Challenging Preconceptions of the

European Iron Age: Essays in Honour

of Professor John Collis Wendy
Morrison (Editor)
Visit to download the full and correct content document:
https://ebookmass.com/product/challenging-preconceptions-of-the-european-iron-age
-essays-in-honour-of-professor-john-collis-wendy-morrison-editor/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Politics, Christianity and Society in Malawi: Essays in

Honour of John McCracken Kenneth R. Ross (Editor)

https://ebookmass.com/product/politics-christianity-and-society-
in-malawi-essays-in-honour-of-john-mccracken-kenneth-r-ross-
editor/

Classical Economics, Keynes and Money Essays in Honour

of Carlo Panico John Eatwell

https://ebookmass.com/product/classical-economics-keynes-and-
money-essays-in-honour-of-carlo-panico-john-eatwell/

Beyond the Fascist Century: Essays in Honour of Roger

Griffin Constantin Iordachi

https://ebookmass.com/product/beyond-the-fascist-century-essays-
in-honour-of-roger-griffin-constantin-iordachi/

Bombay Before Mumbai: Essays in Honour of Jim Masselos

Prashant Kidambi (Editor)

https://ebookmass.com/product/bombay-before-mumbai-essays-in-
honour-of-jim-masselos-prashant-kidambi-editor/
Managing Globalization in the Asian Century: Essays in
Honour of Prema-Chandra Athukorala Hal Hill (Editor)

https://ebookmass.com/product/managing-globalization-in-the-
asian-century-essays-in-honour-of-prema-chandra-athukorala-hal-
hill-editor/

Landscapes of Human Evolution: Contributions in Honour

of John Gowlett John Mcnabb (Editor)

https://ebookmass.com/product/landscapes-of-human-evolution-
contributions-in-honour-of-john-gowlett-john-mcnabb-editor/

Age of Iron: On Conservative Nationalism Colin Dueck

https://ebookmass.com/product/age-of-iron-on-conservative-
nationalism-colin-dueck/

Challenging Economic Journalism: Covering Business and

Politics in an Age of Uncertainty Henrik Müller

https://ebookmass.com/product/challenging-economic-journalism-
covering-business-and-politics-in-an-age-of-uncertainty-henrik-
muller/

Tenacious of Life : The Quadruped Essays of John James

Audubon and John Bachman John James Audubon

https://ebookmass.com/product/tenacious-of-life-the-quadruped-
essays-of-john-james-audubon-and-john-bachman-john-james-audubon/
Challenging Preconceptions
of the European Iron Age
Essays in Honour of Professor John Collis

Edited by

Wendy Morrison
Challenging Preconceptions
of the European Iron Age
Essays in Honour of
Professor John Collis

Edited by

Wendy Morrison

Archaeopress Archaeology
Archaeopress Publishing Ltd
Summertown Pavilion
18-24 Middle Way
Summertown
Oxford OX2 7LG

www.archaeopress.com

ISBN 978-1-80327-006-7
ISBN 978-1-80327-007-4 (e-Pdf)

© the individual authors and Archaeopress 2022

Cover: Extract from Pilbrow’s original illustration (1871, Plate XXIII) of Roman Antiquities from Canterbury.
Hillfort at Berber Hill, Kenn, Devon. Photograph F.M. Griffith, Devon County Council, 29.06.1984.

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic,
mechanical, photocopying or otherwise, without the prior written permission of the copyright owners.

This book is available direct from Archaeopress or from our website www.archaeopress.com
 Contents

Preface������������������������������������������������������������������������������������������������������������������������������������������������������������������ iii

The Atlas of Hillforts of Britain and Ireland: a consideration of the coastal and inland promontory forts and
enclosures of Scotland�������������������������������������������������������������������������������������������������������������������������������������������1
Stratford Halliday and Ian Ralston

A long, largely aceramic, period of Devon’s prehistory������������������������������������������������������������������������������������� 21

Henrietta Quinnell

Deconstructing archaeological databases���������������������������������������������������������������������������������������������������������� 29

Martin Kuna

The Gauls against the State ��������������������������������������������������������������������������������������������������������������������������������� 41

Sophie Krausz

The European Iron Age. John Collis (1984). London: Batsford. a late review������������������������������������������������������� 49
Chris Gosden

Exploring the origins and character of transhumance in England������������������������������������������������������������������� 54

Andrew Fleming

Since John left Devon: some unanticipated outcomes of aerial reconnaissance in the county����������������������� 70
F.M. Griffith and E.M. Wilkes

Mam Tor, Derbyshire: new plans outlining hill and fort, internal platforms and all��������������������������������������� 79
Graeme Guilbert

A rich Late Iron Age burial from Canterbury���������������������������������������������������������������������������������������������������� 127

Timothy Champion

Some reflections on phenomenology, structure, agency and actancy in medieval pottery studies�������������� 136
C. G. Cumberpatch

‘Friendly Hills by Nature Guarded Round’: Recent work at Bathampton Down, Bath������������������������������������������ 143
Lisa Brown

Mapping Celticity����������������������������������������������������������������������������������������������������������������������������������������������� 155

Olivier Buchsenschutz

i
ii
 Preface

Challenging Preconceptions of the European Iron Age is a collection of essays by some of the leading researchers in
the archaeology of the European Iron Age, paying tribute to Professor John Collis. Since the 1960s, John has been
involved in investigating and enriching our understanding of Iron Age society, and crucially, questioning the status
quo of our narratives about the past. He has influenced generations of students and peers alike and has been
one the strongest voices in the demystification of the ‘Celtic’ world. John has never held back on his continual
questioning of the past, and he has been instrumental in helping us unpick the labyrinthine tapestry of myth
and misinterpretation, interrogating the way later prehistory has been traditionally investigated, packaged, and
presented, to reveal a much more interesting past, filled with more nuance, possibility, and humanity than the
uniformly structured Iron Age of the early 20th century narrative.

The idea for this volume was born nearly a decade ago in conversation with my then doctoral supervisor who had
been a student of John’s (and one of the contributors to this volume) Chris Gosden. The list of contributors is not an
exhaustive cohort of those who would wish to honour John, nor of those to whom he has been a profound guide to
their careers and research – indeed that would stretch to many volumes This volume brings together papers from
more than a dozen of Professor Collis’s colleagues and students to mark his 75th birthday. The contributions range
across later prehistory and the European continent, taking in major themes that have been his prime interests -
hillforts, how we use archaeological data, socio-political structures of the past, and of course, the ‘Celts’.

The volume begins with Strat Halliday and Ian Ralston’s treatment of Scottish promontory forts. Drawing on the
wealth of data from the recently published Hillfort Atlas Project,1 they offer a comprehensive treatment of over 500
promontory and coastal sites. They highlight the challenges facing any attempt to classify such a diverse feature
type but the challenges are addressed neatly and convincingly. Through careful and rigorous analysis of their data,
Halliday and Ralston offer a compelling challenge to previous received scholarship on the siting of these features.
Their conclusion invites an exploration of the wider Atlas dataset,2 freely accessible to all, so that multivocality
of interpretations will enrich our growing understanding. In the best tradition of Collis, they are continually
questioning the status quo, backed up by good evidence and a dedication to making that evidence, and the new
interpretations drawn from them, available to wider audiences.

In a similar vein, the second paper by Henrietta Quinnell also calls for a ‘reworking’ of prior scholarship based
on the potential of not only new excavations but also re-interrogating older assemblages in the study of Iron Age
Devon. Opening with a wonderful reminiscence of her first professional interactions with Collis, Quinnell quickly
sets the stage for the current thinking about ceramics as chronological markers in archaeology, and the thorny
issue of dealing with largely aceramic communities. She makes the excellent point that many groups around the
globe have done rather well without ceramics for large swathes of time, and there is no reason to suppose that a
paucity of a ceramic assemblage will undermine our understanding of peoples who used different materials for food
and drink preparation and consumption.

In an age where we are indeed swamped with new excavations and the ability to revisit archived investigations as
well, it is fitting to ask some questions about the massive datasets – the embarrassment of riches – we have access
to. Martin Kuna explores some of the challenges this presents us with regard to such variables as archaeological
visibility, drawing on recent work he has done on the prodigious evidence from Bohemia. Mindful of the reliability
of some quantification models, he makes a clear case for why this big data is an important resource, but also why
we should be cautious in the conclusions we draw from it. The article is a refreshing antidote to some approaches
to ‘Big Data’, which advocate that if there is enough data, we can assume the robust will outweigh the unreliable.

Riffing on the title of Pierre Clastres’ anthropological work The Society against the State, Sophie Krausz looks at
prehistoric and proto-historic Gaul and the way we have thought about their socio-political systems. Challenging
the still-dominant narratives of a Mediterranean-imposed shift in organisation, Krausz invites us to think about
things from a point of view that returns more agency of choice to the Gauls, rather than seeing the formation of a
Gaulish state as a natural step in the linear progression of societal development. Krausz’s assessment of the stages
and influences of urban development in the Iron Age is particularly significance against this refreshing backdrop.

1
Lock, G. and Ralston, I.B., 2022. Atlas of the Hillforts of Britain and Ireland. Edinburgh University Press.
2
https://hillforts.arch.ox.ac.uk

iii
John’s interest in non-linear forms of complex progression in political and urban development is also referenced
in Chris Gosden’s contribution. In what must be the latest review of a publication to be written, Gosden looks at
John’s unrivalled publication The European Iron Age, giving it not only its due in being the only major comprehensive
attempt at such a geographical and temporal span, but also offering critiques along the way, as any honest review
should. That some of the omissions of John’s book were the product of those themes only recently coming into
scholarly vogue can hardly be seen as a true criticism; Gosden has written a fitting tribute to a book which has
influence generations of archaeologists and pointed out how we can build upon it.

Andrew Fleming provides the volume with a call to consider persistence of a tradition often taken for granted in
some periods and regions of Britain, and totally overlooked in others. The practice of transhumance, or the seasonal
cycle of moving livestock from one grazing region to another, may, as Fleming argues, have much deeper and wider
roots on the island. He shows through a range of evidence sources, both scientific and observational, that this sort
of movement of people and animals in a landscape needs to be factored into our prehistoric narratives and suggests
that some of the patterns of land modification and earthworks building we see in later prehistory may be indicative
of transhumant activity. Fleming’s images of modern transhumance in action are particularly compelling.

The call to look more closely at patterns in the landscape is reiterated in the chapter by Frances Griffiths and
Eileen Wilkes. Acknowledging John’s early interest but subsequent distraction from the prehistory of Devon, the
two authors report on four decades of archaeological prospection that have, by looking at the ‘big picture’ of a
wide landscape, produced a more densely exploited and settled prehistoric landscape. Drawing on reconnaissance
from above and below, both aerial photographic campaigns flown over decades, and geophysical surveys, Wilkes
and Griffiths show that not only is their dedication to the archaeology of Devon aiding our understanding of the
past, but also preserving it for th e future; they outline how relationships with landowners and farmers are key, and
how discovering new archaeological features enables farmers to get support to look after them – a collaboration
between researchers and the public that is to be commended and which should be emulated more widely.

The lengthiest chapter in the book represent the first in depth treatment of the plan of Mam Tor, a nationally
significant hillfort in the Peak District. Graeme Guilbert has been researching Mam Tor for decades, since his first
introduction to the site by John Collis in the late seventies, making this contribution a fitting example of the sort
of inspiration John’s enthusiasm and passion can lead to in those with whom he comes into contact. Guilbert’s plan
is a first, recording comprehensively the myriad features of the site, whilst evaluating what previous attempts
have observed. The plans, particularly Figure 11, in this chapter will be beneficial for future researchers of the site.
Guilbert also draws our attention to the fact that whilst Mam Tor is on the outer limits of just how high up a hillfort
settlement can be placed and thrive, our modern perceptions of what constitutes hardship and comfort (and the
ability/luxury to have such considerations) may bear little relation to the lived experiences of the hardy people who
chose to modify this landmark hilltop.

The macro scale of one of the larger hillforts in Britain if followed by the micro scale of a single assemblage. Tim
Champion presents new thoughts on what may be the earliest of the south-eastern Late iron Age ‘richly furnished’
burials. Discovered in the 19th century, and frustratingly reliant on descriptions of that period due to loss of many
of the artefacts, Champion nonetheless created a vibrant reconstruction and analysis of the collection of grave
goods, and crucially, sets them in their significant context of the development of pre-Roman Canterbury.

Staying with the artefactual focus, Chris Cumberpatch writes an insightful paper on the study of medieval pottery,
a topic he states that admittedly may seem at odds with a volume honouring someone who has been the voice of
the European Iron Age. Inspired by his early research with John’s pottery assemblages from Iron Age Auvergne,
Cumberpatch takes us on an exploration of what use may be made of the theoretical approaches more commonly
used in prehistoric archaeology when applied to a medieval dataset. It is a fine example of how John’s influence is
far ranging across areas of study.

In the penultimate chapter, Lisa Brown provides a closer look at a hillfort that has seen much attention, but little
examination. Made famous by peter Gabriel’s pop song, Little Solsbury Hill outside of Bath has long relied on the
scant evidence from earlier 20th century small investigations and the generation of ceramic and faunal data cast up
from animals. Brown puts these finds in the context of new work conducted by the Bath and Counties Archaeological
Society and offers a clearer understanding the nature of the site in relation to the early days of pre-Roman Bath.

iv
Finally, the volume concludes with Oliver
Buchsenschutz’s critical assessment of archaeological
mapping of ‘The Celts’. Fittingly drawing on John’s
body of work and acknowledging the pitfalls and
potentials we have available with such multiple data
sources as placenames and linguistics, Buchsenschutz
offers a way through the labyrinth using data mapping
to get the most out of overlapping strands of evidence.
His admonition to beware the great swathes of time we
are dealing with, encompassed in a single period is a
relevant one, and he rightfully calls for mapping to be
reflective of this.

Together these papers offer readers a cross section

of many of the areas John has contributed to and
influenced over his decades of dedication leadership in
European prehistory. I am grateful to the authors in this
volume for their enthusiastic support of the idea, their
ready willingness to contribute an article, and crucially,
their patience with an editor who was not as alacritous
as envisaged. It is a pleasure and privilege to serve
as editor for this volume in honour of Professor John
Collis. John has been a major influence in the field of
prehistoric archaeology and for me personally has been
a source of stimulating conversation and challenging
debate. Whilst we have not always agreed on everything,
I have never come away from a conversation with John
without understanding the principles and practices of
our discipline better.

I am sure that both contributors and readers alike will

agree that no work like this can hope to capture all the
facets of our subject with which John has engaged, nor
will all the conclusions reached in the chapters jibe
100% with how John might interpret things. Yet the
strength of our discipline is that we remain flexible
to new concepts, always questioning, and that is a
philosophy John has always embodied – challenging John Collis on a visit to Titterstone Clee, Shropshire, during a meeting of
preconceptions. the Hillfort Study Group in April 2009; photographed by Daryl Garton.

v
vi
 The Atlas of Hillforts of Britain and Ireland:
a consideration of the coastal and inland promontory forts and
enclosures of Scotland
Stratford Halliday and Ian Ralston
Introduction
In this paper, which we are delighted to offer to John
Collis, we wish to take stock of a particular subset of the
sites included in the Atlas of Hillforts of Britain and Ireland,
a project funded by the AHRC 2012-16 and published
online in 2017 (Lock and Ralston 20171; https://hillforts.
arch.ox.ac.uk/). The subset we will focus on comprises
the Scottish promontory forts and enclosures, which
make up almost 30% of all the entries for Scotland.
Some of this material will be incorporated in the Atlas
volume to be handed over to the publisher in 2021, but
the promontory sites of Scotland highlight many of
the issues involved in the wider study of enclosed sites
in Britain and Ireland, and the discussion of the sites
themselves is worth airing in greater detail. On the one
hand they expose such prosaic but important issues
as classification and chronology; on the other they
extend to more testing topics such as issues of function
for sites which can occur in highly exposed locations
where routine inhabitation may always have offered
considerable challenges.
At the national scale Scottish promontory forts have
received relatively little attention in recent years.
Some of the more unusual examples, such as the coastal
sites of Burgi Geos in Shetland [4169] and Burghead
in Moray [2925], figure in Harding’s overview (2012)
where he rightly stresses that – as is the case for other
suites of hillforts – topographic position is not a guide
to their chronology (2012, 288). Coastal promontory
forts are also a conspicuous component of Toolis’s
studies of Galloway (2003; 2007; 2015), but the most
recent attempt at a national distribution map, albeit
schematic, was that produced by Raymond Lamb
(1980,5,fig. 1). In common with all these studies, Lamb
focused exclusively on the coastal examples.
In several places in this study we have included
the promontory forts of Northumberland in the
commentary, because in many ways the later prehistoric
1
Individual sites in the database can be accessed by their name and
four-digit numbers, here provided in square brackets. The database
entries include site bibliographies generally not quoted directly in
this paper.
Challenging Preconceptions of the European Iron Age
(Archaeopress 2022): 1–20
enclosed sites of that county – in both numbers and
characteristics – much more closely parallel the world
to the north of them than that of the counties to their
south, where forts of all kinds are altogether rarer.
Thus, Northumberland is the southern flank of a region
that we here term ‘Greater Tyne-Forth’, which extends
eastwards from the North Sea coast of East Lothian and
Berwickshire to include Dumfriesshire and southern
Lanarkshire.2 Atlas data demonstrate that the counties
of this region are in the top quartile in terms of site
numbers per unit area, whereas those to the south are
in the lowest quartile (Lock and Ralston 2020, fig. 4);
this is the starkest such gradient in the British Isles.
In Scotland alone, a total of 502 promontory enclosures
have been included in the Atlas database (see Fig. 1 and
Map 1), and there are a further 19 in Northumberland;3
another 29 fortified enclosures occupy promontory
positions in topographical terms, but these enclosures
belong to other categories and are not considered here.
The Scottish entries are divided between 301 distributed
around the lengthy hard-rock coastline and 201 which
occupy inland positions. Of the overall total, 399 have
been ‘Confirmed’ as forts according to the criteria
agreed at the outset of the project (Lock and Ralston
2017 and 2020; see Halliday 2019 for a full discussion of
the criteria with respect to Scotland): 236 of these are
coastal and 163 inland. These 399 sites represent 41% of
all the Confirmed entries for promontory forts in the
Atlas database for the whole of Britain and Ireland. While
these represent the focus of the paper, it is important to
also explore the character of the additional 97 that are
deemed ‘Unconfirmed’, and the further six marked as
having ‘Irreconciled Issues’. In what follows, these are
discussed first, the issues that they raise serving to both
sharpen and qualify the definition of those Confirmed.
The assessment of the latter then follows, with further
sections on the character of their defences and internal
features and the related series of scarp-edge forts.
2
For this study, we have retained the pre-1974 counties as the most
convenient and best-known geographical divisions. The regions we
have used include Greater Tyne-Forth – by which we designate all of
south-east Scotland plus Northumberland.
3
Northumberland includes 12 Confirmed, 4 Unconfirmed and 3
Irreconciled Issues sites.
 Stratford Halliday and Ian Ralston

Map 1. A maximising view pf the distribution of inland (N= 201) and coastal (N = 301) promontory forts in Scotland drawing on Atlas data (N = 502).

