstone.
pebble
Weathered all over with some
post-depositional damage. Bevel
worn on broad end but obscured
by weathering.
87 1188 Clay
Spread
Bevelled
pebble
Sandstone 97 31 21 Finger-like pebble of sandstone.
Bevels on each face at either
end. The bevels are 11mm-
15mm broad and are smooth
with visible longitudinal
striations. The bevels tend to be
more angled from left up to
right. Width at bevels 21mm and
23mm.
Unstrat 15108 Subsoil Hammerstone Limestone 82 53 30 Oval cobble of limestone. Upper
face with some light pecking
from use. Opposite face may
have been worn smooth from
use.
Unstrat 15107 Bevelled
Sandstone 99 30 25 Finger-like pebble of sandstone.
pebble
On broader end a bevel (13mm
broad) has been pecked and
ground with visible longitudinal
striations which run on for over
10mm past the bevel. Bevel
worked at an angle from left up
to right. On opposite face there
is a smaller bevel (9mm broad),
similarly pecked and ground. On
narrower opposite end there is a
rounded pecked facet. Width at
bevel 28mm.
8BEVELLED PEBBLES, COARSE STONE TOOLS AND OCHREOUS MATERIALS 119
ments were fairly evenly distributed across the three
main structural phases, with 14 occurring in deposits
from Phases 1a and b, 10 from Phase 2 and 14 from
Phase 3.
Ochre is frequently found on Palaeolithic and
Mesolithic sites throughout Europe and it is clear
that hunter-gatherer groups valued this material.
Although it has healing properties and can be used in
the preparation of hides, ochre is also used as a
pigment (Isbister 2000, 191–95), presumably for art,
body decoration or for its symbolic blood-like colour.
The association of red ochre with Palaeolithic and
Mesolithic burials is well attested in North-West
Europe, from the well-known burials in Scandinavia
such as Skateholm (Larsson 1989), to British sites such
as the ‘Red Lady’ of Paviland (Aldhouse-Green and
Pettitt 1998), where the ochre appears to have been
included in the grave as part of a symbolic act. The
discovery of ochreous material at Howick in what is
considered to be ostensibly a domestic setting within
the confines of a dwelling, is of particular interest
because in this case the ochreous material is not
linked to burial. This implies that ochre had wider
uses outside the sphere of ‘burial’ and ‘ritual’ and
that it could have also been used for symbolic
purposes such as art and body ornament as well as
more mundane purposes such as pigment for hide
preparation and medicinal uses. One other interesting
possibility is the use of ochre as a fixative in the
mastic used to seal joints in light water craft. The
American coastal Chumash tribe are recorded as
having used ochre in this way and the presence of
ochre and punching tools associated with hearths has
led Cassidy et al. (2004) to interpret this as evidence
for boatbuilding amongst the early Holocene coastal
groups of California. At Howick, ochre was assoc-
iated with the Phase 1a hearths as well as occupation
layers abutting the hearth areas. Given that Howick
is a coastal site, the use of ochre in boat construction
must remain an important possibility.
120
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS
Karen Hardy and Robert Shiel
Residue Analysis
Residue analysis involves the examination of residues
that adhere to the surface of a tool. A sample of
residue is extracted from the tool’s surface. It is then
dried, mounted on a slide and examined micro-
scopically from a morphological perspective. Par-
ticular attention has been given to the detection of
starch grains and phytoliths (e.g. Barton et al. 1998;
Fullagar 1998) and it is sometimes possible to identify
these up to species level.
A different method for examining residues was
first tested by Jahren et al. (1997). This method
involves semi-quantitative analysis of residues using
an environmental or low-vacuum scanning electron
microscope (SEM). Jahren et al. (1997) however,
carbon coated their samples, which eliminated the
possibility of identifying residues that contained
carbon. In this experimental study, artefacts were not
coated and were inserted without any special pre-
treatment. Jahren’s method is a potentially valuable
non-destructive way to obtain information regarding
the use of a tool. Problems remain to be resolved,
most notably with regard to the potential for post-
depositional contamination, however. One potential
for contamination comes from modern roots and this
is something that affects all buried material. How-
ever, in this case, the C/N ratios that have been
determined are widely different to the expected
norm. In fact the deviation from the norm indicates
the opposite of contamination with modern materials.
It is probable though that material on the tools will
have been affected by subsequent post-depositional
diagenesis.
Method
A selection of tools were first examined using a Meiji
ML 2305 incident and transmitted light microscope
to identify tools with residues and to plan the location
of the residues on the tools’ surface. A small number
of tools were then selected, based on the potential
quality of the residue, for SEM analysis.
A Philips XL30CP SEM instrument with an an-
alytical system capable of detecting light elements
(down to C) was used. Imaging was carried out using
back-scattered electron (BSE) mode, with point
analyses made of features on the rough surface.
Analysis was carried out using PGT (Princeton
Figure 9.1. Tool 4395. Example of ED spectrum for background flint.
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 121

Figure 9.2. Tool 4395. Example of tool surface at low magnification
(x90).

Figure 9.3. C and N proportions derived from energy dispersive
analysis (all samples).
Gamma Tech) analytical software, which provided
an energy dispersive (ED) spectrum that was used as
a fingerprint, to distinguish different material types,
supported by a semi-quantitative analysis (fully
quantitative analyses are impossible with rough
samples). A background analysis was made on the
residue-free flint surface to permit comparison with
spectra and analyses obtained from the residues. In
all cases, peaks reported in the ED spectrum were
identified using the analytical software, taking care
to identify and label Ka, Kb, La
and Lb X-ray peaks
when observed.
Each artefact was examined across one surface and
a series of spectra was taken. The first spectrum for
every piece was a background spectrum of the flint
surface, without residue, which was used as a control.
Further spectra were then taken on a selection of
different residue types located on the surface of the
tool. At magnification, the surface of the flint is
sometimes found to be pock-marked with small holes
and often the residues could be found lying inside
these minute holes. In Figure 9.2 the background flint
can easily be distinguished from the black dots. These
black dots are either empty holes, surface residues or
holes filled with residue. Viewing at higher magnif-
ication enables a distinction to be made between
these.
An interpretation was made as to the nature of the
residue, based on the ED spectrum and chemical
information. For mineral materials, the relative
atomic proportions of Al, Si, Fe, K and Mg were used
to identify tentatively silicates such as biotite,
cordierite, feldspar and clay minerals. Spectra which
showed both Ca and C were attributed to calcite. In
other cases, spectra showing only strong C and N
peaks were attributed tentatively (in the absence of
Ca) to organic matter. Atomic proportions of C:N
were determined to assess variation in the com-
position of the organic residues.
The observed values were compared with ratios
calculated for typical natural materials:
• Wood, straw, oils and fats = C/N >30
• Green plant tissue = C/N around 25
• Wheat and other seed grains = C/N around 15
• Soil and fungi = C/N around 10
• Most animal tissues (excluding fatty acids), bacteria and
actinomycetes = C/N 5 and below.
Animal protein (meat) is very unlikely to survive
without being altered by bacteria. Its small C:N ratio
is unlikely to alter substantially with time, whereas
most plant material begins with a very wide C:N
ratio and narrows progressively. Therefore a nitrogen-
rich residue is an indication of protein-rich materials.
Bulk soil C:N ratio is rarely much below 10 because
of the predominance of resistant plant materials, such
as lignin and cellulose. Ratios as low as 5 are only
found in undisturbed soil at depths greater than 1m
(Jenkinson 1988) and are usually associated with
materials strongly absorbed onto clay surfaces. The
samples analysed here came from lesser depths.
In this experimental study, two result groups
emerged – one in which the C:N ratio was infinite, i.e.
there was no N detected, and the other in which the
C:N ratio lay between 2.5 and 5.5. Neither of these
cases occurs in a normal terrestrial topsoil. Carbon
without nitrogen is found in plant parts but not in
whole plants and will not become attached to stones
without human intervention. Equally, the narrow
ratio falls outside the expected range of values.
Anything below 5 is rare indeed and then is only
found below 1m depth (Dyer 1902). It is therefore
more likely that these materials are the result of the
use of the tools and reflect two groups of uses – one
associated with low C:N ratio materials such as meat
or pulse seeds and the other working wood with its
high to infinite C:N ratio.
Data for all samples containing both C and N are
shown in Figure 9.3. Despite the limitations inherent
in using semi-quantitative analytical information
from EDS spectra, the data obtained in this study are
consistent in the relative proportions of C and N.
122
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 9.4 Example of a carbon/nitrogen-rich spectrum, Tool 3162.
Because other spectra were observed that give a C
but no N signal, it is assumed that the spectra
showing both elements reflect their presence in the
sample, and that their relative proportions are
consistent (although absolute reported quantities may
have little meaning).
The underlying solid geology in the Howick area
consists of a Carboniferous sequence of interbedded
sandstone, limestone and mudstones; dolerite out-
crops less than 1km along the coast. The bedrock is
covered with a veneer of boulder clay that contains
sand pockets, and it is in one of the sand pockets that
the Mesolithic site is situated.

Figure 9.5. Tool 3162. Carbon/nitrogen-rich residue (323x magnif-
ication). The elongated black object at the top left is a hole containing
a nitrogen-rich residue.
Results
Artefacts were placed into sealed bags immediately
as they were being excavated to minimise the risk of
contamination. Bags were not reopened until the
artefacts were ready to be placed in the SEM at which
point their surfaces were lightly brushed to remove
any surface soil. In this section, only the positive,
residue related results will be discussed. Of the
numerous readings that were taken, many were
found to be linked to the background soil while some
were rejected due to poor quality of the spectrum
(low total X-ray counts). These will not be included
in the discussion.
23.69/6.40 = 3.7
23.93/ 4.68 = 5.1
21.36/8.59 = 2.5
All three ratios are very low and fall into the category
of animal protein or seeds. Fat tissue, oil, wax and
carbonised organic material contain carbon but no
nitrogen so they are excluded here.
Lithic No. 3162
Context 236/109. Roasting pit, Phase 2.
Unretouched blade.
Fifteen spectra.
Three spectra had counts of carbon and nitrogen
that are as follows:
Lithic No. 4697
Context 340. Hearth, Phase 1b.
Microlith, point.
Nine spectra.
Two spectra had carbon and nitrogen present and
one spectrum had carbon alone.
Carbon/nitrogen ratios:
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 123

Figure 9.6. Copper-rich spectrum from Tool 4410.

Figure 9.7. Tool 4410. Copper residue (223x magnification). The
copper residue is the bright white mark to the centre right of the
photo.
The analysis is consistent with the presence of a
copper (Cu) mineral in which zinc (Zn) is an impurity,
and with the suggestion that the residue is a second-
ary hydrated copper carbonate mineral.
The copper is weathered; this may have occurred
after the copper became adhered to the tool, or the
flint could have come into contact with a weathered
piece of copper ore.
Secondary copper mineralisation is reported to
occur at outcrops in the Cheviot Hills some 24km
due west, suggesting a possible source (Carruthers et
al. 1932). Additionally, zinc minerals are exposed on
the coast at Howick. Transport from the Cheviots
could have been within glacial till (erratics from the
Cheviots are common at Howick), or could have been
brought by people during the Mesolithic for pro-
cessing and use for colouring purposes. Copper
carbonate minerals can be used as a source of green
or blue pigment.
25.31/5.70 = 4.44
22.97/6.27 = 3.66
Once again the carbon/nitrogen ratios suggest the
presence of protein-bearing material, which is con-
sistent with the tool being in contact with meat or
seeds. The presence of carbon without nitrogen can
be attributed to several different things. It can be
wax, oil, fat or burnt organic material. It is not
currently possible to distinguish between these
different materials.
Lithic No. 4410
Context 316. Occupation layer, Phase 1b.
Microlith, backed blade.
One spectrum.
Lithic No. 4382
Context 320. Large burnt spread, Phase 1b.
Microlith, scalene triangle.
Ten spectra.
Five spectra produced evidence of cordierite,
based on the relative proportions of magnesium (Mg),
aluminium (Al) and silicon (Si). Cordierite is pro-
duced when Mg-bearing clay (a typical constituent of
northern English boulder clay) is fired at high
temperatures (over 800°C). This temperature can
occur in the middle of a hot hearth fire where the
underlying clayey sediment would turn, in part, to
cordierite. The most likely explanation for this is that
this tool inadvertently found itself in a fireplace.
Three spectra had evidence of organic material.
The presence of carbon without nitrogen suggests
that this is either oil, wax, fat or carbonised organic
124
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK


Figure 9.8 Example of cordierite on surface of Tool 4382. Figure 9.9. Tool 1621. Example of wood ash residue (black marks at
upper left of image).
material, or that it is the carbonised (charcoal)
remains after the more reactive nitrogen-containing
compounds have been destroyed, presumably in this
case by heat.
Lithic No. 4986
Context 118. Occupation layer, Phase 2.
Scraper.
Fifteen spectra.
Four spectra had carbon and nitrogen present and six
spectra had carbon alone.
Carbon/nitrogen ratios.
21.38/4.75 = 4.5
24.55/4.67 = 5.25
24.23/4.81 = 5
24.03/5.68 = 4.23
For this piece, the carbon/nitrogen ratios suggest
residues of meat or seeds, and the carbon-only
readings suggest wax, oil, fat or carbonised organic
material. Fat could be produced for example by
scraping the inside of a hide, while wax could be
produced as the result of a glue such as beeswax and
resin. The carbon-only residues are spread across
much of the surface, while the carbon and nitrogen-
bearing spectra occur mostly on or near the edge of
the tool.
Unretouched flake.
Seven spectra.
Four spectra were similar. They contain potassium
(ash contains potassium carbonate), chloride and
sodium. Potassium occurs in chemical fertilisers and
as no fertilisers have been used near the site in recent
memory, it is most likely that the potassium is derived
from ash or charred timber. The presence of sodium
suggests that the burnt woody material might be
derived from or had contact with the sea, such as
seaweed or driftwood.
Interpretation
This piece may have been used to work a piece of
charred timber. Timber can be fired first to make it
easier to work for example in digging out the inside
of dugout canoes. Wood was undoubtedly a crucial
raw material and is likely to have been used to make
a wide range of items.
Lithic No. 2758
Context 49. Occupation layer, Phase 3.
Microlith, scalene triangle.
Six spectra.
Three spectra contained carbon but no nitrogen.
This suggests wax, oil, fat or carbonised organic
material.
Lithic No. 4395
Context 320. Large burnt spread, Phase 1b.
Scraper.
Six spectra.
Three spectra were carbon-rich, suggesting a wax,
oil, fat or carbonised organic material.
Lithic No. 1621
Context 3. Unstratified subsoil.
Lithic No. 4548
Context 320. Large burnt spread, Phase 1b.
Microlith, backed blade.
Five spectra.
Two spectra contain potassium (as said, ash
contains potassium carbonate), chloride and sodium,
possibly suggesting wood ash or charred timber (see
above Tool 1621).
The carbon and nitrogen counts are as follows:
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 125

Figure 9.10 Tool 4548. Example of a wood ash residue spectrum.
23.69/4.36 = 5.4
20.09/5.30 = 3.8
These ratios suggest meat or seeds.
Lithic No. 4294
Context 316. Occupation deposit, Phase 1b.
Scraper.
Thirteen spectra.
Two spectra had carbon and nitrogen present,
while one spectrum had carbon alone.
22.03/7.13 = 3.08
25.18/4.59 = 5.48
These ratios suggest meat or seeds.
Lithic No. 4652
Context 340. Hearth, Phase 1b.
Unretouched blade.
Five spectra.
Two spectra contain potassium, chloride and
sodium, possibly suggesting wood ash or charred
timber (see above Tool 1621).
Lithic No. 4063
Context 232. Occupation layer, Phase 1a.
Coarse stone elongated bevelled pebble tool.
Four spectra.
This is a siltstone. Four spectra were taken from
this piece, three of the background and one from a
residue spot. This produced carbon, suggesting wax,
oil, fat or carbonised organic material.
of Howick during the Mesolithic, though it is im-
portant not to forget the experimental nature of the
study. The residue analysis has identified the possible
use of either driftwood or seaweed. It has also
suggested that copper-rich stones were brought to
Howick, possibly for extraction of pigment. The
number of tools with traces of narrow C:N ratio
materials, typical of having been associated with
much more protein-rich materials such as meat or
seeds, is surprising, particularly given recent ethno-
graphic evidence for the intensive use of lithics in the
preparation of material culture items (Hampton 1999;
Hardy and Sillitoe 2003) and the current focus on
plants as an important food and raw material source
during the Mesolithic (Juel Jensen 1994; Zvelebil 1994;
Owen 2000). Tools that were used for working
protein-poor materials have residues with a wide or
infinite C:N ratio. This ratio has persisted in the
residues (for example Tool no. 2758). The reason for
the difference in residue composition between tools
awaits further investigation. The presence of cor-
dierite shows that the hearth fires sometimes had
very intense heat. Though further work is required,
the development of this method, to enable more
specific and a wider range of interpretations of the
chemical spectra, may provide a new opening for the
study of the organic record, something that has
hitherto been missing from much of the early pre-
historic archaeological record.
Discussion
The residue analysis has produced a series of detailed
findings that can contribute to building up a picture
Use-Wear Analysis
Introduction
The traditional system of classification of lithic
artefacts is based on morphology and an understand-
ing of the technological processes in the manufacture
of lithics. Suggested artefact use is interpretative, and
126
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 9.11. Example of an unpolished, probably unused edge at
100x magnification).

Figure 9.12. Example of a polished point at 50x magnification
(artefact no 2309).
based on morphological characteristics. Use-wear
analysis by contrast uses empirical, measurable and
experimental evidence to examine artefact use.
Artefacts are examined both from a morphological
and physical point of view, using a combination of
measurements and high-power microscopy which
has rarely been carried out on Mesolithic samples
from northern Britain. When it has, there has been a
tendency to concentrate on retouched artefacts, in
particular microliths (Finlayson 1989, 1990; Finlayson
and Mithen 2000) though it is clear that unretouched
pieces were also routinely used (Knuttsson 1988 a
and b, 1990; Hardy 2004, forthcoming a). Addition-
ally, ethnographic studies have suggested that
modern conceptions of what may be deemed a useful
edge or artefact rarely correlate with the perceptions
of the manufacturer/users (White 1968; White and
Thomas 1972; Hayden 1979; Hardy and Sillitoe 2003;
Sillitoe and Hardy 2003).
Use-wear analysis can provide a wealth of in-
formation that contributes greatly to the general
interpretation of a site (e.g. Dumont 1985; Grace
1989). By providing information on artefact use, it
can contribute information about site function as well
as shed light on other elements such as the aims of
artefact modification (were artefacts broken or
modified for use in specific ways?), the knapping
process in general (which unretouched pieces are
waste products and which are not?) and post-
depositional processes of artefact movement. One
intriguing question that can be addressed is whether
traditional classifications (scraper, serration, etc.)
retain meaning at the level of functional use as
opposed to classification.
There are several levels of use-wear analysis which
can contribute a varying range of information on an
assemblage of artefacts. At the most basic level, tools
can be identified as having been used or not;
information then increases to what type of motion
they have been used in (cutting, scraping, etc.), the
type of raw material they have been used on (e.g. soft
material such as meat or hard material such as bone),
and finally what the most likely function of the tool
was (e.g. Keeley 1980; Dumont 1985; Grace 1989;
Finlayson and Mithen 2000). In order to attain this
final level of information, it is necessary to undertake
experimental work in which a range of tasks is carried
out, using a similar raw material. The use-wear traces
are then compared between the experimental and
archaeological assemblages, and a comparative
judgement produces the most likely correlations.
Artefacts are examined first from a morphological
perspective. Measurements are taken on the volume
and shape of an artefact, and the angle of the used
edge (edge angle) measured. Following this, the use-
wear variables are measured. These include: the scope
and type of edge fracturing, the extent of edge
rounding, and polish. Polish is examined from the
point of view of its intensity, its distribution and the
extent of its invasiveness. Different types of fracturing
patterns suggest different movement of the piece, and
lines of polish point to the dominant direction of use,
for example if they all lie perpendicular to an edge
they indicate use in an up/down direction, rather than
longitudinally. Polish that extends deep into an edge
might have been used on a pliable material, such as
hide, while polish that is restricted to the limits of an
edge is more likely to have been used on a hard or
brittle material, such as bone (Figs 9.11–9.13).
Edge fractures can be caused by factors other than
use, notably the thinness of the edge, and stress,
which may occur from being carried around in a
pocket or pouch, being trampled, soil conditions or
post-excavation abrasion, for example bagging with
other artefacts. However, if an artefact shows a
concentration of fractures, often combined with non-
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 127

Figure 9.13. Example of same polished edge at 200x magnification
(artefact 2309).
natural straightening on one edge, or part of an edge,
then it is likely to be due to use. By contrast, if an
artefact, particularly a thin one, has inconsistent or
random fractures around all or most of its edges then
it is more likely to be related to something other than
use. Like edge fracture, polish may be due to many
different factors. Spots of polish, or polish that occurs
at random across a surface, are unlikely to have been
caused by use. A consistent pattern of polish along an
edge is more likely to have been caused by repeated
motion, which usually signifies use. The main
categories into which a tool’s use can be placed
include: cutting, piercing/boring, grinding, whittling,
scraping, and pounding/percussion.
Many different methods have been used to assess
most likely use based on experimental work. These
range from individual, personal comparison of traces,
artefact by artefact (e.g. Juel Jensen 1994) to expert
systems (e.g. Grace 1989) designed to interpret a suite
of variables directly, to produce most likely use. In
this study a method is used that was developed using
an ethnographic assemblage from Papua New Guinea
(Shiel and Hardy 2003; Shiel and Hardy in prep.).
Use-wear variables are numerically placed into a
database, which is then analysed to produce clusters
grouped by like use-wear traces, using a range of
multivariate statistical techniques.
Aims
The aim of this use-wear analysis is to examine tool
use from the Mesolithic site of Howick, using a small
sample selected from the lithic assemblage in order
to identify some of the raw materials worked at the
site as well as some of the probable tasks that were
carried out there.
Method
One hundred lithic artefacts from Howick were
selected for use-wear analysis. Artefacts were sel-
ected from the assemblage found inside the hut.
Selection was not random: artefacts that looked
‘likely’ to provide traces were picked out for use-
wear analysis. The sample included 50 retouched
pieces and 50 unretouched pieces.
Artefacts were washed by soaking in detergent.
Where necessary, edges and surfaces were cleaned
with alcohol. Morphological characteristics were
recorded first, including length, width, thickness and
shape of the artefact, the length and edge angle of
likely used edges, and any non-natural straightening
of an edge, or part of an edge. For the microscopic
analysis, a Meiji ML 2305 incident and transmitted
light microscope was used. An initial scan of the
artefact’s surface and edges was carried out at x40
magnification, followed, where necessary, by a more
detailed examination at x100 and x200 magnifications.
Macroscopic and microscopic features examined
included fractures, edge rounding and blunting,
breakage and the development, invasiveness and
distribution of polish. Examination of all these
features together has resulted in a well-established
method for undertaking use-wear analysis which is
followed here (see Keeley and Newcomer 1977;
Newcomer et al. 1986; Unrath et al. 1986; Grace et al.
1985, 1988; Bamforth 1987; Grace 1989; Hardy and
Sillitoe 2003).
One of the problems with micro-analysis of light-
coloured raw materials, such as many cherts and flint
types, is that it is difficult to detect clearly the surface
polish with the incident light needed for identification
of edge damage like fractures, topography and
striations. Incident light which comes from above at
an angle of approximately 45° is reflected by the
surface and is excellent for detecting irregularities in
the surface and on the edge of the artefact, but it does
not, except very faintly, reveal the presence and
location of polish. Incident light can also detect polish
on the surface of a dark raw material, but this method
of detection becomes much more difficult for detailed
analysis on any lighter raw material.
The polish was therefore examined both by
incident light and using transmitted light through
cross-polarisers. As the edges were normally thin,
this enabled the polish to be clearly detected by
showing it up as very bright, shiny patches along the
used edge. The reason for this is that unpolished
birefrigent, crystalline materials, such as flint and
chert, scatter light such that their surface appears
dull and dark through the microscope. When light
emerges through polished surfaces, it does not scatter,
resulting in these areas appearing brightly ill-
uminated in contrast to the surrounding dark,
unpolished surfaces. This method proved highly
128
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 9.14. Cutting grass as part of the experimental process.

Figure 9.16. Cutting seaweed as part of the experimental process.

Figure 9.15. Whittling wood as part of the experimental process.