2
 The Atlas of Hillforts of Britain and Ireland
Defining the Dataset
The criteria against which all the entries in the Atlas
have been judged rest on three key characteristics,
namely physical advantages of the topographical
position, scale of the enclosing works, and size of the
interior, for which a lower limit of 0.2ha was set (Lock
and Ralston 2020). Applied rigorously, however, it was
foreseen that numerous well-known Scottish forts
would be omitted, and the decision was taken to operate
these qualifications to include enclosures that met at
least two out of the three criteria. This had far reaching
consequences, leading to the inclusion of numerous
small fortified enclosures which otherwise would not
have qualified. What was not foreseen, however, is that
these same criteria also embrace the morphological
character of a series of other promontory enclosures
that have never been conventionally considered as
forts, be that prehistoric or early medieval. To some
extent this may be a reflection of the extensive use of
Scottish promontories for various purposes at different
times, though shades of the same issues are clearly
present in Ireland, but we felt it was better to err on the
side of caution by including this wider range of ancient
promontory enclosures, rather than omitting them and
denying the choice of further exploring these sites to
the user of the Atlas data.
The overall total of 502 entries thus represents all the
promontory works in Scotland that appear to be of
some antiquity, the principal exceptions being most
medieval stone castles exploiting promontory positions.
The emphasis here is placed on the word ‘most’,
because several castles clearly occupy earlier Iron
Age promontory forts, such as Cullykhan, Banffshire
[2982], and Dundarg Castle, Aberdeenshire [2983]. In
other cases, notably at Greenan Castle, Ayrshire [1289],
and Innerwick Castle, East Lothian [3928], cropmarks
have revealed ditch-systems that appear eccentric to
the visible medieval defences and are thus considered
more likely to relate to a different and earlier phase of
activity. This contrasts with several castles where early
medieval texts indicate the likely site of a promontory
fort, but no evidence of any earlier defences is visible
and none has been recovered by small-scale evaluation;
Dunnottar Castle [3112] on the Kincardineshire coast,
examined by the Alcocks (1992), is the prime example,
but others might include the rocky outcrop of Dunaverty,
Argyll [4309], according to the Annals of Iona besieged
in AD 712 (Alcock 1981, 157), and the cliff-girt Castle
Rock of Edinburgh Castle [3713], where again a siege
is noted for the year AD 638. Excavation on Edinburgh
Rock has at least revealed evidence of Late Bronze Age
and early medieval occupation, though any trace of
earlier defences is entirely lost beneath the castle walls.
On balance, Edinburgh Castle has been accepted as a
Confirmed fort, but in the case of Dunnottar Castle there
3
remains a nagging doubt, which led Alcock to consider
the neighbouring and unlikely promontory of Bowduns
as a possible alternative. And yet the impregnable
character of the conglomerate promontory beneath
Dunnottar Castle, rivalled only on this stretch of coast
by the tiny sea stack, presumably once a more extensive
promontory, at nearby Dunnicaer [3111], renders it
almost inconceivable that it was not the location of the
sieges referred to in entries in the Annals of Ulster for the
years AD 680 and 693 (Alcock 1981, 171-2; Alcock and
Alcock 1992, 267). More recently evidence of a timber-
laced defence dating from the 3rd-4th centuries AD has
been recovered from Dunnicaer (Past Horizons 2015),
but it would be surprising if the Dunnottar promontory
had escaped occupation in the pre-Roman Iron Age;
this, after all, is the premier location for a promontory
fort anywhere along this coast. The inclusion of several
other castles in the Unconfirmed category is more
speculative, such as at Borve Castle, Sutherland [2809],
and the Castle of Old Wick, Caithness [2827], where the
character of both the promontories and the defences
re-creates the form of other promontory sites identified
as forts along the coastline of northern Scotland (cf
Lamb 1980, 90-2). Another probable castle included
under this morphological pretext is the earthwork
overlooking the bridge over the River Esk at Gilknockie,
Dumfriesshire [1122], and, for completeness, what
is probably a late medieval stronghold known as
Lady Lindsay’s Castle, Perthshire [3044]. Castle Qua,
Lanarkshire [1568], had already been included in the
Atlas database when fieldwork recognised traces of the
abutments for a bridge crossing the ditch; its status was
altered to Irreconcilable Issues.
This theme of medieval works occupying promontories
that may be the sites of earlier promontory enclosures
also runs through the Irish data in the Atlas, for
example at Ballyoughteragh South, Kerry [1259], or
Downmacpatrick, Cork [0901], this latter where a 15th-
century tower and curtain wall cut off a promontory
identified as an earlier tribal stronghold. Reviewing
the descriptions of all these Irish sites against satellite
imagery, however, also suggests that the ‘promontory
fort’ classification has been employed generously in
both the Archaeological Survey of Ireland Database of the
National Monuments Service and the Northern Ireland
Sites and Monuments Record to include a number of
sites that are more likely to be post-medieval stock or
agricultural boundaries. A 17ha enclosure at Greenane
on Bear Island, Cork [0851], serves as one example,
while the 58ha enclosure on Innishark, Galway [1104],
representing about a quarter of the surface area of
this small island, is another. These Irish examples
are marked Unconfirmed in the Atlas database, but
in Scotland a conscious decision was taken to try to
exclude the majority of such works, which are found
widely on coastal headlands large and small, notably in
 Stratford Halliday and Ian Ralston
Shetland and the Outer Hebrides. In some cases they
have also previously been mistakenly identified as
fortifications and have been included in the Atlas only
to avoid confusion. Weinnia Ness, Shetland [4176], for
example, was first reported as a fort by Raymond Lamb
but later discounted from his synthesis (1980), and yet
it still appears in Canmore – the online National Record
of the Historic Environment of what is now Historic
Environment Scotland – as a fort. Another ‘fort’
recorded at Lambigart, Shetland [4172], was included
on a recent distribution map (Halliday and Ralston
2009, 466, fig 5), but a field visit revealed that the thick
banks cutting off the promontory are no more than
old dykes built of turf stripped from the neck. Likewise
visits to a series of large promontory enclosures noted
by coastal surveys in the Outer Hebrides showed that
the majority were almost certainly agricultural, in
some cases with a 19th-century drystone dyke roughly
replicating the line of an earlier bank flanked by turf-
stripping scars (Halliday and Ralston 2013, 224-5). In
most cases, the relationship of these boundary banks
to the topography at the neck of a promontory is quite
unlike that of defensive works, which almost invariably
adopt the cliffs and scarps on the seaward or outer
side of the neck to enhance the barrier. This sort of
promontory enclosure is not entirely restricted to the
Atlantic coast, and those at Strath Howe, Aberdeenshire
[2984], and Elliot Water, Angus [3080], are possibly no
more than post-medieval field boundaries.
Another group of sites that are more conventionally
considered to be early medieval is a series of what are
thought to be undocumented monastic sites, both large
and small, mainly situated on coastal promontories and
isolated sea stacks. In Shetland these include relatively
large enclosures, notably at Blue Mull, Unst [3723; 8.2ha],
but also including Outer [4195; 0.74ha] and Inner [4196,
4.4ha] Brough, Fetlar, and Brei Holm, Papa Stour [4197;
0.6ha], and the smaller stacks of Clett [4175], Burri
Stacks, Culswick [4177], Kame of Isbister [4182], Aastack,
Yell [4198], and Birrier of West Sandwick, Yell [4189].
The presence of small clustered rectangular structures
on many such sites, best illustrated by Kame of Isbister
and Birrier of West Sandwick, forms the basis for the
monastic interpretation, but it is difficult to demonstrate
conclusively that any of them is not an early medieval
secular settlement, nor, if they are indeed monastic, that
their origins did not lie in secular defended enclosures.
All other forts in the Northern Isles exploit precipitous
promontories and stacks, so it is not unreasonable
to suggest that some of these enclosures that have
been claimed as monastic were initially secular. The
inaccessible Birrier of West Sandwick, for example, has a
wall overlooking the razorback neck and though usually
identified as monastic has been marked Confirmed in
the database. Conversely the promontory enclosure on
Burrier Head [4174], also on Shetland, which was first
4
identified as a possible promontory fort by Raymond
Lamb (1980, 83) but appears to have a row of three
rectangular structures midway along its top, might just
as easily be monastic and has been marked Unconfirmed.
The Landberg promontory fort on Fair Isle [2861] raises
many of these issues, for the rectangular building visible
within it (Hunter 1996, 89-93) was demonstrated by
excavation to be a chapel and there was evidence of
an earlier phase of occupation beneath it from which
pottery comparable to that from brochs was recovered,
along with moulds for copper alloy artefacts. An
evaluation carried out on Brei Holm, Papa Stour [4197],
recovered evidence of a complex sequence of occupation
and two radiocarbon assays returned dates in the 5th-
7th centuries AD, rather earlier than the date suggested
by the rectangular buildings on its summit.
This problem of disentangling secular and monastic
enclosures is by no means unique to Shetland. The
Brough of Deerness on Mainland, Orkney [2840], for
example, with its chapel and rectangular and bow-
sided structures scattered across its interior, has long
been regarded as a monastic site, but the supposed
vallum monasterii is constructed as a major stone-
faced rampart defending the now collapsed neck of
the promontory, with its outer face set on the steep
landward slope. Its status as a fort has therefore been
marked as Confirmed. The Brough of Deerness is a
relatively large enclosure of almost 1ha, but at the
opposite end of the scale there are several smaller
and heavily eroded fortified promontories on Orkney
with traces of rectangular structures on them, such
as The Brough, also on Mainland [2841], and Castle
of Burwick on South Ronaldsay [2813]. The offshore
stack known as The Brough of Burgh Head on Stronsay
[2844] is another with a stout wall along its landward
flank but where there may be some doubt about its
original function. The northern coast of Caithness
and Sutherland includes further examples, on the one
hand with the enclosure of 2.25ha on St John’s Point,
Caithness [2833], traditionally associated with the
remains of a burial-ground and a chapel dedicated to
St John, and on the other two minor promontories
characterised by spectacular cliffs and the narrowest of
razor-backed necks connecting them to the mainland:
the first, Aodann Mhor, Sutherland [2782], is crowded
with small rectangular structures, and less-certain
traces of similar remains are also visible on the second,
An Tornaidh Bhuidhe, Sutherland [2790]. Though
unusual in this part of Scotland, St John’s Point meets
all the criteria for a major fortification with a massive
rampart and ditch, so much so that it is impossible to
deny its Confirmed status. The other two are much
more problematic, and though An Tornaidh Bhuidhe
has a bank facing onto the only access and has been
accepted also as Confirmed, in truth there is no way of
knowing whether either is secular or monastic.
 The Atlas of Hillforts of Britain and Ireland
Aspects of the same problem re-surface in eastern and
south-eastern Scotland, though these examples have
all been accepted as forts, such as the Kirk Hill on St
Abb’s Head, Berwickshire [4150]. Whether or not this
is the site of the documented monastery of St Aebbe,
the place-name Colodaesburg implies a fortified
place and Alcock’s excavations uncovered a complex
early medieval rampart sequence (Alcock et al. 1986).
Elsewhere on this coast at Auldhame, East Lothian
[3921], excavations on a promontory enclosure first
identified from cropmarks have uncovered a chapel and
a long cist cemetery. This has been interpreted as the
remains of another Anglo-Saxon monastery, but there
is nothing inherently in the character of the ditch,
which was not bottomed during the recent excavations
and is not closely dated, to demonstrate that its original
construction was as a monastic enclosure (Crone et al.
2016, 129). Nor is it alone in the association of long cist
burials with a promontory enclosure. In 1831 a long cist
cemetery was discovered not far away at Castle Dykes,
Dunglass Dean, Berwickshire [0486], and also occupies
the site of a multivallate promontory fort, while
other cists were said to have been found before 1853
on the neighbouring promontory fort at Castle Dykes,
Bilsdean, East Lothian [0487]. Yet another promontory
enclosure where cists occur is at Whiting Ness, Angus
[3100], which was traditionally the site of a burial-
ground and a chapel dedicated to St Ninian and marked
as such by the Ordnance Survey.
Another group of Unconfirmed entries in the Atlas
database relating to promontories and sea stacks worth
drawing attention to is in the Outer Hebrides. In their
case, the scope for confusion is not related to the use
of such locations by early monastic communities, but
rather as post-medieval strongholds. The small and
now inaccessible Stac Dhomnuill Chaim at Mangursta,
Lewis [2759], is thought to be the remains of a refuge
constructed in the early 17th century by the Uig
warrior, Donald Cam Macaulay, while Dun Eistean
[2772] on the east coast of Lewis is traditionally a
stronghold associated with the Morrisons since the
16th century (Burgess 2008, 60-2). The excavations
by the Barrowmans at Dun Eistean found no evidence
of occupation before the medieval period, but the
possibility that this is also the site of an earlier fort
cannot be discounted.
All the same issues arise at Dun Eorradail [2773], another
large and inaccessible stack north of Dun Eistean, where
there are traces of at least ten rectangular buildings
and the possibilities for its use range from fort to early
monastic community or post-medieval stronghold.
On reflection, the same could be said of several other
examples that have been accepted as Confirmed forts.
Rudha Shilldinish near Stornoway on Lewis [2765], for
example, carries a suite of large rectangular buildings
which suggest a medieval or later date rather than a
5
prehistoric fort at least for the present configuration
of the site. On Geirum Mor [2484], a cliff-bound islet
of 1ha in the sound between Mingulay and Berneray
at the extreme southern end of the chain of islands,
a wall blocks the only access at the north-east end
and there is a series of rectangular buildings on its
summit. Morphologically this is a fort, but whether
prehistoric, early medieval or post-medieval cannot be
determined without excavation, and it could yet prove
to be monastic. Biruaslum [2483], a tidal islet off the
west coast of Vatersay, which extends to 9.8ha – the
second largest such site in the Outer Hebrides – poses
a similar concern. It is defended by a thick wall facing
the main island; its position recalls the monastic site
on the Brough of Birsay in Orkney, though at 18ha the
latter island is rather larger and there is no evidence of
a perimeter wall overlooking the tidal isthmus linking
it to the mainland.
Another problem found along the Atlantic coasts relates
to the interpretation of the outworks of brochs and
duns, or Atlantic Roundhouses as they are commonly
called, themselves excluded from consideration in the
Atlas database on size grounds. The very term outwork
assigns primacy to the Atlantic Roundhouse, but in
some cases the supposed outworks may have been the
primary fortifications, with the Atlantic Roundhouse
subsequently set within them. This is a sequence
familiar in southern Scotland at Edinshall [4069] and
Torwoodlee [3542], and the excavator of the broch at
Crosskirk, Caithness [4348], Horace Fairhurst, postulated
that the broch there succeeded an earlier promontory
fort (1984). No stratigraphic evidence was advanced
in support of his case, which largely rests on two
radiocarbon dates obtained in the 1970s that purport to
predate the assumed chronology of the broch supplied
by sherds of Samian ware and late Roman pottery
(see discussion by MacKie 2007, 407-26). More recent
excavations at Nybster, Caithness [2820], were unable
to establish the stratigraphic relationship between the
wall across the neck of the promontory and the broch,
and radiocarbon dates for samples underlying the wall
indicate a phase of Late Bronze Age activity rather
than the date of the outwork itself. Four out of nine
Unconfirmed promontory works in Argyll fall into this
category,4 while for Sutherland and Caithness the figure
is 6 out of 16,5 in Orkney 6 out of 10,6 and in Shetland 3
out of 17.7 This uncertainty of the relationship between
Atlantic Roundhouses and their supposed outworks is
4
Dun Bhronaig [2444], Dun Haunn [2503], Dun Aorain [2546] and Dun
Chruban [2550]
5
An Dun, Clachtoll [2793], Altanduin [2806], Poll Gorm [2810],
Scarfskerry [2816], Nybster [2820] and Crosskirk [4348]
6
Weems Castle [2811], Yesnaby, Broch of Borwick [2845], Midhowe
[2846], Riggin of Kami [2847], Lamb Head [2848] and Broch of Burrian
[2849].
7
[Broch of Aithsetter [4187], Noss Sound [4190] and Sna Broch [4260].
 Stratford Halliday and Ian Ralston
not limited to those on coastal promontories and recurs
amongst those in other locations inland.
The rest of the sites that make up the Unconfirmed
promontory enclosures break down into several types
of record. Twelve are long-recorded sites that have
been so heavily degraded that there is insufficient
information to judge their true character with any
confidence.8 Twelve others turn on the interpretation
of cropmarks, either because a ditch is relatively
narrow for a defensive work or the definition of the
cropmarks is too diffuse;9 and finally some are simply
miscellaneous earthworks on promontories, most of
them of uncertain date or purpose. Most spectacular
of these is the multivallate earthwork isolating 54ha on
the Mull of Galloway [0201], which despite excavation
(Strachan 2000) remains undated; it is either the
largest fort in any setting in the whole of Scotland or
an extraordinary enclosure with some other function
that finds its only morphological parallels with
several equally large promontory works in southern
Ireland, such as the 83.5ha enclosure on the headland
at Ballynacarriga, Cork [1970]. Others falling into this
category in southern Scotland are inconsequential by
comparison. Haly Jo, Lumsdaine, Berwickshire [4092], for
example, a slight enclosure on a coastal cliff is probably
no more than a small settlement, though others, such
as Drummoral, Wigtownshire [0223], appear defensive,
here with rock-cut ditches but no evidence of an
accompanying rampart. Elsewhere there are: the tiny
stone-walled enclosure (claimed as unfinished: MacKie
2016) overlying the broch at Leckie, Stirlingshire [1471];
the enclosure beneath a later medieval burial-ground
on Innis Bhuidhe, an island set in the river at Killin,
Perthshire [2609]; the slight boundary, undated despite
limited excavation, enclosing two rectangular buildings
on a low promontory projecting into Loch Kinord,
Aberdeenshire [3075]; the precipitous promontory
known as Tronach Castle, Banffshire [2944], where
there are no visible defences; and Dun Earn, Morayshire
[2918], where a ditch 4m-5m in breadth but with little
trace of a rampart cuts off about 2.5ha on an inland
promontory.
In some cases a visit would solve issues raised by the
existing records, as for example at Lambigart discussed
above, or Hynish, Tiree, Argyll [2486], where fieldwork
(by SH) since the completion of the database clearly
demonstrates that this is not a promontory fort as such,
though there are traces of a fortified enclosure taking
8
Grennan, Grennan Point [0180], Killantrae Bridge [0217], Port o’
Warren [0311], Gunnerton Crag Camps [0520], Ebb’s Nook [0920],
Salter’s Nick [1977], The Heugh [2038], Machrihanish [2222], Keir,
Easter Tarr [2617], Firbush Point [2619], Coldstream [4079] and Siccar
Point [4115]
9
Clanyard Bay [0196], Leffnoll [0342], Loch Quien [1152], Rousland
[1838], Wester Tullynedie [3046], West Lindsaylands [3230], Milton
Mill [3782], Bara [3859], Nether Hailes [3883], Lumsdaine Dean [4098],
Coveyheugh [4101] and Ayton [4142].
6
in the top of the headland. The character of several
others could be resolved likewise, probably including
four relegated to the status of Irreconciled Issues.
Three of these, Court Hill, High Skeog, Wigtownshire
[0227], and An Fang, Craignish Point [2449] and Creag
a’ Chaisteal, Stillaig [2472], Argyll, are where the
observations recorded by the archaeologists who first
visited the sites have been disputed by subsequent
OS surveyors, while the fourth is a site reported in
1993 on the shores of Loch an Iasgaich, Skye [2743],
which appears an unlikely candidate to be a fort on
the grounds of either its topographical position or the
slightness of the supposed defences. In other cases, the
only resolution is by invasive evaluation, though as the
experience of the Mull of Galloway proved, there is no
guarantee of success. Nevertheless, clarifying the date
of the Mull of Galloway is evidently vitally important
for the interpretation of the Iron Age landscape of
south-western Scotland, and the same might be said
locally of Dun Evan in Morayshire if this proved to be a
fortification of Iron Age date.
Confirmed Promontory Forts
This rehearsal of the range of promontory works
found in Scotland and the problems of applying
morphological classifications to identify those that
are at least potentially prehistoric and early medieval
fortifications serves to clarify the definitions that
lie behind those that are attributed the status of
Confirmed forts. In short, they comprise fortified
enclosures where thick walls or earthen ramparts and
ditches bar access on one side, usually the easiest line
of approach, and the rest of the perimeter is apparently
defined by little more than cliff-edges or steep scarps.
In this definition the character of most of these forts
is synonymous with the description of the topography
on which they stand. Originating from its Latin root
to describe a raised headland jutting out into the sea,
it has been adapted more generally to describe other
projecting landforms and raised spits of ground, and
in archaeological terminology further extended to
embrace inland enclosures set on interfluves and thus
often exploiting angles formed in escarpments along
streams and rivers, usually where a tributary has cut
down at its confluence with the main flow. They are
thus largely defined by natural declivities on at least
two sides, often creating a roughly triangular plan in
which the artificial defences form the third side.
This basic format, however, has also led to the
term being applied to the plans of forts that are in
hilltop positions, or on the ends of ridges, where
the defences were apparently constructed only on
one side, complemented by abrupt or at least steep
descents elsewhere. Scottish examples of these tend
to be located in prominent elevated positions, such as
Dumglow, Kinross [3203], Ben Effrey, Perthshire [2648],
 The Atlas of Hillforts of Britain and Ireland
Craik Moor, Roxburghshire [3453], An Sgurr on the
island of Eigg [2527] or Sithean Buidhe, Argyll [2292].
All told, there are only thirteen in this type of setting
in Scotland and they are better analysed with other
hilltop forts.10 Much more problematical in this sort of
morphological classification driven by topographical
descriptors is the exclusion of scarp-edge forts where
the circuit was evidently left incomplete along one side
and the interior simply backs onto the lip of a cliff or
escarpment, and where the distinction between these
and some promontory forts is one of degree. This
issue is discussed further below under the heading of
Scarp-edge Forts. At least 97 other Confirmed forts
share this feature, eight of them in coastal locations,
and their entries in the Atlas database are variously
labelled Contour Fort (40), Hillslope Fort (12), and
Level Terrain Fort (45). While these terms serve as
topographic descriptors, it is unwise to apply any of
them too prescriptively in terms of their archaeological
significance.
Whereas Lamb’s schematic map suggested that
Scottish promontory forts formed discrete
concentrations in Shetland, Orkney, Caithness, Angus
and Galloway (1980, 5, fig 1), the Atlas data reveal
that they are much more widespread and occur
along virtually every coastline where there is some
form of normally rocky escarpment or cliff-line.
Furthermore, this same defensive format is equally
widespread in inland locations. Nevertheless, the
new map hides some general trends. Reference to
the regional and county table (Fig. 1) shows that the
proportions of promontory forts to other forts alter
from region to region. The largest single regional
group of promontory forts is in the western and
northern Highlands, which includes the whole of the
mainland Atlantic coast from Argyll northwards, and
the Inner Hebrides and Outer Hebrides. Across this
region as a whole 159 of the 370 Confirmed forts are
promontory forts, representing 43%, and 32 of them
occur inland. The pattern within this region, however,
varies widely. In the three northernmost counties of
Ross-shire, Sutherland and Caithness, where there
are relatively few forts, the percentages rise to 59%,
64% and 69% respectively. Locally within them, some
of the figures are even higher. In the Outer Hebrides,
for example, which were formerly split between Ross-
shire and Inverness-shire, of the 26 Confirmed forts,
all bar 3 are coastal promontories or tidal islands
(88%); the exceptions are all islands in inland lochs. In
Argyll, too, where the overall number of promontory
forts forms 38% of all Confirmed forts there, no fewer
than 50 of the 68 promontory forts occur on islands
10
The full list comprises Dumyat [1593], Sithean Buidhe [2292], An
Sgurr [2527], Skirley Craig [2633], Ben Effrey [2648], Dun Vallerain
[2715], Phoineas Hill [2887], Dumglow [3203], Little Trowpenny [3378],
Craik Moor [3453], Earn’s Heugh NW [4094], St Abb’s Head [4150] and
An Dun, Cornhill Wood [4381].
7
in the Inner Hebrides. This is perhaps a reflection of
the longer hard coastlines in relation to the surface
areas of the islands and probably aspects of their
geology too. Islay (about 600 sq km and 165 km of
coast), for example, has 38 forts all told, of which no
fewer than 26 are promontory works (68%), and 24
of them coastal, whereas Mull (about 880 sq km and
with 250 km of coast) has equivalent figures of 22, 10
(45%) and 10. Skye (1650 sq km and 650 km of coast),
the northernmost of the Inner Hebrides and formerly
part of Inverness-shire, has figures of 31, 15 (48%) and
12 respectively. It is worth noting in this context that
the 9 Confirmed forts on Orkney, and 15 on Shetland,
are all promontory forts and are all coastal.
Other regions show similar patterns of variation within
them. Thus, while promontory forts make up 34 % of all
forts in the South-West, in Wigtownshire, again with its
relatively long coastline, the figure rises to 65%. The lowest
regional percentage is in the Greater Tyne-Forth region,
where it is no more than 11%, and here exceptionally 87
of the 98 promontory forts occur inland. In the hillier
inland landscapes of Peeblesshire, where forts typically
stand on spurs along the sides of valleys, the figure is
only 1%, representing a single fort at Castlecraig [3636].
In Roxburghshire and neighbouring Northumberland
only 7% of Confirmed forts are on promontories, but
for Berwickshire, with its long predominantly rocky
coastline, it rises to 25%. Without the 8 coastal examples
known there, however, the percentage would be no more
than 18%, a figure more akin to the 17% in Dumfriesshire,
or 15% in Lanarkshire.
In passing, it is worth taking brief note of the
distribution of Unconfirmed promontory works, which
are also recorded in Fig. 1. At a regional scale, these
form between 17% of all promontory works in Greater
Tyne-Forth, but only 2% in comparison to the total
number of forts in the region. The equivalent figures
for the South-West are 13% and 5% respectively, in
Central Scotland 17% and 4%, and in the North-East
15% and 7%. In the northern and western Highlands,
these figures appear to rise overall to 20% and 10%
respectively, but this masks wide regional differences
between Argyll at 14% and 6%, and Caithness at 53%
and 76%. The Northern Isles calculations are even more
extreme, with 54% and 116%. A significant proportion
of these are the supposed outworks of brochs (supra). In
themselves, these figures are of little significance, but
they confirm the general pattern that the proportions
of promontory forts to other forts not only increases
on the islands of the Atlantic seaboard, but also more
generally northwards through the mainland, and
that whereas the alteration of the status of a few of
the Unconfirmed category in the south makes little
difference to the overall proportions of promontory
forts to other types, in the far north it exaggerates the
contrast already observed still further.
 Stratford Halliday and Ian Ralston

Figure 1. Table of promontory enclosures by region and county. Percentages of Confirmed Promontory Forts are calculated as a ratio of All
Confirmed Forts regionally and locally. Sizes column totals include multiple measurements from individual sites; for six others there is no
data.

Region % Conf
All Prom Unconf Irrec Conf All Conf
Prom Coastal Inland Sizes
Historic County Encls Proms Issues Proms Forts
Forts
Greater Tyne-Forth 100 12 1 87 667 13% 11 76 103
Berwickshire 36 7 - 29 117 25% 8 21 35
Dumfriesshire 22 1 - 21 116 18% - 21 24
East Lothian 17 3 - 14 81 17% 3 11 19
Lanarkshire 10 1 1 8 54 15% - 8 8
Midlothian 4 - - 4 41 10% - 4 4
Peeblesshire 1 - - 1 91 1% - 1 1
Roxburghshire 10 - - 10 148 7% - 10 12
Selkirkshire - - - - 19 - - - -
SW Scotland 68 8 1 59 172 34% 49 10 61
Ayrshire 9 - - 9 32 28% 4 5 9
Kirkcudbrightshire 12 1 - 11 80 14% 8 3 11
Wigtownshire 47 7 1 39 60 65% 37 2 41
Central 42 7 - 35 164 21% 5 30 42
Clackmannanshire 1 - - 1 1 100% - 1 1
Dunbartonshire - - - - 4 - - - -
Fife 10 - - 10 39 26% 4 6 13
Kinross-shire 1 - - 1 3 33% - 1 1
Perthshire 21 5 - 16 75 21% - 16 20
Renfrewshire - - - - 8 - - - -
Stirlingshire 7 1 - 6 25 24% - 6 6
West Lothian 2 1 - 1 9 11% 1 - 1
NE Scotland 40 6 - 34 83 41% 20 14 42
Aberdeenshire 7 2 - 5 23 22% 3 2 5
Angus 15 1 - 14 36 39% 8 6 15
Banffshire 7 1 - 6 8 75% 5 1 9
Kincardineshire 8 1 - 7 9 78% 3 4 7
Morayshire 3 1 - 2 7 29% 1 1 6
N & W Highlands 200 37 3 160 371 43% 127 33 159
Argyllshire 79 9 2 68 179 38% 58 10 69
Buteshire (Arran & Bute) 11 2 - 9 19 47% 3 6 9
Caithness 19 10 - 9 13 69% 8 1 9
Inverness-shire 50 7 1 42 102 41% 31 11 40
Ross-shire 24 3 - 21 35 60% 18 3 22
Nairnshire - - - - 6 - - - -
Sutherland 17 6 - 11 17 65% 9 2 10
Northern Isles 52 27 1 24 24 100% 24 - 23
Orkney 19 10 - 9 9 100% 9 - 10
Shetland 33 17 1 15 15 100% 15 - 13
Totals 502 97 6 399 1481 27% 236 163 430

8
 The Atlas of Hillforts of Britain and Ireland

The 399 that are Confirmed include 12 coastal and 17
inland forts with eccentric or wide-spaced lines of
ramparts, taken here to represent separate structural
arrangements probably belonging to successive phases
of construction. In another six cases no size data are
available. This provides a total of 430 measured internal
sizes, which are plotted on a single graph (Fig. 2) in
increments of 0.1ha up to 2ha, and 2ha increments
thereafter, the latter in orange. This shows that almost
half (48%) fall below 0.2ha. The profile of the graph falls
away steeply from 119 (28%) examples below 0.1ha to
about 0.7ha, sites below this threshold representing
87% of all promontory forts. In all, 24 lie between 1ha
and 2ha, and 13 are over 2ha, the largest being two
coastal forts on Vatersay and Mingulay in the Outer
Hebrides enclosing 9.8ha and 10.4ha respectively [2483,
2480].
By breaking this graph (Fig. 2) down into the two
datasets, however, different patterns can be detected
between those in coastal and inland positions. The
profile of the coastal graph of 247 sizes from 236 forts
(Fig. 3) falls away much more steeply to about 0.4ha,
and there are almost twice as many sizes below 0.1ha as
there are in the 0.1ha-0.2ha increment. A total of 185 fall
in the four increment classes below 0.4ha, representing

Figure 2. The Confirmed promontory forts of Scotland by enclosed area. Steps are 0.1 ha to 2ha, thereafter by 2ha divisions. N=430.

9
 Stratford Halliday and Ian Ralston

Figure 3. The Confirmed coastal promontory forts of Scotland by enclosed area. Steps are 0.1 ha to 2ha, thereafter by 2ha divisions.
N=247 from 236 sites.

Figure 4. The Confirmed inland promontory forts of Scotland by enclosed area. Steps are 0.1 ha to 2ha, thereafter by 2ha divisions.
N=183 from 163 sites.

10
 The Atlas of Hillforts of Britain and Ireland
75% of this group. In contrast, the inland graph (Fig.
4) of 183 sizes from 163 forts does not show this peak
in the smallest size band (i.e. below 0.1ha), and falls
steadily from a high point at 0.1ha-0.2ha down to 0.7ha,
beyond which a thin scatter extends to four larger
enclosures over 2ha in extent, the largest of them being
the hilltop fort of Sithean Buidhe, Argyll [2292], at 7ha.
The 168 falling below 0.7ha however represent 92%
of this group (compared with 83% for the equivalent
range in the coastal group).
Two main areas of difference can be identified in these
patterns, the first relating to the incidence of very small
examples, and the second to the forts between 0.2ha
and 0.4ha. In respect to the first, some 56% of coastal
examples (Fig. 3) fall below 0.2ha, compared with 37%
of inland examples. The significance of this observation
is uncertain. Possibly the high incidence of diminutive
coastal sites is little more than an impact of differential
erosion. The exposed Atlantic coasts, for example, are
undoubtedly subjected to far more extreme erosion
than the majority of inland promontories, but the
incidence of small inland examples perhaps indicates
that erosion has not been the unique determining factor
in the overall pattern of relatively small interiors.
The second difference is more clearly defined, and it
is that a greater proportion of the slightly larger band
of promontory forts between 0.2ha and 0.5ha occur
in inland locations. The relative percentages are 21%
of the coastal group, and 43% of the inland group.
Unsurprisingly, 48 out of the 78 forts in this band in the
latter come from Greater Tyne-Forth, and the general
profile of the inland graph is consistent with the wider
pattern of internal sizes in this region.
Character of Promontory Defence Works
Of the 236 Confirmed coastal promontory forts, at least
246 separate defensive schemes can be identified; size
data are also available for all bar four. These schemes
can be broken down into 158 univallate defences (64%),
50 bivallate (20%) and 38 cases of multivallation (16 %),
based on the maximum number of ramparts occurring
in a single sector of the defences. These percentages
can be compared with the overall figures for Confirmed
Scottish forts of 37% univallate, 37% bivallate and
26% multivallate. Evidently the figure of 64% for
promontory forts is well above the general proportion
of univallate forts in Scotland, while the combined
bivallate and multivallate percentage of 36% is well
below the national trend of 63% for all Scottish forts.
The pattern of vallation is displayed against size on
the graph (Fig. 5), which also shows the ghost of the
coastal promontory fort sizes. What may be skewing
the percentages is revealed in the pattern of sizes, since
11
no fewer than 88 of the very smallest forts under 0.2ha
have only a single line of defence, be it a stone wall or
a rampart and ditch. Even allowing for heavy erosion
of promontories, it would appear that the function of
the majority of these small enclosures could be serviced
by a relatively modest investment in artificial works.
Nevertheless, there are also 31 bivallate examples below
0.2ha, and their graph displays a similar profile, while 19
multivallate works occur in the size bands below 0.2ha.
The profile of bivallate examples in Fig. 5 is altogether
flatter, petering out with works below 1ha. Above this
size the only possible bivallate promontory forts are
Meall Lamalum on Colonsay, Argyll [2162], though
with its wide-spaced walls this could also be treated as
two successive univallate works respectively enclosing
1.3ha and 1.6ha (cf Yesnaby, Brough of Bigging, Orkney
[2837]), and Sumburgh Head in Shetland [4184], where
the records of the character of the defences are not
particularly satisfactory.
The larger multivallate works are less contentious. They
comprise Dun a’ Bheirgh (1.2ha) on the rocky Rudha na
Berie [2762], a storm-lashed promontory on the north-west
coast of Lewis in the Outer Hebrides, and the remarkable
Burghead, Moray [2925], where at its maximum extent
the substantially-demolished outer ramparts cut off
some 3.2ha on this sandstone headland. With these
exceptions, the rest of the larger Confirmed promontory
forts are univallate, 14 falling between 1ha and 2ha, these
including the outermost rampart across Isle Head on the
Isle of Whithorn, Wigtownshire [0226], the earliest, outer,
defences at Cullykhan, Banffshire [2982], both walls across
Yesnaby, Brough of Bigging, Orkney [2837] and Mas na
Buaile, Sutherland [2784]. In the Outer Hebrides univallate
sites include a’ Bheirghe, Port of Ness [2771] on Lewis, and
the defended stack of Geirum Mor [2484] in the sound
between Mingulay and Berneray, as well as a series of
coastal promontories on Islay in Argyll, namely Dun Mor
Ghil on the Oa [2071], and along the north-west coast of
the Rhinns, four such promontory forts: Beinn a’ Chaisteal
[2128], Allt nan Ba [2120], Am Burg Coul [2067], and Lossit
Point, Dun na Faing [2062].
There are another seven examples of coastal
promontories over 2ha to consider. In Shetland, Hog
Sound [4191], where the interior has been severed from
its multivallate defences by marine erosion to form
an island still of 2ha; on the north coast of mainland
Scotland, St John’s Point, Caithness [2833], and a rocky
hammerhead in excess of 4ha on Eilean nan Caorach,
Sutherland [2779]; on the Rhinns of Islay, Dun Bheolain
[2117]; and in the Outer Hebrides the spectacular
hammerhead of Gob a’ Chuthail on Lewis [2761], the
9.8ha promontory of Biruaslum [2483] on the west coast
of Vatersay (supra), and the precipitous 10.4ha Dun
Mhiughlaigh on Mingulay [2480], where a short wall
little more than 20m in length at the neck comprises
 Stratford Halliday and Ian Ralston

Figure 5. Univallate, bivallate and multivallate coastal promontory forts mapped against the ghost of the total population of Confirmed
coastal promontory forts in Scotland

the sole artificial defence of an otherwise inaccessible
promontory with cliffs up to 145m high jutting into the
Atlantic (Halliday and Ralston 2013, fig. 3). It should also
be remembered that the 54ha enclosure on the Mull
of Galloway [0201] is multivallate. This last apart, it is
striking how, with relatively few exceptions, these large
coastal promontory forts are found in the North-West
and far North, many of them on islands, and many in
places that are spectacular in their own rights, fringed
by sheer cliffs and looking out across open seascapes.
The positioning of some of the smaller works is certainly
no less spectacular in terms of the adjacent cliff
scenery, irrespective of the character of the artificial
defences. Dun Athad, on the Oa peninsula of Islay
[1877], which was evidently occupied in the medieval
period but where a single piece of vitrified stone has
been recovered from the wall above the narrow and
challenging neck, soars sheer above the foreshore about
100m below. Unsurprisingly, the choice of materials for
the construction of the defences of these sites is largely
driven by the adjacent geology. Ditches are relatively
uncommon along the Atlantic coasts from Kintyre to
Cape Wrath, and elsewhere are mainly found where
there is a covering of till or the rock is relatively soft
and easily quarried. Thus 35 out of 50 promontory forts
in Wigtownshire and Kirkcudbrightshire are equipped
with ditches, whereas in Argyll it is only 1 out of 68. Here
the majority of defences comprise walls, some of which
have evidently been timber-laced and burnt, resulting
in vitrifaction. There is no particular pattern to the sizes
of the forts where vitrifaction occurs, and the examples
qualifying as promontory forts include the diminutive
enclosures of no more than 0.07ha defended by three
walls at Dun nan Gall [2169] and Trudernish Point
[2170] on Islay, and the 0.09ha site of An Dun, on the
mainland at Gairloch [2727]. In north-eastern Scotland,
the early medieval timber-framed – and very locally
vitrified – rampart at Green Castle, Portknockie [2945],
cuts off 0.28ha, but the larger timber-laced ramparts