Figure 9.17. Scraping dry hide as part of the experimental process.
effective for polish detection on all but the thickest of
edges. Incident light was used in recording all micro-
attributes and polish where possible.
The volume of a tool is determined by multiplying
length x width x thickness. The tool’s shape is
determined by standardising the intermediate and
shortest axes by dividing them by length. This gave
lengths coded I and S respectively. The shape (Z) was
then calculated as Z=I2/S, known as the Shiel ratio.
This gives a range from >1, which is planar, ~1, which
is cubic, and <1, which is columnar. The revised
definition of shape gave a negative correlation with
edge angle (P<0.001) and with edge thickness (P<0.05),
which suggests that not only size but also shape is a
factor in determining usable artefacts.
Once all the variables were examined and re-
corded, an estimate of use was also made and
recorded. Using a range of techniques, including
principal component analysis (PCA), chi squared
tests, analysis of variance and cluster analysis, groups
were distinguished and compared statistically. PCA
provides information on which variables are related,
while cluster analysis distinguishes groups, or
clusters, of tools that have similar attributes across a
range of variables. Examination of the variability of
properties of tools within and between groups can
then be carried out with ANOVA, if the measures
compared are continuous, or with chi square if they
are categorical. Using this approach, it is possible to
determine whether a group is made up of tools that
are similar morphologically or are linked by like use-
wear traces. This enables objective examination of
the way individual variables relate to one another.
Experimental Programme
The environmental record of a site and its sur-
roundings provides a good basis for an experimental
programme. A programme which includes working
on a selected range of different raw materials enables
much of the lost organic record, such as the use of
plants, meat, etc., to be reconstructed by comparative
methods, using the experimental tools as the basis for
the use-wear interpretations. By basing the use-wear
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 129
Figure 9.18. Skinning a roe deer as part of the experimental process. Figure 9.19. Scraping fresh hide as part of the experimental process.
Raw material
Type of work Numbers
worked
of tools
Nettles Cutting and shredding 4
Wood Whittling and debarking 3
Seaweed Cutting and slicing 4
Soaked antler Grooving 2
Dry hide Scraping and dehairing 3
Dry hide Piercing 3
Fresh deer Butchering and skinning 2
Fresh hide Removing fat 1
Coarse grass Cutting 6
Bone Scraping and cutting 1
Shell Cutting/piercing 2
Horn Cutting 1
Fish Removing scales 2
Multi-use Grass cutting/piercing wood 1
Multi-use Cutting wood, cutting hide
1
on wood
Total 36
Table 9.1. Experimental artefacts.
Items studied Number of
Number of
pieces
edges
Awl 1 2
Core rejuvenation blade 3 6
Microburin 1 1
Microlith 27 31
Retouched flake 4 4
Scraper 14 17
Unretouched blade 35 63
Unretouched flake 15 24
Total 100 148
Table 9.2. Artefacts examined.
programme on an experimental one, a database of
probable raw materials and the way in which these
were worked can be built up. For the Howick
experimental programme, tools made from the local
beach flint, identical to that used during the Meso-
lithic, were used. A range of tasks was undertaken
using a mixture of retouched and unretouched tools.
(Figs 9.14–9.19).
The aim of each task was to use the artefact for
around 20 minutes; however, many artefacts became
blunt before this time. By contrast, some artefacts
were used for longer; for example one flake (HEP17)
was used to cut coarse grass for over an hour and still
did not become blunt. Artefacts were then washed
and examined using the same methods as for the
examination of the archaeological pieces.
Results
Many artefacts had macroscopic evidence of use on
several edges. Seven artefacts were found to have no
evidence of use while of the remaining 93 artefacts
(65%) had more than one used edge. This brought the
total number of edges examined to 148.
Artefacts with no evidence of use include the
microburin, two microliths (one scalene and one
broken), one retouched flake and three unretouched
flakes. These items were removed from the database
before statistical analysis took place. The final
database that was subjected to statistical analysis
consisted of 141 used edges. The very high percentage
of artefacts showing evidence of use (93%) is un-
surprising, both because they were all found inside
the hut structure and also because of the non-random
way in which the selection for use-wear analysis was
carried out.
PCA analysis was undertaken using all morph-
ological variables as well as the use-wear data.
Following this, analysis of variance and chi-squared
tests were undertaken to test the links between
morphological and use-wear attributes. This suggests
that the amount and type of fracturing on an edge is
linked statistically to the size of edge angle and the
130
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Cluster Number of tools
Cluster 1 24 (17%)
Cluster 2 23 (16%)
Cluster 3 26 (18%)
Cluster 4 42 (30%)
Cluster 5 21 (15%)
Cluster 6 1 (0.7%)
Cluster 7 3 (2%)
Cluster 8 1 (0.7%)
Total 141
Table 9.3. Use-wear tool clusters.
Tool type Numbers
Core rejuvenation blades 2
Microliths 4
Retouched flakes 0
Scrapers 1
Unretouched blades 13
Unretouched flakes 4
Total 24
Table 9.4. Cluster 1, tool types.
Length
mm
Width
mm
Thickness
mm
Length Used
Edge mm
Edge
Angle o
Volume
mm3
Shape (See Shiel
& Hardy 2003)
29.5 12.1 3.7 23.0 32.2 1676 1.45
Table 9.5. Means of measurements of tools in Cluster 1.
width of the edge. Secondly the extent to which polish
extends into an edge (invasiveness) is linked to its
edge angle, with thinner edges having a greater
degree of invasiveness than thicker edges. Further
work is required to determine whether these correl-
ations are due to the morphology of the edge or due
to differential selection for tasks.
Tool type Numbers
Core rejuvenation blades 1
Microliths 1
Retouched flakes 2
Scrapers 4
Unretouched blades 8
Unretouched flakes 7
Total 23
Table 9.6. Cluster 2, tool types.
Cluster analysis
A cluster analysis was carried out only on the use-
wear variables. The analysis was initially run with
two to ten clusters and this showed that using eight
clusters gave groups with relatively homogeneous
tools and all containing more than one tool. The eight
clusters of artefacts are linked exclusively by their
use-wear characteristics and properties of each of the
clusters are described below. Once the use-wear
clusters had been created, the morphological data for
each cluster were examined separately. In this way
the morphology did not interfere with the creation of
the groups.
snaps and steps), edges strongly rounded or ground
down, strong and invasive polish distributed along
the used edge in an even or uneven way. Almost all
tools in this cluster have very heavy diagonal, parallel
and perpendicular lines of polish over and above the
broad stretches of polish along the tools’ edges. The
use assessment of tools in this group is that they were
used for cutting strong grass or fibrous plants. The
experimental data suggest that these use-wear traces
correspond most closely to those observed on tools
used to cut nettles.
At the Sands of Forvie (Hardy forthcoming a) it
was noted that most of the observed inner flakes had
invasive polish. This was interpreted as implying use
on a soft or pliable material. At Forvie, very few
other artefact types had invasive polish and it was
thought possible that unretouched inner flakes were
positively selected for use on soft materials. The
evidence from Howick once again suggests this may
be the case.
Cluster 1
Cluster 1 consists of long, thin, small tools with thin
edges covering a range of technological types, though
with a majority of unretouched flakes and blades
(core rejuvenation, plus unretouched flakes and
blades = 79%).
Artefacts in this group were used exclusively along
their sides, rather than on their ends and contain no
cortex (inner flakes and blades). Edges tend to be
straight. Use-wear characteristics include strongly
fractured edges with multi-fracture patterns (flakes,
Cluster 2
Cluster 2 consists of large squat tools that are short-
edged and wide, with a wide edge angle.
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 131
Length
mm
Width
mm
Thickness
mm
Length Used
Edge mm
Edge
Angle o
Volume
mm3
Shape (See Shiel
& Hardy 2003)
26.3 14.2 4.5 20.6 40.0 2100 1.91
Table 9.7. Means of measurements, Cluster 2.
Tool type Numbers
Core rejuvenation blades 2
Microliths 0
Retouched flakes 1
Scrapers 0
Unretouched blades 17
Unretouched flakes 6
Total 26
Table 9.8. Cluster 3, tool types.
Length
mm
Width
mm
Thickness
mm
Length Used
Edge mm
Edge
Angle o
Volume
mm3
Shape (See Shiel
& Hardy 2003)
30.8 14.7 5.0 14.3 49.4 2784 1.57
Table 9.9. Means of measurements, Cluster 3.
Tool types in this group are again mainly un-
retouched flakes and blades (70%), though there are
also a number of scrapers. Artefacts in this group
were used mostly along their sides, though some
were used on their distal ends. Most tools are non-
cortical (inner) or have a small amount of cortex.
Edges are mostly straight with heavy snap fracturing.
Edges are rounded but rarely ground down and
polish is spread continuously and invasively along
the edge. Polish development is strong and a number
of tools have diagonal lines of polish running away
from the used edge. The use assessment of tools in
this group is less clear, with a range of different uses
being suggested. These include cutting hide, fibrous
plants or fish, scraping hide and cutting wood. The
experimental data suggest that these use-wear traces
correspond most closely to those observed on tools
used to scrape or cut hide.
Tool type Numbers
Core rejuvenation blades 1
Microliths (+ ‘awl’) 7
Retouched flakes 0
Scrapers 12
Unretouched blades 20
Unretouched flakes 2
Total 42
Table 9.10. Cluster 4, tool types.
polish development is weak and lines of polish do
not occur.
The use assessment in this group is mainly for
cutting or whittling wood. The experimental data
suggests that these use-wear traces correspond most
closely to those observed on tools used to cut or
whittle wood.
Cluster 3
Cluster 3 consists of large, wide, thick tools with
wide edge angles and short used edges.
Tool types in this group are almost exclusively
(96%) unretouched. Artefacts in this group were used
mostly along their sides, most edges were straight
and tools were mostly non-cortical. They have
slightly fewer fractures along their used edges than
tools in clusters 1 and 2; fracture type was mostly
snap. Edges are mostly rounded though not ground.
Polish is spread unevenly and is quite invasive but
Cluster 4
Cluster 4 consists of quite small, wide, thick tools
with quite wide edge angles and long used edges.
There are relatively fewer unretouched tools in
this group (54%) and more scrapers. Artefacts in this
group were used in a range of ways, along their sides,
on their distal edges and sometimes all around the
tool. Though many tools were non-cortical, some
tools in this group were heavily cortical, suggesting
some external flakes were used for tools in this group.
132
Length
mm
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Shape (See Shiel
& Hardy 2003)
Width
mm
Thickness
mm
Length Used
Edge mm
Edge
Angle o
Volume
mm3
28.5 13.7 5.3 27.8 49.1b 2220 1.59
Table 9.11. Means of measurements, Cluster 4.
Tool type Numbers
Core rejuvenation blades 0
Microliths 15
Retouched flakes 0
Scrapers 0
Unretouched blades 5
Unretouched flakes 1
Total 21
Table 9.12. Cluster 5, tool types
Length
mm
Width
mm
Thickness
mm
Length Used
Edge mm
Edge
Angle o
Volume
mm3
Shape (See Shiel
& Hardy 2003)
25.2 9.1 3.1 17.8 33.9 1087 1.14
Table 9.13. Means of measurements, Cluster 5.
Edges were straight or convex. Tool edges were on
the whole heavily fractured with most edges having
a mixture of flake, snap and step fracturing. Edges
were mostly rounded but not ground, though a few
are still sharp. Used edges have an even, non-
invasive, intermittent spread of light polish and no
polish lines.
The use assessment in this group is mainly for
cutting/scraping wood or harder materials such as
bone or antler. The experimental data suggest that
these use-wear traces correspond most closely to
those observed on tools used to cut or whittle hard
material such as bone, antler or horn. It is important
to note that the wide edge angles of many pieces in
this group prevented microscopic analysis using
transmitted light through cross-polarisers. For this
reason, none of the scrapers has been given a specific
use assessment due to the difficulty of analysing
them.
some are ground. Polish distribution is mostly
continuous and evenly spread across the edge. Polish
is generally not invasive (<0.5). Polish development
is average, lying between strong and weak polish (B),
and a few tools have diagonal lines.
The use assessment in this group is mainly for
cutting or grooving wood, though a few tools were
assessed as having been used on a harder material
such as bone and one artefact was used on its tip for
piercing. The experimental data suggest that these
use-wear traces do not fall easily into any ex-
perimental group. They fall between seaweed, wood
and hard materials. It is possible that this group
represents a use that was not carried out as part of
the experimental programme.
Cluster 5
Cluster 5 consists almost entirely of microliths so
unsurprisingly these come out as very small, thin
tools with very thin edge angles.
Microliths in this group are all used along their
unretouched side except for one tool that was used
on its tip. Some of the unretouched blades were used
on their distal edge. Tools are almost all non-cortical
and edges are mostly straight. Tool edges were on the
whole heavily fractured with a mixture of flake and
snap fracturing. Edges are mostly rounded though
Cluster 6
Cluster 6 consists of one tool, a scalene microlith that
was used probably for piercing, on its point. It is in a
group of its own mainly because the use-wear was
noted but not recorded on the database due to it
being on a tip rather than on an edge. It is one of two
edges recorded for this tool, which was also used on
its side. This tool should probably fit into Cluster 5.
Cluster 7
Cluster 7 consists of three microliths. These tools
differ significantly from the rest in their use-wear
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 133
Cluster Length
mm
Width
mm
Thickness
mm
Length
Used
Edge
mm
Edge
Angle o
Volume
mm3
Shape
(See
Shiel
&
Hardy
2003)
1 29.5 12.1 3.7 23.0 32.2*< 1676 1.45
2 26.3 14.2 4.5 20.6 40.0 2100 1.91*>
3 30.8 14.7*> 5.0*> 14.3*< 49.4*> 2784*> 1.57
4 28.5 13.7 5.3*> 27.8*> 49.1*> 2220*> 1.59
5 25.2 9.1*< 3.1*< 17.8*< 33.9 1087*< 1.14*<
Table 9.14. Means of properties and measurements not included in the cluster analysis.
Key: * = significantly different to other groups; = less than the mean.
Clusters Numbers of microliths
Total number and
and percentage out of
percentage of all tools
total
per cluster
1 4 (13%) 24 (17%)
2 1 (3%) 23 (16.3%)
3 0 (0%) 26 (18.3%)
4 7 (23%) 42 (30%)
5 15 (50%) 21 (15%)
6 1 (3%) 1 (0.7%)
7 3 (10%) 3 ( 2%)
8 0 (0%) 1 (0.7%)
Total 31 (100%) 141 (100%)
Table 9.15. Microlith groups.
traces. They are heavily fractured with flake, snap
and step fracturing. Their edges are all ground down
and they have continuous invasive, uneven and
heavy polish. They also all have diagonal lines of
polish emanating from their edges. All three tools
were interpreted as having been used for cutting or
skinning in butchery or hide working. The exper-
imental data suggest that the use-wear traces fall
between those observed on tools used to cut shell
and hide.
morphological groups too (Table 9.14). The following
morphological attributes are statistically significant:
Cluster 1 has a small edge angle;
Cluster 2 is square shaped;
Cluster 3 is thick with a wide edge angle and a short used
edge;
Cluster 4 is thick with a wide edge angle and a long used
edge;
Cluster 5 consists of long, thin pieces with short used edges.
This suggests that tools are selected for tasks based
on a combination of their size and their edge angles.
This is in keeping with data collected from ethno-
graphic stone tool users (White and Thomas 1972;
Hardy and Sillitoe 2003).
Cluster 8
Cluster 8 consists of one edge of an unretouched
flake. This is a fairly large tool with a concave edge. It
is heavily fractured with flake and step fracturing.
Polish distribution is continuous, uneven and gen-
erally non-invasive. This edge differs from tools in
Cluster 3 mainly due to its heavy fracturing. It is
similar in all other respects to the tools in Cluster 3.
The use assessment suggested this tool was used for
woodworking and this corresponds well with the
experimental data.
Though the clusters were formed using use-wear
data exclusively, it is clear that they form clear
Microliths
The distribution of microliths suggests that while they
may have been used for a range of tasks, there is clear
differential distribution of them throughout the
clusters, when compared to the overall numbers of
tools. Microliths occur in substantially fewer numbers
in Clusters 2 and 3 and in significantly greater
numbers in Cluster 5. Clusters 5 and 6 are made up
exclusively of microliths.
134
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Cluster Unclassified/broken Point Backed
blade
Scalene Crescent Obliquely
blunted
point
1 1 1 1 1
2 2
4 2 1 2
5 3 3 3 2 2 2
6 1
7 1 1 1
Table 9.16. Microlith classifications according to cluster.
Cluster Unretouched
flakes
Unretouched
blades (incl.
core
Total number of
all unretouched
tools
rejuvenation)
1 4 15 24 19 79%
2 7 9 23 16 69%
3 6 19 26 24 92%
4 2 21 42 23 80%
5 1 5 21 6 40%
Table 9.17. Unretouched tools.
Total and % of unretouched
tools per cluster
Microliths have traditionally been assumed to be
armatures for projectiles of some sort. Recent work
(Hardy forthcoming a; Finlayson and Mithen 2000)
has suggested that, though some microliths may have
been used to tip arrows and so forth, they were also
used along their edges, probably as the cutting edge
of a composite tool of some type. The evidence from
Howick suggests not only that this is the case, but
also that microliths may have been specialised tools
for cutting wood and possibly harder materials, such
as bone or antler, though the residue results suggest a
wider range of uses for microliths.
Morphological classification of microliths into
groups such as scalene, backed blade etc. appears less
meaningful at the level of the use clusters, as the
different types are spread across groups, with no
obvious pattern emerging. However, this study only
included a small sample of these tools (31 in total), so
observations remain preliminary. Furthermore, the
use associated with Cluster 5, which accounted for
50% of these tools, was probably one that was not
undertaken as part of the experiment and so their
true function remains somewhat speculative at this
stage. The one microburin has no evidence of use.
This ties in well with evidence from Sands of Forvie,
where Hardy (forthcoming a) concluded that micro-
burins were not routinely used as tools.
detailed use-wear data on them. All the remaining
scrapers are in Cluster 2. The use assessment for this
cluster is hide working.
Unretouched tools
An examination of all the unretouched tools together
suggest they were multipurpose tools used in a range
of different tasks, though certain differences do occur.
Table 9.17 documents the variability of unre-
touched tools in each cluster. Unretouched pieces are
substantially less common in Cluster 5 than the
average. Cluster 2 has relatively more flakes and
fewer blades than the mean, while Cluster 4 has fewer
flakes and more blades.
Core rejuvenation blades
The use of core rejuvenation blades is interesting in
that it is a clearly defined technological piece linked
to core working. Evidence for their use illustrates
that tools could be selected from any level in the
manufacturing chain.
Scrapers
Of a total of 17 scrapers, most (12) fall into Cluster 4,
though this is likely to be incorrect due to the lack of
Discussion
The use-wear results have produced some specific
indicators about the assemblage, as well as providing
information on the use of tools. Based on the inform-
9RESIDUE AND USE-WEAR ANALYSIS OF STONE TOOLS 135
ation obtained here and at Sands of Forvie (Hardy
forthcoming a), it appears that inner flakes and blades
may have been selected for cutting soft, possibly plant,
materials. Inner flakes, particularly blades, are usually
thin with low edge angles. This makes them ideal for
cutting as they tend to be very sharp.
It is becoming ever more apparent that microliths
were not used exclusively to tip projectiles. Once
hafted, microliths have strong edges. The different
types of microliths do not appear to be linked to
specific types of use and it is possible that these types
are either a stylistic choice, or related to ways of
hafting. Sites such as Fife Ness (Wickham-Jones and
Dalland 1998), with a strong predominance of
crescent microliths, may hold the key to a more
detailed understanding of microlith form.
Scrapers, with their high edge angles, have very
strong rounded edges. Those scrapers that were
completely analysed all fell into the group interpreted
as for hide working. Scrapers are ideal tools for
scraping the fat off the inner part of the hide. Residue
analysis was carried out on three scrapers and the
results for these suggest that they were used on meat
or seeds and fat, wax or burnt organic material.
Though the residue work is still at an experimental
stage, these results do not contradict the suggestion
that scrapers may have been used to process raw
animal hide.
Unretouched pieces are found across the clusters.
This suggests that they were multipurpose cutting
tools, though there is an indication that inner flakes
and blades were used preferentially for cutting softer
material, and that unretouched pieces were also used
for wood working. The fact that core rejuvenation
blades, which are well-understood as forming part of
the technological process rather than being an end
product, were used, suggests that artefacts were
selected for use on the basis of their usefulness rather
than where they occurred in the technological
process. This has been observed ethnographically
among stone tools users across highland Papua New
Guinea, whose main criterion for selection was the
edge angle (Hardy and Sillitoe 2003; White and
Thomas 1972).
The use-wear analysis has suggested that a wide
range of activities was taking place across the site,
exploiting many different raw materials.
Certain experimental uses did not feature in the
archaeological assemblage, for example fish de-
scaling. This does not mean these activities were not
carried out at the site; fish were very likely to have
been eaten but the use-wear sample, which represents
a tiny proportion of tools from the site, did not pick
this up.
A comparison of the residue and use-wear results
is not easy. One hundred artefacts were examined for
use-wear while only ten were examined for residues,
a sample size that was due largely to the level of
funding available. Additionally, the results for the
residue analysis are item-specific, while the results
for the use-wear analysis are broad and spread across
clusters of tools. For some tools, the results appear to
correspond, for example tool number 4382 had
residues suggesting it had been in a very hot fire
(cordierite) and ED spectra suggesting that it had
residues of a burnt organic material. The use-wear
data suggest this tool was used to cut wood. It is
possible that the tool fell inadvertently into a fire
after it had been used on wood or that it was used to
work burnt wood. By contrast, tool no 4410, a point
microlith, had residues of copper, suggesting it was
involved possibly as part of a process aimed at
obtaining pigment. The experimental programme did
not include this use and it was therefore not picked
up in the use-wear analysis.
Tool use and activities at Howick
Fibrous plants such as nettles had a multitude of uses
in the Mesolithic. Cordage, string and netting are
known to have been manufactured since the Upper
Palaeolithic (Wayland Barber 1994; Soffer 2000).
String is the basic raw material for many items that
would have been crucial during the Mesolithic. These
include such items as clothing, belts, carrying nets,
fish nets and twine for tying things together and
lashing structural supports in place. In Denmark, fish
nets made from plant fibre have been found at
Ertebölle sites (Anderson 1995; Malm 1995; Skaarup
1995). String is also likely to have formed an integral
part of many other items. String can be made with
most types of fibrous plants and from different parts
of plants, for example leaf fibres, bast fibres, root
fibres, bark fibres or the whole plant stem (Westcott
1999). String making can be a long and laborious
process that has been relatively ignored in attempts
to reconstruct the Mesolithic way of life.
Phytoliths from grass and leaves, as well as wood,
were found in hearths from the Mousterian and
Natufian site of Hayonim Cave in Israel, suggesting
that these were also used as fuel (Albert et al. 2003).
Coarse grass may also have formed a fuel component
at Howick. Coarse grass was used by the Inuit for
such uses as thatching and for items such as bedding,
matting, clothing and even children’s nappies (Owen
2000). This is perhaps significant in understanding
what type of material/s may have been used at
Howick.
Residues on two tools had possible evidence for
wood ash or charred wood. In both cases, residues of
sodium and chloride suggest a marine origin though
this may be the result of subsequent contamination.
Evidence for hide working (Cluster 2), wood
working (Cluster 3) and working hard materials such
136
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
as bone, antler and horn (Cluster 4) are present. These
materials would all have had very important uses at
Howick; hide was probably used for many things,
such as to make clothing and shelter; wood is likely
to have been used to manufacture many items of
material culture as well as probably the frame for the
hut at Howick. Skins of birds and fish, as well as
animals, were widely used among the Inuit and
Indians of northern North America (Owen 2000),
while the use of birds has been documented both
archaeologically and ethnographically in Tierra del
Fuego (Mamelli and Estevez Escalera 2004). Bone,
antler and horn would have also had many different
uses and it is noteworthy that bone and antler tools
are ubiquitous at Mesolithic shell midden sites on the
west coast of Scotland.
One activity missing from the stone assemblage
examined here is evidence for projectiles. Spears and
arrows would probably have formed an important
part of the hunting tool kit. Among the Wola of Papua
New Guinea, stone was never used to tip arrows and
spears. Extraneous bone causes septicaemia and the
Wola use this for their most deadly weapons. The use
of bone to tip arrows has not been explored for the
Mesolithic in northern Britain, however this has been
seen elsewhere in Europe (Oshibkina 1985; Zagorska
and Zagorski 1985) and ethnographically (Lee and
Devore 1976). It is also possible that arrow tips are
found infrequently on sites. If they were used it is
likely that many are to be found lying distributed
around the landscape. Gould (1980) estimates that
stone tools occurring at habitation base camps among
hunters and gatherers in western Australia represent
only 0.05% of used tools, with an estimated 99.95% of
used tools discarded outside habitation zones.
The presence of copper is interesting as this
suggests that copper may have been extracted to
obtain green and blue colouring. The use of colour is
also suggested by the presence of ochre, which
produces a red colour, at Howick. Evidence for the
use of colour in the Northern British Mesolithic is
scarce but the finds from Howick complement the
presence of ochre and haematite (producing black)
from the Mesolithic site of Sand, Wester Ross (Hardy
and Wickham-Jones 2003), and a copper residue
found on a bone tool from the Mesolithic site of An
Corran (Skye) (Hardy forthcoming b). Together, these
finds form the beginnings of a body of evidence that
points to the use of colour in the Mesolithic of
Northern Britain.
It is therefore interesting to note that five small
fragments of coarse-grained sedimentary rock with
copper mineral visible on the surface were found
inside the Howick hut from occupation layer [316] in
Phase 1b. That this rock and the microlith with copper
residue were found in the same deposit is tantalising.
Though it is never possible to be absolutely sure,
these finds suggest the deliberate collection of copper-
bearing rock, perhaps with the intention of crushing
and using the copper to make pigment or dye.
The lack of any specific evidence for the exploit-
ation of marine food is a gap but is probably
explained by the small sample of tools examined.
People undoubtedly exploited the sea; the shell
midden sites lying along much of the Scottish west
coast are full of fish bones, shells and evidence for the
exploitation of marine mammals (e.g. Mellars 1987;
Hardy et al. forthcoming; Hardy and Wickham-Jones
2003). Sillitoe (1988) has shown that the Wola of
highland Papua New Guinea, despite being sedentary
horticulturalists, still collected and used every
available raw material that was present in their
region, often using them in ways that could not be
imagined today. The people of Howick will have
known their environment intimately and will un-
doubtedly also have used every bit of it, in one way
or another. What remains for us to find today, is but
a tiny glimpse into their material world.
Conclusion
The evidence from the use-wear and residue analyses
suggests, when taken together, that Howick was a
site where a wide range of different activities took
place. Some of these activities are expected and form
a part of everyday survival, such as making fires,
while others such as the use of copper, probably for
pigment, show us that the people of Howick did
more than just survive; they lived. Although only a
preliminary study on a small sample of the lithic
material, this work has demonstrated the applic-
ability of such analyses to lithic collections recovered
from northern British settings.
10 FAUNAL REMAINS
Geoff Bailey and Nicky Milner
Introduction
The quantity of bone and shell material recovered at
Howick is very small. In spite of the extensive
excavations and careful recovery techniques used,
including systematic flotation, the total haul of bone is
approximately 1075 pieces, the great majority being
broken fragments less than 5mm. This is a meagre
collection by any standards and reflects highly
unfavourable conditions of bone preservation resulting
from the relatively acidic pH of the soil. Most pieces
cannot be identified even to anatomical element, let
alone to species or even to more general taxonomic
categories, and this, together with the very small
number of identifiable bones (11 in total that can be
reliably associated with the Mesolithic deposits), puts
severe limitations on the conclusions that can be
drawn. We confine our observations below to the
condition and distribution of the bone material and to
taxonomic identification. The sample is far too small
to allow any reliable assessment of the relative
abundance of different species, or useful interpretation
of the differential frequency of anatomical parts or
butchery practices, and almost totally lacking in
information on the age structure of the exploited
animals or their season of exploitation. The shell
assemblage is also relatively small and comprises shells
and shell fragments widely scattered through the
deposit rather than concentrations of shells that might
deserve the description of a shell midden, or even
isolated shell dumps. There are less than 1000
identified mollusc shells, many of which are whole or
almost whole but some of which are small fragments.
Here too analysis and discussion are largely confined
to the distribution of the shell material and to
taxonomic identifications and their palaeoecological
and palaeoeconomic significance. This chapter will
consider the bone and shell material in turn, with
particular emphasis on taphonomic issues relating to
the poor condition of the material and the likely
influence of natural processes of deposition and post-
depositional destruction. Notwithstanding the poor
quality of the surviving evidence, it is possible to draw
some useful conclusions about the likely palaeo-
economy and function of the site.
Bone analysis
Condition of Bone
The great majority of the bones appear to have been
burnt as well as fragmented, although it is often not
possible with many of the very small fragments to be
sure whether they are derived from burnt or unburnt
bone, since light or partial burning may cause only
slight changes of colouration. Many bone fragments
are heavily burnt, showing the characteristic blue-
white colour and cracking that is typical of material
exposed to high temperature and/or prolonged
periods of burning. Other bones are charred or more
lightly burnt. Many have a whitish-grey colour and
slightly chalky surface texture, which could represent
either moderately prolonged exposure to high burn-
ing temperatures or chemical degradation by soil
acids. No particular significance should be attached
to the fact that so much of the bone appears to be
burnt, since unburnt bone is likely to have been
largely destroyed by a combination of biological and
chemical degradation. However, the belief that burnt
bone is more resistant to chemical attack in acidic
soils than unburnt bone, though widely held, appears
not to be supported by experimental data (Lyman
1994, 391). A small number of bones are unburnt and
these include two whole vertebrae in fresh and
undamaged condition of hare or rabbit, which are
almost certainly recent intrusions. There are also eight
thin, plate-like fragments of unburnt bone, obviously
degraded but not burnt, heavily perforated and with
a waxy surface texture, which were found in the
vicinity of the burial cists. These range in size from
<1cm to 3cm in the maximum dimension, and look
like skull fragments, but it is impossible to say with
confidence whether they are of human or non-human
origin (see Waddington et al. 2005).
It is highly likely that we are dealing here with a
tiny fraction of the bone material originally brought
onto the site by the Mesolithic inhabitants. The high
degree of fragmentation, the generally small size of
the surviving pieces, and the widespread evidence of
burning and/or chemical corrosion, all point in this
direction. Even in more favourable conditions of soil
138
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Phase Hearth or
Burnt
Spread
Pit Occupation
Layer
Post hole/
Stake hole
Linear Discrete
Spread
Totals
Phase 1a
Heavily burnt 100 33 - 10 20 - 163
Partly burnt 57 20 58 - - - 135
Sub Total 157 53 58 10 20 - 298
Heavily burnt 34 133 30 - 10 5 212
Phase 1b
Partly burnt 308 - 2 1 - - 311
Sub Total 342 133 32 1 10 5 523
Heavily burnt - 21 30 40 - - 91
Phase 2
Partly burnt - 8 10 - - 5 23
Sub Total - 29 40 40 - 5 114
Heavily burnt 25 - 20 - - - 45
Phase 3
Partly burnt 25 2 3 - - - 30
Sub Total 50 2 23 - - - 75
Totals 549 217 153 51 30 10 1010
Table 10.1. Distribution of heavily burnt and partly burnt bone by phase and feature type in the Howick hut.
preservation, bone material would have been vul-
nerable to a heavy toll of damage and destruction
before burial, especially if dogs were present, as
seems likely at Howick (see below). Dog-feeding
experiments have demonstrated that if dogs have
access to the remains of the carcases of small and
medium-sized mammals, they leave almost no bones
except a few fragments (Payne and Munson 1985;
White 1968). More general observations of modern
hunting camps suggest that less than 5% of the bone
material of animal carcases survives if dogs are
present (Vereschagin 1967). Bones with large marrow
cavities are particularly vulnerable to destruction by
dog chewing, especially the limb bones and foot
bones. Only the most resistant parts of the bone are
likely to survive, such as the distal articulation of the
humerus (upper front limb) or the phalanges and the
articular ends of the metapodials (toe and foot bones)
(Brain 1981).
Paradoxically, charred or burnt bones are of no
interest to dogs and may be more resistant to
destructive agencies after burial, but they are also
likely, for the very reason that they have been burnt,
to become more brittle and liable to fragmentation.
Bones discarded on human settlements can be
burnt for a variety of reasons: because of exposure to
fire in the course of roasting meat; because of disposal
and burning of waste material; because fires are built
on surfaces where bones have already been discarded
or because of their use as fuel (David 1990). However,
burning changes the composition of the bone. As the
temperature increases, there is progressive loss of
organic material (which comprises about 35 per cent
by weight of bone and gives the bone its strength and
elasticity), shrinkage, and increased brittleness and
thermal cracking and splitting. Bone also goes
through a characteristic series of colour changes with
increasing temperature, from yellow to brown to
black to grey and finally to a bluish-white, although
the correlation between colour change and temper-
ature is only approximate (Shipman et al. 1984; Lyman
1994). The critical temperature for the removal of all
organic material is about 600ºC, and this stage marks
the transformation of the bone from a carbonised
state, with colouration in the yellow to black range,
to calcination, with grey or blue-white colouration.
Between about 600ºC and 950ºC, the mineral crystals
fuse to create a porcelain-like condition, with melting
of the bone occurring at about 1200ºC (McCutcheon
1992). Ordinary campfires can easily generate temper-
atures of up to 800ºC (David 1990; Robins and Stock
1990), and we should therefore expect bones assoc-
10 FAUNAL REMAINS 139
Taxon Element Feature Context Phase Condition
Fox 1st Phalange Hearth 268 1b Partly burnt
Fox Astragalus Hearth 268 1b Partly burnt
Canid 1st Phalange Hearth 355 1a Heavily burnt
Seal 2nd Phalange Hearth 355 1a Partly burnt
Seal 2nd Phalange Hearth 268 1b Partly burnt
Pig 2nd Phalange Hearth 268 1b Partly burnt
Pig Distal Fibula Hearth 355 1a Partly burnt
Small mammal Rib fragment Hearth 355 1a Heavily burnt
Small mammal Distal Phalange Hearth 268 1b Partly burnt
Bird Hearth 355 1a Partly burnt
Bird Hearth 355 1a Partly burnt
Table 10.2. Provenance of identified bone specimens at Howick.
iated with domestic fires to show the full range of
transformations in colour and composition, and this
is what we find with the Howick material. The
duration of burning is also quite critical. David (1990)
has shown experimentally that bones left in a fire for
as little as half an hour at temperatures below the
critical threshold for calcination begin to crack apart.
Long bones in particular show numerous small
thermal cracks and major longitudinal and transverse
fractures of the shaft bone. Some of these fractures
mimic the morphological features associated with the
breakage of bone during butchery of the carcase.
From all these considerations, it follows that burnt
bone is especially vulnerable to fragmentation and to
further physical damage caused by trampling and
weathering on the surface, churning of deposits and
mixing of materials during site maintenance such as
periodic cleaning out of hearths, or subsurface
movement and compression after burial. It is likely
that the Howick bone has been affected by all these
processes, further reducing the survival chances of
any identifiable pieces of bone.
bone in phases 2 and 3 contain introduced deposits
as part of their levelling layers as well as build-up of
occupation debris, with the result that bone from
these contexts cannot be relied on as an accurate
indicator of animal processing and consumption
activities within the hut.
Heavily burnt and partly burnt bones are present
in about equal proportions throughout the excavated
deposits, but there is no straightforward pattern in
their distribution. In most deposits heavily burnt
bones predominate, the notable exception being the
hearth material in Phase 1, which has the highest
quantity of partly burnt bone in the site and by far the
highest ratio of partly burnt to heavily burnt bone
(Table 10.1).
Distribution of bone
Table 10.1 shows the distribution of the bone material
according to the various features and deposits from
which it was recovered. The material that has been
obviously heavily burnt is also distinguished from
the partly burnt material, to see if the two categories
show any evidence of differential distribution.
Most of the surviving bone, accounting for 81% of
the total, and all the identified bones, come from
Phase 1 (1a and 1b) rather than later phases, sug-
gesting that preservation is better in the lower
horizons. Moreover, the great majority of the bone in
Phase 1, some 83% including all the identified bones,
occurs in the hearths and associated pit features,
implying that we are dealing with cooking activities.
The occupation layers that account for most of the
Identification
Table 10.2 summarises the identified material by
provenance and condition.
Family Canidae, dogs
Two species are present. The first species is Vulpes
vulpes, the fox, an animal useful for its pelt, but also
known to have been hunted for its meat in prehistoric
times from evidence of bones showing cut marks and
burning in the Near East (Clutton-Brock 1987).
The 1st phalange of an unspecified canid in Table
10.2 could be either wolf (Canis lupus L.) or dog (Canis
familiaris), but it is impossible to distinguish between
the two on this element. The usual diagnostic criterion
for the domestic dog is the foreshortening of the
mandible and crowding of the teeth. Dogs may also
be smaller than wolves, but not necessarily so, and no
reliable inferences can be made from the size of the
Howick bones because of shrinkage resulting from
burning. Dogs are certainly known to be present at
this time in Britain, the most famous example being
the dog from Star Carr (Degerbøl 1961), belonging to
a slightly earlier period of the Mesolithic than Howick.
140
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Family Phocidae, seals
Seals are represented by two phalanges. It is not
possible to be absolutely certain about the species,
the one comparative specimen available for ident-
ification being a grey seal (Halichoerus grypus). Both
grey seal and common seal (Phoca vitulina) have
breeding populations in British waters and are
especially common around the coastlines of Scotland,
although both may be found on the east coast of
England, and grey seals are present around the
islands off the Northumberland coast today (Corbet
and Harris 1991). Common seal are characteristically
associated with sandbanks and estuaries, whereas
the grey seal favours rocky shorelines and offshore
islands for breeding and hauling out. The ringed seal
(P. hispida) and the harp seal (P. groenlandica) both
have more northerly distributions and are very rare
visitors to British waters (Clark 1946, Corbet and
Harris 1991).
Family Suidae, pigs
The distal end of the fibula and the 2nd phalange are
certainly Sus scrofa, and presumably from wild boar.
The common assumption is that pig remains in
Mesolithic deposits represent hunting of wild boar.
However, there is a long-standing opinion that in
some areas of Europe, Mesolithic people had a closer
relationship with pig populations than is implied by
hunting of wild boar, even if the pigs were not
biologically domesticated in the sense of being
subjected to selective breeding (Rowley-Conwy 1995).
The main criterion for distinguishing domestic pigs
is their smaller size, but even that is unreliable, and
the probable shrinkage of the Howick bone resulting
from burning compromises any deductions drawn
from the size of these specimens.
In addition there are at least two fragments of
mammalian bone that can be identified to anatomical
element though not certainly to species, and a number
of bone fragments, nearly all from the main hearth
area, that are clearly derived from the shafts or
articular ends of the mammalian postcranial skeleton.
There are also two potentially identifiable fragments
of bird bone, one the size of a gull, the other the size
of a bustard, but it has not proven possible to narrow
down the identifications.
Interpretation
Species Representation
The small number of identifiable bones clearly limits
the scope of interpretation, but the range of species is
actually quite wide in relation to the very small
sample size, with evidence of sea-mammal hunting
(seal), hunting on land (boar), hunting for fur-bearing
animals (fox), fowling (bird remains), and the
probable presence of the domestic dog. This variety
of activities is consistent with the interpretation of
the Howick site as a residential settlement, rather
than a specialised hunting encampment.
This diversity is also consistent with the range of
opportunities available within the economic catch-
ment of the site. Wild boar would have been attracted
to feeding opportunities at the forest edge, along the
shoreline and the banks of the Howick Burn, as would
other land mammals. Seal are most vulnerable to
human hunters when hauled out on rocky beaches or
offshore islands, or during the breeding season,
which also takes place on land or in the intertidal
zone, when the seal pups are especially vulnerable.
Small skerries are present just offshore opposite
Howick. They are usually submerged except at very
low tides today, but would have been more prom-
inent small islands at slightly lower sea level, offering
potentially attractive locations for seal haul-outs and
perhaps for breeding. Bird life would also be more
concentrated and accessible along the coastline, and
sea birds in particular would be vulnerable to capture
when nesting on rocky cliffs.
Two types of resources are conspicuous by their
absence. The first is fish, which we might expect to be
associated with maritime activities at a coastal
settlement. No evidence of fish bone has been identified
amongst the fragmentary remains, but this may be due
either to the greater vulnerability of fish bone to the
range of highly destructive processes that have affected
the Howick bone, or to the small sample size, or to
both factors together. The second absence is any
evidence of deer. We would expect both red deer and
roe deer to be present in the Howick environment, and
to be favoured targets of human hunters in the
Mesolithic. Again, however, their absence in the faunal
remains may be due simply to the vagaries of preserva-
tion and the very small number of bones that have
survived in identifiable form.
Representation of Anatomical Elements
Here too, the condition and small sample size of the
bone material limits conclusions, but it is worth
noting the bias towards anatomical elements from
the extremities – foot and toe bones. Foot extremities
are often associated with the removal of waste parts
of the carcase, and hence with butchery sites or
locations of food preparation within a domestic site,
the more productive elements being taken elsewhere
for cooking and consumption. Grigson and Mellars
(1987) have noted that in the much larger sample of
grey seal bones from the Oronsay middens (362
identifiable seal bones), only six are burnt and four of
these are phalanges; they suggest that the seal claws
may have been thrown on the fire as waste material.
Foot bones are also sometimes left attached to a
partially butchered carcase to provide ‘handles’ that
make it easier to carry the carcase from the site of the
kill to the home base, the so-called schlepp effect.
10 FAUNAL REMAINS 141
Species Unstrat./
Surface
Topsoil
(001)
Subsoil
(003)
Stratified
from
Excavation
Stratified
from
Flotation
Stratified
Total
Total
Dogwhelk 117 508 27 17 3 20 672
Flat periwinkle 29 44 11 5 0 5 89
Edible
periwinkle
7 97 9 18 1 19 132
Limpet 2 34 8 20 8 28 72
Topshell 4 1 0 1 0 1 6
Cowrie 2 1 0 0 1 1 4
Mussel 0 1 0 1 2 3 4
Barnacle 0 0 0 2 1 3 3
Crab 0 0 0 0 1 1 1
Oyster 0 0 0 3 0 3 3
Total 161 686 55 64 17 81 966
Table 10.3. Species of shellfish identified from the Howick assemblage and their quantitative distribution by provenance.
However, neither of these interpretations is sustain-
able in the light of the small sample size and
conditions of preservation of the Howick bone
assemblage. It is more likely that the variable survival
of anatomical elements is due to factors of differential
preservation, rather than evidence of differential
treatment of different parts of the animal carcase. The
astragalus is a very dense, tough bone, while phal-
anges are also quite dense bones. They are also small,
compact elements, easily pushed into the ground and
more likely to gain protection by rapid burial from
the destructive agencies operating at the surface.
It is also worth noting the complete absence of
teeth or recognisable tooth fragments. These are
usually quite dense elements and therefore ones that
can be expected to be relatively resistant to most
destructive agencies. We suggest that their absence
from the Howick collection is probably due to the
effects of burning. Differential shrinkage of the
different layers of material that make up teeth, and
cracking, have probably reduced any tooth material
to unrecognisable fragments.
• Littorina littoralis (obtusata), flat periwinkle. Found in the
mid-shore zone living on and among algae and weed,
on which it feeds and lays eggs. The shell is small and
flat-topped and usually colourful: yellow, green, orange,
brown or red and sometimes even banded or che-
quered. It is unlikely that this shell would be collected
as food because of its small size but it could be collected
for its aesthetic qualities.
• Patella vulgata, limpet. These shells are intertidal and
are found in abundance on rocky shores.
• Gibbula sp., topshell. This is found in the mid-lower
shore amongst rocks and seaweed. When damaged the
shells have a mother-of-pearl lustre, which, as with the
flat periwinkle, may be a reason for collecting dead
shells found washed up on the seashore.
• Trivia sp., cowrie. This is found on the lower shore and
below in rock and weed.
• Mytilus edulis, common mussel. This species forms
dense beds from the upper shore to the shallow
sublittoral, often on rocky shores.
• Ostrea edulis, native oyster. These shells are found on
coarse sediments in the sub-littoral, often attached to
rocks or boulders. It is not certain that this species is
present at Howick but a couple of fragments could
possibly be oyster.
Shell analysis
Identification
The following shellfish have been identified:
• Nucella lapillus, dogwhelk. This species is found in the
intertidal zone on rocky shores in crevices and rock
pools.
• Family Littorinidae. Littorina littorea, edible periwinkle.
This is found from upper shore to shallow sublittoral
conditions on rocky shores, especially moderately shel-
tered shores with seaweeds.
Quantification and spatial distribution
Table 10.3 presents the quantities of shellfish. These
have been divided into unstratified and surface finds
(found in molehills and during test pitting), finds from
the topsoil (which was excavated by metre square),
finds from the subsoil (again excavated by metre
square), stratified contexts within the structure and
samples from the flotation of stratified deposits. In all
these contexts, there are small fragments of shell that
have not been identified. Fragments and whole shells
have been counted so the numbers represent the NISP
(number of identified specimens) and not the MNI
142
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
(minimum number of individuals). However with the
main species, the dogwhelk, the vast majority of shells
were found whole. Many unidentified fragments were
found within the flotation samples and it should be
noted that there are also 34 unidentified land snails in
this sample.
It can be seen that dogwhelks are by far the most
common species found at the site (c. 70%). Not all
edible periwinkles or limpets were found whole and
so the MNI is probably considerably lower, which
means the dogwhelk is even more dominant in the
assemblage than it appears to be on first glance. There
are more flat periwinkles than limpets, which is
perhaps surprising considering that this species is
too small for consumption and is more likely to be
collected for its aesthetic qualities. The topshell,
cowrie, mussel, barnacle, crab (a small part of the
claw) and oyster (tentative identification) are found
in very small quantities.
The majority of shells (93%) comes from un-
stratified contexts. When looking at the distribution
in the 2000 season, when the topsoil and subsoil were
excavated in squares, the shells were found right
across the excavated area. Shells were also found
eroding out of molehills across the field away from
the excavation.
Interpretation
The large quantity of dogwhelks found on the site is
unusual and different from most other sites with a
shellfish assemblage to the author’s best knowledge
(and Janice Light pers. comm.). Usually limpets are the
dominant taxon in sites located near rocky shores. A
notable exception is Ferriter’s Cove in Ireland, where
dogwhelks are the most common species, followed
by periwinkle and limpet (McCarthy et al. 1999). It is
also worth noting that dogwhelks are often present in
small numbers in other Mesolithic middens, for
example at Ulva Cave on Mull (Russell et al. 1995), at
Morton (Coles 1971), and in the Oronsay middens,
where damage and removal of the apex appears to be
due to deliberate destruction to remove the meat
(Andrews et al. 1985). It is often suggested in the
literature that this is not an edible species, and the
shells are even considered so unpalatable that fish do
not take them as bait. This is actually a fallacy and
they can be eaten (e.g. Fearnley-Whittingstall 2001). If
they were collected for food, they were probably
boiled, steamed or heated, which then makes it
possible to extract the flesh. Another method is to
break the apex, or the top of the shell, and although
some of the apices have become eroded, it looks as if
this is due to natural processes and not deliberate
breakage as in the Oronsay case. Dogwhelks can also
be used to create purple dye but it is more likely that
they would have been found broken up if utilised in
this way, in order to get to a small vein on the head of
the animal which holds the dye (Gibbons and
Gibbons 2004). It is considered unlikely that dog-
whelks were used for this purpose in the Mesolithic.
The limpets found are also an edible species; it is
interesting to note that these are small and their
morphology suggests they have come from the lower
shore, where the more succulent specimens are
usually found. This could indicate that if the limpets
were being collected for food there was some
selection for the more tender individuals.
The main species of shellfish found at Howick live
on rocky shores and this is exactly the type of
environment found below the archaeological site at
present. Most of these species can still be collected
here, with the exception of the dogwhelk. The lack of
dogwhelk, however, could be a very recent phenom-
enon caused by the effect of tributyl tin, a toxic
substance used in ship’s paint, which causes sterilis-
ation of the dogwhelk. What is interesting from an
analysis of the modern shore is that there were large
quantities of empty shells washed up onto the upper
shore, representing all the species found at the site.
The flat periwinkles are very colourful (red, yellow,
green, purple, orange) and sometimes striped and
the topshells have a mother-of-pearl sheen. They are
also very small shellfish and as said, are unlikely to
have been collected for consumption. This is also
likely to be the case with species such as the cowrie.
One of the cowries does have a perforation in it. This
can be caused naturally by boring predators, and
several natural examples were found on the beach.
However, even if the hole is natural it is possible that
these could have been collected like that and used as
ornamentation. There are parallels to this on several
Mesolithic sites on the west coast of Scotland, such as
the shell middens on Oronsay (Mellars 1987), al-
though these are later in date. Shells are well known
as having been used as currency, gaming pieces and
ornamentation throughout antiquity.
The other possibility that must be considered is
that some, or perhaps even all, of the shells appear on
the site by chance and are not related to the activities
of the Mesolithic inhabitants. One explanation for the
presence of shells on the site is that the land has been
fertilised with seaweed in the recent past, which is a
well known practice, and that shellfish which were
attached to, or scooped up with, the seaweed have
ended up in the soil. The two key indicators are the
spatial distribution of the shells and the level of
preservation. The fact that the majority of shells
comes from the topsoil and other unstratified con-
texts suggests that either there has been a lot of
movement, particularly upwards from the archaeo-
logical strata since the Mesolithic period, or perhaps
also downward movement from the topsoil. In
addition, the fact that shells were found across the
excavation area within the topsoil and in molehills in
other parts of the field is suggestive that there is a
10 FAUNAL REMAINS 143
continuous spread of shells across the cliff top, which
is consistent with the idea of seaweed fertilising.
During the cataloguing of the shells, the level of
degradation was noted using a basic scoring system
of 1, 2 and 3, with 1 being well preserved and 3 being
very degraded. Of the shells from stratified contexts,
22% were well preserved, 69% were slightly de-
graded, but only 9% were very degraded. In sandy
conditions it seems unlikely that the shells would
have survived so well, particularly the smaller
fragmented examples such as the cowries. The mussel
is usually the least resistant to taphonomic processes
and yet here, although there are only a few small
pieces of mussel, they are extremely well preserved
and retain the purple colour, which is very unusual.
What is equally plausible is that there is a mixture
of modern and archaeological shells at the site.
Perhaps there has been some relatively modern
fertilising of the land, which has resulted in the
incorporation of shellfish and particularly dogwhelks
into the soil, while some shellfish, such as those found
within the structure, could have been deposited there
in the Mesolithic period.
General Discussion
The small sample size of organic remains, the
generally poor state of preservation, and the complex
taphonomic history of both the bone and shell
material clearly set limits on interpretation. Neverthe-
less, it is possible to draw certain general conclusions
with reasonable confidence and to suggest additional
possibilities. The material is of particular significance
both because of its early date and geographical
location, and because the bone material in particular
is concentrated in hearth deposits associated with a
substantial dwelling structure.
It is clear that the Mesolithic people living at
Howick had easy access to a wide range of terrestrial
and marine resources, to plants and animals on land,
to fish in the nearby Howick Burn and in the sea, to
birds, molluscs and crustaceans along the littoral, and
to seal on rocky shores and offshore islets. In terms of
actual evidence, we have indications of hunting of
pig, fox and seal, of shell gathering and crab collecting,
and of fowling. The domestic dog was probably
present on the site. The remains thus indicate a range
of subsistence activities, all the more significant given
the small size of the sample, suggesting prolonged
occupation rather than fleeting visits for specialist
activities, and an interest in both the marine and
terrestrial resources of the local environment. The
concentration of bevel-ended stone tools in the area of
the structure is also of interest. Although these are
conventionally described as limpet scoops, Jacobi
(1984) has noted a correlation between the distribution
of these tools and the distribution of seal breeding
grounds in Wales, and suggested that the tools could
have been used for working seal skins. If correct, this
would tend to reinforce the significance of this
resource. Seals provide a range of materials including
meat, skins and sinews. The blubber can be rendered
down into storable food and fuel for eating, heating
and lighting in winter. Grey seal are particularly
vulnerable to predation during the breeding season in
autumn, when the young seals are on shore and can
be easily caught, and again during the spring moulting
season when the adults tend to haul out on land (Clark
1946, Grigson and Mellars 1987). They are thus a
particularly attractive target, providing resources that
can be stored for use during winter, and again during
the spring, a notoriously lean time of the year for
many other food resources. Pig meat is easily cured
on the bone and stored (Rowley-Conwy 1981), and
shellfish can be taken at any time of year, providing a
valuable fall-back when other foods are scarce. The
fox provides fur ideal for winter clothing, although it
could equally well have been hunted for its meat. All
of these indications are consistent with the use of
Howick as a major base in the winter months,
although there are no direct indications in the sur-
viving material that allow us to pinpoint seasonality
of use, or to establish whether the site was used
throughout the year or on a seasonal basis. It is
tempting to conclude from the presence of a durable
dwelling structure that Howick was a permanent site
occupied throughout the year. However, ethnographic
studies demonstrate that there is no necessary link
between substantial structures and year-round occu-
pation. On the Pacific North-West coast, for example,
the classic example of a richly productive marine and
coastal environment, economies dominated by fishing
and sea-mammal hunting supported large villages
with plank-built houses, which give all the appearance
of permanence (Drucker 1955). However, it is also
clear that the inhabitants of these villages moved
around the wider landscape in smaller groups to a
variety of locations at various seasons for other
subsistence activities. These patterns of use range from
fully permanent occupation at one extreme to seasonal
mobility of the whole group at the other, and many
are best described as sedentary-cum-mobile patterns
of settlement (Murdock 1967, and see also Chapter
14).
The next nearest Mesolithic site with faunal
remains, albeit of rather later date, is the Scottish site
of Morton. This site, at the time of its use, was on a
small offshore island near the estuary of the River
Tay. The archaeological evidence suggests a series of
relatively transient but regular visits at various times
of year to collect cobbles from the shore for stone
tools (Coles 1971, Deith 1983). Stake holes suggest
structures no more substantial than a windbreak.
Land mammals, fish, seabirds and molluscs are all
144
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
well represented, with the dominant taxon in each
category being red deer, cod, guillemot and cockle,
respectively.
The Oronsay middens off the west coast of
Scotland are the largest group of well studied shell
middens in Britain, including some of the largest
mounds, and with evidence of hearths and dwelling
structures. Here limpets dominate the shell samples,
with abundant evidence for fishing of saithe (a
member of the same family as the cod), seal hunting,
and deer hunting on larger islands nearby or on the
mainland.
Mt. Sandel in Ireland is of comparable date to
Howick, and has also produced evidence of sub-
stantial dwelling structures and features interpreted
as storage pits. Pig bones dominate the mammalian
assemblage and fish bones are also present in large
numbers. The site is located some 8km inland from
the coast on the River Bann, so that it is not well
placed to access marine resources, but its riverside
location is consistent with fishing for migratory
species such as salmon and eel (Woodman 1978b,
1985, van Wijngaarden-Bakker 1985). The later site of
Ferriter’s Cove is a coastal midden, which also shows
a dominance of pig bones in the mammalian remains,
together with large numbers of fish bones, especially
salmon, and marine molluscs (Woodman et al. 1999).
It is, of course, difficult to judge how far the
differences between these various assemblages reflect
differences of site location, of regional environment,
of seasonality and site function, of taphonomic factors
such as differential preservation and incomplete
sampling of spatially heterogeneous deposits, or
indeed of cultural attitudes. Some of the differences
clearly relate to local environmental conditions, for
example the presence of seal at Oronsay but its
absence at Morton, or the dominance of limpets in
the former case and cockles in the latter. The absence
of red deer at the Irish sites is clearly not the result of
small sample size, as we have suggested for Howick,
but may rather reflect the rarity of deer in Ireland
because of its early isolation from the British main-
land (Woodman et al. 1997). In sites that have been
extensively excavated with analysis of spatial vari-
ation in remains, notably at Morton and Ferriter’s
Cove, there are clearly considerable intra-site differ-
ences in the representation of different classes of food
remains and different taxa, as between shell dumps
and other sorts of surfaces, for example. Some of
these differences reflect differences in the way
different food remains were disposed of, others
appear to reflect local variations in preservation
conditions. Both these factors are likely to be of
importance at Howick, and we have seen hints of
such variation in relation to differences between
hearth deposits and other surfaces and features.
The Howick material does not allow us to pursue
the effects of intra-site spatial variation very far, but
such factors are probably of considerable significance,
especially in explaining some of the differences in
faunal representation compared to the other sites we
have mentioned. The two respects in which the
Howick fauna stands out from the other sites are in
the absence of fish bones and the relative rarity of
molluscs. Given the fairly substantial and prolonged
nature of occupation at Howick, as suggested by
other lines of evidence, we think it unlikely that the
inhabitants would have ignored locally available
resources such as fish and molluscs. The absence of
fish is probably due to the very poor conditions of
bone preservation, reinforced by small sample size,
though we cannot absolutely rule out other poss-
ibilities. It seems unlikely that fish were locally rare,
but not impossible that people chose to ignore them.
The rarity of the molluscs is less easily explained and
calls for some further comment.
Molluscs are widely available in the intertidal zone
as an easily collected complement to the more
unpredictable and intermittent outcomes of hunting
on land and at sea. Even though the shoreline was
further away from the Howick site than it is today, it
is unlikely to have been more than a few hundred
metres away. It is well known that molluscs are quite
expensive to carry any distance because of the high
weight of the shell relative to the edible flesh, and
that the molluscs are often cooked and the shells
removed on the shore, even if the meat is then taken
away for consumption at settlement sites further
inland. However, numerous ethnographic and arch-
aeological records (e.g. Meehan 1982; Bettinger et al.
1997; Bird et al. 1997; Bailey and Craighead 2003)
show that molluscs in the shell are often carried at
least 1km and sometimes more, the critical threshold
being about 5km, depending on the particular species
in question. Even if many shells are cooked on the
beach, it is rare that some are not also taken back to
settlements situated a little further inland, although
there may be a rapid fall-off in quantity with
increasing distance from the shoreline.
With the above considerations in mind, we suggest
that there are five hypotheses that might account for
the relative rarity of shells at Howick
1. Molluscs were collected in considerable quantities but
the shoreline was further away than today because of
lower sea level, and much of the processing of the shells
and perhaps some of the consumption of the mollusc
flesh took place nearer the Mesolithic shoreline, with
relatively few shells being carried back to the Howick
settlement. This remains a possible explanation as we
have discussed above, but we think it insufficient by
itself. The Asturian shell middens of northern Spain,
which are of comparable age to Howick, are on a rocky
coastline with similar molluscan habitats and potential
productivity. They are also associated with a broad-
based subsistence economy including hunting of deer
and boar as well as fishing and shellgathering. There
10 FAUNAL REMAINS 145
are over 50 such sites, most formed in the mouths of
small limestone caves or rockshelters. However, even
though many of these sites are at least several hundred
metres from the contemporaneous shoreline, and some
up to 5km distant, the attractions of shelter and comfort
afforded by the rockshelters clearly made it worthwhile
to carry loads of limpets and topshells back to these
locations. The great majority of the Asturian sites are
proper shell middens, comprising thick deposits of
concentrated shell, some forming quite substantial
mounds that eventually filled the cave entrance (Bailey
and Craighead 2003). However, we should beware of
placing too much emphasis on analogies with other
regions. The main reason why shells were carried
inland from the shoreline in the Asturian case probably
has to do with the shelter provided by the rockshelters
and their attractiveness as bases for hunting and
gathering on land. We are not aware of any similar
examples in Britain, but we might suggest that the hut
structure at Howick exerted a similar attraction to the
rockshelters of Asturias in encouraging the occupants
to carry back some of their mollusc shells from the
nearby shoreline.
2. The Howick site was a specialised camp used for brief
periods at a time to target one or two resources, ignoring
others in the vicinity. We believe this to be highly
unlikely, given the evidence for a wide range of
subsistence activities and the investment in the building
of at least one substantial dwelling structure that was
clearly repeatedly used over a long period. It is worth
emphasising the fact, noted above, that the Morton site
has been interpreted as a task-specific site, primarily
visited to collect stone material from the beach for
artefacts. Yet mollusc shells were collected as food,
along with other resources, even in seasons when the
molluscs were not in their best condition for eating,
precisely because they provided an easy and instant
supply of food for people primarily engaged in other
activities.
3. There was a cultural taboo against shellfood gathering
or consumption. Moss (1993) provides a useful dis-
cussion of food taboos and shows that, amongst the
Indians of the Pacific North-West Coast, there were
widespread taboos against eating shellfish because of
its association with disease, sexuality and low status.
Nevertheless, the taboo was often relaxed, not only for
people of low status with less access to high-status food,
but also for people of higher status when other foods
were scarce. As a resource that is easily accessible and
easy to collect, shellfood is of particular value for the
weaker members of the community – children, mothers
with infants, and old people – and for all members of
the community at times when other resources are in
short supply (see also Meehan 1982). Also, notwith-
standing the evidence of taboos on shellfish, the coastal
settlements of the North-West Coast are associated with
many substantial shell mounds that continued to be in
use into the historical era. We therefore consider this
hypothesis unlikely in the Howick context.
4. The shells were collected and brought back to the site in
quantity but have been subjected to the same heavy toll
of destructive agencies as the animal bone. Mollusc shells
are usually regarded as being far more robust and
resistant to destructive agencies than bone. Dogs have
no interest in them, and if the shells are dumped in
sufficient quantity or sufficiently rapidly in one place,
they can create their own calcium-rich environment that
neutralises the effect of acids in the surrounding
sediment. Nevertheless, they are vulnerable both to
fragmentation if burnt (Robins and Stock 1990), to
deflation and dispersal if located on exposed sandy
sediments, and to removal by chemical action and
leaching out of organic material, even when accum-
ulated as concentrated deposits. Stein (1992) provides
examples and cogent evidence in support of the hy-
pothesis that shells have been completely removed from
the lowermost levels of some shell midden deposits in
British Columbia because of post-depositional chemical
action. Given the hostile conditions for preservation of
organic materials at Howick, we think it quite possible
that many shells have been destroyed, especially if they
were originally accumulated in small concentrations or
scattered through the deposit. The fact that dogwhelks
are the most common species, a shell that has a
particularly robust shell and compact shape, could be
consistent with such an interpretation.
5. There was considerable spatial differentiation of activ-
ities and deposits, with concentrated shell dumps being
located outside the area of the main hearths and
dwelling structure in parts of the site that have now
disappeared because of cliff erosion. In both the Oron-
say shell middens (Mellars 1987) and the Ertebølle shell
mounds of Denmark (Andersen 2000), there is good
stratigraphic evidence to suggest a separation between
hearth areas, associated in the Oronsay case with
dwelling structures some 3m in diameter, and shell
dumps situated outside this zone of intensive domestic
activity. The hearth areas have slowly accumulating
surfaces of heavily fragmented and compacted shell
associated with high proportions of ash and sediments
and abundant bone remains and stone artefacts, while
the shell dumps comprise more rapidly accumulated
concentrations of more loosely packed shells, with little
sedimentary matrix and less cultural debris. This is
particularly evident at the classic site of Ertebølle itself,
a shell mound originally some 140m long by 40m wide
and 2m thick. In addition, there is an area behind the
mound and covering much the same area as the mound,
which is full of bones, stone artefacts and cultural
features, but completely lacking in shells, suggesting
large-scale spatial differentiation of activities (Andersen
and Johansen 1986). Similar variability in spatial
patterning is present at Ferriter’s Cove and at Morton,
as noted earlier. It is possible that there was a similar
spatial differentiation of activities and discard patterns
at Howick, and that the areas of concentrated shell
accumulation were on the seaward side of the dwelling
structure, long since destroyed by erosion at the cliff
edge.
We cannot be certain which of these hypotheses, or
what combination of them, is most appropriate to the
Howick case, but we believe that the factors assoc-
iated with the last two hypotheses (4 and 5), perhaps
reinforced by distance from the contemporaneous
shoreline (as in hypothesis 1), provide the most
146
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
plausible explanation of the relatively small amount
of shell material, and that shellgathering was prob-
ably a regular part of the subsistence schedule, even
though little evidence of it now survives.
significant that all the sites we have discussed, and
especially those with archaeological evidence for
substantial occupation, show good evidence either
for sea fishing (including salmon) or sealing, and in
the case of the Oronsay middens, arguably one of the
biggest concentrations of coastal middens in Britain,
evidence of both subsistence activities.
Conclusions
In conclusion, we re-emphasise the wide range of
subsistence activities represented at Howick, both
marine and terrestrial, and the preponderance of
activities that can be associated with the colder
seasons of the year, or preparation for them. We
highlight in particular the evidence for seal hunting.
Sea mammals and abundant supplies of fatty fish
such as cod and salmon play a crucial role in
providing fat in the human diet, without which
survival in cold, northerly environments is imposs-
ible, especially in sub-Arctic conditions. We suggest
that this factor was of considerable significance in the
Mesolithic of northern Britain, and that the local
availability of such resources is likely to have played
a key role in determining the attractiveness of
particular locations for human settlement and the
extent to which they were capable of sustaining
prolonged occupation for long periods of the year,
especially over the winter months. We consider it
Acknowledgements
We are indebted to the following individuals for
access to comparative material and assistance with
identifications: Umberto Albarella, Department of
Archaeology, University of Durham; Hannah Koon,
Department of Biology, University of York; Terry
O’Connor, Department of Archaeology, University of
York and Jessica Rippengal, Grahame Clark Archaeo-
zoological Laboratory at the McDonald Institute for
Archaeological Research, University of Cambridge.
The Wolfson Bioarchaeology Laboratory, University
of Newcastle upon Tyne, provided facilities for the
analysis. We also gratefully acknowledge the support
of the Leverhulme Trust for the award of a Major
Research Fellowship to GB, which has facilitated the
completion of the research on which this chapter is
based.
11 PLANT REMAINS AND CHARRED WOOD
Jacqueline Cotton
Summary
The initial evaluation of the Howick site during 2000
revealed large quantities of charred archaeobotanical
remains in the Mesolithic hut deposits. This observ-
ation gave rise to the adoption of a 100% sampling
strategy for all excavated Mesolithic deposits. The
entire contents of each deposit identified within the
hut were floated and sieved and both the resulting
flot and residue were scanned for plant macrofossils.
Initial assessment of the samples determined that vast
quantities of charred hazelnut fragments were
preserved within the Mesolithic hut. A detailed
analysis of the data has been undertaken to ascertain
the significance of the remains with regard to
individual features, the status of the site and the
exploitation of plant resources during the Mesolithic
period. Few archaeobotanical remains were pre-
served in the upper stratigraphy of the hut.
The charred hazelnut fragments at Howick rep-
resent processing, waste and incidental deposits,
indicating the deliberate and continual use of hazel-
nuts during the occupation of the site. Hazelnut shells
may have been roasted and, after consumption of the
nut, utilised as fuel. The quantities present in the hut
suggest that nuts formed an important component in
the inhabitants’ diet. The presence of so many charred
nut shells also suggests that nuts may have been
regularly collected to be kept as a stored foodstuff for
consumption during winter and spring months when
food may have been less readily obtained. The
absence of other archaeobotanical remains that may
also have contributed to the diet is attributed to poor
preservation conditions.
The presence of hazelnuts through all three phases
of Mesolithic occupation implies that some degree of
woodland management may have taken place to
ensure a constant supply of nuts over the life of the
hut – some 150 to 200 years (see Chapter 6). Analysis
of the temporal pattern of waste from hazelnut usage
indicated a decline over time. This could be attributed
to a change in reliance upon hazelnuts in the diet, a
reduction in their availability, or more likely, a
shortening in the duration of occupation at the site as
indicated by the radiocarbon dating.
Initial evaluation of the material determined the
presence of large numbers of charred hazelnut
fragments (hereafter termed CHF). The quantity and
age of these remains are significant in the context of
the socio-economic status of the site and with respect
to the characteristics of British and North-West
European Mesolithic settlements. Thus, a detailed
analysis of the environmental data has been under-
taken.
Method Statement
All samples of sediment from the hut were processed
on site using a purpose-built flotation tank which had
a mesh attached to obtain non-organic finds (such as
flint) as well as a series of graduated brass sieves
below the overflow in order to catch organic materials.
The residue was collected in trays from the base of the
tank. Each sample was floated and sieved through a
2mm mesh, with flot residues collected in graduated
sieves with the smallest being 500m (0.5mm). Both
flot and residue were retained, air-dried and scanned
at x40 magnification for environmental remains. The
compositions of the flot and residue matrix were
recorded and total counts of the botanical remains
were made. It is acknowledged that where thousands
of CHF were preserved, there will be a small margin
of error for the totals obtained, but this does not
significantly affect the results and interpretations.
A quantitative assessment of the data has been
undertaken with regard to the relative concentrations
of CHF within each context. The number of CHF per
volume of sediment processed from each context has
been calculated. This produces information relating
to the concentration of CHF in the deposits and
provides an indication as to the nature of accum-
ulation and whether remains are considered to be
incidental deposits, deliberate dumps of material or
the result of in situ activities. The number of CHF per
volume of flot and residue has also been calculated.
This records the concentration of CHF compared to
other small finds and coarse material retained
following sample processing.
148
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 11.1. Charred hazelnut from the Mesolithic structure Phase 1a.
Archaeobotanical Residues
The complete set of results from the environmental
analysis is provided in Table 11.1. The results are
discussed below according to phase of occupation.
Phase 1a
Within the hut structure was a series of small circular
features identified as post sockets. Nine fills from
these circular features were sampled for environ-
mental analysis. One of the nine features contained
no charred archaeobotanical remains, while the
remaining eight contained small to moderate num-
bers. The general trend of limited finds indicates that
the charred hazelnut fragments preserved in the
contexts resulted from the inadvertent deposition of
residual waste, possibly transported into the contexts
from more extensive waste deposition or hazelnut
processing elsewhere inside the hut. The concen-
tration of CHF in the samples was consistent (Fig.
11.1) suggesting similar rates of infilling with waste
over time. Occasional charcoal finds were extracted
from the context flots, but these were insignificant in
number and, like the CHF finds, their presence may
be inadvertent.
One of the stake hole features, context 244, was
sampled for environmental remains. The processing
of the sample produced a small flot, containing no
CHF and small quantities of charcoal fragments.
One of two linear features, context 359, which is
thought to have formed part of the structural
remains, was sampled for environmental analysis.
Little evidence of burning was observed in this
feature, although processing of the sample produced
a small flot containing 255 CHF. The volume of flot
and residue remaining after processing was small,
indicating that most of the fill was fine-grained sandy
material, which was not retained. The relatively low
numbers of archaeobotanical remains preserved
indicate that waste was not deposited directly into
this linear feature.
The central hearth [355] displayed evidence of
multiple burning events and a series of different cuts
resulting in an irregularly shaped pit. Analysis of the
fill revealed the presence of over 42,000 CHF. This
vast number suggests intensive use of hazelnuts
while the evidence of multiple burning events
indicates that hazelnuts were used continuously
during the life of the hearth. To the east of the primary
hearth were two further large hearth pits, [357] and
[379], which also contained significant numbers of
CHF (Fig. 11.2).
The quantity of archaeobotanical remains in the
primary hearth suggests that it was the main area for
roasting or cooking hazelnuts, while the similarly
important quantities in contexts [357] and [379]
indicate that these features were used for similar
purposes. Very little charcoal was preserved in the
samples when compared with the vast numbers of
CHF. This absence is partly the result of the poor
conditions of preservation which have encouraged
degradation so that most charcoal remains are <2mm
in size and were broken down instantly during
excavation. The absence of other edible plant food
items within the hearth is most likely due to the fact
that hazelnut shell is strong and robust, especially
when burnt, and survives well, whereas most other
plant foods would have left little or no trace.
However, it is noteworthy that this hearth [355] did
produce many small burnt bone fragments.
Two later hearths were cut into the primary central
hearth. The fills from these features, [291] and [367],
contained 100 and 80 CHF respectively, indicating
11 PLANT REMAINS AND CHARRED WOOD 149