12
 The Atlas of Hillforts of Britain and Ireland
at Burghead [2925] and Cullykhan [2982] are likely to
have been designed to form complete enclosures that
happen to exploit the natural strength of promontories
and should perhaps be discounted from this reckoning.
As already observed, the general size graph for inland
promontory forts (Fig. 6) is rather different to that for
coastal promontories, but the pattern for univallate
enclosures is similar. This displays a peak of 17 examples
below 0.1ha that upsets the general trend of the inland
graph, which shows a gentler gradient declining from
the 0.1ha-0.2ha increment down to 0.8ha. Bivallate
enclosures, however, climb from the smallest size
category to the biggest cohort of 18 at 0.1ha-0.2ha,
descending relatively evenly to three at 0.5ha-0.6ha
before a spike of eight at 0.6ha-0.7ha, while the
multivallate systems peak with twelve cases in the 0.2ha-
0.3ha size band before descending to only one at 0.6ha-
0.7ha. Notably there is a dearth of univallate systems
enclosing between 0.8ha and 2ha, the exception being a
cropmark recorded at Hatchednize, Berwickshire [4077],
but three out of the four sites over 2ha are univallate –
White Isle, Dumfriesshire [0302], An Sgurr, Eigg [2527],
and Sithean Buidhe, Argyll – the latter two being
conspicuous hilltop enclosures. The fourth site over 2ha
in extent is the multivallate Double Dykes, Sodom Hill,
Lanarkshire [0841], set between the Cander Water and
Avon Water valleys. The seven other promontory works
with more than one rampart and enclosing between
0.8ha and 2ha are mainly cropmarks scattered across
south-east and north-east Scotland, but one, unusually
for a cropmarked site, is at Bridgend, Islay, Argyll [2155].
This set also includes the multivallate fort of 1.6ha on the
summit of Dumglow, Kinross [3203], the commanding
northward-facing summit of 379m OD towards the
western end of the Cleish Hills.
Figure 6. Univallate, bivallate and multivallate inland promontory forts mapped against the ghost of the total population of Confirmed inland
promontory forts (cf Fig. 4) in Scotland
13
Internal Features
Of the 236 Confirmed coastal promontory forts, 148 are
entirely featureless, and of the rest it is reasonable to
suppose that many of the structures that are visible, as
often as not rectangular rather than circular, probably
relate to later occupations. Some of the rectangular
buildings recorded on coastal promontory forts are
substantial structures, at least four being the remains
of churches or chapels – Brough of Deerness, Orkney
[2840], St John’s Point, Caithness [2833], Landberg, Fair
Isle [2861], and St Abb’s Head, Berwickshire [4150].
Others include buildings associated with later castles –
Castle Feather, Wigtownshire [0232], Dun Athad, Islay,
Argyll [1877], and possibly Rudha Shilldinish, Lewis
[2765] and Dun na Muirgheidh, Mull, Argyll [2511]. The
Brough of Deerness is also latterly a monastic site with
numerous traces of rectangular buildings within the
interior, and monastic use possibly accounts too for the
rows of buildings visible on Birrier of West Sandwick,
Yell, Shetland [4189], Castle of Burwick on South
Ronaldsay, Orkney [2813], and An Tornaidh Bhuidhe,
Sutherland [2790], which have all been accepted for the
Atlas as Confirmed promontory forts. The problems of
separating the remains of secular fortifications from
monastic sites occupying precipitous locations that
were almost impregnable before the addition of any
artificial works has already been aired.
Elsewhere, some of the other rectangular buildings
visible within the defences on coastal and inland
promontory forts are probably the remains of medieval
or post-medieval settlements, though without
excavation such remains cannot be dated with any
certainty. In Greater Tyne-Forth examples are Ogle
Linn, Dumfriesshire [3206], and Wrunklaw [4039],
 Stratford Halliday and Ian Ralston
the latter set on a spur deep into the Lammermuir
Hills in Berwickshire and one of the few where map
evidence shows that some of the overlying buildings
and yards were still occupied in the late 18th century.
Nevertheless, in Argyll the building recorded on Am
Burg, Coul, Islay [2067], appears to have been long
deserted by the time of a visit by Thomas Pennant in
July 1772, while the depiction on the 1st edition of the
OS 6-inch map of Dun Mhic Laitheann [2673] in the
Outer Hebrides, on a small tidal islet off Groatay, North
Uist, indicates the buildings there were abandoned by
the late 19th century.
In at least 14 other cases footings of smaller buildings,
huts and pens can be seen on coastal promontories. Such
structures appear widely on all types of fort, particularly
in the North and West Highland region. Generally they
are thought to be the remains of bothies associated
with post-medieval grazing and pasturage, rather than
contemporary buildings within the defences. This
interpretation is to some extent sustained by at least
five instances where the structures appear to butt
against, or overlie, the defences of promontory forts
– Eilean nan Caorach, Sutherland [2779], Dun Channa,
Small Isles [2688], Dun Briste, Berneray [2481], Caisteal
Odair, North Uist [2669], and Ardmenish, An Dunan,
Jura [2185]. Those on Eilean nan Caorach are typical
of local shielings, as are those overlying the inland
promontory forts at Annait, Bay River, on Skye [2692],
and Dunmore, Angus [3067]. At Port Ellen, The Ard, on
Islay, Argyll [2177], little more than three shallow oval
or sub-rectangular hollows mark the probable positions
of structures, but whether these are simply the remains
of heavily degraded post-medieval bothies, or much
older buildings is quite unknown. Traces of drystone
sub-rectangular structures at Gob Eirer on Lewis [2760]
are presented by the excavators (Nesbitt et al. 2011)
as Late Bronze Age / Iron Age in date and associated
with the construction of a wall across the neck of the
promontory, but the contexts of the radiocarbon dates,
centred on the 9th – 4th centuries BC, are not supplied
and the chronological sequence here is most uncertain.
In North-east Scotland, however, on the southern shore
of the Moray Firth, the rectangular building excavated
at Green Castle, Portknockie, [2945], was set parallel
to the early medieval timber-framed rampart and was
probably broadly contemporary with it; Gordon Noble’s
current field project is identifying further instances
of first millennium AD rectilinear architecture within
Burghead, Moray [2925] (Anon 2017).
The distribution of visible traces of round-houses
within promontory forts follows the general pattern for
all forts in Scotland. In the Greater Tyne-Forth region
the complex hilltop fort on Craik Moor, Roxburghshire
[3453], has traces of ring-ditch and ring-groove round-
houses, as does the newly identified fort at Kirktonhill,
14
Berwickshire [3947], while the stone-founded hut-
circles at Earn’s Heugh, Berwickshire [4094], almost
certainly relate to a Roman Iron Age reoccupation, a
common sequence in neighbouring Northumberland
and the eastern Borders. In Dumfriesshire the
platforms visible within Auchencat Burn [3213] may
be contemporary with the twin ramparts with a
medial ditch, but the stone-founded hut-circles at
Dalmakethar [1015] are most unusual in this county
and their relationship to the defences is thus uncertain.
Equally unusual for its district is the hut-circle at Mull
Glen, West Tarbert, Wigtownshire [0200], but not far
away on The Machars at Carghidown Castle [0229] one
of two scooped platforms within the interior of a small
univallate work was excavated to reveal a complex
sequence of round-houses probably dating from the
late 1st millennium BC (Toolis 2007). Only one other
Wigtownshire promontory fort has surface traces of
round-houses within its interior – Dinnans [0235]. The
interiors of a series of other forts on the coastal edge
around the Rhinns of Galloway comprise little more than
bare rock, and it is notable that the tiny promontory
at Dunorroch, West Cairngaan, Wigtownshire [0199],
barely has any occupiable space amongst the jagged
outcrops behind the line of its single wall, so much so
that it seems unlikely that this curious spot tucked away
at the foot of the coastal cliffs was ever intended for
permanent occupation. The excavator of Carghidown
Castle suggested that the round-houses there had been
occupied but fleetingly (Toolis 2007) and the site may
have served as no more than a refuge.
A similar picture emerges northwards up the Atlantic
coast of Argyll and the Inner and Outer Hebrides.
Traces of contemporary structures are unusual in any
fort here, and there are no more than eight coastal
promontory forts containing platforms or stony ring-
banks – in the Inner Hebrides, Dun Uragaig [2161] and
Meall Lamulum [2162] on Colonsay, and Eilean na Ba
[2487] and Nun nan Gall [2488] on Tiree; in the Small
Isles, Shellesder on Rum [2695] and Poll Duchaill on Eigg
[2524]; and in the Outer Hebrides Gob a Chuthail [2761]
and Creag Dubh [2770] on Lewis. In addition there are
two platforms in Sithean Buidhe, Argyll [2292], though,
like Craik Moor, this is best categorized as an inland
hilltop fort. With the exception of Poll Duchaill, which
contains about six circular platforms, most of the rest
are either ring-banks or, in the cases of Dun nan Gall
and Creag Dubh, oval and circular hollows with traces
of stonework around their edges, a description that
also recalls the structures at The Ard, on Islay, Argyll
[2177]. Elsewhere there is a clear mismatch between
the visibility of internal structures and the size of
the fort. The single ring-bank at Gob a Chuthail, for
example, lies within an isolated precipitous headland
extending to about 2ha with a huge and spectacular
bight eroded into its western flank. The row of three
 The Atlas of Hillforts of Britain and Ireland
conjoined hut-circles at Meall Lamulum appears
equally disproportionate to its interior of at least 1.3ha,
while at Dun Uragaig the greater part of the interior of
0.8ha is bare rock. At this latter, there are four or five
structures built immediately behind the wall, though
whether they are contemporary with or butted against
it is unclear. At Shellesder, however, a fort of 0.3ha, one
of the three possible hut-circles overlies the defences
and with numerous shieling huts also in the vicinity the
antiquity of these structures cannot be guaranteed.
Scarp-edge Forts
Scarp-edge forts, where one sector of the circuit is
completed by the lip of a natural escarpment or cliff,
share many similarities with promontory forts. As
observed already, for some it is simply a matter of degree
as to which category they have been attributed. The
two forts on Earn’s Heugh, Berwickshire [4094, 4095],
scene of one of the first of Gordon Childe’s excavations
of Scotland’s later prehistory (Childe and Forde 1932),
have been classified in the Atlas as promontory forts,
but they might just as easily be described as scarp-edge
forts, and while the south-eastern example cuts off an
angle formed where a gully forms a nick in the cliff-edge
on the south-east, the defences of the north-western
enclosure effectively contour around a hilltop position
backing onto the merest stub of a promontory on the
north-east. On the other hand, the fort on Dunman,
Wigtownshire [0177], a hillock backing onto the coastal
cliffs of the Rhinns of Galloway, where a single rampart
has been drawn around three sides, has no claim to
be a promontory topographically, yet on its south-
west flank the interior opens onto a rocky escarpment
falling away to the sea in much the same way as any
other promontory fort. In essence, promontory forts
and scarp-edge forts represent a continuum extending
from D-shaped enclosures backing onto cliff edges
and escarpments, such as Sron Uamha, Argyll [2198],
overlooking the North Channel at the southern end
of Kintyre, or Lour, Peeblesshire [3569], through
relatively shallow projections from a cliff-line seen at
the cropmarked multivallate Barns Mill, Fife [3172],
or the broad triangle of ground at Aytonlaw West,
Berwickshire [4133], to narrow precipitous fingers
with tightly constricted necks, such as Skirza Head,
Caithness [2819].
To generalise about the character of this continuum, it
extends from locations where the whole of the interior
of an enclosure lies in front of, and thus landward of,
the general line of the cliff-edge, to those where the
whole of the interior lies beyond – often seaward of –
the cliff-edge. The latter are clearly more isolated in a
physical sense, and indeed require less investment in
the construction of their defences, but whether this is
any more than a pragmatic approach to the economy
of construction of a fortification or translates into
15
wider cultural or functional differences is difficult to
determine. Nevertheless, such a pragmatic approach
to construction does seem likely for most of the 40
scarp-edge forts that have been also labelled contour
forts, including those already cited on Earn’s Heugh
or Dunman. In these cases the absence of any visible
rampart around a side with a particularly steep slope
is merely an extension of the same reasoning that
led to the construction of two or three outer lines on
gentler slopes elsewhere, either because of the greater
status accrued by being more visible from the easiest
lines of approach, or because tactically these sectors
represented the weaker flanks.
Excluding this subset of contour forts leaves 57
cases where the distinction between a cliff-edge or
promontory location is more blurred, representing
59% of those with incomplete circuits in scarp-edge
positions identified in the Atlas database.11 No more
than two of these are in coastal locations – Barsalloch
Point, Wigtownshire [0219], and Kilspindie Golf Course,
East Lothian [3820] – and the rest are located inland.
Furthermore, no fewer than 43 of the 57 are in the
Greater Tyne-Forth region, contrasting with four in
the South-West, six in the Central region (with another
three in neighbouring Angus) and only a single example
– Balmachree [2903] just east of Inverness – in northern
Scotland. Overwhelmingly, these forts are a feature of
south-eastern Scotland, occurring within the area with
the densest concentration of forts and other enclosures
in the whole country. Broadly speaking, the graph of the
55 sites with measurable sizes (Fig. 8) is an extension of
that described in the inland promontory forts. What is
perhaps more surprising is that, including the two for
which there is no size data, no more than eleven are
univallate, and that bivallate (29) and multivallate (17)
defences make up 81% of this subset; this is significantly
greater than the national figure of 63% in these two
categories for all forts in Scotland. The univallate
examples are concentrated at 0.1ha-0.3ha in the size
range, whereas an admixture of univallate and bivallate
forts include the rest of the range up to 1.1ha. The
largest is the multivallate cropmark above the River
Tyne at East Linton, East Lothian [3870], where the
defences also overlay an earlier palisaded line. By way
of comparison, the 40 contour forts of the overall scarp-
edge group break down into 13 univallate, 15 bivallate
and 12 multivallate, and the combined percentage for
the bivallate and multivallate examples is 68%, quite
close to the overall figure for Scotland.
11
A search of the Atlas will reveal a total of 132 Confirmed forts in
scarp-edge positions, but 16 of these are also annotated as coastal and
inland promontories and are included in the analysis of promontory
forts, and another 19 have complete circuits of artificial defences and
are therefore also discounted for the purposes of this discussion. This
reduces the total of what are termed here ‘scarp-edge forts’ to 97; 40
of these are also labelled contour forts, and of the remaining 57 no
measurable sizes are available for two (see Fig. 8 and Map 2)
 Stratford Halliday and Ian Ralston

Map 2. The distribution of scarp-edge forts in Scotland (N = 97) as defined on p. 15

16
 The Atlas of Hillforts of Britain and Ireland

Figure 7. The Confirmed scarp-edge forts by enclosed area, including 40 also identified as contour forts. N=95.

Figure 8. Subset of those univallate, bivallate and multivallate scarp-edge forts with measurable size data; the ghost excludes those identified
also as contour forts. N=55.

Discussion a wall, on a promontory was a normal local adaptation

In writing about promontory forts almost 40 years ago,
Raymond Lamb contended that they were a tradition
in their own right and not simply a local adaptation
of fort construction to a particular coastal setting
(1980, 6). Even then, the case was hardly compelling,
but in the light of the new data collected in the Atlas
the limitations of his schematic distribution are
starkly exposed. Promontories of all descriptions
were exploited throughout Scotland, notably in inland
settings too, and their natural attributes were merely
one set of locational choices open to the builders of
fortified enclosures. As we have seen in the varying
proportions of promontory forts to other types of fort
in Argyll and its islands, a longer hard-rock coastline
of suitable type in proportion to the interior seems to
have driven more extensive use of precipitous coastal
promontories for forts. This reaches its extreme in
the Northern Isles where all the known forts are in
locations of this type. In effect, the new data turns
Lamb’s argument on its head, to suggest that the
establishment of fortifications, whether earthworks or
wherever suitable topography occurred.
That said, we should be wary of assuming that the
construction of all forts are expressions of the same
drivers. This paper has only referred to chronology
obliquely, if only because the chronological data
for promontory forts is slender at best and hardly
sufficient to construct a detailed chronology at either a
national or regional scale. Nevertheless, it is clear from
the few reliable radiocarbon dates and artefacts from
promontory forts that they range widely in date, for
example, from the mid 1st millennium BC at Cults Loch,
Wigtownshire [0343] (Cavers and Crone 2018, 143-8), to
the late 1st millennium BC at nearby Carghidown [0229],
and from the Roman Iron Age at West Mains of Ethie,
Angus [3093], to the early medieval period at Burghead
[2925] or Green Castle, Portknockie [2945], both in
Morayshire. It is likely that others, such as Cullykhan
[2982], enjoyed longer, or at least recurrent, use. It
would be a mistake to assume that the reasons lying
behind the construction of these fortified enclosures
remained uniform across such a huge span of time, or