Fig. 11.2. Total charred hazelnut fragments from the Mesolithic structure hearths.
that hazelnuts may have been burnt or roasted in
these hearths, but not on the same scale as the
primary hearth [355].
The four other features preserved within Phase 1a
of the Mesolithic structure were pits of varying size
and of probable varying function. The two smaller
pits, [274] and [351], contained 190 and 25 CHF
respectively, while both contained occasional charcoal
fragments. As no evidence for in situ burning was
observed in these fills the presence of CHF and
charcoal suggests that some waste material did
accumulate in these pits. It is possible that the pits
were used for food storage and preparation. The
relatively low numbers of CHF in the pits indicate
that either roasted nuts were not stored in the pits, or
that charred waste fragments were separated from
the edible nuts prior to storage. If food was prepared
or stored prior to roasting then no evidence would be
preserved as the acidic conditions at the site are not
conducive to the preservation of organic material that
has not been charred.
Two larger pits, [365] and [270], were also sampled
for environmental analysis. From the 75 litres of
sediment processed from [365], 2240 CHF were
extracted. The large numbers of remains from [365]
indicate that waste material accumulated within the
pit. As there was no evidence for in situ burning, the
hazelnuts will have been charred prior to deposition.
A significant number of CHF were preserved in pit
[270], 3275 from 68 litres processed, which is very
high compared to finds in other socket or posthole
fills, even when accounting for the large quantity of
sediment processed. Finds in socket or post hole fills
are often inadvertent remains, accumulating within
the depression created by the timber by the actions of
gravity, wind and accidental discard. The large
numbers within [270], however, suggest that waste
was purposely deposited in the pit. Therefore, it may
be that the pit was originally used for storage or
structural purposes and was then later filled in with
debris associated with hazelnut roasting activities. At
some point after this a post was inserted into the
secondary socket.
Phase 1b
Samples [332], [334], [336], [338] and [342], from a
group of small, shallow stake holes, preserved only
low numbers of CHF in four of the stake holes and no
charred remains in the fifth. The stake holes did
contain mineralised wood, confirming their interpret-
ation as the surviving tips of timber stakes, but this
did not survive excavation and sample processing.
A larger stake hole fill, context [310], contained 101
CHF; a moderate number, presumably resulting from
infilling with waste material over time. Charred birch
wood fragments were found within the fill, while
environmental analysis uncovered a single birch seed
in the sample. Due to the aerobic conditions across
the site and the acid nature of the substratum, it is
unlikely that even hardy organic matter could survive
for thousands of years. Birch seeds are particularly
fragile and could not survive for long periods of time
in an aerobic environment. Therefore, it is possible
that the birch remains are intrusive. However, these
deposits were deeply buried below the subsoil and
may have experienced less acidic localised conditions
resulting from the charred ends of the stakes that
allowed for the survival of these botanical remains.
An ‘r’-shaped shallow spread of material, context
[298], contained a relatively large number of CHF,
with 1165 present within the 210ml of flot and
residue. These remains however, may derive from
the underlying deposits within context 340.
150
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Fig. 11.3. Charred Hazelnut from the Mesolithic structure Phase 1b.
The occupation deposit, context [316], was exten-
sively sampled and was found to contain 2065 CHF,
together with small finds of charcoal, shell and
waterlogged seeds. The latter are unlikely to be
contemporary with the context and are likely to have
been introduced by way of mole disturbance. The
numbers of charred remains preserved in the oc-
cupation deposit are high, though it should be
acknowledged that the remains were taken from 316
litres of sediment excavated from a deposit that
covered a large area of the interior of the hut.
Comparison of the numbers present in context [316]
with spreads of burnt occupation deposits (Fig. 11.3)
indicates that remains were uniformly spread across
the hut floor.
Immediately overlying the central hearth [355] was
context [340], an extensive spread of burnt material.
Environmental analysis of the sample extracted from
the spread has determined the presence of over 15,000
CHF. Located above this feature were two additional
burnt spreads or hearth features. One spread, context
[264], contained over 11,000 CHF. The other, [320],
contained 2750 CHF. Together, these large numbers
of hazelnuts represent waste material, perhaps from
extensive processing, roasting and burning. Cutting
into the north-east side of burnt spread [264] was a
smaller burnt deposit or hearth feature, [293]. The
volume of CHF deposition within this feature was
lower than for the feature it cut into, with only 850
remains preserved. It is noted that the flot and residue
from which these CHF were extracted were only
small and therefore the concentration of remains was
relatively high. Context [282], also associated with
the stratified hearth features, produced 950 CHF from
a relatively small quantity of sample.
Hearth feature [268] contained evidence for in situ
burning and a significant concentration of CHF; 8340
from 91 litres of sediment were preserved in its fill.
These remains indicate the continued exploitation of
hazelnuts and the processing, burning and roasting
of nuts within the structure. Bone and flint were also
preserved in this hearth feature.
Pit [285] was cut into burnt spread [264] and
contained 4073 CHF. The feature was initially inter-
preted as a raking pit associated with the hearths.
Although additional degradation of CHF was not
observed in the material from [285], the quantity of
remains present suggests that waste material was
deliberately deposited within this pit. The two other
Phase 1b pits, [214] and [324], contained insignificant
numbers of CHF. Due to the absence of further
botanical remains or other small finds, environmental
analysis cannot further the interpretation of these
pits.
Phase 2
Environmental analysis of large volumes of sediment
from context [210] revealed the presence of over
16,000 CHF. In addition, several other fill sediments
and spreads of material associated with context [210]
contained substantial numbers of CHF. The quantities
of remains are presented in Figure 11.4. The total
number of CHF within all of these contexts is c.
50,000.
The provenance of this large number of nut shells
is not certain other than to say that they derived from
occupation deposits and possibly levelling deposits
introduced from outside the hut (see Chapter 5). If
these charred hazelnuts derived from roasting and
burning waste deposits outside the structure, then it
is probable that they represent only a proportion of
the full extent of nut waste that was originally
produced.
Two of the perimeter stake holes contained min-
eralised wood, attesting to their interpretation as the
11 PLANT REMAINS AND CHARRED WOOD 151

Fig. 11.4. Charred Hazelnut fragments from contexts associated with the levelling layer between phases 1 and 2.

Figure 11.5. Charred hazelnuts from the stakeholes and post sockets of phase 2.
remnants of charred tips of timbers set into the
ground, while the fills of two, [262] and [263], that
were sampled for environmental analysis, contained
450 and 1047 CHF respectively. This large concen-
tration of CHF suggests that these features were near
to areas of cooking, processing or waste deposition.
In comparison (see Fig. 11.5), two post sockets near to
the stake holes contained low numbers of CHF. The
insignificant quantities of CHF in [113] and [224]
indicate the inadvertent accumulation of charred
material.
Pit fill [160] contained 1910 CHF, a moderate
number for the 43 litres of sediment that was
processed. Context 276, a fill associated with this
large pit feature, also contained a moderate number
of CHF. A similar concentration of CHF was preserv-
ed in [206]. In contrast, the pit fill of [192] contained
small numbers of CHF, while insignificant numbers
were preserved in pit fill [214]. The variable quantities
of CHF remains within these pit fills indicate that
each may have performed a different function within
the structure and that hazelnut waste was more
frequent in pits [160], [206] and [276].
One dish-shaped pit, [109], had a clear heat-fused
lining and contained a hazelnut-rich fill. The base of
the pit included a fused surface impregnated with
countless charred hazelnuts while the fill comprised
heat-affected sand with clay fraction and charred
remains. Environmental analysis of the lining and fill
determined the presence of 2040 CHF in the former
and 3590 CHF in the latter, although vast numbers of
charred shells broke down into charred dust during
excavation of these sediments, so the actual number
of CHFs is substantially higher. Furthermore, fill
152
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
[236], which was found to be an extension of pit fill
[109], also contained c. 3600 CHF. The preservation of
the large numbers of charred remains and the
evidence provided by the heat-affected base leads to
the interpretation of this feature as a designated
roasting pit. Hazelnuts can be roasted and dried in
order to preserve them for longer periods. The
exploitation of resources in this way is discussed
further below. An additional fill, [259], associated
with the roasting pit, contained a moderate quantity
of CHF and charcoal, but also included a degraded
charred seed that could be recognised from the
morphology only as a member of a grass family. This
poor state of preservation is an indication of why,
other than hazel, charred species have not been found
at the site.
over the ensuing 10,000 years.
The fill of the largest Phase 3 hearth, [47], included
evidence for in situ burning, but only a relatively
small number of CHF were present in the flot and
residue; 890 from 20 litres of sediment. This could
indicate that waste was removed from this hearth or
that hazelnut roasting was not an important activity
associated with the hearth. The fill of hearth [173]
contained fused sand, indicating in situ burning, and
contained a large concentration of CHF, although the
numbers of remains are not as high as underlying
Phase 1 hearth features.
Two pit features, thought to be associated with the
hearths, were sampled for environmental analysis.
Present in the fill of pit [45] were 1655 CHF and a
moderate quantity of charcoal. The other pit, [53],
contained few remains, but a relatively high concen-
tration of CHF, similar to context [45]. This indicates
that both pits were the recipients of hazelnut roasting,
burning and processing waste, probably associated
with hearth feature [47].
Two additional pits, located towards the edge of
the hut, were also sampled for environmental
residues. The fill of pit [65], an irregularly shaped pit
with no evidence for in situ burning, contained a low
number of degraded CHF and some degraded bone.
The poor state of preservation of these finds suggests
that the remains were moved or trampled prior to
deposition. The fill of post pit [106] contained
relatively few CHF, consistent with its interpretation
as a structural pit for timber [41], and suggesting that
waste from the hearth did not spread to the edge of
the hut.
Phase 3
Environmental analysis of a large volume of sediment
from occupation/levelling layer [49] identified the
presence of over 44,000 CHF. It was acknowledged
by the excavators and during the environmental
analysis that variation in charred nut shell densities
occurred across this occupation layer. This may
potentially reflect variability in the provenance of the
deposits, in that some derived from waste deposits
outside the structure, while others may have accum-
ulated in situ as a result of activities within the hut.
A sample taken from post socket [148] contained
60 CHF. On the north-east side of the hut, post socket
fills [35], [51] and [67] were sampled for environ-
mental analysis. Each of these fills, like that of [148],
contained few CHF, mostly derived from residual
waste material that will have accumulated in the
socket depressions. Post socket [35] also contained a
moderate amount of charcoal, also likely to be
residual fuel waste. The environmental sample from
pit [67] contained an insignificant quantity of CHF.
Post socket fill [266] contained a moderate number of
CHF remains (455) although these may also be
residual waste deposits that steadily worked their
way into the socket depression.
Two major timber stains were preserved from
Phase 3. The fill of one, context [41], included
mineralised wood that was evidently a fallen former
upright. The moderate volume of charcoal in the fill
could have originated from some burning of the
upright, or from fuel waste. Interestingly, the fill also
contained c.2200 CHF. This substantial number of
remains is unlikely to be residual waste, and suggests
either that the depression created by the fallen timber
became in-filled with waste material, or that, more
likely, the CHF residues derive from material that
had become worked into the area of the mineralised
wood due to pedogenic processes and bioturbation
Phase 4
The Phase 4 occupation is believed to have been short
in duration (see Chapter 6). A small burnt feature [55]
from this phase was sampled. This contained c.170
CHF. This is a relatively low number of remains and
could relate to the small size of the feature and the
short period of occupation that comprised Phase 4.
Features outside the Mesolithic hut
Outside the hut a series of pits and scoops was
located which may be linked to the Mesolithic
occupation of the site. Scoop fill [13] was interpreted
as archaeological in origin. Environmental analysis of
the fill found no charred archaeobotanical remains or
small finds that could indicate a chronology. Context
9 was the fill of a clay-lined pit within which stake
holes were uncovered. Apart from the scorched and
fused base of this pit, the only evidence of fire-based
11 PLANT REMAINS AND CHARRED WOOD 153
activity was 17 CHF, although this small number is
not significant and could have blown or washed in
from nearby areas of more extensive debris. More-
over, the presence of hazelnuts does not provide a
positive indication of chronology, as hazelnuts have
been preserved at sites dating from the prehistoric
through to medieval periods in Northern England
(Huntley and Stallibrass 1995).
The fill of a timber stain, Context 72, located
outside of the Mesolithic structure, was analysed and
found to contain 250 CHF alongside occasional flint
and shell fragments. These waste deposits within the
timber stain are similar to those found in the timber
stains within the structure, and probably comprise
residual waste material which slowly worked its way
into the depression formed by the timber as a result
of bioturbation and pedogenic processes.
Charcoal Analysis
Four samples of material from the Mesolithic hut
believed to be charcoal were submitted for identif-
ication. Due to the poor preservation, two of the
charcoal samples could not be identified. The two
charcoal samples that could be identified were hazel
(Corylus avellana). These samples could have been
burnt as fuel, or could have been inadvertently
collected with hazelnuts and charred when the nuts
were burned or roasted. The issues surrounding the
limited quantities of charcoal preserved at the
Howick site are discussed below.
Discussion
Charred archaeobotanical remains at Howick
It is acknowledged that the presence of surviving
charred archaeobotanical material at a site such as
Howick is significant. At Mesolithic sites in the British
Isles, significant quantities of archaeobotanical
remains are rarely found (e.g. Cormack and Coles
1968; Huntley and Stallibrass 1995). This has proved
a major limiting factor for developing an under-
standing of the extent of wild plant usage and
exploitation as well as the impact on the landscape as
a whole. In northern England, significant finds of
charred plant material pre-dating the Iron Age are
seldom found (van der Veen 1982; ASUD 2000),
further enhancing the importance of the finds at
Howick.
Beyond the importance of the preservation of
charred hazelnuts there is a need to address the
question of whether the remains represent debris
from the opportunistic consumption and use of
hazelnuts, or whether they reflect intensive exploit-
ation of a resource (Zvelebil 1994). The sheer quantity
of hazelnut residues at Howick, together with the
roasting pit feature, are suggestive of intensive
exploitation, though given the duration of the site the
volume of nut debris may instead just reflect a steady
accumulation over a century or two. Charred hazel-
nuts preserved in Mesolithic features at Thatcham,
Berkshire were estimated to constitute c. 120 nuts
and suggest opportunistic usage or “snack food”
(Healy et al. 1992). In contrast, the several thousand
remains at Staosnaig, Scotland are interpreted as
deliberate plant usage (Mithen et al. 2001). As the
numbers of fragments at Howick are probably
present in higher numbers than at Staosnaig, and in
some cases occur in concentrated masses in a range
of deposits (e.g. roasting pit, hearths, occupation
layers), they are considered likely to represent
‘intensive’ exploitation of wild plant resources (cf.
Zvelebil 1994).
Charred hazelnut fragments were preserved in the
full range of features at Howick, including hearths,
pits, stake holes, post sockets, timber stains and
occupation debris layers. Analysis of the quantities of
finds, in conjunction with evidence of in situ burning
and the possible function of features leads to a
division of the context fills into three modes of CHF
deposition.
1. Processing deposits. Features where hazelnuts are burnt
or roasted in situ. Large numbers of fragments are often
preserved.
2. Waste deposits. These are features without evidence for
in situ burning where used or ‘failed’ hazelnuts are
deposited. They could also include areas of storage.
Moderate to large numbers of fragments are preserved.
3. Incidental deposits. Depressions or occupation layers
into which fragments have inadvertently accumulated
with no evidence for in situ heating and usually
associated with structural features. Small to moderate
numbers of fragments are preserved.
Nearly all the contexts from the hut fall into one of
these three categories, implying that hazelnut exploit-
ation took place on a significant scale. It creates a
picture of the floor area being permanently covered
with scatters of charred hazelnut fragments. The
presence of the dish-shaped feature with a heat-fused
base with an uncountable quantity of associated nut
shells from the phase 2 hut suggests that hazelnut
roasting was a formalised process that involved an
organised routine for dealing with this plentiful food.
Furthermore, from the fact that hazelnut roasting
appears to have taken place it can be reasonably
inferred that hazelnuts were being prepared for long-
term storage, perhaps over the winter months after
nut collection in the autumn. This seasonal indicator
is important in helping to build up a picture of when
the Howick site was occupied, as well as providing
evidence for what may have been large-scale harv-
154
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
esting of resources. Given the duration of occupation
indicated by the radiocarbon dating, the possibility
arises that long-term management of the surrounding
woodland may have taken place in order to prop-
agate hazel growth and avoid over-exploitation.
However, at present this remains only speculative in
the absence of further independent evidence.
Charcoal
Charcoal was commonly preserved in the Mesolithic
features but never in significant quantities. There were
only rare finds of charcoal large enough to enable
identification. Ideally, at least 1cm3 of charcoal should
be used for identification. This ensures that most
diagnostic features are preserved and recognised to
allow a confident identification. Positive identif-
ications can be made on smaller fragments but in the
case of many of the Howick samples they were still
too small. Numerous features at Staosnaig also had a
limited presence of charcoal (Mithen et al. 2001). The
excavators attributed this to the function of the
features, which were not thought to have experienced
in situ burning. In contrast, many features at Howick
contained heat-fused and fire-reddened fills but still
very little charcoal that did not disintegrate on
excavation. The excavators noted that although
charred wood could clearly be seen in the Howick
deposits, as soon as it was touched, scraped or moved
it disintegrated into powder and consequently few
pieces could be salvaged for identification or dating
purposes.
The low quantities and small sizes of fragments of
charcoal at Howick could be attributed to a number
of causes, the first of which is taphonomy. Flecks of
charcoal were observed in many contexts but not
preserved in the flots. This could suggest that the
fragments were small and degraded and thus not
retained during the processing procedure. Moreover,
when charcoal is saturated it becomes more friable or
‘crumbly’. Thousands of years of wetting and drying
could have restricted preservation.
Secondly, it may be that charcoal was not the sole
source of fuel. Hazelnuts are thought to burn well
and may be used in areas where hazel is the main tree
species and needs to be retained as a food resource
(Mason 1996). Consequently, both the hazelnuts and
the charcoal could be viewed as fuel waste.
Scotland, Mithen et al. (2001) proposed three ways in
which hazelnuts became charred. Firstly, shells from
nuts opened in a raw state are unintentionally charred
in hearths. Secondly, hazelnuts are used as fuel.
Lastly, charred hazelnuts represent burnt residue
from areas of roasting. A fourth possibility is that
none of the above occurs exclusively and debris
results from a combination of the processes.
If the debris represented shells from already
opened nuts, burnt in hearths, this would suggest
that the majority of fragments would be preserved in
hearth features. At Howick, this is not the case. Vast
numbers of remains are also present in pit fills (e.g.
contexts [206] and [270]), burnt spreads of material
(e.g. context [118]) and in debris layers (e.g. contexts
[49], [232] and [210]), suggesting that no single
process such as inadvertent charring in hearths could
have occurred. Furthermore, this explanation fails to
account for the roasting pit that had charred hazelnut
shells embedded in its heat-fused lower lining.
The second suggestion was that hazelnuts were
used as fuel (Mithen et al. 2001). As discussed above,
this is likely given that hazel wood was also present
in the hut deposits. Indeed the dried-out shells may
have provided a good source of fire-lighting material.
However, to use hazelnuts solely as a fuel source
would be wasteful and the kernels will have been
consumed prior to burning the shell. Considering the
vast numbers of hazelnut shells in features where in
situ burning has occurred, it is probable that the shells
were used, either incidentally or deliberately, as a
fuel.
The third possible reason for the charring of
hazelnuts is for roasting. Mason (1996) lists the
reasons why hazelnuts may have been roasted during
the Mesolithic. These include:
1. Storage. Hazelnuts can be stored for up to 6 months if
fresh, or longer if dried. It is likely that roasting would
not significantly prolong the storage period, but may
still have been seen as worthwhile.
2. Releasing oil or changing the oil structure. Hazelnuts
are high in fat and protein, although some of the fat
may not be digestible. It is thought that roasting can
reduce some of the indigestible fat content, and can also
enable oil extraction.
3. Flavour. Roasting can improve the flavour of nuts and
makes them pleasant to eat.
4. Grinding. Nuts may be easier to grind if roasted,
although if left for too long the kernel becomes a greasy
pulp (Mithen and Score 2000).
The use and exploitation of hazel (Corylus spp.)
As we have seen, the majority of Mesolithic deposits
at Howick contained charred hazelnut fragments, in
varying quantities and in varying states of preserv-
ation. In many contexts, thousands of CHF were
present. The primary question is why and how the
hazelnuts came to be charred. Following the dis-
covery of thousands of hazelnuts at Staosnaig,
Each of the above could provide a valid reason for
hazelnut roasting. Furthermore, it is possible that
roasting was done to serve more than one purpose.
The true purpose for the roasting, which has evi-
dently taken place within the hut, can only be
speculated upon. However, the vast quantities of
hazelnuts at Howick, and the array of features in
which the remains were preserved, suggest that the
11 PLANT REMAINS AND CHARRED WOOD 155
Context Feature Phase Vol^ flot
Vol*
(ml)
residue
(ml)
Composition (abundance) Contents (number)
232 Occupation
layer
1a 709,000 8050 Hazelnut frags. (ch)
Coarse sand
Gravel
Modern roots
Fine sand
Shell
Charcoal
Coal
Bone
Flint
Ball bearing
Insect fragments
Ochre?
3
3
2
2
2
1
1
1
1
1
1
1
1
Hazelnut frags. (ch)
Poaceae (w)
Polygonum persicaria
(w)
Chenopodium/Atriplex
spp (w)
Aethusa sp. (w)
29,219
1
12
6
2
270 Linear pit 1a 68,000 1215 Gravel
Hazelnut frags. (ch)
Fine sand
Modern roots
Coarse sand
Charcoal
3
3
3
2
1
1
Hazelnut frags. (ch) 3257
272 Post socket 1a 1000 20 Fine sand
Modern roots
4
2
Hazelnut frags. (ch) 20
274 Pit 1a 4000 57 Hazelnut frags. (ch)
Gravel
Fine sand
Charcoal
Coarse sand
3
3
1
1
1
Hazelnut frags. (ch) 190
291 Hearth 1a 31 Hazelnut frags. (ch)
Gravel
Fine sand
4
2
1
Hazelnut frags. (ch) 100
318 Post socket 1a 2500 2 Fine sand
Charcoal
5
1
330 Post socket 1a 1000 6 Hazelnut frags. (ch)
Gravel
Fine sand
4
2
2
Hazelnut frags. (ch) 16
343 Post socket 1a 1500 10 Hazelnut frags. (ch)
Coarse sand
Gravel
Fine sand
Insect fragments
3
3
2
1
1
Hazelnut frags. (ch) 30
345 Post socket 1a 2000 16 Gravel
Hazelnut frags. (ch)
Coarse sand
Fine sand
3
3
2
1
Hazelnut frags. (ch) 45
349 Post socket 1a 32 Hazelnut frags. (ch)
Gravel
Coarse sand
Shell
Modern roots
Fine sand
Insect fragments
Charcoal
3
2
2
1
1
1
1
1
Hazelnut frags. (ch) 120
351 Pit 1a 5 Hazelnut frags. (ch)
Gravel
Coarse sand
Modern roots
Fine sand
Insect fragments
Charcoal
3
2
2
1
1
1
1
Hazelnut frags. (ch) 25
353 Post socket 1a 12 Hazelnut frags. (ch)
Gravel
Coarse sand
Flint
Modern roots
Fine sand
Insect fragments
Charcoal
4
2
2
1
1
1
1
1
Hazelnut frags. (ch) 40
Table 11.1. Plant macrofossil data by phase. (Continued over the next six pages).
156
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Feature Phase Vol^ flot
(ml)
Vol*
residue
(ml)
Composition (abundance) Contents (number)
355 Hearth 1a 4780 Hazelnut frags. (ch)
Gravel
Flint
Shell
Bone
Charcoal
4
2
1
1
1
1
Hazelnut frags. (ch) 42750
357 Hearth 1a 75,000 590 Hazelnut frags. (ch)
Gravel
Modern roots
Fine sand
Charcoal
4
2
1
1
1
Hazelnut frags. (ch) 2900
359 Linear
feature,
prob timber
stain
1a 40,000 89 Hazelnut frags. (ch)
Gravel
Coarse sand
Flint
Shell
Fine sand
Modern roots
Charcoal
3
2
1
1
1
1
1
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
255
2
365 Pit 1a 75,000 540 Hazelnut frags. (ch)
Gravel
Modern roots
Coarse sand
Charcoal
3
2
2
1
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
2240
2
367 Hearth 1a 32 Hazelnut frags. (ch)
Gravel
Modern roots
Fine sand
Coarse sand
Insect fragments Charcoal
4
2
3
3
2
1
1
Hazelnut frags. (ch) 80
369 Post socket 1a 54 Hazelnut frags. (ch)
Coarse sand
Gravel
5
2
1
Hazelnut frags. (ch) 235
371
same as
357
Hearth 1a 290 Hazelnut frags. (ch)
Gravel
Coarse sand
Flint
Fine sand
Modern roots
Charcoal
4
2
2
1
1
1
1
Hazelnut frags. (ch)
Polygonum persicaria
(w)
1520
1
373 Post hole 1a 12 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
3
3
1
Hazelnut frags. (ch) 25
379 Hearth 1a 1060 Hazelnut frags. (ch)
Gravel
Modern roots
Fine sand
Flint
Ochre?
Shell
4
2
1
1
1
1
1
Hazelnut frags. (ch) 5390
264 Hearth 1b 273,000 3410 Fine sand
Hazelnut frags. (ch)
Coarse sand
Modern roots
Gravel
Charcoal
3
3
2
2
1
1
Hazelnut frags. (ch) 11380
268 Hearth 1b 91,000 1605 Coarse sand
Hazelnut frags. (ch)
Gravel
Fine sand
Modern roots
Bone
Flint
Charcoal
3
3
3
2
2
1
1
1
Hazelnut frags. (ch)
Poaceae (w)
Chenopodium/Atriplex
spp (w)
8340
4
1
282 Burnt fill of
linear [283]
1b 5000 159 Hazelnut frags. (ch)
Gravel
Fine sand
Table 11.1. Continued.
4
1
1
Hazelnut frags. (ch) 950
11 PLANT REMAINS AND CHARRED WOOD 157
Context Feature Phase Vol^ flot
Vol*
(ml)
residue
(ml)
Composition (abundance) Contents (number)
Charcoal 1
285 Pit 1b 36,000 950 Hazelnut frags. (ch)
Gravel
Fine sand
Coarse sand
Charcoal
4
2
2
1
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
4073
1
293 Hearth 1b 155 Hazelnut frags. (ch)
Gravel
Fine sand
Charcoal
4
2
2
1
Hazelnut frags. (ch) 850
298 Linear
feature
1b 29,000 210 Hazelnut frags. (ch)
Gravel
Modern roots
Fine sand
Flint
Charcoal
3
2
2
2
1
1
Hazelnut frags. (ch) 1165
310 Post socket 1b 2000 17 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
3
2
2
1
Hazelnut frags. (ch)
Betula pubescens (w)
101
1
316 Occupation
layer
1b 215,000 775 Gravel
Hazelnut frags. (ch)
Modern roots
Fine sand
Coarse sand
Charcoal
Flint
Shell
3
3
2
1
1
1
1
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
Betula pubescens (w)
Poaceae (w)
2065
4
1
1
320 Hearth 1b 95,000 700 Gravel
Hazelnut frags. (ch)
Modern roots
Fine sand
Charcoal
3
3
2
2
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
2750
1
324 Pit 1b 7000 27 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
3
3
2
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
70
1
332 Stake hole 1b 200 5 Hazelnut frags. (ch)
Gravel
4
2
Hazelnut frags. (ch) 14
334 Stake hole 1b 100 5 Hazelnut frags. (ch)
Gravel
Coarse sand
3
2
2
Hazelnut frags. (ch) 10
336 Stake hole 1b 100 5 Gravel
Coarse sand
Fine sand
3
2
2
Hazelnut frags. (ch) 15
338 Stake hole 1b 200 <5 Coarse sand 5
340 Hearth 1b 510,000 3095 Hazelnut frags. (ch)
Coarse sand
Modern roots
Fine sand
Coarse sand
Insect fragments
Charcoal
Gravel
3
2
2
2
1
1
1
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
Polygonum persicaria
(w)
Aethusa sp. (w)
Betula pubescens (w)
15021
6
1
1
7
342 Occupation
layer
1b 30 Hazelnut frags. (ch)
Gravel
Coarse sand
Modern roots
Fine sand
Insect fragments
Charcoal
4
2
2
2
1
1
1
Hazelnut frags. (ch) 50
107 Roasting pit
fill 109
2 46,000 445 Hazelnut frags. (ch)
Coarse sand
Gravel
Fine sand
4
2
2
2
Hazelnut frags. (ch) 2040
109 Roasting Pit 2 162,000 970 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
Table 11.1. Continued.
3
3
2
2
Hazelnut frags. (ch)
Cirsium spp (w)
3590
1
158
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Feature Phase Vol^ flot
(ml)
Vol*
residue
(ml)
Composition (abundance) Contents (number)
Flint
Shell
Charcoal
1
1
1
111 Occupation
layer
2 5000 15 Gravel
Hazelnut frags. (ch)
3
3
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp. (w)
50
1
113 Post socket 2 25,000 105 Modern roots
Fine sand
Coarse sand
3
3
2
Hazelnut frags. (ch) 231
115 Occupation
layer
2 17,500 227 Hazelnut frags. (ch)
Coarse sand
Gravel
Modern roots
Flint
Shell
Bone
Clinker
Charcoal
3
3
2
2
1
1
1
1
1
Hazelnut frags. (ch) 762
118 Occupation
layer
2 183,000 1107 Hazelnut frags. (ch)
Modern roots
Gravel
Coarse sand
Fine sand
Flint
Bone
Shell
3
3
2
2
2
1
1
1
Hazelnut frags. (ch)
Aethusa sp. (w)
2966
3
159 Occupation
layer
2 11,000 735 Gravel
Hazelnut frags. (ch)
Modern roots
Coarse sand
Fine sand
Charcoal
3
3
2
2
1
1
Hazelnut frags. (ch)
Rubus fruticosa (w)
Chenopodium/Atriplex
spp (w)
Polygonum persicaria
(w)
555
1
1
5
160 Pit 2 43,000 600 Hazelnut frags. (ch)
Modern roots
Fine sand
Coarse sand
Insect fragments
Charcoal
Gravel
3
3
3
2
1
1
1
Chenopodium/Atriplex
spp (w)
Hazelnut frags. (ch)
8
1910
161 Occupation
layer
2 9500 95 Gravel
Hazelnut frags. (ch)
Modern roots
Fine sand
Charcoal
Coarse sand
3
3
2
2
1
1
Hazelnut frags. (ch) 310
192 Pit 2 3000 45 Hazelnut frags. (ch)
Gravel
Fine sand
Modern roots
Flint
Shell
4
2
2
2
1
1
Hazelnut frags. (ch)
Poaceae (w)
200
2
206 Pit 2 50,000 527 Gravel
Hazelnut frags. (ch)
Coarse sand
Modern roots
Fine sand
Charcoal
3
3
2
2
2
1
Hazelnut frags. (ch) 1402
209 Clay lump 2 1000 7 Hazelnut frags. (ch)
Coarse sand
Modern roots
4
2
1
Hazelnut frags. (ch) 8
210 Occupation
layer
2 373,000 4620 Fine sand
Hazelnut frags. (ch)
Soil aggregates
Coarse sand
Modern roots
Insect fragments
Gravel
3
3
2
2
1
1
1
Chenopodium/Atriplex
spp. (w)
Hazelnut frags. (ch)
Aethusa sp. (w)
Poaceae (w)
1
16413
1
1
214 Pit 2 1000 15 Hazelnut frags. (ch) 4 Hazelnut frags. (ch) 40
Table 11.1. Continued.
11 PLANT REMAINS AND CHARRED WOOD 159
Context Feature Phase Vol^ flot
(ml)
Vol*
residue
(ml)
Composition (abundance) Contents (number)
Modern roots
Coarse sand
Gravel
Fine sand
2
2
1
1
224 Post socket 2 3000 30 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
3
3
2
2
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
30
1
228 250 5 Hazelnut frags. (ch)
Gravel
4
3
Hazelnut frags. (ch) 40
234 Occupation
layer
2 10,000 135 Modern roots
Hazelnut frags. (ch)
Coarse sand
Fine sand
4
3
1
1
Hazelnut frags. (ch) 475
236
same as
109
Roasting Pit 2 64,000 660 Coarse sand
Hazelnut frags. (ch)
Fine sand
Modern roots
Gravel
Bone
Charcoal
Shell
3
3
2
2
2
1
1
1
Hazelnut frags. (ch) 3595
244
same as
115/210
Occupation
layer
2 23,000 25 Fine sand
Modern roots
Charcoal
4
2
1
259
same as
111
Occupation
layer
2 2000 20 Modern roots
Hazelnut frags. (ch)
Gravel
Fine sand
Coarse sand
Charcoal
3
3
2
2
2
2
?Charred grain
Hazelnut frags. (ch)
1
45
262 Stake hole 2 7000 85 Gravel
Hazelnut frags. (ch)
Coarse sand
Fine sand
3
3
2
2
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp. (w)
Polygonum persicaria
(w)
450
1
1
262 Stake hole 2 45,000 345 Gravel
Hazelnut frags. (ch)
Coarse sand
Fine sand
Modern roots
3
3
2
2
2
Hazelnut frags. (ch) 1047
276
same as
160
Hearth 2 6000 65 Gravel
Hazelnut frags. (ch)
Coarse sand
Fine sand
3
3
1
1
Hazelnut frags. (ch) 320
35 Post Socket 3 10,000 100 Fine sand
Hazelnut frags. (ch)
Modern roots
Coarse sand
Charcoal
Gravel
3
3
3
2
2
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp. (w)
Aethusa sp. (w)
150
6
1
41 Timber
stain
3 52,000 480 Modern roots
Hazelnut frags. (ch)
Fine sand
Coarse sand
Charcoal
Flint
3
2
2
2
2
1
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp. (w)
2226
1
45 Pit 3 28,000 250 Modern roots
Coarse sand
Hazelnut frags. (ch)
Charcoal
Flint
3
2
3
2
1
Hazelnut frags. (ch) 1655
47 Hearth 3 20,000 325 Hazelnut frags. (ch)
Coarse sand
Flint
Bone
5
1
1
1
Hazelnut frags. (ch) 890
49 Occupation
layer
3 1164,500 14495 Gravel
Hazelnut frags. (ch)
Coarse sand
Table 11.1. Continued.
3
3
2
Hazelnut frags. (ch)
Chenopodium/Atriplex
spp (w)
44367
1
160
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Context Feature Phase Vol^ flot
(ml)
Vol*
residue
(ml)
Composition (abundance) Contents (number)
Modern roots
Fine sand
Flint
Shell
Bone
Ochre?
Insect fragments
2
1
1
1
1
1
1
51 Post socket 3 3000 10 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
Modern roots
Charcoal
Insect fragments
3
3
2
2
1
1
1
Hazelnut frags. (ch)
Poaceae (w)
Urtica dioica (w)
62
1
1
53 Pit 3 7500 90 Gravel
Coarse sand
Fine sand
Modern roots
Charcoal
Flint
Shell
3
3
3
1
1
1
1
Hazelnut frags. (ch) 450
65 Pit 3 56,500 480 Coarse sand
Modern roots
Gravel
Fine sand
Flint
Degraded bone
Coal
3
3
3
2
1
1
1
Hazelnut frags. (ch)
Poaceae (w)
230
1
67 Post socket 3 3000 150 Fine sand
Gravel
Coarse sand
Flint
Hazelnut frags. (ch)
Modern roots
Mollusc
2
2
2
2
1
1
1
Hazelnut frags. (ch) 10
106 Pit 3 24,000 325 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
Flint
4
3
2
1
1
Hazelnut frags. (ch) 515
148 Post socket 3 3000 25 Hazelnut frags. (ch)
Gravel
Flint
4
2
1
Hazelnut frags. (ch) 60
173 Hearth 3 25 Hazelnut frags. (ch)
Gravel
4
2
Hazelnut frags. (ch) 150
266 Post hole
fill of [267]
3 5000 105 Coarse sand
Hazelnut frags. (ch)
Modern roots
Modern wood
Fine sand
Flint
Mollusc
3
3
2
1
1
1
1
Hazelnut frags. (ch) 445
55 Burning pit 4 5000 20 Hazelnut frags. (ch)
Gravel
Coarse sand
Fine sand
Flint
Modern roots
3
3
2
1
1
1
Hazelnut frags. (ch) 170
9 Clay-lined
pit
Outside
Hut
10,000 22 Gravel
Fine sand
Coarse sand
Hazelnut frags. (ch)
4
3
2
2
Hazelnut frags. (ch) 17
13 Pit Outside
Hut
70,000 60 Fine sand
Coarse sand
Gravel
Flint
Shell
Coal
3
3
3
1
1
1
21 Pit Outside 50,000 90 Gravel 3
Table 11.1. Continued.
Context Feature Phase Vol^ flot
11 PLANT REMAINS AND CHARRED WOOD 161
Vol*
(ml)
residue
(ml)
Hut Coarse sand
Fine sand
Flint
Shell
Composition (abundance) Contents (number)
3
1
1
1
57 Clay lining
for
truncated
clay-lined
pit
Unstrat
outside
hut
5000 10 Coarse sand
Fine sand
Coal
Insect fragments
3
2
1
1
72 Linear
timber stain
Unstrat
outside
hut
55 Gravel
Hazelnut frags. (ch)
Modern roots
Flint
Shell
Table 11.1. Continued.
3
3
1
1
1
Hazelnut frags. (ch) 250
nuts may have been roasted, consumed and used as
fuel.
Mithen et al. (2001) assert that charred hazelnut
fragments that derive from roasting events are
‘failures’. Experimental roasting has determined that
such failures constitute only 12–25% of the original
total roasted (Mithen and Score 2000). If the hazelnuts
at Howick solely represent failures then 75–88% of
the hazelnuts collected have not been preserved and
the magnitude of exploitation is greater than that
implied by the data. It is acknowledged that some of
the CHF may have been roasted, then used as fuel,
and hence the numbers preserved will represent a
higher proportion of the original total. Even with this
in mind the hazelnuts at Howick still represent the
collection of perhaps many thousands of nuts each
season. Consequently, it can be concluded that
hazelnuts constituted a significant component in the
Mesolithic diet and will have formed a major source
of protein and fat with special importance as a food
source over the autumn and winter months.
It is somewhat striking that charred hazelnut shells
were virtually the only identifiable archaeobotanical
remains preserved. A single charred seed, possibly a
grain, was found but this was too degraded to
identify and therefore the provenance of the grain
cannot be assured. This absence of other remains is in
contrast with other British Mesolithic sites, where
charred plant material has been preserved. At
Staosnaig, Scotland, the large numbers of charred
hazelnut shells were complemented by finds of
charred crab apple seeds (Malus spp.), which may
have been exploited as a food resource (Mithen et al.
2001). Other charred seeds from local vegetation that
may have been inadvertently charred were also
preserved. At Star Carr, charred starchy plant roots
from emergent aquatic species such as reeds (Phrag-
mites australis) and rushes (Typha spp.) were pre-
served (Hather 1998; Mellars 1998). It is likely that
similar plant resources were available at Howick,
specifically in the nearby valley floor habitats. Could
the absence of such plant remains be the result of
taphonomic processes? It is acknowledged that
charred hazelnut shells are very hardy and may
withstand poor preservation conditions that more
fragile charred material could not. The absence of
other remains could also alternatively result from the
hut deposits accumulating indoors where only
foodstuffs brought into the house could be expected
to survive, and in the case of Howick, only if they
were charred. The presence of a wider range of
species on the other sites could therefore be assoc-
iated with them accumulating outdoors where more
material is likely to become incorporated into the
archaeological deposits.
Conclusions
Analysis of the archaeobotanical remains from
Howick has demonstrated the presence of around
200,000 charred hazelnut fragments and an unquanti-
fiable number that disintegrated on excavation in
features that date to all four phases of the Mesolithic
occupation. The quantity of hazelnuts preserved
suggests the possibility that organised exploitation of
the resource took place. As the charred fragments
uncovered at the site may represent failures, the
original number of hazelnuts collected could have
been many times the number preserved, together
with the unknown quantity that were no doubt
processed and consumed outside, and those frag-
ments that disintegrated during trowelling.
A combination of the limited quantities of charcoal
and the preservation of charred shells in areas of in
situ burning leads to the conclusion that hazelnuts
were, incidentally or purposely, used as a fuel source.
162
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
It is also probable that the hazelnuts were roasted,
although the purpose of roasting, whether for taste,
extraction of the oil or storage, remains unknown.
The continued presence of hazelnuts throughout
successive hut phases suggests that the acquisition of
large quantities of nuts was sustainable, which in
turn implies that strategic husbanding of the wood-
land resource may have taken place.
Context Feature Phase Identification
049 Occuptn. layer 3 Hazel (Corylus avellana)
049 (1960) Occuptn. layer 3 Clinker
210 (2939) Occuptn. layer 2 Too little charcoal available for identification.
232 (3700) Occuptn. layer 1a Hazel (Corylus avellana)
Table 11.2. Charcoal identification results.
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL
ANALYSIS
Ian Boomer, Robert Shiel and Tony Stevenson
Geomorphological Setting
Introduction
A survey was undertaken to produce a broad
classification of the coastline adjacent to the excav-
ation site. This included an assessment of the
geomorphology, distribution of potential archaeo-
logy-bearing deposits and identification of erosion
threats. The survey encompassed a 2.5km stretch with
a range of coastal landforms and possible erosion
threats.
In summary, the hinterland is generally low-lying
with a very gentle (about 2°) slope (Fig. 12.1). The
buffer zone between land and sea ranges from soft
dunes to weakly consolidated glacial deposits (often
expressed as eroding cliffs with significant rotational
slumping in places) and in some areas the local
country rock, Carboniferous Millstone Grit, is ex-
posed as steep cliffs. The intertidal zone is usually
represented by either sand flats or extensive intertidal
wave-cut platforms of the country rock.
This section concerns the geomorphology of the
coastline immediately to the north and south of the
Mesolithic site. A strip of coastline from Long-
houghton Steel (2km south) to Seahouses Farm (1km
north) is described in terms of the main geomorph-
Figure 12.1 View south from the site looking towards Boulmer showing three distinct exposures of wave-cut Carboniferous sediments
exposed at low tide. Note also the slump-edged cliffs and gentle seaward dip of the hinterland.
164
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
ological units, and the potential archaeological
importance highlighted.
Overview
This coastline is characterised by a repeated bay and
headland pattern, with low tide revealing sandy bays
separated by often extensive seaward exposures of
the local Carboniferous country rock (these platforms
are known locally as ‘hards’ or ‘steels’). The hinter-
land predominantly comprises agricultural land on a
relatively planar, gently seaward-dipping surface.
This is particularly well seen to the south of the
archaeological site, looking towards Boulmer (see Fig.
12.1). The major geomorphological feature, which
bisects the coastal plain, is the valley of the Howick
Burn, to the north and east of which the land surface
becomes somewhat more uneven.
The Howick Burn lies in an incised river valley (by
as much as 20 metres) and probably follows a natural
pattern of drainage on the Late-Glacial surface relief.
The burn has quite large valley dimensions for the
relatively small present-day discharge (usually 1–2
m3 s-1). This may be taken to indicate higher discharge
in the past, possibly the result of significant quantities
of local surface water, following the wasting of the
last ice masses after the last deglaciation. Coring
investigations have revealed that below the valley
floor up to 8m of intertidal and fluvially derived
sediments rest upon the bedrock (see below).
glacial or wind-blown sediments. These latter two
types are often affected by wave-driven undercutting,
causing rotational slumping and failure, the most
significant cause of land loss along this stretch of
coast. The low cliffs are usually between 4m and 10m
in height, although the tallest can reach heights of
15m, with a general increase in height northwards.
South of the Howick Burn the coast is dominated by
relatively soft sediments with the occasional appear-
ance of 1–2m of Carboniferous sediments at the cliff
base (often showing evidence of glacial and peri-
glacial disturbance such as frost shattering and
sediment mobilisation). North of the Burn, however,
the Carboniferous rocks become more significant,
creating solid vertical cliffs, particularly at Seahouses
Farm, topped by up to 5m of till.
There are no significant sea defences along this
part of the coast although the remains of some Second
World War anti-tank defences remain near point ‘E’
on the accompanying map.
Description and Classification of Shoreline
and Coastal Plain Morphology
Much of this coastline has a wide intertidal zone
(200–300m at lowest tide). Above the high-tide mark
the beach width varies from a few metres to some
tens of metres. The beach is usually backed by a low
cliff line of hard country rock or, more usually, softer

Figure 12.2. Schematic representation of the main sedimentary units
exposed on the Howick coastline (not all units necessarily present
throughout study area).
Summary of Sedimentary Units
The main sedimentary units exposed along this
coastline comprise, in decreasing age:
1. Carboniferous Rocks (Oldest)
2. Late Quaternary glacial sediments
3. Late Holocene Dunes
4. Soil profile (Youngest)
A summary of the main exposed units is shown in
Figure 12.2.
1. Carboniferous, Millstone Grit Series (approx.
320 million years old). These rocks are relatively
robust compared to the younger, overlying material
and are the most resistant to erosion seen on this
coast. They form extensive seaward platforms,
exposed by low tide, but also form cliffs up to 6m in
height along some sections (particularly at Seahouses
Farm). The rocks represent a series of cyclic deposits
of mudstones, sandstones, limestones and coals or
seat-earths. These diverse lithologies erode at differ-
ent rates and are, to some extent, responsible for the
variable shore face erosion rates along this coast. The
extensive platforms seen at low tide are important in
limiting coastal erosion as they serve to attenuate
some of the wave energy before reaching the shore/
cliff line.
2. Late Quaternary glacial deposits (age approx-
imately 18–12k cal BC). This unit comprises at least
two diamictons, the lower being predominantly fine-
grained, clay to silt grade (laminated in some
sections), but incorporating significant quantities of
larger clasts of up to cobble and boulder size. The
upper (younger) subunit is predominantly sandy and
caps some of the headland cliffs as well as comprising
much of the material excavated on the main site (for
further discussion of these deposits see Fig. 12.3).
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 165

Figure 12.3. Location map of the coastline between Seahouses Farm and Boulmer with the Howick excavation and site of the two Howick
Burn cores indicated. The letters A to T are reference points along the coast and are referred to in the text. © Crown Copyright Ordnance
Survey. An EDINA Digimap/JISC supplied service.
Both of these units appear to be of glacial or glacio-
lacustrine origin and were deposited towards the end
of the last climatic cool phase (Devensian). They result
from the deposition and possible reworking of
material from the retreating ice sheets which had
covered the region between about 22,000–16,000 cal
BC. There is evidence, during the latest period of that
cool phase, of localised periglacial and glacio-tectonic
processes (possibly resulting from the glacial re-
advance during the Younger Dryas stage, between
8500 and 9500 cal BC) which resulted in the deform-
ation of the underlying Carboniferous rocks and
intermixing of the recently deposited glacial dia-
micton and sandy, glacial outwash sediments.
There is also a suggestion, in the area immediately
inland from the Mesolithic site, that some glacial
‘drumlin-like’ features may have been deposited
along this otherwise planar coastal strip. This is
suggested by the occurrence of the steep, natural high
ground upon which the hillfort rests immediately to
the north of the Howick Burn about 500m west of the
excavation. A number of other glacial deposition
features (kames) are also known from the adjacent
coastline, for example at Embleton to the north. The
marked northward-sloping landscape to the south of
the Howick Burn suggests that the course of the river
may have been determined during the Late Glacial
period as it followed, and subsequently enlarged, a
natural feature of the landscape.
The glacial history of this part of the UK remains
poorly understood but a recent review can be found
in Teasdale and Hughes (1999).
3. Dunes (less than 500 years old). Coastal dune
systems are recorded above the high water mark in
166
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 12.4. Detail of shattered country rock immediately below
overlying diamicton.

Figure 12.5. Detail of possible palaeosol developed between the upper
diamicton and sandy sub-soil.
some of the embayments to the south of the Howick
Burn. Evidence from a number of dune systems along
the Northumberland coast (Knight et al. 1998; Orford
et al. 2000) indicates that the oldest of these date to
about the 15th century AD. The dune systems in the
Howick area are less well developed than those
investigated in the Orford study and may, therefore,
be even younger.
4. Soil profile (age uncertain but probably <1000
years). The depth of the soil profile along the coast
varies from a few centimetres to as much as two
metres. This description is based on observations
from the excavation site and from the coastal
exposures. The soils observed in the area are gen-
erally loamy and largely derived from the underlying
glacial deposits. Further details of the soils from the
excavation area are described below.

Figure 12.6. View westward at low tide across the foreshore in front
of Howick Burn.
Description of the Geomorphological Units
The coastline has been spatially divided into a
number of sections for ease of description, although
these do not comprise distinct geomorphological
units; the units are marked on Figure 12.3. Seven
sections of this coast are described below.
A–D: Rock platform to seaward, less significant
towards point D. Sandy beach backed by relatively
low cliffs (generally 3–4m high). The whole section
has a 2–3 metre thick diamicton overlying up to one
metre of Carboniferous sediments. The diamicton is
generally sub-divided into an upper sandy unit and a
lower blocky unit that at its base incorporates large
(10–40cm) clasts of disturbed country rock (sand-
stone, shale; see Fig. 12.4), associated with evidence
of periglacial activity. Localised sections of the lower
diamicton also show evidence of laminations, sug-
gesting sub-aqueous deposition. The southern part of
this section is backed by sand dunes; towards the
north, dunes are recorded on top of the cliff.
D–E: Situated in a relatively sandy embayment,
the cliffs backing this beach are relatively high, rising
to approximately 10 metres. No country rock is visible
in this section; this may explain why the coastline has
eroded more significantly. The section shows sig-
nificant slumping from the higher diamicton. The
sequence is capped by about one metre of possible
aeolian sediments resting on a palaeosol formed on
the diamicton (see Fig. 12.5).
E–G: The section comprises 10m high cliffs of the
diamicton with profiles ranging from quite steep in
fresh sections to about 45° where slumping occurs,
this is most significant in the central part of this
section. There are at least two ‘cuts’ through this
relatively soft sediment; both are probably of anthro-
pogenic origin and are currently used as pathways to
the beach. Exposures of the local country rock are
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 167

Figure 12.7. Massive sandstone country rock overlain by diamicton.
hard country rock, making much steeper cliff lines
than those to the south. This area of the coast north of
the Howick Burn is recognised by local farmers as
being particularly prone to land loss through erosion.
This section includes the coastline immediately below
the main excavation. The cliffs generally comprise a
1–2m thick diamicton on top of the Carboniferous
sediments, topped by approximately 1 metre of sandy
soil. The excavation was made in this upper horizon.
Progressing northwards, this section becomes in-
creasingly covered by heavily vegetated, slumped
sections of the overlying diamicton. It is difficult to
discern whether the base of the cliff is in situ or
slumped diamicton. The foreshore here is dominated
by a wave-cut rock platform.
O–S: Evidence of significant slumping continues
along this section and is supported by the presence of
a number of minor ‘faults’ in the diamicton along the
cliff line. The cliffs range in height from about 6–10m.
The diamicton includes many glacially striated
boulders. Towards the north of this section the
Carboniferous sediments appear once again and
become a dominant feature of the coastline. At some
sites, where the diamicton is seen directly in contact
with the underlying country rock, water is seen
emerging along a spring line. This almost certainly
acts to enhance slumping along this and probably
other parts of this coastline. Further north from points
S to T the coastline is dominated by vertical rock
cliffs of up to 10m with varying thicknesses of
overlying diamicton and soils (see Fig. 12.7), while a
wave-cut platform is present seaward.
visible at the base of the cliffs where they form the
headland at point G. The sequence at the headland is
topped by up to 2m of light buff, fine sand.
G–I: A relatively short section of coastline just to
the south of Howick Burn. Similar to the preceding
unit, this section is mainly composed of 10–12m cliffs
of diamicton overlying country rock. The vertical
extent of the Carboniferous sediments increases
northwards from approximately 1m initially until it
reaches about 5m in total. The ‘hard rock’ cliffs are
generally vertical, while the diamicton is usually
more gently sloped.
I–K: A short stretch which effectively encompasses
that part of the coast which the Howick Burn flows
across (Fig. 12.6). Gently sloping flanks of the local
landscape both to the north and the south of the
burn. Figure 12.6 shows the footbridge that spans
Howick Burn. The local country rock is exposed at
the base of the cliff on both sides of the burn. The cliff
sections are generally vertical with up to 4m of
diamicton resting upon the Carboniferous sediments.
The mouth of the burn itself is fronted by a sandy
beach with rock platforms exposed to the north and
south during low tide.
K–O: The coastline north of Howick Burn is
dominated by significant exposures (up to 7m) of the
Archaeological Potential and Erosion Threats
The rich archaeological heritage of this coastline,
particularly in the immediate vicinity of the Howick
excavation, suggests that ongoing erosion in the area
will inevitably lead to loss of archaeological remains
in the short, medium and long term. The Mesolithic
site was at extremely high risk of loss due to coastal
erosion in the short term.
This mosaic of sediment types and geomorphology
make suggesting broad-scale management options
from a heritage perspective a complex business, and
one that requires a full study in its own right.
However, future strategies might consider focussing
on surveying headlands similar to the Howick site
(although not strictly on a ‘promontory’, the site is on
a local topographic ‘high’ and is protected seaward
by a wave-cut platform). Such prominent sites would
not only have had cultural attractions as landscape
features but may also have suffered lower erosion
rates over time thanks to the ‘protective’ effects of
their associated wave-cut platforms, and may there-
fore have greater archaeological potential. The
168
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 12.8. Illustrating the modelled relationship between changing sea levels and time for three areas of the Northumberland coast during
the Holocene.
actively slumping sections of the coastline and the
eroding soft sediments by direct wave action remain
the most significant areas for the potential loss of
archaeological remains.
The bay and headland morphology of this coastline
is largely the result of differential erosion rates of the
country rock. The chief forms of erosion at Howick
are either rotational slipping of the unconsolidated
diamicton or mass failure of the Carboniferous
sediments; in both cases these result from under-
cutting of the cliffs by wave and/or storm action.
Even within the relatively robust Carboniferous rocks
the great variety of lithologies can cause preferential
erosion of one or more layers, which ultimately
undermines the entire section. Some exposures show
evidence of severe frost shattering/brittle deform-
ation and degradation near the contact with the
overlying diamicton and these sediments would be
readily dissociated by long-term wave/storm action.
In those areas where the cliffs have a good
exposure of Carboniferous rocks at their base, erosion
rates are generally less dramatic than those where the
entire cliff line is composed of softer, non-consol-
idated glacial sediments. Anecdotal evidence from
local farmers indicates farmland has been lost along
this coastline during the last century, although the
scale of such loss is difficult to assess accurately. This
is clearly an ongoing, natural process, which is
bringing about loss of archaeological features. A
Victorian carriageway that ran along the Howick
Burn and up the coast below the excavation site has
now all but disappeared.
The cliff profile along the coast varies depending
on the nature of the sediments; however, most of this
coastline has significant volumes of slumped material
(comprising the unconsolidated glacial material and
soil), with the result that much of the ‘natural’ or in
situ cliff material is obscured. It is therefore difficult
to make detailed surveys of the cliff line for evidence
of archaeological remains along much of this coast.
Late Glacial to Holocene Palaeogeography
and Sea Level Change
A number of research studies over the past 10 years
have detailed the Holocene response of this coastline
to the twin processes of Post-Glacial regional crustal
rebound and rising global sea levels. These studies
indicate that much of the southern North Sea formed
a land bridge across to North-West Europe until
about 5500 cal BC. The ‘Northumberland’ coastline
during the early Holocene, while similar to its
modern outline, would have been several hundred
metres further seaward than it is today, with the
current coastline having been reached by approx-
imately 5500 cal BC. These results are broadly
supported by evidence from the Howick Burn
sedimentary cores outlined below.
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 169
Precise reconstructions of Holocene sea level
change in Northumberland have been made possible
by the careful recording of ‘sea level index points’
based on a combination of sedimentary and micro-
fossil characteristics of sediments from along the
coast. The altitudes of these samples, above or below
ordnance datum, are measured to an accuracy of a
few millimetres while the samples are dated using
radiocarbon techniques. The results are given as an
age/altitude curve for a given location. In the
example below (Fig. 12.8, taken from Shennan et al.
2000), the reconstruction of mean tidal level is shown
for south, mid- and north Northumberland during
the Holocene. The results come from a model based
on a large number of sea level index points from the
east coast of England.
In the context of this model, Howick is close to the
central Northumberland line. Therefore, at the time
the Mesolithic site was occupied, about 8000 cal BC,
the sea level would have been approximately 5m
lower than it is at present. Sea levels would have first
reached their modern position at about 3000 cal BC,
but reached a maximum of about 1m higher than
today by about 2000 cal BC before subsequently
declining to their modern levels. The phenomenon of
gradual Post-Glacial sea level rise is recorded at many
sites throughout the UK and beyond. The increase
above modern levels during the Middle to Late
Holocene is largely a function of the crustal response
of northern Britain following the loss of the ice masses
in Scotland and Northern England. Therefore, we
might expect continual changes in coastal morph-
ology and the local environment throughout the
Holocene. In the following sections we discuss the
evidence for environmental change in the immediate
vicinity of the Mesolithic site from the sediments
within the excavation itself as well as sedimentary,
floral and faunal archives from the river valley
immediately to the south of the site.
Sedimentary Analysis of the Site
Late Glacial to Holocene deposits
The material underlying the excavated site consists
of a chaotic mixture of ‘blocks’ of stratified diamicton
in a substantial matrix of fine sands and silts
containing discontinuous silt/clay-rich planar layers
of 1–3cm. The blocks are typically about 1m across
and the original bedding has been totally dislocated,
although it remains clearly visible within blocks. The
deposit is of limited extent since across the track to
the west, at a distance of less than 20m, the reddish
purple upper till forms a uniform parent material for
the modern soil. As the land surface slopes down-
wards from the excavation site towards the tills to
the west it would appear that the processes forming
the coastal strip of mixed material are very localised.
Robson (1966) provides an illustration of the shore
at Howick, which is now not so well exposed, that
indicates that the underlying solid rocks, and perhaps
the overlying till too, are jumbled, probably as a result
of glacial action which has disturbed the, presumably
frozen, earlier glacial deposits and the underlying
weathered solid rocks. Presumably glacial outwash,
sand and clays, was then washed into the interstices
between the blocks before they fully thawed. It would
appear that the site consists of a low mound of
jumbled sandy diamicton and sands of probably
outwash origin, which provided a freely drained
habitable location convenient to the coast. This
chaotic material across the exposed excavation trench
initially proved problematic to the excavators until it
was appreciated that it was of natural origin.
It appears that in the Late Glacial phase, possibly
during one of the re-advances (such as the Younger
Dryas, or even due to final movements of the
Scandinavian ice), the frozen material of mostly local
origin with a high sand content (possibly combined
with later outwash material) was jumbled together to
form a low ridge or possibly a series of hummocks
along the coast. As this material is somewhat higher
than the surrounding surfaces, and has a relatively
high sand content, it would have been freer draining
than the undisturbed materials to the west and, in
spite of its currently exposed location, would have
formed an attractive settlement location. The lack of
an obvious local source for large amounts of sand
and its absence from nearby areas, together with the
abrupt change to finer-textured deposits inland, is
adequately explained by this origin for the deposits.
The highly mixed nature of these deposits means that
particle-size analysis would not achieve anything that
description alone failed to provide. The former extent
of this deposit seaward of the modern coast remains
unclear. If, as described above, the sea level was up to
1m above the modern level at about 2000 cal BC, then
a considerable extent of pre-Early Bronze Age coastal
deposits may have been lost. This loss results from
the foreshore being gently sloping and the early 8th
millennium cal BC sea level being c. 5m lower than at
present (see Fig. 12.8). Taking cognizance of such
landscape change during the Holocene it can be
inferred that the coastline contemporary with the
Mesolithic hut lay a hundred metres or so further out
than the present shore. However, the hut site should
still be considered a coastal settlement although it
did not occupy a cliff edge situation as it does today.
This kind of position, set back from the shore by a
short distance, is directly comparable to the East
Barns Mesolithic hut site near Dunbar where this
structure was set back from the cliff edge by several
hundred metres (see Gooder 2007). This type of
setting can be considered advantageous as it shelters
170
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
the structure from the full impact of coastal wind and
weather.
very small. Sandy clay loam was the dominant
texture but with slight variation in the coarseness of
the sand and quantity of silt. However there was not
a significant enough change to map patterns over the
field. There was no notable soil colour change over
the field, which could indicate a variation in texture
and associated change in the soil water regime. The
soil colour observed was dark brown (Munsell colour
10YR 3/3 moist).
Cracks in the topsoil were common throughout the
field, frequently 1cm width with a depth of about
10cm. This confirmed the presence of a significant
quantity of clay, which is likely to have a considerable
influence on the soil characteristics such as drainage.
The soil over the whole of the wheat field seems to fit
within that described by Payton and Palmer (1990) as
the Salwick Series. The regular cultivation of this soil
has mixed in any windblown material and obscured
any minor surface accumulation that could have
developed.
Soils
The soil was initially believed, on the limited
evidence of the 2000 excavation, to be developed in
blown sand overlying glacial till; the till was visible
in the scarp slope down to the shore and in the
surrounding fine-textured soil in the field to the east.
However when the larger area was exposed in 2002 it
became immediately apparent that the underlying
material was much more heterogeneous than had
been suspected. There were patches of reddened till
interspersed between areas of sand. As the area
exposed was enlarged it became clear that there was
a chaotic distribution of areas of sand and till; some
of the till was stratified but the blocks of till were
randomly aligned as described above. Much of this
till contained reddened material and was quite sandy.
The soils of the Howick area had been surveyed by
Payton and Palmer (1990) and those surrounding the
site were delineated as the Salwick Series. However,
due to the scale of soil mapping (1:50,000) the units
depicted are likely to include parts of other soil series.
In 2002 the examination was extended outside the
immediate vicinity of the archaeological investigation
to help interpret the site’s current and past physical
conditions (see survey area marked on Figure 14.3).
The total area covered was about 50ha.
The survey examined both Field 1 (see Fig. 1.4)
immediately east of the archaeological site and the
excavation field in which the site was located. Due to
the standing wheat crop, the auger samples were
taken within 1m of the tramlines in Field 1. There
were no such restrictions in the grass field and so
augering occurred along transects covering all the
topographic variation within the grass field.
In the Howick interim report (2000), it was
suggested that soils around the archaeological site
itself may have formed from windblown sands in
which Wilsford Series soils, described by Payton and
Palmer (1990), had formed. If this is the case then it
could be expected that soils would form a coarse –to
fine texture pattern from the coast going inland as the
finer-texture glacial till comes to predominate,
resulting in a transition to the finer-texture Salwick
Series soils mapped by Payton and Palmer. However,
no such pattern was observed in the field survey.
Essentially only two different soils were found while
texturing within the top 20cm. These were one
varying between sandy clay loam and silty clay loam,
and the other a sandy loam.
Field 1 (see Fig. 1.4)
Textural variation in the topsoil within this field
(Centre point Grid Ref. NU 256 167) appeared to be
The Excavation Field (see Fig. 1.4)
Some variation in soil hand texture was found in the
excavation field (Centre point Grid Ref. NU 260 168)
but dark brown (10YR 3/3 moist) sandy clay loam
dominated. However, along a low ridge, between
grid refs NU 2620 1705 and NU 2600 1690, the soil
was slightly coarser than in Field 1. For a few metres
around grid ref. NU 2610 1700 the texture was fine
sandy loam over sand at 20cm depth. Within this
small area, from a depth of 20cm to 1.2m (the length
of the auger) the soil was reddened sand (7.5YR). On
the surface this area was distinguished by a concen-
tration of molehills. Moles are an indicator of the
presence of earthworms, which are sensitive to soil
acidity and are generally not found in soils below
pH5, indicating a moderate pH within soils of this
area. This reddened material may be till transported
down the coast from Berwickshire by the predom-
inant glacial movement. Reddened soil predominates
further north in Northumberland and Berwickshire
and the colours of the local reddened material are
typical of the widespread Whitsome Association in
the Merse (Ragg 1960).
The texture of the rest of the field, heading south
towards the archaeological site, and towards the
wheat field, was sandy clay loam, with a possible
increase in silt and fine sand, similar to the wheat
field. The only other very slight variation was in a
small depression at about grid ref. NU 260 167, with
an increase in clay, but not significant enough to mark
a textural change.
The Salwick Series are said to have reddish brown
clay loam sub-soils (Payton and Palmer 1990), and
this was observed on the cliff edge between the
excavation trench and Sugar Sands. Darker brown –
reddened (nearer 7.5YR hue) patches of soil were
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 171

Figure 12.9. Sedimentary log and LOI (loss on ignition) record of core HEX 11007 from Howick Burn.
observed in the excavation trench. The sandy Wils-
ford soil, common further along the coast, and which
appeared to occur on the original Howick excavation
site, is in fact rare.
The Salwick Series is referred to by Payton and
Palmer (1990) as a stagnogleyic brown earth, dev-
eloped on slightly reddened upper till, and is usually
a sandy loam texture but may be sandy clay loam or
clay loam. It seemed initially that the Salwick Series
soils had been covered with variable amounts of
blown sand, which thins inland, and within which
soils of the Wilsford Series, or intergrades with
172
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 12.10. Lithological and LOI (loss on ignition) record of core HEX 11008 from Howick Burn.
Salwick Series, had developed, but latterly it has
become apparent that the soil over the settlement is
closer to a micro-scale complex of the two soils
because of the local variation in parent material.
Other factors important in soil development include
the pH of both the environment and the parent
material. Wilsford Series are described as being inland
of coastal dunes but it seems that the Howick soils are
not derived from past coastal dunes as there is no
windblown sand containing carbonate-rich shell
fragments only a few metres inland of the dig. The
elevated pH is less due to this source than to the
alkaline influence from the sea. Sea spray has a high
pH and the alkalising effect can extend some distance
inland and would help to slow down the formation of
acid soils along the coastline. Therefore these soils
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 173
may be expected to have a weak acid to alkaline pH.
Evidence for a slightly acid soil was initially thought
to come from the presence of iron movement in the
form of common reddened mottles but this is more
likely due to the admixture of parent materials, some
of which are geologically rather than edaphologically
reddened. As the soils are not strongly bleached there
is no good evidence for more intense soil acidity. There
may have been some spreading of shell sand in the
past to raise pH but this does not seem to have been
extensive on the inland clayey soils where the
advantage would have been greater.
The soil water regime of the Wilsford Series is
described by Payton and Palmer (1990) as deeply
permeable and well drained, but they may suffer
from drought in the dry summer months. Water
retention may be a problem in these soils and this
will have affected plant development. The Salwick
Series has a more problematic water regime because
of restricting downward movement of water. In a
complex of the two the good properties of both, for
once, dominate. So, the sandy area provides good
water movement and the finer textured blocks
provide water retention.
indication of the influence of changing sea levels
along this coastline. In addition to pollen, the
sediments have yielded calcareous microfossils
(ostracods and foraminifera; aquatic Crustacea and
Protista), large plant macrofossils and molluscs,
which together provide a wealth of environmental
detail as well as valuable material that allows the
sequences to be radiocarbon dated.
Summary
The sediments immediately below the excavated
Mesolithic structure must be at least 10,000 years old.
They rest upon Devensian tills, suggesting an age of
approximately 14–8000 cal BC. The occurrence of fine
laminations within these sediments strongly indicates
sub-aqueous deposition. It is suggested that these
sediments filled a minor topographic depression
during the Late Glacial as a result of outwash at some
distance from the retreating ice. It has not been
possible to determine whether these relate to the
Devensian retreat (sensu stricto) or that of a younger
stadial. However, there is clear evidence of Late
Glacial glacio-tectonic action causing different units
to be thrusted as blocks on a metre scale.
The matrix containing the Mesolithic structure
may be an aeolian reworked derivative of the
underlying sediments, indicating a change to much
drier and presumably warmer conditions. However
there is evidence that some of this sequence was
deposited under aqueous conditions.
Howick Burn Sediments
Two cores were recovered (HEX 11007 and HEX
11008, see map, Fig. 12.3) using a modified Stitz
percussion corer, which allows the sediment to be
recovered in one-metre lengths and extruded into
plastic pipes for subsequent logging and analysis
under laboratory conditions. The cores were split,
photographed and logged at the University of
Newcastle. Subsequently, key horizons were ident-
ified for palaeoenvironmental analysis. Sediment
samples of about 3–4cm3 were taken along the core
with a view to interpreting the observed changes in
lithology, sediment colour and grain size. From each
sample, separate sub-samples were made available
for pollen analysis, loss on ignition (LOI) and
microfossils (pollen and calcareous microfossils).
The two cores are composed of mainly clay – silt
grade material with occasional sandy horizons (about
1cm thick, either marine storm or high-discharge
erosional events) and more significant sand – pebble
grade material near the base. Both cores HEX11007
and HEX11008 (Figs. 12.9 and 12.10) are rich in
organic remains, which has allowed for a detailed
multi-proxy palaeoenvironmental reconstruction. A
reconstruction of the depositional environment of
these sediments is fundamental to the interpretation
of the archaeological record, particularly in the
context of sea level change, since this coastline is
known to have witnessed mean sea levels 1m higher
than today at about 2000 cal BC, and sea levels 2 to
4m below present mean sea level between 6000–7000
cal BC. According to Shennan et al.’s model (2000) the
total possible sea level range during the Mesolithic
would have been around 4.5m in the Howick area
(see Fig. 12.8). It is known from studies elsewhere in
Britain that changes in coastal environments can have
significant impacts on patterns of human settlement
and subsistence strategies (Bell 1997).
Holocene Palaeoenvironmental
Reconstruction
The discovery of significant deposits of fine-grained
Holocene sediments beneath the valley floor of the
Howick Burn provided an invaluable opportunity to
reconstruct not only the vegetation record of the
region through the Holocene but also gave an
Core HEX 11007
Core 11007 was situated on the floodplain of the
Howick Burn about 30m upstream from the foot-
bridge and about 5m north of the modern river
channel. The base of the sequence was reached at
8.15m below the floodplain surface and included
174
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Depth
(cm) Sediment description
Microfossil
description
Environmental
interpretation
Unit I
0-190
Very fine sediment, very compact almost
exclusively clay. Sandstone ‘Cobbles’ near
surface have anthropogenic origin.
No Microfossils
recorded
Floodplain. Irregularly
flooded
Unit II
190-291
No Microfossils
recorded
Floodplain. Irregularly
flooded. Subject to
extreme events
mobilising sand.
Unit III
291-502
Foraminifera rare in this
unit but Ostracoda at
their most abundant
and are dominated by
brackish water taxa
although there are also
This unit is interpreted as
a lagoonal phase, with
marine influence perhaps
restricted by a physical
barrier. The thick sand
layer suggests fluvial
Unit IV
502-667
Unit V
667-705
Predominantly laminated fine sands and silt with
significant sand layers of up to 2cm throughout
this unit.
(291-410) Mainly fine-grained sands and silts but
with distinct 45cm medium-coarse sand deposit
(at approx. 360-405cm). Two sand layers (1-
2cm) occur towards the top. Organic content
generally 20-30% throughout this whole Unit.






















































(410-502) Mainly fine-grained sediments
continued from Unit 3 although laminations
disappear at about 445cm.
Fine-grained sediments ranging from grey silts
thorough to organic rich (dark grey-black) silts
with up to 70% organic composition at some
horizons. Sediments show evidence of
lamination down to about 620cm. These may
indicate either tidal marine rhythms or episodic
fluvial events. Chaotic mix of woody debris at
570-580cm.
Dominated by coarse clastics including a basal
medium to fine sand. The sand grains are
entirely composed of quartz (i.e. no marine
carbonate) and are predominantly angular and
are considered to represent re-deposited local
glacial deposits rather than being marine
derived.
a significant number of
freshwater forms.
event.
Unit VI
705-750
Unit VII
750-802
Unit VIII
802-815
A layer of coarse sand and pebbles (including a
large number of angular fragments of local
country rock)
750-774 Medium to dark-grey silts, some dark-
brown spots, possibly plant remains.
802-774 Medium to coarse grey sand with
occasional dark-grey rounded clasts
Mid-grey to yellow sandstones, some shales,
largely degraded
Microfossils rare in this
part of the sequence,
but those present are
indicative of brackish
water and a proximity
to the Mean High Water
Springs mark.
Indicates decreasing
marine influence due
either to a seaward
‘barrier’ or a
sedimentation rate
greater than sea-level
rise.
Dominated by
Foraminifera,
particularly marginal
marine or estuarine
forms. The lowest
sample (702cm)
contained 8 specimens
comprising 4 taxa, 3 of
which are lower marsh
forms.
Marginal marine
conditions. Good
connections with open
marine water but
probably close to Mean
High Water
Not sampled Sediments deposited
after a high-energy
erosional event.
Barren Decreasing energy
suggests change from
fluvial to lagoonal/marsh
conditions
Barren Local Carboniferous
country rock
Table 12.1. Summary of the sedimentary and microfossil record from Core HEX 11007 together with an interpreted palaeoenvironment. All
depths are drill-depth below surface. Further interpretation of the sedimentary sequence based on dating evidence can be found in chapter 6.
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 175
Depth
(cm) Sediment description Microfossil description
Unit I
0-110
Very fine sediment, very
compact almost exclusively
clay. Flat shale ‘slab’ near
surface may have
anthropogenic origin.
No Microfossils recorded
Environmental
interpretation
Vegetated
floodplain.
Irregularly flooded
Unit II
110-300
Weakly laminated mid-grey to
brown silt.
No Microfossils or molluscs recorded but
high quantities of plant macrofossils.
Unit III
300-415
Medium-grey laminated silts
with sand layers up to 1cm
thick. Significant sand body at
382-415cm.
No Microfossils recorded
Floodplain-
wetland. Heavily
vegetated by
Phragmites
stands.
Flood pulse
deposits in
otherwise low-
energy, estuarine
environment.
Unit IV
415-702
Fine-grained sediments,
mainly brown-grey to medium-
grey silts. Sediments show
weak lamination. Plant
macros, throughout,
occasional molluscs visible.
This Unit contains all of the microfossil
recorded in the core. With three sub-units
415-509 Dominated by brackish
ostracods, some freshwater. No
foraminifera, uppermost 50 barren.
_ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _
509-585 No Microfossil recovered. Rich
Phragmites remains may indicate high
pH sediments.
_ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _
585-702 Lowest samples barren.
Assemblages initially dominated by
brackish foraminifera, with brackish
ostracods appearing towards top of sub-
unit.
Weakly laminated
sediments
dominated by
brackish water
organisms
indicates
estuarine/lagoonal
conditions with
decreasing marine
influence through
time, probably as
a result of
sediment
accumulation.
Unit VI
702-744
Dominated by sandstones,
sub-ordinate shales, largely
degraded None Local country rock
Table 12.2. Summary of depths are drill-depth below surface.
the sedimentary and microfossil record from Core HEX 11008 together with an interpreted palaeoenvironment. All
13cm of what were considered to be in situ Carbon-
iferous sediments (local bedrock) at the base, ensuring
that the oldest possible sediments of archaeological
interest were recovered. The succeeding 8m are
therefore considered to be of Holocene (and possibly
latest Glacial) age. A total of 33 microfossil/pollen/
LOI samples and 22 radiocarbon samples was
selected from core 11007. For details of the radio-
carbon analyses see Chapter 6.
The sediments recorded in 11007 are illustrated in
Figure 12.9 together with an indication of the
occurrence of organic remains (both woody plant
macrofossils and molluscs) and a record of the
changing LOI values through the core. A brief
sedimentary description of the core and interpreted
environment of deposition is given in Table 12.1.
Figure 12.9 also gives an indication of the calibrated
ages of the radiocarbon samples taken from the core.
The sedimentological, faunal and dating evidence
from this core allows us to build a picture of changing
environmental conditions at this site through the
Holocene. Dating indicates that although the record
is reasonably complete for the past 11,000 years, there
is a marked discontinuity of approximately 3000 years
near the base of the sequence (Fig. 6.7), which
unfortunately coincides with the dated occupation
176
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
FORAMINIFERA OSTRACODA
Depth
(cm)
Jadammina macrescens
Haynesina depressula
Haynesina germanica
Elphidium williamsoni
Ammonia beccarii
Cypridopsis vidua
Sarscypridopsis aculeata
Cyclocypris ovum
Heterocypris salina
Cyprideis torosa
Candona candida
Hemicythere villosa
Pseudocandona sp.
Leptocythere lacertosa
Elofsonia baltica
Total number of ostracods
% freshwater ostracods
% marine ostracods
% brackish water ostracods
Ostracods /g dry sediment
Total number of forams
% brackish/upper marsh forams
% marine intertidal forams
Forams /g dry sediment
43
112
195
219
258
276
291
304 100 1 85 6 9 175 9 91 113 1 100 1
351 18 38 44 88 63 38 63
405
425 3 37 61 71 63 37 23
455 16 1 33 48 3 185 52 48 179
473 3 413621 39 38 62 13
479 8 28 11 48 5 158 23 77 102
494 25 50 25 12 50 50 4
502 50 50 2 10 2
520 50 50 2 10 1
547 100 1 100 0.3
551 9 91 1 9 91 8
578 100 2 10 1
606
628 100 13 100 4
667 100 1 100 0.3
675 1 88 11 221 12 88 78
702 78 6 16 13 13 13 63 8 13 13 75 3 157 94 6 51
750
754
765
766
771
783
786
803
Table 12.3. Numerical and percentage abundance data for all ostracod and foraminifera recorded in core HEX 11007
levels in the adjacent Mesolithic structure; the occu-
pation of the Mesolithic hut dates to the middle of the
missing period. This coincides with the occurrence of
a 45cm thick deposit of sand and sandstone pebbles,
probably derived from local country rocks rather than
from modern intertidal processes.
The discontinuity is most likely the result of a
relatively short-lived erosional event (possibly high-
energy flood) within the river valley. Dating of the
sediments almost immediately overlying the erosion-
al phase suggests that this event occurred just prior
to about 6100 cal BC. Although there is no direct
evidence as to what might have caused such a
significant event there is widespread evidence of a
marked climatic deterioration at about 6300 cal BC
across northern Europe, Greenland and North Am-
erica (Klitgaard-Kristensen et al. 1998). Whether these
two events are linked cannot be conclusively proven.
Following the deposition of the coarse sand and
pebble layer, most of the rest of the section comprises
fine-grained, blue to grey sediments ranging from clay
to silt grade. Towards the top of the core, particularly
between about 3.5 to 2m core depth, the clay – silt is
developed with small layers of silty sand (ranging
from fine to medium sand). The origin of these sands
is unclear but sedimentologically they appear to be
the result of relatively common marine incursions into
the valley, although there is no supporting evidence
from the faunal records. The upper 2m of sediments
comprise clay and fine silt with sporadic remains of
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 177
FORAMINIFERA OSTRACODS
Drill
Depth
(cm)
Haynesina depressula
Haynesina germanica
Jadammina macrescens
Ammonia sp.
Miliammina fusca
Sarscypridopsis aculeata
Cyclocypris ovum
Heterocypris salina
Candona candida
Other Freswater species
Total number of ostracods
% freshwater ostracods
% brackish water ostracods
Ostracods /g dry sediment
Total Number of forams
% brackish/upper marsh forams
% marine intertidal forams
Forams /g dry sediment
286
348
416
461
470 0.5 3 74 20 1 25 75 120 210
487 5 2 85 8 10 90 73 144
509 100 1 100 0.2
536
559
585 100 1 100 0.6
594 100 71 26 3 21 74 26 89 35 100 53
622 80 20 100 5 100 2 1 80 20 12
646 88 13 13 88 21 8
672
678 96 4 157 96 4 43
681 97 1 1 78 99 1 6.7
692
699
700
Table 12.4. Numerical and percentage abundance data for all ostracod and foraminifera recorded in core HEX 11008.
Phragmites plant macrofossils, and these sediments
were almost certainly deposited in a floodplain setting
similar to that of today.
Core HEX 11008
Situated approximately 50m upstream from 11007,
this core was also taken on the floodplain immed-
iately adjacent to the present-day river channel. A
total of 19 microfossil samples was selected and
analysed from this core, from a total core recovered
of 7.44 m. The sedimentary log of the core and the
LOI record are shown in Figure 12.10.
This core yielded a more homogeneous sequence
than that of HEX 11007, with fewer distinct sediment-
ary events. The core has not been radiocarbon dated
and therefore it is not possible to establish a detailed
chronology of events. The core is not discussed
further in detail but a summary of the sedimentary
and microfossil record is shown in Table 12.2,
together with an interpretation of the depositional
environment.
Calcareous Microfossil Analysis
The record of calcareous microfossils in these cores
includes ostracods, foraminifera and micro-gastro-
pods. The palaeoenvironmental reconstruction is
based on the former two groups. The ostracods
(Crustacea; Ostracoda) are small (usually 0.5–1.5mm
long) bivalved organisms inhabiting almost every
aquatic habitat on earth although most are not able to
survive more than a few hours without water. They
are known to be particularly sensitive indicators of
salinity levels and variability and as such are useful
indicators of changing marine influence in this
marginal marine setting. The foraminifera (Protista)
are small (usually 0.3–2.0 mm) single-celled organ-
isms that secrete a calcareous shell or test (although
some taxa form a test by directly agglutinating
siliceous mineral grains to form their protective test;
these are particularly common in upper saltmarsh
environments). The foraminifera are restricted to
marine and brackish water habitats. Although they
may be of use in reconstructing past salinity changes
they are of particular interest in studies of sea level
change since certain species have become adapted to
different levels of exposure during the tidal cycle and
are therefore indicative of particular elevations within
the tidal frame. They form the main biological index
of past sea level in marginal marine settings.
Although both ostracods and foraminifera are
present in cores 11007 and 11008, they are absent
from about half of the samples studied; this may be
interpreted in one of two ways. Either they were not
originally present (the environment was inimical to
their survival, e.g. irregularly flooded floodplain) or
they were present but have not survived as fossils.
178
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 12.11. Summary of calcareous microfossil data from core HEX 11007 with suggested interpretation of depositional palaeoenvironment.
The latter case is a problem in sediments with high
concentrations of organic material (i.e. high LOI
values) where post-depositional processes result in
elevated pore water pH due to the release of signif-
icant quantities of humic acids from the organic
remains. This results in the loss of most carbonate
remains. (It should be noted that many of the upper
saltmarsh foraminifera species form their shells by
agglutinating detrital quartz grains onto their organic
matrix; these are not affected by dissolution.)
In the faunal analyses the Ostracoda have been
split into ‘freshwater’, ‘brackish’ and ‘marine’ taxa;
although these are not strict definitions, they serve to
indicate the most likely environment of deposition
for a given assemblage. The foraminifera have been
designated as ‘brackish/upper marsh taxa’ or ‘mar-
ine intertidal’ to indicate relative position within the
tidal frame.
Detailed microfossil records from 11007 and 11008
are presented in Tables 12.3 and 12.4 respectively,
while summary plots are given in Figures 12.11 and
12.12 respectively. Combined microfossil and sed-
imentological interpretations from the two cores are
summarised in Tables 12.1 and 12.2, together with an
interpretation of the changing depositional environ-
ments through time.
Pollen Analysis
A total of 26 sub-samples was taken from the fine-
grained sediments in core HEX 11007. Samples of 2–
3g were prepared for pollen using a standard
acetolysis method following sodium polytungstenate
gravity flotation and hydrofluoric acid digestion
(Moore et al. 1989). For most samples at least 400
pollen were enumerated although in a limited
number of samples where pollen concentrations were
extremely low the total fell below this level. The
pollen counts are presented in Figure 12.13 (prepared
using Tilia 2.0 and TGView 1.3.1.1; Grimm 1991).
A total of five local pollen zones is recognised in
this core, zoned by eye. Dates for the zone boundaries
were interpolated from the composite dating model
developed for the core (see Chapter 6). The zones are
numbered from the base to the top.
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 179

Figure 12.12. Summary of calcareous microfossil data from core HEX 11008 with suggested interpretation of depositional palaeoenvironment.
Interpretation of Pollen Analysis
Howick Zone 2-1: 775–610cm (c. 9800 cal BC to
5950 cal BC)
The beginning of the zone is characterised by the
latter stages of the Late Glacial period with an open
Gramineae (30%)/Cyperaceae (50%) and fern-cov-
ered (40%) landscape, together with other ruderal
taxa such as Liguliflorae and Caryophyllaceae. The
initial warming at the beginning of the Holocene is
seen with the increase in Corylus pollen to c. 45%,
although this is interrupted by an expansion of Pinus
and concomitant increases in spores. The strat-
igraphic record at this time (750–705cm) indicates a
significant erosional phase, presumably of the local
Carboniferous Fell sandstones, given the large
amount of spores which are characteristic of these
deposits. Unfortunately, this erosional phase pro-
duces a hiatus, which coincides with the period of
Mesolithic occupation of the site and hampers efforts
to produce a detailed floral environmental context
for the occupation and its impact upon the landscape.
Further increases in temperature are seen with the
increase in oak species (Quercus, c. 20%) from c.
700cm to the extent that the river valley appears to
become dominated by a hazel (60%)/oak woodland
with a small amount of elm by the end of the zone.
Small traces of ruderal taxa indicate woodland
disturbance but because of the coring site’s location
this may be linked to instability in the river channel
and creation of open habitats linked to erosion events
rather than anthropogenic activity. Water ponding in
the valley leads to the establishment of Typha latifolia
communities during the mid-part of the zone.
Howick Zone 2-2: 610–550cm (c. 5950 cal BC to
5700 cal BC)
This zone is marked by a significant increase in Alnus
pollen (85%), probably dating to c. 5500 cal BC, which
comes to dominate the valley floor, and slope
vegetation with an extensive development of local
alder carr. The decreases recorded in the other taxa
probably reflect concomitant percentage effects from
the large rise in alder. A peak in spore values at the
end of the zone indicates another, albeit small, erosion
event, matching an event in the sedimentary profile,
at the same time as the alder peak collapses.
180
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Howick Zone 2-3 550–500cm (c. 5700 cal BC to c.
5500 cal BC)
The woodland vegetation (chiefly Quercus (25%) and
Corylus (40%), together with small amounts of Ulmus)
seen at the end of How 2-1 re-asserts itself at the
beginning of this zone. However, the woodland
appears to become far more disturbed as evidenced
by increases in traces of many ruderal taxa –
Artemisia, Cirsium, Gramineae < 40µm, Plantago
lanceolata, Potentilla and Rumex. Although this may
be due to human interference in the woodland, the
geomorphological instability of the site would tend
to suggest that there may be alternative ‘natural’
explanations.