17
 Stratford Halliday and Ian Ralston
indeed space, let alone that they necessarily mirror the
construction of all other types of fort.
Indeed, the morphological observations recorded in the
Atlas might lead to the very opposite conclusion. While
some of the coastal promontories are in spectacular
situations, the imposing nature of their settings does
not necessarily equate with the character of their
defences. Whereas some contour forts on hills and
hillocks are prominently visible from the surrounding
landscape, creating a commanding impression of
strength in depth by stacking serried banks and ditches
or walls on the slope, this is not an opportunity afforded
many promontories. In this respect, a promontory fort
like Kemp’s Walk, Meikle Larbrax, Wigtownshire [0122],
is quite unusual; an enclosure of 0.37ha with a belt of
three ramparts and ditches at its neck, it can be seen
from over 1km away from across the valley to the east.
The Unconfirmed enclosure on the Mull of Galloway
[0201] would be another exception to the general
pattern of more limited visibility from landward.
Elsewhere along this stretch of coast, however, the
view from the natural lines of access is more confined.
Approached from along the cliff-line, for example, a
series of promontories and inlets will often obscure a
fort until the visitor is quite close by. In other examples,
the approach may be from inland across relatively
level ground or a convex slope, and the fort itself will
remain invisible until the visitor is dropping down
towards the very neck of the promontory. These are
common characteristics of coastal promontories, and
where they jut out into the sea there may be no lateral
viewpoints on land from which to see the defences
from any distance. Consequently, coastal promontory
forts may be relatively intimate places that are only
encountered when the visitor is already close at hand,
rather than being locations that appear to command
the landscape and attention from afar. In this sense,
any visual impact of the defences upon the visitor is
only achieved in the last few metres of the approach,
and in some cases only when passing through the
entrance. These same characteristics can also be found
at some inland promontories, though the patterning of
the size ranges of inland promontories in the Greater
Tyne-Forth region, allied with those of the scarp-edge
forts, would suggest that they are merely the extension
of the settlement organisation represented by the mass
of other forts and enclosures present there.
Having made the case that the selection of a promontory
was a natural choice for a fort in many landscapes, this
sense of intimacy, particularly in the locations of some
coastal promontory forts, hints that perhaps not all
these enclosed sites served the same functions, nor were
they all normal building blocks in the contemporary
settlement pattern. The excavator of Carghidown
suggested that it had been an intermittently and briefly
18
occupied refuge (Toolis 2007, 310). It may be that other
explanations should be sought for tiny enclosures such
as Dunorroch, Wigtownshire [0199], set at the foot of
the main cliff-line and entirely overlooked, and with
jagged outcrops taking up most of its interior.
Some insights into the alternative uses of promontories
is provided by a small group of northern promontory
forts that incorporate structures known as blockhouses.
This terminology not only implies a military function,
but is taken directly from military vocabulary, though
whether this is truly appropriate is debatable. Three
are included in the Atlas, all on Shetland – Ness of Burgi
[4180] and Tonga, Scatness [4181], on Mainland, and
Burgi Geos on Yell [4169] – and have been marked as
Confirmed. Each occupies a promontory with an isolated
rectangular block of masonry set squarely across its
axis and fronted by ditches and other features across
the neck. Taken into Care in 1934 and excavated in 1936,
Ness of Burgi is the archetypal example, measuring
some 23.8m in length by between 5.6m and 6.4m in
breadth and standing behind two ditches with a medial
wall. The blockhouse is pierced by a central entrance
passage furnished with door-checks and a bar-hole, and
there are mural cells built into the thickness of the wall
to either side; there was also a third mural cell, though
it is not known how, or if, this was accessible. More
recent excavation of the surviving west half of a second
blockhouse nearby at Tonga likewise revealed a central
checked entrance-passage and two mural cells, though
the only access to the inner was by means of a creep
no more than 0.2m high. In contrast to Ness of Burgi,
this blockhouse, which had been extensively rebuilt
in a second phase, stands eccentrically behind a broad
ditch with an external bank; indeed the projection of
the arc of the ditch into the eroded east sector implies
the blockhouse was set across the centre of the interior
when it was constructed. The third of this trio, Burgi
Geos, is now one of the most remote forts anywhere in
Britain. The ruinous blockhouse stands astride a narrow
promontory immediately beyond a torturous approach,
which drops down into the neck via a path lined with
stones down one side, and a mound studded with
upright slabs on the other; the latter are interpreted as
the remains of a chevaux de frise.
While at first sight these curious structures appear to
be, and have been interpreted as fortified gatehouses,
this explanation is less than entirely satisfactory. Unlike
the walls and ramparts of other promontory forts,
these blockhouses appear to have stopped short of the
cliff-edges to either side of them, leaving unimpeded
alternative access around their ends. Taken with the
presence of inaccessible chambers within the walls,
this perhaps suggests they were not strictly speaking
fortifications. Possibly they reference other more
esoteric symbolic practices connected with these places,
 The Atlas of Hillforts of Britain and Ireland
located at the hem of the land and the sea. This might
accord with the spectacular character of the natural
world as viewed from many coastal promontories set at
the interface between land, sea and sky. This is a theme
that might also be pursued in examining the distribution
of early monastic sites on the coastal edge. The
superficial similarity between, for example, the Inner
Brough [4196] on Fetlar, Shetland, likely to have served
a monastic community for at least some of its lifespan
as an occupied site, and what is generally accepted as a
substantial promontory fort at Dun Mhiughlaigh [2480]
on Mingulay in the Outer Hebrides, is noteworthy. The
latter, located on the exposed south-western side of
Mingulay and out of all proportion to the size of the
island, stands sentinel over the Atlantic with precipitous
cliffs on virtually every side – a place that must always
have been challenging for any habitation.
At present only five blockhouses are known for certain,
all of them in Shetland, the additional examples being on
small islands in lochs at Clickhimin and Whalsay, Loch of
Huxter. Others have been claimed rather less convincingly
in Orkney and Caithness, where Raymond Lamb (1980)
raised the possibility that some of the promontory
forts with thick inner walls may be hiding equivalent
structures. There was certainly a mural chamber in the
walls of the ‘outworks’ at Nybster [2820], and another
at Crosskirk [4348], both in Caithness, but elsewhere
at Dun Mhairtein, Sutherland [2788], the impression
of a possible blockhouse is partly created by a central
entrance and two successive phases of construction in
which a wall about 4.2m thick has been superimposed
on a massive dump rampart with an external ditch. A
more likely possibility is the inaccessible Stac a’ Chaisteil
on Lewis [2763]. No comparable structures have been
identified any further south, but that is not to say that
the philosophy behind their construction was not more
widely rooted in Iron Age society.
Conclusion
We hope that the foregoing selective treatment of
the promontory forts of Scotland offers some useful
insights into the diversity of the evidence that needs
to be taken into consideration in attempting to impose
some order on these records. Sifting these examples
from within the multifarious records of somewhat
cognate sites that have been built up primarily since the
nineteenth century has been shown recurrently to need
the deployment of professional value judgement. This
also impinges on the operation of particular thresholds
for inclusion or exclusion. Amongst the series of sites
included in the Atlas database, promontory forts are far
from unique in posing these kinds of issues.
What the online database allows however, is for anyone
to extract the data we have employed in the building
of our interpretations of the meaning of components
19
of this data and to reconfigure it as they judge fit. With
what is in some regards such fuzzy data, this is only
appropriate.12
Acknowledgements
The Atlas of Hillforts of Britain and Ireland Project was
funded by the Arts and Humanities Research Council
(2012-16). The Co-I was Professor Gary Lock (Institute of
Archaeology, University of Oxford); and data collection
for Ireland was undertaken with assistance from
Professor William O’Brien, University College, Cork.
For this paper, Dr Val Turner (Shetland Amenity Trust)
provided assistance with data on promontory forts in
Shetland. The maps were prepared by Dr Paula Levick.
Reference list
Alcock, L. 1981 Early historic fortifications in Scotland.
In Guilbert, G. Hillfort Studies: Essays for A. H. A. Hogg,
150-80. Leicester: Leicester University Press.
Alcock, L. and Alcock, E A. 1992 ‘Reconnaissance
excavations on Early Historic fortifications and
other royal sites in Scotland, 1974-84; A, Excavations
and other fieldwork at Forteviot, Perthshire, 1981;
B, Excavations at Urquhart Castle, Inverness-shire,
1983; C, Excavations at Dunnottar, Kincardineshire,
1984’, Proc Soc Antiq Scot 122, 215-87.
Alcock, L., Alcock, E. A. and Foster, S. M. 1986
‘Reconnaissance excavations on early historic
fortifications and other royal sites in Scotland, 1974-
84:1, excavations near St Abb’s Head, Berwickshire,
1980’. Proc Soc Antiq Scot 116, 255-279.
Anon 2017 ‘Pictish longhouse unearthed at Burghead
fort?’, Current Archaeology 331, 12.
Burgess, C. 2008 Ancient Lewis and Harris. Lewis:
Comhairle nan Eilean Siar.
Cavers, G. and Crone, C. 2018 A Lake Dwelling in its
Landscape: Iron Age settlement at Cults Loch, Castle
Kennedy, Dumfries and Galloway. Oxford: Oxbow Books.
Childe, V. G. and Forde, D. 1932 ‘Excavations in two Iron
Age Forts at Earn’s Heugh, near Coldingham’, Proc
Soc Antiq Scot 66, 152-183.
Crone, A. and Hindmarch, E. 2016 Living and dying
at Auldhame. The excavation of an Anglian monastic
settlement and medieval parish church. Edinburgh:
Society of Antiquaries of Scotland.
Fairhurst, H. 1984 Excavations at Crosskirk Broch, Caithness.
Edinburgh: Society of Antiquaries of Scotland
Monograph Series 3.
12
All calculations in this study were made in 2016 before the Atlas
was published online. They are based on spreadsheets derived from
the initial Filemaker Pro database employed on the project. Although
the overall numbers for forts in these spreadsheets have been cross-
checked against comparable searches of the online Atlas, data-
cleaning associated with the transfer to the online format may have
led to some minor discrepancies in searches for points of detail.
 Stratford Halliday and Ian Ralston
Halliday, S. P. 2019 ‘Forts and fortification in Scotland;
applying the Atlas criteria to the Scottish dataset,’
in Lock and Ralston, 2019.
Halliday, S. P. and Ralston, I. 2009 ‘How many hillforts
are there in Scotland?’ in Cooney, G., Coles, J., Ryan,
M., Sievers, S. and Becker, K. (eds) 2009 Relics of old
decency: Archaeological studies in later prehistory.
Festschrift in honour of Barry Raftery, 455-467. Dublin:
Wordwell.
Halliday, S. P. and Ralston, I. 2013 ‘Major forts and
‘minor oppida’ in Scotland: a reconsideration’,
in Krausz, S., Colin, A., Gruel, K., Ralston, I. and
Dechezleprêtre, T. (dir.) L’Age du Fer en Europe:
mélanges offerts à Olivier Buchsenschutz, 219-234.
Bordeaux: AUSONIUS Editions Mémoires 32.
Harding, D. W. 2012 Iron Age hillforts in Britain and
beyond. Oxford: Oxford University Press.
Hunter, J. R. 1996 Fair Isle: the archaeology of an island
community. Edinburgh: HMSO.
Lamb, R. G. 1980 Iron Age promontory forts in the Northern
Isles. Oxford: Brit Archaeol Rep Brit Ser 79.
Lock, G. and Ralston, I. 2017 Atlas of Hillforts of Britain
and Ireland. [ONLINE] Available at: https://hillforts.
arch.ox.ac.uk
Lock, G. and Ralston, I. (eds) 2019 Hillforts: Britain,
Ireland and the Nearer Continent: Papers from the Atlas
of Hillforts of Britain and Ireland Conference, June 2017.
Oxford: Archaeopress.
Lock, G. and Ralston, I. 2020 ‘A new overview of the
later prehistoric hillforts of Britain and Ireland’
in Delfino, D., Coimbra, F., Cardoso, D. and Cruz,
G. (eds) Late Prehistoric Fortifications in Europe:
Defensive, Symbolic and Territorial Aspects from the
Chalcolithic to the Iron Age. (= UISPP Metal Ages in
20
Europe Commission international conference,
Sociedade Martins Sarmento, Guimaraes, Portugal,
November 2017). Oxford: Archaeopress.
MacKie, E. W. 2007 The roundhouses, brochs and
wheelhouses of Atlantic Scotland c. 700 BC – AD 500.
Architecture and material culture Part 2, the northern
and southern mainland and the western islands.
Oxford: Brit Archael Rep Brit Ser 444.
MacKie, E. W. 2016 Brochs and the Empire. The impact of
Rome on Iron Age Scotland as seen in the Leckie broch
excavations. Oxford: Archaeopress.
Nesbitt, C., Church, M. J., and Gilmour, S. M. D. 2011
‘Domestic, industrial, (en)closed? Survey and
excavation of a Late Bronze Age/Early Iron Age
promontory enclosure at Gob Eirer, Lewis, Western
Isles’, Proc Soc Antiq Scot 141, 31-74.
Past Horizons. 2015 ‘Earliest Pictish fort yet
discovered was situated on sea stack’. Past
Horizons. http://www.pasthorizonspr.com/index.
php/archives/07/2015/earliest-pictish-fort-
yet-discovered-was-situated-on-sea-stack. Last
consulted 9th August 2018.
Strachan, R. 2000 ‘Mull of Galloway linear earthworks’,
Discovery and Excavation Scotland 1, 21.
Toolis, R. 2003 ‘A Survey of the Promontory Forts of
the North Solway Coast’, Trans Dumfries Galloway
Nat Hist Antiq Soc 77, 37–78.
Toolis, R. 2007 ‘Intermittent occupation and
forced abandonment: excavation of an Iron Age
promontory fort at Carghidown, Dumfries and
Galloway’, Proc. Soc. Antiq. Scot. 137, 265–318.
Toolis, R. 2015 ‘Iron Age settlement patterns in
Galloway’, Trans Dumfries Galloway Nat Hist Antiq Soc
89, 17-34.13
13
This paper was revised for submission in 2018 and has only been
slightly updated thereafter. The final Atlas has now been published
and is supported by online data as outlined on p. xvi of that volume.
Lock, G. and Ralston, I 2022 The Atlas of the Hillforts of Britain and Ireland.
Edinburgh: Edinburgh University Press.
 A long, largely aceramic, period of Devon’s prehistory
Henrietta Quinnell
I first met John Collis in the summer of 1963 when we
were both excavating Urnfield cemetery sites near
Xanten for the late Laurence Barfield, then working
for the Landesmuseum at Bonn. The excavations
themselves were rather unexciting, just variations on
cremations in pottery vessels. The English group of
excavators included the late Alan Carter, Ros Nisbett,
Jeff Davies and the late Eric Talbot, and we used the
excavations as the base for exploring much of the
adjacent Rhineland. The Deutschmark was then ten to
a pound, enabled travelling, and we only worked a five
day week. At weekends we explored the area by Rhine
ferry, down to Bonn, Cologne, Koblenz and Trier and
up across the border for a night in Arnhem in Holland.
Despite all this John found time to learn to read Russian.
After both John and I had completed our university
studies, we met again at Exeter University in 1970. At the
beginning of that year I was appointed as Archaeology
Lecturer in the Department of Extra-Mural Studies, and
John came in the autumn as the prehistorian and third
archaeologist in the Department of History. Both new
posts were planned by, and campaigned for inside the
University, by Aileen Fox. She envisaged appointees to
both posts engaging with local archaeological work,
and especially in the rescue excavations increasingly
necessary at a time of urban renewal and road building.
She saw this involvement by University staff benefitting
archaeological research in its region. For myself, this
provided a channel of interests which I would develop
through my career and I remained at Exeter until the
millennium. John was firmly directed to running some
of the excavations in Exeter, for which the recently
founded Exeter Museums Archaeological Field Unit
had been set up. He spent the summer vacation of 1971
directing excavations on Guildhall Shopping Centre
site: for background see Fox (2000, 140). This site began
with barrack blocks of the Roman fortress whose
presence at Exeter was rapidly being established, and
continued through the subsequent town and its phases
of medieval and later activity. John moved to Sheffield
at the end of the year, its Department having a more
prehistoric focus. John continued to research aspects
of Devon’s prehistory well after his move to Sheffield
(Collis 1979; 1983; Fleming and Collis 1973).
When I first became involved with the Iron Age in
Devon, directing excavation of a long trench right
through Woodbury Castle in advance of road widening
(Miles 1975), the sparsity of its ceramics in the County
Challenging Preconceptions of the European Iron Age
(Archaeopress 2022): 21–28
was already well recognized. Hillfort excavations by
the Devon Archaeological Society before and just after
WWII at Okehampton (Brailsford 1938) and Stoke Hill
(Radford 1937) had produced few or no ceramics.
At Hembury in East Devon the ceramic sequence
appeared to start in what would now be termed the
Middle Iron Age, despite the hillfort having an obvious
long structural history (Liddell 1935, summarising
the five seasons’ work). One wonders whether the
excavations would have been run for so many seasons
had the underlying Neolithic causewayed enclosure
with its rich ceramic assemblage not been found: the
subsequent history of the Archaeological Society
could have been very different. Particularly in pre-
radiocarbon days, a prefusion of potentially dateable
artefacts was a prerequisite of prehistoric excavation.
Only at Blackbury in East Devon was there a moderate
ceramic assemblage ascribed to Iron A or B (Young and
Richardson 1954-5), and at Milber Down a small Iron Age
B assemblage (Fox et al. 1949-50). At Woodbury Castle in
1971 there was only one sherd of distinctive form (Miles
1975, Fig 11, 1) ascribed to the ‘initial phases of the Iron
Age’.
In the decades since then, and particularly in the last
two decades, I have been studying the whole sequence
of ceramics in Cornwall and in Devon, from their
inception in the Neolithic until the arrival of Roman
influence, and have for both counties seen most
significant assemblages as they were discovered and
examined those nearly all those published before 1970,
a diachronic study of regional prehistoric ceramics. In
general ceramics of all periods were more common in
Cornwall than in Devon, and the study of the one has
fed into the other. From time to time this research
benefits a short piece of synthesis. This volume of
papers for John allows me to offer him this synthesis
for the period between the Middle Bronze Age and the
Middle Iron Age, broadly from 1100 to 300 BC, a small
building block in the developing chronology of the
region. It seems timely, when so many archaeologists
are still unaware of the absence of storage pits in this
region, with their time capsule potential for artefact
study, and of the apparent sparsity of ceramics for
much of the Devon Iron Age.
Scholars involved in the study of prehistoric Europe
generally have taken the use of pottery as normal, from
antiquarians in the eighteenth century to those of the
present generation. Ceramics, especially drinking cups,
 Henrietta Quinnell

Figure 1. Locations of principal sites in Devon referred to in the text.

figure frequently in accounts of classical civilisations
on which the former were reared, and it is surely no
co-incidence that the vessels still known as ‘Beakers’
were first so named back in the nineteenth century (eg
Thurnam 1871), bringing with the appellation all the
connotations of imbibing then current in Europe and
learnt by every schoolboy with a classical education.
Pottery, because of its enormous potential variations
of shape and decoration, has played a central role in
constructing chronologies for prehistoric Europe and
of Britain and in identifying past communities. Even
after some 70 years of increasingly reliable scientific
dating, ceramic chronology is still extremely useful in
providing initial spot dating for sites under excavation.
When pottery is sparse or absent from sites or areas, the
normal processes of modern analysis are slowed, leading
sometimes to the implication that archaeological data
is proving inconvenient (eg Todd 1987, 156)!
There are many parts of the world where communities
have developed successful and often long lived
societies without any use of ceramics, notably in
Polynesia and New Zealand. Here societies thrived until
‘discovery’ by the European world. The consequent
written record of aceramic behaviour could be linked
through observation with successful ways of cooking
and eating without pottery. An excellent account of
aceramic Maori communities in New Zealand is given