Figure 12.13. Summary pollen diagram from core HEX 11007 with local pollen zones.
Howick Zone 2-4: 500–200cm (c. 5550 cal BC to
c. 1930 cal BC)
This zone exhibits the same aspect of woodland
vegetation as the previous zone, with the addition of
significant amounts of alder, probably dominating
the valley floor and indicating relatively stable
conditions. The early part of the zone has much
evidence of disturbance from the many ruderal
indicators present – Bidens, Cirsium, Stellaria holostea,
Cruciferae – and may be linked to the first cultivation
of cereals as evidenced by the presence of some large
graminoid pollen at c. 480cm. The presence of
sporadic peaks in spores suggests occasional geo-
morphological instability in the river valley as a result
of floods and consequent bank erosion and re-
vegetation.
The woodland vegetation seems to fluctuate in
composition throughout the zone, with the most
notable feature being the loss of elm from c. 300cm
and probably dating to c. 5000 BP. Another erosional
phase is recorded in the stratigraphy from c. 300–
250cm and coincides with a temporary reduction in
alder. Increasing evidence of woodland disturbance,
probably anthropogenically determined, is recorded
in the latter third of the zone.
Howick Zone 2-5: 200–0cm (c. 1930 cal BC to
present)
This zone sees the valley side and hill top vegetation
in the vicinity cleared of its woodland cover, with much
evidence of ruderal taxa like Plantago lanceolata,
Pteridium and cereal cultivation (Gramineae > 40µm).
In addition, the large Gramineae < 40µm peak, together
with the macrofossil remains of Phragmites, indicates
the formation of reed swamp within the valley.
Comparison with other sequences
There are surprisingly few other well dated pollen
sequences from the lowlands of East Northumber-
12 GEOMORPHOLOGY AND PALEOENVIRONMENTAL ANALYSIS 181
land, unlike the better studied upland areas (e.g.
Clapperton et al. 1971; Borek 1975; Davies and Turner
1979; Rowell and Turner 1985; Tipping 1996,1998a,
1998b; Moores 1999; Simmons 1996; 2003). Of the
sequences that are available, the study by Bartley
(1966) comes from a series of undated sequences from
a low-lying site at Bradford Kaims, just to the south
of Bamburgh, and would appear to record the
infilling of a Late Glacial and Early Holocene
wetland.
The early Post-Glacial parts of the Bradford Kaims
sequences compare well with the results observed at
Howick, with the dominance of oak, elm, hazel and
eventually alder. Due to the poor resolution of these
sequences it is difficult to ascertain whether these
early Post-Glacial woodlands were subject to dis-
turbance. The only other studied sequences of
significance are those examining the buried coastal
peat bands at Druridge Bay (Innes and Frank 1988;
Farrimond and Flanaghan 1996). The peat sequences
from Druridge Bay only span the latter part of the
Holocene sequence (3430–1000 BC) seen at Howick
(approximately 225–0cm), making detailed inter-site
comparisons difficult.
Although this is a relatively low-resolution study,
it provides an important lowland/coastal palyno-
logical record for the Late Glacial and Holocene of
north-east Northumberland. The geomorphological
setting in an incised coastal river valley means that
the pollen record may well have been more strongly
influenced by local vegetation changes than would
have been the case for open, upland sites.
Latin Name Common Name
Gramineae Grasses
Carophyllaceae Carnation family
Liguliflorae Dandelion
Cyperaceae Sedges
Cruciferae Cabbage family
Pinus Pine
Quercus Oak
Alnus Alder
Corylus Hazel
Ulmus Elm
Artemisia Mugwort
Cirsium Thistle
Plantago lanceolata Narrowleaf Plantain
Potentilla Tormentil
Rumex Sorrel
Bidens Marigold
Stellaria holostea Greater Stitchwort
Pteridium Bracken
Typha latifolia Broadleaf Cattail
Phragmites Common Reed
Table 12.5 Latin pollen identifications and their common names.
Environmental Summary
The area bears evidence of significant glacial in-
fluence towards the end of the Devensian cold period.
Glacial sediments (diamictons) of three distinct types
are recorded along this coastline. Between the
disappearance of the last ice cover and the end of the
glacial period, a series of fine-grained sands was
deposited above the diamictons. These may have
initially been deposited under water, possibly assoc-
iated with glacial outwash, but there is also a
suggestion that these were in part later re-mobilised
by aeolian action. It is into these sediments that the
Mesolithic structure was built.
The Late Glacial to Holocene period is represented
by open grassland. Sedge and fern vegetation pre-
dominated, with aquatic sediments deposited in the
valley adjacent to the site. Thus the area was ice-free
by at least 10,800 cal BC, according to the age of the
oldest organic material recovered in core 11007, and
shortly after, a more typical Early Holocene vege-
tation pattern became established.
Apart from the pollen record there is little other
environmental evidence available to reconstruct
conditions in the Early Holocene. The palaeo-
environmental record in the cores is unfortunately
interrupted by the deposition of a coarse layer of
eroded country rocks. These are interpreted as the
final record of a marked erosional event, which
appears to have removed 3000 years of accumulation,
including the interval that the hut was occupied. The
dating of core HEX 11007 (Chapter 6) indicates that
the erosion event probably occurred at about 6100 cal
BC.
Subsequently the Early to Middle Holocene en-
vironment at Howick was strongly influenced by the
Post-Glacial eustatic rise in sea levels through to a
peak at about 2500 cal BC when sea levels would
have been up to 1m higher than at present. Sediment-
ation in the valley of the Howick Burn during the
Early to Middle Holocene appears to have kept pace
with sea level rise, even exceeding it, since the
environments evolve from marine through brackish
to freshwater by about 4500 cal BC. After this time,
no aquatic microfossils are recovered, suggesting that
by this time conditions in the valley floor were
independent of sea level, either through a seaward
barrier or high accretion rates.
The Late Holocene record is poorly resolved due
to the lack of suitable dating material, the intermittent
nature of sedimentation in a floodplain environment
and the occurrence of anthropogenic alteration of the
valley floor in recent centuries. This sequence remains
one of the most significant, and certainly the most
comprehensively dated palaeoenvironmental records
in Northumberland.
182
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
13 EXPERIMENTAL CONSTRUCTION
Clive Waddington
Introduction
During this study an opportunity arose to ‘recon-
struct’ the Howick hut as part of a BBC documentary
for the ‘Ancestors’ series and, two years later, to build
a second experimental hut on the actual site itself
with the aid of a grant from the Northumberland
Coast Area of Outstanding Natural Beauty. Although
the primary reason for building the first structure
was to visualise what one of the huts may have
looked like for television, it also provided the chance
to undertake an experimental approach to under-
standing how the hut was made. The second hut was
built for experimental and interpretive purposes and
is considered the more faithful of the two, particularly
as it was made having learnt lessons from the initial
construction. Relating observations of the constructed
huts to those provided by the archaeological ex-
cavations has provided important feedback for
understanding the structural form and appearance of
the hut.
The notion of archaeological ‘reconstructions’ has
recently been called into question by Stone and Planel
(1999), who instead advocate the use of the term
‘construction’, as this better describes the result of
building archaeological experimental structures. The
term ‘reconstruction’ is considered loaded as it
implies that the final outcome is the realisation of a
wholly known design whereas in reality it is, in most
cases, a best guess interpretation of the archaeological
remains. Accordingly, the use of the term ‘con-
struction’ is followed here.
Construction 1
The first construction took place during late October
and early November 2002, less than two months after
the excavations had been completed. This allowed
only a limited window of time to assess the results
from the excavation and to propose the structural
form for the construction. All the experimental work
was undertaken by the author and volunteers, who
included staff from the original excavation, students
from the University of Newcastle and members of
the public, over a two-week period. The constructed
hut was built at the Maelmin Heritage Trail in north
Northumberland, a free-access archaeological inter-
pretation site named after the (royal) early medieval
town of ‘Maelmin’, which lies immediately next to
the site (see Waddington 2004 for review and
rationale of the site). The trail is located at the south
end of Milfield village immediately off the A697 trunk
road at NT 940 336 and can be visited any time of the
day, the year round.
It was decided to attempt to construct the Phase 1a
hut as this was considered to be the hut that provided
the least ambiguous archaeological evidence as to
how it was built. Based on the excavated evidence,
the overall design of this hut was considered to
comprise the following:
• A sunken-floored circular pit measuring a maximum of
6m in diameter and averaging around 0.5m deep.
• An internal ring of vertical timber supports, each
measuring c.0.15m–0.2m diameter, set within the inner
edge of the sunken floor area.
• A series of roofing poles measuring around 0.05m
diameter, angled towards the apex of the roof and set in
the ground outside the edge of the sunken floor area.
On the basis of these design principles, evidenced by
the archaeology, the following interpretations were
made:
• It was thought that the inner ring of uprights must have
supported continuous lintels in some form, as these
would provide additional support for the roof poles
and provide stability to the ring of upright timbers.
This type of lintel structure is sometimes known as a
‘roof plate’.
• The angled roofing poles, as evidenced by the stake
holes, were thought to have been laid against the roof
plate so that they formed an apex above the centre of
the hut. This creates a conical-shaped structure when
viewed from the outside, with short vertical walls
provided on the inside by the sunken floor edge.
• As the hut had to have an entrance, a single, speculative
entrance was to be made on one side of the hut.
This overall design was considered to fit best with
the observed archaeological remains for the Phase 1a
hut. One of the issues arising from the observed
13 EXPERIMENTAL CONSTRUCTION 183
Figure 13.1 Examples of different types of roof coverings that may have been used. Illustration courtesy of David Hall.
184
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 13.2 The sunken floor area with surrounding mound after
excavation by machine. The construction of the Milfield North henge
is located in the background.
Figure 13.3 The hut after erection of the timber uprights, lintels and
roofing poles
archaeology, however, was the fact that the stake
holes around the hut edge were generally small in
diameter measuring just 3–5cm. Although these were
clearly the sharpened ends of stakes driven into the
ground at an angle between c.60° and 65°, these are
rather thin for large roofing poles. On the other hand,
as they only represent the very charred ends it is
probable that these sharpened ends continued into
more substantial timber poles, but it is only the very
tips of these stakes that have survived in the
archaeological record, giving these small diameters.
No evidence of the roofing materials was dis-
covered as a result of the excavations, leaving this
component of the structure open to greater spec-
ulation. A variety of roofing materials may be
considered including hides, thatch and turf (Fig. 13.1).
Considering the sturdy nature of the timber uprights
it was considered unlikely that the structure was
roofed with a light covering, such as hides, as this
would not have necessitated stout timber uprights
within the hut. The presence of these timbers
therefore implies that the roof was constructed of
heavier and more durable material/s. Given the
choice between turf and thatch it was decided to opt
for thatch as it was thought that this material would
prove better at keeping the rain out. Although reed
would almost certainly have been readily available to
the inhabitants of the site, from areas such as the
nearby burn, the thatch used for the construction
consisted of wild grass and straw as this could be
acquired at virtually no cost, whereas reed thatch
would have been expensive to purchase.
The materials utilised for the construction con-
sisted of fast-growing softwood timber, kindly
supplied by Forest Enterprise, together with wild
grass and straw supplied by Durham Wildlife Trust
and the local farm. In the event, the roofing poles
were too short to achieve the required 65° angle and
as a result the pitch of the roof was much shallower
than would have been the case with the original hut.
Also, the timbers used as the roofing poles were much
thicker than those suggested by the stake holes in the
archaeological excavation which are more likely to
have been about half their diameter. Obtaining such
slender poles was certainly possible, as fast-growing
hardwood timber, such as birch, can reach heights of
6–7m with trunks only around 0.1m thick. However,
as softwood of only certain sizes was available we
had to settle for thicker poles. Despite this, the overall
principle of the building design could be retained
and it is not thought that either of these compromises
made any particular difference to the functional
qualities of the structure other than to give it an apex
some 2–3m lower than the original and the inclusion
of overly thick roofing poles. The roofing poles were
secured in place by a latticework of thin coppiced
poles tied on to them with plant fibre cord (sizel). The
coppiced poles had been taken from stands of hazel
and alder that had been felled by the group in
conjunction with Durham Wildlife Trust. This lattice
provided a series of concentric rings around the roof
poles on to which the bundles of thatch could be tied.
Constructing the Hut
The sunken floor area was excavated by machine to
save time and the resulting upcast made into a
mound around its edges.
The next stage was to set up the timber uprights
inside the sunken floor area and attach the timber
lintels. This involved shaping the ends of each lintel
to assist with lashing them on to the uprights using
plant fibre cord. A total of 18 timbers was used with
nine used as uprights and nine as lintels, each
measuring around 1.5m in length. The fallen timber
[029] from the phase 3 hut had an observed length of
1.6m and therefore a similar length for the phase 1
timbers can be reasonably suggested. Once the
uprights and lintels had been assembled the timber
roofing poles were erected so that they rested on the
13 EXPERIMENTAL CONSTRUCTION 185
Figure 13.4 Lashing the coppiced poles (purlins) on to the timber
frame.
Figure 13.6 Detailed view of lashing the thatch on to the purlins
using plant fibre cord.
Figure 13.5 Adding the first layer of thatch starting from the base
of the roof.
Figure 13.7 The completed hut with entrance
ring of lintels. Based on the measurements from the
excavated hut the apex of the roof would have
measured around 7.2m high, though it would have
only stood 6.8m above ground as it was built over a
0.5m deep sunken floor. The poles supporting the
conical roof would measure around 7.6m in length.
These nine poles had had their ends charred prior to
being set up in order to preserve the wood and
prevent the ends in contact with the ground from
rotting. They were then lashed onto the lintel frame
using plant fibre chord to keep them stable. Erecting
the frame of the hut took a group of six people one
day once the timbers had been felled, cut to size and
brought to site.
At this point it was realised that the ends of the
stake holes would not have survived in the archaeo-
logical horizons if the surrounding upcast mound
had been in place so it was decided to spread the
upcast out around the hut rather than keep the
surrounding mound.
Once the roof supports had been secured in place
a frame to attach the thatch was added. This involved
cutting lengths of coppiced hazel and alder to size
and then lashing them across the roofing poles to
form ‘purlins’. A speculative doorway was added on
to the south-east side of the structure consisting of a
triangular opening that ran straight back into the cone
of the roof. This was the simplest way of attaching an
opening to the cone shape and was thought to
represent a minimalist, but adequate, entrance.
The next stage of the work involved tying the
thatch into bundles ready for tying on to the purlins.
One group of people prepared the thatch while
another tied it on. Once consecutive layers of thatch
had been added, twisted hazel spikes (spars) were
used like large staples to further secure the thatch. A
small gap was left at the apex of the roof for smoke to
escape. Approximately 1 ton of thatching material
was used to cover the roof.
The bundles of thatch were far shorter than would
have been the case if they had been made from reeds,
with each bundle covering an area of about one tenth
of a square metre. This made the covering of the roof
the most time-consuming part of the construction and
it took a team averaging six people five days to
complete.
186
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Once the hut had been constructed it took on the
appearance of a large haystack; however, inside the
hut it was found to be incredibly spacious and warm.
Lighting only a small fire in the centre of the hut kept
the interior at a very agreeable temperature despite
there being inclement weather outside, which reached
winter temperatures on occasions.
The interior of the hut was fairly well lit with an
open door and small fire going. Once the eyes were
adjusted to the slight gloom, tasks requiring detailed
work could be undertaken inside with ease. The
spatial dimensions of the hut work well, as the floor
space can be used right up to the eaves thanks to the
sunken floor edges acting as walls. Being located
partly within the ground, the hut is well insulated
and draughts from below the eaves could be easily
shut out by wedging thatch material or earth between
the eaves and the ground surface.
The hut looks something like a tepee but it is made
from timber and thatch rather than poles and skins.
Indeed this early hut could be considered to be a
more robust form of a tepee or skin tent. This
observation supports the notion that the original hut
may have resulted from the translation of tepee
architecture into a more solid form and therefore the
hut may have its origins in tepee/tent architecture.
At a height of c.6.8m above ground this hut would
have had a striking visual appearance and would
have stood out as a conspicuous human construction
amongst the rest of the surrounding landscape, even
it was located in a clearing within woodland.
The size and shape of the hut may be described as
ergonomic in the sense that the hut feels very well
suited to the human scale. It provides space to live,
work and sleep around the central focus of the hearth.
The floor space lends itself to a group of around half
a dozen adults and adolescents, and perhaps one or
two more if infants are included, with space left over
for storage, the central hearth area, and a gap in front
of the entrance. It can be aptly described as a family-
size hut and this is significant for gaining a sense of
the size of social units using these types of structures.
The hut survived two winters with no repairs
required. However, an accident during Spring 2004
led to the roof catching fire. A fire was lit by the
author to dry out the damp roof on a day with
exceptionally strong winds. Some sparks got into the
roof thatch and the wind soon brought this to a full
blaze with flames consuming almost the entire thatch
cover. Since then the roof has been repaired using
reed thatch as an alternative. Given that it was always
known that the roof apex was too low this accident
further emphasizes why the original stake holes
indicated such a steep angle of pitch. The taller apex
would not only have reduced the risk of fire but
would also have made a more satisfactory angle for
the thatch to work properly.
Construction 2
The second construction took place during March
2005 immediately next to the site of the excavated
hut. The timber was carefully selected so as to include
long, slender poles of native silver birch together with
timber uprights of the same dimensions as the timber
stains found in the hut. The wood used for the spars
included a mixture of alder, hazel and willow and the
cover, this time of turf, was taken from the ground
immediately next to the hut site. The hut was built in
the same basic way as construction 1. A pit measuring
6m diameter was excavated by hand and the surface
trowelled back in case any new archaeological
remains came to light: none did. The timber uprights
were secured in position by lashing timber lintels
between each upright using plant fibre cord. The
birch poles were then laid against the uprights and
wedged into the ground (Fig. 13.8). This produced a
much taller conical frame than construction 1 with
the apex positioned 6m off the hut floor. Digging out
the hut floor and erecting the timber frame took a
group of 6 people two days to complete.
Timber spars were lashed to the birch poles to
create a latticework for supporting and securing the
turf cover. As the conical frame was connected
together by the spars, the weight in the roofing poles
shifted so that some uprights were pressed into the
ground while others were lifted. This did not cause a
problem other than the need to wedge supports
under those timbers that had lifted so that they bore
some of the weight again. This was most simply
achieved by gathering some flat sandstone pads from
the shore below and wedging them underneath the
lifted timbers. The stone pads found in the excavated
hut would therefore seem to have been wedges rather
than pads for spreading the weight as originally
thought. The sides of the sunken floor area were
somewhat unstable at first but by placing flat timbers
Figure 13.8 Construction 2 used long slender roofing poles made
from birch providing a steep pitch around 60
13 EXPERIMENTAL CONSTRUCTION 187
Figure 13.9 The birch pole superstructure gained immense stability
once the spars were tied across them.
Figure 13.11 Despite initial concerns the turf stayed in place and
was made secure by using twists of willow to peg them over the
underlying lattice.
Figure 13.10 The hut begins to take shape as an entrance is attached
and the turfing begins.
Figure 13.12 Looking south to the completed hut, which is situated
between the site of the excavated hut and the public footpath
against the sides this stabilised the soil. Again
parallels for this can be suggested in the excavated
remains such as the timber stain [359] found near to
the edge of the Phase 1a hut. The soil that gradually
falls in from the sides, however, quickly covers up the
floor areas immediately below the eaves and blurs
the original hut edge. This largely explains the
difficulty encountered trying to determine the hut
edge during excavation.
Attaching the lattice was a time-consuming pro-
cess and took nearly three days to complete, although
at times only two people worked on this task. During
this time a simple triangular doorway slightly taller
than that on the first construction was attached (Fig.
13.10). With the help of an additional ten volunteers,
an area measuring c. 5m by 14m was deturfed over
two days and the turfs piled up next to the hut ready
for setting in place. The upcast from the sunken-
floored area was mounded up around the side of the
hut and used to help bed in the roofing poles. It also
provided an ideal base on which to lay the first row
of turfs and in so doing created a useful windbreak
below the eaves of the roof cover.
The turfs were attached by laying them on to the
latticework and wedging them onto protruding ends
of the lattice. The turfs were further secured by twists
of willow jammed through the turf to rest above the
timber supports (Fig. 13.11). Attaching the turfs and
finishing the hut off took eight people three days to
complete. This made the second construction a five-
day process involving eight people over its duration,
plus the help of ten people to deturf. This is roughly
twice as fast as the first construction and what is
more, the second experimental hut is larger, more
fire-proof and more robust than the first.
The completed hut lies immediately next to the
188
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 13.13 The construction team charring the end of one of the
timber supports before the horizontal snow set in.
excavation site (Fig. 13.12) and can be visited at any
time. It is located adjacent to the coastal path and
National Cycle Route 1. An information panel has
been installed at the site and a self-guide walk leaflet
can be obtained free of charge from most local outlets.
Further information panels explaining the archaeo-
logy of the Howick area can be found at Craster
Tourist Information Centre further along the coastal
path. The second construction was also filmed for the
BBC as part of the ‘Coast’ series, making another
useful record of the construction process.
The second hut construction is more in keeping
with the evidence recovered from the excavation.
Indeed there were no archaeological signatures
produced during the second construction that could
not be directly related to the archaeological remains
observed in the excavated hut. Nearly all the core
team who built the second construction (Fig. 13.13)
excavated on the actual site and so it was important
to note that all of those involved were convinced that
the second construction related well with the ob-
served remains, while the first construction clearly
contained several discrepancies. One of the chief
results of this second construction was production of
a structure that is thought to accurately capture the
scale of the original. The striking proportions of this
hut convey not only the durability and permanence
of the structure, but also its significance as a man-
made feature in the landscape. Whether it was located
in a clearing, open woodland or open ground, this is
a large and imposing feature that would have not
gone unnoticed by other visitors to this area.
Conclusions
The constructions have prompted much discussion
as to how the hut was built amongst a wide range of
people and this has proven enlightening in itself. One
of the most useful comments was made by Roger
Miket who observed that one method of roofing that
he had encountered in parts of Scotland uses a turf
inner and a thatch outer. This system requires the turf
to be inverted and laid on the roof. This serves to
insulate the roof, protect the interior from catching
fire and provide a surface into which the thatch can
be pegged. A thatch cover is then laid over the sods
and pegged into position with spars. Such a roof
would be both proofed and insulated and provide an
extremely durable and effective covering. However,
it would also be heavy, perhaps weighing up to two
tons or so when the thatch is sodden, and this would
require the type of heavy supports indicated by the
inner ring of posts evidenced by the excavation. On
reflection it seems, to this author at least, the most
convincing interpretation of the roof cover so far.
Partly as a result of the construction, and partly as
a result of reflections on the excavated evidence, some
concluding remarks can be made.
• Slender roofing poles 6m long set at c.60º would have
been required and this may have meant careful timber
management prior to the building of the hut.
• Likewise the poles for creating the framework on which
the roofing materials could rest (whether thatch and/or
turf) could have been obtained from coppiced wood-
land.
• The final form of the hut would have resembled, in
shape and size but not in its covering, mobile tent
structures. It raises the possibility that the design of this
hut had its origins in tepee-style architecture.
• The presence of a surrounding upcast mound is con-
sidered likely as it provides a good base for the roof
cover and deflects draughts from entering the hut below
the eaves.
• Despite some deficiencies in our knowledge of the hut
form it remains clear from the position of the stout
timber uprights that these had to be secured in some
way and this could only be achieved if they bore the
weight of a heavy roof. Evidently this was a substantial,
well built and robust structure that required a degree of
planning to construct.
• On the basis of the survival of the constructed huts it is
evident that an aerodynamic hut of this sort will stand
up to northern winters with minimal maintenance.
• The scale of the hut is ideally suited for accommodating
a family-sized group.
14 HOWICK: DISCUSSION AND INTERPRETATION 189
14 HOWICK: DISCUSSION AND INTERPRETATION
Clive Waddington, Geoff Bailey and Nicky Milner, with a
contribution by Ann Clarke to the bevelled pebble discussion
Introduction
Prior to the discovery at Howick, evidence for
Mesolithic activity in Northumberland consisted
almost entirely of stone tool artefact scatters, some of
which were discovered by fieldwalking (Weyman
1975; 1980; Davies 1983; in press; Tolan-Smith 1997a;
Waddington 1999a, 2004), others as clusters of tools
and knapping debris from old land surfaces (Buckley
1922; 1925; Raistrick 1934; Tolan-Smith 1997b; Young
2000a), and yet others from the base of rock shelter
sites with a few structural features (Burgess 1972;
Beckensall 1976; Davies pers. comm.;Waddington
1999a; 1999b). A single midden-type deposit was
partially recorded at Low Hauxley below Bronze Age
burials, situated in an eroding cliff edge at the north
end of Druridge Bay (Bonsall 1984). This is the only
archaeological site to have produced a Mesolithic
radiocarbon determination in the county prior to the
work at Howick. In Bonsall’s short note he states that
it dates to “c.5000 bc” (Bonsall, 1984, 398), but no
further information is given, suggesting this deposit
belongs to the period around 5800 cal BC. Bearing in
mind this somewhat limited understanding of Meso-
lithic settlement and chronology in the North-East
region, it is rather ironic that the Howick site now
provides the most comprehensively dated site of this
period in the British Isles. However, Howick has far
more significance than just pushing back the chron-
ology of the Mesolithic in Northumberland; it has
provided evidence for a largely intact structure that
can be confidently described as a ‘home base’ (see
Smith 1992, 28–32), as well as demonstrating the
potential for Mesolithic structural remains to survive
in unlikely locations, such as beneath a ploughed
field surface. Armed with the data from Howick we
are now in a position to be able to put forward
interpretive models of human settlement during part
of the Mesolithic in this region (see below). Moreover,
the discovery of a directly analogous, and virtually
contemporary site, further up the same stretch of
coast at East Barns near Dunbar (Gooder 2007), will
add yet more to the picture of Mesolithic settlement
during the 8th millennium cal BC in this region.
Site Taphonomy
The Mesolithic hut site at Howick, although truncated
by cliff collapse on its eastern side, is of particular
significance as the deposits within the hut had
experienced relatively little disturbance as a result of
later human activity on the site. All too often early
prehistoric lithic scatters are mixed with later lithic
material, or Mesolithic sites with structural remains
are situated below later features, as was the case at
Mount Sandel (Woodman 1985). In contrast, the
surviving deposits at Howick have only been dis-
turbed by the natural processes of pedogenesis,
bioturbation (particularly mole holes) and cliff
collapse, together with the limited impact of human
interference in the form of sheep burials along the
truncated cliff edge, and the shallow ploughing of the
topsoil that had accumulated over subsequent
millennia. This has meant that the structural features
and finds from within the hut all relate to the
Mesolithic occupation of the site and are not mixed
with later material. These fortunate circumstances of
preservation mean that the recording and interpret-
ation of the site were not clouded by the effects of
intrusive activity that could otherwise contribute to
obscuring understanding of the archaeological r-
emains. It also meant that the finds assemblages and
burnt bone could be analysed as discrete and coherent
assemblages. The key challenge faced in the field
recording of the Howick deposits was the widespread
presence of mole holes. Their occurrence across the
site necessitated careful sampling so as not to include
material from within these holes in samples used for
dating and so forth.
Horizontal ferruginous bands were recognised in
the hut deposits and at first it was thought that these
could represent occupation horizons within the hut.
However, when it was realised that these horizons
not only passed through features, but all the way out
of the hut deposits into the surrounding natural sand,
it was realised that they were in fact the result of
natural iron staining. A similar effect was also
observed around the edge of each of the hut floors.
Along the edge of the cut for the sunken floor areas
could be observed a slight ‘halo’, which appeared as
190
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
a ginger-coloured iron-stained contact lens. Such
effects have been noted at other sites in the North-
East, such as the purple staining that had formed
along the cut of the ditch terminals at the north
entrance of the Coupland Henge (Waddington in
prep.). Despite these potential corruptions to the
archaeological sequence, the Howick deposits were
remarkable for their integrity, especially given their
age, and their location in what was otherwise a fragile
and acidic environment.
Site Recording
The complex stratigraphy could only be recorded in
full by slow and methodical excavation using a single-
context recording system; the employment of a grid
system of excavation would have proven unsatis-
factory, particularly as each deposit had to be
removed in its entirety in order to comprehend the
stratigraphic order of the site. Although grid systems
still tend to be employed for the investigation and
recording of hunter-gatherer sites, usually because of
the large number of lithics associated with them, this
study has shown that full open-area investigation is a
more suitable method for teasing out the complex
relationships of a substantial built site with dis-
cernible contexts, and multiple phases of occupation.
If a test pit or a box had been cut through the hut
deposits as part of an exploratory investigation, this
intervention would have undoubtedly destroyed
relationships between features. As a result of the
independent testing of the recorded stratigraphy
provided by the dating programme it is clear that
open-area excavation can be a highly effective way of
maximising information gained from hunter-gatherer
sites. A further demonstration of this has been
provided by the East Barns site near Dunbar where
open-area investigation has also been used to success-
fully excavate and record another complex Mesolithic
hut site that has evidence for multiple structural
phases (Gooder 2007).
supports took. The Phase 1 and 2 huts appear to have
had a ring of uprights set inside the hut edge,
presumably to support some form of roof plate,
together with poles set around the outside edge of
the hut scoop, angled in towards the apex of the roof
to produce a conical shape around 60º. The roof
covering/s remain/s unknown but is/are likely to
have included one or more of the following: turf,
thatch (which could be made from a variety of
materials such as reed or grass), bark, green foliage
or skins (see Fig. 13.1). Taking into account the
substantial nature of the timbers, the result of the
experimental work and the likely materials available
at the time, it is thought most likely that the hut had
a turf roof, possibly with an outer layer of reed thatch.
Although the details of how the superstructures
fitted together and the way they were roofed remain
open to debate, the overall conical roof design with
hefty internal supports remains a reasonably secure
interpretation of the observed archaeology. This
structural form contrasts, however, with the Phase 3
design, which included what may be a centrally
located post hole, together with evidence for a series
of substantial timber supports set on the edge of the
scoop, with no evidence for angled roofing poles
surviving. Taking into account the rather squared
outline of this hut, the implication in this case is for
some kind of vertical-walled structure rather than the
large conical tent-shaped structure of previous phases.
It would seem that a hut of different constructional
form and physical appearance was built in this phase.
Why this is the case remains problematic as it could
result from any number of reasons. Possible explana-
tions include: change in use, the personal preference
of the Phase 3 residents, or the type of materials
available to hand at the time of its rebuilding. The key
point, though, that ties all three phases of hut con-
struction together is the robust nature of each building
and the use of substantial timbers to support what
must have been a permanently standing structure.
Phase Summaries (see Fig 14.1 for exploded
view of the phase plans)
The Huts
Three successive huts were constructed on the
Howick site (Fig. 14.1). The size of the huts became
slightly smaller on each occasion, with the Phase 1
hut measuring 6m across, the Phase 2 hut 5.9m across
and the Phase 3 hut 5.4m across. The Phase 1 and 2
huts approximate most closely to a circle in plan
whereas the Phase 3 hut, although sub-circular,
appears to have a slightly squared shape. Another
important design difference between the Phase 1 and
2 huts and the Phase 3 hut is the form that the roof
Phase 1a Summary
The Phase 1a hut is broadly circular in plan containing
a central setting of hearths. Structural features are
located around the inner edge of the sunken-floored
area, which averages 0.5m below the contemporary
land surface, and give the impression of a sturdy
superstructure with upright timber supports, to-
gether with slight roofing poles set at an angle to
form a conical roof. Other features include pits of
various shapes and sizes situated around the central
hearth area and towards the edge of the hut. The
14 HOWICK: DISCUSSION AND INTERPRETATION 191

Figure 14.1. Exploded view of the Howick hut phase plans positioned above each other.
function of these pits is uncertain but the position of
elongated pebble tools, microliths and blades in the
north-east area of the hut suggest pit [270] may have
been associated with a range of processing/man-
ufacturing tasks. The large number of hazelnut shells,
together with small pieces of charred bone within the
hearths, indicates the cooking of a variety of food-
stuffs in the central area.
192
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Phase 1b Summary
The Phase 1b deposits, like Phase 1a before it, consist
of a central hearth area with evidence of features
associated with other activities set around the hearth
area and close to the hut edge. There is evidence to
suggest localised rebuilding or repair of the hut on its
west side, in the form of stone pads and stake holes.
The elongated pebble tool distribution is focussed
more on the south and west areas of the hut rather
than the north-east area as in Phase 1a. This suggests
that certain activity loci may have changed position
during this phase of occupation, and this could relate
to structural modifications such as the repositioning
of a door. The large burnt spreads [340], [264] and
[320] have areas of discrete burning visible within
them, suggesting that many fires had been lit around
this central space during the occupation. In addition
to these open fires, hearth pits [293] and [268] had
been cut into the central area and these had evidently
been used on many occasions. According to the
dating model (Chapter 6), the occupation of the Phase
1a and 1b hut lasted for 40–160 years (68% probability;
use 1; Fig 6.4.)
the same basic structural principles that governed the
Phase 1 hut: the sub-circular shape of the sunken-
floored area, a diameter of 5.9m compared with 6m
for the Phase 1 hut, a central burning/hearth area that
contained cooking debris, together with the remains
of a few slight stake hole tips around the hut edge.
According to the dating model (Chapter 6), the
occupation of the Phase 2 hut lasted for 10–80 years
(68% probability; use 2; Fig 6.4).
Phase 3 Summary
The Phase 3 hut had a different structural form from
the two previous structures, consisting of very
substantial timber uprights set around the edge of
the sunken-floored area, together with some smaller
posts. The Phase 3 hut was smaller than the preceding
huts, having an internal diameter of 5.4m. Although
only some of the structural evidence is available to
understand how this hut was built it clearly made
use of large upright timbers around its edge. The use
of the central area of the hut for hearth and fireside
activities was once again in evidence together with
pits suggestive of earth ovens. Elongated pebble tools
were clustered in the north area of the structure,
suggesting, as with Phase 1a, that processing activ-
ities took place in this zone of the hut. According to
the dating model (Chapter 6), the occupation of the
Phase 3 hut lasted for 1–60 years (68% probability; use
3; Fig 6.4).
Phase 2 Summary
The Phase 2 hut differs from the Phase 1 hut on several
counts. Firstly, it does not have the same obvious
evidence for the constructional form of the hut, but
just a few small stake holes and two post sockets. As
with the Phase 1 remains, these are located around
the internal edge of the hut but they are clearly
insufficient to have supported the roofing structure.
Secondly, the Phase 2 hut provides evidence for a
hearth pit, [158], away from the centre of the hut,
which together with the roasting pit abutting the
north-west edge indicates that heating, and possibly
even burning, was taking place at the hut margins as
well as in the burning area at the centre. This must
surely have produced a fire hazard unless the outer
walls of the Phase 2 hut were set further back than the
sunken floored area, or earth walls were built ver-
tically around the edge and the timber superstructure
set into that. This latter idea gains further support
from the analogous site at East Barns near Dunbar
(Gooder 2007). Here, a dark deposit was considered
to have been the remains of a turf/earthen wall that
had collapsed into the hut. The evidence for a
dedicated hazelnut roasting area is another special
feature apparent in this phase at Howick, whereby
sand and clay had been mixed together in a way that
directly compared with the hazelnut roasting remains
identified at the contemporaneous Mesolithic settle-
ment at Duvensee in Northern Germany (Lage 2004).
The clay lumps situated close to the central hearth
area are interpreted here as raw material brought into
the hut for use as part of hazelnut roasting activities.
Overall though, the Phase 2 hut still retains many of
Phase 4 Summary
The modelling of the radiocarbon dates has shown
that around 130–280 years (68% probability; gap; Fig
6.4) after the hut was abandoned (see Chapter 6) it
was utilised again by Mesolithic people, forming the
fourth phase in the stratigraphic sequence. The
radiocarbon dates from the latest burnt features cut
into the upper hut deposits suggest that one or more
short-term re-occupations of the site took place. As
there are no structural remains that can be directly
associated with these latest features it is likely that
this final phase of use relates to (a) short stay/s and
the lighting of fires in the hollow left by the hut,
rather than lengthy occupation inside any kind of
built structure. This is confirmed by the 14C dates
(Chapter 6). As only the uppermost lenses (0.06m
and perhaps up to a maximum of 0.1m) of the
Mesolithic hut deposits appear to have been trun-
cated by natural weathering of the subsoil it is highly
unlikely that a third rebuilding of the structure took
place. It is testament to the quality of the dating
programme that this phase of activity has been
recognised, as it could not otherwise be separated
out from the recorded stratigraphy.
External Activity
Although a large excavation trench was cut, there was
very little surviving evidence for external activity
around the hut, primarily because of later disturbance
and truncation. However, the few pits that were found
on the south side of the hut indicate that this area
formed a focus for other activity. As none of these
features could be dated they could not be reliably
linked to contemporary occupation of the hut. How-
ever, those that contained Mesolithic artefacts are
considered likely to be associated. As discussed
previously (Chapter 5), it is unlikely that there were
many, if any, more huts immediately associated with
the excavated one, as the geophysical survey and
extensive excavation trench and test-pitting revealed
no further structures. The only possible area where
other huts and structures could have been located is
to the east of the hut where the cliff has eroded away.
Other than the meagre evidence for a few truncated
pits, the fieldwalking indicated that other Mesolithic
foci exist in the surrounding landscape at a distance
of a km or so from the excavated hut. Like the hut,
these lithic concentrations tend to be on land above
the freshwater stream but still close to the shore. A
pattern of dispersed family-sized settlements can be
envisaged (see also below, Settlement Organisation
and Economy).
Bevelled Pebble Tools
Bevelled pebbles, variously referred to as ‘bevel-
ended tools, ‘bevelled tools’, ‘limpet hammers’,
‘limpet scoops’ or ‘elongated pebble tools’, have been
recorded at many British Mesolithic sites, particularly
those on the western shores of Britain, in Scotland
(Mellars 1987; Clarke 1990; Barlow and Mithen 2000),
England (Berridge and Roberts 1986; Palmer 1999)
and Wales (Jacobi 1980), with examples also known
from the Isle of Man (Woodman 1987) and Ireland
(Woodman 1978; 1985). The varied terminology
illustrates the confusion surrounding the study of
these tools and this confusion has been exacerbated
further by the search for single-function explanations.
Certainly, amongst the Obanian sites of the west coast
of Scotland, pieces of split bone and antler as well as
pebbles had similar bevels formed on one end, which
indicated that these elongated bevelled tools were
most probably all used in the same way regardless of
material type (Connock et al. 1992). However, the
stone tools from Bolsay Farm, Islay, and Staosnaig,
Colonsay, though of elongated form, demonstrate a
variety of wear patterns, including chipping and
flaking damage, which was not present on the
bevelled pebbles from Howick (see Barlow and
14 HOWICK: DISCUSSION AND INTERPRETATION 193
Mithen 2000). In general, there is a lack of rigour
across the board in describing the wear patterns on
the tools as well as poor, or even no, accompanying
illustrations, and this has made it difficult to identify
and compare the traces of use-wear left on these tools
from other sites.
A look at some previously published pieces shows
that very few demonstrate the wear patterns found
on the Howick bevelled pebbles. In fact the bone and
antler tools from Druimvargie Rockshelter and
Macarthur’s Cave, Oban and Risga (Lacaille 1954, fig
102; Griffitts and Bonsall 2002, figs 1 and 2), have a
bevel that is worked evenly from left to right across
the face of the tool. Even the bevelled pebbles from
Cnoc Sligeach and Caisteal-nan-Gillean, Oronsay
(Lacaille 1954 figs 94 and 88) look different as they
are quite heavily flaked or else bevelled evenly from
left to right across the tool face (with the exception of
fig 94 top row, far right). From Kinloch, Rhum the
bevelled pebbles have rougher bevels with some
flaking and the bevels are worked evenly across the
tool face (Clarke 1990, ill 81). An experimental study
by the Southern Hebrides Mesolithic Project explored
a variety of possible functions for such tools, which
they termed ‘elongated pebble tools’, including their
use as limpet hammers and scoops, knapping ham-
merstones and hide softeners (Barlow and Mithen
2000). These experimentally produced wear patterns
replicated to a certain degree the wear traces that
were present on some archaeological examples from
Islay and Colonsay, which most likely indicates that
the Mesolithic populations, at least at these sites, used
these elongated pebble tools for a range of activities.
These examples serve to illustrate that there is
considerable variation in the wear patterns on these
tools between, and even within, stone assemblages
from a single site, and that interpretations of their use
on one site should, on the basis of wear patterns, only
be applied with caution to other assemblages.
Jacobi has made the key observation that these
bevelled pebble tools are, with just a few exceptions
such as Farnham in Surrey (Clark and Rankine 1939),
confined to coastal or near-coastal settings (Finlayson
1995, 262) and this implies that they are directly
associated with processing raw materials associated
with coastal economies (Jacobi 1980). The main
suggestions for the use of bevelled pebble tools
include shellfish hammers and processing tools
(Lacaille 1954, Griffitts and Bonsall 2002), hammer-
stones for manufacturing flaked tools, and smoothing
tools for the preparation and softening of hides
(Finlayson 1995). The experimental work undertaken
as part of the Southern Hebrides Project revealed that
the effectiveness of these tools as limpet hammers
was reliant on shape and lithology, with thick pieces
of hard rock being most effective (Barlow and Mithen
2000, 516–7). These experimental pieces produced a
number of wear patterns including longitudinal and
194
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
transverse fractures, crushing, pitting and flake
removal but it is noteworthy that this activity did not
produce a bevelled shape, nor require a bevelled
shape, to be effective. The Howick bevelled pebbles
exhibit very little or no evidence for flaking or
crushing and their use as limpet hammers can be
reasonably discounted.
The absence of shellfish from the archaeological
deposits in the Howick hut suggests it is unlikely that
these tools were used as ‘limpet scoops’ or for
otherwise processing shellfish. They are clearly too
large for limpet scoops and they do not have any
remnant polish or gloss on the surface of the tool
which is an observed consequence of scooping
limpets from shells using bone and antler tools
(Griffitts and Bonsall 2002, 211), though such traces
of use may not form on sandstone pebbles in the
same way. The presence of a number of tool blanks
inside the hut at Howick raises the question as to
why they should store these inside the hut, which
was set back from the cliffs and the shellfish beds
beyond, when they were clearly readily available
from the local beaches where the limpets could also
be found. This point is also raised by Roberts
regarding a similar distribution of bevelled pebbles
and tool blanks at a site in Cornwall (cited in Connock
et al. 1992, 33).
Consideration of the Howick bevelled pebble tools
as hammerstones for use in the flaking of tools also
fails to stand up to scrutiny as they are all made from
the local coarse-grained sandstone that fractures
easily and is much softer than the beach flint worked
at the site. The experiments undertaken by the
Southern Hebrides Mesolithic Project showed that
the elongated pebbles provided only a limited degree
of control as knapping tools and that those made
from softer rock tended to be ineffective as hammers
(Barlow and Mithen 2000). The wear patterns re-
sulting from the experimental use of elongated
pebbles as hammerstones were similar to those
identified for limpet removal, including pitting, flake
removal and fracture, although the level of crushing
was more acute and some pieces showed the bevel
development (Barlow and Mithen 2000, 517–20).
Again these wear patterns are at odds with those
identified on the Howick bevelled pebbles (see
Chapter 8). Two typical knapping hammerstones
were found in the Howick deposits (see Chapter 7)
and these were made from hard stone (quartz), were
rounded in shape, and had evidence of crushing and
pitting on the favoured striking surfaces. With the
availability of suitable hammerstones it is unlikely
that specially bevelled tools made from the softer
sandstone were also being employed for this purpose.
Hide working has been suggested by Jacobi (1980),
and more recently by Finlayson (1995), as another use
for bevelled pebbles. Finlayson’s research on bevelled
tools, including those made from bone, antler and
stone, from Obanian sites led him to suggest that
they represented a labour-intensive craft-working
industry, the presence of which could indicate a
degree of social stratification through the production
of prestige hides. Using Hayden’s ethnographic work
on the various stages of hide working, and the
accompanying illustrations of some coarse stone
spalls that were used for hide softening by British
Columbian Indians, Finlayson concluded that the
bevelled pebbles from Obanian sites may have been
formed through their use as hide rubbers/scrapers
(Hayden 1990, ill 5; Finlayson 1995, 263). However,
the illustrations are not actually of whole pebbles but
of split cobbles and spall scrapers, which would most
likely have had a more acute edge angle than the
elongated pebbles. It can also be noted that based on
the, admittedly poor, illustrations the wear on these
tools does not appear to be bevelled so much as edge-
rounded with polish. Experimental hide softening by
the Southern Hebrides Mesolithic Project also re-
sulted in this activity producing a polish as opposed
to an altered bevelled end (Barlow and Mithen 2000,
519–20).
Without recourse to a controlled experimental
study it is not possible to interpret the nature and
extent of wear patterns on the bevelled pebbles from
Howick with certainty, although it remains clear that
they were not used as hammerstones or limpet
hammers/scoops. However, consideration of the
contextual evidence from the site, together with the
result from the residue analysis, presents an argument
for considering their role in relation to skin softeners.
The Howick site has proven exceptional in prov-
iding not only one of the largest assemblages of
bevelled pebble tools and associated tool blanks, but
in this case the bulk of the tools comes from within a
discrete activity setting: a dwelling. Apart from one
bevelled pebble tool found in external pit [21], one
from clay spread [87] and an unstratified piece, all of
which were close to the hut, the other bevelled
pebbles and tool blanks were found inside the
Mesolithic hut, implying that they were used indoors
and close to a fire. The analysis of these specimens
(see above) has shown that they all appear to have
been used in a similar way, suggesting a particular
specialist task.
The bulk of the Howick bevelled pebble tools was
discarded within the confines of the hut structure,
with most of them belonging to Phases 1a and 1b (18)
where the contexts were secure. It is not possible to
reliably infer any significance from the distribution of
the 11 pieces from the upper deposits as these were
found in the occupation and levelling layers [210]
and [49], which may contain material introduced to
level the hut floor.
Consideration of the distribution of the 18 pieces
from Phase 1 is instructive as it reveals a distinct arc
of tools in the north-east sector of the hut floor in
14 HOWICK: DISCUSSION AND INTERPRETATION 195