22
 A long, largely aceramic, period of Devon’s prehistory
by Davidson (1987). Reminding ourselves that pottery
was not used by many human groups prompts us not to
take its presence amongst past communities in Europe
for granted. By and large, most prehistoric European
communities since the development of agriculture did
use pottery, some for a while before this. Its absence or
scarcity at various times and places will have had good
and sufficient reasons which can not (yet) be provided.
But variation in the use of ceramics, despite our lack of
understanding, can still have things to tell us about the
times and places this occurred.
In general the pattern of ceramic use in Devon matches
that found in most other parts of Britain from the Early
Neolithic down to the Middle Bronze Age. And in general
through these three millennia, more ceramics on sites
in Cornwall were used, or at least discarded, than on
those on sites in Devon. This may be due, in part, to the
early discovery and use of the excellent gabbroic potting
clays of the Lizard. (Modern counties are used here as
convenient handling blocks for geographical data). The
same appears to be broadly true of the adjacent parts
of Dorset and Somerset lying to the East and North of
Devon. In the Middle Bronze Age the Trevisker style
was in use across this whole area, with some sub-
regional variations which are still being disentangled.
This regional Middle Bronze Age use of similar ceramics
comes to an end at a date around the eleventh century
BC, a date which still needs closer definition (Quinnell
2012a). The use of profusely decorated Trevisker
ceramics is then replaced by ‘Late Bronze Age Plain
Ware’, a generally undecorated style with simple
forms, not recognised until the 1990s at Brean Down
(Woodward 1990) and on Lundy Island (Quinnell 2010).
This style broadly lasts until the tenth/ninth centuries
BC, profuse on some sites in Cornwall such as Higher
Besore near Truro (Gossip 2021), in Somerset at Brean
Down and in Dorset at Tinney’s Lane (Best et al 2013),
less so in Devon. The replacement of Trevisker ceramics
by LBA Plain Ware appears broadly contemporary with
the now classic abandonment of much of the profuse
hut circle settlement on Dartmoor and its shrinkage
on the other granite landscapes of the South West: this
decrease in ceramics was first highlighted some forty
years ago by Bob Silvester (1979). Are the two to be
linked?
Trevisker vessels come in a variety of sizes, and had
presumably a variety of uses. They are almost always
decorated and therefore easily recognised. The
subsequent plain ware tends to have smaller vessels,
with virtually no decoration, and could easily have been
passed unrecognised in all but very recent excavations
and examinations. Obviously there were changes in
Bronze Age societies we cannot detect which underlie
these ceramic changes. It may also be that movement
of settlement away from Dartmoor was less abrupt than
23
is sometimes assumed. There have been comparatively
few ‘modern’ excavations published of Moorland
settlements. It is worth recalling that activity at Shaugh
Moor has radiocarbon dates which continue down to
the 9th century cal BC (Balaam et al 1982, Table 3). Re-
examination of the ceramics at Kestor, dug by Aileen
Fox in 1951/2 (Fox 1954) has shown that the large hut
circle structure was built and used in the Middle Bronze
Age, with some re-occupation in the Early Iron Age; the
now classic iron working is now radiocarbon dated to
the early post-Roman centuries (Quinnell 2016). The
nearby hut circle at Teigncombe, with the most careful
excavation and record ever to take place on Dartmoor,
has been shown to have a phase with minimal Early
Iron Age ceramics subsequent to its main Middle
Bronze Age phase (Gerrard 2016). This type of re-use
evidence would not have been detected on most earlier
excavations and may be more frequent than is generally
supposed. If for then good reasons, pottery was ceasing
to be used in any quantity during the Late Bronze Age,
it provides some background for its sparsity locally in
the Early Iron Age.
In review of data from Dartmoor and its periphery
written some 25 years back, the author (Quinnell 1994)
suggested a simple term ‘Post-Trevisker’ to cover
ceramics after the Middle Bronze Age and before the
decorated styles of the Middle Iron Age. Late Bronze
Age Plain ware can now be seen as the first of three
successive ceramics styles which were used in this
period, the second and the third Earliest and Early Iron
Age described below. LBA Plain Ware (Figure 2) has
been published from sites on Lundy (Quinnell 2010) and
from Dainton (Sylvester 1980). It is almost certainly
present at Hembury (Honiton) hillfort (see below), and
assemblages of various sizes await publication from
Stowford Rise, Sidmouth (AC Archaeology ACD145);
from Strathculm Road, Hele (AC Archaeology ACD1158);
from Shortlands, Cullompton (SW Archaeology) and
from a site at Hill Barton, Exeter (Quinnell 2019).
In Cornwall the sequence of ceramics, because they
generally survive in greater quantities, is now fairly
well understood (Quinnell 2011), but even in here
ceramics from the 9th to 7th centuries are few in
quantity. The report on Enclosure 1 at Tremough,
Penryn, demonstrates that by the 10th/9th century
BC LBA Plain Wares widen in range to include the
earliest shouldered and carinated forms (Quinnell
2015, 75). These shouldered and carinated, generally
large, vessels continue to become the main types of the
Earliest Iron Age c 800-600 BC. A range of simpler, often
smaller and generally necked jars follow in the Early
Iron Age Plain Jar Group c 600-300 BC (Figures 3 and
4). Both groups have little decoration. Earliest Iron Age
ceramics generally occur in smaller quantities than do
the preceding LBA Plain Ware or the succeeding Early
 Henrietta Quinnell

Figure 2. Late Bronze Age Plain Ware forms, c. 1100-800 BC, based on Cornish examples.

Iron Age Plain Jar group. This sequence and broad
chronology apply well to the more limited Devon
material, with division between Earliest (c 800-600 BC)
and Early Iron Age (c 600-300 BC).
A recent review by the author of the entire assemblage
from Hembury (Honiton) hillfort has identified a small
but previously unrecognised assemblage of Late Bronze
Age to Earliest Iron Age date. Previously the ceramics at
Hembury were thought to start in the Middle Iron Age.
The Gathering Time programme of dating fortuitously
produced two pairs of dates from early features spanning
the 8th to 5th centuries cal BC, which should relate to early
structural phase(s) (Whittle et al 2011, 478-93). Excavations
at Raddon hillfort, Stockleigh Pomeroy, produced a a
small assemblage of Earliest Iron Age material related

24
 A long, largely aceramic, period of Devon’s prehistory
Figure 3. Earliest Iron Age pottery forms, c. 800-600 BC, based on Cornish examples.
to dates within a C14 range 800-400 cal BC: Early Iron
Age material has not be recognised but activity without
surviving ceramics continued through the period 400-
50 BC (Quinnell 1999, 52-3). The recently excavated Hill
Barton earthwork and roundhouse just east of Exeter, with
use centring on the 9th century cal BC, had no associated
ceramics. A roundhouse at Blackhorse a little to the east
has no ceramics but a single radiocarbon date of 770-330
cal BC (Fitzpatrick et al 1999, 191). Woodbury hillfort, in
east Devon, produced only a few sherds from limited
excavations despite the earthworks having a complex
history (Miles 1975). Most of the assemblage from the 1939
excavations at Dainton is probably Early Iron Age (Willis
and Rogers 1951), although the unpublished excavations
of 1986 may extend the range of material (Smith and
Humble 1986). Foal’s Arrishes on Dartmoor has a vessel
which almost certainly belongs here (Radford 1952, Fig 13,
No 6). The promontory site Mount Batten, at the mouth
of the Plym, has Early Iron Age material but, unusually,
appears to have an apparently unbroken sequence
extending down through to the Roman period (Cunliffe
1988, passim), as apparently does the assemblage from
Kent’s Cavern (Sylvester 1986, Fig. 2).
The assemblage from Blackbury hillfort (Young and
Richardson 1954-5) can now be seen as a mixture of Early
25
Iron and Middle Iron Age vessels. The Iron Age material
long known from Kestor, east Dartmoor, and that from
the recently excavated Teigncombe hut circle close by,
both referred to above, are clearly Early Iron Age Plain
Jar Group (Quinnell 2016). Digby Site 2 just east of Exeter
has a single Plain Jar Group vessel from a ditch (Quinnell
and Farnell 2016,130) and the moderate g assemblage
at Hayes Farm 4 km east (Wood 2014) are also of this
Group. Most of the small assemblage associated with
structures predating the south east part of Honeyditches
Roman villa can now be confidently assigned to this
period (Silvester 1981, 61-3, Fig 11). The minimal sherds
of pottery from Holworthy, Parracombe on Exmoor are
likely to belong to the Plain Jar Group: there are also C14
dates from this enclosure site which indicate fourth or
early 3rd century BC activity (Green 2009, 73, 93).
The fabrics throughout this period are generally locally
sourced and there is no record of any gabbroic pottery,
unlike the Middle Bronze Age when it appears to have
been relatively common and the Later Iron Age when it
is also regularly present in small quantities. Generally
ceramics are poorly made, but with inclusions more
comparable in size to those that succeeded than those
that preceded them. It is doubtful how much use and
movement most would have withstood.
 Henrietta Quinnell
Figure 4. Early Iron Age pottery forms, c 600-300 BC, based on Cornish examples.
Scarcity of pottery implies the use of other types of
container and wood and leather containers are likely
to have been used (Earwood 1999, Chapter 1). It is
relevant here that there was a long aceramic period
in Devon between Roman and late Saxon - although
Cornwall maintained a continuous use of local
ceramics throughout. Ceramics were not used in Wales
throughout most of this period and, as in Devon, were
used sparsely in the first millennium BC (Lynch et al.
2000, 199-202). The sparsity of ceramics in Wales in
these periods is generally accepted, perhaps because
the tenth century Law Code of Hywel Dda describes
with authority and in some detail the minutiae of
property which did not involve ceramics (Jenkins
1986).
Recent work on a long pipeline route across South
Devon, from the Otter Valley to Langage west of
Ivybridge revealed a wide range of prehistoric sites,
but none from the period under consideration here
(Mudd and Joyce 2014, 187) despite a number with
radiocarbon dates from the long post-Roman aceramic
period referred to above. The authors suggest that
either settlements were genuinely sparse, or that
26
they lacked sub-soil features which contained
archaeologically detectable material. This period
coincides with the beginning of sparse Dartmoor
settlement, even if a few sites can now be seen to have
been in use there. As Mudd and Joyce point out, the
(some-what sparse) pollen evidence does not support
woodland regeneration which might be expected to
accompany a long term drop in population. There
is over half a millennium in which traces of former
inhabitants are elusive and, when these are present, do
not always demonstrate the use of ceramics. While the
post-Roman period for some 600 years can be described
as aceramic, this longer prehistoric period may be
described as ‘largely aceramic’, for which reasons can
be only speculation, something of a paradox for the
period when hillforts are thought to be have become
established. We need to understand much more about
the ways in which ceramics related to variations in diet
and cooking. Recent rich finds of organic containers, as
at Must Farm, together with the developing analytical
techniques for identifying foodstuffs, should provide
new pathways of understanding and surely, after a
quarter century, a reworking of Earwood’s (1993)
Domestic Wooden Artefacts.
 A long, largely aceramic, period of Devon’s prehistory
References
Balaam, N. D., Smith, K. and Wainwright, G. J. 1982. The
Shaugh Moor Project: fourth report-environment,
context and conclusion. Proceedings Prehistoric
Society 48:203-278.
Best, J., Woodward, A. and Tyler, K. 2013. Late Bronze
Age Pottery Production: Evidence from a 12th to 11th
Century BC Settlement at Tinney’s Lane, Sherborne,
Dorset. Dorchester: Dorset Natural History and
Archaeological Society Monograph 21.
Brailsford, J. W. 1938. Excavations at the Promontory
Fort near Okehampton Station. Proceedings Devon
Archaeological Exploration Society III:2: 86-91.
Collis, J. 1979. Cranbrook Castle Revisited. Proceedings
Devon Archaeological Society 3: 191-4.
Collis, J. 1983. Field Systems and Boundaries on Shaugh
Moor and at Wotter, Dartmoor. Proceedings Devon
Archaeological Society 41: 47-62.
Cunliffe, B. 1988. Mount Batten, Plymouth. A Prehistoric
and Roman Port. Oxford: Oxford Committee for
Archaeology Monograph 26.
Davidson, J. 1987. The Prehistory of New Zealand. Auckland:
Longman Paul.
Earwood, C. 1993. Domestic Wooden Artefacts in Britain
and Ireland from Neolithic to Viking Times. Exeter:
University Press.
Fitzpatrick, A. P., Butterworth, C. A. and Grove, J. 1999.
Prehistoric & Roman Sites in East Devon: the A30 Honiton
to Exeter DBFO Scheme, 1996-9. Salisbury: Wessex
Archaeology Report 16.
Fleming, A. and Collis, J. 1973. A Late Prehistoric Reave
System near Cholwich Town, Dartmoor. Proceedings
Devon Archaeological Society 31: 1-22.
Fox, A. 1954. Excavations at Kestor. Transactions
Devonshire Assocation 96: 21-62.
Fox, A. 2000. Aileen – A Pioneering Archaeologist.
Leominster: Gracewing.
Fox, A., Radford, C. A. R., Rogers, E. H. and Shorter, A. H.
1949-50. Proceedings Devon Archaeological Exploration
Society IV:2: 27-66.
Gerrard, S. 2016. Archaeology and Bracken: the
Teigncombe Prehistoric Roundhouse Excavationn
1999-2005. Proceedings Devon Archaeological Society
74: 1-64.
Gossip, J, 2021. Life Outside the Round: Bronze Age and
Iron Age Settlement at Higher Besore and Truro
College, in Jones, A M and Kirkham, G (eds.), Later
Prehistoric Settlement in Cornwall and the Isles of Scilly:
Evidence from Five Excavations, 149-282. Oxford:
Archaeopress
Green, T. 2009. Excavation of a Hillslope Enclosure
at Holworthy Farm, Parracomber displaying
Bronze Age and Iron Age activity. Proceedings Devon
Archaeological Society 67: 39-98.
Jenkins, D. 1986. The Law of Hywel Dda: law texts of medieval
Wales. Llandysul: Gomer.
27
Liddell, D. M. 1935. Report on the Excavations at
Hembury Fort (1934 and 1935). Proceedings Devon
Archaeological Society II:3: 135-175.
Lynch, F., Aldhouse-Green, S. and Davies, J. L. 2000.
Prehistoric Wales. Stroud: Sutton Publishing.
Miles, H. 1975. Excavations at Woodbury Castle, East
Devon, 1975. Proceedings Devon Archaeological Society
33: 183-208.
Mudd. A. and Joyce, S. 2014. The Archaeology of the South-
West Reinforcement Gas Pipeline, Devon. Investigations
in 2005-7. Cirencester: Cotswold Archaeology
Monograph 6.
Quinnell, H. 1994. Becoming Marginal ? Dartmoor in
Later Prehistory. Proceedings Devon Archaeological
Exploration Society 52: 75-84.
Quinnell, H. 1999. Pottery, in T.H. Gent and H. Quinnell,
Excavations of a Causewayed Enclosure and Hillfort
on Raddon Hill, Stockleigh Pomeroy, Proceedings
Devon Archaeological Society 57: 38-53.
Quinnell, H. 2010. Prehistoric and Roman material from
Lundy. Proceedings Devon Archaeological Society 68: 19-
60.
Quinnell, H. 2011. A summary of Cornish ceramics in
the first millennium BC. Cornish Archaeology 50: 231-
240.
Quinnell, H. 2012a. Trevisker Pottery; Some Recent
Studies, in W. Britnell and R.J. Silvester (eds.)
Reflections on the Past. Essays in honour of Frances
Lynch, 147-71. Welshpool: Cambrian Archaeological
Association.
Quinnell, H. 2012b. Prehistoric Sherds from Shortlands,
Cullumpton. Report to South West Archaeology,
South Molton.
Quinnell, H. 2015. The prehistoric ceramics, in A.M.
Jones, J. Gossip and H. Quinnell, Settlement and
metalworking in the Middle Bronze Age and Beyond, 53-
80. Leiden: Sidestone Press.
Quinnell, H. 2016. The Pottery, in S. Gerrard,
Archaeology and Bracken: the Teigncombe
Prehistoric Roundhouse Excavationn 1999-2005.
Proceedings Devon Archaeological Society 74: 30-43.
Quinnell, H. 2019. Prehistoric pottery, in Mudd, A.,
Garland, N. and Brindle, T. 2019. A Late Bronze Age
Ringwork and other Prehistoric, Roman and later
occupation at Hill Barton, Exeter, Devon. Excavations
2012–2015, Proceedings Devon Archaeological Society
77, 174-178.
Quinnell, H. and Farnell, A. 2016 Prehistoric Sites in the
Digby area of Exeter. Proceedings Devon Archaeological
Society 74: 65-206.
Radford, C. A. R. 1937. Stoke Hill Camp. Proceedings Devon
Archaeological Exploration Society III:1: 24-32.
Radford, C. A. R. 1952. Prehistoric Settlement on
Dartmoor and the Cornish Moors. Proceedings
Prehistoric Society 18:1: 55-84.
Silvester, R. J. 1979. The relationship of First
Millennium BC Settlement to the Upland Areas of
 Henrietta Quinnell
the South-West. Proceedings Devon Archaeological
Society 37: 176-90.
Silvester, R. J. 1980. Excavations at Honeyditches Roman
Villa, Seaton. Proceedings Devon Archaeological Society
38: 17-48.
Silvester, R. J. 1981. The Prehistoric Open Settlement
at Dainton, South Devon. Proceedings Devon
Archaeological Society 39: 37-88.
Silvester, R. J. 1986. The Later Prehistoric and Roman
Material from Kent’s Cavern. Proceedings Devon
Archaeological Society 44: 9-38.
Smith, G. H. and Humble, J. 1986. Excavation at the Late
Bronze Age and Early Iron Age Settlement and Field System
at Dainton, South Devon, Central Excavation Unit Site 414.
A Preliminary Report. London: English Heritage.
Taylor, R. 2016. Neolithic to Early Iron Age fabric, in
Quinnell and Farnell: 120-1.
Thurnam, J. 1871. On Ancient British Barrows, especially
those of Wiltshire and the adjoining counties,
Archaeologia 43: 285-552.
Todd, M. 1987. The South-West to AD 1000. London:
Longmans.
28
Whittle, A., Healy, F. and Bayliss, A. 2011. Gathering Time.
Dating the Early Neolithic Enclosures of Southern Britain
and Ireland. Oxford; Oxbow Books.
Willis, L. and Rogers, E. H. 1951. Dainton Earthworks.
Proceedings Devon Archaeological Exploration Society.
IV:4: 79-101.
Wood, I. 2014. Ceramic and Petrographic report, in
J. Hart, I.Wood, A. Barber, M. Brett., and A. Hardy,
Prehistoric Land Use in the Clyst Valley: Excavations
at Hayes Farm, Clyst Honiton, 1996-2012. Proceedings
Devon Archaeological Society 72: 16-25.
Woodward, A. 1990. Bronze Age Pottery, in M. Bell, Brean
Down Excavations 1983-198: 121-145. English Heritage
Archaeological Report 15.
Woodward, A. and Cane, C. 1991. The Bronze Age
pottery’, in J. Nowakowski, Trethellan Farm,
Newquay: the excavation of a lowland Bronze
Age settlement and Iron Age cemetery. Cornish
Archaeology 30: 103-31.
Young, A. and Richardson, K. M. 1954-55. Report on the
Excavations at Blackbury Castle. Proceedings Devon
Archaeological Exploration Society V:2 & 3: 43-67.
 Deconstructing archaeological databases
Martin Kuna
European archaeology has recently started to employ large-
scale databases of field data for a purposeful research of
settlement history in the long-term perspective. Projects
of this kind target topics like the demography of past
populations, regional differences in settlement density,
environmental impact of settlement, etc. The results of these
efforts are mostly very beneficial casting new light on some
general aspects of historical development of various parts
of Europe; they are not, however, without risks. The present
paper points out some of the pitfalls hidden in straightforward
generalizations of data sets like the national ‘sites and
monuments records’, registers of 14C dates and similar ‘big
data’. Particularly, the varied visibility of archaeological
remains from different periods (caused mainly by secondary
cultural factors in combination with post-depositional
processes) and the course of archaeological fieldwork itself
are among the key issues to be thoroughly considered.1
Introduction
During the last decade, Czech and European
archaeology saw a number of efforts to make use of
the national archaeological databases of field data to
create general models of settlement development in
the past. Here belong the attempts, for example, to use
the Archaeological Map of the CR to model demographic
trends from the Neolithic to the Early Middle Ages in
Bohemia (Dreslerová 2011; Pokorný 2011, pp. 245–51;
Dreslerová – Demján 2015; Demján – Dreslerová 2016)
and Moravia (Kolář et al. 2015). On a European scale,
the study of the demographic trends in the Mesolithic
and Neolithic periods in Western or Southern
Europe by means of large series of radiocarbon dates
(Collard et al. 2010; Shennan et al. 2013; Porčić et al.
2016) has been carried out during the last decade
introducing new theoretical perspectives and valuable
methodological contributions.
The common basis of all these attempts show a general
trust in the quantity of data reflecting some relevant
parameters of the past, such as, in this case, the
changing settlement density and population size. The
authors of the cited projects mostly believe that the
large amounts of data can suppress the individual and
accidental variability of the archaeological record and,
1
Work on this paper has been supported by the Ministry of
Education, Youth and Sports of the CR within the EU Operational
Programme Research, Development and Education (‘Archaeological
Information System of the CR – second generation’ project; ID:
16_013/0001439-01).
Challenging Preconceptions of the European Iron Age
(Archaeopress 2022): 29–40
as a whole, reflect general trends with appropriate
reliability. This was explicitly formulated, for example,
by the British team, according to which a ‘relationship
between the number of dated archaeological sites
falling within a given time interval in a given region…
and population density’ exists (Shennan et al. 2013,
p. 3). P. Demján and D. Dreslerová (2016, pp. 101–2)
basically agree on this, and also assume that the
density of archaeological evidence may be understood
as proxy data for the study of the demographic
development in the past.
The authors of these projects are of course aware of
the fact that secondary factors may influence the
archaeological record and pay considerable attention
to the statistical processing of their data (namely
Timpson et al. 2014; Demján – Dreslerová 2016, p.
102). Statistical methods are, of course, appropriate to
prove that the resulting values (e.g. graphs) represent
archaeological data correctly and reliably describe
their spatial and chronological distribution. Does it
mean that all potential problems have been solved
and the results may be without any doubt taken as
relevant?
According to my opinion, not entirely; to my mind,
the key problem is not whether the quantitative
results reliably represent the available archaeological
record, but mainly, whether the archaeological record
(and the available data on it) reliably represents past
reality. In this respect, not even the most sophisticated
statistical methods can help us; what has to be done,
is a thorough critique of the available record (data)
based on a sound archaeological theory. The following
contribution aims to present some comments and
suggestions on this subject.
The data, which will be used here as an example, come
from the Archaeological Map of the CR (AMCR), which
after more than 20-year work has been implemented
into a publically accessible research infrastructure
(Archaeological Information System of the CR; Kuna et al.
2015). Today, the AMCR contains information on about
90,000 fieldwork events on the territory of Bohemia
(the western part of the CR) from the beginning of
archaeology until today. The data we are using here
come from before 2011 when their final revision
started; this, however, does not lessen their potential
for our purposes. Anyway, the AMCR data, processed
even more accurately and added by thousands of new
 Martin Kuna
entries every year, create a source of information ready
to be exploited in various other research attempts in
the future.
Counting sites from archaeological data sets
Research projects following up the settlement
development in long-term intervals usually do not
work with individual movable objects, but rather with
larger units of archaeological record. These units
are called ‘sites’ or ‘site phases’ (Collard et al. 2010),
‘components’ (e.g. Kolář et al. 2015 and all other
authors adopting the terminology of the ‘settlement
area theory’ of E. Neustupný – e.g. 1998; 2010) or
‘evidence’ (cf. the ‘evidence density estimation’ [EDE]
method used by Demján – Dreslerová 2016). All these
terms are derived from the idea that a certain group
of finds can be viewed as the remains of one human
community in a certain time interval. Under this
condition, the observed number of such units can be
used as a ‘proxy’ figure for the number of communities
(the population size) at a given time.
Considering a particular model of a past society, the
site/site-phases units are usually defined in a formal
way, namely by their spatial and temporal extent.
Usually, if the spatial and temporal distance between
specific finds is smaller than the defined limits, they
are understood as evidence of one and the same
community (as one site/site phase). If in some respect
the distance is larger, we have to understand these as
various ‘sites’, i.e. evidence of non-identical human
communities. Hence, we can rephrase the relevance of
archaeological databases as a question of the reliability
of the number of sites classified to assigned periods
and regions. Since nobody is expecting to capture all
original sites in the landscape, archaeology usually
understands the available data as relative figures, as
samples from which, under specific conditions, the
original number of sites can be estimated and which
can be used to compare the changing density of a
population in various periods and areas.
But is the composition of the available archaeological
databases really so reliable that we may take the
numbers of sites as representative samples and
interpret them this way? I tend to believe that the
distortion contained in the data is much larger than
admitted in most of the cited contributions. In my view,
the patterns observable in our data are influenced
(apart from the actual size of the population in the
past) by two groups of other factors:
• properties of the past culture influencing the
quantity of sites in the archaeological record:
1. average community size;
2. average lifetime of a site (activity area);
30
3. site (activity area) specialization: number of
activity areas simultaneously managed by a
community;
• properties of the past culture and the formation
processes influencing survival and visibility of
sites in the archaeological record:
4. rate of reduction of the site numbers by post-
depositional transformations;
5. methods used in archaeological excavations
and surveys;
6. visibility (preservation) of the distinctive
parts of activity areas (activity remains)
The impact of the first group of factors is obvious.
Populations of the same size may have inhabited a small
number of large settlements or a larger number of
small ones in dependence on their economic strategies
and social structure, environment, relations between
individual populations etc. At the same time, activity
areas (sites) may have been relocated in short time
intervals (by which the resulting number of sites grows);
while in other periods the activity areas may have been
long-living (producing a smaller number of sites).
Necessarily, the number of sites may also be influenced
by the number of activity areas simultaneously used by
one community: if, for example, a community buried
their dead within its residential area, the settlement
and cemetery grew into one site; if for the burials
another place was used, the same community may have
left two sites; something similar can also develop in the
case of seasonal camps, transhumance etc.
In principal, all these factors can be studied, but
there is a long way to go to be able to describe them
with an acceptable probability. By now, some methods
are suitable to describe some of these factors by, for
example, a more accurate dating of individual sites
using radiocarbon dates (Shennan et al. 2013; this
concerns factor 2). However, most of the other aspects
of the prehistoric settlement patterns (factors 1 and
3) yet remain poorly explored. On the other hand, it is
legitimate to assume that the settlement structure and
dynamics of agricultural societies from the Neolithic
to the Early Middle Ages was in most basic aspects
very similar, therefore fundamental differences in the
quantity of sites based on the first group of factors need
not be expected. At least, if chronological demographic
trends within one and the same region are studied, this
assumption may, to a high degree, be justifiable.
Nevertheless, it would be premature to derive too much
optimism from this conclusion. Apart from the factors
of the first group, the composition of archaeological
data is influenced by those of the second group, too.
They include the reduction of the number of sites by
post-depositional processes, the uneven sensitivity of
 Deconstructing archaeological databases