Figure 14.2. Distribution of bevelled pebble tools throughout the different hut phases.
Phase 1a (see Fig. 14.2), suggesting that the activity
for which they were used could be undertaken inside,
where warmth was preferred. Given the size of the
arc, which has a diameter of around 2.5m, this
suggests an activity undertaken by one or possibly
two people at a time. During Phase 1b the distribution
of the bevelled pebbles is not quite so concentrated
although it is clear that there was a preference for the
south sector of the hut and the central hearth area
(see Fig 14.2). This change in the location of the
activity could be associated with structural repairs
that took place and which mark the transition from
Phase 1a to Phase 1b. Perhaps the location of the door
was changed. The group of small stake holes ob-
served in the south sector of the Phase 1b hut could
potentially be associated with the bevelled-pebble
activity in this area.
It appears that the Howick bevelled pebble tools
were associated with a specialised activity that took
place inside, in a specially dedicated part of the hut
where light and heat were important. It is unlikely that
a messy job such as the removal of fat from skins took
place inside, as this is more likely to be an outdoor
task. However, the further working and softening of
skins or furs, in preparation for use in the manufacture
of clothing and bedding, could easily be undertaken
indoors. Such a use gains further credibility when it is
considered that the bevelled pebble tool examined as
part of the residue analysis (see Chapter 9), which came
from the Phase 1a cluster, had carbon residues on it,
indicating that the residue had formed by the tool being
in contact with either fat, oil, wax or a burnt organic
residue. Such a finding is consistent with the use of
these tools for softening, or possibly scraping, skins.
196
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Observing the distribution of bevelled pebble tools
in Wales and South-West England, Jacobi has made
the point that the coastal distribution of these tools
can be correlated closely with the breeding colonies
of grey seal (Jacobi 1980, 189) and draws attention to
the high meat and fat content of these mammals, as
well as the suitability of their pelts for skin boats.
Sealskins are also very warm and make particularly
good clothing. The rock ‘steels’ that lie off the Howick
coastline are well known grey seal basking and
breeding grounds and it is surely noteworthy that in
the limited faunal assemblage recovered from
Howick, grey seal phalange bones were recovered
from hearths [355] in Phase 1a and [268] in Phase 1b.
When the above arguments are considered in com-
bination, together with the wear patterns, residue
analysis, stratigraphic distribution and proximity to
hearths that contain evidence for seal processing, a
case for the preparation of seal skin seems a reason-
able interpretation of the function of the Howick
bevelled pebble tools. If they were used for such a
purpose then it may be surmised that one reason for
the discard of these tools, once they had become
smooth and developed a bevel, was that a certain
degree of roughness was required on their surface in
order to carry out the softening task. This would
account for the accumulation of the tools inside the
hut and their discard, even though they were not
necessarily broken.
The interpretation of the Howick bevelled pebble
tools as tools for hide preparation should not,
however, be considered a certainty, as the presence of
visible striations on the bevels would suggest that
either an additional abrasive was used, such as sand,
or else the tool was used on a coarser or harder
material, perhaps with the skin in between. Although
the Howick bevelled pebble tools were not examined
microscopically there were no macroscopic traces of
a polish or gloss which could be interpreted as a by-
product of hide rubbing, though the post-depositional
conditions may not have been conducive to their
survival or indeed they may not have formed on the
surface of sandstone pebbles. Also, the smaller, more
damaged bevel present on the opposite face of these
tools requires explanation. Since these bevelled
pebbles appear to have been used in a downward
motion angled slightly away from the body against
something that would have enabled the formation of
a slightly convex face on the bevel, perhaps other
uses such as bark stripping, hollowing the inside of
wood or bone for containers, or for some use in
basketry may be other possible functions. However,
these are not tasks that can be considered exclusive to
the coastal economies with which bevelled tools are
associated.
There is a growing body of evidence for individual
Mesolithic sites having areas for specialist processing
activities. At Sand, Skye, another recently excavated
Mesolithic site (Hardy and Wickham-Jones 2003), the
coarse stone tool assemblage showed a preference for
the use of one specific tool type; in this case a facially
dished or dimpled stone, the function of which can
only be guessed at but which would appear to have
been used in a percussive manner, possibly to split
bones, crush nuts and seeds, or as a hammer used
against a narrow tool such as a chisel or borer (Clarke
forthcoming). In contrast, at Kinloch, Rhum, a large
assemblage of 59 coarse stone tools contained a
variety of types, including bevelled pebble tools and
cobbles with ground sides (a type not yet seen from
other sites) (Clarke 1990, table 16), indicating that a
variety of processing activities took place there.
Interestingly too, there was a cache of coarse stone
tools found in a pit which included six bevelled
pebble tools, four plain hammerstones and four
unused cobbles (Clarke 1990, ill 83), suggesting the
storage of specific tools for future use. With the
discovery of activity areas within the hut at Howick,
specific processes also appear to have been spatially
delimited within the structure as well as in outdoor
settings elsewhere.
Permanency
Consideration of the hut sequence is crucial for
understanding the significance of the Howick site.
The lack of any visible hiatus in the huts’ stratigraphy,
together with the precise respect shown by each hut
for the position of its predecessor, imply that there
was no period of sustained abandonment during the
occupation of the huts. Likewise the radiocarbon
results show a smooth curve in the dating sequence
from Phase 1a to Phase 3, which also suggests that
occupation was continuous and not interrupted by
phases of abandonment. In addition the charring of
the roofing pole tips demonstrates the intentionality
of preserving the timbers and thus long term use.
These are significant observations as it means that a
hut was standing on this site for a period of several
generations. Indeed, according to the chronological
modelling presented in Chapter 6, a hut stood on the
site for 100–300 years (68% probability; use hut; Fig 6.4).
Although the hut was rebuilt on two occasions, and
no doubt its appearance changed throughout this
period, the key point is that a substantial timber
building appears to have stood on this site for at least
three, and probably more, generations. In human
terms this is a permanent structure as it outlived the
life of the individuals who used it. It is plausible that
some people may have been born there, lived much
of their life there, and ultimately died there. This
sense of permanence is further attested by the
substantial size of the timber uprights used in its
construction.
14 HOWICK: DISCUSSION AND INTERPRETATION 197
By erecting a permanent structure, intended to
remain standing for generations, this is at one level a
physical expression of the social and economic
organisation of its residents, and in the case of
Howick this points towards a settlement system that
involved not only residential stability over gener-
ations but also one that did not require significant
mobility. At another level, the permanency of the hut
signifies an expression of attachment to a particular
tract of landscape by the group residing there. This
reflects not only on the social organisation of the
group and their wider pattern of settlement, but also
on the reasons for why the construction of permanent
structures was desirable at all at this time. Any
permanent structure signifies the presence of people
in that place and their attachment to it and as such
the Howick huts must have functioned, to some
extent at least, as territorial markers, although the
extent to which it was visible in a probable woodland
setting remains unknown.
The existence of built structures that can be reason-
ably considered to have been permanent has not been
previously acknowledged in the periods prior to the
Neolithic in the British Isles. Indeed much has been
made of the emergence of permanent built monu-
ments in the Neolithic, and particularly with respect
to linking this phenomenon with sedentism and the
emergence of social institutions and organisation (e.g.
Bradley 2001). The linkage between permanent
constructions and a greater degree of sedentism, or a
reduction in residential mobility, is surely of signif-
icance with respect to understanding residential
occupation at Howick.
Site Type
The Howick huts all comprise sunken-floored sub-
circular features between 5.4m and 6m in diameter.
Such huts, often with slightly sunken floors, are
common elsewhere in northern Europe during the
Mesolithic, particularly in Sweden (e.g. Larsson 1996;
Kaliff et al. 1997), Denmark (e.g. Grøn and Sørensen
1995; Sørensen 1996) and Norway (e.g. Hesjedal et al.
1996; Karsten and Knärrström 2001a). However, other
sites directly comparable to Howick have been
recorded in the British Isles at Mount Sandel
(Woodman 1985), Broom Hill (O’Malley and Jacobi
1978) and at East Barns (Gooder 2007). All the
Howick huts have a central hearth area where
burning pits and scoops were situated. Within these
burnt deposits was debris typically associated with
cooking activities including the charred bones of
mammals and bird as well as large quantities of
charred hazelnut shell. Storage of foodstuffs is
attested on the site by the presence of what can best
be described as a hazelnut roasting pit in the Phase 2
hut, and possibly a second in the Phase 3 hut.
Concentrations of lithic artefacts made from locally
available beach flint were distributed across the hut
floors and in the fill of features within each hut. The
range of stone tool types and their possible uses, as
indicated by the morphological, residue and use-wear
analysis, are consistent with their utilisation in a wide
range of tasks that on other Mesolithic sites are
typically associated with processing and butchery
activities. The location of the site in an ecotonal
setting, overlooking a small estuary, makes it ideally
placed for the exploitation of a wide range of food,
clothing and building resources throughout the year.
Bearing these observations in mind the Howick site
can best be perceived as a settlement where long-
term residence took place that may or may not have
been continuous. Consideration has been given to
whether the hut functioned in a different kind of way
from a residential site, such as a sweat lodge or
ceremonial hut and such questions were in the minds
of the excavators from the outset. However, the lack
of any evidence for ritual deposits or disposal, the
inability of the structure to have performed a function
as a sweat lodge and the lack of any human burials
leave no direct evidence to support such ideas.
Considering the site from an ideological per-
spective it is of course possible that the location, form
and activities that took place at the hut included
ideological, religious and symbolic components,
indeed it would be surprising if they did not. Careful
excavation and recording, though, failed to produce
any evidence for specially placed deposits (as
occurred at Culverwell for example, see Palmer 1999,
139–40), stone tools made from exotic materials,
unusual made artefacts or structural details incon-
sistent with domestic occupation, or any evidence for
human bones or burial. The presence of small
quantities of ochreous material can be interpreted
variously, but is most likely to have been used either
as a source of mastic in boat construction, as pigment,
as part of the tanning process, or perhaps for its
medicinal qualities, or perhaps all of these uses.
Likewise the evidence for copper exploitation, as
indicated by the residue analysis and small amount
of copper ore, is perhaps also best considered as a
way of acquiring a pigment source. Such pigment
could have had a multiplicity of uses ranging from
body painting and boat caulking to the dyeing of
skins and ritual activities. However, such uses,
whether functional, symbolic, ritual or otherwise,
remain entirely speculative. The lack of any clear
evidence for symbolic or ritualised behaviour does
not mean that the Howick hut did not possess such
qualities or that associated rites and activities did not
take place there. Rather, such elements of the human
occupation of the site remain steadfastly elusive even
though concerted efforts were made throughout the
excavation to identify any signatures that could imply
198
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
such behaviour. Taking such considerations into
account the physical evidence remains consistent
with that of domestic occupation. It is therefore fair
to conclude that the Howick site is primarily a
settlement and that although activities and structural
features may have occurred with symbolic, religious
and/or ideological concerns in mind, such practices
can at this stage only be surmised and not demon-
strated.
Settlement Organisation and Economy
One of the key issues that remain is the question of
whether occupation at Howick was continuous,
seasonal or perhaps periodic. Although the structure
appears to have been permanent this does not mean
that permanent and continuous residence necessarily
took place. Moreover, as we are dealing with such a
sustained period of use, over perhaps a couple of
centuries, the nature of residency at the site may have
changed through time with duration of occupation
fluctuating according to social, economic and environ-
mental considerations. According to the current data
available, three broad scenarios can be advanced.
Scenario 1 would view the site as a home base with
occupation taking place on a seasonal basis for only
part of the year (Fig. 14.3). This view conforms to the
traditional understanding of Mesolithic settlement
patterns in the British Isles, which are thought to have
been structured around the seasonal availability of
resources across different tracts of landscape. Scenario
2 views the site as being permanently occupied on a
continuous basis throughout the year in what amounts
to fully sedentary occupation with only occasional
overnight forays outside the home base ambit by one
or two group members (Fig. 14.3). Scenario 3 views the
site as being permanently occupied by some members
of the group while other group members moved to
temporary camps to exploit seasonally available
resources in the surrounding landscape (Fig. 14.3).
Variations of all these scenarios are of course possible,
such as continuous occupation for several years
followed by abandonment and then re-occupation, but
at a general level these three scenarios represent the
main ways in which settlement organisation can be
interpreted at the site.
Given the wealth of resources available at the
coast, members of the group would not have had to
travel far to acquire food and other resources. This is
borne out by the fact that shellfish, marine fish, seals,
freshwater fish, nesting birds and their eggs would
have all been available within a few hundred metres
of the site. Furthermore, all the flint utilised on the
site was beach flint and this could also be collected
within a short distance. It is likely that the gathered
hazelnuts did not come from far away either. This
kind of economic strategy whereby most of the group
is involved in hunting, gathering and fishing within
the area around the camp is usually termed ‘resi-
dential foraging’ (see Smith 1992, 17), and it is this
strategy that probably best describes, in broad terms
at least, the economic organisation of the Howick
occupants. This is not to say that some members of
the group did not move further afield from time to
time, but that the general pattern was for most of the
group to be involved with the exploitation of
resources close to the main camp. The kind of more
distant trips that may have been undertaken include
the hunting of deer, pig and other mammals in the
woodland of the interior as well as following salmon
runs up the rivers and streams, taking migrating
wildfowl along the inland river banks and making
contact with, and presumably exchanging goods and
news with, other groups who occupied the interior
and the higher ground inland.
Currently, there is no compelling evidence either
way to determine which of the scenarios represents
the most accurate interpretation of settlement organ-
isation at the site and any of the three could be true.
What can be concluded is that the site is likely to
have formed the main dwelling site utilised in an
annual pattern of settlement, and that a considerable
portion of yearly residency took place there. If
residency was seasonal then the storage of hazelnuts
implies occupation during autumn/winter and
possibly into the spring.
The lack of any hiatus in the stratigraphic sequence
and the absolute dating programme imply that
episodes of prolonged abandonment did not occur at
the site. Therefore, it is most likely that we are dealing
with a structure that was utilised for much of each
year. The resources available in the area of the
Howick estuary make provision for a wide range of
foodstuffs that may have allowed for year-round
occupation in the same place. There is also evidence
for food storage, as represented by the charred
hazelnut shells mentioned above. The estuarine
location of the Howick site is a remarkably rich and
diverse setting, providing access to the full range of
resources available at this time. Figure 14.4 shows the
distribution throughout the year of selected resources
on this stretch of the coast based on modern patterns
of availability.
The types of resources available can be divided
into different categories for convenience. Non-food
resources such as freshwater, flint, timber, reed and
clay were all available locally as well as beach-rolled
pebbles that could be utilised as pebble tools. Ochre
may also have been obtained from the till deposits.
The foodstuffs can be broken down into the following
categories: fish, shellfish, mammals, birds and plant
foods. A wide range of fish has always been available
in the rich fishing grounds of the inshore waters of
the North Sea. These include many varieties of sea
14 HOWICK: DISCUSSION AND INTERPRETATION 199
Figure 14.3. Schematic models of early 8th millennium cal BC settlement at Howick.
200
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK

Figure 14.4. Distribution of resources throughout the year based on modern patterns of availability.
14 HOWICK: DISCUSSION AND INTERPRETATION 201
fish ranging from herring and white fish to more
exotic varieties. Freshwater fish would also be
important, and relatively easy to take, in this
estuarine location. The type of fish available would
include not just the typical freshwater varieties such
as pike, perch and dab but also anadromous fish such
as salmon and sea trout. Shellfish remain common in
these parts and the rock steels give rise to large beds
of molluscs and bivalves including limpets, whelks,
winkles, scallops, mussels, lobster and crab. The
Northumberland coast hosts large colonies of birds
as well as the large flocks of migrating birds and
fowl. It is no surprise, therefore, that the remains of a
large bird, possibly a gull, were identified in one of
the hut deposits. The species present on this stretch
of coast include the usual range of coastal birds,
together with colonies of puffins, eider ducks, geese
and swans. Bird eggs may have also featured in the
Mesolithic diet and these would have been available
in abundance along this stretch of coast, particularly
during the nesting seasons. The proximity of wood-
land would have attracted a different range of
woodland species and these may also have featured
in the Mesolithic diet. The type of mammals that
would have been available include sea mammals such
as grey seal, the presence of which is documented in
the Howick hearths, and which can still be found
basking on the rock steels today. However, the
terrestrial fauna may also have featured as a signif-
icant resource for the Mesolithic inhabitants, who
could not only consume the meat but also use the
horn, bone, skins and sinew of these beasts. Although
wild pig, fox and dog/wolf were evidenced in the
Mesolithic hearth deposits, the other types of fauna
that are likely to have been exploited include red and
roe deer, wild cattle, brown bear and possibly beaver.
Other smaller mammals such as pine marten may
also have been taken from time to time. Plant foods,
which have been argued as playing a more important
role in Mesolithic diets in recent years (e.g. Clarke
1976; Zvelebil 1994), certainly did so in the lives of
the Howick residents as attested by the huge quantity
of charred hazelnut shell fragments present. How-
ever, the role other foods such as fruits, berries, fungi,
seaweed and green leafy plants played in the diet of
the Howick residents remains unknown. All of these
foods would have been easily available, albeit on a
strictly seasonal basis, but it is worth mentioning that
recent carbon and nitrogen isotope studies indicate
that some Mesolithic people appear to have con-
sumed very heavily fish or meat-dominated diets
(e.g. Schulting and Richards 2000; 2002, but see
Milner et al. 2004 for a critical assessment of this
method).
It is clear that an ecotonal setting was the favoured
location for home base sites during the Mesolithic in
North-East England. The Mesolithic flint scatter sites
identified in the Milfield basin, and more recently in
the Lower Tweed valley, are also positioned on the
cusp between diverse areas of the landscape where
seasonally available resources from the different
ecological niches would have been in easy reach from
the centrally located sites (Waddington 1999a, 171–
2). However, Howick is not the only site situated in
an estuarine setting. The prolific flint scatter site at
Crimdon Dene (Raistrick and Westoll 1933; Raistrick
et al. 1935) on the County Durham coast was also
situated above a small estuary and other Mesolithic
flint scatters have been found clustered around the
shallow waters of Budle Bay to the north (Buckley
1922; 1925). Further afield the site at Mount Sandel
may also be considered estuarine, being in easy reach
of the coast by way of the River Bann (Woodman
1985, 159).
The Howick hut can be described as a family-sized
dwelling that would have accommodated a group of
around 6–8 individuals, presumably belonging to the
same family group. Indeed the size of the site is
directly comparable to Mount Sandel, Broomhill and
East Barns, which all have a floor area of about 30
square metres. At these sites only a single structure
exists, or only a single structure was thought to be a
dwelling, and this also appears to have been the case
at Howick. Noting the pattern for the occurrence of
single huts, Tolan-Smith has noted that “most of the
time people seem to have lived in small, isolated
groups” (Smith 1992, 172) and this seems to be the
case at Howick. However, given the results of the
fieldwalking, which suggest that other settlement foci
may exist nearby in locations proximal to the
freshwater stream, it is thought likely that a cluster of
family groups may have lived in isolated huts around
the estuary of the Howick Burn, but spaced several
hundred metres or perhaps a few kilometres apart.
This pattern of dispersed individual settlements,
forming clusters around resource-rich locales, implies
the use of specific tracts of land, and sea, by family
groups related to each other through kinship ties.
Such groups are likely to have had wider kinship
connections and cultural affinities with neighbouring
groups who may have, together, formed a larger
band. It would therefore be expected that aggre-
gations of the larger groups did occur from time to
time to reaffirm social and cultural bonds, as well as
to exchange news, goods and obtain marriage
partners.
The Mesolithic Environment
The sea level model developed by Shennan et al.
(2000) suggests that the sea was around 4.5m lower
than today in the Howick area during the period
when the hut was occupied and this would have
placed the hut in a more sheltered setting than it is
202
MESOLITHIC SETTLEMENT IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
today, set back perhaps a few hundred metres from
the shore. As the section of the sediment sequence
contemporary with the Mesolithic occupation at
Howick appears to have been eroded away by an
event that took place close to 6100 cal BC (see Chapter
12), little can be said with certainty regarding the
contemporary environment around the Howick
settlement. However, it is likely that a mixed tree
cover had developed by this time, as the pollen
diagram shows that species such as hazel, pine,
juniper, willow and even small amounts of oak and
elm were present prior to the occupation of the hut.
Smaller plants present prior to the Mesolithic occu-
pation included grasses, sedges, ferns and dandelion
amongst others (see Fig. 12.13). The presence of hazel
and birch together with a range of smaller plants
including grasses, flowering plants, goosefoot,
stinging nettles, thistle and weeds was identified by
the macrofossil analysis. The weeds and disturbed-
ground indicators show that the area around the hut
is likely to have been cleared back and the prolific
presence of charred hazelnut shells shows that the
woodland under-storey must have contained a large
number of hazel stands. Indeed the volume of nuts is
such that it suggests deliberate propagation and
management of hazel stands for the production of
food, although coppicing for timber poles may have
also been important. Taken together it can be ex-
pected that the Howick settlement occupied a
clearing in an area of relatively open woodland that
included a diverse range of species even by this early
period. Hazel appears to have been dominant and
widely available but other timbers, including some
broad-leaf deciduous trees, were also present by this
time. Summer temperatures had risen to the same as
today’s, or slightly higher, by the time the first hut
was built and the upward trend in temperatures
continued throughout the entire period of occupation
and beyond.
In terms of landscape location the Howick settle-
ment would have occupied a sheltered location on a
kind of peninsula formed by the sea to one side and
the deeply incised valley of the Howick Burn on the
other. It would have overlooked an estuary and
would have been set back from the shoreline by a few
hundred metres. Situated on sandy deposits the site
would have been dry, with relatively mild weather
resulting from the ameliorating effects of the sea. The
wind was probably the key cause for concern, much
as it is today. Inland the gently sloping land of the
coastal plain is likely to have been wooded, with
changes in tree species no doubt taking place when
the coastal plain met the more sharply rising ground
of the Fellsandstone dip slopes that rise up to the
crag lines at 150–200m AOD. In many ways the
inhabitants at Howick occupied a very attractive
landscape; they would have normally had access to
plentiful and tasty food, the climate was good and
getting better, the scenery, as today, would have been
breathtaking and perhaps more so given the prox-
imity of rich woodlands. It is therefore fitting that
today this landscape falls within an area designated
as an Area of Outstanding Natural Beauty and a Site
of Special Scientific Interest.
Conclusions
The longevity and structural form of the settlement
at Howick, together with the evidence for the
exploitation of a wide range of resources and diverse
activities, suggest a level of complexity rarely
evidenced on Mesolithic sites in the British Isles.
Furthermore this site is early, belonging to the
centuries around 7800 cal BC, making it almost 10,000
years old. This pushes back the emergence of complex
hunter-gatherer groups who built permanent struc-
tures within Britain to an earlier stage than has
previously been thought. Such complexity is usually
thought to have emerged in the latter stages of the
Mesolithic when it is traditionally thought that
settlement became less mobile in the lead up to the
introduction/uptake of farming around 4000 cal BC.
Such gradualist notions of increasing sedentism and
socio-economic complexity fail to account for the ebb
and flow of hunter-gatherer organisation and adapt-
ation in different landscape settings, in different
climates, with different social and religious traject-
ories, and at different times. It is possible that in the
centuries around 8000 cal BC, groups living along the
North-East coast of Britain, as evidenced by the sites
at Howick and East Barns, occupied coastal areas
near to estuaries with only limited mobility into the
uplands, whereas other groups occupying the interior
proper may have adopted a more mobile settlement
pattern and economic strategy, which involved
ranging over wider tracts of land. Acknowledging
the potential for diversity in the settlement, economic
and social organisation of Mesolithic groups across
the British Isles at any one time is key to advancing
our understanding of this period. The British Isles is
a hugely diverse and ecologically rich area with great
variations in geology, topography, vegetation and
climate and it must surely follow that the settlement
patterns, economic strategies and social values of
hunter-gather groups occupying this landmass must
have also shown considerable variation. In order to
gain more detailed insights into the kind of lifeways
hunter-gatherers pursued it is appropriate that we
should seek to recognise patterns of hunter-gatherer
behaviour at the regional scale.
15 HOWICK IN ITS NORTH SEA CONTEXT 203
15 HOWICK IN ITS NORTH SEA CONTEXT
Clive Waddington, Geoff Bailey, Alex Bayliss and Nicky Milner
Setting the Scene
In recent years there has been a rapid growth in the
quantity of known Mesolithic settlements around the
North Sea Basin with the excavation of hut sites in
Britain (Woodman 1985; Green 1996; Gooder in
press), Norway (Hesjedal et al. 1996; Karsten and
Knärrström 2001), Sweden (Larsson 1996; Kaliff et al.
1997; Carlsson et al. 1999), Denmark (Grøn and
Sørensen 1995; Sørensen 1996) and the Low Countries
(Crombé 1998), as well as the documentation of
underwater sites in the shallow seas off the Danish
coast (see Fischer 1995). Building on the earlier work
of Newell (1981), and more recently Wickham-Jones
(2004), there is a need for a North Sea Basin-wide
review of the structural evidence for Mesolithic
settlement sites that considers issues such as vari-
ations in size, layout, organisation, landscape setting
and duration of occupation. There is not space in this
volume to tackle this huge topic but it is hoped that
the information presented here will go some way to
informing that future debate.
It was somewhat serendipitous that the East Barns
site should have been found and excavated within a
month of completing the Howick excavations. Full
dialogue between the excavators has ensured that the
sites were compared and the significance of each
recognised. With two sites of virtually identical size,
basic structural form, date, and position on the same
stretch of the North-East British coast, there can be no
question of these sites simply being explained as
anomalies. They are the two best-preserved Mesolithic
dwellings so far discovered on the British mainland,
and as the significance of their remains filters through,
these sites will reshape our understanding of Meso-
lithic settlement in Britain. One of the sites was
excavated as part of a conservation-recording project
(Howick), while the other was excavated as part of a
developer-led project (East Barns). Both sites have
produced high-calibre output, demonstrating that
both conservation-led and developer-led archaeology
in the UK can produce important results for hunter-
gatherer archaeology.
One of the key features of the 8th millennium cal
BC hut sites of the British Isles is the striking
similarities between all the known sites (Fig. 15.1).
The hut sites which date from this period include
Howick, East Barns (Gooder in press), Mount Sandel
(Woodman 1985) and Broom Hill (O’Malley and
Jacobi 1978). Although a later (c. 6500 cal BC) and
considerably smaller site (around 4m diameter with
a corresponding floor space of 13 square metres), the
hut eroding from the cliff edge at Cass-Ny-Hawin on
the Isle of Man (Woodman 1984) has a similar, though
less robust, structural form and layout to these sites.
As can be seen in Figure 15.1, with the exception of
Cass-Ny-Hawin, all these sites are of similar size
measuring between 5.5m and 6m in diameter. They
all have a floor area in the region of 30 square metres
as well as a sunken floor with a central hearth area.
Furthermore they all have evidence for internal
settings of posts located close to the hut edge as well
as clusters of stake holes. Although the Broom Hill
site is located inland it is within a relatively short
distance of the coast, while all the other sites can be
considered either coastal or estuarine.
The North Sea Basin in the Early 8th
Millennium cal BC
Although Ireland had become separated from the
rest of Britain by c. 9000 cal BC as a result of rising sea
level, Britain remained connected to the European
mainland by way of a land bridge, connecting the
area from Flamborough Head southwards with the
German Bight and the Low Countries. This North
Sea Plain was probably extensively settled by human
groups who would have had rich pickings in the
wetlands, mudflats and sandbanks of this low-lying
expanse. However, during the time of the Howick
settlement, the sea level rose rapidly (see Smith 1992;
Coles 1998; Shennan et al. 2000), resulting in large
tracts of land being inundated, and this is likely to
have led to settlement contraction, and ultimately
population dispersals, from this once vast North Sea
Plain.
During the period the Howick site was occupied,
Britain was not an island but a western and northern
204
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.1. Mesolithic hut sites recorded in the British Isles. All date from the 8th millennium cal BC with the exception of Cass-Ny-Hawin
on the Isle of Man which is later and dates to c. 6500 cal BC.
15 HOWICK IN ITS NORTH SEA CONTEXT 205
extension of the European landmass. It formed the
western side of what was a North Sea Basin, open
only to the north. Looking out on to this basin from
Howick, the inhabitants of this site must have had a
very different sense of geography than later gener-
ations, who may have become aware that they had
become cut off from the rest of Europe when the land
bridge was breached. Today in Britain the sense of
being an island nation is so deeply rooted in our
psyche that it is difficult to conceptualise being
attached to Europe. For the residents at Howick, sea
level change was happening around the North Sea
Basin at a very rapid rate that would probably have
been observable within an individual’s lifetime. This
knowledge of a rising sea and loss of land, par-
ticularly in the areas to the south of Howick, may
have been a widely known and perhaps feared
process. Not only would this have led to visible
changes in the coastal landscape and river valleys of
the surrounding lands, but it would also have caused
populations occupying the lowest-lying areas to be
displaced. One of the authors (CW) contends that a
consequence of this may have been the exertion of
pressure on surrounding areas as groups competed
for space and resources, or else looked to colonise
under-populated or empty areas. With movement
across the North Sea Basin, by way of coast hugging
and/or island hopping in skin boats, the Mesolithic
people inhabiting this coastline may have been able
to travel relatively rapidly across long distances. This
could have led to swift information exchange and
rapid movements of people. Rather than being seen
as a barrier, the North Sea may have provided the
means by which groups were able to interact, share
ideas, build reciprocal relationships, exchange goods,
resolve disputes and maintain allegiances on a
regular basis. If people around this vast basin looked
towards the North Sea as the hub around which their
lives were structured, then the cultural and geo-
graphic milieu (as implied by the narrow-blade
micro-triangle techno-complex and its distribution
around this area) in which the Howick people
engaged was one that linked them with groups in
modern-day Scandinavia and the Low Countries and
areas that now lie beneath the North Sea. Rather than
being considered an extension of the Mesolithic
settlement of southern England and part of that
broader cultural and economic milieu, the northern
British, and possibly the early northern Irish sites,
can be better considered as part of a different cultural
and economic sphere centred on the North Sea Basin.
Mesolithic Settlement
Understanding the Howick settlement in relation to
contemporary settlements in the British Isles suggests
it is not as anomalous as perhaps first thought,
particularly in the light of new sites reported in the
companion volume to this monograph (Waddington
and Pedersen 2007). It forms one of a number of
known hut structures in the British Isles that date to
the first half of the 8th millennium cal BC (see Figs.
15.3–15.20b and Table 15.1), all of which are within
short distances of their respective coasts, and all for
family-sized social units. Within the more specific
context of North-East Britain, other dated sites of
similar age to Howick are known at Filpoke Beacon
in County Durham, East Barns near Dunbar, Cram-
ond near Edinburgh and at Fife Ness in Fife (Fig.
15.2). Excepting East Barns, these sites are smaller
than the Howick site, consisting of a few pits, stake
holes and hearths suggestive of small temporary
camps. Although it belongs to the same narrow-blade
industry as the Howick and East Barns material, the
flint assemblage from the Fife Ness site appears to be
highly specialised, with the emphasis on crescentic
microliths, which implies that this was a specialist
extraction/hunting camp of some sort (Wickham-
Jones and Dalland 1998). Other sites with similar
lithic assemblages to the Howick and East Barns
settlements, though not yet radiocarbon dated,
include the scatter of material from an eroding land
surface at Ness End Quarry on Lindisfarne (Young
2000a), as well as the sites known around Newbiggin
(Raistrick 1933), Hart (Weyman 1984) and Crimdon
Dene (Raistrick et al. 1935). Therefore, the 8th
millennium cal BC settlements of the North-East
British coast appear to have been home to a variety of
different sites including home-bases in the more
resource-rich settings such as estuaries, and specialist
small-scale temporary camps in other areas where a
particular resource may have been available at certain
times of the year.
Inland there are few dated Mesolithic sites that
can be compared with the dated coastal settlements.
However, flint scatters containing directly analogous
narrow-blade forms have been found along the major
river valleys of the Tweed (Mulholland 1970), Till
(Waddington 1999a) Tyne (Tolan-Smith 1997) and
Wear (Young 1987), suggesting the spread of narrow-
blade traditions inland from the coast. Rock shelter
sites are known at various points on the sandstone
escarpments of Northumberland (Burgess 1972;
Beckensall 1976; Waddington 1999a; John Davies
pers. comm.) and these may have formed logistical
short-stay camps that worked in association with
home base sites located on river valley gravel
terraces, such as those in the valleys of the Till and
the Tweed (Mulholland 1970; Waddington 1999a), or
those in similar locations to that at Howick such as
Newbiggin (Raistrick 1934), Lyne Hill (Raistrick 1933;
1934) and Crimdon Dene (Raistrick and Westoll 1933;
Raistrick et al. 1935).
Although definable classes of 8th-millennium
206
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.2. Radiocarbon dated 8th millennium cal BC sites and shoreline of the British Isles. The palaeoshorelines are derived from Shennan
et al. 2000.
settlement types are now emerging in the North-East
region (such as home bases, small specialist coastal
camps, and perhaps some of the undated rock shelter
sites), this patchwork of sites suggests that settlement
systems operating over this region included a range
of strategies that may have differed between different
sub-regions. Although potential home base sites have
been identified inland on the raised gravel terraces of
the rivers Till and Tweed, they are currently thought
to have functioned within settlement systems that
involved logistical foraging in the uplands (e.g.
Waddington 1999a). This would contrast with the
settlement pattern suggested for the coast where the
annual range of movement appears to have been
rather restricted and set within a more tightly defined
area. Acknowledging the landscape-specific dimen-
sion of Mesolithic settlement systems in North-East
Britain allows for a more nuanced understanding of
both individual sites and groups of sites within the
widely varying topographies of this region. This
stands in contrast to the use of generic models that
conflate settlement patterns throughout the Meso-
lithic into one model and which also fail to account
for the variation and subtleties that can be seen in
both the archaeological record and the configuration
of what is a highly varied landscape.
15 HOWICK IN ITS NORTH SEA CONTEXT Until recently it has been widely understood that
the human colonisation of northern Britain first
occurred on the west side of the country. Smith
accurately summarised the evidence available in the
early 1990s in the following passage from his syn-
thetic work on the British Mesolithic (Smith 1992,
172):
“The main development during this period [8999bp–
8000bp = 8200 cal BC–7000 cal BC] is the spread of
settlement up the west coast as far as Kinloch, on the
island of Rhum and to Mount Sandel and Lough Boora
in Ireland. No such movement is recorded in eastern
Britain, the most northerly site dating from before
8000bp [7000 cal BC] being Filpoke Beacon in County
Durham. The explosive settlement of north-western
Britain and Ireland merits special attention.”
The figures in square brackets have been added so
that it is clear which periods the author is referring
to.
The explanation given by Smith at the time for this
‘explosive’ settlement expansion up the west coast
was considered to be the increasing marine biomass
made available by the warmer temperatures of the
western seas as a result of the North Atlantic Drift.
This in turn would have attracted human groups
further northwards as they pursued maritime-
oriented economic strategies. This explanation fitted
well with the available data at the time, but given the
new dates that are now available for sites on the
North-East coast of Britain this pattern has now been
entirely reversed.
Based on the currently available radiocarbon
determinations, the earliest dates for northern Britain
are now on east coast sites (see below and Figs. 15.3–
15.20b). The sites at Cramond, East Barns and
Howick are all earlier than any of the North-West
coast sites by several centuries, while other sites such
as Fife Ness and Filpoke Beacon have similar dates to
the North-West sites. The earliest dated site is that at
Cramond, where a series of reliable radiocarbon
determinations has been obtained (Ashmore 2004;
Saville 2004). Modelling of these dates shows this site
is most likely to have been occupied in the centuries
before 8000 cal BC (see below, Fig. 15.12). It now
appears that the initial colonisation, or infilling at
least, may have taken place on the east side of
northern Britain in the coastlands bordering the
North Sea Basin. However, penetration inland along
river valleys to upland locations is perhaps suggested
at this time by the single early date from the Daer
Reservoir site in the Lowther Hills above Nithsdale
near Dumfries (Ashmore 2004; Saville 2004; Figs. 15.2
and 15.4), although this is just a single date on
Pomoideae charcoal from a pit fill associated with a
flint scatter site that also has a later date implying
that this early date could be from residual material.
207
Why Mesolithic Hut Building in the 8th
Millennium cal BC?
It is perhaps no coincidence that permanent hut sites
were constructed around the British Isles in the first
half of the 8th millennium cal BC during precisely the
same period that the North Sea Plain was being
rapidly inundated by rising sea levels. In the view of
one of the authors (CW), the displacement of human
groups by the rising sea is thought to have prompted
movement away from the drowned areas of the
North Sea Plain to higher land, which in turn could
have led to population pressure on neighbouring
areas. One response to such pressure may have been
to build permanent huts, possibly as territorial
markers, in order to protect access to the rich and
already inhabited lands around the North Sea margin.
Once Britain was finally isolated from the continent,
the pressure and stresses resulting from the presence
of incoming groups would have lessened and this
could account for the absence of huts of Howick, East
Barns, Mount Sandel or Broom Hill proportions after
this time. It remains interesting to see whether future
discoveries will fit into this pattern or whether they
will show permanent huts of these proportions being
built over a much longer span. Presently, however,
these huts remain a tightly dated phenomenon and
as such may reflect a particular response to a specific
set of historical, social, economic and environmental
circumstances. Once competition for resources had
settled down the need for permanent structures that
signalled access to territory and resources would
have diminished. Although this explanation for
Mesolithic hut building in the 8th millennium cal BC
may be somewhat simplistic and environmentally
linked, it nevertheless acknowledges the correlation
between the timing of the known hut sites in the
British Isles and the single most significant change to
affect the British landscape since the end of the last
Ice Age, namely sea level rise and the separation of
Britain from the European landmass.
Re-calibration of other 8th
Millennium cal BC sites in the
British Isles
By Alex Bayliss and Clive Waddington
In order to place the Howick site in its broader
temporal context, the radiocarbon determinations
from potentially contemporaneous British archaeo-
logical sites have been assembled (Table 15.1). These
results have been re-calibrated as described above
for the Howick site (see Chapter 6), allowing for
Site Material and context Laboratory
Number
13C (‰) Radiocarbon Age
(BP)
Weighted mean (BP) Calibrated date range (95%
confidence)
Reference
Auchareoch hazelnut shell from pit in quarry face OxA-1601 -25.0 (assumed) 8060±90 - cal BC 7320 – 6680 Affleck et al.1988
Hedges et al. 1989
Aveline’s Hole human bone, disarticulated right distal
humerus
OxA-1070 -21.0 (assumed) 8740±100 - cal BC 8210 – 7580 Hedges et al. 1987
Jacobi 1987
Aveline’s Hole human bone, disarticulated right distal
humerus
OxA-800 -21.0 (assumed) 8860±100 - cal BC 8270 – 7600 Gowlett et al. 1986
Jacobi 1987
charred wood fragments from the
archaeological layer
Q-800 -25.0 (assumed) 8573±110 - cal BC 7940 – 7370 Radley et al. 1974
Switsur et al. 1975
Broom Hill charcoal from the base of pit 3 Q-1192 -25.0 (assumed) 8540±150 8459±87
cal BC 7610 – 7210
T’=1.3; v=2; T’(5%)=6.0
Broom Hill replicate of Q-1192 Q-1528 -25.0 (assumed) 8515±150
cal BC 8160 – 7190
cal BC 7600 – 7040
cal BC 7960 – 7180 O’Malley and Jacobi
1978
O’Malley and Jacobi
1978
O’Malley and Jacobi
1978
Broom Hill charcoal from clay in the top of pit 3 Q-1460 -25.0 (assumed) 7750±120 - cal BC 7040 – 6390 O’Malley and Jacobi
1978
Broom Hill hazelnuts from 5cm above pit 3 infill Q-1191 -25.0 (assumed) 7220±120 - cal BC 6380 – 5810 O’Malley and Jacobi
1978
Broxbourne Site 105 Broomhead Moor
Site 5
Broom Hill replicate of Q-1192 Q-1383 -25.0 (assumed) 8315±150 Birm-343 -25.0 (assumed) 8700±170 - cal BC 8270 – 7380 Shotton and Williams
1973
wood (? Pinus sp.) from below the
Mesolithic horizon
Castleroe bulk charcoal from occupation layer UB-2171 - 8755±135 - cal BC 8270 – 7540 Woodman 1978a: 13-4
Castleroe bulk charcoal from Pit 1 UB-2172 - 8560±75 - cal BC 7750 – 7480 Woodman 1978a: 13-4
Cramond hazelnut from pit 1430 OxA-10178 -23.3 9105±65 - cal BC 8530 – 8230 Lawson 2001
Cramond hazelnut from pit 1425 OxA-10144 -23.1 9110±60 - cal BC 8520 – 8230 Lawson 2001
Cramond hazelnut from pit 1430 OxA-10179 -24.0 9130±65 - cal BC 8530 – 8230 Lawson 2001
Cramond hazelnut from spread 1409, sealing pit
1425 and scoop 1432
OxA-10143 -23.5 9150±45 - cal BC 8530 – 8260 Lawson 2001
Cramond hazelnut from spread 1409 OxA-10145 -24.9 9230±50 - cal BC 8610 – 8280 Lawson 2001
Cramond hazelnut from scoop 1432, cut into pit
1430
OxA-10180 -26.0 9250±60 - cal BC 8690 – 8280 Lawson 2001
Daer Reservoir 2 birch charcoal from a pit in forager
knapping site
AA-30355 -25.1 8055±75 - cal BC 7310 – 6690 Ward 1998
Daer Reservoir 2 Pomoideae charcoal from a pit in forager
flint knapper site
AA-30354 -26.7 9075±80 - cal BC 8520 – 7970 Ward 1998
Druimvargie bone artefact from midden in rock shelter
NMSX.HL 416
OxA-4608 -22.1 8340±80 - cal BC 7580 – 7080 Bonsall et al. 1995
Hedges et al. 1998
East Barns hazelnut shell from post hole sealed
within the hut
AA-54961 -24.0 8830±70 - cal BC 8240 – 7650 Gooder in press
East Barns hazelnut shell from post hole sealed
within the hut
AA-54962 -24.3 8835±65 - cal BC 8240 – 7650 Gooder in press
East Barns hazelnut shell from post hole sealed
within the hut
AA-54960 -23.0 8985±70 - cal BC 8290 – 7960 Gooder in press
Fife Ness hazelnut shell from fill of pit F 84 AA-25202 -26.4 8275±65 - cal BC 7530 – 7080 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 84 AA-25203 -24.5 8340±60 - cal BC 7550 – 7180 Wickham-Jones and
Table 15.1. Radiocarbon determinations from comparable dated sites in the British Isles (c. cal BC 8300–7000 [c. 9200–8000 BP]). (Continued over the next three pages).
208
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Site Material and context Laboratory
Number
13C (‰) Radiocarbon Age
(BP)
Weighted mean (BP) Calibrated date range (95%
confidence)
Reference
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 70 AA-25204 -23.5 8505±75 - cal BC 7650 – 7370 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 70 AA-25205 -24.9 8405±60 - cal BC 7580 – 7200 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 63 AA-25206 -23.6 8355±60 - cal BC 7580 – 7180 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 63 AA-25207 -24.2 8420±65 - cal BC 7590 – 7210 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 61 AA-25208 -23.6 8510±70 - cal BC 7610 – 7380 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from fill of pit F 61 AA-25209 -26.8 8475±75 - cal BC 7600 – 7350 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from occupation layer F46 AA-25210 -21.8 8410±60 - cal BC 7580 – 7210 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from occupation layer F46 AA-25211 -25.7 8460±85 - cal BC 7610 – 7320 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from the lower fill of pit
F41
AA-25212 -22.9 8545±65 - cal BC 7680 – 7480 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from the lower fill of pit
F41
AA-25213 -25.2 8495±65 - cal BC 7610 – 7380 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from the upper fill of pit
F41
AA-25214 -23.2 8510±65 - cal BC 7610 – 7380 Wickham-Jones and
Dalland 1998
Fife Ness hazelnut shell from the upper fill of pit
F41
AA-25215 -24.7 8490±60 - cal BC 7600 – 7380 Wickham-Jones and
Dalland 1998
Filpoke Beacon bulk hazelnut shells from burning spread
with associated flints
Q-1474 -25.0 (assumed) 8760±140 - cal BC 8270 – 7540 Jacobi 1976: 71
Fordhouse barrow hazelnut shell from pit 21 OxA-10059 -23.2 8255±55 - cal BC 7520 – 7080 Proudfoot 2001: 122
Fordhouse barrow hazelnut shell from pit 21 OxA-8225 -23.1 8100±45 - cal BC 7300 – 6860 Proudfoot 1999
Fordhouse barrow hazelnut shell from layer 330 OxA-10058 -25.1 7920±50 - cal BC 7050 – 6640 Proudfoot 2001: 122
Fordhouse barrow hazelnut shell from pit 20 OxA-10057 -23.9 7890±50 - cal BC 7040 – 6590 Proudfoot 2001 122
Greenham Dairy
Farm
red deer bone with a stratified habitation
layer in a cut-off palaeochannel fill
Q-973 -25.0 (assumed) 8779±110 - cal BC 8240 – 7580 Switsur and West 1973:
542
Sheridan et al. 1967
Kettlebury Site 103 hazelnut shell from box H10.7 OxA-378 -25.0 (assumed) 8270±120 - cal BC 7580 – 7050 Gillespie et al. 1985
Reynier 2002: 226
Kettlebury Site 103 hazelnut shell from box 18.9 OxA-379 -25.0 (assumed) 7940±120 - cal BC 7290 – 6470 Gillespie et al. 1985
Reynier 2002: 226
Kettlebury Site 103 hazelnut shell from box 46A at 20cm
depth
OxA-6395 -25.0 (assumed) 7990±90 - cal BC 7290 – 6640 Reynier 2002: 226
Kettlebury Site 103 hazelnut shell from box 16B at 27cm
depth
OxA-6396 -25.0 (assumed) 7890±80 - cal BC 7060 – 6500 Reynier 2002: 226
Kinloch hazelnut shell from hollow BA S2 GU-2150 -25.7 8310±150 - cal BC 7600 – 7040 Wickham-Jones 1990
Kinloch hazelnut shell from lower fill of pit AJ2 GU-2040 -25.1 8560±75 - cal BC 7750 – 7480 Wickham-Jones 1990
Kinloch hazelnut shell from pit AD5 GU-1873 -24.9 8590±95 8575±85 cal BC 7940 – 7480 Wickham-Jones 1990
15 HOWICK IN ITS NORTH SEA CONTEXT 209
Calibrated date range (95%
confidence)
Reference
Kinloch replicate of GU-1873 GU-1874 -23.8 8515±190 T’=0.1; v=1; T’(5%)=3.8
cal BC 7780 – 7480 Wickham-Jones 1990
cal BC 8200 – 7070
Kinloch hazelnut shell from fill of pit BA1 GU-2146 -25.0 8080±50 - cal BC 7300 – 6830 Wickham-Jones 1990
Wickham-Jones 1990
Kinloch hazelnut shell from fill of hollow BA 10 GU-2147 -25.1 7880±70 - cal BC 7050 – 6500 Wickham-Jones 1990
Kinloch hazelnut shell from fill of pit BA3 GU-2145 -25.0 7850±50 - cal BC 7030 – 6530 Wickham-Jones 1990
Site Material and context Laboratory
Number
hazelnut shell from fill of pit complex
BA4/5
Kinloch replicate of GU-2039 GU-2149 -25.3 7570±50
Lussa 1 SRR-160 -28.2 8194±350 - cal BC 8200 – 6380 Mercer 1980;
Harkness and Wilson
1974: 249-50
Longmoor hazelnut shell OxA-376 -25.0 (assumed) 8930±100 - cal BC 8290 – 7740 Gillespie et al. 1985
Reynier 2002: 226
Lough Boora UB-6400 - 8350±70 - cal BC 7580 – 7180
Marsh Benham bulk hazelnut OxA-5195 -23.7 8905±80 - cal BC 8270 – 7740 Hedges et al. 1996: 191-
2
Mount Sandel Lower Mount Sandel Lower 13C (‰) Radiocarbon Age
(BP)
Weighted mean (BP) Kinloch GU-2039 -25.3 7925±65 Wickham-Jones 1990
T’=18.9; v=1; T’(5%)=3.8 cal BC 7060 – 6610
cal BC 6470 – 6260 Longmoor hazelnut shell OxA-377 -25.0 (assumed) 8760±110 - cal BC 8240 – 7580 Gillespie et al. 1985
Reynier 2002: 226
charcoal and Corylus shell from base of
small ring of stones
mixed charcoal and Corylus shell from
two small stone rings
eroded (?) bulk charcoal from cutting A,
layer 2
eroded (?) bulk charcoal from cutting C,
layer 2
Lough Boora bulk charcoal UB-2200 - 8350±70 - cal BC 7580 – 7180 Woodman 1978a: 323-5
Lough Boora bulk charcoal UB-2267 - 8450±70 - cal BC 7600 – 7330 Woodman 1978a: 323-5
Lough Boora bulk charcoal UB-2199 - 8475±75 - cal BC 7600 – 7350 Woodman 1978a: 323-5
Lussa 1 SRR-159 -30.5 7963±200 - cal BC 7520 – 6420 Mercer 1980;
Harkness and Wilson
1974: 249-50
UB-591 - 7720±525 - cal BC 7940 – 5560 Woodman 1985
Mount Sandel Lower bulk charcoal from cutting C, layer 1 UB-592 - 7360±695
UB-532 - 8490±210 Woodman 1985
8412±201
cal BC 8200 – 7050
T’=2.2; v=1; T’(5%)=3.8
cal BC 7960 – 7040
cal BC 7760 – 4800 Woodman 1985
Mount Sandel Upper charcoal and hazelnuts from F56/4 UB-2358 - 8795±135 - cal BC 8270 – 7570 Woodman 1985
Mount Sandel Upper charcoal from F109 UB-2359 - 7885±120 - cal BC 7080 – 6460 Woodman 1985
Mount Sandel Upper from F100/2, higher up than UB-912 GrN-10470 - 8380±50 T’=7.8; v=1; T’(5%)=3.8
cal BC 7580 – 7200 Woodman 1985
Mount Sandel Upper hazelnuts and charcoal from F100/2 UB-912 - 8725±115
cal BC 8230 – 7540 Woodman 1985
T’=6.7; v=2; T’(5%)=6.0
Mount Sandel Upper charcoal and hazelnuts from F56/1 UB-2360 - 8670±100 cal BC 8160 – 7540 Woodman 1985
Mount Sandel Upper charcoal from F56/1 UB-951 - 8790±185 cal BC 8410 – 7480 Woodman 1985
Mount Sandel Upper from F56/1 GrN-10471 - 8430±60
cal BC 7590 – 7330 Woodman 1985
Mount Sandel Upper hazelnuts and charcoal from F27 UB-2008 - 8440±65 - cal BC 7590 – 7330 Woodman 1985
Mount Sandel Upper charcoal from F74 UB-2361 - 8545±165 - cal BC 8200 – 7180 Woodman 1985
Mount Sandel Upper hazelnuts and charcoal from F31/0 UB-913 - 8555±70 - cal BC 7730 – 7480 Woodman 1985
Mount Sandel Upper hazelnuts + charcoal from F31 UB-2007 - 8795±135 - cal BC 8270 – 7570 Woodman 1985
Mount Sandel Upper charcoal from occupation layer L-M-
N/35
UB-2356 - 8765±135 - cal BC 8270 – 7540 Woodman 1985
Mount Sandel Upper charcoal from occupation layer N/31, UB-2357 - 8955±185 - cal BC 8610 – 7590 Woodman 1985
210
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Site Material and context Laboratory
Number
13C (‰) Radiocarbon Age
(BP)
Weighted mean (BP) Calibrated date range (95%
confidence)
Reference
Mount Sandel Upper hazelnuts + charcoal from F56/5 UB-952 - 8960±70 - cal BC 8290 – 7840 Woodman 1985
Mount Sandel Upper charcoal from F99 UB-2362 - 8990±80 - cal BC 8410 – 7940 Woodman 1985
charcoal from J6 Sandbank UB-637 - 8895±125 - cal BC 8300 – 7600 Woodman 1977
Ogof-yr-Ychen human bone OxA-7741 -15.9 8415±65 - cal BC 7590 – 7200 David 2004,
DATELIST
Ogof-yr-Ychen human bone OxA-7690 -15.0 8280±55 - cal BC 7520 – 7080 DATELIST
Ogof-yr-Ychen human bone OxA-7691 -14.4 8210±55 - cal BC 7450 – 7060 DATELIST
P/31-32
Newferry Site 3 Zone
7
Oakhanger Site VII hazelnut Q-1489 - 9225±170 - cal BC 9120 – 7970 Switsur and Jacobi
1979: 54
Oakhanger Site VII wood charcoal Q-1490 - 8995±160 - cal BC 8600 – 7600 Switsur and Jacobi
1979: 54
Oakhanger Site VII wood charcoal Q-1491 - 9100±160 - cal BC 8730 – 7820 Switsur and Jacobi
1979: 54
Oakhanger Site VII charcoal, Pinus and Corylus shell from
Q-1494 - 8885±160 - cal BC 8450 – 7580 Switsur and Jacobi
Level 2
1979: 54
Oakhanger Site VII charcoal, probably Pinus from Level 2 Q-1492 - 8975±160 - cal BC 8550 – 7600 Switsur and Jacobi
1979: 54
Oakhanger Site VII charcoal, probably Pinus from Level 2 Q-1490 - 8995±160 - cal BC 8600 – 7600 Switsur and Jacobi
1979: 54
Prestatyn hazelnut shell from the archaeological
OxA-2268 -23.5 8700±100 - cal BC 8200 – 7540 DATELIST
horizon
hazelnut shell from the archaeological
OxA-2269 -23.6 8730±90 - cal BC 8210 – 7580 DATELIST
horizon
Rhuddlan hazelnut shells BM-691 - 8739±86 - cal BC 8210 – 7580 Burleigh et al. 1976: 27-
8
Rhuddlan hazelnut shells BM-822 - 8528±73 - cal BC 7680 – 7380 Burleigh et al. 1976: 27-
8
Stump Cross bulk charcoal Q-141 - 8450±310 - cal BC 8270 – 6640 Godwin and Willis
1959: 69
Tolpits Lane bulk charcoal Q-1147 - 8260±120 - cal BC 7580 – 7040 Switsur and Jacobi
1979: 57
Switsur and West 1975:
44-5
Potter’s Cave human bone OxA-7688 -17.0 8580±60 - cal BC 7740 – 7530 DATELIST
Potter’s Cave human bone OxA-7687 -16.1 7880±55 - cal BC 7040 – 6590 DATELIST
Prestatyn Sand bevel-ended bone artefact in Spit 8 OxA-10152 -22.1 8470±90 - cal BC 7600 – 7320 Ashmore 2004: 102
Staosnaig hazelnut shell from small pit in Area A AA-21627 -25.1 8110±60 - cal BC 7320 – 6830 Mithen 2000: 373
Trwyn Du hazelnut shell from F13 Q-1385 - 8640±150 - cal BC 8210 – 7370 White 1978: 16-39
Trwyn Du hazelnut shell from F13 HAR-1194 - 8590±90 - cal BC 7940 – 7480 White 1978: 16-39
Warcock Hill III charcoal, oak and birch from Pit 5 Q-789 - 8606±110 - cal BC 7970 – 7480 Radley et al. 1974;
Wetton Mill Minor bone fragments in proximity to
Mesolithic flint artefacts
Q-1127 -21.0 (assumed) 8847±210 - cal BC 8540 – 7520 Switsur and West 1975:
45-6
15 HOWICK IN ITS NORTH SEA CONTEXT 211
212
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.3. Calibrated radiocarbon dates from English archaeological sites potentially contemporaneous with Howick.
direct comparison between sites. More in-depth
discussion of newly discovered settlement sites can
be found in the edited volume by Waddington and
Pedersen (2007). A few archaeological sites with 8th
millennium cal BC radiocarbon dates have been
excluded from this table where the association
between the sample and the archaeology is uncertain
(e.g. Hengistbury Head, Kew Bridge, Redkirk Point
and Thatcham Site II). The locations of the sites in
Table 15.1 are shown in Figure 15.2.
Figures 15.3–15.6 show the calibrated dates for
those sites where few measurements are available,
and where there is insufficient information for
chronological modelling. It must be made clear that
these dates are of mixed quality. Nine of the samples
are of unidentified charcoal, and so may have an ‘old
wood’ offset of up to several centuries. Two measure-
ments were obtained from material bulked from more
than one context (SRR-160 and Q-1127) and so may
represent average ages of material of varying cal-
endar date. A further nine measurements were
obtained from bulked fragments of charred plant
material from single contexts. These latter results may
provide a fair indication of the actual date of these
sites, if all the dated material was of the same age,
freshly deposited, and from short-lived species (e.g.
hazelnuts burnt in a hearth). Seventeen measure-
ments are available from single-entity samples, which
should accurately date activity at the sites concerned.
A range of sites is represented in Figs. 15.3–15.6,
including pits, occupation horizons, burnt deposits,
rock shelters, a midden, caves and the Lussa Wood
settlement. It is likely that the human skeletons dated
from Caldey Island may be rather later than shown
here (Fig. 15.5; Ogof-yr-Ychen and Potter’s Cave).
The stable isotope values from these individuals
indicate a high proportion of marine protein in their
diets, which would introduce a marine offset in the
interpretation of these radiocarbon measurements
(Schulting and Richards 2002, table 1). It is currently
difficult to quantify this effect reliably, however
(Bayliss et al. 2004).
Slightly more extensive dating has been obtained
from three other hut sites, of similar age and structural
form to Howick. Three determinations are currently
available from hazelnut shell fragments from a post
hole within the recently excavated hut at East Barns,
near Dunbar, Lothian, Scotland (Gooder in press).
These fall around 8000 cal BC (Fig. 15.7). At Broom
Hill, Hampshire (O’Malley and Jacobi 1978), three
bulk samples of unidentified charcoal from the
structure (pit 3) provide statistically consistent
measurements. These samples are earlier than those
from the overlying deposits (Q-1460 and Q-1191). This
information is incorporated in the chronological
model shown in Figure 15.8. The radiocarbon results
are consistent with the recorded archaeological
sequence, suggesting that there may not be a signif-
icant old wood effect. The model suggests that the hut
at Broom Hill was used in cal BC 7610–7300 (93%
probability, pit 3 (structure); Fig. 15.8). Fifteen bulk
samples have been dated from the Mount Sandel
settlement (Mount Sandel Upper), Co Londonderry,
Northern Ireland (Woodman 1985) although
additional dates have since been acquired and will be
published in due course (Woodman pers com.). All of
15 HOWICK IN ITS NORTH SEA CONTEXT 213
Figure 15.4. Calibrated radiocarbon dates from Scottish archaeological sites potentially contemporaneous with Howick.
Figure 15.5. Calibrated radiocarbon dates from Welsh archaeological sites potentially contemporaneous with Howick.
Figure 15.6. Calibrated radiocarbon dates from Irish archaeological sites potentially contemporaneous with Howick.
214
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.7. Calibrated radiocarbon dates from the hut at East Barns, East Lothian, Scotland (see Gooder in press).
Figure 15.8. Probability distributions of dates from Broom Hill, Hampshire. The format is identical to that of Figure 6.1. The large square
brackets down the left hand side along with the OxCal keywords define the overall model exactly.
Figure 15.9. Probability distributions of dates from the Mount Sandel settlement (Mount Sandel Upper), Co Londonderry, Northern
Ireland. The format is identical to that of Figure 1. The large square brackets down the left hand side along with the OxCal keywords define
the overall model exactly.
15 HOWICK IN ITS NORTH SEA CONTEXT 215
Figure 15.10. Probability distributions for the dates of occupation of Mesolithic huts analogous to Howick from the British Isles. These
distributions have been obtained from the chronological models shown in Figures 6.2, 15.4, 15.5 and 15.6.
these contained unidentified charcoal which might
suffer from the old-wood effect, although in the only
case where there is a stratigraphic relationship
between samples (in the main hut), this is consistent
with the radiocarbon results (UB-2358 and 56/1; Fig.
15.9). In both cases where there are replicate
measurements from the same deposit, however, these
are statistically inconsistent, suggesting that these
samples may have contained material of varying
calendar age or that there was some problem in
laboratory processing (as suggested by Woodman
[2004]). Because of this problem, the best estimate for
the use of the main hut area is probably provided by
the later samples, falling around 7500 cal BC (see Phase
main hut area; Fig. 15.9). The dated samples from
other features on the site outwith the main hut area
may be contemporary with this structure, or may be
somewhat earlier (see Phase other features; Fig. 15.9).
Because of the possibility of old-wood offsets, this
remains unknown.
Figure 15.10 shows the dates of occupation of huts
analogous to the Howick site. It can be seen that all
four structures fall in a cluster in the first half of the
eighth millennium cal BC (c. 8000–7500 cal BC). The
structure at East Barns is probably the earliest of the
four huts, followed by the Howick hut, and slightly
later by those at Mount Sandel and Broom Hill. We
have seen that the Howick hut was in use for several
generations (Fig. 6.4). The existing radiocarbon
measurements from the other huts do not allow
similar estimates to be made for these structures.
They may have been in use for a similar period, as
the stratigraphy at East Barns and Mount Sandel
could suggest. It should be noted that the scatter on
the measurements from these sites apparent in Figure
15.10 is an artefact of the limited dating evidence
available, and does not suggest a period of prolonged
use for these structures.
A model for the chronology of occupation at the
small Mesolithic camp at Fife Ness, Fife, Scotland is
shown in Figure 15.11. The results form a very tight
group, consistent with a limited period of occupation.
They are also consistent with the only stratigraphic
information included in the model – that the lower fill
of pit F41 is earlier than the upper fill. This model
suggests that the site was occupied between 7600–7520
216
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.11. Probability distributions of dates from the Mesolithic camp at Fife Ness, Fife, Scotland. The format is identical to that of Figure
6.1. The large square brackets down the left hand side along with the OxCal keywords define the overall model exactly.
cal BC (93% probability) and 7540–7400 cal BC (95%
probability). The pit group at Cramond, Edinburgh,
Scotland also seems to have been used for a limited
period of time in the third quarter of the 9th millen-
nium cal BC (Fig. 15.12). Again, the radiocarbon
measurements are consistent with the stratigraphic
information incorporated in the model that pit 1430
was earlier than pit 1432. These dates provide the
earliest evidence so far for the use of a typologically
‘Later Mesolithic’ tool kit in the British Isles (Saville
2004, 206–7).
The chronology of the occupation site at Kinloch,
Rhum, Scotland is shown in Figure 15.13. As sug-
gested by Wickham-Jones (1990, 135), it is possible
that the activity in areas AD and AJ is earlier than the
occupation in area BA. The radiocarbon results from
areas AD and AJ are statistically consistent, and so
may represent a single episode of occupation. The
radiocarbon measurements from area BA, however,
are not consistent, and so must represent a prolonged
phase of activity. It is not clear whether these two
phases were separated by a period of abandonment
or whether they form part of a chronological con-
tinuum of activity across the site. A model that
incorporates this latter interpretation shows good
agreement (Fig. 15.13, Aoverall=87.8%; Bronk Ramsey
1995), and so this is also a plausible explanation. In
either case, activity on the site spanned rather more
than 1000 years, from c. 7800 cal BC–c. 6500 cal BC.
The limited dating currently available for the
Mesolithic pits below the Fordhouse Barrow, Angus,
Scotland is shown in Figure 15.14. These results do
not form a consistent group, suggesting this activity
took place over some period in the centuries around
7000 cal BC.
Figure 15.15 shows the dating of the occupation
site at Lough Boora, Co. Offaly, Ireland. Despite being
bulk samples of unidentified charcoal from unknown
contexts, these results are remarkably consistent,
suggesting the site was occupied for a limited period
c. 7500 cal BC.
The site at Oakhanger VII, Hampshire, England
15 HOWICK IN ITS NORTH SEA CONTEXT 217
Figure 15.12. Probability distributions of dates from the pit group at Cramond, Edinburgh, Scotland. The format is identical to that of Figure
6.1. The large square brackets down the left hand side along with the OxCal keywords define the overall model exactly.
Figure 15.13. Probability distributions of dates from Kinloch, Rhum, Scotland. The format is identical to that of Figure 6.1. The large square
brackets down the left hand side along with the OxCal keywords define the overall model exactly.
also appears to have been occupied for a relatively
short period of time, as it has produced six consistent
radiocarbon measurements (Fig. 15.16; Jacobi 1981).
Two of the samples were of unidentified charcoal,
although the consistency of these measurements with
those from known short-life material (e.g. Q-1489)
suggests that wood-age offsets may not be a signif-
icant problem. However, the samples were of bulked
material from the phase 2 horizon and so may
represent material of a range of actual ages. Never-
218
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.14. Probability distributions of dates from Fordhouse Barrow, Angus, Scotland. The format is identical to that of Figure 6.1. The
large square brackets down the left hand side along with the OxCal keywords define the overall model exactly.
Figure 15.15. Probability distributions of dates from Lough Boora, Co. Offaly, Ireland. The format is identical to that of Figure 6.1. The large
square brackets down the left hand side along with the OxCal keywords define the overall model exactly.
Figure 15.16. Probability distributions of dates from Oakhanger VII, Hampshire, England. The format is identical to that of Figure 6.1. The
large square brackets down the left hand side along with the OxCal keywords define the overall model exactly.
15 HOWICK IN ITS NORTH SEA CONTEXT 219
Figure 15.17. Probability distributions of dates from Kettlebury, Surrey, England. The format is identical to that of Figure 6.1. The large
square brackets down the left hand side along with the OxCal keywords define the overall model exactly.
Figure 15.18. Probability distributions showing the length of time during which the occupation sites at Howick, Cramond, Fife Ness, and
Kinloch were in use. The distributions derive from the models shown in Figures 6.2–6.3, Figure 15.11, 15.12 and Figure 15.13.
theless, the consistency of the results suggests that
the site is likely to fall c. 8000 cal BC.
Figure 15.17 shows that the site at Kettlebury,
Surrey, England is also likely to have been occupied
for a relatively short period, c. 7000 cal BC. The four
results are statistically consistent, and all are single-
entity samples of short-life material.
Only four sites from all those discussed above have
a sufficient number of radiocarbon determinations,
of sufficient contextual quality, to enable a realistic
quantitative assessment of the duration of activity at
the site to be made (at Mount Sandel, it is unclear
whether there is an extended period of activity, or
whether the earlier dates suffer from old-wood
offsets). These estimates are shown in Figure 15.18. It
can be seen that three of these sites were in use for a
relatively short period of time, for a few centuries at
most. Kinloch is different, as activity at this site seems
to have occurred over more than a millennium. The
evidence is sketchy, but the sites at Kettlebury,
Oakhanger VII and Lough Boora also seem to have
been occupied for relatively short periods of time,
whereas the activity beneath the Fordhouse barrow
may have been of longer duration.
Figure 15.19 shows summaries of the dating
available from sites associated with broad-blade, non-
geometric lithic assemblages. At least two of these sites
(Lussa 1, Strathclyde, Scotland and Kettlebury, Surrey,
England) may have been occupied after 7000 cal BC.
Figures 15.20a and 15.20b show summaries of the
dating available for sites associated with narrow-
blade or micro-triangle blade assemblages. The site
at Cramond was probably in use a few centuries
before 8000 cal BC, and those at East Barns and
Howick shortly after this date, but by c. 7500–7000
cal BC, narrow-blade sites had spread further south
and west, with examples from County Durham (e.g.
Filpoke Beacon), Southern England (e.g. Broom Hill,
Hampshire), Wales (e.g. Prestatyn, Clywd), western
Scotland (e.g. Kinloch), and Ireland (e.g. Mount
Sandel Upper, Lough Boora).
There appears to be a period of overlap when both
broad blade and narrow blade lithic assemblages
were in use during the 8th millennium cal BC, and
possibly for as long as two millennia from c. 8500–c.
6500 cal BC.
Lithic Technology and the Adoption
of Narrow-Blade Traditions
The dates from the early North-East British sites
referred to above (Figs. 15.20a and b) are also by far
the earliest dates for narrow-blade stone tool in-
dustries in Britain (as all the east coast sites men-
tioned have an associated narrow-blade assemblage).
Therefore, not only does it seem just possible that the
220
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.19. Probability distributions of dates from sites associated with broad-blade, non-geometric lithic assemblages. The distributions
derive from the models shown in Figures 15.3–15.5, 15.16 and 15.17.
human colonisation/infilling of the north may have
been initiated along the North-East coast but the
narrow-blade techno-complex may have also spread
from this area to the rest of the British Isles. In short,
the narrow-blade industries that come to dominate
the later Mesolithic tool kits of the British Isles appear,
on the basis of the dating currently available, to have
spread into Britain first by way of the North Sea
Basin seaways rather than across the land bridge into
central and southern Britain, as has been suggested
before (e.g. Jacobi 1976).
When these early North-East British dates are
compared with those from elsewhere around the
North Sea Basin there is evidence for widespread
adoption of a narrow-blade technology at this time
(e.g. see Jacobi 1976). For example in western Sweden
Nordqvist (2000, 164) identifies triangle microliths as
first appearing in the Late Hensbacka period and
occurring through the Early Sandarna period and
partly into the Late Sandarna period. The earliest of
these Swedish periods correlate with the time span
represented by the British dated sites at Cramond,
East Barns and Howick while those of the later
periods correlate with the dates from Fife Ness,
Kinloch, Filpoke Beacon, Broomhead Moor site V,
Warcock Hill III, Broom Hill, Mount Sandel and
Lough Boora (see Dating Chapter). Elsewhere early-
mid 8th millennium cal BC dated assemblages
containing narrow-blade triangle microliths have
been documented around the North Sea Basin region
such as at Aardhorst-Vessem III, Rotsterhaule, Warns
and Milheeze II in the Netherlands (see Jacobi 1976),
in Belgium at Verrebroek Dok (Crombé 1988), as well
as at Maglemose sites in Denmark such as Barmose I
(Blankholm 1990), Stallerupholm (Jacobi 1976) and in
Germany sites such as Duvensee Site 13 (Bokelmann
1985). Although not all elements of these assemblages
are shared across this area, such as the lack of core-
axes from northern British sites, the general adoption
of a narrow-blade industry that utilised scalene
triangle forms together with backed blades (some-
times referred to as rods, though this is misleading as
15 HOWICK IN ITS NORTH SEA CONTEXT 221
Figure 15.20a. Probability distributions of dates from sites in England, Wales, and Ireland associated with narrow-blade lithic assemblages.
The distributions derive from the models shown in Figures 6.2, 15.3, 15.5, 15.6, 15.8, 15.9 and 15.15.
222
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Figure 15.20b. Probability distributions of dates from sites in Scotland associated with narrow-blade lithic assemblages. The distributions
derive from the models shown in Figures 15.4, 15.7, and 15.11–15.13.
15 HOWICK IN ITS NORTH SEA CONTEXT 223
they are of different form to the smaller and more
geometric ‘rods’ that belong to the latest phases of
the Mesolithic in Britain (e.g. the dated Pennine sites
excavated by Spikins [2002] and Chatterton [2007]),
indicates that communication networks were in place
that connected groups around the North Sea margin
in such a way as to allow for commonalities in their
choice of stone tool equipment. Whether or not this
reflects social and/or cultural affinities across a large
network of North-West European groups is another
question altogether (see also papers in Waddington
and Pedersen 2007), but at the level of technological
choices it does constitute what may be termed a
shared techno-complex. However, these common-
alities do not appear to have lasted beyond the
breaching of the land bridge c. 7000 cal BC (see
Shennan et al. 2000, 309) because, as Jacobi has
previously pointed out, it is soon after this that
artefact forms common to North-West Europe but
entirely absent from Britain, such as trapezoid
microliths and pressure-flaked points, emerge (Jacobi
1976). From this time on the material culture of the
British Isles follows a rather insular trajectory until
the advent of the Neolithic some 3000 years later.
The earliest dates now available for narrow-blade
sites, all from North and North-East Britain (Cram-
ond, East Barns and Howick), are earlier than some
of the broad-blade sites in western and southern
Britain (e.g. Lussa 1; Kettlebury; Trwyn Du and
Rhuddlan Site M; Figs 15.19, 15.20a and b). The same
narrow-blade sites are contemporary with, or in the
case of Cramond earlier than, other broad-blade sites
in southern and western Britain such as Oakhanger
VII, Longmoor, Greenham Dairy Farm, The Nab
Head Site 1; Daylight Rock (Caldey Island) and
Rhuddlan Site E (Figs. 15.19, 15.20a and b). This
unexpected evidence for the early adoption of
narrow-blade industries in North-East Britain calls
into question the longstanding view that broad-blade
and narrow-blade industries could be explained
primarily as a chronological division (Jacobi 1976;
Switsur and Jacobi 1979), subsequently linked to a
change in hunting strategies associated with an
improving climate and new types of fauna (Myers
1987). Rather, as noted by the recalibrating and
modelling of the available radiocarbon dates (Figs.
15.3–15.20b), there is a period of overlap between
broad-blade and narrow-blade sites during the late
9th and 8th millennia cal BC, and possibly for as long
as two millennia (c. 8500–c. 6500 cal BC). These new
data require a different explanation, as the
chronological argument is no longer adequate. It
remains true that after c. 7000 cal BC, narrow-blade
industries dominate all British Mesolithic assem-
blages, with broad-blade traditions having been
largely abandoned by this time, and it is also true that
there are broad-blade sites with earlier dates than the
earliest of the narrow-blade sites, such as Star Carr,
Thatcham and Marsh Benham. There is undoubtedly
a chronological dimension to this question, but the
division is blurred by a long period of overlap and a
geographical distinction between the location of the
chronologically overlapping sites. Taking this into
account it is suggested here that the broad-blade
industries could be associated with the Early Meso-
lithic groups who occupied Britain. Settlement
appears to have been most dense in southern regions
and northwards to the Tees, north of which settlement
at this time seems to be much less dense and perhaps
episodic, with only a handful of broad-blade sites
known, and virtually no associated dates. On present
evidence, the narrow-blade sites first appear in North
East Britain c. 8400 cal BC on the western shore of the
North Sea Basin (e.g. Cramond) and by c. 8000 cal BC
and shortly after appear to be widespread along this
coastline (e.g. East Barns, Howick), and even reaching
North Wales, perhaps via the northern seaways, as
implied by the dated site at Prestatyn (see also David
and Walker 2004). By c. 7600, narrow-blade sites had
spread north (e.g. Fife Ness), west (e.g. Kinloch) and
south (e.g. Filpoke Beacon) from the early North-East
British sites, and certainly by c. 7500 cal BC had
reached Ireland (e.g. Mount Sandel) and southern
England (e.g. Broom Hill). Therefore it can be
suggested that the narrow-blade techno-complex
perhaps arrived in Britain by way of human move-
ments around and across the North Sea Basin at a
time of rapid sea level rise and inundation of the
North Sea Plain. The adoption of this flaking tradition
could then have spread from North-East Britain to
the rest of what shortly became the British Isles.
Spread of the technology appears to have been rapid
and, at least in the initial phases, resulted from
movement along the seaways of the northern British
littoral. Over an overlapping period of perhaps 1500
years or so, the use of broad-blade forms seems to
have been abandoned in favour of the narrow-blade
technology. It is also noteworthy that it is narrow-
blade industries that are associated with all the
probable ‘permanent’ hut sites, which in itself
suggests a link between a particular type of settle-
ment system and the early adoption of narrow-blade
technology. Whether the initial adoption of narrow-
blade forms in North-East Britain represents the
arrival of displaced newcomers or the adoption by
pre-existing coastal groups of what was becoming a
widespread way of working flint around the North
Sea Basin at this time remains open to question, but
contact amongst North Sea groups is clearly evident.
The broad blade/narrow blade debate is a fascinating
question and sure to run for some time, but as our
understanding of the dating sequence and settlement
patterns in different regions of the British Isles
improves it is hoped greater clarity will emerge.
224
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Conclusion
The Howick results have opened up a new perspective
on Mesolithic settlement in northern England, and
one which has much wider implications. This mono-
graph has provided the results from the fieldwork
together with a discussion of the key topics arising
from this study. We hope that it will encourage further
comparative investigation and integration of recent
research in the British Isles, Scandinavia, Germany,
the Low Countries and France in relation to broader
issues such as the colonisation of northern Europe
and the nature of Early Holocene settlement. With
growing interest in hunter-gatherer marine
archaeology, the Howick results reinforce the impor-
tance of understanding the North Sea Basin, the
English Channel and the Irish Sea/St George’s
Channel not just as seaways but as distinct geographic
arenas around which human groups organised them-
selves and interacted.
Bibliography
Aaris-Sørensen, K. 1984. Om en uroksetyr fra Prejlerup – og
dens sammenstød med Maglemosekulturen. National-
museets Arbejdsmark 1984: 165–173.
Affleck, T.L., Edwards, K. and Clarke, A. 1988. Archaeological
and palynological studies at the Mesolithic pitchstone and
flint site of Auchareoch, Isle of Arran. Proceedings of the
Society of Antiquaries of Scotland 118: 37–59.
Albert, R.M., Bar-Yosef, O., Meignan, L. and Weiner, S. 2003.
Quantitative Phytolith Study of Hearths from the Natufian
and Middle Palaeolithic Levels of Hyonim Cave (Galilee,
Israel). Journal of Archaeological Science 30: 461–80.
Aldhouse-Green, S. and Pettitt, P. 1998. Paviland Cave:
contextualizing the ‘Red Lady’. Antiquity 72 (No. 278):
756–772.
Allen, M.J. 1991. Analysing the landscape: a geographical
approach to archaeological problems. In Schofield, A.J.
(ed.) Interpreting Artefact Scatters. Contributions to Plough-
zone Archaeology. Oxford, Oxbow Monograph 5: 39–57.
Andersen, S.H. 1995. Coastal adaptation and marine exploit-
ation in Late Mesolithic Denmark – with special emphasis
on the Limfjord region. In Fischer, A. (ed.) Man and Sea in
the Mesolithic. Coastal Settlement above and below present sea
level. Oxford, Oxbow Monograph 53.
Andersen, S.H. 2000. ‘Køkkenmøddinger’ (shell middens)
in Denmark: a survey. Proceedings of the Prehistoric Society
66: 361–84.
Andersen, S.H. and Johansen, E. 1986. Ertebølle revisited.
Journal of Danish Archaeology 5: 31–61.
Andrefsky, W. 1998. Lithics: Macroscopic Approaches to An-
alysis. Cambridge, Cambridge University Press.
Andrews, M.V., Gilbertson, D., Kent, M. and Mellars, P. 1985.
Biometric studies of the morphological variation in the
intertidal gastropod Nucella lapillus (L.): environmental
and palaeoeconomic significance. Journal of Biogeography
12: 71–87.
Archaeological Practice. 1996. Middle Warren fieldwalking
assessment: phases 1 and 2. Report by The Archaeological
Practice, University of Newcastle upon Tyne (unpub-
lished).
Ashmore, P. 1999. Radiocarbon dating: avoiding errors by
avoiding mixed samples. Antiquity 73: 124–30.
Ashmore, P. 2004. A datelist (to October 2002) for early
foragers in Scotland. In Saville, A. (ed.) Mesolithic Scotland
and its Neighbours. Edinburgh, Society of Antiquaries of
Scotland: 95–157.
ASUD (Archaeological Services University of Durham).
1998. Turning the Tide Archaeological Assessment. University
of Durham (unpublished report).
ASUD (Archaeological Services University of Durham).
2000. The Breamish valley Archaeology Project. Annual Report
2000. ASUD report 756. University of Durham.
Baales, M. 2001. From lithics to spatial and social organ-
ization: interpreting the lithic distribution and raw
material composition at the final Palaeolithic site of Kettig
(Central Rhineland, Germany). Journal of Archaeological
Science 28 (2): 127–142.
Bahn, P., Pettitt, P. and Ripoll, S. 2003. Discovery of Palaeo-
lithic cave art in Britain. Antiquity 77: 227–31.
Bailey, G.N. and Craighead, A. 2003. Late Pleistocene and
early Holocene coastal palaeoeconomies: a reconsider-
ation of the molluscan evidence from Northern Spain.
Geoarchaeology: an International Journal 18 (2): 175–204.
Balicki, A. 1965. The Netsilik Eskimo. New York, Natural
History Press.
Ball, D. F. 1964. Loss-on-ignition as an estimate of organic
matter and organic carbon in non-calcareous soils. Journal
of Soil Science 15: 84–92.
Bamforth, D.B. 1987. Investigating microwear polishes with
blind tests: the institute results in context. Journal of
Archaeological Science 15: 11–23.
Bang Andersen, S. 1996. The Colonization of Southwest
Norway: An Ecological Approach. In L. Larsson The
Earliest Settlement of Scandinavia and Its Relationship With
Neighbouring Areas. Stockholm, Almqvist and Wiksell
International: 219–234.
Barlow, C. and Mithen, S. 2000. The experimental use of
elongated pebble tools. In Mithen, S. (ed.) Hunter-Gatherer
Landscape Archaeology. The Southern Hebrides Mesolithic
Project 1988–98. Cambridge, McDonald Institute Mono-
graphs Vol 2: 513–521.
Bartley, D.D. 1966. Pollen analysis of some lake deposits
near Bamburgh in Northumberland. New Phytologist 65:
141–165.
Barton, H., Torrence, R. and Fullager, R. 1998. Clues to stone
tool function re-examined. Journal of Archaeological Science
25: 1231–1238.
Bateson, E. 1895. A History of Northumberland Vol II. New-
castle upon Tyne, Northumberland County History
Committee, Austin Reid and Company.
Bayliss, A., Shepherd Popescu, E., Beavan-Athfield, N.,
Bronk Ramsey, C., Cook, G.T. and Locker, A. 2004. The
potential significance of dietary offsets for the interpret-
ation of radiocarbon dates: an archaeologically significant
example from medieval Norwich. Journal of Archaeological
Science 31: 563–75.
Beckensall, S. 1976. The excavation of a rock shelter at
Corby’s Crags, Edlingham. Archaeologia Aeliana 5th Ser. 4:
11–16.
226
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Beckensall, S. 2001. Prehistoric Rock Art in Northumberland.
Stroud, Tempus.
Bell, M.G. 1997. Environmental Archaeology in the coastal
zone. In Fulford, M., Champion, T. and Long, A. (eds)
England’s Coastal Heritage. London, English Heritage: 56–73.
Berridge, P. and Roberts, A. 1986. The Mesolithic period in
Cornwall. Cornish Archaeology 25: 7–34.
Bettinger, R.L., Malhi, R. and McCarthy, H. 1997. Central
place models of acorn and mussel processing. Journal of
Archaeological Science 24: 887–99.
Biggins, J.A., Biggins, J., Coxon, B. and Watson, M. 1997.
Geophysical Survey of the Prehistoric Settlement at
Gardener’s Houses Farm, Dinnington. Durham Archaeo-
logical Journal 13: 43–53.
Biggins, J.A. and Taylor, D.J.A. 2004. A geophysical survey
at Housesteads Roman fort, April 2003. Archaeologia
Aeliana 5th Ser. 33: 1–9.
Biggins, J.A., Waddington, C. and Johnson B. 2002. Survey of
the Iron Age/Romano-British defended settlement at Howick
Haven, University of Newcastle (unpublished report
submitted to Northumberland County Council).
Binford, L.R. 1978a. Dimensional analysis of behavior and
site structure: learning from an Eskimo hunting stand.
American Antiquity 43 (3): 330–361.
Binford, L.R. 1978b. Nunamiut ethnoarchaeology. New York,
Academic Press.
Bird, D.W. and Bliege Bird, R.L. 1997. Contemporary
shellfish gathering strategies among the Meriam of the
Torres Strait Islands, Australia: testing predictions of a
central place foraging model. Journal of Archaeological
Science 24: 39–63.
Bjerck, H. B. 1995. The North Sea Continent and the Pioneer
Settlement of Norway. In A. Fischer (ed.) Man and Sea in
the Mesolithic: Coastal Settlement Above and Below Present
Sea Level. Proceedings of the International Symposium,
Kalundborg, Denmark 1993. Oxford, Oxbow Publishers:
131–144.
Blankholm, H.P. 1985. Maglemosekulturens hyttegrundrids.
En undersøgelser af bebyggelse og adfærdsmønstre i
tidlig mesolitisk tid. Aarbøger for nordisk Oldkyndighed og
Historie 1984: 61–77.
Blankholm, H.P. 1987. Maglemosian hutfloors: an analysis
of the dwelling unit, social unit and intra-site behavioural
patterns in Early Mesolithic Southern Scandinavia. In
Rowley-Conwy, P., Zvelebil, M. and Blankholm, H.P. (eds)
Mesolithic Northwest Europe: Recent Trends. Sheffield,
Department of Archaeology and Prehistory, University
of Sheffield: 109–120.
Blankholm, H.P. 1990. Stylistic Analysis of Maglemosian
Microlithic Armatures in Southern Scandinavia: An
Essay. In P.M. Vermeersch and P. Van Peer (ed.) Contrib-
utions to the Mesolithic in Europe: Papers Presented at the
Fourth International Symposium ‘The Mesolithic in Europe’,
Leuven 1990. Leuven, Leuven University Press: 239–257.
Blankholm, H.P. 1993a. On the Track of a Prehistoric Economy:
Maglemosian Subsistence in Early Postglacial South Scand-
inavia. Århus, Aarhus University Press.
Blankholm, H.P. 1993b. Barmose I revisited. Mesolithic
Miscellany 14 (1 and 2): 12–14.
Boismier, W.A. 1997. Modelling the Effects of Tillage Processes on
Artefact Distributions in the Ploughzone. A simulation study
of tillage-induced pattern formation. Oxford, British Archaeo-
logical Reports, British Series 259.
Bokelmann, K. 1985. A Rest Under the Trees: An Ephemeral
Mesolithic Camp in the Duvensee Peatbog. Mesolithic
Miscellany 6: 7 - 10.
Bonsall, C. 1984. Low Hauxley, Northumberland: Meso-
lithic/Bronze Age coastal site. Proceedings of the Prehistoric
Society 50: 398.
Bonsall, C. 1996. The ‘Obanian’ problem. Coastal adapt-
ations in the Mesolithic of western Scotland. In Pollard, T.
and Morrison, A. (eds) The Early Prehistory of Scotland.
Edinburgh, Edinburgh University Press: 183–197.
Bonsall, C., Tolan-Smith, C. and Saville, A. 1995. Direct
dating of Mesolithic antler and bone artefacts from Great
Britain: new results for bevelled tools and red deer antler
mattocks. Mesolithic Miscellany 16 (1): 2–10.
Boomer, I., Waddington, C., Stevenson, A.C. and Hamilton,
D. (2007.) Holocene coastal change and geoarchaeology
at Howick, Northumberland. The Holocene, 17.1: 89–104.
Borek, M.J.E. 1975. Pollen analysis and vegetational history of the
Akeld Basin. Unpublished MSc thesis, University of
Durham.
Bosanquet, R.C. 1934. Stone Axe found near Longhoughton.
Proceedings of the Society of Antiquaries of Newcastle 4th Ser.
6 (5): 202–203.
Bradley, R. 1987. Flint Technology and the Character of
Neolithic Settlement. Lithic Analysis and Later British
Prehistory. Some problems and approaches. In Brown,
A.G. and Edmonds, M.R. (eds) Lithic analysis in later British
prehistory, Oxford, British Archaeological Reports British
Series 162: 181–185.
Bradley, R. 2001. The birth of architecture. Proceedings of the
British Academy 110: 69–92.
Brain, C.K. 1981. The Hunters or the Hunted? An introduction to
African cave taphonomy. Chicago, Chicago University Press.
Brinch Petersen, E. 1989. Vænget Nord: Excavation, Doc-
umentation and Interpretation of a Mesolithic Site at
Vedbæk, Denmark. In Bonsall, C. (ed.) The Mesolithic in
Europe: Papers Presented at the Third International Sympos-
ium, Edinburgh 1985. Edinburgh, John Donald Publishers:
325–330.
Bronk Ramsey, C. 1995. Radiocarbon calibration and an-
alysis of stratigraphy. Radiocarbon 36: 425–430.
Bronk Ramsey, C. 1998. Probability and dating. Radiocarbon
40: 461–74.
Bronk Ramsey, C. 2001. Development of the radiocarbon
calibration program. Radiocarbon 43: 355–363.
Bronk Ramsey, C. and Hedges, R.E.M. 1997. Hybrid ion
sources: radiocarbon measurements from microgram to
milligram. Nuclear Instruments and Methods in Physics
Research B 123: 539–545.
Bronk Ramsey, C., Higham, T. and Leach, P. 2004. Towards
high precision AMS: progress and limitations. Radiocarbon
46: 17–24.
Brown, A. 1997. Clearances and clearings: Deforestation in
Mesolithic/Neolithic Britain. Oxford Journal of Archaeology
16: 133–146.
Buck, C.E., Cavanagh, W.G. and Litton, C.D. 1996. Bayesian
Approach to Interpreting Archaeological Data. Chichester, J.
Wiley and Son.
Buck, C.E., Christen, J.A., Kenworthy, J.B. and Litton, C.D.
1994. Estimating the duration of archaeological activity
using 14C determinations. Oxford Journal of Archaeology 13:
229–40.
Buck, C.E., Kenworthy, J.B., Litton, C.D. and Smith, A.F.M.
1991. Combining archaeological and radiocarbon inform-
ation: a Bayesian approach to calibration. Antiquity 65:
808–21.
Buck, C.E., Litton, C.D. and Scott, E.M. 1994. Making the
most of radiocarbon dating: some statistical consider-
ations. Antiquity 68: 252–63.
Buck, C.E., Litton, C.D. and Smith, A.F.M. 1992. Calibration
of radiocarbon results pertaining to related archaeo-
logical events. Journal of Archaeological Science 19: 497–512.
Buckley, F. 1922. Early Tardenoisian remains at Bamburgh.
Proceedings of the Society of Antiquaries of Newcastle upon
Tyne 3 (10): 319–23.
Buckley, F. 1925. The microlithic industries of Northumber-
land. Archaeologia Aeliana 4th Ser. 1: 42–47.
Burgess, C.B. 1972. Goatscrag, A Bronze Age rock shelter
cemetery in North Northumberland. Archaeologia Aeliana
4th Ser. 50: 15–69.
Burleigh, R., Hewson, A. and Meeks, N. 1976. British
Museum Natural Radiocarbon Measurements VIII. Radio-
carbon 18: 16–42.
Carruthers, R.G., Burnett, G.A. and Anderson, W. 1932. The
geology of the Cheviot Hills. Memoir of the Geological
Survey of England and Wales, Sheets 3 and 5. London, His
Majesty’s Stationery Office: 174.
Carlsson, T., Kaliff, A. and Larsson, M. 1999. Man and the
landscape: aspects of mental and physical settlement
organization. In Boaz, J. (ed.) The Mesolithic of Central
Scandinavia. Oslo: University of Oldsakssamling: 47–72.
Cassidy, J., Raab, J.M. and Kononenko, A.N. 2004. Boats,
bones and bifaces: The early Holocene mariners of Eel
Point, San Cklemente Island, California. American An-
tiquity 69 (1): 109–130.
Chatterton, R. 2007. South Haw: an upland Mesolithic Site
in its local and regional context. In Waddington, C. and
Pedersen, K. (eds) Mesolithic Studies in the North Sea Basin
and Beyond. Proceedings of a conference held in Newcastle in
2003. Oxford, Oxbow: 69–80.
Clapperton, C.M., Durno, S.E. and Squires, R.H. 1971.
Evidence for the Flandrian history of the Wooler Water,
Northumberland, provided by pollen analysis. Scottish
Geographical Magazine 57: 14–20.
Clark, J.G.D. 1946. Seal-hunting in the Stone Age of north-
western Europe: a study in economic prehistory. Pro-
ceedings of the Prehistoric Society 12: 12–48.
Clark, J.G.D. 1971. Excavations at Star Carr. An Early Mesolithic
Site at Seamer Near Scarborough, Yorkshire. Cambridge,
Cambridge University Press.
Clark, J. G. D. 1972. Excavations at Star Carr: a case study in
bioarchaeology. Reading, Mass., Addison-Wesley Modular
Publications, Module 10.
Clark, J.G.D. and Rankine, W. F. 1939. Excavations at
Farnham, Surrey (1937-38): The Horsham Culture and
the question of Mesolithic dwellings. Proceedings of the
Prehistoric Society 5: 61–118.
Clarke, A. Forthcoming. The coarse stone assemblage from Sand,
Skye.
Clarke, A. 1990. Coarse stone tools. In Wickham-Jones, C.R.
(ed.) Rhum: Mesolithic and Later Sites at Kinloch, excavations
1984–86. Society of Antiquaries of Scotland Monograph
Series 7.
Clarke, D. 1976. Mesolithic Europe: The Economic Basis.
London, Gerald Duckworth Limited.
Clutton-Brock, J. 1987. A Natural History of Domesticated
BIBLIOGRAPHY 227
Mammals. London, British Museum (Natural History).
Coles, J.M. 1971. The early settlement of Scotland: excav-
ations at Morton, Fife. Proceedings of the Prehistoric Society
37: 284–366.
Coles, J.M. 1983. Morton revisited. In O’Connor, A. and
Clarke, D.V. (eds) From the Stone-Age to the Forty-Five.
Edinburgh, John Donald: 9–18.
Coles, B. J. 1998. Doggerland: A Speculative Survey. Pro-
ceedings of the Prehistoric Society 64: 45–81.
Connock K.D., Finlayson, B. and Mills, A.C.M. 1992. The
excavation of a shell midden site at Carding Mill Bay,
near Oban, Scotland. Glasgow Archaeological Journal 17: 25–
38.
Corbet, G.B. and Harris, S. 1991. The Handbook of British
Mammals, 3rd edition. Oxford, Blackwell.
Cormack, W.F. and Coles, J.M. 1968. Mesolithic site at Low
Clone, Wigtonshire. Transactions of the Dumfriesshire and
Galloway Natural History and Antiquities Society 45: 44–72.
Crombé, P. 1998. The Mesolithic in Northwestern Belgium: Recent
Excavations and Surveys. Oxford: British Archaeological
Reports International Series 716.
David, A. 1995. Geophysical survey in archaeological field
evaluation. Ancient Monuments Laboratory. London, English
Heritage.
David, A. and E. Walker 2004. Wales during the Mesolithic
period. Mesolithic Scotland and its Neighbours. In
Saville, A. (ed.) The Early Holocene Prehistory of Scotland, its
British and Irish Context and Some Northern European
Perspectives. Edinburgh, Society of Antiquaries of Scot-
land: 299–337.
David, B. 1990. How was this bone burnt? In Solomon, S.,
Davidson, I. and Watson, D. (eds) Problem Solving in
Taphonomy: Archaeological Studies from Europe, Africa and
Oceania. Queensland, Australia. Tempus, Anthropology
Museum, the University of Queensland: Archaeology
and Material Culture Studies in Anthropology Vol. 2: 65–
79.
Davies, G. and Turner, J. 1979. Pollen diagrams from
Northumberland. New Phytologist 82: 783–804.
Davies, J. 1983. The Mesolithic Sites of Northumberland.
Northern Archaeology 4(ii): 18–24.
Davies, J. 2004. Lithics and small finds from the Bolam and
Shaftoe region of southern Northumberland. Northern
Archaeology 20: 1–132.
Degerbøl, M. 1961. On a find of a Preboreal domestic dog
(Canis familiaris L.) from Star Carr, Yorkshire, with
remarks on other Mesolithic dogs. Proceedings of the
Prehistoric Society 27: 35–55.
Deith, M.R. 1983. Molluscan calendars: the use of growth-
line analysis to establish seasonality of shellfish collection
at the Mesolithic site of Morton, Fife. Journal of Archaeo-
logical Science 10: 423–40.
Donahue, D.J., Beck, J.W., Biddulph, D., Burr, G.S., Courtney,
C., Damon, P.E., Patheway, A.L., Hewitt, L., Jull, A.J.T.,
Lange, T., Lifton, N., Maddock, R., McHargue, L.R.,
O’Malley, J.M. and Toolin, L.J. 1997. The status of the
NSF-Arizona AMS laboratory. Nuclear Instruments and
Methods in Physics Research B. 123: 51 –56.
Drucker, P. 1955. Indians of the Northwest Coast. American
Museum of Natural History, Anthropological Handbook
10.
Dumont, J.V. 1985. A preliminary report on the Mount Sandel
microwear study. In Woodman, P.C. (ed.) Excavations at
228
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Mount Sandel 1973–77. Belfast: Dept. for the Environment
for Northern Ireland. HMSO. (Northern Ireland Archaeo-
logical Monographs No 2): 61–70.
Dyer, B. 1902 Results of investigations on the Rothamsted soils.
USDA Official Experimental Station bulletin 106).
Eliade, M. 1959. The Sacred and Profane: The Nature of Religion.
(translated from the French by Willard R. Trask). New
York, Harcourt Brace Jovanovich.
Eliade, M. 1964. Shamanism: Archaic Techniques of Ecstasy.
(Bollingen Series LXXVI). Princeton, Princeton University
Press.
Farrimond, P. and Flanaghan, R.L. 1996. Lipid stratigraphy
of a Flandrian peat bed (Northumberland, UK): A
comparison with the pollen record. The Holocene 6: 69–74.
Fearnley-Whittingstall, H. 2001. The River Cottage Cookbook.
London, Harper Collins.
Fischer, A. (ed.) 1995. Man and Sea in the Mesolithic: Coastal
Settlement Above and Below Present Sea Level. Proceedings of
the International Symposium, Kalundborg, Denmark 1993.
Oxbow Monograph 53. Oxford, Oxbow Books.
Fischer, A., Vemming Hansen, P. and Rasmussen, P. 1984.
Macro and micro wear traces on lithic projectile points:
experimental results and prehistoric examples. Journal of
Danish Archaeology 5: 7–13.
Finlayson, B. 1989. A Pragmatic Approach to the Functional
Analysis of Chipped Stone Tools. Edinburgh: Edinburgh
University. Unpublished PhD Thesis.
Finlayson, B. 1990. The function of microliths, evidence from
Smittons and Starr SW Scotland. Mesolithic Miscellany 11
(1): 2–6.
Finlayson, B. 1995. Complexity in the Mesolithic of the
Western Seaboard. In Fischer, A. (ed.) Man and Sea in the
Mesolithic. Oxford, Oxbow: 261–264.
Finlayson B. and Mithen, S. 2000. The Morphology and
Microwear of Microliths from Bolsay farm and Gleann
Mor: a Comparative Study. In Mithen, S. (ed.) Hunter-
gatherer landscape archaeology: The Southern Hebrides Meso-
lithic Project 1988–98. Volume 2. Cambridge, McDonald
Institute Monographs: 589–593.
Foley, R. 1981. A model of regional archaeological structure.
Proceedings of the Prehistoric Society 47: 1–17.
Friis Hansen, J. 1990. Mesolithic cutting arrows: functional
analysis of arrows used in the hunting of large game.
Antiquity 64 (4): 494–504.
Fullager, R. 1993. Flakes stone tools and plant food pro-
duction: a preliminary report on obsidian tools from
Talasea, West New Britain, PNG. In Anderson, P.C.,
Beyries, S., Otte, M. and Plisson H. (eds) Traces et Fonction:
les gestes retrouvés 1. Liege, ERAUL 50: 331–337.
Fullager, R. (ed.) 1998. A Closer Look. Recent Australian Studies
of Stone Tools. Sydney: Sydney University Archaeological
Methods Series 6.
Gelfand, A.E. and Smith, A.F.M. 1990. Sampling approaches
to calculating marginal densities. Journal of the American
Statistical Association 85: 398–409.
Gendel, P.A. 1984. Mesolithic Social Territories in Northwestern
Europe. Oxford, British Archaeological Reports Inter-
national Series 218.
Gibbons, M. and Gibbons, M. 2004. Dyeing in the Mesolithic.
Archaeology Ireland 17: (4,66): 28–3.
Gilks, W.R., Richardson, S. and Spiegelhalther, D.J. 1996.
Markov Chain Monte Carlo in practice. London, Chapman
and Hall.
Gillespie, R., Gowlett, J.A.J., Hall, E.T., Hedges, R.E.M. and
Perry, C. 1985. Radiocarbon dates from the Oxford AMS
system: Archaeometry datelist 2. Archaeometry 27: 237–
246.
Godwin, H. and Willis, E.H. 1959. Cambridge University
Natural Radiocarbon Measurements I. Radiocarbon 1: 63–
75.
Gooder, J. 2007. Excavation of a Mesolithic house at East
Barns, East Lothian, Scotland: an interim report. In Wad-
dington, C. and Pedersen, K.L.R. (eds) Mesolithic Studies
in the North Sea Basin and Beyond. Proceedings of a Con-
ference Held at Newcastle in 2003. Oxford, Oxbow: 49–59
Gould, R.A. 1980. Living Archaeology, New Studies in Archaeo-
logy. Cambridge, Cambridge University Press.
Gowlett, J.A.J., Hedges, R.E.M., Law, I.A. and Perry, C. 1986.
Radiocarbon dates from the Oxford AMS system: Arch-
aeometry datelist 4. Archaeometry 28: 206–221.
Grace, R. 1989. Interpreting the Function of Stone Tools. The
quantification and computerization of microwear analysis.
Oxford, BAR International Series 474.
Grace, R., Graham, I.D.G. and Newcomer, M.H. 1985. The
quantification of microwear polishes. World Archaeology
17 (1): 112–120.
Grace, R., Ataman, K., Fabregas, R. and Haggren, C.M.B.
1988. A Multivariate approach to the functional analysis
of stone tools. In Beyries, S (ed.) Industries Lithiques CNRS.
Oxford, British Archaeological Reports International
Series 411 (1): 217–230.
Green, F. 1996. Mesolithic or later houses at Bowman’s Farm,
Romsey Extra, Hampshire, England? In T. Darvill and J.
Thomas (eds.) Neolithic Houses in Northwest Europe and
Beyond. Oxford: Oxbow Monograph 57: 113–122.
Griffitts, J. and Bonsall, C. 2002. Experimental determination
of the function of antler and bone ‘bevel-ended tools’
from prehistoric shell middens in Western Scotland. In
Choyke, A.M. and Bartosiewicz, L. (eds) Crafting Bone:
Skeletal Technologies through Time and Space. BAR Inter-
national Series 937: 207–220.
Grigson, C. and Mellars, P.A. 1987. The mammalian remains
from the middens. In Mellars, P.A. Excavations on Oronsay:
Prehistoric Human Ecology on a Small Island. Edinburgh,
Edinburgh University Press: 243–89.
Grimm, E.C. 1991. Tilia and TiliaGraph. Illinois, University of
Illinois.
Grøn, O. 1983. Social behaviour and settlement structure:
preliminary results of a distribution analysis on sites of
the Maglemose Culture. Journal of Danish Archaeology 2:
32–42.
Grøn, O. 1987. Reconstruction of the social structure of the
Maglemose Culture of Southern Scandinavia and North-
ern Germany. Mesolithic Miscellany 8(1): 18–19.
Grøn, O. 1988. Seasonal Variation in Maglemosian Group
Size and Structure. Current Anthropology 28 (4): 303–317.
Grøn, O. 1989. General spatial behaviour in small dwellings:
a preliminary study in ethnoarchaeology and social
psychology. In Bonsall, C. (ed.) The Mesolithic in Europe:
Papers Presented at the Third International Symposium,
Edinburgh 1985. Edinburgh, John Donald Publishers
Limited: 99–105.
Grøn, O. 1990. A large Maglemosian winter house? Mesolithic
Miscellany 11(1): 7–13.
Grøn, O. 1992. Some comments on the interpretation of
Barmose I. Mesolithic Miscellany 13(2): 12.
Grøn, O. 1994. Barmosen I – continued. Mesolithic Miscellany
15(1): 22–24.
Grøn, O. 1995. The Maglemose Culture: The Reconstruction
of the Social Organization of a Mesolithic Culture in
Northern Europe. Oxford, Tempus Reparatum.
Grøn, O. 1999. A Revision of the model for dwelling
organization in the Southern Scandinavian Mesolithic. In
Thévenin, A. and Bintz, P. (eds.) L’Europe des derniers
chasseurs. Épipaléolithique et mésolithique. Peuplement et
paléoenvironnement de l’Épipaléolithique et du Mésolithique.
Actes du 5e colloque international UISPP, commission XII,
Grenoble, 18–23 septembre 1995. (Documents préhistorique
11). Paris, Éditions du Comité des travaux historiques et
scientifiques: 321–326.
Grøn, O. and Sørensen, S.A. 1993. Møllegabet II – a sub-
merged Mesolithic site and a ‘boat burial’ from Ærø.
Journal of Danish Archaeology 10: 38–50.
Grøn, O. and Sørensen, S.A. 1995. Aggemose: an inland site
from the Early Kongemose Culture on Langeland. Journal
of Danish Archaeology 11: 7–18.
Hampton O.W. 1999. Culture of Stone. Sacred and Profane Uses
of Stone among the Dani. Texas AandM University Press.
Hardy, K. 1993 Preceramic Lithics in Central Mexico: An examin-
ation of the Tehuacan and Oaxaca chronological sequences.
London: University College London. Unpublished PhD
Thesis.
Hardy, K. 2004. Microwear Analysis of a Sample of Flaked
Stone Tools from Camas Daraich. In Wickham-Jones, C.R.
and K. Hardy. Camas Daraich: a Mesolithic site at the Point of
Sleat, Skye. (SAIR). Edinburgh.
Hardy, K. Forthcoming a. Use-wear analysis of a sample of
flaked stone tools from Sands of Forvie. In Warren, G.M.
Fieldwork at the Sands of Forvie, Aberdeenshire.
Hardy K. Forthcoming b. Worked bone. In Hardy, K. and
Wickham-Jones, C.R. (eds.) Mesolithic and later sites
around the Inner Sound, Scotland: the work of the
Scotland’s First Settlers project 1998–2004. Edinburgh. To
be published in SAIR (Scottish Archaeological Internet
Reports).
Hardy, K. and Sillitoe, P. 2003. Material Perspectives: Stone
Tool Use and Material Culture in Papua New Guinea.
Internet Archaeology 14. http://intarch.ac.uk/journal/
issue14/hardy_index.html
Hardy, K. and Wickham-Jones, C.R. 2003. Scotland’s First
Settlers: the Mesolithic seascape of the Inner Sound, Skye
and its contribution to the early prehistory of Scotland.
Antiquity 76: 825–833.
Hardy K. and Wickham-Jones C. Forthcoming. Mesolithic
and later sites around the Inner Sound, Scotland: the
work of the Scotland’s First Settlers project 1998–2004. To
be published in SAIR. Edinburgh.
Hardy, K., Miket, R. and Saville, A. Forthcoming. An Corran,
Staffin, Skye: a rockshelter with Mesolithic and later occupation.
Hayden, B. 1979. Palaeolithic Reflections: Lithic technology of the
Western Desert Aborigines. New Jersey: Humanities Press
Inc. (for the Australian Institute of Aboriginal Studies.
Canberra, Australia).
Harkness, D.D. and Wilson, H.W. 1974. Scottish Universities
Research and Reactor Centre Radiocarbon Measurements
II. Radiocarbon 16: 238–251.
Haselgrove, C.C. and Healey, F. 1992. The prehistory of the
Tyne-Tees lowlands: some recent finds. Durham Archaeo-
logical Journal 8: 1–24.
BIBLIOGRAPHY 229
Hather, J.G. 1998. Identification of macroscopic charcoal
assemblages. In Mellars, P. and Dark, P. (eds) Star Carr in
Context: New Archaeological and Palaeoecological Invest-
igations at the Early Mesolithic site of Star Carr, North
Yorkshire. Cambridge, MacDonald Institute: 183–196.
Hayden, B. 1979. Palaeolithic Reflections: Lithic technology of the
Western Desert Aborigines. New Jersey, Humanities Press
Inc. (for the Australian Institute of Aboriginal Studies.
Canberra, Australia).
Hayden, B. 1990. The right rub: hide working in high
ranking households. In Graslund, B. (ed.) The Inter-
pretative Possibilities of Microwear Studies. Uppsala, Aun 14:
89–102.
Healy, F., Heaton, M. and Lobb, S.J. 1992. Excavations of a
Mesolithic site at Thatcham, Berkshire. Proceedings of the
Prehistoric Society 58: 41–76.
Hedges, R.E.M., Bronk, C.R. and Housley, R.A. 1989. The
Oxford Accelerator Mass Spectrometry facility: technical
developments in routine dating. Archaeometry 31: 99–113.
Hedges, R.E.M., Housley, R.A., Law, I.A., and Bronk, C.R.
1989. Radiocarbon dates from the Oxford AMS system:
Archaeometry datelist 9. Archaeometry 31: 207–34.
Hedges, R.E.M., Housley, R.A., Law, I.A., Perry, C. and
Gowlett, J.A.J. 1987. Radiocarbon dates from the Oxford
AMS system, Archaeometry datelist 6. Archaeometry 29:
289–306.
Hedges, R.E.M., Housley, R.A., Pettitt, P.B., Bronk Ramsey,
C. and van Klinken, G.J. 1996. Radiocarbon dates from
the Oxford AMS system, Archaeometry datelist 21.
Archaeometry 38: 181–207.
Hedges, R.E.M., Pettitt, P.B., Bronk Ramsey, C. and van
Klinken, G.J. 1998. Radiocarbon dates from the Oxford
AMS system: Archaeometry datelist. Archaeometry 26:
437–455.
Hesjedal, A., Damm, C., Olsen, B. and Storli, I. 1996. Arkeologi
pä Slettnes. Dokumentasjon av 11.000 års bosetning. Tromsø,
Tromsø Museums Skrifter XXVI.
Hodson, A. 1971. Concepts in Statistical Mechanics. New York,
Gordon and Breach Science Publishers.
Houtsma, P., Kramer, E., Newell, R.R. and Smit, R.L. 1996.
The Late Palaeolithic Habitation of Haule V: From Excavation
Report to the Reconstruction of Federmesser Settlement Patterns
and Land-Use. Assen, van Gorcum.
Huntley, J.P. and Stallibrass, S.M. 1995. Plant and Vertebrate
Remains from Archaeological Sites in Northern England: Data
Reviews and Future Directions. Durham, Architectural and
Archaeological Society of Durham and Northumberland.
Innes, J.B. and Frank, R.M. 1988. Palynological evidence for
late Flandrian coastal changes at Druridge Bay, North-
umberland. Scottish Geographical Magazine 104: 14–23.
Isbister, A. 2000. Burnished haematite and pigment pro-
duction. In Ritchie, A. (ed.) Neolithic Orkney in its European
Context. Cambridge, McDonald Institute Monographs:
191–195.
Jacobi, R.M. 1976. Britain inside and outside Mesolithic
Europe. Proceedings of the Prehistoric Society 42: 67–84.
Jacobi, R.M. 1980. The Early Holocene settlements of Wales.
In Taylor J.A. (ed.) Culture and Environment in Prehistoric
Wales. Oxford, British Archaeological Reports British
Series 76: 131–206.
Jacobi, R.M. 1981. The last hunters in Hampshire. In
Shennan, S.J. and Schadla-Hall, R.T. (eds) The archaeology
of Hampshire from the Palaeolithic to the Industrial Revolution.
230
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Hampshire Field Club and Archaeol. Soc. Monograph 1:
10–25.
Jacobi, R.M. 1987. The lateglacial archaeology of Gough’s
Cave at Cheddar. In Collcutt, S.N. (ed.) Recent studies in
the Palaeolithic of Britain and its nearest neighbours. Sheffield,
J R Collis.
Jahren, A.H., Toth, N., Schick K., Clark J.D. and Amundson
R.G. 1997. Determining stone tool use: chemical and
morphological analyses of residues on experimentally
manufactured stone tools. Journal of Archaeological Science
24: 245–250.
Jenkinson, D. 1988. Soil Organic matter and its Dynamics. In
Wild, A. (ed.) Russell’s Soil Conditons and Plant growth.
London, Longmans: 565–607.
Jobey, G. and Newman, T.G. 1975. A Collared Urn cremation
on Howick Heugh, Northumberland. Archaeologia Aeliana
5th Ser. 3: 1–16.
Johansen, L. and Stapert, D. 1998. Dense flint scatters:
knapping or dumping? In Aktuelle Forschungen zum
Mesolithikum. (Urgeschichtliche Materialhefte 12). Tü-
bingen, Mo Vince Verlag: 29–41.
Johnson, G.A.L. 1995. Robson’s Geology of North-East
England. Transactions of the Natural History Society of
Northumbria 56: 5.
Jordan, P. 2001. Ideology, Material Culture and Khanty
Ritual Landscapes in Western Siberia. In Fewster, K.J.
and Zvelebil, M. (eds) Ethnoarchaeology and Hunter-
gatherers: Pictures at an Exhibition. Oxford, British Archaeo-
logical Reports International Series 955: 25–42.
Jordan, P. 2003. Material Culture and Sacred Landscape: The
Anthropology of the Siberian Khanty. California, AltaMira
Press.
Juel Jensen, H. 1986. Unretouched blades in the Late
Mesolithic of South Scandinavia. Oxford Journal of Archaeo-
logy 5 (1): 19–33.
Juel Jensen, H. 1994. Flint tools and plant working. Aarhus,
Aarhus University Press.
Kaliff, A., Carlsson, T., Molin, F. and Sundberg, K. 1997.
Mörby. Östergötlands äldsta boplats. Riksantikvarieämbetet,
Rapport UV Linköping 1997:38, Linköping.
Karsten, P. and Knarrström, B. 2001a. Tågerup – fifteen
hundred years of Mesolithic occupation in Western
Scania, Sweden: a preliminary view. Journal of European
Archaeology 4(2): 165–174.
Karsten, P. and Knarrström, B. (eds) 2001b. Tågerup –
specialstudier. (Skånska spår – arkeologi längs Västkustbanan).
Lund, Riksantikvarieämbetet, UV Syd.
Keeley, L.H. 1980. Experimental Determination of Stone Tool
Uses: a Microwear Analysis. Chicago, University of Chicago
Press.
Keeley, L.H. and Newcomer, M.H. 1977. Microwear analysis
of experimental flint tools: a test case. Journal of Archaeo-
logical Science 4: 29–62.
Keeney, G. S. 1939. A pagan Anglian cemetery at Howick.
Archaeologia Aeliana 4th Ser. 16: 120–128.
Keesing, R.M. 1982. Kwaio Religion: The Living and the Dead in
a Solomon Island Society. New York, Columbia University
Press.
Klitgaard-Kristensen, D., Sejrup, H.P., Haflidason, H.,
Johnsen, S. and Spurk, M. 1998. A regional 8200 cal. yr BP
cooling event in northwest Europe, induced by final
stages of the Laurentide ice-sheet deglaciation? Journal of
Quaternary Science 13 (2): 165–169.
Knight, J., Orford, J.D., Wilson, P., Wintle, A.G. and Braley, S.
1998. Facies, age and controls on recent coastal sand dune
evolution in North Norfolk, eastern England. Journal of
Coastal Research 26: 154–161.
Knutsson, K. 1988a. Making and using stone tools, the Analysis of
Lithic Assemblages from Middle Mesolithic Sites with Flint in
Västerbotten, Northern Sweden. Uppsala University, AUN 11.
Knutsson, K. 1988b. Patterns of tool use, scanning electron
microscopy of experimental quartz tools. Uppsala University,
AUN 11.
Knuttsson, K. 1990. A new lithic scene. The archaeological
context of used tools. In Graslund, B. (ed.) The Interpretative
Possibilities of Microwear Studies, Aun, 14, 15–30 (Proceedings
of the International conference on lithic use-wear analysis,
15th–17th February 1989. Uppsala, Sweden).
Lacaille, A.D. 1954. The Stone Age in Scotland. Oxford, Oxford
University Press.
Lage, W. 2004. Zur interpretation der lehmstraten in den
feuerstellen des Duvenseer Moores – lehmplatten als Gar-
und Röstvorrichtungen Wahrend des Mesolithikums in
Schleswig-Holstein. Archäologisches Korrespondenzblatt 34
(3): 293–302.
Larsson, L. 1989. Late Mesolithic settlement and cemeteries
at Skateholm, southern Sweden. In Bonsall, C. (ed.) The
Mesolithic in Europe: Papers Presented at the Third Inter-
national Symposium, Edinburgh 1985. Edinburgh, John
Donald Ltd: 367–378.
Larsson, M. 1996. Högby. Mesolitiska och sennneolitiska boplaster
vid Högby i Östergötland. Bosättningsmönster och materiell
kultur. Riksantikvarieämbetet. Rapport UV Linköping,
1996:35, Linköping.
Lass Jensen, O. 2001. Kongemose og Ertebøllekultur ved
den fossile Nivåfjord. In Lass Jensen, O., Sørensen, S.A.
and Møller Hansen, K. (eds) Danmarks jægerstenalder –
status og perspektiver. Beretning fra symposiet “Status og
perspektiver inden for dansk mesolitikum” afholdt i Vor-
dingborg, september 1998. Hørsholm, Hørsholm Egns
Museum: 115–129.
Lawson, J. 2001. A list of archaeological radiocarbon dates
(Cramond, Edinburgh). In Turner, R. (ed.) Discovery and
Excavation in Scotland new ser 2: 124.
Lee, R., and Devore, I. 1978. Kalahari Hunter-Gatherers: Studies
of the !Kung San and Their Neighbors. Harvard University
Press.
Lyman, R.L. 1994. Vertebrate Taphonomy. Cambridge, Cam-
bridge University Press.
MacKenzie, M.A. 1991. Androgynous Objects: String bags and
gender in central New Guinea. London, Harwood Academic
Publishers.
MacLaughlan, H. 1867. Notes not included in the Memoirs
already published on Roman Roads in Northumberland.
Newcastle Upon Tyne, Privately Printed.
Malm, T. 1995. Excavating submerged stone age sites in
Denmark – the Tybrind Vig example. In Fischer, A. (ed.)
Man and Sea in the Mesolithic. Coastal Settlement above and
below present sea level. Oxford, Oxbow Monograph 53.
Mason, S.L.R. 1996. Hazelnut (Corylus spp.) as a past food
resource? UCL, Institute of Archaeology, Bioarchaeology
Discussion Group, Wednesday 28th February 1996.
Mason, S.L.R. 2000. Fire and Mesolithic subsistence –
managing oaks for acorns in northwest Europe? Palaeo-
geography, Palaeoclimatology, Palaeoecology 164: 139–150.
McCarthy, A., Finlay, N. and McClean, O. 1999. Marine
molluscan remains. In Woodman, P.C., Anderson, E. and
Finlay, N. Excavations at Ferriter’s Cove, 1983–95: Last
Foragers, First Farmers in the Dingle Peninsula. Bray,
Wordwell: 93–102.
McCutcheon, P.T. 1992. Burned archaeological bone. In Stein,
J.K. (ed.) Deciphering a Shell Midden. New York, Academic
Press: 347–70.
Meehan, B. 1982. Shell Bed to Shell Midden. Canberra:
Australian Institute of Aboriginal Studies.
Mellars, P. (ed.) 1987. Excavations on Oronsay: Prehistoric
Human Ecology on a Small Island. Edinburgh, Edinburgh
University Press.
Mellars, P. 1998. Postscript: major issues in the interpretation
of Star Carr. In Mellars, P. and Dark, P. (eds) Star Carr in
Context: New Archaeological and Palaeoecological Invest-
igations at the Early Mesolithic site of Star Carr, North
Yorkshire. Cambridge, MacDonald Institute: 215–241.
Mercer, J. 1980. Laussa 1: the late-glacial and early post-
glacial occupation of Jura. Proceedings of the Society of
Antiquaries of Scotland 110: 1–32.
Milner, N., Craig, O.E, Bailey, G.N., Pedersen, K. and
Andersen, S.H. 2004. Something fishy in the Neolithic? A
re-evaluation of stable isotope analysis of Mesolithic and
Neolithic coastal populations. Antiquity 78: 9–22.
Mithen, S.J. (ed.) 2000. Hunter-Gatherer Landscape Archaeology:
The Southern Hebridean Mesolithic Project, 1988–1998.
Cambridge, McDonald Institute for Archaeological
Research.
Mithen, S.J. and Score, D. 2000. Experimental roasting of
hazelnuts. In Mithen, S.J. (ed.) Hunter-Gatherer Landscape
Archaeology: The Southern Hebrides Mesolithic Project 1988–
1998. Cambridge, MacDonald Institute: 507–512.
Mithen, S.J., Finley, F., Carruthers, W., Carter, S. and
Ashmore, P. 2001. Plant use in the Mesolithic: Evidence
from Staosnaig, Isle of Colonsay, Scotland. Journal of
Archaeological Science 28: 223–234.
Mook, W.G. 1986. Business meeting: Recommendations/
Resolutions adopted by the Twelfth International Radio-
carbon Conference. Radiocarbon 28: 799.
Moore, P.D., Webb, J.A. and Collinson, M.E. 1989. Pollen
Analysis. Oxford, Blackwell.
Moores, A.M. 1999. Holocene Environmental Change, human
activity and archaeological landscapes in the North Tyne Basin.
Unpublished PhD thesis, University of Newcastle upon
Tyne.
Moss, M.L. 1993. Shellfish, gender, and status on the
Northwest coast: reconciling archaeological, ethno-
graphic, and ethnohistorical records of the Tlingit.
American Anthropologist 95 (3): 631–652.
Mulholland, H. 1970. The microlithic industries of the Tweed
valley. Transactions of the Dumfriesshire and Galloway Natural
History and Antiquarian Society 47 (3rd Ser.): 81–110.
Mullan, G.J. and Wilson, L.J. 2004. A possible Mesolithic
engraving in Aveline’s Hole, Burrington Combe, north
Somerset. Proceedings of the University of Bristol Speaeo-
logical Society 23 (2): 75–85.
Murdock, G.P. 1967. Ethnographic Atlas. Pittsburgh, Uni-
versity of Pittsburgh Press.
Myers, A. M. 1987. All shot to pieces? Inter-assemblage
variability, lithic analysis and Mesolithic assemblage
‘types’; some preliminary observations. In Brown, A.
G.and Edmonds, M. R. (eds) Lithic Analysis and Later
British Prehistory. Some problems and approaches. Oxford,
BIBLIOGRAPHY 231
British Archaeological Reports British Series 162: 137–
153.
NERC. 1971. British Regional Geology; Northern England (forth
edition). HMSO.
Newcomer, M.H., Grace, R. and Unger-Hamilton, R. 1986.
Investigating microwear polishes with blind tests. Journal
of Archaeological Science 13: 203–217.
Newell, R.R. 1981. Mesolithic dwelling structures: fact and
fantasy. In Gramsch, B. (ed.) Mesolithikum in Europe. 2.
Internationales Symposium Potsdam, 3. bis 8. April 1978
Bericht. (Veröffentlichungen des Museums für Ur- und
Frühgeschichte Potsdam Band 14/15). Berlin, VEB
Deutscher Verlag der Wissenschaften: 235–284.
Nordqvist, B. 2000. Coastal Adaptation in the Mesolithic: A
Study of Coastal Sites With Organic Remains from the Boreal
and Atlantic Periods in Western Sweden. Göteborg, Depart-
ment of Archaeology and Ancient History, Göteborgs
University.
O’Malley, M. and Jacobi, R.M. 1978. The excavation of a
Mesolithic occupation site at Broom Hill, Braishfield,
Hampshire, 1971–1973. Rescue Archaeology in Hampshire 4:
16–38.
Orford, J.D., Wilson, P., Wintle, A.G., Knight, J. and Braley, S.
2000. Holocene coastal dune initiation in Northumber-
land and Norfolk, eastern UK: climate and sea-level
changes as possible forcing agents for dune initiation. In
Shennan, I. and Andrews, J. (eds) Holocene Land-Ocean
Interaction and Environmental Change around the North Sea.
London, Geological Society of London Special Pub-
lications 166: 197–217.
Oshibkina, S.V. 1985. The material culture of the Veretye-
type sites in the region to the east of Lake Onega. In
Bonsall, C. (ed.) The Mesolithic in Europe. Papers presented at
the Third International Symposium. Edinburgh, Edinburgh
University Press.
Owen, L. 2000. Lithic functional analysis as a means of
studying gender and material culture in prehistory. In
Donald, M. and Hurcombe, L. (eds) Gender and Material
Culture in Archaeological Perspective. London, MacMillan
Press.
Palmer, S. 1999. Culverwell Mesolithic Habitation Site, Isle of
Portland, Dorset, Excavation Report and Research Studies.
Oxford, British Archaeological Reports British Series
287.
Payne, S. and Munson, P.J. 1985. Ruby and how many
squirrels? The destruction of bones by dogs. In. Fieller,
N.R.J., Gilbertson, D.D. and Ralph, N.G.A (eds) Palaeo-
biological Investigations: Research Design, Methods and Data
Analysis. Oxford, British Archaeological Reports Inter-
national Series 266: 31–40.
Payton, R.W. and Palmer, R.C. 1990. Soils of the Alnwick and
Rothbury district (Sheet 81). Silsoe, Soil Survey and Land
Research Centre.
Petersson, M. 1951. Mikrolithen als Pfeilspitzen. Ein Fund
aus dem Lilla Loshult-Moor ksp. Loshult, Skåne. Kungliga
Humanistiska Vetenskapssamfundets i Lund Årsberättelse
IV(1950–1951): 123–137.
Price, T.D. 1978. The spatial analysis of lithic artifact
distribution and association on prehistoric occupation
floors. In Davis, D.D. (ed.) Lithics and Subsistence: The
Analysis of Stone Tool Use in Prehistoric Economies. (Vander-
bilt University Publications in Anthropology No. 20).
Nashville, Vanderbilt University: 1–33.
232
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Proudfoot, E, 1999. Fordhouse barrow, Dun. Discovery and
excavation in Scotland 1999, 111.
Proudfoot, E 2001. Fordhouse barrow, Angus. Discovery and
excavation in Scotland (new series) 2, 122.
Radley, J, Tallis, J H, and Switsur, V R, 1974 The excavation
of three “narrow blade” Mesolithic sites in the southern
Pennines, England. Proceedings of the Prehistoric Society 40,
1–19.
Raistrick, A. 1933. The distribution of Mesolithic sites in the
north of England. Yorkshire Archaeological Journal 31: 141–
156.
Raistrick, A. 1934. Mesolithic sites of the north-east coast of
England. Proceedings of the Prehistoric Society of East Anglia
7: 188–198.
Raistrick, A., Coupland, F. and Coupland, G. 1935. A
Mesolithic site on the south-east Durham coast. Tran-
sactions of the Northern Naturalists Union 1 (4): 207–216.
Raistrick, A. and Westoll, T.S. 1933. A prehistoric site on the
south Durham coast. Vasculum 19: 139–144.
Ragg, J.M. 1960. The soils of the country round Kelso and Lauder.
Edinburgh: HMSO.
Reynier, M.J. 2002 Kettlebury 103: a Mesolithic “Horsham”
type stone assemblage from Hankley Common. Elstead,
Surrey Archaeological Collections, 89, 211–231.
Reynier, M.J. 2005. Early Mesolithic Britain. Origins, Develop-
ment and Directions. Oxford, British Archaeological
Reports British Series 393.
Ripoll, S., F. Muñoz, F., Bahn, P., and Pettitt. P. 2004.
Palaeolithic cave engravings at Creswell Crags, England.
Proceedings of the Prehistoric Society 70: 93–105.
Robinson, J. and Biggins, J.A. 1999. Cushat Wood, Howick,
Northumberland; Survey of an Iron Age Settlement. Com-
missioned by Tyne and Wear Museums for the Howick
Estate, Unpublished Report.
Robins, R.P. and Stock, E.C. 1990. The burning question: a
study of molluscan remains from a midden on Moreton
Island. In Solomon, S., Davidson, I. and Watson, D. (eds)
Problem Solving in Taphonomy: Archaeological Studies from
Europe, Africa and Oceania. Queensland, Australia. Tempus,
Anthropology Museum, the University of Queensland:
Archaeology and Material Culture Studies in Anthro-
pology Vol. 2: 80–100.
Robson, D.A. 1966. A guide to the geology of North-
umberland and the Borders. Transactions of the Natural
History Society of Northumberland, Durham and Newcastle
upon Tyne XVI, No 1.
Rowell, T.K. and Turner, J. 1985. Litho-, Humic and Pollen
stratigraphy at Quick Moss, Northumberland. Journal of
Ecology 73: 11–25.
Rowley-Conwy, P. 1981. Mesolithic Danish bacon: per-
manent and temporary sites in the Danish Mesolithic. In
Sheridan, A. and Bailey, G. (eds) Economic Archaeology:
Towards an Integration of Ecological and Social Approaches.
Oxford, British Archaeological Reports International
Series 96: 51–55.
Rowley-Conwy, P. 1995. Wild or domestic? On the evidence
for the earliest domestic cattle and pigs in South Scand-
inavia and Iberia. International Journal of Osteoarchaeology
5: 115–126.
Russell, N.J., Bonsall, C. and Sutherland, D.G. 1995. The
exploitation of marine molluscs in the Mesolithic of
western Scotland: evidence from Ulva Cave, Inner
Hebrides. In Fischer, A. (ed.) Man and Sea in the Mesolithic:
Coastal Settlement above and below Present Sea Level. Oxford,
Oxbow: 273–288.
Sackett, J.R. 1990. Style and ethnicity in archaeology: the
case for isochretism. In Conkey, M.W. and Hastorf, C.A.
The Uses of Style in Archaeology. Cambridge, Cambridge
University Press: 32–43.
Sanchez, A., Canabate, M.L. and Lizcano, R. 1996. Phos-
phorus analysis at archaeological sites: An optimisation
of the method and interpretation of the results. Archaeo-
metry 38(1): 151–164.
Saville, A. 1980. On the measurement of struck flakes and
flake tools. Lithics, 1: 16–20.
Saville, A. 1997. Palaeolithic handaxes in Scotland. Pro-
ceedings of the Society of Antiquaries of Scotland 127: 1–16.
Saville, A. 2004. The material culture of Mesolithic Scotland.
In Saville, A. (ed.) Mesolithic Scotland and its neighbours.
The early Holocene prehistory of Scotland, and its British and
Irish context and some northern European perspectives.
Edinburgh, Society of Antiquaries of Scotland: 185–220.
Schofield, A. J. 1991 Artefact distributions as activity areas:
examples from south-east Hampshire. In Schofield, A.J.
(ed.) Interpreting Artefact Scatters: Contributions to Plough-
zone Archaeology. Oxford, Oxbow Monograph 5: 117–
128.
Schulting, R. J. and Richards, M. P. 2000. The use of stable
isotopes in studies of subsistence and seasonality in the
British Mesolithic. In Young, R. (ed.) Mesolithic Lifeways. Cur-
rent Research from Britain and Ireland. University of Leciester,
Leicester Archaeology Monographs No. 7: 55–65.
Schulting, R. and Richards, M.P. 2002. Finding the coasting
Mesolithic in southwest Britain: AMS dates and stable
isotope results on human remains from Caldey Island,
south Wales. Antiquity 76: 1011–1025.
Scott, E.M. (ed.) 2003. The Third International Radiocarbon
Intercomparison (TIRI) and the Fourth International
Radiocarbon Intercomparison (FIRI) 1990–2002: results,
analysis, and conclusions. Radiocarbon 45: 135–408.
Shennan, S. 1997. Quantifying Archaeology. (Second Edition).
Iowa City, University of Iowa Press.
Shennan, I., Lambeck, K., Horton, B., Innes, J., Lloyd, J.,
Mcarthur, J. and Rutherford, M. 2000. Holocene isostacy
and relative sea-level changes on the east coast of
England. In Shennan, I. and Andrews, J. (eds) Holocene
Land-Ocean Interaction and Environmental Change around
the North Sea. London, Geological Society of London
Special Publications 166: 275–298.
Shiel, R. and Hardy, K. 2003. Statistical analysis of use-wear
data. In Hardy, K. and Sillitoe, P.
Material Perspectives: Stone Tool Use and Material Culture
in Papua New Guinea.
Internet Archaeology 14. http://intarch.ac.uk/journal/issue14
/hardy_index.html.
Shiel, R. and Hardy, K. (In prep.) A statistical method for
interpreting use-wear analysis data on stone tools.
Shipman, P., Foster, G. and Schoeninger, M. 1984. Burnt bones
and teeth: an experimental study of color, morphology,
crystal structure and shrinkage. Journal of Archaeological
Science 11: 307–325.
Shore, J.S., Bartley, D.D. and Harkness, D.D. 1995. Problems
encountered with the 14C dating of peat. Quaternary
Science Reviews 14: 373–383.
Shotton, F.W., and Williams, R.E.G., 1973 Birmingham
University Radiocarbon Dates VII, Radiocarbon 15, 451–468.
Sillitoe, P. 1988. Made in Niugini. Technology in the highlands of
Papua New Guinea. London, British Museum Publications.
Sillitoe, P. and Hardy, K. 2003. Living Lithics. Ethnography
and Archaeology in Highland Papua New Guinea.
Antiquity 77: 296, 555–566.
Simmons, I.G. 1996. The Environmental Impact of Later Mesolithic
Cultures. Edinburgh, Edinburgh University Press.
Simmons, I.G. 2003. The Moorlands of England and Wales.
Edinburgh, Edinburgh University Press.
Skaarup J. 1995. Hunting the hunters and fishers of the
Mesolithic – twenty years of research on the sea floor
south of Funen, Denmark. In Fischer, A (ed.) Man and Sea
in the Mesolithic. Coastal Settlement above and below present
sea level. Oxford, Oxbow Monograph 53: 397–401.
Slota, Jr P.J., Jull, A.J.T., Linick, T.W. and Toolin, L.J. 1987.
Preparation of small samples for 14C accelerator targets
by catalytic reduction of CO. Radiocarbon 29: 303–306.
Smith, C. 1992. Late Stone Age Hunters of the British Isles.
London, Routledge.
Smith, D.E., Shi, S., Cullingford, R.A., Dawson, A.G.,
Dawson, S., Firth, C.R., Foster, I.D.L., Fretwell, P.T.,
Haggart, B.A., Holloway, L.K., and Long, D. 2004. The
Holocene Storegga Slide tsunami in the United Kingdom.
Quaternary Science Reviews 23: 2291–2321.
Soffer, O., Adovasio, J.M., Illingsworth, J.S., Amirkhanov,
H.A., Preaslov, N.D. and Street, M. 2000. Palaeolithic
perishables made permanent. Antiquity 74: 812–821.
Sørensen, S. A. 1996. Kongemosekulturen i Sydskandinavien.
Færgegården: Egnsmuseet Færgegården.
Spikins, P. 2002. Prehistoric Peoples of the Pennines:
Reconstructing the Lifestyles of Mesolithic Hunter-
Gatherers on Marsden Moor. Leeds, West Yorkshire
Archaeology Service on behalf of English Heritage.
Stapert, D. and Krist, J.S. 1990. The Hamburgian Site of
Oldeholtvolde (NL): Some Results of the Refitting
Analysis. In Cziesla, E. (ed.) The Big Puzzle: International
Symposium on Refitting Stone Artefacts. Studies in Modern
Archaeology 1. Bonn,Holos: 371–404.
Stapert, D., Krist, J.S. and Zandbergen, A.L. 1986. Oldeholt-
volde: A Late Hamburgian Site in the Netherlands. In
Roe, D.A. (ed.) Studies in the Upper Palaeolithic of Britain
and Northwest Europe.. Oxford, British Archaeological
Reports International Series 296: 187–226.
Stein, J.K. 1992. The analysis of shell middens. In Stein, J.K.
(ed.) Deciphering a Shell Midden. New York, Academic
Press:1–24.
Stenhouse, M.J. and Baxter, M.S. 1983. 14C dating repro-
ducibility: evidence from routine dating of archaeological
samples. PACT 8: 147–61.
Stone, P. and Planel, P.G. 1999. Introduction. In Stone, P. and
Planel, P.G. The Constructed Past. Experimental Archaeology,
Education and the Public. London, Routledge: 1–14.
Stuiver, M. and Kra, R.S. 1986. Editorial comment. Radio-
carbon 28(2B): ii.
Stuiver, M. and Polach, H.A. 1977. Reporting of 14C data.
Radiocarbon 19: 355–363.
Stuiver, M. and Reimer, P.J. 1986. A computer program for
radiocarbon age calculation. Radiocarbon 28: 1022–1030.
Stuiver, M. and Reimer, P.J. 1993. Extended 14C data base and
revised CALIB 3.0 14C age calibration program. Radio-
carbon 35: 215–230.
Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S.,
Hughen, K.A., Kromer, B., McCormac, F.G., van der
BIBLIOGRAPHY 233
Plicht, J. and Spurk, M. 1998. INTCAL98 radiocarbon age
calibration, 24,000–0 cal BP. Radiocarbon 40: 1041–1084.
Switsur, V.R. and Jacobi, R.M. 1979. A radiocarbon chron-
ology for the early postglacial stone industries of England
and Wales. In Berger, R. and Suess, H.E. (eds) Radiocarbon
Dating. London, University of California Press: 41–68.
Switsur, V. R., and West, R. G., 1973 University of Cambridge
Natural Radiocarbon Measurements XII. Radiocarbon, 15:
534–44.
Switsur, V R, and West, R G, 1975 University of Cambridge
Natural Radiocarbon Measurements XIII. Radiocarbon, 17:
35–51.
Teasdale, D. and Hughes, D. 1999. The glacial history of
North-East England. In Bridgland, D., Horton, B.P. and
Innes, J.B. (eds) The Quaternary of North-East England.
Quaternary Research Association Field Guide. 10–17.
Tipping, R. 1996. The Neolithic Landscapes of the Cheviot
Hills Hinterland: Paleoenvironmental Evidence. Northern
Archaeology 13/14: 17–33.
Tipping, R. 1998a. Towards an environmental History of the
Bowmont valley and the northern Cheviot Hills. Land-
scape History 20: 41–50.
Tipping, R. 1998b. The chronology of late Quaternary fluvial
activity in part of the Milfield basin, northeast England.
Earth Surface Processes and Landforms 23: 845–856.
Tolan-Smith, C. 1997a. The Stone Age landscape: the
contribution of fieldwalking. In Tolan-Smith, C. (ed.)
Landscape Archaeology in Tynedale. Newcastle upon Tyne,
Department of Archaeology, University of Newcastle
upon Tyne: 79–89.
Tolan-Smith, C. 1997b. Excavations at Brikside Fell Cairn,
1997. University of Durham and University of Newcastle upon
Tyne Archaeological Reports 1997 21: 11–12.
Unrath, G., Owen, L.R., van Gijn, A., Moss, E.H., Plisson, H.
and Vaughan, P. 1986. An evaluation of Use-Wear Studies:
a Multi-analyst Approach. Early Man News 9/10/11: 117–
176.
van der Veen, M. 1982. Carbonised plant remains from
Neolithic Thirlings (Northumberland). Ancient Mon-
uments Laboratory Report 3831: 4.
van Hoek, M. and Smith, C. 1988. Rock Carvings at Goats-
crag Rock Shelters, Northumberland. Archaeologia Aeliana
5th Ser 16: 29–35.
Vang Petersen, P. 1982. Jægerfolket på Vedbækbopladserne.
Nationalmuseets Arbejdsmark 1982: 179–189.
van Wijngaarden-Bakker, L.H. 1985. The faunal remains. In
Woodman, P.C. (ed.) Excavations at Mount Sandel 1973–77.
Belfast, Northern Ireland Archaeological Monographs
No. 2, HMSO: 71–76.
Vereschagin, N.K. 1967. Primitive hunters and Pleistocene
extinctions in the Soviet Union. In Martin, P.S. and
Wright, Jr. H.E. (eds) Pleistocene Extinctions: the Search for a
Cause. New Haven and London, Yale University Press:
365–398.
Waddington, C. 1996. Middle Warren. Fieldwalking Assessment
Phase II. Archaeological Practice, University of Newcastle
Upon Tyne, unpublished report.
Waddington, C. 1998a. Fieldwalking. In Turning the Tide
Archaeological Assessment. Archaeological Services Uni-
versity of Durham. Unpublished Report: 18–28.
Waddington, C. 1998b. Cup and ring marks in context.
Cambridge Archaeological Journal 8 (1): 29–54.
Waddington, C. 1999a. A Landscape Archaeological Study of the
234
MESOLITHIC SETTLEMENTS IN THE NORTH SEA BASIN: A CASE STUDY FROM HOWICK
Mesolithic-Neolithic in the Milfield Basin, Northumberland.
Oxford, British Archaeological Reports British Series
291.
Waddington, C. 1999b. Recent lithic finds from Bowden
Doors, Northumberland. Archaeologia Aeliana 5th Ser. 27:
173–174.
Waddington, C. 2001a Fieldwalking Results from Maiden’s Hall
Archaeological Evaluation. Unpublished Report for the
Archaeological Practice, University of Newcastle upon
Tyne.
Waddington, C. 2001b Milfield Fieldwalking and Test Pit
Results. Department of Archaeology, University of New-
castle upon Tyne. (unpublished).
Waddington, C. 2004. The Maelmin Heritage Trail. Archaeo-
logical research and the public. In Frodsham, P. (ed.)
Interpreting the Ambiguous. Archaeology and Interpretation in
Early 21st Century Britain. Oxford, British Archaeological
Reports British Series 362: 49–57.
Waddington, C., Bailey, G. Boomer, I. and Milner, N.J. 2006.
A Bronze Age Cist Cemetery at Howick, Northumber-
land. The Archaeological Journal 162: 65–95.
Waddington, C. and Passmore, D. (in press) Characterization
and interpretation of surface artefact scatters: a model of
inference. Geoarchaeology.
Waddington, C. and Pedersen, K.L.R. (eds) 2007. Mesolithic
Studies in the North Sea Basin and Beyond. Proceedings of a
Conference Held at Newcastle in 2003. Oxford, Oxbow.
Waraas, T. A. 2001. Vestlandet i tidleg Preboreal tid. Fosna,
Ahrensburg eller vestnorsk tidlegmesolitikum? Unpublished
MA dissertation. Bergen, Universitetet i Bergen.
Ward, G.K. and Wilson, S.R. 1978. Procedures for comparing
and combining radiocarbon age determinations: a crit-
ique. Archaeometry. 20: 19–31.
Ward, T., 1998 Daer Reservoir 1 and 2, Crawford. Discovery
and Excavation in Scotland 1998, 128.
Wayland Barber, E. 1994. Women’s Work: The first 20,000 years.
Women, Cloth and Society in Early Times. New York, W.W.
Norton and Co.
Westcott. D. (ed.) 1999. Primitive technology: A Book of Earth
Skills. Salt Lake City, Gobbs-Smith.
Weyman, J. 1975. Mesolithic occupation at Gallowhill Farm,
Corbridge. Archaeologia Aeliana 5th Ser. 3: 219–220.
Weyman, J. 1980. A flint chipping site at Low Shilford, Riding
Mill, Northumberland. Archaeologia Aeliana 5th Ser. 8: 159–
161.
Weyman, J. 1984. The Mesolithic in North-East England. In
Miket, R. and Burgess, C. (eds) Between And Beyond The
Walls: Essays in Honour of George Jobey. Edinburgh, John
Donald: 38–51.
Whallon, R. 1973. Spatial analysis of occupation floors I: the
application of nearest neighbor analysis of variance.
American Antiquity, 38: 266–278.
Whallon, R. 1974. Spatial analysis of occupation floors II: the
application of nearest neighbor analysis. American An-
tiquity, 39: 16–35.
Whallon, R. 1978. The Spatial Analysis of Mesolithic
Occupation Floors: A Reappraisal. In Mellars, P. A. (ed.)
Mesolithic Settlement in Northwest Europe. London, Duck-
worth: 27–35.
White, C. 1968. Report on Field Survey, June–August 1968.
Canberra: Australian Institute of Aboriginal studies,
Document 68/738 (mimeographed).
White, J.P. 1968. Ston naip bilong tumbuna: the living stone
age in New Guinea. In Bordes, F. and de Sonneville
Bordes, D. (eds) La préhistoire: problèmes et tendances. Paris,
CNRS: 511–516.
White, J.P. and Thomas, D.H. 1972. What mean these stones?
Ethno-taxonomic models and archaeological interpret-
ations in the New Guinea Highlands. In Clarke, D.L. (ed.)
Models in Archaeology. London, Methuen and Co: 275–308.
White, R. B. 1978 Excavation at Trwyn Du, Anglesey, 1974.
Archaeologia Cambrensis 127: 16–39.
Wickham-Jones, C.R. 1990. Rhum: Mesolithic and Later Sites at
Kinloch, Excavations 1984–86. Edinburgh, Society of
Antiquaries of Scotland Monograph Series 7.
Wickham-Jones, C. R. 2004. Structural evidence in the
Scottish Mesolithic. Mesolithic Scotland and its Neigh-
bours. In Saville, A. (ed.) The Early Holocene Prehistory of
Scotland, its British and Irish Context and Some Northern
European Perspectives. Edinburgh, Society of Antiquaries
of Scotland: 229–242.
Wickham-Jones, C.R. and Dalland, M. 1998. A small Meso-
lithic site at Fife Ness, Fife, Scotland. Internet Archaeology
5: 1–32. http://intarch.ac.uk/journal/issue5/wickham
_toc.html
Wiessner, P. 1983. Style and social information in Kalahari
San projectile points. American Antiquity 48 (2): 253–76.
Wiessner, P. 1990. Is There a Unity to Style? In Conkey, W.M.
and Hastorf, C.A. (eds) The Uses of Style in Archaeology.
Cambridge, Cambridge University Press: 105–112.
Woodman, P.C. 1978a. The Mesolithic in Ireland: hunter-
gatherers in an insular environment. Oxford, BAR British
Series 58.
Woodman, P.C. 1978b. The chronology and economy of the
Irish Mesolithic: some working hypotheses. In Mellars,
P.A. (ed.) The Early Postglacial Settlement of Northern Europe.
London, Duckworth: 333–369.
Woodman, P. C. 1984. Excavations of a Mesolithic Site at Cass-
Ny-Hawin, Isle of Man. Mesolithic Miscellany 5 (1): 2–4.
Woodman, P.C. 1985. Excavations at Mount Sandel 1973–1977.
Belfast, Northern Ireland Archaeological Monographs 2
HMSO.
Woodman, P.C. 1987. Excavations at Cass ny Hawin, a Manx
Mesolithic site, and the position of the Manx microlithic
industries. Proceedings of the Prehistoric Society 53: 1–22.
Woodman, P.C. 2004. Some problems and perspectives:
reviewing aspects of the Mesolithic period in Ireland. In
Saville, A. (ed.) Mesolithic Scotland and its Neighbours. The
Early Holocene Prehistory of Scotland, and its British and Irish
Context and Some Northern European Perspectives. Edin-
burgh, Society of Antiquaries of Scotland: 285–297.
Woodman, P.C., Anderson, E. and Finlay, N. 1999. Excavations
at Ferriter’s Cove, 1983–95: Last Foragers, First Farmers in the
Dingle Peninsula. Bray, Wordwell.
Woodman, P.C., McCarthy, M. and Monahan, N. 1997. The
Irish Quaternary Faunas Project, a survey of the 14C
evidence. Quaternary Science Reviews 16: 129–59.
Wymer, J. 1962. Excavations at the Maglemosian sites at
Thatcham, Berkshire, England. Proceedings of the Pre-
historic Society 28: 329–361.
Yellen, J. E. 1977. Cultural patterning in Faunal Remains:
Evidence from the !Kung Bushmen. In Ingersoll, D. and
Yellen, J.E. (eds) Experimental Archaeology. New York,
Columbia University Press: 271–331.
Young, R. 1987. Lithics and Subsistence in North-Eastern
England. Aspects of the Prehistoric Archaeology of the
Wear Valley, County Durham from the Mesolithic to the
Bronze Age. Oxford, British Archaeological Reports.
Young, R. 2000a. Aspects of the ‘coastal Mesolithic’ of the
north east of England. In Young, R. (ed.) Mesolithic
Lifeways. Current Research from Britain and Ireland. Uni-
versity of Leicester, Leicester Archaeology Monographs
7: 179–190.
Young, R.E. 2000b. Investigations at Howick. Archaeology in
Northumberland 1999–2000. Morpeth, Northumberland
County Council: 31.
Zagorska, I. and Zagorski, F. 1985. The Bone and Antler
Inventory from Zvejnieki II, Latvian SSR. In Bonsall, C.
BIBLIOGRAPHY 235
(ed.) The Mesolithic in Europe. Papers presented at the Third
International Symposium. Edinburgh, Edinburgh Uni-
versity Press..
Zvelebil, M. 1994. Plant use in the Mesolithic and its role in
the transition to farming. Proceedings of the Prehistoric
Society 60: 35–74.
Zvelebil, M., Green, S.W. and Macklin, M.G. 1992. Archaeo-
logical landscapes, lithic scatters, and human behaviour.
In Rossignol, J. and Wandsnider, L. (eds) Space, Time and
Archaeological Landscapes. New York, Plenum Press: 193–
226.