various fieldwork methods towards different types of
sites and the variable archaeological visibility of the
remains from different periods.
The influence of factor 4 (post-depositional reduction)
is widely known. In the course of time, many sites
have been destroyed by erosion, agricultural activities
or timber extraction and therefore are lacking in our
databases. Such destructive processes, however, mostly
affect sites from all periods in a similar way and need
not necessarily have produced false patterns on the
chronological axis. In this case, ‘false’ patterning is
rather to be expected in space than in time, since various
regions differ in the landscape type and, consequently,
in the differing types and intensity of post-depositional
processes.
Fieldwork methods (factor 5) obviously affect the
composition of available data, too. Apparently, different
data is usually obtained by field walking (for example,
settlements and shallow cremation graves will be
registered more likely than deeper inhumations) and
large-scale excavations preceded by a coarse topsoil
removing (in which case the resulting image may be
complementary to the field walking evidence). Since
cultural habits differed in various periods of the past (for
example, the preference of inhumation or cremation),
differing methods of fieldwork enable us to capture
sites from various periods with varying probability.
Apart from the methods of fieldwork, our data is also
influenced by the intensity of fieldwork. In the case of
Bohemia, for example, the question remains unclear of
whether the concentration of sites of all periods of the
past around Prague (Figure 1) is rather the result of the
intensity of settlement of this favourable environment
of Central Bohemia in the past, or whether it is the
consequence of the later status of Prague as the
capital town leading to more intensive fieldwork in
the last century, more intensive building activities
(and the following rescue excavations), a higher
number of archaeologists (leading to a larger share of

Figure 1. Map of Bohemia showing the summary of the archaeological record from the Palaeolithic up to the Early Middle Ages. The colour
scale differentiates the number of the main archaeological periods (maximum = 10) recorded per one square of 3000 × 3000m. The source map
has been submitted to a low-pass filtering in 3x3 cells kernel. The position of Prague is marked by a white cicrcle.

31
 Martin Kuna
recorded information) and/or the lower willingness
of archaeologists from more remote regions to submit
information to central databases (latent tensions
between the capital town and regions may have
penetrated many layers of social life in the CR; cf. Kuna
2015).
The core of my message, however, is to point out the
importance of the last mentioned factor. The varying
archaeological visibility of the remains of various
periods of the past (factor 6) afflicts the probability to
record a site by fieldwork and may probably account
for the largest distortion of the available data sets. The
differing archaeological visibility of sites represents a
specific consequence of such properties of artefacts and
activity areas, which are responsible for the number,
state of preservation and probability of identification
of archaeological remains by fieldwork. This factor was
also mentioned by Green et al. (2017, pp. 256–62), who
referred to ‘differences in past settlement type and use
of material culture.’
The understanding of these factors is, among other
things, hindered by the complexity of their operation.
Actually, the archaeological record (data) is rarely
influenced by just one of the factors of the second
group, but rather by a specific combination of all of
them. For example: prehistoric burial mounds covering
shallow cremation graves (a typical expression of the
Middle Bronze Age in Bohemia) are preserved and
easily detectable in traditional forest areas, while on
intensively cultivated land they were destroyed without
any evidence long before archaeological recording
started. ‘Flat’ inhumation burials with deeper grave
pits (e.g. those typical of the La Tène period) are the
opposite case: they are often being found in agricultural
and industrial areas with frequent subsurface
interventions, while mostly missed in forests with little
building activity and few interventions into the terrain.
Hence the existing differences in archaeological data
in different regions may be caused by a joint effect
of various factors: specific properties of artefacts
(shallow vs deep graves), different post-depositional
processes (preserving vs removing past surfaces) and
different fieldwork methods (excavations vs survey).
In other words, post-depositional processes and
fieldwork methods amplify the effects of the original
artefact (site) variability and produce archaeological
records where individual periods and regions may be
represented in a very unequal way.
Indications to be worried about
Varying formal properties of past artefacts and
activities are always at the beginning of the process
of ‘archaeological visibility differentiation’. Such
properties, however, did not originally mean any
32
substantially functional difference in the past culture
adequate to the effect they set up in the archaeological
record. On the contrary, even marginal functional
differences in the original culture may very often have
had decisive impacts on the archaeological record
composition (quantity). In the following, we will
discuss several cases of characteristic data patterning
in the archaeological record from Bohemia exactly
pointing to the varying archaeological visibility of some
specific types of data and warning us of being too much
confident to the data sets being currently at hand.
Residual finds
In large excavations, archaeologists often concentrate
on the interpretation and dating of larger features
(houses, storage pits, graves etc.) while analysing
movable finds not of its own but only in relation to this
goal. Therefore, very little attention is usually paid,
for example, to the individual dating of thousands of
common pottery fragments. A certain share of these
artefacts need not, however, necessarily have any
original (functional, chronological) relation to the
context where they were found in; it may be residual
items or later intrusions. If carried out, analyses of
these finds usually surprise us with evidence of further,
otherwise non-attested archaeological components
(site-phases).
The excavations at Roztoky (Central Bohemia), for
example, uncovered tens of larger sunken features of
seven archaeological cultures (from the Neolithic to
the Early Middle Ages) on an area of about 6,000 m2. A
detailed analysis of the content of 35 features of the Final
Bronze Age (26,000 pottery fragments; Kuna 2002; Kuna
– Němcová et al. 2012) yielded pottery fragments from
nine more archaeological periods that were otherwise
not attested at the site by any of the larger structures!
Often it was only individual pottery fragments among
hundreds of sherds from the ‘main’ periods; their
dating, however, was unequivocal. Obviously, these
were remains of settlement phases, the evidence of
which has not been ‘properly’ preserved because, in
these periods, no sunken features would have been
dug, and the remains of activities disintegrated on the
surface and/or were washed away by erosion (processes
of this kind can, in fact, be assumed on most prehistoric
and early medieval sites in Central Europe).
This would not necessarily affect the composition of
archaeological data in a fatal way, if the discrepancy
between the evidence conveyed by sunken features
and residual finds were only coincidental: one site
would show sunken features of only certain periods,
while on another site these would be missing. This,
however, is not the case, since some archaeological
periods are generally (non-randomly) characterized
 Deconstructing archaeological databases

by the absence or a small number of sunken features; Field walking evidence

therefore, the missing evidence is reoccurring and
systemic.
Even then, the distortion of our data would not need
to be so great, if residual finds were systematically
analysed and registered. In most cases, however, this
is not the case. The amount of pottery fragments in
cultural layers and sunken features of prehistoric and
early medieval sites tends to be large in Central Europe
(tens to thousands of pieces per cubic metre; see Kuna
1991) and their detailed analysis is lengthy, expensive
and demanding special skills. It is therefore normal to
leave out these data from excavation reports, which
necessarily leads to a systematic neglect of a large
portion of relevant information and a proportional
misrepresentation of some parts of the record.
A very similar message is contained in the results of
larger field walking projects. For example, in the course
of the ALRB project, about 6,000 sectors (one-hectare
squares) were walked in Central Bohemia, a territory
with permanent occupation since the Neolithic
(Zvelebil et al. 1993; Kuna 1998; 2001). This area yielded
more than 30,000 fragments of prehistoric pottery, but
the representation of individual periods within this
set was very unbalanced. While, for example, the La
Tène period was recorded in 248 sectors, the Migration
Period only in two and similarly low was the occurrence
of several other periods (Table 1).
Obviously, this image is inaccurate from the point of
view of past reality. This may have several causes, such

Table 1. Number of finds of prehistoric pottery and stone tools from the ALRB landscape survey project in Central Bohemia (cf. Zvelebil et al.
1993; Kuna 2001). Finds of the same category in one traverse of field walking survey are understood as one ‘event’ since their origin from one
and the same sunken feature cannot be excluded.

NUMBER OF
FIND CATEGORY DATE
SECTORS ‘EVENTS’ FINDS
LINEAR POTTERY C. 5600-4900 35 41 67
STROKE POTTERY C. (LATE NEOLITHIC) 4900-4200 26 42 60

Neolithic unsp. 87 129 214

PROTO-ENEOLITHIC 4200-3800 3 3 4
EARLY ENEOLITHIC 3800-3400 31 35 37
MIDDLE ENEOLITHIC 3400-2800 5 5 8
LATE ENEOLITHIC (CORDED WARE & BELL BEAKERS) 2800-2300 12 12 12

Eneolithic unsp. 65 78 90

EARLY BRONZE AGE 2300-1600 30 41 54

MIDDLE BRONZE AGE 1600-1300 17 17 20

Early to Middle Bronze Age 74 109 155

LATE BRONZE AGE 1300-1000 117 174 296

FINAL BRONZE AGE 1000-800 127 210 407

Late to Final Bronze Age 198 384 703

HALLSTATT P. 800-400 106 183 311

Bronze Age to Hallstatt P. unsp. 371 866 1559

LA TÈNE P. 400-0 BC 248 421 798

ROMAN P. 0-400 AD 81 146 374
MIGRATION P. 400-600 2 2 7

prehistory unsp. -- -- 25166

PREHISTORY TOTAL 2106 4764 30342

33
 Martin Kuna
as the different amounts of pottery in the original
activity areas (which seems, except for the beginning
of the Early Middle Ages, less probable) and different
amounts of distinctive pottery attributes (e.g. the La
Tène pottery can be distinguished by technology, visible
on all vessel parts; pottery of other periods, however,
can be identified mostly by decoration, which may be
restricted to small parts of the vessels and, therefore,
visible only on a small number of fragments).
The strongest factor, in my view, is the original
number and size of sunken features in the settlements
influencing the quantity of pottery fragments in
surface scatters. What I believe is that the surface find
scatters contain only those pottery fragments that were
recently ploughed out from sunken features and layers.
Pottery fragments (at least what prehistory concerns)
that were originally left on the surface have decayed
long ago, disintegrated by ploughing and weathering
or fallen prey to erosion.
Frequency of pit digging
Both above-mentioned examples point to the importance
of the same cultural trait: the occurrence (density)
of sunken features on prehistoric and early medieval
sites. A high degree of variability of this kind among
individual periods of the past can be proven directly,
for example, on the results of large-scale excavations.
It is obvious that some periods are characterized by a
large number of sunken houses, workshops, storage
pits, clay pits etc., while sites from other periods show
only few such features. The extreme case is represented
by the Late Eneolithic Corded Ware culture without any
single known settlement feature at all from Bohemia!
Were it not for the cemeteries, Bohemia would appear
as being unoccupied at the end of the Eneolithic. But
it is the cemeteries that witness the contrary, a dense
settlement network, well comparable with the situation
in other parts of prehistory.
The reasons for sunken settlement features not being
built in some periods do not have to be based on
economics or the environment; they may have been
purely symbolic. According to E. Neustupný (1997),
the vertical aspect was loaded with heavy symbolic
connotations in prehistoric cultures, which in some
periods may have prevented people not only from
digging settlement pits, but even post holes for above-
ground buildings. Therefore, the only occasion when
it seems to have been appropriate in the Corded Ware
culture to intervene into the ground was burying.
The opposite may have been the case in some other
periods of prehistory. There are considerable non-
random differences in this cultural trait not only
between various cultures, but also between their
34
individual phases. For example, of the Knovíz-culture
settlement in Kněževes 22 pits fall into the ‘Knovíz 2’
phase, while phase ‘Knovíz 4b’ is marked by 122 pits,
although both phases lasted for approximately the
same time and occupied almost the same area (with
allegedly the same number of households; Kuna –
Němcová et al. 2012, p. 219). Although the size of the
original community most probably was the same, the
number of archaeological remains differs considerably.
Anyway, this fact has a deep impact on the probability
of the site discovery by excavations or field surveys.
Most of the several thousands of rescue excavations
carried out in Bohemia every year consist of small (or
longer but linear) trenches. The probability of finding
a site by a small excavation directly corresponds to the
number of the features that can be hit by the trench.
High numbers of sunken features in some prehistoric
periods not necessarily have to represent a very
peculiar cultural trait. They can be just the consequence
of the habit to renew storage pits more often, perhaps
annually (out of ritual reasons?), and therefore to
produce a large number of them during the lifespan of
a settlement. In other periods or phases, the pits may
have been in use for several years or even replaced by
some above-ground storage facilities.
The number of sunken features on a site also varies
among different regions. In this case, we may consider
the influence of variously attractive subsoils. For
example, South Bohemia generally shows fewer pits:
this may be connected with less favourable properties
of the local subsoils (T. Šálková, pers. comm.).
Nonetheless, a larger number of features per site
influence the probability of their recording, both in the
case of field walking and smaller excavations.
Selective activity representation
Any larger set of archaeological data from Bohemia
shows another suspicious fact: a rather varied
representation of different kinds of past activities. In
Figure 2, we see the number of fieldwork events in the
AMCR registering residential and domestic production
features (houses, pits, ovens etc.) on the one hand and
funeral structures (graves and mounds) on the other.
Even though there are many entries, in which their
authors did not refer to any specific type of activity, it is
improbable that the sample would not represent a kind
of non-random variability.
These differences are really remarkable. While finds
with a domestic function come in the majority from
the period from the Neolithic to the Middle Eneolithic,
the ratio between the residential and funerary remains
in the Beaker cultures from the Late Eneolithic is
reversed. The following Bronze Age (with the exception
 Deconstructing archaeological databases

of the Final Bronze Age) is

characterized by a relatively
balanced share of both types of
evidence, but in the following
periods the share of both types
of activities is varied again.

Since – according to a well-

known archaeological quip – it
is impossible for humans to have
only lived in some periods and
only died in the others, we have
to admit that different societies
built their features for living and
burying in such a way that they
differ in their archaeological
visibility. In theory, there may
be a reason for their reciprocal
manifestation (cf. the opposition
of domus vs agrios mentioned
by Hodder 1990), but we are Figure 2. Frequency of archaeological events registering settlements (residential and domestic
still lacking clear arguments production activities) vs. burial activities in Bohemia by archaeological periods. After AMCR (2011).
for such an interpretation. It
rather seems that apart from
periods characterized by a good
visibility of one or the other
type of activity, there are periods
with a good visibility of both
(Early and Final Bronze Age,
Hallstatt period). This, however,
makes us suspicious: could in
such a case there have been
periods characterized by a bad
visibility of both basic activity
types (e.g. Early Eneolithic, Early
Middle Ages 1–2)? If this is the
case, is it possible to take the
numbers of sites in available data
sources as a reliable base for the
reconstruction of demographic
trends?

Regional variability of the

archaeological record
Figure 3. Frequency of archaeological events registering main types of archaeological features
The situation is even more from all prehistory and early Middle Ages. Split by the main geographical regions of Bohemia.
complicated. The composition After AMCR (2011).
of archaeological data sources
differs not only chronologically, but also regionally. The these categories can be read as follows, namely, from
graph in Figure 3 summarizes the AMCR entries from the view of factors, which are here discussed:
the whole prehistory and Early Middle Ages. We can see
that individual parts of Bohemia differ not only in the • Central Bohemia: a region intensively cultivated
total amount of sites, but also in the shares of the basic on the long term, with a long tradition of
sorts of archaeological remains (the graph differentiates archaeological fieldwork and systematic
between four categories of features: settlement registering of archaeological finds (hence high
pits, ‘flat’ inhumation graves, cremation graves and amount of entries); research conducted by
barrows). The varied composition and total figures in various methods including field walking;

35
Another random document with
no related content on Scribd:
 The drive block or the one-on-one block was developed primarily
for a fast hitting attack, and is one that we use on many occasions.
All blocks consist of the following steps: (1) the approach, (2) the
contact, and (3) the follow through.

Approach:

1. Approach with short, controlled step. First step should be with

the outside foot, and it should not go beyond the forward foot. The
knees should be bent at this time.
2. Split the opponent’s course with the first step, and aim the nose
guard at the middle of the opponent’s target.
3. Keep the eyes on the target at all times.
4. Uncoil off the forward foot. Do not duck the head during the
uncoil.

Contact:

1. Strike a jarring blow with the forehead.

2. Bring the head up as contact is made. (The body will follow the
head; this will help arch the back.)
3. The drive should be forward, then upward. The object is to
knock the opponent off-balance, thus depriving him of his traction;
this will make him easy to move out of the play.
4. Bring the feet up fast with short, choppy, digging steps, but still
keeping the knees bent.

Follow Through:

1. Stay on feet with the knees still bent and keep driving.
2. Keep the feet well spread and under the body.
3. Charge through the opponent.
 4. Keep the body between the opponent and the path of the ball
carrier.
The most important single step in the drive block is keeping the
feet driving the instant contact is made.

The Reverse Shoulder Block

This block is used primarily when the blocker already has position
on a hard charging lineman.

Approach:

1. Aim the far shoulder for the opponent’s far hip. This will help
allow for the opponent’s charge.
2. Shoot the head and shoulders in front of the defensive man.
3. This movement must be sharp and on a straight line.
4. Take a short positive step with the inside foot.

Contact:

1. Uncoil off the near foot.

2. Strike a good blow on the opponent’s hip, squeezing with the
neck and head in the stomach.
3. Bring inside foot up to help cut off penetration.

Follow Through:

1. Bring the legs up fast and keep them driving.

2. The charge of the defensive man will put the blocker in a perfect
blocking position.
3. Keep the body weight into opponent.
 4. If and when the opponent attempts to spin out, go into a crab
position around the outside thigh to prevent pursuit.

Blocking Linebackers

The percentage of time the offense can block the linebackers will
determine to a great extent how successful the offense will be. If the
linebackers are blocked, the defense will usually break down.

Approach:

1. Get off on snap of the ball.

2. Whenever possible release to the inside to block an inside
linebacker; to the outside when blocking an outside linebacker.
3. Stay low, don’t run in a circle, but take the shortest route to the
linebacker.
4. Get to the point of attack in position to block before the
linebacker can get there to defend it.
5. Keep your body between the linebacker and the ball carrier.

Contact:

1. Split the linebacker right down the middle with initial block.
2. Thrust the body forward and upward, hit the linebacker on the
rise and elevate the shoulders on contact.
3. Strike the blow with a low fundamental position, hit with
forehead and forearm.
4. Make contact off top of the feet. Keep the feet underneath the
blocker, thus making it impossible for the blocker to follow-through
properly if he lunges or leaves his feet before or while making
contact.

Follow Through:
 1. As soon as contact is made with the linebacker, continue the leg
drive. The initial blow should knock him off balance and the leg drive
should keep him this way.
2. Keep the weight into the linebacker.
3. Gain position follow through. The blocker should continue to
strive for position to eliminate any possibility of the linebacker being
in on the tackle downfield.

Downfield Blocking

The approach of the downfield block will vary with the play called
and position of the defensive secondary men. The approach we will
discuss will be centered around attacking the defensive halfback.
The downfield block is a definite characteristic of all good football
teams, and it is virtually impossible to have many long runs without
good downfield blocking.

Approach:

1. Release rapidly on the line of scrimmage.

2. Sprint shallowly just beyond the line of scrimmage.
3. When approaching the designated person to block, aim for a
spot in front of his original position; this eliminates running behind
the halfback when he comes up to make the tackle.
4. The blocker must have a clear concept of the point of attack and
the prescribed path of the ball carrier.
5. The blocker should make the defensive man commit himself
before he starts to make contact.

Contact:

1. Once the defensive man has committed himself in a specific

direction, the blocker then starts his contact.
 2. The blocker will get as close as possible to the defensive man
(about two yards) and spring off his inside foot, thus swinging his
outside elbow and arm toward the head of the opponent. This will
cause a lifting motion and make it possible for the blocker to extend
the entire length of his body in a horizontal position.
3. The blocker must aim his body at the opponent’s throat, trying
to make contact with his hip.
4. The defensive man will not allow him to hit him in the throat, but
by aiming there he may hit him around the knees.

Follow Through:

1. The follow through will consist of keeping the body extended as

long as possible and jumping up and trying to make another block.

The Post-Lead Block

The post-lead block, or two-on-one block, is used to insure the

blocking of the most dangerous defensive man, and at the same
time give the ball carrier the feeling of confidence that he can
concentrate his efforts on whipping the one defensive man to the
inside or outside of the man being double-teamed. With this in mind,
the blockers are more conscious of turning the defensive man away
from the play rather than drive him straight down the field. By using
this same method, it makes it possible to cut off the pursuit.

Approach:

(Post Man)

1. Drive straight at the defensive man with a short controlled step.

First step is with the outside foot.
2. Aim the nose guard at the middle of the target.
3. Keep the eyes on the target at all times.
 4. Uncoil off the forward foot.

(Drive Man)

1. Take a position step with the inside foot and then a good step
with the outside foot to get squared away.
2. From this new position aim the nose guard at the middle of the
opponent and use as an apex. From this position the middle of the
opponent would be below the armpit and just above the waist.
3. Keep eyes on target all the way.

Contact:

(Post Man)

1. Strike a good blow with the forehead.

2. Drive straight through the opponent to stop his charge.
3. Hit on the rise and keep the pressure applied, keeping the
knees bent at all times.

(Drive Man)

1. Strike a jarring blow with the forehead below the armpit of the
opponent.
2. After good contact, slide head past body and continue applied
pressure with the inside shoulder.
3. Keep the tail down, head up, and feet under the body.

Follow Through:

(Post Man)

1. After good contact and feet are well up under the body, turn the
tail toward the drive man.
 2. Keep pressure applied, feet moving in short, choppy steps, and
drive man down the line.

(Drive Man)

1. Keep pressure with shoulder and neck, and do not let man spin
out to outside.
2. Keep the tail to inside and do not allow defensive man to split.
3. Keep feet moving in short, choppy steps and move man down
the line.
4. Force defensive man into pursuit to help eliminate it.

The Trap Block

Approach:

1. It is important to take the same stance as if going straight

ahead, but just before the ball is snapped, the weight is shifted to the
outside foot.
2. Take a quick, short, 6-inch step with inside foot and point toe
directly toward spot to be trapped.
3. Never raise the body up. Stay down and in a semicoiled
position.
4. Drive off the forward foot, still taking short steps.
5. Anticipate the person to be trapped, to be filling the inside or
floating in the hole.

Contact: (see below)

Follow Through:

The contact and follow through are exactly the same as on our
drive block, which I explained previously.
 The Pass Protection (Drop back pass)

This technique is used by the guards and tackles when protecting

the passer on a straight dropback pass. It is good to go from a pre-
shift position when using this type of block.

Approach:

1. Take a short step backwards with the inside foot, then a longer
step backwards with the other foot.
2. While doing this, turn the tail slightly to the inside and remember
the inside gap must be protected.
3. After the second step backwards, the blocker should be in a
good football position (tail down, head up, knees bent, back straight,
and the body in a cocked position ready to strike a blow).

Contact:

1. Look at the man to be blocked, make him come to the blocker.

When close enough, spring and butt him with the forehead, trying to
make contact with him at the numbers.
2. After contact is made and the charge of the offensive man has
been stopped, the blocker will take a step back, regain his football
position and make contact again.
3. At all times the blocker will be forcing the defensive man to the
outside.

Follow Through:

1. When the blocker sees that he can no longer strike a blow with
his head, due to the position of the defensive man, he will fake a
blow with his head, then slide it and his body in front of the defensive
man.
 2. He will lock him with his leg and at the same time apply
pressure with his body.
3. At all times the blocker will be in a crab position and working his
feet around to force the defensive man to the outside, thus keeping
him away from the passer.

The Crack Back Block

This block is used primarily by the ends when they are blocking an
inside linebacker. It can also be used by anyone who might be
blocking directly in front of the ball carrier.

Approach:

1. Release from line of scrimmage as quickly as possible and run

straight at linebacker.
2. Aim slightly above the knees of the linebacker.
3. Get within a yard and a half before going into block.

Contact:

1. Throw outside arm with a quick, jerky motion in front of

linebacker. This quick motion will cause the head and shoulder to get
in front of defensive man.
2. Release off of inside foot to start block.
3. Make good solid contact with the outside hip just above the
defensive man’s knees.

Follow Through:

1. After contact is made, go into a four-point position with the arms

extended, and tail high into the air.
 2. Work the feet around and force the defensive man to go in front
of blocker’s head.
3. Do not let defensive man get out behind the blocker’s feet
because this will allow him to pursue the ball from a proper angle.
4. Keep tail up and feet working.

The Junction Block

This block is used primarily by the fullback when he is blocking a

true end or a corner man on a wide play.

Approach:

1. Aim for a spot about a yard in front and outside of man to be

blocked.
2. Run low and hard at spot.

Contact:

1. Extend the head and shoulders in front of and past the

defensive man. (The blocker’s head is pointing downfield.)
2. The blocker will use his inside leg and hip to catch the outside
leg and hip of the defensive man.
3. The blocker will then have his head and shoulders past the
defensive man and his inside leg firmly against the outside leg of the
defensive man, with his hands down and his tail up.

Follow Through:

1. Force defensive man to go inside by keeping feet moving and

driving downfield.
2. Do not extend body and fall flat on the ground.
3. Do not let man spin to blocker’s outside.
 4. Keep pressure on defensive man as long as possible.

The Roll Block

This block is used when a small back, in particular, is trying to

block a true end or corner man on the ground.

Approach:

1. Aim for a spot about a yard in front and outside of the man to be
blocked.
2. Run low and hard at spot.

Contact:

1. With the head pointed at the outside hip of the defensive man,
the blocker will get about a yard and a half from him. As the blocker
springs, he will drop his inside shoulder and turn his head down and
to the outside.
2. The blocker will hit the outside leg of the defensive man with the
part of his shoulder pad that covers the scapula of his inside
shoulder.

Follow Through:

1. After contact has been made with the back of shoulder, the
blocker will roll to his inside and give a corkscrew effect.
2. Stay as high as possible (around the knees) and make two or
three complete rolls in the corkscrew fashion, trying to knock the legs
out from under the defensive man.
3. At all times the blocker must keep his head pointing downfield
or no less than a 60 degree angle.

NUMBERING THE DEFENSIVE ALIGNMENTS

 Figure 102a

In making our blocking rules or assignments for all of the players

versus all different defensive alignments, there are several factors to
take into consideration. The rules must be simple, and secondly,
they must be brief. We have used several different kinds of rules,
such as, “Inside gap, over, linebacker,” and others. This was a good
method, but it amounted to quite a bit of memory work for the players
because the majority of the blocks were all different and the players
were required to learn a number of different sequences. Trying to
adhere to the theory, “the simpler, the better,” we started numbering
the defensive men as illustrated in Figure 102a versus the 5-4
defensive alignment; 102b versus the wide tackle 6; 102c versus the
gap 8 defense; and Figure 102d versus the Eagle defense.

Figure 102b
 Figure 102c

Figure 102d

We start counting with the man over the center and number him
zero, and from there go both left and right numbering every man
within two yards of the line of scrimmage, as illustrated in Figures
102a-b-c-d. We also number the men in the secondary. We do this
by merely continuing with our numbering beyond the end lineman in
the direction the ball is going to go, as illustrated in Figure 103a
versus the Oklahoma 5-4 defense, and Figure 103b versus the wide
tackle 6-2 defensive alignment.

Figure 103a
 Figure 103b

Application of Our Blocking Rules

By numbering the defensive men it is simple when making out the

blocking rules to assign each man a number, and the offensive man
merely learns which numbered man he will block. There will be
certain plays in which a particular position must have one or two
options. As an example, a rule might read, “#3 unless outside, then,
#2.”
The following are examples of blocking rules by numbers:

Regular Block—straight ahead

On-E #3
On-T #2
On-G #1
C #0, off-side
Off-G #1
Off-T #5
Off-E #6

Trap Block—
 On-E #3
On-T L.B.
On-G Slam man #1 or #2, N/T #0
C Off-side
Off-G Trap
Off-T 1st outside guard
Off-E #6

Counter Block—

On-E #6
On-T #2
On-G #1
C #1, N/T #0, N/T off-side
Off-G #1
Off-T #2
Off-E #3

By using this particular method it is also very easy to incorporate

the backs into the blocking scheme. When using the back to the side
the ball is going, in order to have a backfield man block a particular
defensive player, we merely add two zeroes to the number which
has been assigned to the defender, and the back gets this man. As
an illustration, if we want our halfback to block the #3 defensive man,
the halfback’s block is 300; and the #4 man, the block would be
called 400. Several complete illustrations would be as follows:

300 Block—

Off-E #6
Off-T #5
Off-G #1
C #0, N/T off-side
On-G #1
On-T #2
On-E slam #3, then #5
H.B. #3

400 Block—

Off-E #6
Off-T #5
Off-G #1
C #0, N/T off-side
On-G #1
On-T #2
On-E #3
H.B. #4

Under Block—

Off-E #6
Off-T #5
Off-G #1
C #0, N/T off-side
On-G #1
On-T #2
 On-E #3
H.B. #4

Over Block—

Off-E #6
Off-T #5
Off-G #1
C #0, N/T off-side
On-G 1st man on line of scrimmage
On-T 2nd man on line of scrimmage
On-E 3rd man on line of scrimmage
H.B. 1st inside linebacker

ATTACKING THE DEFENSIVE ALIGNMENT

Probably like many other football teams, we have too much
offense. However, in order to do an intelligent job of planning our
attack versus the numerous defensive alignments, one must have
sufficient offense since not all plays are good against all defenses.
Regardless of one’s offense, the first approach is to establish a
good sound middle attack that is based on the trap, fullback hand-
off, and a pass off of the same action. Unless a team can force the
defense to respect the inside power and force the defensive
linebackers and guards to stay “at home,” it is almost an impossibility
to perfect the outside attack. After we establish the inside attack,
then we want to run far enough inside the defensive end to cut down
his quick containment. The next step then is to perfect the wide
attack and the corner passes, in order to have a well-rounded
offense. Bootleg passes, reverses, and an occasional trick play are
also needed in order to keep the defense “honest,” and to make the
above-mentioned plays more effective.
 GOING WIDE
It is an offensive must for a team to be able to go wide and to get
the long gainer. Occasionally all of us get a good gainer from the
inside attack, but most long gainers are from passes or some form of
wide attack.
Previously we operated on the assumption if we could gain four
yards on each play we would score with a sustained drive. Statistics
will prove a team will generally stop itself by some error, or the
defense will stop the offense, before the attack can make four
consecutive first downs or gain 50 yards, a majority of the time.
Consequently a team must perfect its wide game.
There are several ways of going wide, but regardless of the
manner attempted, the defensive end or corner man must be
eliminated either by blocking him, optioning him, or throwing the
football over his area.

The Run-Pass Option

When trying to get wide against a corner man, the run or pass
option is one of the best methods to employ. Figure 104 illustrates
the on-guard blocking #1, on-tackle #2, and the fullback blocking the
#3 man. The near halfback and the on-end will run their pass routes.
The #4 man is not blocked, and the passer is going to option from
him. As soon as the passer has possession of the ball, he watches
the #4 man. If he comes up to tackle the passer, the latter throws to
the open receiver. If the #4 man drops back to play pass defense,
the passer is instructed to turn upfield and run with the football. If our
passer feels he can gain at least four yards on the play, we want him
to run with it most of the time. The play is much better against a
team that uses a 9-man front than against an 8-man front alignment.
 Figure 104

The Quarterback Option

There are several kinds of option plays, but the one designed to
get wide is the two-way option, in which the quarterback either keeps
or pitches to the far halfback, depending upon the reaction of the
defensive man from whom he is optioning.
The blocking rules on the quarterback option play are very simple,
as illustrated in Figure 105. The offensive guard and tackle to the on-
side block their numbers, respectively. The on-end will block to his
inside, N/T (No One There) halfback, and the fullback will block the
#4 man with a junction block. The quarterback must watch his man
all the time in order for the play to turn the corner successfully. He
must option the #3 man and not permit the latter to option the
quarterback.
 Figure 105

The Three-Way Option

Figure 106

The other option play, illustrated in Figure 106, is the three-way

option with the quarterback playing the #4 defender. This option is in
the same series with the outside belly play. The quarterback has the
option of (1) giving the ball to the fullback, who will hit off-tackle, (2)
keeping the ball himself, after faking the fullback off-tackle, or (3)